Part III — Local Fields

Based on lectures by H. C. Johansson

Notes taken by Dexter Chua

Michaelmas 2016

These notes are not endorsed by the lecturers, and I have modified them (often
significantly) after lectures. They are nowhere near accurate representations of what
was actually lectured, and in particular, all errors are almost surely mine.

The p-adic numbers Qp (where p is any prime) were invented by Hensel in the late 19th
century, with a view to introduce function-theoretic methods into number theory. They
are formed by completing Q with respect to the p-adic absolute value | − |p , defined
for non-zero x ∈ Q by |x|p = p−n , where x = pn a/b with a, b, n ∈ Z and a and b are
coprime to p. The p-adic absolute value allows one to study congruences modulo all
powers of p simultaneously, using analytic methods. The concept of a local field is an
abstraction of the field Qp , and the theory involves an interesting blend of algebra and
analysis. Local fields provide a natural tool to attack many number-theoretic problems,
and they are ubiquitous in modern algebraic number theory and arithmetic geometry.

Topics likely to be covered include:

The p-adic numbers. Local fields and their structure.

Finite extensions, Galois theory and basic ramification theory.
Polynomial equations; Hensel’s Lemma, Newton polygons.
Continuous functions on the p-adic integers, Mahler’s Theorem.
Local class field theory (time permitting).

Pre-requisites

Basic algebra, including Galois theory, and basic concepts from point set topology
and metric spaces. Some prior exposure to number fields might be useful, but is not
essential.

1
Contents III Local Fields

Contents
0 Introduction 3

1 Basic theory 4
1.1 Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.2 Rings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3 Topological rings . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
1.4 The p-adic numbers . . . . . . . . . . . . . . . . . . . . . . . . . 11

2 Valued fields 15
2.1 Hensel’s lemma . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Extension of norms . . . . . . . . . . . . . . . . . . . . . . . . . . 18
2.3 Newton polygons . . . . . . . . . . . . . . . . . . . . . . . . . . . 22

3 Discretely valued fields 27

3.1 Teichmüller lifts . . . . . . . . . . . . . . . . . . . . . . . . . . . 30
3.2 Witt vectors* . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33

4 Some p-adic analysis 38

5 Ramification theory for local fields 43

5.1 Ramification index and inertia degree . . . . . . . . . . . . . . . 43
5.2 Unramified extensions . . . . . . . . . . . . . . . . . . . . . . . . 47
5.3 Totally ramified extensions . . . . . . . . . . . . . . . . . . . . . 49

6 Further ramification theory 51

6.1 Some filtrations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
6.2 Multiple extensions . . . . . . . . . . . . . . . . . . . . . . . . . . 55

7 Local class field theory 63

7.1 Infinite Galois theory . . . . . . . . . . . . . . . . . . . . . . . . . 63
7.2 Unramified extensions and Weil group . . . . . . . . . . . . . . . 65
7.3 Main theorems of local class field theory . . . . . . . . . . . . . . 68

8 Lubin–Tate theory 71
8.1 Motivating example . . . . . . . . . . . . . . . . . . . . . . . . . 71
8.2 Formal groups . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 75
8.3 Lubin–Tate extensions . . . . . . . . . . . . . . . . . . . . . . . . 79

Index 88

2
0 Introduction III Local Fields

0 Introduction
What are local fields? Suppose we are interested in some basic number theoretic
problem. Say we have a polynomial f (x1 , · · · , xn ) ∈ Z[x1 , · · · , xn ]. We want to
look for solutions a ∈ Zn , or show that there are no solutions at all. We might
try to view this polynomial as a real polynomial, look at its roots, and see if
they are integers. In lucky cases, we might be able to show that there are no
real solutions at all, and conclude that there cannot be any solutions at all.
On the other hand, we can try to look at it modulo some prime p. If there
are no solutions mod p, then there cannot be any solution. But sometimes p is
not enough. We might want to look at it mod p2 , or p3 , or . . . . One important
application of local fields is that we can package all these information together.
In this course, we are not going to study the number theoretic problems, but
just look at the properties of the local fields for their own sake.
Throughout this course, all rings will be commutative with unity, unless
otherwise specified.

3
1 Basic theory III Local Fields

1 Basic theory
We are going to start by making loads of definitions, which you may or may not
have seen before.

1.1 Fields
Definition (Absolute value). Let K be a field. An absolute value on K is a
function | · | : K → R≥0 such that
(i) |x| = 0 iff x = 0;

(ii) |xy| = |x||y| for all x, y ∈ K;

(iii) |x + y| ≤ |x| + |y|.
Definition (Valued field). A valued field is a field with an absolute value.

Example. The rationals, reals and complex numbers with the usual absolute
values are absolute values.
Example (Trivial absolute value). The trivial absolute value on a field K is the
absolute value given by (
1 x 6= 0
|x| = .
0 x=0

The only reason we mention the trivial absolute value here is that from
now on, we will assume that the absolute values are not trivial, because trivial
absolute values are boring and break things.
There are some familiar basic properties of the absolute value such as
Proposition. ||x| − |y|| ≤ |x − y|. Here the outer absolute value on the left
hand side is the usual absolute value of R, while the others are the absolute
values of the relevant field.
An absolute value defines a metric d(x, y) = |x − y| on K.
Definition (Equivalence of absolute values). Let K be a field, and let | · |, | · |0
be absolute values. We say they are equivalent if they induce the same topology.
Proposition. Let K be a field, and | · |, | · |0 be absolute values on K. Then
the following are equivalent.
(i) | · | and | · |0 are equivalent

(ii) |x| < 1 implies |x|0 < 1 for all x ∈ K

(iii) There is some s ∈ R>0 such that |x|s = |x|0 for all x ∈ K.
Proof. (i) ⇒ (ii) and (iii) ⇒ (i) are easy exercises. Assume (ii), and we shall
prove (iii). First observe that since |x−1 | = |x|−1 , we know |x| > 1 implies
|x|0 > 1, and hence |x| = 1 implies |x|0 = 1. To show (iii), we have to show that
log |x|
the ratio log |x0 | is independent of x.

4
1 Basic theory III Local Fields

Suppose not. We may assume

log |x| log |y|
0
< ,
log |x| log |y|0
and moreover the logarithms are positive. Then there are m, n ∈ Z>0 such that
log |x| m log |x|0
< < .
log |y| n log |y|0
Then rearranging implies  n  n 0
x 
 < 1 < x  ,
 

 ym   ym 
a contradiction.
Exercise. Let K be a valued field. Then equivalent absolute values induce the
same the completion K̂ of K, and K̂ is a valued field with an absolute value
extending | · |.
In this course, we are not going to be interested in the usual absolute values.
Instead, we are going to consider some really weird ones, namely non-archimedean
ones.
Definition (Non-archimedean absolute value). An absolute value | · | on a field
K is called non-archimedean if |x + y| ≤ max(|x|, |y|). This condition is called
the strong triangle inequality.
An absolute value which isn’t non-archimedean is called archimedean.
Metrics satisfying d(x, z) ≤ max(d(x, y), d(y, z)) are often known as ultra-
metrics.
Example. Q, R and C under the usual absolute values are archimedean.
In this course, we will only consider non-archimedean absolute values. Thus,
from now on, unless otherwise mentioned, an absolute value is assumed to be
non-archimedean. The metric is weird!
We start by proving some absurd properties of non-archimedean absolute
values.
Recall that the closed balls are defined by

B(x, r) = {y : |x − y| ≤ r}.

Proposition. Let (K, | · |) be a non-archimedean valued field, and let x ∈ K

and r ∈ R>0 . Let z ∈ B(x, r). Then

B(x, r) = B(z, r).

So closed balls do not have unique “centers”. Every point can be viewed as
the center.
Proof. Let y ∈ B(z, r). Then

|x − y| = |(x − z) + (z − y)| ≤ max(|x − z|, |z − y|) ≤ r.

So y ∈ B(x, r). By symmetry, y ∈ B(x, r) implies y ∈ B(z, r).

5
1 Basic theory III Local Fields

Corollary. Closed balls are open.

Proof. To show that B(x, r) is open, we let z ∈ B(x, r). Then we have

{y : |y − z| < r} ⊆ B(z, r) = B(x, r).

So we know the open ball of radius r around z is contained in B(x, r). So B(x, r)
is open.
Norms in non-archimedean valued fields are easy to compute:

Proposition. Let K be a non-archimedean valued field, and x, y ∈ K. If

|x| > |y|, then |x + y| = |x|. P
∞
More generally, if x = c=0 xi and the non-zero |xi | are distinct, then
|x| = max |xi |.

Proof. On the one hand, we have |x + y| ≤ max{|x|, |y|}. On the other hand,
we have
|x| = |(x + y) − y| ≤ max(|x + y|, |y|) = |x + y|,
since we know that we cannot have |x| ≤ |y|. So we must have |x| = |x + y|.
Convergence is also easy for valued fields.
Proposition. Let K be a valued field.
(i) Let (xn ) be a sequence in K. If xn − xn+1 → 0, then xn is Cauchy.

If we assume further that K is complete, then

(ii) Let (xn ) be a sequence in K. If xn − xn+1 → 0, then a sequence (xn ) in
K converges.
P∞ P∞
(iii) Let n=0 yn be a series in K. If yn → 0, then n=0 yn converges.

The converses to all these are of course also true, with the usual proofs.
Proof.
(i) Pick ε > 0 and N such that |xn − xn+1 | < ε for all n ≥ N . Then given
m ≥ n ≥ N , we have

|xm − xn | = |xm − xm−1 + xm−1 − xm−2 + · · · − xn |

≤ max(|xm − xm−1 |, · · · , |xn+1 − xn |)
< ε.

So the sequence is Cauchy.

(ii) Follows from (1) and the definition of completeness.

(iii) Follows from the definition of convergence of a series and (2).

The reason why we care about these weird non-archimedean fields is that
they have very rich algebraic structure. In particular, there is this notion of the
valuation ring.

6
1 Basic theory III Local Fields

Definition (Valuation ring). Let K be a valued field. Then the valuation ring
of K is the open subring
OK = {x : |x| ≤ 1}.
We prove that it is actually a ring
Proposition. Let K be a valued field. Then

OK = {x : |x| ≤ 1}

is an open subring of K. Moreover, for each r ∈ (0, 1], the subsets {x : |x| < r}
×
and {x : |x| ≤ r} are open ideals of OK . Moreover, OK = {x : |x| = 1}.
Note that this is very false for usual absolute values. For example, if we take
R with the usual absolute value, we have 1 ∈ OR , but 1 + 1 6∈ OR .
Proof. We know that these sets are open since all balls are open.
To see OK is a subring, we have |1| = |−1| = 1. So 1, −1 ∈ OK . If x, y ∈ OK ,
then |x + y| ≤ max(|x|, |y|) ≤ 1. So x + y ∈ OK . Also, |xy| = |x||y| ≤ 1 · 1 = 1.
So xy ∈ OK .
That the other sets are ideals of OK is checked in the same way.
×
To check the units, we have x ∈ OK ⇔ |x|, |x−1 | ≤ 1 ⇔ |x| = |x|−1 = 1.

1.2 Rings
Definition (Integral element). Let R ⊆ S be rings and s ∈ S. We say s is
integral over R if there is some monic f ∈ R[x] such that f (s) = 0.
Example. Any r ∈ R is integral (take f (x) = x − r).
Example. Take Z ⊆ C. Then z ∈ C is integral over Z if√it is an algebraic
integer (by definition of algebraic integer). For example, 2 is an algebraic
integer, but √12 is not.

We would like to prove the following characterization of integral elements:

Theorem. Let R ⊆ S be rings. Then s1 , · · · , sn ∈ S are all integral iff
R[s1 , · · · , sn ] ⊆ S is a finitely-generated R-module.
Note that R[s1 , · · · , sn ] is by definition a finitely-generated R-algebra, but
requiring it to be finitely-generated as a module is stronger.
Here one direction is easy. It is not hard to show that if s1 , · · · , sn are all
integral, then R[s1 , · · · , sn ] is finitely-generated. However to show the other
direction, we need to find some clever trick to produce a monic polynomial that
kills the si .
The trick we need is the adjugate matrix we know and love from IA Vectors
and Matrices.
Definition (Adjoint/Adjugate matrix). Let A = (aij ) be an n × n matrix with
coefficients in a ring R. The adjugate matrix or adjoint matrix A∗ = (a∗ij ) of A
is defined by
a∗ij = (−1)i+j det(Aij ),
where Aij is an (n − 1) × (n − 1) matrix obtained from A by deleting the ith
column and the jth row.

7
1 Basic theory III Local Fields

As we know from IA, the following property holds for the adjugate matrix:
Proposition. For any A, we have A∗ A = AA∗ = det(A)I, where I is the
identity matrix.
With this, we can prove our claim:
Proof of theorem. Note that we can construct R[s1 , · · · , sn ] by a sequence

R ⊆ R[s1 ] ⊆ R[s1 , s2 ] ⊆ · · · ⊆ R[s1 , · · · , sn ] ⊆ S,

and each si is integral over R[s1 , · · · , sn−1 ]. Since the finite extension of a finite
extension is still finite, it suffices to prove it for the case n = 1, and we write s
for s1 .
Suppose f (x) ∈ R[x] is monic such that f (s) = 0. If g(x) ∈ R[x], then there
is some q, r ∈ R[x] such that g(x) = f (x)q(x) + r(x) with deg r < deg f . Then
g(s) = r(s). So any polynomial expression in s can be written as a polynomial
expression with degree less than deg f . So R[s] is generated by 1, s, · · · , sdeg f −1 .
In the other direction, let t1 , · · · , td be R-module generators of R[s1 , · · · , sn ].
We show that in fact any element of R[s1 , · · · , sn ] is integral over R. Consider
any element b ∈ R[s1 , · · · , sn ]. Then there is some aij ∈ R such that
d
X
bti = aij tj .
j=1

In matrix form, this says

(bI − A)t = 0.
We now multiply by (bI − A)∗ to obtain

det(bI − A)tj = 0
P
for all j. Now we know 1 ∈ R. So 1 = cj tj for some cj ∈ R. Then we have
X X
det(bI − A) = det(bI − A) cj t j = cj (det(bI − A)tj ) = 0.

Since det(bI − A) is a monic polynomial in b, it follows that b is integral.

Using this characterization, the following result is obvious:

Corollary. Let R ⊆ S be rings. If s1 , s2 ∈ S are integral over R, then s1 + s2
and s1 s2 are integral over R. In particular, the set R̃ ⊆ S of all elements in S
integral over R is a ring, known as the integral closure of R in S.

Proof. If s1 , s2 are integral, then R[s1 , s2 ] is a finite extension over R. Since

s1 + s2 and s1 s2 are elements of R[s1 , s2 ], they are also integral over R.
Definition (Integrally closed). Given a ring extension R ⊆ S, we say R is
integrally closed in S if R̃ = R.

8
1 Basic theory III Local Fields

1.3 Topological rings

Recall that we previously constructed the valuation ring OK . Since the valued
field K itself has a topology, the valuation ring inherits a subspace topology.
This is in fact a ring topology.
Definition (Topological ring). Let R be a ring. A topology on R is called a
ring topology if addition and multiplication are continuous maps R × R → R. A
ring with a ring topology is a topological ring.

Example. R and C with the usual topologies and usual ring structures are
topological rings.
Exercise. Let K be a valued field. Then K is a topological ring. We can see
this from the fact that the product topology on K × K is induced by the metric
d((x0 , y0 ), (x1 , y1 )) = max(|x0 − x1 |, |y0 − y1 |).

Now if we are just randomly given a ring, there is a general way of constructing
a ring topology. The idea is that we pick an ideal I and declare its elements to
be small. For example, in a valued ring, we can pick I = {x ∈ OK : |x| < 1}.
Now if you are not only in I, but I 2 , then you are even smaller. So we have a
hierarchy of small sets
I ⊇ I2 ⊇ I3 ⊇ I4 ⊇ · · ·
Now to make this a topology on R, we say that a subset U ⊆ R is open if every
x ∈ U is contained in some translation of I n (for some n). In other words, we
need some y ∈ R such that
x ∈ y + I n ⊆ U.
But since I n is additively closed, this is equivalent to saying x + I n ⊆ U . So we
make the following definition:
Definition (I-adically open). Let R be a ring and I ⊆ R an ideal. A subset
U ⊆ R is called I-adically open if for all x ∈ U , there is some n ≥ 1 such that
x + In ⊆ U .

Proposition. The set of all I-adically open sets form a topology on R, called
the I-adic topology.
Note that the I-adic topology isn’t really the kind of topology we are used
to thinking about, just like the topology on a valued field is also very weird.
Instead, it is a “filter” for telling us how small things are.

Proof. By definition, we have ∅ and R are open, and arbitrary unions are clearly
open. If U, V are I-adically open, and x ∈ U ∩ V , then there are n, m such that
x + I n ⊆ U and x + I m ⊆ V . Then x + I max(m,n) ⊆ U ∩ V .
Exercise. Check that the I-adic topology is a ring topology.

In the special case where I = xR, we often call the I-adic topology the x-adic
topology.
Now we want to tackle the notion of completeness. We will consider the case
of I = xR for motivation, but the actual definition will be completely general.

9
1 Basic theory III Local Fields

If we pick the x-adic topology, then we are essentially declaring that we take
x to be small. So intuitively, we would expect power series like
a0 + a1 x + a2 x2 + a3 x3 + · · ·
to “converge”, at least if the ai are “of bounded size”. In general, the ai are
“not too big” if ai xi is genuinely a member of xi R, as opposed to some silly thing
like x−i .
As in the case of analysis, we would like to think of these infinite series as a
sequence of partial sums
(a0 , a0 + a1 x, a0 + a1 x + a2 x2 , · · · )
Now if we denote the limit as L, then we can think of this sequence alternatively
as
(L mod I, L mod I 2 , L mod I 3 , · · · ).
The key property of this sequence is that if we take L mod I k and reduce it mod
I k−1 , then we obtain L mod I k−1 .
In general, suppose we have a sequence
(bn ∈ R/I n )∞
n=1 .

such that bn mod I n−1 = bn−1 . Then we want to say that the ring is I-adically
complete if every sequence satisfying this property is actually of the form
(L mod I, L mod I 2 , L mod I 3 , · · · )
for some L. Alternatively, we can take the I-adic completion to be the collection
of all such sequences, and then a space is I-adically complete it is isomorphic to
its I-adic completion.
To do this, we need to build up some technical machinery. The kind of
sequences we’ve just mentioned is a special case of an inverse limit.
Definition (Inverse/projective limit). Let R1 , R2 , , · · · be topological rings, with
continuous homomorphisms fn : Rn+1 → Rn .

R1 f1
R2 f2
R3 f3
R4 ···

The inverse limit or projective limit of the Ri is the ring

( )
Y
lim Rn = (xn ) ∈ Rn : fn (xn+1 ) = xn ,
←−
n

with coordinate-wise addition and multiplication,

Q together with the subspace
topology coming from the product topology of Rn . This topology is known as
the inverse limit topology.
Proposition. The inverse limit topology is a ring topology.
Proof sketch. We can fit the addition and multiplication maps into diagrams

lim Rn × lim Rn lim Rn

←− ←− ←−

Q Q Q
Rn × Rn Rn

10
1 Basic theory III Local Fields

By the definition
Q of the subspace topology, it suffices to show that the correspond-
ing maps on Rn are continuous.Q By the Quniversal property of the product, it
suffices to show that the projects Rn × Rn → Rm is continuous for all m.
But this map can alternatively be obtained by first projecting to Rm , then doing
multiplication in Rm , and projection is continuous. So the result follows.
It is easy to see the following universal property of the inverse limit topology:
Proposition. Giving a continuous ring homomorphism g : S → lim Rn is the
←−
same as giving a continuous ring homomorphism gn : S → Rn for each n, such
that each of the following diagram commutes:
gn
S Rn
gn−1 fn−1

Rn−1

Definition (I-adic completion). Let R be a ring and I be an ideal. The I-adic

completion of R is the topological ring

lim R/I n ,
←−
where R/I n has the discrete topology, and R/I n+1 → R/I n is the quotient map.
There is an evident map

ν : R → lim R/I n
←−
r 7→ (r mod I n ).

This map is a continuous ring homomorphism if R is given the I-adic topology.

Definition (I-adically complete). We say that R is I-adically complete if ν is a
bijection.
Exercise. If ν is a bijection, then ν is in fact a homeomorphism.

1.4 The p-adic numbers

For the rest of this course, p is going to be a prime number.
We consider a particular case of valued fields, namely the p-adic numbers,
and study some of its basic properties.
Let x ∈ Q be non-zero. Then by uniqueness of factorization, we can write x
uniquely as
a
x = pn ,
b
where a, b, n ∈ Z, b > 0 and a, b, p are pairwise coprime.
Definition (p-adic absolute value). The p-adic absolute value on Q is the
function | · |p : Q → R≥0 given by
(
0 x=0
|x|p = .
p−n x = pn ab as above

Proposition. The p-adic absolute value is an absolute value.

11
1 Basic theory III Local Fields

Proof. It is clear that |x|p = 0 iff x = 0.

