United States
Environmental Protection
Agency
Industrial Environmental Research
Laboratory
Research Triangle Park NC 27711
Research and Development
EPA-600/S2-83-066 Dec. 1983
v°/ERA Project Summary
Cyanide Removal from Coke
Making and Blast Furnace Waste
Waters
G.W. Lower and D.J. Spottiswood
The objective of this research program,
supported jointly by the American Iron
and Steel Institute and the U.S. Envi-
ronmental Protection Agency (EPA),
was determining the feasibility of
removing cyanide from coke making
and blast furnace waste waters by ion
flotation or column precipitate flotation
of iron ferrocyanides. Ion flotation was
reasonably effective on ferricyanide,
but not on cyanide or ferrocyanide;
therefore, efforts were concentrated on
the formation and flotation of iron
ferrocyanide precipitates. (Note: A
readily available source of ferrous iron is
waste pickle liquors.)
An experimental program was de-
signed for precipitate flotation to
evaluate the effect of operational
variables (iron addition, reaction time to
form the iron cyanide complex, precipi-
tation time to form the iron precipitate,
collector type and dosage, conditioning
time, flotation time, and air flow
rates). Synthetic ferrocyanide solutions,
free cyanide solutions, and two coke
plant waste waters [crude ammonia
liquor (CAL) and intercepting sump
water (ISW)] were tested, both in batch
and continuous column flotation tests.
Wet oxidation tests were conducted on
the froth product
Results of the tests showed that 95-
99% cyanide could be recovered from
the synthetic solutions containing 100
mg/l cyanide, and 91% from ISW. The
most effective flotation reagent was a
primary amine (dodecylamine acetate),
and the most important variable in the
process was the pH of flotation. A set of
operating conditions (iron addition,
collector addition, pH, reaction time,
conditioning time, and flotation time)
were developed for both the synthetic
solutions and a coke plant ISW. Prelim-
inary wet oxidation tests indicated
that the froth product could be converted
to ammonia and ferric oxide by wet
oxidation and the solid product would
meet EPA/Office of Solid Waste
extraction procedure (EP) toxicity
standards. Results showed that precipi-
tate flotation could be used as a primary
process to remove most of the cyanide
and could meet effluent limitations
under certain conditions depending on
feed concentration and volume.
Ion exchange tests on synthetic
solutions produced effluents which met
discharge standards. This may be one
possible secondary method of treatment.
However, no work was done on flotation
effluents or plant liquors.
It was also noted in this test program
that the iron-iron cyanide precipitate
settled quite rapidly once it was formed.
Although no work was done in this area,
most of the cyanide could possibly be
removed by precipitation and thickening,
followed by wet oxidation of the
thickener underflow.
This Project Summary was developed"
by EPA's Industrial Environmental
Research Laboratory, Research Triangle
Park, NC, to announce key findings of
the research project that is fully
documented in a separate report of the
same title (see Project Report ordering
information at back).
Introduction and Summary
Fifty years ago, coke production from
by-product ovens surpassed coke produc-
tion from beehive ovens, the advantages
of by-product processing were that it
reduced the air pollution problems
-------
associated with coking, and recovered
marketable by-product chemicals. How-
ever, the gas processing introduced a
significant water pollution problem. The
problem was formally defined by the U.S.
EPA, following passage of the Federal
Water Pollution Control Act Amendment
in 1972 (P.L. 92-500). In May 1982, the
EPA issued final regulations defining iron
and steel manufacturing point source
category effluent limitations guidelines
(Federal Register Vol. 47, No. 103, pp.
23135-23404). For by-product coke
making, these regulations set a 30-day
average limitation of 0.00351 kg of
cyanide per tonne of coke produced by
application of the best technology eco-
nomically achievable (BAT).
The level of cyanide in coke plant waste
water streams varies from 20 to 400
mg/l. In addition to the cyanide, the plant
liquors contain fixed and free ammonia,
free and emulsified oils, phenols, and
various suspended solids. These con-
taminants are the by-product of pyrolysis
reactions during the carbonization of
coal. The process waters are contaminated
during the processing of the coke oven
gas for the recovery of coal chemicals.
