WorksPhysics and Astronomy
OGLE Pre-discovery Observations of the Dwarf Planet Candidate 2026 ED138
Abstract
The newly discovered trans-Neptunian object, 2026 ED138, is a strong dwarf planet candidate detected by the Vera C. Rubin Observatory. Despite its high intrinsic brightness, it evaded detection by previous sky surveys as it hid near crowded Galactic disk fields. Here, we present an analysis of precovery observations of 2026 ED138 in archival data from the Optical Gravitational Lensing Experiment (OGLE). We identify 102 reliable detections of the object as it moved across the dense regions of the Galactic plane and bulge between 2011 and 2024. We provide precise astrometric and photometric measurements of the object. Modeling its phase curve yields absolute magnitudes of H_I = 2.224 +/- 0.034 mag and H_V = 3.154 +/- 0.035 mag. Period analysis of the phase-angle-corrected light curve reveals a candidate rotation period of 1.18077 days and its 1-day alias 0.54066 days, although a low signal-to-noise ratio renders these periodicities tentative.
Introduction
The newly detected trans-Neptunian object (TNO), 2026 ED138, is one of the intrinsically brightest known objects in its class, making it a strong dwarf planet candidate. It was discovered in data from the Vera C. Rubin Observatory by Holman et al. [7], who derived an absolute magnitude of the object of mag, corresponding to an estimated diameter of 612–1530 km for geometric albedos between 0.25 and 0.04. By combining Rubin observations with archival DECam data spanning 2016 to 2023, Holman et al. [7] determined a semi-major axis of au, an orbital eccentricity of , and an inclination of deg, placing the object near the edge of the 12:7 mean-motion resonance with Neptune.
Shortly after the discovery, Han et al. [6] reported detections of the object in full-frame images from the Transiting Exoplanet Survey Satellite (TESS) taken between 2019 and 2026. Murtagh et al. [10] analyzed publicly available archival images from various sky surveys, identifying 2026 ED138 on 13 photographic plates (taken between 1937 and 2000) and in several CCD images from surveys such as VISTA, SkyMapper, and ATLAS (from 2010 on). Incorporating these historical observations significantly refined the orbital solution for the object. Murtagh et al. [10] argued that about half of this improvement can be tracked to precise astrometry from VISTA and SkyMapper, highlighting the usefulness of archival CCD observations.
One reason why 2026 ED138 escaped detection in previous TNO searches is the fact that it lies deep in the southern hemisphere near the Galactic plane, a region typically avoided by traditional TNO surveys. An exception to this rule was the OGLE-Carnegie Kuiper Belt Survey (OCKS; Sheppard et al., [13]), which was conducted as part of the fourth phase of the Optical Gravitational Lensing Experiment (OGLE-IV; Udalski et al. [15]). The OCKS survey was carried out between 2010 March and 2011 January using the 1.3-m Warsaw Telescope equipped with a wide-field mosaic CCD camera.
The OCKS survey led to the discovery of several large TNOs, including 471143 Dziewanna (a.k.a. 2010 EK139). With an absolute magnitude of mag, this object was initially estimated to have a diameter as large as 1000 km. However, subsequent thermal observations with the Herschel satellite indicated a much smaller diameter of km (Pál et al., [11]) or km (Lellouch et al., [9]), while a stellar occultation observed on May 17, 2019 indicated a minimum major-axis length of 507 km (Burzyński, [3]).
The OCKS survey observed regions within 20 deg of the ecliptic plane. Curiously, at that time, 2026 ED138 was located at an ecliptic latitude of approximately deg, thereby evading detection in 2010. Nevertheless, over the past 16 years, 2026 ED138 remained close to the Galactic plane and Galactic bulge fields that were regularly observed as part of the OGLE-IV survey. This article presents an analysis of these pre-discovery OGLE observations of 2026 ED138.
Data
All data reported in this paper were collected as part of the OGLE-IV survey (Udalski et al., [15]). The survey operates the 1.3-m Warsaw Telescope located at Las Campanas Observatory1, Chile, which is equipped with a mosaic CCD camera providing a field of view of . The camera consists of thirty-two 2048 4102 detectors with a pixel scale of per pixel. Most observations were taken through the -band filter, which closely resembles that of the standard Cousins system. Approximately 10% of images were taken through the -band filter, whose transmission curve resembles that of the Johnson filter.
The survey observes a pre-defined grid of fields covering regions of the sky with the highest stellar density: the Galactic bulge, the Galactic plane, and the Magellanic Clouds. To identify images that might contain 2026 ED138, we first generated its ephemeris for the years 2010–2026 at 1-day intervals using the JPL Horizons System.2 The ephemeris was generated based on the orbital solution that was available on November 2, 2026.