Suppose we have
a c
x = pn , y = pm .
b d
We wlog m ≥ n. Then we have
 ac 
|xy|p = pn+m  = p−m−n = |x|p |y|p .

bd
So this is multiplicative. Finally, we have
 ab + pm−n cb 
 
|x + y|p = pn  ≤ p−n = max(|x|p , |y|p ).
bd 

Note that we must have bd coprime to p, but ab + pm−n cb need not be. However,
any extra powers of p could only decrease the absolute value, hence the above
result.
Note that if x ∈ Z is an integer, then |x|p = p−n iff pn || x (we say pn || x if
p | x and pn+1 - x).
n

Definition (p-adic numbers). The p-adic numbers Qp is the completion of Q

with respect to | · |p .

Definition (p-adic integers). The valuation ring

Zp = {x ∈ Qp : |x|p ≤ 1}

is the p-adic integers.

Proposition. Zp is the closure of Z inside Qp .

Proof. If x ∈ Z is non-zero, then x = pn a with n ≥ 0. So |x|p ≤ 1. So Z ⊆ Zp .
We now want to show that Z is dense in Zp . We know the set

Z(p) = {x ∈ Q : |x|p ≤ 1}

is dense inside Zp , essentially by definition. So it suffices to show that Z is dense

in Z(p) . We let x ∈ Z(p) \ {0}, say
a
x = pn , n ≥ 0.
b
It suffices to find xi ∈ Z such that xi → 1b . Then we have pn axi → x.
Since (b, p) = 1, we can find xi , yi ∈ Z such that bxi + pi yi = 1 for all i ≥ 1.
So    
xi − 1  =  1  |bxi − 1|p = |pi yi |p ≤ p−i → 0.
   
 b p  b p


So done.
Proposition. The non-zero ideals of Zp are pn Zp for n ≥ 0. Moreover,

Z ∼ Zp
= n .
pn Z p Zp

12
1 Basic theory III Local Fields

Proof. Let 0 6= I ⊆ Zp be an ideal, and pick x ∈ I such that |x|p is maximal.

This supremum exists and is attained because the possible values of the absolute
values are discrete and bounded above. If y ∈ I, then by maximality, we have
|y|p ≤ |x|p . So we have |yx−1 |p ≤ 1. So yx−1 ∈ Zp , and this implies that
y = (yx−1 )x ∈ xZp . So I ⊆ xZp , and we obviously have xZp ⊆ I. So we have
I = xZp .
Now if x = pn ab , then since ab is invertible in Zp , we have xZp = pn Zp . So
I = pn Zp .
To show the second part, consider the map
Zp
fn : Z → n
p Zp

given by the inclusion map followed by quotienting. Now pn Zp = {x : |x|p ≤ p−n .

So we have
ker fn = {x ∈ Z : |x|p ≤ p−n } = pn Z.
Now since Z is dense in Zp , we know the image of fn is dense in Zp /pn Zp .
But Zp /pn Zp has the discrete topology. So fn is surjective. So fn induces an
isomorphism Z/pn Z ∼= Zp /pn Zp .
Corollary. Zp is a PID with a unique prime element p (up to units).
This is pretty much the point of the p-adic numbers — there are a lot of
primes in Z, and by passing on to Zp , we are left with just one of them.
Proposition. The topology on Z induced by | · |p is the p-adic topology (i.e.
the pZ-adic topology).
Proof. Let U ⊆ Z. By definition, U is open wrt | · |p iff for all x ∈ U , there is
an n ∈ N such that
{y ∈ Z : |y − x|p ≤ p−n } ⊆ U.
On the other hand, U is open in the p-adic topology iff for all x ∈ U , there is
some n ≥ 0 such that x + pn Z ⊆ U . But we have

{y ∈ Z : |y − x|p ≤ p−n } = x + pn Z.

So done.
Proposition. Zp is p-adically complete and is (isomorphic to) the p-adic com-
pletion of Z.
Proof. The second part follows from the first as follows: we have the maps

ν (fn )n
Zp lim Zp /(pn Zp ) lim Z/(pn Z)
←−

We know the map induced by (fn )n is an isomorphism. So we just have to show

that ν is an isomorphism
To prove the first part, we have x ∈ ker ν iff x ∈ pn Zp for all n iff |x|p ≤ p−n
for all n iff |x|p = 0 iff x = 0. So the map is injective.
To show surjectivity, we let

zn ∈ lim Zp /pn Zp .
←−

13
1 Basic theory III Local Fields

We define ai ∈ {0, 1, · · · , p − 1} recursively such that

n−1
X
xn = ai pi
i=0

is the unique representative of zn in the set of integers {0, 1, · · · , pn − 1}. Then

∞
X
x= ai pi
i=0

exists in Zp and maps to x ≡ xn ≡ zn (mod pn ) for all n ≥ 0. So ν(x) = (zn ).

So the map is surjective. So ν is bijective.
Corollary. Every a ∈ Zp has a unique expansion
∞
X
a= ai pi .
i=0

with ai ∈ {0, · · · , p − 1}.

More generally, for any a ∈ Q× , there is a unique expansion
∞
X
a= ai pi
i=n

for ai ∈ {0, · · · , p − 1}, an 6= 0 and

n = − logp |a|p ∈ Z.

Proof. The second part follows from the first part by multiplying a by p−n .

Example. We have
1
= 1 + p + p2 + p3 + · · · .
1−p

14
2 Valued fields III Local Fields

2 Valued fields
2.1 Hensel’s lemma
We return to the discussion of general valued fields. We are now going to introduce
an alternative to the absolute value that contains the same information, but is
presented differently.
Definition (Valuation). Let K be a field. A valuation on K is a function
v : K → R ∪ {∞} such that
(i) v(x) = 0 iff x = 0
(ii) v(xy) = v(x) + v(y)
(iii) v(x + y) ≥ min{v(x), v(y)}.
Here we use the conventions that r + ∞ = ∞ and r ≤ ∞ for all r ∈ ∞.
In some sense, this definition is sort-of pointless, since if v is a valuation,
then the function
|x| = c−v(x)
for any c > 1 is a (non-archimedean) absolute value. Conversely, if | · | is a
valuation, then
v(x) = − logc |x|
is a valuation.
Despite this, sometimes people prefer to talk about the valuation rather than
the absolute value, and often this is more natural. As we will later see, in certain
cases, there is a canonical normalization of v, but there is no canonical choice
for the absolute value.
Example. For x ∈ Qp , we define

vp (x) = − logp |x|p .

This is a valuation, and if x ∈ Zp , then vp (x) = n iff pn || x.

Example. Let K be a field, and define
(∞ )
X
i
k((T )) = ai T : ai ∈ k, n ∈ Z .
i=n

This is the field of formal Laurent series over k. We define

X 
v ai T i = min{i : ai 6= 0}.

Then v is a valuation of k((T )).

Recall that for a valued field K, the valuation ring is given by

OK = {x ∈ K : |x| ≤ 1} = {x ∈ K : v(x) ≥ 0}.

Since this is a subring of a field, and the absolute value is multiplicative, we

notice that the units in O are exactly the elements of absolute value 1. The

15
2 Valued fields III Local Fields

remaining elements form an ideal (since the field is non-archimedean), and thus
we have a maximal ideal

m = mK = {x ∈ K : |x| < 1}

The quotient
k = kK = OK /mK
is known as the residue field .
Example. Let K = Qp . Then O = Zp , and m = pZp . So

k = O/m = Zp /pZp ∼
= Z/pZ.

Definition (Primitive polynomial). If K is a valued field and f (x) = a0 + a1 x +

· · · + an xn ∈ K[x] is a polynomial, we say that f is primitive if

max |ai | = 1.
i

In particular, we have f ∈ O[x].

The point of a primitive polynomial is that such a polynomial is naturally,
and non-trivially, an element of k[x]. Moreover, focusing on such polynomials is
not that much of a restriction, since any polynomial is a constant multiple of a
primitive polynomial.
Theorem (Hensel’s lemma). Let K be a complete valued field, and let f ∈ K[x]
be primitive. Put f¯ = f mod m ∈ k[x]. If there is a factorization

f¯(x) = ḡ(x)h̄(x)

with (ḡ, h̄) = 1, then there is a factorization

f (x) = g(x)h(x)

in O[x] with
ḡ = g, h̄ = h mod m,
with deg g = deg ḡ.

Note that requiring deg g = deg ḡ is the best we can hope for — we cannot
guarantee deg h = deg h̄, since we need not have deg f = deg f¯.
This is one of the most important results in the course.
Proof. Let g0 , h0 be arbitrary lifts of ḡ and h̄ to O[x] with deg ḡ = g0 and
deg h̄ = h0 . Then we have

f = g0 h0 mod m.

The idea is to construct a “Taylor expansion” of the desired g and h term by

term, starting from g0 and h0 , and using completeness to guarantee convergence.
To proceed, we use our assumption that ḡ, h̄ are coprime to find some a, b ∈ O[x]
such that
ag0 + bh0 ≡ 1 mod m. (†)

16
2 Valued fields III Local Fields

It is easier to work modulo some element π instead of modulo the ideal m, since
we are used to doing Taylor expansion that way. Fortunately, since the equations
above involve only finitely many coefficients, we can pick an π ∈ m with absolute
value large enough (i.e. close enough to 1) such that the above equations hold
with m replaced with π. Thus, we can write

f = g0 h0 + πr0 , r0 ∈ O[x].

Plugging in (†), we get

f = g0 h0 + πr0 (ag0 + bh0 ) + π 2 (something).

If we are lucky enough that deg r0 b < deg g0 , then we group as we learnt in
secondary school to get

f = (g0 + πr0 b)(h0 + πr0 a) + π 2 (something).

We can then set

g1 = g0 + πr0 b
h1 = h0 + πr0 a,

and then we can write

f = g1 h1 + π 2 r1 , r1 ∈ O[x], deg g1 = deg ḡ. (∗)

If it is not true that deg r0 b ≤ deg g0 , we use the division algorithm to write

r0 b = qg0 + p.

Then we have
f = g0 h0 + π((r0 a + q)g0 + ph0 ),
and then proceed as above.
Given the factorization (∗), we replace r1 by r1 (ag0 + bh0 ), and then repeat
the procedure to get a factorization

f ≡ g2 h2 mod π 3 , deg g2 = deg ḡ.

Inductively, we constrict gk , hk such that

f ≡ gk hk mod π k+1
gk ≡ gk−1 mod π k
hk ≡ hk−1 mod π k
deg gk = deg ḡ

Note that we may drop the terms of hk whose coefficient are in π k+1 O, and the
above equations still hold. Moreover, we can then bound deg hk ≤ deg f − deg gk .
It now remains to set

g = lim gk , h = lim hk .
k→∞ k→∞

17
2 Valued fields III Local Fields

Corollary. Let f (x) = a0 + a1 x + · · · + an xn ∈ K[x] where K is complete and

a0 , an 6= 0. If f is irreducible, then
|a` | ≤ max(|a0 |, |an |)
for all `.
Proof. By scaling, we can wlog f is primitive. We then have to prove that
max(|a0 |, |an |) = 1. If not, let r be minimal such that |ar | = 1. Then 0 < r < n.
Moreover, we can write
f (x) ≡ xr (ar + ar+1 x + · · · + an xn−r ) mod m.
But then Hensel’s lemma says this lifts to a factorization of f , a contradiction.
Corollary (of Hensel’s lemma). Let f ∈ O[x] be monic, and K complete. If f
mod m has a simple root ᾱ ∈ k, then f has a (unique) simple root α ∈ O lifting
ᾱ.
Example. Consider xp−1 − 1 ∈ Zp [x]. We know xp−1 splits into distinct linear
factors over Fp [x]. So all roots lift to Zp . So xp−1 − 1 splits completely in Zp .
So Zp contains all p roots of unity.
√
Example. Since 2 is a quadratic residue mod 7, we know 2 ∈ Q7 .

2.2 Extension of norms

The main goal of this section is to prove the following theorem:
Theorem. Let K be a complete valued field, and let L/K be a finite extension.
Then the absolute value on K has a unique extension to an absolute value on L,
given by q
|α|L = n
|NL/K (α)|,
where n = [L : K] and NL/K is the field norm. Moreover, L is complete with
respect to this absolute value.
Corollary. Let K be complete and M/K be an algebraic extension of K. Then
| · | extends uniquely to an absolute value on M .
This is since any algebraic extension is the union of finite extensions, and
uniqueness means we can patch the absolute values together.
Corollary. Let K be a complete valued field and L/K a finite extension. If
σ ∈ Aut(L/K), then |σ(α)|L = |α|L .
Proof. We check that α 7→ |σ(α)|L is also an absolute value on L extending the
absolute value on K. So the result follows from uniqueness.
Before we can prove the theorem, we need some preliminaries. Given a finite
extension L/K, we would like to consider something more general than a field
norm on L. Instead, we will look at norms of L as a K-vector space. There
are less axioms to check, so naturally there will be more choices for the norm.
However, just as in the case of R-vector spaces, we can show that all choices of
norms are equivalent. So to prove things about the extended field norm, often
we can just pick a convenient vector space norm, prove things about it, then
apply equivalence.

18
2 Valued fields III Local Fields

Definition (Norm on vector space). Let K be a valued field and V a vector

space over K. A norm on V is a function k · k : V → R≥0 such that

(i) kxk = 0 iff x = 0.

(ii) kλk = |λ| kxk for all λ ∈ K and x ∈ V .
(iii) kx + yk ≤ max{kxk , kyk}.
Note that our norms are also non-Archimedean.
0
Definition (Equivalence of norms). Let k · k and k · k be norms on V . Then
two norms are equivalent if they induce the same topology on V , i.e. there are
C, D > 0 such that
0
C kxk ≤ kxk ≤ D kxk
for all x ∈ V .
One of the most convenient norms we will work with is the max norm:
Example (Max norm). Let K be a complete valued field, and V a finite-
dimensional K-vector space. Let x1 , · · · , xn be a basis of V . Then if
X
x= ai xi ,

then
kxkmax = max |ai |
i

defines a norm on V .

Proposition. Let K be a complete valued field, and V a finite-dimensional

K-vector space. Then V is complete under the max norm.
Proof. Given a Cauchy sequence in V under the max norm, take the limit of each
coordinate to get the limit of the sequence, using the fact that K is complete.

That was remarkably easy. We can now immediately transfer this to all other
norms we can think of by showing all norms are equivalent.
Proposition. Let K be a complete valued field, and V a finite-dimensional
K-vector space. Then any norm k · k on V is equivalent to k · kmax .
Corollary. V is complete with respect to any norm.

Proof. Let k · k be a norm. We need to find C, D > 0 such that

C kxkmax ≤ kxk ≤ D kxkmax .

We set D = maxi (kxi k). Then we have

X
kxk = ai xi ≤ max (|ai | kxi k) ≤ (max |ai |)D = kxkmax D.

We find C by induction on n. If n = 1, then kxk = ka1 x1 k = |a1 | kxk =

kxkmax kx1 k. So C = kx1 k works.

19
2 Valued fields III Local Fields

For n ≥ 2, we let

Vi = Kx1 ⊕ · · · ⊕ Kxi−1 ⊕ Kxi+1 ⊕ · · · ⊕ Kxn

= span{x1 , · · · , xi−1 , xi+1 , · · · , xn }.

By the induction hypothesis, each Vi is complete with respect to (the restriction

of) k · k. So in particular Vi is closed in V . So we know that the union
n
[
xi + V i
i=1

is also closed. BySconstruction, this does not contain 0. So there is some C > 0
n
such that if x ∈ i=1 xi + Vi , then kxk ≥ C. We claim that

C kxkmax ≤ kxk.
P
Indeed, take x = ai xi ∈ V . Let r be such that

|ar | = max(|ai |) = kxkmax .

i

Then
−1
kxkmax kxk = a−1

r x


a1 ar−1 ar+1 an
= x1 + · · · +
xr−1 + xr + xr+1 + · · · + xn
ar ar ar ar
≥ C,

since the last vector is an element of xr + Vr .

Before we can prove our theorem, we note the following two easy lemmas:
Lemma. Let K be a valued field. Then the valuation ring OK is integrally
closed in K.

Proof. Let x ∈ K and |x| > 1. Suppose we have an−1 , · · · , a0 ∈ OK . Then we

have
|xn | > |a0 + a1 x + · · · + an−1 xn−1 |.
So we know
xn + an−1 xn−1 + · · · + a1 x + a0
has non-zero norm, and in particular is non-zero. So x is not integral over OK .
So OK is integrally closed.
Lemma. Let L be a field and | · | a function that satisfies all axioms of an
absolute value but the strong triangle inequality. Then | · | is an absolute value
iff |α| ≤ 1 implies |α + 1| ≤ 1.

Proof. It is clear that if | · | is an absolute value, then |α| ≤ 1 implies |α + 1| ≤ 1.

Conversely, if this holds, and |x| ≤ |y|, then |x/y| ≤ 1. So |x/y + 1| ≤ 1. So
|x + y| ≤ |y|. So |x + y| ≤ max{|x|, |y|}.
Finally, we get to prove our theorem.

20
2 Valued fields III Local Fields

Theorem. Let K be a complete valued field, and let L/K be a finite extension.
Then the absolute value on K has a unique extension to an absolute value on L,
given by q 
|α| = n NL/K (α),
L

where n = [L : K] and NL/K is the field norm. Moreover, L is complete with

respect to this absolute value.
Proof. For uniqueness and completeness, if | · |L is an absolute value on L, then
it is in particular a K-norm on L as a finite-dimensional vector space. So we
know L is complete with respect to | · |L .
0 0
If | · |L is another absolute value extending | · |, then we know | · |L and | · |L
are equivalent in the sense of inducing the same topology. But then from one of
the early exercises, when field norms are equivalent, then we can find some s > 0
s 0
such that | · |L = | · |L . But the two norms agree on K, and they are non-trivial.
So we must have s = 1. So the norms are equal.
To show existence, we have to prove that
q 
|α|L = n NL/K (α)

is a norm.
(i) If |α|L = 0, then NL/K (α) = 0. This is true iff α = 0.
(ii) The multiplicativity
√ of |α| and follows from the multiplicativity of NL/K ,
| · | and n · .
To show the strong triangle inequality, it suffices to show that |α|L ≤ 1 implies
|α + 1|L ≤ 1.
Recall that

OL = {α ∈ L : |α|L ≤ 1} = {α ∈ L : NL/K (α) ∈ OK }.

We claim that OL is the integral closure of OK in L. This implies what we

want, since the integral closure is closed under addition (and 1 is in the integral
closure).
Let α ∈ OL . We may assume α 6= 0, since that case is trivial. Let the
minimal polynomial of α over K be

f (x) = a0 + a1 x + · · · + an−1 xn−1 + xn ∈ K[x].

We need to show that ai ∈ OK for all i. In other words, |ai | ≤ 1 for all i. This
is easy for a0 , since
NL/K (α) = ±am 0 ,

and hence |a0 | ≤ 1.

By the corollary of Hensel’s lemma, for each i, we have

|ai | ≤ max(|a0 |, 1)

By general properties of the field norm, there is some m ∈ Z≥1 such that
NL/K (α) = ±am0 . So we have
  
|ai | ≤ max NL/K (α)1/m  , 1 = 1.
 

21
2 Valued fields III Local Fields

So f ∈ OK [x]. So α is integral over OK .

On the other hand, suppose α is integral over OK . Let K̄/K be an algebraic
closure of K. Note that
!d
Y
NL/K (α) = σ(α) ,
σ:L,→K̄

for some d ∈ Z≥1 , and each σ(α) is integral over OK , since α is (apply σ to the
minimal polynomial). This implies that NL/K (α) is integral over OK (and lies
in K). So NL/K (α) ∈ OK since OK is integrally closed in K.
Corollary (of the proof). Let K be a complete valued field, and L/K a finite
extension. We equip L with | · |L extending | · | on K. Then OL is the integral
closure of OK in L.

2.3 Newton polygons

We are going to have a small digression to Newton polygons. We will not make
use of them in this course, but it is a cute visual devices that tell us about roots
of polynomials. It is very annoying to write down a formal definition, so we first
look at some examples. We will work with valuations rather than the absolute
value.
Example. Consider the valued field (Qp , vp ), and the polynomial

t4 + p2 t4 − p3 t2 + pt + p3 .

We then plot the coefficients for each power of t, and then draw a “convex
polygon” so that all points lie on or above it:

valuation of coefficient

power of t
0 1 2 3 4

Example. Consider (Q2 , v2 ) with the polynomial

7 9
4t4 + 5t3 + t + .
2 2
Here there is no t2 term, so we simply don’t draw anything.

22
2 Valued fields III Local Fields

valuation of coefficient

power of t
0 1 2 3 4

−1

We now go to come up with a formal definition.

Definition (Lower convex set). We say a set S ⊆ R2 is lower convex if

(i) Whenever (x, y) ∈ S, then (x, z) ∈ S for all z ≥ y.

(ii) S is convex.
Definition (Lower convex hull). Given any set of points T ⊆ R2 , there is a
minimal lower convex set S ⊇ T (by the intersection of all lower convex sets
containing T – this is a non-empty definition because R2 satisfies the property).
This is known as the lower convex hull of the points.
Example. The lower convex hull of the points (0, 3), (1, 1), (2, 3), (3, 2), (4, 0) is
given by the region denoted below:

Definition (Newton polygon). Let f (x) = a0 + a1 x + · · · + an xn ∈ K[x], where

(K, v) is a valued field. Then the Newton polygon of f is the lower convex hull
of {(i, v(ai )) : i = 0, · · · , n, ai 6= 0}.
This is the formal definition, so in our first example, the Newton polygon
really should be the shaded area shown above, but most of the time, we only
care about the lower line.
Definition (Break points). Given a polynomial, the points (i, v(ai )) lying on
the boundary of the Newton polygon are known as the break points.

