Galois module structure of local unit groups

Size: px
Start display at page:

Download "Galois module structure of local unit groups"

Transcription

1 Galois module structure of local unit groups Romyar T. Sharifi Abstract We study the groups U i in the unit filtration of a finite abelian extension K of Q p, for an odd prime p. We determine explicit generators of the U i as modules over the Z p -group ring of Gal(K/Q p ). We work in eigenspaces for powers of the Teichmüller character, first at the level of the field of norms for the extension of K by p-power roots of unity and then at the level of K. 1 Introduction Fix an odd prime p and a finite unramified extension E of Q p. We use F n to denote the field obtained from E by adjoining to E the p n th roots of unity in an algebraic closure of Q p. The ith unit group in the unit filtration of F n will be denoted by U n,i. The object of this paper is to describe generators of the groups U n,i as modules over the Z p -group ring of G n = Gal(F n /Q p ). We express these generators in terms of generators of the pro-p completion D n of F n as a Galois module. In fact, one consequence of our work is a rather elementary proof of an explicit presentation of D n as such a module, as was proven by Greither [Gr] using Coleman theory. Instead of working with all of D n at once, we find it easier to work with certain eigenspaces of it. For this and several purposes, it will be useful to think of the Galois group G n as a direct product of cyclic subgroups G n = Γ n Φ, where Γ n = Gal(F n /E) with = p 1 and Γ n = p n 1, and Φ is isomorphic to Gal(E/Q p ). We then decompose D n into a direct sum of p 1 eigenspaces for powers of the Teichmüller character ω : Z p. For any integer r, the ω r -eigenspace D n (r) of D n is the subgroup of elements upon which σ acts by left multiplication by ω(σ) r. This definition depends only on r modulo p 1, so we fix r with 2 r p. Note that D n (r) is a module over the group ring A n = Z p [Γ n Φ]. In fact, as we shall see in Section 3.1, the A n -module D n (r) has a generating set with just one element if r p 2, three elements if r = p 1, and two elements if r = p. 1

2 Galois module structure 2 We will be interested in the A n -module structure of the groups It turns out that V (r) n,i V (r) n,i = D (r) n U n,i. V (r) n,i+1 = V (r) n,i+2 = = V (r) n,i+p 1 for all i r mod p 1 (see Lemma 2.1), so we will consider only such i and set V n,i = V (r) n,i. Our main results, Theorems and 4.3.3, provide a small set of at most n + 1 generators of V n,i as an A n -module and state that any proper generating subset of it has cocardinality 1. The elements of this set are written down explicitly as A n -linear combinations of elements of the generators of D n (r). In Section 4.2, elements of a special form are constructed so as to lie as deep in the unit filtration as possible. In Section 4.3, these are refined to elements of the same form that instead lie just deep enough to be in V n,i, which are in turn the generators that we use. It is convenient to work first in the field of norms F of Fontaine-Wintenberger for the tower of extensions F n of E. This is a field of characteristic p, the multiplicative group of which is the inverse limit of the F n. We prove analogues of all of the above-mentioned results first at this infinite level, prior to applying them in descending to the level of F n. The fact that the pth power map is an automorphism of F simplifies some of the computations. Moreover, the structure of the eigenspaces of the pro-p completion of F, which we study in Section 3.1, is somewhat simpler than that of the D n (r). We construct special elements in the eigenspaces of the groups in the unit filtration in Section 3.2, refine them in Section 3.3, and prove generation and a minimality result in Section 3.4. We see a number of interesting potential applications for the results of this paper. To mention just one, it appears to make possible the computation of the conductors of all degree p n Kummer extensions of F n in terms of the Kummer generator of the extension. The problem of making this computation, which was approached by the author in three much earlier papers, has until now seemed beyond close reach in this sort of generality. Acknowledgments. The idea for this paper originated with the author s 1999 Ph.D. thesis, and initial computations were performed on an evening in June 2001 during a visit to the University of Nottingham. The author thanks Ivan Fesenko for his hospitality. A short draft was written in 2002, when the author was supported by an NSF Postdoctoral Research Fellowship. In August 2006, some additions were made, and that work was funded in part by an NSERC Discovery Grant and the Canada Research Chairs program. As the paper tripled in size in the summer of 2011, the author was supported in part by NSF award DMS The author thanks Richard Gottesman for his interest in this work, which inspired him to finish the paper.

3 Galois module structure 3 2 Preliminaries We maintain the notation of the introduction and introduce some more. Recall from [W] that the field of norms F for the extension F = n F n of E is a local field of characteristic p with multiplicative group F = lim F n, the inverse limit being taken with respect to norm maps. Let ζ = (ζ p n) n be a norm compatible sequence of p-power roots of unity, with ζ p n primitive p n th root of unity in F n. Then a λ = 1 ζ = (1 ζ p n) n is a prime element of F. For m n, let N m,n : F m F n be the norm map. Recall that the addition on F is given by (α + β) n = lim m N m,n(α m + β m ) for α = (α n ) n and β = (β n ) n in F. We fix an isomorphism of the residue field of E (and thereby each F n ) with F q, with q the order of the residue field. Using this, the field F q is identified with a subfield of F via the map that takes ξ F q to ( ξ p n ) n F, where ξ is the (q 1)th root of unity in E lifting ξ. The field F may then be identified with the field of Laurent series F q ((λ)). If F is the union of the F n, then G = Gal(F /Q p ) acts as automorphisms on the field F. As with G n, we may decompose G = Gal(F /Q p ) into a direct product of procyclic subgroups G = Γ Φ, where Gal(F /E) = Γ, the group has order p 1, the group Γ is isomorphic to Z p, and Φ is isomorphic to Gal(E/Q p ). Let γ denote the topological generator of Γ such that γ(ζ p n) = ζ 1+p pn for all n. The pro-p completion D of F decomposes into a direct sum of eigenspaces for the powers of the Teichmüller character ω on. For an integer r, we let D (r) = D εr, where ε r is the idempotent ε r = 1 ω(δ) r δ Z p [ ]. p 1 For i 1, let U i denote the ith group in the unit filtration of F. We then set V (r) δ i = U i D (r) and (V (r) i ) = V (r) i V (r) i+1. The following is Lemma 2.3 of [Sh] (with F n replaced by F ).

4 Galois module structure 4 Lemma 2.1. We have V (r) i i r mod p 1. /V (r) i+p 1 = F q for every i 1, and (V (r) i ) if and only if From now on, we set V i = V (r) i and V i = (V (r) i ) if i r mod p 1. As a consequence of Lemma 2.1, an element z V i is determined modulo λ i+p 1 by its expansion z 1 + ξλ i mod λ i+1 (2.1) with ξ F q. The following is Lemma 2.4 of [Sh] (with F n replaced by F ). Lemma 2.2. Let z V i. If p i, then z γ 1 V i+p 1. Otherwise, z γ 1 V i+2(p 1). We identify Λ = Z p [[Γ]] with the power series ring Z p [[T ]] via the continuous, Z p - linear isomorphism that takes γ 1 to T, and we use additive notation to describe the action of Z p [[T ]] on D. Ramification theory would already have told us that T V i V i+p 1 for all i. On the other hand, explicit calculation will yield the following two lemmas and proposition, which provide more precise information on how powers of T move elements of V i. For ξ F q, we let V i (ξ) denote the set of z V i for which z has an expansion of the form in (2.1). We use [k] to denote the smallest nonnegative integer congruent to k Z modulo p. Lemma 2.3. Let z V i (ξ) for some i. Then, for 0 j [i], we have ( ) [i]! T j z V i+j(p 1) ([i] j)! ξ. Proof. Note that λ γ = 1 ζ 1+p = 1 (1 λ)(1 λ p ) = λ + λ p λ p+1. (2.2) Using this, we see, for any i 1, that Hence, (1 + ξλ i ) γ iξλ i+p 1 1 λ 1 + ξλ i mod λi+2p 2. (2.3) Applying (2.4) recursively, we obtain the result. (1 + ξλ i ) γ iξλ i+p 1 mod λ i+p. (2.4) Lemma 2.4. Let z V pi p+1 (ξ) for some i 2. If j is a nonnegative multiple of p 1, then T j+1 z V p(i+j) (ξ).

5 Galois module structure 5 Proof. Let us begin by proving slightly finer versions of (2.3) in two congruence classes of exponents modulo p. For any t 1, we have (1 + ξλ pt ) γ 1 = 1 + ξλpt (1 + λ p(p 1) λ p2 ) t 1 + ξλ pt 1 mod λ p(t+p 1) and (1 + ξλ pt+1 ) γ 1 = 1 + ξλ pt+1 pt+1 m=1 ( pt+1 ) m (λ p 1 λ p ) m 1 + ξλ pt ξ(λ p(t+1) λ p(t+1)+1 ) mod (λ p(t+p 1)+1, λ p(2t+1)+1 ), the latter congruence following from the fact that p ( ) pt+1 m for 2 m < p. Via some obvious inequalities, we conclude that (1 + ξλ pt ) γ 1 1 mod λ p(t+2), (2.5) (1 + ξλ pt+1 ) γ 1 (1 + ξλ p(t+1) )(1 ξλ p(t+1)+1 ) mod λ p(t+2). (2.6) Let x = 1 + ξλ pi p+1. Recursively applying (2.5) and (2.6), we see that x (γ 1)k+1 (1 + ( 1) k ξλ p(i+k) )(1 + ( 1) k+1 ξλ p(i+k)+1 ) mod λ p(i+k+1), for any positive integer k, as (2.4) implies that U γ 1 p(i+k) U p(i+k+1). The result now follows by application of ε i, since z 1 x ε i V pi, T j+1 x ε i V p(i+j) (ξ), and T j+1 V pi V p(i+j+1) 1 by Lemma 2.2. Let us use {k} to denote the smallest nonnegative integer congruent to k Z modulo p 1. For i 1 with p i, we define a monotonically-increasing function φ (i) : Z 0 Z by φ (i) (0) = i and φ (i) (a) = pa + (i [i]) + {[i] a} (2.7) for a 1. Proposition 2.5. Let z V i (ξ) for some i 2 with p i. Then, for j 1, we have ( ) [i]! T j z V φ (i) (j) {[i] j}! ξ. Proof. Lemma 2.3 implies that T [i] 1 z V φ (i) ([i] 1)([i]! ξ), and note that φ (i) ([i] 1) 1 mod p. Set k = {[i] j}. Since j + k [i] is divisible by p 1, Lemma 2.4 then implies that T j+k z V φ (i) (j+k)([i]! ξ). (2.8)

6 Galois module structure 6 It follows from (2.7) that φ (i) (j + k) i = p(j + k [i]) + (p 1)[i], and so, given (2.8), Lemma 2.2 forces T l z V for all l j + k. In particular, φ (i) (l) applying Lemma 2.3 with j replaced by k and z replaced by T j z, we see that for (2.8) to hold, T j z must have the stated form. Remark 2.6. The obvious analogues of the results of this section all hold at the level of F n for n 2, with λ replaced by λ n = 1 ζ p n. In fact, Lemmas 2.1 and 2.2 were originally proven in that setting in [Sh]. That the other results hold breaks down to the fact that p is a unit times λ pn 1 (p 1) n in F n, which in particular tells us that (2.2) can be replaced by 3 The infinite level λ γ n λ n + λ p n λ p+1 n 3.1 Structure of the eigenspaces mod λ p(p 1)+1 n. In this subsection, we fix choices of certain elements that will be used throughout the paper. From now on, we let ξ denote an element of F q with Tr Φ ξ = 1, the conjugates of which form a normal basis of F q over F p. Let ϕ Φ denote the Frobenius element. Let N Φ Z p [Φ] denote the norm element. Let ζ = (ζ p n) n be a norm-compatible system of primitive p n th roots of unity as before. Let r be an integer satisfying 2 r p. If 2 r p 2, we simply fix an element u r V r (ξ). In the case that r = p 1, generation of D (p 1) requires one additional element π D (p 1), a non-unit, chosen along with u p 1 V p 1 (ξ) in the lemma which follows. The case of r = p shall require more work, but we will fix elements w V 1 ( ξ) and u p V p (ξ) as in Proposition below. Lemma There exist elements π D (p 1) and u p 1 V p 1 (ξ) such that π ϕ = π and π γ 1 = u N Φ p 1. Proof. Set π = λ ε p 1, which satisfies π ϕ = π and π γ 1 V p 1 (1). Since every unit is a norm in an unramified extension, there exists u p 1 D (p 1) such that (u p 1) N Φ = π γ 1, and such an element must lie in V p 1 (ξ ) for some ξ F q with Tr Φ ξ = 1. Hilbert s Theorem 90 tells us that ξ = ξ + (ϕ 1)η for some η F q. Let z V p 1 (η), and set u p 1 = u p 1z 1 ϕ. In fact, one could have chosen u p 1 V p (ξ) arbitrarily and then taken π to satisfy the relations, as can be seen using the results of the following section.

