Introduction.

My initial draft of this this article started with a caveat describing my sad lack of education and a plea for someone with the right letters after their name to come forward. Happily just such a person has emerged, Mike Siverns, a member of the Royal and a retired consultant metallurgist. I have to thank him for his corrections and confirmations in the preparation of this piece. It's always nice to be able to use big words in their proper context and to have an adult nearby if needed. Thank you, Mike.

A key point to remember with this subject is the relatively primitive state of metallurgical knowledge of the time. Those craftsmen only had a very basic idea of the variation in the chemical make up of different plates, they lacked accurate methods of gauging temperature and were only just starting to understand the impact of different quench speeds of the work. I understand that most philatelists will want to see hard and fast rules that can be applied in all cases, but in this instance, probably the best we can hope for are words and phrases like 'will usually', 'probably' and 'possibly'. Is this ideal? No, but I feel it is unavoidable. We are limited to looking at the practices used by those plate makers and trying to overlay what we know know to be correct now.

Steel in the 19^th Century.

Neither Perkins Bacon nor their steel suppliers would have been able to ascertain the chemical make-up let alone the exact carbon content of a piece of steel. The best that could be achieved would have been a comparative assessment of that piece of metal with other pieces. Relative hardness could have been estimated by applying files of a progressive hardness to the metal to give a rough idea. We see Perkins using such terminology as “best” or “softest” when ordering steel from their suppliers.

Temperature could not be measured directly. It could only be estimated by the colour radiating from a piece of metal, and the hotter it became, the more difficult it became to estimate. It needs to be remembered that white hot metal (about 1100'C) was bright enough to blind, in a similar way to looking directly at a welder's work. Coloured glass could have been used to diminish the possibility of damage to eyesight, but only at the expense of a more accurate estimate. Mike Siverns has pointed out that radiating colours (rather like the colour you see from a lightbulb) are different from quenching colours which derived from the oxidised surface metal (similar to a covering of paint). Thus the description of quenching described in Perkin's patent of 1819 should not be confused with tables describing radiating colours of steel.

The plates as used by Perkins Bacon

The supply of good quality steel at this period was a problem. The first order that Perkins made for plates large enough for postage stamps was from Richard Bayley of Sheffield, from whom they ordered 40 plates. It would seem that the quality was not quite up to scratch, with correspondence showing that Bayley was having problems sourcing good quality steel. Later orders were made to John H Winder whom they used (not totally exclusively) until the end of Perkin's contract to supply stamps. Swedish steel was considered to be the best available which was largely free from common contaminants such as phosphorous or sulphur, but was very expensive. Perhaps this was the source that Winder used. The plates were made by casting an ingot in a mould and subsequently hot-rolling it.

The plate was then softened by using the “Perkins Method” as described in the Perkins Patent of 1819. Michael Siverns describes this process as being possibly unique. In short, the plate was encased in iron filings and heated to a cementing heat (around 1050*C) for a period and then allowed to cool slowly in an attempt to draw out the carbon by oxidation at the plate surface. In this state, as long as the plate was still square and perfectly flat, the stamp impressions were rolled in. Once all of the engraving process was complete, the plate would have been case-hardened again as described in the Perkins Patent. This was a hit a miss affair which produced results that varied considerably, even within a single plate. Thus we see plates that were used for some years, and others that were withdrawn after only two or three months.

The metallurgy of the Perkins method.

I have presented a shortened discussion of the probable metallurgical events that took place. If you wish to go further into this aspect, please look up Michael Sivern's article in the London Philatelist (Ref) on which much of this essay is based.

Here is a table of the main chemical protagonists involved in the story, which has been reproduced by Kind permission of Michael Siverns.

The plate as it came out of the mould and after hot-rolling would have has a probable carbon content of around 0.8%, and have a structure of course pearlite. If the plate had gone through the Perkins method of plate softening, then most of the surface would have been almost pure ferrite, which would have provided a perfect surface into which to roll in stamp impressions.

The plate would then have been hardened by the Perkins method. It would have been sealed in an iron box packed with a carbon rich material. After some time it would have removed from the box and quenched by dipping into cold water. Under normal circumstances, this would have produced a case hardened plate with a high proportion of martensite. Perhaps predictably, Perkins had their own way of quenching the hot plate: rather than just leaving the plate in the water until it was cool, they practised a two stage quench. By this I mean that the plate was cooled by quenching in the normal way until the plate was “a pale yellow or straw” colour, that is still quite hot, then withdrawn from the water and allowed to cool slowly from that point. One of the points that Mr Granzow did not address in his recent book was the speed of the quench used. A very fast quench would have produced a high proportion of martensite in the plate surface as he describes. The main features of martensite is that it is very hard although brittle, and it has a larger volume than the metal from which it is formed. Brittle is not a good property to have in this context, but even worse is a volume increase. This would lead to mechanical stresses within the plate and would be possibly one of the main reasons for plate cracking or plate bending. A two stage quench would have allowed time for the martensite to settle into a more stable constituent such as bainite, which has a similar volume as other constituents that would have normally been found within a plate. To quote Michael Siverns “A touch of genius, Perkins had discovered the isothermal transformation of austenite to bainite; a century before the research metallurgists had recognised it and introduced it as one of the most useful techniques for the modern heat treatment of steels.”

So what was the structure of a typical Perkins plate? A pearlite core with a skin of hard bainite? Something similar, but with a mix of structures at the surface such as bainite, pearlite and cementite? What about a core of pearlite surrounded by relatively soft ferrite, itself surrounded by a harder skin of bainite, pearlite and cementite? I am guessing that all of these are possibles, and the results would vary from plate to plate, as well as within the same plate.

Durability

Mike Siverns ended his article with some comments about hardness of a plate and wear. A very hard metal such as martensite would have worn comparatively quickly, since although it was very hard, it did not have a great deal of ductility. Under pressure, it would therefore break rather than bend. Softer metallic crystals, that is those with a greater ductility, or indeed hard crystals encased in a softer structure would have exhibited a greater resilience to wear. Also, he points out that we do not know how the chemicals within the ink or paper would have reacted with the plate. This may well have added another dimension to the question.

AP

June 2014

As a postscript to this article, Michael's last point about chemicals in the ink reacting with the metal of the plate was very possibly bang on the money. Recent research by David Slattery, Mike Williams and others have pointed to the Prussiate of Potash that was added to the ink that turned the paper blue, may have accelerated plate wear. When that chemical ceased to be added to the ink after the fire in 1857, there was a sharp improvement in the life-span of the printing plates. It would appear to be separate from the "breakthrough in hardening" as put forward by Gary Granzow, though I am not discounting that point made.

AP

March 2016