Effect of different ions on dissolution rates of silica and feldspars at high pH

The dissolution kinetics of silica-containing minerals in aggregates influences strongly the process of ASR in concrete. In this paper, the effect of different ions on the dissolution rates of SiO2 (amorphous and quartz) and feldspars at high pH values was studied by following the increase of silicon concentrations in dissolution experiments and with a novel approach of measuring the evolution of scratches of polished surfaces. The second method avoided the problem of precipitation in some systems, such as the formation of C-S-H when calcium was present and lithium silicates in the presence of lithium. At high pH values, lithium, calcium and sulfate increased the dissolution rates of silica and feldspars, while iron, magnesium and additional NaCl, KCl or CsCl showed no significant effect. In contrast, aluminium slowed down significantly the dissolution rates of quartz, amorphous silica and Na and K-feldspar at for all temperatures studies: 20, 40 and 60 degrees C.

High vacuum chemical vapor deposition (HV-CVD) of alumina thin films

Xavier Multone

We analyzed along this work the feasibility to produce high quality alumina thin films by High Vacuum Chemical Vapor Deposition (HV-CVD). We study the influence of various parameters on the growth process and on the film quality, such as substrate temperature, gas flow or pressure. We aim to highlight the advantages and the limitations of the HV-CVD technique. A high vacuum CVD reactor has been designed, built and optimized. A novel effusing source has been designed to guarantee a molecules distribution uniformity of 95% on the substrate surface. The main technical challenges faced were to reach the maximum substrate temperature in vacuum, to control the precursor flux and to control the temperature of the gas lines to avoid condensation. For the first time, films of Al2O3 are obtained by HV-CVD from aluminum isopropoxide as precursor in a novel HV-CVD reactor. The HV-CVD approach stands out of all the other CVD methods as it allows excellent prediction of growth rates and film thickness homogeneities if the decomposition probability of the precursor is once determined. In this work, the latter was determined to increase with increasing substrate temperature to 0.2, further increase of temperature reduces the decomposition probability due to increased desorption rate. Pure alumina films are obtained in a reliable and reproducible way. The presence of oxidizing agents, as decomposition partners, reduces the activation energy of deposition process from 33.1 ± 8.2 kJ/mol to 11.4 ± 5.3 kJ/mol in presence of sufficient oxidizing agent. The Al2O3 films are growing at deposition rate from 1 to 50 nm/min. It has been found that, the lower the deposition temperature, the higher the density of the deposited films therefore the higher refractive index. The index of refraction varies from 1.65 ± 0.01 to 1.35 ± 0.01 at 632 nm wavelength from lowest to highest deposition temperature. At low deposition temperature, Al2O3 can contain residual OH-groups and Al=O entities. This content is independent of the total flow rates of the precursor in the range of 5·1015 molecules/cm2·s to 3·1016 molecules/cm2·s impinging on the substrate. Substrate effects have not been observed in the limited range of tests on natural oxidized silicon, 3 µm SiO2 on silicon, quartz, stainless steel, fused silica and 100 nm silicon nitride on top of silicon. The chemical composition of the deposited alumina films measured by EDX and XPS is nearly stoichiometric with 35 ± 5 at% Al and 65 ± 5 at% O, without carbon contamination. Remaining hydrogen in the films has not been studied in detail, but the difference of OH- absorption peaks by FTIR indicates that, a low temperature deposition, hydrogen incorporation is possible. Concerning the optical properties of the HV-CVD deposited films, absorption is very low from 250 nm to 1800 nm wavelength, but due to the porosity and granular structure of the films, light scattering can take place. But propagation loss smaller than 2 dB/cm were measured in channel waveguides fabricated from the HV-CVD alumina films. Planar and channel waveguides demonstrate good guiding properties at 670 nm and 1.55 µm, as well. This work also opens new possibilities to deposit in situ local structures of or in transparent alumina on large surface. Light induced high vacuum chemical vapor deposition of alumina microstructures by 248 nm excimer laser is presented and shows an activation energy of 5.2 ± 0.4 kJ/mol, much smaller than for thermal deposition. The influence of the fluence and the repetition rate is discussed. Electron beam assisted HV-CVD of alumina is demonstrated and proves the feasibility of in situ structuration under HV-CVD conditions.

