Maritime Archaeology: New Research on the Metal Finds from Akko 1 Shipwreck

Our associate editor of maritime archaeology, Nicolás Ciarlo, has brought to our attention the following research:

Metallurgical characterization of metal finds from the nineteenth century Akko 1 shipwreck

D. Ashkenazi
School of Mechanical Engineering, Tel Aviv University, Ramat Aviv 6997801, Israel.

D. Cvikel
Leon Recanati Institute for Maritime Studies and Department of Maritime Civilizations, University of Haifa, Haifa 3498838, Israel.

The Akko 1 shipwreck is the remains of a 26-m-long naval auxiliary brig, sailing under the Egyptian flag. Various metal objects were retrieved from the shipwreck, among them three cannonballs, a box containing carpenter’s accessories, six flintlock muskets with brass fittings, and 158 brass cases. These objects were studied by various archaeometallurgical testing methods, which revealed the composition and manufacturing processes of the objects. Moreover, examination of some of the objects also exposed their period of production and possible provenance. The metallurgical study of the metal finds from the Akko 1 shipwreck, combined with the historical background and archaeological evidence, suggests that the ship was built during the first half of the nineteenth century, and apparently wrecked during the 1840 naval bombardment of Akko.

The Akko 1 shipwreck was found in 4 m of water inside the ancient harbour of Akko, Israel, and was excavated over three seasons. The finds included metal, organic and ceramic objects. The metal finds comprised ammunition and related artefacts, such as cannonballs, musket fittings, and brass cases. The results of the archaeological research and the study of the historical background suggest that the Akko 1 shipwreck is the remains of a naval auxiliary brig, built at the beginning of the nineteenth century, and sailing under the Egyptian flag. In light of the naval bombardment of Akko in 1840, it is suggested that the ship was wrecked during this event (Cvikel and Kahanov, 2013). A shipwreck represents a unique moment frozen in time. Therefore, the study of the Akko 1 shipwreck, including its metal finds, contributes valuable information and enriches our understanding of the technological changes that occurred in the early nineteenth century. 

Experimental methods and tests
The various metal objects retrieved from the Akko 1 shipwreck include 9-, 12-, and 24-pdr cannonballs (Ashkenazi et al., 2012; Cvikel et al., 2013; Mentovich et al., 2010), a box containing iron nails and two split pins (Cvikel et al., 2016), six flintlock muskets with their brass fittings (Cvikel et al., 2017), and 158 brass cases (Ashkenazi et al., 2011; Ashkenazi et al., 2014) (Fig. 1).

Figure 1. The metal objects retrieved from the Akko 1 shipwreck: (a) 24-pdr cannonball; (b) 12-pdr cannonball; (c) 9-pdr cannonball; (d) iron bolt; (e) and (f) iron nails; (g) two split pins; (h) remains of a flintlock musket with brass fittings; (i) ramrod pipe; (j) front view of brass case with lid in place; and (k) crescent-shaped stamped mark observed on the back of two brass cases (Photo credit: (a) to (c) and (k) taken by J.J. Gottlieb, (d) to (g) by N. Iddan, (h) by S. Levy, (i) by D. Ashkenazi, (j) by J.J. Gottlieb and D. Cvikel).

These objects were studied by various non-destructive testing (NDT) and destructive testing methods, as described below: 

(a) Visual testing (VT) to detect visible details that may indicate the state of preservation and manufacturing process.
(b) Chemical analysis with handheld (HH) X-ray fluorescence (XRF) instrument to identify general composition. 
(c) Destructive metallographic testing of specimens according to ASTM E3 standard. Specimen preparation stages were: grinding the surface with silicon carbide papers, followed by polishing with alumina and diamond pastes, and final polish with 30 nm colloidal silica suspension paste. After each stage the specimens were cleaned in an ultrasonic bath. Finally, the iron specimens were etched with Nital acid, and the brass specimens were etched with hydrochloric acid in ferric chloride solution. All etched specimens were cleaned with ethanol before microstructural examination.
(d) Metallographic examination of specimens by light microscopy (LM) to determine the connection between microstructure and manufacturing processes.
(e) Vickers microhardness tests of homogeneity.
(f) Scanning electron microscopy (SEM) equipped with an energy dispersive spectroscopy (EDS) analysis to correlate between local microstructure and composition.

