Perhaps one of the more interesting questions is where did Earth's volatiles come from? The generally accepted theory is that Earth formed by the catastrophic collisions of planetary embryos (Mars-sized bodies), which effectively turned earth into a giant ball of magma, at which time the iron settled to the core though having a greater density, and took various siderophile elements with it. At this stage, the Earth would have been reasonably anhydrous. Subsequently, Earth got bombarded with chondritic material from the asteroid belt that was dislodged by Jupiter's gravitational field (including, in some models, Jupiter migrating inwards then out again), and it is from here that Earth gets its volatiles and its siderophile elements. This bombardment is often called "the late veneer". In my opinion, there are several reasons why this did not happen, which is where these papers become relevant. What are the reasons? First, while there was obviously a bombardment, to get the volatiles through that, only carbonaceous chondrites will suffice, and if there were sufficient mass to give that to Earth, there should also be a huge mass of silicates from the more normal bodies. There is also the problem of atmospheric composition. While Mars is the closest, it is hit relatively infrequently compared with its cross-section, and hit by moderately wet bodies almost totally deficient in nitrogen. Earth is hit by a large number of bodies with everything, but the Moon is seemingly not hit by wet bodies or carbonaceous bodies. Venus, meanwhile, is hit by more bodies that are very rich in nitrogen, but relatively dry. What does the sorting?
 
The first paper (Nature 501: 208 – 210) notes that if we assume the standard model by which core segregation took place, the iron would have removed about 97% of the Earth's sulphur and transferred it to the core. If so, the Earth's mantle should exhibit fractionated ³⁴S/³²S ratio according to the relevant metal-silicate partition coefficients, together with fractionated siderophile metal abundances. However, it is usually thought that Earth's mantle is both homogeneous and chondritic for this sulphur ratio,  consistent with the acquisition of sulphur  ( and other siderophile elements) from chondrites (the late veneer). An analysis of mantle material from mid-ocean ridge basalts displayed heterogeneous ³⁴S/³²S ratios that are compatible with binary mixing between a low ³⁴S/³²S  ambient mantle ratio and a high ³⁴S/³²S recycled component. The depleted end-member cannot reach a chondritic value, even if the most optimistic surface sulphur is added. Accordingly, these results imply that the mantle sulphur is at least partially determined by original accretion, and not all sulphur was deposited by the late veneer.
 
In the second (Geochim. Cosmochim. Acta 121: 67-83), samples from Earth, Moon, Mars, eucrites, carbonaceous chondrites and ordinary chondrites show variation in Si isotopes. Earth and Moon show the heaviest isotopes, and have the same composition, while enstatite chondrites have the lightest. A model of Si partitioning based on continuous planetary formation that takes into account T, P and oxygen fugacity variation during Earth's accretion. If the isotopic difference  results solely from Si fractionation during core formation, their model requires at least ~12% by weight Si in the core, which exceeds estimates based  on core density or geochemical mass balance calculations. This suggests one of two explanations: Earth's material started with heavier silicon, or (2) there is a further unknown process that leads to fractionation. They suggest vaporization following the Moon forming event, but would not this lead to lighter or different Moon material?
 
One paper (Earth Planet. Sci. Lett. 2013: 88-97) pleased me. My interpretation of the data related to atmospheric formation is that the gaseous elements originally accreted as solids, and were liberated by water as the planet evolved.  These authors showed that early early degassing of H₂ obtained from reactions of water explains the "high oxygen fugacity" of the Earth's mantle. A loss of only 1/3 of an "ocean" of water from Earth would shift the oxidation state of the upper mantle from the very low oxidation state equivalent to the Moon, and if so, no further processes are required. Hydrogen is an important component of basalts at high pressure and, perforce, low oxygen fugacity. Of particular interest, this process may have been rapid. On the early Earth, over 5 times the amount of heat had to be lost as is lost now, and one proposal (501:501 - 504 ) heat pipe volcanism such as found on Io would manage this, in which case, the evolution of water and volatiles may have also been very rapid.
 
