Dark Oxygen Production (DOP) was a hypothesis published by Sweetman et al. (2024)[1] to explain anomalous data from experiments in the Clarion-Clipperton Zone, Pacific ocean. These experiments recorded rising oxygen for the first few hours, when previous experiments of the same type had only ever recorded reducing oxygen due to consumption by benthic organisms.[2][3][4][5][6] In order to explain this phenomenon, Sweetman et al.[1] claimed that manganese nodules, which were present in many of the experiments, must be responsible for the oxygen production. It was thus given the name "dark oxygen production", borrowing the term "dark oxygen" that had previously been used for oxygen production by bacteria without sunlight. Sweetman et al.[1] claimed that in this case the oxygen was being produced by electrolysis, using measured voltages from the surface of the nodules as evidence. The article, published in nature geoscience, was widely reported in mainstream media and social media, and contributed to the ongoing debate around deep sea mining.[7] However, it has since emerged that there are many problems with the theory, and may even be an example of scientific misconduct since contrary results and pertinent information were omitted from the article.[7] The omitted data showed that manganese nodules could not have been responsible for the phenomenon. A collaborative rebuttal published by leading scientists and industry experts instead claims that the rising oxygen levels are just experimental artefact, and not due to any natural phenomenon.[7] Sweetman et al. have publicly stepped back from the two main conclusions in the article; that manganese nodules produce oxygen, and that they do this by electrolysis.[8]
Benthic chamber landers are a form of benthic lander where a box, or chamber, is lowered to the seafloor and partially penetrates the sediment, creating a sealed volume of water which is then separate from the surrounding water column. In this chamber, various sensors and sampling devices are used, including oxygen sensors which measure the uptake of dissolved oxygen from the seawater by organisms within the chamber. However, Sweetman et al.[1] published data from benthic chamber landers that appeared to show oxygen rising in many of their chambers during the first few hours of the experiment. Subsequently, further experiments were done by taking samples of the seafloor and incubating them in the lab, and these appeared to confirm rising oxygen in the presence of manganese nodules. This led to the hypothesis that manganese nodules were somehow causing or producing the rising oxygen. The mechanism was not explained until voltages were measured on the surface of the nodules, and a reported voltage of 0.95 V was used as evidence that the manganese nodules were splitting water via electrolysis, producing oxygen and, presumably, hydrogen (although no data on hydrogen was released). And so the theory was formed, that manganese nodules can split water by electrolysis and produce oxygen, and this was published in Nature Geoscience in 2024.[1] The article made bold claims about how this could influence our understanding of the oxygenation of the Earth, the emergence of life and even life on other planets. It quickly attracted a lot of attention, and the article was widely reported, becoming one of Nature Geoscience's most read articles.
Rebuttals
The Dark Oxygen hypothesis quickly attracted doubt from industry leaders and independent scientists, stating numerous issues with the research.[9] Some of the initial issues included that no energy source for electrolysis had been identified, and no other studies had observed the same phenomenon despite being conducted in similar places and in the presence of manganese nodules. Shortly, pre-print rebuttals emerged which brought solid claims against the findings, including that crucial data was omitted from the article in order to support the dark oxygen hypothesis. Direct rebuttals to the article are summarised here.
Downes et al. 2024 [10] revealed that some of the lander experiments conducted in 2018 did not contain manganese nodules, and yet still recorded rising oxygen. This directly contradicts the hypothesis that manganese nodules were responsible for the rising oxygen. They also revealed a control experiment where rising oxygen was recorded in the absence of sediment and nodules, further contradicting the hypothesis.
Nakamura 2024 [11] noted that there is no "missing" oxygen source in the deep sea that would require such a process to exist, and there isn't elevated oxygen above manganese nodule fields.
Tengberg et al. 2024 [12] also highlight that earlier experiments that recorded increasing oxygen did so without nodules present. They further analyse the data from all the benthic chamber landers and find that only 2 out of 32 might have produced reliable data. They note that rising oxygen is a known experimental artefact, and an indication that the data from that chamber should be discarded. They explicitly call on Nature Geoscience to retract the article.
Cuesta and Jaspars, 2025 [13] go into great detail about the thermodynamics of the electrolysis hypothesis, concluding that it violates the 1st and 2nd laws of thermodynamics.
Trellevik et al. 2024 [14] note many flaws in the data that are apparent, including inconsistencies in the oxygen measurements and that the voltages recorded were not sufficient to split water.
Subsequently, many of the authors involved in the rebuttals collaborated to produce a comprehensive and peer reviewed critique of the original article, published in Frontiers in Marine Science in 2025.[7]
Problems
This is a list of some of many issues raised with the research.
Some landers did not contain manganese nodules. In probably the leading problem for the hypothesis, which relies on manganese nodules being present, some of the lander experiments conducted in 2018 either didn't contain any nodules or only contained small manganese oxide particles: "No nodules were collected as they were either not present or were too small to be relevant".[15] Manganese oxide particles are extremely common in deep sea oxic sediments.[16] The fact that some of the landers did not contain nodules was not discussed or disclosed in the article, and the research that states this fact was not cited. Since some of the landers that recorded increasing oxygen didn't contain manganese nodules, then manganese nodules cannot be responsible for the rising oxygen.[7][12]
The timeline of events suggests a deliberate decision to attribute dark oxygen production to manganese nodules. The landers that recorded rising oxygen without manganese nodules were conducted in 2018[15], and yet further work was conducted on at least four separate expeditions in 2021, 2022 and 2023, including shipboard incubation of manganese nodules.[1] The fact that rising oxygen was observed without nodules in 2018 was seemingly disregarded during this further work, as other scenarios such as sediment-only incubations were not conducted, or if they were then they were not presented. This demonstrates a bias towards finding manganese nodules responsible for rising oxygen, when evidence pointed to the contrary. Furthermore, the lack of citation for the article that stated there were no nodules present[15], and the lack of discussion of this point, demonstrates a concerted effort to hide the fact that oxygen rose without nodules present. This is an example of data falsification, where data is manipulated, cherry picked or omitted to support a certain conclusion - a form of scientific misconduct.
Control experiments were omitted. On at least one occasion, the benthic chamber lander was lowered to just above the seafloor, the chamber doors were closed, and the incubation proceeded with just seawater present.[10][7] Two of the three chambers recorded rising oxygen levels in this scenario. This control experiment ruled out sediment and manganese nodules causing the rising oxygen. However, this control experiment was not included in the article by Sweetman et al.[1], and thus important information that rejects the primary hypothesis was omitted - another example of data falsification.
The findings are anomalous in comparison to previous work. The paper states "Our findings contrast with all published deep-sea benthic O2 flux studies" and yet this inconsistency is not expanded or commented upon.[1] No attempt is made to explain why rising oxygen was seen in this case but not in others, despite the fact that anomalous results are an indication of experimental artefact. See previous work for more details.
The electrolysis hypothesis has been shown to be thermodynamically impossible.[13][7]. Although the original article by Sweetman et al.[1] did state that an energy source had not been identified, this is a violation of the 1st law of thermodynamics[13], and the article passed the review process with this significant flaw in logic. Furthermore, Cuesta et al. state that "[the] suggestion that polymetallic nodules... act like batteries that split water violates the 2nd law of thermodynamics" because the reaction that produces oxygen would go uphill in Gibbs free energy.
The voltages recorded were not great enough to split water. Although the article presents the maximum value of 0.95 V in the text as supporting the argument for electrolysis, this value is still far below that needed to split water.[13] The 0.95 V value is a clear outlier, with the average of all the readings being around 0.024 V.[1] No explanation is given for why these values, which are insufficient to split water, were used as evidence to support the hypothesis rather than reject it. This would appear to be incompatible with the scientific method, particularly falsifiability, which states that a hypothesis must be capable of being falsified by conceivable data. Furthermore, Sweetman et al. also mention that the structure or chemical makeup of the manganese nodules may lower the voltage required for electrolysis, which is an example of ad hoc hypothesis, a way of protecting a hypothesis from falsification by adding extra assumptions that have not been tested. There are also questions around the entire concept of measuring voltage on the surface of the nodules. Downes et al. [7] states "Any two conductive materials immersed in an electrolyte exhibit a potential difference—this alone is not indicative of electrolysis".
Tengberg et al.[12] note that rising oxygen is a known artefact in benthic chamber lander experiments, and is an indication that the data from that chamber should be discarded. In an interview with Science magazine, Matthias Haeckel stated that they had previously used the same equipment that Sweetman et al.[1] used, and had detected trapped air bubbles.[9]
There is no "missing" oxygen source. Oxygen production by manganese nodules is not required to explain oxygen levels in the deep sea,[11] and there is no evidence of elevated oxygen levels above manganese nodule fields.[17]
Inaccurate oxygen measurements. The article presents various measurements of bottom water oxygen concentrations which vary wildly and do not accurately reflect the known oxygen concentration in that location.[10][7] This calls into question the accuracy of the oxygen sensors or the equipment used, and thus the validity of any conclusions made based on this data.