23
2 Valued fields III Local Fields

Definition (Line segment). Given a polynomial, the line segment between two
adjacent break points is a line segment.
Definition (Multiplicity/length). The length or multiplicity of a line segment
is the horizontal length.
Definition (Slope). The slope of a line segment is its slope.
Example. Consider again (Q2 , v2 ) with the polynomial
7 9
4t4 + 5t3 + t + .
2 2
valuation of coefficient

power of t

The middle segment has length 2 and slope 1/2.

Example. In the following Newton polygon:

The second line segment has length 3 and slope − 31 .

It turns out the Newton polygon tells us something about the roots of the
polynomial.
Theorem. Let K be complete valued field, and v the valuation on K. We let
f (x) = a0 + a1 x + · · · + an xn ∈ K[x].
Let L be the splitting field of f over K, equipped with the unique extension w
of v.
If (r, v(ar )) → (s, v(as )) is a line segment of the Newton polygon of f with
slope −m ∈ R, then f has precisely s − r roots of valuation m.

24
2 Valued fields III Local Fields

Note that by lower convexity, there can be at most one line segment for each
slope. So this theorem makes sense.

Proof. Dividing by an only shifts the polygon vertically, so we may wlog an = 1.

We number the roots of f such that

w(α1 ) = · · · = w(αs1 ) = m1
w(αs1 +1 ) = · · · = w(αs2 ) = m2
..
.
w(αst ) = · · · = w(αn ) = mt+1 ,

where we have
m1 < m2 < · · · < mt+1 .
Then we know

v(an ) = v(1) = 0
X 
v(an−1 ) = w αi ≥ min w(αi ) = m1
i
X 
v(an−2 ) = w αi αj ≥ min w(αi αj ) = 2m1
i6=j
..
.
 
X
v(an−s1 ) = w  αi1 ...αis  = min w(αi1 · · · αis1 ) = s1 m1 .
1
i1 6=...6=is1

It is important that in the last one, we have equality, not an inequality, because
there is one term in the sum whose valuation is less than all the others.
We can then continue to get

v(αn−s1 −1 ) ≥ min w(αi1 · · · αis1 +1 ) = s1 m1 + m2 ,

until we reach
v(αn−s1 −s2 ) = s1 m1 + (s2 − s1 )m2 .
We keep going on.
We draw the Newton polygon.

(n, 0)

(n − s1 , s1 m1 )
···
(n − s1 − s2 , s1 m1 + (s2 − s1 )m1 )

25
2 Valued fields III Local Fields

We don’t know where exactly the other points are, but the inequalities imply
that the (i, v(ai )) are above the lines drawn. So this is the Newton polygon.
Counting from the right, the first line segment has length n − (n − s1 ) = s1
and slope
0 − s1 m1
= −m1 .
n − (n − s1 )
In general, the kth segment has length (n − sk−1 ) − (n − sk ) = sk − sk−1 , and
slope
 Pk−2   Pk−1 
s1 m1 + i=1 (si+1 − si )mi+1 − s1 m1 + i=1 (si+1 − si )mi+1
sk − sk−1
−(sk − sk−1 )mk
= = −mk .
sk − sk−1

and the others follow similarly.

Corollary. If f is irreducible, then the Newton polygon has a single line segment.
Proof. We need to show that all roots have the same valuation. Let α, β be in
the splitting field L. Then there is some σ ∈ Aut(L/K) such that σ(α) = β.
Then w(α) = w(σ(α)) = β. So done.

Note that Eisenstein’s criterion is a (partial) converse to this!

26
3 Discretely valued fields III Local Fields

3 Discretely valued fields

We are now going to further specialize. While a valued field already has some
nice properties, we can’t really say much more about them without knowing
much about their valuations.
Recall our previous two examples of valued fields: Qp and Fp ((T )). The
valuations had the special property that they take values in Z. Such fields are
known as discretely valued fields.
Definition (Discretely valued field). Let K be a valued field with valuation v.
We say K is a discretely valued field (DVF) if v(k × ) ⊆ R is a discrete subgroup
of R, i.e. v(k × ) is infinite cyclic.
Note that we do not require the image to be exactly Z ⊆ R. So we allow
scaled versions of the valuation. This is useful because the property of mapping
into Z is not preserved under field extensions in general, as we will later see. We
will call those that do land in Z normalized valuations.
Definition (Normalized valuation). Let K be a DVF. The normalized valuation
VK on K is the unique valuation on K in the given equivalence class of valuations
whose image is Z.
Note that the normalized valuation does not give us a preferred choice of
absolute value, since to obtain an absolute value, we still have to arbitrarily pick
the base c > 1 to define |x| = c−v(x) .
Definition (Uniformizer). Let K be a discrete valued field. We say π ∈ K is
uniformizer if v(π) > 0 and v(π) generates v(k × ) (iff v(π) has minimal positive
valuation).
So with a normalized valuation, we have vK (π) = 1.
Example. The usual valuation on Qp is normalized, and so is the usual valuation
on k((T )). p is a uniformizer for Qp and T is a uniformizer for k((T )).
The kinds of fields we will be interested are local fields. The definition we
have here might seem rather ad hoc. This is just one of the many equivalent
characterizations of a local field, and the one we pick here is the easiest to state.
Definition (Local field). A local field is a complete discretely valued field with
a finite residue field.
Example. Q and Qp with vp are both discretely valued fields, and Qp is a local
field. p is a uniformizer.
Example. The Laurent series field k((T )) with valuation
X 
v an T n = inf{n : an 6= 0}

is a discrete valued field, and is a local field if and only if k is finite field, as the
residue field is exactly k. We have
(∞ )
X
n
Ok((T )) = k[[T ]] = an T : an ∈ k .
n=0

Here T is a uniformizer.

27
3 Discretely valued fields III Local Fields

These discretely valued field are pretty much like the p-adic numbers.
Proposition. Let K be a discretely valued field with uniformizer π. Let S ⊆ OK
be a set of coset representatives of Ok /mk = kK containing 0. Then
(i) The non-zero ideals of OK are π n OK for n ≥ 0.
(ii) The ring OK is a PID with unique prime π (up to units), and mK = πOK .
(iii) The topology on OK induced by the absolute value is the π-adic topology.

(iv) If K is complete, then OK is π-adically complete.

(v) If K is complete, then any x ∈ K can be written uniquely as
∞
X
x= an π n ,
n−∞

where an ∈ S, and
|x| = |π|− inf{n:an 6=0} .

(vi) The completion K̂ is also discretely valued and π is a uniformizer, and

moreover the natural map

Ok ∼ OK̂
π n Ok π n OK̂

is an isomorphism.
Proof. The same as for Qp and Zp , with π instead of p.

Proposition. Let K be a discretely valued field. Then K is a local field iff OK

is compact.
Proof. If OK is compact, then π −n OK is compact for all n ≥ 0 (where π is the
uniformizer), and in particular complete. So
∞
[
K= π −n OK
n≥0

is complete, as this is an increasing union, and Cauchy sequences are bounded.

Also, we know the quotient map OK → kK is continuous when kK is given the
discrete topology, by definition of the π-adic topology. So kK is compact and
discrete, hence finite.
In the other direction, if K is local, then we know OK /π n OK is finite for
all n ≥ 0 (by induction and finiteness of kK ). We let (xi ) be a sequence in OK .
Then by finiteness of OK /πOK , there is a subsequence (x1,i ) which is constant
modulo π. We keep going, choosing a subsequence (xn+1,i ) of (xni ) such that
(xn+1,i ) is constant modulo π n+1 . Then (xi,i )∞
i=1 converges, since it is Cauchy as

|xii − xjj | ≤ |π|j

for j ≤ i. So OK is sequentially compact, hence compact.

28
3 Discretely valued fields III Local Fields

Now the valuation ring OK inherits a valuation from K, and this gives it a
structure of a discrete valuation ring. We will define a discrete valuation ring in
a funny way, but there are many equivalent definitions that we will not list.
Definition (Discrete valuation ring). A ring R is called a discrete valuation
ring (DVR) if it is a PID with a unique prime element up to units.
Proposition. R is a DVR iff R ∼
= OK for some DVF K.
Proof. We have already seen that valuation rings of discrete valuation fields are
DVRs. In the other direction, let R be a DVR, and π a prime. Let x ∈ R \ {0}.
Then we can find a unique unit u ∈ R× and n ∈ Z≥0 such that x = π n u (say,
by unique factorization of PIDs). We define
(
n x 6= 0
v(x) =
∞ x=0

This is then a discrete valuation of R. This extends uniquely to the field of

fractions K. It remains to show that R = OK . First note that
 
1
K=R .
π

This is since any non-zero element in R π1 looks like π n u, u ∈ R× , n ∈ Z, and

 

is already invertible. So it must be the field of fractions. Then we have

v(π n u) = n ∈ Z≥0 ⇐⇒ π n u ∈ R.

So we have R = OK .
Now recall our two “standard” examples of valued fields — Fp ((T )) and
Qp . Both of their residue fields are Fp , and in particular has characteristic p.
However, Fp ((T )) itself is also of characteristic p, while Qp has characteristic 0.
It would thus be helpful to split these into two different cases:
Definition (Equal and mixed characteristic). Let K be a valued field with
residue field kK . Then K has equal characteristic if

char K = char kK .

Otherwise, we have K has mixed characteristic.

If K has mixed characteristic, then necessarily char K = 0, and char kK > 0.
Example. Qp has mixed characteristic, since char Qp = 0 but char kQp =
Z/pZ = p.
We will also need the following definition:
Definition (Perfect ring). Let R be a ring of characteristic p. We say R is
perfect if the Frobenius map x 7→ xp is an automorphism of R, i.e. every element
of R has a pth root.
Fact. Let F be a field of characteristic p. Then F is perfect if and only if every
finite extension of F is separable.
Example. Fq is perfect for every q = pn .

29
3 Discretely valued fields III Local Fields

3.1 Teichmüller lifts

Take our favorite discretely valued ring Zp . This is p-adically complete, so we
can write each element as

x = a0 + a1 p + a2 p2 + · · · ,

where each ai is in {0, 1, · · · , p − 1}. The reason this works is that 0, 1, · · · , p −

1 are coset representatives of the ring Zp /pZp ∼ = Z/pZ. While these coset
representatives might feel like a “natural” thing to do in this context, this is
because we have implicitly identified with Zp /pZp ∼ = Z/pZ as a particular subset
of Z ⊆ Zp . However, this identification respects effectively no algebraic structure
at all. For example, we cannot multiplying the cosets simply by multiplying the
representatives as elements of Zp , because, say, (p − 1)2 = p2 − 2p + 1, which is
not 1. So this is actually quite bad, at least theoretically.
It turns out that we can actually construct “natural” lifts in a very general
scenario.
Theorem. Let R be a ring, and let x ∈ R. Assume that R is x-adically
complete and that R/xR is perfect of characteristic p. Then there is a unique
map [−] : R/xR → R such that

[a] ≡ a mod x

and
[ab] = [a][b].
for all a, b ∈ R/xR. Moreover, if R has characteristic p, then [−] is a ring
homomorphism.
Definition (Teichmüller map). The map [−] : R/xR → R is called the Te-
ichm uller map. [x] is called the Teichmüller lift or representative of x.
The idea of the proof is as follows: suppose we have an a ∈ R/xR. If we
randomly picked a lift α, then chances are it would be a pretty “bad” choice,
since any two such choices can differ by a multiple of x.
Suppose we instead lifted a pth root of a to R, and then take the pth power
of it. We claim that this is a better way of picking a lift. Suppose we have picked
−1
two lifts of ap , say, α1 and α10 . Then α10 = xc + α1 for some c. So we have

(α10 )p − α1p = α1p + pxc + O(x2 ) − α1p = pxc + O(x2 ),

where we abuse notation and write O(x2 ) to mean terms that are multiples of
x2 .
We now recall that R/xR has characteristic p, so p ∈ xR. Thus in fact
pxc = O(x2 ). So we have

(α10 )p − α1p = O(x2 ).

So while the lift is still arbitrary, any two arbitrary choices can differ by at most
x2 . Alternatively, our lift is now a well-defined element of R/x2 R.
We can, of course, do better. We can lift the p2 th root of a to R, then take
the p2 th power of it. Now any two lifts can differ by at most O(x3 ). More
generally, we can try to lift the pn th root of a, then take the pn th power of

30
3 Discretely valued fields III Local Fields

it. We keep picking a higher and higher n, take the limit, and hopefully get
something useful out!
To prove this result, we will need the following messy lemma:
Lemma. Let R be a ring with x ∈ R such that R/xR has characteristic p. Let
α, β ∈ R be such that
α = β mod xk (†)
Then we have
αp = β p mod xk+1 .
Proof. It is left as an exercise to modify the proof to work for p = 2 (it is actually
easier). So suppose p is odd. We take the pth power of (†) to obtain
p−1  
p p
X p
α −β + αp−i β i ∈ xp(k+1) R.
i=1
i

We can now write

p−1
p−1   2  
i p i p
X X
p−i i
(αβ)i αp−2i − β p−2i


(−1) α β = (−1)
i=1
i i=1
i
= p(α − β)(something).

Now since R/xR has characteristic p, we know p ∈ xR. By assumption, we know

α − β ∈ xk+1 R. So this whole mess is in xk+2 R, and we are done.
−n
Proof of theorem. Let a ∈ R/xR. For each n, there is a unique ap ∈ R/xR.
We lift this arbitrarily to some αn ∈ R such that
−n
αn ≡ ap mod x.

We define n
βn = αnp .
The claim is that
[a] = lim βn
n→∞

exists and is independent of the choices.

Note that if the limit exists no matter how we choose the αn , then it
must be independent of the choices. Indeed, if we had choices βn and βn0 ,
then β1 , β20 , β3 , β40 , β5 , β60 , · · · is also a respectable choice of lifts, and thus must
converge. So βn and βn0 must have the same limit.
Since the ring is x-adically complete and is discretely valued, to show the
limit exists, it suffices to show that βn+1 − βn → 0 x-adically. Indeed, we have
p n n
βn+1 − βn = (αn+1 )p − αnp .

We now notice that

p −n−1 −n
αn+1 ≡ (ap )p = ap ≡ αn mod x.

So by applying the previous the lemma many times, we obtain

p n n
(αn+1 )p ≡ αnp mod xn+1 .

31
3 Discretely valued fields III Local Fields

So βn+1 − βn ∈ xn+1 R. So lim βn exists.

To see [a] = a mod x, we just have to note that
n −n n
lim αnp ≡ lim (ap )p = lim a = a mod x.
n→∞ n→∞

(here we are using the fact that the map R → R/xR is continuous when R is
given the x-adic topology and R/xR is given the discrete topology)
The remaining properties then follow trivially from the uniqueness of the
above limit.
For multiplicativity, if we have another element b ∈ R/xR, with γn ∈ R
−n −n
lifting bp for all n, then αn γn lifts (ab)p . So
n n n n
[ab] = lim αnp γnp = lim αnp lim γnp = [a][b].
−n −n −n
If R has characteristic p, then αn + γn lifts ap + bp = (a + b)p . So
pn n n
[a + b] = lim(αn + γn ) = lim αnp + lim γnp = [a] + [b].
Since 1 is a lift of 1 and 0 is a lift of 0, it follows that this is a ring homomorphism.
Finally, to show uniqueness, suppose φ : R/xR → R is a map with these
−n −n
properties. Then we note that φ(ap ) ≡ ap mod x, and is thus a valid choice
of αn . So we have
−n n
[a] = lim φ(ap )p = lim φ(a) = φ(a).
n→∞

Example. Let R = Zp and x = p. Then [−] : Fp → Zp satisfies

[x]p−1 = [xp−1 ] = [1] = 1.
So the image of [x] must be the unique p − 1th root of unity lifting x (recall we
proved their existence via Hensel’s lemma).
When proving theorems about these rings, the Teichmüller lifts would be
very handy and natural things to use. However, when we want to do actual
computations, there is absolutely no reason why these would be easier!
As an application, we can prove the following characterization of equal
characteristic complete DVF’s.
Theorem. Let K be a complete discretely valued field of equal characteristic p,
and assume that kK is perfect. Then K ∼
= kK ((T )).
Proof. Let K be a complete DVF. Since every DVF the field of fractions of
its valuation ring, it suffices to prove that OK ∼
= kK [[T ]]. We know OK has
characteristic p. So [−] : kK → OK is an injective ring homomorphism. We
choose a uniformizer π ∈ OK , and define
kK [[T ]] → OK
by
∞
X ∞
X
an T n 7→ [an ]π n .
n=0 n=0
Then this is a ring homomorphism since [−] is. The bijectivity follows from
property (v) in our list of properties of complete DVF’s.
Corollary. Let K be a local field of equal characteristic p. Then kK ∼
= Fq for
some q a power of p, and K ∼
= Fq ((T )).

32
3 Discretely valued fields III Local Fields

3.2 Witt vectors*

We are now going to look at the mixed characteristic analogue of this result. We
want something that allows us to go from characteristic p to characteristic 0.
This is known as Witt vectors, which is non-examinable.
We start with the notion of a strict p-ring. Roughly this is a ring that satisfies
all the good properties whose name has the word “p” in it.
Definition (Strict p-ring). Let A be a ring. A is called a strict p-ring if it is
p-torsion free, p-adically complete, and A/pA is a perfect ring.
Note that a strict p-ring in particular satisfies the conditions for the Te-
ichmüller lift to exist, for x = p.
Example. Zp is a strict p-ring.
The next example we are going to construct is more complicated. This is in
some sense a generalization of the usual polynomial rings Z[x1 , · · · , xn ], or more
generally,
Z[xi | i ∈ I],
for I possibly infinite. To construct the “free” strict p-ring, after adding all these
variables xi , to make it a strict p-ring, we also need to add their pth roots, and
the p2 th roots etc, and then take the p-adic completion, and hope for the best.
Example. Let X = {xi : i ∈ I} be a set. Let
∞
−∞ [ −n
B = Z[xpi | i ∈ I] = Z[xpi | i ∈ I].
n=0

Here the union on the right is taken by treating

−1
Z[xi | i ∈ I] ⊆ Z[xip | i ∈ I] ⊆ · · ·

in the natural way.

We let A be the p-adic completion of B. We claim that A is a strict p-ring
p−∞
and A/pA ∼ = Fp [xi | i ∈ I].
Indeed, we see that B is p-torsion free. By Exercise 13 on Sheet 1, we know
A is p-adically complete and torsion free. Moreover,
−∞
A/pA ∼
= B/pB ∼
p
= Fp [xi | i ∈ I],

which is perfect since every element has a p-th root.

If A is a strict p-ring, then we know that we have a Teichmüller map

[−] : A/pA → A,

Lemma. Let A be a strict p-ring. Then any element of A can be written

uniquely as
X∞
a= [an ]pn ,
n=0

for a unique an ∈ A/pA.

33
3 Discretely valued fields III Local Fields

Proof. We recursively construct the an by

a0 = a (mod p)
−1
a1 ≡ p (a − [a0 ]) (mod p)
..
.

Lemma. Let A and B be strict p-rings and let f : A/pA → B/pB be a ring
homomorphism. Then there is a unique homomorphism F : A → B such that
f = F mod p, given by
X  X
F [an ]pn = [f (an )]pn .

Proof sketch. We define F by the given formula and check that it works. First of
all, by the formula, F is p-adically continuous, and the key thing is to check that
it is additive (which is slightly messy). Multiplicativity then follows formally
from the continuity and additivity.
To show uniqueness, suppose that we have some ψ lifting f . Then ψ(p) = p.
So ψ is p-adically continuous. So it suffices to show that ψ([a]) = [ψ(a)].
−n −n
We take αn ∈ A lifting ap ∈ A/pA. Then ψ(αn ) lifts f (a)p . So
−n −n
ψ([a]) = lim ψ(αnp ) = lim ψ(αn )p = [f (a)].

So done.
There is a generalization of this result:
Proposition. Let A be a strict p-ring and B be a ring with an element x
such that B is x-adically complete and B/xB is perfect of characteristic p. If
f : A/pA → B/xB is a ring homomorphism. Then there exists a unique ring
homomorphism F : A → B with f = F mod x, i.e. the following diagram
commutes:
F
A B
.
f
A/pA B/xB

Indeed, the conditions on B are sufficient for Teichmüller lifts to exist, and
we can at least write down the previous formula, then painfully check it works.
We can now state the main theorem about strict p-rings.
Theorem. Let R be a perfect ring. Then there is a unique (up to isomorphism)
strict p-ring W (B) called the Witt vector s of R such that W (R)/pW (R) ∼
= R.
Moreover, for any other perfect ring R, the reduction mod p map gives a
bijection
HomRing (W (R), W (R0 )) ∼ HomRing (R, R0 ) .

Proof sketch. If W (R) and W (R0 ) are such strict p-rings, then the second part
follows from the previous lemma. Indeed, if C is a strict p-ring with C/pC ∼ =
R∼ = W (R)/pW (R), then the isomorphism ᾱ : W (R)/pW (R) → C/pC and its
inverse ᾱ−1 have unique lifts γ : W (R) → C and γ −1 : C → W (R), and these
are inverses by uniqueness of lifts.

34
3 Discretely valued fields III Local Fields

To show existence, let R be a perfect ring. We form

−∞
Fp [xpr | r ∈ R] → R
xr 7→ r
−∞
Then we know that the p-adic completion of Z[xpr | r ∈ R], written A, is a
strict p-ring with
−∞
A/pA ∼= Fp [xpr | r ∈ R].
We write −∞
I = ker(Fp [xpr | r ∈ R] → R).
Then define
∞
( )
X
n
J= [ak ]p ∈ A : an ∈ I for all n .
n=0

This turns out to be an ideal.

J A R

0 I A/pA R 0

We put W (R) = A/J. We can then painfully check that this has all the required
properties. For example, if
∞
X
x= [an ]pn ∈ A,
n=0

and
∞
X
px = [an ]pn+1 ∈ J,
n=0

then by definition of J, we know [an ] ∈ I. So x ∈ J. So W (R)/J is p-torsion

free. By a similar calculation, one checks that
∞
\
pn W (R) = {0}.
n=0

This implies that W (R) injects to its p-adic completion. Using that A is p-adically
complete, one checks the surjectivity by hand.
Also, we have
W (R) ∼ A
= .
pW (R) J + pA
But we know ( )
X
J + pA = [an ]pn | a0 ∈ I .
n

So we have −∞
W (R) ∼ Fp [xpr | r ∈ R] ∼
= = R.
pW (R) I
So we know that W (R) is a strict p-ring.

35
3 Discretely valued fields III Local Fields

Example. W (Fp ) = Zp , since Zp satisfies all the properties W (Fp ) is supposed

to satisfy.
Proposition. A complete DVR A of mixed characteristic with perfect residue
field and such that p is a uniformizer is the same as a strict p-ring A such that
A/pA is a field.
Proof. Let A be a complete DVR such that p is a uniformizer and A/pA is
perfect. Then A is p-torsion free, as A is an integral domain of characteristic 0.
Since it is also p-adically complete, it is a strict p-ring.
Conversely, if A is a strict p-ring, and A/pA is a field, then we have A× ⊆
A \ pA, and we claim that A× = A \ pA. Let
∞
X
x= [xn ]pn
n=0

with x0 6= 0, i.e. x 6∈ pA. We want to show that x is a unit. Since A/pA is a

field, we can multiply by [x−1
0 ], so we may wlog x0 = 1. Then x = 1 − py for
some y ∈ A. So we can invert this with a geometric series
∞
X
x−1 = pn y n .
n=0

So x is a unit. Now, looking at Teichmüller expansions and factoring out multiple

of p, any non-zero element z can be written as pn u for a unique n ≥ Z≥0 and
u ∈ A× . Then we have (
n z 6= 0
v(z) =
∞ z=0
is a discrete valuation on A.
Definition (Absolute ramification index). Let R be a DVR with mixed charac-
teristic p with normalized valuation vR . The integer vR (p) is called the absolute
ramification index of R.
Corollary. Let R be a complete DVR of mixed characteristic with absolute
ramification index 1 and perfect residue field k. Then R ∼
= W (k).
Proof. Having absolute ramification index 1 is the same as saying p is a uni-
formizer. So R is a strict p-ring with R/pR ∼
= k. By uniqueness of the Witt
vector, we know R ∼
= W (k).
Theorem. Let R be a complete DVR of mixed characteristic p with a perfect
residue field k and uniformizer π. Then R is finite over W (k).
Proof. We need to first exhibit W (k) as a subring of R. We know that id : k → k
lifts to a homomorphism W (k) → R. The kernel is a prime ideal because R is
an integral domain. So it is either 0 or pW (k). But R has characteristic 0. So it
can’t be pW (k). So this must be an injection.
Let e be the absolute ramification index of R. We want to prove that
e−1
M
R= π i W (k).
i=0

36
3 Discretely valued fields III Local Fields

Looking at valuations, one sees that 1, π, π, · · · , π e−1 are linearly independent

over W (k). So we can form
e−1
M
M= π i W (k) ⊆ R.
i=0

We consider R/pR. Looking at Teichmüller expansions

∞
X e−1
X
[xn ]π n ≡ [xn ]π n mod pR,
n=0 n=0

we see that 1, π, · · · , π e−1 generate R/pR as W (k)-modules (all the Teichmüller

lifts live in W (k)). Therefore R = M + pR. We iterate to get

R = M + p(M + pR) = M + p2 r = · · · = M + pm R

for all m ≥ 1. So M is dense in R. But M is also p-adically complete, hence

closed in R. So M = R.
The important statement to take away is
Corollary. Let K be a mixed characteristic local field. Then K is a finite
extension of Qp .
Proof. Let Fq be the residue field of K. Then OK is finite over W (Fq ) by the
previous theorem. So it suffices to show that W (Fq ) is finite over W (Fp ) = Zp .
Again the inclusion Fp ⊆ Fq gives an injection W (Fp ) ,→ W (Fq ). Write q = pd ,
and let x1 , · · · , xd ∈ W (Fq ) be lifts of an Fp -bases of Fq .. Then we have
d
M
W (Fq ) = xd Zp + pW (Fq ),
i=1

and then argue as in the end of the previous theorem to get

d
M
W (Fq ) = xd Zp .
i=1

37
4 Some p-adic analysis III Local Fields

4 Some p-adic analysis

We are now going to do some fun things that is not really related to the course.
In “normal” analysis, the applied mathematicians hold the belief that every
function can be written as a power series
∞
X
f (x) = an xn .
n=0

When we move on to p-adic numbers, we do not get such a power series expansion.
However, we obtain an analogous result using binomial coefficients.
Before that, we have a quick look at our familiar functions exp and log, which
we shall continue to define as a power series:
∞ ∞
X xn X xn
exp(x) = , log(1 + x) = (−1)n−1
n=0
n! n=1
n

The domain will no longer be all of the field. Instead, we have the following
result:

Proposition. Let K be a complete valued field with an absolute value | · | and

assume that K ⊇ Qp and | · | restricts to the usual p-adic norm on Qp . Then
exp(x) converges for |x| < p−1/(p−1) and log(1 + x) converges for |x| < 1, and
then define continuous maps

exp : {x ∈ K : |x| < p−1/(p−1) } → OK

log : {1 + x ∈ K : |x| < 1} → K.

Proof. We let v = − logp | · | be a valuation extending vp . Then we have the

dumb estimate
v(n) ≤ logp n.
Then we have
xn
 
v ≥ n · v(x) − logp n → ∞
n
if v(x) > 0. So log converges.
For exp, we have
n − sp (n)
v(n!) = ,
p−1
where sp (n) is the sum of the p-adic digits of n. Then we have
 n  
x n 1
v ≥ n · v(x) − = n · v(x) − →∞
n! p−1 p−1
n

if v(x) > 1/(p − 1). Since v xn! ≥ 0, this lands in OK .
For the continuity, we just use uniform convergence as in the real case.
What we really want to talk about is binomial coefficients. Let n ≥ 1. Then
we know that  
x x(x − 1) · · · (x − n + 1)
=
n n!

38
4 Some p-adic analysis III Local Fields

x


is a polynomial in x, and so defines a continuous function Zp → Q p by x →
7 n .
x


When n = 0, we
set n = 1 for all x ∈ Zp .
x
x

We know n ∈ Z if x ∈ Z≥0 . So by density of Z≥0 ⊆ Zp , we must have
n ∈ Zp for all x ∈ Zp .
We will eventually want to prove the following result:
Theorem (Mahler’s theorem). Let f : Zp → Qp be any continuous function.
Then there is a unique sequence (an )n≥0 with an ∈ Qp and an → 0 such that
∞  
X x
f (x) = an ,
n=0
n

and moreover
sup |f (x)| = max |ak |.
x∈Zp k∈N

We write C(Zp , Qp ) for the set of continuous functions Zp → Qp as usual.

This is a Qp vector space as usual, with

(λf + µg)(x) = λf (x) + µg(x)

for all λ, µ ∈ Qp and f, g ∈ C(Zp , Qp ) and x ∈ Zp .

If f ∈ C(Zp , Qp ), we set

kf k = sup |f (x)|p .
x∈Zp

Since Zp is compact, we know that f is bounded. So the supremum exists and

is attained.
Proposition. The norm k · k defined above is in fact a (non-archimedean)
norm, and that C(Zp , Qp ) is complete under this norm.
Let c0 denote the set of sequences (an )∞
n=0 in Qp such that an → 0. This is
a Qp -vector space with a norm

k(an )k = max |an |p ,

n∈N

and c0 is complete. So what Mahler’s theorem gives us is an isometric isomor-

phism between c0 and C(Zp , Qp ).
We define
∆ : C(Zp , Qp ) → C(Zp , Qp )
by
∆f (x) = f (x + 1) − f (x).
By induction, we have
n  
i n
X
n
∆ f (x) = (−1) f (x + n − i).
i=0
i

Note that ∆ is a linear operator on C(Zp , Qp ), and moreover

|∆f (x)|p = |f (x + 1) − f (x)|p ≤ kf k.

39
4 Some p-adic analysis III Local Fields

So we have
k∆f k ≤ kf k.
In other words, we have
k∆k ≤ 1.
Definition (Mahler coefficient). Let f ∈ C(Zp , Qp ). Then nth-Mahler coeffi-
cient an (f ) ∈ Qp is defined by the formula
n  
X n
an (f ) = ∆n (f )(0) = (−1)i f (n − i).
i=0
i

We will eventually show that these are the an ’s that appear in Mahler’s
theorem. The first thing to prove is that these coefficients do tend to 0. We
already know that they don’t go up, so we just have to show that they always
eventually go down.
Lemma. Let f ∈ C(Zp , Qp ). Then there exists some k ≥ 1 such that
k 1
k∆p f k ≤ kf k.
p
Proof. If f = 0, there is nothing to prove. So we will wlog kf k = 1 by scaling
(this is possible since the norm is attained at some x0 , so we can just divide by
f (x0 )). We want to find some k such that
k
∆p f (x) ≡ 0 mod p
for all x. To do so, we use the explicit formula
k
p  k
pk
X p i
∆ f (x) = (−1) f (x + pk − i) ≡ f (x + pk ) − f (x) (mod p)
i=0
i
k

because the binomial coefficients pi are divisible by p for i 6= 0, pk . Note that
k
we do have a negative sign in front of f (x) because (−1)p is −1 as long as p is
odd, and 1 = −1 if p = 2.
Now Zp is compact. So f is uniformly continuous. So there is some k such
that |x − y|p ≤ p−k implies |f (x) − f (y)|p ≤ p−1 for all x, y ∈ Zp . So take this
k, and we’re done.
We can now prove that the Mahler’s coefficients tend to 0.
Proposition. The map f 7→ (an (f ))∞
n=0 defines an injective norm-decreasing
linear map C(Zp , Qp ) → c0 .
Proof. First we prove that an (f ) → 0. We know that
kan (f )kp ≤ k∆n f k.
So it suffices to show that k∆n f k → 0. Since k∆k ≤ 1, we know k∆n f k is
monotonically decreasing. So it suffices to find a subsequence that tends to 0.
To do so, we simply apply the lemma repeatedly to get k1 , k2 , · · · such that

≤ 1 kf k.
pk1 +...+kn
∆
pn

40
4 Some p-adic analysis III Local Fields

This gives the desired sequence.

Note that
|an (f )|p ≤ k∆n k ≤ kf k.
So we know
k(an (f ))n k = max |an (f )|p ≤ kf k.
So the map is norm-decreasing. Linearity follows from linearity of ∆. To finish,
we have to prove injectivity.
Suppose an (f ) = 0 for all n ≥ 0. Then

a0 (f ) = f (0) = 0,

and by induction,we have that

f (n) = ∆k f (0) = an (f ) = 0.

for all n ≥ 0. So f is constantly zero on Z≥0 . By continuity, it must be zero

everywhere on Zp .
We are almost at Mahler’s theorem. We have found some coefficients already,
and we want to see that it works. We start by proving a small, familiar, lemma.

Lemma. We have      
x x x+1
+ =
n n−1 n
for all n ∈ Z≥1 and x ∈ Zp .
Proof. It is well known that this is true when x ∈ Z≥n . Since the expressions
are polynomials in x, them agreeing on infinitely many values implies that they
are indeed the same.

Proposition. Let a = (an )∞

n=0 ∈ c0 . We define fa : Zp → Qp by

∞  
X x
fa (x) = an .
n=0
n

This defines a norm-decreasing linear map c0 → C(Zp , Qp ). Moreover an (fa ) =

an for all n ≥ 0.
Proof. Linearity is clear. Norm-decreasing follows from
X    
 x   x 
|fa (x)| = 
 an ≤ sup |an |p   ≤ sup |an |p = kan k,
n  n n p n

where we used the fact that nx ∈ Zp , hence  nx p ≤ 1.




Taking the supremum, we know that

kfa k ≤ kak.

For the last statement, for all k ∈ Z≥0 , we define

a(k) = (ak , ak+1 , ak+1 , · · · ).

41
4 Some p-adic analysis III Local Fields

Then we have

∆fa (x) = fa (x + 1) − fa (x)

∞    
X x+1 x
= an −
n=1
n n
∞  
X x
= an
n=1
n − 1
∞  
X x
= an+1
n=0
n
= fa( 1) (x)

Iterating, we have
∆k fa = fa(k) .
So we have
an (fa ) = ∆n fa (0) = fa(n) (0) = an .

Summing up, we now have maps

F
C(Zp , Qp ) c0
G

with

F (f ) = (an (f ))
G(a) = fa .

We now that F is injective and norm-decreasing, and G is norm-decreasing

and F G = id. It then follows formally that GF = id and the maps are norm-
preserving.

Lemma. Suppose V, W are normed spaces, and F : V → W , G : W → V are

maps such that F is injective and norm-decreasing, and G is norm-decreasing
and F G = idW . Then GF = idV and F and G are norm-preserving.
Proof. Let v ∈ V . Then

F (v − GF v) = F v − F GF v = (F − F )v = 0.

Since F is injective, we have

v = GF v.
Also, we have
kvk ≥ kF vk ≥ kGF vk = kvk.
So we have equality throughout. Similarly, we have kvk = kGvk.
This finishes the proof Mahler’s theorem, and also finishes this section on
p-adic analysis.

42
5 Ramification theory for local fields III Local Fields

5 Ramification theory for local fields

From now on, the characteristic of the residue field of any local field will be
denoted p, unless stated otherwise.

5.1 Ramification index and inertia degree

Suppose we have an extension L/K of local fields. Then since mK ⊆ mL , and
OL ⊆ OL , we obtain an injection
OK OL
kK = ,→ = kL .
mk mL
So we also get an extension of residue fields kL /kK . The question we want to ask
is how much of the extension is “due to” the extension of residue fields kL /kK ,
and how much is “due to” other things happening.
It turns out these are characterized by the following two numbers:
Definition (Inertia degree). Let L/K be a finite extension of local fields. The
inertia degree of L/K is
fL/K = [kL : kK ].
Definition (Ramification index). Let L/K be a finite extension of local fields,
and let vL be the normalized valuation of L and πK a uniformizer of K. The
integer
eL/K = vL (πK )
is the ramification index of L/K.
The goal of the section is to show the following result:
Theorem. Let L/K be a finite extension. Then

[L : K] = eL/K fL/K .

We then have two extreme cases of ramification:

Definition (Unramified extension). Let L/K be a finite extension of local fields.
We say L/K is unramified if eL/K = 1, i.e. fL/K = [L : K].
Definition (Totally ramified extension). Let L/K be a finite extension of local
fields. We say L/K is totally ramified if fL/K = 1, i.e. eL/K = [L : K].
In the next section we will, amongst many things, show that every extension
of local fields can be written as an unramified extension followed by a totally
ramified extension.
Recall the following: let R be a PID and M a finitely-generated R-module.
Assume that M is torsion-free. Then there is a unique integer n ≥ 0 such that
M∼ = Rn . We say n has rank n. Moreover, if N ⊆ M is a submodule, then N is
finitely-generated, so N ∼
= Rm for some m ≤ n.
Proposition. Let K be a local field, and L/K a finite extension of degree n.
Then OL is a finitely-generated and free OK module of rank n, and kL /kK is
an extension of degree ≤ n.
Moreover, L is also a local field.

43
5 Ramification theory for local fields III Local Fields

Proof. Choose a K-basis α1 , · · · , αn of L. Let k · k denote the maximum norm

on L. n
X
xi αi = max |xi |


i=1,...,n
i=1

as before. Again, we know that k · k is equivalent to the extended norm | · | on

L as K-norms. So we can find r > s > 0 such that

M = {x ∈ L : kxk ≤ s} ⊆ OL ⊆ N = {x ∈ L : kxk ≤ r}.

Increasing r and decreasing s if necessary, we wlog r = |a| and s = |b| for some
a, b ∈ K.
Then we can write
n
M n
M
M= Ok bαi ⊆ OL ⊆ N = OK aαi .
i=1 i=1

We know that N is finitely generated and free of rank n over OK , and so is M .

So OL must be finitely generated and free of rank n over OK .
Since mk = mk ∩ OK , we have a natural injection
OK OL
,→ = kL .
mk mL
Since OL is generated over OK by n elements, we know that kK is generated by
n elements over kK , so it has rank at most n.
To see that L is a local field, we know that kL /kK is finite and kK is finite,
so kL is finite. It is complete under the norm because it is a finite-dimensional
vector space over a complete field.
Finally, to see that the valuation is discrete, suppose we have a normalized
valuation on K, and w the unique extension of vK to L. Then we have
1
w(α) = vK (NL/K (α)).
n
So we have
1 1
w(L× ) ⊆ v(K × ) = Z.
n n
So it is discrete.
Note that we cannot just pick an arbitrary
√ basis of L/K√and scale√it to give
a basis of OL /OK . For example, Q2 ( 2)/Q2 has basis 1, 2, but | 2| = √12
and cannot be scaled to 1 by an element in Q2 .
Even if such a scaled basis exists,
√ it doesn’t necessarily give a basis√ of the
integral
√ rings. For example, Q 3 ( −1)/Q 3 has a Q 3 -basis 1, 1 + 3 −1 and
|1 + 3 −1| = 1, but √ √
−1 6∈ Z3 + Z3 (1 + 3 −1).
So this is not a basis of OQ3 (√−1) over Z3 .
Theorem. Let L/K be a finite extension. Then

[L : K] = eL/K fL/K ,

and there is some α ∈ OL such that OL = OK [α].

44
5 Ramification theory for local fields III Local Fields

Proof. We will be lazy and write e = eL/K and f = fL/K . We first note that
kL /kK is separable, so there is some ᾱ ∈ kL such that kL = kK (ᾱ) by the
primitive element theorem. Let

f¯(x) ∈ kK [x]

be the minimal polynomial of ᾱ over kK and let f ∈ OL [x] be a monic lift of f¯

with deg f = deg f¯.
We first claim that there is some α ∈ OL lifting ᾱ such that vL (f (α)) = 1
(note that it is always ≥ 1). To see this, we just take any lift β. If vL (f (β)) = 1,
then we are happy and set α = β. If it doesn’t work, we set α = β + πL , where
πL is the uniformizer of L.
Then we have

f (α) = f (β + πL ) = f (β) + f 0 (β)πL + bπL

2

for some b ∈ OL , by Taylor expansion around β. Since vL (f (β)) ≥ 2 and

vL (f 0 (β)) = 0 (since f¯ is separable, we know f 0 (β) does not vanish when we
reduce mod m), we know vL (f (α)) = 1. So f (α) is a uniformizer of L.
We now claim that the elements αi π j for i = 0, · · · , f − 1 and j = 0, · · · , e − 1
are an OK -basis of OL . Suppose we have
X
aij αi π j = 0
i,j

for some aij ∈ K not all 0. We put

f
X −1
sj = aij αi .
i=0

We know that 1, α, · · · , αf −1 are linearly independent over K since their re-

ductions are linearly independent over kK . So there must be some j such that
sj 6= 0.
The next claim is that if sj 6= 0, then e | vL (sj ). We let k be an index for
which |akj | is maximal. Then we have
f
X −1
a−1
kj sj = a−1 i
kj aij α .
i=0

Now note that by assumption, the coefficients on the right have absolute value
≤ 1, and is 1 when i = k. So we know that

a−1
kj sj 6≡ 0 mod πL ,

because 1, ᾱ, · · · , ᾱf −1 are linearly independent. So we have

vL (a−1
kj sj ) = 0.

So we must have

vL (sj ) = vL (akj ) + vL (a−1 × ×

kj sj ) ∈ vL (K ) = evL (L ) = eZ.

45
5 Ramification theory for local fields III Local Fields

Now we write
X e−1
X
aij αi π j = sj π j = 0.
j=0
j
6 0, then we have vL (sj π ) = vL (sj ) + j ∈ j + eZ. So no two non-zero
If sj =
Pe−1 Pe−1
terms in j=0 sj π j have the same valuation. This implies that j=0 sj π j 6= 0,
which is a contradiction.
We now want to prove that
M
OL = OK αi π j .
i,j

We let M
M= OK αi π j ,
i,j

and put
f −1
M
N= OL α i .
i=0
Then we have
M = N + πN + π 2 N + · · · + π e−1 N.
We are now going to use the fact that 1, ᾱ, · · · , ᾱf −1 span kL over kK . So we
must have that OL = N + πOL . We iterate this to obtain
OL = N + π(N + OL )
= N + πN + π 2 OL
= ···
= N + πN + π 2 N + · · · + π e−1 N + π n OL
= M + πK OL ,
using the fact that πK and π e have the same valuation, and thus they differ by
a unit in OL . Iterating this again, we have
OL = M + πkn OL
for all n ≥ 1. So M is dense in OL . But M is the closed unit ball in the subspace
M
Kαi π j ⊆ l
i,j

with respect to the maximum norm with respect to the given basis. So it must
be complete, and thus M = OL .
Finally, since αi π j = αi f (α)j is a polynomial in α, we know that OL =
OK [α].
Corollary. If M/L/K is a tower of finite extensions of local fields, then
fM/K = fL/K fM/L
eM/K = eL/K eM/L
Proof. The multiplicativity of fM/K follows from the tower law for the residue
fields, and the multiplicativity of eM/K follows from the tower law for the local
fields and that fM/K eM/K = [M : K].

46
5 Ramification theory for local fields III Local Fields

5.2 Unramified extensions

Unramified extensions are easy to classify, since they just correspond to extensions
of the residue field.
Theorem. Let K be a local field. For every finite extension `/kK , there is a
unique (up to isomorphism) finite unramified extension L/K with kL ∼= ` over
kK . Moreover, L/K is Galois with
Gal(L/K) ∼
= Gal(`/kK ).
Proof. We start with existence. Let ᾱ be a primitive element of `/kK with
minimal polynomial f¯ ∈ kK [x]. Take a monic lift f ∈ OK [x] of f¯ such that
deg f = deg f¯. Note that since f¯ is irreducible, we know f is irreducible. So we
can take L = K(α), where α is a root of f (i.e. L = K[x]/f ). Then we have
[L : K] = deg f = deg(f¯) = [` : kK ].
Moreover, kL contains a root of f¯, namely the reduction α. So there is an
embedding ` ,→ kL , sending ᾱ to the reduction of α. So we have
[kL : kK ] ≥ [` : kL ] = [L : K].
So L/K must be unramified and kL ∼= ` over kK .
Uniqueness and the Galois property follow from the following lemma:
Lemma. Let L/K be a finite unramified extension of local fields and let M/K
be a finite extension. Then there is a natural bijection
HomK - Alg (L, M ) ←→ HomkK - Alg (kL , kM )
given in one direction by restriction followed by reduction.
Proof. By the uniqueness of extended absolute values, any K-algebra homomor-
phism ϕ : L ,→ M is an isometry for the extended absolute values. In particular,
we have ϕ(OL ) ⊆ OM and ϕ(mL ) ⊆ mM . So we get an induced kK -algebra
homomorphism ϕ̄ : kL → kM .
So we obtain a map
HomK-Alg (L, M ) → HomkK -Alg (kL , kM )
To see this is bijective, we take a primitive element ᾱ ∈ kL over kK , and take a
minimal polynomial f¯ ∈ kK [x]. We take a monic lift of f¯ to Ok [x], and α ∈ OL
the unique root of f which lifts ᾱ, which exists by Hensel’s lemma. Then by
counting dimensions, the fact that the extension is unramified tells us that
kL = kK (ᾱ), L = K(α).
So we can construct the following diagram:
reduction
ϕ HomK-Alg (L, M ) HomkK -Alg (kL , kM ) ϕ̄
∼
= ∼
=

ϕ(α) {x ∈ M : f (x) = 0}reduction{x̄ ∈ kM : f¯(x̄) = 0} ϕ̄(ᾱ)

But the bottom map is a bijection by Hensel’s lemma. So done.

47
5 Ramification theory for local fields III Local Fields

Alternatively, given a map ϕ̄ : kL → kM , we can lift it to the map ϕ : L → M

given by X  X
ϕ [an ]πkn = [ϕ̄(an )]πkn ,

using the fact that πkn is a uniformizer in L since the extension is unramified. So
we get an explicit inverse.
Proof of theorem (continued). To finish off the proof of the theorem, we just
note that an isomorphism ϕ̄ : kL ∼= kM over kK between unramified extensions.
Then ϕ̄ lifts to a K-embedding ϕ : L ,→ M and [L : K] = [M : K] implies that
ϕ is an isomorphism.
To see that the extension is Galois, we just notice that

| AutK (L)| = | AutkK (kL )| = [kL : kK ] = [L : K].

So L/K is Galois. Moreover, the map AutK (L) → AutkK (kL ) is really a
homomorphism, hence an isomorphism.
Proposition. Let K be a local field, and L/K a finite unramified extension,
and M/K finite. Say L, M are subfields of some fixed algebraic closure K̄ of K.
Then LM/M is unramified. Moreover, any subextension of L/K is unramified
over K. If M/K is unramified as well, then LM/K is unramified.
Proof. Let ᾱ be a primitive element of kK /kL , and f¯ ∈ kK [x] a minimal polyno-
mial of ᾱ, and f ∈ Ok [x] a monic lift of f¯, and α ∈ OL a unique lift of f lifting
ᾱ. Then L = K(α). So LM = M (α).
Let ḡ be the minimal polynomial of ᾱ over kM . Then ḡ | f¯. By Hensel’s
lemma, we can factorize f = gh in OM [x], where g is monic and lifts ḡ. Then
g(α) = 0 and g is irreducible in M [x]. So g is the minimal polynomial of α over
M . So we know that

[LM : M ] = deg g = deg ḡ ≤ [kLM : kM ] ≤ [LM : M ].

So we have equality throughout and LM/M is unramified.

The second assertion follows from the multiplicativity of fL/K , as does the
third.
Corollary. Let K be a local field, and L/K finite. Then there is a unique
maximal subfield K ⊆ T ⊆ L such that T /K is unramified. Moreover, [T : K] =
fL/K .
Proof. Let T /K be the unique unramified extension with residue field extension
kL /kK . Then id : kT = kL → kL lifts to a K-embedding T ,→ L. Identifying T
with its image, we know
[T : K] = fL/K .
Now if T 0 is any other unramified extension, then T 0 T is an unramified extension
over K, so
[T : K] ≤ [T T 0 : K] ≤ fL/K = [T : K].
So we have equality throughout, and T 0 ⊆ T . So this is maximal.

48
5 Ramification theory for local fields III Local Fields

5.3 Totally ramified extensions

We now quickly look at totally ramified extensions. Recall the following irre-
ducibility criterion:
Theorem (Eisenstein criterion). Let K be a local field, and f (x) = xn +
an−1 xn−1 + · · · + a0 ∈ OK [x]. Let πK be the uniformizer of K. If πK |
2
an−1 , · · · , a0 and πK - a0 , then f is irreducible.
Proof. Left as an exercise. You’ve probably seen this already in a much more
general context, but in this case there is a neat proof using Newton polygons.
We will need to use the following characterization of the ramification index:
Proposition. Let L/K be an extension of local fields, and vK be the normalized
valuation. Let w be the unique extension of vK to L. Then the ramification
index eL/K is given by

e−1
L/K = w(πL ) = min{w(x) : x ∈ mL },

Proof. We know w and vL differ by a constant. To figure out what this is, we
have
1 = w(πK ) = e−1
L/K vL (πK ).

So for any x ∈ L, we have

w(x) = e−1
L/K vL (x).

In particular, putting x = πL , we have

w(πL ) = e−1 −1
L/K vL (πL ) = eL/K .

The equality
w(πL ) = min{w(x) : x ∈ mL },
is trivially true because the minimum is attained by πL .
Definition (Eisenstein polynomial). A polynomial f (x) ∈ OK [x] satisfying the
assumptions of Eisenstein’s criterion is called an Eisenstein polynomial .
We can now state the proposition:
Proposition. Let L/K be a totally ramified extension of local fields. Then
L = K(πL ) and the minimal polynomial of πL over K is Eisenstein.
Conversely, if L = K(α) and the minimal polynomial of α over K is Eisenstein,
then L/K is totally ramified and α is a uniformizer of L.
Proof. Let n = [L : K], vK be the valuation of K, and w the unique extension
to L. Then
1
[K(πL ) : K]−1 ≤ e−1
K(πL )/K = min w(c) ≤ ,
x∈mK(πL ) n
where the last inequality follows from the fact that πL ∈ mL(πL ) .
But we also know that

[K(πL ) : K] ≤ [L : K].

49
5 Ramification theory for local fields III Local Fields

So we know that L = K(πL ).

Now let f (x) = xn + an−1 xn−1 + · · · + a0 ∈ OK [x] be the minimal polynomial
of πL /K. Then we have
n n−1
πL = −(a0 + a1 πL + · · · + an−1 πL ).

So we have
 
n n−1 i
1= w(πL ) = w(a0 + a1 πL + · · · + an−1 πL ) = min vk (ai ) + .
i=0,...,n−1 n

This implies that vK (ai ) ≥ 1 for all i, and vK (x0 ) = 1. So it is Eisenstein.

For the converse, if K = K(α) and n = [L : K], take

g(x) = xn + bn−1 xn−1 + .. + b0 ∈ OK [x]

be the minimal polynomial of α. So all roots have the same valuation. So we

have
1 = w(b0 ) = n · w(α).
1
So we have w(α) = n. So we have

1
e−1
L/K = min w(x) ≤ = [L : K]−1 .
x∈mL n
So [L : K] = eL/K = n. So L/K is totally ramified and α is a uniformizer.

In fact, more is true. We have OL = OK [πL ], since every element in OL can

be written as X
i
a i πL ,
i≥0

where ai is a lift of an element in kL = kK , which can be chosen to be in OK .

50
6 Further ramification theory III Local Fields

6 Further ramification theory

6.1 Some filtrations
×
If we have a field K, then we have a unit group UK = OK . We would like to
come up with a filtration of subgroups of the unit group, namely a sequence
(2) (1) (0)
· · · ⊆ UK ⊆ UK ⊆ UK = UK
of subgroups that tells us how close a unit is to being 1. The further down we
are in the chain, the closer we are to being 1.
Similarly, given a field extension L/K, we want a filtration on the Galois
group (the indexing is conventional)
· · · ⊆ G2 (L/K) ⊆ G1 (L/K) ⊆ G0 (L/K) ⊆ G−1 (L/K) = Gal(L/K).
This time, the filtration tells us how close the automorphisms are to being the
identity map.
The key thing about these filtrations is that we can figure out information
(s) (s+1)
about the quotients UK /UK and Gs (L/K)/Gs+1 (L/K), which is often easier.
Later, we might be able to patch these up to get more useful information about
UK and Gal(L/K).
We start with the filtration of the unit group.
Definition (Higher unit groups). We define the higher unit groups to be
(s)
UK = U (s) = 1 + πK
s
OK .
We also put
(0) ×
UK = UK = U (0) = OK .
The quotients of these units groups are surprisingly simple:
Proposition. We have

UK /UK ∼
(1) ×
= (kK , · ),
(s) (s+1) ∼
U /U
K = (kK , +).
K

for s ≥ 1.
× ×
Proof. We have a surjective homomorphism OK → kK which is just reduction
(1)
mod πK , and the kernel is just things that are 1 modulo πK , i.e. UK . So this
gives the first part.
(s)
For the second part, we define a surjection UK → kK given by
s
1 + πK x 7→ x mod πk .
This is a group homomorphism because
s s
(1 + πK x)(1 + πK y) = 1 + π S (x + y + π s xy),
and this gets mapped to
x + y + πs x + y ∼
=x+y mod πK .
(s+1)
Then almost by definition, the kernel is UK .

51
6 Further ramification theory III Local Fields

The next thing to consider is a filtration of the Galois group.

Definition (Higher ramification group). Let L/K be a finite Galois extension
of local fields, and vL the normalized valuation of L.
Let s ∈ R≥−1 . We define the sth ramification group by
Gs (L/K) = {σ ∈ Gal(L/K) : vL (σ(x) − x) ≥ s + 1 for all x ∈ OL }.
So if you belong to Gs for a large s, then you move things less. Note that we
could have defined these only for s ∈ Z≥−1 , but allowing fractional indices will
be helpful in the future.
Now since σ(x) − x ∈ OL for all x ∈ OL , we know
G−1 (L/K) = Gal(L/K).
We next consider the case of G0 (L/K). This is, by definition
G0 (L/K) = {σ ∈ Gal(L/K) : vL (σ(x) − x) ≥ 1 for all x ∈ OL }
= {σ ∈ Gal(L/K) : σ(x) ≡ x mod m for all x ∈ OL }.
In other words, these are all the automorphisms that reduce to the identity when
we reduce it to Gal(kL /kK ).
Definition (Inertia group). Let L/K be a finite Galois extension of local fields.
Then the inertia group of L/K is the kernel of the natural homomorphism
Gal(L/K) → Gal(kL /kK )
given by reduction. We write this as
I(L/K) = G0 (L/K).
Proposition. Let L/K be a finite Galois extension of local fields. Then the
homomorphism
Gal(L/K) → Gal(kL /kK )
given by reduction is surjective.
Proof. Let T /K be maximal unramified subextension. Then by Galois theory,
the map Gal(L/K) → Gal(T /K) is a surjection. Moreover, we know that
kT = kL . So we have a commutative diagram

Gal(L/K) Gal(kL /kK )

∼
Gal(T /K) Gal(kT /kK ).

So the map Gal(L/K) → Gal(kL /kK ) is surjective.

Then the inertia group is trivial iff L/K is unramified. The field T is
sometimes called the inertia field .
Lemma. Let L/K be a finite Galois extension of local fields, and let σ ∈ I(L/K).
Then σ([x]) = [x] for all x.
More generally, let x ∈ kL and σ ∈ Gal(L/K) with image σ̄ ∈ Gal(kL /kK ).
Then we have
[σ̄(x)] = σ([x]).

52
6 Further ramification theory III Local Fields

Proof. Consider the map kL → OL given by

f : x 7→ σ −1 ([σ̄(x)]).

This is multiplicative, because every term is multiplicative, and

σ −1 ([σ̄(x)]) ≡ x mod πL .

So this map f has to be the Teichmüller lift by uniqueness.

That’s all we’re going to say about the inertia group. We now consider the
general properties of this filtration.
Proposition. Let L/K be a finite Galois extension of local fields, and vL the
normalized valuation of L. Let πL be the uniformizer of L. Then Gs+1 (L/K) is
a normal subgroup of Gs (L/K) for s ∈ Z≥0 , and the map
(s)
Gs (L/K) UL
→ (s+1)
Gs+1 (L/K) UL

given by
σ(πL )
σ 7→
πL
is a well-defined injective group homomorphism, independent of the choice of
πL .
Proof. We define the map
(s)
UL
φ : Gs (L/K) → (s+1)
UL
σ 7→ σ(πL )/πL .

We want to show that this has kernel Gs+1 (L/K).

First we show it is well-defined. If σ ∈ Gs (L/K), we know
s+1
σ(πL ) = πL + πL x

for some x ∈ OL . So we know

σ(πL ) s (s)
= 1 + πL x ∈ UL .
πL
So it has the right image. To see this is independent of the choice of πL , we let
× s+1
u ∈ OL . Then σ(u) = u + πL y for some y ∈ OL .
Since any other uniformizer must be of the form πL u, we can compute
s+1 s+1
σ(πL u) (πL + πL )(u + πL y)
=
πL u πL u
s s+1 −1
= (1 + πL x)(1 + πL u y)
s
≡ 1πL x (mod ULs+1 ).
(s) (s+1)
So they represent the same element in in UL /UL .

53
6 Further ramification theory III Local Fields

To see this is a group homomorphism, we know

σ(τ (πL )) σ(τ (πL )) τ (πL )

φ(στ ) = = = φ(σ)φ(t),
πL τ (πL ) πL

using the fact that τ (πL ) is also a uniformizer.

Finally, we have to show that ker φ = Gs+1 (L/K). We write down

ker φ = {σ ∈ Gs (L/K) : vL (σ(πL ) − πL ) ≥ s + 2}.

On the other hand, we have

Gs+1 (L/K) = {σ ∈ Gs (L/K) : vL (σ(z) − z) ≥ s + 2 for all z ∈ OL }.

So we trivially have Gs+1 (L/K) ⊆ ker φ. To show the converse, let x ∈ OL and
write
X∞
n
x= [xn ]πL .
n=0

Take σ ∈ ker φ ⊆ Gs (L/K) ⊆ I(L/K). Then we have

s+2
σ(πL ) = πL + πL y, y ∈ OL .

Then by the previous lemma, we know

∞
X
σ(x) − x = [xn ] ((σ(πL ))n − πL
n
)
n=1
X∞
s+2 n n


= [xn ] (πL + πL y) − πL
n=1
s+2
= πL (things).

So we know vL (σ(x) − x) ≥ s + 2.
Corollary. Gal(L/K) is solvable.
Proof. Note that \
Gs (L/K) = {id}.
s

So (Gs (L/K))s∈Z≥−1 is a subnormal series of Gal(L/K), and all quotients are

(s)
UL ∼
abelian, because they embed into (s+1) = (kK , +) (and s = −1 can be checked
UL
separately).
Thus if L/K is a finite extension of local fields, then we have, for s ≥ 1,
injections
Gs (L/K)
,→ kL .
Gs+1 (L/K)
Since kL is a p-group, it follows that

|Gs (L/K)|
|Gs+1 (L/K)|

54
6 Further ramification theory III Local Fields

is a pth power. So it follows that for any t, the quotient

|G1 (L/K)|
|Gt (L/K)|

is also a pth power. However, we know that the intersection of all Gs (L/K)
is {id}, and also Gal(L/K) is finite. So for sufficiently large t, we know that
|Gt (L/K)| = 1. So we conclude that
Proposition. G1 (L/K) is always a p-group.
We now use the injection
G0 (L/K) ×
,→ kL ,
G1 (L/K)
×
and the fact that kL has order prime to p. So G1 (L/K) must be the Sylow
p-subgroup of G0 (L/K). Since it is normal, it must be the unique p-subgroup.
Definition (Wild inertia group and tame quotient). G1 (L/K) is called the wild
inertia group, and the quotient G0 (L/K)/G1 (L/K) is the tame quotient.

6.2 Multiple extensions

Suppose we have tower M/L/K of finite extensions of local fields. How do the
ramification groups of the different extensions relate? We first do the easy case.
Proposition. Let M/L/K be finite extensions of local fields, and M/K Galois.
Then
Gs (M/K) ∩ Gal(M/L) = Gs (M/L).
Proof. We have

Gs (M/K) = {σ ∈ Gal(M/L) : vM (σx − x) ≥ s + 1} = Gs (M/K) ∩ Gal(M/L).

This is trivial, because the definition uses the valuation vM of the bigger field
all the time. What’s more difficult and interesting is quotients, namely going
from M/K to L/K.
We want to prove the following theorem:
Theorem (Herbrand’s theorem). Let M/L/K be finite extensions of local fields
with M/K and L/K Galois. Then there is some function ηM/L such that

Gs (M/K)
Gt (L/K) ∼
=
Gs (M/L)

for all s, where t = ηM/L (s).

To better understand the situation, it helps to provide an alternative charac-
terization of the Galois group.
Definition (iL/K ). We define

iL/K (σ) = min vL (σ(x) − x).

x∈OL

55
6 Further ramification theory III Local Fields

It is then immediate that

Gs (L/K) = {σ ∈ Gal(L/K) : iL/K (σ) ≥ s + 1}.

This is not very helpful. We now claim that we can compute iL/K using the
following formula:
Proposition. Let L/K be a finite Galois extension of local fields, and pick
α ∈ OL such that OL = OK [α]. Then

iL/K (σ) = vL (σ(α) − α).

Proof. Fix a σ. It is clear that iL/K (σ) ≤ vL (σ(α) − α). Conversely, for any
x ∈ OL , we can find a polynomial g ∈ OK [t] such that
X
x = g(α) = bi α i ,

where bi ∈ OK . In particular, bi is fixed by σ.

Then we have

vL (σ(x) − x) = vL (σg(α) − g(α))

n
!
X
i i
= vL bi (σ(α) − α )
i=1
≥ vL (σ(α) − α),

using the fact that σ(α) − α | σ(α)i − αi for all i. So done.

Now if M/L/K are finite Galois extensions of local fields, then OM = OK [α]
implies OM = OL [α]. So for σ ∈ Gal(M/L), we have

iM/L (σ) = iM/K (σ).

Going in the other direction is more complicated.

Proposition. Let M/L/K be a finite extension of local fields, such that M/K
and L/K are Galois. Then for σ ∈ Gal(L/K), we have
X
iL/K (σ) = e−1
M/L iM/K (τ ).
τ ∈Gal(M/K)
τ |L =σ

Here eM/L is just to account for the difference between vL and vM . So the
real content is that the value of iL/K (σ) is the sum of the values of iM/K (τ ) for
all τ that restrict to σ.
Proof. If σ = 1, then both sides are infinite by convention, and equality holds.
So we assume σ 6= 1. Let OM = OL [α] and OL = OK [β], where α ∈ OM and
β ∈ OL . Then we have

eM/L iL/K (σ) = eM/L vL (σβ − β) = vM (σβ − β).

Now if τ ∈ Gal(M/K), then

iM/K (τ ) = vM (τ α − α)

56
6 Further ramification theory III Local Fields

Now fix a τ such that τ |L = σ. We set H = Gal(M/L). Then we have

 
X X Y
iM/K (τ 0 ) = vM (τ g(α) − α) = vM  (τ g(α) − α) .
τ 0 ∈Gal(M/K),τ 0 |L =σ g∈H g∈H

We let
b = σ(β) − β = τ (β) − β
and Y
a= (τ g(α) − α).
g∈H

We want to prove that vM (b) = vM (a). We will prove that a | b and b | a.

