Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region

G Model

EXIS-23; No. of Pages 13 The Extractive Industries and Society xxx (2014) xxx–xxx

Contents lists available at ScienceDirect

The Extractive Industries and Society journal homepage: www.elsevier.com/locate/exis

Original Article

Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region Kim A.A. Hein *, Todani A. Funyufunyu School of Geosciences, University of the Witwatersrand, PBag 3, WITS, Johannesburg 2050, South Africa

A R T I C L E I N F O

A B S T R A C T

Article history: Received 28 February 2014 Received in revised form 11 April 2014 Available online xxx

In sub-Saharan Africa, artisanal and small-scale mining (ASM) has changed very little for hundreds of years and today is the most important economic activity in many of the region’s rural communities. It is rare to find a near-complete record of mining practice and skill, but in the Dem region of Burkina Faso, a history of extraction, processing and smelting of iron ore is recorded in 3 opencast mines, 2 underground mines, waste dumps, processing sites, and 11 furnaces that host fragments of furnace wall clay, tuyeres, and rare crucibles. The immediate mining footprint covers 1.9 by 1.3 km. Mine site mapping and petrographic study of the Dem site has shown that selective extraction of magnetite-hematite primary ore from fractures in quartz veins, and secondary hematite-goethite ore in iron-rich ferricrete, took place, along with smelting in furnaces to produce iron metal. Limited carbon dating of furnace charcoal supports ethnographic research that points to iron smelting being more than two centuries old, extending as far back as the Songhai Empire in the 15th century. The Dem ASM site should, therefore, be protected and preserved as a historic and cultural monument by the relevant authorities in Burkina Faso at state and provincial levels. ß 2014 Elsevier Ltd. All rights reserved.

Keywords: Artisanal and small-scale mining (ASM) Burkina Faso Iron mining Mining value chain Petrography, Carbon isotopes

1. Introduction On a ferricrete capped ridge to the west of the village of Dem in Burkina Faso (Fig. 1), a number of opencast workings and underground mines show clear evidence of extensive historic artisanal mining for iron. In difference to artisanal and small scale mining (ASM) of iron in the modern Sahelian-Saharan rural sector, which has often been characterised by one man with one small furnace, one pump and one crucible, and perhaps is more often dominated by iron recycling, the Dem ASM site of Burkina Faso covers an area in excess of 2 km2 and presents a significant scale of mining, with 11 large furnaces, and a large mining footprint. Importantly, the site is not modern as evidenced by small-grassed fossil dunes of sand that cover the forging sites, and which are presently experiencing erosion, in the process exposing details of the mining area. It is unique as a new heritage site in West Africa. Research undertaken in the Dem region has attempted to uncover the historic extraction, processing and production of iron

* Corresponding author. Tel.: +27 11 7176623. E-mail addresses: [email protected] (Kim A.A. Hein), [email protected] (T.A. Funyufunyu).

ore. Because the exploration and mining history are arguably lost, the extraction techniques are not known, the reasons for production have been forgotten and the age of mine workings is no longer known, this research recorded the size and dimensions of the mine sites, waste sites, orebody and furnaces. It recorded the techniques used as best could be interpreted from the site, and studied the metallurgy of slag and type of ore deposit, while constraining the geological setting of the iron ore deposits. Several interviews were conducted with senior elders in the village of Liligomde near Kaya (Fig. 2a–f), with the assistance of staff from the Museum of Kaya. These elders were the former blacksmiths of the Chief in the Kaya region. The elders explained that although iron at the Dem site was known, iron extraction, processing and working in the Dem region had not occurred in the last 80–90 years. Historically, there was virtually no use of metals in sub-Saharan Africa before 800–500 BC, when both iron and copper came into use in the savannah and forest regions of West Africa (Smith and Timothy, 1993; Holl, 2009). Iron reached East Africa by 200 BC, and by 200 AD had been carried to the south by ancestors of the modern Bantu peoples, alongside the expansion of farming. However, the age of iron in the Dem region is not known and establishing its age may be important to the mining history and heritage of West Africa.

http://dx.doi.org/10.1016/j.exis.2014.04.004 2214-790X/ß 2014 Elsevier Ltd. All rights reserved.

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 2

K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

Fig. 1. Location of the study area in Burkina Faso, northeast of the capital city of Ouagadougou and northwest of the regional centre of Kaya. The study area is located near the village of Dem and west of the Lac (Lake) Dem in the Goren greenstone belt.

With this in mind, the research undertaken at Dem, originally targeted for detailed geological research, included the geological mapping of the mine site to broaden understanding of the mining footprint and verify the details of the ore-bearing rocks. Engineering aspects focussed on the methods of extraction, processing and production of iron ore as best as could be reconstructed. The mineralogy and metallurgy of slag was investigated to constrain furnace conditions, and a first pass estimate of the absolute age of iron mining and the age of the furnaces was attempted using limited carbon dating. The results of those investigations are reported for the first time, with the expectation that the Dem site will focus attention towards preservation and facilitate further research of the historic mining practices of the peoples of West Africa.

2. Artisanal and small-scale mining In sub-Saharan Africa and Europe, ASM is the oldest form of mineral extraction and processing (cf., Agricola, 1556) and in the former, continues to be widespread, despite the advent of largescale mechanised mining, beginning in the late 19th century. It is a point of note that for a significant number of mineral occurrences in sub-Saharan Africa, the discovery of large-scale mineable reserves is attributed to ASM, with exploration strategies often focussed on extending or delineating resources already discovered by the artisanal miner as small-scale reserves. Commercial mining has often displaced ASM with government-sanctioned collective management of mineral resources; in many cases, this has meant that traditional and historic sites of ASM activity and resource value chains have been lost from Africa’s heritage landscape. Most ASM activity is labour-intensive, populated by persons or groups who use traditional techniques and/or low-tech equipment (Anon., 1999; Hein, 2007). Modern ASM has strong continuities with historic activity (ECA, 2012) particularly in the cases of gold, copper and iron. The principal drivers of ASM may include economic hardship, poverty, failures of cash crops under harsh climatic conditions, and greed and exploitation (Hein and Luning, 2009), but overall, it is a traditional activity (Dueppen, 2008; Holl, 2009 and others). In general, artisanal operators use simple methods and processes to extract in excess of 30 different mineral substances world-wide (ECA, 2012); there is huge variation in the amount of success and degree of windfall. Mining practices vary according to the type of the deposit being mined and the location (Aryee et al., 2003), but the majority of ASM operators rely on manual methods

of extraction using simple equipment such as shovels, pans, chisels, pick-axes and hammers. This manual aspect of ASM has remained unchanged for hundreds, if not thousands, of years in sub-Saharan Africa and around the world and is probably how iron ore was mined at Dem. On the modern ASM site, the more sophisticated operators purchase small, cheap, readily mobile excavation equipment to facilitate production, or use forms of underground ventilation, such as wind assisted plastic ventilation tubes, to bring air to depth so that extraction can proceed to 60 m or more. Torches and LED lamps strapped to the heads of miners aid with vision underground; plastic ropes and metal pulleys, or bucket hauling systems are all modernisations, but extraction is typically manual and labour-intensive. Unfortunately, modern ASM is characterised by simple and sometimes illicit marketing arrangements mainly due to poor or excessive government policing (Sinding, 2005; Hoadley and Limpitlaw, 2004; Hein and Nhlengetwa, submitted for publication). Funds for the allocation of proper market research are limited by low or no investment capital; this encourages effective illegal trading and smuggling (Mutemeri and Petersen, 2002). The process of finding adequate markets for minerals in the ASM sector is also difficult and may be disorganised (United Nations, 2002). Trade tends to follow along traditional or informal trade routes. There is a range of criteria that are generally used to categorise ASM practices, which may or may not have relevance to understanding historic practices. Hilson (2002) placed ASM activities into two categories, namely: (1) high value mineral extraction including iron ore, gold, silver, precious stones and quarry mining; and (2) the mining of industrial minerals and construction materials. In the first category, as an example, iron ore requires mining, furnacing, smelting and production where the final product can be used to manufacture tools such as handpicks and hoes. In contrast, the mining of industrial minerals and construction materials (such as sand) generally do not require any form of processing technique (although ASM of quartz for gravel in Burkina Faso requires considerable beneficiation and handcrushing). Lovitz (2006) attempted to classify ASM practices according to modern mining techniques, markets, and productivity drivers, and included categories spanning investment costs, mine output, labour, productivity, size of concessions, reserves and annual sales. However, these categories are insignificant for ASM when compared to commercial mining, and are often difficult to quantify. For example, production volume is considerably less, ore grade is highly variable, and output is irregular in ASM relative to commercial mining operations. Productivity from day to day is difficult to quantify at an ASM operation. For ore reserves, ASM is generally characterised by small and poorly defined reserves, and thus ASM produces small concessions (ECA, 2012). This is an important factor in understanding why ASM is often overlooked or negated as an important contributor to local and regional economies. For ASM in sub-Saharan Africa, Hein (2007) preferred a classification system based on livelihood/society contribution in which three categories are recognised, namely: (1) Long-term stable sites of ASM activities with focussed, ordered communities and a relatively sustained economic contribution, such as the historic gold mines of the Ashanti Kingdom of Ghana (Ayensu, 1997), iron mining and smelting in Yatenga in Burkina Faso (ECA, 2012), or copper-iron mines of southern Africa (Friede, 1980). These ASM sites are characterised by long-term production of ore, with activities lasting from years to (rarely) centuries; processing and refining takes place on the ASM site in a value chain; and a formal ASM settlement(s) is

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

3

Fig. 2. Photographs of the elders of the village of Liligomde who provided the ethnographic background to this study (a, b, e). Their information indicated that no artisanal mining for iron had taken place at Dem in their lifetimes. The senior elders ranged in age from 75 to 101 years old. They were the blacksmiths of the chief of the Kaya region. Consequently, it can be argued that the Dem site is older than 80–90 years. Artisanal mining for iron is currently conducted at a site west of the village of Liligomde as shown in (c) with smelting in the village on small crucibles. Iron rich ferricrete is used for smelting (d) and makes tools for local agricultural use. The elderly gentleman in (e) made his own gun from smelting local iron in this manner from Liligomde ore. (f) Ethnographic data was collected by interview and was recorded in a notebook with translations from Mossi to French and English as facilitated by Mr Bamogo Makido of the Museum of Kaya (squatting). Photographs are presented with permission of the village elders.

established with an internally regulated and ordered social structure. (2) Seasonal sites where ASM is balanced against agrarian livelihoods and crop cycles, and/or forms an integral part of the function in the rural village (e.g., making of iron tools for agriculture). Overall volume of ore production is low and varies from day to day; processing and refining normally takes place

at the home-village and not on-site; and an informal and temporary ASM settlement may be established but the social structure is home-village regulated and ordered. (3) Short-term sites typified by invasion and casino mining, particularly during periods of economic stress. These are characterised by mass invasion and boom-bust entrepreneurs (Hein and Nhlengetwa, submitted for publication); production

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 4

K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

cycles are irregular and limited to days or months; processing and refining is off-site; and illegal trading and smuggling structures are common. Many ASM sites are also associated with health, safety and environmental problems, particularly those lacking internal regulation. As Lovitz (2006) explains, poor occupational health and safety, and environmentally destructive mining and processing practices, attracts criticism. The occupational health, safety and environmental problems are often aggravated by inadequate regulatory frameworks governing ASM in many developing countries, especially those in sub-Saharan Africa. Nonetheless, although ASM may be associated with these negative impacts, it provides a livelihood, jobs and business opportunities to millions of people across the world (ECA, 2012; Hilson and McQuilken, 2014) and in many instances is a significant contributor to local and regional economies. This has long been the case in Burkina Faso. 3. Artisanal mining, processing and production of iron Sub-Saharan Africa is abundantly supplied with ores of differing qualities such as iron, copper and gold. Iron is found in nature as iron ore but iron oxides are the most abundant. In modern societies, iron ore is a key ingredient in structural material such as steel due to its availability, low cost and strength, but was used in ancient cultures largely for domestic and agricultural tools (Ross, 2002; Dueppen, 2008), as well as weapons and jewellery. Historically, iron was a resource more common than bronze or copper and an easier metal to work (Phillipson, 1975). Iron workings may have reached West Africa as early as the 6th century BC (Insoll, 1997; Pleiner, 2000; Ross, 2002) and a culture of iron smelting and smithing is known from archaeological studies at Kirikongo in Western Burkina Faso from the period 100–1450 AD (Dueppen, 2008), with evidence of iron smelting and smithing to at least the early 1900s in the same region (ECA, 2012). The Iron Age civilizations in the southeast part of Burkina Faso and southwest of Niger were epitomised by the Bura people, who practiced smelting and forging of tools and weapons from 300 to 1200 AD (Miller and Van Der Merwe, 1994). There was a rapid expansion of iron technology from 1400 to 1600 AD across sub-Saharan Africa (Ross, 2002). Charcoal featured as a primary fuel in sub-Saharan Africa for smelting, and successful smelting required resilient woods that would burn slowly at high temperature (Childs, 1991). Ore preparation involved picking lean ores, followed by sorting and crushing to remove the matrix and gross impurities (Kense, 1983). The ore was frequently pounded or ground to concentrate the minerals. Ore was also roasted to eliminate excess moisture and increase rock permeability (Paynter, 2011). Studies by Friede and Steel (1986) show that the ore was sometimes made into balls that could be fed into furnaces. Flux was added in the form of slag, carbonate or quartz rock, or similar materials to help remove impurities, but flux was also present in the ore/fuel ratio and may not have been specifically added (pers. comm., Thornton, 2014). Each smelting episode required several days of constant work and monitoring, and a continuous supply of resources. The chemistry of the slag is important to the smelting process as well as the chemistry of the iron ore. Smelting in sub-Saharan Africa depended exclusively on the Bloomery process (Childs, 1991). The Bloomery process removes or extracts iron (Fe) from its ore (i.e., separates Fe and O in Fe2O3 to isolate Fe). Herbert (1993), in explaining the concept of smelting, points to how smelting requires three sources, namely: (1) an available source of iron which is the iron ore, (2) a source of fuel to produce high temperatures which is charcoal from wood and/or coke from coal, and (3) carbon to reduce iron by combining with oxygen in iron ore

to form CO2. The characteristics of the slag are dependent on the carbon (Scott-Garrett, 1956). The addition of silica and/or calcium carbonate helps to remove impurities from the slag, and/or alters slag viscosity. The process of removing fluxed impurities to produce iron is known as smelting. As described by the curator at the Museum of Kaya in Burkina Faso, the smelting of iron ore in artisanal iron workings begins when the furnace is loaded with charcoal and iron ore. Air is directed into the furnace from below with a tuyere (blow pipe). During the process, iron oxides in the ore are reduced to an iron-slag mix within the furnace. Three vital chemical reactions take place inside the furnace during smelting: (1) the formation of carbon monoxide (CO) via the reaction of oxygen in the air and carbon present in the fuel, (2) the reduction of iron oxide to metallic iron, and (3) the formation of slag. The chemical reaction in which CO is formed takes place when the furnace is packed and preheated by burning grass and wood, and when air is blown into the furnace. Charcoal is added to bring the furnace to smelting temperatures. Ancient furnaces would not have reached temperatures above approximately 1200 8C, although Okafor, 2002 in Holl (2009) reported fusion temperatures of 1780 8C for slag samples from ancient furnaces in Nigeria. The heat oxidises carbon to CO2 through an endothermic reaction that causes the temperature to rise drastically. As carbon dioxide is transferred throughout the furnace, it combines with the charcoal to form CO through an endothermic reaction that reduces the temperature. The key to smelting is maintaining the iron bloom in a CO– reducing environment for a given critical temperature (pers. comm., Thornton, 2014). The second chemical reaction reduces the iron oxide to iron metal, which is the main purpose of smelting, but the ratio of CO to CO2 and O2 determines whether the process yields useable Fe or not (pers. comm., Thornton, 2014). The reduction process deals with the separation of Fe and O in Fe2O3 (hematite) and Fe3O4 (magnetite), and probably goethite (FeO(OH)), to isolate Fe for subsequent use. The iron oxides are reduced to wu¨stite (FeO), and from wu¨stite to metal iron (Tylecote et al., 1971; Doherty et al., 1985). Ulvospinel (Fe2TiO4) may also form in the slag courtesy of a reaction between magnetite and titanium in the host rocks under reducing conditions (Verhoogen, 1962) and in the approximate temperature range of 980–1100 8C (Taylor et al., 1972; Medenbach and ElGoresy, 1982). At high temperature (<1200 8C), any silica present in the furnace is likely to combine with iron oxide to form fayalite (Fe2SiO4) in slag (Killick and Gordon, 1989; Paynter, 2006). Based on the conditions inside the furnace, the reduction of iron oxide may take place in stages within the slag, or independently within the individual ore particles. At high pressure, fayalite transforms to spinels in the slag. At normal pressure, fayalite is the only stable compound within the Fe-O-SiO2 system (Schuhmann et al., 1953). The third reaction, the formation of slag, occurs between the flux(es) and impurities in the iron ore. The purpose of the flux is to assist with the removal of impurities to leave iron metal. The slag may be tapped through a temporary opening on the lower side of the furnace (Herbert, 1993), or the furnace may be dismantled and left to cool. 4. Methodology The project conducted in the Dem study area featured five aspects: mapping, sampling, database development, laboratory studies, and analysis of the results. A reconnaissance geological survey of the Dem site was undertaken in early November 2010 to establish traverse lines for a detailed geological study of iron workings; at that time, the heritage potential of the Dem site was not realised.

