EPA-600/2-77-161
August 1977
ELECTRODIALYSIS FOR CLOSED LOOP CONTROL
OF CYANIDE RINSE WATERS
by
George W. Bodamer
International Hydronics Corporation
Princeton, New Jersey 06540
for
Keystone Lamp Manufacturing Company
Slatington, Pennsylvania 18060
Grant No. S-803304
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
60SJ4
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research Laboratory
Cincinnati, U.S. Environmental rVotection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the views and policies of
the U.S. Environmental rVotection 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 increasingly 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.
The subject of the report was to evaluate a full-scale, closed-loop electrodialysis
system for brass plating cyanide rinse waters. The system proved to be inefficient and
therefore unsuitable for this application. To avoid future failures, this report stresses the
importance of membrane testing in solution to be treated on a laboratory scale before a
full-scale demonstration. For further information concerning this subject the Industrial
Pollution Control Division should be contacted.
David G. Stephen
Director
Industrial Environmental Research Laboratory
Cincinnati
111
-------
ABSTRACT
Full-scale trials of electrodialysis of brass plating rinse solutions were conducted in
the plant of the Keystone Lamp Manufacturing Company at Slatington, Pennsylvania.
The installation was intended to deplete dissolved salts in the rinse waters, thereby
rendering the water suitable for re-cycle to the rinsing operations and returning the salts
to the plating bath. Zero discharge would thus be achieved.
The system worked to a certain extent but was only about one-fourth as effective
as had been anticipated from prior work in which electrodialysis was used to effectively
treat sodium copper cyanide rinse waters. The reason for this was not immediately
apparent, and considerable time was spent exploring various hypotheses of a mechanical
or electrical nature.
Eventually it appeared that the inefficiency was caused by a reduction in perm-
selectivity of the an ion membranes as a result of retention of an insoluble 2:inc compound
or of a tightly adsorbed zinc complex anion. No remedy that would not disturb the
chemistry of the plating bath could be found for this situation, so the work was
terminated.
The lack of success in this particular application should not discourage attempts to
apply electrodialysis to other waste treatment processes. Small-scale preliminary tests
should be made, however, if there is any question about the interaction of membranes
and dissolved salts.
This report was submitted in fulfillment of Grant No. S-803304 by Keystone Lamp
Manufacturing Company under the sponsorship of the U.S. Environmental Protection
Agency. This report covers the period July 1, 1974 to December 31, 1975, and work
was completed as of January 12, 1976.
Iv
-------
CONTENTS
Foreword I??
Abstract iv
Figures vi
Tables vii
Acknowledgments viii
1. Introduction 1
2. Conclusions 4
3. Recommendations 5
4. Development and Demonstration Studies
International Hydronics Corporation, Electrodialysis Cells 6
Asahi Glass Co., Ltd., Electrodialysis Cell 14
Laboratory Investigations 17
5. Results and Discussion
Performance of Hydronics Stacks on Keystone Brass Plating Rinses ... 19
Performance of Asahi Stacks on Brass Plating Solutions 24
Laboratory Investigations 25
-------
FIGURES
Number Page
1 Schematic Flow Diagram of Electrodialysis of
Electroplating Rinse Waters 3
2 Schematic of Installation using HydronicsE. D. Stacks 7
3 Schematic of Installation using Asahi E.D. Stacks 15
4 D . C. Amperes vs. Applied Voltage for Solutions and
Membranes 28
5 Three-Chamber Cell Electrodialysis 29
VI
-------
TABLES
Number Page
1 HydpMiics Stacks: 80 Chambers in First Stage,
40 Chambers in Second Stage 9,10
2 Hydronics Stacks: 44 Chambers in First Stage,
40 Chambers in Second Stage 12
3 Drag-Out in Keystone Brass Plating Line 13
4 Asahi Stacks: 99 Chambers in Both Stages 16
5 Asahi Stacks: Both Stacks in Parallel Operating as First Stage .... 16
6 Ranges of Concentration 20
7 Concentration of Copper and Zinc in Anion Membrane
MA-3475 After Use in Brass Electrodialysis:
Comparison with Brass Solution 26
8 Three-Chamber Cell E.D. Characteristics of Copper, Zinc, and
Brass Solutions 30
9 Three-Chamber Cell E.D. Concentration Changes in
Center Chamber 31
10 Efficiency of Electrodialytic Depletion 32
VII
-------
ACKNOWLEDGMENTS
The writer thanks Dr. Sidney B. Tuwiner, consultant, for his advice and
suggestions.
The encouragement, perseverance, and diligence of James T. Stevens,, Vice
President, Manufacturing, and of George Unangst, Plating Supervisor, both of Keystone
Lamp Manufacturing Company, as well as members of their staff, was invaluable and is
deeply appreciated by the writer.
We thank Asahi Glass Company, Ltd., for the opportunity to use their excellent
electro-dialysis cells in this application .
-------
SECTION 1
INTRODUCTION
In January 1973, International Hydronics Corporation, Princeton, New Jersey,
entered into an agreement with Dr. Sidney B. Tuwiner, New York, to develop and
market Dr. Tuwiner's process for the concentration of electrolyte from dilute washings by
electrodialysis in a closed system as described in his U.S. Patent No. 3,674,669,
July 4, 1972, and (later) in his report "Investigation of Treating Electroplaters Cyanide
Waste by Electrodialysis", EPA-R2-73-287, December, 1973.
As applied to an electroplating line, the Tuwiner process operates as follows. The
first rinse solution is circulated through the depleting chambers of an electrodialysis
stack, and the bath solution is passed through the concentrating chambers of the same
stack. The second rinse solution is passed through the depleting chambers of a second
electrodialysis stack, and the first rinse solution is put through the concentrating
chambers of the second stack. Salts are thereby transferred from the second rinse to the
first rinse, and from the first rinse to the bath. The concentration of the rinse -solutions
is thus maintained at a low enough level that they can be continuously recycled and
accomplish the desired rinsing effect while the plating salts are returned to the bath. A
schematic flow diagram is shown in Figure 1.
The Tuwiner process has the evident advantages, first, that waste rinses containing
polluting materials do not have to be treated prior to discharge since there will be no
discharge and, second, that valuable chemicals and process water will be conserved.
It is well known that in order for a multichamber electrodialysis stack to achieve
the greatest possible efficiency in depleting the salt concentration of one stream while
increasing the concentration of the other, it is necessary to eliminate or at least minimize
any cross-leakage of the two streams within the stack. Dr. Tuwiner indicated that he
felt an improvement in this feature of stack construction was needed, and International
Hydronics began to design, fabricate, and test various configurations early in 1973.
The design ultimately adopted consisted of forty PVC frames 0.159 cm thick with
an open area of 0.074 square meters. On each side of a frame was a pure gum rubber gasket
0.079cm thick. Anion and cation membranes were MA-3475 and MC-3470, manufactured
by lonac Chemical Co., Birmingham, N.J. The construction of the channels leading
from the manifolds into the chambers was subsequently found to be very similar to that
described in Patent Nos. 3,219,572 (1965) and 3,235,481 (1966) assigned to A.M.F. by
the inventor B. M. Zwart, Jr., although Hydronics was unaware of these patents at the
1
-------
time. Cross-leakage within a 40-chamber stack of this design amounted to less than 5 cc
per minute at a pressure differential between chambers of 0.14 kg/sq cm and at a flow
rate through one set of chambers of 22.7 liters per minute.
Stacks of this type were tested in the Hydronics1 pilot plant on sodium copper
cyanide solutions during the latter part of 1973. What the Tuwiner process refers to as
"first stage" operation-depletion of ions from a first rinse with corresponding transfer of
ions to a bath solution-was simulated using a "bath" solution of 60,000 to 70,000 ppm
total cyanide. The "first rinse" solution was typically depleted to 0.02 of those
concentrations when the drag-out from "bath" to "rinse" was 3.79 liters per hour. The
flow rate of both streams was about 37.85 liters per minute; the pressure of the depleted
stream was maintained about 0.14 kg/sq fm higher than that of the concentrated stream.
Voltage applied to the stack was 125 volts; amperage was 25-40 amps, depending on
temperature.
Similarly, "second stage" operation was studied using a "first rinse" of 600-700
ppm total. Using the same stack and a 3.79 liters per hour drag-out rate, the depleted
stream ("second rinse") was again maintained at 0.02 of the concentrated sttream. The
amperage was about 0.3 amps at an applied potential of 125 volts across the stack.
In general, the performance of the stacks described above was felt to confirm
Dr. Tuwiner's findings in a cell of fewer chambers as reported in the above-mentioned
EPA report.
At this stage, Hydronics wanted to find a plating plant that would permit
installation of the system and testing under actual operating conditions. A particular aim
was accumulation of data on the life expectancy of the membranes. Hydronics was
fortunate to be given the opportunity to conduct such trials at the Slatington,
Pennsylvania plant of the Keystone Lamp Manufacturing Company. The solutions to be
treated were the first and second rinses following a typical brass bath in a continuous,
automatic plating line. At the time, it was assumed that a sodium copper zinc cyanide
solution would behave just as had a sodium copper cyanide solution in our pilot runs.
The equipment was installed at Keystone in February, 1974.
During the next 2 months, the system was run during the normal 8-hour shift on 16
days. Performance was poorer than expected to the extent that the rinse concentration
rose steadily instead of being depleted or being held at a constant low value. This was
blamed in part on various mechanical and chemical problems which were encountered and
corrected, and in part on the assumption that the drag-out rate was much higher than the
system had been designed to handle: 3.79 liters per hour.
Problems of electrode corrosion, buildup of brass on the first stage cathode requiring
frequent removal, and precipitation of copper and zinc cyanide (particularly in the
second stage depletion chambers) were especially troublesome and required a major
reorientation of our concepts.
-------
The stacks were returned to Hydronics1 pilot plant for revision. Platinized titanium
electrodes were installed at both ends of the stacks instead of stainless steel. Switches
were provided to reverse the polarity of the potential applied to the stacks, and three-
way valves were installed to interchange the depleted and concentrated streams. These
measures permitted dally potential reversal and stream interchange and solved the above-
mentioned problems.
It was this system that was in operation when the present EPA grant went into effect
on July 1, 1974. The grant was to extend to January 31, 1975, and was to provide for
accumulation of data indicating the reliability of the system, the life of the membranes,
and the economics of the process.
PLA
TH
Fl
Rl
ST
SE
SECOND
RINSE
a
LLJ
£
Q
LU
Q_
ULJ
a
0
u
o
LU
FIRST
ELECTRODIALYSIS
STACK
SECOND
ELECTRODIALYSIS
STACK
Figure 1. Schematic flow diagram of electrodialysis of electroplating rinse waters.
3
-------
SECTION 2
CONCLUSIONS
From extended plant trials of electrodialysis of brass plating rinse waters for return
of chemicals to the bath and demineralization of the rinses for reuse, the following
conclusions were drawn:
1. The electrodia lysis of brass solutions carried out under plant conditions was
only about one-fourth as effective as that achieved with sodium copper cyanide solutions
in prior inhouse pilot plant studies.
2. Various possible mechanical and electrical malfunctions were ruled out as
causes of this inefficiency.
3. Two sets of electrodialysis stacks, one developed by International Hydronics
Corporation, the other manufactured by Asahi Glass Company, Ltd., differing primarily
in chamber design, performed similarly.
4. Increasing the membrane area by using two stacks in parallel on the first rinse
did not improve deionization rate.
5. Laboratory experiments indicated that the inefficiency of brass solution
electrodialysis could be attributed to retention of some insoluble zinc compound or
complex ion on the anion permeable membranes, thus reducing their permselectivity.
6. In view of the unsatisfactory performance for this particular application, it was
considered unwarranted to continue trials for a longer period which would have been
necessary to fulfill the original aims of the U.S. Environmental Protection Agency (EPA)
grant: establishment of system reliability, membrane life, and economics.
-------
SECTION 3
RECOMMENDATIONS
Before undertaking full-scale trials in the future, it would be well to determine the
permselectivity of membranes in the solutions to be used and to conduct small-scale,
three-chamber cell tests as described in this report.
The experience described in this report on the electrodialysis of actual brass
solutions should not discourage further attempts to apply the principle to other rinse
streams. The potential for conserving chemicals and water and attaining zero discharge
remains very great.
-------
SECTION 4
DEVELOPMENT AND DEMONSTRATION STUDIES
INTERNATIONAL HYDRO NIC S ELECTRODIALYSIS CELLS
When the present EPA grant became effective on July 1, 1974, the system in
operation at the Keystone Lamp Manufacturing Co. was essentially as previously
described and shown in Figure 2 with certain exceptions as follows.
1. The first stage stack consisted of 80 chambers (40 cell pairs) and the second
stage stack of 40 chambers (20 eel I pairs). The increase in chambers in the first stage had
been made in order to reduce amperage which was causing overheating of this unit.
2. The simple, two-way valves (V-l c, V-l d, V-2 c, V-2 d, etc.) were replaced
with three-way valves which were interconnected in such fashion as to permit inter-
changing the depleted and concentrated streams in each stage on a daily basis.
3. The rectifier was connected to the electrodes through double-pole, double-
throw switches so that the polarity could be reversed at the same time the streams were
interchanged.
Operation with the electrodes having a certain condition of polarity and the
concentrating and depleting streams flowing through the proper chambers was said to be
"Forward" operation. The reversal of polarity and the corresponding interchange of
streams was said to be "Reverse" operation .
