[guided]We want to show two things for the given nonzero ideal $\mathfrak{a} \unlhd \mathcal{O}_L$: first, that $\mathfrak{a}$ contains a nonzero rational integer, and second, that $\mathcal{O}_L/\mathfrak{a}$ is finite. This step handles the first — and the integer we produce is precisely what will make the quotient finite in the next step.
**Strategy.** An element $\alpha \in \mathcal{O}_L$ is, by definition, a root of a monic polynomial with $\mathbb{Z}$-coefficients. If that polynomial has a nonzero constant term $a_0$, then the relation $p_\alpha(\alpha) = 0$ can be rearranged to express $a_0$ as a multiple of $\alpha$; and multiples of $\alpha$ lie in $\mathfrak{a}$. The key subtlety is that we need $a_0 \neq 0$, which is why we use the *minimal* polynomial.
**Picking $\alpha$.** Since $\mathfrak{a}$ is nonzero, choose any $\alpha \in \mathfrak{a} \setminus \{0\}$.
**The minimal polynomial has integer coefficients.** By the [Integrality Criterion via Minimal Polynomial](/theorems/1570): $\alpha \in \mathcal{O}_L$ if and only if the minimal polynomial $p_\alpha \in \mathbb{Q}[x]$ of $\alpha$ over $\mathbb{Q}$ has all coefficients in $\mathbb{Z}$. Since $\alpha \in \mathfrak{a} \subseteq \mathcal{O}_L$, Theorem 1570 applies and
\begin{align*}
p_\alpha(x) &= x^m + a_{m-1}x^{m-1} + \cdots + a_1 x + a_0
\end{align*}
with $a_0, \ldots, a_{m-1} \in \mathbb{Z}$.
**Why $a_0 \neq 0$.** If $a_0$ were zero, we could factor $p_\alpha(x) = x \cdot q(x)$ with $q \in \mathbb{Q}[x]$ nonconstant and of positive degree. A minimal polynomial is irreducible over its base field (otherwise, one of its factors would annihilate $\alpha$ with smaller degree, contradicting minimality). But $p_\alpha = x \cdot q$ is reducible over $\mathbb{Q}$ when $\deg q \geq 1$. The only escape is $\deg p_\alpha = 1$, giving $p_\alpha = x$, which would mean $\alpha = 0$ — contrary to our choice. Hence $a_0 \neq 0$.
**Extracting $a_0$ from the annihilation equation.** Evaluating $p_\alpha$ at $\alpha$:
\begin{align*}
\alpha^m + a_{m-1}\alpha^{m-1} + \cdots + a_1 \alpha + a_0 = 0.
\end{align*}
We solve for $a_0$, bringing all other terms to the other side, and factor an $\alpha$ out of each remaining term:
\begin{align*}
a_0 &= -\alpha^m - a_{m-1}\alpha^{m-1} - \cdots - a_1 \alpha \\
&= -\alpha \cdot \big(\alpha^{m-1} + a_{m-1}\alpha^{m-2} + \cdots + a_2 \alpha + a_1\big).
\end{align*}
**Why the bracketed factor lies in $\mathcal{O}_L$.** By [Ring of Integers is a Ring](/theorems/1567), $\mathcal{O}_L$ is a commutative ring containing $\mathbb{Z}$. Since $\alpha \in \mathcal{O}_L$ and $a_i \in \mathbb{Z} \subseteq \mathcal{O}_L$, the expression $\alpha^{m-1} + a_{m-1}\alpha^{m-2} + \cdots + a_1$ is a polynomial in $\alpha$ with $\mathcal{O}_L$-coefficients, hence lies in $\mathcal{O}_L$.
**Why $a_0 \in \mathfrak{a}$.** The displayed equation says $a_0 = \alpha \cdot \beta$ for some $\beta \in \mathcal{O}_L$. Since $\mathfrak{a}$ is an ideal and $\alpha \in \mathfrak{a}$, we have $\alpha \cdot \beta \in \mathfrak{a}$, i.e., $a_0 \in \mathfrak{a}$.
**Assembling the conclusion of this step.** $a_0 \in \mathfrak{a}$ and $a_0 \in \mathbb{Z}$ (coefficients of $p_\alpha$ are rational integers), hence $a_0 \in \mathfrak{a} \cap \mathbb{Z}$. And $a_0 \neq 0$ as shown. So $\mathfrak{a} \cap \mathbb{Z}$ is a nonzero ideal of $\mathbb{Z}$; in particular, $\mathfrak{a} \cap \mathbb{Z} \neq \{0\}$.[/guided]