Is Hg recycling into the Earth’s mantle negligible?
Affiliations | Corresponding Author | Cite as | Funding information
Published by the European Association of Geochemistry
under Creative Commons License CC BY-NC-ND 4.0
Keywords: Mercury (Hg) concentrations, Hg devolatilisation, Hg subduction, Earth’s mantle, mantle-derived basalts and mafic rocks, collision and subduction-related rocks
- Share this article
-
Article views:341Cumulative count of HTML views and PDF downloads.
- Download Citation
- Rights & Permissions
top
Abstract
Figures
 Figure 1 Box plot diagram showing the mercury (Hg) concentrations (ng/g) in mantle rocks, mantle derived basalts and mafic rocks, depleted mantle (DM), MORB, bulk continental crust, CI chondrite and primitive mantle. Numbers in parentheses correspond to references in Table S-1, which are also reported in the SI. |  Figure 2 Mercury (Hg) concentrations (ng/g) vs. temperature (°C) in metamorphic samples from Otago and Alpine schists, New Zealand, California Coastal Ranges, USA, Dabie and Sumdo belts, China, and Central Slave Craton, Canada. |  Figure 3 P-T diagram showing the devolatilisation of Hg. Mercury concentrations (ng/g) are median values discussed in the text. Tonga and Cascadia (Syracuse et al., 2010) represent the coldest and hottest subduction zones (after Leeman, 2020). The geothermal gradient of 150 °C/GPa (Palin et al., 2020) is reported solely for reference. |
| Figure 1 | Figure 2 | Figure 3 |
top
Introduction
Sediments and seawater altered oceanic basalts undergo significant loss of mercury (Hg) during prograde metamorphism due to its high volatility, resulting in high grade metamorphic rocks (>500 °C and ∼0.5–3.0 GPa) having Hg concentrations <3 ng/g (Marowsky and Wedepohl, 1971
Marowsky, G., Wedepohl, K. (1971) General trends in the behavior of Cd, Hg, Tl and Bi in some major rock forming processes. Geochimica et Cosmochimica Acta 35, 1255–1267. https://doi.org/10.1016/0016-7037(71)90114-1
; Pitcairn et al., 2006Pitcairn, I.K., Teagle, D.A.H., Craw, D., Olivo, G.R., Kerrich, R., Brewer, T.S. (2006) Sources of metals in orogenic gold deposits: insights from the Otago and Alpine Schists, New Zealand. Economic Geology 101, 1525–1546. https://doi.org/10.2113/gsecongeo.101.8.1525
, 2010Pitcairn, I.K., Olivo, G.R., Teagle, D.A.H., Craw, D. (2010) Sulfide evolution during prograde metamorphism of the Otago and Alpine Schists, New Zealand. The Canadian Mineralogist 48, 1267–1296. https://doi.org/10.3749/canmin.48.5.1267
, 2015Pitcairn, I.K., Craw, D., Teagle, D.A.H. (2015) Metabasalts as sources of metals in orogenic gold deposits. Mineralium Deposita 50, 373–390. https://doi.org/10.1007/s00126-014-0547-y
; Stepanov, 2021Stepanov, A.S. (2021) A review of the geochemical changes occurring during metamorphic devolatilization of metasedimentary rocks. Chemical Geology 568, 120080. https://doi.org/10.1016/j.chemgeo.2021.120080
; Chen et al., 2025Chen, D., Luo, A., Zhendong, T., Changzhou D., Stephen E.G., Chen, Y-X., Yin, R. (2025) Mercury isotopes in eclogites constrain the sources and fates of mercury during oceanic subduction and continental collision. Science Bulletin 70, 657–660. https://doi.org/10.1016/j.scib.2024.12.025
). On a global scale, such low Hg concentrations in these rocks question the efficacy of Hg recycling into the mantle via subduction. To date, however, the large variation in Hg isotope mass independent fractionation values (typically as Δ199Hg) between basaltic rocks (−0.6 to +0.4 ‰) and terrestrial and marine sediments (−1 to +0.4 ‰) is taken as evidence of the Hg exchange between the Earth’s lithosphere and the atmosphere-land-ocean systems (e.g., Yin et al., 2024Yin, R., Wang, X., Sun, R., Gao, L., Deng, C., Tian, Z., Luo, A., Lehmann, B. (2024) Linking the mercury biogeochemical cycle to the deep mercury cycle: A mercury isotope perspective. Chemical Geology 654, 122063. https://doi.org/10.1016/j.chemgeo.2024.122063
; Xu et al., 2025Xu, R., Yin, R., White, W.M, Bizimis, M., Cai, Y., Zhang, J., Chen, C., Tian, Z., Ruan, T., Li, Y., Zhang, G., Liu, L., Bai, Z., Zhou, M., Liu, Y. (2025) Ancient storage of anomalous mercury isotope signatures in the Earth’s transition zone. Nature Communications 16, 11714. https://doi.org/10.1038/s41467-025-66917-z
and references therein).Such an apparent contrasting behaviour between elemental Hg and its isotopes defies whether the Hg concentrations observed in mantle rocks and mantle derived magmas reported in the literature support the recycling of Hg into the Earth’s mantle. Degassing processes in magmatic and volcanic systems modulate the final Hg concentration in rocks (e.g., Bouilliung et al., 2025
Boulliung, J., Wood, B.J., Mather, T.A. (2025) Volatility of mercury and related volatile metals at magmatic temperatures. Chemical Geology 695, 123018. https://doi.org/10.1016/j.chemgeo.2025.123018
). Also, analytical artefacts linked to low instrumental reproducibility and/or sample contamination can alter the measurement precision of Hg concentration in rocks (Narduzzi et al., 2025Narduzzi, F., Stefano C., Floreani, F., Pavoni, E., Petranich E., Jantzi S.C., Pistone, M., Černok, A., Venier, M., Crosera, M., Ziberna, L. (2025) Strengths and weaknesses of the analytical techniques used for measuring low mercury concentrations (<10 ng/g) in crystalline rocks: Direct Mercury Analyzer versus Cold-Vapour-Atomic-Fluorescence-Spectroscopy. Geostandards and Geoanalytical Research 49, 295–314. https://doi.org/10.1111/ggr.12606
). Here, we propose an alternative scenario for the Hg geochemical cycle by reviewing Hg concentrations in mantle peridotites, mantle derived basalts and mafic rocks (Mg# [Mg/(Mg + Fetot) mol. %] ≥ 60), and rocks from variably metamorphosed collision and subduction related settings. We aim to demonstrate the necessity of establishing new constraints on Hg concentration in protoliths and its behaviour during metamorphism. Details of data filtering procedure are reported in the Supplementary Information, along with the Hg and major element concentrations of the used samples (Table S-1). Data are presented as median ± interquartile range or as median ± standard deviation (1σ) when Hg analyses in the data set were ≤4. In a few cases, only one analysis was reported.top
Mercury in the Earth’s Mantle and Mantle Derived Rocks
Mantle peridotites, mantle derived basalts and mafic rocks exhibit similar Hg concentrations ranging from 1.0+0.5−0.4 to 2.1+3.0−1.1 ng/g (Fig. 1). However, most of these data are acquired via thermal decomposition atomic absorption spectrometry (TD-AAS) using DMA-80 and Lumex RA 915+, which can potentially produce relatively inaccurate and imprecise results when Hg concentrations in rocks are <10 ng/g (Narduzzi et al., 2025
Narduzzi, F., Stefano C., Floreani, F., Pavoni, E., Petranich E., Jantzi S.C., Pistone, M., Černok, A., Venier, M., Crosera, M., Ziberna, L. (2025) Strengths and weaknesses of the analytical techniques used for measuring low mercury concentrations (<10 ng/g) in crystalline rocks: Direct Mercury Analyzer versus Cold-Vapour-Atomic-Fluorescence-Spectroscopy. Geostandards and Geoanalytical Research 49, 295–314. https://doi.org/10.1111/ggr.12606
). Most of the previous studies rarely provide statistically robust results and a comprehensive description of the employed analytical methodology in terms of specifics regarding a single or set of analyses per sample, the related standard deviations, sample weights associated to the data, and procedural blanks prior to the sample measurements. This lack of information casts doubts on the uncertainty and repeatability of the results.Figure 1 Box plot diagram showing the mercury (Hg) concentrations (ng/g) in mantle rocks, mantle derived basalts and mafic rocks, depleted mantle (DM), MORB, bulk continental crust, CI chondrite and primitive mantle. Numbers in parentheses correspond to references in Table S-1, which are also reported in the SI.
For instance, the Hg concentration in mid-ocean ridge basalts (MORBs) (12.5 ng/g; Arevalo and McDonough, 2010
Arevalo, R.Jr., McDonough, W.F. (2010) Chemical variations and regional diversity observed in MORB. Chemical Geology 271, 70–85. https://doi.org/10.1016/j.chemgeo.2009.12.013
) appears to be unreliable (Fig. 1) because it was derived from Salters and Stracke’s (2004)Salters, V., Strake, A. (2004) Composition of the depleted mantle. Geochemistry, Geophysics, Geosystems 5, https://doi.org/10.1029/2003GC000597
uncertain depleted mantle (DM) Hg value. Salters and Stracke (2004)Salters, V., Strake, A. (2004) Composition of the depleted mantle. Geochemistry, Geophysics, Geosystems 5, https://doi.org/10.1029/2003GC000597
suggested 10 ng/g of Hg for the DM (Fig. 1) based on the similar Hg/Mn ratio between the primitive mantle (PM) of McDonough and Sun (1995)McDonough, W.F., Sun, S. (1995) The composition of the Earth. Chemical Geology 120, 223–253. https://doi.org/10.1016/0009-2541(94)00140-4
and the bulk continental crust (CC) of Rudnick and Gao (2003)Rudnick, R.L., Gao, S. (2003) 3.01 - Composition of the continental crust. In: Holland, H.D., Turekian, K.K. (Eds.) Treatise on Geochemistry 3, 1–64. http://dx.doi.org/10.1016/b0-08-043751-6/03016-4
. Such a Hg concentration in the DM represents only an approximation because Hg is rarely measured in basalts and peridotites, and the CC does not always complement the DM (Salters and Stracke, 2004Salters, V., Strake, A. (2004) Composition of the depleted mantle. Geochemistry, Geophysics, Geosystems 5, https://doi.org/10.1029/2003GC000597
). Additionally, the Hg value of the PM of McDonough and Sun (1995)McDonough, W.F., Sun, S. (1995) The composition of the Earth. Chemical Geology 120, 223–253. https://doi.org/10.1016/0009-2541(94)00140-4
(∼10 ng/g) derives from limited flameless AAS analyses on basaltic and peridotitic reference materials (Flanagan et al., 1982Flanagan, F.J., Moore, R., Aruscavage, P.J. (1982) Mercury in geologic reference samples. Geostandandards Newsletters 6, 25–46. https://doi.org/10.1111/j.1751-908X.1982.tb00343.x
) and CV-AAS analyses, after dissolution with aqua regia/HF, on orogenic peridotites from the Ivrea-Verbano Zone (Italy, Garuti et al., 1984Garuti, G., Gorgoni, C., Sighinolfi, G.P. (1984) Sulfide mineralogy and chalcophile and siderophile element abundances in the Ivrea-Verbano mantle peridotites (Western Italian Alps). Earth and Planetary Science Letters 70, 69–87. https://doi.org/10.1016/0012-821X(84)90210-3
), the latter being likely contaminated (Palme and O’Neil, 2014Palme, H., O’Neill, H.S.C. (2014) Cosmochemical estimates of mantle composition. In: Holland, H.D., Turekian, K.K. (Eds.) Treatise on Geochemistry (2nd edition), Elsevier, Oxford, 1–39. http://dx.doi.org/10.1016/B978-0-08-095975-7.00201-1
; Canil et al., 2015Canil, D., Crockford. P.W., Rossin, R., Telmer, K. (2015) Mercury in some arc crustal rocks and mantle peridotites and relevance to the moderately volatile element budget of the Earth. Chemical Geology 396, 134–142. http://dx.doi.org/10.1016/j.chemgeo.2014.12.029
; Narduzzi et al., 2025Narduzzi, F., Stefano C., Floreani, F., Pavoni, E., Petranich E., Jantzi S.C., Pistone, M., Černok, A., Venier, M., Crosera, M., Ziberna, L. (2025) Strengths and weaknesses of the analytical techniques used for measuring low mercury concentrations (<10 ng/g) in crystalline rocks: Direct Mercury Analyzer versus Cold-Vapour-Atomic-Fluorescence-Spectroscopy. Geostandards and Geoanalytical Research 49, 295–314. https://doi.org/10.1111/ggr.12606
).The estimate of Hg in the PM (∼6 ng/g) provided by Palme and O’Neill (2014)
Palme, H., O’Neill, H.S.C. (2014) Cosmochemical estimates of mantle composition. In: Holland, H.D., Turekian, K.K. (Eds.) Treatise on Geochemistry (2nd edition), Elsevier, Oxford, 1–39. http://dx.doi.org/10.1016/B978-0-08-095975-7.00201-1
considers that Hg is chalcophile and the Hg/Se ratio, calculated using the CC of Gao et al. (1998)Gao, S., Luo, T.C., Zhang, B.R., Zhang, B-R., Zhang, H-F., Han, Y-w., Zhao, Z-D., Hu, Y-K. (1998) Chemical composition of the continental crust as revealed by studies in East China. Geochimica et Cosmochimica Acta 62, 1959–1975. https://doi.org/10.1016/S0016-7037(98)00121-5.
, is constant during mantle melting. However, the Hg concentration in the PM might be biased because the Gao et al. (1998)Gao, S., Luo, T.C., Zhang, B.R., Zhang, B-R., Zhang, H-F., Han, Y-w., Zhao, Z-D., Hu, Y-K. (1998) Chemical composition of the continental crust as revealed by studies in East China. Geochimica et Cosmochimica Acta 62, 1959–1975. https://doi.org/10.1016/S0016-7037(98)00121-5.
CC Hg concentration (∼9 ng/g) is based on East China rocks rather than on a global data set (e.g., Rudnick and Gao, 2003Rudnick, R.L., Gao, S. (2003) 3.01 - Composition of the continental crust. In: Holland, H.D., Turekian, K.K. (Eds.) Treatise on Geochemistry 3, 1–64. http://dx.doi.org/10.1016/b0-08-043751-6/03016-4
). Conversely, Moynier et al. (2020)Moynier, F., Chen, J., Zhang, K., Cai, H., Wang, Z., Jackson M.G., Day, J.M.D. (2020) Chondritic mercury isotopic composition of Earth and evidence for evaporative equilibrium degassing during the formation of eucrites. Earth and Planetary Science Letters 551, 116544. https://doi.org/10.1016/j.epsl.2020.116544
obtained ∼2 ng/g of Hg for the PM, assuming that Hg was delivered by late accretion. However, it is unknown whether this value would change if Hg was partially or entirely partitioned into the core and/or devolatilised during planetary accretion. The work of Canil et al. (2015)Canil, D., Crockford. P.W., Rossin, R., Telmer, K. (2015) Mercury in some arc crustal rocks and mantle peridotites and relevance to the moderately volatile element budget of the Earth. Chemical Geology 396, 134–142. http://dx.doi.org/10.1016/j.chemgeo.2014.12.029
provides an exemplary approach to carefully assessing the uncertainties with Hg concentrations in rocks and how these are related to sample preparation procedures and instrumental errors. Indeed, Canil et al. (2015)Canil, D., Crockford. P.W., Rossin, R., Telmer, K. (2015) Mercury in some arc crustal rocks and mantle peridotites and relevance to the moderately volatile element budget of the Earth. Chemical Geology 396, 134–142. http://dx.doi.org/10.1016/j.chemgeo.2014.12.029
proposed 0.4−0.6 ng/g of Hg for the primitive upper mantle (PUM), based on TD-AAS analyses of six spinel lherzolite xenoliths from British Columbia showing depletion/enrichment trends. Besides the TD-AAS derived uncertainties (see Narduzzi et al., 2025Narduzzi, F., Stefano C., Floreani, F., Pavoni, E., Petranich E., Jantzi S.C., Pistone, M., Černok, A., Venier, M., Crosera, M., Ziberna, L. (2025) Strengths and weaknesses of the analytical techniques used for measuring low mercury concentrations (<10 ng/g) in crystalline rocks: Direct Mercury Analyzer versus Cold-Vapour-Atomic-Fluorescence-Spectroscopy. Geostandards and Geoanalytical Research 49, 295–314. https://doi.org/10.1111/ggr.12606
), the PUM value is based on a limited and local mantle xenolith dataset, and therefore not likely to be representative of the whole Earth’s mantle.Mercury concentrations in mantle peridotites are below ∼2.5 ng/g (i.e. maximum value from ophiolites; Canil et al., 2015
Canil, D., Crockford. P.W., Rossin, R., Telmer, K. (2015) Mercury in some arc crustal rocks and mantle peridotites and relevance to the moderately volatile element budget of the Earth. Chemical Geology 396, 134–142. http://dx.doi.org/10.1016/j.chemgeo.2014.12.029
). Note that here the analytical methods employed were mostly TD-AAS. As reliable results of cold vapour atomic fluorescence spectroscopy (CV-AFS) usually correspond to the minimum TD-AAS values (Narduzzi et al., 2025Narduzzi, F., Stefano C., Floreani, F., Pavoni, E., Petranich E., Jantzi S.C., Pistone, M., Černok, A., Venier, M., Crosera, M., Ziberna, L. (2025) Strengths and weaknesses of the analytical techniques used for measuring low mercury concentrations (<10 ng/g) in crystalline rocks: Direct Mercury Analyzer versus Cold-Vapour-Atomic-Fluorescence-Spectroscopy. Geostandards and Geoanalytical Research 49, 295–314. https://doi.org/10.1111/ggr.12606
