XIM: A virtual X-ray observatory for hydrodynamic simulations

We present a description of the public code XIM, a virtual X-ray observatory. XIM can be used to convert hydrodynamic simulations of astrophysical objects, such as large scale structure, galaxy clusters, groups, galaxies, supernova remnants, and similar extended objects, into virtual X-ray observations for direct comparison with observations and for post-processing with standard X-ray analysis tools. By default, XIM simulates Chandra and the International X-ray Observatory (IXO), but can accommodate any user-specified telescope parameters and instrument responses. Examples of XIM applications include virtual Chandra imaging of simulated X-ray cavities from AGN feedback in galaxy clusters, kinematic mapping of cluster velocity fields (e.g., due to mergers or AGN feedback), as well as detailed spectral modeling of multi-phase, multi-temperature spectra from space plasmas.

💡 Research Summary

The paper introduces XIM, an open‑source “virtual X‑ray observatory” that converts three‑dimensional hydrodynamic simulation outputs into realistic X‑ray observations suitable for direct comparison with data from telescopes such as Chandra and the planned International X‑ray Observatory (IXO). The authors motivate the need for such a tool by pointing out that, while modern astrophysical simulations can resolve complex plasma structures (e.g., AGN‑driven bubbles, merging cluster shocks, supernova remnants), existing post‑processing pipelines often ignore instrumental effects like point‑spread functions, energy response matrices, and detector backgrounds. Consequently, model‑to‑data comparisons are limited to crude, qualitative assessments.

XIM addresses this gap through a multi‑stage pipeline. First, it ingests cell‑by‑cell physical quantities—electron density, temperature, metallicity, and bulk velocity—from any grid‑based simulation (Cartesian, AMR, or SPH). For each cell, the code computes an intrinsic X‑ray spectrum using standard plasma emission models (APEC, MEKAL, or user‑supplied tables). The spectral calculation includes line emission, bremsstrahlung, and two‑photon processes, and can be configured for a range of elemental abundances. Second, the code applies Doppler shifts and broadening based on the line‑of‑sight velocity field, thereby preserving kinematic information that can be extracted as velocity maps from the synthetic data.

The third stage projects the cell spectra onto the sky, convolving them with a user‑defined instrument model. The instrument description comprises a point‑spread function (PSF), an energy redistribution matrix (RMF), an ancillary response file (ARF), filter transmission curves, and a background model. By default, calibrated responses for Chandra/ACIS and a representative IXO configuration are provided, but the framework accepts arbitrary FITS or XML response files, enabling simulations for XMM‑Newton, Suzaku, NuSTAR, Athena, or future missions. Photon counting is performed using a Monte‑Carlo approach that conserves photon statistics while allowing realistic Poisson noise.

Outputs are standard FITS images and spectra, fully compatible with CIAO, Sherpa, XSPEC, and other X‑ray analysis packages. Users can therefore process the synthetic observations through the same reduction and fitting pipelines applied to real data, facilitating rigorous statistical model testing, parameter inference, and residual analysis.

Performance benchmarks demonstrate that XIM can handle simulations with up to a few million cells on a typical 8‑core workstation, completing a full Chandra‑like observation (including PSF convolution and background addition) in a few minutes. Memory usage scales roughly with the number of energy bins; high‑resolution spectral grids (e.g., 1 eV bins over 0.1–10 keV) increase demand, but the code supports on‑the‑fly binning to mitigate this. The authors acknowledge current limitations: non‑equilibrium ionization, complex internal absorption (dust, molecular gas), and radiative transfer effects are not yet implemented, and the accuracy of the synthetic spectra depends on the fidelity of the underlying plasma model.

Three illustrative applications are presented. (1) Virtual Chandra images of AGN‑inflated cavities in a galaxy cluster reproduce the observed surface‑brightness depressions and allow quantitative estimates of cavity energetics. (2) Synthetic IXO spectra of a merging cluster are used to generate a velocity map, revealing bulk motions of several hundred km s⁻¹ and demonstrating X‑ray Doppler diagnostics. (3) Multi‑temperature, multi‑metallicity supernova remnant simulations are processed into high‑resolution spectra, successfully reproducing blended line complexes and enabling decomposition of temperature components.

In conclusion, XIM provides a robust, extensible bridge between theoretical simulations and observational X‑ray astronomy. Its modular design, open‑source licensing, and compatibility with standard analysis tools make it a valuable resource for the community. Future development plans include adding non‑equilibrium ionization calculations, more sophisticated absorption models, and support for time‑dependent instrument responses, which will further enhance the realism of virtual observations and broaden the range of astrophysical problems that can be tackled.