3

I am wondering about a proof of the fact that If $f\in L^1(\mathbb{R})$ is such that $\int_{\mathbb{R}}f\phi=0$ for all continuous compactly supported $\phi$, then $f\equiv 0$. I am familiar with the proof given here http://www.academia.edu/740170/An_Introduction_to_Weak_Derivatives on page 4, but I am wondering if there is another proof using density.

Basically something like take a sequence of $\phi_n\in C_c({\mathbb{R}}$) such that $\phi_n\to f$ in $L^1$, then $\displaystyle0=\int_\mathbb{R} f\phi_n\to\int_{\mathbb{R}}f^2$, hence $f=0$ a.e. But I need to justify this last limit. If $f$ is bounded it is fine, since $\displaystyle|\int_\mathbb{R} f\phi_n-f^2|\le\int_\mathbb{R} |f||\phi_n-f|\le\|f\|_{\infty}\int_\mathbb{R}|\phi_n-f|\to 0$. I am having trouble doing it without the assumption that $f$ is bounded though.

JLA
  • 6,932

1 Answers1

5

Take $\phi\in \mathcal C^\infty_c(\Bbb R)$, such that $\phi\ge 0$ and $\int_{\Bbb R}\phi(x)dx=1$. Let's call $\phi_n(x) = n\phi(nx)$.

On the one hand, this is a regularising sequence. Therefore, $\phi_n\ast f\to f$ in the sense of $L^1$ (the proof of this fact, however, relies on density of $\mathcal C _c$ in $L^1$.)

On the other hand, since $\phi_n(y-x)\in \mathcal C^\infty_c(\Bbb R)$ for a fixed $y$, we can write by our hypothesis $$0=\int_{\Bbb R}\phi_n(y-x)f(x)dx.$$

Thus, $f=0$ in the sense $L^1$.

TZakrevskiy
  • 23,308