-
. . . . .
i .
... .'.?.
0
TY, K.J. DR.; KUBAREWICZ, J.W. REVERSE OSMOSIS FIELD TEST:
CLOSED LOOP RECOVERY OF ZINC CYANIDE RINSEWATER AES RESEARCH
PROJECT 45 SUR/FIN 79, SESSION L.
CLOSED-LOOP TREATMENT OF ZINC CYANIDE RINSEWATERS USING REVERSE
OSMOSIS FOR CONCENTRATION OF THE DILUTE RINSEWATERS FOLLOWED BY
THE CONCENTRATE TO THE PLATING BATH. THE TECHNICAL FEASIBILITY
OF THE CONCEPT WAS SUCCESSFULLY DEMONSTRATED USING THE PA-300
MEMBRANE
k
A FIELD TEST HASBEEN CONDUCTED TO DEMONSTRATE THE FEASIBILITY
OF
, EVAPORATION TO ACHIEVE THE VOLUME REDUCTION REQUIRED TO
RETURN
,*...Î. -+. -. .__II_C
I
Ir\ 3
-
I a ,
ABSTRACT
A f i e l d t e s t was conducted t o demonstrate closed-loop
recovery of zinc , cyanide rinsewater a t a job shop plating f a c
i l i t y . Since the zinc cyanide
b a t h operates a t room temperature with very l i t t l e
evaporation from the bath, reverse osmosis (RO) treatment of the
rinsewater m u s t be supplemented by
.. evaporation i n order to achieve the volume reduction
necessary for return of ' a concentrate t o the plating bath. The
permeate from the RO u n i t was recycled t o the f i r s t rinse a
f t e r p l a t i n g while the d i s t i l l a t e from the
evaporator was recycled to the second rinse a f t e r plating.
Continuous, unattended operation of this system was demonstrated w
i t h no adverse e f f ec t s on plating quali ty.
%.
%
Spiral-wound PA-300 membrane modules were used i n t e R O u n '
' , tests were conducted throughout the demonstration t o char P e
r i ze meubrane
performance under standard conditions. These t e s t s indicated
a gradual loss in membrane f l u x and rejection. After 3,000 hours
of exposure t o the rinse- water, the membranes were cleaned by
flushing w i t h a cleaning solut ion. The cleaning resulted i n
nearly complete restoration o f f l u x and reject ion. The gradual
loss i n membrane performance i s thus a t t r i bu tab le t o
fouling of the membrane by par t iculates i n the rinsewater. Such
fouling can be reduced by better pre-fi 7 t ra t ion and reversed
by periodic cleaning.
Peri odi c ."
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 r inse a t the demonstration s i
t e . The analysis showed t h a t the total operating cost
(including amor t i za t ion ) was somewhat l e s s f o r the
combined RO evaporation system t h a n for evaporation alone, The m
i n i m u m cost occurred for 90% water recovery i n the RO system.
recovery were insuf f ic ien t t o completely of f - se t the to ta
l operating "cost of the recovery system.
However c red i t s f o r rinsewater
1
I
-
I
i n acceptable l i f e f o r the conventional commercial
membranes' (cel lulose ace- t a t e and aromatic polyamide). I?
addition, nickel baths operate a t elevated temperatures where
substanti a1 evaporation occurs, and' closed-loop operation can be
achieved with RO alone.
INTRODUCTION 1 Wastewater treatment technologies fo r the
electroplat ing indus t ry can
be broad1.y c l a s s i f i ed a s end-of-pipe destruction
processes or i n - p l a n t recOv, ery processes. eff luent t o
remove a mixture of heavy metals. A t present i t i s neither tech-
nical ly nor economically feas ib le t o recover and recycle metals
from the end- of-pipe processes (1). On the other hand, in-plant
recover processes treat rinsewater from a spec i f ic plating b a t
h (o r other operation J making i t pas- s ib l e t o recover and
return the heavy metals t o the plat ing bath.
Because of the inherent disadvantage of end-of-pi pe treatment
-- loss of valuable plat ing chemicals, cost o f treatment
chemicals, cost o f sludge disposal -- increasing at tent ion has
been focused on closed-loop recovery methods. favorable resulting i
n rapid payback on the capi ta l investment for recovery equipment
(2 ) .
The end-of -pipe destruction processes t r e a t a to ta l
shop
In many cases, the economics of closed-loop recovery have been
very
Aside from a few applications i n which closed-loop recovery can
be achieved by countercurrent r i n s i n g along, some technique m
u s t be used to remove the dis. solved plating chemicals from the
rinsewater. under development, evaporation, reverse osmosis ( R O )
, and i on exchange are the most commonly used processes for
rinsewater recovery (1 ,3 ) . Each of these techniques has par t
icu lar advantages and disadvantages, and the best technique o r
combination of techniques will depend on factors spec i f i c to
each a p p l i - cation.
A1 though other techniques are
A number of advantages can be ci ted fo r the use of R O i n
rinsewater
I recovery. and minimal space requirements. two major l imitat
ions fo r RO a r e : These include low capi ta l cos t , low energy
and operating costs,
However, there are a lso some limitations. The
1 . The membrane modules deter iorate w i t h time and require
periodic replacement. The r a t e of deter iorat ion depends on the
type o f membrane, the rinsewater pH and temperature, and the
concentration of other reactants i n the r k e w a t e r such as
oxidants.
Reverse osmosis cannot produce a highly concentrated stream f o
r recyle t o the plating b a t h . baths, RO mus t be supplemented
w i t h some o ther concentration technique, such as evaporation, i
n order t o c lose the loop.
t - - 2
T h u s f o r ambient temperature
I
2
-
Several programs, j o i n t l y sponsored by EPA and AES, have
been conducted to evaluate the appl icabi l i ty of RO to plating
baths other than nickel Laboratory t e s t s were conducted w i t h
a variety of newly developed membranes and rinsewaters w i t h
extreme pH levels (6). These t e s t s indicated t h a t of the
mem- branes tested, the PA-300 was superior t o the other membranes
fo r treatment of copper cyani de I) zi nc cyandi e , and chromi c
acid rinsewaters. has since been commercialized (current ly
designated TFC-PA; manufactured by Fluid Systems Division of LOP)
and i s avai lable i n a spiral-wound modular configuration.
(4,5,6).
The PA-300 membrane
: i
t ’ ’ I
A f i e l d t e s t was undertaken to evaludte the PA-300
membrane module f o r recovery of z i nc cyani de ri nsewater under
r e a l i s t i c condi ti ons. was selected because of the large
volume of zinc cyanide plating done by the industry and because the
h i g h pH of the rinsewaters would provide a ”worst case” test of
the membrane f o r res is tance t o a lkal ine conditions. Since
the zinc cyanide bath operates &t room temperature, i t was
necessary to use an evaporator t o supplement RO treatment and
achieve the level o f concentration necessary for closed-loop
operation. this f i e ld t e s t .
Z i nc cyanide
T h i s paper presents and discusses the r e su l t s o f
I I I
I
! i
!
