Volatile metal emissions from volcanic degassing and lava-seawater interactions at Kīlauea Volcano, Hawai'i

The study investigates how volatile trace metals and metalloids are emitted from basaltic volcanism, focusing on differences between magmatic plumes and laze plumes generated when lava enters the ocean. Volcanic emissions can rival or exceed anthropogenic metal emissions during major eruptions and impact air quality and ecosystems. Basaltic ocean island settings like Kīlauea can generate additional emissions via lava–seawater interactions that produce acidic, chlorine-rich plumes, yet these laze plumes are comparatively understudied. The 2018 Kīlauea lower East Rift Zone eruption, with exceptionally high SO2 fluxes and extensive coastal entries, provided a natural laboratory to quantify metal volatility, particle size distributions, and gas speciation, and to determine how seawater-derived chlorine alters metal degassing compared to the magmatic source vent.

Prior work has documented volcanic metal emissions at numerous volcanoes since the 1960s–1970s, showing that volcanoes are major natural metal sources and that some metals act as nutrients at low levels but pollutants at high levels. Basaltic plumes can expose large populations to gases and metal-bearing particulates. Laze plumes from lava–seawater interaction are known to be dominated by HCl after water and can rival industrial sources of HCl, but their detailed chemistry is less constrained than magmatic plumes. Comparative studies show tectonic setting influences volatile metal abundances, with arc volcanoes often exhibiting higher chloride-associated metal emissions due to more oxidized, halogen-rich gases. Experimental and modelling studies have highlighted roles of chloride and sulphide ligands in metal speciation, but intraplate basaltic systems (e.g., Kīlauea) have been less explored with comprehensive speciation modelling.

Field sampling during the 2018 Kīlauea LERZ eruption combined ground-based and unoccupied aircraft system (UAS) measurements to characterize gas and particulate matter (PM) in magmatic plume (Fissure 8) and laze plume (ocean entry) emissions. Instruments and approaches: (1) Multi-GAS for in situ CO2, SO2, H2S (and H2O via T and RH) at 1 Hz, calibrated pre/post campaign. CO2/SO2 ratios computed with Ratiocalc. (2) Filter packs to collect gases (alkali-impregnated filters capturing SO2 as sulfate and HCl as chloride) and bulk PM on PTFE filters; operated on ground and via UAS. (3) Sioutas cascade impactor on UAS for size-segregated PM in five bins (D50: 2.5, 1.0, 0.5, 0.25 µm, and after-stage <0.25 µm). UAS sampling heights were ~300 m over Fissure 8 and ~100 m over the ocean entry; ground-based sampling deployed via a drop-and-run frame near Fissure 8. (4) UV spectroscopic SO2 flux measurements with a low-cost PiSpec DOAS spectrometer traversing beneath the plume; four traverses produced a weighted mean SO2 emission rate of 39 ± 11 kt/day (with correction for saturation and wind from GDAS). Additional literature-constrained rates reported earlier in the eruption exceeded 200 kt/day. Laboratory analyses: PM and gas filters were extracted under clean-lab conditions. Water-soluble fractions were analyzed by ion chromatography (IC) for Cl− and SO4 2−; total digests analyzed by ICP-MS and ICP-OES for major and trace elements. Detection limits and blanks were assessed; propagated uncertainties for IC: ±33% (Cl−), ±35% (SO4 2−); ICP methods: ±10–18%. Data processing: - Ash (silicate) correction used a contemporaneous Kīlauea basaltic matrix glass composition and refractory elements (Fe, Al, Ti, selected REEs) to derive weighted ash fractions (WAF) and isolate non-silicate aerosol contributions. - Emanation coefficients (ε) were estimated from ash-corrected X/S in the plume and an assumed total S degassed from the melt (1250 ± 300 ppm) to quantify relative metal volatility. - Elemental emission rates for trace metals were derived by multiplying X/SO2 mass ratios by independently measured SO2 fluxes. - Gas speciation modelling employed Gibbs Energy Minimisation (HSC Chemistry v9.9.2). The model simulated simultaneous mixing of magmatic gases with air (Va/Vm = 0–0.33) and associated cooling (1145 °C to 1016 °C) to capture the compositional discontinuity from rapid oxidation. Input major gas compositions followed published Kīlauea datasets; trace metals used ash-corrected filter pack data corrected to source via total sulphur. Sensitivity tests varied HCl by 0.001–1000× to assess chloride’s effect on metal speciation at reduced (Va/Vm = 0) and oxidized (Va/Vm = 0.33) endmembers. Sampling constraints: Two non-saturated ground-based magmatic samples (FP_8_5 and FP_8_6) formed the basis for quantitative gas/PM partitioning and X/SO2 ratios; laze plume was sampled via UAS at Isaac Hale Park (filter packs and three consecutive impactor flights). Background (2019) atmospheric measurements were used for comparison.

