The first brackish water desalination water treatment plant in Massachusetts will be constructed along the Taunton
River in Dighton, Massachusetts. The source water is seasonally brackish and salt water intrusion at the proposed
intake location normally takes place during the period of July through November. The source water is high in
natural organic matter (NOM) and color, with wide variations in temperature and salinity. Thus, the full-scale
process being considered must not only be adequate in providing proper pretreatment for the reverse osmosis
system, but must also be able to provide full treatment when the reverse osmosis system is not required for
removal of total dissolved solids. Specifically, the pretreatment process must be effective in removing NOM,
color, turbidity and particles for compliance with the Stage 1 & 2 Disinfection/Disinfection Byproduct Rule and
Long Term 2 Enhanced Surface Water Treatment Rule.
An alternatives analysis was conducted to identify the most effective process to satisfy these demands. A dual
membrane approach was ultimately selected, involving immersed membrane ultrafiltration (UF) followed by
reverse osmosis (RO). A three season pilot study was initiated in September 2002 and completed in June of 2003
and was conducted to evaluate the performance of the proposed dual membrane system and to provide design
criteria for the full-scale facility. The major focus of this study was to identify the optimal coagulation scheme for
efficient operation of the UF system, and to evaluate the ability of the UF system to provide pretreatment for the
RO system. The coagulants that were investigated included alum, two brands of polyaluminum chloride (PACL),
and ferric chloride.
This paper presents the results of the pilot study, with the major focus on the UF coagulation process. The pilot
study demonstrated that the UF system, operated in the enhanced coagulation mode, was very effective in
providing upstream pretreatment for the RO system, and the UF permeate quality allowed for stable operation of
the RO pilot unit. After optimization of the UF process, silt density index (SDI) values of UF permeate were
below 2.5 units average, total particles were maintained below 5 per mL, and permeate turbidity averaged 0.02
NTU. In addition, the UF system achieved up to 71% removal of total organic carbon (TOC), up to 88% removal
of UV-254, and > 90 % removal of apparent color. The use of ferric chloride as primary coagulant proved most
effective for NOM removal. Includes 9 references, tables, figures.