Figure 1 shows the sky trajectory of 2026 ED138 along with the grid of nearby OGLE fields. The path of the object crosses the eastern Galactic disk fields (GD1064, GD1073, GD1081, GD1082, GD1089, GD1090, GD1097, GD1098, GD1105, GD1106, GD1112, GD1113, GD1901, GD1905, GD1909) as well as the western fields of the shallow Galactic bulge survey (BLG940, BLG943, BLG946, BLG947, BLG953, BLG961, BLG962, BLG983, BLG985). We identified all available OGLE images of these fields, calculated the expected position of 2026 ED138 at the time of each observation, and checked whether the position fell within a given field. In total, we identified 128 images that were expected to contain 2026 ED138.
Figure 1. Sky trajectory of 2026 ED138 from 2010 to 2026 (blue line). Black polygons mark the OGLE fields located along the path of the object. Pre-discovery OGLE observations are indicated by brown dots. Arrows indicate the direction of motion.
The OGLE photometric pipeline is based on the Difference Image Analysis (DIA) technique [14, 1] as implemented by Woźniak [16]. A deep, nearly noiseless reference image is constructed for each field and filter by stacking several high-quality individual frames. This reference image is then subtracted from each science frame, producing a difference (or subtracted) image. As a result, all constant stars are subtracted out, leaving only objects that have changed their brightness relative to the reference image or changed their position.
We examined the subtracted images for all frames that we identified in the previous step. We were able to detect the target near the position predicted by the JPL Horizons ephemeris in 112 of these images (for the remaining 16 images, the object was located slightly outside the footprint of the detector or in a gap between detectors of the CCD mosaic). The observed positions matched the predicted ephemeris very well, typically within 1–2 pixels (or –). Example images are shown in Figure 2, where the left column shows cutouts of the reference image, the middle column shows the science images, and the right column shows the difference images. The expected position of the object is indicated by white tick marks.

Figure 2. Example pre-discovery images of 2026 ED138 taken on the night of April 9, 2011 ( band; upper panel) and May 30, 2011 ( band; lower panel) with the exposure time of 180 s. The left column shows cutouts of the reference images, the middle column shows the science images, and the right column shows the difference images. The expected position of 2026 ED138 is marked by white tick marks. All cutouts are , North is up and East is to the left.
In most cases, in particular in the shallow survey frames, the object was too faint to be detected by the standard OGLE DIA pipeline. Therefore, to measure its position and flux, we fitted a 2D Gaussian profile to a pixel cutout around
the expected position:
where is the position of the centroid, is the standard deviation of the profile, is the flux of the object, and is the background flux. The best-fit parameters and their uncertainties were found using the Levenberg-Marquardt algorithm.
We were able to obtain reliable flux and position measurements for 2026 ED138 in 102 images. For the remaining ten images, the object was either marginally detected or blended with a brighter background star that was not perfectly subtracted, biasing the measurements.
All 102 measurements are reported in Table 1, which includes 91 epochs in the band and 11 in the band. The 2011 observations (six in and six in ) were obtained with exposure times of 180 s. The remaining images were collected as part of the OGLE-IV Galaxy Variability Survey [15] with much shorter exposure times (typically, 25 s in and 30 s in ).
| JD (UTC) | RA (deg) | Decl. (deg) | (″) | (″) | Filter | (mag) | (mag) |
| 2455660.69345 | 252.4650651 | −45.5974260 | 0.075 | 0.077 | I | 18.863 | 0.047 |
| 2455662.70740 | 252.4381975 | −45.6084694 | 0.077 | 0.077 | V | 19.830 | 0.061 |
| 2455686.60781 | 251.9990365 | −45.6993743 | 0.074 | 0.073 | I | 18.817 | 0.039 |
| 2455686.62997 | 251.9985916 | −45.6994135 | 0.075 | 0.074 | V | 19.716 | 0.045 |
| 2455701.70291 | 251.6366537 | −45.7131462 | 0.079 | 0.079 | I | 18.954 | 0.070 |
| 2455701.72507 | 251.6361107 | −45.7131488 | 0.090 | 0.091 | V | 19.947 | 0.119 |
| 2455701.74844 | 251.6354925 | −45.7130971 | 0.075 | 0.075 | I | 18.875 | 0.048 |
| 2455711.50197 | 251.3831321 | −45.7029074 | 0.076 | 0.077 | V | 19.721 | 0.061 |
| 2455712.51962 | 251.3562755 | −45.7009425 | 0.078 | 0.079 | I | 18.987 | 0.070 |
| 2455712.54178 | 251.3557448 | −45.7009017 | 0.075 | 0.076 | V | 19.735 | 0.052 |
Table 1. Positions and apparent magnitudes of 2026 ED138 measured from archival OGLE images. Note: This table is available in its entirety in machine-readable form.
