DECam Rogue Earths & Mars Survey (DREAMS)

• DREAMS is a gravitational microlensing survey using DECam on the Blanco 4m to detect sub-Earth-mass bound and free-floating planets.
• It employs minute-cadence imaging in SDSS-like r and z bands to capture short-duration events approaching Mars and Moon masses.
• The survey targets Galactic bulge fields to yield statistical constraints on low-mass exoplanet populations and resolve modeling degeneracies.

The DECam Rogue Earths and Mars Survey (DREAMS) is a high-cadence gravitational microlensing program operating on the Blanco 4 m telescope at CTIO since June 2025. Its primary objective is to extend microlensing sensitivity into the regime of extremely low-mass planetary bodies, targeting both bound analogs—"rogue Earths" in wide orbits—and unbound, free-floating planets (FFPs), with detection thresholds approaching the mass of Mars and, in favorable conditions, the Moon. DREAMS leverages minute-scale imaging cadence, a large-aperture telescope, and the wide field of view of the DECam imager to address parameter spaces inaccessible to conventional microlensing surveys, thus enabling statistical constraints on sub-Earth-mass planetary populations in the Galactic bulge and disk (Yang et al., 16 Jan 2026).

1. Scientific Motivation

DREAMS addresses both the detection of faint, low-mass bound exoplanets and the census of unbound planetary-mass objects. Traditional microlensing surveys (e.g., OGLE, MOA, KMTNet, PRIME) provide wide-area coverage (∼100 deg²) but are limited to cadences of at most ∼6 hr⁻¹, rendering them photon-starved for high-magnification, short-duration events characteristic of $q \lesssim 10^{-4}$ mass ratios or sub-Earth-mass FFPs. Previous studies (Sumi et al. 2011; Mroz et al. 2017; Gould et al. 2022) hint at a substantial FFP population below 1 $M_{\oplus}$, but Mars-mass ($M \lesssim 0.1 M_{\oplus}$) events—typically with Einstein crossing times $t_E \lesssim 0.5$ hr—remain at the limits of survey detectability. DREAMS’s approach fuses high temporal cadence (1–2 min), millimagnitude photometric precision, and focused Galactic bulge fields, thereby enabling the detection of FFPs down to $\sim$0.03 $M_{\oplus}$ and bound systems at extremely low mass ratios.

2. Instrumentation and Technical Configuration

DREAMS utilizes the Blanco 4 m telescope at CTIO, configured with an f/2.7 prime focus. The Dark Energy Camera (DECam) provides a $3\,\mathrm{deg}^2$ field of view across 62 science CCDs with a pixel scale of $0.263^{\prime\prime}$ per pixel. Observations are conducted in SDSS-like $r$ and $z$ filters, optimizing for the extinction properties and expected source colors in the bulge. These instrumental specifications are critical to supporting minute-cadence, high-precision photometric monitoring across the selected fields (Yang et al., 16 Jan 2026).

3. Survey Design and Observing Strategy

The primary fields encompass $M_{\oplus}$0 in the Galactic bulge, selected for moderate extinction ($M_{\oplus}$1 and $M_{\oplus}$2) and elevated microlensing event rates, with longitude and latitude ranges $M_{\oplus}$3 to $M_{\oplus}$4, $M_{\oplus}$5 to $M_{\oplus}$6. The 2025 "pilot I" campaign (June 29–July 4) implemented five $M_{\oplus}$7-blocks ($M_{\oplus}$8 s exposures) plus one $M_{\oplus}$9-block ($M \lesssim 0.1 M_{\oplus}$0 s) per hour per field, totaling 20 $M \lesssim 0.1 M_{\oplus}$1 and 4 $M \lesssim 0.1 M_{\oplus}$2 exposures hr⁻¹. The "pilot II" (September) run adjusted to four $M \lesssim 0.1 M_{\oplus}$3-blocks ($M \lesssim 0.1 M_{\oplus}$4 s) and one $M \lesssim 0.1 M_{\oplus}$5-block ($M \lesssim 0.1 M_{\oplus}$6 s) per hour. The long-term (2026–2028) baseline calls for two adjacent, partially overlapping fields sampled with four $M \lesssim 0.1 M_{\oplus}$7-blocks ($M \lesssim 0.1 M_{\oplus}$8 s) and one $M \lesssim 0.1 M_{\oplus}$9-block (80 s) per hour over $t_E \lesssim 0.5$030–40 nights per bulge season.

