It is a strange one. Metals are shiny, and this one certainly is, in a dark-grey, silvery way. It is so shiny that it can even be used as a telescope mirror. Metals tend to be hard, but this one certainly is not. It isn’t even a solid. Mercury is one of only two elements that are liquid at room temperature (the other one is bromine). It freezes (an unexpected word to use for a metal) at a chilly -38 C. A liquid state may seem to be a bit of a problem when using it as a mirror: pour it on the parabolic surface of a telescope mirror base, and it collects at the bottom. But uncooperative physics never stopped astronomers. If you let the mirror rotate (at just the right rate), the centrifugal force will make the mercury spread out, so that it covers the whole surface in a perfectly reflective layer. A slight problem is that it only works while the telescope is vertical: any other angle, and the mercury will sink down. This is a mirror that can’t point. You just have to hope that there is something interesting to look at at the zenith.
There are other limitations: you can’t use it in Antarctica or in the Polar regions (the mercury would freeze), and there is a tendency for your mirror to evaporate with rather poisonous vapour. But you could build a massive ten-meter mirror in your back garden for just a few million. When astronomers dream, they dream of mercury.
But mercury has more uses than just in astronomy, and it reveals more than just the Universe. It also provides a window to the past and its ancient volcanism.
Let’s first look at the chemistry. The periodic table is a good place to start. It can take a while to find mercury: it is quite far down, in the 6th row: look for element number 80. It is in illustrious company, coming immediately after platinum and gold. The two following elements are thallium and lead, if you are interested.
Many of the elements of the 6th row are very dense. Mercury is not the most dense (Osmium has that claim), but it still 13 times the density of water, and 20% denser than lead. This is a true heavy metal – it rocks.
Looking at the vertical columns in the periodic table, Mercury is in the so-called ‘group 12’ elements of the periodic table (zinc, cadmium, mercury). These share low melting points. They also have low reactivity – they are reluctant to form molecules, because their main electron shell is full. If you can imagine your heavy metal band being a solo artist, that is mercury.
But this does stop it from forming molecules. Compounds with mercury include HgF2, HgCl2, HgBr2, HgO2 and HgS2. The last one is the most important one. More complex molecules can also form, such as Hg2SO4. But there are no reactions with iron, and therefore mercury is normally kept in iron containers. On the other hand, mercury reacts well with aluminium, and for this reason it is not allowed on airplanes.
Mercury is a poisonous element. But put it in a compound and it becomes highly poisonous. The worst of those compounds is methylmercury, which contains CH3Hg. This neurotoxin readily forms when mercury circulates in water, and it enters the food chain through fish. The most common cause of methylmercury poisoning is indeed from eating fish. Whale meat can be particularly high in methylmercury.
One gram of methylmercury can be fatal, and because it takes many months to leave the body, this dose can build over months. In the US, a safe dose is considered less than 1 microgram per day per person. Afflictions caused by methylmercury can include fatigue, vertigo, headaches, mental decline, progressing into shaking and eventually death. The French physicist Blaise Pascal, who used mercury in his research, may have succumbed to this. He suffered from poor health, with worsening headaches, and he complained about losing his ability to concentrate. Pascal is often said to have suffered a low grade brain tumour, but mercury poisoning seems to fit the symptoms as well. Quem Mercurius perdere vult, dementat prius. In Japan, methylmercury poisoning is known as Minamata disease, named after a particularly nasty case involving an unscrupulous company which at one point stopped investigations by installing a fake water treatment system. In the Roman Empire, people working in the mercury mines developed a shaking which became known as ‘mercurial’.
In the past, the closely related substance ethylmercury was used as a pesticide. It wasn’t the most toxic of the pesticides in use, but it came close. Eventually its use was banned.
Once ingested, mercury stays in the body for a long time. After the intake of mercury is stopped, recovery can therefore take several years.
Methylmercury builds up at the apex of food chain of the oceans. This is why whales can reach high levels (their longevity may also play a role). You do wonder whether they suffer from the same symptoms as we do: vertigo, fatigue, and headaches.