Fluxes were calibrated into standard - and -band magnitudes following standard OGLE procedures [15]. The typical magnitude precision for deep 2011 observations is 0.045–0.070 mag, whereas 90% of shallow observations have photometric precision ranging from 0.087 to 0.290 mag, with a median of 0.133 mag.
The positions of the object measured on the subtracted images were transformed into equatorial coordinates using transformations similar to those described by Udalski et al. (2015). These transformations were derived by cross-matching bright stars detected in the OGLE images with the Gaia Data Release 3 catalog [4, 5] and calculating the third-degree polynomial transformations between the image coordinates and Gaia equatorial coordinates.
Positional uncertainties reported in Table 1 were calculated by adding in quadrature the centroid fitting uncertainty () and the root-mean-square (rms) error of the coordinate transformation ():
The value of ranges from 0.045 to 0.070 pixels for the deep observations and from 0.12 to 0.52 pixels for the shallow observations (with a median of 0.22 pixels). The rms of the coordinate transformations in one direction was typically –, depending on the field. In addition, represents a constant error term that was added in order to make sure that /d.o.f. for the best-fit astrometric models.
The data are available in the electronic form via:
https://www.astrouw.edu.pl/ogle/ogle4/2026ED138
https://zenodo.org/records/24144451
Orbital Parameters
We used LAYUP v1.0.0 ([8]) to derive the orbital elements of 2026 ED138 from the OGLE astrometric dataset. The resulting barycentric ecliptic J2000 orbital elements are presented in the second column of Table 2. The best-fit model yields /d.o.f. , indicating a good fit to the data. The astrometric residuals (in the North and East direction) from the best-fit model are shown in Figure 3.
| Element | OGLE only | Holman et al. | OGLE + Holman et al. |
| (au) | |||
| (deg) | |||
| (deg) | |||
| (deg) | |||
| (deg) | |||
| /d.o.f. | |||
| Epoch | |||
| (MJD TDB) |
Table 2. Orbital elements of 2026 ED138.
Figure 3. Astrometric residuals from the OGLE-only orbital fit.
In order to validate the accuracy of the OGLE astrometric measurements, we used LAYUP to fit the orbit to the astrometric measurements reported by Holman et al. [7], which are based on DECam and Rubin observations collected from 2016 through 2026. This fit yielded /d.o.f. . The best-fit parameters are reported in the third column of Table 2 and they match the parameters from the OGLE-only fit within their uncertainties (except for the mean anomaly , which was calculated for a different epoch).
We then used the DECam+Rubin model to calculate the predicted positions of 2026 ED138 at the time of OGLE observations. Evaluating the OGLE observations against this DECam+Rubin orbit yielded , which was larger than for the OGLE-only fit. This difference can be mostly explained by the fact that the OGLE dataset includes 41 pre-2016 observations, which significantly extend the observational arc. These 41 epochs contribute to , while the remaining difference can be attributed to the statistical noise in the data. Figure 4 shows the cumulative difference between the astrometric model that was fit to the Holman et al. [7] data and that fit to the OGLE-only data.
Figure 4. Cumulative difference between the astrometric model that was fit to the Holman et al. (2026a) data and that fit to the OGLE-only data.
The fourth column of Table 2 presents the orbital solution derived from the combined OGLE, DECam, and Rubin datasets.
Finally, we compared our astrometric observations with the orbital parameters calculated by Murtagh et al. [10], Tables 4 and 5. We used their model to calculate the predicted positions of 2026 ED138 at the time of OGLE observations. For that model, we found , i.e., a difference of compared to the OGLE-only model, indicating a slight tension. This difference cannot be attributed to data from a specific night or field, but rather accumulates throughout the entire duration of the observations.
Phase Curve and Search for Periodic Variability
The OGLE dataset spans over 13 years of observations across a relatively broad range of phase angles, ranging from 0.45 to 1.23 deg. Moreover, because all observations were obtained using standard - and -band filters, we can directly determine the absolute magnitudes of 2026 ED138. To this end, we fitted the following model to the OGLE - and -band data:
where and are the heliocentric and geocentric distances, is the Sun–target–observer phase angle, is the phase slope (common to both bands), and is the absolute magnitude in a given band. Figure 5 shows the dependence of the reduced magnitude, , on the phase angle ().
Figure 5. OGLE phase curve of 2026 ED138. Red and blue lines mark the best-fit models for the - and -band observations, respectively.
A linear least-squares fit yields:
These values are consistent with those reported by Holman et al. [7] ( mag and mag deg) and Murtagh et al. [10] ( mag and mag deg). Unlike those studies, which derived indirectly from - and -band observations, our results are based on direct -band photometry.
We derived a color index of the object to be mag. (Note that the uncertainty on is smaller than the uncertainty on either or because these values are strongly correlated.) Comparing this color index with the average values for taxonomic classes from Barucci et al. [2], we place 2026 ED138 in the BR group (intermediate blue-red), consistent with the classification of Murtagh et al. [10].
