🔭 astrophysics

You Shall Not Pass (Without Modeling): High-Resolution Analysis of KMT-2019-BLG-0253 using MORIA

This paper introduces MORIA, an automated high-resolution imaging pipeline that analyzes HST observations of the microlensing event KMT-2019-BLG-0253 to precisely characterize its planetary system and demonstrate the tool's critical role in informing light curve modeling for future Roman Galactic Bulge Time Domain Survey discoveries.

Original authors: T. Dex Bhadra, Sean K. Terry, David P. Bennett, Aparna Bhattacharya, Ian A. Bond, Jon Hulberg, Stela Ishitani Silva, Przemek Mróz, Aikaterini Vandorou

Published 2026-05-12

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: T. Dex Bhadra, Sean K. Terry, David P. Bennett, Aparna Bhattacharya, Ian A. Bond, Jon Hulberg, Stela Ishitani Silva, Przemek Mróz, Aikaterini Vandorou

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine trying to solve a cosmic mystery where a tiny, invisible planet passes in front of a distant star, briefly brightening it like a lighthouse beam. This is gravitational microlensing. For years, astronomers could detect these bright flashes and know a planet was there, but they were like detectives trying to solve a crime with only a blurry photo. They could guess the planet's size relative to its star, but they couldn't be sure of the star's actual weight, the planet's true mass, or how far away the whole system was.

This paper introduces a new tool called MORIA (Microlensing Object high-Resolution Imaging Analysis) and uses it to finally get a "high-definition" look at a specific mystery: KMT-2019-BLG-0253.

Here is the story of how they solved it, using simple analogies:

1. The Problem: The "Blurred" Photo

For a long time, ground-based telescopes looked at this event. Because the atmosphere makes the stars twinkle and blur, the light from the target star, the hidden host star, and a nearby "neighbor" star all got mashed together into one fuzzy blob. It was like trying to hear three people talking in a crowded room from a mile away; you knew voices were there, but you couldn't tell who was saying what.

Previous studies guessed the host star was about 5,000 light-years away and had a planet about 9 times the mass of Earth. But the math had too many "maybe" answers.

2. The Solution: The "Cosmic Zoom Lens"

To clear up the blur, the team used the Hubble Space Telescope (HST), which acts like a super-powered zoom lens sitting above the Earth's atmosphere. They took two types of pictures:

The "Old Photo" (2005): A lucky snapshot taken years before the event happened. It showed a slightly stretched-out star, hinting that something was blended with it, but the picture wasn't sharp enough to separate them.
The "New Photo" (2025): A dedicated, high-definition photo taken 19 years after the event. By this time, the host star and the background star had drifted apart in the sky, like two cars that were once bumper-to-bumper but have now driven in different directions.

3. The New Tool: MORIA

The team didn't just look at the new photo; they built a new software robot named MORIA to analyze it. Think of MORIA as a super-smart photo editor that doesn't just sharpen an image, but mathematically "un-mixes" the colors.

The Analogy: Imagine a smoothie made of strawberries, bananas, and blueberries. If you just look at the purple liquid, you can't tell the ingredients. MORIA is like a machine that can taste the liquid and mathematically separate it back into the exact amount of strawberry, banana, and blueberry juice.
What it did: MORIA analyzed the Hubble image and proved there weren't just two stars blended together, but three. It separated the light of the Source (the background star), the Lens (the host star with the planet), and a Blend (a random neighbor star that was just in the way).

4. The Results: Solving the Mystery

By separating the stars, the team could finally measure the "Lens" star directly.

The Host Star: They measured its brightness and used a known rule (like a "weight-to-size" chart for stars) to determine its mass. It turned out to be a small, dim star (about 65% the mass of our Sun), located much closer to us than previously thought: 2.64 kiloparsecs (about 8,600 light-years) away.
The Planet: With the host star's mass known, they could finally calculate the planet's true weight. It is a "Super-Earth" or "Mini-Neptune," weighing between 7 and 9 times the mass of Earth.
The Distance: They confirmed the system is relatively close in cosmic terms, resolving the uncertainty about whether it was 5,000 or 8,000 light-years away.

5. Why This Matters for the Future

The paper explains that a new space telescope, the Nancy Grace Roman Space Telescope, is launching soon (in 2026). It will find thousands of these planetary events. However, without a tool like MORIA, astronomers would be stuck with blurry guesses again.

The authors are essentially saying: "We built this new robot (MORIA) to clean up the messy photos from the Hubble telescope. We used it to solve one specific mystery (KMT-2019-BLG-0253), and now we are ready to use this same robot to solve thousands of mysteries when the new Roman telescope starts sending us data."

In short: They took a blurry cosmic puzzle, built a new digital tool to separate the pieces, and found that the "monster" they were looking for was actually a smaller, closer, and more precise system than anyone had guessed before.

