United States
Environmental Protection
Agency
Water Engineering Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA/600/S2-85/080 Aug. 1985
Project Summary
Fabrication and Pilot Scale
Testing of a Prototype Donnan
Dialyzer for the Removal of Toxic
Metals from Electroplating
Rinse Waters
Henry F. Hamil
An initial program was conducted to
develop anion-exchange membranes to
be used in the removal of copper,
cadmium, and zinc, as their complex
cyanide anions, from cyanide process
electroplating rinse waters by a Donnan
dialysis process. For these laboratory
studies, simulated rinse waters prepared
by diluting electroplating bath solutions
to the desired metal content were
utilized.
A series of anion-exchange mem-
branes based on radiation grafted poly-
ethylene films were prepared. The graft-
ing monomers used were vinylpyridines
or vinylbenzyl chloride. These grafted
membranes were converted to anion-
exchange membranes by quaternization
with alkyl halides or trialkylamines,
respectively. The series of membranes
exhibited varying ion-exchange capaci-
ties and varying hydrophilicity.
A follow-on program was conducted
to fabricate a prototype Donnan dialyzer
to be used in testing the previously
developed anion exchange membranes.
The prototype dialyzer was to be evalu-
ated in actual electroplating shops in
order to determine its engineering and
economic feasibility for these rinse
waters.
Synthesis of the required quantity of
membranes for the dialyzer produced a
product membrane which was unstable
in the dialyzer stripping solution. At-
tempts to resolve the instability problem
were unsuccessful. An alternate mem-
brane, therefore, was selected for use in
the prototype dialyzer. The dialyzer was
fabricated and laboratory tests were
carried out prior to field evaluation. A
series of techncial problems were en-
countered with the plate and frame
hardware for the dialyzer. The most
serious involved development of inter-
nal leaks between the rinse water and
stripping solution. These problems in-
dicated that a major redesign of the
dialyzer was required. After considera-
tion of financial and time requirements
for the redesign effort, a decision was
made to terminate the program.
This Project Summary was developed
by EPA's Water Engineering Research
Laboratory, Cincinnati. Ohio, to an-
nounce 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
In the past several years, legislation
has been passed at state and federal
levels leading to the limiting of allowable
emissions of many trace metals in water
from industrial sites. Among the metals
are copper, zinc, nickel, chromium, cad-
mium, lead, and mercury. Removal of
these metals, present as ions, from the
untreated industrial outflows will there-
-------
fore be required. In certain cases, the
removal may take place by using various
means such as reverse osmosis, dialysis,
and electrodialysis. All of these methods
have certain advantages and disadvan-
tages.
An approach which incorporates the
advantages of ion-exchange systems is a
membrane transfer process known as
Donnan dialysis. This process is basically
a continuous ion-exchange procedure
utilizing an ion-selective membrane to
establish a Donnan equilibrium between
the two solutions of electrolytes separated
by the membrane. In the case of ananion-
exchange membrane, the cations in the
two solutions are prevented from inter-
diffusing across the membrane, but the
anions will redistribute themselves be-
tween the two solutions until equilibrium
is reached and the ratios of all similarly
charged anions are equal. The driving
force for anion exchange is the system's
displacement from the equilibrium ratios
and can be controlled by manipulation of
the solution concentrations.
The simplest Donnan dialysis cell com-
prises an ion-selective membrane and a
space on each side of the membrane. The
toxic heavy metal ions in a rinse water
feed can be extracted across the mem-
brane and concentrated in a stripping
solution. The only energy required in
such a process is to pump the feed and
stripping solutions across the cell. Large
hydraulic pressures as required by re-
verse osmosis or large electric current
flows as required by electrodialysis are
not required in Donnan dialysis. A com-
pact Donnan dialysis stack can be de-
signed to contain many membrane-spacer
units to economically treat large quanti-
ties of rinse water.
The application of Donnan dialysis is
being extended to the removal of trace
metals front electroplating wastes. Cad-
mium, copper, and zinc are important as
they are present in electroplating wastes
as their complex cyanide anions. Remov-
al of these trace metals will therefore
require the development of suitable
anion-exchange membranes, in contrast
to the removal of nickel, which is present
as its free cation and could thus be
removed with cation-exchange mem-
branes. Conventional rinse water treat-
ment involves chlorination for cyanide
destruction, followed by precipitation of
the metals as their hydroxides or sulf ides.
