Field reflectance spectrometry for detection of swelling clays at construction sites

Citation
Afh. Goetz et al., Field reflectance spectrometry for detection of swelling clays at construction sites, FIELD A C T, 5(3), 2001, pp. 143-155
Citations number
16
Categorie Soggetti
Environment/Ecology,"Spectroscopy /Instrumentation/Analytical Sciences","Instrumentation & Measurement
Journal title
FIELD ANALYTICAL CHEMISTRY AND TECHNOLOGY
ISSN journal
1086900X → ACNP
Volume
5
Issue
3
Year of publication
2001
Pages
143 - 155
Database
ISI
SICI code
1086-900X(2001)5:3<143:FRSFDO>2.0.ZU;2-B
Abstract
We have developed techniques to map the distribution and composition of cla y-rich soils with portable field spectrometers on the ground. Spectral refl ectance measurements in the 1800-2400-nm region with 10-nm resolution can d istinguish smectites, which cause swelling, from illites and kaolinite that do not swell significantly. Illites and smectites are often mixed in the s oil and result in varying swell potential. Standard engineering soil tests are too time consuming and costly to be used in areas where there is highly varying swell potential. Therefore, in many instances in regions of steepl y dipping sedimentary layers, beds of swelling clay go undetected. We show that it is possible to determine smectite content with a standard cross-val idation error of 10% based on partial least-squares analysis of second-deri vative reflectance spectra. Loadings show that the 1800-2000- and 2150-2250 -nm regions contain the most relevant information for the detection and qua ntification of smectite content and these correspond to the bound water in the clay lattice interlayer and the Al-OH combination band, respectively. C orrelations as high as 87% were obtained with the Seed swell-potential inde x. The correlations are higher when the samples were dried rather than meas ured in their moist condition shortly after collection in the field, as in an earlier study. Correlation with other swell indices shows that reflectan ce spectroscopy would be a reliable indicator that could divide samples int o low, medium, and high swell potential. The effect of sample moisture was studied and the results show that the surface of the field samples must fir st be dried before measurement, in order to obtain a reliable swell potenti al value from the model. (C) 2001 John Wiley & Sons, Inc.