EPA-600/2-76-261
September 1976
Environmental Protection Technology Series
OF ELECTROPLATING
WASTES BY REVERSE OSMOSIS
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate instrumentation, equipment, and methodology to repair or prevent
environmental 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-600/2-76-261
September 1976
TREATMENT OF ELECTROPLATING WASTES
BY REVERSE OSMOSIS
Submitted by
The American Electroplater's Society
East Orange, New Jersey 07017
Prepared by
Richard G. Donnelly
1 Robert L. Goldsmith
Kenneth J. McNulty
Donald C. Grant
Michael Tan
Walden Research Division of Abcor, Inc.
Cambridge, Massachusetts 02139
Grant Contract No. R-800945-01
Project Officer
John Ciancia
Industrial Pollution Control Branch
Industrial Environmental Research Laboratory
Edison, New Jersey 08817
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH ;.UD DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
For sale by the Superintendent of Documents, U.S. Government
Printing Office, Washington, D.C. 20402
-------
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 recom-
mendation for use.
11
-------
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 con-
trol 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. These studies were un-
dertaken to evaluate the feasibility of commercially available reverse
osmosis membranes for achieving closed-loop pollution abatement of metal fin-
ishing rinse wastewaters. Reverse osmosis pilot plant testing was carried
out to recover the .chemicals while purifying the water for reuse on nine
major metal finishing rinse waters- The Metals and Inorganic Chemicals
Branch, Industrial Pollution Control Division may be contacted for further
information on this subject.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
111
-------
ABSTRACT
Reverse osmosis treatment of plating bath rinse waters has been examined.
Emphasis has been placed on closed-loop operation with recycle of purified
water for rinsing, and return of plating chemical concentrate to the bath.
Three commercially available membrane configurations have been evaluated exper-
imentally; tubular (cellulose acetate membrane), spiral-wound (cellulose ace-
tate membrane), and hollow-fiber (polyamide membrane). Tests were conducted
with nine different rinse wastes prepared by dilution of actual plating baths.
Advantages and limitations of the reverse osmosis process and specific mem-
branes and configurations are discussed. Promising, as well as unattractive,
applications are indicated.
This report was submitted in partial fulfillment of EPA Project Number
R800945 (formerly 12010 HQJ) by the American Electroplaters1 Society, Inc. un-
der the partial sponsorship of the Environmental Protection Agency.
IV
-------
CONTENTS
Page
Abstract iv
List of Figures V1
List of Tables viii
Acknowledgments ix
Sections
I Conclusions 1
II Recommendations 4
III Background 5
IV Experimental 22
V Results and Discussion 28
VI References 95
-------
FIGURES
No. Page
1 Generalized Process Flow Schematic 7
2 Tubular Membrane Module (photo) 13
3 Spiral Wound Membrane Cartridge 14
4 Permasep Hollow-Fiber Permeator 15
5 Experimental Apparatus: Flow Schematic 25
6 Flux In Neutralized Chrome Bath 32
7 Flux In Neutralized Chrome Bath 33
8 Solids Rejection In Neutralized Chrome Bath 35
+fi
9 Cr Rejection In Neutralized Chrome Bath 36
10 Flux In Unneutralized Chrome Bath 39
11 Flux In Unneutralized Chrome Bath 40
12 Solids Rejection In Unneutralized Chrome Bath 41
13 Cr Rejection In Unneutralized Chrome Bath 42
14 Flux In Copper Pyrophosphate Bath 44
15 Flux In Copper Pyrophosphate Bath 45
16 Solids Rejection In Copper Pyrophosphate Bath 46
+2
17 Cu Rejection In Copper Pyrophosphate Bath 47
-4
18 P207 Rejection In Copper Pyrophosphate Bath 48
19 Flux In Nickel Sulfamate Bath 51
20 Flux In Nickel Sulfamate Bath 52
21 Solids Rejection In Nickel Sulfamate Bath 53
22 Ni+2 Rejection In Nickel Sulfamate Bath 54
23 Br" Rejection In Nickel Sulfamate Bath 55
24 TOC Rejection In Nickel Sulfamate Bath 56
25 Boric Acid Rejection In Nickel Sulfamate Bath 57
26 Flux In Nickel Fluoborate Bath (Hollow Fiber Module) 60
27 Flux In Nickel Fluoborate Bath (Tubular Module) 61
VI
-------
FIGURES (CONTINUED)
No. Page
28 Solids Rejection In Nickel Fluoborate Bath 62
29 Ni+2 Rejection In Nickel Fluoborate Bath 63
30 Flux In Zinc Chloride Bath 66
31 Flux In Zinc Chloride Bath 67
32 Solids Rejection In Zinc Chloride Bath 68
33 Cl" Rejection In Zinc Chloride Bath 69
34 Flux In Cadmium Cyanide Bath 72
35 Solids Rejection In Cadmium Cyanide Bath 73
36 Cd+2 Rejection In Cd(CN)2 Bath 74
37 CN~ Rejection In Cd(CN)2 Bath 75
38 Flux In Zinc Cyanide Bath 77
39 Solids Rejection In Zinc Cyanide Bath 78
40 Zn+2 Rejection In Zn(CN)2 Bath 79
41 CN" Rejection In Zn(CN)2 Bath 80
42 Flux In Copper Cyanide Bath 83
43 Solids Rejection In Copper Cyanide Bath 84
44 Cu+ Rejection In CuCN Bath 85
45 CN" Rejection In CuCN Bath 86
46 Flux In Rochelle Copper Cyanide Bath 88
47 Rejection In Rochelle Copper Cyanide Bath < . 89
48 Life Data of B-9 Permeator #3 91
49 Flux Data In Zn(CN)2 Life Test 93
50 Total Solids Rejection In Zn(CN)2 Life Test 94
VII
-------
TABLES
No. Page
1 Capabilities and Limitations of RO Systems Tested 2
2 Commercially-Available Membrane Systems 11
3 Some Impurities and Their Effect 18
4 Summary of Experiments 23
5 Chemical Analyses 27
6 Experiment # 1 Chrome Bath 31
7 Experiment # 2 Chrome Bath 38
8 Experiment # 3 Copper Pyrophosphate Bath 43
9 Experiment # 4 Nickel Sulfamate Bath 50
10 Experiment # 10 Nickel Fluoborate Bath 59
11 Experiment # 5 Zinc Chloride Bath 64
12 Experiment # 6 Cadmium Cyanide Bath 71
13 Experiment # 7 Zinc Cyanide Bath 76
14 Experiment # 8 Copper Cyanide Bath 82
15 Experiment # 11 Rochelle Copper Cyanide Bath 87
16 Guide To Figure 48 92
Vlll
-------
ACKNOWLEDGMENTS
Direction was received throughout the program from members of the Ameri-
can Electroplaters' Society Project Committee: Charles Levy (District Supervi-
sor), Lawrence E. Greenberg (Committee Chairman), Arthur A. Brunei!, Joseph
Conoby, and Robert Michaelson. The project officer, John Ciancia, and two mem-
bers of the AES Research Board, Dr. Martin S. Frant and Mr. Robert Duva, have
also contributed substantially to the program direction.
Financial Support for this research from the American Electroplaters' So-
ciety (AES Project 32) and from the Office of Research and Development of
EPA is gratefully acknowledged.
IX
-------
SECTION I
CONCLUSIONS
1. Reverse osmosis (RO) appears to be an attractive process for closed-
loop treatment of plating rinse wastes. Baths for which technical
feasibility has been demonstrated are given in Table 1, together with
any limitations. In addition, systems which cannot be treated by RO at
present are also indicated.
2. At present, the spiral-wound and hollow-fiber membrane module configura-
tions are preferred for plating applications because they are more com-
pact and less expensive than tubular modules. Membrane selection is
based primarily on the pH of the rinse-water concentrate: the cellulose
acetate membrane can be used from pH 2.5-7; the polyamide membrane can
be used from pH 4-11. In the region of pH overlap, neither membrane has
an overriding advantage over the other.
3. The degree to which the recycled plating chemicals would be concentrated
by RO in a commercial-scale installation must be determined on a case-
by-case basis. Important factors include: the ratio of water loss from
plating bath evaporation to dragout; the maximum concentration limit
physically obtainable by RO; and the relative economic attractiveness of
RO concentration at high solids (low flux) vs. auxiliary evaporation.
4. The degree to which rinse waters would be purfied in a commercial
installation would also be determined on a case-by-case basis and depend
on the water purity required for effective rinsing. Considerations for
achieving the required rinsing include adding a final rinse stage out-
side the closed-loop system, staging the permeate stream in RO units,
and/or adding a final purification step, e.g., ion exchange, to treat
the permeate before reuse.
1
-------
Table 1. Evaluation of RO for Systems Tested
Attractive Systems
(For Treatment)
Limitations
Watts-Type Nickel
Nickel Sulfamate
Copper Pyrophosphate
Nickel Fluoborate
Zinc Chloride
Copper Cyanide
Zinc Cyanide
Cadmium Cyanide
Boric acid selectively permeates mem-
branes.
Boric acid selectively permeates mem-
branes .
Possible decomposition of pyro-
phosphate.
Boric acid selectively permeates mem-
branes.
Need evaporation to close loop.
Need low-pH bath for current mem-
branes.
Need low-pH bath for current mem-
branes; need evaporation to close
loop.
Need low-pH bath for current mem-
branes; need evaporation to close
loop.
Unattractive Systems
(Not For Treatment)
Limitations
Chromic Acid
Very-high pH Cyanide Baths
Attacks and destroys all membranes
unless neutralized.
Attack and destroy all membranes
commercially available. Newer
membranes under development show
promise for treating high-pH
cyanide baths.
-------
5. The effect of impurity buildup in RO closed-loop operations varies for
each bath, and will require pilot-scale studies for determination. It
is believed, however, that purification techniques used for evaporative
closed-loop systems will be equally suitable for RO systems, since the
purification problems are the same.
6. The effect of the selective permeation of certain plating bath components
in some systems will require proportionally larger additions to the bath
of those chemicals selectively permeated.
-------
SECTION II
RECOMMENDATIONS
While the results of this research indicate that reverse osmosis (RO)
appears promising for the treatment of electroplating rinse waters, it is
necessary to demonstrate the capabilities of RO under realistic conditions
before recommending its use in the plating industry. Therefore, it is re-
commended that field demonstrations be conducted with the plating baths that
appear to be attractive candidates for treatment by RO. The field demon-
stration should evaluate:
1. membrane life under practical operating conditions
2. effect of closed-loop treatment on the level of bath
impurities, and their effect on plating characteristics
3. means of controlling bath impurities
4. economics of closed-loop treatment by RO.
