Construction and Building Materials 95 (2015) 355–365

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: www.elsevier .com/locate /conbui ldmat

Microtextural and microstructural influence on the changes of physicaland mechanical proprieties related to salts crystallization weatheringin natural building stones. The example of Sabucina stone (Sicily)

http://dx.doi.org/10.1016/j.conbuildmat.2015.07.1310950-0618/� 2015 Elsevier Ltd. All rights reserved.

⇑ Corresponding author at: Department of Biological Geological and Environ-mental Science, Corso Italia 57, 95129 Catania, Italy.

E-mail address: pmazzol@unict.it (P. Mazzoleni).

G. Barone, P. Mazzoleni ⇑, G. Pappalardo, S. RaneriUniversity of Catania, Department of Biological, Geological and Environmental Sciences, C.so Italia, 57, Catania, Italy

h i g h l i g h t s

� A complete characterization of a calcarenite from Sabucina (Sicily) has been provided.� Changes in microtextural and microstructural due to salt weathering have been studied.� Variations of physical-mechanical parameters due to salt weathering have been studied.� Correlation between degradation mechanism and engineering parameters has been obtained.� The obtained results are useful in planning conservative action on buildings.

a r t i c l e i n f o

Article history:Received 9 March 2015Received in revised form 22 June 2015Accepted 14 July 2015

Keywords:Sabucina calcareniteMinero-petrographic analysisSalts crystallizationUltrasonic velocityPhysical-mechanical proprietiesPorosimetry

a b s t r a c t

The main aim of this research is to highlight the effects of salts crystallization on physical and mechanicalproprieties of natural building stones in consideration of petrographic and porosimetric features. In orderto achieve the goals of this work, numerous standard tests have been performed on both fresh and weath-ered samples of a Sicilian calcarenite named Sabucina stone, widely used as building and replace stonesin local Cultural Heritages. The obtained results allow to establish interesting correlation between theintensity of degradation processes and the studied parameters; in fact, a modification of the mechanicalbehavior has been observed in concomitance with a change in the degradation mechanism occurred inthe pore network, highlighting the influence of microtextural and microstructural modifications due tothe salt crystallization in the engineering properties of the studied rock.

� 2015 Elsevier Ltd. All rights reserved.

1. Introduction

Even if a large bibliography exists about durability esteem pro-cedures for building stones [1–7], really a few examples of inte-grate study on both microtextural, microstructural andphysical-mechanical proprieties changes due to weathering areavailable [8–9].

As it is well known, physical and mechanical proprieties ofrocks represent fundamental knowledge in engineering applica-tions and are strong influenced by petrographic features such astexture and structure [10–13]. For aforementioned, the knowledgeof the relationship among these features allow to better under-stand the behavior of the stones, especially in the framework of

restoration works. In particular, during restoration the choice of areplace stone is strictly related to proprieties as availability, aes-thetical compatibility, mechanical resistance and durability.

The Sabucina stone is a yellowish coarse grained calcarenitequarried since ancient times in the area of Caltanissetta (Sicily)and employed in several important Sicilian monuments and sculp-tures [14] (Fig. 1). Furthermore, due to its excellent durability andaesthetic features, it is widely used as a replacement stone, espe-cially in Baroque monuments of the Noto Valley (belonging tothe UNESCO Heritage List). An example of this application is repre-sented by the restoration works carried out on the S. NicolòCathedral in Noto (Sicily) (UNESCO World Heritage Site), in whichthe Sabucina Stone has been used in several architectural ele-ments, especially in dome and other covering parts [15]. In viewof the widely application of this stone in Cultural Heritages (bothas building and replace material) is of relevant interest the under-standing of the changes on its physical-mechanical proprieties due

Fig. 1. The use of Sabucina Stone in some representative monuments of southern Italy. (a) S. Spirito church; Caltanissetta (1151); (b) Capodarso bridge (1556); (c) NotoCathedral (1703-2007); (d) Caltanissetta Cathedral.

356 G. Barone et al. / Construction and Building Materials 95 (2015) 355–365

to salts damages, also considering the results obtained in a previ-ously research [16] in which the application of innovative andnon-destructive methodology for studying the porous network ofthe stone before and after accelerating aging test allow to highlighthow salts are responsible of a strongly modifications of its porestructure.

For aforementioned, a multidisciplinary approach has beenapplied with the aim at outline the relation between physicaland mechanic characteristics considering the stone degradationdegree. Therefore, comparative measurements of water absorption,mechanical resistance, density, porosity and ultrasonic speedvelocity have been carried out before and after accelerated weath-ering tests. In detail, samples have been subjected to weatheringdue to salts crystallization, as this process is considered the mostdestructive for porous building stones [17].

