DEVILS: Deep Extragalactic Legacy Survey

• DEVILS is a high-completeness spectroscopic survey targeting ~60,000 intermediate-redshift galaxies (0.3<z<1.0) to bridge local and deep extragalactic observations.
• The survey employs deep near-IR selection and advanced photometric extraction with dynamic redshift feedback to robustly identify galaxy groups and halos.
• It enables precise measurements of group statistics, merger rates, and environmental influences on galaxy evolution, providing a keystone dataset for the ΛCDM paradigm.

The Deep Extragalactic VIsible Legacy Survey (DEVILS) is a high-completeness, intermediate-redshift spectroscopic campaign targeting $\sim$60,000 galaxies to $Y<21.2$ mag over $\sim$6 deg$^2$ in three premier extragalactic fields: COSMOS (D10), Extended Chandra Deep Field South (ECDFS, D03), and XMM-Large Scale Structure (XMM-LSS, D02). DEVILS is designed to bridge the gap between local ($z\sim0$) spectroscopic surveys and high-redshift pencil-beam programs, creating a legacy dataset that enables the direct measurement of group and halo properties, galaxy evolution, star formation histories, and the environmental processes shaping the Universe over the last $\sim$8 Gyrs (Davies et al., 2018).

1. Scientific Motivation and Survey Rationale

DEVILS is constructed to deliver one of the first very high-completeness ($>95\%$) spectroscopic surveys at the previously underexplored intermediate redshifts ($0.3

• Measuring the late-time evolution of the dark matter halo mass function (HMF), a direct prediction of $\Lambda$CDM, by robustly identifying galaxy groups and clusters down to low halo masses. This enables constraints on the growth of structure, especially at the group and cluster scale.
• Investigating how the environment—group, filament, and cluster membership—influences galaxy evolution. This encompasses both global properties (stellar mass function, star formation history) and local processes (e.g., gas stripping, minor mergers), thereby tying local and large-scale evolution (Davies et al., 2018).

DEVILS fills a critical redshift regime, linking local surveys like SDSS and GAMA to deep, narrow-field high-$z$ efforts, and is uniquely configured to provide robust group catalogs, merger rates, and environmental metrics at $0.3

2. Experimental Design and Target Selection

Field Selection

DEVILS targets three well-studied, deep-drilling fields overlapping with multi-wavelength ancillary data and future survey footprints:

Field	Common Name	RA, Dec (J2000)	Area (deg²)
D02	XMM-LSS	~02h22m, –04°42′	3.0
D03	ECDFS	~03h32m, –28°00′	1.5
D10	COSMOS	~10h00m, +02°13′	1.5

The total unmasked area is approximately 6 deg$^{2}$ (Davies et al., 2018).

Photometric Selection

DEVILS uses deep near-IR VISTA Y-band imaging (VIDEO for D02 and D03, UltraVISTA for D10) to select targets. The Y-band limit at $Y<21.2$ mag is empirically chosen via simulations (TAO lightcones) to enable:

• Detection of $M^{*}$ galaxies out to $z\sim1$
• Detection of major merger pairs to $z\sim0.8$
• Identification of galaxy groups ($M_{\rm halo}\sim10^{13}{\rm M}_{\odot}$) to $z\sim0.7$ with high completeness.

Selection is performed using ProFound, an advanced source finder, which provides robust total fluxes via isophotal segment dilation and optimizes color- and total-flux extraction, significantly improving deblending and low surface brightness photometry (Davies et al., 2018, Davies et al., 2021).

Star-Galaxy Separation

A key step involves NIR color and surface-brightness thresholds:

$$(H-Ks)-(Y-J) > -0.26 \quad \text{and} \quad \langle\mu_{Y,90}\rangle > 18~\text{mag arcsec}^{-2}$$

where $\langle\mu_{Y,90}\rangle$ is the Y-band mean surface brightness within the 90% flux radius. Visual inspection corroborates automated classification (Davies et al., 2018).

3. Observational Methodology and Redshift Acquisition

Spectroscopy

DEVILS is undertaken at the 3.9 m Anglo-Australian Telescope (AAT) using AAOmega with the Two-degree Field (2dF) fibre positioner, offering 400 fibres across a 2° diameter field. The spectral coverage spans 3750 Å to 8850 Å at $R\sim1000$–$1600$, enabling secure redshift measurements via prominent features across $0.3Davies et al., 2018).

