WASP-121b: Ultra-Hot Transiting Jupiter

• WASP-121b is an ultra-hot Jupiter that transits an F-type star every 1.27 days, showing extreme tidal distortion and inflated radii.
• Its atmosphere reveals a rich chemical inventory—including water, neutral and ionized metals, and signs of thermal inversion—through diverse spectroscopic techniques.
• High-precision phase curves and high-resolution spectroscopy uncover inefficient heat redistribution, dynamic atmospheric escape, and evidence of refractory condensation.

Searching arXiv for recent and foundational WASP-121b papers to ground the article. WASP-121b is a transiting ultra-hot Jupiter orbiting the F-type star WASP-121 every $1.2749255$ days at a separation of $\sim 0.025$ AU, with a mass near $1.18\,M_{\rm Jup}$ and a radius near $1.8\text{--}1.9\,R_{\rm Jup}$, placing it among the most inflated and intensely irradiated giant planets known (Delrez et al., 2015). Its atmospheric characterization spans transmission spectroscopy, secondary eclipses, full-orbit phase curves, and high-resolution optical and infrared spectroscopy, establishing a dayside thermal inversion, strong non-blackbody emission, water emission in the near-infrared, H$^-$ continuum opacity at short wavelengths, a rich inventory of neutral and ionized metals, and evidence for atmospheric escape (Mikal-Evans et al., 2020). Because WASP-121b lies only $\sim 1.15$ times above its Roche limit and follows a nearly polar orbit, it is also a key system for studies of tidal distortion, orbital evolution, and migration (Delrez et al., 2015).

1. System architecture, discovery, and tidal configuration

WASP-121b was identified as a short-period transiting hot Jupiter orbiting an F6V star with $V = 10.44$, $T_{\rm eff} = 6460 \pm 140$ K, stellar mass $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$, and stellar radius $R_\star = 1.458 \pm 0.030\,R_\odot$ (Delrez et al., 2015). In the discovery analysis, the planet’s mass was measured as $\sim 0.025$0, its Roche-corrected mean radius as $\sim 0.025$1, and its orbital semi-major axis as $\sim 0.025$2 AU (Delrez et al., 2015). Later studies using updated system parameters reported $\sim 0.025$3, $\sim 0.025$4, and $\sim 0.025$5 days, showing that the system’s basic characterization is stable at the level relevant for atmospheric interpretation (Pelletier et al., 2024).

A defining property of WASP-121b is its proximity to the Roche limit. The discovery paper derived $\sim 0.025$6 AU and $\sim 0.025$7, identifying the planet as orbiting only about $\sim 0.025$8 above tidal disruption (Delrez et al., 2015). Roche geometry modeling further showed strong asphericity, with $\sim 0.025$9, $1.18\,M_{\rm Jup}$0, and $1.18\,M_{\rm Jup}$1 (Delrez et al., 2015). This makes WASP-121b one of the most tidally distorted known exoplanets.

The orbit is nearly circular in the discovery analysis, with $1.18\,M_{\rm Jup}$2 at $1.18\,M_{\rm Jup}$3 (Delrez et al., 2015). A later secondary-eclipse timing study derived $1.18\,M_{\rm Jup}$4 from SMARTS K-band data and $1.18\,M_{\rm Jup}$5 from TRAPPIST $1.18\,M_{\rm Jup}$6 data, both consistent with low eccentricity and compatible with a circular orbit within uncertainties (Kovacs et al., 2019). This near-circular configuration is consistent with strong tidal circularization in an ultra-short-period system.

WASP-121b also exhibits an extreme spin-orbit configuration. From the Rossiter-McLaughlin effect, the discovery paper measured a sky-projected obliquity of $1.18\,M_{\rm Jup}$7 deg, implying a nearly polar orbit (Delrez et al., 2015). A plausible implication is that the planet’s migration involved strong dynamical excitation rather than exclusively smooth disk migration, although the exact pathway is not uniquely determined by these data alone.