2. Geological setting

The study area is located 4 kilometers away from the urban cen-ter of Caltanissetta, in central Sicily, at the foot of the SabucinaMountain (Fig. 2). This sector of Sicily holds several syntectonicsedimentary basins, formed since Upper Tortonian [18]. The ana-lyzed calcarenite belong to the Piacenzian Enna Formation whichis formed by marls and marly clays in the lower part andsands-calcarenites in the upper part. This Formation lies uncon-formably on the lower Pliocene deposits and it is covered by themarly clays of the Geracello Formation [19].

3. Material and methods

For this study a total of 50 samples have been obtained from aunique freshly quarried block sampled at the North-Western slopeof the quarry placed in Sabucina Mountain (Caltanissetta, Sicily).

Dimension and shape of the samples are related to the recommen-dation of the applied standard tests.

3.1. Minero-petrographic and chemical analysis

Petrographic analysis have been carried out on thin sections byusing a Zeiss optical polarized microscope. Textural and structuralfeatures have been determined following Folk [20] and Dunham[21] classifications.

The semi-quantitative mineralogical composition have beendetermined by using a Siemens D-5000 Diffractometer on ran-domly oriented powders and on insoluble residue of bulk rock afteracid attack with a solution of chloride acid 0.1 N in concentration.Spectra have been acquired using the range 2–65�2h, step-size of0.02�h and step-time of 2 s/step. The presence of swelling clay min-erals in the fine fraction of insoluble residue has been determinedon oriented slides by treating samples with ethylene glycol at 60 �Cfor 12 h.

Chemical analysis have been performed by using Tescan VegaLMU scanning electron microscope (SEM), equipped with anEDAX NeptuneXM4-60 micro-analyzer, characterized by anultra-thin Be window. Data have been collected by usingmicro-map and spot mode analyses on polished thin section with20 kV accelerating voltage and 0.2 mA beam current.

3.2. Physical test

Physical features have been studied by applying a routine basedon NORMAL and UNI EN tests. Information on real and apparentdensity and porosity have been obtained following UNI EN 1936standard [22] and by performing Mercury Intrusion Porosimetry(MIP) analyses. In detail, the distribution of pore size has beendetermined by using a Thermoquest Pascal 240 macropore unitin order to explore a porosity range �0.0074 lm < r < �15 lm (r

Fig. 2. (a) Photographs of the excavation front of Sabucina quarry and (b) geological setting of the quarried area (modified by ISPRA, 2014). Symbols in geological map: SalsoRiver alluvium (b; Holocene); Lacustrine deposits (e; Upper Plestocene-Holocene); Lannari sands (NNL; Gelasian-Calabrian); Geracello Formation (GER; Gelasian); EnnaFormation marls and marly clays (ENNa; Piacenzian) and sands-calcarenites (ENNb; Piacenzian);Pliocene deposits (TRB);Terravecchia Formation (TRV; Lower Tortonian-Lower Messinian).

G. Barone et al. / Construction and Building Materials 95 (2015) 355–365 357

being the radius of the pores), and by a Thermoquest Pascal 140porosimeter instrument in order to investigate a porosity rangefrom �3.8 lm < r < �116 lm, with maximum pressure of 400 kPaand 200 MPa, respectively. Measurements have been performedon small samples similar in dimension and shape. For each sample,almost three specimens sampled from the same cubic one havebeen analyzed, in order to verify the reproducibility of test. Datahave been collected on fresh and weathered samples in order tohighlight changes in pore structure due to degradation processes.

3.3. Hydric test and durability

In order to study the behavior of the stone against water, capil-larity [23] and total immersion [24] absorption tests have been car-ried out on cubic samples of 7 cm in edges, according to thestandard recommendations.

Accelerated aging test according to UNI EN 12370 standard test[25] have been carried out with the aim at study the durability ofthe stone against salts crystallization and obtain weathered sam-ples representative of different degree of crystallization damages.In detail, test have been performed on 20 cubic samples of 7 cmin edge; with step of almost two cycles, consisting a cycle inimmersion in decahydrate sodium sulphate solution at 14% v/vfor 2 hours at 20 ± 5 �C and dry at 105 ± 5 �C for 16 hours, two sam-ples have been moved from the analyzed set and preserved forphysical, mechanical and MIP tests. According to the standard rec-ommendation, at the end of the test samples have been immersedin deionized water for 24 hours and consequently dried at105 ± 5 �C in order to remove the soluble salts fraction. It is worthnoting that the UNI EN 12370 test have been stopped at the 8th

cycle in order to maintain the cubic shape of samples in view ofperforming mechanical tests after weathering.

3.4. Ultrasonic test

The propagation speed of longitudinal ultrasonic waves hasbeen measured for each specimen according to the European rec-ommendation NORMAL 22/86 [26] along three directions (x, yand z), arbitrarily chosen since neither preferential orientation inthe quarry, nor sedimentary planes in the stone were observed.The test has been performed by a pulser/receiver device (M.A.E.A5000U) and a couple of transducers operating at 200 kHz, allow-ing the input of the sound energy in the specimen and the mea-surement of its travel time. In addition, both dry and wet freshspecimens have been analyzed, as well as samples previously sub-ject to the salts crystallization test. Measurements allowed the cal-culation of the related seismic anisotropy, by using the formulaA% = 100 * (Vmax�Vmin)/Vmean [27].

