Let $X$ be a infinite dimensional Hilbert Space that is separable and has $\{e_{n}\}_{n \in \mathbb N}$ a countable Orthonormal Basis.
I have been given the map:
$F: X \to \ell^{2}, x\mapsto (\langle e_{n},x\rangle)_{n \in\mathbb N}$
I need to show that: $2.$ $F$ is surjective
Idea on $2$:
I am supposed to use the following hint:
if for $(c_{n})_{n\in \mathbb N}\subset\mathbb R$ we know that $\sum\limits_{n \in \mathbb N}\vert c_{n}\vert^{2}<\infty$ then $\sum\limits_{n\in\mathbb N}c_{n}e_{n}$ is well defined, where $\{e_{n}\}_{n \in \mathbb N}$ is an orthnormal set.
How is this supposed to help me?
Edit, to show surjectivity:
Let $y \in \ell^{2}$ arbitrary: We know that therefore $\sum\limits_{n \in \mathbb N}\vert y_{n}\vert^{2}<\infty$ from the tip this then implies that $\sum\limits_{n \in \mathbb N}y_{n}e_{n}$ is well-defined in $X$, in other words, there exists a $x\in X$ so that $x=\sum\limits_{n \in \mathbb N}y_{n}e_{n}$.
Now by definition of $F$, it follows that $Fx=(\langle e_{n},x\rangle)_{n \in\mathbb N}=(\langle e_{n},\sum\limits_{n \in \mathbb N}y_{n}e_{n}\rangle)_{n \in\mathbb N}=(y_{n})_{n \in \mathbb N}=y$. But now I have a problem: what there if there is another $z \in \ell^{2}$ which is unequal to $y$ so that $x =\sum\limits_{n \in \mathbb N}z_{n}e_{n}$, then I'd have $Fx=y\neq z =Fx$. How does well-definedness alleviate this problem? Because well-definedness of $\sum\limits_{n \in \mathbb N}y_{n}e_{n}$ according to my understanding leads only to the fact that there cannot be two different $x,w$ where $x \neq w$ so that $x = \sum\limits_{n \in \mathbb N}y_{n}e_{n}$ and $w = \sum\limits_{n \in \mathbb N}y_{n}e_{n}$. Is my understanding of well-definedness correct?
$1.$ Question: What does well-defined in $X$ mean in this regard, does it mean there can be only one $x\in X$ so that $x=\sum\limits_{n \in \mathbb N}y_{n}e_{n}$?
To answer the surjectivity, I would argue in the following way:
If we would assume that $\exists c:= (c_n)_n \in {\ell}^2: c \notin \operatorname{ran}(F)$, then $\sum_{n \in \Bbb N} c_n e_n$ is well-defined by your hint, but $\sum_{n \in \Bbb N} c_n e_n \notin \overline{\operatorname{span}\{e_n\}}$ because otherwise $\sum_{n \in \Bbb N} c_n e_n = \sum_{n \in \Bbb N} \langle x, e_n \rangle e_n $ for some $x \in X$ which of course leads to a contradiction because $\{e_n\}$ is a ONB.