Dissolution mechanisms of albite and hornblende, and of ...

STOCKHOLMS UNIVERSITETS INSTITUTIONför

GEOLOGI och GEOKEMI N:r 297

Dissolution mechanisms of albite and hornblende, and of calcite in sandstone

Paul Frogner

Stockholm, 1998

MEDDELANDEN FRÅN STOCKHOLMS UNIVERSITETS INSTITUTION

FÖR GEOLOGI och GEOKEMI

N:r 278 DARCE, M. Mineralogical Alteration Patterns, Chemical Mobility and Origin of the La Libertad Gold Deposits, Nicaragua. Stockholm 1989.N:r 279 ÖSTLUND, P. Early Diagenesis Processes in Baltic Sediments, with Emphasis in the Distribution of Fallout Plutonium. Stockholm 1990.N:r 280 KANO, A. Deposition, Paleoecology, and Diagenesis of the Silurian Reef-like Limestones on Gotland. Stockholm 1990.N:r 281 SCHWEDA, P. Kinetics and Mechanisms of Alkali Feldspar Dissolution at Low Temperatures. Stockholm 1990.N:r 282 CARMAN, R. Biogeochemical Constraints on Phosphorous Exchange and Related Diagenetic Processes in Baltic Sea Sediments. Stockholm 1990.N:r 283 OLAFSSON, G. Late Oligocène through Late Miocene Calcareous Nannofossil Biostratigraphy and Biochronology. Stockholm 1991.N:r 284 ANDERSSON, P. Hydrogeochemistry of Iron, Manganese and Sulphur- and Strontium Isotopes in a Coniferous Catchment, Central Sweden. Stockholm 1991.N:r 285 BLOMQVIST, S. Geochemistry of Coastal Baltic Sediments: processes and Sampling Procedures. Stockholm 1992.N:r 286 BROMAN, C. Origin of Massive Sulfide Ores in the Skellefte District, as Indi­cated by Fluid Inclusions. Stockholm 1992.N:r 287 BODEN, P Biostratigraphic Implications of Neogene Diatom Abundances in the Norwegian Sea, the North Atlantic and the Western North Pacific. Stockholm 1992.N:r 288 GEBEYEHU, M. Geochemistry, Pb, C-0 Isotope Data and Genesis of Some Lower Proterozoic Massive Sulfide Deposits in the Bergslagen Ore Province, South Central Sweden. Stockholm 1992.N:r 289 MANSFELD, J. Crustal Evolution in the Southeastern Part of the Fennoscandian Shield. Stockholm 1995.N:r 290 MÖRTH, C.-M. Sulfur Isotopes Used as a Tracer of Acidification Reversal, Dispersion of Acid Mine Drainage and Sulfur Dynamics in Small and Large Catchments. Stockholm 1995.N:r 291 ANDERSSON, C. Pliocene Foraminiferal Biostratigraphy and Paleoceanography of the Ontong Java Plateau, Western Equitorial Pacific Ocean. Stockholm 1995.N:r 292 LEI, G. Geochemical Processes Controlling Transition Metal Distributions in Marine Manganese Concreations and Sediments. Stockholm 1996.N:r 293 STERNBECK, J. Manganese and Related Elements in Aquatic Environments: Pathways for Redox Reactions and Environmental Significance. Stockholm 1996.N:r 294 ESTMARK KALINOWSKI, B. Dissolution Kinetics and Alteration Products of Micas and Epidote in Acidic Solutions at Room Temperature. Stockholm 1997.N:r 295 ANDERSSON, E. Hydrothermal alteration of organic matter at spreading centers. Stockholm 1998.Nr 296 STURKELL, E. The Origin of the Marine Lockne Impact Structure, Jämtland. Stockholm 1998.Nr 297 FROGNER, P. Dissolution mechanisms of albite and hornblende, and of calcite in sandstone. Stockholm 1998.

Dissolution mechanisms of albite and hornblende, and of calcite in sandstone

Paul Frogner

Thesis to be publicly defended for the degree of Doctor of Philosophy

in lecture room Nordenskjöldsalen, Svante Arrheniusväg 8b, Friday, April 24, 1998, at 10.00 a.m

Department of Geology and Geochemistry Stockholm University S-106 91 Stockholm Sweden

Stockholm, 1998 ISBN 91-7153-749-X ISSN 1101-1599

Abstract

This thesis discusses the weathering mechanisms of two common silicates, albite and hornblende (Part A) and also weathering and conservation of building materi­als made of calcite cemented sandstone (Part B).

Part A; The role of feldspars and their capacity in soils to buffer against acid rain has together with amphiboles been regarded as important due to their high release of elements and thereby consumption of protons.

In feldspars interstitial cations are released due to an exchange by protons followed by subsequent breakage of Si-O-AI bonds. Raman spectroscopy on albite shows that this weathering mechanism can be studied through the band at 976 cm1. Phis band is considered sensitive to the Al-ordering in feldspars and decreases in intensity for weathered samples.

From dissolution kinetic experiments of hornblende it is concluded that a Si- enriched residual layer is formed after the loss of AI and to a minor extent also Mg and Fe. This is based on the apparent coupling between the initial non-stoichiom- etry and the gradual decrease in dissolution rates until steady state is obtained in experiments at pH 1 and 2. This residual layer and its surface properties are of great significance as it seems to control the dissolution rate of hornblende.

Part B; The Gotland sandstone has been used in ornaments and buildings since the early middle ages. Today this material suffer to chemical weathering due to the acid rain water that dissolves the calcareous cement that binds the sand grains together. Studies of the calcite-cemented sandstone from the Royal Palace balus­trade, Stockholm, show a weathering front 2mm in thickness. Modelling of the carbonate chemistry in the pores of the sandstone indicates that saturation with respect to calcite is reached after only 100 seconds. The weathering rate of the stone is controlled by the transport of these ions from the pores. The preservation of calcite cemented sandstone may thereby be focused on restricting this trans­port.

Tfetraethylortosilicate, TEOS, is an organic compound, which is frequently used as a replacement for the calcite cement in sandstone to reduce the damage by weath­ering. It was found that TEOS did not possess the conservation qualities for sandstone as the dissolution rate of TEOS increased with increasing electrolyte concentration and with increasing pH in the range 1-5.

© Paul Frogner ISBN 91-7153-749-X ISSN1101-1599 Akademitryck Aß, Edsbruk, 1998

STOCKHOLMS UNIVERSITETS INSTITUTIONför

GEOLOGI och GEOKEMI N:r 297

Dissolution mechanisms of albite and hornblende, and of calcite in sandstone

Paul Frogner

Stockholm, 1998

Abstract

This thesis discusses the weathering mechanisms of two common silicates, albite and hornblende (Part A) and also weathering and conservation of building materi­als made of calcite cemented sandstone (Part B).

Part A; The role of feldspars and their capacity in soils to buffer against acid rain has together with amphiboles been regarded as important due to their high release of elements and thereby consumption of protons.

In feldspars interstitial cations are released due to an exchange by protons followed by subsequent breakage of Si-O-Al bonds. Raman spectroscopy on albite shows that this weathering mechanism can be studied through the band at 976 cm1. This band is considered sensitive to the Al-ordering in feldspars and decreases in intensity for weathered samples.

From dissolution kinetic experiments of hornblende it is concluded that a Si- enriched residual layer is formed after the loss of A1 and to a minor extent also Mg and Fe. This is based on the apparent coupling between the initial non-stoichiom­etry and the gradual decrease in dissolution rates until steady state is obtained in experiments at pH 1 and 2. This residual layer and its surface properties are of great significance as it seems to control the dissolution rate of hornblende.

Part B; The Gotland sandstone has been used in ornaments and buildings since the early middle ages. Today this material suffer to chemical weathering due to the acid rain water that dissolves the calcareous cement that binds the sand grains together. Studies of the calcite-cemented sandstone from the Royal Palace balus­trade, Stockholm, show a weathering front 2mm in thickness. Modelling of the carbonate chemistry in the pores of the sandstone indicates that saturation with respect to calcite is reached after only 100 seconds. The weathering rate of the stone is controlled by the transport of these ions from the pores. The preservation of calcite cemented sandstone may thereby be focused on restricting this trans­port.

Tetraethylortosilicate, TEOS, is an organic compound, which is frequently used as a replacement for the calcite cement in sandstone to reduce the damage by weath­ering. It was found that TEOS did not possess the conservation qualities for sandstone as the dissolution rate of TEOS increased with increasing electrolyte concentration and with increasing pH in the range 1-5.

© Paul Frogner ISBN 91-7153-749-X ISSN1101-1599Akademitryck AB, Edsbruk, 1998

Contents

Abstract

Introduction 7

Part A: Weathering mechanisms of albite and hornblende

Part B: Deterioration and conservation of sandstone

8

12

Discussion 15

Acknowledgements

References

17

18

Appendices

Paperi (page 23-42). Frogner, P. and Schweda, P, (In press), Hornblende dissolu­tion kinetics at 25°C, Chemical Geology.

Paper II (page 43-56). Frogner, P, Broman, C., and Lindblom, S., (In press), Weathering detected by Raman-spectroscopy using Al-ordering in albite, Chemical Geolog)'

Paper III (page 57-74). Frogner, P, Mörth, M., Lindblom, S. and Nilsson, Ö., Weathering of calcite cemented sandstone. Manuscript.

Paper IV (page 75-92). Frogner P. and Sjöberg, L., (1996) Dissolution of tetraethyl orthosilicate in acid solution, Proceeding volume, 8th International Congress on Deterioriation and Conservation of Stone, Berlin. Proceedings volume 3: 1233-

Introduction

The aim of this thesis is to investigate the weathering mechanisms of silicate minerals through dissolution kinetics of albite and hornblende. It also concerns the weathering and conservation of calcite cemented sandstone. Two silicate minerals, albite and hornblende are chosen for this study to reach a better understanding of the silicate residual layers and their surface properties. These features control the dissolution rates of silicates and shed light on the properties necessary for consoli­dation products.

Raman spectroscopy is one method that can be used to detect minor structural variatons in feldspars (Velde et al. 1989; Heymann and Hörz, 1990). The large amount of dissolution kinetic data available for feldspars and their fairly well constrained chemistry and structure facilitate interpretations of spectroscopic results of weathered mineral surfaces.

The relation between silicate surface properties and products used in consolida­tion of stone determines the ability of organo-silane based compounds (the most frequently used today) to create bonds to the silicate surfaces. Their assumed tendency to be strongly bonded to the mineral surfaces has been considered important in order to integrate the consolidant with stone to reach a long-term chemical stability (Wheeler et ah, 1992; Goins, 1995).

This work mainly explains weathering mechanisms of albite and hornblende but is also concerned with the weathering and conservation of calcite cemented sand­stone. The dissolution rates of feldspars are well established but few dissolution kinetic data of amphiboles are available and mineral dissolution rates published show large discrepancies. Also, few calcite dissolution studies presents a mechanis­tic description that can be used to explain the calcite dissolution in rocks, e.g. in calcite-cemented sandstone (Plummer et ah, 1978; Sjöberg and Rickard 1983; Arakaki and Mucci, 1995). Furthermore, to increase our knowledge about the chemical and physical parameters of importance for the conservation of stone, there is a need for a long-term chemical stability-test of rock consolidants.

Different aspects are therefore considered as important within this thesis:

1) Dissolution kinetic data of silicates as compared to structural information about reacted mineral surfaces investigated by Raman spectroscopy.

2) Investigation of what mechanisms that control the dissolution rate of calcite in calcite cemented sandstone.

3) Investigation of the chemical stability of alkoxysilane under pH conditions found in rainwater and electrolytic conditions for sandstone.

Part A: Weathering mechanisms of albite and hornblende

Background

The major part of weathering research aims to understand large-scale processes in a geological context. It is well known that weathering processes affect the pH in soils and influence the marine and atmospheric chemistry.

Feldspars which are some of the most abundant minerals in the Earth’s crust supply K, Na, Ga, Al, and Si to the pore solutions in soils and sediments and to the oceans. Feldspars are important as buffers against acid rain due to their abundance and their easy release of elements by consuming protons (Stillings and Brantley,1995) .The same applies for amphiboles and due to their high dissolution rate they are a major source of Ga, Fe and Mg (Sverdrup and Warfvinge, 1992; Brantley and Chen, 1995).

The release rate of Ga and Mg from feldspars and amphiboles can also effect the long-term climate control. Atmospheric C02 dissolved in the oceans combine with released Ga and Mg to form carbonates in marine sediments (Berner and Maasch,1996) . Calcite and its reaction kinetics is therefore one important factor for the regulation of the atmospheric levels of C02due to its equilibrium with marine carbonate deposits. A better understanding of these mechanisms would help to evaluate the accuracy of the dissolution rates obtained from laboratory experiments for models describing the global elemental flux.

Dissolution experiments of feldspars were initiated already in the 19th century by Daubrée, (1867). Mainly two hypothetical mechanisms have been proposed in the literature for laboratory dissolved feldspars and amphiboles, one diffusion con­trolled and one surface reaction controlled (re feldspars, see Correns and Engelhardt, 1938; Lagache, 1965, 76; Wollast, 1967; Helgesson, 1971, 72; Luce et al. 1972; Paces, 1973; Petrovic et al., 1976; Holdren and Berner, 1979; Chou and Wollast, 1985; Heilman, 1995, 97; Schweda et al., 1997; re amphiboles, see Schott et al., 1981; Berner and Schott, 1982).

A diffusion controlled Si-enriched leached surface layer for feldspars was first proposed by Correns and Engelhardt (1938). This suggestion was based on experi­ments of dissolved feldspars where the rates were proportional to t1/2, which could be described with the parabolic rate expression:

dm / dx = k t 1/2 (1),

where m is mass and t stands for the reaction time. They also obtained a preferen­tial release of K and Al relative to silica. They therefore concluded that the forma­tion of the residual layer would explain the initially high preferential release and the gradually decreasing rates. Dissolution experiments were at that time per-

formed in short studies. This resulted in observed release rates that were not in steady state and the leached layer hypothesis was rejected in favour of the surface reaction hypothesis.

Lagache (1965), stated that the dissolution of feldspars is surface reaction con­trolled. This implies that the dissolution is a linear function of time. He thought that any deviation from linear kinetics was due to the precipitation of secondary phases. The surface reaction controlled dissolution of feldspars was further sug­gested by Lagache (1976), Petrovic et al. (1976), Holdren and Berner (1979), where the parabolic rate appearance at that time was explained as due to rapidly reacting fine particles as they couldn’t detect any leached layers or precipitates by Electron spectroscopy for chemical analysis, ESCA. Berner and Schott (1982), suggested a surface reaction control for the dissolution of pyroxenes and amphib­oles as the residual layer that they obtained was too thin to be diffusion limiting.

Recent studies show, however, that not all chemical interactions and weathering processes are restricted to the surface of silicates. The controversy at present in dissolution kinetics research of silicates is about the formation of the leached surface layers as a result of weathering. The structure of these surface layers seems to be very important because they are rate controlling.

The chemical interactions that occur on mineral surfaces during dissolution in aqueous solution involve adsorption, ion exchange, diffusion, hydrolysis, oxidation- reduction and condensation, but the protons are also diffusing into the silicate framework as hydronium ions (Lanford et ah, 1979). The common belief is that protons are exchanged for interstitial cations (Oelkers and Schott, 1992) but protons also counterbalance exceeding negative charge around oxygen bridging bonds. Xiao and Lasaga (1994) showed, using ab initio molecular orbital calcula­tions, that Al-0 bonds are even more weakened than Si-0 bonds by the adsorption of hydronium ions, which will result in a sequential breaking of bonds of the silicate framework producing a leached layer (Gout et ah, 1997).

Dissolution experiments of albite and hornblende have shown that interstitial cations and framework forming A1 is preferentially released relative to silica during early dissolution (re albite, see Heilman, 1995; Gout et ah, 1997; re hornblende, see Zhang et ah, 1996). This initial preferential release is generally followed by a decrease in reaction rates for these elements until stoichiometry is obtained. A formation of a silica-rich residual layer during dissolution in acid solution is there­fore proposed (re K-feldspars, see Oelkers et ah, 1994; re albite, see Heilman,1995; re labradorite, see Schweda et ah, 1997; re Wollastonite, see Casey et ah, 1993; re hornblende, see Zhang et ah, 1996).

The dissolution rate for silicates is dependent on the protonation or deprotonation of the mineral surface. The dissolution rate as a function of pH for a range of

silicates, results is a U-shaped plot, where the dissolution rate is lowest at the point of zero charge pH(izc. The residual layer of silicates resulting from acidic dissolution may therefore initially contain Si-OH groups at acidic conditions and Si-O' groups under alkaline conditions. Recent studies propose a condensation of these groups releasing water and producing a polymerized amorphous silica layer (re wollastonite, see Casey et ah, 1993, re labradorite, see Schweda et ah, 1997, re hornblende, see Zhang et ah, 1996). Casey et ah (1993) explained this as an amorphous domain enriched in silicon and hydrogen, which has been confirmed by Raman spectroscopy and ion beam elemental analyses. Schweda et ah (1997) showed in their experiments of acid leached labradorite that the O/Si ratio was reduced in the near-surface layer compared to fresh laboradorite. They explained the lower O/Si ratio as a spontaneous condensation of adjacent silanol groups and removal of oxygen through the release of water. It has also been proposed that the thickness of this layer and the hydrogen penetration depth increases with the hydrogen activity, ionic strength and temperature of the solution (Chou and Wollast, 1985; Casey et ah, 1988; Petit et ah, 1989; Schweda, 1990; Nesbitt et ah, 1991; Stillings and Brantley, 1995).

Objectives

Raman spectroscopy has become increasingly used as an analytical technique in several geological and mineralogical studies, e.g. for the identification of different mineral phases and the detection of minor chemical and structural variations. It is a non-destructive method; the same sample may be analyzed several times and in combination with other methods. The sample can be placed directly under the microscope without preparation, which is very important in the study of altered minerals, since preparation artefacts are eliminated.

Feldspars constitute a major fraction of the Earth's crust, but only a few spectro­scopic studies have been made of them, including two oriented single crystal studies have been published (von Stengel, 1977, infrared and Raman; Iiishi et ah, 1971, infrared). Several structural alteration studies are addressed to metamor­phosed minerals as e.g. the transition from crystalline to diaplectic glass state as a function of shock stress (Velde et ah 1989; Heymann and Hörz, 1990). Structural differences of glasses as a function of different alkali cation types on the silicate network have also been studied by Raman spectroscopy (Matson et ah, 1986).

The weathering study of A1 leaching in feldspars using Raman spectroscopy was initiated inspired by the large amount of dissolution kinetic data published on feldspars. Several dissolution studies show that A1 leaching and thereby the break­age of Si-O-Al bonds is the main weathering mechanism of the feldspar framework (Blum, 1994; Sjöberg et ah, 1995; Gout et ah, 1997). Albite was chosen in this Raman study because of the structural properties, the Si-O-Si bonds which are

absent in the anorthite unit cell, where half the sites contain Al, half contain Si and all are thus T-O-T bonds of the type Al-O-Si. The Si-O-Si bonds in an ideally ordered unit cell of low albite keep the framework together even though Al-0 bonds are broken. It was also possible with the highly Al-ordered albite to study the specific bands in a Raman spectrum related to the Al-ordering (Velde and Boyer, 1985) that indicate the greater destabilization effect of H+ adsorption in breaking the Si-O-Al bond compared to the Si-O-Si bond (Xiao and Lasaga, 1994).

The dissolution kinetic study of hornblende was initiated to compare the dissolu­tion mechanisms of hornblende to that of other silicates as e.g. albite. Hornblende was chosen partly because of discrepant dissolution rates published in various studies (Sverdrup, 1990; Zhang et ah, 1996), partly because amphiboles are abun­dant mafic minerals in the Earth's crust. Amphibole has been considered to produce high dissolution rates relative to feldspar although less abundant and has thereby been regarded as important for the chemistry of soils (Sverdrup, 1990).

Results

Paper I, establishes the need for much longer reaction times of hornblende to reach a steady state than proposed in earlier studies. The dissolution rate of hornblende was found to be somewhat lower than that reported by Zhang et al. (1996), but more than 100 times lower than that calculated by Sverdrup (1990). These long-term experiments have shown that the residual layer thickness is independent of pH but that it was faster developed at a lower pH. The release of ions did not reach steady state until the residual layer was formed at maximum depletion. It is therefore concluded that the residual layer controls the dissolution of hornblende and hence the release of the interstitial cations.

Paper II shows that Raman spectroscopy is a powerful complementary method in weathering and dissolution studies. It is possible to identify the crystal orientation of pristine albite crystals through specific bands in the wavenumber range 150 - 1200 cm1. In the identification of crystal faces on weathered samples the intensity ratios of bands in the range 150 - 300 are used. The spectral range most sensitive to structural alterations is found at 700 - 1200 cm1. The decrease in intensity found for the band 976 cm1, that is referred to as sensitive to the Al-ordering confirms the breakage of Si-O-Al bonds as the principle weathering mechanism of the feldspar framework. This study also reveals that the knowledge of the crystal orientation and its influence on the band intensity is necessary for a successful description of this weathering mechanism.

Part B: Deterioration and conservation of sandstone

Background

The high anthropogenic production of carbon dioxide has lately augmented the greenhouse effect and the combustion of fossil fuels has raised the levels of sulfu­ric and nitrogen gases in the atmosphere. These emissions produce acid rainwater that are of major concern for the deteriorated stone monuments.

Through the ages, a wide range of rock types have been used as building materials and monuments. Easy access, workability and esthetic appearance have normally caused the choice of material. Galcite-cemented sandstone has been chosen in many parts of Europe in buildings and monuments, in spite of its low' resistance to chemical weathering, because of its availability and workability. Gotland sandstone consists mainly of quartz with minor quantities of feldspar and mica and has been widely used in the northern part of Europe (Löfvendahl, 1996). Its high porosity and the relatively low amount of calcite, that hold the mineral grains together, makes it probably more exposed to the weathering processes than most other building materials. A large amount of calcite dissolution studies are done but few' presents a mechanistic description that can be used to explain the calcite dissolu­tion in rocks, like in the calcite-cemented sandstone (Plummer et ah, 1978;Sjöberg and Rickard, 1983; Arakaki and Mucci, 1995).

Several types of sandstone treatments have been used in the past to protect or replace the natural calcite cement lost by weathering. Lime mortar, cement, waxes, acrylic resins and epoxies are some of the known treatments tried. Waxes, acrylic resins and epoxies were however found to discolor the material and to restrict water vapor to pass in and out through the rock pores, which causes spalling (exfoliation). Acrylic resins was also found unstable to solar radiation. These materials are therefore no longer considered as acceptable.

Alkoxysilane is an organic compound, which is frequently used as a replacement for the calcite cement in sandstone to reduce the damage by weathering. Alkoxysilane denotes a group of compounds where the most commonly used is tetraethylorthosilicate, TEOS.

Alkoxysilane based treatments are not regarded as a general solution for all variet­ies of stone material as the physical as w'ell as chemical characteristics of stones vary much.

The desired qualities of consolidants should be stated as follows: 1) During consolidation, pores and cracks in the stone must be filled to prevent direct transport of water which might lead to further deterioriation; 2) A low viscosity is needed for a deep penetration; 3) The product must adhere to the mineral sur-

faces of the stone and be durable; 4) It must also resemble the stone closely in physical properties allowing water vapour to pass in and out of the pores and closely resemble the stone in color appearance.

An ethoxysilane based consolidant used for sandstone reacts with moisture and an organo tin catalyst as follows:

Sn-OCOH, + H20-> CH,COOH + RSn-OH (2),

R Si-OR +R Sn-OH-»R Sn-O-SiR + ROH (3),

R Sn-O-SiR + H,0->R Si-O-SiR + Sn-OH (4),

Where R denotes ethyl groups and n is the index.

The exact mechanisms of the condensation reaction are poorly understood, but the reaction requires a catalyst in combination with water. It is assumed byVan der Weij (1980) that the reaction is initiated in contact with water which hydrolyses the carboxylate group of the catalyst under the release of carboxylic acid (see reaction 2). Ethyl groups could thereafter be hydrolysed and separated from the orthosilicate in the presence of tin hydroxide groups. This releases ethanol and formes Si-O-Sn bridges (see reaction 3 and 4). These bridges are easily hydrolysed at the presence of protic agents like water forming the more chemically stable siloxane bonds.

The reasons why alkoxysilanes have been considered best for sandstone conserva­tion (Wheeler et ah, 1992; and Goins et al., 1995), are based on: 1) The higher mechanical strength reached for sandstone as compared to carbonate stones consolidated with alkoxysilane. Discrepancies in mechanical strength between these materials are explained as due to the lack of cross linking bonds to the carbonate surface; 2) A high ability to penetrate deeply into the material. The deep penetration is due to its low viscosity that is further enhanced by the large pore volume of sandstone; 3) Its chemical stability, which is a result of how well the monomers are polymerized because of the silicon-oxygen-silicon bonds (silox­ane bonds) forming the skeletal structure of the polymer. The siloxane bonds (unlike C-G bonds) have no tendency to form conjugate double bonds (which are highly reactive sites).

The assumed advantages with alkoxysilane for sandstone conservation are based on its chemical and physical properties and the increase in mechanical strength reached at the use on sandstone. However very little is known about the long-term chemical stability of the resulting polymer.

This part of the thesis considers the resistance of polymerized alkoxysilane against chemical weathering for a carbonate hosted rock. Previous studies on alkoxysilane have mainly focused on the increased mechanical strength due to the consolida­tion. The aim with this study was to investigate the dissolution rates of alkoxysilane at pH conditions, 1 - 5. It covers the prevailing pH of rain-waters (pH 4-5) which is expected on stone surfaces. But the trend in pH dependence can also be used to evaluate its long-term chemical stability at higher pH values as in the sandstone pores. The influence of different electrolytes on the dissolution rate was investigated because a high ionic strength is expected in the pore-waters of natural stone. It was also investigated if the polymerization rate was influenced by a variation in humidity. The two most important reactions during polymeriza­tion of etoxysilane are hydrolysis of ethyl groups bonded initially to orthosilicates and the succeeding condensation of orthosilicates producing siloxane bonds. These two reactions were investigated during the polymerization at varied humidity by Raman spectroscopy. As quartz is the prevailing mineral in sandstone it was also important to see how the alkoxysilane based consolidant was bonded to the min­eral surface of quartz.

The aim is furthermore to increase the knowledge of the dissolution meachanisms of calcite in sandstone. This background information is needed to illuminate the advantages and drawbacks in using alkoxysilane based products in the consolida­tion of stone.

Results

In paper III, a first attempt to model the calcite dissolution in a calcite-cemented sandstone is presented. The model predictes that saturation in the pore waters with regard to calcite is reached wirhin 100 seconds and at a pH of about pH 8.3. The dissolution of calcite would therefore be restricted by the transport of ions from the sandstone pores to the surface. The diffusion control is further supported by the observed weathering front, 2 mm thick, found in the microscopic investiga­tion. The mineral grains in the affected surface layer are still kept together. The open porosity in this layer serves as a buffer zone, which prevents direct contact of the low pH of rainwater.

Paper IV investigates the reaction kinetics of tetraethylorthosilicate (TEOS, Steinfestiger-OH, a common product used to consolidate sandstone) as a function of relative humidity, and shows that polymerization went faster at higher relative humidity. Dissolution experiments of TEOS in electrolytic solutions and in the pH range 1-5, revealed an increase in dissolution rates towards a higher pH, and that the dissolution rate also increases at higher ionic strengths. The effect on the dissolution rate for different electrolytes was similar between NH Cl, Nad and

KCl. Exchanging NaCl with solutions of Na2S04 and NaNO, shows that the anions take no active part in the weathering mechanism as no change in dissolution rate was observed.

Discussion

Surface properties of silicates and consolidation of stone

The theory to chemically link the alkoxysilanes to hydroxyl groups on the surface of silicates is based on the discovery 1936 of hydroxyl groups on silica by means of a selective reacting compound (Vansant et ah, 1995). Among a range of reacting compounds organosilane was used (Vansant et ah, 1995).

However, the dissolution study of TEOS surface coatings on silica presented in paper IV shows that bonding to the silica surfaces does not reach the high stability that was advocated by the manufacturer. This can be seen in SEM microphoto­graphs where the quartz shows a lack of adhesive products attached to the grain surfaces after dissolution at pH 5 in 0.01M NaCl solution at 175 hours.