Other membranes should be tested as they become available. Of particular
interest are membranes that can tolerate very high or low pH's and withstand
oxidizing conditions.
-------
SECTION III
BACKGROUND
GENERAL
Industrial pollution of our streams and waterways has become an increas-
ingly acute problem. Electroplating and metal finishing waste streams are
significant contributors to this problem, either directly, due to their content
of toxic or corrosive chemicals; or indirectly, due to the deleterious effect
of these chemicals on biological waste treatment systems. The Federal Water
Quality Control Act Amendments of 1972 call for the application of "best prac-
ticable" technology by 1977 and "best available" technology by 1983 to allevi-
ate the problem. In addition, the act declares that it is the national goal to
eliminate the discharge of pollutants into the navagable waters of the United
States by 1985.
There is ample technology available for treating metal finishing wastes
to any required degree of detoxification. The problem of reducing contaminants
to a specified level is, therefore, one of economic feasibility rather than
technological feasibility. One single process will obviously not solve all
pollution problems for all platers. A variety of waste treatment facilities
will be installed in the future with the exact facility for any plater depend-
ing on a balance of economic factors for his particular plating operation.
The treatment of plating wastes can be broadly classified according to
whether plating chemicals are destroyed or recovered. Destructive processes
include chemical or electrolytic treatment and are aimed primarily at reducing
contaminant levels in effluent streams rather than reclaiming chemicals that
are lost in bath dragout and subsequent rinsing operations. On the other hand,
5
-------
recovery processes treat effluent streams in such a way that ionic species are
recovered in a form suitable either for recycle to the plating operation or
for other reuse. Examples of recovery processes are ion exchange (if the ex-
change resins are regenerated to yield recyclable materials), evaporation, RO,
and, more recently, electrodialysis. In many cases, waste streams that would
ordinarily be discharged can be treated to reclaim useful metals.
Recovery processes offer considerable potential for an economically viable
solution to the plating waste problem. As plating chemicals and disposal of
solid wastes from destructive systems become increasingly more expensive, chem-
ical recovery will be preferred to chemical destruction. Ideally, complete
closed-loop operation is feasible with the recycle of purified water to rinsing
operations and concentrated chemicals to plating baths.
Figure 1 shows a typical block diagram for closed-loop treatment of elec-
troplating rinse water by RO. Any other recovery technique, e.g., evaporation,
would be applied in a similar manner. Rinse water from the first rinse tank,
which would otherwise be discharged to drain, is first treated to remove impur-
ities, primarily particulates, and then pumped to the RO membrane concentrator.
The concentrator separates the feed stream into a "concentrate" stream, con-
taining a relatively high concentration of plating chemicals, and a "permeate"
stream, containing purified water. Before returning the concentrate stream to
the plating bath, it may be necessary to remove additional water by evapora-
tion. The permeate stream is recycled to the final rinsing stage where makeup
water is added to compensate for bath evaporation. It is necessary to use de-
ionized makeup water in order to avoid a buildup in the bath of impurities
contained in untreated water. The operation of a closed-loop process such as
shown in Figure 1 not only eleminates the problem of rinse water disposal but
also recovers valuable plating chemicals and reduces the amount of purchased
water required for rinsing.
Over the past several years RO has received increased attention as an
attractive recovery process for electroplating waste streams ^'~4'. Applica-
tion has been largely limited to the treatment of Watts-type nickel rinses,
although more recently, the treatment of total plating shop effluents and a few
-------
PLATING
CHEMICALS EVAPORATION
DRAG IN
BATH
PURIFI-
CATION
DRAG OUT
PLATING BATH
POSSIBLE I
EVAPORATION
r^n
I I
CONCENTRATED
PLATING
CHEMICALS
(CONCENTRATE)
COUNTER-CURRENT
RINSING
PURIFICATION
(FILTRATION)
RINSE WATER
(FEED)
PURIFIED
WATER
MAKE-UP
WATER TO
ACCOUNT FOR
EVAPORATION
(PERMEATE)
REVERSE OSMOSIS
CONCENTRATOR
Figure 1. Generalized Process Flow Schematic
-------
other selected rinse waters have been considered. However, published data on
the RO treatment of plating wastes are limited.
This report presents data for the RO treatment of a number of actual
plating bath rinse wastes using three different commercially available RO mod-
ules. These data can be used to determine the current performance characteris-
tics and to estimate the economic feasibility of RO for treatment of a specific
plating waste.
PRINCIPLES OF REVERSE OSMOSIS
A brief summary of reverse osmosis theory follows. A more detailed treat-
ment can be found in Reference 5. Membrane performance is evaluated in terms
of the quantity (flux) and the quality (rejection) of the permeate obtained
under conditions of the experiment.
Flux, (J) is the volume flow of permeate (water) per unit membrane area
and is proportional to the effective pressure driving force:
J = K (AP-AII) (1)
where K = the membrane constant;
AP = difference in applied pressure across the membrane
All = difference in osmotic pressure across the membrane.
Flux is an important factor since the size of the membrane system, for a
given plant capacity, is inversely proportional to flux. From equation (1) it
can be seen that flux decreases with increasing feed concentration; since II is
approximately proportional to feed concentration. At a given feed solids
level, flux is higher for plating chemicals which form high molecular weight
complexes, since II depends on molarity. Since it is generally desirable to
concentrate a rinse water to as high a solids level as possible, it is impor-
tant that the effective pressure driving force be as large as possible. This
favors the use of high pressure membrane equipment.
8
-------
The percent rejection, R, is defined as:
R = Cfeed " Permeate
Lfeed
Rejection depends primarily on membrane type. In practice, for plating appli-
cations, one selects the highest rejection membrane available in order to max-
imize the recovery of chemicals and to maximize the product water purity.
Rejection decreases with increasing feed concentration, since the solute
flux is, to a first approximation, a function only of solute concentration.
However, water flux decreases with increasing solute concentration (see Refer-
ence 5 for details).
APPLICATION OF REVERSE OSMOSIS TO PLATING WASTE TREATMENT
Previous studies have shown that RO treatment is both technically and
(1-4)
economically feasible for some plating waste streams. v ' Particular ad-
vantages of RO over other recovery processes are:
1. Low capital cost. The modular nature of RO units makes them particu-
larly well-suited for small-scale installations.
2. Low energy cost. Only power for pumping is required.
3. Low labor cost. The process is simple to operate since it involves
primarily the pumping of liquids.
4. Low space requirements. RO equipment is compact and operates con-
tinuously, requiring minimal tankage.
There are, however, some important limitations of RO specifically, and
closed-loop recovery processes in general:
1. There is a limited pH range (about 2.5 - 12) over which current mem-
branes can operate for extended periods. Treatment of streams out-
side this range requires neutralization, and the concentrate cannot,
therefore, be directly recycled to the plating bath. As new mem-
branes become available, the treatable pH range will undoubtedly
9
-------
broaden. However, at present, the economics are not particularly
attractive for the RO treatment of highly acidic or highly alkaline
waste streams.
RO is incapable of concentrating solutions to very high concentra-
tions. Concentration can be achieved by reverse osmosis only so
long as the operating pressure exceeds the osmotic pressure of the
solution. Thus, the degree of concentration provided by a partic-
ular module is limited by its maximum operating pressure (Table 2).
The severity of this limitation depends on the specific osmotic
pressure of the bath and the ratio of bath evaporation to dragout.
For plating baths with substantial evaporation (e^g., Watts nickel,
nickel sulfamate, and copper cyanide) a high degree of concentration
is not required and RO is capable of providing a concentrate stream
which can be recycled to the bath directly. For ambient-temperature
plating baths, it may be necessary to use evaporation in conjunction
with RO in order to achieve sufficient concentration for direct
recycle.
No membrane is completely effective in rejecting ionic solutes.
Specific solutes may permeate the membrane at rates sufficient to
give unacceptably high solute concentrations in the permeate for
final rinsing (or discharge). Additional treatment of the permeate,
for example by an additional RO stage or ion exchange, may be re-
quired. The seriousness of this limitation must, of course, be
evaluated on a case-by-case basis.
As in any closed-loop recovery process, a buildup in bath impurities
must be anticipated. Impurities that were formerly removed by bath
dragout and rinsing are, in closed-loop operation, recovered along
with plating chemicals and recycled to the bath. This could result
in a long-term buildup of impurities in the system which might have
an adverse effect on plating quality. As in Figure 1, it may be
necessary to use various purification systems, such as filtration,
adsorption, chemical treatment, electrolysis, and aeration, to keep
impurities at acceptably low levels. These systems are further dis-
cussed under "Purification Techniques" (page 16). It is anticipated
that the same purification schemes used for evaporative closed-loop
10
-------
Table 2. Commercially Available Membrane Systems
Type
Hollow Fiber b
(Polyamide
Membrane)
Spiral Wound
(Cellulose
Acetate
Membrane)
Tubular
(Generally
Cellulose
Acetate
Membrane
Allowable
pH range
4 to 12
2.5 to 7
2.5 to 7
Maximum
operating
pressure
kg/cm2
29
(400 psig)
58
(800 psig)
109
(1500 psig)
Approximate
cost,a
$/ liter/day
$0.084
($ 0.32/gal/day)
$0.084
($ 0.32/gal/day)
$ 0.26 - 0.53
($ 1 to 2/gal/day)
Susceptibil-
ity to plug-
ging by sus-
pended solids
very high
high
low
Space
require-
ment
very low
very low
moderate
a Cost of membrane element alone, without pump, controls, piping, instrumentation, etc.
A high-pressure, hollow-fiber module (duPont B-10) has recently become commercially
available. It is capable of operating at 58 kg/cm2 (800 psi) and uses the same membrane
(polyamide) as the lower pressure module (duPont B-9).
-------
operations will be suitable and adequate for use with RO closed-up
systems.
MEMBRANE PROCESS EQUIPMENT
There are essentially three types of commercially available RO membrane
configurations, all of which were examined in this study. The simplest type
is the tubular module, which consists of a porous tubular support with the
membrane cast in place or inserted into the support tube. The feed solution
is pumped through the tube; the concentrate is removed downstream; and the
permeate passes through the membrane/porous support composite into the sur-
rounding collection shell. Figure 2 shows a tubular membrane module which
consists of a bundle of tubes in one collection shell.
The spiral wound module contains a large membrane sheet(s) which, in order
to obtain a compact design, is wound around a central permeate-collector tube.
The feed solution is passed over one side of the sheet, and the permeate passes
through the membrane, flows through the backing material and into the permeate-
collector tube. The membrane cartridge construction is shown in Figure 3.
