EPA 670/2-74-023
March 1974
Environmental Protection Technology Series
ELECTROCHEMICAL REMOVAL
OF HEAVY METALS
FROM ACID MINE DRAINAGE
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-670/2-74-023
May 1974
ELECTROCHEMICAL REMOVAL OF HEAVY METALS
FROM ACID MINE DRAINAGE
By
Nicholas B. Franco
Robert A. Balouskus
ECOTROL, Inc.
Columbia, Maryland 21044
Grant No. 802614
Program Element 1BB040
Project Officer
Ronald D. Hill
Mining Pollution Control Branch
Industrial Waste Treatment Research Laboratory
National Environmental Research Center
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
REVIEW NOTICE
The National Environmental Research Center ~
Cincinnati has reviewed this report and approved
its publication. Approval does not signify that the
contents necessarily reflect the views and policies
of the U. S. Environmental Protection Agency, nor
does mention of trade names or commercial products
constitute endorsement or recommendation for use*
ii
-------
FOREWORD
Man and his environment must "be protected from the adverse
effects of pesticides, radiation, noise and other forms of
pollution, and the unwise management of solid waste. Efforts to
protect the environment require a focus that recognizes the inter-
play between the components of our physical environment—air,
water, and land. The National Environmental Research Centers
provide this multidisciplinary focus through programs engaged in
o studies on the effects of environmental
contaminants on man and the "biosphere, and
o a search for ways to prevent contamination
and to recycle valuable resources.
This project focused on the development of a new method for
treating acid mine drainage from coal mines. An electrochemical
technique was developed and field tested for converting ferrous
iron to ferric iron to enhance rapid removal of the iron pollutant
from mine drainage.
A. W. Breideribach, Ri.D.
Director
National Environmental
Research Center, Cincinnati
iii
-------
ABSTRACT
Laboratory and field studies were conducted to determine the
economics of ferrous iron oxidation in a cell containing a
bed of conductive particles in the space between the cathode
and the anode. The effects of the process on other heavy
metals present in acid mine drainage (AMD) and on the
character of solids precipitated during treatment of low
acidity water were also observed.
A 18.9 liter/min (5 gal./min ) pilot plant was operated at
an actual mine site to evaluate treatment of 40 and 250 ppm
ferrous iron AMD at pH levels of 2 and 5. A conventional
aeration system was also included to generate comparative
data. Approximately 86 percent of the ferrous iron was
oxidized during electrolysis of the low pH water. The con-
version rate was less in the pH 5 AMD due to coating of
electrode sites with ferric hydroxide. A 4 to 10% decrease
from 4 ppm occurred in manganese concentration, while
aluminum feed values of approximately 1 ppm were reduced by
40 to 60% especially in the pH 2 water.
Estimates for a 473 liter/min (125 gal./min ) plant based
on the pilot data for oxidation only indicate that capital
and operating costs for electrochemical treatment would be
higher than those for aeration by factors of 5 and 1.7
respectively.
This report was submitted in fulfillment of Project Number
14010 GAO, under the partial sponsorship of the Office of
Research and Monitoring, Environmental Protection Agency
and Pennsylvania Department of Environmental Resources
(Project CR-110) by Ecotrol, Inc., 7379 Route 32, Columbia,
Maryland, 21044.
iv
-------
CONTENTS
Page
Foreword
Abstract
List o£ Figures vi
List of Tables viii
Acknowledgments x
Sections
I Conclusions 1
II Recommendations 3
III Introduction 4
IV Phase I - Laboratory Studies 10
V Phase II - Field Studies 56
VI Cost Analysis and Discussion 77
VII References 83
-------
FIGURES
NO. PAGE
1 Rectangular Cells 11
2 Circular Cells 13
3 Flow System for Evaluation of AMD
Treatment at High Flow Rates 17
4 Effect of pH on Percent Reduction in
Iron Concentration 27
5 Effluent Concentration versus Feed
Concentration of Fe(II) for Treatment in
Small Rectangular Cell 30
6 Applied Cell Potential versus Temperature
of AMD 35
7 Comparison of Sludge Settling Rates for
Electrolyzed and Aerated AMD before Lime
Addition 37
8 Comparison of Sludge Settling Rates for
Electrolyzed and Aerated AMD after Lime
Addition 38
9 Percent Bed Expansion of Norit Bed in
Rectangular Column 41
10 Percent Fe(II) Conversion versus Flow
Rate in Rectangular Cell 44
11 Percent Expansion of Norit Bed in
Cylindrical Columns 47
12 Percent Fe(Il) Conversion versus Flow Rate
in Large Circular Cell 50
vi
-------
FIGURES
NO. PAGE
13 Effluent Concentration versus Feed
Concentration of Fe(II) for Treatment
in Large Circular Cell 53
14 Variation in Deep-well AMD Composition
with Time 58
15 Flow Diagram of Pilot Plant 60
16 Unit Cell Used in Pilot Plant 61
17 Operation with Six Unit Cells 63
18 Complete Electrochemical System 64
19 Power Section of Electrochemical Plant 65
20 Effect of Electrode Polarity on Fe(Il)
Conversion 66
21 Effect of Bed Compaction for Four Liter
Per Minute Feed 68
22 Effect of Bed Compaction for Nineteen
Liter Per Minute Feed 69
23 Effect of Flow Rate on Fe(II) Conversion 70
24 Sludge Settling Rates 76
vii
-------
TABLES
No. Page
1 Electrode Bed Materials 15
2 Analytical Methods 19
3 Data for AMD Treatment in Two-Inch
Rectangular Cell with Static Norit Bed 21
4 Evaluation of Electrode Bed Materials 23
5 Data on Effects of AMD pH 25
6 Data on Effects of Fe(II) and Fe(III)
Concentrations 28
7 Data on Treatment of AMD Containing Other
Metals Besides Iron 32
8 Effect of Temperature on AMD Treatment 34
9 Sludge Characteristics 39
10 Data on AMD Treatment at High Flow Rates
in Rectangular Cell 42
11 Data on AMD Treatment in Small Circular
Cells 46
12 Data on AMD Treatment at High Flow Rates
in Circular Cells 49
13 Comparison of Rectangular and Circular
Cell Performance 51
14 Analysis of AMD from Delaware Pumping
Station 57
15 AMD Treatment by Electrolysis and
Neutralization 72
16 AMD Treatment by Neutralization and
Aeration 73
viii
-------
TABLES
No.
17 Capital Cost Estimate for
Electrochemical System
18 Capital Cost Estimate for
Aeration System
19 Estimates of Operating Costs for
Electrochemical and Aeration Systems
lx
-------
ACKNOWLEDGMENTS
The financial support of the project by the Officer of
Research and Monitoring, Environmental Protection Agency,
and Pennsylvania Department of Environmental Resources is
acknowledged with sincere thanks. Thanks are due to
Mr. R. D. Hill, the Grant Project Officer, and to
Mr. D. F. Fowler, Dr. J. J. Demchalk, and Mr. R. Buhrman of
the Department of Environmental Resources for their valuable
technical and administrative assistance.
Dr. R. B. Rozelle of Wilkes College provided much informa-
tion on the nature of AMD in the Wilkes-Barre area.
Witco Chemical Corporation, West Virginia Pulp and Paper,
Pittsburgh Activated Carbon, Union Carbide Corporation,
Berol Corporation, Blue Coal Corporation, and the Pittsburgh
Bureau of Mines are acknowledged for providing samples of
carbon, graphite and coal used in the evaluation of elect-
rode bed materials.
Other personnel who contributed in part to the project are
Mr. J. Shockcor, Mr. W. E. Litz, Mr. D. B. Smith, Mr. R. Hess,
Mr. F. Ferraro, and Mr. E. T. Woodruff.
-------
SECTION I
CONCLUSIONS
The technology of electrochemical oxidation in participate
electrode cells to reduce ferrous iron concentrations in AMD
to low levels has been demonstrated. However, based on the
data obtained and the scale-up factors applied, the electro-
chemical approach was found to be more expensive in capital
and operating costs than conventional aeration.
An important factor affecting the performance of the cells
during treatment of low pH AMD is the occurrence of back
reduction of ferric ion by electrolytic reaction at the cell
cathode and by chemical reaction with hydrogen, which is
also generated in the high acidity water. The steps taken
during this study to minimize these reactions apparently
were not adequate. More effective cell membranes and more
efficient means for gas venting should be incorporated in
future systems.
The back reduction effect should be minimized during oxida-
tion of AMD at pH 3 or higher since the acidity is lower in
these cases and ferric ion is precipitated from solution.
However, deposits of the usual gelatinous ferric hydroxide
in the cell reduce its efficiency by coating electrode sites.
This has a more deleterious effect than the back reduction
problem as indicated by the lower ferrous conversion in high
pH waters. During the pilot study, attempts were made to
remove solids from the cells by applying high feed velocities
to expand the beds. However, the maximum velocity that could
be used was limited by the retention time requirement for
ferrous oxidation. Periodic maintenance procedures involving
vigorous purging of the bed with an inert gas or flushing of
the cells with dilute sulfuric acid to dissolve the precipitate
should be tried.
Throughout the course of the laboratory work, it was repeatedly
observed that during electrolysis of synthetic AMD at a pH
of 3 or greater, a small fraction of the total solids formed
consisted of a dark reddish-brown material with rather
desirable properties. It appeared to be crystalline and
definitely settled at a rapid rate compared to the usual
ferric hydroxide sludge, All attempts in the lab and in the
-------
field to generate the solid in large quantities and to
identify it were unsuccessful. It is concluded that the
formation of this solid in significant quantity probably
requires close control of cell operating conditions.
Electrochemical treatment of AMD has essentially no effect
on calcium and magnesium concentrations. Additional studies
with high manganese AMD are required to confirm current
observations of low percentage removal for this metal.
Indications are that reductions of about 50% in aluminum
concentration may be effected under conditions which also
cause 86% conversion of the iron. Aluminum removal there-
fore should also be investigated further.
-------
SECTION II
RECOMMENDATIONS
In order to improve the economics of electrochemical treat-
ment:
1. Re-evaluate procedure for scale-up from
laboratory data.
2. Modify cells and operational conditions for
more efficient gas and solids expulsion.
3. Evaluate different cell membrane materials.
4. Employ multi-outlet rectifiers to sectionalize
power to the cells system for more efficient
power utilization.
5. Establish periodic bed cleaning procedures.
Pilot studies should be conducted with 378 liter/min (100
gal./min ) plant to provide more reliable scale-up data for
commercial-size systems.
Investigate further the effects of electrochemical treatment
on manganese and aluminum removal.
Determine operating conditions for maximum generation of
the rapid settling solid described in this report.
Investigate the economics of electrochemical conversion
followed by limestone neutralization.
-------
SECTION III
INTRODUCTION
This is the final report on the Pennsylvania Department of
Environmental Resources Project CR-110. The project was
initiated on September 1, 1971 with financial support from
Resource Control, Inc. , the Pennsylvania Department of
Environmental Resources, (D.E.R.) and the U. S. Environmental
Protection Agency (E.P.A.) Office of Research & Development
through Grant 14010 GAO to the Commonwealth of Pennsylvania.
Objectives and Phasing
The primary objective of the overall project was to demon-
strate the economic advantage of using the Ecotrol, Inc.
Electrochemical Process (formerly called the Resource Control,
Inc. Process) to oxidize ferrous iron to ferric iron in acid
mine drainage (AMD). Subordinate objectives include:
1. Process parameter definition for future
application;
2. Determination of effect of the process on
ionic species, other than iron, that can
normally be expected in the waste water; and
3. Determination of effect of the process on
removal of suspended solids in the discharge
from the electrolytic cell by settling and
filtration.
The overall project was divided into two main parts:
1. Phase I - Laboratory Studies
2. Phase II - Field Studies
Phase I activities, which were geared toward developing
laboratory data for the design of a 18.9 liter/min (5 gal./
min ) pilot plant, were completed by mid-April of 1972, when
authorization was given by the E.P.A. and D.E.R. to proceed
with Phase II.
The pilot plant was installed at the Delaware Pumps Station
in Wilkes Barre, Pennsylvania during the last week of June,
1972. However, due to various operational restrictions and
problems resulting from the disastrous June flood, initial
-------
systems testing could not be started until mid-August and the
project was not completed by November as originally planned.
In view of the potential for damage to equipment from freezing
temperatures at the station, it was decided on November 10,
1972, that activity be halted and essential equipment be
removed from the site until April, 1973, when the project
would be resumed. The field work was completed on July 30,
1973.
On August 6, 1973, the D.E.R. service contract with Resource
Control, Inc. was assigned to Ecotrol, Inc., a wholly owned
subsidiary of W. R. Grace & Co., which on July 10, 1973,
acquired substantially all of the assets and technology of
Resource Control, Inc.
Background
The nature, extent and impact of the acid mine drainage (AMD)
problem are amply documented in numerous articles and reports-'-,
Basically, the problem occurs when, in the process of mining
coal, the surrounding rock strata is exposed to air and water.
This rock strata contains metal sulfides , the most common of
which is iron pyrite (FeS2) . The sulf ides of aluminum (A^Sg)
and manganese (MhS) are also found to a lesser extent. When
in contact with air and moisture, these sulfides are oxidized
to sulfates via a reaction that also generates acid. The
basic reaction involving iron pyrite can be expressed as
follows:
2FeS2 + 2H20 + 702 —> 2FeS04 + 2H2S04
The drainage from coal mines will vary in degree of con-
tamination from area to area. A typical AMD will have
approximately the following composition:
Fe+2 100 - 500 ppm
+3
Fe 90 - 300 ppm
S04= 1,000 - 4,000 ppm
Mn+2 15 ppm
Al+3 50 ppm
pH 2.9 - 4.0
-------
Untreated AMD leads to stream pollution in many ways. The
acid is not only harmful to the ecological balance in the re-
ceiving stream, but is also corrosive to man-made structures,
such as bridge abutments and concrete dams. The ferrous
iron is gradually oxidized to the ferric state with a simul-
taneous reduction of dissolved oxygen in the stream. Both
the ferrous and ferric iron, while soluble in acidic water,
will precipitate out of solution as the pH is increased through
dilution. The precipitated yellow-brown hydrous oxides
gradually settle out and cause enough silting to cover the
vegetation in the stream bed. Although precipitated, the
ferrous species will continue to be oxidized to the ferric
state by the dissolved oxygen. Aluminum and manganese in
the AMD will also precipitate out as the pH is increased.
Various physical-chemical processes for the removal of con-
taminants from AMD have been evaluated. Purification by
crystallization, distillation, ion-exchange and reverse
osmosis produce high-quality water^~°. However, it must be
recognized that each of these techniques is not a complete
treatment. In addition to pure water there is also produced
a concentrate of the contaminating salts and acid which
subsequently must be treated by chemical means before dis-
posal. Studies on the treatment of AMD by sulfide or carbon-
ate precipitation, adsorption on mill tailings and by foam
separation of contaminants have revealed problems in the
control of process parameters and have shown unfavorable
cost
By far the most common method for pollution control of AMD
is treatment with alkaline reagents to neutralize the acid
and precipitate metal oxides or hydroxides. While lime
neutralization has been used in most cases, studies with
limestone show the latter to be more desirable on the basis
of cost, availability, hazards of handling and quality of
the precipitated sludge 13-17 ^ jjot a^i types of limestone
work equally well. In general, the more effective materials
are those with a high calcium content, low magnesium content,
small particle size and large surface area-*-^ »•*•".
The major disadvantage to the alkaline neutralization approach
has always been the production of waste sludges. These
sludges are generally quite voluminous and require relatively
long retention times to permit settling out prior to discharge
from the treatment facility. Conventional filtration has
-------
not been too successful, since the form in which the metal
precipitates from solution readily clogs the pores in the
filters. The use of various flocculating agents, coagulant
aids and inert waste materials as additives to increase the
density and enhance the filterability of the sludges has
been studiedl-4-,18,19. One report on the treatment of AMD
sludges includes a procedure for converting sludges with
high ferrous contents (mainly Fe(OH)2> to a ferro-magnetic
form which is denser and faster settling than the material
normally encountered19. All of the aforementioned approaches
for sludge handling exhibit some potential for success, but
the problem is still far from being solved economically.
The most promising sludge dewatering techniques appear to
be centrifugation, conventional rotary vacuum filtration,
and rotary precoat vacuum filtration2^>21.
In the neutralization of AMD from various locations there is
an appreciable difference in the ease of treatment depending
mainly upon the ratio of ferrous iron to ferric iron in the
raw waste water1^17. With high ferric iron AMD, free acid
and the combined acid resulting from hydrolysis of ferric
iron are neutralized, and a maximum degree of iron precipita-
tion is achieved as the pH is increased to three22. The pH
is subsequently adjusted to a vlaue in the range of 5-7 where
precipitate quality and rate of coagulation are best. Mine
drainage with a high ferrous iron concentration is the most
difficult to treat by neutralization. Although ferric iron
is removed from the solution ideally at a pH of about 6.5,
ferrous iron requires a pH of 10 for maximum precipitation.
This necessitates the use of large amounts of lime and
introduces appreciable alkalinity to the treated water. In
addition, the Fe(OH>2 precipitate is light, more voluminus
and, therefore, more difficult to handle than the sludge with
predominately ferric iron.
Attempts have been made to oxidize and hydrolyze the ferrous
iron by aeration after neutralization to pH 6.5. However, a
decrease in pH occurs due to the protons liberated. Thus,
an excess of alkali is still required with this technique.
Finally, investigators have concentrated on the oxidation of
ferrous iron before neutralization as another possible solution
to the problem. Although aeration is probably the easiest
and least expensive method, the conversion of Fe(II) to Fe(III)
by air oxidation is known to be extremely slow in solutions
-------
99 9"^ 9A
with a pH less than six » J» . Various transition metal
ions catalyze the reaction, but still not to an effective
degree. Laboratory tests with various solid substances
have revealed high rates of Fe(II) oxidation when aerated
synthetic and actual AMD at a pH of about 3 were passed
through beds of coal-base activated carbons 14,25,26^
Adsorption, as well as oxidation, is the mechanism involved.
