# $\mathbb{Z}^*_p$ is a group under multiplication

Show that the set $$\mathbb{Z}^*_p=\{1,2, \dots, p-1\}$$ where $$p \in \mathbb{N}$$ is prime, is a group under multiplication.

Attempt:

Associativity and identity:

Obviously, multiplication $$\bmod p$$ is associative and $$1 \in \mathbb{Z}^*_p$$ (identity element).

Inverses:

Let $$k \in \mathbb{Z}^*_p$$. Since $$k, $$\gcd(k,p)=1$$ and therefore Bezout's Lemma ensures that there exist $$a,b \in \mathbb{N}$$ s.t. $$ak+bp=1\iff ak+bp \equiv 1 \space\bmod p \iff ak \equiv1 \bmod p$$

However, is the fact $$a guaranteed, so that $$a=k^{-1}\in \mathbb{Z}^*_p?$$

Is it because the congruence class $$\bar{a}$$ with $$a contains every $$a_i$$ you choose s.t. $$a_ik \equiv 1 \bmod p$$?

• I think you're mixing up integers and classes of integers modulo $p$. – Junkyards Jan 30 at 15:40
• If that $a\in \Bbb Z$ is not in the list of the explicitly declared representatives, choose the one representative which is $a$ modulo $p$. – dan_fulea Jan 30 at 15:41
• @Junkyards So, it would make more sense to work with classes: since there exists $a \in \mathbb{N}$ s.t. $$a \bar{k} \equiv 1 \bmod p$$ the inverses exist? – LoneBone Jan 30 at 15:44
• Yes, this is exactly the idea behind reasoning modulo $p$ – Junkyards Jan 30 at 15:46
• If you show that if $gcd(a,p)=1$ and $gcd(b,p)=1$ than $gcd(a \cdot b,p)=1$ you are done, are'nt you? – Shaq Jan 30 at 15:49

I think you are mixing up integers and classes of integers modulo $$p$$. The set you just gave should not be set of integers, but a set of classes of integers modulo $$p$$. Indeed, it if was a set of integers, it would not be a group : $$(p-1) \times 2$$ is not in this set !
But if you consider these elements as classes of integers modulo $$p$$, then all of your reasoning makes sense, and the element $$a$$ you get is indeed the inverse you are looking for : it's a class of integers. It just turns out that there is a representative of it between $$1$$ and $$p-1$$.