地球最古老的生命跡象或許就潛藏在加拿大的岩石當中

原文網址：http://www.nature.com/news/oldest-traces-of-life-on-earth-may-lurk-in-canadian-rocks-1.22685

地球最古老的生命跡象或許就潛藏在加拿大的岩石當中

研究人員發表了存活在39.5億年前的生物留下的化學訊號，但有眾多學者對此抱持懷疑

Alexandra Witze

顯微鏡下的黑色石墨顆粒，可能具有地球遠古生命留下的訊號。 Credit: Komiya et al. Nature, 10.1038/nature24019

一項新研究提出加拿大東北方的古老岩石含有超過39.5億年前的生物留下的化學痕跡。如果此發現屬實，就會成為地球所有已知生命跡象中最早的紀錄。

對於某些人來說，此研究增添了更多證據暗示地球初期充斥各種不同形式的生物。「接受這項事實吧！」倫敦大學學院的地球化學家Dominic Papineau如此表示。他在三月跟其他學者發表的一篇研究中，於加拿大魁北克發現可能為微生物的化石，其年代至少可以追溯至37.7億年前。

但是其他人則懷疑這項最新研究是否禁得起更仔細的檢驗。先前許多聲稱找到遠古生命的研究都遭到相當猛烈的質疑聲浪——部份是因為數十億年前形成的岩石歷經劇烈高溫作用且嚴重擠壓，使得其中蘊含的地質語言變得難以解讀；另一部份則是因為生物留下的化學訊號跟沒有生物參與的化學反應，實際上相當難以區分開來。

「當我讀到這篇研究我想說『又來了。』」斯德哥爾摩瑞典自然史博物館的地質學家Martin Whitehouse表示。2002年他參與的研究批評了一篇在格陵蘭發現遠古生命的類似報告。

關鍵岩塊

這篇最新研究調查了位在加拿大北方的一套岩石，統稱為Saglek岩塊。由東京大學小宮剛和佐野有司領導的團隊在2011至2013年間探訪此區，在武裝警衛持續警戒北極熊的同時，團隊分散在露頭各處的岩石蒐集樣品。

於9月28日當期的《自然》，研究人員發表他們對Saglek地區進行的碳同位素分析結果，樣品類型分別是整個磨碎的岩石和個別石墨顆粒。在某些樣品中，他們發現其含有的碳-13對碳-12的比例相對來說較低。碳-12是比較輕的同位素，它的原子核比碳-13少了一顆中子。生物傾向於利用碳-12來製造有機化合物，因此微生物曾經居住過的場所，像是Saglek的岩石，就會含有豐富的輕同位素。

小宮表示在這些古老且高度變形的岩石中找到生命跡象「令他們相當驚訝且興奮」。他說他和團隊排除了其他會讓碳同位素比例產生偏差的可能解釋，像是菱鐵礦的分解。此外，石墨結晶時的溫度似乎跟他們附近的岩石在擠壓加熱的過程中經歷的溫度大致相同，顯示石墨並非是之後才進入岩石內部的汙染物質。

研究人員從加拿大拉布拉多稱作Saglek岩塊的地區採集岩石樣品。Credit: Komiya et al. Nature, 10.1038/nature24019

在之前發表的論文中，小宮的團隊描述了Saglek岩塊的地質史，並利用一種變質岩――片麻岩中的古老鋯石晶體進行鈾―鉛定年，結果顯示片麻岩的年代為39.5億年。他們肯定含有生命訊息的石墨至少有這麼老，因為石墨位在看似被39.5億年的片麻岩穿過的岩石之中，所以團隊推測石墨的年代會早於片麻岩。

但此描述反而促使其他科學家提出警語。「這些石墨沉積物的年代比作者宣稱的年輕許多。」華沙波蘭科學院地質科學研究所的地質學家Monika Kusiak，與澳洲伯斯科廷大學的地質學家Daniel Dunkley表示。他們都曾經進行研究企圖闡明Saglek岩塊的地質史，並主張研究人員推論中被片麻岩穿過的岩石，實際年代比片麻岩還要年輕。

