I got stuck in the following lemma:
Lemma 1: Let $B$ be the unit ball in the 4 dimensional Euclidean space. Suppose that $u\in W^{2,2}(B)$, then $e^{u}\in L^{q}$ for any $q\geq1$$q>1$.
Actually, I realized that the Lemma 1 follows from Trudinger's theorem. https://en.wikipedia.org/wiki/Trudinger%27s_theorem
---------------------------------------------------------
Below is my little thoughts about this problem in the beginning and just want to keep it for myself. As we know this is exactly the critical case of Sobolev's imbedding where we fail to get $L^{\infty}$ bounds. Any suggestion and help would be appreciated.
My idea of this that fails: Note $\int_{B} e^{pu}\leq \sum_{k=1}^{\infty}\frac{p^{k}}{k!}\int_{B} |u|^{k}$. By Sobolev imbedding we know that $W^{2,2}(B)\hookrightarrow L^{k}$ for any $k\geq 1$. So we know that $$\parallel u\parallel_{L^{k}(B)}\leq C_{k} \parallel u \parallel_{W^{2,2}} $$ We can then get : $\int_{B} e^{pu}\leq \sum_{k=1}^{\infty}\frac{p^{k}}{k!} (C_{k})^{k} \parallel u \parallel_{W^{2,2}}^{k}$. So that it suffices to control $C_{k}$. But based on my calculation, $C_{k} \sim k$, from which it seems this argument would fail.
After carefully thinking about what I need, I get the following.:
Lemma2: Suppose $u\in W^{2,2}(B(r_{0}))$ where $B_{r_{0}}$ is a ball with radius Note $r_{0}$ in the 4 dimensional Euclidean space, then$\int_{B} e^{pu}\leq \sum_{k=1}^{\infty}\frac{p^{k}}{k!}\int_{B} |u|^{k}$. By Sobolev imbedding we know that $e^{u}\in L^{q}(B_{\frac{r_{0}}{2}})$$W^{2,2}(B)\hookrightarrow L^{k}$ for any $q\geq 1$$k\geq 1$. So we know that
Proof: by multiplying a cut off function$$\parallel u\parallel_{L^{k}(B)}\leq C_{k} \parallel u \parallel_{W^{2,2}} $$
We can then get $\eta$ which is equal:
$\int_{B} e^{pu}\leq \sum_{k=1}^{\infty}\frac{p^{k}}{k!} (C_{k})^{k} \parallel u \parallel_{W^{2,2}}^{k}$.
So that it suffices to 1 oncontrol $B_{\frac{r_{0}}{2}}$, we can apply the theorem 7$C_{k}$.15 But based on Gilbert&Trudinger's bookmy calculation, i.e. Elliptic partial differential equations of second order (which can be also found on the link provided by Pedro Lauridsen Ribeiro in the comment ) to get $$ \int_{B_{r_{0}}}e^{(\frac{\eta u}{C_{1}||D^{1}(\eta |u|)||^{4}})^{\frac{4}{3}}}\leq C_{2}|B_{r_{0}}| $$ which then help control any $||e^{u}||_{L^{q}(B_{\frac{r_{0}}{2}})}$ for any $q\geq 1$$C_{k} \sim k$, from which it seems this argument would fail.