1: \begin{abstract}

2: We extend finite-temperature tensor network methods to compute Matsubara imaginary-time correlation functions,

3: building on the minimally entangled typical thermal states (METTS) and purification algorithms. 

4: While imaginary-time correlation functions are straightforward to formulate with these methods, 

5: care is needed to avoid convergence issues that would result from naive estimators.

6: As a benchmark, we study the single-band Anderson impurity model, 

7: even though the algorithm is quite general and applies to lattice models.

8: The special structure of the impurity model benchmark system and our choice of basis enable 

9: techniques such as reuse of high-probability METTS for increasing algorithm efficiency. 

10: The results are competitive with state-of-the-art continuous time Monte Carlo.

11: We discuss the behavior of computation time and error as a function of the number of purified sites in the Hamiltonian.

12: 

13: \end{abstract}

14: