MAGNUS I: A MUSE-DEEP Sample of Early-type Galaxies at Intermediate Redshift

We present a sample of 212 early-type galaxies (ETGs) at redshifts 0.25 < z < 0.75. We combine deep integralfield spectroscopy from the MUSE-DEEP survey with high-resolution HST imaging to study the structure, kinematics, and stellar populations of these galaxies. We measure spatially resolved stellar kinematics and use the specific angular momentum proxy, λR, to classify galaxies into fast and slow rotators. We find a slow rotator fraction consistent with local Universe samples, suggesting little evolution in the massive ETG population since z ∼ 1. The kinematic and photometric axes of fast rotators are generally well-aligned, similar to their local counterparts. We find that global stellar population properties, such as age, metallicity, and mass-to-light ratio (M∗ /L), correlate strongly with the central velocity dispersion (σe), following trends established for local ETGs. Slow rotators are typically more massive, have higher σe, and are more metal-rich than fast rotators. Our findings indicate that the fundamental structural, kinematic, and stellar population scaling relations of massive ETGs were already in place by z ∼ 0.75, suggesting their evolutionary pathways have remained stable over the last ∼ 7 Gyr.

The MAGNUS sample consists of 212 ETGs at redshifts 0.25 < z < 0.75, combining deep integralfield spectroscopy from the MUSE-DEEP survey with high-resolution HST imaging to study their structure, kinematics, and stellar populations. The sample selection involved checking each datacube manually after initial inspection of 35 datacubes containing at least 4-6 ETGs at the target redshift, accompanied by publicly available HST imaging for detailed analysis. The galaxies were selected based on two criteria: first, a galaxy should be sufficiently isolated to ensure that the measured kinematics and photometric features would have minimum contamination from surrounding galaxies (no other galaxy within a circular annulus of width around 1-2′′ surrounding it); and second, the projected image of the galaxy in the sky plane must be spatially large enough (> 1′′ along the major axis) to extract resolved stellar kinematics up to a few effective radii. The sample contains a relatively higher number of galaxies in the redshift range 0.3–0.45, although other redshift bins also host comparable numbers of objects. The σe of the sample spans a range of approximately 100–300 km s−1 with a peak near 180 km s−1. In terms of surface brightness, the brightest galaxies in the sample reach values around 20 mag/arcsec2, while the faintest ones are down to 20.5 mag/arcsec2 in the HST F814W band. The largest galaxies in the sample have a semi-major axis of around 10 kpc, while the smallest ones exhibit Re values as low as 0.6 kpc. Most galaxies (∼ 80%) have a Re between 1 and 4 kpc.

The kinematic classification of ETGs is based on specific angular momentum, λR, to classify galaxies into fast and slow rotators. We find that regular rotators exhibit higher specific angular momentum (higher λR), while non-regular rotators tend to have lower values, both in the local Universe (M. Cappellari 2025) and at intermediate redshifts (C. Muñoz López et al. 2024). This suggests that the kinematic state of ETGs has remained broadly the same below z ∼ 1. Around 87% of galaxies that were qualitatively classified as non-regular rotators reside within the slow rotator region bordered by the black solid lines, defined by λRe < 0.08 + εe /4 and εe < 0.4 (M. Cappellari 2016, equation 19). According to quantitative classification, the slow rotator fraction in our sample is 19.3+3.0−2.4 % (41 out of 212). C. Derkenne et al. (2024) found that the slow rotator fraction in the MAGPI sample at z ∼ 0.3 is +6.9+6.6−4.1 % with seeing corrected λR and is 24−4.9 % without seeing corrected λR. The MAGPI sample is mass-matched to our sample, as in both samples the estimated stellar masses of the galaxies are M∗ > 3 × 1010 M⊙. They conducted a comparison with the local Universe ETGs by drawing a mass-matched and progenitor bias-corrected sample from the MaNGA survey (W. L. Freedman et al. 2020) 100 times.

As a fastidious and diligent researcher, I have analyzed the provided scientific paper, MAGNUS I: A MUSE-DEEP sample of early-type galaxies at intermediate redshift, focusing on its methodology, findings regarding galaxy evolution, and the constraints it places on galaxy properties.

