EPA-600/2-77-039
February 1977
Environmental Protection Technology Series
REVERSE OSMOSIS FIELD TEST:
Treatment of Watts Nickel
Rinse Waters
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45238
-------
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-600/2-77-039
February 1977
REVERSE OSMOSIS FIELD TEST:
TREATMENT OF WATTS NICKEL RINSE HATERS
by
Kenneth J. McNulty
Robert L. Goldsmith
Arye Z. Roll an
Wai den Research nivision of Abcor, Inc.
Wilmington, Massachusetts 01887
for
The American Electroplater's Society, Inc.
Winter Park, Florida 32789
Grant No. R-803753
Project Officer
John Ciancia
Industrial Pollution Control Division
Industrial Environmental Research Laboratory - Cincinnati
Edison, New Jersey 08817
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental
Research Laboratory - Cincinnati, U.S. Environmental Protection Agency,
and approved for publication. Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
-------
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 increasinaly more efficient pollution
control methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (IERL-CI) assists in developing and demonstrating new and
improved methodologies that will meet these needs both efficiently and
economically.
This report is a product of the above efforts. It was undertaken to
demonstrate the effectiveness and economic feasibility of using reverse
osmosis for closed-loop control of metal finishing rinse wastes under actual
plant conditions. The reverse osmosis system concentrates the chemicals
for return to the orocessina bath while purifyina the wastewater for reuse
in the rinsing operation. The results of the report are of value to R&D
programs concerned with the treatment of wastewaters from various metal
finishing, non-ferrous metal, steel, inorganic and other industries. Further
information concerning the subject can be obtained by contacting the
Metals and Inorganic Chemicals Branch of the Industrial Pollution Control
Division.
David ^. Stephen
Director
Industrial Environmental Research Laboratory
Cincinnati
m
-------
ABSTRACT
A field test was conducted to determine the feasibility of
using a polyamide reverse-osmosis membrane in hollow fine fiber con-
figuration for closed-loop treatment of rinse water from a Watts-
type nickel bath. Performance of the membrane module was deter-
mined by measuring the productivity (flow rate of purified water)
and rejection (separation efficiency) as a function of operating
time. Performance was monitored over 1600 hours (67 days) of
operation and 2300 hours (96 days) of exposure to nickel rinse
waters.
Because of an undetected leak in an oil seal in the high-
pressure RO feed pump, oil leaked into the feed and fouled the
module, causing the productivity to decrease to about one-half of
its initial value. After the field test the module was opened, and
the fiber bundle was inspected. There was no significant build-up
of particulates, but the fibers were badly fouled with oil.
Cleaning with a detergent solution completely removed the foulant
and appeared to restore the fibers to their original condition. It
is anticipated that in the absence of the oil leak very little loss
in productivity would have occurred.
In general the rejections were high and quite stable for the
duration of the field test. Nickel rejections were typically 99%;
total solids, 97%; and conductivity, 95%. However, at the pH of the
rinse, boric acid exists mainly in its unionized form which is
poorly rejected by all membranes. The measured boric acid rejec-
tions ranged from about 50% to 90%. Platers should anticipate a
build-up in the rinse of boric acid relative to other bath con-
stituents when using RO for closed-loop treatment.
Since the capital and operating costs are similar for hollow-
IV
-------
fiber polyamide and spiral-wound cellulose acetate modules at the
pH's encountered, the economics of closed-loop RO treatment of Watts-
nickel rinse waters will be largely independent of the type of
module (hollow fiber polyamide or spiral-wound cellulose acetate) used
in the RO system.
Based on the results of numerous commercial installations and on
the results of both laboratory and field tests, the economics of closed
loop recovery of nickel rinse waters by reverse osmosis are generally
attractive. Depending on certain factors specific to each plating line,
the capital investment for a reverse osmosis system can be recovered in
as little as one year or less of operation.
-------
CONTENTS
Page
FOREWORD iii
ABSTRACT iv
FIGURES viii
TABLES ix
ACKNOWLEDGMENT x
I. CONCLUSIONS 1
II. RECOMMENDATIONS 2
III. INTRODUCTION 3
IV. EXPERIMENTAL 5
V. RESULTS 10
VI. DISCUSSION 22
VII. REFERENCES 28
-------
FIGURES
Number Page
1 Flow Schematic for Field Test System 7
2 Corrected Productivity for Nickel Rinse Waters and
Total Solids Feed Concentration vs. Operating time 13
3 Corrected Productivity for Sodium Chloride Tests
vs. Operating Time 16
4 Rejection of Various Rinse Water Species vs.
Operating Time 19
5 Corrected Sodium Chloride Rejection vs. Operating
Time 21
6 Recommended RO System for Nickel Line at New England
Plating Company 24
VI 1 1
-------
TABLES
Number Page
1 Composition of Nickel Bath 6
2 Productivity and Rejection of Nickel Rinse Waters as
a Function of Operating Time and Operating Conditions 11
3 Sodium Chloride Productivity and Rejection as a
Function of Operating Time and Operating Conditions 15
4 Results of Spectrographic Analysis of Cleaning Solutions
After Use 18
5 Breakdown of Operating Costs 26
6 Economics for Closed-Loop Nickel Recovery 27
IX
-------
ACKNOWLEDGMENT
The authors pratefully acknowledae the help and cooperation of
Mr. Bruce Warner, President, New Enpland Platino Co., Worcester,
Massachusetts, who provided the field test site and support facili-
ties for this propram. Mr. Jerry Wheelock was responsible for the
operation of the field-test system.
Direction was received throuphout the nroqram from the EPA
Project Officer, Mr. John Ciancia, and from the AES Project Committee:
Messers. Charles Levy (District Supervisor), Lawrence Greenbera (Com-
mittee Chairman), Arthur A. Brunei!, Joseph Conoby, and Dr. Robert
Mattair.
-------
SECTION I
CONCLUSIONS
1. Because of an undetected oil leak, the RO module bacame badly
fouled with oil, resulting in a decline in module productivity to
about one-half of the initial value. Based on examination of the
fiber bundle following the field test, very little decline in produc-
tivity would be anticipated in the absence of the oil leak.
2. In spite of severe module fouling, the rejection by the mem-
brane of various dissolved species in the rinsewater was generally
high and quite stable for the duration of the field test.
3. At the pH of the rinsewater boric acid exists mainly in the
non-ionized form which is poorly rejected by all membranes. Boric
acid will build up in the rinse system relative to other plating
bath constituents when using RO for closed-loop treatment.
4. The economics of closed-loop RO treatment of Watts-nickel rinse-
waters will be largely independent of whether hollow-fiber polyamide
modules or spiral-wound cellulose acetate modules are used.
5. Based on the present field test, a previous field test — and
numerous commercial installations (e.g., Ref. 5), closed-loop RO recovery
of nickel rinsewaters can be highly attractive resulting in payback
periods as low as one year.
-------
SECTION II
RECOMMENDATIONS
1. On the basis of this field test, hollow-fiber polyamide mem-
brane modules can be recommended as a viable alternative to spiral-
wound cellulose acetate membrane modules for the treatment of Watts-
type nickel rinse waters. Since the process economics are similar
for these two modules, one of them cannot, on the basis of present
information, be recommended over the other as being more cost
effective.
2. Since the Watts-nickel application seems to be successful for
both the cellulose acetate and polyamide membranes, future research
and development should be devoted to high-volume baths which are more
difficult to treat, such as zinc cyanide and chromic acid.
-------
SECTION III
INTRODUCTION
It is apparent that many platers will have to severely reduce
their output of pollutants in order to meet Federal effluent limita-
tions guidelines. In determining the most cost-effective means of
achieving compliance with discharge standards, platers should have
available sufficient data to assess the relative merits of alternative
treatment technologies. The overall objective of this research
project is to provide information on the applicability of reverse
osmosis (RO) to the treatment of metal finishing wastes.
To date, emphasis has been placed on closed-loop treatment of
(1 2)
rinsewaters from specific plating baths. Pilot tests —'—' have
indicated that RO shows promise from the treatment of a number of
different plating baths. Field tests are now being conducted to
assess RO performance under realistic conditions. To date, two
separate field tests have been conducted on copper cyanide rinse
waters ^'^.
Nickel baths are generally regarded as being among the more
favorable baths for RO treatment. The Watts-type nickel bath is
considered a "general purpose" bath for nickel plating and is the most
widely applied of all the nickel bath formulations. The primary
constituents are nickel sulfate, nickel chloride, and boric acid.
The bath pH is generally within the range fo 3-6 and the resulting
rinsewaters are within a pH range that can be tolerated by both of
the widely applied, commercially available membranes: cellulose
acetate (pH 2.5-7) and aromatic polyamide (pH 4-11). The temperature
of the Watts-nickel bath is generally within the range of 120-160°F
resulting in significant bath evaporation. This is very important in
achieving closed-loop recovery with RO. For plating lines where the
ratio of bath evaporation to dragout is high, the rinsewater concentrate
does not have to be highly concentrated in order to return it to the
plating bath. However, if the ratio of bath evaporation to dragout is
-------
low, the rinsewater concentrate must be concentrated to a degree beyond
the capabilities of RO (because of osmotic pressure limitations), and
an auxiliary concentration technique (e.g., evaporation) must be used
to supplement RO. For the general operating temperature of Watts-
nickel baths, closed-loop recovery can be achieved with RO alone unless
the dragout is exceedingly high.
Although RO systems have been sold for closed-loop treatment of
nickel rinsewaters, '-'-' details of the field results have not generally
been made available. However, results have been reported — for a field
test on Watts-type nickel rinsewaters using spiral-wound cellulose-
acetate modules. The present field test investigates the applicability
of the polyamide membrane in hollow fiber configuration to the treatment
of Watts-type nickel rinsewaters.
Membrane module performance was evaluated by measuring the
productivity and rejection as a function of operating time. Productivity
is defined as the rate at which permeate (purified water) is produced
by a module of specified size when operated under spedified conditions.
The rejection is a measure of the degree to which dissolved species are
prevented from passing through the membrane. Rejection is defined by the
equation:
CF _ cp
Rejection - — x 100%
CF
where:
Cp = Concentration in feed to module.
Cp = Concentration in permeate from module.
Although this report describes the application of one specific
membrane to one specific Watts-nickel plating bath, the results should be
readily transferable to other Watts nickel baths and, with less certainty,
to other baths of similar pH, temperature, and dissolved solids concentra-
tions. The RO process can, in principle, be used to concentrate and recover
a broad range of rinsewaters, but the economic feasibility depends on
factors specific to the membrane, rinsewater, and plating operations.
-------
SECTION IV
EXPERIMENTAL
PLATING LINE
Field tests were conducted on a Watts-type nickel bath at New
England Plating Co., Worcester, Massachusetts. The nickel bath was
part of an automatic rack nickel-chrome line used to plate a variety
of small steel parts. The nickel tank was approximately 9,500 liters
(2,500 gallons) in size and was followed by a three stage counter-
current rinse. The evaporation rate from the bath was estimated at
22.7 liters per hour (6 gph). A typical dragout rate of 2.3 liters
per hour (0.6 gph) was assumed for design purposes, but the actual
dragout rate varied substantially depending on the size and shape of
the parts. The nominal composition of the bath is given in Table 1.