7 Galois module structure 7 Lemma There exist elements w V 1 ( ξ) and u p V p (ξ) with w N Φ u ϕ 1 p = w γ 1 p. = ζ and Proof. First, local class field theory yields the existence of an element w D (p) with (w ) N Φ = ζ. Since ζ V 1 ( 1), we must have w V 1 ( ξ ) for some ξ F q with Tr Φ ξ = 1. Since ξ = ξ + (ϕ 1)η for some η F q, we choose any y V 1 (η), and then w = w y 1 ϕ V 1 ( ξ) satisfies w N Φ = ζ as well. Next, note that (w γ 1 p ) N Φ = 1, and so Hilbert s Theorem 90 allows us to choose an element u p D (p) with (u p) ϕ 1 = w γ 1 p. A simple computation using (2.4) tells us that w γ 1 p V p (ξ p ξ), and therefore u p V p (ξ + a) for some a F p. We may then choose z V p (a) with z ϕ = z and take u p = u pz 1 V p (ξ). We need slightly finer information on the relationship between w and u p inside the unit filtration, as found in the following proposition. Proposition There exist elements w V 1 ( ξ) and u p V p (ξ) with w N Φ and u ϕ 1 p = w γ 1 p such that the element y = u p w pϕ 1 lies in V 2p 1 ( ξ). = ζ Proof. For now, fix any choices of u p and w as in Lemma We must have u p = (1 + ξλ p ) ε 1 α with α V 2p 1 and w = (1 ξλ) ε 1 β with β V p. Note that (1 + ξλ p ) ϕ (ξ p ξ)λ p mod λ 2p and (1 ξλ) γ 1 p = 1 ξ(λ + ( ζλp ) (1 ξλ)(1 ξ p λ p ) 1 + ξ p ξ 1 λ ) λ p mod λ 2p. 1 ξλ We then have (1 ξλ) γ 1 p ξ(1 ξ) 1 + (1 + ξλ p ) ϕ 1 1 ξλ λp+1 mod λ 2p. (3.1.1) We denote the quantity on the right-hand side of (3.1.1) by θ. By Lemma 2.2, we have β γ 1 p V 3p 2, from which it follows that α ϕ 1 θ ε 1 V 3p 2. On the other hand, by Lemma 2.1, we have yα 1 = (1 + ξλ p ) ε 1 (1 ξλ p ) ε 1 β pϕ 1 V 3p 2, so in fact we have y ϕ 1 θ ε 1 V 3p 2. If we can show that θ ε 1 V 2p 1 (ξ ξ p ), we will then have y V 2p 1 ( ξ + a) for some a F p. As in the proof of Lemma 3.1.2, we can then choose an element z V 2p 1 (a) with z ϕ = z and replace u p by u p z 1 to obtain the result.

8 Galois module structure 8 By Proposition 2.5, we see that to show that θ ε 1 V 2p 1 (ξ ξ p ), it suffices to show that θ ε 1(γ 1) p 1 V p 2(ξ p ξ). Since p 2 1 mod p 1, for this, it suffices to show that θ (γ 1)p (ξ p ξ)λ p2 mod λ p2 +1. This is a simple consequence of Lemma 3.1.4, which follows. That is, in the notation of said lemma, Fermat s little theorem and the binomial theorem tell us that d p 1,k = 1 for all positive integers k p 1. Lemma For each positive integer j p 1, one has ( ) ( (γ 1) j j ) ξ(1 ξ) ξλ λp d j,k ξ k (1 ξ) λ (j+1)p mod λ (j+1)p+1, where for positive integers k j. Proof. We make the expansion θ = 1 + d j,k = k=1 k ( ) k ( 1) j+h h j F p h h=1 p 1 ξ(1 ξ) 1 ξλ λp+1 (1 + ξ k (1 ξ)λ p+k ) mod λ 2p. k=1 Since Us γ 1 U s+p 1 for all s, as follows from (2.4), to compute θ (γ 1)j modulo λ (j+1)p+1, it suffices to compute (1 + ξ k (1 ξ)λ p+k ) (γ 1)j modulo λ (j+1)p+1. Fix a positive integer k p 1. We claim that the coefficient of λ (j+1)p in the expansion of (1 + ξ k (1 ξ)λ p+k ) (γ 1)j as a power series in F q [[λ]] is 0 if j < k and ξ k (1 ξ)d j,k if j k. As a consequence of (2.3), one sees that (1 + ξλ t ) γ 1 (1 + tξλ t+p 1 )(1 tξλ t+p ) mod λ t+2p 2 for any t p 1. Using this and the finer congruence (2.6) when possible, an induction yields that the expansion in question is determined by min(j,k) m=0 where (a i ) P j,k,m ( 1 + ξ k k! (1 ξ) (k m)! j m i=1 a i λ (j+1)p+k m ) ( 1)j m mod λ (j+1)p+k+1, P j,k,m = {((a 1, a 2,..., a j m ) Z j m k m a 1 a 2 a j m k}

9 Galois module structure 9 if j > m and P j,k,j = {0}, and we consider the empty product to be 1. In particular, the coefficient in question is indeed 0 for j < k and is ξ k (1 ξ)c j,k for j k, where c j,k = ( 1) j k k! j k a i. (a i ) P j,k,k i=1 It remains to verify that c j,k = d j,k for j k. Let D denote the differential operator x d on F dx p[x]. By the binomial theorem, we have that k ( ) k D j ((1 x) k ) x=1 = ( 1) h h j x h x=1 = ( 1) j d j,k. h h=1 On the other hand, repeated application of the product formula for the derivative yields that min(j,k) D j ((1 x) k ) x=1 = ( 1) k h=1 for all j k, hence the result. k! (k h)! (a i ) P j,h,h i=1 j h a i (x 1) k h x h x=1 = ( 1) j c j,k In the next section, we will obtain the following very slight refinement of what is essentially a result of Greither s [Gr, Sections 2-3] (see also [Sh, Corollary 2.2]). Theorem For r p 2, the A-module D (r) is freely generated by any u r V r (ξ). The A-module D (p 1) has a presentation D (p 1) = π, u p 1 π ϕ = π, u N Φ p 1 = π γ 1, for some u p 1 V p 1 (ξ) and π D (p 1). The A-module D (p) has a presentation D (p) = u p, w w γ 1 p = u ϕ 1 p, for some u p V p (ξ) and w V 1 ( ξ) such that w N Φ = ζ. 3.2 Special elements Fix r with 2 r p, and define φ: Z 0 Z by φ(a) = φ (r) (a) for a 1. Set { 0 if 2 r p 1, δ = 1 if r = p. For all a 1, we have φ(a) = p(a + δ) + {r δ a},

10 Galois module structure 10 so φ(a) is the smallest integer that is at least p(a + δ) and congruent to r modulo p 1. From now on, i will be used solely to denote a positive integer congruent to r modulo p 1. We will write α β to denote that both α and β lie in V i (ξ) for some i and ξ F q. We use additive notation for the action of A = Z p [Φ][[T ]] on D (r). We begin with the following useful lemma. Lemma Let j be a positive integer. a. We have ( ) T j [r]! u r V φ(j) {r δ j}! ξ. b. If j r δ mod (p 1) so that T j u r pz for some z D (r), then T j u r pz V φ(j)+p 1 ( [r]!ξ). Proof. For r < p, part a is a direct consequence of Proposition 2.5 and the fact that u r V r (ξ). For r = p, Proposition 2.5 and the fact that φ = φ (2p 1) on positive integers would tell us more directly that T j 1 y V φ(j) ( ξ) for j 1, for y as in Proposition { j}! Note, however, that T u p = T y pϕ 1 T w T y, since pt w V p 2. This is also the key point of part b. That is, we have as pt z V pφ(j) and Since T (T j u r pz) T j+1 u r φ(j + 1) φ(j) + 2(p 1) < pφ(j). T j+1 u r V φ(j)+2(p 1) ([r]!ξ), a final application of Proposition 2.5 tells us that T j u r pz had to be in the stated group. For a nonnegative integer m, let us define φ m : Z 0 Z 0 by We remark that φ m = p m (φ + 1) 1. pφ m = φ (φ m δ). (3.2.1) From now on, we set ρ = pϕ 1 for brevity of notation. We define special elements in the unit filtration of D (r).

11 Galois module structure 11 Theorem Let m and j be nonnegative integers. Define ( ) α m,j = 1 m {r δ j}!ρ m T j ρ m k T φ k 1(j) δ u r, [r]! unless j = 0 and r = p 1, in which case we replace {r δ j}! with 1 in the formula. Then α m,j V φm(j)(ξ). Furthermore, k=1 (p m bt j + c)u r / V φm(j)+p 1 for all b Z p [Φ] with b 0 mod p and c T j+1 A. Proof. We work by induction, the case of m = 0 being Lemma 3.2.1a, aside from the case j = 0, in which case it is simply the definition of u r. Assume we have proven the first statement for m. Then pα m,j V pφm(j)(ξ p ) and, using Lemma 3.2.1a and (3.2.1), we have Lemma 3.2.1b then tells us that T φm(j) δ u r V pφm(j)( [r]!ξ ). α m+1,j = ρα m,j 1 [r]! T φm(j) δ u r V pφm(j)+p 1(ξ). Now assume the second statement is true for m. (For m = 0, this is a consequence of the fact that the conjugates of ξ are F p -linearly independent.) Suppose that α = (p m+1 bt j + c)u r V i with i φ m+1 (j), b Z p [Φ] pz p [Φ] and c T j+1 A. We write c = (pc + T h ν)u r for some c, ν A with ν 0 mod (p, T ) and h j + 1. By induction, we have that (p m bt j + c )u r / V φm(j)+p 1. Since φ m+1 (j) = pφ m (j) + p 1 and α V φm+1 (j) by assumption, this forces p(p m bt j + c )u r T h νu r, which tells us by Lemma 3.2.1a that φ(h) pφ m (j). On the other hand, Lemma 3.2.1b tells us that α V φ(h)+p 1, which forces i = φ m+1(j). The second statement of Theorem insures, in particular, that au r 0 for all nonzero a A. We therefore have the following corollary.

12 Galois module structure 12 Corollary The A-submodule of D (r) generated by u r is free. In the exceptional case that r = p, we require additional elements. First, we modify the function φ m for this r. For nonnegative integers m and j, we set φ m(j) = φ m (j) unless r = p and j = p l 1 for some l 0, in which case we set φ m(p l 1) = p m+l+1 + p m+1 1 = φ m (p l 1) + p m (p 1). Theorem Let l and m be nonnegative integers. Define ( ) m β m,l = ρ m T pl 1 + ρ m k T φ k 1 (pl 1) 1 u p + ρ m+l+1 w. Then β m,l V φ m (p l 1)( ξ). Moreover, for any j 0, we have k=1 (p m bt j + c)u p + dw / V φ m (j)+p 1 for all b Z p [Φ] pz p [Φ], c T j+1 A and d Z p [Φ]. Proof. The proof is similar to that of Theorem definition of w tell us that Since Lemma 3.2.1a and the T pl 1 u p V p l+1(ξ) and ρ l+1 w V p l+1( ξ), Lemma 3.2.1b yields β 0,l = ρ l+1 w + T pl 1 u p V p l+1 +p 1( ξ). For any m 0, we have By induction and Lemma 3.2.1a, we have β m+1,l = ρβ m,l + T φ m (pl 1) 1 u p. ρβ m,l V pφ m (p l 1)( ξ) and T φ m(p l 1) 1 u p V pφ m (p l 1)(ξ). Since φ m+1(p l 1) = pφ m(p l 1) + p 1, that β m+1,l V φ m+1 (p 1)( ξ) is just another l application of Lemma 3.2.1b. Let j 0, b Z p [Φ] pz p [Φ], c T j+1 A, and d Z p [Φ]. First, suppose that α = (bt j + c)u p + dw V i for some i φ 0(j). Note that (bt j + c)u p V φ(j) by Lemma 3.2.1a, while dw V for some l 0. Since α V p l φ 0 (j), we must have p l φ(j). We then have α V φ(j) unless φ(j) = pl. This occurs if and only if l 1 and j = p l 1 1, in which case φ 0(j) = p l + p 1. For this to hold, we must have (bt j + c)u p dw. Lemma 3.2.1b then implies that α V, so i = p l +p 1 φ 0(j) in all cases.

