Can someone please check my proof?

Let $n \in \mathbb{N}$ such that n is a non-perfect square. Show that $\sqrt{n}$ is irrational.

Let's prove it by contradiction, that is, suppose there are $n,a,b \in \mathbb{N}$ such that $\frac{a}{b}=\sqrt{n}$, n is a non-perfect square and there are no common factors for a and b, that is, they are coprime. n is either even or odd. Suppose n is even, that means $n=2k$ for some $k \in \mathbb{N}$. Therefore: $$ \begin{align*} \frac{a^2}{b^2}&=2k\\ a^2 &= 2k\cdot b^2 = 2(k\cdot b^2) \end{align*} $$ Which means that $a^2$ is even and therefore a is even and that means that there is some $j\in \mathbb{N}$ such that $a=2j$. Plugging in that value for a: $$ \begin{align*} \frac{2j}{b^2}&=2k\\ 4j^2 &= 2k\cdot b^2 \rightarrow b^2 = 2\big(\frac{2j^2}{k}\big) \end{align*} $$ Using the same reasoning b is also even. That contradicts the fact that a and b are coprime and therefore $\sqrt{n}$ is irrational.

Now suppose n is odd, that means that $n=2k+1$ for some $k\in \mathbb{N}$. Therefore: $$ \begin{align*} \frac{a^2}{b^2}&=2k+1\\ a^2 &= (2k+1)\cdot b^2 \rightarrow (2k+1)|a^2 \end{align*} $$ Since $(2k+1)|a^2$ we can conclude that $a^2 = q(2k+1)$. Plugging in that value for a: $$ \begin{align*} \frac{q(2k+1)}{b^2}&=2k+1\\ q(2k+1) &= (2k+1)\cdot b^2 \rightarrow q=b^2 \end{align*} $$ That way q is a common factor for a and b, which contradicts the fact that they are coprime, hence $\sqrt{n}$ is irrational.

Is my reasoning correct? I've seen easier and shorter proofs, by I tried to generalize the reasoning that I used to prove that $\sqrt{2}$ and $\sqrt{3}$ are irrational...

Thank you for reading it!

  • $\begingroup$ Why are you doing separate even and odd cases? $\endgroup$ – fleablood Sep 11 '18 at 1:02
  • $\begingroup$ A more direct way would be to write $n = p_1^{r_1} \dots p_k^{r_k}$ where the $p$'s are prime and some $r_i$ is odd (since $n$ is not a perfect square). $\endgroup$ – Austin Mohr Sep 11 '18 at 1:04
  • $\begingroup$ Also $2kb^2$ does not mean $b$ is even. What if $k$ is even? then you have $4mb^2$ and that's not a contradiction. $\endgroup$ – fleablood Sep 11 '18 at 1:04
  • $\begingroup$ "Using the same reasoning b is also even. " Not if $k$ is divisible by $4$. But more important what other primes is $k$ divisible by? Apparently none if $k$ and $b$ are co-prime. That is really more to the point. $\endgroup$ – fleablood Sep 11 '18 at 1:08
  • $\begingroup$ " I've seen easier and shorter proofs, by I tried to generalize the reasoning that I used to prove that 2–√ and 3–√ are irrational..." But those use even and divisible by $3$ BECAUSE they are square roots of 2 and 3. It's not relevent for other values of $n$. $\endgroup$ – fleablood Sep 11 '18 at 1:10

As the comments suggest, you overanalyze. If $\frac{a}{b}=\sqrt{n}$ then $\frac{a^2}{b^2}=n$ and $a^2=nb^2$. Thus, as with the cases of $2$ and $3$, $n\mid a^2$. Now, for every prime factor $p_i$ of $n$, if $p_i\mid a^2$, then $p_i\mid a$. Thus $(p_i)^2\mid a^2$. Remove that $p_i$ from each occurrence of $n$ and $a$, and you are left with the fact that $\frac{n}{p_i}\mid \frac{a^2}{(p_i)^2}$. Repeat for every other prime factor of $n$ to arrive at the fact if $n\mid a^2$ then $n\mid a$. So $a=kn,a^2=k^2n^2$, and from here the proof is the same as for $2$ and $3$ in that you can show that $n$ must be a factor of $b$.


Don't bother with even or odd cases. Just do:

$\frac {a^2}{b^2} = n$ so $a^2 = nb^2$. So for any prime, $p$ that divides $b$ then $p|a^2$ so $p|a$. That's impossible because $a$ and $b$ are coprime. So there are no primes that divide $b$ (!!!!!!). So $b = 1$. And $a^2 = n$ and $n$ is a perfect square.

That's it.

  • $\begingroup$ I've seen this proof, I was Just trying to get another one using cases, but it's not necessary. Thanks anyway $\endgroup$ – Bruno Reis Sep 11 '18 at 1:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.