「這篇論文中其他所有內容都構築在地質年代之上，但地質年代本身並沒有建造得十分穩固。」Whitehouse補充，「整篇論文就像是一棟搖搖欲墜的危樓。」

小宮表示他支持由他的團隊做出的解釋。

跟拉布拉多相關的研究

對此成果的爭議跟1996年開始發生的爭論十分相似，當時一組美國―澳洲―英國的聯合團隊發表他們在格陵蘭西南部外海Akilia島取得的岩石中，內部的磷灰石顆粒含有生物作用過的石墨，年代至少有38.6億年之久。這項研究在許多方面都受到批判，但還是有些人堅稱Akilia島的岩石含有遠古生命留下的蹤跡。

拉布拉多包括Saglek岩塊在內的古老岩石可以直接跨過拉布拉多海連結到格陵蘭。「它們有許多相似之處。」哥本哈根丹麥自然史博物館的地質學家Minik Rosing表示。他的研究領域為格陵蘭的岩石中可能由生命留下來的化學訊號。「這邊的地質狀況極為複雜。」

他在讚賞小宮團隊的同時表示仍有許多研究尚待完成。他說：「這是一個良好例子呈現出相當精細且品質高超的分析成果。」然而，他也表明當地岩石十分複雜，使得研究用的石墨「就算通過各種方法驗證，其年代可能僅有27億年這般年輕。」

Oldest traces of life on Earth may lurk in Canadian rocks

Researchers report chemical evidence of organisms that lived 3.95 billion years ago, but scepticism abounds.

The microscope view shows dark grains of graphite that could contain traces of ancient life on Earth. Credit: Komiya et al. Nature, 10.1038/nature24019

Ancient rocks in northeastern Canada could contain chemical traces of life from more than 3.95 billion years ago, a new study suggests¹. If confirmed, the finding would be among the earliest known signs of life on Earth.

To some, the work adds to growing evidence that the young Earth was teeming with many different kinds of organism. “Accept it!” says Dominic Papineau, a geochemist at University College London who, in March, co-authored a report of possible fossilized microbes from Quebec that date back at least 3.77 billion years².

But others are sceptical that the latest work will hold up to scrutiny. Many previous claims of ancient life have been hotly contested — in part because rocks that formed billions of years ago have been severely heated and squished, making the geological context hard to interpret, and in part because the chemical traces of life can be difficult to distinguish from reactions that do not involve living organisms.

“When I read this I thought, ‘here we go again,’” says Martin Whitehouse, a geologist at the Swedish Museum of Natural History in Stockholm who, in 2002, co-authored a study³ that criticized a similar report of ancient life in Greenland.

Rock block

The latest work investigated a set of rocks from northern Labrador, known collectively as the Saglek block. A team led by Tsuyoshi Komiya and Yuji Sano of the University of Tokyo visited the area between 2011 and 2013, fanning out among the rocky outcrops to gather samples while armed guards kept watch for polar bears.

In the 28 September issue of Nature, the researchers report analysing carbon isotopes in powdered rock and in individual graphite grains from the Saglek area. In some of the samples, they found relatively low amounts of the isotope carbon-13 compared to carbon-12, a lighter isotope with one less neutron in its nucleus. Organisms prefer to use carbon-12 to make organic compounds, and so material in which microbes once lived — like the Saglek rocks — becomes enriched in the lighter isotope.

Finding evidence of life in these ancient, highly deformed rocks “is surprising and exciting”, says Komiya. He says that he and his group have ruled out other possible explanations for the skewed ratio of carbon isotopes, such as decomposition of the mineral siderite. And the graphite seems to have crystallized at roughly the same temperature as that experienced by the rocks around it as they were squeezed and heated, which suggests the graphite isn’t just contamination that arrived later.

Researchers sampled rocks from an area in Labrador, Canada, called the Saglek block. Credit: Komiya et al. Nature, 10.1038/nature24019

In previous papers^4, 5, Komiya’s team described the geological history of the Saglek block, and used uranium–lead dating on ancient zircon crystals inside a type of metamorphic rock called gneiss to conclude that the gneiss was 3.95 billion years old. They say that the graphite containing hints of life must be at least that old, because it lies within rocks that are apparently shot through with — and thus presumably older than — the 3.95-billion-year-old gneiss.

Yet that scenario prompts other scientists to raise a warning flag. “The graphite is in much younger sediment than the authors claim,” say geologists Monika Kusiak of the Institute of Geological Sciences at the Polish Academy of Sciences in Warsaw and Daniel Dunkley of Curtin University in Perth, Australia, who have been working to unravel the geological history of the Saglek block. They argue that the rocks that are supposedly shot through with the gneiss are not, in fact, older than it.

