MICROSCOPY RESEARCH AND TECHNIQUE 65:270 –275 (2004)
Determination of the Distribution of Consolidants and Interpretation of Mercury Porosimetry Data in a Sandstone Porous Network Using LSCM KARIMA ZOGHLAMI
´ MEZ-GRAS* DAVID GO
Departament de Geologia, Universitat Auto`noma de Barcelona, Spain
sandstones; 3D distribution; LSCM; consolidants; porous network
ABSTRACT Knowing the 3D distribution of a consolidant within the porous network of a rock is essential for understanding the porosity quantitative data obtained by mercury porosimetry and for observing the effect of consolidants on pore interconnection. In this work, we show for the ﬁrst time that the distribution of consolidant in the porous network can be determined using laser scanner confocal microscopy (LSCM). Results indicate that consolidants are concentrated in pore throats of less than 40 m in diameter, affecting both the porous interconnection and the circulation of ﬂuids. LSCM allowed demonstration of the fact that the increase in microporosity detected by mercury porosimetry is due to the development of ﬁssures within the consolidants. No consolidant that produces this kind of ﬁssure can be used in the consolidation of building stones, since it would increase microporosity and, in consequence, vulnerability to weathering agents. Microsc. Res. Tech. 65:270 –275, 2004. © 2005 Wiley-Liss, Inc. INTRODUCTION The utility of a consolidant lies in reestablishing the cohesion of the particles in a deteriorated stone (Dukes, 1972; Torraca, 1975; Alessandrini et al., 1975). In addition, a good consolidant should meet performance requirements concerning durability, depth of penetration (Young et al., 1999), effect on stone porosity, effect on moisture transfer (Borselli et al., 1990; Dell’Agli et al., 2000) compatibility with stone, and effect on appearance (Biscontin et al., 1975). The porosity and pore size distribution of a stone may have a major effect on its durability. For example, the resistance of a given type of stone to frost and salt damage decreases as the proportion of ﬁne pores increases (Hudec, 1978; Camaiti and Amoroso, 1997). A stone consolidant that reduces pore size without plugging may therefore be harmful. Changes in pore-size distribution are usually caused in rocks treated with consolidants (Esbert, 1993; Villegas et al., 1995; Alvarez De Buergo et al., 2004). Mercury porosimetry is probably the most commonly used indirect technique by which to characterize the porous network conﬁguration both of untreated and treated rocks with conservation products (Sasse et al., 1993). With this technique, the porous network is incorrectly simulated as a collection of cylindrical noninterconnected tubes that is very far from true pore shapes and network conﬁguration. Therefore, the geometric complexity of pore space based on pore throats and pore bodies leads to ambiguities in the physical interpretation of mercury data and of other indirect methods. Information provided by mercury porosimetry is thus insufﬁcient and should be complemented with other techniques. Scanning electron microscopy (SEM) has been the most commonly applied direct technique used in determining the distribution of conservation products in ©
2005 WILEY-LISS, INC.
porous rocks (Esbert et al., 1990; Piacenti et al., 1993; Paterno and Charola, 2000; Alvarez De Buergo and Fort, 2001). Nevertheless, SEM allows analysis of sample surfaces only (rock fragments or thin sections). Hence, in previous work only 2D images of the surface of a treated sample were obtained, the effect of the consolidant in the porous network having to be deduced from mercury porosimetry data. In addition, the use of 2D images introduces further problems related to the interconnection of the porous network, as these often mask the true 3D topology (Petford et al., 1999). The application of a consolidant may cause a decrease or increase of micropores, but as the quantitative porosity data are obtained by mercury porosimetry, both effects are usually interpreted simplistically in the same way. When an increase in microporosity is detected, it is interpreted as a total sealing of the missing pores. When a decrease in microporosity occurs, it is interpreted as a partial sealing of originally larger pores (Esbert and Dı´az-Pache, 1993). This general interpretation, applied to different kinds of rocks, is due to a lack of information about the real 3D conﬁguration of the porous network in each type of rock. Without knowing the initial conﬁguration of the porous network or the spatial distribution of the consolidant, it is very difﬁcult to correctly interpret the effect of this consolidant. Laser scanning confocal microscopy (LSCM) is an optical imaging technique that has been used success-
*Correspondence to: David Go´mez-Gras, Dpt. Geologia, Ediﬁci Cs. Universitat Auto`noma de Barcelona, 08193 Bellaterra, Spain. E-mail: [email protected]
Received 28 June 2004; accepted in revised form 12 October 2004 Contract grant sponsor: Agencia Espan˜ola de Cooperacio´n Internacional (AECI) (to K.Z.). DOI 10.1002/jemt.20119 Published online in Wiley InterScience (www.interscience.wiley.com).