3.5. Mechanical test

The mechanical properties of the rock, and in particular theUniaxial Compressive Strength (UCS) and the Flexural Strengthhave been estimated through laboratory tests according to [28]and [29] recommendations, respectively. UCS, which is consideredone of the basic geomechanical parameters for the characterizationof a rock [30], has been measured on 12 unweathered cubical spec-imens 7 cm in edge, both in wet and dry conditions, to have a ref-erence value. Subsequently, in order to evaluate the influence ofthe salt crystallization on the mechanical behavior of the rock, uni-axial compression tests have been performed on 13 specimens

Fig. 3. Microphotographs of studied stone; the main textural and structuralfeatures are shown in figure. All: allochems; Sp: sparite; Mc: micrite; Inter_v:inter-particle voids; Intra_v: intra particle voids.

Table 1Semi-quantitative mineralogical composition of the bulk rock and insoluble residue.Cc = calcite; Dol = dolomite; Qz = quartz; Hm = hematite; Goeth = goethite; C.M. = clay minerals. xx = abundant; x = scarce; tr = traces; – = absent.

Cc Dol Qz Hm Goeth C. M.

Bulk rocks xxx xx x tr tr xInsoluble residue – – xxx xx x xx

Fig. 4. SEM images of Sabucina Stone. Textural features analyzed are shown. Mc = micrite

358 G. Barone et al. / Construction and Building Materials 95 (2015) 355–365

subjected to different salts weathering cycles. The FlexuralStrength has been determined on 10 specimens with size15 � 7.5 � 2.5 cm; the analysis have not been carried out onweathered samples because the requested shape is different fromthat suggested by normative for salt crystallization.

4. Results and discussion

4.1. Characterization of the Sabucina stone

Petrographically, Sabucina stone can be classifies as biosparite[20]. Allochems (200 lm–1.5 mm, in dimension) are small frag-ments of foraminifera, bryozoa, lamellibranchia, crinoid, echinidae,algae and gasteropoda; orthochems are manly represented byspathic calcite and small amount of micrite. A siliciclastic compo-nent composed by sub-rounded grains of quartz (2–3 vol%; about100 lm in diameter) has been also identified. According toChoquette et al. [31], porosity (20–30 vol%) is classified as interpar-ticle, moldic and vuggy.

Microscopic observations allow to obtain detailed informationon microtextural features. In particular, worth of noting is that tex-ture varies between grainstone and packestone [21], due to theirregular distribution of sparite and micrite, the latter one enrichedin iron oxides (Fig. 3).

X-ray diffraction patterns of studied stone highlight the pres-ence of (in order of abundance) calcite, dolomite and quartz as

; Sp = sparite; Dol = dolomite; FeTiO3 = iron titanium oxide; FeS2 = iron sulfide oxide.

G. Barone et al. / Construction and Building Materials 95 (2015) 355–365 359

main mineralogical phases. The semi-quantitative mineralogicalcomposition of insoluble residue obtained from acid attack of bulkrock (�1.74%) reveals the presence of quartz, clay minerals, hema-tite and goethite (Table 1). Referring to clay minerals in the finegrain size fraction, the XRD patterns highlight the presence of pre-dominant low-crystallinity smectite and subordinate illite.

Chemical analysis carried out with SEM-EDS on polished thinsection (Fig. 4) allow to determine the average composition of spa-tic calcite (CaO�97%), micrite (mainly composed by SiO2�50%,CaO�15% and small amount of Al2O3, K2O and FeOtot) and oxides(ilmenite and pyrite, due to the presence of SO3,FeOtot, TiO2) pre-sent in the sample (Table 2). Moreover, dolomite crystals dispersedin the micrite fraction have been identified (MgO�30%, CaO�70%;see Fig. 4b), suggesting an initial dolomitization of the stone.

Table 2Average values and standard deviations of the chemical composition of the principal texturalmost three point analyses.

Average values MgO Al2O3 SiO2 K2O

Spatic calcite 2.38 (±0.841) – – –Micrite 4.68 (±0.85) 18.52 (±1.09) 55.73 (±9.25) 3.1Fe–S oxides 1.25 (±0.48) – 5.99 (±0.56) –Fe–Ti oxides – – 0.76 –Dolomite 32.52 – – –

Fig. 5. Water absorption by capillarity: (a) curves related to the three investigated directtotal immersion: average curve (c) and water absorption in the first 30 minutes (d).

