Pore Size Distributions in Hardened Cement Paste by Sem Image Analysis

  • PDF / 2,077,664 Bytes
  • 10 Pages / 414.72 x 648 pts Page_size
  • 4 Downloads / 168 Views

DOWNLOAD

REPORT


217 Mat. Res. Soc. Symp. Proc. Vol. 370 ©1995 Materials Research Society

Separate portions of the same specimens were oven dried and subjected to MIP analysis using a Micromeritics Autopore 9220 porosimeter with a maximum pressuring capacity of 414 MPa (60,000 psi). A contact angle of 1300 was assumed. SPECIMENS PREPARED The cement paste specimens were all prepared from a single Type I portland cement of normal chemical characteristics. The pastes were mixed with de-aired water using an evacuated chamber to prevent the inclusion of air voids, at w:c ratios of 0.40 and 0.25. After demolding at 24 hours, separate portions of each paste were hydrated in saturated calcium hydroxide solution for an additional 6 days or an additional 27 days, thus providing specimens 1, 7, and 28 days old. At the conclusion of the designated period of hydration the pastes were oven dried at 1050 C until constant weight was attained. Specimens for image analysis were prepared by cutting a thin slice from the middle third of each paste specimen as cast, embedding it with a low viscosity epoxy resin while under evacuation, and curing at 70' C The surface to be examined was ground and polished using successively finer diamond paste grades, with the final polish accomplished using 1 pmrdiameter diamond paste. The prepared surfaces were then sputter coated with palladium to prevent charging in the SEM. Companion specimens for MIP were broken off also from the middle third of each paste specimen, in such a way that all surfaces of the fragments to be intruded were fracture surfaces. DEVELOPMENT OF PROCEDURES FOR IMAGE ANALYSIS DISTRIBUTION MEASUREMENTS

PORE SIZE

A number of considerations needed to be addressed in developing an appropriate method for measuring pore size distributions by image analysis. These involved such factors as obtaining an appropriate binary segmentation of the BEI to include the pores and exclude the solids, setting the lower limit of the pores that can be tallied, rigorously defining what is meant by "size" of an irregular pore, and deciding on the method of conversion of two dimensional image data to an acceptable approximation of three dimensional reality. Finally, an appropriate way of assessing the three dimensional size distribution so as to permit direct comparison with MIP results needed to be developed. Binary Segmentation The gray scale histogram of BEI images normally obtained with cement paste do not show a clearly separate peak for the dark (epoxy-filled) pores. Thus the specific gray level used for the binary segmentation of the pores from the solid features must be arbitrary. In this work the gray level was selected for each image by the operator after repeated observations of the results of trials at various candidate threshold levels. A single operator, the second-named author, performed all of the segmentations, thus avoiding any operator bias. Diligence and extreme care to maintain consistency between images greatly reduced quantitative variation. Size Limits The capillary pores in hcp are seen in