May 29, 2002

Arsenic in Drinking Water - Part 4

Arsenic Removal Methods

Conventional Treatment: Precipitation with Iron, Aluminum

As early as 1972, prior to the start of the U.S.
Environmental Protection Agency (EPA) review of the arsenic standard, arsenic
removal studies (Gulledge and O’Connor) demonstrated that following
oxidation to arsenate ion, As V, arsenic could be readily removed (90 to 98
percent) by conventional water treatment processes utilizing chemical
coagulants, such as iron and aluminum. Iron precipitates were found to be
particularly effective in adsorbing or coprecipitating arsenic. Subsequent
studies have repeatedly confirmed effective removal by precipitation with iron.
As a result, where arsenic removal is practiced, it generally is removed along
with precipitated iron. Where dissolved (ferrous) iron is absent in the source
water, the addition of an iron coagulant, followed by iron hydroxide
precipitation and filtration is used to remove arsenic with great efficiency
and at low cost.

Numerous variations on the direct use of iron and aluminum
coagulants have been or are currently being tested. Whether the
coagulant/adsorbent is dissolved, previously precipitated or attached to a
proprietary medium, most of the methods currently under development reportedly
are successful in achieving effective arsenic removals.


Alternative Arsenic Removal Techniques

Less commonly used arsenic removal techniques include anion
exchange and adsorption on activated alumina. Both produce arsenic-bearing
brines or rinse waters that may require evaporation for volume reduction prior
to their ultimate disposal. While suitable for small, household treatment
systems where the medium is replaced on a prescribed schedule, greater process
control and more comprehensive monitoring would be required to maintain
continuously effective arsenic removal using either of these processes in a
community water treatment system.

Even more costly water treatment techniques such as reverse
osmosis and electrodialysis have been used to develop some cost estimates of
the potential economic impact of a national program of arsenic removal.
Membrane separation processes might prove economically feasible if other water
quality objectives such as the reduction of dissolved solids and nitrate ion
concentrations were required in addition to arsenic removal.


USEPA-Designated Best Available Treatment Processes

USEPA offers the following seven processes as “best
available treatment” (BAT) for the removal of arsenic V. While arsenic
may be predominately in the oxidized As V form in surface waters, any reduced
As III first should be oxidized to As V by pretreatment. As III oxidation is
particularly important for groundwater sources where reducing conditions
normally prevail. It may be accomplished by prechlorination, chloramination,
the addition of potassium permanganate or contact with previously precipitated
manganese dioxide, MnO2.

Conventional Treatment

Aeration / Filtration

Treatment of groundwaters for iron and manganese removal is
highly effective for arsenic removal since precipitated ferric and manganic
oxides readily adsorb As V. The arsenic is recovered in the relatively small
volume of sludge settled from the filter backwash water. As little as 1 or 2 mg
Fe/L in or added to the well water should result in meeting the maximum
contaminant level for arsenic. Monitoring for iron removal might be used to
confirm effective arsenic removal.


Coagulation / Filtration

Conventional surface water treatment employing iron or alum
coagulants forms hydrous ferric or aluminum oxides. Consistent arsenic removals
should be achieved using low dosages of iron or conventional dosages of alum.
Operationally, monitoring of finished water turbidity should confirm effective
arsenic removal.


Lime Softening

Precipitation of hardness by lime with supplementary
coagulant addition may result in the formation of a mixed precipitate of iron,
aluminum and magnesium oxides along with calcium carbonate. Arsenic would be
merely a trace component of the large quantity of softening sludge produced.
Again, a relationship between arsenic and turbidity in the finished water
should be established for purposes of operational monitoring of effective
arsenic removal.


Adsorption Media

Anion Exchange

AsO2– (As III) and AsO3–(As V) are exchanged for
chloride ion on strongly basic anion exchange resin. The use of a salt (NaCl)
regenerant results in the production of an arsenic-bearing brine. The
breakthrough of sulfate ion may signal the loss of treatment unit capacity for
As V removal.


Activated Alumina

Adsorption of As V on aluminum oxide (AlO2) at pH 6 may
require relatively low hydraulic surface loading rates. Following backwash, the
medium is regenerated first with sodium hydroxide, then rinsed with sulfuric
acid. If fluoride is present in the source water, its removal may be helpful in
confirming arsenic removal.


Membrane Processes

Reverse Osmosis

Under high pressure, reverse osmosis (RO) membranes reject a
portion of arsenic, sulfate, nitrate and other ions from relatively
particle-free, pretreated water. Arsenic removal by reverse osmosis membranes
may be related to overall total dissolved solids reduction. The arsenic removed
is recovered in a large volume of RO concentrate. Membrane effectiveness is
affected by suspended solids, scaling by carbonates, silicates and sulfates,
hydrogen sulfide and the tendency of the source water to support microbial


Electrodialysis Reversal

Similar to reverse osmosis.

Coagulation/filtration and lime softening are not designated
as BAT for systems with fewer than 500 service connections. Instead, EPA has
defined “small system compliance technologies” (SSCTs) that limit
the treatment technology for arsenic removal for these smaller communities.
These limitations are based on the presumption that smaller communities will
not be able to provide “appropriate” operation and maintenance.
Accordingly, reverse osmosis and electrodialysis reversal also are not BAT for
these systems.


POE/POU Treatment

EPA will allow small water utilities to utilize
point-of-entry/point-of-use devices to meet enforceable drinking water
standards. These utilities would own, operate and maintain the devices plus ensure
compliance with the MCLs. Pilot testing on the source water would be required.
The utilities might have to seek revisions to local ordinances to require
consumers to provide access to the installed devices for maintenance. Frequent
sampling and additional staff may be required. Currently, USEPA asserts that
POE/POU treatment can be less expensive than central treatment for communities
up to 250 people.


Proprietary Media

Water treatment equipment manufacturers are offering package
plants containing proprietary media for arsenic removal. Filtronics’
“Electromedia” and General Filter’s GFH (granular ferric
hydroxide) media are examples. The former medium is backwashed and continually
reused; the latter is replaced after exhaustion.

Another medium undergoing testing at the University of
Missouri-Columbia utilizes triple reverse burn (TRB) char prepared from a
subbituminous coal. After service, this medium can be reduced to ash.


USEPA Arsenic Removal Study: “Plant A” Results

Starting in 1998, EPA began monitoring arsenic removals at a
central Illinois water treatment plant identified only as “Plant
A.” The water system for this community has treated groundwater for the
removal of iron and manganese since 1970. EPA monitoring results showed that,
unintentionally but incidental to the treatment provided for iron removal,
arsenic concentrations were reduced from 20.3 µg/L to less than 3
µg/L, the lowest arsenic standard under consideration by EPA. The
incremental cost for the highly effective (91 percent average) removal of
arsenic at Plant A was zero. It appears that this utility, with three identical
conventional iron removal plants, has been effectively removing arsenic from
its groundwater for more than 30 years. In the Midwestern United States, thousands
of iron removal plants essentially use the same standard process as Plant A,
one of the three operated by the Village of Morton.


Disposal of Arsenic-Bearing Treatment Plant Residues

As expected, the arsenic removed at Morton’s Plant A
is recovered as a minor constituent of the iron hydroxide sludge settled from
filter backwash water. This has created a secondary concern over the practice
of land disposal of the water treatment plant residues and raised the question
of whether the Plant A sludge required regulation as a hazardous waste. Until
recently, the water treatment plant residue was discharged to the municipal
wastewater collection system and subsequently was combined with the wastewater
sludge. The blended sludges ultimately were disposed of on farm land.

The EPA Toxicity Characteristic Leaching Procedure (TCLP)
test has indicated that the sludge from Plant A is not readily leached and
should be accepted in nonhazardous landfills.

The results from the Plant A study demonstrate that
coprecipitation of arsenic with iron provides a stable and inexpensive means
for effectively removing arsenic from groundwater. This treatment procedure
also minimizes the costs of disposing of the treatment plant sludge residues.


Implications for the Feasibility of Meeting More Stringent Arsenic

As of June, 2001, EPA has cited a $200 million annual cost
to municipalities, states and industry of meeting the 10 µg/L arsenic
standard by 2006. One assumption made is that the cost of meeting still more
restrictive (lower) standards would be exponentially greater for a given
utility. This is not likely true in the case of arsenic when more than 95
percent removals are being observed even at older, conventional plants not
specifically designed for arsenic removal. Still further reductions in arsenic
may require only increased iron coagulant dosages and, as a result, arsenic
removal may be essentially independent of the physical facilities required.

Alternately, as indicated in Part 3 of this series, a more
restrictive, state-mandated arsenic standard would require a greater proportion
of water utilities to reduce influent arsenic concentrations. Therefore, to
more accurately assess the impact of a more stringent arsenic standard, a
nationwide, multi-year program of finished water monitoring is required to
fully define the number of supplies potentially impacted. This should be
accompanied by detailed data on existing treatment facilities plus other
pending finished water quality problems (e.g., DBP, radium) that have yet to be
addressed. These latter data would facilitate a rational assessment of the
appropriate treatment modifications required and permit a more equitable
assessment of the portion of the cost that should be allocated to arsenic
removal alone.

Many of the smaller (predominantly, rural) utilities
impacted by the arsenic rule face other MCL violations often due to the
presence of agricultural chemicals such as nitrite, nitrate and atrazine. Their
compliance needs might be addressed by providing treatment, nominally, to
address their most critical health issue. A dispassionate assessment of small
system treatment needs may show that these water utilities require the
construction of treatment facilities to meet the requirements of existing
regulations independent of the need to reduce arsenic concentrations.


Arsenic Removal in Albuquerque, N.M.

Since the groundwater source for Albuquerque, N.M., contains
52 µg/L of arsenic, treatment facilities are being designed to rectify
this situation. Following pilot scale evaluation of three options
(coprecipitation with iron, anion exchange and activated alumina)
coprecipitation of arsenic with ferric hydroxide formed from the addition of
ferric chloride was selected. Pilot studies indicated that arsenic
concentrations could be reduced to less than 2 µg/L (96 percent arsenic
removal) by the addition of 5 to 22 mg/L of ferric chloride. This result is
consistent both with the laboratory results and with the full-scale arsenic
removal experience at Morton, Ill.

While the treatment plant planned for Albuquerque is more
elegant (e.g., skid-mounted microfiltration units, lamella thickeners, recessed
plate filter press) than Morton’s Plant A, the process and the predicted
results are essentially the same.

Anion exchange was rejected because it requires the use of
large quantities of sodium chloride. The arsenic-bearing brine produced in
regenerating the anion exchange resin was classified as a hazardous waste,
further increasing brine disposal costs.

The use of activated alumina would require the addition of
sulfuric acid to lower the pH of Albuquerque’s alkaline water. Final
treatment then would require the addition of sodium hydroxide to neutralize the
treated water. As with anion exchange, arsenic adsorption on activated alumina
would generate a brine classified as a hazardous waste.

Not only is the estimated capital cost of an iron
precipitation facility lower, but also the estimated annual operation and
maintenance costs reportedly would be 60 percent of the other alternatives.
While a great deal of research has been done in recent years on the
effectiveness of alternate processes for arsenic removal, little attention
previously was focussed on the potential capital and operating costs of some of
these alternatives.


Summary of Treatment Options

A range of appropriate technology for the removal of arsenic
from many, if not most, community water supplies exists. Among the options is
iron removal, one of the most common processes used in the treatment of both
ground and surface waters in the United States. Coagulation with aluminum
sulfate also is effective although requisite chemical dosages may be higher and
the effective pH range somewhat narrower.

The incremental costs for the removal of arsenic in those
communities already having filtration facilities should represent only a small
fraction of the total cost of producing and delivering safe drinking water.

In many cases, where citizens and community leaders have
been reluctant or financially unable to support the cost of providing treatment
to meet existing regulatory requirements, the need for arsenic removal may
provide an additional incentive to either provide comprehensive treatment, seek
a more secure alternate water source or purchase water from a regional water
commission. The latter option is being adopted rapidly throughout the
midwestern United States as smaller communities confront the cost and legal
implications of meeting an increasing number of health-based regulatory
requirements. Progressively, experience is confirming the expected economic
benefits of the regionalization of these water supplies. As in other United
States industries, in drinking water production and distribution, substantial
economies of scale are achieved with the management and operation of
larger-capacity water facilities.

About the author

John T. O’Connor, EngD, P.E., is CEO of H2O’C
Engineering, Columbia, Mo. Phone 877-22-WATER; e-mail: [email protected].