22nd International Symposium on Plasma Chemistry July 5-10, 2015; Antwerp, Belgium Plasma chemical treatment for metal artefacts: conservation approach H. Grossmannová1 and F. Krcma2 1 Methodical Centre for Conservation, Technical Museum in Brno, Purkynova 105, CZ-61200 Brno, Czech Republic 2 Brno University of Technology, Faculty of Chemistry, Purkynova 118, CZ-61200 Brno, Czech Republic Abstract: Article describes the possibilities and conditions of the hydrogen plasma treatment application for the iron artefacts conservation from the conservation-restoration approach. Specific groups of samples were treated to describe the whole process – pure artificial minerals, model samples, real archaeological samples. Electron microscopy SEM-EDS and XRD techniques were used to understand the transformation of compounds. Keywords: low pressure plasma, metal restoration, conservation-restoration 1. Introduction - technology and conservation This work is dealing with the possibilities and conditions of the plasma treatment application for the iron artefacts conservation. Generally, it was proved that the application of hydrogen/argon plasma allows reduction of chlorinated products as well as oxides from the corrosion layers of archaeological objects [1]. Plasma chemical treatment has some advantages compare to the other methods, as abbreviation of the conservation procedures and relative regardful of the artefacts. The corrosion removal process is very complex and it is very important to understand well the treatment mechanisms and to find the optimal conditions for the whole procedure. This technology itself is well defined and verified [2 – 4], but still some important tasks are related to the conservation issues. For these reasons, we cooperated with conservator specialists, to understand fully the process details and its specific parameters. Our apparatus, usable for the large amount of the treated objects (kind of "mass conservation techniques" could be applied) or for the big archaeological object such as swords or hatchets, needs to be studied for the plasma homogeneity (i.e. homogeneity of the process, represented by the sample temperature). Next important factor of this technique is how to combine effectively plasma with the conservation pre-treatment and/or after treatment. An important part of the experiments is performed also to understand the changes in material composition and migration of corrosion activators (chlorides, sulphates). Fig. 1. Plasma reactor, model samples treated in plasma. The plasma is produced in cylindrical reactor (0.75 m3) by 13.56 MHz generator at operating pressure of 30 Pa, P-III-6-21 total gas flow of 400 ml/min (200 ml/min H 2 , 200 ml/min Ar). Reactor is equipped with gas inlets and mass flow controllers for hydrogen and argon and a pumping system. Whole technology is fully automated, supplied by Programmable Logic Controller unit allowing a digital recording of all treatment parameters. 2. Experimental Model iron samples or newly excavated originals used for all the presented experiments were always pre-dried by molecular sieve only, treated in plasma reactor and cooled down in inert atmosphere. Optical emission spectrometer (Ocean Optics HR4000) was used to monitor the process during some measurements. Optical fibre was focused axially to the central part of the reactor. The OH radicals are generated during the plasma chemical treatment and can be used for the process monitoring [5]. We also tried to understand the desalination process induced by plasma. As the original artefacts are mostly contaminated by chlorides and sulphates, is quite challenging to determine and analyse the decrease of chlorides in metal samples (solution is to use higher volume of the samples or to use model samples with defined chloride concentration). Despite this, migration of chloride ions is confirmed by XRD analysis. Sample temperature during the plasma treatment is one of the most important parameters, because it is necessary to avoid the transformation of the metal phases during the eutectoid transformation. Moreover, the higher temperature corresponds to the activity of the chemical species and it also helps to the migration of contaminants. The temperature strongly depends on the supplied power, its mode (continuous or pulsed) and on the sample position in the reactor. Generally, it would basically be taken into account as the maximum allowed process temperature of the eutectoid transformation for a particular metal - from the phase diagrams (Fe approx. 700 °C.). However, due to the object manufacturing technology (turbidity, carburising) specific phases are formed in the metal, which are very important for the authenticity of the object. This means it 1 is very difficult to determine one exact value of the maximum allowed temperature. Because of various impurities etc., the most sensitive material, which should be the part of the object (verified at least by X-rays analysis before the treatment) must be taken into account. Thus contemporary practically used maximal temperature for iron objects should not exceed about 150 °C, only. Fig. 2. Sample temperature during 2 h treatment. C – continuous, P – pulsed (duty cycle 50 %). Fig. 3. Sample temperature during 90 min treatment, 400 W, continuous. Coordinates present the sample position in plasma. The sample temperature contact monitoring system was developed and successfully tested. It allows measurement only at one spot, but simultaneously with plasma operation. The results give the information about typical temperature profile – growing up during about 2 hours of the process. The maximum temperature achieved under the different electrical parameters (RF discharge generation at 200–700 W in continuous and pulsed regime) was varied between 71 and 194 °C. The application of pulsed discharge regime could maintain the corrosion removal process at lower mean power and thus lower heating stress affected the treated samples. The use of pulsed regime helps to regulate the temperature but when the mean supplied power is similar (200 W/C versus 400 W/P) the achieved temperature is almost the same. The plasma reactor is 1.5 m long and outer electrodes surround the cylinder in the length 1 m 2 around its centre. Active plasma can be visually observed particularly in the area between the electrodes, but its intensity is not homogenous. Due to this fact, we tried to determine this non-homogeneity on the samples treatment determined via their temperature. The samples located in several positions. Point [-20,0] presented the openings of the reactor, point [0,0] – X (start of the electrodes), Y - centre of the reactor diameter, point [100,0] – X (end of the electrodes), Y - centre of the reactor diameter (see the immersed sketch in Fig. 3). The highest maximum temperature was achieved at the reactor centre as it was expected. Based on the results shown in Fig. 2, it can be concluded that only the part of the reactor between the cylindrical electrodes can be used as the working space for the objects treatment 3. Results and conclusions Treatment of pure minerals To verify the processes leading to the corrosion removal, the pure corrosion model samples were used. Pure lepidocrocite (γ-FeOOH) was formed in laboratory, and treated together with the whole real samples (pieces of the agricultural tools), to understand the process of the Cl- migration and mineral reduction. Lepidocrocite is an orange iron corrosion product that is close to tone of the mineral akaganeite. These two types of corrosion products are generally less stable compared to goethite (α-FeO(OH)) or magnetite Lepidocrocite creates a secondary (Fe 3 O 4 ). corrosion coating in the neutral water during conservation (desalination process). These types of corrosion products are commonly found on the surface at points or sockets, and they are the typical products of the so-called chloride corrosion. Its colour allows approximately distinguish it from the black-brown corrosion products oxides with hydrated forms Fe 2 O 3 .H 2 O and goethite. The plasma induced reduction of lepidocrocite and particularly of akaganeite, too, leads to the mechanical stabilization (upper layers of the treated object destabilized by plasma and are easily removable). Experimental observation showed that treated lepidocrocite (amount of 1 g) is visually transformed to goethite on the powder sample surface; original lepidocrocite orange products remained unchanged in the powder bulk. Significant amount of chlorine containing mineral akaganeite is formed. This supports the theory of chloride ions trapping (from other real samples) at the lattice of akaganeite [6]. Table. 1. Results of XRD analysis before and after plasma treatment of pure lepidocrocite. L – lepidocrocite, L h – lepidocrocite (homogenized L after treatment). Sample Treatment Lepidocrocite γ-FeO(OH) Pure L No 100% Akaganeite β-FeO(OH,Cl) - Goethite α-FeO(OH) - P-III-6-21 Pure L h 6h cycle 51.4 % 17.3 % 31.29 % Treatment of model and archaeological samples Two groups of samples were used for other series of experiments. At the first, the effect of the plasma treatment of artificial corrosion layer (without the core, chlorinated) was tested. Secondary, the effect of the plasma and separately of the temperature, only (at the same time profile as in plasma), was studied under 6 h treatment of model corrosion layers without any metallic core and corrosion on archaeological samples with metallic core. Archaeological samples were determined as parts of the sickles, excavated in BrnoŽebětín, track "U Újezda" extinct medieval village from 15th century. Layered artificial corrosion was powdered and analysed by XRD on a Bruker D8 Advance apparatus with Cu anode (λ Kα = 1.54184 Å) and variable divergence screens at ΘΘ Bragg-Brentano. The qualitative phase analysis was done to determine the phase (mineral) composition changes due to the plasma treatment. Simultaneously, part of the samples was analysed using a scanning electron microscope with an energy-dispersive micro analyser (SEM-EDX). Analytical work was carried out on electron microscope PHILIPS XL 30th. Surface micro analyses were performed on the analytical complex PHILIPS-EDAX. On the real samples, detailed microscopy of the chloride and sulphate nests was performed. Table. 