1

I feel like I'm meant to use the uniqueness of prime factorization and show that since $\text{lcm}(3,2)=6$ each prime should appear 6 times, but how would I justify this?

Will Fisher
  • 5,112
  • Can you prove the simpler statement that if $n = x^2$ then the maximum exponent of any prime factor of $n$ is even? – JMoravitz Dec 18 '18 at 23:34

4 Answers4

1

Use the unique prime factorization.

First not that if $p$ is a prime factor of $x$ then $p|x^2$ so $p|y^3$ so $p|y$. And if $q$ is a prime factor of $y$ then $q|y^3$ so $q|x^2$ so $q|x$. So $x$ and $y$ have the exact same prime factors.

So if the distinct prime factors of $x$ are $P=\{p_1, p_2, ..... p_n\}$ the $x = \prod\limits_{k=1}^n p_k^{a_i}$ for some $a_i \ge 1$ and $y = \prod\limits_{k=1}^n p_k^{b_i}$ for some $b_i \ge 1$.

Now $x^2 = y^3$ so

$(\prod\limits_{k=1}^n p_k^{a_i})^2 = (\prod\limits_{k=1}^n p_k^{b_i})^3$ so

$\prod\limits_{k=1}^n p_k^{2a_i}=\prod\limits_{k=1}^n p_k^{3b_i}$ so (because the is a unique prime factorization of the value $x^2$ which is $y^3$) we have:

$2a_i = 3b_i$ which means $2|3b_i$ for each $b_i$ and $3|2a_i$ for each $a_i$. And as $2,3$ are primes that means $2|b_i$ and $3|a_i$ for each $b_i$ and each $a_i$.

So each $b_i = 2c_i$ for some integer $c_i$. So $y = \prod\limits_{k=1}^n p_k^{b_i} = \prod\limits_{k=1}^n p_k^{2c_i}= (\prod\limits_{k=1}^n p_k^{c_i})^2$. So $y$ is a perfect square.

So each $a_i = 3d_i$ for some integer $d_i$. So $x = \prod\limits_{k=1}^n p_k^{a_i} = \prod\limits_{k=1}^n p_k^{3d_i}= (\prod\limits_{k=1}^n p_k^{d_i})^3$. So $x$ is a perfect square.

That's it.

fleablood
  • 124,253
0

Write $x^2 = \prod_{i=1}^k p_i^{l_i}$, for some positive finite integer $k$ where $p_i$ is the $i$-th prime and $l_i$ is a nonegative integer. Then as $\prod_{i=1}^k p_i^{l_i}$ is a square [because $x$ is an integer] each $l_i$ must be a multiple of 2.

As $\prod^k_{i=1} p_i^{l_i}$ is also a cube [because $y$ is an integer] each $l_i$ must also be a multiple of 3. Thus each $l_i$ must be a multiple of 6.

Then $x = \prod_{i=1}^k p_i^{\frac{l_i}{2}}$ [because $x^2 = \prod_{i=1}^k p_i^{l_i}$] and $y = \prod_{i=1}^k p_i^{\frac{l_i}{3}}$ [because $y^3 = \prod_{i=1}^k p_i^{l_i}$], which as $l_i$ is a multiple of 6, implies that $\frac{l_i}{2}$ is a multiple of 3, which implies that $x$ is a cube; namely $x= \left(\prod_{i=1}^k p_i^{\frac{l_i}{6}}\right)^3$ with each $\frac{l_i}{6}$ an integer. Likewise $y$ is a square, namely $y= \left(\prod_{i=1}^k p_i^{\frac{l_i}{6}}\right)^2$ with each $\frac{l_i}{6}$ an integer..

Mike
  • 20,434
0

Here's an alternate proof which uses unique factorization in a different way: Given $x^2 = y^3$ with $x,y \ne 0$, define $a := \frac{x}{y}$. Then $a^2 = \frac{x^2}{y^2} = \frac{y^3}{y^2} = y$ and $a^3 = \frac{x^3}{y^3} = \frac{x^3}{x^2} = x$.

It remains to show that $a \in \mathbb{Z}$. We do know that $a \in \mathbb{Q}$ and that $a^2 = y \in \mathbb{Z}$. Now, look at one of the standard proofs that $\sqrt{2}$ is irrational using unique factorization; this will generalize in a fairly straightforward manner to show the desired fact that if $a \in \mathbb{Q}$ and $a^2 \in \mathbb{Z}$, then $a \in \mathbb{Z}$.

(This can also be viewed as a corollary of the fact that $\mathbb{Z}$ is integrally closed, since $\mathbb{Z}$ is a unique factorization domain and any unique factorization domain is integrally closed.)

0

Here's a proof by infinite descent that $x$ is necessarily a perfect cube. The proof that $y$ is a perfect square is of the same exact spirit.

Without loss of generality assume that $x,y$ are positive. Suppose there were to exist pairs of positive integers $(x,y)$ with $x^2=y^3$ but $x$ not a perfect cube. Choose the pair $(x_0,y_0)$ with $x$ minimal. Clearly $x_0,y_0>1$. Consider some prime $p$ dividing $x_0$ and write $x_0=px_1$. Then $p\mid y_0^3\implies y_0=py_1$ for some $y_1\in\mathbb Z_{\geq 0}$. So $$p^2x_1^2=p^3y_0^3\implies x_1^2=py_1^3\implies p\mid x_1$$ Write $x_1=px_2$, again with $x_2$ a positive integer. Then $$p^2x_2^2=py_1^3\implies px_2^2=y_1^3\implies p\mid y_1$$ Next let $y_1=py_2$. We have that $$px_2^2=p^3y_2^3\implies x_2^2=p^2y_3^3\implies p\mid x_2$$ Finally, take $x_2=px_3$. Our equation becomes $$p^2x_2^2=p^2y_2^3\implies x_3^2=y_2^3$$ So $(x_3,y_2)$ is a pair of positive integers satisfying $x_3^2=y_2^3$, but clearly $x_3<p^3x_3=x_0$. A contradiction $\blacksquare$

Rafi
  • 1,531
  • 6
  • 11