CONSOLIDANT DISTRIBUTION IN SANDSTONES (LSCM)
fully in the 3D pore structure reconstruction of rock material (Petford and Miller, 1990, 1993; Montoto et al., 1995; Pironon et al., 1998; Fredrich, 1999; Petford et al., 1999; Mene´ndez et al., 1999). Computer controlled CSLM allows a rapid, nondestructive serial sectioning to a resolution of 0.1 m in all planes, and the resultant volumetric image data allow a 3D reconstruction of the porous network. In the present work, LSCM was used in order to improve the observation of the 3D distribution of a consolidant within the porous network of a building sandstone and to better understand mercury porosimetry data, usually misinterpreted due to a lack of 3D distribution for consolidants. MATERIALS AND METHODS The building sandstone studied here is a noncemented quartz-arenite, lithiﬁed by compaction and composed of quartz grains (69 – 84%), feldspars (mainly orthoclase; 0 –1.1%), porosity (17–25%), and clay minerals as matrix (0 –11%) (Go´mez-Gras and Zoghlami, 2003; Zoghlami et al., 2004b). The quartzose composition of the rock, absence of unstable (unreactive) minerals and cements, together with its excellent hydric behavior provides very high resistance to chemical alterability. Nevertheless, due to its low mechanical resistance, this rock is particularly vulnerable to weathering agents whose mechanisms involve disruptive mechanical forces, such as salt crystallization, to freezing processes, and to changes caused by thermal expansion (Zoghlami et al., 2004a). The 3D porous network conﬁguration was studied by LSCM using a Leica TCS-SP2-AOBS microscope. Images were obtained at a resolution of 1024 ⫻ 1024 pixels and at depth steps of 0.5 m. The laser was ﬁxed to outputs of 25% and ⫽ 488 nm. Samples were prepared in accordance with the method of Fredrich (1999). Oligomeric ethyl silicates, Tegovakon V (TV) and Keim-Silex OH (KSOH), were selected as consolidants. This choice was based on the good results obtained in previous studies of sandstone (Garcı´a Garmilla et al., 2002; Wheeler, 1992) and of other types of rocks (Esbert, 1993; Villegas and Vale Parapar, 1993; Villegas et al., 1995; Rivas et al., 1998). In order to determine the distribution of the consolidant within the porous network, samples were prepared in accordance with the following procedure: ● Addition of powdered ﬂuorochrome to the consolidant solution (5 g/l 1 of consolidant). ● The obtained solution was left for 1 hour to allow total dissolution of the ﬂuorochrome in the consolidant solution. ● Brush application of the doped consolidant solution to the sandstone samples (5 ⫻ 5 ⫻ 5 cm). ● Curing treated samples for 1 month with doped consolidant to permit complete polymerization of the consolidant. ● Preparation and mounting of polished planar sections (thickness ⬎ 100 m) on a glass slide. The pore-size distribution of the untreated and treated sandstone samples was determined by mercury intrusion porosimetry, following RILEM recommendations.
TABLE 1. Characteristics and parameters of the studied samples obtained from the mercury intrusion porosimetry Samples 3
Real density (g/cm ) Apparent density (g/cm3) Total porosity (%) Macroporosity (%) Microporosity (%)(*) Average diameter (m) Mode (m)
2.60 2.03 22.35 89.18 8.65 33.40 30–40
2.60 1.96 24.79 85.38 12.82 23.43 20–30
2.60 1.97 24.57 85.52 12.87 21.20 20–30
2.60 1.97 25.86 83.47 14.65 20.80 20–30
*Microporosity: pore diameter ⬍ 15 m.
Fig. 1. access.
Distribution of the porosity based on the diameter of pore
RESULTS Porous Network Before the Application of Consolidants Pore-Size Distribution. Total porosity of the rock, determined by mercury intrusion porosimetry, showed high values (17–25%). Macroporosity (pore diameter ⬎ 15 m), according to Gon˜i et al. (1968) and Bousquie´ (1980) represents a percentage of 83.38 – 89.18%, whereas microporosity ranges from 8.65–14.78%, indicative of the sandstone’s macroporous character (Table 1; Fig. 1). The mode of the pore-access size varied from 20– 30m, representing 60 – 80% of total porosity. Mercury porosimetry results show that most of the pores (⬎80%) have a pore access diameter of 20– 40m, depending on the rock’s grain size. The rest of the pores (⬍20%) present a diameter pore access less than 15 m (micropores). However, this method only allows obtainment of quantitative porosity data; it does not provide information on the arrangement of porosity, nor the way that pores are interconnected. For these reasons, in addition to achieving a 3D reconstruction of the sandstone’s porous network, ﬂuorescence and confocal microscopy were used. Laser Scanning Confocal Microscopy (LSCM). Fluorescence images of thin sections showed that sandstone samples had a very simple porous structure constituted by pore throats and pore bodies (Fig. 2A). In general, pores displayed channel-like shapes with diameters smaller than 40 m (Fig. 2B). Megapores reaching up to 300 m in diameter were also observed (Fig. 4B). As the images of thin sections obtained by ﬂuorescence microscopy were 2D, it was not possible to infer the degree of interconnection between both pore types; the real conﬁguration of the porous network could therefore not be determined.
´ MEZ-GRAS K. ZOGHLAMI AND D. GO
Fig. 2. LSCM images. A: General aspect of the sandstone porosity. Bar length: 80 m. B: Detailed image of pores (communicated channels). Scale bar ⫽ 40 m. C: 3D reconstruction of the porous network. Scale bar ⫽ 40 m. D: 3D pore detail. Scale bar ⫽ 40 m.
As mercury porosimetry only measures pore access, real pore size was measured using LSCM. The obtained results showed that in ﬁne-grained sandstones, pore sizes varied from 50 – 60 m, and could reach up to 120 m. In the medium-grained sandstones, the average main pore size was around 200 m, reaching up to 600 m, giving a macroporous character to the rock that allowed optimal interconnection between the pores that facilitated ﬂuid circulation. The 3D reconstruction of the porous network (Fig. 2C) showed that it was constituted by a single pore system whose conﬁguration only depends on grain arrangements and degree of compaction. The porous network was constituted by channels (⬍40 m in diameter) that may occasionally expand, giving rise to megapores of up to 300 m in diameter.
Combined LSCM and mercury intrusion porosimetry data allowed recognition of the fact that porosity was present as large pores, intercommunicated by channels that constitute pore accesses. Although of smaller size with respect to the main pores, these pore accesses were still large (20 – 40 m). On the other hand, the 3D reconstruction of the porous network in the sandstone facilitated not only an understanding of the effect of consolidants on pore-size distribution and on the network conﬁguration of the rock, but also allowed correct interpretation of the porosimetry data. Effect of Consolidants on the Porous Network Pore-Size Distribution Deduced From Mercury Intrusion Porosimetry. Compared to untreated samples (MB, Table 2), samples treated with consolidants
CONSOLIDANT DISTRIBUTION IN SANDSTONES (LSCM) TABLE 2. Untreated (MB) and treated sample porosity (mercury intrusion porosimetry)
Samples MB TV KSOH
Total porosity (%)
Average diameter (m)
22.35 18.99 19.42
89.98 69.61 78.60
8.65 29.06 19.90
33.40 31.93 31.73
Fig. 3. Pore-size distribution of the sandstone treated with KSOH consolidant (A) and treated with TV consolidant (B).