The phase-angle-corrected OGLE photometry of 2026 ED138 was subjected to period-finding procedures. We searched for possible stable periods that could be interpreted as rotation periods of the dwarf planet candidate. We used two period-search algorithms: FNPEAKS3 and TATRY [12]. Both searches yielded a similar peak frequency at d. However, the of the frequency peak is at a relatively low level (), making this detection very uncertain. Moreover, when two significant outliers were removed from the light curve, an alternative periodicity of d showed up with similar strength. This is likely a 1-day alias of the former period. Nevertheless, we are unable to definitively conclude if these periodicities are real, and both must be treated as very tentative. The OGLE -band data phased with these two periods are presented in Figure 6.
Figure 6. OGLE phase-angle-corrected -band light curve of 2026 ED138 folded with possible periods of d (left panel) and d (right panel).
Note that Murtagh et al. [10] reported tentative detection of periodic variability with a period of 0.4317 d at based on the analysis of DECam, Rubin, SkyMapper, VISTA, and ATLAS data. We do not find any significant signal at similar period in the OGLE data. Continuous observations of 2026 ED138 conducted over several consecutive nights will be required to confirm whether coherent rotational variability is present.
Conclusions
The discovery of a TNO as bright and potentially large as 2026 ED138 by Holman et al. [7] was unexpected. Wide-field sky surveys conducted over the past two decades, including the OGLE OCKS, covered and searched the sky regions most likely to harbor large TNOs, including dense fields in the Galactic bulge. However, this discovery demonstrates that orbital inclinations of TNOs may be higher than previously expected, suggesting that some such large objects may still remain undiscovered.
Although 2026 ED138 evaded detection by the OGLE OCKS in June 2010 because it was located approximately 3 deg south of the survey’s ecliptic latitude limit, it crossed several OGLE fields over the subsequent 16 years. The 102 OGLE pre-discovery detections of this dwarf planet candidate (including 41 epochs before 2016) almost doubled the available observational dataset. Extending the observational arc back to 2011 using these high-precision CCD measurements significantly refines the orbital solution. Furthermore, OGLE standard photometry led to the direct determination of the planet’s -band absolute magnitude of mag, and the color index equal to mag, as well as a tentative rotation period of 1.18077 d or 0.54066 d.
The high intrinsic brightness of 2026 ED138 suggests that this dwarf-planet candidate may have a large size, with an estimated diameter ranging from 612 to 1530 km ([7]) depending on its unknown albedo. Over the coming years, the object will continue to drift across dense stellar fields south of the Galactic bulge. Thus, the chances of stellar occultations, which may lead to a direct size estimate, are relatively large.
Murtagh et al. [10] provided a list of such phenomena in the coming years based on their orbital parameters. As these predictions are extremely sensitive to the accuracy of the orbit, they can be verified with our determination of orbital parameters of 2026 ED138. This is, however, out of the scope of this paper.
Acknowledgments
The OGLE project has received funding from the Polish National Science Centre grant OPUS 2024/55/B/ST9/00447 awarded to A.U.
References
- [1]Alard, C., and Lupton, R.H. 1998, ApJ, 503, 325.
- [2]Barucci, M.A., Belskaya, I.N., Fulchignoni, M., and Birlan, M. 2005, AJ, 130, 1291.
- [3]Burzyński, W. 2024, Journal for Occultation Astronomy, 2024-03, 14.
- [4]Gaia Collaboration, et al. 2016, A&A, 595, A1.
- [5]Gaia Collaboration, et al. 2021, A&A, 649, A1.
- [6]Han, T., Audenaert, J., Muthukrishna, D., Holman, M., and Ricker, G.R. 2026, arXiv e-prints, arXiv:2610.40106.DOI
- [7]Holman, M.J., et al. 2026a, arXiv e-prints, arXiv:2610.38693.arxiv.org/abs/2610.38693
- [8]Holman, M.J., et al. 2026b, Layup: Orbit Fitting at LSST Scale.DOI
- [9]Lellouch, E., et al. 2013, A&A, 557, A60.
- [10]Murtagh, J., et al. 2026, arXiv e-prints, arXiv:2611.02371.arxiv.org/abs/2611.02371
- [11]Pál, A., et al. 2012, A&A, 541, L6.
- [12]Schwarzenberg-Czerny, A. 1996, ApJL, 460, L107.
- [13]Sheppard, S.S., et al. 2011, AJ, 142, 98.
- [14]Tomaney, A.B., and Crotts, A.P.S. 1996, AJ, 112, 2872.
- [15]Udalski, A., Szymański, M.K., and Szymański, G. 2015, Acta Astron., 65, 1.
- [16]Woźniak, P.R. 2000, Acta Astron., 50, 421.




