1: \begin{abstract}

2: Approximate full configuration interaction (FCI) calculations have

3: recently become tractable for systems of unforeseen size thanks to

4: stochastic and adaptive approximations to the exponentially scaling

5: FCI problem.  The result of an FCI calculation is a weighted set of

6: electronic configurations, which can also be expressed in terms of

7: excitations from a reference configuration.  The excitation amplitudes

8: contain information on the complexity of the electronic wave function,

9: but this information is contaminated by contributions from

10: disconnected excitations, i.e. those excitations that are just

11: products of independent lower-level excitations. The unwanted

12: contributions can be removed via a cluster decomposition procedure,

13: making it possible to examine the importance of connected excitations

14: in complicated multireference molecules which are outside the reach of

15: conventional algorithms. We present an implementation of the cluster

16: decomposition analysis and apply it to both true FCI wave functions,

17: as well as wave functions generated from the adaptive sampling CI

18: (ASCI) algorithm. The cluster decomposition is useful for interpreting

19: calculations in chemical studies, as a diagnostic for the convergence

20: of various excitation manifolds, as well as as a guidepost for

21: polynomially scaling electronic structure models. Applications are

22: presented for (i) the double dissociation of water, (ii) the carbon

23: dimer, (iii) the $\pi$ space of polyacenes, as well as (iv) the

24: chromium dimer. While the cluster amplitudes exhibit rapid decay with

25: increasing rank for the first three systems, even connected octuple

26: excitations still appear important in \ce{Cr2}, suggesting that

27: spin-restricted single-reference coupled-cluster approaches may not be

28: tractable for some problems in transition metal chemistry.

29: \end{abstract}

30: