Experimental Characterization and Quantification of Cement-Bentonite Interaction using Core Infiltration Techniques Coupled with X-ray Tomography

Deep geological storage of radioactive waste foresees cementitious materials as reinforcement of tunnels and as backfill. Bentonite is proposed to enclose spent fuel canisters and as drift seals. Sand/bentonite (s/b) is foreseen as backfill material of access galleries or as drift seals. The emplacement of cementitious material next to clay material generates an enormous chemical gradient in pore-water composition that drives diffusive solute transport. Laboratory studies and reactive transport modeling predicted significant mineral alteration at and near interfaces, mainly resulting in a decrease of porosity in bentonite. The goal of this thesis was to characterize and quantify the cement/bentonite interactions both spatially and temporally in laboratory experiments. A newly developed mobile X-ray transparent core infiltration device was used to perform X-ray computed tomography (CT) scans without interruption of running experiments. CT scans allowed tracking the evolution of the reaction plume and changes in core volume/diameter/density during the experiments. In total 4 core infiltration experiments were carried out for this study with the compacted and saturated cores consisting of MX-80 bentonite and sand/MX-80 bentonite mixture (s/b; 65/35%). Two different high-pH cementitious pore-fluids were infiltrated: a young (early) ordinary Portland cement pore-fluid (APWOPC; K+–Na+–OH-; pH 13.4; ionic strength 0.28 mol/kg) and a young ‘low-pH’ ESDRED shotcrete pore-fluid (APWESDRED; Ca2+–Na+–K+–formate; pH 11.4; ionic strength 0.11 mol/kg). The experiments lasted between 1 and 2 years. In both bentonite experiments, the hydraulic conductivity was strongly reduced after switching to high-pH fluids, changing eventually from an advective to a diffusion-dominated transport regime. The reduction was mainly induced by mineral precipitation and possibly partly also by high ionic strength pore-fluids. Both bentonite cores showed a volume reduction and a resulting transient flow in which pore-water was squeezed out during high-pH infiltration. The outflow chemistry was characterized by a high ionic strength, while chloride in the initial pore water got replaced as main anionic charge carrier by sulfate, originating from gypsum dissolution. The chemistry of the high-pH fluids got strongly buffered by the bentonite, consuming hydroxide and in case of APWESDRED also formate. Hydroxide got consumed by mineral reactions (saponite and possibly talc and brucite precipitation), while formate being affected by bacterial degradation. Post-mortem analysis showed reaction zones near the inlet of the bentonite core, characterized by calcium and magnesium enrichment, consisting predominately of calcite and saponite, respectively. Silica got enriched in the outflow, indicating dissolution of silicate-minerals, identified as preferentially cristobalite. In s/b, infiltration of APWOPC reduced the hydraulic conductivity strongly, while APWESDRED infiltration had no effect. The reduction was mainly induced by mineral precipitation and probably partly also by high ionic strength pore-fluids. Not clear is why the observed mineral precipitates in the APWESDRED experiment had no effect on the fluid flow. Both s/b cores showed a volume expansion along with decreasing ionic strengths of the outflow, due to mineral reactions or in case of APWESDRED infiltration also mediated by microbiological activity, consuming hydroxide and formate, respectively. The chemistry of the high-pH fluids got strongly buffered by the s/b. In the case of APWESDRED infiltration, formate reached the outflow only for a short time, followed by enrichment in acetate, indicating most likely biological activity. This was in agreement to post-mortem analysis of the core, observing black spots on the inflow surface, while the sample had a rotten-egg smell indicative of some sulfate reduction. Post-mortem analysis showed further in both cores a Ca-enrichment in the first 10 mm of the core due to calcite precipitation. Mg-enrichment was only observed in the APWOPC experiment, originating from newly formed saponite. Silica got enriched in the outflow of both experiments, indicating dissolution of silicate-minerals, identified in the OPC experiment as cristobalite. The experiments attested an effective buffering capacity for bentonite and s/b, a progressing coupled hydraulic-chemical sealing process and also the preservation of the physical integrity of the interface region in this setup with a total pressure boundary condition on the core sample. No complete pore-clogging was observed but the hydraulic conductivity got rather strongly reduced in 3 experiments, explained by clogging of the intergranular porosity (macroporosity). Such a drop in hydraulic conductivity may impact the saturation time of the buffer in a nuclear waste repository, although the processes and geometry will be more complex in repository situation.

Encapsulation of caesium-loaded Ionsiv in cement

The microporous material Ionsiv is used for 137Cs removal from aqueous nuclear waste streams. In the UK, Cs-loaded Ionsiv is classed as an intermediate-level waste; no sentencing and disposal route is yet defined for this material and it is currently held in safe interim storage on several nuclear sites. In this study, the suitability of fly ash and blast furnace slag blended cements for encapsulation of Cs-Ionsiv in a monolithic wasteform was investigated. No evidence of reaction or dissolution of the Cs-Ionsiv in the cementitious environment was found by scanning electron microscopy and X-ray diffraction. However, a small fraction (<= 1.6 wt.%) of the Cs inventory was released from the encapsulated Ionsiv during leaching experiments carried out on hydrated samples. Furthermore, it was evident that K and Na present in the cementitious pore water exchanged with Cs and H in the Ionsiv. Therefore, cement systems lower in K and Na, such as slag based cements, showed lower Cs release than the fly ash based cements.

Quantification of water content across a cement-clay interface using high resolution neutron radiography

In many designs for radioactive waste repositories, cement and clay will come into direct contact. The geochemical contrast between cement and clay will lead to mass fluxes across the interface, which consequently results in alteration of structural and transport properties of both materials that may affect the performance of the multi-barrier system. We present an experimental approach to study cement-clay interactions with a cell to accommodate small samples of cement and clay. The cell design allows both in situ measurement of water content across the sample using neutron radiography and measurement of transport parameters using through-diffusion tracer experiments. The aim of the high- resolution neutron radiography experiments was to monitor changes in water content (porosity) and their spatial extent. Neutron radiographs of several evolving cement-clay interfaces delivered quantitative data which allow resolving local water contents within the sample domain. In the present work we explored the uncertainties of the derived water contents with regard to various input parameters and with regard to the applied image correction procedures. Temporal variation of measurement conditions created absolute uncertainty of the water content in the order of ±0.1 (m3/m3), which could not be fully accounted for by correction procedures. Smaller relative changes in water content between two images can be derived by specific calibrations to two sample regions with different, invariant water contents.

Reduction of cement leakage by sequential PMMA application in a vertebroplasty model

PURPOSE Leakage is the most common complication of percutaneous cement augmentation of the spine. The viscosity of the polymethylmethacrylate (PMMA) cement is strongly correlated with the likelihood of cement leakage. We hypothesized that cement leakage can be reduced by sequential cement injection in a vertebroplasty model. METHODS A standardized vertebral body substitute model, consisting of aluminum oxide foams coated by acrylic cement with a preformed leakage path, simulating a ventral vein, was developed. Three injection techniques of 6 ml PMMA were assessed: injection in one single step (all-in-one), injection of 1 ml at the first and 5 ml at the second step with 1 min latency in-between (two-step), and sequential injection of 0.5 ml with 1-min latency between the sequences (sequential). Standard PMMA vertebroplasty cement was used; each injection type was tested on ten vertebral body substitute models with two possible leakage paths per model. Leakage was assessed by radiographs using a zonal graduation: intraspongious = no leakage and extracortical = leakage. RESULTS The leakage rate was significantly lower in the "sequential" technique (2/20 leakages) followed by "two-step" (15/20) and "all-in-one" (20/20) techniques (p < 0.001). The RR for a cement leakage was 10.0 times higher in the "all-in-one" compared to the "sequential" group (95 % confidence intervals 2.7-37.2; p < 0.001). CONCLUSIONS The sequential cement injection is a simple approach to minimize the risk for leakage. Taking advantage of the temperature gradient between body and room temperature, it is possible to increase the cement viscosity inside the vertebra while keeping it low in the syringe. Using sequential injection of small cement volumes, further leakage paths are blocked before further injection of the low-viscosity cement.

Lavage prior to vertebral augmentation reduces the risk for cement leakage

PURPOSE This study aimed at assessing the cement leakage rate and the filling pattern in patients treated with vertebroplasty, kyphoplasty and stentoplasty with and without a newly developed lavage technique. STUDY DESIGN Retrospective clinical case-control study. METHODS A newly developed bipedicular lavage technique prior to cement application was applied in 64 patients (45.1 %) with 116 vertebrae, ("lavage" group). A conventional bipedicular cement injection technique was used in 78 patients (54.9 %) with 99 levels ("controls"). The outcome measures were filling patterns and leakage rates. RESULTS The overall leakage rate (venous, cortical defect, intradiscal) was 37.9 % in the lavage and 83.8 % in the control group (p < 0.001). Venous leakage (lavage 12.9 % vs. controls 31.3 %; p = 0.001) and cortical defect leakage (lavage 17.2 % vs. controls 63.3 %; p < 0.001) were significantly lower in the lavage group compared to "controls," whereas intradiscal leakages were similar in both groups (lavage 12.1 % vs. controls 15.2 %; p = 0.51). For venous leakage multivariate logistic regression analysis showed lavage to be the only independent predictor. Lavage was associated with 0.33-times (95 % CI 0.16-0.65; p = 0.001) lower likelihood for leakage in compared to controls. CONCLUSIONS Vertebral body lavage prior to cement augmentation is a safe technique to reduce cement leakage in a clinical setting and has the potential to prevent pulmonary fat embolism. Moreover, a better filling pattern can be achieved.