The hollow-fiber module consists of thousands of fine hollow fiber mem-
branes (40-80y dia) which are arranged in a bundle around a central porous tube
as shown in Figure 4. The feed solution is introduced through this tube,
passes over the outside of the fibers and is removed as concentrate. Water
permeates through the fiber walls, flows through the hollow fibers, and is
collected at one end of the unit.
Each of these modules has particular advantages and disadvantages as
summarized in Table 2. The hollow-fiber and spiral-wound systems show cost
advantages (in units of dollars per liter permeate per day) over the tubular
system; however, they are more susceptible to plugging and fouling by suspended
solids, requiring careful filtration of the feed. The tubular system has the
advantage of a higher operating pressure which makes it a preferred system
for feeds having a high osmotic pressure.
12
-------
o
Figure 2. Tubular Membrane Module
-------
ROLL TO
ASSEMBLE
FEED SIDE
SPACER
PERMEATE FLOW
(AFTER PASSAGE
THROUGH MEMBRANE)
X
PERMEATE OUT
PERMEATE SIDE BACKING
MATERIAL WITH MEMBRANE ONX
EACH SIDE AND GLUED AROUND \
EDGES AND TO CENTER TUBE \
Figure 3. Spiral Wound Membrane Cartridge
14
-------
SNAP RING
•0- RING SEAL
FEED
END PLATE
PERMEATE
FIBER
SHELL
CONCENTRATE
•0' RING SEAL
POROUS FEED END PLATE
DISTRIBUTOR TUBE
Figure 4. Permasep Hollow-fiber Permeator
15
-------
There are a number of membrane materials presently under development,
but only two are in commercial use. Of these two, (cellulose acetate and
polyamide), cellulose acetate is the more widely applied. This membrane was
originally developed for water desalination but has since been adopted for
many industrial waste treatment applications. It exhibits excellent water per-
meation rates and high rejection of ionic species, but unfortunately it is
limited to a fairly narrow pH range (2.5 - 7). Operation beyond this range
hydro!izes the membrane and changes its structure, thereby destroying its
ability to selectively pass water.
The other membrane material, duPont's polyamide, has been observed to
show long operating life (three years) when operated over a pH range of 4
to 11. Thus this membrane, commercially available in the hollow-fiber config-
uration, appears to be especially attractive for the treatment of high pH
wastes, such as cyanides, while the cellulose acetate membrane, commercially
available in both the tubular and spiral-wound configurations, would be pre-
ferable for treatment of low pH wastes.
Since these kinds of considerations must be evaluated on a case-by-case
basis, all three membrane systems were tested for all the plating baths
studied (except where pH limitations were obvious and insurmountable).
PURIFICATION TECHNIQUES^6"10)
General
Regardless of what concentration process a closed-loop recovery system is
based on (reverse osmosis, evaporation, ion exchange, or electrodialysis), a
build-up in bath impurities can be expected. Impurities formerly kept at a
low level by dragout from the bath are, in closed-loop operation, recovered
and recycled to the bath along with the plating chemicals.
Purification is presently required for long-term baths and bright-dip
baths, and current purification techniques should be adequate, if not ideal,
because RO adds no new impurities to the system. Generalizations on purifica-
tion are virtually impossible because each bath, each process, and
16
-------
each ship, have different impurity problems. The impurities can, however, be
divided into three general categories: Suspended solids, organic contaminants,
and inorganic contaminants.
Suspended solids are a problem in all baths because they result in
roughness and pitting. Sources of suspended solids include anodes, chemical
additions, improperly cleaned work, and airborne dust. Suspended solids are
routinely removed from plating baths by various methods of filtration.
Organic contaminants in the bath affect the appearance of the deposit,
especially in bright plating, and lower the efficiency of the bath. Sources
include, organic decomposition products of brighteners and improperly cleaned
work. The most common technique for removing organic contaminants is adsorp-
tion on activated carbon.
Inorganic contaminants show vastly different effects in different types
of baths. Cyanide baths have a built-in protective mechanism because many
metals complex with free cyanide (two exceptions are hexavalent chromium and
lead). Acid baths have no such complexing agents, and any contaminant near or
below the deposition potential of the plating material will co-deposit. Table
3 shows some of the problem inorganics for several baths. Inorganic contamin-
ants can be removed by chemical or electrolytic purification.
Removal of Suspended Solids by Filtration
One type of filtration involves the use of depth-type cartridge filters
to remove suspended solids. These filters retain particles as small as l-30y
and consist of twisted yarn which is wound around a core to form diamond-shaped
openings. Subsequent windings of yarn hold the fibers in place, and the open-
ings get larger as winding continues because the same number of diamond shaped
openings are included in each layer as the diameter increases. Because of this
geometry, larger particles are retained on the exterior of the filter and
smaller particles on the interior, thereby yielding a high surface area filter
in a small volume. The filter-cartridge lifetime is, of course, dependent on
the application. In general, filter systems are designed to have two 378
liters/hour (100 gallons/hour) cartridges for each 378 liters (100 gallons) of
17
-------
Table 3. Some Impurities and Their Effect
Bath
Impurity
Effect
Chromium
Copper Sulfate
Nickel Baths
Zinc Cyanide
Copper, Iron
Nickel, Zinc, Cadmium
Trivalent Chromium
Chlorides
Arsenic, Antimony
Nickel, Iron
Silicates (soluble)
Hexavalent Chrome,
Copper, Zinc
Cadmium, Lead
Copper
Co-Precipitation
Insenitive
Increases bath resistance and
may cause gray deposits
Catalytic effect
Brittle and rough deposits
Reduces bath conductivity and
may cause rough deposits
May precipitate on work
Co-precipitate
Dull plate
Dull plate and darkening
of bright plate when nitric
acid dipped
18
-------
tank capacity. Typical lifetimes of 4-6 weeks can be expected for plating
applications. The primary advantage of cartridge filters is their convenience
of use.
Another type of filtration uses precoat filters and will retain particles
as fine as 0.5-5y. The filter consists of a porous support which is treated
with a filter aid such as diatomaceous earth or some fibrous-type filter aid.
Porous supports include; paper, cloth, ceramic, stone, sintered metal, wire
mesh, and wound cartridges. The filter aid is precoated onto the porous sup-
port by forming a slurry of the filter aid with water or plating solution and
filtering the slurry. During operation, the filter is back-washed manually
when the pressure drop becomes excessive.
Granular depth filters (sand filters) can also be used for suspended
solids removal. Sand filters are typically 61-76 cm. (24-30 in.) in depth
with sands ranging from 0.45 to 1mm. in diameter. Both gravity-flow and
pressurized vessels are available. Suspended solids are trapped in the inter-
stices of the filter medium, and in some cases, formation of a surface cake at
the top of the filter can aid in the removal of solids. As the filter bed be-
comes loaded with particulates, the burden or removal gradually shifts from the
upper layers of the bed to the lower. Eventually, solids removal or pressure
drop becomes unacceptable and the filter must be cleaned by backwashing. The
advantages of the granular depth filter are its simplicity of design, opera-
tion, and maintenance. The disadvantage is that the quality of the filtrate
generally decreases with time.
Adsorption
The most common adsorbent used in bath purification is activated carbon.
It is commonly used to remove oil, grease, additive agents, and decomposition
products of brighteners. Different types of carbon remove different organics,
and suppliers of proprietary plating bath solutions will frequently supply
specially mixed carbons for adsorbing the specific impurities found in such
baths.
19
-------
Activated carbon can be applied to the plating bath in several different
ways:
1. Throw-away cartridges and canisters are commercially available.
2. Carbon can be used as the granular filter medium in a granular depth
filter.
3. Carbon can be used as the filter aid in a precoat filter.
4. Batch treatment of a portion of the bath can be carried out in a
separate mixing tank.
5. A carbon column can be used following a separate filtration system.
The final option is the most versatile since the filtration and adsorption
systems can be designed, controlled, and maintained separately.
Another adsorbent is activated clay which is used to remove NH. from
cobalt-nickel baths. Specific adsorbents are also available to remove certain
inorganics.
Chemical Treatment
Two types of chemical treatment are oxidation and precipitation. Oxida-
tion in this case, is performed using potassium permanganate or hydrogen per-
oxide. Potassium permanganate (KMnCL) oxidizes many organics, including
brightener decomposition products, and is reduced to MnCk in the process.
Mn02 has adsorptive properties, so this treatment preceding carbon adsorption
is often effective. It is not universally effective, however, and sometimes
is detrimental. Chemical treatment with KMnO^ is performed with the solution
at a pH of 1.5-2.5. The carbon is added while raising the pH of the solution
to 4.5-5.0, and the solution is then filtered.
Hydrogen peroxide (H202) is a weaker oxidant than potassium permanganate,
but has favorable decomposition products. The treatment is performed by adding
0.1-3.0 ml of 30% H202 to each liter of solution, agitating at 60°C (140°F) for
3 hours, adding carbon, and filtering.
Precipitation is used as a last resort in removing inorganics, and methods
vary from bath to bath. In zinc cyanide plating, lead and cadmium can be
20
-------
precipitated by maintaining a small excess of sodium polysulfide in the bath
and copper can be removed by adding zinc dust and filtering after 2-3 hours.
Electrolytic Purification at Low Current-Density
This treatment is based on the phenomenon that many impurities cause spe-
cific effects at certain low current-density sites during plating. In fact,
preferential co-deposition occurs at these low current-densities, so that a
"dummy" cathode operated at low current will have deposited relatively more
impurity than the plating material. Corrugated and flat plates are used for
purification, and should be plated with the plating material first. The volt-
age used must not be too low, or no deposition takes place. Also, counter-
potentials could be set up since only certain areas of the plate deposit im-
purities. The current must remain on at all times to prevent dissolution of
impurities.
2 2
Treatment begins at 0.0028-0.0043 amps/cm (0.02-0.03 amps/in ) and is
lowered gradually. Metals deposit at 0.00084-0.0043 amps/cm2 (0.006-0.03
o
amps/in ) and organics at 1/10 of these values. Both pH and additive agents
can accelerate or impede deposition; temperature and agitation accelerate
deposition; and ultrasonics may speed impurity deposition.
While a comprehensive review of purification techniques is beyond the
scope of this report, the above summary indicates some of the types of treat-
ment available to the plater. The effect of impurities on plating character-
istics should be studied in pilot-scale and full-scale operations for each
bath, and purification techniques should be recommended on a case-by-case
basis.