Results and discussion
The cannonballs
XRF analysis of the 24-pdr and 9-pdr cannonballs (Fig. 1a, 1c, respectively) revealed they were mostly composed of iron, with less than 1.0 wt% of silicon, manganese and phosphorus (Mentovich et al., 2010). The manganese concentrations higher than 0.5 wt% found in both cannonballs must have been the result of intentional addition, implying a post-1839 manufacturing date (Mentovich et al., 2010, pp. 2524–2526; Wayman, 2000, p. 265; Wiltzen and Wayman, 1999, p. 119).

Metallographic examination by LM and SEM-EDS of the etched surfaces of the 24-pdr cannonball and samples demonstrated different microstructures of different parts of the cannonball (Fig. 2a‒b). Grey cast iron containing dark graphite flakes surrounded by pearlite matrix was observed on the exterior of the 24-pdr cannonball (Fig. 2a), while a white cast iron, made of cementite precipitates (dark grey) surrounded by pearlite matrix, was observed in its interior (Fig. 2b). This was the result of variations in both the local chemical composition and the cooling rate during the casting process (Ashkenazi et al., 2012; Mentovich et al., 2010). 

Figure 2. Metallographic observation of etched microstructure: (a) grey cast iron containing graphite flakes near  the surface of 24-pdr cannonball (SEM); (b) white cast iron in the interior of 24-pdr cannonball (SEM); (c) slag inclusions surrounded by a ferrite matrix of 12-pdr wrought iron cannonball (LM); and (d) carpenter's split pin, with large ferrite grains in the interior and Widmanstatten ferrite and pearlite on the surface (LM). 

Metallographic examination by LM and SEM-EDS of the 9-pdr cannonball revealed a microstructure of white cast iron. Petrographic examination of sand found inside the 9-pdr cannonball (remains of the casting process) revealed presence of quartz, opaque minerals and riebeckite. Riebeckite granite mineral is found in several locations in Egypt, indicating that the workshop of the 9-pdr cannonball was probably located in Egypt (Mentovich et al., 2010).

XRF analysis of the 12-pdr cannonball revealed that it was composed of iron, with less than 1.0 wt% of silicon, molybdenum, nickel and copper (Cvikel et al., 2013). LM and SEM analyses (Fig. 1b) revealed a homogeneous microstructure of relatively pure α-ferrite phase, containing glassy wüstite and fayalite slag inclusions, typical of an annealed wrought iron product (Fig. 2c). This was unexpected and unique, since producing wrought iron cannonballs by hammering is a complex process. The chemical composition, microstructure and microhardness values show that the 12-pdr cannonball was produced by a completely different technology and probably at a different place from the cast iron 9-pdr and 24-pdr cannonballs. It is suggested that the 12-pdr cannonball might have been used as ballast (Cvikel et al., 2013).

The carpenter’s accessories
One bolt, 12 nails and two split pins were found in the carpenter’s wooden tool box (Fig. 1d‒g). The XRF analysis of the bolt (Fig. 1d) indicated it was made of iron, with 0.1 wt% chromium; and the nails (Fig. 1e‒f) were made of iron, containing small amounts (up to 1.5 wt%) of silicon, nickel, copper, chromium, and cobalt(Cvikel et al., 2016). 

The metallographic LM and SEM observations of the bolt and nails, combined with the microhardness values, indicated a heterogeneous microstructure. The carbon concentration in the nails varied between about 0.01 wt% in the interior and up to 0.7 wt% at the surface (Cvikel et al., 2016). The difference between the internal and external microstructures indicates that the carpenter’s accessories were carburized.

The XRF analysis of the two split pins demonstrated that they were made of iron, containing 0.1 wt% copper. Metallographic LM and SEM observations showed ferrite phase in the interior of the split pins, and Widmanstätten ferrite and pearlite microstructures near the surface (Fig. 2d). The results indicate that the split pins were also carburized.

The brass fittings
XRF and EDS chemical analyses of the musket fittings: side-plate, butt-plate and ramrod pipe, revealed that they were all made of α-brass (Fig. 1h‒i). The presence of more than 30 wt% zinc indicates that the musket fittings were most probably made from speltered brass, and manufactured in the early nineteenth century (Cvikel et al., 2017). LM showed a dendritic network with shrinkage cast defects in the side-plates (Fig. 3a). The composition, combined with LM and SEM metallographic observations of coarse dendritic microstructure (Fig. 3b‒c), indicate that the fittings were manufactured by casting, probably at the same workshop. 

