The 2017 Release of Cloudy

TL;DR

Cloudy C17 introduces a major shift to external atomic data databases, enabling vastly larger and more accurate line catalogs while offering a default subset optimized for speed and memory. The release preserves Cloudy’s NLTE, collisional-radiative framework for H-like and He-like ions, while applying a scalable two-level approximation for more complex species, and implements isotropic-radiation corrections, time-dependent cooling, and non-equilibrium dynamics. It expands H2 and grain chemistry, PAH treatments, and LAMDA/CHIANTI/Stout data integration, with updated Gaunt factors, H2 formation on grains, updated LAMDA molecular data, and refined Lyman-pumping handling in PDRs. Practical gains include dramatically more emission lines predicted by default, improved convergence and runtime performance, and enhanced capabilities for PDRs, X-ray plasmas, and time-dependent cooling, making Cloudy more versatile for modeling diverse astrophysical environments. The combination of large, maintainable data repositories, improved reporting, and multi-core grid execution positions Cloudy as a robust, scalable tool for spectroscopic predictions across ISM, PDR, and AGN contexts.

Abstract

We describe the 2017 release of the spectral synthesis code Cloudy. A major development since the previous release has been exporting the atomic data into external data files. This greatly simplifies updates and maintenance of the data. Many large datasets have been incorporated with the result that we can now predict well over an order of magnitude more emission lines when all databases are fully used. The use of such large datasets is not realistic for most calculations due to the time and memory needs, and we describe the limited subset of data we use by default. Despite the fact that we now predict significantly more lines than the previous Cloudy release, this version is faster because of optimization of memory access patterns and other tuning. The size and use of the databases can easily be adjusted in the command-line interface. We give examples of the accuracy limits using small models, and the performance requirements of large complete models. We summarize several advances in the H- and He-like iso-electronic sequences. We use our complete collisional-radiative models of the ionization of these one and two-electron ions to establish the highest density for which the coronal or interstellar medium (ISM) approximation works, and the lowest density where Saha or local thermodynamic equilibrium can be assumed. The coronal approximation fails at surprisingly low densities for collisional ionization equilibrium but is valid to higher densities for photoionized gas clouds. Many other improvements to the physics have been made and are described. These include the treatment of isotropic continuum sources such as the cosmic microwave background (CMB) in the reported output, and the ability to follow the evolution of cooling non-equilibrium clouds.

Figures (28)

• Figure 1: This compares the number of lines that fall into 1000 ${\rm km}\hbox{${\rm s}^{-1}\,$}\,$ velocity bins in C13 (black) and the default C17 setup (red).
• Figure 2: This shows the number of lines that fall into 1000 km/s velocity bins, across the spectrum. The red points for default setup and the black points give the number of lines the are predicted when the databases are made as large as possible. The upper panel shows the full spectral range considered by the code, while the lower panel shows the peak of the line density.
• Figure 3: Experimental energy levels NIST_ASD for some species present in an ionized gas. The energies are given relative to the ionization potential (IP). Of these ions, only O III and Mg II have data for autoionizing levels, shown as the levels above the ionization limit indicated by the red hashed box. The autoionizing levels of Mg II are not visible since they are far above the ionization limit. The vertical bar in the middle, corresponding to E/IP = 0.05, is a typical gas kinetic energy in a photoionized plasma and is shown to indicate which levels are energetically accessible from the ground state.
• Figure 4: Ratios of He I lines using the different datasets with respect to PS-M, see text for details. Figure from Guzman.II.2017.
• Figure 5: This shows ratios of our predicted H i emission to the 1995MNRAS.272...41S Case B tables. Calculations are for $T_e = 10^4$ K and the indicated densities. The upper panel, our test case limit_caseb_h_den4_temp4.in., shows that we reproduce their results to high accuracy when a large model is used. The default model, chosen as a compromise between speed and accuracy, is shown in the lower two panels. The default model is designed to give higher accuracy for the brighter optical and near-IR lines, plotted as the larger filled circles. Note that each panel has a different vertical scale.