• Magmatic vs laze plume composition: The magmatic plume contains abundant volatile metals and metalloids; the laze plume is significantly enriched in chlorine and exhibits elevated copper (and Ag) relative to both seawater and the magmatic plume. Total molar S/Cl is 32–35 in the magmatic plume and ~0.1 in the laze plume.
• Gas/particle partitioning and size distributions: In the near-source magmatic plume, <1% of total S and Cl are in PM; S is ~91% in D ≤ 0.25 µm PM; Cl is bimodal (~60% ≤0.25 µm; ~27% ≥2.5 µm). In the laze plume, ~49% of S and ~50% of Cl are in PM and are more uniformly distributed across size bins.
• Volatility (emanation coefficients, ε): Elements were grouped as refractory (ε < 0.001%) and volatile (ε > 0.001%). Order of increasing volatility (ε > 0): ε < 0.001%: Gd, La, Sm, Tb, Nd, Pr, Ce, Fe, Eu, Al, Ti, Mg, Ta; ε > 0.001%: Ba, Na, Cu; ε > 0.01%: Zn, Ag; ε > 0.1%: In, Sn, Pb; ε > 1%: As; ε > 10%: Bi, Cd, Cl, Se, Re, Te, S. Refractory elements are predominantly coarse and ash-derived (>90% accounted for by ash for most), while volatile elements are predominantly fine (D < 0.25 µm) with WAFs generally <0.3% (exceptions: Ba 55%, Cu 4%, Na 4%, Zn 1%).
• Speciation modelling: At emission (no air mixing), many volatile elements complex with sulphide (e.g., TeS, SeS, AsS, PbS, SnS) or appear in elemental form; some (In, Ag, Cu) are present as chlorides. After atmospheric mixing and oxidation (Va/Vm ~0.33), sulphide species shift to oxides/hydroxides and chloride complexes become more prevalent. Increasing HCl by ~100× drives nearly complete chloride complexation of Cu, Ag, Zn, Cd, and Bi at Va/Vm = 0.33; Se, Te, As remain non-chloride even at high HCl.
• Laze plume sources: Comparing X/Cl in laze to seawater shows seawater contribution dominates for major seawater ions (Ca, Mg, Na, K), but laze is enriched by 1–6 orders of magnitude over seawater for Al, Ti, Fe, REEs, and volatile metals (Cd, Bi, Cu, Ag, Zn, Re). After ash correction, several refractory elements remain elevated—likely due to particle scavenging/lofting in seawater—and volatile metals (Cd, Zn, Ag, Cu, Bi, Re) remain elevated, indicating degassing from distal lavas in a Cl-rich environment. Cu emission in laze may equal or exceed magmatic plume Cu.
• Tectonic setting contrasts: Kīlauea’s element volatility order matches Erta 'Ale (rift basalt), while Holuhraun (Iceland) shows lower X/SO2 and ε, plausibly due to lower chlorine (S/Cl mass: Holuhraun 40–52 vs Kīlauea 32–35; Erta ‘Ale 6–15). Arc volcanoes show higher In and Cu relative abundances, consistent with chloride-driven speciation in more oxidized, Cl-rich arc gases.
• Emission rates: SO2 emission rate measured 39 ± 11 kt/day on 31 July 2018; earlier in June–early July exceeded 200 kt/day. Scaling Kīlauea 2018 X/SO2 to a hypothetical single Deccan Traps eruption of 1000 km3 implies average emissions of 300–7800 kg Cu/day and 3100–72000 kg Se/day over 10–100 years, potentially exceeding modern national anthropogenic Se emissions by up to three orders of magnitude.

