UNIVERSITA' DEGLI STUDI DI PADOVA
Sede Amministrativa: Università degli Studi di Padova
DIPARTIMENTO DI GEOSCIENZE
DOTTORATO DI RICERCA IN SCIENZE DELLA TERRA CICLO XX
CARBONATE LITHOTYPES EMPLOYED IN HISTORICAL MONUMENTS: QUARRY MATERIALS, DETERIORATION AND RESTORATION TREATMENT
Coordinatore: Ch.mo Prof. Bernardo Cesare Supervisore: Ch.mo Prof. Gianmario Molin Cotutori: Prof.ssa Lucia Baccelle Scudeler, Dott. Vasco Fassina, Prof.ssa Cristina Stefani
Dottorando: Simone Benchiarin
31 gennaio 2008
1
2
Ai miei genitori
3
4 RIASSUNTO Negli ultimi anni, la crescente sensibilizzazione verso il Patrimonio Culturale, ha determinato lo sviluppo di un’intensa attività di ricerca nell’ambito architettonico - monumentale costituendo una vera e propria “scienza della conservazione dei monumenti”, una disciplina basata su conoscenze di tipo storico - umanistico, scientifico e tecnologico. Numerose metodologie analitiche tipiche di scienze quali la geologia, la chimica, la fisica e la biologia sono state impiegate nello studio delle pietre impiegate in queste opere; lo scopo è quello di esaltarne il valore favorendo la loro conservazione attraverso idonee metodiche di preservazione e di restauro, adottate sulla base dei dati scientifici prodotti. In particolare, nello studio dei materiali lapidei impiegati in tale ambito, è possibile distinguere tre principali tematiche di ricerca, riguardanti rispettivamente la caratterizzazione dei materiali originali e di cava, lo studio delle cause e dei meccanismi di deterioramento nonché la valutazione dei prodotti e degli interventi di restauro. Il presente lavoro rappresenta un esempio di “scienza della conservazione” applicata allo studio di alcune problematiche riguardanti una serie di litotipi carbonatici largamente impiegati in Veneto grazie alla relativa facilità di lavorazione, a cui si sommano doti estetiche di pregevole valore: si tratta delle cosiddette “Pietre tenere” del vicentino e del Rosso Ammonitico Veronese. Capitolo 1. Questa sezione affronta l’analisi di alcuni calcari ampiamente usati nell’entroterra veneto sia in contesto storico-artistico che architettonico. Lo studio mira a definire le caratteristiche di questi materiali per comprendere il loro comportamento in termini di durabilità e con il fine di correlare il materiale di cava con quello delle opere. L’identificazione di materiali di cava consente inoltre l’individuazione di siti estrattivi per eventuali opere di sostituzione, qualora si rendessero necessarie. Il termine “Pietra di Vicenza” viene comunemente impiegato per indicare i materiali lapidei naturali provenienti dal comparto estrattivo dei Colli Berici, raggruppando sotto tale definizione pietre di età, caratteristiche e provenienza diverse. Cave attive e dismesse di questi materiali sono state individuate per mezzo di mappe e censimento dei siti estrattivi, a cui è seguito un campionamento sistematico. Le pietre sono state inizialmente suddivise in tre gruppi in base al loro colore: pietre bianche, pietre gialle e pietre grigie. Una dettagliata caratterizzazione minero- petrografica e chimica è stata condotta sui materiali provenienti dalle diverse cave, al fine di potenziare le conoscenze su questi materiali e identificare parametri significativi per definirne la provenienza e permettere la correlazione con le pietre impiegate nei monumenti. Due principali varietà all’interno delle pietre bianche sono state identificate: San Gottardo e Costozza. La prima è caratterizzata dall’importante presenza di alghe coralline di tipo incrostante (Litotamni), mentre la seconda per la presenza di significative quantità di matrice e di specifici foraminiferi bentonici: i Miliolidi. Il gruppo delle pietre gialle contiene due varietà: la pietra di Nanto, caratterizzata da grana medio-fine ed elevato grado di compattazione. L’associazione faunistica comprende foraminiferi bentonici (Nummuliti) e planctonici, echinodermi e serpulidi. La varietà San Germano di differenzia dalla precedente per la grana generalmente più grossa e l’assenza di foraminiferi planctonici, oltre che per la contemporanea presenza di Nummuliti e Discocicline. La varietà grigia è identificabile nella varietà San Germano denotando identiche caratteristiche ad eccezione della colorazione. Le diverse varietà sono state differenziate anche in base a composizione chimica, percentuale e composizione del residuo insolubile. Le informazioni raccolte costituiranno un data-base con lo scopo di disporre di informazioni sulle sorgenti dei materiali. Capitolo 2. Le proprietà petrofisiche (porosità, distribuzione percentuale del diametro dei pori, bulk density, capacità di imbibizione, assorbimento capillare) delle principali varietà precedentemente individuate e di due varietà di Rosso Ammonitico sono state determinate mediante test di laboratorio. L’assorbimento d’acqua per immersione totale e per capillarità sono stati eseguiti per valutare il comportamento idrico delle singole pietre; le pietre tenere hanno
5 evidenziato valori molto marcati rispetto al calcare del veronese. Sono state ricercate delle correlazioni tra caratteristiche intrinseche delle pietre e proprietà misurate. Test accelerati di laboratorio mediante simulazioni di gelo disgelo e cristallizzazione di sali sono stati condotti sui diversi litotipi con il fine di correlare le proprietà tessiturali con la loro durabilità. Una buona resistenza al gelo è stata osservata per quasi tutti i materiali. L’esame visivo dopo 16, 32 e 45 cicli ha rilevato l’assenza di un degrado significativo nei calcari del vicentino mentre alcuni campioni di Rosso Ammonitico evidenziano dei danni, soprattutto in corrispondenza delle discontinuità della roccia. Il coefficiente di imbibizione aumenta tuttavia in varie proporzioni in tutti i campioni all’aumentare dei cicli. Test di cristallizzazione mediante l’impiego di solfato di sodio sono stati eseguiti attraverso cicli di idratazione ed essiccamento, accertando correlazioni tra porosità e danno riscontrato. Capitolo 3. L’uso dei polimeri nell’ambito della conservazione lapidea è estremamente diffuso nelle operazioni di restauro. Attualmente, uno dei problemi che rappresenta un’innovazione scientifica e tecnologia in questo contesto, è la valutazione dell’efficacia dei trattamenti a base di polimeri organici eseguiti nel recente passato. Particolare attenzione va rivolta agli effetti indesiderati causati da questi prodotti sintetici sui vari materiali e la loro durabilità nel tempo. Fino ad oggi, solo alcuni studi sono stati condotti direttamente sulle opere interessate dagli interventi. Lo scopo di questa sezione è quello di valutare i trattamenti protettivi e consolidanti applicati su due monumenti della città di Padova, la Loggia Cornaro e la Stele della Minerva a diversi anni dal restauro. La loggia Cornaro, in pietra di Nanto, opera del 16° secolo dell’architetto Giovanni Maria Falconetto, è uno dei più importanti monumenti rinascimentali della città del santo. La Stele della Minerva, costruita nel 1941 da Paolo Boldrin su commissione dell’Università di Padova, è invece in Pietra gialla di San Germano. La Loggia Cornaro è stata restaurata due volte dal 1979 al 2003. In tutte le fasi il prodotto impiegato è stato una resina polisilossanica, ma il tipo è cambiato in funzione della disponibilità del mercato. Un metilfenilsilossano, denominato Rhodorsil 11309 è stato impiegato per il primo trattamento mentre il Rhodorsil RC90 per il secondo. Prodotti simili sono stati impiegati nella Stele della Minerva, in cui tuttavia è stato utilizzato anche il Rhodorsil RC70, un etilsilicato. L’esame macroscopico ha rivelato un buono stato di conservazione per la Loggia Cornaro, mentre le condizioni sono più critiche per la Minerva, con diffuse forme di degrado quali micro fessure, depositi superficiali e croste nere. Prove di assorbimento d’acqua per capillarità (Normal 44/93) sono state eseguite sulle superfici della Loggia per testare l’idrorepellenza delle superfici, ottenendo buoni risultati. Osservazioni al SEM delle sezioni trasversali dei campioni superficiali hanno permesso di constatare la presenza del polimero sia come strato superficiale sia come riempimento di fessure e pori in profondità. L’osservazione di campioni tal quali ha evidenziato le peculiarità morfologiche dei polimeri e la loro presenza su diversi porzioni di superficie. Sali solubili quali solfati, nitrati e cloruri sono stati accertati in diversi quantitativi. Le analisi FT-IR hanno permesso la caratterizzazione dei polimeri e lo studio di possibili interazioni polimero-substrato. L’analisi porosimetrica dei campioni trattati ha evidenziato i cambiamenti microstrutturali indotti dai trattamenti. I dati comparati con quelli dei materiali non trattati hanno evidenziato una minore porosità e uno shift della curva di distribuzione dei pori verso quelli di dimensioni minori. Questo fatto risulta più evidente per i campioni della Loggia, meno per quelli della Minerva. I risultati rivelano la presenza del polimero in vari spessori sulle superfici. La concentrazione di SiO2 generalmente decresce dall’esterno verso l’interno. Nel caso della Loggia Cornaro il trattamento ha prevenuto la formazione di nuove fasi, cosicché solo alcuni prodotti di degrado sono stati accertati. Uno stato conservativo peggiore è stato rilevato per la Minerva. Sulla base dei risultati ottenuti è stata suggerita l’esecuzione in tempi brevi di nuove applicazioni dei prodotti utilizzati, al fine di preservare al meglio le due opere.
6 SUMMARY Conservation of monuments is a cultural, artistic and technical activity based on humanistic and scientific knowledge. In recent years, the generalised deterioration of the Italian cultural heritage has led to the development of conservation science applied to monuments, i.e., the application of scientific methods to the study of stone materials. The aim of conservation science is to prolong the life of cultural heritage artefacts, and this goal is reached by reinforcing, through applied research, the preservation, conservation and restoration of monuments in order to enhance their aesthetic value. In the study of stone materials used in monuments, the main fields of research concern characterisation of original and quarry materials, study of the causes and mechanisms of deterioration, and evaluation of restoration products. The present work is an example of the conservation science of monuments applied to the study of various aspects of carbonate lithotypes often used in the Veneto Region (North-East Italy) for their aesthetic qualities and ease of working: the most frequently used are the so-called “soft stone” of the Berici Hills or “Vicenza Stone”, and “Rosso Ammonitico Veronese”. Chapter 1. This section concerns the limestone widely used in the Veneto Region, in both historical and architectural contexts. It aims at studying the characteristics of these materials in order to understand their behaviour in terms of durability, and at correlating materials from quarries and artefacts. Identification of the quarries of provenance aims at obtaining historical information on the exploitation of stone materials, and at finding new resources for substitution purposes. The name “Vicenza Stone” is given to a group of rocks with differing petrographical, mineralogical and chemical characteristics, and this section deals with examination of these varieties of bio-calcarenitic materials, which outcrop in the province of Vicenza and are quarried along the slopes of the Berici Hills. Thanks to their characteristic softness, they are relatively easy to extract and cut, and have therefore been used since ancient times as construction materials in local historic buildings. Ancient and active quarries were located by means of maps, identification, and a census of extraction sites. Mineralogical, petrographic and chemical study of representative samples from various quarries in the area was carried out. The stone was preliminarily grouped into three categories, based on colour: White, Yellow and Grey Vicenza Stone, and detailed minero-petrographic and chemical examination of the different stones was then made. For improved knowledge of these materials and in order to identify significant parameters regarding the provenance of the stone used for monuments, differences were sought between stone in any of the three main groups. Two main varieties of White stone were identified: San Gottardo and Costozza. The former is characteristic with respect to the latter due to encrusting coralline algae (Lithothamnion) and the latter due to the presence of one type of benthic foraminifera (Miliolids). The Yellow group contains two varieties: Nanto Stone, characterised by fine to medium grain size and an elevated degree of compaction, containing many fragmented fossils, classified as grainstone. Microfossils are benthic foraminifera (prevalently Nummulites) and echinoderms in equal proportions, planktonic foraminifera and bryozoans. Instead, San Germano Stone contains prevalent benthic foraminifera (entire Discocyclinas and Nummulites, the former generally prevailing), echinoderms and bryozoa. An Ash-Grey variety may also be found in San Germano Stone and, in this case, is called Ash-Grey San Germano Stone. Nanto Stone and San Germano Stone belong to different formations. The several varieties were also characterised by percentage and composition of insoluble residue. All information collected will go towards making up a data-bank aimed at further knowledge and documentation of stone resources. Chapter 2. The petro-physical properties (porosity, bulk density, imbibition capacity, capillary absorption) of the main types of limestone described above were analysed and evaluated by various methods. Two varieties of Rosso Ammonitico Veronese were also considered. Porosity was evaluated by mercury porosimetry, and differing values were found for some of the soft Vicenza types. However, the values are very high when compared with those of Rosso
7 Ammonitico (average 0.8%). Total pore size distribution showed particular trends for the main varieties. Water absorption, measured by total immersion and capillarity, was carried out on three samples for each variety (Normal 7/81). Soft Vicenza stone had imbibition coefficients and water absorption values markedly higher than those of Rosso Ammonitico. Correlations between these measured properties were made. In order to improve the properties of stone, to understand its behaviour in terms of durability, and to define correlations between intrinsic characteristics and durability, limestone deterioration was simulated by laboratory tests, including freeze-thaw cycles and crystallisation of soluble salts. Generally good performance in terms of durability during freeze/thaw tests emerged for all samples. Macroscopic inspection after 16, 32 and 45 cycles revealed good frost resistance for all varieties of soft stone from the Berici Hills, whereas some samples of Rosso Ammonitico showed damage near discontinuities, worsened by repeated cycles. Imbibition capacity increased with number of cycles in all samples. Salt crystallisation tests with sodium sulphate were carried out by hydration/drying cycles and showed a correlation between porosity parameters and damage. Chapter 3. The use of polymers to protect stone materials from the effect of decay is widespread in the field of conservation of architectural monuments. Today, one of the problems which represent a scientific and technological challenge in this sphere is evaluation of organic polymer treatment carried out in the recent past. Particular attention should be devoted to the side-effects caused by these synthetic products on various materials, and their durability. Until now, only a few studies on the in-depth behaviour of such polymers have been carried out on artistic monuments. The aim of this section is to evaluate the consolidation and protective treatment applied to two monuments in Padova, the Loggia Cornaro and Stele di Minerva, several years after restoration operations. The Loggia Cornaro, built in Nanto Stone at the beginning of the 16th century by the architect Giovanni Maria Falconetto, is one of the most important historical buildings in Padova. The Stele di Minerva was constructed in 1941 by Paolo Boldrin in Yellow San Germano Stone. The Loggia Cornaro was restored twice from 1979 to 2003, and treatment was effected in several stages. In all phases, a polysiloxane resin was used, although the actual type changed according to market availability. A methyl phenyl siloxane, trade name Rhodorsil 11309, was used for the first treatment, and Rhodorsil RC-90 for the last one. Similar products was also used for the restoration of the Stele di Minerva in 1995, where Rhodorsil RC70 was also employed. Macroscopic examination revealed a good state of conservation for the Loggia Cornaro and a worse one for the Stele di Minerva, and forms of deterioration such as microcracks, superficial deposits and black crusts had spread to a great extent in the latter. Water capillarity absorption (Normal 44/93) was used to assess the penetration of water into the bulk of the stone materials of the Loggia Cornaro, and revealed good water-repellent properties of surfaces. SEM observations of thin cross-sections of stone fragments revealed the presence of the polymer, both as superficial flakes and as infill for cracks and pores at depth. SEM (SEI) analysis of 3-D samples showed that the stone surfaces were often covered by the polymer, more evidently in the Loggia Cornaro. Ion chromatography revealed salts such as sulphates, nitrates and chlorides in differing amounts. FT-IR was used to estimate the depth of penetration of the polymer and the possible formation of bonds between product and substrate. Mercury intrusion porosimetry evaluated micro-structural modifications inside the stone due to consolidation of applied materials. These data were compared with those of quarry materials, and revealed lower porosity and a higher number of micropores in the Loggia Cornaro samples, less evident in the Stele di Minerva. Results highlighted various depths of penetration into the stone; the concentration of Si generally decreased from the external surface inwards. For the Loggia Cornaro, treatment had prevented the formation of new phases, so that only a few decay products were observed. Conditions are less good in the Stele di Minerva. Systematic maintenance, with further applications of the protective coating on the whole surface, should be carried out in the near future.
8 Note to the thesis As the chapters of the thesis are to be viewed as “independent” scientific papers, there is, inevitably, a certain degree of repetition.
ACKNOWLEDGEMENTS Special thanks are due to the persons who collaborated in this thesis: Prof. Gianmario Molin, Prof. Lucia Baccelle Scudeler, Dr. Vasco Fassina and Prof. Cristina Stefani. Thanks go to the Soprintendenza per i Beni Storico-Artistici ed Etnoantropologici del Veneto and the Soprintendenza per i Beni Architettonici e Paesaggistici del Veneto and to the owners of the Vicenza Stone and Rosso Ammonitico quarries. Lastly, appreciation is expressed to all those working in the Dipartimento di Geoscienze of the University of Padova.
9
10 CONTENTS Riassunto 5 Summary 7 Note to the thesis and Acknowledgements 9 Contents 11
Chapter 1 “Vicenza Stone”: mineralogical-petrographical and geochemical definition of the main varieties of “Soft Stone of the Berici Hills”. Parameters for identification. 13 1. Introduction 13 2. Berici Hills: Geographical Location, Geomorphology and Tectonics 15 3. Stratigraphical Setting 16 4. Quarry Stratigraphy 21 5. Quarry Activity 21 6. Quarry Locations 22 7. Analytical Procedures and Methods 24 8. Preliminary Differentiation 24 9. Results and Discussion 26 9.1 Petrography 26 9.2 Evidence of Diagenesis 33 9.3 Palaeo-Environmental Interpretation of Microfacies 36 9.4 Mineralogical Composition 37 9.5 Chemical Composition 39 9.6 Geochemistry and Palaeo-Redox Information Data 39 10. Discussion 43 11. Conclusions 45 References 46
Chapter 2 Durability of carbonate lithotypes employed in historical monuments of Veneto (North-East Italy): a laboratory study. 49 1. Introduction 49 2. Materials and Methods 50 2.1. Types of Limestone Studied Here 50 2.2. Analytical Techniques and Methods 51 2.2.1. Petro-Physical Properties 51 2.2.2. Experimental Laboratory Weathering Simulations 53 2.2.3. Measurement Strategy 55 3. Results and Discussion 55 3.1. Characterisation of Lithologies 55 3.2. Pore Structure 65 3.3. Water Absorption Properties 69 3.4. Durability Tests 77 4. Conclusions 94 References 96
11 Chapter 3 In situ evaluation of restoration treatments on Nanto Stone and Yellow San Germano Stone in two important monuments in Padova: Loggia Cornaro and Stele di Minerva. 101 1. Introduction 101 2. Synthetic Products 103 3. Analytical Techniques 104 4. Loggia Cornaro 106 4.1. Historical Notes 106 4.2. History of Treatment 107 4.3. Sampling 107 4.4. Preliminary Work 108 4.5. Results 110 4.5.1. Site Test – Water Capillarity Absorption Results 110 4.5.2. Minero-Petrographic and Chemical Characterisation of Stone 111 4.5.3. Porosity and Results of Pore Size Distribution analysis 117 4.5.4. Drilling Techniques 118 4.5.5. Polymer-Substrate Interactions 120 4.5.6. SEM Observations 122 5. Stele di Minerva 136 5.1. Historical Notes 136 5.2. Restoration 136 5.3. Preliminary Activities 137 5.4. Sampling 144 5.5. Results 145 5.5.1. Site Test – Water Capillarity Absorption Results 145 5.5.2. Minero-Petrographic and Chemical Characterisation of Stone 145 5.5.3. Porosity and Results of Pore Size Distribution analysis 147 5.5.4. Deterioration Products 148 5.5.5. Study of Polymers 150 5.5.6. SEM Observations 152 6. Discussion 162 7. Conclusions 166 References 169
12 CHAPTER 1
“Vicenza Stone”: mineralogical-petrographical and geochemical definition of the main varieties of “Soft Stone of the Berici Hills”. Parameters for identification.
1. Introduction In recent years, growing concern in Italy’s Cultural Heritage and in particular the safeguarding of the architectural and monumental patrimony has aroused great interest in the ornamental stone employed in these artworks - above all because their integrity is severely threatened by sometimes advanced deterioration processes. The conservation, maintenance, restoration and substitution of stone artefacts, both worked pieces and buildings, or entire countries or historical centres of cities, require not only historical- artistic studies but also proper knowledge of the materials used, in relation to their provenance, chemical-mineralogical-petrographic characters, physico-mechanical properties, and state of alteration. Such knowledge about Vicenza Stone is generally incomplete, so that systematic analysis of these materials is appropriate. Although types of limestone often have similar appearances and properties, their varying geological background may give rise to differences in colour, mineral composition, granulometric properties, pressure strength and/or deterioration behaviour. Because of the close relation between the petrography of limestone and its physical properties, sound assessment of its quality, apart from very general statements, is impossible without a minimum of petrographic data. For instance, the causes and mechanisms of stone deterioration can only be studied on the basis of good knowledge of mineral composition and texture (including pore structure, pore-filling materials, etc.) of the building stone in question. Such petrographic properties of a stone influence deterioration processes at least to the same extent as environmental factors (climatic conditions, air pollution). At the same time, in studying the causes and mechanisms of the decay of stone monuments and/or possible conservation products which can be applied as supports or to avoid further deterioration, it sometimes necessary to have available stone with similar characteristics to that used in the monuments, in order to carry out laboratory studies (tests of accelerated alteration simulating environmental conditions), and also with the aim of possible restoration work involving replacement of badly damaged areas. The ideal situation would be to use stone from the same quarries which supplied the original stone but, failing this, if that stone is not accessible or its quarries no longer exist, it is necessary to use stone from other quarries with similar properties. In any case, a comparative study should be made of the stone both from the quarry and from the monument, in order to establish similarities. Italy has a wide range of calcareous materials which, thanks to their ample availability, aesthetic qualities, and ease of working, are optimal not only for building but also for creating monumental artworks. The architecture of many Italian cities has been deeply influenced by the stone found in or near them; in many cases, its extensive use made it a peculiar feature of the city or the nearby area, at least within a well-defined historical period. Limestone has been a traditional building material worldwide for centuries and is still quarried for this purpose today. It has been extensively used, mainly for decorative purposes, throughout the history of the Veneto region. The centres of cities such as Vicenza, Padova, Verona and, marginally, Venice, were built mainly with materials from local quarries (Rodolico, 1963). Within the general context of extraction of Veneto stone, “Vicenza Stone”, in its different varieties, thanks to its characteristic softness, aesthetic qualities and ease of working, has been of great importance to regional development, and has made this region known all over the world. Indeed, it boasts a centuries-old tradition in the culture and history of the Veneto. Its
13 widespread occurrence, together with its good technical features, has made this type of limestone one of the building stones in greatest demand in Veneto architecture. Although it is mainly used locally, it is also exported to Germany, France and the USA. The limestone outcropping in the Berici Hills (province of Vicenza, North-East Italy) has been extensively quarried since Palaeo-Venetic times, although its large-scale exploitation became very important only with the Romans, and has continued until today. The impact on the Berici Hills has been devastating over time and, in order to lessen the impact on the environment, quarrying has recently been limited to only a few sites. During the Middle Ages and the Renaissance, huge volumes of this stone continued to be quarried and employed in the Veneto and in the Po Plain; its features were particularly appreciated in those times, when the possibilities of stone cutting and working were more limited than they are nowadays. Several of the “Ville Venete” designed by the architect Andrea Palladio probably represent the main proof of how this stone was exploited. Various kinds of “Vicenza Stone” were also used in Padova (e.g., the statues in Prato della Valle, Loggia Cornaro, and St. Anthony’s Basilica). In general, their easy workability also favoured their use for sculptures (Chemello and Fabiani, 1929, 1939; Cornale, 1994). In spite of its extensive use, little attention has been devoted to studying the mineralogical, petrographical, chemical and petro-physical features of this stone and the interpretation of its behaviour when used as building or ornamental material. The name “Vicenza Stone” is now given to a series of bio-calcarenitic materials, grouping under this name rocks with differing colours, ages, provenance, and petrographical, mineralogical and chemical characteristics. The complex historical vicissitudes and scientific lack of material characterisation of the soft stone from the Berici Hills have given rise to a variety of names that often complicate its identification. Several attempts (Cattaneo et al., 1976; Cornale, 1994) have been made to use detailed mineralogical and petrographical data to study its provenance and properties, but they are not exhaustive. The same type of stone is often known by several names, related to various phases and corresponding competences under which the material passed (quarry operators, stone-cutters, architects, collectors). Two main causes have influenced the nomenclature of the Berici stones. The oldest names frequently derive from the names of the nearest locality; others have been added in recent periods, describing their aesthetic and textural characters. This process has led to the use of a large number of names such as “Costozza Stone”, “San Gottardo Stone”, “Nanto Stone”, “Yellow Nanto Stone”, “San Germano Stone”, “Zovencedo Stone”, etc. (Pieri, 1966), none of which can be associated with a certain stone with certain characteristics, adding to problems of nomenclature and identification. Another problem is that limestones from different localities in the Berici Hills often show similar features in terms of macroscopic characters, especially colour, making it hard to discriminate between them. The aim of this research is improved knowledge of these stone materials, in order to define some main varieties through identification of significant features as regards determination of provenance and, at the same time, some diagnostic petrographical, chemical and mineralogical indicators to establish the quarry provenance of various limestones and the area of the original outcrop of each lithotype. Petrographic information is also needed for more detailed classification of monotonous limestones, which is necessary in the case of stone replacement. Many quarries from various localities in the Berici Hills were examined, to provide evidence in discriminating unequivocally between various limestones in terms of microscopic properties. In this light, study of their stratigraphic and microfacies characteristics, diagnostic fossils in facies, depositional and diagenetic textures, quantitative-qualitative determination of insoluble residue and chemical composition was undertaken. These parameters are used to correlate materials of monuments to quarry materials, allowing identification of their provenance.
14 The results of the project will be well-reasoned knowledge at the disposal of stone dealers and restorers, so that stone for producing new artefacts can be correctly chosen for its planned use, quarried respecting the environment, corresponding to local historical traditions, and with good durability.
2. Berici Hills: Geographical Location, Geomorphology and Tectonics The Berici Hills form an isolated rise in the Po Plain south of Vicenza, characterised by a karstic plateau (Figs. 1, 2).
Fig. 1. Geographical setting of Berici Hills and nearby areas.
In plan, they are a parallelogram-shaped relief about 24 km long, with its main axis oriented SE- NW, developing over an area of approximately 165 km2. The Hills are bounded by steep slopes to the east; a gentler morphology occurs in the southern sector. The Berici Hills are cut by the Liona valley and by the Fimon valley system, which divide the area into two distinct sectors: a wide plateau, very abrupt along the S-W margin, and a western sector characterised by gentler morphology, with slight undulations. The maximum altitudes of the Hills are found in the first sector, with Monte Alto (444 m a.s.l.), Monte della Cengia (428 m a.s.l.) and Monte Tondo (415 m a.s.l.).
15 As explained in detail later, the Berici Hills essentially consist of carbonate rocks, whose role is particularly important in the development of karst phenomena. The morphological appearance of the Hills is linked to litho-stratigraphy, involving a calcareous-marly unit, easily erodible and outcropping extensively in the western area, and a calcareous unit forming a large portion of the eastern area (Mietto, 1988). From the tectonic point of view, the area is characterised by sub-horizontal to slightly inclined strata, with small folds (running SE-NW). Those cut by faults, running perpendicularly from the previous ones, show prevalently horizontal movements (Benvenuti and Norinelli, 1967). Structurally, the area is very simple, although tectonic activity has produced an upthrust to the south-west sector (containing the oldest units of the stratigraphical succession, Nanto and Sossano); the NE sector has more recent units.
Fig. 2. Berici Hills.
3. Stratigraphical Setting The Berici Hills, in the southern Vicenza plain, are considered the south-western extension of the Lessini Hills. They show stratigraphical analogies with the Palaeogene formations of the Lessini Shelf (Corsi and Gatto, 1967; Mietto et al., 1981; Luciani, 1989), a palaeo-geographic unit of the Southern Alps (Bosellini, 1989). The stratigraphical succession mainly consists of carbonate deposits and igneous rocks (basalts), ranging from the Late Cretaceous to the Early Miocene (Piccoli, 1966; Mietto, 1988). The stratigraphy of the area has recently been assessed by Bassi et al. (2000) (Figs. 3, 4). The most ancient formation is the “Scaglia Rossa”, Upper Cretaceous in age, which consists of thickly stratified, fine-grained, red or pink limestone, sometimes containing white or red flint strips, with planktonic foraminifera (Globigerinoids, Globotruncane), revealing a pelagic environment. The formation outcrops in isolated ribbons only at the base of the flank between Sossano and Nanto. At the end of the Cretaceous, the physico-chemical characteristics of the sedimentation environment radically changed, probably due to tectonic movements, precursors of the Tertiary eruptive cycle in the eastern Veneto (Piccoli, 1966), so that sediment deposition was suspended and Palaeogene deposits are not found in this area. The Berici Hills have a large stratigraphic hiatus that includes the Palaeocene (Massari et. al., 1976), as testified by the presence of hardgrounds at the bottom of the Scaglia Rossa. During the Lower Eocene in the Lessini area, basalt-like volcanic manifestations began,
16 persisted into the Middle Eocene, and then resumed in the Middle Oligocene, also involving the
Fig. 3. Stratigraphical section of Berici Hills (from Cornale, 1994, mod.). Berici Hills. Volcanic activity was directly linked to tectonics and, more precisely, to the presence of the “Alpone-Chiampo graben”, a rift which received all volcanic and volcano-clastic materials produced in situ or in nearby areas. This rift, lengthened to the NNW-SSE, started its
17 action at the end of the Late Cretaceous, and persisted throughout the Middle Eocene, allowing the accumulation of basaltic vulcanite in which calcareous sedimentary rocks of the ancient Eocene sea, known as “Marmi di Chiampo” are inserted (Mietto, 1988). The graben is bounded by deep fracture lines (Barbieri, 1972), particularly west of the Castelvero fault running NNW- SSE, and north of the pedemontane line running ENE-WSW. The eastern margin is not directly observable, but it presumably runs NNW-SSE (along the alignment Castelgomberto - Montecchio Maggiore - Bocca d'Ansiesa - Val Liona) (Mietto, 1988). The eastern Berici Hills thus constitute the south-eastern part of the Alpone-Chiampo graben, so that the quarrying area of soft stone can be geologically divided into two main sectors, each characterised by several stratigraphic differences from the Middle Eocene onwards.
Fig. 4. Stratigraphical diagram of eastern and western Berici Hills (from Girardi and Mezzalira, 1991 mod.). In this period, the eastern part of the Berici Hills was influenced by the volcanic-tectonic activity of the Alpone-Chiampo graben, which gave rise to a shallow-water sedimentation environment, strongly influenced by terrigenous and volcanic-detrital inputs (Mietto, 1988). This is revealed by the presence of a carbonate-marly arenaceous complex of Middle Eocene
18 age between Villaga and Nanto (Massari et al., 1976), inside which, nearer Nanto, a specific yellow carbonate arenaceous facies - Nanto stone p.n. (properly named) - was quarried as precious ornamental stone. Gradually, an instable marine environment, characterised by areas with slow bottom movements, became a shallow sea (40-60 m deep) with epineritic sedimentation and warm water (Girardi and Mezzalira, 1991). In its bed organisms typical of this environment such as Nummulites, Discocyclinas, echinoderms, red algae, gastropods, bryozoa and molluscs were deposited (Broglio Loriga, 1982). This led to biocalcarenite rich in macroforaminifera, forming the Calcari Nummulitici Formation, about 100 m thick in the Liona valley, where it is still quarried as building stone (Mietto, 1988). This is the San Germano Stone p.n.. In the sea that covered the entire Berici area, numerous volcanoes were concentrated to the east inside the graben, which had given rise to basaltic and tufitic deposits, locally forming shallow lagoons and brackish lakes. Volcanic and volcano-clastic rocks belonging to this eruptive phase outcrop in the eastern Berici near Sarego, Lonigo and Brendola. At the end of the Middle Eocene, this area emerged as a consequence of the longest accumulation of volcanic products. Instead, in the eastern sector, above an ideal line joining Alonte, Spiazzo and Grancona, the sedimentary sequence was not interrupted, so that the Luteziano-Bartoniano-Priaboniano stratigraphic succession continues (Girardi and Mezzalira, 1991). In that period, the eastern Berici area was a strip of land lapped by the sea both to the west, where the volcanic products of the Alpone-Chiampo graben were deposited, and to the east, where the Calcari Nummulitici accumulated. Subsequently, in the Late Eocene (40-36 m.y., Priabonian), a new transgressive phase started, during which the sea again occupied higher land above sea level, originating a new sedimentary cycle. The transgression started from the east, an shown by a particular level in the eastern Berici, which is a proper basal conglomerate: the “Cerithium diaboli” horizon, near Meledo. It is composed of yellowish calcarenite characterised by many minute pebbles, originating from the disintegration of basaltic soil and rich in coastal environment fossils (Bartolomei, 1958). Following the deposition of the Priabona Formation, a marly calcareous complex lies directly on the Calcari Nummulitici Formation in the eastern Berici. It corresponds to a sedimentation environment composed of a wide platform covered by a warm shallow sea, muddy and rich in organisms, in which some terrigenous inputs most probably come from the present-day area of Recoaro and Valdagno. In some sectors of the basin, where these inputs were limited or absent, colonies of calcareous algae (Melobesie) grew (Ungaro and Bosellini, 1965). The thickness of the Priabona Formation sometimes reaches 200 m. Its palaeontological contents include foraminifera such as Nummulites and Discocyclinas, molluscs, gastropods, bryozoans, corals and echinoderms. The highest part of the formation is composed of a clay complex named “marl with bryozoa” outcropping near Brendola and Montecchio Maggiore. In the Oligocene (36-30 m.y.), environmental conditions changed again, and a carbonate platform with coral and algal reefs developed in the Berici Hills area, as testified by the Calcareniti di Castelgomberto Formation, a very pure calcareous complex, white or pale yellow in colour, in which Sclerattini corals, hydrozoa, foraminifera and algae are frequent. Many studies have been carried out on the Oligocene reef complex (Rossi and Semenza, 1958; Geister and Ungaro, 1977; Ungaro, 1972), but the best interpretation is that of Frost (1981) who suggested the presence in the Oligocene of an extensive lagoon, bounded to the NW by the pedemontane Valdagno area and the Recoaro band and, to the SE, by a coral-algal reef, from Lumignano to Mossano (Figs. 5, 6). Beyond the coral reef that corresponds to the SE margin of the Berici Hills, the sea extended where the Oligocene deposits correspond to the Euganean marl. Oligocene rocks are lacking west of the lagoon in the Lessini Hills, and this supports the idea that the area emerged in the east near Marostica. A lagoon very similar to that described above
19 occurs in the reef complex of Queen Cay, along the Caribbean coast of Belize (Frost, 1981). A similar sedimentary environment produced three main types of rock in the Calcareniti di Castelgomberto Formation: . A massive limestone reef; . A back-reef with Nullipore calcarenite, generated in a high-energy environment near tide channels, which created the typical facies of Vicenza Stone, according to Mietto (1988); . Lagoon back-reef calcarenite, a typical facies of the Calcareniti di Castegomberto Formation.
Fig. 5. Reconstruction of Oligocene lagoon in Berici area (Frost, 1981).
Fig. 6. Geologic section of Berici Hills (Frost, 1981).
20 The massive limestone corresponds to biohermal structures with prevailing corals, divided by tidal channels that allowed oxygenation of the back-lagoon environment and proliferation of benthic organisms. The back-reef limestone is stratified and less compact than that of the reef, and originated near tide channels in high-energy areas. It is composed of very white Nullipore calcarenite with calcareous algal fragments. These rocks may also form inside the lagoon. The lagoon back-reef calcarenite is composed of white and straw-yellow calcareous rocks, stratified and very rich in fossils such as foraminifera, echinoids, bryozoans, molluscs, corals and benthic organisms. At the end of the Oligocene, volcanic activity resumed and produced many necks. In the Late Oligocene (30-24 m.y.), the lagoon was progressively filled in, and a fine-grained calcarenite grew until its complete emersion. In the Early Miocene (24-18 m.y.), the environmental conditions changed yet again, and a new discontinuous marine succession, closed with siliceous sands, probably aeolic, was deposited. Above this are the “Arenarie di S. Urbano”, formed of coarse-grained calcareous arenite, stratified and white to yellow in colour, with abundant fossils such as echinoids and Lepidocyclinas, characteristic of shallow-water deposits. They outcrop prevalently in the eastern Lessini area. The marine transgression continued until the Middle Miocene, with deposition of marly limestone and green marly clay (due to the presence of glauconite) called the “Marne argillose del M. Costi”, and outcropping in the Lessini area. Later, the Alpine orogeny caused the complete emersion of these formations, which have gradually been shaped by erosion.