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

Field mapping of the study area was completed in November– December 2010. At each station point, lithological data were recorded and the GPS coordinates were established (UTM; WGS 84). Moreover, where appropriate, a photographic record was made and geological samples were collected for petrographic study (36 samples of different rock types) but are not described herein. An aerial photographic mosaic map of the Dem site was constructed as a base map, and a geology map of the same region was constructed (Fig. 3). Importantly, certain geographical features at Dem suggested that the site was old and discussions with officials at the Museum of Kaya confirmed this to be the case as there are no historic records of an iron site at Dem. It became prudent to undertake an ethnographic study before further research was undertaken and to first establish the possible immediate history of the artisanal mining of iron in the Dem region. Interviews with numerous Mossi elders (Fig. 2a–f) in the region of Kaya were arranged with the assistance of officials from the Museum of Kaya and conducted in the towns of Liligomde and Kaya, which are the two local centres of artisanal mining for iron (and gold). No interviews were held without the expressed permission of the senior elders. Data were recorded in a notebook; translations from Mossi to French and English were facilitated by Mr Bamogo Makido of the Museum of Kaya, and all notes were recorded in English. A detailed photograph record was kept. Semi-structured interviews were conducted with two groups of senior elders (ages ranged from 75 to 101 years of age) (Fig. 2a–f), all of whom are the former blacksmiths of the regional chief and therefore significant persons in social and cultural structures. The results of the interviews clearly suggest that significant extraction

5

of iron in the Dem region, and any major artisanal mining in the Dem region for any metal commodity, had not occurred in the last 80–90 years. The interviewees reported that iron extraction for primarily agricultural purposes only took place at a site immediately west of the village of Liligomde, the principal site of regional iron production in the region, and not at Dem. With the permission of staff at the Museum of Kaya, six carefully chosen samples were collected from furnaces at the Dem site for a pilot carbon study, and 11 samples of slag for petrographical analysis. The six carbon samples in the form of furnace charcoal were studied using radiometric and accelerator mass spectrometry (AMS) at Beta Analytic Laboratories in Miami, Florida, US. The main aim of the pilot study was to constrain the age of forging as a proxy for artisanal mining activity at the Dem region through the dating of furnace charcoal, which was assumed to have formed at the same time that mining and processing took place. Thin sections and polished blocks of slag were prepared at the SGS Laboratory in Booysens in South Africa. The Dem site was subsequently remapped with a focus on the geology of the iron orebody, the type and number of mines, waste sites, ore dressing and processing techniques (as reasonably could be established) and the greater mining footprint. Underground mine sampling was limited because the adits and drives were deemed unsafe and unstable to access. Geographical features such as dry river profiles, sand dunes and the geomorphology of the terrain were also mapped. Station point data used a GPS (Garmin) instrument, WGS84, UTM. Artefacts were photographed whenever found and their GPS location, recorded. The locations of burial sites were noted.

Fig. 3. Geology and layout of the Dem site, Dem village and Lac (Lake) Dem. Open cast mines 1–3 (OC1–3) and underground mines 1–2 (UG 1–2) are located on the eastern side of a ridge of ferricrete duricrust. Furnaces 1–11 (F 1–11) are situated on the alluvium plain below the open cast mines. Slag waste dumps are found nearby furnace sites. Mining waste dumps are located immediately east of the mines. A large remnant processing sites (Pr 1) is located between the mines and furnaces 8–9.

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 6

K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

5. The mine site at Dem The mining of iron ore in the Dem study area produced a limited number of underground workings and surface open pits. These mine workings exhibit no signs of modern mechanised operational methods. The surveyed ASM site covers an area of 1.9 km by 1.3 km (Fig. 3). Five ore extraction areas and eleven furnace sites were located during field studies but a broader study of the Dem region is needed. The proximal mining footprint is characterised by underground mine openings and semi-circular open casts (OC1–3). Waste rocks were dumped next to both the surface and underground mine excavations. Slag was dumped near the smelting sites (furnaces). The host rocks in the iron ore deposits included metamorphosed shale, siltstone, laterite-ferricrete and iron-rich quartz veins.

development was parallel to the mineralised vein type ore but at UG1, two 1–2 m wide adits allowed direct access to the ore. These also improved ventilation for deeper mining. After the ore was extracted from the open cast and underground mines, it was likely transported to the processing sites (within 500 m of the mine sites), with waste rock being dumped proximal to mining operation, similar to what occurs at ASM and commercial mine sites around the world today. The ore may have been man-hauled to the surface in bags or sacks particularly because there is no evidence of mechanical aids for hoisting and hauling ore. It is possible that the open cast surface operations were mined by women and/or children, similar to what is witnessed at ASM sites throughout West Africa today, with the underground operations being the domain of men. 5.3. Ore dressing and processing

5.1. Mine geology and ore deposits The Dem study area is geologically situated in the Goren volcano-sedimentary belt and hosted in rocks of the Palaeoproterozoic Birimian Supergroup which is dated at 2.3–2.1 Ga (Hein et al., 2004; Hein, 2010; Peters and Hein, 2013). The lithologies of the study area include: (1) metamorphosed basalt, and (2) shale and siltstone that are variably enriched in iron, carbonate, silica and manganese. These rocks are unconformably overlain by laterite and ferricrete duricrust that is Tertiary in age (Brown et al., 1994; Beauvais et al., 2008). The western half of the study area is characterised by fine grained siltstone and shale. The eastern half of the study area is composed of metamorphosed basalt (Fig. 3). The geology is unconformably overlain by deposits of orangeyellow aeolian sand and clay that is interbedded with rare stream conglomerate beds. The sand–clay and conglomerate beds are incised by ephemeral streams. The Dem mine site is crosscut by buck quartz veins, which in the shale and siltstone beds are associated with networks of fractures and fine quartz veins enriched in iron in the form of visible magnetite and hematite (Fig. 4a). Magnetite-rich ore in massive siltstone crops out as discrete lens that are 30–100 cm long and 10 cm wide; they were selectively mined using the gallery method. The vein-hosted iron was clearly a primary source of high-grade iron ore. Across the West African Craton, an extensive laterite-ferricrete duricrust plateau is locally mined as a source for iron ore, such as near the village of Liligomde near Kaya. Ferricrete duricrust is generally a nodular arrangement of rock fragments cemented by iron oxides that are progressively leached from the underlying insitu rock. In the Dem region, a 6 m iron-rich duricrust that unconformably overlies iron-rich shale was also mined artisanally for iron (Fig. 4b and c). Thus vein-hosted and laterite-ferricrete iron ores were mined at Dem. The grade of ore mined varied from highgrade magnetite to low-grade laterite. 5.2. Mines The surface mines at the Dem site consist of two semi-circular open cast pits (OC1, OC2) that are approximately 500–600 m in diameter (Fig. 4c), with small channels running a short distance into the sides of the pit following the iron-rich vein-type ore. OC1 and OC2 mined iron-rich ferricrete duricrust boulders that rolled down the hill from the duricrust plateau above. A third site (OC3) is situated approximately 650 m south-southeast of these two opencast mines near furnaces 10–11 (Fig. 3), but was not studied in detail and is not discussed further. The Dem site also presents two underground mines (UG1 and UG2) that feature drifts, galleries, adits and tunnels (Fig. 4a). The mines worked quartz vein-hosted iron ore in the form of magnetite and hematite hosted in shale and siltstone rock units. Mine

After extraction, the ore component was likely separated from the waste component by handpicking, sorting and crushing as a first-stage beneficiation process. After sorting and dressing, the ore was transported to the furnace sites. The means of transport is not known, but it has been speculated that the ore was transported by man-labour, or perhaps on livestock such as donkeys, as this is the technique that continues to feature today at many ASM sites in Burkina Faso. Transport from the sorting site and ultimately to the furnaces would have been across short distances, as the processing and furnace sites are not far from the mine sites. Furnaces 1 and 5 are the only processing sites that are relatively far from the mine sites, at 1.6 and 1.7 km distance, respectively (Fig. 3). Because stockpiles of high grade ore were not found in the study area, it is assumed that ore was processed and passed directly into the furnaces as needed. However, small ore stockpiles of iron rich ferricrete-duricrust are found near processing sites not far from furnaces 1–2 and 4–5, suggesting that stockpiling of secondary ore may have occurred to some degree. The evidence of iron smelting in the study area includes the remains of at least 11 furnaces (mostly floor fragments with rare remnants of the furnace walls), and hundreds of fragments of tuyeres, some crucibles, burnt clay and iron slag (Fig. 5a–d). The diameter of furnaces is generally 2 m; they are assumed to have been tall chimney in shape. The floor designs of these furnaces were shallow bowl-shaped or flat. They exhibited similar characteristics; it is not clear if they worked as entities, or together. Fragments of tuyeres made of clay and sand (often holding remnant slag), or burnt clay-sand and slag were found in and around all furnaces. The average tuyere fragment size found in the study area was 7 cm long but it is clear from reconstructed fragments that tuyeres measured >50 cm in length and approximately 10 cm in diameter. Some crucibles were also found in the furnaces and nearby but their usage is not clear as crucibles were not used in iron making (pers. comm., Thornton, 2014). The presence of tuyere fragments in all furnaces indicated that a bellows system (perhaps made of goat skin) was used to allow air into the furnaces in order to increase the furnace temperature, as described by Childs (1989). The source of the material used to make the tuyeres, crucibles and furnace wall clay can only be speculated; however, windblown aeolian sand and clay, which is found locally on the alluvial plan east of the mines (Fig. 3), may have been harvested to produce (probably by hand) tuyeres, crucibles and furnace wall clay. Thus, iron and sand mining operations may have been allied industries. The waste rock at the mine sites was dumped near the open casts and underground mines. Open casts tend to produce more waste than underground mines because a large quantity of waste is removed to gain access to the ore deposit (Hudson et al., 1999). Consequently, the largest waste piles are located next to the opencast mines.

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

7

Fig. 4. (A) Photograph of vein-hosted iron ore from adits in the underground mines of the Dem iron site. Numerous fine vein rich veinlets crosscut the host rock. (B) A nodular ferricrete duricrust crops out across the Dem region and is a rich source of secondary iron ore. Several layers of ferricrete development can be interpreted from the duricrust profile, but all are iron rich as the cement that binds rock fragments. (C) Two surface mines (OC1, OC2) consist of semi-circular opencast pits that are approximately 500–600 m in diameter. The most significant is OC1 which worked a considerable volume of primary ore dislodged from outcrops above the opencast, and in a natural valley. Iron rich quartz veins also crop out across the floor and above the opencast mines. Underground workings (UG 1) occur between the two open casts with adit/entrance located high on a rock face. OC2 worked mainly ferricrete boulders that rolled down the slope from the laterite plateau above. The processing site (Pr 1) is located not far from the mine sites.

In contrast, waste from the furnaces included slag and furnace fragments, and these were dumped within metres of the furnaces. Processing sites were also located proximal to the furnace sites, with the largest number situated between the mines and furnaces 8–9.

Different types of slag (Fig. 4a and b) were dumped at different locations in the Dem area. Due to the high silica content of the slag, it is resistant to weathering and thus large amounts of slag fragments were found at the ASM site. A significant quantity of slag waste was found around furnace 5

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 8

K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

Fig. 5. Iron smelting at the Dem site is evidenced from the remains of at least 11 furnace (furnace wall and floor fragments), and hundreds of fragments of tuyeres, rare crucibles, burnt clay and iron slag. (A) A well exposed furnace wall of clay from furnace 11. (B) Furnace 2 which is representative of all furnaces in the Dem mine site. They are typically round with flat or bowl-shaped floors. Furnace 2 is flat bottomed and is lined with fired clay. (C) Tuyere fragments are scattered around many of the furnaces or welded into the floor of the furnaces by slag. (D) Crucibles made of clay are also found near or in furnaces, and welded together with slag as is the case at furnace 1. (E) Photograph of Type 1 low viscosity glassy slag with flow texture from furnace 5. A large tuyere fragment is trapped in the slag mass. (F) Photograph of Type 2 moderate viscosity vesicular green-black blocky slag from furnace 10.

and extended up to 25–40 m away from the furnace; it may have produced more iron metal and slag, because it is bigger than all the other furnaces, or was less efficient in producing usable iron. The above evidence of ore dressing and processing clearly shows that mining in the study area was important. It can be concluded that the processing of ore must have been viable because a considerable amount of human effort and skill was expended at the site.

5.4. The greater mining footprint According to Haaland et al. (2004), the successful cycle of historic iron smelting involved: (1) selecting the smelting site with respect to settlement and resources (iron ore, charcoal, clay, water and flux), (2) gathering trees for making charcoal for the processing site, (3) clay collection for furnace construction and production of tuyeres and crucibles, and (4) distribution of the final product after smelting. Presumably, historic smelters could choose a smelting

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

site near at least one resource. Certainly, the iron-rich resources at Dem may have given reason for the other industries to exist; tuyere, crucible and furnace wall clay production is interpreted to have taken place locally from orange aeolian sand and clay that covers the Dem study area. Charcoal was the most appropriate fuel for smelting iron ore in ancient times but the type of wood used in the Dem study area is the subject of speculation. Charcoal is a by-product of dry distillation of wood and contains high carbon content (Karbowniczek, 2005). It is advantageous because it is easy to produce, has low ash content, contains no sulphur and has a high caloric value, meaning it can maintain very high temperatures in a small volume (Craddock, 1995). At Dem, an immediate source of charcoal could have been from the valleys to the east of the ASM site, which today hosts 2 large dams (Lac Dem) for irrigation and extensive market gardens (Fig. 3). The vegetation in the valley is presently dominated by large trees and savannah forest, and is sourced by wood-cutters for the local charcoal market. It seems plausible that the historic Dem forest could also have provided a ready source of fuel (charcoal) during iron smelting, and perhaps a ready source of water. A source of flux is also important in iron ore production. Flux is used as an agent to remove impurities, facilitates the chemical reaction of smelting, separates the molten metal from the waste, and/or alters the viscosity of slag. Some fluxes used in smelting include carbonate and quartz, but carbon in charcoal is also important. In the study area a ready source of clean carbonate rock does not exist (Hottin and Ouedraogo, 1992; Hein et al., 2004; Castaing et al., 2003) although many of the local siltstones are carbonaceous in composition; they may have been used as an impure source of flux during smelting. A considerable amount of crushed quartz and outcrops of quartz veins occurs both in the mine sites themselves and regionally (Hein et al., 2004; Hein, 2010). It is possible that quartz was the main flux, but further studies are needed to clarify this aspect of iron ore production at Dem. However, it is more likely that neither carbonate nor quartz was added, but that the combination of carbon from charcoal, and carbonate and silica impurities in the ore, together acted as fluxing agents. An important aspect of the greater mining footprint of the Dem ASM site is the many fossil tracks that exposed on the eroding fossil aeolian landscape. Tracks in desert aeolian sands are difficult to eradicate and can remain for decades as trace fossils, as evidenced from rehabilitation programmes around the world which want to remove tire and track damage in deserts (Burke and Cloete, 2004; Wassenaar et al., 2012). In the Dem study area, tracks are pressed into the fossil alluvial soil (that is currently being eroded) and these lead from mines to furnaces and processing sites, and from the east of the mine site region. These are interpreted as representing the routes or paths that the ASM operators used to convey material between, mines, processing sites and/or perhaps charcoal from the Dem forest to the furnaces. Significant modern tracks for human movement and/or livestock do not occur in the study area (apart from the main road from Kaya through the village of Dem and heading northwest). In addition, a number of stone artefacts of unknown use and origin litter the study area. Burial sites are also located close to furnaces 5, 6–7 and 10–11. 6. Petrography of ore-bearing rocks and slag 6.1. Ore-bearing rocks Iron and manganese rich shale and siltstone are the dominant host rocks for primary in-situ ore, while iron rich ferricrete duricrust forms a secondary ore. Primary in-situ ore is hosted in fractured quartz veins (Fig. 4a). The fractures are filled with iron oxide minerals and form an iron-rich quartz breccia. The iron

9

oxides are dominated by hematite (Fe2O3) and/or magnetite (Fe3O4). The secondary ore is composed of goethite (FeO(OH)) and hematite (Fig. 4b). Goethite is by far the most abundant iron oxide phase and is significant in quantity. 6.2. Slag geology and petrology Slag is found at several furnace sites at Dem. The slag contains primary slag phases including oxides, silicates, sulphides and the metal phase. Common primary minerals include hematite and fayalite glass (Fe2SiO4), with accessory wu¨stite (FeO), ulvospinel (Fe2TiO4), quartz, iron metal, pyrite (FeS2), and pyrrhotite (FeS1 x). The macroscopic texture and shape of the slag depends on the cooling rate and nature of the furnace where the slag solidified. The slags include Type 1, which is a dense black glassy slag with flow texture (Fig. 5e), and Type 2, which is vesicular green-black blocky slag (Fig. 5f). Type 1 slag dominates furnaces 1–5 while Type 2 dominates furnaces 8–10. Intermediate slag types also exist. Type 1 slag is a low viscosity type and similar in texture to ropey lava, or the volcanic rock obsidian. It is interpreted to have resulted from a fast to rapid cooling rate from high temperature (not greater than 1200 8C) where gas is mostly released during the cooling process, and a slag of very fine skeletal crystals or amorphous glass is produced. Type 2 slag is a moderate viscosity type and is interpreted to have resulted from moderate cooling rate from a lower temperature, where gas bubbles were trapped in the cooling slag, and a slag of large hematite crystals (>300 mm) is produced. The entrapment of gas left permanent voids within the solidified slag. The different tapped-slag types suggest that furnace temperatures were variable from site to site, and perhaps between each furnacing episode. This suggests that the quality of the usable iron may have varied considerably. Macroscopic slag samples from furnaces 1, 2, and 5 have comparatively high hematite content to that of furnaces 9, 10 and 11. Samples from furnace 5 had no vesicles; this is interpreted to be the result of processing using a slow cooling rate at that site. Visible pyrite occurs in samples from furnace 5, suggesting low oxygen fugacity, or impurities in the feed ore into the furnace. Accessory iron metal is present in samples collected from furnace 10. The common iron oxide mineral in slag was hematite, with accessory wu¨stite and magnetite (Fig. 6a–c). Interlocking crystals of spinifex and herringbone textured hematite are present in samples from all furnaces (Fig. 6b and c). The hematite hosts ocelli of quartz and fayalite (Fig. 6b). Cross-shaped wu¨stite is present in samples that host massive and semplectic intergrowths of hematite (Fig. 6b), and accessory euhedral magnetite. Fayalite ocelli dominate in slag samples FOR01-03, FOR09-03 and FOR1010. The cloth-weaved texture of accessory ulvospinel is interpreted to be the result of phase unmixing during initial melting (cf., Spry, 1987). It occurs as an extremely fine, dark isotropic exsolution of bow-tie and wispy textured crystals. Ocelli in the slag include fayalite glass (Fe2SiO4) and quartz (Fig. 6b). It fills or lines vesicles (as amygdales) that formed from entrapped gases. Moreover, it fills fractures that crosscut the hematite matrix of the slag. Fayalite results from the reaction of iron oxide (FeO) and silicate gangue minerals in the ore, and is the common Fe-rich endmember of the olivine solid-solution series that forms at high temperature (<1200 8C) and reducing conditions. It is present in all samples of slag from the Dem study area and confirms that temperatures in the furnaces did not exceed approximately 1200 8C. Partially melted grains of quartz are also present in all samples because much of the primary iron ore mined at Dem is hosted in quartz veins; it is likely that it was difficult to liberate the quartz

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 10

K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

Fig. 6. Photomicrographs of slag from the Dem mine site. (A) Fine to coarse skeletal hematite with spinifex in a hematite matrix from furnace 4. (B) Herringbone textured hematitewu¨stite (Wu) with ocelli of fayalite (Fa) and quartz (Qtz), and vesicles (V) in a hematite matrix. (C) Beautiful skeletal structure of hematite spinifex in hematite matrix.

from the iron in the smelting process. Alternatively (and as stated before), quartz may have been added as flux to improve the viscosity of the slag. Pyrite and Fe metal occurs as rare inclusions in the slag samples and sometimes fills veins and vesicles. The presence of pyrite is interpreted to mean that the iron ore feed into the furnaces was weakly sulphide-bearing, or contaminated with host rocks that were sulphide-bearing (i.e., sulphidic and ferruginous shale and siltstone). Summarily, the smelting processes attained temperatures high enough to form iron-silicate slag and not pure iron (>1500 8C). This is confirmed by the presence of fayalite, rare ulvospinel and wu¨stite, which form in the range 980–1200 8C, and perhaps rare iron metal in the slag. The iron ore feed into the furnace was contaminated with quartz and sulphide (pyrite and pyrrhotite) suggesting that ore dressing was not an exact process. 7. Radiocarbon dating pilot study The main aim of dating charcoal samples from furnaces in the Dem study area, using AMS radiocarbon dating, was to constrain

the age of smelting as a proxy for artisanal mining activity at the Dem mine site, with the assumption that furnace charcoal formed at the same time as mining and processing of iron. The sampling of furnaces was undertaken as a pilot study in 2011 with the premise of undertaking further studies if sufficient funds could be found to mount a fuller study, and if valuable results could be obtained. Six surface and near surface charcoal samples were selected and collected from 11 different furnaces. Sampling was undertaken in such a way that impact on the furnace structures was minimised. Samples were prepared at the University of Witwatersrand Johannesburg and radiocarbon dating was performed at Beta Analytic Laboratories in Miami Florida (USA). The results of the six samples analysed are presented in Table 1, which lists the measured and conventional radiocarbon ages. Table 2 presents the results of two samples with 1s (AD/BP) and 2s (AD/BP) calibrated ages derived, following Talma and Vogel (1993). All samples submitted for radiocarbon dating were immediately subjected to substantial quality control measures. They were cross-checked throughout the process to maximise precision. The measurements of each 14C were followed by a statistical error of

Table 1 Results of samples processed and calibrated by Beta Analytic using the Pretoria Calibration Procedure program (Talma and Vogel, 1993). Sample name

Analysis method

Measured radiocarbon age (pMC/BP)

FOR02-01CAR FOR05-02CAR FOR06-02CAR FOR08-01CAR FOR09-01CAR FOR11-01CAR

AMS-standard AMS-standard AMS-standard AMS-standard AMS-standard AMS-standard

107.6  0.3 pMC 105.2  0.3 pMC 230  30 BP 110  30 BP 103.8  0.4 pMC 104.8  0.4 pMC

delivery delivery delivery delivery delivery delivery

13

C/12C ratio

Conventional radiocarbon age (pMC/BP)

20.5% 16.3% 12.9% 18.3% 16.0% 22.3%

106.7  0.3 pMC 103.4  0.3 pMC 430  30 BP 220  30 BP 101.9  0.4 pMC 104.3  0.4 pMC

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx Table 2 Radiocarbon ages from the Dem study area. The dates have been calibrated with the INTCAL04 Radiocarbon Age Calibration after Talma and Vogel (1993). Sample name

1 sigma (AD) 68%

1 sigma (BP) 68%

2 sigma 95%

Cal BP

FOR06-02CAR FOR08-01CAR FOR08-01CAR

1440–1460 1650–1670 1780–1800

510–490 300–280 170–150

1430–1480 1640–1680 1740–1800

520–470 310–270 210–150

uncertainty (1s and 2s). The first statistical error (1s) indicated that there is a 68% possibility that the true result may fall within 1s. The second statistical error (2s) showed that there is a 95% chance that the true result may fall within 2s (Bowman, 1990). The conventional radiocarbon ages of the selected samples were calculated using the Pretoria Calibration Procedure programme (after Vogel et al., 1993). They are reported together with the standard deviation of the laboratory measurement and then doubled to obtain the 95% probability interval. Conversion of 14C calendar age is dependent on the calibration curve. The radiocarbon ages and errors were rounded with respect to the conventions of Stuiver and Polach (1977). The conventional radiocarbon ages (95% probability) from the Dem iron mine site cover the Roman calendar ages of 1430–1480, 1640–1680, 1740–1800 and 1940–1950 AD. Samples FOR0201CAR, FOR05-02CAR, FOR09-01CAR, and FOR11-01CAR from furnaces 2, 5, 9 and 11, respectively, provided modern dating results, which are 0 BP, thus indicating that modern material had been introduced during the last 60 years or so. Based on ethnographic evidence that mining, processing or smelting of iron did not occur in the Dem region in the last 80–90 years, these samples were assumed to have been contaminated with modern organic material and were excluded, having been deemed insignificant. However, the 2s calibrated ages (95% probability) for samples FOR06-02CAR and FOR08-01CAR arguably suggested that smelting of iron may have been active in the Dem region at furnace 6 during the period 1430–1480 AD, and at furnace 8 during the periods 1640–1680 AD and 1740–1800 AD. These are the only samples that were collected from deep in the furnace pile, with access being made easy because the sides of the furnaces had been breached and collapsed into a small rivulet. Consequently, the furnace stratigraphy was fully exposed over approximately 60 cm of depth. Furnaces 6 and 8 are situated immediately east of the open cast and underground mines (Fig. 3). Clearly, it will be necessary to sample the full furnace stratigraphy in any future study of the Dem site. Interestingly, although quite limited in number, the oldest age corresponds broadly with the rapid expansion of iron technology from 1400 to 1600 AD across sub-Saharan Africa (Ross, 2002) and the rise of the Songhai Empire in the 15th century. The radiocarbon ages may represent the historical ages of iron working in this part of West Africa; a fuller investigation of the Dem site is thus warranted. Certainly, there is currently little evidence of iron smelting for the period 1430–1480 AD from the Dem region, although Insoll (1997) and Ross (2002) advocated that iron technology in the period 1400–1600 AD was one of a series of fundamental social assets that assisted the growth of significant centralised kingdoms, particularly the Songhai Empire (under the rule of Sonni Ali) which was seated in Djenne or Timbuktu or Gao (Fig. 1). These towns are respectively situated 360 km northwest, 450 km north-northwest and 370 km northeast of Dem across the international border of present day Burkina Faso and Mali. By 1500 AD, the Songhai Empire extended from central Niger in the east to Senegal in the west, and from central Burkina Faso in the south to northern Mali.

11

Olson (1979) reported that the Songhai economy was a clan based system linked to artisanal skill (craft guilds), with the most important being metalworks, mechanics and carpentry. Trade in gold, iron and salt were central to the Songhai economy and fundamental to its power base (Masonen, 1997; Park, 2011). Although speculative, the limited radiocarbon results for sample FOR06-02CAR may suggest a link between the Dem site and the demand for metal in the region during the 15th century. In contrast, the periods 1640–1680 AD and 1740–1800 AD are difficult to interpret given the limited historical data for the Dem\Kaya region in the 17th–18th century. However, the history of iron mining and the work of blacksmiths is known from early workings across West Africa (Dueppen, 2008; Holl, 2009) and were central in their societies, up until the 20th century, as a maledominated practice of considerable importance (Helfrid, 2004; Birba, 2010). There is no reason to believe that iron working diminished in cultural significance in that time. Thus, the radiocarbon ages may represent historical ages of reworking or continued practice from the 15 century onwards. 8. Discussion and conclusion In this study, the dynamics of an extensive iron mining, processing and production site in Burkina Faso that may link with demand for metal in the region during the 15th century, and through to the 17th–18th centuries, were examined. The site features a limited number of underground workings and surface open pits. These mine workings show no modern mechanised mining techniques. Ore extraction areas and 11 smelting sites were located. The evidence of iron smelting in the study area includes the remains of at least 11 furnaces (furnace wall and floor fragments), and hundreds of fragments of tuyeres, rare crucibles (perhaps not used in iron smelting), burnt clay and iron slag. Small scale mining of iron ore was conducted in the Dem study area as evidenced as open cast and underground mines, waste dumps, processing sites, furnaces sites and a relatively large mining footprint. The ASM operators employed both surface and underground methods, which yielded two medium bowl-shaped surface mines (OC1 and OC2) and perhaps other smaller workings such as OC3, and two underground mines (UG 1 and 2) with small galleries, adits and raises. The underground mine workings and opencasts were clearly mined using selective mining methods, and iron ore from iron-rich fractures in quartz veins was extracted. Iron-rich ferricrete duricrust was also mined from open casts. Waste rocks included quartz, silicrete and iron-poor ferricrete, and these were dumped proximal to mining operations, similar to what occurs at ASM sites around the world today. The ore might have been man-hauled to the surface in bags or sacks: there is no evidence at the Dem site of mechanical aids for hoisting and hauling ore. Traditional methods of ore processing were used and included handpicking, sorting and crushing, which is similar to non-mechanised techniques still employed today in Burkina Faso at many artisanal mines. The host rocks mined at Dem consisted of iron-rich shale, siltstone and ferricrete duricrust. In the iron-rich shales and siltstone, mining focussed on iron-rich fractures in quartz veins as the primary ore. The ore was dominantly composed of magnetite (Fe3O4) and hematite (Fe2O3), with impurities of SiO2. The processing sites featured at least 11 furnaces. The furnaces are littered with tuyere fragments and rare crucibles made of sandy clay, which was possibly sourced locally from the alluvial plain on which the furnaces are situated (Fig. 3). Tuyere fragments are sometimes welded together with slag in the base of the furnace structure. Waste areas are situated adjacent to most furnaces, suggesting a modest level of industrial activity ranging from the extraction of ore, processing and dressing, building of furnaces,

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 12

K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx

making of tuyeres and crucibles, and processing of slag. This physical evidence suggests that the Dem site may have been an important historic iron-producing region and a long-term or seasonal site of ASM activity. A source of fuel remains speculative but the location of the mine site east of Lac (Lake) Dem (Fig. 3), which was historically a wooden valley prior to damming, may have meant that a source of fuel in the form of timber was readily available. In any case, charcoal may have been transported to the site in the same way that charcoal (or timber for charcoal production) is transported to the growing urban centres nearby from the local forests (Kra¨mer, 2002; Ouedraogo et al., 2010). A source of flux may have the carbonaceous rocks and quartz that formed part of the host rocks of the region. A considerable amount of clean crushed quartz and outcrops of quartz veins occurs in the mines themselves, as well as regionally (Hein et al., 2004). A source of clean carbonate rock, which would make an ideal flux, does not occur in the region according to regional maps assembled by Hottin and Ouedraogo (1992), Hein et al. (2004) and Castaing et al. (2003), although the rocks host impure carbonate, which may have been used. Alternatively, the combination of carbon from charcoal, and quartz-carbonate impurities in the iron ore feed, together acted as fluxing agents. Iron smelting at Dem converted magnetite-hematite ore from iron-rich fractures in quartz veins, and hematite-goethite ore from ferricrete, to slag-iron in furnaces reaching temperatures of 980– 1200 8C. The character of the slags in the region show that furnace temperatures varied from furnace to furnace and with each smelting episode so that slags ranged from Type 1 low viscosity glassy or ropey slag to Type 2 moderate viscosity blocky vesicular slag. Slags were dumped proximal to the furnace sites. The distinguishing behaviour of slag under any operational environment is based on the slag composition and the range of components in it, where the flux materials and the type of ore used are most important (Muszer, 2000). Based on the petrographic descriptions, fayalite, quartz, magnetite and spinifex-shaped hematite were the dominant minerals in all slag samples with minor sulphides and iron metals. Ulvospinel was identified in a few samples. The presence of fayalite and silica in all slag samples suggests a moderate to good quality slag-iron product was produced during smelting, or perhaps an iron-silicate slag (Craddock and Meeks, 1987). Furnace temperatures no greater than 1200 8C were reached (cf., Paynter, 2006). Iron production at the Dem site was bloomery furnace smelting (cf., Childs, 1991). A limited number of radiocarbon dates were obtained from charcoal samples collected from the base of furnaces 2, 5, 6, 8, 9 and 11. The 2s calibrated results (95% probability) cover the Roman calendar ages of 1430–1480 AD, 1640–1680 AD and 1740– 1800 AD. Findings from ethnographic research suggested that iron mining in the Dem region, or any significant artisanal mining in the Dem region for any metal commodity, had not occurred in the last 80–90 years and there are no records of significant (largescale) artisanal iron extraction, processing or smelting in the Dem region known to staff at the Museum of Kaya. Although speculative, the limited radiocarbon results for sample FOR0602CAR may suggest a link between the Dem site and the expansion of iron use in West Africa. Iron was initially forged in West Africa as early as the 6th century BC (Insoll, 1997; Pleiner, 2000; Ross, 2002; Holl, 2009) and iron smelting was practiced in the southeast of Burkina Faso (250 km southeast of Dem) by the Bura culture. By the 15th century, forging and smelting of iron (and gold) were widespread in West Africa and particularly, the Songhai economy, which dominated trade in the region. It is possible that the limited radiocarbon results for sample FOR06-02CAR suggest a link between the Dem site and the demand for metal in the region

at that time, but further research is needed to strengthen this interpretation. The multiple ages calculated in sample FOR0801CAR show that re-furnacing occurred. In any event, the Dem site shows evidence of an African society at work with skills in metal extraction, production and forging technology (pyrometallurgy). These skills must have included: (1) exploration for iron resources and recognition of reservable iron ore grades, (2) mining methodologies both in underground and open cast mining, (3) ore processing techniques and refining practice, and (4) smelting and beneficiation of iron ore. There is evidence of a mining value chain. The production of charcoal and the manufacture of tuyeres, crucibles and furnace wall clay may have occurred in parallel to iron ore production as ancillary industries. The Dem ASM site should therefore be protected and preserved as a geological, mining and cultural monument by the relevant authorities in Burkina Faso at the state and provincial levels. This will prevent degradation and destruction of the heritage, and consequently preserve the values and history of the area for present and future generations in sub-Saharan Africa. Acknowledgements This project was supported by the National Research Foundation of South Africa and Professorial Funds. We acknowledge the contribution of the Museum of Kaya and villagers of Liligomde (Pe`re` Continwagre, Grampre Conviba, and Arve`re Grape`re Yonka) for their help on the insight on historic artisanal mining of iron ore in Burkina Faso. The manuscript was considerably improved by the careful reviews of R. Thorton, G. Hilson, and two anonymous reviewers.

References Agricola, G., 1556. De Re Metallica. Translated from Latin Edition of 1556 with Biographical Introduction, Annotations and appendices upon the development of Mining Methods, metallurgical Processes, Geology, Mineralogy and Mining Law from the earliest times to the 16 century. (H.C. Hoover, L.H. Hoover, Trans.)Published for the Translators by the Mining Magazine, Salisbury House, London, E.C. 389 pp. Anon., 1999. Social and Labour Issues in Small-Scale Mines. Report for discussion at the Tripartite Meeting on Social and Labour, Issues in Small-Scale Mines, Geneva, 17–21 May 1999. International Labour Office, Geneva. Aryee, B.N.A., Ntibery, B.K., Atorkui, E., 2003. Trends in the small-scale mining of precious minerals in Ghana: a perspective on its environmental impact. J. Clean. Prod. 11, 131–140. Ayensu, E.S., 1997. Ashanti Gold: The African Legacy of the World’s Most Precious Metal. Published by the Marshall Editions Development Ltd. 10 pp. Beauvais, A., Ruffet, G., He´nocque, O., Fabrice, C., 2008. Chemical and physical erosion rhythms of the West African Cenozoic morphogenesis: the 39Ar–40Ar dating of supergene K–Mn oxides. J. Geophys. Res. 113,, 15 pp. Birba, S.N., 2010. La pale´ome´tallurgie du fer dans la province du Bam (Burkina Faso): L’identite´ des acteurs et la mobilite´ des techniques. In: Presentation at the colloquim, Me´tallurgie du fer et socie´te´s africaines Bilans et nouveaux paradigmes dans la recherche´ anthropologique et arche´ologique, Universite´ de Provence, Maison Me´diterrane´enne des Sciences de l’Homme, France, 23–24 April 2010. Bowman, S.G.E., 1990. Radiocarbon Dating; Interpreting the Past. British Museum Publications 64 pp. Brown, E.T., Bourle`s, D.L., Colin, F., Safno, Z., Raisbeck, G.M., Yiou, F., 1994. The development of iron crust lateritic systems in Burkina Faso, West Africa examined with in-situ-produced cosmogenic nuclides. Earth Planet. Sci. Lett. 124, 19–33. Burke, A., Cloete, J., 2004. Restoring tire tracks: lessons from the southern Namib Desert. Ecol. Restor. 22, 269–274. Castaing, C., Billa, M., Mile´si, J.P., Thie´blemont, D., Le Mentour, J., Egal, E., Donzeau, M. (BRGM) (coordonnateurs) et Guerrot, C., Cocherie, A., Chevremont, P., Tegyey, M., Itard, Y. (BRGM), Zida, B., Ouedraogo, I., Kote, S., Kabore, B.E., Ouedraogo, C. (BUMIGEB), Ki, J.C., Zunino (ANTEA), 2003. Notice explicative de la Carte ge´ologique et minie´re du Burkina Faso a` 1/1 000 000. Childs, S.T., 1989. Style, technology and iron smelting furnaces in Bantu-speaking Africa. J. Anthropol. Archaeol. 10, 332–359. Childs, S.T., 1991. Style, technology, and iron furnaces in Bantu-speaking Africa. J. Anthropol. Archaeol. 10, 332–359. Craddock, P.T., 1995. Early metal mining and production. Smithsonian, Edinburgh 4, 898–905.