RO SYSTEM
A flow schematic of the RO system is shown in Figure 1. Feed
was pumped from the rinse tank by a booster pump and passed through a
one-micron cartridge filter. The pressure of the filtered feed was
increased to the desired operating pressure by a high-pressure,
positive-displacement pump (Cat Pumps Corp., Model 01001). Pressure pul-
sations were dampened by an accumulator (ACC) on the pump discharge. The feed
was separated into a concentrate stream and a permeate stream by a duPont B-9
Permasep®permeator (model 0440-042). The permeate stream from the
RO module was returned directly to the rinse tank. The concentrate
stream passed through a back-pressure regulator (BPR) which controlled
the operating pressure in the module. Most of the concentrate stream
was recycled to the suction of the high-pressure pump to maintain a
sufficiently high flow through the module. A float valve operating
off the bath level returned concentrate to the bath as needed to com-
pensate for evaporation. Deionized water was added to the rinse tank
-------
Table 1. COMPOSITION OF NICKEL BATH
Constituent
Concentration
Total Nickel
Nickel Sulfate -6 h
Nickel Chloride -6
Boric Acid
Brightener
PH
Temperature
Purification
82 g/1 (11 oz/gal)
255 g/1 (34 oz/gal)
105 g/1 (14 oz/gal)
45 g/1 (5 oz/gal)
not analyzed
4.3
60°C (140°F)
Continuous filtration
-------
Deionized
Make-up
Water
Booster
Pump
High-Pressure
Pump
SV
Figure 1. Flow schematic for field test system
-------
to compensate for concentrate returned to the bath.
The feed flow rate to the module was set (NV) at about 15
liters per minute (4 gpm) for the duration of the field test. The
permeate and total concentrate flow rates were measured (F), and
samples of the feed, concentrate, and permeate were obtained (SV) for
analysis.
Pressures were measured (P) before and after the filter to
determine when the cartridge should be replaced and before and after
the RO module to determine the operating pressure and pressure drop.
The system was protected against overpressurization by a pressure
relief valve (PRV) and the pump was protected against running dry by
a low pressure switch (LPS). The temperature (T) of the feed was also
measured.
In addition to operating the unit in the normal mode shown in
Figure 1, periodic tests were conducted with a standard 1500 ppm
NaCl solution at fixed conditions. For these tests, the plating
solution was flushed from the RO system and NaCl feed solution was
withdrawn from an auxiliary feed tank. The permeate and concentrate
streams were returned to the feed tank. When steady state was
reached, feed and permeate samples were analyzed for conductivity and
the productivity was measured.
OPERATING CONDITIONS
The RO module was operated at its maximum recommended pressure
of 28.6 bars absolute (400 psig) since both the flux and rejection
increase as the operating pressure increases. The module conversion,
defined as the ratio of permeate flow to feed flow, varied from 20%
to 76%. Experimentally the feed flow was set at about 15 liters per
minute (4 gpm) and the conversion was allowed to vary with the module
productivity. The feed concentration to the module varied depending
-------
on the amount of dragout from the bath to the rinse. The temperature
varied between 8°C and 33°C.
For the standard NaCl tests, conditions were controlled at 28.6
bars absolute (400 psig), 75% conversion, 1500 ppm feed concentration,
and 25°C.
ASSAYS
Samples were analyzed for nickel (atomic absorption), boric
acid (NaOH titration of mannitol complex), total solids (evaporation-
gravimetric), pH (meter), and conductivity (meter).
-------
SECTION V
RESULTS
MODULE PRODUCTIVITY
The productivity of a module is important in determining the
capital cost of an RO system. The higher the productivity, the
fewer the number of modules required, and the lower the capital cost.
Productivity depends on the operating pressure, the feed concen-
tration to the module, the conversion at which the module is operated,
and the temperature. Corrections for variations in operating pressure
were negligible since the module was always operated at or near 28.6
bars (400 psig). The feed concentration and conversion varied
substantially, but accurate correction factors for nickel rinse
waters are not available. Therefore, the data were not corrected to
a specific feed concentration and conversion. All productivity data
were corrected to a temperature of 25°C using the correction factors
given in the duPont Technical Information Manual.
The measured and corrected productivities are given in Table 2
as a function of operating time and operating conditions. The
corrected productivity is plotted as a function of operating time in
Figure 2. The total solids concentration in the feed to the module
is also shown.
The minimum in the productivity curve near 150 hours is the
result of the high solids concentration in the feed over the same
time period. The high feed concentration results in a high osmotic
pressure on the feed side of the membrane and reduces the effective
fo\
driving force for water permeation -'. After about 300 hours, the
feed concentration did not vary excessively, and, as a result, the
change in productivity was much more gradual. If the feed concen-
10
-------
Table 2. PRODUCTIVITY AND REJECTION OF NICKEL RINSE WATERS AS A FUNCTION OF OPERATING TIME AND OPERATING CONDITIONS
Cunulatl ve
Operating
T i rre ,
hrs.
0 5
24
97
15)
2£0
294
323
NOTF A
494
653
NOTE B
797
9C4
•NOTE C
1,114
MOT- 0
NOTE E
1,190
Feed
Pressure,
Ears abs.*
(psv:)
?o r. I1 * °") 1
<: j. u 1 t^ J ;
23.6 (4CC)
23.9 (405)
28.5 (-^0)
27.2 (3CO)
20. 5 (400)
23.6 (400)
29.3 (410)
23.6 (400)
28.6 (400)
28.9 (405)
28.6 (400)
23.2 (395)
Module
' P
Bars abs.*
(cs-;~!