We start with a general observation about elements in OL . Given z ∈ OL ,
we can write
Xn
z= zi β i , zi ∈ OK .
i=1

Then we know
n
X
τ (z) − z = zi (τ (β)i − β i )
i=1

is divisible by τ (β) − β = b.
Now let F (x) ∈ OL [x] be the minimal polynomial of α over L. Then explicitly,
we have Y
F (x) = (x − g(α)).
g∈H

Then we have Y
(τ F )(x) = (x − τ g(α)),
g∈H

where τ F is obtained from F by applying τ to all coefficients of F . Then all

coefficients of τ F − F are of the form τ (z) − z for some z ∈ OL . So it is divisible
by b. So b divides every value of this polynomial, and in particular
Y
b | (τ F − F )(α) = (α − g(α)) = ±a,
g∈H

So b | a.
In other direction, we pick f ∈ OK [x] such that f (α) = β. Then f (α)−β = 0.
This implies that the polynomial f (x) − β divides the minimal polynomial of α
in OL [x]. So we have
f (x) − β = F (x)h(x)
for some h ∈ OL [x].
Then noting that f has coefficients in OK , we have

(f − τ β)(x) = (τ f − τ b)(x) = (τ F )(x)(τ h)(x).

Finally, set x = α. Then

−b = β − τ β = ±a(τ h)(α).

So a | b.

57
6 Further ramification theory III Local Fields

Now that we understand how the iL/K behave when we take field extensions,
we should be able to understand how the ramification groups behave!
We now write down the right choice of ηL/K : [−1, ∞) → [−1, ∞):
!
X
−1
ηL/K (s) = eL/K min(iL/K (σ), s + 1) − 1.
σ∈G

Theorem (Herbrand’s theorem). Let M/L/K be a finite extension of local

fields with M/K and L/K Galois. We set

H = Gal(M/L), t = ηM/L (s).

Then we have
Gs (M/K)H
= Gt (L/K).
H
By some isomorphism theorem, and the fact that H ∩ Gs (M/K) = Gs (M/L),
this is equivalent to saying
Gs (M/K)
Gt (L/K) ∼
= .
Gs (M/L)

Proof. Let G = Gal(M/K). Fix a σ ∈ Gal(L/K). We let τ ∈ Gal(M/K) be an

extension of σ to M that maximizes iM/K , i.e.

iM/K (τ ) ≥ iM/K (τ g)

for all g ∈ H. This is possible since H is finite.

We claim that

iL/K (σ) − 1 = ηM/L (iM/K (τ ) − 1).

If this were true, then we would have

Gs (M/K)H
σ∈ ⇔ τ ∈ Gs (M/K)
H
⇔ iM/K (τ ) − 1 ≥ s

Since ηM/L is strictly increasing, we have

⇔ ηM/L (iM/K (τ ) − 1) ≥ ηM/L (s) = t

⇔ iL/K (σ) − 1 ≥ t
⇔ σ ∈ Gt (L/K),

and we are done.

To prove the claim, we now use our known expressions for iL/K (σ) and
ηM/L (iM/K (τ ) − 1) to rewrite it as
X X
e−1
M/L iM/K (τ g) = e−1M/L min(iM/L (g), iM/K (τ )).
g∈H g∈H

We then make the stronger claim

iM/K (τ g) = min(iM/L (g), iM/K (τ )).

58
6 Further ramification theory III Local Fields

We first note that

iM/K (τ g) = vM (τ g(α) − α)
= vM (τ g(α) − g(α) + g(α) − α)
≥ min(vM (τ g(α) − g(α)), vM (g(α) − α))
= min(iM/K (τ ), iM/K (g))
We cannot conclude our (stronger) claim yet, since we have a ≥ in the middle.
We now have to split into two cases.
(i) If iM/K (g) ≥ iM/K (τ ), then the above shows that iM/K (τ g) ≥ iM/K (τ ).
But we also know that it is bounded above by m. So iM/K (τ g) = iM/K (τ ).
So our claim holds.
(ii) If iM/K (g) < iM/K (τ ), then the above inequality is in fact an equality as
the two terms have different valuations. So our claim also holds.
So done.
We now prove an alternative characterization of the function ηL/K , using a
funny integral.
Proposition. Write G = Gal(L/K). Then
Z s
dx
ηL/K (s) = .
0 (G0 (L/K) : Gx (L/K))

When −1 ≤ x < 0, our convention is that

1
= (Gx (L/K) : G0 (L/K)),
(G0 (L/K) : Gx (L/K))
which is just equal to 1 when −1 < x < 0. So
ηL/K (s) = s if − 1 ≤ s ≤ 0.
Proof. We denote the RHS by θ(s). It is clear that both ηL/K (s) and θ(s) are
piecewise linear and the break points are integers (since iL/K (σ) is always an
integer). So to see they are the same, we see that they agree at a point, and
that they have equal derivatives. We have
|{σ ∈ G : iL/K (σ) ≥ 1}|
ηL/K (0) = − 1 = 0 = θ(0),
eL/K
since the numerator is the size of the inertia group.
If s ∈ [−1, ∞) \ Z, then
0 −1
ηL/K (s) = eL/K (|{σ ∈ G : iL/K (σ) ≥ s + 1}|)
|Gs (L/K)|
=
|G0 (L/K)|
1
=
(G0 (L/K) : Gs (L/K))
= θ0 (s).
So done.

59
6 Further ramification theory III Local Fields

We now tidy up the proof by inventing a different numbering of the ramifica-

tion groups. Recall that

ηL/K : [−1, ∞) → [−1, ∞)

is continuous, strictly increasing, and

ηL/K (−1) = −1, ηL/K (s) → ∞ as s → ∞.

So this is invertible. We set

Notation.
−1
ψL/K = ηL/K .

Definition (Upper numbering). Let L/K be a Galois extension of local fields.

Then the upper numbering of the ramification groups of L/K is defined by

Gt (L/K) = GψL/K (t) (L/K)

for t ∈ [−1, ∞). The original number is called the lower numbering.

To rephrase our previous theorem using the upper numbering, we need a

little lemma:
Lemma. Let M/L/K be a finite extension of local fields, and M/K and L/K
be Galois. Then
ηM/K = ηL/K ◦ ηM/L .
Hence
ψM/K = ψM/L ◦ ψL/K .
Proof. Let s ∈ [−1, ∞), and let t = ηM/L (s), and H = Gal(M/L). By Her-
brand’s theorem, we know

Gs (M/K)H ∼ Gs (M/K) Gs (M/K)

Gt (L/K) ∼
= = = .
H H ∩ Gs (M/K) Gs (M/L)

Thus by multiplicativity of the inertia degree, we have

|Gs (M/K)| |Gt (L/K)| |Gs (M/L)|

= .
eM/K eL/K eM/L

By the fundamental theorem of calculus, we know that whenever the derivatives

make sense, we have
0 |Gs (M/K)|
ηM/K (s) = .
eM/K
So putting this in, we know
0 0 0
ηM/K (s) = ηL/K (t)ηM/L (s) = (ηL/K ◦ ηM/L )0 (s).

Since ηM/K and ηL/K ◦ ηM/L agree at 0 (they both take value 0), we know that
the functions must agree everywhere. So done.

60
6 Further ramification theory III Local Fields

Corollary. Let M/L/K be finite Galois extensions of local fields, and H =

Gal(M/L). Let t ∈ [−1, ∞). Then
Gt (M/K)H
= Gt (L/K).
H
Proof. Put s = ηL/K (t). Then by Herbrand’s theorem, we have

Gt (M/K)H GψM/K (t) (M/K)H

=
H H
∼
= GηM/L (ψM/K (t)) (L/K)
= Gs (L/K)
= Gt (L/K).

This upper numbering might seem like an unwieldy beast that was invented
just so that our theorem looks nice. However, it turns out that often the upper
numberings are rather natural, as we could see in the example below:
Example. Consider ζpn a primitive pn th root of unity, and K = Qp (ζpn ). The
minimal polynomial of ζpn is the pn th cyclotomic polynomial
n−1 n−1
Φpn (x) = xp (p−1)
+ xp (p−2)
+ · · · + 1.

It is an exercise on the example sheet to show that this is indeed irreducible.

So K/Qp is a Galois extension of degree pn−1 (p − 1). Moreover, it is totally
ramified by question 6 on example sheet 2, with uniformizer

π = ζpn − 1

is a uniformizer. So we know

OK = Zp [ζpn − 1] = Zp [ζpn ].

We then have an isomorphism

 ×
Z
n
→ Gal(L/Qp )
p Z
obtained by sending m → σm , where

σm (ζpn ) = ζpmn .

We have

iK/Qp (σm ) = vK (σm (ζpn ) − ζpn )

= vK (ζpmn − ζpn )
= vK (ζpm−1
n − 1)

since ζpn is a unit. If m = 1, then this thing is infinity. If it is not 1, then ζpm−1
n

is a primitive pn−k th root of unity for the maximal k such that pk | m − 1. So

by Q6 on example sheet 2, we have
pn−1 (p − 1)
vK (ζpm−1
n − 1) = = pk .
pn−k−1 (p − 1)

61
6 Further ramification theory III Local Fields

Thus we have
vK (ζpm−1
n − 1) ≥ pk ⇔ m ≡ 1 mod pk .
It then follows that for
pk ≥ s + 1 ≥ pk−1 + 1,
we have
Gs (K/Qp ) ∼
= {m ∈ (Z/pn )× : m ≡ 1 mod pk }.
Now m ≡ 1 mod pk iff σm (ζpk ) = ζpk . So in fact

Gs (K/Qp ) ∼
= Gal(K/Qp (ζpk )).

Finally, when s ≥ pn − 1, we have

Gs (K/Qp ) = 1.

We claim that
ηK/Qp (pk − 1) = k.
So we have
Gk (K/Qp ) = Gal(K/Qp (ζpk )).
This actually looks much nicer!
To actually compute ηK/Qp , we have notice that the function we integrate to
get η looks something like this (not to scale):

1
p−1

1
p(p−1)
1
p2 (p−1)

p−1 p2 − 1 p3 − 1

The jumps in the lower numbering are at pk − 1 for k = 1, · · · , n − 1. So we have

1 1
ηK/Qp (pk − 1) = (p − 1) + ((p2 − 1) − (p − 1))
p−1 p(p − 1)
1
+ · · · + ((pk − 1) − (pk−1 − 1)) k−1
p (p − 1)
= k.

62
7 Local class field theory III Local Fields

7 Local class field theory

Local class field theory is the study of abelian extensions of local fields, i.e. a
Galois extension whose Galois group is abelian.

7.1 Infinite Galois theory

It turns out that the best way of formulating this theory is to not only use finite
extensions, but infinite extensions as well. So we need to begin with some infinite
Galois theory. We will mostly just state the relevant results instead of proving
them, because this is not a course on Galois theory.
In this section, we will work with any field.
Definition (Separable and normal extensions). Let L/K be an algebraic exten-
sion of fields. We say that L/K is separable if, for every α ∈ L, the minimal
polynomial fα ∈ K[α] is separable. We say L/K is normal if fα splits in L for
every α ∈ L.

Definition (Galois extension). Let L/K be an algebraic extension of fields.

Then it is Galois if it is normal and separable. If so, we write

Gal(L/K) = AutK (L).

These are all the same definitions as in the finite case.

In finite Galois theory, the subgroups of Gal(L/K) match up with the
intermediate extensions, but this is no longer true in the infinite case. The
Galois group has too many subgroups. To fix this, we need to give Gal(L/K) a
topology, and talk about closed subgroups.
Definition (Krull topology). Let M/K be a Galois extension. We define the
Krull topology on M/K by the basis

{Gal(M/L) : L/K is finite}.

More explicitly, we say that U ⊆ Gal(M/K) is open if for every σ ∈ U , we can

find a finite subextension L/K of M/K such that σ Gal(M/L) ⊆ U .

Note that any open subgroup of a topological group is automatically closed,

but the converse does not hold.
Note that when M/K is finite, then the Krull topology is discrete, since we
can just take the finite subextension to be M itself.
Proposition. Let M/K be a Galois extension. Then Gal(M/K) is compact
and Hausdorff, and if U ⊆ Gal(M/K) is an open subset such that 1 ∈ U , then
there is an open normal subgroup N ⊆ Gal(M/K) such that N ⊆ U .
Groups with properties in this proposition are known as profinite groups.
Proof. We will not prove the first part.
For the last part, note that by definition, there is a finite subextension of
M/K such that Gal(M/L) ⊆ U . We then let L0 be the Galois closure of L over
K. Then Gal(M/L0 ) ⊆ Gal(M/L) ⊆ U , and Gal(M/L0 ) is open and normal.

63
7 Local class field theory III Local Fields

Recall that we previously defined the inverse limit of a sequence rings. More
generally, we can define such an inverse limit for any sufficiently nice poset
of things. Here we are going to do it for topological groups (for those doing
Category Theory, this is the filtered limit of topological groups).
Definition (Directed system). Let I be a set with a partial order. We say that
I is a directed system if for all i, j ∈ I, there is some k ∈ I such that i ≤ k and
j ≤ k.

Example. Any total order is a directed system.

Example. N with divisibility | as the partial order is a directed system.
Definition (Inverse limit). Let I be a directed system. An inverse system (of
topological groups) indexed by I is a collection of topological groups Gi for each
i ∈ I and continuous homomorphisms

fij : Gj → Gi

for all i, j ∈ I such that i ≤ j, such that

fii = idGi

and
fik = fij ◦ fjk
whenever i ≤ j ≤ k.
We define the inverse limit on the system (Gi , fij ) to be
( )
Y Y
lim Gi = (gi ) ∈ Gi : fij (gj ) = gi for all i ≤ j ⊆ gi ,
←−
i∈I i∈I i∈I

which is a group under coordinate-wise multiplication and

Qa topological space
under the subspace topology of the product topology on i∈I Gi . This makes
limi∈I Gi into a topological group.
←−
Proposition. Let M/K be a Galois extension. The set I of finite Galois
subextensions L/K is a directed system under inclusion. If L, L0 ∈ I and L ⊆ L0 ,
then we have a restriction map
0
0
· |L
L : Gal(L /K) → Gal(L/K).
0
Then (Gal(L/K), · |L
L ) is an inverse system, and the map

Gal(M/K) → lim Gal(L/K)

←−
i∈I
σ 7→ (σ|L )i∈I

is an isomorphism of topological groups.

We now state the main theorem of Galois theory.

64
7 Local class field theory III Local Fields

Theorem (Fundamental theorem of Galois theory). Let M/K be a Galois ex-

tension. Then the map L 7→ Gal(M/L) defines a bijection between subextensions
L/K of M/K and closed subgroups of Gal(M/K), with inverse given by sending
H 7→ M H , the fixed field of H.
Moreover, L/K is finite if and only if Gal(M/L) is open, and L/K is Galois
iff Gal(M/L) is normal, and then
Gal(L/K)
→ Gal(L/K)
Gal(M/L)
is an isomorphism of topological groups.
Proof. This follows easily from the fundamental theorem for finite field extensions.
We will only show that Gal(M/L) is closed and leave the rest as an exercise. We
can write [
L= L0 .
L0 ⊆L
L0 /K finite

Then we have \
Gal(M/L) = Gal(M/L0 ),
0
L ⊆L
L0 /K finite

and each Gal(M/L0 ) is open, hence closed. So the whole thing is closed.

7.2 Unramified extensions and Weil group

We first define what it means for an infinite extension to be unramified or totally
ramified. To do so, we unexcitingly patch up the definitions for finite cases.
Definition (Unramified extension). Let K be a local field, and M/K be alge-
braic. Then M/K is unramified if L/K is unramified for every finite subextension
L/K of M/K.
Note that since the extension is not necessarily finite, in general M will not
be a local field, since chances are its residue field would be infinite.
Definition (Totally ramified extension). Let K be a local field, and M/K be
algebraic. Then M/K is totally ramified if L/K is totally ramified for every
finite subextension L/K of M/K.
Proposition. Let M/K be an unramified extension of local fields. Then M/K
is Galois, and
Gal(M/K) ∼ = Gal(kM /kK )
via the reduction map.
Proof. Every finite subextension of M/K is unramified, so in particular is Galois.
So M/K is Galois (because normality and separability is checked for each
element). Then we have a commutative diagram
reduction
Gal(M/K) Gal(kM /kK )
∼ ∼
reduction
lim Gal(L/K) ∼ lim Gal(kL /kK )
←− ←−
L/K L/K

65
7 Local class field theory III Local Fields

The left hand map is an isomorphism by (infinite) Galois theory, and since all
finite subextensions of kM /kK are of the form kL /kK by our finite theory, we
know the right-hand map is an isomorphism. The bottom map is an isomorphism
since it is an isomorphism in each component. So the top map must be an
isomorphism.
Since the compositor of unramified extensions is unramified, it follows that
any algebraic extension M/K has a maximal unramified subextension

T = TM/K /K.

In particular, every field K has a maximal unramified extension K ur .

We now try to understand unramified extensions. For a finite unramified
extension L/K, we have an isomorphism
∼
Gal(L/K) Gal(kL /kK ) ,

By general field theory, we know that Gal(kL /kK ) is a cyclic group generated by

FrobL/K : x 7→ xq ,

where q = |kK | is the size of kK . So by the isomorphism, we obtain a generator

of Gal(L/K).
Definition (Arithmetic Frobenius). Let L/K be a finite unramified extension
of local fields, the (arithmetic) Frobenius of L/K is the lift of FrobL/K ∈
Gal(kL /kK ) under the isomorphism Gal(L/K) ∼ = Gal(kL /kK ).
There is also a geometric Frobenius, which is its inverse, but we will not use
that in this course.
We know Frob is compatible in towers, i.e. if M/L/K is a tower of finite
unramified extension of local fields, then FrobM/K |L = FrobL/K , since they both
reduce to the map x 7→ x|kK | in Gal(kL /kK ), and the map between Gal(kL /kK )
and Gal(L/K) is a bijection.
So if M/K is an arbitrary unramified extension, then we have an element

(FrobL/K ) ∈ lim Gal(L/K) ∼

= Gal(M/K).
←−
L/K

So we get an element FrobM/K ∈ Gal(M/K). By tracing through the proof of

Gal(M/K) ∼ = Gal(kM /kK ), we see that this is the unique lift of x 7→ x|kK | .
Note that while for finite unramified extensions M/K, the Galois group is
generated by the Frobenius, this is not necessarily the case when the extension is
infinite. However, powers of the Frobenius are the only things we want to think
about, so we make the following definition:
Definition (Weil group). Let K be a local field and M/K be Galois. Let
T = TM/K be the maximal unramified subextension of M/K. The Weil group
of M/K is

W (M/K) = {σ ∈ Gal(M/K) : σ|T = FrobnT /K for some n ∈ Z}.

We define a topology on W (M/K) by saying that U is open iff there is a finite

extension L/T such that σ Gal(L/T ) ⊆ U .

66
7 Local class field theory III Local Fields

In particular, if M/K is unramified, then W (M/K) = FrobZT /K .

It is helpful to put these groups into a diagram of topological groups to see
what is going on.

Gal(M/T ) W (M/K) FrobZT /K

Gal(M/T ) Gal(M/K) Gal(T /K)

Here we put the discrete topology on the subgroup generated by the Frobenius.
The topology of W (M/K) is then chosen so that all these maps are continuous
homomorphisms of groups.
In many ways, the Weil group works rather like the Galois group.
Proposition. Let K be a local field, and M/K Galois. Then W (M/K) is dense
in Gal(M/K). Equivalently, for any finite Galois subextension L/K of M/K,
the restriction map W (M/K) → Gal(L/K) is surjective.
If L/K is a finite subextension of M/K, then
W (M/L) = W (M/K) ∩ Gal(M/L).
If L/K is also Galois, then
W (M/K) ∼
= Gal(L/K)
W (M/L)
via restriction.
Proof. We first prove density. To see that density is equivalent to W (M/K) →
Gal(L/K) being surjective for all finite subextension L/K, note that by the
topology on Gal(M/K), we know density is equivalent to saying that W (M/K)
hits every coset of Gal(M/L), which means that W (M/K) → Gal(L/K) is
surjective.
Let L/K be a subextension. We let T = TM/K . Then TL/K = T ∩ L. Then
we have a diagram

Gal(M/T ) W (M/K) FrobZT /K

Gal(L/T ∩ L) Gal(L/K) Gal(T ∩ L/K)

Here the surjectivity of the left vertical arrow comes from field theory, and the
right hand vertical map is surjective because T ∩ L/K is finite and hence the
Galois group is generated by the Frobenius. Since the top and bottom rows are
short exact sequences (top by definition, bottom by Galois theory), by diagram
chasing (half of the five lemma), we get surjectivity in the middle.

To prove the second part, we again let L/K be a finite subextension. Then
L · TM/K ⊆ TM/L . We then have maps
∼
=
FrobZTM/K /K Gal(TM/K /K) Gal(kM /kK )

∼
=
FrobZTM/L /L Gal(TM/L /L) Gal(kM /kL )

67
7 Local class field theory III Local Fields

So the left hand vertical map is an inclusion. So we know

FrobZTM/L /L = FrobZTM/K /K ∩ Gal(TM/L /L).