21
-------
SECTION IV
EXPERIMENTAL
SCOPE OF EXPERIMENTS
The scope of the experimental study is summarized in Table 4. A total of
nine different plating baths were studied. Bath solutions were obtained from
actual in-field plating operations. Chromium and cyanides were of particular
interest because of their high-volume usage in the industry. Bath properties
(total dissolved solids and pH) are listed in Table 4. Also shown are the con-
centration and pH ranges for the test solutions, prepared by diluting the bath
samples with deionized water. Three commercially available modules were
tested:
1. A duPont B-9 hollow-fiber module containing a polyamide membrane (E.I.
DuPont de Nemours and Co., Inc., Permasep Products, Wilmington, Dela-
ware, 19898).
2. A T.J. Engineering 97H32 spiral-wound module containing a cellulose
acetate membrane (T.J. Engineering, Inc., Downey, Cal.).
3. An Abcor TM5-14 tubular module containing a cellulose acetate membrane
(Abcor, Inc., 341 Vassar St., Cambridge, Massachusetts, 02139).
Because of the pH limitations of the cellulose acetate membrane, only the
polyamide membrane (duPont hollow-fiber module) was used in tests with cyanide
solutions.
Two types of tests are required in order to assess the ability of RO to
treat a specific plating waste. First, short-term tests are required to de-
termine intrinsic membrane performance, and second, long-term life tests are
required to determine the stability of performance over long periods.
22
-------
Table 4. Summary of Experiments
Baths
Plating
Baths
Chromic Acid
(Neutralized)
Chromic Acid
(Unneutralized)
Copper Pyroshosphate
Nickel Sulfamate
Nickel
Fluoborate
" Zinc Chloride
Cadmium Cyanide
Zinc Cyanide
Copper Cyanide
Rochelle Copper
Cyanide
Zinc Cyanide Life
Test
Source of
Bath
Whyco Chromium
Company
Whyco Chromium
Company
Honeywell
(M & T)
Honeywell
(Harstan)
Hampden
Colors & Chemicals
General Electric
(Conversion
Chemical )
American Electro-
plating Co.
American Electro-
plating Co.
American Electro-
plating Co.
Whyco Chromium
Company
American Electro-
plating Co.
Properties of Bath
TDS-mg/1
37.1
27.5
31.9
31.0
25.7
19.8
26.3
11.4
37.0
12.7
26.3
PH
--
0.53
8.8
4.2
3.5
4.5
13.1
13.9
13.3
11.2
13.1
Properties of Test
Solutions
Concentration
Range(TDS-mg/l)
0.3 - 15.0
0.4 - 9.0
0.2 - 21.0
0.5 - 26.0
0.9 - 17.0
0.2 - 12.0
0.3 - 10.0
0.5 - 4.0
0.6 - 8.0
0.13 - 3.3
0.3 - 10.0
PH
Range
5.3 - 6.4
0.9 - 1.9
7.1 - 8.5
4.6 - 6.1
1.9 - 4.9
4.7 - 6.1
11.4 - 12.9
12.3 - 13.7
11.8 - 12.9
9.8 - 10.6
11.4 - 12.9
Membrane
Modules
Tested*
A, B, C
A, B, C
A, B, C
A, B, C
A, C
A, B, C
A
A
A
A
A
* A - DuPont B-9 permeator, polyamide hollow fiber membrane.
B - T.J. Engineering 97 H 32 spiral wound module; cellulose acetate membrane.
C - Abcor TM 5-14 module, tubular configuration; cellulose acetate membrane.
-------
Short-term tests were conducted with all the wastes listed in Table 4, however,
because of the time involved, life tests were conducted with zinc cyanide only.
TEST SYSTEM
A flow schematic for the pilot plan memerane test system used is shown in
Figure 5. A feed sample was charged to a 76 liter (20 gallon) surge tank. A
level switch served as a safety device to shut down the system if the surge
tank contents were depleted. A temperature switch controlled the flow of ei-
ther hot or cold water through a coil installed in the tank so as to permit
operation at the desired test temperature. Feed solution was pumped through a
20-micron cartridge filter by a low-pressure booster pump. The filter solution
was then pumped through the membrane modules by a high-pressure positive-
displacement pump; a pressure relief valve prevented system overpressurization.
Inlet and outlet pressures for the three membrane modules were measured, and
the outlet pressures were individually controlled with back pressure regula-
tors. Concentrate and permeate flows were measured before being returned to
the surge tank. Samples of the various permeates and concentrates were collec-
ted through the sample valves shown.
OPERATING PROCEDURE
The test system was operated in a "differential" mode. Samples were di-
luted to rinse water concentration in the surge tank and pumped through the
membrane elements. Both concentrate and permeate were returned to the surge
tank, so that feed concentration was time-invariant. This permitted the evalu-
ation of membrane performance at a constant feed concentration. Different feed
concentrations were tested by simply changing the dilution of the bath sample.
During all tests, operation proceeded at fixed conditions for at least one
hour, by which time steady state was attained. At that point, membrane capac-
ity (flux) was measured and samples were collected for chemical analyses.
There are a number of process parameters important in influencing perform-
ance and costs of a full-scale unit. These important process parameters are;
feed concentration, pH, operating pressure, operating temperature, circulation
24
-------
LEGEND
LS • LEVEL SWITCH
TC - TEMPERATURE SWITCH/CONTROLLER
Tl - TEMPERATURE INDICATOR
PRV - PRESSURE RELIEF VALVE
SV - SAMPLE VALVE
P - PRESSURE GAUGE
BPR • BACK PRESSURE REGULATOR
ro
en
HOT OR COLD
WATER
MEMBRANE
UNIT
DRAIN
TANK BOOSTER FILTER FEED
PUMP
PUMP
CONCENTRATE
ROTAMETER
PERMEATE
ROTAMETER
Figure 5. Experimental Apparatus: Flow Schematic
-------
rate, and time over the life of the membrane. The effects of some of these pa-
rameters on general membrane performance are well known, and for these (operat-
ing pressure, operating temperature, circulation rate, and to some extent pH)
optimum conditions have been established. Of the more freely variable param-
eters, feed concentration, pH, and life times are of greatest interest.
ANALYSES
The methods of analysis used are listed in Table 5. All are from referer-
ence 11.
26
-------
Table 5. Chemical Analyses
Constituent
Method
Procedure
(Reference No. 11)
Cadmi um
Total chrome
Hexavalent chrome
Copper
Nickel
Zinc
Bromide
Chloride
Pyrophosphate
Cyanide
Boric acid
Total organic
carbon
Total dissolved
solids
Atomic absorption
Atomic absorption
Colorimetric
Atomic absorption
Atomic absorption
Atomic absorption
Colorimetric
Hg (N03)2 Titration
Colorimetric
Titration method
As Boron; Semi Colorimetric
Combustion-infrared
Gravimetric
129
129
211-IID
129
129
129
108
112B
223
207B
107A
138A
148B
27
-------
SECTION V
RESULTS AND DISCUSSION
GENERAL
Results are presented and discussed below for each system tested. In the
short-term tests, the independent variable was the feed concentration, and the
dependent variables were the flux and rejection. Flux is reported in terms of
gallons per square foot of membrane surface per day (gfd) for the tubular and
spiral-wound modules and in terms of gallons per minute per module for the
hollow-fiber module. Rejections are plotted as the logarithm of (100-R) in or-
der to clarify the rejection behavior at low feed concentrations. The data are
summarized in Tables 6 through 16 and Figures 6 through 50. It should be noted
that the use of the high-pressure B-10 hollow fiber module (which was not avail-
able at the time of these tests) would give substantially higher fluxes and re-
jections than reported here for the lower-pressure B-9 module.
In general, operating conditions were held approximately constant. The
hollow fiber and spiral wound modules were operated at their pressure limit
(29.1 and 43.2 kg/cm [400 and 600 psi] respectively) while the tubular module
o
was operated at 48.6 to 58.2 kg/cm (650 to 800 psi), considerably below its
o
limit of 106 kg/cm (1500 psi). Operation of the tubular module at higher
pressures would give higher fluxes and higher rejections than observed in this
study.
The feed temperature was maintained at approximately 30°C (86°F). In gene-
ral, flux increases with temperature at about 3.5% per degree C, but rejection
remains essentially independent of temperature. A maximum temperature of 35°C
(95°F) is recommended for the duPont polyamide membrane, and a temperature of
about 30°C (85°F) is recommended for the cellulose acetate membrane, although
28
-------
higher tempertures can be tolerated at the expense of a more rapid irreversible
loss of flux because of membrane compaction.
The conversion, defined as the ratio of permeate flow to feed flow, is
also an important operating variable. When the conversion is low, the entire
module "sees" approximately the same concentration, i.e., the feed concentra-
tion.. However, for high conversions, the average concentration within the mod-
ule can be considerably greater than the feed concentration. Since both flux
and rejection decrease with increasing feed concentration, a module operated
under conditions of high conversion is subjected to more severe conditions than
a module operated under low-conversion conditions.
The range of conversions differed for the three modules: Tubular, 0-10%;
Spiral, 1-30%; Hollow Fiber, 5-75%. Ideally, the flux and rejection should be
measured at 0% conversion so the measured flux and rejection will correspond to
the known feed concentration rather than to some unknown average concentration.
For conversions in the 0-30% range, the measured flux and rejection as a func-
tion of feed concentration will be reasonably accurate and in any case conser-
vative (low) compared to the case for 0% conversion. In the range of 5-75%
conversion for the hollow fiber module, measurement of the flux and rejection
for a given feed concentration may be quite conservative (low) as compared to a
measurement at 0% conversion. This is because at 75% conversion, the average
concentration seen by the module is on the order of two times the feed concen-
tration.
In general, higher conversions were employed at low feed concentrations.
Experimentally, the feed to a module was held constant, and at low feed concen-
trations, the high flux resulted in high conversions. Conversely, at high feed
concentrations the conversion was low, and the concentration dependence of flux
and rejection are given more accurately. Therefore, the effect of conversion
on the flux and rejection data is: the results are conservative at low feed
concentrations and more accurate at higher feed concentrations.
Several trends in the data are obvious. First, the flux decreases with
increasing feed concentration. This behavior follows equation (1) since an
29
-------
increase in feed concentration is accompanied by an increase in osmotic pres-
sure, and the driving force for permeation (AP - All) is reduced. Second, the
rejection also generally decreased with increasing feed concentration. This is
because the flux of the solute increases with feed concentration, but the flux
of water decreases with increasing feed concentration. In many cases, e.g.,
Figure 8, the rejection decreased at low concentrations. This is attributed to
dissociation, at low concentration, of higher molecular weight complexes into
lower molecular weight ionic species, for which the membrane rejection is lower.
CHROMIC ACID BATH (NEUTRALIZED)
Neutralization of chromic acid bath rinse waters was found necessary in
order to prevent the eventual destruction, by hydrolysis, of the membrane
material for all three modules tested. This neutralization was accomplished by
addition of NaOH to give a pH of 4.5 to 6, and resulted in a significant exten-
sion of membrane life. This pH range was chosen because the rate of hydrolysis
of cellulose acetate membranes is at a minimum when the pH value is near 5.
Operation at the very low pH values of the unneutralized samples is known to
considerably shorten the life of both membranes tested. The question of mem-
brane life is a crucial one, and will be discussed subsequently.
All three membrane modules performed satisfactorily in neutralized chromic
acid rinse waters with respect to flux and rejection (see Table 6). Flux de-
creased with increasing feed concentration as expected, according to the theo-
retical treatment presented above. This decrease was most pronounced for the
9
hollow-fiber module, since the operating pressure was only 29.1 kg/cm (400
2 2
psig) as compared to 43.2 kg/cm and 48.6 kg/cm (600 and 650 psig) for the
spiral wound and tubular modules respectively. The decrease in flux is shown
graphically in Figures 6 and 7. Note that these pressures were the maximum
operating pressures for all but the tubular module, which would have shown
significantly improved performance if pressure were increased toward the 106
kg/cm2 (1500 psig) limit. Flux was good up to about 10% of plating bath con-
centration (5.4% TDS) and acceptable up to approximately 25% of bath concentra-
tion (12.5% TDS). Note in this regard that the particular bath sample chosen
for study is of much higher solids content (450 g/1 CrOj than typical. Thus,
30
-------
Table 6. Experiment # 1 Chrome Bath
(Neutralized with NaOH)
Feed Solution
% Solids(1'3)
.28
1.55
2.70
4.50
14.9
37.0
% Bath^2'4)
.57
2.91
5.16
9.5
28.4
100
Membrane Module
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Operating Conditions
Pressure
(psi)
400
600
650
400
600
650
400
600
650
400
600
650
400
600
650
Temperature
(°c)
20
30
39
28
29
PH of
Feed
6.1
4.5
4.7
5.5
4.4
Flux
(gfd)
3.02
14.0
10.4
2.26
11.6
7.12
1.68
9.01
6.33
1.10
4.62
3.96
0.46
4.23
0.96
% Rejection
Total Dissolved
Solids
97.9
95.1
96.7
98.8
95.7
97.9
98.7
96.1
97.5
97.6
90.8
96.4
40.2
76.7
89.2
Cr+6
99.4
94.8
97.5
97.4
96.2
98.6
97.6
96.0
97.4
95.0
90.8
96.7
51.7
76.7
94.8
(1) % Total Dissolved Solids (TDS)
(2) % of Plating Bath TDS Concentration
(3) Includes NaOH added
(4) Excludes NaOH added
-------
30
25
A T. J. ENGINEERING
SPIRAL WOUND MODULE
(36 FT2 MEMBRANE)
• ABCOR TUBULAR MODULE
(9.1 FT2 MEMBRANE)
20
•&
*" 15
10
10 15 20
TOTAL DISSOLVED SOLIDS
25
30
Figure 6. Flux in Neutralized Chrome Bath
32
-------
3.0
2.S
2.0
DUPONT POLYAMIDE
HOLLOW FIBER PERMEATOR
E
O)
x"
1.S
1.0
0.5
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
Figure 7. Flux in Neutralized Chrome Bath
33
-------
flux and rejection results corresponding to 25% of this bath strength are typi-
cal of rinse waters of 60 to 75% of normal bath strength.
The quality of the permeate is illustrated in Figure 8 which shows re-
jection as a function of feed solute concentration for the three membrane sys-
tems. These data are based on measurements of the percentage total solids
content of the various samples. The corresponding data based on percentage
hexavalent chromium content are given in Figure 9. At low to moderately high
concentrations (up to 10% of actual bath concentration) rejections were
excellent (> 97%). At 25% of actual bath concentrations, rejections were good
(> 90%). Only at the sample concentration of 50% of bath concentration did
rejection fall to low values. This fall, at the higher feed solute concentra-
tions, was due to the build-up of very high concentrations of solute just at
the membrane surface. This build-up was much greater for the higher feed
solute concentrations, and the net effect is that greater amounts of solute
were permeated.
Note that the rejections for the tubular and spiral wound system increased
with concentration at low concentrations before falling at high concentrations.
This is believed due to the formation of the larger and more easily rejectable
dichromate ion, the formation of which is favored under low acidity/high chro-
mate concentration conditions. Under highly acidic conditions, the formation
of chromic acid is favored.
Based on flux and rejection data, reverse osmosis (RO) appears to be an
attractive method for concentrating chromium plating bath rinse waters to
obtain a concentrate containing about 7 to 9% TDS, if neutralization is prac-
ticed. However, the costs of neutralization (equipment and chemicals) and of a
process to remove added chemicals, if reuse of the chromic acid is desired (ion
exchange), are significant. The development of an oxidant resistant membrane,
which could be used to concentrate chromic acid rinse waters without neutrali-
zation, is clearly desirable. The availability of such a membrane would likely
result in widespread use of RO for concentrating chromium rinses.
34
-------
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
OPEN POINTS GIVE THE
REJECTION FOR '/2% NaCI
SOLUTION BEFORE (LEFT
AXIS) AND AFTER (RIGHT
AXIS) THE TEST.
10 15 20
% TOTAL DISSOLVED SOLIDS
Figure 8. Solids Rejection in Neutralized Chrome Bath
35
-------
u
LU
K
CO
O
s?
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
10 15 20
% TOTAL DISSOLVED SOLIDS
Figure 9. Cr Rejection in Neutralized Chrome Bath
36
-------
CHROMIC ACID BATH (UNNEUTRALIZED)
The flux and rejection results (Table 7) for the unneutralized chromic
acid rinse waters were similar to those for the neutralized chromic acid rinse
waters, as is evident from Figures 10 and 11 (flux data), Figure 12 (rejection
data based on TDS) and Figure 13 (rejection data based on Cr+6). The maximum
in the rejection behavior was apparently absent, suggesting that dichromate
formation is inhibited in low pH samples. Because of the hydrolysis and even-
tual destruction of the membranes tested with unneutralized samples, these re-
sults are only of interest for the indication that RO will be effective if
suitable hydrolysis-resistant membranes can be formulated.
COPPER PYROPHOSPHATE
Permeate fluxes and rejections were, for the most part, very favorable for
this system (see Table 8). As is seen in Figures 14 and 15, flux was excellent
for the tubular module up to the highest feed concentration tested (approxi-
mately 40% of bath concentration) and excellent for the spiral wound module up
to approximately 25% of bath concentration (10.5% TDS). Flux was good for the
hollow fiber module up to 15 to 20% of bath concentration (7-8% TDS).
Rejection results were also highly favorable, as can be seen in Figures
+2 -4
16, 17, and 18, based on TDS, Cu , and PJ^j concentrations, respectively.
In particular, rejections for the spiral wound module were excellent up to 25%
of bath concentration (10.5% TDS) and good up to approximately 40% of bath
concentration (15% TDS). Maximum rejection
and 99.7%, respectively, for this membrane.
+2 -4
concentration (15% TDS). Maximum rejections based on Cu and P0 were 99.6
The hollow fiber membrane also performed very well. Less attractive was
the performance of the tubular membrane. This is unexpected based on the per-
formance of this configuration with other plating bath samples and is believed
due to the use of the membrane module in the immediately preceding unneutral-
ized chromic acid tests. The membrane was apparently partially hydrolyzed, as
was evident from the rejection of 0.5% NaCl solution at the completion of the
copper pyrophosphate experiments. Results comparable to, or better than, the
spiral wound membrane would be expected for a fresh tubular membrane.
37
-------
Table 7. Experiment # 2 Chrome Bath
(Unneutralized)
Feed Solution
% Solids(1)
.398
1.83
4.11
9.43
% Bath^2^
1.45
6.65
14.9
34.3
Membrane Module
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Operating Conditions
Pressure
(psi)
400
600
650
400
600
650
400
600
650
400
600
650
Temperature
(°C)
29
29
29
28
PH of
Feed
1.9
1.2
1.2
0.9
Flux
(gfd)
2.59
15.3
10.0
1.97
13.2
8.58
1.20
10.6
7.31
leaks
leaks
6.60
% Rejection
Total Dissolved
Solids
84.3
97.3
99.4
95.0
94.0
96.9
89.9
91.7
95.2
leaks
leaks
94.2
Cr+6
97.0
95.6
97.8
86.7
86.1
90.6
91.1
92.2
95.6
leaks
leaks
97.0
co
00
(1), (2) See Table 6
-------
30
25
A T. J. ENGINEERING
SPIRAL WOUND MODULE
(36 FT2 MEMBRANE)
• ABCOR TUBULAR MODULE
(9.1 FT2 MEMBRANE)
20
•6
x 15
V
^m
\
10
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 10. Flux in Urfneutralized Chrome Bath
39
-------
3.0
2.5
• (LEAK)
• DUPONT POLYAMIDE
HOLLOW FIBER PERMEATOR
2.0
E
a.
01
x* 1.5
1,0
0.5
I L
10 15 20
% TOTAL DISSOLVED SOLIDS
25 30
Figure 11. Flux in Unneutralized Chrome Bath
40
-------
I
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
OPEN POINTS GIVE THE
REJECTION FOR '/*% NaCI
SOLUTION BEFORE (LEFT
AXIS) AND AFTER (RIGHT
AXIS) THE TEST.
10 15 20
% TOTAL DISSOLVED SOLIDS
25
Figure 12. Solids Rejection in Unneutralized Chrome Bath
41
-------
o
UJ
oc
ID
U
ss
10
20
30
40
50
60
70
80
90
95
96
97
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
98
99
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 13. Cr Rejection in Unneutralized Chrome Bath
42
-------
Table 8. Experiment # 3 Copper Pyrophosphate Bath
Feed Solutions
% Sol ids ^
.177
1.09
2.47
5.22
8.49
11.4
14.5
21.4
% Bath^2^
.55
3.42
7.74
16.4
26.6
35.1
45.5
67.1
Membrane Module
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Hollow Fiber
Spiral
Operating Conditions
Pressure
(psi)
400
600
800
400
600
800
400
600
800
400
600
800
400
600
800
400
600
800
400
600
400
600
Temperature
(°c)
28
31
29
30
30
31
34
28
pH of
Feed
6.8
7.0
7.3
7.3
8.0
8.5
8.4
8.3
Flux
(gfd)
2.88
21.7
26.8
2.45
18.2
20.9
2.11
16.7
18.0
1.34
13.4
16.5
.64
10.3
13.8
.088
7.4
leak
0.016
3.8
0.0061
.53
% Rejection
Total Dissolved
Solids
92.4
91.2
97.9
98.5
97.8
98.8
99.0
98.0
97.2
98.8
96.8
95.7
93.1
96.7
82.5
86.8
95.8
leak
77.7
91.5
12.4
60.4
Cu+2
99.8
99.5
-100
99.9
99.6
99.7
99.6
99.6
99.3
99.8
99.3
98.3
96.0
98.7
97.7
87.9
98.1
leak
83.8
85.9
23.1
70.8
P 0
P2U7
98.1
99.1
=100
99.6
99.7
99.7
99.7
99.6
99.4
99.5
99.4
98.6
96.4
98.9
98.7
91.2
98.2
leak
80.9
94.3
50.9
51.8
-pi
co
(1), (2) See Table 6
-------
30
25
20
x 15 -
10 -
5 -
A T. J. ENGINEERING
SPIRAL WOUND MODULE
(36 FT2 MEMBRANE)
ABCOR TUBULAR MODULE
(9.1 FT2 MEMBRANE)
10 15 20
% TOTAL DISSOLVED SOLIDS
Figure 14. Flux in Copper Pyrophosphate Bath
44
-------
3.0
2.5
2.0
DUPONT POLYAMIDE
HOLLOW FIBER PERMEATOR
|
1.5
t.O
0.5
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25 30
Figure 15. Flux in Copper Pyrophosphate Bath
45
-------
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
OPEN POINTS GIVE
THE REJECTION FOR
'/*% NaCI SOLUTION
BEFORE (LEFT AXIS)
AND AFTER (RIGHT
AXIS) THE TEST.
O
A
10 15 20
% TOTAL DISSOLVED SOLIDS
25
Figure 16. Solids Rejection in Copper Pyrophosphate Bath
46
-------
99.5
99.9
SPIRAL WOUND
TUBULAR
HOLLOW FIBER
99
10 15 20
% TOTAL DISSOLVED SOLIDS
+2
Figure 17. Cu Rejection in Copper Pyrophosphate Bath
47
-------
10
20
30
40
SO
60
70
80
90
o
P
u
IU
3
K
f
I*.
O
CM
95
96
97
98
99
99.S
99.8
SPIRAL WOUND
TUBULAR
HOLLOW FIBER
% TOTAL DISSOLVED SOLIDS
10
25
30
Figure 18.
-4
Rejection in Copper Pyrophosphate Bath
48
-------
Based on the flux and rejection results, the copper pyrophosphate system
appears to be a most attractive candidate for RO treatment. No pH adjustment
would be necessary for operation with any of the three commercially available
membranes tested, and no evidence of fouling was observed. One limitation
which was not evaluated is the possible conversion of pyrophosphate to ortho-
phosphate, and the effect on bath performance of its build-up in a closed loop
system.
NICKEL SULFAMATE
All three membrane modules performed satisfactorily with nickel sulfamate
plating bath rinse waters (see Table 9). Flux data are presented in Figure 19
for the spiral wound and tubular modules, and in Figure 20 for the hollow fiber
module. Flux was excellent up'to 40% of bath concentration (15% IDS) for the
tubular module and good nearly to this level for the spiral wound module. Flux
was acceptable for the hollow fiber module up to only about 20% of bath
concentration.
Rejection results were quite good for all three modules based on concen-
trations of IDS (Figure 21), Ni+2 (Figure 22), Br~ (Figure 23), and total
organic carbon (Figure 24). Rejection of boric acid was only fair (tubular) to
poor (spiral wound and hollow fiber) as seen in Figure 25. In general, there
were no distinct differences in rejection performance for the three module
types, unlike the behavior for most other systems. Rejections were good up to
approximately 15 to 20% of bath composition (7-8% IDS) for all components
except total organic carbon, for which they were acceptable, and boric acid,
for which they were only fair. Similar behavior has been reported for treat-
ment of Watts Nickel baths.^1-2^
Nickel sulfamate plating bath rinse waters may well be suitable for RO
treatment. No pH adjustment would be necessary and rejection results are quite
attractive for all modules. Flux results favor the tubular and spiral wound
configurations, but an economic study would be needed to make a final determin-
ation of membrane configuration. In addition, nickel sulfamate baths are
operated at elevated temperatures where evaporation is substantial, and a high
49
-------
Table 9. Experiment # 4 Nickel Sulfamate Bath
Feed Solutions
% Sol ids ^
.499
2.39
4.11
6.17
12.0
23.6
% Bath(2)
1.60
7.68
13.2
19.8
38.6
75.9
Membrane
Module
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Operating Conditions
Pressure
(psi)
400
600
800
400
600
800
400
600
800
400
600
800
400
600
800
400
600
800
Temperature
(°C)
30
30
29
29
29
30
pH of
Feed
6.1
5.8
5.5
5.3
4.9
4.6
Flux
(gfd)
2.02
17.1
25.1
1.58
15.6
20.9
.96
12.9
17.6
.38
9.83
15.5
.048
6.34
10.9
.015
.79
1.67
% Rejection
Total Dissolved
Solids
94.8
96.8
91.6
95.8
93.3
93.8
96.7
94.6
90.5
92.4
89.2
89.6
73.9
90.0
88.0
17.4
51.3
13.1
Ni+2
93.5
95.3
95.4
95.9
95.6
95.8
97.6
94.7
94.8
94.8
92.4
92.0
86.6
84.1
90.4
Br"
=100
=100
97.7
99.3
97.2
97.0
91.2
95.6
92.2
89.2
92.5
89.6
39.5
86.0
81.8
H3B03
40.0
16.7
66.7
44.3
13.6
70.5
61.9
38.1
55.0
5.0
39.6
50.5
26.3
35.1
40.4
TOC
96.1
95.1
94.2
92.5
86.0
98.5
91.3
83.9
83.4
82.8
73.9
80.2
56.9
72.5
70.8
9.1
1.5
1.5
en
o
(1), (2) See Table 6
-------
30
25
T. J. ENGINEERING
SPIRAL WOUND MODULE
(36 FT2 MEMBRANE)
ABCOR TUBULAR MODULE
(9.1 FT2 MEMBRANE)
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 19. Flux in Nickel Sulfamate Bath
51
-------
3.0
2.5
2.0
. 1.5
1.0
0.5
DUPONT POLYAMIDE
HOLLOW FIBER PERMEATOR
I I
±
10 15 20
% TOTAL DISSOLVED SOLIDS
25 30
Figure 20. Flux in Nickel Sulfamate Bath
52
-------
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
OPEN POINTS GIVE
THE REJECTION FOR
J4 % NaCI SOLUTION
BEFORE (LEFT AXIS)
AND AFTER (RIGHT
AXIS) THE TEST.
10 15 20
% TOTAL DISSOLVED SOLIDS
Figure 21. Solids Rejection in Nickel Sulfamate Bath
53
-------
o
u
10
20
30
40
50
60
70
80
90
95
96
97
98
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
99
I
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
+2
Figure 22. Ni Rejection in Nickel Sulfamate Bath
54
-------
10
20
30
40
50
60
70
80
o
3 90
K
95
96
97
98
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 23. Br" Rejection in Nickel Sulfamate Bath
55
-------
o
u
U
a
10
20
30
40
SO
60
70
80
90
95
96
97
98
SPIRAL WOUND
TUBULAR
HOLLOW FIBER
99
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 24. TOC* Rejection in Nickel Sulfamate Bath
* (Total Organic Carbon)
56
-------
u
LU
10
20
30
40
50
60
70
80
90
95
96
97
98
'A—A-
SPIRAL WOUND
TUBULAR
HOLLOW FIBER
I
_L
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 25. Boric Acid Rejection in Nickel Sulfamate Bath
57
-------
degree of concentration by RO is not required. The poor rejection of boric
acid is a potential limitation to RO treatment of this bath, but the loss of
boric acid by selective permeation will not present a significant economic dis-
advantage when compared to the advantage of nickel recovery. However, the
effect of its build-up in the rinse tanks should be given further consideration.
NICKEL FLUOBORATE
Tests were conducted on the nickel fluoborate bath using the tubular and
hollow fiber modules. The data is summarized in Table 10.
Flux results are plotted in Figures 26 (hollow fiber module) and 27 (tub-
ular module). The flux remained acceptable for both modules up to about 70% of
bath concentration.
Rejection data are shown in Figures 28 (total solids) and 29 (nickel).
Rejections for the hollow fiber module were significantly below those for the
tubular module. This was because the hollow fiber module used in this test was
the same module used in the life test with zinc cyanide, which was conducted
prior to the nickel fluoborate tests. During the life test (discussed below)
a decrease in rejection was observed and accounts for the poorer rejection
performance of the hollow fiber module.
Based on the tubular module rejections, good nickel rejections were ob-
tained up to about 70% of bath concentration (the highest concentration mea-
sured) and may well remain good to 100% of bath concentration. The total solids
rejections were less favorable. Although boric acid rejections were not mea-
sured, selective boric acid permeation should be anticipated for this bath.
Nevertheless, nickel fluoborate appears to be a very attractive candidate for
RO treatment.
ZINC CHLORIDE
Flux and rejection data for the zinc chloride sample are given in Table 11.
Flux and rejection for the zinc bath showed a rapid decline with increasing
58
-------
Table 10. Experiment # 10 Nickel Fluoborate Bath
Feed Solutions
% Solids(1)
.88
1.7
2.5
5.8
17
% Bath^
3.4
6.6
10
23
66
Membrane Module
Hollow Fiber
Tubular
Hollow Fiber
Tubular
Hollow Fiber
Tubular
Hollow Fiber
Tubular
Hollow Fiber
Tubular
Operating Conditions
Pressure
(psi)
400
800
400
800
400
800
400
800
400
800
Temperature
(°C)
19
19
20
20
23
23
23
23
25
25
pH of
Feed
6.1
6.1
4.9
4.9
4.0
4.0
3.4
3.4
3.0
3.0
Flux
(gfd)
20.6
1.5
19.2
1.2
18.4
1.1
10.9
.55
5.6
.33
% Rejection
Total Dissolved
Solids
65
95
67
93
64
92
60
85
44
72
Ni+2
78
95
74
95
70
95
74
94
72
93
en
10
(1), (2) See Table 6
-------
3.0
2.5
2.0
1.5
1.0
0.5
-O
O
I
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 26. Flux in Nickel Fluoborate Bath
( Hollow Fiber Module )
60
-------
30
25
20
15
10
I
J_
I
10 15 20
TOTAL DISSOLVED SOLIDS
25
30
Figure 27. Flux in Nickel Fluoborate Bath
( Tubular Module )
61
-------
u
IU
3 90
o
v>
. D
•-*
J
O HOLLOW FIBER, 400 PSI
O TUBULAR, 800 PSI
99
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 28. Solids Rejection in Nickel Fluoborate Bath
62
-------
3 901
HOLLOW FIBER, 400 PSI
TUBULAR, 800 PSI
99!
I
_L
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
+2
Figure 29. Ni Rejection in Nickel Fluoborate Bath
63
-------
Table 11. Experiment # 5 Zinc Chloride Bath
Feed Solutions
% Sol ids ^
.163
.644
1.58
4.19
4.86
11.82
% Bath(2)
.82
3.25
8.0
21.2
24.5
60.0
Membrane Module
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Hollow Fiber
Spiral
Tubular
Operating Conditions
Pressure
(psi)
400
600
800
400
600
800
400
600
800
400
600
800
400
600
800
400
600
800
Temperature
(°c)
27
27
28
29
30
30
pH of
Feed
6.1
5.8
5.5
5.3
5.0
4.7
Flux
(gfd)
2.06
16.6
22.6
1.68
12.8
19.9
.96
9.33
15.1
.11
5.07
9.62
.03
2.54
5.23
.014
.58
.230
% Rejection
Total Dissolved
Solids
84.1
84.6
82.9
91.3
89.0
93.4
93.7
92.2
92.7
95.6
93.7
88.5
73.9
77.6
88.6
30.0
42.7
53.9
Cl~
52.0
54.4
50.4
76.1
70.7
80.9
84.3
81 .7
82.6
90.0
86.8
78.3
65.9
69.0
83.4
17.1
31.7
41.5
(1), (2) See Table 6
-------
feed concentration. Since zinc chloride complexes are of relatively low-
molecular weight, this demonstrates that molarity is the important concentra-
tion parameter in controlling flux and rejection. The fluxes for the tubular
and spiral wound modules (Figure 30) were high at low concentrations, but fell
rapidly to good and acceptable values at 20% of bath concentration (5% IDS).
Flux was acceptable for the hollow fiber module (Figure 31) up to only one-
half this concentration.
Rejections showed a pronounced maximum for all three modules based on
both IDS concentration (Figure 32) and Cl~ concentration (Figure 33). The
three modules behaved very similarly with no improvement under increasing op-
erating pressure, except at the unacceptably low rejections at higher concen-
trations. The increased rejection with increasing concentration is attributed
to the following phenomenon. At low concentration, the zinc chloride complex
is highly dissociated, and a major proportion of the zinc and chloride exists
as free ions. At intermediate concentrations, a much greater proportion is
complexed as a higher molecular weight species, more readily rejected by the
membrane. Rejections were good up to about 20% of bath concentration (5% TDS)
based on TDS concentration, but were only acceptable to fair up to this feed
strength based on Cl~ concentration.
Based on these results, zinc chloride does not appear to be as attractive
a candidate as copper pyrophosphate, nickel sulfamate, and nickel fluoborate.
Rinse water cannot be concentrated to as great a degree and there is no sub-
stantial evaporative loss from the zinc chloride bath (which operates at
ambient temperature). Therefore, substantial auxiliary evaporation would be
required for closed-loop treatment.
CADMIUM CYANIDE
The high pH range of all cyanide baths excludes the use of cellulose
acetate membranes. Consequently, only the hollow-fiber polyamide membrane
module was employed in these experiments.
65
-------
30
25
20
x 15
10
A TJ. ENGINEERING
SPIRAL WOUND MODULE
(36 FT2 MEMBRANE)
• ABCOR TUBULAR MODULE
(9,1 FT2 MEMBRANE)
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 30. Flux in Zinc Chloride Bath
66
-------
3.0
2.5
DUPONTPOLYAMIDE
HOLLOW FIBER PERMEATOR
2.0
E
* 1.5
w
1.0
0.5
I I
10 15 20
%TOTAL DISSOLVED SOLIDS
25 30
Figure 31. Flux in Zinc Chloride Bath
67
-------
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
OPEN POINTS GIVE THE
REJECTION FOR Yz% NaCI
SOLUTION BEFORE (LEFT
AXIS) AND AFTER (RIGHT
AXIS) THE TEST.
99
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 32. Solids Rejection in Zinc Chloride Bath
68
-------
o
UJ
10
20
30
40
50
60
70
80
••
90
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
95
96
97
98
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
30
Figure 33. Cl" Rejection in Zinc Chloride Bath
69
-------
The cadmium cyanide bath tested had a solids content intermediate between
that of the zinc cyanide, on the low side, and the copper cyanide bath, on the
high side. The flux and rejection results were intermediate, accordingly, in
line with the theory discussed above. A summary of the data is shown in
Table 12.
Fluxes for this system are given in Figure 34 and were acceptable up to
approximately 10% of the bath concentration (3.5% TDS). Rejection based on
TDS (Figure 35), cadmium (Figure 36), and cyanide (Figure 37) were good up to
nearly double this concentration.
While these results indicate that cadmium cyanide rinse waters can be
treated with reverse osmosis, it is evident that the full bath strength cannot
be obtained. Since cadmium cyanide baths operate at close to ambient tempera-
ture where evaporative losses are small, it is likely that auxiliary evapora-
tion would be required for closed-loop treatment of cadmium cyanide.
ZINC CYANIDE
On the basis of solids content, the results of flux and rejection for the
zinc system (Table 13) were nearly identical for those of the cadmium system.
However, because the solids content of the zinc plating bath was low compared
to the cadmium and copper systems, concentrates of a higher fraction of bath
strength can be successfully generated by RO.
The flux data shown in Figure 38 indicate that performance was acceptable
up to 25 to 30% of bath concentration (approximately 3.2% TDS). Rejection data
given in Figures 39 (total solids), 40 (Zn+2), and 41 (CN~), indicate that re-
jections were good up to 20% of bath strength, and acceptable up to 30%.
Based on these results, zinc cyanide appears to be an attractive candidate
for RO treatment. However, as in the case of cadmium cyanide, zinc cyanide is
an ambient-temperature bath, and auxiliary evaporation would probably be re-
quired for closed-loop treatment.
70
-------
Table 12. Experiment # 6 Cadmium Cyanide Bath
Feed Solutions
% Sol ids W
.31
1.03
2.43
3.12
9.82
% Bath^2^
1
4
9
12
37
Membrane Module
Hollow Fiber
Hollow Fiber
Hollow Fiber
Hollow Fiber
Hollow Fiber
Operating Conditions
Pressure
(psig)
400
400
400
400
400
Temperature
(°c)
28
28
27
27
27
pH of
Feed
11.5
IT. 8
12.2
12.5
12.9
Flux
(gpm)
2.1
1.6
.67
.24
.028
% Rejection
Total Dissolved
Solids
98
89
97
96
9.2
Cd++
99+
99+
99+
99
78
CN"
83
97
95
92
10
(1), (2) See Table 6
-------
3.0
2.5
DUPONTPOLYAMIDE
HOLLOW FIBER PERMEATOR
2.0
0.
1.5
1.0
0.5
10 15 20
% TOTAL DISSOLVED SOLIDS
25
Figure 34. Flux in Cadmium Cyanide Bath
72
-------
z
o
u
UJ
oc
CO
o
If)
10
20
30
40
50
60
70
80
90
94
95
96
97
98
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
OPEN POINTS GIVE THE
REJECTION FOR 1/2% NaCI
SOLUTION BEFORE (LEFT
AXIS) AND AFTER (RIGHT
AXIS) THE TEST.
- /
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
Figure 35. Solids Rejection in Cadmium Cyanide Bath
73
-------
90
o
u
«M
O
ss
99
99.9
I I
10 15
% TOTAL DISSOLVED SOLIDS
20
25
Figure 36. Cd+2 Rejection in Cd(CN)2 Bath
74
-------
90
u
a?
99
99.9
I
I
10 IS
% TOTAL DISSOLVED SOLIDS
20
25
Figure 37. CN~ Rejection in Cd(CN)2 Bath
75
-------
Table 13. Experiment # 7 Zinc Cyanide Bath
Feed Solutions
% Sol ids ^
.47
.77
1.27
2.44
4.05
% Bath(2)
4
7
n
21
36
Membrane Module
Hollow Fiber
Hollow Fiber
Hollow Fiber
Hollow Fiber
Hollow Fiber
Operating Conditions
Pressure
(psig)
400
400
400
400
400
Temperature
(°C)
27
27
27
27
27
pH of
Feed
12.3
12.6
12.8
13.3
13.7
Flux
(gpm)
1.8
1.6
1.2
.58
.21
% Rejection
Total Dissolved
Solids
97
97
96
90
70
Zn
98
99+
99+
99+
98
CN"
85
99+
99
97
97
(1), (2) See Table 6
CD
-------
3.0
2.5
2.0
"fe
X* 1.5
1.0
0.5
DUPONTPOLYAMIDE
HOLLOW FIBER PERMEATOR
I I I I
10 15 20
% TOTAL DISSOLVED SOLIDS
25 30
Figure 38. Flux in Zinc Cyanide Bath
77
-------
V)
60
70
80
95
97
98
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
OPEN POINTS GIVE THE
REJECTION FOR 1/2% NaCI
SOLUTION BEFORE (LEFT
AXIS) AND AFTER (RIGHT
AXIS) THE TEST.
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
Figure 39. Solids Rejection in Zinc Cyanide Bath
78
-------
90
u
UJ
3
K
N
S?
99
99.9
O
O
- O
I
10 15
% TOTAL DISSOLVED SOLIDS
20
25
Figure 40. Zn+2 Rejection in Zn(CN)2 Bath
79
-------
90
K
I
CJ
99
99.9 ' ' '
10 15 20 25
TOTAL DISSOLVED SOLIDS
Figure 41. CN~ Rejection in Zn(CN)2 Bath
80
-------
COPPER CYANIDE
The copper cyanide plating bath, having a solids content 65% higher than
the cadmium cyanide bath and 4.5 times that of the zinc cyanide bath, showed
markedly poorer flux and rejection behavior as a function of fraction of plat-
ing bath strength (see Table 14).
As shown in Figure 42, fluxes for the copper cyanide rinse waters were
acceptable up to a feed concentration of only 7 to 8% of bath strength (approx-
imately 4% TDS). Rejections, shown in Figures 43 (total solids), 44 (Cu ),
and 45 (CN~), were seen to be good up to only 10% of bath strength (5.5% TDS).
Although these results indicate that a high degree of concentration can-
not be obtained with RO, copper cyanide is still an attractive application.
Copper cyanide baths are operated at temperatures between 60 and 82°C (140 and
180°F), and evaporative losses are significant. Therefore, it is still pos-
sible to close the loop with RO even though a highly concentrated stream is
not produced.
