1

I'm currently working on a proof, but didn't get the right idea, a hint would be appreciated:

Show that $\mathbb{Z}/n^2\mathbb{Z}$ is not isomorphic to $\mathbb{Z}/n \times \mathbb{Z}/n$

So, lets assume there is an isomorphism between these groups, my first idea to was to construct a contradiction with the homomorphism property. But I don't know how, there should be an element in $\mathbb{Z}/n \times \mathbb{Z}/n$ which is not in $\mathbb{Z}/n^2$ to reach this kind of contradiction. Of course there is $n \in \mathbb{Z}/n^2$ which is not in the other group but I don't know if this helps.

Thanks

Derato
  • 33
  • Think about the order of an element of greatest order in each of the two groups. – Gerry Myerson Mar 06 '18 at 08:12
  • The first group is cyclic the latter one is not, indeed the order of $(a,b)\in \mathbb{Z}_n\times \mathbb{Z}_n$ cannot exceed $n$. This is exactly what Gerry Myerson says. – Mathematician 42 Mar 06 '18 at 08:22
  • Thanks but we haven't yet defined the terms cyclic and order. I googled those terms and while that makes sense, I would like to use another approach. – Derato Mar 06 '18 at 08:26

2 Answers2

3

Suppose $\phi:\mathbb{Z}_{n^2}\to\mathbb{Z}_n\times \mathbb{Z}_n$ is an isomorphism. Notice that $$\phi(x)=\phi(\sum_{i=1}^x1)=\sum_{i=1}^x\phi(1).$$ Now write $\phi(1)=(a,b)$ for certain $a,b$. Then $\sum_{i=1}^x\phi(1)=(\sum_{i=1}^xa,\sum_{i=1}^xb)=x\cdot(a,b)$. It follows that $\phi(1)=(a,b)=(n+1)\cdot(a,b)=\phi(n+1)$. Hence $\phi$ is not injective, a contradiction!

Remark that $a=(n+1)\cdot a$ in $\mathbb{Z}_n$ by definition of $\mathbb{Z}_n$, this explains the equality $(a,b)=(n+1)\cdot(a,b)$ above.

Mathematician 42
  • 13,176
1

Z/(n^2Z) is cyclic but Z/(nZ) × Z/(nZ) is non cyclic (as every element of this group has order n). Hence they are not isomorphic.

Biman Roy
  • 335