ORIGINAL PAPER

Spectroscopic evaluation of the environmental impacton black crusted modern mortars in urban–industrial areas

N. Prieto-Taboada & M. Maguregui &I. Martinez-Arkarazo & M. A. Olazabal & G. Arana &

J. M. Madariaga

Received: 30 July 2010 /Revised: 8 October 2010 /Accepted: 10 October 2010 /Published online: 30 October 2010# Springer-Verlag 2010

Abstract A multianalytical characterisation of blackcrusted modern construction materials from buildingslocated in the Bilbao Metropolitan area (North Spain) wascarried out. According to the mineral composition deter-mined by Raman spectroscopy, calcite and hematite werethe major compounds found while aragonite, limonite,rutile, quartz and some aluminosilicates such as obsidian oramazonite (KAlSi3O8) were also present in minor percen-tages. As deterioration products, gypsum and anhydritewere widely found not only in the surface but also in theinner part of strongly deteriorated samples. Coquimbite(Fe2(SO4)3·9H2O) was identified as well in the mostprotected facade where high amounts of Fe, havingprobably an anthropogenic origin, were measured by microX-ray fluorescence (μ-XRF). Zn was found to be in highamounts while Cu, Pb, Ti, Mn, Sr and K were identified asminor elements. Considering the non-expected concentra-tions found for some anthropogenic elements, a sequentialextraction was carried out in order to determine theirchemical form by means of ion chromatography andinductively coupled plasma mass spectrometry. Theorientation of the facades, which had a different influencefrom rain washing and industrial and traffic impact, wasshown to affect the accumulation of different compounds

in the black crust. Finally, the MEDUSA software wasused to simulate the reactions among the original com-pounds, deposited pollutants and the atmospheric acidgases in order to explain the presence of the decayingspecies found.

Keywords Acid gases . Black crust . Coquimbite . Modernmortars . Raman spectroscopy . Sequential extraction

Introduction

The Bilbao metropolitan area has supported, for more than100 years, an important economic and industrial activity.This region has the highest population density of the Northof Spain although most of the highly polluting industrialand human activities have been stopped or renewed.Nowadays, it is still the most industrialised area, especiallyimportant for the steel industries, chemical plants producingacids and metallic compounds, metal processing andindustrial port activities [1]. However, the high traffic isbecoming the main pollution source in the area [2].

Some of the buildings located in Metropolitan Bilbaoand the surrounding areas show substantial damage on theirfacade materials. Despite natural weathering parametersbeing the main factor accelerating stone decay processes[3], atmospheric contamination plays also an importantrole. The most relevant atmospheric pollutants affecting thestone decay are SOx and NOx gases that result in theformation of soluble salts such as sulphates [4–7] or nitrates[8] after the acidic attack of the corresponding acid aerosolsagainst the stone material.

In the particular case of carbonated stones, calciumsulphate (gypsum) is frequently formed as the final product

This paper was published in the special issue Analytical Chemistry forCultural Heritage with Guest Editors Rocco Mazzeo, Silvia Prati, andAldo Roda.

N. Prieto-Taboada (*) :M. Maguregui : I. Martinez-Arkarazo :M. A. Olazabal :G. Arana : J. M. MadariagaDepartment of Analytical Chemistry,University of the Basque Country (EHU/UPV),P.O. Box 644, 48080 Bilbao, Spaine-mail: nagore_prieto@ehu.es

Anal Bioanal Chem (2011) 399:2949–2959DOI 10.1007/s00216-010-4324-1

[9, 10]. However, particulate matter usually including sootand carbonaceous particles, where heavy metals aretrapped, as well as the marine aerosol typical from coastalareas, should be also considered in this case [11]. Thecombination of gypsum and particulate matter depositionsforms the harmful and unsightly black crust [12, 13].Therefore, the stone surface acts as a pollutant accumulatorand thus, the presence and the concentration level of certainchemicals such as heavy metals, persistent organic com-pounds or soluble salts should be taken into account [1, 14],especially in those facades not washed by the rain. Inaddition, the chemical form in which pollutants anddeterioration products are present determines their hazard-ousness; consequently, the chemical form (molecularspeciation) conditions the security measures to be taken inrestoration or demolition works, and the maximum contentsassumable for the construction materials to be reused [8,15]. In this sense, a sequential extraction is a good choice inorder to perform the speciation analyses since it providesinformation about the composition of the different fractionsof solid samples [16].

Different analytical techniques have been used in recentyears in the field of material characterisation in CulturalHeritage. The use of non-destructive analytical techniquesis preferred in this field, most of times combining elementaland molecular information. For instance, the combinationof X-ray fluorescence (XRF) and Raman spectroscopy isone successful combination [17–19]. Although such tech-niques are commonly used in a non-quantitative way, theyallow to determine the conservation state of the materialsunder study in a first stage.

For more exhaustive studies, methodologies based on theapplication of thermodynamic simulations together withchemometric tools to analyse quantitative data have beensatisfactorily used up to now in the characterisation ofconstruction materials [1, 19, 20].

In this work, the cited methodology is combined with asequential extraction protocol, in which the analytes arequantified by inductively coupled plasma/mass spectroscopy(ICP/MS) and ion chromatography (IC) to perform the wholespeciation analysis. This innovative speciation methodologyis more robust since it minimises the limitations of thedifferent individual analytical techniques and the uncertaintyin the identification of the species. The aim of this work is toapply the new methodology on the characterisation of modernmortars (first half of the twentieth century) of a buildinglocated in an urban–industrial area. Moreover, the damagemechanisms that have given rise to the degradation productsare proposed as well. Finally, the influence of the facadesorientation regarding the accumulation of pollutants anddegradation products, in both inside and outside the mortar,is also examined.