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004

G Model

EXIS-23; No. of Pages 13 K.A.A. Hein, T.A. Funyufunyu / The Extractive Industries and Society xxx (2014) xxx–xxx Craddock, P.T., Meeks, N.D., 1987. Iron in ancient copper. Archaeometry 29, 187– 204. Doherty, R.D., Hutchings, K.M., Smith, J.D., Yo¨ru¨k, S., 1985. The reduction of Hematite to Wustite in a laboratory fluidized bed. Metall. Trans. B 16B, 425–432. Dueppen, S.A., 2008. Reinventing equality: The archaeology of Kirikongo, Burkina Faso. (doctor of philosophy thesis)University of Michigan 328 pp. ECA, 2002, December. Compendium on Best Practices in Small-Scale Mining in Africa. Economic Commission for Africa. Published by the United Nations Economic Commission for Africa Publishing Unit, Addis Ababa, Ethiopia 230 pp. ECA, 2012. Minerals and Africa’s Development: The International Study Group Report on Africa’s Mineral Regimes. Published by the United Nations Economic Commission for Africa Publishing Unit, Addis Ababa, Ethiopia 230 pp. Friede, H.M., 1980. Iron Age mining in the Transvaal. J. S. Afr. Inst. Min. Metall. (April), 156–165. Friede, H.M., Steel, R.H., 1986. Traditional smithing and forging of South African Bloomery iron. S. Afr. Archaeol. Bull. 41, 81–86. Haaland, G., Haaland, R., Dea, D., 2004. Furnace and Pot: Why the Iron Smelter is a Big Pot Maker, A Case Study from South-Western Ethiopia 20 pp. Hein, K.A.A., 2007. Artisanal mining, its place and role in the mining sector. women in mining – taking stock of the present, shaping the future.In: Colloquium. 8–9 November 2007, University of the Witwatersrand Johannesburg, 15 pp. Hein, K.A.A., 2010. Succession of structural events in the Goren Greenstone belt (Burkina Faso); implications for West African tectonics. J. Afr. Earth Sci. 56, 83–94. Hein, K.A.A., Luning, S., 2009. Artisanal mining in (West) Africa, unexplored research challenges – worlds of debts, encounters for co-habitation: gold mining sites in West Africa. In: Symposium on Research and Training for Sustainable Development in Africa, African University Day, 12 November 2009, Hospital Auditorium, Medical School, University of the Witwatersrand Johannesburg. Hein, K.A.A., Morel, V., Kagone´, O., Kiemde, F., Mayes, K., 2004. Birimian lithological succession and structural evolution in the Goren segment of the Boromo-Goren Greenstone Belt, Burkina Faso. J. Afr. Earth Sci. 39, 1–23. Hein, K.A.A., Nhlengetwa, K., submitted for publication. Zama Zama Mining in the Durban Deep-Roodepoort Area, Johannesburg, South Africa: An Invasive or Alternative Livelihood? Extractive Industries and Society (submitted for publication). Helfrid, S., 2004. Towards Gender Equality in Burkina Faso. Published by Swedish International Development Cooperation Agency. Department for Policy and Methods 87 pp. Herbert, E., 1993. Iron gender and power, Indiana University. Precious metals of the age of expansion.In: Papers of the 14th International Congress of the Historical Sciences. , 123 pp. Hilson, G., 2002. Small-scale mining and its socio-economic impact in developing countries. Nat. Resource Forum 26, 3–13. Hilson, G., McQuilken, J., 2014. Four Decades of Support for Artisanal and SmallScale Mining in Sub-Saharan Africa: A Critical Review. The Extractive Industries and Society. Hoadley, M., Limpitlaw, D., 2004. The artisanal and small-scale mining sector and sustainable livelihoods. In: Mintek Small Scale Mining Conference, Book of Proceedings. 9 September, NASREC, Johannesburg, 9 pp. Holl, A.F.C., 2009. Early West African metallurgies: new data and old orthodoxy. J. World Prehistory 22, 415–438. Hottin, G., Ouedraogo, O.F., 1992. Carte Ge´ologique du Burkina Faso, 2nd edition. Bureau de Recherches Ge´ologiques et Minie`res, Burkina Faso. Hudson, T.L., Fox, F.D., Plumlee, G.S., 1999. Metals, Mining and the Environment. AGI Environmental Awareness Series 368 pp. Insoll, T., 1997. Iron Age Gao: an archaeological contribution. J. Afr. History 38, 1–30. Karbowniczek, M., 2005. Metallurgical process in ancient shaft furnace, theoretical consideration. Assoc. Metall. Eng. Serbia 5, 145–155. Kense, J.F., 1983. Traditional African Iron Workings. African Occasional Papers. Department of Archaeology, University of Calgary 25 pp. Killick, D., Gordon, R.B., 1989. The mechanism of iron production in the bloomery furnace. In: Farquhar, R.M., Hancock, R.G.V., Pavlish, L.A. (Eds.), Proceedings of the 26th International Archaeometry Symposium, Held at University of Toronto, Toronto, Canada, May 16th to May 20th 1988. University of Toronto, Toronto, pp. 120–123.

13

Kra¨mer, P., 2002. The Fuel Wood Crisis in Burkina Faso. In: Solar Cooker as an Alternative. Solar Cooker Archive, Ouagadougou, Burkina Faso, 19 pp. Lovitz, B.A., 2006. Scales Responsible for Gold Mining: Overcoming Barriers to Cleaner Artisanal Mining in Southern Ecuador. The Faculty of the Graduate College of the University of Vermont 111 pp. Masonen, P., 1997. Trans-Saharan trade and the West African discovery of the Mediterranean. In: M’hammad Sabour, Vikør, K.S. (Eds.), Ethnic Encounter and Culture Change. Papers from the Third Nordic Conference on Middle Eastern Studies, Bergen 1997, vol. 3. Nordic Research on the Middle East, pp. 116–142. Medenbach, O., ElGoresy, A., 1982. Ulvo¨spinel in native iron-bearing assemblages and the origin of these assemblages in basalts from Ovifak, Greenland, and Bu¨hl, Federal Republic of Germany. Contrib. Mineral. Petrol. 80, 358–366. Miller, D.E., Van Der Merwe, N.J., 1994. Early metal working in Sub Saharan Africa. J. Afr. History 35, 1–36. Muszer, A., 2000. An Outline of Ore Microscopy. University of Wroclaw, Poland 233 pp. Mutemeri, N., Petersen, F.W., 2002. Small-scale mining in South Africa: past, present and future. Nat. Resources Forum 26, 286–292. Okafor, E.E., 2002. La re´duction du fer dans les Bas-Fourneaux—Une industrie vieille de 2500 ans au Nigeria. In: Bocoum, H. (Ed.), Aux origines de la me´tallurgie du fer en Afrique. Editions UNESCO, Paris, pp. 35–48. Olson, J.S., 1979. The Ethnic Dimension in American History, 1st edition. Published by St Marins Press 440 pp. Ouedraogo, I., Tigabu, M., Savadogo, P., Compaore Oden, P.C., Ouadba, J.M., 2010. Land cover change and its relation with population dynamics in Burkina Faso, West Africa. Land Degr. Dev. 21, 453–462. Park, D.P., 2011. Climate change, human response and the origins of urbanism at Timbuktu. (doctoral thesis)Yale University, Department of Anthropology 577 pp. Paynter, S., 2006. Regional variation in Bloomery smelting slag of the Iron Age and Romano-British periods. Archaeometry 48, 271–292. Paynter, S., 2011. Introduction to Heritage Assets, Pre-industrial Ironworks 6 pp. Peters, L.F.H., Hein, K.A.A., 2013. Volcanic facies and preliminary geochemical observations of the Goren greenstone belt, Burkina Faso. In: EGU General Assembly 2013, 7–12 April 2013, Vienna, Austria. Phillipson, D.W., 1975. The chronology of the Iron Age in Bantu Africa. J. Afr. History 16, 321–342. Pleiner, R., 2000. Iron in Archaeology: The European Bloomery Smelters. Published ˇ R, Prague, ISBN: 80-86124-26-6 400 pp. by Archeologicky´ u´stav AVC Ross D E.G., 2002. The age of iron in West Africa. In: Heilbrunn Timeline of Art History. The Metropolitan Museum of Art, 2000–, New York. , http://www.metmuseum.org/toah/hd/iron/hd_iron.htm (accessed October 2002). Schuhmann, R., Powell, R.J., Michal, E.J., 1953. Constitution of FeO–Fe2O3–SiO2 system at slag making temperatures. Metall. Mater. Trans. 197, 1097–1101. Scott-Garrett, C., 1956. Romano-British sites at Chestnuts Hill and Popes Hill, Forest of Dean. Trans. Bristol Gloucestershire Archaeol. Soc. 75, 199–202. Sinding, K., 2005. The dynamics of artisanal and small-scale mining reform. Nat. Resources Forum 29, 243–252. Smith, P., Timothy, G., 1993. Iron and Steel Industry. Editor, Steel times and Steeltimes Interval 45 pp. Spry, P.G., 1987. The chemistry and origin of zincain spinel associated with the Aggeneys Cu–Pb–Zn–Ag deposits, Namaqualand, South Africa. Mineral. Deposita 22, 262–268. Stuiver, M., Polach, H.A., 1977. Reporting of 14C data. Radiocarbon 19, 355–363. Talma, A.S., Vogel, J.C., 1993. A simplified approach to calibrating C14 dates. Radiocarbon 35, 317–322. Taylor, L.A., Williams, R.J., McCallister, R.H., 1972. Stability relations of ilmenite and ulvospinel in the Fe–Ti–O system and application of these data to lunar mineral assemblages. Earth Planet Sci. Lett. 16, 282–288. Tylecote, R.F., Austin, J.N., Wraith, A.E., 1971. The mechanism of the Bloomery process in shaft furnaces. J. Iron Steel Inst. 209, 342–363. Verhoogen, J., 1962. Oxidation of iron-titanium oxides in igneous rocks. University of California, Berkeley. J. Geol. 70, 168–181. Vogel, J.C., Fuls, A., Visser, E., 1993. Pretoria calibration for short lived samples. Radiocarbon 35, 73–85. Wassenaar, T.D., Henschel, J.R., Pfaffenthaler, M.M., Mutota, E.N., Seely, M.K., Pallett, J., 2012. Ensuring the future of the Namib’s biodiversity: ecological restoration as a key management response to a mining boom. J. Arid Environ. 93, 126–135.

Please cite this article in press as: Hein, K.A.A., Funyufunyu, T.A., Artisanal mining in Burkina Faso: A historical overview of iron ore extraction, processing and production in the Dem region. Extr. Ind. Soc. (2014), http://dx.doi.org/10.1016/j.exis.2014.04.004