TheoremComplete

Dedekind Domains and Factorization - Main Theorem

The fundamental theorem of Dedekind domains establishes unique factorization at the ideal level, restoring structure lost at the element level.

TheoremUnique Factorization of Ideals

Let RR be a Dedekind domain. Every nonzero ideal II of RR can be written uniquely as a product of prime ideals: I=p1e1prerI = \mathfrak{p}_1^{e_1} \cdots \mathfrak{p}_r^{e_r} where pi\mathfrak{p}_i are distinct prime ideals and ei1e_i \geq 1. The factorization is unique up to reordering of factors.

Moreover, every nonzero fractional ideal has a unique factorization I=p1e1prerI = \mathfrak{p}_1^{e_1} \cdots \mathfrak{p}_r^{e_r} where now eiZe_i \in \mathbb{Z} (possibly negative).

This theorem is the cornerstone of algebraic number theory. The proof relies on showing that:

  1. Every nonzero prime ideal is maximal
  2. Every ideal can be written as a product of prime ideals
  3. Cancellation law holds: if IJ=IKIJ = IK with I0I \neq 0, then J=KJ = K
TheoremCharacterization of Dedekind Domains

For an integral domain RR, the following are equivalent:

  1. RR is a Dedekind domain
  2. Every nonzero fractional ideal of RR is invertible
  3. RR is Noetherian, integrally closed, and every nonzero prime ideal is maximal
  4. RR is Noetherian and the localization RpR_\mathfrak{p} at each prime p\mathfrak{p} is a discrete valuation ring (DVR)
ExamplePrime Ideal Structure

In OQ(5)=Z[1+52]\mathcal{O}_{\mathbb{Q}(\sqrt{-5})} = \mathbb{Z}\left[\frac{1+\sqrt{-5}}{2}\right], the prime ideals lying over (2)(2) are: p2=(2,1+5),pˉ2=(2,15)\mathfrak{p}_2 = (2, 1+\sqrt{-5}), \quad \bar{\mathfrak{p}}_2 = (2, 1-\sqrt{-5})

We can verify (2)=p22=pˉ22(2) = \mathfrak{p}_2^2 = \bar{\mathfrak{p}}_2^2 (these are the same ideal). The ramification index is e=2e = 2, and inertia degree is f=1f = 1, satisfying ef=2=[K:Q]ef = 2 = [K : \mathbb{Q}].

TheoremFundamental Exact Sequence

The multiplicative structure of fractional ideals gives an exact sequence: 1OKKFrac(OK)Cl(K)01 \to \mathcal{O}_K^* \to K^* \to \text{Frac}(\mathcal{O}_K) \to \text{Cl}(K) \to 0

where Frac(OK)\text{Frac}(\mathcal{O}_K) is the group of fractional ideals and Cl(K)\text{Cl}(K) is the ideal class group. This sequence encodes:

  • Elements vs. principal ideals: KFrac(OK)K^* \to \text{Frac}(\mathcal{O}_K) sends α(α)\alpha \mapsto (\alpha)
  • Class group as quotient: Cl(K)=Frac(OK)/K\text{Cl}(K) = \text{Frac}(\mathcal{O}_K) / K^*
Remark

The class number hK=Cl(K)h_K = |\text{Cl}(K)| measures obstruction to unique factorization. When hK=1h_K = 1, every ideal is principal and OK\mathcal{O}_K is a PID. The Minkowski bound provides computational method for determining hKh_K by finding representatives of each ideal class.

TheoremKummer-Dedekind Theorem

Let K=Q(α)K = \mathbb{Q}(\alpha) where α\alpha satisfies monic f(x)Z[x]f(x) \in \mathbb{Z}[x]. For prime pp not dividing [OK:Z[α]][\mathcal{O}_K : \mathbb{Z}[\alpha]], if f(x)i=1gfˉi(x)ei(modp)f(x) \equiv \prod_{i=1}^g \bar{f}_i(x)^{e_i} \pmod{p} where fˉi\bar{f}_i are distinct irreducibles in Fp[x]\mathbb{F}_p[x], then (p)=i=1gpiei,pi=(p,fi(α))(p) = \prod_{i=1}^g \mathfrak{p}_i^{e_i}, \quad \mathfrak{p}_i = (p, f_i(\alpha))

where fif_i is a lift of fˉi\bar{f}_i to Z[x]\mathbb{Z}[x], and the inertia degree of pi\mathfrak{p}_i is deg(fi)\deg(f_i).

This computational tool allows explicit factorization of primes by factoring polynomials modulo pp.