Experimental

Samples

Rendering and embellishing mortar samples were collectedfrom a house located in Leioa (Metropolitan Bilbao area,see Fig. 1). This building might be considered a modernconstruction and it has not been restored since it wasconstructed. The house is located in the middle of anindustrial area close to a road with strong industrial trafficin the past, but still significant in the present. Figure 1shows that the SW-oriented facade of the house suffers thedirect impact of the atmospheric pollutants from theindustrial area (mainly steel industries and chemical plantsproducing acids and metallic compounds) as well as fromthe road just in front of the SE-oriented facade.

Removing samples from the facade was not a realproblem in this case. Sampling was performed on the three

Fig. 1 Location of the buildingin the industrial area of Leioa(Metropolitan Bilbao) and viewof the SE-oriented facade

2950 N. Prieto-Taboada et al.

facades most protected from rain washing; that is, thoseoriented to the south (SW and SE) and northeast (NE),taken into account that the predominant winds, and thus,rain action in this area come from the NW. Renderingmortar samples belonging to the SW- and SE-orientedfacades and embellishing mortar (used as decorativeelement in the NE-oriented facade) samples were extractedfrom the building.

The samples had a similar size, approximately about35×35 mm, 12 g of weight and 6–8 mm thickness. Thesamples range the total depth of the rendering layer in orderto consider also the pollutant migration through the facade.All the samples showed a clear deterioration (detachment ofgrains) having a notable black crust, which covers thewhole building.

Instruments and methods

The Raman measurements were performed using an InViaRaman spectrometer (Renishaw, UK), coupled to a LeicaDMLM (UK) microscope; the system is provided with adiode laser at a 785-nm excitation wavelength and a Peltier-cooled charge-coupled device (CCD) detector. The equip-ment was daily calibrated with the 520.5 cm−1 silicon line.In order to avoid thermal decomposition of the samples, thelaser power (350 mW) was varied on 1%, 10%, and 100%depending on measurements. The spectra were obtainedwith a resolution of 1 cm−1 in the range 2200–100 cm−1,accumulating several scans from each spectrum to improvethe signal-to-noise ratio. The ×20 microscope lens was usedto perfect focusing of the laser beam’s spot (approximately10–200 μm) thanks to a TV microcamera. Data acquisitionwas carried out by the Wire 3.0 software package ofRenishaw and the analysis of the results undertaken byGRAMS 8.0 software (both from Thermo Fisher ScientificInc., Waltham, USA). The interpretation of the results wascarried out by comparison with reference spectra includedin e-VISART and e-VISARCH Raman spectral databases[21, 22].

The μ-XRF ArtTax model by Rontec (currently Bruker)was used to determine semiquantitatively the elementalcomposition of the depositions (external face of thesamples) as well as the original elements of the mortars(internal face). The equipment is composed of an X-raytube with a Mo anode working at 50 kV voltage and0.6 mA current. The X-rays were collimated by a 0.65-mmdiameter tantalum collimator. The equipment has a specialXflash detector (5 mm2) and the measuring head of theequipment implements a CCD camera that allows us tofocus on the sample by a motorised XYZ positioning unitcontrolled by the computer. The daily calibration of theequipment was made with a bronze reference standard(Bruker). He flow was not used during the analysis since

the lightest elements belong to the original composition ofthe mortar.

The MEDUSA software was used in order to simulatethe chemical processes that have given rise to thedegradation compounds identified by Raman spectroscopy[23], or proposed from the sequential extraction tests.

The sequential extraction protocol used in this work isbased on three sequential steps. First, soluble salts andexchangeable ions are determined by the ultrasound-assisted extraction with deionised water as alternative tothe Italian norm NORMAL 13/83 [24] for extraction ofsoluble salts from buildings materials [8]. The second stepconsists on the identification of the carbonate relatedspecies by the extraction with acetic acid 0.11 M [25] andthe third on a microwave digestion with aqua regia basedon the adaptation of EPA 3051A method for soils [26] toextract the oxidised material, linked organic matter andresidual fraction. The concentration of 18 elements extractedin each of the three steps of the protocol was determined bymeans of an Elan 9000 ICP-MS (PerkinElmer, Ontario,Canada), provided with a Ryton cross-flow nebulizer, aScott-type double pass spray chamber and standardnickel cones. Preparation of calibrates was carried outby using standard solutions from Alfa Aesar, Specpure®(Germany) and analysis were done inside a class 100clean room (Burdinola, Spain). Argon (99.999%, Praxair,Spain) was used as carrier gas in the ICP/MS measure-ments. A microwave-assisted acid digestion was carriedout prior to the analysis in the third step, in a Multiwave3000 (Anton Paar, Graz, Austria) system provided withan 8XF-100 microwave digestion rotor and 100 mLfluorocarbon polymer microwave vessels. The softwareused was ELAN 3.2.

The quantification of the anions and some cations fromthe salts present in the three extracts was carried out by aDionex ICS 2,500 ionic chromatograph with a suppressedconductivity detector ED50. An IonPac AS23 (4×250 mm)column and IonPac AG23 (4×50 mm) precolumn wereused for the separation of anions (fluoride, chloride,sulphate and nitrate). The quantification of cations (sodium,potassium and calcium) was conducted by using an IonPacCS12A (4×250 mm) column and IonPac CG-12A (4×50 mm) precolumn from Vertex; 5 mM Na2CO3/0.8 mMNaHCO3, 25 mA and 1 ml·min−1 flow were used as mobilephase, suppression current and flow, respectively, for theanalysis of the anions. In the case of cations, 20 mMCH4SO3 as mobile phase, 75 mA of suppression currentand 1 ml·min−1 flow were used.

The UnScrambler v9.2 software was used to perform thecorrelation analysis of the quantitative data from the waterextraction and from the acetic acid extraction. The cationsand anions considered in the calculation of correlationcoefficients were those having concentrations statistically

Black crusted modern mortars in urban–industrial areas 2951

different from zero in the samples from at least two of thethree facades.