It is to be noted that the concentrated stream for the first stage was contained in a
separate recirculating drum independent of the actual plating bath. This permitted the
study of different concentrations in this stream and also allowed for checking the
suitability of this solution for return to the plating bath by withdrawing a sample for Hull
cell tests. In the first stage the solution from the recirculating drum was passed through
the electrode chambers in two separate streams. In the second stage, the first rinse was
passed through the anode chamber, and the second rinse through the cathode chambers.
Previous experience at Keystone had shown that the inefficiency of the system
could be compensated for by transferring part of the rinses into the plating bath each
morning. Daily evaporation from the plating bath was of such magnitude that about half
of the first rinse could be transferred to the bath as makeup. The same volume of second
-------
-------
rinse was then transferred to the first rinse tank. The second rinse tank was then filled
with softened water. In previous runs the electrodialysis system transferred enough salts
out of the rinse streams that the second rinse was held at an acceptably low concentration
for the subsequent plating operation that the parts were subject to.
A typical day's operation covered the regular 8-hour shift in the plating
department. The three-way valves and the double-pole, double-throw switches were set
for the opposite direction of the previous day's run. All pumps and the rectifier were
turned on simultaneously by means of a single "start" button on a central control panel.
On data-logging forms the following information was recorded: time, first and second
stage amperage, volts, flow rates in gpm for all streams (from rotameters), pressure in psi
In all streams (from pressure gauges), the temperature of the solutions in therecirculating
drum and both rinse tanks, and the conductivity of all solutions. These data were
collected periodically throughout the day's run and at shut-down.
Conductivity had been related to total cyanide concentration by analysis of
suitable solutions over a wide range of concentration. For concentrations in the range
of the recirculating drum (20,000-70,000 ppm) the solutions were diluted 1:200 with
distilled water to get them into a suitable range on the conductivity meter,. Similarly,
first rinse solutions (1000-7000 ppm) were diluted 1:20 while second rinse solutions were
measured without dilution.
Forty-four runs were made with the system as described above. In nine cases the
data were not complete. For the other 35, the most relevant data are presented in
Table 1. Certain parameters in these runs were fairly constant and fell in the ranges
indicated as follows.
Flow rate:
First stage, both streams 37.85 to 49.21
liters per minute
Second stage, both streams 30.28 to 37.85
liters per minute
Pressure of feed streams:
First stage 1.26to 1.48 kg/sq cm
Second stage 1.26to 1.48 kg/sq cm
(Depleted streams were maintained at the same pressure as the concentrated stream
in the same stack or at pressures up to 2 psi higher than the concentrated stream).
8
-------
LU
O
h-
CO
Q
z
o
sOlO«OlOOCM' CM
V
c
CM
.
KO co«nix;9FcocoocNis.a«co
CM> CO" CM CMCOCM CM
»no*oi>o<>>MD«oorooorv
-- CNCNCMCMCOCOCO
I 8888S8888888?888888§88888888
o o o o o o o o o o o o o
rxCOlO CMCMCM CM--
00000000000
O-^-CNOOOOOO3>O
hsCM'-OlO^'*^t' O-
COCOCNCMCMCNCNCOCM
CO o o 5
- CM
ir> CM
O-st-TtCOCOCMCOCOCOCMCOCOCOCOCM
CO *
I I
«o
CNCMCMCMCNCOCNCOCOCMCMCOCOCOCOCOCMCM
OOMJOOCM'OIXO'OOoOOOt'sO-'tCOO' «OOO^"'Of>OOO^t' W)OO
_. COCOCOCOCMCMCOCOCM" CNCMCMCO^-COCOCMCMCM CM CM
COCOCOCMCM- CMCM'' CM1 CMOOCNCOCM CM1 CM"
8
CMCMCMCNCNCNCMCM
S
-------
U-l
<
CO
O
Z
o
u
LU
CO
z
co
eg
LU
co
X
U
o
<*
*
LU
O
CO
to
eg
Z
CO
eg
eo
5
X
U
o
co
U
<
to
co
U
Z
O
eg
O
N
X
TJ
D
^C
4-
o
u
i
i
^*~
LU
CQ
H-
u
c
U
%
^c
eg
TJ
C
0
u
CO
*
u
c
o
U
8
ft
LU
E
O
J
0)
ol
ex
E
^(
ft
i
§
Of
u
^-
a
a
i
c
in
t
£
to
a
a
I§
u_
*-
o
CO
*
u
o
U
V
CN
co
*~"
CO
O> ^O'~CMCO
^O Lrt f^ fj JN^ <\j |^ f\i
eo|olocorx*orxi
*""
S-^t CN *O O O CO «O
O CN O> CO Ov CO >O
CN CN CO CM CN CN ,
88888888
r*s ^s ^K co ^K >o ^o c^
CO CO CO ^* CO CO CO CO
83888838
" rx co co "" *o rx fN
CN CNCN« CO
§!§§§§§§
00 r< 00 C3 r^ ^K C^ ^3
CN CNCOCOlOCN
W| Vl *tj
COCOCOCO-^-CNCOCN
1 u!,J> ' ' ' J, '
*
CN CNCOCOCNCNCO
O CN LO O
CN
1 1 1 1 1 1 1
rxoorNOoocorx
^N CN CO tN CO IT) CO
O'^O^OCOCNOO'O
rxrxcorxrxrxrxix
^3
p
L
0 - CO T* ^
CNCOCOCOCOCOCO Q
V
TO
1
u
"o
o
4-
E
a.
a.
.£
c
JO
"c
0)
u
c
o
*
10
-------
Temperatures:
Recirculating drum 80 to 117° F
First rinse tank 90 to 114° F
Second rinse tank 74 to 110 F
(Temperatures depended partly on ambient temperature; concentration of streams,
which had some effect of amperage; and whether or not cooling was applied to the
recirculating drum by means of a cooling coil. Temperatures rose 5 to 15° F during a
run).
Voltage:
Both stages 125-130 volts
Variations within the ranges of the above parameters were not observed to influence
the performance of the installation in a systematic way; therefore the corresponding data
on individual runs is not shown in Table 1.
The actual hours elapsed between the first and last sample is tabulated as "Test
Hours", although the runs usually continued throughout the entire plating shift, 8.5
hours. Amperage In each stage is given at the time of the first and last sample.
The concentration of the recirculating drum is given as an average value. Although
this concentration usually rose by 1000 to as much as 5000 ppm, such increases frequently
did not correspond to transfers out of the first rinse stream and seemed to be due to
evaporation.
The total cyanide concentration of the first and second rinse streams at start and
finish was tabulated and the changes in that concentration per hour, " ^ ppm/hr ", was
calculated. This latter value is of interest when compared to drag-out rates, and to the
operation of the installation when there was no drag-out, as discussed later.
In order to investigate the effect of the number of chambers in a stack, the first
stage was reduced from 80 chambers to 44 chambers, the second stage remaining at 40
chambers. The flow rates in the first stage were now in the range of 34 to 45.4 liters per
minute while the pressure on the feed streams, the temperature experience, and the
voltage were essentially as described previously. The other data obtained during four
runs with this assembly are given in Table 2.