13 Galois module structure 13 Suppose now that α = (p m+1 bt j + c)u p + dw V i for some i φ m+1(j). Rewrite c as pc + T h ν for some h j + 1 and c, ν A with ν 0 mod (p, T ). If we are to have α V p, we may also write d = pd for some d Z p [Φ]. By induction, we have that (p m bt j + c )u p + d w / V φ m (j)+p 1, and so in order that α V φ m+1 (j), we must have that (p m+1 bt j + pc )u p + dw T h νu p which tells us using Lemma 3.2.1a that φ(h) pφ m(j). On the other hand, Lemma 3.2.1b tells us that α V φ(h)+p 1, so we must have i = φ m+1(j). Theorem may now be proven as a consequence of the description of the above elements and their place in the unit filtration. Proof of Theorem For r p 1, the union of the disjoint images of the functions φ m is exactly the set of positive integers congruent to r modulo p 1. Therefore, Theorem implies that there exists an element of the A-module generated by u r in V i (ξ) for each i r mod p 1. In particular, u r therefore clearly generates V r as an A-module, which equals D (r) for r p 2, and it is free by Corollary Every element of D (p 1) may then be written in the form π m u a p 1 with m Z p and a A, and such an element can clearly only be trivial if m is, and therefore a is as well. Noting that our choices of π and u p 1 as in Lemma satisfy the desired relations, the presentation for r = p 1 is as stated. For r = p, the union of {1} and the images of the functions φ m and φ m is the set of positive integers that are congruent to 1 modulo p 1. Theorem and Theorem imply that there exists an element of the A-module generated by u p and w in V i (ξ) for each i 1 mod p 1. Thus, this A-module is D (p). Our choices of u p and w satisfy the relations of Lemma 3.1.2, and it follows from the second statement of Theorem that if either c A or d Z p [Φ] is nonzero, then so is cu p + dw. 3.3 Refined elements In this section, we provide refinements of the elements constructed in Theorem and Theorem We maintain the notation of Section 3.2. We begin by constructing certain one-sided inverses to the monotonically increasing functions φ and φ m. For any nonnegative integer a and positive integer t, let us set a t = max(a + {t a}, t).

14 Galois module structure 14 Therefore, a t is the smallest integer greater than or equal to t and a and congruent to t modulo p 1. Define ψ : Z 0 Z 0 by a r + 1 ψ(a) = δ p except for r = p 1 and a p 1, in which case we set ψ(a) = 0. For m 0, define ψ m : Z 0 Z 0 by ( ) a + 1 ψ m (a) = ψ 1. Note that ψ 0 = ψ. Lemma We have ψ m (φ m (j)) = j for all nonnegative integers j. Moreover, for all such j and positive integers a, we have φ m (j) a if and only if j ψ m (a). Proof. First, note that φ(j) is congruent to r modulo p 1, so we have φ(j) + 1 p(j + δ) + {r δ j} + 1 ψ(φ(j)) = δ = δ = j p p unless r p 1 and j = 0, but one checks immediately that ψ(φ(0)) = ψ(r) = 0 if r p 1 as well. It follows that we have ( ) p m (φ(j) + 1) ψ m (φ m (j)) = ψ 1 = ψ(φ(j)) = j. p m Therefore, if φ m (j) a, then j = ψ m (φ m (j)) ψ m (a), since ψ m is nondecreasing. To finish the proof, we need only show that φ m (ψ m (a)) a, since φ m is nondecreasing (in fact, strictly increasing). First, note that the definition of ψ is such that p m ψ(a) = ψ( a r ). For i r mod p 1 with i 1, p 1, the value φ(ψ(i)) is the unique integer between p i+1 i+1 and p + p 2 which is congruent to r mod p 1. This implies that p p { a r if a r 1 mod p, or a r = p 1, φ(ψ(a)) = a r + p 1 otherwise. (3.3.1) which is, in particular, at least a. By definition of φ m and ψ m, we then have that a + 1 φ m (ψ m (a)) = p m (φ(ψ m (a)) + 1) 1 p m 1 a. p m

15 Galois module structure 15 We actually need a version of Lemma with φ m replaced by φ m and ψ m replaced by an appropriate function ψ m : Z 0 Z 0, which we now define. Set ψ m = ψ m if r p 1 and, if r = p, let { ψm (a) 1 if p m+l+1 + p m a p m+l+1 + p m+1 1 for some l 0, ψ m(a) = ψ m (a) otherwise. Note that ψ m(a) = ψ m (a) 1 for r = p if and only if φ m (p l 1) < a φ m(p l 1) for some l 0, in which case ψ m(a) = p l 1. One then easily checks the following. Corollary We have ψ m(φ m(j)) = j for all nonnegative integers j. Moreover, for all such j and positive integers a, we have φ m(j) a if and only if j ψ m(a). For the rest of this section, we fix a positive integer i with i r mod p 1. Remark Lemma and Theorem (resp., Corollary and Theorem 3.2.4) tell us that each α m,ψm(i) (resp., β m,l with ψ m(i) = p l 1) lies in V i. These elements have the form (p m bt j + c)u r + dw for j = ψ m(i), where b Z p [Φ] pz p [Φ], c T j+1 A, and d Z[Φ], with d = 0 if r p. The same results also show that no such element with j < ψ m(i) can lie in V i. For any m 0, define θ m : Z 1 Z 0 by ( ) a r θ m (a) = ψ. By Lemma 3.2.1a and Lemma 3.3.1, the value θ m (i) for i r mod p 1 is the minimal integer j such that p m T j u r V i. In particular, θ m (i) ψ m (i) for all i. Lemma For all positive integers m and k with k m, we have with equality if and only if p m φ k 1(ψ m(i)) δ θ m k (i) 1, p m k+1 φ k 1(ψ m(i)) < i. (3.3.2) Moreover, we have ψ m(i) θ m (i) 1, with equality if and only if the above equivalent conditions hold for k = 1. Proof. Let us check the case that r = p and ψ m(i) = p l 1 for some l 0 separately. First, suppose that p m+l+1 < i < p m+l+1 +p m+1. In this case, we have ψ m(i) = θ m (i) 1. We also have φ k 1(ψ m(i)) = φ k 1(p l 1) = p k+l + p k 1 = ψ(p k+l+1 + p k+1 ) θ m k (i),

16 Galois module structure 16 with equality if and only if p m k+1 φ k 1(ψ m(i)) = p m+l+1 + p m+1 p m k+1 < i. (3.3.3) Moreover, in the case that p m+l+1 p m+1 + 2p m < i p m+l+1, the values φ k 1 (ψ m(i)) and p m k+1 φ k 1 (ψ m(i)) are the same as in the previous case, while θ m k (i) 1 and i are smaller. So, we may assume from this point forward that r and i are such that ψ m(i) = ψ m (i) and φ k 1 (ψ m(i)) = φ k 1 (ψ m (i)) for all k. We claim that ρ m k T φ k 1(ψ m(i))+1 δ u r lies in V i+p 1 (resp., ρ m k T α k,ψm(i) lies in V i+p 1 ) for all positive (resp., nonnegative) k m. Note that T α m,ψm(i) V i+p 1 as a consequence of Theorem Suppose that ρ m k T α k,ψm(i) V i+p 1 for some positive k m. We then have that ρ m k T φ k 1(ψ m(i))+1 δ u r [r]!ρ m k T α k,ψm(i) V i+p 1, which also forces ρ m k+1 T α k 1,ψm(i) V i+p 1, since ρ m k+1 T α k 1,ψm(i) = ρ m k T α k,ψm(i) + 1 [r]! ρm k T φ k 1(ψ m(i))+1 δ u r, proving the claim. In particular, since ρ m T α 0,ψm(i) V i+p 1, we have ρ m T ψm(i)+1 u r V i+p 1 as well. The definition of θ m k (i) now yields the desired inequalities. Equation (3.3.2) holds for a given k if and only if ρ m k+1 α k 1,ψm(i) / V i. Since [r]!ρα k 1,ψm(i) T φ k 1(ψ m(i)) δ u r, this occurs if and only if ρ m k T φ k 1(ψ m(i)) δ u r / V i and, therefore, if and only if φ k 1 (ψ m (i)) δ θ m k (i) 1, which must then be an equality. Note also that ψ m (i) < θ m (i) if and only if ρ m α 0,ψm(i) / V i, which holds by Lemma 3.2.1a if and only if p m φ(ψ m (i)) < i, the same condition as (3.3.2) for k = 1. From now on, we set i m = i p m for all m 0. Lemma For any pair of positive integers m and k with k m, we have with equality if and only if φ k 1(ψ m(i)) δ θ m k (i) 1, 1. i m+ɛ 0 mod p, or r = p 1 and i m = p, 2. i m+ɛ r + 1 mod p 1, but not r = p 1 and i m = 1, and

17 Galois module structure i j mod p m+ɛ for some 0 < j < p m+1 k, where ɛ = 0 unless r = p and i m+1 = p l + 1 for some l 0, in which case we set ɛ = 1. Moreover, we have ψ m(i) θ m (i) 1, with equality if and only if the above conditions hold with k = 1. Proof. The case that r = p and ψ m(i) = p l 1 for some l 0 follows from the proof of Lemma 3.3.4, noting that if i m+1 = p l + 1, then it is both nonzero modulo p and congruent to p + 1 modulo p 1, and the third condition of the lemma holds exactly when (3.3.3) does. On the other hand, for the remaining i with ψ m(i) = p l 1, we have i m+1 = p l, and the fact that the inequality is strict was shown in the proof of Lemma So, we again assume that r p or i is such that ψ m(i) p l 1 for all l 0. By Lemma 3.3.4, it suffices to determine the precise conditions under which (3.3.2) holds. Let us set a = (i + 1) m. It follows from (3.3.1) that we have p m k+1 φ k 1 (ψ m (i)) = { p m a r+1 p m k+1 if p a r+1, p m a r+1 + p m (p 1) p m k+1 otherwise, (3.3.4) unless r = p 1 and a r+1 = p, in which case p m k+1 φ k 1 (ψ m (i)) = p m+1 p m k+1. Aside from this exceptional case, (3.3.4) implies that p cannot divide a r+1 if (3.3.2) is to hold. Moreover, if a r+1 > a, then again (3.3.2) cannot hold, so for it to hold, we must have a r + 1 mod p 1, but not r = p 1 and a = 1. Assuming that these necessary conditions hold, the condition that p m k+1 φ k 1 (ψ m (i)) = p m a p m k+1 < i is exactly that i j mod p m with 0 < j < p m k+1. For m 0, we will define new elements κ m,i of V i that involve fewer terms and easier-to-compute exponents of powers of T than the expressions for α m,ψm(i) and β m,l. In preparation, set σ(m, i) = log p (p m i m i) for any m 0 such that p m i. Note that 0 σ(m, i) m 1 when it is defined and σ(m + 1, i) is defined and greater than or equal to σ(m, i) whenever σ(m, i) is defined. First, supposing either that r p 1 or that r = p and i m+1 1 is not a power of p, we set κ m,i = ρ m T θm(i) u r (3.3.5) if i m r + 1 mod p 1, p i m, i < p m, or p m i, unless r = p 1 and i m = p, and κ m,i = ρ m T θm(i) 1 a m,i m 1 k=σ(m,i) ρ k T θ k(i) 1 u r (3.3.6)

18 Galois module structure 18 otherwise, where a m,i denotes the least positive residue of the inverse of {r+1 δ θ m (i)}! modulo p unless r = p 1 and θ m (i) = 1, in which case we take a m,i = 1. In the remaining case that r = p and i m+1 1 is a power of p, we set κ m,i = ρ m T θm(i) 1 + m 1 k=σ(m+1,i) ρ k T θ k(i) 1 u r + ρ m+log p (im+1 1)+1 w. (3.3.7) For consistency, we let a m,i = 1 for such m. Note that Lemma tells us that each κ m,i has the form (ρ m T ψ m (i) + c)u r + dw for some c T ψ m (i)+1 A and d Z p [Φ], with d taken to be zero if r p 1. We give two examples for p = 5 and particular values of i. Example Suppose that p = 5, r = 3, and i = Then we have κ 0,i = T 2380 u 3, κ 1,i = ρt 476 u 3, κ 2,i = (ρ 2 T 95 ρt 475 T 2379 )u 3, κ 3,i = (ρ 3 T 19 ρ 2 T 95 )u 3, κ 4,i = ρ 4 T 4 u 3, κ 5,i = (ρ 5 ρ 4 T 3 ρ 3 T 19 )u 3. Example Suppose that p = 5, r = 5, and i = Then we have κ 0,i = T u 5, κ 1,i = (ρt 3708 T )u 5, κ 2,i = ρ 2 T 741 u 5, κ 3,i = (ρ 3 T 147 ρ 2 T 740 ρt 3708 )u 5, κ 4,i = ρ 4 T 29 u 5, κ 5,i = (ρ 5 T 4 + ρ 4 T 28 )u 5 + ρ 7 w, κ 6,i = ρ 6 u 5 + ρ 7 w. Remark It is not hard to see from the definition of σ(m, i) that σ(m, i) k for k < m if and only if p m k i k. Moreover, if for a given k there exists m > k such that σ(m, i) is less than k or not defined, then p i k so κ k,i = ρ k T θ k(i) u r unless r = p and i k+1 1 is a power of p or r = p 1 and i k = p. The previous examples illustrate some of this. Let us show that the κ m,i are actually elements of V i. In the process, we see how they compare to the elements α m,ψm(i) and β m,l previously defined. Proposition The elements κ m,i lie in V i for all nonnegative integers m. Proof. Suppose first that r p or i does not satisfy i m = p l + 1 for any l 0 (and omitting the case r = p 1 and ψ m (i) = 0, for which one should take the fractions in the following two equations to be 1). If ψ m (i) = θ m (i), then we have that κ m,i = [r]! {r δ ψ m (i)}! ρm α 0,ψm(i),

19 Galois module structure 19 and this lies in V i by the definition of θ m (i). If ψ m (i) = θ m (i) 1, we claim that κ m,i [r]! {r δ ψ m (i)}! ρσ(m,i) α m σ(m,i),ψm(i). (3.3.8) To see this, note that m σ(m,i) κ m,i = ρ σ(m,i) ρ m σ(m,i) T ψm(i) a m,i ρ m σ(m,i) k T θ m k(i) 1 u r. Lemma tells us that θ m k (i) 1 = φ k 1 (ψ m (i)) δ if and only if p m k+1 > p m i m i, and therefore if k m σ(m, i), proving the claim. (Note that we the reason we do not have actual equality in (3.3.8) is simply that we took a m,i to be an inverse to {r δ ψ m (i)}! modulo p, not in Z p.) Moreover, we have by Theorem that κ m,i V t with t = p σ(m,i) φ m σ(m,i) (ψ m (i)). k=1 Since p σ(m,i) p m i m i, Lemma implies that φ m σ(m,i) (ψ m (i)) δ θ σ(m,i) 1 (i), and Lemma then states that t i. Finally, if r = p and i m+1 = p l +1 for some l 0, then Lemma similarly implies that κ m,i = ρ σ(m+1,i) β m σ(m+1,i),l. By Theorem 3.2.4, we have in this case that κ m,i V t with t = p σ(m+1,i) φ m σ(m+1,i)(p l 1) i, the inequality again following from Lemmas and Generating sets In this subsection, we give explicit minimal generating sets of all of the A-modules V i in terms of the elements κ m,i of the previous section. We begin with generation. Recall that δ {0, 1} is 1 if and only if r = p. Theorem We let S i = {κ m,i 0 m s} for ( ) i + 1 s = log p. r δ(p 1) If 2 r p 1, then S i generates V i as an A-module, while if r = p, then S i {p log p (i) w} generates V i as an A-module.

20 Galois module structure 20 Proof. Let t = (i + 1) m 1. In the case that 2 r p 1, we have ψ m (i) = ψ(t) and ψ(t) > 0 if and only if i+1 > r + 1, or m < log p m p ( i+1 ). The smallest m such that r+1 ψ m (i) = 0 is therefore s. If r = p, then ψ m(i) = ψ(t) ɛ t, where ɛ t {0, 1} is 1 if and only if p l t p l+1 + p 1 for some l 0. In particular, we have ψ(t) > ɛ t if and only if t 2p, so the smallest m such that ψ m(i) = 0 is again s. It suffices to show that the images of our elements generate V i /V i+p 1. Suppose that α = (ρ k bt j + c)u r + dw V i for some j, k 0, b Z p [Φ] pz p [Φ], c T j+1 A, and d Z p [Φ] (with d = 0 if r p). Let m = min(k, s). Then j ψ m(i) by Theorems and and Corollary (and the fact that ψ s(i) = 0), and we set α = α ρ k m bt j ψ m (i) κ m,i V i A(T j+1 u r, w). If r p 1, we may repeat this process recursively until we obtain an element of V i+p 1. If r = p, either κ m,i Au p or κ m,i ρ m+l+1 w + Au p for some l 0 with i < p m+l+1 + p m+1. Since (T, p)p m+l+1 w V p m+l+2, there exists an element α V i (T j+1 Au r + Z p [Φ]w) with α α V i+p 1, and again we may repeat the process until we obtain an element of V i+p 1 plus an element of V i Z p [Φ]w = Z p [Φ]p log p (i) w. We will require the following lemma. Lemma If m 1 is such that θ m (i) 1, then θ m 1 (i) θ m (i) + 2. Proof. First, suppose that θ m (i) 1, and note that i m 1 p(i m 1) + 1. Therefore, 2 + {r im } θ m 1 (i) ψ(p(i m 1) + 1) = i m 1 + δ. (3.4.1) p On the other hand, im {r i m } θ m (i) = δ. (3.4.2) p In particular, we have that θ m (i) = 1 exactly when r + 1 i m r + (δ + 1)(p 1). In this case, θ m 1 (i) r + 1 δ 3 = θ m (i) + 2. In general, (3.4.1) and (3.4.2) tell us that θ m 1 (i) i m 1 δ and i m p + 1 δ θ m(i), and we have i m 1 δ i m p + 3 δ if and only if i m 4 p, which holds for i p 1 m r + p unless i m = 5, r = 2, and p = 3, in which case θ m 1 (i) 5 and θ m (i) = 2.

21 Galois module structure 21 For each m 0, let us set ɛ m (i) = θ m (i) ψ m(i), which lies in {0, 1} by Lemma and the remark before it. The following corollary is useful in understanding the form of our special elements. Corollary For every m 0, we have ψ m(i) ψ m+1(i), with equality if and only if ψ m(i) = 0. Proof. If θ m+1 (i) 1, Lemma and the fact that ɛ k (i) {0, 1} for all k imply that ψ m(i) > ψ m+1(i). Otherwise, ψ m+1(i) = 0, and the inequality holds automatically, with equality exactly if ψ m(i) = 0. We next show that the sets given in Theorem are minimal unless r = p. It is in the proof of this result that the refined elements κ m,i first hold an advantage of ease of use over the elements of Section 3.2. Theorem For r p 1, no proper subset of S i generates V i as an A-module. For r = p, every proper subset of S i {p log p (i) w} that generates V i as an A-module must contain S i. Proof. Assume first that 2 r p 1. Suppose that s c m κ m,i = 0, (3.4.3) m=0 where c m A for m s. We must show that no c m is a unit. We prove the somewhat stronger claim that c m (p, T ɛm(i)+1 ) for each m. Fix an nonnegative integer m s. If ɛ m (i) = 0, then κ m,i = ρ m T θm(i) u r by (3.3.5). If ɛ m (i) = 1, then (3.3.6) tells us that noting Lemma Set κ m,i ρ m T θm(i) 1 u r mod AT θm(i)+1 u r, X m = {k Z m < k s, ɛ k (i) = 1, σ(k, i) m}, (3.4.4) which is actually a set of cardinality at most one, though we do not need this fact. Let k s. If k X m, then (3.3.6) and Lemma together imply that κ k,i a k,i ρ m T θm(i) 1 u r mod (p m+1, T θm(i)+1 )u r, and if k / X m, they and (3.3.5) similarly imply that κ k,i (p m, T θm(i)+1 )u r. Thus, (3.4.3) yields the congruence c m ρ m T ψm(i) c k a k,i ρ m T θm(i) 1 mod (p m+1, T θm+1(i)+1 ). k X m (3.4.5)

22 Galois module structure 22 If the claim holds for all k > m, then we have c k (p, T 2 ) for each k X m, so c m (p, T ɛm(i)+1 ), as desired. If r = p, a completely analogous argument shows that at most p log p (i) w is unnecessary for generation, if one works modulo Aw = Z p [Φ]w + A(ϕ 1)u p throughout. Here, one should replace X m by X m = {k Z m < k s, ɛ k (i) = 1, σ (k, i) m}, (3.4.6) where we set σ (k, i) = σ(k, i) unless i k+1 = p l + 1 for some l 0, in which case we set σ (k, i) = σ(k + 1, i). For the purpose of completeness, we also give the precise condition on i under which no proper subset of S i {p log p (i) w} generates V i in the case that r = p. Proposition For r = p, the set S i generates V i if and only if i s = p + 1. Proof. To determine whether p log p (i) w is or is not necessary, we work in distinct ranges of i separately. Note that the definition of s forces 2p s < i < 2p s+1. Case 1: 2p s < i p s+1. In this case, all of the elements κ m,i lie in Au p, and therefore p s+1 w is necessary. Case 2: p s+1 < i p s+1 + p s p s 1. In this range, we have κ s,i = ρ s u p + ρ s+1 w and κ s 1,i = ρ s 1 T p 1 u p + ρ s+1 w. Note that so (T p)κ s,i = ρ s (T p)u p + ρ s+1 (ϕ 1)u p = ρ s (T ρ)u p, ρ s T u p ρ s+1 u p mod Aκ s,i and ρ s u p ρ s+1 w mod Aκ s,i. (3.4.7) Applying these to ρκ s 1,i, we obtain ρκ s 1,i ρ s+p 1 u p + ρ s+2 w ( ρ s+p + ρ s+2 )w mod Aκ s,i, which in particular tells us that p s+2 w A(κ s 1,i, κ s,i ). Case 3: p s+1 + p s p s 1 < i p s+1 + p s. In this range, we have κ s,i = ρ s u p + ρ s+1 w and κ s 1,i = ρ s 1 T p 1 + s 2 k=σ(s,i) with σ(s, i) s 2. Moreover, κ m,i Au r for m s 2. ρ k T θ k(i) 1 u p + ρ s+1 w

23 Galois module structure 23 Set ν k,i = ρ k T θ k(i) u r for all nonnegative k. We note that ν m,i A(κ 0,i,..., κ m,i ) for m s 2: if κ m,i ν m,i, which is to say ɛ m (i) = 1, then ν m,i = T κ m,i + a m,i m 1 k=σ(m,i) ν k,i. Let j = θ s 2 (i) p, and note that j p Since T j κ s 1,i ρ s 1 T θ s 2(i) 1 u p + ρ s+1 T j w mod A(ν σ(s,i),i,..., ν s 2,i ) and ρ k+1 T θ k(i) 1 u p Aν k+1,i for all k with σ(s, i) k s 3, we therefore have that (ρ T j )κ s 1,i ρ s T p 1 u p + ρ s+1 (ρ T j )w mod A(ν σ(s,i),i,..., ν s 2,i ). (3.4.8) Using (3.4.7) to reduce (3.4.8), we see that (ρ T j )κ s 1,i ρ s+2 (1 ρ p 2 ρ j 1 )w mod A(ν σ(s,i),i,..., ν s 2,i, κ s,i ), which implies that p s+2 w A(κ 0,i,..., κ s,i ). Case 4: p s+1 + p s < i < 2p s+1. In this case, all of the κ m,i with m s 1 lie in AT p u p, and so for p s+2 w to be unnecessary, there would have to exist c A such that Note that cκ s,i p s+2 w mod AT 2 u p. (3.4.9) cκ s,i c(ρ s u p + ρ s+1 w) mod AT 2 u p, and this forces c T 2 c mod (ϕ 1) for some c A. This means that cκ s,i c p s+3 w mod A(u p, (ϕ 1)w), but p s+2 w / A(p s+3 w, (ϕ 1)w, u p ), so (3.4.9) cannot hold. 4 The finite level 4.1 Norms and eigenspace structure In this section, we explore the consequences of the results of Section 3 for unit groups of actual abelian local fields of characteristic 0. Fix a positive integer n. Recall from the introduction that F n is the field obtained from E by adjoining the p n th roots of unity and that U n,t denotes the tth unit group of F n for t 1. As before, we set Γ n = Gal(F n /F 1 ). For positive integers m n, let N m,n and Tr m,n denote, respectively, the norm and trace from F m to F n. We also let N n denote the restriction map N n : F F n on norm compatible sequences. Recall that λ n = N n (λ) = 1 ζ p n, where ζ p n = N n (ζ) is a primitive p n th root of unity. We will require a few preliminary lemmas.

24 Galois module structure 24 Lemma One has for all k 1 and ɛ {0, 1}. Proof. An easy calculation shows that for every t 0. Since ( ) ( pk ɛ k ɛ = pj j for any j 0, we have the result. Tr n+1,n (λ pk ɛ n+1 ) pλ k ɛ n mod p 3 t p ( ) t Tr n+1,n (λ t n+1) = p ( ζ p n) j pj ) p(k j) s=1 p s j=0 ( 1 + pj ) s ( ) k ɛ mod p 2 j Let e n = p n 1 (p 1) denote the ramification index of F n. In applying Lemma 4.1.1, it is useful to make note of the fact that p λ en n mod λ pn n. (4.1.1) Lemma For t 1 and any unit η in E, one has 1 + η p λ t n mod λ t+1 n if t < p n 1, N n+1,n (1 + ηλ t n+1) 1 + (η p η)λ pn ɛ n 1 ηλ en+k ɛ n mod λ pn +1 ɛ n if t = p n ɛ, ɛ {0, 1}, mod λ en+k+1 ɛ n if t = pk ɛ > p n, ɛ {0, 1}. Moreover, we have for all t > p n. N n+1,n (1 + ηλ t n+1) 1 mod λ en+ t p n. Proof. The jump in the ramification filtration of Gal(F n+1 /F n ) occurs at p n 1. By [Se, Lemmas V.4 and V.5], we have and N n+1,n (1 + ηλ t n+1) 1 + η Tr n+1,n (λ t n+1) + η p λ t n mod λ en+ 2t/p n Tr n+1,n (λ t n+1) 0 mod λ en+ t/p n. The result is then a corollary of Lemma 4.1.1, upon applying (4.1.1).

25 Galois module structure 25 Let D n be the pro-p completion of F n, and let D n (r) = Dn εr for any r Z. As before, we fix r with 2 r p, and i will always denote a positive integer with i r mod p 1. Let V n,i = U εr n,i = U n,i D n (r) for any such i. These V n,i are all modules over A n = Z p [Γ n Φ]. As in Lemma 2.1, we have isomorphisms V n,i /V n,i+p 1 F q that send 1 + xλ i n for some x in the valuation ring of F n to the element x of F q that is identified with the image of x in the residue field of E under the isomorphism fixed in Section 2. We may then set V n,i = V n,i V n,i+p 1 and define V n,i (η) for η F q as the set of elements 1 + xλ i n with x = η. We have the following consequence of Lemma Lemma For any t 1, we have with equality for t 0. Proof. Note that Lemma tells us that N n+1,n (V n+1,p n +t) V n,p n +t (p 1) t+1 p, N n+1,n (U n+1,p n +pk ɛ) = U n,p n +k ɛ for all k 0 and ɛ {0, 1} with k ɛ, since every element in U n,p n +k ɛ can be written as a product of elements of the form 1 + η t λ pn +t n with t k ɛ and η t F q. (For k = 0 and ɛ = 1, it tells us just that any element of U n+1,p n 1 has a norm in U n,p n 1.) Note that U εr n,p n +k ɛ = V n,p n +k ɛ+{r k+ɛ 1} and U εr n+1,p n +pk ɛ = V n,p n +pk ɛ+{r k+ɛ 1}. For any t 0, we may write t = pk ɛ + {r k + ɛ 1} for some k, ɛ, and r, and we have t + 1 t (p 1) = k ɛ + {r k + ɛ 1}, p hence the result. The next corollary is almost immediate from Lemmas and 4.1.3, so we leave it to the reader. Corollary For any unit η in E, one has V n,i (η p ) if i < p n 1, N n+1,n (V n+1,i (η)) V n,i (η p η) if i = p n 1, V n,p n +k 1( η) if i = p n + pk 1 for some k > 0, with equality if r p 1 or i > p n.

26 Galois module structure 26 As for the p-power map, we have the following well-known and easy-to-prove fact. Lemma Suppose that i > p n 1. Then the pth power map induces an isomorphism V n,i V n,i+en, and we have V n,i (η) p = V n,i+en ( η) for all η F q. Next, we discuss the restriction map from the field of norms to the finite level. Proposition The map N n induces maps N n : D (r) D n (r) that are surjections for r p 1 and which have procyclic cokernel for r = p 1. For η F q, we have V n,i (η p n ) if i p n 2, N n (V i (η)) V n,p n 1(η p n η p n 1 ) if i = p n 1, V n,p n +k 1( η p n 1 ) if i = p n + pk 1 for some 0 < k < e n. Moreover, we have induced maps V i /V i+1 V n,i /V n,i+1 for all i < p n, and these are isomorphisms for i p n 1. For i p n, we have that V n,i = N n V i if r p 1, and V n,i /N n V i is procyclic if r = p 1. Proof. That the cokernel of N n is trivial (resp., procyclic) if r p 1 (resp., r = p 1) follows easily from local class field theory, but it is also a consequence of the argument that follows. The first jump in the ramification filtration of Gal(F /F n+1 ) is at p n+1 1. In particular, for t less than this value, repeated application of Lemma tells us that N n+1 (1 + ηλ t ) = lim N m,n+1(1 + η p m λ t m) 1 + η p n 1 λ t n+1 mod λ t+1 n+1. m Moreover, repeated application of Corollary followed by two applications of Lemma tells us that N n (V p n+1 1+{r}) V n,p n +e n 1+{r}. An application of Corollary then yields the stated containments. Since η p n and η p n 1 run through all elements of F q as as η F q varies, we obtain V n,i = N n V i + V n,i+p 1 for all i p n but p n 1. Noting Lemma 4.1.5, this implies V n,i+ken = N n V p k i + V n,i+ken+p 1 for p n 1 < i p n with i p n 1 and k 0. Note that every element of every V n,i may be written as an infinite product over j 0 of one element from each of a fixed set of representatives of the V n,i+j(p 1) /V n,i+(j+1)(p 1). Thus, we have that N n V i = V n,i so long as r p 1.

27 Galois module structure 27 If r = p 1, we can choose an element z n of V n,p n 1(ξ) that is not a norm. By the above-proven formula for N n (1 + ηλ pn 1 ) modulo λ pn n, we have that V n,p n +ke n 1 = N n V p k (p n 1) + V n,p n +ke n+p 2 + z pk n for k = 0, and then for all k 0 by taking powers. Therefore, V n,i /N n V i is generated by z n for all i < p n with i 0 mod p 1. The following structural result is again essentially found in [Gr], without the stated congruences. Here, we derive it from more basic principles. Theorem For r p 2, the A n -module D n (r) by an element u n,r V n,r (ξ). The A n -module D n (p 1) D (p 1) n = π n, u n,p 1, v n π ϕ n = π n, π γ 1 n is freely generated as an A n -module has a presentation = u N Φ n,p 1, v γ n = v n, u N Γn n,p 1 = v 1 ϕ n, where v = vn ϕ2 n 1 + pξ mod p 2 is independent of n and u n,p 1 V n,p 1 (ξ) for n 2, while u 1,p 1 V n,p 1 (ξ ξ p 1 ). The A n -module D n (p) has a presentation D (p) n = u n,p, w n w γ 1 p n = u ϕ 1 n,p with u n,p V n,p (ξ) and w n V n,1 ( ξ) such that w N Φ n = ζ p n. Proof. We set u n,r = (N n u r ) ϕn, π n = N n π, and w n = (N n w) ϕn with u r, π, and w as in Theorem It follows from the surjectivity of N n for r p 1 in Proposition that the element u n,r generates D n (r) for r p 2, while the elements w n and u n,r generate D n (p). By Hilbert s Theorem 90, the kernel of N n consists exactly of elements of the form α γpn 1 1 with α D, and therefore it follows that D n (r) is free of rank 1 on u n,r over A n for r p 2 and that D n (p) has the stated presentation. (That u 1,p V 1,p (ξ) requires a simple check using Propositions and ) The elements π n and u n,p 1 automatically satisfy the first two relations in the desired presentation of D n (p 1). In particular, u N Γn Φ n,p 1 = π (γ 1)N Γn n = 1, so Hilbert s Theorem 90 tells us that u N Γn n,p 1 = vn 1 ϕ for some v n in the pro-p completion of E. By Proposition 4.1.6, we have that u N Γn n,p (ξ pn 1 ξ pn 2 )λ p 1 1 mod λ p 1. Noting (4.1.1), we may in fact choose v n 1 + pϕ n 2 (ξ) mod p 2 with v = vn ϕ2 n independent of n. HiIbert s theorem 90 and Theorem tell us that the A n -module generated by u n,p 1 is isomorphic to A n /(N Γn Φ). By Proposition 4.1.6, the cokernel of N n on D (p 1)

28 Galois module structure 28 is isomorphic to Z p. We claim that the image of v topologically generates this cokernel. If this is the case, then clearly D n (p 1) is generated by π n, u n,p 1, and v, and any solution with b, d Z p and c A n to πnu b c n,p 1v d = 1 must satisfy b = d = 0 and c Z p N Γn Φ. It remains only to demonstrate the claim. Suppose by way of contradiction that there exists a A n such that x = vu a n,p 1 is a pth power in D n (p 1). This implies that x γ 1 = u a(γ 1) n,p 1 is a pth power in the A n -module generated by u n,p 1. It follows that a(γ 1) A n (p, N Γn Φ), which forces a(γ 1) 0 mod p, so a A n (p, N Γn ). It then suffices to show that vu bn Γn n,p 1 = v 1+bϕn 2 (1 ϕ) is not a pth power in F n for any b Z p [Φ]. If it were for some b, then v N Φ and hence 1 + p would be a pth power in F n as well, but this is clearly not the case. 4.2 Special elements We assume for the rest of the paper that n 2, the case that n = 1 being slightly exceptional but also completely straightforward. In this subsection, we construct special elements in the groups in the unit filtration of F n. Aside from the case that r = p 1, these arise as restrictions of the elements introduced in Section 3.2. Note that Z p [Γ n ] = Z p [T ]/(f n ), where f n = (T +1) pn 1 1. Of course, we can then speak of the action of T on an element of D n (r). Once again reverting to additive notation, the following is now an immediate corollary of Theorem and Proposition Proposition Let m and j be nonnegative integers with φ m (j) < p n 1. Define ( ) α n,m,j = 1 m {r δ j}!ρ m T j ρ m k T φ k 1(j) δ u n,r, [r]! unless j = 0 and r = p 1, in which case we replace {r δ j}! with 1 in the formula. Then α n,m,j V n,φm(j)(ξ). Furthermore, k=1 for all b Z p [Φ] pz p [Φ] and c T j+1 A n. (p m bt j + c)u n,r / V n,φm(j)+p 1 For nonnegative m n 2, define φ n,m : Z 0 Z 0 by φ n,m(j) = φ m(j) unless r = p 1 and j = e n m 1, in which case we set φ n,m(e n m 1 ) = e n + p m+1 1 = φ m (e n m 1 ) + p m (p 1).

