Archaeological Iron

I treated seven pieces of iron from a larger collection excavated at Ham Hill hillfort in Somerset, an Iron Age site. The archaeologists wanted the iron to be cleaned for interpretation rather than display, so we were able to practice on them without too much pressure. To clean them, most of us used an air abrasive machine.

Air abrasion is like sandblasting. We use aluminium oxide, which is a finer powder than sand. It is dangerous to breathe the powder, which gets everywhere, so masks must be worn. By the end, your clothes are covered with a fine layer of powder. The objects to be treated are placed boxes with a viewing window and hand holes, which is hooked up to an air extractor. The aluminium oxide is blown through a small nozzle, which is held at a shallow angle to the metal surface. The airflow is pedal-powered, and the air pressure and amount of powder can be controlled at the machine.

Top: the air abrasive machine (top left), box (bottom middle), and extractor tube (right). Bottom: horseshoe being abraded.

The iron objects had large, bulky corrosion products. The rusty orange or brown layers were removed by air abrasion. The fine aluminium oxide powder knocks off corrosion, turning it into dust and blowing it away. Its hard to see the powder working, so it looks like the corrosion is melting off, worn down by the powder.

There is a dark grey layer that follows the original surface, called magnetite. The corrosion products have to be removed, stopping at the magnetite layer but not hit the metal core underneath. This is fairly hard in practice, as its hard to see in the boxes and with the poor lighting. The stylus is held at a shallow angle so that higher areas are removed. This can produce a relatively smooth surface if done right, but my surfaces were fairly rough. Every once in awhile, you have to pull out the object to see if you are at the right level or not. There are also often voids and blisters in the iron. These areas can pop off if you are not careful, revealing the metal beneath.

One of my objects was a scythe. The scythe had dusty orange corrosion and large blisters. I slowly wore away at the corrosion with the air abrasive machine, and removed a couple of blisters by snapping them off with dental pliers. The corrosion abraded off fairly easily. It took about 4 hours to abrade the entire piece. Although it looked like a dolphin at the beginning, the shape and edges of the blade were visible after treatment. The surface still had voids and blisters, which I had avoided, but the object could be identified and its shape studied without much problem. The end of the piece had fallen off at some point, and I adhered it back on with epoxy.

Top: Scythe before treatment. Center: X-ray of the scythe, showing dark voids, a thinner edge, and an irregular surface. Bottom: after treatment.

Another object I treated was a broken horseshoe. I knew it had holes from the X-ray, and I knew that three of the four had dense metal inside. Therefore, when I abraded the piece, I was able to look out for them. The corrosion was harder than on the scythe, and so it took longer to treat the horseshoe. In the end, it took over 5 hours to treat. The surface was rougher and more uneven than the scythe. However, the shape and characteristics were evident.

Top: horseshoe before treatment. Upper middle: X-ray, showing the holes and metal in the first three holes. Lower three images: after treatment.

I cannot say that I enjoyed this treatment. I found the air abrasive machines difficult to use well. It was slow, hard to see, and dirty. The set up caused back pain because I had to bend over the boxes for hours. I was quite glad to be done. However, it was a good skill to learn. I certainly am not great at air abrading yet, but with practice, I will improve.

Copper Alloy Fragments : Treatment

After determining that the pale blue-green corrosion was just carbonates and oxides, I decided not to remove it. I could have removed the corrosion without effecting the stability of the piece. However, if I removed it, the reddish copper oxide layer would be exposed. This is visible in the pits I excavated for sampling. Although the bright spots are not aesthetically pleasing, red spots would not blend into the smooth green patinas either. The aesthetic and stability considerations were the same whether I removed it or not. Therefore, I chose to leave the corrosion as it was and just clean up the pits from which I took samples.

I removed the dirt from the surface of the fragments. The fragments, having been buried at one point, still had dirt caught in the crevices and incisions. The pieces with a rougher patina had quite a bit of dirt remaining. Dirt can soak up moisture and pollution from the air and hold it against the metal surface. This can cause reactions and negatively effect the stability of the copper alloy pieces.

The dirt was removed mechanically. Under the microscope, I gently scraped off the dirt with a scalpel. A rounded blade was used because it was smaller, easier to control, and would be less likely to scratch the surface than a pointed one. I softened the worst of the dirt with a cotton swab dampened with a solvent, IMS (industrial methylated spirits). The solvent was not used to remove the dirt because it could have just as easily pushed the dissolved dirt into the porous corrosion products. However, by dampening the dirt slightly with the solvent, it was easier to take off the dirt. I felt like I had more control, and needed to apply much less force.

A half completed fragment: cleaned on the left, with the soil remaining on the right.

There was a question regarding two of my pieces, broken fragments from the same bracelet. I did not know whether the incisions were filled with dirt or if it was deteriorated enamel or inlay. I analyzed them and another piece with SEM again to evaluate which was more likely. Enamel is mainly made of silicon and oxygen, much like dirt. In fact, it is fairly hard to accurately distinguish degraded enamel from dirt. However, most inlays would have had a colorant added, such as copper or manganese. Most colorants are either not found in soil or only found in small quantities. I compared the material in the decorative incisions on the bracelet with dirt found on another piece. I looked at them using SEM-EDS, backscatter imaging, and secondary electron imaging.

The material on the bracelet did not have any large, distinct peaks indicating a colorant. Unfortunately, I could not be sure from this test as it does contain elements that could have potentially been colorants, though in small enough quantities that they probably came from the soil. In fact, its composition was fairly similar to the dirt found on the other pieces. I decided that it was far more likely to be dirt, and removed it from the bracelet.

Compared spectra: the yellow is the spectrum of the dirt found on another piece, and the red line is the spectrum from the material in the bracelet's decoration.

The tested bracelet: the top bracelet fragment has not been cleaned, and the dirt has been removed from the incisions on the bottom bracelet fragment.

Using the SEM, I was able to use high magnification to see the surface. A spot that looked completely smooth and undamaged, even under an optical microscope, revealed to be cracked at x 700 magnification. This agrees with my theory that the smooth tin oxide patina had dehydrated and cracked, allowing for continued reaction with the copper metal beneath.

Once the fragments were clean, I treated them with benzotriazole (BTA). BTA is commonly used in the conservation of copper alloys as an inhibitor, though usually as a barrier to chlorides. The BTA complexes with the copper so that the metal cannot react with chlorides. In this case, I wanted the copper to be protected from further oxidation. I immersed the fragments in jars of 3% w/v BTA in IMS for 24 hours in the fume hood.

The BTA darkened the surface of the patina slightly. This is one of the downsides of BTA. The other downside is that the fragments are now toxic. BTA is thought to be a carcinogen. Gloves must be worn at all times when handling objects that have been treated with BTA.

I negated the toxicity slightly by coating the fragments. I applied two coats of 3% w/v Paraloid B-72 in toluene with fumed silica. Paraloid B-72 is a clear adhesive. Dissolved in toluene, and at such a low percentage, the Paraloid mixture was more like water and did not dry out too quickly when thin coats were applied to the metal surfaces. The fumed silica was added to the Paraloid B-72 so that there was a matte finish, rather than a smooth and shiny appearance. This, too, slightly changed the appearance of the fragments. However, it is fairly small and now they can be handled.

Top: after treatment. Bottom: before treatment.

In the end, the copper alloy pieces really did not need much treatment. They would have been fine without me cleaning, inhibiting, and coating them. It was a good experience for using analytical techniques, but I hope to do more treatment next time.