Of prime interest were actual measurements of drag-out in the Keystone brass
plating line when it was running normally and when the electrodialysis equipment was
not in operation. This information was collected on eleven different days and is
tabulated in Table 3.
11
-------
LU
0
^
I/
Q
Z
o
U
LU
to
Z
to
LU
CO
^
X
U
o
LU
O
<
to
to
£
1 1
LL.
Z
to
Qi
LU
CD
5
X
5
t/5
U
^
i
to
to
U
Z
o
Q
S^_
X
CN
CO
^i
t
*
U
O
U
(U
in
C
S
^j
O
U
l_
*
1
a
fl
V)
iZ
"t
p
^
Q.
(X
xj
00 *O O IO
CO ^^ "^^ NQ
o hx ^j* rx
O IX 00 00
*7J ON 00 ON
CO CO ^ CO
CN IX CO IX
ON CN O NQ
CN CN CN
i i i i
^C
u.
1
to
8888
sf O O tr>
CN CO CO CO
0000
co o co -^f
*O O^ ^0 ^O
r«* ^M P«. | .
E
3
Q
u
(U
u.
%.
E
^
J{
u
c
o
U
(U
CN
to
^_
to
IA
|l
- x
O Q 0 O
O O 0 O
-------
LU
Z
_J
o
z
1
^f
1
Q_
oo
oo
Cki
CO
LLJ
Z
o
^^
oo
LU
Jbl
Z
^
l_
Z)
o
1
o
C£
Q
CO
LU
CO
1
c
c
U
*
u
c
0
U
OOCMOOCNO-CNCO'<}-> CN
OOOOOO- O
IO ^- hs CS
1 1 1 1 1 1 1 ^^ 1
00*0
1 ' ' ' ' ' * ,£; ^ J^ 1 I
M ^ ""*
Cs555o2§CNCKCN^8S^
CO CNCNCNCNCOCOCNCN
88888888888 ,
CN'^tCOlOCS^t' O-^-CNCM '
OOCO^-KOs-CO000000^
§^^ *^ ^? C? ^^ ^5 C^ ^^ C^ ^^
^5 ^5 ^? ^5 ^5 ^5 CO ^^ ^0 ^^
*" ^ i1^ ^^ ^o ^o ^T r^s c^ ^" ^ i
CMCNCOKhxCN-^ CM
10
'O 1^ w) *»/ *O i«O Wj IO *O *O **)
>o rx rx rx tx rv rx i\ ix N! N! '
<0
CO
n
u
v* ^^ co ^r ^o ^o ^^ oo ^^ *ij ^^ ^
|^B fMH .^J"
-s.
O)
o
c
o
_.
t_
J_
1C
**
_
o
CO
o
CN
P
.1
.^
1
-------
Only one opportunity presented itself to determine the way in which the Hyon
installation (with 80 chambers in the first stage and 40 chambers in the second)functioned
in the absence of any drag-out, i.e., when the plating line was not in operation. This
run was made following Run 29, Table 1. The pertinent data are given in Table 1 and
discussed hereafter.
ASAHI GLASS COMPANY, LTD., ELECTRODIALYSIS CELL
Asahi Glass Company, Tokyo, Japan, generously loaned International Hydronics
two of its Model DW-O electrodialyzers for testing in the brass rinse application. Each
of these units had the following characteristics.
Fifty concentrating chamber frames; 49 depleting chamber frames.
Frames 2 mm thick, made of resilient rubber. Open area 511 sq. cm.
Distributors leading from feed manifolds into chambers made of multi-
tubular polypropylene.
Cation exchange membranes: Selemion CMV; Anion exchange membranes:
Selemion AMV (products of Asahi Glass Co .).
Electrode chambers fed with streams different from those passing through the
main portion of the stack.
Stainless steel cathode and anode .
These cells were tested hydraulically by passing water through one set of chambers
and looking for the emergence of water from the other set of chambers; no internal cross-
leakage was observed. One of the stacks was tested for electrodialysis of sodium-
copper-cyanide solutions in the Hydronics pilot plant, and was found to deplete a
simulated first rinse solution whjle a "bath" solution was being dragged into it at a rate
of 3.79 liters per hour. Performance here was comparable to that found with the
Hydronics stacks mentioned in the Introduction.
On the strength of these tests, the two Asahi DW-O stacks were installed at
Keystone on the brass plating line in place of the Hydronics stacks. The system is
sketched in Figure 3. It is seen that separate streams are fed to the electrode chambers
in both stages. In the first stage, solutions of 52.4 gm NaCN/liter were reciirculated to
both electrodes; in the second stage, solutions of 7.49 gr NaCN/liter were used.
Three runs were made, and the data are tabulated in Table 4.
14
-------
H EH [gl-
15
-------
to
LU
O
to
X
t
O
CQ
Z
CO
eg
CHAMBE
8:
ft
CO
V
U
CO
X
$
Tf
LU
I
CQ
<
h~
*
O
o
U
Q>
>0 10 0
«o co
+ + Hr
o o o
S2 t 3
^ «o ^-
10 o o
tx to o
CO CN
*
u
c
o
u
1
eg
«
u
LU
E
D
Q
u
o>
Cg
in
0)
UP
q>
DJ
<
!_
a
a
-5
^
^c
LU
*
U
c
U
0)
CN
CO
,_
to
in
In D
OJ O
1- X
O 00 00
CO O CN
lO CN
1 + +
o^ ON"CN*
CM
III
». *. V
co hx
co
§Q
o
*. ^ V
* "* "O *
1 1 1
§ <
>^ H-
o to
° 5
p ^
E <
&
c u->
g "
O co
" _jf
I -
c
0)
u
c
o
u
*
*
u
c
o
U
(u
^c
'of.
en
LU
l_
1
<
±
'jc
LU
j-
J5
to
00 O
CO 00
co
+ +
88
00 CN
cvTo"
o o
o o
CN 00
CN"
1
<
CN
CO
,_
co
lO
CO CO
o
o oo
V)
fe ^
aj o
I X
CO K
CO <
^ ^
t_
V
§i
eg 3
CN
-------
When it appeared that the performance was tending in the same direction as that of
the Hydronics stacks, the second stage Asahi stack was re-piped so as to treat the
solutions in the recirculating drum and the first rinse tank.(separate solutions of 52.4 gm.
NaCN/liter were now recirculated to these electrode chambers also). Thus both stacks
were now arranged to function in parallel in "first stage" operation. The effect was that
of doubling the cross-sectional area of exposed membrane without decreasing voltage per
cell pair (one depleted and one concentrated chamber) and performance should have been
materially improved. The data for two runs is given in Table 5.
LABORATORY INVESTIGATIONS
As it became apparent that electrodialysis of brass solutions proceeded less
efficiently than that of sodium copper cyanide solutions, explanations were sought by
means of various laboratory experiments.
Analysis of Anion Permeable Membrane from Brass E.D. Cell
A sample from one of the an ion exchange membranes from the HyaVonics first stage
stack when it was cut down from 80 chambers to 44 chambers (after 45 days of operation)
was analyzed for its copper and zinc content. The membrane was washed thoroughly in
distilled water and then cut into pieces about 0.318 cm square. These were digested in
concentrated HNO3 until only a scanty, granular residue, presumably the fluorocarbon
binder, was left. The acid extract was diluted with distilled water and the solution
analyzed by atomic adsorption spectrophotometry for copper and zinc. Brass plating
solution was also analyzed for copper and zinc for comparison
Membrane Resistance in Copper Cyanide and in Brass Solutions
A small PVC cell, built in two halves, was used for simple measurements of the
resistance of membranes while direct current was passing through them. A steel cathode
and a graphite anode were used. The cross-sectional area of the cell was 12.7 cm wide
x 7.62 cm deep (96.62 sq cm), and the electrodes were 1.91 cm apart.
The cell was first filled with the solution of interest - either a sodium-copper-
cyanide solution or a sodium-copper-zinc-cyanide solution. (In both cases, the
solutions used contained approximately 60,000 ppm of total cyanide). Voltage from a
small rectifier with voltage continuously variable from 0-15 volts was applied and was
raised so as to make potential readings at 0.5, 1.0, 1.5, 2.0, 2.5, 3.0, 4.5, and 5.0
amperes. The entire course of readings was made in 2 minutes in order to minimize as
far as possible the changes in composition and concentration which occurred as a result
of electrode reactions.
The second step was to insert, between the two halves of the cell, an lonac
membrane which had previously been equilibriated with the desired solution by soaking
17
-------
in that solution for two days. The voltage-amperage readings were repeated as before.
The data were plotted on coordinate graph paper in order to facilitate a qualitative
evaluation of the possible interaction of membranes and ionic species.
Small, Three-Chamber Cell Tests of Electrodialysis
The same PVC cell described above, but modified to provide a 0.635 cm thick
center chamber, was used to conduct some simplified electrodialyses on cyanide
solutions. In these experiments, an lonac cation permeable membrane was placed on the
cathode side and an lonac anion permeable membrane on the anode side, both membranes
having been previously equilibriated for two days by soaking in plating bath solutions of
interest. The bath solutions were placed in the electrode compartments, while 200 cc of
a mixture made by diluting the bath 1:50 with distilled water was placed in the center
chamber. The voltage was set at 12 volts. Amperes were read periodically, and the run
was terminated when the amperage fell to 0.3 amps or less. The center chamber was
analyzed at the start and at the end, and ampere efficiencies were calculated.
18
-------
SECTION 5
RESULTS AND DISCUSSION
PERFORMANCE OF HYDRONICS STACKS ON KEYSTONE BRASS PLATING RINSES
From the very outset, the installation was incapable of depleting the first rinse
fast enough to keep up with the drag-out rate. Table 1 shows increases in concentration
of the first rinse from start to finish ranging from a low of 81 ppm/rir (Run 17) to a high
of 382 ppm/hr (Run 27). Run-to-run variation could not be readily attributed to any
one of the following parameters.
1. Flow rate of the streams.
2. Pressure in the feed lines.
3. Temperature.
4. Amperage (which is a function of applied voltage, temperature, and stream
concentrations).
5. Absolute concentration.
6. Concentration differential between concentrated and depleted streams.
7. Whether the stacks were operating in "forward" or "reverse" fashion.
It is believed that the day-to-day variation in performance was primarily related
to drag-out variation. Over a period of several weeks Keystone made available the
figures on square meters of parts plated per day. However, it developed that these
figures could not be related to degree of depletion either, and it is obvious that the
shape of parts (whether flat or cup-shaped) is as important as simple projected area in
determining the amount of drag-out. We have not attempted to quantify this shape
factor because of the great variety and complexity of shapes plated on this line.
The above comments generally apply to the second stage performance, also.
However, the second stage stack appeared to be depleting the second rinse somewhat
more effectively. Five runs actually resulted in over-all decreases in concentration
during the operating period.
19
-------
Run Number A ppm/hr
1 _8.6
2 -0.8
5 - -2.1
16 -1.0
25-- -2.7
Run No. 9 showed the same concentration at start and finish. However, the other
29 runs showed increases in second rinse concentration ranging from 0.2 (Run No. 20) to
12.5 ppm/hr (Run No. 28).
The 80-chamber (first stage) stack at 125 volts resulted in approximately 3 volts
per unit cell. Much of the prior work (including some of Tuwiner's) was carried out at
about 6 volts per unit cell. The latter condition, of course, provides a higher potential
over the back-EMF due to concentration differential. In order to see if this would have
significant benefit, the first stage Hydronics stack was reduced to 44 chambers, while
the second stage stack remained at 40 chambers.
The data for four runs collected with this configuration, shown in Table 2, reveals
essentially the same performance as found with the larger first stage stack. The range of
concentration increases observed during these runs may be summarized in the following
table:
TABLE 6. RANGES OF CONCENTRATION
A ppm/hr
Run Number First Rinse Second Rinse
36 92 3.8
39 267 6.5
During most of the work described above we were not alerted to the true state of
affairs for two reasons.
1. From a practical standpoint, it was possible to maintain the second rinse at an
acceptably low level by transferring part of the rinses toward the plating bath each day
as described in the Experimental section.
2. It was assumed that the drag-out greatly exceeded the design capability of 3.79
I iters per hour.
20
-------
On two different occasions, the electrodialysis system was shut down for several
days for inspection of the stacks and servicing of auxiliary equipment such as pumps and
filters. During those periods the drag-out was measured by sampling the rinse tanks near
the beginning and end of each shift. The resulting data are shown in Table 3 and
indicate considerable day-to-day variation in concentration rise. However, the average
was 242 ppm/hr. If one calculates the increase in concentration of the first rinse
solution on the basis that 3.79 liters per hour of bath containing 64,000 ppm total
cyanide is dragged into the first rinse tank, the volume of which is 984 liters, a rise of
1/260 x 64,000 = 246 ppm CN/hr would be anticipated . Thus, the average drag-out
rate is in the electrodialysis system's design range of 3.79 liters per hour, and on only
three days did the drag-out appreciably exceed that rate,
It is of interest to note that, in this line, it might be possible to maintain the
second rinse at a sufficiently low concentration, without electrodialysis or other
treatment, simply by starting with water in the first rinse tank, and, each day there after,
pumping as much of the first rinse into the plating bath as evaporation will permit,
similarly transferring an equal amount of second rinse into the first rinse tank, and
making up the second rinse with water (note especially runs 8, 9, and 10, Table 3).