Technical Summary: High-Resolution Analysis of KMT-2019-BLG-0253 using MORIA

Problem Statement
Gravitational microlensing is a unique method for detecting cold, low-mass exoplanets beyond the snow line. However, a primary limitation of this technique is that standard lightcurve modeling typically yields only the mass ratio ( $q$ ) of the lens system, leaving physical parameters such as host mass and distance unconstrained. While "higher-order effects" in lightcurves or high-angular resolution imaging (e.g., HST, Keck AO) can resolve these degeneracies, the latter often requires manual, labor-intensive analysis of crowded fields. Furthermore, the upcoming Nancy Grace Roman Galactic Bulge Time Domain Survey (GBTDS) is expected to detect thousands of microlensing events, necessitating an automated, end-to-end pipeline to reduce high-resolution imaging data and perform multi-component Point Spread Function (PSF) fitting to constrain lens properties.

Methodology
The authors present MORIA (Microlensing Object high-Resolution Imaging Analysis), a Python-based, automated pipeline designed to process high-resolution images of microlensing targets. The workflow involves:

Data Reduction: Converting HST Wide-Field Camera 3 (WFC3) exposures into a common coordinate system, applying geometric distortion corrections, and generating star lists.
PSF Modeling: Constructing empirical PSF models using reference stars with similar magnitudes and colors to the target, selected via Color-Magnitude Diagrams (CMDs).
Multi-Star PSF Fitting: Performing simultaneous (non-iterative) fitting of one, two, or three stellar components to the target's pixel grid using Markov Chain Monte Carlo (MCMC) sampling. This determines the positions, fluxes, and residuals of blended sources.
Photometric Calibration: Calibrating instrumental magnitudes to the OGLE-III system by cross-matching HST stars with the OGLE catalog.
Image-Constrained Modeling: Integrating the derived lens-source separation, relative proper motion ( $\mu_{rel}$ ), and flux constraints into the lightcurve modeling code eesunhong. This restricts the parameter space to solutions consistent with the high-resolution imaging.

The pipeline was applied to the microlensing event KMT-2019-BLG-0253 (KB190253). The analysis utilized:

Archival Data: 2005 HST/ACS observations (GO-10353), which serendipitously captured the field but were not optimized for high-precision astrometry.
Follow-up Data: 2025 HST/WFC3 observations (GO-17834) taken ~19.5 years after the 2005 epoch and ~6 years after the 2019 peak magnification.
Lightcurve Data: Photometry from OGLE, MOA, and KMTNet.

Key Results

Resolution of Blending: The 2025 HST imaging revealed three distinct stars at the target location. Simultaneous three-star PSF fitting significantly improved the $\chi^2$ over single- and two-star models ( $\Delta\chi^2 \approx 346$ and $41$, respectively).
Identification of Components: By combining lightcurve modeling with image constraints and Galactic kinematic priors (Koshimoto et al. 2021), the authors identified Star 2 as the source, Star 3 as the lens (host star), and Star 1 as a blended field star.
Physical Parameters:
- Host Mass: $M_{host} = 0.65 \pm 0.04 M_\odot$ .
- Planet Mass: $m_p = 7.18 \pm 0.40 M_\oplus$ (close solution) or $9.48 \pm 1.13 M_\oplus$ (wide solution).
- Lens Distance: $D_L = 2.64 \pm 0.22$ kpc.
- Relative Proper Motion: $\mu_{rel,H} = 7.98 \pm 0.21$ mas/yr.
- Separation: Projected separation $a_\perp = 3.06 \pm 0.12$ au; 3D separation $a_{3d} = 3.74 \pm 0.15$ au.
Degeneracy Reduction: The image-constrained modeling reduced the number of viable solutions from four (as reported in Hwang et al. 2022) to two, resolving the $u_0 < 0$ ambiguity. However, the classic close/wide degeneracy remains.
Precision: The direct flux measurements allowed for tighter constraints on the angular Einstein radius ( $\theta_E$ ) and microlensing parallax ( $\pi_E$ ), resulting in significantly smaller uncertainties compared to the original Bayesian analysis.

Significance and Claims
The paper demonstrates that MORIA is a viable, automated tool for extracting precise physical parameters from crowded-field microlensing targets. By directly measuring lens flux and relative proper motion, the pipeline enables the use of empirical mass-luminosity relations to infer host masses and distances, bypassing the large uncertainties associated with standard Bayesian priors.

The authors claim that this work highlights the critical importance of automated high-resolution imaging tools for the upcoming Roman GBTDS. They posit that MORIA (or a Roman-adapted version) will be essential for characterizing the expected ~1,400 bound exoplanets and hundreds of free-floating planets, as well as for breaking degeneracies in lightcurve modeling that ground-based surveys cannot resolve. The study serves as a proof-of-concept that automated, image-constrained modeling can transform microlensing detections into well-characterized planetary systems.

Limitations and Future Work
The authors note that while the close/wide degeneracy was not fully broken, the solution space was halved. Definitive identification of the third star (Star 1) as a field star versus a bound companion requires an additional epoch of high-resolution imaging (e.g., from Roman GBTDS) to measure its proper motion relative to the lens-source pair. Future development of MORIA is planned to support ACS and JWST data, simultaneous fitting of $n > 3$ objects, and batch processing for the large sample sizes expected from Roman.

1. The Problem: The "Blurred" Photo

2. The Solution: The "Cosmic Zoom Lens"

3. The New Tool: MORIA

4. The Results: Solving the Mystery

5. Why This Matters for the Future

More like this