Project Summary

In this study, slag, a byproduct of the iron smelting process, was excavated from an ancient iron smelting site in northern England. Dr. Carl Blair, the archeologist who led the excavation in England, also produced iron with his own replica furnace at the University of Minnesota. These two groups of slag were compared were compared and there was a greater range of densities for the group from Minnesota. This information will allow people to see the differences between ancient and modern iron smelting and give a broader view of the characteristics of slag in general.

Background

Around AD 900 in northern England, the Nordic people were trading vast amounts of iron. The profitable iron trade allowed the Nordic people to trade iron for food and other necessities that they could not easily produce due to the large amount of time it takes to smelt iron. The physical and chemical processes of iron production today differs greatly from what it was hundreds of years ago. The process used in early iron making is called direct process, due to the fact that after the reaction, bloom, a combination of slag and iron, can be removed directly from the furnace and forged into iron. The iron production involves a reduction reaction in which iron oxide (FeO) is brought into contact with carbon monoxide (CO) in an environment with plentiful energy (heat). When this occurs, the iron (Fe) and carbon dioxide (CO2) are the products of the reduction. There are several basic equations of direct process iron smelting. The two equations below display the necessary energy and carbon monoxide for the reduction reaction.

The following three equations show the process of iron oxide being reduced to metallic iron.

Most early furnaces operated at temperatures between 1150 - 1300 degrees Celsius because ore begins to be reduced around 800 degrees Celsius. At these temperatures, a liquid slag forms which largely drains away from the pieces of iron and weld together forming bloom. Two equations below show the aspects of slag formation during iron reduction.

This mixture is produced because direct process furnaces operate at temperatures below the melting point of iron, 1540 degrees Celsius. Iron is then acquired by forging the bloom, a mixture of slag and iron, and banging out the slag to get the pure iron. (Blair, et al. 1992; Plambeck, et al. 1998) There were three main types of furnaces being used in early iron production. The earliest furnace was a clay-lined hole called a basic bowl furnace which was simply dug into the ground (see Figure 1). Over time, the bowl furnace evolved into the domed furnace, a structure much like the bowl furnace, but the sides extend higher so it looked like an egg (see Figure 2). This type of furnace is rarely uncovered due to the fact that the sides were very thin and were not able to last hundreds of years. The last furnace created for early iron making was the shaft furnace. It is the best known and the most diverse. Looking much like the dome furnace from which it evolved, the shaft furnace had higher, straighter walls and was longer lasting. (Blair, et al. 1992)

Hypothesis

It is highly probable that the two groups will differ in physical properties. Just by looking at samples from both sites, it is apparent that the slag from the University of Minnesota is, as a majority, less dense than the northern England slag. Also, both groups have different color variations - the slag from northern England has many different colors, the slag from the northern England is more uniform in color. The chemical analysis will most probably show that the slag from both groups are different because the oxide ores made to produce the iron are not the same.

Method

Physical Properties The site in northern England was chosen by Dr. Carl Blair in 1991. He picked the site because after walking around the area, he saw the evidence of artificial structures beneath the dirt. It was fortunate that in this area, a valley in northern England, there is very little accumulation of organic soil material. With the help of remote sensing, he and other archaeologists began digging in one site and each year the project grew. Eventually, there were numerous dig sites, most of which were situated on one hill. The land is actually farm land used for raising sheep and is very fertile. The soil mainly consists of silty loam which is has a little sand in it, but is much like planting soil. It is very rocky with lots of roots.

The dig was divided into several sites, I worked mostly at the main site, where the project began. At this site there was already an uncovered furnace and there were thought to be more furnace structures in the vicinity. The digging began at nine each day, after removing the tarps from the site, each person was assigned a one meter by one meter square in which to dig. Everything that was removed from the unit was kept in a bucket to sift when it was full. The sifting process is when the materials dug from the unit are put onto a sifter and the smaller things sift through, while the larger, and possibly valuable items, remain on the screen. Usual one would find slag, charcoal, and iron ore left on top of the screen. These artifacts would be placed in bags labeled with the site, unit, level number, date, and the name of the worker. At the end of the day, the bags would be brought down the lab which was located near the site, so they could be washed and cataloged. Cataloging artifacts is quite simple, but vital to an archaeological study. There is no standard technique for recording artifacts because each dig is different. My advisor had the artifacts divided into specific groups (charcoal, iron ore, tap slag, etc.) and then after counting the amount of artifacts in each group, the individual groups were weighed. The artifact type, number of each artifact per group, and weight for each group, were all recorded into the computer and will be analyzed at the later time. To start finding the physical properties, each slag piece was weighed and recorded. Since there were no graduated cylinders, the density of the slag was found in a different manner. A bowl of water was placed on the scale and then the scale was zeroed. Next, the slag piece, being tied with string to a pencil, was lowered into the water until all of it was barely beneath the surface of the water. The number that was displayed on the scale was the density.

