New techniques to investigate the AGN-SF connection with integral field spectroscopy

This paper introduces a robust, automated method using Integral Field Spectroscopy to disentangle AGN and star-formation contributions in local galaxies, revealing a moderate correlation between AGN accretion rates and recent nuclear star formation that supports the co-evolution of these processes.

The Gist

The Big Mystery: The Black Hole and the Baby Stars

Imagine a galaxy as a bustling city. In the very center of this city lives a Supermassive Black Hole (the AGN). It's like a giant, hungry vacuum cleaner that eats gas and dust. When it eats, it glows incredibly bright.

Surrounding this central vacuum cleaner are Star-Forming Regions (SF). These are like nurseries where new stars are being born.

For decades, astronomers have been arguing about the relationship between the hungry black hole and the baby stars. Does the black hole's feeding frenzy help the stars grow? Or does it blow the gas away and stop the stars from forming? This is the "AGN-SF connection."

The Problem: A Blurry Photo

To solve this mystery, astronomers usually look at the light coming from the galaxy. But for a long time, they only had "single-lens" cameras (like old fiber-optic telescopes).

Imagine trying to study a city by looking at it through a single, tiny straw. You can see the total light, but you can't tell if the light is coming from the busy factory in the center (the black hole) or the construction site on the edge (the stars). The light from both gets mixed together, making it impossible to tell who is doing what.

The New Tool: A High-Definition Map

This paper introduces a new way to look at 54 nearby galaxies using Integral Field Spectroscopy (IFU).

Think of IFU not as a single straw, but as a high-definition mosaic. Instead of one blurry dot, the telescope breaks the galaxy into thousands of tiny pixels (like a digital photo). Now, astronomers can look at the center, the edges, and every spot in between individually.

The New Trick: The "Mixing Sequence" Recipe

The core of this paper is a new mathematical recipe to separate the "Black Hole Light" from the "Star Light."

1. The BPT Diagram (The Flavor Chart): Astronomers use a special chart (called a BPT diagram) to identify what kind of gas is glowing.

   • Pure Star Light is like a "Vanilla" flavor.
   • Pure Black Hole Light is like a "Spicy Chili" flavor.
   • Mixed Light is somewhere in between, like "Vanilla with a hint of Chili."

2. The Old Method (The Flawed Chef): Previous methods tried to draw a straight line between the "Vanilla" and "Chili" points on the chart to guess the mixture. But if the data was messy or had outliers (like a chef accidentally dropping a rock in the soup), the line would be wrong.

3. The New Method (The Smart Sorter): The authors (Chopra, Zovaro, and Davies) invented a new way to find the "Vanilla" and "Chili" points.

   • They use a statistical tool called the Mahalanobis distance. Think of this as a smart sorter that ignores the "rock in the soup" (outliers) and finds the truest "Vanilla" and "Chili" samples, even if the data is messy or curved.
   • Once they know what pure Vanilla and pure Chili look like, they can mathematically "un-mix" every single pixel in the galaxy. They can say, "This specific spot is 80% stars and 20% black hole."

What They Found: The Timing Game

Once they separated the light, they asked: When does the black hole get hungry relative to the star birth?

They looked at the "age" of the stars in the galaxy (specifically, how many young stars were born in the last 100 million years).

• The Result: They found a moderate link. When the black hole is eating at its fastest (high "Eddington ratio"), there are often young stars nearby that were born very recently (within the last 100 million years).
• The Analogy: It's like a party. The black hole and the star factories seem to be throwing their parties at the same time. The gas that fuels the baby stars also seems to be the gas that feeds the black hole.

However, the connection isn't perfect. It's a "weak" link. Sometimes the black hole is eating, but the stars aren't born yet, or vice versa. This suggests that while they are related, other factors (like the black hole blowing gas away) might be interfering.

Why This Matters

This paper is a "how-to" guide for the future.

