Icelandic Rocks Suggest Meteorites Brought Gold To Earth 82
sciencehabit writes "Gold, platinum, and other precioius metals were sucked into Earth's molten iron core soon after our planet formed. So where did all of the material for our fancy jewelry come from? According to high-precision measurements of two isotopes, or atomic variants, of tungsten in 4-billion-old rocks from Greenland published online today in Nature [the abstract adds a bit more; the full version is paywalled, though], precious-metal-bearing meteorites struck Earth around this time, coating the planet in a veneer containing gold, platinum, and other elements long after their native counterparts had disappeared into the planet's core."
More like a "dusting" (Score:5, Informative)
Just a nitpick, but they use the word "veneer" several times in the abstract. It makes it sound like a thin solid sheet of precious metal. That's not what they're trying to imply. They're trying to emphasize the "thinness" of it, but not really getting the "scattered" part.
Probably "dusting" has some specific connotation to geologists. Maybe "scattering" would suit the situation.
Gold moves extensively through the crust (Score:5, Informative)
I've read the abstract but it's not clear that they're talking about enormous quantities of added gold/platinum/whatever. One interesting thing about gold, silver, copper, platinum, and some of the other precious metals is that they're soluble in hot water, so what you form is these huge underground plumes of rising hot water, over local hot magma areas, and the plumes are filled with dissolved metals. When the water rises enough it cools and the metals precipitate out -- primarily in cracks through which the water moves, forming veins that contain very high concentrations of precious metals. These plumes can be many, many miles high, and can pull up/concentrate metals from significant depths, so it's not clear to me that early gravity sorting of heavy metals downwards would result in no heavy metals at the surface. (An interesting side-note is that since each metal has a different solubility in water, as the water rises and cools, different metals precipitate out at different points, so if you find silver you're likely to find at least some gold nearby, but most likely not at exactly the same spot.) Note that I'm not a geologist, just an amateur gold hunter, but this is the explanation I've been given by my geologist friends.
Iceland isn't the star of the paper (Score:4, Informative)
Iceland != Greenland.
In fact, it wouldn't make a speck of sense if it was Iceland being studied, because Iceland is a geologically very young volcanic island with rocks no more than ~40 million years old, whereas the rocks being studied in this paper are 4 billion years or so and among the oldest on Earth. The whole point of the paper is to show that tungsten isotopes have changed over Earth history, and that the change happened quite early. They do compare the old Greenland tungsten isotopic measurements to more recent igneous rocks such as the ones from Iceland, but you could have as easily mentioned Hawaii, the Azores, the Canary Islands, and several other "recent" locations used for the comparison. Iceland isn't special in that respect.
The premise of this paper is that the difference can be explained if the early Earth (>4 billion or so) chemically differentiated initially and most of the siderophile elements (things like tungsten, gold, platinum, etc.) sank to the core during that process, leaving the surface rocks more depleted. That's the time the Greenland samples may represent. Then at a younger time, speculated to be near the ~3.8 billion year late heavy bombardment [wikipedia.org], a bunch more meteoritic stuff was dumped on the top (more siderophile-enriched), mixed into the upper part of the mantle, and igneous rocks have been generated mainly from that upper mantle source ever since (including the more modern samples they are comparing to, and also the ~2 billion-year-old samples they also show). There are other scenarios, but it is plausible and ties in with other evidence about the late heavy bombardment (such as Nd isotopic data from Sm/Nd and Hf/W dating). They model the effects of some alternative models and show those models can't easily be used to explain what is seen. It's a pretty testable hypothesis as people continue to do tungsten isotope studies on rocks of a variety of ages before and after the late heavy bombardment. This is a pretty bold paper.
Interesting Proposition, But I Don't Buy It (Score:4, Informative)
Re:Iceland isn't the star of the paper (Score:3, Informative)
"I don't understand. Why should tungsten isotopic abundances change over time on Earth as compared to in space?"
Do you *really* want me to try to explain that? :-) Well, hey, this is slashdot. People here aren't stupid and I'm up for a challenge, although I'm not an expert on this stuff either.
Okay, first of all there are several isotopes of tungsten (W). The ones relevant in this paper are 182W and 184W. These are both stable, non-radioactive isotopes, but the 182W is also produced by the decay of radioactive 182Hf -- that is, over time the amount of 182Hf in a batch of solar system stuff (like the entire Earth) is going to decline due to decay, and the product is 182W. Thus, the amount of 182W is going to correspondingly increase, causing the 182W/184W ratio to climb. Wherever the 182Hf is present you'll get changing tungsten isotopes in that batch of material -- shifting the tungsten ratios over time from whatever it was initially. Wherever the 182Hf is chemically depleted from rocks or their melts, the 182W/184W ratio remains static (because you're not adding any new 182W from the decay process). So, in a way you have a tungsten isotopic system that will indicate the presence of 182Hf even if the 182Hf has eventually decayed away entirely. If Hf was ever there, the W ratios will be perturbed.