Mercury is less dangerous to us in its metal form, because it is not easily absorbed by the skin – although contact should still be avoided. But where there is liquid, there is also vapour, and the mercury vapour is far more dangerous. That is what can make liquid mercury dangerous to have around: the vapour can be inhaled.
The ecology of mercury
Mercury has a complex life cycle: it circulates between the atmosphere, the soil and the oceans. In the atmosphere, mercury lasts for about a year. It exists at very low concentrations, though. Eventually it oxidizes and in that form is dissolves into water. So rain removes it and it enters the soil and oceans. Oceans slowly lose it back to the atmosphere, mainly through waves. Both on land and in the sea, mercury binds readily to living organisms. In the sea, this brings it into the food chain. On land, a lot ends up in trees. The concentration of mercury in living organisms is called bio-amplification.
Trees take in much of the mercury and they are good at keeping hold of it. Forest fires release it back into the atmosphere. The trees that formed the coal reserves brought their mercury with, and so coal has a higher amount of mercury than typical rock.
The natural input of mercury into the atmosphere comes from oceans and trees. Human emissions exceed the natural ones. Much of this has come from coal burning. Although mercury levels in coal are not very high, we burn a lot of coal and the total adds up. There are international agreements in place that have considerably reduced this, and these emissions have declined by 90% since 1970. Not all news is bad! At the moment, coal burning is estimated to account for 25% of our mercury emission, and it is no longer the largest source. The largest contribution nowadays comes from gold mining.
However, mercury stays in the environment for a long time. Trees still have high levels from our past emissions, and they release this into the atmosphere. At the moment, mercury emissions into the atmosphere are still at about twice the natural level largely because of the memory effect.
The total natural emission is around 1000 tons per year, anthropocentric emission is around 2000 tons and legacy emission amounts to about 6000 tons. Our history clearly still matters. But our improved mercury management is paying off: mercury levels in the Atlantic Ocean are now falling.
The figure shows the full cycle, with the natural and anthropocentric emissions indicated. The red/orange striped arrows indicate ‘re-emissions’ which are caused by our much larger emissions of the past. For instance, mercury in trees remains in storage until someone lights a match. Forest fires then become major source of re-emission of the stored mercury. Emissions from the oceans, as mentioned above, are also largely legacy mercury. The re-emitted mercury remains in the atmosphere for around a year before it re-enters the sea or the soil.
Eventually, the mercury ends up under ground. This happens mainly through buried organic material. Sediment and decaying vegetation become mercury’s final resting place. Sediments can build up over time to amazingly thick layers. Over ten kilometer of sediment is not impossible, as the sea floor sinks under the weight. Mercury can therefore become deeply buried, albeit at low concentrations, within much of the continents.
In the Earth’s crust mercury mainly combines with sulfur. Mercurysulfide does not dissolve in water. It remains solid, and forms a mineral called cinnabar. Cinnabar can be stunning. In crystal form it is bright scarlet to red, and it looks like a deeply coloured quartz. This is rare, however, and normally it is a soft, brick-red rock. It is a hydrothermal mineral which is deposited on the surface of other rock.
It is the basis of vermillion, a pigment that can range in colour from orange to purple, and was commonly used in paintings. Nowadays it has been largely replaced by cadmium red, equally red but less poisonous.
Some less than fully trustworthy web shops offer cinnabar for its healing properties (the words ‘crystal healing’ are the warning sign). These claims should be taken with more than just a grain of salt – a pillar would be more like it. If you feel that ‘crystal healing’ is harmless, not in this case. Cinnabar is potentially the most toxic mineral known to us. Mercurysulfide itself is not as poisonous as other forms of mercury, but skin contact should be limited (mercury rash is a warning sign), and although it does not vaporize, inhaling the dust is dangerous. Google, Amazon and Ebay remain happy to advertise cinnabar’s ‘healing’ (but life-shortening) properties. One company even has named a perfume after it; although it contains no mercury, the name is about as appealing and healthy as breathing in oil (also known as vaping).
