SEARCH

SEARCH BY CITATION

Keywords:

  • archaeological iron;
  • chloride ions extraction;
  • conservation;
  • electrochemical methods;
  • SEM/EDS;
  • XRD

Abstract

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

The conservation of archaeological marine iron artefacts requires chloride ions removal. In this study, the removal of chloride ions was undertaken by two electrochemical methods: the electrolytic and the galvanic reduction in alkaline media. The results were compared with those obtained by the washing and the sulphite reduction methods, under identical conditions. The experiments were performed on samples coming from an 18th century cast iron cannon-ball, found in the archaeological context of a shipwreck, l'Océan, which sank near the southern Portuguese coast, in 1759. The extraction of chloride ions was monitored by ionic chromatography (IC). The results allow to conclude that the sulphite reduction experiments using the mixture 0.5 M NaOH/0.5 M Na2SO3 presents the higher efficiency in the first week, being further overcome by both electrochemical methods. After 40 days of treatment, the electrolytic reduction is the most efficient method.

1 Introduction

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

Archaeological objects from marine environments are necessarily subjected to degradation processes. The phenomenon is particularly relevant when the objects are removed from a wet medium with low contents of oxygen to a dry and oxygenated atmosphere [1-7].

Concerning to the conservation of iron marine archaeological objects, the removal of chloride ions is an important step to prevent or at least to reduce the rate of further corrosion, since chloride ions have been reported as one of the major factors which leads to localized corrosion, due to local acidification [3-10].

The first experiments for the treatment and conservation of archaeological iron artefacts, conducted by electrochemical techniques, are dated from the 19th century [11] and have been accurate during the last two decades [12-17].

The corrosion products on the surface of an archaeological artefact depend on different factors, particularly on the conjugation between its composition and the characteristics of the environment in which the object was kept. Depending on factors like the aeration, pH, temperature and time, different corrosion products can be formed (see North and MacLeod [2], Selwyn [3], North [18], Argo [19], among others).

Techniques addressed to the stabilization of metallic marine artefacts, such as the simply immersion in alkaline aqueous solutions and the so-called alkaline sulphite reduction method have been intensively tested and data reported in the literature [7, 20-28].

Electrochemical techniques, like cyclic and linear sweep voltammetry, have been recognized as powerful tools for the characterization of archaeological metallic objects [12-17].

This study deals with samples coming from an iron cannon-ball, belonging to the archaeological context of a shipwreck from 1759 [29-31], taken from a profundity of 4/5 m of the sea, at the South of Portugal (Lagos, Algarve). The ball was covered with a concretion with a thickness of 0.9 cm. The aim of the work consisted in the comparison of the efficiency of two classical techniques (simple immersion in non de-oxygenated alkaline medium and reduction in alkaline sulphite solution at 50 °C) and two electrochemical techniques [electrolytic reduction (ER) and galvanic reduction (GR) in alkaline medium], for the extraction of chloride ions from samples of the iron cannon-ball (genuine archaeological iron). The amount of chloride ions removed by the different methods was monitored by ionic chromatography, which is a less time consuming methodology when compared with the classical methods of titration [6, 21].

Samples from the iron cannon-ball have been characterized by cyclic voltammetry (CV), X-ray powder diffraction (XRD) and scanning electron microscopy coupled to X-ray microanalysis (SEM/EDS). It was concluded that the iron used for the fabrication of the cannon-ball was cast iron which is in agreement with the literature concerning to the production and applications of iron in the 18th century [32, 33].

The efficiency of each method, measured by the amount of chloride ions extracted per unit area in a certain period of time, was studied for different reduction methods: electrochemical techniques (ER and GR) and the sulphite reduction method (SR1 and SR2). These studies were performed using small samples of archaeological iron and have been compared with the data obtained from simple immersion (IM) in alkaline solution. The extraction rates, Vextr, measured over the six weeks of each treatment, have been analysed and compared.

2 Experimental

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

This work was conducted on small samples carefully cut from the cannon-ball. Samples included a portion from the nucleus and another from the surface of the ball.

For the electrochemical tests, the samples have been mounted using epoxy resin (Araldite® standard – CEYS®) and a metallic connection (insulated copper wire soldered with tin solder, Sn60), which were used as working electrodes. An image of such a sample is given in Fig. 1.

image

Figure 1. Photograph of the working electrode

Download figure to PowerPoint

After being constructed, these electrodes were carefully kept under wet conditions in seawater taken from the archaeological site. The surface of all the electrodes was photographed in a plan parallel to the objective of the camera, and further vectorized using the CorelDraw 12 software. The superficial areas were calculated using the ‘Get Area’ (IsoCalc.com) tool. Values between 2.0 and 3.6 cm2 were obtained. All results were normalized using the area of the corresponding working electrode.

The aqueous solutions were prepared with pure KOH from Pronalab ([Cl] < 0.01%); NaOH, QP, from Panreac ([Cl] < 0.01%) and Na2SO3 PA, from Riedel-de Haën ([Cl] < 0.005%). Salts were dissolved in ultrapure water from Millipore® Milli-Q and all the experiments (with the exception of the alkaline sulphite methods) were performed in non-deoxygenated solutions, at room temperature.

The open circuit potential (EOCP) measurements were performed using a digital multimeter, HP 34401A, with automatic acquisition, through Microsoft® Excel, using the tool ‘Excel IntuiLink for Multimeters Toolbar Addin’ (Agilent Technologies). All pH measurements have been performed with an HANNA® HI 1131 electrode coupled to a pH meter also from HANNA®, model HI 3222.

The electrochemical experiments for the removal of chloride ions were conducted in polyethylene bottles covered with its caps in which holes for the passage of the contacts of each electrode have been drilled.

The reference electrode used was a Ag/AgCl, KCl 3 M electrode with double junction from Metrohm® (E0 = +0.210 V vs. SHE). Before each experiment the electrode was maintained during 24 h in a KOH 1% aqueous solution and its potential checked against another Ag/AgCl (KCl 3 M) commercial electrode from Metrohm®.

The simple immersion method, also known as washing method, was performed by the introduction of the iron sample in the alkaline solution (KOH 1%). The evolution of the process was followed by measuring the OCP (EOCP) of the working electrode against a reference electrode.

The removal of chloride by the alkaline sulphite reduction method has been performed in two different solutions with the following compositions: (SR1) 0.1 M NaOH + 0.05 M Na2SO3, pH 13.3, and (SR2) 0.5 M NaOH + 0.5 M Na2SO3, pH 13.8. In this case, the experiments were carried out with the iron samples placed in bottles filled with 50 mL of the corresponding solution, sealed in order to avoid contact between the solution and the air and placed in an oven (Lenton WF 30) at 50 °C.

The removal of chloride ions by the ER method was performed by a three electrode cell, with Pt acting as counter electrode. A potentiostat AUTOLAB® PGSTAT10 (Eco Chemie B.V.) controlled the potential of −0.950 V against the reference electrode. The evolution of the chloride ions removal was followed by recording the current passing between the working and the secondary electrode during six successive weeks. According to the literature [13-15] the potential value of −0.950 V versus Ag/AgCl is the correct value for the chloride extraction from archaeological iron, avoiding the evolution of hydrogen, which could provoke damage in the graphitized layer, where is usually the information of archaeological interest.

For the extraction of chloride ions by the GR method a three electrode cell (a working electrode made with the iron sample and a zinc coupon of 8.8 cm2 (E0(Zn2+/Zn) = +0.763 V vs. SHE) acting as sacrificial anode giving cathodic protection to the iron (E0(Fe2+/Fe) = +0.440 V vs. SHE) was used. The cell was filled with 50 mL of KOH 1% non-deoxygenated solution and the evolution of the process was followed by recording the potential of the working electrode against the reference electrode. For all the methods tested, at least two replicas have been performed and very good reproducibility was obtained.

The quantification of chloride ions removed in each experiment was performed by ionic chromatography with a chromatograph Dionex® DX-500 equipped with an isocratic pump Dionex® IP20 and columns IonPac AG9-HC and IonPac AS9-HC, a suppressor of ions Dionex® ASRS-ULTRA II plus, and a conductivity detector Dionex® CD20. The data acquisition was performed by the Peaknet® software. The eluent used was a solution of Na2CO3 3.5 mM + NaHCO3 1 mM. Due to the sensitivity of the column, the solutions have been diluted (1:100) using ultrapure water.

