United States
Environmental Protection
Agency
Water Engineering
Research Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-87/109 Mar. 1988
&EPA Project Summary
Reverse Osmosis Treatment to
Remove Inorganic
Contaminants from Drinking
Water
Martin R. Huxstep and Thomas J. Sorg
A pilot study by the Charlotte
Harbor Water Association determined
the removal of a number of inorganic
contaminants from drinking water
using five state-of-the-art reverse
osmosis membrane elements. The
test system consisted of a small, 19
m3/day (5,000 gal/day) capacity,
reverse osmosis system that was
modified to accept the five different
membranes.
A Florida ground water was spiked
with the contaminants and test runs
of 1 to 13 days were made with each
membrane. Removal data were
collected for each membrane for all
spiked contaminants and for several
naturally occurring substances.
The five state-of-the art
membranes in the study were Toray
SC 3100*. Filmtec BW 30-4021, Dow
5K, Dupont B-9, and Hydranautics
P/N 4040 LSY-IFCI. The spiked
contaminants consisted of fluoride,
cadmium, mercury, chromium,
arsenic, selenium, nitrate, and lead.
Limited tests were also conducted
on nitrite, molybdenum, copper, and
two naturally occurring radio-
nuclides - uranium and radium.
During each test run, removal data
was collected on several naturally
occurring substances, hardness,
chloride, sulfate, total dissolved
solids (TDS), and, on a limited basis,
on calcium and sodium.
"Mention of trade names or commercial products
does not constitute endorsement or
recommendation for use.
Although the membranes operated
under different conditions (pressure,
recovery) according to manufactur-
er's specifications, rejection of the
spiked and natural substances was
generally in agreement for all
membranes. Highest removals (above
95%) were achieved on arsenic V,
calcium, cadmium, chromium III and
VI, copper, lead, molybdenum,
sodium, radium, selenium IV and VI,
uranium, hardness, and total
dissolved solids. Moderate removals
(85% to 94%) were obtained on
fluoride, chloride, nitrate, and nitrite
and lowest removals (below 85%) on
arsenic III and inorganic mercury.
Wide variations in removal
occurred with four contaminants:
arsenic III, mercury, fluoride, and
nitrate. Because nitrite tests were so
limited, no conclusion could be made
on variability. Variability in removal of
these contaminants was attributed to
the chemistry of the contaminants
and the test waters, membrane
material, and test conditions.
This Project Summary was
developed by EPA's Water Engineering
Research Laboratory, Cincinnati, OH,
to announce key findings of the
research project that is fully
documented in a separate report of
the same title (see Project Report
ordering information at back).
Introduction
Reverse osmosis (RO) is a relatively
new water treatment process, having
been applied successfully in desalting
-------
L—Low Pressure Shutoff Switch
H—High Pressure Shutoff Switch
P—Pressure Valve
S—Sample Valve
F—Flowmeter
pH—pH Meter
Permeate
Heat
Exchanger
Concentrate
0.38 m3
(100 gal)
Stainless
Steel
Tank
5 Micrometer
Cartridge
Filter
Booster
Pump
High Pressure
Pump
Concentrate
Flow
Control Valve
FeedPresure
Control Valve
Shutoff
Valve
Figure 1. Flow diagram of CHWA 19m3/day reverse osmosis research units.
brackish water for domestic use for less
than two decades. RO systems typically
operate at 5,520 kPa (800 psig) for
seawater applications (35,000 mg/L total
dissolved solids [IDS]) and at 2,760 kPa
(400 psig) for brackish water applications
with IDS ranging from 1,000 to 15,000
mg/L. In the late I970's, tremendous
progress was made in membrane
technology during which advancements
not only occurred with the traditional
2,760 kPa (400 psig), 90% TDS rejecting
membranes, but even more significantly
with the reduced pressure membranes
that require approximately 1,380 kPa
(200 psig) to achieve TDS rejection in
excess of 90%. These membranes also
operate in a wider range of feedwater pH
and thus are capable of increasing
applications. The major advantage,
however, is the reduced energy
requirement and therefore lower
operating cost.
RO is effective for the removal of
most dissolved solids. Extensive studies
have been conducted to ascertain the
efficacy of RO to reject common water
constituents such as sodium, chloride,
sulfate, TDS, calcium, etc., however, very
limited experimentation has been
performed to evaluate the effectiveness
of RO to remove many of the heavy
metals and other inorganic contaminants
listed in the National Interim Primary
Drinking Water Regulations (NIPDWR).
The investigations have generally
consisted of laboratory studies, and most
results have not been verified on either a
pilot plant or full-scale level.
The primary objective of this research
project was to determine the rejection of
most of the inorganic contaminants listed
in the NIPDWR using several state-of-
the-art RO membrane elements.
Several other contaminants being
considered for future regulations were
also studied. Because of various
problems associated with the specific
chemistry of the raw water, some
contaminants were not investigated. This
project was a continuation of a similar
project reported by Huxstep in 1981.
Reverse Osmosis Test System
The reverse osmosis pilot plant
system was housed in the Charlotte
Harbor Water Association, Inc. (CHWA)
water treatment plant facilities located in
Harbour Heights, Florida. The sys
(Figure 1) consisted of a 19 m3/
(5,000 gal/day) reverse osmosis mo<
with a high pressure pump, a 0.38
(100 ga!) stainless steel tank with a
pressure pump as a feedwater sou
pretreatment in the form of 5 pm filter
and a cooling unit for temperal
stabilization. The system was alte
from a standard flow configuration
no recirculation to a continm
recirculation mode of operation
returning both the product <
concentrate waters to the feedwi
holding tank. The RO test system he
single fiberglass reinforced pla
pressure vessel into which a single
cm membrane element could be loac
This aspect of the system was <
adopted to handle other size press
vessels because several of the
membranes were provided with their i
pressure vessel.
Test Water
Initially, the test water was the
water used by CHWA with a TDS
1,900 to 2,000 mg/L. However, a
experiencing some problems caused
-------
'he relatively high sulfate content (550
.ng/L) of this well water, the test water
was changed to CHWA finished potable
water (IDS 600 to 800 mg/L). CHWA
finished water was used in almost all
cases except for the radium and uranium
tests.
The test water for the radium
experiments was CHWA raw water
containing natural radium. Well water
containing naturally occurring uranium
was obtained for the uranium tests from
a small community in southern Florida.
RO Membranes
Five RO membranes were used in the
study: (I) Toray SC 3100; (2) Filmtec BW
30-4021; (3) Dow 5K; (4) Dupont B-9
(0440-42); and (5) Hydranautics P/N
4040 LSY-IFC. Each membrane
element was tested under conditions as
close as possible to the manufacturer's
specifications for the element. Thus, the
product water flow rates, feed water
pressures, and the specific recoveries
differed between membrane elements.
Procedures
The basic test procedure consisted of
spiking the test water with contaminant
and operating the system according to
manufacturer's operating specifications.
Initially, the test water was spiked with
one contaminant and the system was
operated continuously for 6 to 8 hrs per
day during a regular, 5-day work week.
Because this schedule required an
inordinate amount of time to complete a
full series of tests with all the inorganic
contaminants, the program was
shortened by combining 2 or 3
contaminants. Each contaminant of a
contaminant group was generally tested
over a time period sufficient to permit the
collection of 12 or more samples per test
run. Occasionally a test was repeated to
verify unrealistic or inconsistent data,
particularly in the case of mercury, where
analytical results were quite varied.
During the last study with the
Hydranautics membrane, several short
term tests were added for copper, nitrite,
and molybdenum. These tests provided
limited data of only 2 to 8 samples.
The operation of the RO system was
monitored by direct and continuous
measurements of feed water pH, and
product water and concentrate flows.
Pressure gauges installed on the
feedwater, product water, and
concentrate streams were referenced on
an hourly basis during each test run,
Performance data were collected
immediately prior to test water sample
collection. Due to feedwater temperature
fluctuations, the desired system recovery
tended to drift and, therefore, very
frequent fine-tuning of the feed and
concentrate flows was necessary.
Test water samples consisted of
feedwater, product water, and reject
water. All three samples were collected
at each sampling in the order of product
water, reject water, and then feedwater
so as not to disturb the system recovery
by lowering the feedwater pressure prior
to product water and reject water
sampling.
The amount of water collected varied
according to the analyses to be
performed but was generally around 1 L.
Routine chemical analyses were
performed onsite immediately following
sample collection and consisted of pH,
total hardness, and chloride. Analyses of
the spiked inorganic contaminants, such
as fluoride, nitrate, arsenic, and
selenium, were conducted by an
approved private laboratory or by the
EPA laboratory in Cincinnati. Both
analytical laboratories followed standard
laboratory quality control procedures in
analyzing the water samples.
