United States
Environmental Protection
Agency
Industrial Environmental Research EPA-600/2-80-1 38
Laboratory June 1980
Cincinnati OH 45268
Research and Development
Foam Flotation
Treatment of
Industrial
Wastewaters
Laboratory and
Pilot Scale
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution-sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-80-138
June 1980
FOAM FLOTATION TREATMENT OF INDUSTRIAL KASTEWATERS:
LABORATORY AND PILOT SCALE
by
David J. Wilson and Edward L. Thackston
Vanderbilt University
Nashville, Tennessee 37235
Grant No. R-804438
Project Officers
Hugh B. Durham and Donald L. Wilson
Industrial Pollution Control Division
Industrial Environmental Research Laboratory
Cincinnati, Ohio 45268
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Industrial Environmental Research
Laboratory-Cincinnati, U. S. Environmental Protection Agency, and approved
for publication. Approval does not signify that the contents necessarily
reflect the views and policies of the U. S. Environmental Protection Agency,
nor does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
ii
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
health often require that new and increasingly more efficient pollution con-
trol methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
This report deals with bench and pilot plant scale studies of the
removal of toxic inorganic compounds from aqueous systems by adsorbing col-
loid flotation. These results are relevant to the treatment of industrial
wastewaters when a high degree of removal of toxic inorganics is necessary,
and should be of use to the secondary lead smelting industry, brass mills,
electroplating shops, copper and zinc smelters, and other sources of waste-
waters containing toxic inorganics. The Industrial Pollution Control Divi-
sion is to be contacted for further information on the subject.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
111
-------
ABSTRACT
A floe foam flotation pilot plant was shown to remove lead and zinc in
dilute aqueous solution to qud,te low concentrations. The results suggest sev-
eral design improvements. These include: (1) a larger mixer-flocculation
tank to increase the detention time of the floe before flotation; (2) in-
creased baffling of the stripping column section, to decrease channelling and
foam overturn at high loadings; and (3) a decrease in the length of the foam
drainage section of the column, to decrease the tendency of the foam to col-
lapse before discharge. Axial dispersion studies indicate that a simple dif-
fusion-type model is not accurate; the large scale of the turbulences pre-
cludes their modelling by diffusion. Hydraulic loading, column alignment,
baffles, and influent dispersion head geometry affect axial dispersion.
The floe foam flotation of zinc is readily carried out with A1(OH)3 and
sodium lauryl sulfate (NLS). Cr(OH)3 is floated with NLS, but adsorbing col-
loid flotation of Cr(III) with Fe(OH)3 or A1(OH)3 yielded better results.
Cobalt and nickel levels are reduced to around 1 mg/£ by flotation with
A1(OH)3 and NLS. Mn(II) levels can be reduced to 1-2 mg/;. by flotation with
Fe(OH)3 and NLS. Floe foam flotation of copper was compatible with several
precipitation pretreatments (soda ash, lime, Fe(OH)j and Al(OH)3), although
modifications were needed to prevent interference from excessive Ca or car-
bonate. Floe foam flotation can therefore be used as a polishing treatment.
The flotation of mixtures of copper(II), lead(II) and zinc(II) was carried out
using Fe(OH)3 and NLS. The flotation of simple and complexed cyanides and
mixtures of metal cyanide complexes was also carried out with Fe(OH)-z and NLS;
a pH of around 5 is optimum.
The flotation of Fe(OH)3 floes with NLS is impeded by several polyvalent
anions, some of which occur in industrial cleaners. These anions displace
surfactant from the floe, rendering it unfloatable. This phenomenon, an in-
terference in waste treatment, could be used to reclaim surfactant from flota-
tion sludges.
A surface adsorption model for floe foam flotation was analyzed and found
to account for the effects of surfactant concentration, ionic strength, spe-
cifically adsorbed ions, and surfactant hydrocarbon chain length.
This report was submitted in fulfilment of Grant No. R-804438 under the
sponsorship of the U. S. Environmental Protection Agency. This report covers
the period May 1, 1976 to November 1, 1978 and work was completed as of
October, 1978.
IV
-------
CONTENTS
Foreword iii
Abstract iv
Figures vi
Tables viii
Acknowledgments x
1. Introduction 1
2. Conclusions 5
3. Recommendations 7
4. Objectives of the Research Program 8
5. Pilot Plant Work (10-cm Dia Column) 10
6. Pilot Plant Design and Construction (30-cm Dia Column) 17
7. Pilot Plant Results (30-cm Dia Column) 35
8. Continuous Flow Axial Mixing Studies 42
9. Batch Technique Laboratory Studies 62
Flotation of Zinc(II) 62
Flotation of Nickel(II), Manganese (II), Chromium(111),
and Cobalt (II) 65
Compatibility of Floe Foam Flotation with Precipitation
Pretreatments 71
Interferences with Floe Foam Flotation Resulting from
Foreign Ions 75
Simultaneous Floe Foam Flotation of Cu(II), Pb(II), and
Zn(II) 76
Floe Foam Flotation of Cyanides 78
References 86
Appendices
A. Data on Lead Removal from Wastes and Simulated Wastes with the
10-cm and 30-cm Columns 92
B. Absorption Isotherms 106
C. Theory of Surfactant Displacement of Salts 121
-------
FIGURES
Number Page
1 The 10-cm pilot plant 11
2 Schematic diagram of 30-cm pilot plant 18
3 Adsorbing colloid flotation pilot plant 19
4 Plan view of pilot plant 20
5 Mixing chamber components 22
6 Flotation column 23
7 Flow dispersion head 25
8 Typical baffle installation 26
9 Foam breaker 27
10 Clarifier 28
11 Support structure 31
12 The foam flotation apparatus used in the axial dispersion studies... 43
13 Dispersion head designs 44
14 High-speed foam breaker design 45
15 The setup for photography of the axial dispersion experiments 47
16 Axial dispersion data, run 1 49
17 Axial dispersion data, run II 50
18 Axial dispersion data, run III 51
19 Axial dispersion data, run IV 52
20 Axial dispersion data, run V 53
21 Axial dispersion data, run VI 54
22 Axial dispersion data, run VII 55
23 Axial dispersion data, run VIII 56
24 Axial dispersion data, run IX 57
25 Center of mass motion for Figures 17-20 59
26 Pulse width versus time for Figures 17-20 61
VI
-------
FIGURES (Continued)
Number Page
B-l Model of floe-bubble attachment ..................................... 107
B-2 Adsorption isotherms of surfactant on floe with condensation
B-3 Dependence of a • on w ............................................ 116
B-4 Dependence of a . on iK ........................................... 118
F crit °
B-5 Dependence of a . on Q ........................................... 119
crit °°
C-l Effect of competing salt concentration c^' on the adsorbtion
isotherm ........................................ •. ................. 127
C-2 Effect of eg on surfactant condensation concentration .............. 128
C-3 Effect of xg on tne adsorption isotherm ............................. 129
C-4 Effect of 10 on the adsorption isotherm .............................. 130
C-5 Effect of XA on the adsorption isotherm ............................. 131
C-6 Effect of temperature on the adsorption isotherm .................... 132
vn
-------
TABLES
Number Page
1 Proposed Experimental Design for 10-cm Flotation Column ............. 13
2 Effects of pH on Lead Removal ....................................... 15
3 Effects of Increasing Ionic Strength on Lead Removal ................ 15
4 Recommended Design Parameters for Further Studies ................... 16
5 Flow Rates for Various Settings of Chemical Feed Pump 7016 .......... 29
6 Effects of pH on Lead Removal ....................................... 36
7 Effects of Increasing Ionic Strength on Lead Removal ................ 37
8 Optimum Operating Parameters, 30-cm Column .......................... 39
9 Parameters for Figure 16 ............................................ 48
10 The Effect of Fe(OH) on Zn(II) Removal ............................. 63
o
11 The Effect of pH on Zn(II) Removal with Fe(III) and N'LS ............. 63
12 The Effect of Al(III) on Zn(II) Removal ............................. 64
13 Floe Flotation of Zinc with Al(OH) and NILS. Effect of Ionic
Strength ........................ . ................................. 64
14 Floe Flotation of Zinc with Al(OH)^ and NLS. Effect of Electrolyte
Identity ........................ ? ................................. 65
15 Effect of pH on Nickel (II) Floe Foam Flotation ...................... 65
16 Effect of Ionic Strength on Nickel (I I) Floe Foam Flotation .......... 66
17 Ni(II) Flotation at Decreased Ni(II) Concentration .................. 66
18 Precipitate Flotation of Chromic Hydroxide .......................... 67
19 Effect of pH on Chromium(III) Floe Foam Flotation ................... 68
20 Effect of Iron(III) Concentration on Chromium(III) Floe Foam
Flotation [[[ 69
21 Effect of Inert Salt Concentration on Chromium(III) Floe Foam
Flotation [[[ 69
22 Floe Foam Flotation of Cobalt (II) with Al(OH) or Re(OH)3 and NLS... 70
23 Effect of Ionic Strength on Flotation of Co(II) with A1(OH)_ and
-------
TABLES (Continued)
Number Page
25 Effects of Air Sparging at Low pH on Cu(II) Removal after
Precipitation with Na CO 73
26 Effect of Foaming pH on Cu(II) Removal after Precipitation with
Ca(OH) 74
27 Effect of Foaming pH on Cu(II) Removal after Coprecipitation with
Al (OH) or Fe (OH) 75
28 Effect of Various Added Salts and Glycerol on the Flotation of
Ferric Hydroxide ' '
29 Results of the Floe Foam Flotation of Cu(II), Pb(II), and Zn(II)
with Fe(OH)3 and NLS 78
30 Stability Constants for Cyanide Complexes 79
31 Solubilities of Cyanide Compounds 79
32 Cyanide Precipitate Flotation Runs 81
33 Molarity arid mg/£ Conversions 81
34 Standard Operating Conditions: Determining Runs 82
35 Cyanide Ionic Strength Runs 83
56 Heavy Metal Runs: Average Residual Concentrations 85
37 Percent Residual Metals: Run Types A and D 85
A-l Data on Lead Removal from Wastes and Simulated Wastes with the 10-cm
and 30-cm Columns 93
A-2 Data on Lead Synthetic Waste Removal with the 30-cm Column 98
A-3 Data on Zinc Removal (Plating Waste) with 30-cm Column 103
B-l Floe-Bubble Binding Energy as a Function of Contact Angle 110
B-2 Effect of Temperature on a 117
r crit
IX
-------
ACKNOWLEDGMENTS
The experimental work presented in this report was carried out by Joseph
C. Barnes, Ann N. Clarke, Ben L. Currin, Jo S. Hanson, Douglas L. Miller, Jr.,
F. John Potter and Ronald P. Robertson. The theoretical work was done with
the aid of Ben L. Currin. The advice of Hugh B. Durham and Donald L. Wilson
(who served as the first EPA project officer) is gratefully acknowledged.
-------
SECTION 1
INTRODUCTION
The removal of trace levels of toxic metals has become a problem of cur-
rent interest in the area of advanced wastewater treatment. Toxic metals are
not efficiently removed by biological treatment or by conventional drinking
water treatment techniques. They not only interfere with the efficient oper-
ation of secondary (biological) treatment plants, but their presence in bio-
logical sludge interferes with the disposal of this material by some methods,
especially land application. The extent of the problem is detailed in our
previous report (1) and in other government publications (2,3,4, for example).
Any metal removal technique should be: (a) rapid; (b) cheap in terms of
labor, energy, equipment, and chemicals; (c) capable of application in small,
intermediate, and large scales; (d) productive of small volumes of liquids or
solids highly concentrated in the contaminant(s); and (e) capable of produc-
ing effluents well within the standards established or anticipated. Data
presented here and in our earlier report (1) indicate that adsorbing colloid
foam flotation may well meet these criteria.
LITERATURE REVIEW
The brief literature review given here can be supplemented by a more
comprehensive review given in our first report (1), and by a very comprehen-
sive review of the recent literature (5). Lemlich's book (6) provides an ex-
cellent introduction to principles and applications. Somasundaran published
extensive general reviews in 1972 (7) and 1975 (8). Grieves (9), Ahmed (10),
Bahr and Hanse (11), Panou (12), Richmond (13) and Bakerzak (14) have also
written recent general review articles on foam flotation. In 1977, at least
three articles reviewed the use of foam flotation techniques in wastewater
treatment (15-17).
An extensive collection of symposium papers covering most aspects of
particulate flotation was edited by Somasundaran and Grieves (18) .
The operation of foam stripping columns has been analyzed by a number of
workers. The earlier work was reviewed by Goldberg and E. Rubin (19), who
also referenced work on foam drainage, a matter of some importance in foam
separations. Wang and his co-workers worked out a theory for continuous
-------
bubble fractionation columns which requires local equilibrium between surface
and bulk phases (20,21,22). Goldberg and Rubin have analyzed stripping
columns without solute transfer in the countercurrent region (23) . Cannon and
Lemlich presented a detailed analysis which assumes linear isotherms (24), and
Lee has given a somewhat similar treatment (25). Sastry and Fuerstenau
analyzed the differential equations modeling a countercurrent froth flotation
column, obtaining formulas for column efficiencies in some limiting cases (26).
Wilson and his co-workers analyzed stripping column operation with axial
diffusion, non-linear isotherms, and finite rate of solute transport between
surface and bulk liquid phases (27) ; they solved the differential equations
describing mass transport by iterative use of a quasi-linearization method.
Grieves and his co-workers have developed an approach which permits the use
of batch foam fractionation data from one system to predict the performance
of other systems, both batch and continuous (28).
Next, we will address the attachment of particles to bubbles. Petrakova,
et al. (29) examined the mechanism of interaction of gas bubbles with mineral
particles, and Abramov (30) has also worked on the physico-chemical modeling
of flotation systems. Reay and Ratcliff (31) provided a theoretical treat-
ment of the benefits of using smaller bubbles or larger particle sizes in
dispersed air flotation systems. When particles are large enough that they
are unaffacted by Brownian motion, flotation rate increases with increasing
diameter.
Scheludka, el al. (32) has given an elegant analysis in terms of capil-
larity of the attachment of a spherical particle to a planar liquid surface
after contact is made. The minimum particle size (M0~4 cm) for which flo-
tation can occur is calculated as well as the maximum particle size which can
remain attached. The maximum size depends on the contact angle. Particle
detachment may be the result of gravity or the kinetic energy of collision
with a bubble; the former yields the larger particle radius, so the latter
determines the upper limit of flotability for a given system. The maximum
radius calculated for a representative system is 2.7 x 10~2 to 5.5 x 10"*- cm
for gravity detachment; the smaller size corresponds to those particles
found in practice to be removed by flotation. Deryagin, et al. (33) also
studied the criteria for bubble attachment using the theory of heterocoagula-
tion. Bleier, et al. (34) also used heterocoagulation to examine adsorption
and critical flotation conditions. They claim that surface tension data show
a large surface charge density on the bubble during flotation. Hetercoagula-
tion suggests that the oppositely-charged double layer of a system reacts to
produce a large potential energy of attraction. These electrostatic forces
lead to the rupture of the aqueous film around the particles, and then these
electrostatic forces decrease. The ionic species in the gas-liquid interface
then desorb. The relative hydrophobicity of the particle surface then de-
termines whether the particle and bubble will stay attached (will hetero-
coalesce).
Lai and Fuerstenau (35) proposed a model for the surface charge dis-
tribution and its effect on flotation response. In water, an oxide surface
exhibits, simultaneously, positive, negative, and electrically neutral sites,
due to the interplay of H+ and OH" with the surface. The distributions of
these sites were calculated as functions of pH, and the point of zero charge
-------
determined. Various oxide minerals and collectors (surfactants) which ab-
sorbed either physically or chemically were studied. It was expected that
flotation efficiency would depend on surface site distributions since flota-
tion had been shown to correlate with adsorption phenomena. Also, one expects
that, at low collector concentrations, a physically adsorbing cationic col-
lector would adsorb at a negative site; an anionic collector, at a positive
site. Flotation responses correlated well with the site distribution model
predictions, allowing, in some instances, for concomitant ionization of
neutral flotation collectors.
Floe foam flotation isotherms and local rate effects have been investi-
gated by Wilson and his co-workers by means of a modified Gouy-Chapman model
in which the binding force between particles and the air-water interface is
due to coulombic attraction between the negatively (positively) charged air-
water interface and the positively (negatively) charged particles. The
charge on the air-water interface is assumed to be due to the formation of a
hemimicelle layer of ionic surfactant on it, and the coulombic interaction is
modified by the ionic atmospheres in the vicinities of the two charge den-
sities (36,37,38). They have extended the statistical mechanical treatment
of floe adsorption isotherms to include the effects of non-ideal floes and
salts (59), the electrical repulsions between floe particles (40), the oc-
currence of cooperative phenomena in the surfactant film (41), and the effects
on surface potentials of the specific adsorption of ions by floes (42).
More recently, this group has also examined a model used by Fuerstenau,
Somasundaran, Healy, and others for some time in interpreting the results of
ore flotation experiments (43,44,45,46,47,48,49,50, for example). In this
model it is assumed that the particle surface is made hydrophobic by the ab-
sorption of the ionic heads of the surfactant ions in the primary adsorption
layer of the particle, leaving the hydrocarbon tails presented to the water.
Van der Waals interactions between the hydrocarbon tails permit the possibility
of surface condensation; particles with surfaces densely covered with surfactant
are hydrophobic, with non-zero air-water contact angles; this results in bub-
ble attachment. Wilson's group investigated the statistical mechanics of
this phase change (41); the effects of specific adsorption of foreign ions on
displacement of surfactant from the floe (51) ; and the calculation of adsorp-
tion isotherms for this non-coulombic model for the attachment of floe parti-
cles to the air-water interface (52). They also examined the magnitude of
the viscous drag forces tending to detach floe particles from the air-water
interface (51,52) .
We turn next to recent work on adsorbing colloid flotation. Our first
report (1) presented data on the precipitate and adsorbing colloid flotation
of lead, copper, cadmium, mercury, arsenic and fluoride; many of these data
are published elsewhere (53-57). Zeitlin and his co-workers have published
very extensively on adsorbing colloid flotation separations; they have used
the technique for concentrating the following from sea water: zinc (58),
copper (59), molybdenum (60,61), uranium (62,63), mercury (64), phosphate
and arsenate (65), silver (66), and vanadium (67). They also carried out a
study of the foam separation of six floes suitable for adsorbing colloid
flotation [Fe(OH)-, A1(OH)3, Th(OH)4, Mn02, HgS, and CdS] with nine different
surfactants (68); this work should prove invaluable, particularly to those
-------
developing techniques for concentrating trace quantities of material for
analysis.
Zhorov, el al. (69) studied the flotation of copper, molybdenum, and
uranium with Fe(OH)3 as the floe, stearic acid and indolebutyric acid as col-
lectors and Stearox 6 as the foaming agent. An application of adsorbing col-
loid flotation is suggested by a study of the sorption of arsenite by fer-
rous sulfide carried out by Grigor'ev and his co-workers (70).
Grieves and Bhattacharyya, very active workers in foam separation
techniques, have investigated a number of waste treatment applications, many
of which are reviewed in our first report. Studies of the flotation of
Cr(OH)3 (71), and of the precipitate coflotation of CaS05 and CaCC>3 (of
interest in treating wet scrubber slurries formed during SC>2 removal from
stack gases) (72) are of particular interest here.
A problem which arises in the scaling up of foam flotation separations
is that of collapsing the foam. Our experience has been that the high-speed
spinning disc technique described some years ago by E. Rubin and his co-
workers (73,75) is very effective.
Although we are not concerned with ore flotation techniques as such,
their principles are quite significant for precipitate flotation and adsorbing
colloid flotation. M. C. Fuerstenau edited a recent extensive collection of
papers on ore flotation, which contains a wealth of relevant information (75).
-------
SECTION 2
CONCLUSIONS
Floe foam flotation methods were investigated on a pilot-plant scale and
demonstrated to reduce lead and zinc from dilute aqueous solution to very low
concentrations. The pilot-plant experience suggests several improvements in
the design of the apparatus which should result in an effluent of higher
quality and a column of increased hydraulic loading capacity. These modifi-
cations include: (1) a larger mixer-flocculation tank to increase the deten-
tion time of the floe before it is removed by flotation; (2) increased baf-
fling of the stripping column section, to decrease chaneling and foam
overturn at high hydraulic loadings; and (3) a decrease in the length of the
foam drainage section of the column, to decrease the tendency of the foam to
collapse before it has been discharged from the column.
Studies indicate that a simple diffusion-type model is not very accurate
when column performance becomes seriously limited by axial dispersion. The
scale of the turbulences (channeling, overturn, and eddies) is too large to
permit their accurate modeling by simple diffusion. Hydraulic loading and
alignment of the column, baffles, and influent dispersion head are the most
important factors affecting axial dispersion.
The floe foam flotation of zinc(II) is readily carried out with aluminum
hydroxide as a carrier floe and sodium lauryl sulfate (NLS) as surfactant;
ferric hydroxide is much less effective as a carrier floe. Chromic hydroxide
is floated with NLS, but the residual Cr(III) concentration was generally
greater than 5 mg/£; adsorbing colloid flotation with Fe(OH)3 and AlCOH)^
yielded much better results. Cobalt(II) and nickel(II) levels are reduced to
around 1 mg/£ by flotation with A1(OH)3 and NLS; Fe(OH)3 is not so effective
with nickel.
Floe foam flotation of copper(II) was shown to be compatible with a
number of precipitation pretreatments (soda ash, lime, ferric and aluminum
hydroxides), although some modifications are needed to prevent interference
from excessive calcium or carbonate ion concentration. This makes possible
the use of floe foam flotation as a polishing treatment for relatively con-
centrated wastewaters which are not adequately treated by precipitation
alone.
The flotation of ferric hydroxide floes with NLS is profoundly affected
by a number of polyvalent anions, some of which are commonly found in in-
dustrial cleaners (silicates, phosphates). These can cause serious inter-
ference with floe foam flotation unless they are disposed of by preliminary
-------
treatment, such as lime precipitation. These anions apparently are able to
displace surfactant from the floe, thereby rendering it hydrophilic and
unfloatable. This phenomenon, at times a troublesome interference in waste-
water treatment, can apparently be used to recover surfactant from floe foam
flotation sludges for recycling.
The flotation of mixtures of copper(II), lead(II) and zinc(II) was
successfully carried out using Fe(OH)3 as the floe and NLS as the collector
surfactant. The flotation of simple and complexed cyanides and mixtures of
metal cyanide complexes was also carried out with Fe(OH)3 and NLS; a pH of
around 5 is optimum. Zinc cyanide complex is not as well removed as are
those of copper, chromium, nickel and cobalt.
A surface adsorption model for floe foam flotation was examined by
statistical mechanical techniques and found to account for the effects of
surfactant concentration, ionic strength, specifically adsorbed ions, and
surfactant hydrocarbon chain length. The model proved to be a helpful tool
for the designing and trouble-shooting of floe foam flotation separations.
-------
SECTION 3
RECOMMENDATIONS
The results of the adsorbing colloid flotation pilot plant studies re-
ported here justify a recommendation that the following design modifications
in the pilot plant be investigated.
(1) The size of the mixer-flocculator tank should be increased to
allow sufficient time for optimum floe formation and adsorption of the dis-
solved ion onto the floe. This should result in improved effluent quality.
Laboratory scale kinetic studies of the rates of coprecipitation and/or ad-
sorption of copper, zinc, and lead with ferric hydroxide are needed to deter-
mine the detention time required.
(2) More baffling should be added to the stripping section of the
column to decrease channeling and foam overturn at high hydraulic loadings.
(3) The length of the foam drainage section of the column should be
decreased to lessen the tendency of the foam to collapse and drop sludger*
through the column before it has been discharged from the top of the column.
(4) The pilot plant clarifier assembly should be modified to permit the
use of sodium carbonate (or other salts) to displace surfactant from the
sludge in the collapsed foamate in order to recover this surfactant for
recycling.
The bench scale work supports the following recommendations:
(1) The use of mixed floes (aluminum hydroxide ferric hydroxide, for
example) for the removal of mixtures of metal ions should, be studied.
(2) Additional work on interferences resulting from the presence of
polyvalent anions likely to be present in waste streams is needed to deter-
mine the extent of these interferences and to develop techniques to counter-
act them.
Additional field testing of the pilot plant is necessary to facilitate
improvements in equipment design, flotation procedures, and trouble-shooting
techniques.
-------
SECTION 4
OBJECTIVES OF THE RESEARCH PROGRAM
The objectives of our work included both pilot-scale work and bench-
scale studies to demonstrate the feasibility of adsorbing colloid flotation
for industrial wastewater treatment and to extend the applicability of the
technique to substances and combinations of substances not dealt with in our
earlier report (1) . These objectives can be listed as follows:
A. Pilot-scale work
1. Laboratory testing of a 10-cm continuous flow pilot plant
removing lead from simulated wastewaters with ferric
hydroxide and sodium lauryl sulfate (NLS) to estimate optimal
design parameters for a 30-cm pilot plant to be constructed.
2. Construction of a mobile 30-cm continuous flow adsorbing col-
loid flotation pilot plant suitable for field testing.
3. Operation of the 30-cm pilot plant with simulated and actual
wastewater samples to determine optimum ranges of operating
conditions and to indicate needed modifications of the design.
4. Examination of column performance characteristics (particularly
axial dispersion) and their dependence on the operating para-
meters of the 10-cm continuous flow column.
B. Bench-scale work
1. Development of feasible methods, if possible, for the adsorbing
colloid flotation of chromium(III) , manganese (NM^), cobalt(II),
zinc(II), nickel(II), and cyanides. (Initially we included
antimony and thallium on this list; we could obtain no waste -
waters containing these, so we replaced them with cyanides.)
2. Investigation of the compatibility of various precipitation
pretreatments [Na2C05, Ca(OH)2, Na2S, etc.] with subsequent
adsorbing colloid foam flotation of copper(II). Demonstration
of such compatibility would establish the feasibility of using
adsorbing colloid foam flotation as a polishing technique to
upgrade the quality of fairly concentrated wastewaters after
their pretreatment by precipitation methods.
-------
3. Investigation of the adsorbing colloid foam flotation of mix-
tures of zinc(II), lead(II), and copper(II) to determine the
feasibility of using the technique on mixed wastes.
4. Investigation of the recovery of surfactant from collapsed
foamate and sludge separated from collapsed foamate. Sur-
factant recovery would markedly improve the economics of the
process and also reduce the likelihood of contamination of
surface and ground waters with surfactant if the sludge is
deposited in a landfill.
5. Investigation of interferences with adsorbing colloid flotation
[using Fe(OH)3 and NLS] from the presence of other ions such
as phosphate, oxalate, EDTA, cyanide, etc.
