Preliminary study on analysis and removal of wax from a Carrara marble statue

Abstract This preliminary study has mainly focused on the wax identification by nuclear magnetic resonance (NMR) and removal. Wax is used for many purposes in the field of art as protective coatings on wooden, stone or metal objects. From the comparison of the spectra H NMR and in particular with the correspondence of the resonance peaks of the samples taken from the statue and beeswax and paraffin, we can conclude that the wax applied on the statue surface is beeswax. From our data, it can be concluded that, to remove the beeswax, from any stone support, the more effective solvent is the mixture of cyclohexane/ethyl acetate. The removal percentages ranged from 19 to 99%. Lower percentages of removal have been observed in the case of yellow marble, probably because of its high porosity. We can affirm that, this solvent mixture can be employed in real art objects using cotton swabs to remove protective wax. Graphical abstract

Due to their irreversibility, these interventions must be conducted carefully and thus call for particular control and selectivity. Organic and inorganic compounds can be present on a stone surface as residues of biological activities or past treatments with natural materials, such as waxes, proteins or synthetic products (siloxanes, acrylic resins, etc.; Price 1996). Generally, cleaning operation refers to the removal of different materials (biological, organic and inorganic) due to environmental pollution, alteration or degradation processes. Furthermore, the limits between the different layers, often, are not clearly delineated and thus the extent and level of desired cleaning may be complicated and tricky.
Removing wax from stone artifacts is a problem that many conservators showed and tried to solve by trial and error, because there are insufficient data in the literature. For this reason, it is essential to optimise a method to speed up the wax removal and make it safe for the health of the operator and for the artifact itself. The identification of constituents (Menu 2006;Mannino & Orecchio 2011;Orecchio 2011;Orecchio 2013), contaminants and alteration products (Barreca et al. 2014) on the stone surface (Orecchio 2010) is usually part of the analytical research carried out in preparation for a conservation plan. Recent developments in the archaeometry and analytical chemistry are providing important contributions towards a better understanding of cultural heritages. The use of analytical techniques (Martınez-Arkarazo et al. 2007;Mannino et al. 2013) to the study of historical and cultural heritage has led to a considerable improvement in the knowledge of human civilizations.
Wax produced by animals and plants, mainly, are mixtures of waxy esters and contain different percentages of other compounds, for example fatty acids, fatty alcohols and hydrocarbons. Due to the complex composition of natural wax, their detailed analysis is not an easy task.
Although for 50 years nuclear magnetic resonance (NMR) has been a very powerful tool in many fields, its application to cultural heritage is rather recent (Stamatakis et al. 2010). One of the first researches on the use of NMR spectroscopy was in 1974, to identify a sample obtained from an archaeological find (Beck et al. 1974). A liquid contained in a flask found in Syria, dated sixth-fourth century BC, was investigated by 1 H NMR spectroscopy. This paper illustrates the capabilities of NMR in identification of organic materials on marble statue surfaces. The study has mainly focused on wax identification and removal. In this context, wax is used for many purposes in the field of art, especially employed as protective coatings on wooden, stone or metal objects. Wax is commonly employed on Italian monuments made with Carrara marble.
On average, the Carrara marble contains about 98% of calcium carbonate, about 1.8% of dolomite, less than 1% of silica, traces of magnesium oxide and strontium compounds.
One of the properties that a wax must have is the ability to penetrate the pore stone. This, in turn, requires a low viscosity and a low contact angle. Next, the wax needs to stiffen in order to couple with the stone. These requirements can be met by considering different properties: first, one could think of applying a substance that is liquid at high temperature and stiffens as it cools down. The second approach is to use a wax dissolved in a solvent. Unfortunately, due to these properties, the removal of wax from stone during restoration procedures, can be considered as the most important step. In this context, it is necessary to identify the best solvent able to remove wax from stone.
Thus, the aim of this preliminary study was to identify the best solvent to remove wax (beeswax, paraffin, etc.) from a stone surface.

Results and discussion
The colour of the film on the marble was from ochre to dark brown ( Figure 1) and prevented full appreciation of the sculpture. The waxy deposits on the statue appeared different: the first, homogeneously distributed over the entire statue ( Figure 1), probably, related to the practice of 'gànosis' , the last, due to the accidental dripping of votive candles and present mostly in the lower part of the handwork. Gànosis indicates the process used by Greek sculptors, at least from sec. V a. C., to produce on the surface of marble sculptures a golden patina or to prepare them for a polychrome decoration. This procedure treats the sculptures' surface with mixtures of oil and wax. The hypothesis of the restorers was that the first type was beeswax, while the second paraffin. To verify the hypothesis of restorers, we collected and analysed samples casually taken from different points of the statue and samples from its bottom part (casting of candles). Contrary to the hypothesis of the restorers, all samples taken from the statue (Figure 1) were made up of beeswax, which became apparent from the comparison of their spectra H NMR ( Figures S2 and S3) with that obtained from a sample of true beeswax that clearly shows the correspondence of the resonance peaks. In particular, in the NMR spectra (Table S1), it is interesting to note the presence of three different types of ester groups: primary (R-CH 2 O-COR, δ = 4.04 t), secondary (R-CH-O(COR)-R, δ = 4.80 m) and of ω−1 alcohols (CH 3 -CH-O(COR)-R, δ = 4.90 m), together with the signals of the corresponding α and β-methylenes (δ = 2.29 t and δ = 1.62 m, respectively). There is also a signal relating to the free carboxylic acid (δ = 2.38 t) and intense signals produced by the alkyl chains (δ = 1.25 m) and of terminal methyl groups (δ = 0.88 t).
A preliminary systematic study of the solvent extracted from test samples of wax was carried out. The experiments were carried out applying paraffin or beeswax on tests of stone with different characteristics and removing them using different solvents (Figures 2 and 3).
Generally, there are two ways to remove material from a stone surface: liquid cleaning or dry cleaning. In the first case, distilled water is used to clean and as a neutral rinse, white nATURAL PRoDUCT RESEARCH spirit is adapted for removing wax and oily dirt and for common grime, nicotine build-up and fingerprints can be cleaned using white spirit, or a mixture of distilled water with up to 2% non-ionic surfactant . Surfactants tend to leave a residue behind on the surface which can increase the rate of redeposit of dirt and may affect the surface in other ways in the long term. If liquid cleaning is not recommended, solid cleaning material (natural rubber sponges, etc.) can be used (www.vam.ac.uk/content/articles/c/cleaning-marble/).
In our case, we use white spirit or cyclohexane/ethyl acetate mixture. The percentages of beeswax and paraffin removed from different types of stone material and with the various solvents (or solvents mixtures) are shown respectively in Figures 4 and 5. From the experimental data, it can be preliminarily concluded that, to remove the beeswax, from any stone support, the more effective solvent is the mixture of cyclohexane/ethyl acetate. The removal percentages ranged from 19 to 99%. Lower percentages of removal have been observed in the case of yellow marble, probably because of its high porosity. While, as regards the paraffin, all solvents used, except the cases of pink Portugal marble (A) and green marble (B) treated with white spirit (a) or cyclohexane/ethyl acetate mixture (b), remove only a low percentage of surface coating.
The tests of removal of wax from marble were repeated at least three times. The standard deviations were between 5 and 25% and the average of all the measures was 14%.
In the light of the preliminary results obtained, we can propose a protocol to remove the wax from the surface of a real great marble artefact: to use solvents or mixtures mentioned above by dampened cotton swabs and begin in a small area (e.g. 4-5 sq cm at a time), renewing the swabs once soiled to avoid driving dirt into the marble instead of removing it.

Laboratory equipment
All glassware and sample containers were thoroughly washed with hot detergent solution followed by rinsing with Milli-Q water and acetone (analytical grade), respectively. These were finally kept in the oven at 85 °C overnight.

Chemical
Analytical reagent grade cyclohexane and ethyl acetate (Carlo erba, Milano) were used as solvents.
White spirit and turpentine were bought in a shop for restoration at Palermo (Italy). Deuterated chloroform was purchased from Aldrich, USA (225789).

Materials and sampling
Micro shallow film samples were collected, using a small scalpel, from a statue of Carrara marble depicting a Madonna with Child by Antonello Gagini dated 1503 (Figure 1). The statue is located at Palazzo Abatellis in Palermo. For subsequent analysis, the collected micro samples (few milligrams) were dissolved in deuterated chloroform, extracted by sonication and analysed by NMR spectrometry.

Paraffin wax
Paraffin wax was first produced commercially in 1867. Paraffin wax is a colourless or white product, somewhat translucent consisting of a mixture of solid straight-chain hydrocarbons ranging in melting point from about 48 to 66 °C.

Beeswax
Beeswax is a natural wax produced by bees of the genus Apis. Beeswax is formed from a mixture of several compounds: hydrocarbons, monoesters, diesters, triesters, hydroxy monoesters, hydroxy polyesters, acid esters, acid polyesters, free fatty acids, free fatty alcohols, etc. An estimated chemical formula for beeswax is C 15 H 31 COOC 30 H 61 . Beeswax can be classified generally into european and Oriental types. Beeswax has a relatively low melting point range of 62-64 °C. If beeswax is heated above 85 °C discoloration occurs.

Cleaning tests
The experiments were carried out, in triplicate, applying two different waxes (paraffin and beeswax) on 42 tests of stone with different characteristics and removing them using different solvents.
In detail the tests stone were:

Procedure
Marble simulated specimens, with surface dimension 3 cm × 3 cm × 0.7 cm, were carefully polished using Milli Q water and acetone. The clean samples were dried in an air oven at 105 °C to constant weight. After cooling in a desiccator containing silica gel, two different types of melted wax (microcrystalline paraffin wax and beeswax) were manually applied, by a brush, on the different specimens. By weighing, using the analytical balance (Sartorius, mod. entrius sens. 0.1 mg), the wax amount deposited on the surface of each sample test was determined. The samples were subjected to a initial ageing process by exposing to natural sunlight for at least 30 days (medium solar irradiance: > 28 MJ/m 2 ; temperature under sun exposure: about 60 °C; average relative humidity: 75-50%) in order to evaluate the first step of ageing. At this time, for the removal of the superficial wax, the specimens were left immersed in a beaker containing the solvent (or solvents mixtures) for 30 min. For each of the tests (stones and solvents), this operation was repeated three times. The amount of wax removed was evaluated by weighing (sens. 0.1 mg).

NMR analysis
Twenty milligrams of each sample were dissolved in 0.8 mL of CDCl 3 , sonicated and then filtered. 1 H NMR spectra were recorded on a Bruker Avance series 300 MHz spectrometer, using the residual solvent signal (δ 7.27 for CDCl 3 ) as reference.

Conclusion
From our preliminary data, the wax samples investigated gave very characteristic fingerprints, which make it easy to distinguish between the single materials without any doubt.
Removing wax from artistic stone handwork is a problem that many conservators have identified and tried to solve by trial and error, because there are no scientific publications on the subject. On the basis of our preliminary results, the removal of beeswax, from any stone support, can be obtained in a efficient way with the mixture of cyclohexane/ethyl acetate, whereas, the best solvent for removing paraffin has been shown to be white spirit.
In real artifacts (statue, large surfaces, etc.), the solvents or mixtures of them mentioned above can be used, from time to time in limited areas, by dampened cotton swabs and changing the swabs once soiled.

Supplementary data and research materials
The supplementary material for this paper is available online at http:// dx.doi.10.1080/14786419.2015.1113411.

Disclosure statement
No potential conflict of interest was reported by the authors.