), we suggest that the actual Hg concentrations in mantle peridotites range between 0.25 and 1.0 ng/g. This is below the Moynier et al. (2020)Moynier, F., Chen, J., Zhang, K., Cai, H., Wang, Z., Jackson M.G., Day, J.M.D. (2020) Chondritic mercury isotopic composition of Earth and evidence for evaporative equilibrium degassing during the formation of eucrites. Earth and Planetary Science Letters 551, 116544. https://doi.org/10.1016/j.epsl.2020.116544
PM Hg value and partially overlaps with the Canil et al. (2015)Canil, D., Crockford. P.W., Rossin, R., Telmer, K. (2015) Mercury in some arc crustal rocks and mantle peridotites and relevance to the moderately volatile element budget of the Earth. Chemical Geology 396, 134–142. http://dx.doi.org/10.1016/j.chemgeo.2014.12.029
Hg range in the PUM. The generally lower Hg concentrations in peridotites compared to most PM models are consistent with the incompatible behaviour of Hg during mantle melting and its progressive partitioning into the crust. Based on the above considerations on the analytical uncertainties, the Hg concentration in mantle derived magmas is most likely between ∼0.5 and 2 ng/g. Such a small difference between peridotites and mantle derived basalts would suggest relative partition coefficients for Hg (KD values for peridotite/melt are between 0.1 and 2.0) higher than previous estimates (Canil et al., 2015Canil, D., Crockford. P.W., Rossin, R., Telmer, K. (2015) Mercury in some arc crustal rocks and mantle peridotites and relevance to the moderately volatile element budget of the Earth. Chemical Geology 396, 134–142. http://dx.doi.org/10.1016/j.chemgeo.2014.12.029
). Alternatively, this might reflect the use of Hg concentrations from magmatic rocks that underwent extensive Hg loss due to degassing processes before complete solidification. This would be consistent with the experimental observations of Bouilliung et al. (2025)Boulliung, J., Wood, B.J., Mather, T.A. (2025) Volatility of mercury and related volatile metals at magmatic temperatures. Chemical Geology 695, 123018. https://doi.org/10.1016/j.chemgeo.2025.123018
, who reported 99 % of Hg degassing from basaltic melts at atmospheric pressure.top
Mercury in Metamorphic Terrains
Here, we report the metamorphic terrains in which Hg analyses were obtained for samples that underwent different peak metamorphic conditions and along different geothermal gradients (Table S-1 and Fig. S-1). Unless otherwise discussed, the studies summarised below used samples from areas where post-metamorphic magmatism has not occurred and represent prograde metamorphic rocks.
The Otago and Alpine Schists, New Zealand. These terrains form a <180 Ma accretionary complex where unmetamorphosed (100–200 °C, ≤0.1 GPa) sediments and seawater altered seafloor basalts underwent similar losses of Hg during prograde metamorphism up to amphibolite facies conditions (550–650 °C, 0.8–1.0 GPa; Pitcairn et al., 2006
Pitcairn, I.K., Teagle, D.A.H., Craw, D., Olivo, G.R., Kerrich, R., Brewer, T.S. (2006) Sources of metals in orogenic gold deposits: insights from the Otago and Alpine Schists, New Zealand. Economic Geology 101, 1525–1546. https://doi.org/10.2113/gsecongeo.101.8.1525
, 2010Pitcairn, I.K., Olivo, G.R., Teagle, D.A.H., Craw, D. (2010) Sulfide evolution during prograde metamorphism of the Otago and Alpine Schists, New Zealand. The Canadian Mineralogist 48, 1267–1296. https://doi.org/10.3749/canmin.48.5.1267
, 2015Pitcairn, I.K., Craw, D., Teagle, D.A.H. (2015) Metabasalts as sources of metals in orogenic gold deposits. Mineralium Deposita 50, 373–390. https://doi.org/10.1007/s00126-014-0547-y
). Unmetamorphosed and amphibolite facies rocks have Hg concentrations of 75+10−13 ng/g and 2.6+0.3−1.2 ng/g, respectively (CV-AFS analyses). Such data implies a potential Hg loss of ∼97 % with increasing metamorphic grade (Fig. S-1a). Specifically, the greatest loss of Hg occurs at or before 300–500 °C (greenschist facies), where Hg decreased to 8.1+8.8−4.2 ng/g, equivalent to ∼89 % decrease of Hg concentration.California Coastal Range, USA. Mercury was analysed (cold vapour MC-ICP-MS) in some lithologies near Clear Lake (Smith et al., 2008
Smith, C.N., Kesler, S.E., Blum, J.D., Rytuba, J.J. (2008) Isotope geochemistry of mercury in source rocks, mineral deposits and spring deposits of the California Coast Ranges, USA. Earth Planetary Science Letters 269, 399–407. https://doi.org/10.1016/j.epsl.2008.02.029
). The metamorphic P-T peak conditions were retrieved from other works (Fig. S-1b). Sediments from Great Valley (163 ± 18 °C, ∼0.1 GPa) and Franciscan Complex (225 ± 35 °C, ∼0.3 GPa) share similar Hg concentrations (65+22−10 and 64+21−22 ng/g, respectively) but are high compared to the Franciscan blueschists (200 ± 50 °C, 1 ± 0.5 GPa) and Coastal Ranges ophiolitic serpentinites (350 °C, ∼0.4 GPa), which have 43 ± 11 and 24 ± 6 ng/g of Hg, respectively.Dabie and Sumdo Belts, China. Dabie (2.8–3.6 GPa, 700–800 °C) and Sumdo (∼2.7 GPa, ∼730 °C) retrogressed eclogites (Fig. S-1c,d) share similar Hg concentrations (Hg = 0.3+0.4−0.0 and 0.3+0.1−0.1 ng/g, respectively) as analysed via the Lumex RA 915+ technique (Chen et al., 2025
Chen, D., Luo, A., Zhendong, T., Changzhou D., Stephen E.G., Chen, Y-X., Yin, R. (2025) Mercury isotopes in eclogites constrain the sources and fates of mercury during oceanic subduction and continental collision. Science Bulletin 70, 657–660. https://doi.org/10.1016/j.scib.2024.12.025
). Such Hg concentrations are lower than their assumed protoliths, with Hg mean concentrations of 13.5 ± 8.9 ng/g and 1.36 ± 0.55 ng/g for the Dabie and Sumdo protoliths, respectively, (Chen et al., 2025Chen, D., Luo, A., Zhendong, T., Changzhou D., Stephen E.G., Chen, Y-X., Yin, R. (2025) Mercury isotopes in eclogites constrain the sources and fates of mercury during oceanic subduction and continental collision. Science Bulletin 70, 657–660. https://doi.org/10.1016/j.scib.2024.12.025
). Accordingly, 70 % of Hg was released from the subducting slab while at crustal depths, and <30 % of Hg entered the mantle (Chen et al., 2025Chen, D., Luo, A., Zhendong, T., Changzhou D., Stephen E.G., Chen, Y-X., Yin, R. (2025) Mercury isotopes in eclogites constrain the sources and fates of mercury during oceanic subduction and continental collision. Science Bulletin 70, 657–660. https://doi.org/10.1016/j.scib.2024.12.025
). However, the percentages of Hg release reported in this paper are meaningless as the protoliths are unrelated in space and/or time to the Sumdo and Dabie eclogites. The protoliths they use in this study are basalts from present day mid-oceanic ridge settings and granitoids from the northeastern Central Asian Orogenic belt and South China Craton (cf. Fig. 1a in Chen et al., 2025Chen, D., Luo, A., Zhendong, T., Changzhou D., Stephen E.G., Chen, Y-X., Yin, R. (2025) Mercury isotopes in eclogites constrain the sources and fates of mercury during oceanic subduction and continental collision. Science Bulletin 70, 657–660. https://doi.org/10.1016/j.scib.2024.12.025
), and their Hg concentrations do not reflect those of the protoliths of the studied eclogites.Central Slave Craton. Mercury analyses (LA-ICP-MS) were also performed on sulphide inclusions in eclogite diamonds, which formed during Paleoproterozoic subduction of an oceanic crust interacting with serpentinite derived fluids (Fig. S-1e; Aulbach et al. 2012
Aulbach, S., Stachel, T., Seitz, H-M., Brey, G.P. (2012) Chalcophile and siderophile elements in sulphide inclusions in eclogitic diamonds and metal cycling in a Paleoproterozoic subduction zone. Geochimica et Cosmochimica Acta 93, 278–299. https://doi.org/10.1016/j.gca.2012.04.027