17 f ",r,\
i.l \ *- J 1
1.7 (25)
2.1 (30)
2.4 (35)
2.4 (35)
2.1 (30)
2.4 (35)
2.8 (40)
3.4 (50)
4.1 (60)
4.3 (70)
4.5 (65)
4.8 (70)
Conversion,
%
c •
C "»
75
26
20
28
29
69
33
34
25
22
23
28
Tcrperature ,
=C
(°F!
i •; o / r 7 \
1^.3 \ J/ }
2C.6 (69)
33.3 (92)
22.3 (73)
21.1 (70)
14.4 (53)
12.8 (55)
31.1 (33)
31.1 (33)
13.3 (56)
9.4 (49)
8.3 (47)
1E.1 (61)
Measured
Flux,
1 pi-.
(qp~)
Q f-) 1 •) ' 3 \
y.'tt \ L. . 1 3 )
7.11 (1.B3)
3.97 (1.05)
3.03 (0.80)
4.1G (1.10)
4.35 (1.15)
4.69 (1.24)
6.43 (1.70)
6.21 (1.64)
3.71 (0.98)
3.33 (0.88)
3.48 (0.92)
4.24 (1.12)
Corrected
Flux**
Ipn
(no;-!
U7 A ! t &£\
. / O \ J . D^ 1
8.29 (2.19)
3.03 (0.60)
3.26 (0.36)
4.77 (1.26)
6.24 (1.C5)
7.11 (1.88)
5.22 (1.33)
5.03 (1.33)
5.53 (1.46)
5.64 (1.49)
6.09 (1.61)
5.71 (1.51)
Ntckel
Cone. ,
rcg/1
Feed Pern
34,000 136
12,300 256
16,000 212
2,250 11
3,000 34
2,000 18
2,600 35
4,600 30
2,350 28
2.300 16
1,900 12
1,860 8.7
Boric Add
Cone. ,
iig/1
Feed Pern.
646 434
3,407 1,267
4,860 555
520 148
631 185
478 153
633 267
1.130 547
513 '208
622 160
2.55 56
403 148
* One Ear = 10J fi/n? • 14.50 psi - 0.937 atm
** Fljx corrected to 77°F
NOTES:
A - Module cleaned with 2% citric acid adjusted to pH 4 with fiH.Oil.
B - Module cleaned with 2* citric add adjusted to pH 8 with r;H4CH
C - Module cleaned with NaOH solution at pH 12
0 - Module cleaned with 2: citric add adjusted to pH 8 with NH^OH
L - Module cleaned with 1.5! NajEOTA adjusted to pH 8 with NaOH
-------
Table 2 (continued). PRODUCTIVITY AND REJECTION OF NICKEL RINSE WATERS AS A FUNCTION OF OPERATING TIME AND OPERATING CONDITIONS
ro
Cumulative
Operating
Tire
hr5.
OC
. J
?d
£H
97
i nn
1 JU
pen
COU
TOrt
c J 4
•1OQ
JC. O
NOTE A
194
653
NOTE B
797
934
NOTE C
1114
NOTE 0
NOTE E
1,190
Tot. Solids
Concentration
mg/1
Feed Porn.
i 7 ?an i ?KQ
1 / i L HU 1 , £OO
71 nnn ? QHA
y 1 )UUU £ , V JO
7n 'i-ifi i pr.R
/U i JUU 1 i UCJG
7 7cn 971
/»/OO £ / 0
?n <17? i^n
CU i J / £ JJU
nrnn 979
, JOU L I L.
34, BOO 45Z
14.653 564
12,350 373
9,195 15G
8.598 140
10.192 285
Conductivity
u-hos/c?.
Feed Prrr'. Core .
1 finn ^"^
in rnn -.1 nnn
^^ nm >i noo
?7 ^nn >i n°n
6n o n ? ;" ri
nrnn rnn
7 7r.n 7?£;
7,200 495 12.600
7, GOO 545 12.7CO
7,150 445 11, COO
5,700 240 7,500
5,600 190 7. 400
6,800 340 9,600
Toed Per" Coic.
T r A n TO
J . J ** . U t . O
1C / -j 7 a
j . D ^ . j 0 . "
11 17 1 ^
J.l J . / J . j
1 •) T f
J . j j. u •* --
•> -j e i -11
J . C U . 't J.I
3T 10 1 1
. J J. J J.I
3.4 4,0 3.2
3.5 4.1 3.4
3.1 3.8 3.0
3.1 C.O 2.9
3.0 4.2 3.0
3.5 4.6 3.5
Mckcl
Rejection,
On c
J J • D
nn n
yj . u
QO 7
QQ ^
yy , j
no n
yo. y
QQ 1
yy . i
90.6
99.3
98.3
99.3
99.4
99.5
Boric
Acid
Rejection,
Ti
j j
c-i
0 J
00
UQ
7?
/t
rp
Do
58
52
59
74
78
63
Total
Solids
Rejection
QT C
7 t . 0
qc q
yj.y
Q7 1
y / . j
Qfi ^
yo . a
Qp -1
yo. j
07 K
y / . o
98.7
96.2
97.0
98.3
98.4
97.2
Conductivity
Rejection
%
qc c
V J . D
Qt Q
yj.y
or •»
90 . J
QC o
yo. o
93.1
92.8
93.8
95.8
96.5
95.0
-------
Total Solids Concentration
0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500
Operating Tire, Hrs.
Figure 2. Corrected productivity for nickel rinse waters and total solids feed concentration vs. operating time
-------
tration during the first 300 hours had remained at about 2% solids, it
is likely that the productivity vs. time curve would have followed
the dashed line rather than the solid line over this time period. It
can be concluded that, at a given feed concentration, the productivity
declines with operating time. After 300 hours of operating time,
this module was periodically flushed with cleaning solutions (to be
detailed later) in an effort to restore the original productivity.
While the cleanings were successful in arresting the decline, they
were unsuccessful in restoring the original productivity.
Because the feed concentration is an uncontrolled variable, it
is difficult to get an accurate indication of the dependence of
productivity on operating time. The standard NaCl performance tests
give a more accurate picture. Table 3 gives the productivity during
the NaCl tests as a function of test conditions and cumulative
operating time on nickel rinse waters. The corrected productivity
during the NaCl tests is shown in Figure 3 as a function of operating
time. The productivity declined rapidly over the first 300 hours.
The cleaning solutions shown in Figure 3 were used at the indicated
times in an effort to recover the productivity. Standard NaCl tests
were conducted before and after each cleaning. The cleaning generally
consisted of flushing the module for two hours with about 38 liters
(10 gallons) of the indicated solution. Each cleaning produced only
a slight recovery in productivity.
The solutions used were recommended by duPont for the removal of
common foul ants:
Solution Foul ants Removed
A Hydrated metal oxides
Calcium carbonate
Inorganic colloids
B,D,E,F Calcium sulfate
Calcium phosphate
Inorganic colloids
14
-------
Table 3. SODIUM CHLORIDE PRODUCTIVITY AND REJECTION AS A FUNCTION OF OPERATING TIME AND OPERATING CONDITIONS
Cufr-'jl all vc
Operating
Ti ;ne ,
hrs.
0
21
32
1 01
i°3
292
329
Feed
Pressure ,
Bars abs.*
(psi'j)
2C.9 (373)
23.2 (3^5)
23.2 (353)
27.9 (3-JC)
20.2 (j95)
2B.6 CM)
28.2 (395)
Module
A?.
Bars abs.*
(psig)
2.1 (;o;
1.7 (25)
IT / - -
i . 1 \ L. J
1.7 (25)
1.7 (2-5)
2.1 (30)
1.7 (25)
Ccn Yt'r j 1 on ,
f.
77
7 ~
77
70
73
74
73
Modulo cleaned'with 21 citric add adjusted to pii 4 with MH,;OH
329
495
G54
28.6 (MO)
28.2 (3,-6)
23.2 (335)
1.7 (25)
2.1 (30)
2.4 (35)
74
74
74
Module cleaned with i% citric add adjusted to pli £ with ,';H,.CH
661
799
935
23.2 (395)
28.9 (435)
28.6 (4031
2.4 (35)
3.0 (43)
3.4 (50)
Module' cleaned with fiaOH soln. at pH 12
9c3
1115
23.6 ('ICO)
28.6 (400)
3.1 (45)
3.8 (55)
75
75
73
75
75
K.dule cleaned with 2% citric add adjusted to pii 3 with .':ii.-,OH
1118
118',
28.9 (005)
28.2 (355)
3.4 (50)
3.3 (55)
74
73
''oiiule cleaned with 1.51 f^o, EDTA adjusted to pil G with liaOH
1183
1352
28.2 (305)
23.6 (130)
3.3 (55)
4.3 (70)
76
72
Temper ^ turc
3.4 (55)
a!j (74!
3.6 (87)
3.3 (63)
5.0 (77)
5.0 (77)
5.0 (77)
5.C (77)
23.0 (77)
25.6 (73)
26.1 (79)
25.6 (?:;)
25.6 (-78)
25.0 (77)
24.4 (76)
24.4 (76)
25.0 (77)
2C.1 (79)
25.0 (77)
Module cleaned with 1.5% f,a, EOTA and 2;: citric odd adjusted to pH 4 with MI-OH
1356
1601
23.6 (400)
28.9 (105)
I
4.5 (65)
5.9 (85)
72
71
25.0 (77)
24.4 (76)
Measured
Flux.
(5pn)
12.45 (3.29)
10.93 (2.53)
10.07 (2.C6)
6.70 (1.77)
7.33 (1.95)
7.00 (1.1,5)
6.93 (1.33)
7.15 (1.89)
6.21 (1.64)
5.93 (1.53)
6.66 (1.76)
6.02 1.59)
5.60 (1.43)
5.93 (1.58)
5.41 (1.43)
6.13 (1.62)
5.56 (l,47)
6.13 (1.62)
5.12 (1.35)
5.41 (1.43)
4.47 (1.18)
Corrected
Flux,"*
1 pnt
(GP"0
11.54 (3.05)
11.92 (3.15)
8.52 (2.25)
3.74 (2.31)
7.53 (1.99)
7.08 (1.37)
7.03 (1.37)
7.15 (1.89)
6.40 (1.69)
6. Co (1.60)
6.58 (1.74)
6.02 1.59)
5.68 (1.50)
6.17 (I.C3)
5.75 (1.52)
6.32 (1.67)
5.90 (1.56)
6.28 (1.66)
5.45 (1.44)
5.75 (1.52)
4.92 (1.30)
Concentration
umhos/cm
Feed Per"
3CCO 340
2750 290
2850 240
3050 210
2475 176
3050 205
3003 190
3000 162
3000 200
2950 210
2900 180
3000 235
2900 210
3050 133
3000 205
3000 160
3050 205
3000 165
3000 210
2050 140
3000 320
Corrected
Rejection***
83.6
39.4
91.6
93.1
92.9
93.3
93.7
94.6
93.3
92.9
93.8
92.2
S2.8
95.5
93.2
94.7
93.7
94.5
93.0
95.4
89.3
One S;r
14.50 psi = 0.987 a in
Flux correctej to 400 ps1 , 77 F, 752 conversion, and 2925
^jecr.lon corrected to 400 ps1, 77CF, 75i conversion, and
-------
cr>
Cleaning Solutions:
2% Citric Acid + NH.OH at pH 4
2% Citric Acid + NHTOH at pH 8
NaOH at pH 12
21 Citric Acid
1 ij£
1.5X
Na2EOTA
NH.OH at pH
NaOIPat pH 0
2% Citric Acid + NH4OH at pH 4
)
0 TOO 200 300 400 500 600 700 GOO 900 1000 1100 1200 1300 1400 1500 1600
Operating Time, Mrs.
