Introduction

During the last century, the industrial process and the burning of coal in cities released increasing amounts of sulphur dioxide into the atmosphere, causing a severe pollution problem which became a subject of great concern in a variety of different scientific fields. Sulphur and nitrogen emissions are responsible for the formation of “acid rain” (or acid deposition), which influences climate change by modifying atmospheric and freshwater environments and damaging ecosystems and forests (Kellog et al. 1972; Tommervik et al. 1995; Seinfeld and Pandis 1998). Buildings, in general, and historical monuments, in particular, are also seriously affected by airborne pollution (Viles 1990; Török et al. 2011; Ghobadi and Momeni 2011).

Many researchers have focussed on gas emissions and have demonstrated that air pollution is a key factor in the decay of the materials used in the construction of our architectural heritage, which in many cases cause chromatic, chemical, biological and physical changes (Winkler 1966; Amoroso and Fassina 1983; Camuffo 1995; Antill and Viles 1999; Kontozova-Deutsch et al. 2011).

A large part of our architectural heritage is located in the centre of historic cities, areas in which there is a high concentration of motor vehicles and, in some cases, industries. Vehicles are the main source of aerosols enriched in C, S and N in the form of acids and, when they come into contact with construction materials (stone, brick, mortar, bronze, glass, etc.), start reacting on the surface (Fassina 1991; Rodríguez-Navarro and Sebastián 1996; Lefévre and Ausset 2002; Cultrone et al. 2008). The principal effect of this reaction is the formation and development of chemical weathering on the surface of the material, which enhances its decay and causes irreparable damage (Saiz-Jimenez et al. 2004).

The most frequently used construction and ornamental stones in historical monuments are carbonate based, because they were easily quarried and in abundant supply. They are however very sensitive to atmospheric pollution, especially sulphur dioxide. A great deal of research has been done on the effects of atmospheric pollution on old and new buildings made of limestones and/or calcarenites (Török 2002, 2008, Török et al. 2011; Luque et al. 2008a, b), dolostones (Olaru et al. 2010; López-Arce et al. 2008) and marbles (Feddema and Meierding 1987; Fassina et al. 2002; Lan et al. 2005). It has been demonstrated that the final effect of airborne SO2 on carbonate materials is the development of sulphated black crusts on the stone surface (Brimblecombe 2004) and the loss of material due to solubilisation (López-Arce et al. 2008; Olaru et al. 2010).

Numerous researchers have investigated the mechanisms involved in the reaction and oxidation of SO2 in calcitic carbonates (Böke et al. 2002; Gauri et al. 1973; Elfving et al. 1994) and, in many of them, it has been well established that calcium sulphite is the first stage in the process of calcium sulphate formation on the calcitic substrate. However, there are relatively few works that describe the initial chemical sulphation process on dolostones, and a full, detailed description of the reaction between SO2 and dolomitic carbonates has so far not been provided (Gauri et al. 1992; López-Arce et al. 2009; Kulshreshtha et al. 1989).

According to Böke et al. (1999), the product formed by the reaction of SO2 with calcareous stones under high relative humidity conditions is CaSO3·1/2H2O (calcium sulphite hemidrate) which, by oxidation of bisulphite ions, is transformed into CaSO4·2H2O (gypsum) as follows:

$$ \begin{gathered} {\text{CaCO}}_{ 3}\;\left( {{\text{SO}}_{ 2} /{\text{H}}_{ 2} {\text{O}}} \right) \, \to {\text{ CaSO}}_{ 3} \cdot 0. 5 {\text{H}}_{ 2} {\text{O}} \hfill \\ {\text{CaSO}}_{ 3} \cdot 1/ 2 {\text{H}}_{ 2} {\text{O }}\left( {{\text{O}}_{ 2} /{\text{H}}_{ 2} {\text{O}}} \right) \, \to {\text{ CaSO}}_{ 4} \cdot 2 {\text{H}}_{ 2} {\text{O}} \hfill \\ \end{gathered} $$
(1)

On the contrary, when the sulphation process detailed above occurs on dolostones, it results in the formation of gypsum and epsomite as follows (Tambe et al. 1991):

$$ {\text{CaMg}}\left( {{\text{CO}}_{ 3} } \right)_{ 2} + {\text{ 2SO}}_{ 2} + {\text{ 9H}}_{ 2} {\text{O }} + {\text{ O}}_{ 2} \to {\text{ CaSO}}_{ 4} \cdot 2 {\text{H}}_{ 2} {\text{O }} + {\text{ MgSO}}_{ 4} \cdot 7 {\text{H}}_{ 2} {\text{O }} + {\text{ 2CO}}_{ 2} $$
(2)

But, previous to the crystallisation of epsomite, an intermediate phase, i.e. magnesium sulphite, could have been developed. Magnesium sulphite and hydrates of magnesium sulphite are well-known salts in wet flue gas desulphurisation technology, and are used as reagents to absorb the SO2 generated by coal-fired industrial processes or metal works (e.g. in wood, pulp and paper production) (Hagisawa 1993).

To this purpose, Przepiórski et al. (2012) noticed that MgSO3 forms in porous carbon when SO2 interacts with basic oxide anions on the surface of MgO particles, according to the reaction:

$$ {\text{MgO }} + {\text{ SO}}_{ 2} \to {\text{MgSO}}_{ 3} $$
(3)

Moreover, it is well described that the reaction of SO2 from flue gases with an aqueous suspension of MgO results in the formation of MgSO3·6H2O or MgSO3·3H2O phases, depending on the prevailing conditions (Söhnel and Rieger 1994): below 40–42.5 °C magnesium sulphite hexahydrate is the stable equilibrium solid phase, while above this temperature the stable phase is the trihydrate (Pinaev 1964). However, due to the metastability of both phases, sulphur can react and form magnesium bisulphite, Mg(HSO3)2, making the desulphurisation process more difficult (Schöggl et al. 2006). Data on the solubility of magnesium sulphites are therefore of great importance, and previous researches have described how this solubility is directly related to the increase in magnesium sulphate content (MgSO4) (Nývlt 2001).

In addition, the easy oxidation of MgSO3 into MgSO4 is a well-known effect (Rowland and Abdulsattar 1978; Liu et al. 2010). Hence, the same sulphation process that occurs on calcitic marbles can occur on dolomitic marbles as follow:

$$ {\text{CaSO}}_{ 3} \cdot 1/ 2 {\text{H}}_{ 2} {\text{O }} + { 1}/ 2 {\text{O}}_{ 2} + { 3}/ 2 {\text{ H}}_{ 2} {\text{O }} \to {\text{ CaSO}}_{ 4} \cdot 2 {\text{H}}_{ 2} {\text{O,}} $$
(4)
$$ {\text{MgSO}}_{ 3} + { 1}/ 2 {\text{O}}_{ 2} + {\text{ 7H}}_{ 2} {\text{O }} \to {\text{ MgSO}}_{ 4} \cdot 7 {\text{H}}_{ 2} {\text{O}} . $$
(5)

Przepiórski et al. (2012) noticed that the available literature does not contain appropriate X-ray diffraction patterns to confirm the presence of MgSO3, while by means of X-ray photoelectron spectroscopy the presence of sulphur species (sulphite and y sulphate) was confirmed on the surface of porous carbon.

The aim of this paper is to identify the sulphated compounds that form on the surface of different types of marbles by using X-ray photoelectron spectroscopy. We also demonstrated that the mineralogical composition of marbles (calcitic and dolomitic) was the main factor that influenced the formation rate of sulphite and sulphate species. To this end, Malaga-Starzec et al. (2004) observed that the stability of sulphites on the surface of calcite was twice as high as on the surface of dolomite, which is why the sulphation process had a more serious effect on calcitic marbles than on dolomitics. Finally, we found out if marble grain size and grain boundaries were other variables that played a role in sulphate development.

