Multianalytical Study of Archaeological Iron Nails of the Roman Period (First to Third Century) in Northern Spain

Abstract

In the archaeometallurgical study of iron nails to investigate Roman manufacturing processes, multi-analyte characterization provides information on alloy composition and microstructure. Nails from the Roman sites of Forua, Aloria, and Iuliobriga (northern Spain) were studied. To characterize the iron phases and microstructures of the nails, optical microscopy (OM), scanning electron microscopy coupled with electron-dispersive spectroscopy (SEM-EDX), micro-Raman spectroscopy, electron backscatter diffraction (EBSD) realized in environmental mode, and microhardness measurements were carried out. The chemical composition of the metal was determined by X-ray fluorescence (XRF). The corrosion mineralogical composition was determined by powder X-ray diffraction (XRD). Aggressive burial conditions had a significant effect on the forms of corrosion of the Forua nails to the point of complete iron loss. Examination of the metal of the nails from the Aloria site revealed that most of the ironwork was made in the villa’s own forge. In the case of the Iuliobriga nails, different degrees of forging were identified associated with different workshops.

1. Introduction

The application of modern characterization techniques in archeometallurgical research on ancient artefacts provides information on alloy composition and microstructure. The analysis of archaeological iron artefacts is essential to elucidate manufacturing processes in ancient times [1,2,3,4,5,6,7]. The microstructure and composition of iron artefacts shed light on the technological skills of ancient blacksmiths and attest to the particular manufacturing processes in each workshop. Knowledge of metallurgical techniques can attest to the existence of commercial contacts among different sites, but also to the assimilation of different production techniques.

In Roman times, iron was the predominant metal used for tools and weapons. Advances in iron and steel technology increased iron production, leading to the expansion of forges [8,9,10]. Innovations in ore processing led to a remarkable increase in iron products and to the diversification and standardization of product manufacture.

The production of iron tools in the eastern Cantabrian region (northern Iberian Peninsula) during the Roman period was organized in several ways [11,12]. Forge remains found in iron ore mining and processing centers were involved in the production and repair of tools used in the manufacture and maintenance of equipment [13,14,15]. Iron tool production workshops were also common in Roman cities or civitates and villages, to produce tools not only for their own consumption but also for trade [16,17]. Evidence of archeometallurgical activity has also been found in Roman military camps [18,19]. In addition, forges in rural villas (villa rustica) repaired and manufactured objects used in villa work [15,20].

The collection of iron artefacts is large and varied and consists of different types of tools, ranging from weapons to agricultural implements and other utensils associated with trade and other activities. Archaeometallurgical studies require the cutting and damaging of artefacts to analyze their composition and microstructure, which limits the availability of artefacts for study. However, the abundance of nails at Roman sites facilitates the access to these objects for study, while the wide distribution of nails allows for the comparison of manufacturing processes between different workshops. The present study analyzes iron nails from forges and workshops in the Roman civitas of Iuliobriga (Cantabria), the village of Forua (Biscay), and the villa of Aloria (Alava).

The Latin name of Iuliobriga means “Fortified Civitas of Julius”. The most important civil civitas in Roman Cantabria (northern Iberia) was located on the route between the Cantabrian coast and the Castilian plateau, with access to the sea through the so-called Portus Victoriae Iuliobrigensium, probably the present-day Santander. Two rooms for forging were identified with two hearths with a water channel inside [16,21]. The iron assemblage consists mainly of nails and tools and more elaborate items, such as five andirons. These artefacts came from an abandonment floor near the hearth of the forge. The andirons are currently on display in the Archaeological Museum of Cantabria (Spain). The chronology of these productive areas corresponds to the second century AD.

The village of Forua was a forum consisting of a place with a market and a tribunal. Located at the estuary of the river Oka, on the left bank of the Urdaibai estuary, on an old and wide meander, it was an important fluvial port. The port of Forua was used to import and export products, including local iron in the form of finished products or bars and agricultural surpluses. In addition to port activities, iron metallurgy was the main activity in this settlement during the second and third centuries AD. Archaeological work attested to at least eight buildings with forges indicating an intense activity in the transformation of iron [22,23,24]. The assemblage of iron artefacts consists mainly of nails, but also highly deteriorated forging tools difficult to identify and remains of the forging process, such as shards of iron, were found. All the iron artefacts were found in the forge rooms.

