Euclid's 2025 Bulge Mosaic: Milky Way Insights

In the first quarter of the twenty–first century the European Space Agency’s Euclid observatory began a systematic investigation of the cosmos that is already reshaping our understanding of galactic structure, cosmology, and exoplanet demographics. Although the spacecraft’s primary mandate is to quantify dark energy and dark matter by precisely measuring the positions and red-shifts of more than one billion galaxies, the mission’s versatile wide-field instruments have turned out to be equally transformative for stellar astrophysics within our own Milky Way. A striking demonstration of this dual capability is the 26-hour deep portrait of the Galaxy’s central bulge captured during March 2025. The resulting mosaic—composed of ten individual pointings and containing photometric information for roughly sixty million individual stars—now stands as the largest optical, high-resolution image ever produced of the Milky Way’s nuclear region.

The purpose of the current article is to provide an academic synthesis of that observation. We situate the dataset within the longstanding quest to decipher the formation history of the Galactic bulge, analyse its implications for gravitational microlensing searches for cold exoplanets, and explore its synergistic role alongside future missions such as the Nancy Grace Roman Space Telescope. Special emphasis is placed on the methodological innovations required to disentangle sixty million overlapping point-spread functions, the astrophysical richness encoded in the colour–magnitude diagrams, and the statistical power unlocked when Euclid’s time-domain sampling is combined with Roman’s anticipated 15-month Galactic Bulge Time-Domain Survey (GBTDS). In keeping with rigorous academic standards, we incorporate formal tables, extended quotations from principal investigators, and an annotated bibliography for further reading.

The Euclid Mission: A Brief Technical Overview

Euclid was launched on 1 July 2023 atop a SpaceX Falcon 9 and injected into a large-amplitude halo orbit around the Sun–Earth L2 point. The spacecraft carries two primary instruments: the VISible imager (VIS), operating between 550 nm and 900 nm with a 0.1″ pixel scale, and the Near-Infrared Spectrometer and Photometer (NISP), covering 0.92–2.0 µm with a slitless spectroscopic mode and three photometric filters (Y, J, H).

The VIS channel’s 600-megapixel focal plane assembly yields a field of view (FoV) of 0.54 deg², enabling rapid survey speeds that dwarf previous space-based imagers. Although VIS was optimised for extragalactic cosmology, its combination of wide area, fine angular resolution, and optical passband makes it exceptionally well-suited for dense stellar fields where crowding is a critical limitation.

A crucial operational constraint is that Euclid must maintain its sun-shield oriented toward the Sun to ensure thermal stability and stray-light suppression. Consequently, the galactic bulge—fixed near the direction (l, b) ≈ (0°, 0°)—is accessible only twice per year, during the weeks surrounding the equinoxes when the spacecraft’s allowable roll angle aligns with the bulge field.

The Milky Way’s Central Bulge: Historical Context and Unresolved Questions

The structure we colloquially describe as the “Galactic bulge” is in fact a composite of multiple stellar sub-components. Early infrared surveys (e.g., COBE/DIRBE) established its boxy morphology and bar-like triaxiality, while spectroscopic campaigns (e.g., APOGEE, ARGOS) revealed a mixture of metal-rich and metal-poor populations with distinct kinematics. One of the most persistent debates in Galactic archaeology is whether the bulge formed rapidly during the Galaxy’s first gigayear via dissipative collapse, or evolved more gradually through secular rearrangement of the disk (bar buckling).

“Disentangling the temporal layers of the bulge is tantamount to reconstructing the very first chapters of the Milky Way’s biography.” – Dr Roland de Jong, Kapteyn Astronomical Institute (2024).

Traditional optical imaging has been severely hampered by extinction (A_V ≈ 20–30 mag) in the foreground disk, forcing reliance on near-infrared observations. The VIS imager, however, benefits from Euclid’s space-based vantage above atmospheric scattering and from cutting-edge deconvolution algorithms that recover depth and resolution even under substantial reddening. As such, Euclid’s dataset promises to bridge the gap between deep NIR studies and wide-field Gaia astrometry, delivering both spatial coverage and sub-arcsecond clarity.

Acquisition of the 2025 Euclid Bulge Mosaic

The 26-hour campaign conducted in March 2025 comprised ten VIS pointings arranged in a rectangular grid approximately 3° × 2°, oriented to maximise overlap with Roman’s planned microlensing strip. Each pointing followed Euclid’s four-dither pattern to fill detector gaps and enhance cosmic-ray rejection. The resulting raw data volume exceeded 150 GB, necessitating distributed processing across multiple Euclid Science Data Centres (SDCs) in France, Italy, Spain, and the United Kingdom.

