Computational investigation on mass diffusivity in Portland cement paste based on X-ray computed microtomography (lCT) image Mingzhong Zhang a,⇑ , Yongjia He b , Guang Ye a , David A. Lange c , Klaas van Breugel a a Microlab, Faculty of Civil Engineering and Geosciences, Delft University of Technology, Stevinweg 1, 2628 CN Delft, The Netherlands b Key Laboratory for Silicate Materials Science and Engineering, Wuhan University of Technology, Wuhan 430070, China c 2129 Newmark Civil Engineering Laboratory, University of Illinois, 205 N. Mathews, Urbana, IL 61801-2352, USA article info Article history: Received 17 February 2011 Received in revised form 27 June 2011 Accepted 18 July 2011 Available online 23 August 2011 Keywords: X-ray computed microtomography Transport property Cement paste Microstructure Percolation abstract The mass diffusivity in Portland cement paste plays an important role in the durability design and assess- ment of cement-based materials. This paper presents a link between mass diffusivity in cement paste and its microstructure. X-ray computed microtomography (lCT) was applied to derive the three-dimensional (3D) images of cement paste specimens with water-to-cement (w/c) ratio 0.50 at curing ages of 1, 3, 7, 28 and 120 days at a resolution of 0.485 lm/voxel. By choosing the image threshold value based on the gray level histogram, the phases, i.e., capillary pores, hydration products and unhydrated cement grains in the microtomography images of each specimen were segmented and the 3D microstructure and pore struc- ture were obtained. The degree of pore connectivity and percolation of each specimen were analyzed in detail on the basis of cluster-labeling algorithm. In addition, the finite element method (FEM) was applied to simulate the diffusion process of tritiated water through the extracted microstructure and quantify the diffusivity of tritiated water by associating with Fick’s law. The simulated diffusivity was compared with the measured value and seemed consistent with the experimental investigation. The results suggest that X-ray lCT is a reliable non-destructive technique and suitable tool to investigate the 3D microstructure. The obtained microstructure can be considered as an input to predict the transport properties of cement- based materials. Ó 2011 Elsevier Ltd. All rights reserved. 1. Introduction The service life prediction and durability design of cement- based materials and structures has been an important topic of re- search for many years. Most of durability research focuses on the diffusion behavior of fluids and potentially harmful ionic species (e.g., chloride, sulfates, etc.) in cement-based materials due to its significant influence on the process of leaching, corrosion and car- bonation of cement constituents [1]. Diffusive transport of solutes in cement-based materials is intimately related to the microstruc- ture of cement paste, especially the pore structure characteristics. It is of importance to take into account the three-dimensional (3D) complex microstructure of cement paste in the estimation of diffu- sion coefficient (diffusivity) of solutes in cement-based materials. In order to characterize the microstructure, a variety of techniques including computer-aided simulation and experimental tests are proposed and developed in the past decades. With regard to experimental techniques, various approaches such as optical microscope, scanning electron microscope (SEM) and mercury intrusion porosimetry (MIP) are widely applied. Optical microscope and SEM can be used to observe the topogra- phy and distribution of phase components of cement paste. How- ever, the acquired images just reflect the information on one of the cross-sections. The spatial distribution of the solid phases and pores still cannot be acquired. Their differences in content, topography and distribution varying with time cannot be traced. Furthermore, the sample preparation processes in these methods, such as polishing, drying may damage the microstructure of the samples before or during the testing. Therefore, what the research- ers watched may not be the real structure of the materials. Although the environmental scanning electron microscope (ESEM) method can in some extent realize the continuous and non- destructive observation of the materials, but the strict testing con- ditions and the drawback of lacking the ability of acquiring 3D information also restrict its application [2]. MIP is another impor- tant approach for researching the pore structure, of which the mea- suring range is about 1 nm to 1000 lm. But there are also inevitable drawbacks existing in this testing method. For example, in the analysis process of this method, Washburn model is used to transfer the acquired data to the characteristics of the pores, but the model assumes the pores are all cylinder-like, which is not in agreement with the actual fact. Another drawback of MIP is that the mercury is not easy to enter the ink-bottle like pores with 0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.conbuildmat.2011.07.017 ⇑ Corresponding author. E-mail address: m.zhang@tudelft.nl (M. Zhang). Construction and Building Materials 27 (2012) 472–481 Contents lists available at SciVerse ScienceDirect Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat