I was asked to prove for $n \in \mathbb{N}$ that $T_n=2^{2^n+1}+1$ is composite.
One solution by induction is as follows:
Base case $T_1$: $2^{2^n+1}+1=2^3+1=9=3^2$, so $T_1$ is composite with a factor of $3$.
Inductive step: $T_{n+1}=2^{2^{n+1}+1}+1 =2^{2^{n}}2^{2^{n}+1}+1$
Now, the first few terms of $T_n$ are $9, 33, 513$, suspiciously all multiples of $3$. If we reframe this inductive step as $T_n \equiv 0 \mod 3$, with a quick intermediate proof that for $n \in \mathbb{N}, 2^{2^n} \equiv 1 \mod 3$, we have:
modulo $3$: $2^{2^{n}}2^{2^{n}+1}+1 = 2^{2^{n}}(T_n-1)+1 \equiv (1)(2)+1\equiv 0$
This proves that $T_n$ is always a multiple of $3$ for $n \geq 1$. Since all $T_n > 3$, all $T_n$ are composite.
But here is my question: the proof above uses the fact that $T_n$ has a factor of $3$, more specifically than just that $T_n$ is composite. Is there a proof that doesn't use the factor of $3$ as an assumption? I've not found it. The inductive step can also be represented as $T_{n+1}=\frac{(T_n-1)^2}{2}+1$ with the assumption that $T_n$ is composite and odd.