WATER POLLUTION CONTROL RESEARCH SERIES • 12010 EIE 11/71
An Investigation of Techniques
for Removal of Cyanide from
Electroplating Wastes
U.S. ENVIRONMENTAL PROTECTION AGENCY
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Series describes
the results and progress in the control and abatement
of pollution in our Nation's waters. They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Environmental
Protection Agency through inhouse research and grants and
contracts with Federal, State, and local agencies, research
institutions, and industrial organizations.
Inquiries pertaining to Water Pollution Control Research
Reports should be directed to the Head, Publications Branch
(Water), Research Information Division, R&M, Environmental
Protection Agency, Washington, B.C. 20460.
-------
AN INVESTIGATION OF TECHNIQUES FOR REMOVAL
OF CYANIDE FROM ELECTROPLATING WASTES
Sponsored by
OFFICE OF RESEARCH AND MONITORING
ENVIRONMENTAL PROTECTION AGENCY
and
METAL FINISHERS' FOUNDATION
Prepared by
BATTELLE
Columbus Laboratories
Columbus, Ohio 43201
Industrial Pollution Control Section
Program #12010 EIE
Grant #WPRD 201-01-68
November 1971
-------
EPA Review Notice
This report has been reviewed by the Environmental Protection
Agency and approved for publication. Approval does not signify
that the contents necessarily reflect the views and policies of
the Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or
recommendation for use.
For sale liy the Sii|KTintciiik>nt of Documents, l.'.S. (iovernnicnt ITinting Ollicc, Washington. P.O. 2
-------
ABSTRACT
This report describes work which was conducted on the removal of
cyanide from plating rinse waters employing various treatment pro-
cesses. The study consisted of an initial phase in which information
was sought by questionnaire and by wastewater analyses on the type of
waste produced by smaller electroplating plants. Laboratory studies
were conducted on several nonconventional methods for treatment of
these wastewaters, including ion flotation, adsorption on activated
carbon, acidification volatilization, and solvent extraction. A
demonstration pilot-plant study also was conducted on the activated
carbon process employing actual rinse waters from a zinc cyanide
plating operation.
The results of the various phases of the study indicated that activated
carbon adsorption for cyanide removal may have practical application in
many small plating plants. When combined with pretreatment stages for
solids removal, effective elimination of heavy metals also can be
expected. Further development of the process was recommended in actual
plating plant installations.
This report was submitted by Battelle's Columbus Laboratories in
partial fulfillment of Grant Project No. 12010 EIE by the Industrial
Pollution Control Section, OR&M of the Environmental Protection Agency
to the Metal Finishers' Foundation.
Key Words:
Electroplating wastes
Cyanide
Waste treatment
Activated carbon.
111
-------
CONTENTS
Section Page
I CONCLUSIONS 1
II RECOMMENDATIONS 3
III INTRODUCTION 5
IV BACKGROUND INFORMATION 7
State of the Art of Metal-Finishing Waste
Treatment 7
Current Practices in Small Electroplating
Plants 7
V EXPERIMENTAL WORK 15
Phase 1: Preliminary Experimental Laboratory
and Bench-Scale Studies 15
General Scope of Investigation 15
Ion Flotation Studies 15
Liquid-Liquid Extraction Studies 21
Studies on Acidification and Volatilization
of Cyanide 29
Activated Carbon Adsorption Studies 31
Preliminary Evaluation of Treatment
Costs 60
Phase 2: Pilot-Scale Investigation 63
Equipment and Procedures 63
Preliminary Results 66
Results of Multiple Cycle Adsorption and
Regeneration Runs 71
Discussion of Pilot-Plant Runs 71
VI ECONOMIC EVALUATION OF THE CARBON SORPTION PROCESS . . 75
VII ACKNOWLEDGEMENTS 81
VIII APPENDIX: CONVENTIONAL METHODS FOR TREATMENT OF
CYANIDE WASTEWATERS 83
Complete Destruction of Cyanide 83
Conversion of Cyanides to Cyanates 85
Other Methods for Treating Cyanide Wastes .... 86
-------
FIGURES
No. Page
1 CALCULATED LEVELS OF CHROMIUM AND CYANIDE DISCHARGED BY 10
NAMF MEMBER PLANTS
2 WATER USAGE BY NAMF MEMBER PLANTS 11
3 SKETCH OF FLOTATION CELL 17
4 EXAMPLE OF TWO-STAGE EXTRACTION, SINGLE-STAGE STRIPPING 22
LIQUID-LIQUID EXTRACTION PROCESS
5 BENCH-SCALE MORRIS-TYPE CONTACTOR FOR CONTINUOUS LIQUID- 26
LIQUID EXTRACTION EXPERIMENTS
6 SPECULATIVE ACIDIFICATION-VOLATILIZATION PROCESS FOR CYANIDE 30
WASTES
7 BENCH-SCALE CONTINUOUS COLUMNS FOR ADSORPTION EXPERIMENTS 35
8 EFFECT OF STIRRING TIME ON CYANIDE EVOLUTION 38
9 EFFECT OF POWDERED ACTIVATED CARBON ON THE REMOVAL OF 41
CYANIDE FROM A SOLUTION CONTAINING CADMIUM CYANIDE
10 EFFECT OF POWDERED ACTIVATED CARBON ON THE REMOVAL OF 42
CYANIDE FROM A SOLUTION CONTAINING NICKEL CYANIDE
11 RATE OF COMPLEX CYANIDE FORMATION WITH VARIOUS METALS 44
12 EFFECT OF pH ON COMPLEX CYANIDE FORMATION WITH ADDITION OF 45
FERROUS SULFATE
13a AUTOMATIC BARREL PLATING LINE AT SUPERIOR PLATING COMPANY 64
13b COLLECTION POINT FOR CYANIDE RINSE WATER 64
13c PILOT-PLANT CARBON ADSORPTION SYSTEM 64
14 SYSTEM I: FLOW DIAGRAM OF BATCH OR MANUALLY OPERATED 76
CARBON PROCESS
15 SYSTEM II: FLOW DIAGRAM OF CONTINUOUS CARBON PROCESS 77
VI
-------
TABLES
No.
1 SUMMARY OF NAMF MEMBER PLANTS IN WHICH CHROMIUM AND/OR 8
CYANIDES ARE USED
2 SUMMARY OF RINSING PRACTICES IN NAMF MEMBER PLANTS 8
3 SUMMARY OF DISPOSAL METHODS USED BY NAMF MEMBER PLANTS 8
4 CONCENTRATIONS OF CHROMIUM AND CYANIDE IN VARIOUS WASTE 13
STREAMS FROM NAMF MEMBER PLANTS
5 TYPICAL CONCENTRATION OF HEAVY METALS IN COMBINED EFFLUENTS 13
FROM SEVERAL NAMF MEMBER PLANTS
6 TECHNIQUES EVALUATED EXPERIMENTALLY 15
7 EXPERIMENTAL DATA ON VARIOUS COLLECTORS FOR FLOTATION OF 19
CADMIUM CYANIDE AND NICKEL CYANIDE COMPLEXES
8 FLOTATION DATA ON IRON CYANIDE 'SOLUTIONS 20
9 SUMMARY OF BATCH EXTRACTION EXPERIMENTS ON CYANIDE SOLUTIONS 27
USING ALIQUAT 336
10 MIXER-SETTLER EXPERIMENTS ON CYANIDE SOLUTIONS USING ALIQUAT 28
336
11 REMOVAL OF HCN FROM COMPLEX CYANIDES BY ACIDIFICATION- 32
VOLATILIZATION AT PH 3
12 TYPES OF ACTIVATED CARBON BEING EVALUATED 34
13 EFFECT OF PH ON THE REMOVAL OF CYANIDE WITH ACTIVATED 36
CARBON
14 EFFECT OF CARBON CONTACT TIME ON CYANIDE REMOVAL 37
15 AMOUNTS OF CONTAMINANT REMOVED IN BATCH EXPERIMENTS USING 39
VARIOUS ACTIVATED CARBONS
16 EXPERIMENTAL DATA ON COMPLEX CYANIDE FORMATION WITH 47
ADDITIONS OF ZINC SULFATE FOLLOWED BY FERROUS SULFATE
17 EXPERIMENTAL DATA ON COMPLEX CYANIDE FORMATION WITH 47
ADDITIONS OF ZINC SULFATE FOLLOWED BY NICKEL SULFATE
VII
-------
TABLES (continued)
No. Page
18 EXPERIMENTAL DATA ON COMPLEX CYANIDE FORMATION WITH 47
ADDITIONS OF ZINC SULFATE FOLLOWED BY COPPER SULFATE
19 CONTINUOUS ADSORPTION DATA FOR THE COPPER CYANIDE SYSTEM 51
20 STRIPPING CONDITIONS AND RESULTS 52
21 LABORATORY STRIPPING OF COPPER AND CYANIDE FROM 54
ACTIVATED CARBON
22 CONDITIONS AND RESULTS OF MULTIPLE CYCLE ADSORPTION- 57
REGENERATION EXPERIMENTS
23 EFFICIENCY AND COSTS OF ION FLOTATION OF COMPLEX 60
CYANIDES
24 COMPARISON OF COSTS FOR SELECTED TREATMENT METHODS 62
25 PRELIMINARY PILOT-PLANT ADSORPTION DATA ON MIXED CYANIDE 67
RINSE WATERS
26 PRELIMINARY PILOT-PLANT ADSORPTION DATA ON THE COPPER 68
CYANIDE SYSTEM
27 PILOT-PLANT DATA ON STRIPPING OF CYANIDE COMPLEXES 70
28 RESULTS OF MULTIPLE CYCLE ADSORPTION AND REGENERATION 72
RUNS ON CONCENTRATED ZINC CYANIDE WATERS
29 CONDITIONS AND RESULTS OF REGENERATION FOR MULTIPLE 73
CYCLE EXPERIMENT
30 PURCHASED EQUIPMENT COSTS FOR EXAMPLE CARBON SYSTEMS 79
31 ESSENTIAL OPERATING COSTS FOR EXAMPLE CARBON SYSTEMS 80
viii
-------
SECTION I
CONCLUSIONS
(1) Although the use of waste treatment systems by small electro-
plating plants is not yet required in many locations, there is little
doubt that more and more stringent treatment of discharges to streams
and sewers will be required in the near future. Adequate technology
is available for treating these wastes to any required degree of
detoxification using conventional processes.
(2) On the basis of the results of this study, several nonconventional
processes are technically feasible for the treatment of cyanide waste-
waters, including activated carbon adsorption, acidification and
volatilization of hydrogen cyanide, and ion flotation. Several other
recently developed processes, such as electrolytic oxidation of cyanide
and chemical oxidation using peroxides, also appear to be practical for
use by the small plater.
(3) The carbon process developed in this study through laboratory
and pilot-plant operations utilized actual zinc cyanide plating rinse
waters obtained from a small plating company. The systems were oper-
ated with removal efficiencies greater than 99 percent for total cyanide.
Removal of copper which was added to the rinse water to effect adsorp-
tion also was complete. These removal efficiencies compare favorably
with other known methods for treating cyanide wastewaters. The system
provided for the complete regeneration and recovery of copper and the
elimination of cyanide from the wastewater. The cost of chemicals--a
major index of process operating costs — was estimated from the early
laboratory and pilot-scale work to be in the range of $0.20-0.40 per
pound of cyanide treated. With further development, it is considered
likely that the cost of chemicals can be reduced to or below $0.30 per
pound of cyanide, especially if the process is applied to cyanide waste
solutions which contain copper.
(4) Estimation of capital requirements and net operating costs were
prepared for both a manually controlled and for a semicontinuous carbon
sorption process for removing cyanides from cadmium and copper as well
as zinc plating rinse waters. These estimates were based on a waste-
water flow of 10 gallons per minute containing 300 ppm of cyanide:
Manually
Operated Semicontinuous
System System
Estimated Capital Required, $ 12,000 16,000
Estimated Cost of Chemicals, $/lb CN 0.30 0.30
Estimated Net Operating Costs, $/lb CN 0.21-0.87 0-0.82
-------
The range in net operating costs depends upon the amount of the credit
that can be taken for recovery of copper as copper sulfate. The lower
estimate of net operating cost accounts for the recovery of copper from
the treatment of copper plating rinse waters.
(5) From a practical and economic standpoint, an extension of the
present treatment of zinc cyanide wastes using the activated carbon
adsorption process has promise of excelling the conventional methods,
when applied to the treatment of copper cyanide plating rinse waters
and copper cyanide strip solutions, in which case the process might be
used for copper recovery as well as cyanide treatment.
-------
SECTION II
RECOMMENDATIONS
This program has involved an evaluation of both conventional and non-
conventional processes for the treatment of cyanide rinse waters pro-
duced in small electroplating plants. The study has demonstrated
that several new approaches for treating cyanide wastewaters show
promise as simple and economic processes, especially for use by the
small plater. However, it was not within the scope of the program to
develop these processes to commercial utilization.
For accurate definition of the full potential of this study in the
cyanide waste treatment field, the work of this study should be extended
to a demonstration scale. The pilot-scale investigation has in fact
demonstrated the feasibility of an activated carbon process for treat-
ment of zinc cyanide rinse waters; several important questions need to
be answered:
(1) To what extent will the feasibility of the process be effected
for operation with other cyanide wastewaters peculiar to the plating
industry? Indications are that the process may be simplified and have
greater economic merit if applied to copper plating, brass plating, or
copper cyanide stripping waste solutions. The method in its present
development has been demonstrated only for zinc cyanide wastewaters.
(2) How can regeneration of the carbon after adsorption of copper
cyanides be made more effective and less time consuming? The present
method obviously must be improved to fit into a practical plant
operating schedule. Moreover, to what extent will improved regenera-
tion or alternative procedures enhance the practical and economic
feasibility of the process. Battelle has proposed to Metal Finishers'
Foundation a program to investigate other means of regenerating the
carbon, including electrolytic methods.
(3) What are the precise economic criteria such as labor, instru-
mentation, and analysis governing the economic evaluation of this
process, and how will it compare to conventional treatment after the
required further development?
It is recommended that these factors be established firmly by extending
the studies of the carbon adsorption process in full-scale demonstration
equipment. It also is recommended that a logical next step be an
extension of the method to treatment of copper plating rinse waters.
This application appears to be feasible and one which may result in
simplified procedures and improved economics. Very probably the devel-
opment would involve some minimum of additional laboratory study in
conjunction with the operation of the demonstration plant.
-------
SECTION III
INTRODUCTION
Many industries in the heavily industrialized areas of the United
States contribute substantially to the growing problem of water
pollution. In many cases the wastes generated by industrial sources
are carbonaceous and can be treated by biological methods or by
incineration techniques. In other cases, the wastes are largely in-
organic materials which are difficult to dispose of by these tech-
niques. Some of the most troublesome wastes are those generated in
the production of inorganic chemicals and in other metallurgical
operations. In the latter category, considerable attention has been
directed toward abatement of pollution from waste pickle liquor
from the steel industry, and from metal-finishing wastes.
Electroplating and metal-finishing waste streams can contribute
directly to stream pollution resulting from the content of toxic
and corrosive materials such as cyanide, acids, and metals. Indirect
effects also have been noted in the deleterious effect these com-
ponents exert on biological sewage treatment systems. Federal,
state, and municipal regulations fixing the allowable concentrations
of the waste components to these discharges already have been
established. The restrictions are fairly rigorous at present. There
is indication that they will, in many places, be made more rigorous
in the future. Enforcement of regulations may be expected to become
increasingly strict.
There is ample technology available for treating cyanide wastewaters
to any required degree of detoxification. The conventional treatment
methods have been developed primarily for large electroplating
plants where the relatively high cost of treatment can be absorbed
more readily. Economical methods have not been developed for the
treatment of wastes from the relatively small electroplating shops.
Currently many of these plants discharge their wastes into the city
sewers, and so depend upon the municipal sewage plant for the removal
of toxic materials.
More restrictive sewer ordinance and receiving water standards are
major factors that make development of economical techniques for treat-
ing the effluents from small electroplating facilities highly desirable,
Another factor is the potential recovery of valuable metals from these
waste streams.
-------
The Pollution Abatement Committee of the Metal Finishers' Foundation
being acutely aware of this problem, authorized Battelle to undertake
a study of existing hydrometallurgical techniques to determine their
applicability to the treatment of wastes from the smaller electroplating
plants. This study, as proposed to the Metal Finishers' Foundation,
was comprised of two phases:
Phase 1. Preliminary Experimental Study
Phase.2. Demonstration Plant Study.
In each of these phases, chromium wastes and cyanide wastes were given
separate study. This report presents the results, conclusions, and
recommendations of the study on cyanide wastes. A similar report dated
July 31, 1970, was submitted on chromium wastes.
The first phase of this work on cyanide containing wastes was, in effect,
a screening study, the overall objective of which was to pinpoint the
processes best adapted technically and economically to the treatment of
cyanide wastes typical of those generated by the smaller plater. In
this phase of the study, a thorough review of the present state of the
art in the treatment of cyanide waste was conducted; wastes from a
selected sample of smaller plating establishments were characterized for
volume and composition; and several nonconventional approaches available
to the smaller plater for treating his wastes were evaluated experimentally
As a result of this preliminary phase study, it was concluded that a
novel technique—activated carbon adsorption—showed promise of
excelling the conventional process as a practical and economic method for
treating cyanide wastes from the typical smaller shop.