Component	Specification	Purpose/Benefit
Blanco 4 m	f/2.7 prime focus	Large aperture for faint source recovery
DECam	$t_E \lesssim 0.5$1, 62 CCDs, $t_E \lesssim 0.5$2/pixel	Wide field, high spatial resolution
Filters	SDSS-like $t_E \lesssim 0.5$3, $t_E \lesssim 0.5$4	Mitigate extinction, sample source color
Cadence	1–2 min exposures	Capture short-duration FFP events

4. Photometric Reduction, Calibration, and Performance Metrics

DECam raw frames are processed by the NOIRLab DECam Community Pipeline (bias, flat-field, cross-talk, WCS correction). Difference-image analysis (DIA) is conducted via pySIS (for KMTNet and DREAMS) and PRIME’s custom pipeline. For each source $t_E \lesssim 0.5$5 and time $t_E \lesssim 0.5$6:

$t_E \lesssim 0.5$7

where $t_E \lesssim 0.5$8 is the microlensing magnification (computed via VBBinaryLensing), $t_E \lesssim 0.5$9 the source flux, and $\sim$0 blend flux. Error normalization follows Yee et al. (2012): $\sim$1 for non-variable stars.

The empirically determined signal-to-noise relationships are:

• $\sim$2
• $\sim$3

For a 60 s $\sim$4-band exposure at $\sim$5, $\sim$6 ($\sim$7 mag). In the September 2025 pilot, DREAMS achieved $\sim$80.04 mag rms at $\sim$9 and $M_{\oplus}$00.07 mag rms at $M_{\oplus}$1, defining 3$M_{\oplus}$2 finite-source FFP detection limits for duration $M_{\oplus}$310 min at Mars mass for typical bulge parameters, or Moon mass ($M_{\oplus}$4) in favorable (low extinction or overlap) regions.

5. Early Discovery: KMT-2025-BLG-1616Lb

The high-magnification ($M_{\oplus}$5) microlensing event KMT-2025-BLG-1616 was independently detected by KMTNet and DREAMS. KMTNet data presented a short U-shaped anomaly near maximum, but binary-lens model solutions (central versus resonant caustics) were highly degenerate ($M_{\oplus}$6). DREAMS minute-cadence $M_{\oplus}$7 and $M_{\oplus}$8 data resolved this, significantly preferring the "wide-resonant" solution ($M_{\oplus}$9, $3\,\mathrm{deg}^2$0, $3\,\mathrm{deg}^2$1 over central models). DREAMS provided precise source color via time-correlated $3\,\mathrm{deg}^2$2–$3\,\mathrm{deg}^2$3 blocks:

• $3\,\mathrm{deg}^2$4 (DECam)
• $3\,\mathrm{deg}^2$5 (transformed)
• Extinction-corrected $3\,\mathrm{deg}^2$6, $3\,\mathrm{deg}^2$7
• Angular source size $3\,\mathrm{deg}^2$8 μas
• Normalized source radius $3\,\mathrm{deg}^2$9
• Einstein radius $0.263^{\prime\prime}$0 mas, relative proper motion $0.263^{\prime\prime}$1 mas yr⁻¹

Bayesian Galactic-model analysis yields:

• $0.263^{\prime\prime}$2
• $0.263^{\prime\prime}$3
• $0.263^{\prime\prime}$4 kpc
• $0.263^{\prime\prime}$5 au

This constitutes the first bound-planet detection from DREAMS and demonstrates the scientific advantage of high-cadence sampling in degeneracy resolution (Yang et al., 16 Jan 2026).

6. Sensitivity to Free-Floating Planets: Simulations and Thresholds

DREAMS sensitivity to sub-terrestrial-mass FFPs has been quantified through end-to-end simulation of point-lens, finite-source light curves using the post-2025 cadence. For a bulge Mars-mass FFP ($0.263^{\prime\prime}$6, $0.263^{\prime\prime}$7 kpc, $0.263^{\prime\prime}$8 kpc, $0.263^{\prime\prime}$9 mas yr⁻¹), sources at $r$0 mag with photometric noise and extinction ($r$1, $r$2) satisfy detection criteria:

1. At least six data points exceed baseline by %%%%83$t_E \lesssim 0.5$84%%%%
2. $r$5

For a disk Moon-mass case ($r$6, $r$7 kpc, $r$8 kpc, $r$9 mas yr⁻¹, source $z$0), simulated events yield ten points $z$1, $z$2, confirming sensitivity to $z$3 lenses. Cadence overlap within the $z$4 intersection further enhances this capability.

7. Role in Broader Exoplanet Demographics and Prospects

DREAMS fills a niche in exoplanet microlensing by:

• Detecting faint, low-mass bound exoplanets inaccessible to wider, slower surveys;
• Achieving, for the first time, robust sensitivity to Mars- and Moon-mass FFPs;
• Enabling color measurement of extremely faint sources ($z$5) via multi-band, high-cadence data.

Over the 2026–2028 bulge seasons, DREAMS will collect tens of thousands of DECam exposures across $z$610 deg² at minute cadence, facilitating a statistical census of terrestrial-mass and sub-terrestrial-mass rogue planets and extending bound-planet detection toward $z$7. When combined with KMTNet, OGLE, MOA, PRIME, and anticipated space-based missions (Roman, Euclid), DREAMS is poised to support a nearly comprehensive mapping of galactic planetary demographics from gas giants through Moon-mass wanderers (Yang et al., 16 Jan 2026).