SPACE: Sub-Neptune Atmosphere Study

• SPACE is a comparative exoplanet-atmosphere experiment targeting sub-Neptunes that lack Solar System analogues.
• It combines HST/WFC3 and STIS near-infrared and ultraviolet observations with high-resolution ground-based spectroscopy to uniformly address atmospheric composition and cloud properties.
• The program’s findings inform on atmospheric evolution, volatile retention, and the role of aerosols in shaping sub-Neptune physical and chemical diversity.

Sub-neptune Planetary Atmosphere Characterization Experiment (SPACE) is a comparative exoplanet-atmosphere program focused on sub-Neptunes, a planet class with no Solar System analogue and persistent mass-radius-composition degeneracies. In its 2025 programmatic form, SPACE is described as a Hubble treasury survey targeting eight sub-Neptunes across planet radius $2\text{--}3.5\,R_\oplus$ and equilibrium temperature $300\text{--}1400$ K, combining HST/WFC3 near-infrared transmission spectroscopy with HST/STIS ultraviolet host-star characterization to map how atmospheric composition and aerosols depend on size, temperature, and irradiation (Kahle et al., 17 Jul 2025). In a later mission-concept formulation, SPACE is also presented as a uniform, large-sample transmission-spectroscopy survey for the Nautilus Space Observatory, motivated by the view that sub-Neptune atmospheric diversity is intrinsically a population-level problem that requires sample size and a uniform observing strategy (Welbanks et al., 30 Jun 2026).

1. Scientific rationale and scope

SPACE is motivated by the fact that sub-Neptunes are the most abundant type of planet known today, yet they do not have a Solar System counterpart, leaving major open questions about their composition, formation, cloud physics, atmospheric escape, and evolutionary pathways (Kahle et al., 17 Jul 2025). The same radius and mass can correspond to rocky or metal-rich bodies wrapped in substantial H/He envelopes, or to water-rich worlds with only thin H/He atmospheres, so atmospheric composition is treated as the key observable for breaking interior degeneracies (Kumar et al., 26 Jan 2026).

The programmatic logic is therefore comparative rather than purely single-target. Published SPACE discussions ask whether sub-Neptunes form a single continuous family or several distinct atmospheric classes, including H$_2$-rich envelopes, H$_2$/H$_2$O stratified interiors, well-mixed metal-rich atmospheres, steam worlds, and Hycean-like planets (Welbanks et al., 30 Jun 2026). They also frame clouds and hazes as a population problem: previous studies suggested a temperature-linked aerosol trend, yet hot and temperate sub-Neptunes already show mutually inconsistent atmospheric outcomes, so a uniform survey is needed to determine whether there are real transition boundaries in chemistry, cloudiness, and escape (Kahle et al., 17 Jul 2025).

A second scientific axis is temporal evolution. SPACE-relevant work repeatedly ties atmospheric characterization to the radius valley, the Neptune desert, and the first gigayear of planetary evolution, where photoevaporation, core-powered mass loss, and migration are expected to sculpt the sub-Neptune population (Orell-Miquel et al., 2023). This suggests that SPACE is not only a spectroscopy program, but also an empirical framework for linking observed atmospheres to the formation and erosion histories of small planets.

2. Target regime and benchmark planets

The published target set emphasizes transiting sub-Neptunes whose host stars and orbital geometries make transmission spectroscopy efficient. Nearby M dwarfs are especially prominent because small stellar radii produce large transit depths and favorable atmospheric signal amplitudes, but bright hotter hosts also appear when the planetary scale height or irradiation state is scientifically diagnostic (Kawauchi et al., 2022). The selected systems span temperate planets near habitable-zone boundaries, hot planets near the radius valley or Neptune desert, and transitional worlds whose bulk properties already imply retained volatiles but whose atmospheric composition remains unconstrained (Timmermans et al., 2024).

Several benchmark targets illustrate the range of the SPACE landscape.