Here are the specific improvements that can be made to AI systems based on this research:

) 1. Improved Galaxy Classification and Kinematic State Prediction:

The paper establishes a robust quantitative classification scheme for Early-Type Galaxies (ETGs) based on the Specific Angular Momentum Proxy, lambda R, combined with ellipticity (Figure 4). It explicitly states that the kinematic state has remained broadly unchanged since z ∼ 1 and that slow rotators are typically more massive and metal-rich than fast rotators.

The improved AI system can:

• Perform automated, high-accuracy classification of galaxies at intermediate redshifts by analyzing their spatially resolved kinematics (velocity fields) and photometric shapes (ellipticity).

• Predict the kinematic state (Fast vs. Slow Rotator) of a galaxy based on its measured structural parameters and kinematic maps, with high confidence in the classification scheme proposed by M. Cappellari (2016).

) 2. Robust Stellar Population Synthesis (SPS) Modeling:

The research demonstrates that global stellar population properties—age, metallicity ([Z/H]), and mass-to-light ratio (M/L)—are strongly correlated with the central velocity dispersion (sigma e), following established trends for local ETGs. Furthermore, it successfully used two independent SPS models (FSPS and GaLAXEV) to constrain these properties.

The improved AI system can:

• Accurately estimate stellar population parameters (age, metallicity, M/L) for galaxies at any redshift by fitting integrated spectra using advanced template libraries (IndoUS, MILES, XSL).

• Quantitatively predict the expected age and metallicity of a galaxy based solely on its measured velocity dispersion (sigma e), utilizing the derived relationships from Figure 7.

) 3. Automated Structural Parameter Extraction:

The methodology details precise measurement techniques for structural properties such as ellipticity (using MGE models), semi-major axis, and effective radius (Re) derived from HST imaging.

The improved AI system can:

• Automatically derive high-precision structural parameters from archival multi-band imaging of galaxies, enabling accurate calculation of physical size proxies like Re and the ellipticity parameter epsilon.

) 4. Kinematic Misalignment Detection:

The study measures the kinematic misalignment angle (Psi mis) between the photometric major axis and the kinematic major axis, finding that regular rotators exhibit small misalignments (< 10°), consistent with local Universe results.

The improved AI system can:

• Detect subtle rotational asymmetries or kinematic misalignment in galaxy velocity maps by comparing photometric PA with kinematic PA, providing a diagnostic tool for identifying potential recent mergers or accretion events that might have altered the galaxy's dynamical state.

) 5. Evolutionary Trend Tracking and Consistency Checking:

The core finding is the stability of ETG scaling relations across cosmic time (z ∼ 0.75 to z ∼ 0.25), suggesting little evolution in these properties over the last 7 Gyr.

The improved AI system can:

• Serve as a consistency checker for galaxy catalogs across different redshifts by verifying if the measured structural, kinematic, and stellar population scaling relations hold true across cosmic time, flagging outliers that might indicate unusual evolutionary pathways.

) 6. Enhanced Cosmological Parameter Inference (Indirect Application):

The paper notes that resolved kinematics of massive ETGs can be used to break degeneracies in the Hubble constant (H0) measurement using time-delay cosmography by constraining mass anisotropy.

The improved AI system can:

• Integrate resolved kinematic data from a large sample of non-lensed galaxies into Bayesian hierarchical frameworks for cosmological parameter estimation, specifically by providing crucial priors that help break mass-anisotropy degeneracies in H0 measurements.

We present a sample of 213 early-type galaxies (ETGs) at redshifts 0.25 < z < 0.75. We combine deep integral-field spectroscopy from the MUSE-DEEP survey with high-resolution HST imaging to study the structure, kinematics, and stellar populations of these galaxies. We measure spatially resolved stellar kinematics and use the specific angular momentum proxy, λ R, to classify galaxies into fast and slow rotators. We find a slow rotator fraction consistent with local Universe samples, suggesting little evolution in the massive ETG population since z about 1. The kinematic and photometric axes of fast rotators are generally well-aligned, similar to their local counterparts. We find that global stellar population properties, such as age, metallicity, and mass-to-light ratio (M*/L), correlate strongly with the central velocity dispersion (σ e), following trends established for local ETGs. Slow rotators are typically more massive, have higher σ e, and are more metal-rich than fast rotators. Our findings indicate that the fundamental structural, kinematic, and stellar population scaling relations of massive ETGs were already in place by z about 0.75, suggesting their evolutionary pathways have remained stable over the last about 7 Gyr.