METHODS AND MATERIALS
! I ,
A mobile RO t e s t system was leased from Wbcor, Inc. and an
evaporator was leased from ljlastesaver Corporation f o r the
duration o f the f i e l d t e s t . These two units were Instal
led on an automatic rack, zinc cyanide plating line a t New England
Plating Co. i n Worcester, Massachusetts. The overal schematic of
the ins ta l la t ion i s shown i n Figure 1 . Rinse T a n k if”l
and separated by the RO system into a permeate stream and a con-
centrate stream. For purposes of design, i t was assumed t h a t
the RG system would produce about 2 gpm of permeate and would
operate a t 90% conversion. (Conver- sion i s defined as the r a t
i o of permeate flow to feed flow.) would be fed a t the rate of
2.22 gpm and would produce concentrate a t the ra te of 0.22 gpm.
uas fed to the evaporator.
Feed t o the RO system was w i t h d r a w n from
T h u s the RO system
The permeate was returned t o Rinse Tank #’I and the
concentrate
t ! I ‘
Since d r a g i n and dragout were e s sen t i a l ly ident ical
fo r the plat ing b a t h and the r a t e of evaporation was negl
igible , there was no room in the pla%i-wj b a t h for a
concentrate stream, it would have to evaporate i t t o dryness i n
order to prevent eventual overflow Of the b a t h . I n order to
prevent precipi ta t ion of plating chemicals i n the evaporator a
7 gpm purge stream was circulated from the plating b a t h t h r o
u g h the evaporator and carried the plating s a l t s introduced w
i t h the RO concentrate back t o the plating bath. T h a t i s ,
the evaporator concentrate was higher i n concentration t h a n the
plating bath by the amount added by the RO concentrate. The d i s t
i l l a t e from the evaporator was col lected i n a holding t a n
k and added a t a controlled r a t e t o Rinse Tank #2. in Rinse
Tank $1 insured tha t the ra te of ROt~oncentrate production was
exactly balanced by the r a t e of d i s t i l l a t e returned t o
Rinse Tank $2. A s l i g h t excess of d i s t i l l a t e was
produced t o insure t h a t the holding tank would always remain f
u l l ; and the excess was permitted to overflow into the plating b
a t h (0.02 gpm). The steam ra te was cut back to minimize overflow
from the holding t a n k .
- I f the evaporator were fed only RO concentrate,
A f l o a t valve operating off the level
3
r
I
-
JAPORATOR INC E NTRAT E
L (1.0 gpm)
PLATING BATH
I
H O L D I N G I t i DISTILLATE
CONCENTRATE SYSTEM
CONV.
RO
F i g u r e 1 . O v e r a l l S c h e m a t i c of RC/Evapora to
r O p e r a t i o n
I . 4
-
A flow schematic of the RO system i s shown in Figure 2. Feed
from the f i r s t ri:nse tank was wi thdrawn by a booster pump and
passed through two s t r ing- would cartridge f i l t e r s i n pa
ra l l e l . Both 1-rr and 20-p f i l t e r s were used a t d i f
ferent times d u r i n g the f i e l d t e s t . returned t o the r
inse tank. After p re - f i l t r a t ion , the feed was
pressurized t o 700 psi w i t h a multi-stage centrifugal feed pump
and passed through three 4-inch diamter, spiral-wound, PA-300
modules arranged i n se r ies . Most of the concen- t ra te from
the t h i r d module was recycled t o the suction of the feed pump
i n order t o maintain the required feed flow r a t e through the
modules. A heat exchanger i n the recirculat ion loop removed heat
generated by the energy i n p u t of the pumps. A small flow o f
concentrate from the th i rd module was fed t o the evaporator (see
Figure l ) , and the permeate from the three modules was combined
and returned to the f i r s t rinse tank. The instrumentation and
con- t rols f o r the RO system are shown in Figure 2.
Excess flow from the booster pump was
In order to characterize membrane performance w i t h a standard
feed solu- tion, the RO system was periodically operated i n a
total recycle mode using the auxiliary feed t a n k . l ine was
closed o f f , the concentrate l i ne to the evaporator was opened,
and *the permeate was returned to the auxi l iary t a n k rather t
h a n the rinse t a n k . The standard solution (generally a
portion of plating b a t h diluted to 10% by volume of original b a
t h s t rength) was charged to the auxi l iary t a n k and the
system was operated w i t h to ta l recycle unt i l steady s t a t
e was achieved. A t steady s t a t e , the permeate flow ra te f o
r each module was measured, a n d samples of the feed and permeate
from each module were obtained for analysis.
For this mode o f operation, the booster pump recycle
Typical operating conditions f o r both closed-loop a n d total
recycle were:
Feed Pressure 700 psi
Concentrate Flow Rate
Recirculation Flow Rate 10 QPm Temperature 70-90 O F
0.2 gpm (closed-loop only)
The flow schematic f o r the evaporator i s shown in Figure 3.
Steam bras fed through a pressure reducing valve t o a tube bundle
submerged in the boiler sec- t i o n of the evaporator, and steam
condensate was returned to the plant b o i l e r . For most i n s t
a l l a t ions , a cooling tower i s used to cool the water which i
s recirculated through the condenser section of the evaporator.
this i n s t a l l a t ion i t was more convenient t o use
recirculated chil led water since i t was readily available a t the
i n s t a l l a t i o n s i t e and the c h i l l e r had suf f ic
ien t excess capacity. water t h r o u g h an eductor. Cooling
water was added to the eductor t a n k t o remove the energy i n p
u t of the eductor c i rculat ion pump. controlled by a level
switch (LS) and solenoid valve. the solenoid valve opened and feed
was drawn by vacuum into the evaporator. The d i s t a l l a t e
from evaporation of the feed condensed, was collected i n a tray
below the condenser, and was continuously pumped back t o the
second rinse a f t e r Plating (see Figure 7 ) . was continuously
pumped back t o the plating bath.
H o w e m fur
The evaporator was maintained under vacuum by ci rculat i nq
Feed to the evaporator was Upon low level s ignal ,
The concentratelfrom the boi le r section of the evaporator
5
-
4- L 0 c, a L 0 II a > w 0 -c,
Q, c-, a L -c, c aJ V S 0 V
I-*
o m 0 3 0
l r
Y S ca 1-
X 3 <
8
- 0 t- L 0 ce
o m C L )
m L Q, c, r--. ‘7
L L
2 m a
5 u
S 0
*? u rg L c, m C
2 a C
0 C
L 0 4-
aJ f - 0,
L .r-
-
1 7
I - ; 3.aJ o >
0 0 0
E
m L
'r
n
0 c.,
I
I
L
@-
e--
I I t J
aJ i-,
7
-
@ Typical opera t i ng condi t ions fo r the evaporator
were:
Vacuum 26-27 i n . Hg Steam Pressure 4 psi Temperature
100-110°F
Concentrate Flow Rate 1 gpm
During the f i e l d t e s t , the RO modules were cleaned us
ing a cleaning
A 0.1% by sequence recommended by the membrane manufacturer. The
modules were f i r s t * ,~ flushed w i t h 50 gal of water to
remove the plating chemicals. volume solution of Tr i ton X-100, a
non-ionic sur fac tan t , was prepared and recirculated through the
modules a t a pressure of 700 p s i , a flow ra te of 10 gpm, and a
temperature of 120°F f o r 45 minutes. After flushing w i t h
another 50 gallons of water, a 2% c i t r i c acid solution was
prepared a n d adjusted to pH 3.0 w i t h ammonium hydroxide.
through the modules a t the same condions and for the same time as
the Triton X-100. Fol lowing the c i t r i c acid cleaning the
system was again flushed w i water and returned t o treatment o f
zinc cyanide rinsewater, Since the PA- 300 membrane i s rapidly
degraded by chlorine, a l l water used for flushing and preparing
cleaning solutions was dechlorinated by the addi t i on of sodium s
u l f i t e .