System	Reported properties	SPACE relevance
TOI-2136b	$d=33.361\pm0.019$ pc, M3 dwarf host, $J=10.184\pm0.024$, $R_p=2.20\pm0.07\,R_\oplus$, $P=7.851925\pm0.000016$ d, TSM $\sim 93$	Strong target for atmospheric work; potentially a hycean world (Kawauchi et al., 2022)
TOI-4336 A b	$300\text{--}1400$0, $300\text{--}1400$1 d, $300\text{--}1400$2 K, $300\text{--}1400$3 pc, $300\text{--}1400$4, TSM $300\text{--}1400$5	Temperate sub-Neptune near the inner edge of the empirical habitable zone (Timmermans et al., 2024)
HD 77946 b	$300\text{--}1400$6, $300\text{--}1400$7, $300\text{--}1400$8 K, TSM $300\text{--}1400$9	Compositionally ambiguous benchmark around a bright F5 star (Palethorpe et al., 2024)
TOI-4602 b	$_2$0 d, $_2$1, $_2$2, $_2$3 K, TSM $_2$4	Transitional sub-Neptune just above the radius valley (Maio et al., 13 Apr 2026)

This target space is not homogeneous in host-star type or insolation. Instead, it is physically stratified. Temperate planets such as TOI-4336 A b and TOI-270d probe small, cool atmospheres in regimes relevant to Hycean or water-rich interpretations, while hotter systems such as HD 86226 c, TOI-4602 b, and TOI-5800 b probe metal enrichment, refractory clouds, tidal heating, and atmospheric escape (Mikal-Evans et al., 2022). A plausible implication is that SPACE is designed to sample parameter space where distinct atmospheric processes should dominate rather than to maximize signal-to-noise alone.

3. Observational architecture and validation chain

SPACE-relevant studies use a recurring multi-instrument architecture that couples planet validation to atmospheric follow-up. In the validation of TOI-2136b, the chain consisted of TESS photometry in Sectors 26 and 40, ground-based multicolor photometry with MuSCAT, MuSCAT2, MuSCAT3, and LCO/Sinistro, high-resolution radial velocities from Subaru/IRD, and Gemini North/‘Alopeke speckle imaging; this combination established the object as a bona fide planet rather than an eclipsing binary or blended false positive (Kawauchi et al., 2022). Comparable validation logic appears in TOI-4336 A b, where TESS, resolved multi-band photometry, spectroscopy, speckle imaging, and archival images were required because of the hierarchical triple M-dwarf architecture (Timmermans et al., 2024).

For space-based atmospheric measurements, the core SPACE configuration is HST/WFC3 transmission spectroscopy plus HST/STIS ultraviolet characterization of the host star (Kahle et al., 17 Jul 2025). WFC3 provides the near-infrared transmission spectrum, while STIS constrains the stellar ultraviolet environment relevant to photochemistry, haze production, and atmospheric escape. TOI-270d exemplifies this mode: HST/WFC3 G141 covered $_2$5, and HST/STIS G140M provided a $_2$6 credible upper limit of $_2$7 for stellar Ly$_2$8 emission (Mikal-Evans et al., 2022).

Ground-based high-resolution transmission spectroscopy is a complementary branch of the SPACE methodology. For helium escape searches, Subaru/IRD and CARMENES target the He I infrared triplet at $_2$9 Å, $_2$0 Å, and $_2$1 Å, with telluric $_2$2 correction, masking of $_2$3 emission, rest-frame alignment, and co-addition of in-transit spectra (Kawauchi et al., 2022). For temperate sub-Neptunes, Gemini-South/IGRINS extends the methodology to molecular high-resolution cross-correlation spectroscopy in the H and K bands over $_2$4 at $_2$5, with SVD detrending and cross-correlation against H$_2$6O, CH$_2$7, and NH$_2$8 templates (Cabot et al., 2024). In that framework, the cross-correlation function is written as

$_2$9

and the expected planetary velocity as

$_2$0

This combination of validation, low-resolution transmission spectroscopy, ultraviolet stellar characterization, and high-resolution line-resolved spectroscopy gives SPACE a multi-scale observational structure. It also means that non-detections are interpretable in context, because stellar multiplicity, activity, and planetary bulk parameters are constrained before spectral interpretation is attempted.

4. Empirical atmospheric results

The first dedicated SPACE results paper, on HD 86226 c, reported a featureless $_2$1 transmission spectrum consistent within $_2$2 with a constant transit depth of $_2$3 ppm (Kahle et al., 17 Jul 2025). Its amplitude is only $_2$4 scale heights for a H/He-dominated atmosphere, which excludes a cloud-free solar-metallicity atmosphere at $_2$5. The paper finds that the flat spectrum can be explained either by a cloudless atmosphere with $_2$6 at $_2$7, or by refractory clouds of MgSiO$_2$8, Fe, or MnS. Because the planet is too hot for methane-based organic haze formation to be efficient, the result explicitly challenges the previously inferred aerosol trend for sub-Neptunes.