4. Quarry Stratigraphy The quarrying area of the “Berici Hills soft stone” appears on sheet 49 (Verona) and sheet 50 (Padova) of the Geological Map of Italy (scale1:100000), and comprises the Berici Hills and the eastern part of the Lessini Hills. The “soft stone” quarries are located at various stratigraphic levels but inside two main geological formations of different age: “Calcari Nummulitici” and “Calcareniti di Castelgomberto”. Nanto Stone forms 40-150-cm thick beds in a particular carbonate-marly arenaceous complex at the base of the Calcari Nummulitici Formation (Fig. 4). The quarries are located at about 80 m a.s.l. near the village of Nanto (eastern sector). San Germano Stone p.n. is quarried at Pederiva di Grancona (central-western Berici Hills) at about 120 m. a.s.l. and belongs to the Calcari Nummulitici Formation (Fig. 4). “Vicenza Stone” is quarryied over a large area, and all quarries are included in the Calcareniti di Castelgomberto Formation (Fig. 4), extending as far as the Lessini Hills, where it displays lagoonal facies. This change means less quarrying, since the well-known facies is a back-reef, in a particular level generally located between 250 and 350 m a.s.l.. The thickness of the quarried bed does not reach 15 m (Cornale, 1994).
5. Quarrying Activity The ornamental stone of the Berici Hills generally lies under a cover, so that mining is necessary. There were some open-pit quarries in the past, but they are now all closed. Mining allows the environmental impact to be reduced. Quarrying is carried out by machinery, which cuts the stone without producing violent vibrations which could damage it. Particular attention is paid to the geometrical development of quarries, to obtain maximum profit, although static requirements and proper functional methods of excavation are respected. Quarry design provides for the development of large “rooms” (maximum width 8 m) excavated from top to bottom. Quarrying methods mainly depend on the desired size and shape of the stone and its physical characteristics. For limestone, since explosives cannot be used (because they would disturb
21 adjacent workings or produce instability), pneumatically operated channellers are used. After vertical cuts have been made, gadding machines (working on the same principle) are used to make horizontal cuts. Wedges are then split off long blocks, which are cut and removed (Cornale, 1994).
6. Quarry Locations Prior to field sampling, some present-day and ancient quarries were identified on the basis of geologic surveys, oral histories, historical observations, and of course direct field surveys. Field inspections are a necessary and vital element in any quarry study, principally because of the misleading or vague information transmitted through oral histories and historical reminiscences. Local tradition, for example, may state that a specific site was a quarry. Other factors that influence actual field identification of possible sources of stone are vegetal cover, soil development, precipitous landforms, and subjective preconceptions as to what an ancient quarry might look like. In some instances, a quarry may already have been exhausted, obscured by later workings, or weathered, rendering it unrecognisable as such and thus facilitating its integration into the natural landscape. In other instances, a quarry may actually have exploited secondary deposits, where raw materials are widely dispersed; evidence of quarrying, for example, occurs in the form of fairly well-defined open-pit mines. Evidence of use and extraction is, of course, a necessary condition in any provenance study. But not all quarries retain such obvious features as pit mines. Other indications of exploitation are the form of tool-marks left on rock outcrops, large blocks of stone removed from nearby outcrops, work stations in adjacent areas and established on levelled platforms within the quarry tailings or excavated in the outcrops, and local stock-piles of raw material. All these features, singly or in combination, comprise the unmistakable although sometimes subtle evidence of a quarry. Once identified as such, an efficacious sampling programme focused on provenance determination can quickly be developed. In sampling rock bodies, it is necessary to take enough samples to obtain an estimate sufficiently precise or representative of the deposit. During a systematic survey of quarries in the Berici Hills, approximately 30 quarry sites were visited, almost always subdivisible into local sites, each exploited during various historical periods. In this way, about 50 quarry sites were listed. Examination of ancient quarries was limited to a few well-preserved sites considered most significant for the purposes of this study; more recent sites were also examined and, although only numbering a few, are representative of the main stone types quarried in the area. Samples were taken from ancient and present-day quarries in the Berici area, near the villages of Costozza, Nanto, Grancona, San Gottardo, San Germano and Zovencedo, from which the names of the extracted stone are derived. In the case of sub-varieties within one outcrop, samples of each of the varieties were taken. These materials should allow more detailed classification and provide more information for a database of all limestone varieties. No stone quarried in these areas and not used for building but for other purposes (e.g. cement factories, etc.) is considered here. Sampling was carried out on lithotypes of Palaeogene formations belonging to the Veneto succession. The 20 quarries sampled for limestone characterisation are shown in Figure 7. Sites where quarrying was particularly important were sampled in detail, in order to demonstrate possible lithologic variability within individual quarries.
22
Fig. 7. Location of quarries (from Ginevra et al., 1999, mod.)
White Stone Nanto Stone Yellow San Germano Stone Grey San Germano Stone
23 7. Analytical Procedures and Methods To study limestone from differing localities, which often show very similar properties and are difficult to distinguish, a specific method, carried out on the basis of macroscopic and microscopic analyses of the collected samples, was developed, to complement conventional petrographic limestone analysis (structural and textural properties, types and abundance of single components, etc.). This procedure is mainly based on the combination of macroscopic rock description, geochemical analysis, X-ray diffraction analysis, thin-section polarising microscopy, qualitative and quantitative determination of the non-carbonate fraction, and porosity analysis. The combination of all these data allows complete characterisation of the materials.
Optical microscopy Observations of thin sections by U-polarised microscopy at different magnifications were carried out to define textural parameters. Three perpendicular thin sections for each sampling point were prepared in order to evaluate possible variations in structure. The percentages of various skeletal grains were estimated by means of comparative charts developed for limestone by Baccelle and Bosellini (1965). The petrographic classification adopted is that of Dunham (1962), with modifications by Embry and Klovan (1971).
Insoluble residue - sample preparation Each sample was treated in order to extract its insoluble residue as follows: carbonate was dissolved in hydrochloric acid (8%); the resulting solution was filtered, and the insoluble fraction thus obtained was dried at 60°C. Quantitative and qualitative data on composition were obtained.
X ray powder diffraction X-ray powder diffraction (XRPD) was performed on bulk samples and insoluble residue by means of a Philips X’Pert powder diffractometer, with Ni-filtered Ka radiation of Cu from a graphite monochromator. Measurements were performed at 40 kV and 40 mA. In particular, development of the Rietveld refinement allows study of complex mineral compositions with low phase contents. Therefore, XRD analyses were carried out on selected limestone samples to differentiate these rocks by their mineral composition.
Chemical analyses Geochemical analyses of the bulk limestone materials were carried out to prove whether or not these data can help to distinguish between different rocks. Chemical composition (major, minor and trace elements) was ascertained by XRF analysis with correction of matrix effects, except for FeO, which computed by titration with KMnO4, and water content (LOI: loss of ignition) by gravimetric analysis. Results were calibrated against several international standards, and reproducibility was better than 0.5%. For all trace elements, analytical uncertainty was lower than 5%. Measurements were carried out by sequential spectrometry (WDS Philips PW2400).
8. Preliminary Differentiation The limestone samples were first characterised according to their visual macroscopic appearance (texture, colour, homogeneity, peculiarities, etc.). On the basis of colour, three main groups were identified: Vicenza White Stone, Yellow Stone and Grey Stone. Although the optical properties of the limestone materials inside these groups were very similar, most of the samples appear homogeneous, so that initial conclusions allowed further differentiation (Table 1). The optical features of limestone may play a significant role, especially in the case of stone replacement. Comparison of raw limestone with the historical materials used in monuments can often provide preliminary information on whether to assign or exclude the
24 potential material.
Table 1. Limestone quarry samples from Berici Hills.
Municipal Stratigraphic Macroscopic Sample Name of quarry Colour Grain size district position features San Gottardo Area Milk-white 1 WCe Cengio Zovencedo Oligocene White Medium embellishment Milk-white 2 WBa Badia Zovencedo Oligocene Whitish Medium to coarse embellishment Milk-white 3 WGa Gazzo Zovencedo Oligocene Whitish Medium to coarse embellishment Milk-white 4 WSG San Gottardo Zovencedo Oligocene White Medium to coarse embellishment Milk-white 5 WCP Col di Pava Zovencedo Oligocene White Medium to coarse embellishment Milk-white 6 WAr Arcari Zovencedo Oligocene Whitish Medium to coarse embellishment Milk-white 7 WCB Cà Bertoldi Zovencedo Oligocene White Medium to coarse embellishment Costozza Area White 8 WVo Volto Longare Oligocene Straw- Fine to medium - yellow 9 WCi Ciole Longare Oligocene White Fine to medium - 1 WOl Olivari Longare Oligocene White Fine to medium - 0 San Germano Area 1 S. Germano Milk-white WSt Strenghe Oligocene White Medium to coarse 1 dei Berici embellishment Zovencedo Area 1 WSg La Sengia Zovencedo Oligocene White Fine to medium - 2 1 WCv Cava Vecchia Zovencedo Oligocene White Fine to medium - 3 Nanto Area 1 Yellowish- YCN Nanto Nanto Middle Eocene Fine to medium - 4 brown 1 Yellowish- YCG Grotte Nanto Nanto Middle Eocene Fine to medium - 5 brown 1 Yellowish- YCP Nanto 2 Nanto Middle Eocene Fine to medium - 6 brown Grancona Area (Pederiva) Frequent 1 YAq Le Acque Grancona Middle Eocene Yellow Medium to coarse macroscopic 7 allochems Frequent 1 YCe Cengelle Grancona Middle Eocene Yellow Medium to coarse macroscopic 8 allochems Few 1 YSc Scioso Grancona Middle Eocene Yellow Medium to coarse macroscopic 9 allochems Frequent 2 GAq Le Acque Grancona Middle Eocene Grey Medium to coarse macroscopic 0 allochems Frequent 2 GCe Cengelle Grancona Middle Eocene Grey Medium to coarse macroscopic 1 allochems Villabalzana Area 2 WVi Villabalzana Villabalzana Oligocene White Fine to medium - 2 Mossano Area 2 WMa Mantoan Mossano Oligocene Whitish Fine to medium - 3
25 Macroscopic examination alone is insufficient. Because of the very monotonous mineral composition and grain-size distribution of limestone varieties, a complex analytical scheme had to be developed to provide more detailed information on mineralogy and geochemistry. To complement macroscopic rock descriptions with a thorough micro-scale analysis, samples were subdivided into different portions to obtain material for geochemical analysis and XRD studies, and to prepare thin sections for microscopic analysis, respectively.
9. Results and Discussion 9.1 Petrography Results concern approximately 30 thin sections from each quarry. Thin-section studies included description of grain type, structure, texture, fossil content, frequency estimation of allochems by comparison charts (Baccelle and Bosellini, 1965), presence and type of interstitial material, type and frequency of contacts, and porosity. A summary of these characteristics for all examined quarries is shown in Table 2. On the basis of observed characters and quarry site, four main facies were identified. The main characteristics and diagnostic petrographic parameters for identification are reported. The names of the different varieties correspond to the traditional name given by that of the nearest village. However, some quarries in different localities show similar features, and these are grouped under the same variety name, corresponding to the village with the highest number of quarries with that stone.
San Gottardo Stone This is a white to straw-yellow biocalcarenite, very badly sorted (from 100 μm to a few millimeters in grain size) and coarse-grained (average grain-size 1-1.5 mm, but several grains reach pluri-millimetric to centimetric length). The texture is clastic-organogenic with grain support. Grain contact types are point and tangential contacts. Less frequent are concavo-convex and sutured contacts. The structure is essentially isotropic, although some portions show a slight preferential orientation of lengthened allochems (especially articulated coralline algae). Porosity is very high (20-30% by comparison chart) and is mainly of primary inter-granular type. Large pores are very frequent; the average value is 300 μm, but many reach pluri- millimetric size. The main components (> 40% by comparison chart) of this facies are allochems of coralline red algae, both encrusting coralline algae (Lithothamnion, ECA) and articulated coralline algae (ACA), the former generally prevailing. Benthic foraminifera (Nummulites, 10-20%), echinoderms (10%), bryozoans, molluscs, ostracods and corals are the other main components. Pore-filling materials are very scarce; the matrix is present in a few isolated portions. Another component is carbonate cement. Three main types occur: isopachous cement (multiple cement rims growing with equal thickness around grains; the rims consist of fibrous and microcrystalline crystals) around coralline red algal clasts; isolated intergranular and intergranular blocky sparry and micro-sparry carbonate cement; and syntaxial overgrowth rim cement on echinoid fragments. Quartz, feldspar and goethite-limonite are accessory constituents. Petrographically, the San Gottardo Stone is classified as rudstone-grainstone, locally grading to bindstone (Fig. 8).
26
Table 2. Diagram of textural and structural parameters of quarry stone.
RECAPITULATORY DIAGRAM OF TEXTURAL AND STRUCTURAL PARAMETERS OF QUARRY SAMPLES
Grain Types
Intra-basin components
Carbonate Extra-basin Component
Not Carbonate Skeletal Not Skeletal
Not carbonate Carbonate
Porosity
filling material
-
Grain Grain contact types Pore
Allochems
Classification size
- (Duhnam 1962, modified by
Embry & Klovan 1971)
Colour
Sorting
Sample
Texture
Structure
Grain
Benthic Foraminifera
Conservation State
limonite
-
Pyrite
Quartz
Peloids
Feldspar
Intraclasts
Glauconite Extraclasts
Phosphates
s
Goethite
pulids
Convex contacts
Corals
Matrix
Sutured Cement
Bryozoa
–
Mollusc
Ser
Intergranular
Intragranular
Echinoderms
Point contactsPoint
(Lithothamnion)
Tangential Tangential contacts
Miliolids
Nummulites
Dyscociclinas
Plancktonic Foraminifera
Concavo
Encrusting Corallines Algae Articulated Articulated Corallines Algae
WCe W M P I CO +++ +++ + + 15% 85% ++ ++ G +++ +++ ++ - - - ++ + + ------+ - - Rudstone-Packstone
Rudstone-Grainstone WBa Wh M-C VP I-A CO +++ +++ + + 20% 80% ++ ++ G ++++ +++ ++ - - - ++ ++ + + ------+ - - (Bindstone)
Rudstone-Grainstone WGa Wh M-C VP I CO +++ +++ + + 10% 90% ++ ++ G ++++ +++ ++ - - - ++ ++ + ------+ - - (Bindstone)
Rudstone-Grainstone WSG W M-C VP I-A CO +++ +++ + + 10% 90% ++ ++ G ++++ ++ ++ - - - + ++ + ------+ - - (Bindstone)
WCP W M-C VP I CO +++ +++ + + 20% 80% ++ + G ++++ +++ ++ - - - ++ ++ + ------+ - - Rudstone-Grainstone
WAr Wh M-C VP I CO +++ +++ + + 50% 50% ++ + G ++++ + ++ - - - ++ + + + ------+ - - Rudstone-Grainstone
WCB W M-C VP I CO +++ +++ + + 30% 70% ++ ++ G +++ +++ ++ - - - ++ + + ------+ - - Rudstone-Grainstone
WVo WY F-M P-M I CO ++ + ++ ++ 20% 80% ++++ ++ G - ++++ ++ - +++ - + + ++ ------+ - - Packstone
WCi WY F-M P-M I CO + + ++ ++ 20% 80% ++++ + G - ++++ ++ - +++ - + ++ ++ ------+ - - Packstone
27 WOl WY F-M P-M I CO ++ + ++ ++ 20% 80% ++++ ++ G - ++++ ++ - +++ - + + ++ ------+ - - Packstone
WSt W M-C VP I CO +++ ++ + + 20% 80% ++ ++ G ++++ ++ ++ - - - + ++ + ------+ - - Rudstone-Grainstone
WSg W F-M P I CO + ++ ++ + 30% 70% ++++ ++ G - ++++ ++ - +++ - ++ + ++ ------+ - - Packstone
WCv W F-M P I CO + ++ ++ ++ 30% 70% ++ + G + ++++ ++ - +++ - ++ + ++ ------+ - - Packstone-Grainstone
++ Packstone YCN YB F-M P I-A CO + + +++ 30% 70% +++ ++ B - + +++ + - ++++ +++ + + - ++ + + ++ + - + +++ + - ++
++ Packstone YCG YB F-M P I CO + + +++ 20% 80% +++ ++ B - ++ +++ - - ++++ +++ + + - ++ + + ++ + - + ++++ + - ++
++ Packstone YCP YB F-M P I CO + + +++ 30% 70% +++ + B - + ++++ + - ++++ +++ + + - ++ + + ++ + + + ++++ + - ++
++ Rudstone-Grainstone YAq Y M-C VP I CO + + +++ 40% 60% +++ ++ B - + ++++ ++++ - - ++ + - - - ++ + ++ ++ - - +++ - - ++ (Packstone)
++ RudstoneGrainstone YCe Y M-C VP I-A CO + + +++ 30% 70% +++ ++ B - + ++++ ++++ - - ++ + - - - ++ - ++ ++ + - +++ - - ++ (Packstone)
++ YSc Y M-C VP I CO + + +++ 30% 70% +++ ++ B - + ++++ ++++ - - ++ + - - - ++ + ++ ++ + - +++ - - Rudstone-Grainstone ++
++ GAq G M-C VP I CO + + +++ 30% 70% +++ ++ M - + ++++ ++++ - - ++ + - - - ++ - +++ ++ + + ++ - - Rudstone-Grainstone ++
++ GCe G M-C VP I CO + + +++ 30% 70% +++ ++ M - + ++++ ++++ - - + + - - - ++ - +++ ++ - + ++ - - Rudstone-Grainstone ++
WVi W F-M P I CO + ++ + + 20% 80% ++++ ++ G - ++++ + - +++ - ++ - ++ ------+ - - Packstone
WMa Wh F-M P I CO + ++ ++ + 30% 70% ++ ++ G - +++ + - - - ++ + +++ ------+ - - Packstone-Grainstone
Colour: W= white; Y= yellow; G= Grey; WY= white-straw yellow; Wh=whitish; YB= Yellowish - brown Grain Size: C= coarse; M= Medium; F= Fine Sorting: G=Good; M=Moderately; P=Poorly; VP= Very Poorly Structure: I= isotropic; A= anisotropic Texture: CO= Clastic - Organogenic Conservation state (on the base of oxides-hydroxydes content): G=good (oxides rare) M=Moderately (oxides frequent) Bad=B(oxides very frequent) [+] = rare [++] = scarce [+++] = frequent [++++] = very frequent
28
Fig. 8. Rudstone-grainstone with encrusting and articulated coralline algae, benthic foraminifera and echinoderms. San Gottardo Stone. Nicols //. Quarries of this stone include: WCe, WBa, WGa, WSG, WCp, WAr, WCB and WSt.
Costozza Stone This is a medium- to fine-grained limestone (average grain size 300-700 µm, with a few clasts of millimetric length). It is white to straw-yellow in colour, with moderate sorting (0.3–0.8mm). The texture is clastic-organogenic with grain support. The main grain contact types are point and tangential contacts; in some portions concavo-convex and sutured contacts are present. The structure is essentially isotropic. Porosity is high (15-20% by comparison chart) and of intraparticle and interparticle type, the latter prevailing. Pore size ranges between a few microns to 1 mm, with an average of 0,2mm; the distribution is isotropic. The main components of this facies are bioclasts (> 50% by comparison chart) and are represented by prevailing articulated coralline algae (ACA, 30-40%), benthic foraminifera (main Miliolids and Nummulites, 10-15%), bryozoans (10%), echinoderms and molluscs. Interstitial material is abundant, mainly composed of a fine-grained intergranular carbonate matrix which surrounds clasts and is uniformly distributed. Intergranular and intragranular carbonate sparite and microsparite cement fill the matrix microporosities. Isopachous fibrous and microcrystalline crystal cement occurs around some coralline red algal bioclasts. More evident is the clear calcite syntaxial overgrowth cement on rare plates and spines of echinoderms. Quartz, feldspar and goethite-limonite are accessory constituents. Petrographically, the Costozza Stone is classified as packstone (Fig. 9). Quarries of this stone include: WVo, WCi, WOl. Although WVi, WSg and WMa do have some characters in common with Costozza Stone, they were excluded to prevent possible confusion.
29
Fig. 9. Packstone with articulated coralline algae, benthic foraminifera (Miliolids, Nummulites) and bryozoans. Costozza Stone. Nicols //.
Nanto Stone Nanto Stone is a yellowish-brown medium- to fine-grained limestone of Middle Eocene age, outcropping near Vicenza and quarried along the slopes of the south-western sector of the Berici Hills. Often used locally for its aesthetic qualities and easy working, when exposed to an urban environment, it undergoes severe deterioration (Cattaneo et al., 1976; Fassina and Cherido, 1985). Nanto Stone is petrographically classified as grainstone-packstone; the matrix is frequent, porosity is high, and cementation is fair. The texture is clastic-organogenic, with grain support. The structure is essentially isotropic. The skeleton is mainly due to benthic (Nummulites, Assilines) and planktonic plurilocular foraminifera (Globigerinoids) (30-40%), plates and spines of echinoderms (20-30%), red algae and micron-sized undeterminable skeletal debris. Many highly fragmented allochems are present. Skeletal grains, with sutured grain boundaries, are cemented by sparite and micro-sparite of carbonate crystals and rare goethite. More evident is the clear syntaxial overgrowth calcite cement on echinoderm plates and spines. The matrix is quite abundant (10%), mainly intergranular, and in a few cases derives from deformed intraclasts (pseudomatrix). The siliciclastic fraction is made up of a few isolated grains of biotite, glauconite, chlorite, phosphates, pyrite, quartz, feldspars and clay minerals. Apart from echinoid allochems, other components appear as very fragmented bioclasts, and entire particles are quite isolated. Grain contact types are mainly concavo-convex and sutured contacts; point and tangential contacts are scarce. Porosity is medium to high (15% by comparison chart) and preferentially of intragranular type. Large pores are very scarce; their average value is 100-200 μm. Mesopores are occupied by drusy calcite. Iron sulphides occur as small pyrite framboids or clusters of framboids, finely disseminated in the limestone within foraminifera chambers. Glauconite is also frequent as filling in foraminifera pores. Microscopic observations in incident 30 light clearly demonstrate the oxidation of glauconite into goethite. Petrographical evidence points to migration of goethite out of the original grains into adjacent areas, which are both stained with iron hydroxides, and a beige to orange-brown discoloration of the limestone occurs. Iron oxides-hydroxides are sometimes present in fossil shells. Quartz occurs as small grains (average 60 µm) dispersed in carbonate (Fig. 10).
Fig. 10. Grainstone-packstone with planctonic and benthic foraminifera, echinoderms and red algae. Nanto Stone. Nicols //. The quarries of this stone are all near the village of Nanto, and include YCN, YCG and YCP.
San Germano Stone This limestone has two well-defined varieties, termed Yellow San Germano Stone (YSGS) (Fig. 11), because of its pale yellow colour, and Grey San Germano Stone (GSGS) (Fig. 12), ash-grey in colour. The main difference between the two varieties is, precisely, colour; in YSGS, oxidation of ferrous minerals (glauconite and to a lesser extent pyrite) is mainly responsible for pigmentation. In GSGS, ferrous minerals are less weathered and alteration products (goethite- limonite) are very scarce. Both are quarried in the same localities, but have evidently been subjected to different oxidation-reduction conditions, as also confirmed by the high organic content in the grey variety. These limestones are coarse- to very coarse-grained (average grain size 1-1.5 mm). In addition, the presence of macro-bioclasts (macroforaminifera such as Nummulites and Discocyclinas) from a few millimeters to several centimeters in size, is very characteristic. The texture is clastic-organogenic with grain support. The main grain contact types are concavo-convex and sutured contacts; point and tangential contacts are less frequent. The structure is essentially isotropic, although some portions show preferential orientation in the largest bioclasts; a tiled roof structure also occurs in some transversal sections. 31 Porosity is high (15-20%) and of intraparticle and interparticle type, as it develops within and between fossils. Pore sizes range from a few microns to 1.5 mm, with an average value of 0.1 mm. The matrix is frequent (5-10%) and mainly intergranular. Examples of a pseudomatrix are also present. The cement, both intragranular and intergranular, is not very abundant, and mainly consists of microcrystalline calcite mosaics, but also sparite and microsparite. Clear syntaxial overgrowth calcite cement on plates and spines of echinoderms is present.
Fig. 11. Rudstone-packstone with benthic foraminifera (Nummulites and Discocyclinas), red algae and echinoderms. Yellow San Germano Stone. Nicols //. The stone consists predominantly of bioclasts of macroforaminifera (40-50% by comparison charts) such as Discocyclinas and Nummulites, followed by red algae (5-10%), and echinoderms (5%). Intraclasts and/or micritised grains are frequent (10-20%). Detrital quartz grains (3–5%), green grains (glauconite), clay minerals and oxides-hydroxides of iron (goethite-limonite) are other non-carbonate constituents. Frequent fragmented allochems are also present. Petrographically, San Germano Stone is classified as rudstone-packstone. Iron sulphides occur as small pyrite framboids or clusters of framboids, finely disseminated in the rock, especially inside foraminifera. Glauconite also is present. Microscopic observations in incident light clearly demonstrate the oxidation of glauconite and pyrite into goethite. In the yellow variety, petrographical evidence points to migration of goethite into adjacent areas. The quarries of Yellow San Germano Stone include YAq, YCe and YSc. Those of Grey San Germano Stone include GAq and GCe.
32
Fig. 12. Rudstone-packstone with benthic foraminifera (Nummulites and Discocyclinas), red algae and echinoderms. Grey San Germano Stone. Nicols //. 9.2 Evidence of Diagenesis Porosity in carbonate rocks results from many processes, both depositional and post- depositional. An understanding of these processes and of the textural history of porosity is necessary for a full appreciation of the history of the rock. Several mechanisms appear particularly important in producing or changing porosity and pore size distribution in carbonate rocks. Primary interparticle porosity is formed by the deposition of well-sorted calcareous sand or gravel under the influence of strong currents or waves, or by local production of calcareous sand-sized particles with sufficient rapidity to deposit particle on particle, with little or no interstitial mud. Dissolution of interstitial mud in calcareous sand may produce microvuggy porosity resembling interparticle pore space. Simple cementation by calcite, anhydrite or dolomite destroys porosity and pore size. Calcite cement appears to be especially common where the particles, such as crinoid fragments, are monocrystalline. Primary structures are produced by the formation of a rigid or semi-rigid framework which may be organic or inorganic, and inter-framework pockets may be filled with sediment or, later, with cement. Modification of porosity is a burial process, and is directly related to the intensity of chemical compaction, which itself is controlled by three factors: (1) degree of early cementation, (2) grain type, (3) grain size. Early-cemented intervals give rise to moderate compaction, regardless of grain size or type. Early diagenetic cement stabilised the limestone framework and prevented mechanical compaction. Where early cement is absent, mechanical compaction severely reduced primary porosity during a first burial phase. Ductile grains were squeezed between rigid grains, obstructing pore connections and giving rise to distinct porosity patterns in the limestone bodies. In the absence of early cementation, grain type and size were the dominant factors of porosity control. Coarse-grained intervals composed of micritic grains underwent less compaction than fine-grained intervals (Flugel, 2004).
33 San Gottardo Stone The most important processes occurring in subsurface carbonates are compaction and pressure solution. Burial of non-cemented carbonate sediments under an increasing burden results in compaction and, if continued, in pressure solution. Mechanical compaction can eventually continue on to chemical compaction if the grains begin to dissolve at their contacts. During mechanical compaction, strain is concentrated at grain contacts. The specific criteria of these processes may be recognised on different scales, including thin sections. Diagenesis, creating situations of different cementation and compaction, has contributed to the specific textural characters of this type of stone. After deposition of coarse-grained skeletal debris with high primary porosity and accumulation of red algae, foraminifers and corals, cementation started in a marine-phreatic environment. This is shown by isopachous cement rims, commonly observed in the intraskeletal pores of fossils and around bioclasts. The cement derives from precipitation of oxidising fluids, typical of early diagenetic marine cements (Tucker & Wright, 1990).
Isopachous rim cement
Fig. 11. Evidence of isopachous rim cement around algae bioclasts. Primary pores partly filled with isopachous equant calcite cement and the dominant point and tangential contact types between grains suggest that this lithotype suffered relatively earlier cementation, which later opposed the compaction process, since early cementation protects grains against compaction and pressure solution. Point contacts indicate initial compaction, and tangential contacts increasing compaction. The few sutured and concavo-convex contact points reveal no effects of pressure solution at grain contacts. The preservation of open interparticle porosity was also favoured by the relatively stable mineralogy of some skeletal grains (high Mg-calcite) and probably to burial solution and
34 enhancement of interparticle pores. The preservation of primary pores requires post-depositional diagenesis to be limited in its pore-destroying effects, compaction to be kept at a minimum, and fluctuations between exposure and sub-emergence to create a balance between the formation of solution pores and the destruction of porosity by shallow-burial cementation. Preservation of porosity in shallow-burial environments is a consequence of minimal burial, reduced burial stress, increased framework rigidity, exclusion of pore water, low calcite mineralogy, permeability barriers, and pore resurrection. Skeletal architecture probably also influenced the preservation of primary porosity, with low percentages of microcrystalline skeletal grains (e.g., bryozoans) which are more susceptible to mechanical and chemical diagenesis than single-crystal echinoderm grains, which are more frequent (Meyers, 1980). Grain size and sorting are two other factors influencing porosity; coarse grains and poor sorting favour large pores. The high porosity of these rocks is clearly linked to their specific diagenetic features.
Costozza Stone. Cement rims are lacking in the micrite-rich low-energy facies, suggesting that the quantity of cement around skeletal grains is directly related to the hydrodynamic energy level of the depositional environment. Some primary pores partly filled with isopachous equant calcite cement, and point and tangential contact types, but also sutured and concavo-convex contact points between grains, suggest that this lithotype suffered relatively earlier cementation, only partially opposed by the compaction process. The large numbers of Miliolids, organisms typically found in areas near of emerged land, and of a low-energy depositional environment, as confirmed by large amounts of matrix, are due to quite stable sediments, subjected to only weak currents. The presence of this matrix decreases primary interparticle porosity and probably contrasted high compaction phenomena. Skeletal architecture probably also influenced the non-preservation of primary porosity, with large amounts of microcrystalline skeletal grains (e.g., bryozoans) which are more susceptible to mechanical and chemical diagenesis than less frequent single-crystal echinoderm grains (Meyers, 1980).
Nanto Stone. Diagenesis, creating situations of different cementation and compaction, contributed to modifying the textural character. This rock is more compacted, as testified by penetrated and sutured contacts among the grains and many fragmented allochems. This lithotype does not seem to have suffered earlier cementation opposing compaction, and the presence of a pseudomatrix confirms compaction processes. Large pores are also very scarce. Clay-bearing rocks are often more susceptible to strong compaction than clay- free rocks. In limestone/marl sequences, limestone is compacted (Munnecke, 1990). However, the precise effect of clay (e.g., decreasing permeability to cementing pore waters, enhanced diffusion of dissolved CaCO3) requires further study.
San Germano Stone. Considerations for this stone are similar to those of Nanto Stone. Differences in grain size in comparison with Nanto Stone correspond to different degrees of energy in a depositional context. Variations in oxidation-reduction conditions appear in the two varieties of San Germano Stone. Pyrites occur both within foraminifer chambers and in the matrix. Glauconite and pyrite together indicate their authigenic precipitation at the oxic/anoxic or at least oxic/suboxic boundary in an extremely slow sedimentation environment where Fe3+ was transiently available as Fe2+ in solution (Odin and Matter, 1981; Ireland et al., 1983). Later, pervasive oxygenated fluids led to the formation of goethite by oxidation of these ferrous minerals (especially glauconite), and caused the different colours of the two varieties.
35 9.3 Palaeo-Environmental Interpretation of Microfacies The distributional pattern of the microfacies types is based on the following main parameters: faunal association, planktonic ratio in foraminiferal associations, interstitial material. The grouping of foraminifera according to their living affinities provides an indication of the water depth of their depositional environments. Planktonic foraminifera have great potential in palaeo- ecological studies (Murray, 1991). Changes in faunal associations reflect the differentiation of biotopes into several smaller areas, each controlled by bottom morphology and current patterns (Piller and Pervesler, 1989).
San Gottardo Stone This stone is characterised by large amounts of encrusting and articulated red algae, with different proportions of Nummulites and echinoderms. These sediments probably indicate a high to moderate energy environment within or near coral patch reefs, as shown by poorly sorted rudstone to grainstone textures and algal assemblages. Moderate energy characterises the environment and results in poor sorting, abrasion of grains, enrichment of components resistent to abrasion (echinoderms, Lithothamnion) and low micrite contents, conditions which promote high primary permeability. These carbonates were supported by high primary porosity and permeability, facilitating rapid water flow through the rocks.
Costozza Stone This micrite-rich packstone, with articulated algae and the significant presence of benthic foraminifera such as Miliolids, suggests that it was deposited in a lagoonal environment with little wave movement (below storm wave base/fair weather wave base). Due to the high micrite content (up to 20%), the primary permeability of these sediments was rather low. Diagenesis is also characterised by limited cementation and sometimes by mechanical compaction. Chemical compaction features like grain-to-grain or concave–convex contacts are more frequent than in the San Gottardo variety, although the presence of micrite may have contrasted extreme compaction. Low primary permeability and the presence of micrite in interparticle pore spaces probably reduced the rate of fluid flow in the sediment, which resulted in limited cementation. Moreover, echinoderms, which are the nuclei for the crystallisation of the volumetrically most important epitaxial cements, are rather scarce in this stone. Micrite was later squeezed between the components and lithified the sediment.
San Germano Stone This microfacies was deposited in low energy conditions, above the storm wave base in the middle ramp setting, at a water depth estimated to have been between 30 and 60 m (middle neritic). Sediments may have been affected by bottom currents. The large amount of matrix and mode of preservation of the larger foraminifera indicate a low-energy outer ramp setting, with events of heavy erosion and floating transport from coastal areas. Glauconite grains occurring as infillings of chambers in larger foraminifera support this interpretation (Odin and Matter, 1981). The presence of feldspar and quartz grains indicates sporadic terrigenous influx to the environment.
Nanto Stone This variety is represented by fossiliferous grainstones, and the microfacies is characterised by the most diverse foraminiferal fauna. The Nummulitidae family is represented by Assilines and Nummulites. Discocyclina was also identified in some samples. The dominance of fine-grained sediments and the presence of fragmented allochems indicate a high-energy depositional environment, probably an outer ramp, because of the presence of
36 planktonic foraminifera. Larger foraminifers in particular suffered great in situ fragmentation, because they are fragile and large enough to have several grain-to-grain contacts with neighbouring components. The planktonic foraminiferal assemblage contains surface to deep- welling species. The presence of glauconite indicates deeper water and less oxygenated or reducing conditions with a low sedimentation rate (Odin and Matter, 1981). The variable, unequal preservation of mud and larger bioclasts attest to high- to moderate- energy settings.
Fig. 14. Reconstructed block diagram of Middle Eocene carbonate ramp, showing microfacies settings, representing a wave-dominated carbonate ramp with minor storm activity (Cosovic et al., 2004, mod.).
9.4 Mineralogical Composition Mineral composition was characterised in selected limestone samples from the Berici Hills, to examine the possibility of their differentiation by mineral content. For this purpose, the non- carbonate fraction was separated, to investigate main rock composition and insoluble materials individually. The results (Table 3) clearly show the predominance of calcite in the bulk samples. In general, all white stones are very pure carbonates, with small amounts of accessory minerals, mainly quartz and goethite, which are found in the residual fraction. Also, no differences in terms of mineralogical content between the two varieties previously defined by petrographic examination (San Gottardo and Costozza Stone) were noted. Their differing compositions emerge from the yellow varieties of soft stone. A large non- carbonate fraction testifies to significant amounts of accessory minerals, especially in Nanto Stone, which has insoluble residues of about 12%, mainly composed of goethite, K-feldspar, quartz and traces of clay minerals (smectite). Yellow San Germano Stone has a lower percentage of non-carbonate fraction (average 6%, although one quarry only had about 2%), and mainly contains goethite and K-feldpar. No important traces of clay minerals were found. Instead, the grey variety is characterised by large amounts of glauconite and lesser amounts of goethite, which may explain its different colour, as different weathering stages of glauconite
37 may produce circumgranular goethite crusts responsible for yellow pigmentation. Other minerals in the grey variety are the same as those in the yellow one. There are indications that, in some samples, the presence or absence of unaltered glauconite or clay minerals can be used as criteria for specific limestone types (GSGS and Nanto Stone). However, the percentage of insoluble residue distinguishes different varieties: White stone with insoluble residue < 2.5% and mainly composed of quartz and goethite: of these, the Costozza variety has a slightly higher non-carbonate fraction than the San Gottardo Stone; Yellow San Germano Stone, with average insoluble residue of 6%, composed of goethite, K-feldspar and quartz; Grey San Germano Stone, with insoluble residue ranging from 7% to 8.5%, composed of glauconite, K-feldspar, goethite and pyrite; Nanto Stone, with insoluble residue around 11-12%, composed of goethite, K-feldspar, quartz and clay minerals (smectite);
Table 3. Mineralogical composition (by XRD) of bulk samples and non-carbonate fraction (in order of abundance).
Sample XRD Bulk sample % Residue XRD Residue San Gottardo Area WCe Calcite 0.45±0.12 Quartz, Goethite WBa Calcite, traces of Quartz 0.55±0.04 Quartz, Goethite WGa Calcite 0.48±0.09 Quartz, Goethite WSG Calcite 0.58±0.10 Quartz WCP Calcite 0.49±0.05 Quartz, Goethite WAr Calcite, traces of Quartz 0.52±0.06 Quartz, Goethite WCB Calcite 0.61±0.13 Quartz, Goethite Costozza Area WVo Calcite, traces of Quartz 1.65±0.34 Quartz, Goethite WCi Calcite 0.68±0.22 Quartz, Goethite WOl Calcite 1.04±0.39 Quartz, Goethite San Germano Area WSt Calcite 1.22±0.12 Quartz, Goethite Zovencedo Area WSg Calcite 2.09±0.92 Quartz, Goethite WCv Calcite 1.87±0.23 Quartz, Goethite Nanto Area YCN Calcite, K-feldspar, Goethite 11.5±0.67 Goethite, K-feldspar, traces of Clay minerals YCG Calcite, Goethite, Quartz 12.5±0.55 Goethite, Quartz YCP Calcite, K-feldspar, Goethite 11.8±0.72 Goethite, K-feldspar, traces of Clay minerals Grancona Area (Pederiva) YAq Calcite, K-feldspar 6.24±0.89 Goethite, K-feldspar YCe Calcite, K-feldspar 6.34±1.02 Goethite, K-feldspar, Quartz YSc Calcite 2.86±0.34 Goethite, K-feldspar, Quartz GAq Calcite, K-feldspar 8.49±1.09 Glauconite, K-feldspar, Goethite, Pyrite GCe Calcite, K-feldspar 7.02±0.93 Glauconite, K-feldspar, Goethite, Pyrite Villabalzana Area WVi Calcite 1.64±0.45 Quartz, Goethite Mossano Area WMa Calcite 1.19±0.55 Goethite
Considering the age and formation of the various types of stone, a direct correlation between the percentage of insoluble fraction and age was observed. The oldest stone (Nanto) has the highest content of insoluble residue; the youngest (White Vicenza) the lowest. This may be due to the different terrigenous contributions at the times of formation of these sediments, and in particular to different environmental conditions of deposition.
38 9.5 Chemical Composition Geochemical analyses of the bulk quarry materials from the Berici area were carried out to prove whether or not these data can help to distinguish between the various limestones. The results, listed in Table 4, show that, in general, the composition of stone from individual quarries is chemically quite homogeneous, like that of the major elements. There is rather wide chemical variability among the different sites, allowing good discrimination between the stone of the three main groups (white, yellow, grey). As previously determined for mineralogical composition, it is clear that differentiation on the basis of chemical composition reflects the differing mineralogical content of the various stones. White varieties are in general very monotonous, with typically high contents of CaO and very low concentrations of other elements. There are therefore almost no important differences between quarries. Lower contents of SiO2 and Al2O3 are due to small amounts of feldspar and quartz, and Mg to the presence of many bioclasts of red algae, which may contain up to 30% of MgCO3. In general, however, even the trace element compositions of the limestone are very monotonous, because of the low concentrations of minor minerals. In other varieties, the high concentration of silica, particularly in Nanto Stone, is due to quartz, feldspar and clay minerals. Mica and clay minerals can also cause a slight increase in Al2O3, MgO, TiO2, K2O, Fe2O3, Cr and V. High Fe2O3 in the yellowish-brown varieties fits the high amounts of hydroxide as goethite- limonite; in the grey variety, it is due to glauconite and goethite. In conclusion, the chemical composition of these limestones only provides more or less partially useful data for detailed characterisation and differentiation.
9.6 Geochemistry and Palaeo-Redox Data Although carbonate geochemistry may be considered a rather conservative parameter in diagenetic processes, it does yield important palaeo-environmental information. In particular, Sr content is helpful in understanding the origin and diagenesis of carbonates, since calcareous- dolomitic samples generally contain less Sr than limestone ones. This behaviour fits the fact that Sr substitutes for Ca rather than Mg in the dolomite lattice. The Sr contents of our samples allow some further comments as, according to current understanding, the Sr/Mg ratio may used as a palaeo-environmental indicator. In neritic and coastal environments, the Sr and Mg contents of carbonates show a positive correlation and in pelagic environments an inverse correlation. In our samples, Sr correlates positively with Mg, suggesting that both were deposited in shallow water in the coastal platform domain, fitting geological and palaeo-ontological data. Other significant elements yielding palaeo-environmental and diagenetic information are redox- sensitive trace elements (V, Cr, Ni, Co, Zn, Cu), widely used as indicators of palaeo- environmental conditions on the sea bed. In particular, the Ni/Co ratio is considered an index of palaeo-oxygenation conditions. In our samples, the values of this ratio are <5, indicating that the sediments were deposited under an oxic water column. Only some specimens of the grey variety show values up to 5 and do not fit oxic conditions, which are well correlated with low amounts of oxidised minerals. Taking into account the relative Ni enrichment of Nanto Stone and Grey San Germano Stone, it may be assumed that these groups were deposited in less oxic conditions than the others. Vanadium may also be used as a palaeo-redox indicator, and confirms these considerations. Recent culture experiments on foraminifers have shown that the V/Ca ratio in foraminifer shells is incorporated in direct proportion to that in seawater (Hastings et al., 1996). This, together with the fact that anoxic sediments also act as a sink of marine V (Breit and Wanty, 1991), makes this tracer a potential tool in defining redox conditions in ancient sediments. Other trace metals like Ni, Co and Zn have also been found to be incorporated at elevated levels in laminated anoxic sediments (Dean et al., 1997; Jacobs et al., 1985). Even Cr, a lithophile
39 element, behaves like a chalcophile one in reducing conditions and is precipitated. Concentrations of these elements in Nanto Stone and San Germano Stone, in comparison with the white variety, confirm the depositional conditions of these rocks. Petrographic study shows pyrite as fine aggregates in Nanto Stone and San Germano Stone, indicating authigenic precipitation. Pyrite also occurs within foraminifer chambers, together with glauconite. In general, the presence of authigenic pyrite is an indicator of a reducing depositional environment (Berner, 1984). In oxic water, Fe is mainly present as a particulate oxide because Fe3+ compounds are quite insoluble, whereas in a reducing environment iron undergoes reductive dissolution to Fe2+ species. An alternative hypothesis is that, after sediment deposition, element redistribution took place. This process, occurring during diagenesis or weathering, is well-known. Element remobilisation usually occurs towards redox boundaries, and the fixing of mobilised elements is controlled by a redox-sensitive diagenetic phase (e.g., pyrite). Element redistribution may take place if sufficient organic matter and time are available to allow full redistribution, and if redox- sensitive elements are in a form that allows their transport. As organic matter is the main source of such elements, its destruction by oxidising reactions causes mobilisation of elements in dissolved ionic form in interstitial fluids. Thus, fixing of trace elements is controlled by the presence of reduced sulphur and, if sulphide compounds undergo oxidation, some of the elements remains dissolved in seawater or is removed by interstitial fluids. We believe that the different distributions of significant trace elements in our samples reflect different palaeo- oxygenation conditions in the sedimentary basin, although element redistribution during diagenesis cannot be excluded.
40
Table 4. Chemical composition of quarry samples.
Major Elements (%Ox) Ox% WCe WBa WGa WSG WCP WAr WCB WVo WCi WOl WSt WSg WCv YCN YCG YCP YAq YCe YSc GAq GCe WVi WMa SiO2 0.14±0.04 0.14±0.05 0.22±0.10 0.16±0.03 0.30±0.05 0.50±0.09 0.17±0.06 0.26±0.08 0.23±0.05 0.28±0.03 0.09±0.04 0.16±0.07 0.20±0.08 5.42±0.62 4.90±0.57 5.40±0.87 2.62±0.98 3.00±1.18 1.57±0.49 5.01±1.03 3.70±0.89 0.09±0.04 0.06±0.05 TiO2 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.02±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.34±0.06 0.29±0.09 0.37±0.05 0.26±0.04 0.23±0.09 0.11±0.12 0.42±0.24 0.23±0.13 0.01±0.00 0.01±0.00 Al2O3 0.12±0.03 0.12±0.04 0.16±0.03 0.12±0.05 0.19±0.02 0.22±0.03 0.13±0.01 0.15±0.03 0.14±0.02 0.12±0.01 0.10±0.04 0.12±0.05 0.15±0.03 2.13±0.34 1.91±0.87 2.47±0.23 0.96±0.43 1.05±0.38 0.55±0.46 1.72±0.41 1.19±0.35 0.07±0.04 0.07±0.05 FeO 0.16±0.02 0.16±0.02 0.24±0.03 0.21±0.01 0.24±0.02 0.26±0.04 0.23±0.03 0.27±0.03 0.25±0.01 0.22±0.02 0.20±0.02 0.21±0.03 0.24±0.03 0.62±0.39 0.63±0.13 0.27±0.46 0.44±0.08 0.4±0.1 0.33±0.21 0.66±0.24 0.69±0.43 0.24±0.07 0.23±0.04 Fe2O3 0.01±0.01 0.01±0.01 0.01±0.01 0.01±0.01 0.02±0.01 0.02±0.01 0.02±0.01 0.03±0.01 0.03±0.00 0.03±0.00 0.03±0.01 0.02±0.00 0.01±0.00 2.9±0.51 2.52±0.78 3.38±0.35 0.48±0.23 0.61±0.45 0.36±0.34 1.37±0.24 1.19±0.36 0.03±0.01 0.02±0.00 MnO 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.01±0.00 0.02±n.d. 0.02±n.d. 0.02±n.d. 0.01±0.01 0.02±0.02 0.05±0.02 0.01±0.01 0.01±0.00 0.01±0.00 0.01±0.00 MgO 0.42±0.02 0.44±0.02 0.40±0.03 0.44±0.01 0.46±0.00 0.46±0.03 0.45±0.02 0.52±0.04 0.49±0.03 0.42±0.04 0.45±0.02 0.43±0.03 0.46±0.01 0.85±0.17 0.68±0.47 0.95±0.34 0.69±0.32 0.7±0.17 0.51±0.25 1.02±0.26 1.08±0.14 0.41±0.01 0.41±0.02 CaO 55.48±0.68 55.80±0.72 56.14±0.46 56.07±0.56 55.60±0.72 52.64±0.43 55.85±0.43 54.79±0.61 55.07±0.45 54.89±0.48 54.78±0.34 52.83±0.78 56.42±0.46 47.01±1.10 48.11±0.98 46.40±1.02 49.96±1.32 51.22±1.68 53.48±1.56 47.82±1.49 50.54±1.78 54.99±0.34 54.60±0.48 Na2O 0.04±0.01 0.02±0.01 0.01±0.00 - - - 0.01±0.00 - - - - - 0.02±0.01 0.03±0.04 0.01±n.d 0.01±n.d 0.04±0.01 0.02±0.02 0.02±0.01 0.04±0.01 0.08±.02 - - K2O 0.01±0.00 0.02±0.01 0.02±0.01 0.02±0.01 0.02±0.01 0.06±0.01 0.02±0.01 0.03±0.01 0.03±0.01 0.02±0.00 0.02±0.01 0.03±0.01 0.02±0.01 0.47±0.11 0.41±0.11 0.43±0.09 0.33±0.31 0.39±0.21 0.13±0.14 0.68±0.22 0.46±0.16 0.02±0.01 0.02±0.01 P2O5 0.02±0.01 0.03±0.01 0.04±0.01 0.03±0.010.01 0.04±0.01 0.06±0.02 0.03±0.01 0.03±0.00 0.02±0.00 0.03±0.01 0.04±0.01 0.03±0.01 0.03±0.01 0.18±0.06 0.17±0.04 0.23±0.06 0.04±0.03 0.05±0.02 0.04±0.04 0.07±0.05 0.04±0.03 0.02±0.01 0.03±0.01 L.O.I 43.90±0.49 43.89±0.23 43.83±0.46 43.79±0.67 43.76±0.38 46.30±0.49 43.87±0.63 43.99±0.29 44.05±0.36 44.78±0.28 44.73±0.43 46.59±0.48 43.89±0.34 40.02±0.24 39.97±0.34 39.85±0.27 44.49±1.07 42.59±1.55 43.16±0.98 40.57±1.54 40.61±1.34 43.92±0.34 44.47±0.48 Tot 100.31±0.31 100.64±0.45 101.10±0.25 100.86±0.33 100.65±0.48 100.54±0.31 100.80±0.15 100.09±0.38 100.32±0.41 100.81±0.36 100.46±0.54 100.43±0.49 101.45±0.45 99.99±0.59 99.62±0.67 99.78±0.64 100.33±0.09 100.28±0.05 100.31±0.23 99.37±0.87 99.81±0.59 99.80±0.74 99.91±0.38 Trace Elements (ppm) Sc <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <2 <5 <5 <5 <5 <5 <5 <5 <5 V <5 8±2 6±2 13±3 11±2 6±2 11±3 6±2 5±2 6±2 5±1 7±2 10±3 89±29 76±16 119±26 28±6 34±5 39±5 43±7 29±4 6±2 <5 Cr <6 <6 <6 <6 9±4 8±2 14±3 9±4 12±4 8±3 <6 21±5 10±4 67±14 56±8 80±12 35±9 38±10 32±7 48±6 39±7 13±4 13±4 Co 26±4 <3 16±4 4±1 7±2 13±3 6±4 11±4 10±5 9±2 4±1 54±11 4±1 14±13 3±1 16±2 40±6 28±13 38±15 13±7 9±2 6±3 7±2 Ni <3 <3 6±1 <3 3±1 9±1 4±1 7±2 4±2 7±2 10±1 11±2 4±1 29±13 24±11 42±13 12±2 134±4 11±4 18±5 60±11 7±2 4±1 Cu <3 <3 <3 <3 5±3 9±2 4±1 25±5 20±4 30±5 13±6 12±4 5±2 18±6 12±6 23±11 5±3 6±5 <3 9±3 <3 13±5 12±5 Zn <3 3±2 <3 <3 <3 3±1 <3 12±3 9±1 15±2 <3 4±1 <3 30±2 30±2 32±3 5±1 9±6 <3 25±3 17±6 7±2 6±2 Ga <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 <3 <5 <5 <5 <5 <5 <5 <5 8±1 Rb <3 <3 6±2 <3 <3 13±2 <3 11±3 9±2 10±3 14±2 10±2 5±1 5±6 <2 11±2 <3 4±4 <3 <3 <3 7±2 11±3 Sr 429±17 439±13 380±23 413±23 431±34 407±56 396±28 486±35 493±24 479±25 419±27 431±56 425±35 457±43 424±67 451±23 431±34 478±43 362±47 436±36 586±29 316±25 327±46 Y 3±1 3±1 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 6±3 8±1 3±0 4±0 5±2 5±0 7±0 6±1 <3 <3 Zr 8±2 9±2 8±5 8±2 8±4 18± 7±2 24±6 18±4 22±2 18±5 19±3 9±1 30±25 6±2 47±7 22±6 25±9 13±1 33±5 27±4 12±2 13±4 Nb <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 7±3 <2 6±2 5±2 5±2 3±3 7±1 5±2 <3 <3 Ba 34±4 31±3 27±3 38±3 35±5 17±4 33±6 <10 <10 <10 19±5 15±4 30±4 49±15 64±14 39±11 49±12 38±14 29±9 49±14 41±11 <10 <10 La <10 13±4 <10 <10 <10 <10 21±4 <10 <10 <10 13±2 <10 <10 <10 <10 10±2 <10 11±1 11±3 12±2 12±2 <10 <10 Ce <10 11±1 <10 <10 <10 <10 <10 13±1 11±2 14±1 <10 14±1 11±1 14±6 10±3 18±2 <10 18±4 17±1 <10 14±3 10±2 12±2 Nd 16±1 <10 11±2 15±3 <10 <10 12±2 <10 <10 <10 19±2 <10 13±1 16±11 8±3 27±2 13±1 16±7 11±2 15±1 <10 <10 <10 Pb <5 <5 <5 <5 <5 11±2 <5 10±2 8±2 12±2 9±1 6±1 <5 20±8 25±3 15±4 5±1 7±2 <5 5±2 5±1 11±3 11±4 Th 12±4 <3 25±3 21±3 19±5 17±4 26±4 24±4 28±4 29±2 17±3 15±2 11±1 26±20 45±11 20±8 26±4 26±7 <3 29±2 3±1 24±4 22±4 U 5±1 5±0 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 <1 <3 <3 <3 <3 <3 <3 <3 <3
41
42 10. Discussion Vicenza Stone, in its different varieties, has been widely used for building in the historical and modern architecture of many towns in the Veneto region. This massive use of “soft stone” is due to the large deposits of this material on the slopes of the Berici Hills, as well as for its ease of working, availability, and aesthetic features. Summarising the analytical results, certain criteria (parameters) were evaluated for limestone differentiation. Characteristic features were compiled for the main limestone type, and the results of all analytical methods (Table 5) provides a first overview and classification of the various types. In particular, the combination of allochem types and frequency, qualitative and quantitative non- carbonate mineral contents, and chemical composition were useful in successfully differentiating the main varieties of soft stone from several districts. Some specific peculiarities may also be observed in the various types (Table 5). These peculiarities, together with the presence/absence and frequency of certain specific features, allowed the differentiation of the monotonous limestone into several subtypes and the assignment of samples to definite quarry regions. For white stone, the prevalence of coralline algae as allochems is a specific feature. Very pure carbonate composition, with small amounts of accessory minerals (mainly quartz and goethite) and the consequently lower percentage of insoluble residue (<2.5%) are other diagnostic qualities. The chemical composition in all quarries is also quite monotonous, with Ca as major element and very low values of others, mainly due to scarce other non-carbonate minerals. The varying proportions of coralline algal types (encrusting and articulated coralline algae) and other additional specific features (grain size, matrix content, foraminifera content) are helpful in further subdividing the white limestone into two subvarieties: San Gottardo Stone and Costozza Stone. The former is characterised by the prevalence of encrusting coralline algae (Lithothamnion) and coarse grain, and the latter by medium-fine grain, absence of these types of algae, large amounts of matrix, and benthic foraminifera (Miliolids). The two varieties reflect different energy conditions in the depositional environment, as also testified by the diagenetic features of the two rocks. Other conclusions may be drawn for the yellow varieties. Nanto limestone is fine-grained, yellowish-brown in colour, and characterised by the presence of benthic (Nummulites and a few Discocyclinas) and planktonic foraminifera (Globigerinoids), echinoderms, algae and serpulids. A characteristic feature is the high percentage of planktonic foraminifera and serpulids, which establish the depositional context. At the same time, the high non-carbonate fraction, about 12%, mainly composed of goethite, quartz, and traces of clay minerals, is characteristic. The other variety, Yellow San Germano Stone, often considered the same as Nanto Stone, can be distinguished first for the presence of macroscopic allochems of benthic foraminifera. However, if an initial distinction is not clear, this may be because the stone is cut along an unfavourable plane, since other features are diagnostic. This variety can be identified in particular by its grain size, generally coarser than that of Nanto Stone, but especially due to the absence of planktonic foraminifera, and to the equal proportions of two types of benthic foraminifera, Nummulites and Discocyclinas, both very common. Nanto Stone also has an elevated proportion of skeletal debris. In addition, the insoluble fraction is diagnostic, with values of about 6%, whereas composition seems quite similar, except for the lower amounts of clay minerals. This difference is also reflected in chemical composition. Lastly, the grey varieties have been demonstrated to have the same petrographical features as Yellow San Germano Stone, and reflect differences in diagenetic processes undergone. However, some peculiar features are noted; the percentage of insoluble residue, about 7-8%, and its composition, with high amounts of glauconite with respect to goethite, and the presence of pyrite. Some variations are also noted in chemical composition. For all quarry samples, microscopy provided the most useful information on typical rock
43
Table 5. Parameters for identification of main varieties of Vicenza Stone. PARAMETERS FOR IDENTIFICATION OF MAIN VARIETIES OF VICENZA STONE MINERALOGY INSOLUBLE STONE VARIETY PETROGRAPHY MINERALOGY OF INSOLUBLE RESIDUE RESIDUE SAN Coarse to medium grain. Encrusting (Lithothamnion) and articulated coralline Calcite <1% Quartz, Goethite WHITE GOTTARDO algae, benthic foraminifera (Nummulites), echinoderms and ostracods. VICENZA Medium to fine grain. Articulated coralline algae, benthic foraminifera STONE COSTOZZA (Miliolids and Nummulites) and bryozoans. Calcite <2.5% Quartz, Goethite High matrix contents. Goethite, Medium to fine grain. Planktonic (Globigerinoids) and benthic (Nummulites) Calcite, Goethite, K-feldspar, Quartz, NANTO 11-12% YELLOW foraminifera, echinoderms, bryozoans, and serpulids. K-feldspar. Quartz traces of Clay VICENZA minerals (Smectite) STONE YELLOW SAN Coarse to medium grain. Benthic foraminifera (Nummulites and Discocyclinas), Calcite, Goethite, 6-6.5% GERMANO echinoderms and intraclasts. K-feldspar K-feldspar, Quartz GREY Glauconite, GREY SAN Coarse to medium grain. Benthic foraminifera (Nummulites and Discocyclinas), Calcite, VICENZA 7-8.5% K-feldspar, GERMANO echinoderms and intraclasts. K-feldspar STONE Goethite, Pyrite
44 characteristics. Petrographic, mineralogical and chemical analyses carried out for the present research on samples previously collected, revealed a quite different composition from that of the most common facies, with different micro-environments of sedimentation and diagenetic patterns. Differences in the minor elements most useful for discriminating oxide-reducing conditions were also noted. On the basis of our database of potential types of limestone for building, it is now possible to match historically used material with potential source outcrops. Several case studies have shown that this is successful in detecting provenance. Examples from historical buildings in the city of Padova (unpublished data) demonstrate the usefulness of such an approach for both basic research and practical applications. The data can also contribute to interpreting weathering damage on historical monuments and providing means for their successful conservation and reconstruction.
11. Conclusions New analytical data presented here concerning the microscopic features of some limestones from the Berici Hills (Veneto, North-East Italy) were differentiated in detail. Neither their mineralogy nor their petrography is as monotonous as it may seem at first sight, relying only on descriptions with traditional study methods. This is significant in finding the sources of stone materials needed for historical buildings, and also the different weathering behaviour and weathering forms of several varieties. Although the above data must be interpreted with caution, because of possible petrographic variations in limestone types within a specific quarry area, the results emphasise that the combination of macroscopic rock descriptions, thin sections, mineralogical and chemical composition, and qualitative and quantitative evaluation of non-carbonate fractions is very useful in differentiating limestone materials. The distinction between the various types used for building becomes clearer than on the basis of observations by conventional optical microscopy. All field and laboratory data are stored in a database, formed of a main table, which contains essential information on the study sites. The resulting petrographical atlas provides a powerful tool for identification and decision-making in restoration.
45 References Baccelle, L., Bosellini, A. (1965): “Diagrammi per la stima visiva della composizione percentuale nelle rocce sedimentarie”. Ann. Univ. Ferrara, Sez. 9 (N.S.) 4, 59-62. Barbieri G., (1972): “Sul significato geologico della Faglia di Castelvero (Lessini veronesi)”. Atti Mem. Acc. Patav. SS.LL.AA., Vol. 84, 2, Padova, 297-302. Bartolomei G., (1958): “Resti di un carsismo terziario nei Colli Berici”. Actes Deuxième Congrès International de Spéléologie. Bari-Lecce-Salerno, 5-12 October 1958, Vol. 1, sez. 1, Putignano, 216-219. Bassi, D., Cosovic, C., Papazzoni, C.A., Ungaro, S. (2000): “The Colli Berici”. In: Bassi, D. (Ed.), Field Trip Guidebook. Shallow Water Benthic Communities at the Middle-Upper Eocene Boundary. Southern and North-eastern Italy, Slovenia, Croatia, Hungary. 5th Meeting of the IGCP 393. Annali dell’Università di Ferrara, (Suppl. 8), 43-57. Benvenuti G., Norinelli A. (1967): “Contributo geofisico alla conoscenza delle strutture sepolte tra I Colli Euganei e I Berici”. Boll. Geofis. Teorica Appl. , 9/36, Trieste. Berner R.A, (1984): “Sedimentary pyrite formation: an update”. Geochim. Cosmochim. Acta 48, 605-615. Bosellini, A. (1989): “Dynamics of Tethyan carbonate platforms”. In: Crevello, P.D., Wilson, J.L., Sarg, J.F., Read, J.F. (Eds.), Controls on Carbonate Platform and Basin Platform. Society of Economic Paleontologists and Mineralogists, Special Publication 44, 3-13. Breit G.N., Wanty R.B., (1991): “Vanadium accumulation in carbonaceous rocks: a review of geochemical controls during deposition and diagenesis”. Chem. Geol., 91, 83-97. Broglio Loriga C., (1982): “Geologia del territorio”. In AA.VV., Storia di Vicenza, Vol. I, Neri Pozza Edit., Vicenza, 1-13. Cattaneo A., De Vecchi G. P., Menegazzo Vitturi L. (1976): “Le pietre tenere dei colli Berici”. Atti e Memorie dell’Accademia Patavina di Scienze, Lettere ed Arti, LXXXVIII, 79- 100. Chemello F., Fabiani R. (1929): “ La pietra tenera: sue applicazioni artistiche ed edilizie”, 1° ed., Stab. Tip. Brunello, Vicenza. Chemello F., Fabiani R. (1939): “ La pietra tenera: sue applicazioni artistiche ed edilizie”, 2° ed., riv. E compl. Off. Tip. Vicentine, Vicenza. Cornale P., Rosanò P. (1994): “Le pietre tenere del vicentino”, La Grafica & Stampa Ed. s.r.l., Vicenza. Corsi M., Gatto G.O. (1967): “Tettonica”. In: Bosellini, A., Carraro, F., Corsi, M., De Vecchi, G.P., Gatto, G.O., Malaroda, R., Sturani, C., Ungaro, S., Zannetin, B. (Eds.), Note Illustrative della Carta geologica d’Italia, Foglio. Cosovic´V., Drobne K., Moro A., (2004):” Paleoenvironmental model for Eocene foraminiferal limestones of the Adriatic carbonate platform (Istrian Peninsula)”. Facies, 50, 61-75. Dean W.E., Gardner J.V., Piper D.Z., (1997): “Inorganic geochemical indicators of glacial– interglacial changes in productivity and anoxia on the California continental margin”. Geochim. Cosmochim. Acta 61, 4507-4518. Dunham, R. J. (1962): “Classification of carbonate rocks according to depositional texture”, in Ham, W. E. (Ed.), Classification of Carbonate Rocks: American Association of Petroleum Geologists, Memoir 1, 108-121. Embry, A. F., Klovan, J. E. (1971): “A Late Devonian reef tract on North-Eastern Banks Island”, N.W.T: Canadian Petroleum Geology Bulletin, v. 19, p. 730-781. [revision of Dunham classification]. Fassina V., Cherido M. (1985): “The Nanto stone deterioration and restoration of Loggia Cornaro in Padova”. Preprints of the Vth Int. Congr. “Deterioration and Conservation of Stone”, Lausanne, 313-324.
46 Flugel E. (2004): “Microfacies of Carbonate rocks”, Springer-Verlag Berlin Heidelberg. Frost, S.H. (1981): “Oligocene reef coral biofacies of the Vicentin, northeast Italy”. In: Toomey, D.F. (Ed.), European Fossil Reef Model, Spec. Publ. SEPM, vol. 30, 483-539. Geister J., Ungaro S. (1977): “The Oligocene coral formations of the Colli Berici (Vicenza, northern Italy). Eclog. Geol. Helv., 70/3, Basel. Girardi A., Mezzalira F., (1991): “Il Lago e le Valli di Fimon”, Publigrafica Editrice, Tavernelle (Vicenza), 31-47. Hastings D.W., Emerson S.R., Erez J. and Nelson B.K., (1996): “Vanadium in foraminiferal calcite. Evaluation of a method to determine paleo-seawater vanadium concentrations”. Geochim. Cosmochim. Acta 60 , 3701-3715. Ireland, B.J., Curtis, C.D. and Whiteman, J.A., (1983): “Compositional variation within some glauconites and illites and implications for their stability and origins”. Sedimentology, v. 30, 769-786. Jacobs L., Emerson S., Skei J., (1985): “Partitioning and transport of metals across the O2/H2S interface in a permanently anoxic basin: Framvaren Fjord, Norway”. Geochim. Cosmochim. Acta 49, 1433-1444. Luciani, V. (1989): “Stratigrafia sequenziale del Terziario nella catena del Monte Baldo (Provincie di Verona e Trento)”. Memorie di Scienze Geologiche 41, 263-351. Massari F., Medizza F., Sedea R. (1976): “L’evoluzione geologica dell’area euganea tra il Giurese superiore e l’Oligocene inferiore”, Mem. Ist. Geol. Min. Univ. Padova, v. 30, 174- 197, Padova. Meyers, W.J., (1980): “Compaction in Mississippian skeletal limestones, Southwestern New Mexico“. - J. Sed. Petrol., 50/2, 457-474, Tulsa. Mietto P. (1988), “Aspetti geologici dei Monti berici”, in “I Colli Berici-Natura e Civiltà”, Signum Ed., 13-23, Vicenza. Mietto, P., Sedea, R., Ungaro, S. (1981): “Note Illustrative del Foglio 50, “Padova””. In: Castellarin, A. (Ed.), Carta tettonica delle Alpi Meridionali (1:200.000). Servizio Geologico d’Italia, Roma, 99-103. Murray, J.W., (1991): Ecology and Paleoecology of Benthic Foraminifera. Longman Scientific and Technical, Essex. Odin G.S., Matter A., (1981): “De glauconiarum origine”, Sedimentology 28, pp. 611-641. Piccoli G. (1966): “ Studio geologico del vulcanesimo paleogenico Veneto”, Mem. Ist. Geol. Miner. Univ. Padova, 26, Padova. Pieri M., (1966): “Marmologia. Dizionario di marmi, graniti Italiani ed Esteri”. Hoepli editore, Milano. Piller WE, Pervesler P., (1989): “The northern bay of Safaga (Red Sea, Egypt): an actuopalaeontological approach”. I. Topography and bottom facies. Beitr Paläont Österr 15, 103-147. Rodolico F. (1963): “Le pietre delle città d’Italia”, Le Monnier, Firenze. Rossi D., Semenza E. (1958): “Le scogliere oligoceniche dei Colli Berici”. Ann. Univ. Ferrara, (NS), s.9, Sc. Geol e Miner. 3/3, Ferrara. Tucker, M.E. and Wright, V.P., (1990): Carbonate Sedimentology. In: Blackwell, Oxford. Ungaro S. (1972) – “L’Oligocene dei Colli Berici”. Rivista Italiana di Paleontologia e Stratigrafia, vol. 84, 199-278. Ungaro S., Bosellini A. (1965): “Studio micro paleontologico e stratigrafico sul limite Eocene-Oligocene nei colli Berici occidentali”, Ann. Univ. Ferrara, (NS), s.9, Sc. Geol e Miner., 3/9, Ferrara.
47
48 Chapter 2
Durability of carbonate lithotypes employed in historical monuments of Veneto (North-East Italy): a laboratory study
1. Introduction Natural stone has been used as a material for thousands of years to built monuments and other structures such as churches, castles and magnificent houses in cities, as well as in simple country houses and architectural creations. Ornamental stone was considered a durable material of high aesthetic value, but - especially today - it is clear that not even rock is a material that lasts forever. Degradation of natural stone is a process influenced by temperature changes, wind, biological factors, salt crystallisation and frost (Amoroso and Fassina, 1983; Camuffo, 1998). During the last few decades, attention has been paid to the increasing deterioration of old buildings and monuments built of natural stone. The problem has now been quite well studied and documented. It is generally accepted that rapid industrialisation and urbanisation have caused a significant increase in the rate of decay, as a result of human action causing increased pollution in the atmosphere, especially of gaseous and particulate sulphurous compounds. Therefore, much of the research on this subject has focused on chemical reactions between stone and the products of air pollution (Fassina, 1988, 1992, 1999). Deterioration of old stone buildings and monuments is giving rise to considerable costs to private and public owners all over the world, as it is considered important to preserve these relics of the past for future generations. To be able to do this properly, knowledge of how local stone deteriorates must be enhanced, especially in a region like the Veneto which contains many precious monuments. Rosso Ammonitico Veronese and several varieties of Vicenza Stone are two rock types that have been extensively used in the past in this area, for both buildings and sculptures (Rodolico, 1963). Recent repair plans for some monuments made of these types of stone require materials for substitution purposes. The original stone type is difficult to find because, over time, many quarries have opened and closed, and extensive exploration of abandoned quarries lost over the centuries is costly. The other possibility is to use alternative types of stone, particularly those actually taken from the same ancient quarries, i.e. with similar or identical petrographic, mineralogical and chemical characteristics, and requires numerous laboratory tests including long-term durability tests (freeze-thaw cycles, resistance to salt crystallisation) and evaluations (measurement of porosity, micro-porosity, water absorption) of actual quarry materials. According to test results, certain quarried stones are more or less resistant to weathering agents than others, and it is thus possible to establish which are most suitable for replacement purposes. However, as quarrying in the Berici and Lessini Hills (the districts providing Vicenza Stone and Rosso Ammonitico Veronese, respectively) has been devasted over the centuries, work has been limited recently to only a few sites, in order to lessen the environmental impact. The result is that a single variety is currently quarried in only one place, and in some cases nowhere. Fortunately, some ancient sites have been found and their materials have been examined. In the Po Plain, various degradation processes, both natural and artificial, interact with each other in a complicated way, so that deterioration patterns are more complex, involving the combined actions of water, freezing/thawing, salt damage, and chemical reactions (Amoroso and Fassina, 1983). Hydric behaviour, the sensitivity of the stone to ageing by weathering, and how each single factor accelerates the degradation of selected rock types are still not well understood, not having been systematically studied before. Estimation of durability in rocks with differing textural parameters (in terms of grain-supported or mud-supported; type of interstitial material; porosity) and correlations with intrinsic properties and depositional and diagenetic features has not been exhaustively discussed in the 49 literature, although Myrsini Varti-Matarangas and Dionysis Matarangas (2000) discussed correlations between microfacies characters and deterioration, and Nicholson and Nicholson (2000) and Nicholson (2001) reported the behaviour of mudstone and sandstone with lithic clasts and nodules subjected to experimental freeze-thaw cycles, wetting and drying, and salt weathering. The mineralogical composition, texture and characteristics of voids are the main factors involved in estimating the intensity of physical and chemical damage in rocks when subjected to new environmental conditions (Hudec, 1998). When we recall that water is the main weathering agent and that open microcracks are the natural way for water to penetrate inside rock, the importance of understanding microcrack characteristics and their physical properties (e.g., open porosity) becomes clear. For more information about the dependence of the fabric on the petro-physical properties of natural limestone (porosity, pore size distribution, water absorption) and correlations between them and frost and salt resistance, various types of limestone with differing textural properties have been studied. Due to the complex nature of rocks, frost action and salt crystallisation mechanisms, all rock fabrics, textures and structural elements should be considered. Rock texture includes size, shape, preferred orientation of clasts, their contacts and interstitial materials; and pore structure, defined by porosity, pore size and fissure distribution (Durrast and Siegesmund, 1999; Prikryl, 2006). We place particular emphasis on the influence of the size and spatial distribution of pores and fissures of rocks on decay mechanisms and their relationship with diagenetic processes. Evaluation of the endogenic causes of decay was also attempted, since most of them (porosity, texture, etc.) are microfacies features, controlled by sedimentary and diagenetic processes. All this information is useful for planned action on the most sensitive factors influencing stone deterioration, compatibly with single cases.
2. Materials and Methods 2.1 Types of Limestone Studied Here Vicenza Stone and Rosso Ammonitico Veronese have older origins as building material, and have indeed been used since Roman times in the architectural heritage of the Veneto. The Basilica Palladiana and the Ville Venete (Vicenza, white stone), the Prato della Valle sculptures (Padova, white stone), the Loggia Cornaro (Padova, Nanto Stone) and the Arena of Verona (Rosso Ammonitico) are some examples of their use. The name “Vicenza Stone” is given to a group of rocks outcropping in the province of Vicenza and quarried along the slopes of the Berici Hills. The group comprises several varieties belonging to various geological formations, with various names and petrographical, mineralogical and chemical characteristics. Thanks to their typical softness, they are relatively easy to extract and cut, and have therefore been used since ancient times. The Oligocene white limestone of the Calcareniti di Castelgomberto Formation, which provides the “White Vicenza Stone” was quarried extensively in the Berici Hills, and the stone was often given the name of the nearest locality. All types of quarried stone have similar characters but, in some cases, in some localities distinctive petrographic properties distinguish two sub-varieties of White Vicenza Stone: San Gottardo Stone and Costozza Stone. The other main variety is Nanto Stone (quarried near Nanto, SW sector), a yellowish-brown limestone of Middle Eocene age, stratigraphically placed at the bottom of the Calcari Nummulitici Formation (Massari et al., 1976). San Germano Stone, quarried in the central-eastern area of the Berici Hills, comprises two sub- varieties characterised by different colours, Yellow San Germano Stone and Grey San Germano Stone. They are currently quarried at Pederiva di Grancona, inside the Calcari Nummulitici Formation (Mietto, 1988). Until recently, Nanto Stone and Yellow San Germano Stone were often considered to be the
50 same variety, but detailed petrographic examination has demonstrated their differences. Rosso Ammonitico Veronese (RAV) is a fine-grained nodular limestone, varying in colour from red to yellow or white, locally rich in internal moulds of ammonites, characteristic of the Middle -Upper Jurassic (Upper Bajocian-Tithonian), with frequent discontinuities. It is highly appreciated as an ornamental stone and extensively employed in monuments. The major quarries are located in the Lessini Hills and near Asiago; in this work, we examined two quarries of the Lessini district. Five heterogeneous limestones quarried in the Berici Hills, San Gottardo Stone (SGS), Costozza Stone (CS), Nanto Stone (NS), Yellow San Germano Stone (YSGS) and Grey San Germano Stone (GSGS), and two homogeneous anisotropic limestones of the Lessini Hills, Rosso Ammonitico Veronese V (RAVV) and Rosso Ammonitico Veronese F (RAVF), were chosen, due to their complex and at the same time different rock fabric, and also for the economic impact of their use as building materials. Chosen for their differing textural characteristics, they include carbonate rock types differentiated by stratigraphic interval, size, type and percentage of allochems, pore-filling component (interstitial material, e.g., cement or matrix) and depositional and diagenetic textures.
2.2 Analytical Techniques and Methods Quarried materials were qualitatively and quantitatively characterised by minero-petrographic and physico-chemical diagnostic studies. First, preliminary characterisation was performed on samples by X-ray diffraction (XRD); Philips X’Pert Pro with θ-θ geometry with CuKα (1.54 Å) radiation filtered by nickel in operational conditions: 40KV and 40mA. The structural and textural properties of the samples were observed in thin section, on a transmitted and reflected light optical microscope (OM), a stereoscope, and back-scattered electrons on a Camscan MX2500 scanning electron microscope equipped with an energy dispersive microanalyser (SEM-EDS). Characterisation of all samples and insoluble residue (obtained by 8% HCl attack) was performed by X-ray diffraction (XRD). Elemental composition was performed by X-ray fluorescence (XRF) by WDS Philips PW2400 sequential spectrometry. The petrographic classification adopted is that of Dunham (1962), modified by Embry and Klovan (1971); the percentage of various skeletal grains was estimated with the comparative charts developed for limestone by Baccelle and Bosellini (1965).
2.2.1 Petro-Physical Properties The basic petro-physical properties of fresh, unweathered materials were ascertained, and included total and open porosity, bulk density, and water absorption.
Pore structure. As regards the porosity range, in order to facilitate data description, interpretations refer to the following classification of pore spaces based on water absorption properties: - microporosity (pores with radius < 1 μm); - mesoporosity (pores with radius 1-150 μm); - macroporosity (pores with radius > 150 μm). Mesoporosity and microporosity was determined on a mercury porosimeter. Macroporosity was studied by analysis on thin sections. First, porosity, pore size distribution, pore shape and discontinuities visible at optical microscopy on thin sections (30 μm thick) were qualitatively evaluated.
Open porosity and real density. The pores of the original samples accessible to water were evaluated by the following procedure. Porosity determination is essentially very simple: samples were dried in a ventilated oven and specimen mass, dried at 100 (±5°C) to constant weight, is
51 measured (md). Then the specimen is placed in a vacuum and immersed in distilled water until all open pores are completely filled with water. Its mass is then measured, both under water (mw) and normally in air (mm). According to Archimedes’s law, open porosity can be calculated from: mm md OP% 100 mm mw
For these tests, ten stone samples of each variety with dimensions 5×5×2 cm were employed.
Total porosity, bulk density and pore size distribution. Approximately 1 cm cubes of intact stone material were taken from each sample and subjected to mercury porosimetric testing. Prior to testing, the material was dried to constant mass in a ventilated oven at 105±5°C. Total porosity, pore volume, bulk density and pore size distribution were obtained on a Carlo Erba 2000 porosimeter. Pressure ranges were 1–2000 bar, and data for pore sizes between 0.01 μm and 200 μm were obtained. The volume of intruded mercury for each pressure step allowed determination of total open pore volume and pore size (volume) distribution. The range of measured pore diameters was given by Washburn’s equation: P = -4 σ cos φ / d where d is pore diameter, σ the surface tension of mercury, φ the mercury wetting angle of the pore surface (φ>90°) and P the absolute pressure exerted. The smallest pore diameter is limited by maximum applied pressure. The instrument calculates bulk density from the sample weight and volume at the initial mercury filling pressure. Data were logged automatically and computer-processed into spreadsheets. Also in this case, ten specimens of each stone were considered, in order to have significant data.
Water absorption properties. Water absorption properties on fresh, unweathered material were ascertained.
Water absorption by total immersion. This was carried out following Italian NORMAL 7/81 on 6 specimens of 5×5×2 cm for each quarried lithotype. Test procedure. Water absorption experiments (with distilled water) were carried out at ambient temperature and humidity. Samples were dried in a ventilated oven at 60±5°C to constant weight and put in a desiccator to allow them to cool down. A plastic container was set up, with glass balls on its bottom surface, and samples were placed with one square surface on them. Distilled water was added until the upper surfaces of the samples were completely covered by 2 cm of water. The container was closed (but not sealed, in order to avoid condensation). Samples were weighed at timed intervals until they were saturated as indicated by a difference in mass between the two last weighings of less than 0.01 g. After saturation, they were placed in an oven at 60±5°C to constant mass and then weighed. Results from total absorption tests are expressed in a graph as average water absorption versus time for each set of specimens. The water absorption of each sample at time ti was calculated according to:
= where Mwi is the wet mass (g) of the sample at time ti, and Md is the mass (g) of dried sample. Imbibition capacity (IC) was calculated by the following formula:
100
52 where Ms is the mass (g) of the saturated sample and Md the final mass (g) of the dried sample.
Water absorption by capillarity. The UNI 10859:2000 procedure was followed. Experiments were undertaken to determine the capillarity absorption properties of each type of limestone considered here. Materials were quarried stone used for construction work. Three parallelepiped specimens of 5×5×2 cm were cut from each limestone and washed in distilled water to remove all traces of dust. A standard experimental set-up was used, in which one face of the specimen was placed in contact with a liquid surface, and the amount of liquid absorbed as a function of time was determined by weighing the specimen at appropriate intervals. Specimens were first dried at 60±5°C to constant mass, and put in a desiccator to allow them to cool down. Water absorption experiments (distilled water) were carried out at ambient temperature and humidity. Test procedure. Glass balls were placed on the bottom of a plastic container and samples were arranged on them. Distilled water was added until only the bottom surfaces of the samples were in contact with the water. The container was closed (but not sealed, to avoid condensation). Samples were weighed at timed intervals until they were saturated. Before they were weighed, the wet surface was wiped with a damp sponge to remove surface water. Saturation was indicated by a difference in mass between the two last weighings of less than 0.01 g. Readings of mass weight were taken at the following intervals (in minutes): 0, 15, 30, 60, 120, 180, etc. Results of capillarity tests are expressed as absorbed water per unit of area at time ti, Qi 2 (mg/cm ), versus the square root of time. For all samples, Qi was calculated as:
Qi where Mwi is the wet mass (g) of the sample at time ti, Md the mass (g) of the dried sample, and 2 A the area of the sample in contact with water, expressed in cm . Qtf (Qi for t=final moment of experiment) was also calculated. Capillarity (CA) coefficients were calculated by determining the slope of the respective curves in the initial linear section.
2.2.2 Experimental Laboratory Weathering Simulations Resistance to weathering (durability) is one of the crucial factors considered when specifying stone for use in outdoor environments. Lengthy experience with some stone types was the only available measure of susceptibility to deterioration before some standardised test procedures were introduced for quantitative and qualitative data on their behaviour. They involve freeze- thaw cycles and salt crystallisation. Freeze-thaw mechanisms have long been recognised as one of the most widespread and significant causes of stone decay, and direct accelerated freeze-thaw tests are one of the most frequently used procedures for prediction of stone durability. Although overall various methods are used, freeze-thaw tests usually involve relatively small samples of stone, which are sooked in water for a given length of time, and then subjected to cycles of freezing in air and thawing in air or water. In the Veneto region, continental climatic conditions prevail. The numbers of freeze-thaw cycles throughout the year is considered to be about 20, so that freeze-thaw processes are quite effective. In this study, up to 45 such cycles were performed to assess the resistance to frost of the stone. Some physical properties were recorded at different test cycles and compared with those of fresh samples. Relative changes in properties were applied to assess test effects. The other main cause of weathering of building stone is crystallisation of soluble salts. These usually originate in the ground or in the mortar, but may also form part to the stone itself, since they flow as a solution through a complex capillary system towards the exterior. In favourable conditions, these salts crystallise as efflorescences on the stone surface, or as subefflorescences below it. This mechanism, which can lead to the complete decay and destruction of the stone, is
53 also influenced by the nature of the substrate, climatic conditions, and salt-system characteristics.
Decay test description Accelerated ageing tests (freeze-thaw and salt crystallisation cycles) were performed to quantify susceptibility to deterioration of stone samples.
FROST TEST 1. Preparation of specimens Six cube specimens (5×5×5 cm) of each limestone facies were prepared by cutting blocks with diamond saw and washing them in distilled water to remove all traces of dust. 2. Drying specimens Specimens were dried in a ventilated oven at a temperature of 60±5°C to constant mass. This was assumed to have been attained when the difference between two weighings at an interval of 24 h was not greater than 0.01 ‰ of the last measurement. 3. Immersion of specimens Specimens were placed in a container at least 15 mm from adjacent ones. Then tap water was added (20±10°C) up to half their height. Twenty-four hours later, tap water was added to three- quarters of the height of the specimens. After 48 h, tap water was added until they were completely immersed under 20 mm of water. They were then left completely immersed for 48 h. 4. Test procedure 4.1 Arrangement of specimens in climatic chamber inside water bath. Specimens were placed in a tank so that they did not touch either each other or the sides of the tank. They were positioned at least 10 mm apart, and at least 20 mm away from the tank sides. Glass balls were set on the bottom. The tank was then filled with distilled water. 4.2 Description of freeze/thaw cycles. Each cycle was as follows:
Time Start of cycle Saturated specimens are put into tank T0 Stage 1 Temperature reaches -20±0.5°C T0+2h Stage 2 Temperature remains at -20±0.5°C T0+4h Stage 3 Temperature reaches 20±0.5°C T0+6h Stage 4 Temperature remains at 20±0.5°C T0+8h
The cycle was repeated 45 times.
30
20 10 T (°C) 0 -10 0 2 4 6 8 -20 -30
Time (h)
Fig. 1. Frost resistance cycle.
54 RESISTANCE TO SALT CRYSTALLISATION This test is one of several used to assess the durability of natural stone, and describes one method for determining its relative resistance to salt crystallisation. Solution: a 14% solution of sodium sulphate decahydrate (i.e., 14g of Na2SO4·10H2O for every 86 g of deionised water) was used (specific gravity at 20°C, is 1.055). Procedure: 1. Three specimens of each lithotype (considered representative of the body of stone tested) were selected from a homogeneous batch. 2. Samples of 5×5×2 cm were prepared and cut with a diamond saw; surface irregularities were removed by grinding. All loose material was washed away with distilled water. 3. Specimens were dried in an oven at 60±5°C to constant mass, which was assumed to have been attained when the difference between two weighings at an interval of 24±2h was not greater than 0.01‰ of the mass of the specimen. They were then cooled to room temperature, weighed to ±0.01g (Mi), and labelled with a durable tag wired on. 4. The procedure entails the use of a 14% solution of sodium sulphate decahydrate. The specific density of the solution was checked prior to use. 5. Dried samples were placed in a suitable container and covered with the sodium sulphate solution to a depth of 20±2 mm, i.e., immersing them entirely. There was a minimum of 10 mm between samples and at least 20 mm between the samples and the sides of the container. Glass spheres were placed at the bottom of the tank. 6. After 4 h immersion, samples were removed and dried in an oven, set to provide high relative humidity in the early stages of drying and to raise sample temperature to 60±5°C in not less than 10 and not more than 15 hours. Initial high relative humidity was obtained by placing a tray of water in the cold oven, and switching on the heater for 30±5 minutes before putting in the samples. 7. Samples were left in the oven for 18±2 h and then cooled to room temperature for 2±0.5h before weighing and resoaking in fresh sodium sulphate solution for the next cycle. 8. These cycles were carried out six times or to failure of samples. 9. Samples were weighed after drying to constant mass (at 105±5°C) if they were sufficiently solid to determine their weight.
2.2.3 Measurement Strategy Pre- and post-deterioration simulation measurements of specific parameters were conducted on the same samples, to allow accurate observation of changes without increasing measurement error by the natural variation of properties (heterogeneity) commonly recorded in natural stone. Damage induced was evaluated first by visual inspection of material loss and weight changes on different runs. Frost test. Weight, and Imbibition Capacity (Ic) measurement variations were estimated on samples subjected to freeze-thaw testing, prior to freeze-thaw cycling, after 16 and 32 cycles and again after 45 cycles. Salts. Weight loss at each cycle and accumulation of salts were measured. Petrographic examination was carried out on all stone types prior to salt cycling and also after six cycles in cases where the tested sample was sufficiently intact to allow a thin section to be prepared.
3. Results and Discussion 3.1 Characterisation of Lithologies Abbreviations: encrusting coralline algae (Lithothamnion, ECA), articulated coralline algae (ACA), benthic foraminifera (BF), planktonic foraminifera (PF), echinoderms (E), molluscs (M), bryozoans (B).
55 San Gottardo Stone (SGS) Petrography. This is a white to straw-yellow bio-calcarenite, very badly sorted (100 μm to few millimeters in grain size) and coarse-grained (average size 1-1.5 mm, but several grains reach pluri-millimetric to centimetric length). The texture is clastic-organogenic with grain support. Grain contact types are point and tangential contacts. Less frequent are concavo-convex and sutured contacts. The structure is essentially isotropic, although in some portions slight preferential orientation of lengthened allochems (especially ACA) was observed. Porosity is very high (20-30% by comparison chart) and mainly found in samples as primary inter-granular type. Large pores are very frequent; average value is 300 μm, but many reach pluri-millimetric size. Most of this porosity is formed of interconnected pores, as examination of perpendicular sections of samples showed. The main components (>40% by comparison chart) of this facies are allochems of coralline red algae, both encrusting coralline algae (Lithothamnion, ECA) and articulated coralline algae (ACA), the former generally prevailing. Benthic foraminifera (Nummulites, 10-20%), echinoderms (10%), bryozoans, molluscs, ostracods and corals are the other main components. Pore-filling materials are scarce; the matrix is present in a few isolated portions. Three main types of cement are present: isopachous cement (multiple cement rims growing with equal thickness around grains; the cement rims consist of fibrous and microcrystalline crystals) around coralline red algal clasts; isolated intergranular and intergranular blocky sparry and micro-sparry carbonate cement; syntaxial overgrowth rim cement on echinoid fragments. Quartz, feldspar and goethite-limonite are accessory constituents. Petrographically, the SGS is classified as rudstone-grainstone, in some cases passing to bindstone (Fig.2). Diagenesis. The most important processes occurring in subsurface carbonates are compaction and pressure solution. Burial of non-cemented carbonate sediments under an increasing burden results in compaction and, if continued, in pressure solution. Mechanical compaction may eventually continue on to chemical compaction if the grains begin to dissolve at their contacts. The specific criteria of these processes can be recognised on different scales, including thin sections. Diagenesis, creating situations of differing degrees of cementation and compaction, has contributed to the specific textural characters of this stone type. The presence of primary pores, partly filled with isopachous equant calcite cement, and the dominant point and tangential contact types between grains suggest that this lithotype suffered relatively earlier cementation, opposing compaction. Early cementation does protect grains against compaction and pressure solution. Point contacts indicate initial compaction, and tangential contacts increasing compaction. The lower number of sutured and concavo-convex contact points reveal no effects of pressure solution at grain contacts. The preservation of open interparticle porosity was also favoured by the relatively stable mineralogy of some skeletal grains (high-Mg calcite) and probably to burial solution, with enhancement of interparticle pores. The preservation of primary pores requires post-depositional diagenesis to be limited in its pore-destroying effects, compaction to be kept to a minimum, and fluctuations between exposure and sub-emergence to create a balance between the formation of solution pores and the destruction of porosity by shallow burial cementation. Preservation of porosity in shallow burial environments is a consequence of minimal burial, reduced burial stress, increased framework rigidity, exclusion of pore water, low calcite mineralogy, permeability barriers, and pore resurrection. Skeletal architecture probably also influenced the preservation of primary porosity, with low percentages of microcrystalline skeletal grains (e.g., bryozoans) which are more susceptible to mechanical and chemical diagenesis than more frequent, single-crystal echinoderm grains (Meyers, 1980). Grain size and sorting are other two factors influencing porosity; coarse grain and badly sorting favour large pores. It is clear that the high porosity of these rocks is linked to their specific
56 diagenetic features. Mineralogy. XRD analyses of bulk sample show only calcite. The percentage of insoluble residue is lower than 1% and qualitative mineralogical analysis showed that it consists of quartz, feldspars and goethite. Geochemistry (major and trace elements). Results (Table 1) show that the composition of major elements is generally very monotonous, with typical high contents of CaO and very low concentrations of other elements. Only Mg is dissimilar, and is due to the large amounts of coralline algae, which may contain up to 30% of MgCO3. Very low contents of SiO2 are related to trace amounts of quartz and feldspar. In general, even the trace element composition of the limestone is quite unvarying, because of the low concentrations of minor minerals; the only elevated contents are those of Sr and Ba, due to carbonate allochem contents.
Fig. 2. Rudstone-grainstone with encrusting and articulated coralline algae, benthic foraminifera and echinoderms. San Gottardo Stone. Nicols //. Costozza Stone (CS) Petrography. This is a medium- to fine-grained limestone (average size 500-700 µm, with some clasts of millimetric length). It is characterised by very white to straw-yellow colour, with moderate sorting (0.3–0.7mm). The texture is clastic-organogenic with grain support. The main grain contact types are point and tangential contacts; in some portions, concavo-convex and sutured contacts occur. The structure is essentially isotropic. Porosity is high (15-20% by comparison chart) and of intraparticle and interparticle type, the latter dominant. Pore sizes range between a few microns to 1 mm, with an average of 0.2mm; distribution is isotropic. The main components of this facies are bioclasts (>50% by comparison chart) and are represented by prevailing ACA (30-40%), benthic foraminifera (mainly Miliolids and Nummulites, 10-15%), bryozoans (10%), echinoderms and molluscs. Interstitial material is
57 abundant, mainly composed of fine-grained intergranular carbonate matrix which surrounds clasts and is uniformly distributed. Intergranular and intragranular carbonate, sparite and microsparite cement are present, and fill the micropores of the matrix. Isopachous fibrous and microcrystalline crystal cement are found around some coralline red algal bioclasts. More evident is the clear calcite syntaxial overgrowth cement on plates and spines of echinoderms. Quartz, feldspar and goethite-limonite are accessory constituents. Petrographically, CS is classified as packstone (Fig. 3). Diagenesis. The presence of some primary pores partly filled with isopachous equant calcite cement, and point, tangential, and also sutured and concavo-convex contact points between grains suggest that this lithotype suffered relatively earlier cementation, only partially opposing compaction. The presence of large numbers of Miliolids, organisms typically found in areas near emerged land, suggest a low-energy depositional environment, as confirmed by large amounts of matrix, due to quite stable sediments subjected to only weak currents. The presence of this matrix decreases primary interparticle porosity. Skeletal architecture probably also influenced the non-preservation of primary porosity, with high percentages of microcrystalline skeletal grains (e.g., bryozoans) which are more susceptible to mechanical and chemical diagenesis than more frequent, single-crystal, echinoderm grains (Meyers, 1980). Grain size and sorting are other two factors influencing porosity; medium grain and scarce sorting probably favour compaction processes. Mineralogy. This rock type is mainly composed of calcite, with traces of quartz. The non- carbonate fraction is very low, ranging from 1 to 2%, and includes quartz and goethite. Geochemistry. No meaningful differences were observed with respect to San Gottardo Stone.
1000 µm
Fig. 3. Packstone with articulated coralline algae, benthic foraminifera (Miliolids, Nummulites) and bryozoans. Costozza Stone. Nicols //.
58 Nanto Stone (NS) Petrography. Nanto Stone is a yellowish-brown medium- to fine-grained limestone of Middle Eocene age, outcropping near Vicenza and quarried along the slopes of south-western sector of the Berici Hills. This stone, often used locally for its aesthetic qualities and easy working, undergoes severe decay processes when exposed to an urban environment (Cattaneo et al., 1976; Fassina and Cherido, 1985). Nanto Stone is petrographically classified as grainstone-packstone (Fig.4); the matrix is absent, porosity is high, and cementation is fair. The texture is clastic-organogenic, with grain support. The structure is essentially isotropic. The skeleton is mainly due to benthic (Nummulites, Assilines) and planktonic foraminifera (Globigerinoids) (30-40%), plates and spines of echinoderms (20-30%), red algae, and a little rare, micron-sized skeletal debris. Skeletal grains, with sutured grain boundaries, are cemented by sparite and micro-sparite of carbonate crystals, and rare goethite. More evident is the clear calcite syntaxial overgrowth cement on plates and spines of echinoderms. More frequent is matrix (10%), mainly intergranular, and frequent cases of pseudomatrix (from intraclasts) occur. The mesopores are cemented by drusy calcite. The siliciclastic fraction is made up of a few isolated grains of biotite, glauconite, chlorite, phosphates, pyrite, quartz, feldspars, and clay minerals. Apart from echinoid allochems, other components appear as very fragmented allochems (bioclasts) and the entire structure is quite isolated. Grain contact types are mainly concavo-convex and sutured contacts; point and tangential contacts are scarce. Porosity is medium-high (15% by comparison chart) and preferentially of intragranular type. Large pores are very scarce; their average value is 100-200 μm.
1000 µm
Fig. 4. Grainstone-packstone with planktonic and benthic foraminifera, echinoderms and red algae. Nanto Stone. Nicols //. Diagenesis. By creating situations of different cementation and compaction, diagenesis has 59 contributed to modifing the textural character. This rock is more compacted, as shown by the penetrating and sutured contacts among the granules and the high number of fragmented allochems. This lithotype does not seem to have suffered earlier cementation opposing compaction. The presence of a pseudomatrix confirms compaction. Large pores are also very scarce. Clay-bearing rocks are often more susceptible to strong compaction than clay-free ones. In limestone/marl sequences, limestone is more or less compacted (Munnecke and Samtleben, 1996). The exact effects of clay (e.g., decreasing permeability to cementing pore waters, enhanced diffusion of dissolved CaCO3) require further study. Mineralogy. The samples are mainly composed of calcite, although goethite is also abundant. The insoluble residue (non-carbonate minerals) is high, around 12%. Qualitative analyses showed that it mainly consists of goethite and swelling clay minerals of the smectite group (i.e., montmorillonite, smectite–illite). Minerals of the chlorite group are also present, together with feldspars and quartz. Geochemistry. Results of geochemical analyses of the bulk limestone material (Table 1) show that major elements have high contents of SiO2, followed by allumina and magnesium. These elements correlate well with clay minerals present in the stone. Fe2O3 has also large amounts in relation to iron oxide-hydroxide (goethite-limonite). Slightly elevated contents of Al2O3 and K2O are due to small amounts of feldspar and illite. Mica and clay minerals can also cause a slight increase in the concentrations of Al2O3, K2O, Fe2O3, Cr and V. In general, trace element composition is quite variable because of the concentrations of minor minerals; the highest contents of Sr and Ba are due to carbonate allochem contents.
San Germano Stone (SGS)
1000 µm
Fig. 5. Rudstone-packstone with benthic foraminifera (Nummulites and Discocyclinas), red algae and echinoderms. Yellow San Germano Stone. Nicols //.
60 Petrography. This limestone has two well-defined varieties, Yellow San Germano Stone (YSGS), in view of its pale yellow colour, and Grey San Germano Stone (GSGS), ash-grey in colour. The main differences between the two varieties are colour and origin; in the yellow variety, oxidation of ferrous silicate (glauconite) is the main agent responsible for pigmentation. In the grey variety, glauconite is more or less weathered, and alteration products (goethite, limonite) are very scarce. These two stones are quarried in the same place, but have evidently been subjected to different oxide-reducing conditions, as confirmed also by the high content of organic fraction in the grey stone. These limestones are coarse- to very coarse-grained (average size 1-1.5 mm). The presence of macro-bioclasts (macroforaminifera, Nummulites and Discocyclinas), from a few millimeters to several centimeters in size, is also very characteristic. The texture is clastic-organogenic with grain support. The main grain contact types are concavo-convex and sutured contacts; point and tangential contacts are rare. The structure is essentially isotropic, although some portions show preferential orientation of the largest bioclasts; tiled roof structure is also present in some transversal sections.
1000 µm
Fig. 6. Rudstone-packstone with benthic foraminifera (Nummulites and Discocyclinas), red algae and echinoderms. Grey San Germano Stone. Nicols //. Porosity is high (15-20%), and of intraparticle and interparticle type, as it develops within and between fossils. Pore size ranges between a few microns to 1.5 mm, with an average of 0.1 mm. The matrix is frequent (5-10%) and mainly intergranular. Examples of pseudomatrix also occur. The cement, both intragranular and intergranular, is not very abundant and mainly consists of microcrystalline calcite mosaics, but also sparitic and microsparitic material. Clear calcite syntaxial overgrowth cement is found on plates and spines of echinoderms. The stone consists predominantly of bioclasts of macroforaminifera (40-50% by comparison charts) such as Discocyclinas and Nummulites, followed by red algae (5-10%) and echinoderms (5%). 61 Intraclasts and/or micritised grains are frequent (10-20%). Detrital quartz grains (3–5%), green grains (glauconite), clay minerals and oxides-hydroxides of iron (goethite-limonite) are other non-carbonatic constituents. Fragmented allochems are frequent. Petrographically, San germane Stone is classified as rudstone-packstone (Figs. 5, 6). Diagenesis. Considerations for these varieties are similar to those for Nanto Stone. The difference is the variation in oxide-reducing conditions. Mineralogy. The samples are mainly composed of calcite, although goethite is also abundant. The insoluble residue (non-carbonate minerals) is different in the two varieties, about 6% for yellow stone and 7-8% for the grey one. Qualitative analyses show that it consists mainly of goethite, quartz and K-feldspar in the yellow variety. The grey variety contains significant amounts of glauconite and also pyrite. Geochemistry. The same considerations as for Nanto Stone are made.
Rosso Ammonitico Veronese (RAV) Rosso Ammonitico Veronese is a red nodular limestone, locally rich in internal moulds of ammonites, characteristic of the Middle-Upper Jurassic (upper Bajocian-Tithonian). It was deposited on the distal Trento Plateau (Gaetani, 1975; Bosellini and Martinucci, 1975; Bosellini and Winterer, 1975; Winterer and Bosellini, 1981) on the southern continental margin of the Tethyan Sea. Several facies occur, and differ in structure (presence and type of nodularity) and texture (nature of components, grain vs mud support). Many discontinuities are recognised and consist of stylolites, locally increasing in amplitude near calcite veins. RAV is quarried in many localities of the Veneto, but the main area is that along the slopes of the Lessini Hills, in the province of Verona.
Fig. 7. Wackestone-packstone with molluscs, radiolaria and echinoderms. RAVV. Nicols //. In the Verona area, the RAV is less than 30 m thick and is subdivided into three units and eight
62 different facies (pseudonodular, mineralised, bioclastic, nodular, thin-bedded limestone, thin- bedded cherty limestone, subnodular, stromatolitic) (Martire 1992, 1996; Clari and Martire 1996). The lowest unit, the Rosso Ammonitico Inferiore or RAI (upper Bajocian-lower Callovian) is formed of pseudonodular, mineralised and massive facies; the middle unit, the Rosso Ammonitico Medio or RAM (upper Callovian-middle Oxfordian) is formed of thin bedded, cherty and subnodular limestone; the upper unit, the Rosso Ammonitico Superiore or RAS (lower Kimmeridgian-upper Tithonian) is composed of stromatolitic, pseudonodular and nodular limestone. The pseudonodular facies (RAI) consists of wackestones with bivalves (Bositra), planktonic and benthic foraminifers (Protoglobigerina, Lenticulina), ostracods, radiolarians and peloids. The bioclastic facies are formed of peloidal packstone and rare grainstone with bivalves and crinoids. The thin-bedded facies (RAM) consists of red chert lenses, represented by bivalve wackestone–packstone and mudstone with radiolarian moulds. The nodular facies (RAS) consists of bioclastic packstone with bivalves, foraminifers and crinoids. Facies of RAI (RAVV) and RAS (RAVF) respectively was considered in this study.
Fig. 8. Packstone with molluscs, radiolaria and echinoderms. RAVF. Nicols //. Petrography. Both RAVV and RAVF are petrographically classified as wackestone-packstone; the matrix is abundant; porosity is very low, and cement occurs primarily as a filler of veins with blocky sparite cement. The structure is anisotropic, with frequent discontinuities and veins, especially in the RAVF variety. The texture is clastic-crystalline-organogenic, with mud and/or grain support. The skeleton is mainly due to molluscs (Posidonia sp., Bositra sp.), with an elevated degree of packing. Echinoderms (crinoids), sponge spicules and radiolaria are also present. In the confluence of calcite veins, macrocrystalline calcite masses and large pores associated with dissolution processes also occur.
63 RAVF (Fig. 8) is characterised by a red colour and nodule-rich fabric. Other peculiar features are the presence of bioclasts with hematite infillings. The petrographic features of pink-red samples (RAVV, Fig. 7) are generally comparable with those of the reddish ones, although some differences are observed: pink slightly bioturbated matrices; or bioclasts not completely filled or surrounded by iron oxides, as observed in the red samples. Discontinuities are few. Diagenesis. Unidirectional pressure and consequent compaction may provide the pore solution migration necessary for stylolitisation. Solution pressure acts before complete reduction of pore spaces by cementation, and is indicative of continuity during diagenetic stages. Thus, the presence of these structures clearly indicates high compaction of sediment, also suggested by overpacking of mollusc shells and very low porosity.
Table 1. Chemical Composition of Stone varieties (SGS: San Gottardo Stone; CS: Costozza Stone; NS: Nanto Stone; YGSG: Yellow San Germano Stone; GSGS: Grey San Germano Stone; RAVV: Rosso Ammonitico Veronese V; RAVF: Rosso Ammonitico Veronese F.
Major Elements (Ox%) Ox% SGS CS NS YSGS GSGS RAVV RAVF SiO2 0.14±0.04 0.26±0.08 5.42±0.62 5.40±0.87 5.01±1.03 2.13±0.23 1.65±0,34 TiO2 0.01±0.00 0.01±0.00 0.34±0.06 0.37±0.05 0.42±0.24 0.02±0.00 0.03±0,01 Al2O3 0.12±0.03 0.15±0.03 2.13±0.34 2.47±0.23 1.72±0.41 0.45±0.11 0.55±0,05 FeO 0.16±0.02 0.27±0.03 0.62±0.39 0.27±0.46 0.66±0.24 0.23±0.03 0.37±0,05 Fe2O3 0.01±0.01 0.03±0.01 2.9±0.51 3.38±0.35 1.37±0.24 0.02±0.00 0.02±0,00 MnO 0.01±0.00 0.01±0.00 0.02±n.d. 0.02±n.d. 0.01±0.01 0.05±0.01 0.03±0,00 MgO 0.42±0.02 0.52±0.04 0.85±0.17 0.95±0.34 1.02±0.26 0.46±0.09 0.43±0,07 CaO 55.48±0.68 54.79±0.61 47.01±1.10 46.40±1.02 47.82±1.49 53.92±0.45 54.24±0,48 Na2O 0.04±0.01 - 0.03±0.04 0.01±n.d 0.04±0.01 - - K2O 0.01±0.00 0.03±0.01 0.47±0.11 0.43±0.09 0.68±0.22 0.17±0.03 0.15±0.04 P2O5 0.02±0.01 0.03±0.00 0.18±0.06 0.23±0.06 0.07±0.05 0.08±0.01 0.10±0.01 L.O.I. 43.90±0.49 43.99±0.29 40.02±0.24 39.85±0.27 40.57±1.54 42.75±0.24 42.89±0.36 Tot 100.31±0.31 100.09±0.38 99.99±0.59 99.78±0.64 99.37±0.87 100.28±0.31 100.46±0.23 Trace Elements (ppm) Sc <5 <5 <5 <5 <5 <5 <5 V <5 6±2 89±29 119±26 43±7 6±1 10±2 Cr <6 9±4 67±14 80±12 48±6 <6 <6 Co 26±4 11±4 14±13 16±2 13±7 7±2 11±3 Ni <3 7±2 29±13 42±13 18±5 4±1 6±2 Cu <3 25±5 18±6 23±11 9±3 <3 <3 Zn <3 12±3 30±2 32±3 25±3 9±2 9±1 Ga <5 <5 <5 <5 <5 <5 <5 Rb <3 11±3 5±6 11±2 <3 <3 <3 Sr 429±17 486±35 457±43 451±23 436±36 107±21 120±14 Y 3±1 <3 6±3 3±0 7±0 14±2 17±2 Zr 8±2 24±6 30±25 47±7 33±5 10±4 12±5 Nb <3 <3 7±3 6±2 7±1 <3 <3 Ba 34±4 <10 49±15 39±11 49±14 38±3 36±4 La <10 <10 <10 10±2 12±2 <10 25±2 Ce <10 13±1 14±6 18±2 <10 10±3 15±3 Nd 16±1 <10 16±11 27±2 15±1 <10 30±5 Pb <5 10±2 20±8 15±4 5±2 <5 <5 Th 12±4 24±4 26±20 20±8 29±2 46±2 35±6 U 5±1 <3 <3 <3 <3 5±1 7±3
Mineralogy. The composition of bulk samples shows calcite, quartz, K-feldspar, haematite and illite as main phases. The non-carbonate fraction (1-2%) is mainly composed of quartz, K- feldspar, haematite and illite. Geochemistry. The siliciclastic fraction is very low and matches silica and alumina contents. No other significant evidence was noted.
64 3.2 Pore Structure Observations by optical microscopy. In thin-section microscopy with polarised light, San Gottardo Stone shows the largest pores with respect to other varieties. Average pore size is 500 µm; in the other white variety, porosity seems to be less frequent and pores do not exceed 200- 300 µm at most. The other varieties, YSGS, GSGS and NS, show porosity not visible under the light microscope, except for many pores of 100 µm. Therefore, their porosity must be caused by very small pores, either invisible in thin-section microscopy (less than 20 µm) or invisible because of other complications in interpreting thin sections. Detailed estimation of porosity in thin sections is reported in the previous section on petrography.
Mercury intrusion porosimetry. The different porosity characteristics of limestone were partially confirmed by Mercury Intrusion Porosimetry (MIP) studies (Table 2). San Gottardo Stone has the highest total porosity (28%) with respect to other Berici stones, and the relative volumes of the pore diameters show bi-modal distribution, with peaks around 0.2- 0.3 μm and 20-30 μm (Fig. 9). The height of the two maximum values differs. In general, pores of large diameter are more frequent. This material has the greatest number of mesopores. Costozza Stone has a total porosity of 19.35±0.65%, which is lower than that of SGS. The distribution of pore diameters show some similarities, pointing to bi-modal distribution with peaks around 0.2-0.3 μm and 10-15 μm; smaller pores are more frequent (Fig. 10). Nanto Stone has high porosity, with 25.34±0.67%. The distribution in this case is uni-modal with a peak value around 1-1.5 μm, very high with respect to the other percentages (Fig. 11) which are concentrated only in small pore areas (90% of total porosity is constituted of pores in the range 0.02-3 μm). Yellow San Germano Stone has total porosity of 21.44±0.55%. Relative volume vs pore diameters show tri-modal distribution, with two main peaks at about 2-3 and 5-6 μm (the former higher) and a small peak around 0.2-0.3 μm (Fig. 12). Grey San Germano Stone shows a similar value of total porosity with respect to the Yellow variety. However, the pore graph shows bi-modal distribution, with the highest peak around 3-4 μm and a smaller one around 0.2-0.3 μm. These peaks are concentrated in the zone of small pores (Fig. 13). Rosso Ammonitico reveals lowers values of porosity in comparison with the above limestones. The distribution for both these types is uni-modal, with peaks concentrated at minimum values of pore diameters (Figs. 14, 15). This material has the lowest total open porosity of those studied.
65
Fig. 9. Pore diameter ranges of San Gottardo Stone. Bi-modal distribution.
Fig. 10. Pore diameter ranges of Costozza Stone. Bi-modal distribution.
66
Fig. 11. Pore diameter ranges of Nanto Stone. Uni-modal distribution.