EPFL2009

Aluminium and alkali uptake in calcium silicate hydrates (C-S-H)

Calcium silicate hydrate (C-S-H) is the main hydrate phase in Portland cements. Its composition varies depending on the Ca/Si ratio and on the presence of other ions such as aluminium (called CA- S-H) or alkalis. This thesis aimed to study the composition and the structure of C-S-H. An experimental data base was realised on the effect of different Ca/Si and Al/Si ratios, alkali concentrations (potassium and sodium hydroxide) and of different temperatures on C-S-H after equilibration times up to 1 year. The composition of the solid phases was analysed by TGA, XRD/Rietveld analysis and 29Si and 27Al MAS NMR. The solutions were analysed by ionic chromatography and pH measurements. In the absence of alkali, only C-A-S-H is formed at Al/Si ≤ 0.1. At higher Al/Si ratios, katoite, stratlingite and/or AH3 precipitate in addition to C-A-S-H, such that a part of the aluminium is consumed and limits the Al/Si ratio in C-S-H to 0.15±0.05 regardless of the Ca/Si ratios. A strong correlation between the aqueous aluminium concentration and the aluminium uptake in the solid phase is observed. Aluminium in C-S-H is observed mainly as tetrahedrally, Al(IV) or octahedrally coordinated, Al(VI) aluminium. At high Ca/Si ratios, Al(VI) is favoured and the aluminium uptake in C-S-H increases with the aqueous aluminium concentration. At low Ca/Si ratio (≤ 0.8), aluminium is observed mainly as Al(IV) and the low aqueous aluminium concentrations (below detection limit) indicate a higher affinity of aluminium towards C-S-H. The uptake of aluminium increases slightly the mean basal spacing slightly at low Ca/Si ratio and up to ≈2Å at higher Ca/Si ratios due to incorporation of aluminium and calcium in the interlayer. The presence of alkali hydroxide increases the pH leading (i) to a destabilisation of stratlingite and (ii) to higher aqueous silicon and aluminium concentrations and to lower calcium concentrations. As the aluminium uptake in C-S-H is strongly related to the aqueous aluminium concentrations, the aluminium uptake in the C-S-H increases significantly in the presence of alkali hydroxides, as aluminium is more soluble at higher pH. The increase of the temperature from 7 to 80 °C increases the phase-purity and long-range order of the C-S-H. The mean chain length remains stable up to 50°C, while at 80°C chain length increases and crosslinking is observed in the presence of aluminium. At all temperatures studied, the presence of aluminium has little effect on the calculated solubility of C(-A)-S-H. The incorporation of alkali ions in C-S-H increases with the initial concentration of alkali and with the decrease of Ca/Si ratio. The aqueous calcium concentration play a dominant role in alkali uptake, as calcium competes with the alkali ions to charge balance the negatively charged silanol groups of the C-S-H. A low concentration of calcium favours the alkali uptake in C-S-H. The uptake of alkali ions shortens the silica chain length as calcium is redistributed from the interlayer to the main sheet. The alkali uptake is not significantly influenced by the presence of aluminium in C-S-H, the hydration time or the type of cation (sodium or potassium).

EPFL2014

Typologie des eaux souterraines de la molasse entre Chambéry et Linz (France, Suisse, Allemagne, Autriche)