). According to nitrogen systematics of diamonds and a 1.85 Ga formation age for the sulphides, diamond mantle residence T was ∼1050 °C, corresponding to a depth of ∼170 km (∼5.5 GPa; Aulbach et al., 2012Aulbach, S., Stachel, T., Seitz, H-M., Brey, G.P. (2012) Chalcophile and siderophile elements in sulphide inclusions in eclogitic diamonds and metal cycling in a Paleoproterozoic subduction zone. Geochimica et Cosmochimica Acta 93, 278–299. https://doi.org/10.1016/j.gca.2012.04.027
). Here, although >30 % of Hg analyses were below the LA-ICP-MS detection limit (0.158 ± 0.115 μg/g), Aulbach et al. (2012)Aulbach, S., Stachel, T., Seitz, H-M., Brey, G.P. (2012) Chalcophile and siderophile elements in sulphide inclusions in eclogitic diamonds and metal cycling in a Paleoproterozoic subduction zone. Geochimica et Cosmochimica Acta 93, 278–299. https://doi.org/10.1016/j.gca.2012.04.027
calculated on the remaining data that the bulk eclogite xenolith had ≤1 ng/g of Hg (Fig. S-1e).top
Implications and Conclusions
The estimated concentrations of Hg in mantle derived rocks (Fig. 1) cast doubts on the possibility that significant amounts of Hg are recycled into the Earth’s mantle. Accordingly, the Hg concentrations in both collision and subduction related metamorphic rocks show a systematic decrease with increasing metamorphic grade (Fig. 2). Indeed, Hg devolatilisation appears to increase with temperature while being independent of pressure (Pitcairn et al., 2006
Pitcairn, I.K., Teagle, D.A.H., Craw, D., Olivo, G.R., Kerrich, R., Brewer, T.S. (2006) Sources of metals in orogenic gold deposits: insights from the Otago and Alpine Schists, New Zealand. Economic Geology 101, 1525–1546. https://doi.org/10.2113/gsecongeo.101.8.1525
, 2010Pitcairn, I.K., Olivo, G.R., Teagle, D.A.H., Craw, D. (2010) Sulfide evolution during prograde metamorphism of the Otago and Alpine Schists, New Zealand. The Canadian Mineralogist 48, 1267–1296. https://doi.org/10.3749/canmin.48.5.1267
, 2015Pitcairn, I.K., Craw, D., Teagle, D.A.H. (2015) Metabasalts as sources of metals in orogenic gold deposits. Mineralium Deposita 50, 373–390. https://doi.org/10.1007/s00126-014-0547-y
; Stepanov, 2021Stepanov, A.S. (2021) A review of the geochemical changes occurring during metamorphic devolatilization of metasedimentary rocks. Chemical Geology 568, 120080. https://doi.org/10.1016/j.chemgeo.2021.120080
and references therein). Consequently, by inference, we propose that any lithology along the modern day subduction P-T paths may lose Hg due to an increase in temperature (Fig. 3), and that more than 97 % of Hg can be devolatilised from any lithology forming a subducting slab before crossing the sediment and oceanic crust wet solidus curves, hence before reaching the slab-mantle wedge interface zone beneath arc volcanoes (Fig. 3; Grove et al., 2012Grove, T.L., Till, C.B., Krawczynski, M.J. (2012) The role of H2O in subduction zone magmatism. Annual Review of Earth and Planetary Sciences 40, 413–439. https://doi.org/10.1146/annurev-earth-042711-105310
; Leeman, 2020Leeman, W.P. (2020) Old/new subduction zone paradigms as seen from the Cascades. Frontiers in Earth Science 8, 535879. https://doi.org/10.3389/feart.2020.535879
). This observation is well supported by the Dabie and Sumdo belt eclogites which have negligible Hg (∼0.3 ng/g) and plot within the P-T range conditions of modern day subduction and the slab-mantle wedge interface zones (Fig. 3). Based on these observations, and in the absence of data that proves otherwise, it can be reasonably proposed that only a negligible amount of Hg (≤1 ng/g) can be recycled into the Earth’s mantle nowadays. The high solubility of Hg in crude oil (Wilhelm and Bloom 2000Wilhelm, S.M., Bloom, N. (2000) Mercury in petroleum. Fuel Processing Technology 63, 1–27. https://doi.org/10.1016/S0378-3820(99)00068-5
) indicates that methane-rich fluids and petroleum forming within the oil-gas window (∼60–225 °C, ≤0.25 GPa; Stepanov, 2021Stepanov, A.S. (2021) A review of the geochemical changes occurring during metamorphic devolatilization of metasedimentary rocks. Chemical Geology 568, 120080. https://doi.org/10.1016/j.chemgeo.2021.120080
; Goldfarb and Pitcairn, 2023Goldfarb, R.J., Pitcairn, I. (2023) Orogenic gold: Is a genetic association with magmatism realistic? Mineralium Deposita 58, 5–35. https://doi.org/10.1007/s00126-022-01146-8.
) and their migration towards shallower levels are likely responsible for the tremendous loss of Hg (up to ∼89 %) during metamorphism up to greenschist facies conditions (Stepanov, 2021Stepanov, A.S. (2021) A review of the geochemical changes occurring during metamorphic devolatilization of metasedimentary rocks. Chemical Geology 568, 120080. https://doi.org/10.1016/j.chemgeo.2021.120080
; Goldfarb and Pitcairn, 2023Goldfarb, R.J., Pitcairn, I. (2023) Orogenic gold: Is a genetic association with magmatism realistic? Mineralium Deposita 58, 5–35. https://doi.org/10.1007/s00126-022-01146-8.
).Figure 2 Mercury (Hg) concentrations (ng/g) vs. temperature (°C) in metamorphic samples from Otago and Alpine schists, New Zealand, California Coastal Ranges, USA, Dabie and Sumdo belts, China, and Central Slave Craton, Canada.
Figure 3 P-T diagram showing the devolatilisation of Hg. Mercury concentrations (ng/g) are median values discussed in the text. Tonga and Cascadia (Syracuse et al., 2010
Syracuse, E.M., Van Keken, P.E., Abers, G.A. (2010) The global range of subduction zone thermal models. Physics of the Earth and Planetary Interiors 183, 73–90. https://doi.org/10.1016/j.pepi.2010.02.004
) represent the coldest and hottest subduction zones (after Leeman, 2020Leeman, W.P. (2020) Old/new subduction zone paradigms as seen from the Cascades. Frontiers in Earth Science 8, 535879. https://doi.org/10.3389/feart.2020.535879
). The geothermal gradient of 150 °C/GPa (Palin et al., 2020Palin, R.M., Santosh, M., Cao, W., LI, S-S., Hernández-Uribee, D., Parsons, A. (2020) Secular change and the onset of plate tectonics on Earth. Earth-Science Review 207, 103172. https://doi.org/10.1016/j.earscirev.2020.103172
) is reported solely for reference.High upper mantle temperatures during the Archean (∼1450–1800 °C; e.g., Palin et al., 2020
Palin, R.M., Santosh, M., Cao, W., LI, S-S., Hernández-Uribee, D., Parsons, A. (2020) Secular change and the onset of plate tectonics on Earth. Earth-Science Review 207, 103172. https://doi.org/10.1016/j.earscirev.2020.103172
) that would have hindered oceanic lithosphere subduction (e.g., Brown et al., 2020Brown, M., Johnson, T., Gardiner, N.J. (2020) Plate Tectonics and the Archean Earth. Annual Review of Earth and Planetary Sciences 48, 291–320. https://doi.org/10.1146/annurev-earth-081619-052705
; Palin et al., 2020Palin, R.M., Santosh, M., Cao, W., LI, S-S., Hernández-Uribee, D., Parsons, A. (2020) Secular change and the onset of plate tectonics on Earth. Earth-Science Review 207, 103172. https://doi.org/10.1016/j.earscirev.2020.103172
), and the observation that ∼97 % of Hg is lost before oceanic crust and sediment wet-melting (Fig. 3), imply that recycling of Hg into the early Earth’s mantle was also negligible. For instance, the bulk Hg calculated for the ∼1.85 Ga Central Slave Craton eclogite xenolith (Aulbach et al., 2012Aulbach, S., Stachel, T., Seitz, H-M., Brey, G.P. (2012) Chalcophile and siderophile elements in sulphide inclusions in eclogitic diamonds and metal cycling in a Paleoproterozoic subduction zone. Geochimica et Cosmochimica Acta 93, 278–299. https://doi.org/10.1016/j.gca.2012.04.027
) indicates that ≤1 ng/g of Hg was recycled into the Paleoproterozoic mantle. Eventually, the limited amount of Hg recycled into the mantle over time (≤1 ng/g in the rocks of the subducted slabs), suggests that most of the Hg in present day volcanic rocks (>0.5–2.0 ng/g) might either derive from a primordial mantle reservoir, from assimilation of relatively Hg enriched crustal material, or both.In summary, this review highlights that the behaviour of Hg in different geological conditions and settings still remains to be well constrained, a problem exacerbated by the unforeseen uncertainties of the analytical methods used to analyse this element. While Hg isotopes suggest recycling of Hg into the mantle (e.g., Yin et al., 2024
Yin, R., Wang, X., Sun, R., Gao, L., Deng, C., Tian, Z., Luo, A., Lehmann, B. (2024) Linking the mercury biogeochemical cycle to the deep mercury cycle: A mercury isotope perspective. Chemical Geology 654, 122063. https://doi.org/10.1016/j.chemgeo.2024.122063
; Xu et al., 2025Xu, R., Yin, R., White, W.M, Bizimis, M., Cai, Y., Zhang, J., Chen, C., Tian, Z., Ruan, T., Li, Y., Zhang, G., Liu, L., Bai, Z., Zhou, M., Liu, Y. (2025) Ancient storage of anomalous mercury isotope signatures in the Earth’s transition zone. Nature Communications 16, 11714. https://doi.org/10.1038/s41467-025-66917-z
), our review suggests that Hg is almost entirely devolatilised before the subducting slab reaches the P-T conditions that allow fluids to migrate towards sub-arc mantle melting regions. Because Hg recycling was likely also inhibited in the Archean and Paleoproterozoic mantle, we then expect that Hg exchange between Earth’s lithosphere and the atmosphere-land-ocean systems was very limited over time. Hence, it appears that Hg isotopes only track a minimal part of the Hg cycle, whereas the majority of Hg lost during devolatilisation remains unsampled. A more robust way of calculating the cycling of Hg into the mantle might be using the existing data set highlighted in this manuscript, with Hg concentrations in forearc sedimentary and metamorphic lithologies and in volcanic rocks from arc settings. We emphasise the importance of using methodologies that allow for measurements of Hg with uncertainties ≤1 ng/g, and of reporting the complete analytical procedure that allows for assessing the uncertainty of the measurements. Only with such a data set it will be possible to test the hypothesis suggested in this work and to quantitatively estimate the distribution of Hg in the subduction cycle.top
Acknowledgements
FN was supported by Project 101066580 - STECALMY - HORIZON-MSCA-2021-PF-01, founded by the European Research Executive Agency (REA) – Marie Skłodowska-Curie Actions and Support to Experts A.2 – MSCA European Postdoctoral Fellowships. FN would like to express its gratitude to Sonja Aulbach, Jean-François Moyen, Davide Mariani, Marco Andreoli, Nonkuselo Madlakana, and Marthinus Cloete. MP acknowledges the support from the Teaming for Interdisciplinary Research Pre-Seed Program of the Office of Research of the University of Georgia (USA) and from the National Science Foundation (EAR 2322935). LZ acknowledges the “Fondazione CRTrieste”, project ref. 24112-2024.0028. The authors would like to express their sincere gratitude to Aleksandr Stepanov and the anonymous reviewer, whose invaluable comments and criticisms greatly improved the quality of this manuscript. The authors would also like to thank Raúl Fonseca for his excellent editorial handling and for recognising the value of this work.
Editor: Raúl Fonseca
top
References
Arevalo, R.Jr., McDonough, W.F. (2010) Chemical variations and regional diversity observed in MORB. Chemical Geology 271, 70–85. https://doi.org/10.1016/j.chemgeo.2009.12.013
Show in context For instance, the Hg concentration in mid-ocean ridge basalts (MORBs) (12.5 ng/g; Arevalo and McDonough, 2010) appears to be unreliable (Fig. 1) because it was derived from Salters and Stracke’s (2004) uncertain depleted mantle (DM) Hg value.
View in article
Aulbach, S., Stachel, T., Seitz, H-M., Brey, G.P. (2012) Chalcophile and siderophile elements in sulphide inclusions in eclogitic diamonds and metal cycling in a Paleoproterozoic subduction zone. Geochimica et Cosmochimica Acta 93, 278–299. https://doi.org/10.1016/j.gca.2012.04.027
Show in context Mercury analyses (LA-ICP-MS) were also performed on sulphide inclusions in eclogite diamonds, which formed during Paleoproterozoic subduction of an oceanic crust interacting with serpentinite derived fluids (Fig. S-1e; Aulbach et al. 2012).
View in article
According to nitrogen systematics of diamonds and a 1.85 Ga formation age for the sulphides, diamond mantle residence T was ∼1050 °C, corresponding to a depth of ∼170 km (∼5.5 GPa; Aulbach et al., 2012).
View in article
Here, although >30 % of Hg analyses were below the LA-ICP-MS detection limit (0.158 ± 0.115 μg/g), Aulbach et al. (2012) calculated on the remaining data that the bulk eclogite xenolith had ≤1 ng/g of Hg (Fig. S-1e).
View in article
For instance, the bulk Hg calculated for the ∼1.85 Ga Central Slave Craton eclogite xenolith (Aulbach et al., 2012) indicates that ≤1 ng/g of Hg was recycled into the Paleoproterozoic mantle.
View in article
Boulliung, J., Wood, B.J., Mather, T.A. (2025) Volatility of mercury and related volatile metals at magmatic temperatures. Chemical Geology 695, 123018. https://doi.org/10.1016/j.chemgeo.2025.123018
Show in context Degassing processes in magmatic and volcanic systems modulate the final Hg concentration in rocks (e.g., Bouilliung et al., 2025).
View in article
This would be consistent with the experimental observations of Bouilliung et al. (2025), who reported 99 % of Hg degassing from basaltic melts at atmospheric pressure.
View in article
Brown, M., Johnson, T., Gardiner, N.J. (2020) Plate Tectonics and the Archean Earth. Annual Review of Earth and Planetary Sciences 48, 291–320. https://doi.org/10.1146/annurev-earth-081619-052705
Show in context High upper mantle temperatures during the Archean (∼1450–1800 °C; e.g., Palin et al., 2020) that would have hindered oceanic lithosphere subduction (e.g., Brown et al., 2020; Palin et al., 2020), and the observation that ∼97 % of Hg is lost before oceanic crust and sediment wet-melting (Fig. 3), imply that recycling of Hg into the early Earth’s mantle was also negligible.
View in article
Canil, D., Crockford. P.W., Rossin, R., Telmer, K. (2015) Mercury in some arc crustal rocks and mantle peridotites and relevance to the moderately volatile element budget of the Earth. Chemical Geology 396, 134–142. https://dx.doi.org/10.1016/j.chemgeo.2014.12.029
Show in context Such a Hg concentration in the DM represents only an approximation because Hg is rarely measured in basalts and peridotites, and the CC does not always complement the DM (Salters and Stracke, 2004). Additionally, the Hg value of the PM of McDonough and Sun (1995) (∼10 ng/g) derives from limited flameless AAS analyses on basaltic and peridotitic reference materials (Flanagan et al., 1982) and CV-AAS analyses, after dissolution with aqua regia/HF, on orogenic peridotites from the Ivrea-Verbano Zone (Italy, Garuti et al., 1984), the latter being likely contaminated (Palme and O’Neil, 2014; Canil et al., 2015; Narduzzi et al., 2025).
View in article
The work of Canil et al. (2015) provides an exemplary approach to carefully assessing the uncertainties with Hg concentrations in rocks and how these are related to sample preparation procedures and instrumental errors.
View in article
Indeed, Canil et al. (2015) proposed 0.4−0.6 ng/g of Hg for the primitive upper mantle (PUM), based on TD-AAS analyses of six spinel lherzolite xenoliths from British Columbia showing depletion/enrichment trends.
View in article
Mercury concentrations in mantle peridotites are below ∼2.5 ng/g (i.e. maximum value from ophiolites; Canil et al., 2015).
View in article
Note that here the analytical methods employed were mostly TD-AAS. As reliable results of cold vapour atomic fluorescence spectroscopy (CV-AFS) usually correspond to the minimum TD-AAS values (Narduzzi et al., 2025), we suggest that the actual Hg concentrations in mantle peridotites range between 0.25 and 1.0 ng/g. This is below the Moynier et al. (2020) PM Hg value and partially overlaps with the Canil et al. (2015) Hg range in the PUM.
View in article
Such a small difference between peridotites and mantle derived basalts would suggest relative partition coefficients for Hg (KD values for peridotite/melt are between 0.1 and 2.0) higher than previous estimates (Canil et al., 2015).
View in article
Chen, D., Luo, A., Zhendong, T., Changzhou D., Stephen E.G., Chen, Y-X., Yin, R. (2025) Mercury isotopes in eclogites constrain the sources and fates of mercury during oceanic subduction and continental collision. Science Bulletin 70, 657–660. https://doi.org/10.1016/j.scib.2024.12.025
Show in context Sediments and seawater altered oceanic basalts undergo significant loss of mercury (Hg) during prograde metamorphism due to its high volatility, resulting in high grade metamorphic rocks (>500 °C and ∼0.5–3.0 GPa) having Hg concentrations <3 ng/g (Marowsky and Wedepohl, 1971; Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021; Chen et al., 2025).
View in article
Dabie (2.8–3.6 GPa, 700–800 °C) and Sumdo (∼2.7 GPa, ∼730 °C) retrogressed eclogites (Fig. S-1c,d) share similar Hg concentrations (Hg = 0.3+0.4−0.0 and 0.3+0.1−0.1 ng/g, respectively) as analysed via the Lumex RA 915+ technique (Chen et al., 2025).
View in article
Such Hg concentrations are lower than their assumed protoliths, with Hg mean concentrations of 13.5 ± 8.9 ng/g and 1.36 ± 0.55 ng/g for the Dabie and Sumdo protoliths, respectively, (Chen et al., 2025).
View in article
Accordingly, 70 % of Hg was released from the subducting slab while at crustal depths, and <30 % of Hg entered the mantle (Chen et al., 2025).
View in article
The protoliths they use in this study are basalts from present day mid-oceanic ridge settings and granitoids from the northeastern Central Asian Orogenic belt and South China Craton (cf. Fig. 1a in Chen et al., 2025), and their Hg concentrations do not reflect those of the protoliths of the studied eclogites.
View in article
Flanagan, F.J., Moore, R., Aruscavage, P.J. (1982) Mercury in geologic reference samples. Geostandandards Newsletters 6, 25–46. https://doi.org/10.1111/j.1751-908X.1982.tb00343.x
Show in context Such a Hg concentration in the DM represents only an approximation because Hg is rarely measured in basalts and peridotites, and the CC does not always complement the DM (Salters and Stracke, 2004). Additionally, the Hg value of the PM of McDonough and Sun (1995) (∼10 ng/g) derives from limited flameless AAS analyses on basaltic and peridotitic reference materials (Flanagan et al., 1982) and CV-AAS analyses, after dissolution with aqua regia/HF, on orogenic peridotites from the Ivrea-Verbano Zone (Italy, Garuti et al., 1984), the latter being likely contaminated (Palme and O’Neil, 2014; Canil et al., 2015; Narduzzi et al., 2025).
View in article
Gao, S., Luo, T.C., Zhang, B.R., Zhang, B-R., Zhang, H-F., Han, Y-w., Zhao, Z-D., Hu, Y-K. (1998) Chemical composition of the continental crust as revealed by studies in East China. Geochimica et Cosmochimica Acta 62, 1959–1975. https://doi.org/10.1016/S0016-7037(98)00121-5
Show in context The estimate of Hg in the PM (∼6 ng/g) provided by Palme and O’Neill (2014) considers that Hg is chalcophile and the Hg/Se ratio, calculated using the CC of Gao et al. (1998), is constant during mantle melting. However, the Hg concentration in the PM might be biased because the Gao et al. (1998) CC Hg concentration (∼9 ng/g) is based on East China rocks rather than on a global data set (e.g., Rudnick and Gao, 2003).
View in article
Garuti, G., Gorgoni, C., Sighinolfi, G.P. (1984) Sulfide mineralogy and chalcophile and siderophile element abundances in the Ivrea-Verbano mantle peridotites (Western Italian Alps). Earth and Planetary Science Letters 70, 69–87. https://doi.org/10.1016/0012-821X(84)90210-3
Show in context Such a Hg concentration in the DM represents only an approximation because Hg is rarely measured in basalts and peridotites, and the CC does not always complement the DM (Salters and Stracke, 2004). Additionally, the Hg value of the PM of McDonough and Sun (1995) (∼10 ng/g) derives from limited flameless AAS analyses on basaltic and peridotitic reference materials (Flanagan et al., 1982) and CV-AAS analyses, after dissolution with aqua regia/HF, on orogenic peridotites from the Ivrea-Verbano Zone (Italy, Garuti et al., 1984), the latter being likely contaminated (Palme and O’Neil, 2014; Canil et al., 2015; Narduzzi et al., 2025).
View in article
Goldfarb, R.J., Pitcairn, I. (2023) Orogenic gold: Is a genetic association with magmatism realistic? Mineralium Deposita 58, 5–35. https://doi.org/10.1007/s00126-022-01146-8
Show in context The high solubility of Hg in crude oil (Wilhelm and Bloom 2000) indicates that methane-rich fluids and petroleum forming within the oil-gas window (∼60–225 °C, ≤0.25 GPa; Stepanov, 2021; Goldfarb and Pitcairn, 2023) and their migration towards shallower levels are likely responsible for the tremendous loss of Hg (up to ∼89 %) during metamorphism up to greenschist facies conditions (Stepanov, 2021; Goldfarb and Pitcairn, 2023).
View in article
Grove, T.L., Till, C.B., Krawczynski, M.J. (2012) The role of H2O in subduction zone magmatism. Annual Review of Earth and Planetary Sciences 40, 413–439. https://doi.org/10.1146/annurev-earth-042711-105310
Show in context Consequently, by inference, we propose that any lithology along the modern day subduction P-T paths may lose Hg due to an increase in temperature (Fig. 3), and that more than 97 % of Hg can be devolatilised from any lithology forming a subducting slab before crossing the sediment and oceanic crust wet solidus curves, hence before reaching the slab-mantle wedge interface zone beneath arc volcanoes (Fig. 3; Grove et al., 2012; Leeman, 2020).
View in article
Lauretta, D., Devouard, B., Buseck, P. (1999) The cosmochemical behavior of mercury. Earth and Planetary Science Letters 171, 37–47. https://doi.org/10.1016/S0012-821X(99)00129-6
Leeman, W.P. (2020) Old/new subduction zone paradigms as seen from the Cascades. Frontiers in Earth Science 8, 535879. https://doi.org/10.3389/feart.2020.535879
Show in context Tonga and Cascadia (Syracuse et al., 2010) represent the coldest and hottest subduction zones (after Leeman, 2020).
View in article
Marowsky, G., Wedepohl, K. (1971) General trends in the behavior of Cd, Hg, Tl and Bi in some major rock forming processes. Geochimica et Cosmochimica Acta 35, 1255–1267. https://doi.org/10.1016/0016-7037(71)90114-1
Show in context Sediments and seawater altered oceanic basalts undergo significant loss of mercury (Hg) during prograde metamorphism due to its high volatility, resulting in high grade metamorphic rocks (>500 °C and ∼0.5–3.0 GPa) having Hg concentrations <3 ng/g (Marowsky and Wedepohl, 1971; Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021; Chen et al., 2025).
View in article
McDonough, W.F., Sun, S. (1995) The composition of the Earth. Chemical Geology 120, 223–253. https://doi.org/10.1016/0009-2541(94)00140-4
Show in context Salters and Stracke (2004) suggested 10 ng/g of Hg for the DM (Fig. 1) based on the similar Hg/Mn ratio between the primitive mantle (PM) of McDonough and Sun (1995) and the bulk continental crust (CC) of Rudnick and Gao (2003).
View in article
Such a Hg concentration in the DM represents only an approximation because Hg is rarely measured in basalts and peridotites, and the CC does not always complement the DM (Salters and Stracke, 2004). Additionally, the Hg value of the PM of McDonough and Sun (1995) (∼10 ng/g) derives from limited flameless AAS analyses on basaltic and peridotitic reference materials (Flanagan et al., 1982) and CV-AAS analyses, after dissolution with aqua regia/HF, on orogenic peridotites from the Ivrea-Verbano Zone (Italy, Garuti et al., 1984), the latter being likely contaminated (Palme and O’Neil, 2014; Canil et al., 2015; Narduzzi et al., 2025).
View in article
Moynier, F., Chen, J., Zhang, K., Cai, H., Wang, Z., Jackson M.G., Day, J.M.D. (2020) Chondritic mercury isotopic composition of Earth and evidence for evaporative equilibrium degassing during the formation of eucrites. Earth and Planetary Science Letters 551, 116544. https://doi.org/10.1016/j.epsl.2020.116544
Show in context Conversely, Moynier et al. (2020) obtained ∼2 ng/g of Hg for the PM, assuming that Hg was delivered by late accretion. However, it is unknown whether this value would change if Hg was partially or entirely partitioned into the core and/or devolatilised during planetary accretion.
View in article
Note that here the analytical methods employed were mostly TD-AAS. As reliable results of cold vapour atomic fluorescence spectroscopy (CV-AFS) usually correspond to the minimum TD-AAS values (Narduzzi et al., 2025), we suggest that the actual Hg concentrations in mantle peridotites range between 0.25 and 1.0 ng/g. This is below the Moynier et al. (2020) PM Hg value and partially overlaps with the Canil et al. (2015) Hg range in the PUM.
View in article
Narduzzi, F., Stefano C., Floreani, F., Pavoni, E., Petranich E., Jantzi S.C., Pistone, M., Černok, A., Venier, M., Crosera, M., Ziberna, L. (2025) Strengths and weaknesses of the analytical techniques used for measuring low mercury concentrations (<10 ng/g) in crystalline rocks: Direct Mercury Analyzer versus Cold-Vapour-Atomic-Fluorescence-Spectroscopy. Geostandards and Geoanalytical Research 49, 295–314. https://doi.org/10.1111/ggr.12606
Show in context Also, analytical artefacts linked to low instrumental reproducibility and/or sample contamination can alter the measurement precision of Hg concentration in rocks (Narduzzi et al., 2025).
View in article
However, most of these data are acquired via thermal decomposition atomic absorption spectrometry (TD-AAS) using DMA-80 and Lumex RA 915+, which can potentially produce relatively inaccurate and imprecise results when Hg concentrations in rocks are <10 ng/g (Narduzzi et al., 2025).
View in article
Such a Hg concentration in the DM represents only an approximation because Hg is rarely measured in basalts and peridotites, and the CC does not always complement the DM (Salters and Stracke, 2004). Additionally, the Hg value of the PM of McDonough and Sun (1995) (∼10 ng/g) derives from limited flameless AAS analyses on basaltic and peridotitic reference materials (Flanagan et al., 1982) and CV-AAS analyses, after dissolution with aqua regia/HF, on orogenic peridotites from the Ivrea-Verbano Zone (Italy, Garuti et al., 1984), the latter being likely contaminated (Palme and O’Neil, 2014; Canil et al., 2015; Narduzzi et al., 2025).
View in article
Besides the TD-AAS derived uncertainties (see Narduzzi et al., 2025), the PUM value is based on a limited and local mantle xenolith dataset, and therefore not likely to be representative of the whole Earth’s mantle.
View in article
Note that here the analytical methods employed were mostly TD-AAS. As reliable results of cold vapour atomic fluorescence spectroscopy (CV-AFS) usually correspond to the minimum TD-AAS values (Narduzzi et al., 2025), we suggest that the actual Hg concentrations in mantle peridotites range between 0.25 and 1.0 ng/g. This is below the Moynier et al. (2020) PM Hg value and partially overlaps with the Canil et al. (2015) Hg range in the PUM.
View in article
Palin, R.M., Santosh, M., Cao, W., LI, S-S., Hernández-Uribee, D., Parsons, A. (2020) Secular change and the onset of plate tectonics on Earth. Earth-Science Review 207, 103172. https://doi.org/10.1016/j.earscirev.2020.103172
Show in context The geothermal gradient of 150 °C/GPa (Palin et al., 2020) is reported solely for reference.
View in article
High upper mantle temperatures during the Archean (∼1450–1800 °C; e.g., Palin et al., 2020) that would have hindered oceanic lithosphere subduction (e.g., Brown et al., 2020; Palin et al., 2020), and the observation that ∼97 % of Hg is lost before oceanic crust and sediment wet-melting (Fig. 3), imply that recycling of Hg into the early Earth’s mantle was also negligible.
View in article
Palme, H., O’Neill, H.S.C. (2014) Cosmochemical estimates of mantle composition. In: Holland, H.D., Turekian, K.K. (Eds.) Treatise on Geochemistry (2nd edition), Elsevier, Oxford, 1–39. https://dx.doi.org/10.1016/B978-0-08-095975-7.00201-1
Show in context Such a Hg concentration in the DM represents only an approximation because Hg is rarely measured in basalts and peridotites, and the CC does not always complement the DM (Salters and Stracke, 2004). Additionally, the Hg value of the PM of McDonough and Sun (1995) (∼10 ng/g) derives from limited flameless AAS analyses on basaltic and peridotitic reference materials (Flanagan et al., 1982) and CV-AAS analyses, after dissolution with aqua regia/HF, on orogenic peridotites from the Ivrea-Verbano Zone (Italy, Garuti et al., 1984), the latter being likely contaminated (Palme and O’Neil, 2014; Canil et al., 2015; Narduzzi et al., 2025).
View in article
The estimate of Hg in the PM (∼6 ng/g) provided by Palme and O’Neill (2014) considers that Hg is chalcophile and the Hg/Se ratio, calculated using the CC of Gao et al. (1998), is constant during mantle melting. However, the Hg concentration in the PM might be biased because the Gao et al. (1998) CC Hg concentration (∼9 ng/g) is based on East China rocks rather than on a global data set (e.g., Rudnick and Gao, 2003).
View in article
Pitcairn, I.K., Teagle, D.A.H., Craw, D., Olivo, G.R., Kerrich, R., Brewer, T.S. (2006) Sources of metals in orogenic gold deposits: insights from the Otago and Alpine Schists, New Zealand. Economic Geology 101, 1525–1546. https://doi.org/10.2113/gsecongeo.101.8.1525
Show in context Sediments and seawater altered oceanic basalts undergo significant loss of mercury (Hg) during prograde metamorphism due to its high volatility, resulting in high grade metamorphic rocks (>500 °C and ∼0.5–3.0 GPa) having Hg concentrations <3 ng/g (Marowsky and Wedepohl, 1971; Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021; Chen et al., 2025).
View in article
These terrains form a <180 Ma accretionary complex where unmetamorphosed (100–200 °C, ≤0.1 GPa) sediments and seawater altered seafloor basalts underwent similar losses of Hg during prograde metamorphism up to amphibolite facies conditions (550–650 °C, 0.8–1.0 GPa; Pitcairn et al., 2006, 2010, 2015).
View in article
Indeed, Hg devolatilisation appears to increase with temperature while being independent of pressure (Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021 and references therein).
View in article
Pitcairn, I.K., Olivo, G.R., Teagle, D.A.H., Craw, D. (2010) Sulfide evolution during prograde metamorphism of the Otago and Alpine Schists, New Zealand. The Canadian Mineralogist 48, 1267–1296. https://doi.org/10.3749/canmin.48.5.1267
Show in context Sediments and seawater altered oceanic basalts undergo significant loss of mercury (Hg) during prograde metamorphism due to its high volatility, resulting in high grade metamorphic rocks (>500 °C and ∼0.5–3.0 GPa) having Hg concentrations <3 ng/g (Marowsky and Wedepohl, 1971; Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021; Chen et al., 2025).
View in article
These terrains form a <180 Ma accretionary complex where unmetamorphosed (100–200 °C, ≤0.1 GPa) sediments and seawater altered seafloor basalts underwent similar losses of Hg during prograde metamorphism up to amphibolite facies conditions (550–650 °C, 0.8–1.0 GPa; Pitcairn et al., 2006, 2010, 2015).
View in article
Indeed, Hg devolatilisation appears to increase with temperature while being independent of pressure (Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021 and references therein).
View in article
Pitcairn, I.K., Craw, D., Teagle, D.A.H. (2015) Metabasalts as sources of metals in orogenic gold deposits. Mineralium Deposita 50, 373–390. https://doi.org/10.1007/s00126-014-0547-y
Show in context Sediments and seawater altered oceanic basalts undergo significant loss of mercury (Hg) during prograde metamorphism due to its high volatility, resulting in high grade metamorphic rocks (>500 °C and ∼0.5–3.0 GPa) having Hg concentrations <3 ng/g (Marowsky and Wedepohl, 1971; Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021; Chen et al., 2025).
View in article
These terrains form a <180 Ma accretionary complex where unmetamorphosed (100–200 °C, ≤0.1 GPa) sediments and seawater altered seafloor basalts underwent similar losses of Hg during prograde metamorphism up to amphibolite facies conditions (550–650 °C, 0.8–1.0 GPa; Pitcairn et al., 2006, 2010, 2015).
View in article
Indeed, Hg devolatilisation appears to increase with temperature while being independent of pressure (Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021 and references therein).
View in article
Rudnick, R.L., Gao, S. (2003) 3.01 - Composition of the continental crust. In: Holland, H.D., Turekian, K.K. (Eds.) Treatise on Geochemistry 3, 1–64. https://dx.doi.org/10.1016/b0-08-043751-6/03016-4
Show in context Salters and Stracke (2004) suggested 10 ng/g of Hg for the DM (Fig. 1) based on the similar Hg/Mn ratio between the primitive mantle (PM) of McDonough and Sun (1995) and the bulk continental crust (CC) of Rudnick and Gao (2003).
