Colloid Concentration Effect on Natural Media Filtration Efficiency
Sherri Cook, Dr. Stefan Grimberg, Dr. Thomas Holsen, and Aiman Jaradat
Clarkson University ’s Department of Civil and Environmental Engineering
Introduction
Stormwater runoff quality is important due to its ability to affect water supply quality. It is necessary to treat contaminated runoff, but depending on the contaminants present this process can be very costly especially when compared to its benefits (Kalman et al., 2000). Contaminants can include particles, hydrophobic organic compounds which sorp to the particles and pass through the filter, and metals. Natural media filtration (NMF) may offer a cost effective alternative to current best available technology (BAT); natural media filters can be heterogeneous natural porous materials such as compost and other organic materials. Other advantages of using natural media filtration are the potential for biodegradation of the contaminants as well as the sustainability of the filter media. Before using this type of filtering medium can replace other technologies, though, the efficiency of natural media needs to be examined. This research focuses on determining the efficiency of the filter when the amount of particles in the stormwater is increased and how these particles can help transport contaminants through the filter material.
Polychlorinated biphenyl (PCB) contaminated stormwater is the motivation of this research, while the study can be applied to other hydrophobic organic compounds. PCBs are toxic pollutants that have several health affects, as reported by the EPA, which include the potential to cause cancer and to impact the immune system, as well as to disrupt the endocrine and reproductive systems (USEPA 2007). Current treatment of PCB contaminated stormwater uses sand and granulated activated carbon filtration (Jaradat, Grimberg, Holsen, 2007)
Some prior research found that NMF could be effective in removing colloidal PCBs. These experiments have been conducted at one particle concentration. Therefore there is a need to determine the effect of stormwater particle concentration on contaminant transport in NMF systems. If greater colloid concentrations make the filter inefficient, by allowing greater amounts of facilitated-transport of the contaminants, then the research will show at what levels of colloid concentrations that the NMF will be feasible. In order for NMF to be applied to various situations, different independent variables need to be changed to determine the overall efficiency of the NMF; these changes can also show limitations of NMF so its feasible applications are known.
For this experiment, leaf compost was used as the natural media. Also used for this research were sulfate latex particles to represent the model particles in the stormwater as well as phenanthrene to be a substitute for PCBs. Colloids, small particles, not only decrease water quality by increasing turbidity, but they can also allow for the facilitated transport of other hydrophobic compounds; this experiment uses sulfate latex particles as the colloids present in the stormwater (Jaradat, Grimberg, Holsen, 2007). Phenanthrene is an aromatic molecule that sorbs comparably to PCBs; it sorbs to hydrophobic surfaces and therefore represents a reasonable, less toxic, model contaminant. The compounds are similar, and phenanthrene is also a common hydrophobic compound, like PCBs, that has been used in past experiments dealing with colloid-facilitated transport (Roy and Dzombak, 1997). This synthetic stormwater solution was used in column experiments.
The efficiency of the filter media can be determined through quantification of phenanthrene and latex particle breakthrough. The filter should remove the colloids and phenanthrene from the water. Breakthrough indicates that the filter media life to trap particles and contaminants is exhausted.
The goal of this experiment was to examine the effect of different colloid concentration of the transport of hydrophobic organic compounds, like PCBs and phenanthrene, through the filter. Past experiments have shown that effluent concentrations of phenanthrene are greater with the presence of colloids. This experiment can show the degree of colloid influence on effluent quality and contaminant transport.
Materials and Methods
In order to increase the validity of the project, parameters from previous experiments at Clarkson were repeated. The experimental setup consists of several different aspects:
Columns : The columns were one centimeter long with an inner diameter of half a centimeter. Leaf compost was packed into these glass containers, which were then held upright while the solutions were pumped at a specified flowrate through the filter material. Tubing for the columns was stabilized before the sample collection began; a 0.5 mg/L phenanthrene solution was passed through the tubing until the effluent and influent concentrations were similar in order to fully saturate the tubing and decrease sorbing during experimentation. The packed columns were also purged by running de-aired water and a calcium chloride solution through them for several days in order to remove remaining organic material as well as to remove all air in the filter to ensure full sorption ability of the leaf compost.
Stormwater : There were four different synthetic stormwater solutions. These solutions were the influent for the experiment. Each contained 0.5 mg/L phenanthrene and 100 mg/L calcium chloride. The phenanthrene is our contaminant indicator while the calcium chloride increases the ionic strength of the solution. Since de-aired and de-ionized water was used to make these solutions, ions were necessary to make the stormwater solution more consistent with precipitation found in the environment, which contains ions. The last component of the solutions is the sulfate latex particles; these colloids had varying concentrations: 0, 5, 10, 30 mg/L sulfate latex particles. These four solutions were continually mixed during the pumping and sample collection.
Measurements : A Perkin Elmer Luminescence Spectrometer (LS 50B) was u sed to determine concentrations of the phenanthrene and sulfate latex particles. The wavelength of interest for phenanthrene is 364 nm and 515 nm for the sulfate latex particles used in this experiment. Samples of 10-20 mL were taken from the effluent of the columns, and periodically from the influent solutions, in order to normalize the final results. Phenanthrene intensity data was corrected due to the diminishing fluorescence effect of the sulfate latex particles on the fluorescence of the phenanthrene.
