WATER POLLUTION CONTROL RESEARCH SERIES • 14010 FUI 10/71
  Foam Separation of
   Acid  Mine Drainage
U.S. ENVIRONMENTAL PROTECTION AGENCY

-------
           BATER POLLUTION CONTROL RESEARCH SERIES

the Water Pollution Control Research Series describes the
results and progress in the control and abatement of pollution
in our Nation*s waters.  They provide a central source of
information on the research, development, and demonstration
activities in the Environmental Protection Agency, through
inhouse research and grants and contracts with Federal, State,
and local agencies, research institutions, and industrial
organizations.

Inquiries pertaining to Water Pollution Control Research Reports
should be directed to the Head, Publications Branch (Water),
Research Information Division, R&M, Environmental Protection
Agency, Washington, D.C. 20460.

-------
Foam Separation  of Acid Mine Drainage
          Horizons Incorporated
          23800 Mercantile Road
          Cleveland, Ohio 44122
                 for the
       Environmental Protection Agency
          Project No. 14010 FUI

          Contract No. 14-12-876
              October 1971

-------
                   EPA Review Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.   Approval
does not signify that the contents necessarily reflect
the views and policies of the Environmental Protection
Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendations for
use.
                              ii

-------
                       ABSTRACT


Laboratory studies of continuous flow foam separation were
conducted to determine the optimum operating conditions for
maximum extraction of dissolved metal cations  (Fe, Ca, Mg,
Mn and Al) from acid mine drainage.  Foaming experiments
were condircted in a 6 in.-diameter glass column capable of
liquid flow rates of 3-12 gal. per hour.  The approach to
foam separation taken was the production of persistent
foams which allowed protracted foam drainage to reduce
liquid carry-over in the foam.  The effects of pH, chelate
addition, surfactant type and concentration, air sparging
rate, metal concentration and foam drainage were investi-
gated in relation to metal extraction.
                                               _y
The average extraction rate obtained was 1.9x10   moles total
metal per cm  column cross-section area per minute which is
approximately 4.0xlO~  equivalents per cm  per minute.
Operation in simple and countercurrent foaming modes produced
similar extraction rates for acid mine drainage.

The low extraction capacity of foam fractionation renders the
process economically unfeasible for the treatment of acid
mine drainage.  The principal chemical cost is for surfactant
followed by air; either alone makes the process noncompetitive,

Surfactant regeneration from collapsed foam by the addition of
base was investigated as a means for surfactant reuse and
cost reduction.

This report was submitted in fulfillment of Project Number
14010 FUI, Contract 14-12-876, under the sponsorship of the
Water Quality Office, Environmental Protection Agency-
                          iii

-------
                       CONTENTS






Section                                              Page



    I     Conclusions                                  1



   II     Recommendations                              3



  III     Introduction                                 5



   IV     Foam Separation Apparatus                   11



    V     Analytical Methods                          19



   VI     SAMD Experiments-Results and Discussion     21



  VII     AMD Experiments-Results and Discussion      41



 VIII     Surfactant Regeneration                     43



   IX     Economic Evaluation and Summary             47



    X     Acknowledgements                            51



   XI     References                                  53



  XII     Notations and Glossary                      57

-------
                       FIGURES


Number                                              Page

   1     Modes of Continuous Foam Separation          8

   2     Separation Column                           12

   3     Equilibrium Column                          13

   4     Column Nomenclature                         14

   5     Sparging Head                               16

   6     Distribution Ratios                         30

   7     Volume Reduction versus Sparging Rate       33

   8     Foam Density versus Foam Residence Time     34

   9     Dynamic Surface Excess versus Pool
         Concentration                               36

  10     Total Metal Fraction Removed and Total
         Metal Extraction Rate versus Total
         Metal Feed Rate                             38

  11     Process Flow Scheme for Foam Separation
         of Metals and Surfactant Regeneration       48
                          VI

-------
                         TABLES
No.                                                   Page

 1     Independent Control Variables in Continuous
       Foam Separation Processes                        9

 2     Dynamic Surface Excess Values for Selected
       Surfactants in SAMD                             24

 3     Effect of pH on Metal Separation from SAMD      26

 4     Dilute SAMD                                     29

 5     Comparison of Simple and Stripper Mode
       Experiments                                     31

 6     Summary of Results for SAMD Experiments and
       Comparison with AMD Experiments                 39

 7     Grassy Run Acid Mine Discharge                  41

 8     Recovery of NaDS from Collapsed Foam by
       Addition of NaOH                                44
                          vii

-------
                       SECTION I

                      CONCLUSIONS
1.  Continuous foam separation of mixtures of acid mine
drainage (AMD) and anionic surfactants achieved the ex-
traction of dissolved metal cations at a practical rate
of 1.9xlO~7 moles total metal per cm  column cross-sec-
tion area per minute which is approximately 4.0xlO~7
equivalents per cm2 column cross-section area per minute.
These rates represent 16 percent of the estimated theo-
retical maximum extraction capacity of foam separation.

2.  Operation of foam fractionation in a persistent foam
regime in order to minimize foam density requires exce's-
sive surfactant consumption which results in increased
chemical costs.

3.  The low extraction capacity of foam fractionation
renders the process economically unfeasible for the treat-
ment of AMD.  The principal chemical cost is for surfactant
followed by sparging air; either alone makes the process
noncompetitive.

4.  Bubble production by sparging air through porous
materials, although convenient and effective, is not en-
tirely satisfactory for control of optimum bubble diameter
and a narrow bubble distribution.  Bubbles tend to be
larger than optimum.

5.  Synthetically prepared AMD proved to be an accurate
substitute for AMD in foaming tests.

6.  Increasing the pH above the 2.2 - 2.3 of synthetic AMD
tended to reduce metal separation, particularly for iron.

7.  The use of EDTA with anionic and cationic surfactants
did not improve metal separation over anionic surfactants alone.

8.  Foam column operation in simple and countercurrent mode pro-
duced comparable metal removals.

-------
                      SECTION II

                    RECOMMENDATIONS
The economic unfeasibility of foaming AMD in a persistent
foam regime precludes further study of that system.  The
high chemical costs for foaming AMD in a persistent foam
regime could be reduced by foaming in a transient foam
regime which requires less surfactant.  The low extraction
capacity of foam fractionation, which is governed by
stoichiometric adsorptive interaction of-metal cation and
surfactant anions, could be increased by operation in a
non-stoichiometric system, such as in froth flotation, which
may require less surfactant and sparging air per mass of
metal removed from AMD.  Further research on foam separa-
tion treatment of AMD should be directed to non-stoichio-
metric, transient systems.

The use of waste surfactant and air should be investigated.
The combination of waste surfactant in municipal sewage,
the tendency for autopurification of AMD and sewage mix-
tures, and the high pressure method of foam separation
(bubble formation by degasification of air supersaturated
liquid) appears to have favorable technical and economic
feasibility and should be investigated.

-------
                       SECTION III

                       INTRODUCTION
Environmental degradation caused by the acidic, iron rich
effluent from coal producing regions has required assiduous
efforts to find cost-effective'-abatement and treatment
methods.  The treatment method for acid mine drainage (AMD)
under investigation in this study is classically known as
foam fractionation and often called foam separation.
Separation is based on selective adsorption of dissolved
surface active material (surfactant) at the liquid-gas
interface of rising bubbles.  Surface inactive material can
be separated by coadsorption with surfactant.  Separation of
adsorbed and coadsorbed material from liquid is accomplished
by the production and collection of foam.

The purpose of this investigation is to determine the tech-
nical and economic effectiveness of continuous-flow foam
separation as a treatment method for AMD.  Towards this goal,
a laboratory scale, pilot plant apparatus capable of 3-12 gal.
per hour   was constructed and operated to define and
evaluate the variables which control separation and, thus,
the practical application of the method to AMD on a field
scale.  For expediency, the majority of work was conducted
using a synthetically prepared acid mine drainage (SAMD)
followed by confirming experiments using freshly sampled AMD.

Numerous small scale, laboratory investigations of foam
separation, generally of a theoretical nature utilizing
simplified solutions, have empirically developed the basic
theory on which this practical investigation of a complex
solution (AMD) is based (1-22).  Even so, the complexity
of a foam separation system has not permitted a single uni-
fying theory and thus an empirical approach is necessitated.
Applied investigations of foam separation for treatment of
various polluting liquid wastes, a few on a pilot plant scale,
have appeared (20, 23-27).  Several reviews of foam separa-
tion processes are available (28-30).

This research program is divided into two main efforts; the
construction phase in which the foaming apparatus was
assembled and tested, and the experimental phase in which
SAMD and AMD were treated.  Once the apparatus was opera-
tional, the operating variables which control separation
efficiency were determined within the practical limits of a
potential pollution control process which in large part are
restricted by hard economics.

-------
Foam separation is based on the phenomenon of surface
activity which results from the ability of certain solutes
(surfactants) to reduce the surface free energy of their
solutions, and therefore the total free energy of the
system, by accumulating at an interface.  Surface activity
as it relates to foam separation process is described using
the concept of Gibbs surface excess.

         ~         fl
     - d /    = RT£ I; d In a;                      (D
            ~     M  L       L
     where  / = interfacial surface tension

            R = gas constant

            T = absolute temperature

            n = number of solute components

            I; = surface excess of the i component

            a = chemical activity of the i component
In practice foam separation consists of passing bubbles
through a solution of surface active solute(s) with the aim
to adsorb the solute(s) onto the gas-liquid interfaces and
to remove these surfaces intact as foam, thus effecting a
separation.  Further, by coadsorption of non-surface active
solute with surface active solutes, the former can be
separated from solution with the latter.  This is the case
in the treatment of AMD.

An understanding of the experimental approach used here can
be gained by considering three factors:  (1)  the particular
goal of foam separation as applied to acid mine drainage,
(2) the separation characteristics of the various operational
modes of foam separation processes, and (3) practical con-
sideration of the final application of this research.

Foam separation can operate in four basic modes, with each
mode offering potential separation characteristics appli-
cable to different goals (Figure 1).  These goals can be
described as:

     1.  Decontamination, in which the goal is separation
         of constituents from the feed to produce a puri-
         fied bottom flow.

     2.  Recovery, in which the goal is separation and
         recovery in the foam of constituents in the feed.

-------
     3.  Separation, in which the goal is sepp.ration of
         multiple constituents, one from another, by
         fractionation in the foam.
Obviously, the prime goal for the treatment of acid mine
drainage is decontamination and, thus, the operational
mode of foam separation is determined by this goal.

In comparison to simple mode operation which can be con-
sidered single theoretical plate operation, the stripper,
enricher and combined modes can be considered multiple
plate operations.  Theoretically, the stripper mode should
enhance the decontamination of the feed by providing ad-
sorption equilibrium between bubble film and interstitial
liquid of feed concentration instead of pool concentration
as in the simple mode.  This is accomplished by counter-
current flow of feed and foam, as depicted in Figure 1.
The purpose of recycling (external reflux) a portion of the
collapsed foam back into the rising foam, as in the enricher
and combined modes, is to further enhance the bubble film-
interstitial liquid equilibrium by providing an even richer
interstitial liquid than in the stripper mode.  The effect
of external reflux should be a higher concentration in the
net overhead product, Lc-r (Figure 1) and thus a more
efficient operation for recovery and concentration of feed
constituents.  Thus, the stripper mode should enhance de-
contamination of the feed, the enricher mode should enhance
the recovery in the foam and the combined mode should pro-
vide the benefits of both the stripper and enricher modes.

Since the prime goal in the treatment of acid mine drainage
is decontamination, the stripper mode should be given prime
consideration.  However, concentration of the overhead pro-
duct is still desirable, not for the purpose of recovering
the contaminants in acid drainage but rather for reducing
the liquid volume carried over in the foam.  A foam of low
liquid volume is produced by drainage of the interstitial
liquid out of the foam.  Provision for conditions which pro-
mote foam drainage is thus required for efficient overall
operation designed for the decontamination of mine drainage.

In a practical sense, the use of complex modes is often re-
stricted by operational difficulties.  For example, the
distribution of liquid feed or reflux into a rising foam
without causing channeling or bubble rupture is a real
problem.  Also, the effect of micelles introduced in con-
centrated reflux streams can adversely affect the bubble
film-interstitial liquid equilibrium and cause a reduction
of concentration in the overhead flow.

-------
            t
          Foam
          Pool
                  Jb
            tG
       Simple Mode
               c


               t



             Foam
             Pool



             T75"
         Stripper  Mode
            t
           Foam
           Pool
                    •+L
                       c-r
            TG
       Enricher Mode
               f
             Foam
             Pool
                      •*• L
                                                         c-r
               TG
          Combined  Mode
                          FIGURE 1


             Modes of Continuous Foam Separation
Lf  Liquid Feed  Flow


L,   Liquid Bottom Flow


G   Gas Sparging Flow
 c ™~
    Total Liquid Overhead Flow


      Net Liquid Overhead Flow
L   Liquid Recycle Flow
                          8

-------
Certain independent operating variables are known to sig-
nificantly affect the separation achieved in a foam column
(7) (Table 1).  Not every variable is continuously variable,
as bubble diameter is restricted by the flow dynamics of
the system (1) and available bubble producing equipment.
                        TABLE 1

           Independent Control Variables, in
         Continuous Foam Separation Processes

Geometric Variables  Solution Variables  Operation Variables
1.  Bubble diameter
2.  Column diameter
3.  Column length
1.  Ratio of the
concentration of
surfactant to
contaminant.

2.  Chemical
character of
the feed.

3.  Specific
chemical and
physical inter-
ferences .
1.  Ratio of the
flow rate of gas to
liquid feed.
2.  Liquid feed rate
as a function of
column dimensions.

3.  Foam drainage
time.
                                         4.  Modification of
                                         column to allow
                                         operation in various
                                         modes (Figure 1).
For this program, the solution variables are of secondary
importance since a standard SAMD is used.  The exceptions are
the amount and type of surfactant required and pH effects.
The main variables considered here are bubble diameter;
column length, particularly as it effects liquid and foam
residence times in the column; surfactant concentration or
better stated for the continuous flow system the mass flow
rate; gas flow rates particularly in relation to bubble
diameter, solute throughput, foam density and foam stability;
contaminate (metal ion) mass flow rate and operation in
simple and stripper modes.

-------
                      SECTION  IV

              FOAM SEPARATION  APPARATUS
Foam Column Description - Foam separation processes are
not in wide use.  With the exception of particle flotation
methods used in  the mining industry, which are not com-
pletely analogous to  foam separation of dissolved consti-
tuents in mine drainage, the state-of-the-art in foam
separation knowledge  and equipment  is still in its infancy.
As a result, not only must the physical-chemical and pro-
cess engineering aspects of applying foam separation to
mine drainage be studied, but, also, equipment must be
designed and fabricated.  The development of equipment is
generally costly and  time consuming, thus every effort was
made to make use of commercially available equipment and/or
components in order to speed fabrication.

The foam column  (Figure 2) was constructed of 6 in.-ID pyrex
pipe using teflon lined, stainless  steel couplings at the
joints.  The couplings are constructed with a nearly smooth
inside surface which  can be significant to foam stability
in the drainage sections.  The foam originates at the
sparging head in the  lower part of  the vertical section
and flows up and then to the right  in the horizontal drainage
section.  The foam passes from the  drainage section down into
the foam breaker (not visible) behind the control panel.
Structural support for the column and auxiliary equipment is
provided by an open pipe structure  of dimension 15 ft. L. x
9 ft. H. x 3 ft. D.   The horizontal drainage section is pro-
vided with valved ports which allow foam drainage to be re-
moved from the column, collected in a plastic pipe which
hangs below the glass pipe and either removed from the system
or reintroduced at a  desired point.  Also, shown are liquid
storage tanks, air and liquid control equipment, liquid
pumps and timing devices.

The column was constructed of multiple short lengths of
pipe to provide rapid manipulation  of such parameters as
pool depth and drainage section length which in turn control
bubble residence time in the liquid and foam drainage time,
respectively.  Referring to Figures 2 and 3 the column has
two basic configurations, the separation column and the
equilibrium column.   The separation column is of general
design, adaptable to  operation in any of the four modes
(Figure 1), and was used to optimize the geometric and
operating variables (Table 1) once  the soluble variables
were defined.  The equilibrium column was designed to study
the solution variables which in this case are determinations
of the dynamic surface excess (an equilibrium measurement
                           11

-------
Figure 2.
tion.
Separation Column.   The column is shown in opera-
from which the term equilibrium column was derived) for the
metallic constituents of mine drainage as a function of
both various types and concentrations of surfactants in the
feed.  The subject of dynamic surface excess will be dis-
cussed later.

Filtered, humidified laboratory air was used for sparging.
Air pressure control is provided by adjustable spring dia-
phragm-type regulators and gauges.   Air flow was controlled
with neec'le valves and rotameters.   The various types of
sparging heads tested are discussed later.

SAMD, AMD and surfactant solutions  were held in poly-
ethylene tanks.  The solutions were pumped into the foaming
column by positive displacement diaphragm metering pumps.
The liquid flow rates are controlled by adjustment of the
stroke rate and stroke length of the pumps, by needle valves
and rotameters.
                           12

-------
FIGURE 3.  Equilibrium Column.  The column is shown in
operation.  The absence of an extended drainage section
is the only significant structural difference from the
separation column.  The foam breaker is shewn.
Column Operation - Column operation was begun by pumping
feed solution(Surfactant and SAMD or AMD) into the vertical
foaming section, the pool volume of which is controlled by
height adjustment of the overflow discharge tube (Figure 4).
The surfactant and SAMD solutions can be mixed in line or
in batch prior to pumping; the later method was found more
reproducible and used almost exclusively.  When the pre-
determined static pool volume was reached, sparging was
begun and the column was allowed to operate until a steady
state was reached.  Ostensive indications of steady state
operation were stability in the height of the dynamic flooding
                           13

-------
a
o
•H
-P
O
0
K!
a
•H


o
EH
si
o
•H
-P
                                Horizontal Drainage Section
Counter-
current
Length
    Dynamic
    Flooding
    Level
             Bottom
             Exit
                                           Foam
                        Foam
                       Pool
                                          Drainage Collector
                                      Stripper Mode
                                      Feed Entry
                                                                       Foam Exit
                                       Feed Entry
                         t
                     Sparging
                     Air Entry
                                           FIGURE 4


                                     Column Nomenclature

-------
 level  and  steady  rate  of  foam drainage return.   As  the  foam
 proceeds up  the vertical  section,  drainage  returns  counter-
 current to foam flow.   In the horizontal  section drainage
 is perpendicular  to  foam  flow and  is  removed  through  valved
 ports  in the bottom  of the section (Figure  2).   Drainage
 is collected in the  drainage  return pipe  (Figure 2) and
 returned to  the vertical  section by gravity or  pumping.
 Usually from one  to  two hours were required to  reach
 steady state;  separation  results are  based  on samples
 taken  at steady state.

 Sparging Heads -  Production of a stable and persistent
 foam is a  fundamental  requirement  for achieving efficient
 separation of metal  ions  from acid mine drainage.   Since
 the goals  of separation are to decontaminate  the feed and
 to collect the contaminating  metal ions in  as small a
 foam volume  as possible,  a foam of sufficient stability to
 allow  protracted  drainage is  desirable.   As a foam  drains
 its persistence   decreases due to  thinning  of bubble  films.
 Thus in practice  it  is often  difficult to separate  foam
 drainage and bubble  coalescense.   Since the foam becomes
 progressively more susceptible to  rupture as  drainage pro-
 ceeds, the drainage  section of the column must  reduce to a
 minimum the  chances  of physical shock.  A horizontal  drainage
 section provides  a minimum countercurrent flow  of interstitial
 liquid through the foam by allowing gravitational drainage
 perpendicular to  foam  flow instead of parallel  through  the
 length of  the column.   By reducing countercurrent drainage
 flow,  the  chances of bubble rupture are reduced especially
 at low foam  densities.  Drainage sections must  be designed
 to provide minimum constriction or expansion  of the drainage
 foam,  and  as smooth  an inside surface as  practical.

 A foam is  produced from a number of individual  bubbles; how-
 ever the structure of  a foam  is best  visualized as  a  network
 of interconnecting films  surrounding  gas  spaces.  The pres-
 sure difference across a  bubble film  is
        is the surface tension and r is the radius of curva-
ture.  Thus it is evident that the pressure difference in-
creases as the radius decreases.  This explains the tendency
for larger bubbles to grow at the expense of smaller ones in
a foam of varying bubble sizes.  A foam with a wide distribu-
tion of bubble sizes tends to be self destructive.  As the
bubbles rupture, surface adsorbed material is released into
the interstitial liquid and the efficiency of separation is
reduced.  Thus it is necessary to produce a foam with as
narrow a bubble size distribution as possible.

One practical method to produce a foam for continuous foam
separation is by sparging air through an prifice(s) into a
                           15

-------
liquid and have the foam form above the liquid.  Since bubble
size and size distribution are important to foam separation,
considerable time was spent designing and testing sparging
heads.

An explicit relationship for the bubble size obtained by
sparging as a function of pore size, solution density and
surface tension has not been determined for multi-pore
devices.  Thus, it is necessary to conduct experiments to
find the sparging device most suited to the conditions.

Seven different sparging heads were constructed and tested.
Two heads were made from glass Buechner style filtering
funnels (Corning No. 36060) with 125 mm and 90 mm diameter
fritted discs of coarse porosity.   Porosity varied over
the discs which resulted in channeling and thus streams of
large bubbles.  The small diameter disc produced less
channeling, but still the bubble size distribution was too
wide for a stable foam.

Three and four glass gas dispersing tubes (Corning No. 39533)
with fritted cylinders of extra coarse porosity were mounted
in specially constructed manifolds and adapted to the 6 in.-
dia. column for use as sparging heads (Figure 5).  The tube
stems were shortened from 250 to 57 mm.   These heads showed
very little channeling and produced a stable foam.
FIGURE 5.  Sparging Head.  The head is made from four pyrex gas
dispersing tubes with extra coarse fritted cylinders.  The
diameter of the fritted cylinder is 2 mm.  The support struc-
ture and a column adapter coupling is made of PVC and poly-
ethylene plastic.
                           16

-------
A special holder was fabricated from PVC plastic to hold
2 in.-dia. discs of scintered stainless steel  (Pall Corp.
Type 316L).  Six porosity grades ranging from 5 to 165 u
mean pore size were tested and found to be less uniform
and to channel more than the fritted glass discs in the
Buchner funnels.  The foam produced was very unstable.

A sparging head was fabricated from stainless steel to hold
four cup-shaped spinnerettes which contained precision
drilled holes on their upper faces.  The spinnerettes were
designed for previous unrelated research but were tested to
determine if holes of one known size and spacing would pro-
duce a uniform, narrow bubble size distribution.  The four
spinnerettes had 60 holes (15 each) of 125 u-dia.  The
bubbles formed were very uniform and the foam was stable.
However, the sparging air capacity was too small for use
in the 6 in.-dia. column.  Quotes for fabrication of a
larger sparging head of sufficient air capacity were approxi-
mately 1.5 dollars per hole.  Since at least 2000 holes of
80 u-dia. would be necessary, the cost of 3000 dollars was
considered excessive at the present stage of research.