- %
.:*
T h i s solut ion was recirculated
. Samples collected d u r i n g the f i e l d t e s t were
analyzed for zinc (atomic absorption), f r ee cyanide ( se lec t
ive ion e lec t rode) , total solids (gravimetric) determination of
residue) , conducti vi ty (conducti v i t y br idge) , and pH
(electrob(
The nominal composition of the plating bath was:
Zn (as metal) 20,000 mg/l 2.7 oz/gal CN (as NaCN) 60,000 mg/l
8.0 oz/gal Caus t i c 75,000 mg/l 10.0 otlgal Srightener (700
Special 4 m l / l 4 ga1/1000 gal
In addition to these compounds, polysulfide was reguJarly added
to the b a t h fm pur i f ica t ion , and the bath a l so contained
a large quantity of carbonates. The to ta l sol ids concentration
of the bath was i n the v i c in i ty o f 350,000 mg/l (35% by
weight).
- - RESULTS AND DISCUSSION
Field a c t i v i t i e s covered the months of June through
December of 1978. Dur ing t h i s time, the system was operated
primarily i n the closed-loop mode shown i n Figure 1. However,
closed-loop operation was periodically interrupt& t o conduct
to ta l recycle t e s t s on the RO system. The performance of RO
modules i s generally assessed i n terms of the module flux ( r a t
e of permeate p r o d d o n Per u n i t membrane area) given i n
gallons per f t2 per day (g fd ) and the rejectior (defined for a
par t icu lar dissolved species as one minus the r a t i o of
pemeate t o feed concentrations) given i n percent. Since both f
lux and rejection depend
.( I
8 s
-
on feed concentration, i t i s necessary to conduct RO
performance t e s t s a t a fixed feed concentration i n order to
be able t o accurately in te rpre t trends in the flux and reject
ion data. tration t o the RO system can vary considerably
(depending on the par ts being plated and their dragout); hence to
ta l recycle t e s t s a t a fixed feed concen- tration (10% of
bath strength) were periodically conducted t o accurately assess
trends in the RO system performance. rigorously during operation i
n the closed-loop mode.
Dur ing closed-loop operation the feed concen-
System performance was monitored less
The total operating time for the field demonstration (time
during which permeate was being produced) was approximately 1000
hours. The to ta l exposure time of the modules b the rinsewater,
however,was about 4,200 hours. longer exposure time re f l ec t s
the system down-time dur ing which the modules were s i t t i n g i
n contact w i t h the concentrated zinc cyanide rinsewater, Var-
ious fac tors contributed t o the system dwn-time, including:
electro-mechanical fa i lure of various system components, high or
low alarm shut-down of the system, various problems w i t h the
plating l ine (unrelated to the recovery system), and weekends and
holidays. t h a n operating time i n controll ing the degree to
which the modules are attached chemically by consti tuents i n the
rinsewater ( e -g . , caus t i c ) . hand, operating time i s more
s igni f icant in controll ing the degree t o which the modules
become fouled w i t h par t icu la tes i n the feed. below a re
based on exposure time. evenly spread over the t e s t program, and
correlat ions on an operating-time basis muld be s imilar .
The
The to ta l exposure time i s probably mare s igni f icant
On the other
Correlations presented However, the operating time was
reasonably
Closed-Loop Operation
The RO/evaporator system was designed t o operate continuously,
\ : i t h no operator a t tent ion, between s ta r t -up Monday
morning and shut-down Fr iday a f t e r - . noon. However, d u r i
n g most of the f i e l d t e s t program, various
electro-mechanical and other problems occurred which prevented
unattended week-long operation. Each f a i lu re generally resulted
in several days down time because of the logis- t i cs of gett ing
project personnel t o the f i e l d s i t e , diagnosing the
problem, and implementing remedial action. Eventually these
problems were solved and week- long unattended operation was
demonstrated. - -
During demonstration of the closed-loop system, no adverse e f f
ec t s were noted on the qual i ty of the plated par t s . p l a t
i n g were followed by an acid d i p , a flowing r inse , and other
surface f inish- ing s teps before the parts were f i n a l l y
inspected. rinsing following zinc cyanide plating was probably not
of c r i t i c a l importance t o quali ty control.
However, the r inse tanks a f t e r zinc
Therefore th,e degree of
Both the RO system and evaporator were under-designed as a d i r
ec t replace- ment f o r the two-stage countercurrent r inse a t a
nominal rinsewater flow ra te Of 2 gpm (see "Economics" below).
Posed on program costs , a v a i l a b i l i t y of FA-300 modules
a t the time the system was fabricated, and lack of design data fo
r the zinc cyanide application. Never- theless, the system was of
su f f i c i en t size t o obtain meaningful design and economic
data.
The system design reflected l imitat ions im-
9
-
RO performance d u r i n g closed-loop operation was monitored
by measuring the productivity of each module ( r a t e of permeate
production) and by monitori the conductivity of the combined RO
permeate. I n general, the productivity d u r i n g closed-loop
operation was s imilar ( b u t s l i g h t l y higher) than the
pro- duct ivi ty d u r i n g the to t a l recycle t e s t s (see
discussion below). of the combined permeate generally ranged
between 2,000 and 4,500 u-mhos/cm. T h i s is equivalent to a total
sol ids concentration of approximately 1,000 to 2,000 mg/l .
The evaporator was operated a t about one-half of i t s rated
capacity (15 gal per hour vs. a capacity o f 25 g p h ) . Samples
of d i s t i l l a t e and evaporator concentrate were obtained and
analyzed f o r zinc, f r e e cyanide, total solids, conductivity
and pH. The results o f these analyses are presented i n Table 1.
During closed-loop operation a t the time the samples were taken, a
stream o f about 1.5 gpn from the bath was circulated through the
evaporator and back to the bath. in order t o prevent precipi ta t
ion o f pla t ing s a l t s i n the evaporator, T h u s , the
evaporator was operating on a feed very s imilar i n composition t
o the bath and producing a concentrate which was more concentrated
than the plating bath. (In addition, the plating bath during this t
e s t appears to be signifi- cant ly higher in zinc than the
nominal level given above). T h e quali ty of dis- t i l l a t e
produced was qui te good. below t h e i r respective detection 1
imi t s , b u t cyanide, conductivity, and hydro- xide i o n were
detectable. The reject ions or removal e f f ic ienc ies were q u i
t e high; par t icular ly f o r zinc and to ta l so l ids . The qua
l i ty of d i s t i l l a t e was considered qui te su i tab le for
f ina l r ins ina .