Figure 3. Metallographic observation of etched microstructure: (a) brass side-plate of flintlock musket, showing casting defects (shrinkage) formed during solidification (LM); (b) corse dendritic microstructure formed in the interior cast brass of side-plate (LM); (c) brass ramrod pipe, with coarse dendritic microstructure (SEM); and (d) the microstructure of brass case, showing alpha-brass phase with twins (LM). 

The brass cases
The XRF and EDS analysis results show that the 158 cases were made of brass containing about 30 wt% zinc. The presence of tin-lead filler indicated that a soldering process was used in their manufacture. LM observation revealed an alpha-brass phase with presence of annealing twins (Fig. 3d). The uniform thickness of the cases and their microstructure indicate that they were made of rolled sheets, and hand-made by using simple tools. The composition and microstructure suggest that the brass cases were manufactured during the first half of the nineteenth century. 

The archaeometallurgical study of the various metal objects retrieved from the Akko 1 shipwreck revealed their composition and manufacturing processes, and supported the dating of the ship, indicating that she was in service no later than the first half of the nineteenth century, and apparently wrecked during the 1840 naval bombardment of Akko. The study of the metal finds retrieved from the shipwreck contributes valuable information to our understanding of the context of the shipwreck, as well to technological changes occurring in the Industrial Revolution during the early nineteenth century. 

The underwater excavation and research of the Akko 1 shipwreck were supported by the late Ron Marlar, the Yaacov Salomon Foundation, the late Reuven Sadnai—Coral Maritime Services Ltd, the Halpern Foundation, a Sir Maurice Hatter Fellowship, the Hecht Trust, a Jewish National Fund Fellowship, the President, Rector, Dean and Faculty of Humanities, University of Haifa, and anonymous donors, to whom the authors are grateful. The authors would like to thank J.B. Tresman for the English editing.

Ashkenazi, D., Cvikel, D., Iddan, N., Mentovich, E., Kahanov, Y., Levinstain, M., 2011. Archaeometallurgical study of the brass cases from the Akko 1 Shipwreck. Journal of Archaeological Science 38:2410–2419.
Ashkenazi, D., Mentovich, E., Kahanov, Y., Cvikel, D., Barkai, O., Aronson, A., 2012. Archaeometallurgical investigation of iron artifacts from shipwrecks: A review. In: I. Ollich-Castanyer (Ed.), Archaeology, New Approaches in Theory and Techniques, pp. 169–186. InTech Publisher.
Ashkenazi, D., Cvikel, D., Stern, A., Klein, S., Kahanov, Y., 2014. Metallurgical characterization of brass objects from the Akko 1 shipwreck, Israel. Materials Characterization 92:49–63.
Cvikel, D., Kahanov, Y., 2013. The 19th-Century Akko 1 Shipwreck, Israel: hull-construction report. International Journal of Nautical Archaeology 42 (1):167–187.
Cvikel, D., Mentovich, E.D., Ashkenazi, D., Kahanov, Y., 2013. Casting techniques of cannonballs from the Akko 1 shipwreck: Archaeometallurgical investigation. Journal of Mining and Metallurgy, Section B: Metallurgy 49 (1):107–119.
Cvikel, D., Ben-Artzi, T., Ashkenazi, D., Iddan, N., Stern, A., Kahanov, Y., 2016. A box containing carpenter’s accessories from the Akko 1 Shipwreck, Israel: Archaeometallurgical analysis of surviving ironwork. Archaeometry 58 (3):427–440.
Cvikel, D., Ashkenazi, D., Spiegelman, V., Stern, A., Klein, S., Rimer, G., 2017. Flintlock brass fittings from the 19th-century Akko 1 shipwreck, Israel. Arms & Armour 14 (2):138–164.
Mentovich, E., Schreiber, D., Goren, Y., Kahanov, Y., Goren, H., Cvikel, D., Ashkenazi, D., 2010. New insights regarding the Akko 1 shipwreck: A metallurgic and petrographic investigation of the cannonballs. Journal of Archaeological Science 37:2520–2528.
Wayman, M.L., 2000. Archaeometallurgical contributions to a better understanding of the past. Materials Characterization 45:259–267.
Wiltzen, T.S., Wayman, M.L., 1999. Steel files as chronological markers in North American fur trade sites. Archaeometry 41:117–135.