Results and discussion

The objective of this analysis was to identify the degrada-tion compounds present in the materials sampled. However,it was also necessary to determine their original composi-tion in order to propose the processes (chemical reactions)that have given rise to the degradation products currentlypresent. For this purpose, the external and internal surfacesof the samples were analysed; the comparison of thecomposition found in both surfaces by non-destructiveanalytical techniques will allow to establish which com-pounds are original (presence in both sides with similarconcentrations) and which have impacted and migrated tothe inside part of the mortars (lower concentration inside).Therefore, the conservation state of the materials of thefacade can be evaluated.

Raman spectroscopy

The sampled mortar was a compact mass of fine grains,mixture of sandstone and cement where the external andinternal side were clearly differentiable due to the blackcrust covering the surface. The Raman results showedcalcite (CaCO3, Raman bands at 1085 (s), 712 (m) and 281(m) cm−1) as well as different iron (III) (oxy) hydroxidessuch as limonite (FeO(OH), 552, 394, 297 and 240 cm−1)or hematite (Fe2O3, Raman bands at 612, 410, 293 and226 cm−1) in a higher proportion, as the compounds with amajor presence; both iron species have been frequentlyfound in sandstones of the area which have a high ironcontent [27]. Aragonite with main bands at 205(m), 705(m)and 1085(s) cm−1 and titanium oxide in the rutile form werealso identified by Raman spectroscopy thanks to itscharacteristic bands at 608 and 447 cm−1. These com-pounds are referred to be common in this kind of samplesas well [27]. Finally, vitrified carbon (sharp bands at 1310and 1600 cm−1), quartz (SiO2, with main Raman band at464 cm−1) and a variety of silicates including obsidian(main Raman bands at 1645, 1518 and 1333 cm−1) andamazonite (KAlSi3O8, Raman bands at 1124 (broad), 814(m), 750 (m), 650 (broad), 512 (vs), 474 (s), 453 (m), 403(m), 370 (w), 355 (w), 329 (w), 283 (s) and 256 (m) cm−1)were also identified. The absence of small bioclasticmaterials and the presence of aluminium-silicates suggestthat the raw material for the preparation of the renderingmortars used in the building was sandstone from the nearbyquarries [27, 28].

With regard to the exposed surface of the mortars, onlythose corresponding to the SE facade (the one shown in

Fig. 1) revealed the presence of gypsum CaSO4·2H2Oclearly identified by the characteristic bands at 1132 (m),1008 (s), 673 (m), 619 (m), 492 (m) and 413 (m) cm−1.Occasionally, anhydrite (CaSO4, 1130 (m), 1017 (vs), 675(m), 627 (m), 609 (m), 500 (m) and 417 (m) cm−1) was alsofound as a result of the dehydration of gypsum in zonesmost exposed to the sun and wind which favoured the lossof hydration water molecules. In addition, gypsum was alsodetected in the internal parts of these mortars. Therefore,pollutants were able to reach deep through the entire sample(6–8 mm), probably due to the infiltration of sulphuric acidaerosols in the porous material. The Raman spectra ofgypsum and anhydrite determined in the mortar are shownin Fig. 2.

As already mentioned, gypsum frequently composes theblack crusts together with carbon particles that are theresponsible for its black colour. Amorphous carbon show-ing broad Raman bands centred at 1600 and 1310 cm−1 wasalso identified indeed in the external surface of the mortarsamples.

Moreover, coquimbite (Fe2(SO4)3·9H2O) was anothersulphate compound identified in mortar samples thanks toits characteristic bands at 1198 (m), 1092 (vs), 598 (m) and497 (m) cm−1. According to the literature, it is considered anatural compound which is usually formed by the evapo-ration of iron-rich acid water [29–31]. In fact, only onereference mentions the formation of coquimbite as adecaying compound in cultural heritage. In that case, thecoquimbite was proposed to be formed from the reactionbetween iron oxides (hematite particles) and the solublesulphate from the formation of calcium sulphate, after theattack of sulphuric aerosols (resulting from SOx gaseshydration) against calcium carbonate of the stone [32],following reactions (1) and (2). That mechanism agreedwith the Raman spectroscopy results shown in Fig. 3 wherethe joint presence of hematite, gypsum and coquimbite wasstated; the shoulder corresponds to the main gypsum band

Fig. 2 Raman spectra of gypsum (1) and anhydrite (2) determined inthe SE mortars

2952 N. Prieto-Taboada et al.

(1008 cm−1) and it was identified by deconvolutionperformed by the GRAMS 8.0 software. Thus, the sameset of reactions could explain the presence of coquimbite inthe rendering mortar samples analysed in this work.

H2SO4ðatmosphericÞ þ CaCO3 , CaSO4 � 2H2OðsÞ þ HCO�3 þ Hþ

ð1Þ

CaSO4 � 2H2OðsÞ þ Fe2O3ðsÞþ 2H2SO4ðatmÞþH2CO3 , Fe2ðSO4Þ3 � 9H2Oþ CaCO3

ð2ÞDue to the fact that the samples were taken after a rainy

period nitrates were not found in the surface at concen-trations high enough to produce a Raman signal. However,although nitrates are referenced to unlikely be present oridentified [33] and hence they have been only occasionallyfound as degradations products by Raman spectroscopy,they can achieve high concentration levels especially inporous materials [1, 8, 20], being one of the main reasonsfor their degradation.

X-ray fluorescence

The lightest elements were not determined since theybelong to the original composition of the mortar. Therefore,their identification was not considered essential in order todetermine the deterioration level of the materials. Theanalysis confirmed Ca and Fe as the major elements insamples. This fact agrees with the molecular character-isation, which revealed high amounts of calcite and ironoxides. The naturally occurring stones of the area areindeed iron-rich and carbonate-based sandstones.

