United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Cincinnati OH 45268
Research and Development
EPA-600/S2-81-132 Sept. 1981
Project Summary
Demonstration of Zinc
Cyanide Recovery Using
Reverse Osmosis and
Evaporation
Kenneth J. McNulty and John W. Kubarewicz
A field test was conducted to
demonstrate closed-loop recovery of
zinc cyanide rinsewater at a job shop
plating facility. Since the zinc cyanide
bath operates at room temperature
with very little evaporation from the
bath, reverse osmosis (RO) treatment
of the rinsewater must be supple-
mented by evaporation in order to
achieve the volume reduction neces-
sary for return of a concentrate to the
plating bath. The permeate from the
RO unit was recycled to the first rinse
after plating, while the distillate from
the evaporator was recycled to the
second rinse after plating. Contin-
uous, unattended operation of this
system was demonstrated with no
adverse effects on plating quality.
Spiral-wound PA-300 membrane
modules were used in the RO unit.
Periodic tests were conducted through-
out the demonstration to characterize
membrane performance under stan-
dard conditions. These tests indicated
a gradual loss in membrane flux and
rejection. After 3,000 hours of ex-
posure to the rinsewater, the mem-
branes were cleaned byflushing with a
cleaning solution. The cleaning resulted
in nearly complete restoration of flux
and rejection. The gradual loss in
membrane performance is thus attrib-
utable to fouling of the membrane by
particulates in the rinsewater. Such
fouling can be reduced by better pre-
filtration and reversed by periodic
cleaning.
The economics of the combined RO
evaporation system were assessed for
a system designed to provide rinsing
equivalent to the present two-stage
counter-current rinse at the demon-
stration site. The analysis showed that
the total operating cost (including
amortization) was somewhat less for
the combined RO evaporation system
than for evaporation alone. The mini-
mum cost occurred for 90% water
recovery in the RO system. However,
credits for rinsewater recovery were
insufficient to completely off-set the
total operating cost of the recovery
system.
This report was submitted in fulfill-
ment of EPA Grant Number R805300
by The American Electroplaters'
Society, Inc. (AES) under the partial
sponsorship of the U.S. Environmental
Protection Agency. This report covers
the period August 24, 1977, to
November 24. 1978, and work was
completed as of February 7, 1979.
This Project Summary was devel-
oped by EPA's Industrial Environmen-
tal Research Laboratory. Cincinnati,
OH. to announce key findings of the
research project that is fully docu-
mented in a separate report of the
same title (see Project Report ordering
information at back).
-------
V -/
? ' *
e:-.
f "*"
Introduction
Wastewater treatment technologies
for the electroplating industry can be
broadly classified as end-of-pipe de-
struction processes or in-plant recovery
processes. The end-of-pipe destruction
processes treat a total shop effluent to
remove a mixture of heavy metals. At
present it is neither technically nor
economically feasible to recover and
recycle metals from the end-of-pipe
processes (1). In-plant recovery
processes, however, treat rinsewater
from a specific plating bath (or other
operation) making it possible to recover
and return the heavy metals to the
plating bath.
Because of the inherent disadvantage
of end-of-pipe treatment — loss of
valuable plating chemicals, cost of
treatment chemicals, cost of sludge
disposal — increasing attention has
been focused on closed-loop recovery
methods. In many cases, the economics
of closed-loop recovery have been very
favorable, resulting in rapid payback on
the capital investment for recovery
equipment (2).
Aside from a few applications in
which closed-loop recovery can be
achieved by countercurrent rinsing
alone, some technique must be used to
remove the dissolved plating chemicals
from the rinsewater. Although other
techniques are under development,
evaporation, reverse osmosis (RO), and
ion exchange are the most commonly
used piocesses for rinsewater recovery
.(1,3). Each of these techniques has
particular advantages and disadvantages
and the best technique or combination
of techniques will depend on factors
specific to each application.
A number of advantages can be cited
for the use of RO in rinsewater recovery.
These include low capital cost, low
energy and operating costs, and minimal
space requirements. However, there
are also some limitations. The major
limitations for RO are:
1. The membrane modules deterio-
rate with time and require periodic
replacement. The rate of deterio-
ration depends on the type of
membrane, the rinsewater pH and
temperature, and the concentration
of other reactants in the rinsewater
such as oxidants.