“Everything else in this paper is resting on the geochronology, and the geochronology is just not well-founded,” adds Whitehouse. “It’s a house of cards.”

Komiya says that he stands by his team’s interpretation.

Labrador links

The debate over the results resembles one that began in 1996, when a US–Australian–British team reported finding biologically altered graphite from at least 3.86 billion years ago inside grains of a mineral called apatite, in rocks on Akilia island off southwestern Greenland⁶. That work was criticized on many fronts⁷, although some still argue that Akilia's rocks contain traces of ancient life.

Labrador’s ancient rocks, including the Saglek block, lie directly across the Labrador Sea from Greenland. “There are a lot of similarities,” says Minik Rosing, a geologist at the Natural History Museum of Denmark in Copenhagen who has studied possible chemical traces of life in Greenland rocks⁸. “It's extremely complex geology.”

He applauds Komiya's team, but says that more remains to be done. “It's a good example of very sophisticated and high-quality analytical work,” he says. But the rocks are so complicated, he explains, that the graphite being studied “could be as young as 2.7 billion years old, and still pass all the tests”.

原始文章：Alexandra Witze. Oldest traces of life on Earth may lurk in Canadian rocks. Nature, 2017. doi:10.1038/nature.2017.22685

火山學：海水減少，岩漿增多

原文網址：https://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo3040.html

火山學：海水減少，岩漿增多

減壓作用是地函發生部分熔融的主要方式之一。一般來說，在中洋脊下方或者隱沒帶後方，地函岩石上湧的時候便會發生此種熔融現象。對此作用一項合理的推論是，岩石感受到的壓力變化或許可以無關於穩定而緩慢的地函對流，而是施加於地球表面的壓力發生迅速變化。過往科學家曾特別提出冰層後退或是海平面變動導致的壓力變化，可以調節大範圍火山系統產生與噴發的岩漿多寡。然而，它在不同地體構造的影響規模有多大，以及在地質紀錄中會留下什麼樣的痕跡，卻都還是爭論中的議題。Pietro Sternai和其同僚在《自然―地質科學》撰寫的論文中，呈現出的證據指出大約600萬年前的墨西拿鹽度危機(Messinian salinity crisis)期間，一部分地中海蒸發的同時火山活動也跟著增加。他們提出當時海平面降低數公里造成地函受到減壓作用影響，是最能解釋岩漿活動為何激增的原因。

科學家長久以來認為全球水圈變化連帶使地表承受的荷重發生改變，可以從上往下擾動地函受到的壓力，使得減壓熔融發生。舉例來說，厚達2公里的冰層在2000年內消失可以迅速減輕地函受到的壓力，其效率是地函持續上湧有關的減壓作用的十倍以上。此效應能解釋冰島在10000至8000年前記錄到的火山活動增加現象。類似地，科學家現在認為冰河期―間冰期循環中，跟海平面動盪有關的壓力變化，可以對中洋脊產生的熔岩量有10%左右的影響。

Sternai等人探討墨西拿鹽度危機對岩漿活動造成的作用，而把全新觀點帶入這項議題之中。墨西拿鹽度危機事件標記了一段時期，當時地中海大量蒸發並在盆地內部堆積大量鹽分。究其原因為直布羅陀海峽關閉，但海平面大幅下降也可能牽涉其中。地中海區域的地體構造相當複雜，原因為非洲板塊和歐亞板塊在此聚合而產生了許多小型隱沒系統，提供底下的高溫地函許多可以上湧、熔融，然後在地表點燃火山活動的空間。不過，彙整火山活動發生時間的現有定年數據之後，研究人員指出在560萬年前墨西拿鹽度危機期間，這些火山系統發生岩漿噴發和入侵事件的數目變得異常地多。

這段火山活動高峰期大多數發生在地中海盆地邊緣的火山。Sternai等人提出其為下方地函因為海平面變化導致的減壓作用而產生的反應。為了建立兩者之間關聯的可信度，他們首先計算以公里為單位的海平面下降對地表荷重造成的改變有多大，接著確認這些異常火山活動的岩漿庫和來源所處的地函和地殼，其中是否有部分能感受到此荷重變化。雖然跟全球海平面變化對中洋脊造成的效應比起來，此荷重變化顯得微小許多，但是發生在地中海中央附近的減壓作用仍然可以透過間接方式影響遠處位在盆地邊緣的火山系統。地質動力學的模擬結果顯示出如果應力變動的幅度夠大且速度夠快，甚至還能透過黏彈性的岩石圈和上部軟流圈傳遞到距地中海海岸數百公里遠的內陸地區。