According to the analysis performed on the exposedsurface of the mortars, Zn, Cu, Pb, Ti, Mn, Sr and K were

considered to have an anthropogenic source since theirproportion in the external surface was by far higher than inthe inner areas as shown in Fig. 4a. In fact, their occurrencein the internal parts of the samples was not significant formost of them. Figure 4a shows that the iron quantity in theexposed surface was around three times higher than in theinternal part, which indicates that there is an anthropogeniccontribution for this element. Iron mining as well as steeland shipyard industries had a great importance in the firsttwo thirds of the twentieth century in Bilbao surroundingsand some of these kinds of industries are still at work.Therefore, the accumulation of iron particles together withother related elements in the facades of the buildingsconstructed in that period seems to be quite logical.However, their presence can promote the formation ofsubsequent degradation products. In this sense, this input ofiron particles could maybe explain the formation of someiron species (like the referred coquimbite) that otherwisewould not be possible due to the stability of the originaliron oxide of samples.

In the case of Zn, Cu and Mn their higher quantity in theexposed surface can be related again with the industrialactivity. For instance, according to the Spanish Register ofEmission and Pollutant Sources (PRTR), high amounts ofZn, Cu and Mn are emitted as particulate matter to the

Fig. 4 a μ-XRF spectra of the internal (1) and external (2) side of amortar sample. b Comparison between a mortar of the SW facade(oriented to the metallurgical industries) and a mortar of the SE one

Fig. 3 Raman spectrum of coquimbite in a hematite and gypsummatrix (1), together with the Raman spectrum of coquimbite (2) andhematite (3)

Black crusted modern mortars in urban–industrial areas 2953

atmosphere by the factories located in the industrial areawhere the studied building is located. Actually, theseelements were found in higher quantities in the facadeoriented to the cited industries; that is, the one oriented tothe SW. Figure 4b compares mortar samples of the SW- andSE-oriented facades, where this trend is shown. However,and due to the location of the building, the contribution ofroad traffic in particular for lead and marine traffic shouldnot be ruled out. Even though the use of lead in the gasolineis forbidden since 2000, high amounts of lead have beenaccumulated for years in areas suffering from high trafficimpact.

The summary of the molecular and elemental compositionof the mortars is shown in Table 1.

Speciation analysis

In order to identify the chemical species present in thesamples, the sequential extraction described above wascarried out. Table 2 summarises the mean results obtainedby ion chromatography and ICP/MS for each of thefacades and steps of the protocol. Note that nitrate andchloride were not determined in the third step sincehydrochloric and nitric acid were used as the extractants.Results show that the major elements are Ca, Mg, Fe, Al,Zn, Mn, Ba and Pb. The anthropogenic-consideredelements, including the most hazardous ones, are extractedin the hardest conditions corresponding to the third step inall of the facades. However, their concentrations are higherin the SW facade than in the others, although it is the mostrain-washed of the three, probably because it suffers amore direct impact of the metallurgical industries. On thecontrary, soluble salt content is by far lower in that facadedue to the rain washing effect. For instance, the solublesulphate content in this facade is around 100 times lowerthan in the other two.

To define which cations and anions are grouped, acorrelation analysis was performed with the datacorresponding to the three facades; the concentrations weretransformed to milliequivalents·kg−1 units to perform thecorrelation analysis. Moreover, the carbonate concentrationwas theoretically determined using the electroneutralityprinciple since it was not possible to be determinedexperimentally due to the composition of the mobile phase.

Another purpose of the correlation analysis was to knowif the different oriented materials had the same behaviour interms of pollutants accumulation. For this purpose, the dataof the SE and NE facades were treated separately from thedata of the SW-oriented facade. According to the results,the difference among the facades oriented to the SE and NEshowed the same behaviour as the SW-oriented facaderegarding the salts and metals present in the solublefraction. Moreover, the three facades showed the samebehaviour for the fractions corresponding to the carbonaterelated phase (extractable by acetic acid) and the fractionextractable in aqua regia. The correlation matrixes for themost significant elements (only analytes showing concen-trations statistically different from zero in at least two of thefacades were considered) of the first and second steps, forthe whole set of data from the three facades, are shown inTables 3 and 4, respectively.

The presence of some soluble salts can be concludedfrom the correlations between cations and anions in thewater-soluble fraction (first step of the sequential extractionprotocol, see Table 3). For example, the high correlationparameter between iron and sulphate is an indication of thesolubilisation of coquimbite (Fe2(SO4)3·9H2O). Sulphateand calcium are also highly correlated in agreement withgypsum presence in the samples; gypsum is partiallysoluble in water. The species proposed from the correlationanalysis agree with the Raman identification of gypsum andcoquimbite.

The nitrate compounds, which according to the correla-tion analysis are probably present as Na, K, Ca, Mn, Feand/or Cu salts, could not be detected due to their lowRaman scattering and their low concentrations. Differentchlorides (especially Na, K, Ca and Mg) are surely presentin the samples due to impact of the marine aerosols formingsaline depositions, but unfortunately they do not produce aRaman signal.

In addition, the hydrogen carbonate species is positivelycorrelated with Al and Sr, but not significantly with Ca; thisrequires a clear explanation. The mortars cannot havehydrogen carbonate unless the original carbonate speciesare neutralised by acids, forming the respective solublehydrogen carbonate salts. The SOx acid attack on calciumcarbonate transforms the insoluble salt in gypsum, hydro-gen carbonate and proton according to reaction 1; that

Table 1 Summary of the elements found in the mortars by μ-XRF and the compounds identified by Raman spectroscopy

μ-XRF Embellishing mortar Rendering mortarCa, Fe, Sr, K, Ti, Mn, Cu, Zn, Pb Ca, Fe, Zn,Ti, Mn, Cu, Sr, Pb,

Raman Original Calcite, aragonite, hematite, quartz, aluminosilicates, rutile Calcite, aragonite, limonite, hematite, quartz, aluminosilicates

Deposition Carbon black Carbon black

Degradation Gypsum Gypsum, anhydrite, coquimbite

2954 N. Prieto-Taboada et al.

proton can react with another solid carbonate (of Na, Ca,Sr, etc.) promoting the soluble hydrogen carbonate andcation species. Since magnesium correlates negatively withhydrogen carbonate, magnesium carbonate was concludedto be absent in the raw materials. The considerable amountsof aluminium determined in this fraction are probably dueto halide depositions; actually, a high correlation with F−isobserved in Table 3.

In the second step, where the carbonates and relatedcompounds are extracted, calcium and iron are bothcorrelated with carbonate and sulphate. Barium is onlypositively correlated with the original considered compo-nents, that is, Ca, Fe and CO3

2−, so it should have a naturalsource, even though it seems to have an anthropogeniccontribution in the SW facade. Finally, the correlationvalues obtained for Zn, Mn and Sr indicate a commonsource (probably from the metallurgical industry).

Considering the strong extraction conditions in the thirdstep, most of the cations should have been as oxides andinsoluble hydroxides, except for barium and lead, which,considering the high sulphate concentrations found, couldhave been also present as barium and lead sulphates.

Barium and lead are associated with wheel wear and fuelcombustion, respectively [34].

Simulation of the processes to form the degradation products

Theoretical simulations by chemical modelling were carriedout to explain the formation of the degradation productsresulting from the interaction of the original components withthe environment, particularly with combustion and green-house gases. The MEDUSA software was used, althoughother software having similar characteristics like RUNSALThave also proved to be suitable for this purpose [8, 35–37].

The environmental conditions of the area are affected byboth road and marine traffic as well as by industry nearby.Combustion acid gases and greenhouse gases (mainly CO2,NO2 and SO2) are strongly present in the area as well asMetropolitan Bilbao surroundings, and the problems withPM10 exceeded the limits established by Spanish legislation(R.D. 1073/2002) [2].

In order to perform the chemical simulations, themaximum concentration values for those pollutants wereconsidered. The main acids interacting with the surfaces

Table 2 Mean results in mg·kg−1 obtained by ion chromatography and ICP/MS for each of the facades and steps of the protocol

SW facade NE facade SE facade

First Second Third First Second Third First Second Third

F− 134±10 – – 171±14 – – 289±19 – –

Cl− 594±46 152±12 – 3,119±248 255±17 – 1,392±73 282±15 –

NO3− 2,169±170 263±33 – 6,218±523 574±52 – 3,486±277 471±46 –

SO42− 624±23 5,185±285 4,741±261 71,856±1,405 3,975±345 3,071±273 53,033±2,372 4,220±421 3,049±64

Na 566±40 611±43 – 3,217±420 975±69 – 681±41 614±14 –

K 587±88 570±85 – 2,788±387 1,094±34 – – 215±16 –

Ca 6,876±687 129,138±11,914 – 39,608±2,703 109,960±7,620 – 37,002±1,419 122,184±6,632 –

Mg 226±27 1,581±32 7,819±391 240±32 1,690±188 3,937±369 109±12 2,878±117 3,393±225

Al 208±17 6.2±0.2 43,184±864 80±36 36±13 20,621±2,876 995±4 – 14,280±1,723

Mn 1.23±0.09 274±3 1,032±52 5.6±0.6 228±20 481±67 – 113±10 333±25

Fe 22±4 969±165 20,362±204 97±6 689±160 13,969±1,091 75±6 889±56 19,867±1,821

Cu 3.4±0.8 34±3 610±12 18±2 75±14 288±50 1.3±0.2 9±1 132±20

Zn 15.6±0.6 1,467±29 9,656±386 – 604±63 1,325±181 – 53±10 378±51

Sr 38.0±0.6 471±5 335±3 73±6 415±43 228±45 86±19 301±39 141±16

Ba 2.5±0.5 85.4±0.9 454±9 0.94±0.02 63±4 201±21 18±2 92±20 388±37

Pb – 0.35±0.03 216±4 – 1.5±0.3 381±18 – – 148±9

V 5.7±0.2 10.2±0.2 77±2 0.71±0.09 2.1±0.2 24±3 1.7±0.3 5.1±0.7 22±2

Cr 0.85±0.03 0.99±0.05 45.9±0.5 – 0.9±0.2 21±3 1.0±0.2 3.6±0.8 18±1

Co – 2.23±0.02 13.5±0.1 – 0.9±0.1 5.4±0.8 – 1.06±0.08 3.2±0.5

Ni – 3.60±0.04 59±2 – 2.0±0.2 24±1 – 4.9±0.3 20±2

As 0.80±0.02 1.55±0.02 17 ±3 0.9±0.1 5.0±0.7 17±2 0.4±0.2 1.1±0.3 34±3

Mo – 0.230±0.009 – – 0.12±0.01 – – 0.18±0.02 –

Cd – 3.7±0.2 14.0±0.6 – 0.53±0.07 – – 0.38±0.05 –

Sn – – 6.5±0.1 – – 4.7±0.1 – – 7.1±0.6

Sb 0.660±0.007 0.97±0.01 3.6±0.1 0.69±0.09 0.71±0.06 3±1 0.5±0.2 0.25±0.02 2.8±0.8

Black crusted modern mortars in urban–industrial areas 2955

Tab

le3

Correlatio

nvalues

amon

gthespeciesrelatedto

thewater-solub

lefractio

n(firststep

ofthesequ

entialprotocol)

