1: \begin{abstract}

2: 

3: Coupled-cluster theory with single and double excitations (CCSD) is a promising

4: \textit{ab initio} method for the electronic structure of three-dimensional metals, for which

5: second-order perturbation theory (MP2) diverges in the thermodynamic limit.

6: However, due to the high cost and poor convergence of CCSD with respect to basis size,

7: applying CCSD to periodic systems often leads to large basis set errors.

8: In a common ``composite'' method, MP2 is used to recover the missing dynamical correlation energy

9: through a focal-point correction,

10: but the inadequacy of MP2 for metals raises questions about this approach.

11: Here we describe how high-energy excitations treated by MP2 can be ``downfolded'' into a low-energy

12: active space to be treated by CCSD.

13: Comparing how the composite and downfolding approaches perform for the uniform electron

14: gas, we find that the latter converges more quickly with respect to the basis set size.

15: Nonetheless, the composite approach is surprisingly accurate because

16: it removes the problematic MP2 treatment of double excitations near the Fermi surface.

17: Using the method to estimate the CCSD correlation energy in the combined complete basis set and thermodynamic

18: limits, we find CCSD recovers over 90\% of the exact correlation energy at $r_s=4$.

19: We also test the composite and downfolding approaches with the random-phase approximation used in place of MP2,

20: yielding a method that is more effective but more expensive. 

21: 

22: \end{abstract}

23: