Figures

top

Introduction

Aqueous fluids released from the subducted slab are likely the main agent for melting and mass transport in subduction zones (e.g., Tatsumi, 1989

Tatsumi, Y. (1989) Migration of fluid phases and genesis of basalt magmas in subduction zones. Journal of Geophysical Research: Solid Earth 94, 4697–4707. https://doi.org/10.1029/JB094iB04p04697

; Manning and Frezzotti, 2020

Manning, C.E., Frezzotti, M.L. (2020) Subduction-zone fluids. Elements 16, 395–400. https://doi.org/10.2138/gselements.16.6.395

; Rustioni et al., 2021

Rustioni, G., Audetat, A., Keppler, H. (2021) The composition of subduction zone fluids and the origin of the trace element enrichment in arc magmas. Contributions to Mineralogy and Petrology 176, 51. https://doi.org/10.1007/s00410-021-01810-8

). They also return subducted volatiles, such as water, carbon, and nitrogen back to the surface. Over long geologic time, this fluid flux, therefore, is involved in the processes regulating global sea level (e.g., Rüpke et al., 2004

Rüpke, L.H., Phipps Morgan, J., Hort, M., Connolly, J.A.D. (2004) Serpentine and the subduction zone water cycle. Earth and Planetary Science Letters 223, 17–34. https://doi.org/10.1016/j.epsl.2004.04.018

), the carbon dioxide content of the atmosphere (e.g., Plank and Manning 2019

Plank, T., Manning, C.E. (2019) Subducting carbon. Nature 574, 343–352. https://doi.org/10.1038/s41586-019-1643-z

), and climate. Traditionally, the fluids released by the dehydration of hydrous minerals, such as amphibole and serpentine, were considered to consist of simple solvent molecules, i.e. mostly H₂O and CO₂, plus some dissolved inorganic species (e.g., Manning, 2004

Manning, C.E. (2004) The chemistry of subduction zone fluids. Earth and Planetary Science Letters 223, 1–16. https://doi.org/10.1016/j.epsl.2004.04.030

). However, recently theoretical predictions have emerged which suggest that, in high pressure subduction fluids, most of the carbon may under some redox and pH conditions be present as organic molecules, such as acetate (Sverjensky et al., 2014

Sverjensky, D.A., Stagno, V., Huang, F. (2014) Important role for organic carbon in subduction-zone fluids in the deep carbon cycle. Nature Geoscience 7, 909–913. https://doi.org/10.1038/ngeo2291

, 2020

Sverjensky, D., Daniel, I., Vitale Brovarone, A. (2020) The changing character of carbon in fluids with pressure. In: Manning, C.E., Lin, J.-F., Mao, W.L. (Eds.) Carbon in Earth’s Interior. Geophysical Monograph 249. American Geophysical Union, Washington, D.C., John Wiley and Sons, Inc., Hoboken, NJ, 259–269. https://doi.org/10.1002/9781119508229.ch22

; Sverjensky and Huang, 2015

Sverjensky, D.A., Huang, F. (2015) Diamond formation due to a pH drop during fluid–rock interactions. Nature Communications 6, 8702. https://doi.org/10.1038/ncomms9702

). In particular, acetate was predicted to be the predominant carbon species in aqueous fluids at 5 GPa, 600 °C, and intermediate redox conditions (log fO₂ = 10⁻¹⁶ to 10⁻¹⁹ bar, one to four log units below the QFM = quartz fayalite magnetite buffer). Organic species in aqueous high pressure fluids have also been invoked to explain elevated solubilities of the forsterite + enstatite and magnesite + enstatite assemblages in the presence of carbon (Tiraboschi et al., 2018

Tiraboschi, C., Tumiati, S., Sverjensky, D.A., Pettke, T., Ulmer, P., Poli, S. (2018) Experimental determination of magnesia and silica solubilities in graphite-saturated and redox-buffered high-pressure COH fluids in equilibrium with forsterite + enstatite and magnesite + enstatite. Contributions to Mineralogy and Petrology 173, 2. https://doi.org/10.1007/s00410-017-1427-0

). Similarly, Tumiati et al. (2017)

Tumiati, S., Tiraboschi, C., Sverjensky, D.A., Pettke, T., Recchia, S., Ulmer, P., Miozzi, F., Poli, S. (2017) Silicate dissolution boosts the CO₂ concentrations in subduction fluids. Nature Communications 8, 616. https://doi.org/10.1038/s41467-017-00562-z

attributed the increase of CO₂ molar fraction in COH fluids upon in the presence of silica to the formation of some organic complexes involving Si.

Evidence for the stability of acetate and similar organic species either from observations in natural samples or from high pressure experiments is up to now rather limited. Frezzotti (2019)

Frezzotti, M.L. (2019) Diamond growth from organic compounds in hydrous fluids deep within the Earth. Nature Communications 10, 4952. https://doi.org/10.1038/s41467-019-12984-y

studied fluid inclusions in diamond-bearing rocks from the Alps by Raman spectroscopy. She concluded that the data provide evidence that diamond surfaces are coated by sp²- and sp³-bonded amorphous carbon containing functional groups of carboxylic acids. This conclusion is, however, based on the deconvolution of rather broad Raman bands into various components and on band assignments that may not be unique. Huang et al. (2017)

Huang, F., Daniel, I., Cardon, H., Montagnac, G., Sverjensky, D.A. (2017) Immiscible hydrocarbon fluids in the deep carbon cycle. Nature Communications 8, 15798. https://doi.org/10.1038/ncomms15798.

heated sodium acetate solutions in an externally heated diamond cell to 300 °C and 2.4–3.5 GPa and observed the partial decomposition of acetate to immiscible isobutane. However, they did not reach the predicted P-T stability field of acetate and it remains unclear whether the isobutane observed in these experiments is a stable species or just an intermediate, metastable decomposition product of the acetate. Li (2016)

Li, Y. (2016) Immiscible C–H–O fluids formed at subduction zone conditions. Geochemical Perspectives Letters 3, 12–21. https://doi.org/10.7185/geochemlet.1702

studied the speciation in aqueous C–H–O fluids using synthetic fluid inclusions. He observed traces of ethane and perhaps higher hydrocarbons together with methane at 2.5 GPa, 600 °C, and Fe–FeO buffer conditions, but no indication of acetate or other organic acid anions. We, therefore, carried out some exploratory experiments to test the predicted stability of acetate at 5 GPa and 600 °C.

top

Methods

All experiments were carried out with a solution of 10 wt. % of sodium acetate (CH₃COONa · 3 H₂O) in water. We made no attempt to externally buffer oxygen fugacity or pH. Rather, we assumed that if acetate were stable, it should, at high concentrations, buffer these parameters to its intrinsic stability range. We carried out two types of experiments:

   1. Piston cylinder experiments using very thick-walled silver capsules. Silver is poorly permeable for hydrogen at 600 °C (Chou, 1986

Chou, I.M. (1986) Permeability of precious metals to hydrogen at 2 KB total pressure and elevated temperatures. American Journal of Science 286, 638–658. https://doi.org/10.2475/ajs.286.8.638

), such that hydrogen loss from the capsule should be minimal and the intrinsic oxygen fugacity of the solution should be preserved during the experiments. Runs were designed to form a times series, from a nominally “zero-time experiment” (quenching 8 minutes after heating to 600 °C at 5 GPa) to 24 hours run duration. For mechanical stabilisation, capsules were filled with silica powder in addition to the acetate solution. The use of silica is realistic for subduction zone fluids, as even MORB eclogites usually contain a trace of free quartz or coesite (Sisson and Kelemen, 2018

Sisson, T.W., Kelemen, P.B. (2018) Near-solidus melts of MORB + 4 wt% H₂O at 0.8–2.8 GPa applied to issues of subduction magmatism and continent formation. Contributions to Mineralogy and Petrology 173, 70. https://doi.org/10.1007/s00410-018-1494-x

). Run products were investigated by Raman spectroscopy and by powder-ray diffraction.

   2. In situ experiments using externally heated diamond anvil cells. Here, the solution was directly observed under a microscope and studied by Raman spectroscopy during different heating and cooling paths. In one of these experiments, we reached 4.65 GPa and 600 °C, essentially the predicted stability field of acetate.

Since Raman spectroscopy was used as the main tool for identifying acetate and other organic species, Figure 1 shows the Raman spectra of both crystallised sodium acetate and of the 10 wt. % sodium acetate solution used in our experiments. Band assignments are after Ito and Bernstein (1956)

Ito, K., Bernstein, H.J. (1956) The vibrational spectra of the formate, acetate, and oxalate ions. Canadian Journal of Chemistry 34, 170–178. https://doi.org/10.1139/v56-021

. For identifying acetate in the experiments, we mostly relied on the strong C–C stretching band near 930 cm⁻¹ and the C–H stretching vibration of the methyl group at 2935 cm⁻¹.

Full size image

Further details on the experimental methods are given in the Supplementary Information.

top

Results from Piston Cylinder Experiments

The piston cylinder experiments were all carried out at 5 GPa and 600 °C, with run durations of 0.1, 0.5, 2, 6, and 24 hours. In all runs, the SiO₂ powder had been converted to coesite, as indicated by X-ray diffraction. No other crystalline phases were detected. Raman spectra of the quenched run products (Fig. 2) failed to detect any acetate. Instead, bands similar to a highly disordered graphite or perhaps kerogen-like material were observed. This is consistent with the grey discoloration observed in the run products (Fig. 2). There was no fundamental difference between the results of the “zero-time” experiment and the experiments with longer run durations up to 24 hours; only the bands of the carbonaceous material became narrower with increasing run duration.

Full size image

top

Results from in situ Experiments in the Diamond Cell

A total of four experiments were carried out by heating the acetate solution in the diamond anvil cell. Detailed pressure-temperature paths for all experiments are given in the Supplementary Figures S-1 and S-2. Run AB1 reached a maximum of 125 °C and 2.25 GPa before losing pressure. The Raman spectrum of acetate was observed up to the maximum temperature; the solution remained clear and no decomposition products of acetate were detectable. After cooling to room temperature, pressure was increased by tightening the cell such that the solution solidified to ice. Run AB2 was then conducted with this charge and reached 350 °C and 2.93 GPa before the gasket cracked. In this experiment, some dark precipitate started to form around 150 °C. However, as shown in Figure 3a, even at the highest temperatures, the C–C stretching bands (930 cm⁻¹) and the C–H stretching bands (2935 cm⁻¹) of acetate were still detectable in the Raman spectrum of the solution, although shifted to slightly higher wave numbers due to the elevated pressure. After cooling experiment AB2 to room temperature, bubbles of an immiscible liquid were observed inside the sample chamber. Raman spectra (Fig. 3b) suggest that they consist of newly formed hydrocarbons, in particular propane and isobutane, in very good agreement with similar observations made by Huang et al. (2017)

Huang, F., Daniel, I., Cardon, H., Montagnac, G., Sverjensky, D.A. (2017) Immiscible hydrocarbon fluids in the deep carbon cycle. Nature Communications 8, 15798. https://doi.org/10.1038/ncomms15798.

. The formation of propane and isobutane suggests that, already under these rather mild conditions, new C–C bonds may form rapidly, probably via some radical mechanism. In particular, the reaction of acetate to isobutane requires a complete rearrangement of the C–C bonds.

Full size image

Experiment AB3 reached maximum conditions of 400 °C and 3.50 GPa. After cooling to room temperature, some patches of dark material were visible in the solution and a small crystal of an alkali carbonate (not further identified) was detected by Raman spectroscopy, similar to observations made by Huang et al. (2017)

Huang, F., Daniel, I., Cardon, H., Montagnac, G., Sverjensky, D.A. (2017) Immiscible hydrocarbon fluids in the deep carbon cycle. Nature Communications 8, 15798. https://doi.org/10.1038/ncomms15798.

. The carbonate (with C⁴⁺) likely formed together with some hydrocarbons (containing C⁴⁻) by partial disproportionation of the acetate (with an average carbon oxidation state of zero). Despite these results, which indicate partial decomposition, some acetate was still detectable in the solution. The charge from experiment AB3 was then rerun in experiment AB4, which reached 600 °C and 4.65 GPa, essentially the predicted stability field of acetate. However, already at 3.99 GPa and 500 °C, no acetate was detectable in the fluid anymore (Fig. 4). Instead, the Raman spectrum of a carbonaceous material resembling highly disordered graphite was detected. No acetate could be detected in the fluid after cooling it back to room temperature; instead, abundant dark, carbonaceous material was visible inside the cell (Fig. S-3), which yielded Raman spectra resembling highly disordered graphite.

Full size image

Clear evidence for methane or CO₂ was not seen in any of the Raman spectra of the in situ experiments; however, traces may have remained undetected due to the proximity of the CH₄ stretching bands to those of the CH₃-group of acetate and of the CO₂ Fermi dyad to the main band of the diamond anvils.

top

Discussion

The results both from the piston cylinder and the in situ diamond cell experiments appear very consistent and suggest that, above 500 °C, acetate is not stable anymore in aqueous fluids and decomposes to a carbonaceous material. This observation has important consequences for the fate of subducted carbon. While carbon in oxidised form, i.e. as carbonates, may be readily mobilised during the release of aqueous fluids, the fate of reduced carbon during subduction is much less understood (Plank and Manning, 2019

Plank, T., Manning, C.E. (2019) Subducting carbon. Nature 574, 343–352. https://doi.org/10.1038/s41586-019-1643-z

). Reduced carbon in the form of former organic material is abundant in many sediments. If acetate, which is highly water-soluble, were stable in aqueous fluids, such fluids would likely be very efficient in returning subducted carbon back to the surface. In the absence of such mobile organic species, graphitic material is likely almost insoluble in water and very poorly soluble in silicate melts (e.g., Eguchi and Dasgupta, 2017

Eguchi, J. , Dasgupta, R. (2017) CO₂ content of andesitic melts at graphite-saturated upper mantle conditions with implications for redox state of oceanic basalt source regions and remobilization of reduced carbon from subducted eclogite. Contributions to Mineralogy and Petrology 172, 12. https://doi.org/10.1007/s00410-017-1330-8

), at least in an intermediate range of oxygen fugacities, where neither the oxidation to CO or CO₂ nor the reduction to CH₄ is thermodynamically favourable. Such graphitic material may therefore be recycled deep into the mantle during subduction. However, fluxing by external, particularly oxidised, fluids may be an efficient mechanism to dissolve reduced carbon and return it to the surface. Field evidence for the efficiency of such a process was presented by Vitale Brovarone et al. (2020)

Vitale Brovarone, A., Tumiati, S., Piccoli, F., Ague, J.J., Connolly, J.A.D., Beyssac, O. (2020) Fluid-mediated selective dissolution of subducting carbonaceous material: Implications for carbon recycling and fluid fluxes at forearc depths. Chemical Geology 549, 119682. https://doi.org/10.1016/j.chemgeo.2020.119682

. Moreover, at least at relatively shallow depth corresponding to pressures of 1 GPa, organic carbon may be significantly more fluid-soluble than well ordered graphite (Tumiati et al., 2020

Tumiati, S., Tiraboschi, C., Miozzi, F., Vitale-Brovarone, A., Manning, C.E., Sverjensky, D.A., Milani, S., Poli, S. (2020) Dissolution susceptibility of glass-like carbon versus crystalline graphite in high-pressure aqueous fluids and implications for the behavior of organic matter in subduction zones. Geochimica et Cosmochimica Acta 273, 383–402. https://doi.org/10.1016/j.gca.2020.01.030

).

The causes for the discrepancy between our experimental results and theoretical predictions (Sverjensky et al., 2014

Sverjensky, D.A., Stagno, V., Huang, F. (2014) Important role for organic carbon in subduction-zone fluids in the deep carbon cycle. Nature Geoscience 7, 909–913. https://doi.org/10.1038/ngeo2291

, 2020

Sverjensky, D., Daniel, I., Vitale Brovarone, A. (2020) The changing character of carbon in fluids with pressure. In: Manning, C.E., Lin, J.-F., Mao, W.L. (Eds.) Carbon in Earth’s Interior. Geophysical Monograph 249. American Geophysical Union, Washington, D.C., John Wiley and Sons, Inc., Hoboken, NJ, 259–269. https://doi.org/10.1002/9781119508229.ch22

) are uncertain. One possibility is that the causes are in the parametrisation of the Helgeson-Kirkham-Flowers model (Shock et al., 1992

Shock, E.L., Oelkers, E.H., Johnson, J.W., Sverjensky, D.A., Helgeson, H.C. (1992) Calculation of the thermodynamic properties of aqueous species at high pressures and temperatures. Effective electrostatic radii, dissociation constants and standard partial molal properties to 1000 °C and 5 kbar. Journal of the Chemical Society, Faraday Transactions 88, 803–826. https://doi.org/10.1039/FT9928800803

). Here, the caloric and volumetric parameters for organic species may be poorly constrained. Uncertainties in some of the most basic properties of water at high pressure and high temperature may also limit the reliability of thermodynamic models. The predicted stability of ionic species in aqueous fluids very strongly depends on the precise value of the dielectric constant. However, up to now, direct experimental measurements of the dielectric constant have been limited to 550 °C and 0.5 GPa (Heger et al., 1980

Heger, K., Uematsu, M., Franck, E.U. (1980) The static dielectric constant of water at high pressures and temperatures to 500 MPa and 550 °C. Berichte der Bunsengesellschaft für Physikalische Chemie 84, 758–762. https://doi.org/10.1002/bbpc.19800840814

). Moreover, while the stability of acetate as a main carbon species in deep subduction fluids appears rather unlikely in the light of the experimental results presented here, the stability of minor concentrations of organic species is still possible and requires further investigation. For example, oxalate (COO)₂²⁻, the most simple dicarbonic acid ion, is known to form very strong complexes with many cations in aqueous solutions (e.g., Krishnamurty and Harris, 1961

Krishnamurty, K.V., Harris, G.M. (1961) The chemistry of the metal oxalato complexes. Chemical Reviews 61, 213–246. https://doi.org/10.1021/cr60211a001

). Already minor concentrations of oxalate could therefore have a significant effect on the chemical transport properties of subduction zone fluids.

top

Acknowledgements

We thank Matthieu Galvez and an anonymous referee for constructive reviews which helped to improve the manuscript.

Editor: Horst R. Marschall

top

References

top

Supplementary Information

The Supplementary Information includes:

• Experimental Methods
• Figures S-1 to S-4
• Supplementary Information References

Download the Supplementary Information (PDF).