The plant contains three major process
streams: crude ammonia liquors (CAL),
barometric condenser liquors (BCL), and
intercepting sump water (ISW). The CAL
(also referred to as waste ammonia
liquors) are the condensed water vapors
separated from tar in the tar plants. The
BCL result from the direct contact of
cooling water with vapors released in the
crystallizing and concentrating of ammonia
sulfate by vacuum evaporation. The ISW
is a combination of the water from the
oil/water separators and the various
plant drains.
Cyanide concentrations in the combined
effluent flow vary from plant to plant,
depending on the total effluent volume. In
one plant, producing 5442 tonnes (6000
short tons) of coke per day, the total
effluent volume is 1452 I/tonne of coke,
and the effluent cyanide concentration
is 50.3 mg/l. BAT limitations would
require reducing the effluent cyanide
concentrations to no more than 2.42
mg/l. In a second plant, process improve-
ments drastically reduced the total
effluent volume, primarily by altering the
ammonium sulfate condenser system: in
this plant, producing 2721 tonnes (3000
short tons) of coke per day, the total
effluent volume is 681 I/tonne of coke,
and the cyanide concentration is 56.8
mg/l. Therefore the BAT effluent cyanide
limitation would be no greater than 5.16
mg/l.
The primary objective of this study was
to determine the feasibility of removing
the cyanide from these waste waters by
either ion flotation or by precipitate
flotation of an iron-iron cyanide precipitate.
Secondary studies were conducted on ion
exchange removal of ferrocyanide.
The objective of the ion flotation
studies was to complex the various
cyanide species with a quaternary amine
and then remove the resulting hydro-
phobic complex by flotation. The precipitate
flotation phase involved complexing the
cyanide with ferrous iron, precipitating
the ferrocyanide complex as an iron
ferrocyanide, and removing the precipitate
(and thus the cyanide) by microparticulate
flotation. (Note: Previous investigators
have shown that (1) when ferrous iron is
added, the cyanide level is at minimum at
about pH 8 as a result of complex
formation and precipitation of Fe2Fe(CN)e;
(2) the rate of complex formation is
greatest above pH 7.5; and (3) precipitated
iron cyanide can be floated with a cationic
collector at pH 6. Consequently, current
efforts were directed toward determining
the conditions necessary for the formation
of ferrocyanide complexes, precipitation
of the complexes, and flotation of the
precipitate.
Experimental
Continuous ion flotation and batch
precipitate flotation tests were carried
out in a 4.7 cm diameter by 70.0 cm tall
column equipped with a froth overflow
launder. Continuous precipitate flotation
tests were run in a 25 mm (1 in.(column
shown schematically in Figure 1.
Ion flotation tests were run on synthetic
solutions of K4Fe(CN)6 containing 36.8
mg/l Fe(CN)e3by dispersing the amine in a
small volume of the feed solution prior to
its addition to the bulk of the feed
solution. Following a conditioning period
(10-12 minutes), the feed was pumped
through the flotation column countercur-
rent to the air flow, and the froth product
was removed.
Batch precipitate flotation tests were
run on 11 samples of synthetic solutions
of K4Fe(CN)6 and KCN as well as on coke
plant effluents. The solutions were
reacted in a conditioning vessel with
FeS04-7H2O, the pH adjusted, the collector
added immediately, and the slurry
conditioned. The slurry was then pumped
into the column and floated at an air flow
rate of 0.4 volumes air per volume of
solution. The synthetic solutions contained
100 mg/l total cyanide, and some
solutions also contained 5000 mg/l NaCI.
Continuous precipitate flotation tests
were run on synthetic K4Fe(CN)6solutions
by first adding the amine, A-336, to
complex the ferrocyanide, and then
adding FeSO4-7H2O to form the precipitate.
Ion exchange tests on synthetic K4Fe
(CN)s solutions containing 25 mg/l total
cyanide plus 5000 mg/l NaCI were run in
a 40 cm diameter column containing 200
ml of wet settled resin (Rohm and Haas,
IRA-958). All cyanide assays were
performed using the standard ASTM
method for cyanide in water.
Ion Flotation
Distribution tests using a chloroform
solution containing 1% by volume of a
quaternary amine, Aliquat-336 (General
Mills, A-336), demonstrated that the
amine would complex cyanide, ferrocya-
nide, and ferricyanide. Ferricyanide had
the highest distribution coefficient.