The geochemical typology of natural subsurface waters of the Molasse basin is part of the AQUITYP project, initiated by the Geology Laboratory (GEOLEP) of the Swiss Federal Institute of Technology in Lausanne (EPFL). This study provides an overview of the variation in space of the hydrochemical parameters, especially of trace elements, in order to determine the geological origin of their aquifers. Our data base is mainly made up of two observation networks: a subsurface water network and a deep groundwater network. Complete analyses of all samples have been executed using standard methods for anions and silicium, and high resolution inductively coupled mass spectrometry (HR-ICP-MS) for major cations and trace elements down to the detection limit of about 0.2 µg/l. Subsurface water network It is the main observation network of this investigation and comprises 112 water samples from different springs, collected all over the Molasse basin between Chambéry (France) and Linz (Austria), but mainly in the Swiss Molasse basin. For this region, four groups are generally distinguished, according to the sediment deposition in a continental or marine environment, in ascending stratigraphic order: (1) the Lower Marine Molasse (UMM and North Helvetic "flysch" p.p.), (2) the Lower Freshwater Molasse (USM), (3) the Upper Marine Molasse (OMM) and (4) the Upper Freshwater Molasse (OSM). Our water samples originate from typical sedimentary deposits which we divided in about 30 different sub-type aquifers according to their geological origin and their depositional environment. They are defined on the basis of lithological and structural homogeneity. Therefore, only catchments that completely drain a single, lithologically homogenous, Molasse terrain unit are considered here. Spring catchments selected were generally not affected by runoff from towns, or paved roads. Agricultural activities have been avoided as well as possible. Such spring water should be free of major human influences and only in few cases contain a certain amount of fertilizer products (p.ex: nitrates). Leachate tests of crushed rocks and comparison with snow and soil water compositions allowed to decipher the geogene origin of the dissolved ions in natural groundwaters. The water withdrawn can be generally classified as a calcium-magnesium-hydrogenocarbonate (Ca-Mg-HCO3)-type water. The trace analyses allowed to distinguish several subtypes within this homogenous water type group. The general spacial distribution of the trace elements show, that spring waters in the Bavarian Molasse basin are caracterized by elements as lithium and uranium linked to acid rocks, while those of the Swiss Molasse basin are preferentially influenced by elements derived from mafic or ultramafic rocks, such as chromium and cobalt. The following elements show specific links between aquifer rock type and groundwater. Sulfates: Total mineralisation of spring waters issued from the gypse-bearing Molasse (USM) and the "Glimmersande" (OSM) are highly influenced by high sulfate contents. It can be demonstrated that sulfate ions are produced from two different sources within this two sub-type aquifers: from the dissolution of gypsum in water, producing high calcium contents as well, and from the oxydation of sulfurous ore minerals, especially pyrite, which are abundant in the "Glimmersande". The "Glimmersand" aquifers in NE Switzerland are furtheron caracterized by high lithium, molybdenum and uranium concentrations, reflecting the granitic composition of this sandstones. High barium contents (>150 µg/l) have been found in spring waters issued from aquifers with a dominant fissure permeability ("Grès de la Cornalle", "Gibloux-Delta" and "Rigi-Schüttung"). Its origin is supposed to be the mineral barite, which tends to precipitate in fractures. Therefore, high barium concentrations may indicate the presence of fracture flow within the aquifer. Considerably high silicon contents are linked to the quartzitic and fine grained aquifers of the brackish marine molasse in Bavaria. Both, petrographical composition and slow flow rate are favouring the water-rock interaction. Caracteristic lithium footprints are found in several waters, as in the aquifers composed of the "Glimmersande", the gypse-bearing molasse, the Rigi-conglomerates, the brackish and the estuarine marine molasse. Chromium strongly marks the waters originating from the Swiss marine Molasse aquifers. It is probably related to ophiolithic detritus derived from the flysch of the Prealpes. A couple of waters show weak anomalies of cobalt, which are nevertheless intimately linked to the presence of ophiolithic components in the sediments. Cobalt might be used as a hydrostratigraphical tool to identify greenstone inputs. Analysed colloidal aluminum may serve as a hydrodynamical parameter, indicating microturbidity within the aquifer. High amounts of aluminum varying in time, have been found in estuarin deposits of the OMM. Generally speaking, we propose three main origins of trace elements in the spring water of the Molasse basin: presence of certain petrographical components within the sediments such as basic and ultrabasic rocks, heavy minerals and evaporites. a small flow rate through interstitial porosity of fine grained sediments, favouring water-rock interaction. the abundance of colloidal material in the waters which depends on the petrographical composition of the aquifer and on sample treatment, e.g. filtration techniques. Deep groundwater network A second observation network composed of 21 borehole waters has been added for comparison. Compared to the subsurface waters, deep water chemistry might differ. The following reasons seem to be responsible: Long residence times of water in the underground favours slow reaction kinetics. Chemistry of very deep water can be influenced by mixing with older formation waters. Most of the deep boreholes in the Swiss Molasse basin are situated in the OMM and therefore represent a likely homogenous lithology. Borehole waters, because of different aquifer rock types and unknown surface and borehole influences, are least likely to represent typical Molasse water composition.

EPFL1995