Suppose $e^A = A$, prove that $A$ is diagonalizable, where A is a matrix.
What I have tried to do is write $A= D + N$, where $D$ is diagonalizable, $N$ is nilpotent and $DN = ND$.
Since $N$ is nilpotent, there exist a minimal $n$ such that $N^n=0$.
Then $e^A=e^{D+N}=e^De^N=e^D(I+N+\frac{N^2}{2}+...+\frac{N^{n-1}}{(n-1)!})=A=D+N$.
If we times $N^{n-1}$ on both side, then what remain is $e^DN^{n-1}=DN^{n-1}$.
And then I don't know how to carry on.
Please help! Do I need a new method to do this question? Thanks a lot!
You write $$D+N=e^{D}e^{N}=e^{D}+e^{D}(e^{N}-I)$$
As the matrix $e^{D}(e^{N}-I)$ is nilpotent (because it is of the form $e^{D}NQ(N)$ and everything commutes), $e^D$ is diagonalizable (because $D$ is), and these two matrices commutes, from uniqueness in Dunford decomposition, you get:
$$D=e^D, \;\;\;\;N=e^{D}(e^{N}-I)$$
Therefore you get, multiplying by $N^{k-1}$ where $k$ is the smallest integer such that $N^{k+1}=0$ (and assuming by contradiction $k\geq 1$):$$N^k=DN^k,$$ in other words $$(I-D)N^k=0.$$ Now, is $(I-D)$ inversible? Yes, as $D=e^D$, 1 cannot be an eigenvalue for $D$, and you conclude $N^k=0$, which by definition of $k$, constitute a contradiction, and finish the proof as $k=0$ and $N=0$.