View in article
The estimate of Hg in the PM (∼6 ng/g) provided by Palme and O’Neill (2014) considers that Hg is chalcophile and the Hg/Se ratio, calculated using the CC of Gao et al. (1998), is constant during mantle melting. However, the Hg concentration in the PM might be biased because the Gao et al. (1998) CC Hg concentration (∼9 ng/g) is based on East China rocks rather than on a global data set (e.g., Rudnick and Gao, 2003).
View in article
Salters, V., Strake, A. (2004) Composition of the depleted mantle. Geochemistry, Geophysics, Geosystems 5, https://doi.org/10.1029/2003GC000597
Show in context For instance, the Hg concentration in mid-ocean ridge basalts (MORBs) (12.5 ng/g; Arevalo and McDonough, 2010) appears to be unreliable (Fig. 1) because it was derived from Salters and Stracke’s (2004) uncertain depleted mantle (DM) Hg value.
View in article
Salters and Stracke (2004) suggested 10 ng/g of Hg for the DM (Fig. 1) based on the similar Hg/Mn ratio between the primitive mantle (PM) of McDonough and Sun (1995) and the bulk continental crust (CC) of Rudnick and Gao (2003).
View in article
Such a Hg concentration in the DM represents only an approximation because Hg is rarely measured in basalts and peridotites, and the CC does not always complement the DM (Salters and Stracke, 2004). Additionally, the Hg value of the PM of McDonough and Sun (1995) (∼10 ng/g) derives from limited flameless AAS analyses on basaltic and peridotitic reference materials (Flanagan et al., 1982) and CV-AAS analyses, after dissolution with aqua regia/HF, on orogenic peridotites from the Ivrea-Verbano Zone (Italy, Garuti et al., 1984), the latter being likely contaminated (Palme and O’Neil, 2014; Canil et al., 2015; Narduzzi et al., 2025).
View in article
Smith, C.N., Kesler, S.E., Blum, J.D., Rytuba, J.J. (2008) Isotope geochemistry of mercury in source rocks, mineral deposits and spring deposits of the California Coast Ranges, USA. Earth Planetary Science Letters 269, 399–407. https://doi.org/10.1016/j.epsl.2008.02.029
Show in context Mercury was analysed (cold vapour MC-ICP-MS) in some lithologies near Clear Lake (Smith et al., 2008).
View in article
Stepanov, A.S. (2021) A review of the geochemical changes occurring during metamorphic devolatilization of metasedimentary rocks. Chemical Geology 568, 120080. https://doi.org/10.1016/j.chemgeo.2021.120080
Show in context Sediments and seawater altered oceanic basalts undergo significant loss of mercury (Hg) during prograde metamorphism due to its high volatility, resulting in high grade metamorphic rocks (>500 °C and ∼0.5–3.0 GPa) having Hg concentrations <3 ng/g (Marowsky and Wedepohl, 1971; Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021; Chen et al., 2025).
View in article
Indeed, Hg devolatilisation appears to increase with temperature while being independent of pressure (Pitcairn et al., 2006, 2010, 2015; Stepanov, 2021 and references therein).
View in article
The high solubility of Hg in crude oil (Wilhelm and Bloom 2000) indicates that methane-rich fluids and petroleum forming within the oil-gas window (∼60–225 °C, ≤0.25 GPa; Stepanov, 2021; Goldfarb and Pitcairn, 2023) and their migration towards shallower levels are likely responsible for the tremendous loss of Hg (up to ∼89 %) during metamorphism up to greenschist facies conditions (Stepanov, 2021; Goldfarb and Pitcairn, 2023).
View in article
Syracuse, E.M., Van Keken, P.E., Abers, G.A. (2010) The global range of subduction zone thermal models. Physics of the Earth and Planetary Interiors 183, 73–90. https://doi.org/10.1016/j.pepi.2010.02.004
Show in context Tonga and Cascadia (Syracuse et al., 2010) represent the coldest and hottest subduction zones (after Leeman, 2020).
View in article
Wilhelm, S.M., Bloom, N. (2000) Mercury in petroleum. Fuel Processing Technology 63, 1–27. https://doi.org/10.1016/S0378-3820(99)00068-5
Show in context The high solubility of Hg in crude oil (Wilhelm and Bloom 2000) indicates that methane-rich fluids and petroleum forming within the oil-gas window (∼60–225 °C, ≤0.25 GPa; Stepanov, 2021; Goldfarb and Pitcairn, 2023) and their migration towards shallower levels are likely responsible for the tremendous loss of Hg (up to ∼89 %) during metamorphism up to greenschist facies conditions (Stepanov, 2021; Goldfarb and Pitcairn, 2023).
View in article
Xu, R., Yin, R., White, W.M, Bizimis, M., Cai, Y., Zhang, J., Chen, C., Tian, Z., Ruan, T., Li, Y., Zhang, G., Liu, L., Bai, Z., Zhou, M., Liu, Y. (2025) Ancient storage of anomalous mercury isotope signatures in the Earth’s transition zone. Nature Communications 16, 11714. https://doi.org/10.1038/s41467-025-66917-z
Show in context To date, however, the large variation in Hg isotope mass independent fractionation values (typically as Δ199Hg) between basaltic rocks (−0.6 to +0.4 ‰) and terrestrial and marine sediments (−1 to +0.4 ‰) is taken as evidence of the Hg exchange between the Earth’s lithosphere and the atmosphere-land-ocean systems (e.g., Yin et al., 2024; Xu et al., 2025 and references therein).
View in article
While Hg isotopes suggest recycling of Hg into the mantle (e.g., Yin et al., 2024; Xu et al., 2025), our review suggests that Hg is almost entirely devolatilised before the subducting slab reaches the P-T conditions that allow fluids to migrate towards sub-arc mantle melting regions.
View in article
Yin, R., Wang, X., Sun, R., Gao, L., Deng, C., Tian, Z., Luo, A., Lehmann, B. (2024) Linking the mercury biogeochemical cycle to the deep mercury cycle: A mercury isotope perspective. Chemical Geology 654, 122063. https://doi.org/10.1016/j.chemgeo.2024.122063
Show in context To date, however, the large variation in Hg isotope mass independent fractionation values (typically as Δ199Hg) between basaltic rocks (−0.6 to +0.4 ‰) and terrestrial and marine sediments (−1 to +0.4 ‰) is taken as evidence of the Hg exchange between the Earth’s lithosphere and the atmosphere-land-ocean systems (e.g., Yin et al., 2024; Xu et al., 2025 and references therein).
View in article
While Hg isotopes suggest recycling of Hg into the mantle (e.g., Yin et al., 2024; Xu et al., 2025), our review suggests that Hg is almost entirely devolatilised before the subducting slab reaches the P-T conditions that allow fluids to migrate towards sub-arc mantle melting regions.
View in article
top
Supplementary Information
The Supplementary Information includes:
- Data Filtering
- Supplementary Figure S-1
- Metamorphic rocks
- Supplementary Figure S-2
- Supplementary Table S-1
- References, including those listed in supplementary Table S-1
Download the Supplementary Information (PDF)
Download Table S-1 (xlsx)
Figures
Figure 1 Box plot diagram showing the mercury (Hg) concentrations (ng/g) in mantle rocks, mantle derived basalts and mafic rocks, depleted mantle (DM), MORB, bulk continental crust, CI chondrite and primitive mantle. Numbers in parentheses correspond to references in Table S-1, which are also reported in the SI.
Figure 2 Mercury (Hg) concentrations (ng/g) vs. temperature (°C) in metamorphic samples from Otago and Alpine schists, New Zealand, California Coastal Ranges, USA, Dabie and Sumdo belts, China, and Central Slave Craton, Canada.
Figure 3 P-T diagram showing the devolatilisation of Hg. Mercury concentrations (ng/g) are median values discussed in the text. Tonga and Cascadia ( Syracuse et al., 2010
Syracuse, E.M., Van Keken, P.E., Abers, G.A. (2010) The global range of subduction zone thermal models. Physics of the Earth and Planetary Interiors 183, 73–90. https://doi.org/10.1016/j.pepi.2010.02.004
) represent the coldest and hottest subduction zones (after Leeman, 2020Leeman, W.P. (2020) Old/new subduction zone paradigms as seen from the Cascades. Frontiers in Earth Science 8, 535879. https://doi.org/10.3389/feart.2020.535879
). The geothermal gradient of 150 °C/GPa ( Palin et al., 2020Palin, R.M., Santosh, M., Cao, W., LI, S-S., Hernández-Uribee, D., Parsons, A. (2020) Secular change and the onset of plate tectonics on Earth. Earth-Science Review 207, 103172. https://doi.org/10.1016/j.earscirev.2020.103172
) is reported solely for reference.