圖一：墨西拿鹽度危機期間造成地中海火山活動增加的機制

(i) 地中海盆地在大約570至550萬年前發生的大規模蒸發現象和海平面迅速下降，可以減少地殼和下方地函受到的壓力。(ii) 地函減壓導致部分熔融。(iii) 另一方面，地殼減壓的同時產生拉伸，讓更多岩漿以岩脈的形式往地表移動。(iv) 最終導致地表火山活動增加。Sternai和其同僚利用數值模型顯示這一連串的事件，可以解釋地中海盆地的火山活動為何在大約560萬年前出現一段異常高峰期。

模型也測試了不同的負載路徑，也就是岩石圈因為鹽分沉澱而受到的壓力增加，以及蒸發導致水位下降而造成的壓力減輕，兩者的綜合效應。模型顯示唯有快速且大規模的海平面下降可以對岩漿來源地區和岩漿庫受到的壓力產生夠大的影響，這和墨西拿鹽度危機與海平面劇烈下降有密切關係的說法一致。他們提出發生在地表的減壓作用不只能提升地函的熔融作用，同時也減少地殼受到的應力，有利於岩漿發生岩脈入侵作用。遺憾的是，可以同時處理剛性地殼變形和岩漿在其中如何移動的地質動力學自洽(self-consistent)模型還很稀少，且額外產生的熔融物質到達地表需要多少時間也還沒有確定下來。若要精確定量岩漿傳輸和入侵對瞬變作用力的敏感度，還需要發展出更先進的模型基礎。

要辨識出從地表傳遞至下方的力量對岩漿活動的調節作用還有其他難處，像是岩漿活動的時間序列有所缺失或是取樣可能出現偏差。由於我們只能研究我們能取得的地質紀錄，因此要降低這些爭議並不容易。另外，要在有許多內部變因的複雜系統當中辨識出特定來源的外在變因——在此例中為減壓作用——也格外具有挑戰性。研究墨西拿鹽度危機的優勢之一是，它就像是一起突發事件，發生在其他對岩漿活動有影響的地質作用力大都可視為平穩狀態的時期。Sternai和同僚採用的方法為彙整大範圍地區的數據，以辨識出整個泛地中海地區不同火山事件的共時性，這讓他們可以宣稱其發生原因並非個別岩漿系統本身具有的波動。要進一步證實這些說法需要對其背景特性有通盤瞭解；找出地中海整區未受影響下的岩漿產量；並納入減壓熔融以外的作用來解釋不同的熔岩產生模式。

Sternai和共同研究人員顯示墨西拿鹽度危機期間迅速且大規模的海平面下降可能會減少地函受到的壓力，而促進岩漿生成並在地表引發更頻繁的火山活動。他們提出的關係可以激勵研究人員蒐集高解析度的野外數據來更加精確地定年地中海的火山活動，並且發展新方法建構岩石圈—岩漿的耦合動力學模型。地表荷重對岩漿活動的影響在地質動力學中仍然是倍受爭議的問題。認識此交互作用背後運行的物理機制的重要性不只是能讓我們更加瞭解岩漿系統，也可以利於我們評估當地表發生變化火山活動會有何反應，這或許能讓我們找出全球氣候系統當中尚未發現的反饋作用。

Jean-Arthur Olive

Volcanology: When less water means more fire

Decompression is the primary way to partially melt Earth's mantle. Such melting occurs routinely as mantle rocks upwell beneath a mid-ocean ridge or behind a subduction zone. A corollary of this process is that rocks may also feel changes in pressure that have nothing to do with slow and steady mantle convection, but instead relate to rapid shifts in the loads that weigh on Earth's surface. In particular, pressure changes due to ice-sheet retreat or sea-level variability have been proposed to modulate the amount of magma produced and erupted in a wide range of volcanic systems^{1, 2, 3, 4}. The magnitude of this modulation across different geodynamic settings, and its signature in the geological record, however, remain a matter of debate^{5, 6}. Writing in Nature Geoscience, Pietro Sternai and colleagues⁷ present evidence for enhanced volcanism coincident with partial evaporation of the Mediterranean Sea during the Messinian salinity crisis, about six million years ago. They argue that this boost in magmatism is best explained by decompression of the mantle caused by a kilometre-scale drop in sea level at that time.