In connection with these lab studies, we note that industrial wastes
are usually complex mixtures of markedly varying composition, and therefore
neither studies on simple solutions of a single contaminant nor studies on
a single type of waste can be expected to give an accurate assessment of the
capabilities and limitations of any waste treatment technique. A broad
range of bench scale tests on samples of an equally broad range of known
compositions should provide invaluable guidance in trouble-shooting field
tests and industrial operations employing this or any other treatment tech-
nique.
-------
SECTION 5
PILOT PLANT WORK (10-on DIAMETER COLUMN)
DESCRIPTION OF APPARATUS
Our first continuous-flow apparatus approaching pilot plant size was the
flotation system diagrammed in Figure 1. The column was a lucite cylinder 10
cm in diameter and 186 cm in length. The influent was introduced through a
tube passing through the top cover plate of the column, and was sprayed into
the rising foam through a spider-like dispersion head mounted at the end of
this tube about 90 cm below the top to the column. A foam discharge port was
mounted in the top coverplate. Bubbles were generated by admitting air
through a cylindrical bubbling stone (YWR Scientific Catalog No. 32568-007)
mounted on the bottom coverplate. The air flow rate was measured with a
variable area flow meter and the wastewater flow was determined by measuring
the time required to discharge one liter. House air was humidified and pass-
ed through glass wool before being admitted to the column. The influent was
prepared by adding the appropriate chemicals to approximately 50 £ of water
in a large plastic container. The solution was stirred continuously
while the pH was adjusted to the desired value, and the solution
was then pumped to a smaller constant head reservoir above the column,
from which the influent flow to the dispersion head was controlled by a screw
clamp. The effluent was discharged from the column through a plastic tube
through the bottom coverplate, the vertical position of which could be adjust-
ed to control the level of liquid in the bottom of the column. Four stainless
steel wire mesh (0.05-cm mesh) baffles were mounted between 95 to 113 cm below
the top of the column to aid in reducing turbulence and overturning in the
foam. The foam was discharged to either a spinning screen foam breaker or a
hot surface foam breaker, and collected in a glass beaker below the foam
breaker.
Several designs of influent dispersion heads were tested before satis-
factory results were obtained. Reduction of column cross-sectional area must
be minimized, the distribution of influent over the column cross-section must
be quite uniform, and the linear velocity of the liquid exitting the dis-
charge ports must be small enough that the foam is not broken. If these con-
ditions are not met, foam channelling and overturn occur at relatively low
influent hydraulic loading. The design finally used was a spider made by
bundling together seven 5-cm pieces of 1/16-in ID copper tubing and soldering
these into the end of a 3.8-cm piece of 3.8-in ID copper tubing. The six
outer tubes were bent at right angles to the axis of the distribution head
and two 1/16-in holes were drilled into the top side of each tube. The ends
of these tubes were pinched shut, and the center small tube was cut short
10
-------
n
1. Air Flow Meter
2. Glass Wool Column
3. Humidifier
4. Air Stone
5. Influent Reservoir and Stirrer
6. Pump
7. Constant Head Tank
8. Influent Stopcock
9. Influent Dispersion Head
10. Influent Dispersion Head
11. Foam Baffles
12. Effluent Drain
13. Effluent
14. Foam Discharge Pipe
15. Spinning-wire Foam Breaker
16. Hot-Wire Foam Breaker
17. Collapsed Foamate
FIGURE 1. The 10-cm pilot plant.
11
-------
and pinched shut. Wire struts to keep the distribution head centered radially
in the column were soldered onto the bottoms of alternate tubes.
It is essential that the dispersion head and the baffles be carefully
levelled and that the column be vertical. A 1° vertical misalignment of the
column resulted in consistent channelling of the foam down the lower side.
Slight tilting of the baffles or dispersion head also caused channelling.
The foam discharge port in the top of the column directed the foam to a
spinning screen and wire spider foam breaker in an inverted gallon milk jug,
the bottom of which had been cut out. This device was usually overwhelmed by
foam, and a second inverted bottomless milk jug containing a hot wire gird was
added below the first to catch the overflow. Neither of these foam breakers
was really satisfactory; we have since used high speed, spinning disc foam
breakers of the type described in Section 6. These are extremely effective.
PROCEDURES
Our objective was to investigate the removal of lead by adsorbing colloid
flotation with ferric hydroxide and sodium lauryl sulfate (N'LS) . The proce-
dures used were as follows.
Stock solutions of ferric chloride (Felll, 10 g/£), lead nitrate
(Pbll, 10.0 g/£), NLS (10.0 gm/£), sodium hydroxide (5 molar) and sulfuric
acid (1.25 molar) were prepared using tap water. Nitric acid (5 ml/£ of con-
centrated acid) was added to the lead nitrate stock solution. Reagent grade
lead nitrate, nitric acid, and sulfuric acid were used; NLS was laboratory
grade; and ferric chloride and sodium hydroxide were technical grade. Ionic
strength of the simulated waste was adjusted by the addition of the appro-
priate weight of solid reagent grade sodium nitrate to the waste reservoir.
To make up a wastewater sample, the desired volumes of lead nitrate and
ferric chloride were added to roughly 10 I of tap water in the waste reser-
voir, sodium nitrate was added as needed, and the solution stirred until all
solids were dissolved. IVe then added tap water until the desired volume was
reached (25 or 50 £) and adjusted the pH by adding sodium hydroxide solution
or sulfuric acid as needed. NLS solution was then added, the pH again ad-
justed, and the pump feeding the constant head reservoir turned on.
One liter of 200 mg/£ NLS solution was poured into the column initially
to permit rapid generation of foam. During this process an air flow rate of
approximately 3 SL/min was maintained. Influent was permitted to enter the
column when the foam reached the dispersion head; at this point some chan-
nelling and overturn invariably occurred. The foam usually quickly stabilized,
however, and the influent flow rate could then be increased somewhat and the
air flow rate decreased.
The stirrer in the main waste reservoir and the two foam breakers were
kept in operation during the entire run. Effluent samples were taken at ap-
proximately 5-minute intervals during the run.
-------
Reagent grade lead nitrate and concentrated nitric acid (5 ml/a) in de-
ionized water were used to make a stock solution and standards for analysis.
Lead concentrations were determined on a Perkin-Elmer model 305B atomic ab-
sorption spectrophotometer operating at 217.0 nm.
RESULTS AND DISCUSSION
The objectives of this research were to investigate the feasibility of
using adsorbing colloid flotation to treat lead-bearing wastes on a larger
scale than the previous bench-scale experiments and to determine whether or
not experimental results were consistent with the theoretical model presented
in Section 1. Results of some work with a smaller (4.8-cm ID x 121 cm) con-
tinuous flow column (1), warranted a trial with a larger column. A larger
column also seemed necessary to determine the effects of channeling and foam
overturn which could not be estimated from the work with the 4.8-cm column.
In the work with the smaller column, Fe(III) concentration was usually
set at 100 mg/£, NLS at 50 mg/£, and Pb(II) at 50 mg/£. At low ionic strength
(Na2S04 less than 4000 mg/£) optimum pH was in the range of 6.0-6.5. These
conditions were used as a starting point for the experiments with the 10.2-cm
column. Variables studied in this work were Fe(III) concentration, NLS con-
centration, pH, influent flow rate, air flow rate, and ionic strength. The
influent Pb(II) concentration was 50 mg/£ in all runs.
Table 1 shows the experimental design proposed for this work. The values
for the variables were chosen to include what were thought to be the optimum
operating conditions for the column. Air flow rate and influent flow rate
were not specified in this proposed outline because they could be (and were)
varied within each run.
As the work progressed, some of the proposed values were changed to agree
with newly indicated optimum conditions. The values for parameters actually
used in each run and the results from each run are listed in Appendix A.
TABLE 1.
Run
Numbera
1
2
3
4
5
6
7
8
9
10
11
12
PROPOSED
pH
6.0
5.5
5.0
6.5
7.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
EXPERIMENTAL
Fe(III)
mg/i
150
150
150
150
150
150
150
100
200
150
150
150
DESIGN
NLS
mg/H
40
40
40
40
40
35
30
40
40
40
40
40
FOR 10 -cm FLOTATION COLUMN
NaN03
moles/2,
_
-
-
-
-
-
-
-
-
.10
.20
.30
Added
mg/£
_
-
-
-
-
-
-
-
-
8500
17000
25500
All runs made with initial Pb(II) concentration of 50 mg/£.
13
-------
The aim of this study was not to reach the maximum hydraulic loading pos-
sible in the column, but to balance several variables to achieve a procedure
which could produce consistently good effluent quality at low cost. Both in-
fluent flow rate and air flow rate could easily be adjusted during a run.
Fe(III) concentration, NLS concentration, and ionic strength were held con-
stant throughout any one run.
Constant pH monitoring and/or adjustment was impossible with this ap-
paratus. Invariably the pH changed slightly during the course of a run.
These changes were always small (less that 0.4 pH units) and seemed to be
always toward a value around pH 5.5. When the initial pH was less than 5.5,
the pH increased during the course of the run; if the initial pH was greater
than 5.5, the pH decreased.
In very early runs (before the dispersion head problem was worked out)
100 mg/£ Fe(III) seemed to be insufficient to achieve the lead removals ex-
pected. We therefore used a concentration of 150 mg/£ Fe(III) in further
preliminary runs.
It is desirable to use the smallest amounts of NLS and Fe(III) possible
(while still maintaining good effluent quality) in order to keep chemical
costs to a minimum. The wider the pH range in which good effluent quality can
be maintained, the less accurate (and therefore less expensive) the pH control
system on a full-scale operation can be. Increasing ionic strength has been
shown (53,54,55,56,67) to seriously decrease the separation efficiency. Three
runs were made in this work at higher ionic strengths. Influent and air flow
rates were adjusted frequently during each run.
The effects of channelling ard overturning in the foam proved to be the
controlling factor in almost every run made. Effluent lead concentrations
less than 0.15 mg/£ could be attained within fairly wide ranges of pH and
influent flow rates when foam overturning could be eliminated. When channel-
ling and overturning occurred effluent concentrations in the range of 0.5 to
1.5 mg/£ Pb(II) were common.
Influent flow rate seemed to affect the removal efficiency only when over-
turning produced excessive axial dispersion in the foam. In one run in which
the foam was particularly stable, an influent flow rate to 0.95 £/min (15
gal/hr), producing a hydraulic loading rate of 2.9 gal/min-ft^ or 171 nrVday m,
was reached, and the effluent lead concentration was 0.05 mg/£. In most runs,
serious overturning occurred above about 0.63 £/min (10 gal/hr), a hydraulic
loading of 1.9 gal/min-ft^ or 112 nrVday-m .
The air flow rate used in most runs was 2.0 £/min (1.24 m-Vmin-m )- At
air flow rates below this value it was almost impossible to prevent overturn-
ing of the foam, even at the lowest influent flow rates. At air flow rates
greater than 2 £/min, somewhat higher influent flow rates could be tolerated,
but the foam coming out the top of the column was very wet. At an air flow
rate of 2 £/min the volume of collapsed foamate was always 3% or less of the
influent volume. At an air flow rate of 4 £/min (700 m3/day-m2), the col-
lapsed foamate was as much as 9% of the influent volume. Since a smaller vol-
ume of more concentrated solution is easier to dispose of, the drier foam is
14
-------
preferable, and the lower air flow rate was chosen.
Considerably lower concentrations of NLS than the 50 mg/£ used in the
smaller column (1) were sufficient to produce a stable foam. The minimum
concentration which could be relied upon to produce a consistently stable foam
at influent flow rates above 0.50 £/min (8 gal/hr), or 1.5 gal/min-ft2 (88
m-Vday-m ) was 35 mg/£. Occasionally, 30 mg/£ NLS produced a sufficiently
stable foam. Even at 30 mg/£, there was some dissolved NLS being carried out
with the effluent. This was evident from the foam which built up on the sur-
face of the liquid in the effluent bucket.
A somewhat wider range of pH than the 6.0-6.5 recommended earlier (1)
produced effluents of 0.1 mg/£ or less, as is shown in Table 2.
TABLE 2. EFFECTS OF pH ON LEAD REMOVAL
pH
Influent Flow Rate
(gal/hr) (£/min)
Residual Lead
(mg/£)
5.0
5.5
6.0
6.4
7.1
12.5
5.5
8.0
8.0
8.5
.78
.35
.50
.50
.54
.10
.08
.10
.10
.84
Note: All runs were made with 50 mg/£ Pb(II), 150
mg/£ Fe(III) and 35 mg/£ NLS. Air flow
rate was 2 £/min.
a!0 gal/hr = 1.9 gal/min-ft2:
or 112 m /day-m .
Optimum Fe(III) concentration in the larger column was 150 mg/£. At 100
mg/£ Fe(III), the effluent concentration was somewhat higher (0.3 mg/£ at 10.5
gal/hr) than at 150 mg/£ Fe(III) (0.10 mg/£ at 12.5 gal/hr). The effluent
lead concentration at 200 mg/£ Fe(III) was essentially the same (0.10 mg/£ at
10.5 gal/hr) as that at 150 mg/£ Fe(III). Therefore, 150 mg/£ seemed to be
the optimum Fe(III) concentration in this study.
As expected, increasing ionic strength seriously reduced the removal ef-
ficiency of the column. Three runs were made at increasing ionic strengths
(addition of NaNO^) at otherwise optimum conditions. Some of the results are
shown in Table 5.
TABLE 3. EFFECTS OF INCREASING IONIC STRENGTH ON LEAD REMOVAL
NaN03 Added
moles/£ mg/£
Influent Flow Rate
gal/hra £/min
Residual Lead
mg/£
0.10
0.10
0.20
0.20
0.30
0.30
8500
8500
17000
17000
25500
25500
7.0
10.5
7.0
8.5
6.5
8.5
.44
.66
.44
.54
.41
.54
0.44
0.38
1.80
2.10
2.80
3.50
Note: All runs made with 50 mg/£ Pb(II), 150 mg/£ Fe(III), and
35 mg/£ NLS. pH was 5.7 to 5.8, air flow rate was 2 £/min,
15
-------
As mentioned in the description of the apparatus, having the screen baf-
fles and dispersion head level was very important in preventing overturning
in the foam. The slightest misalignment of either led to overturning of the
foam, which would then not stabilize. When the baffles and dispersion head
were level and overturning of the foam occurred, the foam would often recover
in a few minutes and return to a stable rise pattern.
The reddish-brown color of the ferric hydroxide floe was very obvious in
the foam at the high concentrations present from the dispersion head to the
top of the column. Very seldom was there any discernible color below the two
two screens; overturning of the foam was most common below the lowest screen.
On some runs, especially those at pH's greater than 6.5 and those at high
ionic strength, the ferric hydroxide was quite obvious in the effluent as a
yellow color even when there seemed to be no color breaking through below the
screens .
The glass rods supporting the screen baffles must be kept well away from
the walls of the column. Toward the conclusion of the work with this ap-
paratus the screens had deteriorated enough to allow one of the three support
rods to move toward the column wall. This caused serious channelling of the
influent between the rod and the wall, which was disastrous to effluent
quality. Removing the whole screen-support-rod assembly, taking the screens
off the rods, and punching new holes for the rods at least 1/2 inch from the
edge remedied the problem.
Bubble size was not measured quantitatively in this work, but at one
point during the work the air stone which had been in use was replaced be-
cause it was clogged by oil present in the compressed air. The new stone,
which looked identical to the old one, produced bubbles which were obviously
larger than those produced by the original stone. The effluent from a run
began with the second stone in place was very yellow and the run was not con-
tinued. When the original stone was replaced after backwashing with a com-
mercially available detergent, the bubbles were smaller, and the effluent
from the same batch of waste was colorless.
Channelling and foam overturning proved to be the major determinants of
effluent quality in this work. Increasing ionic strength seriously decreases
separation efficiency. Above an NLS concentration of 55 or 40 rug/ 2, in-
creasing the surfactant concentration does not improve removal efficiency,
and at higher concentrations, more NLS is lost with the effluent. pH values
outside of the range 5.0-6.5 increase residual lead concentrations. Tempera-
ture effects were not considered here. Table 4 summarizes recommended design
parameters .
TABLE 4, RECOMMENDED DESIGN
pH 5.5-6.5a
Fe(III) concentration 150 mg/£
NLS concentration 55-40 mg/J,
Hydraulic loading rate 2-5 gal/min-ft^ (118-176
Air supply 0.2-0.5 m-Ymin-m-
a
For high ionic strength wastes, a lower pH than 5.5 may be
necessarv.
16
-------
SECTION 6
PILOT PLANT DESIGN AND CONSTRUCTION (30-cra DIAMETER COLUMN)
The principal objective of the present study was the design, construction
and testing of a mobile adsorbing colloid foam/flotation facility suitable for
both laboratory studies and field testing. Several major changes from the de-
sign used for the continuous flow apparatus described in Section V were neces-
sary. These included (a) a high-capacity spinning disk foam breaker to deal
with the very large volumes of foam generated; (b) a monitoring and control
system to provide continuous adjustment of the pH; (c) a chemical feed system
to meter the process chemicals into the waste stream at the proper locations;
(d) a mixing and flocculation chamber; and (e) a small clarifier to separate
sludge from the collapsed foamate.
The foam flotation apparatus used in this study is shown in Figures 2 and
3. The system was built on a steel frame mounted on casters so that it would
be portable. The equipment mounted on this support frame includes the pumping
system, mixing and flocculation chamber, flotation column, foam breaker, foam-
ate clarifier, chemical feed system, pH monitoring and control system, air and
waste stream flowmeters, and electrical control panel. A separate waste stor-
age tank was also used in the laboratory studies. All of these are described
in detail below.
PUMPING SYSTEM AND RELATED PIPING
The pumping system and related piping are shown in Figures 2, 3 and 4.
In order to achieve minimal friction losses at flow rates of 1 to 5 gal/min
(.0631 to .315 £/sec), all flexible hose and rigid PVC piping is 3/4-inch.
The influent line to the main pump is a PVC line with a tee which is adapted
for NaOH injection. The other end of this line is fitted with an adapter for
attaching the 3/4-inch influent hose. The pump is a Teel Model IP 700 stain-
less steel centrifugal pump discharging 10 gal/min at 15 ft head. The pump
discharge passes through a 3/4-inch butterfly valve used to control the liquid
flow through the system. On the other side of this valve is a second tee used
for injection of ferric chloride. The PVC pipe leading from this tee goes to
the influent nozzle of the mixing chamber.
The discharge from the mixing chamber leads to a tee at which surfactant
(generally sodium lauryl sulfate, NLS) is injected. From this tee, a PVC pipe
leads to a flowmeter (Dwyer Ratemaster Flowmeter No. RMC-SSV-145) used to mon-
itor the flow. A section of PVC pipe leading from the flowmeter connects to a
flexible hose leading to the inlet at the top of the flotation column.
A section of PVC pipe is connected from the outlet at the bottom of the
17
-------
I
28
1. Waste Tank
2. Waste Tank Valve
3. Sodium Hydroxide Injection Tee
4. Main Pump
5. Flow Control Valve
6. Ferric Choride Injection Tee
7. Mixing Chamber
8. Control pH Electrode
9. NLS Injection Tee
10. Waste Flow Rotometer
11. NaOH Solenoid Valve
12. Electrical Junction Box
13. Control pH Meter
14. NaOH Tank
15. Ferric Chloride Fee; Pump
16. NLS Feed Pump
17. Ferric Chloride Tank
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
NLS Tank
Flow Dispersion Head
Column
Baffle
Air Diffuser
Air Supply Line
Air Pressure Regulator
Air Flow Rotometer
Monitoring pH Electrode
Column Liquid Level Control
Effluent Line
Monitoring pH Meter
Foam Breaker Motor
Foam Breaker
Clarifier
Clarifier Liquid Level Control
Broke Foam Container
FIGURE 2 - SCHEMATIC DIAGRAM OF 30 - CM PILOT PLANT
18
-------
FIGURE 3 - ADSORBING COLLOID FLOTATION PILOT PLANT
L9
-------
14
17
16
COMPONENT LIST
1. Waste Pump
2. Flow Control Valve
3. Mixing Chamber
4. Liquid Flowmeter
5. Flotation Column
6. NaOH Injection Tee
7. Ferric Chloride Injection Tee
8. NLS Injection Tee
9. Chemical Feed Pump
10.
11.
12.
13.
14.
15.
16.
17.
Ferric Chloride Container
NLS Container
Monitoring pH Meter
Monitoring pH Electrode
NaOH Container
Solenoid Valve
pH Control Meter
Control pH Electrode
18. Air Flowmeter
FIGURE 4 - PLAN VIEW OF PILOT PLANT, SHOWING
PUMPING, CHEMICAL FEED, AND pH CONTROL
AND MONITORING SYSTEMS.
20
-------
flotation column to a tee which houses an electrode for pH monitoring. The
other end of the tee attaches to a flexible hose drain line. This drain line
contains a tee open to the atmosphere; this prevents siphoning from occuring
when this section of the drain line is raised or lowered to control the level
of the liquid pool in the bottom of the column.
MIXING CHAMBER
The mixing chamber, shown in Figures 2 and 5, is constructed from a 30.5-
cm x 29.2-cm ID (12 x 11.5-inch) section of clear lucite pipe. The total vol-
ume of the mixing chamber is 20.45, (0.72 ft3) . This gives a detention time of
about 1 minute when the pilot plant is operated at a flow rate of 0.32 £/sec
(5 gal/min). The bottom of the chamber is a 1.90-cm (3/4-inch) solid piece of
lucite which is cemented in place. The top of the chamber is fitted with a
3.8-cm (1 1/2-inch) flange made from 3/4-inch lucite and glued in place. A
3/4-inch solid top is bolted to this flange and sealed with an 0-ring. A pH
probe is mounted in the top of the mixing chamber for control of the pH in the
flotation column. Inlet and discharge ports are mounted in the bottom of the
chamber. The inlet to the chamber is connected to a PVC elbow inside the
chamber, which mounts a nozzle made by drawing a piece of 3/4 inch PVC pipe
down to a 1/4-inch opening at one end. This constriction increases the dis-
charge velocity into the chamber, enhancing the mixing action.
FLOTATION COLUMN
The flotation column is shown in Figures 2, 3 and 6. A 29.2-cm (11.5-
inch) ID (30.5-cm, 12-inch OD) lucite pipe was chosen for the construction of
the flotation column, because this size was small enough to result in a rea-
sonably mobile unit, large enough to indicate the feasibility of possible
full-scale future systems, and this size was readily available. The column
was constructed from two 122-cm (4-ft) x 29.2-cm ID sections of clear lucite
pipe. Each end of both sections is fitted with a 3.8-cm (1 1/2-inch) flange
made of 1.9-cm (3/4-inch) lucite which is glued in place. The two sections
are bolted together by 8 bolts through the flanges, and are sealed with an 0-
ring. The two remaining ends of the joined sections have solid 3/4-inch
lucite covers bolted and sealed in the same way.
The bottom cover plate of the flotation column is fitted with an effluent
discharge line and an air inlet line, which leads to an air diffuser for gen-
erating the foam. House air is used for the column. This air passed through
a pressure regulator and then through a 0.64-cm (1/4-inch) tube to an air
flowmeter (Dwyer Ratemaster Flowmeter No. RMC-SSV-102) mounted on the support
frame of the apparatus. The discharge line from the flowmeter is a 3/4-inch
PVC pipe which leads to the bottom cover plate of the flotation column. The
air diffuser is a 12.7-cm (5-inch) fritted glass disc (modified fine fritted
Biichner funnel) .
The top cover of the flotation column is fitted with a 10-cm (4-inch) PVC
wye, which is the discharge point for the foam generated in the column. A 3/4
inch PVC pipe is run down through the center of this wye by use of adapters
and a compression fitting. (See Figure 6) The top end of this pipe is con-
nected by a hose adapter to the flexible hose bringing the wastewater from the
21
-------
7
12"
11.5"
MIXING CHAMBER COMPONENTS
1. 12-ln. x ir/2-ln. ID Lucite Section 5.
2. 3/4-ln. Lucite Bottom 6.
3. 3/4-ln. Lucite Flange 7.
4. 3/4-ln. Lucite Cover 8.
Typical Mounting Bolt (8 Total)
Mixing Chamber Inlet
Mixing Chamber Effluent
Control pH Electrode
FIGURE 5 - MIXING CHAMBER
22
-------
8
9-
13
11
i 4
FLOTATION COLUMN COMPONENTS
1. Flotation Column Section
2. Top Cover Plate
3. Flange
4. Bottom Cover Plate
5. Air Inlet
6. Air Diffuser
7. Waste Inlet
8. Compression Fitting
9. Adapter
10. Flow Dispersion Head
11. Effluent Line
12. Baffle Structure
13. 4-lndi PVC Wye
14. Foam Discharge Pipe
FIGURE 6 - FLOTATION COLUMN
-------
influent flowmeter. The pipe extends down the axis of the column for about
80-120 cm (31-47 inches) , and the influent dispersion head is mounted on its
lower end.
The dispersion head was designed to have a linear discharge velocity of
40 cm/sec (1.3 ft/sec) at its orifices at a flow rate of 7.6 £/min (2 gal/min);
the total orifice area required is 3.16 cm^ . The optimum number and size of
the openings were found by trial and error; the best design (which was used in
the work reported here) is shown in Figure 7. Each of the eight radial arms
contains nine holes 0.238 cm (3/32 in.) in diameter.
A system of baffles was constructed and mounted in the lower (stripping)
section of the column to reduce foam overturn and channelization. The baffles
were cut out of 1.27-cm (1/2-in.) plastic "egg carton" - a decorative grating
for lighting fixtures. The assembly consists of seven baffles 27.9 cm (11 in)
in diameter mounted 15.2 cm (6 in.) a part on three 0.64-cm (1/4-in.) threaded
nylon rods. Nylon nuts and washers were used to secure the baffles to the
rods. See Figure 8.
FOAM BREAKER
The foam breaker, shown in Figure 9, operates by allowing the foam to im-
pinge on a disc spinning at 2-3000 rpm. The shearing action at the contact
destroys the foam very effectively. The housing for the foam breaker was con-
structed from a 15.2-cm (6-in.) x 29.2-cm ID (11.5-in.) clear lucite pipe fit-
ted with a 1.9-cm (3/4-in.) thick lucite cover. A 4-in. wye is mounted on the
top of the cover and connected by a 4-in. PVC pipe to the wye on the top of
the flotation column. A 1/8-HP, 3000-rpm permanent split capacitor motor
mounted on top of the wye drives the disc. A 1.6-cm (5/8-in.) aluminum shaft
with bearings is connected to the motor shaft; the other end extends into the
foam breaker housing where it drives a 0.32-cm (1/8-in.) x 27.9-cm (11-in.)
diameter nylon disc. Clearance between the disc and the top of the housing is
1 .3 cm (1/2 in.).
