Question: Find the minimal polynomial of $\alpha=\sqrt{3+2\sqrt{2}}$ over $\mathbb{Q}$

Thoughts: the "standard" method of starting by squaring (twice) to get rid of the square roots, because I don't have a nice way of showing the resulting polynomial is irreducible. So..

Attempt: It would be great if I could get our $\alpha$ in the form $$(a+b)^2=a^2+2ab+b^2=\sqrt{3+2\sqrt{2}}.$$ So, $$3+2\sqrt{2}=(\sqrt{2})^2+2\sqrt{2}+1=(\sqrt{2}+1)^2.$$So, $$\alpha=\sqrt{3+2\sqrt{2}}=\sqrt{2}+1.$$So, $$(\alpha-1)^2=2\\ \alpha^2-2\alpha+1=2\\ \alpha^2-2\alpha-1=0.$$So let $f(x)=x^2-2x-1$. Since $f(x)$ is irreducible over $\mathbb{Q}$ by the Rational Roots Test (since it has degree $2$), $f(x)$ is monic, and $f(\alpha)=0$, we conclude that $f(x)$ is the minimal polynomial of $\alpha$ over $\mathbb{Q}$.

Does this look okay?

  • 2
    $\begingroup$ What you have done is good. If you were to try squaring twice to remove the square roots you would get a quartic which would be reducible and would have the minimal polynomial as an irreducible factor. $\endgroup$
    – Ben
    Jul 16, 2021 at 22:10
  • 1
    $\begingroup$ @Ben Yeah, I was stuck there for a bit when good ol' Eisenstein's didn't work. $\endgroup$
    – User7238
    Jul 16, 2021 at 22:12

2 Answers 2


OP's proof is good, and there is not much room left to simplify. The following are just alternatives.

  • Let $\beta = \sqrt{3 - 2\sqrt{2}}$ and note that $\alpha\beta=1$ and $\alpha^2+\beta^2=6$. Given that $0 \lt \beta \lt \alpha$ it follows that $\alpha-\beta=\sqrt{(\alpha-\beta)^2}=\sqrt{\alpha^2+\beta^2-2\alpha\beta}=\sqrt{6-2}=2$.

    Then $\alpha(-\beta)=-1$ and $\alpha+(-\beta)=2$, so $\alpha, -\beta$ are the roots of $x^2 - 2 x - 1\,$.

  • The "standard" method can also work. Suppose "starting by squaring (twice) to get rid of the square roots", then $x^4 -6x^2+1=0=(x^2-1)^2-4x^2=(x^2-2x-1)(x^2+2x-1)$. Only the first quadratic factor has a root larger than $2$, which must be $\alpha$.

  • $\begingroup$ Why would having a "root larger than $2$" justify that it MUST be $\alpha$. I intuitively understand that, it is just one of those things that as soon as I think it is true, I can see myself finding some trivial counterexample. $\endgroup$
    – User7238
    Jul 16, 2021 at 23:36
  • $\begingroup$ @User7238 Because $\alpha$ must be the root of one of the factors, and we know $\alpha \gt 2$ since $\alpha^2 = 3+2\sqrt{2}\gt 4$. (Actually, $\alpha \gt 1$ would suffice, since the other factor has both roots smaller than $1$.) $\endgroup$
    – dxiv
    Jul 16, 2021 at 23:38

Since the comment has answered the OP's stated question, it is open season.

Alternative approach

If $a$ is rational and $b$ is irrational, then $(a+b)$ is irrational.

Further, if $a$ is irrational, then $\sqrt{a}$ is also irrational.

The proof to the 2nd assertion above is that if $\sqrt{a}$ can be rationally expressed as $\frac{P}{Q}$, this implies that $a$ has the rational expression $\frac{P^2}{Q^2}$, which contradicts the premise that $a$ is irrational.

Using the two assertions, you have that since $\sqrt{2}$ is known to be irrational, so is $\left[3 + 2\sqrt{2}\right]$, and therefore, so is $\sqrt{3 + 2\sqrt{2}}.$

This implies that there can not be any polynomial equation of degree 1, with form $x + b = 0,$ with $(b)$ rational, whose root is $\sqrt{3 + 2\sqrt{2}}.$


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.