X-ray photoelectron spectroscopy (XPS) is a technique for surface analysis that provides information about the elemental and chemical composition of the uppermost atomic layers. In combination with low energy ion bombardment, which is used for depth profiling, this technique can be used for compositional and chemical analysis at different depths (Briggs and Seah 1983). In this case, secondary effects such as ion bombardment can induce chemical and compositional changes. As a consequence, only indirect information is obtained regarding the chemical and compositional state of the material being studied. From a technological point of view, ion bombardment can also be used for the modification of chemical and physical properties of surfaces, thus producing changes in the solid surface that can be exploited for certain applications of stone (Torrisi 2008) or metal (Santamarina et al. 2007) and even in research into nanoparticles (Kelly 1989; González-Elipe et al. 1993).

Materials used

We began by selecting seven marbles commonly used in Spain as construction and ornamental stones. Three of them (White Macael, Tranco Macael and Yellow Macael) came from quarries in the Sierra de los Filabres (Almeria), two (Aroche and Fuenteheridos) from the Sierra de Aracena (Huelva), one (White Iberico) from Sierra Tejeda (Granada), and the last (White Mijas) from Sierra Blanca (Malaga).

From a mineralogical point of view, all these marbles belong to either the calcitic, White Macael (WM), Tranco Macael (TM), Aroche (AR) and Fuenteheridos (FH), or the dolomitic, Yellow Macael (YM), White Iberico (WI) and White Mijas (MI) marble groups. However, within these groups, each marble has its own distinctive textural variations (Luque et al. 2011).

Methodology

Samples characterisation and experimental design

Mineralogical characterisation and petrographic features of the seven marbles prior to the sulphation test were determined by means of polarised optical microscopy (OM, Olympus BM-2).

To get a quick chemical reaction on marble surface, accelerated sulphation test was performed in a Kesternich chamber [details of the experimental setup are reported in Luque et al. (2008a, b)], at constant atmospheric pressure (1 atm), 25 °C, 90 % RH and 400 ppm SO2 concentration for 24 h. A container full of water was introduced into the chamber to maintain high RH concentration. Marble samples were cut into two slabs of 10 × 10 × 0.3 cm and dried for 48 h at 50 °C before being placed in the chamber.

Colour variations

Before and after the sulphation test, colour measurements were carried out with a MINOLTA CR-210 colorimeter. Measurements were expressed using the CIE (Commission International de l’Eclairage) L* a* b* system (Billmeyer and Salzman 1981), where L* represents the lightness and a* and b* are the chromatic coordinates. The overall colour variation (ΔE) was evaluated using the following equation:

$$ \Updelta E \, = \, (\Updelta L^{*2} \, + \Updelta a^{*2} \, + \Updelta b^{*2} )^{ 1/ 2} . $$

Mineralogical determination

To identify the new phases developed on marbles surfaces, mineralogical analysis on unaltered and altered samples was performed by powder X-ray diffraction (XRD). We used a Philips PW-1710 diffractometer with graphite monochromator, automatic slit and CuKα radiation (λ = 1.5405 Å). Data were collected in step-scanning mode with 0.02º goniometer rate and 2θ from 3 to 60º. XRD goniometric calibration was performed using a silicon standard and the resulting data were interpreted using XPowder software (Martín 2004).