Liquid mercury can be obtained from cinnabar by heating in an oven, followed by condensation of the mercury vapour and collection in iron containers. However, do not try this at home.
Cinnabar is a fairly common mineral, and is mined for instance in the Red Devil mine in Alaska. About a third of the world’s production comes from Spain. It is found in deposits associated with hot springs, and in veins within rock. As a hydrothermal mineral, it is associated with circulating hot water. This establishes a firm link with volcanoes.
The typical level of mercury in ordinary rock is 0.1 ppm, a negligible amount. In contrast, cinnabar can have 86% concentration HgS. The way to achieve such concentrations involves volcanoes. And plants.
Mercury is a volatile element and can readily end up in melt inclusions. An example is granite, partially melted rock formed at depth underneath mountains. As it comes up, granite can bring the mercury with it, and indeed granite can have twice the mercury concentrations of other rocks (still only 0.2 ppm). Magma acts similarly, and as it rises in the crust it too carries mercury. Magma reservoirs become the main accessible reservoir of mercury.
Now there are two ways to the surface. One is by water circulation, and indeed cinnabar mines are often located in deposits of ancient hot springs. The second is directly from the magma itself, during eruptions. Mercury is too volatile to remain in lava, and it is released in the gas emission, both during an eruption and in steam vents (solfataras). Etna and Hawai’i are both confirmed mercury emitters. So is Yellowstone, with the proviso that USGS measurements found that the wildfires around the park emit more mercury than the geysers do.
The figure shows the mercury concentration measured in an ice core from the Upper Fremont Glacier in Wyoming. The green spikes are volcanic emissions, orange is the gold mining, and red is the industrial emission. The human emissions dominate over volcanic ones. This already shows the importance of our emissions. The gold rush was a local event which was less significant on a global scale, but that it true for the St Helens eruption as well.
The ice core above gives an idea of the resulting mercury levels: the spikes due to volcanoes correspond to about 10 ppb. St Helens caused an equally large peak as Krakatau, although it was a less significant eruption. That shows that proximity is important: more mercury is deposited near the eruption. But Tambora also shows up: mercury can travel a long way.
But there are times when nature wasn’t so well behaved. Recent work has shown that at times, volcanic mercury was rampant. And this had consequences.
We live in volcanically quiet times. At times a volcano erupts and causes havoc, but only the immediate surroundings are covered in lava. Occasionally a supereruption goes off and affects half the world, but that is just one of those things and the planet recovers soon enough. But there are times when life is more exciting (and potentially short). A major rift opens up and lava comes out in huge flows, covering large areas in kilometers-thick layers. These eruptions are not instantaneous: they keep erupting intermittently over perhaps 100,000 years or more. They are called LIPs, for Large Igneous Provinces. Typically they last for half a million years. Several of these enormous eruptions coincide with world-wide mass extinctions. The largest of them all was the Siberian Traps, 252 million years ago, which covered much of Siberia (the name gives a clue) and wiped out 99% of life. The extinction, almost a global sterilization, acted through a combination of global superwarming and oxygen depletion.
In our current, quiet times, volcanoes emit around 80 tons of mercury per year. One might expect that in LIPs, this will have increased. Indeed, so it was found. Because a LIP can substantially change the world, they often define the transition of one geological era to another. At these boundaries, mercury levels spike. This is the case for the Permian-Triassic boundary, the Trassic-Jurassic boundary, and Cretaceous-Tertiary boundary (now known as the Cretaceous-Paleocene boundary). These transitions coincide with the Siberian Traps (252 million years ago), the Central Atlantic Magmatic Province (201 million years ago), and the Deccan Traps (66 million years ago).
All three coincided with mass extinctions. For the Deccan Traps the extinction was not solely caused by the volcanoes, as the huge impact in Mexico provided the dinosaur killer blow, but it did not help. All three boundaries coincide with a significantly enhanced level of mercury in the environment.