Fig. 12. Pore diameter ranges of Yellow San Germano Stone. Bi-modal distribution.
67
Fig. 13. Pore diameter ranges of Grey San Germano Stone. Bi-modal distribution.
Fig. 14. Pore diameter ranges of RAVV. Uni-modal distribution.
68
Fig. 15. Pore diameter ranges of RAVF. Uni-modal distribution.
Open Porosity. Porosity values calculated by MIP are higher than those of open porosity (Table 2). This is logical, as the intrusion of mercury was forced, whereas water absorption was free, and could be related to pore network connectivity. Comparing these results, it is clear that all the limestones have high values of open porosity with respect to total porosity, suggesting good interconnections between pores. The highest values are those of Yellow San Germano Stone, Grey San Germano Stone and Nanto Stone.
Bulk density. Bulk density values (Table 2) are quite similar in the Berici Hills samples, although some differences are noted in relation to their mineralogy and porosity. SGS has the lowest value, like YSGS: these results correlate well with high values of open porosity. In comparison, CS and YSGS, with lower open porosity values, have the highest bulk density values; for GSGS, this result is also correlated with large amounts of non-carbonate minerals, especially ferrous (glauconite). The high compaction of Rosso Ammonitico explains the higher values in comparison with the Vicenza Stones, also due to extensive occurrence of heavy minerals such as haematite.
Table 2. Open Porosity, Bulk Density and Total Porosity results. Stone Open Porosity (%) Bulk Density (g/cm3) Total porosity (%) San Gottardo Stone (SGS) 23.73±0.77 2.04±0.06 28.67±0.84 Costozza Stone (CS) 15.03±0.43 2.27±0.04 19.35±0.63 Nanto Stone (NS) 21.66±0.49 2.09±0.09 25.34±0.67 Yellow San Germano Stone (YSGS) 18.67±0.86 2.04±0.08 21.44±0.55 Grey San Germano Stone (GSGS) 19.60±0.55 2.20±0.08 22.42±0.65 Rosso Ammonitico V (RAVV) 0.33±0.11 2.80±0.02 0.82±0.11 Rosso Ammonitico F (RAVF) 0.36±0.13 2.77±0.04 0.86±0.12
3.3 Water Absorption Properties Water absorption by total immersion. Table 3 shows the results of the water absorption test expressed as Imbibition Capacity (IC%), and Figs. 16, 17, 18 and 19 illustrate resulting absorption curves. Both curves for initial time of absorption (a) and total time (b) are reported, to evaluate absorption at the initial stages of the test and to have an idea of the velocity of initial
69 absorption. As expected from porosity results, the various samples have different water absorption affinity. Vicenza Stone varieties and Rosso Ammonitico show very different hydric behaviour. Berici Hills limestones absorb more water and reach very high absorption values like those of Rosso Ammonitico. Inside the Vicenza Stone group, the San Gottardo variety has the highest values, followed by Nanto Stone, Grey and Yellow San Germano Stone. The lowest absorption is that of Costozza Stone. The maximum water absorbtion value, as recommended by Normal 7/81, is that of San Gottardo Stone, with an IC of 10.43±0.44 %. Almost all water is absorbed in several materials in very early stages of the test. Exceptions are Nanto Stone and RAV.
Table 3. Water absorption results by total immersion. Stone Imbibition Capacity % San Gottardo Stone (SGS) 10.43±0.44 Costozza Stone (CS) 7.64±0.17 Yellow San Germano Stone (YSGS) 9.11±0.35 Grey San Germano Stone (GSGS) 9.34±0.45 Nanto Stone (NS) 10.3±0.39 Rosso AmmoniticoV (RAVV) 0.267±0.021 Rosso Ammonitico F (RAVF) 0.287±0.044
Measurements on samples were prolonged well over the times recommended in Normal 7/81. Water absorption continues for 8 days, indicating that very small pores become saturated slowly. Therefore, the difference between total and integral open porosity is not attribuible only to the class over 200 μm, but probably also to that under 0.01µm. San Gottardo Stone has the highest total (28.67±0.84%) and open porosity (23.73±0.77), and also the highest number of mesopores. Considering that only mesoporosity allow water to be held, whereas microporosity does not allow it to enter, and macroporosity allows it to flow but does not hold it (Barsottelli et al., 1998), SGS absorbs the highest amount of water (Fig. 16b). The lower total absorption of Costozza Stone in comparison with SGS is explained by the lower value of total porosity (about 10% less). For both stones, however, the bi-modal distribution of mesoporosity (Figs. 9, 10) accounts for the flattening in the water absorption curve through total immersion 10 minutes after the beginning of the test (Fig. 16a) and the following increment after 8 days, when the smaller pores fill up. This consideration also applies to Yellow and Grey San Germano Stone (Barsottelli et al., 2001). Comparison of the water absorption curves of CS, YSGS and GSGS, materials with bi-modal pore distribution and similar values of total porosity, the lower value of final absorption of CS with respect to the other two varieties is to be noted. This result matches the lower open porosity value of this white stone with respect to the coloured varieties (Table 2). RAV has the lowest values, matching the porosity values and uni-modal distribution of the pores, which cause the lack of flattening of typical bi-modal distribution. Looking at the water absorption curves (Figs. 16-19), high absorption velocity is evident, so that, after 10 minutes, over 90% of the final amount of water has been absorbed by SGS, CS, YSGS, GSGS. Figs. 16a, 17a, 18a, 19a show an enlargement of the first parts of the curves: for all materials, water absorption is clearly rapid, because the final values are almost reached in the first 10 minutes, particularly in SGS and GSGS. Different behaviour is shown by Nanto Stone and Rosso Ammonitico, which needs more time to reach 90% of the total absorption value: about 3 hours is needed for Nanto Stone and 150 hours for RAV. For the latter, the lowest values are explained by the lower percentage of mesopores and the lowest total porosity; for Nanto Stone, one answer is that, initially, the presence of clay minerals does not allow fast absorption, partly because they retain water in the crystal lattice. However, it must be noted that these the two materials have uni-modal pore distribution.
70
Fig. 16a. Water absorption curves by total immersion test (initial times) of Costozza Stone (CS) and San Gottardo (SGS) Stone.
Fig. 16b. Water absorption curves by total immersion test of Costozza Stone (CS) and San Gottardo (SGS) Stone.
71
Fig. 17a. Water absorption curves by total immersion test (initial times) of Nanto Stone (NS).
Fig. 17b. Water absorption curves by total immersion test of Nanto Stone (NS).
72
Fig. 18a. Water absorption curves by total immersion test (initial times) of Yellow (YSGS) and Grey (GSGS) San Germano Stones.
Fig. 18b. Water absorption curves by total immersion test of Yellow (YSGS) and Grey (GSGS) San Germano Stones.
73
Fig. 19a. Water absorption curves by total immersion test (initial times) of Rosso Ammonitico V (RAVV) and Rosso Ammonitico F (RAVF).
Fig. 19b. Water absorption curves by total immersion test of Rosso Ammonitico V (RAVV) and Rosso Ammonitico F (RAVF).
74 Water absorption by capillarity. The trends of the capillarity water absorption curves (Figs. 20- 23) are similar to those of water absorption through total immersion, due to the large number of micropores with radius <1 µm, which generate the highest capillary forces (Barsottelli et. al., 2001). For each limestone, the cumulative capillarity absorption increases with t1/2, the square root of elapsed time, are shown in Figs. 20-23. San Gottardo Stone has the highest ability to absorb water by capillarity (Table 4). GSGS and NS are the other two stones with the highest total absorption by capillarity, and they are the rocks with the highest values of open porosity. Rosso Ammonitico has a very low capacity for transporting water, although its pore diameters are concentrated in small pores, but the lowest porosity value is the datum that influenced this trend. As previously noted, also for this property, the lowest value of absorption of the Vicenza Stone varieties is shown by Costozza Stone. Vicenza Stones show quick initial suction, especially the more porous SGS, YSGS and GSGS. For all materials, the inflection point in the initial part of the absorption curve is reached between 3 and 4 hours. From this point of view, these stones behave almost like sponges. After the “nick-point”, the pore system continues to fill for a long time until it becomes completely saturated. The reason for these increases, even after the nick-point is reached, is that the pores are further filled with water when air in enclosed bubbles, contained in pores longer than about 1 µm, is dissolved (Fagerlund, 1994). The total amount of absorbed water of Nanto Stone is higher than that of the San Germano Stones, although they have similar open porosity values. This is explained by the uni-modal pore diameter distribution (around 1 µm) which, as previously mentioned, aids capillarity. The asymptotic value of absorption, evaluated according to UNI 10859:2000, is recorded after 8 days. For a more thorough investigation into the behaviour towards water on non-aged samples, the test lasted much longer than that required by the UNI document. The asymptotic value was considered to have been achieved when the difference between two successive values of the quantity of water absorbed was 0.01%. Complete saturation occurs after about 3 months. Comparing the capillarity curves of SGS and CS, the low absorption rate of SGS is probably a consequence of the large number of macropores, which slow down capillary forces; these forces are in inverse relation to pore radius. The capillarity test shows slower absorption for Nanto Stone, although the total amount of absorbed water is similar to that of SGS. This may be due to the presence of clay minerals which hinder absorption at initial times. Rosso Ammonitico has the lowest values and slow rates of absorption. Flow rates decrease with the increase in fine-grained matrix. Pure micrite exhibits the lowest permeability. Flow rates depend strongly on pore distribution: interparticle pores should provide higher permeability than moulds and intraparticle pores (Pray, 1960). Pore geometry, especially the size and shape of the interconnections between adjacent pores, is a major controlling factor. Permeability decreases during diagenesis because of compaction and cementation. Intergranular cement causes a reduction in permeability (Panda and Lake, 1995).
Table 4. Water absorption by capillarity parameters, Qtf: water absorption by capillarity at final time of experiment; CA: Capillary absorption coefficient. 2 2 -1/2 Stone Qtf (mg/cm ) CA (Capillary absorption coefficient, mg/cm s ) San Gottardo Stone (SGS) 481.41±5.23 5.38±1.05 Costozza Stone (CS) 379.36±19.58 6.44±0.29 Nanto Stone (NS) 479.44±11.78 2.61±0.19 Yellow San Germano Stone (YSGS) 425.57±18.30 4.22±0.32 Grey San Germano Stone (GSGS) 441.95±17.89 5.76±0.50 Rosso Ammonitico V (RAVV) 16.34±0.95 0.041±0.009 Rosso Ammonitico F (RAVF) 16.74±2.68 0.049±0.004
75 600
500
) 400 2
300
(mg/cm SGS i
Q 200 CS
100
0 0 200 400 600 800 Time (s1/2)
Fig. 20. Capillarity water absorption curves of Costozza Stone (CS) and San Gottardo (SGS) Stone.
Fig. 21. Capillarity water absorption curve of Nanto Stone (NS).
76
Fig. 22. Capillarity water absorption curves of Yellow (YSGS) and Grey (GSGS) San Germano Stones.
Fig. 23. Capillarity water absorption curves of Rosso Ammonitico V (RAVV) and Rosso Ammonitico F (RAVF).
3.4 Durability Tests Salt crystallisation test results. The salt crystallisation test had considerable effects on sample stability. All samples underwent a marked weight loss, due to granular disintegration and
77 cracking (Table 5), and all showed significant loss of material as the number of ageing cycles increased. Salt weathering was quantified by the percentage of dry weight loss ΔW (%) during the salt crystallisation test and visual observation of damage. The ranking of each sample by weight loss after the test is shown in Table 5. All samples showed weight loss and a decrease in rock strength after the durability test. RAVV and RAVF showed the greatest resistance to salt weathering; SGS, GSGS and NS were the least resistant. The type of decay in each sample varied greatly, depending on rock and petrophysical properties. Many studies have established the dependence of salt weathering of rocks on composition, porosity and hydric behaviour. Thus, in general terms, the durability of these heterogeneous rocks to salt crystallisation mechanisms diminishes in types of stone with high porosity. Except for Rosso Ammonitico, of the Vicenza Stones, YSGS was the most resistant to salt weathering, and showed slight superficial disintegration, mainly in the matrix. These considerations also apply to Nanto and Costozza, with generally good resistance to salt weathering. At the end of this test, the average highest weight loss was observed in San Gottardo Stone, the lithotype which suffered salt crystallisation most severely, leading to the failure of two samples out of three. The weight loss of the only entire samples is about 40%. The other stones most susceptible to salt action were GSGS and NS, with one failure for each; also, CS lost an average of 9%. In general, all Vicenza Stone samples showed a disgregation form of erosion and in some cases cracks and microcracks. These forms developed after only one test cycle and worsened as the number of cycles increased (Figs. 31-35). The behaviour of the various materials suggests that the differences in terms of deterioration depend not only on the salt involved (the same for all materials) but also on pore size (Colston et al., 2001). An increase in sample weight was generally observed during the first few cycles, as a result of salt accumulation (Martınez Hernando and Suarez del Rıo, 1989; Ihalainen and Uusinoka, 1994), followed by a period during which weight loss occurred (Figs. 24-30).
Table 5. Weight loss (ΔW% with respect to initial weight) during crystallisation tests. Weight loss vs. original mass (%) Cycle 1 2 3 4 5 6 SGS 1.09 -2.30 -13.66 -24.12 -30.67 -39.16 CS 1.15±0.35 0.90±0.25 0.25±0.76 -3.76±0.60 -6.73±0.75 -9.22±1.65 YSGS 1.57±0.07 1.11±0.22 -2.12±0.34 -3.13±0.44 -3.57±0.65 -4.87±1.00 GSGS 1.08±0.06 0.58±0.12 -2.43±0.09 -5.57±3.81 -10.75±7.55 -14.72±9.59 NS 1.29±0.14 0.93±0.22 -1.56±1.38 -3.61±1.10 -5.70±1.18 -9.61±0.91 RAVV 0.012±0.006 -0.007±0.004 -0.002±0.004 -0.037±0.059 -0.089±0.149 -0.170±0.294 RAVF 0.012±0.003 -0.022±0.021 -0.017±0.037 -0.065±0.085 -0.243±0.390 -0.380±0.608
Rosso Ammonitico did not show any significant forms of deterioration (Figs. 36, 37). Three more cycles were performed for this stone, and some cracking along discontinuities was observed, in both RAVV and especially RAVF (Figs. 38, 39). Microscopic studies confirmed substantial changes in the thin surface layers, involving the expansion of microscopic fractures parallel to the exposed surface (beginning of spalling) and surface layer discoloration. Microscopy enables study of the texture of the components, the shape and size of the pores and/or fissures, and possible mineralogical-petrographical modifications. After six cycles, epoxy-impregnated thin sections were cut perpendicular to the cubes. Thin sections were studied in transmitted light to evaluate potential disintegration. Microscopic changes had occurred in the uppermost part of the samples. The depth of observed changes below exposed surfaces was not very variable, and depended on the intrinsic physical properties of the materials.
78 The salt used to study damage by salt crystallisation was sodium sulphate (Na2SO4). This mineral is common in salt crystallisation damage, and is frequently used in laboratory tests (Sperling and Cooke, 1985; Goudie, 1993). In room conditions, the sodium sulphate-water system has two stable phases: mirabilite (Na2SO4·10H2O) and thenardite (Na2SO4); and a metastable phase (Na2SO4·7H2O), which has not been identified in nature. Mirabilite is the most stable mineral phase in sodium sulphate brine in room conditions (Gmelin, 1966; Rodriguez Navarro et al., 2000). The stability of these minerals depends mainly on the ionic strength of the solution, temperature, and relative humidity. The strong dependence of mirabilite saturation on temperature allows it to reach high degrees of supersaturation with only small temperature changes, and therefore explains the severe levels of damage produced by mirabilite crystallisation due to solutions in porous materials (Flatt, 2002). Thenardite is stable in solution at temperatures up to 32.4°C, revealing its low level of dependence on temperature (Gmelin, 1966). The origin of the extensive damage that sodium sulphate can cause in durability testing, which involves impregnation and drying cycles, is due to mirabilite precipitation, which occurs during the wetting step. Damage is severe because, when water enters the thenardite-containing material, dissolution of the mineral creates a solution highly supersaturated with respect to mirabilite. The concentration is high enough for mirabilite precipitation to generate stresses in excess of the tensile strength of most types of stone. This process does not rely on thenardite having been obtained by mirabilite dehydration, implying that thenardite recrystallisation does not contribute to damage in such conditions (Tsui et al., 2003). The weathering effect of sodium sulphate in the present experiments is considered to be a result of the transformation of water-free thenardite (Na2SO4) to the hydrated phase mirabilite (Na2SO4·10H2O). Salt hydration is coupled with a volume increase of about 300% (Price and Brimblecombe, 1994). Both the expected hydration pressure and the crystallisation pressure generated are greater than the tensile strength of porous natural stone. In addition, rock fabric may also affect the durability of natural stone, and pore radius distribution is even more important (Snethlage, 1984). The results of salt action showed that the samples of all stone types submitted to the sodium sulphate test showed slight granular disintegration at the surfaces already after the first cycle. Soluble salt crystallisation inside porous materials generates crystallisation and/or hydration pressures likely to exceed the elastic limit of the material, causing its failure. Salt damage is closely correlated to pore size. Crystallisation occurs initially in the larger pores, forming large crystals from solutions supplied by the smaller pores. The maximum pressure is reached when a large crystal grows within a pore with small entries. It cannot penetrate to the surrounding small pores until high saturation is achieved, producing great stress which damages the material (Steiger, 2005). This is probably the cause of severe damage undergone by San Gottardo Stone, which has the highest porosity and large pores. Parameters such as water absorption and microporosity affect stone durability to a large extent (Benavente et al., 2001), especially when the cause of damage is salt crystallisation. However, compared with other rocks, limestones are known to have complex pore geometry, which commonly results in fast, intense alterations of carbonate sediments, which greatly influence pore characteristics (Anselmetti et al., 1998). At the same time, water is one of the most important agents of deterioration, and also facilitates the damaging action of other agents (Amoroso and Fassina, 1983). There is direct correlation between water absorption and salt damage, so that stone with high water absorption can accumulate more thenardite inside pores, leading to even more serious decay. This explains the general deterioration of all the porous Vicenza stones with respect to Rosso Ammonitico varieties. Damage appears to increase abruptly, as soon as enough salt is present to prevent unrestrained precipitation of mirabilite during impregnation. This implies that stresses are exerted over a large fraction of the sample volume - a situation that would lead to different
79 patterns of damage with respect to field conditions, where the presence and amount of smaller flaws requiring higher stresses to be propagated dictate the durability of materials. The ability of this durability test to produce a classification of resistance that is relevant for in situ exposure must be given further consideration. In our test, samples were totally immersed in the solution. In samples with small pores, the flow of the solution through the porous media is blocked by air bubbles which partially occupy the pores (Hammecker and Jeannette, 1994). If the brine does not fill the porous media completely, the efficiency of pressure crystallisation decreases strongly, and the test therefore gives different results from the degradation caused by salt weathering. Only in stones with large pores (mesoporosity and macroporosity) can air bubbles be expelled easily, permitting the solution to fill all pores completely. Different data from macroporous and microporous samples were also observed by Richardson (1991). Hence, a porous stone with higher micro- and mesoporosity will suffer more damage due to crystallisation (Fitzner and Snethlage, 1982). Thus, the differences in weight loss data in our test (Table 5) are due to the type of porosity and pore size distributions of the samples. In samples with the highest number of large pores, such as SGS, the solution enters easily and moves within the porous media. As this percentage decreases with decreased porosity, damage is reduced. So it is possible to classify stone according to its resistance to salt crystallisation: SGS shows the worst behaviour, followed by GSGS, NS, CS and YSGS. RAV varieties show the best behavior, although damage is produced along discontinuities. YSGS and GSGS have similar characteristics in terms of porosity, but the yellow variety is more resistant to salt damage. The grey variety contains organic matter, and this may explain its behaviour. Considering the clast characters of this stone and the susceptibility of these components to salt crystallisation mechanisms closely connected with its strength, large amounts of strong, well-preserved allochems resist deterioration and protrude from the surface after testing, due to grain loss from the weaker whole material. Probably also the grey variety shows similar behaviour if it does not contain easily deteriorated components (organic matter). As a consequence, the durability of these heterogeneous materials, particularly that of San Germano Stone, depends not only on their whole properties, such as porosity and hydric behaviour, but also on the size and spatial distribution of textural elements within the rock. In RAV, discontinuities are associated with dissolution processes and act as preferential paths for solutions. This fabric anisotropy is particularly important in the RAV deterioration pattern, which appears to be mainly located in these discontinuities in different directions. As a result, the rock undergoes progressive detachment and fragmentation along preferred anisotropic surfaces. In other words, salt crystallisation pressure produces salt-induced microfissures and the coalescence and enlargement of pre-existing pores and fissures, both of which contribute to further weakening of rock discontinuities. Fabric greatly influences trends in the salt weathering of limestone rocks, even those of nearly identical mineralogical composition. The combination of porosity and fissure properties explain the durability of these rocks. In addition, in our case, the disruptive effects of salt crystallisation may have been partially amplified by the presence of clay minerals, in stone that contains them (Nanto Stone). McGreevy and Smith (1984) demonstrated that swelling and non-swelling clay minerals enhance salt-related breakdown and that the phenomenon is augmented by the abundance of clay in rocks.
80
Fig. 24. Weight Loss (ΔW%) versus number of cycles for San Gottardo Stone.
Fig. 25. Weight Loss (ΔW %) versus number of cycles for Costozza Stone.
81
Fig. 26. Weight Loss (ΔW %) versus number of cycles for Nanto Stone.
Fig. 27. Weight Loss (ΔW %) versus number of cycles for YSG Stone.
82
Fig. 28. Weight Loss (ΔW %) versus number of cycles for GSG Stone.
Fig. 29. Weight Loss (ΔW %) versus number of cycles for RAVV.
83
Fig. 30. Weight Loss (ΔW %) versus number of cycles for RAVF.
Visual appearance after salt crystallisation tests. Typical examples of stone degradation are shown in Figs. 31-39, highlighting damaged stone parallelepipeds before and after salt tests. Throughout the tests, all samples were placed as in the photographs. In the total immersion test, the whole sample was put into the saline solution at the immersion stage, so that the solution penetrated the samples from all directions. During the drying stage, salt precipitation occurred in all samples. Scaling was unevenly distributed and the most severe decay was seen in San Gottardo Stone, with the failure of two out of three samples. Failure also occurred in one sample of Grey San Germano Stone and one of Nanto Stone.
Fig. 31. Appearance of San Gottardo Stone (SGS) after six salt crystallisation cycles.
84
Fig. 32. Appearance of Costozza Stone (CS) after six salt crystallisation cycles.
Fig. 33. Appearance of Nanto Stone (NS) after six salt crystallisation cycles.
Fig. 34. Appearance of Yellow San Germano Stone Stone (YSGS) after six salt crystallisation cycles.
85
Fig. 35. Appearance of Grey San Germano Stone (GSGS) after six salt crystallisation cycles.
Fig. 36. Appearance of Rosso Ammonitico Veronese V (RAVV) after six salt crystallisation cycles.
Fig. 37. Appearance of Rosso Ammonitico Veronese F (RAVF) after six salt crystallisation cycles.
86
Fig. 38. Appearance of RAVV and RAVF after nine salt crystallisation cycles.
Fig. 39. Visible forms of deterioration after nine cycles on Rosso Ammonitico V (RAVV) (left) and Rosso Ammonitico F (RAVF) (right). Freeze-thaw test. Mechanisms of frost decay. There are currently three main theories explaining frost damage: 1. Volumetric expansion of water on freezing; 2. Crystallisation or “ice lens theory”; 3. Hydraulic pressure theory. Theory (1) attributes frost damage to the fact that, when water freezes to form ice, the change is accompanied by volumetric expansion and, if the ice in a porous material is restricted from growing, then potentially damaging internal pressure arises. Theory (2) was proposed by Tabor (1929, 1930), who demonstrated by laboratory tests that heave in soils was caused by the formation of ice lenses. He showed that they could be made to grow under pressure and that water could be drawn through a porous substrate from a reservoir, to allow continued growth. The resulting expansion was much greater than that which would have resulted from the expansion in volume of the water without additional water introduced as a result of freezing. One explanation for this was proposed by Everett (1961): when a saturated porous material freezes, ice crystals begin to form in the larger pores and water is withdrawn from the smaller pores, and mechanical disruption may accompany the growth of these crystals. Theory (3) relies on the volume expansion due to freezing, but ice is not the damaging agent. Powers (1945) described a mechanism by means of which, as ice grows in a pore, the expansion caused by freezing expels unfrozen water from the pore. If water is expelled into fine pores at a high rate, then the result is considerable resistance to flow, leading to hydraulic pressure in the fluid, which could cause damage to porous materials. A second mechanism is suggested for materials
87 totally saturated with water (Chatterji and Cristensen, 1979): if the water cannot escape (e.g., if the entire outer surface of the stone is sealed with ice), then hydrostatic pressure develops, causing eventual damage. The pressure exerted in such a closed system may be as high as 2100 kg/cm3 (Cooke and Doornkamp, 1974). The difference between these two mechanisms is that the first requires a high rate of freezing but the second does not (Ross et al., 1991). In Europe, buildings are subjected to relatively slow rates of freezing during winter. In a particular field situation, frost damage may be a combination of all three mechanisms, one generally emerging as dominant, depending on environmental conditions. Field observations and laboratory trials indicate that there are a number of factors that increase the amount of frost damage caused to stone. These are: high degrees of saturation; repeated freezing; rapid freezing rates; small mean pore diameters; sample size used for laboratory frost tests (greater damage caused to large specimens at higher freezing rates). It appears that ultimate freezing temperature has little effect on the amount of frost damage (Ingham, 2005). The response of our samples to the freeze–thaw test was different from their response to sodium-sulphate attack. The freeze-thaw test apparently changed the physical properties of all stone types to some degree. The form and severity of these changes varied greatly, some stones exhibiting significant signs of deterioration after only a few cycles and others appearing visually unchanged even after 45 cycles. A comparison of freeze-thaw test performance with known historical performance was not possible, due to the lack of detailed scientific results about the stones examined. The Vicenza Stones did not show any damage visible to the naked eye after 45 test cycles (Figs. 41-45), except for one sample of YSGS, which had damage on one corner (Fig. 48). Rosso Ammonitico F underwent appreciable deterioration (Figs. 49-51). The loss of weight (Table 7) was greatest in GSGS, followed by YSGS and Nanto Stone. Rosso Ammonitico and the White varieties (SGS, CS) lost little weight, and also showed the smallest variation in IC (Table 6). GSGS and NS showed the highest variation values and RAV also revealed significant differences during cycles (Fig. 40). The results for the seven limestones (SGS, CS, YSGS, GSGS, NS, RAVV, RAVF) follow a clear pattern, although some of them showed unforeseen behaviour considering their performance in the field. In particular, several specimens of Rosso Ammonitico which has a long history of successful use and has proved to be durable in the field, were the ones that performed worst in the frost resistance test. Deterioration was rapid and severe, with failure modes that bore little relation to those seen on stone monuments. By contrast, one porous limestone (Nanto Stone), long known to have only moderate to poor frost resistance, exhibited relatively good durability (although no specific tests were carried out, but this hypothesis concerned general severe deterioration in the field). Results clearly show that, from a general point of view, the five stone samples of Berici stones performed well in response to the test. An unexpected observation was the occurrence of significant damage in very compact Rosso Ammonitico, most visible on samples with many original discontinuities (stylolites). The amplitude of these cracks turned out to be exclusively induced by the freeze-thaw cycles: the hypothesis that the cracks were widened by swelling of clay minerals in discontinuities was not confirmed, as these minerals, mainly illite, do not possess great expansion properties. The possible effect of discontinuities on cracking in severe climatic conditions appears to be very important in predicting the durability of these types of facies.
88 It must also be recalled that, during the freeze-thaw test, the samples passed from a temperature of -20±5°C to +20±5°C in 2 h, so the possibility that cracking was induced by thermal shock should be considered. In particular, good understanding of pore properties seems to be the key element in successful prediction of freeze-thaw durability. The combination of total porosity, open porosity and pore diameter data was very informative in the case of porous limestone materials. Mercury porosimetry indicated that stone with a bi-modal distribution of pore sizes and no significant presence of accessory minerals had good resistance to frost (SGS, CS). Instead, two other types with bi-modal distribution showed lower resistance, as testified by a large variation in the IC, especially for GSGS. For San Germano Stones, the presence of more sensitive accessory minerals, and for GSGS, the presence of an organic fraction, were responsible for the lower resistance. However, the uni-modal pore distribution Nanto Stone also results in less resistance than in the white varieties. Large mean pore size allows outward drainage and the escape of water from the frontal advance of the ice line (Amoroso and Fassina, 1983). Thus, in meso- and macroporous stones, free draining is produced, and they are less susceptible to damage. This matches the frost resistance of all the Vicenza Stones, particularly the best behaviour of SGS, with the highest values of open porosity and many mesopores. Pore size distribution does help to explain why these stones are so durable, because ice crystals preferentially form in the larger pores in all but the highest levels of saturation, thus reducing damage caused by the confinement of growing crystals. In addition, the high content of well interconnected macropores allows these limestones to drain freely. Combined with the results of petrographic examinations, pore size distribution data become a powerful investigative tool. The preferential damage direction in RAV caused the formation of macrofissures of varying widths and lengths, associated with dissolution processes, with stylolites and calcitic veins from well-cemented to large dissolution pores. Near the fissures, variations in grain and pore size, porosity and mineralogy often coincided with grain loss or disintegration, intensifying the weakening of rock surfaces. The effects of weathering on RAVF were detachment of portions of rock near the discontinuities, so that pre-existing fissures are particularly important in RAV deterioration. These rock fabric properties are closely related, as already mentioned, to durability.
Table 6. Imbibition Capacity (IC) variations with increasing freeze-thaw cycles. Variations in Imbibition Capacity (IC-mg/cm2) Number of cycles 0 16 32 45 San Gottardo Stone (SGS) 10.28±0.26 10.32±0.18 10.33±0.22 10.34±0.23 Costozza Stone (CS) 7.58±0.18 7.61±0.15 7.64±0.18 7.66±0.21 Yellow San Germano Stone (YSGS) 9.68±0.45 9.76±0.36 9.81±0.29 9.85±0.41 Nanto Stone (NS) 9.71±0.37 9.91±0.26 9.99±0.31 10.00±0.35 Grey San Germano Stone (GSGS) 8.52±0.57 8.75±0.43 9.09±0.48 9.71±0.52 Rosso AmmoniticoV (RAVV) 0.243±0.030 0.245±0.030 0.247±0.032 0.255±0.030 Rosso Ammonitico F (RAVF) 0.281±0.030 0.303±0.030 0.317±0.031 0.325±0.030
Table 7. Weight loss (ΔW%) at various step intervals of freeze-thaw test. Weight Loss (%) Number of cycles 16 32 45 San Gottardo Stone (SGS) 0.02±0.01 0.04±0.01 0.04±0.01 Costozza Stone (CS) 0.03±0.01 0.04±0.01 0.06±0.01 Nanto Stone (NS) 0.09±0.02 0.10±0.01 0.12±0.03 Yellow San Germano Stone (YSGS) 0.02±0.01 0.07±0.02 0.13±0.02 Grey San Germano Stone (GSGS) 0.52±0.13 0.83±0.17 1.38±0.35 Rosso Ammonitico V (RAVV) 0.01±0.01 0.02±0.01 0.03±0.02 Rosso Ammonitico F (RAVF) 0.03±0.01 0.03±0.01 0.05±0.02
89
Fig. 40. Imbibition Capacity (IC) variations versus number of freeze-thaw cycles.
Photographs of samples exposed to frost resistance test. The photographs show only a few damaged stone cubes after the test. During the test all cubes were placed as shown, and were completely submerged. Only RAVF showed significant damage near discontinuities; other varieties did not appear to have changed.
Fig. 41. Appearance of San Gottardo Stone (SGS) samples after freeze-thaw test.
90
Fig. 42. Appearance of Costozza Stone (CS) samples after freeze-thaw test.
Fig. 43. Appearance of Nanto Stone (NS) samples after freeze-thaw test.
Fig. 44. Appearance of Yellow San Germano Stone (YSGS) samples after freeze-thaw test.
91
Fig. 45. Appearance of Grey San Germano Stone (GSGS) samples after freeze-thaw test.
Fig. 46. Appearance of Rosso Ammonitico Veronese V (RAVV) samples after freeze-thaw test.
Fig. 47. Appearance of Rosso Ammonitico Veronese F (RAVF) samples after freeze-thaw test.
92
Fig. 48. Damage to Yellow San Germano Stone (YSGS) after 32 freeze-thaw cycles (lower left).
Fig. 49. Appearance of Rosso Ammonitico F (RAVF-1) before and after freeze-thaw cycles. Note difference in discontinuities.
Fig. 50. Appearance of Rosso Ammonitico F (RAVF-2) before and after freeze-thaw test. Note damage near discontinuity.
93
Fig. 51. Damage in discontinuities of Rosso Ammonitico RAVF (RAVF-2).
4. Conclusions The behaviour of seven stone varieties extensively used in monumental buildings in the Veneto region was studied with analytical methods typical of applied petrography. Experimental data on water absorption, by both total immersion and capillarity, match the data from measurements of open and total porosity, and highlight strong correlations between absorption mechanisms and porosimetric characteristics in each material. The test of water absorption through total immersion confirmed that the amount of water saturating the materials increases as a function of total open porosity and of the percentage of mesopores. Instead, the capillarity absorption test turned out to depend greatly on the degree of connection among the pores and on the presence of capillary pores (dimension about 1 µm). At the same time, other factors such as the presence of clay minerals and percentage of matrix influenced the rate of water absorption, as observed in Nanto Stone, where clay minerals reduced initial absorption rates, and in Rosso Ammonitico. Flow rates decreased with the increase in fine-grained matrix. Pure micrite exhibited the lowest permeability. Flow rates greatly depend on pore distribution: interparticle pores provide better permeability than moulds and intraparticle pores (Pray, 1960). Pore geometry, especially the size and shape of interconnections between adjacent pores, is a major controlling factor. Permeability decreases during diagenesis because of compaction and cementation. The low rate of absorption between SGS and CS may also related to the plentiful matrix in the latter. These characteristics of water absorption are an important factor in deterioration, especially frost and damage due to salts. In particular, the durability of materials is closely linked to their behaviour with respect to water absorption, which in turn depends on material porosity. Pore size also favours greater, faster absorption of water, as demonstrated in the hydric tests. Experimental trials were conducted to examine the resistance of these stones in accelerated tests, and to predict their durability to frost and salt crystallisation. Weight loss showed only a very moderate increase during freeze-thaw experiments, but more significant changes in the salt crystallisation test, considerably affecting sample stability. All samples suffered a marked weight loss, due to grain disintegration and cracking, and all showed significant loss of material as ageing cycles increased in number. The response of samples to freeze-thaw cycles was different from that to sodium sulphate attack. The Vicenza Stones did not show any damage visible to the naked eye after 32 freeze-thaw cycles; only RAV suffered an appreciable loss of material. Our experiments showed that the stones are more sensitive to crystallising salts, which caused sudden failure of all samples after six crystallisation cycles. The freeze/thaw cycles caused no significant damage to the samples, except RAV. The combination of total porosity and water absorption data used here was found to provide an apparently dependable indication of stone performance. It appears that good understanding of
94 pore properties is the key element in successful prediction of durability. The findings of the petrographic part of this project were very informative, both in determining intrinsic rock characteristics and as a means by which to observe changes caused by test action. Petrography is a fundamental tool for geologists, which is currently under-exploited in the field of stone durability research. Petrographic examination allows thorough understanding of the geological processes which form and change rock in nature, and which are also key factors in calculating the durability of building stone. Petrography allows correlations between textural parameters and diagenetic features of stone and its durability. In this sense, stone that has undergone early cementation (isopachous cement) leading to conserved primary porosity in sediments has higher water absorption properties and is more susceptible to salt weathering and less to frost (San Gottardo Stone). Stone with relatively early cementation but rich in matrix has less capacity to absorb water and shows less damage, revealing the fact that interstitial material plays an important role as regards hydric properties. Stone with high contents of matrix which, during diagenesis, underwent high compaction with the formation of pressure-solution structures (stylolites) are more susceptible to salt and frost action along these discontinuities. Clearly, laboratory simulations represent extreme conditions, not occurring normally. This is why the deterioration of RAV in natural conditions does not lead to severe damage, as our tests showed. Study of these stones is important in understanding the processes and factors that damage historic buildings.