The same sparging head used to hold the spinnerettes was
also used to hold discs of collimated hole structure
(Brunswick Corp.).  One disc contained 5,500 holes of
125 u-dia. and the other contained 335,000 holes of 12.5 u-dia;
both discs were tested as currently available without special
fabrication.  Both discs produced streams of bubbles widely
varying in size, not because of channeling but due to bubble
coalescence in the swarm.  The hole spacing  (length between
hole centers) is less than three hole radii apart; this
allows for bubble coalescence and the resulting wide bubble
size distribution and unstable foam.  Collimated hole struc-
ture can be made in a wide variety of hole sizes and spacings
and may be a less expensive alternative for drilled spin-
nerettes.

The results of the test indicate that the sparging head
constructed of four dispersing tubes (Figure 5) is adequate
for producing a reasonably stable foam; it was used in all
subsequent tests.  Its low cost is an attractive feature.

Feed Distributors - Operation of the foam column in the
countercurrent mode requires that descending liquid feed be
introduced into ascending foam without causing significant
bubble rupture or liquid channeling.  The feed stream must,
therefore, impinge the foam stream with minimum momentum.
Uniform distribution of the feed stream across the column
diameter is desired to  provide liquid-foam contact over
a minimum downflow length.  To meet these requirements,
special liquid distributors were constructed and tested.
                           17

-------
Previous work with countercurrent foam separation has lead
to liquid distributors of various designs.  Spaced, packed
columns have been used to distribute feed streams  (9).
Distributors constructed of single tubes, multiple tubes
and screen mesh in various configurations have been used
(3) (21).  In general, the multiple tube distributors are
reported to have operated most effectively.

Three multiple tube distributors were designed, constructed,
and tested with the 6 in.-ID column.  Based_on the expected
liquid feed rates (approximately 500 ml-min 1), the number
of distributor tubes of various sizes necessary to provide
the required pressure drop, was calculated such that the
feed liquid would impinge the foam at near ambient pressure.
This would provide minimum momentum transfer to the foam
and, thus, reduce foam collapse.  The number of tubes should
be sufficient to provide effective feed distribution over
the column cross-section, but not so numerous that their
cumulative surface area exposed to the foam would be large
enough to disrupt foam flow.  Based on these considerations,
the three distributors had 8, 12 and 21 tubes each of
0.062 in.-ID, 0.045 in.-ID and 0.030 in.-ID tubing,
respectively.

The distributors are constructed entirely of polyethylene
tubing and silicon rubber potting material.  The smaller
distributing tubes are potted directly into 0.25 in.-ID
tubing which connects directly to the feed stream tubing.
The silicon rubber provided an economical and durable
method for distributor construction.

While all three distributors operated satisfactorily, the
8-tube distributor caused some channeling, and the 21-tube
distributor appeared to disrupt the foam flow more than the
others.  The 12-tube distributor provided adequate distri-
bution with minimal channeling and foam collapse.  The
12-tube distributor was used in all subsequent countercurrent
experiments.

-------
                       SECTION V

                   ANALYTICAL METHODS
Total Fe, Ca, Mg, Al and Mn - These metals in AMD are deter-
mined by atomic absorption spectrophotometry using standard
comparator methods  (31).  The samples are aspirated directly
into the flame with dilution necessary only to bring the
samples into the concentratiob range of the particular
resonance line.  A high solids burner head (Perkin-Elmer
three slot) is used and no interferences from the surfactant
present has resulted.

Surfactant - Most of the surfactants used to date have been
anionic and possess a sulfonate or sulfate hydrophylic
group.  Such surfactants form strongly colored complexes
with the triphenylmethane dyes which thus provide the basis
for a spectrophotometric method of surfactant determination
(32).  In this case, the dye crystal violet is complexed with
the surfactant, the colored complex is extracted by benzene
and the absorbance of the benzene layer determined at 615 mu.

Ferrous Iron - Since the Fe II/Fe III ratio present in mine
water can theoretically have an effect on the foam separa-
tion of iron, determination of this variable was considered
important.  Fe(II) is determined spectrophotometrically (33),
and Fe(III) is determined by difference from total iron de-
termined by atomic absorption.  Fe(II) in aqueous solution
is complexed by 4,7-diphenyl-l-10-phenanthroline at pH 4,
the colored complex is extracted by n-hexyl alcohol and the
absorbance of the alcohol layer measured at 533 mu.

pH - The surfactants under test are weak acids and, thus,
their solubility and speciation in solution is a function
of pH.  pH is monitored by a glass electrode - calomel
reference electrode pair and a pH meter.

Bubble Diameter Measurement - As will be seen in Section VI,
determination of the average bubble diameter in the foam is
necessary for calculation of surface excess.  Bubble dia-
meters can be determined by direct or indirect methods.  An
example of an indirect method is the empirical relationship
of optical density to bubble diameter.  A direct measurement
would be bubble photographs.  It is known that photographs
of a foam taken through a glass surface can yield useful
measurements of bubble diameters (34).  The bubble diameters
at the periphery of the foam column are not significantly
different from those in the interior.  Thus a method of ob-
taining photographs of the bubbles was developed.  A 35 mm
                           19

-------
single-lens reflex camera with a bellows unit and 55 mm lens
was used to obtain 1-2 x photographs of the moving foam.  A
strob-flash provided the light.

Bubble diameters are measured directly from the negative
with an optical comparator which magnifies the bubbles and
allows visual comparison of the bubble to a graduated
reticle.  The measured diameters, d. are used to calculate
a mean  diameter,

                   3 -
                           /^
                          i i
where n. is the_number of bubbles of .various diameters d..
Calculation of d was by computer.
                           20

-------
                      SECTION VI

       SAMD EXPERIMENTS  - RESULTS AND DISCUSSION
Surfactant Screening  - A great number of surfactants are
commercially available from an equally great number of
companies.  Surfactants are commonly sold under trade
instead of generic names because their chemical composi-
tions are considered  proprietary.  In addition, test in-
formation on the chemical and physical properties of sur-
factants is often fragmentary and/or inconsistent or
noncomparable from one manufacturer to another.  Therefore,
the screening of surfactants for those suitable to foam
separation under the  conditions encountered in acid mine
drainage presents a formidable task.

Since the metal ions  in acid drainage are not surface
active, a surfactant  must be added to allow coadsorption
of metal.  Coads-orption of metal may arise from charge
interactions between  adsorbed surfactant and a diffuse
double layer of metal counterions, or from bonded inter-
actions between surfactant and metallic species.  Both
methods for obtaining oeadsorption have been considered.

Since the major metallic constituents of acid mine drainage,
i.e., Fe, Al, Ca, Mg  and Mn, occur as cations at low pH, an
anionic surfactant would be required for coadsorption by
the charge interaction mechanism.  Thus, of the three charge
types, i.e., nonionic, cationic, and anionic, anionic sur-
factant has received  greatest attention.

Metal-ligand complexes of high bond strength and metal speci-
ficity are formed by  a class of multidentate ligands known
as chelates.  Ethylenediamine tetraacetic acid (EDTA) and
diethylene triamine pentaacetic acid (DTPA) (and their
associated salts) are well known, examples of ligands which
form water soluble, multidentate complexes (chelates) with
a variety of metal ions.  These chelates are characterized
by exceptional complex stability and metal specificity; how-
ever they have minimal surface activity and are not directly
useful in foam separation.  The chelating ability of EDTA
and the surfactant properties of the dodecyl-benzyl func-
tional group are combined in dodecylbenzyl-diethylene triamine
tetraacetic acid (DBDTTA) to form a water soluble, chelating
surfactant.  Another  chelating surfactant is dodecylimino
diacetic acid (DIDAA).

To screen the large number of available surfactants by
laboratory experiments would be very costly in time.  Thus,
                           21

-------
an extensive literature search was considered the most prac-
tical method for picking the few best surfactants for labora-
tory testing.  Research and commercial data was consulted;
in addition, surfactant manufacturers and suppliers were
consulted and asked to consider surfactant requirements and
recommend suitable materials.  By these methods surfactants
were chosen for testing.

Anionic type surfactants were chosen which produce stable
foams which have sufficient solubility at_low pH and in
hard water to resist precipitate and/or micelle formation
and which are low cost and readily available.  Unfortunately,
chelating surfactants are not commercially available and can
be obtained only in research quantities at considerable cost.
Thus, the testing ©f chelating surfactants has not been further
considered.

Surfactant Tests - The potential for effective foam separa-
tion can be expressed by the surface excess concentration
gamma P (equation 1, page 7) for surfactant and coadsor-
bants, in general the greater the P values, the greater the
potential.

The equilibrium surface excess of a solute is usually deter-
mined by measuring surface tension as a function of solute
concentration.  Solute concentration is kept low so that
the activity coefficient can be assumed equal to one and,
therefore, the chemical activity is assumed the same as
concentration.   Rvalues measured in this manner are not
representative of dynamic foam separation systems (35) and,
thus, have little value.  Instead, useful equilibrium values
can be determined for surfactant and coadsorbed metals
under dynamic conditions in a foam column designed for
that purpose.  The equilibrium column (Figure 3) was de-
signed to allow measurement of  P values.  The column
operates in the continuous, simple mode.

By mass balance and surface equilibrium considerations
around the foam, the system can be described by
                           22

-------
                             6GP
where  [x]  is  the  concentration  of  surfactant or  coadsorbed
           component

      LC   is  the  liquid  volume  rate  of  foam

      G    is  the  air  volume sparging rate

      d    is  the  average bubble diameter
      P
       x   ls  the  surface excess in equilibrium with  [x],


and the subscripts c  and b refer to  the foam exit and pool
exit flows, respectively.

Rearrangement of  equation  (2) gives  an expression for I



           P    ,r  i   r  , ,5 LC
           I  =  (l-xl -f-vl ^	—                       (1}
           •v    ^!-Jr«LjAJv1'crl                        v«J /
           X       CD  oLr
Thus, by measuring  [x],  Wfcj ^» Lc and G' values of '
can be calculated for x equal to surfactant, Fe, Ca, A'l,
Mn and Mg.  The basic requirement is that sufficient bubble
liquid contact time be provided to assure attainment of sur-
face adsorptive and coadsorptive equilibrium and that foam
coalescence is minimal.   Such equilibria appear to have
rapid kinetics (36), and  approximately one foot of pool
depth gives sufficient bubble contact time (29).

Five surfactants, determined in the initial screening to
have properties suitable  for foam separation at low pH,
were selected for testing.  The tests were carried out in
the equilibrium column to determine (values for surfactant
and metal constituents.   Each of the surfactants were tested
at four feed concentrations ranging from 100-400 ppm in
standard SAMD.

The surfactants tested were sodium dodecyl sulfate, Alkanol
189-S (DuPont), Alipal EO-526 (GAF), Alipal CD-128  (GAF)
and Aerosol AY (Cyanamid).  Alkanol 189-S is a sodium alkyl
sulfonate with an average carbon chain length of 15.  Alipal
EO-526 is a sodium salt of a sulfated alkylphenoxy poly(ethyl
eneoxy)  ethanol with an undisclosed chain length.  Alipal
                           23

-------
CD-128 is a sulfate-ester type surfactant based on a linear
chain (hydrophobe) of undisclosed length.  Aerosol AY is
sodium diamyl sulfosuccinate.