The site of Aloria was a rural villa, consisting of a farmhouse or villa set in the countryside, with an agricultural and livestock section. The activity of the villa extended between the first and the fifth centuries AD [15]. The villa was located in the valley of Orduña (Alava, northern Iberia), which was also an inland route for transit and a communication route between the Castilian plateau and the Cantabrian Sea. A small room for forging with a hearth was identified. The iron assemblage was found in this area and mainly consists of nails and highly degraded tools, difficult to identify and probably corresponding to farming tools; also, two iron ingots were found.

Nails from these three archeometallurgical contexts were analyzed to characterize iron phases and microstructures using optical microscopy (OM), scanning electron microscopy coupled with electron-dispersive spectroscopy (SEM-EDX), micro-Raman spectroscopy, electron backscatter diffraction (EBSD) realized in environmental mode, and microhardness measurements. In addition, the chemical composition of the metal was determined using X-ray wavelength fluorescence spectroscopy (WDXRF), and the mineral composition of the corrosion was determined by powder X-ray diffraction (XRD). The microstructure and composition provide evidence of iron artefact manufacturing processes such as hammering, heating, and quenching. Both the observed metal composition and the microstructure are directly related to the metallurgical state of the material resulting from the manufacturing process and the use of the object.

This study focuses on the metallographic analysis of iron nails from the sites of Aloria (Alava), Forua (Biscay), and Iuliobriga (Cantabria), in order to gain insights into the manufacturing techniques of the artefacts and to compare the technological skills of rural and urban Roman blacksmiths.

2. Materials and Methods

The small number of samples in this study was due to the restriction in obtaining archaeological material from museums. Only about 10% of the requested material was provided, and only nails were supplied, as they were the most common in the sites studied. Eleven nails from three sites in the north of the Iberian Peninsula, the rural site of Aloria (Alava) and the urban sites of Forua (Biscay) and Iuliobriga (Cantabria), were studied for metallographic analysis. (Figure 1), i.e., four nails (Al-03, Al-06, Al-07 and Al-11) from the Aloria site, five nails (F-1 to F-5) from Forua, and two nails (R-01 and R-02) from Iuliobriga.

The shape of the nails varies from site to site. The nails from Aloria and Forua have a circular cross section with a sharp point at one end, and most have a disc-shaped head (Figure 2a,b). In contrast, the nails from Iuliobriga have a square cross section (Figure 2c). The length of the nails, measured before the metallographic analysis, varies between 2 and 5 cm in the Forua samples, 3 and 5 cm in the Aloria samples, and up to 8 cm in the Iuliobriga samples. All the nails show severe wear and corrosion due to burial.

The metallographic analyses were carried out along a longitudinal section from head to tip, mounted in epoxy resin. The nails were cut longitudinally with a diamond wire saw, using a mineral oil (Nujol) capable of forming a protective film during cutting and preventing the resumption of corrosion during sampling. Two sections of each nail were examined, one for metallographic observations and the other for microhardness measurements and SEM and EBSD analyses. The surface was prepared using standard metallographic techniques consisting of grinding with SiC papers to the P4000 grade (5 μm grain size), followed by polishing with non-aqueous diamond suspensions to 1 μm grain size. Finally, the specimens were etched with 3 vol% nitric acid in ethanol (3% Nital solution). For the EBSD experiments, the dedicated sections were polished again from P4000 to ¼ μm (DP-paste M, Struers, Copenhagen, Denmark) and finally with 0.03 μm colloidal silica (Suspension SPM colloidal silica 0.03 μm, Buehler, Whitby, ON, Canada) for several hours on a vibrating polisher using a Buehler VibroMet™ 2 (Whitby, ON, Canada).

The metallographic specimens were examined by polarized-light optical microscopy (OM) in reflected light mode using a Nikon Eclipse LV100POL (Nikon Instruments Inc., Melville, NY, USA). Phase microstructure and non-metallic inclusion composition were analyzed by environmental scanning electron microscopy (ESEM-EDX) using a FEI Quanta 200 FEG/ESEM (Philips, Amsterdam, The Netherlands) coupled to an EDAX Genesis EDX X-ray microanalysis system. Analytical SEM parameters such as 20 KV acceleration voltage were used. For observations and EDS, the electron beam current was 0.1 nA, and the water vapor pressure was 0.9 mbar. Iron phase identification and grain orientation maps were obtained by electron backscatter diffraction (EBSD) using an environmental scanning electron microscopy (Philips FEI Quanta 200FEG/ESEM, Philips, Amsterdam, The Netherlands) in environmental mode coupled with an EDAX/TSL OIM EBSD, so that the surfaces were not coated with a sputtered conductor film. For EBSD, the electron beam current was 5 nA, and the water vapor pressure was 0.3 mbar. OIM Analysis 5 software was used to collect the EBSD diffraction data and calculate the average grain size in each phase.

For phase identification, micro-Raman spectroscopy analysis was also carried out using a Jobin Yvon High-Resolution (HR LabRAM) (HORIBA, Kyoto, Japan) spectrometer coupled to a microscope (Olympus BX41, Olympus Corporation, Tokyo, Japan). A He-Ne laser at a wavelength of 632.8 nm was used for the analysis with a nominal power at the source of 0.6 mW to avoid heat-sensitive mineral phase transformations. The beam focused through a × 50 objective (long focus lens) had a diameter of ~3 μm. Spectra were recorded with a resolution of 0.2 cm⁻¹ and recollected using LabSpec 6 software (HORIBA Scientific, Longjumeau, France).

To evaluate the mechanical properties of each sample, Vickers microhardness measurements were taken using a Qness Micro Hardness Tester M variant (Q10). The Brinell hardness scale conversion was used because of the low values expected.

The elemental chemical composition of major elements of the metal area was determined using the Panalytical Axios mAX advanced, Sequential Fluorescence Sequence X-ray Wavelength (WDXRF), equipped with an automatic sample changer, a vacuum measurement system, a Rh tube, and three detectors (gas flow, scintillation, and sealing of Xe). Each sample was placed in the appropriate sample holder, with an irradiated area diameter of 10 mm, designed to allow the sample to be measured in a vacuum atmosphere.

The semi-quantitative software used is based on the measurement of a series of scans using different analyzer crystals, to determine the presence of different elements. The (semi)-quantitative assessment is based on the fundamental parameters method and the instrumental calibration/sensitivity determined by the software on the basis of a small set of standard samples associated with the software itself. In addition, the appropriate corrections were made for the solid sample preparation and for the diameter of the irradiated area.

The mineralogical composition of the corrosions was determined by means of powder X-ray diffraction (XRD) using a Philips X’Pert diffractometer (Malvern PANalytical, Almelo, The Netherlands) equipped with monochromatic Cu-kα1 X-radiation (λ = 1.5405 Å) operating at 40 kV and 20 mA. The data collection was performed on powder samples and on an area diameter of 10 mm by a continuous scan in the range of 5–90º 2theta, at an acquisition rate of 0.02° theta per second. The diffraction patterns were interpreted with X’Pert HighScore Plus 3.0 software (Malvern PANalytical, Almelo, The Netherlands), on the basis of the characteristic space of each mineral, reconstructing the mineral profiles of the compounds and comparing the experimental peaks with experimental patterns in the ICDD and ICSD diffraction databases.

All the metallographic examinations were carried out at the Laboratory of Engineering Sciences for the Environment of the University of La Rochelle (France). The XRD and XRF analyses were performed in the Materials and Surface Unit of the Advanced Research Facilities (SGIker) of the Basque Country University (UPV/EHU).

3. Results

3.1. Forua Samples

All the five nails from the Forua site showed an external appearance similar to that of the samples from the other sites studied. However, when they were cut longitudinally for metallographic examination, all the nails showed total degradation due to corrosion, and it was not possible to perform any metallographic analysis (Figure 2a and Figure 3). Due to the total degradation of the metal, alteration phases could only be identified by micro-Raman spectroscopy and X-ray diffraction. The dominant black compound on the outside was identified as magnetite (Fe₃O₄), and the dominant yellowish-brown compound on the inside corresponded to goethite (α-FeOOH) (Figure 3). The XRD analysis identified goethite and magnetite and small amounts of quartz, corroborating the micro-Raman results.