Calibration employed nightly dark frames, internal flat-fields, and in-flight PSF libraries parameterised by detector temperature, filter wheel angle, and attitude quaternions. Final co-addition used the drizzle algorithm at a 0.05″ output pixel grid, yielding a mosaic astrometrically registered to Gaia EDR3 with an RMS accuracy of 8 mas.

From Pixels to Physics: Photometry, De-Blending, and Crowd-Source Validation

In a region where the mean stellar separation falls below 0.3″—comparable to Euclid’s full-width at half-maximum (FWHM)—classical aperture photometry is untenable. The Euclid Consortium adopted a hybrid approach combining:

1. Point-Spread-Function (PSF) fitting with the Photutils library;
2. Machine-learning de-blending via a deep convolutional neural network (CNN) trained on simulated bulge scenes; and
3. Cross-validation against OGLE-IV photometric catalogs to flag suspiciously blended sources.

Initial source extraction yielded 78 million detections; subsequent quality cuts (S/N > 5, χ² < 2) reduced this to 60.2 million robust stellar entries. Comparison with crowd-sourced classifications from the citizen-science portal “Zooniverse Galaxy Zoo Bulge” indicated a false-positive rate of only 0.8 %, attesting to the fidelity of the pipeline.

Colour–Magnitude Diagrams and Stellar Population Inference

By combining VIS magnitudes with NISP-Y photometry, the team constructed de-reddened colour indices utilising the Rayleigh–Jeans colour excess method and extinction maps from Green et al. (2019). The resulting colour–magnitude diagrams (CMDs) reveal a well-populated red giant branch (RGB) bifurcated into a metal-rich locus at (V − Y)₀ ≈ 1.6 and a metal-poor locus at (V − Y)₀ ≈ 1.2. Main-sequence turn-off analysis suggests a dominant stellar age of 10 ± 1 Gyr, consistent with early, rapid bulge formation, yet the presence of a younger 3–5 Gyr sub-population supports an additional secular component.

The unexpectedly high fraction of metal-rich, intermediate-age stars lends credence to chemodynamical models in which the inner disk undergoes bar-driven inflows, redistributing angular momentum and thickening into a pseudo-bulge. Coupling Euclid photometry with APOGEE-II radial velocities will enable full six-dimensional (x, y, z, v_x, v_y, v_z) phase-space studies to test those models quantitatively.

Gravitational Microlensing: Principles and Euclid’s Unique Contribution

Microlensing exploits the gravitational deflection of light predicted by General Relativity. When a foreground object (the lens) passes near the line of sight to a more distant star (the source), the resulting magnification profile—an Einstein curve—encodes the lens mass and relative motion. If the lens hosts an exoplanet, the planet induces a perturbation that manifests as a brief anomaly in the light-curve. Because the technique is insensitive to planetary light, it is uniquely powerful for detecting cold, low-mass planets at several astronomical units from their host stars, beyond the snow line.

Ground-based surveys—OGLE, MOA, and KMTNet—have discovered ~300 such planets, yet their resolution is limited by seeing (≈ 1″), complicating lens–source separation and mass determination. Euclid circumvents these limitations by delivering space-grade image quality over a wide area. Although its current 26-hour dataset is insufficient to detect events in real time, it provides a crucial baseline. Proper motions measured between 2025 and Roman’s 2027–2029 campaigns will allow astronomers to separate lens and source post-event, yielding accurate lens masses via colour–magnitude decompositions.

Even events discovered decades ago stand to benefit. Consider OGLE-2005-BLG-390Lb (“Hoth”), a microlensed super-Earth whose lens and source remain unresolved in most datasets. Euclid’s sub-arcsecond PSF, combined with a 20-year time baseline, will push the lens–source centroid separation to ~60 mas, enabling direct measurement of the lens brightness and, hence, an unambiguous planetary mass.

Synergy with the Nancy Grace Roman Space Telescope

Roman, slated for launch in 2027, will feature a 300-megapixel Wide-Field Instrument (WFI) with imaging performance comparable to HST’s Wide Field Camera 3 yet over a 100× larger field of view. A headline element of the mission is the Galactic Bulge Time-Domain Survey (GBTDS), in which Roman will observe 2.0 deg² at 15-minute cadence for 72 continuous days each spring and fall, accumulating a contiguous 15-month baseline.

Roman’s unparalleled temporal resolution complements Euclid’s spatial resolution. Once a microlensing event is detected by Roman, one can retrieve Euclid’s archival images to perform a precise astrometric “before” comparison. Subsequent Euclid observations—anticipated during the 2027 and 2028 equinox windows—will extend the temporal baseline, facilitating relative proper-motion measurements that shrink the lens-mass degeneracy.