This eruption is famous for causing the PT extinction event, the closest the Earth has come to sterilization. It caused extreme global warming where the oceans reached 40C. Photosynthesis ceased and large oxygen declines occurred in the Tethys seas. A bit later, a second pulse spread the oxygen-poor waters across the globe. The Siberian Traps were a huge eruption, but it is not completely clear why it had such a massive impact. There was a major spike in CO2 levels and this explains the heat, but there may have a secondary source of greenhouse heating, perhaps methane.
The extinction stands out in the fossil record. The teeming life of the Permian disappears almost in the blink of the eye, worldwide. The disappearance coincide with the indications for heat and a change in the carbon isotopic ratio. But how about mercury? This also showed a spectacular increase.
In the figure above, the coloured bands shows the number of genera with their extreme decline at the boundary. The points show the mercury levels in various samples across the world. They are measured as fraction of total organic carbon (TOC). Background levels both during the Permian and Triassic were around 0.20-0.5 ppb. (Note that the plot gives the number in per cent). But during the Siberian Traps eruption, this increased by a factor of 10, and in some sediments by a factor of 100. The timing coincides closely with the extinction event. Over a 40,000 year period with intermittent eruptions, the Siberian Traps may have emitted 10 tons of mercury per year on average, ten times the normal rate.
A second major extinction event happened 201 million years ago, at the transition from the Triassic to the Jurassic. This is the event which cemented the dominance of the dinosaurs. Before this, mammals were common and competitive. But they apparently were not able to cope with the extinction event, or did not recover fast enough. The proto-dinosaurs saw their chance and mammals withdrew into their hiding places. They would have to wait there for 130 million years, while dinosaurs evolved, diversified and ruled.
The extinction is attributed to CO2 emissions from the Central Atlantic Magmatic province, also known as CAMP. This LIP accompanied the first break-up of Pangea, Africa began to separate from Europe and Central America. The lava and ash flows are found in South America (esp. Brazil), Africa, but also eastern North America (Gettysburg), and in southern Europe. As a break-up it didn’t full succeed. It extended into central Brazil, but clearly that part failed to become an ocean. Sometimes, a rift doesn’t rift.
The extinction event was smaller than the PT event but it was still massive. The abundant tropical reefs disappeared, perhaps related to the acidification of the oceans. This all happened during the first major pulse of the CAMP. The CAMP eruptions continued for another million years, and the biodiversity remained very low while the CAMP was in progress. Only after the eruption had ended did the biosphere begin to recover – but it was too late for those poor mammals. Over the next tens of millions of years reefs re-evolved, but now they consisted of very different types of organisms. The play must go on – but with different actors.
The CAMP eruption polluted the world, and this pollution included mercury. Old seabed sediments in Nevada , which were deposited during this time, show elevated mercury levels for about a million years, i.e. the duration of the CAMP. The main extinction event happens early in this period, and there was a sharp peak in Hg at this time. It seems plausible to associate this with an early peak in the CAMP volcanic activity. The break-up began violently, whilst later the activity slowly petered out.
The Deccan Traps
66 million years ago, while India was moving across the Tethys ocean on its way to a collision with Asia, it met with a hot spot. The mantle plume melted through the Indian crust and erupted on the surface. Out came the Deccan Traps (literally: southern stairs), an immense lava field of 1.1 million cubic kilometers. That may be relatively small compared to the Siberian Traps (4 million km3) and the CAMP (2 million km3). However, the Deccan was five times the volume of the Columbia flood basalt. A high sulfur output did not help. This happened in the middle of an ocean. Life on India itself must have been wiped out. Did it impact the rest of the world? The Deccan Traps coincided with the demise of the dinosaurs. However, so did the exceptional impact of the Chicxulub asteroid. The arguments about which of the two was really responsible have gone back and forth. In fact, neither event seems quite sufficient in itself for the magnitude of the extinction event.
The main phase of the Deccan Traps eruption is called phase-2. There was also a phase 1 and phase 3, but these were much smaller affairs. Phase 2 lasted for three quarters of a million year, and formed a lava sheet which reached a thickness of 3.4 km. In general, the layers are thick and horizontal and remarkably uniform. The flow was not continuous: there are many separate layers, sometimes interspersed with other layers, perhaps weathering of the top surface. Look closely and you’ll find different structures. Uranium dating has indicated that phase-2 had four major pulses, lasting up to 100,000 years each and reaching eruption rates of 10 km3 per year on average. There may have been short-lived phases with much higher eruption rates.
The asteroid impacted while the Deccan Traps eruption was on-going. It hit about 250,000 years into phase-2, about 20-30,000 years after the second major Deccan pulse and 100,000 years before the third pulse. These dates are not absolutely certain but considered likely at the 90% confidence level (Schoene et al., Science 363, 862–866 (2019)).
Mercury deposits show the Deccan Traps eruption well, but different locations do not always give the same results. In France, enhanced mercury levels has been found for about 150,000 years around the extinction event. A Danish site also show this. But measurements in India find two mercury spikes, near the beginning and the end of phase-2. The picture agrees with the pulse seen in uranium dating, but the mercury data lacks the high accuracy dates: they tend to be derived from argon dating, which is less accurate than uranium.
Combining the two, the middle mercury peak seen in European sites seem to be a merger of the second and third Deccan pulse. This is also the time of the extinction event, but it appears that this event happened while the Deccan Traps entered a quiet interlude.
Could the asteroid have caused mercury pollution? This seems unlikely as those rocks have very low levels of mercury. However, if it hit in a region with enhanced mercury, the vaporized rock evacuated from the crater may have spread mercury across the world. This is still a matter of discussion. It would be useful to have mercury measurements from American deposits. However, it is also possible that the mercury of this time came not from India, but from burning and dying vegetation.
So what caused the extinction? The jury remains out: it is a race between the second pulse of the Deccan phase 2, called Poladpur, and Chicxulub. The former poured 150,000 km3 of lava across India over millennia, while the second instantaneously vaporized perhaps 100,000 km3 of rock. It was like a modern election, with a choice between two evils and no good.
The role of mercury
Mercury is associated with major volcanic events. Eruptions drive it into the atmosphere, where it remains for 6 months to two years. Once it comes out, it is readily incorporated in organic matter, and ends up in sediments together with the carbon. This makes it a suitable tracer of ancient eruptions, more so than sulphates, which also rain out but do not show up in organic matter. Sulphates are excellent in ice cores but these go back no more than a million years. Over geological times, mercury holds sway.
Volcanoes cause a large increase in mercury in the atmosphere, and this remains visible in the sediments that form at that time. Even at levels measured in ppb (parts per billion) it is an identifiable tracer. And because the levels respond so quickly to the eruptions, it gives a very good view of how the eruption developed.
In the three cases discussed here, mercury remains elevated during the eruption and this shows how long the LIP event lasted. In general, this seems to be a bit less than a million years. But the eruptions do not happen at a constant rate. There are short-lived pulses in the eruption. For the Siberian Traps, these pulses seems to have been extreme, and lasted no more than 10,000 years. The main extinction event is clearly associated with the peak. The CAMP shows several peaks, with one very strong peak which coincides with the extinction event. The Deccan Traps are a less clear cut cases. It too shows several pulses but the relation to the extinction event is not exact. The mercury levels also are not as high, and even without evidence for the asteroid impact, we might well have been looking for a secondary cause for why the associated extinction event was so severe.
But can we be sure these mercury peaks were volcanic? There is a lot of mercury stored in the biosphere, and it is also possible to increase mercury emissions by going after the biosphere. In a mass extinction, the biosphere shrinks – could this cause the mercury peaks? In a way, that is what we a doing now by burning coal: it removes organic (or post-organic) matter and this causes a major of our current high mercury levels. How can we tell whether mercury comes from organic matter or from volcanoes?