XPS analyses

In order to characterise the chemical phases developed on the surface of the seven marbles, X-ray photoelectron spectroscopy (XPS) analysis was performed; 4 keV Ar+ bombardment was used to enable chemical analyses to be performed at greater depth. XPS spectra were recorded using a Physical Electronics PHI 5701 spectrometer with a multichannel hemispherical electroanalyser (SCAI, University of Malaga). Non-monochromatic MgKα X-ray (300 W, 15 kV, 1253.6 eV) was used as the excitation source. The spectrometer energy scale was calibrated using Cu 2p3/2, Ag 3d5/2 and Au 4f7/2 photoelectron lines at 932.7, 368.3 and 84.0 eV, respectively. The binding energy of photoelectron peaks was referenced to C 1s core level for adventitious carbon at 284.8 eV. High-resolution spectra were recorded at a given take-off angle of 45º by a concentric hemispherical analyser operating in the constant pass energy mode at 29.35 eV and using a spot size 720 μm diameter aperture. The residual pressure in the analysis chamber was maintained below 1.33 × 10−7 Pa during the spectra acquisition. Marbles were mounted on a sample holder without adhesive tape and kept overnight under high vacuum in the preparation chamber before being transferred to the analysis chamber of the spectrometer for testing. Each spectral region was scanned in two areas with several sweeps until a good signal-to-noise ratio was observed. The PHI ACCESS ESCA-V8.0C software package was used for acquisition and data analysis. Recorded spectra were fitted using Gauss–Lorentz curves to determine the binding energy of the different element core levels more accurately (>90 %) (Briggs and Seah 1983). Atomic concentration percentages (A.C.  %) of the characteristic marble elements were determined from high-resolution spectra after the subtraction of a Shirley-type background, and taking into account the corresponding area sensitivity factor for every photoelectron line (Moulder et al. 1992). Survey and multiregion spectra were recorded of C 1s, O 1s, Ca 2p, S 2p and Mg 2p photoelectron peaks.

A depth profiling study was carried out by 4 keV Ar+ bombardment. The at-depth scale of 2.4 nm/min is assumed to be equivalent to the sputter rate of Ta2O5 under the same sputter conditions. Differences in sputtering yield between the sample being studied and Ta2O5 were not considered. Two depths were considered, after 2 min of Ar+ bombardment (which corresponds to ~4.8 nm depth), and after 19 min of Ar+ bombardment (~45.6 nm depth).

VPSEM observation

Visual observation of the marbles after the sulphation test was performed by means of a variable pressure scanning electron microscope (VPSEM) LEO 1430-VP, and the chemical composition of the crystals that developed on the surface was analysed with EDX microanalysis Inca 350 version 17 Oxford Instrument (CIC, University of Granada), which enables the identification of elements with low atomic numbers, including carbon.

Slab samples (10 × 10 × 3 cm) were dried for 48 h at 40 °C before being metalised with a mix of gold and platinum. The study was done in high vacuum at 20 kV and the images were acquired in a range from 10 to 20 kV.

Results and discussion

Samples characterisation

Table 1 summarises the petrological characteristics of the seven marbles.

Table 1 Petrological features of the studied marbles
Full size table

Starting with the calcitic marbles, White Macael (WM) is characterised by a white pearl colour and a saccharoid texture. However, depending on the quarrying level, it may show a marked grey band with varying numbers of veins. Tranco Macael (TM) is a white marble with a heterogeneous grey banding and a smaller crystal size than WM. Aroche (AR) is a heterogeneous marble with extreme variability of grain size. It is saccharoid white in colour with some green/grey veins. Fuenteheridos (FH) differs from AR mainly in its smaller grain size and marked heterogeneous greenish banding.

As for the dolomitic marbles, Yellow Macael (YM), as its name suggests, is yellow and is characterised by its small grain size; White Iberico (WI) is a white marble with a marked grey band and small grain size; and White Mijas (MI) is a translucent white marble with the largest grain size of the dolomitic marbles.

Colour variations

Colour parameters were determined before and after exposure of the marbles to SO2 and L* a* b* values were plotted on two diagrams: the main chromatic changes were observed in a 2D diagram, while the lightness variations were better checked on a 3D diagram (Figs. 1, 2).