(KSOH, TV) showed a moderate decrease in total porosity. Consolidants affected pore-size distribution by producing a decrease in macroporosity and an increase in microporosity, especially for pores of less than 10 m in diameter (Fig. 3A,B). Figure 3 shows that a decrease occurred in the amount of pores of 20 – 40 m, the most abundant range in this rock (Figs. 1, 3A,B). Additionally, there was the appearance of pores having a smaller diameter (⬍10 m). The presence of this new pore population might be explained by a partial sealing of pores of 20 – 40 m diameter. The effect of partial sealing in different treated rocks has been discussed by Esbert and Dı´az-Pache (1993). Distribution of Consolidants from LSCM. In thin sections, LSCM observations showed that grain surfaces were covered by a discontinuous coating ﬁlm. When grains were very close, consolidants usually formed meniscus plugging only in the small throats (diameter ⬍ 40 m) (Fig. 4A,B). 3D reconstruction (Fig. 4C) showed that consolidants ﬁlled the pore-network throats (pore diameter of less than 40 m), whereas megapores were covered by a very thin (1–2 m) coating ﬁlm of consolidant (Fig. 4B). It is worth observing that the consolidants used in our experiments developed cracks reaching up to 10 m in diameter. These cracks developed a network of small
channels interconnected between both themselves and the megapores (Fig. 4D). DISCUSSION Mercury porosimetry data from samples treated with consolidants showed a microporosity increase with respect to untreated samples (Table 2), a decrease in pores with diameters of 20 – 40 m, and the development of new pores with diameters smaller than 10 m (Fig. 3). Nevertheless, LSCM images demonstrated that the main pores sizes (megapores) were not affected by the application of consolidants, due to the development of a very thin coating ﬁlm around the grains. 3D reconstruction showed that consolidants are concentrated within throats having a pore diameter smaller than 40 m and that cracks developed in the polymer could reach up to 10 m in size. Therefore, the increase in microporosity can be attributed to the formation of microﬁssures, probably due to the type of catalyst used; this probably destabilizes the polymer used as consolidant, as reported by Brus and Kotlik (1996). These results also showed that a correct interpretation of the mercury porosimetry data needs to consider the formation of microcracks during the application of a consolidant. Some authors have demonstrated that the presence of microporosity made the rock more vulnerable to weathering agents whose mechanisms involve stress, such as salt crystallization and hydration (Hudec, 1978; Camaiti and Amoroso, 1997). As the development of microﬁssures in the polymer was the cause of an increase in microporosity, the consolidants used (TegovakonV and Keim Silex H) may not be suitable for improving the rock’s resistance. CONCLUSIONS LSCM was used for the ﬁrst time to determine the 3D distribution of consolidants in porous rocks. It proved to be a useful technique in achieving a 3D reconstruction of a rock’s porous network, and in understanding quantitative porosity data obtained by mercury porosimetry. LSCM also allowed the study both of consolidant distribution within the rock’s pores and of the consolidant’s effect on the pore-network conﬁguration and interconnection, both of these being important properties affecting rock durability, since they control the circulation of ﬂuids. LSCM demonstrated that consolidants were concentrated within throats with diameters smaller than 40 m, affecting the porous interconnection and complicating the circulation of ﬂuids. However, in megapores (greater than 40 m), the consolidant was dispersed only as a thin ﬁlm around the pores. Consolidants developed microﬁssures that explain the origin of the new porosity detected by mercury porosimetry. LSCM showed that an increase in microporosity was due to polymer cracking and not to partial pore sealing or to the reduction in pore diameter by the consolidant coating ﬁlm. These consolidants will increase microporosity, thus increasing the vulnerability of the stone to the weathering agents whose mechanisms involve disruptive mechanical forces, such as salt crystallization and freezing processes. The usefulness of LSCM in building-stone preservation studies has therefore been demonstrated.
´ MEZ-GRAS K. ZOGHLAMI AND D. GO
Fig. 4. Fluorescence microscope images. Distribution of consolidant (A) and (B). Red arrow: porosity; yellow arrow: grain; white arrow: consolidant. Scale bar ⫽ 80 m. 3D reconstruction. C: Image shows that consolidant is concentrated in throats. Scale bar ⫽ 80 m. D: Image shows microﬁssures in consolidant. Scale bar ⫽ 40 m.