Finally, in (Icarus 226: 1489 -1498), near-infrared spectra show the presence of hydrated poorly crystalline silica with a high silica content on the western rim of Hellas. The surfaces are sporadically exposed over a 650 km section within a limited elevation range. The high abundances and lack of associated aqueous phase material indicate high water to rock ratios were present, but the higher temperatures that would lead to quartz were not present. This latter point is of interest because it is often considered that the water flows on Mars in craters were due to internal heating due to impact, such heat being retained for considerable periods of time. To weather basalt to make silica, there would have to be continuous water of a long time, and if the water was hot and on the surface it would rapidly evaporate, while if it was buried, it would stay super-heated, and presumably some quartz would result. This suggests extensive flows of cold water.

In a previous post, I questioned whether gold showed relativistic effects in its valence electrons. I also mentioned a paper of mine that proposes that the wave functions of the heavier elements do not correspond exactly to the excited states of hydrogen, but rather are composite functions, some of which have reduced numbers of nodes, and I said that I would provide a figure from the paper once I sorted out the permission issue. That is now sorted, and the following figure comes from my paper.


 
 
The full paper can be found at http://www.publish.csiro.au/nid/78/paper/PH870329.htm  and I thank CSIRO for the permission to republish the figure. The lines show the theoretical function, the numbers in brackets are explained in the paper and the squares show the "screening constant" required to get the observed energies. The horizontal axis shows the number of radial nodes, the vertical axis, the "screening constant".
 
The contents of that paper are incompatible with what we use in quantum chemistry because the wave functions do not correspond to the excited states of hydrogen. The theoretical function is obtained by assuming a composite wave in which the quantal system is subdivisible provided discrete quanta of action are associated with any component. The periodic time may involve four "revolutions" to generate the quantum (which is why you see quantum numbers with the quarter quantum). What you may note is that for ℓ = 1, gold is not particularly impressive (and there was a shortage of clear data) but for ℓ = 0 and ℓ = 2 the agreement is not too bad at all, and not particularly worse than that for copper.
 
So, what does this mean? At the time, the relationships were simply put there as propositions, and I did not try to explain their origin. There were two reasons for this. The first was that I thought it better to simply provide the observations and not clutter it up with theory that many would find unacceptable. It is not desirable to make too many uncomfortable points in one paper. I did not even mention "composite waves" clearly. Why not? Because I felt that was against the state vector formalism, and I did not wish to have arguments on that. (That view may not be correct, because you can have "Schrödinger cat states", e.g. as described by Haroche, 2013, Angew. Chem. Int. Ed. 52: 10159 -10178). However, the second reason was perhaps more important. I was developing my own interpretation of quantum mechanics, and I was not there yet.
 
Anyway, I have got about as far as I think is necessary to start thinking about trying to convince others, and yes, it is an alternative. For the motion of a single particle I agree the Schrödinger equation applies (but for ensembles, while a wave equation applies, it is a variation as seen in the graph above.) I also agree the wave function is of the form
ψ = A exp(2πiS/h)
So, what is the difference? Well, everyone believes the wave function is complex, and here I beg to differ. It is, but not entirely. If you recall Euler's theory of complex numbers, you will recall that exp(iπ) = -1, i.e. it is real. That means that twice a period, for the very brief instant that S = h, ψ is real and equals the wave amplitude. No need to multiply by complex conjugates then (which by itself is an interesting concept –where did this conjugate come from? Simple squaring does not eliminate the complex nature!) I then assume the wave only affects the particle when the wave is real, when it forces the particle to behave as the wave requires. To this extent, the interpretation is a little like the pilot wave.
 
If you accept that, and if you accept the interpretation of what the wave function means, then the reason why an electron does not radiate energy and fall into the nucleus becomes apparent, and the Uncertainty Principle and the Exclusion Principle then follow with no further assumptions. I am currently completing a draft of this that I shall self-publish. Why self-publish? That will be the subject of a later blog.
 

In the latest Chemistry World, Derek Lowe stated that keeping up with the literature is impossible, and he argued for filtering and prioritizing. I agree with his first statement, but I do not think his second option, while it is necessary right now, is optimal. That leaves open the question, what can be done about it? I think this is important, because the major chemical societies around the world are the only organizations that could conceivably help, and surely this should be of prime importance to them. So, what are the problems?
 