Upon addition of Alkanol 189-S to SAMD, a coarse reddish
precipitate formed.  Alkanol189-S is normally a high
foaming, acid stable surfactant, but mixed with SAMD it
produced only a transient froth.  A sample of the precipi-
tate was dried at 100°C, producing a viscous substance
obviously containing a large amount of surfactant.  The
surfactant is incorporated in the precipitate and the
precipitate is not solely the result of oxidation of Fe(II)
and hydrolysis of Fe(III) due to pH increase.

Aerosol AY would not produce a stable foam in SAMD over a
concentration range of 100-500 ppm surfactant.

In the tests using sodium dodecyl sulfate, Alipal EO-526,
and Alipal CD-128, stable foams were produced and the data
necessary for \  calculations were obtained.

For purposes of surfactant comparison, individual |  values
for surfactant and SAMD metals were averaged for each sur-
factant and surfactant concentration; the results are in
Table 2.

                        TABLE 2

          Dynamic Surface Excess Values for
            Selected Surfactants in SAMD
        Equilibrium Column,  G = 10 liters-min
        Lf = 700 ml-min 1,  [S]f = 100-400 ppm
                                             _i
r r r
1 s ' Fe ' Ca
Surfactant, S (Units: x
Sodium dodecyl
sulfate, NaDS
(M.W. = 288)
Alipal EO-526
(M.W. = 586)
Alipal CD-128
(M.W. = 352)
Aerosol AY
(M.W. = 360)
Alkanol-189S
(M.W. = 354)
26.1 5.22 5.11
27.7 3.22 1.92
48 2.5 1.5
r r r r /r
1 Mn ' Mg ' tm ' tm/ ' s
1011 moles -cm 2)
0.131 0.768 11.2 0.430
0.1 0.166 5.40 0.195
0.14 0.99 5.09 0.106
No foam produced over surfactant concentration
range 100-500 ppm
Precipitate formed;
produced
only transient foam
                           24

-------
The economics of foam separation are greatly dependent on
the magnitudes of the adsorptive J""1 for surfactant and the
coadsorptive \  for the metals to be separated.  Increased
adsorption of surfactant to bubble walls increases the
capacity to separate coadsorbed metals and thus promotes
the most efficient use of surfactant.  Because foam
separation is inherently of low exchange capacity compared
to ion exchange or solvent extraction  (1), the economics
of foam separation will depend on maximizing exchange
capacity (i.e., maximizing coadsorptive) per consumption
of materials and power (i.e., surfactant and sparging air).

Generally a stoichiometric relationship is believed to
exist for foam separation of dissolved constituents, but
the exact relationship has not been clearly defined.  In
foams produced from pure aqueous solutions of ionic sur-
factants, the surface concentrations of adsorbed surfac-
tant and coadsorbed metal will be equal (1), e.g., for
NaDS in deionized water  PN + =  f^q- and tne ratio
•Na+/ IDS" = 1"  Considering metal coadsorption to arise
from charge interactions with adsorbed surfactant, and
considering the requirement of electroneutrality for the
system, a one-to-one ratio of surfactant and metal in the
foam would be expected on an equivalent basis.  Ratios less
than one indicate coadsorptive competition from metal im-
purities.

At a pH of 2.2, the hydrogen ion concentration in SAMD is
6.3 x 10~3M compared to about 7 x 10~3M total added metal.
Sodium in SAMD derived from surfactant and tap water
(approximately 13 ppm in tap water) is at least equal molar
with surfactant (sodium analyses were not conducted).  Iron
in SAMD is initially greater than 98 percent Fe(II); the
extent of oxidation to Fe(III) during sparging and, thus
the exact equivalence of iron in the foam system is unknown.
For these reasons, equivalence ratios of coadsorbed metal
to adsorbed surfactant are not determined.  Molar ratios
are used to provide information on the relative effective-
ness of a particular surfactant for foam separation in
standard SAMD.

An optimum value for adsorptive \  , the value to which ex-
perimental values for surfactants can be compared, can be
estimated at least as to order of magnitude (1).  A (value
of about 3 x 10~10 moles-cm"  would be expected under
equilibrium conditions for the surfactant sodium dodecyl
sulfate (M.W. = 288).  Comparing the average \s values in
Table 2 with this theoretical value, it is apparent that
adsorptive equilibrium was reached for NaDS, EO-526 and
CD-128.  Considering the individual surface excess values
for each metal |  , and then the sum of the individual values
               1 m'
                           25

-------
P. ?it is clear that NaDS with a molar ratio of total
coadsorbed metal to adsorbed surfactant P   /p  equal
to 0.430 is the most efficient surfactant Tor metal
separation.

Effect of pH - Experiments were conducted at pH 2.4, 2.7
and 2.9 to determine the effect of hydrogen ion concen-
tration on foam separation of SAMD metals with NaDS.  pH
was adjusted from 2.2 withNaOH, the upper limit being
restricted by iron precipitation at about pH 3.0.  The re-
sults in Table 3 indicate a slight decrease in P  and a
more significant decrease in P,  with increasing pH.  The
decrease in R  and ft
coadsorption 01 iron.
                              ,                     .
                            results mostly from decreased
                         TABLE 3
       Effect of pH on Metal Separation from SAMD

         Equilibrium Column, G = 10 liters-min"1
         Lf = 700 ml-min"1, [S]f = 100 ppm
     PH
                                          r.
                                           mt
                                                 _2
                          (Units:  x 101  moles-cm" )
                                                      r /r
                                                      1 mt/' s
NaDS and SAMD
2.2
2.4
2.7
2.9
EO-526 and SAMD
without iron
2.8
3.5
4.4

2.61
2.40
2.22
2.34

1.56
1.52
1.41

1.12
1.08
0.761
0.542

0.303
0.212
0.202

0.430
0.450
0.343
0.232

0.194
0.139
0.143
Three additional experiments were conducted over a wider pH
range at 2.8, 3.5 and 4.4 with EO-526 and SAMD without iron.
The results in Table 3 also indicate a slight decrease in
   and a more significant decrease in R .   The relative
            ~
decrease of

                                         .
                with pH is greater with iron present.
                           26

-------
Effect of the Chelating Agent EDTA - A number of experiments
were conducted to determine the effect of the chelating
agent EDTA on the coadsorption of metals in SAID with both
anionic and cationic surfactants.  EDTA alone will not pro-
duce a foam in SAMD, but in combination with a surfactant
it is possible that its ability to form strong, selective,
ionic metal complexes coupled with its tendency to associate
with and stabilize foams could enhance selective foam
separation of metals.  The stability of EDTA complexes with
SAMD metals is:  Al=» Fe(II) =- Mn=~ Ca=- Mg.  In general, ex-
cept for the alkaline earths, the metal chelates are stable
to a pH of near 1.0.  Decreasing pH  favors formation of
less highly ionized, less strongly chelated complexes.  The
effect of increasing pH is to stabilize the completely
ionized chelatent species, but at high pH this advantage
is balanced and eventually nullified by metal hydroxide
formation.

Three experiments using NaDS and SAMD at pH 2.3 were con-
ducted with EDTA to NaDS molar ratios of 0.59, 1.02 and 1.62;
[NaDS]f was 4.38 x 10~ M.  P  for the three experiments was
       10 M ; for \tm it was 3.51 x  10"11 which gives a
       ratio of 0.193.  Comparing these values to those in
TalSle 2 for similar conditions except without EDTA, it is
apparent that adsorption of surfactant and coadsorption of
metal is reduced by EDTA.

Six additional experiments were conducted to test higher con-
centrations of EDTA with both anionic NaDS and cationic
Hyamine-1622 surfactants.  Hyamine-1622  (Rohm and Haas) is
di-isobutyl phenoxy ethoxy ethyl dimethyl benzyl ammonium
chloride, monohydrate.  These experiments were conducted
under conditions similar to those above except with an EDTA
to total SAMD metal ratio of 1 and pH at 2.9 - 3.1, 6.0 -
6.2 and 7.8 - 7.9.  In three experiments [NaDS]f was
6.89 x 10~4M; in the other three [Hyamine-1622]f was
7.22 x 10~4M.

At pH 7.8 - 7.9 Hyamine-1622 and SAMD produced a dark red-
brown turbid solution containing a grey-black precipitate;
at pH 6.0 - 6.2 the solution was yellow-green and turbid;
at pH 2.9 - 3.1 the solution was green and clear.  In all
three pH ranges foam production and stability was marginal;
and no measurable amounts of metal were removed.

With NaDS and EDTA, metal separation at all three pH_ranges
was lower than without EDTA as indicated by r^/Ps ratios
of less than 0.190.  The results of these tests indicate
no enhancement of metal coadsorption with EDTA compared to
surfactant alone.  Hyamine-1622 partially precipitated with
                           27

-------
 SAMD  and EDTA,  thus results  from  its  tests were  questionable.
 The possibility of enhanced  separation  and metal specificity
 with  chelating  surfactants remains  to^be  examined.

 Effect  of  Surfactant Concentration  -  Results  of  testing
 several surfactants in SAMD  at various  pH values with and
 without EDTA  indicate that NaDS without EDTA  at  the  2.2 -
 2.3 pH  of  standard SAMD are  the most  efficient conditions
 tested.  Therefore, NaDS was used in  all  subsequent  ex-
 periments .

 The I    values for a surfactant will increase  with surfactant
 concentration until complete saturation of the air/solution
 interface  is  reached.  Further increase of surfactant con-
 centration will not increase [""" .  As  surfactant  concentration
 increases  the critical micelle concentration  (CMC) is
 reached, above  which additional surfactant will  form micelles
 which are  bulk  rather than surface  active.  Above the CMC,
 the "  for surfactant will remain constant while the | ,
 for coadsorbed metal will decrease  due  to coadsorption com-
 petition between surfactant  adsorbed  at bubble interfaces
 and surfactant  present as micelles.   The  maximum surfactant
 concentration for efficient  coadsorptive  separation  will  be
 determined by the CMC and/or the  concentration at which
 surface saturation is maximum (theoretically  these two
 concentrations  are the same).  The  lower  limit of surfactant
 concentration is usually that necessary to produce a stable
 foam.   Thus,  an optimum surfactant  concentration or  concen-
 tration range will exist at which maximum coadsorptive
 separation of metal will be  achieved.

 SAMD  was foamed in a series of experiments with  NaDS over a
 surfactant feed concentration range of  180-300 ppm.   The
 column  was operated in the simple continuous  mode with an
 extended horizontal drainage section.   A  second  series of
 experiments were conducted under  similar  conditions  but
 with  a  more dilute mine drainage  DSAMD, the composition of
 which is given  in Table 4.  For reasons of analytical
 sensitivity,  the metal concentrations are not in the same
 relative proportions as in SAMD.  The results are shown
 in Figure  6.

The distribution coefficients for both SAMD and DSAMD,
which differ by a factor of approximately 200 in  total
molar metal concentration,  indicate the optimum  [NaDS],
is in the range 4.7 to 6.1 x 10  M.
                           28

-------
                         TABLE 4

                       Dilute SAMD
     Constituent            Metal Concentration, M x 106
     FeS04-7H20

     CaSO4-2H20

     MgSO4-7H20

     MnSO4-H20

     A12(S04)3-18H20
 2.24

11.5

 1.42

 0.574
20.4
Effect of Countercurrent (Stripping Mode) Column Operation -
Two series of experiments were conducted with SAMD under
conditions similar to those reported in Figure 6 except
that the column was operated in the stripping instead of
the simple mode.  Two countercurrent lengths were tested
to determine the effect on separation of increased contact
time between counterflowing feed and foam.  [NaDS]f was
set such that [NaDS],  would be in the optimum range
(Figure 6).