Lemma 1.1. The field K embeds as a subfield of Q(ζ D ).

Lemma 1.1. The field K embeds as a subfield of Q(ζ D ). Math 248A. Quadratic characters associated to quadratic fields The aim of this handout is to describe the quadratic Dirichlet character naturally associated to a quadratic field, and to express it in terms

More information

Galois groups with restricted ramification

Galois groups with restricted ramification Galois groups with restricted ramification Romyar Sharifi Harvard University 1 Unique factorization: Let K be a number field, a finite extension of the rational numbers Q. The ring of integers O K of K

More information

NOTES ON FINITE FIELDS

NOTES ON FINITE FIELDS NOTES ON FINITE FIELDS AARON LANDESMAN CONTENTS 1. Introduction to finite fields 2 2. Definition and constructions of fields 3 2.1. The definition of a field 3 2.2. Constructing field extensions by adjoining

More information

CLASS FIELD THEORY WEEK Motivation

CLASS FIELD THEORY WEEK Motivation CLASS FIELD THEORY WEEK 1 JAVIER FRESÁN 1. Motivation In a 1640 letter to Mersenne, Fermat proved the following: Theorem 1.1 (Fermat). A prime number p distinct from 2 is a sum of two squares if and only

More information

= 1 2x. x 2 a ) 0 (mod p n ), (x 2 + 2a + a2. x a ) 2

= 1 2x. x 2 a ) 0 (mod p n ), (x 2 + 2a + a2. x a ) 2 8. p-adic numbers 8.1. Motivation: Solving x 2 a (mod p n ). Take an odd prime p, and ( an) integer a coprime to p. Then, as we know, x 2 a (mod p) has a solution x Z iff = 1. In this case we can suppose

More information

The Kronecker-Weber Theorem

The Kronecker-Weber Theorem The Kronecker-Weber Theorem November 30, 2007 Let us begin with the local statement. Theorem 1 Let K/Q p be an abelian extension. Then K is contained in a cyclotomic extension of Q p. Proof: We give the

More information

A BRIEF INTRODUCTION TO LOCAL FIELDS

A BRIEF INTRODUCTION TO LOCAL FIELDS A BRIEF INTRODUCTION TO LOCAL FIELDS TOM WESTON The purpose of these notes is to give a survey of the basic Galois theory of local fields and number fields. We cover much of the same material as [2, Chapters

More information

ON GALOIS GROUPS OF ABELIAN EXTENSIONS OVER MAXIMAL CYCLOTOMIC FIELDS. Mamoru Asada. Introduction

ON GALOIS GROUPS OF ABELIAN EXTENSIONS OVER MAXIMAL CYCLOTOMIC FIELDS. Mamoru Asada. Introduction ON GALOIS GROUPS OF ABELIAN ETENSIONS OVER MAIMAL CYCLOTOMIC FIELDS Mamoru Asada Introduction Let k 0 be a finite algebraic number field in a fixed algebraic closure Ω and ζ n denote a primitive n-th root

More information

Lecture 8: The Field B dr

Lecture 8: The Field B dr Lecture 8: The Field B dr October 29, 2018 Throughout this lecture, we fix a perfectoid field C of characteristic p, with valuation ring O C. Fix an element π C with 0 < π C < 1, and let B denote the completion

More information

(1) A frac = b : a, b A, b 0. We can define addition and multiplication of fractions as we normally would. a b + c d

(1) A frac = b : a, b A, b 0. We can define addition and multiplication of fractions as we normally would. a b + c d The Algebraic Method 0.1. Integral Domains. Emmy Noether and others quickly realized that the classical algebraic number theory of Dedekind could be abstracted completely. In particular, rings of integers

More information

ON TAMELY RAMIFIED IWASAWA MODULES FOR THE CYCLOTOMIC p -EXTENSION OF ABELIAN FIELDS

ON TAMELY RAMIFIED IWASAWA MODULES FOR THE CYCLOTOMIC p -EXTENSION OF ABELIAN FIELDS Itoh, T. Osaka J. Math. 51 (2014), 513 536 ON TAMELY RAMIFIED IWASAWA MODULES FOR THE CYCLOTOMIC p -EXTENSION OF ABELIAN FIELDS TSUYOSHI ITOH (Received May 18, 2012, revised September 19, 2012) Abstract

More information

p-adic fields Chapter 7

p-adic fields Chapter 7 Chapter 7 p-adic fields In this chapter, we study completions of number fields, and their ramification (in particular in the Galois case). We then look at extensions of the p-adic numbers Q p and classify

More information

Formal power series rings, inverse limits, and I-adic completions of rings

Formal power series rings, inverse limits, and I-adic completions of rings Formal power series rings, inverse limits, and I-adic completions of rings Formal semigroup rings and formal power series rings We next want to explore the notion of a (formal) power series ring in finitely

More information

Reciprocity maps with restricted ramification

Reciprocity maps with restricted ramification Reciprocity maps with restricted ramification Romyar Sharifi UCLA January 6, 2016 1 / 19 Iwasawa theory for modular forms Let be an p odd prime and f a newform of level N. Suppose that f is ordinary at

More information

A SIMPLE PROOF OF KRONECKER-WEBER THEOREM. 1. Introduction. The main theorem that we are going to prove in this paper is the following: Q ab = Q(ζ n )

A SIMPLE PROOF OF KRONECKER-WEBER THEOREM. 1. Introduction. The main theorem that we are going to prove in this paper is the following: Q ab = Q(ζ n ) A SIMPLE PROOF OF KRONECKER-WEBER THEOREM NIZAMEDDIN H. ORDULU 1. Introduction The main theorem that we are going to prove in this paper is the following: Theorem 1.1. Kronecker-Weber Theorem Let K/Q be

More information

Higher Ramification Groups

Higher Ramification Groups COLORADO STATE UNIVERSITY MATHEMATICS Higher Ramification Groups Dean Bisogno May 24, 2016 1 ABSTRACT Studying higher ramification groups immediately depends on some key ideas from valuation theory. With

More information

KUMMER S LEMMA KEITH CONRAD

KUMMER S LEMMA KEITH CONRAD KUMMER S LEMMA KEITH CONRAD Let p be an odd prime and ζ ζ p be a primitive pth root of unity In the ring Z[ζ], the pth power of every element is congruent to a rational integer mod p, since (c 0 + c 1

More information

FIELD THEORY. Contents

FIELD THEORY. Contents FIELD THEORY MATH 552 Contents 1. Algebraic Extensions 1 1.1. Finite and Algebraic Extensions 1 1.2. Algebraic Closure 5 1.3. Splitting Fields 7 1.4. Separable Extensions 8 1.5. Inseparable Extensions

More information

FIXED-POINT FREE ENDOMORPHISMS OF GROUPS RELATED TO FINITE FIELDS

FIXED-POINT FREE ENDOMORPHISMS OF GROUPS RELATED TO FINITE FIELDS FIXED-POINT FREE ENDOMORPHISMS OF GROUPS RELATED TO FINITE FIELDS LINDSAY N. CHILDS Abstract. Let G = F q β be the semidirect product of the additive group of the field of q = p n elements and the cyclic

More information

Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra

Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra Course 311: Michaelmas Term 2005 Part III: Topics in Commutative Algebra D. R. Wilkins Contents 3 Topics in Commutative Algebra 2 3.1 Rings and Fields......................... 2 3.2 Ideals...............................

More information

GALOIS THEORY. Contents

GALOIS THEORY. Contents GALOIS THEORY MARIUS VAN DER PUT & JAAP TOP Contents 1. Basic definitions 1 1.1. Exercises 2 2. Solving polynomial equations 2 2.1. Exercises 4 3. Galois extensions and examples 4 3.1. Exercises. 6 4.

More information

Workshop on Serre s Modularity Conjecture: the level one case

Workshop on Serre s Modularity Conjecture: the level one case Workshop on Serre s Modularity Conjecture: the level one case UC Berkeley Monte Verità 13 May 2009 Notation We consider Serre-type representations of G = Gal(Q/Q). They will always be 2-dimensional, continuous

More information

MA257: INTRODUCTION TO NUMBER THEORY LECTURE NOTES

MA257: INTRODUCTION TO NUMBER THEORY LECTURE NOTES MA257: INTRODUCTION TO NUMBER THEORY LECTURE NOTES 2018 57 5. p-adic Numbers 5.1. Motivating examples. We all know that 2 is irrational, so that 2 is not a square in the rational field Q, but that we can

More information

LECTURE 2. Hilbert Symbols

LECTURE 2. Hilbert Symbols LECTURE 2 Hilbert Symbols Let be a local field over Q p (though any local field suffices) with char() 2. Note that this includes fields over Q 2, since it is the characteristic of the field, and not the

More information

Problems on Growth of Hecke fields

Problems on Growth of Hecke fields Problems on Growth of Hecke fields Haruzo Hida Department of Mathematics, UCLA, Los Angeles, CA 90095-1555, U.S.A. A list of conjectures/problems related to my talk in Simons Conference in January 2014

More information

CYCLOTOMIC FIELDS CARL ERICKSON

CYCLOTOMIC FIELDS CARL ERICKSON CYCLOTOMIC FIELDS CARL ERICKSON Cyclotomic fields are an interesting laboratory for algebraic number theory because they are connected to fundamental problems - Fermat s Last Theorem for example - and

More information

Math 210B. Artin Rees and completions

Math 210B. Artin Rees and completions Math 210B. Artin Rees and completions 1. Definitions and an example Let A be a ring, I an ideal, and M an A-module. In class we defined the I-adic completion of M to be M = lim M/I n M. We will soon show

More information

Algebra SEP Solutions

Algebra SEP Solutions Algebra SEP Solutions 17 July 2017 1. (January 2017 problem 1) For example: (a) G = Z/4Z, N = Z/2Z. More generally, G = Z/p n Z, N = Z/pZ, p any prime number, n 2. Also G = Z, N = nz for any n 2, since

More information

24 Artin reciprocity in the unramified case

24 Artin reciprocity in the unramified case 18.785 Number theory I Fall 2017 ecture #24 11/29/2017 24 Artin reciprocity in the unramified case et be an abelian extension of number fields. In ecture 22 we defined the norm group T m := N (I m )R m

More information

ALGEBRA 11: Galois theory

ALGEBRA 11: Galois theory Galois extensions Exercise 11.1 (!). Consider a polynomial P (t) K[t] of degree n with coefficients in a field K that has n distinct roots in K. Prove that the ring K[t]/P of residues modulo P is isomorphic

More information

DONG QUAN NGOC NGUYEN

DONG QUAN NGOC NGUYEN REPRESENTATION OF UNITS IN CYCLOTOMIC FUNCTION FIELDS DONG QUAN NGOC NGUYEN Contents 1 Introduction 1 2 Some basic notions 3 21 The Galois group Gal(K /k) 3 22 Representation of integers in O, and the

More information

A talk given at University of Wisconsin at Madison (April 6, 2006). Zhi-Wei Sun

A talk given at University of Wisconsin at Madison (April 6, 2006). Zhi-Wei Sun A tal given at University of Wisconsin at Madison April 6, 2006. RECENT PROGRESS ON CONGRUENCES INVOLVING BINOMIAL COEFFICIENTS Zhi-Wei Sun Department of Mathematics Nanjing University Nanjing 210093,

More information

Notes on p-divisible Groups

Notes on p-divisible Groups Notes on p-divisible Groups March 24, 2006 This is a note for the talk in STAGE in MIT. The content is basically following the paper [T]. 1 Preliminaries and Notations Notation 1.1. Let R be a complete

