seminormal weakly Krull monoids and domains
A. Geroldinger∗ and F. Kainrath and A. Reinhart
International Meeting on Numerical Semigroups
Cortona, September 2014
Outline
(Unions of) Sets of Lengths Krull and weakly Krull monoids
Arithmetic: Precise versus Qualitative Results Seminormal Monoids and Domains
Main Results
Methods: Transfer Homomorphisms
Sets of lengths in monoids
Let H be a multiplicatively written, commutative, cancellative semigroup, and let a ∈ H be a non-unit.
• If a = u1· . . . ·uk where u1, . . . ,uk are irreducibles (atoms), then k is called thelengthof the factorization.
• LH(a) = {k | a has a factorization of length k} ⊂ N is the set of lengthsof a.
• If L(a) = {k1,k2,k3, . . .}with k1<k2 <k3 < . . ., then
∆ L(a) = {k2−k1,k3−k2, . . .}
is the set of distancesof L(a).
• If |L(a)| ≥ 2, then |L(am)| >m for each m ∈ N.
Sets of distances and unions of sets of lengths
We call
∆(H) = [
a∈H
∆ L(a) ⊂ N theset of distances of H. For k ∈ N, we call
Uk(H) = [
k∈L(a)
L(a)
= {` ∈ N | there is an equation u1· . . . ·uk =v1· . . . ·v`} theunion of sets of lengths containing k.
An atomic monoid H is calledhalf-factorialif one foll. equiv. holds:
(a) |L(a)| = 1 for each a ∈ H.
(b) ∆(H) = ∅.
(c) Uk(H) = {k} for each k ∈ N.
Outline
(Unions of) Sets of Lengths Krull and weakly Krull monoids
Arithmetic: Precise versus Qualitative Results Seminormal Monoids and Domains
Main Results
Methods: Transfer Homomorphisms
Denition of Krull monoids
H is called aKrull monoidif one of the foll. equiv. holds : (a) H is v-noetherian and completely integrally closed.
(b) H has a divisor theory ϕ: H → F(P) = F :
• ϕis a divisor homomorphism:
For all a, b ∈ H we have a | b if and only if ϕ(a) | ϕ(b) .
• For all p ∈ P there is a set X ⊂ H such that p = gcd(ϕ(X ))).
(c) There is a divisor homomorphism into any free abelian monoid.
Thedivisor class groupG is isomorphic to thev-class group:
G = q(F )/q ϕ(H) = {aq ϕ(H) = [a] | a ∈ F } ∼= Cv(H) . Let R be a domain.
• R is a Krull domain i • is a Krull monoid.
• Integrally closed noetherian domains are Krull by Property (a).
Primary monoids and domains
1. An element q ∈ H is called primary if q /∈ H× and, for all a, b ∈ H, if q | ab and q - a, then q | bn for some n ∈ N.
2. H is called primary if m = H \ H×6= ∅ and one of the following equivalent statements are satised :
(a) s-spec(H) = {∅, H \ H×}. (b) Every q ∈ m is primary.
(c) For all a, b ∈ m there exists some n ∈ N such that a | bn. 3. Let R be a domain.
Then R• is primary i R is one-dimensional and local.
Finitely primary monoids and domains
A monoid H is callednitely primary (of rank s and exponent α) if one of the following equivalent conditions holds:
(a) There exist s, α ∈ N with the following properties :
H is a submonoid of a factorial monoid F = F××[p1, . . . ,ps] with s pairwise non-associated prime elements p1, . . . ,ps s.t.
H \ H×⊂p1· . . . ·psF and (p1· . . . ·ps)αF ⊂ H . (b) H is primary, (H : bH) 6= ∅ and bHred∼= (Ns0, +).
Clearly, numerical monoids are nitely primary of rank 1.
Let R be a domain.
• If R is a one-dimensional local Mori domain such that (R : bR) 6= {0}, then R• is nitely primary.