Results
A summary tabulation of test results
for each membrane is shown in Tables 1
through 5. Some operation control
problems occurred during limited periods
of time with several of the membranes,
therefore, direct comparison of
membrane results for specific
contaminants is not recommended.
Although the majority of summary results
are averages of about 12 samples, a few
values are averages of only 2 to 8
samples. These results are from tests
that were added at the end of the project
(copper, molybdenum, and nitrite).
A review of the TDS removal data for
all the membrane tests generally showed
very steady removals except for about a
30-day period (10 to 40 test days) for
the Filmtec membrane when the TDS
rejection slowly declined from around
98% to about 85%. After test day 40,
removals returned to around 98% for the
remaining 35 days of testing. The reason
for the slow decline is not known, but
because of it, removals of the specific
contaminants tested during this period
are probably lower than what could be
achieved under proper operating
conditions.
Summary
Considering the test data from all five
membranes as a whole, Table 6 shows
the contaminants (natural and spiked)
grouped into three categories according
to removal capability:
(1) Highly removed (above 95%);
(2) Moderately removed (85% to 94%);
and
(3) Poorly removed (below 85%).
Wide variation in removals occurred
with four contaminants: As + 3, Hg, F, and
NOs. Because nitrite tests were limited to
a 2-day test with one membrane, no
general conclusion on variability for nitrite
removal can be made. The variation in
removals of these contaminants occurred
among membranes and with each
membrane test. The reason for the
variation is concluded to be the
chemistry of the contaminants and water
matrix, membrane material, and test
conditions. In the case of mercury,
analytical procedures may also have
contributed to the variation in results.
The full report was submitted in
fulfillment of Cooperative Agreement No.
CR-808358 by the Charlotte Harbor
Water Association, Inc., under the
sponsorship of the U.S. Environmental
Protection Agency.
-------
Table 1.
Run
Days
1-13
16-32
34-35
36-57
67-71
72-77
78-89
92-97
100-104
1-104
1-104
1-104
Summary of Contaminant Removal with Toray Membrane
Wn nf
Contaminant Samples
F
N03(N)
As(-i-3)
Cd
Hg
Cr( + 3)
As( + 3)
SeC* 4)
As( + 3)
Pb
U
Cr( + 6)
As( + S)
Se( + 6)
TDS
Hardness
Chloride
Avg. recovery
Avg. feed
Avg. feed
Table 2
Of in
r\ui i
Days
1-10
11-20
21-37
39-49
50-62
63-66
67-70
71-74
1-74
1-74
1-74
water pressure -
water flow
26
35
4
46
0
0
10
10
11
12
-
6
12
12
-
-
-
9.8%
7970 Wa (284
Feedwater Concentration (mg/L)
Min
3.0
1.7
0.03
0.02
-
-
0.05
0.12
0.15
0.24
-
0.31
0.12
0.26
-
-
-
psig)
Max
10.0
25.3
0.34
0.54
-
-
0.68
0.74
0.68
1.3
-
0.96
0.74
1.0
-
-
-
Avg
6.1
11.8
0.14
0.23
-
-
0.30
0.33
0.30
0.55
-
0.60
0.35
0.61
-
-
-
Percent Resection
Min Max
73 94
35 82
58 70
95 99
-
-
44 79
96 99
46 76
97 99
-
97 98
97 >99
99 >99
94 97
97 99
90 94
Avg
90
69
63
99
-
-
66
97
64
98
-
97
99
>99
95
99
93
25 Umin (6.7 gal/min)
Summary of Contaminant Removal with Filmtec Membrane
Contaminant
F
As( + 3)
Se( + 4)
N03(N)
Pb
As( + 5)
Sef* 6;
Cr( + 6)
Cd
Hg
Cr( + 3)
N03(N)
Pb
Cd
Hg
Cr( + 3)
U
TDS
Hardness
Chloride
Wn nf
ITU. \JI
Samples
22
7
21
20
32
5
16
9
11
10
0
0
8
9
0
9
7
-
-
-
Feedwater
Min
8.4
0.04
0.02
12.8
0.04
0.10
0.58
0.04
0.28
0.002
-
-
0.19
2.5
-
0.05
0.533
-
-
-
Concentration
Max
10.2
0.18
0.08
14.3
0.13
0.47
2.6
1.3
0.36
0.109
-
-
7.32
2.6
-
0.29
0.879
-
-
-
(mg/L)
Avg
8.9
0.10
0.04
13.7
0.07
0.26
1.2
0.73
0.32
0040
-
-
0.4 1
2.6
-
0.72
0.682
-
-
-
Percent Rejection
Min Max
72 92
55 83
>85 >96
77 78
65 94
98 >99
96 >99
87 >99
Avg
83
69
-
75
89
99
99
97
>99 > 99 > 99
60 89
-
-
78 >99
99 >99
-
94 98
99 99
84 99
86 99
81 96
78
-
-
97
99
-
96
99
95
98
92
Avg. recovery
Avg. feed water pressure
Avg. feed water flow
- 70.4%
- 7378 kPa (191 psig)
- 15 Umin (4.0 gallmin)
-------
Table 3.
Days
1-17
78-26
27-34
34-47
42-50
57-63
64-67
68-69
70-77
72-73
7-73
7-73
7-73
Summary of Contaminant Removal with Dow Membrane
Contaminant
F
N03(N)
Pb
Cd
Hg
Cr( + 3)
As(-i-5)
Se( + 6)
Cr( + 6)
As( + 3)
Se( + 4)
As( •*• 3)
Se( + 4)
U
Hg
Ra(pCilL)
Hg
As( + 3)
TDS
Hardness
Chloride
Avg. recovery
Avg. feed water pressure
Avg. feed water flow
Table 4.
Run
iUJ'l
Days
1-5
6-10
11-16
17-22
23-28
29-34
34-35
36-47
42-43
7-43
7-43
7-43
7-43
7-43
Summary of
Contaminant
F
Cd
Hg
Cr( + 3)
As( + 5)
Se( + 6)
Cr( * 6)
Pb
N0398
99
99
99
99
Avg
91
85
96
98
74
97
98
98
96
98
99
75
98
99
64
97
76
83
96
98
93
Avg
92
99
-
99
99
99
98
> 96 > 99 > 98
93
46
97
96
96
95
97
89
92
98
86
95
84
99
97
99
95
99
99
97
99
98
94
71
98
96
98
95
98
99
94
99
96
Avg. recovery - 50%
Avg. feed water pressure - 2650 kPa (384 psig)
Avg. feed water flow - 17 Umin (4.5 gal/min)
-------
Table s.
Run
Days
1-3
4-6
7-9
10-13
14-16
17-20
21-23
24-25
26-27
28-29
1-29
1-29
1-29
1-29
1-29
Summary of Contaminant Removal with Hydranautics Membrane
Contaminant
Cd
Hg
Cr( + 3)
F
As( + S)
Se( + 6)
Cr( + 6)
Mo
N03(N)
Pb
U
As( + 3)
Se( + 4)
Ra(pCi/L)
Cu
NOtfN)
TDS
Hardness
Chloride
Calcium
Sodium
No. of
Samples
12
12
12
12
12
12
15
12
12
12
11
12
6
6
4
-
-
-
-
-
Feedwater
Min
1.17
0.86
14.0
1.3
2.0
4.16
1.6
18.1
1.7
0.252
0.8
1.0
7.86
4.8
4.8
-
-
-
-
-
Concentration
Max
1.36
1.46
16.0
2.0
3.2
5.96
4.3
43.1
4.8
0.310
1.1
2.4
9.83
5.9
4.8
-
-
-
-
-
(mg/L)
Avg
1.31
1.23
14.5
1.7
2.7
4.46
2.4
27.1
2.6
0.277
0.92
1.5
8.91
5.1
4.8
-
-
-
-
-
Min
99
99
98
96
99
97
88
96
98
99
5
93
96
97
90
95
98
91
96
92
Percent Rejection
Max
99
99
98
99
99
98
Avg
99
99
98
98
99
98
>98 >97
98
99
99
75
98
98
98
92
99
99
97
98
99
97
99
99
46
95
97
97
92
96
99
95
98
96
Avg. recovery
Avg. feed water pressure
Avg. feed water flow
- 10.7%
-1953 kPa (283 psig)
- 23 Umin (6.2 gallmin)
Table 6. Removal of Specific Contaminants by
Reverse Osmosis
Moderately
Highly Removed Removed
(above 95%) (85% to 94%)
Poorly Removed
(below 85%)
Arsenic V
Calcium
Cadmium
Chromium III
Copper
Lead
Molybdenum
Sodium
Radium
Selenium IV
Selenium VI
Uranium
Total Hardness
Total Dissolved
Solids
Fluoride
Chloride
Nitrate
Nitrite
Arsenic III
Mercury (Inorganics)
U. S. GOVERNMENT PRINTING OFFICE: 1988/548-158/67106
------- |