IRUS Total

Characterising the microstructure of cement-based materials using laser scanning confocal microscopy

File Description SizeFormat 
Yio-MHN-2017-PhD-Thesis.pdfThesis20.72 MBAdobe PDFView/Open
Title: Characterising the microstructure of cement-based materials using laser scanning confocal microscopy
Authors: Yio, Marcus Heo Nong
Item Type: Thesis or dissertation
Abstract: Three-dimensional (3D) pore characterisation of cement-based materials is essential for understanding the influence of topological pore parameters such as connectivity and tortuosity on transport processes. The main objective of this thesis was to develop laser scanning confocal microscopy (LSCM) for 3D imaging and quantification of pore structure of cement-based materials at submicron resolution. To enable this, a novel approach to reconstruct large volumes of cement-based materials at submicron resolution was developed by combining serial sectioning with LSCM. The method uses a series of Z-stacks with overlapping regions for stitching based on phase correlation. With this method, no information is lost and the spatial resolution is maintained with increase in image size. The effects of axial distortion in LSCM images caused by mismatch of refractive indices between immersion medium and different phases within cement-based materials on various pore attributes were examined. Results indicated that parameters including porosity, specific surface area, percolation connectivity, scalar diffusion tortuosity and formation factor are not significantly affected by axial distortion. A generic correction method was proposed based on measuring the aspect ratio of pulverised fuel ash (PFA) particles in hardened blended pastes. The representative elementary volume for 3D pore characterisation of different cementitious systems was also investigated using a statistical approach. For a given number of realisations, an image volume of 1003 μm3 was found to give comparable porosity to that measured by 2D backscattered electron (BSE) microscopy. BSE signal variation across pore-solid boundaries was simulated using a 3D Monte Carlo technique to enhance image analysis of the pore structure. It was found that a single pore of down to 1 nm can be resolved with field emitters under ideal imaging conditions. The Overflow method was also found to be able to accurately segment pores larger than 1 μm with errors of ~1% and randomly inclined pores with an average error of ~5.2%. Effects of supplementary cementitious materials including silica fume (SF), pulverised fly ash (PFA) and ground granulated blastfurnace slag (GGBS) on the 3D pore structure of cement pastes were investigated using LSCM in conjunction with BSE microscopy. Generally, results from both techniques showed that SF enhances the pore structure (i.e. decreased porosity and percolation connectivity, and increased diffusion tortuosity) from early ages whereas PFA and GGBS show improvements at later ages. The percolation connectivity decreases while diffusion tortuosity increases drastically, as porosity reduces to ~15%. Measured 3D pore characteristics were used as inputs to simple analytical equations for predicting transport properties. Predicted results agreed reasonably well with measured values, mostly within a factor of five. An exploratory study into the application of fluorescence LSCM for real-time imaging of early cement hydration is also presented. Qualitative and quantitative analyses of microstructural developments in different hydrating cementitious systems were made. The advantages and limitations of LSCM for such application are also discussed.
Content Version: Open Access
Issue Date: Dec-2016
Date Awarded: Jun-2017
URI: http://hdl.handle.net/10044/1/55294
DOI: https://doi.org/10.25560/55294
Supervisor: Wong, Hong
Buenfeld, Nick
Department: Civil and Environmental Engineering
Publisher: Imperial College London
Qualification Level: Doctoral
Qualification Name: Doctor of Philosophy (PhD)
Appears in Collections:Civil and Environmental Engineering PhD theses

Unless otherwise indicated, items in Spiral are protected by copyright and are licensed under a Creative Commons Attribution NonCommercial NoDerivatives License.

Creative Commons