COST-EFFECTIVENESS RELATIONSHIPS FOR THE REMOVAL
OF CADMIUM, MERCURY, AND CYANIDE FROM INDUSTRIAL WASTESTREAMS
Preliminary Analysis
For
Economic Analysis Division
Environmental Protection Agency
Washington, D. C.
By
John b. _Kooo^ Project Manager
Y. Argaman, Project Engineer
Cheryl D. Magee, Project Engineer
Carl E. Adams, Jr., Project Consultant
ASSOCIATED WATER AND AIR RESOURCES ENGINEERS, INC.
2907 12th Avenue South
Nashville, Tennessee 37204
November, 1973
-------
COST-EFFECTIVENESS RELATIONSHIPS FOR THE REMOVAL
OF CADMIUM, MERCURY, AND CYANIDE FROM INDUSTRIAL WASTESTREAMS
Scope of the Investigation
The objective of this work was to estimate the cost of removing
cyanide, mercury, and cadmium from industrial wastes and to obtain
estimates of the residual concentrations which can be obtained using
best available treatment. Because estimates were assembled from readily
available literature data and from communication with researchers and
equipment manufacturers, the values used in this report cannot be taken
as the most accurate estimates available. Considerable more time in
evaluating data from all sources would be required to arrive at final
judgements. Treatment costs were obtained from industrial waste treat-
ment cost relationships, costs for municipal wastewater treatment
processes, and from the best judgement of equipment manufacturers. In
many cases it was necessary to use attainable effluent concentrations
given by researchers and equipment manufacturers without having access
to the detailed conditions under which the results were obtained.
Several assumptions were made regarding the wastestream to be
treated. These conditions may be enumerated as follows:
1. Waste flow = 3 mgd.
2. Influent concentrations of the species of concern = 2 mg/1.
3. Waste constituents to be removed con not be isolated in more
concentrated streams. Therefore, it was assumed necessary to
1
-------
treat the entire waste flow.
4. The wastestream would contain some suspended solids and organic
constituents ncbt removed in upstream treatment processes.
5. The effect of complexing agents could not be thoroughly con-
sidered due to lack of applicable information.
Analytical Limits of Detection
In order to add perspective to the residual concentrations of these
toxic substances reported, the limits of detection for each constituent
were reviewed. Briefly, the findings are as follows:
1. Mercury. The flameless atomic absorption procedure recommended by
the EPA is capable of detecting mercury at the 0.01 ppb level; EPA,
however, recommends that 0.2 ppb be set as the practical detection limit
for mercury in natural waters.
2. Cadmium. The detection limit for cadmium in water when determined
by flame atomic absorption at 228.8 nm is about 40 ppb, using the Boling
burner as recommended in Standard Methods (13th Ed). An adaptation
known as the Delves cup method, which involves analysis of evaporated
samples, permits detection limits for blood samples (and presumably for
wastewater samples) of approximately 2 ppb. Methods based on atomic
fluorescence and neutron activation are capable of detection limits
several orders of magnitude lower, but their suitability for wastewater
analysis has not been established.
3. Cyanide. The most sensitive practical methods for cyanide are color-
imetric. The procedure suggested by Standard Methods involves conversion
2
-------
of cyanide from a distilled sample to cyanogen chloride, which then
reacts with a pyridine-pyrazolone reagent to form a blue dye. The
latter is determined spectrophotometrically at 620 nm. Standard
Methods gives the effective range as 1 - 5 pg in the wastewater; this
would correspond to 40 - 200 ppb, depending on how the sample was pre-
pared. A five-fold increase in sensitivity is said to be achieved by
extracting the dye with butyl alcohol.
Treatment Processes
Treatment processes which were considered for the removal of each
ion are discussed below. This discussion is limited to process para-
meters which have the greatest effect on treatment performance.
A. Cadmi um
1. Precipitation. Precipitation of CatOHjg by lime treatment at
pH 11 is expected to reduce the cadmium concentration from 2.0 nig/1
to about 0.2 mg/1. The lime dose, which depends on the waste alkalinity,
was estimated at 400 mg/1. The effluent pH was readjusted to 7 - 8
by recarbonation. Sludge production was estimated at 500 mg/1.
2. Precipitation and Filtration. The precipitated and settled effluent
was applied to a granular media filter where the cadmium concentration
is further reduced from 0.2 mg/1 to about 0.02 mg/1. This removal was
based on data from the treatment of municipal wastes containing similar
cadmium concentrations. No specific data for the treatment of industrial
wastes was available.
3
-------
3. Ion Exchange. Cadmium can be removed to less than 10 ppb with
ion exchange. A resin which is selective for cadmium or a mixed bed
resin can be used. However, the mixed bed process is more practical.
There are several types of resins which could be utilized including
carboxylic acid resins and chelating resins. In this case, a chelating
resin may be better because of the low salt splitting capacity. The
degree of removal and the type of resin utilized is dependent on the
complex being removed.
4. Reverse Osmosis. Generally, 97 to 99 percent rejection of cadmium
can be obtained by reverse osmosis in a single-stage process. However,
in this application, because of the low influent concentration assumed,
i.e., 2 ppm, imperfections in the membrane, would result in a lower
percent rejection. Because of the higher surface area and wider opera-
tional pH range, hollow fiber membranes would be recommended.
5. Activated Carbon. Cadmium levels of <50 ppb are obtainable by the
use of activated carbon. Although these low levels are achievable,
the carbon loading is very low, i.e., 3 lb Cd/100 lb carbon. A
residence time of 35 to 40 minutes is required for good removal.
B. Mercury
1. Lime Precipitation. Precipitation of mercuric oxide by lime treat-
ment at pH 11 is expected to reduce the mercury concentration from 2 mg/1
to 0.4 mg/1. The lime dose was estimated at 400 mg/1 for waste with
200 mg/1 alkalinity as CaCO^. The effluent pH must be readjusted to
7 - 8 by recarbonation. Sludge production was estimated at 500 mg/1.
4
-------
2. Sulfide Precipitation and Precoat Filtration. Precipitation of
mercuric sulfide by adding a sulfide source such as sodium hydro-
sulfide followed by precoat filtration can reduce mercury concentrations
from 2.0 mg/1 to 0.03 mg/1. This information was obtained from an existing
pilot plant treating a chlor-alkali industrial waste. In the absence
of other sulfide consuming compounds, the sulfide dose was estimated
to be 10 mg/1 in excess of the stoichiometric requirement, i.e.,
about 18 mg/1 of NaHS for 2 mg/1 mercury. The excess sulfide was
then precipitated as FeS by adding ferrus sulfate (about 75 mg/1).
3. Ion Exchange. Mercury can be removed to approximately 10 ppb
by ion exchange. Again, as in the case of cadmium, the type of resin
employed depends on the metal complex to be removed and the other
constituents in the wastewater.
4. Reverse Osmosis. A 95 to 97 percent rejection of mercury can
be obtained by a single-stage reverse osmosis system assuming the
mercury is in the proper form. The complexed metal or free ion is
rejected; however, organic mercury would not be rejected. Again, this
percent rejection may be lower, due to membrane imperfections at these
low levels. The hollow fiber membrane would also tend to give better
results in this application because of the larger specific surface
and wider pH range.
5. Activated Carbon. Mercury levels of 1 to 5 ppb have been reported
from the treatment of chlor-alkali wastes. However, effluent levels of
34 ppb were obtained in a study with municipal wastes. Again, the
5
-------
carbon loading is low, i.e., 3 lb Hg/100 lb carbon. The adsorbability
of mercury varies significantly with pH with the best results obtained
at pH 4 to 6. The adjustment of pH with HC1 enhances removals probably
due to the precipitation of Hg CI2- The current facilities which
utilize activated carbon for mercury removal do so on a "throw-away"
basis. When regenerating carbon with heat or hot gases, problems
develop with mercury vapors.
C. Cyanides
1. Alkaline Chlorination. Cyanides are completely oxidized to
carbonates in a two-stage process by adding caustic soda and chlorine.
Excess chlorine is reduced by the addition of SC^. The process is
carried out in a continuous flow system with the chemicals being fed
automatically through pH and ORP control units. For a cyanide concen-
tration of 2 mg/1 the chemical doses are estimated as follows: Clg -
20 mg/1, NaOH - 25 mg/1, and SOg - 5 mg/1. Effluent concentration of
cyanide is reported as 0 with the analytical method used. For the pre-
sent calculations a conservative value of 0.2 mg/1 was assumend. (See
section on analytical method for a more detailed discussion of dectable
limits.)
2. Ion Exchange. Free cyanide can be removed to 10 ppb/£y the use of
a strong base anion exchanger. Again, various resins can be employed.
Removals are affected by the presence of cyanide in complexed forms.
3. Activated Carbon. Cyanide can be removed to 50 ppb by the use of
activated carbon. The process involved is catalytic oxidation with
6
-------
CuSO^ and oxygen which converts the cyanide to CO^ and N2- One
disadvantage associated with this process is that copper appears in
the effluent in concentrations up to 0.1 ppm. The copper is capable
of displacing the majority of the heavy metals which are complexed
with cyanide but copper is not capable of displacing the iron. There-
fore, any iron which is in the wastewater must be removed prior to
the carbon system. Removal of iron was not considered in this investi-
gation. To prevent the formation of metal hydroxides and subsequent
filter clogging, the pH of the influent stream must be between 6.5 and
8. While this process appears to be a feasible method for treating
cyanide-bearing wastes, the technological success of the process has
not been fully established.
4. Ozonation. Cyanide can be removed to less than 10 ppb with ozone
alone if the cyanide is not complexed with iron. However, if the
cyanide is complexed, pH control and heat or ultraviolet light must be
employed in combination with the ozone for adequate cyanide removal.
Treatment Costs
All cost estimates were adjusted to January 1972 levels (EPA
Sewage Treatment Plant Index = 175). Allowances were made for engineering,
legal, and administrative fees and miscellaneous construction costs.
Special conditions assumed for each process are discussed below.
1. Precipitation and Filtration. Costs were estimated for one-stage
precipitation processes. Sludge handling was estimated using gravity
thickening and vacuum filtration. The sulfide precipitation process
7
-------
included NaHS addition, filtration through a diatomaceous earth
filter, Fe SO^ addition, flocculation, and sedimentation.
2. Oxidation Processes. Facilities for cyanide removal by alkaline
chlorination and ozonation included flocculation basins and chemical
feed facilities. For chlorination the conventional two-step process
was used which employs pH and ORP controls; additional facilities
were added for dechlorination using
3. Reverse Osmosis. Reverse osmosis was used for the removal of
cadmium and mercury. Cyanide may be removed if present as CN" or
cyanide complexes: however, it appears that any unionized HCN can
pass through membranes. All capital and operating costs were obtained
from manufacturers based on "best judgement" estimates of costs to
achieve the given residual metals concentrations. Filtration costs
were also added for pretreatment in all cases. Because cost data are
not from independently reported sources, the adequacy of membrance
cleaning and replacement costs cannot be determined. It is possible
that higher costs might be obtained upon application of this process
to specific wastestreams; however, no basis exists for increased costs
in this investigation.
4. Ion Exchange. Ion exchange costs were estimated from manufacturer's
data using the "0.6 rule" to make adjustments in plant capacities.
Pretreatment included activated carbon and filtration, although this
pretreatment might not be necessary for all wastes. Due to lack of
8
-------
data and the complexity of selecting operating parameters, these
costs are speculative.
5. Activated Carbon. Metal removal results and costs using activated
carbon were estimated from one reported investigation as interpreted
by two carbon manufacturers. Acid washing of spent carbon was included
to restore the metals capacity of the carbon. An additional 10 percent
was added to O&M costs to cover the purchase of acid. No costs were
added for thermal regeneration to destroy organics which would adsorb
to the resin. However, it was assumed that carbon would be completely
replaced twice per year. Costs for the catalytic destruction of
cyanide were based on the process developed by Calgon Corporation.
No facilities for iron removal prior to this process were assumed.
When iron is present, this process might not be feasible. Although
the treatment costs using activated carbon in all cases are attractive,
it is felt that the technical feasibility of these prcoesses has not
been proven and therefore, that costs are somewhat speculative. Attainable
effluent levels for mercury which have been reported varied from 1 - 5
ppb to 54 ppb. This ranqe is indicated in the cost-effectiveness relation-
ship. However, the lower residual values are subject to question.
Cost effectiveness relationships are shown in Figure 1-3.
9
-------
£
LU
tr
0.70
o
CJ>
O
O
O
\
I—
CO
o
o
^ 0.30
LlI
0.30
0.40
0.20
0.10
0
0
Ion Exchange
Rcve-re Osmosis
i— *—j
Precipitation and Filtration
Precipitation-
Activated Carbon
50 100 150
EFFLUEfOT CONCENTRATION (j^o/\)
200
F13. I. C03T-cFFECTiV
tiv-x-S '»¦** I 'i I I ii * 1 u«t
rro**"m ^
I i Lr* ( > t J W ii 'J i
* "~V» /"\ I
* I V (f 'i fc*
-------
Ion Exchange
0.60!-
0.50
0.40
o
C-">
O
o
g
\
¦&>
I—
CO
o
o
H 0.30
~Z".
L'J
«. ~
H
lTj 0.20
cr
h-
¦
~ i
I
Rovcr39 Oct.ocIs
0.!C\-
6 <-
S-Jlfido Process
«
Activated CarLon
I
Precipitation end Hitrotion
Precipitation
0
0
¦D.
)
z:.r\r\
N—' V V*/
~ • Vi*
I
I tM
EFFLUENT CC!'CEiJTRATiON Jfln/D
— v N ~ •) " pr3l ^tTj-y r>
j 1 . ^ j t * J* 5 * J it W t «v J i
-------
0.70 r-
ro
\
o.c:-
I
i
o
o
O
O
o
o
h—
CO
O
O
H
~2L
LjlJ
H-
<
U
m
H-
0.CC-
0.40 j-
i
I
I
0.50 -
0.20-
I
0.1 C< *'
>
Alkaline Chlorinatioiv
zonation
I
Activet;d Ccrlon Catalysis
.'j iCO 150
EFFLUENT CONCENTRATION
200
i' i ?.
>
. ¦ . a .'*** * p*"™* e " r * ^
V* ' • I kuB I I Uw*» I 1 rf J . . mt I 4 9
:;:? fc.i cya;;:dz removal
------- |