Got that so far?
Another important point: 182Hf has a half-life of "only" 9 million years. After about 10 half lives (i.e. ~90 million years), the amount left will fall into the range of measurement uncertainties and effectively can be treated as zero, and in practice the limit is usually less than that (turns out to be ~60 million for the Hf-W system). Even 60 million years is darn short compared to the multi-billion-year age of the Earth. How much 182Hf should be left in the Earth? None. The only way to change the 182W/184W ratio after that is to mix batches of stuff that have different 182W/184W ratios together.
Ok, now we start making planetary bodies. Take a batch of material that got blown out of a supernova somewhere, swept up due to gravity, and eventually start accreting into the sun, planets, and other bits. It starts off with a certain amount of 182Hf which starts decaying, and with a particular 182W/184W ratio. Everything is happily tracking along together until you start chemically separating things, and in this case Hf and W have very different chemical behaviors in typical solar system materials and conditions. Specifically, Hf tends to go with "lithophile" elements -- that is silicate rocks and melts. By contrast, tungsten tends to go with the "siderophile" elements -- i.e. common metals such as iron and nickel. In small meteorites that didn't glom together into bodies hundreds of km in diameter, chemical differentiation never really happens because they didn't melt. They've been "frozen" since the time the solar system formed. In those the Hf decays, alters the W isotopes, and pretty much reflects the mix of Hf and W in the original, average material that formed the solar system. We can sample this by analyzing the meteorites for their W isotopes, but you have to be sure it is the undifferentiated ones (generally speaking, these are the carbonaceous chondrites [wikipedia.org]). There is a lot of diversity to meteorites.
Meanwhile, in the larger bodies, like big asteroids (think Vesta size), moons, or planets, the stuff starts melting. The metals start sinking to the cores due to their density, taking the W with them. Very little Hf goes along. At the same time, the silicates separate/float into the upper portions, taking most of the Hf. This doesn't happen instantly, but the bottom line is, you end up with silicate materials enriched in Hf compared to the undifferentiated meteorites, and a core that is depleted in Hf compared to the undifferentiated meteorites. Therefore, you would expect silicate rocks in differentiated bodies to have their W ratio
Re:Gold moves extensively through the crust (Score:5, Informative)
Your comments about the non-zero solubility of [anything] in hot water are not incorrect, but not relevant to this work.
The LHB veneer idea doesn't claim that there would be no siderophile elements at the Earth's surface after segregation of the core. It simply states that the early mantle and the early segregating core would have been in (approximate) chemical equilibrium, and consequently the concentrations of different elements in the iron-rich phase compared to the silicate-rich phase would have approximated towards the concentrations predicted from the partition coefficients of the relevant species in the relevant conditions. (This is a tautology - that is what "partition coefficient" means ; whether the conditions approached equilibrium is a more moot point, but they would probably have satisfied my Mantle Petrology tutor's criteria of being to more than 100km depth for more than 100 million years, which is a regime not terribly amenable to experiment.)
However, when one does the sums, and plugs in the experimental data that one has, one finds that there is, in most mantle samples from 3+ billion years ago and 100+km below the surface, more of various elements, including tungsten, gold, PGEs, etc., which should have been taken core-wards with the differentiation of the planet. So, either the segregation process and the rates of diffusion were less efficient than we have reason to believe, or there is something peculiar going on.
It gets more peculiar though ... some sources of samples (diamonds, to be precise) do show mantle materials that have been depleted in siderophile elements to the expected extent (compared to meteorite materials).
So in some places the process works as expected, and in others, it doesn't. Which is damned peculiar. And that peculiarity is what makes them come up with the model of late accretion of a modest amount of (undepleted chondritic) material onto the upper surface of the mantle after the accumulation of 90%+ of the Earth and the segregation of the core.
That puts it into the time scale appropriate for the (probable) Moon-forming Giant Impact, and for the considerably later "Late Heavy Bombardment", but I've not yet seen anyone explicitly linking the 3 events into two or even one event.
Memo to self : must make time to attend the next public lecture I hear of being given by Moorbath; this is a dead-interesting topic.