More information

Cover Page. The handle holds various files of this Leiden University dissertation

Cover Page. The handle   holds various files of this Leiden University dissertation Cover Page The handle http://hdl.handle.net/1887/25833 holds various files of this Leiden University dissertation Author: Palenstijn, Willem Jan Title: Radicals in Arithmetic Issue Date: 2014-05-22 Chapter

More information

Quadratic reciprocity (after Weil) 1. Standard set-up and Poisson summation

Quadratic reciprocity (after Weil) 1. Standard set-up and Poisson summation (September 17, 010) Quadratic reciprocity (after Weil) Paul Garrett garrett@math.umn.edu http://www.math.umn.edu/ garrett/ I show that over global fields (characteristic not ) the quadratic norm residue

More information

Definitions. Notations. Injective, Surjective and Bijective. Divides. Cartesian Product. Relations. Equivalence Relations

Definitions. Notations. Injective, Surjective and Bijective. Divides. Cartesian Product. Relations. Equivalence Relations Page 1 Definitions Tuesday, May 8, 2018 12:23 AM Notations " " means "equals, by definition" the set of all real numbers the set of integers Denote a function from a set to a set by Denote the image of

More information

Class Field Theory. Anna Haensch. Spring 2012

Class Field Theory. Anna Haensch. Spring 2012 Class Field Theory Anna Haensch Spring 202 These are my own notes put together from a reading of Class Field Theory by N. Childress [], along with other references, [2], [4], and [6]. Goals and Origins

More information

FORMAL GROUPS OF CERTAIN Q-CURVES OVER QUADRATIC FIELDS

FORMAL GROUPS OF CERTAIN Q-CURVES OVER QUADRATIC FIELDS Sairaiji, F. Osaka J. Math. 39 (00), 3 43 FORMAL GROUPS OF CERTAIN Q-CURVES OVER QUADRATIC FIELDS FUMIO SAIRAIJI (Received March 4, 000) 1. Introduction Let be an elliptic curve over Q. We denote by ˆ

More information

NUNO FREITAS AND ALAIN KRAUS

NUNO FREITAS AND ALAIN KRAUS ON THE DEGREE OF THE p-torsion FIELD OF ELLIPTIC CURVES OVER Q l FOR l p NUNO FREITAS AND ALAIN KRAUS Abstract. Let l and p be distinct prime numbers with p 3. Let E/Q l be an elliptic curve with p-torsion

More information

Math 429/581 (Advanced) Group Theory. Summary of Definitions, Examples, and Theorems by Stefan Gille

Math 429/581 (Advanced) Group Theory. Summary of Definitions, Examples, and Theorems by Stefan Gille Math 429/581 (Advanced) Group Theory Summary of Definitions, Examples, and Theorems by Stefan Gille 1 2 0. Group Operations 0.1. Definition. Let G be a group and X a set. A (left) operation of G on X is

More information

Maximal Class Numbers of CM Number Fields

Maximal Class Numbers of CM Number Fields Maximal Class Numbers of CM Number Fields R. C. Daileda R. Krishnamoorthy A. Malyshev Abstract Fix a totally real number field F of degree at least 2. Under the assumptions of the generalized Riemann hypothesis

More information

Algebraic Number Theory Notes: Local Fields

Algebraic Number Theory Notes: Local Fields Algebraic Number Theory Notes: Local Fields Sam Mundy These notes are meant to serve as quick introduction to local fields, in a way which does not pass through general global fields. Here all topological

More information

On Galois cohomology and realizability of 2-groups as Galois groups II. Author copy

On Galois cohomology and realizability of 2-groups as Galois groups II. Author copy Cent. Eur. J. Math. 9(6) 2011 1333-1343 DOI: 10.2478/s11533-011-0086-z Central European Journal of Mathematics On Galois cohomology and realizability of 2-groups as Galois groups II Research Article Ivo

More information

The primitive root theorem

The primitive root theorem The primitive root theorem Mar Steinberger First recall that if R is a ring, then a R is a unit if there exists b R with ab = ba = 1. The collection of all units in R is denoted R and forms a group under

More information

Some algebraic number theory and the reciprocity map

Some algebraic number theory and the reciprocity map Some algebraic number theory and the reciprocity map Ervin Thiagalingam September 28, 2015 Motivation In Weinstein s paper, the main problem is to find a rule (reciprocity law) for when an irreducible

More information

Chapter 8. P-adic numbers. 8.1 Absolute values

Chapter 8. P-adic numbers. 8.1 Absolute values Chapter 8 P-adic numbers Literature: N. Koblitz, p-adic Numbers, p-adic Analysis, and Zeta-Functions, 2nd edition, Graduate Texts in Mathematics 58, Springer Verlag 1984, corrected 2nd printing 1996, Chap.

More information

PERFECT POLYNOMIALS OVER F p WITH p + 1 IRREDUCIBLE DIVISORS. 1. Introduction. Let p be a prime number. For a monic polynomial A F p [x] let d

PERFECT POLYNOMIALS OVER F p WITH p + 1 IRREDUCIBLE DIVISORS. 1. Introduction. Let p be a prime number. For a monic polynomial A F p [x] let d PERFECT POLYNOMIALS OVER F p WITH p + 1 IRREDUCIBLE DIVISORS L. H. GALLARDO and O. RAHAVANDRAINY Abstract. We consider, for a fixed prime number p, monic polynomials in one variable over the finite field

More information

Introduction to Arithmetic Geometry Fall 2013 Lecture #24 12/03/2013

Introduction to Arithmetic Geometry Fall 2013 Lecture #24 12/03/2013 18.78 Introduction to Arithmetic Geometry Fall 013 Lecture #4 1/03/013 4.1 Isogenies of elliptic curves Definition 4.1. Let E 1 /k and E /k be elliptic curves with distinguished rational points O 1 and

More information

Winter School on Galois Theory Luxembourg, February INTRODUCTION TO PROFINITE GROUPS Luis Ribes Carleton University, Ottawa, Canada

Winter School on Galois Theory Luxembourg, February INTRODUCTION TO PROFINITE GROUPS Luis Ribes Carleton University, Ottawa, Canada Winter School on alois Theory Luxembourg, 15-24 February 2012 INTRODUCTION TO PROFINITE ROUPS Luis Ribes Carleton University, Ottawa, Canada LECTURE 2 2.1 ENERATORS OF A PROFINITE ROUP 2.2 FREE PRO-C ROUPS

More information

Finite Fields. [Parts from Chapter 16. Also applications of FTGT]

Finite Fields. [Parts from Chapter 16. Also applications of FTGT] Finite Fields [Parts from Chapter 16. Also applications of FTGT] Lemma [Ch 16, 4.6] Assume F is a finite field. Then the multiplicative group F := F \ {0} is cyclic. Proof Recall from basic group theory

More information

Contradiction. Theorem 1.9. (Artin) Let G be a finite group of automorphisms of E and F = E G the fixed field of G. Then [E : F ] G.

Contradiction. Theorem 1.9. (Artin) Let G be a finite group of automorphisms of E and F = E G the fixed field of G. Then [E : F ] G. 1. Galois Theory 1.1. A homomorphism of fields F F is simply a homomorphism of rings. Such a homomorphism is always injective, because its kernel is a proper ideal (it doesnt contain 1), which must therefore

More information

Algebraic Number Theory

Algebraic Number Theory TIFR VSRP Programme Project Report Algebraic Number Theory Milind Hegde Under the guidance of Prof. Sandeep Varma July 4, 2015 A C K N O W L E D G M E N T S I would like to express my thanks to TIFR for

More information

Modules Over Principal Ideal Domains

Modules Over Principal Ideal Domains Modules Over Principal Ideal Domains Brian Whetter April 24, 2014 This work is licensed under the Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License. To view a copy of this

More information

CYCLOTOMIC POLYNOMIALS

CYCLOTOMIC POLYNOMIALS CYCLOTOMIC POLYNOMIALS 1. The Derivative and Repeated Factors The usual definition of derivative in calculus involves the nonalgebraic notion of limit that requires a field such as R or C (or others) where

More information

Dirichlet Series Associated with Cubic and Quartic Fields

Dirichlet Series Associated with Cubic and Quartic Fields Dirichlet Series Associated with Cubic and Quartic Fields Henri Cohen, Frank Thorne Institut de Mathématiques de Bordeaux October 23, 2012, Bordeaux 1 Introduction I Number fields will always be considered

More information

Cover Page. The handle holds various files of this Leiden University dissertation.

Cover Page. The handle   holds various files of this Leiden University dissertation. Cover Page The handle http://hdl.handle.net/1887/20310 holds various files of this Leiden University dissertation. Author: Jansen, Bas Title: Mersenne primes and class field theory Date: 2012-12-18 Chapter

More information

On the image of noncommutative local reciprocity map

On the image of noncommutative local reciprocity map On the image of noncommutative local reciprocity map Ivan Fesenko 0. Introduction First steps in the direction of an arithmetic noncommutative local class field theory were described in [] as an attempt

More information

Cover Page. The handle holds various files of this Leiden University dissertation

Cover Page. The handle   holds various files of this Leiden University dissertation Cover Page The handle http://hdl.handle.net/1887/32076 holds various files of this Leiden University dissertation Author: Junjiang Liu Title: On p-adic decomposable form inequalities Issue Date: 2015-03-05

More information

The group (Z/nZ) February 17, In these notes we figure out the structure of the unit group (Z/nZ) where n > 1 is an integer.

The group (Z/nZ) February 17, In these notes we figure out the structure of the unit group (Z/nZ) where n > 1 is an integer. The group (Z/nZ) February 17, 2016 1 Introduction In these notes we figure out the structure of the unit group (Z/nZ) where n > 1 is an integer. If we factor n = p e 1 1 pe, where the p i s are distinct

More information

and this makes M into an R-module by (1.2). 2

and this makes M into an R-module by (1.2). 2 1. Modules Definition 1.1. Let R be a commutative ring. A module over R is set M together with a binary operation, denoted +, which makes M into an abelian group, with 0 as the identity element, together

More information

On the Grassmann modules for the symplectic groups

On the Grassmann modules for the symplectic groups On the Grassmann modules for the symplectic groups Bart De Bruyn Ghent University, Department of Pure Mathematics and Computer Algebra, Krijgslaan 81 (S), B-9000 Gent, Belgium, E-mail: bdb@cage.ugent.be

More information

1 Rings 1 RINGS 1. Theorem 1.1 (Substitution Principle). Let ϕ : R R be a ring homomorphism

1 Rings 1 RINGS 1. Theorem 1.1 (Substitution Principle). Let ϕ : R R be a ring homomorphism 1 RINGS 1 1 Rings Theorem 1.1 (Substitution Principle). Let ϕ : R R be a ring homomorphism (a) Given an element α R there is a unique homomorphism Φ : R[x] R which agrees with the map ϕ on constant polynomials

More information

MATH 205B NOTES 2010 COMMUTATIVE ALGEBRA 53

MATH 205B NOTES 2010 COMMUTATIVE ALGEBRA 53 MATH 205B NOTES 2010 COMMUTATIVE ALGEBRA 53 10. Completion The real numbers are the completion of the rational numbers with respect to the usual absolute value norm. This means that any Cauchy sequence

More information

ON 2-CLASS FIELD TOWERS OF SOME IMAGINARY QUADRATIC NUMBER FIELDS

ON 2-CLASS FIELD TOWERS OF SOME IMAGINARY QUADRATIC NUMBER FIELDS ON 2-CLASS FIELD TOWERS OF SOME IMAGINARY QUADRATIC NUMBER FIELDS FRANZ LEMMERMEYER Abstract. We construct an infinite family of imaginary quadratic number fields with 2-class groups of type (2, 2, 2 whose

More information

School of Mathematics and Statistics. MT5836 Galois Theory. Handout 0: Course Information

School of Mathematics and Statistics. MT5836 Galois Theory. Handout 0: Course Information MRQ 2017 School of Mathematics and Statistics MT5836 Galois Theory Handout 0: Course Information Lecturer: Martyn Quick, Room 326. Prerequisite: MT3505 (or MT4517) Rings & Fields Lectures: Tutorials: Mon

More information

A connection between number theory and linear algebra

A connection between number theory and linear algebra A connection between number theory and linear algebra Mark Steinberger Contents 1. Some basics 1 2. Rational canonical form 2 3. Prime factorization in F[x] 4 4. Units and order 5 5. Finite fields 7 6.

