Canonical: https://knowlpedia.com/real-analysis/positive-powers-of-flat-weighted-profiles/

# Positive powers of flat weighted profiles

Two-sided flat-weight control and weighted derivative bounds give smooth zero extension of every fixed positive power.

Let \(w>0\) be a [flat interval weight](https://knowlpedia.com/real-analysis/flat-logarithmic-interval-weight/) in an open interval and let \(\delta\) denote its positive boundary margin. Suppose a smooth positive \(y\), possibly with additional compact parameters, satisfies
\[
y\ge c w,\qquad |\partial^\alpha y|\le C_\alpha w\delta^{-N_\alpha}
\]
for every fixed multi-index, including zero. For every fixed \(r>0\), the function \(y^r\) extends smoothly by zero across the interval endpoints.

## Derivative estimate

Repeated chain and product rules express each derivative as a finite sum of terms
\[
c\,y^{r-j}\prod_{\nu=1}^j\partial^{\alpha_\nu}y,
\qquad |\alpha_\nu|\ge1.
\]
The upper bound on \(y\) controls positive powers; the lower bound controls negative powers. Each term is bounded by \(Cw^r\delta^{-N}\) for some finite \(N\). A positive power of the flat weight beats every such inverse margin power, so all mixed derivatives tend to zero at the boundary. The zero-extension criterion applies. In particular \(r=1/2\) justifies square roots under these quantitative hypotheses.