ACKNOWLEDGMENTS The authors thank Esteve Cardellach for critical review of earlier drafts of the manuscript. The authors thank M. Martı´, F. Bohils, O. Castell (Servei de Microsco`pia de la Universitat Auto`noma de BarcelonaUAB), J. Ques (Laboratori La`mines Primes, UAB), and D. Parcerisa for help during the work. We also thank David Owen for revision of the English version. REFERENCES Alessandrini G, Del Fa CM, Rossi-Doria P, Tabasso M, Vannucci S. 1975. Treatment of stone in monuments. A review of principles and processes. In: The conservation of stone I. Bologna, Italy: Proceeding of the International Symposium. p 635– 650.
Alvarez De Buergo M, Fort R. 2001. Basic methodology for the assessment and selection of water-repellent treatments applied on carbonatic materials. Prog Organ Coat 43:258 –266. Alvarez De Buergo M, Fort R, Go´mez-Heras M. 2004. Contributions of scanning electron microscopy to the assessment of the effectiveness of stone conservation treatments. Scanning 26:41– 47. Biscontin G, Frascati S, Marchesini L. 1975. Colour variations in old bricks and stones as a result of consolidation with resin. In: The conservation of stone I. Bologna, Italy: Proceedings of the International Symposium. p 741–747. Borselli G, Camaiti M, Pasetti A, Maravelaky P, Matteoli U. 1990. Protettivi impiegati nella conservazione dei materiali lapidei: storia, impieghi, sviluppi. L’Edilizia 1/2:67–78. Bousquie´ P, Pellerin FM, Struillou R, Arnould M. 1980. Contribution de la porosimetrie au mercure a l’etude de quelques proprietes de roches carbonatees. In: Wolters R, editor. Materials and engineering geology. Bull Int Assoc Eng Geol 22:225–232.
CONSOLIDANT DISTRIBUTION IN SANDSTONES (LSCM) Brus J, Kotlı´k P. 1996. Consolidacio´n de la piedra con mezclas de Alcoxisilanos y polı´meros acrı´licos. Cuadernos sobre conservacio´n 41:N°2. Camaiti M, Amoroso GG. 1997. Scienza dei materiali e restauro. La pietra: dalli mani degli artisti e delli scalpellini a quelle dei chimici macromolicolari Alinea Ed., Firenze, p 320. Dell’Agli G, Ferone C, Mascolo G, Marino O, Vitale A. 2000. Durability of tufaceous stones treated with protection and consolidation products. In: 9th International Congress on Deterioration and Conservation of Stone. Venice. p 379 –386. Dukes WH. 1972. Conservation of stone: chemical treatments. Architects’ Journal Information Library. p 433– 438. Esbert RM. 1993. The use of surface protectants: the current situation in Spain. Mater construcc 43:5–16. Esbert RM, Dı´az-Pache F. 1993. Inﬂuence of petrographical characteristics in the penetration of consolidants in porous building stones. Mater construcc 43:25–36. Esbert RM, Grossi CM, Valdeon J, Ordaz FJ, Alonso FJ, Marcos RM. 1990. Laboratry study for stone conservation at the Cathedral of Murcia. Mater construcc 40:5–15. Fredrich JT. 1999. 3D imaging of porous media using laser scanning confocal microscopy with application to microscale transport process. Phys Chem Earth 24:551–561. Garcı´a Garmilla F, Rodrı´guez-Maribona I, Cano M, Zalbide M, Iba´n˜ez-Go´mez JA, Osa-Chans E, Garı´n S. 2002. Comparacio´n analı´tica de dos consolidantes comerciales aplicados en areniscas eocenas de monumentos de los siglos XVI y XIX en San Sebastia´n (norte Espan˜a). Mater construcc 52:5–18. Go´mez-Gras D, Zoghlami K. 2003. Procedencia de las areniscas de la unidad superior de la Fm. Fortuna en el NE de Tu´nez. Geotemas 5:109 –111. Gon˜i J, Ragot JP, Sima A. 1968. Methode d’etude du champ microﬁssural des mineraux et des roches et possibilities d’application en geologie. Bull Bur Rech Geol Min Section 2 Geologie des Gites Mineraux 4:51– 86. Hudec PP. 1978. Rock weathering on the molecular level. In: Decay and preservation of stone, engineering geology case histories. Geological Society of America. Boulder, Colorado. 11:47–51. Mene´ndez B, David C, Darot M. 1999. A study of the crack network in thermally and mechanically cracked granite samples using confocal scanning laser microscopy. Phys Chem Earth 24:627– 632. Montoto M, Martinez-Nistal A, Rodriguez-Rey A, Fernandez-Merato N, Soriano P. 1995. Microfractography of granitic rocks under confocal scanning laser microscopy. J Microsc 177:138 –149. Paterno MC, Charola AE. 2000. Preliminary studies for the consolidation of Guadalupe tuff from the Philippines. In: 9th International Congress on Deterioration and Conservation of Stone. Venice. p 155–163.