Concessions by the original authors
The criticisms of the study has lead to the authors of the article retreating from the two main findings. The first claim was that manganese nodules were responsible for rising oxygen. However, as early as September 2024, Andrew Sweetman conceded that some of the landers did not contain nodules, but only small particles of manganese oxide.[9] Later, in a 2025 AGU conference abstract about the dark oxygen hypothesis, manganese nodules are not mentioned, and instead only "manganese oxides" and "polymetallic oxides" are referred to.[8] This has likely been forced due to the revelation that the lander experiments from 2018 did not contain manganese nodules, and only sometimes contained manganese oxide grains. This is a hugely significant concession as it now implicates manganese oxide found anywhere in the ocean, not just manganese nodules, which are found in limited areas. However, since much of the evidence presented in Sweetman et al.[1] worked to link manganese nodules to the rising oxygen, it is not clear what evidence supports linking rising oxygen to other forms of manganese oxide, which is a common material in the deep ocean.[16]
The second concession is that electrolysis is not the mechanism that explains the rising oxygen, despite it being fundamental to the original paper. The 2025 AGU abstract states that "It is presently unclear what the mechanism behind DOP is", and electrolysis is not mentioned.[8]
Problems with claimed significance
In the paper and elsewhere in the media, Sweetman et al. made many claims, sometimes contrary, about the process and how important it might be to the benthic ecosystem, to the oxygenation of the Earth, and even to the emergence of life.[18]. However, there is a clear problem with these claims since the presence of oxygen is a pre-requisite for the formation of manganese oxides[7] - and thus, they could not exist without the Earth first being oxygenated by other means. Currently, the oxygenation of the Earth, known as the Great Oxidation Event, is thought to have been caused by photosynthetic bacteria, and therefore life was already present at that time. Therefore, the theory is "unlikely at best".[19]
In terms of the importance to the benthic ecosystem, the original article by Sweetman et al.[1] simultaneously suggests its importance whilst also saying that the results can not be extrapolated, and that the equipment used may have somehow caused or triggered the phenomenon, and so “we shouldn’t think these nodules are bubbling oxygen all the time”.[9] There is currently no evidence that dark oxygen production is important to the benthic ecosystem.[11][17]
Previous work
Sweetman et al. state that "Our findings contrast with all published deep-sea benthic O2 flux studies", and states that more work is needed to study the phenomenon.[1]. It is therefore important to examine the body of work that has already been done. Many previous studies have measured oxygen consumption rates in the deep sea, including in areas with manganese nodules present. None of these studies have reported rising oxygen levels, or any evidence of oxygen production. Some of these studies include:
Smith et al. (1983) measured oxygen consumption rates in the central and eastern North Pacific.[2] No evidence of oxygen production was reported.
Khripounoff et al. (2006) reported oxygen consumption from nodule fields in Clarion Clipperton Zone.[3] No evidence of oxygen production was reported.
Stratmann et al. (2018) and Vonnahme et al. (2020) reported oxygen consumption rates from nodule fields in the Peru Basin.[4][5] No evidence of oxygen production was reported.
An et al. (2024) reported oxygen consumption rates from nodule-rich areas of the Pacific Ocean.[6] No evidence of oxygen production was reported.
Anomalous results in the context of so much previous work is a strong indication of experimental artefact - an example of Twyman's law, which states that "the more unusual or interesting the data, the more likely they are to have been the result of an error of one kind or another". It is established that when a result deviates significantly from a body of previous work, the default interpretation should be that there is an artefact, not that a new phenomenon has been discovered.[20] However, although Sweetman et al.[1] note that their results are anomalous, and they attempt to address some possible sources of errors that might have occurred, they do not discuss or attempt to explain why their results differ so greatly from all previous work. In general, the scientific method requires that extraordinary results be repeated in order to be accepted by the scientific community, to build scientific consensus, but in this case the experiments have already been repeated multiple times by multiple groups of researchers, without seeing the same phenomenon. Furthermore, the study published by Sweetman et al. itself repeats the experiments multiple times, and yet still only sees the phenomenon in some of the experiments.[1] There is no explanation given as to why the phenomenon occurs in some experiments but not others. The most likely explanation is that sometimes the equipment functioned correctly, and oxygen production was not observed.[7]
Future work
On the basis of these findings, Andrew Sweetman received a grant from the Nippon Foundation for £2M over three years for a project that "seeks to understand the actual conditions surrounding dark oxygen [and] will involve surveying 11,000 meters below the ocean surface using newly developed, specialized experimental equipment, and in addition to identifying the source of dark oxygen, will attempt to clarify its role in deep-sea ecosystems."[21] The collaborative rebuttal by Downes et al.[7] states "key omissions, including data from control experiments showing that the DOP signal originates from the benthic chamber lander itself—not the nodules—strongly indicates that the reported oxygen increases are experimental artefact rather than a genuine geochemical process" and that the hypothesis "is incompatible with established knowledge and inconsistent with experimental evidence", and so this funding may have been "potentially misallocated on the basis of speculative assertions".