2. Orbital geometry, deformation, and interior response

The planet’s transits are nearly central. The discovery analysis reported $1.18\,M_{\rm Jup}$8, $1.18\,M_{\rm Jup}$9, and impact parameter $1.8\text{--}1.9\,R_{\rm Jup}$0 (Delrez et al., 2015). A TESS reanalysis using 2-minute cadence found $1.8\text{--}1.9\,R_{\rm Jup}$1, $1.8\text{--}1.9\,R_{\rm Jup}$2, and $1.8\text{--}1.9\,R_{\rm Jup}$3, emphasizing that coarse 30-minute cadence biases the inferred inclination and radius ratio low unless oversampling corrections are applied (Yang et al., 2020). This established WASP-121b as a standard cautionary case for cadence-dependent biases in precision transit analyses.

The same extreme tidal environment that distorts the planet also makes it a candidate for direct constraints on fluid response through Love numbers. An HST/STIS-based study modeled the transit shape of a tidally distorted planet and found a tentative measurement of the quadrupolar radial Love number,

$1.8\text{--}1.9\,R_{\rm Jup}$4

using two STIS G430L transits (Hellard et al., 2019). That work estimated a fractional quadrupolar radial deformation of

$1.8\text{--}1.9\,R_{\rm Jup}$5

for $1.8\text{--}1.9\,R_{\rm Jup}$6, corresponding to a deformation of $1.8\text{--}1.9\,R_{\rm Jup}$7 km (Hellard et al., 2019). The analysis also concluded that the impact of noise modeling on $1.8\text{--}1.9\,R_{\rm Jup}$8 is stronger than the impact of limb-darkening modeling, and that the wavelet method for correlated noise analysis can mask limb brightening (Hellard et al., 2019).

The tentative $1.8\text{--}1.9\,R_{\rm Jup}$9 value is consistent with a fluid, centrally condensed, inflated hot Jupiter (Hellard et al., 2019). This suggests that WASP-121b is a potentially informative target for connecting atmospheric inflation to interior density structure, although that connection remains contingent on tighter photometric constraints than currently available. The STIS study estimated that about 12 complete transits with G430L would be needed to reach the white-noise precision required for a robust $^-$0 measurement (Hellard et al., 2019).

3. Transmission spectrum, terminator chemistry, and cloud asymmetry

Transmission spectroscopy first established WASP-121b as a chemically rich ultra-hot atmosphere. HST/WFC3 transit observations over $^-$1 detected the $^-$2 water band at $^-$3, while reanalysis of ground-based $^-$4, $^-$5, and $^-$6 transit photometry showed significantly deeper optical transits than in the near-infrared (Evans et al., 2016). In that framework, scattering by high-altitude haze alone was found unlikely to explain the optical-to-near-IR contrast, and the spectrum was interpreted as evidence for titanium oxide and vanadium oxide absorption, with possible enhanced opacity from FeH across $^-$7 (Evans et al., 2016). The best-fit transmission model gave relative abundances of TiO/H$^-$8O $^-$9 solar, VO/H$\sim 1.15$0O $\sim 1.15$1 solar, and FeH/H$\sim 1.15$2O $\sim 1.15$3 solar, although the paper stressed that these are relative opacity contributions rather than absolute abundances because of the reference-pressure degeneracy (Evans et al., 2016).

High-resolution optical transmission spectroscopy later revised the status of TiO and VO at the limb. A UVES/VLT search found no evidence for TiO or VO in the terminator spectrum and set rough detection limits of $\sim 1.15$4 and $\sim 1.15$5, while emphasizing that these limits are degenerate with scattering properties and the position of the cloud deck (Merritt et al., 2020). That study concluded that TiO is almost certainly not a major opacity source at the terminator, while VO could not be conclusively ruled out because the available VO line list may not be accurate enough for high-resolution cross-correlation (Merritt et al., 2020). This tension between low-resolution and high-resolution constraints became one of the central interpretive issues in WASP-121b atmospheric studies.

A later ESPRESSO transmission retrieval established consistent relative abundance constraints across multiple epochs and instruments. Neutral metals Fe I, Mg I, Cr I, V I, Na I, and Ca I were strongly detected in all ESPRESSO transits, and refractory abundance ratios such as $\sim 1.15$6, $\sim 1.15$7, and $\sim 1.15$8 were found to be broadly consistent with stellar values (Maguire et al., 2022). The same work inferred upper-atmosphere temperatures of $\sim 1.15$9 K at pressures below about $V = 10.44$0 bar, a net blueshift $V = 10.44$1 km s$V = 10.44$2 for the combined neutral-metal signal, and line broadenings consistent with rotation plus modest additional dynamics (Maguire et al., 2022). These measurements indicate a hot, metal-rich, and dynamically active terminator.

Three-dimensional cloud modeling provides a physical framework for the pronounced asymmetry expected across the limb. A SPARC/MITgcm plus kinetic cloud microphysics study classified WASP-121b as a low-gravity ultra-hot Jupiter with an essentially cloud-free dayside, a cloudy nightside, and a strongly asymmetric terminator in which the morning limb is cloudier and the evening limb more atomic and clear (Helling et al., 2021). In that model, the dayside upper atmosphere reaches $V = 10.44$3 K, the nightside sits near $V = 10.44$4 K, and the mean molecular weight changes from $V = 10.44$5 on the dayside upper atmosphere to $V = 10.44$6 on the nightside and terminators (Helling et al., 2021). The same work found enhanced C/O in cloud-forming regions, a deep thermal ionosphere on the dayside with $V = 10.44$7 approaching unity at very low pressures, and morning-evening terminator asymmetry that should produce ingress/egress asymmetries in transmission (Helling et al., 2021). This suggests that one-dimensional limb retrievals are inherently incomplete for WASP-121b.

4. Dayside emission spectrum and thermal inversion

Dayside emission spectroscopy is the observational basis for the canonical picture of WASP-121b’s thermal inversion. HST/WFC3 G102 eclipse spectroscopy extended the dayside spectrum down to $V = 10.44$8, which, combined with G141, ground-based $V = 10.44$9 and $T_{\rm eff} = 6460 \pm 140$0, and Spitzer/IRAC photometry, yielded a continuous dayside spectrum from $T_{\rm eff} = 6460 \pm 140$1 to $T_{\rm eff} = 6460 \pm 140$2 (Mikal-Evans et al., 2019). In that analysis the full dataset rejected a blackbody at $T_{\rm eff} = 6460 \pm 140$3, with the best-fit blackbody temperature $T_{\rm eff} = 6460 \pm 140$4 K but a poor fit $T_{\rm eff} = 6460 \pm 140$5, while a chemical-equilibrium retrieval with thermal dissociation and ionization provided an excellent fit with reduced $T_{\rm eff} = 6460 \pm 140$6 (Mikal-Evans et al., 2019). The short-wavelength part of the spectrum was interpreted as H$T_{\rm eff} = 6460 \pm 140$7 emission shortward of $T_{\rm eff} = 6460 \pm 140$8, and the dayside thermal profile was found to rise from $T_{\rm eff} = 6460 \pm 140$9 K at $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$0 mbar to $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$1 K at $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$2 mbar (Mikal-Evans et al., 2019).

This picture was sharpened by improved G141 eclipse measurements. Four new HST/WFC3/G141 secondary eclipses, combined with reanalysis of the original eclipse, achieved a median precision of $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$3 ppm across 28 spectroscopic channels spanning $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$4, compared with $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$5 ppm in the earlier single-eclipse dataset (Mikal-Evans et al., 2020). The joint white-light fit yielded

$M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$6

and the updated spectrum clearly resolved the water emission band at $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$7 (Mikal-Evans et al., 2020). Using the full optical-to-IR dataset, blackbody fits with the B19 and D19 TESS points gave $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$8 K and $M_\star = 1.353^{+0.080}_{-0.079}\,M_\odot$9 K, respectively, but were rejected at $R_\star = 1.458 \pm 0.030\,R_\odot$0 and $R_\star = 1.458 \pm 0.030\,R_\odot$1 (Mikal-Evans et al., 2020). The emission spectrum therefore requires non-isothermal structure and wavelength-dependent opacity.