For comparison, results of the stripping experiments are
shown in Table 5 with those from an analogous experiment
conducted in the simple mode.  Extension of the countercur-
rent length from 15 to 29 inches produced an absolute in-
crease in both surfactant adsorption and metal coadsorption;
however, the relative metal coadsorption per unit surfac-
tant adsorbed remained the same.  Comparison of \\.j/\s
and VR values for stripping and simple operation suggests
the stripping mode to be less efficient at metal separation.
Greater liquid carry-over with countercurrent operation
would be expected considering the higher adsorbed and inter-
stitial liquid surfactant concentrations compared to simple
operation.  Extension of the horizontal drainage section
would increase VR to about 20 without significant reduction
in separation due to foam collapse.
                           29

-------
 ID'1-,
s
u
10
  -3.
            o  SAMD VR = 11-20  	>
             •  Dilute SAMD VR = 15-23
         .-"r
         ,AT
    4.5
5.0
7.0
                                                 10
                                                                -3
                                                -10
                                                   o
                                                 10
                                                              10'
7.5
                             6.0
                   [NaDS] b, ¥ (xlO 4)
              FIGURE 6  DISTRIBUTION  RATIOS
Simple Mode, 18 in. Vertical Section, 72 in.  Horizontal Section,
G = 8 liter •  min '1,  Lf = 500 ml-  min -1  .  Each data  point
represents the average  of 4 to 5  measurements.
                               30

-------
                         TABLE 5
  Comparison of Simple and Stripper Mode Experiments
  O A TUTT-k — « *3 XTn T"\O  /~* 	 O 1-34-.n..»»__-!u ^~  T  	 rr f\/\	T 	! 	 ~1
 (SAMD and NaDS, G = 8 liter-min
,  Lf  = 500  ml-min   )
Variable
Column Mode
Vert. Sect. , in.
Horz. Sect. , in.
C.C. Length, in.
VR
FRT, min.
, , moles • cm
tm'
1 , moles -cm
s '
^n/Q
Stripper,.
24 in.
72 in.
15 in.
13.6
6.02 min.
9.06 x 10""11
2.30 x 10"1
0.394
Stripper
36 in.
72 in.
29 in.
14.2
5.94 min.
10.3 x 10'11
2.70 x 10'1
0.381
Simple
18 in.
72 in.
-
19.0
5.99 min.
11.1 x 10'11
2.22 x 10"1
0.500
 ' Because adsorptive  phenomena  are  in  equilibrium with  feed,
 liquid in  the  stripping  mode  rather  than  pool  liquid  as  in
 the simple mode,  the  proper form  of  equation 3 for
 calculation  in the  stripper mode  is  (29) :
               r  =
                                3 L
                x
                             f'  6.59 G
In  the simple  mode,  adsorptive phenomena are in equilibrium
with  pool  liquid;  in the stripping mode equilibrium is with
feed  liquid.   Since  surfactant optimization tests (Figure 6)
were  conducted in  the simple mode, and since [NaDS]f  [NaDSJb,
the surfactant concentration in the stripping tests was
above optimum  for  maximum coadsorption of metal.  A reduction
of  [NaDSK from 8.7  - 8.9 x 10~4M to 4.7 x 6.1 x 10  M would
increase   fL/P  to  about °'5 and thus increase separation
by  about 3 to  21 percent.  The small relative magnitude of
this  increase  with respect to metal separation from SAMD
precluded  allocation of time for further tests.
                            31

-------
Volume Reduction - Production of dry foams is required to
reduce interstitial liquid carry-over in the foam and, thus,
to reduce the waste volume for a foam column operating to
decontaminate the feed.  Volume reduction factors (Lf/Lc)
were determined as a function of sparging rate G and sur-
factant concentration  (Figure 7).  Volume reduction factors
equivalent to approximately 1.7 to 3.8 percent liquid carry-
Over are easily achieved at low gas rates.  Further increase
in volume reduction factors may be possible through increase
of foam residence time FRT which is accomplished by reducing
sparging rate and/or by lengthening the drainage section.

The practical factors which operate to limit volume reduction
in a 6 in.-dia. column are the sparging heads at high gas
rates and foam collapse at low gas rates, i.e., long FRT.
At high gas rates, bubble diameters increase and their size
distribution widens.  Thus surface throughput per volume of
gas and foam stability declines markedly.  At low gas rates
a maximum FRT is reached above which further drainage to
reduce interstitial liquid results in foam collapse and re-
duction in separation efficiency.  With protracted drainage
the foam becomes dry and its suceptibility to rupture in-
creases until mechanical shock from flow in contact with the
column wall produces complete collapse.

The inverse log relationship of foam density and drainage
time (1) and the practical limits for foam drainage in the
6 in. column are shown in Figure 8.  For [NaDS]f = 200 ppm
the foam density decreases with drainage for FRT values
approximately three and greater; this section of the curve
follows the expected exponential relationship.  At low FRT
values, which result experimentally from high sparging
rates, the average bubble diameter is increased and the
diameter size distribution is widened. The result is more
rapid drainage and a significant reduction in foam stability.
Also, at high gas rates turbulence in the pool causes voids
to form in the foam at the foam-liquid interface.  These
voids progressively enlarge and destroy plug flow in the
drainage section.  The low foam densities at low FRT values
(Figure 8) are caused by extensive foam collapse which, of
coarse, destroys separation efficiency.  The destructive
effect of high sparging rates can be seen in Figure 7 where
the curve of decreasing VR with increasing gas rate changes
shape above 10 liters-min

These data indicate the operational restrictions for the 6 in,
dia. column.  A minimum surfactant concentration is required
for foam production; under present conditions the minimum
can be expressed as 100 ppm-= [NaDS] f*= 200 ppm.  Gas rates of
approximately 10 liters-min"1 or less are required with the
EC fritted glass sparging heads.  With these restriction
foam densities in the range of 2 - 5 x 10 3 ml liquid per ml
                           32

-------
    70

    60

    50


    40


    30




    20
    10-
     9-
     8

     7-

     6-

     5-


     4-
D
                             [NaDS]
                                    f
                               o 100 ppm
                               D 200 ppm
                               a 300 ppm
                               A 400 ppm
                        10        15

                             G, liter • min -1
                             20
25
30
                               FIGURE 7
                 Volume Reduction versus Sparging Rate
(Simple  Mode, 18 in. Vertical  Section, 36 in. and  72  in. Horizontal Sections,
 Lf = 500  ml • min"1)
                                 33

-------
   10-
   9
   8-
/-^
no 7-
i—i
- 6
 6
 o 5
 g 4-
    3-
 •H
 S  2-
 0
 p

 Cj
[KADS]f
                                               100 ppm
                                               200 ppm
                                               300 ppm
                                               400 ppm
                                                   8
                    10
                               FRT, min
                              FIGURE 8
              Foam Density versus Foam Residence Time
                               34

-------
foam can be easily obtained; lower values could result from
further development of drainage sections.  However, since
drainage rate and extraction rate are opposite functions
of bubble diameter, a compromise must be made between
separation efficienty and foam density.  For this study
the controlling factor is the narrow range of sparging
rates over which a stable foam can be produced, the
sparging head determining the upper limit and a minimum
gas flow to produce a stable -foam determining the lower
limit.

Effect of Metal Concentration on Coadsorptive Separation -
In the presence of dissolved electrolytes, the surface
tension and critical micelle concentration of ionic sur-
factants are reduced.  Ions of charge opposite to the
surface active species tend to reduce repulsion between
surfactant ions and allow closer packing of surfactant
in the surface film.  Thus, the presence of cations in
solutions of anionic surfactants, e.g., SAMD and NaDS,
should tend to increase j   values.  Experimental data
showing this effect is plotted in Figure 9;  [~*  values tend
to be greater in SAMD than in DSAMD.         S

As shown in Figure 6, the decline in  |, /[M],  with in-
creasing [NaDS],  indicates that the CMC is reached at [NaDS],
greater than approximately 6x10  M which is equivalent to
[NaDS]f greater than 7.5xlO~4M under the operating condition
shown.  The reduction in CMC by dissolved salts is evident by
comparing these values to 7x10  M, the CMC for NaDS in pure
aqueous solution (4) .

Literature data for surface excess of NaDS in pure aqueous
solution as measured by the foaming method in included in
Figure 9 (4) (37) .  This data suggests that l      in SAMD
                         "1          "2
would be larger than SxlO"   moles -cm"  and possibly
larger than 4X10"1  moles -cm   .  Our data (Figures 6^ and
9) show that practical PNaDS values equal to 2-3x10"
moles • cm"2 are the equilibrium values for conditions in
the 6 in. column.  As indicated above (Figure 6), the
foam column is most efficiently operated with a surfactant
concentration below the CMC and with the concentration
optimized for metal coadsorption under existing conditions.

Coadsorptive P values for total metals show the characteris
tic equilibrium relationship with increasing metal concen-
tration in the pool (Figure 9) .  The total metal \  values
increase by a factor of approximately 10 while total metal
concentration in the pool increases by a factor of 700; sur
factant P values increase by a factor of 3 or less over
the same increase in metal concentration.
                           35

-------
                                                                            x = NaDS in Dist. H?0 (37)
                                 x = NaDS in SAMD
9  3'
                                                                          x = NaDS in H20 (4)
Gi
                        S
                        o
                        •
                        01
                        0)
                        rH
                        O
                        E
                           1-
                                                      x = Mt  SAMD
                                                     -H_g_JL
                                                                              7       8       9
                                                        [X]b,  M (xlO3)
                                                            FIGURE 9
                                         Dynamic Surface Excess versus Pool Concentration
                                       (For  x = NaDS in  SAMD: • is Dilute SAMD, • is SAMD]

-------
Optimum Operating Conditions for the 6 in.-Dia. Column -
The metal separation rate increases with higher metal feed
rates  (Figure 10) as would be expected from P.  data
                                             "tin
(Figure 9); the separation rate increases by a factor less
than 5 while the metal feed rate increases by a factor of
about 200.  Optimum surfactant feed concentrations were
found to have a narrow range over the wide range of total
metal concentrations in SAMD and dilute SAMD (Figure 6).
Thus, the range of surfactant feed rates (approximately
3.5 - 4,5x10 4 moles-min  ) was considerably narrower
than the metal feed rates.

Date in Figure 10, which includes all simple and stripping
experiments with standard and dilute SAMD, shows that the
metal separation rates do not significantly differ as a'
function of operating mode and surfactant type (CD-128 and
EO-526 experiment are included).  The fraction of total
metal removed depends strongly on the metal feed rate.
For standard SAMD, an average metal separation rate is
about 3.5xlO~5 moles-min"  which is equivalent to 1.93x10"
moles-cm 2 of column_cross section area per minute or
approximately 4.0x10   equivalent cm 2-min  .  The corres-
ponding experimental rate for surfactant is 4.8x10"
equivalent cm" -min" .  The difference between total metal
and surfactant capacities represents unaccounted metal
coadsorption such as from sodium derived from surfactant
and metal impurities.

Based on theoretical calculations (1) and empirical measure-
ments  (4) the maximum exchange capacity for a foam separa-   _
tion system would be approximately 3x10"  equivalent cm" -min
based on an optimum bubble diameter of 0.08 cm, a maximum
gas rate of 9.4 cm3-min~1 per cm  of column cross section
and a packing density of 40 sq-8 per NaDS molecule on the
bubble surface.  Thi| theoretical maximum sparging rate is
equivalent to 1.7x10  cm -rain"1 for the 6 in.-dia. column
and is 2.1 times larger than the practical rate of 8x10
cm3-min   which was restricted by foam instability at high
rates using the EC fritted glass sparging heads.  The
average bubble diameter obtained in this study at gas rates
of 8xl03 cm -min   was 0.13 cm and would reduce the surface
area throughput by a factor of 1.6.