Now if σ ∈ Gal(M/L), then we have

σ ∈ W (M/L) ⇔ σ|TM/L /L ∈ FrobZTM/L /L
⇔ σ|TM/K /K ∈ FrobZTM/K /K
⇔ σ ∈ W (M/K).
So this gives the second part.

Now L/K is Galois as well. Then Gal(M/L) is normal in Gal(M/K). So

W (M/L) is normal in W (M/K) by the second part. Then we can compute
W (M/K) W (M/K)
=
W (M/L) W (M/K) ∩ Gal(M/L)
∼ W (M/K) Gal(M/L)
=
Gal(M/L)
Gal(M/K)
=
Gal(M/L)
∼ Gal(L/K).
=
The only non-trivial part in this chain is the assertion that W (M/K) Gal(M/L) =
Gal(M/K), i.e. that W (M/K) hits every coset of Gal(M/L), which is what
density tells us.

7.3 Main theorems of local class field theory

We now come to the main theorems of local class field theory.
Definition (Abelian extension). Let K be a local field. A Galois extension
L/K is abelian if Gal(L/K) is abelian.
We will fix an algebraic closure K̄ of K, and all algebraic extensions we will
consider will be taken to be subextensions of K̄/K. We let K sep be the separable
closure of K inside K̄.
If M/K and M/K are Galois extensions, then LM/K is Galois, and the map
given by restriction
Gal(LM/K) ,→ Gal(L/K) × Gal(M/K).
is an injection. In particular, if L/K and M/K are both abelian, then so is
LM/K. This implies that there is a maximal abelian extension K ab .
Finally, note that we know an example of an abelian extension, namely the
maximal unramified extension K ur = TK sep /K ⊆ K ab , and we put FrobK =
FrobK ur /K .
Theorem (Local Artin reciprocity). There exists a unique topological isomor-
phism
ArtK : K × → W (K ab /K)
characterized by the properties

68
7 Local class field theory III Local Fields

(i) ArtK (πK )|K ur = FrobK , where πK is any uniformizer.

(ii) We have
ArtK (NL/K (x))|L = idL
for all L/K finite abelian and x ∈ L× .

Moreover, if M/K is finite, then for all x ∈ M × , we know ArtM (x) is an

automorphism of M ab /M , and restricts to an automorphisms of K ab /K. Then
we have ab
ArtM (x)|K
K = ArtK (NM/K (x)).
Moreover, ArtK induces an isomorphism

K× M ∩ K ab
 
→ Gal .
NM/K (M × ) K

To simplify this, we will write N (L/K) = NL/K (L× ) for L/K finite. From
this theorem, we can deduce a lot of more precise statements.
Corollary. Let L/K be finite. Then N (L/K) = N ((L ∩ K ab )/K), and

(K × : N (L/K)) ≤ [L : K]

with equality iff L/K is abelian.

Proof. To see this, we let M = L ∩ K ab . Applying the isomorphism twice gives

K× ∼ K×
= Gal(M/K) ∼
= .
N (L/K) N (M/K)

Since N (L/K) ⊆ N (M/K), and [L : K] ≥ [M : K] = | Gal(M/K)|, we are

done.
The theorem tells us if we have a finite abelian extension M/K, then we
obtain an open finite-index subgroup NM/K (M × ) ≤ K × . Conversely, if we are
given an open finite index subgroup of K × , we might ask if there is an abelian
extension of K whose norm group is corresponds to this subgroup. The following
theorem tells us this is the case:
Theorem. Let K be a local field. Then there is an isomorphism of posets
   
open finite index finite abelian
subgroups of K × extensions of L/K
.
H (K ab )ArtK (H)

N (L/K) L/K

In particular, for L/K and M/K finite abelian extensions, we have

N (LM/K) = N (L/K) ∩ N (M/K),

N (L ∩ M/K) = N (L/K)N (M/K).

69
7 Local class field theory III Local Fields

While proving this requires quite a bit of work, a small part of it follows from
local Artin reciprocity:

Theorem. Let L/K be a finite extension, and M/K abelian. Then N (L/K) ⊆
N (M/K) iff M ⊆ L.
Proof. By the previous theorem, we may wlog L/K abelian by replacing with
L ∩ K ab . The ⇐ direction is clear by the last part of Artin reciprocity.
For the other direction, we assume that we have N (L/K) ⊆ N (M/K), and
let σ ∈ Gal(K ab /L). We want to show that σ|M = idM . This would then imply
that M is a subfield of L by Galois theory.
We know W (K ab /L) is dense in Gal(K ab /L). So it suffices to show this for
σ ∈ W (K ab /L). Then we have

W (K ab /L) ∼
= ArtK (N (L/K)) ⊆ ArtK (N (M/K)).

So we can find x ∈ M × such that σ = ArtK (NM/K (x)). So σ|M = idM by local
Artin reciprocity.
Side note: Why is this called “class field theory”? Usually, we call the field
corresponding to the subgroup H the class field of H. Historically, the first type
of theorems like this are proved for number fields. The groups that appear on
the left would be different, but in some cases, they are the class group of the
number field.

70
8 Lubin–Tate theory III Local Fields

8 Lubin–Tate theory
For the rest of the course, we will indicate how one can explicitly construct the
field K ab and the map ArtK .
There are many ways we can approach local class field theory. The approach
we use, using Lubin–Tate theory, is the most accessible one. Another possible
approach is via Galois cohomology. This, however, relies on more advanced
machinery, namely Galois cohomology.

8.1 Motivating example

We will work out the details of local Artin reciprocity in the case of Qp as a
motivating example for the proof we are going to come up with later. Here we
will need the results of local class field theory to justify our claims, but this is
not circular since this is not really part of the proof.
Lemma. Let L/K be a finite abelian extension. Then we have
× ×
eL/K = (OK : NL/K (OL )).

Proof. Pick x ∈ L× , and w the valuation on L extending vK , and n = [L : K].

Then by construction of w, we know

vK (NL/K (x)) = nw(x) = fL/K vL (x).

So we have a surjection

K× vK Z
.
N (L/K) fL/K Z

The kernel of this map is equal to

× × ×
OK N (L/K) ∼ OK OK
= × = × .
N (L/K) OK ∩ N (L/K) NL/K (OL )

So by local class field theory, we know

× ×
n = (K × : N (L/K)) = fL/K (OK : NL/K (OL )),

and this implies what we want.

Corollary. Let L/K be a finite abelian extension. Then L/K is unramified if
× ×
and only if NL/K (OL ) = OK .
Now we fix a uniformizer πK . Then we have a topological group isomorphism

K× ∼ ×
= hπK i × OK .

Since we know that the finite abelian extensions correspond exactly to finite
index subgroups of K × by taking the norm groups, we want to understand
subgroups of K × . Now consider the subgroups of K × of the form
m (n)
hπK i × UK .

71
8 Lubin–Tate theory III Local Fields

We know these form a basis of the topology of K × , so it follows that finite-index

open subgroups must contain one of these guys. So we can find the maximal
abelian extension as the union of all fields corresponding to these guys.
Since we know that N (LM/K) = N (L/K) ∩ N (M/K), it suffices to further
specialize to the cases
(n)
hπK i × UK
and
m
hπK i × OK
separately. The second case is easy, because this corresponds to an unramified
extension by the above corollary, and unramified extensions are completely
characterized by the extension of the residue field. Note that the norm group and
the extension are both independent of the choice of uniformizer. The extensions
corresponding to the first case are much more difficult to construct, and they
depend on the choice of πK . We will get them from Lubin–Tate theory.
Lemma. Let K be a local field, and let Lm /K be the extension corresponding
m
to hπK i × OK . Let [
L= Lm .
m

Then we have
K ab = K ur L,
Lemma. We have isomorphisms

W (K ab /K) ∼
= W (K ur L/K)
∼
= W (K ur /K) × Gal(L/K)
∼
= FrobZK × Gal(L/K)

Proof. The first isomorphism follows from the previous lemma. The second
follows from the fact that K ur ∩ L = K as L is totally ramified. The last
isomorphism follows from the fact that TK ur /K = K ur trivially, and then by
definition W (K ur/K ) ∼
= FrobZK .
Example. We consider the special case of K = Qp and πK = p. We let

Ln = Qp (ζpn ),

where ζpn is the primitive pn th root of unity. Then by question 6 on example

sheet 2, we know this is a field with norm group
(n)
N (Qp (ζpn )/Qp ) = hpi × (1 + pn Zp ) = hpi × UQp ,

and thus this is a totally ramified extension of Qp .

We put
∞
[
Qp (ζp∞ ) = Qp (ζpn ).
n=1

Then again this is totally ramified extension, since it is the nested union of
totally ramified extensions.

72
8 Lubin–Tate theory III Local Fields

Then we have

Gal(Qp (ζp∞ )/Qp ) ∼

= lim Gal(Qp (ζpn )/Qp )
← −
n
= lim(Z/pn Z)×
← −
n
= Z×
p.

Note that we are a bit sloppy in this deduction. While we know that it is true
that Z× ∼ (Z/pn Z)× , the inverse limit depends not only on the groups
p = lim
←−n
n ×
(Z/p Z) themselves, but also on the maps we use to connect the groups together.
Fortunately, from the discussion below, we will see that the maps

Gal(Qp (ζpn )/Qp ) → Gal(Qp (ζpn−1 )/Qp )

indeed correspond to the usual restriction maps

(Z/pn Z)× → (Z/pn−1 Z)× .

It is a fact that this is the inverse of the Artin map of Qp restricted to Z×

p.
Note that we have W (Qp (ζp∞ )/Qp ) = Gal(Qp (ζp∞ )/Qp ) because its maximal
unramified subextension is trivial.
We can trace through the above chains of isomorphisms to figure out what
the Artin map does. Let m = Z× p . Then we can write

m = a0 + a1 p + · · · ,

where ai ∈ {0, · · · , p − 1} and a0 6= 0. Now for each n, we know

m ≡ a0 + a1 p + · · · + an−1 pn−1 mod pn .

By the usual isomorphism Gal(Qp (ζpn )/Qp ) ∼

= Z/pn Z, we know m acts as
a +a1 p+...+an−1 pn−1
ζpn 7→ ζpn0 “=” ζpmn

on Qp (ζpn ), where we abuse notation because taking ζpn to powers of p greater

than n gives 1. It can also be interpreted as (1 + λpn )m , where λpn = ζpn − 1 is
a uniformizer, which makes sense using binomial expansion.
So the above isomorphisms tells us that ArtQp restricted to Z× p acts on
Qp (ζp∞ ) as
−1
ArtQp (m)(ζpn ) ≡ σm−1 (ζpn ) = ζpmn .
The full Artin map can then be read off from the following diagram:
ArtQp
Q×
p W (Qab
p /Qp )

∼
= ∼ restriction

hpi × Z×
p W (Qur
p /Qp ) × Gal(Qp (ζp∞ )/Qp )

where the bottom map sends

hpn , mi 7→ (FrobnQp , σm−1 ).

73
8 Lubin–Tate theory III Local Fields

In fact, we have
Theorem (Local Kronecker-Weber theorem).
[
Qab
p = Qp (ζn ),
n∈Z≥1
[
Qur
p = Qp (ζn ).
n∈Z≥1
(n,p)=1

Not a proof. We will comment on the proof of the generalized version later.
Remark. There is another normalization of the Artin map which sends a
uniformizer to the geometric Frobenius, defined to be the inverse of the arithmetic
Frobenius. With this convention, ArtQp (m)|Qp (ζp∞ ) is σm .
We can define higher ramification groups for general Galois extensions.
Definition (Higher ramification groups). Let K be a local field and L/K Galois.
We define, for s ∈ R≥−1

Gs (M/K) = {σ ∈ Gal(M/K) : σ|L ∈ Gs (L/K) for all finite

Galois subextension M/K}.

This definition makes sense, because the upper number behaves well when
we take quotients. This is one of the advantages of upper numbering. Note that
we can write the ramification group as the inverse limit

Gs (M/K) ∼
= lim Gs (L/K),
←−
L/K

as in the case of the Galois group.

Example. Going back to the case of K = Qp . We write Qpn for the unramified
extension of degree n of Qp . By question 11 of example sheet 3, we know that

Gal(Qpn (ζpm )/Qp )
 s = −1

Gal(Q n (ζ m )/Q n )
p p p −1 < s ≤ 0
Gs (Qpn (ζpm )/Qp ) = ,


 Gal(Q p n (ζpm )/ζ k )
p k−1<s≤k ≤m−1
1 s>m−1


which corresponds to

hpi × U (0)
s = −1





 hpn i × U (m)
n (0)
 hp i × U



n (m)
−1 < s ≤ 0
hp i × U
hpn i × U (k)


k−1<s≤k ≤m−1


hpn i × U (m)





1 s>m−1

under the Artin map.

74
8 Lubin–Tate theory III Local Fields

By taking the limit as n, m → ∞, we get

Theorem. We have

Gs (Qab k
p /Qp ) = ArtQp (1 + p Zp ) = ArtQp (U
(k)
),

where k is chosen such that k − 1 < s ≤ k, k ∈ Z≥0 .

Corollary. If L/Qp is a finite abelian extension, then

N (L/Qp )(1 + pn Zp )
 
Gs (L/Qp ) = ArtQp ,
N (L/Qp )

where n − 1 < s ≤ n.
Here ArtQp induces an isomorphism

Q×
p
→ Gal(L/Qp ).
N (L/Qp )

So it follows that L ⊆ Qp (ζpm ) for some n if and only if Gs (L/Qp ) = 1 for all
s > m − 1.

8.2 Formal groups

The proof of local Artin reciprocity will be done by constructing the analogous
versions of Ln for an arbitrary local field, and then proving that it works. To
do so, we will need the notion of a formal group. The idea of a formal group is
that a formal group is a rule that specifies how we should multiply two elements
via a power series over a ring R. Then if we have a complete R-module, then
the formal group will turn the R-module into an actual group. There is then a
natural notion of a formal module, which is a formal group F with an R-action.
At the end, we will pick R = OK . The idea is then that we can fix an
algebraic closure K̄, and then a formal OK -module will turn mK̄ into an actual
OK -module. Then if we adjoin the right elements of mK̄ to K, then we obtain
an extension of K with a natural OK action, and we can hope that this restricts
to field automorphisms when we restrict to the unit group.

Notation. Let R be a ring. We write

 
 X 
R[[x1 , · · · , xn ]] = ak1 ,...,kn xk11 · · · xknn : ak1 ,...,kn ∈ R
 
k1 ,...,kn ∈Z≥0

for the ring of formal power series in n variables over R.

Definition (Formal group). A (one-dimensional, commutative) formal group
over R is a power series F (X, Y ) ∈ R[X, Y ] such that
(i) F (X, Y ) ≡ X + Y mod (X 2 , XY, Y 2 )
(ii) Commutativity: F (X, Y ) = F (Y, X)
(iii) Associativity: F (X, F (Y, Z)) = F (F (X, Y ), Z).

75
8 Lubin–Tate theory III Local Fields

This is most naturally understood from the point of view of algebraic geometry,
as a generalization of the Lie algebra over a Lie group. Instead of talking about
the tangent space of a group (the “first-order neighbourhood”), we talk about its
infinitesimal (formal) neighbourhood, which contains all higher-order information.
A lot of the seemingly-arbitrary compatibility conditions we later impose have
such geometric motivation that we unfortunately cannot go into.
Example. If F is a formal group over OK , where K is a complete valued field,
then F (x, y) converges for all x, y ∈ mK . So mK becomes a (semi)group under
the multiplication
(x, y) 7→ F (x, y) ∈ mk
Example. We can define
Ĝa (X, Y ) = X + Y.
This is called the formal additive group.
Similarly, we can have
Ĝm (X, Y ) = X + Y + XY.
This is called the formal multiplicative group. Note that
X + Y + XY = (1 + X)(1 + Y ) − 1.
So if K is a complete valued field, then mK bijects with 1 + mk by sending
x 7→ 1 + x, and the rule sending (x, y) ∈ m2K 7→ x + y + xy ∈ mK is just the
usual multiplication in 1 + mK transported to mK via the bijection above.
We can think of this as looking at the group in a neighbourhood of the
identity 1.
Note that we called this a formal group, rather than a formal semi-group. It
turns out that the existence of identity and inverses is automatic.
Lemma. Let R be a ring and F a formal group over R. Then
F (X, 0) = X.
Also, there exists a power series i(X) ∈ X · R[[X]] such that
F (X, i(X)) = 0.
Proof. See example sheet 4.
The next thing to do is to define homomorphisms of formal groups.
Definition (Homomorphism of formal groups). Let R be a ring, and F, G be
formal groups over R. A homomorphism f : F → G is an element f ∈ R[[X]]
such that f (X) ≡ 0 mod X and
f (F (X, Y )) = G(f (X), f (Y )).
The endomorphisms f : F → F form a ring EndR (F ) with addition +F given by
(f +F g)(x) = F (f (x), g(x)).
and multiplication is given by composition.

76
8 Lubin–Tate theory III Local Fields

We can now define a formal module in the usual way, plus some compatibility
conditions.
Definition (Formal module). Let R be a ring. A formal R-module is a formal
group F over R with a ring homomorphism R → EndR (F ), written, a 7→ [a]F ,
such that
[a]F (X) = aX mod X 2 .
Those were all general definitions. We now restrict to the case we really care
about. Let K be a local field, and q = |kK |. We let π ∈ OK be a uniformizer.
Definition (Lubin–Tate module). A Lubin–Tate module over OK with respect
to π is a formal OK -module F such that
[π]F (X) ≡ X q mod π.
We can think of this condition of saying “uniformizer corresponds to the
Frobenius”.
Example. The formal group Ĝm is a Lubin–Tate Zp module with respect to p
given by the following formula: if a ∈ Zp , then we define
∞  
X a
[a]Ĝm (X) = (1 + X)a − 1 = X n.
n=1
n

The conditions
(1 + X)a − 1 ≡ aX mod X 2
and
(1 + X)p − 1 ≡ X p mod p
are clear.
We also have to check that a 7→ [a]F is a ring homomorphism. This follows
from the identities
((1 + X)a )b = (1 + X)ab , (1 + X)a (1 + X)b = (1 + X)ab ,
which are on the second example sheet.
The objective of the remainder of the section is to show that all Lubin–Tate
modules are isomorphic.
Definition (Lubin–Tate series). A Lubin–Tate series for π is a power series
e(X) ∈ OK [[X]] such that
e(X) ≡ πX mod X 2 , e(X) ≡ X q mod π.
We denote the set of Lubin–Tate series for π by Eπ .
Now by definition, if F is a Lubin–Tate OK module for π, then [π]F is a
Lubin–Tate series for π.
Definition (Lubin–Tate polynomial). A Lubin–Tate polynomial is a polynomial
of the form
uX q + π(aq−1 X q−1 + · · · + a2 X 2 ) + πX
(1)
with u ∈ UK , and aq−1 , · · · , a2 ∈ OK .
In particular, these are Lubin–Tate series.

77
8 Lubin–Tate theory III Local Fields

Example. X q + πX is a Lubin–Tate polynomial.

Example. If K = Qp and π = p, then (1 + X)p − 1 is a Lubin–Tate polynomial.
The result that allows us to prove that all Lubin–Tate modules are isomorphic
is the following general result:
Lemma. Let e1 , e2 ∈ Eπ and take a linear form
n
X
L(x1 , · · · , xn ) = ai Xi , ai ∈ OK .
i=1

Then there is a unique power series F (x1 , · · · , xn ) ∈ OK [[x1 , · · · , xn ]] such that

F (x1 , · · · , xn ) ≡ L(x1 , · · · , xn ) mod (x1 , · · · , xn )2 ,

and
e1 (F (x1 , · · · , xn )) = F (e2 (x1 ), e2 (x2 ), · · · , e2 (xn )).
For reasons of time, we will not prove this. We just build F by successive
approximation, which is not terribly enlightening.
Corollary. Let e ∈ Eπ be a Lubin–Tate series. Then there are unique power
series Fe (X, Y ) ∈ OK [[X, Y ]] such that

Fe (X, Y ) ≡ X + Y mod (X + Y )2
e(Fe (X, Y )) = Fe (e(X), e(Y ))

Corollary. Let e1 , e2 ∈ Eπ be Lubin–Tate series and a ∈ OK . Then there exists

a unique power series [a]e1 ,e2 ∈ OK [[X]] such that

[a]e1 ,e2 (X) ≡ aX mod X 2

e1 ([a]e1 ,e2 (X)) = [a]e1 ,e2 (e2 (X)).

To simplify notation, if e1 = e2 = e, we just write [a]e = [a]e,e .

We now state the theorem that classifies all Lubin–Tate modules in terms of
Lubin–Tate series.
Theorem. The Lubin–Tate OK modules for π are precisely the series Fe for
e ∈ Eπ with formal OK -module structure given by

a 7→ [a]e .

Moreover, if e1 , e2 ∈ Eπ and a ∈ OK , then [a]e1 ,e2 is a homomorphism from

Fe2 → Fe1 .
×
If a ∈ OK , then it is an isomorphism with inverse [a−1 ]e2 ,e1 .
So in some sense, there is only one Lubin–Tate module.
Proof sketch. If F is a Lubin–Tate OK -module for π, then e = [π]F ∈ Eπ by
definition, and F satisfies the properties that characterize the series Fe . So
F = Fe by uniqueness.
For the remaining parts, one has to verify the following for all e, e1 , e2 , e3 ∈ Eπ
and a, b ∈ OK .

