Roman Concrete

I hope all of you had a Merry Christmas and a Happy New Year, and 2023 is shaping up well for you. They say the end of the year is a time to look back, so why not look really back? Quite some time ago, I visited Rome, and I have always been fascinated by the Roman civilization, so why not start this year by looking that far back?

Perhaps one of the more rather remarkable buildings is the Pantheon, which has the world’s largest unreinforced concrete dome. That was built under the direction of Marcus Vipsanius Agrippa, the “get-things-done” man for Augustus. No reinforcement, and it lasted that long. Take a look at modern concrete and as often as not you will find it cracks and breaks up. Concrete is a mix of aggregate (stones and sand) that provides the bulk, and a cement that binds the aggregate together. We use Portland cement, which is made by heating limestone and clay (usually with some other material but the other material is not important) in a kiln up to about 1450 degrees Centigrade. The product actually depends to some extent on what the clay is, but the main products are bellite (Ca2SiO4) and alite (Ca3SiO5). If the clays contain aluminium, which most clays do, various calcium aluminosilicates are formed. Most standard cement is mainly calcium silicate to which a little gypsum is added at the end, which makes the end surface smoother.

Exactly what happens during setting is unknown. The first thing to note is that stone does not have a smooth surface at close to the molecular level, and further, stones are silicates, in which the polymer structure is perforce terminated at the surface. That would mean there are incomplete bonds. An element like carbon would fix this problem by forming double bonds but silicon cannot do that so these “awkward” surface molecules react with water to form hydroxides. What I think happens is the water in the mix hydrolyses the calcium silicate and forms silica with surface hydroxyls, and these eliminate with hydroxyls on the stone, with the calcium hydroxide also taking part, in effect forming microscopic junctions between it and stone. All of this is slow, particularly when polymeric solids cannot move easily. So to make a good concrete, besides getting the correct mix you have to let it cure for quite some time before it is at its best.

So what did the Romans do? They could not make cement by heating clay and lime up to that temperature easily, but there were sources where it was done for them: the silicate around volcanoes like Vesuvius. The Roman architect and engineer Vitruvius used a hot mix of quicklime (calcium oxide) that was hydrated and mixed with volcanic tephra. Interestingly, this will also introduce some magnesiosilicates, which are themselves cements, but magnesium may fit better than calcium onto basaltic material. For aggregate Vitruvius used fist-sized pieces of rock, including “squared red stone or brick or lava laid down in courses”. In short, Vitruvius was selecting aggregate that was much better than ordinary stone in the sense of having surface hydroxyl groups to react. That Roman concrete lasted so long may in part be due to a better choice of aggregate.

A second point was the use of hot mixing. One possibility is they used a mix of freshly slaked lime and quicklime and by freshly slaking the mix became very hot. This speeds up chemical reactions, and also allows compound formation that is not possible at low temperatures. By reacting so hot it reduced setting times. But even more interestingly, it appears to allow self-healing. If cracks begin to form, they are more likely to form around lime clasts, which can then react with water to make a calcium-rich solution, which can react with pozzolanic components to strengthen the composite material. To support this, Admir Masic, who had been studying Roman cement, made concretes using the Roman recipe and a modern method. He then deliberately cracked the samples and ran water through them. The Roman cement self-healed completely within two weeks, while the cracks in the modern cement never healed.


Rocky Planet Formation

In the previous posts I have argued that the evidence strongly supports the concept that the sun eliminated its accretion disk within about 1 My after the star formed. During this 1 My, the disk would be very much cooler than while the sun was accreting, and the temperatures were probably not much different from those now at any given distance from the star in the rocky planet zone. Gas was still falling into the star, but at least ten thousand times slower. We also know (see previous posts) that small solid objects such as CAIs and iron bearing meteorites are much older than the planets and asteroids. If the heavier isotope distributions of xenon and krypton are caused by the hydrodynamic loss to space, which is the most obvious reason, then Earth had to have formed before the disk cleanout, which means Earth was more or less formed within about 1 My after the formation of the sun.