In the second phase of the study, this carbon adsorption process was
investigated on the pilot-plant scale. The pilot plant was set up in an
actual plating plant (a member of the National Association of Metal
Finishers ) and operated on this plant's normal zinc-cyanide rinse waters.
This overall study on chromium and cyanide emphasized only these two
contaminants in the evaluation of experimental processes. This, of
course, overlooks other important contaminants, particularly heavy metals,
which would be carried over in the combined wastewater stream. Although
techniques were developed which might also provide effective removal of
heavy metals in addition to chromium and cyanide, the experimental
development of the techniques as an overall waste treatment approach
was beyond the scope of the current program.
-------
SECTION IV
BACKGROUND INFORMATION
State of the Art of Metal-Finishing Waste Treatment
One of the initial efforts in this program was devoted to a survey of
the open literature pertaining to waste treatment in the metal-finishing
industry. This state-of-the-art survey emphasized such aspects as the
nature of electroplating and metal-finishing wastes, current restrictions
on their disposal, and conventional methods available for treatment of
these wastes. The survey is presented in detail as a separate progress
report on this program, designated 12010 EIE, dated November 15, 1968.
It is not included within this report. However, for the purpose of
comparing waste-treatment processes, a description of the conventional
methods determined by this survey is included in the appendix. The
review of these methods is intended to provide facts for the guidance
of the smaller plater in the selection of a waste treatment process.
It should be pointed out, however, that these methods were developed
for use in electroplating plants having large volumes of wastewater.
The use of these methods for treating low volumes of wastewater would
certainly be feasible technically, but could be impractical or un-
economical for the smaller plater.
Current Practices_in_ Small Electroplating Plants
Production Characteristics
One of the major efforts during the initial phase of the program was
the accumulation of considerable data on the operating characteristics
of small electroplating plants. These data were obtained by surveying
member plants in NAMF via questionnaires. The survey emphasized pri-
marily the extent of chromium- or cyanide- plating operations and
certain aspects of the operations such as rinsing methods, disposal methods,
chemical and water usage, etc. Questionnaires were sent to 655 U. S.
members (foreign members were not included). In addition, 50 firms
were asked to submit samples of combined and segregated rinse waters
from their plant. Although only about 200 questionnaires were returned,
a large number of firms were not involved in chromium or cyanide plating.
The basic production characteristics of NAMF member plants based on this
survey are summarized in Tables 1, 2, and 3. Table 1 shows the number
of plants which have chromium and/or cyanide plating operations. These
-------
TABLE 1. SUMMARY OF NAMF MEMBER PLANTS IN WHICH
CHROMIUM AND/OR CYANIDES ARE USED
Plants with chromium only
Plants with cyanide only
Plants with both chromium and cyanides
Total
Number
of Plants
24
11
151
186
Percent
of Total
13
6
81
100
TABLE 2. SUMMARY OF RINSING PRACTICES IN NAMF MEMBER PLANTS
Separate chromium and cyanide rinse circuits
Combined chromium and cyanide rinse circuits
Practical to separate wastes
Impractical to separate wastes
Number
of Plants
84
94
71
100
Percent
of Total
47
53
42
58
TABLE 3. SUMMARY OF DISPOSAL METHODS USED BY
NAMF MEMBER PLANTS
Municipal sanitary sewer
Storm sewer
Land disposal
Lagooning
Natural stream, lake, etc.
Others
Total
Number
of Plants
143
24
5
3
10
6
191
Percent
of Total
75
12
3
2
5
3
100
-------
data indicate that a large majority of the plants (81 percent) do both
chromium and cyanide plating.
Table 2 shows the number of plants which currently have separate or
combined rinsing circuits and whether or not it would be practical to
separate these wastes. Table 3 indicates the current distribution
of plants that employ sanitary sewers, lagoons, natural water bodies,
etc., for disposal of wastewaters. These data indicate that most of
the plants use sanitary or storm sewers for disposal.
It is reasonable to presume that the response to the questionnaires con-
stitutes a valid sample and that the information obtained is applicable
to the association as a whole and possibly to the entire smaller shop
plating industry.
Waste Effluent Volumes and Compositions
The level of contaminants in wastewater from electroplating operations
and the total volume of these wastes are also important factors in the
evaluation of waste-treatment processes for small plating plants. Since
this information is almost unavailable in the literature, the waste
survey was used to provide data on these factors. NAMF member companies
were asked to report their annual consumption of plating chemicals con-
taining chromium and cyanide and the total annual water usage within
the plant. For the purpose of estimating the amount of these plating
chemicals which eventually appear in the wastewater, a loss factor of
80 percent was selected for both the chromium and cyanides (U. S.
Bureau of Mines Information Circular 8058).
Based on this assumption, calculations were made of the total chromium,
cyanide, and water discharged by each particular plant. A summarization
of these data for all plants surveyed is shown in Figures 1 and 2.
These graphs are important in establishing the approximate position
of a particular plant within the industry and can be used to provide
the following generalizations with regard to waste treatment methods.
They are of importance in the assessment of:
(1) The general levels of chromium and cyanides likely to be discharged
in the plant's wastewater
(2) The approximate annual cost for destruction of these contaminants
within various segments of the industry
(3) Areas where recovery of plating chemicals and water should be
considered.
-------
5 10 20 30 40 50 60 70 80 90 95 98 99
Plants with Less Than the Indicated Amount of Contaminant, %
FIGURE 1. CALCULATED LEVELS OF CHROMIUM AND CYANIDE
DISCHARGED BY NAMF MEMBER PLANTS
10
-------
5 10 20 30 40 50 60 70 80 90 95 98
Plants With Less Than the Indicated Amount of Water, %
99
FIGURE 2. WATER USAGE BY NAMF MEMBER PLANTS
11
-------
In addition to the data shown in Figures 1 and 2, considerable data
also were obtained on the actual composition of rinse waters generated
by NAMF member plants. Selected plants submitted samples of their waste
streams, and these samples were analyzed to determine concentrations
of cyanide, chromium, and other metals. Analytical data on chromium and
cyanide levels in various waste streams are summarized in Table 4.
Typical analyses for heavy metals in effluents from various plants are
shown in Table 5. The data on chromium and cyanide indicate that
cyanide concentrations are two to three times higher than chromium con-
centrations for combined effluents, which supports those data shown
previously in Figure 1. Generally, it can be stated that concentrations
will range from 10 to 100 ppm for either chromium or cyanide. Heavy
metal concentrations, by contrast, generally fall below 10 ppm with the
possible exceptions of copper and nickel. These elements were the major
heavy metals found besides chromium in the particular samples which
were analyzed.
12
-------
TABLE 4. CONCENTRATIONS OF CHROMIUM AMD CYANIDE IN VARIOUS
WASTE STREAMS FROM NAMF MEMBER PLANTS(a)
Total Chromium,
Type of Sample
Separate chromium stream
Separate cyanide stream
Combined effluent
Number of
Samples
9
7
28
ppmAb)
Range
15-70
—
0-49
Average
41
—
11
Total
Cyanide,
ppm(c)
Range
9-115
1-103
Average
72
28
(a) Based on actual analyses by atomic adsorption and standard chemical
techniques.
(b) Includes chromium contained in both the liquid and solid fractions
of the sample.
(c) Cyanides in the liquid fraction only.
TABLE 5. TYPICAL CONCENTRATION OF HEAVY METALS IN COMBINED
EFFLUENTS FROM SEVERAL NAMF MEMBER PLANTS
Concentration of Indicated Component, ppm
Cu
Plant
Plant
Plant
Plant
Plant
No.
No.
No.
No.
No.
4
7
12
15
21
31
2
.0
.0
10.0
36
7
.0
.0
Zn
5.
10.
3.
<0.
0.
0
0
0
5
2
Cd
<0
1
4
1
<0
.5
.0
.0
.0
.5
Fe
2
2
<2
4
2
.0
.0
.0
.0
.0
Ni
9.0
5.0
58.0
24.0
23.0
(a) Based on analyses of samples by atomic adsorption techniques.
13
-------
SECTION V
EXPERIMENTAL WORK
Phase L: Preliminary Experimental Laboratory and
Bench-Scale Studies
General Scope of Investigation
The initial phase of experimental work was limited to the investigation
of nonconventional methods that might be applicable to the treatment of
cyanide rinse waters; the conventional methods of treatment having been
thoroughly studied by numerous investigators for many years. At the out-
set of the program, the two most promising of the nonconventional
approaches appeared to be:
(1) Ion flotation
(2) Liquid-liquid extraction.
These two approaches and a third—activated carbon adsorption--received
major attention in the experimental program. During the course of the
study, other nonconventional approaches were suggested and examined briefly,
The nonconventional processes tested experimentally during this program
are listed in Table 6, and are described in the following sections of the
report.
TABLE 6. TECHNIQUES EVALUATED EXPERIMENTALLY
Free Cyanide Applicability, complex cyanides
Sodium Zinc Cadmium Iron Nickel Copper
Ion Flotation
Liquid-Liquid Extraction
Acidification Volatilization
Activated Carbon Adsorption
X
X
X
X
X
X
X
XXX
X
X X
X X
Ion Flotation Studies
General Description of Method. Of the various techniques studied during
this program, ion flotation probably is the most recently developed pro-
cess. The technique of separating ions from aqueous solutions by
15
-------
flotation has been recognized only for about 20 years. During this time,
however, it has received significant development effort and is now be-
coming known as one of the basic chemical engineering unit processes.
Ion flotation is basically a combination of conventional mineral flota-
tion and ion-exchange processes. In mineral flotation, for example,
finely divided solid particles are separated from a bulk solution by
attachment to small air bubbles introduced into the liquid. The bubbles
rise to the liquid surface, collapse, and form a froth which contains the
solid material in concentrated form. An identical procedure is used in
ion flotation to separate and concentrate ions from solution rather than
solid particles.
The mechanism of bubble attachment in both flotation techniques is
accomplished through the addition of a suitable collector. These collec-
tors usually are some type of organic chemical having surface-active
properties and are selective for only certain compounds. In addition
to surface activity, the collectors for ion flotation have an inorganic
group which ionizes in aqueous solutions and tends to make the collector
partially soluble. The exchange of ions between the collector and solu-
tion is the basis for separation of certain ions from the solution.
Because ion flotation is especially effective for extremely dilute solu-
tions, it was given initial emphasis during the experimental program.
Its simplicity and low equipment costs enhanced its attractiveness for
meeting the needs of the smaller plater.
Equipment and Procedure. The experimental apparatus for the study of
the ion flotation technique consisted of a specially designed glass flo-
tation cell (see Figure 3).
The glass cell was fabricated from a standard 500-cc graduated cylinder
by installing ports for injecting flotation collector solution via a
syringe and for withdrawing samples of the bulk solution. The cell con-
tained an air dispersion tube which was made of fritted glass with a
porosity of 25 to 50 microns. During flotation, the air was admitted at
a constant rate and was measured by a rotameter. The cell also was pro-
vided with an overflow tube which could be used to collect any foam pro-
duced during an experiment.
Generally, experiments were conducted with a 300-cc volume of solution to
be floated. The procedure used in conducting the majority of experiments
was as follows:
(1) The cell was filled with 300 cc of the particular solution under
study.
(2) The required flow of air was started (in most cases, a rate of 450
cc/minute was used).
16
-------
Air In
Syringe
Injection
Port
ooO
ooo
-,~orO
O 00000
0 0
0
0
O
O 0
o r
Air Out
OOOOO
°°ooo
oOoo
1
1 ____ 1
Foam Level
Solution Level
Sample Port
FIGURE 3. SKETCH OF FLOTATION CELL
17
-------
(3) A preselected volume of collector solution was injected via the
syringe port.
(4) After sufficient time had elapsed (generally 10 or 20 minutes),
samples of the treated solution were obtained and saved for chemical
analyses.
(5) Steps (3) and (4) were repeated until no further change occurred or
until a specified total quantity of collector had been added.
(6) Samples were analyzed and the percentage extraction calculated.
Results and Discussion. The initial experiments on ion flotation were
conducted with the objective of screening several possible collectors
which might be effective for removal of complex metal cyanides. The com-
plex cyanide ion, being negatively charged, required the addition of an
anionic collector in order to effect flotation. For the purpose of
selecting a suitable collector, a total of seven anionic collectors were
evaluated during this preliminary experimentation. The compounds selected
for study consisted of primary, tertiary, or quaternary-ammonium compounds
containing a long chain organic group of between 10 and 16 carbon atoms.
The preliminary series of experiments was conducted with initial solu-
tions containing 10 ppm cyanide and either 10.8 ppm cadmium or 5.64
ppm nickel as the complexing material. Results are shown in Table 7.
It was found that nickel cyanide complexes could be removed much more
effectively than cadmium cyanide complexes. With nickel cyanide solutions,
extractions over 90 percent were obtained with several flotation collec-
tors; the best extraction obtained with cadmium cyanide solutions was 57
percent. Although some cyanide was extracted from cadmium cyanide solu-
tions, it was not certain whether this occurred by flotation or because
of volatilization of HCN from the solution. Comparative tests (described
in a subsequent section of this report) indicated that simple aeration at
pH 6 can result in a 50 percent removal of cyanide as HCN in 1 hour from
cadmium solutions initially containing 10 to 100 ppm of cyanide; with
nickel solutions at comparable concentrations no removal of cyanide as HCN
was observed.
Flotation runs also were made on solutions containing iron cyanide com-
plexes formed by the addition of either ferrous or ferric sulfate to a
sodium cyanide solution. Typical data for several experiments are shown
in Table 8, Note that high extractions were obtained only when the solu-
tion was prepared by adding ferrous iron to a basic cyanide solution.
In this case the characteristic blue ferrocyanide complex or "Prussian
blue" was formed and this material appeared to float easily. In slightly
acid solution, the complex between ferrous or ferric iron and cyanide
did not form to an appreciable extent and low extractions were obtained.
18
-------
TABLE 7. EXPERIMENTAL DATA ON VARIOUS COLLECTORS FOR FLOTATION OF CADMIUM
CYANIDE AND NICKEL CYANIDE COMPLEXES
Cadmium Cyanide Runs
(1)
(2)
(3)
(4)
(5)
(6)
(7)
Collector Used
Dodecylamine HC1
Tetradecylamine HC1
Hexadecylamine HC1
N,N-Dimethyldodecylamine HC1
Decyltrimethylammonium bromide
Ethylhexadecyldimethylammonium bromide
Hexadecylpyridinium chloride
Amount
Used,
cc(a)
4.0
2.0
2.0
4.0
4.0
0.5
0.5
ppm CN
Initial (b)
10.
10.
10.
10.
10.
10.
10.
0
0
0
0
0
0
0
Indicated
Final
5.25
6.25
7.50
6.75
4.25
7.75
7.50
Percent
Removal
47
37
25
32
57
22
25
Nickel Cyanide Runs
Amount
Used,
cc
2.
3.
2.
0.
1.
(a)
0
0
0
5
0
ppm CN
Initial <^c) Final
10
10
10
10
10
.0 0.75
.0 0.75
.0 7.50
.0 0.50^d'
.0 i.oo«:
Indicated
Percent
Removal
93
93
25
1 95
> 90
(a) Solutions were made by dissolving collector in isopropanol and adjusting to 20 gpl,
were neutralized with HC1 to a pH of 7.
(b) Initial solutions also contained 10.8 ppm cadmium as cadmium chloride.
(c) Initial solutions also contained 5.64 ppm nickel as nickel sulfate.
(d) Excessive foaming occurred during these runs causing loss of some solution.
The amine collectors also
-------
TABLE 8. FLOTATION DATA ON IRON CYANIDE SOLUTIONS
Expt.
No.
16A
16B
16C
17A
17B '
17C
17D
18A
18B
18C
18D
Initial Solution
ppm
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
10.
CN
0
0
0
0
0
0
0
0
0
0
0
ppm Fe
3.
3.
3.
3.
3.
3.
3.
3.
5.
7.
3.
3.
58
58
58
58
58
58
58
58
37
16
58
58
ferrous
ferrous
ferric
ferrous
ferrous
ferrous
ferrous
ferrous
ferrous
ferrous
ferrous
ferric
pH During
pH(a) Flotation (a
Basic
Acid
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Basic
Acid
4.0
4.3
4.0
8.4
6.5
5.1
4.0
=4
=4
=4
=4
Final
Solution
Analyses , (b>
' ppm CN
1.
7.
6.
2.
1.
2.
2.
2.
1.
1.
9.
25
75
25
35
95
10
65
70
05
25
50
Apparent
' Percent
Extraction
87.
22.
37.
76.
80.
79.
73.
73.
89.
87.
5.
5
5
5
5
5
0
5
0
5
5
0
(a) Adjustments of pH were made by adding dilute HC1 or NaOH.
(b) Solutions were floated by adding 0.5 cc of tetradecylamine collector
and aerating for 10 minutes.
20
-------
In subsequent experiments, attempts were made to obtain more efficient
removal of cyanide from solutions containing either a nickel cyanide
complex or an iron cyanide complex. Variations were made in flotation
conditions, such as concentration, pH, and type of collector. To expe-
dite the evaluation of these experiments, results were based primarily
on visual observation of the extent of removal of the colored cyanide
complexes. Selected samples showing high color removal then were analyzed
for residual cyanide.