Donnan dialysis, provided high transport
membranes are available, could provide a
simple, efficient alternative electroplating
rinse water treatment process.
The objectives of this research included
the development of optimized anion-
exchange membranes for the removal of
copper, zinc, and cadmium, as their
complex cyanide anions, from electroplat-
ing rinse waters and the fabrication and
field evaluation of a prototype Donnan
dialyzer for field evaluation in electro-
plating shops. It was intended for this
research to provide engineering data to
allow technical and economic analysis of
Donnan dialysis as a means of removing
toxic metals from electroplating rinse
waters.
During the course of this laboratory
study, all membrane evaluations were
conducted with simulated electroplating
rinse waters. These were prepared by
diluting plating bath formulations to give
the desired metal ion concentration (50
ppm or 500 ppm).
Experimental Procedures
Film Grafting Procedures
A desired length of the polymer films
was backed with absorbent crepe paper
toweling or cheesecloth and rolled onto a
12.7-mm aluminum pipe that was capped
at one end. The roll of film and backing
material were placed in a hydrometer jar,
which was connected to a vacuum system
and pumped down to approximately 10to
12 Torr Hg. A grafting solution of the
desired monomer in a suitable solvent
(usually benzene, methanol, or water)
was drawn into the evaluated jar. The jars
were placed on turntables in the irradi-
ation facility and exposed to a uniform
cobalt-60 source adjusted to give the
desired dose rate.
The film was removed from the reactor,
unrolled, and separated from the paper
toweling or cheesecloth. The film was
then washed in solvents suitable for re-
moving homopolymer, which was deposi-
ted on the film and allowed to dry on
paper toweling.
Film Quaternization Procedures
Vinylpyridine films were quaternized
with methyl iodide (2% in methanol, 48
hours) or with 1 -bromobutane (2% in
methanol, 48 hours). Vinylbenzyl chloride
films were quaternized with trimethyl-
amine (10% in water, 24 hours) or with
tri-n-butylamine (10% in methanol, 48
hours). After quaternization, the films
were washed with 5% hydrochloric acid,
washed with deionized water, and air
dried.
Film Characterization
The membranes were characterized for
equilibrium water content, ion-exchange
capacity, osmotic water transport, and
metal complex anion transport. The mem-
branes were pneumatically leak tested.
The metal complex anion transport rate
constants were determined on a Donnan
dialysis system. A schematic of the test
system is shown in Figure 1. The feed
solution (simulated rinse water) was
pumped through the cells on a once-
through basis to waste, and the strip
solution was pumped through the system
to the strip reservoir for recycle. The rate
constant for metal ion transport was
calculated using the expression:
k =
1n
where C0 is the concentration of metal ion
in the cell inlet, C is the metal concen-
tration in the cell outlet, and t is the
residence time in the cell.
Initial membrane evaluations indicated
that flow rates were too short to give
metal complex anion removal rates cap-
able of being accurately determined. The
membrane test system was modified to
provide cell contact times on the order of
0.5 to 0.7 minute. With the modified test
system, variable metal ion transport rates
h H h -I h H
T F L J F L J F
® <£>
A - Feed Reservoir
B - Strip Reservoir
C - Strip Pump
D - Feed Pump
E - Sample Points
F - Test Cells
C - Flow Meters
Waste
Figure 1. Membrane test system.
-------
End Plate Flow Spacer Membrane Flow Spacer End Plate
'S
-e
Feed
Strip to
Recycle
Effluent
to Waste
Strip
Note: The assembly bolt holes in the end plates are not shown.
Figure 2. Details of dialysis stack components.
were encountered during the initial period
of each run until the system came to
steady-state conditions.
Prototype Dialyzer Design
The prototype dialyzer was based upon
a plate and frame design and was a
scaled-up version of a design reported in
the literature. The unit consists of two
end plates, flow spacers, and sheet
membranes. The flow spacers were fab-
ricated using polypropylene mesh. The
border of each spacer was impregnated
with silicone rubber and was faced on
each side with neoprene rubber gaskets.
