Euler’s Coprimes

Problem 1. Given integers a, b, k, show that \gcd(a,b) = \gcd(a+kb, b).

(Click for Solution)

Solution. Let c = \gcd(a,b) and d = \gcd(a+kb, b). Then

c \mid a\quad \text{and} \quad c \mid b \quad \Rightarrow \quad c \mid (a+kb),

so that c \mid d. Similarly,

d \mid (a+kb)\quad \text{and} \quad d \mid b \quad \Rightarrow \quad d \mid ((a+kb) - kb)= a,

so that d \mid c.

Let n > 1 denote any positive integer.

Definition 1. Define

\Phi(n) := \{ 1 \leq x \leq n : \gcd(x, n) = 1\}

and Euler’s totient function \varphi by \varphi(n) := |\Phi(n)|.

Problem 2. Show that

G_n := \{ [x]_n : x \in \Phi(n)\} \subseteq \mathbb Z_n^*

forms an Abelian group of order \varphi(n) under multiplication.

(Click for Solution)

Solution. To establish closure, we need to show that for x, y \in \Phi(n), [xy]_n = [m]_n for some m \in \Phi(n). To that end, use Bézout’s lemma to furnish integers r_1, s_1, r_2, s_2 such that

\begin{aligned} r_1 \cdot x + s_1 \cdot n &= 1,\\ r_2 \cdot y + s_2 \cdot n &= 1. \end{aligned}

Multiplying both equations together,

(r_1 r_2) \cdot xy + (r_1 s_2 x + r_2 s_1 y + s_1 s_2 n) \cdot n = 1.

By Bézout’s lemma again, \gcd(xy, n) = 1.

The identity is clearly [1]_n and the inverse of [x]_n exists from \gcd(x, n) = 1.

Problem 3 (Euler’s Theorem). Show that for any integer a such that \gcd(a, n) = 1,

[a]_n^{\varphi(n)} = [1]_n.

(Click for Solution)

Solution. Assume 1 \leq a \leq n for simplicity, so that a \in \Phi(n). If a \neq 1, then \langle [a]_n \rangle \leq G_n. By Lagrange’s theorem,

m := | \langle [a]_n \rangle | \mid |G_n| = \varphi(n).

Find an integer k such that \varphi(n) = km. Then

\begin{aligned} [a]_n^{\varphi(n)} &= [a]_n^{km} \\ &= ([a]_n^m)^k \\ &= [1]_n^k = [1]_n. \end{aligned}

For the general case, use the division algorithm to find integers q,r with 0 \leq r < n such that a = qn + r. By Problem 1,

\begin{aligned} \gcd(r,n) &= \gcd(a-qn,n) \\ &= \gcd(a,n)  \\ &= 1 . \end{aligned}

By the special case, [a]_n^{\varphi(n)} = [r]_n^{\varphi(n)} = [1]_n.

Problem 4. Verify Fermat’s little theorem: for any integer a and prime p,

[a]_p^p = [a]_p.

(Click for Solution)

Solution. The result for a = 0 is trivial. For 1 \leq a \leq p-1, \varphi(p) = p-1, so that

\begin{aligned} [a]_p^p &= [a]_p \cdot [a]_p^{p-1} \\ &= [a]_p \cdot [a]_p^{\varphi(p)} \\ &= [a]_p \cdot [1]_p \\ &= [a]_p. \end{aligned}

For the general case, write a = qp + r where 0 \leq r \leq p-1:

[a]_p^p = [r]_p^p = [r]_p = [a]_p.

—Joel Kindiak, 20 Jan 26, 1940H

,

Published by


Leave a comment