It has long been thought that rapid shifts in surface loads due to changes in the global hydrological cycle could perturb mantle pressure from the top down, causing decompression melting. For example, the removal of an ice sheet approximately 2-km thick within about 2,000 years would decompress the mantle ten times faster than decompression associated with steady mantle upwelling. This effect can account for the increase in volcanic productivity documented in Iceland between 10,000 and 8,000 years ago^{1, 2}. Similarly, pressure changes associated with glacial/interglacial sea-level fluctuations are now thought to modulate the magma supply at mid-ocean ridges by about 10% (refs 4,5).

Sternai et al.⁷ bring an original perspective into this discussion by considering the magmatic consequences of the Messinian salinity crisis. This crisis marked an episode of massive evaporation and salt deposition in the Mediterranean basin that was triggered by the closure of the Gibraltar Straight, and was potentially associated with spectacular sea-level drop⁸. The Mediterranean region is tectonically complex because the convergence between the African and Eurasian plates has created several small-scale subduction systems⁹ that provide numerous opportunities for hot mantle to ascend, melt and fuel volcanism at the surface. However, using a compilation of existing age data that constrain the timing of volcanism, the researchers show that an anomalously high number of eruptive and intrusive events occurred in these systems during the Messinian salinity crisis, around 5.6 million years ago.

Sternai et al.⁷ propose that this pulse in volcanism, which occurred mostly along the edges of the Mediterranean basin, was a response to sea-level-driven decompression of the underlying mantle. To convincingly establish this connection, they first calculate the magnitude of the change in surface load associated with a kilometre-scale sea-level drop, and then determine whether this change in load would have been felt in parts of the mantle and crust that sourced and stored the anomalous volcanism. Although the latter may seem trivial when considering the effect of global sea-level change on a mid-ocean ridge, it is not straightforward that unloading near the centre of the Mediterranean Sea would have consequences for volcanic systems located far away from the centre, along the edge of the basin (Fig. 1). Geodynamic modelling, however, reveals that if the stress fluctuations were sufficiently large and rapid, they could have been transmitted through the visco-elastic lithosphere and upper asthenosphere, hundreds of kilometres inland beyond the Mediterranean shores.

Figure 1: Mechanisms for enhanced volcanism in the Mediterranean during the Messinian salinity crisis.

Large-scale evaporation and rapid sea-level fall in the Mediterranean basin about 5.7 to 5.5 million years ago would have unloaded the crust and underlying mantle (i). Mantle decompression leads to partial melting (ii), whereas crustal unloading and extension allows more magma to migrate towards the surface via dykes (iii), generating increased volcanism at the surface (iv). Sternai and colleagues⁷ use a numerical model to show that this sequence of events could explain an anomalous pulse of volcanism in the Mediterranean basin about 5.6 million years ago.

Various loading paths, representing scenarios of salt precipitation that loads the lithosphere and evaporative drawdown that unloads the lithosphere, were tested in the models. Only a rapid and large magnitude sea-level drop can exert a sizeable pressure modulation on the magma production and storage areas, which is consistent with the idea that the Messinian salinity crisis was associated with significant sea-level change. Unloading at the surface is proposed to not only generate additional melt in the mantle, but also to promote volcanic eruptions because reduced stress in Earth's crust creates conditions that are favourable for initiating magma-filled dyke intrusions. Unfortunately, geodynamic models that self-consistently handle deformation and melt transport through the brittle lithosphere are rare¹⁰, and the time necessary for any extra melt to reach the surface is debated. To accurately quantify the sensitivity of magma transport and extrusion to transient forcings warrants the development of novel modelling frameworks.

Additional difficulties in identifying top-down controls on magmatism include incomplete preservation and potential sampling bias in the time-series of magmatic activity. Mitigating these issues is difficult to do because we can only work with the geological record that is available to us. Identifying a particular source of external variability — in this case surface unloading — in a complex system with multiple sources of internal variability is especially challenging. One advantage of studying the Messinian salinity crisis is that this was a sudden pulse-like event during which other geodynamic forcings on magmatic activity can be considered mostly steady. Sternai and colleagues adopt a large-scale data collection approach to identify synchronicity in volcanic events across the entire pan-Mediterranean region, which argues against intrinsic fluctuations specific to each magmatic system. To further test these ideas will require a thorough characterization of the background, unperturbed magma output of the Mediterranean domain, and to account for the diversity of melt-generation modes beyond just decompression melting.