ROCHELLE COPPER CYANIDE
Test results for a Roche!le copper cyanide bath are shown in Table 15.
The flux, plotted in Figure 46, remained acceptably high even at 26% of bath
concentration (3.3% TDS).
Rejections, plotted in Figure 47, were exceptionally good, particularly
at the higher feed concentrations. At 26% of bath concentration rejections of
total solids, conductivity, copper, and cyanide were all above 90%. In addi-
tion, this bath operates at elevated temperatures so that a high degree of
concentration is not required. Thus, the standard Rochelle copper cyanide
bath appears to be a very attractive application for closed-loop RO treatment.
RESULTS FOR CYANIDE LIFE TEST
Membrane processing of electroplating rinse waters will be of interest
only if an economically viable membrane life can be obtained. What is
81
-------
Table 14. Experiment # 8 Copper Cyanide Bath
Feed Solutions
% Solids(1)
.57
1.93
3.71
7.98
% Bath(2)
2
5
10
22
Membrane Module
Hollow Fiber
Hollow Fiber
Hollow Fiber
Hollow Fiber
Operating Conditions
Pressure
(psig)
400
400
400
400
Temperature
(°C)
26
26
26
27
pH of
Feed
11.8
12.2
12.5
12.9
Flux
(gpm)
1.8
1.2
.62
.076
% Rejection
Total Dissolved
Solids
98
98
97
77
Cu++
99+
99+
99+
84
CN"
92
99+
99+
92
(1), (2) See Table 6
CO
IN3
-------
3.0
2.5
DUPONT POLYAMIDE
HOLLOW FIBER PERMEATOR
2.0
1.5
1.0
0.5
I I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25 30
Figure 42. Flux in Copper Cyanide Bath
83
-------
10
20
30
40
50
60
80
A SPIRAL WOUND
• TUBULAR
• HOLLOW FIBER
OPEN POINTS GIVE THE
REJECTION FOR 1/2% NaCI
SOLUTION BEFORE (LEFT
AXIS) AND AFTER (RIGHT
AXIS) THE TEST.
£ 90
CO
a
1MB
_l
e
CO
3? 94
95
96
97
9.'
99
I
I
10 15 20
% TOTAL DISSOLVED SOLIDS
25
Figure 43. Solids Rejection in Copper Cyanide Bath
84
-------
90
oc
+
3
CJ
ss
99
99.9
I
I
10 15
% TOTAL DISSOLVED SOLIDS
20
25
Figure 44. Cu Rejection in CuCN Bath
85
-------
90
-O
I
z
o
99
99.9
I
I
10 15
% TOTAL DISSOLVED SOLIDS
20
25
Figure 45. CN~ Rejection in CuCN Bath
86
-------
Table 15. Experiment # 11 Rochelle Copper Cyanide Bath
Feed Solutions
% Solids
3.34
1.35
.132
% Bath
26.3
10.6
1.04
Membrane
Module
Hollow Fiber
Hollow Fiber
Hollow Fiber
Operating Conditions
Pressure
(psig)
400
400
400
Temperature
(°C)
25
28
28
pH of
Feed
10.6
10.1
9.8
Flux
(gpm)
1.60
2.06
2.54
%
Total Dissolved
Solids
98.5
98.1
96.0
Rejection
Conductivity
93
92
82
Cu+
96.6
97.3
89.6
CN"
94.9
94.7
64.7
CO
-------
2.0
1.0
0.5
j | L
10 15 20
% TOTAL DISSOLVED SOLIDS
25 30
Figure 46. Flux in Rochelle Copper Cyanide Bath
-------
0
30
60
HOLLOW FIBER MODULE
z
o
o
90
93
96
99
99.3
99.6
99.9
I
O TOTAL DISSOLVED SOLIDS
D CU+
A CN~
CONDUCTIVITY
I
I
10 15
TOTAL DISSOLVED SOLIDS
20
25
Figure 47. Rejection in Rochelle Copper Cyanide Bath
89
-------
economically attractive depends, of course, on overall process economics. As
a guide, however, a membrane life of six months or more would seem desirable.
In order to study long-term irreversible changes in membrane performance,
an extended test was performed using a zinc cyanide bath at one-tenth bath con-
centration. Usually, the system was operated for nine hrs/day, five days/week.
However, the permeator remained in contact with the plating solution at all
other times. For most of the life tests, pH was adjusted to 11.5 by addition
of H2S04.
Figure 48 summarizes the history of the hollow-fiber module used in the
life test. The module was exposed to various plating bath solutions ranging in
pH from 3.3 to 13.9, and ranging in solute concentrations from 0.3% up to 24%,
for a total of over 800 hrs. During approximately half of this period, the
2
system was in operation at 29.1 kg/cm (400 psig). Flux and rejection data for
a 0.5% NaCl solution were obtained periodically throughout the life test, and
are shown in Table 16. Flux and rejection remained essentially unchanged un-
til sometime between points H and I (i.e., about 500 hrs of exposure to various
plating bath solutions), when a decrease in both occurred.
During the zinc cyanide life test, membrane flux and total solids rejec-
tion were measured daily. These are shown in Figures 49 and 50. The gaps
between groups of points indicate interruptions in the life tests for various
reasons. Figure 49 shows that flux remained relatively unchanged over the en-
tire life test period. However, the total solids rejection data (Figure 50)
showed a significant decline starting at about 500 hrs of exposure.
These life test data with the cyanide bath are considered to be promising,
and suggest that if operation is at controlled pH, membrane life would be
satisfactory.
90
-------
14
12
0
S 10
o
W
K 8
0
^fff
D E
WMWM
< A
UJ
ec
S 6
X
o.
2
p
i
0
|
1
|
1
L
!
7
/
mt
t
•••
^H
k ,
F
mm
7
/
mm
T
y
f
/
••i
4
1
7
>
r
mti
5
G
/» j
s\
G
6
2
100
H 1
1
T
'//////X \/////////)(
7 8
I
J K
10
9
1 1 1 1
1 1
200 300 400 500 600 700 80
NUMBER OF HOURS OF EXPOSURE
Figure 48. Life Data of B-9 Permeator #3
-------
Table 16. Guide to Figure 48
System pressurized and in operation
System exposed to solution at ambient conditions
1
2
3,6
4,7,8,10=
5
9
= Ni (OS02NH2)2 solution
ZnCl2 solution
Cd(CN)2 solution
Zn(CN)2 solution
CuCN solution
Ni (BF4)2 solution
0.5% NaCl Flux, gpm
0.5% NaCl Rejection
A =
B =
C =
D =
E =
F =
G =
H =
I =
J =
K =
2.50
2.10
1.80
2.60
2.40
2.35
2.20
2.15
1.45
1.20
1.60
96%
97%
98%
98%
98%
98%
97%
97%
90%
90%
88%
92
-------
co
1.3
1.2
1.1
a.
a
x"
1.0
0.9
0.8
A
L AA
100
AA A AAA A
-AA—A AAAA
A AA
AA
AA
A AAA
AA A
AA
J_
J_
I
l
200 300 400 500 600
NUMBER OF HOURS OF EXPOSURE TO PLATING SOLUTIONS
700
800
Figure 49. Flux Data in Zn(CN)5 Life Test
-------
98
96
o
" 94
in ~~
-------
SECTION VI
REFERENCES
1. Coulomb, A. Plating. 57:1001. 1970.
2. Coulomb, A. ibid. 59:316, 1972.
3. Spatz, D.D. Product Finishing. 36:79, August 1972.
4. Spatz, D.D. Finisher's Management. 17, July 1972.
5. Merten, U. "Desalination by Reverse Osmosis". The M.I.T. Press,
Cambridge, Massachusetts, 1966.
6. Metal Finishing Guidebook for 1971, (39th edition) Metals and
Plastics Publications, Incorporated, Westwood, New Jersey, 1971.
7. Arinet, R.C. The Modern Electroplating Laboratory Manual.
Robert Draper Ltd., publishers. Teddington 1965.
8. Cyanide Zinc Plating. American Electroplater's Society,
Incorporated. East Orange, New Jersey, 1970.
9. Berg, J.H. "Clarification and Purification of Solutions".
Metal Finishing. 12:48-51, December 1972.
10. Filtration and Carbon Treatment of Plating Solutions. American
Electroplating Society, Incorporated. East Orange, New Jersey.
11. "Standard Methods for the Examination of Water and Waste Water",
APHA, 13th Ed., 1971.
95
-------
TECHNICAL REPORT DATA
(Please read Inuructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-76-26.
3. RECIPIENT'S ACCESSION-NO.
-I. TITLE AND SUBTITLE
TREATMENT OF ELECTROPLATING WASTES BY REVERSE
OSMOSIS
5. REPORT DATE (Issuing
September 1976
6. PERFORMING ORGANIZATION CODE
7-AUTHOR(3) Richard G. Donnelly, Robert L. Goldsmitl
Kenneth J. McNulty, Donald C. Grant, Michael Tin
8. PERFORMING ORGANIZATION REPORT NO.
5*
9. PERFORMING ORG "VNIZATION NAME AND ADDRESS
The American Electroplaters' Society
56 Melmore Gardens
East Orange, New Jersey 07017
10. PROGRAM ELEMENT NO.
1BB03.6
11. CONTRACT/GRANT NO.
R-800945-01
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Laboratory - Gin., OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
* Walden Research Division of Abcor, Inc.
201 Vassar Street, Cambridge, Massachusetts
02139
16. ABSTRACT
Reverse osmosis treatment of plating bath rinsewaters has been examined.
Emphasis has been placed on closed-loop operation with recycle of
purified water for rinsing, and return of plating chemical concentrate
to the bath. Three commercially available membrane configurations have
been evaluated experimentally;tubular (cellulose acetate membrane),
spiral-wound (cellulose acetate membrane), and hollow-fiber (polyamide
membrane). Tests were conducte'd with nine different rinsewaters preparec
by dilution of actual plating baths. Advantages and limitations of the
reverse osmosis process and specific membranes and configurations are
discussed. Promising, as well as unattractive applications are indicatec
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
*Electroplating, Industrial Waste Treat-
ment, Industrial Water,* Membranes,*
Osmosis,* Semipermeability, Water
Pollution
Polymer membranes
Reverse osmosis
Electroplating Wastewatei
13/B
3. DISTRIBUTION STATEMENT
.Release to Public
19. SECURITY CLASS (This Report)
ITnp-
i f i
21. NO. OF PAGES
106
>0. SECURITY CLASS (Thispage)
22. PRICE
EPA Form 2220-1 (9-73)
96
. GOVERNMENT PRINTING OFFICE: 1977-757-056/5526 Region No. 5-11
------- |