Phosphine Gas in the Cloud Decks of Venus

The study investigates whether Venus’ atmosphere contains phosphine (PH₃), a highly reducing gas that on rocky planets is expected to be scarce and rapidly destroyed in oxidizing environments. PH₃ can be a potential biosignature gas because on Earth it is associated with anthropogenic and microbial activity and is not known to be produced in significant quantities abiotically in oxidized settings. Previous solar system explorations have identified intriguing trace gases on other bodies, but often with ambiguous or inaccessible sources. The authors target the PH₃ 1–0 rotational transition in the millimeter regime, motivated by the idea that any detectable PH₃ in a rocky-planet atmosphere could indicate unusual chemistry or possibly life. The research question: Does Venus’ cloud deck atmosphere contain PH₃, and if so, can known abiotic processes account for it? The purpose is to detect and quantify PH₃, assess spectral contamination, and evaluate all plausible abiotic production pathways. The significance lies in understanding Venusian atmospheric chemistry, potential geochemical or photochemical processes, and implications for biosignature searches.

The paper frames PH₃ as a proposed biosignature gas for rocky planets due to its association with life on Earth and lack of known abiotic production in oxidizing environments. It contrasts this with PH₃ in gas giants, where it is produced under high temperature and pressure and brought upward by convection. Prior studies have reported trace gases like methane on Mars with ambiguous sources and organics in icy moon plumes, highlighting challenges in interpretation. Venus’ potential for aerial habitability has been discussed for decades, with speculative microbial life in temperate cloud layers. Spectroscopic challenges exist as many PH₃ features are absorbed by Earth’s atmosphere. The work builds on prior PH₃ detections in Saturn and theoretical studies on PH₃ as a biosignature, as well as photochemical and thermodynamic models of Venus’ atmosphere.

Observations and data reduction:

• Telescopes and transition: Targeted PH₃ J=1–0 rotational transition near 266.9445 GHz (λ≈1.123 mm). Initial observations with the James Clerk Maxwell Telescope (JCMT) over 5 mornings in June 2017; follow-up with the Atacama Large Millimeter/submillimeter Array (ALMA) in March 2019.
• JCMT: Single-dish, whole-planet spectra (limb down-weighted ~50%). Major limitation was spectral baseline ripple from reflections. Employed multi-step polynomial baseline removal: fitted low-order polynomials to remove broad ripples, then addressed higher-frequency ripples; finally interpolated a polynomial across a velocity interval around the expected line to avoid overfitting out the absorption. Explored interpolation windows (|Δv|=2–8 km/s) to bracket systematics; adopted Δv=5 km/s for mid-range solution. Co-added 140 spectra, produced line-to-continuum (l:c) spectra and per-channel uncertainties.
• ALMA: Interferometric imaging and spectral extraction. To mitigate baseline-dependent spectral ripples due to Venus’ brightness and large angular size, flagged all baselines <33 m, improving dynamic range but causing spatial filtering of smooth, large-scale absorption. Performed standard calibration with Callisto as bandpass/flux calibrator, time-dependent phase self-calibration on continuum channels, and continuum subtraction. Extracted spectra as zonal averages (equator, mid-latitudes, poles) to cancel residual ripples. Applied 12th-order polynomial baseline fits across ±40 km/s with interpolation over ±5 km/s around line center. Verified robustness via alternative lower-order fits and by checking adjacent bandpass regions. Also analyzed simultaneous wideband data for line reproducibility and searched for other species (e.g., SO₂ lines) to constrain contamination.
• Radiative transfer and abundance retrieval: Spherical, multi-layer model using Venus International Reference Atmosphere (VIRA) temperature/pressure profiles; included line-by-line opacity (CO₂ continuum, molecular lines) and geometry (limb path lengths). Modeled disk-averaged spectra for JCMT and ALMA beams. Assumed vertically constant PH₃ mixing ratio. Uncertainty dominated by unknown CO₂ pressure broadening for PH₃ 1–0; considered range 0.186–0.286 cm⁻¹/atm (theoretical PH₃-based estimate to NH₃ proxy), propagating to abundance range.
• Photochemical and thermochemical analysis: Built a 1D photochemistry–diffusion model (ARGO code) including H/C/N/O chemistry and added S/Cl/P networks relevant to Venus, with PH₃ reactions (with H, O, OH, Cl), photolysis, and thermal decomposition using Lindemann kinetics. Included an empirical ‘mysterious absorber’ in UV transport and cloud SO₂ rainout approximation. Validated model against observed species profiles. Estimated PH₃ lifetimes versus altitude and required source fluxes to sustain observed levels.
• Assessment of abiotic sources: Evaluated photochemical production via a forward-only kinetic network from oxidized P species to PH₃ using known and homologous nitrogen reaction rates; compared maximum production to destruction rates. Thermodynamic feasibility assessed for atmospheric/surface reactions and subsurface redox (oxygen fugacity) conditions over broad T–P–composition space. Quantified contributions from lightning, volcanism, meteoritic delivery, and other energetic processes.
• Line identification and contamination checks: Compared JCMT and ALMA spectra over common passbands pre-final baseline fitting to rule out coincident artefacts; tested polynomial order for artefact line creation; searched for nearby species lines (notably SO₂ J=31 transition offset by +1.3 km/s) in simultaneous ALMA data; modeled maximum SO₂ contribution consistent with non-detections of favorable SO₂ lines.