In general, the results of these experiments did not show an increase in
the level of cyanide removed over previous experiments. The extent of
cyanide removal appeared to be related more to variables affecting forma-
tion of the complexes than to variables affecting flotation. The forma-
tion of the blue ferrocyanide complex, particularly, was affected by
concentration in addition to pH. At concentrations of 100 ppm cyanide,
this material tended to flocculate more readily and could not be floated
effectively with any collectors. At 10 ppm cyanide, the complex appeared
more colloidal and high removals were achieved. Although these studies
did not develop an optimum set of conditions for the removal of cyanides,
the flotation method showed considerable promise and was considered as
an alternative process for further investigation during the remainder of
the program.
Liquid-Liquid Extraction Studies
General Description of Method. The liquid-liquid extraction process as
it might be applied to metal-finishing wastes is shown in Figure 4. The
waste or aqueous phase follows the solid line path through the system
into the first-stage mixer (Item 2 in the diagram), then to the first-stage
settler (4). Here the phases separate and the aqueous phase now partially
depleted of its contaminant passes through the second-stage mixer and
settler (5 and 3) where virtually all the remaining contaminant is removed.
The organic phase containing the extracting compound passes through the
two-stage extraction system countercurrent to the aqueous phase along the
path indicated by the dashed lines through the two stages of mixing and
settling (5, 3, 2, 4, and 8).
As it issues from the settler (4) it is sent to a regeneration stage (8
and 9) where it is stripped of its load of contaminant and restored to
its original composition. The regenerated organic is then recycled back
to the extraction stages (6). In the stripping or regeneration section,
the stripping agent is constantly recirculated through the mixer-settler
(8 and 10). Periodically enough of the stripping solution now loaded with
the contaminant is removed from the circuit for subsequent treatment (11).
The process outlined in Figure 4 is merely an example. Numerous variations
in flow patterns, number of stages, etc., are possible.
Equipment for liquid-liquid extraction too can be varied. The equipment
implied in Figure 4 is the so-called horizontal type, mechanically agitated
21
-------
„ ,
Make-up
Organic Exbroctant
To Compensate for
Losses
(1)
Waste Feed
(Aqueous)
N5
Partially
Extracted
Aqueous
Phase to
Second St
Stage
Extraction.
Make-up
Stripping
Solution
(Aciueous)
_„ , „ ,
1
1' *
,,, j 1 1 Light
! 'rtsX^ 1 Organic
|>s7'*N Phase
do \ (3)
Mixed Heavier
Phases | kv,.n,,a
Phase
--
^ Mixer Settler s
Second-Stage Extraction
i
\
r
„ j
~* (2) 1 -^1
1 — »
do
Mixed
Phases
V Mixer
N
Lig
Org£
Pha
(t
\\ea\
Aque
Ph«
1
Settl
First-Stage Extractic
?
(9)
..:
Light
Organic
Phase
(10V
Heavier
Aqueous
Phase
Stripping or Regeneration
Fully or Adequately
Extracted Aqueous
Phase
"Strip" Solution-^
Containing substantially
all, or most of, the
material extracted
from the waste feed
To Disposal (7)
To Chemical Destruction (11)
or Recovery
FIGURE 4. EXAMPLE OF TWO-STAGE EXTRACTION, SINGLE-STAGE STRIPPING LIQUID-LIQUID
EXTRACTION PROCESS
-------
mixer-settler, in which in each stage the phases are intimately dispersed
in one vessel by agitation and then permitted to flow by gravity to an-
other settling vessel for phase separation. Each stage, therefore, re-
quires two separate vessels. Other types of mechanical mixer-settlers
may be used. These include the vertical type in which the stages are
superimposed in a single vertical shaft with a central axle for driving
the agitators in each mixing section and baffled tanks similar in
nature to the apparatus used in the laboratory. There are several other
types of liquid-liquid extraction equipment which probably also would be
applicable to the treatment of metal-finishing waste. These include
(1) Various types of columns in which no agitation is employed, such
as perforated plate columns, bubble cap columns, packed columns, and
spray columns. This apparatus might possibly be applicable to the treat-
ment of metal-finishing wastes.
(2) Mechanically agitated columns, such as the so-called pulsed column.
These would be applicable to the treatment of metal-finishing wastes
also, but would be more costly than spray columns.
(3) Centrifuge types in the contacting or mixing. Settling is greatly
accelerated making possible high treatment rates and consequently re-
duced equipment sizes. These types generally are expensive, but require
less floor space than the other types.
Chemistry of Liquid-Liquid Extraction^ The chemistry involved in liquid-
liquid extraction can probably be most clearly explained by an example.
For the extraction of cyanide, a number of organic compounds would be
satisfactory. The one investigated most thoroughly in the laboratory cam-
paign was the so-called Alamine 336, manufactured by General Mills. Another
which was investigated to a lesser extent, but which exhibited as much
promise, was Amberlite LAI, manufactured by Rohm and Haas. The former is
a tertiary amine, the latter a secondary amine.
The chemistry of cyanide removal by liquid-liquid extraction is not well
defined. One reason for this is that the cyanide radical may exist in
a number of species in waste solutions, either as free dissolved hydro-
cyanic acid, as the so-called free cyanide (e.g., sodium cyanide), or as
a number of metallic cyanide complexes with such metals as copper, nickel,
iron, zinc, and cadmium. Each of these species could be expected to react
somewhat differently with the organic extractants and to combine with them
in differing proportions.
A brief review of the literature on the chemistry of cyanide adsorption
by ion exchange processes, which is analogous to the liquid-liquid ex-
traction process using amines, indicates that the chemistry is not known
very precisely. It can be surmised that the reactions which would occur
between any of the species in which cyanide would exist in water solutions
would react conventionally according to such equations as:
23
-------
R.NC1 +
4
Quaternary
amine
chloride
compound
such as
NaCN ->
Free sodium
cyanide as
in waste
liquor
R, NCN +
4
Compound
of amine
with
cyanide
NaCl
Sodium
chloride
dissolved
in aqueous
phase
Aliquat 336
4R.NC1
4
Quaternary
amine
chloride
+ Na.Fe(CN).
4 6
Sodium
ferrocyanide
as might be
intentionally
formed in
waste
(R.N) Fe(CN), +
44 o
Compound of
amine with
ferrocyanide
ion
4NaCl
Sodium
chloride
The extent to which such reactions take place has not been determined.
There is evidence in the literature on the ion exchange absorption of
cyanide and cyanide complexes that the chemistry involved is not clear
cut and probably not precisely understood.
Requirements for Feasibility. For liquid-liquid extraction to be feasi-
ble in the treatment of metal-finishing wastes, these conditions would
have to be met:
(1) The extraction of cyanide from rinse waters should be virtually com-
plete. Although there may at present be a certain laxity in legislation
and enforcement, the day may come when every plater may be required to
satisfy limits as low as 1 ppm of cyanide. An acceptable liquid-liquid
extraction process would have to be capable of meeting such limits.
(2) Reagent recovery by stripping should be efficient.
(3) The stripping operation should produce a greatly concentrated solu-
tion of the contaminant either for recovery or chemical destruction.
(4) The treated effluent solution should be sufficiently free from
organic contaminants such as oil to satisfy local restrictions. Present
restrictions on oil are variable, ranging from zero to some nonspecific
quantity described in such terms as "none making the water unsuitable for
the use indicated". It is the very nature of liquid-liquid extraction
in which oily substances, such as kerosene, are intimately dispersed in
the waste, to introduce some oil into the water phase. The process
should operate, therefore, in such a manner that the "oily substance"
content of the discharged waste is within permissible limits. The control
of the oil content of the treated effluent might be important if the
water were reused, even in part, for rinsing, owing to the possible
effects on work quality.
24
-------
(5) The operation should be relatively simple and routine requiring a
minimum of time and attention by the plater:
(6) Costs, both capital and operating, should be reasonable.
Equipment and Procedure. Preliminary batch experiments were made in
separatory funnels, by the conventional single-stage procedure. In these
runs, measured quantities of the aqueous and organic phases adjusted to
the desired pH level were mixed. The mixture was permitted to stand for
20 minutes to effect phase separation, and the treated aqueous phase was
analyzed to determine its cyanide content.
Subsequent mixer-settler runs were made in the apparatus shown in Figure
5, which provided three stages of extraction, with the aqueous phase
being fed into the righthand mixing compartment. Agitation by laboratory
stirrers was provided in each stage. The treated aqueous solution was
overflowed from the lefthand settling compartment and the "loaded"
organic was aspirated by suction from the righthand settling compartment.
The runs were made at various flow rates, volumetric ratios of aqueous
to organic, and composition of aqueous and organic solutions.
Results and Discussion. Initially a series of screening experiments to
determine suitable reagents was made. These indicated that incomplete
extraction of cyanide occurred with primary, secondary, and tertiary
amines, as indicated by qualitative testing of treated solutions with
silver nitrate. A tertiary amine salt, Aliquat 336, manufactured by Rohm
and Haas, however, indicated some promise in these experiments, and was
selected for subsequent testing. The results of batch experiments using
this on various soulutions are summarized in Table 9. These results indi-
cated that Aliquat 336 has potential as an extractant for cyanide at high
pH values. The results also indicated that complexed cyanides were more
effectively extracted than the simple cyanide.
To determine the possibility that multiple-stage extraction might produce
an effluent below 1 ppm in cyanide, several experiments were made in the
mixer-settler. These runs were made on solutions prepared from reagent-
grade chemicals to simulate rinse waters that might be obtained from
copper and zinc cyanide plating lines. The conditions and results of
these mixer-settler experiments are shown in Table 10.
These results indicated that high extractions were possible on both zinc
and copper cyanide solutions at elevated pH. However, in neither case
was the cyanide consistently reduced to the desired 1 ppm level. In
Experiment 75-76 (zinc cyanide solution) which was run for 210 minutes,
extremely low cyanide values in the effluent showed as much as 5 ppm of
cyanide. A similar situation occurred in Experiment 82 (copper cyanide
solution). This erratic behavior indicates that some factor in the
laboratory system was not under proper control. While it is believed
25
-------
FIGURE 5. BENCH-SCALE MORRIS-TYPE CONTACTOR FOR
CONTINUOUS LIQUID-LIQUID EXTRACTION
EXPERIMENTS
26
-------
TABLE 9. SUMMARY OF BATCH EXTRACTION EXPERIMENTS ON CYANIDE
SOLUTIONS USING ALIQUAT 336
NJ
Aqueous Phase
Expt.
No.
48-1
4g_2(b)
49-2 (O
43-2(d)
43_3(d)
43-4(d)
43-5Cd)
47-l(e)
47-2(e)
47.4(6)
47-5
-------
Ni
CXI
TABLE 10. MIXER-SETTLER EXPERIMENTS ON CYANIDE SOLUTIONS
USING ALIQUAT 336
Feed Solution
Expt.
No.
81
75-76
82
Type
Zinc Cyanide (a)
Zinc Cyanide (a)
Copper Cyanide (b)
ppm
CN
100
100
100
PH
7
10
10
Feed
Rate,
ml/min
200
200
200
Duration
o f Run ,
min
210
210
210
Organic
Feed
Rate,
ml/min
20
20(0
20
End
Solution
ppm Extraction,
CN percent
9.7-13.0 87-92
-------
that cyanide probably could be reduced to the desired level of less than
1 ppm consistently, the laboratory work failed to demonstrate this.
The results of studies on the reagent regeneration by stripping also in-
dicated that this was a problem. In no case was adequate stripping of
cyanide from "loaded organic solution" satisfactorily achieved. It is
believed that a more intensive and systematic campaign would be necessary
to evaluate fully the feasibility of stripping cyanide or cyanide com-
plexes from the organic solutions.
Studies on Acidification and Volatilization of Cyanide
Acidification-aeration as a means for eliminating cyanide from metal-
finishing wastes once was practiced commercially. Reportedly, the pro-
cess is no longer being used. In the commercial process the waste was
acidified, heated, and then swept with large volumes of air to eliminate
the cyanide as hydrogen cyanide gas. The gas was discharged up a stack
and presumably was sufficiently dispersed so as to be harmless. The suit-
ability of such a process for the smaller plater likely to be located in
heavily populated areas and subject to local government restrictions seems
dubious. However, a variation of this process that might fit the require-
ments of smaller establishments was conceived and tested in a very pre-
liminary fashion during this program. The envisaged process is depicted
in Figure 6.
The process would consist of passing the slightly acidified and possibly
heated waste through a vacuum system. The exhaust from a mechanical
vacuum pump would be introduced into a sodium hydroxide absorbing solu-
tion which would readily absorb the cyanide gas and whatever CO was ex-
hausted by the pump. Alternately, the exhaust from the vacuum pump could
be directed to a flare which would destroy cyanide by combustion.
Periodically some or all of the loaded absorbent would be removed and
fresh sodium hydroxide solution added. The spent solution either could
be treated by a conventional destruction method or could be sent to a re-
covery step. Depending on the carbon dioxide content of the exhaust gases,
various amounts of sodium carbonate would be present in the absorbent
solution along with sodium cyanide. If too much were present for reuse
of the cyanide solution in plating, it might be treated with calcium
cyanide to precipitate calcium carbonate and to form sodium cyanide
according to the reactions:
Na CO + Ca(CN) -»• ZNaCN + CaCO .
Sodium Calcium Sodium Calcium
carbonate cyanide cyanide carbonate
precipitate
A process of this type, i:'. workable, would eliminate the shortcomings of
simple acidification aeration such as:
29
-------
Cyanide Rinse Water
Sulfuric acid
to adjust pH
to about 2
Heat (if required)
Steam, electric,
etc.
Vacuum System
CN, CO,
Mechanical
Vacuum Pump
Effluent to Disposal
with pH Adjustment by
Na2C03, etc., if
Necessary
Exhaust Line
10 Percent Sodium
Hydroxide Solution
to Absorb C02, CN,
etc.
Sodium
Rypochlorite
Ca(CN)2
Recovery of
Cyanide
V
Destruction
NaCN Solution
to Reuse -
CaCO-
Filtration or
Settling
/\
Disposal
FIGURE 6. SPECULATIVE ACIDIFICATION- VOLATILIZATION PROCESS FOR
CYANIDE WASTES
30
-------
(1) The discharge of cyanide to the atmosphere
(2) The use of large quantities of air which would make absorption and
hence recovery of cyanide difficult if not impossible.
The technical feasibility of this type of process was investigated on
a batch scale using sodium cyanide solutions and, in other experiments,
solutions containing complex metal cyanides. In these experiments, solu-
tions of cyanide were acidified to a pH of 3, transferred to a glass filter
flask, heated to about 100 F, and maintained at that temperature throughout
the run during which they were stirred mechanically. A vacuum of about
25 to 28 inches of mercury was applied. Samples were removed at various
intervals of time and analyzed for cyanide.
The data developed during this series of experiments are summarized in
Table 11. As shown the percentage of cyanide that was volatilized
varied considerably, depending on the time for evacuation, the concentra-
tion, and the type or form of cyanide in the solution. At a concentration
of about 50 ppm cyanide, good removals (greater than 95 percent) were ob-
tained after 30 minutes from solutions containing free cyanide (sodium
cyanide) or from solutions containing the weaker cyanide complexes (zinc
and cadmium cyanides). The percentage removals decreased slightly to
about 92 percent in 30 minutes for solutions containing 260 ppm cyanide
in these same forms. For solutions containing either copper or nickel,
however, much lower removals were observed, indicating that these metals
form stronger complexes with cyanide and would require a much longer
treatment time to reduce cyanide to an acceptable level.
On the basis of these experiments, it was concluded that the process is
technically feasible at pH = 3 for most of the major types of plating
rinse waters (zinc, cadmium, and free cyanide solutions). If copper or
nickel (and probably iron) is present, the method would be prohibitively
long and for all practical purposes not feasible for the smaller plater.
Activated Carbon Adsorption Studies
General Description of Method. During the search of the literature for
the state-of-the-art review, references were found in which the use of
activated carbon for the adsorption of chromium from solution was cited.
Activated carbon has been studied by various investigators for the ter-
tiary treatment of domestic wastewaters. It also has been used for the
adsorption of various materials from solution, including metal ions.
For these reasons it was concluded that adsorption on activated carbon
should be one of the techniques to be evaluated during this phase of the
program.
Modern theories hold that the adsorption of materials from solution by
activated carbon is accomplished by van der Waal or dispersion forces.
These forces exist among all molecules and atoms, whether or not they are
31
-------
TABLE 11. REMOVAL OF HCN FROM COMPLEX CYANIDES BY ACIDIFICATION-VOLATILIZATION AT pH 3
Type of
Solution(a)
Sodium Cyanide
(52 ppm CN)
Sodium Cyanide
(260 ppm CN)
Zinc Cyanide
(52 ppm CN)
Cadmium Cyanide
(52 ppm CN)
Cuprous Cyanide
(52 ppm CN)
Cupric Cyanide
(52 ppm CN)
Nickel Cyanide
(52 ppm CN)
Mixed Cyanides
(260 ppm CN)
Evacuation Time,
min
15
30
15
30
15
30
15
30
15
30
15
30
15
30
15
30
Cy an i de Remo ve d ,
rag
46.0
52.5
197.4
229.8
45.5
50.0
40.6
49.1
31.7
37.1
27.9
31.1
11.3
20.2
187.4
213.6
Residual Cyanide,
mg
0.2
18.5
1.0
2.6
10.2
11.4
26.8
20.6
Apparent Percent
Removal
87
99.6
80
93
89
98
79
95
67
78
66
73
24
43
80
91
(a) Initial solutions were made up to 1 liter and maintained at 100 F during evacuation.
-------
chemically combined, and may be considered analogous to the gravitational
force the earth exerts upon objects near it.