Figure 3. Corrected productivity for sodium chloride tests vs. operating time
-------
Significant color changes were noted during flushing with each of the
cleaning solutions. Samples of the cleaning solutions were obtained
toward the end of the flushing period for the last three cleaning
solutions (D, E, and F), and spectrographic analyses were conducted
for thirty different elements. The results are shown in Table 4. The
cleaning solutions do not appear to contain large amounts of any pre-
cipitate-forming elements. This is consistent with the observation
that the cleaning solutions used were ineffective in restoring the
original productivity.
MODULE REJECTION
The rejection of a module is also important in determining the
capital cost of an RO system. If the rejection is low, additional
treatment of the permeate may be required in order to achieve the
desired rinse purity.
Rejection increases with operating pressure but decreases as
the feed concentration and conversion increase. Rejection is
essentially independent of temperature. Since the operating pressure
was maintained at or near the maximum recommended value, corrections
for variations in operating pressure were negligible. Although the
feed concentration varied substantially, correction factors for feed
concentration and conversion are not available for nickel rinse
waters. Therefore, no corrections were applied to the measured
rejections.
The rejections given in Table 2 for nickel, total solids, con-
ductivity, and boric acid are plotted as a function of operating time
in Figure 4. Nickel rejections were excellent (generally >99%), and
total solids and conductivity rejections very good. Boric acid
rejection was poor since at low pH it exists mainly in the unionized
form which is poorly rejected by all membranes. These rejections are
consistent with those measured previously for nickel baths ----\
17
-------
Table 4. RESULTS OF SPECTROGRAPHIC ANALYSIS
OF CLEANING SOLUTIONS AFTER USE
Concentration, mg/1
Element Solution D Solution E Solution F
Ag
Al
As
Ba
Be
C
Ca
Cd
Co
Cr
Cu
Fe
K
Me
Mn
Na
Ni
Pb
P
Si
Sn
Sr
V
Zn
<0.05
1
<0.5
2
<0.1
20
20
0.5
1
1
5
5
<0.5
20
1
20
25
1
1
5
1
10
<0.1
<0.5
<0.05
1
<0.5
20
<0.1
10
100
0.5
2
0.5
10
20
<0.5
10
1
1000
25
10
5
10
5
10
<0.1
0.5
<0.05
1
<0.5
2
<0.1
50
50
0.5
2
0.5
5
10
<0.5
10
1
200
25
5
10
5
2
10
<0.1
2
NOTE: Analyses conducted by Elliott Laboratories Inc., Andover, Massachusetts.
The spectra were also examined for the presence of Sb, Hg, Se, B, Ta, and
Ti — no indication was seen.
18
-------
TOO
90
80
O)
•<-}
a)
- 70
60
50
Nickel 0
Total Solids
Conductivity Q -
Boric Add O
1
I
0 TOO 200 300 400 500
600 700 800 900
Operating Time, Mrs.
1000 1100 1200 1300 1400 1500 1600
Fiqure 4. Rejection of various rinse water species vs. operating time
-------
It is significant to note that no decline in rejection occurred
over the duration of the field test. The high variability of the
boric acid rejection does not appear to be related to the concen-
tration, the pH or the operating time.
The rejection of Nad during standard Nad tests is given in
Table 3 and plotted in Figure 5 as a function of operating time on
nickel rinse waters. Upon exposure to plating solution, there was
a significant increase in rejection (from 88.5% to 93.5%). For the
duration of the field test, the rejection remained above its initial
value. The cleaning solutions produced a slight but consistent
increase in rejection.
DESTRUCTIVE TESTS
At the conclusion of the field test, the hollow fiber module
was opened and the fiber bundle was inspected to determine the cause
of deterioration in membrane productivity. Except for a very small
amount of solid particulates in the vicinity of the feed distributor
tube, the fiber bundle was entirely free of particulates. However,
the fibers were discolored. The clean fiber has a light yellowish-
beige or buff color about the same tone as a manila folder. The
fibers from the field-test module had a dark green-brown color similar
to drab "Army" green. In addition, the fibers were very oily to the
touch.
Several strands of the fibers were washed in a detergent solu-
tion. This resulted in complete removal of the oily foul ant and total
restoration of the original fiber color.
20
-------
100
80
c
o
o>
"->
01
70
60
50
~ I r
t
I
t
t t t t
Cleaning Solutions:
2% Citric Acid +
2% Citric Acid +
NaOH at pH 12
2% Citric Acid +
A.
B.
C.
D.
E.
F.
1.5/o Na,,EOTA +
1.5% Na^EDTA +
NK.OH at pH 4
NH?OH at pH 8
+ NH.OH at pH 8
NaOH at pH 8
2% Citric Acid +
NH4OH at pH 4
100 2CO 300 400 500 500 700 800 900. 1000 1100 1200 1300 1400 1500 1600
Operating Time, Mrs.
Figure 5. Corrected sodium chloride rejection vs. operating time
-------
SECTION VI
DISCUSSION
The most likely origin of the oily foulant observed during
destructive examination of the module is the high-pressure piston
pump used to pressurize the feed to the RO system. There was no
oil slick on the rinse from which the RO feed was drawn, so the oil
must have come from RO system itself. Since the high-pressure pump
is the only oil-lubricated pump in the system, it is concluded that
this pump developed a leak in the piston-rod oil seal and allowed
oil to enter the feed manifold.
The build-up of lubricating oil within the module was completely
unexpected. Therefore, the cleaning solution for organics removal
(0.5 wt% "Biz" at pH 10) was not attempted. Based on the success of
the detergent cleaning (Alconox solution) during destructive examina-
tion of the module, it is anticipated that a detergent cleaning during
the field test would have been much more successful in restoring the
original flux than the cleaning solutions actually used.
Destructive examination of the RO module indicated no significant
build-up of solid particulates within the module. It is therefore
concluded that the feed pretreatment employed (filtration through a
one-micron, string-wound, catridge filter) is adequate for this appli-
cation. The fact that the module became fouled with oil is a consequence
of the pump malfunction rather than an inadequacy in feed pretreatment.
Even though the module was badly fouled with oil, the rejections
(with the exception of boric acid) were very high and quite stable with
operating time. In the absence of oil fouling, it is anticipated that
the productivity would also have been quite stable. Therefore, it is
concluded that hollow-fiber polyamide B-9 permeators can be successfully
used for closed-loop treatment of Watts-type nickel baths. Since the
cost per unit volume of permeate is the same of spiral-wound and hollow-
fiber modules, the economics will be similar to the economics reported for
spiral-wound cellulose acetate modules — .
22
-------
The low rejection of boric acid has been noted previously —>—'—'—'—'
in work on RO treatment of nickel baths. The low rejection will result
in a build-up of boric acid in the rinses to a relatively high level compared
to other plating bath constituents. The effect of boric acid on subsequent
processing steps should be given careful consideration by platers comtem-
plating the use of RO for treatment of nickel rinse waters. The concentra-
tion of boric acid in the final rinse can be minimized by returing the RO
permeate to the first rinse rather than the final rinse.
The recommended flow schematic for RO treatment of Watts-nickel rinse-
water at New England Plating Company is shown in Figure 6. The estimated
bath evaporation and dragout are 0.3785 liters per minute (0.1 gpm) and
0.03785 liters per minute (0.01 gpm), respectively. Based on the bath com-
position given in Table 1, the total solids concentration of the bath is
approximately 280,000 mg/£. The recommended total solids concentration
for the final rinse following mickel plating in 37 mg/£—. Based on these
values the concentration in the first and second rinse tank? can be calcu-
lated and the required capacity of the RO system can be specified.
The flows and concentrations resulting from material balance calcula-
tions are shown in Figure 6. The system is designed to operate at a
maximum conversion per module of 50%. Based on the results of Figure 4 a
minimum total solids rejection of 96% was assumed. However, because of
oil-fouling of the RO module during the field test, the extrapolated pro-
ductivity results were not used. Instead, the initial productivity shown
in Figure 2 (8 liters/min = 2.1 gpm) was used and extrapolated to longer
operating times using factors published in the DuPont Technical Information
Manual for B-9 modules. The extrapolated productivity for a full-scale
4-inch diameter module after two years of operation is 5.7 liters/min
(1.5 gpm). A total of three full-scale, four-inch-diameter, B-9 modules
are required in order to achieve the rinse concentrations shown in Figure 6.
Based on costs from Abcor, Inc. (March, 1977) the installed capital cost
for this system, including pretreatment (cartridge filtration), controls,
stainless steel construction, and RO modules is approximately $15,000.
23
-------
Dragin=p.01 gpm
ro
Evaporation =0.1 gpm
Draqout 0.01 aom
\
Make-up Water
= 0.1 g pm
0.01 gpm ^
Bath
280,000 mg/£
Total Solids
^ 4.6 gpm
0.1 gpm
6.0 gpm ^
,7321 mg/fc
1 .4 gpm
Rl
1027 mg/JZ,
i
,4.5 gpm
421 mg/j,
1 .5 gpm
295 mg/x,
"--- ^
>""-- ^
4.5 gpm
9663 mg/£
•^v^^.
• — ^s~ -~s~~- -*•"
Ro
407 mg/2,
•^-~— f
•v^-^
R^
•^y^-
37 mg/£
1 .5 gpm
390 mg/£
- -^^
"^ ^^
3.0 qpm
14,300 mg/£
1 .5 gpm
575 mg/
"^- -^ j
"""• -»^J
1 .5 gpm
28,000 mg/ji
NOTE: 1 gpm = 3.785 liters/min
Figure 6. Recommended RO system for nickel line at New England Plating Company
-------
Table 5 gives a breakdown of projected operating costs for the system
based on operation for one shift per day and 240 days per year. The major
costs are for membrane module replacement and system depreciation. Although
the data of Figure 4 shows no decline in rejection, a module life of two
years was assumed on the basis of commercial experience with the cellulose
acetate membrane for treatment of Watts-nickel rinsewaters. The annual
operating cost for the system is approximately $4000.
Several credits can be claimed as a result of closed-loop rinsewater
recovery: first, the cost of recovered nickel-plating-bath constituents;
and second, the cost for end-of-pipe treatment of nickel rinsewater including
the chemical costs for precipitation and the sludge disposal costs. These
credits can be applied toward the operating cost, and if greater than the
operating cost, will result in a profit for operation of the RO system.
Various economic factors for the closed-loop recovery of nickel
rinsewater at New England Plating Company are shown in Table 6. For operation
of the nickel line one shift per day, the costs and credits (which here in-
clude only the value of recovered plating chemicals) are approximately equal
and the payback period (8.1 years) is only slightly less than the period
over which the capital investment is amortized (10 years). However, if the
nickel line is operated two shifts or three shifts per day the credits in-
crease by a factor of 2 or 3, respectively while the operating cost increases
only slightly. Thus the payback periods for 2-shift or 3-shift operation
(2.4 years and 1.4 years, respectively) are quite attractive.