2. Comparison of XRD analysis according treatment procedure. (MCL - model corrosion layer, AS archaeological sample, NT - no treatment, P - plasma, T temperature). MCL P AS NT AS P AS T 3.5 % - 1,5 % - - 3.7 % 4.2 % - - - 50.3 % 44.2 % 58.9 % 69.3 % 66.0 % 42.5 % 51.6 % 9.5 % 21.8 % 13.6 % - - 17.4 % 6.0 % 11.9 % - - 7.3 % 2.9 % 6.6 % - - 5.4 % - - MCL Lepidocrocite γ-FeO(OH) Akaganeite β-FeO(OH,Cl) Goethite α-FeO(OH) Magnetite Fe 3 O 4 Silica SiO 2 Albite NaAlSi 3 O 8 Microcline KAlSi 3 O 8 NT Analysing XRD data (see Table 2), it was confirmed that the influence of the plasma results in the transformation of corrosion products. The significant increase of magnetite concentration, both in model and real samples, was observed. For model samples, lepidocrocite and goethite are transformed in to magnetite. The results for original archaeological objects are affected by a high content of soil silicate P-III-6-21 minerals and thus it is difficult to precede the correct methodology of the sampling for the XRD analysis. Data are distorted by different percentage of soil minerals, but calculating the ratio of ferrous compounds without the soil minerals confirms the formation of magnetite also for these samples. Decreasing lepidocrocite concentration confirms the results obtained during the treatment of the pure mineral. To better understanding these chemical processes, new experimental series using samples of other pure minerals will be prepared. As it is shown on the Fig. 4, 5 and 6, SEM-EDX (TESCAN MIRA3) was experimentally used to understand the changes in morphology and in distribution of chemical elements. At these first experiments, we focused on the analyzes of chloride and sulphate nests on the surface of real archaeological samples. Macroscopically, nest is a hollow space in the corrosion layer filled with tiny crystals shining brownish coloration. Fig. 4. Left - untreated corr. nest, right - treated corr. nest. Fig. 5. Detailed SEM pictures (edge of the nest) with pointed areas for EDX microanalyzes, untreated nest. Comparison of treated and untreated corrosive nest visually confirm a morphological changes - elimination of flat oriented crystals. EDX 3 surface microanalyses (spots spectrum 3 - 6) indicate the presence of chlorides and sulfur (sulphate crystals are presented, not only chloride FeCl 3 crystals). However, the analysis compares two different corrosion nests with different representation stimulators corrosion, therefore no loss in weight of sulfur and chlorine is clearly proved. Morphological changes - the absence of similar looking nests on the corrosion of untreated subjects verify the changes in samples composition. Changes in surface morphology can significantly influence the analysis with respect to the disclosure of the surface of the electron beam, (less diffraction and reflection phenomena of characteristic X-rays at crystalline surface of corrosion products). [4] S. Veprek, Ch. Eckmann and J Th Elmer. Plasma Chemistry and Plasma Processing, 445, 4 (1998) [5] Z. Rašková, F. Krčma, M. Klíma and J. Kousal. Czechoslovak Journal of Physics, 927, 52 (2002) [6] I. Kotzamanidi, P.Vassiliou, Em. Sarris, A. Anastassiadis, L. Filippakis, S. E. Filippakis. AntiCorrosion Methods and Materials 256, 49 (2002) Fig. 6. Detailed SEM pictures (edge of the nest) with pointed areas for EDX microanalyzes, treated nest. In future experiments, we would like to focus on the effect of plasma applications on the sulphate concentration by analyzing corrosion by EDX before and after the treatment. An important research would be also performed on the artificially prepared pure corrosion minerals to understand well the effect of the transformation and migration process. 4. Acknowledgement The work was supported by the Ministry of Culture of the Czech Republic, project No. DF11P01OVV004. 5. References [1] J. Patscheider and S. Vepřek. Studies in Conservation, 31 ,29 (1986) [2] V. Sázavská, F. Krčma, T. Šimšová and N. Zemánek. Journal of Physics: Conference Series, 207, 1742 (2010) [3] K. Schmidt-Ott and V. Boissonnas. Studies in Conservation, 81, 47 (2002) 4 P-III-6-21
© Copyright 2026 Paperzz