Weathering of albite and hornblende minerals are, in contrast to quartz, forming a leached surface layer enriched in silica (Casey et ah, 1993; Schweda et ah, 1997; Heilman et ah, 1997; Frogner and Schweda, 1998) that seems to control the diffusion of variable cations to different extent. In paper I, it is suggested that there is no positive correlation between the cation-depleted depth of the residual layer and the hydrogen activity. However the maximum Al-depletion depth is faster developed at a lower pH. This implies that some other property determines the thickness of this layer as no relation is found with it and the hydrogen activity. The diffusion control is addressed as due to the condensation of abandoned OH groups. These groups become abandoned due to an exchange of interstitial cations by hydrogen at tetrahedral vacancies in feldspars (Schweda et ah, 1997) and at octahedral positions in amphibole (Zhang et ah, 1996). The most important mechanism for the residual layer formation of albite is probably the preferential breaking of Al-O-Si bonds in the silica framework. This is also confirmed by Raman spectroscopy to be the principle weathering mechanism of albite in paper II. This study also reveals that knowledge of the crystal orientation and its influ­ence on the band intensity is crucial for a successful description of this weathering mechanism.

It is difficult to make any final conclusion about the residual layer of feldspars and the effect of it in contact with alkoxysilane in sandstone. The weathering of silicates includes a re-polymerization of abandoned silanol and oxide groups in the residual layer (Casey et ah, 1993). The bonding abilities of the alkoxysilanes are thereby only possible at the surface. But the surface layer of feldspars should be

more brittle in contrast to the surface of a quartz that has been dissolved by an ideal controlled surface mechanism.

Alkoxysilane needs a nearby catalyst in order to condensate siloxane bonds fast, which otherwise is a very slow process (van der Weij, 1980). I propose that the thickness of the residual layer of hornblende describes the density of re-polymer­ized siloxane bonds needed to restrict the transport of cations. This implies that the the reaction is diffusion controlled until a steady-state is reached However, a surface controlled reaction dominates when this layer becomes an effective barrier for the ions at the mineral/residual layer -interface to diffuse. The hydrolysis of Si- O-Si bonds at the immediate surface will then become the limiting reaction step (re: albite, see Gout et ah, 1997).

In the study of TEOS in paper I\( an organic-tin catalyst is included in the prod­uct. Eventhough tin compounds have been known for a long time, the mechanism of its catalytic reaction is little known. Some authors also suggest that tin still could be bonded to the structure of the final product of alkoxysilanes after a completed polymerization (van der Weij, 1980; Frogner and Sjöberg, 1996). But this would probably not explain the fast release rates of silica found in paper IV towards a higher pH. The Si-O-Sn bonds are easier to hydrolyse than Si-O-Si bonds (van der Weij, 1980) and a higher dissolution rate should therefore be expected in the lower pH experiments, whereas the opposite was found.

It may finally be concluded that alkoxysilane does not possess the conservation qualities that were intended. In the pH range 1 - 5, the rates increased with increasing electrolyte concentration. The effect of electrolytes also increased with increasing pH. The increase in dissolution rate is an effect of the cation concentra­tion. Due to these results a high weathering rate of alkoxysilane is expected in sandstones, because the pH is expected not to fall below pi I 4 where TEOS is chemical stable.

Surface properties of silicates have not been studied in enough detail to explain the attachment of silanol groups that should be considered as important in the conservation of stone at present. Here I propose that the consolidants in use are not able to form sufficiently effective bonds with the silicate mineral surfaces, in order to resist against chemical weathering. But the most important aspect may be the low chemical stability of alkoxysilane at the pH of rainwater and the trend of an increase in dissolution rates towards a higher pH. By careful consideration of this problem, development of more effective and chemically stable consolidants can be made, because the concept is in the right direction, but research has been incomplete so far.

Remarks and suggestions for further studies

My purpose is to further investigate the residual layers on hornblende with Raman spectroscopy to identify its structure and in a second article discuss its surface properties in more detail as this study only explains the residual layer formation from a dissolution kinetic point of view. It may also be interesting in future to investigate the interaction of a organo-tin catalysts at the development of residual layers in dissolution kinetic experiments of silicates.

The protection of sandstones could perhaps be obtained by a hydrophobic agent to restrict the transport of ions in solution to and from the sandstone surface. As the hydrophobic agents are thought to introduce tension in the material, they are not considered as acceptable in Sweden at present. The evidence to support this is however still poor.

Acknowledgments

These studies were initiated in discussion with my former supervisor Lennart Sjöberg. I am sincerely grateful for the advantage to be in his research group and for the time that he spent with us, to give his view of research and how to enjoy things in life.

I would like to express my gratitude to Prof. Kurt Boström (who became my supervisor after the passing away of my former supervisor Lennart Sjöberg) for his help, support and improvements of my articles through many discussions.

I would also like to express my gratitude to Curt Broman, Sten Lindblom and Carl- Magnus Mörth that closely followed, advised and encouraged me through out my graduate work. Carl-Magnus Mörth for introducing me in to calcite dissolution kinetics. Curt Broman and Sten Lindblom for all their efforts during the graduate student time and for introducing me into Raman spectroscopy and discussions in all other matters.

Vladimir Baumruck, Department of Physics, University of Prag for teaching me about orientational aspects, in Raman spectroscopic analysis of minerals. Peter Schweda for teaching me about dissolution kinetics of silicates and for his and Sten Linbloms support when I had my first international talk. Marianne Ahlbom and Birgitta Boström for their kindness, guidance and support in the laboratory, and Lilly Johansson for help with analysis. Anders Sundberg for his skillful con­structions of all the experimental equipment that I needed. Carl-Magnus Mörth, and Martin Jacobsson for bringing me the computers that I used. Arne Lif for technical support of the computers. Lars-Erik Bågander, Nils Holm, Erik Sturkell and Kjell Wannäs are also thanked for the advice and helpful comments on the manuscript

I would also like to give my appreciation to my colleagues and friends Birgitta Estmark-Kalinowski, Heléne Strandh and Elin Carlsson. Kerstin Gille-Johnson and Runa Jacobson for their kind support in all other matters at the department. Kjell Wannäs who introduced me to the sports club “barnen i boulerbyn” made the time during the studies even more enjoyable.

And to my family that have been very understanding during this period and that always is supporting me.

This work was financed by: Stockholm University (SU), the Central Board of National Antiquities (RAA), the Royal Academy of Seiendes (KVA) foundations: Hierta-Retzius and C F Liljevalch J:ors, European Science Foundation (ESF) and the Geochemical Society of America (GSA).

References

Arakaki, T. and Mucci, A., 1995, A continuous and Mechanistic Representation of Calcite Reaction-Controlled kinetics in Dilute Solutions at 25°C and 1 Atm Total Pressure.

Berner, R.A. and Maasch, K.A., 1996. Chemical weathering and controls on atmo­spheric 02 and C02: Fundamental principles were enunciated by J.J. Ebelmen in 1845, Geochim. Cosmochim. Acta, 60:1633-1637.

Berner R.A. and Schott, J., 1982; Mechanism of pyroxenes and amphibole weather- ing-II, Observation of soil grains, Am. Jour. Sei., 282, 1214-1231.

Blum, A.E., 1994. Feldspars in weathering. In: Parsons I. (editor) Felspars and their reactions. NATO Advanced Study Inst, Kluwer Academic Pub, Netherlands. 595- 629.

Brantley S.L. and Chen, Y., 1995, Chemical weathering rates of pyroxenes and amphiboles. Reviews in Mineralogy, 31: 565-583.

Casey, W,FI., Westrich, H.R. and Arnold, G.W, 1988, Surface chemistry of laboradorite feldspar reacted with aqueous solution at pH = 2, 3, and 12.Geochim. Cosmochim. Acta, 52;2795-2807.

Casey, WH., Westrich, H.R., Banfield, J.E, Ferruzzi, G. and Arnold, G.W, 1993, Leaching and reconstruction at the surfaces of dissolving chain-silicate minerals. Nature, 366: 253-255.

Chou, L and Wollast, R., 1985, Steady-state kinetics and dissolution mechanisms of aibite. Am J Sei., 285:963-993.

Correns, G.W and von Engelhardt, W, 1938, Neue Untersuchungen uber die Verwitterung des Kalifeldspates. Chemie der Erde, Bd. XII, 1-22.

Daubrée, A., 1867, Expériencies sur les décompositions chimiques provoquées par les actions méchniques dans divers minéraux tels que le feldspath. Acad. Sei. Paris, Comptes Rendus 64, 339-346.

Frogner, R, Broman, C. and Lindblom, S., 1998, Weathering detected by Raman spectroscopy using Al-ordering in Albite, Chemical Geology. In press.

Frogner, P. and Schweda, R, 1998, Hornblende dissolution kinetics at 25°C. Chemical Geology, on the geochemistry of crystal fluids. Chemical Geology. In press.

Frogner, P. and Sjöberg, L, 1996, Dissolution of tetraetylorthosilicate coated quartz grains in acid solution, “ the 8 th International Conference on Deterioration and Conservation of Stone, Berlin. Proceedings volume 3: 1233-1242.

Goins, E.S., 1995. Alkoxysilane stone consolidants: The effect of the stone sub­strate on the polymerization process. Doctoral dissertation. University College London.

Gout, R., Oelkers, E., H., Schott, J. and Zwik, A., 1997. The surface chemistry and structure of acid-leached albite: New insight on the dissolution mechanism of the alkali feldspars. Geochim. Cosmochim. Acta, 61 ;3013-3018.

Flelgeson, H.C., 1971, Kinetics of masstransfer among silicates and aqueous solutions. Geochimica et Cosmochimica Acta 35;421 -469.

Helgeson, H.C., 1972, Kinetics of masstransfer among silicates and aqueous solutions. Correction and clarification. Geochim. Cosmochim. Acta 36;1067-1070.

Flellman, R., 1995, The albite-water system: Part II. The time-evolution of the stoichiometry of dissolution as a function of pFI at 100, 200 and 300°C. Geochim. Cosmochim. Acta, 59; 1669-1697.

Heilman, R., 1997, The albite-water system: Part III. Characterization of leached and hydrogen-enriched layers formed at 300OC using MeV ion beam techniques. Geochim. Cosmochim. Acta, 61;1575-1594.

Heymann, D. and Hörz, E, 1990, Raman-Spectroscopy and X-ray Diffractometer Studies of Experimentally Produced Diaplectic Feldspar Glass, Phys Chem Miner­als, 17:38-44.

Holdren, G.R. Jr and Berner, R.A., 1979, Mechanism of feldspar weathering-I. Experimental studies. Geochim. Cosmochim. Acta, 43:1161-1171.

Iiishi, K., Tomisaka, T., Kato, T, and Umegaki, Y., 1971, Isomophous substitution and infrared and far infrared spectra of the feldspar group. Neues Jahrbuch. Miner., 115:98-119.

Lagache, M., 1965, Contribution l’etude de l’alteration des feldspaths dans l’eau entr 100 et 200OC, sous diverses pressions de C02 et applications a la synthese des minéraux, Bull. Soc. Fr. Minerai Cristallogr. 88:223-253.

Lagache, M., 1976, New data on the kinetics of the dissolution of alkali feldspares at 200°C in C02charged waters. Geochim. Cosmochim. Acta 40, 157-161.

Lanford, WA., Davies, K., LaMarche, R, Laursen, T, Groleau, R. and Doremus, R.,H., 1979, Hydration of soda-lime glass. Journal of Non-Crystalline Solids 33, 249-266.

