Discussion of “Measurements of suction versus water content for bentonite–sand mixtures”Appears in Can. Geotech. J. 47(5): 583–594 (2010)

Abstract

The authors have presented a very interesting research paper on the comparison of various techniques for suction measurements of a compacted bentonite–sand mixture. One of the important techniques discussed in the paper is the filter paper method, and the authors used the filter paper calibration equations presented by Agus (2005). Different aspects of using the calibration equations proposed by the authors are examined and discussed here in light of previously published data and our laboratory test results. According to Agus (2005), the calibration curves for the Whatman No. 42 filter paper were obtained using the pressure plate technique for suctions less than 1500 kPa and using a desiccator (vapour equilibrium technique) for suctions higher than 2000 kPa. To determine what the authors called the ‘‘matric suction calibration curve’’ (eq. [2]), a filter paper was been placed directly over a ceramic disc in a pressure plate during matric suction equalization, and a dead weight was placed on top of the filter paper to maintain good contact between the filter paper and the ceramic disc. An equilibration period of 3 weeks was adopted for this calibration test. The noncontact filter paper calibration data (eq. [1]) was obtained using different concentrations of NaCl solutions, and a 5 week equilibration time was adopted for this technique. The authors performed the calibration test for two different sizes of filter paper: a standard-sized (55 mm diameter) and a small-sized (25 mm diameter) filter paper. Figure D1 presents the calibration curves for the Whatman No. 42 filter paper obtained by Fawcett and CollisGeorge (1967) at 22 8C, Chandler and Gutierrez (1986) at 21 8C, Chandler et al. (1992) at 21 8C, Leong et al. (2002) at 25 8C, the proposed calibration curves by Agus (2005) at 22 8C, and also the writers’ experimental data at 25 8C. The difference in the filter paper calibration curves may be due to factors such as equilibration time, temperature, and the use of different batches of filter paper as indicated by Marinho and Oliveira (2005). Using eq. 1 proposed by Agus (2005) for suctions less than 2000 kPa can cause significantly misleading results as the equation is based on experimental data by Agus (2005) for suctions greater than 4000 kPa. The proposed calibration curve for suctions less than 2000 kPa strongly disagrees with the calibration curves found in literature and also shows a very small sensitivity of suction in relation to filter paper water content (see Fig. D1). Although the authors reported the use of the proposed calibration curve only for suctions greater than 2000 kPa, it seems the proposed calibration curve has been used for total suction measurements lower than 2000 kPa, as data presented on Figs. 4a, 4b, and 5a (in the paper under discussion), which is not applicable. The writers obtained the filter paper (Whatman 42) calibration curve using two procedures. One was via vapour equilibrium using sodium chloride solutions at different concentrations, for suctions higher than 300 kPa.3 The calibration was performed in a temperature-controlled chamber, which maintained the desired temperature to ±0.1 8C. The other procedure used the pressure plate with the filter paper to be calibrated placed between two soil samples. The filter paper was protected by two larger-size filter papers. The arrangement of the soil samples and the filter papers was subsequently placed on a fully saturated high air-entry value ceramic disc in the pressure plate. An equilibration period of 2 weeks was adopted for both calibration procedures. The results obtained from the two procedures are shown in Fig. D1. From the results it can be observed that the data from the vapour equilibrium method for inducing suction agree with the calibration equation proposed by the authors only for suctions higher than approximately 5000 kPa. The data presented by the writers agree well with the data presented by the authors, when compared with the calibration curve represented by eq. [2], regardless of the method used for generating the suction (i.e., vapour flow or capillary flow). Therefore it can be concluded that a 2 week equilibration time is sufficient for a capillary flow of any level and for a vapour flow higher than 500 kPa. From Fig. D1 it can be observed that the last two points obtained using the vapour flow method (suction lower than 500 kPa) require a longer time to reach equilibrium. This is in agreement with Received 24 September 2010. Accepted 26 November 2010. Published on the NRC Research Press Web site at cgj.nrc.ca on 1 February 2011.