Sternai and co-workers⁷ show that rapid and large-scale sea-level fall during the Messinian salinity crisis could have decompressed the mantle, boosted magma production and increased surface volcanism. This proposed link will motivate the collection of high-resolution field data that better constrain the timing of volcanism in the Mediterranean, along with the development of novel approaches for coupled lithosphere–magma dynamics. The influence of surficial loading on magmatic activity remains a contentious question in geodynamics. Knowledge of the physical processes at play during such interactions is not only critical for our understanding of magmatic systems, but also because appreciation of the volcanic response to changes at Earth's surface may allow the identification of previously unrecognized retro-actions in the global climate system³.

原始文章：Jean-Arthur Olive. Volcanology: When less water means more fire, Nature Geoscience, 2017. doi:10.1038/ngeo3040

地球化學：地殼改變造成大氣氧化

地球化學：地殼改變造成大氣氧化

J. Elis Hoffmann

大氣中的氧對於複雜生命的存活與生長來說至關重要。在地球歷史中有將近一半的時光，地球大氣處於缺氧狀態。接著，於23至24億年前左右發生的大氧化事件中(Great Oxygenation Event ,GOE)氧氣含量急遽上升，開啟了大氣氧濃度逐步上升的趨勢。是什麼原因造成了大氧化事件時氧氣濃度飆升，至今仍未明瞭。Smit和Mezger在《自然―地質科學》(Nature Geoscience )撰寫的文章中，提出約莫在大氧化事件發生的10億年之前，上部大陸地殼總成分產生了改變，造成地表化學反應消耗的氧氣減少，使得大氣中的氧氣可以累積起來。

地球大氣中的氧氣大部分是經由植物(包括浮游植物)或藍綠菌進行的光合作用產生。因此大氧化事件或許跟藍綠菌族群增加有關。然而，在大氧化事件時藍綠菌已經能形成稱作「氧氣綠洲」的小規模有氧環境，所以它們演化出來的時間很可能比事件早上許多。較早之前的大氣氧氣增加可能會被海水和大陸的成分抵銷：像是從太古宙海底熱泉流入海洋的還原鐵，以及年輕地球特有鐵含量較豐富的基性地殼。任何出現在大氣中的自由氧都會跟這類還原物質反應，造成氧被吸收而阻止它們在大氣中累積。因此，只有地殼成分出現變化之後，氧氣才能在大氣中聚積。不過，妨礙大氣氧化的吸收過程其確切性質仍然眾說紛紜。

Smit和Mezger彙整出全球碎屑沉積物的化學成分資料庫，年代涵蓋了地球歷史大部分時間。它們利用稀有元素鉻和鈾的比例(以太古宙之後來自大陸地殼的沉積物當作現今平均值進行標準化)，重建出地殼礦物組成隨時間如何變化。具體來說，研究人員是利用了難熔礦物(refractory mineral)碎屑中鉻和鈾的分異作用彼此不同的特性。除非受到有氧風化作用而影響到溶解度，不然這兩種元素在岩石裡都是屬於不活動元素。有氧風化進行時，鉻會跟鉻鐵礦之類的礦物結合，鈾則會跟鋯石之類的礦物結合。但是，因為鋯石幾乎不會存在於基性或超基性地殼中，所以沉積物的鉻/鈾比可以用來指示源區成分為何。研究人員的觀察結果跟過去研究一致：在大約33至24億年前地殼成分發生了改變，上部大陸地殼的性質從基性―超基性岩為主轉變成酸性岩為主；礦物組成也從橄欖石和輝石為主，變成含有石英和長石之類的礦物。

Smit和Mezger根據他們建立的鉻/鈾資料庫，提出基性―超基性地殼中的礦物在熱水置換過程中會因為水合作用形成蛇紋石類礦物。此類礦物在進行脫氧作用時會把OH原子團納入自己的結晶構造並釋放出氫氣。這種反應造成母岩轉變成典型太古宙綠岩帶中出露的蛇紋岩(圖一)。由於蛇紋岩化會促成還原性物質產生，因此有蛇紋岩在的地表水鹼度可能會變得相當高並影響藍綠菌的生長環境，使其可以吸收細菌產生的氧氣。

圖一：西格陵蘭南方伊蘇阿綠岩帶的蛇紋岩，其年代超過37億年。

Smit and Mezger現在提出這種太古宙上部大陸地殼的常見成分或許能吸收氧氣，因而抑制早期地球大氣中的氧含量。(Source: Elis Hoffmann)