Other SPACE-relevant targets show resolved molecular structure rather than flat spectra. For TOI-270d, the combined TESS and HST/WFC3 transmission spectrum provides $_2$9 evidence for molecular absorption relative to a featureless spectrum and $d=33.361\pm0.019$0 evidence for H$d=33.361\pm0.019$1O, although the H$d=33.361\pm0.019$2O significance drops to $d=33.361\pm0.019$3 when stellar heterogeneity is included and only WFC3 is used (Mikal-Evans et al., 2022). The retrieved water abundance is approximately $d=33.361\pm0.019$4, and the atmosphere is described as hydrogen-rich rather than steam-dominated. For GJ 3470 b, a combined Hubble/Spitzer analysis of 12 transits and 20 eclipses yielded a $d=33.361\pm0.019$5 water detection, a low-metallicity hydrogen-dominated atmosphere with $d=33.361\pm0.019$6 solar, strong methane depletion with $d=33.361\pm0.019$7 at $d=33.361\pm0.019$8 confidence, and a sharp drop in cloud opacity around $d=33.361\pm0.019$9 interpreted as Mie scattering by particles of $J=10.184\pm0.024$0 (Benneke et al., 2019).

Helium escape results are similarly heterogeneous. HD 235088 b shows a robust He I detection with absorption $J=10.184\pm0.024$1, equivalent width $J=10.184\pm0.024$2 mÅ, and blueshift $J=10.184\pm0.024$3, and the preferred 1D hydrodynamic interpretation implies $J=10.184\pm0.024$4 K and $J=10.184\pm0.024$5 in the photon-limited escape regime (Orell-Miquel et al., 2023). By contrast, TOI-2136b shows no statistically significant helium absorption: the $J=10.184\pm0.024$6 confidence limits are $J=10.184\pm0.024$7 mÅ and absorption $J=10.184\pm0.024$8, while the only tentative feature near $J=10.184\pm0.024$9 Å is at the $R_p=2.20\pm0.07\,R_\oplus$0 level and is not considered robust (Kawauchi et al., 2022). Within the SPACE framework, this contrast is important because it shows that atmospheres can be both spectroscopically accessible and physically diverse even among small planets of similar radius.

5. Interpretation, degeneracies, and classification

A recurring conclusion in SPACE-relevant work is that bulk density alone does not determine atmospheric architecture. HD 77946 b is a clear example: its $R_p=2.20\pm0.07\,R_\oplus$1 and $R_p=2.20\pm0.07\,R_\oplus$2 are consistent either with a sub-Neptune with a $R_p=2.20\pm0.07\,R_\oplus$3 H/He atmosphere or with a water-world composition, and the paper explicitly states that a strong degeneracy exists between water-world and silicate/iron-hydrogen models (Palethorpe et al., 2024). In TOI-2141 b, refined mass and radius measurements favor “a significant volatile envelope atop an Earth-like core,” while PASTA modeling under an H-dominated atmosphere implies that the planet likely lost only about $R_p=2.20\pm0.07\,R_\oplus$4 of its total mass over its lifetime as atmospheric escape; the atmosphere is therefore inferred to be retained, but not yet directly characterized (Luque et al., 31 Aug 2025).

Theoretical modeling further shows that even spectral interpretation is parameter-degenerate. In a self-consistent PICASO+VULCAN grid for K2-18b analogs, varying $R_p=2.20\pm0.07\,R_\oplus$5 from $R_p=2.20\pm0.07\,R_\oplus$6 to $R_p=2.20\pm0.07\,R_\oplus$7 K and $R_p=2.20\pm0.07\,R_\oplus$8 from $R_p=2.20\pm0.07\,R_\oplus$9 to $P=7.851925\pm0.000016$0 significantly changes CH$P=7.851925\pm0.000016$1, CO$P=7.851925\pm0.000016$2, CO, NH$P=7.851925\pm0.000016$3, and HCN abundances, with H$P=7.851925\pm0.000016$4O comparatively unaffected (Kumar et al., 26 Jan 2026). The paper’s central claim is that single-parameter assumptions can misclassify planetary interiors. Its diagnostic logic is molecule-specific: CH$P=7.851925\pm0.000016$5 and NH$P=7.851925\pm0.000016$6 trace thermochemical state and quench depth, CO$P=7.851925\pm0.000016$7 identifies intermediate thermal regimes, CO traces oxidized carbon favored by hot interiors, HCN is a strong disequilibrium tracer, and H$P=7.851925\pm0.000016$8O acts mainly as a bulk composition anchor.