Fig. 1
figure 1

Chromatic parameters (a* and b*) for unaltered and altered marbles (the arrow indicates the change from unaltered to altered samples). WM White Macael, TM Tranco Macael, AR Aroche, FH Fuenteheridos, YM Yellow Macael, WI White Iberico, MI White Mijas

Full size image
Fig. 2
figure 2

CIE L* a* b* parameters for unaltered (square) and altered (circle) marbles: chromaticity (a* and b*) versus lightness (L*). WM White Macael, TM Tranco Macael, AR Aroche, FH Fuenteheridos, YM Yellow Macael, WI White Iberico, MI White Mijas

Full size image

The YM sample appears separately from the other marbles (b* is approximately 18) because of its distinctive yellow colour. The other samples fall near the origin of the axes (between −2 and 1 a* values and 0 and 10 b* values). The lightness (L*) of all the marbles was very high, reaching values over 90.

After a weathering test (24 h of exposure to SO2), significant variations could be observed in all samples. Chromatic values for all the marbles had shifted approximately one unit towards –a*, while b* remained almost unchanged. As for L* values, all marbles showed a noticeable decrease (the average ΔL* decrease was ~17.3) (Fig. 2). The marble that suffered the highest chromatic changes was YM (3.76) and the highest lightness change was AR (19.07), while the smallest chromatic changes were for FH (1.65) and the smallest lightness change was in WM (13.91).

The colour change undergone by all the samples (∆E ≥ 14) denotes a significant alteration on the surface of all the marbles (Table 2), even if these changes are very small and are not visible to the naked eye.

Table 2 Overall colour change (∆E) in marbles after sulphation test
Full size table

Mineralogical determination

XRD analyses corroborated that there were no sulphate minerals on the surface of unaltered marbles. After 24 h of exposure to SO2, all samples showed the presence of sulphates (Table 3). More in detail, calcium sulphite hydrate (CaSO3·H2O) was the most common new phase and formed in all marbles. Lesser amounts of gypsum (CaSO4·2H2O) were identified in WM, WI and MI samples, while only on two dolomitic marbles surfaces, YM and WI, hexahydrite (MgSO4·6H2O) was found.

Table 3 Mineralogical phases detected on marbles surface by XRD after SO2 exposition
Full size table

XPS analyses

A first survey spectra recorded in each marble surface showed that the main chemical elements present on unaltered marble were C, O, Ca and Mg, while on altered marbles C, O, Ca, Mg and S were present.

The high-resolution spectra of selected peaks allowed us to determine their different contributions in terms of the atomic concentration percentage (A.C.  %) of C 1s, O 1s, Ca 2p 3/2, Mg 2p 3/2 and S 2p 3/2. From the binding energy and intensity of each photoelectron peak, the elemental identity and chemical state of these elements can be determined (Briggs and Grant 2003).

A detailed study of the photoelectron peak S 2p signal recorder in both marble groups can determine the contribution that identifies the different sulphur compounds developed on calcitic and dolomitic marbles surfaces after their exposure to SO2.

High-resolution spectra of selected peaks

From high-resolution spectra, we measured the atomic concentration percentage (A.C.  %) of selected peaks of C, O, Ca, Mg and S, detected on the surface of unaltered and at a depth of ~45.6 nm of altered marbles (Table 4).

Table 4 Atomic concentration of O, Ca, Mg, and S elements relative to C concentration (in  %) calculated for seven unaltered and altered marbles (on the surface and after 19 min of sputtering)
Full size table

The A.C. percentages of C, O, Ca and Mg values fit quite well with the mineralogical (calcitic and dolomitic) composition of unaltered marbles. In addition, the low S value measured in all samples (<1 %) confirmed that sulphur compounds were not present on marble surfaces (Table 4) after argon ion bombardment.

When altered calcitic and dolomitic marbles were analysed on the surface and at depth (after 19 min of argon ion bombardment), XPS analyses showed a different trend. Table 5 shows different chemical compounds with their binding energies determined by the chemical state of C, O, Ca, Mg and S.

Table 5 Binding energy (in eV) of O Oxygen, C Carbon, Ca Calcium, Mg Magnesium and S Sulphur peaks for different oxidation states of chemical compounds
Full size table
Calcitic marbles

Surface analysis of altered marbles showed that the values of S content (~16 %) measured in all samples was approximately three times higher than the C (as carbonate) content (~6 %), while Ca plus Mg values (~15 %) versus S content were constant. When analyses were carried out at depth (~45.6 nm), the S content (~9.5 %) decreased significantly with respect to C content (~4 %) and the sum of Ca and Mg content (~25 %) was now approximately twice the S content (Table 4).

Dolomitic marbles

Surface analysis showed that the values of the S content (~6 %) was approximately half the C content (14 %) and the sum of the Ca and Mg values (~19 %) was clearly twice that of the S content. At a greater depth (~45.6 nm), the S content fell by half (~3 %), C slightly decreased (~10 %) and the Ca plus Mg content (~28 %) increased noticeably up to 20 times higher than the S content (Table 4).

This analysis shows that the sulphur content (S %) measured in both marble groups is higher in all calcitic marbles than in dolomitics, regardless of their grain size and boundaries.

High-resolution spectra of S 2p on marbles

The interpretation (curve shape and binding energy) of high-resolution spectra of the photoelectron peak S 2p signal recorder in all marbles allowed us to identify the different contributions of sulphur compounds present on calcitic and dolomitic surfaces.

Deconvoluted spectra were adjusted with an average value of the peak width at half height (FWHM) at 1.75 eV value for MgKα radiation and the average Gauss peak-shape of 90.

After the subtraction of Shirley background, the data were fitted with model lines corresponding to two chemical species. Each chemical species was modelled with a spin–orbit doublet and the shift between S 2p 3/2 and S 2p 1/2 components was set on 1.18 eV and intensity ratio was 2:1 (Moulder et al. 1992).

Regarding the curve shape, we can mention that all S 2p spectra recorder on the surface of calcitic and dolomitic marbles were similar for each marble group even when these analyses were carried out at depth.

S 2p 3/2 spectra obtained on calcitic marbles showed their maximum binding energy at 166.7 eV and asymmetrical shape to higher values (~168.7 eV), suggesting that two contributions can be deconvoluted in this region. However, the same spectra collected in dolomitic marbles, more irregular and wider than that in calcitic marbles, suggests that more contributions can be deconvoluted in this region.

Regarding the S 2p 3/2 core level spectra obtained in seven marbles after 19 min (~45.6 nm depth) of Ar+ bombardment, another photoemission at a binding energy of 161.5 eV is attributed to reduced sulphur compounds.

Calcitic marbles

The S 2p spectra obtained on the surface and after 19 min (~45.6 nm depth) of Ar+ bombardment were deconvoluted into two contributions (Figs. 3, 4): CaSO3 (166.8 ± 0.1 eV) and CaSO4 (168.8 ± 0.2 eV). The measurement of CaSO3 (sulphite) and CaSO4 (sulphate) photoemissions from the photoelectron peak S 2p allowed us to determine their contents (in  %) from different layers (on surface and at depth).

Fig. 3
figure 3

S 2p core level spectra region (dashed lines) deconvoluted into each contribution (CaSO3 and CaSO4) on the surface of calcitic marbles. Binding energy (eV) versus intensity (a.u.)

Full size image
Fig. 4
figure 4

S 2p core level spectra region (dashed lines) deconvoluted into each contribution (CaSO3 and CaSO4) in-depth (~45.6 nm) of calcitic marbles. Binding energy (eV) versus intensity (a.u.). The component corresponding to reduced sulphur (dotted lines) is clearly evident after the ion bombardment

Full size image

In general, the values of sulphites obtained on calcitic marbles are always higher than values of sulphates on surface and at depth, although for both sulphur compounds values decrease with depth (Table 6). On the surface, CaSO3 contents (~83 %) are four times higher than the CaSO4 values (~17 %), while at depth a strong decrease in calcium sulphite content (~53 %) and a slight increase in calcium sulphate content (~34 %) can be observed.

Table 6 Calcium sulphite, calcium sulphate, and reduced sulphur content in calcitic marbles on the surface and after 19 min of sputtering (45.6 nm) Ar+ bombardment
Full size table

As one can see, low values of reduced sulphur contents were detected (binding energy at 161.1 eV) when calcitic marbles were analysed at depth (Table 6).

Dolomitic marbles

In these marbles, the S 2p spectra obtained on the surface and after 19 min (~45.6 nm depth) of Ar+ bombardment were deconvoluted into four contributions (Figs. 5, 6): CaSO3 (166.8 ± 0.1 eV), MgSO3 (167.4 ± 0.2 eV), CaSO4 (168.8 ± 0.2 eV) and MgSO4 (169.4 ± 0.3 eV).

Fig. 5
figure 5

S 2p core level spectra region (dashed lines) deconvoluted into each contribution (CaSO3, CaSO4, MgSO3, and MgSO4) on the surface of dolomitic marbles. Binding energy (eV) versus intensity (a.u.)

Full size image
Fig. 6
figure 6

S 2p core level spectra region (dashed lines) deconvoluted into each contribution (CaSO3, CaSO4, MgSO3, and MgSO4) at depth (~45.6 nm) for dolomitic marbles. Binding energy (eV) versus intensity (a.u.). The component corresponding to reduced sulphur (dotted lines) is clearly apparent after the ion bombardment

Full size image

With the exception of YM and similarly to calcitic marbles, values of sulphites (CaSO3 and MgSO3) detected on surface and at depth were higher than those of sulphates (CaSO4 and MgSO4). Nevertheless, some differences can be noticed when values of calcium (Ca-sulphite and sulphate) and magnesium species (Mg-sulphite and sulphate) are compared on surface and at depth (Table 7).

Table 7 Magnesium and calcium sulphate, magnesium and calcium sulphite and reduced sulphur content in dolomitic marbles on the surface and after 19 min sputtering (45.6 nm) Ar+ bombardment
Full size table

Surface analysis showed that calcium species concentrations (>46 %) measured in dolomitic marbles were slightly higher than magnesium species concentrations (<44 %). When analysis was carried out at depth (~45.6 nm), the calcium species concentrations (<39 %) decreased significantly with respect to magnesium species concentrations (>51 %) which even showed a slight increase.

Likewise to calcitic marbles, low values of reduced sulphur contents were detected (binding energy at 161.4 eV) when dolomitic marbles were analysed at depth (Table 6). However, considering that this phase was not detected on surface analysis, its presence can be attributed to sputtering processes. Moreover, considering that in dolomitic marbles, MgSO3 concentrations slightly increased in depth, reduced sulphur can only be due to a reduction of sulphites. Ion bombardment in metal sulphates has evidenced a partial susceptibility to damage of hydrated calcium sulphate from radiation, whereas CaSO4 proved to be more resistant (Coyle et al. 1981).

Reduction of oxides during ion bombardment is known and numerous research have related how the oxide ion is reduced preferably during analyses sputtering (Briggs and Seah 1983). According to Christie et al. (1981), the oxygen photoelectron peak having the lower binding energy may be assigned to the oxide formed as a result of ion bombardment, and the intensity of the peak is thus a measure of the extent of decomposition.

VPSEM observations

VPSEM images corroborate the sulphation process in marble surfaces after a very short period (24 h) of exposure to SO2. The morphology and the size of the new sulphate phases can also be identified. Calcitic marbles show the biggest crystal size of the sulphated products scattered on the surface of the sample (see White Macael and Tranco Macael, Fig. 7a and b), whereas dolomitic marbles have few crystallised areas but have a high concentration of small crystals of sulphated products (see White Iberico and White Mijas, Fig. 7c and d).

Fig. 7
figure 7

Microphotographs showing the development of different crystal shapes on the surfaces of four marbles (a White Macael, b Aroche, c White Iberico and d White Mijas) at the end of the sulphation test

Full size image

Finally, the main difference between the calcitic (WM and TM) and the dolomitic marbles (WI and MI) is in the morphology of the new crystals that develop. In calcitic marbles the crystals are like tabular aggregates, whereas in dolomitic marbles two different morphologies can be observed: rosette and tabular crystals in WI and a radiating cluster of needle-like crystals in MI.

Conclusions

In this work, we have demonstrated that the surfaces of calcitic and dolomitic marbles suffer chemical attack after just 24 h of exposure to SO2, The combined techniques used in this research have identified and confirmed early stages of sulphation on the surface of the marbles.

Air pollution is a decay factor that produces several colour changes in the stone surface, even with minimal exposure to air pollution. The chromatic parameters and, above all, the lightness measured before and after exposure to SO2 suggest that some chemical processes have occurred on the surface of the marbles.

Only CaSO3, CaSO4 and MgSO3 phases were detected by XRD, while XPS analyses showed great accuracy in the identification and quantification of calcium and magnesium sulphite and sulphate phases formed on the seven marbles. These phases were then observed under scanning electron microscopy (VPSEM) showing the different morphologies, sizes and the population density of the new minerals that developed.

We have noticed that calcitic marbles show higher rates of sulphation than dolomitic marbles, and also that in all marbles (calcitic and dolomitic), sulphite concentrations are always higher than sulphate concentrations. According to Malaga-Starzec et al. (2004), this is due to the stability of sulphite on calcite which is twice as high as on dolomite.

The sulphation process in calcitic marbles begins with the development of calcium sulphite which is then transformed by oxidation into calcium sulphate. As we have seen on the surface and at 45.6 nm depth, this reaction occurs early and at shallow depth. However, when sulphation process occurs in dolomitic marbles, although it begins with the development of calcium and magnesium sulphite, the subsequent transformations into calcium and magnesium sulphate are not clear.

Surface analysis by XPS shows that calcium sulphite and sulphate concentrations are higher than magnesium sulphite and sulphate. When XPS is carried out at depth, calcium sulphite and sulphate concentrations show a slight decrease compared to magnesium sulphite and sulphate. In addition to these phases, reduced sulphur was also detected after ion bombardment of calcitic and dolomitic marbles. Coyle et al. (1981) suggested that the development of reduced sulphur was due to the sensitivity of hydrated calcium sulphite to sputtering processes. Conversely, Christie et al. (1983) have not evidenced this phenomenon. Therefore, we consider that further analyses must be carried out to better explain the sensitivity of calcium sulphite to ion bombardment process.

In addition, although the sulphation processes affecting calcitic materials have been well described in literature, this is not the case for dolomitic materials, in which the development of magnesium sulphite has not been identified so far. Even if magnesium sulphite was not detected by X-ray diffraction, the excess Mg atomic concentration with respect to Ca atomic concentration (see Table 4) was not comparable to the rates of magnesium and calcium sulphate formation, which led us to think that part of this magnesium can be present as magnesium sulphite (Przepiórski et al. 2012).

The absence of magnesium sulphite in marbles damaged by exposure to SO2 may be caused by both, the higher solubility of MgSO3 in the presence of MgSO4 (Nývlt 2001) and the early oxidation of MgSO3 to MgSO4, thus preventing the formation of magnesium sulphite (Rowland and Abdulsattar 1978; Liu et al. 2010). The absence of sulphite may even explain why less magnesium sulphates (epsomite) develop on dolomite substrates than calcium sulphates (gypsum).

It is evident that the mineralogical composition of marbles is the main factor influencing the formation of sulphites and sulphates on the surface, while the grain size and grain boundaries do not have much influence on these reactions. We have observed that the sulphation trend is the same for all calcitic and dolomitic marbles, regardless of their particular fabric.

Our research confirms that XPS is a great tool to help us understand the different chemical processes that occur in stone surfaces after a very short period of exposure to air pollution, because it is able to detect the development of decay crusts measuring only a few micrometres.