95 References Amoroso G.G., Fassina V. (1983): “Stone decay and conservation”. Elsevier (Amsterdam). Anselmetti F.S., Luthi S., Eberli, G.P., (1998): “Quantitative characterization of carbonate porosity by digital image analysis”. AAPG Bulletin, v. 82, 1815-1836. Baccelle L., Bosellini A., (1965): “Diagrammi per la stima visiva della composizione percentuale nelle rocce sedimentarie”. Ann. Univ. Ferrara, Sez. 9 (N.S.) 4, 59-62. Barsottelli M., Cellai G. F., Fratini F. and Manganelli C., (2001): “The hygrometric behaviour of some artificial stone materials used as elements of masonry walls”. Materials and Structures, v. 34, Number 4, 211-216. Barsottelli M., Fratini, F. Giorgetti G. and Manganelli C., (1998): “Microfabric and alteration in the Carrara marble: a preliminary study”. Science and Technology for Cultural Heritage, v. 7 (2), 109-120. Benavente D., Garcia del Cura M.A., Bernabéu A., Ordonez S., (2001): “Quantification of salt weathering in porous stones using an experimental continuous partial immersion method”. Engineering Geology, v. 59, 313-325. Bosellini A, Martinucci M., (1975): “Annegamento delle piattaforme carbonati che”. Ann Univ Ferrara IX 5 10, 181-193. Bosellini A., Winterer E. L., (1975): “Pelagic limestone and radiolarite of the Tethyan Mesozoic: a genetic model”. Geology v. 3, 279-282. Camuffo D. (1998): Microclimate for cultural heritage. Elsevier (Amsterdam). Cattaneo A., De Vecchi G. P., Menegazzo Vitturi L. (1976): “Le pietre tenere dei colli Berici”. Atti e Memorie dell’Accademia Patavina di Scienze, Lettere ed Arti, LXXXVIII, 79-100. Chatterji S., Cristensen P., (1979): “A mechanism of breakdown of limestone nodules in a freeze-thaw environment”. Cement and Concrete Research, 9, 741-746. Clari P, Martire L., (1996): “Interplay of cementation, mechanical compaction, and chemical compaction in nodular limestones of the Rosso Ammonitico Veronese (Middle-Upper Jurassic, Northeastern Italy)”. J Sediment Res 66, 447-458. Colston B.J., Watt D.S., Munro H.L., (2001): “Environmentally-induced stone decay: the cumulative effects of crystallization-hydration cycles on a Lincolnshire oopelsparite limestone”. Journal of Cultal Heritage 2, 297–307. Cooke R.U., Doornkamp J.C., (1974): Geomorphology in Environmental Management. Clarendon Press, Oxford, 281-283. Dunham, R. J., (1962): “Classification of carbonate rocks according to depositional texture”, in Ham, W. E. (Ed.), Classification of Carbonate Rocks: American Association of Petroleum Geologists Memoir 1, 108-121. Durrast H., Siegesmund S., (1999): “Correlation between rock fabrics and physical properties of carbonate reservoir rocks”. International Journal of Earth Sciences 88, 392-408. Embry, A. F., Klovan, J. E., (1971): “A Late Devonian reef tract on North-Eastern Banks Island”, N.W.T: Canadian Petroleum Geology Bulletin, v. 19, 730-781[revision of Dunham classification]. Everett D.M., (1961): “The thermodynamics of frost damage to porous solids”. Trans Faraday Soc 57, 2205-2211. Fagerlund G., (1994): “Frost resistance of high performance concrete-some theoretical considerations”. Durability of High Performance Concrete, Proceedings of RILEM International Workshop (ed. H. Sommer), Vienna, February 12–15, 112-140. Fassina V., (1988): “Environmental pollution in relation to stone decay”. Durability of Building Materials, 5, 317-368.
96 Fassina V., Cherido M., (1985): “The Nanto stone deterioration and restoration of loggia Cornaro in Padova”. Preprints of Vth International Congress on Deterioration and Conservation of stone", Lausanne, 313-324. Fassina V., Stevan A., (1992): “Fogwater composition in Venice in relation to stone decay”, Proc. of the 7th International Symposium on Deterioration and Conservation of Stone, Lisbon 15-18 June 1992, 365-374. Fassina V.; (1999): “Analisi dei prodotti di alterazione delle superfici dei monumenti mediante cromatografia ionica”, in “La cromatografia ionica nell'analisi ambientale”, a cura di S. Cavalli e C. Sarzanini, 201-220. Fitzner B., Snethlage R., (1982): “Ueber Zusammenhange zwischen salzkristallisationsdruck und Porenradienverteilung”. GP Newsletter 3, 13-24. Flatt R.J.(2002): “Salt damage in porous materials: how high supersaturations are generated”, Journal of Crystal Growth 242, 435-454. Gaetani M., (1975): “Jurassic stratigraphy of the Southern Alps: a review”. In Geology of Italy (ed. Squyres, C. H.), v. 1, 377–402 (The Earth Sciences Society of the Libyan Arab Republic, Tripoli). Gmelin L., (1966): “Handbuch der anorganischen chemie”. Aufl. Natrium Erg. Bd. 3, Verlag Chemie, Weinheim. Goudie, A.S. (1993): “Salt weathering simulation using a single immersion technique”, Earth Surface Processes and Landforms, 18, 369-376. Hammecker C., Jeannette D., (1994): “Modelling the capillary imbibition kinetics in sedimentary rocks: role of the petrographical features”. Transport in Porous Media 17, 285- 303. Hudec P., (1998): “Rock properties and physical processes of rapid weathering and deterioration”. In: Moore, D.P. and Hungr, O., (Eds), 1998. Proceedings of 8th International Congress of IAEG vol. 1, Balkema, Rotterdam, 335-341. Ihalainen P.E., Uusinoka R.P.J., (1994): “Comparison of weathering resistance of some building stones based on treatments simulating different external conditions”. In: R. Oliveira et al., (Eds), Proceedings 7th International Congress of IAEG vol. 5, Balkema, Lisboa, 3613- 3618. Ingham J.P., (2005): “Predicting the frost resistance of building stone˝. Quarterly Journal of Engineering Geology and Hydrogeology; v. 38; 387-399. Martínez Hernando A., Suárez del Río L.M., (1989): “Alterabilidad de algunos materiales graníticos de la Comunidad de Castilla y León utilizables como rocas ornamentales”. Simposio sobre Rocas Industriales y Ornamentales, Mieres del Camino (Asturias,Spain), 11 Martire L., (1992): “Sequence stratigraphy and condensed pelagic sediments. An example from the Rosso Ammonitico Veronese, northeastern Italy”. Palaeogeogr. Palaeoclimatol. Palaeoecol. 94, 169-191. Martire L., (1996): “Stratigraphy, facies and synsedimentary tectonics in the Jurassic Rosso Ammonitico Veronese (Altopiano di Asiago, NE Italy)”. Facies 35, 209-236. Massari F., Medizza F., Sedea R., (1976): “L’evoluzione geologica dell’area euganea tra il Giurese superiore e l’Oligocene inferiore”, Mem. Ist. Geol. Min. Univ. Padova, v. 30,174- 197, Padova. McGreevy J.P., Smith B.J., (1984): “The possible role of clay minerals in salt weathering”. Catena 11, 169-175. Meyers W.J., (1980): “Compaction in Mississippian skeletal limestones, Southwestern New Mexico“. Journal of Sedimentary Petrolology, 50/2, 457-474, Tulsa. Mietto P., (1988): “Aspetti geologici dei Monti berici”, in “I Colli Berici-Natura e Civiltà”, Signum Ed., 13-23, Vicenza.
97 Munnecke A., Samtleben C. (1996): “The formation of micritic limestones and the development of limestone-marl alternations in the Silurian of Gotland“. Facies, 34, 159-176. Myrsini Varti-Matarangas, Dionysis Matarangas, (2000): “Microfacies analysis and endogenic decay causes of carbonate building stones at the Asklepieion Epidaurus monuments of Peloponnessos, Greece”. Journal of Cultural Heritage 1, 165-177. Nicholson D.T., (2001): “Pore properties as indicators of breakdown mechanisms in experimentally weathered limestones”. Earth Surface Processes and Landforms 26, 819-838. Nicholson D.T., Nicholson F.H., (2000): “Physical deterioration of sedimentary rocks subjected to experimental freeze-thaw weathering”. Earth Surface Processes and Landforms 25, 1295-1307. NORMAL 7/81 (1981): “Assorbimento d'acqua per immersione totale Capacità di imbibizione”, CNR-ICR, Rome. Panda M.N. and Lake L.W., (1995): “A physical model of cementation and its effects on single-phase permeability”. American Association of Petroleum Geologists Bulletin 79, 431- 443. Panda M.N., Lake L.W., (1995): “A physical model of cementation and its effects on single- phase permeability“. Amer. Ass. Petrol. Geol. Bull., 79, 431-443. Powers T.C., (1945): “A working hypothesis for further studies of frost resistance of concrete”. Journal of the American Concrete Institute, 16, 245-272. Pray L.C., (1960): “Compaction in calcilutites, an abstract“. Bull. Geol. Ass. Amer., 71, p.1946, New York. Price C.A., Brimblecombe P., (1994): “Preventing salt damage in porous materials”. In: A. Roy and P. Smith (Eds), Preventive Conservation, Practice and Theory, International Institute for Conservation of Historic and Artistic Works, London, 90-93. Prikryl, R., (2006): “Assessment of rock geomechanical quality by quantitative rock fabric coefficients: limitations and possible source of misinterpretations”. Engineering Geology 87, 149-162. Richardson B.A., (1991): “The durability of porous stones”. Stone Industries, 26, 22-25. Rodolico F., (1963): “Le pietre delle città d’Italia”, Le Monnier, Firenze. Rodriguez-Navarro C., Doehne E., Sebastian E. (2000): “How does sodium sulfate crystallize? Implications for the decay and testing of building materials", Cement and concrete research, 30, 1527-1534. Ross K.D, Hart D., Butlin R.N., (1991): “Durability tests for natural building stone”. In: Baker J.M., Nixon P.J., Majumdar J., Davies H. (eds). Durability of Building Materials and Components, Proceedings of the Fifth InternationalConference. Chapman and Hall, London, 97-111. Snethlage R., (1984): “Steinkonservierung. Bayerisches Landesamt für Denkmalpflege”. Arbeitshefte 22, p. 203. Sperling C.H.B., Cooke R.U., (1985): “Laboratory simulation of rock weathering by salt crystallisation and hydration processes in hot, arid environments”. Earth Surface Processes and Landforms 10, 541-555. Steiger M., (2005): “Crystal growth in porous materials. I: The crystallisation pressure of large crystals”. J Cryst Growth 282, 455-469 Tabor S., (1929): “Frost heaving”. Journal of Geology, 37,428-461. Tabor S., (1930): “The mechanics of frost heaving”. Journal of Geology, 38, 303-317. Tsui N., Flatt R. J., Scherer G. W., (2003): “Crystallization damage by sodium sulphate”. Journal of Cultural Heritage, 4, 109-115. UNI 10859:2000: “Beni culturali - Materiali lapidei naturali ed artificiali - Determinazione dell'assorbimento d'acqua per capillarità”. Beni Culturali-Normal.
98 Winterer E. L., Bosellini A., (1981): “Subsidence and sedimentation on Jurassic passive continental margin, Southern Alps, Italy”. American Association of Petroleum Geologists, Bulletin 65, 394-421. Ginevra M., Saralli M., Sedea R., (1999): “Il bacino estrattivo dei Colli Berici”. Assessorato alle politiche per l’ambiente. Regione del Veneto, Giunta Regionale. Quaderno n.1. Venezia.
99
100 Chapter 3
In situ evaluation of restoration treatments on Nanto Stone and Yellow San Germano Stone in two important monuments in Padova: Loggia Cornaro and Stele di Minerva.
1. Introduction Increasing environmental pollution in urban areas has been endangering the survival of carbonate stone in monuments and statuary for many decades. Numerous conservation treatments have been applied to protect and consolidate these works of art, before extensive granular disintegration causes loss of surface material and irreversible damage. Surface treatments to protect stone materials from the effect of decay are widespread in the field of conservation of buildings. Various types of protective treatments have been developed over the years, varying from natural materials such as lime washes to polymer coatings. Inorganic materials have frequently been used, due to their chemical relevance to authentic building materials, but many types of sophisticated polymers and protective coatings are now commonly used in many countries, to counter the problem of stonework degradation due to natural weathering or to prevent premature corrosion. Polymers used in restoration are differently applied, according to their use as strengthening agents (consolidants) or water repellents. To increase the cohesion of weathered stone, a consolidant must penetrate inside the porous network; a protective product is usually applied as a surface coating to prevent contact between the substrate and agents of deterioration. The main function of stone-consolidating materials is to re-establish cohesion between particles of deteriorated stone. In addition to consolidation performance, requirements include durability, depth of penetration, effect on stone porosity, effect on moisture transfer, compatibility with stone, and effect on appearance. These are some of the essential requirements that this type of product must fulfil to be considered suitable for the protection of stone: treatments must be liquid-water repellent and permeable to water vapour, have good adhesion to the stone, and cover pores as a transparent film, not altering the colour of the surface to which they are applied (Botteghi et. al, 1992). There is now a wide variety of specific products available, designed to prevent or reduce as much as possible the penetration of liquid water into stone. Although some of these products afford good results, the degree of water repellency provided by others is not satisfactory. In some cases, they can even alter the petro-physical features of the material, leading to its decay or accelerating specific degradation processes. Synthetic polymers have been widely employed to treat stone materials, as both consolidants and water repellents (Amoroso, 2002). Synthetic polymers should guarantee that treated stone does not change over time or that, at least, complete removal of applied products whenever problems arise is possible (Horie, 1987; Amoroso and Camaiti, 1997; Price, 1996). At the beginning of modern conservation, i.e., the second decades of the last century, acrylic and vinyl polymers were the most widely used (Brandi, 1977; Stanley Price et. al., 1997). Later, silicon- based products were introduced as consolidants and water repellents, and their physical and chemical properties have been widely studied to evaluate their conservative effects on stone. Siliceous consolidants have been used to consolidate sandstone and limestone through the formation of silica or insoluble silicates (Amoroso, 2002). When early polymer treatment of stone substrates was carried out, the theory and practice of conservation considered reversibility of polymers as an essential requirement. As the years went by and the behaviour of synthetic products with all their drawbacks became more evident (i.e., yellowing, detachments, etc.), retreatability became the new priority (AA.VV, 2003). Silicone ester and ethyl silicate consolidants have been used for well over a century. A form of
101 ethyl silicate was used in London as early as 1861. Wacker Chemie was the first company to market ethyl silicates in 1972; the company is based in Germany and has been represented in the United States by ProSoCo since the mid-1980s. In 1996, the Eighth International Congress on the Deterioration of Stone focused discussion on the topic of ethyl silicates and how they perform differently on silicate versus carbonate stone. The focus in 1996 reflected recent interest in the architectural conservation field, and conservators continue to examine the failure or incompatibility of ethyl silicate consolidants with calcareous stone. Some of the tests performed in the mid-1990s focused on individual calcite grains to examine the bonding of various silicate- based products to crystals (Correia, 2005). There are many reports on the performance of water repellents and consolidants in laboratory- based situations (Apollonia et. al., 2001; Mosquera et. al., 2002; Alvarez and Fort Gonzalez, 2001; D’Orazio et al., 2001; Cardiano et al., 2002; Favaro et al., 2006, 2007), but conservationists should be aware that results from stone samples treated in the laboratory are not necessarily representative of those obtained on real buildings. In the laboratory, samples can be effectively impregnated, whereas it is seldom possible to achieve the same level of control on real buildings. Laboratory-based applications therefore produce better, more reproducible results than are generally expected in the field. Despite these difficulties, laboratory-based tests can provide useful indications of the potential performance of a treatment in cases where the volume of sample and its situation are a good simulation of the building façade. The literature on this subject reveals an array of previous research and published works, mostly focusing on laboratory-based and destructive site methods. A few standard recommendations exist for non- destructive site assessment of surface-treated structures. However, many of these are based on the use of a single test method, which can only provide a partial and sometimes ambiguous picture of the condition of the treated surface. More recently, comprehensive studies on the chemical deterioration of these synthetic polymers have been performed, with particular emphasis on their behaviour when applied to stone surfaces, especially considering the progressive deterioration of building materials treated with polymers (Favaro et al., 2006, 2007). Today, one of the problems which represents a scientific and technological challenge in this sphere is evaluation of organic polymer treatment carried out in the recent past. Particular attention should be devoted to the side-effects caused by these synthetic products on various materials and their durability. Until now, only a few studies on their in-depth behaviour have been carried out on artistic monuments (Moropoulou et al., 2000; Favaro et al., 2001). To know the conservation state of treated stone, in order to plan a new method, it is essential to identify and characterise the polymer applied previously and the related by-products of deterioration, in order to understand the deterioration pathways undergone during exposure to the environment. On one hand, polymer deterioration in outdoor conditions can modify both chemical composition and physical properties of stone: chemical decay leads primarily to the formation of oxidised species which quite often produce yellowing of treated surfaces. On the other hand, physical changes induce stiffening and brittleness in polymers, often resulting in polymer fissures, detachment from the stone substrate, and worsening of mechanical properties (Melo et al., 1999). A simple but practical suite of wholly non-destructive and destructive complementary tests has already been employed following the results of a previous study on a variety of stone types and concrete mixes by several authors (Sunny et al., 2001; Moropoulou et al., 2000; Favaro et al., 2001). One test comprises water absorption measurements. When possible, cores should be taken, so that more sophisticated laboratory-based tests such as total porosity, pore size distribution and electron microscopy can be carried out. Since definite procedures and methods for evaluating interventions do not yet exist, research aims at applying new techniques and methodologies to evaluate treatments, to prepare efficient
102 monitoring procedures to evaluate restoration of architectural and monumental artworks in all cases. This allows comparisons between different cases and represents a useful tool for planning maintenance operations by authorities overseeing conservation. Another aim is to carry out a series of tests in order to predetermine restoration effectiveness and, above all, tests that certify that the material to which they are applied will not suffer irreparable damage. The final goal is the development of appropriate methods for long-term evaluation of consolidation interventions on site. There has been considerable interest in developing such methods which can enable owners of structures to appreciate when treatment has been properly applied or when the need arises to re-treat an already treated but subsequently weathered surface. In Padova, since 1975, many restorations have been carried out on external stone surfaces with silicon polymers; considering only those of Berici Hills provenance, they involved entire façades and minor architectural elements such as portals, pilasters or colonnades. The Loggia Cornaro is undoubtedly one of the most important monuments restored in this period. The artwork, in Nanto Stone, is considered a symbol of Paduan architecture of the Renaissance. Twenty years after the first interventions and four from the latest ones, the state of conservation and residual efficiency of treatments were evaluated. This chapter reports observations, results and assessments. The treatment of Nanto Stone has been a frequently studied problem in recent years, and laboratory tests have yielded partial results. Another important artwork restored about one decade ago is the Stele di Minerva, made in Yellow San Germano Stone. The long period which has passed since intervention is shown by the significant deterioration patterns on the façade of the monument, great differential decay, white and black patinas, superficial deposits, etc.. For this monument too, the actual action of the products employed in the restoration was evaluated.
2. Synthetic Products The history of synthetic polymers is not well explained. Scientific papers deal with this subject only partially and in fragmented terms, mostly following individual approaches that cannot easily be integrated into intelligible papers that would be of extreme value for the practice of conservation. Most of the available literature is produced in an academic environment which seeks scientific recognition but which does not necessarily match the doubts and needs of the professional conservation field. The absence of a common approach leads to a situation in which most of the scientific literature is not used for the benefit of conservationists. The ideal consolidant is a product able to restore both strength and other physical properties of decayed stone layers to the level of the sound stone that existed before, and to achieve this without any harmful side-effects. Currently, consolidation processes are not entirely mastered and, in practice, we know very little about how to restore decayed stone to its original state. And, to make things more complex, harmful side-effects are very frequent, suggesting that there is still a long road ahead until full mastery of the process is reached. Mastering the process implies dealing with numerous variables, which makes this objective a complex problem. The ethyl silicate is probably most frequently used consolidant. After hydrolysis and condensation, ethyl silicates originate colloidal silica that is deposited inside the porous structure, precipitating an amorphous SiO2 layer which uniformly covers pore walls and capillaries without sealing them (Rodrigues Delgado, 2001). Silica molecules are chemically similar to silicate minerals, and therefore exhibit very good compatibility with stone of silicate- based composition. Conversely, they show no affinity for carbonate minerals, and some authors have shown that calcite may even act as an inhibitor of polymerisation (Danehey et al., 1992; Goins et al., 1996 a, b). In spite of this drawback, ethyl silicates have frequently used been as consolidants for carbonate materials. This apparent contradiction illustrates one of the current problems of stone consolidation: that it is not the optimum consolidation product is a known fact, but it is used in the absence of anything better. Recently, some research groups have
103 attempted to improve the behaviour of ethyl silicates as consolidants for carbonate stone, and two promising research lines are now giving their first results: elastomerisation of ethyl silicate by means of the addition of some silane molecules (Boos et al., 1996) and pre-treatment with an organic reactant, in order to prepare carbonate minerals to adhere to silica molecules (Weiss et al., 2000). Experiments on carbonate rocks have shown that current ethyl silicates induce a very slight increase in strength, but the main positive characteristics are their very high capacity for migration inside stone, slight reduction in water vapour permeability, and the absence of strong interfaces between treated and non-treated zones. These are important arguments to justify the extensive use that has been made of them, and they will be taken as an incentive to continue research to improve their properties. To overcome the disadvantages of these consolidants, silicon-based products were chosen, since they have good compatibility with the stone and chemical stability, owing to the silicon-oxygen- silicon (Si-O-Si) bond, thus avoiding absorption of solar radiation and discoloration. These products also penetrate deeply, due to their low viscosity.
3. Analytical Techniques Treatments were evaluated with a non-destructive technique, together with laboratory methods requiring samples from treated parts.
Non-destructive technique - capillarity water absorption test at low pressure (Normal 44/93). This water absorption test is a simple method for measuring the volume of water absorbed by a material within a specified time, and represents a measure of the susceptibility of the stone to take up water through the exposed surface. The equipment needed is simple. The test was carried out using a pipe-like apparatus, the standard Karsten tube (Fig.1), graded from 0 to 4 ml (each grade representing an increment of 0.5ml), with a diameter of 2.7 cm. The bottom has an opening of 2.4 cm, which makes contact with the test surface. The total height of the column of water applied to the surface, measured from the central point of the flat, circular brim to the top gradation, is 10 cm. The test is conducted by attaching the Karsten tube to the test surface with ordinary modelling clay and pressing it down to ensure adhesion. Water is then added to the upper, open end of the pipe, until the column reaches the “0” gradation mark. The quantity of water absorbed by the material during a specified period of time is read directly from Fig. 1. Karsten tube. the graduated tube. The periods of time appropriate for the test depend on the porosity of the material being measured; generally intervals of 5, 10, 15, 20, 30 and 60 minutes provide the most useful data. This is one of the specific measurements used to evaluate the efficacy of water-repellent treatments on stone. Effective treatment should substantially reduce the permeability of the stone to water, and thus reduce its vulnerability to water-related deterioration. A comparison of test results on treated versus untreated samples provides information about the degree of protection that can be provided. Data give immediate and clearcut indications. 2 Test results are plotted on a graph showing Qi (water absorption per surface unit in mg/cm ).
Selection of test area and sequence of procedures In fieldwork, it is difficult to work to standard conditions because of the effect of weather, problems of access to test areas, and variations in stone surfaces. Consequently, the value of
104 many studies has been spoilt, and it is rarely possible to compare results from different sources. In order to tackle these problems in the present study, the programme and procedures for carrying out site activities were designed to ensure uniformity in test procedures as well as trustworthy results. A method statement has been written for the suite of tests selected for site work. The next section describes the procedures carried out in accordance with the main recommendations contained in the method statement. For the test to be effective, a relatively smooth surface, free of dust, debris and organic growth is required. The exact size and position of each test area was chosen so as to enable the full suite of tests to be performed on the same area and each area, was clearly defined and recorded. At least three sets of readings were taken at each test area. More were taken if it was considered necessary to achieve consistent readings. To prevent water on the surface affecting the subsequent tests, non-wetting tests were carried out first. Tests involving the application of water to the surface of the structure (such as water absorption) were carried out afterwards. The areas of stonework to be tested were selected to ensure that the site equipment could conveniently be used at each position. Typically, four points along the perpendicular, at different heights from the ground, were selected; these measures were repeated on three vertical lines on the façade.
Destructive methods These tests involve the removal of core specimens, and can only be performed when the structure is not highly sensitive and when the owner has given permission. Coring must be undertaken with the greatest care, and any damage done to the structure at the end of the test must be repaired. This precaution was followed in the present study. The cores removed from the structures are 2 cm in diameter and are a maximum of 2 cm thick.
Optical microscopy Observations by U-polarisation microscopy at various magnifications on thin sections were carried out to define textural parameters and obtain detailed petrographic characterisation of the stone. Reflected light microscopy observations were also performed. The petrographic classification adopted is that proposed by Dunham (1962), with modifications by Embry & Klovan (1971).
X-Ray Power Diffraction (XRPD) X-ray power diffraction (XRPD) was performed to determine the mineralogical composition of the bulk samples and insoluble residue on a Philips X’Pert Pro with a θ-θ geometry powder diffractometer, with CuKα (1.54 Å) radiation filtered by nickel in 40KV and 40mA operational conditions. Reflection patterns were compared with standard reflections. The detection limit for trace minerals was 3%.
Scanning Electron Microscopy (SEM) The texture, distribution and penetration depth of the polymer of treated samples were observed on a Camscan MX2500 scanning electron microscope equipped with an energy dispersive micro-analyser (SEM-EDS). Both stone surfaces (SEI images) and sections perpendicular to the surfaces (BEI images) were analysed. All specimens were coated with a graphite film before SEM-EDS investigations.
X-ray Fluorescence (XRF) Chemical analyses were performed to determine element concentrations in bulk samples; measurements were carried out by X-ray fluorescence (XRF) on a WDS Philips PW2400
105 sequential spectrometer, calibrated by means of standard samples.
Fourier Transform Infrared Spectroscopy (FT-IR) The powder FT-IR spectra of the samples were measured in the wave-number region 480-4000 cm-1 by the KBr method with a Thermonicolet spectrometer. A total of 1 mg of ground sample was mixed with 200 mg KBr and then pressed under vacuum to a pellet. The spectra were recorded with 32 scans at a resolution of 2 cm-1.
Mercury Intrusion Porosimetry (MIP) Total porosity and pore size distribution were measured by mercury porosimetry on a Carlo Erba 2000 Porosimeter. Pressure ranges were 1–2000 bar, and porosimetric data for pore sizes between 0.02 μm and 200 μm were obtained. The volume of intruded mercury for each pressure step allowed determination of total open pore volume and pore size (volume) distribution. The range of measured pore radii was given by Washburn’s equation: P = -4 σ cos φ / d where d is pore diameter, σ the surface tension of mercury, φ the mercury wetting angle of the pore surface (φ>90°) and P absolute pressure exerted.
Ion Chromatography Anion concentrations of sulphates, oxalates, nitrates and chlorides were measured by ion chromatography on a Dionex DX100 equipped with an AS4A-SC column.
4. Loggia Cornaro 4.1 Historical Notes The Loggia Cornaro (Fig. 2), built in Nanto Stone at the beginning of the 16th century by the famous architect Giovanni Maria Falconetto (Verona 1468 – Padova 1535), is one of the most important historical buildings in the city of Padova.
Fig. 2. Loggia Cornaro. It was built for private theatrical performances by Alvise Cornaro (1484-1566), a patron of the arts, who wished to institute a cultural circle to which intellectuals and friends could congregate.
106 It is a clear transposition of the stage part of the ancient theatre described by Vitruvius, represented in “De Architectura” edited by Fra’ Giocondo in 1511 and in modern editions of Terentius’ comedies. The Loggia Cornaro was planned for plays following humanists’ increased interest in ancient theatre at that time; it is the first work in the Veneto of a frons scenae of Roman age. The Loggia, the first building truly of Renaissance age in the city of Padova, dates back to 1524. The date and architect’s name are carved on the architrave of the central arc, now not very legible: “1524, IO MARIA FALCONETVS ARCHITECTVS VERONENSIS”. The Loggia may have been built in two phases, first the lower part and, a few years later, the upper part (Semenzato, 1987; Litardi, 1998).
4.2 History of Treatment In 1968, when the city of Padova became the owner of the important complex of the Loggia, damage to the monuments was so advanced that immediate restoration was required. Little is known of the condition and treatment history of the Loggia Cornaro until that time. It was restored twice from 1979 to 2003, in several operational stages. In all phases, a polysiloxane resin was used, but the actual type changed according to market availability. In 1983, a multi-phase preservation effort began, involving authorities from the Soprintendenza of the Veneto, the city of Padova, and the International Centre for the Study of the Preservation and Restoration of Cultural Properties (ICCROM). Experimental tests were performed in the laboratory and in situ to choose the most suitable product for restoration; consolidation first involved the lower external part of the artwork, and the product used, a methyl phenyl siloxane (trade name Rhodorsil 11309) yielded good results. In 1989, on the basis of these results, a new management programme was established, partly thanks to new municipal funding. The project focused on the most severely deteriorated parts of the upper part of the monument. In the same programme, to prevent water penetrating into the interior, some structural interventions was made and the roof was repaired. At the same time, the courtyard in front of the monument was well-tended, to allow tourists to visit. In the latest restoration intervention, Rhodorsil RC-90 was used, and work was terminated in 2003. In all phases, the method adopted involved: Pre-consolidation with Rhodorsil 11309 (or RC90) spray applied to surfaces showing exfoliation and disgregation forms of decay, with subsequent application of Japanese paper; Fixing of major fragments and scabs by epoxy resin injections; Brushing superficial deposits with soft brushes and/or local aspiration; Consolidation by slow percolation of Rhodorsil 11309 (or RC90) solutions in toluene, alternating these applications and long treatments with solvent only, to increase the depth of penetration; Removal of excess resin from the surface by solvents with low boiling point; Cleaning off deposits and reducing discoloration by aero-abrasive systems; Stucco work with a mixture of lime, rock dust and sand.
4.3 Sampling About 40 samples from the monument were taken (Fig. 3), following Italian guidelines Normal 3/80. After careful macroscopic observation, several micro-flake samples, representative of all the parts of the monument were taken. Most of the samples belong to the façade which, for evaluation purposes, was the most interesting area. Surface samples were collected from this area: damaged layers were scraped off, fragments of rock were removed, and dust at different depths was taken by drilling. Samples were dried, ground and stored in sealed containers until characterisation by physico-chemical analytical techniques (samples not shown in Fig. 3 were taken from the inner parts of the monument).
107
Fig. 3. Sample locations.
4.4 Preliminary Work Visual observations. Prior to carrying out any tests, details of the whole structure were recorded and documented. Where possible, previously treated and untreated areas were identified and any similarities and differences between them were noted. Any surface features such as forms of deterioration, biological growth, etc. were also ascertained. Macroscopic evidence showed that the most advanced and severe deterioration was in the lower part of the monument, which remains the area in most need of attention for evaluation of treatment. However, this great damage, which compromised sculptural forms, had occurred before restoration, and today, comparing photographic documentation before and after intervention, no recent decay seem to have taken place. If the deterioration of the artwork had been allowed to continue without intervention, not only the finely carved surface but also the overall architectural legibility and structural stability of the entire building would have been lost. The monument faces south. The whole surface is exposed to rain and only a small band under a parapet is protected from direct wetting. There are other small areas protected from precipitation by architectural objects. They show some blackening and black scabs. Disgregation has been observed in the most exposed areas. There is no evidence of rising damp along the entire base of the wall and no chromatic differences. Various microcracks (Figs. 4, 5, 6, 10) are present on the surface and sometimes evolve to scaling of fragments (Figs. 7, 9). Green bio-growth is visible on small portions of the east and west walls and is greatest in the protected areas, due to water saturation. In some areas, thin plaquettes have become detached, leaving a powdery white layer underneath (Figs. 8, 9).
108 10 cm 2 cm
Fig. 4. Microcracks. Fig. 5. Microcracks.
10 cm 10 cm
Fig. 6. Extensive microcracks. Fig. 7. Microcracks and scaling.
1 cm 2 cm
Fig. 8. Scaling and white layer. Fig. 9. Scaling and white layer.
5 cm
Fig. 10. Extensive cracks.
109 4.5 Results 4.5.1 Site test - Water capillarity absorption results The 60-minute water absorption test results of each point of the three different lines along the perpendicular façade show the following trends (Figs. 11-13). Each graph reports results for the points at various heights from the floor for the same vertical line.
0,06
0,05 ) 2 0,04
0,03 LC7
0,02 LC9 Qi (ml/cm Qi LC13 0,01 LC30 0,00 0 10 20 30 40 50 60 70 Time (min.)
Fig.11. Qi (capillarity water absorption per surface unit) versus time.
0,12
0,1 ) 2 0,08
0,06 LC31 LC32 0,04 Qi (ml/cm Qi LC33 0,02 LC34 0 0 10 20 30 40 50 60 70 Time (min.)
Fig. 12. Qi (capillarity water absorption per surface unit) versus time.
The following considerations were used in assessing results: Low water absorption indicates that the surface has been treated; high absorption that the it is probably untreated, that treatment has become ineffective, or simply that moisture is entering the substrate; Rapid absorption of water and water drops is indicative of no ineffectual treatment; slow or absent absorption denotes treatment; A long incubation phase (period during which no water is absorbed by the surface), is indicative of effective treatment. The total absence of an incubation period or a very short one indicates that surface treatment is absent or was ineffectual. 110
0,12
0,1 ) 2 0,08
0,06 LC35
0,04 LC36 Qi (ml/cm Qi LC37 0,02 LC38 0 0 10 20 30 40 50 60 70 Time (min.)
Fig. 13. Qi (capillarity water absorption per surface unit) versus time.
The lack of previous studies of water absorption values for a well-treated Nanto Stone surface do not allow any comparisons to be made with the data obtained, but, considering the absorption value for untreated stone at 30 minutes (about 110 ml/cm2) and the absence of absorption in some areas, it is possible to conclude that: The water absorption values of the surface are still quite small compared with those of the untreated stone, and this difference is imputable to treatment already performed; The total absence of an incubation period (time prior to water being absorbed) in some tested points is evidence that these areas are still efficient in repelling water; The presence of an incubation period (even short) and water absorption values in some of the vertical locations suggest that these areas have become less efficient in repelling water and may need to be re-treated in the near future; Areas evidencing water absorption are mainly located on the lower part of the façade, which is the most severely damaged; In the upper part, probably due to a generally good state of conservation, treatment seems to guarantee a water-repellent surface. The shape of the water absorption curve against time also provides information about the effectiveness of treatment. The observed trend is compatible with the nature of the (consolidating) treatment applied. Test data suggest that the treatment is still effective in places, and most effective in the upper part of the monument. However, water is still able to penetrate into some parts, albeit at a slow rate. The built-up of large amounts of water behind portions of the structure probably indicates that the effectiveness of treatment has deteriorated in these areas and may need attention in the near future. The most informative parameters used in judging performance are water absorption and the incubation period prior to water being absorbed. On the whole, the suite of tests developed is practical. An initial judgement can be made as to whether a surface has been treated or if a previously treated surface has become ineffective. It is important to note, however, that tests provide mostly qualitative results and require careful interpretation.
4.5.2 Minero-Petrographic and Chemical Characterisation of Stone Identification of stone These sections report the results of mineralogical, petrographic and chemical analyses carried 111 out on the yellowish-brown-grey lithotype used in the monument. Although the samples have differing chromatic intensity, they are petrographically classified as grainstone (Fig. 14); the matrix is not very abundant, porosity is high, and cementation is fair. The texture is clastic-organogenic, with grain support. The skeleton is mainly due to benthic and planktonic foraminifera, plates and spines of echinoderms, red algae, and a little rare, micron- sized skeletal debris. The skeletal grains, with sutured grain boundaries, are cemented by sparite and micro-sparite of carbonate crystals and goethite. More evident is the clearcut calcite syntaxial overgrowth cement on plates and spines of echinoderms. The pores are cemented by drusy calcite. Thin sections show some euhedral dolomite rhombi casually disseminated in the rock. The siliciclastic fraction is significant and made up of grains of biotite, glauconite, phosphates, pyrite, quartz, feldspars and clay minerals (mainly smectite), as partially confirmed by XRD analysis. The samples are mainly composed of calcite (ascertained by XRD; XRF shows average CaO values of 40.52%, Table 2) but proto-dolomite, gypsum and traces of quartz and K-feldspar were also detected (Table 1). The XRD of the insoluble residue (17-19%) shows K-feldspar, goethite, smectite, illite, or mixed-strata smectite-illite, quartz and traces of probable chlorite.
Table 1. Mineralogical composition of samples from façade and lower inner part of Loggia Cornaro. Sample Mineralogical Composition LC2 Calcite, Gypsum, Quartz, traces of Feldspar LC4 Calcite; Quartz, traces of Feldspars LC7 Calcite; Quartz, traces of Feldspars LC8 Calcite, Quartz LC9 Calcite, Gypsum, Proto-dolomite, traces of Quartz LC11 Calcite, Gypsum, traces of Quartz LC12 Calcite, Gypsum, Proto-dolomite, traces of Quartz LC12b Quartz, Calcite, Hatrurite, traces of Talc LC13 Calcite, Quartz, Proto-dolomite, Gypsum, traces of Feldspar LC17 Calcite, Gypsum, Proto-dolomite, traces of Quartz LC18 Calcite, Gypsum, traces of Quartz LC1 residue K-feldspar, goethite, smectite, illite, or mixed-strata smectite-illite, quartz LC13 residue K-feldspar, goethite, smectite, illite, or mixed-strata smectite-illite, quartz
Evidence of diagenesis Diagenesis, creating situations of differing cementation and compaction, has contributed to modifying textural features. All samples are more compacted, as testified by penetrated and sutured contacts among grains. These lithotypes do not seem to have suffered earlier cementation opposing compaction. Isochemical diagenetic phase. In all samples, when present, early cement is calcitic. Allochemical diagenetic phase. Dissolution of early cement is followed by local precipitation of clear euhedral dolomite rhombus cement in the rock, and occasional changes in the chemistry of the local circulating solutions is responsible for silicisation of some bioclasts. This dolomite, probably not stechiometric, was detected by XRD as proto-dolomite.
