[Paper Review] Quantifying Feedback from Narrow Line Region Outflows in Nearby Active Galaxies. IV. The Effects of Different Density Estimates on the Ionized Gas Masses and Outflow Rates

Mitchell Revalski, D. M. Crenshaw+10 more|arXiv (Cornell University)|Mar 14, 2022

Galaxies: Formation, Evolution, Phenomena73 references37 citations

TL;DR

This study evaluates simplified methods for estimating ionized gas masses and outflow rates in nearby active galactic nuclei (AGN) by comparing them to high-precision multi-component photoionization models. It finds that single-component photoionization models with radius-dependent ionization parameters (log U) derived from [O III]/Hβ ratios provide the most reliable estimates—within ~3× (±0.5 dex) of the benchmark models—making them suitable for large-scale surveys with limited spectral data.

ABSTRACT

Active galactic nuclei (AGN) can launch outflows of ionized gas that may influence galaxy evolution, and quantifying their full impact requires spatially resolved measurements of the gas masses, velocities, and radial extents. We previously reported these quantities for the ionized narrow-line region (NLR) outflows in six low-redshift AGN, where the gas velocities and extents were determined from Hubble Space Telescope long-slit spectroscopy. However, calculating the gas masses required multi-component photoionization models to account for radial variations in the gas densities, which span $\sim$6 orders of magnitude. In order to simplify this method for larger samples with less spectral coverage, we compare these gas masses with those calculated from techniques in the literature. First, we use a recombination equation with three different estimates for the radial density profiles. These include constant densities, those derived from [S II], and power-law profiles based on constant values of the ionization parameter ($U$). Second, we use single-component photoionization models with power-law density profiles based on constant $U$, and allow $U$ to vary with radius based on the [O III]/H$\beta$ ratios. We find that assuming a constant density of $n_\mathrm{H} =$ 10$^2$ cm$^{-3}$ overestimates the gas masses for all six outflows, particularly at small radii where the outflow rates peak. The use of [S II] marginally matches the total gas masses, but also overestimates at small radii. Overall, single-component photoionization models where $U$ varies with radius are able to best match the gas mass and outflow rate profiles when there are insufficient emission lines to construct detailed models.

Motivation & Objective

• To assess the reliability of simplified techniques for estimating ionized gas masses and outflow rates in AGN narrow-line regions when high-resolution, multi-line spectroscopy is unavailable.
• To identify systematic biases introduced by common assumptions such as constant density or [S II]-derived densities in outflow mass calculations.
• To determine whether single-component photoionization models with variable ionization parameters can accurately reproduce results from complex, multi-component models.
• To guide future surveys by identifying methods that balance computational efficiency with acceptable accuracy for large samples of AGN.

Proposed method

• Compare gas masses and outflow rates derived from multi-component photoionization models (Revalski et al. 2021) with those from simplified techniques.
• Use a recombination equation (M ∝ L / nH) with three density estimates: constant nH = 10² cm⁻³, [S II] doublet-derived densities, and power-law density profiles based on constant ionization parameter (U).
• Apply single-component photoionization models with both constant and radius-variable log(U), where variable log(U) is derived from observed [O III]/Hβ flux ratios.
• Use the photoionization code Cloudy to compute ionized gas properties under different density and ionization assumptions.
• Validate results against benchmark multi-component models by comparing radial profiles of gas mass and outflow rate across six nearby AGN (Mrk 3, Mrk 34, Mrk 78, Mrk 573, NGC 1068, NGC 4151).
• Quantify discrepancies in mass and outflow rate estimates using dex-based comparisons to assess systematic biases.

Experimental results

Research questions

• RQ1 How do constant-density assumptions (nH = 10² cm⁻³) affect the accuracy of ionized gas mass and outflow rate estimates in AGN NLRs?
• RQ2 To what extent do [S II] doublet-based electron density estimates reproduce the true gas mass profiles derived from multi-component photoionization models?
• RQ3 Can single-component photoionization models with radius-dependent ionization parameters (log U) accurately reproduce the gas mass and outflow rate profiles when multi-component modeling is not feasible?
• RQ4 What are the systematic biases introduced by using spatially integrated spectra instead of spatially resolved data for outflow energetics?
• RQ5 How do assumptions about density and ionization structure impact the reliability of outflow rate estimates in low-S/N or limited-wavelength data sets?

Key findings

• Assuming a constant density of nH = 10² cm⁻³ overestimates gas masses and outflow rates across all six AGN, particularly at small radii where outflow rates peak.
• Using [S II]-derived densities marginally reproduces total gas masses and peak outflow rates within ±0.8 dex, but still overestimates masses at small radii due to tracing partially ionized zones.
• Recombination-based mass estimates with power-law density profiles assuming constant U show significant disagreement when optically thick and thin components coexist or high-ionization gas is present.
• Global outflow rates derived from spatially integrated spectra are strongly biased toward low-density, large-radius emission, leading to overestimation of inner-region masses and energetics.
• Single-component photoionization models with radius-dependent log(U) derived from [O III]/Hβ ratios provide the most accurate approximation, with uncertainties within ~3× (±0.5 dex) of the benchmark multi-component models in most cases.
• The best-performing simplified method—single-component models with variable log(U)—achieves acceptable accuracy (up to ~5× or ±0.7 dex at some radii) and is recommended for large-sample studies with limited spectral coverage.