Findings demonstrate that ligand availability and oxidation state strongly control volcanic metal degassing and speciation. At Kīlauea’s vent, reduced, S-rich gases favor sulphide complexes for several metalloids and metals; rapid atmospheric oxidation increases chloride complexation potential. In laze plumes, seawater-derived chlorine and reduced sulphur inventory in distal lavas shift degassing toward chloride-complexing metals (e.g., Cu, Ag, Zn, Cd, Bi), while sulphide-complexing elements (Se, Te, As) are minimal due to prior S loss and low residual concentrations. Thus, lava–seawater interactions produce a distinct metal fingerprint relative to magmatic plumes, with implications for air quality and marine ecosystems because laze plumes efficiently deliver metals directly to the ocean. Comparisons across tectonic settings reinforce that higher Cl and more oxidized conditions at arcs promote chloride-speciated metal emissions (e.g., In, Cu), whereas intraplate basalts like Kīlauea are more controlled by sulphide complexation at emission. The results provide analogues for large igneous province impacts, suggesting potentially massive trace metal releases that could influence environmental and biogeochemical systems on regional to global scales.

This work provides direct, size-resolved measurements and speciation modelling that distinguish metal emissions from magmatic versus lava–seawater interaction plumes at Kīlauea (2018). Key contributions are: (1) quantification of relative metal volatilities and particle-size associations; (2) demonstration that magmatic plumes are dominated by sulphide/metalloid species at emission with minimal PM partitioning for S and Cl; (3) identification of laze plumes as chlorine-rich environments that enhance chloride-complexing metal degassing, especially Cu, leading to a fundamentally different trace metal fingerprint and potentially higher Cu emissions than the magmatic source; and (4) elucidation of how ligand availability (S2− vs Cl−) and oxidation state govern speciation and variability across tectonic settings. Future work should target real-time gas speciation (including kinetics), expanded laze sampling across diverse ocean entries, quantification of seawater-evaporation fractionation effects on X/Cl, and coupled ocean–atmosphere metal deposition and bioavailability studies.

• Sampling constraints limited non-saturated ground-based magmatic samples to two, and laze sampling to UAS operations; some filter packs experienced gas filter saturation, restricting interpretation to PM on first filters. - Potential biases from UAS rotor turbulence and thermal updrafts on PM ingestion and apparent size distributions are not fully quantified. - Several elements were at or below detection limits, affecting completeness of laze metal inventories (e.g., Pb, In in filter packs). - Ion chromatography quantification can be impacted by residual H2O2 and glycerol, with large propagated uncertainties for Cl− and SO4 2−; F− was not reported. - Ash correction depends on assumed matrix glass composition and refractory tracers; small ash contributions can strongly affect apparent concentrations. - Emanation coefficient estimates rely on assumptions (e.g., linear scaling with S degassing, adopted Sdegassed = 1250 ± 300 ppm) and do not fully capture halide-driven degassing. - Speciation modelling assumes equilibrium, neglects kinetic effects, and uses historical major gas compositions; it assumes no compositional fractionation between source and sampling. - Evaporation-induced fractionation of X/Cl in seawater during laze formation is unconstrained and not accounted for. - The oxidation state of distal lavas at ocean entry is uncertain and approximated via endmember mixing in models.