Weakly Krull monoids and domains
A monoid H isweakly Krull if
H = \
p∈X(H)
Hp and {p ∈ X(H) | a ∈ p} is nite for all a ∈ H ,
Weakly Krull domains: Anderson, Mott, and Zafrullah, 1992 Weakly Krull monoids: Halter-Koch, Boll. UMI 1995
• A domain R is weakly Krull i R• is a weakly Krull monoid.
• H v-noetherian: H weakly Krull ⇐⇒ v-max(H) = X(H).
• H Krull ⇒ H seminormal v-noetherian weakly Krull a. H = bH.
• We suppose that all weakly Krull monoids are
• v-noetherian
• Hp are nitely primary for each p ∈ X(H).
• (H : bH) = f 6= ∅.
• Example: 1-dim. noeth. domains R s.t. R is a f.g. R-module
Outline
(Unions of) Sets of Lengths Krull and weakly Krull monoids
Arithmetic: Precise versus Qualitative Results Seminormal Monoids and Domains
Main Results
Methods: Transfer Homomorphisms
Arithmetic of Krull monoids: Precise Results
Let H be a Krull monoid with class group G such that each class contains a prime divisor.
1. (Carlitz 1960) H is half-factorial if and only if |G| ≤ 2.
2. Let 2 < |G| < ∞. Then
• ∆(H) is a nite interval with min ∆(H) = 1.
• All Uk(H) are nite intervals.
• .... and much more .... for example ....
• If G is cyclic of order n, then ∆(H) = [1, n − 2], max U2k(H) = kn, and max U2k+1(H) = kn + 1.
3. If G is innite, then ∆(H) = Uk(H) = N≥2 for all k ∈ N.
Arithmetic of weakly Krull monoids: Qualitative Results
Let R be a non-principal order in an algebraic number elds with Picard group G.
• Apart from quadratic number elds (Halter-Koch 1983), there is no characterization of half-factoriality.
• ∆(R) is nite. If |G| ≤ 2, then it is open whether 1 ∈ ∆(R).
• For each k ∈ N≥2 the following are equivalent:
• Uk(R) is nite.
• The natural map X(bR) → X(R) is bijective.
• There is no information
• on the structure of the set of distances ∆(R)
• nor on the structure of the unions Uk(R).
Outline
(Unions of) Sets of Lengths Krull and weakly Krull monoids
Arithmetic: Precise versus Qualitative Results Seminormal Monoids and Domains
Main Results
Methods: Transfer Homomorphisms
Seminormality: Denitions and Remarks
TheseminormalizationH0 of H is dened by
H0 = {x ∈ q(H) | there is some N ∈ N such that xn∈H for all n ≥ N}
Then
• H ⊂ H0⊂ bH ⊂ q(H).
• H is seminormal if H = H0. Equivalently, if x ∈ q(H) and x2,x3 ∈H, then x ∈ H.
A domain R isseminormal if one of the foll. equiv. holds:
(a) R• is seminormal.
(b) Pic(R) → Pic R[X ] is an isomorphism.
Traverso (1970), Swan (1980); Survey by Vitulli (2010)
Seminormal nitely primary monoids
Let H ⊂ bH = F = F××[p1, . . . ,ps]be nitely primary.
• H0 =p1· . . . ·psF ∪ H0×.
• If F×= {1}, then H0 ∼= (Ns∪ {0}, +) ⊂ (Ns0, +).
• A(H0) =εp1k1 · . . . ·psks | ε ∈F×,min{k1, . . . ,ks} =1 .
• H0 is seminormal, v-noetherian, and
nitely primary of rank s and exponent 1.
For a domain R the following statements are equivalent : (a) R is a seminormal one-dimensional local Mori domain.
(b) R• is seminormal nitely primary.
Outline
(Unions of) Sets of Lengths Krull and weakly Krull monoids
Arithmetic: Precise versus Qualitative Results Seminormal Monoids and Domains
Main Results
Methods: Transfer Homomorphisms
Algebraic Structure of seminormal weakly Krull monoids
Let H be a seminormal weakly Krull monoid with nontrivial conductor f = (H : bH) ( H, and let P∗= {p ∈ X(H) | p ⊃ f}.
