I suspect that graphite is precipitating on bifilms, sitting mainly at grain boundaries. Bifilms are introduced into steels during the turbulence of the pour of the liquid metal into the mould. The oxide on the surface of the liquid folds or impinges against itself, forming an oxide-to-oxide interface, defining a void or a crack between the opposed films of solid ceramic. Oxide bifilm are to be expected even when steel is poured in vacuum as a result of the 'vacuum' containing sufficient residual air.

The graphite precipitates on the outer surfaces of the bifilms because these interfaces are in atomically perfect contact with the matrix (having grown atom by atom off the liquid surface). The bifilm is a preferred substrate as a result of its associated 'air gap' or void between the two films; the strain energy resulting from the significant expansion of the graphite during precipitation is most easily accommodated by the adjacent void.

The action of high S and P to reduce graphitization is interesting. I speculate that these highly surface active elements will segregate to the surface of the liquid steel during the pour, forming oxysulphides and perhaps oxyphosphides with the oxide. These compounds will have low melting points, almost certainly lower than the casting temperature for steel in the region of 1580 to 1600 C. Thus the surface solid oxide will be converted to a liquid phase by the presence of high S and P. The folding or impingement of the liquid steel surface will now no longer form bifilms; the impingement of the liquid films will result in mutual assimilation and rapid formation of droplets which will tend to float out. No bifilms will be produced. Thus no graphitization should result.

This proposed mechanism is, of course, not mainstream metallurgical thinking. However, it is relatively easily tested. The techniques for the casting of steels and other liquid metals without turbulence is outlined in some detail in my book "Complete Casting Handbook". This is easily and quickly implemented technology at low cost.

The advantages to steels greatly outweigh the benefits of avoiding graphitization. For instance I have proposed that bifilms constitute the Griffith cracks required for failure by cracking. Thus their elimination should produce steels of unique properties, which would not be capable of failure by cracking, but only by plastic deformation to 100 per cent reduction in area. Other difficult to understand phenomena such as various corrosion modes, including stress corrosion cracking, and perhaps even hydrogen embrittlement might also be understandable for the first time.

I look forward to the day when all metals will be cast appropriately, and thus be free from defects. Fundamental considerations predict that condensed phases should be unable to nucleate non-condensed phases.  Thus  in a pure liquid at normal stresses  the nucleation of pores and voids should be impossible during the phase change we call solidification (being merely a tiny rearrangement of atoms by less than an atom radius)  Only entrainment from outside the metal, creating bifilms and bubbles, can explain the phenomenon of porosity and cracking in metals.

Some traditional foundries around the world are currently making good progress using my 10 Rules towards lowering bifilm content, with significant benefits to properties and costs. However, foundries are being built and are projected to eliminate bifilms. When these manufacturing systems start to produce we shall be rewarded with astonishing metals. It promises a revolution in metallurgy and engineering.  I look forward to it.

Hi and thank you professor Campbell. I want to ask a question. Dear prashant, please specify the grade of steel. Is your considered steel a low carbon or a medium carbon steel subjected to batch annealing process? In this case we see carbide formation and not graphite formation. Moreover, maximum annealing temperature of carbon steels is 700°C and below. The temperature is lower than the A1 temperature and we seldom observe solution of carbides. The only thing that we can observe is spheroidisation of carbides in severely deformed cold rolled coils. Your mentioned elements may change the A1 temperature and therefore, change the carbide formation behavior of the material. You cannot increase the annealing temperature more than 700°C because you will see sticking problems in the wraps of the coil. Therfore, I request you to give us moreinformation about your steel grade and the annealing cycle.

The mechanism I have suggested will, I expect, not be over-sensitive to the grade of steel, although I can imagine many steels will precipitate a carbide in place of the graphite if carbide stabilizers (such as Mn) are present in sufficient quantity.  The cure for this effect will remain the same: the steel needs to be cast without excessive turbulence. At the present time all steel castings, with the possible exception of continuous casting, are poured turbulently, so all will be expected to contain a huge population of bifilm cracks on which graphite or carbide will precipitate during subsequent heat treatments.