• Robustness: Their new method works even when the data is messy, which is great because real astronomical data is rarely perfect.
• Scalability: Because the method is automated and doesn't require a human to manually tweak every galaxy, it can be used on massive surveys containing thousands of galaxies (like the upcoming SDSS-V project).

The Bottom Line

The authors built a better "un-mixer" for galaxy light. Using this tool, they confirmed that black holes and new stars often grow together in the same neighborhood, likely sharing the same fuel supply. However, the relationship is complex, and to fully understand the physics, we need to look at even more galaxies in the future.

In short: They figured out how to separate the noise from the signal in a crowded room, proving that the "monster" in the center and the "babies" on the outside are often part of the same family event.

Technical Summary

1. Problem Statement

The relationship between Active Galactic Nuclei (AGN) and Star Formation (SF) is a fundamental but unresolved issue in astrophysics. While cosmological simulations suggest AGN feedback regulates star formation, observational evidence is often conflicting.

• Limitations of Previous Studies: Prior research largely relied on single-fiber spectroscopy (e.g., SDSS), which integrates light over a fixed aperture (~1 kpc). This approach suffers from:
  • Spatial Mixing: Inability to distinguish between nuclear AGN activity and circumnuclear star formation.
  • Systematic Uncertainties: Converting line ratios to AGN fractions depends heavily on gas metallicity, ionization parameters, and the hardness of the radiation field, which vary spatially within galaxies.
  • Contamination: Difficulty in separating AGN ionization from shocks or Hot Low-Mass Evolved Stars (HOLMES).
• Goal: To develop a robust, spatially resolved method to decompose AGN and SF contributions in emission lines, allowing for accurate calculation of AGN accretion rates and Star Formation Rates (SFRs) to test the AGN-SF connection.

2. Methodology

Data Sample

• Survey: Siding Spring Southern Seyfert Spectroscopic Snapshot Survey (S7) Data Release 2.
• Instrument: Wide Field Spectrograph (WiFeS) on the ANU 2.3m telescope.
• Sample: 54 local active galaxies ($z < 0.02$).
• Selection Criteria: Galaxies were categorized into "clean" (25), "ambiguous" (13), and "AGN-dominated" (16) based on their distribution in optical diagnostic diagrams (BPT and WHAN). "Unsuitable" galaxies (77) were excluded due to shock dominance or lack of mixing sequences.

Novel Decomposition Method (BPT Mixing Sequences)

The authors introduce a new, non-parametric method to separate AGN and SF contributions using the [NII]/H$\alpha$ vs. [OIII]/H$\beta$ BPT diagram mixing sequence.

1. Outlier Removal: Uses the Mahalanobis distance to identify and remove outliers in the multidimensional line ratio space, accounting for correlations between variables. This is more robust than Euclidean distance.
2. Basis Spectra Calculation: Instead of fitting a linear line (which is sensitive to scatter and uneven data distribution), the method bins data in $\log_{10}([NII]/H\alpha) + \log_{10}([OIII]/H\beta)$ space into 10 equal-width bins.
   • The SF Basis Spectrum is derived from the average of the lowest bin.
   • The AGN Basis Spectrum is derived from the average of the highest bin.
3. Linear Decomposition: For each spaxel, the observed line luminosity ($L_i$) is modeled as a linear combination of the basis spectra:
   $$L_i = m_{AGN} \times L_{i, AGN} + n_{SF} \times L_{i, SF}$$
   Superposition coefficients ($m, n$) are solved using the Levenberg-Marquardt algorithm (LMFIT) on extinction-corrected line fluxes (H$\alpha$, [NII], [SII], [OIII]).

Stellar Population Analysis (PPXF)

• Tool: Penalised PiXel Fitting (PPXF) code.
• Setup: Fits the stellar continuum using 222 Simple Stellar Population (SSP) templates (González Delgado et al. 2005) covering ages $10^6$–$10^{10}$ yr and three metallicities.
• AGN Continuum Handling: Includes power-law templates ($F_\nu \propto \nu^{-\alpha_\nu}$) to model and subtract AGN continuum contamination.
• Recovery Tests: Extensive Monte Carlo (MC) simulations were performed to quantify systematic errors.
  • Result: Light-weighted (LW) ages have systematic uncertainties of ~0.2 dex (100 Myr cutoff) and ~0.1 dex (1 Gyr cutoff) without AGN contamination, increasing to ~0.4 dex with strong AGN continuum.
• Apertures: Analysis performed in three apertures: 1 kpc (constant physical scale), 1 $R_e$ (effective radius), and the S7 aperture (approx. SDSS fiber size).

Correlation Analysis

• AGN Proxy: Eddington ratio proxy defined as $L_{[OIII]}/\sigma_*^4$, where $L_{[OIII]}$ is the total AGN [OIII] luminosity (summed over the full field of view) and $\sigma_*$ is the stellar velocity dispersion.
• SFR: Calculated from the decomposed H$\alpha$ flux using the Kennicutt (1998) relation.
• Statistical Test: Spearman rank-order correlation coefficients calculated using Monte Carlo sampling (10,000 iterations) to account for measurement uncertainties and the biased nature of the radio-selected sample.

3. Key Contributions

1. Robust Decomposition Algorithm: A new, automated method for calculating AGN/SF basis spectra that is resilient to outliers, varying mixing sequence shapes, and metallicity gradients. It requires minimal manual intervention, making it scalable for large IFU surveys (e.g., MaNGA, SAMI, HECTOR).
2. Spatially Resolved AGN-SF Separation: Successfully separates AGN and SF contributions on a spaxel-by-spaxel basis, providing more accurate global SFRs and AGN luminosities than single-aperture methods.
3. Rigorous Error Quantification: Detailed recovery tests using mock spectra to define the reliability limits of stellar age measurements in the presence of AGN continuum and extinction.
4. Multi-Aperture Analysis: Systematic comparison of correlations across different physical scales (nuclear vs. global) to isolate the regions where AGN-SF coupling is strongest.

4. Results

• AGN-SFR Correlation: A moderately strong positive correlation was found between the AGN Eddington ratio and the Star Formation Rate (SFR) across all apertures.
• AGN-Stellar Age Correlation:
  • A statistically significant ($p < 0.01$) stronger correlation exists between the Eddington ratio and the light-weighted stellar age under 100 Myr compared to ages under 1 Gyr (specifically in 1 kpc and 1 $R_e$ apertures).
  • This suggests AGN activity is linked to very recent nuclear star formation rather than older populations.
• Aperture Dependence: Correlations are stronger in the smaller 1 kpc and 1 $R_e$ apertures compared to the larger S7 aperture. This implies the AGN-SF connection is most tightly coupled in the nuclear regions where gas inflow occurs.
• Timescales: The results support a timeline where AGN accretion peaks roughly 50–250 Myr after the onset of a starburst, consistent with gas inflow models.
• Fueling Mechanism: Calculations suggest stellar mass loss from massive stars is insufficient (by an order of magnitude) to fuel the observed AGN accretion rates, implying external gas inflow is the primary driver.

5. Significance

• Physical Insight: The findings support the "co-evolution" scenario where AGN and star formation are fueled by the same reservoir of cold gas, with AGN activity typically lagging slightly behind the peak of nuclear star formation.
• Methodological Advancement: The proposed decomposition technique overcomes the limitations of previous empirical fitting methods, allowing for the analysis of complex, scattered mixing sequences common in large IFU surveys.
• Future Applications: The method is designed to be applied to upcoming large-scale surveys (e.g., SDSS-V Local Volume Mapper, CARS) to better constrain the physical mechanisms and timescales of AGN feedback and fueling, particularly by including radio-quiet galaxies which were underrepresented in this radio-selected sample.

Conclusion: The paper establishes a robust framework for disentangling AGN and star formation in integral field data, confirming a weak but significant link between high AGN accretion rates and very young nuclear stellar populations, consistent with a shared gas fueling mechanism.