&EPA
United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
EPA-600/2-80-059
April 1980
Research and Development
Laboratory and Field
Evaluation of IMS-100
Reverse Osmosis
Membrane
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-80-059
April 1980
LABORATORY AND FIELD EVALUATION
OF NS-100 REVERSE OSMOSIS MEMBRANE
by
Kenneth J. McNulty, Donald C. Grant,
John R. Harland and Robert L. Goldsmith
Walden Division of Abcor, Inc.
Wilmington, Massachusetts 01887
for
American Electroplaters' Society, Inc.
Winter Park, Florida 32789
Grant No. R803753
Project Officer
Mary K. Stinson
Industrial Pollution Control Division
Industrial Environmental Research Laboratory
Edison, New Jersey 08817
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory-Cincinnati, U.S. Environmental Protection Agency, and approved for
publication. Approval does not signify that the contents necessarily reflect
the views and policies of the U.S. Environmental Protection Agency, nor does
mention of trade names or commercial products constitute endorsement or
recommendation for use.
ii
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on
our health often require that new and increasingly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory-
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
This report is a product of the above efforts. It was undertaken to
demonstrate the effectiveness and economic feasibility of using reverse
osmosis for closed-loop control of metal finishing rinse wastes uder actual
plant conditions. The reverse osmosis system concentrates the chemicals for
return to the processing bath while purifying the wastewater for reuse in the
rinising operation. The results of the report are of value to R & D programs
concerned with the treatment of wastewaters from various metal finishing,
non-ferrous metal, steel, inorganic, and other industries. Further informa-
tion concerning the subject can be obtained by contacting the Metals and
Inorganic Chemicals Branch of the Industrial Pollution Control Division.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
-------
ABSTRACT
Laboratory-'life tests were conducted with B-9 and NS-100 reverse osmosis
(RO) membranes treating zinc cyanide plating solution at ten percent of bath
strength. The B-9 membrane was degraded by the high pH of the solution,
which was beyond the upper pH limit (pH 11) recommended for this membrane.
The NS-100 membrane showed little deterioration in performance over the first
500 hours of operation. A reduction in permeate flux and rejection after 500
hours appeared to be caused by the precipitation of salts that resulted from
operation in the closed loop test system and would not be expected in actual
field operation.
At the New England Plating Company, field tests were conducted treating
rinsewater from the zinc cyanide plating operation with a module of seven
tubular NS-100 reverse osmosis membranes. Conductivity rejections of 80-96
percent and zinc rejections of 96-99 percent were measured at flux levels of
0.20-0.37 m3/m2/day (5-9 gal/ft2/day). During 2,300 hours of exposure
to the rinsewater, the membrane showed no degradation in performance as
determined by NaCl performance tests and standard tests with a solution of
zinc cyanide plating bath diluted to five percent of bath strength.
Because of the high cost per unit membrane area of tubular RO modules, a
number of attempts were made to fabricate a NS-100 spiral-wound module.
These attempts were all unsuccessful, and it is concluded that a more exten-
sive development program will be required before the NS-100 membrane can be
offered commercially in an economically attractive configuration. Recently,
a new membrane similar to the NS-100 has been developed in a spiral-wound
configuration. This membrane, designated PA-300, shows promise of super-
ceding the NS-100 membrane for plating waste applications.
This report was submitted in fulfillment of Grant No. R803753 by the
American Electr'oplaters' Society, Inc. under the sponsorship of'the U.S.
Environmental Protection Agency. This report covers the period June 1, 1975
to April 15, 1976 and work was completed as of June 1, 1978.
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables T vi
Acknowledgment vii
1. Introduction 1
2. Conclusions 3
3. Recommendations 4
4. Laboratory Life Tests with NS-100 and B-9 Membranes . . 5
5. Field Test of the NS-100 Membrane 13
6. Fabrication of NS-100 Spiral-Wound Modules 26
References 27
-------
FIGURES
Number Page
1 Flow schematic for laboratory life test system .... 6
2 Schematic drawing of a B-9 mini-permeator 7
3 Rejections and productivities vs. operating time for
the NS-100 and B-9 membranes 10
4 Abcor spiral wound membrane module 14
5 Flow schematic for field test system 15
6 Rejection for membrane Jl at 5% of bath strength ... 18
7 Rejection for membrane J2 at 5% of bath strength ... 19
8 Rejection for membrane J3 at 5% of bath strength ... 20
9 Flux at 5% of bath strength 21
10 Conductivity rejections obtained during standard tests
with sodium chloride 22
11 Fluxes obtained during standard sodium chloride
tests 24
TABLES
Number Page
1 Membrane Performance During Operation on Actual Zinc
Cyanide Rinsewater . . . .< 25
vi
-------
ACKNOWLEDGMENT
The authors gratefully acknowledge the cooperation of Mr. Bruce Warner,
New England Plating Co., Worcester, MA for providing samples of zinc cyanide
plating bath for the laboratory life test and for providing the site for
field tests of the NS-100 membrane.
Technical direction was received throughout the program from the EPA
Project Officer, Ms. Mary K. Stinson, and from the members of the American
Electroplaters1 Society Project Committee: Mr. Charles Levy, Mr. Jack Hyner,
Mr. Lawrence Greenberg, Mr. Joseph Conoby, Dr. Robert Mattair, Mr. James
Morse, Mr. Herbert Rondeau, and Mr. George Scott.
vii
-------
SECTION 1
INTRODUCTION
The discharge of used rinsewaters from metal finishing operations is a
major source of water pollution in the electroplting industry. Various
techniques are available for treating the rinsewater generated in the electro-
plating process. During recent years, increased attention has been focused
on closed-loop systems used to reclaim rinsewater from individual plating
baths. In these systems, the chemicals dissolved in the rinsewater discharge
are concentrated for return to the plating bath. The purified water produced
in the process can be reused for rinsing.
Reverse osmosis (RO) is one of several concentration techniques that can
be used for closed-loop treatment of electroplating rinsewaters. When
pressurized rinsewater (feed) is brought in contact with a semi-permeable
membrane, water passes through the membrane at a much higher rate than the
dissolved salts. The process produces a low concentration/high volume
"permeate" stream, which is recycled to the rinsing operation, and a high
concentration/low volume "concentrate" stream, which is returned to the
plating bath. Membrane performance is generally characterized in terms of
flux (the flow rate of permeate produced per unit membrane area) and
rejection (the percent concentration differecne between the feed and permeate
streams), both measured at specified conditions. Advantages and limitations
of RO for closed-loop treatment of electroplating rinsewaters have been
discussed previously (1).
Preliminary tests have been conducted (1) to assess the applicability of
the commercially available membranes (cellulose acetate and polyamide) to a
variety of different electroplating rinsewaters. While the membranes were
effective in concentrating dissolved species for all the plating baths,
membrane life was judged to be insufficient for use in certain rinsewaters
with extremes of pH or high oxidant (chromic acid) levels.
Several RO field tests have been conducted to determine the applica-
bility of RO under practical conditions. These have included:
— spiral-wound cellulose acetate for treatment of Watts-nickel
rinsewater (2).
~ hollow fiber polyamide for treatment of Watts-nickel rinsewater (3).
~ hollow fiber polyamide for treatment of copper cyanide rinsewater (4).
1
-------
These field tests indicate that either the cellulose acetate or polyamide
membrane can be used to treat Watts-nickel rinsewaters, and the economics of
this application can be quite attractive. On the other hand, treatment of
copper cyanide rinsewaters cannot be considered a proven application since,
during one of two field tests, the rate of membrane deterioration would
result in excessive costs for membrane replacement.
In general, previous work has indicated that membrane life is critical
in determining the applicability of RO to rinsewater recovery, and that pH is
an important parameter in membrane life. For the cellulose acetate membrane,
life is generally adequate over a pH range of 2.5-7; for the polyamide
membrane, life is generally adequate over a pH range of 4-11.
The objective of the work covered by this report was to, extend the
applicability of RO to the high-pH cyanide rinsewaters. In particular, zinc
cyanide was selected because of its high pH and its high volume usage in the
plating industry. Because of the pH limitations of the commercially avail-
able membranes, a non-commercial membrane, NS-100, was to be tested and
developed in a spiral-wound configuration for electroplating applications.
Previous tests with this membrane have indicated good stability over a pH
range of 1-13 (5).
-------
SECTION 2
CONCLUSIONS
Laboratory life tests indicated that the NS-100 membrane is chemically
resistant to zinc cyanide rinsewater at ten percent bath strength (rinsewater
pH = 12.0-12.5). On the other hand, the performance of the commercially
available B-9 membrane deteriorated with time, probably as a result of chemical
degradation, since the pH of the test solution was beyond the recommended
range (4-11) for this membrane.
Membrane fouling, which was observed in the laboratory life tests after
500 hours of operation, was probably the result of the buildup and precipita-
tion of a sparingly soluble salt within the closed-loop test system. This
type of fouling would not be anticipated in a practical system.
During a 2300-hour field test with a tubular NS-100 module, flux levels
of 0.20-0.37 m3/m2/day, conductivity rejections of 80-96 percent, and zinc
rejections of 96-99.7 percent, were obtained and remained stable throughout
the test. There was no degradation of the membrane as determined by a standard
NaCl solution test and a standard test with zinc cyanide solution diluted to 5
percent of bath strength.
Attempts to fabricate the NS-100 membrane in a spiral-wound configuration
were unsuccessful. A more extensive development effort will be required
before the NS-100 membrane can be offered in an economically attractive
configuration for electroplating applications.
-------
SECTION 3
RECOMMENDATIONS
Considering both the imminent commercialization of the PA-300 membrane in
a spiral-wound configuration and the similarity of the PA-300 to the NS-100 in
type and chemistry, tests should be conducted with the PA-300 membrane to
evaluate its performance and stability for treatment of zinc cyanide rinse-
waters. If the results look promising, field tests should be conducted to
evaluate the performance of spiral-wound PA-300 modules for treatment of
actual zinc cyanide rinsewaters. (These recommendations are implemented
under EPA Grant Nos. R8043311 and R805300. Full results will be included in
the final reports for these grants).
Further development of the NS-100 membrane should be suspended until the
PA-300 and other membrane materials have been evaluated for the treatment
of various electroplating rinsewaters.
-------
SECTION 4
LABORATORY LIFE TESTS WITH NS-100 AND B-9 MEMBRANES
The objective of this program was to demonstrate the feasibility of using
RO for closed-loop treatment of the rinsewaters resulting from zinc cyanide
plating operations. Because of the high pH of zinc cyanide rinsewater, a
laboratory life test was conducted to determine whether the du Pont B-9
polyamide membrane could be used in the field tests. When this program was
initiated, the polyamide membrane was the only commercially available mem-
brane that could withstand alkaline conditions (pH 4-11). The other commer-
cially available membranes, namely cellulose acetate and cellulose triace-
tate, are limited to pH values below about 8.0.
In parallel with these tests, a laboratory life test was conducted with
the NS-100 developmental membrane to determine if this membrane could be used
for the field test in the event that the B-9 membrane was degraded by the
zinc cyanide rinsewater. The NS-100 membrane has exhibited good stability
over a pH range of 1-13 (5).
EXPERIMENTAL PROCEDURE
Laboratory Life Test System
A simplified flow schematic of the test system used for the laboratory
life tests is shown in Figure 1. A positive displacement pump (Yarway
Cyclophram Model 072) was used to withdraw solution from the feed tank and
pressurize it to 2.8 x 10" N/m^ (400 psig). The flow rate to the modules
was controlled at 0.12 m3/hr (0.53 gpm) by adjusting the displacement
volume of the pump. The pump discharge pressure was measured, and pressure
pulsations were dampened by an accumulator. The modules were protected
against over-pressurization by a pressure relief valve and a high pressure
switch. In addition, the pump was protected against running dry by a low
pressure switch. Feed passed through a NS-100 tube and a B-9 mini-permeator
in series, with the operating pressure controlled by a back pressure regula-
tor. Concentrate and both permeates were returned to the feed tank (total
recycle) so that the feed concentration would not vary with time. The feed
temperature was measured at the pump suction. Frictional heat input from the
pump was removed by a cooling coil placed in the feed tank.
Membrane Modules
The B-9 polyamide membrane was in the form of a mini-permeator. The
design of a mini-permeator is shown in Figure 2. The active portion of the
permeator consists of one strand (150 filaments) of polyamide hollow fibers.
-------
Concentrate
Cooling
Water
BPR
/Positive
,Displacement
Pump
Key:
T Temperature Gauge
ACC Accumulator
PRV Pressure Relief Valve
HPS High Pressure Switch
IPS Low Pressure Switch
P Pressure Gauge
BPR Back Pressure Regulator
Figure 1. Flow schematic for laboratory life test system.
-------
Permeate
Epoxy Pot
1 Strand of
150 Filaments
Stainless
Steel Fitting
Shell
Concen-
trate
VA
^— Stainless
Steel
Fitting
Feed
Figure 2. Schematic drawing of a B-9 mini-permeator.
-------
The strand is looped as shown, and both ends are sealed in an epoxy pot.
Permeate is withdrawn from one end of the fiber strand after slicing exposes
the open fiber ends. A normal four-inch diameter permeator contains about
900,000 filaments; so the mini-permeator contains less than 2 x 10~4 of the
surface area of a full-scale module.
The NS-100 membrane is formed by treating a polysulfone membrane support
with a solution of PEI (polyethylenimine) in water, a solution of TDI (2,4-
tolylene diisocyanate) in hexane, and heat curing to cross-link the PEI.
Before reaction with the PEI and TDI, the polysulfone membrane is essentially
an ultrafiltration membrane capable of removing suspended solids and macro-
molecules. The PEI/TDI/heat-cure treatment causes a skin to form at the
membrane surface that is capable of rejecting smaller molecules and dissolved
salts. The NS-100 membrane is supported on the interior wall of an epoxy-
impregnated fiberglass tube 12.7 mm'(0.5 in) in diameter and 0.61 mm (2.0
ft) long.
Performance Parameters
Productivity—
The productivity of a given module is the rate at which permeate is
produced under specific conditions. Productivity is dependent on tempera-
ture, pressure, and feed concentration at the membrane surface. Where slight
variations in the operating conditions occurred, the measured productivity of
the mini-permeator was corrected to 2.8 x 106 N/m2 (400 psig) and 25°C
(normal operating conditions for the du Pont module) using data from the
du Pont Technical Information Manual. Flux is defined as productivity per
unit membrane area.
Conversion—
The conversion is the ratio of permeate flow to feed flow. For a module
operating at near-zero conversion (as were both modules in the life test),
the concentrations of the feed and concentrate streams are nearly the same.
Thus, the concentration on the feed/concentrate side of the membrane is very
nearly the same as the feed concentration. On the other hand, a module
operated at high conversion will produce a concentrate stream having a higher
concentration than the feed stream. In this case, the average concentration
•on the feed/concentrate side of the membrane will be substantially higher
than the feed concentration. For two modules operated at the same pressure,
temperature, and feed concentration, the flux will be lower for the module
operated at the higher conversion, since flux decreases with increasing
average concentration on the feed/concentrate side of the membrane.
Rejection—
The rejection (r) measures the degree to which plating salts are pre-
vented from passing through the membrane. The rejection depends on the oper-
ating pressure, the conversion, and the feed concentration, and is defined by:
8
-------
100 [Cf - Cp]
Cf
where Cf = feed concentration (conductivity)
Cp = permeate concentration (conductivity).
Operating Conditions
The feed solution used in the life tests was prepared by diluting a
sample of zinc cyanide plating bath from New England Plating Company to ten
percent of bath strength (volumetric basis). The nominal bath composition is
given below:
zinc (as metal) 15.0 g/1 (2.0 oz/gal)
free cyanide 19.5 g/1 (1.6 oz/gal)
caustic 75.0 g/1 (10 oz/gal)
carbonate 50-60 g/1 (7-8 oz/gal) avg
brightener 0.5% (vol)
The test system was operated continuously on the zinc cyanide feed solution,
except during interruptions for maintenance and NaCl performance tests.
The decline in rejection performance is readily indicated by the NaCl
tests, since rejection is most sensitivve for small univalent ions. The NaCl
performance tests were conducted every two days to obtain control data on the
membrane flux and rejection. After disconnecting the zinc cyanide feed tank
and draining the remainder of the system, a feed tank containing a 1,500 mg/1
solution of NaCl was connected, and the system was operated at 1.8 x 106
N/m2 (400 psig), 2.0 1/min flow rate, and 25°C until steady state was reached.
At this time the NaCl productivity and NaCl rejection were measured for each
module. The system was drained and returned to closed-loop operation on the
zinc cyanide solution.
RESULTS AND DISCUSSION
NaCl rejections and productivities for the two membranes, measured
approximately every 48 hours, are shown in Figure 3 as a function of operat-
ing time. The productivity for the NS-100 increased over the first 500 hours
from 7 to 8.5 cm3/min. However, between 500 and 600 hours, the productivity
decreased substantially to only 3.6 cm3/min. This decrease is largely the
result of membrane fouling. When the membrane "was cleaned with a two percent
citric acid solution (A), the productivity was restored to its initial value.
However, when the productivity again declined rapidly, a second cleaning with
citric acid (B) increased the productivity but did not restore it to its
initial value. After treatment with PT-B, a du Pont proprietary membrane
-------
A Cleaned with 2% citric acid at pH 4
B Cleaned with 2% citric acid at pH 4
C After treatment with PT-B soln.
D After treatment with PEI soln. (NS-100 only)
300 400 500
Operatinq Time (hrs)
700 800
10
9
8
Z 4
*->
u
a
-a ,
o 3
B.
2
1
0
NS-100
-£*•
_L
_L
t U
100 200 300 400^ 500 600 700H 800
Operatinq Time (hrs)
Figure 3. Rejections and productivities vs. operating time
for the NS-100 and B-9 membranes.
10
-------
"tightening agent," the productivity increased almost to its initial value
(C), but dropped after treatment with a PEI solution (D).
The productivity for the B-9 membrane decreased rapidly to about one-
third of its initial value and appeared to be leveling off until about 500
hours. A rapid decline in productivity was again observed between 500 and
600 hours and between 600 and 700 hours. Only modest improvements in pro-
ductivity were obtained after cleaning and PT-B treatment.
The rejection for the NS-100 was very stable over the first 500 hours,
but there was a substantial decline in rejection between 500 and 600 hours
and between 600 and 700 hours. This decline corresponds exactly with the
productivity decline discussed above. Citric acid cleaning and PEI treatment
restored the rejection of the NS-100 to its initial value.
The rejection for the B-9 decreased substantially over the life test and
could not be restored by citric acid cleaning or PT-B treatment. Even over
the first 500 hours, where the rejection for the NS-100 remained constant,
the rejection for the B-9 declined substantially.
One of the most striking features of these data is the difference in
membrane performance before and after 500 hours of operating time. Appar-
ently, some type of membrane fouling occurred after 500 hours. The cleaning
solution (two percent citric acid at pH 4) was selected to remove iron
hydroxide precipitate, which could have resulted from corrosion of stainless
steel components in the test system. However, if this were the foul ant,
performance should have declined gradually from the beginning of the life
test. A more satisfactory explanation is that some constituent in the feed
builds up to its solubility limit over the first 500 hours of operation and,
thereafter, continues to foul the membrane by precipitation. This hypothesis
is supported by the presence of a precipitate in the feed tank that was quite
evident after 720 hours and may have started to accumulate at about 500
hours. The feed solution was analyzed for carbonates, with the thought that
the dissolution of C02 could increase the carbonate level and cause preci-
pitation of Na2CO^. The analysis showed that Na2C03 was considerably
below its solubility limit; however, precipitation of some less soluble
carbonate species cannot be ruled out.
If the hypothesis is correct that the total-recycle mode of operation
resulted in an accumulation of sparingly soluble salt, the data beyond 500
hours operating time would not be characteristic of membrane performance
under actual operating conditions. Even if the bath were saturated with this
salt, the drag-out would be diluted in the rinse and concentrated in the RO
system to only about ten percent of bath strength, that is, ten percent of
saturation. Hence, the sparingly soluble salt would not accumulate in the
reverse osmosis system as it does during total recycle.
Based on the first 500 hours of operation, the decline in both flux and
rejection for the B-9 is too rapid to make this membrane economically attrac-
tive for zinc cyanide recovery. This conclusion is supported by the results
of static tests conducted by du Pont. In these tests the polyamide hollow
fiber was soaked in a zinc cyanide solution at ten percent of bath strength.
11
-------
Significant deterioration of the physical strength of the fibers was observed
during several weeks of testing.
Over the first 500 hours of operation, the flux and rejection for the
NS-100 were good. It was concluded that the NS-100 membrane has potential
for the treatment of zinc cyanide rinsewaters and should be field tested on
actual rinsewaters from a zinc cyanide plating operation.
12
-------
SECTION V
FIELD TEST OF THE NS-100 MEMBRANE
ATTEMPTS TO FABRICATE AN NS-100 SPIRAL-WOUOND MODULE
Following the laboratory life tests with the B-9 mini-permeator and the
NS-100 tubular membrane it was decided that the most meaningful field test
would be one conducted with the NS-100 membrane in the spiral-wound configur-
ation. The spiral-wound configuration is preferred for treating plating
rinsewaters since the capital and operating costs are lower, and the advan-
tages of the tubular configurations (namely, resistance to fouling by
suspended solids and operation at higher pressures) are not required in this
application. The construction details of a spiral-wound module are shown in
Figure 4.
The conditions and procedures used for formation of the NS-100 were
similar to those used for preparation of NS-100 tubular membranes (6).
Following the laboratory life test, four attempts were made to fabricate a
workable NS-100 spiral-wound module. The first three attempts failed because
of leaks in the module glue seams. Better results were obtained with the
fourth module. This module, which contained a membrane surface area of about
2.3 m^, was tested on a total-recycle system similar to that shown in
Figure 1. The module was operated on a 1,500 ppm NaCl feed solution at 2.8 x
10° N/m^, 25°C, and approximately zero percent conversion. No gross
leaks were observed, but a considerable break-in period was required in order
to achieve the final steady state flux and rejection. After 24 hours of
operation, the flux was approximately 0.53 m3/m2/day (13 gfd) and the
rejection was 75 percent. While these results are encouraging, the perform-
ance of this module is inadequate for a meaningful field test. The NS-100
membrane is capable of achieving rejections in excess of 98 percent under the
conditions at which the module was tested.
Since attempts to fabricate a NS-100 spiral-wound module were unsuccess-
ful, it was decided to proceed with field testing the NS-100 membrane in the
tubular configuration. It was reasoned that these tests would illustrate the
performance of the membrane over an extended period of exposure to rinsewater
under actual field conditions. If these tests were successful, further
development of a spiral-wound module would be warranted.
EXPERIMENTAL PROCEDURE
A flow schematic of the reverse osmosis system used to field test the
tubular NS-100 membrane is shown in Figure 5. Feed was pumped from the rinse
tank by a booster pump (Worthington Model D-520) and filtered through a
30-micron and a 1-micron cartridge filter arranged in series. The pressure
13
-------
FEED SOLUTION
FEED SOLUTION
MODULE SEAL (SEALS AGAINST THE INSIDE WALL OF A PRESSURE
, VESSEL TO FORCE THE FEED SOLUTION THROUGH THE MODULE)
PERMEATE COLLECTION HOLES
CONCENTRATE
PERMEATE OUT
CONCENTRATE
FEED FLOW
ACROSS FEf D
CHANNEL SPACER
PERMEATE FLOW (AFTER PASSAGE
THROUGH MEMBRANE INTO PERMEATE
COLLECTION MATERIAL)
COVERING
FEED CHANNEL
SPACER
MEMBRANE
PERMEATE COLLECTION
MATERIAL
MEMBRANE
FEED CHANNEL
SPACER
'ADHESION LINE
Figure 4. Abcor spiral-wound membrane module.
-------
Plated Parts
Plated Parts 8 Drag out
Drag out
Rlnsewater •——
Parts
To
Plant
Waste
Disposal
System
Booster
Pump
High Pressure
Pump
> To Plant Haste
Disposal System
Key:
T Temperature Gauge
P Pressure Gauge
ACC Accumulator
LPS Low Pressure Switch
NV Needle Valve
PRV Pressure Relief Valve
SV Sample Valve
BPR Back Pressure Regulator
F Flow Meter
FCV Flow Control Valve
Figure 5. Flow schematic for field test system.
-------
of the filtered feed was then increased to the desired operating pressure of
4.2 x 106 N/m2, (600 psig) by a positive displacement diaphragm pump
(Yarway Cyclophram Model 072). Pressure pulsations were dampened by an
accumulator located on the pump discharge. The membrane tubes were connected
in series with a flow rate of 2.5 m3/day (0.45 gpm) in each tube. The
permeate stream from the tubular NS-100 module flowed directly to the plant
wastewater treatment system. The concentrate stream passed through a back
pressure regulator, which was used to control the operating pressure. Most
of the concentrate stream was returned to the suction side of the booster
pump while a small portion (sufficient to yield a 90 percent conversion) was
bled through a control valve to the plant wastewater system.
Pressures were measured before and after the cartridge filters, to
determine when replacement was necessary, and before and after the reverse
osmosis module to determine the operating pressure and pressure drop along
the tubes. The system was protected from over-pressurization by a pressure
relief valve.
Feed and concentrate flow rates were measured using flow meters, while
the permeate flow rate from each individual tube was measured by the "bucket
and stopwatch" technique. Feed, concentrate, and permeate samples were taken
periodically for analysis. Samples were analyzed for zinc, total solids, pH,
and conductivity using the following analytical methods:
Parameter Method Procedure
Zinc Atomic absorption 301A*
Total solids Evaporation-gravimetric 208A*
pH Meter reading 424*
Conductivity Meter reading 205*
The membrane module consisted of seven tubular NS-100 membranes arranged
in series. Each tube had a diameter of 12.7 mm (0.5 in), a length of 0.61 mm
( 2 ft), and a surface area of 0.024 m2 (0,26 ft2), (tubular module total
area = 0.17 m2). Two different types of NS-100 membrane were installed in
the module: four tubes of Type I membrane cast on a 4-mil-thick polysulfone
backing, and three tubes of Type J, cast on 8-mil-thick polysulfone backing.
Membrane performance was determined by periodically interrupting the
field test and conducting a total recycle test with a standard solution at
standard operating conditions. The standard solution was contained in an
auxiliary feed tank and, after draining and flushing the system, feed was
withdrawn from the auxiliary tank and both concentrate and permeate were
returned to the tank (see Figure 1). At steady state, the flux of each tube
*Standard Method for the Examination of Water and Wastewater, 14th ed.,
American Public Health Association, Washington D.C. 1976.
16
-------
was measured and samples of feed and permeate were obtained to determine the
rejection for each tube. Following the test, the system was drained and
returned to operation on the actual zinc cyanide rinsewater.
Standard performance tests were conducted with two different solutions:
a 1,500 ppm Nad solution, and a solution of the actual zinc cyanide plating
bath diluted to five percent of bath strength. During these tests the system
was operated at 4.2 x 106 N/nr (600 psig), 25°C C77°F), and 2.5 m3/day/tube
(0.45 gpm/tube) feed rate.
Results of previous field tests have shown that performance data obtained
during operation on the actual rinsewater are difficult to interpret since
there is no way to control the concentration of the feed to the RO system.
Since feed concentration affects both flux and rejection, data obtained at
different feed concentrations cannot be directly compared to determine
whether the flux and rejection remain stable with time. This problem,
varying feed concentration, is obviated by using the standard zinc cyanide
solution and total recycle operation.
RESULTS AND DISCUSSION
During the tests, the module with seven tubular membranes was exposed to
zinc cyanide rinsewaters for 2,300 hours, with an actual operating time of
1,300 hours. The shorter operating time reflects uncontrollable system
interruptions at the site, as well as minor mechanical problems that mainly
occurred during the first 1,000 hours of membrane exposure. Permeate fluxes
and solute rejections for both types of membrane showed little decline in
performance throughout the test. The Type J membranes (8-mil backing)
displayed higher rejections than the Type I membranes (4-mil backing),
presumably because of the improved physical properties. Accordingly, only
data for the Type J membranes are presented.
The rejections obtained during standard tests with zinc cyanide rinse-
waters (plating bath diluted to five percent of bath strength) for the three
Type J tubular membrane assemblies are presented in Figures 6 through 8.