“9
The COnduct ivf
Zinc and to t a l so l ids concentrations were
Total Recycle Tests a t 10% of B a t h Concentration
Total recycle t e s t s were periodically conducted i n order t
o evaluate membrane module performance under well defined
conditions o f feed concen- t ra t ion (10% of b a t h
concentration),pressure (700 p s i ) , temperature (75- 85OF), and
recirculat ion ra te t h r o u g h the modules (10 gpm) . as a
function of time are shown i n Figure 4. The flux was calculated
from the measured permeate flow ra t e using a surface area of 70 f
t 2 per module. (Actual surface areas measured a f t e r the f i e
l d t e s t f o r Modules iy2 a n d fi’3 were 73 and 69 f t 2 ,
respectively.) The data were corrected t o 75°F using the inverse
re la t ion between f lux and water v i scos i ty .
Results for flux
For Modules n”1 and #2 the f lux levels were nearly ident i ta l
throughout the f i e l d demonstration. over the f i r s t 3,000
hours. Following cleaning a t 3,000 hours, the flux increased to 12
g f d , which i s very close t o the or iginal level . cleaning
procedure employed was qui te successful in res tor ing the flux. I
t may therefore be concluded t h a t the major reason fo r flux
decline i s fouling of the membranes.
The flux gradually declined from 13 gfd t o 7 gfd
Thus the
.
10
-
TABLE 1. REJECTION R E S U L T S FOR EVAPORATOR DURING
CLOSED-LOOP OPERATION r. --
CONCENTRATE D I ST I LLATE R E J E C T I O N
Zinc 46,000 mg/l 99.9998%
Cyanide - 90,000 mg/l 35 mg/l 99.96%
Total Solids 409,000 mg/l 99.9992
Conductivity 160,000 mg/l 300 mg/l 99.81%
P H 13.4 10 .3
NOTE: Operating conditions 104"F, 27 in Hg vacuum, 1 7 %
conversion.
11
-
I I I
4
0- - -0
/ /
/
I 1 I I I 1 I 1
O I
1 I I
a 0
/
I
0
v 0 Ln
d e
0 0 0
d n
0 0 Ln
m 0.
0 0 0
m L
0 0 m c\f
L
0 0 0
iu c
0 0 L9
1
c
0 0 0- - .-
0 0 Ln
0
I
I
12
-
Results for Module $3 a re also presented i n Figure 4 (dashed l
i n e ) . The f l u x s t a r t s a t a s ignif icant ly higher
level (20 qfd) b u t declines more rapidly t o a value of 5 gfd a t
3,000 hours. effective i n restoring the flux of Module #3.
3,000 hours.
The cleaning procedure was i n - Because of i t s low f l u x
and
. ION rejections (discussed below) Module #3 was removed from
the system a t
After cleaning, the system was returned t o closed-loop
operation on the a c t u a l rinsewater. During the next 1,200
hours the flux fo r Modules #l and $2 declined to about 7 gfd. The
r a t e of f lux decline was more r a p i d d u r i n g this time
period since a coarser grade (20~ vs 111) car t r idge f i l t e r
was used t o pre- f i l t e r the feed d u r i n g t h i s segment
o f the demonstration.
Results for zinc reject ion d u r i n g the f i e l d
demonstration a re shown i n Figure 5. and are correlated w i t h a
s ingle curve (sol id l i n e ) . declined gradually from an i n i
t i a l value of 99% to 97% a f t e r 3,000 hours. Upon cleaning,
the rejection f o r Modu’le $1 increased t o nearly 99% while the
rejection f o r Module #2 increased to a b o u t 98%. T h u s the
cleaning procedure was successful in restoring the zinc rejection.
Therefore i t i s reasonable to conclude t h a t the l o s s o f
zinc rejection i s primarily the resu l t of membrane fouling.
occurred.
For Modules #l and # Z the zinc rejections agree reasonably well
The zinc reject ion
After cleaning the rejection again declined as fouling
For Module $3 the zinc rejection dropped o f f rapidly t o a
value of 90:; a f t e r only 700 hours and remained reasonably
constant between 700 and 3,000 hours. Cleaning produced a s ign i f
icant increase i n rejection (from 90 t o 952).
The rejection of f ree cyanide i s shown i n Figure 6 . A g a i
n the r e r u l t s for Modules $1 and +Z agree qui te well
throughout the t e s t . For these modules the cyanide rejection
declined s l igh t ly -- from 98% t o 97% -- during the f i r s t
3,000 hours. However, because of the small loss in re ject ion,
cleaning would be ant ic ipated to have only a minor e f f ec t on
rejection levels . cyanide rejection declined a t a more rapid ra
te , probably as the r e s u l t L the increased rate of
fouling.
Cleaning had l i t t l e e f fec t on the rejection leve l .
Following cleaning, the -
6
The cyanide reject ion r e su l t s for Module $ 3 are a l s o
shown i n Figure 6 . As for zinc, the cyanide rejection dropped
quite rapidly, from 99 to 852, during the f i r s t 700 hours a n d
appeared t o s t ab i l i ze a t a b o u t 25% unti l 3,000 hours.
The d a t a point a t 3,000 hours (93% reject ion) indicates a ’
substantial recovery i n re ject ion before cleaning. was observed
fo r zinc and total sol ids re ject ions, i t i s possible t h a t
the high cyanide rejections a t 3,000 hours could be at t r ibuted
t o a sampling Or analytical error.
Since no similar recovery
13
-
I I Ln w I-
I
m 3= /
I W n
I / n
U
/ --I 3 E / / /
I n 0 N
Lo c
- c u m * * a w w w
0 ?
L n 0
14
co d ..
0 c 0
d n
0 0 m m c
0 0 0
m -
0 0 Ln
N c
0 0 0
N
0 C tn .. c
-0 - c, c u 3 c 5 & c Ce
ro m ro
X 2 - U
0 0 Ln
Lo
w L 53 c,
.r- L
c
-
k I I I I
a: !I- -D I \
0
r - " * * * w w w
o a o
\ \ \ \ 0 I
/ /
/ I I .( I ,'-
u? or 0 01 L n co 0 c3
0 m .p'
.-.
0 0 0
9' n
0 0 Lo
m n
0 0 0
m n
0 0 m
cu n
0 0 0
0.J
0 0 LD
z
7
c 0 0
c
0 C Ln
0
W -e, z c_ G t- L
I3 CT C G X CJ
ln > c 0
+J V w '3 "J L
w -0
c n3 h V
aJ a, L LL
*-
--
! '
I
I I
-
Total sol ids rejections a re presented i n Figure 7. The r e su
l t s for Modules #l and f 2 are , as before, almost ident ical .
There was a gradual loss i n to ta l sol ids re ject ion f o r
Modules #1 and #2 from 95% i n i t i a l l y to 90% a f t e r 3,000
hours. Cleaning a t 3,000 hours produced an apparent loss in
rejection. (A similar loss was a l so noted fo r conductivity re-
jection. ) 90% -- the same level as before cleaning. T h u s the
loss of re ject ion upon cleaning was only temporary. indicate
fouling and a decline i n membrane performance between 3,000 and
4,200 hours, i t i s reasonable to postulate t ha t a similar
decline i n total sol ids rejection occurred d u r i n g t h i s
period. By virtue o f the f a c t t h a t the rejection a t 4,200
hours i s the same as a t 3,000 hours, i t can be concluded t h a t
the cleaning actual ly improved the total sol ids rejection. The
observed loss i n re ject ion i s probably the r e su l t of an
interaction between the c i t r i c acid cleaning solution and the
membrane surface. Similar resu l t s have been observed w i t h c i
t r i c acid i n cleaning t e s t s conducted by the manufacturer (
7 ) . These t e s t s consistently showed a loss in rejection a f t
e r cleaning, b u t the re ject ion t h e n increased over a rela-
t ively short period (5-24 hours) t o the level expected f o r a
clean module. The mechanisms of t h i s interact ion i s not well
understood, b u t the rejection loss appears t o t a l ly
reversible by extensive flushing or by returning the system to
operation on the normal feed. Thus the most reasonable expla-
nation of the to ta l sol ids re ject ion behavior i s t h a t ,
when the system was returned to normal operation a f t e r
cleaning, the rejection increased t o a level probably close t o
95% w i t h i n 24 hours and then decreased gradually t o 90% a t
4,200 hours as f o u l i n g of the module occurred.
Dur ing operation a f t e r cleaning, the rejection increased
to
Indeed, since the r e su l t s of Figures 4-6
The t o t a l solids re ject ions for Module $3 a re quite
similar t o the z i n c and cyanide reject ions for t h i s module.
The rejection decreased from 952 t o about 80% during the f i r s t
700 hours and remained a t about 80% u n t i l the module was
removed a t 3,000 hours. change i n to ta l sol ids re ject ion f o
r Module 2 3 .
Cleaning produced no s ignif icant
The close agreement between the r e su l t s for Modules R"1 and
32 would be anticipated for two identical modules operated i n se r
ies . son the resu l t s for Module # 3 are qui te poor. per pass
was low (high recirculat ion flow re l a t ive t o permeate f low)
the feed concentrations t o the three modules were approximately
the same; therefore a higher feed concentration would not account
for the..&oor- performance of Module #3.
By compari- Since the system conversion
A similar r a p i d loss i n performance has been occasionally
noted for This some of the e a r l i e r PA-300 modules used i n
water desalination ( 7 ) .
problem was traced to a procedure used i n manufacturing the
modules. procedure has since been changed and the problem thereby
eliminated. ever, the modules used i n t h i s f i e l d t e s t
were manufactured before this change, and i t i s believed tha t
Module 83 was defective from the outset of the test.
The How-
% t
16
-
/ I 1 1 t
L
P
cn I 1 I 1
D
1 I
d 1 I I 1 I I I I 1
U
3 I I
0 c Ln
d c
0 0 0
d c
0 0 m cr) ..
0 0 0
0 c
0 0 Ln
c\1 ..
0 0 0
N c
0 0 Ln c
r--
0 0 0
n
7
0 0 iR
C
v) v) rc L
c, r c =
v) >
-
. The level o f suspended sol ids in the zinc cyanide b a t h
was high compared A pur i f ie r (PO~ysU~fjd,)
During the f i r s t 3,000 hours o f
t o the levels observed f o r copper cyanide baths (5) . was
regularly added t o the zinc cyanide bath and produced a mud-like
sludge tha t was removed by f i l t r a t i o n . the r inse tanks
and thus into the RO system. the tes t , l p car t r idge f i l t e
r s were used and had t o be changed approximately once per week
(assuming 100 operating hours per week). hours, 2Op car t r idge f
i l t e r were used. f i l t e rs was substant ia l ly longer
(estimated service time = one month) b u t the r a t e of fouling
of the modules was greater w i t h the coarser f i l t e r s .
recommended tha t two cartridge f i l t e r s in se r i e s -- a
2Op f i l t e r followed by a lp f i l t e r -- be used fo r pre- f
i l t ra t ion of zinc cyanide rinsewater.
Some of t h i s sludge was carried over i n t o
After about 3,000 The service time fo r the coarser
I t i s
Following the f i e l d t e s t Module #2 was cut open and
unwound for inspection. w i t h the degree of fouling increasing
toward the center of the module. layer could be very eas i ly wiped
from the membrane surface. a t 3,000 hours, a d i s t i nc t odor
of H2S was noted during recycling of the c i t r i c acid solution.
This i s presumed to be the r e s u l t of solubilization of some
of the sulf ide foulant. was the ac t ive ingredient i n the
cleaning procedure and tha t the T i t ron X-1co surfactant could
be eliminated. Cleaning agents other t h a n c i t r i c acid could
work as well o r bet ter and should be tes ted. O f par t icu lar
in te res t would be the use of oxidizing agents t o oxidize the su
l f ide foulant layer. t h e membrane i s succeptible t o r a p i d
attack by some oxidizing agents such as chlorine, i t has been
shown t o be r e s i s t an t t o others such as chlorine dioxide
and chromic acid ( 6 , 7 ) .
The membrane was fouled w i t h a t h i n layer of su l f ide
sludge T h i s
D u r i n g cleaning
T h i s suggests t h a t the ammoniated c i t r i c acid
While
. I n addition to f o u l i n g of the membrane, examination o f
the module i n - t e rna ls revealed some possible deter iorat ion
of the membrane backing material . Samples of the membrane and
backing were returned to the manu- f ac turer for examination, the
finding of which confirmed t h a t the backing had been deter
iorated, probably as a r e su l t of the h i g h concentrations of
hydroxide ion i n the rinsewaters ( 7 ) . However, i t i s n o t c
lear w h a t effect deter iorat ion of the backing would have on
membrane module performance. No gross e f f ec t s on performance
were observed during the f i e l d t e s t . Subst i tute backing
materials could be used, b u t a development program would be
required t o commercialize the PA-300 on a more r e s i s t a n t
backing.
Total Recycle Tests a t Other Feed Concentrations - - Following
i n i t i a l s t a r t u p of the RO system, to ta l recycle t e s
t s were
conducted to determine the dependence of f l u x and re jec t
ion on feed concen- t r a t ion . ranging i n nominal concentration
from zero to 20% of the b a t h concentration.
Tests were conducted a t various d i lu t ions of the plating b
a t h
Flux r e su l t s as a funct feed a r e shown f o r Modules #
two modules a re i n very good
on of the total so l ids concentration of the and 82 i n Figure
8. The r e su l t s for these
agreement. (Results f o r Module 4 3 were not
18
-
I
' !
-
consis tent with those shown in Figure 8 a n d have n o t been
included). expected, the f lux decreases with increasing feed
concentration and approaches a level generally considered
uneconomical a t a feed concentratjon i n the v i c in i ty of
100,000 mg/l. approximately 350,000 mg/l i n t o t a i sol ids .
These d a t a i l l u s t r a t e the problem of u s i n g RO t o
achieve very h i g h concentrations. The d a t a Presented in
Figure 8 can be used i n optimizing the degree of concentration t h
a t should be obtained i n the RO system p r i o r t o
evaporation.
Rejection as a function o f the to ta l solids concentration i n
the feed is shown i n F igu re 9 f o r Modules #l and #2. The zinc,
cyanide, and t o t a l so l id re ject ions remain essent ia l ly
constant a t 99%, 98%, and 95%, respec- t ive ly f o r feed
concentrations below about 40,000 mg/l t o t a l so l ids . A t h
ighe r feed concentrations the reject ions d r o p off as
expected.
As
By comparison the bath concentration i s
%
4 .g
Economics
One o f the main objectives of conducting a f i e ld
demonstration i s t o provide a basis f o r evaluating the
economics of the process. shows the flow schematic for the original
two-stage countercurrent rinsing system. The rinsewater flow r a t
e was measured a t 2 . 0 gpm. The r a t e of dragout from the b a t
h was estimated by t u r n i n g off the rinsewater and measuring
the buildup o f zinc and total sol ids in the f i r s t r inse as a
function of time. T h e calculated d r a g o u t ra te was 0.01 t o
0.015 gpm. However the par ts b e i n g plated during th i s
determination were f l a t a n d free-draining and a somewhat
higher dragout r a t e of 0.02 gpm was considered more typical fo r
purposes of design. Given these f low ra tes a n d ideal mixing in
the r inse tanks, the zinc concentration i s reduced from 20,000
mg/l in the b a t h t o 2.0 mg/l i n the f inal r inse.
Figure l o a
Figure 10b shows the flow schematic, flow rates ( in gpm) , a n
d zinc concentrations ( i n mg/l) for ar! RO/Evaporator system
designed to give the same zinc concentration in the f inal r inse.
The flow ra tes given in Figure 1 0 b a r e f o r 90% water
recovery in the RO system ( r a t i o of system permeate t o system
feed: 5.85/6.5 = 0.90). All of the concentrate from the RO system i
s converted to d i s t i l l a t e i n the evaporator, and an equal
flow (0.65 gpm) of plat ing b a t h i s used t o purge the
concentrated chemicals from the evaporator and prevent precipi ta t
ion. The stream returning t o the b a t h i s about 3% higher in
concentration t h a n the b a t h , and+wecipitatio* in the
evaporator would occur only i f there were some const i tuent i n
the bath ( e - g . , carbonates) a t a concentration very close to
i t s so lubi l i ty 'limit. 0 .65 gpm, respect ively) a re
considerably greater t h a n the corresponding flows i n the system
demonstrated (1-1 .5 and 0 .2 gpm, respect ively) .
The permeate and d i s t i l l a t e flow rates given in Figure
l o b (5.85 a n d
ir-
z C c
t- c
L, C
20
-
100
95
90
* z 2 85 t- o w 3 W C
80
7 5
A
I Zinc f I
C y a n i de O&. OA on
I -0 -8 0-
0 tlODULE $1
A MODULE $2
i-
20,000 40,090 60,00(1 80,000 100,000 0
T O T A L S O L I D S F E E D C O N C E N T R A T I O N ,
mg/l
f igure 9. Dependence o f re ject ion on feed concentration.
v
-
Zn 20,000 mg/l Zn 200
TS 350,000 mg/l TS 350
Zn 20,000 mg/l TS 350,000 mg/1
Z10 gpm
I
Zn 2.0 TS 1,516 6.5 gpm
(4
5.85 gpm
Zn 6.06
RO
5.65 gpm
1
- - 5.0 gpm Zn 617
' r0 .65 gpm I
0.65 gpm Zn 20,511 '
0.65 gpm Zn Z C , O O O l
0.65 gpm
Zn 0 EVAPORATOR
b. C l q s e d - l o o p r i n s i n g sysgem
F i g u r e 1 0 . Flows a n d c o n c e n t r a t i o n s f o r
o p e n and c l o s e d - l o o p r i n s i n a
22
-
One of the most important of design c r i t e r i a i s the flux
f o r the RO system. concentration, and the flow ra t e of
concentrate from the module. system would be designed to operate a
t 800 psi, maximum recommended operating pressure f o r the
modules. The design temperature would be 77°F. temperatures the f l
u x would increase, and the reject ion would remain essentially the
same. w i t h increasing temperature, although the membrane 1 i f e
may be shortened. For the design case shown i n Figure 10b ( i .e .
90% RO system conversion) the to ta l sol ids concentration i n the
feed t o the f i r s t RO module i s calculated t o be 5,500 mg/l
and the concentrate withdrawn from the f ina l module i s 10,600
mg/i . T h u s the average feed-side to ta l sol ids concen- t r a
t i o n is about 8,000 mg/l. From Figure 8 the f lux fo r a new
module a t t h i s feed concentration would be about 20 gfd ( a t
700 psi tion flow r a t e of 10 gpm). of i t s i n i t i a l value
over an operating period of about 700 hours (exposure time: 3,000
hours) . Thus i t would seem reasonable t o design the system for a
minimum flux of 10 gfd. i s approximately 0.5 gpm for each module
(area = 70 f t z ) . recommended r a t i o of concentrate flow t o
permeate flow (10 t o 1 ) the ra te o f recirculation for the RO
system would be 5 gpm as shown in Figure lob. (This recirculation r
a t e i s only half of t ha t used during the f i e l d
demonstration and could r e su l t i n a s l i gh t ly lower flux t
h a n measured; however the higher operating pressure -- 800 vs 700
psi -- would compensate by increasing the f lux . )
The f l u x depends on the operating pressure, temperature, feed
The
A t higher
T h u s the performance of the RO system would improve
and a recircula- From Figure 4 the f l u x declines to about
half
A t t h i s f lux the r a t e of permeate production Based on
the
Rejection i s a l so a n important parameter in the design.
Based on the resul ts of Figure 9 ai: a to ta l so l ids
concentration of 8,000 mg/l, the z i n c and total sol ids
rejections would be 99% and 95%, respectively. From Figures 5 and 7
, fouling of the modules can be expected t o r e su l t in a de-
crease i n zinc rejection from 99 to 98%. T h u s , f o r the
design, the minimum zinc a n d total sol ids rejections were
selected as 98% and 90%, respectively.
The RO system can be designed t o operate a t any desired
conversion
A t some intermediate conversion,
within reason. A t lo^ conversion the capi ta l and operating
costs a r e dominated by the evaporator; a t high conversion, capi
ta l and operating costs are dominated by the RO system. the total
operating cost should pass through a minimum. Material m a n c F
calculations were performed fo r the system of Figure 10b using
various R O system conversions. (Results a re shown only for the
90% conversion case.) Capital and operating costs were then
developed fo r each conversion in order t o determine the
optimum.
I !
Table 2 gives the capi ta l costs fo r the RO un i t , the
evaporator un i t , and the total system, fo r various RO system
conversions. The material balance relationships and the RO module
reject ion were used to calculate the permeate flow and the
evaporator qapacity required to give a zinc concentration of 2 mg/l
in the f ina l r inse. A t zero conversion in the RO system ( i .e
. no RO system) the e n t i r e treatment load i s handled by
the
23
-
.. . - . .- - . . _- ... .- r.'
T A B L E 2. CAPITAL COSTS FOR V A R I O U S R O SYSTEM
CONVERSIONS
Reqii red Required Membrane Total Cost Required Total Cost Total
RO Permeate Membrane Required Module Housi ng f o r RO e.
Evaporator f o r System System Flow Area2 ( a ) No. of cos t ( c )
c o s t ( d ) System Capacity ( f ) Evaporator Cost Conversion gpm
f t Modules ( b ) $
0 120 44,129 44,129
0.70 2.575 371 6 3,780 1,700 21,780 66.2 39,199 60,979
0.80 3.625 5 2 2 8 5,040 2,550 23,890 54.4 39,199 63,089
0.90 5.85 842 12 7,560 3,400 28,560 39 .O 33,880 62,440
0.95 8.91 1,283 18 11,340 5,100 34,040 28.1 33,880 67,920 *
I
p ( a ) Design f l u x = 10 g f d . 2 ( b ) Based on area o f 70
f t per module,
( c )
( d )
( e )
Based on $630 per module (Abcor, I n c , )
Based on $850 per 3-module housing (The Permutit Co., Inc.).
Based on system cos t o f $?5,000 (Osmonics, Inc.) for s y s t e
m W/O modules, housings, and high-pressure pump. Pump/motor cos t =
$1,300 f o r .c4 gpm permeate; $2,600 f o r .4 gpm permeate (Wanner
Engineering, I n c . ) ,
100 g p h ($39,199), and 50 gph (,$33,880). ( f ) Double e f f e
c t evaporator with d ooling tower package. Based on rated capac i
t ies of 200 g p h ($44,129),
"
-
I
evaporator. provide the same degree of r insing as the or iginal
2-stage countercurrent system. Cost de t a i l s a re noted a t the
foot of Table 2 . cos t occurs when the evaporator i s used to
handle the e n t i r e treatment load. For RO system conversion
between 70 and 90%, the to ta l system cost remains a t about
$62,000.
For this case the evaporator m u s t produce 2.0 gpm i n order
to
The lowest capital
Total annual operating costs were a l so calculated f o r each
RO system
The costs a re based on operation 100 hours per week (Monday
conversion. shown i n Table 3. Table 3. morning through Friday
afternoon) and 50 weeks per year f o r a to ta l of 5,000 operating
hours per year. other RO system conversions.
The breakdown of operating costs f o r a conversion of 90% i s
Specific costs and assumptions a r e noted a t the foot o f
Similar operating costs were developed for
The e l ec t r i ca l costs for the RO system a re strongly
dominated by the requirements for the h i g h pressure pump. The
use of a posit ive dis- placement pump was assumed w i t h a
combined pump/motor efficiency of 75%. I n addition, since a new
(or clean) module would produce twice the design flux, the RO
system could be operated a t half the design pressure ( i . e . ,
400 p s i ) . As fouling occurs the operating pressure would be
increased t o maintain the design f lux , a n d the system would be
cleaned when the operating pressure reached 800 psi . T h u s an
average feed pressure of 600 psi was used in calculating the power
costs o f Table 3.
When operating w i t h the l p car t r idge p r e - f i l t e r
s , i t was necessary t o change the two paral le l cartridges a f
t e r about 100 hours of operation. the 2Op f i l t e r s , the
pressure drop was s t i l l qui te low ( < 3 ps i ) even a f t e
r 300 hours of operation. 1~ car t r idge in se r i e s , i t i s
estimated tha t the f o u r f i l t e r s would l a s t for about a
month (400 hours). for the economic calculations.
W i t h
W i t h two paral le l passes each h a v i n g a 202 a n d a
Monthly replacement was therefore assumed
W i t h l p p re - f i l t r a t ion , cleaning was required a f
t e r about 700 hours of operation. design cases shown i n Table 2,
anywhere from 6 to 18 modules would be used. I t i s reasonable t o
postulate t h a t the extent of fouling would-mry mversely with the
membrane surface area. f o u r months (1600 hours) was assumed.
During this time three modules were in use. However for the
T h u s a cleaning frequency o f once every
Membrane l i f e i s a very important parameter i n the
economics. As mentioned above, chemical a t tack of the membrane o
r other module components would be expected t o depend on to ta l
time of exposure to the rinsewaters rather than operating time per
se. were exposed t o the rinsewaters fo r a t c t a l of 4,200
hours. T h i s i s close t o the 5,000 hours operating time per
year taken as the basis for calcul- ating the operating costs .
modules w i t h d i s t i l l a t e on shut-down, the e f fec t ive
exposure time would a l so be about 5,000 hours.
The two f i e l d demonstration modules
I f the system were designed to purge the
T h u s a membrane l i f e on one year would
25
-
TABLE 3 . TOTAL AN0 N E T OPERATING COSTS FOR 90% RO SYSTEX
CONVERSiON (5,000 OPERATING HOURS PER YEAR)
RO Evaporator Total System System System
Item 5 5 5
2,100 3,388 5,488
Steam E, $3.50/1000 l b s ( b ) 3,760 3,760
Capi ta l Amortization (a)
Direct Operating Costs - E l e c t r i c a l @ $ 0 . 0 4 7 4 / k
~ - h r ( ~ ) 1,008 1,100 2,108
Operation & Maintenance @ $lO/hr
Cartr idge Filters & Cleaning ( e ) including f r i n g e b
e n e f i t s (d) 960 960 1,920
- Chemicals 240 2 40 Merrbrane Replacement (2-year l i f e )
3,780 3,780
T o t a l Direc t Operating Cost 5,998 5,820 i i ,808
17,296
(2,820)
(361 1
( 6 , 2 4 0 )
-%%> 9*208 Total Annual Operating Cost Credi t f o r 3ath
Recovery @ $0.47/gal lfl
Credi t f o r Cyanide Destruct ioa 13 54.051lb CFi")
3 Credi t f o r Water Recovery P 90.45/100 f t
Credi t for Sol id k a s t e Disposal @ S29/drun ( h ) and 25% s
o l i d s i n s ludge
Total Annual Credi t s (9,075)
Net Annual Operating Cost 57,321
S t r a i g h t - l i n e deprec ia t ion over 10 years with
zero sa lvage value.
Based on ac tua l c o s t f o r No. 4 fuel o i l o f S 0 . 3 ? 3
/ ~ a l , heat ing value of 140,000 8 tu /ga l , and a b o i l e r
e f f i c i e n c y of 80%.
Based on ac tua l c o s t for Cctober 1978.
Based on a c t u a l maintenance labor r a t e s .
Four c a r t r i d g e f i l t e r s changed once per mon th a t
a v e r a g t t t o s t of 54.68 each. Cleaning three times per
year w i t h 6 l b s c i t r i c ac id per c leaning a t 50.82 per
pound.
Chemical costs: zinc $0.445/lb; sodium cyanide SO.Sl/lb; c a u s
t i c $0.185/lb; b r ightener $6.792/ga1.
Based on 8 lbs NaOCl per l b of C N w i t h no addi t ion of c a
u s t i c (a l ready present i n r insewater ) .
Based on s ludge cent r i fuged t o 25% s o l i d s and hauled i
n lots of 80 55-gal drums.