For the color of the slag, a Munsell color chart was used. Usually, a Munsell chart is utilized for determining soil type, but it was also helpful in finding the colors of the slag pieces. In the Color Chart, each color has three attributes: value, hue, and chroma. The value indicates the lightness of the color. The range is 0 for pure black and 10 for solid white. The blacks, whites, and grays between them have no value and are therefore called chromatic colors. The hue represents the attribute of the color by which we separate blue from red. Each hue is represented by a symbol and the Munsell Chart is divided into ten hue sections: R, YR, GY, G, BG, B, PB, P, and RP. The chroma is the degree of departure of a color from the neutral color of the same value. Colors of low chroma are usually called weak, while colors of high color are designated as strong. The scale begins at zero, but there is no real end to the scale. With each rock, the Munsell Color Chart was used in order to find the proper color for them. (GretagMacbeth et al. 1996)

Chemical Properties Three slag samples from each group were sent to Washington D.C. and analyzed for the chemical content.

Discussion Physical Results The graph below titled Figure 1 displays a comparison of the varying slag densities from the two groups. The slag from England varied much more in density than its counterpart from the University of Minnesota. Both groups had only one piece that had a density less than or equal to one g/cm3. The slag from England had one more piece with a density less than or equal to two g/cm3. Both groups had the greatest number that had a density near 3 g/cm3. The slag from the University of Minnesota still had a greater amount of slag pieces in the next column for pieces with a density near 4 g/cm3. For densities for 5 g/cm3, the University still had more than the group from England. For densities near 6 g/cm3, 7 g/cm3, and 8 g/cm3, there was no slag from Minnesota with those densities, only from England. The difference in densities show that the slag from the University of Minnesota had a much bigger variance than the group from England. In a way, the group from Minnesota was artificial, an attempt to copy what had been done hundreds of years ago. Dr. Blair, the professor who ran the iron smelting at the University of Minnesota probably used different kinds of ores and different sizes of coal in his furnace. He may have changed these variables in order to know the results of different types of iron smelting since he was unsure exactly how the Nordic people produced iron in AD 900. The slag densities for the group from England shows more consistency because of the consistent practice - these people knew what they were doing, Dr. Blair, with his replica furnace could only guess. Since the ancient iron smelters were well accustomed to their trade, their results are more constant.

Chemical Results

The chart below titled Figure 2 shows the results from the chemical analysis of the slag from both groups. Surprisingly, the data is quite similair which is strange since the slag came the sources of iron from the two groups were different. It is interesting to notice that in almost every element except Chromium, there is twice as many parts per million in the slag from England, than in the slag from the University of Minnesota. The most notable difference in the graph is that the University of Minnesota slag has three times as much Chromium than the England slag.

Conclusion

In conclusion, my hypothesis proved correct. The physical properties of both groups differed, but not as much as I had expected. The group from the University of Minnesota showed no consistent patterns which is because my advisor, Dr. Blair, used different components when he did each smelting cycle. On the other hand, although the slag from England had densities that were over a larger range, the most abundant density was 3 g/cm3 and 4 g/cm3. This group was more uniform because the ancient iron smelters smelted their iron the same way every time. The results from the chemical analysis were the most surprising. Despite the fact that the iron ore was from a different source for the two slag groups, their chemical content were quite similair. With this information, one is given a better understanding of slag, something which is overlooked largely because it is the waste product of the smelting process. Possibly, a study like this could help find the technique of the Nordic people a thousand years ago.

Acknowledgments

I would like to thank everyone I worked with this past summer at the archaeological field school. More specifically, most of my thanks is due to my advisor, Dr. Carl Blair. Another thank you must be paid to my junior chemistry teacher, Ms. Fruen, who helped me so much with science and got me to love it. I would also like to thank my research class teacher, Dr. Miller. He went through a tough year with us, but made us believe that we are all winners. I would have never come to research class if he wasn't teaching it.

The Munsell System of Color Notation. 1996. GretagMacbeth. 4 November 1999.

Blair, Carl Edward. The Scale and Organization of the Iron Industry at Celtic Oppida: As Characterized by the Oppidum at Kelheim, Bavaria. Diss. Minnesota U, 1992.

Metals: Industrial Process. 1998. Ed. Dr. James A. Plambeck. Introductory University Chemistry II. 14 September 1999.