CLARIFIER
The clarifier (see Figure 10) receives the collapsed foamate from the
foam breaker and separates the (floating) solids from the liquid. The clar-
ifier was made by cutting the bottom from a 49-2- (13-gal) polyethylene carboy
and mounting a level control and drain line (1.9-cm, 3/4-in. PVC) in the screw
on cap of the inverted carboy.
CHEMICAL FEED SYSTEM
The chemical feed system is shown in Figures 2 and 3. Ferric chloride
and sodium lauryl sulfate (NLS) solutions are stored in two 7.6-fc (2-gal)
polyethylene jugs on the shelves of the pilot plant support frame. The jugs
are connected by 0.32-cm (1/8-in.) plastic tubing to two pump heads (Cole
Parmer Instrument Co. Nos. 7016 and 7016-20) driven by a Masterflex Variable
Speed Pump Drive (Cole Parmer No. 7545-10). The discharges from the pump
heads are injected into tees in the influent line to the column as mentioned
earlier. Table 5 shows the feed rates of the 7016 at \arious pump speed settings.
24
-------
COMPONENTS
3/32-ln. Orifice
3/4-ln. PVC Inlet
5-ln. x 1/4-ln. PVC Pipe Nipple
1-1/2-ln. PVC Cap
Nylon Plug
1-1/2-ln. x 3/4-ln. Adapter
^
^
^
3
/
I
FIGURE 7 - FLOW DISPERSION HEAD
25
-------
I I I I I I
i i i i i rn
COMPONENTS
1. Flotation Column Wall
2. Nylon Support Rod
3. Plastic Baffle
4. Nylon Washer
5. Nylon Nut
FIGURE 8 - TYPICAL BAFFLE INSTALLATION
26
-------
COMPONENTS
1. Foam Breaker Housing
2. 4-ln. PVCWye
3. Foam Breaker Drive Motor
4. 5/8-ln. Aluminum Drive Shaft
5. 11-In. Nylon Disc
6. 4-ln. PVC Elbow
7. 4-ln. PVC Foam Inlet
FIGURE 9 - FOAM BREAKER
27
-------
COMPONENTS
1. Clarifier
2. Level Control
3. Clarifier Discharge Line
FIGURE 10 - CLARIFIER
28
-------
TABLE 5. FLOW RATES FOR VARIOUS SETTINGS OF
CHEMICAL FEED PUMP 7016
Pump Speed Feed Rate
Setting (ml/min)
10 100
9 84
8 75
7 61
6 54
5 52
4 16
-------
pH CONTROL AND MONITORING SYSTEM
The pH systems are shown in Figures 2 and 4. The pH electrode (Kruger
and Eckels No. 80278) mounted in the mixing chamber is connected to a pH meter
controller (Horizon Model 5650) mounted on a support frame shelf. This pH
controller operates a solenoid valve (Peter Paul No. 22R9DCV-12060) which pro-
vides on-off control of the flow of sodium hydroxide into one of the tees
mounted in the influent line before the mixing chamber and the main pump. A
7.6-5, (2-gal) plastic jug is connected to the solenoid valve by a 0.32-cm (1/8
in.) plastic tube. The jug is mounted on the top support frame shelf, which
allows the sodium hydroxide to flow by gravity into the influent line.
The pH of the flotation column effluent is monitored by a glass electrode
mounted in a tee in the column effluent line. This electrode is connected to
a monitoring pH meter (Sargent-Welch Model LSX) mounted on a support frame
shelf.
ELECTRICAL CONTROL PANEL
The electrical control panel is shown in Figure 3. All the power cords
to the electrical units of the pilot plant are connected to this panel. The
panel has a toggle switch, indicator light, and fuse for each electrical unit.
Other electrical controls, such as those for the pH meters and the chemical
feed pump control, are located on the individual units. The entire system
(and the support frame) is grounded, since otherwise the electrical hazard
could be quite serious.
SUPPORT STRUCTURE
The pilot plant support structure is shown in Figure 11. The frame con-
sists basically of a 91-cm x 152-cm (3-ft x 5-ft) vertical rectangular steel
frame mounted on a 91 x 152-cm (3 x 5-ft) steel base frame covered withe 3/4-
in. plywood. There are also two 91 x 46-cm (3x1 1/2-ft) shelves mounted on
a 91-cm (3-ft) frame; this trame is also mounted on the base. The support
structure is mounted on casters.
WASTE PRETREATMENT AND STORAGE TANK
A 1136-& (300-gal) cylindrical polyethylene tank fitted with a 1.9-cm
(3/4-in.) gate valve at the base was used for mixing of simulated wastes and
pretreatment and storage of industrial wastewaters. The valve at the base of
the tank is connected to the inlet hose of the pilot plant.
WORK ON SIMULATED LEAD-BEARING WASTES PROCEDURES
Chemical feed solution and the simulated waste were prepared for each
run. The simulated waste was made by dissolving the required weight of reagent
grade lead nitrate in about 4£ of tap water acidified with roughly 5 ml of
nitric acid. This solution was then mixed in the required volume of tap water
in the waste storage tank. In all runs the waste solution was prepared to
contain approximately 20 mg/£ of lead. When sodium chloride was used to in-
crease the ionic strength, the appropriate amount of commercial rock salt was
30
-------
o
COMPONENTS
1. 3-ft x 5-ft Base
2. 3-ft x 5-ft Support Frame
3. Equipment Shelves
FIGURE 11 - SUPPORT STRUCTURE
31
-------
weighed out and added to the storage tank with sufficient water to dissolve
it. After it was dissolved with stirring, water and lead nitrate solution
were added to make up the desired final volume.
The chemical feed solutions of technical grade ferric chloride and lab-
oratory grade ferric chloride and laboratory grade NLS were made in 7-£ quan-
tities in the containers, from which they were fed to the influent line. The
concentrations of the solutions were dependent on the waste flow rate, chem-
ical feed rate, and the concentrations of reagents desired in the waste. A
sample calculation for determining these solution concentrations is shown
below.
Given: Waste Flow - 2 gal/min (7.57 1/min)
Chemical Feed Pump Setting - 6 (54 m 1/min)
Desire: NLS Concentration - 50 mg/1
Fed 3 Concentration - 100 mg/1
7 liters of each of the above solutions.
XLS Calculation:
Waste Desired Volume Solution
Flow (1/min) X Cone, (mg/1) X Desired (1) X 1000 ml/1 .
Chemical Feed Pump Rate (ml/min) X 1000 mg/g ~ bolute Weig«t UJ
7.57 1/min X 50 mg/1 X 7 1 X 1000 ml/1
54 ml/min X 1000 mg/g g
FeClj Calculation:
Waste Desired Volume Solution
Flow (1/min) X Cone, (mg/1) X Desired (1) X 1000 ml/1 c . ^ . ,
i—L i \ °j.—<. :—i '.— — So Infp WPI ant
Chemical Feed Pump Rate (ml/min) X 1000 mg/g X 0.5443 g
7.57 1/min X 100 mg/1 X 7 1 X 1000 m/1 _ 2qt- „,
54 ml/min X 1000 mg/g X 0.3443 " ' g
From results of these calculations, it is seen that, if the waste flow
rate is 2 gal/min and the chemical feed pump feed rate is 54 ml/min, it
would require 49.06 g of XLS dissolved in 7 liters of water to give a concen-
tration of NLS in the waste of 50 mg/1. It would also require 285.01 g of
FeCl^ dissolved in 7 liters of water to give a concentration of 100 mg/1
Fe(III) in the waste.
-------
Seven liters of 0.25-molar (M) sodium hydroxide solution for pH control
were also prepared for each run. From preliminary experimentation this was
found to be the optimum concentration for pH control in the apparatus. Tech-
nical grade flake sodium hydroxide was used.
After all the solutions were prepared, the pH meters were calibrated
against two or three standard buffer solutions. The pH electrodes required
occasional cleaning with HC1 followed by soaking in KC1 solution to remove
ferric hydroxide deposits which gradually decreased the rate of response of
the pH meters as it built up.
The first step in actually making a run with the column was to turn on
the chemical feed pump, check its flow rates, and make any adjustment neces-
sary to obtain the desired reagent solution flow rates. After this check,
the chemical feed pump, air supply, main pump, and sodium hydroxide solenoid
valve were turned on, and the pH meters were turned from standby to pH. The
waste flow control valve was then adjusted to maintain the correct flow and
the air flow rate was set to the desired level. After the main pump had run
for about two minutes, the liquid level in the column was set by adjusting
the level of the outlet hose. The pH controller was checked and set to main-
tain the desired pH range. The foam breaker was turned on when the column
was filled with foam.
Throughout each run, all of the parameters were frequently checked and
control settings adjusted to maintain the desired values. Influent flow
rate, liquid level in the column, and column pH required frequent adjustment;
pH control needed closest attention.
In a properly functioning run, the liquid in the bottom of the column
would quickly become clear, and the ferric hydroxide would be carried upward
in the foam immediately after leaving the dispersion head.
Samples of column effluent were taken at various times during the course
of the run; samples of the influent were taken at the beginning and end of
each run. All samples were stored in polyethylene bottles and were acidified
with nitric acid before analysis. Lead analyses were carried out on a
Perkin-Elmer Model 305 B atomic absorption spectrophotometer at 217 nm.
Stock solutions and standards were prepared from reagent grade lead nitrate
and nitric acid (5 ml concentrated acid per liter of solution) in deionized
water.
Out objectives in this phase of the work were to continue previous con-
tinuous flow system studies of lead removal by adsorbing colloid flotation
with ferric hydroxide and NLS and to determine the feasibility of developing
this treatment method on a larger scale than that of Hanson's 10.2-cm ID x
186-cm continuous flow column described in Section 5. Her work involved ad-
justing the pH and adding the process chemicals to a container of lead-
bearing waste; this was then pumped to the column. We felt that automatic pH
control and a continuous chemical feed system should be developed and demon-
strated to show the feasibility of practical large-scale systems.
33
-------
Hanson suggested a set of optimum operating parameters determined from
her work on the smaller continuous flow column; these are shown in Section 5
and we used them as a starting point with the 30-cm column. Some preliminary
runs were made with lead-free influent, but most runs were made with simula-
ted waste containing approximately 20 mg/£ of lead added as lead nitrate.
34
-------
SECTION 7
PILOT PLANT RESULTS (30-cm DIAMETER COLUMN)
SYNTHETIC LEAD-CONTAINING WASTES
In this study we wished to achieve an optimum balance between the various
parameters which would yield good quality effluent under economical operating
conditions. Influent flow rate, air flow rate, and pH were easily adjusted
during the course of individual runs; NLS concentration, ferric iron concen-
tration, and ionic strength were also varied during the course of individual
runs, but these parameters are not as flexible as the first three. The para-
meter values used in each run and the results from each run are listed in
Appendix B. We next discuss the individual parameters and the results of
varying them.
The pH was monitored and controlled at the mixing chamber, which was lo-
cated on the influent line just downstream of the point at which sodium hydrox-
ide was injected for pH control. pH was also monitored in the column effluent.
In our apparatus, 0.25 molar was the optimum concentration of the sodium
hydroxide solution used for pH control. If a more concentrated solution were
used, the pH control system over-compensated and the pH oscillated excessively.
At much lower concentrations, the sodium hydroxide solution had to be reple-
nished too frequently. A time lag between the two pH displays was observed.
This lag was due to the time required for liquid to flow through the apparatus.
There was also a roughly constant difference of about one pH unit between the
two displays, due to the introduction of NLS downstream from the monitor-
controller electrode. The alkaline NLS caused the pH of the column effluent
to be higher than that of the influent in the mixing chamber. We adjusted the
pH monitor-controller to a lower pH setting to yield the desired pH in the
effluent.
The response time of the monitor-controller was observed to increase
after several runs had been made, resulting in excessive oscillation of the pH.
This resulted from the gradual formation of a deposit of ferric hydroxide on
the glass electrode. We reconditioned the electrodes by cleaning them in
hydrochloric acid (1 part concentrated acid to 3 parts deionized water) and
then soaking them in concentrated KC1 solution overnight. With a freshly pre-
pared electrode, the pH control system could maintain the pH within a range
of 0.4 pH units for several hours.
The operating pH was varied over a wide range in the pilot plant runs.
In order to obtain the best separation and good foaming conditions, the pH
should be in the range 6.0 to 7.0. (This range is somewhat higher than that
recommended by Hanson, 5.5-6.5.) The dependence of effluent lead
35
-------
concentration on pH is shown in Table 6. At pH's much above 7.2, the flo-
tation of the ferric hydroxide floe is poor.
TABLE 6. EFFECTS OF pH CN LEAD REMOVAL
pH Residual Lead (mg/£)
5.6
6.4
6.6
6.7
6.9
7.0
7.2
7.0
1.5
0.9
0.5
0.6
0.8
0.9
The maximum feasible hydraulic loading rate was determined by the geo-
metries of the column and the influent dispersion head, and by the character-
istics of the foam itself. The linear velocity of flow from the dispersion
head could not be too great or it would break the foam and cause overturning.
The influent also had to be spread evenly across the column without being
directed against the column sides. With our best influent dispersion head,
an influent flow rate of 7.57 £/min (2.0 gal/min) was the maximum which con-
sistently did not produce overturning and/or channelling. This corresponds
to a hydraulic loading rate of 163 m-Vday-m- (2.77 gal/ft- min) . Several
different configurations and sizes of holes in the radial arms of the dis-
persion head were used over a range of flow rates. With all of these, the
maximum hydraulic loading rates which could be tolerated fell in the range
120-180 ml/min-m2 min (2-3 gal/ft2 min).
The quality of effluent was very strongly affected by hydraulic loading
rate. Lead concentrations increased dramatically at hydraulic loading rates
at which channeling and overturning occurred.
We found that inclusion of a baffle system in the lower (stripping)
section of the column improved column performance at high hydraulic loadings.
In preliminary runs (without lead) it was visually apparent that the baffle
system was effective to a limited extent in controlling foam overturn and in
reducing the size of the eddies when overturn did occur. Xear the baffles
the foam appeared to move in a plug flow pattern even when overturning was
taking place in the regions between the baffles. Plug flow could probably
be further stabilized by the inclusion of more baffles. We also observed
that the baffles and their support rods tended to cause channelling down the
wall of the column if they were allowed to touch the wall.
The depth of liquid in the bottom of the column was varied in the preli-
minary runs. When the surface of the liquid was more than 2-3 cm (1 in)
above the air diffuser, foam production was reduced (bubble size was increas-
ed), and the larger bubbles and turbulent motion of the liquid surface had a
detrimental effect on the plug flow pattern of the foam upward through the
column.
36
-------
We destroyed the first air diffuser by permitting the air pressure to
become excessive as the diffuser fouled with oil from dirty compressed air.
A filter and pressure regulator were installed, and the air pressure was kept
below 10 psi (0.7 kg/cm^). In the preliminary runs, an air flow rate of 1.1
m^/hr (40 SCFH) was used; this was satisfactory in some runs, but was not
sufficient to push the foam up more than 3/4 of the length of the column in
others. Usually, when the foam collapsed at this point ferric hydroxide de-
posited thickly on the wall, building up until eventually heavy clots of
ferric hydroxide-rich dry foam broke loose from the wall, fell down through
the rising, less df-nse foam, and contaminated the effluent pool at the bottom
of the column.
We found that an air flow rate of 1.7-2.0 m3/h (60-70 SCFH), corres-
ponding to 0.4-0.5 m^/min-m^, markedly improved column performance. We note
that Hanson's recommended range of 0.2-0.3 m-Vmin-nr was significantly lower,
and believe that the air flow required by the present column could be reduced
if the upper section of the column were shortened. We originally thought
that this length was necessary for adequate drainage of the foam, but obser-
vation of the operating column leads us to believe that there was very little
benefit from draining in the upper one-fourth, roughly, of the column.
Increasing ionic strength (dissolved salts concentration, essentially)
has been shown experimentally (56,57) to markedly decrease lead removal ef-
ficiency, as is predicted theoretically (39,40). Therefore, a number of runs
were made with all variables except ionic strength held constant. (Some un-
avoidable minor variations in pH also occurred.) Some of the results of
these runs, shown in Table 7, show the detrimental effects of elevated ionic
strength. Commercial rock salt was used to vary the ionic strength.
TABLE 7. EFFECTS QF_INCREASING IONIC STRENGTH ON LEAD REMOVAL
NdZi Added Residual Level
(molesA) (mg/£) pH (mg/fc)
0.01
0.01
0.05
0.05
0.05
0.07
0.07
0.07
690
690
3450
3450
3450
4830
4830
4830
6.6
6.4
6.5
6.3
6.4
6.5
6.7
6.7
0.86
1.57
2.29
1.86
2.43
2.57
3.14
3.29
Note: All runs made with 20 mg/£ Pb(II), 100 mg/fc
and 50 mg/£ NLS. Air flow rate was 40 SCFH and liquid
flow was 2.0-2.5 gal/min.
We could not consistently produce stable; foams at an influent air flow
rate of 7.6 2,/min (2 gal/min) if concentrations of NLS less than 40 mg/£ were
used. Hanson (Section 5) had recommended an NLS concentration of 35-40 mg/£,
and the results of this work are consistent with that figure. Some dissolved
NLS was carried out in the effluent, as indicated by the formation of a light,
relatively non-persistent foam on shaking an effluent sample.
37
-------
The problem, mentioned earlier, of foam collapsing halfway up the drain-
age section of the column was thought to possibly be due to decreased temper-
ature of the influent and column during the winter months. Use of simulated
wastewater at a temperature of 27°C (80°F) yielded marginal benefits, if any.
The optimum Fe(III) (ferric iron) concentration was found to be in the
range of 100-150 mg/£. At Fe(II) concentrations less than 100 mg/£, lead re-
moval was markedly less complete, and, at concentrations greater than 150
mg/£, at times not all the ferric hydroxide (containing lead) was removed.
These results are in agreement with Hanson's recommended Fe(III) concen-
tration of 150 mg/£.
The spinning disc foam breaker, built on the principles of one described
by E. Rubin (73,74), easily handled the maximum flow of foam that the column
would deliver. The disc must be carefully balanced to avoid excessive vi-
bration and wear on the bearings. After the foam was broken, the collapsed
foamate drained down the inside of the disc housing and dripped into the
clarifier unit. The ferric hydroxide, lead, and much of the NLS formed a
scum and floated on top of the liquid in the clarifier. The clear liquid
drained from the clarifier contained somewhat less than half of the NLS in
the influent; the bulk of the rest was in the floating ferric hydroxide-lead
sludge, from which it could be displaced by sodium carbonate. This com-
plicates surfactant recycling, but does not make it impossible. The adsorp-
tion of surfactant on floes, and its displacement by salts, is discussed in
much more detail in Section 9 and Appendix D.
We were disappointed somewhat to note that the levels of lead removal
attained with this pilot plant were not as favorable as those obtained in lab
scale batch equipment, a lab scale continuous flow apparatus, and the 10-cm
continuous flow pilot plant described in Section 5. Our earlier work had
fairly routinely achieved residual lead levels of 0.1 mg/£ or less if the
ionic strength of the solution was not too great; in this study we were not
able to reduce lead levels below 0.5 mg/£. It was some time before we re-
alized the cause of this discrepancy. In all of the laboratory scale work
and in the 10-cm pilot plant the simulated waste was mixed with ferric
chloride and the pH adjusted to the desired level at least 5 to 10 min before
entering the column itself. In the present (30-cm) apparatus, the time in-
terval between the formation of the ferric hydroxide floe and the entrance of
the influent into the column was of the order of only 1 min. Preliminary lab
scale studies on the adsorption and/or coprecipitation of lead with ferric
hydroxide indicate that 1 min is not long enough to permit adsorption to
reach equilibrium. We are currently modifying the apparatus to increase the
time interval during which flocculation and adsorption can take place. We
believe that this should improve the performance of the apparatus substant-
ially.
In summary, the study of simulated lead-bearing wastes in the 30-cm
pilot plant leads us to the following conclusions:
1. Channeling and foam overturn are the determining factors of
effluent quality at high hydraulic loadings. They can be con-
trolled somewhat by a dispersion head which spreads the feed
38
-------
uniformly at low linear velocity over the cross-sectional
area of the column without excessively reducing the cross-
sectional area of the column at that point. Baffles and
dispersion head must be level, and the column must be pre-
cisely vertical.
2. Increasing ionic strength above about 0.05 molar (about
3000 mg/£ of NaCl) decreases lead removal.
3. NLS concentrations in excess of 40 mg/Jl are not necessary;
concentrations much below this level often did not pro-
duce sufficiently stable foams.
4. The optimum pH range is about 6.0-7.0; substantial vari-
ations outside of this range result in severe deterioration
of the separation.
5. The most efficient Fe(III) concentration is from 100 to 150
mg/£.
6. Hydraulic loading rates up to 150-180 m-Vday-m2 are feas-
ible} and air flow rates of 0.4-0.5 mS/min-m2 seem to be
optimal.
7. Design of the apparatus could be improved by increasing the
size of the mixing chamber to permit a detention time of at
least 5 min, by decreasing the length of the foam drainage
section of the column to about 50 cm, and probably by the
inclusion of more baffles in the stripping section.
The optimum operating parameters for this apparatus are summarized in
Table 8.
TABLE 8. OPTIMUM. OPERA TING PA H\ METERS
pH 6.0-7.0a
Fe(III) concentration 100-150 mg/2,
NLS concentration 150 35-40 mg/£
Hydraulic Loading Rate 150-180 m3/day-m2 (2.4-3.0 gal/min-ft2)
Air Flow 0.4-0.5 m3/min-m2 (1.3-1.6 ft3/min-ft2)
aFor high ionic strength wastes, a lower pH than 6.0 may be
necessary.
FIELD TESTS ON A ZINC AND COPPER-CONTAINING WASTEWATER
Arrangements were made to carry out field tests at a plant manufacturing
water heater components. The plant conducts an automated copper descaling and
zinc plating operation, with cleaning and pickling steps and countercurrent
rinses. A zinc chloride bath is used for the plating and a hydrochloric acid
pickling step is employed. Two commercial cleaners are used; these contain
salts of polyvalent anions which undergo extensive hydrolysis, and a variety
39
-------
of anionic and nonionic surfactants. Their exact compositions are proprietary.
Spent pickle and spent plating solution are discharged from time to time to a
holding tank, from which a small flow of acidic concentrated waste is contin-
uously blended in a mixing tank with the alkaline overflow from the counter-
current rinsing operations. After mixing and adjustment of the pH to approxi-
mately 8.5, the wastewater is sent to an upflow clarifier, where zinc hydrox-
ide floe is removed. The clarifier effluent is then discharged to the city
sewer mains. The waste in the mixing tank was used for our field tests. It
showed considerable fluctuation in composition, ranging in Zn(II) concen-
tration from roughly 30 to roughly 400 mg/£. The concentration of Cu(II) was
substantially lower, generally around 1 mg/£. Total solids were 1-2 mg/£.
Waste was pumped from the mixing tank to our pretreatment and storage
tank, from which it was then pumped through the foam flotation pilot plant as
described in Section VI. Our original plan was to treat this wastewater by
the same procedure used for the simulated lead wastes - to add 100-150 mg/£ of
Fe(III), adjust the pH to 6.0-6.5, add 30-35 mg/£ of NLS, and run the waste
through the column. This proved a failure. The presence of polyvalent hydro-
lyzable anions from the cleaners (phosphates and silicates) in the wastewater
resulted in a change in sign of the zeta potential of the floe (normally posi-
tive, in these wastes the floe zeta potential was usually negative). These
ions sorb very strongly on the Fe(QH)3 floe and prevent attachment of the floe
to the air-water interface. (The theory of this competition is discussed in
Appendix D.)
An alternative approach was therefore developed in which the interfering
anions were precipitated with lime (approximately 5 g of lime per liter of
waste) and the supernatant waste, containing 8-20 mg/£ of Zn(II) and 0.1-0.2
mg/£ of Cu(II) was then foamed with 50 mg/£ of Fe(III) and the polyethoxy-
ethanol surfactant present as a contaminant in the wastewater. No NaOH was
added. The pH's were around 11. Effluent levels of Zn(II) in the range 0.7-
11 mg/£ were achieved; Cu(II) was removed to below our limit of detectability
(O.05 mg/i). Insufficient surfactant remained in the effluent to maintain a
stable foam on shaking. The data are presented in Appendix B.
We found that the upper (foam draining) section of the column was some-
what too long for the concentrations of surfactant which occurred in this
wastewater; the accumulation of heavy, sludge-bearing material in the long
upper section of the column was occasionally too great to be supported by the
surfactant-lean foam, resulting in lumps of this heavier material occasionally
falling through the foam and contaminating the effluent. The collapsed
foamate readily separates into a floating sludge and a clarified surfactant-
rich bottom layer, which could advantageously be recycled into the influent,
so there is no reason to strive for an extremely well-drained foam. A foam
drainage section of the column of 50-60 cm length should be ample, and should
prevent the occasional accumulation of sludge, which occurred in the top 30-40
cm of the column with this waste. (The present length of this section is 120
cm).
A small amount of work was carried out with wastewater obtained from the
electroplating line of an automobile bumper recycling operation. The com-
bined wastewater contained 350-1900 mg/£ total solids; had a conductivity of
40
-------
365-1300 micromhos/cm; contained CrO^" (10-120 mg/S.) , C u(II) (from a cyanide
bath, 18-230 mg/£), Ni(II) (7-110 mg/£) and cleaners. Two runs were made with
hauled waste (space restrictions precluded setting up the apparatus at the
plant), with rather erratic results. Pretreatment with sodium metabisulfite
was necessary to convert Cr(VI) toCr(III). Lime precipitation was necessary
to remove polyvalent anions in the cleaners. Separation in both instances
were erratic at pH 6.5-7.0, 150 mg Fe(III), 40 mg/£ MIS. The best results ob-
tained wereCr(III) of 0.5 mg/£, Cu(II) < 0.5 mg/SL, and Ni(II) of 0.7 mg/£.
These were not achieved with reliability, and the Fe(CH)3 frequently washed
through the column and contaminated the effluent. We believe that the diffi-
culty was caused by the rather high ionic strength of this waste after pre-
treatment with 300 mg/£ of sodium metabisulfite and 1000 mg/£ of lime.
Preliminary batch tests indicated that smaller concentrations of the pretreat-
ment chemicals were not adequate to convert Cr(VI) toCr(III) and to precipi-
tate interfering polyvalent anions from the cleaners. Our first report
demonstrated the inability of flotation with NLS to deal with synthetic mix-
tures of high ionic strength, and we believe that our difficulties here reflect
this fact.