The basic problem for forming rocky planets is how does the rocky material stick together? If you are on the beach, you may note that sand does not turn into a solid mass. A further problem is the collisions of large objects involve huge energies. Glancing collisions lead to significant erosion of both objects, and even direct hits lead to local pulverization and intense heat, together with a shock wave going through the bodies. When the shock wave returns, the pulverized material is sent into space. Basically craters are formed, and a crater is a hole. Adding holes does not build up mass. Finally, if the two are large enough and about equal sized, they each tend to shatter as a consequence of the shock waves. This is why I believe the Monarchic growth makes more sense, where what collides with the major body is much smaller. Once the forming object is big enough, it accretes all small objects it collides with, due to gravity, but the problem is, how do small bodies stick together?

The mechanism I developed goes like this. While the star is accreting, we get very high temperatures and anything over 1000 degrees will lead to silicates softening and becoming sticky. This generates pebbles, stones and boulders that get increasingly big as we get closer to the star, because more of the silicates get more like liquids. At 1550 degrees C, iron melts, and the iron liquids coalesce. That is where the iron meteorites come from. By about 1750 – 1800 degrees silicates get quite soft, and it may be that Mercury formed by a whole lot of “liquids” forming a sticky mass. Behind that would be a distribution of ever decreasingly sized silicate masses, with iron cores where temperatures got over 1550. This would be the origin of the cores for Earth, Venus and Mercury. Mars has no significant iron core because the iron there was still in the very small particulate size.

The standard theory says the cores separated out with heavier liquids sinking, but what most people do not realize is that the core of the Earth does not comprise liquid silicates, at least not the mobile sort. You have no doubt heard that heat rises by convection at hot spots, but it is not a sort of kettle down there. The rate of movement has been estimated at 1 mm per year, which would mean the silicates would rise 1000 km every billion years. We are still well short of one complete turnover. Further an experiment where two different silicates were heated to 2000 degrees C under pressure of 26 Gpa showed that the silicates would only diffuse contents a few meters over the life of the Earth. They may be “liquid” but the perovskite silicates are so viscous nothing moves far in them. So how did the core form so quickly? In my opinion, the reason is the iron has already separated from the silicates, and the collision of a whole lot of small spherical objects do not pack well; there will be channels, and molten iron that already exists in larger masses will flow down them. Less-viscous aluminosilicates will flow up and form the continents.

The next part unfortunately involves some physical chemistry, and there is no way around it. I am going to argue that the silicates that formed the boulders separated into phases. An example is oil and water. Molecules tend to have an energy of association, that is all the water molecules have an energy that tends to hold them all together as a liquid as opposed to a gas, and that tends to keep phases separate because one such energy between like molecules is invariably stronger than the energy between different ones. There is also something called entropy, which favours things being mixed. Now the heat of association of polymers is proportional to the number of mers, while the entropy is (to a first approximation) proportional to the number of molecules. Accordingly, the longer the polymers, the less likely they are to blend, and the more likely to phase separate. That is one of the reasons that recycling plastics is such a problem: you cannot blend them because if the polymers are long, they tend to separate in processing, and your objects have “faults” running through them.

The reason this is important, from my point of view, is that at about 1300 degrees C, calcium silicate tends to phase separate from the rest, and about 1500 degrees C, a number of calcium aluminosilicates start to phase separate. These are good hydraulic cements, and my argument is that after cool down, collisions between boulders makes dust, and the cements are particularly brittle. Then if significant boulders come together gently, e.g. as in the postulated “rubble piles”, the cement dust works it way through them, and water vapour from the disk will set the cement. This works up to about 500 degrees C, but there are catches. Once it gets significantly over 300 degrees C, less water is absorbed, and the harder it is to set it. Calcium silicate only absorbs one molecule of water, but some aluminosilicates can absorb up to twenty molecules per mer. This lets us see why the rocky planets look like they do. Mars is smaller because only the calcium silicate cement can form at that distance, and because iron never melted it does not have an iron core. It has less water because calcium silicate can only set one molecule of water per cement molecule, and it does not have easily separable aluminosilicates so it has very little felsic material. Earth is near the optimum position. It is where the iron core material starts, and because it is further from the sun than the inner planets, there is more iron to sweep up. The separated aluminosilicates rise to the surface and form the felsic continents we walk on, and provided more water when setting the cement. Venus formed where it was a little hot, so it was a slow starter, but once going, it will have had bigger boulders to grow with. It has plenty of iron core, but less felsic material, and it started with less water than Earth. This is conditional on the Earth largely forming before the disk gases were ejected. If we accept that, we have a platform for why Earth has life, but of course that is for later.