地殼從基性為主轉變成較為酸性可能跟全球地質作用的動力機制改變有關，從靜止蓋層(stagnant lid)的構造作用型態變成開始跟今日運行的板塊構造作用較為類似。隨著隱沒作用開始進行，基性上部地殼受到侵蝕並循環回地函，造成地殼中層演化程度較高的酸性地殼露出地表。基性地殼減少造成露出地表的蛇紋岩變少，使得蛇紋岩化作用再也無法成為大量消耗氧氣的機制，因此讓大氣的氧含量得以增加。

氧氣主要受到蛇紋岩化作用而消耗的模型跟過往提出的假說有所區隔。之前的研究提出藍綠菌族群擴大之後過一段時間大氣氧氣才上升的主因，是地表有許多類型的還原鐵可供氧氣進行還原作用。實際上，Smit和Mezger顯示在某些種類的岩石中鉻/鈾會跟成分有類似的變化，但它們的鐵含量卻都相同，意謂鐵含量並非控制大氣自由氧有多少的主要因素。此外，蛇紋岩化模型也允許太古宙晚期的上部大陸地殼成分存有某種程度的變異範圍。太古宙結束之際蛇紋岩對陸地逕流化學性質的影響逐漸淡去，或許造成海水上層部分處於有氧狀態，而海床之下發生的海洋地殼蛇紋岩化作用則讓太古宙之後的海水底層仍然處在缺氧狀態。

雖然蛇紋岩化作用確實是在大氧化事件發生前將還原作用的反應物帶到環境中的重要作用，但它並非唯一一種。其他可以把氧氣從大氣移除的重要消耗作用包括火山玻璃，基性與超基性岩中含硫礦物的風化；火山氣體變化、水下火山活動增加、以及海中熱泉噴出的鐵發生的沉澱反應。經過30億年之後，較為酸性的海床沉積物進入隱沒帶後可能會改變此處的氧化還原狀態，造成島弧跟陸弧火山噴出更多氧化氣體，而讓大氣進一步氧化。

Smit和Mezger顯示蛇紋岩化作用或許能緩衝太古宙地球的氧化還原狀態，減緩大氣氧氣的累積速率。研究人員現在應該要用數值模型來定量此種過程的重要性，並且跟硫同位素的紀錄比對――其為世上唯一涵蓋此年代的氧氣定量紀錄。如果太古宙時蛇紋岩化作用確實是相當重要的地質作用，則蛇紋岩的低密度可能也會對其他地質作用有深遠影響，像是早期地球的地殼循環過程和大陸地殼的穩定性。

Geochemistry: Oxygenation by a changing crust

Atmospheric oxygen is vital for the development and habitability of complex life. Earth's atmosphere lacked oxygen for nearly half of the planet's history. Oxygen levels then rapidly increased about 2.3 to 2.4 billion years ago, during the Great Oxygenation Event (GOE)^1, 2, 3, which initiated a stepwise rise in atmospheric oxygen concentrations¹. The reason for this increase during the GOE is unclear. Writing in Nature Geoscience Smit and Mezger⁴ propose that a change in the bulk composition of the upper continental crust about one billion years before the GOE would have caused a decrease in oxygen-consuming reactions at Earth's surface, allowing oxygen to build up in the atmosphere.

Most of Earth's atmospheric oxygen is produced via photosynthesis by plants (including phytoplankton) and cyanobacteria. The GOE is thus probably related to an increasing population of cyanobacteria. However, cyanobacteria are likely to have evolved much earlier than the GOE⁵, when they formed oxygen oases — small-scale oxidized environments. An earlier increase in atmospheric oxygen could have been mitigated by the composition of the oceans and continents^1,6, 7, 8: reducing agents such as reduced iron introduced by hydrothermal fluids into the Archaean oceans, as well as the more iron-rich mafic crust that characterized the younger Earth, may have reacted with any free oxygen, creating oxygen sinks that prevented its accumulation in the atmosphere. Only with a change in crustal composition could oxygen concentrate in the atmosphere. However, the precise nature of the sinks that hampered oxygenation of the atmosphere is ambiguous.