• Spectral detections: JCMT co-added spectrum shows candidate PH₃ 1–0 absorption with S/N ~3–7 depending on processing window; ALMA confirms absorption at the PH₃ wavelength with S/N ~13 for whole-planet spectrum.
  • JCMT (whole planet): l:c = −2.5 ± 0.8 ×10⁻⁴; centroid −0.2 ± 1.1 km/s; FWHM 3.6 ± 1.2 km/s; S/N up to 6.7 under alternate processing.
  • ALMA (whole planet): l:c = −0.87 ± 0.11 ×10⁻⁴; centroid +0.7 ± 0.3 km/s; FWHM 4.1 ± 0.5 km/s; S/N 13.3. Zonal spectra: mid-latitudes stronger (−1.26 ± 0.14 ×10⁻⁴; S/N 14.5), equator weaker (−0.39 ± 0.14 ×10⁻⁴; S/N 5.0), poles undetected (3σ limit −0.29 ×10⁻⁴ in 10 km/s bins).
  • Velocity agreement: Line centroids consistent with Venus’ velocity within ~0.1–0.7 km/s systematics; coincidence requirement for alternative species would demand rest-frequency match to within ~10⁻⁶.
• Consistency and artefact checks: Independent facilities and reduction methods yield comparable line depths and widths; overlap of pre-final-processed JCMT and ALMA spectra across the full passband shows only the PH₃-candidate feature coincident; polynomial fitting does not produce spurious PH₃-like lines when applied away from the expected position.
• Spatial filtering at ALMA: Omission of <33 m baselines dilutes large-scale absorption; correcting for maximum plausible filtering, ALMA whole-planet l:c could be as deep as −4.9×10⁻⁴, bracketing JCMT’s −2.5×10⁻⁴. Zonal corrections suggest equator and mid-latitudes may be consistent after filtering; polar cap remains weak.
• Abundance estimate: Radiative transfer modeling of JCMT data yields PH₃ mixing ratio ~20 ppb (range ~20–30 ppb depending on CO₂ broadening). JCMT l:c uncertainty contributes ~±6 ppb (30%), with systematics of −2/+5 ppb possible.
• Line contamination: The nearest plausible interloper, SO₂ (J=31) offset by +1.3 km/s, is constrained by non-detection of favorable SO₂ transitions in simultaneous ALMA wideband data. Modeling indicates SO₂ contributes <10% to integrated l:c over ±5 km/s and would shift the centroid by <0.1 km/s; cannot reproduce the observed feature without violating SO₂ limits.
• Vertical sensitivity: Absorption arises against continuum from ~53–61 km (upper/middle cloud deck), implying PH₃ is at least at these altitudes; narrower line widths consistent with upper-atmosphere absorptions.
• Required source flux and lifetime: PH₃ lifetime very short above ~80 km (<10³ s), long near surface (~10⁸ s), and uncertain but limited in between; overall atmospheric lifetime ≤10³ years considering transport to destructive regions. To sustain ~10 ppb, required outgassing/production flux is ~10⁶–10⁷ molecules cm⁻² s⁻¹.
• Abiotic production insufficiency:
  • Photochemical pathways: Maximum forward-only network rates under Venus conditions are 4–6 orders of magnitude too small to balance destruction.
  • Thermodynamics: ~75 candidate reactions across thousands of conditions are endergonic by ~10–400 kJ/mol (avg ~+100 kJ/mol); subsurface oxygen fugacity too high by 8–15 orders to allow phosphate reduction; hydrolysis of meteoritic phosphide and formation/disproportionation of phosphorous acid ruled out quantitatively.
  • Energetic events: Lightning yields are ≥10⁷ times too low; volcanism would require >200× Earth’s activity (up to ~10⁸, depending on mantle assumptions); meteoritic delivery leads to f(PH₃) ~6.5×10⁻¹³—far below observations; other processes (solar wind protons, tribochemical) negligible.
• Ancillary detections: Simultaneous HDO line detected with Venus-normal water abundance when modeled; supports data fidelity.