This would eliminate the running rinse normally employed in tank 2.
Table 4 shows how the Hydronics stacks performed during one 7.5-hour period
(following run 29, Table 1) after the plating line was shut down. Here it is seen that
the first stage was removing only about one-fourth (66.7/242) of the average drag-out
required to hold a steady concentration in the first rinse, and the second stage about
three-fifths (10.7/16.9) of that required to stabilize the second rinse.
During this period of study, four main hypotheses were advanced (by various
interested parties) to explain the relatively poor performance at Keystone on brass
solutions as compared to that previously experienced on copper cyanide solutions. These
were:
1. Physically defective membranes.
2. External current loss between electrodes in the first stage.
3. Poor flow distribution among chambers.
4. Some unexpected effect of the presence of zinc.
These hypotheses will be discussed individually.
Physically Defective Membranes
Obviously a hole of any sort in any one of the membranes will work against
efficient dialysis. The lonac membranes employed were the standard materials MC-3470
21
-------
and MA-3475. They were cut to approximate size, soaked in hot water, and then in
cold water for several days. They were then die-cut to exact dimensions while wet. At
this point they were inspected closely against a light source for pinholes, slits, or thin
areas. The occurrence of such defects was rare, but when they were found, the
membranes were rejected. Using this technique, there was the same degree of quality
control exerted prior to cell assembly both before the pilot plant studies and before the
Keystone Lamp work.
Further, newly assembled stacks were tested for internal cross-leaks (one cause of
which would be perforated membranes) by passing water through one set of chambers at
pressures up to 0.14 kg per sq cm while the other set of chambers was empty. Stacks
were considered satisfactory if less than 5 cc per minute issued from the empty chambers.
Prior experience had shown that when membranes were damaged as a result of some
serious malfunction, such as severe overheating by contact of membranes with an
electrode (in earlier designs) or puncture of membranes by metal deposited on the cathode
(before daily polarity reversal was introduced), there was no mistaking the condition.
The concentrations of the two streams rapidly approached each other as complete mixing
began to occur, and the levels in the two tanks feeding the stack would often rise and
fall, respectively, at easily observable rates. When such a stack was subsequently
tested hydraulically, the cross-leakage into an empty set of chambers would far exceed
5 cc per minute . Upon disassembly, visual inspection of the membranes usually would
quickly reveal perforations.
Since none of these effects was observed for the runs described above, it seems
most unlikely that the inefficiency of the cells at Keystone can be ascribed to physically
defective membranes.
External Current Loss Between Electrodes in the First Stage
Two routes for current passage, other than through the stack, existed for the first
stage. These were through the 0.635 cm feed lines and the 1.27 cm return lines to the
electrode chambers, both carrying concentrated "bath" solution. The total distance from
cathode to point of origin on the feed line and thence to anode was nine feet. The
corresponding return line loop was longer since it carried back to the recirculating drum
and totalled 8.53 meters.
When the concentrated solution contained 50,000 ppm total cyanide, it had a
conductivity of 280,000>umhos/cm or a specific resistance of 3.6 ohm-cm,. The current
that would pass through these two routes is calculated as follows.
22
-------
Current in feed line: Ip = E/Rp
RF - f> x JF
AF
= 3.6 x 9' x 12"/ft x 2.54 cm/in
TTx 0.32
= 3520 ohms
IF = E/RF
= 125 volts/3520 ohms
= 0.036 amperes
Current in discharge line; IQ = E/Rp
RD = P * JjL
AD
= 3.6x28x 12x2.54
TTx 0.6*
= 2720 ohms
ID = E/RD
= 125 volts/2720 ohms
= 0.046 amperes
Total current through fhese routes;
IF + ID = 0.036 amperes + 0.046 amperes
= 0.082 amperes
This is a very minor portion of the current passing through the first stage. No
other possible routes of current loss could be found.
The same loops did not exist in the second stage because the solution supplied to
the cathode chamber was always the dilute stream (second rinse) while that fed to the
anode chamber was always the concentrated stream (first rinse).
23
-------
Thus, it did not seem plausible that poor performance of the plant installation was
due to failure of the current to pass through the stacks.
Poor Flow Distribution Among Chambers
It is well known that ff the flow in one or more of the depleted chambers is
appreciably reduced with respect to the others, the effects will be detrimental. The
slower flow leads to greater de-ionization in the chamber involved, causing, in turn,
lower amperage through the entire stack and less de-ionization of the other chambers.
The over-all result, therefore, is less effective depletion.
The chamber design and assembly at Keystone was identical to that used in
successful runs on sodium copper cyanide in the pilot plant, so that it was difficult to
believe that flow distribution had become a problem between the two series of trials.
Nevertheless, a means of shedding some light on the problem was afforded by the
availability of the Asahi Glass Co. stacks. These cells had an altogether different type
of distribution system and one which had been tested and found satisfactory over many
years. Results obtained with the Asahi cells are discussed in the following section.
Some Unexpected Effect of the Presence of Zinc
The laboratory experiments designed to investigate the significance of this
hypothesis are discussed later.
PERFORMANCE OF ASAHI STACKS ON BRASS PLATING SOLUTIONS
The data in Table 4 show an encouraging decrease in concentration of 1530 ppm
CN/hr in the first stage depleting stream during the first day's run. However, the
depleting stream in the second stage increased 66.6 ppm CN/hr during that run, as did
both dilute streams for the next two days.
The rather high A ppm CN/hr values for the second stage dilute stream are
probably due to the correspondingly high concentration of the concentrating stream in
that stage; it had not been possible to start the first rinse solution at as low a
concentration as desired at the beginning of this trial.
Except for the short-lived (and as yet unexplained) good performance of the first
stage during Run 1, it appeared that the Asahi cells were showing the same inefficiency
in treating brass solutions, as compared to sodium copper cyanide solutions, as had been
observed for the Hydronics cells.
The operation of the two Asahi stacks in parallel on the first rinse alone is
summarized in Table 5. As stated before, this arrangement ought to have improved
depletion of the rinse because the effective membrane area was doubled without a
24
-------
reduction in voltage per unit cell. The data for two runs shows there was no improvement.
The respective increases in concentration were 138 and 389 ppm CN/hr.
With these results, one must conclude that either poor distribution of flow through
the chambers is not a controlling factor in our observations on brass rinse treatment or
that carefully designed cells are at present incapable of giving good enough uniformity
of flow distribution. If the latter is true, then electrodialysis is of questionable
technical feasibility. Success in other applications has long shown that this is no!1 so.
Elimination of the aforementioned hypotheses concerning inefficient electrodialysis
of brass solutions gave added impetus to exploration of the effect that zinc might have.
