[Math] gcd(a,b)=d then show that gcd(a/d,b/d)=1 (without euclids lemma or bezout’s theorem)

elementary-number-theory

I was solving the problem $\gcd(a,b)=d$ then show that $\gcd\left(\frac{a}{d},\frac{b}{d}\right)=1$ (without Euclid's lemma or Bézout's theorem) then stumbled upon the fact that if we say
$$\gcd\left(\frac{a}{d},\frac{b}{d}\right)= k \tag{1}$$
then $a=kdn$ and $b=kdm$ then,
$$\gcd(a,b)=\gcd(m,n) \, kd \tag{2}$$
but I seem to not be able to prove that $(1)$ holds. If this holds then from this it’s clear that $d = \gcd(m,n)\,kd$ then $\gcd(m,n) = \frac{1}{k}$

Since $k$ is an integer and $\gcd(m,n)$ is also an integer then only value of $k$ that satisfies such a relation is $k = 1$ then by $(1)$ our question is answered.

So the question here is if two numbers $a = A \cdot B$ and $b = M \cdot N$, then
is it true that,
$$\gcd(a,b) = \gcd(A,M) \cdot \gcd(B,N)?$$

Best Answer

Let $a,b>0$, $d=(a,b)$. Let $a'=ad^{-1},b'=bd^{-1}$. Suppose that $e\mid a',b'$. Then $de\mid a,b$, so $de\mid (a,b)=d$. This means $e\mid 1$. It follows that $(a',b')=1$.