F-

Cl−

NO3−

SO42−

Na

KCa

Mg

Al

Mn

Fe

Cu

Sr

HCO3−

F−

1

Cl−

−0.889

1

NO3−

0,03

10.72

71

SO42−

0.47

60.97

60.89

41

Na

−0.252

1.00

00.95

90.73

11

K−0

.474

0.99

90.86

50.54

80.97

11

Ca

0.63

10.711

0.79

50.98

30.59

20.38

41

Mg

−0.947

0.99

00.29

2−0

.168

0.55

00.73

3−0

.347

1

Al

0.93

6−0

.873

−0.323

0.13

6−0

.577

−0.754

0.31

7−0

.999

1

Mn

−0.434

0.99

70.88

70.58

50.98

10.99

90.42

50.70

1−0

.724

1

Fe

0.44

80.89

20.90

81

0.75

20.57

40.97

6−0

.136

0.10

50.611

1

Cu

−0.390

0.99

60.90

80.62

40.98

90.99

60.46

80.66

6−0

.690

0.99

90.64

81

Sr

0.85

2−0

.543

0.55

10.86

60.29

30,06

0.94

4−0

.637

0.61

20.10

20.85

00.15

01

HCO3−

0.98

3−0

.815

0.21

50.63

0−0

.069

−0.304

0.76

3−0

.871

0.85

5−0

.261

0.60

5−0

.214

0.93

41

Tab

le4

Correlatio

nvalues

amon

gthespeciesbelong

ingto

thefractio

nextractedby

acetic

acid

(secon

dstep

oftheprotocol)

NO3−

SO42−

Na

KCa

Mg

Al

Mn

Fe

Cu

Zn

Sr

Ba

Pb

CO32−

NO3−

1

SO4=

−0.990

1

Na

0.75

9−0

.662

1

K0.42

7−0

.299

0.91

31

Ca

−0.939

0.88

3−0

.936

−0.711

1

Mg

0.26

5−0

.396

−0.426

−0.758

0,08

1

Al

0.64

3−0

.531

0.98

70.96

7−0

.867

−0.568

1

Mn

−0.456

0.57

40.23

30.61

00.12

3−0

.979

0.38

81

Fe

−0.907

0.84

0−0

.963

−0.769

0.99

60.16

6−0

.906

0,04

1

Cu

0.45

3−0

.326

0.92

41

−0.731

−0.739

0.97

40.58

7−0

.787

1

Zn

−0.747

0.83

1−0

.133

0.28

20.47

3−0

.839

0,02

90.93

30.39

70.25

51

Sr

−0.498

0.61

30.18

60.57

10.17

1−0

.968

0.34

40.99

90,08

0.54

70.94

91

Ba

−0.582

0.46

4−0

.971

−0.984

0.82

60.63

0−0

.997

−0.458

0.87

1−0

.989

−0.106

−0.415

1

Pb

0.43

1−0

.303

0.91

51

−0.714

−0.755

0.96

80.60

6−0

.771

10.27

80.56

7−0

.985

1

CO32−

−0.919

0.85

6−0

.954

−0.749

0.99

80.13

7−0

.893

0,06

80.80

0−0

.768

0.42

40.116

0.85

6−0

.752

1

2956 N. Prieto-Taboada et al.

were H2CO3 and H2SO4 because of their involvement inthe air, despite the incipient increase of NO2 in theatmosphere as revealed the evidences of nitrate presencementioned above. These pollutants were introduced inMEDUSA together with the compounds belonging to theoriginal composition, determined by Raman spectroscopy;that is, calcite and hematite.

The results indicated that only calcite was susceptible tothe attack of the acid gases, due to the high stability of theoriginal iron oxide (III) which needed extreme acidconditions to react. Figure 5a shows that under thesulphuric acid concentration range, corresponding to theSO2 emissions recorded by the PRTR, the formation of thedihydrated calcium sulphate (gypsum) is favoured. Thetrend shown in this chemical equilibrium diagram agreeswith the Raman results, which revealed the massivepresence of gypsum in comparison to the non-hydratedcalcium sulphate (anhydrite) formed under lower environ-mental relative humidity conditions or as transformationproduct of gypsum under drier conditions. The meanrelative humidity of the region is around 75% [38],favouring the formation of gypsum, which has been foundto be the main compound in black crusts on differentsupports in previous works [1, 14, 39].

With regard to the coquimbite (Fe2(SO4)3·9H2O) forma-tion, Fig. 5a shows that this salt is formed after completionof gypsum formation by the reaction of sulphuric acid withcalcite of the support. The iron is surely that coming withthe particulate matter trapped in the surface, which is morereactive than the original hematite and showed concen-trations three times higher in the exposed surface than inthe inner parts according to μ-XRF results.

Finally, in an attempt to explain the formation of PbSO4

as the most probable lead species according to thesequential extraction results, the chemical simulation shownin Fig. 5b was carried out. The major emission of lead is inform of salts, both by industry as well as by combustion offuels [40]. Those compounds give rise to a fast reaction oflead with acid gases. Under the environmental conditions ofthe area, the simulation suggests that original lead salts,deposited on the building surface, become Pb(NO3)2 and asthe SO2 concentration increases it is transformed into thestable PbSO4, which can be accumulated on the stonesurface.

A similar process could be proposed for the formation ofBaSO4. Although this compound has not been determinedby Raman spectroscopy, barium is known to accumulate bya dry deposition as BaO, which can react with acid gasesgiving rise to the formation of the corresponding bariumsulphate salt [8, 41].

Conclusions

The presence of elements with anthropogenic origin anddecaying compounds identified even in the internal side ofthe building materials studied revealed the substantialdeterioration suffered in a short time period.

The high amounts of nitrates and sulphates determinedin the black crust concluded the high impact of the acidgases present in the atmosphere. In this sense, apart fromgypsum typically composing the black crusts (sometimeswith presence of anhydrite), evidences of coquimbiteformation were found by Raman spectroscopy in themortars belonging to the most protected facade. Besides,the correlation analysis of the results obtained for the water-soluble fraction as well as the thermodynamic simulation ofthe formation conditions corroborated the formation of thisiron sulphate. In the same way, although lead and bariumsulphates were not identified, their major presence in thefraction extracted in aqua regia together with the chemicalsimulation results indicated again the sulphate form as themost probable for barium and lead.