SEM/EDS studies have been performed on samples from the nucleus of the ball and from the corrosion products taken from the surface (in powder form). This study was undertaken using a scanning electronic microscope (JEOL JSM 35C) coupled to a spectrometer (Noran Voyager).

3 Results and discussion

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

3.1 The archaeological site and the cannon-ball

At 24 of August 2009 a cannon-ball was recovered from the archaeological site of the shipwreck of l′Océan (Salema beach, Lagos, Portugal). The ball (Fig. 2) was raised from a profundity of about 5 m at a spot where the seabed is covered with similar cannon-balls. The artefact had a weight of 8.171 kg and a diameter of 13.2 cm. Measurements of the properties of the seawater collected at the archaeological site have given the following values: pH = 7.8 and dissolved O2 = 64.9%. The photograph of the ball shows a concretion including a zone of orange colour corrosion products observed on the left side. Grey and black corrosion products are also visible all over the ball surface. According to Argo [19], the orange products are most probably iron (II) hydroxide-oxide, namely, lepidocrocite (γ-FeOOH). The grey black and black colours may be attributed to FeO and Fe3O4, respectively (Selwyn [3] and Argo [19]), while, the white colour may be related with the formation of iron (II) chloride (FeCl2) (Selwyn [3], Turgoose [34], Askey et al. [35], among others). However, none of such phases could be identified by XRD.

image

Figure 2. The ball after being retrieved from the archaeological site (photo by José Paulo Ruas)

Download figure to PowerPoint

3.2 Nucleus and surface (cannon-ball) characterization

Figure 3 gives a representative cyclic voltammogram (CV) of a sample obtained from the core of the cannon-ball, in a KOH 1% aqueous solution, polarized between +0.65 and −1.5 V versus Ag/AgCl.

image

Figure 3. Cyclic voltammogram of a sample of the cannon-ball in KOH 1%. ν = 0.10 V/s; Ei = +0.65 V; Eλc = −1.5 V versus Ag/AgCl

Download figure to PowerPoint

The cyclic voltammogram shows during the cathodic scan and before the H2 evolution, a cathodic peak, A′, at −1.08 V (−0.87 V vs. SHE). Then, during the anodic scan a peak A is observed at −0.72 V (−0.51 V vs. SHE), followed by a decay in the current leading to a passive region between −0.25 and 0.50 V. At +0.50 V (+0.71 V vs. SHE) the oxidation of Fe(II) or even the Fe(III) compounds and/or the evolution of O2 may take place leading to the abrupt increase in the anodic current of the order of 45 µA/cm2, about ten times higher than the passivation current.

According to the Pourbaix diagram of the Fe-Cl-H2O, at 25 °C [36, 37], the potential of peak A and the pH value of 13 define a point located in the passive region. The passivity can be attributed to the Fe3O4 oxide species. However, as under anodic polarization kinetics should also be considered, peak A could be related with the oxidation of the Fe(II) into Fe(III) species and peak A′ (−0.87 V vs. SHE) probably due to the reduction of Fe (III) oxide/hydroxide to Fe (II) [38, 39].

Concerning to the analysis of the cannon-ball (core) Fig. 4 presents the EDS spectrum corresponding to the global analysis. From the analysis of the EDS spectrum it was concluded that the main constituent elements of the nucleus are Fe, C, P and Si. Apart from the global analysis, local regions have also been analysed by SEM/EDS, some micrographs being presented in Fig. 5. The zones where the EDS analysis has been performed are indicated (Z1 to Z5). The micro-constituents of the grey colour zone (Z1) were predominantly iron and carbon in amounts corresponding to cementite; in zone Z2, of white grey colour, a mixture α-ferrite and cementite could be identified; zone Z3 contains clearly a phosphorous mixture and zone Z4 inclusions of MgS. Finally, in zone Z5, the inclusions of black colour have been assigned to lamellar graphite.

image

Figure 4. EDS spectrum of a sample from the nucleus of the cannon-ball

Download figure to PowerPoint

image

Figure 5. Micrographs of a sample from the nucleus (zones Z1 to Z5)

Download figure to PowerPoint

The relatively high percentages of carbon (6–7% w) and Si (0.5–0.6% w), as well as the white grey colour have allowed to conclude that the nucleus was made of cast iron, must probably white cast iron. According to Mentovich et al. [33] the concentration of Si is the principal factor allowing to distinguish between white and grey cast iron. The relevant literature [3, 32, 33] reports also that the white cast iron consisting of a mixture of α-ferrite and cementite, is very hard and presents high resistance against cutting. These properties were effectively verified during the cutting of the various small samples representative of the cannon-ball. In addition, as stated by Scott [32], this type of alloy presents a structure including inclusions of lamellar graphite, as observed in zone Z5 of the studied samples.

Concerning the corrosion products on the surface of the cannon-ball, many factors may lead to transformations in the corrosion products even during its analysis (see Selwyn [3]). In the present work, the EDS analysis performed on powders removed from one sample of the cannon-ball allowed the identification of several elements as resumed in Table 1, where besides a global analysis, the results of the analysis performed in different points were also included (1–5). For a better comparison of the relative amount of these elements, C and O, presented in all cases, were excluded. Besides the elements previously identified in the nucleus, Cl and Na were also identified, as expected.

Table 1. EDS results (at%) of the corrosion products on the surface of the cannon-ball
AnalysisAlSiPSClCaFeNaMg
Global6.34.60.90.987.3
Point 16.03.23.51.70.378.94.71.7
Point 26.02.00.50.988.42.2
Point 311.012.20.61.670.13.90.7
Point 46.54.329.22.853.63.6
Point 50.68.99.91.31.272.94.21.1

3.3 Removal of chloride ions

3.3.1 Simple immersion in alkaline solutions

The simple immersion of iron samples in alkaline solutions such as of KOH 1%, allow the removal of chloride ions of the iron artefacts by exchange of Cl with hydroxide ions (OH) [3-10].

The experiment was followed by the OCP (EOCP) evolution and by the determination of the amount of chloride ions removed after each period during the six successive weeks of the washing process. These results are depicted in Fig. 6.

image

Figure 6. Evolution of the EOCP and total amount of chloride ions removed by simple immersion in KOH, during 6 successive weeks

Download figure to PowerPoint

During the first instants of immersion, the EOCP is displaced in the negative direction reaching the value of −0.60 V, but in few hours an inversion is observed and the potential is displaced in the anodic direction reaching a value of about −0.45 V, after about 12 days, and then decreases abruptly reaching the minimum of −0.475 V followed by another increase till a value of −0.37 V. At the end of the chemical treatment the EOCP presents a value of −0.42 V. According to the Pourbaix diagram for the system Fe-Cl-H2O [36] the observed behaviour may be related with the formation of a passive film of an oxide/hydroxide, Fe3O4 species, due to the contact of the iron sample with a high concentration of hydroxyl ions in solution. The initial decrease of EOCP can be attributed to the reaction between the corrosion layer and the hydroxyl ions in solution [14].

The total amount of chloride ions removed during the 40 days of treatment was 4.8 mg cm−2 (97 ppm/cm2 of the superficial area).

3.3.2 Reduction by the alkaline sulphite method

The removal of chloride ions by the alkaline sulphite solution is promoted by the conjugation between the decrease of O2 due to the action of sulphite ions and by the ion-exchange action of the alkaline solution [20-22, 24-26]. In the present work, these experiments were carried out under 50 °C, and the temperature certainly influences the kinetic's of the removal process. In this case and taking into account previous works [20-23] two different alkaline solutions have been used: 0.1 M NaOH + 0.05 M Na2SO3 (SR1) and 0.5 M NaOH + 0.5 M Na2SO3 (SR2). The experiments were also conducted over six successive weeks and the solutions changed every week. The results were followed by the determination of the EOCP, when the sample was taken out from the oven (weekly). The amount of chloride ions was determined in each solution (see data in Fig. 7).

image

Figure 7. EOCP evolution and total amount of chloride ions removed from an iron sample immersed in the following alkaline sulphite solutions: 0.1 M NaOH + 0.05 M Na2SO3 (SR1), pH = 13.3; 0.5 M NaOH + 0.5 M Na2SO3 (SR2), pH = 13.8