The water emission band itself is a central diagnostic. In the updated G141 spectrum, eclipse depths rise from about $R_\star = 1.458 \pm 0.030\,R_\odot$2 ppm near $R_\star = 1.458 \pm 0.030\,R_\odot$3 to $R_\star = 1.458 \pm 0.030\,R_\odot$4 ppm around $R_\star = 1.458 \pm 0.030\,R_\odot$5, including $R_\star = 1.458 \pm 0.030\,R_\odot$6 ppm at $R_\star = 1.458 \pm 0.030\,R_\odot$7 and $R_\star = 1.458 \pm 0.030\,R_\odot$8 ppm at $R_\star = 1.458 \pm 0.030\,R_\odot$9 (Mikal-Evans et al., 2020). Because eclipse depths in the band exceed those in adjacent continuum regions, the water band is in emission rather than absorption, implying that the temperature increases upward over the pressures where water lines form, roughly $\sim 0.025$00 mbar (Mikal-Evans et al., 2020).

The retrieval framework used in that study parameterized the thermal profile with the ATMO semi-analytic radiative-equilibrium model, with free parameters $\sim 0.025$01, $\sim 0.025$02, and $\sim 0.025$03, plus elemental abundance parameters $\sim 0.025$04, $\sim 0.025$05, and $\sim 0.025$06 (Mikal-Evans et al., 2020). In the favored Case 3 retrieval, the marginalized posteriors gave

$\sim 0.025$07

with reduced $\sim 0.025$08 (Mikal-Evans et al., 2020). The retrieved PT profiles rose from $\sim 0.025$09 K at $\sim 0.025$10 mbar to $\sim 0.025$11 K at $\sim 0.025$12 mbar, and the large $\sim 0.025$13 confirmed that visible opacity exceeds infrared opacity, the canonical condition for a stratosphere-like inversion (Mikal-Evans et al., 2020).

The interpretation of the $\sim 0.025$14 feature changed decisively with the improved dataset. The original single-eclipse spectrum contained a localized bump around $\sim 0.025$15, tentatively attributed to VO emission, but the combined five-eclipse spectrum no longer shows a bump there; the $\sim 0.025$16 channel has an eclipse depth $\sim 0.025$17 ppm consistent with neighboring bins (Mikal-Evans et al., 2020). The authors concluded that the $\sim 0.025$18 bump was either a statistical fluctuation or a systematic artefact specific to the original dataset (Mikal-Evans et al., 2020). This substantially weakened the earlier case for VO emission at that wavelength, even though VO remained relevant in broader discussions of inversion drivers.

5. Phase curves, dayside–nightside contrast, and circulation

Optical and infrared phase curves show that WASP-121b has very inefficient longitudinal heat redistribution. The TESS optical phase curve yielded a secondary-eclipse depth of $\sim 0.025$19 ppm in one analysis, dayside brightness temperature $\sim 0.025$20 K, and nightside brightness temperature $\sim 0.025$21 K, with the hotspot located at the substellar point and $\sim 0.025$22, consistent with negligible hotspot offset (Bourrier et al., 2019). A later full-orbit TESS analysis that simultaneously modeled primary transit, secondary eclipse, thermal emission, reflection, and ellipsoidal variation found a secondary eclipse depth of $\sim 0.025$23 ppm, dayside temperature $\sim 0.025$24 K, nightside temperature $\sim 0.025$25 K, and geometric albedo $\sim 0.025$26 (Eftekhar, 2022). Both studies concluded that heat transport from dayside to nightside is inefficient and that the optical emission is dominated by thermal flux rather than reflection.

Spitzer/IRAC phase curves extended this picture into the mid-infrared. At $\sim 0.025$27, the phase amplitude is $\sim 0.025$28 ppm and the phase offset is $\sim 0.025$29; at $\sim 0.025$30, the phase amplitude is $\sim 0.025$31 ppm and the offset is $\sim 0.025$32 (Davenport et al., 16 Mar 2025). The corresponding dayside brightness temperatures are $\sim 0.025$33 K and $\sim 0.025$34 K, while the nightside temperatures are $\sim 0.025$35 K and $\sim 0.025$36 K (Davenport et al., 16 Mar 2025). These large day–night contrasts, combined with near-zero phase offsets, indicate a hot dayside, a much cooler nightside, and minimal longitudinal displacement of the thermal hotspot.