The 6 in.-dia. column operates at a practical capacity of
16 percent of theoretical.  Improvement in sparging head
design, possibly the use of precision drilled spinnerettes,
or change in the method of sparging, such as bubble
                           37

-------
OJ
00
                  1.0



                   .9


                   .8-



                   .7


                   .6
.4


.3


.2



•H
                     10-5
                                               L  [M,]f,  moles-min
10
                                                                    -i
                                                                      -3
                         10



                          9


                          8



                          7


                          6



                          5


                          4


                          3



                          2
                                                                        10
                                                                          -2
                                                                                                       c
                                                                                                       •H
                                                                                                       e

                                                                                                       en
                                                                                                       0)
                                                                                                       rH
                                                                                                       O
                                                                                                       e
                                                                                                       o
                                                      FIGURE 10
               Total Metal Fraction Removed and Total  Metal  Extraction Rate Versus Total Metal Feed Rate.

-------
formation by  degasification of supersaturated solutions,
may allow closer approach to optimum gas rates and bubble
diameters.  Allowing these improvements, the capacity
would increase to 54 percent of theoretical.

Derived from the experimental results with SAMD, the opti-
mum conditions for practical operation of the 6 in.-dia.
column demonstrate the effectiveness of continuous foam
separation for treating acid mine drainage  (Table 6).
Average values for VR, [NaDS],, [M,]f, SR  and SR,  de-
termined within the practical operating values for G and
L- in the 6 in.-dia. column were used to calculate the
average material balances for all SAMD experiments.
Approximately 6 percent metal removal with 5 percent
liquid carryover indicates that 1 percent metal is separa-
ted as coadsorbed metal.  Twenty-six percent of the added
surfactant is removed in the foam, 23 percent is removed
as adsorbed material and 74 percent remains in the bottoms
stream.
                        TABLE 6

        Summary of Results  for SAMD Experiments
          and  Comparison with AMD Experiments
                   SAMD  and  NaDS
AMD and NaDS
  (3 runs)
Column Mode
G, liter -min
L.., ml-min
VR
[NaDS],, _i
moles -liter
moles -liter"1
SR , moles -min"
s '
SRtm» moles"min~
[NaDS]f, _i
moles- liter
Simple and
Stripping
8
500
20
G.OxlO"4
7.0xlO~3
8.8xlO"5
3.5xlO"5
7.7xlO~4
Simple
4
250
27(11)
4
5.4x10
7.9xlO~3
4.3xlO~5
1.7xlO~5
7.5xlO~4
Simple
8
500
13
4.8xlO~4
7.5xlO~3
7.3xlO~5
4.3xlO~5
7.1xlO~4
Stripping
8
500
23
5.3xlO~4
7.9xlO~3
2.8xlO~5
1.8xlO~5
7.4xlO~4
                            39

-------
TABLE 6 (continued)
SAMD and NaDS
AMD and NaDS
  (3 runs)
[NaDS]c,
moles • liter"
[Mt]b>
moles - liter
[Mt]c>
moles • liter
„ , . -i
Rf _ , moles -mm
R, , moles -min"
R , moles -min"
cs '
Rftm' m°les'min
R, . , moles -min"
R , , moles -min
Material Balance
Metal, % diff.
H20, % diff.
NaDS, % diff.
SRtm/SRs
SRs/Rfs.102
Rbs/Rfs-1()2
Rcs/Rfs-lo2
R , /R». -102
c tm i tm
SR, /R_,
tm f tm
4.1xlO~3
6.9xlO~3
8.3xlO~3
3.8xlO~4
2.8xlO~4
l.OxlO"4
3.5xlO~3
3.3xlO~3
2.0xlO~4

-
-
-
0.40
23%
74%
26%
6.0%
1.0%
5.2xlO~3
7.7xlO~3
9.6xlO~3
1.9xlO~4
1.2xlO~4
5.2xlO~5
2.0xlO~3
1.7x10"
8.9xlO~5

7
5
9
0.40
23%
63%
27%
4.5%
0.85%
2.4x10 3
7.4xlO~3
8.6xlO~3
3.5xlO~4
2.2xlO~4
9.0xlO~5
3.7xlO~3
3.4xlO~3
3.2xlO~4

1
0.4
13
0.59
21%
63%
26%
8.6%
1.2%
1.8xlO~3
7.8xlO~3
8.6xlO~3
3.7xlO~4
2.7xlO~4
4.0xlO~5
3.9xlO~3
4.0xlO~3
1.9xlO~4

8
6
16
0.62
7.7%
73%
11%
4.8%
0.45%
        40

-------
                      SECTION VII

         AMD EXPERIMENTS-RESULTS AND DISCUSSION


To verify the results with SAMD, AMD was foamed with NaDS
in the 6 in.-dia. column operating in both simple and
countercurrent modes.  The AMD originated from Grassy Run
(Sample Site GT6-1) near Elkins, W. Va. and was trucked to
Cleveland; the tests began approximately 24 hours after
sampling.  Analyses of the water was conducted approximately
30 hour's after sampling (Table 7) .


                        TABLE 7

        Grassy Run Acid Mine Discharge  (GT6-1)
              Total Fe                 112   ppm

              Ca                       114   ppm

              Mg                        32   ppm

              Mn                         3.8 ppm

              Al                        34   ppm

              pH                       2.75


Two simple mode experiments at different liquid feed and air
sparging rates were conducted on the second and third days
after delivery of the Grassy Run AMD.  The results of these
experiments compare favorably with those using SAMD (Table 6).
Total metal removal was 8.6 and 4.5 percent with 1.2 and 0.85
percent separated as coadsorbed metal.  Approximately 63
percent of the added surfactant remained in the bottoms stream.
The error in material balances was typical of SAMD experiments.
Separation efficiencies were essentially the same at the two
different feed and sparging rates.

An additional experiment in the countercurrent mode was
conducted on the fourth day after delivery of the acid
drainage.  The results  (Table 6) are consistent with pre-
vious tests except for the abnormally low measured surfactant
concentration and surface throughput in the foam.  The
measured surface throughput of coadsorbed metal also appears
to be low.  The surfactant balance between feed and bottom
                           41

-------
flows indicates that surfactant analysis of the foamate
was incorrect due most probably to surfactant loss onto
sampling containers at the high VR of 23.  As with SAMD,
no advantage for operation in the stripping mode compared
to the simple mode was found with AMD.  SAMD appears to
be a valid laboratory substitute for low pH, high Fe(II)
AMD, at least for foam separation studies.
                          42

-------
                      SECTION VIII

                SURFACTANT REGENERATION


Surfactant represents the principle expense in foam separa-
tion of AMD and thus its recovery and reuse in a cyclic
pattern would be required for the process to be economically
competitive.  As a preliminary investigation of surfactant
recovery, the addition of base to collapsed foam was tested
to determine if surfactant and hydrolyzable metals (Fe,
Al and Mn) could be separated by pH increase.

The hydrous oxide of Fe(III) begins to precipitate in weak
acidic solution of pH about 4; precipitation is generally
complete at or before pH 7.  Hydrous aluminum oxide begins
to precipitate at pH 3 or higher and is completely precipi-
tated at or before pH 7; at about pH 9 resolution is evident.
Thus for initial tests, sufficient base was used to bring
the pH up to approximately 7.

Hydrous ferric and aluminum oxides tend to coprecipitate
cations at higher pH, which would be desirable in this case.
However, excess base would also be a cost factor.  Copre-
cipitation of cationic species can be promoted by rapid
addition of base which promotes precipitation in local ex-
cess of hydroxide.

Experiments were conducted with the collapsed foam collected
from SAMD and AMD experiments.  Standardized NaOH was added
to filtered and unfiltered foamate samples.  The samples
(three replicates) were continuously stirred and pH was
continuously recorded.  Once the precipitate was formed it
was allowed to settle briefly before supernatant aliquots
were taken for surfactant and metals analysis.  Some ali-
quots were taken up to 3 days after precipitate formation to
determine if aging had any effect on supernatant composition.
Supernatant samples were either contrifuged or filtered
(0.45 u membrane filter) to remove solids before analysis.
Some foamate samples were filtered to determine the amount
and nature of solid material.

Tests results for SAMD and AMD, although different in detail,
generally agree (Table 8).  From 41 to 62 weight percent
of the collapsed foam is solid surfactant.  Fe, Al and Mn
are precipitated to a greater extent than Ca and Mg.
                           43

-------
                        TABLE 8
          Recovery of NaDS from Collapsed Foam
                  by Addition of NaOH
                   NaDS    Fe
              Ca    Mg    Mn
                (Units; ppm)
Al
pH
Foam from SAMP

  Collapsed foam

  Collapsed foam,
  filtrate
  42 Whatman
  Supernatant
  solution
              (1)
  Percent removed
  filtration,
  42 Whatman

  Percent removed
  filtration plus
  precipitation
Foam from AMD

  Collapsed foam

  Collapsed foam,
  filtrate
  42 Whatman

  Collapsed foam,
  filtrate
  0.45  membrane
487    209    85    28   8.5     27      2.3
286    206    70    25   8.2     25
 88   0.75    53   8.3   0.03   0.33
 41   1.4    6.2   9.0   3.5     10
 82     99    38    70   >• 99  =» 98
        2.3


      8.0-8.5
510    136   123    35   3.8     46
194     99    99    35   3.8     43
194    102    99    35   3.8     44
        2.7
        2.7
        2.7
(1)
    Supernatant solution from addition of 0.150N NaOH to
    filtered collapsed foam (42 Whatman)
                           44

-------
                  TABLE 8  (continued)

                   NaDS    Fe    Ca   Mg    Mn
                                   (Units; ppm)
Al
pH
(2)
Supernatant
solution,
centrifuged
replicate 1
2
3
Supernatant
solution,
filtered (0.45)
replicate 1
2
3
Percent removed
filtration
42 Whatman
Percent removed
filtration
0.45 membrane
Percent removed
precipitation




265 0.10
270 0.20
265 0.15



265 0.12
265 <.l
265 -=.1


62 27


62 25

48 =-99




72
82
78



75
77
77


20


20

37




25
27
27



26
27
28


0


0

21




0.92 «= 1 7.2
1.2 -= 1 6.9
1.2 «= 1 7.0



0.85 •= 1 7.2
1.3 •= 1 6.9
1.2 «cl 7.0


0 6


0 4

71 >98
(2)
    Supernatant solution from addition of 0.147N NaOH to
    unfiltered collapsed foam
The solution resulting from base treatment of foamate con-
tains on the average 265 ppm NaDS, less than 2 ppm total
Fe, Mn and Al, 77 ppm Ca and 27 ppm Mg.  This represents
52 percent recovery of surfactant present in the foamate
with the removal of =» 99 percent iron, =-98 percent Al and
71 percent Mn.

No noticeable difference in separation occurred between rapid
(14 min.) or slow (90 min.) base titratlon of foamate.
                           45

-------
Supernatant pH values tended to drop slightly during the
first two hours after which they remained constant.  For
example, with the AMD foamate samples (Table 8) the pH
of the first replicate dropped from 7.2 to 6.3, the
second replicate from 6.9 to 6.2 and the third replicate
from 7.0 to 6.3 all within two hours; 24 hours later the
pH had not changed.