In fact, we have some information on this. Mercury comes in a range of isotopes. Most common are the isotopes with mass numbers 204, 202, 201, 200, 199. (Other isotopes are rare, radioactive, of both.) Volcanic mercury has the isotopic ratios of the mantle and crust, which differ slightly from those of the atmosphere and biosphere. Volcanoes show a small excess of 202Hg, and a more restricted range of 201Hg. The differences are not large, and there is some overlap.
The results confirm that there is a large component of volcanic mercury during a LIP. However, some samples show a more complex pattern. Several samples from the Deccan era show different isotopic ratios. And for the Siberian Traps, most samples show volcanic mercury, but samples from shallow seas of modern China have a low 199Hg fraction, and this points at wildfires releasing mercury from the biosphere. There have been suggestions that the Siberia Traps erupted through a major coal field and set it on fire, and this could have given a further mercury source. It seems that both disasters had double trouble.
The idea has been pushed further. A series of papers over the past 2-3 years have found mercury spikes at 15 different epochs, including the Palaeocene-Eocene thermal maximum (55 million years ago, related to the opening of the North Atlantic), the Aptian-Albian low-oygen ocean event, 120 million years ago and related to the Southern Kerguelen Plateau and/or the Greater Ontong Java Plateau, the Upper Jurassic, 183 million years and part of the Karoo event, and as far back as the SPICE event of the Cambrian, 500 million years (the last from Puss et al 2019). Even the end of the Snowball Earth has been found to show a mercury spike. There are some uncertainties regarding these. It is for instance not always clear whether the ‘spike’ is significantly above background levels. Mercury is often measured relative to the amount organic carbon, and if the amounts of the latter in the sediment are low, the mercury can appear to spike just because of measurement uncertainties. In some cases there is no evidence for a volcanic episode. The ‘SPICE’ event data suffers from both these problems.
The method is promising but the data is difficult. But a clear result is that all of the 5 known mass extinctions are associated with mercury enhancements. Although they are not in themselves proof of volcanic events, it seems likely that LIPs are the main cause of deadly environments.
The study of volcanic mercury has exploded in the past few years. The many studies indicate that most of geological time, mercury is at stable background levels. The identified spikes above this background level are fairly short lived (geologically speaking) and they seem to coincide with LIPs. The LIPs stand accused, and they have no strong defence.
LIPs emit mercury, and because gaseous mercury can last as long as a year in the atmosphere, this pollution spreads across the globe. Levels during LIPs are ten times or more above background levels. Typical mercury levels in limestone are between 30 and 50 microgram per kilogram of rock. During LIPs, this increases to levels as high as 500 microgram per kilogram. And because living matter builds up mercury, the levels in organic matter can become even higher.
This leaves one question unanswered. Can the level become so high that the living matter becomes non-living matter? Mercury, after all, is highly poisonous. Flood basalt eruptions affect the ecosystem pretty badly, as is evident from the fact that they coincide with mass extinctions. These extinctions are attributed to a number of effects which act worldwide, far from the eruption itself. They included sulfur pollution, global heating, and low oxygen levels in the oceans. Mercury has not been considered – but could it contribute to the toll?
The mercury amounts that are measured seem small. But this may be deceptive. The level detected in sediments formed during the Siberian Traps are similar to those found in the badly polluted San Pablo Bay in San Francisco. In people affected by Minamata disease in Japan, mercury levels in their body were 150-200 times those in people in other areas. A 50 times enhancement of mercury emissions may therefore not be insignificant. Is it possible that mercury played a role in the mortality during the extinctions? Several papers cautiously suggest that this should be investigated. Based on the numbers, mercury levels could have been dangerous, and locally it could have been deadly. But it seems for a global mass extinction, more may be needed. Mercury is unlikely to be the main cause. But a contribution can not be ruled out.
An interesting finding is that pollen and spores in sediment deposited during LIPs show a high rate of damage and deformation. That is normally attributed to UV-damage, coming from a thinned or absent ozone layer. But mercury pollution can have a similar effect.
It remains this – a suggestion. But perhaps the deadliest of the volcanoes had a helping hand from this deadliest of elements.
Albert, November 2019
of vestigial material
of ancient volcanicity