1

My problem:

Let $n$ be a natural number and $p$ be a prime. Prove that there exists an irreducible polynomial of degree $n$ in $\mathbb{Z}_p[x]$ (note: $Z_p$ is the quotient field $\mathbb{Z}/p\mathbb{Z}$)

Basically I don't have a clear idea to even think about. But somehow I think this relates to the splitting field of $f(x) = x^{p^n} - x$ in $Z_p[x]$, which serves as the extension of $Z_p$ to $p^n$ elements. Is it possible that there exists a factor of this polynomial which has degree $n$?

Please give me a hint. Anything is greatly appreciated.

I have basic knowledge about field extensions, spliting fiels and finite fields. Thank you.

ElementX
  • 922
  • 1
    This gotta be one of the most frequently asked questions about finite fields on this site. I'm a bit disappointed that veteran users would not search first. – Jyrki Lahtonen Dec 25 '18 at 08:22
  • 1
    If you are also interested in counting how many such polynomials exist, look at this. – Jyrki Lahtonen Dec 25 '18 at 08:26
  • 1
    This thread covers both the existence and the count, but is a bit more taxing to follow. Honestly, I don't know the earliest thread handling this question. Even back in 2012 some prolific answerers would pay too little attention to site hygiene in their eagerness to answer. But, such things happen to most of us. Back then duplication wasn't as severe a problem as it is today. – Jyrki Lahtonen Dec 25 '18 at 08:29
  • In my case... not a veteran user here; I only joined a month ago. On the other hand, I have posted this count before on AoPS, in 2005 and again in 2018 – jmerry Dec 25 '18 at 21:07

2 Answers2

4

This is all standard theory. The main steps are:

  1. Show that the splitting field $K$ of $x^{p^n}-x$ has order $p^n$,
  2. Show that $K$'s multiplicative group is cyclic, so has a generator $\alpha$,
  3. Observe that $K=\Bbb Z_p(\alpha)$,
  4. Show the minimal polynomial of $\alpha$ over $\Bbb Z_p$ has degree $n$.
Angina Seng
  • 158,341
3

OK, now that the standard theory's up, I'll add my favorite argument. We prove that irreducible polynomials of every degree exist by counting them.

Let $F$ be a field with $q$ elements, for some finite $q$. We count the number of monic polynomials of degree $n$ over $F$ in two ways.
First, a monic polynomial $x^n + a_{n-1}x^{n-1}+\cdots + a_1 x + a_0$ has $n$ coefficients that can freely vary among $q$ values each, for a total of $q^n$ polynomials.
Second, a monic polynomial can be uniquely factored as a product of monic irreducible polynomials. We choose the power each polynomial is raised to, make sure the sum of the degrees is $n$ - it's an intimidating counting problem, but we can make it easier by thinking in terms of generating functions. Let $P_k$ be the number of monic irreducible polynomials of degree $k$, and $M_n$ the total number of monic polynomials of degree $n$. This model gives us that $$\sum_{n=0}^{\infty} M_n t^n = \prod_{k=1}^{\infty}\frac1{(1-t^k)^{P_k}}$$ (Each $\frac1{1-t^k}=1+t^k+t^{2k}+\cdots$ factor represents the choice of how many copies of a particular irreducible polynomial of degree $k$ to include)
So, what now? First, we use that earlier count to simplify the left side: $$\sum_{n=0}^{\infty} M_n t^n = \sum_{n=0}^{\infty} q^n t^n = \frac1{1-qt}$$ Next, take logarithms* and differentiate with respect to $t$: \begin{align*}\frac1{1-qt} &= \prod_{k=1}^{\infty}\frac1{(1-t^k)^{P_k}}\\ -\log(1-qt) &= \sum_{k=1}^{\infty} -P_k\log(1-t^k)\\ \frac{q}{1-qt} &= \sum_{k=1}^{\infty} \frac{kP_kt^{k-1}}{1-t^k}\\ \sum_{n=1}^{\infty} q^nt^n = \frac{qt}{1-qt} &= \sum_{k=1}^{\infty} \frac{kP_kt^k}{1-t^k} = \sum_{n=1}^{\infty}t^n\sum_{k|n}kP_k\\ q^n &= \sum_{d|n}dP_d\end{align*} Now, we can extract the values of $P_n$ by a standard technique; Mobius inversion. We get $$nP_n = \sum_{d|n} \mu(d)q^{\frac{n}{d}}$$ where $\mu$ is the Mobius function of number theory.

Now, what we really wanted to know is that $P_n$ is positive for $n\ge 1$. For this, one key fact: $\mu(d)\in \{-1,0,1\}$ for all natural $d$ and $\mu(1)=1$, so \begin{align*}nP_n &= q^n +\sum_{k|n, 0<k<n}\mu(\frac{n}{k})q^k\\ nP_n &\ge q^n - \sum_{k|n, 0<k<n} q^k \ge q^n - \sum_{0<k<n}q^k = q^n - \frac{q^n-q}{q-1} > 0\end{align*} We use that $q\ge 2$ here; we do need at least two elements to have a field, after all. And there it is: any finite field with $q$ elements has irreducible polynomials for every degree. This includes both the prime fields $\mathbb{Z}/p$ and fields with prime power order. That such fields exist to have irreducible polynomials over - well, we'll just have to apply this result over the prime fields first.

*Taking a logarithm? Technically, it's all still formal power series manipulation, that doesn't rely on any form of convergence - but if you feel uncomfortable, we're just taking a derivative and then dividing by the original.

jmerry
  • 19,403
  • What do you mean by the line Each $\frac1{1-t^k}=1+t^k+t^{2k}+\cdots $ factor represents the choice of how many copies of a particular irreducible polynomial of degree k to include – Nothing Nov 23 '19 at 12:28
  • @Nothing ...to include... in the factorization of the monic polynomial into irreducible factors. – Sungjin Kim Nov 02 '23 at 10:45