Fig. 6. (a) Average curve of weight loss percentage during crystalliz

The average curves on Fig. 5 describe the behavior of stoneagainst water by capillarity (Fig. 5a) and total immersion(Fig. 5c). In particular, the obtained results give back values of0.02 g/cm2⁄ps and 11.07% for capillarity and total absorption coef-ficients, respectively. In Fig. 5b and d the behavior of the stone inthe first 30 minute of capillary absorption and total immersionare shown, respectively. Note worthy is that no significant varia-tions of the water absorption coefficient by capillarity along threeinvestigated directions are detected, according to the isotropic tex-ture of the stone.

Referring to resistance to salts crystallization test [25] deter-mined for eight cycles, the weight loss percentages calculated dur-ing the weathering cycles allow to draft a curve that describe thedegradation style of the studied stone (Fig. 6a). During the first

al and petrographic component of the studied stone are reported. Values are assess on

SO3 CaO TiO2 FeOtot

0.57 (±0.01) 96.69 (±1) – –7 (±0.64) – 11.47 (±2.43) 5.03 (±1.35)

21.18 (±1.03) 1.90 (±0.26) – 69.62 (±0.76)– 1.99 52.7 43.01– 67.48 – –

ions and average curve; (b) absorption in the first 30 minutes. Water absorption by

ation cycles and (b) picture of sample at the end of 8th cycle.

Fig. 7. (a) Cumulative pore volume (cm3/g) vs. pore radius (l) and (b) pore size distribution dV/dlog(R) (cm3/g) vs. pore radius (l) curves of fresh samples. Measurementshave been performed on three samples (gray scale curves). The average curves are shown in black.

Table 3Porosimetric data of analyzed fresh samples.

Sample Total intruded volume(cm3/g)

Bulk density(g/cm3)

Real density(g/cm3)

Porosity(%)

Total surface area(m2/g)

Average poreradius (lm)

Mode pore radius(lm)

Median poreradius (lm)

Fresh_1 0.1283 2.05 2.77 26.24 1.68 0.1527 5.1102 2.1119Fresh_2 0.1429 1.96 2.71 27.94 2.07 0.1379 9.4878 2.3103Fresh_3 0.1356 1.95 2.65 26.41 1.91 0.1419 11.0720 2.3245Average 0.1356 1.98 2.71 26.86 1.89 0.1442 8.5567 2.2489St. Dev. 0.0073 0.05 0.06 0.94 0.20 0.0077 3.0880 0.1189

Table 4Physical and mechanical proprieties determined thought laboratory tests on fresh samples of Sabucina Stone with indication of samples ID, average values and standard deviation(St.Dev).

Sample ID Imbibition coefficient (%) Apparent density (g/cm3) Porosity (%) Ultrasonic velocity(m/s)

Anisotropy (%) UCS (MPa)

Total Efficient Dry Wet Dry Wet Dry Wet

7 15.23 1.87 29.85 28.86 2063.22 3088.78 2.91 14.58 5.5511 14.85 1.89 29.24 28.39 2053 3173.44 5.78 5.99 7.0316 15.02 1.87 29.96 28.42 1998 3175.56 3.15 9.45 5.4444 13.73 1.92 28.10 26.67 2110 3245 10.97 12.17 6.8149 15.53 1.84 30.92 29.00 1961 2955 7.73 14.31 5.2158 15.96 1.84 31.08 29.71 1953.33 2959.56 3.75 4.35 6.6538 15.07 1.88 29.70 28.63 2030.78 3045.56 4.87 15.21 5.7340 15.19 1.87 30.12 28.69 2028.8 2875.8 12.86 15.71 8.0942 14.58 1.89 29.36 27.83 2100 3154 3.11 1.57 5.4845 15.38 1.86 30.32 28.97 2008 3031.44 17.15 4.37 5.4946 15.69 1.85 30.90 29.31 2010 2787.33 8.97 6.69 4.2648 15.83 1.84 31.00 29.52 2020 3019 3.88 4.65 5.76Min 13.73 1.84 28.10 26.67 1953.33 2787.33 2.91 1.57 5.21 4.26Max 15.96 1.92 31.08 29.71 2110.00 3245.00 17.15 15.71 7.03 8.09Average 15.17 1.87 30.05 28.67 2028.01 3042.54 7.09 9.09 6.12 5.80St. Dev. 0.61 0.02 0.89 0.81 48.18 134.35 4.58 5.09 0.80 1.25

360 G. Barone et al. / Construction and Building Materials 95 (2015) 355–365

six cycles, a mass increasing due to crystallization of salts withinthe pore structure is described by positive values of DM%; on thecontrary, from the seventh weathering cycle negative DM% hasbeen recorded. Overall, samples show an average weight loss of�4.16% at the end of 8th cycle. The average mass loss after thecleaning of samples in deionised water in order to remove saltsis �6.15%,The main degradation forms are represented by granulardisintegration and powdering, in association with slight chromaticchanges (Fig. 6b).