3.2. Aloria Samples

The metal in most of the Aloria nails showed a heterogeneous microstructure with a grain size decreasing from coarse at the head to fine at the tip, as in the Al-06 nail, or from coarse on the inside to fine on the outside, as in the Al-07 and Al-11 nails. The AL-03 sample showed a grain size distribution in the examined domains with small size variations (Figure 4). Microscopic observation showed that the Al-03 nail was composed of pearlite and ferrite (Figure 4a). Widmanstätten patterns were observed (Figure 4a). The Al-06 nail appeared composed of ferrite only (Figure 4b), whereas the Al-07 nail appeared composed of ferrite and minor cementite (Figure 4c). Finally, the Al-11 nail was found to be composed of ferrite surrounded by pearlite grains (Figure 4d).

SEM-EDX allowed the non-metallic inclusions to be examined. Most of the slag inclusions were elongated and were aligned with the longitudinal axis of the nails. The non-metallic inclusions appeared to be mainly composed of aggregates of two different phases: a dendritic FeOx-rich phase and a globular FeO·SiO₂ phase (Figure 5). Micro-Raman spectroscopy analyses identified the nature of the inclusions. The slags of samples Al-06, Al-07 and Al-11 were found to be composed of wüstite (FeO) and fayalite (Fe₂SiO₄) trapped in a glassy matrix. In contrast, the Al-03 inclusions appeared formed by a single dark phase consisting of a glassy matrix.

Vickers microhardness was measured on the longitudinal section of the nails, from the tip to the head. The Brinell microhardness values are summarized in

Table 1. Brinell microhardness test results.

Site	Sample	Average	Minimum	Maximum	Std Dev
Aloria villa	Al-03 (n = 37)	192.8	174	214	7.7
	Al-06 (n = 43)	79.4	66	92	6.8
	Al-07 (n = 26)	80.6	66	107	9.8
	Al-11 (n = 23)	99.5	82	124	11.6
Iuliobriga civitas	R-01 (n = 39)	128.5	114	152	9.3
	R-02 (n = 26)	95.3	84.7	120	10.8

. Al-06 and Al-07 showed lower microhardness values with an average of 80 HB, consistent with their dominant ferritic composition. The Al-11 nail showed an average hardness of 100 HB due to the predominance of ferrite together with minor amounts of pearlite. The nail Al-03 showed the highest average hardness, 193 HB, with a maximum of 214 HB, due to the predominance of pearlite and ferrite.

The EBSD study of the nails from Aloria revealed very complex microstructures (Figure 6). At a scale comparable to that of the optical microscopy micrographs, it was often difficult to process the data to obtain good-quality mappings clearly showing all the grains (Figure 6a–c). This artefact was due to a scale effect in the acquisition of the maps, as at a large scale, a size of one or a few pixels is not enough to highlight the details of a microstructure. Thus, acquired at a much lower scale (Figure 6a′–c′), new maps revealed that inside the grains, observable by optical microscopy, there was a substructure consisting of clusters of nanocells or grains of nanoscopic size, impossible to highlight by classic observation methods. Such submicronic structures are characteristic of a high level of deformation in grains. However, the substructure of Al-06 appeared different from that of the samples Al-07 and Al-11, although for each sample, maps were realized in areas corresponding to large grains (Figure 4). Al-07 and Al-11 showed a higher density of low-misorientation interfaces (orange grain boundaries on Figure 6b,c′), probably due to higher dislocation density further preventing carbon diffusion and grain restoration. Al-06 appeared composed of ferrite (Figure 4b), and because of a low carbon content, in spite of a probably intense hammering, grains could have been restored at a submicronic scale. So, the different patterns may be linked to the carbon content of the alloy and the way this element could diffuse following the blacksmith work (hammering, quenching). The microstructure of sample Al-03 was found to be different again from that of the other two samples. In sample Al-03 EBSD, the crystallographic orientation mappings for iron α (inverse pole figures) showed grains in random colors, meaning a polycrystalline microstructure without preferential orientation and no texture. In addition, Widmanstätten patterns were observed (Figure 6d). Whatever the microstructure patterns, the grains of all samples were not uniform and showed color gradients meaning misorientations lower than 5° (Figure 6a″–c′,d′), testifying of an intense hammering.

