1: \begin{abstract}

2: {Let $W_N(\b) = \DE_0\left[e^{ \sum_{n=1}^N \b\omega(n,S_n) - N\b^2/2}\right]$ be the partition function of a two-dimensional directed polymer in a random environment,

3: where $\omega(i,x), i\in \mathbb{N}, x\in \mathbb{Z}^2$ are i.i.d.\ standard normal and $\{S_n\}$ is the path of a random walk. With $\beta=\beta_N=\hat\b \sqrt{\pi/\log N}$ and $\hat \b\in (0,1)$ (the subcritical window),

4: $\log W_N(\b_N)$ is known to converge in distribution to a Gaussian  law of mean $-\lambda^2/2$ and variance $\lambda^2$, with $\lambda^2=\log (1/(1-\hat\b^2))$

5:  (\textit{Caravenna,  Sun, Zygouras, Ann.\ Appl.\ Probab.\ (2017)}).

6: We study in this paper

7: the moments $\IE [W_N( \b_N)^q]$ in the subcritical window, and prove a lower bound that matches for $q=O(\sqrt{\log N})$ the upper bound derived by 

8: us in \textit{Cosco, Zeitouni, arXiv:2112.03767}. 

9: The analysis is based on appropriate decouplings and a Poisson convergence

10: that uses the method of ``two moments suffice''.}

11: \end{abstract}

12: