by Daniel N. Creek & James M. Davidson

Ongoing research by the California MTBE Research Partnership shows clearly that there are applications for GAC as a cost-effective choice for MTBE removal from water.  This paper addresses those applications by discussing feasibility, design, and operational considerations for MTBE removal using GAC.

Feasibility Considerations

Specific characteristics of GAC play a part in the adsorption effectiveness for MTBE.  Different pore sizes within the GAC matrix will attract different contaminant molecules to fill adsorption sites.  Due to coconut’s higher density of micropores (high-energy adsorption sites), coconut shell GACs are expected to be more effective than standard coal-based GACs for MTBE removal (California MTBE Research Partnership, 2000). 

Site conditions should be considered when making a preliminary evaluation of GAC cost effectiveness for MTBE removal.  Important variables to consider are the concentration range of MTBE to be treated and the background water quality conditions that are expected at the site.  Computer modeling and cost estimates by the California MTBE Research Partnership (2000) show that GAC is most likely to be cost-effective for lower concentrations of MTBE.  Results of this recent study show that, for example, unit treatment costs for a 600-gpm system increase from $0.77 to $2.37/1,000 gallons as influent MTBE concentration rises from 20 ppb to 2,000 ppb.  This cost increase is caused by the higher GAC usage rate needed for higher influent MTBE concentrations.  The reader should note that these cost estimates were developed based on a detailed set of assumptions not given here; for further information, please see California MTBE Research Partnership (2000).

GAC removal effectiveness is also highly dependent on natural organic matter (NOM) and the presence of other contaminants such as BTEX.  Recent column testing has shown that GAC usage rates vary substantially depending on the background water conditions (California MTBE Research Partnership, ongoing study).  MTBE removal from surface water (Lake Perris) resulted in a carbon usage rate approximately 400% higher than that for a low NOM ground water (South Lake Tahoe area).  Similarly, column testing and computer modeling have shown that the presence of other contaminants (e.g., BTEX) can increase usage rates by 50% or more.  As such, GAC is most likely to be cost-effective for waters that are low in background NOM (e.g., typical groundwater) and clean of other contaminants.

Design and O&M

Although site conditions will determine the ultimate effectiveness of GAC for MTBE removal, there are several recommendations for system design and O&M that can be generalized for all MTBE sites.  These include site-specific testing, in-series operation, GAC variability, and water sampling requirements.

Because GAC effectiveness is dependent on background water quality, bench- or pilot-scale testing using the site water and dissolved-phase contaminants is usually required for confident system design.  Isotherm tests, which give equilibrium adsorption capacity, are relatively fast, cheap, and can readily be used to evaluate relative adsorption effectiveness of different carbons.  For more accurate estimates of carbon usage rate under dynamic conditions, column testing with site-specific water is recommended.  There currently are two well-established column-testing methods for GAC.  The rapid small-scale column test (RSSCT), developed by Crittenden et al. (1989), is currently undergoing consideration as a testing standard for the American Society for Testing and Materials (ASTM).  The second method, developed by Calgon Carbon Corporation, is known as the Accelerated Column Test, or the ACT.  Both of these methods utilize scaling relationships to predict full-scale GAC usage from data developed with bench-scale columns.

In-series operation of two or more GAC vessels is recommended due to MTBE’s relatively weak adsorption to GAC, which causes an extended mass transfer zone within the GAC matrix.  Operation of two or more GAC vessels in-series allows for higher MTBE removal rates by allowing the lead GAC vessel to reach saturation prior to changeout.

As discussed previously, it appears that coconut shell GAC is more effective for MTBE removal than coal-based GACs.  However, because of the variability of coconut source materials, the effectiveness of coconut shell carbon is likely to be more variable than the effectiveness of more uniform coal-based GACs.  As such, material quality assurance/quality control (QA/QC) is of importance and should be given appropriate attention during initial GAC selection and vessel changeout.

Frequent sampling and analytical testing of influent, midfluent, and effluent water is recommended to monitor changing influent conditions, system removal effectiveness, and to anticipate changeout requirements.  Desorption of MTBE can occur as the influent MTBE concentrations drop or with the arrival of more strongly adsorbed compounds such as BTEX.  As such, monitoring plans should account for the impact of changing influent conditions on MTBE removal effectiveness.

Strengths and Weaknesses

The use of GAC for MTBE removal has the following weaknesses:

• Changing influent conditions will impact removal effectiveness;

• GAC removal effectiveness is dependent on natural organic matter;

• Monitoring of influent, midfluent, and effluent is required; and,

• Cost effectiveness decreases for higher MTBE influent levels.

The use of GAC for MTBE removal has the following strengths:

• Reliable and technically straightforward;

• Well-established and easy to implement;

• Expected to be effective for relatively low MTBE influent levels; and,

• Effective as polishing step after primary treatment.

References

California MTBE Research Partnership, 2000.  Treatment Technologies for Removal of Methyl Tertiary Butyl Ether (MTBE) from Drinking Water: Air Stripping, Advanced Oxidation Processes, Granular Activated Carbon, Synthetic Resin Sorbents.  Center for Groundwater Restoration and Protection, National Water Research Institute, Fountain Valley, CA.

Crittenden, J.C., Reddy, P.S., Hand, D.W., and Arora, H., 1989.  “Prediction of GAC Performance Using Rapid Small-Scale Column Tests”, AWWA Research Foundation, September.

McKinnon, R.J., and Dyksen, J.E., 1984.  “Removing Organics from Groundwater Through Aeration Plus GAC”, Journal of AWWA, May, pp. 42-47.

Top