Where did the rocky planets get their atmospheres from? This question is not trivial. Planets accrete by some mechanism whereby dust particles form larger objects and sooner or later these form planets. However, when they are small, they are either in a vacuum, or earlier they are in the gas that is falling into the sun and which will make the sun. If they are in a vacuum there is no gas to accrete. If they are in the gas streaming into the star they will absorb some gas more or less in proportion to what is in the gas stream, with some preference of heavier gas per unit concentration. However, that preference will not mean much because the concentration of hydrogen is so high it will swamp out most of the rest. When the rocky planet gets big enough, it will form an atmosphere from the accretion disk gas, so these two mechanisms predict either no atmosphere (accretion after the disk gas is gone) or gas that is predominantly hydrogen and helium.
When the sun ejected its accretion disk, it continued to send out a flux of high-energy UV radiation. What is expected to happen then is this would boil the hydrogen atmosphere into space, and this hydrodynamic outflow would take most of the other gases with it. None of the rocky planets in our solar system has enough gravity to hold hot hydrogen and helium for long. So any gas accreted so far is either underground or lost to space. The rocky planets start without an atmosphere, except maybe residual heavy gas that was not blown away by the strong UV. The only gases that are likely to have been so held are krypton and xenon, and they have an excess of heavy isotopes that indicate they may be such residues.
The next possibility is the gases were trapped underground and emitted volcanically after the extreme UV from the sun had stopped. Now the hydrogen and helium could leak away to space slowly and leave everything else behind. But we know that our atmosphere is not a remnant of gas from the accretion disk held by gravity or absorption because if it were, neon is about as common as nitrogen in those gases, and they would be absorbed at about the same rate and both would be held equally by gravity. If our atmosphere was delivered that way, it should contain at least 0.6 bar of neon, which is many orders of magnitude greater than what we see. Neon is a very rare gas on Earth.
Attempts to answer this question have mixed results, and tend to divide scientists into camps, wherein they defend their positions vigorously. One school of thought has the gases were forced into a magma ocean that arises from the heat of the collisions of entities about the size of Mars. I disagree with this. Should this have happened, the time taken to get the collisions going (originally estimated as 100 million years, subsequently reduced to about 30 million years with some unspecified correction to the calculations to accommodate the planet being here when the Moon-forming collision occurred) the gas would have long gone. And if the calculations were so wrong and it did happen, we are back to the neon problem.
The usual way out of this is to argue the gases came from carbonaceous chondrites, which are supposedly bits knocked off asteroids from the outer part of the asteroid belt. Such chondrites sometimes have quite reasonable amounts of water in them, as well as solids containing carbon and nitrogen. The idea is that these hit the earth, get hot, and the water oxidises the carbonaceous material to liberate carbon dioxide and nitrogen gas. Ten years ago I published the first edition of my ebook “Planetary Formation and Biogenesis”, which contained evidence that this could not be the source of the gases. The reasons were numerous and some of them complex, but one simple reason is the three rocky planets all have different proportions of the different elements. How can this happen if they came from a common source?
Now, a paper has appeared (Péron and Mukhopadhyay, Science 377: 320 – 324) that states that the krypton gas in the Chassigny meteorite, shows Mars accreted chondritic volatiles before nebular gases. I have a logic problem with this: the nebula gases were there before Mars even started forming. There was never any time that there was a Mars and the nebular gases had yet to arrive. They then found the krypton and xenon had isotope ratios that fell on a line between cosmogenic and what they assigned as trapped Martian mantle gases. There is a certain danger in this because the rock would have been exposed to cosmic rays, which lead to spallation and isotope alteration. Interestingly, the xenon data contradicts a previous report by Ott in 1948 (Geochim Cosmochim Acta 52: 1937 – 1948), who found the xenon was solar in nature. It may be that these differences can be simply explained because these are taken from a meteorite and only very small amounts of the meteorite are allowed to be taken. The samples may not be representative. Interestingly Péron and Mukhopadhyay consider the meteorite to have come from the Martian interior, based on the observation by Ott that the sample had been heated to a high temperature and was presumably of volcanic nature. The problem I see with that is that Ott came to the same conclusion for a number of other meteorites that have quite different isotope ratios. It is usually wrong to draw major conclusions from an outlier result. Anyway, based on the argument that Ott thought this meteorite was igneous, this latest paper concludes that its rare gases came from the interior of Mars, and hence show the volatiles did not come from carbonaceous chondrites.
In my opinion, the conclusion is valid, but not for the right reasons. What annoys me is the example that a previous researcher thought the sample might have been volcanic rock is assume to have come from deep in the interior now, while the previous results that do not fit the proposition are put to one side. I think that small differences from two tiny samples show you should not draw conclusions. I know there are funding pressures on scientists to publish papers, but surely everything in their work and previous work they quote should be self-consistent or reasons be found for discrepancies.