Download figure to PowerPoint

The OCP of the sample immersed in the solution with lower concentration of OH and inline image ions (SR1) has gradually changed from −0.70 to −0.95 V versus Ag/AgCl after approximately 3 weeks, being almost constant after that. In the case of SR2 experiment, the value of approximately −0.95 V was attained at the third week, and slightly changed until a value close to −0.85 V −0.92 in the last week. The maintenance of the open circuit voltage after a certain time can be explained considering that chemical reactions occur especially at the beginning of the experiment, and that after some time the chemical composition of the electrode is almost unchanged. Such behaviour has also been observed by other authors [3, 13, 26, 27]. The evolution of the open circuit voltage observed on SR1 and SR2 experiments are in accordance with the evolution of chloride ions extraction, as, an almost constant value is observed after 3 weeks in the case of SR1, while the value determined for SR2 abruptly increased after ∼20 days, and, although slowly, continued to increase until the end of the experiment (6 weeks).

The total amount of chloride ions removed during the six successive weeks was ∼5 mg/cm2 (101 ppm/cm2) for the weaker solution (SR1) and 15.6 (309 ppm/cm2) for the stronger alkaline solution (SR2).

3.3.3 Electrolytic reduction

The ER was performed in KOH 1%, at a constant potential of −0.950 V versus Ag/AgCl. In this case, the experiment was followed by recording the current transients and by the determination of the chloride ions removed. The results obtained along the experiment are presented in Fig. 8.

image

Figure 8. Current transients and total amount of chloride ions removed from the iron sample polarized at −0.950 V, in KOH 1%

Download figure to PowerPoint

The i versus t curve shows an abrupt decrease of the reduction current value during the first seven days of treatment. Then, a slight increase is observed, with oscillations ranging from −125 to 75 µA (62.5 and 37.5 µA/cm2). It should be emphasized that those oscillations in the negative direction are observed at the beginning of each period, just after immersing the electrode in the new fresh solution. Such behaviour may be explained by the fact that during the transference of the electrode from the old to the new solution, the reduction of the corrosion products (such as Fe(III) to Fe(II) compounds) may occur simultaneously with the removal of chloride ions. Effectively, a colour change from red-brown to black was observed on the sample during the ER.

At the end of the experiment (39 days) the total amount of chloride removed was 22.4 mg/cm2 (439 ppm/cm2). The reduction potential of −0.95 V versus Ag/AgCl (−0.74 V vs. SHE) applied to the sample (working electrode) certainly promotes the reduction of Fe(III) to Fe(II) species. North and Pearson [23] and Selwyn [3] stated that the ER promotes the reduction of the Fe(III) oxide/hydroxide leading to magnetite, a less dense compound, thus a more porous film on the surface of the iron sample.

3.3.4 Galvanic reduction

The potential between the sample from the iron cannon-ball (acting as cathode) and the zinc electrode (acting as anode), both in the same solution of KOH 1%, over the period of treatment is presented in Fig. 9. From the analysis of this figure, a decay in the potential in the cathodic direction can be observed during the first 5 days, being further maintained at an almost constant value of about −1.0 V. Considering the pH value of the solution, this potential leads to a point in the passivity region of the Pourbaix diagram of the Fe-Cl-H2O system at 25 °C [36]. Similarly to what has been observed in the ER method, a big displacement of the potential in the negative direction from −0.58 to −1.0 V is observed during the first week, and then, this value is kept as an average value. Slight decays at the beginning of the renewal of the solution, just when the sample is introduced in the new fresh solution, are visible. This behaviour may be associated with modifications in the corrosion layer and/or to changes in the chloride ions gradient since at the beginning of each treatment, the concentration of chloride ions in the fresh solution is almost zero.

image

Figure 9. Potential values (E) and total amount of chloride removed from the iron sample subjected to GR

Download figure to PowerPoint

After the end of the GR treatment (39 days), the total amount of chloride removed was 17.5 mg/cm2 (352 ppm/cm2).