Porosity and pore size distribution of studied stone (fresh_#)have been investigated by mercury intrusion porosimetry tests(MIP). Cumulative pore volume vs. pore radius and pore size

distribution curves are shown in Fig. 7. The obtained results(Table 3) highlight a total accessible porosity of 26.86% and a totalpore volume of 0.1356 g/cm3. Referring to pore size distribution,average pore radius of 0.1442 lm and mode peak at 8.5567 lmhave been obtained.

Physical parameters have been also determined following UNIEN standard tests; the obtained results are summarized in Tables4 and 5.

In detail, the average real density, refereed to the volume of thespecimen with no voids and determined through the pycnometermethod, is 2.67 g/cm3, while the average apparent density mea-sured, considering the presence of pores, is 1.87 g/cm3. Referring

Table 5Flexural strength resistance determined thought laboratory tests on fresh samples ofSabucina Stone with indication of samples ID, average values and standard deviation(St.Dev).

Sample ID Rtf (MPa)

2L 5.593L 8.837L 6.488L 7.3610L 5.5911L 4.1213L 5.5914L 6.4815L 7.0618L 8.24Min 4.12Max 8.83Average 6.53St. Dev. 1.4

Fig. 8. Mode of failure of the specimens under uniassial compressive strength.

G. Barone et al. / Construction and Building Materials 95 (2015) 355–365 361

to porosity, average total and effective porosity are 30% and 29%,respectively; a comparison between this two latter data and theimbibition coefficient obtained (ranging from 13% to 16%) suggesta really high permeability of the rock.

The average Vp values have been measured both in dry and wetconditions with the aim to evaluate the structural anisotropy of thestudied stone [32]. The mean velocity measured in the dry speci-men is 2040 m/s (±46.16) with a mean anisotropy value of 7.09%,while for wet specimens a mean velocity of 3042 m/s (±77.2) with9.09% of anisotropy have been obtained, suggesting a quite isotro-pic arrangement of structural and textural elements.

Finally, the most significant mechanical resistance parameterfor Sabucina Stone have been determined by applying compressiveuniaxial (both in dry and wet conditions) and flexural stresses.

Referring to compressive strength, the UCS values in dry condi-tion ranges from a minimum of 5.21 to a maximum of 7.03 MPa,while in wet condition ranges between 4.26 and 8.09 MPa, withaverage values quite similar each other (6.12 ± 0.8 MPa and5.80 ± 1.25 MPa, respectively). This evidence suggest that themechanical behavior of the stone is not strongly affected by the

presence of water, even if a slight decrease of compressive resis-tance can be observed in saturated samples. The failure mode ischaracterized by fractures with orientation from parallel to 30� ofinclination with respect to the axis of load, as can be observed inFig. 8, and by a quasi-brittle behavior; in fact, the load-time curvesshow a linear trend until the failure of the specimen, which is fol-lowed by a rapid unloading phase.

Finally, the evaluation of resistance to flexural mechanicalstress in dry conditions give back values ranging between 4.12and 8.83 MPa, with average flexural strength of 6.53 ± 1.4 MPa,slightly higher than compressive strength.

The correlations between the physical and mechanical parame-ters determined (Fig. 9) allow to better understand how the vari-ability in term of porosity and heterogeneity in textural featuresinfluence the mechanical behavior of the studied stone. Asexpected, an increasing of porosity determine a strongly decreaseof both compressive strength and ultrasound velocity, testifiedby the linear negative trends for UCS values and ultrasonic velocityvs. porosity shown in Fig. 9a and d. Moreover, the greater is flexu-ral strength the more resistant is the material to application ofcompressive load (positive linear trend; see Fig. 9c).

The previous observation support that the study of this stoneshould considered the range of values in spite of the average valuesin order to take in account its micro-textural and micro-structuralheterogeneity.

4.2. Physical and mechanical characterization of samples after saltsweathering

The changes in physical and mechanical proprieties of the stud-ied stone after salts weathering have been esteemed by comparingmercury intrusion porosimetric curves and ultrasound, physicaland mechanical data before and after accelerated aging tests.

In detail, mercury intrusion porosimetry tests have been per-formed on weathered samples at 2nd, 4th, 6th and 8th weatheringcycles, in order to highlight the difference in pore network arrange-ment due to salts weathering over different degradation degree. InTable 6 results obtained on analyzed samples are summarized,while in Fig. 10 cumulative pore volume vs. pore size and pore sizedistribution curves (dV/dlog(R) vs. pore size) collected forunweathered (in red) and weathered samples (gray scale) areshown.

As can be seen in Fig. 10a, the total pore volume generallydecrease from quarried sample to weathered ones, suggesting adecrease of the global volume accessible to mercury due to thepresence of salts crystals.

According to Angeli et al. [33] a comparison between fresh andweathered pore size distribution curves (dV/dlog(R) vs. pore size)is able to supply information on weathering mechanism (Fig. 11).In detail, if salt crystals obstruct a pore entry (case 1, Fig. 11a),the entire volume is seen with a lower diameter (e.g., pore radiusmode is low in weathered samples) and the global porosity remainunchanged; otherwise, if salt crystals partially fill a pore (case 2,Fig. 11b), the pore size entry is not modified (e.g., pore size modeis the same for fresh and altered samples) but the global volumewill be lower. Finally, if both phenomena arise and salt crystalsobstruct a pore entry and partially fill it (case 3, Fig. 11c), a shiftof modal pore size toward lower values and a decrease of globalvolume can been observed.

