379Arsenic Removal from Drinking Water by Iron Removal Plants
EPA 600-R-00-086
The study was divided into three phases: source water sampling, preliminary sampling,
and longterm evaluation. The first phase, source water sampling, was conducted to
evaluate source water characteristics at each plant. The second phase, preliminary
sampling, was initiated at Plant A in April 1998 and at Plant B in May 1998. This phase
consisted of a four-week sampling period to refine procedures for subsequent events
during the third phase.
The third phase, long-term evaluation, consisted of weekly sample collection and analysis
beginning in June 1998 and continuing through June 1999 at Plant A and through
December 1998 at Plant B. Plant personnel conducted all sampling during the long-term
evaluation phase and Battelle coordinated sampling logistics. Sludge samples also were
collected at Plant A during a single sampling event from an outdoor settling pond in
November 1998. Samples of supernatant discharge (Plant A) and recycle supernatant
(Plant B) were collected monthly beginning in November 1998 and continuing until June
1999 at Plant A and until January 1999 at Plant B.
Results from the long-term evaluation phase were varied regarding the ability of the iron
removal process to consistently achieve low-level arsenic concentrations (e.g., <5 μg/L in
the finished water). The total arsenic concentrations at Plant A were reduced by an
average of 87%, which represents a decrease in average arsenic concentration from 20.3
μg/L to 3.0 μg/L.
Adsorption and coprecipitation with iron hydroxide precipitates are believed to be the
primary arsenic removal mechanisms. The total arsenic concentrations at Plant B were
reduced by an average of 74%, which represents a decrease in average arsenic
concentration from 48.5 μg/L to 11.9 μg/L. At Plant B, it appeared that only the particulate
arsenic in the source water was removed. This particulate arsenic was most likely
associated with the oxidized iron particles present in the source water (i.e., arsenic sorbed
onto iron particles).
The primary difference in arsenic removal efficiency at Plants A and B is believed to be
the amount of iron in the source water. Source water at Plant A averaged 2,284 μg/L of
iron, while Plant B averaged 1,137 μg/L. Increasing the iron in the source water at Plant B
using a coagulant, such as ferric chloride, would likely enable Plant B to consistently
achieve lower levels of arsenic.
None of the sludge samples collected at Plant A qualified as a hazardous waste based on
the Toxicity Characteristic Leaching Procedure (TCLP) test for metals. Therefore,
nonhazardous waste landfills should be able to accept the sludge generated by this
treatment facility. Stricter hazardous waste classification regulations in some states, such
as California, on total arsenic concentrations in solid waste also were met at Plant A.
Sludge samples were not collected at Plant B; however, analytical results were provided
from a 1994 sludge sampling event.