3.3.5 Comparison of the efficiency of the various methods

Figure 10 presents the curves of the total amount of chloride ions removed during the different treatments, namely the two classical ones and the two electrochemical methods, after successive 6 weeks of immersion. Under identical conditions (∼40 days, 50 mL of alkaline aqueous solution, samples with areas ranging between 2.0 and 3.6 cm2) the efficiency of the four methods tested followed the order: ER 22.4 mg/cm2 (439 ppm/cm2) > GR 17.5 mg/cm2 (352 ppm/cm2) > reduction by the alkaline sulphite method, with the equimolar solution (SR2) 0.5 M NaOH + 0.5 M Na2SO3, 15.6 g/cm2 (309 ppm/cm2) > reduction by the alkaline sulphite solution (SR1) 0.1 M NaOH + 0.05 M Na2SO3, 5.1 mg/cm2 (101 ppm/cm2) > simple immersion in KOH 1% 4.8 mg/cm2 (97 ppm/cm2). When the simple immersion (IM) and the alkaline sulphite reduction (SR1) treatments were used, the total amounts of extracted chloride ions are significantly lower.

image

Figure 10. Evolution of the total amount of chloride removed by the different methods

Download figure to PowerPoint

Figure 11 represents the evolution of the rate of the chloride ions extraction (measured as μg per area and per day) for each week, for all the methods. Significant differences are observed on the extraction rates profiles (Vextr per week): while the sulphite reduction method defines a curve which indicates a diffusive model (higher values of rate extraction for the first week), both electrochemical methods ER and GR define a peak shaped curve with maximums at the third week of treatment. The simple immersion method in 1% KOH (IM) presents almost constant and quite low rates all over the 6 weeks.

image

Figure 11. Extraction rates (Vextr) of chloride ions removal for all the methods during the six successive periods. SR1: sulphide reduction (0.1 M NaOH + 0.05 M Na2SO3); SR2: sulphide reduction (0.5 M + 0.5 M Na2SO3); IM: immersion in KOH; ER: electrolytic reduction; GR: galvanic reduction

Download figure to PowerPoint

4 Conclusions

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

In this work, different methods were used to remove chloride ions from iron samples obtained from a cannon-ball, which was identified as being cast iron, as expected.

Under identical conditions (∼40 days, samples with areas ranging between 2.0 and 3.6 cm2) the efficiency of the tested methods followed the order: ER > GR > reduction by the alkaline sulphite method (SR2) > reduction by the alkaline sulphite method (SR1) ≈ simple immersion in KOH 1% (IM).

The higher efficiency of the reduction methods (ER, GR and SR2) for the removal of chloride ions from marine iron archaeological objects, during a 40 days period, when compared to the simple washing in alkaline solution (IM) and the less concentrated sulphite (SR1) has been well demonstrated in this study.

Both the simple immersion (IM) and the SR1 methods present very low efficiencies and very low rates of chloride ions extraction. It is also noteworthy that the extraction rate profiles (Vextr vs. time) for the sulphide reduction methods (both SR1 and SR2) define a curve which indicates processes most probably controlled by diffusion, while the electrochemical methods ER and GR both define a peak shaped curve with maximums extraction rates after the third week.

Acknowledgments

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References

The authors acknowledge the financial support from FCT (Fundação para a Ciência e Tecnologia) to CCMM (Centro de Ciências Moleculares e Materiais). The authors thank also to IGESPAR (Instituto de Gestão do Património Arquitectónico e Arqueológico, IP) for providing the samples of the cannon-ball and to SUBNAUTA, S.A. for the logistic support during the recovery of the ball.

References

  1. Top of page
  2. Abstract
  3. 1 Introduction
  4. 2 Experimental
  5. 3 Results and discussion
  6. 4 Conclusions
  7. Acknowledgments
  8. References
  • 1
    H. Knight, Conservation of Iron, Maritime Monographs and Reports No. 53, National Maritime Museum, London 1982.
  • 2
    N. A. North, I. D. MacLeod, in: C. Pearson (Ed.), Conservation of Marine Archaeological Objects, Butterworth, London 1987.
  • 3
    L. Selwyn, Proceedings of Metals 2004, National Museum of Australia, Camberra 2004.
  • 4
    W. A. Woddy, Int. J. Naut. Arch. Und. Exp. 1975, 42, 367.
  • 5
    W. S. Robinson, Int. J. Naut. Arch. Und. Exp. 1982, 11, 221.
  • 6
    Q. Wang, Stud. Conserv. 2007, 52, 125.
  • 7
    M. Rimmer, PhD Thesis, School of History and Archaeology, Cardiff, UK 2010.
  • 8
    A. A. Al-Zahrani, PhD Thesis, University of Wales, UK 1999.
  • 9
    D. Watkinson, in: A. Roy, P. Smith (Eds.), Copenhagen Congress, 26–30 August, International Institute for Conservation, London 1996.
  • 10
    D. Watkinson, A. Al-Zahrani, Conservator 2008, 31, 75.
  • 11
    K. M. R. Gilberg, J. Am. Inst. Conservat. 1987, 26, 105.
  • 12
    A. Doménech-Carbó, M. T. Doménech-Carbó, V. Costa, in: F. Scholz (Ed.), Archaeometry, Conservation and Restoration, Springer, Berlin 2009.
  • 13
    C. Degrigny, J. Solid State Electrochem. 2010, 14, 353.
  • 14
    C. Degrigny, A. F. Silva, P. M. Homem (Eds.), Faculdade de Letras da Universidade do Porto, University of Porto, Portugal 2008.
  • 15
    J. C. G. Coelho, MSc Thesis, University of Lisbon, Lisboa, Portugal 2009.
  • 16
    F. Dalard, Y. Gourbeyre, C. Degrigny, Stud. Conserv. 2002, 47, 117.
  • 17
    W. Carlin, D. Keith, J. Rodriguez, Stud. Conserv. 2001, 7, 68.
  • 18
    N. A. North, in: C. Pearson (Ed.), Conservation of Marine Archaeological Objects, Butterworths, London 1987, pp. 207252.
  • 19
    J. Argo, Stud. Conserv. 1981, 26, 42.
  • 20
    M. R. Gilberg, N. J. Seeley, Stud. Conserv. 1982, 27, 180.
  • 21
    S. Burshneva, N. Smirnova, in: G. Eggert, B. Schmutzler (Eds.), Archaeological Iron Conservation Colloquium 2010 – Extended Abstracts Session 3: Alkaline Chloride Extraction, State Academy of Art and Design, Extended Abstracts, 24th to 26th June Stuttgart, Germany 2010.
  • 22
    K. Schmidt-Ott, N. Oswald, Archaeological Metal Finds – From Excavation to Exhibition, Mannheim, Germany 2006.
  • 23
    N. A. North, C. Pearson, Stud. Conserv. 1978, 23, 174.
  • 24
    B. Scmutzler, B. N. Ebinger-Rist, Mater. Corros. 2008, 59, 248.
  • 25
    J. Liu, Y. Li, Mater. Corros. 2008, 59, 52.
  • 26
    M. Rimmer, D. Watkinson, Q. Wang, Stud. Conserv. 2012, 57, 29.
  • 27
    Q. Wang, Conservator 2008, 31, 67.
  • 28
    A. Rinuy, F. Schweizer, Conservation of Iron, Maritime Monographs and Reports No. 53, National Maritime Museum, London 1982.
  • 29
    J.-Y. Blot, M. L. Blot, Private Communication, 1982.
  • 30
    F. Alves, Abstracts of the 4th Congress of Algarve, Montechoro, Portugal 1986.
  • 31
    F. Alves, Arq. Port. 1987, 1990–1992, 455.
  • 32
    D. A. Scott, Metallography and Microstructure of Ancient and Historic Metals, J. Paul Getty Trust, Los Angeles 1991.
  • 33
    E. D. Mentovich, D. S. Schreiber, Y. Goren, H. Kahanov, H. Goren, D. Cvikel, D. Ashkenazi, J. Arch. Sci. 2010, 37, 2520.
  • 34
    S. Turgoose, Stud. Conserv. 1985, 27, 97.
  • 35
    A. Askey, S. B. Lyon, G. E. Thompson, J. B. Johnson, G. C. Wood, M. Cooke, P. Sage, Corros. Sci. 1993, 34, 233.
  • 36
    M. Pourbaix, Atlas D'équilibres Électrochimiques à 25°C, Gauthier-Villars & Cie, CEBELCOR, Paris 1963.
  • 37
    C. Rémazeilles, D. Neff, F. Kergourlay, E. Foy, E. Conforto, E. Guilminot, S. Reguer, Ph. Refait, Ph. Dillmann, Corros. Sci. 2009, 51, 2932.
  • 38
    C. M. Rangel, I. T. E. Fonseca, R. A. Leitão, Electrochim. Acta 1996, 31, 1659.
  • 39
    J. Urbabniak, J. M. Skowronski, B. Olejnik, J. Solid State Electrochem. 2010, 14, 1629.