Interactions between iron and carbon- crystal structure in OES analysis

The development of ferrous metallurgy is inextricably linked to the advance of civilization. It can also be truthfully said that without carbon there would be no cast iron or steel, as it has been the reducing agent used to liberate metallic iron from its ores since the earliest times. A few tenths of a percent difference in carbon content can have a dramatic effect on the mechanical properties of iron and steel, so its accurate measurement is critical to ferrous metallurgy. Chemical and spectroscopic methods have been developed for the measurement of carbon in iron and steel. One of the most popular is Optical Emission Spectrometry (OES) using an electric spark source. When measuring carbon in cast iron, however, this method can be prone to errors traceable to the granular nature of the material and to the presence of particles of “free” carbon in the form of graphite.

The “blast” furnace appeared in China in about 500 BC, charged with ore and charcoal and using phosphorus minerals as a flux. Similar processes were in use in India at about the same time. This produced “pig” or “cast” iron, that could be cast but due to its relatively high carbon content – typically 2-5% - was very hard and brittle. This technology did not reach Europe for almost 2000 years. A major step forward was the use of coke rather than charcoal as fuel for the furnace, first introduced in England in 1779. Without expensive charcoal, iron could now be produced cheaply and on an industrial scale. Carbon is clearly fundamental to iron and steel metallurgy. In early times, development was largely by trial and error, as the chemical and metallurgical mechanisms were not understood, but during the 19th century, the complicated interactions between iron and carbon were studied. Le Chatelier and others demonstrated that iron and steel have a crystalline, or “grain” structure that has a huge effect on the mechanical and other properties of the metal. This grain structure depends to a great extent (but not exclusively) on the carbon content, so the ability to control the concentration of carbon accurately and precisely is vital to the iron and steel production process.

The great advantage of spark OES is that in addition to carbon, it can also measure other elements of importance in iron and steel metallurgy, including nitrogen, silicon, sulfur and alloying elements like manganese, nickel and chromium. This would seem to render the combustion analyzer redundant, but with high carbon levels the sample taking technique can have a significant effect on OES carbon results. For good accuracy, it is especially important that the samples be taken with no formation of graphite.

Sample taking for iron and steel analysis is not straightforward: The test sample is normally only a tiny fraction of the total melt, but should be as representative of the whole as possible. Care must be taken to avoid contamination by slag. Molten metal is highly reactive, and sampling techniques should be designed so that chemical reactions that might take place after sampling, thus changing the composition away from that of the melt, are minimized. Sampling can be either single-stage, where the sampling device is also the sample mold, or two-stage, when a sample is first taken with a suitable spoon or ladle and then poured into a mold. Single stage sampling using an immersion sampler or “lance” is more suited to automation, which can help with repeatability of sampling. The cooling rate as the sample solidifies is very important: As noted above, fast cooling reduces the formation of free graphite, which can affect the OES carbon analysis. Single stage sampling tends to cause problems for cast iron applications. In double sampling, the sample is often cast as a thin disc in a heavy copper mold to cool the sample quickly. In spite of these precautions, in such a dynamic situations samples from the same melt can still show differences in crystal structure by the time they are presented for analysis.

Cast Iron – Cubes and Crystals

Under the microscope cast iron and steel are not homogeneous but granular. The structure of a given sample depends on a number of factors, but mainly on the carbon content and the thermal and mechanical processes to which it has been subjected. Iron and carbon can form a number of compounds, each with its own microstructure and hence mechanical properties. At room temperature, commercial grades of iron are composed of granular mixtures of ferrite, austenite, and iron carbide Fe3 C, with or without particles of free carbon (graphite).

The crystal structure of ferrite is an example of “Body Centered Cubic” or BCC structure; austenite is “Face Centered Cubic” or FCC. In both cases carbon atoms can enter the iron lattice as the melt cools to form a stable crystal: In ferrite they can only take up a position in the center of the cube, and in austenite positions in the center of its faces. Clearly this limits the maximum concentration of carbon in each crystal type, and we find that the maximum carbon concentration in ferrite is 0.025%; in austenite it is 2.06%. The regular structure of these materials is what makes it possible for them to be rolled or cold-worked, as slip-planes are possible between adjacent crystal faces. This is an important property of steel, and conventionally if the material contains less than the 2.06% maximum of austenite it is classified as steel, if more, then as cast iron.

At higher carbon concentrations iron carbide Fe3 C forms. This contains 6.7 weight% carbon and is also known as cementite. Free carbon can also form as the melt cools slowly and be deposited in the grain boundaries as graphite. Whether the melt cools to solidify as cementite or a mixture of cementite and graphite depends largely on the cooling rate: fast cooling promotes the formation of cementite, giving “white” cast iron without free graphite; whereas under slower cooling conditions free graphite can form to produce “grey” cast iron. The presence of certain alloying elements also has an effect: Carbon atoms cluster around single Mg or Ce atoms, forming small globes with diameters between a few micrometers and 150 μm. Thus, the amount of free carbon formed varies with relatively small variations in cooling rates and the presence of alloying elements.