Geophysicists have found that the primary magma ocean was either moderately oxidizing or moderately reducing. Scientists came to this conclusion when explaining the abnormally low proportion of niobium relative to tantalum in the studied magma samples. The work was based on experiments with a mixture of metals and silicates, which were compressed and heated using a diamond anvil to conditions similar to magma. The article is published in the Proceedings of the National Academy of Sciences.
Niobium and tantalum are located under each other in the periodic table and have similar chemical properties. These metals were brought to the young Earth by meteorites, and it would be logical to assume that since their properties are similar and they behave almost identically, then their ratio in the mantle should remain the same as in meteorites. At a minimum, this is true for other similar pairs, such as cyron-hafnium or cerium-lead.
But it turned out that in all known mantle samples , the ratio of niobium to tantalum is lessthan in meteorites, and a couple of decades ago, scientists called this the niobium paradox. There could be two explanations for this paradox: either for some reason there was a sampling error, and in other places the niobium on the contrary is larger, or something causes it to behave in a different way from tantalum and descend lower to the metal core.
In General, chemical elements can be lithophilic — have an affinity for silicates, or siderophilic-to iron. Both niobium and tantalum are lithophiles, but scientists know that, for example, vanadium at very high pressure (25 gigapascals) begins to behave like a siderophile. It was assumed that niobium behaves in a similar way, and that this is what causes it to separate from tantalum and concentrate closer to the iron core. However, this assumption imposes limitations on the chemical conditions of the primary basalt ocean: its environment must be very reductive, since at a high concentration of iron oxide, the increase in pressure ceases to increase siderophilicity. The model of an extremely reductive primary basalt ocean may be hypothetically correct, but it requires additional agreement with the rest of the knowledge about the evolution of planets.
Meanwhile, after the spread of presses with diamond anvils and laser heating, which can achieve a pressure of tens of gigapascals and a temperature of more than four thousand degrees Celsius, an unexpected effect was discovered. At very high temperatures and pressures, oxygen dissolves in the liquid metal, and radically changes the distribution of elements between silicate and metal melts.
Based on this, Dongyang Huang from the University of Paris and his colleagues decided to take a new approach to the niobium paradox. Their work consisted of two parts: experiments with these metals with them in conditions as close as possible to the young Earth, and computer modeling of the earth's interior based on this data.
An alloy of iron, Nickel, vanadium, and chromium that was most relevant to the simulated conditions was used as the metal part, and silicon dioxide mixed with manganese oxide and other elements, including niobium and tantalum oxides, was used as the mantle. In a series of experiments, this mixture was heated to 4,200 degrees Celsius and compressed to 75 gigapascals. The analysis showed that as such, heating and pressure either have little effect on the siderophilicity of niobium and tantalum, or on the contrary, reduce their affinity for iron. However, the oxygen coming from the oxides, which dissolved in the liquid metal, changed everything.
In its presence, both niobium and tantalum increase their cravings for iron, but niobium did so to a higher limit and with greater speed. Based on the data obtained, the scientists decided to update the model of the formation of the earth's interior and use it for calculations. They set different initial amounts of iron oxide in the primary magma ocean, from very small to 20 percent, fixed the current state (6 percent) as the end point, and watched for evolution. Previously, the same model required reducing conditions for the observed distribution of substances, but now it turns out the opposite: models with a low concentration of iron oxide give a lower proportion of niobium relative to tantalum than in the studied samples. Models that assume either moderately oxidizing or moderately reducing magma, with a concentration of iron oxide from 2 to 18 percent, are matched with the observed picture. This allows us to solve the niobium paradox without using unrealistically extreme assumptions.
Diamond anvils allow you to simulate the conditions in different planets, not just Earth. For example, they found out that the bowels of carbon-rich planets should be full of diamonds, and at Harvard University they obtained metallic hydrogen, similar to that which should be found inside gas giants.