The observations provide evidence for PH₃ at ~20 ppb in Venus’ cloud decks, strongest at mid-latitudes and not detected at the poles within limits. The spectral line’s velocity, width, depth, and reproducibility across independent facilities and reductions, plus the lack of viable contaminants (SO₂ constrained to minor contribution), support the PH₃ identification. Interpreting the origin, comprehensive assessments of photochemical, thermochemical, and energetic abiotic pathways fail by 4–8 orders of magnitude to account for the required production flux (~10⁶–10⁷ cm⁻² s⁻¹). Thus, PH₃ may indicate either unknown photochemistry (e.g., droplet-phase reactions in H₂SO₄ clouds), unknown geochemistry, or potentially biological production, by analogy to Earth where PH₃ is associated with life in an oxidizing atmosphere. The latitudinal pattern—enhanced at mid-latitudes and weak/absent over poles—might relate to Hadley circulation boundaries and cloud dynamics, but longitudinal analyses were limited by spectral ripple. Vertical distribution remains uncertain beyond the constraint that PH₃ must be above ~53 km; lifetime–transport interplay likely confines PH₃ to cloud altitudes. While the detection alone is not evidence for life, it reveals anomalous chemistry on Venus that current models cannot explain, motivating further observations and laboratory/theoretical work.

The study reports and confirms a millimeter-wave absorption feature at the PH₃ 1–0 transition in Venus’ atmosphere using JCMT and ALMA, inferring a PH₃ abundance of ~20 ppb at cloud-deck altitudes. Robustness checks exclude known contaminants (notably SO₂) as primary contributors and confirm line properties across instruments and reductions. Exhaustive evaluations of steady-state photochemical, thermodynamic, geological, and exogenous mechanisms cannot produce the observed PH₃ within many orders of magnitude, indicating unknown atmospheric chemistry or geochemistry, or potentially biological processes, may be at play. Future work should: (1) search for additional PH₃ transitions across mm/submm and IR (recognizing ground-based challenges), (2) improve mapping and vertical profiling with higher dynamic range and minimized ripple, (3) undertake laboratory measurements (e.g., CO₂ pressure-broadening, PH₃ reaction kinetics, droplet-phase photochemistry), (4) refine photochemical–transport models with better radical profiles and cloud microphysics, and (5) pursue in situ cloud and surface sampling or aerosol return to directly assess sources.

• Single-transition detection: Only the PH₃ 1–0 line was observed; higher-J and IR lines were not available, limiting spectroscopic confirmation.
• Instrumental systematics: Significant spectral baseline ripple required high-order polynomial removal; residual ripple increases channel-to-channel noise and complicates wing measurements.
• Interferometric filtering: ALMA’s omission of short baselines dilutes smooth, large-scale absorption, biasing l:c depths and spatial distribution; corrections depend on unknown spatial scales of PH₃.
• Abundance retrieval uncertainties: Dominated by unknown CO₂ pressure-broadening for PH₃ 1–0 (0.186–0.286 cm⁻¹/atm) and systematics in l:c; vertical profile assumed constant.
• Photochemical model uncertainties: Radical abundances (H, OH, O, Cl) and cloud microphysics uncertain; PH₃ lifetimes at intermediate altitudes poorly constrained by orders of magnitude; UV ‘mysterious absorber’ parameterization.
• Spatial analysis constraints: Zonal averaging used to mitigate ripple; polar spectra noisiest; longitudinal contrasts could not be robustly assessed.
• Alternative species: While known contaminants are ruled out to first order (SO₂ <10% contribution), an unknown species with an extremely close rest frequency cannot be entirely excluded without multi-line confirmation.