Environ. Sci. Technol. 2006, 40, 5930-5936
Noninvasive Quantitative Measurement of Colloid Transport in Mesoscale Porous Media Using Time Lapse Fluorescence Imaging JONATHAN W. BRIDGE,* STEVEN A. BANWART, AND A. LOUISE HEATHWAITE† Groundwater Protection and Restoration Group, Department of Civil and Structural Engineering, Kroto Research Institute, North Campus, University of Sheffield, Broad Lane, Sheffield, S3 7HQ, U.K.
We demonstrate noninvasive quantitative imaging of colloid and solute transport at millimeter to decimeter (meso-) scale. Ultraviolet (UV) excited fluorescent solute and colloid tracers were independently measured simultaneously during co-advection through saturated quartz sand. Pulse-input experiments were conducted at constant flow rates and ionic strengths 10-3, 10-2 and 10-1 M NaCl. Tracers were 1.9 µm carboxylate latex microspheres and disodium fluorescein. Spatial moments analysis was used to quantify relative changes in mass distribution of the colloid and solute tracers over time. The solute advected through the sand at a constant velocity proportional to flow rate and was described well by a conservative transport model (CXTFIT). In unfavorable deposition conditions increasing ionic strength produced significant reduction in colloid center of mass transport velocity over time. Velocity trends correlated with the increasing fraction of colloid mass retained along the flowpath. Attachment efficiencies (defined by colloid filtration theory) calculated from nondestructive retained mass data were 0.013 ( 0.03, 0.09 ( 0.02, and 0.22 ( 0.05 at 10-3, 10-2, and 10-1 M ionic strength, respectively, which compared well with previously published data from breakthrough curves and destructive sampling. Mesoscale imaging of colloid mass dynamics can quantify key deposition and transport parameters based on noninvasive, nondestructive, spatially high-resolution data.
Introduction and Background Colloids in soil and groundwater act as vectors for contaminant transport or as contaminants themselves (1). Colloid transport and filtration in homogeneous porous media is well understood (2, 3). Research on colloid movement in porous media is increasingly focused on colloid-media interactions in dynamic or heterogeneous systems (4); for example, hydraulically unsaturated/draining systems (5), bacteria and biofilms (6, 7), or complex porous media containing fractures, layers, or other permeability discontinuities (8). Breakthrough curves and invasive or destructive * Corresponding author e-mail:
[email protected]; tel: +44 (0)114 222 5785; fax: +44 (0)114 222 5701. † Director, Centre for Sustainable Water Management, Lancaster Environment Centre, Lancaster University, LA1 4YQ. 5930
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 40, NO. 19, 2006
sampling of closed column experiments integrate colloidmedia interactions over time or space, losing information relating to these complex environments. Direct imaging (visualization) of tracers within porous media at pore and mesoscale (millimeter to meter) enables high spatio-temporal resolution measurement of reactive transport and hence is the subject of considerable current research (e.g., 9). Pore-scale studies using etched glass micropore networks (10-12) or small flow cells containing translucent glass beads or quartz sand (13-16) are now well developed using light microscopy techniques. Visualization of colloids and solutes at the mesoscale provides continuous measurement over millimeter to decimeter resolutions in near-real time. It thus represents a key link between porescale visualization and the low-resolution results of “black box” column and field studies. Mesoscale visualization techniques have been used to observe NAPL migration (17), saline infiltration (18), reactive solute transport and biodegradation (19-21), and biofilm growth under saturated and unsaturated conditions (2224). To date, very few published studies have observed colloid or bacterial transport phenomena at the mesoscale (25-27). The most common visualization method is light transmission (28-30). A translucent porous medium is placed between a controlled light source and a detector and spatial variations in transmitted intensity are used to measure liquid saturation and tracer mass distribution. Weisbrod et al. (25) combined light transmission with fluorescent colloid tracers to measure changes in colloid transport with colloid size and input concentration. Huang et al. (31) developed a fluorescence system in which fluorescent solute within a translucent sand bed is excited by ultraviolet (UV) light and emits