2. Reverse osmosis cannot produce a
highly concentrated stream for
recycle to the plating bath. Thus
for ambient temperature baths,
RO must be supplemented with
some other concentration tech-
nique, such as evaporation, in
order to close the loop.
To date, RO has been applied primarily
to the recovery of nickel rinsewaters.
For nickel, the rinsewaters are relatively
mild in pH (4-6) resulting in acceptable
life for the conventional commercial
membranes (cellulose acetate and
aromatic polyamide). In addition, nickel
baths operate at elevated temperatures
where substantial evaporation occurs,
and closed-loop operation can be
achieved with RO alone.
Several programs, jointly sponsored
by EPA and AES, have been conducted
to evaluate the applicability of RO to
plating baths other than nickel (4,5,6).
Laboratory tests were conducted with a
variety of newly developed membranes
and rinsewaters with extreme pH levels
(6). These tests indicated that of the
membranes tested, the PA-300 was
superior to the other membranes for
treatment of copper cyanide, zinc
cyanide, and chromic acid rinsewaters.
The PA-300 membrane has since been
commercialized (currently designated
TFC-PA; manufactured by Fluid Systems
Division of UOP) and is available in a
spiral-wound modular configuration.
A field test was undertaken to
evaluate the PA-300 membrane module
for recovery of zinc cyanide rinsewater
under realisitic conditions. Zinc cyanide
was selected because of the large
volume of zinc cyanide plating done by
the industry and because the high pH of
the rinsewaters would provides "worst
case" test of the membrane for resist-
ance to alkaline conditions. Since the
zinc cyanide bath operates at room
temperature, it was necessary to use an
evaporator to supplement RO treatment
and achieve the level of concentration
necessary for closed-loop operation.
This report summarizes and discusses
the results of this field test.
Methods and Materials
A mobile RO test system was leased
from Abcor, Inc., and an evaporator was
leased from Wastesaver Corporation for
the duration of the field test. These two
units were installed on an automatic
rack, zinc cyanide plating line at New
England Plating Co. in Worcester,
Massachusetts. The overall schematic
of the installation is shown in Figure 1.
Feed to the RO system was withdrawn
from Rinse Tank #1 and separated by
the RO system into a permeate stream
and a concentrate stream. For purposes
of design, it was assumed that the RO
system would produce about 7.5 Ipm (2
gpm) of permeate and would operate at
90% conversion. (Conversion is defined
as the ratio of permeate flow to feed
flow.) Thus the RO system would be fed
at the rate of 8.4 Ipm (2.22 gpm) and
would produce concentrate at the rate
of .8 Ipm (0.22 gpm). The permeate was
returned to Rinse Tank #1 and the
concentrate was fed to the evaporator.
Since drag-in and drag-out were
essentially identical for the plating bath
and the rate of evaporation was neg-
ligible, there was no room in the plating
bath for a concentrate stream. If the
evaporator were fed only RO concen-
trate, it would have to evaporate it to
dryness in order to prevent eventual
overflow of the bath. In order to prevent
precipitation of plating chemicals in the
evaporator a 3.8 Ipm (1 gpm) purge
stream was circulated from the plating
bath through the evaporator and carried
the plating salts introduced with the RO
concentrate back to the plating bath.
That is, the evaporator concentrate was
higher in concentration than the plating
bath by the amount added by the RO
concentrate. The distillate from the
evaporator was collected in a holding
tank and added at a controlled rate to
Rinse Tank #2. A float valve operating
off the level in Rinse Tank #1 insured
that the rate of RO concentrate pro-
duction was exactly balanced by the rate
of distillate returned to Rinse Tank #2. A
slight excess of distillate was produced
to insure that the holding tank would
always remain full; and the excess was
permitted to overflow into the plating
bath (.08 Ipm). The steam rate was cut
back to minimize overflow from the
holding tank.
A flow schematic of the RO system is
shown in Figure 2. Feed from the first.
rinse tank was withdrawn by a booster
pump and passed through two string-
wound cartridge filters in parallel. Both
1-/u and 20-fj filters were used at
different times during the field test.
Excess flow from the booster pump was
returned to the rinse tank. After pre-
filtration, the feed was pressurized to
48.6 atm (700 psi) with a multi-stage
centrifugal feed pump and passed
through three 102 mm (4-inch)diameter,
spiral-wound, PA-300 modules arranged
in series. Most of the concentrate frpm
the third module was recycled to the
suction of the feed pump in order to
maintain the required feed flow rate>
through the modules. A heat exchanger*
-------
Distillate
0.90 Ipm
(0.24 gpm)
Overflow
0.075 Ipm
(0.02 gpm)
Plating Bath
Evaporator
Concentrate
3.8 Ipm
(1.0 gpm)
Distillate
0.83 Ipm
(0.22 gpm)
Rinse
Tank #2
Evaporator
Feed
3.86 Ipm
(1.02 gpm)
EO
Feed
8.4 Ipm
(2.22 gpm)
Permeate
7.8 Ipm
(2 gpm)
RO
Concentrate
0.83 Ipm
(0.22 gpm)
Figure 1. Overall schematic of RO/evaporator operation.
in the recirculation loop removed heat
generated by the energy input of the
pumps. A small flow of concentrate
from the third module was fed to the
evaporator (see Figure 1), and the
permeate from the three modules was
combined and returned to the first rinse
tank. The instrumentation and controls
for the RO system are shown in Figure
2.
In order to characterize membrane
performance with a standard feed
solution, the RO system was periodically
operated in a total recycle mode using
the auxiliary feed tank. For this mode of
ration, the booster pump recycle line
was closed off, the concentrate line to
the evaporator was opened, and the
permeate was returned to the auxiliary
tank rather than the rinse tank. The
standard solution (generally a portion of
plating bath diluted to 10% by volume or
original bath strength) was charged to
the auxiliary tank and the system was
operated with total recycle until steady
state was achieved. At steady state, the
permeate flow rate for each module was
measured, and samples of the feed and
permeate from each module were
obtained for analysis.
Typical operating conditions for both
closed-loop and total recycle were:
Feed Pressure 48.6 atm (700 psi)
Recirculation
Flow Rate 37.8 Ipm (10 gpm)
Temperature 21-32°C (70-90°F)
Concentrate 0.75 Ipm (0.2 gpm)
Flow Rate closed-loop only
The flow schematic for the evaporator
is shown in Figure 3. Steam was fed
through a pressure reducing valve to a
tube bundle submerged in the boiler
section of the evaporator, and steam
condensate was returned to the plant
boiler. For most installations, a cooling
tower is used to cool the water which is
recirculated through the condenser
section of the evaporator. However, for
this installation it was more convenient
to use recirculated chilled water since it
was readily available at the installation
site and the chiller had sufficient excess
capacity. The evaporator was main-
tained under vacuum by circulating
water through an eductor. Cooling
water was added to the eductor tank to
remove the energy input of the eductor
circulation pump. Feed to the evaporator
was controlled by a level switch (LS) and
solenoid valve. Upon low level signal,
the solenoid valve opened and feed was
drawn by vacuum into the evaporator.
The distillate from evaporation of the
feed condensed, was collected in a tray
below the condenser, and was contin-
uously pumped back to the second rinse
after plating (see Figure 1). The con-
centrate from the boiler section of the
evaporator was continuously pumped
back to the plating bath.
Typical operating conditions for the
evaporator were:
Vacuum 0.87 atm abs.
(26-27 in. Hg)
Temperature 38-43°C
(100-110° F)
<1.3 atm abs.
«19.7psig)
Steam
Pressure
Concentrate
Flow Rate
3.8 Ipm (1 gpm)
During the field test, the RO modules
were cleaned using a cleaning sequence
recommended by the membrane manu-
facturer. The modules were first flushed
with 189 liters (50 gal) of water to
remove the plating chemicals. A 0.1%
by volume solution of Triton X-100, a
non-ionic surfactant, was prepared and
recirculated through the modules at a
pressure of 48.6 atm abs (700 psi), a
flow rate of 37.9 Ipm (10 gpm), and a
temperature of 49°C (120°F) for 45
minutes. After flushing with another
-------
Permeate
Concentrate to Evaporator
A ux.
Tank
P - Pressure Gauge
AP - Differential Pressure Gauge
PS - Pressure Switch
T - Temperature Gauge
TS - Temperature Switch
TS - Temperature Wtitch
F - Flow Meter
CC - Conductivity Cell
Figure 2. Flow schematic for RO demonstration system.