Smit and Mezger⁴ compile a global database of clastic sediment chemical compositions that covers most of Earth's history. Using the ratio of the trace elements Cr and U (normalized to the modern average value for post-Archaean sediment derived from continental crust), they reconstruct changes in crustal mineral composition through time. Specifically, the researchers take advantage of the different partitioning behaviours of Cr and U in refractory detrital minerals — both elements are immobile unless oxidative weathering influences the solubility. During oxidative weathering, Cr combines with minerals such as chromites, while U combines with minerals such as zircon. However, zircon is largely absent in mafic and ultramafic crust, so the Cr/U ratio can be used as a fingerprint of the composition of the source area for the sediments. In line with past work^6, 7, 8, the researchers observe a change in crustal composition about 3.3 to 2.4 billion years ago from a dominantly mafic–ultramafic upper continental crust that consisted mainly of minerals such as olivine and pyroxene, to a dominantly felsic upper crustal composition, comprising minerals such as quartz and feldspar.

Based on their Cr/U database, Smit and Mezger propose that minerals within the mafic–ultramafic crust were hydrated during hydrothermal alteration, forming serpentine minerals. Serpentines are a class of minerals that incorporate OH groups into their crystal lattice and release H₂ in oxygen-scavenging reactions⁴. These reactions transform the host rock into a serpentinite, which is found exposed in typical Archaean greenstone belts (Fig. 1). Since serpentinization triggers the production of reducing species⁴, surface waters in the presence of serpentinites would have been highly alkaline and could have influenced cyanobacterial habitats, acting as a sink for the bacteria-produced oxygen.

Figure 1: Serpentinites from the Isua Greenstone Belt of southern West Greenland, which is more than 3.7 billion years old.

Smit and Mezger⁴ now argue that these typical components of Archaean upper continental crust may have acted as an oxygen sink, suppressing oxygen levels in Earth's early atmosphere.   Source: Elis Hoffmann

The transition from a dominantly mafic to a more felsic crust was probably related to a changing style of global geodynamics, from a stagnant-lid tectonic regime towards the onset of plate tectonic processes similar to those active today⁹. With the onset of subduction, the upper mafic crust was eroded and recycled into the mantle, leaving the more evolved felsic crust from mid-crustal levels exposed at the surface of the continents. Removal of the mafic crust would reduce the amount of exposed serpentinites, minimizing serpentinization as an important sink for oxygen thereafter, and allowing oxygenation of the atmosphere.

The model of serpentinization as a primary oxygen sink stands independent from previous suggestions that the higher abundance of reduced iron species available for oxygen reduction was responsible for the delay in the rise of atmospheric oxygen following the expansion in cynobacterial populations. Indeed, Smit and Mezger show that the Cr/U ratios are similar in a number of rock types with variable compositions, yet the same iron abundance, implying that iron content was not the primary control on free oxygen availability. The serpentinization model also leaves some freedom for compositional heterogeneities within late Archaean upper continental crust. The dilution of the influence of serpentinites on the chemistry of continental runoff by the end of the Archaean may have led to partial oxidation of the upper ocean water column⁴, while sub-seafloor serpentinization in the oceanic crust kept the lower water column anoxic beyond the Archaean¹.

Although serpentinization is certainly an important process in introducing reduced reagents into the environment before the GOE, it is not the only process. Other key sinks for extracting oxygen from the atmosphere include weathering of volcanic glasses⁸, sulfur-bearing minerals in mafic and ultramafic rocks⁷, a change in volcanic gas composition¹⁰, increased subaerial volcanism¹¹, and hydrothermal Fe precipitation in seawater³. After three billion years, the subduction of more felsic ocean floor sediments may have changed the redox conditions in subduction zones, potentially leading to more oxidized gases being emitted at arc volcanoes, triggering further oxygenation⁷.

Smit and Mezger⁴ show that serpentinization may have buffered the redox conditions on the Archaean Earth, mitigating the accumulation of atmospheric oxygen. The importance of this process should now be quantified using numerical models and compared with the sulfur isotope record — the only quantitative record of oxygen on Earth across this time period². If serpentinization was indeed a critical process during the Archaean, the low density of serpentine may also have had a profound impact on geodynamic processes such as crustal recycling and stabilization of continental crust on the early Earth.

原始論文：Matthijs A. Smit and Klaus Mezger. Earth’s early O₂ cycle suppressed by primitive continents. Nature Geoscience, 2017.  doi:10.1038/ngeo3030.

引用自：J. Elis Hoffmann. Geochemistry: Oxygenation by a changing crust. Nature Geoscience, 2017. doi:10.1038/ngeo3038

https://www.nature.com/ngeo/journal/vaop/ncurrent/full/ngeo3038.html