Each figure presents the rejection data obtained for one of the three Type J
tubes tested. The rejections were excellent. They appeared to increase
during the first 1,500 hours, then remained constant or decreased slightly
during the remainder of the test. Conductivity rejections were 80-96 per-
cent, total solids rejections 87-96 percent, and zinc rejections 96-99.7
percent.
Fluxes obtained during the tests at five percent of bath strength were
also very stable. They are presented in Figure 9, which indicates flux
levels of 0.20-0.37 m3/m2/day.
The conductivity rejections obtained with 1,500 ppm NaCl solutions are
given in Figure 10. The rejection appeared to decrease with time, but this
is believed to have resulted from an ion exchange or adsorption phenomenon
occurring when the membranes were in contact with the highly alkaline zinc
cyanide rinsewaters. This hypothesis was substantiated by operating the
system on a 10,000 ppm NaCl solution for 170 hours and again measuring the
17
-------
00
99.9
99.8
99.6
99.4
99.2
99
98
96
94
92
90
80
60
40
20
Key:
O Conductivity
D Total SolIds
A Zinc
200 400 600 800 1000 1200 1400
Contact Time, Hours
1600 1800 2000 2200 2400
Figure 6. Rejection for membrane Jl at 5% of bath strength.
-------
vo
o
•r"
+J
99.
99.8
99. (
99.4
99.2
99
98
96
94
92
90
80
60
40
20
0
Key:
O Conductivity
D Total SolIds
A Zinc
200 400 600 800 1000 1200 1400 1600 1800 2000 2200 2400
Contact Time, Hours
Figure 7. Rejection for membrane J2 at 5% of bath strength.
-------
ro
o
»«
ft
g
3
99.%
99. £-
99.6
99.4
99.2
99
91 -
96
80
60
40
20
0
Key:
O Conductivity
DTotal Solids
Azinc
200 400 600 800
1000 1200 1400
Contact Time, Hours
1600 1800 2000 2200 2400
Figure 8. Rejection for membrane 03 at 5% of bath strength.
-------
1.0
0.8
0.6
0.4
CM
«*£ 0.2
0.1
0.08
0.06
I
Key:
O Membrane Jl
n Membrane J2
A Membrane J3
I I
I
I
I
I
I
200 400 600 800
1000 1200 1400
Contact Time, Hours
1600 1800 2000 2400
Figure 9. Flux at 5% of bath strength.
-------
IN)
INS
o
5
99.9
99.8 -
Key:
O Membrane Jl
rj Membrane J2
A Membrane J3
Closed symbols represent data
obtained after the 1% Nad
treatment.
200 400
600 800 1000, 1200 1400 1600 1800 ,2000 2200 2400
Contact Time., Hours
Figure 10. Conductivity rejections obtained during standard tests
with sodium chloride.
-------
conductivity rejection with a 1,500 ppm Nacl solution. It was felt that by
operating the system at a high salt concentration, the adsorption or ion
exchange that had occurred could be reversed. This was indeed the case, as
the rejection for all three membranes increased after this treatment (closed
symbols, Figure 10). Perhaps further recovery of NaCl rejection might be
achieved by operating the system with higher salt concentrations for a longer
period of time.
Flux levels remained stable during the tests except for an apparent
increase after an elapsed exposure time of 1,600 hours (Figure 11). This may
indicate membrane degradation, but it is more likely to be connected with
some other phenomenon since no other data indicate degradation of the mem-
brane.
Data obtained while the system was operating on actual zinc cyanide
rinsewaters indicate high fluxes and rejections, as shown in Table 1.
Membrane fluxes were 0.25-0.41 m^/m^/day; conductivity rejections were
76-94 percent; total solids rejections were 85-87 percent; and zinc rejec-
tions were 95.5-99.3 percent. It should be noted that although the system
was operated at 90 percent water recovery, concentrations in the membrane
loop were only 2-4 times those in the feed because of the low rejections of
the Type I membranes.
Results of the 2,300-hour field test with the tubular NS-100 module
indicated adequate flux and rejection and no degradation for the Type J
membranes. It is, therefore, concluded that the NS-100 membrane can be used
for the treatment of zinc cyanide rinsewaters.
23
-------
E
«l
X
1.0
0.8
0.6
0.4
0.2
0.1
0.08
0.06
I
Key:
O Membrane Jl
Q Membrane J2
A Membrane J3
I
I
I
200
400 600 800
1000 1200 1400
Contact Time, Hours
1600 1800 2000 2200
* Figure 11. Fluxes obtained during standard sodium chloride tests.