3.3. Iuliobriga Samples

The two nails from Iuliobriga differ from the Aloria ones because they are longer with a roughly square cross section, as they were used for different purposes. Both nails (R-01 and R-02) showed a corroded surface and an irregular shape of the head. Only nail R-01 showed slight oxidation inside the head, which did not affect the matrix. The R-01 nail head showed a heterogeneous microstructure with the grain size larger in the inner areas than on the surface. However, the shank showed a relatively more homogeneous grain size, with an overall larger grain size towards the surface (Figure 7a).

The microstructure appeared composed of ferrite and pearlite in the inside and mainly Widmanstätten ferrite–pearlite on the outside (Figure 7b). Nail R-02 was bent in the middle and the microstructural analysis was performed on the upper half. The nail displayed a more pronounced orientation and elongation of the microstructure in the areas where the shape of the nail underwent a change (Figure 7c). Nail R-02 showed a prevalent ferritic microstructure characterized by a non-uniform grain size and evident orientations defining a banded microstructure (Figure 7d,e).

The non-metallic inclusions were relatively few and aligned along the forging direction. Particularly, the inclusions were concentrated in the head of the nail. The slag inclusions were significantly more abundant in coarse grain size areas and roughly defined the boundary between layers (Figure 7c–e). The nature of inclusions in the Iuliobriga nails appeared similar to that in the Aloria samples. In both cases, the inclusions were composed of two phases, predominantly wüstite and less abundantly fayalite, as identified by micro-Raman spectroscopy analyses.

The microhardness measurements revealed notable differences between the two nails, with the pearlite-dominated R-01 nail averaging 130 HB, and the ferrite-dominated R-02 nail averaging 95 HB (Table 1). The two studied nails from Iuliobriga showed different metallographic characteristics, including composition and microstructure, despite their external and formal similarities. The bending of the R-02 nail can be attributed to the use of the nail and was favored by the ferritic microstructure and low microhardness.

The EBSD analysis of both nails revealed different microstructural patterns. Nail R-01 showed a microstructure similar to that of the Aloria Al-03 nail, consisting of a polycrystalline microstructure with no preferential orientation and no texture and Widmanstätten patterns (Figure 8a). The EBSD study of the R-02 nail revealed a more complex microstructure with characteristics similar to nails Al-07 and Al-11. The mapping of Figure 8b was realized in a large-grain-size area of the nail, since the map of Figure 8b′ was realized in a small-grain-size area (Figure 7). Here again, a substructure formed by clusters of nanocells or nanosized grains was revealed.

4. Discussion

The high degree of corrosion of the Forua nails affected completely the metallic matrix, making it impossible to obtain information about the iron manufacturing techniques. The complete absence of iron and the presence of a hole in the interior of all samples raised questions about the mechanisms involved in corrosion. A particularly aggressive burial context could have had a significant impact on the forms of corrosion of the Forua nails. The chemical composition of the analyzed nails (

Table 2. Chemical analysis of major elements in nails from Forua (F), Aloria (Al), and Iuliobriga (R) sites determined by X-ray fluorescence.

Sample	Si	Cl	Fe	Al	As	F	Mg	P	S
F	0.84	0.34	98.16	0.25			0.15	0.09	0.17
Al-03	0.91	2.20	95.40	0.14	0.19	0.95	0.07	0.14
Al-06			100
Al-07	0.24	0.65	99.11
Al-11	0.24	0.65	99.53
R-01			100
R-02	0.44		99.32

) revealed that only the Forua nails contained a significant amount of sulfur. This could be linked to a long-term burial in anoxic (and waterlogged) conditions favorable to the development of anaerobic sulfate-reducing bacteria (SRB). These microorganisms are responsible for microbiologically induced corrosion (MIC) leading to the corrosion of iron into iron sulfides (mackinawite FeS, greigite Fe₃S₄, pyrite FeS₂) [25,26]. These phases are mainly ferrous and very sensitive to oxygen. Goethite is one of their transformation products once exposed to air, as already observed on archaeological iron artefacts [27,28]. Thus, an external layer of magnetite could have formed during burial, constituting an oxygen barrier for the rest of the metal and therefore a favorable environment for the development of SRB. The metal could then have corroded into iron sulfides, this being reflected by the remaining sulfur in the F nail. Then, the transformation of iron sulfides into goethite could have occurred after excavation and during the storage of the nails, carried out without any specific conditions preventing air, while the magnetite layer remained stable.