189 liters of water, a 2% citric acid
solution was prepared and adjusted to
pH 3.0 with ammonium hydroxide. This
solution was recirculated through the
modules at the same conditions and for
the same time as the Triton X-100.
Following the citric acid cleaning the
system was again flushed with water
and returned to treatment of zinc
cyanide rinsewater. Since the PA-300
membrane is rapidly degraded by
chlorine, all water used for flushing and
preparing cleaning solutions was de-
chlorinated by the addition of sodium
sulfite.
Samples collected during the field
test were analyzed for zinc (atomic
absorption), free cyanide (selective ion
electrode), total solids (gravimetic
determination of residue), conductivity
(conductivity bridge), and pH (electrode).
The nominal composition of the
plating bath was:
CN (as NaCN)
Caustic
Brightener
(700 Special)
60,000 mg/l
8.0 oz/gal
75,000 mg/l
10.0 oz/gal
4 ml/I
4 gal/1000 gal
Zn (as metal)
20,000 mg/l
2.7 oz/gal
In addition to these compounds, poly-
sulfide was regularly added to the bath
for purification, and the bath also
contained a large quantity of carbonates.
The total solids concentration of the
bath was in the vicinity of 350,000 mg/l
(35% by weight).
Conclusions
Closed-loop recovery of zinc cyanide
rinsewaters can be achieved with a
combined RO/evaporator system. Con-
tinuous and unattended operation of the
system was demonstrated over one-
week periods (Monday through Friday,
three shifts per day). No adverse effects
on plating quality were noted during the
demonstration.
The single-effect vacuum evaporatoi
was operated at about one-half of its
rated capacity, and the vacuum was
sufficient to keep the vaporizatior
temperature below 43°C (110°F). Tht
quality of the distillate was considerec
quite suitable for final rinsing.
Tests were conducted periodically t<
determine the flux and rejection per
formance of the three PA-300 spiral
wound modules used in the RO system
Results for Modules #1 and #2 were it
close agreement throughout the dem
onstration. Module #3 was concluded t
be defective from the outset of thi
program and was removed after 3,00l
hours.
The flux and rejection of Modules #
and #2 gradually declined during th
demonstration. Over the first 3,00
hours of exposure, the flux declined t
about one-half of its original value an
the zinc rejection, for example, decline
from 99% to 97%. At 3,000 hours, th
membranes were cleaned by flushin
with a cleaning solution. The clean!
-------
T - Temperature Gauge
P - Pressure Gauge
F - Flow Meter
LW - Level Switch
Chilled
Water
Steam
Supply
Condensate
Return
Pressure
Reducing
Valve
\
r
Eductor
Tank
r-X— Cooling Water
Drain
Distillate to
2nd Rinse
Plationg
Bath
Concentrate 4*
Feed T
Concentrate
from RO
Figure 3. Flow schematic for evaporator.
resulted in nearly complete restoration
of flux and rejection. The gradual loss in
membrane performance is thus attrib-
utable to fouling of the membrane by
participates in the rinsewater. Such
fouling can be reduced by better pre-
filtration and reversed by periodic
cleaning.
Following the field test, Module #2
was cut open and unwound for in-
spection. The membrane was fouled
with a thin layer of what appeared to be
sulfide sludge. (Polysulfide is used as a
bath purifier). Examination of the
module internals revealed that some
deterioration in the strength of the
membrane backing material had oc-
curred. However, this produced no
gross effects on the performance of the
modules as observed during the field
test.
The economics of the combined RO
evaporation system were assessed for a
system designed to provide rinsing
equivalent to the present two-stage
counter-current rinse at the demonstra-
tion site. The analysis"bhowed that the
total operating cost (including amorti-
zation) was somewhat less for the
combined RO evaporation system than
for evaporation alone. Total operating
costs were calculated for various RO
system water recoveries, and the
minimum cost occurred for 90% water
recovery in the RO system. However,
credits for rinsewater recovery were
insufficient to completely off-set the
total operating cost of the recovery
system. Energy costs for evaporation
using a double effect evaporator were 4-
5 times greater than those for RO.
