Introduction
With increasing regulation, the global water treatment market has been tasked with finding commercially available technologies suitable for treating and removing arsenic contamination in drinking water to levels below 10 µg/l. At Severn Trent, initial evaluation of arsenic removal technologies centered on a variety of technologies which were thought to be the best suited for this application (Table 1: Comparison Technologies). However, through detailed lab, pilot and full scale research, the use of iron oxide adsorptive media proved itself as a viable technology for reducing arsenic levels across varying drinking water supplies.
Adsorption is a continuous process conducted at a specific flow rate or velocity, normally about 7 gpm/ft, downward through a fixed bed adsorber. Empty bed contact time (EBCT), which dictates the amount of water resident within the bed required to effect complete arsenic adsorption, is another key process parameter. An attractive characteristic of adsorption technology is its simplicity and relatively low cost. For example, coagulation filtration has higher initial capital costs and is labor intensive, with labor costs often not adequately accounted for in operating cost estimates. In addition, this technology is more complex than adsorption, a key factor for utilities without centralized treatment plants.
Methods
At the onset of developing an arsenic removal research program, Severn Trent approached LANXESS (formerly Bayer AG, Germany) to develop a media that could be used to treat high levels. After lab testing different iron oxide media samples, Bayoxide® E33 demonstrated that it had the most important aspects of a viable iron oxide media, namely: it has a high capacity for arsenic, is mechanically robust, is stable with a uniform grain size, has a low leaching potential, has minimal head-loss build-up and is immediately effective in a start-stop process. Severn Trent also initiated a lab-based research program to characterize the performance of the media in a broad array of waters. A statistically significant array of tests was performed with a background water assay based upon NSF 53 water.
After the successful completion of laboratory testing, pilot plant work was undertaken to further research arsenic removal rates, effect of pH, pre-oxidation requirements, impact on disinfection and the effect of other ions. One of Severn Trent’s most challenging pilot programs on the performance of the Bayoxide® E33 media was conducted on a potable water source in New Mexico, United States. The water source was considered challenging due to its high arsenic levels, high pH and high levels of vanadium, a metal that is co-adsorbed by the media. The water analysis, adsorption data and graph for the New Mexico pilot program, which includes a program summary, are shown in Figure 1.
Understanding the effects of other ions is important to the design of an adsorption process because water sources that contain iron, manganese, phosphate, silica, sulfate and vanadium, have been shown to affect process performance. Table 2 details the variations in water quality evaluated during pilot plant testing undertaken by Severn Trent to further refine the predicted full-scale performance of the Bayoxide® E33 media. Hydraulic performance was also studied; evaluating media grain size, empty bed contact time (EBCT), head-loss, differential pressure, bed expansion and backwash volume requirements.
Results and Discussion
The SORB 33™ system, as the adsorption process came to be called, has a relatively small footprint, making it suitable for retrofitting or upgrading existing treatment plants. The system consists of simple adsorber vessels normally operated in parallel flow configuration, (Figure 2: Standard SORB 33™ Adsorption Process). The primary operator functions for the system are monitoring flow, pressure, pressure differential and total flow treated data; collecting effluent samples for arsenic and other analyses; and ensuring each adsorber vessel is backwashed on a periodic basis.
The SORB 33™ systems are designed with an EBCT range of 3.3 – 4.5 minutes. Routine media backwash or service washes – done normally on a monthly basis – can be initiated automatically on a preset date and time, by volume of water treated, differential pressure readings or by operator initiation. Service washes are important as they stratify the media bed and remove fine particulate material, which could cause increased differential pressures during the normal downflow operational mode.
Full scale SORB 33™ arsenic removal systems have been in commercial operation since 1999, beginning with 16 arsenic removal treatment facilities treating over 46 million gallons per day (MGD) in the United Kingdom. In the United States, over 45 SORB 33™ adsorbers are installed across 14 sites.
Through this extensive commercial application, the knowledge base for how this adsorptive media works and how best to optimize its performance has grown steadily. Service washing has been extended at some sites from 28 days to greater than 50 days. In fact, some full scale plants have achieved 90 days before experiencing significant increases in differential pressure and requiring a service wash. In addition to reducing service wash frequency, significant backwash volume reductions of up to 65% have also been made through process optimization.
Conclusions
Years of lab and field tests have shown that Bayoxide® E33 iron oxide media is a viable product with a high capacity to remove arsenic contamination in potable water sources. Still, continuous improvement is essential. Additional research is focusing on improving the Bayoxide® E33 media in order to manage difficult water qualities and increase process efficiencies.
To this end, LANXESS has developed a pelletized version of their media which is currently undergoing full scale evaluation at one of Severn Trent’s well water sites. The trials to date have shown that the pelletized version of the Bayoxide® E33 media has the same high capacity for arsenic removal as the original media, its handling is better, it has lower associated fine levels with low solids release during backwash and it remains 'dust free' when being loaded into a vessel in the dry state.
In addition, the composition of a new media addresses the problems posed by complex water sources in both drinking and non-drinking water applications. Some of the advantages of this new media are predicted to include a higher capacity for arsenic adsorption together with greater robustness. The new media composition, which has increased adsorption capacity and faster kinetics, will help to address difficult water qualities, where high concentrations of arsenic and heavy metals may occur. A media with a higher mechanical stability leads to better handling and overall process efficiencies. Pilot plant testing on this new media is about to be undertaken.
| Table 1-Comparison Technologies |
| Technology |
Process |
Chemical Use |
Waste Generated |
Water Wasted |
| Iron Oxide Adsorption |
Simple |
None |
Low |
<0.1% |
| Reverse Osmosis |
Moderate |
Cleaning chemicals |
Low |
10-25% |
| Ion Exchange |
Complex |
Regeneration chemicals |
High |
2% |
| Activated Alumina |
Complex |
Regeneration chemicals |
High |
5% |
| Coagulation Microfiltration |
Complex |
Cleaning, coagulation chemicals |
Moderate |
5% |
| Table 2-Variations in Water Quality Assay Range From Pilot Programs |
| Assay |
Range |
| pH |
6.5-8.9 |
| Alkalinity |
60-400 mg/L |
| Hardness |
7-350 mg/L |
| Fluoride |
<0.1-2.0 mg/L |
| Phosphate |
<0.01-0.90 mg/L |
| Silica |
5-100 mg/L |
| Sulfate |
5-150 mg/L |
| Total dissolved solids |
100-800 mg/L |
| Metals: |
| Arsenic |
11-200 µg/L |
| Chromium |
2-50 µg/L |
| Iron |
<50-1,500 µg/L |
| Vanadium |
<5-100 µg/L |
Figure 1
Figure 2
For more information, contact Severn
Trent Services at info@severntrentservices.com.