The ability of activated carbon to adsorb a given material depends largely
on its surface area. Each particle of activated carbon has a vast inter-
connecting network of many-sized pores, providing a very large surface
area for adsorption. Consequently, the pore structure of activated car-
bons is extremely important in determining their adsorptive properties.
Although the reasons are not completely understood, it is known that
other substances have an influence on the adsorptive properties of acti-
vated carbons. Oxygen combined with the carbon can be particularly
important and can increase its affinity for polar compounds and decrease
its affinity for nonpolar compounds. In some cases, the inorganic-ash
portion of the carbon can also influence the adsorption process. ^These
latter factors may play a major role in the adsorption of inorganic
materials.
Activated carbons are usually classified according to their physical
form (e.g., powdered or granular) and according to their use (e.g.,
water grade, decolorizing, liquid phase, or gas phase). Granular car-
bons are those materials which are over 150 mesh in particle size and
powdered carbons are those which are smaller in particle size.
Activated carbons are produced from various carbonaceous raw materials
(e.g., bituminous coal, nut shells, lignite, pulp-mill residue, and
wood). During this experimental program, many types and sizes of acti-
vated carbon were studied for the removal of chromium from waste
solutions.
Equipment and Procedure. In preliminary batch experiments on the adsorp-
tion of cyanide, various amounts of carbon were mixed with a synthetic
cyanide waste solution and the mixture was stirred for a predetermined
time at room temperature. The mixture then was filtered on glass filter
paper and the filtrate analyzed for total cyanide.
During the preliminary experiments, several sizes of activated "^bon as
well as those produced from various materials, were evaluated. The car-
bons studied are listed in Table 12.
In subsequent continuous-adsorption experiments a series of f0" Slass
columns arranged as shown in Figure 7, was employed. The columns were
on t cted of 2-lnch-diameter Pyrex glass tubing and -re approximately
30 inches in length. By pumping the solution through the four columns
in series, a total carbon bed depth of 10 feet was attainable The
granular carbon was supported on a 65-mesh stainless steeJ 8"<^ °scil-
Lting pumps were used to pump the solution down through the columns The
columns were constructed with a sampling port at the Attorn of each one
so that effluent samples could be obtained for analysis The type carbon
selected for the continuous experiments was Pittsburgh Activated Carbon
33
-------
TABLE 12. TYPES OF ACTIVATED CARBON BEING EVALUATED
Trade Name
Company
Source Material
Type and/or Size
Ul
Nuchar
ii
Pittsburgh
Absorbite
Cliffchar
II
West Va. Pulp & Paper
ditto
Pittsburgh Act. Carbon Co.
ditto
Barneby-Cheney
ditto
Royal Oak Charcoal Co.
ditto
Pulp mill residue
ditto
Bituminous coal
ditto
II
II
Nut shells
ditto
Wood charcoal
ditto
Granular (8 x 30) (WVL)
Pulverized
Pulverized (RC)
8 x 30 (SGL)
12 x 40 (CAL)
20 x 50 (OL)
Pulverized, nonactive (XB)
Granular, nonactive (BB)
Pulverized, low-active (YD)
Granular, low-active, 12x30 (PA)
Pulverized, high-active (XZ)
Granular, high-active, 10x50 (PC)
Pulverized, acid wash (JF)
Granular, acid wash, 10x50 (PK)
Pulverized
Granular, 10 x 20
-------
FIGURE 7. BENCH-SCALE CONTINUOUS COLUMNS FOR ADSORPTION EXPERIMENTS
35
-------
Company's Type OL which is a granular carbon with a particle size of 20
x 50 mesh. This selection was based on the overall results of prelimi-
nary batch experiments.
The procedure used in the continuous experiments was to pump a synthetic
waste solution containing varying amounts of cyanide through the column
at a given rate with periodic analysis of effluent samples. In most
experiments only one column was used so that the time of "breakthrough"
could be observed as well as the time when the carbon was completely
loaded. The major variables studied during these experiments were:
(1) pH cf the feed solution
(2) Concentration of the feed solution
(3) Rate of flow or residence time
(4) Ionic form of the element in solution (e.g., free cyanide, complexed
cyanide, or cyanide/chromium mixtures).
Results of Preliminary Batch Studies. The initial experimentation on the
removal of cyanides by adsorption on activated carbon was carried out on
solutions made from distilled water and sodium cyanide containing 100 ppm
of cyanide.
The first series of experiments was concerned with a study of the effect
of pH on cyanide removal. The pH value was varied by the addition of
either dilute sulfuric acid or sodium hydroxide solution prior to con-
tacting with the carbon. The conditions and results obtained from this
series of experiments are shown in Table 13. It can be seen that maximum
removal amounted to only about 50 percent at pH values of 8 or below.
TABLE 13. EFFECT OF pH ON THE REMOVAL OF
CYANIDE WITH ACTIVATED CARBON
pH Cyanide Removal, percent
12 9
10(a) 18
8 50
6 48
4 50
2 50
1 45
(a) No acid or caustic required to obtain pH 10
since this is normal value of 100 ppm cyanide.
Conditions: Aqua Nuchar A carbon; carbon concentration, 5 g/1;
contact time, 5 minutes; agitated by stirring in an open beaker
36
-------
Another series of experiments was run to determine the effect of contact
time on cyanide removal. Conditions and results are shown in Table 14.
The results showed that increasing the contact time from 5 to 40 minutes
increased the percentage of removal from 50 to 98. Longer contact time
did not increase the percentage of cyanide removal.
TABLE 14. EFFECT OF CARBON CONTACT TIME ON
CYANIDE REMOVAL
Contact Time,
minutes
Cyanide Removal,
percent
5
10
20
40
60
50
65
88
98
98
Conditions: Aqua Nuchar A Carbon; carbon concentration,
5 g/1; pH 2.0; agitated by stirring in an open beaker.
It was found that significant amounts of cyanide were being evolved as
HCN rather than being adsorbed at the low pH values. Several tests were
made to determine the relative amount of cyanide that was volatilized
and so not adsorbed. Sodium cyanide solutions containing 100 ppm of
cyanide were stirred for various time periods without carbon addition.
One sample was stirred at pH 2 and another sample at pH 10 (the normal
pH value of the sodium cyanide). The results are shown in Figure 8.
A large percentage of cyanide was evolved from the low pH solution even
after only 10 minutes of stirring while the solution at pH 10 did not lose
a significant amount of cyanide until after 10 minutes. This observed
loss at the high pH value may have been due to oxidation rather than
volatilization. However, at both pH levels the amount of cyanide increased
with increased stirring time. In view of these results, it was decided
that the preceding experiments probably did not show actual adsorption
characteristics and that an alternative procedure would be needed to pre-
vent cyanide losses by volatilization.
Subsequent experiments using an alternate procedure were conducted to
determine which of the various types of carbon (listed previously) were
more suitable for cyanide adsorption. For these experiments, the acti-
vated carbon and the solution were agitated by rolling in a sealed jar
to prevent oxidation of the cyanide or its loss by volatilization. The
mixtures were then filtered after the desired time had elapsed and the
filtrate analyzed. These experiments were run on pure sodium cyanide
solutions. The conditions and results are shown in Table 15. These
data indicate that the Pittsburgh carbon gave better overall per-
formance than the other types used. The percentage removal of cyanide
37
-------
100
(U
0
10
20 30 40
Elapsed Time, minutes
50
FIGURE 8. EFFECT OF STIRRING TIME ON CYANIDE EVOLUTION
38
-------
TABLE 15. AMOUNTS OF CONTAMINANT REMOVED IN BATCH EXPERIMENTS
USING VARIOUS ACTIVATED CARBONS(a)
Type of Activated Carbon
Nuchar, pulverized
(pulp mill residue)
Nuchar, granular
(pulp mill residue)
Pittsburgh, pulverized
(bituminous coal)
Pittsburgh, granular
(bituminous coal)
Absorbite, pulverized
(nut shell)
Absorbite, granular
(nut shell)
Cliffchar, pulverized
(wood charcoal)
Cliffchar, granular
(wood charcoal)
Cyanide (b)
35
38
33
35
25
20
30
27
Removal, percent
Mixture
Chromium
41
48
60
60
6
3
—
6
(c)
Cyanide
53
45
65
58
58
25
38
40
(a) Conditions: Carbon concentration: 10 g/1
Contact time: 10 minutes
Agitation: Rolling in sealed jar.
(b) Starting solution contained 100 ppm cyanide added as sodium cyanide;
pH 10.
(c) Starting solution contained 10 ppm chromium and 10 ppm cyanide added
as potassium dichromate and sodium cyanide; pH 5.9.
39
-------
from the mixed solution was higher, in all cases, than that from the
pure cyanide solution. Because different levels of cyanide were in-
volved (10 and 100 ppm, respectively) and because two different pH levels
were employed (5.9 and 10), it cannot be stated positively that the mixed
waste responded better. It does suggest the possibility that the presence
of chromium may have assisted in removing some cyanide.
Several series of experiments also were conducted with complex cyanides
of cadmium and nickel. In the first series, the solution contained 36
ppm cadmium and 64 ppm cyanide, and the activated carbon concentration
was varied from 5 to 40 grams per liter. Contact time was held constant
at 5 minutes at pH = 9.6. The results of these experiments are shown in
Figure 9. They indicate that carbon concentrations of 25 grams per liter
and higher effected a maximum of about 90 percent cyanide removal.
In another series of experiments, a solution containing 56.4 ppm nickel
and 100 ppm cyanide was treated under conditions similar to those used
for cadmium cyanide, except that the pH was held at 7.0. The results of
this series shown in Figure 10 indicate that 30 grams per liter of carbon
gave almost complete removal of cyanide.
On the basis of the results obtained in these preliminary batch experi-
ments, only tentative conclusions could be drawn:
(1) The adsorption of uncomplexed cyanide (free cyanide) by activated
carbon probably does not occur to an appreciable extent, although some
adsorption apparently takes place at the lower pH levels. The magnitude
of the adsorption is difficult to pinpoint, owing to uncertainty about
the extent that volatilization, oxidation, or other unknown processes
may be occurring.
(2) The complexing of the cyanide with metals such as nickel and cadmium
(and iron, zinc, and copper) enhances the adsorptive power. The experi-
mental results suggested that the nickel complex adsorbs completely and
rapidly and the cadmium complex somewhat less completely and only under
specific conditions of pH and carbon concentration. More laboratory work
would be required to determine whether a practical process based on
cyanide cotnplexation-adsorption is feasible.
(3) Certain carbons appear to be superior to the other carbons tested.
Studies on Formation of Complex Metal Cyanides. On the basis of the pre-
liminary studies just described, it was concluded that the development of
a workable process for cyanide removal probably depended on the use of
metal cyanj.de complexes rather than free cyanides. It also was evident
that more basic information was needed above that appearing in the litera-
ture on the formation, behavior, and properties of metal cyanide complexes
as produced in or from electroplating rinse waters. A brief study there-
fore was undertaken to provide this information.
40
-------
10Q
c
01
o
1-1
o
6
OJ
PS
T3
•H
C
efl
5 10 15 20 25 30 35
Concentration of Powdered Activated
Carbon, grams/liter
40
FIGURE 9. EFFECT OF POWDERED ACTIVATED CARBON ON THE REMOVAL
OF CYANIDE FROM A SOLUTION CONTAINING CADMIUM CYANIDE
41
-------
100
5 10 15 20 25 30 35
Concentration of Powdered Activated
Carbon, grams/liter
40
FIGURE 10. EFFECT OF POWDERED ACTIVATED CARBON ON THE REMOVAL
OF CYANIDE FROM A SOLUTION CONTAINING NICKEL CYANIDE
42
-------
The emphasis in this work was to develop conditions and procedures for
the formation of specific complex cyanides (e.g., iron, zinc, copper,
nickel) which then might be effectively removed by one or more of the
removal techniques under consideration such as carbon adsorption or
ion flotation.
The experiments in this study were conducted with the aid of an Orion
Model 401 specific ion analyzer using electrodes for measuring free
cyanide in aqueous solutions. This instrument indicates the free cyanide
concentration in a range of 0.01 to 100 ppm but does not apparently in-
dicate cyanide which is bound in a metal complex or cyanide present as
HCN. This characteristic provided a convenient method for studying the
rate of complex formation under various experimental conditions.
Typical data which were obtained initially on the complexing of cyanide
with four different metals are shown in Figures 11 and 12. The data in
Figure 11 illustrate the rate of complexing with each metal in 0.1 molar
NaOH solution (pH = 12). As shown, the free cyanide concentration de-
creased with time after the addition of stoichiometric amounts of the
various metals added as sulfates to sodium cyanide solutions. For zinc,
copper, and nickel, a stoichiometric ratio of 4 parts cyanide to 1 part
metal was selected. For iron, a ratio of 6:1 was selected. For the runs
with iron, zinc, and copper, the amounts of metal were doubled after 15
minutes had elapsed during complexing. In the case of nickel, only a 20
percent excess of this metal was subsequently added after 30 minutes had
elapsed.
The results shown in Figure 11 illustrate the marked difference observed
in complexing ability between the individual metals. Nickel, for example,
exhibited a rapid and complete reaction in that free cyanide was reduced
to less than 1 ppm in 3 minutes and to less than 0.2 ppm in 40 minutes.
Copper was the second most effective reagent and reduced free cyanide to
about 1.5 ppm in 60 minutes. By contrast, the reactions with iron and
zinc were exceedingly slow; only about 50 percent of the free cyanide
was complexed in 60 minutes.
The results obtained with zinc, copper, and nickel were expected since
these metals form progressively more stable complexes with cyanide and
thus should show more complete reaction proceeding from zinc to copper to
nickel. However, since iron is known to form very stable complexes simi-
lar to that of nickel, the very slow and incomplete reaction observed
above was somewhat unexpected. In order to further investigate this
aspect, a subsequent series of runs was made with iron at various pH
levels. In these experiments, solutions containing 100 ppm cyanide were
maintained at constant pH values (7.0, 8.0, 9.0, 10.0, and 11.0) and in-
cremental amounts of ferrous sulfate corresponding to 0.4 to 2.0 times
the theoretical amount were added. The solutions were agitated for 10
minutes after each addition of ferrous sulfate. Samples were then taken
and diluted by a factor of 10 with 0.1 molar NaOH to allow direct measure-
ment by the specific ion meter. Figure 12 presents the data which were
obtained from this series of experiments. As shown, changes in pH produced
43
-------
Zinc
Copper
V
_L
20 30 40
Time, minutes
50
60
FIGURE 11. RATE OF COMPLEX CYANIDE FORMATION WITH VARIOUS METALS
44
-------
2
U
6
{X
p.
c
o
•H
5 1.0
0)
u
c
o
u
-------
a marked effect on the extent of the complexing reaction. Also evident
is the effect of excess ferrous sulfate in increasing the amount of
complex formed. The data explain the very slow reaction at pH 12 ob-
served initially and indicate that a close control of pH and concentra-
tion is important to provide maximum formation of complex iron cyanides.
Although the above data indicated that optimum formation of iron cyanides
could be effected under certain conditions, the results were found
applicable only to simple cyanide solutions containing sodium cyanide.
During subsequent experiments, the investigation was extended to include
more representative waste cyanide solutions which would contain cyanide
already partially complexed with metals such as zinc or cadmium. It was
found that the presence of these metals, which form mild complexes with
cyanide, greatly interfered in some cases with the formation of the stronger^
complexes such as those of nickel, copper, and iron. Typical data showing
this effect are given in Tables 16, 17, and 18.
In Table 16, data are shown for the residual free cyanide concentration
remaining at various time intervals after addition of zinc sulfate,
followed by the addition of ferrous sulfate. The amount of zinc added
was equal to the theoretical amount (4 parts cyanide to 1 part zinc)
for complete formation of the complex. Iron additions corresponded to 5
times the theoretical amount for complete complex formation. In Tables
17 and 18 comparative data are shown for additions of zinc sulfate followed
by either copper or nickel sulfate except that the amount of nickel added
was only 1.1 times the theoretical amount and the amount of copper added
was 1.5 times the theoretical amount.
The three sets of data suggest a clearcut difference in the complexing
ability of iron versus nickel and copper. With iron--even at 5 times the
theoretical amount — complex formation was greatly retarded by the previous
addition of zinc to the solution. The reaction with nickel--at 1.1 times
the theoretical amount—was only slightly affected by zinc. This occurred
at all pH values investigated with a pH of 10 being optimum for greatest
formation of the cyanide complex of either metal. Similar data also were
obtained when cadmium was used in place of zinc as the interfering metal.
On the basis of the results of these experiments, it was concluded that:
(1) Nickel and copper probably are the best complexing agents for free
cyanide and also for cyanide that has been previously complexed with zinc
or cadmium and might be incorporated into a removal process using either
ion flotation or carbon adsorption. Economics considerations suggest
that wastes generated within the plant would be required to provide
sources of these metals.
