Let O⊆OK\mathcal O\subseteq\mathcal O_K be an in a KK, with OK\mathcal O_K. Its conductor is

f={x∈OK:xOK⊆O}.\mathfrak f=\{x\in\mathcal O_K:x\mathcal O_K\subseteq\mathcal O\}.

It is the largest OK\mathcal O_K- contained in O\mathcal O, and is also an ideal of O\mathcal O. Since 1∈OK1\in\mathcal O_K, every element of f\mathfrak f already belongs to O\mathcal O.

Why it is nonzero

The additive quotient OK/O\mathcal O_K/\mathcal O is finite. If its order is mm, then mOK⊆fm\mathcal O_K\subseteq\mathfrak f. Thus the conductor records a finite obstruction to the order being maximal; it equals OK\mathcal O_K exactly when O=OK\mathcal O=\mathcal O_K.

Quadratic orders

If KK is quadratic and O=Z+fOK\mathcal O=\mathbb Z+f\mathcal O_K with f≥1f\ge1, then f=fOK\mathfrak f=f\mathcal O_K. The integer ff is also called the conductor, whereas f\mathfrak f is the conductor ideal.

For K=Q(−3)K=\mathbb Q(\sqrt{-3}) and O=Z[−3]\mathcal O=\mathbb Z[\sqrt{-3}], this gives

f=2OK=(2,1+−3).\mathfrak f=2\mathcal O_K=(2,1+\sqrt{-3}).

It is the noninvertible ideal from the .

Away from the conductor

For an integral ideal I⊆OI\subseteq\mathcal O, being prime to the conductor means I+f=OI+\mathfrak f=\mathcal O. Such nonzero ideals are invertible. Localization outside the primes containing f\mathfrak f removes the difference between O\mathcal O and OK\mathcal O_K.

References
  1. Andrew Sutherland, MIT 18.785 Lecture 7, §7.1, Lemma 7.1, Theorem 7.8, and Corollary 7.10.