Results
Previous Trends: Past experiments, using similar material and methods, have shown trends of the phenanthrene and latex particle concentrations compared to pore volumes. The phenanthrene increases until around 1,000 pore volumes, at which point it levels off. A second increase then appears around 3,500 pore volumes and continues until the normalized concentration of phenanthrene reaches its maximum amount, 1.0. This maximum is reached around 10,000 pore volumes.
For latex particles, there is a steady increase in concentration until about 1,000 pore volumes. Around this time the effluence concentration is the same as the influent concentration and all of the latex particles are passing through the filter and no longer being sorbed to the natural media. Colloidal breakthrough dominants the first part of the phenanthrene breakthrough curve, while breakthrough of adsorbed phenanthrene limits phenanthrene breakthrough at larger pore volumes. Increase in colloidal concentration therefore should only affect the earlier portion of the breakthrough curve assuming no particle filtration is occurring.
Previous Trials: These trends were expected for this experiment. The only difference should have been the normalized concentrations and their varying magnitudes. When these trends were not seen in the data during a run using 0, 10, 25, and 50 mg/L sulfate latex particles, notable problems and method consistencies were identified,.
A previous experimental run allowed the current experiment to be modified and improved. The tubing was allowed to reach equilibrium by running a phenanthrene solution through the tubing for several days; this equilibrium will decrease the loss of the hydrophobic organic compound to the tubing. Also, the leaf compost packing was made lighter to achieve a pore volume around 4 minutes instead of 2-3 minutes; this increased porosity will help to prevent the physical filtering of the colloids. Less glass wool, located at the ends of the columns to maintain the compost in the columns, was used to prevent the physical blocking at the influent e d of the column. By decreasing the amount of physical blocking, the data will better represent the other mechanisms of contaminant transport through the filter. Other parameters that were modified include the decreased length of the tubing, in order to decrease sorption of the colloids and phenanthrene on the tubing, and the increased flowrate to reduce travel time in the tubes and coagulation with the greater velocity.
Phenanthrene: The breakthrough curves of phenanthrene for the current experiment only includes to approximately 3000 pore volumes; a completed data set should be around 10,000 pore volumes. The following figure is the most recent compiled data:
Figure 1. Recent Phenanthrene Effluent Data (as of July 24, 2007)
With the experiment still running, most of the data has not be collected and analyzed. Thus far, the shape of the breakthrough curves for phenanthrene is similar for each column. Phenanthrene breakthrough occurs first for the column receiving no latex particles (Figure 1). In previous experiments, columns receiving latex particles broke through earlier presumably do to the facilitated transport of latex particles in the compost media. These results, however, show that colloidal particles are effectively removed in the filter media (Figure 2), thus facilitated transport is inhibited. Since experiments were conducted with an alternate batch of leaf compost it may be possible the surface properties of this batch are different than the ones used earlier. A positively charged filter media would attract negatively charged latex particles, thus efficiently remove the colloidal phase in the NMF filter. Unfortunately no surface charge measurements were conducted on the media to verify this hypothesis.
Latex Particles : Again, the samples only provide data to around 3000 pore volumes and the normalized concentrations are shown in the following figure:
Figure 2. Recent Latex Particle Effluent Data (as of July 24, 2007)
The particles are not breaking through the NMF as soon as expected, but they continue to increase in concentration. Also, there is qualitative data of the particles filtering through the columns as the pore volumes increase, a trend seen in past experiments.
Conclusion
The compiled results of this experiment will show the relationship between phenanthrene breakthrough as a function of colloid particle concentration. Since stormwater contains colloids, the efficiency of the filter will be based on its efficiency to retain the contaminant even while the water contains large concentrations of the colloid. Filtering of the colloids themselves is also of importance; the filtered water should not be turbid, so the particles should not pass through the natural media filter either.
If conclusive data is collected for this varying colloid concentration experiment, then future examinations can change other variables to continue to determine the efficiency of natural media filters. Some modifications include changing the type of compost, leaf or mushroom, as well as altering the ionic strength of the stormwater solution in order to determine the affect on effluent quality.
REFERENCES :
“Health Effects of PCBs” Polychlorinated Biphenyls (PCBs). USEPA, 2007: http://www.epa.gov/pcb/pubs/effects.html
Jaradat, Aiman Q.; Grimberg, Stefan J.; Holsen, Thomas M. “Colloidal Transport Through the Natural Media Filter.” Manuscript to Journal of Environmental Engineering 2007, 1-38.
Jaradat, Aiman Q.; Grimberg, Stefan J.; Holsen, Thomas M. “Transport of Colloids and Associated HOCs in NMF.” 2007, 1-34.
Kalman, Orit et al. (2000). “Benefit-cost Analysis of Stormwater Quality Improvements.” Environmental Management 2000, 26, 615-628.
Roy, Sujoy B.; Dzombak, David A. “Sorption Nonequilibrium Effects on Colloid-Enhanced Transport of Hydrophobic Organic Compounds in Porous Media.” Journal of Contaminant Hydrology 1998, 30, 179-200.