Evaluation of cleaning methods for granite based on petrographic examinations

Transcription

1 Materials and Structures/Matériaux et Constructions, Vol. 29, April 1996, pp Evaluation of cleaning methods for granite based on petrographic examinations S. Pavía Santamaría, P. O Brien, T.P. Cooper University of Dublin, Trinity College, Dublin, Ireland TECHNICAL NOTES A B S T R A C T The decay of granite blocks in buildings results in microfracturing, mineral alteration and salt crystallisation. The conservation of granite historic buildings involves the application of cleaning methods to reduce these effects and to improve the general condition of the stone. Strong cleaners applied without any control have sometimes caused damage to the materials. This study determines the effect of various cleaning methods on microfractures, mineralogy and soluble salt content in granite. Samples of weathered granite from buildings in Trinity College, Dublin and unweathered freshly-quarried granite were cleaned by physical and chemical methods. The condition of the granite was assessed before and after cleaning by examining thin sections under a petrographic microscope. Fracturing, mineral transformation and soluble salt content were used as indicators of the condition of the stone. This paper concludes that the cleaning methods tested do not significantly alter the intensity and depth of fracturing in weathered granite. It also concludes that chemical cleaning does not cause significant mineralogical alteration to weathered granite and that salts remaining after cleaning are remnant weathering products and not reaction products. Finally, the paper highlights the effectiveness of a poultice containing detergent and chelating agents in removing salts from weathered granite. R É S U M É La détérioration des blocs de granit dans la construction entraîne des microfissures, une altération minérale et une cristallisation saline. La conservation des bâtiments historiques en granit implique l emploi de méthodes de nettoyage destinées à réduire ces effets et à améliorer l état général de la pierre. L application incontrôlée de produits de nettoyage décapants a également parfois endommagé les matériaux de construction. Cette étude a pour objet de déterminer l effet de diverses méthodes de nettoyage sur les microfissures, la minéralogie et le contenu en sels solubles du granit. Des échantillons de granit érodé en provenance de bâtiments de Trinity College à Dublin, ainsi que des échantillons de granit récemment taillé et non exposé aux intempéries, ont été nettoyés à l aide de méthodes physiques et chimiques. L état du granit a été évalué avant et après le nettoyage, par l examen de fines sections au microscope pétrographique. La fissuration, les transformations minérales et le contenu en sels solubles ont été utilisés comme indicateurs de l état de la pierre. On conclut que les méthodes de nettoyage essayées ne modifient pratiquement pas l intensité et la profondeur des fissures dans le granit érodé, et que les sels présents après le nettoyage sont les restes des produits de l action des intempéries et non des produits de réaction. Pour finir, cet article met en évidence l efficacité d un cataplasme à base de détergent et d agents chélatants pour éliminer le sel dans le granit érodé par des intempéries. 1. INTRODUCTION The decay of granite blocks in buildings has been likened to microfracturing determining breakdown into particle sizes greater than 63 um [1]. The significant role played by soluble salt crystallisation has been hypothesised [2], expounded vigorously [3] and demonstrated in the laboratory and field [4 and 5]. The conservator of granite historic buildings is thus provided with much information and evidence of the detrimental role played by microfractures, mineralogy and soluble salts but with little guidance on how to reduce these effects. Leinster granite is one of the main building stones used in the Dublin area. It was drawn from the nearby coastal and hill quarries in Dublin and Wicklow Counties, and used in most of the buildings constructed during the 17th century. The Georgian buildings of the historic west front of Trinity College are typical of the era with granite ashlar framed and trimmed by Portland limestone. The granite is a medium to coarse rock composed essentially of quartz (30%), feldspar (microcline 35%), Na and Ca plagioclase (25%), muscovite (7%) and biotite (3%). It contains minor amounts of sericite, kaolinite, chlorite, opaques, zooisite, tourmaline, rutile and sphene [6]. The /96 RILEM 185

2 Materials and Structures/Matériaux et Constructions, Vol. 29, April 1996 Table 1 Cleaners tested Cleaner name Composition and properties PHDR (ProSoco Heavy Duty 14.6% HF, strongly acidic masonry cleaner Restoration cleaner) not suitable for limestone or concrete DIJK HF HF P poultice Detergent and chelating agents (ProSoco T-1260 ph 8.8 poultice) dwell time hours ProSoco 766 (masonry prewash) solubilises crusts prior to application of restoration cleaner ph 14 contains NaOH > 5% needs to be neutralised dwell time 30 mins - 1 hour Stenex paste contains NaO anti-static additives suitable for limestone Murodex AR contains NaOH Monument paste EDTA (ethylenediamine tetraacetate) chelating agent Blasting (grit blasting) aluminium and silicone oxide particles Table 2 Scores of cleaning methods. ***-acid cleaner; **-alkaline cleaner; *-strongly-alkaline cleaner Cleaning method Fracturing Fracturing Mineral Salts and granite intensity depth (mm) transformation condition Fresh stone uncleaned none blasted none blasted + HF none Weathered stone uncleaned gypsum blasted gypsum PHDR*** calcite, gypsum DIJK HF*** gypsum DIJK HF*** gypsum P poultice** none ProSoco 766* gypsum Stenex paste* gypsum, anhydrite Murodex AR* gypsum Monument paste* gypsum porosity of unweathered, recently-quarried Leinster granite is approximately 0.03%, indicating that it should be resistant to water penetration and thus a durable building material. Despite this, the granite of the buildings of Trinity College exhibits many forms of severe decay, thereby indicating that it is susceptible to attack in an urban environment. The decay is typical of severely degraded granite with increased micro-fracturing leading to granular disintegration, mineral alteration and severe staining and salt encrustation. Strong acid and alkaline-based cleaners along with wet and dry blasting have been used extensively for the removal of encrustations and stains. The chemical and physical stresses associated with these processes are assessed by examining their effect on the nature and extent of fracturing, on changes in mineralogy and on the removal, formation and mobilisation of soluble salts. 2. MATERIALS AND METHODS Samples of weathered granite taken from the buildings of Trinity College and samples of freshly-quarried granite were examined before and after cleaning. The weathered granite was cleaned in situ using a range of chemical cleaners and one type of grit blasting see Table 1. The cleaners are described with reference to technical data sheets provided by the manufacturers or their agents. The chemical cleaners were chosen from a range of widely-used proprietary cleaners. The grit blasting was carried out in accordance with the method suggested in previous research [7]. Saftigrain, an aluminium and silicone oxide, was chosen because of its chemical stability. A small particle size was used to minimise the amount of energy dissipated at the surface of the stone. 50-mm diameter cores 13cm in length were taken using a water-cooled core drill before and after cleaning and sectioned for examination by petrographic microscope. Thin sections of about 30 microns were cut and polished without water in order to preserve water-soluble salts. The sections were examined using polarised-light petrographic microscopy and either cross or parallel nicols. The cleaning methods were assessed in terms of their effect on fracturing and on mineralogy. All observations were made in triplicate. Three sections from each sample were examined, and the results were used to calculate an average numerical evaluation of observed alterations. The results obtained are set out in Table RESULTS 3.1 Fresh uncleaned samples The fresh samples had a distinct low-intensity fracture pattern associated with primary fractures. The primary fractures were both inter-granular and intra-granular, very narrow with no observed separation. Original texture of granite shows extensive feldspar alteration (sericitization). Plagioclase was more altered than was K-feldspar. It also shows muscovite changing into chlorite. These mineralogical changes were caused by hydrothermal alteration during the last stages of the granite genesis. No salts were observed in these samples. 186

3 Pavía Santamaría, O Brien, Cooper 3.2 Fresh chemically-cleaned and blasted samples The sections from the fresh samples showed no significant alteration after cleaning. 3.3 Weathered uncleaned samples The weathered uncleaned samples showed extensive secondary fracturing with separation at grain boundaries. Mineral transformation included the greatly increased hydrothermal alteration with sericitization of plagioclase and K-feldspar as well as the formation of secondary minerals - mainly gypsum. A large proportion of the enlarged fractures were filled with gypsum. Gypsum crystals were found as flakes in some cracks and between altered mica plates. 3.4 Weathered chemically-cleaned samples Most of the samples exhibited medium to heavy intensity of fracturing. Gypsum, clay minerals and brown staining were found in the outer 3mm. Calcite and anhydrite were also present. Gypsum crystals were found in the fractures of all the samples (Figs.1 and 2) except those cleaned with ProSoco T-1260 poultice (P poultice). These samples showed no evidence of gypsum. 3.5 Weathered blasted samples These were similar to the chemically-cleaned samples (Fig.3). 3.6 Comparison of fractures There was a marked difference between the fresh and weathered samples. The fresh samples had a distinct low intensity fracture pattern associated with primary fractures. The primary fractures were both inter-granular and intra-granular, very narrow with no observed separation. The weathered samples had extensive secondary fracturing. These were more intense and wider. A clear separation could be observed (Figs.1-3). Primary trans-granular fractures were observed in quartz and feldspars. Secondary trans-granular fractures were observed mainly in feldspars and mica. Fig. 1 Enlarged fracture filled with gypsum flakes in PHDR cleaned granite. Gypsum flakes start breaking biotite plates. Left parallel nicols, right crossed nicols. 187