41
-------
SECTION 8
CONTINUOUS FLOW AXIAL MIXING STUDIES
In our first report we described a mathematical model, analyzed in more
detail elsewhere (27), of foam flotation stripping column operation. One of
the terms included in that model was an empirical correction term to take
axial dispersion into account. The dispersion constant which appears in this
term is not easily assessed on theoretical grounds. Our experimental work on
continuous flow stripping columns indicated that axial dispersion is an im-
portant factor in column performance, especially at high hydraulic loadings,
at which channeling and foam overturn occur. We here describe experimental
work carried out on a 10-cm continuous flow stripping column to give more de-
tailed insight into the factors affecting axial dispersion.
APPARATUS AND l^ETHODS
Figure 12 diagrams the flotation system used in these studies. The
column, 10.2 cm ID x 183 cm, was made of lucite. Influent was introduced at
or slightly above the midpoint of the column through a 0.64-cm tube, entering
through the top coverplate of the column and exiting through the dispersion
head detailed in Figure 13. The most important aspect of the dispersion
head design seemed to be that one must avoid spraying the influent onto the
wall of the column, which results in channeling A 1.90-cm I.D. lucite pipe
mounted in the top coverplate of the column permitted discharge of the foam to
the foam breaker, diagrammed in Figure 14.
Foam was generated by passing filtered, humidified air through a porous
cylindrical air stone mounted on the bottom coverplate. The air flow rate was
measured and controlled by a flowmeter. An exit port 1.5 cm above the bottom
of the column allowed discharge of effluent through a rubber tube and screw
clamp. Inside the column, three or four baffles made of 1/8-in. mesh galva-
nized hardware cloth were used to control foam overturn. These were held in
place by metal washers made from large hose clamps.
A 190-£ polyethylene cylindrical tank was used to mix and store the water
to be treated. A stirring motor provided mixing adequate to keep the ferric
hydroxide floe in suspension, and a small centrifugal pump supplied a constant
head tank 180 cm above the dispersion head. Liquid flow rates to the column
were measured and controlled with a flowmeter.
The following method was used to examine axial dispersion. The column
was back-lighted by a 4-ft fluorescent lamp mounted vertically as close as
possible behind the column. Stray light was blocked by aluminum foil taped
42
-------
COMPONENTS
1. Constant Head Tank 9.
2. Influent Flowrneter 10.
3. Stirrer 11.
4. 3000 rpm Motor 12.
5. High Speed Foam Breaker 13.
6. Foamate 14.
7. Pump 15.
8. Liquid Reservoir 16.
FIGURE 12 - THE FOAM FLOTATION APPARATUS USED
IN THE AXIAL DISPERSION STUDIES
Dispersion Head
Air Flowrneter
Air Inlet
Glass Wool Filter
Baffles
Humidifier
Effluent Drain
Air Stone
43
-------
FIGURE 13 - DISPERSION HEAD DESIGNS
44
-------
COMPONENTS
1. Foam Inlet Tube
2. 3000 rpm Motor
3. Housing
4. Polyethylene Disc
FIGURE 14 - HIGH SPEED FOAM BREAKER DESIGN.
45
-------
from the lamp fixture to the column. The lucite walls of the column refracted
light around the column, obscuring events within; we therefore cut a slot from
top to bottom on each side of the lucite tube and filled these with an opaque
epoxy cement. This almost completely eliminated scattered light. Figure 15
shows the photographic setup.
Photographs of event sequences were taken on 35-mm film (Kodak TRI-X ASA
400, with a yellow filter). The developed negatives were scanned on a den-
sitometer capable of measuring up to three absorbance units (A.U.) with a pre-
cision of -0.01 A.U. These absorbance values were plotted by a strip chart
recorder. A "time zero" photograph was taken at the beginning of each run be-
fore the injection of ferric hydroxide was made; subsequent shots were com-
pared to this, and the difference between the initial absorbances and the
absorbances at time -£ plotted as functions of vertical position in the column
for various values of £.
NLS was used as the surfactant in all runs, and ferric hydroxide was the
floe which was removed by flotation. Ionic strength was adjusted with reagent
grade sodium nitrate, and the pH was adjusted with sodium hydroxide and nitric
acid, reagent grade. All chemicals were added to the water in the 190-£
storage tank; NLS was added after the precipitation of ferric hydroxide. The
parameters for the runs reported here are given in Table 9.
Two types of runs were made: (1) "pulse" type runs, in which 2 ml of
ferric hydroxide suspension containing 1000 mg/£ of Fe (III) was suddenly in-
jected into the feed line, and (2) "step change" type runs, in which ferric
hydroxide was present in the feed at all times and an abrupt change in influ-
ent hydraulic loading or ionic strength was made at some point during the
course of the run. Pulse type runs were more informative about axial dis-
persion of floe in the foam, while the step change type runs more closely
approximated actual operating conditions for foam flotation columns. Before
a photographic sequence was initiated for either type of run, the column was
brought to a steady state with a feed containing surfactant and sodium
nitrate.
EXPERIMENTAL RESULTS AND DISCUSSION
The experimental results for the runs described in Table are shown in
Figures 16 through 24. A number of other pulse runs made at higher hydraulic
loadings showed extremely rapid axial mixing by channelling and foam overturn.
When all conditions are nearly optimal, a pulse of ferric hydroxide is removed
very quickly, as seen in Figures 17 through 19. In Figures 17 through 20 the
ionic strength (0.062/molar) was near its maximum feasible value, and the foam
tended to be quite dry, enabling it to support a relatively high specific air
flow rate of 0.84 m-Vmin-m • At an ionic strength of 0.031 molar, however,
the foam was quite a bit wetter, especially at the higher hydraulic loading
rates, and the maximum feasible specific air flow rate dropped to 0.66m /min-
m2.
Figure 16 shows the results of an NLS concentration which is too high to
effect a good separation. At 200 mg/£ NLS, the ferric hydroxide floe is
46
-------
COMPONENTS
1. Clamp
2. Aluminum Foil Reflector
3. Opaque epoxy-filled slot
in column
4. Fluorescent Lamp
5. Camera
6. Clamp
FIGURE 15 - THE SETUP FOR PHOTOGRAPHY OF THE AXIAL
DISPERSION EXPERIMENTS.
47
-------
TABLE 9. PARAMETERS FOR FIGURES 16 - 24.
Figure
Number
16
17
18
19
20
21
22
23 .
24
£1,
pulse*
pulse*
pulse*
pulse*
pulse*
50
50
35
50
PH
5.5
5,5
5.5
5.5
5.5
5.5
5.5
4.0
5.6
(raol/S.)
.062
.062
.062
.062
.062
.031
.031
0+0.09
0+0.12
NLS
200
50
50
50
50
SO
SO
200+50
200+50
Air
(St/m^-min'
172
215
237
333
441
215
215
215
333
Flow
! (cm/sec)
.15
.19
.21
.29
.38
.19
.19
.19
.29
Water
(£/m2-min)
140
140
108
129
127
14+83
83+142
73
140
Flow
(cm/sec)
.12
.12
.09
.12
.11
.01+. 06
.06+. 12
.06
.12
* Each pulse was a 2-ml suspension of ferric hydroxide at a concentration of 1000 mg/fc.
-------
h^
.mill, inn.
hinllllilll Illllll..
iiiii.iiiilniiiiilli Hllllllii.
.illliii..in.> iHillll iillllllllllllllll
Ml Hi I .illlllllll Mill.
FIGURE 16 - AXIAL DISPERSION DATA. See Table 9 for run parameters.
Run I
49
-------
flu
iiiimiiiii
..ulil
O.
liln
.Illlllllh.
ll IlllHlhn.
Illfflllll..
Illl
O.
..ii.iinil.llllUlllllllllii
1
FIGURE 17
Run II
- AXIAL DISPERSION DATA. See Table 9.
50
-------
.llffllll..
o
.ill. .1
O j
...mi. Ilin.
o
i Jin.
Ulllllh
Illlllllllll,
o
.IHIIllllllllllllllll...
O
..MM.n.ililllllllllllli.
O
......illlllllllli,
FIGURE 18 - AXIAL DISPERSION DATA. See Table 9.
Run III
51
-------
-ilk.
o
...llnillllllll.
J_L
O
..Ililllllllll.
O
..nl Ililllllllll
mill
FIGURE 19
Run IV
- AXIAL DISPERSION DATA. See Table 9.
52
-------
1°
III
mlllllll
III!
.O
..i.nilh
o
.i.inili Illiiilliilliih.
o
......in mini i illlllliiin.
o
.ill! lllllllll..llll 111 Illlllll.
..I I
11
FIGURE 20 - AXIAL DISPERSION DATA. See Table 9.
Run V
53
-------
.ililllln lln
Hi. i in illHlllllliilihii ..in i
.. i.
Mill
,1.
....iiiiiiiiiiiii
.1
Illlllllllllllliln
Iniiii
FIGURE 21
Run VI
- AXIAL DISPERSION DATA. See Table 9,
54
-------
....I
ll 111 IIIII III III
Mil
III III li
.nniiH
Illiinlllin
ll
FIGURE 22
Run VII
- AXIAL DISPERSION DATA. See Table 9,
55
-------
.......iin.iilil
..MI
1
Mil
Illlllllii IIIIH nil . 1
iliiiiiillluli Inn
FIGURE 23 - AXIAL DISPERSION DATA. See Table 9
Run VIII
56
-------
mill
ILL
llii
u.
liillllllllll
FIGURE 24 - AXIAL DISPERSION DATA. See Table 9.
Run IX
57
-------
coated with a micellar double layer of surfactant ions, with the outer layer
presenting ionic heads (the sulfate groups) to the surrounding water. Floe
particles so coated are hydrophilic, air bubbles are unable to attach to them,
and flotation does not readily occur. The pulse of ferric hydroxide therefore
spreads throughout the column, and only a partial separation is achieved.
The consequences of having too large a specific air flow rate in an
otherwise stable floe removing run are show in Figure 20. There is so much
turbulence in the upper half of the column (above the influent dispersion
head) that the pulse is very rapidly spread out. The upper edge of the pulse
is rapidly carried out the top of the column, but the pulse is smeared out so
badly by turbulence that its lower portion clears the column quite slowly.
Actually; the specific air fldw rate is quite important to the flow pattern of
the foam (ideally plug flow), but seems to be less influential on the effect-
iveness of separation. Several runs were made at high specific air flow rates
at which turbulence occurred in the upper half of the column and nevertheless
ferric hydroxide removal was quite effective. Evidently, the deleterious
effect of increased axial dispersion is counterbalanced by the increase'd
linear velocity of the rising interfaces. Increased specific air flow rates
also resulted in larger bubble sizes and in an increase in the wetness of the
foam, due to decreased drainage time.
Step change runs are shown in Figures 21 through 24. Here, column oper-
ation was allowed to stabilize at a particular set of parameters, and then
either the influent hydraulic loading or the specific air flow rate was
changed to a new value. The dynamic response of the system to such changes is
seen in these figures.
Figure 21 portrays a change in the influent hydraulic loading from 8.6 to
52 m /day-m . In both cases the system restabilized in approximately 2 min,
and the separation was not impaired.
The runs depicted in Figures 23 and 24 exhibit the response of the system
to a sudden increase in ionic strength. As was usual for a continuous flow
test run, the column was primed with 2 to 4 £ of a 200 mg/£ NLS-tap water sol-
ution and air was bubbled into the column for several minutes to charge it with
foam. The influent solution containing ferric hydroxide and NLS (50 mg/£) was
then fed to the column until it was apparent that effective separation was
taking place. The ionic strength of the feed was then increased. In both
cases separation of the floe remained satisfactory initially, but then rapidly
deteriorated. After only minutes a new steady state was achieved with most of
the floe being discharged in the effluent from the bottom of the column. In-
terpretation of these two runs was complicated somewhat by a tendency of the
floe to adhere to the lucite wall of the column.
Figure 25 plots position of the "center of mass" of the pulse versus time
for the runs exhibited in Figures 17 - 20, in which these "centers of mass"
are indicated by encircled dots. The specific airflow rate and the mean floe
velocity are approximately equal for the three runs which exhibited plug flow.
The data presented here plus the results of other runs not shown indicate that
in this apparatus the specific airflow rate has a relatively small working
ranoe compared to the influent hydraulic loading, which was variable over a
much larger percentage change.
58
-------
£
Q
LU
I
z
O
CO
cc
LLI
D-
CO
Q
O
CC.
LU
O
D_
CO
Q
Fig. 19
Data
2.0 4
^ 1.54
1.04
Fig. 18
Data
0.54
0
-0.5 4
Fig. 17
Data
120
IGURi; 25 - CHNTHR OF MASS MOTION TOR FICURHS 17-20
-------
We had hoped to estimate axial dispersion constants by plotting pulse
spread as a function of time. The root-mean-square width of the pulse and its
width at half maximum should, according to diffusion theory, be given by
X% = 4(loge2
where D is the axial dispersion constant. Diffusion theory also predicts that
the pulses should be gaussian in shape. It is apparent from Figures 16
through 20 that the pulses deviate from gaussian form substantially; and we
see from Figure 26 that pulse width does not appear to increase proportional
to t^. Evidently, the axial dispersion processes taking place are not ade-
quately described by simple diffusion. We speculate that the size scale of
the eddies and turbulence contributing to axial dispersion is comparable to
the size of the pulse itself, which would invalidate the use of simple diffu-
sion theory except as a very rough approximation.
We conclude from this work on axial dispersion that, under conditions
where axial dispersion seriously affects column performance the mechanisms
responsible for axial dispersion, which include foam overturn and channelling,
are not well modelled by simple diffusion. At present, we can do little more
than employ specific air flow rates and influent hydraulic loadings which do
not yield excessive visible turbulence, channelling, and overturn. For this
apparatus, influent hydraulic loadings must be less than 170m3/day-m2, and
specific air flow rates should be in the range 0.18-0.3 m3/min-m2.
60
-------
1,5
1.0
0.5
"
-
-
i
Fig. 17
Data
2.0
1.5
1.0
0.5
40 80
TIME (Sec)
40
40
80
80
120
1.5
1.0
0.5
-
•
i
i
Fig. 18
Data
120
1.5
1.0
0.5
"
-
•
i
Fig. 19
Data
120
Fig. 20
Data
0
40
80
120
FIGURE 26 - PULSE WIDTH VERSUS TIME FOR FIGURES 17-20.
61
-------
SECTION 9
BATCH TECHNIQUE LABORATORY STUDIES
BATCH FLOTATION OF ZINC(II)
We report here experimental results on batch floe foam flotation of
zinc [I I) with aluminum and ferric hydroxides, with sodium lauryl sulfate
(NLS) as collector. Dependence of the separations on pH and ionic strength
was studied in some detail, and specific ion effects on the separation of
zinc(II) with AirOH)^ were investigated.
The equipment used was two batch columns essentially identical to the
apparatus described by us earlier (1,54,57). House air was passed through
ascarite, water (for rehumidification) , and glass wool, then through a "fine"
glass gas dispersion tube at the bottom of the column. Laboratory grade NLS
or HTA (hexadecyltrimethylammonium bromide) was used as the collector; all
other chemicals were of reagent grade. Stock solutions (1000 g/fc) of the
metal ions and surfactant were mixed in the desired amount and diluted with
deionized water to nearly 200 ml; ionic strength was adjusted by the addition
of sodium nitrate solution, and the pH adjusted by addition of dilute NaOH
and HN03- The sample was diluted to 200 ml and added to the column. The air
flow rate was measured with a soap film flowmeter. Air flow rates of approxi-
mately 65 ml/min were normally used. The pH was monitored during the course
of each run, and 7 ml samples were withdrawn from the bottom of the column at
5 min intervals. Zinc analyses were carried out by atomic absorption spectro-
scopy.
Kim and Zeitlin's studies (59) on the removal of zinc from sea water with
dodecylamine and ferric hydroxide at pH 7.6 resulted in 94% recovery from an
initial concentration of 3.2 ppb. Our work was carried out at a substantially
higher initial concentration of zinc; ferric or aluminum hydroxide was used as
floe, NLS was used as the collector, and pH and ionic strength were varied.
Ferric hydroxide is speedily removed by foam flotation with NLS at pH
5.5, but with pH's as high as 7.5 the separation was found to be slow and in-
complete. The precipitation point for Zn(OH)2 is roughly at pH 8 (5), so one
could not expect ferric hydroxide and NLS to be an optimum system for zinc
removal. Runs made with solutions initially containing 50 ppm NLS and 50 ppm
Zn(II) are listed in Table 11, and bear out this surmise; inclusion of 200 ppm
Fe(III) appears to interfere with the separation. The air flow rates in all
of the work on zinc are 60 to 68 ml/min unless otherwise specified. The
samples listed in Table 11 were taken after 5 min of treatment, at which time
Fe(OH)^ removal was essentially complete. There appears to be a competition
62
-------
TABLE 10. THE EFFECT OF Fe(OH)3 ON Znfl I) REMOVAL
Fe(III) concn (ppra) pH Final Zn(II) concn (ppm)
0
0
200
200
6.6
7.1
6.6
7.1
20
7.2
25
19
TABLE 11. THEJBFFECT OF pH ON Zn(II) REMOVAL WITH Fe(III) AND NLS
pH Final Zn(II) concn (ppm)
5.5
6.2
6.6
7.1
7.5
42
37
25
25
8
between the Zn(II) ions and the ferric hydroxide floe for the collector. The
effect of pH on Zn(II) removal is shown in Table 11. Initial concentrations
were 50 ppm Zn(II), 200 ppm Fe(III), and 50 ppm NLS. At a pH of 7.5 the iron
was incompletely removed.
It was found that pulse additions of Fe(OH)3 floe during the course of
the runs did not significantly improve the separations, nor did varying the
air flow rate. Increasing the duration of the run and/or the amount of NLS
added during the course of the run resulted in some improvement in separation,
presumably through an ion flotation mechanism. Our results indicate that this
is not an industrially feasible technique for removal of zinc from wastewaters,
inasmuch as the use of long time periods and high collector concentrations
resulted in final Zn(II) concentrations of about 3 ppm.
Huang (55) and Clarke (54) have removed A1(OH)3 with NLS at pH's close
to 8, the precipitation point of Zn(OH)2- We therefore attempted the removal
of Zn(II) with A1(OH)5 and NLS. The data listed in Table 13 indicate that
zinc is effectively flocculated with A1(OH)3 at pH's at which A1(OH)3 is re-
moved by foaming with anionic surfactants such as NLS. In these runs the
initial Zn(II) concentration was 50 ppm and the run duration was 15 rain.
Final Zn(II) concentrations of less than 0.5 ppm are attained at pH's between
7.7 and 9.2 by adsorbing colloid flotation, and the residual Zn(II) concen-
trations are roughly an order of magnitude lower than those obtained by ion
flotation alone with a similar quantity of surfactant.
The effect of increasing ionic strength at various pH's is shown in
Table 13. All runs lasted 30 min; the initial Zn(II) concentration was 50 ppm,
the intital Al(III) concentration was 100 ppm; in most of the runs (not star-
red) the initial NLS concentration was 100 ppm, with 50 ppm added after 15
min. The starred runs (see Table 14) had an initial NLS concentration of 150
63
-------
TABLE 12. THE EFFECT OF Alfllll ON Znflll REMOVAL
Final Zn(II)
NLS (ppm) concn (ppm)
PH
Al(III) (ppm)
6.9
7.4
7. 7
8.0
8.3
8.6
8.9
9.2
0
200
0
200
0
100
200
0
100
0
100
200
0
100
0
100
0
100
150
150
150
150
150
100
150
100
100
100
100
100
100
100
100
100
100
100
19.6
4.7
5.5
0.7
8.3
0.39
0.2
6.2
0.20
4.8
0.17
0.0
6.0
0.16
1.1
0.41
1.6
0.42
TABLE 13. FLOC FLOTATION OF ZINC WITH Al(OH) AND NLS. EFFECT OF
IONIC STRENGTH
Added NaN03 (mole/£)
0.5
0.10
0.15
0.20
0.30
0.40
Residual zinc (ppm)
9.2
8.9
8.6
8.3
8.0
7 . 7
7.4
0.06
0.03
0.09
0.07
0.10
0.27*
1.1*
1.
0.
0.
0.
0.
0.
18
59
11
13
22
32*
4.
0.
0.
0.
0.
0.
1
33
11
13
21
37*
9.2
1.9
0.15
0.12
0.42
0.50*
11
2
0
0
0
0
1
.2
.8
.60
.9
.40
.56*
.3*
1
0
0
0
.2
.21*
.42*
.41*
4.7*
2.3*
0.9*
0.77*
1.2*
ppm. We see the usual pattern of decreasing separation efficiency with in-
creasing ionic strength, but good separations can be obtained at ionic
strengths below 0.15 mole/SL of sodium nitrate.
The effects of different electrolytes on the separation of zinc with NLS
and A1(OH)3 are illustrated in Table 14. In all these runs the initial
Al(III), NLS, and Zn(II) concentrations were 100, 150, and 50 ppm, respective-
ly; the pH was adjusted to 8.3, 50 ppm of NLS was added after 15 min; the air
flow rate was 60 to 68 ml/min; and samples were taken for analysis after 30
min of treatment. At pH 8.3 equilibrium calculations indicate that ~93% of
64
-------
the phosphate is present as HPO^~ , 7% as
as H3P04- For arsenate at this pH, 96
0.1% as As043-, and 10"6% as H3AS04.
as
is present as HAsC>4
2~
3~, and
4% as H2As04"
The decreased removal of zinc with sulfate, as opposed to nitrate as the
added anion, is probably due to the increased negative charge on the anion, as
is indicated by the theoretical results below. The effects of phosphate and
arsenate are much too large to be due to merely a change in the charge of the
anion. Presumably these effects are due to strong adsorption of these anions
onto the floe, which presumably neutralized the charge of the floe.
TABLE 14. FLOC FLOTATION OF ZINC WITH A1(OH)3 AND NLS :
EFFECT OF ELECTROLYTE IDENTITY
Electrolyte
identity
NaN03
Na2S04
Na2HAs04
Na2HP04
0.05
0.13
0.6
47
46
Ionic strength (mole/5,)
0.10 0.15 0.20
Residual zinc (ppm)
0.13 0.12 0.22
2.2 3.6
48
>45
0.30
3.4
3.1
BATCH FLOTATION OF NICKEL(II), MANGANESE(II), CHROMIUM(III), AND COBALT(II)
Nickel
Essentially the same procedures employed for the flotation of zinc(II)
were used to carry out the flotation of nickel(II). One group of runs was
made using Al(III) or Fe(III) at 50 mg/£, NLS at 50 mg/£, and Ni(II) at 100
mg/£. Sample volumes were roughly 200 ml, and air flow rates of 60-67 ml/min
were used. In the pH range 5.5-8.0 the floes were rapidly and completely re-
moved, but Ni(II) removals were very poor; Ni(OH)2 has a relatively large
solubility product (1.6 x 10 ^ mole3/£3). Analyses were carried out by
atomic absorption spectrophotometry at 231.8 nm. The data for these runs are
as follows:
TABLE 15. EFFECT OF pH ON NICKELfII) FLOC FOAM FLOTATION5
Runs with A1(OH)3
Residual Ni (mg/£)
Runs with Fe(OH)3
Residual Ni (mg/£)
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
9.5
>50
>50
>50
>50
17.
10.
5.
3.
•
3
0
5
8
8
>50
>50
>50
>50
>50
38
21
1.
1.
25
25
Residual Ni measured after 15 min of treatment.
65
-------
The effect of ionic strength on flotation of Ni(II) with A1(OH)3 and NLS
is shown in Table 16.
TABLE 16. EFFECT OF IONIC STRENGTH ON NICKEL(II) FLOC
FOAM FLOTATION5
pH
Added NaN03 (moles/£)
.025
.050
.075
.250
8.5
9.0
9.5
6.6
5.6
1.4
1.3
1.5
1.4
4.2
2.5
0.6
19
16
>50
Residual Ni measured after 10 min of treatment.
Another set of runs made under somewhat different conditions. Initial
Ni(II) concentrations were 20 mg/£, and 100 mg/£ of Fe(III) or Al(III) and
50 mg/£ NLS were used, and the air flow rate was about 65 ml/min. The re-
sults are tabulated below.
TABLE 17. Niflll FLOTATION AT DECREASED Midi") CONCENTRATION5
(added NaN03, m/£) pH
0 6.5
7.0
7.5
8.0
.02 6.5
7.0
7.5
8.0
.05 6.5
7.0
7.5
8.0
.075 6.5
7.0
7.5
8.0
.10 6.5
7.0
7.5
8.0
Residual Ni
[A1(OH)3]
6.5
4.9
3.2
0.43
3.3
3.8
0.97
0.68
4.5
3.7
2.9
2.4
2.9
2.0
1.5
1.4
4.6
2.2
0.75
0.40
Residual Ni
[Fe(OH)3]
7.5
5.7
3.4
1.9
9.2
6.0
5.3
3.2
7.3
8.8
4.8
5.3
>10
8.0
9.3
6.6
-
-
-
-
Residual Ni measured after 10 mir of treatment.
In conclusion we have found that residual Ni(II) concentrations below 1
mg/£ are achievable at pH's of 8 to 9.5 using A1(OH)3 and NLS provided the
ionic strength is not too great (< .1
66
-------
Manganese
The relatively high solubility product of Mn(OH)2, 10~19, suggests that
the floe flotation of Mn(OH)2 with ferric or aluminum hydroxides should not
yield very high levels of removal. In preliminary runs this was found to be
the case. Samples were therefore made alkaline (pH^9) and aerated for various
lengths of time. The pH decreased during aeration as oxidation took place,
so 0.1 N NaOH was added as needed to bring the pH back to 9. After aeration,
flotation runs were made with 100 mg/X, of Fe(III), 50 mg/£ of NLS, an air
flow rate of 60 ral/min, and sample volumes of 200 ml. Duration of treatment
was 10 min.
At a flotation pH of 6.0, 1 h of aeration, and an initial Mn concen-
tration of 10 mg/£, removal of the floe was visibly quite incomplete. At
a flotation pH of 5.5, 1 hr of aeration, and other conditions as before,
removal of the floe was complete after 2 min, and residual Mn concentrations
were a fifth to a third of the initial values over a range of initial Mn
concentrations from 5 to 50 mg/Jl. Residual Mn was present as Mn(II), in-
dicating incomplete oxidation to Mn(OH)3 or Mn02. Samples'were therefore
aerated at pH 9 for 24 hrs and then treated by floe foam flotation at pH
5.5 as described above. Samples initially containing 30 mg/£ of Mn yielded
residuals of less than 2.0 mg/£; samples initially containing 15 mg/£
yielded residuals of less than 1.0 mg/£. Additional work on pH optimization
and the effects of ionic strength are being carried out. We conclude that
Mn(II) can be effectively removed by floe foam flotation with ferric hydro-
xide and NLS only if adequate aeration at an alkaline pH is provided.