The high concentrations of sulphate and nitrate in thewhole body of the analysed materials are a clear indicationof the potential risk for recycling and reusing them asconstruction materials when such kinds of buildings are

Fig. 5 Chemical simulation by MEDUSA software of the impact ofsulphuric aerosols a on the studied mortar composed of calcite andhematite; b on lead salt depositions to form the corresponding leadsulphate

Black crusted modern mortars in urban–industrial areas 2957

demolished. Nowadays, there is no specific legislationregarding the concentration level assumable for thispurpose. However, in the case of soluble salts, only themaximum value of sulphate is legislated for differentapplications. For example, the maximum sulphate concen-tration value for road filling is 0.7% w/w [15]. Thus, takinginto account that sulphate concentration in NE and SE is7% and 5%, respectively, the mortars from these facadesshould not be reused for this application. Besides, consid-ering the soluble aluminium amounts found especially inthe mortars of the SW facade, the reuse of this materialafter demolition is not recommended as well in order toavoid aluminium-related problems.

The building facades could be considered as staticindicators of the pollution suffered by the area in the lastyears. Moreover, the orientation of the facade demonstratedto be a factor affecting the accumulation of pollutants.Actually, the SW facade, which is the one oriented to theindustrial area and most affected by the action of rain,shows the highest heavy metal concentrations but thelowest soluble salt amount.

On the whole, Raman spectroscopy and XRF datacoupled with information from sequential extractionprocedures and thermodynamical simulation is a suitablemethodology for a molecular speciation analysis in orderto evaluate and diagnose the deterioration level ofmodern buildings. This methodology allows to justify orcorroborate the presence of the decaying compounds andto identify the chemical forms of the elements in othercases. This identification is a challenge that should besolved, especially for heavy metals, in order to perform arisk assessment when the restoration or demolition of abuilding is raised.

Acknowledgments This work has been financially supported by theIMDICOGU project (ref.:BIA2008-06592) from the Spanish Ministryof Science and Innovation (MICINN). N. Prieto-Taboada and M.Maguregui gratefully acknowledge their grants from the MICINN andthe University of the Basque Country (UPV-EHU), respectively.Assistance from LASPEA services (SGIker UPV-EHU) from theUniversity of The Basque Country is gratefully acknowledged.

References

1. Martínez-Arkarazo I, Angulo M, Bartolomé L, Etxebarria N,Olazabal MA, Madariaga JM (2007) An integrated analyticalapproach to diagnose the conservation state of building materialsof a palace house in the metropolitan Bilbao (Basque Country,North of Spain). Anal Chim Acta 584:350–359

2. Labein Fundation (2006) Action Plan for air quality in the regionof the Bajo Nervion. Diagnosis of air pollution in the town ofGetxo. Basque Goverment, Basque Country

3. Colston BJ, Watt DS, Munro HL (2001) Environmentally-inducedstone decay: the cumulative effects of crystallization–hydration cycleson a Lincolnshire oopelsparite limestone. J Cult Herit 2:297–307

4. Massey SW (1999) The effects of ozone and NOx on thedeterioration of calcareous stone. Sci Total Environ 227:109–121

5. Charola A, Pühringer J, Steiger M (2007) Gypsum: a review of itsrole in the deterioration of building materials. Environ Geol52:339–352

6. Grossi C, Bonazza A, Brimblecombe P, Harris I, Sabbioni C(2008) Predicting twenty-first century recession of architecturallimestone in European cities. Environ Geol 56:455–461

7. Dotsika E, Psomiadis D, Poutoukis D, Raco B, Gamaletsos P(2009) Isotopic analysis for degradation diagnosis of calcitematrix in mortar. Anal Bioanal Chem 395:2227–2234

8. Maguregui M, Sarmiento A, Martínez-Arkarazo I, Angulo M,Castro K, Arana G, Etxebarria N, Madariaga JM (2008)Analytical diagnosis methodology to evaluate nitrate impact onhistorical building materials. Anal Bioanal Chem 391:1361–1370

9. Bakaoukas N, Kapolos J, Koliadima A, Karaiskakis G (2005)New gas chromatographic instrumentation for studying the actionof sulfur dioxide on marbles. J Chromatogr A 1087:169–176

10. Bai Y, Thompson GE, Martinez-Ramirez S, Brüeggerhoff S(2003) Mineralogical study of salt crusts formed on historicbuilding stones. Sci Total Environ 302:247–251

11. Chabas A, Jeannette D (2001) Weathering of marbles and granitesin marine environment: petrophysical properties and special roleof atmospheric salts. Environ Geol 40:359–368

12. Maravelaki-Kalaitzaki P (2005) Black crusts and patinas on Pentelicmarble from the Parthenon and Erechtheum (Acropolis, Athens):characterization and origin. Anal Chim Acta 532:187–198

13. Sabbioni C, Zappia G, Ghedini N, Gobbi G, Favoni O (1998)Black crusts on ancient mortars. Atmos Environ 32:215–223

14. Martínez-Arkarazo I, Sarmiento A, Maguregui M, Castro K,Madariaga JM (2010) Portable Raman monitoring of moderncleaning and consolidation operations of artworks on mineralsupports. Anal Bioanal Chem 397:2717–2725

15. Association Française de Normalisation (AFNOR) (1992) NFP11-300: Earthworks. Classification of materials for use in theconstruction of embankments and capping layers of roadinfrastructures

16. Zhang J, Liu J, Li C, Nie Y, Jin Y (2008) Comparison of thefixation of heavy metals in raw material, clinker and mortar usinga BCR sequential extraction procedure and NEN7341 test. CemConcr Res 38:675–680

