In proposition 2.5 (p77) of Hartshorne's book, Hartshorne considers a graded ring $S = \bigoplus_{d =0}^\infty$ and the localisation $S_f$ of a homogeneous element $f \in S^+= \bigoplus_{d=1}^\infty S_d$ (i.e. $f \in S_d$ for some $d \geq 1$).
I.e. we localise $S$ w.r.t. the multiplicatively closed set $\{1,f,f^2, f^3, \dots\}$. How are we sure this localisation makes sense? I.e. why is $f$ not nilpotent?
Let's recall the relevant portion of the proposition in question:
If $f$ is nilpotent, then $S_f=S_{(f)}=0$, and $\operatorname{Spec} S_{(f)} = \emptyset$, which is a perfectly fine open subset.