Where to put the information is not a problem because we now seem to have almost unlimited digital storage capacity. Similarly, organizing it is not a problem provided the information is correctly input, in an appropriate format with proper tags. So far, easy! Paying for it? This is more tricky, but it should not necessarily be too costly in terms of cash.
 
The most obvious problem is manpower, but this can also be overcome if all chemists play their part. For example, consider chemical data. The chemist writes a paper, but it would take little extra effort to put the data into some pre-agreed format for entry into the appropriate data base. Some of this is already done with "Supplementary information", but that tends to be attached to papers, which means someone wishing to find the information has to subscribe to the journal. Is there any good reason why data like melting points and spectra cannot be provided free? As an aside, this sort of suggestion would be greatly helped if we could all agree on the formatting requirements, and what tags would be required.
 
This does not solve everything, because there are a lot of other problems too, such as "how to make something". One thing that has always struck me is the enormous wastage of effort in things like biofuels, where very similar work tended to be repeated every crisis. Yes, I know, intellectual property rights tend to get in the way, but surely we can get around this. As an example of this problem, I recall when I was involved in a joint venture with the old ICI empire. For one of the potential products to make, I suggested a polyamide based on a particular diamine that we could, according to me, make. ICINZ took this up, sent it off to the UK, where it was obviously viewed with something approaching indifference, but they let it out to a University for them to devise a way to make said polyamide. After a year, we got back the report, they could not make the diamine, and in any case, my suggested polymer would be useless. I suggested that they rethink that last thought, and got a rude blast back, "What did I know anyway?" So, I gave them the polymer's properties. "How did I know that?" they asked. "Simple," I replied, and showed them the data in an ICI patent, at which point I asked them whether they had simply fabricated the whole thing, or had they really made this diamine? There was one of those embarrassed silences! The institution could not even remember its own work!
 
In principle, how to make something is clearly placed in scientific papers, but again, the problem is, how to find the data, bearing in mind no institute can afford more than a fraction of the available journals. Even worse is the problem of finding something related. "How do you get from one functional group to another in this sort of molecule with these other groups that may interfere?" is a very common problem that in principle could be solved by computer searching, but we need an agreed format for the data, and an agreement that every chemist will do their part to place what they believe to be the best examples of their own synthetic work in it. Could we get that cooperation? Will the learned societies help?
 

One concern I have as a scientist, and one I have alluded to previously, lies in the question of computations. The problem is, we have now entered an age where computers permit modeling of a complexity unknown to previous generations. Accordingly, we can tackle problems that were never possible before, and that should be good. The problem for me is, the reports of the computations tell almost nothing about how they were done, and they are so opaque that one might even question whether the people making them fully understand the underlying code. The reason is, of course, that the code is never written by one person, but by rather a team. The code is then validated by using the computations for a sequence of known examples, and during this time, certain constants of integration that are required by the process are fixed. My problem with this follows a comment that I understand was attributed to Fermi: give me five constants and I will fit any data to an elephant. Since there is a constant associated with every integration, it is only too easy to get agreement with observation.
 
An example that particularly irritated me was a paper that tried "evolved" programs on molecules from which they evolved (Moran et al. 2006. J. Am Chem Soc. 128: 9342-9343). What they did was to apply a number of readily available and popular molecular orbital programs to compounds that had been the strong point of molecular orbital theory, such as benzene and other arenes. What they found was that these programs  "predicted" benzene to be non-planar with quite erroneous spectral signals. That such problems occur is, I suppose, inevitable, but what I found of concern is that nowhere that I know was the reason for the deviations identified, and how such propensity to error can be corrected, and once such corrections are made, what do they do to the subsequent computations that allegedly gave outputs that agreed well with observation. If the values of various constants are changed, presumably the previous agreement would disappear.
 
There are several reasons why I get a little grumpy over this. One example is this question of planetary formation. Computations up to about 1995 indicated that Earth would take about 100 My to accrete from planetary embryos, however, because of the problem of Moon formation, subsequent computations have reduced this to about 30 My, and assertions are made that computations reduce the formation of gas giants to a few My. My question is, what changed? There is no question that someone can make a mistake, and subsequently correct it, but surely it should be announced what the correction was. An even worse problem, from my point of view, was what followed from my PhD project, which involved, do cyclopropane electrons delocalize into adjacent unsaturation? Computations said yes, which is hardly surprising because molecular orbital theory starts by assuming it, and subsequently tries to show why bonds should be localized. If it is going to make a mistake, it will favour delocalization. The trouble was, my results, which involved varying substituents at another ring carbon and looking at Hammett relationships, said it does not.
 
Subsequent computational theory said that cyclopropane conjugates with adjacent unsaturation, BUT it does not transmit it, while giving no clues as to how it came to this conclusion, apart from the desire to be in agreement with the growing list of observations. Now, if theory says that conjugation involves a common wave function over the region, then the energy at all parts of that wave must be equal. (The electrons can redistribute themselves to accommodate this, but a stationary solution to the Schrödinger equation can have only one frequency.) Now, if A has a common energy with B, and B has a common energy with C, why does A not have a common energy with C? Nobody has ever answered that satisfactorily. What further irritates me is that the statement that persists in current textbooks employed the same computational programs that "proved" the existence of polywater. That was hardly a highlight, so why are we so convinced the other results are valid? So, what would I like to see? In computations, the underpinning physics, the assumptions made, and how the constants of integration were set should be clearly stated. I am quite happy to concede that computers will not make mistakes in addition, etc, but that does not mean that the instructions for the computer cannot be questioned.

Once again there were very few papers that came to my attention in August relating to my ebook on planetary formation. One of the few significant ones (Geochim Cosmochim Acta 120: 1-18) involved the determination of magnesium isotopes in lunar rocks, and these turned out to be identical with those of Earth and in chondrites, which lead to the conclusion that there was no significant magnesium isotopic separation throughout the accretion disk, nor during the Moon-forming event. There is a difference in magnesium isotope ratios between magnesium found in low and high titanium content basalts, but this is attributed to the actual crystallization processes of the basalts. This result is important because much is sometimes made of variation in iron isotope variations, and in variations for some other elements. The conclusion from this work is that apart from volatile elements, isotope variation is probably more due to subsequent processing than in planetary formation, and the disk was probably homogeneous.
 
Another point was that a planet has been found around the star GJ 504, at a distance of 43.5 A.U. from the star. Commentators have argued that such a planet is very difficult to accommodate within the standard theory. The problem is, if planets form by collision of planetesimals, and as these get bigger, collisions between embryos, the probability of collision, at least initially, is proportional to the square of the concentration of particles, and the concentration of particles depends to some power between 1 and 2, and usually taken as to the power 1.5, of the radial distance from the star. Now standard theory argues that it in our solar system, it was only around the Jupiter-Saturn distance that bodies could form reasonably quickly, and in the NICE theory, the most favoured computational route, Uranus and Neptune formed closer and had to migrate out through gravitational exchanges between them, Jupiter, Saturn, and the non-accreted planetesimals. For GJ 504, the number density of planetesimals would be such that collision probability would be about 60 times slower, so how did they form in time to form a planet four times the size of Jupiter, given that, in standard theory in our system, growth of Jupiter and Saturn was only just fast enough to get a giant?
 
In my opinion, the relative size compared with Jupiter is a red herring, because it also depends on when the gas disk is cleaned out by a stellar outflow. The reason is, in my model, bodies do not grow largely by collision of equally sized objects, but rather they grow by melt accretion of ices at a given temperature, and the rate of growth depends on the initial concentration of solids in the disk only, and of course, the gas inflow rate because that, together with the initial gas temperature and the position of the star within a cluster, determines the temperature, and the temperature determines the position of the planet. If GJ 504 formed under exactly the same conditions as Earth, this planet lies about midway between where we might expect Neptune and Uranus to lie, and which one it represents can only be determined by finding inner planets. In previous computations, the planet should, not form; in my theory, it is larger than would normally be expected but it is not unexpected, and there should be further planets within that orbit. Why is only one outer planet detected so far? The detection is by direct observation of a very young planet that is still glowing over red hot through gravitational energy release. The inner ones will be just as young, but the closer to the star, the harder it is to separate their light from that of the star, and, of course, some may appear very close to the star by being on certain orbital phases.