1

I want to show that no group of order $90=2 \cdot3^2 \cdot5 $ is simple. Part (iii) of Sylow's Theorem gives

$$n_2 \in\{1,3,5,9,15,45\} $$ $$n_3 \in \{1,10\} $$ $$n_5 \in \{1,6\} $$

I am aware of the solution that appears here, but I wonder if a more "brute force" approach is possible, as follows:

Assuming $G$ is a simple group of order $90$, we have

$$n_2 \in\{3,5,9,15,45\} $$ $$n_3=10 $$ $$n_5=6 $$

G then has $6(5-1)=24$ elements of order $5$, and $n_2$ elements of order $2$. The remaining $66-n_2$ elements must be exactly those in the union $$\bigcup_{i=1}^{10} P_i $$ of the Sylow 3-subgroups.

We now try to estimate the size of that union: The inclusion-exclusion principle gives

$$\left| \bigcup_{i=1}^{10} P_i \right|=\sum_{n=1}^{10} (-1)^{n+1} \sum_{i_1 <i_2<\dots <i_k} \left|P_{i_1} \cap P_{i_2} \cap \dots \cap P_{i_k} \right| $$

Obviously, the $n=1$ term in the sum above is $90$, and every intersection has possible order $1$ or $3$ (According to Lagrange's Theorem). Is is possible to find the minimal value of this expression?

I thought about splitting each inner sum into two, the sum over trivial intersections, and intersections of size $3$. But the calculations seem hard to handle.

Thank you!

user1337
  • 24,859

1 Answers1

0

Well, in general we have this, but it might be to sophisticated.

Theorem Let $P \in Syl_p(G)$, then $|G -\bigcup_{g \in G}P^g| \equiv 0 \text { mod } |P|$.

See here for a proof. If you apply this it follows that $n_2=3$ and this gives a contradiction since $G$ can then be embedded in $S_3$ (index$[G:N_G(S)]=3$ for an $S \in Syl_2(G)$).

Nicky Hekster
  • 52,147