# Equivalence of Sobolev Norms via Fourier Transformation

I am working through a proof which shows that the usual Sobolev norm $\|\cdot\|_{k,2}$ on $W^{k,2}(\Omega)$ is equivalent to the norm

$$\|u\|_{H^m} := \| (1 + |\cdot|^2)^{\frac{k}{2}}\mathcal{F}u \|_2$$

At one point we use the inequality

$$2^{-k} \leq \frac{1 + |\xi|^{2k}}{(1 + |\xi|^2)^k} \leq 2.$$

Since the fraction always is smaller than one, the right inequality is not a problem. But I can't find a way to prove the left inequality. Is there an elementary proof to this inequality?

• $\|f\|^2_{W^{k,2}} = \sum_{|m|\le k} \|\partial_m f\|^2_{L^2} = \sum_{|m|\le k} \|\ |\xi|^m \hat{f}\|^2_{L^2} \sim\|\ (1+|\xi|)^{k} \hat{f}\|^2_{L^2}$ (where $\sim$ means the two norms are obviously equivalent) – reuns Sep 21 '16 at 17:28

Yes. Distinguish the cases $|\xi|\leq 1$ and $|\xi|> 1$ and use crude bounds.