Consequently, ion flotation tests were
run on synthetic ferricyanide solutions.
Flotation results as a function of the
system variables were as follows:
Amine Concentration
In chloride-free solutions, cyanide
recoveries reached at maximum of about
83% at an amine/ferricyanide mole ratio
of 3.75/1. In high chloride solutions,
recoveries continually increased but
were lower than in chloride-free solutions
except at very high concentrations of
amine.
Flotation Variables
Increased conditioning times increased
recovery until a steady state was reached
at about 13 minutes. Variation of the
flotation time from 1.07 to 2.3 minutes
increased recoveries from 75 to 83%,
whereas changes in the air flow rate from
0.08 to 0.22 l/min/cm2 had only a minor
effect.
Feed Concentration
Nearly constant recoveries of about
82% were obtained over a feed concen-
tration range of 30-70 mg/l ferricyanide.
Recoveries decreased belowthis range to
about 65% at 5.0 mg/l ferricyanide. The
pH of the solution had little effect over a
pH range of 4-7.
Limited work on ion flotation of
cyanide and ferrocyanide was discontinued
because cyanide did not float well and
ferrocyanide required very long condi-
tioning times. Consequently efforts were
directed toward precipitate flotation of
iron ferrocyanides.
-------
Precipitate Flotation
Batch precipitate flotation was studied
initially on synthetic solutions of ferro-
cyanide, followed by studies on free
cyanide solutions and on plant liquors.
Ferrocyanide Solutions
Initial tests were run on ferrocyanide
solutions to study the effect of iron
addition, collector addition, and pH. The
collector used was a quaternary amine,
Aliquat-336 (General Mills, A-336).
Conditioning time was 15 minutes, and
flotation time was 5 minutes. Two levels
of iron, 7 molar (90 mg/l) and 70 molar
(144 mg/l) percent excess (based on
formation of Fe^FefCNJeh were used at
collector levels of 283 and 567 mg/l A-336
and pH values of 3.0 and 7.0. Results
showed that pH was the most important
variable. Flotation, using 144 mg/l iron
and 283 mg/l A-336 over a pH range of 3-
8, gave essentially constant recoveries of
90-95% from pH 3 to 6. Above pH 6,
recovery decreased rapidly.
Addition of collector above 100 mg/l A-
336 gave relatively constant recoveries of
over 90% at pH 4 using 144 mg/l Fe and 5
minutes conditioning time. Below 100
mg/l A-336, recovery dropped off rapidly.
In solutions containing NaCI, higher
collector additions were required. More-
over, recovery increased steadily from
85% to 95% as the collector or concentra-
tion was varied from 50 to 800 mg/l.
Free Cyanide Solutions
For free cyanide solutions, a complexing
reaction must take place prior to the
precipitation reaction. Reaction times of 5
and 10 minutes, after addition of iron but
before collector addition, were studied at
two iron levels — 144 and 215 mg/l
using 283 mg/l A-336 at pH 4. Both the
conditioning and flotation times were 5
minutes. At the lower iron level, recoveries
increased with increased reaction time;
however, at the higher iron level,
increased reaction time had little effect.
Similar results were obtained on chloride-
containing solutions, although overall
recoveries were slightly lower.
Collector Type
A second cationic collector, dodecyla-
mine acetate (12-D), and an anionic
collector, sodium lauryl sulfate (NLS),
were investigated as a function of pH at
an iron level of 215 mg/l. Collector levels
were 31 mg/l for 12-D and 45 mg/l for
NLS. The effectiveness of the NLS fell off
sharply above pH 6, indicating that above
this pH the particles are negatively
charged. In contrast, the positively
charged 12-D was effective up to pH 9
with recoveries as high as 98%. This
primary amine was more effective than
the quaternary amine A-336, particularly
above pH 7. In addition, much lower
levels of the primary amine were required:
recoveries of over 97% were obtained
with additions of only 24 mg/l of 12-D.
The froth volume and stability was good
over all pH ranges with 12-D; whereas,
little froth was obtained above pH 7 with
A-336.