MINERALOCORTICOID REGULATION OF ION CONDUCTIVE PATHWAYS OF THE CORTICAL COLLECTING DUCT

Mineralocorticoids (DOCA) are known to increase Na('+) absorption and K('+) secretion in the rabbit cortical collecting duct (CCD). However, the mechanism of regulation of the apical and basolateral cell membranes and tight junction ion conductive pathways (G('a), G('b), and G('tj), respectively) by mineralocorticoids are only partially understood. Using electrophysiological techniques and microelectrodes it was demonstrated that the apical cell membrane contained a dominant Ba('2+) sensitive K('+) conductive pathway, G(,K)('a), and an amiloride sensitive Na('+) conductive pathway, G(,Na)('a). The basolateral membrane contained a dominant Cl('-) conductive pathway, G(,Cl)('b), and a significant Ba('2+) sensitive K('+) conductive pathway, G(,K)('b). Upon elevating the mineralocorticoid levels of rabbits with intact adrenal glands it was found that V('te) was significantly increased after 1 day with a further increase after 13-16 days. These results indicated both primary and secondary effects of mineralocorticoid elevation. After 1 day of DOCA treatment, G(,Na)('a), I(,Na)('a) and I(,K)('a) increased by more than 2-fold and were maintained at high levels after 13-16 days of DOCA treatment. Secondary (chronic) effects of mineralocorticoids were evident after 4 days or more of DOCA treatment. These included a significant increase in G(,K)('a) from 4.0 to 10.2 mS.cm('-2) and a hyperpolarization of V('b) by -20 mV after 4 days of treatment. After 13-16 days of DOCA treatment V('b) remained hyperpolarized at -98.1 mV and G('tj) decreased from 5.6 to 4.2 mS.cm('-2). The hyperpolarization of V('b) was due to an increase in electrogenic Na('+) pump activity as the pump current, I(,act)('b), increased significantly from 35.7 to 195.2 (mu)A.cm('-2). Whereas net passive K('+) current across the basolateral membrane, I(,K)('b), was near zero in the control group of animals, i.e., K('+) near equilibrium, I(,K)('b) was approximately -40 (mu)A.cm('-2) in chronic DOCA treated animals. These results demonstrate that the initial effect of mineralocorticoid elevation is to increase G(,Na)('a). The ensuing depolarization of the apical membrane increases the driving force for K('+) exit into the lumen. Between 1 and 4 days of elevation, G(,K)('a) more than doubles in magnitude and at the same time the electrogenic activity of the Na('+) pump increases. This results in a hyperpolarization of V('b) which increases the driving force for K('+) uptake from the bath to the cell through a basolateral membrane conductive pathway. After 13-16 days G('tj) decreases thereby serving to maintain high electrochemical gradients across the epithelium. Therefore, the long term effects of mineralocorticoid elevation on the CCD appear to be adaptive mechanisms that serve to maintain high levels of K('+) secretion and Na('+) absorption. ^

Optimizing Performance of Conductive Nanoparticle-filled Membranes

Conductive membranes were prepared by magnetic alignment of graphite-coated iron nanoparticles (GCINs) in a polyisobutylene (PIB) matrix, which was cast onto an interdigitated surface electrode. Toluene and tetrahydrofuran (THF) were used as solvents for the casting solution. Different molecular weights of PIB and solutions with concentrations ranging from 0.05- to 50-wt% were explored to optimize the mechanical and physical properties of the membrane. The amount of GCINs used in the membranes ranged from 0.1- to 2.5-wt%, and a sonicator was used to disperse the particles in the membrane. Sedimentation and surfactant studies were conducted to investigate the dispersion of GCINs in solutions. Progresses were made to find an optimal combination of various parameters in order to attain thin homogenous membranes for fast response.

(Table 2) Oxygen, carbon, and strontium isotopic composition of carbonate-brucite material from the Lost City hydrothermal field

The isotopic (dD, d18O, d13C, and 87Sr/86Sr) and geochemical characteristics of hydrothermal solutions from the Mid-Atlantic Ridge and the material of brucite-carbonate chimneys at the Lost City hydrothermal field at 30°N, MAR, were examined to assay the role of the major factors controlling the genesis of the fluid and hydrothermal chimneys of the Lost City field. The values of dD and d18O in fluid samples indicates that solutions at the Lost City field were produced during the serpentinization of basement ultramafic rocks at temperatures higher than 200°C and at relatively low fluid/rock ratios (<1). The active role of serpentinization processes in the genesis of the Lost City fluid also follows from the results of the electron-microscopic studying of the material of hydrothermal chimneys at this field. The isotopic (d18O, d13C, and 87Sr/86Sr) and geochemical (Sr/Ca and REE) signatures indicate that, before its submarine discharging at the Lost City field, the fluid filtered through already cold altered outer zones of the Atlantis Massif and cooled via conductive heat loss. During this stage, the fluid could partly dissolve previously deposited carbonates in veins cutting serpentinite at the upper levels of the Atlantis Massif and the carbonate cement of sedimentary breccias underlying the hydrothermal chimneys. Because of this, the age of modern hydrothermal activity at the Lost City field can be much younger than 25 ka.