Then we have
1. H is Krull and Pb ∗ is nite.
2. The monoid Iv∗(H) of v-invertible v-ideals satises Iv∗(H) ∼= F (P) × Y
p∈P∗
(Hp)red,
and it is seminormal, v-noetherian, and weakly factorial, 3. There is an exact sequence
1 → bH×/H×→ a
p∈X(H)
Hbp×/Hp×
→ Cε v(H)→ Cϑ v( bH) → 0 .
Arithmetic Structure
Suppose in addition that G = Cv(H) is nite, and that every class contains a p ∈ X(H) with p 6⊃ f.
1. Suppose the natural map X(bH) → X(H) is bijective.
1.1 Uk(H) is a nite interval for all k ≥ 2.
1.2 Suppose that ϑ: Cv(H) → Cv( bH) is an isomorphism.
Then there is a transfer homomorphism θ : H → B(G).
In particular, (unions of) sets of lengths and (monotone) catenary degrees of H and B(G) coincide.
2. Suppose the natural map X(bH) → X(H) is not bijective.
Then for all k ≥ 3, we have
N≥3 ⊂ Uk(H) ⊂ N≥2.
Characterization of Half-Factoriality
Suppose in addition that the class group G = C(H) is nite, and that every class contains a p ∈ X(H) with p 6⊃ f.
Then the following statements are equivalent : (a) c(H) ≤ 2.
(b) H is half-factorial.
(c) |G| ≤ 2, the natural map X(bH) → X(H) is bijective, and the homomorphism ϑ: Cv(H) → Cv( bH) is an isomorphism.
where
π : X( bH) → X(H), is dened by π(P) = P ∩ H for all P ∈ X(bH)
Outline
(Unions of) Sets of Lengths Krull and weakly Krull monoids
Arithmetic: Precise versus Qualitative Results Seminormal Monoids and Domains
Main Results
Methods: Transfer Homomorphisms
Transfer Homomorphisms
Consider
H −−−−→ D = F(P)×T ∼= Iv∗(H)
β
y βe
y B = B(G, T , ι) −−−−→ F = F(G)×T where
• H ,→ D is saturated, and the class group G = C(H, D) satises G = {[p] | p ∈ P} ⊂ G.
• ι : T → G is dened by ι(t) = [t].
• β :e D → F be the unique homomorphism satisfying eβ(p) = [p]
for all p ∈ P and eβ | T = idT.
1. The restriction β = eβ | H : H → B is a transfer hom.
2. Transfer homomorphisms preserve sets of lengths. In particular, unions of sets of lengths and half-factoriality.
Combinatorial weakly Krull monoids: B(G, T , ι)
Let G be a nite abelian group and T = D1× . . . ×Dn a monoid.
Let
• ι : T → G a homomorphism, and
• σ : F (G) → G satisfying σ(g) = g.
Then
B(G, T , ι) = {S t ∈ F(G)×T | σ(S) + ι(t) = 0 } ⊂ F(G)×T the T -block monoid over G dened by ι.
Special Cases:
• If G = {0}, then B(G, T , ι) = T = D1× . . . ×Dn is a nite product of nitely primary monoids.
• If T = {1}, then
B(G, T , ι) = B(G) = {S ∈ F(G) | σ(S) = 0} ⊂ F(G) is the monoid of zero-sum sequences over G.
Saturated submonoids inherit the properties under consideration
Consider a saturated submonoid
H ⊂ D = F(P)×Yn
i=1
Di,
where P ⊂ D is a set of primes, n ∈ N0, and D1, . . . ,Dn are primary monoids. Then we have .
1. If C(H, D) is a torsion group, then H is a weakly Krull monoid.
2. If D1, ...,Dn are seminormal nitely primary, then H is seminormal and v-noetherian with (H : bH) 6= ∅.