-------
TABLE 1. MEMBRANE PERFORMANCE DURING OPERATION
ON ACTUAL ZINC CYANIDE RINSEWATER
Time (hrs)
Rinse concentration
Conductivity (ymho/cm)
Total solids (mg/1 )
Zinc (mg/1)
Membrane loop concentration
Conductivity (umho/cm)
Total solids (mg/1)
Zinc (mg/1)
Membrane Jl
Conductivity rejection (%)
Total solids rejection (%)
Zinc rejection (%)
Flux (m3/m2/day)
Membrane J2
Conductivity rejection (%)
Total solids rejection (%)
Zinc rejection (%)
Flux (m3/m2/day)
Membrane J3
Conductivity rejection (%)
Total solids rejection (%)
Zinc rejection (%)
Flux (m3/m2/day)
1,850
9,800
6,843
375
17,000
14,606
980
94.1
96.2
99.3
0.32
75.9
85.3
96.1
0.24
86.5
92.0
98.7
0.26
2,100
8,000
4,560
204
21,000
18,350
665
94.3
97.1
99.2
0.39
77.1
86.5
95.5
0.28
88.1
94.1
98.5
0.38
25
-------
SECTION 6
FABRICATION OF NS-100 SPIRAL-WOUND MODULES
Following the successful field tests with the NS-100 membrane in tubular
configuration, further efforts were directed toward fabricating an NS-100
spiral-wound module.
The preparation of NS-100 spiral-wound membrane modules based on forma-
tion procedures developed for tubular configurations required adaptations in
two main areas:
1. the casting of a continuous flat sheet polysulfone substrate on a
cloth backing (rather than the casting of discrete tubular sections)
2. the formation of the NS-100 ultrathin membrane on the flat sheet
polysulfone substrate.
A flat sheet of polysulfone, 0.91 m wide by 92 m long, was cast, and two
basic approaches to the formation of the NS-100 spiral-wound module were
attempted. In the first approach, a polysulfone module was wound and then
the NS-100 skin was formed in situ. In the second approach, the NS-100
ultrathin membrane was formed on the polysulfone substrate and then the
membrane was wound into a spiral module. One module using the first approach
and four modules using the second approach were fabricated. The best per-
formance characteristics obtained during a standard 1,500 mg/1 NaCl test were
a flux of 0.33 m3/m2/day and a rejection of 60 percent. Following these
attempts it was concluded that the additional development required to fabri-
cate a workable NS-100 spiral-wound module was beyond the resources available
to this program.
Although commercialization of a spiral-wound NS-100 module does not
appear imminent, the Fluid Systems Division of Univeral Oil Products, Inc.
(San Diego, CA) has recently developed a similar membrane, designated PA-300,
in a spiral-wound configuration. This membrane module should be commercially
available in the near future and may well supercede the NS-100 as the most
promising new membrane for treatment of zinc cyanide and other electroplating
rinsewaters. Tests being conducted at the Walden Division of Abcor, Inc.,
under a separate EPA grant, indicate excellent performance characteristics
for PA-300 flat-sheet membranes during exposure to zinc cyanide rinsewaters.
Based on these tests and the results presented in Sections 4 and 5 for the
NS-100 membrane (given its similarity to the PA-300), field testing of a
spiral-wound PA-300 membrane on zijic cyanide rinsewater is recommended. Such
a field test is being conducted under a separate EPA grant; results will be
published in a separate report.
26
-------
REFERENCES
1. Donnelly, R.6., R.L. Goldsmith, K.J. McNulty, D.C. Grant, and M. Tan.
1976. Treatment of electroplating wastes by reverse osmosis. EPA-600/2-
76-261, U.S. Environmental Protection Agency, Cincinnati, Ohio. 96 pp.
2. Golomb, A. 1973. Application of reverse osmosis to electroplating waste
treatment, Part III. Plating 60(5):482-486. 1977.
3. McNulty, K.J., R.L. Goldsmith, and A.Z. Gollan. 1977. Reverse osmosis
field test: treatment of Watts nickel rinse waters. EPA-600/2-77-039,
U.S. Environmental Protection Agency, Cincinnati, Ohio. 29 pp.
4. McNulty, K.J., R.L Goldsmith, A.Z. Gollan, S. Hossian, and D. C. Grant.
1977. Reverse osmosis field test: treatment of copper cyanide rinse
waters. EPA-600-2-77-170, U.S. Environmental Protection AGency, Cincin-
nati, Ohio. 89 pp.
5. Rozelle, L.T., J.E. Cadotte, C.V. Kopp, and K.E. Cobian. 1973. NS-1
membranes: potentially effective new membranes for treatment of washwater
in space cabins. ASME Paper 73-ENAS-19. ASME, New York, N.Y. 5 pp.
6. Zakak, A., P. Hoover, A.Z. Gollan, and R.L. Goldsmith. 1975. Development
of a low-cost tubular reverse osmosis module for one-pass sea water
desalting. U.S. Department of Interior, Office of Saline Water, OSW
Contract No. 14-30-3251.
27
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
RLPORT NO.
EPA-600/2-80-059
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
LABORATORY AND FIELD EVALUATION OF
NS-100 REVERSE OSMOSIS MEMBRANE
5. REPORT DATE
April 1980 issuing date
6. PERFORMING ORGANIZATION CODE
K
------- |