Comparisons with general circulation models point to magnetic drag as the leading explanation for these phase-curve properties. In the RM-GCM simulations, models without magnetic effects develop strong eastward equatorial jets and significant hotspot shifts, whereas models with a prescribed $\sim 0.025$37 G magnetic field introduce a drag term,

$\sim 0.025$38

that weakens dayside zonal flow, suppresses the hotspot shift, and increases day–night contrast (Davenport et al., 16 Mar 2025). The paper concluded that magnetic drag is required to match Spitzer’s low offsets and large amplitudes (Davenport et al., 16 Mar 2025). This suggests that WASP-121b is in the regime where radiative timescales are short and the hot, partially ionized dayside couples sufficiently to magnetic fields to inhibit efficient recirculation.

Single-band eclipse measurements are consistent with the same physical picture. A SMARTS 1.3 m K-band study measured

$\sim 0.025$39

with $\sim 0.025$40, and found that the K, $\sim 0.025$41, and $\sim 0.025$42 depths all favor a hot dayside with poor heat transport (Kovacs et al., 2019). A blackbody with $\sim 0.025$43 K could approximately fit the three single-band points, but the full HST spectrum required more involved atmosphere models with species producing emission and absorption features rather than smooth blackbody emission (Kovacs et al., 2019). This established that inefficient redistribution and non-blackbody emission were already evident before the more detailed phase-curve and multi-eclipse analyses.

6. High-resolution spectroscopy, metals, escape, and composition

High-resolution spectroscopy revealed that WASP-121b’s atmosphere contains an unusually rich inventory of neutral and ionized species. A reanalysis of ESPRESSO transit data confirmed H, Li, Na, Mg, Ca, Ca$\sim 0.025$44, V, Cr, Mn, Fe, Fe$\sim 0.025$45, and Ni, and newly detected Ba$\sim 0.025$46, Co, and Sr$\sim 0.025$47, with a tentative detection of Ti$\sim 0.025$48 (Silva et al., 2022). Ba$\sim 0.025$49 was identified as the heaviest detected element in any exoplanet atmosphere to date, with amplitudes $\sim 0.025$50 ppm and $\sim 0.025$51 ppm on the two nights analyzed, and detection significances $\sim 0.025$52 and $\sim 0.025$53 (Silva et al., 2022). These detections show that the upper atmosphere is both extremely hot and strongly ionized.

The same ESPRESSO analysis highlighted the especially extended and asymmetric Ca$\sim 0.025$54 signal. Ca II H&K absorption is strongly blueshifted, with fitted centers at $\sim 0.025$55 km s$\sim 0.025$56 and $\sim 0.025$57 km s$\sim 0.025$58, and very large widths of $\sim 0.025$59 km s$\sim 0.025$60 and $\sim 0.025$61 km s$\sim 0.025$62 (Silva et al., 2022). Using the relation

$\sim 0.025$63

the authors inferred effective Ca$\sim 0.025$64 radii of $\sim 0.025$65 and $\sim 0.025$66, beyond the Roche lobe, and concluded that the signal may arise from atmospheric escape (Silva et al., 2022). Strong and broad H$\sim 0.025$67 absorption, with amplitudes $\sim 0.025$68 ppm and $\sim 0.025$69 ppm on the two nights, supports the picture of an extended upper atmosphere undergoing mass loss (Silva et al., 2022).

A dedicated high-resolution transmission retrieval quantified the extent of several escaping species. Using ESPRESSO and a model-filtering framework, one study derived effective altitudes of $\sim 0.025$70 for H$\sim 0.025$71, $\sim 0.025$72 for Fe II, and $\sim 0.025$73 for Ca II, with the transit equivalent Roche limit at $\sim 0.025$74 (Maguire et al., 2022). H$\sim 0.025$75 and especially Ca II therefore extend beyond the Roche limit, confirming that at least part of the upper atmosphere is escaping (Maguire et al., 2022). The same study found that neutral-metal abundance ratios are stable across multiple epochs and instruments, implying that the lower thermosphere and terminator composition are not strongly variable on timescales of years (Maguire et al., 2022).