4 Materials and Structures/Matériaux et Constructions, Vol. 29, April 1996 Mineral transformation : 0- Granite showing primary hydrothermal alteration : Bright quartz and K-feldspar, but sericitized plagioclase and biotite changing into chlorite. 1- Increase in the hydrothermal alteration. Strongly sericitized plagioclase and K-feldspar. 2- Plagioclase and K-feldspar totally clouded because of sericitization ; opaque quartz; muscovite changing into kaolinite ; new formation of gypsum and/or calcite and/or anhydrite. Brown staining by iron release following biotite attack. Fracturing scores depend on the primary fracture pattern of the granite. Mineral transformation is related to the primary texture and degree of decay of the granite. 4. DISCUSSION AND CONCLUSION 4.1 Chemical cleaning Fig. 2 General view of granite after Monument paste cleaning. Fractures filled with gypsum ran parallel to the exposed surface. Up parallel nicols, down crossed nicols. The results obtained indicate that there is no significant difference in fracturing and mineral transformation between uncleaned and cleaned weathered granite. There was no evidence of salt formation resulting from the cleaning products. Weathering products (gypsum, calcite and anhydrite) remained embedded in fractures after cleaning in all samples apart from those cleaned with a poultice. No reaction salts (CaF 2, MgF 2, NaF, KF) were observed. 3.7 Scoring To assist with the assessment of the samples, the severity of fracturing and mineral transformation were scored as follows : Fracturing : 0- Either unfractured or showing primary cracks only. Primary cracks have virtually no measurable width. They sometimes include iron oxides. 1- Medium fracturing : Increase in the number and width of fractures in the outer 3 millimetres of the granite. Fractures affect mainly feldspars and micas. 2- Heavy fracturing (Figs.1-3) : Extensive fissuration, opening up of grain boundaries and occurrence of canals (very wide cracks). Opening up of mica plates. Some fissures occur linking micas. 4.2 Blasting Contrary to expectations, there was no evidence of increased fracturing in the blasted samples. The cracking observed appeared to be almost entirely the result of weathering. 4.3 Recommendations The results obtained point towards the selection of primary cleaning methods on the basis of other criteria. The cleaning options should be considered for each type of stone and then tested in restricted surfaces before carrying out the cleaning of the whole building. Effectiveness in removing stains is clearly a priority. Other factors such 188

5 Pavía Santamaría, O Brien, Cooper as material loss, effect on water uptake, time of wetness and brown staining should be investigated before and after cleaning. If these are deemed to be similar, environmental and economic issues may become the determinant factors. Whatever treatment is selected should be applied at minimal concentration/pressures to keep possible increases in secondary fracturing and mineral transformation to a minimum. Secondary treatment with detergent/chelating poultice should be considered as a means of removing salt deposits. REFERENCES [1] Haneef, S.J., Johnson, J.B., Jones, M., Thompson, G.E., Wood, G.C. and Azzaz, S.A., A laboratory simulation of degradation of Leinster granite by dry and wet deposition processes. Corrosion Science, 34, (3), (1993), [2] Rainey, M.M. and Whalley, W.B., 'The importance of microfractures in granitic rock breakdown with reference to building stone decay, in Granite Weathering and Conservation, Proceedings of a European Conference, Dublin, Ireland, 1993 (E. Bell and T.P. Cooper, Dublin, 1994), [3] Cooper, T.P., Dowding, P., Lewis, J.O., Mulvin, L., O'Brien, P., Olley, J. and O'Daly, G., Contribution of calcium from limestone and mortar to the decay of granite walling, in Science, Technology and European Cultural Heritage, Proceedings of a European Symposium, Bologna, Italy, June 1989 (ECSC-EEC-EAEC, Brussels- Luxembourg, 1991), [4] Ashurst, J. and Dimes, F.G., Conservation of Building and Decorative Stone, (Butterworths, London, 1995). [5] Haneef, S.J., Dickinson, C., Johnson, J.B., Thompson, G.E. and Wood, G.C., Simulation of the degradation of coupled stones by artificial acid rain, Studies in conservation, 37, (1992), [6] Warke, P.A. and Smith, B.J., Inheritance effects on the efficacy Fig. 3 Gypsum crystals filling transgranular open fractures in weathered blasted granite. Up parallel nicols, down crossed nicols. of salt weathering mechanisms in thermally cycled granite blocks under laboratory and field conditions in Granite Weathering and Conservation, Proceedings of a European Conference, Dublin, Ireland, 1993 (E. Bell and T.P. Cooper, Dublin, 1994), [7] O'Brien, P.F. and Cooper, T.P., Comparative study of granite cleaning techniques in Granite Weathering and Conservation, Proceedings of a European Conference, Dublin, Ireland, 1993 (E. Bell and T.P.Cooper, Dublin, 1994). 189