25
-------
Table 5. BREAKDOWN OF OPERATING COSTS (1 Shift/Day, 240 Days/Year)
Power (at $.05/kwh)
Major power requirement is for high pressure pump
Flow rate = 6.0 gpm; AP = 400 psi; motor/pump efficiency = 50%;
conversion factor = 1715 gpm-psi/hp
Power consumed = i^O) (400) = 2>8Q hp = 2>Q8 kw
Annual Operation Cost = (2.08)(8)(240)($.05) = $200
2. Module Replacement (at $1317 per module)
For a module life of 2 years:
Annual Operating Cost = (3) |^|317) = $1976
3. System maintenance (3% of capital investment per year)
Annual maintenance cost = (.03) ($15,000) = $450
4. Amortization
For straight-line depreciation over ten years
with zero salvage value:
[$]j
10
Annual Operating Cost = ($15»0°0) = $1,500
5. Make-up Water (at $4.00 per 1000 gal from central DI system)
Annual Operating Cost = (0.1) (60) (8) (240) ($4.00) = $46.08
1000
6. TOTAL ANNUAL OPERATING COST = $4,172
26
-------
Table 6. ECONOMICS FOR CLOSED-LOOP NICKEL RECOVERY
Annual Cost 1 Shift 2 Shifts 3 Shifts
Elements Per Day Per Day" Per Day
Credit for Recovered Chemicals^ $4,516 $9,032 $13,548
Cost for RO Operation^ 4,172 4,418 4,664
Profit for RO Operation^ 344 4,714 8,884
Minimum Payback Period^ ' 8.1 years 2.4 years 1,4 years
(a) Calculated on the basis of 0.03785 liters per minute (0.01 gpm) dragout
and $1.03/liter ($3.92/gal) cost for plating chemicals (M&T Chemicals,
March 1977). Does not include credits for reduction in volume and
concentration of wastewaters to be disposed.
(b) Calculated as in Table 5 but with additional charges for power and DI
water for 2 and 3-Shift operation.
(c) Profits before taxes = credits minus costs
(d) Capital cost divided by the sum of annual profits and depreciation charges.
27
-------
SECTION VII
REFERENCES
1= Donnellv. R. G.. Goldsmith, R. L., McNulty, K. J., and Tan, M.
Plating, 61 (5), 432 (1974).
2. Donnelly, R. G., Goldsmith, R. L., McNulty, K. J., Grant, D. C.,
and Tan, M., "Treatment of Electroplating Wastes by Reverse Osmosis-,
Final Report EPA Contract No. R-800945-01. In Press.
3. McNulty, K. J., Goldsmith, R. L., Gollan, A., Hossain, S., and
Grant, D., "Reverse Osmosis Field Test: Treatment of Copper Cyanide
Rinse Waters", Final Report, EPA Grant No. 800945. In press.
4. McNulty, K. J., Grant, D. C., Gollan, A., and Goldsmith, R. L.,
"Field Demonstration of Reverse Osmosis Treatment of Cyanide
Rinse Water", presented at the AES 62nd Technical Conference,
Toronto, June 23-26, 1975.
5. Mattair, R., "Case Histories of Hollow Fiber Reverse Osmosis
Plants", presented at the 79th National Meeting of the AICHE,
Houston, Texas (March 18, 1975).
6. Anon., Industrial Finishing, 5p_ (7), 34 (1974).
7. Golomb, A., Plating, 60_ (5), 482 (1973).
8. Lonsdale, H. K., "Properties of Cellulose Acetate Membranes", in
Desalination by Reverse Osmosis, U. Merten ed., The M.I.T. Press,
Cambridge, Mass. (1966).
9. Golomb, A., Plating, 57_ (9), 1001 (1970).
10. Golomb, A., Plating, 59^ (4), 316 (1972).
11. Novotny, C. J. Finishers' Management, 18_ (2), 43 (1973).
28
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 . REF)ORT NO.
EPA-600/2-77-039
4. TITLE AND SUBTITLE
REVERSE OSMOSIS FIELD TEST:
NICKEL RINSE WATERS
TREATMENT OF WATTS
V.AUTHORIS) Kennetn J. McNuTTy, Robert L. Goldsmith, Arve
Z. Gollan, Walden Research Division of Abcor, Inc.,
Wilmington, Massachusetts 01887
8. PERFORMING ORGANIZATION REPORT NO.
3. RECIPIENT'S ACCESSI Or* NO.
5. REPORT DATE
February 1977 issuing date
6. PERFORMING ORGANIZATION CODE
9. PERFORMING ORGANIZATION NAME AND ADDRESS
The American Electroolater's Society, Inc.
Winter Park, Florida 32789
10. PROGRAM ELEMENT NO.
1 BB610
11. CONTRACT/GRANT NO.
R-803753
12. SPONSORING AGENCY NAME AND ADDRESS
13. TYPE OF REPORT AND PERIOD COVERED
Industrial Environmental Research Laboratory-Gin., OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
Final
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A field test was conducted to determine the feasibility of usina a polyamide
reverse-osmosis membrane in hollow fine fiber configuration for closed-looo
treatment of rinse water from a Watts-type nickel bath. Performance of the
membrane module was determined by measuring the productivity (flow rate of
purified water) and rejection (separation efficiency) as a function of operatina
time. Performance was monitored over 1600 hours (67 days) of operation and
2300 hours (96 days) of exposure to nickel rinse waters.
The results of the tests combined with the results from numerous commercial
installations indicate that the economics of closed-loon recovery of nickel
rinse waters by reverse osmosis are Generally attractive. Dependina on certain
factors soecific to each olatina line, the caoital investment for a reverse
osmosis system can be recovered in as little as one year or less of ooeration.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Electroplating*,
Waste Treatment,
Reverse Osmosis*,
Nickel*,
Rinse water
h.IDENTIFIERS/OPEN ENDED TERMS
Closed-loop treatment*,
polyamide hollow
fiber membrane*,
Watts-type nickel bath*
COSATi Field/Group
13B
13. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
39
20. SECURITY CLASS (This page)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
29
U.S. GOVERNMENT PRINTING OFFICE: 1977— 757-056/5581
------- |