Cost f o r 15% NaOCl s o l u t i o n = 50.665/gal.
26
-
be v i r tua l ly assured. However, i t i s l i ke ly t h a t
the membrane l i f e would be considerably longer than one year. D
u r i n g the f i e ld demonstration the modules produced only a
fraction of the permeate flow t h a t would be produced by a system
designed to meet the rinsing constraints of Figure l o b . Thus the
feed concentration t o the f i e l d demonstration modules was con-
siderably greater t h a n would be seen by a la rger capacity
system. more, d u r i n g periods when t h e demonstration system
was not operating, the concentration across the membranes
equilibrated result ing i n very h i g h concentrations on the
permeate s ide. of the membrane backing material and other
peremate-side components. P u r g i n g the system on shut-dhow
would largely eliminate t h i s source of module de te r iora t ion
. Considering these fac tors , i t i s not unreasonable to project
a two-year membrane l i f e a s assumed i n Table 3.
Further-
This s i tua t ion would accelerate a t tack
Operating and maintenance labor f o r the R O system was assumed
to require 1 day per month. shut-down on Fridays, car t r idge f i
l t e r replacement, membrane cleaning, a n d other maintenance a s
required.
This would include system s ta r t -up on Mondays,
Steam consumption f o r a double e f f e c t evaporator
operating as shown i n Figure 10b with temperatures of 130°F and
110°F i n the f i r s t and second ef fec ts , respectively was
calculated to be 0.66 lbs steam/lb d i s t i l l a t e . Electrical
requirements f o r operation of the cooling tower, evaluation of
the evaporator, pumping the various output streams, a n d
miscellaneous usage Nere obtained from the manufacturer (4.60 k w ,
8.64 kw, a n d 16.64 k w f o r double e f f e c t capaci t ies of
50, 100, and 200 g p h , r espec t ive ly) . As for RO, operation
and maintenance fo r the evaporator was assumed to reauire one
man-day per month.
As shown i n Table 3, the d i r ec t operating costs (excludinc;
amort'zation) are about $6,000 per year each fo r the RO system a n
d the evaporator f c r a total of about S12,OOO. The to ta l a n n
u a l operating c ~ s i including amorti- zation i s $17,300. T a b
l e 3 a l so gives the breakdown of c red i t s resu l t ing from
recovery o f the rinsewater. in the amount of cyanide t o be
destroyed. The c red i t for sludge disposal was calculated on the
basis of the weight of zinc hydroxide produced by precipitation of
the dragout and i s probably s ignif icant ly l e s s t h a n +
plating shops which do not have so l ids dewatering equipment. The
to t a l annual c r e d i t f o r rinsewater recovery i s about S l
O , O O O , which leaves a net operating expense of $7,300 per
yeaI- f o r the recovery system.
The la rges t c r e d i t r e su l t s from a reduction
-
Operating cos ts were a l so determined fGr other 20 system
conversions and a r e shown i n Figure 11. operating cos t s
increase and the evaporator operating cos ts decrease. The total o
p e r a t i n g cos t passes through a minimum a t a n RO systefn
conversion of 90:;. Using a n evaporator alone for rinsepjater
recovery (zero percent RO system conversion) the to t a l annual
operating cost including amortization is about $20,900. total
annual operating cost i s approximately S17,300 which represents a
n annual savings of 53,600.
For conversions ranging from 70'1 to 95"' the RO
For the optimum combination of RO and evaporation, the
27 1
-
22,000
20,000
18,000
16,000
14,000
12,000
10,000
8,990
6,000
L
0
TOTAL
\ \
\
\ \ \ \ \
I I I
I
1 I I i / ------?
/ / *
/
I
I
i i I
--1
i ! I
I I
- ! - ,
! I I
i I j i I
V 0 3.7 9.8 0.9 1 .o
RO SYSTEil CONVERSIOlJ
F i g u r e 11. Annua l o p e r a t i n g cos t s f o r v a r i
o u s RO c G n v e r s i O n s
-
I
In considering the impact Of these numbers i t should be
ernphasjzed t h a t the recovery system was designed to meet a
given rinsing c r i t e r i a f o r a two- stage countercurrent
rinsing system. Recovery system costs could be reduced
Significantly by u s i n g more r i n s i n g stages. rinse t anks
could be inserted i n the l i ne , and for automatic operations,
over-the-tank spray rinses could be used. I n addition,
consideration should be given t o working with higher rinse
concentrations where rinsing i s n o t c r i t i ca l . The costs
for recovery provide a s u b s t a n t i a l incentive for reducing
the rinsewater flow by simple, inexpensive techniques.
For manual operations, a d d j t i o , i a l
Acknowl edgements
The authors grateful ly acknowledge the cooperation and support
o f Bruce %arner, president o f New E n g l a n d P l a t i n g Co.
i n providing the f ie ld s i t e for the demonstration, support
personnel to a s s i s t i n i n - s t a l l a t ion , and cost
information f o r operation and waste treatment a t >4ew England
Plating Co. Financial support f o r the program was provided by E P
A ( G r a n t Pro. R835330) a n d ACS (Research Project No. 45 ) .
Tech- nical support d u r i n g the prooram was received from the E
P A Project Officer, I larv S t i n s o n , a n d from the AES
Project Committee: Jack H;lfier, Jose2h Conoby, Charles Levy, James
Yorse, and George Scott.
29
f
-
? . . . .
I
R E F E R E N C E S
1 . Skovronek, ti. S . , and id. K . Stinson, Advanced Treatment
Apgroaches for ideta1 Finishing Yastewaters ( P a r t I I ) .
Plating and Surface Finishing, 6 4 ( 1 1 ) : 24-31, 1 9 7 7 .
2 . Anon., Recovery Pays: Plating and Surface Finishing, 6€ 48,
1379.
2) :45-
3 . H a l l , E . P . , D. J . Lizdas: and E . E . Auerbach,
Platinr; and Sur- face F i n i s h i n o , 66 ( 2 ) : 49-53,
1973.
4. Donnelly, R . G . , R . L . Goldsmith, !:. J . W ? l u l t y
, a n d LI. T a n , Reverse 3smosis Treatment o f Electroplating
'Aastes, P l S t i n E j . 61 (5): 432-422, 1974.
5. ilCXU'lty, 5 . J . , R . L . Goldsmith, A. Gollan, S .
Hossain, and 0 . Grant, geverse Csciosis Field Test: i reawent o f
Coaaer C'anjde i?inse Waters. EPA-603/2-77-179, U. S . E n v i
ronnental Protection. Agency, Cincinnati, Ohio, 1 9 7 7 . 33
pp.
6. ;Ic:lu?:y, K. J . , P . 2. Hoover, and R . L . Goldsr:ith, E
v a l u a t i c q of Advanced Reverse Osmosis iierr,branes for che
Treatment o f E 7 ec t rr, p l a t: i ncj !,la s t es . tection
Acienc;/, Cincinnati. Ohio.
E P A - 603/ 8- 73-51 13. E n v i r a n Ken ti! 1 Pro -
7 . Personal Connunications, Or. Robert L . Riley, Flsid System
3ivision o f il i lP, San Diego, CA.
30