For aforementioned, an inspection of pore size distributioncurves in weathered samples respect to unweathered one(Fig. 10b) allow to understand the pore network modification ofstone due to salts action. In detail, at the second cycle, total porevolume is quite unchanged, while pore radius mode is lower thanfresh sample; this behavior suggest that, in a first stage, salt crys-tals obstruct pore entries (case 1). From the fourth weathering

Fig. 9. Correlations between the main physical-mechanical properties determined: (a) UCS (MPa) vs. porosity (%); (b) UCS (MPa) vs. Ultrasonic velocity (m/s); (c) UCS (MPa)vs. flexural strength (MPa); (d) ultrasonic velocity (m/s) vs. porosity (%).

Table 6Porosimetric data obtained on analyzed weathered samples. Average parameters asses on almost three specimens and standard deviations are reported.

Nweatheringcycles

Total intrudedvolume (cm3/g)

Real density(g/cm3)

Apparentdensity (g/cm3)

Porosity (%) Total surfacearea (m2/g)

Average poreradius (lm)

Modal poreradius (lm)

Median poreradius (lm)

2 cycle 0.096 (±0.72) 1.978 (±0.039) 2.698 (±0.047) 26.687 (±0.223) 2.191 (±0.234) 0.124 (±0.011) 7.799 (±1.168) 2.005 (±0.154)4 cycle 0.126 (±0.014) 2.148 (±0.296) 2.939 (±0.431) 26.860 (±0.609) 1.743 (±0.182) 0.145 (±0.007) 14.693 (±2.81) 3.231 (±0.598)6 cycle 0.117 (±0.117) 2.056 (±0.036) 2.705 (±0.095) 23.973 (±1.383) 1.739 (±0.210) 0.135 (±0.012) 12.843 (±3.043) 2.653 (±0.256)8 cycle 0.131 (±0.131) 1.949 (±0.109) 2.614 (±0.163) 25.383 (±3.194) 1.976 (±0.224) 0.132 (±0.010) 11.636 (±2.159) 2.757 (±0.919)

Fig. 10. (a) Average cumulative pore volume (cm3/g) vs. pore radius (l) and (b) average pore size distribution dV/dlog(R) (cm3/g) vs. pore radius (l) curves of fresh (red) andweathered samples (gray scale). (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

362 G. Barone et al. / Construction and Building Materials 95 (2015) 355–365

Fig. 11. Weathering mechanism due to salts crystallization and porosimetric curves modified from Angeli et al. [33].

Table 7Average values of the main physical-mechanical parameters of Sabucina Stone at different salt weathering cycles.

Apparent density (g/cm3) Real density (g/cm3) Ultrasonic velocity (m/s) Porosity (%) UCS (MPa)

Freshly quarried Sabucina stone 2.04 2.77 2028.01 ± 12.71 26.24 6.12 ± 0.802 cycle 2.02 2.76 1443.50 ± 8.23 26.70 6.53 ± 1.344 cycle 1.94 2.63 1463.00 ± 4.62 26.47 7.72 ± 1.166 cycle 2.01 2.61 1423.67 ± 8.16 22.81 6.17 ± 0.277 cycle 1.97 2.68 1455 ± 12.99 25.85 6.44 ± 2.158 cycle 1.90 2.45 1200.00 ± 6.14 22.32 7.41 ± 4Residuals 1.83 2.63 1110.88 ± 6.2 30.55 4.11 ± 1.1

Fig. 12. Variation of average values of (a) real density (g/cm3); (b) porosity (%); (c) ultrasonic velocity (m/s) and (d) UCS (MPa) parameters in function of weathering cycles.

G. Barone et al. / Construction and Building Materials 95 (2015) 355–365 363

364 G. Barone et al. / Construction and Building Materials 95 (2015) 355–365

cycle, samples show a lower total pore volume values and a shift ofmodal pore radius towards higher values. These evidences suggestthat in weathered samples salts crystals fill pores characterized bylarger pore entries than in fresh one with an enlargement of thethroats (case 2 with larger pore entries).

The variation on structural arrangement after salt crystalliza-tion in pores influence also the physical proprieties and themechanical behavior of the stone (Table 7, Fig. 12). In fact, adecrease of real density determined by pycnometer method(Fig. 12a) and total porosity (Fig. 12b) can be observed; in detail,the decrease of total porosity is due to the presence of salts fillingthe pores, while referring to real density, such behavior has to becorrelated to the dissolution of salts during the water pycnometermeasurements and the consequently increasing of porosity in therange below 30 lm inside the powered material used for the test.In addition, a decrement of ultrasound velocity (Fig. 12c) has beenobserved, especially up to the 6th salt crystallization cycle, due tothe occurring of macro- and micro-fractures generated by thegrowth of salts within the pore network.