Definition and provenance Petrographic characterisation, composition and percentages of insoluble residue, and chemical analyses of bulk rock indicate that the monument is made of Nanto Stone. The generally homogeneous composition indicates a shallow-water deposit, subject to temporary energy and/or chemical changes. The differences do not involve a change of facies, but oscillations in the quantity of skeleton, are ascribed to weak fluctuations in morphology, energy and chemical composition in a depositional basin. Nanto Stone is a yellowish-brown, marly-arenaceous limestone of Middle Eocene age, 112 outcropping near Vicenza and quarried along the slopes of the south-western sector of the Berici Hills. Often used locally for its aesthetic qualities and easy working, when exposed to an urban environment, this stone undergoes severe decay processes (Cattaneo et al., 1976; Fassina and Cherido, 1985). Structural, textural and compositional characteristics suggest that all samples belong to the same lithotype. The rock used to build the Loggia Cornaro may all have come from the same quarry, although the differentiated decay between the lower and upper parts (the two parts seem to have been built at different periods) indicate that the layer of rock was different. In particular, the upper part was probably built of rock from a higher stratigraphic layer, less rich in clay minerals (Figs. 15, 16, 17), due to the lesser influence of the Alpone Chiampo graben, which was active during the depositional period of Nanto Stone (Mietto, 1988) (confirmed by SEM results on thin sections). Fig. 18 shows a section of a sample from the lower part, where the higher quantities of clay minerals are visible, together with the resulting deterioration. Fig. 19 shows a typical sample of the inner part, with fewer clay minerals and lesser effects of decay in depth. These microscopic observations reveal the phenomenon which is considered one of the main causes of the deterioration of Nanto Stone.
Table 2. Chemical composition of samples from Loggia Cornaro (LC) and Nanto Stone quarry (NSQ). Major Elements (%Ox) Ox% NSQ LC8 LC11 LC12A LC14 LC16 LC18 LC20 LC22 LC25 LC28 SiO2 5.42±0.62 11.03 7.34 11.87 15.20 13.51 10.18 12.61 13.36 8.28 11.92 TiO2 0.34±0.06 0.47 0.24 0.49 0.47 0.29 0.26 0.39 0.71 0.49 0.59 Al2O3 2.13±0.34 2.72 3.35 2.83 3.13 2.40 2.03 3.14 3.70 2.42 3.10 FeO 0.62±0.39 n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. n.d. Fe2O3 2.9±0.51 2.82 2.05 3.12 3.29 5.36 2.10 2.77 3.83 3.32 3.37 MnO 0.02±n.d. 0.02 0.02 0.02 0.04 0.03 0.02 0.02 0.02 0.02 0.02 MgO 0.85±0.17 1.14 0.80 0.95 1.42 1.19 0.72 1.05 1.04 1.03 0.99 CaO 47.01±1.10 41.65 44.36 40.36 37.90 37.76 42.63 39.22 38.19 42.78 40.34 Na2O 0.03±0.04 0.07 0.05 0.05 0.14 0.13 0.03 0.08 0.13 0.14 0.08 K2O 0.47±0.11 0.62 0.34 0.66 0.72 0.41 0.34 0.53 0.98 0.82 0.76 P2O5 0.18±0.06 0.20 0.16 0.18 0.21 0.22 0.14 0.56 0.25 0.25 0.22 L.O.I. 40.02±0.24 38.39 39.53 37.67 36.79 36.96 39.86 37.96 37.05 37.07 37.89 Tot 99.99±0.59 99.14 98.24 98.19 99.30 98.27 98.30 98.35 99.27 96.62 99.27 Trace Elements (ppm) S 559±322 5204 10940 15575 7350 16834 17486 16777 6863 35219 8694 Sc <5± <5 <5 <5 <5 <5 <5 <5 <5 <5 <5 V 89±29 86 60 83 85 116 55 86 122 91 108 Cr 67±14 82 46 77 80 85 48 58 103 83 92 Co 14±13 6 3 7 7 3 3 6 13 8 11 Ni 29±13 34 18 37 38 31 20 36 66 33 50 Cu 18±6 45 <3 7 7 4 4 8 6 7 3 Zn 30±2 37 21 28 28 39 32 26 39 44 24 Ga <5 <5 <5 <5 <5 <5 <5 <5 5 <5 <5 Rb 5±6 <3 <3 <3 12 <3 <3 8 10 6 7 Sr 457±43 511 468 504 724 344 422 514 521 534 528 Y 6±3 12 9 11 9 10 8 15 9 13 11 Zr 30±25 46 28 43 47 39 30 39 64 43 56 Nb 7±3 10 6 11 9 8 6 9 16 10 13 Ba 49±15 72 58 55 96 73 56 62 77 102 65 La <10 20 <10 <10 13 13 <10 <10 17 11 <10 Ce 14±6 25 <10 <10 34 20 <10 30 20 <10 12 Nd 16±11 25 <10 12 21 18 29 12 13 21 15 Pb 20±8 6 5 7 10 5 6 10 9 8 8 Th 26±20 12 11 7 31 29 22 27 32 13 <3 U <3 <3 <3 <3 <3 <3 <3 <3 <3 <3 <3
113
500 µm
Fig. 14. Grainstone with planktonic and benthic foraminifera, echinoderms, red algae and serpulids. Nanto Stone. Nicols //.
Fig. 15. BEI image of cross-section of sample LC3. Note layer of clay minerals (upper red line).
114
Fig. 16. BEI image of cross-section of sample LC3. Layer of clay minerals.
Fig. 17. Sample LC3. Micro-analyses of clay minerals.
115
Fig. 18. BEI image of cross-section of sample LC14. Typical appearance of stone of upper part of monument.
Fig. 19. BEI image of cross-section of sample LC23. Typical appearance of stone of lower part of monument, with deep microcracks parallel to external surface.
116
4.5.3 Porosity and Results of Pore Size Distribution Analysis
Pore Diameter range distribution
Quarry LC21
LC14 Relative 30,0 Volume (%) LC9 20,0 LC8 10,0
0,0 LC3
…
…
…
-
-
-
3
-
6
0.3
8
0.5
0.7
0.9
1.5
-
-
-
-
4
0.09
4.1
-
0.15
4.3
0.03
-
4.5
0.05
-
4.7
10
0.08
4.9
20
-
-
40
-
-
7
60
-
9
80
-
-
-
-
1
-
2
-
-
100
5
0.4
0.6
0.8
-
15
0.1
4.2
30
0.2
4.4
50
4.6
70
4.8
90 200 Pore diameter ranges (μm)
Fig. 20. Comparison of pore diameter ranges.
Table 3. Mercury Intrusion Porosimetry data of treated and untreated specimens from Loggia Cornaro. Bulk Density Sample State Position Total Porosity (%) (g/cm3) Lower part, inside sheltered LC3 Untreated 21.05 2.08 area of arcade Lower part, exposed area of LC8 Treated 6.78 2.57 façade Lower part, semi-exposed LC9 Treated 4.95 2.29 area of façade Upper part, exposed area of LC14 Treated 11.34 2.57 façade Lower part, inside sheltered LC21 Untreated 19.73 2.14 area of arcade Nanto Untreated - 25.34±1.52 2.09±0.08 Quarry
Mercury intrusion porosimetry was applied to examine changes in the total porosity and pore distribution ranges of samples belonging to treated surfaces with respect to untreated and quarry samples. Consolidation materials aim to change the micro-structural characteristics of stone and possible modifications are very important, considering the relationship between pore structure and the susceptibility of rock to specific forms of deterioration. Incongruous results due to changes in 117 stone porosity after consolidation have been found in previous studies. In some papers, decreased micro-porosity is reported, with a consequent increase in the percentage of larger pores. In other cases, authors report the relative increase of both smaller and larger pores after consolidation. The total porosity results of treated, untreated and quarry samples (Nanto Stone) are listed in Table 3. Comparisons of pore size distribution of the same samples are shown in Fig. 20. Results reveal that the untreated stone of the monument had total porosity values of approximately 20.5 %, that is, lower than currently quarried stone (25.34±1.52%), also considering that the increased porosity of the Loggia samples is reasonable when related to deterioration phenomena. A substantial decrease in total porosity was observed in treated samples belonging to the façade, especially in sample LC9, compatible with its location in a semi-sheltered position. Sample LC8 has a lower value, although it was sampled in an exposed area. The relatively high value of sample LC14, belonging to the upper part (which seemed well protected by in previous tests) may be due to its more exposed position (explained by the presence of many nearby cracks). Pore distribution range results indicate that, after treatment, the micro-structural characteristics of the stone changed with respect to those of untreated and quarry specimens. A shift towards smaller pore radius was observed in treated samples (more visible in samples LC8 and 9), perhaps due to the formation of a film on the surfaces of coarse pores. In this case, increased micro-porosity in treated samples was observed. Average pore diameters were also decreased, indicating deposition of conservation materials into the pores of the stone. Comparisons of Loggia and quarry sample distributions reveal general bi-modal pore distribution ranges in the former, with the first peak on micro-pore intervals and with the peak between 20 and 40 µm in the latter. Instead, the quarry samples reveal uni-modal distribution, centred on smaller pores. These differences are probably due to deterioration, with an increase in larger pores in that range. The increase in larger pores may also be attributed to the detection range of the porosimeter, concerning pore radius >220 μm which, before treatment, could not be measured, as the radii were too large. However, due to their reduction in radius (owing to the formation of films), they can now be detected. The second hypothesis regarding the porosity observed in Nanto Stone quarry samples at microscopy seems less probable, although the formation of large pores due to decay processes is plausible. Bulk density values showed an increase after treatment, indicating improved cohesion. However, it must be mentioned that great numbers of samples are required for porosimetry measurements, in view of the definite inhomogeneity of porous stone. A in-depth study of this explanation is required, because the suitability of mercury porosimetry for characterising stone micro-structural changes after consolidation is debatable.
4.5.4 Drilling Techniques Drilling tests were carried out for two main reasons: characterisation of decay profiles, and detection and tracing of treatments. The degree of penetration of restored products inside the porous matrix of stone is an important parameter when evaluating the efficacy and durability of treatment. Developing reliable testing methods for determining penetration depth is crucial. One method is tracing penetration profiles by means of micro-drilling techniques, in which pulverised samples taken from treated parts at different depths are then analysed by FT-IR to verify the presence of polymers. At the same time, it is possible to verify the presence of decay products and obtain quantitative data about them, by means of Ion Chromatography analyses. The equipment consists of a hand-drill; rotation speed and advancing rates are kept slow and constant throughout the test. The bit is placed in contact with the surface and drilling is started. Holes, typically 3 mm in diameter, to a depth of 1 cm, are executed in four steps and dust 118 produced by the bit is collected at every 2.5mm of depth.
Table 4. Ion Chromatography results from dust at different depths. Sample Height from g.l. Depth Sulphates Nitrates Chlorides LC 7.1 0 - 2.5 mm 1.55 0.25 0.06 LC 7.2 2.5 - 5 mm 1.34 0.25 0.04 1.20 m LC 7.3 5 - 7.5 mm 0.82 0.30 0.05 LC 7.4 7.5 - 10 mm 0.80 0.33 0.05 LC 10.1 0 - 2.5 mm 1.73 0.01 - LC 10.2 2.5 - 5 mm 1.56 0.02 - 2.5 m LC 10.3 5 - 7.5 mm 1.00 0.02 - LC 10.4 7.5 - 10 mm 0.67 0.01 - LC 15.1 0 - 2.5 mm 1.38 0.03 0.04 LC 15.2 2.5 - 5 mm 1.02 0.03 0.04 3.5 m LC 15.3 5 - 7.5 mm 0.56 0.02 0.02 LC 15.4 7.5 - 10 mm 0.32 0.02 0.03 LC 30.1 0 - 2.5 mm 1.22 0.05 0.03 LC 30.2 2.5 - 5 mm 1.01 0.02 0.03 6 m LC 30.3 5 - 7.5 mm 0.54 0.02 0.03 LC 30.4 7.5 - 10 mm 0.23 0.02 0.03 LC 31.1 0 - 2.5 mm 1.03 0.04 0.06 LC 31.2 2.5 - 5 mm 0.86 0.04 0.04 1.20 m LC 31.3 5 - 7.5 mm 0.23 0.05 0.04 LC 31.4 7.5 - 10 mm 0.13 0.03 0.05 LC 32.1 0 - 2.5 mm 2.10 0.05 0.01 LC 32.2 2.5 - 5 mm 1.57 0.04 0.01 2.5 m LC 32.3 5 - 7.5 mm 0.67 0.05 - LC 32.4 7.5 - 10 mm 0.21 0.04 - LC 33.1 0 - 2.5 mm 1.14 0.03 0.01 LC 33.2 2.5 - 5 mm 0.94 0.03 0.02 3.5 m LC 33.3 5 - 7.5 mm 0.87 0.02 0.01 LC 33.4 7.5 - 10 mm 0.24 0.01 - LC 34.1 0 - 2.5 mm 1.13 0.01 0.02 LC 34.2 2.5 - 5 mm 0.85 0.01 - 6 m LC 34.3 5 - 7.5 mm 0.67 0.01 - LC 34.4 7.5 - 10 mm 0.34 0.01 - LC 35.1 0 - 2.5 mm 1.47 0.06 0.05 LC 35.2 2.5 - 5 mm 1.23 0.07 0.04 1.20 m LC 35.3 5 - 7.5 mm 0,78 0,04 0,06 LC 35.4 7,5 - 10 mm 0,65 0,05 0,04 LC 36.1 0 - 2.5 mm 1.98 0.05 0.03 LC 36.2 2.5 - 5 mm 1.76 0.04 0.02 2.5 m LC 36.3 5 - 7.5 mm 0.93 0.04 0.02 LC 36.4 7.5 - 10 mm 0.78 0.04 0.02 LC 37.1 0 - 2.5 mm 1.25 0.05 0.01 LC 37.2 2.5 - 5 mm 1.03 0.04 - 3.5 m LC 37.3 5 - 7.5 mm 0.79 0.05 - LC 37.4 7.5 - 10 mm 0.74 0.04 - LC 38.1 0 - 2.5 mm 1.19 0.04 0.02 LC 38.2 2.5 - 5 mm 0.89 0.03 0.01 6 m LC 38.3 5 - 7.5 mm 0.76 0.03 0.02 LC 38.4 7.5 - 10 mm 0.55 0.03 0.02
Typically, drill bits are 2-3 mm in diameter, but smaller or larger ones respectively are used in special cases of very hard or very soft materials. Currently, tungsten bits are used, but flat-tipped diamond bits are recommended for their better precision and reliability. Drill holes were made all over the façade at different heights from the ground. Several other 119 holes were drilled in areas where traces of treatment were not so evident, or even suspected. The test provides two types of results: Estimates of the presence of consolidating material at specific depths; Salt concentrations vs depth.
Results of salt concentrations tests are listed in Table 4. FT-IR spectroscopy showed some problems, which will be discussed in the next paragraphs. Due to the triangular shape of the bit, the drilled surface has a conical shape and therefore involves a range of depths and not exactly the depth indicated in the table. Results of the presence of soluble salts on and in the façade revealed sulphates, widespread at all detected points, with a peak at 2.1 %. In general, sulphates on the surface always exceeded 1%. These values cannot be considered negligible, but they are very low when compared with a typical decayed stone (e.g., monuments made of untreated Nanto Stone in the same exposure conditions give values of 14-16% of sulphates). Other salts (nitrates and chlorides) occur in lesser amounts. Clearly, therefore polymers did not stop salt formation but they certainly reduced it. In addition, the “memory” of the stone must be borne in mind, and these decay products were probably due to soluble salts present in the stone before treatment and not properly removed during intervention.
4.5.5 Polymer-Substrate Interactions Information on possible bond formations between consolidation and stone materials was investigated by FT-IR spectroscopy. Data on chemical and/or physico-chemical interactions between porous stone and polymer product remain scarce (Danehey et al., 1992; Elfving and Jäglid, 1992; Kumar, 1995; Spoto et al., 2000; Rizzarelli et al., 2001). The main difficulty in direct determination of polymers in treated stones by the FT-IR technique is often the dominating signal of the inorganic substrate, which masks the weaker signals of organic compounds. In the present study, examination of the typical IR spectrum of pulverised samples shows strong absorption bands of carbonates at 1429, 875 and 712 cm-1. The presence of the polymer is indicated by the band in the range 1100-1000 cm-1, centred on 1029 cm-1, attributed to the stretching Si-O-Si of cycling and open siloxane chains. However, the same band is also common to silica minerals: as other polymer bands are not evident and there were large amounts of clay minerals in the stone, the attribution of this band is not certain. One analytical strategy involves eliminating the calcareous stone matrix by acid attack with HCl solution (8%), followed by analysis of residues by FT-IR spectroscopy. By removing all interfering CaCO3, whose intense absorption bands often mask peaks belonging to polymer products, it is generally possible to highlight the polymer bands. An example of the resulting spectra is shown in Fig. 21, where the strong band at 1100-1000 cm-1 centred on 1006 cm-1 is attributed to the sum of the bands due to the stretching Si-O-Si of cycling and open siloxane chains and to the Si-O-Si bond of silica chains of clay minerals. But, with respect to the previous spectra, other bands of the polymer were identified here. In particular, there is a strong absorption band at 1264 cm-1 (characteristic of methylsiloxane) due to the symmetric stretching -1 of the Si-C bond of the SiCH3 group, so that the weaker absorption at 1131 cm is attributed to -1 stretching ʋSi-OCH3. Strong absorption at 844 cm is attributed to both rocking of the Si-CH3 -1 group and CH3 stretching. The band at 909-912 cm can also be attributed to the ʋ Si-O stretching of the Si-OH groups (Lin-Vien et al., 1991). Bands in the range 699-844 cm-1 are also characteristic of these polymers (comparison with IR spectra of RC90; Fig. 22). No evidence of polymer-substrate interactions was ascertained.
120
%Reflectance 1366.9 95 1734.9 1632.0 2334.1 3361.7 1430.4 1262.7 90 2359.5
85
80
75 844.0 734.8 70 1131.4 801.5 909.7 65 696.3 60
55 525.7 50
45 1006.8 40
35 407.3
30 4000 3000 2000 1500 1000 500 Wavenumbers (cm-1) Fig. 21. IR spectra of insoluble fraction of sample LC8.
%Reflectance 95 1634.3 1392.0 90 2901.6 2357.0 1264.0 2977.7 3389.8 1431.1 85
80
75
70 845.7 65
60 737.9 1129.1 55 791.0 697.5 50
45
40
35
30 1023.7
25 4000 3000 2000 1500 1000 500 Wavenumbers (cm-1) Fig. 22. IR spectra of Rhodorsil RC90.
This procedure was also employed for dust samples obtained by the micro-drilling technique, and the presence of polymer to a depth of 1 cm was ascertained.
121
4.5.6 SEM Observations In order to evaluate the condition of treated surfaces and their state of conservation, samples were collected and examined, and the presence, distribution and morphology of the polymer were detected. All samples revealed two main groups: 1. Samples coated with a wide, uniform, thick layer of resin (confirmed by EDS; Fig. 23) visible in both SEI and BEI images. The latter clearly discriminate between the inorganic substrate (brighter because of the higher average atomic number) with respect to siloxane (grey, due to its low average atomic number). As an example, Figs. 24 and 25 compare BEI and SEI images, respectively, of the same area. As regards the morphology of the polymer surface, several cracks were noted. 2. Samples coated with a wide, uniform, thin layer of resin. BEI images do not show this coating, although EDS analyses do reveal the abundance of Si on all surfaces. Only SEI images confirm the presence of thin coatings, because the carbonate crystals are characterised by a smooth surface appearance. These are due to the thickness of the surface layers (Figs. 26, 27). Considering all samples, those of the first group are in a minority: about 20% of all samples taken from semi-sheltered areas reveal a thick layer of resin.
Fig. 23. EDS micro-analyses of polymer.
122
Fig. 24. BEI image of sample LC11. Thick resin coating on stone surface.
Fig. 25. SEI image of sample LC11. Morphology of thick resin coating on stone surface. 123
Fig. 26. BEI images of surface with thin polymer layer. Sample LC17.
Fig. 27. SEI images of surface of sample LC 17. Note smooth appearance of carbonate crystals, due to polymer layer. 124
Detailed examination of the polymer surface revealed cracks (Figs. 28, 29) of various length, and small particles. Micro-analyses of these materials showed S and Ca as main elements, clearly related to gypsum (Figs. 30-32).
Fig. 28. BEI image of polymer cracks of sample LC 9. Fig. 29. SEI image of polymer microcracks.
Fig. 30. BEI image of polymer and gypsum Fig. 31. SEI image of polymer and gypsum particles of sample LC 19. particles of sample LC19.
Fig. 32. EDS micro-analyses of particles above polymer.
125
Cross-sections Cross-sections of surface samples were examined by BSE. The deterioration pattern of Nanto Stone was evident in all samples as open cracks and microcracks, usually parallel to the external surface, often very long and deep. The frequent presence nearly of clay minerals confirms that expandable clays are one of the most important causes in decay of this type of stone (Fig. 33). In the examined samples, SEM was capable of discriminating between polymer and inorganic matrix, characterised as previously described by different average atomic weights, and yields an overall picture of the performance of the applied product.
Clay minerals
Fig. 33. BEI image of cross-section of sample LC23. Note typical cracks due to expansion and contraction of clay minerals. All sections revealed an external layer of Si composition, i.e., polysiloxane, ranging from 10 to 100 µm in thickness. However, this covering layer was not observed in all samples and was often not regular in thickness and continuity on the surface (Figs. 34, 35). Evidence of penetration of resin inside the stone was frequent in several samples, although only some cracks were filled. In some samples, the penetration path was visualised by X-ray maps on cross-sections of Ca and Si, as markers of CaCO3 substrate and silicon product, respectively. These images show deposition inside some pores and a higher content of silicic component on the external resin layers (Figs. 36, 36a, 37, 37a, b, c, 38, 38a, b, c). Examples of resin filling deep cracks are shown in Figs. 39a and b.
126
Fig. 34. BEI image of cross-section of sample LC18. Note external resin layer on stone surface. Resin penetrated in a crack under surface.
127
Fig. 35. BEI image of cross-section of sample LC18. Note superficial and internal presence of resin.
Fig. 36. BEI image of mapped area of sample LC18. Fig. 36α. SiKα map of sample LC18. Note superficial Si-rich layer.
128
Fig. 37. BEI image of mapped area of sample LC20 in cross-section.
Fig. 37a. CaKα map of sample LC20.
129
Fig. 37b. SiKα map of sample LC20. Note Si-rich zones due to resin, and a few grains of silica minerals (sub- rounded particles).
Fig. 37c. SKα map of sample LC20.
130
Fig. 38. BEI image of mapped area of sample LC16 in cross-section.
Fig. 38a. Cakα map of sample LC16.
131
Fig. 38b. SiKα map of sample LC16. Note Si-rich areas near cracks.
Fig. 38c. SKα map of sample LC16.
132
Fig. 39a. BEI image of cross-section of sample LC16. Crack in depth filled with resin.
Fig. 39b. BEI image of cross-section of sample LC24. Crack in depth filled with resin.
In some sections the resin coating was observed under gypsum layers (Fig. 40, 40 a, b, c). In some cases both gypsum and resin were present simultaneously in the same areas.
133
Fig. 40. BEI image of mapped area of sample LC24.
Fig. 40a. Cakα map of sample LC17.
134
Fig. 40b. SiKα map of sample LC17. Note Si-rich areas due to resin.
Fig. 40c. SKα map of sample LC17.
135
5. STELE DI MINERVA
5.1 Historical Notes The plan of the Stele di Minerva must be set within new building at the main complex of the University of Padova, made necessary in the early 20th century, as the lecture halls had became inadequate owing to the great numbers of students. Financial support for building the monument resulted from agreements between the Rector Carlo Anti and the Ministry responsible. Some historical documents, which can be consulted in the University archives, allow us to reconstruct the events leading to the design and construction of the monument, particularly the correspondence between Anti (Rector), Gio Ponti (architect responsible for the work) and Paolo Boldrin (sculptor). The contract for the execution of a marble stele was signed on 13 December 1940, and the Rector, in his official capacity, entrusted the sculptor Paolo Boldrin with carving a large stone stele, in memory of the students who had died in the name of Revolution and the Fascist cause. The sculpture was to be placed in the new hall of the main building of the University of Padova. In 1941, the Stele di Minerva was placed in the Sala Eroi (Hall of Heroes), but in 1958 it was moved to the east wall of the new main building, where it still is (historical documentation of University of Padova).
Fig. 41. Stele di Minerva.
5.2 Restoration In 1995, severe damage to all parts of the Stele di Minerva required restoration, to save the work from irreparable deterioration. Describing the details of the restoration is too difficult in this case, because it was not precisely 136 documented. Unfortunately, this is a common problem in conservation science, because only in recent years has the proper documentation and cataloguing of artworks and operations executed on them been started. However, sufficiently detailed information was available to develop this research project. The product extensively used for stone consolidation was Rhodorsil RC70, 70% ethyl silicate, density 0.9 Kg/dm3, a transparent liquid diluted in white spirit (manufactured by Rhone Poulenc). The damaged right hand of the sculpture was re-attached by epoxy resin and pinned. Stucco work, with lime having low salt contents (Lafarge), rock dust and acrylic resin in an aqueous emulsion, was carried out. The product used for protection was Rhodorsil 6.9% in white spirit, which is a water-repellent methylsiloxane-based product.
5.3 Preliminary Work Visual observations Prior to carrying out any tests, details of the whole sculpture were recorded and documented. Where possible, previously treated and untreated areas were identified, and any similarities and differences between them were noted. Surface features such as forms of deterioration, biological growth, etc., were also ascertained. In situ study of alteration damage to the monument included surveying, classification and recording of alteration forms. NORMAL 1/88 terminology was used to describe the macroscopic forms of decay affecting the stone, and their detailed recording yielded a “decay map” (Fig. 56), for the purpose of planning suitable diagnostic sampling to evaluate intervention. Macroscopically, clearcut differential deterioration forms were evident, since limestone tends to form reliefs between dense macroscopic fossil shells and a less dense matrix, due to acid rain, frequent in an urbanised area such as Padova. Macrofossils are shown in Figs. 46, 48 and 52. Dissolution processes and precipitation of carbonate minerals cause frequent white deposits in the form of percolation on the surface of monuments (Figs. 42, 45, 47, 49). This decay pattern was still visible in 1995, as documented by photographs, and comparisons of photographs in the two periods suggest that the phenomenon has continued to cause decay and a progressive loss of ornamentations (Fig. 51, 53a, b, 54, 55). The stone of the monument is in a relatively poor state of conservation. The most serious damage is visible in the upper parts, particularly on the most exposed surfaces, such as the head and shoulders. These portions show particularly frequent phenomena of scaling and fragmentation (Figs. 43, 44). Black crusts are widespread on protected areas (Fig. 51), and disgregation, exfoliation, surface deposits and lacunae are visible on various parts of the surface (Figs. 42, 43, 49, 50). There is no evidence of rising damp along the entire base of the wall and no discoloration. No biological growth was observed.
137
Fig. 42. Disgregation. Fig. 43. Scaling.
Fig. 44. Scaling and disgregation. Fig. 45. Percolation and white deposits.
138
Fig. 46. Differential deterioration. Fig. 47. Percolation and white deposits.
Fig. 48. Differential deterioration. Fig. 49. Lacunae and white deposits.
Fig. 50. Surface deposit. Pigeon guano. Fig. 51. Erosion and black scabs.
139
1 cm
Fig. 52. Sample SM1. Note relief between dense macroscopic fossil shells and less dense matrix.
Fig. 53a. Detail of shield, 1995. Fig. 53b. Detail of shield, 2007.
140
Fig. 54. Bust of Stele di Minerva. 1995.
Fig. 55. Various parts of Stele di Minerva, 2007.
141
Fig. 56. Decay map of Minerva. 142
LEGEND
Differential deterioration
Stucco
Exfoliation
Scaling
Crust
Surface deposit
Lacuna e
Erosion
143
5.4. Sampling The monument was examined following Italian guidelines (Normal 3/80). Surface samples (30 specimens) (Fig. 57) were collected from those areas which showed severe symptoms of stone deterioration: damaged layers were scraped off, fragments of black crusts and rock removed, and care taken to ensure that samples were representative of the whole monument, rather than its single parts. Samples were dried, ground, and stored in sealed containers until characterisation by a combination of physico- chemical analytical techniques.
. Fig. 57. Location of samples.
144
5.5 Results 5.5.1 Site test - Water capillarity absorption results The test was not carried out for this monument, because of a special surface decoration bushhammered by the sculptor. Edges and micro-depressions made the surfaces too irregular for good adhesion of the Karsten tube, even with ordinary clay.
5.5.2 Minero-petrographic and chemical characterisation Identification The sections below report the results of mineralogical, petrographic and chemical analyses carried out on the yellow lithotype used to carve the monument. Macroscopically, all samples, coarse to very coarse grained (average 1-1.5 mm), contain macro- allochems (macroforaminifera, Nummulites, Discocyclinas) from a few millimeters to several centimeters in size.
500 µm
Fig. 58. Sample SM12. Rudstone-grainstone-packstone with benthic foraminifera, red algae and echinoderms. Yellow San Germano Stone. Nicols //. The texture is clastic-organogenic with grain support. The structure is essentially isotropic. Porosity is high (15-20%) and of intraparticle and interparticle type, as it develops within and between fossils. Pore size ranges from a few microns to 1.5 mm (average 0.1 mm). The matrix is frequent (5-10%) and mainly intergranular. Examples of pseudomatrix also occur. The cement, both intragranular and intergranular, is not very abundant, and mainly consists of microcrystalline calcite mosaics but also sparite and microsparite. Calcite syntaxial overgrowth cement is clearcut on echinoderm plates and spines. The stone mainly consists of bioclasts of macroforaminifera (40-50% by comparison charts) such as Discocyclinas and Nummulites, followed of red algae (5-10%) and echinoderms (5%).
145
Intraclasts and/or micritised grains are frequent (10-20%). Detrital quartz grains (3–5%), green grains (glauconite), clay minerals and oxides-hydroxides of iron (goethite-limonite) are other non-carbonatic constituents.
Table 5. Mineralogical composition of surface samples of Stele di Minerva. Sample Mineralogical Composition SM3 Gypsum, Calcite, Quartz SM8 Calcite, Gypsum, Quartz SM9 Calcite, Gypsum, Quartz SM10 Calcite, Gypsum, Quartz, Feldspar SM18 Calcite SM26 Calcite, Gypsum, Quartz, Goethite SM27 Calcite, Gypsum, Quartz, Goethite SM29 residue Goethite, Feldspar
Table 6. Chemical composition of Stele di Minerva and Yellow San Germano Stone Quarry(YSGSQ) samples. Major Elements (Ox%) Ox% YSGSQ SM2 SM12 SM28 SiO2 3.00±1.18 7.81 5.37 4.41 TiO2 0.23±0.09 0.35 0.26 0.21 Al2O3 1.05±0.38 1.52 1.09 0.87 FeO 0.4±0.1 n.d. 0.67 0.55 Fe2O3 0.61±0.45 2.16 1.77 0.97 MnO 0.02±0.02 0.01 0.02 0.01 MgO 0.7±0.17 0.90 0.79 0.79 CaO 51.22±1.68 46.63 48.44 50.61 Na2O 0.02±0.02 0.03 0.02 0.01 K2O 0.39±0.21 0.61 0.42 0.31 P2O5 0.05±0.02 0.08 0.04 0.04 L.O.I. 42.59±1.55 39.62 39.00 41.55 Tot 100.28±0.05 99.73 97.89 100.32 Trace Elements (ppm) S 557±121 6825 39828 6886 Sc <5 <5 <5 <5 V 34±5 36 30 27 Cr 3810 49 47 20 Co 28±13 6 <3 <3 Ni 134±4 17 15 8 Cu 6±5 105 137 <3 Zn 9±6 37 59 93 Ga <5 <5 <5 <5 Rb 4±4 <3 <3 5 Sr 478±143 589 547 467 Y 5±2 7 7 <3 Zr 25±9 32 24 19 Nb 5±2 7 5 4 Ba 38±14 65 44 36 La 11±1 <10 11 <10 Ce 18±4 14 <10 13 Nd 16±7 13 <10 <10 Pb 7±2 7 8 5 Th 26±7 74 16 24 U <3 2 12 28
Petrographically, the stone is classified as rudstone-grainstone-packstone (Fig. 58). 146
The samples are mainly composed of calcite (XRD; XRF shows average values of 48.56% CaO, Table 6), but gypsum and quartz were also detected (Table 5). The XRD of the insoluble residue (5%) shows K-feldspar, goethite and feldspar. The diffractograms of these insoluble fractions have a low-intensity peak between 10° and 30°, characteristic of an amorphous phase, and due to the presence of an organic polymer. This explains the higher value of insoluble fraction (5%) in comparison with the quarry sample (4%).
Evidence of Diagenesis Diagenesis, creating situations of differing cementation and compaction, has modified textural features. All samples are more compacted, as testified by penetrating and sutured contacts among the grains. These lithotypes do not seem to have suffered earlier cementation opposing compaction. Isochemical diagenetic phase. In all samples, the early cement, when present, is calcitic. Late diagenetic phase. Occasional dissolution of the early cement is followed by local re- precipitation.
Definition and Provenance Petrographic characterisation, composition and percentages of insoluble residue and chemical analyses of the bulk rock indicate that the monument is made of Yellow San Germano Stone. Composition, generally homogeneous, indicates a shallow-water deposit, subject to temporary energy and/or chemical changes. Yellow San Germano Stone (YSGS) is a yellow limestone of Middle Eocene age, outcropping near Vicenza and quarried along the slopes of the central-eastern sector of the Berici Hills. Structural, textural and compositional characteristics suggest that all the samples belong to the same lithotype and came from the same site, located near Pederiva di Grancona, where this stone is still quarried.
5.5.3 Porosity and Results of Pore Size Distribution Mercury intrusion porosimetry was applied to examine changes in total porosity and pore distribution ranges of samples belonging to treated surfaces with respect to quarry samples. Only two samples of the monument were taken for this analysis, due to the historic value of the artwork and its relatively small size. Total porosity results of treated and quarry samples (now Yellow San Germano Stone quarries) are listed in Table 7. Comparison of pore size distributions of the same samples are shown in Fig. 59. Results show that the treated stone of the monument has total porosity values higher than those of quarry samples, due to deterioration phenomena. A decrease in total porosity was not observed in treated samples of the surface, and higher values correspond to samples from completely exposed areas.
Table 7. Mercury intrusion porosimetry data of samples (SM) from Stele di Minerva and San Germano Quarry (YSGSQ). Bulk Density Sample State Position Total Porosity (%) (g/cm3) SM2 Treated Upper part, exposed area 25.7214 2.22575 Medium-high part, semi- SM15 Treated 24.5757 2.56652 exposed area YSGSQ Untreated - 21.44±0.55 2.04±0.08
147
Pore diameter range distribution
Relative Volume (%) 20,0 San Germano Quarry
10,0 SM15
0,0
SM2
3
0.02
0.3
0.04
0.5
-
0.07
0.7
-
0.09
-
0.9
-
0.15
-
1.5
-
-
-
6
4
-
-
4.1
8
-
4.3
-
4.5
-
-
4.7
-
1
4.9
2 -
10
- 7
20
-
9
0.4
40
-
0.6
60
0.8
-
0.1
80
-
0.2
0.03
5
-
0.05
-
0.08
4.2
4.4
4.6
100
4.8
15
30
-
50
70
90 200
Pore diameter ranges (μm)
Fig. 59. Comparison of pore diameter ranges.
Pore distribution ranges results indicated that, after treatment, changes sometimes occurred in micro-structural characteristics in comparison with quarry specimens. A shift towards smaller pore diameters was observed; in particular in sample SM15, the low percentage of 5-20 µm- sized pores and the increase in the range 0.08-1µm with respect to quarry samples was interpreted as a change from of larger to smaller ones due to polymer filling coarse pores.
