I'm studying $p$-adic modular forms, and in particular I was trying to understand Swinnerton-Dyer's proof on the structure of the algebra of modular forms (of level 1) modulo $p\geq 5$.
In his proof, he states that given any modular form $f=\sum c_{a,b}E_4^aE_6^b$ with rational and $p$-integer coefficients in its $q$-expansion, the $c_{a,b}$ are necessarily rational and $p$-integer. This seems very natural, but getting my hands on it, I can't seem to find a proof.
In his "Formes modulaires et fonctions zêta $p$-adiques", Serre hints that this may be done by induction on the weight, using that $\Delta$ has this property; however this hasn't helped me much.
Thank you for your help!
I’m assuming $p\geq 5$. In weight at most $14$, I’ll leave it to you to check it.
Let $f \in \mathbb{Z}_{(p)}[[q]]$ the $q$-expansion of a modular form of weight $w >14$, and write $f=\sum_{a,b}{c_{a,b}E_4^aE_6^b}$, and we want to show that the $c_{a,b}$ are $p$-integers. Write $\gamma \in \mathbb{Z}_{(p)}$ the value at infinity of $f$, and consider $f_2=f-\gamma E_4^aE_6^b$: it has a $p$-integral and rational $q$-expansion.
It’s easy to see that $\frac{\Delta(q)}{q} \in \mathbb{Z}[[q]]$ has constant value $1$, so that it has an inverse $i(q) \in \mathbb{Z}[[q]]$. Now, let $f_3=i(q)\frac{f_2(q)}{q}=\frac{f_2}{\Delta}$, then $f_3$ is the $p$-integral $q$-expansion of a modular form of weight $w-12$ so it is a $\sum_{a,b}{d_{a,b}E_4^aE_6^b}$ with rational $p$-integral $d_{a,b}$. Now $f_2=f_3\frac{E_4^3-E_6^2}{1728}$ and $1728=12^3$ is a unit in $\mathbb{Z}_{(p)}$.