Broad wavelength coverage changes retrieval outcomes as well. In a $P=7.851925\pm0.000016$9 synthesis of HST and JWST transit spectra for K2-18b, hydrocarbon haze models support an H$\sim 93$0-dominated mini-Neptune atmosphere with mean molecular weight $\sim 93$1 Daltons and lower inferred CH$\sim 93$2 and CO$\sim 93$3 abundances than haze-free studies (Liu et al., 13 Sep 2025). The paper argues that haze reduces the need for high-$\sim 93$4 solutions and that stellar-parameter uncertainties propagate directly into planetary density and gravity. This suggests that SPACE classification is not a binary clear-versus-cloudy exercise; rather, it is a coupled inference problem involving stellar characterization, gravity, temperature structure, transport, and aerosol microphysics.

6. Evolutionary context and future forms of SPACE

SPACE-style science extends from mature benchmark systems to planets observed during active evolution. V1298 Tau b, at $\sim 93$5 Myr, shows a $\sim 93$6 water-vapor detection, an inferred scale height of $\sim 93$7 km, a mass upper limit of $\sim 93$8, a density upper limit of $\sim 93$9, and a retrieved atmospheric metallicity of $300\text{--}1400$00 solar (Barat et al., 2023). The same study reports no methane feature near $300\text{--}1400$01, with $300\text{--}1400$02 volume mixing ratio constrained to $300\text{--}1400$03 or below, and interprets this as evidence for a warm, strongly mixed interior-atmosphere system rather than simple equilibrium chemistry. In SPACE terms, this is a progenitor case: a young, low-density planet whose observable atmosphere may evolve into the mature Neptune/sub-Neptune population.

At the opposite end of the evolutionary spectrum, TOI-5800 b is presented as a sub-Neptune likely actively undergoing tidal migration into the evaporation desert (Jenkins et al., 15 May 2025). Its orbital period is $300\text{--}1400$04 d, eccentricity $300\text{--}1400$05, equilibrium temperature $300\text{--}1400$06 K, and TSM $300\text{--}1400$07. The received flux varies by a factor of $300\text{--}1400$08 over the eccentric orbit, the estimated tidal luminosity is $300\text{--}1400$09, and energy-limited plus PASTA-based estimates give mass-loss rates from $300\text{--}1400$10 to $300\text{--}1400$11. This makes the system a direct test of how migration, tidal heating, and photoevaporation interact in shaping close-in sub-Neptunes.

The future, explicitly named SPACE architecture appears in the Nautilus concept papers. There, the program becomes a sub-Neptune population survey with four primary objectives: determining the timescales over which planets evolve into sub-Neptunes and super-Earths; tracking the temporal evolution of atmospheric mass-loss rates; characterizing the evolution of atmospheric mean molecular weight and C/O ratio; and identifying the emergence of helium-dominated worlds (Pascucci et al., 24 Jun 2026). The proposed survey scales with mission class: about $300\text{--}1400$12 sub-Neptunes for a Probe configuration, several hundred for a Flaglet configuration, and $300\text{--}1400$13 for a Flagship configuration (Welbanks et al., 30 Jun 2026). Required wavelength coverage is described as $300\text{--}1400$14 with a high-value extension to $300\text{--}1400$15, baseline spectral resolution $300\text{--}1400$16, and an optional $300\text{--}1400$17 mode for He I $300\text{--}1400$18 and related line-resolved work.

Taken together, these strands define SPACE less as a single instrument program than as a coherent experimental strategy for sub-Neptune atmospheres. Its present form uses comparative HST spectroscopy, stellar-UV characterization, and high-resolution ground-based follow-up to establish that sub-Neptunes can exhibit flat spectra, hydrogen-rich transmission features, refractory clouds, methane depletion, hydrocarbon hazes, and both detectable and undetectable helium escape. Its projected future form extends the same logic to statistical population mapping, with the explicit aim of turning atmospheric diversity into a calibrated framework for planetary classification, interior inference, and evolutionary reconstruction.