5.5.4 Deterioration Products A close examination of the type of decay of the monument showed soluble salts as an important cause of the damage. Sulphates, nitrates and chlorides are known to be the most harmful soluble salts for porous stone. Ion Chromatography analyses of stone materials taken at different depths from stone surfaces by microdrilling (Amoroso and Fassina, 1983; Camuffo, 1998; Fassina, 1999) identified sulphates as the most abundant anions, with trends decreasing from the surface inwards. These sulphates are clearly related to the presence of gypsum, a product of recent alteration, linked to sulphation processes, according to the characteristics of the atmosphere in contact with the stone surface (Amoroso and Fassina, 1983; Fassina 1983, 1987, 1988). In our case, IC analyses show gypsum ranging from 13% to 17% on the surface, and falling to 0.5% in depth (Table 8). Differing amounts of nitrates widespread in stone were found by IC. Macroscopically, the basement shows no evidence of water migration through the stone (absence of efflorescences) but significant amounts were probably due to capillary ascent from moist subsoil. Nitrates in the upper part of the monument were probably caused by pigeon droppings (Amoroso & Fassina, 1983). Chlorides were carried by wind from the Adriatic coast (Amoroso and Fassina, 1983). SEM observations (back-scattered electron images) confirmed the presence of large quantities of gypsum on the surfaces, with accumulation in some superficial areas, not in the form of a compact, continuous layer, but isolated in some portions, indicating irregular aggregates of needle-shaped gypsum crystals, lenticular and minute. Three-dimensional images of surface samples, observed by SEI at SEM, show the typical plate morphological features of gypsum crystals (Figs. 60a, b) and also compacted portions of gypsum (Figs. 61a, b). 148
Table 8. Anion concentrations (%) in stone materials taken at various depths, measured by Ion Chromatography. Sample Height Depth Sulphates Nitrates Chlorides SM 24.1 0 - 2.5 mm 17.25 0.63 0.77 SM 24.2 2.5 - 5 mm 17.16 0.56 0.79 1.00 m SM 24.3 5 - 7.5 mm 0.74 0.65 1.05 SM 24.4 7.5 - 10 mm 0.48 0.49 0.80 SM 20.1 0 - 2.5 mm 15.52 0.82 1.12 SM 20.2 2.5 - 5 mm 9.64 0.63 1.17 2.5 m SM 20.3 5 - 7.5 mm 0.73 0.67 1.38 SM 20.4 7.5 - 10 mm 0.68 0.65 1.29 SM 21.1 0 - 2.5 mm 13.67 0.97 1.02 SM 21.2 2.5 - 5 mm 10.45 0.92 0.93 4.5 m SM 21.3 5 - 7.5 mm 0.58 0.79 0.95 SM 21.4 7.5 - 10 mm 0.46 0.81 0.85
Fig. 60a. BEI image of lenticular gypsum Fig. 60b. SEI image of lenticular gypsum of sample SM13. of sample SM 13.
Fig. 61a. BEI images of compact gypsum Fig. 61b. SEI image of compact gypsum of sample SM8. of sample SM8.
149
Particulate matter embedded in gypsum surfaces was also visible in the form of: i) Carbonaceous spherical and porous particles, emitted by combustion of heavy oils (Sabbioni, 1995); ii) Smooth spherical particles, mainly composed of iron and other metals, derived from combustion of fossil fuels (Sabbioni, 1995) (Fig.62b); iii) Particles and organic fibres, probably vegetal (Fig.62a). This particulate matter was not very abundant, and only occurred in a few isolated cases.
Fig. 62a. BEI image. Organic particles. Fig. 62b. BEI image. Metal particle.
5.5.5 Study of Polymers Information on bond formations between consolidation and stone materials was investigated by FT-IR spectroscopy. Data on chemical and/or physico-chemical interactions between porous stone and polymer product remain very scarce (Danehey et at., 1992; Elfving and Jaglid, 1992; Kumar, 1995; Spoto et al., 2000; Rizzarelli et.al., 2001). The main difficulty in direct determination of polymers in treated stones by the FT-IR is often the dominating signal of the inorganic substrate, which masks the weaker signals of organic compounds. In the present study, examination of the typical IR spectrum of pulverised samples (Fig. 63a, b) shows strong absorption bands of carbonates at 1415, 872 and 712 cm-1. The presence of the polymer is indicated by the band in the range 1100-1000 cm-1, centred on 1024- 1034 cm-1, attributed to the stretching Si-O-Si. However, the same band is also common to silica minerals: as other polymer bands are not evident and there were significant amounts of silica minerals (quarts, feldspar and clay minerals) in the stone, the attribution of this band is not certain. One analytical strategy involves eliminating the calcareous stone matrix by acid attack with HCl solution (8%), followed by analysis of residues by FT-IR spectroscopy. By removing all interfering CaCO3, whose intense absorption bands often mask peaks belonging to polymer products, it is generally possible to highlight the polymer bands. An example of the resulting spectra is shown in Fig. 64, where the strong band at 1100-1000 cm-1 centred on 1054 cm-1, is attributed to the sum of the bands due to the symmetrical vibration Si-O-Si of polymer and to Si-O-Si bond of silica minerals. But, with respect to the previous spectra, other bands of the consolidant (RC70) were identified here. The vibrations of Si-O-Si bond (symmetrical, anti- symmetrical and bending) are presented like this: the symmetrical vibration in the range of 1050 cm-1, the anti-symmetrical vibration in the range of 790 and the bending vibration in the range of 440. the anti-symmetrycal vibration band at 793 cm-1 (Lin-Vien et. al., 1991). No evidence of polymer-substrate interactions was ascertained. Pulverized samples taken from treated stone parts from different depths were analysed with the aid of FT-IR spectroscopy. Measurements indicate the presence of consolidants materials up to a 150 depth of 2mm.
%Reflectance
100 1794.3
95 797.7
90 661.3 598.6523.4 85 469.6 443 419.4
1090.3 80 1151.7 1034.6 712.1
75
70
65
60 872.1
55 1415.5
50
4000 3000 2000 Wavenumbers (cm1500- 1000 500 Fig. 63a. FT-1)IR spectra of sample SM10.
%Reflectance
100 1792.0 95
90 800.0 662.9 598.9 85 845.7 521.1 80 427.4 468.7 75 1131.4 712.2 70 1042.1
65
60
55
50
45 872.6
40 1415.0 35 4000 3000 2000 1500 1000 500 Wavenumbers (cm-1) Fig. 63b. FT-IR spectra of sample SM17.
151
%Reflectance
100
95 1362.3 1737.1 2334.1 90 2359.8 85 667.4 80
75 793.4 544.0
70 945.3 65
60
55
50
45
40 1054.5 439.2 35
4000 3000 2000 1500 1000 500 Wavenumbers (cm-1)
Fig. 64. FT-IR spectra of insoluble residue of sample SM29.
5.5.6 SEM Observations 3-D surface samples were examined to detect any polymer on surfaces and its morphology. No clear evidence of any polymer was found. In some portions, organic materials with Si as main component were identified, together with morphological features concerning the polymer and the consolidant in particular (Figs. 67a, b). In all samples, however, BEI images revealed a cover layer identified by smooth single crystals (Figs. 65, 66). Filaments on the surface may be due to the protective product, but no confirmation is given.
152
Fig. 65. SEI image of treated surface of sample SM23. Filaments may be due to protective products. Note smooth appearance of carbonate crystals.
Fig. 66. SEI image of treated surface of sample SM13. Note smooth appearance of carbonate crystals. 153
Fig. 67a. BEI image of sample SM17. Consolidant material.
Fig. 67b.. SEI image of sample SM17. Consolidant material.
Fig. 67c. Sample SM17. EDS spectra of consolidant material of Figs. 68a, b. 154
Cross-sections observed by BSE electron revealed the presence in several samples (>40%) of a discontinuous superficial layer, with low average atomic number in comparison with the carbonate substrate, not compacted and characterised by frequent lacunae and voids (Figs. 68, 69). A substance with different structure and a low atomic number is also present at depth, filling some pores and cracks. A typical such structure is composed of sub-rounded cells. The presence of the fill substance was only ascertained to a depth of 800 µm. It seems to be loose from the substrate, but incorporates many carbonate particles. In superficial layers, this substance is often present with gypsum, crystallised on it or inside it. X-ray maps of cross-sections of Ca and Si, markers respectively of CaCO3 substrate and silicon products, show deposition inside pores and more silica on the surface (Figs. 70-73-abcd). Superficial layers with the highest concentrations of Si are the most visible (Figs. 72c, 73c). In depth, pore concentrations are not so evident.
Fig. 68. BEI image of sample SM22. Cross-section of treated surface. Note presence of low atomic number material widespread in rock.
155
Fig. 69. BEI image of cross-section of sample SM15. Note superficial layer of polymer.
Fig. 70a. BEI image of cross-section of sample SM18. Mapped area.
156
Fig. 70b. CaKα map.
Fig. 70c. SiKα map. Si-rich areas due to polymer presence are very scarce. Numerous grains of silica minerals.
157
Fig. 70d. SKα map.
Fig. 71a. BEI image of cross-section of sample SM 19. Mapped area.
158
Fig. 71b. CaKα map.
Fig. 71c. SiKα map. Note presence of Si-rich area in some portions of section.
Fig. 71d. SKα map.
159
Fig. 72a. BEI image cross-section of sample SM16. Mapped area.
Fig. 72b. Cakα map. Fig. 72c. SiKα map. Note Si-rich layer due to polymer coating.
Fig. 72d. SKα map.
160
Fig. 73a. BEI image of cross-section of sample SM14. Mapped area.
Fig. 73b. Cakα map. Fig. 73c. SiK α map. Note Si-rich layers due to polymer coating.
Fig. 73d. SKα map.
161
6. Discussion Evaluation of restoration treatments carried out with synthetic products in the recent past yielded results on the performance of some restoration materials available on the market and correlations between their theoretical and real behaviour. Two main types of products were used for restoration. The first is Rhodorsil RC70 (from Rhodia), ethyl silicate, at a concentration of 70% and a density of 0.9 Kg/dm3, in the form of a transparent liquid, diluted in white spirit and designed for pre- and post-consolidation operations. As it is not water-repellent, it must be used in combination with a repellent product. After consolidation, a final application of methylsiloxane as protective agent was applied to the Stele di Minerva. The second type of restoration product, a methylphenyl-polysiloxane resin, Rhodorsil RC 90, designed for consolidation and protective operations, was applied in the Loggia Cornaro. In the present study, not only the consolidating and protective properties but also the durability of restoration products were investigated. The different periods of restoration seem to be adequate for a complete overview of the performance of these materials. Non-destructive and destructive tests were performed in situ as well as in the laboratory. In the Loggia Cornaro, in situ water capillarity absorption tests (Normal 44/93) showed that moisture absorption uptake differed according to area. Generally, moisture absorption measured on the vertical surfaces of the lower part of the monument was higher than that measured on the upper parts. This suggests that surfaces close to the basement are more susceptible to environmental weathering than upper ones, and, indeed, different deterioration patterns were found. The basement and lower parts have a strong marly component ascribed to significant amounts of clay minerals (see SEM observations of cross-sections). Clay minerals occur near cracks and microcracks, very frequent in samples from areas near the basement and extensively found at depth. The stone of the upper part belongs to a stratigraphic horizon with lower contents of clay minerals, thus explaining the good macroscopic state of conservation and probably lower moisture uptake. Measurements made vertically at various points showed moisture uptake values between 0 and 0.04 ml/cm2. In general, these values are still quite low compared with those of the blank reference limestone (110 ml/cm2). The lower values in most of the vertical locations indicate that, after four years, the treated surfaces are still sufficiently water-repellent. However, water is clearly able to penetrate into the various parts of the building, albeit slowly. The build-up of small amounts of water behind some surface areas probably indicates that the effectiveness of the treatment is deteriorating in these areas, and they may need attention in the near future. The surfaces of Stele di Minerva could not be examined, owing to the fact that the Karsten tube could not be used on its vertical surfaces, even when commercial clay was used as sealing agent. Small specimens from the surface of the two monuments were taken in order to analyse stone behaviour in the laboratory. Surface observations of bulk samples revealed different situations. In the Loggia Cornaro, some areas have widespread resin coating, visible on BEI and SEI images, which present a smooth skin with occasional cracks and microcracks. Instead, other areas show coating layers recognisable by the smooth appearance of carbonate crystals, visible in SEI images. The latter situation also occurs in the Stele di Minerva surfaces, although the layer is less evident. No appreciable traces of resin were observed on this monument, except in some isolated cases. The depth of penetration of the resin into the stone and its internal distribution were studied by SEM. Cross-sections of treated specimens proved to be particularly good in detecting polymers. The technique was capable of discriminating polymer distribution from the inorganic matrix. Cross-sections are very useful in observing polymer and substrate relationships. SEM images prove that, in the Stele di Minerva, ethyl silicate precipitates as an amorphous SiO2 layer, which covers some portions of the stone surface where the presence of the polymer is
162 more evident and concentrated; in depth, pore and capillary walls were partially coated by amorphous silica, which does not always seal them. In the Loggia Cornaro, the substrate is covered by a very compact, heterogeneous film, from 10 to 100 µm thick. This external layer is formed of the surface coating; the product was also found under the surface layers and inside some cracks at depth. In a few samples, the coating appeared to be absent. The situation for the Stele di Minerva samples was quite different. Here, only some samples had external coatings, due to the protective substance, not adhering well to the substrate and showing pores and voids between coating and substrate. In depth, consolidant material was found in various portions filling pores, and the typical structure was formed of sub-rounded cells. In the Loggia Cornaro, the samples do not exhibit the typical external deposition layer, due to accumulation of atmospheric pollutants, mainly composed of carbonaceous particles with gypsum crystals. In the Stele di Minerva, the extensive presence of gypsum as both lenticular crystals and compact crusts was evident. The former occur in areas where the polymer has probably disappeared and consequently the decay process is starting again. X-ray silicon distribution maps indicate a silicon-based treatment. Silicon distribution is very different in the two monuments. All samples show that silicon is homogeneously distributed to a depth of 1 mm, but Loggia Cornaro samples have 5 to 10 times higher more silicon than those of the Stele di Minerva. In both monuments, however, many cracks in depth (to 3 mm) are empty. In particular, in the Loggia Cornaro, a series of microcracks parallel to the surface is not filled by the resin, although others at a greater depth are filled. The depth of penetration of newly formed decay products was also studied by means of a thin drilling core micro-sampling system. Various locations were chosen and every 2.5 mm to a depth of 10 mm were taken. Sulphates range from a maximum of 17% in the first 2.5 mm to a minimum of 0.46% at 10 mm in the Stele di Minerva; in the Loggia Cornaro, the range is from about 2% to 0.2% at depth, probably due to incomplete cleaning treatment; in the Stele di Minerva they are ascribed to decay phenomena. There is evidence of decay ascribable to atmospheric pollutants. Since the main difficulty in direct determination of polymers in treated stone is often the dominating signal of the inorganic substrate, which masks the weaker signals of organic compounds, one analytical strategy involves eliminating the calcareous stone matrix by acid attack with HCl solution, followed by analysis of residues by FT-IR spectroscopy. All CaCO3, the intense absorption bands of which often mask peaks belonging to the polymer product can be removed. FT-IR analyses on all superficial and in depth samples of both artworks show absorbance spectra of various contemporaneously present materials. Spectra obtained after acid treatment revealed patterns quite similar to those of commercial methylpolysiloxane products for the Loggia Cornaro and ethyl silicate products in the Stele di Minerva. Peaks of carbonates and gypsum were also found in the spectra of bulk samples. The former were obviously due to stone material; the latter is a product of stone decay. No evidence of polymer-substrate interactions was ascertained. A shift in pore size distribution towards smaller pore radii, as well as percentage reduction of large pores, was observed in treated stone samples, and was attributed to the deposition mechanism of the consolidant on pore wall surfaces (formation of films). Results indicate that, after treatment, the microstructural characteristics of all types of specimens were detected. Average pore diameters were decreased, indicating deposition of conservation materials into the pores of the stone. For some authors, reduction of fine pores is desirable. Water in fine pores tends to be absorbed and retained for long periods, and therefore condensation is more likely to occur in them (Camuffo, 1983). 163
The deposition mechanism of a consolidant leads to more pronounced changes inmicrostructural characteristics, and this was especially true for Loggia Cornaro samples. XRF results showed that all samples were characterised by considerable amounts of Si, which was higher than in quarry samples. This difference is partly explained by the large amounts of clay minerals in Loggia Cornaro samples with respect to quarry samples, but it is improbable that all these anomalies were due to the presence of clay, because other elements typical of clay minerals would also have shown marked differences. Alumina and magnesium showed insignificant increases. Other elements varied, but not to any great extent; increases in the concentrations of some trace elements were due to pollution. Considering the highest values of SiO2 concentrations in all samples of both monuments in comparison with quarry materials, doubt arose as to whether XRF had revealed the Si of the polymer. Some experimental tests were carried out to ascertain this possibility. In order to prepare tampons with a similar composition of that of the treated stone, RC90 dust and a pure carbonate (99.9%) were mixed in appropriate proportions. The tampons were used in this case because of the inflammability and instability of the polymer. Test results are shown below.
Polymer RC90 dust CaCO3 SiO2 (%) Parts 1 5 20.6 Parts 1 10 9.48
Polymer RC90 dust Nanto Stone SiO2 (%) Parts 0 1 9.79 Parts 1 3 14.47
The experiments were also performed on Nanto Stone. Results clearly showed that X-ray fluorescence can detect the Si of polymers and that this is an important point in developing a method for in situ evaluation of treatments. In fact, after calibration in the laboratory with different proportions of polymer, portable XRF equipment could directly detect the amount of Si present and ascribe it to the presence of resin on the stone surface (obviously, the method was tested in particular on stone containing silica minerals). This technique must be tested by collecting more data on site applications, but its possible development, together with laboratory studies, would lead to improved instrumentation and more precise interpretation of acquired data.
Assessment of treatment effectiveness There are a number of factors involved in assessing the performance of any consolidant or water-repellent treatment. They include changes in:
water absorption and retention characteristics; depth of penetration of treatment; porosity; permeability; colour; thermal expansion and contraction changes. The ideal polymer for use in restoration of stone must slow down decay processes and possibly return it as much as possible to its original state of conservation. In order to achieve this, treated 164 stone should mimic sound stone in as many characteristics as possible. However, some characteristics are more important than others. The most important are strength, porosity, permeability, thermal dilation, and colour. Of all polymers, siloxanes seem to be the most promising, although they may not be suitable for every situation. For modern products, these issues still remain, but the most important are depth of penetration and water repellency. As regards ethyl silicate, the end-product of hydrolysis is silica, which is present as a cementing mineral in many types of sandstone and may simulate the behaviour of natural cement more closely than many other compounds. Since water plays such an important role in the decay of stonework, research into polymers has often focused on their ability to make stone water-repellent. However, this may prevent the harmless effects of natural weathering as well as the more damaging ones, causing the appearance of the stone to change. Water repellents are usually marketed as being vapour- permeable, suggesting that moisture will be able to escape to the surface and evaporate. Unfortunately, water uptake may continue as vapour on the surface, as water escapes through as rising damp or by transfer from surrounding stonework, often entering the structure faster than it can escape. Water repellents can reduce the amount of moisture in stone, but they cannot guarantee to prevent moisture completely. For the consolidant treatments tested here, after four years the Loggia Cornaro has good water-repellent behaviour, so in this case the polymer seems to be suitable. However, after 12 years, this characteristic has been lost and water can enter the stone. Clearly, the life-span of synthetic polymers is limited, like their action. Another important factor in evaluation of treatment is depth of penetration. There is common agreement that treatment confined to outer surfaces is dangerous, since it can accelerate spalling and exfoliation. However, there is no agreement on what would be an appropriate depth of treatment, beyond the fact that it is obviously necessary to treat the stone as deeply as possible, to consolidate the whole thickness of the decayed zone. The degree of penetration of polymers inside porous matrixes is an important parameter when evaluating the efficacy and durability of treatments. Developing reliable testing methods for determining penetration depth is crucial. Honeyborne reported that “a common cause of failure of stone preservatives is that, even in porous materials, and under the most favourable conditions, the preservative penetrates only to a relatively small depth and a surface skin is formed which differs in physical properties from the underlying material” (in Ashurst and Dimes, 1990). Similarly, in a detailed review of current research on stone conservation in 1996, Price stated that “little attention has been given to the distribution of products within stone at the microscopic level. Little is known about the bonding, if any, that takes place between treatment and the substrate, and much is left to chemical intuition”. Scaling was not observed in the monuments examined here. In the Loggia Cornaro, the evaluated depth of penetration is over 2 mm, and also in the Stele di Minerva the consolidant RC70 showed good penetration. However, depth of penetration is an important factor, although the most significant is the path of penetration; but only if all cracks in the stone are filled with polymer. This is a crucial problem, because even if some typical commercially available polymer products can penetrate to depth, but cannot fill all spaces, their function may be compromised. Ethyl silicate can penetrate quite deeply into porous limestone, but its consolidating capacity is never very pronounced. In general, for Nanto Stone, the penetration depth may reach more than 2 mm, and for the Stele di Minerva too penetration is very good. However, especially for Nanto Stone, the fact that repetitive empty fractures occur at different depths cannot be considered in relation to type of product, but to deterioration patterns in the stone. This should be considered when deciding the conservation action to be taken. Results clearly show that consolidating these fractures cannot be made directly from the surface, probably because the cracks do not communicate with each other. On the contrary, a special approach - of indirect consolidation, with the creation of specific access to the interior of the stone by micro-drilled holes - will be required. 165
Where a treatment product undergoes a reduction in volume due to solvent evaporation, stresses may be set up in the stone after setting, if the product adheres to the mineral grains before shrinkage is complete. There was no evidence of this behaviour in the monuments examined here. Porosity and pore size distribution are sometimes greatly affected by treatment; at other times they are little affected. This influences resistance to salt crystallisation and freeze/thaw action. Although reduction in pore diameter is good for all decay phenomena, this subject needs to be further studied. According to some authors, if a consolidant increases the proportion of fine pores in a substrate, the result may be increased susceptibility to frost damage; for others, reduction of fine pores is desirable. Water in fine pores tends to be absorbed and retained for long periods, and condensation is more likely to occur in them (Camuffo, 1983). Another problem which emerged from the results is that soluble salts are often present under the surface layer. When they occur in untreated stone, capillary forces generated by water evaporation at the surface cause them to be deposited on or near the surface. In a water-repellent treated stone, the water cannot move out to the surface and must evaporate inside the stone, beyond the limit of penetration of the repellent. This may result in the deposition and concentration of soluble salts within the stone and damage to it through crystal growth pressure or volume changes may then result. Evidence of these phenomena were ascertained for the cases studied here. The problem must be related to the “memory” of the stone, which always contains a certain amount of salts after cleaning. At the same time, where there is a new area of stabilised decayed material, moisture evaporation must take place within the stone, and this may lead to salt crystallisation at the boundary between treated and non-treated stone. However, considering the two interventions, it is clear that, although some side-effects of applied products do occur, they do not seem to be significant with respect to the benefits obtained. Without doubt, treatments do not stop but certainly slow down deterioration. All these considerations and results are very interesting in evaluating the performance of polymers applied to stone artefacts, because there is a clear decrease in their consolidant/waterproofing behaviour. However, above all, there is the quite significant difficulty and or impossibility of removing them from the substrate in the case of new restoration. As planning new treatment must deal with the massive presence of insoluble polymers extensively distributed inside and on the stone surface, evaluation of the amount of insoluble material and identification of deterioration processes are important issues to be taken into account when choosing the most appropriate new restoration treatment.
Conclusions In situ evaluation of restoration interventions was performed on two monuments in the city of Padova, respectively 4 and 12 years after restoration operations. In the present work, this evaluation is estimated by the use of the above-mentioned non-destructive and micro-destructive techniques. Performance evaluation of treatments took place at various periods after restoration and aimed at verifying the effectiveness and development of appropriate methods for evaluating treatments after their execution on monuments in situ. In order to be considered suitable to protect stone, some of the essential requirements that this type of product must fulfil are: treatment must be liquid-water-repellent and permeable to water vapour; it should have good adhesion to stone, and cover pores as a transparent film, without altering the colour of the stone surface on which it is applied. This characterisation of old treated surfaces has also led to some significant conclusions. First of all, we have shown that a series of complementary techniques is capable of revealing the occurrence of these old treatments, and of yielding further information about them. The most immediate informative parameters which may be used in judging performance are: 166 water capillary absorption and polymer distribution inside the stone. It can be seen that, apart from a few local variations, the absorption tests carried out on the Loggia Cornaro yield results which are generally within acceptable ranges. Applying the test results, it is possible to make initial judgements as to whether a surface has been treated or if a previously treated surface has become ineffective. It is most important to note, however, that tests provide mainly qualitative results, which require careful interpretation. The information provided is complementary and may be very useful in making initial decisions regarding the need to apply treatment to a previously untreated surface, assessing the performance of a treated surface, or the need to re- treat an already but subsequently weathered surface. A shift in pore size distribution towards smaller pore radii, as well as the percentage reduction of larger pores, was observed in our treated stone samples, and is attributed to the deposition mechanism of the consolidant on the pore surfaces (formation of films or precipitation as amorphous SiO2). Consolidant penetration depth and new formation decay products show that Si concentration generally decreases from the external surface inwards in both monuments. Si concentrations are very high on the surface and remain quite high to a depth of 300 µm. As regards the products of new decay formation, we conclude that sulphur pollution has partly influenced further gypsum formation in the Loggia Cornaro. In fact, no very important decay was observed, and cross-section analyses showed the widespread presence of residual amounts of still active consolidant and protective coating. The salts currently present were probably insufficiently removed before consolidation. In the Stele di Minerva, decay phenomena were found, with much gypsum on surfaces. Residual traces of treatments were found, but not abundantly. Cross-sections of treated samples proved to be particularly useful in detecting synthetic products applied. Distribution and changes in structure and adherence to the substrate were ascertained with good precision. The possible use of fluorescence analyses in situ was considered: such techniques improve and augment the precise interpretation of acquired data, together with data from laboratory investigations. Polymer application has certainly slowed down but not completed halted the deterioration processes of the stone materials examined here. Decay due to atmospheric pollutants has been greatly reduced, but it has rarely been possible to prevent the penetration of salt solutions or mitigate the effects of moisture and temperature changes, which undoubtedly cause disruption within the pore structure as a consequence of mechanical stresses. These results indicate that some of the treatments applied to the monuments have deteriorated to the extent that it is now timely to consider re-treatment. The results obtained in the case of Loggia Cornaro are still generally within acceptable values. It can be seen that the test methods used are complementary and adequate for a complete overview of consolidant performance. They are quite useful in verifying the actual state of conservation of an artefact and therefore in helping to detect the performance of applied polymers, in order to propose a suitable maintenance programme for both monuments. We suggest that maintenance should be carried out in the near future by renewing the protective coating on the whole surface and by consolidating the stone in some areas which are not in a good state of conservation. In both cases, however, re-treatment may need to be renewed after approximately 6 years. Among conservation activities, consolidation is certainly the most difficult. Conservationists face complex materials with diversified mechanisms of evolution, which made them virtually unique at each practical intervention. They must take into account the nature and state of decay of the stone and its water contents, and they know that good results depend on the type and concentration of consolidant, type of solvent, application technique, and contact time. In such a complex domain, conservationists need the help of the scientific community but, when 167 examined from their viewpoint, most scientific results cannot easily be transferred into practical ideas or solutions, or solve problems which others do not have or which they experience in rather different terms. Better and more frequent dialogue among scientists and restorers would bring great benefits for the practice of conservation.
168
References AAVV., (2003): “La Reversibilita` nel Restauro: Riflessioni, esperienze, percorsi di ricerca”. Atti del XIX Convegno di Studi, Bressanone 1-4 luglio. Alvarez de Buergo Ballester M., Fort Gonzalez R., (2001): “Basic methodology for the assessment and selection of water-repellent treatments applied on carbonatic materials”. Progress in Organic Coatings, 43, 258-266. Amoroso G, Camaiti M., (1997): “Scienza dei materiali e restauro. La pietra dalle mani degli artisti e degli artisti a quelle dei chimici macromolecolari”. Firenze: Alinea. Amoroso GG, Fassina V., (1983): “Stone decay and conservation”. Amsterdam: Elsevier. Amoroso GG., (2002): “Trattato di Scienza della Conservazione dei Monumenti”. Firenze: Alinea Editrice s.r.l.; 1-415. Apollonia L., Borgia G.C., Bortolotti V., Brown R.J.S., Fantazzini P., Rezzaro G., (2001): “Effects of hydrophobic treatments of stone in pore water studied by continuous distribution analysis of NMR relaxation times”. Magnetic Resonance Imaging, 19, 509-512. Ashurst, J., and F. G.Dimes, eds. (1990). Conservation of building and decorative stone. London: Butterworth-Heinemann. Boos M., Grobe J., Hilbert G. and Muller-Rochholz, J., (1996): “Modified elastic silicic- acid ester applied on natural stone and tests of their efficiency”. Proc. 8th Cong. on Deterioration and Conservation of Stone, Berlin, 1179-1185. Botteghi C., Matteoli U., Paganelli S., Arbizzani R., Ciardelli F., Aglietto M., Taburoni E., Fassina V, Sci. Technol. Cult. Heritage 1 (1992), 111. Brandi C., (1977): “Teoria del Restauro”. Torino: Piccola Biblioteca Einaudi. Camuffo D. (1998) – “Microclimate for cultural heritage”. Elsevier (Amsterdam). Camuffo D., (1993): “Condensation-evaporation cycles in pore and capillary systems according to the Kelvin model” .Water Air Soil Pollut., 21, 151-159. Cardiano P., Sergia S., Lazzari M., Piraino P., (2002): “Epoxy–silica polymers as restoration materials”. Polymer 43, 6635–6640. Cattaneo A., De Vecchi G. P. & Menegazzo Vitturi L. (1976): “Le pietre tenere dei colli Berici”. Atti e Memorie dell’Accademia Patavina di Scienze, Lettere ed Arti, LXXXVIII, 79-100. Correia J.M., (2005): “Treatment evaluation for the micritic limestone at mission San Josè San Miguel De Aguayo, San Antonio, Texas”. Presented to the Faculties of the University of Pennsylvania in Partial Fulfillment of the Requirements of the Degree of master of science in historic preservation. D’Orazio L., Gentile G., Mancarella C., Martuscelli E., Massa V., (2001): “ Water- dispersed polymers for the conservation and restoration of Cultural Heritage: a molecular, thermal, structural and mechanical characterization. Polymer Testing 20, 227-240. Danehey C., Wheeler G. S., and Su S. C. H., (1992):. “The influence of quartz and calcite on the polymerization of methyl-trimetoxysilane”. In Proceedings of the Seventh International Congress on the Deterioration and Conservation of Stone, Lisbon, June 15–18, ed. J.Delgado Rodrigues, F.Henriques, and F.Telmo Jeremias. Lisbon: Laboratorio nacional de engenharia civil. 1043–1052. Dunham R. J., (1962): “Classification of carbonate rocks according to depositional texture”, in Ham, W. E. ed., Classification of Carbonate rocks: American Association of Petroleum Geologists Memoir 1, 108-121. Elfving P., and Jäglid U., (1992): “Silane bonding to various mineral surfaces”. Department of Inorganic Chemistry, Göteborg, Sweden. Embry A. F. and Klovan J. E., (1971): “A Late Devonian reef tract on North-Eastern Banks Island”, N.W.T: Canadian Petroleum Geology Bulletin, v.19, 730-781. [revision of Dunham classification]. 169
Fassina V. (1999) – “Analisi dei prodotti di alterazione delle superfici dei monumenti mediante cromatografia ionica”, in “La cromatografia ionica nell'analisi ambientale”, a cura di S. Cavalli e C. Sarzanini, 201-220. Fassina V. Cherido M., (1985): “The Nanto stone deterioration and restoration of Loggia Cornaro in Padova”. Preprints of the Vth Int. Congr. "Deter. and Conserv. of stone", Lausanne, 313-324.
Fassina V., (1983): “L'ossidazione dell’SO2 in atmosfera e la formazione delle croste sulle superfici lapidee”, Atti del Convegno "Le piogge acide", Milano 16 apr. 1983, 116-133, Franco Angeli Editore, Milano. Fassina V., (1988): “Environmental pollution in relation to stone decay, Durability of Building Materials”, 5, 317-368. FassinaV. (1987): “Influenza dell'inquinamento atmosferico sui processi di degrado dei materiali lapidei”, Bollettino d'Arte, Supplemento al n.41 "Materiali lapidei - problemi relativi allo studio del degrado e della conservazione”, 19-47. Favaro M., Mendichi R., Ossola F., Russo U., Simon S., Tomasin P., Vigato P.A., (2006): “Evaluation of polymers for conservation treatments of outdoor exposed stone monuments. Part I: Photo-oxidative weathering”. Polymer Degradation and Stability, 91, 3083-3096 Favaro M., Mendichi R., Ossola F., Simon S., Tomasin P., Vigato P.A., (2007): “Evaluation of polymers for conservation treatments of outdoor exposed stone monuments. Part II: Photo-oxidative and salt-induced weathering of acrylic-silicone mixtures”. Polymer Degradation and Stability , 92, 335-351. Favaro M., Menichelli C., Bassotto L., Fassina V., (2001): “Preliminary results on the behaviour of restoration materials used in the past on monuments in relation to their durability and to decay processes case study: “Porta della Carta” in Venice”. In 3rd International Congress on “Science and Technology for the Safeguard of Cultural Heritage in the Mediterranean Basin”, v.1, Alcala de Henares (Spain). Goins E.S., Wheeler G.S. and Wypyski, M.T., (1996a): “Alkoxysilane film formation on quartz and calcite crystal surfaces”. 8th Cong. on Deterioration and Conservation of Stone, Berlin, 1255-1264. Goins E.S., Wheeler G.S., Griffiths D. and Price C.A., (1996b): “The effect of sandstone, limestone, marble and sodium chloride on the polymerisation of MTMOS solutions”. Proc. 8th Cong. On Deterioration and Conservation of Stone, Berlin, 1243-1254. Horie C.V., (1987): “Materials for conservation, organic consolidants, adhesives and coatings”. London: Butterworth. Kumar R., (1995): “Fourier transform infrared spectroscopic study of silane-stone interface”. In Materials Issues in Art and Archaeology, Materials Research Society Symposium Proceedings 352, eds. P.Vandiver et al., Pittsburgh: Materials Research Society, 341–347. Lin-Vien D., Colthup N.B., Fateley W.G., Grasselli J.G., (1991): “The handbook of infrared and Raman characterisation of organic molecules”. New York: Academic Press. Litardi A.,(1998): “La Loggia e l’Odeo Cornaro, in Padova ritrovata. Itinerari guidati alla scoperta della città restaurata” (brochure edited by Comune di Padova, “Notturni d’arte” 1998), p. 96. Melo MJ, Bracci S, Camaiti M, Chiantore O, Piacenti F., (1999): “Photodegradation of acrylic resins used in the conservation of stone”, Polymer Degradation and Stability; 66, 23- 30. Mietto P., (1988): “Aspetti geologici dei Monti Berici”, in “I Colli Berici-Natura e Civiltà”, Signum Ed., pp 13-23, Vicenza.
170
Moropoulou A., Haralampopoulos G., Tsiourva Th., Theoulakis P., Koui M., (2000): “Long term performance evaluation of consolidation treatments in situ”. Scienza e Beni Culturali XVI, Bressanone. Mosquera M.J., Pozo J., Esquivias L., Rivas T., Silva B., (2002): “Application of mercury porosimetry to the study of xerogels used as stone consolidants”. Journal of Non-Crystalline Solids 311, 185-194. Normal 1/88 (1988): “Alterazioni macroscopiche dei materiali lapidei: lessico”. CNR - ICR (Rome). Normal 3/80 (1980): “Materiali lapidei: campionamento”. CNR -ICR (Rome). Normal 44/93: “Assorbimento d'Acqua a Bassa Pressione”. CNR -ICR (Rome). Price C., (1996): “Stone conservation: an overview of current research”. Los Angeles: Getty Conservation Institute. Rizzarelli P., La Rosa C., and Torrisi A., (2001): “Testing a fluorinated compound as protective material for calcarenite”. Journal of Cultural Heritage 2, 55–62. Rodrigues Delgado J., (2001): “Consolidation of decayed stones. A delicate problem with few practical solutions”, Historical Constructions, P.B. Lourenço, P. Roca (eds.), Guimarães. Sabbioni C., (1995): “Contribution of atmospheric deposition to the formation of damage layers”. The Science of the Total Environment, N° 167, 49-55. Semenzato C., (1987): “La teatralità dell’architettura veneta e la Loggia Cornaro, in Convegno Internazionale di Studi sul Ruzante”, a cura di Giovanni Calendoli e Giuseppe Vellucci (Padova, 26/27/28 maggio 1983), Venezia, Corbo e Fiore, 15-19: Spoto G., Rizzarelli P., and Torrisi P., (2000): “Fluorinated phosphoric ester-based protective material for limestone-made ancient monuments, buildings, and artifacts: An x- ray photoelectron spectroscopy study”. Applied Spectroscopy 54 (12), 1817–1823. Stanley Price N., Kirby Tally M., Melucco Vaccaro A., (1997): “Historical and philosophical issues in the conservation of cultural heritage”. Los Angeles: Getty Conservation Institute. Sunny O., Nwaubani J. and Dumbelton, (2001): “A pratical approach to in situ evaluation of surface treated structures”, Construction and Building Materials, 15, 199-212. Weiss N., Slavid I. and Wheeler G., (2000): “Development and assessment of a conversion treatment for calcareous stone”. Proc. 9th Int. Cong. On Deterioration and Conservation of Stone, Venice, 533- 540.
171