(2) The use of iron for complex formation and subsequent removal of
the cyanide does not appear feasible. The results indicate that good
complex formation with iron occurs only by careful control of the solu-
tion pH and with simple cyanide solutions containing free cyanide. The
experiments with more representative waste cyanide solutions containing
46
-------
TABLE 16. EXPERIMENTAL DATA ON COMPLEX CYANIDE FORMATION WITH
ADDITIONS OF ZINC SULFATE FOLLOWED BY FERROUS SULFATE
Elapsed Time,
min
0
1
2
5
10
20
30
pH 9
10.0
7.0
6.8
6.2
6.2
5.8
5.5
Free Cvanide
pH 9.5
10.0
5.8
5.7
5.4
5.3
4.8
4.4
Concentration, ppm CN
pH 10
10.0
4.4
4.2
3.9
3.7
3.4
3.2
pH 10.5
10.0
6.5
6.3
5.8
5.4
4.8
4.5
pH 11
10.0
8.0
8.0
8.0
7.6
7.0
6.6
TABLE 17. EXPERIMENTAL DATA ON COMPLEX CYANIDE FORMATION WITH
ADDITIONS OF ZINC SULFATE FOLLOWED BY NICKEL SULFATE
Elapsed Time,
min
0
1
2
5
10
20
30
Free Cyanide Concentration, ppm
pH 9 pH 10
10.0 10.0 -
0.27 <0.1
0.17 <0.1
<0. 1 <0. 1
<0.1 <0.1
<0. 1 <0 .1
<0. 1 <0 . 1
pH 11
10.0
2.2
2.0
1.7
1.4
1.2
1.1
CN
pH 12
10.0
3.8
2.7
1.9
1.3
0.9
0.6
TABLE 18. EXPERIMENTAL DATA ON COMPLEX CYANIDE FORMATION WITH
ADDITIONS OF ZINC SULFATE FOLLOWED BY COPPER SULFATE
Elapsed Time,
min
0
1
2
5
10
20
30
Free Cyanide Concentration, ppm CN
pH 9
10.0
1.0
1.1
1.6
1.6
1.5
1.4
pH 10
10.0
0.5
0.5
0.5
0.5
0.5
0.5
pH 11
10.0
0.9
0.8
0.8
0.7
0.7
0.7
pH 12
10.0
3.7
3.3
5.0
2.8
2.6
2.5
47
-------
zinc or cadmium indicate that these metals seriously interfere with sub-
sequent complexing by iron.
Continuous Column Adsorption Studies. The results of previous work on the
adsorption of cyanide by activated carbon indicated that complexation of
the cyanide by certain metals very probably would be required for virtu-
ally complete adsorption and adequate capacity of the carbon. Early ex-
periments made during the first phase of this project showed that free
cyanide and the relatively weak zinc cyanide were very incompletely ad-
sorbed and that the capacity of the carbon was too low. They also showed
that carbon adsorption of ferrocyanide offered some promise. In those
experiments, up to 95 percent of the cyanide (in the ferrocyanide form)
was adsorbed in the continuous carbon column. Although no adsorption
studies were made on copper or nickel cyanides in the belief that if iron
could be made to work, the others would be a poor economic competitor,
it was believed that complex copper or nickel cyanides might be adsorbed
as readily as ferrocyanide.
A more basic study of the requirements for complexing was undertaken, the
results of which indicated that the presence of other metals (such as
zinc, which would normally occur in plating rinse waters) seriously affects
complexation by ferrous iron (Table 16). Accordingly, the plans for car-
bon adsorption experiments on the ferrocyanide system were abandoned. On
the other hand, the presence of zinc was shown not to interfere with the
formation of complex copper or nickel cyanides (Tables 17 and 18).
Parallel to the complexation studies described in the previous section,
several continuous experiments were made to fill in a gap in the existing
data on adsorption. Continuous experiments had been made on free cyanide,
zinc cyanide, and ferrocyanide, but no data had been developed for the
copper cyanide system itself. The purpose of the experiments described
below was to determine the behavior of the copper cyanide system in the
adsorption process.
Studies on the copper cyanide system. Two experiments were made con-
currently by separate investigators to accelerate the work. One experi-
ment was carried out in the apparatus shown in Figure 7 and by the general
procedures described previously. This run was made on a solution con-
taining 100 ppm each of cyanide and copper preadjusted to pH 10.
The second experiment was done in a 1-1/2-inch-diameter glass column con-
taining 100 grams of activated carbon, with arrangement for continuously
feeding solution through a large separatory funnel. It was made on a
feed containing 50 ppm of cyanide and 64 ppm of copper at pH 3.0. At this
pH a white salt, presumably cuprous cyanide [Cu(CN)2] was precipitated.
The feed rates employed in each experiment were reasonably proportionate
to the volume of carbon involved in each case. In the experiment run in
the 2-1/2-foot column which contained 700 grams of activated carbon, the
48
-------
overall feed rate which held fairly constant throughout the run was about
450 ml per minute. The initial feed rate in the second experiment made
with 100 grams of carbon was 70 ml per minute. The precipitated Cu(CN)9,
however, which collected on top of the bed, slowed the feed rate to about
40 ml per minute after several hours of operation. When this precipitate
was removed the original feed rate was restored. Both experiments were
begun with fresh carbon and were continued until the effluents approached
closely to the pH values of the original feed solution. During each
experiment, effluent solutions were periodically analyzed for cyanide,
copper (qualitatively), and pH values.
Cyanide adsorption in the run made on the solution containing 100 ppm of
cyanide and copper at pH 10 was initially good. Effluents for the first
24 hours of operation contained from 1 to 2 ppm of cyanide, equivalent
to 98 to 99 percent removal. After this period, however, the cyanide
content of the effluent gradually mounted to about 4 ppm after 3 hours
and 12 ppm after 4 hours, holding at that level for 2 additional hours.
In the experiment made at pH 3 on the solution containing 52 ppm of
cyanide and 64 ppm of copper, almost complete removal of cyanide was
obtained in the first half hour, but thereafter rose to 3-4 ppm after 1
hour. It remained at this level for an additional hour and then dropped
to about 1.5 ppm when the feed rate was decreased to about 40 ml/minute
by the formation of a precipitate on top of the carbon bed. When the
precipitate was removed and the normal flow reestablished, the cyanide
content of the effluent quickly rose to 5-6 ppm.
The data from the experiment at pH 10 were promising. In this run about
70 liters of solution containing 100 ppm of cyanide was passed through
700 grams of carbon with about 98 percent removal of cyanide. The indi-
cated capacity of the carbon for cyanide, about 1.1 percent, was consider-
ably greater than has been obtained in any previous experiments. Opera-
tion with a feed at pH 3 did not appear as promising because the degree
of cyanide removal was not as great, and the precipitate of Cu(CN)» which
forms at this pH would tend to plug the column.
Although the results obtained in these preliminary runs were promising,
several unknown factors remained to be investigated before the feasibility
of the technique could be confirmed. These included:
(1) The effect of pH on adsorption. In the experiment made at pH 10,
maximum adsorption (98 percent) was obtained when the effluent pH was
about 7. As the effluent in this experiment approached 10, adsorption
decreased. In the experiment made with the feed at pH 3 such a pattern
was not observed and the poorest adsorption occurred with effluents of
about pH 7.
(2) The regeneration of the carbon after loading could be difficult.
After the adsorption experiments, attempts were made to regenerate the
carbon with strong caustic. During regeneration copper removal was
found to be only about 50 percent complete by this means and yielded
49
-------
solutions containing only about 1.3 grams per liter of copper. The
remainder of the copper was readily removed from the carbon with 10
percent nitric acid.
In subsequent experiments, the emphasis was to investigate the effect
of pH values intermediate between 3 and 10. It was found that at pH
values between 3 and 6, the process was complicated by the presence of
precipitated Cu(CN)?. Although good adsorption was obtained, the column
tended to become clogged and the required flow rate of feed solution
could not be maintained. At higher pH values of 10, the operation also
was affected by the presence of precipitated solids, either zinc hydroxide
or copper hydroxide--the latter coming from the use of excess copper
during complexation. Both of these conditions were not consistent with
a practical process in the downflow mode of operation.
An evaluation of cyanide adsorption at pH 7 then was conducted using
the continuous column apparatus. The adsorption experiment was begun
with a feed solution containing 104 ppm total CN, 66 ppm zinc, and 95 ppm
cupric copper. The solution was initially adjusted to a pH of 10 with
sodium hydroxide to provide maximum complexing of the copper and cyanide.
After agitating for 5 to 10 minutes, the solution was adjusted to pH 7
with sulfuric acid for the adsorption cycle. Previous experiments on
complex formation indicated that the copper complex forms readily under
these conditions and results in a solution containing less than about
3.5 ppm cyanide as free cyanide. The remaining cyanide was believed to
react with cupric copper according to the following equation:
2Cu+2 + 20H~ + 7CN~ ->• 2Cu(CN)~2 + CNO" + H20.
The presence of cyanate definitely was established in the feed and
effluent solutions used during this experiment. The results of chemical
analysis showed about 26 ppm cyanate in both the feed and effluent sam-
ples indicating essentially no adsorption of this component.
During the adsorption cycle, the carbon column was operated with a feed
rate of between 300 and 400 cc/minute over a period of 11 days. A total
of about 275 gallons of solution was fed to the column. Experimental
data obtained as the run progressed are shown in Table 19.
The results obtained during the .experiment were encouraging. Cyanide
removal initially was greater than 99 percent and gradually decreased to
'about 85 percent near the end of the experiment. The operation was dis-
continued when the level of cyanide in the effluent increased above 10
ppm of cyanide.
For the entire experiment, a total of about 70 grams of cyanide was ad-
sorbed on the 700 grams of carbon contained in the column. Approximately
30 grams of cyanide were converted to cyanate and not adsorbed by the
carbon. The indicated loading capacity for cyanide of about 10 percent
was several times higher than achieved in previous work. Additionally,
50
-------
complete removal of copper was obtained throughout the experiment indi-
cating no losses of this component. The amounts of zinc adsorbed however
were not established.
TABLE 19. CONTINUOUS ADSORPTION DATA FOR THE
COPPER CYANIDE SYSTEM
Total Feed . .
Solution Treated,
gal
5
40
70
90
110
130
155
180
195
225
250
265
275
Effluent
PH
7.9
7.7
7.7
7.3
6.6
7.1
7.0
7.0
6.8
6.9
6.5
7.1
—
Conditions
ppm CN
0.4
0.4
0.2
0.9
2.0
2.2
2.7
5.0
4.7
5.3
7.5
9.9
14.5
Apparent
Percent
Removal
99.6
99.6
99.8
99.1
98.0
97.8
97.3
95.0
95.3
94.7
92.5
90.1
85.5
(a) Initial solution contained 75 ppm cyanide, 26 ppm cyanate, 66 ppm
zinc, and 95 ppm copper.
To complete the evaluation of the copper cyanide system, preliminary
experiments then were conducted on the regeneration of the loaded car-
bon. This work was done on both the entire volume of carbon used in the
adsorption experiment and on small samples taken from the column.
Initial attempts to regenerate the carbon column were made with sulfuric
acid and hydrochloric acid solutions. With 10 percent sulfuric acid
regenerant, significant amounts of cyanide were removed from the column
but the stripping required several hours. Approximately 34 grams of
cyanide were removed which corresponded to about 50 percent of the
adsorbed cyanide. Virtually none of the copper was removed.
In order to cover a wider range of regeneration conditions, samples of
carbon were removed from the column after the sulfuric acid stripping
and tested separately. The results of these small-scale tests are given
in Table 20.
The results of these small-scale experiments indicated that several pro-
cedures could be used to effect efficient stripping. These included
51
-------
TABLE 20. STRIPPING CONDITIONS AND RESULTS
Weight
Expt. Carbon
No. g'a) Stripping Solution
Temp.
Weight %
-------
treatment with hydrochloric acid, alkaline cyanide solution, sodium
hypochlorite, sulfuric acid, etc. These experiments were made on small
samples of loaded carbon that had been previously dried at 110 C. Work
done at this time on carbon that had not been dried, however, was not
as promising. For example, in an experiment in which 10 percent sulfuric
acid was circulated through a bed of loaded carbon for 20 hours only
about 50 percent of the cyanide and only a few percent of the copper was
removed.
Because stripping with sulfuric acid in a system that would permit the
recovery of copper and cyanide appeared to be more practical and possibly
more economical for the smaller plater, emphasis was placed on investi-
gating such a process.
A fresh lot of Pittsburgh OL carbon (500 grams) was charged into the
adsorption column and loaded with copper cyanide by feeding 100 gallons
of a solution containing 76 ppm of cyanide, 100 ppm of copper, and an
estimated 25 to 30 ppm of cyanate (CNO) (arising from the oxidation of
cyanide by cupric copper and/or air). The latter was used for agitation
during solution preparation. Adsorption efficiency was excellent (99+
percent of the copper and 96-99 percent of the cyanide) during the adsorp-
tion cycle. The loaded carbon, by calculation, contained 36.1 grams of
copper and 29.0 grams of cyanide. This carbon was then treated for 13
hours by recirculating a solution of 10 percent sulfuric acid through it
at a flow rate of 300 ml per minute. During the recirculation, the pump
reservoir of acid was constantly aerated. Exit gas from the reservoir was
continuously passed through a caustic trap to collect cyanide vapors. At
the end of the run, the acid solution was analyzed for copper and the
caustic trap solution for cyanide. The results are shown in Table 21
(Run 1). They indicated that after 13 hours only 41 percent of the
copper and 57 percent of the cyanide was removed.
Run 2 (Table 21) was made on the partially stripped carbon from Run 1.
In this run, which was carried out for 18 additional hours, with 10 per-
cent sulfuric acid under conditions of aeration, 100 percent of the re-
maining adsorbed copper was removed from the carbon. The conclusion that
the additional 18-hour treatment alone was responsible for complete removal
was not justified however. This was because the partially loaded carbon
from Run 1 was used in an unsuccessful experiment to adsorb cyanide from
an uncomplexed or free cyanide solution before attempting Run 2. It is
conceivable that the treatment with free cyanide solution may have contri-
buted to the complete removal of copper experienced in Run 2. Another
possible contributing cause for the complete stripping achieved in Run 2
was that previous to the run, the carbon bed was drained and partially
dried in the column.
Run 3 (Table 21) was made on the same 500-gram lot of carbon, which had
been reloaded with copper and cyanide. The purpose of this run was to
disclose the role played by the aeration procedure in the stripping of
copper and cyanide. In this run, the agitation required to keep cyanide
gas swept out of the circulating acid solution was supplied by bubbling
nitrogen instead of air through the reservoir. The results indicated that
53
-------
TABLE 21. LABORATORY STRIPPING OF COPPER AND CYANIDE
FROM ACTIVATED CARBON
Stripping Conditions
Run Time,
No. hr
1
2 (a)
3(b)
5(0
13
18
16
58
Stripping
Medium
10% H2S04
10% H2S04
10% H2S04
10% H2S04
Aeration
Yes
Yes
No
Yes
Loaded Carbon Percent
Carbon,
e
500
500
500
500
Copper,
36.1
19.9
36.3
42.7
Cyanide. Stripped
e Copper
29.0 41
100
29.0 15
100
Cyanide
57««
—
35 (d)
ioo(e)
(a) This run was made on the carbon left from stripping Run No. 1; which
still contained considerable copper and cyanide. However, before
beginning this run, the carbon left from Run 1 had been subjected to
treatment with 20 gallons of free cyanide solution containing 100 ppm
of cyanide, in an experiment to determine whether copper-bearing carbon
would adsorb free cyanide. This pretreatment may have contributed to
the complete removal of copper.
(b) This run was made on the completely stripped carbon from Run 2, which
had been reloaded by the adsorption of 100 gallons of solution containing
76 ppm of cyanide and 100 ppm of copper. No air was employed in this
run and agitation to sweep cyanide out of the adsorbing acid was provided
by bubbling a stream of nitrogen through the solution.
(c) This run was made on the partially stripped carbon from Run 3, to
which additional copper had been added as copper sulfate. However,
before this run, the column of copper-rich carbon had been unsuccess-
fully tested as a medium for adsorbing free cyanide. It is possible
that this treatment with free cyanide may have contributed to the com-
pleteness of copper removal obtained.
(d) Based on analyses of absorbing caustic solution. Does not take into
account CNO formation. Values shown, therefore, are somewhat low.
(e) Based on analyses of carbon after stripping, which indicated complete
removal.
54
-------
considerably less copper was stripped under these conditions than when
air was used as in Run 1 (15 vs. 41 percent of the copper). It is be-
lieved that the oxidation of cuprous to cupric copper by air contributed
significantly to the oxidation of cuprous cyanide and the consequent
solubilization of copper.
In Run 5 (Table 21) complete stripping of copper and cyanide was again
achieved. This run was made on the partially stripped carbon from Run 3
to which additional copper had been added. As in the case of Run 2, it
was not justifiable to attribute the complete removal of copper and cyanide
simply to the longer treatment with aerated acid. This was because prior
to Run 5 the copper enriched partially loaded column had been used to
determine its effectiveness as an adsorbent for free cyanide. Again, it
is conceivable that the treatment with free cyanide solution may have
contributed to copper removal. After Run 5, a sample of the carbon was
removed from the column and analyzed for cyanide. The analyses (0.05
percent) indicated that cyanide removal also was virtually complete.
The results of Runs 2 and 5 showed that it was possible to strip com-
pletely both copper and cyanide from the carbon. What was not yet known
was the relative importance of the conditions employed in these runs, i.e.,
time, aeration, pretreatment with dilute free cyanide solution, and par-
tial drying of the bed. Also, in the stripping runs, a good material
balance for copper was always obtained. The balances for cyanide, however,
were poor with up to 50 percent of this component being unaccounted for.
It is believed that during stripping, the-missing portion of the cyanide
was oxidized, either to cyanate, or carbonate, or was converted to
ammonia; however, no analytical verification of these possible reactions
was completed. . . .
