National
Drinking Water Clearinghouse
West Virginia University
P.O. Box 6064
Morgantown, WV
26506-6064
How
To Operate and Maintain Manganese Greensand Treatment Units
Story and Photos by Larry Rader,
NDWC Environmental Consultant
Recently, I had the opportunity to provide onsite technical assistance to
a small elementary school. On my initial visit, I encountered a sign in large,
block letters saying, “DO NOT DRINK THE WATER.” The 177 students
and 28 faculty members were using bottled water for drinking, signs were posted
everywhere warning about the water, and a local TV crew was on the way. Although
the school had its own water treatment system, poor water quality forced the
state health department to issue a “boil water” order. This was
not a good situation, particularly for an elementary school.
This school was built outside an area served by municipal drinking water.
Drilling a well and installing a greensand treatment process was the only
alternative to assure a sufficient supply of safe water. The treatment process
consisted of three cylindrical units, 14 inches in diameter and 65 inches
in height. Each fiberglass unit contained 40 pounds of support gravel and
3.25 cubic feet of manganese greensand, which was being regenerated during
backwash with potassium permanganate. A 300-gallon galvanized tank provided
chlorine contact time. Soda ash for pH adjustment and calcium hypochlorite
for disinfection completed the chemical treatment. At times, they were able
to obtain a trace chlorine residual throughout the school. This was the exception,
however, and the water was usually discolored, with an objectionable taste
and odor, and had the ability to stain anything it contacted.
Poor Water Was One Challenge
The school’s raw water quality would cause problems even for a larger,
more sophisticated treatment plant. An onsite analysis showed 16.3 parts per
million (ppm) of iron, 1.19 ppm manganese and a 6.5 pH. There was also a small
amount of hydrogen sulfide present. Each of the three treatment units had
approximately one square foot of surface area for a total surface area of
three square feet, and the pump was producing 21 gallons per minute (gpm).
One unit was on stand-by, causing surface loading greater than 10 gallons
per minute per square foot (gpm/sq. ft) on the other two. At some point, an
attempt to convert the chlorine contact tank into a pre-treatment basin had
only made the problems worse.
I will be the first to admit that, for many years, I paid little attention
to these small treatment units. The size and design just did not look like
serious water treatment to me. However, the more I learned about them, the
more I realized they required the same operation, maintenance, water treatment
knowledge, and experience as units many times their size.
 Photo Caption-Three slightly larger grensand units located at Douthat State Park ,Virginia |
Manganese
Greensand
The manganese greensand process has been used effectively for removing iron,
manganese, and hydrogen sulfide since
the
1950s in the U. S. Manganese greensand is processed from what is commonly
known as “New Jersey greensand,” but is more correctly identified
as glauconite. For iron and manganese removal, the naturally occurring singular
grains of glauconite are washed and classified to produce a filtration media
having a sieve analysis of 18 x 60 mesh with a resulting effective size of
0.3–0.35 millimeters (mm) and a uniformity coefficient of 1.60 or less,
giving the media excellent filtration characteristics.
The glauconite is first stabilized then coated with manganese oxide. This
coating provides the glauconite with its special chemical oxidation-reduction
properties for the removal of iron and manganese, as well as small quantities
of hydrogen sulfide.
The greensand process for removing iron and manganese is usually accomplished
using one of two procedures: continuous regeneration (CR) or intermittent
regeneration (IR), or, in some cases, both.
|
Photo Caption-This photo shows a typical setup to feed potassium permanganate consisting of a tank,mixer anf feed pump. |
Continous
Regeneration(CR)
CR is primarily used when iron removal is the main objective.
The CR method is exactly what the name implies, continuously feeding an oxidizer,
such as chlorine, potassium permanganate (KMn04), or a combination of the
two, into the raw water prior to the filter. This process can remove 15 milligrams
per liter (mg/L) or more of soluble iron. However, with high concentrations,
the flow rate through the filter must be reduced to 1.5–2.5 gpm/sq. ft.
At these higher concentrations, the filter runs can also be reduced to as
little as four to six hours. Waters having iron concentrations in the lower
range of 0.5–3.0 mg/L could expect filter runs of 18 to 36 hours at a
higher flow rate of 3–5 gpm/sq. ft. If high concentrations of iron are
being oxidized prior to the filter, capping the filter with anthracite to
a depth of 18 inches or greater will increase run times.
Filters removing large amounts of particulate matter, such as oxidized iron,
benefit greatly from the addition of several inches of anthracite as a cap.
Anthracite is less dense than greensand or normal filter sand, allowing particles
to penetrate deeper before being trapped. In fact, if the main consideration
is removing oxidized iron, the greensand layer itself may be reduced slightly
to allow for the addition of a full 18 inches of anthracite. Larger greensand
units that have a trained operator present may use KMn04 in addition to chlorine,
in the CR method.
If both iron and manganese are present, chlorine should be injected into the
raw water first, followed by KMn04. These feed points should be spaced to
allow the chlorine—the cheaper of the two chemicals—time to oxidize
as much iron as possible before the KMn04 is injected. The KMn04 will then
oxidize the manganese as well as any remaining soluble iron. The accepted
procedure is to carry a slight excess of permanganate onto the filter. Manganese
greensand will reduce the excess permanganate to manganese oxide. The manganese
oxide will then precipitate on the grains of media, maintaining them in a
continually regenerated state. When using this process for filter regeneration,
a thin line exists between regeneration and break-through. For this reason,
smaller treatment units, specifically those without full-time operators, should
consider the CR process using only chlorine as the pre-oxidant.
All new greensand must be fully regenerated prior to placing it into service.
This procedure is printed on each bag of greensand and consists of soaking
the media for several hours in a 2–3 percent solution of KMn04. The procedure
that I have used for years consists of thoroughly mixing one pound of KMn04
to five gallons of water, applying the solution to the filter until the media
is covered, then allow to “marinate” overnight. The following morning,
the filter is backwashed until no permanganate remains in the wash water,
then placed into service. From this point on, maintaining free chlorine residual
greater than 0.5 at the filter effluent will insure the media retains the
all-important manganese oxide coating. Because the school already had intermittent
regeneration equipment in place, we decided to combine the CR with IR.
 Photo Caption-The Norton,Harding,Jimtown(West Virginia)PSD serves approximately,1000 customers in two communities.On the right is a greensand filter and a carbon filter,used for low levels of phenols in the raw water,is on the left.This plant has two additional greensand filters that remove 3.3 mg/L iron and 0.4 mg/L at a pH of 6.5 |
Intermittent
Regeneration (IR)
IR is normally used when the problem is mostly manganese with lesser quantities
of iron.
In this process, manganese oxidation occurs directly using the properties
of the freshly regenerated manganese greensand. After treating a specific
amount of water, the oxidation capacity of the media will be consumed and
regeneration is required. Following a normal backwash cycle, the bed is regenerated
by the down-flow passage of a dilute KMn04 solution through the filter bed
using 1.5 ounces of KMn04 per cubic foot of media. This solution is allowed
to remain in contact with the media for several minutes. Following regeneration,
the filter will require rinsing until all the excess permanganate is gone.
The rinse water containing the excess permanganate can be directed to a container
for use in the next regeneration. This method minimizes permanganate disposal
problems and reduces chemical costs. One drawback in the IR process is the
extra time the filter is out of service for backwashing. The CR process requires
approximately 15 minutes for backwashing and rinsing, rather than a complete
regeneration required for the IR process (usually 75 minutes).
For operations using sodium hydroxide as a pH adjustment, along with the CR
process, the pre-pH correction should not exceed 6.8-7.0 to prevent the formation
of a non-filterable colloid, which sometimes occurs.
Backwash Rates
The recommend backwash rate for manganese greensand is 12 gpm/sq. ft. of filter
area at 60 degrees Farenheit. This rate is sufficient to expand the bed 35–40
percent. Please note that backwash rates versus filter loading rates can cause
serious problems in smaller treatment units. For example, a small installation
with one 12-inch inside diameter filter, will require the well pump to deliver
12 gpm to properly backwash. However, if high levels of iron are present,
that same unit may only be capable of filtering 2 gpm.
Minimum Bed Depth
The minimum depth of the greensand bed is 24–30 inches when using the
IR method and 18–20 inches when using the CR method. If you are removing
large amounts of pre-oxidized iron, reducing the depth of the greensand will
allow for a deeper cap of anthracite. Before changing the configuration of
any filter bed, contact your supplier or the manufacturer.
Removing Iron and Manganese
Even with small treatment units, removing iron and manganese requires knowledge,
skill, and occasional access to a few simple pieces of lab equipment—no
matter what the sales representative told you. A certified laboratory must
perform the initial analysis for any new water supply. If the lab determines
minimal or no treatment is necessary, sending an occasional sample to the
same lab will make certain that changes have not taken place.
If, on the other hand, iron, manganese, hydrogen sulfide, etc., are present,
and you spend the money for a treatment unit to remove them, then protect
your investment and learn to operate it properly. Simple lab tests designed
for small operations on a tight budget can check levels of iron, manganese,
chlorine, pH, and any number of other parameters on site.
If you are using the treatment unit in a school, restaurant, or other situation
where your liability is increased, I would invest in some simple equipment.
The lab equipment that I rely on more than any other when working in well
supplies both large and small, is a filter flask, a hand vacuum pump, and
a package of 0.22 micron filters. When analyzing samples for iron and/or manganese
content, always perform two separate analyses, one before filtering and one
after. Many of the less sophisticated tests used to determine iron and manganese
levels measure total concentrations, both soluble and insoluble.
When measuring levels of a specific inorganic in raw water, it is important
to know what amount, if any, is already insoluble so you can accurately oxidize
what remains. If the finished water has objectionable levels of iron or manganese,
it is important to know whether the problem is caused by incomplete oxidation
or particles of oxidized material that have passed through the filter.
Pressure gauges must be installed on the raw water line before it enters the
filter (influent) and on the finished water line as it leaves the filter (effluent).
Following a backwash, when the filter is clean, the two gauges should have
approximately the same pounds per square inch (psi) reading.
As the filter begins to collect oxidized material, you will begin to see a
pressure differential between the two. The more iron and manganese in the
raw water reaching the filter, the quicker this pressure differential will
occur. This is called “loss of head” and indicates the filter is
becoming dirty. When the pressure difference between the two gauges reaches
8–9 psi, the filter needs backwashing.
Failure to backwash a dirty filter can force iron or manganese, either in
the soluble or insoluble state, through the filter just like toothpaste being
squirted from its tube. In some instances, the precipitated iron is of such
a nature that it filters in depth and leaks into the effluent, even after
a 2–5 psi differential pressure increase. In this case, backwashing should
be initiated on a gallons-treated basis using experience.
Back to School
What happened at the elementary school that was the seed for this article?
Luckily, everyone from the county board of education to the school’s
maintenance people was willing to do whatever it took to fix the problem.
At this school, the design required that the treatment process meet the high
periodic demand for water. Because of the demand, reducing the flow rate was
not an option. Instead, we installed three additional treatment units, bringing
the total to six. Each unit was fitted with an orifice plate that reduced
the influent rate to 2.5 gpm. This provided a combined filter capacity of
15 gpm. All six filter beds were then rebuilt using manufacturer’s specifications
and properly regenerated with potassium permanganate.
Because of the high levels of both iron and manganese, the decision was made
to combine the manufacturer’s IR process with CR using pre-chlorine only.
We are attempting to maintain a free residual of 1.0 ppm at the filter effluent,
which produces an acceptable residual throughout the school.
Pressure gauges were installed on both the influent and effluent lines to
measure head loss. The well pump rate was adjusted, a new 300-gallon chlorine
contact tank replaced the old one, and lines throughout the school were flushed.
Because of the time required to complete an IR backwash cycle, the automatic
controls were set to backwash two filters each night. Six weeks into the school
year, bacteria samples, as well as overall water quality, has been very acceptable.
|
Invest
in Basic Lab Equipment
Perhaps the single factor that contributes most to poor water quality—particularly in smaller treatment facilities—is the lack of minimal basic lab equipment. No operator should be expected to solve a problem unless he or she knows what is actually causing the problem. Although most facilities (even smaller ones) are required to occasionally send samples to a certified lab, this is usually only a one-liter snapshot out of the thousands of gallons that have passed through your treatment process. If you are one of the lucky few with good quality raw water and no complaints, this snapshot may be sufficient. However, for those of us who have listened to the daily complaints, from regulators and customers alike, or have struggled through one problem after another, there is a better way. Hach, LaMotte, and other companies offer low-cost, low-tech equipment to analyze chlorine, iron, manganese, pH, etc. A complete pre-analysis filter kit, containing a hand-operated vacuum pump, filter holder, 0.22 micron filters, and complete instructions is available through Control Equipment Company in Roanoke, Virginia. Call them at (800) 572-3220 or visit their Web site at www.cec-online.com. |
Acknowledgement
The author would like to thank Howard Schwartz, lab technician with Hungerford
& Terry, Inc., for information used in this article.
About The Author
Larry Rader has more than 25 years in the water industry. If you have
a question for Rader, he can be reached through e-mail at lrader@meer.net.