Finally, a slight and irregular increasing of compressive strengthcan be observed (Fig. 12d). About to the latter one aspect, a corre-lation between the change in mechanical behavior of the stone andthe modification of the degradation mechanism determined byporosimetric analyses has been observed. In fact, up to the 2ndto the 4th crystallization cycles, salts fill the pore throats, deter-mining a slight and progressive increment of compressive strengthresistance. Otherwise, when the enlargement of the throats and thefilling of the pores have occurred (i.e. up to 6th crystallizationcycle) a decrease of UCS values has been observed, claiming astrictly relation between the location of salts and the mechanicalresistance of the stone.

The presence of salts in the pore network modify also themechanical behavior of the stone, from quasi-brittle toquasi-plastic. On the contrary, the removal of salts from the sam-ples (residuals in Fig. 12) determine a significant worsening ofphysical and mechanical proprieties, due to an intense degradationaction, as testified by accelerated aging tests previously performed.

5. Conclusions

In this work a complete characterization of the Sabucina Stone,both from the physical and mechanical point of view, have beensupplied, highlighting also features related to salt weathering phe-nomena. In detail, the results allow us to obtain a good explanationof the degradation style of the stone, giving back an interesting cor-relation between the mass loss curves and the weathering mecha-nism suggested by porosimetric data. In fact, from the first to thesixth cycles the curve is ‘‘salt controlled’’ (positive mass variation)and the porosimetric curve suggests the crystallization of salts inthe entry of pores, while from the seventh cycle the curve is‘‘weathered controlled’’ (negative mass variation), with an increas-ing of the throats radius and crystallization of salts inside theenlarged pores. Such a behavior affect also thephysical-mechanical properties. In fact, a globally decrease of den-sity and ultrasound velocity can be observed, especially from the6th salts crystallization cycle, even if a non homogeneous trendmay be delineated. Referring to compressive strength, the irregularincreasing of UCS values in samples with salts and the subse-quently decrease of them can be explained only thanks to a deep-ened knowledge of the weathering mechanism occurred into thepore network of the stone. In addition, the significant decrease ofUCS values after cleaning the stone strongly suggests the necessityto remove salts from weathered samples in order to have a correctesteem of mechanical proprieties in degraded stones.

Finally, the variability of the experimental parameters can beunderstood only considering the micro-textural andmicro-structural heterogeneity observed during the characteriza-tion of the freshly quarried stone, in which the ranging of petro-graphic classification determine a significant range of value, themaximum and the minimum ones equally representative of itsbehavior; therefore, the necessity of a multi-methodologicalapproach in investigating the physical-mechanical characteristicsof a building stone for allowing the correct interpretation of data,in consideration of its microtextural and microstructuralproprieties.

Acknowledgments

The authors are very grateful to Prof. E. Curcuruto and to Ing. M.Dell’Aera for supplying freshly quarried blocks of Sabucina Stone.

References

[1] B.A. Richardson, The durability of porous stones, Stone Ind. 26 (1991). 25-22.[2] B.K. Mod’d, R.J. Howarth, C.H. Bland, Rapid prediction of building research

establishment limestone durability class from porosity and saturation, Q. J.Eng. Geol. Hydrogeol. 29 (1996) 285–297.

[3] Y. Matsukura, N. Matsuoka, The effect of rock properties on rates of tafonigrowth in coastal environments, Zeitschriftfür Geomorphologie 106 (1996)57–72.

[4] S. Ordóñez, R. Fort, M.A. García del Cura, Pore size distribution and thedurability of a porous limestone, Q. J. Eng. Geol. Hydrogeol. 30 (1997) 221–230.

[5] D. Benavente, M.A. García del Cura, R. Fort, S. Ordóñez, Durability estimation ofporous building stones from pore structure and strength, Eng. Geol. 74 (2004)113–127.

[6] M. Angeli, J.P. Bigas, D. Benavente, B. Menéndez, R. Hébert, C. David, Saltcrystallization in pores: quantification and estimation of damage, Environ.Geol. 52 (2007) 205–213.

[7] S. Yu, C.T. Oguchi, Role of pore size distribution in salt uptake, damage, andpredicting salt susceptibility of eight types of Japanese building stones, Eng.Geol. 115 (2010) 226–236.

[8] A. Tugrul, The effect of weathering on pore geometry and compressive strengthof selected rock types from Turkey, Eng. Geol. 75 (2004) 215–227.

[9] M. Ludovico-Marques, C. Chastre, Effect of salt crystallization ageing on thecompressive behavior of sandstone blocks in historical buildings, Eng. FailureAnal. 26 (2012) 247–257.