Calculations based on the foamate from the AMD experiments
indicate an average of 1.5xlO~  moles NaOH are required to
precipitate (at pH 7) the metals from a liter of foamate.
Titration curves indicate that possibly 9.4x10   moles
NaOH may be sufficient at pH 4-5.
                           46

-------
                      SECTION  IX

           ECONOMIC EVALUATION AND  SUMMARY
The effectivness  of  foam separation  for  the  extraction of
metals from AMD is low  (Figure  10).  Approximately 6 per-
cent of  the metals are  removed  per pass  under  the practical
operating  limits  of  the 6  in.-dia. column; of  this 6 percent
approximately  1 percent is removed as coadsorbed metal, the
remainder  as interstitial  liquid  in  concentration equal to
that in  the pool.  Minimum surfactant concentration is de-
termined by that  required  to  produce a foam  of sufficient
persistence to remain intact  during  protracted drainage
which is necessary to reduce  foam density and  liquid carry-
over in  the tops.  The  inherent low  foamability of AMD  re-
quires a surfactant  concentration of about 200 ppm, the
majority of which remains  in  the  bottoms.

With surfactant concentration equal  to 265 ppm in the super-
natant liquid  after  the surfactant regeneration step, re-
usable surfactant amounts  to  6  percent of-that in the feed,
dissolved  in liquid  equal  to  approximately 5 percent of the
feed stream.   Loss of surfactant, 74 mole percent to the
bottoms  and 20 mole  percent to  the sludge, represents a
major technical and  economic  obstacle to efficient operation.

Chemical and air  process costs  can be estimated based on
the flow scheme presented  in  Figure  11.  A ratio of 1.90
moles surfactant  used per mole  total metal removed leads to
a surfactant cost of 235 dollars  per Ib-mole metal removed
assuming 0.43  dollars per  Ib  surfactant and  no surfactant
recovery and reuse.  This  is  equivalent  to 0.78 dollars per
1000 gal.  per  foam separation stage  which removes approxi-
mately 6 mole  percent total AMD metals in the ratio 1 mole
Fe:0.61  mole Ca:0.32 mole Mg:0.05 mole Mn:0.27 mole Al.
To reduce  the  Fe  concentration  to 10 ppm, assuming an initial
Fe concentration  of  200 ppm as  in SAMD, would require 19
process  stages at a  cost of 0.20  dollars per stage per 1000
gal. for each  of  the first 18 stages plus 0.78 dollars for
the last stage giving total cost  of  4.50 dollars per 1000
gal. for surfactant.  This calculation is based on no sur-
factant  recovery  from the overheads  but with make-up sur-
factant  added  to  each successive  stage.  This calculation
assumes  no significant  reduction  in  metal extraction rate
at reduced metal  concentration  and thus  is only an esti-
mate (Figure 10).

At a sparging  air-to-liquid feed  volumetric  ratio of 16,
the cost of air would range between  0.04 and 0.12 dollars
per 1000 gal.  per stage depending on power costs, or be-
tween 0.76 and 2.28 dollars for 19 stages.
                           47

-------
so
Bottoms


 Rbg=2.8xlO



 Rbtm=3'3xl(

 Lb=0.95 Lf
                   A
                    3
                              Tops
                         R  =1.0x10
                          cs


                         R ,  =2.0x10
                          ctm


                         L =0.05 Lf
                          O       J-
                                          ~4
                                           ~4
  Foam

Separa-

  tion
                                             -4
                           Feed


                            Rf =3.8x10



                            Rftnr3-5xl°

                            Lf=500 ml-min
                                              -4
                                                -i
                   Gas

                   G=8 liter-min

                                                    Surfactant

                                                    Recycle
 Surfactant

Regeneration
                                                                                  Base  (NaOH)


                                                                                    R =1.5xlO~2-L
                                                                                     u           c

                                                                                      moles-min"1
                              Rrtm=7-7xl°"

                              L =L
                               r  c
      Sludge


        R  =7.7x10
         ws

        R  . =1.2x10
         wtm
                                                                                     ~5
                                                                                      ~4
                                                   FIGURE 11


                                    Process Flow Scheme for Foam Separation

                                     of Metals and Surfactant Regeneration

-------
It is evident from these cost figures  that the inherent low
extraction capacity of foam fractionation disallows economic
feasibility for treatment of waters as concentrated as AMD.
Air costs for foaming are comparatively low since the process
operates near ambient pressure with little loss.  Even so
the air costs alone would be prohibitive.

Six mole percent of the total surfactant  (23 mole percent
of surfactant in the overheads) can be recovered for reuse
at a cost of 0.06 dollars per 1000 gal. per stage based on
a NaOH cost of 0.25 dollars per Ib.  This degree of re-
covery has no economic advantage.

Reduction in surfactant costs may be effected by operation
with less surfactant and recovery of a greater percentage
of surfactant.  This research program  investigated the
extraction of dissolved metals from AMD by foam separation
under conditions which allowed production of persistent
foams and protracted foam drainage.  Operating with persis-
tent foams requires surfactant concentrations and sparging
rates large enough to support a stable foam without sur-
passing the critical micelle concentration or allowing air
velocities high enough to cause foam rupture.  At lower sur-
factant concentrations and sparging rates a transient foam
regime is generally produced in which foam drainage and
collapse are rapid but controllable.  Extraction in a transient
foam regime must be lower than for a persistent foam regime
due to collapse of surface film and loss of adsorbed material
to interstitial liquid followed by drainage back to the pool.
Reduction in extraction efficiency could be offset by reduc-
tion in surfactant requirements.  An optimum sparging rate
appears to exist for maximum extraction in a transient foam
regime (15); the exact rate being a complex function of
chemical and physical parameters must be determined em-
pirically.  The maximum occurs at a sparging rate such that
the foam drainage rate is significantly greater than the
rate of foam collapse (surface loss).  This would tend to
minimize liquid carryover while maintaining a reasonable
metal extraction rate.  Such a transient foam system would
be characterized by low foam residence times.

A method for surfactant recovery which appears feasible is
solvent extraction of surfactant from aqueous solution
followed by distillation or spray drying to concentrate
surfactant.  Such a system has been mentioned (36) but no
information on its effectiveness was reported.  Surfactant
loss to the bottoms flow could be reduced by transient
foaming or bubble fractionation, but again air costs would
be prohibitive for the treatment of concentrated AMD.
                           49

-------
Economic feasibility of foam separation for AMD could be
improved with the use of waste air and surfactant.  A major
source of waste surfactant is sewage, which suggests the
waste-plus-waste approach for the synergistic treatment of
AMD with alkaline municipal sewage.  Studies of the auto-
purification of AMD and treated sewage effluent have
achieved reductions of 80-95 percent iron, 94-98 percent
phosphate, 60-76 percent surfactant, 3-63 percent COD and
70-84 percent acidity in the combined wastes (23) (38).
The process consisted of aeration and settling unit
operations.

The feasibility of foam separation for sewage treatment
has been investigated with the general conclusion that
the foamability of sewage is too low since the change from
alkyl benzene sulfonate (ABS) to linear alkylate sulfonate
(LAS) detergents.  Recent research into foam separation of
primary and secondary sewage effluents utilizing a combina-
tion of foam fractionation and froth flotation has achieved
reductions of 70 percent COD, 90 percent phosphate and 40-80
percent suspended solids (27).   Air bubbles for particulate
flotation and adsorptive extraction were produced by spon-
taneous degasification of air supersaturated sewage upon
pressure release.  The best results were achieved with the
addition of 300-400 ppm ferric chloride or alum as coagulants;
ferrous sulfate was less successful as a coagulant.

These two independent lines of research suggest that if mix-
tures of AMD and sewage were foamed by degasification, a
process which produces fine bubbles capable of maintaining
a transient foam in sewage, a significant reduction in COD,
phosphate, AMD metals such as Fe and Al, suspended solids,
and surfactant could result in addition to neutralization
of the wastes.  Waste constituents in this type of system
are removed non-stoichiometrically with surfactant which
tends to reduce treatment costs compared to foam separation
of only dissolved, non-surface active constituents of AMD.
Preliminary costs for high pressure foaming of sewage along
appear promising (27).
                           50

-------
                               SECTION  X


                           ACKNOWLEDGEMENTS
The research program was conducted under the direction  of
Dr. Peter J. Hason.  Construction and operation of the  foaming
apparatus and analytical work were performed by Messrs. Paul  H. Hartman
and Jerome H. Jacobs.  The report was prepared by Dr. Hanson.
A significant objective of this project was to investigate practical
means of abating mine drainage pollution.   Such research  projects,
intended to assist in the prevention of pollution  of water by
industry, are required by Section 6 b of the Water Pollution Control
Act, as amended.  This project of EPA was  conducted under the
direction of the Pollution Control Analysis Section, Ernst P.  Hall,
Chief, Dr. James M. Shackelford, Project Manager,  and  Ronald D.  Hill,
Project Officer.
                                        51

-------
                        SECTION XI

                        REFERENCES
 1.  Wace, P. F. , P. J. Alder and D. L. Banfield, 1968, Foam
     Separation Process Design, In:Unusual Methods of
     Separation, Chem. Eng. Progress Symposium Series
     (A. I. Ch. E.) 65:19-28.

 2.  Weinstock, J. J., et al., 1963, Fission Product Separa-
     tion by Foam Extraction, (Radiation Applications Inc.,
     Long Island City, N. Y., NYO-10038) 123 p.

 3.  Schonfeld, E. and A. H. Kibbey, 1967, Improving Strontium
     Removal from Solution by Controlled Reflux Foam Separa-
     tion, Nucl. Appl. 3:353-359.

 4.  Kishimoto, H., 1963, The Foam-Separation of Surface-Active
     Substances, Part I, Fundamental Treatments, Kolloid-
     Zeitschrift and Zeitschrift fur Polymer, 192:66-101.

 5.  Newson, I. H., 1966, Foam Separation:The Principles
     Governing Surfactant Transfer in a Continuous Foam Column,
     J. Appl. Chem. 16:43-49.

 6.  Rubin, E., C. R. LaMantia and E. L. Gadan, Jr., 1967,
     Properties of Dynamic Foam Columns, Chem. Eng.  Sci.
     22:1117-1125.

 7.  Grieves, R. B., 1968, Studies on the Foam Separation Process,
     British Chem. Eng. 13:77-82.

 8.  Rubin, A. J. , J. D. Johnson and J. C. Lamb III, 1966,
     Comparison of Variables in Ion and Precipitate Flotation,
     Ind. Eng. Chem. Process Design Develop. 5:368-375.

 9.  Bauer, D. J. , 1962, Foam Concentration of Scandium,  (U. S.
     Dept. Interior, Bureau Mines, Wash., D. C., RI-5942) 15 p.

10.  Rubin, A. J. and W. L. Lapp, 1969, Foam Separation of
     Lead (II) with Sodium Lauryl Sulfate, Anal. Chem. 41:1133-1135,

11.  Rubin, A. J., 1968, Removal of Trace Metals by Foam Separa-
     tion Processes, J. Am. Water Works Assoc. 60:832-846.

12.  Wace, P. F. and D. L. Banfield, 1966, Foam Separation,
     Chem. and Process Eng. 47:70-76 and 90.

13.  Fineman, M. N., et al., 1952, Foaming of Non-Ionic Surface
     Active Agents,  J.  Phys. Chem. 56:963-966.
                            53

-------
14.  Saturnine, F.  and R.  Lemlich,  1965,  Predicting the Per-
     mance of Foam Fractionation Columns, A.  I.  Ch. E.-I.
     Chem. E. Symposium Series 9:75-78.

15.  Robertson, G.  H.  and T.  Vermeulen,  1969,  Foam Fractiona-
     tion of Rare-Earth Elements (California  Univ.,
     Berkeley, Lawrence Radiation Lab.,  UCRL-19525) 138 p.