Chromium
Chromium(III) itself forms a highly insoluble hydroxide floe (KSp =
7 x 10~31 m4/£4-)j so we carried out its flotation without the addition of
Fe(III) or Al(III). Initial Cr(III) and NLS concentrations of 50 mg/£ were
used, and the air flow rate was about 67 ml/min. Sample volumes were about
200 ml. Samples were taken for analysis after 10 min of treatment; chromium
was determined by atomic adsorption at 357.6 nm. Results were as follows:
TABLE 18, PRECIPITATE FLOTATION OF CHROMIC HYDROXIDE
Added NaN03 (m/£)
pH 0 .025 .050 .075
5.5
6.0
6.5
7.0
7.5
8.0
23
18
8.6
22
>40
>40
>40
11
5.5
16
-
-
>40
15
4
5
-
-
.2
.4
>40
10
12
13
-
-
67
-------
The data indicate that simple precipitate flotation of Cr(OH)3 is not
able to reduce Cr(III) down to the 1 mg/£ range.
We therefore investigated the coprecipitation of Cr(III) with ferric hy-
droxide; if this provides effective removal, one would anticipate that adsorb-
ing colloid flotation of Cr(III) with ferric hydroxide and NLS should also be
effective. Solutions were prepared containing the desired concentrations of
Cr(III) and Fe(III) as nitrates, the pH was adjusted to the desired level (5,
6, or 7) with NaOH and HN03, the solutions slowly stirred for about 5 min, the
precipitates allowed to settle, and samples taken for analysis. At pH 5.0
solutions initially containing Cr(mg/£) : Fe(mg/£) of 50:50, 50:100, 50;150,
and 25:100 all contained more than 1 mg/£ Cr(III) after precipitation; a
solution of initial composition 20:100 contained 0.5 mg/£ of Cr(III) after
precipitation. At pH's 6.0 and 7.0 the supernates of all the solutions con-
tained less than 0.5 mg/£ of Cr.
On the basis of these results we carried out adsorbing colloid flotation
runs on solutions containing 20 mg/£ of Cr(III) and 100 mg/£ of Fe(III) at pH's
in the range 5.5-7.25. The solutions were slowly stirred for 5-10 min after
precipitation; they were then made 50 mg/£ in NLS and transferred to the flo-
tation column. Air flow rates of about 65 m£/min were used. The solutions
became clear within 2 min after the initiation of flotation. The results
(after 10 min of foaming) are shown in Table 19.
TABLE 19.
EFFECT OF pH ON CHRONlTUMf II II FLOG FOAM FLOTATION
pH Residual Cr(mg/£)
5.5
6.0
6.15
6.5
7.0
7.25
<.25
<.25
<.25
<.25
.5
1.3
The effects of varying the initial Fe(III) concentration are shown in
Table 20.
TABLE 20. EFFECT OF IRON(III) CONCENTRATION ON CHROMIUM(III)
FLOC FOAM FLOTATION5
Initial Fe(III) (mg/£)
pH
Residual Cr (mg/£)
100
80
40
20
6.15
6.1
6.1
6.4
6.4
<.25
<.25
<.25
<.25
.4
lnitial Cr(III) concentration 20 mg/£, air flow rate 65 ml/min,
run time 10 min.
68
-------
The effects of increasing the ionic strength by the addition of NaNOj are
shown in Table 21'.
TABLE 21. EFFECT OF INERT SALT CONCENTRATION ON CHROMIUM(III)
___ FLOG FOAM FLOTATION3
Added NaN03 (moles/£) pH Residual Cr (mg/£)
0
.05
.10
6.15
6.25
6.50
6.30
.25
.3
. 3
2.2, 3.6
o
Initial Cr(III) concentration 20 mg/i, air flow rate 65 ml/min,
run time 10 min.
We noticed that if, in the course of adjusting the pH before flotation,
one overshot and the pH got up to 9-10 or more (after which it was adjusted
back to the desired level with 0.1N HN03), flotation did not remove all the
precipitate from the column, and the solution contained more than 1 mg/£ of
chromium after flotation. We explored this by carrying out four runs in
which the pH was held at 10 for various periods of time, then reduced to 6.2
for another period of time before flotation was carried out. In all of these
runs 20 mg/£ of Cr(III), 100 mg/£ of Fe(III), and 50 mg/£ of NLS were used;
the air flow rate was in the range 60-65 ml/min; and the solution volume was
200 ml.
In the first run the pH was held at 10 for 1 min; it was then adjusted
to 6.2 and flotation was immediately carried out. After 10 min of flotation
the residual Cr(III) level was 0.4 mg/SL, somewhat in excess of the values
of <.25 mg/£ which were routinely obtained in the absence of pH overshoot.
After 25 min of flotation the residual Cr(III) level was 0.2 mg/i.
The pH was held at 10 for 4 min for the second run; flotation was carried
out immediately after adjustment of the pH of 6.2. After 10 min of flotation
the residual Cr(III) concentration was 8.1 mg/£; after 25 min, 6.4.
The pH was held at 10 for 1 hr for the third run; the pH was then re-
duced to 6.2 and the solution permitted to stand for 2 hr. The pH tended
to increase, and was readjusted to 6.2 several times during this period.
After 10 min of flotation the residual Cr(III) concentration was 5.0 mg/£;
after 25 min, 4.3.
In the fourth run the pH was kept at 10 for 8 min; the pH was then re-
duced to 6.2 and the solution permitted to stand overnight. The next morning
the pH was again adjusted down to 6.2 and flotation carried out. After 10
min of flotation the residual Cr(III) level was 2.8 mg/£; after 25, 2.0.
We interpret these results as indicating the formation of a soluble, re-
latively non-labile hydroxide complex ion of Cr(III) at pH 10. Chromium(III)
complexes generally tend to be rather non-labile. We have not observed this
effect with other metals. Its occurrence with chromium dictates the careful
69
-------
avoidance of high pH's at all times if effective separations by floe foam flo-
tation or by precipitation techniques are to be achieved at reasonably rapid
rates.
We conclude that Cr(III) may readily be removed from aqueous solution by
adsorbing colloid flotation with ferric hydroxide and NLS, provided that the
pH is kept low enough to avoid the formation of soluble, relatively non-labile
chromium complexes with hydroxide.
Cobalt
The removal of Co(II) by floe foam flotation is similar to that of Ni(II).
The solubility product of Co(OH)2 is 2.5 x 10"15 m3/£3, slightly larger than
that of Ni(OH)2- 200 ml samples were treated; air flow rates of about 67
ml/min were used; and samples were taken for analysis after 15 min of treat-
ment. Data showing the pH dependence of the separations are shown in Table
22.
TABLE 22. FLOC FOAM FLOTATION OF COBALT(II) WITH A1(OH)3 OR Fe(OH)3
AND NLS
A1(OH)3 runsa
pH Residual Co(II) (mg/£)
pH
Fe(OH)3 runsb
Residual Co(II) (mg/£)
6.5
7.2
7.4
7.8
8.3
8.8
>8
7.6
5.4
3.5
1.2
1.4
6.0
6.5
7.0
7.5
8.0
8.5
9.0
>8
>8
7.2
3.0
2.4
1.9
6
alnitial Al(III) concn. = 100 mg/£. blnitial Fe(III) = 20 mg/£.
Initial Co++ concn. = 50 mg/£.
The effect of ionic strength is shown in Table 23.
TABLE 23. EFFECT OF IONIC STRENGTH ON FLOTATION OF
Co fill WITH AlfOH13 AND NLS
Added NaN03
pH
.025
.050
.075
7.2
7.8
8.3
8.8
9.3
7.6
3.5
1.2
1.4
6.
5.3
2.4
1.5
3.6
8.1
2.7
2.4
2.2
8.2
-
6.3
2.5
2.7
-
-
70
-------
We conclude, that residual Co(II) levels in the range 1-2 mg/£ can be
achieved at a pH of about 8.5 provided that the ionic strength is less than
.05 m/£.
COMPATIBILITY OF FLOC FOAM FLOTATION WITH PRECIPITATION PRETREATMENTS
We noted in our first report that foam flotation techniques appear to be
best adapted to the treatment of wastewaters, which are rather dilute in the
substance being removed. We also note that precipitation treatment at times
does not yield an effluent of adequate quality. These facts motivated our
investigation of the compatibility of several precipitation pretreatment
procedures for the removal of copper(II) by adsorbing colloid flotation
with Fe(OH)3 as the floe and NLS as the collector. This separation was used
because it had previously been shown to be a very effective one, (1) and
because the atomic absorption spectrophotometric analysis for copper at 324.8
nm is quite sensitive.
1. Use of NaoCO-r (soda ash) as a precipitating agent.
Z. O
Batch runs were made as follows. A solution of 500 mg/£ of Cu(II) as
Cu(N03)2 was prepared, and to this was added sufficient 0.5 M Na2C03 solution
to precipitate the copper and achieve the desired pH. (Precipitation was
carried out at pH's of 7.0, 8.0, and 9.0.) After the CuC03 precipitate set-
tled out, the supernatant liquid was decanted. Foam flotation was carried
out on this supernate, using 100 mg/£ of Fe(III), 100 mg/£ of NLS (50 mg/£
initially, 25 mg/Jc initially, 25 mg/£ after 6 rain, and 25 mg/£ after 11 min) .
Flotation was carried out at pH's of 5.5, 6.0, 6.5 and 7.0.
After the precipitation step the Cu(II) concentration was in the range
5-15 mg/£, depending on the settling pH. (Typically ~15 ml of Na2C03 solution
was added to 300 ml of 500 mg/£ Cu(II) solution.) The carbonate ion in the
supernatant solution severely hindered subsequent Cu(II) removal by foam
flotation. Since carbonates form C02 at the lower pH's and can then be
sparged from solution, the flotation runs made at acidic pH gave much better
results, especially if the solution was allowed to sparge in the column for
10 min before flotation was begun. Flotation runs made at pH 7.0 gave poor
Cu(II) removal; flotation runs at 6.5 caused a reduction in Cu(II) from 9.0
mg/£ to 0.11 mg/£ after 20 min. Similar runs in which the flotation pH was
6.0 gave a Cu(II) concentration of 0.20-0.40 mg/£ after 25 min, depending
on conditions. Similar runs in which the flotation pH was 5.5 gave somewhat
poorer results; residual Cu(II) was 1.0-2.0 mg/£ after 25 min of foaming. We
also found that, if the solution is placed in the column and sparged with air
at a very low pH (2.5-3.0) for 5-10 min, foam flotation at pH 6.5 readily
produced residual Cu(II) concentrations in the range 0.10-0.20 mg/£ after 10
min of foaming.
Data on the effects of settling pH and foaming pH are shown in Table 25.
Data on the effects of a preliminary air sparging at low pH are shown in
Table 26.
2. Use of Ca(OH)7 (lime) as a precipitating agent.
71
-------
TABLE 24. EFFECTS OF SETTLING AND FOAMING pH's ON Cu(II) REMOVAL AFTER
PRECIPITATION WITH Na2C03
Settling pH
Flotation pH Time, min
0
5
7.0 JP
20
25
0
5
, - 10
6"° 15
20
25
0
5
6.5* j|?
20
-> r
0
5
6.0 j°
20
25
0
5
IS
20
25
0
5
. r 10
15
20
25
0
5
- -* 10
0.3* , -
13
20
25
7.0
Cu(II) (mg/£)
7.9 (10.5)f
7.8
7.8
7.8
8.0
7.8
9.1 (12.1)
7.7
5.9
3 . 7
0.85
0.44
9.0 (12.1)
2.57
1.47
0.59
0.11
0.12
10.1 (13.5
4.5
1.06
0.54
0.49
0.47
11.5 (15.3)
1.22
0.85
0.68
0.63
0.77
10.8 (14.4)
3. 33
2.79
2,64
2.42
2.41
9.5 (12.6)
2.46
2.11
1.94
1.75
2.00
8.0
7.0 (9.3)
6.9
6.7
6.3
6.2
5.7
- (-)
6.3
5.8
4.3
2.29
1.20
4.6 (6.1)
4.6
2.58
0.71
0.31
0.30
5.0 (6.7)
3.6
1.88
0.41
0.27
0.24
5.4 (7.1)
1.08
0.35
0.37
0.35
0.37
4.6 (6.1)
1.65
1.27
1.17
1.09
1.14
7.6 (10.2)
2.05
2.09
1.86
1.79
1.62
9.0
4.3
4.2
4.3
4.2
4.3
4.3
6.3
4.7
4.2
2.4
1.67
0.89
3.9
-
1.46
0.53
0.24
3.15
3.3
1.03
0.36
0.34
0.31
3.14
0.79
0.46
0.34
0.28
-
3.29
1.40
1.04
0.96
1.00
0.99
3.4
1.03
0.99
0.95
0.95
0.94
(5.7)
(8.5)
(5.2)
(4.2)
(4.2)
(4.4)
(4.6)
'Sample sparged with air for 10 min before foaming.
"Parenthetic values are Cu(II) concentrations in the decantate before addi-
tion of other reagent solutions.
72
-------
TABLE 25. EFFECTS OF AIR SPARGING AT LOW pH ON Cu(II) REMOVAL AFTER
PRECIPITATION WITH Na2C05*
Sparging time,
min Time, min Cu(II) concentration (mg/£)
2
4
6
10
0
5
10
15
20
25
0
5
10
15
20
25
0
5
10
15
20
25
0
5
10
15
20
25
0
5
10
15
20
25
10.1 (13.
5.3
1.71
0.44
0.35
0.26
8.7 (11.
0.96
0.26
0.26
-
0.21
8.9' (11.
1.30
0.21
0.15
0.15
0.15
10.6 (14.
0.52
0.15
0.15
0.15
0.14
10.1 (13.
0.27
0.18
0.10
0.14
0.12
5)
6)
8)
2)
4)
*Solutions were sparged with air at pH 2.5-5.0, pH of precipitation
and settling was 7.0 (Na?C05), initial Cu(II) concentration was 500
mg/£. 150 ml of supernate was made 100 mg/£ in Fe(III), placed in
column, acidified, and sparged for the desired period. The pH was
then raised to 6.5 (NaOH), 50 mg/£ of NLS added initially, 25 mg/£
after 6 min, 25 mg/£ after 11 min. Air flow rate were about 60 ml/min.
Batch runs were as follows. Solid Ca(OH)9 was added with vigorous stir-
ring to 200 ml of a solution containing 500 mg/>: of Cu(II) (as the nitrate)
until a pH of ~12 was reached. This required approximately 1.2 gm of Ca(OH)^
per liter of solution. The solution was allowed to settle for 10 min and the
supernatant decanted for further treatment by floe foam flotation with Fe(III)
and NLS as described previously. We found that the high Ca(II) concentration
-------
prevented foaming by forming a scum with the NLS, and that foam flotation was
not possible under these circumstances. The problem was solved by bubbling
C02 through the solution until the pH drops to about 10.0; this precipitates
CaCOj but does not leave excessive C03= or HC03~ in the solution to interfere
with the foam flotation step. The resulting solution is readily treated by
floe foam flotation [100 mg/£ of Fe(II); 50, 25, and 25 mg/£ of NLS initially,
after 6 min, after 11 min; air flow rate about 60 ml/min]. The results are
shown in Table 26. They indicate that Cu(II) levels, already quite low after
TABLE 26. EFFECT OF FOAMING pH ON Cu(II) REMOVAL AFTER PRECIPITATION
WITH Ca(OH)3
Flotation pH 5.5 6.0 6.5 7.0
time, min Cu(II) (mg/£)
0
5
10
15
20
25
0
0
0
0
0
0
4-
.23 (.31)'
.10
.11
.08
.06
.05
0.
0.
0.
0.
0.
-
23 (.31)
05
03
01
00
0
0
0
0
0
0
.19 (.25)
.13
.04
.06
.05
.04
0
0
0
0
0
0
.25
.20
.05
.08
.08
.05
(.35)
TParenthetic values are Cu(II) concentrations in the decantate before
addition of other reagent solutions.
precipitation with lime, are readily reduced to extremely low values by floe
foam flotation with Fe(OH)^ and NLS, provided that excessive Ca(II) is re-
moved.
3. Use of Al(III) or Fe(III) as coprecipitating agents.
In these runs solutions were prepared containing 500 mg/£ of Cu(II) as
the nitrate; aluminum nitrate or ferric nitrate and sodium hydroxide were used
to generate the coprecipitating A1(OH)3 or Fe(OH)3 floes; coprecipitation was
carried out at a pH of 7.0 in all runs. After coprecipitation and settling,
supernate was decanted, the pH was adjusted to the desired value with NaOH
and HN03; Fe(III) was added (100 mg/£); and NLS was added (50 mg/'I initially,
25 mg/£ after 6 min and 25 mg/£ after 10 min of flotation). The air flow
rate was approximately 60 ml/min.
In the first set of runs 100 mg/£ of Al(III) was used in the coprecipita-
tion step. After coprecipitation and settling the supernate contained ap-
proximately 1.00 mg/£ of Cu(II). Flotation of the supernate as described
above resulted in residual Cu(II) concentrations of 0.05 mg/£ (flotation pH
7.0), 0.03 mg/£ (6.5), 0.03 mg/£ (6.0) and 0.16 mg/£ (5.5) after 25 min.
The second set of runs was made using 50 mg/£ of Al(III) in the copre-
cipitation step; after the resulting floe had settled, the supernate contained
typically 1.5-2.4 mg/£ of Cu(II). Flotation of the supernate as described
above reduced residual Cu(II) levels to 0.05 mg/£ (flotation pH - 7.0), 0.03
mg/£ (6.0) and 0.15 mg/£ (5.5) after 25 min.
74
-------
Fe(III) at 50 tng/X, was used as the coprecipitating agent in the third set
of runs. After settling, the supernate contained 3.0-4.0 mg/2, of Cu(II) .
Flotation of this solution as described above produced residual Cu(II) levels
of 0.01 mg/£ (flotation pH = 7.0), 0.05 mg/X. (6.5), 0.03 mg/£ (6.0), and 0.23
mg/£ (5.5). It was found that the Cu(II) concentrations after coprecipitation
with 100 mg/£ of Fe(III) were approximately the same as those resulting when
50 mg/£ of Fe(III) was used. See Table 27.
TABLE 27. EFFECT OF FOAMING pH ON Cu(II) REMOVAL AFTER COPRECIPITATION WITH
A1(OH)3 or Fe(OH)3
Precipitating agent
and concentration
Flotation pH Time, min
100 mg/£ Al(III) 50 mg/£ Al(III) 50 mg/S, Fe(III)
0
5
•7 n 10
7 0
/-U 15
20
25
0
5
10
15
20
25
0
5
10
6.0
20
25
0
5
10
15
20
25
.95 (1.27)+
.21
.07
.07
.07
.05
.89 (1.19)
.20
.03
.03
.03
.03
1.05 (1.40)
.16
.03
.03
.03
.03
1.13 (1.51)
.20
.19
.17
.17
.16
2.40 (3.20)
.25
.13
.07
.07
.07
2.26 (3.01)
.13
.04
.03
.03
.03
1.80 (2.40)
.07
.03
.03
.03
.03
1.37 (1.83)
.17
.15
.15
.15
.15
4.5
.15
.01
.01
.01
.01
4.0
.30
.06
.06
.05
.05
2.70
.08
.05
.03
.03
.03
2.82
.34
.26
.24
.23
.23
(5.9)
(5.3)
(3.60)
(3.76)
Parenthetic values are Cu(II) concentrations in the supernate before addition
of other reagent solutions.
We conclude that coprecipitation with Al(OH)3 or Fe(OH)3 is quite com-
patible with adsorbing colloid flotation.
INTERFERENCES WITH FLOC FOAM FLOTATION RESULTING FROM FOREIGN IONS
One of the factors which can very markedly affect the efficiency of
adsorbing colloid flotation and surfactant recovery from flotation sludges is
75
-------
the extent to which other ions are adsorbed into the primary layer of the floe
The theory of this effect is discussed in references (42) and (51), and in
Appendix D. The very large changes in surface potential and in surface con-
centration of surfactant result (in the model analyzed) from varying
the salt concentration and identity of added salts. Earlier in this section
we demonstrated very marked differences in the ability of different anions to
interfere with the flotation of zinc(II) with A1(OH)3 and NLS; in order of
increasing interference we found N03~ < S0^= « Hr04= ~ HAs04=. [See also
(56).] In the present work we examine the effects of various added salts on
the batch flotation of ferric hydroxide floes with sodium lauryl sulfate at
pH 5.0. We choose this system and these conditions [100 rng/£ of Fe(III),
50 mg/£ of sodium lauryl sulfate] because (1) in the absence of added salts
flotation is very rapid; (2) the system is quite effective for a number of
separations; and (3) the flotation of the strongly colored Fe(OH)3 is readily
observed visually.
Batch runs of about 200 ml were made in an apparatus of the sort pre-
viously described (1,54,57). Removal rates were graded as rapid [removal
of Fe(OH)3 visually complete in 5 min], slow [visual evidence of FefOH)^ in
the foam, but removal not complete in 5 min], or none [no visual evidence
of Fe(OH)3 in the foam]. A variety of anions were chosen having different
charges and coordination affinities for Fe(III) (15). Glycerol, which co-
ordinates readily with Fe(III), was also investigated. The results are shown
in Table 29.
We find that phosphate, hexaphosphate, arsenate, EDTA, and oxalate are
extremely effective in suppressing the flotation of ferric hydroxide under
conditions at which, in the absence of these ions, it floats rapidly and
completely. In all the runs a possibly interfering substance was added to
the solution after the ferric hydroxide was precipitated to avoid possible
loss of the interfering ion by coprecipitation in the bulk of the solid
where it would presumably be ineffective. It was somewhat surprising to us
that cyanide and thiocyanate, both of which complex readily with Fe(III) in
solution, were nowhere nearly as effective in blocking flotation as the ions
mentioned above. Neither did we anticipate that ClO^' would be somewhat
more effective than NOj" and Cl in blocking flotation, since it is one of
the weakest-binding ligands known.
These results suggest a number of potential applications. Addition of
interfering anions might be used to make precipitate flotations more selective
and could also facilitate the recovery of surfactant from foam flotation
sludges. Our findings also introduce a complication into the use of foam
flotation techniques for the removal of metals from wastewater which may
contain interfering ions. The behavior of the floes in the presence of
interfering ions suggest that these ions may interfere severely with pre-
cipitation separations also.
SIMULTANEOUS FLOG FOAM FLOTATION OF Cu(II), Pb(II), AND Zn(II)
Most metal-containing industrial wastes contain several different metal
ions, so we carried out some work on the simultaneous removal of copper, lead
and zinc from solutions containing these metals. The concentration of each
-------
TABLE 28. EFFECT OF VARIOUS ADDED SALTS AND GLYCEROL ON THE FLOTATION OF FERRIC HYDROXIDE
Concentration (moles/£)a
Substance
NaClO.
NaNO
NaCl
KCN
KCNS
NaF
Glycerol
Na2S04
Na2HP04
Na2HAs04
Concentration
Na2EDTA
Concentration
Na2G2°4
Concentration
(NaP03)6
0.0025
R
R
N
N
0.0005
N
2.5 x 1CT5
R 4
4.2 xlO
N
R
S
S
N
N
0.005
N
6.25 x 10-5
S
8.3 x 10
N
0.01
R
R
R
S-R
S
N
S
S
N
N
1.25 x 10-4
1.67 x 10~3
N
0.03
S
R
R
S
S
S
N
3.13 x 10"4
5 x 10"3
N
0.05
S
R
R
S
S
S
6.25 x lO-4
N .,
8.3 x 10 '
N
N
R
R
N
S
S
1.25 x 10~3
N
~".JB "! T===7— -=r.i. !.- .-s ... ^-^
R = rapid removal; S = slow and/or incomplete removal after 5 min; N = no visible removal.
Operating conditions: pH = 5.0 ± 0.1; 100 ppm Fe(III); 50 ppm NLS; air flow 85 ml/min.
-------
of the metal ions was 20 mg/1, 100 mg/£ of NLS was used as the collector,
sample volumes were about 200 ml, air flow rates were about 67 ml/min, and
flotation was carried out for 20 min. Copper, lead and zinc analyses were
carried out by atomic adsorption. Ionic strength was varied by the addition
of sodium nitrate.
The results of this work are shown in Table 29, and indicate that
simultaneous removal of several metals by floe foam flotation is possible.
TABLE 29. RESULTS OF THE FLOW FOAM FLOTATION OF Cu(II), Pb(II),
AND Zn(II) WITH Fe(OH)3 AND NLS
Added
NaN03
(m/£)
.02
.02
.02
.05
.05
.05
.10
.10
.10
.075
.075
.075
PH
6.1
6.5
7.02
6.05
6.55
7.05
6.05
6.56
7.02
8.05
8.05
8.05
Fe(III) Residual
100
100
100
100
100
100
100
100
100
100
150
200
Cu(II)
(mg/JO
.30
.25
.20
.27
.26
.30
.35
.26
1.25
7.75
>8.0
>8.0
Pb(II)
(mg/JO
.20
.15
.20
.15
.15
.30
.4
.30
1.2
>10
>10
>10
Zn(II)
(nig/ JO
3.2
2.1
1.8
2.9
2.9
2.2
3.2
2.8
2.9
3.3
3.5
3.5
FLOC FOAM FLOTATION OF CYANIDES
Various treatment technologies (physical, chemical and biological) exist
for the removal of cyanide from wastewaters. These include: alkaline
chlorination; electrolytic decomposition; ozonation; complexation with metals;
ion exchange; reverse osmosis; dialysis, irradiation; permanganate oxidation;
peroxide oxidation; complexation with polysulfides; the Kastone process;
liquid-liquid extraction with primary and secondary amines; copper-catalyzed
activated carbon adsorption; and biological oxidation using trickling filters
and activated sludge. Many of these processes are well adapted to treating
waste streams of low volume and high concentration (e.g. wastewater from an
electroplating facility). Some exhibit technical difficulties, while others
at present lack full-scale demonstration. The alkaline chlorination process
is most commonly employed.