While the preceding studies were favorable to the continued development
of a process based on the adsorption of a copper cyanide complex, uncer-
tainty about various important aspects of such a process still existed.
These were:
(1) The effect of multiple cycles of adsorption and stripping on the
adsorptive capacity of the carbon.
(2) The requirements for effective stripping with sulfuric acid. The
previous work did not rule out the possibility that for sulfuric acid
stripping to be effective, it would have to be augmented by such opera-
tions as predrying of the carbon or pretreatment with alkaline cyanide
solution before stripping.
(3) Cyanate formation. In the previous work about 20 percent of the
original cyanide was converted to cyanate. It was believed that the
major mechanism of cyanate formation was oxidation by the cupric copper
sulfate used for complexing according to the equations:
55
-------
3CuSO. +
4
Copper
sulfate
Y
(CN)2 +
Cyanogen
6NaCN •*
Sodium
cyanide
2NaOH -»•
Sodium
hydroxide
2CuCN-Cu(CN)2 + (CN)2 -
Cupro-cupric Cyanogen
cyanide A
NaCN +
Sodium
cyanide
NaCNO + H20
Sodium
cyanate
Sodium
sulfate
In the earlier runs, however, the solution was agitated by a stream of
air during the entire period of solution preparation. It was considered
possible that some cyanate may have been formed by the simple oxidation
of cyanide according to the equation:
NaCN + 1/200 -» NaCNO
Sodium
cyanide
Sodium
cyanate
During subsequent work, a series of multiple adsorption and regeneration
cycles was made to resolve these uncertainties.
In order to expedite the completion of the multiple cycle series, it was
decided to scale down the system. Instead of the 2-1/2-foot columns
containing about 500 to 700 grams of carbon, which required about 100
gallons of solution per cycle, a smaller adsorption column containing only
125 grams of carbon was used. Conditions and results of the experiment
are summarized in Table 22.
The Pittsburgh OL granular carbon showed excellent adsorption efficiency
for the first 3 cycles adsorbing respectively 94, 100, and 100 percent of
the copper and 91, 97, and 94 percent of the cyanide. The lower adsorp-
tion percentages obtained in Cycle 1 compared to those in Cycles 2 and 3
are believed due to the extremely high feed rate used in that cycle
(corresponding to 30 to 40 gpm per 100 pounds of carbon. In Cycles 2
and 3, where lower feed rates were used (10 to 20 gallons per minute per
100 pounds of carbon) adsorption efficiency for both copper and cyanide
improved. The short bed (12 inches) employed in these tests to save time
contributed to some unmeasured extent in the failure to obtain 100 percent
adsorption.
According to technical information supplied by the Pittsburgh Carbon
Company which manufactures the carbon used in these experiments, a 10-foot-
long bed of carbon is the minimum required for the accurate translation
of laboratory data to practice. It is believed that had the recommended
10-foot-long columns been used, even higher adsorption efficiencies would
have been obtained in Cycles 1 through 3.
The relatively poor adsorption obtained in Cycle 4 (80 percent for both
copper and cyanide) is probably the result of a combination of factors:
56
-------
TABLE 22. CONDITIONS AND RESULTS OF MULTIPLE CYCI.K ADSORPTION-REGENERATION EXPERIMENTS
Ul
Cycle
No.
1
2
3
4
5
Feed Solution
Compos! tion(a>
Cu, CN,
ppm ppm pll
98 78 7.0
98 78 7.0
98 78 7.0
118 107 7.0
98 78 7.0
Feed
Kate,
ml/min
(g/min/
Ib C)
300-400
(29-38)
100-200
(15-20)
80-100
(10-15)
80-150
(10-20)
100-200
(15-20)
Vol ume
Fed,
1 i ters
(gal)
94
(25)
94
(25)
94
(25)
100
(27)
112
(30)
Wt C
in
Column,
g
125(8)
120(8)
U5(S)
110(K)
105(8)
Effluent
Composi ti on (^'
Cu, CN,
ppm ppm pfl
5 7 2.3-7.000
<1 2 2.2-7.20')
'1 4.5 2.2-7.lO>)
24 21 2.2-8.5(»>
15 16 2.3-7.7<'>)
Weight Adsorbed(c) . Stripping
Adsorption, Cu CN Wt1-6^
;;(c) lbs/100 lbs/100 Aeration Intermediate Stripped Percent Stripped
Cu CN g Ib of C g Ib of C df Acid l>rying(d) Hours Cu CN Cu (.X^ >
94 91 8.74 7.0 6.67 5.34 Yes No 18 7.9 6.9 93 100+
100 97 9.21 7.7 7.12 5.95 Yes Mo 37 6.8 6.4 83 93
100 94 9.21 8.0 6.83 5.96 Yes No 28 7.1 5.9 77
-------
Ul
Footnotes for Table 22.
(a) Standard feed solutions for Cycles 1 through 3 and 5 were made up by adding cupric sulfate to sodium
cyanide solution, adjusting pll to 10 and agitating with air for 10 minutes to effect complexing; and
then adjusting pH to 7,0 with dilute l^SO^. Compositions shown based on analysis of a single solution
made by the standardized procedure. Feed solution for Cycle 4 was a composite of two different types
of solutions. The first third of the run was trade on a solution prepared by attempting to complex the
free^cyanide with, metallic .copper under conditions of aeration. When this appeared to be impractically
slow in reducing free cyanide, cupric sulfate was added in small increments until free cyanide, by the
specific CN ion meter showed 4 npm. After 9 gallons of this material had been fed through, a break
occurred in the solution lines during the night. The run was completed by feeding 18 gallons of the
standard feed solution through the column. The composition shown for this feed solution is based on
the analysis of a composite sample.
(b) For Cycles 1 through 3, Cu and CN compositions are based on average analyses of effluent samples. For
Cycles 4 and 5, in which the effluents contained sufficient copper to interfere with the routine analytical
procedure, compositions are based on analysis of the combined effluent.
(c) Based on feed and effluent analyses and on weight of carbon in column.
00 (d) Drying of the carbon in Cycle 4 done by passing a slow stream of air downward through the column for 12
hours. After drying, stripping with acid was begun.
(e) liased on analyses of acid stripping solution and caustic scrubbing solution.
(f) Weight stripped x 100-r weight adsorbed.
(g) Reduction in weight of carbon from cycle to cycle due to removal of samples.
(h) pH at beginning of each cycle was as low as 2 to 3, but rose to the pli of the feed after a few gallons
had been fed. In Cycle 4, the effluent pil exceeded the feed pll for an as yet unexplained reason. A
similar though less pronounced situation occurred in Cycle 5.
-------
(1) Feed solution composition. Two feed solutions were used in this run.
The first third of the run was made with a solution in which the cyanide
was complexed partially by metallic copper (cement copper and copper wool)
and then, when this procedure was found to be too time consuming, by
cupric sulfate. About 9 gallons of this solution had been passed through
the column with an effluent containing 5 to 7 ppm of cyanide, and 3 to 5
ppm of copper, when a leak in the pump lines occurred and the remainder
of the 25-gallon batch leaked to the sewer and escaped treatment. Follow-
ing this a batch of the standardized solution as used in Cycles 1 through
3 was passed through the carbon. Almost from the outset, the effluent
exhibited an anomalous behavior. The pH rose to 8.5 and significant
amounts of both copper and cyanide began to appear in the effluent.
These results point up the possibility that the attainment of proper
complexing conditions—conditions that will give good adsorption—may
be touchy and the desirability of specific studies to determine the work-
able limits within which solution compositions can be varied.
(2) Incomplete stripping of copper deposited on the carbon in previous
cycles. Material balance data based on the data shown in Table 22
indicate that in the first three cycles 27.10 grams of copper were
adsorbed by the carbon, whereas only 21.8 grams were removed by stripping,
a difference of 5.3 grams. Analysis of the carbon after Cycle 3 strip-
ping indicated that it still contained 3.1 grams of copper. The amount
of copper stripped after Cycle 4 adsorption, 125 percent of that calcu-
lated to have been adsorbed in Cycle 4, reflects this continual buildup
of copper on the carbon in the first three cycles. Previous work has
shown that incomplete stripping leads to poor adsorption. It is quite
probable that such incomplete stripping was a significant factor in the
poor adsorption in Cycle 4.
(3) Decreased adsorptive efficiency of the carbon. It is possible that
the adsorptive efficiency of the carbon deteriorates through some unex-
plained mechanism, with repeated adsorption and regeneration cycles. That
serious deterioration in adsorptive powers occurs, however, is not con-
sidered likely, considering that the capacity of the carbon in Cycle 4
for copper and cyanide is great or greater than it was in Cycles 1 through
3.
The lower adsorption percentages obtained in Cycle 5 which was run with
the standardized feed solution were somewhat disappointing, and might
indicate a loss in the adsorption efficiency of the carbon. However,
considering that 20 grams of the carbon had been removed by sampling and
that the length of the bed was accordingly decreased by about 20 percent,
it is believed that channeling of the solution contributed significantly
to the amount of copper and cyanide in the effluent. As shown in Table
22, the loaded carbon from Cycle 4 contained 8.8 pounds of copper and
6.61 pounds of cyanide per 100 pounds of copper, which is comparable to
the loadings obtained in the previous cycles. This fact provides some
indication that the capacity of the carbon for copper cyanide was not
seriously impaired.
59
-------
Preliminary Evaluation of Treatment Costs
An attempt was made to assess the relative merits of the candidate
cyanide treatment processes studied during the initial phase of the
program. It was assumed that the cost of chemicals would be the most
important indicator of the relative cost for these processes.
Ion Flotation. The costs involved in operating an ion flotation system
were assumed to include primarily the flotation collectors and complex-
ing chemicals which were added to effect removal. It was assumed that
regeneration with recovery of chemicals was not feasible; the spent
collector loaded with cyanide simply would be collected as a sludge and
disposed of in some way, i.e., incineration, chemical oxidation, etc.
During the experiments, two approaches involving flotation of complex
nickel and iron cyanides appeared to show good removals of cyanide. The
efficiency of these approaches and the estimated reagent costs are shown
in Table 23. On the basis of a typical plant having a waste flow of
15 gallons per minute containing 100 ppm of cyanide, it was estimated
that the equipment costs for the ion flotation process would be about
$4,000.
TABLE 23. EFFICIENCY AND COSTS OF ION FLOTATION OF
COMPLEX CYANIDES
Removal Reagent
Attained, (a) Requirement, (a) Costs
Reagent TL lb/lb CN $/lb CN $/1000 gal
Iron Cyanide Solutions <3 10 ppm CN
Ferrous Sulfate (FeSO,) 1.0 0.01 0.001
@ $0.01/lb
Amine Collector 88 3.3 1.09 0.09
@ $0.33/lb
Totals 1.10 0.09
Nickel Cyanide Solutions @ 10 ppm CN
Nickel Sulfate (NiSO,)
@ $0.37/lb
Amine Collector 93
@ $0.33/ Ib
1.5
.13.3
Totals
0.56
4.39
4.95
0.04
0.35
0.39
(a) Based on experimental data shown in Tables 7 and 8.
60
-------
Activated Carbon Adsorption. In determining operating costs for the
carbon process for cyanide removal, a number of important assumptions
had to be made. Based on projected development and refinement of the
process, it seemed reasonable to assume that the copper sulfate used
for complexing could be almost entirely recovered and reused in the
system. It also was assumed that the carbon could be regenerated in-
definitely with no loss in adsorptive capacity; thus carbon replacement
costs would be minimal. The major chemical consumed, therefore, would
be sulfuric acid which was used for regeneration and pH adjustment in
the complexing step.
On the basis of the laboratory experiments, about 5 pounds of sulfuric
acid were used to adsorb and regenerate 1 pound of cyanide, indicating
a cost of $0.10 per pound of cyanide removed (@ $0.02/lb H2S04). Con-
servatively estimating a loss of 10 percent of the copper sulfate per
regeneration, additional cost of about $0.07 per pound of cyanide would
be incurred. Total chemical costs, therefore, could approach $0.17 per
pound of cyanide treated. For the typical plant (15 gallons per minute
at 100 ppm cyanide) equipment costs were estimated at $7,000.
Acidification-Volatilization. Preliminary estimates of the capital and
operating costs for the acidification-volatilization process were made
using the same estimating data and techniques employed above. Capital
cost for a plant capable of treating 15 gallons per minute of waste was
estimated at $8,500. The estimate is believed to be conservative.
The cost of reagents for this treatment were estimated based on labora-
tory data which indicated that a pH of 3 and temperature of 100 F would
be used during treatment. The following tabulation shows how the reagent
costs were calculated.
Reagent
Sulfuric Acid @ $0.02/lb
Soda Ash(a) @ $0.025/lb
Sodium Hydroxide(b) @ $0.03/lb
Amount
Required,
Ib/lb CN
18
9
3
Total
Estimated
Cost, $/lb CN
0.36
0.22
0.09
0.67
(a) Assume half the weight of H2S04 to bring the solution back to a
higher pH before disposal.
(b) This is intended to provide enough caustic to adsorb HCN and CC^
if it is not flared to the atmosphere.
Comparison of Processes. In addition to the above cost estimates for
the experimental processes, the conventional method for cyanide
treatment using sodium hypochlorite solution was evaluated. Using
61
-------
the same basis mentioned previously, reagent and equipment costs were
estimated at $1.35 per pound of cyanide and $6,500, respectively.
A comparison of these costs with those determined for the experimental
techniques is shown in Table 24. This economic comparison indicates
that the three experimental methods have somewhat lower costs than the
conventional method with carbon adsorption showing the greatest
advantage.
TABLE 24. COMPARISON OF COSTS FOR SELECTED
TREATMENT METHODS
Ion Flotation
Carbon Adsorption
Acidification- Volatilization
Conventional Method with NaOCl
Reagent
Cost,
$/lb CN
1.10
0.17
0.67
1.35
Capital
Cost,
-------
Phase 2: Pilot-Scale Investigation
Equipment and Procedures
The second phase of experimental work on the activated carbon process was
conducted in pilot-plant equipment employing two 1-foot-diameter columns
each 6 feet tall and packed with activated carbon. The system was in-
stalled at a plating plant in Columbus, Ohio, and operated on actual
waste zinc cyanide rinse waters produced at this plant. The plant had
available two zinc cyanide plating lines from which rinse water could be
obtained—one an automatic barrel line and the other an automatic rack
conveyor. During operation of the lines, the entire waste rinse water
flow was mixed in a floor drain and discharged to a common sump leading
to the sewer. Combined flow was 25 to 50 gpm with cyanide concentrations
up to 50 ppm.
The system was operated initially on combined rinse waters from both
lines. For the majority of later runs, only concentrated cyanide rinse
water from the barrel line was used. Figure 13 shows photographs of the
barrel plating line, the pilot-plant waste treatment system, and the
point of collection of the concentrated cyanide rinse waters.
The equipment shown in Figure 13c consisted basically of two 1000-gallon
rubber-lined tanks, an adsorption module containing two carbon columns,
and associated pumps, piping, and instruments to operate the system.
Each column was filled with 100 pounds of Pittsburgh Type OL activated
carbon. The unit includes piping, closed top polyethylene tanks, and
valves to permit recirculation of the regenerating solutions.
The procedures used in operating the pilot-plant system were developed
and modified as the study progressed and several modifications were
employed.
In preliminary adsorption runs, mixed rinse waters from both plating
lines were used as feed to the system. These rinse waters contained about
50 ppm of cyanide and smaller quantities of zinc, iron, and hexavalent
chromium. The objective of these preliminary trials was to evaluate
adsorption efficiency on the waste solutions as they were actually pro-
duced by the plant and then compare the results with an improved procedure
which involved addition of copper to the rinse water and subsequent ad-
sorption of complex copper cyanides. As discussed below, these runs
showed unexpected results, so that it was decided to use only concentrated
cyanide rinse waters produced by the barrel plating line. These rinse
waters were collected as shown in Figure 13b from a primary rinse tank
immediately following the plating tank in the normal operating sequence
of the line. The rinse waters contained 300 to 400 ppm of cyanide.
Following a preliminary operation with the concentrated rinses, the follow-
ing general procedure was used for subsequent detailed adsorption runs:
63
-------
FIGURE 13a. AUTOMATIC BARREL PLATING
LINE AT SUPERIOR PLATING COMPANY
FIGURE 13b. COLLECTION POINT FOR
CYANIDE RINSE WATER
FIGURE 13c. PILOT-PLANT CARBON
ADSORPTION SYSTEM
64
-------
(1) Concentrated rinse water was pumped from the plating line at 2-3
gpn. and collected in one or both of the 1000-gallon holding tanks.
(2) The solution was agitated by high pressure air and analyzed for
cyanide content by the standard analytical procedure.
(3) Copper sulfate solution (60 gpl Cu) was added and the solution
agitated for 10 minutes to effect formation of the complex copper cyanide.
A pH of 10 was maintained during this time by addition of sulfuric acid,
if necessary. For most runs, a copper to cyanide ratio of 0.75 by
weight was selected.
(4) After complexing, the solution was adjusted to a pH of 6-7 by add-
ing sulfuric acid; the precipitate which was formed was allowed to
settle for a period of at least one hour. The supernatant clear liquid
was pumped to the adsorption column at a rate of 2-3 gpm.
(5) Effluent from the adsorption column was discharged to the sewer.