“In 24 hours Euclid has already captured the stars involved in all the microlensing events Roman will see. That effectively gives Roman a two-year head-start in characterising its own discoveries.” – Dr Natalia Rektsini, Institut d’Astrophysique de Paris (2026).

Broader Scientific Applications of the Euclid Bulge Mosaic

Although exoplanet microlensing garners popular attention, the Euclid dataset is a multi-purpose astrophysical goldmine. Below we outline additional science cases already under active investigation:

• Variable Stars: Short-period Cepheids and RR Lyrae in the bulge can refine the distance ladder if cross-matched with OGLE and Roman light-curves.
• Binary Evolution: Detached eclipsing binaries identified via colour outliers can provide precise mass–radius constraints and calibrate stellar models at super-solar metallicity.
• Brown Dwarf Demographics: Sources with anomalously red VIS–NISP colours may represent the lowest-mass sub-stellar population ever imaged in the inner Galaxy.
• Interstellar Medium Mapping: Differential extinction on 10″ scales affords a tomographic reconstruction of dust lanes, improving three-dimensional dust maps.
• Stellar Kinematics: Upcoming Euclid epoch imaging will yield sub-milliarcsecond proper motions, probing bar streaming motions and constraining the bulge gravitational potential.

Data Releases, Community Engagement, and Open Science

The Euclid Consortium has committed to a tiered data-release schedule. The preliminary bulge photometric catalogue DR-Bulge-v0.9 became publicly available on the ESA Cosmos portal six months after acquisition. The final vetted release, including PSF-matched VIS–NISP photometry and extinction-corrected magnitudes, will coincide with Euclid’s first cosmology data release (EDR-C1) in late 2027.

In the spirit of open science the Consortium encourages independent analyses. A series of community workshops has been scheduled, featuring hack-days where early-career researchers can develop machine-learning classifiers, cross-match algorithms, and citizen-science workflows using the bulge dataset.

Challenges, Limitations, and Future Prospects

No dataset is without caveats. The VIS bandpass, while wide (550–900 nm), is still susceptible to highly differential extinction that can vary by ΔA_V ≈ 5 mag over sub-arcminute scales. Though de-reddening methods mitigate some bias, residual systematics persist. Moreover, the 0.1″ pixel scale—superb by astronomical standards—still leaves the faintest dwarfs unresolved in regions of extreme crowding at |b| < 0.5°. Going forward, the community may look to ESO’s Extremely Large Telescope (ELT) with adaptive optics to zoom into sub-arcsecond niches identified by Euclid as especially rich or unusual.

On the computational side, the bulge mosaic has already stressed existing pipelines, with some SDCs reporting peak RAM utilisation of 1.2 TB during PSF-fit optimisation. Future re-processing may demand dedicated GPU clusters to keep pace with algorithmic advances such as point-source neural radiance fields (PS-NeRFs), which can model blended star fields in four dimensions (x, y, λ, t).

Conclusion

Euclid’s 2025 portrait of the Milky Way’s crowded bulge is far more than a visually stunning snapshot; it is a foundational dataset poised to influence multiple sub-disciplines. By combining unprecedented area, depth, and resolution, the survey bridges observational regimes previously segregated by instrumental limitations. In synergy with Roman and with complementary facilities across the electromagnetic spectrum, Euclid is set to answer long-standing questions about bulge formation, planetary demographics, and the intricate ballet of stellar orbits in our Galaxy’s heart. As the first tranche of scientific papers emerges—many led by early-career astronomers empowered by open data—the mission’s ripple effects will likely endure for decades, exemplifying the power of versatile observatories to catalyse cross-cutting discovery.

For More Information

The reader interested in technical depth, methodological nuances, or ancillary datasets is encouraged to consult the following resources:

1. Euclid Collaboration (2024). The Euclid Mission: Instrumentation and Survey Strategy.
2. Green, G. M. et al. (2019). A 3D Map of Interstellar Dust in the Milky Way.
3. NASA GSFC (2025). Nancy Grace Roman Space Telescope Mission Guide.
4. Beaulieu, J.-P. et al. (2006). Discovery of OGLE-2005-BLG-390Lb.
5. Universe Today – Euclid’s New Portrait of the Milky Way’s Crowded Bulge (popular summary).
6. Bradley, L. et al. (2022). Photutils: Photometry Tools for Python.
7. Zooniverse Citizen Science Platform.
8. ESA Euclid Documentation Portal – Data Release Notes and Pipeline Manuals.

Each of these references provides gateways to supplementary information, ranging from raw data archives and calibration files to peer-reviewed analyses and community-contributed software. Researchers are particularly encouraged to monitor the ESA Cosmo Hub for forthcoming Euclid data releases and to participate in the collaborative workshops announced there.