The mineralogical and chemical composition and microstructural features of Aloria nails showed that the Al-03 nail differs significantly from other nails (Table 2, Figure 7a). The microstructural patterns of Aloria nails indicated a high degree of grain deformation due to internal stresses resulting from a combination of the manufacturing process, the amount of iron carbides, and the use of the object. The anisotropy in the grain size distribution between the head and the tip and from the outside to the inside suggests that the nails were manufactured by hot plastic deformation. The mineralogy of the nails is indicative of low-carbon steels with variable carbon contents, resulting in nails with different ferrite and pearlite contents (Figure 4). Between nails composed entirely of ferrite (Al-06) and nails composed almost entirely of pearlite (Al-03), other nails were observed, and each type of nail had a different mechanical behavior.

The remains of a forge at the rural villa of Aloria suggest that there were craftsmen specialized in the repair and manufacture of iron tools. In addition, the lack of an iron ore in the surrounding area and the iron ingots found at the site suggest the purchase of iron. The Al-3 sample, with a different chemical and mineralogical composition, may correspond to a product of non-domestic origin similar to that of the ingots found.

Both Iuliobriga nails showed a similar chemical composition but different microstructures. The silica content of sample R-02 was associated with a higher number of non-metallic inclusions in the head and along the shank (Table 2, Figure 8). The microstructural patterns indicated a high degree of deformation during the manufacturing process. The flow lines observed in the R-02 nail suggest that the head was probably formed by bending the shank, whereas in the R-01 nail, it was probably formed by compression in the shank [29]. The different microstructures and mineral composition of the two nails suggest different forging processes in the various workshops of the civitas of Iuliobriga, while the similar chemical composition suggests the use of the same iron ore from the surroundings. In fact, Iuliobriga was located to the south of several iron deposits in Cantabria [30]. Pliny the Elder (23–79 AD) documented these iron mining deposits and stated “Of all the metalliferous veins, the most abundant in Cantabria is iron” in Naturalis Historia bk. 34, 149 [31].

The EBSD maps revealed different patterns of microstructures, all characterized by a submicronic structure composed of nanocells or grains of nanoscopic size. This could have been an effect of the surface preparation before the analysis, but misorientations appeared sufficiently high to consider that what was observed was not an artefact. This type of patterns was never observed in archaeologic ferrous alloys before and may be characteristic of an intense forging work. The slight differences between these nanostructure patterns could be due to the different contents of carbon, diffusing more of less easily according to the densities of the dislocations induced by hammering and allowing the grains to be more or less well restored after cooling.

In the sites studied, nails with different compositional and microtextural characteristics were identified, suggesting a poor control of the nail manufacturing techniques. Similarly, nails with different characteristics were observed at the nearby site of Loiola, suggesting a lack of standardization in the production of iron tools during this Roman period in the north of the Iberian Peninsula [6].

5. Conclusions

Despite the small number of samples studied, some conclusions can be drawn about the method of iron product manufacturing in different archaeological sites in northern Spain. The combination of several analytical techniques allowed for the comparison of the manufacturing processes of Roman iron nails between rural (Aloria site) and urban forges (Forua and Iuliobriga sites). The chemical composition of the nails allowed for identifying two iron types of different composition and genesis. The aggressive burial conditions that affected the nails of Forua promoted corrosion until the metal disappeared completely. Conversely, the burial conditions at Aloria and Iuliobriga prevented the complete corrosion of the metal, allowing the metallographic study. The composition and microstructure of the Aloria nails suggest that the iron tools were mainly made and repaired in their own forges from trade iron ingots. The Iuliobriga nails studied suggest different workshops in the village with different levels of craftsmanship.