The addition of a frother, 20 mg/l of 2-
methyl-4 pentanol (MIBC), gave 97%
recovery with only 12 mg/l of 12-D.
Plant Liquors
A drum of intercepting sump water
(ISW) and a drum of crude ammonia
liquor (CAL) were used for the tests. The
CAL was dark colored and had a strong
odor due to high concentrations of
ammonia and phenol. The total cyanide
content was 10 mg/l. The ISWwas a dirty
brown color and contained 89 mg/l total
cyanide.
IS W Liquors - Effect of pH
Since the work on synthetic solutions
showed that pH was the most important
variable, a series of tests were run on the
ISW at various pH levels. Test conditions
were 24 mg/l of collector 12-D, 200 mg/l
of iron, and reaction, adsorption, and
flotation times of 15, 10, and 5 minutes,
respectively. Recoveries increased with
increasing pH to a maximum of 91'% at
about pH 7. Variations in reaction time
over the range of 5-30 minutes yield
nearly constant recoveries of about 91%.
Recoveries decreased with reaction
times longer than 30 minutes.
CAL Liquor
The CAL I iq uor was floated at pH 7 with
24 mg/l of 12-D and 200 mg/l iron.
Reaction, adsorption, and flotation times
of 15, 5, and 5 minutes, respectively,
were used. Cyanide recovery was about
40%. Additional coagulation resulted
when amine was added: these solids
were also reported in the froth product.
Flotation under the same conditions,
except without amine, resulted in a 60%
recovery. This indicates that the oils
present in the liquor may be acting as
collectors.
The above results indicate that a batch
flotation process:
(1) Will remove over 98% of the
cyanide from synthetic solutions
containing 100 mg/l cyanide.
(2) Will remove over 90% of the
cyanide from plant liquors contain-
ing approximately 90 mg/l cyanide.
If the same percentage removal can
be achieved at feed concentrations
of the order of 50 mg/l, the higher
of the two effluent standards can be
met.
(3) Has the potential of a primary
process for bulk removal of cyanide
and possibly a process for meeting
effluent standards if additional
flotation stages are added. In
addition the underflow is relatively
clean, indicating the possible re-
moval of other contaminants.
Continuous Precipitate
Flotation
Continuous precipitate flotation tests
were run on synthetic K4Fe(CN)esolutions
using reverse addition of amine and iron;
i.e., the amine, A-336, was added as a
complexing agent followed by the addition
of iron as FeSO4-7HgO. Initial two-phase
batch extraction tests using CCU as the
amine solvent indicated an average
3.5/1 amine/ferrocyanide stoichiometry.
Dispersion of the amine in ferrocyanide
solution produced a finely dispersed wax.
Subsequent addition of iron produced a
floatable amine-iron-cyanide precipitate.
Minimum iron addition was found to be
250 mg/l. Results of batch tests on
solutions containing 10 and 75 mg/l
cyanide showed recovery was not mater-
ially affected by amine additions greater
than a 3/1 mole ratio of amine/ferro-
cyanide. Using an amine/ferrocyanide
ratio of 3/1 plus 250 mg/l iron, very high
cyanide recoveries (95-98%) were ob-
tained in batch tests on feed solutions
containing cyanide.
Using the above system of reverse
amine-iron addition, continuous tests
were run in the system shown in Figure 1.
Column performance (1 in. column) was
not affected significantly by changes in
feed flow rate, specific air rate, or
retention time.
The effect of feed concentration on
cyanide effluent concentration is shown
in Figure 2. The effluent limitations of 2.4
and 5.1 mg/l for the operation described
earlier are shown by the dotted lines.
These results indicate that:
(1) Substantial removal of ferrocyanide
can be achieved.
-------
Head
Tank
Feed
Tank
A-336 Fe
Filter
Regulator
Ml*
^ iin
Conditioning Tanks
Manometer
Feed
Flowmeter
Feed
Valve
o Valve
L {J}«
Underflow Out
Figure 1. Continuous column flotation system.
\ Feed
r*\ Pump
18
15
*"
s
I
C/V. Effluent Limit
57 mg/l
24 mg/l
0 25 50 75 WO
Cyanide in Feed, mg/l
Figure 2. Effluent cyanide performance of
flotation column.