LABORATORY INVESTIGATIONS
Analysis of An ion Permeable Membrane from Brass E .D . Cell
The most obvious feature of the analytical data presented in Table 7 is that, while
the amount of copper in brass plating solution is three times that of zinc, quite the
reverse is true on the membrane where there is 1.5 times as much zinc as copper. The
membrane evidently has a strong selectivity for zinc cyanide complex in this mixture or
is a repository for a zinc oxide precipitate.
It may be pertinent to note that the sum of the number of milligram atomic weights
of copper and zinc per gram of membrane is close to the ion exchange capacity of the
membrane.
No. of mg at wts Cu/g membrane 0.284
No. of mg at wts Zn/g membrane 0.414
O98
Ion exchange cap. MA-3475, meq/g 0.743
This suggests that each complex metal cyanide anion may be attached by only
about one negative charge per quaternary group on the membrane, leaving the remaining
negative charges free to act as a "pathway" for migration of cations in the counter-
direction through the membrane. The selectivity of the membrane for anion transport
could thus be in large part destroyed and the entire electrodialysis process frustrated.
De Korosy and co-workers* found that ferrocyanide ion converted an anion
permeable membrane into a cation permeable membrane, and that precipitates of certain
metal oxides on membranes had similar effects. The writer believes that a phenomenon
of this kind is at work in the electrodialysis of brass.
*De Korosy, F, et. al. J . Phys. Chem. 71, 3706 (1967); Desalination. 1970, 8 (2)
195-220 (C.A. 74, 25380 q.)
25
-------
TABLE 7. CONCENTRATION OF COPPER AND ZINC IN ANION MEMBRANE
MA-3475 AFTER USE IN BRASS ELECTRODIALYSIS:
COMPARISON WITH BRASS SOLUTION
Item Copper 2Iinc
In Membrane:
Weight % 1.8 2.7
Weight ratio 1.0 1.5
Number of mg atomic weights
per g of membrane 0.284 0.414
Ratio of number of mg atomic
weights 1.0 1.46
In Brass Plating Solution:
Ounces/gallon 4.96 1.67
Weight ratio 2.97 1.0
Ratio of number of gm
atomic weights 3.06 1.0
26
-------
Membrane Resistance in Copper Cyanide and in Brass Solutions
The plots of amperes vs. volts for solutions and for cells containing membranes are
shown in Figure 4. Note that the data for the solutions alone is nearly identical. If one
observes the proximity of the circles to the crosses and x's, it is seen that a cation
permeable membrane equilibrated with a brass solution adds very little to the resistance
of the system used in this experiment. (The same was true of a cation permeable
membrane equilibrated with sodium copper cyanide solution). This is not surprising
because the ion exchange capacity of the cation membrane on a volume basis is fairly
high and conductivity in these solutions is primarily due to the rapidly moving sodium
ion.
On the other hand, the interposition of an anion membrane (equilibrated with
sodium copper cyanide) between electrodes in a sodium copper cyanide solution raises
the resistance considerably. An even greater increase in resistance is seen as a result of
placing a brass-solution-equilibrated anion membrane between electrodes in the brass
solution.
These experiments were felt to show that there was indeed a difference in behavior
of anion membranes in copper cyanide and in copper, zinc cyanide solutions, although
they did not in themselves explain the less efficient electrodialysis of brass solutions.
Small, Three-Chamber Cell Tests of Electrodialysis
1 These experiments were another attempt at a simplified demonstration of some
difference between copper cyanide and brass solutions. In addition, sodium zinc cyanide
solutions (without copper) were included in the work. The solutions studied are described
in Table 8.
The course of typical electrodialysis runs is shown in Figure 5 where, at a constant
12 volts applied potential, the change of current with elapsed time is depicted.
In the case of both sodium copper cyanide and sodium zinc cyanide, there was a
short initial rise in amperage, followed by a fairly rapid drop and a final leveling off.
The concentrated brass solution showed an initial drop, followed by a long, gradual
increase in current and then an almost symmetrical decrease; only after 110 minutes had
the amperes decreased to the level (0.2 a) that the sodium copper cyanide and sodium
zinc cyanide solutions had reached in 36 to 42 minutes. The diluted brass solution
displayed no initial current rise, but only a gradual decline, reaching 0.29 a in 61
minutes.
The slow rate of current fall in the case of the brass solutions indicates slow
depletion of the center chamber, and this is shown quantitatively in the analyses
recorded in Table 9. The column headed A ppm/hr shows that the rate of concentration
decline with time in the brass solution is only about one-quarter to one-half that in the
27
-------
CD
s
o
CO CM
S3iJ3dWV
I/I
0)
CL
E
U
O
(U
5
.S*
u.
.28
-------
0)
(U
fe
Q.
.2
1
u
V
V
U
I
o
V
V
o
(U
o>
-------
to
Z
g
to t
to "
$8
to
Q to
22
|_ CC
U Q
az
LU <;
nl u
UZ
os N
LU
CO -
< Q£
LU
I °r
UO
LLjU
LU U_
£0
I to
y
oo to
ni r*
_J LU
eo h-
< U
t- <
0£
<
"32
°i
^ E^
'£ _c ^
r u<
" u. 0
^ OJ _C
' * E
C C fc
S^
^
*° c
V 3
P
^^
O
"c
(U
u
c
o
u
o
.2
"o
0)
t3 "5
N?*
LLJ j^«
O - '
0 "
'£ -Q
'» E
O O
D- f*"
E U
o
U
:§
.£
u_
0
^r
I
v>
"5
c
»
^E
8
lO
CM
U
*
4:
u
_0>
^1 1
U
i_
-^>
0)
^
O
X?
*~~
Z
U
o
Z
00
CM
Z
y
N
0
Z
u
u
in
"c
(1)
c
8.
E
o
U
Z
u
1
u
1
Z
CO
' CO
o
Z
c
D
7
CO
o
8 8
r~
8 8
^ ^^
^ oo
tx o
0 0
*~ *~
*
^ 1
0
lO If)
0 0
Z
u
Z ^
U N
i i
C D
N U
i i
f f
"- CM
CO CO
c^ 7>
CO CO
o o
0
CO
*~
1
~-
0
~
N,
IX
CM
>O
Z
U
1
c
N
i
D
U
1
Z
o
CM
CO
CO
O
O
o
N
O
's
(U
I
13
(U
.2
^
J?
O)
Z D
o z:
a! Q
O
(U
(U
V
* *.
30
-------
TABLE 9. THREE-CHAMBER CELL E.D. CONCENTRATION CHANGES
Item
133-Cu:
Cond.
Cu
No
CN
PH
131-Zn:
Cond.
Zn
No
CN
pH
132-Cu,Zn:
Cond.
