Originally Posted by

**Lipticboven** "Assuming there exist two integers $\displaystyle p$ and $\displaystyle q$ such that $\displaystyle a^p=b^q=0$, can you find $\displaystyle n$ such that $\displaystyle (a+b)^n=0$ ?"

Is it when n = max(p,q)?

No : for $\displaystyle p=q=2$ one has $\displaystyle (a+b)^{\max(p,q)}=(a+b)^2=2ab$ which may or may not equal 0, depending on how $\displaystyle a$ and $\displaystyle b$ were chosen.

A sufficient condition which gives us $\displaystyle (a+b)^n = 0$ is

for all $\displaystyle k$ such that $\displaystyle 0\leq k\leq n$, either $\displaystyle a^k=0$, either $\displaystyle b^{n-k}=0$.

In other words :

for all $\displaystyle k$ such that $\displaystyle 0\leq k\leq n$, either $\displaystyle k\geq p$, either $\displaystyle n-k\geq q$.

(you don't have to find the lowest integer $\displaystyle n$ such that $\displaystyle (a+b)^n=0$, you have to find *an* integer $\displaystyle n$ such that $\displaystyle (a+b)^n=0$)

Originally Posted by

**flyingsquirrel** You can look for this counterexample in $\displaystyle \mathcal{M}_2(\mathbb{R})$ (the ring of real $\displaystyle 2\times 2$ matrices).

I'd like to add that strictly triangular matrices are nilpotent. They look like $\displaystyle \begin{bmatrix} 0 & \alpha \\ 0 & 0\end{bmatrix}$ or $\displaystyle \begin{bmatrix} 0 & 0 \\ \beta & 0\end{bmatrix}$.