(2) For feed cyanide concentrations
less than 62 mg/l, the effluent
limitation of 5.1 mg/l can be met,
and for feed cyanide concentrations
less than 35 mg/l the lower
limitation of 2.4 mg/l can be
achieved.
Disposal of Flotation
Concentrate
Wet oxidation of the flotation concen-
trate using the Wetox® Process appears
to be a technically feasible process for
detoxifying the concentrate. In this
process, cyanide is converted to NHaand
iron to iron oxide. Preliminary tests on a
concentrate containing 2000 mg cyanide
produced a liquor containing less than
1.0 mg/l of cyanide and an off-gas
containing less than 0.5 mg cyanide.
Further work in this area could optimize
the process.
The levels of all metallic contaminants
(D004 to D011) in the solid residue were
less than the maximum allowable con-
centrations as determined by EP Toxicity
Test Procedure (Federal Register, Vol. 45,
No. 98, May 19, 1980).
Ion Exchange
A limited series of tests were run on
ferrocyanide solutions containing 25
mg/l cyanide using IRA-958. The break-
through point was arbitrarily set at an
effluent concentration of 0.25 mg/l
cyanide. Successive cycles were run
with elution after each cycle using 15%
NaCI solution at pH 12. Excellent
adsorption was obtained at flow rates of
19.4 and 16.0 bed volumes per hour.
Volume throughputs to breakthrough
were about 380 at 19.4 bed volumes per
hour and 500 at 16.0 bed volumes per
hour. Very little decrease in throughput
was noted over three cycles. No tests
were run on plant liquors; therefore, the
effect of other anions present in these
liquors is unknown.
Recommendations
Further investigation of this process for
the removal of cyanide from coke making
and blast furnace waste waters by
flotation of iron cyanide precipitates
would be of value, particularly if it
involved semi-pilot scale tests on a
multistage continuous basis on plant
liquors. In addition to column flotation,
other flotation processes (e.g., submerged
air flotation) bear investigating. Further
studies on wet oxidation of the flotation
concentrate would help determine opti-
mum operating conditions.
Conclusions
The results of experimental studies on
the removal of cyanide from coke making
and simulated blast furnace waste
waters by amine flotation of iron ferro-
cyanide precipitates indicate that most of
the cyanide can be removed by this
process. In certain cases, depending on
the feed cyanide concentration and feed
volume, effluent limitations can be met.
Cyanide recoveries of 95-99% were
obtained from both ferrocyanide solutions
and non-complexed cyanide solutions
containing 100 mg/l cyanide plus high
NaCI concentrations. Cyanide recoveries
of as high as 91% were obtained from a
plant liquor (intercepting sump water)
containing 89 mg/l cyanide.
The most important variable in the
process was pH, and the most efficient
collector was a primary amine (dodecyla-
mine acetate) at pH 8. A quaternary
amine (Aliquat-336) was effective up to
pH 6, but recoveries dropped sharply at
higher pH values.
The cyanide in the froth product can be
effectively destroyed by wet oxidation.
Ion exchange using a strong base resin
was capable of producing an acceptable
effluent (less than 0.25 mg/l cyanide)
from synthetic solutions containing 25
mg/l cyanide as ferrocyanide plus 5000
mg/l chloride; however, the effectiveness
of this process as a secondary stage
following flotation was not determined.
-------
G. W. Lower is with Michigan Technological University, Houghton, Ml 49931, and
D. J. Spottiswood is presently with Colorado School of Mines, Golden, CO
80401.
Robert C. McCrillis is the EPA Project Officer (see below).
The complete report, entitled "Cyanide Removal from Coke Making and Blast
Furnace Waste Waters," (Order No. PB 83-259 671; Cost: $ 10.00, subject to
change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
1W
w \
\\\\
PS 0000329
U S ENVIR PROTECTION AGENCY
REGION 5 LIBRARY
330 S DEARBORN STREET
CHICAGO IL 60604
US. OFF
)P£NAPY
Fi*
FfMTt
USE S3QQ
*
=ICIAL MAIL
liS-PDbTAGE
- n *, *,
~ U,u /
SMFfTs
503205
,
4-
* US GOVERNMENT PRINTING OFFICE- 1984-759-102/818
------- |