The iron cyanide complexes are so stable (see Table 50) that standard
alkaline chlorination does not affect them. Since they exhibit little dis-
sociation, they have acquired "non-toxic" labels. Table 31 lists the
solubilities of some complex cyanide salts. Iron complexes are capable of
releasing cyanide ion through photo dissociation in strong sunlight. Also,
bacterial decomposition of the complex in the receiving water to form CN~ is
possible as well as increased solubility under alkaline conditions. Conse-
quently the deliberate complexing of simple cyanides with iron salts as an
78
-------
TABLE 3D. STABILITY CONSTANTS FOR CYANIDE COMPLEXES
Complex Ks (25°C)
_ _
Co(CN)6 1 x 10
Cu(CN) "2 1 x 1025
-4 24
Fe(CN)4 1 x 10
Fe(CN)6~3 1 x 1031
Ni(CN)4~2 1 x 1022
Zn(CN)4~2 8.3 x 1017
(from A. J. Bard, Chemical Equilibria, Harper and
Row, New York, 1966)
TABLE 31. SOLUBILITIES OF CYANIDE COMPOUNDS
Compound Solubility, g/£ Temp, °C
Ni(CN)2
Zn(CN)2
Fe4[Fe(CN)6]3
Zn2Fe(CN)6
Zn3[Fe(CN)6]2
Cu(CN)
5.92
5.8
2.5
2.6
2.2
0
x 10 2
x 10"3
x 10"4
x 10~3
x 10 5
.014
18
18
22
N/A
N/A
20
(from ASTM, 1975 and Linke, 1958, 1965)
economical wastewater treatment should be unacceptable. The insoluble iron
cyanide in a solid waste can best be treated by burial or landfill in an area
where acid conditions are common. The other metallo-cyanide complexes are
susceptible to chlorine oxidation but proceed at different rates.
Grieves, Bhattacharyya, and co-workers studied the removal of iron-com-
plexed cyanide by foaming with a cationic surfactant, ethylhexadecyldi-
methylammonium bromide (81,82). After treatment the free residual cyanide
averaged 7.5 mg/£; residual complexed cyanide, 2.9 mg/£. The reduction in
free cyanide ranged from approximately 80 to 90%. Other studies involving
metallo-cyanide complexes include: batch foam fractionation experiments con-
cerning the selectivity of several chloride vs cyanide complex ions (81); and
a similar determination of the selectivity coefficients for Ag(CN)2~ and
Au(CN)2~ vs I" (87).
Here we report on a study of the removal of metallo-cyanide complexes
using NLS. We have considered the precipitate and/or adsorbing flotation of
cobalt, copper, chromium, iron, nickel and zinc systems, individually and
in combination. Their removal is addressed from two aspects: (1) the re-
moval of cyanide ion itself; and (2) the removal of complexes which could be
79
-------
present or readily formed in the wastewater. Particular interest was paid to
iron-iron-cyanide systems because of their low toxicity and low cost.
Experimental
Batch foam separations were carried out in the batch columns described
earlier. Metals analyses were performed on acidified aqueous samples by
atomic absorption. Cyanide determinations were performed on basic (pH 13)
solutions with an Orion specific ion probe electrode, model 96-06. The stan-
dard curves were linear in the range studied, 0.05-10.0 mg/£. The average
correlation coefficient (for the first order fit of log concentration vs_ MV)
was 0.9970.
A semi-quantitative method was developed for the determination of residual
iron-cyanide complex in solution. A solution was prepared as described below,
but not foamed. The first metal added was Fe+^ (100 mg/£); the second, Fe
(150 mg/£). The additional pH adjustments to prepare for the cyanide and
metals analyses were also performed on the sample. This insured comparable
ionic strength of the solutions. The solution was then permitted to stand,
covered, to simulate the oxidation experienced in the column. Aliquots were
diluted to yield a series of standards representing 0-20% of the initial con-
centration. Absorbance vs percent of initial concentration were run on a
Beckman DB spectrophotometer at a wavelength of 732 my. The results were
linear in the range studied with a correlation coefficient of 0.9996 for a
first order fit.
Stock solutions of cyanide (50 mg/£) from KCN were prepared daily. Two
hundred ml were placed in a beaker and the appropriate amount of the "first
metal" was added. In all cases 1 ml of stock metal solution resulted in a 50
mg/£ concentration. It was presumed that the "first metal" forms the metal-
cyanide complex anion. The stock metal solutions were prepared from the cor-
responding nitrate salts except for the Fe+2 solution, for which ferrous am-
monium sulfate was used. All metal solutions were kept at acidic pH. The
metal cyanide solution was stirred for approximately 5 min, pH was monitored
throughout. The "second metal" was added and the solution stirred for ap-
proximately 10 min. Longer stirring periods were tested but yielded no in-
crease in CN removal. The pH was adjusted to the desired value by adding NaOH
or HNC>3 as necessary. Five ml of 1000 mg/£ NLS were added before the test
solution was poured into the column. Additional surfactant was added in a
series of 5 ml injections throughout the run as required. Air flow rate in
the column was approximately 60 ml/min.
Results and Discussions
A series of runs was performed to determine the optimum operating para-
meters for the foam removal of cyanide. Optimization required the assessment
not only of residual cyanide concentrations but also residual levels of
cyanide complex and iron. The parameters which were varied were: Fe(II)
concentration; Fe(III) concentration; NLS concentration.; pH; and duration of
foaming. The manner of NLS addition was also studied. Second order effects
were also briefly assessed. These included the effect of increased stirring
time after iron addition and use of dry weight iron salts additions (instead
80
-------
of acidified stock solutions). The latter was tried in an effort to maintain
ionic strength as low as possible. Both second order effects produced no more
than minor variations, well within the experimental precision of the results.
Table 32 shows the results of preliminary tests in which precipitate
flotation was employed to remove cyanide. (Earlier runs determined optimum
pH for removal.) The precipitation with Fe(III) effected no removal. Ferric
hydroxide was formed and an odor of HCN was noticed escaping from the column,
so the run was terminated. Precipitation with Fe(II) produced better results.
Foaming effected an average of 82.8% free CN reduction and a 91.1% iron
reduction.
TABLE 32. PRECIPITATE FLOTATION RUNS
Initial Cone.
Run
A
B
C
Fe(II),
50
50
0
Fe(III),
mg/£
0
0
50
PH
4.9
4.9
4 . 9+6 . 5
Foaming
Time
40
40
*•
Residual
CN, mg/£
8.6 '
8.6
17.1
Residual
Fe, mg/£
4.1
4.8
N/A
Aborted - see text.
TABLE 55. MOLARITY AND mg/£ CONVERSIONS
Species
at
mg/ £
yields Molarity (xlO ) Solution
CN
Fe
Cu
Cr
Co
Ni
Zn
Species
Fe(CN)6'3
Fe(CN)6~4
Fe[Fe(CN)6]
Fe4[Fe(CN)6]3
50 ppm Fe § 50 ppm CN
50
50
50
50
50
50
50
ratio mg/£
in moles
.358
.358
0.716
0.835
1.00
19.2
8.95
7.87
9.61
8.48
8.52
7.65
ratio Fe/CN
in moles
0.167
0.167
0.333
0.389
0.466
For 50 ppm CN total conversion to Fe4[Fe(CN)£]3 yields 9.45 mg/£ complex
salt or 1.1 x 10~-> moles complex salt.
Table 33 presents a summary of required molar ratios and mg/£ equivalents
for the iron cyanide complexes and precipitated compounds. Similar informa-
81
-------
tion for the heavy metals studied (and discussed later) is also included. It
was assumed that the Fe(II) formed the stable complex Fe(CN)£~4 anc[ excess
iron formed Fe4[Fe(CN)g]3 - the excess Fe(II) being oxidized to Fe(III) upon
stirring and/or foaming.
In an effort to increase the removal of cyanide, a series of runs was
made using adsorbing colloid flotation. The precipitate formed was still
assumed to be predominantly Fe4[Fe(CN)§]3. This was then adsorbed onto ferric
hydroxide floe and foamed with pulsed NLS additions. For 60 min runs, 50 ml
(25 mg/£) additions of NLS were added at time = 0, 10, 25 and 40 min. Table
35 shows the results of these runs which determined the standard operating
conditions. A fresh ascarite air filtering system adequately removed CC>2 so
TABLE 34. STANDARD OPERATING CONDITIONS: DETERMINING RUNS
Initial Cone.
Operating Conditions
+ 2
+ 3
Fe Fe
Run* mg/£ mg/£
foaming dur- Total NLS Free Complex,
pH ration, min. added, mg/£ CX, mg/£ (% initial cone.)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
100
100
100
100
100
100
100
100
100
100
100
100
100
100
200
200
200
200
200
200
150
150
150
150
150
150
150
150
3.
4.
5.
5.
5.
6.
4.
4.
4.
4.
4.
5.
5.
5.
85
7
3
5
7
0
2
2
2
2
2
5
5
5
45
45
60
60
120
aborted-poor
60
90
90
120
120
60
60
60
75
100
100
100
175
floe removal
125
125
125
150
150
100
100
100
3.
3.
2.
4.
1.
6.
5.
6.
5.
5.
3.
4.
4.
9
2
8
1
9
5
9
5
9
1
6
1
1
0.
0
0.
0
0.
0
0.
0
0
0.
0
0
0
8
5
5
2
2
that initial and final pH in the columns were essentially the same. Two ad-
ditional pH considerations were included in this study: (1) ferric hydroxide
floe floated best with NLS over a pH range of 5 to 7; and (2) most discharge
regulations require the pH of plant effluent to be between 6 and 9.
The first series of runs (1-6) employed 100 mg/£ Fe(II) to form the fer-
rocyanide complex and precipitate and 200 mg/£ excess Fe(III) to form the
adsorbing floe. Overall best results were obtained in the pH 5.3-5.5 range
considering duration of foaming and amount of NLS required.
It should be noted that according to Grieves and Bhattacharyya (88), as
suggested by Legros (89,90), complete conversion of "free" cyanide to com-
plexed cyanide is impossible. A "reasonable" percentage of non-complexed
cyanide was reported to be 20%. This results from the hydrolysis of the fer-
rocyanide complex ion to form a ferro aquo penta cyanide complex,
[Fe(CN)5H20]--5, and free cyanide, CN~.
82
-------
The next series of runs (7-14) optimized the iron concentration and
foaming time. On the basis of the results of series 1, the pH employed was
5.3-5.5. At pH 4.2 the average CN removal was 88.0%. The average iron
removal was 87.7%. There was evidence of trace residual iron complex at this
pH. At pH 5.5 the CN removal was 92.2% (3.9 ppm average residual) and the
average iron removal was 97.8% (5.4 ppm average residual). There was no
evidence of residual iron complex when the samples were analyzed as described
above. Extended foaming did not decrease the residual free cyanide concen-
tration but did reduce the iron concentration from 5.0 mg/£ at 60 min to 3.0
mg/£ after 90 min. The foamate volume at the end of the 60 min runs was 5-6%
of the initial sample volume for one system; this quantity was 7-8% in
another column.
As mentioned previously, the iron cyanide complex is known to photode-
compose. A run was made which was shielded from the laboratory fluorescence
lights. No decrease in free cyanide concentration was noted, but there was a
slight increase in residual complexed cyanide, 0.76%. (Operating conditions
were as in runs 12-14.)
The operating conditions from runs 12-14 were selected as the standard
operating conditions. Another series of runs using these conditions investi-
gated the effect of increasing ionic strength on residual concentrations.
The results are seen in Table 35. As anticipated, the residual concentrations
TABLE 55. IONIC STRENGTH RUNS*
Run
Ave. of
Stnd . Runs
18
19
20
21
22
23
NaN03
Molarity
0
0.01
0.01
0.1
0.1
0.25
0.25
CN, mg/£
3.9
8.8
8.8
7.3
6.2
3.4
3.2
Fe, mg/x
5.4
11.3
19.5
44.3
54.8
73.6
145.6
Complex
(% initial cone.)
0
0.2
0
N/A
32
100
100
*Standard operating conditions; pH 5.3-5.4; 100 mg/£ Fe+ ; 150 mg/£
Fe+3; duration of foaming 60 min; total NLS added, 100 mg/&.
of iron and complex increased with increasing salt concentration. The con-
centration of free cyanide at the end of the run was found to decrease with
increasing salt concentration.
Next, standard operating conditions were used in a study of cyanide re-
moval in the presence of various heavy metals, singly and in combination.
The metals selected for study were cobalt, copper, chromium, nickel and zinc,
Cobalt was included since it also forms a very stable cyanide complex (see
Table 30). The other metals are commonly used in electroplating facilities
83
-------
and are found in wastewater. Cyanide can be used as a major anion in the
plating baths for zinc, nickel or copper. These plated metals often form the
basis surface for subsequent chromium plating. The average results for
cyanide removal are given in Table 36.
In a Type A run, one first adds the metal of interest, M, to the free
cyanide solution to form the complex and then adds the Fe(III). This is
completely analogous to the standard runs. In a Type B run the complex is
made with Fe(II) and then the metal, M, is added. The operating pH was se-
lected as optimum from an earlier series of test runs. A Type C run re-
verses the order of the addition used in Type B runs. It is run at the
standard pH. The Type D runs combine the five metals in equal concentrations,
20 mg/£ each for a total concentration of 100 mg/£ metal. Runs of Type D are
then handled identically to Type A runs.
For the cobalt study, Type B was the least effective method, although
the residual free cyanide is low. It is assumed that the overall lower free
cyanide concentration (cf. standard runs) reflects the absence or reduction
of the hydrolysis reaction. Run Types A and C give comparable results in
cobalt and cyanide residual concentrations but not iron. This could be
attributed to the only partial oxidation of Fe(II) to Fe(III). For cobalt,
and the other metals, there is general agreement in the percent residual
metal between run Types A and D (see Table 37).
Copper, however, gave generally better results with Type B runs. Nickel
and chromium gave generally better results with Type A runs. Studies using
zinc gave the worst removals of the five metals. This was expected because
of zinc's amphoteric character and the comparatively small stability constant
of the zinc cyanide complex (See Table 30). Again the increased residual
iron concentrations obtained from the A and C Type runs most likely reflect
incomplete conversion of Fe(II) to Fe(III) during the stirring and foaming
periods.
The residual concentrations of iron and cyanide in the mixed metal runs
are lower than those obtained in the standard runs. The average 0.2 mg/£ Fe
residual is a 99.9% reduction. The average 1.5 mg/£ CN residual is a 97%
reduction.
Conclusions
The adsorbing colloid flotation of free cyanide by Fe(II)/Fe(III) results
in roughly 92.2% removal of free cyanide, and 100% removal of iron/cyanide
complex and 97.8 removal of iron after 60 min at a pH of 5.5. NLS was the
surfactant used. Increased ionic strength reduces the percent removal of the
complex and iron but decreases the presence of free cyanide after the initial
increase. Removal of cyanide in the presence of heavy metals, other than
iron, can be effected by the addition of Fe(III) to provide the adsorbing
precipitate of Fe(OH)j. The concentration of heavy metal is also reduced.
Optimum removal conditions must be determined for each metal or combination.
84
-------
TABLE 56. HEAVY METAL RUNS : AVERAGE RESIDUAL CONCENTRATIONS
Metal
Cobalt
Copper
Chromium
Nickel
Zinc
Run
Type
A
B
C
D
A
B
C
D
A
B
C
D
A
B
C
D
A
B
C
D
M, ppm
70.2
100
68.6
8.2
34.6
29.2
14.1
4.9
0.3
100
40.0
0.1
41.8
38.7
99.1
8.7
80.0
91.0
89.0
15.0
Fe, ppm
0.4
39.1
30.3
0.2
1.6
0.4
5.0
0.2
0.3
100
40.0
0.1
0.4
N/A
43.1
0.2
0.4
33.8
74.9
0.2
CN, ppm
0.08
0.13
0.07
1.5
7.6
0.06
0.69
1.5
19.2
- 7.4
45.4
1.5
0.23
0.23
0.43
1.5
43.4
10.4
40.3
1.5
Run Type A: M + Fe+3; [M] = 100 ppm; [Fe+3] = 150 ppm; [CN] =
50 ppm; pH 5.4-5.5
B: Fe+2 + M; [Fe+2] = 100 ppm; [M] = 100 ppm; [CN] =
50 ppm; pH 4.7-4.8
C: M + Fe+2; [M] = 100 ppm; [Fe+2] = 100 ppm; [CN] =
50 ppm; pH 5.4-5.5
D: Ml + M2 + M3 + M4 + MS + Fe+3; [Ml] etc. = 20 ppm;
[Fe+3] = 150 ppm; [CN] = 50 ppm; pH 5.4-5.5
TABLE 37. PERCENT RESIDUAL METALS. RUN TYPES A & D
Run Type
A
D
Co
70.2
41.0
Cu
34.6
24.5
Cr
0.3
0.5
Ni
41.8
43.5
Zn
80.0
75.0
-------
REFERENCES
1. Wilson, D. J., Foam Flotation Treatment of Heavy Metals and Fluoride-
Bearing Wastewaters, Environmental Protection Agency EPA-600/2-77-072,
U. S. Government Printing Office, Washington, D.C. (1977).
2. Hallowell, J. B. , Shea, J. F., Smithson, G. R., Jr., Tripler, A. B., and
Gonser, B. W., Water Pollution Control in the Primary Nonferrous-Metals
Industry, Vols. 1 and 2, Environmental Protection Agency EPA-R2-73-247a,
b, U. S. Government Printing Office, Washington, D.C. (1973).
3. Olsen, A. E., and Hauf, E. N., In-Process Pollution Abatement: Upgrading
Metal Finishing Facilities to Reduce Pollution, EPA Technology Transfer
Seminar Publ. No. 1 (1973).
4. Lancy, L. E., and Rice, R. L., Waste Treatment: Upgrading Metal-
Finishing Facilities to Reduce Pollution, EPA Technology Transfer Seminar
Publ. No. 2 (1973).
5. Clarke, A. N., and Wilson, D. J., Separation by Flotation, Separ. Purif.
Methods, 7 55 (1978).
6. Lemlich, R., ed. Adsorptive Bubble Separation Techniques, Academic Press,
New York (1972).
7. Somasundaran, P., Foam Separation Methods, Separ. Purif. Methods, 1 117
(1972).
8. Somasundaran, P., Separation Using Foaming Techniques, Separat. Sci., 10,
93 (1975).
9. Grieves, R. B., Foam Separations: A Review, Chem. Eng. J., 9, 93 (1975).
10. Ahmed, S. I., A Survey of the Laws of Foam Formation and Its Applications
in Separation Processes, Separat. Sci., 10, 649 (1975).
11. Bahr, A., and Hense, J., Flotation, Fortschr. Verfahrenstech., Abt. F, 14
503 (1976) .
12. Panou, 0., Flotation, Rapp. Tech., Cent. Beige Etude Corros., 129 (RT.
234), 153 (1976).
13. Richmond, R., Some Fundamental Concepts in Flotation, Chem. Ind. (London),
19, 792 (1977).
86
-------
14. Balcerzak, W., Foaming Processes-Foam Utilization and Elimination, Gaz,
Woda Tech. Sanit., 51 (5), 131 (1977).
15. Ecklund L., and Anderson, C., Wastewater Purification by Flotation, Ind.-
Anz., 99 (49), 875 (1977).
16. Hyde, R. A., Miller, D. G., Packham, R. F., and Richards, W. N., Water
Clarification by Flotation, J. Am. Water Works Assoc. , 69 (7), 369 (1977).
17. Marks, R. H., and Thurston, R. J., Application of Flotation to Industrial
Effluent Treatment, Environ. Pollut. Manage., 7 (4), 94 (1977).
18. Somasundaran, P., and Grieves, R. B., eds., "Advances in Interfacial Phe-
nomena of Particulate/Solution/Gas Systems; Applications to Flotation
Research," AIChE Symp. Ser. No. 150, Vol. 71 (1975).
19. Goldberg, M., and Rubin, E., Foam Fractionation in a Stripping Column,
Separat. Sci., 7, 51 (1972).
20. Wang, L. K.-P., Granstrom, M. L. and Kown, B. T., Continuous Bubble Frac-
tionation: Part I, Theoretical Considerations, Environmental Letters, 3
251 (1972) .
21. Ibid., Continuous Bubble Fractionation: Part II, Effects of Bubble Size
and Gas Rate, Environmental Letters, 4, 233 (1973).
22. Ibid., Continuous Bubble Fractionation: Part III, Experimental Evalu-
ation of Flow Parameters, Environmental Letters, 5, 71 (1973).
23. Goldberg, M., and Rubin, E., Foam Fractionation in a Stripping Column,
Separat. Sci., 7, 51 (1972).
24. Cannon, K. D., and Lemlich, R., A Theoretical Study of Bubble Fraction-
ation, AIChE Symposium Series 68 (124), Water-1971, 180 (1972).
25. Lee, S. J., Bubble Tower Performance in Concentrating Dilute Aqueous Sur-
factant Ph.D. Dissertation. Univ. of Tenn., 1969, Diss. Abstr. Int. B.,
31 (4), 1927 (1970).
26. Sastry, K. V. S. and Fuerstenau, D. W., Theoretical Analysis of a Counter-
current Flotation Column Trans. AIME, 247, 46 (1970).
27. Wilson, J. W., Wilson, D. J., and Clarke, J. H., Electrical Aspects of
Adsorbing Colloid Flotation. IV. Stripping Column Operation, Separat.
Sci., 11, 223 (1976).
28. Grieves, R. B., Ogbu, I. U., Bhattacharyya, D., and Conger, W. L., Foam
Fractionation Rates. Separat. Sci., 5, 583 (1970).
29. Petrakova, A. G., Golovanchikov, A. B., and Mamakov, A., Mechanism of
Interaction of Gas Bubbles with Mineral Particles During Flotation and
Electroflotation, Chem. Abstr., 85, 16244a (1976).
87
-------
30. Abramov, A. A., and Avdokhin, V. M., Physicochemical Modeling of Flotation
Systems, Chem. Abstr., 87, 204644a (1977).
31. Reay, D., and Ratcliff, G. A., Removal of Fine Particles from Water by
Dispersed Air Flotation: Effects of Bubble Size and Particle Size on Col-
lection Efficiency, Can. J. Chem. Eng., 51, 178 (1973).
32. Scheludko, A., Toshev, B. V., and Bojadjiev, D. T., Attachment of Particles
to a Liquid Surface (Capillary Theory of Flotation), J. Chem. Soc.,Faraday
Trans. I, 72 (12), 2815 (1976).
33. Deryagin, B. V., Rulev, N. N. , and Dukhin, S. S., Effect of Particle Size
on Heterocoagulation in an Elemental Flotation Act, Kolloidn. Zh., 39 (4),
680 (1977).
34. Bleier, A., Goddard, E. D., and Kulkarni, R. D., Adsorption and Critical
Flotation Conditions, J. Colloid Interface Sci., 59 (3), 490 (1977).
35. Lai, R. W. M. and Fuerstenau, D. W., Model for the Surface Charge of Ox-
ides and Flotation Response, Soc. of Min. Eng.-AIME, Trans.,260,104(1976).
36. Wilson, J. W., and Wilson, D. J., Electrical Aspects of Adsorbing Colloid
Flotation, Separat. Sci., 9, 381 (1974).
37. Huang, S.-D., and Wilson, D. J., Electrical Aspects of Adsorbing Colloid
Flotation. II. Theory of Rate Effects, Separat. Sci., 10, 405 (1975).
38. Wilson, J. W., and Wilson, D. J., Electrical Aspects of Adsorbing Colloid
Flotation. III. Excluded Volume Effects, Separat. Sci., 11, 89 (1976).
39. Wilson, D. J., Electrical Aspects of Adsorbing Colloid Flotation, V. Non-
ideal Floes and Salts, Separat. Sci., 11, 391 (1976).
40. Wilson, D. J., Electrical Aspects of Adsorbing Colloid Flotation. VI.Elec-
trical Repulsion between Floe Particles, Separat. Sci.,12, 231 (1977).
41. Wilson, D. J., Electrical Aspects of Adsorbing Flotation. VII. Cooperative
Phenomena, Separat. Sci., 12, 447 (1977).
42. Clarke, A. N., Wilson, D. J., and C.arke, J. H., Electrical Aspects of Ad-
sorbing Colloid Flotation. VIII. Specific Adsorption of Ions by Floes,
Separat. Sci. Technol.} 13, 573 (1978).
43. Fuerstenau, D. W., Healy, T. W., and Somasundaran, P., The Role of the
Hydrocarbon Chain of Alkyl Collectors in Flotation, Trans. AIME, 229, 321
(1964).
44. Gaudin, A. M., and Fuerstenau, D. W., Quartz Flotation with Cationic Col-
lectors, Trans. AIME, 202, 958 (1955).
45. Somasundaran,?., and Fuerstenau, D.W., Mechanism of Alkyl Sulfonate Ad-
sorption at the Alumina-Water Interface, J. Phys. Chem.,70, 90 (1966).
88
-------
46. Wakamatsu, T., and Fuerstenau, E. W., Effect of Hydrocarbon Chain Length
on the Adsorption of Sulfonates at the Solid-Water Interface, Adv. Chem.
Ser., 79, 161 (1968).
47. Somasundaran, P., and Fuerstenau, D. W., On Incipient Flotation Conditions
Trans. AIME, 241, 102 (1968).
48. Somasundaran, P., The Relationship between Adsorption at Different Inter-
faces and Flotation Behavior, Trans. AIME, 241, 105 (1968).
49. Somasundaran, P., Healy, T. W., and Fuerstenau, D. W., Surfactant Adsorp-
tion at the Solid-Liquid Interface-Dependence of Mechanism on Chain length
J. Phys. Chem., 68, 3562 (1964).
50. Fuerstenau, D. W., and Healy, T. W., Principles of Mineral Flotation, in
ref. 6, p. 92.
51. Currin, B. L., Potter, F. J., Wilson, D. J., and French, R.H., Surfactant
Recovery in Adsorbing Colloid Flotation, Separat. Sci. Technol.,13, 285
(1978).
52. Wilson, D. J., A Non-Coulombic Model for Adsorbing Col.loid Flotation, Sep-
arat. Sci. Technol., 13, 107 (1978).
53. Ferguson, B., Hinkle, C., and Wilson, D. J., Foam Flotation of Lead and
Cadmium in Industrial Wastes, Separat. Sci., 9, 125 (1974).
54. Clarke, A. N., and Wilson, D. J., The Adsorbing Colloid Flotation of Fluo-
ride Ion by Aluminum Hydroxide in Aqueous Media, Separat. Sci, 10, 417 (1975).
55. Huang, S.-D., and Wilson, D. J., Foam Separation of Mercury (II) and Cad-
mium (II) from Aqueous Systems, Separat. Sci., 11, 215, (1976).
56. Robertson, R. P., Wilson, D. J., and Wilson, C. S., The Adsorbing Colloid
Flotation of Lead (II) and Zinc (II) by Hydroxides, Separat. Sci., 11,
569, (1976).
57. Chatman, T. E., Huang. S.-D., and Wilson, D. J., Constant Surface Charge
Model in Floe Foam Flotation. The Flotation of Copper (II), Separat. Sci.,
12,461 (1977).
58. Zeitlin, H., and Kim, Y. S., Seperation of Trace-Metal Ions from Seawater
by Adsorptive Colloid Flotation, J. Chem. Soc.D.(Chem. CommJ^lS,672(1971)
59. Kim, Y. S., and Zeitlin, H. The Separation of Zinc and Copper from Sea-
water by Adsorbing Colloid Flotation, Separat. Sci.,7, 1 (1972).
60. Kim, Y. S., and Zeitlin, H. Thorium Hydroxide as a Collector for Moly-
bdenum from Seawater, Anal. Chim. Acta, 51, 516 (1970).
61. Kim, Y. W., and Zeitlin, H., A Rapid Adsorbing Colloid Method for the Sep-
aration of Molybdenum from Seawater, Separat. Sci., 6, 505 (1971).
89
-------
62. Kim, Y. S., and Zeitlin, H., Separation of Uranium from Seawater by Ad-
sorbing Colloid Flotation, Anal. Chem., 43, 1390 (1971).
63. Leung, G., Kim, Y. S., and Zeitlin, H., An Improved Separation and Deter-
mination of Uranium in Seawater, Anal. Chim Acta, 60, 229 (1972).
64. Voyce, D. and Zeitlin, H., The Separation of Mercury from Seawater by Ad-
sorption Colloid Flotation and Analysis by Flameless Atomic Absorption,
Anal. Chim. Acta, 69, 27 (1974).
65. Chaine, F. E., and Zeitlin, H., The Separation of Phosphate and Arsenate
from Seawater by adsortior^ Colloid Flotation, Separat. Sci., 9, 1 (1974).
66. Rothstein, N., and Zeitlin, H., The Separation of Silver from Seawater by
Adsorption Colloid Flotation, Anal. Lett., 9, (5), 461 (1976).
67. Hagadone, M., and Zeitlin, H. , The Separation of Vanadium from Seawater by
Adsorption Colloid Flotation, Anal. Chim. Acta, 86, 289 (1976).
68. Matsuzaki, C., and Zeitlin, H., The Separation of Collectors Used as Co-
recipitants of Trace Elements in Seawater by Adsorping Colloid Flotation,
Separat. Sci., 8, 185 (1973).
69. Zhorov, V. A., Barannik, V. P., Lyashenko, S. V., Kirchanova, A. I., and
Kobylyanskaya, A. G., Use of Adsorption Colloid Flotation for Separating
Trace Elements from Seawater, Chem. Abstr., 86, 143432d (1976).
70. Grigorev, Yu. 0., Tyurin, N. G., Pushkarev, V. V., Pustovalov, N. N., and
Perederii, 0. G., Sorption of Arsenic Ions by Ferrous Sulfide, Zh. Fiz.
Khim., 50, 1004 (1976).
71. Bhattacharyya, D., Carlton, J. A., and Grieves, R. B., Precipitate Flota-
tion of Chromium, AIChEJ, 17, 419 (1971).
72. Grieves, R. B., Schwartz, R. M., and Bhattacharyya, D., Precipitate Co-
flotation of Calcium Sulfite and Calcium Carbonate: Application to Solids
Removal from S02 Wet-Scrubbing Slurries Separat. Sci., 10, 777 (1975).
73. Goldberg, M., and Rubin, E., Mechanical Foam Breaking, Ind. Eng. Chem.,
Process Des. Develop., 6, 195 (1967).
'*' Rubin, E., and Goly, M., Foam Breaking with a High Speed Rotating Disk,
Ind. Eng. Chem., Process Des. Develop., 9, 341 (1970).
75. Fuerstenau, M. C., ed., Flotation, A. M. Gaudin Memorial Volume, AIME, New
York, 1976.
76. Fowler, R., and Guggenheim, E. A., Statistical Thermodynamics, Cambridge
University Press, 1952, pp. 429-443.
77. Macdonald, J. R., and Brachman, M. K., Exact Solution of the Debye-Hiickel
Equations for a Polarized Electrode, J. Chem. Phys., %2, 1314 (1954).
90
-------
78. Patterson, J. W., Technical Inequities in Effluent Limitation Guidelines,
JWPCF, 49 (7), 1586 (1977).
79. Patterson, J. W., Wastewater Treatment Technology, Ann Arbor Science Pub-
lishers Inc., Mich., 1975, ch. 9.
80. Bard, A. J. , Chemical Equilibria, Harper and Row, New York, 1966.
81. Grieves, R. B., Ghosal, J. K., and Bhattacharyya, D., Foam Separation of
Complexed Cyanide: Studies of Rate and Pulsed Addition of Surfactants, J.
Appl. Chem. (London), 19, 115 (1969).
82. Grieves, R. B., Ghosal, J. K., and Bhattacharyya, D., Ion Flotation of Di-
chromate and of Complexed Cyanide: Surfactants for Qualitative Analysis,
J. Amer. Oil Chem. Soc., 45, 591 (1968).
83. Grieves, R. B., and Bhattacharyya, D., "Ion, Colloid and Precipitate Flo-
tation of Inorganic Anions," in ref. 6, p. 187.
84. Walkowiak, W., and Grieves, R. R., "Foam Fractionation of Cyanide Complex
Anions of Zn(II), Cd(II), Hg(II), and Au(III), " J. Inorg. Nucl. Chem.,
38, 1351 (1976).
85. Hill, T. L., Introduction to Statistical Thermodynamics, Addison-Wesley,
Reading, Mass., 1960, pp. 130-132.
86. Adamson, A. W., A Textbook of Physical Chemistry, Academic Press, New York
1973, p. 25.
87. Walkowiak, W., and Rudnik, Z./'Selectivity Coefficients for Ag(CN)2"and
Au(CN)2~from Continuous Foam Fractionation with a Quaternary Ammonium Sur-
factant," Separat. Sci. Technol., 13, 127 (1978).
88. Grieves, R. B., and Bhattacharyya, D., "Foam Separation of Cyanide Com-
plexed by Iron," Separat. Sci., 3, 185 (1968).
89. Emschwillee, G., and Legros, J., "Cinetique et Equilibre de Dissociation
des Ferrocyanures en Solution Aqueuse," Compt. Rend., 241, 44 (1955).
90. Legros, J., "Etudes Cinetiques sur 1'Hydrolyse de quelques Complexes
Cyanes du Fer Bivalent," J. Chim. Phys., 61, 909 (1964).
91
-------
APPENDIX A
DATA ON LEAD REMOVAL FROM WASTES AND SIMULATED
WASTES WITH THE 10-CM AND 30-CM COLUMNS
92
-------
TABLE A-l. DATA ON LEAD REMOVAL FROM WASTES AND SIMULATED WASTES WITH THE 10-CM AND 30-CM COLUMNS
Run Fe(III) NLS Air Flow Influent Flow
No. pll mg/£ mg/£ £/mina Rate (gal/hr)b
1 5.9 150 40 4 10
15
12
12
12
12
12
11
12.5
12.5
12.5
2 6.0 200 40 4 8
11
11
10.5
5 10
10
10
10
Hydraulic
Loading
gal/min-ft^
1.9
2.9
2.3
2.3
2.3
2.3
2.3
2.1
2.4
2.4
2.4
1.5
2.1
2.1
2.0
1.9
1.9
1 .9
1.9
Effl.
Pb(II)
mg/J>
.05
.10
.10
.05
.05
<.05
.12
.05
.08
.31
.31
.20
.37
.24
.44
.57
.37
.47
.47
Comments
Foam very
stable, but
wet. Foamate
volume -6% of
influent vol-
ume .
Foam over-
turning
Foam over-
turning
Overturning
occurred
throughout
most of the
run. One of
the screen
baffles was
tilted
slightly
(continued)
-------
Run Fe(III) NLS Air Flow Influent Flow
No. pll mg/£ mg/£ 2./mina Rate (gal/hr)b
3 6.1 150 35 3 8
2 9
12
12
15
15
13
13
13
13
4 6.0 150 35 1 7.5
1 9
1.5 12.5
1.5 12.5
5 6.4 150 35 2 5.5
5.5
8
10
10
12
12
6 7.1 150 35 2 8.5
10.0
12.5
7 5.8 150 35 2 5.5
10.5
12.5
("continued) 14.5
Hydraulic
Loading
gal/min-ft^
1.5
1.7
2.3
2.3
2.9
2.9
2.5
2.5
2.5
2.5
1.4
1.7
2.4
2.4
1.0
1.0
1.5
1.9
1.9
2.3
2.3
1.6
1.9
2.4
1.0
2.0
2.4
2.8
Ef f 1 .
Pb(II)
mg/£.
.10
.14
.30
.64
.57
1.3
1.0
.44
.61
.47
.10
.13
1.5
.96
.05
.06
.10
.36
.68
.84
1.0
.84
2.3
3.2
.08
.17
.32
.49
Comments
Some overtur-
ning in foam
Foam recover-
ing
Severe over
turning
Recovering
Beginning to
turn
No overtur-
ning evident
Slight over-
turning
-------
TABLE A-l. (continued)
Ln
Run Fe(III)
No. pH mg/£
8 5.0 150
9 4.9 150
10 4.3 150
11 5.2 100
NLS Air Flow Influent Flow
mg/£ Vmina Rate (gal/hr)b
35 2 5.5
10.5
10.5
12.5
12.5
15
15
30 2 4
8
10.5
10.5
10.5
30 2 8
10.5
10.5
12.5
15
16.5
16
16
30 2 3
8.5
10.5
12.5
15
15
Hydraulic
Loading
gal/min-ft^
1.0
2.0
2.0
2.4
2.4
2.9
2.9
0.8
1.5
2.0
2.0
2.0
1.5
2.0
2.0
2.4
2.9
3.2
3.1
3.1
0.6
1.6
2.0
2.4
2.9
2.9
Effl.
Pb(II)
mg/H
.38
.28
.11
.10
.07
.05
.13
.11
.72
.87
1.4
1.7
1.8
2.8
2.9
2.3
2.2
3.0
1.6
2.2
.64
.51
.37
.41
.42
.40
Comments
Some overtur-
ning
Recovering
Overturning
beginning
Severe over-
turning
throughout
run
Little or no
overturning
present
(continued)
-------
ON
Run
No.
12
13
14
15
16
Fe(III) NLS Air Flow Influent Flow
pll mg/2, mg/£ £/mina Rate (gal/hr)b
5.4 200 30 2 8.5
10.5
10.5
12.5
12.5
12.5
5.8 100 35 2 8.5
10.5
10.5
12.5
12.5
16.5
5.7 200 35 2 8.5
10
12.5
12.5
12.5
5.7 150 35 2 5
7
7
8.5
8.5
10.5
10.5
5.8 150 35 3 5.5
7
8.5
8.5
8.5
Hydraulic
Loading
gal/min-ft^
1.6
2.0
2.0
2.4
2.4
2.4
1.6
2.0
2.0
2.4
2.4
3.2
1.6
1.9
2.4
2.4
2.4
.96
1.3
1.3
1.6
1.6
2.0
2.0
1.0
1.3
1.6
1.6
1.6
Effl.
Pb(II)
mg/£
.30
.55
.69
.95
.90
.77
.45
.15
.34
.45
.56
.81
.05
.10
.36
.39
.45
.68
.44
.44
.53
.53
.38
.35
1.1
1.8
2.1
3.2
2.9
Comments
Some overtur-
ning present
Very little
overturning
evident
Slight over-
turning
0.1 moles/£
NaN03 added.
Very little
overturning.
0.2 moles/£
NaN03 added.
(continued)
-------
TABLE A-l. rcontinuedl
Run Fe(III)
No. pH mg/£
17 5.8 150
NLS Air Flow
mg/& £/mina
35 2
Influent Flow
Rate (gal/hr)b
7
6.5
8.5
8.5
Hydraulic
Loading
gal/min-ft2
1.3
1.2
1.6
1.6
Effl.
Pb(II)
1.0
2.8
3.8
.3.9
Comments
0.3 moles/£
NaN03 added.
I 7
2 Ji/min = 0.24 m^/rain-
m
gal/hr =1.9 gal/min-ft2 = 112 m3/day-m2.
-------
TABLE A-2. DATA ON LEAD REMOVAL (SYNTHETIC WASTE) WITH THE 30-cm COLUMN
EXPERIMENTAL DATA FOR RUN 1
VD
oo
Sample
Number
1
2
3
4
5
6
7
8
pH
4.4
4.5
5.0
4.2
4.0
4.5
EXPERIMENTAL
1
2
3
4
5
6
7
4.7
5.0
4.5
4.4
4.2
Fe(III)
(mg/Jl)
70
70
70
70
70
70
DATA FOR
70
70
70
70
70
NLS
70
70
70
70
70
70
RUN 2
70
70
70
70
70
Air Flow
Rate
(SCFH)
50
50
50
50
50
50
50
50
50
50
50
Hydraulic
Loading Rate
(gal/min-ft2)
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
3.47
Influent
Flow Rate
(gal/min)
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
2.5
Waste
Ionic Pb(II)
Strength Sample
(moles) (mg/£)
16.8
0
0
0
0
0
0
11.5
16.9
0
0
0
0
0
15.1
Effluent
Pb(II)
Sample
(mg/£)
8.7
7.6
2.1
7.1
13.0
6.3
5.2
3.3
8.5
4.1
7.2
(continued)
-------
vo
VO
TABLE A-2. (continued)
EXPERIMENTAL DATA FOR RUN 3
Sample
Number
1
2
3
4
5
6
7
8
9
PH
5.9
6.3
6.2
5.6
5.0
5.5
5.8
EXPERIMENTAL
1
2
3
4
5
6
7.0
6.7
6.6
7.1
Fe(III)
(mg/A)
70
70
70
70
70
70
70
DATA FOR
100
100
100
100
NLS
(mg/£)
70
70
70
70
70
70
70
RUN 4
40
40
40
40
Air Flow
Rate
(SCFH)
60
60
60
60
60
60
60
60
60
60
60
Hydraulic
Loading Rate
(gal/min-ft2)
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
Influent
Flow Rate
(gal/min)
2
2
2
2
2
2
2
2
2
2
2
Waste
Ionic Pb(II)
Strength Sample
(moles) (mg/£)
19.8
0
0
0
0
0
0
0
16.9
15.8
0
0
0
0
14.1
Effluent
Pb(II)
Sample
(mg/4)
1.7
1.7
3.1
3.7
5.3
3.6
3.6
0.9
1.1
1.1
0.7
(continued)
-------
TABLE A-2 . (continued)
EXPERIMENTAL DATA FOR RUN 5
o
o
Sample
Number
1
2
3
4
5
6
7
8
pH
6.5
6.9
6.9
7.0
7.0
7.2
EXPERIMENTAL
1
2
3
4
5
6
7
6.5
6.5
6.7
6.9
5.6
EXPERIMENTAL
1
2
3
6.7
7.3
Fe(III)
(mg/ )
100
100
100
100
100
100
DATA FOR
150
150
150
150
150
DATA FOR
100
100
NLS
(mg/ )
40
40
40
40
40
40
RUN 6
40
40
40
40
40
RUN 7
50
50
Air Flow Hydraulic
Rate Loading Rate
(SCFH) (gal/min-ft2)
60
60
60
60
60
60
60
60
60
60
60
50-60
50-60
2
2
2
2
2
2
2
2
2
2
2
2
2
.77
.77
.77
.77
.77
.77
.77
.77
.77
.77
.77
.77
.77
Influent
Flow Rate
(gal/min)
2
2
2
2
2
2
2
2
2
2
2
2
2
Ionic
Strength
(moles)
0
0
0
0
0
0
0
0
0
0
0
0.01
0.01
Waste Effluent
Pb(II) Pb(II)
Sample Sample
(mg/ ) (mg/ )
18.5
0
0
0
0
0
0
15.4
19.7
1
19. 7(mixing
1
1
7
19.2
19.0
0
0
.6
.6
.6
.9
.8
.9
.5
chamber)
.5
.6
.0
.5
.4
(continued)
-------
TABLE A-2. (continued)
EXPERIMENTAL DATA FOR RUN 8
Sample
Number
1
2
3
4
5
PH
7.2
6.7
6.4
6.2
7.2
EXPERIMENTAL
1
2
3
4
5
6
7
6.5
6.8
6.5
6.3
6.3
6.5
Fe(III)
(mg/JO
100
100
100
100
100
DATA FOR
100
100
100
100
100
100
NLS
(mg/JO
50
50
50
50
50
RUN 9
50
50
50
50
50
50
Air Flow
Rate
(SCFH)
40-50
40-50
40-50
40-50
40-50
40
40
40
40
40
40
Hydraulic
Loading Rate
(gal/min-ft?~
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
Waste Effluent
Influent Ionic Pb(II) Pb(II)
Flow Rate Strength Sample Sample
(gal/min) (moles) (mg/JO
2
2
2
2
2
2
2
2
2
2
2
0.05 20*
0.05 20
0.10 20
0.10 20
0.10 20
0
0 20
0
0
0
0
0
(ing/ JO
2.6
2.4
6.1
12.8
11.0
1.14
1.14
1.14
1.85
3.00
3.14
(continued)
-------
TABLE A-2. (continued)
EXPERIMENTAL DATA FOR RUN 10
Sample
Number
1
2
3
4
5
6
7
8
9
PH
6.6
6.4
6.5
6.3
6.4
6.5
6.7
6.7
Fe(III)
(mg/S.)
100
100
100
100
100
100
100
100
NLS
(mg/S.)
50
50
50
50
50
50
50
50
Air Flow
Rate
(SCFH)
40
40
40
40
40
40
40
40
Hydraulic
Loading Rate
(gal/min-ft2)
2.77
2.77
2.77
2.77
2.77
2.77
2.77
2.77
Influent
Flow Rate
(gal/min)
2
2.5
2
2.5
2
2
2
2
Waste
Ionic Pb(II)
Strength Sample
(moles) (mg/&)
0.01 10
0.01
0.01
0.05
0.05
0.05
0.07
0.07
0.07
Effluent
Pb(II)
Sample
(mg/A)
0.86
1.57
2.29
1.86
2.43
2.57
3.14
3.29
-------
TABLE A-3. DATA ON ZINC REMOVAL (PLATING WASTE) WITH THE 30-cm COLUMN
EXPERIMENTAL DATA FOR RUNS 1. 2 and 5
o
u>
Description
of sample
Raw waste, run 1
After liming
Column effluent
it
it
Raw waste, run 2
After liming
Column effluent
1 1
Raw waste, run 3
After liming
Column effluent
ti
ii
ii
ir
it
Influent Flow
pH Rate, GPM
8.8
11.4
11.3
11.3
11.3
1.9
11.4
11.3
11.2
10.3
11.3
11.3
11.3
11.3
11.3
11.3
11.3
_
-
2.5
2.5
2.5
_
-
2.0
2.0
_
-
2.5
2.5
2.5
2.5
2.5
2.5
Air Flow Rate NLS
(SCFH) (mg/A)
_
-
40
40
40
_
-
40
40
_
-
40
40
40
40
40
40
_
-
0
0
0
_
-
33
33
_
-
0
0
0
0
0
0
Fe(III)
(mg/£)
—
-
80
80
80
„
-
67
67
_
-
100
100
100
100
100
100
Zn
(mg/£)
114
6
.7
.7
2
420
7.4
7.4
5.6
134
15
7
8
5
6
8
9
Cu
(mg/£)
1.2
0(<.01)
0
0
0
32
0
0
0
3.3
.10
0
0
0
0
0
0
(continued)
-------
o
-O
TABLE A-3. (continued)
BXPBRIJjMTALJJATA,FOR RUNS 4. 5. AND 6
Description
of sample
Raw waste, run 4
After liming
1!
Column effluent
1 1
n
n
Raw waste, run 5
After liming
Column effluent
"
"
"
"
Raw waste, run 6
After liming
Column effluent
(poor .floe)
separation)
PH
2.7
11.4
11.4
11.3
11.3
11.3
11.3
6.5
11.2
11.2
11.2
11.2
11.2
11.2
9.4
11.4
11.3
Influent Flow
Rate, GPM
_
-
-
2.0
2.0
2.0
2.0
_
-
2.5
1.5
2.0
2.0
2.0
_
-
1.5
Air Flow Rate
(SCFH)
_
-
-
40
40
40
40
_
-
40
decreasing
to
18
_
-
25
NLS
(mg/£)
_
-
-
33
33
0
0
_
-
0
0
0
0
0
-
-
33
Fe(III)
(mg/£)
_
-
-
67
67
80
80
-
-
80
110
100
100
100
-
-
67
Zn
(mg/fc)
320
14
18
11
6
10
11
214
19
13
12
12
14
14
.8
<.2
<. 2
Cu
(mg/£)
14
.3
.2
0
0
0
0
.4
0
0
0
0
-
-
0
0
0
(continued)
-------
TABLE A-3. (continued)
EXPERIMENTAL DATA FOR RUNS 7 AND 8
o
Ul
Description
of sample
Raw waste, run 7
After liming
Column effluent
it
11
Raw waste, run 8
After liming
Column effluent
11
"
"
11
11
11
"
PH
8.8
11.3
11.3
11.3
11.2
9.0
11.2
11.2
11.2
11.2
11.2
11.2
11.2
11.2
11.2
Influent Flow
Rate, GPM
_
-
2.0
2.0
2.0
_
-
2.0
2.0
2.0
2.0
2.0
2.0
2.0
2.0
Air Flow Rate
(SCFH)
_
-
45
45
45
_
-
54
54
54
54
54
54
54
54
NLS
(mg/A)
_
-
50
0
0
_
-
0
0
0
0
0
0
0
0
Fe(III)
Cmg/ «)
_
-
100
100
66
_
-
80
80
80
80
80
80
80
80
Zn
fmg/*)
15
4.3
.3
.3
.3
21
3.2
.2
.2
.2
.2
.2
.2
.2
.2
Cu
(ing/ A)
.8
0
0
0
0
.8
0
0
0
0
0
0
0
0
0
-------
APPENDIX B
ADSORPTION ISOTHERMS
In our previous report (l) and other papers, (27,36-40,42) we examined a
Gouy-Chapman type model for the binding of a floe particle to an air-water
interface. Fuerstenau, Somasundaran, and others have used an alternate model
to describe ore flotation (43-50), and it adequately describes quite a number
of observed effects. In their model the ionic heads of the surfactant ions
adsorb into the primary layer of the solid particle; the hydrocarbon chains
of the surfactant ion then present a hydrophobic surface to the solution.
This results in changes in the surface free energies; if these result in non-
zero bubble contact angles on the solid, bubbles attach and flotation occurs.
We here examine first, the adsorption isotherms of particles to air-water
interfaces resulting from finite contact angles (52), and second, the adsorp-
tion isotherms of surfactant on the surfaces of the solid particles (41).
1. Particle Adsorption Isotherms
Let the binding energy between a floe particle and the air-water inter-
face be due only to differences in the surface free energies yaw (air-water),
Yas (air-solid) and YSW (solid-water). We approximate our floe particles as
spheres, and assume that floe particle-bubble attachment is as shown in Figure
27. We assume that the air bubble is much larger than the floe particle.
Let 6 be the contact angle of the air-water, air-solid, and solid-water
interfaces (see Figure 27), and r be the floe particle radius. As the parti-
cle attaches, the loss in air-water interface is given by
L = irr2sin2e (B-l)
aw v '
106
-------
LIQUID
AIR —*- LIQUID
AIR
FIGURE B-l - MODEL OF FLOC-BUBBLE ATTACHMENT
-------
The loss in solid-water interface (and the gain in air-solid interface) is
given by
2
L , = 2-irr (1 - cose) (B-2)
The free energy of attachment is therefore
AG = -Y irr sin 6 + (Y - Y )2irr (1 - cos 6) (B-3)
aw as sw
The contact angle is given by
cos6 = (Y - Y )/Y (B-4)
as 'sw' aw *• '
at equilibrium, so that Eq. (B-3) can be written as
AG = -Yav/r2(l - cose)2 (B-S)
We next determine the adsorption isotherm of the floe particles on the
air-water surface by use of the grand partition function; we follow Hill (85)
Consider an area of the air-water interface capable of accommodating one
particle, 6r2//3~, and a column of liquid based on this area and capable of
containing a maximum of m floe particles. This constitutes a site, which
contains m cells; we take the sites to be independent of each other.
The grand partition function for a site is then given by
m
I = ini(l + qAexp[-3Vi]} (B-6)
where q = partition function for the internal degrees of freedom of a floe
particle, assumed independent of the cell in which the particle
is located
A = absolute activity of the floe particles, exp (y/kT), where y is
the chemical potential of a floe particle
e = i/kT
V-L = potential energy of a floe particle in cell i.
Vj_ = AG, given by Eq. (B-5) ;
Vj_ = 0, i > 1.
The average number of floe particles per site is then given by
m
s" = y
Aqexp(-gAG) (m-l)Aq
1 + Aqexp(-gAG) 1 + Aq
In the absence of the interface,
mAq
-5- = mo - -;—^-rr
(B-8)
108
-------
which we obtain from Eq. (B-8) by setting AG = 0. Here a is the average
number of floe particles per cell in the bulk solution. Eq. (B-9) yields
Aq = a/(l - a) (B-10)
and using this in Eq. (B-8) yields
_ 1
s = — + (m-l)cj (B-ll)
a
The surface excess of floe particles in the site is given by
s" = T - ma (B-12)
excess *• '
which is given by
_ a(l-a)[1 - exp(BAG)]
Sexcess = a + (l-a)exp(BAG)" (B-13)
Generally, as we shall see, exp(3AG) « 1, so Eq. (B-13) can be well approxi-
mated by
(B-14)
excess 1 + exp(BAG)/o
The basal area of a site is 2/3i-2 (on assuming hexagonal close-packing in the
planer, and the volume of a cell is given by
V = 4/3r3 (B-15)
The concentration of floe particles in the bulk solution, CD, is therefore
related to a by
4/3r3cb = a (B-16)
o
and the excess surface concentration in particles/cm , c is related to
s" by
excess
c = IT /(2/3r2) (B-17)
s excess' v J
From Eqs. (B-13) and (B-16) we may readily show that the bulk floe particle
concentration at which the surface is half saturated Cs"QV^oec = -2) is given by
r • = n n Sexp(BAG)
b L u l-exp(BAG)J
= exp(BAG)/(4/3r3) (B-18)
109
-------
We next examine the sizes of the floe particle-bubble energies for
representative values of the parameters. We set XaW = 40 dynes/cm, correspon-
ding to the surface tension of NLS solutions at roughly the critical micelle
concentration. We assume floe particles 200 nm in diameter. The floe-bubble
binding energy as a function of contact angle is then given in Table B-l.