visible light at discrete wavelengths. The wide separation of excitation and emission wavelengths eliminates background noise when combined with bandpass filters at the detector. This study extends the UV-fluorescence technique to measure colloid and solute transport phenomena by timelapse fluorescence imaging (TLFI). Using tracers with discrete emission spectra, paired measurements are made of conservative solute and colloid tracers passing simultaneously through a thin sand bed. We present a series of pilot experiments in steady-state saturated conditions to assess the feasibility of using TLFI to study colloid transport at mesoscale. We measure carboxylate-modified latex (CML) microspheres at monovalent ionic strength from 10-3 to 10-1 M, in the pH range 4-6, and at volumetric flow rates of 1 and 2 mL min-1. The effects of ionic strength on colloid deposition can be described by Gouy-Chapman diffuse double layer and DLVO theories for particle-surface interactions in an electrolyte (3, 32). In DLVO theory total interaction energy is determined by the sum of electrostatic repulsion and van der Waals attraction. As electrolyte ionic strength increases, electrostatic repulsion is reduced and the probability of interaction leading to deposition increases. This results in reduced colloid mass breakthrough and increased colloid mass retained in porous media columns (e.g., 33). We demonstrate that these phenomena can be detected and quantitatively analyzed by mesoscale visualization techniques.
Methods and Materials Reflexive Ultraviolet Fluorescence Imaging. In UV fluorescence imaging, tracers within a transparent porous medium in a flow chamber with UV-transparent (quartz) viewing surfaces are illuminated by UV light in a darkroom. 10.1021/es060373l CCC: $33.50
2006 American Chemical Society Published on Web 08/30/2006
FIGURE 1. Schematic diagram showing major components of the UV reflexive fluorescence imaging system. The tracers fluoresce at visible wavelengths and emitted light is detected by a charge-coupled device (CCD) detector. The intensity of fluorescence is directly proportional to the tracer concentration at any point. A detailed description is provided by Huang et al. (31). This study uses reflexive light geometry in which both light source and detector are placed in front of the chamber (Figure 1). Such geometry permits installation of discrete sampling ports in the rear of the flow chamber and simplifies its construction. However, at high tracer concentrations the fluorescence signal is strongly biased to the front surface of the sand bed due to rapid absorption of excitation light. The Beer-Lambert law states that for a typical “strong” fluorophore with molar absorptivity 75 000 cm-1 M-1 the transmitted excitation intensity reduces exponentially by >80% in the first mm at concentrations above 10-4 M. At 10-5 M, the rate of reduction in transmitted intensity of excitation light is lowered to 15% per mm path length. The minimal background noise characteristic of UV-fluorescence imaging means it is particularly suitable for use with tracers at these low concentrations (20) and thus reflexive imaging geometry becomes feasible. A detailed analysis of reflexive fluorescence imaging for mesoscale colloid transport studies is provided in the Supporting Information. Flow Chamber and Imaging System. The flow chamber was constructed of three Perspex plates bolted tightly together and separated by rubber seals. The rear plate was solid. The middle and front plates were cut away to accommodate a quartz glass view plate. When assembled, the internal void measured 200 × 100 × 6.7 mm. Two upper and two lower ports in the rear plate permitted input and extraction of aqueous phase materials from the chamber. Wire gauze separators covered the lower ports to prevent clogging by sand. Blacklight tubes (Sylvania 350BL) positioned above and below the detector line of sight produced UV excitation light (main peak 350 nm, minor peak at 470). Tracer fluorescence (visible light) was imaged by an 8-bit color CCD camera (Hitachi KP-D581, Hitachi-Denshi UK) placed 900 mm from the front of the chamber and connected to PC image capture software (Data Translation, UK). The pixel resolution in images of the sand bed was 0.478 mm2. Manually interchangeable 10 nm bandpass filters (25.4 mm diameter, Ealing Catalog, UK) separated fluorescence signals from colloid
(excitation 350 nm, emission 612 nm) and solute (excitation 475 nm, emission 530 nm) tracers. Materials. The granular porous medium was quartz sand (99.999% SiO2, Multi-labs Ltd, UK). The mean grain size was