12345678910111213141516Sweetman, Andrew K.; Smith, Alycia J.; de Jonge, Danielle S. W.; Hahn, Tobias; Schroedl, Peter; Silverstein, Michael; Andrade, Claire; Edwards, R. Lawrence; Lough, Alastair J. M.; Woulds, Clare; Homoky, William B.; Koschinsky, Andrea; Fuchs, Sebastian; Kuhn, Thomas; Geiger, Franz; Marlow, Jeffrey J. (August 2024). "Evidence of dark oxygen production at the abyssal seafloor". Nature Geoscience. 17 (8): 737–739. Bibcode:2024NatGe..17..737S. doi:10.1038/s41561-024-01480-8.
12Smith, K. L.; Laver, M. B.; Brown, N. O. (September 1983). "Sediment community oxygen consumption and nutrient exchange in the central and eastern North Pacific1". Limnology and Oceanography. 28 (5): 882–898. Bibcode:1983LimOc..28..882S. doi:10.4319/lo.1983.28.5.0882.
12Khripounoff, Alexis; Caprais, Jean-Claude; Crassous, Philippe; Etoubleau, Joël (September 2006). "Geochemical and biological recovery of the disturbed seafloor in polymetallic nodule fields of the Clipperton-Clarion Fracture Zone (CCFZ) at 5,000-m depth". Limnology and Oceanography. 51 (5): 2033–2041. Bibcode:2006LimOc..51.2033K. doi:10.4319/lo.2006.51.5.2033.
12Stratmann, Tanja; Voorsmit, Ilja; Gebruk, Andrey; Brown, Alastair; Purser, Autun; Marcon, Yann; Sweetman, Andrew K.; Jones, Daniel O. B.; van Oevelen, Dick (September 2018). "Recovery of Holothuroidea population density, community composition, and respiration activity after a deep-sea disturbance experiment". Limnology and Oceanography. 63 (5): 2140–2153. Bibcode:2018LimOc..63.2140S. doi:10.1002/lno.10929.
12Vonnahme, T. R.; Molari, M.; Janssen, F.; Wenzhöfer, F.; Haeckel, M.; Titschack, J.; Boetius, A. (May 2020). "Effects of a deep-sea mining experiment on seafloor microbial communities and functions after 26 years". Science Advances. 6 (18) eaaz5922. Bibcode:2020SciA....6.5922V. doi:10.1126/sciadv.aaz5922.
12An, Sung-Uk; Baek, Ju-Wook; Kim, Sung-Han; Baek, Hyun-Min; Lee, Jae Seong; Kim, Kyung-Tae; Kim, Kyeong Hong; Hyeong, Kiseong; Chi, Sang-Bum; Park, Chan Hong (May 2024). "Regional differences in sediment oxygen uptake rates in polymetallic nodule and co-rich polymetallic crust mining areas of the Pacific Ocean". Deep Sea Research Part I: Oceanographic Research Papers. 207 104295. Bibcode:2024DSRI..20704295A. doi:10.1016/j.dsr.2024.104295.
123Cecchetto, Marta M.; Moser, Annabell; Smith, Craig R.; van Oevelen, Dick; Sweetman, Andrew K. (May 2023). "Abyssal seafloor response to fresh phytodetrital input in three areas of particular environmental interest (APEIs) in the western clarion-clipperton zone (CCZ)". Deep Sea Research Part I: Oceanographic Research Papers. 195 103970. Bibcode:2023DSRI..19503970C. doi:10.1016/j.dsr.2023.103970.
↑Trost, Katharina; Gennis, Robert B.; Allen, John F.; Mills, Daniel B.; Martin, William F. (April 2026). "Oxygen reductase origin followed the great oxidation event and terminated the Lomagundi excursion". Biochimica et Biophysica Acta (BBA) - Bioenergetics. 1867 (2): 149575. doi:10.1016/j.bbabio.2025.149575. PMID41338429.{{cite journal}}: CS1 maint: article number as page number (link)
This page is based on this Wikipedia article Text is available under the CC BY-SA 4.0 license; additional terms may apply. Images, videos and audio are available under their respective licenses.