More information

CYCLOTOMIC POLYNOMIALS

CYCLOTOMIC POLYNOMIALS CYCLOTOMIC POLYNOMIALS 1. The Derivative and Repeated Factors The usual definition of derivative in calculus involves the nonalgebraic notion of limit that requires a field such as R or C (or others) where

More information

MINIMAL GENERATING SETS OF GROUPS, RINGS, AND FIELDS

MINIMAL GENERATING SETS OF GROUPS, RINGS, AND FIELDS MINIMAL GENERATING SETS OF GROUPS, RINGS, AND FIELDS LORENZ HALBEISEN, MARTIN HAMILTON, AND PAVEL RŮŽIČKA Abstract. A subset X of a group (or a ring, or a field) is called generating, if the smallest subgroup

More information

Field Theory Qual Review

Field Theory Qual Review Field Theory Qual Review Robert Won Prof. Rogalski 1 (Some) qual problems ˆ (Fall 2007, 5) Let F be a field of characteristic p and f F [x] a polynomial f(x) = i f ix i. Give necessary and sufficient conditions

More information

Quadratic reciprocity (after Weil) 1. Standard set-up and Poisson summation

Quadratic reciprocity (after Weil) 1. Standard set-up and Poisson summation (December 19, 010 Quadratic reciprocity (after Weil Paul Garrett garrett@math.umn.edu http://www.math.umn.edu/ garrett/ I show that over global fields k (characteristic not the quadratic norm residue symbol

More information

Galois theory of fields

Galois theory of fields 1 Galois theory of fields This first chapter is both a concise introduction to Galois theory and a warmup for the more advanced theories to follow. We begin with a brisk but reasonably complete account

More information

18 The analytic class number formula

18 The analytic class number formula 18.785 Number theory I Lecture #18 Fall 2015 11/12/2015 18 The analytic class number formula The following theorem is usually attributed to Dirichlet, although he originally proved it only for quadratic

More information

On The Weights of Binary Irreducible Cyclic Codes

On The Weights of Binary Irreducible Cyclic Codes On The Weights of Binary Irreducible Cyclic Codes Yves Aubry and Philippe Langevin Université du Sud Toulon-Var, Laboratoire GRIM F-83270 La Garde, France, {langevin,yaubry}@univ-tln.fr, WWW home page:

More information

Local Class Field Theory I

Local Class Field Theory I CHAPTER 4 Local Class Field Theory I In this chapter we develop the theory of abelian extensions of a local field with finite residue field. The main theorem establishes a correspondence between abelian

More information

Graduate Preliminary Examination

Graduate Preliminary Examination Graduate Preliminary Examination Algebra II 18.2.2005: 3 hours Problem 1. Prove or give a counter-example to the following statement: If M/L and L/K are algebraic extensions of fields, then M/K is algebraic.

More information

COMMUTING ELEMENTS IN GALOIS GROUPS OF FUNCTION FIELDS. Fedor Bogomolov and Yuri Tschinkel

COMMUTING ELEMENTS IN GALOIS GROUPS OF FUNCTION FIELDS. Fedor Bogomolov and Yuri Tschinkel COMMUTING ELEMENTS IN GALOIS GROUPS OF FUNCTION FIELDS by Fedor Bogomolov and Yuri Tschinkel Abstract. We study the structure of abelian subgroups of Galois groups of function fields. Contents Introduction................................................

More information

Sets of Completely Decomposed Primes in Extensions of Number Fields

Sets of Completely Decomposed Primes in Extensions of Number Fields Sets of Completely Decomposed Primes in Extensions of Number Fields by Kay Wingberg June 15, 2014 Let p be a prime number and let k(p) be the maximal p-extension of a number field k. If T is a set of primes

More information

Subrings of Finite Commutative Rings

Subrings of Finite Commutative Rings Subrings of Finite Commutative Rings Francisco Franco Munoz arxiv:1712.02025v1 [math.ac] 6 Dec 2017 Abstract We record some basic results on subrings of finite commutative rings. Among them we establish

More information

Up to twist, there are only finitely many potentially p-ordinary abelian varieties over. conductor

Up to twist, there are only finitely many potentially p-ordinary abelian varieties over. conductor Up to twist, there are only finitely many potentially p-ordinary abelian varieties over Q of GL(2)-type with fixed prime-to-p conductor Haruzo Hida Department of Mathematics, UCLA, Los Angeles, CA 90095-1555,

More information

KUMMER THEORY OF ABELIAN VARIETIES AND REDUCTIONS OF MORDELL-WEIL GROUPS

KUMMER THEORY OF ABELIAN VARIETIES AND REDUCTIONS OF MORDELL-WEIL GROUPS KUMMER THEORY OF ABELIAN VARIETIES AND REDUCTIONS OF MORDELL-WEIL GROUPS TOM WESTON Abstract. Let A be an abelian variety over a number field F with End F A commutative. Let Σ be a subgroup of AF ) and

More information

SOME AMAZING PROPERTIES OF THE FUNCTION f(x) = x 2 * David M. Goldschmidt University of California, Berkeley U.S.A.

SOME AMAZING PROPERTIES OF THE FUNCTION f(x) = x 2 * David M. Goldschmidt University of California, Berkeley U.S.A. SOME AMAZING PROPERTIES OF THE FUNCTION f(x) = x 2 * David M. Goldschmidt University of California, Berkeley U.S.A. 1. Introduction Today we are going to have a look at one of the simplest functions in

More information

Hamburger Beiträge zur Mathematik

Hamburger Beiträge zur Mathematik Hamburger Beiträge zur Mathematik Nr. 270 / April 2007 Ernst Kleinert On the Restriction and Corestriction of Algebras over Number Fields On the Restriction and Corestriction of Algebras over Number Fields

More information

2 More on Congruences

2 More on Congruences 2 More on Congruences 2.1 Fermat s Theorem and Euler s Theorem definition 2.1 Let m be a positive integer. A set S = {x 0,x 1,,x m 1 x i Z} is called a complete residue system if x i x j (mod m) whenever

More information

D-MATH Algebra II FS18 Prof. Marc Burger. Solution 26. Cyclotomic extensions.

D-MATH Algebra II FS18 Prof. Marc Burger. Solution 26. Cyclotomic extensions. D-MAH Algebra II FS18 Prof. Marc Burger Solution 26 Cyclotomic extensions. In the following, ϕ : Z 1 Z 0 is the Euler function ϕ(n = card ((Z/nZ. For each integer n 1, we consider the n-th cyclotomic polynomial

More information

Mod p Galois representations of solvable image. Hyunsuk Moon and Yuichiro Taguchi

Mod p Galois representations of solvable image. Hyunsuk Moon and Yuichiro Taguchi Mod p Galois representations of solvable image Hyunsuk Moon and Yuichiro Taguchi Abstract. It is proved that, for a number field K and a prime number p, there exist only finitely many isomorphism classes

More information

Class groups and Galois representations

Class groups and Galois representations and Galois representations UC Berkeley ENS February 15, 2008 For the J. Herbrand centennaire, I will revisit a subject that I studied when I first came to Paris as a mathematician, in 1975 1976. At the

More information

RUDIMENTARY GALOIS THEORY

RUDIMENTARY GALOIS THEORY RUDIMENTARY GALOIS THEORY JACK LIANG Abstract. This paper introduces basic Galois Theory, primarily over fields with characteristic 0, beginning with polynomials and fields and ultimately relating the

More information

Local Fields. Chapter Absolute Values and Discrete Valuations Definitions and Comments

Local Fields. Chapter Absolute Values and Discrete Valuations Definitions and Comments Chapter 9 Local Fields The definition of global field varies in the literature, but all definitions include our primary source of examples, number fields. The other fields that are of interest in algebraic

More information

Spanning and Independence Properties of Finite Frames

Spanning and Independence Properties of Finite Frames Chapter 1 Spanning and Independence Properties of Finite Frames Peter G. Casazza and Darrin Speegle Abstract The fundamental notion of frame theory is redundancy. It is this property which makes frames

More information

The Kummer Pairing. Alexander J. Barrios Purdue University. 12 September 2013

The Kummer Pairing. Alexander J. Barrios Purdue University. 12 September 2013 The Kummer Pairing Alexander J. Barrios Purdue University 12 September 2013 Preliminaries Theorem 1 (Artin. Let ψ 1, ψ 2,..., ψ n be distinct group homomorphisms from a group G into K, where K is a field.

More information

FERMAT S LAST THEOREM FOR REGULAR PRIMES KEITH CONRAD

FERMAT S LAST THEOREM FOR REGULAR PRIMES KEITH CONRAD FERMAT S LAST THEOREM FOR REGULAR PRIMES KEITH CONRAD For a prime p, we call p regular when the class number h p = h(q(ζ p )) of the pth cyclotomic field is not divisible by p. For instance, all primes

More information

55 Separable Extensions

55 Separable Extensions 55 Separable Extensions In 54, we established the foundations of Galois theory, but we have no handy criterion for determining whether a given field extension is Galois or not. Even in the quite simple

More information

Polynomials with nontrivial relations between their roots

Polynomials with nontrivial relations between their roots ACTA ARITHMETICA LXXXII.3 (1997) Polynomials with nontrivial relations between their roots by John D. Dixon (Ottawa, Ont.) 1. Introduction. Consider an irreducible polynomial f(x) over a field K. We are

More information

Groups of Prime Power Order with Derived Subgroup of Prime Order

Groups of Prime Power Order with Derived Subgroup of Prime Order Journal of Algebra 219, 625 657 (1999) Article ID jabr.1998.7909, available online at http://www.idealibrary.com on Groups of Prime Power Order with Derived Subgroup of Prime Order Simon R. Blackburn*

More information

Infinite Galois theory

Infinite Galois theory Treball final de grau GRAU DE MATEMÀTIQUES Facultat de Matemàtiques i Informàtica Universitat de Barcelona Infinite Galois theory Autor: Ignasi Sánchez Rodríguez Director: Realitzat a: Dra. Teresa Crespo

More information

SPEAKER: JOHN BERGDALL

SPEAKER: JOHN BERGDALL November 24, 2014 HODGE TATE AND DE RHAM REPRESENTATIONS SPEAKER: JOHN BERGDALL My goal today is to just go over some results regarding Hodge-Tate and de Rham representations. We always let K/Q p be a

More information

Fields and Galois Theory. Below are some results dealing with fields, up to and including the fundamental theorem of Galois theory.

Fields and Galois Theory. Below are some results dealing with fields, up to and including the fundamental theorem of Galois theory. Fields and Galois Theory Below are some results dealing with fields, up to and including the fundamental theorem of Galois theory. This should be a reasonably logical ordering, so that a result here should

More information

NONABELIAN GROUPS WITH PERFECT ORDER SUBSETS

NONABELIAN GROUPS WITH PERFECT ORDER SUBSETS NONABELIAN GROUPS WITH PERFECT ORDER SUBSETS CARRIE E. FINCH AND LENNY JONES Abstract. Let G be a finite group and let x G. Define the order subset of G determined by x to be the set of all elements in

More information

RIEMANN S INEQUALITY AND RIEMANN-ROCH

RIEMANN S INEQUALITY AND RIEMANN-ROCH RIEMANN S INEQUALITY AND RIEMANN-ROCH DONU ARAPURA Fix a compact connected Riemann surface X of genus g. Riemann s inequality gives a sufficient condition to construct meromorphic functions with prescribed

More information

KUMMER S CRITERION ON CLASS NUMBERS OF CYCLOTOMIC FIELDS

KUMMER S CRITERION ON CLASS NUMBERS OF CYCLOTOMIC FIELDS KUMMER S CRITERION ON CLASS NUMBERS OF CYCLOTOMIC FIELDS SEAN KELLY Abstract. Kummer s criterion is that p divides the class number of Q(µ p) if and only if it divides the numerator of some Bernoulli number

More information