16.  Grieves, R. B., 1966, Foam Separation for the Treatment of
     Low-Quality Waters (Illinois Institute of Tech., Chi-
     cago, AD 478452L) 24 p.

17.  Wallace, G. T. and D. F. Wilson,  1969, Foam Separation
     as a Tool in Chemical Oceanography  (Naval Research Lab.,
     Wash. D. C., NRL-6958) 17 p.

18.  Weinstock, J.  J., 1965,  The Removal of Metallic Ions by
     Foaming Agents and Suspensions:Laboratory and Engineering
     Studies, Annual Report,  July 1,  1963-June 30, 1964
     (Radiation Applications  Inc.,  Long  Island City, N. Y. ,
     RAI-350) 41 p.

19.  Davis, W. , Jr., £t a_l. ,  1965,  Laboratory Demonstration
     of the Two-Step Process  for Decontaminating Low-Radio-
     activity-Level Process Waste Water  by Scavenging-Pre-
     cipitation and Foam Separation (Oak Ridge National Lab.,
     Tenn., ORNL-3811) 32 p.

20.  Bikerman, J. J. ,  et, a_l. , 1970, Treatment of Acid Mine
     Drainage, (Horizons Inc., Cleveland, Ohio,  Water Pollution
     Control Research Series  14010 DEE,  FWQA)  88 p.

21.  Hass, P. A., 1965, Engineering Development of a Foam
     Column for Counter-Current Surface-Liquid Extraction of
     Surface-Active Solutes (Oak Ridge National Lab., Tenn.,
     ORNL-3527) 245 p.

22.  Banfield, D. L. and I. H. Newson, 1967,  Improvements in
     or Relating to Foam Deionization Apparatus, British
     Pat. No. 1,062,346.

23.  Streeter, R. C. and D. C. McLean, 1966,  A Study of the
     Interactions and Foam Fractionation of Sewage Effluent-
     Acid Mine Drainage Mixtures (Pennsylvania State Univer-
     sity, Water Resources Research Publication No. 6-66) 51 p.

24.  Rubin, E. , et a.1. , 1963, Contaminant Removal from Sewage
     Plant Effluents by Foaming, (Radiation Applications, Inc.,
     Long Island City, New York, USPHS Publ.  999-WP-5) 56 p.

25.  Brunner, C. A. and D. G. Stephan, 1965,  Foam Fractionation,
     Ind. Eng. Chem. 57:40-48.
                           54

-------
 26.  King. L. J. , ^t  al. ,  1968,  Pilot Plant  Studies  of  the
     Decontamination  of  Low-Level  Process  Waste  by a
     Scavenging-Precipitation Foam Separation Process  (Oak
     Ridge National Lab.,  Tenn., ORNL-3803)  57 p.

 27.  Miller, J.  P. K., et  al.,  1970,  Investigation of a
     High-Pressure FoanflVastewater Treatment Process  (Garrett
     Research and Development Inc., LaVerne, California,
     EPA, Water  Pollution  Control  Research Series 17020) 39 p.

 28.  Schoen, H.  M. , 1966,  Foam Separation  as a Purification
     and Preparative  Tool,  Annals  New York Acad. Sci.
     137:148-161.

 29.  Lamlich, R., 1968,  Principles of Foam Fractionation,
     In:Progress in Separation  and Purification, E.  S. Perry
     (ed), Interscience, New  York,  p.  1-56.

 30.  Lemlich, R., 1968,  Adsorptive Bubble  Separation Methods,
     Ind. Eng. Chem.  60:16-29.

 31.  Perkin-Elmer Corp., 1968, Analytical  Methods for Atomic
     Absorption  Spectrophotometry,  Norwalk,  Conn., looseleaf.

 32.  Hedrick, C. E. and  B.  A. Berger,  1966, Extraction of Anions
     Using Triphenylmethane Dyes,  Anal. Chem. 38:791-793.

 33.  Lee, G. F.  and W. Stumm,  1960, Determination of Ferrous
     Iron in the Presence  of  Ferric Iron with Bathophenantb.ro-
     line, Jour. AWWA 52:1567-1574.

 34.  Chang, R. C., H. M. Schoen and C. S.  Grove, Jr., 1956,
     Bubble Size and  Bubble Size Distribution, Ind. Eng.
     Chem. 48:2035-2039.

 35.  Banfield, D. L., et al., 1966, Surface Excess in Solutions
     of Surface ActiveTgents, I:A Comparison of Static and
     Dynamic Results  (United Kingdom  Atomic Energy Authority,
     AERE-R-5124) 30  p.

 36.  Grieves, R. B. ,  et al. ,  1969, Optimization of the Ion
     Flotation of Dichromate, Jour. Sanit. Eng. Div.  Am. Soc.
     Civil Engrs. SA3:515-525.

37.  Wilson, A., et al. , 1957, The Adsorption of Sodium Lauryl
     Sulfate and Lauryl Alcohol at  the Air-Liquid Interface,
     J. Colloid Sci.  12:345-355.

38.  McLean, D. C. and J. A. Wernham,  1968, A Pilot Plant Study
     of the Autopurification of Sewage Effluent-Acid Mine
     Drainage Mixtures (Pennsylvania  State University, Research
     Publication No. 55) 37 p.
                                55

-------
                      SECTION XII



                NOTATIONS AND GLOSSARY






AMD      acid mine drainage



SAMD     synthetic acid mine drainage



CMC      critical micelle concentration



d        average bubble diameter



FRT      foam residence time, min



L        volumetric liquid rate, ml-min~



G        volumetric air sparging rate, liter-min



M        metals in mine drainage, Fe, Ca, Mg, Mn and Al



R        material mass rate, moles-min



S        surfactant



SR       surface material mass rate, moles-min~



VR       volume reduction (Lf/L )
                            -L  O


[  ]       concentration



\         dynamic surface excess, moles-cm



lx/[X]   distribution ratio, cm







Subscripts



    b    refers to liquid pool or bottom flow



    c    refers to overhead, foamate or tops flow



    f    refers to feed flow



    r    refers to recycle of regenerated surfactant



    s    refers to surfactant



    t    refers to total,  e.g.,  [Mt]f refers to total

         metal concentration in the feed flow

-------
Subscripts
    tm   refers to total AMD or SAMD metals, e.g.,
         R.P.  refers to feed rate of total metal
          ± tm

    u    refers to base in surfactant regeneration

    w    refers to sludge in surfactant regeneration

    x    refers to any variable denoted in text
Adsorption - The concentration of a surface active component
of a solution at a phase boundary.  In the case of foam
separation, the adsorption of surfactant at the gas-liqiiid
interface.

Chelation - An equilibrium reaction between a metal ion and
a complexing agent, characterized by formation of multiple
bonds between the metal and a molecule of complexing agent
and resulting in the formation of a ring structure incorpora-
ting the metal ion (see p. 21 and 27).

Coadsorption - The concentration of a surface inactive com-
ponent of a solution at a phase boundary as a result of its
association with a surface active component of the solution.
In the case of foam separation, the coadsorption of metal
ions with adsorbed surfactant at the gas-liquid interface.

Critical Micelle Concentration - The surfactant concentra-
tion at which significant aggregation of surfactant molecules
into micelles begins.

Dynamic Surface Excess, I   - Defined by equation (2), p. 23.
The surface excess measured in a flowing foam column (see
p. 22-23).

Exchange Capacity - The theoretical maximum time-rate of
material removal per unit cross-sectional area of foam
column.   (See p. 37-39 for discussion and derivation.)

Extraction Rate - The time-rate of adsorbed or coadsorbed
material being removed from solution by the foam.  (See
p. 37-40 and H.gure 10.  Extraction rate and separation rate
used synonymously.)

Foam Coalescence - The rupture of bubble walls resulting in
foam collapse and a change from a two phase ;liquid-gas
system to a single liquid phase.
                           58

-------
Foam Density - The ratio of the gas-free volume of liquid
contained in a foam, to the volume of the original foam.

Micelles. - Aggregates of about a hundred monomer surfactant
units of roughly spherical shape which form as surfactant
concentration is increased.  The change from monomeric to
polymeric units marks the boundary between true and col-
loidal solutions.

Surface Excess, |)  - Defined by Gibbs model, equation  (1),
p. 6.  The surf acre excess of solvent is zero.  Thus, a
solute £ with a positive surface excess means that the
liquid containing a unit area of surface contains ["/  more
of solute than the volume of bulk liquid which con-
tains the same amount of solvent.

Surfactant - An acronym for "surface active agent".  Often
characterized by a linear molecular structure which is com-
posed of solvent compatible radicals at one end and solvent
incompatible radicals at the other end, these solutes in-
variably reduce the surface energy of their solvents by
collecting at an interface.  (See p. 6, 21, and 22 for
further discussion of surfactants as they apply to this
study.

Synthetic Acid Mine Drainage, SAMP -
CaS04-2H20

MnS04-H20

A12(S04)3-18H20

FeS04-7H20

MgS04-7H2O
0.344 g

0.024 g

0.186 g

0.997 g

0.246 g )
                                       Dissolved in one

                                       liter of 0.01 N H2S04
                           59

-------
    Accession Number
                         Subject Field & Group
                                         SELECTED WATER RESOURCES ABSTRACTS
                                              INPUT TRANSACTION FORM
    Organization
             HORIZONS  INCORPORATED, Cleveland,  Ohio
    Title
             FOAM  SEPARATION OE ACID MINE DRAINAGE
 1 Q
    Authors)
     Hanson, Peter  J.
                              16
                                   Project Designation
                                  EPA Project  14010 FUI, Contract  14-12-876
                                21
                                 Note
  22
   Citation
     Water Pollution Control  Research Series, 14010  FUI  10/71
     Environmental Protection Agency, Washington, D.  C.
 23
     Descriptors (Starred First)
      Acid Mine  Drainage,  Foam Separation,   Mine Drainage Treatment,
     *              *       *
      Surfactants,   Iron,  Economics, Metal  Cations, Surfactant

      Regeneration
  25
     Identifiers (Starred First)
27
     -4t>s(rac(Laboratory studies of continuous  flow foam separation  in a persis-
     foam regime were conducted to determine the optimum operating conditions
for maximum  extraction of dissolved metal  cations (Fe, Ca, Mg,  Mn and Al)
from acid mine  drainage.  Foaming experiments  were conducted  in  a 6 in.-dia-
meter glass  column capable of liquid  flow  rates of 3-12 gal.  per hour.  The
effects of pH,  chelate addition, surfactant type and concentration,  air
sparging rate,  metal concentration and  foam drainage were investigated in
relation to  metal  extraction.
  The average extraction rate obtained  was 1.9xlO~  moles total metal per
cm  column cross-section area per minute which is approximately 4.0xlO~7
equivalents  per cm2  per minute.  Operation in simple and countercurrent
foaming modes produced similar extraction  rates for acid mine drainage.
  The low extraction capacity of foam fractionation renders  the process
economically unfeasible for the treatment  of  acid mine drainage.  The prin-
cipal chemical  cost  is for surfactant followed by air.
  Surfactant regeneration from foam by  addition of base was  investigated
for surfactant  reuse and cost reduction.    (Hanson-Horizons)
 Abstractor
 Peter J. Hanson
                         Horizons Incorporated,  Cleveland, Ohio 44122
  WR:I02 (REV. JULY 1969)
  WRSI C
                                       SEND TO: WATER RESOURCES SCIENTIFIC INFORMATION CENTER
                                             U.S. DEPARTMENT OF THE INTERIOR
                                             WASHINGTON, D. C. 20240


                                                                  » CPO: 1969-359-339
                       For sale by the Superintendent of Documents, U.S. Government Printing Office
                                Washington, B.C., 20402 - Price 65 cents

-------