Examination of Crucible Steel Blades

It is difficult to distinguish between crucible steel, that has a corroded surface or does not exhibit a Damascus pattern, from other types of steel.  There is not a singular characteristic that can be used to conclusively distinguish crucible steel from other types of steels. However, the appearance of characteristics including: a lack of slag inclusions, an homogenous carbon content across the sample, a mottled appearance across the sample, and spheroidal/globular cementite in a pearlitic/ferritic matrix, all point towards the use of crucible steel because these features are not commonly observed together in other types of steel. Apart from the mottling effect, all these other features are also characteristic of Damascus steel.

According to these criteria, four blades from Kislovodsk,  # 1, # 2, # 10 and # 15 are made of crucible steel.  A fifth blade, from Jety Asar (RAS # 2) is heavily corroded but retains features suggesting it too is made of crucible steel.

The latest blade examined was a late 11^th century AD sabre from Koltso Gora. The handle, guard, two-point suspension points and the tip guard are gold plated over a copper alloy.  The gold plate is decorated using repoussè or a similar method, into intricate geometric patterns. The sabre is attributed to the Saultovo Mauaskaya culture, a culture related to the Khazar Turks, before the invasion of the Tatar-Mongols (Arzhantseva, pers. com.). KIS # 2 was one of the earliest blades examined. It is associated with a Sarmatian/early Alani burial at Klin Yar, dated to the 3^rd – 4^th century AD. The sword has a handle made of piled steel which was riveted to the blade. KIS #10 and #15 are only pieces of blades. Both blades are attributed to the Alani culture. KIS #10 is a fragment of a presumably double-edged sword from the 7^th century AD. The blade was found in a horse burial.  Horse burials were a common practice among pagans for centuries across Central Asia and in Europe and are usually associated with nomadic groups.

Examining the four crucible steel blades under low magnification after etching in 3% nital, revealed that each exhibited a mottled pattern consisting of elongated light and dark areas in a preferred orientation parallel to the blade. This feature was also noted on crucible steel blades reported by Lang et al., (1998), Allan and Gilmour (2000), and France-Lanord (1969).

The clarity, size and concentration of the mottled areas differ between samples. KIS #1 has a clear pattern over the entire sample with mottled areas around 0.5 mm long and 0.05 mm wide.  The mottled pattern of KIS # 2 is not as clear. The mottled areas are difficult to measure as they are so faint, but appear longer and much thinner than in the previous sample. In KIS #10 and KIS #15 the mottling appears to be more random than in the previous samples. The mottling tends to be more spherical and unevenly distributed throughout the sample. This mottling was also apparent after etching with Oberhoffers’s etch therefore, signifying that the effect is due to segregation of minor and trace elements that occurred during solidification. The mottled pattern may reflect the original dendritic structure that has become flattened and elongated during forging.

Under higher magnification the differences in microstructures become more visible. KIS #1 contains elongated inclusions. These were identified as manganese sulphide inclusions by EPMA spot analyses as Mn 20%, S 19%, and Fe 40%. The high iron content may be from the surrounding matrix due to their small size.  The microstructure is very fine and difficult to observe. It consists of cementite needles that are beginning to break-up and form elongated globular cementite. There is no preferred orientation of the cementite.  Some of the cementite seems to be located at prior austenite grain boundaries. The matrix is composed of irresolvable pearlite. The hardness was determined to be 345 VH. The hardness and structure is comparable to a 1.5% C experimental sample made by Furrer (Harsh, 2001).

The microstructure of KIS #2 is comparable to that seen in some blades with a Damascus pattern. The microstructure is composed of globular cementite in a DET matrix (divorced eutectoid transformed matrix of cementite particles in ferrite, see Verhoeven et al., 1998). The DET microstructure indicates that the blade was air-cooled (Verhoeven et al., 1998). The cementite has a diameter from about 1 - 5 µm. This is consistent with Verhoeven et al.’s (1998) finding that Damascus steel has an average cementite diameter of about 6µm. The alignment of the cementite, however, is not very strong indicating that the blade would not have had a crisp Damascus pattern. The microstructure is similar to Zschokke’s sword # 7 which has a Damascus pattern.

The fragments of KIS blades #10 and #15 have a similar microstructure. In both blades, elongated cementite needles and prior austenite grain boundary cementite are beginning to break up and becoming more globular appearance. The matrix is composed of very fine pearlite.

The average elemental composition is presented.  Elements analysed for but not detected in high enough quantities above the noise limit to confidently state they are present include: C, P, Al, S, Ni, As, Ag, Ti, Au, V, Ba, and Cr.

                    Average elemental Composition of Blades

The presence of manganese in three of the blades (KIS #1, #10, and #15) is significant because it may have been deliberately added to the crucible charge. The absence of manganese in KIS #2 was unexpected because in this blade the cementite was beginning to align and manganese is one of the elements that promotes cementite banding. Other impurity elements, which promote banding such as vanadium, could be causing the cementite to begin to form bands.

One of the blades from Jety Asar (RAS # 2) may be made of crucible steel. The extent of the corrosion is too advanced to detect any mottled structure under low magnification. The sample is composed primarily of corrosion with elongated areas of preserved cementite. The cementite appears to have begun to break up into globules when the blade was being forged. The microstructure of the cementite is not unlike that found in KIS #15, both have cementite needles that are beginning to break up and form spheroids.

Discussion

 The identification of four crucible steel blades is particularly important because the only other comparatively early crucible steel objects known from an archaeological context are those from Taxila.

Only six archaeological objects, the four blades from Kislovodsk, the Sasanian sword, and the blade attributed to Luristan have enough preserved steel and are documented in enough detail to compare the blades’ microstructures with the aim of determining characteristics specific to crucible steel from archaeological contexts. The examination using optical microscopy indicated that they have four common characteristics:

1)  A mottled appearance (after etching and under low magnification);

2)  An homogenous carbon content throughout the sample;

3)  Spheroidal or globular cementite well formed or just beginning;

4)   No slag, or very little slag that contains less than 4% iron oxide.

All six blades had a mottled appearance after etching in Nital and examination at low magnification (around x60). At higher magnification spheroidal/globular cementite was observed. Gilmour also noted this mottled structure and spheroidal/globular cementite on historic objects made of crucible steel (Allan and Gilmour, 2000, 491). The mottled appearance is due to slight variations in the elemental compositions of the different regions, corresponding to the original location of dendrites and interdendritic regions that formed during the initial solidification of the metal. This feature is apparent on the six archaeological examples and historic objects, in addition to being a necessary feature of hypereutectoid Damascus steel, although the mottled pattern is not observed in Damascus steel because the cementite is now aligned in the interdendritic regions.  The absence of this mottled pattern does not, however, necessarily rule out the use of crucible steel. Samuels (1980, 183) noted that in modern cast steels, the dendritic segregation could be eliminated if the steel undergoes a homogenizing annealing. It is reasonable to assume that the same mechanisms would apply to crucible steel made by traditional methods. However, Samuels (1980, 183) states that the time and temperature needed for full homogenisation is at least 45 hours at 1400^OC for phosphorus and longer for other elements such as manganese. It is unlikely that ancient craftsmen reached this high a temperature for that long a time.  In addition, although the dendritic segregation may be present, it might not be readily observed. How the object was forged and how the object was sampled in relationship to the dendrites, could affect how well the dendrites are observed (Samuels, 1980, 181). These characteristics were also noted in historical crucible steel objects and therefore, these four characteristics could be used to discriminate between crucible steel and other types of steel that were once liquid. Further research on objects and replicated crucible steel may reveal additional traits.