The Production of Crucible steel and the Damascus Pattern

There are two fundamental factors that will profoundly influence the final characteristics of the steel product: the crucible charge and the forging method. The materials and methods used to produce and forge the ingot will directly affect whether or not a pattern can be produced. Modern replication experiments, historical and ethnographic accounts demonstrate that there are many possible ingredients that can be used for the crucible charge to produce a crucible steel ingot.  They have also determined particular factors which are necessary to produce a pattern. 

Al-Beruni stated that farand (the Damascus pattern) was not the result of industry and design, but was an accidental product (Said, 1989, 218). Curiously, Wilkinson (1937, 193) made a similar statement a thousand years later, “…the figure of the genuine ancient and modern Damascus sword-blades is the result of nature, and not of art”.  More recent research by various scholars has determined the factors which affect the formation of the pattern and it is now known that the pattern is indeed the result of the nature of steel, although a certain amount of “art and industry” of forging is also required.

Essentially crucible steel can be produced from an infinite number of possible crucible charge ingredients containing iron and carbon. The presence of minor and trace elements in the crucible charge, via the source of iron, carbon or additional substances added to the charge, will also affect the steel ingot. These elements can affect the forging of the ingot (e.g. in rendering it “hot short” due to phosphorus) in addition to the performance and appearance of the final product. 

The percentage of the carbon content of the crucible steel is significant for the creation of different types of patterns and the performance of the blade. Hypoeutectoid (< 0.8% C) and hypereutectoid (> 0.8% C) steel can produce a pattern, but the microstructure and, therefore, the pattern will be noticeably different. Hypoeutectoid steel will produce a banded pattern (e.g. Sham pattern), however, the most characteristic Damascus steel patterns (e.g. Kara Khorasan pattern) is produced from hypereutectoid steel.

Hypoeutectoid ingots produce ferrite-pearlite banding. A factor in the production of the banding is the presence of elements, which during the solidification of the liquid ingot, remain in the interdendritic region  (Samuels, 1980, 129). Pearlite will form in the interdendritic band, possibly influenced by the presence of manganese. According to Samuels (1980, 129) the dendrite itself is composed primarily of ferrite and very slow cooling will produce bigger bands.

Studies, primarily lead by Verhoeven (e.g. 2001) have found that the formation of the pattern in hypereutectoid steels is due to the alignment of globular/spherical cementite in the interdendritic zones. The cementite aligns because of the presence of impurity elements present in the interdendritic zone. Verhoeven et al. (1998) determined that elements such as vanadium and molybdenum, even in quantities as low as 0.003%, promote the alignment of cementite. Other elements, which also promote banding, are chromium, niobium, and manganese (Verhoeven et al., 1998, 63).

The effect of the cooling rate on the forging of the ingot and the resulting pattern has not been studied in any depth. Verhoeven and Jones (1987, 170) note that cementite at the prior austenite grain boundary form during slow cooling, whereas faster cooling rates promote Widmanstätten cementite.  Richard Furrer (pers. com.) noted that during his replication experiments quickly cooled ingots were easier to forge than slowly cooled ingots. This is probably the result of the different cementite locations. It seems reasonable to assume that the cooling rate affects the appearance of the pattern. This is because the faster the ingot cools, the smaller the dendrites are, and therefore, the closer the interdendritic zones. The closer the interdendritic zones, the closer the aligned globular/spheroidal cementite are, and therefore, the finer the final surface pattern.  Therefore, a blade forged from a slowly cooled ingot would have a coarser pattern than a blade forged from a quickly cooled ingot, assuming that the blades require a similar amount of forging. In addition, Verhoeven and Jones (1987, 177) suggest that the grain boundary cementite grows coarser with each forging cycle, opposed to the Widmanstätten cementite, which becomes finer. It is the large cementite particles responsible for the thicker “thread” of the Damascus patterns. The extent of forging and consequently the extent of deformation of the dendrites would also affect the fineness and appearance of the pattern. The influence of the cooling rate was also noted by ethnographic accounts. Bronson (1986, 38) states that many ethnographic observers suggested that the Damascus pattern “is an effect of cooling the original crucible contents at an extremely slow rate”.   Therefore it seems likely that the fineness or coarseness of the final pattern would depend on the cooling rate of the liquid steel in addition to the amount of forging. A slowly cooled ingot could make a coarse pattern or, if forged for a long period, a fine pattern, but a quickly cooled ingot could never make a coarse patterned blade but only a fine patterned one.

Verhoeven and Pendray’s (1992, 210) experiments found that the as-cast ingot was “hot short” due to microsegregation of phosphorus and sulphur. Although few ancient steels contain sulphur, they often contain phosphorous. Since the ingots solidified from a liquid, they have areas particularly high in phosphorus appearing as the iron-carbon, phosphorous phase steadite rather than being evenly distributed, thus the ingots are “hot short”. Whether ancient blades were also “hot short” and if this decarburization procedure would have been needed if the crucibles cooled slowly in the furnace or is necessary for all crucible steel is uncertain, however, the crucible steel blades examined did contain areas with around 0.1% P. The findings by Verhoeven, that the crucible steel ingots were “hot short”, are important for three reasons:

1)  It supports the fact that Moxon among others noted that “hot shortness” was a feature of crucible steel.

2)  Being “hot short”, the blades required a different forging technique than used for other types of steel.

3)  The low temperature forging would produce spheroidal cementite.

The phosphorous in the ingots caused the ingots to be “hot short” and therefore they had to be forged at low temperatures. Verhoeven (2001, 65) found that during forging at the necessary low temperatures, below the austenite transition temperature, the cementite collects in the interdendritic regions, perhaps nucleating on the impurity elements, which are concentrated in the interdendritic regions. The austenite transition temperature (A_cm) is the temperature at which ferrite and cementite begin to separate during slow cooling (Samuels, 1980, 43). The austenite transition temperature depends upon the elemental composition of the steel, particularly the carbon content. The transition temperature begins in the region of 730^OC, around the eutectoid composition (0.8% C). The austenite transition temperature increases with the carbon content until the carbon content reaches around 2% (cast iron) where the temperature is over 1100^OC (see Samuels, 1980, 43).

The time and temperature of the forging are major factors in the formation of the pattern. Verhoeven and Pendray’s replication experiments heated the blades to 50^OC below the austenite transition temperature and then forged the blade while it slowly air-cooled to around 250^OC below the austenite transition temperature (Verhoeven, 2001, 64-65). They record that initially the carbides are randomly distributed but after additional heating and forging at these temperatures the cementite began to align. The more cycles they performed, the more distinct the banding became.

In order for the pattern to be readily observed on the surface of the blade, the decarburized and oxidized layer had to be ground off, the blade had to be cleaned and polished before it was etched. Wilkinson records that wood-ashes and water were used in India, or chalk and water to remove any surface grease (1837, 191).  Other materials used to clean the steel include dry lime with water and tobacco ash (Sachse, 1994, 83). To etch the blades, Wilkinson (1837,191) discusses the use of dilute nitric and sulphuric acids at Cutch. He also records that a better effect is produced when the blade is immersed in a bath of copper sulphate in water for ten to thirty minutes (Wilkinson, 1937, 190-191). Sachse (1994, 84) refers to the use of ferric sulphate and ferrous sulphate to etch the blades. The etching reacts preferentially to the iron and carbide regions and the effect depends on the type of etchant used and the amount of time it reacts with the metal. According to Verhoeven and Jones (1987, 155) the white component (a.k.a. threads, see Classification of Damascus Patterns) of hypereutectoid Damascus patterned blades is the cementite. On hypoeutectoid blades the ferrite is the white or lighter component. The darker “background” colour is often a form of pearlite which appears darker, or having a pearl–like appearance, hence the name. However, which phases appear lighter or darker also depends on the microstructure and the etchant used.