Identifying and Distinguishing Quenching Galaxies with Spatially Resolved Star Formation in the Hubble Frontier Fields

Mapping the Death of Star Formation

In the grand narrative of cosmic evolution, galaxies are not static islands of light. They are dynamic engines that eventually run out of fuel. This process, known as quenching, describes how a galaxy transitions from the "blue cloud"—a population of vibrant, star-forming systems—to the "red sequence" of quiescent, aging stellar populations. Astronomers have long sought to understand why this happens. They debate whether the cause is internal, such as feedback from a supermassive black hole, or external, such as the harsh environment of a massive galaxy cluster.

Until now, the debate has lacked a clear view of the spatial progression of this shutdown. We knew galaxies quenched, but we did not know if they died from the "inside-out," like a candle burning down its wick, or "outside-in," like a frost creeping across a windowpane. This paper provides that missing spatial dimension by analyzing the Hubble Frontier Fields. It reveals that quenching is not a monolithic event. Instead, it is a bifurcated process dictated by mass and environment.

The Spatial Blind Spot in Quenching Models

The fundamental difficulty in studying galaxy quenching lies in the mismatch between global averages and local physics. Traditionally, astronomers have treated galaxies as single points of data. They measure an integrated Star Formation Rate (SFR)—the total mass of new stars produced per year—and a total stellar mass. While these numbers tell us if a galaxy is dying, they hide the how.

Current models struggle to reconcile the diverse ways star formation ceases. Internal mechanisms, such as Active Galactic Nucleus (AGN) feedback (energy injected into the interstellar medium by a central black hole), are theorized to suppress star formation from the center outward. Conversely, environmental effects like ram pressure stripping (the removal of a galaxy's gas by the pressure of the surrounding intra-cluster medium) are expected to truncate star formation from the outskirts inward. Because standard observations often lack the resolution to map these processes across the galaxy's radius, the field has been stuck in a stalemate. Researchers have been unable to definitively categorize which mechanism dominates in different cosmic habitats.

Decoding the Radial Signature

To move beyond global averages, the authors implement a highly granular method centered on spatially resolved Spectral Energy Distribution (SED) fitting. They decompose each galaxy into 20 concentric circular annuli (ring-like segments). These rings extend out to five times the effective radius ($R_e$, the radius containing half the galaxy's total light).

The workflow follows a rigorous three-stage pipeline:

1. Annular Photometry: The researchers extract light from each of the 20 rings across multiple Hubble filters. This allows them to build a radial profile of both stellar mass and star formation.
2. Spatially Resolved SED Fitting: Using the FAST++ code, they fit these ring-specific light profiles to stellar population models. This determines the specific star formation rate (sSFR)—the star formation rate per unit of stellar mass—for every segment of the galaxy. As shown in, this reveals distinct radial "shapes" for different quenching modes.

3. Machine Learning Classification: The researchers utilize a $k$-nearest neighbors (kNN) algorithm. This algorithm is trained on morphological metrics derived from the TNG50 (IllustrisTNG) simulations. These metrics include the concentration of star formation ($C_{SF}$) and the specific radii where star formation abruptly drops off ($R_{inner}$ and $R_{outer}$).

By feeding these structural descriptors into the kNN, the authors can objectively classify a real galaxy into an "inside-out" or "outside-in" pathway. This classification is based on how its star formation is distributed spatially.

Two Paths to Silence

The application of this pipeline to 1,437 galaxies in the Hubble Frontier Fields yields a stark divergence in evolutionary trajectories. The authors identify 129 galaxies following an inside-out pathway and 70 following an outside-in pathway.

The most significant finding is the mass and environment dependency of these modes. The authors report that inside-out quenching galaxies are significantly more massive. They are more massive by $0.8^{+0.2}_{-0.1}$ dex (roughly a six-fold increase in mass) than their outside-in counterparts. Furthermore, the inside-out population is overwhelmingly found in clusters (106 out of 129). In contrast, the outside-in population is more common in the field. As demonstrated in, these galaxies primarily inhabit the "green valley," the transitional zone between active star formation and total quiescence.

The prevalence of these modes also scales with mass. In cluster environments, the fraction of inside-out quenching galaxies increases sharply with stellar mass. This population represents approximately 30% of the non-quenched population at the highest mass scales. Conversely, the outside-in fraction remains relatively constant and low ($\lesssim$10%) regardless of mass or environment, as seen in . This suggests that while environmental stripping is a real phenomenon, it is a secondary player in the overall census of quenching compared to the massive, internally-driven inside-out process.

Constraints on the Model

While the study provides a robust framework for classification, it is not without limitations. First, the methodology is computationally expensive and sensitive to signal-to-noise ratios. To ensure reliable radial profiles, the authors imposed a strict H-band signal-to-noise threshold. This resulted in the loss of approximately 45% of their initial candidate sample. This highlights a significant trade-off. The pursuit of spatial detail necessitates a much smaller, higher-quality sample. This may bias results toward brighter, more easily resolved objects.

Second, the reliance on the TNG50 simulation for the kNN training set introduces a layer of "model dependency." The authors acknowledge that simulations may overstate the strength of AGN feedback. They may also overestimate the efficiency of environmental gas removal. Although they mitigated this by including the distance from the star-forming main sequence ($\Delta\text{SFMS}$) as a classification feature, the possibility remains that the morphological boundaries learned from simulations might not perfectly match real-world physics. Finally, the authors note they were unable to use kNN regression to predict the progress of quenching. They could only identify its mode. This leaves the exact timing of the transition still partially obscured.

A New Standard for Large-Scale Surveys

The verdict is clear: spatially resolved morphology is the essential metric for the next era of extragalactic astronomy. This work proves that the morphological signatures of quenching are not just theoretical constructs from simulations. They are observable realities in the Hubble Frontier Fields. The ability to distinguish between inside-out and outside-in pathways allows us to move from asking if a galaxy is quenching to asking why.

The methodology is primed for the massive data volumes expected from upcoming missions. The authors explicitly point to the Nancy Grace Roman Space Telescope and the proposed CASTOR mission as ideal candidates to scale this approach. While this paper identifies the pathways, the next frontier will be using these tools to map the precise timescales of the transition. This will turn a snapshot of a "dying" galaxy into a cinematic reconstruction of its demise.

Figures from the paper