Search

The optimum sulfate content of Portland cement is usually determined empirically but the logic underlying selection of the "optimum" is open to challenge. Quantification is difficult until the relations between, on the one hand, paste mineralogy and on the other, amount of sulfate added, are known. Sulfate in hardened paste is known to be contained in pore fluid and three solids; AFm, AFt and C-S-H. Recent studies have characterised AFm and AFt but it is difficult directly to determine sulfate contents of the C-S-H of commercial cements. The objective of the current work is to quantify the sorption potential of C-S-H for sulfate supplied by the phases relevant for Portland cement hydration, i.e., gypsum, AFt and AFm+AFt. The potentially relevant factors such as partial substitution of silicon by alumina in C-S-H (C-A-S-H), effect of changing Ca/Si ratio, temperature, aqueous sulfate concentration, pH and ageing of C-(A)-S-H have been considered when designing the experiments. The solid-liquid sulfate distribution ratios (Rd) were calculated and served as a basis for (i) calculation of thermodynamic parameters of sulfate/C-(A)-S-H interaction as well as for (ii) re-evaluation of earlier models of sulfate distribution in hydrated Portland cement. In this study, synthetic C-S-H and C-A-S-H phases were prepared in the range of bulk Ca/Si ratios from 0.9-1.8 (target Al/Si ratio of C-A-S-H = ~0.04) by the direct reaction method and the modified sol-gel method respectively. The distribution of sulfate between C-(A)-S-H, aqueous phase and the relevant sulfate-containing phases was determined adopting the two experimental protocols: known amounts of C-(A)-S-H have been contacted either (i) with various amounts of gypsum-undersaturated aqueous solutions, measuring the sorption indirectly by the change of sulfate concentration in solution or (ii) with an excess of solid gypsum, AFt or AFm+AFt in an aqueous environment, measuring the sorption by direct quantitative analysis of C-(A)-S-H. These two methods allowed the sorption study to be performed for a range of aqueous sulfate concentrations from about 1-14 mmol/l. The procedures were repeated with C-S-H and C-(A)-S-H of each Ca/Si ratio at 5, 20, 55 and 85°C. The effect of ageing as well as the effect of presence of alkali (0.1M NaOH) on resulting sulfate affinity of C-(A)-S-H was tested on selected samples where practical, desorption was also measured. The sulfate concentrations were measured by HPIC. The sulfate sorption on C-S-H was found to be increasing with increasing aqueous sulfate concentration, Ca/Si ratio and temperature but decreasing with increasing aqueous alkali concentration. Nearly complete reversibility of sulfate sorbed on C-S-H in the range of temperatures from 5-85°C was subsequently demonstrated at 20°C suggesting, that if sufficient water is available, sulfate in C-S-H in hydrated Portland cement is potentially mobile. Ageing of C-S-H for up to 10 years at 20°C had only a negligible effect on resulting sulfate affinity. The data obtained on both C-S-H and C-A-S-H showed that Rd is independent of sulfate concentration of solution but varies systematically with temperature of sorption, Ca/Si ratio, pH and, in the case of C-A-S-H, with previous thermal treatment. The sulfate affinity of C-A-S-H was found to be decreasing as a result of (i) ageing process as well as (ii) with increasing temperature of synthesis. Fresh C-A-S-H gels exhibited higher sulfate uptake than aged C-A-S-H gels, almost certainly due to formation of sulfoaluminate phases. Alumina, when introduced into C-S-H, reacts over a period of several months before reaching a steady state at 20°C or several days at 55°C. Before reaching a steady state, alumina is, at least partially, distributed between C-A-S-H and a secondary alumina-rich phase, possibly OH-AFm. This phase, if contacted with sulfate from gypsum, forms sulfoaluminate phases, possibly SO4-AFm and/or AFt. As a result, the affinity of fresh inhomogeneous C-A-S-H for sulfate is a sum of two parallel processes: sorption and precipitation. On the other hand, homogeneous C-A-S-H gels exhibited much lower sulfate affinity but there are uncertainties about the nature of interaction with sulfate ions. The sulfate uptake by C-A-S-H was found to be increasing or decreasing with increasing temperature, depending on temperature applied during the ageing process. Moreover, the reversibility of sulfate uptake process is also dependent on previous thermal treatment. The sulfate sorbed on C-A-S-H gel at 5 and 20°C is subsequently only partially desorbable at 20°C but the sulfate sorbed on C-A-S-H at 55°C is almost completely desorbable at 20°C. Evaluated together, further research is necessary to elucidate the nature of C-A-S-H and to explore the possible formation of sulfoaluminate phases. However, in absolute numbers, the affinity of homogeneous C-A-S-H for sulfate was found to be similar to that of C-S-H of similar Ca/Si ratio. The data presented in this Thesis were utilized for calculation of quantitative distribution of sulfate between relevant phases in hydrated Portland cement of a typical composition. In general, the C-S-H phase in hydrated Portland cement sorbs substantially less sulfate than previously reported.