Petford N, Miller JA. 1990. SLM confocal microscopy: an improved way of viewing ﬁssion tracks. J Geol Soc Lond 147:217–218. Petford N, Miller JA. 1993. The study of ﬁssion tracks and other crystalline defects using CSLM. J Microsc 170:201–212. Petford N, Davidson G, Miller JA. 1999. Pore structure determination using confocal scanning laser microscopy. Phys Chem Earth 24: 563–567. Piacenti F, Camaiti M, Mangalli del Fa` C, Scala A. 1993. Fluorinated aggregating materials for stone. In: Proceedings of the International Rilem/Unesco Congress. Conservation of stone and other materials. Hrsg. Thiel MJ ed. Großbritannien Bd. 2, S, London, p 740 –747. Pironon J, Canals M, Dubessy J, Walgenwitz F, Laplace-Builhe C. 1998. Volumetric reconstruction of individual oil inclusions by CSLM. Eur J Mineral 10:1143–1150. Rivas T, Silver B, Prieto B. 1998. Medida de la eﬁcacia de dos hidrofugantes aplicados a rocas granı´ticas. Mater construcc 48:5–24. Sasse HR, Honsinger D, Schwamborn B. 1993. “PINS” — new technology in porous stone conservation. In: Proceedings of the International Rilem/Unesco Congress. Conservation of stone and other materials. Hrsg. Thiel MJ ed. Großbritannien Bd. 2, S, London, p 705–715. Torraca G. 1975. Treatment of stone in monuments. A review of principles and processes in the conservation of stone I. In: Proceeding of the International Symposium. Bologna. p 297–315. Villegas R, Vale Parapar JF. 1993. Evaluacio´n de tratamientos de hidrofugacio´n aplicados a piedras de catedrales andaluzas. III. Ensayo de alteracio´n acelerada en atmo´sfera contaminada. Mater construcc 43:25–38. Villegas R, Martı´n L, Vale JF, Bello MA. 1995. Characterization and conservation of the stone used in the cathedral of Granada, Spain. Mater construcc 45:17–35. Wheeler G. 1992. Comparative strengthening of several consolidants on Wallace sandstone and India limestone. Proceedings of the 7th International Congress on Deterioration of Stone, Lisbon. p 1033– 1042. Young ME, Murray M, Cordiner P. 1999. Stone consolidants and chemical treatments in Scotland. Report to historic Scotland. Historic Scotland. Edinburgh. Longmore House p 298. ´ lvarez A. 2004a. Petrophisical characZoghlami K, Go´mez-Gras D, A terisation and durability of miocenic sandstones used in the roman aqueduct of Zaghouan–Carthage building. In: 6th International symposium on the conservation of monuments in the Mediterranean Basin. Lisbon. p 385–389. ´ lvarez A, De Luxa´n MP. 2004b. IntrinZoghlami K, Go´mez-Gras D, A sic factors that condition the physical behaviour and the durability of the Miocene sandstones used in the construction of the Roman aqueduct of Zaghouan-Carthage. Mater construcc 54:29 – 41.