The above procedure was employed in batchwise manner to effect loading
of the carbon column until a breakthrough was achieved. The breakthrough
point was selected as the time at which copper first appeared in effluent
samples from the system. This normally was measured by qualitatively
testing the effluent with dilute acid and observing the presence of
colloidal Cu(CN)_. Throughout the entire adsorption cycle, samples of
both the feed and effluent solutions were obtained periodically for
chemical analysis.
Regeneration runs on the loaded carbon bed normally were carried out after
each adsorption cycle. The procedures used for regeneration also were
subject to modification during the pilot-plant campaign. Based on previous
work, the required methods consisted of recirculating a dilute aerated
acid solution through the carbon bed for a period of about 24 hours. This
effectively stripped both copper and cyanide from the carbon, leaving
copper dissolved in the acid solution as cupric sulfate and producing
hydrogen cyanide vapors ;which were then swept out of the acid solution.
The use of air injected into the carbon column along with the sulfuric
acid was found to be essential to oxidize cuprous to cupric copper, there-
by breaking the copper cyanide complex to liberate cyanide. (In pilot-
plant runs, air introduced into the acid solution did not accomplish this
end.)
It was necessary to supplement the acid treatment of the carbon with a
final caustic wash and to readjust the column effluent to pH 7. It is
believed that this final treatment eliminated any residual adsorbed acid
and effected greater adsorption efficiencies in the subsequent adsorption
step.
65
-------
Preliminary Results
Initial pilot-scale experiments on cyanide adsorption were begun with
mixed rinse waters from both plating lines being used as feed to the
system. These rinse waters contained about 50 ppm of cyanide and smaller
quantities of zinc, iron, and hexavalent chromium. The objective was to
evaluate adsorption efficiency on waste solutions as they were presently
produced at the plant, without addition of copper to effect complexing
of the cyanide. ( Pertinent operating data for these runs are shown in
Table 25,
It was found that an unexpectedly high adsorption of cyanide was obtained
from these mixed rinses. In Run 1, for example, a total of about 3100
gallons of rinse water was fed to the system before cyanide was observed
in the effluent samples. The apparent adsorption capacity of the column
during this run was 1.2 pounds of cyanide per 100 pounds of carbon, a
value which was several times greater than that achieved in laboratory
studies using synthetic zinc cyanide rinse waters.
t
It was suspected that the unusually high adsorption capacity in Run 1
was caused by the presence of iron in the mixed rinse waters used as
feed. During a preliminary regeneration with sulfuric acid, a blue-
colored solution was formed which indicated the presence of complex iron
cyanides. Chemical analysis of this regenerating solution indicated
1570 ppm total iron, an amount equivalent to 16 ppm total iron in the
feed solution used during the adsorption.
Although the presence of iron in the rinse waters seemed beneficial in
the form of a complexing agent, at least some of the iron was not com-
plexed and formed a ferric hydroxide precipitate which tended to clog
the carbon column and thus prevent operation at the desired flow rates.
In Run 2, a high level of adsorption was again achieved with mixed rinse
waters at a pH of 9.5; however, the run was terminated when flow rates
decreased to 1.0 gpm at a pressure drop of 8 psig through the column.
Normal flow rates (2.0 gpm) were restored only by flushing the column
with a dilute sulfuric acid solution.
Subsequent experiments then were conducted employing copper sulfate to
complex cyanide in the rinse water solutions. Due to uncertainties re-
garding the mixed rinse waters and the inability to control this stream,
it was decided to employ a separate concentrated cyanide rinse water
produced only from one location in the plant. This rinse stream was
collected from the overflow of a 2-stage countercurrent rinse tank immedi-
ately following the zinc cyanide barrel plating tank (see Figure 13b).
The adsorption data on initial pilot-plant runs with complexed copper
cyanide feed solution are shown in Table 26. These runs were conducted
initially by diluting the strong cyanide rinse stream from the plating
line to about 100 ppm cyanide using tap water. Copper sulfate then was
added using the procedures described previously.
66
-------
TABLE 25. PRELIMINARY PILOT-PLANT ADSORPTION DATA ON MIXED CYANIDE RINSE WATERS
Effluent Data Cyanide
Gallons
Run 1 3130
Run 2 280
Feed Data
pH ppm CN
6.4-7.7 42-49
9.5 65
Typical
ppm Cu pH ppm CN
0(a) 6.8-9.5 <1.0
0(a) 8.1-9.5 <1.0
Final Adsorbed
ppm CN Ib %
17 1.2 >98
<1.0 0.1 >98
Indicated
Loading Capacity,
Ib. CN/100 Ib carbon
1.2
— —
(a) No copper was added in these initial runs to evaluate adsorptive capacity on waste waters ' as
produced" at the plating plant.
-------
TABLE 26, PRELIMINARY PILOT-PLANT ADSORPTION DATA ON THE COPPER CYANIDE SYSTEM
00
~4BS»
Run(a
No.
3a
•IK
4a
4b
4c
4d
4e
)
Gallons
807
450
1630
1317
1623
1370
703
Feed Data
Hours
5
6
26
12
16
16
g
PH
6.7
7.6
7.1
7.4
11.2
6 5-11 2
CN, ppm
68
91
96
44
98
336
455
Cu , ppm
100
100
113
113
113
189
189
Effluent Data
Typical Final
pH CN, ppm CN, ppm
8.1 <1.0 " <1.0
8.9 <1.0 <1.0
Run 3 Totals
7.0-8.9 <1.0 <1.0
8.9 <1,0 <1.0
8.1 <1.0 <1.0
10.2 >347 347
9.2-11.1 400 153
Run 4 Totals*' '
Weight
Ib CN
0.46
0.34
0.80
1.31
0.48
1.33
Assumed
In
Adsorbed Loadi
Ib Cu
0
0
1
1
1
1
Assumed
3.12
4
.67
.38
.05
.54
.24
.53
to be
to be
.31
% Cu Ib, C
>98
>98
>98
>99
>98
>99
zero
zero
>98
dicated
ng Capacity,
N/100 Ib carbon
—
—
0.8
—
—
— —
— —
—
3.1
(a) These runs were performed on a fresh carbon bed of 100 Ib Pittsburgh OL Carbon.
(b) Based only on first part of run in which high adsorption was achieved.
-------
In adsorption Run No. 3, effluents from the carbon column containing
less than 1.0 ppm of cyanide were consistently produced from feed solu-
tions containing 68 to 91 ppm cyanide and about 100 ppm of copper.
General operation in this run, however, was unsatisfactory due to an
increase in pressure drop through the carbon column caused by solids
which had precipitated from the feed solution. Run No. 3 was terminated
after feeding about 1250 gallons. The column was washed for 3 hours
with 10 percent sulfuric acid. Run No. 4 was started immediately follow-
ing the acid wash, which had effectively eliminated the high pressure
drop through the column.
In the first part of Run No. 4, adsorption was again carried out on dilute
solutions. Approximately 3.12 pounds of copper were adsorbed during this
period with adsorption efficiencies greater than 98 percent. In the later
stages of the run, an attempt was made to feed strong cyanide solutions
as collected from the plating line without dilution. Operating conditions
in this latter part of the run were not optimum and very poor adsorption
was observed. The inability to closely control the feed composition (pH
and ratio of copper to cyanide) was believed responsible for the poor
operation. This run was terminated when high levels of adsorption could
not be reestablished.
Results obtained in preliminary regeneration studies on Runs 1 through 4
are summarized in Table 27. As previously indicated, attempts to regen-
erate the carbon after Run 1, in which mixed rinse waters were used, con-
firmed the observation that significant adsorption of cyanide occurred
during this run, presumably as the iron complex. As can be seen, a total
of 0.65 pound of cyanide and 0.33 pound of iron was stripped from the
carbon during 48 hours of regeneration. Also of significance was the
fact that measurable quantities of chromium also had been adsorbed during
the operation.
Three difficulties were encountered: (1) the stripping of cyanide from
the column for Run 1 was slow, requiring 48 hours to remove about 50
percent of the adsorbed cyanide, (2) the stripping of cyanide was incom-
plete, and (3) cyanide was not effectively and rapidly removed from the
sulfuric acid regenerating solution by air sparging. A similar behavior
also was observed in Regeneration Run 2. These preliminary runs point
out that iron cyanide complexes very probably cannot be effectively
stripped by a simple treatment with sulfuric acid, and that additional
steps would be needed to achieve a successful approach.
The preliminary data on regeneration of copper cyanide complexes (Runs 3
and 4 shown in Table 27) also indicate the very slow stripping of cyanide
and copper from the carbon that was observed in the pilot-plant system.
Results from Run No. 4, for example, indicate that only about 40 percent
of the cyanide was removed in 32 hours. It was found, however, that
copper removal was dependent primarily on the method of introducing air
into the system. The use of air was found to be a critical factor in
laboratory studies and was believed to oxidize cuprous to cupric copper,
thereby breaking the copper cyanide complex and liberating the cyanide.
Air introduced into the acid solution in the pilot plant was not effectively
69
-------
TABLE 27. PILOT-PLANT DATA ON STRIPPING OF CYANIDE COMPLEXES
Run
No.
1
2
3
4a
4b
4c
4d
Column (a)
Number
1
1
2
2
2
2
2
Regenerating Solution
Volume,
gal
25
25
25
25
25
25
25
H2S04,
Ib-
20
20
16
20
3
6
4
Type of
Aeration Hours
Acid sparging 48
Acid sparging 8
Acid sparging 3
Acid sparging 16
Acid sparging 16
Column Injection 40
Column Injection 40
Run 4
Intermediate
Drying of
Carbon
No
No
No
No
No
Yes
No
Totals
Weight Stripped, Ib Percent
CN
0.65
Cu
0
Fe Cr CN
0.33 0.05 54
Stripped
Cu
Negligible
0.03
0.68
0.35
0.19
0.08
1.30
trace
0.02
0.04
0.84
3.49
4.39
0.14 — 4
0.07
0.02
_ _ — __
— _ __ __
0.09 42
0
—
__
__
100
(a) Both columns charged with 100 Ib Pittsburgh OL Carbon.
-------
used as shown by the results for Runs 4a and 4b in Table 25. Much better
results were obtained in Run 4c when the column was predried with air and
when air was injected directly into the column along with the acid
regenerant.
Results of Multiple Cycle Adsorption and Regeneration Runs
Following the preliminary operation, a series of multiple cycle adsorp-
tion and regeneration runs were conducted on a single carbon bed using
concentrated cyanide rinse waters containing from about 200 ppm to over
350 ppm cyanide. The data obtained during this multiple cycle operation
are presented in Tables 28 and 29.
For adsorption cycles, the standardized procedure described previously
was used for all runs. One additional step was included in an attempt
to eliminate the formation of precipitates from the feed to the column.
This consisted of allowing the solution to settle in the holding tanks
for a period of at least one hour after adjusting the pH to a value of
7. This settling procedure effectively reduced the amounts of the zinc
hydroxide and iron compounds which previously had caused clogging of the
carbon column. The settled sludges produced in the feed tanks were
analyzed to determine what percentage of the contaminants were removed by
settling.
As can be seen from the data in Table 28, adsorption efficiencies for
cyanide and copper were good through 5 cycles of operation. Copper re-
moval remained virtually complete (greater than 99 percent) in each cycle
up to the point of breakthrough. The adsorption of cyanide decreased in
Cycles 2 and 3 to a low of about 80 percent and then improved to about
98 percent in Cycles 4 and 5. This poorer performance in Cycles 2 and
3 was subsequently traced to an improvement discovered in regeneration
for preceding cycles. Following Run 3, the regeneration procedure was
modified to include a final treatment of the carbon with caustic to
readjust the column pH to a value of 7 just prior to the succeeding ad-
sorption. This final caustic treatment improved adsorption efficiencies
for Cycles 4 and 5.
Discussion of Pilot-Plant Runs
The results of the pilot-plant study demonstrated that real zinc cyanide
rinse waters can be effectively treated by a carbon adsorption process
and that the method will produce effluents which will meet current regu-
lations governing discharges of cyanide to municipal sewers. One of the
major goals of the study was to demonstrate that the carbon does not lose
its adsorptive capacity after repeated adsorption-regeneration cycles.
During the study, the adsorptive capacity of the carbon remained high
after 5 cycles of adsorption and regeneration, indicating that the column
71
-------
TABLE 28. RESULTS OF MULTIPLE CYCLE ADSORPTION AND REGENERATION RUNS ON CONCENTRATED ZINC CYANIDE WATERS
•vl
IS3
Tynical Concentrations, ppm
Cycle 1, 2460 gal
treated
Cycle 2, 1770 gal
treated
Cycle 3, 1800 gal
treated
Cycle 4, 3540 gal
treated
Cycle 5, 4100 gal
treated
Component
Cyanide , CM
Copper, Cu
Iron, Fe
Cyanide, CN
Copper, Cu
Iron, Fe
Cyanide, CN
Copper, Cu
Iron, Fc
Cyanide , CN
Copper, Cu
Iron, Fe
Cyanide, CN
Copper, Cu
Iron, Fe
In Feed
220-340
240-317
6-17
239-364
145-252
0.1-16.7
203-270
164-214
1.3-8.5
234-436
170-365
0.3-33
333-468
224-330
0.3-21
Effluent
Average
1
<1
2
30
<1
2
50
<1
2
<1
0.3
2-3
<1
0.3
Effluent at
Breakthrough
61
30
12
94
43
0.2
60
32
0.2
31
5.8
—
Weiaht,
Adsorbed
5.88
5.47
0.33
(4.93)
4.10
3.73
3.06
7.36
6.72
0.26
10.15
9.57
0.33
pounds
Stripped
1.67
5.06
0.03
1.05
3.09
0.01
0.97
2.87
0.18
0.89
5.59
0.13
—
Efficiency, percent
Adsorption
>99
>99
=90
>99
=80
>99
98-99
>99
98-99
>99
Receneration
28
92
10
21
75
26
94
12
83
50
—
-------
TABLE 29. CONDITIONS AND RESULTS OF REGENERATION FOR MULTIPLE CYCLE EXPERIMENT
u>
Cycle
No
la
Ib
2
3a
3b
4a
4b
Gallons
25
25
25
25
30
25
30
Regenerating Solution
Ib H2S04 Ib NaOH Hours
.12 0 9
16 0 104
Cycle
20 0 104
20 0 120
0 9(a) 2
Cycle
26 0 =200
0 10(a) 1
Cycle
Wash Water,
gallons
—
20
1 Totals
35
35
3 Totals
35
15
4 Totals
Amount Stripped, Ib
CN
1.43
0.24
1.67
1.05
0.52
0.45
0.97
0.52
0.37
0.89
Cu
0
5.06
5.06
3.09
2.79
0.08
2.87
5.42
0.17
5.59
Fe
0.005
0.029
0.034
0.011
0.026
0.150
0.176
0.031
0.102
0.133
(a) In these runs, a final caustic wash was given to the carbon bed after draining the acid
regenerant. Column pH finally was adjusted to 7.
(b) Samples of the loaded carbon near the top of the bed was removed after 24 hours had elapsed in
this regeneration cycle and analysed for total cyanide. The analvses showed 3.5 Ib CN/100 Ib carbon.
-------
performance was maintained provided the acid regeneration was followed
by a caustic wash prior to the next sorption step. Not enough cycles
were conducted during the campaign to fully evaluate the life of the
carbon bed, although these results indicate little or no diminution of
the affinity of the carbon for the copper cyanide complex.
Another aspect of the process which was not fully investigated involves
stripping of the copper and cyanide from the carbon during regeneration.
Complete stripping of copper generally occurred during all runs with a
minimum of 90 percent recovery. However, the stripping procedure re-
quired successively longer periods as the number of cycles increased,
an as yet unexplained effect. In addition to longer stripping periods,
the accountability of cyanide stripped from the carbon became progress-
ively poorer as the number of cycles increased. This effect was not as
noticeable in previous laboratory studies.
74
-------
SECTION VI
ECONOMIC EVALUATION OF THE CARBON SORPTION PROCESS
On the basis of pilot-scale studies conducted in Phase 2, capital and
operating costs for the carbon adsorption process were reevaluated.
The results of Phase 2 indicated the desirability of including several
modifications and/or alternatives which were not considered originally.
Two hypothetical treatment systems were considered: (1) a manually
operated system similar to the unit used in pilot-scale work, and (2)
a semicontinuous process designed to require a minimum of attention by
the user. Both systems were based on a wastewater flow of 10 gallons
per minute containing 300 ppm of cyaaiue. It is believed that both
systems can be applied successfully to the treatment of cyanide rinse
waters which contain cadmium or cooper as well as to the zinc rinse
waters used in the experimental work. Small amounts of other heavy
metals such as iron or chromium also may be present.
Flowsheets for these proposed carbon systems are illustrated in Figures
14 and 15. During operation, the rinse water first would be introduced
into the tank where the pH would be adjusted to a value of 10. Enough
copper sulfate would be added to completely complex any free cyanide in
the solution. After mixing for 10 to 20 minutes, the solution would be
neutralized to pH 7 with sulfuric acid. In the manual system (Figure
14), these steps all would be conducted in the feed collection tank by
a plating plant operator. In the continuous system (Figure 15), it is
believed that appropriate controls and instruments could be used to
conduct the steps automatically and continuously.
Following these initial steps, the solution would be allowed to settle
to minimize the amount of suspended solids sent to the carbon columns.
If the solids are settled, a sludge would be produced and allowed to
accumulate in the feed tanks for removal at weekly intervals. The precise
disposition of this sludge was beyond the scope of this work. It is be-
lieved that the sludge could be dewatered and disposed of without the
need for excessive additional treatment.
The clear effluent after settling then is passed through carbon adsorption
columns which remove all copper and at least 99 percent of the cyanide.
The treated effluent is discharged to the sewer.
Periodically, the carbon beds must undergo regeneration to remove the
load of adsorbed copper cyanides. It is believed that with some further
development work, the adsorption and regeneration cycle could be accom-
plished with a two-column system on weekly intervals. The regeneration
step produces a waste gas stream, an acidic solution of copper sulfate,
and a caustic solution containing small amounts of ferrocyanides if iron
75
-------
Cyanide
Rinse Water,
10 gpm
t><9
Feed
Collection
Tank
5000 gal
Sludge i
Dewatering '
(optional) '
7
A
0
1
O
Feed
Collection
Tank
5000 gal
S
T
— • J
<3
r
.h»
I
1 '
Carbon
Unit
(2 beds)
t 1
Feed Pump
Sludge
to
Disposal
H2S04
NaOH
Air
"Effluent to sewer
Regenerant
System
Vapors to Burner
CuSO.
4f
Recycle to feed
or recovery
FIGURE 14. SYSTEM I: FLOW DIAGRAM OF BATCH OR MANUALLY OPERATED CARBON PROCESS
-------
"J-QJL: CN. jcontr ol_|
~~> ph control i
Cyanide —
Rinse Water,
10 gpm
£
Complexing
pH 10
Neutrali-
zation
pH 7
I
I
Isiudge
Recycle
Carbon
Unit
(2 beds)
Effluent
"to Sewer
Sludge Dewatering
(optional)
NaOH
Air
J
1
Regenerant
System
Vapors
"to Burner
Sludge to Disposal
CuSO,
Recycle to Feed
or Recovery
FIGURE 15. SYSTEM II: FLOW DIAGRAM OF CONTINUOUS CARBON PROCESS
-------
is present in the feed solutions. The waste gases, which were found
in pilot-scale work to contain about 1 percent of hydrogen cyanide (HCN)
and varying amounts of carbon dioxide, would be passed through a burner
before being exhausted to the atmosphere. Spent caustic would be either
neutralized or fed back to the feed solution tanks. The copper sulfate
solution would be used as recycle for subsequent complexing of the feed
stream. Alternately, it is believed that a portion of the copper sulfate
could be sent to a recovery step if sufficient copper was originally
present in the cyanide rinse waters. This recovery was assumed to
provide a credit toward the cost of operating the process.
Equipment 'and operating costs were estimated for each of the two waste
treatment systems described above. As shown in Tables 30 and 31, the
estimated plant equipment costs are about $12,000 for the manual system
and about $16,000 for the continuous system. The estimate for the con-
tinuous operation is more speculative because the required procedures,
instrumentation, and controls have not been developed experimentally
during this program.
In summarizing operating costs, the important aspects are that the cost
of chemicals totals about $0.30 per pound of cyanide removed for each
of these processes. Estimated total operating costs (including labor,
power, fuel, supplies, maintenance, and amortization) range from $10 to
$14 per day. These estimates were based on the assumption that no
copper would be present in the feed rinse and that the amount of copper
used for complexing would be completely recovered during regeneration
and recycled within the process. The value of the copper sulfate used
in the process is considerable and could represent a credit to the
operation if sufficient copper were originally present in the feed rinse
waters. The net operating costs shown in this table indicate the effect
of applying a credit for copper sulfate recovery. It is significant
to note that net operating costs probably could be reduced nearly to zero
in the case of 16-hour-per-day operation for System II.
78
-------
TABLE 30. PURCHASED EQUIPMENT COSTS FOR EXAMPLE
CARBON SYSTEMS
Cost, $
System I: Manual Operation
Essential Equipment:
(1) Carbon unit; two columns, regenerant 7,000
tanks, feed pump, piping, and valves
(2) Feed tanks; two 5000 gallon 3,000
(3) Pump; 10 gpm 250
(4) pH meter
(4) Pump; 10 gpm
400
Subtotal 10,650
Optional Sludge Removal:
(1) Filter i.000
(2) Pump 25°
Total 11,900
System II: Continuous Operation
Essential Equipment:
(1) Carbon unit; two columns, regenerant 7,000
tanks, feed pump, piping, and valves
(2) Tanks; two 250 gallon, agitated, and 1,500
one 2000 gallon
(3) Reagent feeders; two each 750
25
Subtotal 9,500
Optional Sludge Removal:
(1) Filter i'000
(2) Pump 25°
Instrumentation and Controls:
,—
Total 15,750
79
-------
TABLE 31. ESSENTIAL OPERATING COSTS FOR EXAMPLE CARBON SYSTEMS
oo
o
System I Cost, $/dav
Chemicals Consumed:
(1) H2S04 (complexing) 3
3c/lb
(2) H2S04 (regeneration)
3c/lb
(3) NaOH (regeneration)
Other Operating Costs:
.3 Ib/lb CN @
4 Ib/lb CN @
1 Ib/lb CN @ 6.5c/lb
Subtotal
(1) Labor @ $3.50/man-hr
(2) Power @ $0.010/kwhr
(3) Fuel @ $0.25/106 Btu
(4) Supplies and maintenance @ 0.003% of plant
(5) Amortization @ 0.0224% of plant cost
Total operating cost: $/day
Cyanide removal, lb CN/day
Copper sulfate recovered: lb
Possible credits @ $0.345/lb
Net Operating Cost: $/lb CN
=====================^
(3) up t^^ppm^peV0^
Subtotal
CuS04/day^
CuS04: $/day
>er was originally pres
8-hr day
1.20
1.44
0.78
3.42
3.50
0.20
0.30
cost 0.36
2.69
7.05
10.47
12
0-23
0-7.93
0.21-0.87
ent in the rinse water
System II
8-hr' day
1.20
1.44
0.78
3.42
1.75
0.40
0.30
0.47
3.53
6.45
9.87
12
0-23
0-7.93
0.16-0.82
Cost, $/day
16-hr day
2.40
2.88
1.56
6.84
1.75
0.80
0.60
0.47
3.53
7.15
13.99
24
0-46
0-15.86
0-0.58
and might range from 0 ppm
-------
SECTION VII
ACKNOWLEDGEMENTS
This research program was conducted during the period of April, 1968,
through May, 1971. Battelle personnel participating in the program
were A. K. Reed, T. L. Tewksbury, J. F. Shea, R. G. Brown, J. G. Price,
M. F. Nichols, R. H. Cherry, Jr., and G. R. Smithson, Jr.
The cooperation and assistance of the following is gratefully
acknowledged:
Metal Finishers' Foundation Pollution Abatement Committee:
Walter V. Turner (Eric S. Turner & Co., New Rochelle, N.Y.)
Chairman
A. T. Leonard (Superior Plating, Inc., Minneapolis, Minn.)
Charles Levy (Swift Laboratories, Inc., Waltham, Mass.)
Raymond E. Morris (Remco Finishing Corp., Fernwood, Pa.)
A. T. Marinaro (Masters' Electro-Plating Association, Inc.,
Long Island City, N.Y.)
C. W. Shriver (Superior Plating Co., Columbus, Ohio)
P. Peter Kovatis (National Association of Metal Finishers)
Mr. William J. Lacy, Office of Research and Monitoring, EPA
Mr. Edward Dulaney, Office of Research and Monitoring, EPA
Mr. John Ciancia, Office of Research and Monitoring, EPA.
81
-------
APPENDIX
CONVENTIONAL METHODS FOR TREATMENT OF CYANIDE WASTEWATERS
-------
SECTION VIII
APPENDIX
CONVENTIONAL METHODS FOR TREATMENT OF CYANIDE WASTEWATERS
Because of the toxicity of most cyanide compounds, the elimination of
cyanides from rinse waters produced in metal-finishing operations
has been standard practice for many years in the larger installations.
The techniques which have been developed can be categorized as follows:
(1) Complete destruction of the cyanide ion with the formation
of nitrogen,
(2) Conversion of the cyanide ion to the considerably less toxic
cyanate ion, and
(3) Conversion of the cyanide ion to some other less toxic form
such as the ferrocyanide ion.
Complete Destruction of Cyanide
Cyanide in rinse waters can be completely destroyed by treatment with
chlorine gas in an alkaline solution at room temperature. The
cyanide radical, CN, is disrupted with the carbon fragment being
converted to carbonate and the nitrogen to nitrogen gas. The reaction
involved is:
ZNaCN + 5C12 + 12NaOH N2 + 2Na2C03 + lONaCl + 6H20.
Sodium Chlorine Sodium Nitrogen Sodium Sodium Water
cyanide hydroxide carbonate chloride
In the chlorination step, which is conducted under highly alkaline con-
ditions, most of the heavy metals that accompany cyanide in the rinse
waters precipitate as hydroxides, or possibly as ferro or ferricyanides.
These latter compounds respond slowly to the chlorine treatment and, if
permitted to settle out, probably receive little or no treatment. For
this reason, vigorous agitation is required in the process.
83
-------
The process can be carried out on a batch or continuous basis.
If done batchwise, it requires the installation of rather large
tanks for storing the rinse waters before treatment and usually re-
quires the installation of duplicate tankage—each capable of undergoing
treatment while the other is filling. Recommended tank sizes vary,
ranging in capacities corresponding to from 2 to 24 hours of opera-
tion. If done continuously on a flow-through basis, this process
must be instrumented to control the reagent additions and the quality
of the effluent.
Sludge formation nearly always accompanies the chlorination process.
The sludge is composed of the precipitated metal hydroxides and,
if the lime is used to provide the alkalinity, it also contains
calcium carbonate and possibly calcium sulfate. In some cases the
volume of sludge produced may be too great for disposal to sewers and
provisions for dewatering the sludge must be incorporated in the
plant.
The suitability of this process for the smaller plater has been
questioned by some authorities, primarily because of the potential
hazards, the difficulty in handling, metering, and distributing the
chlorine gas, and the need for maintaining the proper conditions of
alkalinity in the solution to prevent formation and evolution of
poisonous cyanogen chloride. Nevertheless, this process is widely used in
the large metal-finishing plants and is described frequently in the
literature.
Cyanide also may be completely destroyed by hypochlorites such as
sodium hypochlorite (NaOCl), calcium hypochlorite [Ca(OCl)2], or
bleaching powder (CaOC^) . An example of the probable reaction—
analogous to the reactions with chlorine gas in an alkaline medium—
is shown in the following equation:
2NaCN + 5NaOCl + H20 •* N2 + 2NaHC03 + 5NaCl .
Sodium Sodium Water Nitrogen Sodium Sodium
cyanide hypochlorite bicarbonate chloride
The process is relatively simple and consists essentially of adding the
hypochlorite, either as a solution or as a solid to the rinse water.
No additional alkalinizing agent, such as sodium hydroxide or hydrated
lime, is required, as in the chlorine gas process.
Agitation also is required with hypochlorites. The process with
hypochlorites may be conducted batchwise or continuously. This process
probably is best suited to the needs of the smaller plater who is faced
with the necessity for complete destruction of cyanide.
-------
Advantages of hypochlorite processes over the chlorination process
are:
(1) If sodium hypochlorite or calcium hypochlorite are used, the
theoretical chlorine consumption is only half as great as that if
chlorine gas is used. If bleaching powder is used, the theoretical
chlorine consumption will be the same as for the chlorine gas process,
(2) Handling and metering of the hypochlorites is relatively simple
and nonhazardous.
(3) The reaction !.s more rapid than that with chlorine.
(4) If sodium hypochlorite is used, sludge production is minimized.
This process also is used commercially.
Conversion of Cyanides to Cyanates
The cyanate radical (CNO) reportedly is only about one-thousandth as
toxic as cyanide, and many regulatory agencies will accept rather high
cyanate concentrations in effluents. The conversion of cyanide to
cyanate by chlorine gas is a rapid reaction, requiring only minutes.
The reaction which is an intermediate stage in the process for the com-
plete destruction of cyanide by chlroine is:
NaCN + C12 + 2NaOH -»•' NaCNO + ZNaCl + H20 .
Sodium Chlorine Sodium Sodium Sodium Water
cyanide hydroxide cyanate chloride
The process can be carried out in the same type of equipment as that in-
volving the complete destruction of cyanide by chlorine. It also may
be conducted batchwise or continuously. Since sludge formation occurs
in this process, too, provisions for removing the solids from the treated
liquor may be required.
Hypochlorites also can be used for converting cyanides to cyanates by a
similar reaction:
NaCN + NaOCl -> NaCNO + NaCl .
Sodium Sodium Sodium Sodium
cyanide hypochlorite cyanate chloride
Other hypochlorites, such as calcium hypochlorite or bleaching powder
can be used in place of sodium hypochlorite. The process is straight-
foward and rapid.
85
-------
For the smaller plater aiming at the partial destruction of cyanide,
i.e., the conversion of cyanide to cyanate, the hypochlorite method
probably is more suitable than the chlorine method.
Among the other oxidants which have been proposed for the conversion
of cyanides to cyanates is hydrogen peroxide.
Hydrogen peroxide reacts with cyanides to form cyanates according to
the reaction:
NaCN
Sodium
cyanide
Hydrogen
peroxide
NaCNO
Sodium
cyanate
Water
Hydrogen peroxide has been used in Europe for treating cyanide wastes
from metal- finishing plants. It has an advantage over chlorine or
hypochlorite in that additional pollutants (chlorides) are not intro-
duced. However, its use is more expensive .than the alkaline chlorination
and usually somewhat more; expensive than hypochlorite treatment.
Other Methods for Treating Cyanide Wastes
The formation of less toxic cyanide complexes such as ferro and
ferricyanides also has been used as a method for disposing of cyanide
wastewaters. This process involves the use of iron salts to form
complex compounds with the free cyanide in the wastes. Eventually
these cyanide complexes are precipitated and removed as a sludge. The
use of ferrous sulfate, for example, produces the characteristic dark
blue sludge or Prussian blue.
The major advantage of this treatment method is that it is relatively
inexpensive where waste ferrous sulfate is available. However, con-
siderable quantities of sludge may be formed and the treated solutions
are strongly colored. There also is evidence that ferrocyanides may
be decomposed to free cyanide by sunlight. The regeneration of the
cyanide under these conditions would contaminate the receiving stream.
This method has received very little acceptance by the industry in this
country, but appears to be used in Europe. The complexing process
apparently does not completely destroy cyanide under practical operating
conditions. Cyanide levels in treated solutions may be as great as 5
to 10 ppm.
Other methods which are described in the literature for the treatment
of cyanide wastes include: the formation of cyanide complexes by the
use of polysulfides to form relatively nontoxic sulfocyanates, the
86
-------
conversion of cyanide to cyanate by ozone, the complexation of cyanide
by nickel salts to form the highly stable nickel cyanide complex, the
biological destruction of cyanide on trickling filters, the destruction
of cyanide by irradiation, and the use of techniques such as dialysis and
reverse osmosis.
These processes are in various stages of development. Some, such
as the use of polysulfides to complex cyanides, or the use of perman-
ganate to destroy cyanides, have been reduced to commercial practice.
Some have been merely suggested, i.e. , the complexation of cyanide by
nickel.
«U.S. GOVERNMENT PRINTING OFFICE: 1972 484-483/13 1-3
87
-------
1
w
2 i Subject Field 6s. Group
05D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Metal Finishers' Foundation, Upper Montclair, New Jersey
Title
An Investigation of Techniques for Removal of Cyanide from Electroplating Wastes
10
Authors)
A. K. Reed
J. F. Shea
T. L. Tewksbury
R. H. Cherry, Jr.
G. R. Smithson, Jr.
•tf. I Project Designation
Project No. 12010 EIE
JVoto
This report, 12010 EIE 05/71 is the final
report on Phase II Cyanide System Work. Phase I
report 12010 EIE 11/68, "A State-of-the-Art Review
of Metal Finishing Waste Treatment is available from
GPO
22
Citation
Battelle, Columbus Laboratories, Columbus, Ohio
22 I Descriptors (Starred First)
*Activated Carbon,*Cyanide, *Waste Water Treatment
25
Identifiers (Starred First)
*Electroplating Wastes
27
Abstract
This report describes work which was conducted on the removal of cyanide from
plating rinse waters employing various treatment processes. The study con-
sisted of an initial phase in which information was sought by questionnaire
and by wastewater analyses on the type of waste produced by smaller electro-
plating plants. Laboratory studies were conducted on several nonconventional
methods for treatment of these wastewaters, including ion flotation, and sol-
vent extraction. A demonstration pilot-plant study also was conducted on the
activated carbon process employing actual rinse waters from a zinc cyanide
plating operation.
The results of the various phases of the study indicated that activated carbon
adsorption for cyanide removal may have practical application in many small
plating plants. When combined with pretreatment stages for solids removal,
effective elimination of heavy metals also can be expected. Further develop-
ment of the process was recommended in actual plating plant installations.
This report was submitted by Battelle's Columbus Laboratories in partial
fulfillment of Grant Project #WPRD 201-01-68, Program #12010 EIE by the
Water Quality Office, Environmental Protection Agency to the Metal Finishers'
Foundation.
/n.s-?j/u//on
SEND. Wl TH CO:>Y
^- DOCUMENT TO: WATtR ^tSCL/RCES SCIENTIFIC INFORMATION CENTER
V'.S. DEPARTMENT OF THE INTERIOR
WASHING! ON, D. C. 20240
GPO: 1070 - 407 -8»t
------- |