Cu
Zn
No
CN
pH
Start
5900
1700 ma/1
1200
3050
10.3
4900
1100
1050
2450
11.0
8000
840
500
1700
2850
11.1
Finish
460
92
85
150
9.8
600
100
105
450
10.3
1100
70
50
180
350
9.5
Finish
Start x 100 (%)
7.5
5.6
7.1
4.3
"
12.2
9.1
10.0
18.4
"""
13.75
8.3
10.0
10.6
12.3
-
A ppm
-
1608
1115
2900
"
-
1000
945
2000
-
770
450
1520
2500
-
A ppm
-
2300
1640
4150
"
-
1280
1210
2560
^
-
420
245
830
1360
-
31
-------
sodium copper cyanide solution, depending on which species of ion is compared.
Similarly, the calculations shown below for the following Table 10 indicate! a lower
efficiency of electrodialytic depletion of all species of ions from a brass solution as
compared to a sodium copper cyanide solution.
TABLE 10. EFFICIENCY OF ELECTRODIALYTIC DEPLETION
% Ampere Efficiency for Ion Depletion
Element Na-Cu-CN Solu. Brass Solu.
Na 47.4 38.2
Cu 47.4 13.4
Zn - 7.6
CN 104.5 53.1
The amperes vs. minutes curve for the sodium zinc cyanide solution is quite close
to that for the sodium copper cyanide solution, but the analytical data and ampere
efficiency calculations lie between the values for the copper solution and the
concentrated brass solution. The writer can offer no explanation for this.
In general, the writer believes that the laboratory experiments have shown that
there is enough difference between sodium copper cyanide and brass solutions to account
for the difference observed in the performance of the Hydronics and Asahi E .D. stacks
when operating on these two solutions. To determine exactly the mechanism by which
brass solutions cause poorer efficiency of electrodialysis is beyond the scope of this
project. We can only suggest that the effect must be due to a phenomenon like one or
more of those observed by De Korosy, mentioned above.
32
-------
Calculations^
Faraday
1 MillhFaraday
133-Cu
131-Zn
96,500 amp-sees
1,608 amp-min
1.608 amp-min transports 1 meq
amp-min = 34.28
milli-F = 21.35
mg/l Na 1115 x .2
mg/l Cu 1608 x .2
mg/l CN2900x.2
= 223.Omg
^ 22
10.1 meq Na
= 321.6 mg
-31.8
10.1 meq Cu*
= 580 mg
4-26
""SOmeqCN
AE (Na) =
AE (Cu) =
AE (CN) =
10.1
TT35
10.1
TL35
22.3
7O5
= 47.4%
= 47.4%
= 104.5%
amp-min - 34.56
m?l!i-F = 21.5
A mg/l Na 945 x .2
A mg/l Zn 1000 x .2
A mg/l CN2000x.2
AE (Na) = 8.6
= 8.6 meq Na
= 6.1 meq Zn *
= 15.4 meq CN
AE (Zn) = 6.1
= 189.0 -J- 22
= 200 f 32.7
= 400 f 26
= 40.0%
= 28.3 %
= 71.5 %
(continued)
* Assuming the migrating onions are principally Cu(CN)3 and Zn (CN)4
AE (CN) = 15.4
33
-------
Calculations (Continued)
132-Cu, Zn
amp-min = 58.21
m!lli-F = 36.2
A mg/l Na 1520 x .2
A mg/l Cu 770 x .2
A mg/l Zn 450 x .2
Amg/ICN2500x .2
AE (Na) = 13.8
3O
AE (Cu) = 4.85
AE (Zn) = 2.75
AE (CN) = 19.2
36.2
304
154
90
500
22 = 13.8 meq Na
31.8 = 4.85 meq Cu*
t.7 = 2.75 meq Zn*
- J1.8 = 4.85 meq Cu
32.7 = 2.75 meq Zn
-26 = 19.2 meq CN
= 38.2%
= 13.4%
io.*t 'O^^
= 7.60%^
= 53.1 %
AE(Cu&Zn) = 21.0%
* Assuming the migrating onions are principally Cu(CN)3 and Zn
34
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-77-161
2.
3. RECIPIENT'S ACCESSION NO.
4 TITLE AND SUBTITLE
Electrodialysis for Closed Loop Control of Cyanide
Rinse Waters
5. REPORT DATE
August 1977 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
George W. Bodamer
International Hydronics Corporation, Princeton, NJ. 08540
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Keystone Lamp Manufacturing Company
Slatington, Pennsylvania 18080
10. PROGRAM ELEMENT NO.
1BB610 (01-01-07A)
11. CONTRACT/GRANT NO.
S-803304
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Laboratory - Cin., OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Full scale demonstration of electrodialysis for closed loop treatment of brass plating
cyanide rinse waters was conducted in the Keystone Lamp Manufacturing plant at Slatington, Pa,
In treatment of actual rinse water, the system was only one-quarter as effective as anticipated.
Numerous attempts to improve the efficiency of the installation were unsuccessful and the work
was terminated. Laboratory studies indicated that the failure was caused by a reduction in the
permselectivity of the anion membranes. It is believed that an insoluble zinc compound or zinc
complex anion was retained on the membrane and reduced its permselectivity. To avoid future
failures, the membranes need to be laboratory tested on actual waste waters before a full-scale
demonstration. In this study the electrodialysis system was tested on sodium copper cyanide
solutions, whereas the actual rinse waters contained sodium copper zinc cyanide.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS c. COSATI Field/Group
Metal finishing
Cyanides
Electroplating
Waste Water
Waste treatment
Electrodialysis
Chemical wastes
Heavy metals
Water pollution control
13 B
13 DISTRIBUTION STATEMENT
Release to public
19. SECURITY CLASS (ThisReport}
Unclassified
21. NO. OF PAGES
43
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
35
*US GOVBMKNTPRINTING OFFICE. 1977- 720-117/2020
-------
DATE DUE
-------
-------
Q
<
d
U)
HI
UJ
U.
Q
Z
<
UJ
CD
t-
(0
O
a
>
u
z
a)
CD
4
Z
0
h-
U
UJ
H
0 m
a: n
a
_i <
< a
h UJ
Z
UJ
^r
i
Z
o
£
>
z
UJ
0)
D
fll
U*
4^
fl
a:
in
in
fl
U
.C .*
r o
3 O
0 DO
u_
(0
o
0)
a
(A
w/
O
z
LJJ ~
o 5
< E
^ 2 oo
O (D
LJJ a
DC
« co Q
ro E -c
c o O
if c ~
CD CD
C
Z W CO
UJ 0) ^
^ DC 'c
DC Q
> ^
LU O
O £
O >
co 0
tf) -i
- a.
UJ -c
0)
UJ
z
(/)
D
03
d
a:
0
LL
UJ
>
h
Z
D
I-
K
O
0.
Q.
O
IS
m w
a z
IJT-
V
O
"X - C C
0: CDvC
c oro
a:
"5 ?> -Q
to "5
« -S o (5
§ J> ~~
ill?
O -Q t fO
« * C S
J> a> o .
fc -c o, '
SJ P
.C
;n
S
8 5«
1
S >
5) -C
to
to
o la 5i
«8 ...SO
^ ^ -
2 O 3
------- |