Puddling (metallurgy)

Pig iron contains much free carbon and is brittle. Before it can be used, and before it can be worked by a blacksmith, it must be converted to a more malleable form as bar iron, the early stage of wrought iron.

Abraham Darby's successful use of coke for his blast furnace at Coalbrookdale in 1709[2] reduced the price of iron, but this coke-fuelled pig iron was not initially accepted as it could not be converted to bar iron by the existing methods.[3] Sulphur impurities from the coke made it 'hot short', or brittle when heated, and so the finery process was unworkable for it. It was not until around 1750, when steam powered blowing increased furnace temperatures enough to allow sufficient lime to be added to remove the sulfur, that coke pig iron began to be adopted.[4] Also, better processes were developed to refine it.[3]

Abraham Darby II, son of the blast furnace innovator, managed to convert pig iron to bar iron in 1749, but no details are known of his process.[5] The Cranage brothers, also working alongside the River Severn, achieved this experimentally by using a coal-fired reverbatory furnace, in which the iron and the sulphurous coal could be kept separate.[5] They were the first to hypothesise that iron could be converted from pig iron to bar iron by the action of heat alone. Although they were unaware of the necessary effects of the oxygen supplied by the air, they had at least abandoned the previous misapprehension that mixture with materials from the fuel were needed. Their experiments were successful and they were granted patent Nº851 in 1766, but no commercial adoption seems to have been made of their process.

In 1783, Peter Onions at Dowlais constructed a larger reverbatory furnace.[5] He began successful commercial puddling with this and was granted patent Nº1370.

The puddling furnace was improved by Henry Cort at Fontley in Hampshire in 1783–84 and patented in 1784. Cort added dampers to the chimney, avoiding some of the risk of overheating and 'burning' the iron.[5] Cort's process consisted of stirring molten pig iron in a reverberatory furnace in an oxidising atmosphere, thus decarburising it. When the iron "came to nature", that is, to a pasty consistency, it was gathered into a puddled ball, shingled, and rolled (as described below). This application of grooved rollers to the rolling mill, to roll narrow bars, was also Cort's invention.[6] Cort was successful in licensing this process and receiving revenue from it, even from ironworks like Cyfarthfa, where the earlier process had been invented previously by their neighbours at Dowlais. Cort is now generally regarded as the 'inventor' of puddling.

Ninety years after Cort's invention, an American labor newspaper recalled the advantages of his system:

"When iron is simply melted and run into any mold, its texture is granular, and it is so brittle as to be quite unreliable for any use requiring much tensile strength. The process of puddling consisted in stirring the molten iron run out in a puddle, and had the effect of so changing its anotomic arrangement as to render the process of rolling more efficacious."[7]

Cort's process (as patented) only worked for white cast iron, not grey cast iron, which was the usual feedstock for forges of the period. This problem was resolved probably at Merthyr Tydfil by combining puddling with one element of a slightly earlier process. This involved another kind of hearth known as a 'refinery' or 'running out fire'.[8] The pig iron was melted in this and run out into a trough. The slag separated, and floated on the molten iron, and was removed by lowering a dam at the end of the trough. The effect of this process was to desiliconise the metal, leaving a white brittle metal, known as 'finers metal'. This was the ideal material to charge to the puddling furnace. This version of the process was known as 'dry puddling' and continued in use in some places as late as 1890.

The alternative to refining gray iron was known as 'wet puddling', also known as 'boiling' or 'pig boiling'. This was invented by a puddler named Joseph Hall at Tipton. He began adding scrap iron to the charge. Later he tried adding iron scale (in effect, rust). The result was spectacular in that the furnace boiled violently. This was a chemical reaction between the oxidised iron in the scale and the carbon dissolved in the pig iron. To his surprise, the resultant puddle ball produced good iron.

One big problem with puddling was that almost 50% of the iron was drawn off with the slag because sand was used for the bed. Hall substituted roasted tap cinder for the bed, which cut this waste to 8%, declining to 5% by the end of the century.[9]

Hall subsequently became a partner in establishing the Bloomfield Iron Works at Tipton in 1830, the firm becoming Bradley, Barrows and Hall from 1834. This is the version of the process most commonly used in the mid to late 19th century. Wet puddling had the advantage that it was much more efficient than dry puddling (or any earlier process). The best yield of iron achievable from dry puddling is a ton of iron from 1.3 tons of pig iron (a yield of 77%), but the yield from wet puddling was nearly 100%.

The production of mild steel in the puddling furnace was achieved circa 1850 in Westphalia, Germany and was patented in Great Britain on behalf of Lohage, Bremme and Lehrkind. It worked only with pig iron made from certain kinds of ore. The cast iron had to be melted quickly and the slag to be rich in manganese. When the metal came to nature, it had to be removed quickly and shingled before further carburisation occurred. The process was taken up at the Low Moor Ironworks at Bradford in Yorkshire (England) in 1851 and in the Loire valley in France in 1855. It was widely used.

The puddling process began to be displaced with the introduction of the Bessemer process, which produced steel. This could be converted into wrought iron using the Aston process for a fraction of the cost and time. For comparison, an average size charge for a puddling furnace was 800–900 lb (360–410 kg)[10] while a Bessemer converter charge was 15 short tons (13,600 kg). The puddling process could not be scaled up, being limited by the amount that the puddler could handle. It could only be expanded by building more furnaces.