TABLE B-l. FLOC-BUBBLE BINDING ENERGY AS A FUNCTION
OF CONTACT ANGLE5
cosO -AG(erg) x 1010
1.0
0.9
0.8
0.6
0.4
0.2
0
-1.0
0
1.257
5.027
20.11
45.24
80.42
125.66
502.64
akT = 4.14 xlO-14erg at 25°C
Even for these rather small particles the floe-bubble binding energy is
roughly a thousand times larger than kT; AGJ remains much larger than kT for
contact angles of any appreciable size for particles of 20 nm diameter.
Floe-bubble attachment is evidently an interaction which overwhelms the effects
of random thermal motions unless the contact angle is essentially zero, so
floe foam flotation should be an extremely efficient process as long as the
floe is hydrophobic. This requires that the surfactant concentration and the
floe surface potential be such that the surfactant can form a condensed sur-
face phase (hemimicelle) with the hydrophobic hydrocarbon chains of the
surfactant presented to the liquid phase; conditions for this to occur are
explored later in this Appendix.
Estimation of contact angles in this model is likely to be difficult.
The floes of interest are often hydrous oxides, extensively and probably
irreproducibly hydrated, and their characteristics are surely changing with
time. Contact angle measurements on well-characterized crystalline materials
can be no more than a rough guide in estimating contact angles on these fresh,
imaged floes.
We have also examined in detail an alternative model, in which small
bubbles interact with plane solid surfaces (corresponding to large particles).
The analysis, presented elsewhere (52) is slightly more tedious, and yields
conclusions basically similar to those of the model treated above. The
particle-bubble binding energy is orders of magnitude larger than kT for
bubbles of diameter 20 nm or greater. Thus, with either of the two models,
the binding energy is ample to produce efficient separations, despite random
thermal forces.
In stripping columns there is a countercurrent flow of liquid downward
between rising air-water interfaces. We will next demonstrate that the
110
-------
viscous drag forces resulting from this countercurrent flow are insufficient
to tear loose floe particles bound to the air-water interface (52).
We equate the gravitational pull down on an element of liquid Icmx 1cm x
thickness dx to the viscous drag in the opposite direction to obtain
Pg = -nd2v/dx2 (B-19)
where p = liquid density
g = gravitational constant
t) = viscosity
v(x) = velocity of the liquid in the downward direction a distance x
from one side of a vertically oriented lamella.
Integration of Eq. (B-19) and use of the boundary conditions v(0) = v(£) = 0
yields
v - |&(£x - x2) (B-20)
where £ is the thickness of the lamella.
We estimate the viscous drag force on the floe particle by assuming that
the spherical particle is attached to the air-water interface and is moving
with it, and that the relative velocity of the liquid streaming past it is
given by v(x). We let the particle radius be r and estimate an upper limit
to the viscous drag on the particle
v(x = 2r) = ^(2£r - 4r2) . (B-21)
Viscous force = f = Girnrv (B-22)
fy - 67rpgr2(£ - 2r) (B-23)
_2
If the diameter of our particle is 200 nm and the film thickness is 10
cm, the viscous force is 1.84 x 10"^ dynes. We roughly estimate the binding
force of the floe particle to the air-water interface as
(B_24)
i ,-. ~~ *? ~~
10 cm
at least three order of magnitude greater than the viscous drag. The binding
force increases proportional to r, while the viscous force increases roughly
proportional to r^, so viscous drag may impede floe foam flotation when
particle radii are of the order of 10"^ cm or larger, or if the foams are
extremely wet (large £).
We have presented elsewhere (51) a discussion of the viscous drag forces
on floe particles attached to spherical bubbles rising in liquid in batch foam
floation work.
Ill
-------
2. Surfactant Adsorption Isotherms
We here examine a model for the formation of hydrophobia surfaces on floe
particles by adsorption of the surfactant ionic heads directly into the pri-
mary layer on the surface of the floe (41). The hydrocarbon chain tails of
these surfactant ions then, at sufficiently high surface concentration, form a
hydrophobic surface. Surface free energies are thereby changed such that non-
zero contact angles of air bubbles on the solid surface occur, and flotation
takes place.
Of particular interest is the formation of a condensed surface phase of
surfactant (a hemimicelle) on the floe particle surface at sufficiently high
surfactant concentrations in the bulk solution. This cooperative phenomenon
results from the Van der Waals stabilization energy of the surfactant hydro-
carbon chains when they are able to pack closely together; and Fuerstenau,
et al. (50) have showed experimentally that, as one would expect, it occurs
at lower surfactant concentrations as surfactant chain length increases.
This cooperative phenomenon accounts for the abrupt and large increase of
flotation efficiency which is observed with increasing surfactant concentra-
tion at intermediate surfactant concentrations. The competition of non-
surfactant ions with surfactant ions for space in the primary layer accounts
for the effects of varying ion identities and ionic strength, as discussed
in Appendix C.
We calculate the adsorption isotherm of surfactant on the solid by means
of an approximate method described by Fowler and Guggenheim (76); exact
treatments of even quite simple two-dimensional cooperative phenomena are
extremely complex and difficult. The binding energy of an isolated surfactant
ion to the solid is calculated by calculating the electric potential in the
vicinity of the solid surface by means of a modified Poisson-Boltzmann
equation. Fuerstenau and Healy (50) estimate the free energy of removal of
hydrocarbon chains from water as approximately -0.6 kcal/mole of CH2 groups;
this information is needed in connection with the surface condensation
phenomenon.
We take as our starting point the following adsorption isotherm, the
derivation of which is given in Fowler and Guggenheim's book (76).
J
«6>-r?% rfi^J CB-
= [1 - 46(1-0) {1 - exp(-2w/zkT)}]ls (B-26)
/ZrrmkT \
\h2 /
C0 = (^m*± J kT J^ii exp(-Xn/kT) (B-27)
JA(T)
Here
0 = fraction of surface sites on the solid which are occupied by
surfactant ions
c(0) = concentration of surfactant in bulk solution in equilibrium
112
-------
with solid surface having a fraction 0 of its sites occupied
by surfactant
z = maximum number of nearest neighbors of a surfactant ion on the
surface, taken here as six
2w/z = increase in energy when a new pair of nearest neighbors is
formed
X0 = increase in energy when a surfactant ion is adsorbed on an
isolated site on the solid surface
m = mass of surfactant ion
k = Boltzmann's constant
h = Planck's constant
„ T = absolute temperature
j (T) = partition function for the internal degrees of freedom of a
surfactant ion in solution
j (T) = partition function for the internal degrees of freedom of an
adsorbed surfactant ion
Assume that j /j^ is essentially independent of the inert salt concentra-
tion, and that XQ can ^e calculated by multiplying the charge on an adsorbed
surfactant ion by the electric potential at that point. We write c =
c'exp(-XQ/kT), where c' is assumed to be independent of ionic strength and no
more than weakly temperature-dependent. For dodecyl sulfdte at room tempera-
ture, c' = 2.85 x 10~4 mole/£. We define a reduced concentration, a, as
(2-20 \
3 + 1 - 20 I
o(e) = ^- = ^ *°"" Yr-e I « : i "~™ J CB'28)
As noted above, the free energy for the removal of hydrocarbon chains
from water has been estimated as -0.6 kcal/mole of CH2 groups. This yields
Eq. (B-29) for w, where n is the number of CH2 groups in the surfactant tail
plus 1. We note that roughly six new pairs of nearest neighbors are formed
when a surfactant ion goes into the condensed surface phase.
w = -n x 2.070 x .0~14ergs (B-29)
We previously showed (39) how the method of Macdonald and Brachman (77)
could be used to calculate electric potentials in the electric double layer
for solutions in which the finite volume of the electrolyte ions was taken
into account. The Poisson-Boltzmann equation for the case of planar geometry
is
d2iKx) _ A sinh (z'e»/kT) fR „,
,2 1 + B cosh rzf~'./tTt >-D ™)
dx
where ty(x) = electric potential a distance x from the solid-liquid interface
|z'e| = magnitude of the charge of the ions of z'-z' electrolyte
STTZ'eCoo
A =
(1-2 cro/c ) D
^ °° max
B = 2 c /(c - 2c )
°°' v max °°'
113
-------
c = anion (or cation) concentration in bulk solution, ions per
c = maximum possible concentration of ions in solution, ions
max -i ^ '
per cm0
D = dielectric constant
The solution to Eq. (B-30) is given by
. B cosh
log
\."
where ^Q is the potential at the solid-liquid interface. We assume that the
center of charge of the surfactant ion is at a distance £ from the charged
floe surface; ^(£) is then obtained from Eq. (B-31) . The binding energy of
an isolated surfactant ion on the solid surface is given by
XQ = zse a(^) for 6 < %,
we set a(6) = aft); if a(6) < a (%), 9 > %, we set a(6) = aft). The situation
is essentially identical to that arising with Van der Waals isotherms for
nonideal gases below the critical temperature (86) .
Results
Some representative adsorption isotherms are shown in Fig. B-2 which
exhibits the effects of varying w, the surfactant-surfactant interaction
energy. The values of w used here are somewhat smaller than would be used,
say, for sodium lauryl sulfate (2.48 x 10"* erg) in order to nake the fea-
tures of the plots clearer. The reduced concentration at which hemimicelle
formation occurs, ocrit> is seen to decrease rather rapidly with increasing
values of w (and of the hydrocarbon chain length) , as noted by Fuerstenau and
co-workers (50), and the critical temperature increases, as indicated by Eq.
(B-33).
114
-------
3x 10
r 2
CT
0.5
1.0
e
FIGURE B-2 - ADSORPTION ISOTHERMS OF SURFACTANT ON FLOG WITH
CONDENSATION. FOR ALL PLOTS T = 298°K, X° = 8.0 x 1CT14 erg.
FROM THE TOP DOWN: - w = 1.0, 1.04, 1.1, 1.2, and 1.4 x lO'13 erg.
115
-------
The dependence of acrit on w is indicated over a much wider range of w
in Fig. -3. On noting that acrit = &(%) and using Eqs. (B-29) and (B-34),
we obtain
5 r
O)
o
FIGURE B-3 - DEPENDENCE OF a on W. 4/0 = 50 mV, T = 320°K,
-3 crit
1 = 10 cm, a =10 mole/1, z = -1, c = 1.0 mole/1.
^° J ITlciA
116
-------
-31og acrit 1_ -14
3n kT x z.u/u x iU
-31og1Q acrit = 65_2/T = 0_2919 (B
3n
at 298°K. The value of this derivative, which we obtain from the results of
Fuerstenau, Somasundaran, and his co-workers (50), is approximately 0.32.
Increasing the magnitude of the interaction energy of the hydrocarbon chains
from 0.6 to 0.88 kcal/mole of CH2 groups eliminates the discrepancy; in view
of uncertainty in the figure of 0.6 and the approximations inherent in the
theory, the agreement is better than one might expect.
The effects of surface potential 4'Q on the value of ocrit are shown in
Fig. B-4, and the dramatic decrease in ocr^t with increasing surface potential
is what one would intuitively anticipate. Ionic strength also exhibits a.
marked effect upon acrit> as shown in Fig. B-5, as is well established experi-
mentally. We note that our analysis here is predicated upon the assumption
that the concentration of surfactant in the bulk solution is at all times
sufficiently low that the surfactant is an ideal solute; in particular, we
have ignored the possibility of micelle formation in the bulk solution.
This surely leads to spurious results at the upper end of the curve in Fig.
B-4, where one would expect acrit to be a very rapidly increasing function of
l°§10Coo- The binding energy of the surfactant to the solid, XQ* is reduced
to values less than kT with increasing Co,, and the formation of micelles then
should become competitive with the formation of hemimicelles on the solid sur-
face. At this point, between roughly cm = 0.1 and 1.0 in Fig. B-5, foam
flotation should cease.
Experimentally, we did observe that foam flotation is less efficient at
higher salt concentrations and that increasing the surfactant concentration
above the levels which are effective at lower ionic strengths is of little
or no help. The ionic strength at which failure occurs depends somewhat on
pH, and therefore upon IJJQ, as one would anticipate (56).
TABLE B-2.
T
298
320
340
Q
In these
mole/£, ij
. EFFECT OF TEMPERATURE
ON a "
crit
crit
1
2
2
runs, £ = 10~7 cm,
!»n = 50 mV, zs = -1
.33 x
.08 x
.99 x
Coo =
, and
io-3
ID'3
io-3
io-3
z' = 1.
The effect of temperature on ocrit is shown in Table B-2. We note that
acrit increases with increasing temperature rather rapidly; as long as oci>.t
does not approach the reduced critical micelle concentration and as long as
the reduced surfactant concentration exceeds C7cr^t, flotation should take
place. However, these results indicate that the high-temperature cut-off of
117
-------
50
75
100
125
(mv)
FIGURE B-4 - DEPENDENCE OF
ON
T = 298°K, £ 10
"7 cm, c^ = ICT3 m/2, ,
~13
c = 1.0 moleA, w = -2 x 10 erg.
max &
118
-------
8x10"
O~cr\t
-3
-2
-1
10
FIGURE B-5 - DEPENDENCE OF acrit ON e T = 298°K, 1 = 10"7 cm, Cmax = 10
moles/1,
= 50 mV, w = -2 x 10
~
erg.
119
-------
foam flotation should depend markedly on surfactant concentration. This type
of temperature dependence does not appear in our earlier approach (35-40), and
an experimental study of the interaction of temperature and surfactant concen-
tration at relatively low surfactant levels may provide a means of distinguish-
ing between the two models. This result also suggest the importance of
investigating the temperature dependence of the air-water surface potential
within the framework of our other model. We have discussed this elsewhere (41).
120
-------
APPENDIX C
THEORY OF SURFACTANT DISPLACEMENT BY SALTS
A matter of considerable impact on the economics of foam flotation is
the extent of which surfactant can be recovered for recycling. Within the
framework of the Fuerstenau-Somasundaran-Healy model (43-50) , one expects that
a very substantial fraction of the surfactant in the collapsed foamate from
floe foam flotation will be absorbed on the floe sludge, interfering with the
settling of the sludge and the recovery of the surfactant. An analysis of the
adsorption isotherms of the surfactant on the floe and the effect on these of
varying ionic strength is given in Appendix B, in which the adsorption of
nonsurfactant ions into the primary layer on the floe, and the competition
of surfactant and nonsurfactant ions for sites in the primary layer, is
neglected. We here present an analysis which takes these effects into account.
Our results indicate the feasibility of displacing surfactant ions from the
floe sludge by the addition of nonsurface-active salts.
ANALYSIS
We shall first look at the simple case in which the surfactant ions do
not interact with each other; this is readily seen to yield Langmuir-type
adsorption isotherms for the surfactant ion and the nonsurface-active com-
peting ion.
^y
NS = number of surface sites/cm^
0^ = fraction of sites occupied by surfactant ions A
Og = fraction of sites occupied by non-surface-active ions B
CA = bulk concentration of A
eg = bulk concentration of B
At equilibrium the rates of adsorption and desorption of A (and also of
B) are equal, yielding
- vw1 - SA - v (C-2)
Solution of these equations leads in the usual way to
b.cA
6 = _A_A _ (C-3)
6A (1 + bc + bc)
121
-------
be
a _ " " _ ff ,«•>
B " C1 + VA + VB> C }
where bA = k^/k^f and bg = kgr/kgf. We see, as expected, that increasing eg
results in decreasing 6^, displacing surfactant from the floe particle surface
The model, however, is rather unrealistic, in that it neglects the Van der
Waals interactions of the hydrocarbon chains of the surfactant ions which, as
Fuerstenau and his co-workers have observed experimentally (43-50), have a pro-
found effect on the shapes of the adsorption isotherms. Surface condensation
may occur, resulting in a sudden increase in QA from slightly greater than
zero to slightly less than one, with a slight decrease in temperature or a
slight increase in surfactant aoncentration.
We shall next attack the problem of competition for surface sites when
interactions between the surfactant ions are significant. We use a generali-
zation of an approximate method described by Fowler and Guggenheim (76) .
We let
z = number of nearest neighbors of a site
N^ = number of sites occupied by A
NB = number of sites occupied by B
^XY ~ average number of pairs of sites occupied by X and Y; X, Y = A, B
or 0 (empty)
We take into account the A-A pair interaction energy, 2w/z, as follows:
4NAAN00 = NA02exP(-2w/zkT) (C-5)
4NBBN00 = NB02 (C-6)
2N00NAB = NAONBO ^
2NM + NAO + NAB = ZNA (C-8>
2N00 + NAO + NBO = *(NS - NA - NB) (C-9)
2NBB + NBO + NAB = ZNB (C
(This is a straightforward extension of the approach used for a single ad-
sorbed species in Ref. 76.) A lengthy series of successive eliminations
finally yields a remarkably simple equation for NAA:
4(D-1)NM2 - [4(D-1)NA + 2Ns]zNM + Dz^2 - 0 ' (C=l
where D = exp(-2w/zkT) . This is identical to Fowler and Guggenheim's Eq.
(1010,1). We follow these authors to obtain
122
-------
NAA
6 = [1 - 4eA(l-0A)(l-D)]is (C-13)
The expression for the chemical potential of A is given by (76)
w = logAA = logT7Q-^ - log aA°(T)
(B-l+26 )(l-0 )
'
where a/^°(T) is the partition function for the internal degrees of freedom of
the surfactant ion. We note that
aA°(T) = exp(xA/kT)JA(T) , . (C-15)
where j^fT) is the partition function for the internal motions of a surfactant
ion in bulk solution and XA is the binding energy of an isolated surfactant
ion to the floe -water interface.
In solution we assume the chemical potential of a surfactant ion to be
given by
PA 5
~ = const, -j log! + log CA - log JA(T) ' (C-16)
Equating Eqs. (C-14) and (C-16) then yields
9A X. (6-l+26J(l-ej
A Aw z , v AJ *- AJ
,
log
i-eA-eB - kf kf
= log-4 - logc ' (C-17)
CA A
where c.° is determined by the contant in Eq. (C-16).
f\
In similar but simpler fashion,
- * 10 T '
We can calculate adsorption isotherms from Eqs. (C-17) and (C-18) cal-
culating c^' and eg1 as functions of 6^ and 6g; it is more convenient to
calculate c^' and 83 as functions of B^ and Cg'. We solve Eq. (C-18) for
0g, obtaining
123
-------
(1-8.) r/?
> b = T
Some typical isotherms are plotted in Fig. C-l. We see that unstable
phases may occur [where (8y^/98^) cfi = kT(Slogc^'/3 6^' )CR is negative], so that
the system splits into two stable phases. We next determine the values of 6^
for these stable phases; again we follow Fowler and Guggenheim. If SJQ and
6X2 rePresent the values of 6^ for the two surface phases in equilibrium with
each other (X = A, B) , then we must have
XA1 S V
AB1 ~= V6
and (6A23^2) , (C-22)
where 41 is the spreading pressure of the surface phase.
We generalize Fowler and Guggenheim's Formula (1008,5) for d (76) to get
N kT
and note that Eq. (C-22) can be written as
J2d
-------
From Eq. (C-25) we have
90,
and
l-i
A B
(C-28)
(C-29)
From Eq. (C-14) we have
1
i-e.-en
(C-30)
and
1 -
2 -
93
+ 2-r-
36
A 1
(C-31)
33
36.
(C-32)
where 3 is defined by Eq. (C-13) .
We now wish to find values 6Aj and 6^2 such that Eqs. (C-20) and (C-26)
are satisfied. [Use of Eq. (C-19) in Eq. (C-25) guarantees that Eq. (C-21)
is satisfied.] We proceed as follows. For fixed eg' we calculate a table
of logXA as a function of 8A = nA0, n = 1, 2, ..., N[9g is determined by Eq.
(C-19)]. Then we find those values of nA6 for which logXA[(n + 1)A8] -
logAA[nA6] change sign. If there are none, our table and Eq. (C-16) give us
the adsorption isotherm, logcA' as a function of 8A. If there are two, we
have a loop in our isotherm, and a phase transition occurs. In this case let
the two values of n be n^ and nr, and calculate logAA[n£ + nr)/2](A6). Then
increment n above (n£ + nr)/2 until logXA(nA6) - logAa[(n^ + nr)/2](A0)
changes sign; call the value of nA6 for which this occurs 62- Similarly,
decrement n below (n^ + nr)/2 until logAA(nA8) - logAA[[n^ + nr)/2](A8)
changes sign; call this value of nA 6 8. .
We use 0j and 62 as trial limits for the integral Eq. (C-26), which we
evaluate numerically with the aid of Eqs. (C-27) to (C-32). If the integral
is greater than (less than) zero, we replace 6^ by 0^ + A0(6j - AB), determine
the new value of logAA(8^) from the table, and we then increase (decrease) 8A
from 82 until logA^t 6A) - logAA( 6^) changes sign; this value of 0A is our new
upper limit. We use the new values of 0j and 62 as trial limits for the
integral, Eq. (C-26), and continue this process until the integral changes
sign. The values of 6^ and 62 for which this occurs are our desired 6^1 and
8. 2 5 which specify the fractions of surface coverage by surfactant in the two
phases in equilibrium with each other.
125
-------
RESULTS
Adsorption isotherms were calculated by the procedure outlined above on
an XDS Sigma 7 computer; about 3 sec of machine time was required per isotherm
The results are plotted as logeA^ vs 9^ for various values of the parameters.
In Fig. C-l we see the effect of increasing the reduced concentration of
salt, eg'. As the salt concentration increases, the activity of surfactant
required to bring about condensation of the surfactant on the floe surface
(A^con) increases, too, although the shape of the isotherm is unaffected. The
dependence of logeX^con on eg' is shown in Fig. C-2. The larger the binding
energy of the salt ion to the floe, Xg, the more effective it is in preventing
condensation of the surfactant on the floe, or (equivalently) in displacing
sorbed surfactant from the floe; this is shown in Figs. C-2 and C-3.
Thus, with positively charged ferric hydroxide floes, one would expect
that univalent anions would be fairly effective in displacing anionic sur-
factants such as the alkyl sulfates, but that divalent anions would be com-
parably effective at substantially lower concentrations. One would also
anticipate that complexing or chelating ions would be especially effective
in displacing surfactant.
The effect of the magnitude of w, the energy of interaction between the
surfactant ions, is shown in Fig. C-4, and is in agreement with our earlier
calculations on a similar but simpler model [Appendix B (41)]. The effect
of the binding energy holding an isolated surfactant ion to the floe surface,
X^, is shown in Fig. C-5, and is also similar to our earlier results (41).
The results of varying temperature are shown in Fig. C-6; the bulk concentra-
tion of surfactant required to produce a condensed surface phase increases
with temperature, as one would expect.
126
-------
-3
-5
0.2
0.4
0.6
0.8
1.0
FIGURE C-l - EFFECT OF COMPETING SALT CONCENTRATION Oa") ON TH
~13
ADSORPTION ISOTHERM. T = 298°K, X^ = l-2 x 10> XB = l-2 x
z = 6, c' = 30, 10, 3, and
w = -1.234 x 10~13erg,
1 X 10 (TOP TO BOTTOM) .
127
-------
loge XAcon
-6
0
3x 10s
FIGURE C-2 - EFFECT OF
ON SURFACTANT CONDENSATION CONCENTRATION. T =
298°K, XA = 1-2 x 10~13, W = -1.2 x 10~13 ERG, z = 6. XB =
1.2 x 10~13 (LOWER) and 1.8 x 10"13 (UPPER) ERG.
128
-------
-2 i-
-3
logeX
'e A
-5
-6
0
0.2 0.4 0.6 0.8 1.0
FIGURE C-3 - EFFECT OF XB ON THE ADSORPTION ISOTHERM.
= 105; XD = 2.4, 1.8, and 1.2 x 10"13 erg (top to bottom);
D
other parameters as in Fig. 1.
129
-------
-3 r
-4
log. X
e A
-7
I
I
I
I
0 0.2 0.4 0.6 0.8 1.0
^A
FIGURE C-4 - EFFECT OF w ON THE ADSORPTION ISOTHERM, w = -1.0, -1.2,
"13
-1.4, and -1.6 x 10
105
ERG (TOP TO BOTTOM);
OTHER PARAMETERS AS IN FIG. 1.
130
-------
-5
-6
log,X
'e A
-7
-8
0.2 0.4 0.6 0.8 1.0
a.
FIGURE C-5 - EFFECT OF XA ON ™E ADSORPTION ISOTHERM. GE" = 10 ; XA =
1.2, 1.8, and 2.4 x 10~13 ERG (TOP TO BOTTOM); OTHER PARAMETERS AS IN FIG. 1,
131
-------
-4
0.2
0.4
0.6
0.8
1.0
FIGURE C-6 - EFFECT OF TEMPERATURE ON THE ADSORPTION ISOTHERM,
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-80-138
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
FOAM FLOTATION TREATMENT OF INDUSTRIAL
WASTEWATERS: LABORATORY AND PILOT SCALE
5. REPORT DATE
June 1980 issuing date,
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
David J. Wilson, Edward L. Thackston
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Vanderbilt University
Nashville, TN 37235
10. PROGRAM ELEMENT NO.
IBB610
11. CONTRACT/GRANT NO.
R-804438
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Lab-Cincinnati, OH
Office of Research & Development
U.S. Environmental Protection Agency
Cincinnati, OH 45268
13. TYPE OF REPORT AND PERIOD COVERED
14. SPONSORING AGENCY CODE
EPA/600/12
15. SUPPLEMENTARY NOTES
16. ABSTRACT
A floe foam flotation pilot plant reduced lead and zinc in dilute solution to
very low concentrations. The results suggest a number of design improvements. A
simple diffusion model does not adequately describe axial dispersion at high column
leadings. The floe foam flotation of zinc, cobalt, nickel, chromium (III), and simpli
and complexed cyanides was carried out. Modified procedures make floe foam flota-
tion of copper compatible with several precipitation pretreatments. The flotation
of ferric hydroxide floes is profoundly affected by polyvalent anions such as sil-
icates and phosphates. The flotation of mixtures of copper, lead and zinc was
successfully carried out. A surface adsorption model for floe foam flotation was
analyzed and found to account for the effects of ionic strength, specifically ad-
sorbed ions, surfactant concentration, and surfactant hydrocarbon chain length.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.IDENTIFIERS/OPEN ENDED TERMS
COSATI Field/Group
Flotation
Wastewaters
Metals
Surfactants
Floe foam flotation
Flotation column
Pilot plant
Continuous flotation
13B
3. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
143
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
133
U.S. GOVERNMENT PRINTING "FFICF: 1080--557-165/0010
------- |