High-resolution dayside emission spectroscopy brought the composition question into a volatile-versus-refractory framework. Combined ESPRESSO and CRIRES+ observations directly measured C and O as volatile elements and Fe and Ni as refractory elements, obtaining

$\sim 0.025$76

and $\sim 0.025$77, corresponding to a volatile-to-refractory enrichment of about $\sim 0.025$78 times the stellar value (Pelletier et al., 2024). The same study measured $\sim 0.025$79 in chemical equilibrium and $\sim 0.025$80 in the hybrid free retrieval, with TiO/VO metallicity constrained to $\sim 0.025$81 at $\sim 0.025$82 (Pelletier et al., 2024). These results imply a volatile-rich, refractory-poor atmospheric composition relative to the host star and were interpreted as favoring formation near the CO snowline followed by inward migration (Pelletier et al., 2024). This suggests that WASP-121b can be used not only as an atmospheric physics benchmark but also as a tracer of giant-planet formation history.

7. JWST-era synthesis and current interpretation

Broadband JWST dayside emission spectroscopy pushed WASP-121b into a new regime of characterization. A panchromatic $\sim 0.025$83 dayside spectrum from archival JWST/NIRISS and NIRSpec/G395H observations yielded statistically significant detections of H$\sim 0.025$84O at $\sim 0.025$85, CO at $\sim 0.025$86, SiO at $\sim 0.025$87, TiO at $\sim 0.025$88, and VO at $\sim 0.025$89, together with a robust $\sim 0.025$90 detection of Titanate $\sim 0.025$91 clouds (Saha et al., 27 Aug 2025). The retrieved free-chemistry abundances implied a super-solar $\sim 0.025$92, a sub-solar $\sim 0.025$93, and a metallicity of $\sim 0.025$94 solar (Saha et al., 27 Aug 2025). The study also found strong evidence of TiO depletion relative to equilibrium, interpreted as sequestration into refractory condensates such as Titanate clouds (Saha et al., 27 Aug 2025).

This result modifies earlier debates about TiO and VO. In transmission at the terminator, high-resolution observations did not detect TiO or VO (Merritt et al., 2020). In earlier low-resolution dayside data, VO-like explanations were weakened when the $\sim 0.025$95 bump disappeared in the combined five-eclipse G141 spectrum (Mikal-Evans et al., 2020). The JWST dayside analysis, however, reported both TiO and VO in emission and simultaneously found evidence that at least part of the Ti reservoir is depleted into CaTiO$\sim 0.025$96 clouds (Saha et al., 27 Aug 2025). A plausible implication is that the apparent contradictions between datasets reflect geometry, vertical structure, and condensation chemistry rather than a single globally uniform abundance field. The limb, dayside, and high-altitude escaping regions need not share the same molecular inventory.

Even before the JWST panchromatic result, WASP-121b had already emerged as a prototype for the ultra-hot Jupiter regime: a clear dayside inversion, H$\sim 0.025$97-dominated short-wavelength continuum, water and CO in emission, strong day–night asymmetry, poor heat redistribution, extensive metal chemistry, and a dynamically escaping upper atmosphere (Mikal-Evans et al., 2019). The new cloud detection adds refractory condensation directly to that picture, indicating that even on an ultra-hot dayside, condensate formation can be spectroscopically important (Saha et al., 27 Aug 2025).

The main unresolved issues now concern the exact identity of the inversion drivers, the three-dimensional distribution of clouds and molecules, and the connection between present-day atmospheric composition and formation pathway. High-resolution optical and infrared spectroscopy has already shown that volatile and refractory elements can be constrained simultaneously (Pelletier et al., 2024), while cloud microphysics models predict strong morning-evening asymmetries and deep ionospheres (Helling et al., 2021). This suggests that future progress on WASP-121b will depend on combining panchromatic emission data, limb-resolved transmission constraints, phase-resolved spectroscopy, and multidimensional retrieval frameworks rather than relying on one-dimensional interpretations alone.