78
8 Lubin–Tate theory III Local Fields

(i) Fe (X, Y ) = Fe (Y, X).

(ii) Fe (X, Fe (Y, Z)) = Fe (Fe (X, Y ), Z).
(iii) [a]e1 ,e2 (Fe (X, Y )) = Fe1 ([a]e1 ,e2 (X), [a]e1 ,e2 (Y )).
(iv) [ab]e1 ,e3 (X) = [a]e1 ,e2 ([b]e2 ,e3 (X)).
(v) [a + b]e1 ,e2 (X) = [a]e1 ,e2 (X) + [b]e1 ,e2 (X).
(vi) [π]e (X) = e(X).
The proof is just repeating the word “uniqueness” ten times.

8.3 Lubin–Tate extensions

We now use the Lubin–Tate modules to do things. As before, we fixed an
algebraic closure K̄ of K. We let m̄ = mK̄ be the maximal ideal in OK̄ .
Proposition. If F is a formal OK -module, then m̄ becomes a (genuine) OK
module under the operations +F and ·
x +F y = F (x, y)
a · x = [a]F (x)
for all x, y ∈ m̄ and a ∈ OK .
We denote this m̄F .
This isn’t exactly immediate, because K̄ need not be complete. However,
this is not a problem as each multiplication given by F only involves finitely
many things (namely two of them).
Proof. If x, y ∈ m̄, then F (x, y) is a series in K(x, y) ⊆ K̄. Since K(x, y) is
a finite extension, we know it is complete. Since the terms in the sum have
absolute value < 1 and → 0, we know it converges to an element in mK(x,y) ⊆ m̄.
The rest then essentially follows from definition.
(n)
To prove local class field theory, we want to find elements with an UK /UK
(n)
action for each n, or equivalently elements with an OK /OK action. Note that
the first quotient is a quotient of groups, while the second quotient is a quotient
of a ring by an ideal. So it is natural to consider the following elements:
Definition (π n -division points). Let F be a Lubin–Tate OK -module for π. Let
n ≥ 1. The group F (n) of π n -division points of F is defined to be
F (n) = {x ∈ m̄F | [π n ]F x = 0} = ker([π n ]F ).
This is a group under the operation given by F , and is indeed an OK module.
Example. Let F = Ĝm , K = Qp and π = p. Then for x ∈ m̄Ĝm , we have
n
pn · x = (1 + x)p − 1.
So we know
Ĝm (n) = {ζpi n − 1 | i = 0, 1, · · · , pn − 1},
where ζpn ∈ Q̄p is the primitive pn th root of unity.
So Ĝm (n) generates Qp (ζpn ).

79
8 Lubin–Tate theory III Local Fields

To prove this does what we want, we need the following lemma:

Lemma. Let e(X) = X q + πX. We let

fn (X) = (e ◦ · · · ◦ e)(X).
| {z }
n times

Then fn has no repeated roots. Here we take f0 to be the identity function.

Proof. Let x ∈ K̄. We claim that if |fi (x)| < 1 for i = 0, · · · , n − 1, then
fn0 (X) 6= 0.
We proceed by induction on n.
(i) When n = 1, we assume |x| < 1. Then
 q 
f10 (x) = e0 (x) = qxq−1 + π = π 1 + xq−1 6= 0,
π
since we know πq has absolute value ≤ 1 (q vanishes in kK , so q/π lives in
OK ), and xq−1 has absolute value < 1.
(ii) in the induction step, we have

0
 q 
fn+1 (x) = (qfn (x)q−1 + π)fn0 (x) = π 1 + fn (x)q−1 fn0 (x).
π
By induction hypothesis, we know fn0 (x) 6= 0, and by assumption |fn (x)| <
1. So the same argument works.
We now prove the lemma. We assume that fn (x) = 0. We want to show that
|fi (x)| < 1 for all i = 0, · · · , n − 1. By induction, we have
n
fn (x) = xq + πgn (x)

for some gn (x) ∈ OK [x]. It follows that if fn (x) = 0, then |x| < 1. So |fi (x)| < 1
for all i. So fn0 (x) 6= 0.
The point of the lemma is to prove the following proposition:
Proposition. F (n) is a free OK /π n OK module of rank 1. In particular, it has
q n elements.
Proof. By definition, we know

π n · F (n) = 0.

So F (n) is indeed an OK /π n OK -module.

To prove that it is free of rank 1, we note that all Lubin–Tate modules
for π are isomorphic. This implies that all the honest OK modules F (n) are
isomorphic. We choose F = Fe , where e = X q + πX. Then F (n) consists
of the roots of the polynomial fn = en (X), which is of degree q n and has no
repeated roots. So |F (n)| = q n . To show that it is actually the right thing, if
λn ∈ F (n) \ F (n − 1), then we have a homomorphism

OK → F (n)

80
8 Lubin–Tate theory III Local Fields

given by A 7→ a · λn . Its kernel is π n OK by our choice of λn . By counting, we

get an OK -module isomorphism
OK
→ F (n)
π n OK
as desired.
Corollary. We have isomorphisms
OK ∼
= EndOK (F (n))
π n OK
UK ∼
(n)
= AutOK (F (n)).
UK

Given a Lubin–Tate OK -module F for π, we consider

Ln,π = Ln = K(F (n)),

which is the field of π n division points of F . From the inclusions F (n) ⊆ F (n + 1)

for all n, we obtain a corresponding inclusion of fields

Ln ⊆ Ln+1 .

The field Ln depends only in π, and not on F . To see this, we let G be another
Lubin–Tate OK -module, and let f : F → G be an isomorphism. Then

G(n) = f (F (n)) ⊆ K(F (n))

since the coefficients of f lie in K. So we know

K(G(n)) ⊆ K(F (n)).

By symmetry, we must have equality.

Theorem. Ln /K is a totally ramified abelian extension of degree q n−1 (q − 1)
with Galois group
UK
Gal(Ln /K) ∼
= AutOK (F (n)) ∼
= (n) .
UK
(n)
Explicitly, for any σ ∈ Gal(Ln /K), there is a unique u ∈ UK /UK such that

σ(λ) = [u]F (λ)

for all λ ∈ F (n). Under this isomorphism, for m ≥ n, we have

(n)
U
Gal(Lm /Ln ) ∼
= K (m)
.
UK

Moreover, if F = Fe , where

e(X) = X q + π(aq−1 π q−1 + · · · + a2 X 2 ) + πX,

81
8 Lubin–Tate theory III Local Fields

and λn ∈ F (n) \ F (n − 1), then λn is a uniformizer of Ln and

en (X) n−1
φn (X) = = X q (q−1) + · · · + π
en−1 (X)
is the minimal polynomial of λn . In particular,
NLn /K (−λn ) = π.
Proof. Consider a Lubin–Tate polynomial
e(X) = xq + π(aq−1 X q−1 + · · · + a2 X 2 ) + πX.
We set F = Fe . Then
en (X)
φn (X) = = (en−1 (X))q−1 + π(aq1 en−1 (X)q−2 + · · · + a2 en−1 (X)) + π
en−1 (X)
is an Eisenstein polynomial of degree q n−1 (q − 1) by starting at it long enough.
So if λn ∈ F (n) \ F (n − 1), then λn is a root of φn (x), so K(λn )/K is totally
ramified of degree q n−1 (q − 1), and λn is a uniformizer, and
NK(λn )/K (−λn ) = π
as the norm is just the constant coefficient of the minimal polynomial.
Now let σ ∈ Gal(Ln /K). Then σ induces a permutation of F (n), as these
are the roots of en (X), which is in fact OK -linear, i.e.
σ(x) +F σ(y) = F (σ(x), σ(y)) = σ(F (x, y)) = σ(x +F y)
σ(a · x) = σ([a]F (x)) = [a]F (σ(x)) = a · σ(x)
for all x, y ∈ mLn and a ∈ OK .
So we have an injection of groups
UK
Gal(Ln /K) ,→ AutOK (F (n)) = (n)
UK
But we know
 
U 
 K 
 (n)  = q n−1 (q − 1) = [K(λn ) : K] ≤ [Ln : K] = | Gal(Ln /K)|.
U 
K

So we must have equality throughout, the above map is an isomorphism, and

K(λn ) = Ln .
It is clear from the construction of the isomorphism that for m ≥ n, the
diagram
∼ (m)
Gal(Lm /K) UK /UK
restriction quotient

∼ (n)
Gal(Ln /K) UK /UK
commutes. So the isomorphism
(m)
UK
Gal(Lm /Ln ) ∼
= (n)
UK
follows by looking at the kernels.

82
8 Lubin–Tate theory III Local Fields

Example. In the case where K = Qp and π = p, recall that

Ĝm (n) = {ζpi n − 1 | i = 0, · · · , pn−1 − 1},

where ζpn is the principal pn th root of unity. The theorem then gives
Gal(Qp (ζpn )/Qp ) ∼
= (Z/pn )×
given by if a ∈ Z≥0 and (a, p) = 1, then
σa (ζpi n − 1) = [a]Ĝm (n) (ζpi n − 1) = (1 + (ζpi n − 1))a − 1 = ζpain − 1.
This agrees with the isomorphism we previously constructed.
Back to the general situation, setting
∞
[
L∞ = Ln ,
n=1

we know L∞ /K is Galois, and we have isomorphisms

∼ ∼ (n) ∼
Gal(L∞ /K) lim Gal(Ln /K) limn UK /UK = UK
←− ←−
σ (σ|Ln )n

This map will be the inverse of the Artin map restricted to L∞ .

To complete the proof of Artin reciprocity, we need to use the following
theorem without proof:
Theorem (Generalized local Kronecker-Weber theorem). We have
K ab = K ur L∞
(for any π).
Comments on the proof. One can prove this from the Hasse-Arf theorem, which
states that in an abelian extension, the jumps in the upper ramification groups
occur only at integer values. This, together with the calculation of ramification
groups done later, easily implies the theorem. Essentially, L∞ maxed out all
possible jumps of the upper ramification groups. However, the Hasse-Arf theorem
is difficult to prove.
Another approach is to prove the existence of the Artin map using other
techniques (e.g. Galois cohomology). Consideration of the norm group (cf. the
next theorem) then implies the theorem. The content of this section then
becomes an explicit construction of a certain family of abelian extensions.
We can characterize the norm group by
Theorem. We have
(n)
N (Ln /K) = hπi × Uk .
Comments on the proof. This can be done by defining Coleman operators, which
are power series representations of the norm. Alternatively, assuming the
(n)
description of the local Artin map given below and local Artin reciprocity, Uk
(n)
is in the kernel of Art|Ln , so hπi × Uk ⊆ N (Ln /K). The result follows by
comparing order.

83
8 Lubin–Tate theory III Local Fields

We can then construct the Artin map as follows:

Theorem. Let K be a local field. Then we have an isomorphism Art : K × →
W (K ab /K) given by the composition
Art
K× W (K ab /K)
∼ ∼

hπi × UK FrobZK × Gal(L∞ /K)

where the bottom map is given by (π m , u) 7→ (Frobm

K , σu−1 ), where

σu (λ) = [u]F (λ)

S∞
for all λ ∈ n=1 F (n).
The inverse shows up in the proof to make sure the map defined above is
independent of the choice of uniformizer. We will not prove this, nor that the
map obtained has the desired properties. Instead, we will end the course by
computing the higher ramification groups of these extensions.
Theorem. We have

Gal(Ln /K) −1 ≤ s ≤ 0

Gs (Ln /K) = Gal(Ln /Lk ) q k−1 − 1 < s ≤ q k − 1, 1 ≤ k ≤ n − 1

1 s > q n−1


Proof. The case for −1 ≤ s ≤ 0 is clear.

For 0 ≤ s ≤ 1 (which we may wlog is actually 1), we know that

Gal(Ln /Lk ) ∼
(k) (n)
= UK /UK

under the isomorphism Gal(Ln /K) ∼

(n)
= UK /UK . On the other hand, we know
G1 (Ln /K) is the Sylow p-subgroup of Gal(Ln /K). So we must have

G1 (Ln /K) ∼
(1) (n)
= UK /UK .
So we know that G1 (Ln /K) = Gal(Ln /L1 ). Thus we know that Gs (Ln /K) =
Gal(Ln /K) for 0 < s ≤ 1.

(1) (n)
We now let σ = σu ∈ G1 (Ln /K) and u ∈ UK /UK . We write
u = 1 + επ k
for some ε ∈ UK and some k = k(u) ≥ 1. Since σ is not the identity, we know
k < n. We claim that
iLn /K (σ) = vLn (σ(λ) − λ) = q k .
Indeed, we let λ ∈ F (n) \ F (n − 1), where F is a choice of Lubin–Tate module
for π. Then λ is a uniformizer of Ln and OLn = OK [λ]. We can compute
σu (λ) = [u]F (λ)
= [1 + επ k ]F (λ)
= F (λ, [επ k ]F (λ))

84
8 Lubin–Tate theory III Local Fields

Now we can write

[επ k ]F (λ) = [ε]F ([π k ]F (λ)) ∈ F (n − k) \ F (n − k − 1),
since [ε]F is invertible, and applying [π n−k ]F to [π k ]F (λ) kills it, but applying
[π n−k−1 ]F gives [π n−1 ]F , which does not kill.
So we know [επ k ]F (λ) is a uniformizer of Ln−k . Since Ln /Ln−k is totally
×
ramified of degree q k , we can find ε0 ∈ OL n
such that
k
[επ k ]F (λ) = ε0 λq
Recall that F (X, 0) = X and F (0, Y ) = Y . So we can write
F (X, Y ) = X + Y + XY G(X, Y ),
where G(X, Y ) ∈ OK [[X, Y ]]. So we have
σ(λ) − λ = F (λ, [επ k ]F (λ)) − λ
k
= F (λ, ε0 λq ) − λ
k k k
= λ + ε0 λq + ε0 λq +1
G(λ, ε0 λq ) − λ
k k k
= ε0 λq + ε0 λq +1
G(λ, ε0 λq ).
In terms of valuation, the first term is the dominating term, and
iLn /K (σ) = vLn (σ(λ) − λ) = q k
So we know
iLn /K (σk ) ≥ s + 1 ⇔ q k(u) − 1 ≥ s.
So we know
Gs (Ln /K) = {σK ∈ G1 (Ln /K) : q k(u) − 1 ≥ s} = Gal(Ln /Lk ),
where q k−1 − 1 < s ≤ q k − 1 for k = 1, · · · , n − 1, and 1 if s > q n−1 = 1.
Corollary. We have

Gal(Ln /K) −1 ≤ t ≤ 0

t
G (Ln /K) = Gal(Ln /Lk ) k − 1 < t ≤ k, k = 1, · · · , n − 1

1 t>n−1


In other words, we have

(
t Gal(Ln /Ldte ) −1 ≤ t ≤ n − 1
G (Ln /K) = ,
1 t>n−1

where we set L0 = K.
Once again, the numbering is a bit more civilized in the upper numbering.
Proof. We have to compute the integral of
1
.
(G0 (Ln /K) : Gx (Ln /K)
We again plot this out

85
8 Lubin–Tate theory III Local Fields

1
q−1

1
q(q−1)
1
q 2 (q−1)

q−1 q2 − 1 q3 − 1

So by the same computation as the ones we did last time, we find that

s

 −1 ≤ s ≤ 0
s−(q k−1 −1)
ηLn /K (s) = (k − 1) + q k−1 (q−1)
q k−1 − 1 ≤ s ≤ q k − 1, k = 1, · · · , n − 1
s−(q n−1 −1)

(n − 1) + n−1
− 1.

q n−1 (q−1) s>q

Inverting this, we find that


t
 −1 ≤ t ≤ 0
ψLn /K = q dte−1 (q − 1)(t − (dte − 1)) + q dte−1 − 1 1<t≤n−1.

 n−1
q (q − 1)(t − (n − 1)) + q n−1 − 1 t>n−1
Then we have
Gt (Ln /K) = Gψ(Ln /K)(t) (Ln /K),
which gives the desired by the previous theorem.
So we know that
(
dte (n)
UK /UK −1 ≤ t ≤ n
Art−1 t
K (G (Ln /K)) = .
1 t≥n
Corollary. When t > −1, we have
Gt (K ab/K ) = Gal(K ab /K ur Ldte ),
and
Art−1 t ab
K (G (K /K)) = U
(dte)
.
Proof. Recall the following fact from the examples class: If L/K is finite un-
ramified and M/K is finite totally ramified, then LM/L is totally ramified, and
Gal(LM/L) ∼ = Gal(M/K) by restriction, and
Gt (LM/K) ∼
= Gt (M/K).
via this isomorphism (for t > −1).
Now let Km /K be the unramified extension of degree m. By the lemma and
the previous corollary, we have
(
t
G (Km Ln /K) = ∼ G (Ln /K) = Gal(Ln /Ldte ) −1 < t ≤ n
t
1 t≥n
(
Gal(Km Ln /Km Ldte ) −1 < t ≤ n
=
1 t≥n

86
8 Lubin–Tate theory III Local Fields

So we have

Gt (K ab /K) = Gt (K ur L∞ /K)
= lim Gt (Km Ln /K)
←−
m,n

= lim Gal(Km Ln /Km Ldte )

←−
m,n
n≥dte

= Gal(K ur L∞ /K ur Ldte )
= Gal(K ab /K ur Ldte ),

and
 

Art−1 ab ur −1 
K (Gal(K /K Ldte )) = ArtK  lim Gal(Km Ln /Km Ldte )

←−
m,n
n≥dte

= lim Art−1


← − K Gal(Km Ln /Km Ldte )
m,n
n≥dte
(dte)
U
= lim K(n)
←−
m,n UK
n≥dte

= U dte .

Corollary. Let M/K be a finite abelian extension. Then we have an isomor-

phism
K× ∼ Gal(M/K).
ArtK : =
N (M/K)
Moreover, for t > −1, we have
!
(dte)
UK N (M/K)
Gt (M/K) = ArtK .
N (M/K)

Proof. We have
!
(dte)
t Gt (K ab /K)G(K ab /M ) UK N (M/K)
G (M/K) = = Art .
G(K ab /M ) N (M/K)

87
Index III Local Fields

Index

C(Zp , Qp ), 39 DVR, 29
I-adic completion, 11
I-adic topology, 9 Eisenstein criterion, 49
I-adically complete, 11 Eisenstein polynomial, 49
I-adically open, 9 equal characteristic, 29
K ur , 66 equivalence of norm, 19
L/K, 52 equivalent absolute values, 4
N (L/K), 69
(s) formal additive group, 76
Uk , 51
formal group, 75
W (M/K), 66
homomorphism, 76
W (R), 34
module, 77
ηL/K , 58
formal Laurent series, 15
OK , 15
formal module, 77
mK , 16
formal multiplicative group, 76
π n -division points, 79
fundamental theorem of Galois
ψL/K , 60
theory, 65
c0 , 39
iL/K , 55 Galois extension, 63
kK , 16 generalized local Kronecker-Weber
p-adic absolute value, 11 theorem, 83
p-adic integers, 12 geometric Frobenius, 74
p-adic numbers, 12
sth ramification group, 52 Hasse-Arf theorem, 83
x-adic topology, 9 Hensel’s lemma, 16
Herbrand’s theorem, 58
abelian extension, 68 Higher ramification groups, 74
absolute ramification index, 36 higher unit groups, 51
absolute value, 4 homomorphism
archimedean, 5 formal group, 76
non-archimedean, 5
absolute values inertia degree, 43
equivalence, 4 inertia field, 52
adjoint matrix, 7 inertia group, 52
adjugate matrix, 7 integral element, 7
algebraic integer, 7 integrally closed, 8
archimedean absolute value, 5 inverse limit, 10, 64
arithmetic Frobenius, 66 inverse limit topology, 10
inverse system, 64
break points, 23
Krull topology, 63
class field, 70
Coleman operators, 83 length, 24
line segment, 24
directed system, 64 local field, 27
discrete valuation ring, 29 Local Kronecker-Weber theorem, 74
discretely valued field, 27 local Kronecker-Weber theorem, 83
DVF, 27 lower convex hull, 23

88
Index III Local Fields

lower convex set, 23 residue field, 16

lower numbering of ramification ring topology, 9
group, 60
Lubin–Tate module, 77 separable extension, 63
Lubin–Tate polynomial, 77 slope, 24
Lubin–Tate series, 77 strict p-ring, 33
strong triangle inequality, 5
Mahler coefficient, 40
Mahler’s theorem, 39 tame quotient, 55
maximal ideal, 16 Teichmüller lift, 30
mixed characteristic, 29 Teichmüller map, 30
module Teichmüller representative, 30
formal group, 77 topological ring, 9
multiplicity, 24 totally ramified extension, 43, 65
triangle inequality, 4
Newton polygon, 23
trivial absolute value, 4
non-archimedean absolute value, 5
norm
ultrametric, 5
equivalent, 19
uniformizer, 27
norm on vector space, 19
unramified extension, 43, 65
normal extension, 63
upper numbering of ramification
normalized valuation, 27
group, 60
perfect ring, 29
primitive polynomial, 16 valuation, 15
profinite groups, 63 normalized, 27
projective limit, 10 valuation ring, 7, 15
valued field, 4
ramification group, 52 discretely, 27
lower numbering, 60
upper numbering, 60 Weil group, 66
ramification index, 43 wild inertia group, 55
rank, 43 Witt vector, 34

89