[10] A. Tugrul, I.H. Zarif, Correlation of mineralogical and textural characteristicswith engineering properties of selected granitic rocks from Turkey, Eng. Geol.51 (1999) 303–317.

[11] G.F. Andriani, N. Walsh, Physical properties and textural parameters ofcalcarenitic rocks: qualitative and quantitative evaluations, Eng. Geol. 67(2002) 5–15.

[12] N.K. Tamrakar, S. Yokota, S.D. Shrestha, Relationships among mechanical,physical and petrographic properties of Siwalik sandstones, central Nepal sub-himalayas, Eng. Geol. 90 (2007) 105–123.

[13] K. Zorlu, C. Gokceoglu, F. Ocakoglu, H.A. Nefeslioglu, S. Acikalin, Prediction ofuniaxial compressive strength of sandstones using petrography-based models,Eng. Geol. 96 (2008) 141–158.

[14] A. Bellanca, E. Curcuruto, S. Lo Bue, Neri R. Petrografia, Geochimica eriferimenti all’impiego storico delle calcareniti plioceniche in Sabucina,Sicilia centrale, Mineralogica et petrographica acta 42 (1999) 193–210.

[15] R. De Benedictis, S. Tringali, La Ricostruzione Della Cattedrale Di Noto, L.C.T.Edizioni, Ragusa, 2000.

[16] S. Raneri, V. Cnudde, T. De Kock, H. Derluyn, G. Barone, P. Mazzoleni, X-raycomputed micro-tomography to study the porous structure and degradationprocesses of a building stone from Sabucina (Sicily), Eur. J. Mineral. (2015),http://dx.doi.org/10.1127/ejm/2015/0027-2433.

[17] M. Steiger, S. Asmussen, Crystallization of sodium sulfate phases in porousmaterials: the phase diagram Na2SO4–H2O and the generation of stress,Geochimica et Cosmochimica Acta 72 (2008) 4291–4306.

[18] A. Di Grande, V. Giandinoto, Plio-Pleistocene sedimentary facies and theirevolution in centre-south-eastern Sicily: a working hypothesis EGU StephanMueller, Spec. Publ. 75 (2001). 221-211.

[19] L. Tortorici, M. Grasso, S. Catalano, C. Monaco, G. De Guidi, G. Pappalardo, E.Curcuruto, M. Brescia, A. Di Stefano, R. Maniscalco, G. Barreca, G. Sturiale, CartaGeologica d’Italia. Scala 1:50.000 Foglio 631 Caltanissetta 2010.

[20] R.L. Folk, Practical petrographic classification of limestones, Am. Assoc. Pet.Geol. Bull. 43 (1959) 1–38.

[21] R.J. Dunham, Classification of carbonate rocks according to their depositionaltexture, in: E. Ham (Ed.), Classification of Carbonate Rocks, AmericanAssociation of Petroleum. Geologists Memoir, Tulsa, 1962, pp. 21–108.

[22] UNI EN 1936 Natural stones test methods: Determination of real density andapparent density and of total and open porosity, UNI ed., Milano, 2001.

G. Barone et al. / Construction and Building Materials 95 (2015) 355–365 365

[23] NORMAL 11/88 Natural stones test methods: determination of waterabsorption coefficient by capillarity.

[24] NORMAL 7/81 Natural stones test methods: determination of waterabsorption by total immersion.

[25] UNI EN 12370:2001 Natural stone test methods – Determination of resistanceto salt crystallization. Ed. UNI (Ente Nazionale Italiano di Normalizzazione),Milano.

[26] NORMAL 22/86 Natural stones test methods: Determination of ultrasonicwaves velocity.

[27] F. Birch, The velocity of compressional waves in rocks to 10 Kbar: Part 2, J.Geophys. Res. 66 (1961) 2199–2224.

[28] UNI EN 1926 Natural stones test methods: determination of compressionstrength, UNI ed., Milano, 2000.

[29] UNI EN 12372, Natural stones test methods: determination of flexural strengthresistance under concentrated load, UNI ed., Milano, 2001.

[30] G. Pappalardo, S. Mineo, G. Marchese, Effects of cubical specimen sizing on theUniaxial Compressive strength of Etna volcanic rocks (Italy), Ital. J. Eng. Geol.Environ. 2 (2013) 45–54.

[31] P.W. Choquette, L.C. Pray, Geological nomenclature and classification ofporosity in sedimentary carbonates, Am. Assoc. Pet. Geol. Bull. 54 (1970)207–250.

[32] G. Barbera, G. Barone, P. Mazzoleni, A. Scandurra, Laboratory measurement ofultrasound velocity during accelerated aging tests: Implication for thedetermination of limestone durability, Constr. Build. Mater. 36 (2012) 938–977.

[33] M. Angeli, D. Benavente, J.P. Bigas, B. Menéndez, R. Hébert, C. David,Modification of the porous network by salt crystallization in experimentallyweathered sedimentary stones, Mater. Struct. 41 (2008) 1091–1108.