17. Vandenabeele P, Garcia-Moreno R, Mathis F, Leterme K, VanElslande E, Hocquet F, Rakkaa S, Laboury D, Moens L, StrivayD, Hartwig M (2009) Multi-disciplinary investigation of the tombof Menna (TT69), Theban Necropolis, Egypt. Spectrochim Acta A73:546–552

18. Chaplin TD, Clark RJH, Martinón-Torres M (2010) A combinedRaman microscopy, XRF and SEM–EDX study of three valuableobjects—a large painted leather screen and two illuminated titlepages in 17th century books of ordinances of the WorshipfulCompany of Barbers, London. J Mol Struct 976:350–359

19. Castro K, Sarmiento A, Martínez-Arkarazo I, Madariaga JM,Fernández LA (2008) Green copper pigments biodegradation incultural heritage: from malachite to moolooite, thermodynamicmodeling, X-ray fluorescence, and Raman evidence. Anal Chem80:4103–4110

20. Sarmiento A, Maguregui M, Martínez-Arkarazo I, Angulo M,Castro K, Olazábal MA, Fernández LA, Rodríguez-Laso MD,Mujika AM, Gómez J, Madariaga JM (2008) Raman spectroscopyas a tool to diagnose the impacts of combustion and greenhouseacid gases on properties of Built Heritage. J Raman Spectrosc39:1042–1049

21. Castro K, Pérez-Alonso M, Rodríguez-Laso MD, Fernández LA,Madariaga JM (2005) On-line FT-Raman and dispersive Ramanspectra database of artists’ materials (e-VISART database). AnalBioanal Chem 382:248–258

2958 N. Prieto-Taboada et al.

22. Pérez-Alonso M, Castro K, Madariaga JM (2006) Vibrationalspectroscopic techniques for the analysis of artefacts withhistorical, artistic and archaeological value. Curr Anal Chem2:89–100

23. Puigdomenech I (2009) MEDUSA (Make Equilibrium DiagramsUsing Sophisticated Algorithms). 32 bit version

24. CNR-ICR (1983) Raccomandazioni Normal 13/83: Dosaggio deiSali solubili Normal 13/83 recommendation: Determination ofTotal Amount of Soluble Salts., Rome

25. Rauret G, López-Sánchez JF, Sahuquillo A, Rubio R, Davidson C,Ure A, Quevauviller P (1999) Improvement of the BCR three stepsequential extraction procedure prior to the certification of newsediment and soil reference materials. J Environ Monit 1:57–61

26. Carrero JA, Goienaga N, Barrutia O, Artetxe U, Arana G,Hernández A, Becerril JM, Madariaga JM (2010) In: Rauch S,Morrison GM, Monzón A (eds) Highway and urban environment.Springer, Netherlands

27. Ibañez-Gómez JA, Yusta I, García-Garmilla F, Cano M, Rodríguez-Maribona I, Beraza K, Garín S (2001) Chemical and mineralogicalstudy of restoration mortars applied to the Eocene sandstones ofGipuzkoa used for building constructions. Geogaceta 30:223–226

28. Herrero JM, Gil PP, García-Garmilla F (200) Characterization ofstone and mortars from the facades of Arriaga Theatre of Bilbao.Cuad Lab Xeol Laxe 25:419–421

29. Meixiang Z, Wei T (1987) Surface hydrothermal minerals andtheir distribution in the Tengchong geothermal area, China.Geothermics 16:181–195

30. Fernández-Remolar DC, Morris RV, Gruener JE, Amils R, KnollAH (2005) The Río Tinto Basin, Spain: Mineralogy, sedimentarygeobiology, and implications for interpretation of outcrop rocks atMeridiani Planum, Mars. Earth Planet Sci Lett 240:149–167

31. Buckby T, Black S, Coleman ML, Hodson ME (2003) Fe-sulphate-rich evaporative mineral precipitates from the Río Tinto,southwest Spain. Mineral Mag 67:263–278

32. Maguregui M, Knuutinen U, Castro K, Madariaga JM (2010)Raman spectroscopy as a tool to diagnose the impact andconservation state of Pompeian second and fourth style wallpaintings exposed to diverse environments (House of MarcusLucretius). J Raman Spectrosc. doi:10.1002/jrs.2671

33. Oms MT, Cerdà A, Cerdà V (2007) In: Nollet LML (ed)Handbook of water analysis, 2nd edn. CRC, United States

34. Zafra CA, Temprano J, Tejero JI (2007) Contamination by urbansuperficial runoff: accumulated heavy metals on a road surface.Ing Investig 27:4–10

35. Price C (2007) Predicting environmental conditions to minimisesalt damage at the Tower of London: a comparison of twoapproaches. Environ Geol 52:369–374

36. Klenz Larsen P (2007) The salt decay of medieval bricks at a vaultin Brarup Church, Denmark. Environ Geol 52:375–383

37. Price C (2000) Expert chemical model for determining theenvironmental conditions to prevent salt damage in porousmaterials. Archetype, London

38. http://www.ign.es/espmap/mapas_clima_bach/Mapa_clima_07.htm39. Pérez-Alonso M, Castro K, Martínez-Arkarazo I, Angulo M,

Olazabal MA, Madariaga JM (2004) Analysis of bulk andinorganic degradation products of stones, mortars and wallpaintings by portable Raman microprobe spectroscopy. AnalBioanal Chem 379:42–50

40. Corey G, Galvao LA (1989) Lead. Metepec, Mexico41. Pérez-Rodríguez J, Del Carmen Jimenez de Haro M, Maqueda C

(2004) Isolation and characterisation of barium sulphate and titaniumoxides in monument crusts. Anal Chim Acta 524:373–377

Black crusted modern mortars in urban–industrial areas 2959