Dolomite is shown to precipitate in laboratory experiments that simulate the microbiogeochemical conditions prevailing during the later stages of evaporation in the ephemeral, hypersaline, dolomitic lakes of the Coorong region, South Australia, where microbially-dominated ecosystems support intensive bacterial sulphate reduction. Analyses of numerous lake- and pore-water samples from Coorong lakes document rapid geochemical changes with depth and time. Extremely high sulphate and magnesium ion concentrations occur in lake waters and decline rapidly with depth in pore waters through the sulphate reduction zone. Ultimately sulphate ions are totally consumed but magnesium ions are replenished, presumably from desiccated cyanobacterial sheaths. Carbonate concentrations in pore waters reach levels up to 100 times those of normal seawater. Most-probable-number counts show that large populations of sulphate-reducing bacteria are present in sediment cores, while sulphur isotope analysis of residual lakewater sulphate indicates that microbial fractionation takes place in all the study lakes. Microbes from the lakes were cultured in the laboratory under anoxic conditions and viable populations were injected into vials containing a sterilised granular substrate immersed in a simulated lakewater solution. Falls in the levels of sulphate ion concentration and rising pH in selected vials were interpreted as due to active bacterial sulphate reduction accompanied by increased concentrations of carbonate. After 1 month, subspherical nano-grains of dolomite were precipitated. This study proposes that bacterial sulphate reduction overcomes kinetic constraints on dolomite formation by removing the sulphate ions and releasing magnesium and calcium ions from neutral ion pairs, and by generating elevated carbonate concentrations in a hypersaline and strongly electrolytic solution. Microbiogeochemical and isotopic data therefore demonstrates that bacterial sulphate reduction controls dolomite precipitation in both the laboratory experiments and the lake sediments. It is proposed that dolomite formation through bacterial sulphate reduction provides a process analogue that is applicable to thick platformal dolostones of the past, where benthic microbial communities were the dominant colonisers of the shallow marine environment.

The accumulation of undesired matter at heat transfer surfaces (fouling) is a severe problem to industry. The growth of calcium sulphate dihydrate on heated metal surfaces has been examined and a mechanism for the effect of surface roughness on the amount of deposition has been proposed. A novel piece of equipment was designed and constructed in which the growth of crystals on heat transfer surfaces with different surface roughnesses could be observed under controlled solution conditions. The test section was a transparent rectangular channel into Rhich three 25 mm diameter test pieces could be inserted such that the polished faces were planar with the rear face of the channel. The back faces of these test pieces were heated by contact with hot Mater. Using Reynolds numbers of 300 to 13,500 and calcium sulphate 2+ solutions with bulk concentrations from 20 to 50 mH Ca , no gross effect due to surface roughness was seen. However a limited effect, which distinguished grit-blasted surfaces from polished surfaces, was 2+ found in experiments with a bulk concentration from 28 to 33 mH Ca . In all of the experiments it was observed that the presence of bubbles enhanced crystal growth. It was also found that the amount of deposition formed on any surface decreased with decreasing dissolved oxygen content of the bulk solution. It is suggested that a bulk concentration of approximately 33 mH 2+ Ca is a critical level of supersaturation, which corresponds with the so-called metastable limit of supersaturation. The surface roughness effect may be associated with two factors. Firstly, as the critical supersaturation is approached crystal growth is enhanced at certain sites. in particular the edges of bubbles. Secondly, very rough surfaces. such as grit-blasted surfaces. more readily support and initiate bubble formation and consequently the grit-blasted surface shows greater growth. Additional work with a different test rig. using a stagnant solution, indicated that suppression of bubble formation during an experiment, either by initially degassing the surface or by pressurising the system, decreased the amount of crystal growth. This work suggests that a surface roughness effect exists which is related to the presence of bubbles at the surface. Consequently the amount of deposition can be decreased either by controlling the formation of the bubbles (e.g. by completely degassing both the solution and the surface. and pressurising the system). or by polishing the surface. However. the surface roughness effect was most apparent between the very coarse, grit-blasted surfaces (R ~ 2.90 um) and the a medium finish 180 or 240 Grit surfaces (R ~ 0.23 um). a Little or no further improvement was obtained in polishing to a mirror finish. i. e. 0.25 or 0.1 um (R ~ 0.03 um). a further polishing beyond removing gross roughness would not appear to Therefore. for industrial purposes, yield sufficient benefit to justify the large cost involved. It is recommended that further work be undertaken to determine whether the deposition which occurs in the absence of bubbles is influenced by the surface roughness, and whether other chemical system and surfaces behave similar to the calcium sulphate - stainless steel system studied in this work. Moreover, additional work should be undertaken to examine the effect of surface physical properties on crystal and bubble nucleation and growth.