Redshift Feedback Strategy

A defining aspect is the dynamic “redshift feedback” approach:

• All targets are initially observed with a 1 hr unit exposure, chosen to be sky-limited yet efficient for rapid reconfiguration.
• Each night, spectra are reduced and redshifts measured promptly using the TAZ pipeline.
• Successfully redshifted targets are removed; those without reliable redshifts are reobserved.

This feedback loop maximizes observational efficiency by concentrating subsequent time on faint or otherwise “hard” targets, enabling $>95\%$ spectroscopic completeness without unnecessary overexposure (Davies et al., 2018).

4. Spectroscopic and Catalog Properties

• Initial target sample: $\sim$57,000 galaxies to $Y<21.2$ across 6 deg$^{2}$.
• During early observations, $\sim$8,856 sources were observed yielding $\sim$4,353 secure redshifts despite $\sim$33% time lost to weather, with $\sim$50% redshift-success rate (consistent with prioritization of faint, hard-to-type objects).
• Early datasets already demonstrated substantial improvements to redshift completeness, with, e.g., field D10's median completeness rising from $54\%$ to $67\%$ post-initial campaign (Davies et al., 2018).
• Catalog entries include astrometry, total and surface photometry, morphological parameters, and cross-matching with auxiliary datasets. Fields were chosen to enable legacy combination with data from LSST, Euclid, MeerKAT, and others.

5. Scientific Capabilities and Early Results

DEVILS is explicitly optimized for group and halo identification, as robust group finding and halo occupation distribution (HOD) analysis requires highly complete slit/fibre redshift coverage. Fields are sufficiently deep to characterize group-scale ($\sim10^{13}M_{\odot}$) dark matter haloes, galaxy-galaxy pairs, and merger rates with minimal projection effects.

Early science cases include:

• Mapping the $z=0.3$–$1.0$ evolution of the group and cluster population and the environmental dependence of star formation.
• Quantifying the major merger rate through the spectroscopic identification of close-pairs.
• Enabling cross-disciplinary time-domain science via simultaneous observation of transients (supernovae, AGN reverberation) within well-characterized environments (Davies et al., 2018).

Representative extracted spectra (absorption-line and emission-line systems at a range of magnitudes and redshifts) demonstrate the high S/N achievable and the efficacy of the dynamic feedback approach.

6. Data Integrity, Completeness, and Public Release

DEVILS establishes a uniform, high-quality, and statistically robust sample in three deep fields, supported by state-of-the-art multi-wavelength photometry (22 bands, FUV–FIR, ProFound extraction) and careful masking of artefacts. Detailed completeness maps (2′ × 2′ grids) allow users to construct statistically rigorous, volume-limited samples for downstream analyses (Davies et al., 2021).

Catalogues and images are scheduled for public release after the completion of key internal science analyses, ensuring reproducibility and legacy value.

7. Legacy Value and Future Prospects

DEVILS is positioned to become a fundamental reference survey for structure and galaxy evolution in the $\Lambda$CDM paradigm at $0.3

• Quantitative halo and group statistics with minimal redshift failure bias.
• Empirically mapping the impact of mergers, environment, and cosmic web filaments on the stellar mass function and galaxy quenching processes.
• Providing an anchor for spectral energy distribution modeling, halo occupation distribution inference, and calibration of next-generation cosmological simulations.

The survey’s design—combining deep, uniform near-IR selection, high spectroscopic completeness, innovative feedback strategies, and precise photometry—enables critical tests of galaxy evolution theory, the emergence of environmental quenching and dynamical assembly, and continuous comparison from $z\sim1$ to the local Universe (Davies et al., 2018).

---

In summary, the Deep Extragalactic VIsible Legacy Survey (DEVILS) constitutes a process-optimized, high-completeness, near-infrared selected spectroscopic survey specifically designed for transformative measurements of group environments, mergers, and galaxy evolution at intermediate redshift, providing a keystone dataset for the coming decade of extragalactic research.