Continuing escalation of energy costs
would provide further incentive for a
combined RO evaporation system.
Recommendations
On the basis of this field test and
previous laboratory tests (6) the PA-300
membrane can be recommended for the
treatment of cyanide rinsewaters on a
commercial scale. Application to copper
cyanide recovery appears to be partic-
ularly attractive because of: 1 )the lower
pH of copper cyanide relative to zinc
cyanide, 2) the higher value of copper
relative to zinc, and 3) the higher
evaporation rate from the copper bath
which would permit closed-loop opera-
tion with RO alone (no evaporator
required).
Additional development of the PA-
300 or similar membranes is recom-
mended with particular emphasis on
the development of modules containing
materials which are highly resistant to
pH extremes. In particular, a more
alkaline-resistant membrane backing
material should be developed for
cyanide applications.
Data should be obtained on the life of
PA-300 (or similar) modules for longer
exposure times (>4,200 hours). This
data can be conveniently and inex-
pensively obtained by "soaking" the
module in a solution of the plating bath
diluted to simulate concentrated rinse-
water. The flux and rejection of the
modules can then be determined
-------
periodically. These soak tests will
provide information on the resistance of
various module components to attack by
the major constituents of the bath (e.g.
OH').
In view of the good acid resistance
exhibited by the PA-300 membrane in
laboratory tests (6), it is recommended
that a field test and long-term soaktests
be conducted using PA-300 modules
and low pH rinsewater such as acid
copper. (In future field tests, the feed to
the modules should be pretreated by
passage through a ]/j filter. The objec-
tive of these tests would be to extend the
applicability of RO to pH levels below 2.5
(the lower limit for cellulose acetate
membranes).
For evaporative recovery, develop-
ment of a low capacity, low cost,
mechanical vapor recompression
evaporator is recommended. This type
of evaporator is highly efficient and
could be operated at a significantly
lower cost than the combined RO/
evaporator system tested in this program.
References
1. Skovronek, H.S., and M.K. Stinson.
Advanced Treatment Approaches for
Metal Finishing Wastewaters (Part
II). Plating and Surface Finishing, 64
(11): 24-31, 1977.
2. Anonymous. Recovery Paysl Plating
and Surface Finishing, 66 (2): 45-48,
1979.
3. Hall, E.P., D.J. Lizdas, and E.E.
Auerbach. Plating and Surface
Finishing, 66 (2); 49-53, 1979
4. Donnelly, R.G., R.L. Goldsmith, K.J.
McNulty, and M. Tan. Reverse
Osmosis Treatment of Electroplating
Wastes. Plating, 61 (5): 422-432,
1974.
5. McNulty, K.J., R.L. Goldsmith, A
Gollan, S. Hossain, and D. Grant.
Reverse Osmosis Field Test: Treat-
ment of Copper Cyanide Rinse
Waters. EPA-600/2-77-170, U.S.
Environmental Protection Agency,
Cincinnati, Ohio, 1977. p. 89.
6. McNulty, K.J., P.R. Hoover, and R.L.
Goldsmith. Evaluation of Advanced
Reverse Osmosis Membranes for
the Treatment of Electroplating
Wastes. In: First Annual Conference
on Advanced Pollution Control for
the Metal Finishing Industry. EPA-
600/8-78-010, U.S. Environmental
Protection Agency, Cincinnati, Ohio,
1978. pp. 66-75.
Kenneth J. McNulty and John W. Kubarewioz are with Walden Division of Abcor,
Inc., Wilmington. MA 01887.
Mary Stinson is the EPA Project Officer (see below).
The complete report, entitled "Demonstration of Zinc Cyanide Recovery Using
Reverse Osmosis and Evaporation," (Order No. PB 81-231 243; Cost: $6.50,
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:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
•ft U S GOVERNMENT PRINTING OFFICE, 1981 — 757-012/7360
-------
United States Center for Environmental Research Feesfpaid"
Environmental Protection Information cn?,i,«nm»r>»
Agency Cincinnati OH 45268 Protection
Agency
EPA 335
Official Business
Penalty for Private Use $300
RETURN POSTAGE GUARANTEED _. . . _.
Third-Class
Bulk Rate
US tHA KtblUN V
23U b UtAHdUHiM ST
CHICAGO 1L 6060a
------- |