Entry and exit ports were cut into the
gaskets so inlet and outlet manifolds
were formed upon assembly of the dial-
ysis stack. Holes were also cut into the
gasketed edges of the spacers to accom-
modate locating pins. Matching ports and
locating pinholes were cut in the mem-
branes. It was found necessary to use a
liquid rubber adhesive to assemble the
Table 1. Membrane Transport Data
unit to avoid leaks. Details of the dialysis
stack component design are shown in
Figure 2.
The assembled dialysis unit used two
positive displacement gear pumps with
variable speed electric motors. Each feed
pump was connected to a 5-micron in-
line filter, a flow meter, a pressure gauge
and was connected to the appropriate
inlet port on the dialyzer stack. The entire
unit was mounted on a base designed for
handling with a forklift. The assembled
unit is shown in Figure 3.
Results and Discussion
Several membranes were evaluated for
metal ion transport. Results are shown in
Table 1.
The trend that is readily apparent in the
data in Table 1 is a positive correlation of
metal ion transport rate with ion-ex-
change capacity. Increased ion-exchange
capacity appears to lead to increased
metal ion transport rates. Additionally, it
Membrane
number
E11Q4
E12Q4
E12Q5
E14Q4
E15QJ
E16Q1
E16Q2
E18Q1
E18Q2
Ion-exchange
capacity
meg/dry g
2.5
2.6
1.1
0.7
1.1
2.2
1.2
2.2
1.4
Osmotic water
flow rate
mL/hr/cm2*
0.040
0.053
0.010
0.010
0.012
0.102
0.016
0.054
0.009
Metal complex anion
transport rate constant**
Cu
3.1
2.3
0
1.1 (1.1ft
1.0
2.5
0
2.2
O
Cd
2.8
2.2
--
0.3
0.4
--
--
1.4
-
Zn
1.9
2.9
--
0.2
0.1
--
--
0.6
-
•0.2N NaCI versus deionized HtO—cell effective area 122 cm2.
* "Feed-nominal 50 ppm in metal of interest; stripping solution 0.2N NaCI.
^Duplicate determinations of the rate constant.
was determined that there was a positive
correlation between osmotic water flow
rate and ion-exchange capacity.
The four membranes that showed the
highest metal ion transport rate constants
when operating with low levels of metal
ion (about 50 ppm) in the feed were
selected for further evaluation with high-
er concentrations of metal ion in the feed.
These four membranes, whose properties
are shown in Table 2, all have ion-
exchange capacities in excess of 2 milli-
equivalents per gram. Two membranes
are based on vinylpyridine-grafted poly-
ethylene, and two are based on vinyl-
benzyl chloride-grafted polyethylene film.
The results obtained with the four mem-
branes are presented in Table 3.
All four membranes showed good
transport rate constants with the low
level feed solutions. However, Mem-
branes E12Q4 and E16Q1 showed signif-
icantly lower rate constants with the high
level feeds. Membrane E11Q4 showed
decreased rate constants with the high
level feed, but the rate constants still are
acceptably high.
This membrane was selected for use in
the prototype dialyzer. Preparation of a
sufficient quantity for dialyzer fabrication
was undertaken.
The initial synthesis utilized 11 rolls of
film. The film was grafted and quaternized
as described above. Laboratory evaluation
indicated nonuniform grafting was ob-
tained.
A second grafting run utilized six rolls
of film. Infrared spectra of film samples
indicated uniform grafting was obtained.
These films were quaternized and sub-
jected to laboratory evaluation.
It was found that the metal ion transport
rate constant decreased with time. Initial
values of 1.4-1.6 mirf1 were obtained,
but these values declined to <0.5 min'1 in
several hours of dialyzer operation.
Infrared spectra of the unused samples
of the membranes showed very strong
absorption bands for the quaternized
polyvinylpyridine.
Sa mples of the quaternized membrane
were washed with 10% hydrochloric acid
and with 10% sodium hydroxide. Acid
washing produced no change in the infra-
red spectra, but caustic washing led to
loss of most of the absorption band
intensity of quaternized polyvinylpyridine.
It was concluded that the films had not
grafted appreciably. The strong infrared
absorption bands were attributed to oc-
cluded polyvinylpyridine homopolymer
within the film. After quaternization, the
resulting water-soluble polyvinyl-N-
-------
methylpyridinium iodide leached out of
the film.
Inasmuch as good membranes were
prepared on the previous project by
grafting 4-vinylpyridine onto polyethylene
followed by quaternization with methyl
iodide using the above procedures, a
review of possible differences in this
work and the previous work was under-
taken. The results of this review follow.
1. Polyethylene Film—The film used
in the previous and present studies
was from a special lot purchased
several years ago.
2. Cheesecloth Interleaving—The
cheesecloth used in the present
and previous studies was from a lot
purchased several years ago.
3. Solvent—ACS grade methanol was
used as the grafting solvent. GC/
MS analysis of the methanol did not
reveal any significant impurities
which might affect the grafting
reaction.
4. Monomer—Monomer on the cur-
rent program was from a different
source than the monomer on the
previous program. GC/MS analysis
indicated 99.8% plus 4-vinylpyri-
dine with no differences in impuri-
ties (mostly alkylpyridines).
5. Radiation Dose Rate—The dosim-
etry at SwRI had been done for
several years using the Bausch and
Lomb cobalt chip technique. Dos-
imetry was changed to lithium
fluoride crystal dosimeters when
Bausch and Lomb discontinued
production of cobalt glass chips. A
comparison of dosimetry measure-
ments was made using a few
remaining cobalt glass chips and
the lithium fluoride dosimeters to
ensure that dose rates in the pre-
vious work and current were the
same. Identical results were ob-
tained.
A series of experi ments were conducted
in an attempt to resolve the problem of
membrane stability. Study parameters
involved included the effect of source of
the monomers, irradiation dose rate and
total dose, inclusion of small amounts (1 -
4%) of other monomers known to enhance
grafting in the grafting solution, and
crystalline/amorphous ratio of the poly-
ethylene. None of these parameters af-
fected the stability of the product mem-
branes.
A peer review was conducted to assess
the status of the program and to make
1
Figure 3. Prototype Donnan dialysis unit.
recommendations for further work. The
review group included two representa-
tives of the U.S. Environmental Protection
Agency, a representative of the American
Electroplaters Society, two industry con-
sultants, and two representatives of
Southwest Research Institute. It was
concluded that the reasons for the insta-
bility of all 4-vinylpyridine-grafted mem-
branes prepared subsequent to the
synthesis of Membrane E11Q4 were not
known. The peer reviewers also con-
cluded that while the above-described
attempts at resolving the problem were
reasonable and diligently pursued, further
efforts at resolving the problem were not
justified at the present time. It was
recommended that no further work on the
problem of membrane instability be
undertaken. It was further recommended
that the best available, stable SwRI mem-
branes be compared with representative,
commercially available membranes. The
best membrane available, either SwRI
produced or commercially available,
should be selected. The prototype should
be fabricated using this membrane and
the field evaluation conducted. Three
commercially available membranes were
obtained from the manufacturer. These
Table 2. Properties of High Transport Rate Membranes
Membrane
number
£11 Q4
E12Q4
E16Q1
E18Q1
Membrane type
(4-VP)CH3l
(4-VP/N-VP)CH3l
(VCB) (CH3)3 N
(VBC/N-VPKCH^N
Equilibrium
water content
9H-,0/g
1.08
2.45
1.52
1.00
Ion-exchange
capacity
meq/dry g
2.5
2.6
2.2
2.2
Osmotic water
flow rate*
mL/hr/cm2
0.040
0.048
0.102
0.090
*0.2N NaCI versus deionized H20—cell effective area 122 cm2.
-------
Table 3. Effect of Metal Ion Concentration on Ion Transport
Metal ion transport rate. min~
Low concentration'"
Membrane
£)JQ4
E12Q4
E16Q1
E18Q1
Cu
3.1
2.3
2.5
2.2
Cd
2.8
2.2
1.6
1.4
Zn
1.9
2.9
0.7
0.6
High
Cu
1.5
0.7
0.6
1.41
concentration"*
Cd
1.5
0.8
0.9
0.6
Zn
1.5
0.6
-t
0.4
* Nominal 50 ppm in metal ion.
**Nominal500 ppm in metal ion.
t Membrane ruptured.
membranes, along with several SwRI
membranes, were evaluated in the labo-
ratory dialysis system. Results are pre-
sented in Table 4.
SwRI Membrane SW-3-3 was selected
for use in the dialysis unit.
Six rolls of film were grafted using a
solution of 4-vinylpyridine (24.4%) and
styrene (3.6%) in methanol. After quater-
nization, samples (SW-R-3 and SW-R-5)
were evaluated. Osmotic water transport
rates were slightly lower than for SW-3-
3, and the copper transport rate constant
was about 25% higher, as shown in Table
4.
These membranes were selected for
incorporation into the prototype dialyzer.
The membranes were cut to size and leak
tested. Assembly of the dialyzer stack
was completed, and the dialysis unit was
assembled.
After assembly the unit was checked
for operability by pumping water through
both the feed and stripping solution
pumps. After a short period of operation,
leaks developed in the dialyzer stack. The
stack was disassembled and examined. It
was found that the leaks developed in the
Table 4. Membrane Evaluation
flow spacer gaskets. These gaskets are a
laminated structure consisting of neo-
prene rubber/silicone rubber/neoprene
rubber, the silicone rubber serving to
impregnate and seal the periphery of the
polypropylene mesh flow spacers. The
neoprene rubber/silicone rubber inter-
faces had delaminated in some areas and
led to leaks.
The flow spacers were reassembled
using a liquid rubber adhesive. Addition-
ally, as each flow spacer-membrane
assembly was placed in the dialysis stack,
the adhesive was used to ensure good
spacer-to-spacer bonding.
The reassembled dialysis stack was
subjected to flow testing at 1 .Oto 1.2 gpm
flow rate through both the feed and
stripping solution pumps. There were no initial
problems; however, after approximately
33 hours of operation (5 to 6 hours per
day), it was noticed that the volume of
water in the striptank was increasing and
the volume of water in the feed tank was
decreasing, indicating internal leaks. The
pressure drop across the dialysis stack
was 35 to 40 psi at flow rates of 1.0 to 1.2
gpm. It was not possible to control the
Membrane
number Source Type
C- 1 Commercial VBC
C-2 Commercial VBC
C-3 Commercial 4-VP
SW-3-3 SwRI 4-VP/ST
SW-3-6 SwRI 4-VP/ST
SW-3-2 SwRI 2-VP
SW-3-4 SwRI VBC
SW-R-3 SwRI 4-VP/ST
SW-R-5 SwRI 4-VP/ST
"Deionized H20 versus 1 .ON Nad
"500 ppm Cu feed; 1. ON NaCI strip
t/Vof determined — membrane unstable
Osmotic water
transport"
mL/hr/cm2
0.026
0.017
-t
0.071
0.068
0.052
0.054
0.059
0.061
Copper removal
rate constant**
min'^
0.20
0.31
4-
0.61
0.28
0.30
0.18
0.75
0.77
VBC vinylbenzyl chloride
4-VP 4-vinylpyridine
2-VP 2-vinylpyridine
ST styrene
feed and strip pressure at identical values
with the manually controlled feed pumps.
A pressure differential between the feed
and strip of about 3 to 5 psi was normally
seen, and the leaks were from the high
pressure stream to the low pressure
stream.
It was not determined whether the
leaks were due to defects in the spacer
assemblies or due to physical defects
developing in the membranes. Since the'
spacer-membrane assemblies were bond-
ed together during reassembly of the
dialysis unit, an attempt to disassemble
the unit would probably result in a certain
amount of physical damage to the com-
ponents and would obscure the cause of
the leaks; however, it does appear that
the present design is not viable for a
dialysis unit with the dimensions em-
ployed.
Review of the project led to the conclu-
sion that a major redesign of the dialysis
unit would be required. After considera-
tion of both the financial and time factors
involved in such a redesign effort, it was
decided to terminate the project.
Conclusions
The results obtained in the initial study
indicated that anion-exchange mem-
branes with transport properties suitable
for removal of metal complex anionsfrom
electroplating rinse waters by Donnan
dialysis can be prepared by irradiation
grafting of polymer films.
The results further indicate that mem-
branes prepared by grafting with 4-vinyl-
pyridine followed by quaternization with
methyl iodide are superior to membranes
prepared by grafting with vinylbenzyl
chloride followed by quaternization with
trimethylamine.
The data show that the ion transport
rates across the membrane are propor-
tional to membrane ion-exchange capac-
ity, i.e., the higher the ion-exchange
capacity, the higher the transport rate.
However, equilibrium water content also
increases the increasing ion-exchange
capacity.
Two major technical problems were
encountered in fabrication and evaluation
of the prototype dialyzer. The first involved
the inability to reproduce the high trans-
port rate membrane developed in the
initial study. All attempts to prepare this
membrane led to a product which was
unstable in the basic stripping solution.
The problem of the membrane stability
was not resolved.
The second problem encountered in-
volved the development of internal leaks
-------
in the dialysis unit. The exact cause of
these internal leaks could not be ascer-
tained but probably was due to insufficient
support for membranes under the applied
hydraulic loads resulting in damage to the
membranes.
As a result of these problems, no field
evaluation data on metals removal from
electroplating rinse waters were obtained.
Therefore, no technical or economic
evaluation of Donnan dialysis for this
application can be made.
Recommendations
The data obtained in the initial study
indicate that the best membranes pre-
pared in this study show acceptable
performance for removal of metal-cyanide
complex anions from simulated electro-
plating rinse waters under laboratory
conditions.
It was recommended that further
studies be conducted to evaluate the
most promising membranes under field
conditions.
A prototype dialyzer would be evaluated
on electroplating rinse waters in com-
mercial plating shops to obtain engineer-
ing data. These data would allow technical
and economic evaluation of Donnan
dialysis as a means of controlling effluent
levels of copper, cadmium, and zinc from
commercial plating lines. It would also
provide field experience on membrane
life and maintenance. Quantities of spent
stripping solution would be available for
characterization and study as to the best
means of disposal.
The results of the follow-on study did
not allow the technical or economic eval-
uation of Donnan dialaysis as a pollution
control device in electroplating shops.
The lack of success of this program is
attributable to inability to reproduce high
transport rate membranes and to mechan-
ical problems with the dialysis unit. The
dialyzer problems were probably due in
part to the complexity of the dialyzer
design.
Any future work in this area would
benefit from careful consideration of
dialyzer design. Tube- and shell-type
units, fabricated from bundles of small
diameter ion-exchange tubing sealed into
a tubular housing at the ends, would
minimize seal problems. Procedures for
isolating leaks within individual ion-
exchange tubes followed by plugging of
any leaking tubes could provide a rela-
tively simple and easy means of elimina-
ting internal leaks. Finally, the technology
for preparing ion-exchange tubular
membranes by irradiation-initiated graft-
ing of small diameter, thin wall polyolefin
tubing exists and could be readily opti-
mized. This approach should provide a
dialyzer which is more easily constructed
and maintained and is less susceptible to
internal leaks. Consideration of a system
of this type is recommended for consider-
ation in any future program in this area.
The full report was submitted in fulfill-
ment of Cooperative Agreements
CR807456 and CR809761 by Southwest
Research Institute under the partial
sponsorship of the U.S. Environmental
Protection Agency, the American Elec-
troplaters Society, and Southwest Re-
search Institute.
Henry F. Hamil is with Southwest Research Institute, San Antonio. TX 78284.
M. Lynn Apel is the EPA Project Officer (see below).
The complete report, entitled "Fabrication and Pilot Scale Testing of a Prototype
Donnan Dialyzer for the Removal of Toxic Metals from Electroplating Rinse
Waters. "(Order No. PB 85-227 890/AS; Cost: $8.50. subject to change) will be
available only from:
National Technical Information Service
5285 Port Royal Road
Springfield, VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Water Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
irU. S. GOVERNMENT PRINTING OFFICE: 1985/557-111/20664
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
BULK RATE
POSTAGE & FEES PAID
EPA
PERMIT No. G-35
Official Business
Penalty for Private Use $300
EPA/600/S2-85/080
3063240 *t,
LOU U TILLEY
RFGIDN V EPA
LIBRARIAN
230 S D6
CHICAGO
ST
IL 60604
------- |