Luce, R. W, Bartlett, R.W and Parks, G.A., 1972 Dissolution kinetics of magne­sium silicates. Geochm. Cosmochim. Acta 36;35-50.

Löfvendahl, R., 1996, Gotland sandstone in Swedish buildings ” the 8 th interna­tional conference of deterioration and consolidation of stone. Proceeding volume I:11-19.

Matson, D.W, Sharma, S.K. and Philpotts, J.A., 1983, Raman spectra of some tectosilicates and of glasses along the orthoelase-anorthite and nepheline-anorthite joins. Am. Mineral. 71:694-704.

Nesbitt, H.W, Marae, N.D., and Shotyk, W, 1991, Congruent and incongruent dissolution of labradorite in dilute acidic salt solutions. J.Geol. 99, 429-442.

Oelkers, E., H., Schott, J., 1992. The dissolution rate of albite as a function of chemical affinity and the stoichiomery of activated complexies in aluminosilicate dissolution reactions. GSA abstr. Progr. 24, A-207.

Oelkers, E., H., Schott, J. and Dévidai, J-L., 1994, The effect of aluminium, pH, and chemical affinity on the rates of aluminosilicate dissolution reactions. Geochim. et Cosmochim. Acta, 9: 2011-2024.

Paces, T, 1973, Steady state kinetics and equlibrium between ground water and granitic rock. Geochim. Cosmochim. Acta, 37:2641-2663.

Parks, G.A., 1967, Aqueous surface chemistry of oxides and complex oxide miner­als. Advances in Chemistry Series 67, 121-160. American Chemical Society.

Petit, J.-C., Dran, J.C., Paccaguril, A. and Della Mea, G., 1989, Structural depen­dence of crystalline silicate hydration during aqueous dissolution. Earth Planet.Sei. Lett. 93:292-298.

Petrovic, R., Berner, R.A. and Goldhaber, M.B., 1976, Rate control in dissolution of alkali feldspars-I. Study of residual feldspar grains by X-ray photoelectron spectros­copy. Geochim. Gosmochim. Acta, 40:537-548.

Plummer, L.N. and Wigley, T.M.L. and Parkhurst, D.L., 1978. The kinetics of calcite dissolution in C02-water systems at 5-60°G and 0.0-1.0 atm COr Am. Jour. Sei., 278:179-216.

Schott, J., Berner, R.A. and Sjöberg, L.E., 1981, Mechanism of pyroxenes and amphibole weathering-I. Experimental studies of iron-free minerals, Geochim. Gosmochim. Acta, 45:2123-2135.

Schweda, R, 1990, Kinetics and mechanisms of alkali feldspar dissolution at low temperatures. Ph.D dissertation, Stockholm Univ.

Schweda, R, Sjöberg, L., and Söderwall, U., 1997, Near surface composition of acid-leached laboradorite investigated by SIMS. Geochim. Gosmochim. Acta, 45:2123-2135.

Sjöberg E.L. and Rickard, D.T., 1983, Calcite dissolution kinetics: Surface spécia­tion and the origin of the variable pH dependence. Chem. Geol. 42:119-136.

Sjöberg, L., Strandh, H. and Schweda, R, 1995. Diffusion of protons and cations in leached feldspar surface layers. Terra abstracts, Abstract supplement No. 1 to Terra Nova 7, 65, EUG 8.

Stillings, L.L. and Brantley, S.L., 1995, Feldspar dissolution at 25°C and pH 3: Reaction stoichiometry and the effect of cations. Geochim. Gosmochim. Acta 59:1483-1496.

Sverdrup, H.U., 1990. The Kinetics of base cation release due to chemical weath­ering. Lund University Press, 245 pp.

Sverdrup H.U. and Warfvinge, R, 1992, Caculation of field weathering rate, using a mechanistic geochemical model-PROFILE, Applied Geochemistry 27.

Vansant, E.E Van der Voort, P. and Vrancken, K.C., 1995, Characterization and chemical modification of the silica surface, studies in surface science and catalysis 93, Elsevier, pp 1 - 556.

Van der Weij, F W, 1980, The action of tin compounds in the condensation-type R'lV silicone rubbers. Makromol. Chem. 181; 2541-2548.

Velde, B. and Boyer, H. J., 1985. Raman microprobe spectra of naturally shocked microcline feldspars. Geophys. Research. 90: 3675-3682.

Velde, B., Syono, Y, Kikuchi, M. and Boyer, H., 1989. Raman microprobe study of synthetic diapletic plagioclase feldspars. Phys. Chem. Minerals, 16: 436-441.

von Stengel, M.O., 1977. Normalschwingungen von alkalifcldspäten. Zeitschrift für kristallographie, 146:1-18.

Wheeler, G. Segan, Fleming, S.A., Ebersole, S., 1992. Comparative strengthening effect of several consolidants on Wallace sandstone and Indiana limestone. In “The Deterioration and Conservation of stone ,vol.3 “Rodrigues, J.D., Henriques,E, Jeremias, ET., editores. Laboratorio Nacional de Engenharia CivikLisbon, 1033- 1041.

Wollast,R., 1967, Kinetics of the alteration of K-feldspar in buffered solution at low temperature. Geochim. Cosmochim. Acta, 31: 635-648.

Xiao, Y. and Lasaga, A.C., 1994, Ab initio quantum mechanical studies of the kinetics and mechanisms of silicate dissolution: H+(H,0+) catalysis. Geochim. Cosmochim. Acta, 58: 5379-5400.

Zhang, H., Bloom, P.R., Nater, E.A. and Erich, M.S., 1996, Rates and stoichiometry of hornblende dissolution over 115 days of laboratory weathering at pH 3.6-4.0 and 25°C in lithium acetate. Geochim. Cosmochim. Acta, 60: 941-950.

© Paul Frogner ISBN 91-7153-749-X ISSN1101-1599Akademitryck AB, Edsbruk, 1998