The major disadvantage of the process is the cost of the
activated carbons which apparently became inactive after
relatively short periods of use.
Favorable results have been obtained from a bench-scale
examination of AMD treatment by ozone oxidation before neutral-
99
ization'1 . An engineering design and economic study to
evaluate this system concludes that an ozone process is
feasible and compares economically with existing processes27.
Field tests with on-site generation of ozone by electrical
discharge are recommended.
Laboratory studies have indicated that acidophilic iron
bacteria can effectively accelerate the oxidation of Fe(II)
in AMD, but the bacteria metabolism is temperature dependent,
being greatly reduced at low temperatures28. The biochemical
iron-oxidation limestone process has been successfully demon-
strated during the treatment of AMD streams containing up to
450 ppm Fe(II)29. Presumably consideration is being given
to the design of treatment plants based on this process.
Electrolysis has been considered as a technique for oxidizing
ferrous iron in AMD and for removing total iron by plating
and precipitation^. However, standard electrolytic methods
have never been practical when applied to the removal of trace
contaminants from water (less than 1000 ppm). The long
retention time required for the ions to migrate to the
electrode surface, together with the high power consumption
necessary for treating the increasingly purer water, have
seriously restricted the use of electrochemical technology.
To date, conventional electrolysis has been used only in
waste treatment cases where high concentrations are reduced
to lower concentrations prior to chemical treatment.
The present report covers the use of electrolytic cells
containing a semiconductive bed of particles between the
electrodes. The bed particles, while not reacting chemically
8
-------
with the contaminants in solution, behave as small inter-
mediate electrodes in the electric field between the major
electrodes. That is, ends of the particles facing the major
cathode exhibit anodic characteristics, and ends facing the
major anode exhibit cathodic characteristics. The particles
themselves become bipolar. The result is a cell with a
multiplicity of anodic and-cathodic sites. The cell can also
be modified to provide a bed with predominantly anodic or
cathodic character. When the solution to be treated is
passed through this medium, the distance contaminant ions
have to migrate to any one of the numerous electrode sites is
drastically reduced. This feature overcomes the excessive
residence time required by conventional cells in the treatment
of dilute waste water. Since the particles are conductive,
and because of their proximity in the cell, abnormally high
voltages are not needed to oxidize or reduce trace quantities
of dissolved impurities.
While initial development of the Ecotrol cell was directed
towards applications in the metal finishing industry,
exploratory tests with synthetic AMD containing up to 500 ppm
ferrous iron showed that the iron could be oxidized at an
acceptable rate in the cell. In addition, it was found that
in AMD with a pH of 3 or higher, the ferrous iron was converted
to a relatively dense red solid that appeared to be readily
filterable.
In February 1972, Tyco Laboratories, Inc. completed a
laboratory study on the electrochemical oxidation of ferrous
AMD in a packed bed electrode cell similar to the Ecotrol
Cell^O. Solutions with pH 2 or less were used. Calculations
from the laboratory data indicate favorable economics for the
operation of commercial scale systems.
For the present project, the laboratory work was followed by
a field study on the treatment of pH 2 and pH 5 AMD.
-------
SECTION IV
PHASE I - LABORATORY STUDIES
The purpose of this part of the overall project was to examine
the effects of cell configuration, electrode bed material,
power, AMD composition, and temperature on the degree of treat-
ment in bench-scale cells in order to gain some idea of the
range of operational capabilities which had to be incorporated
in the design of the Phase II system. While consideration was
given to wide variations in AMD composition, some emphasis was
placed on those conditions which were to be encountered at the
mine site chosen for the field demonstration.
Experimental Details
Laboratory Cells and Associated Equipment
Cells were designed in two different configurations. The
Ecotrol Electrochemical System in its basic form, used, for
example, in treating waste waters from the metal finishing
industry, consists of a series of cells having rectangular
plane-parallel anodes and cathodes with the semi-conductive
bed of particles dispersed in between. The basic rectangular
configuration was, therefore, considered first for evaluation
of AMD treatment. From other studies in this laboratory,
cylindrical cells with concentric anodes and cathodes were
found to give more favorable results than the rectangular
type in certain applications. Thus, the cylindrical con-
figuration was also chosen for AMD study.
A typical rectangular cell is illustrated in Figure la. It
consists of a stainless steel cathode (#304) and graphite
anode (Union Carbide Electrolytic Graphite) supported and
spaced from each other at a desired distance with polyvinyl
chloride (PVC) or clear Lucite side walls. This assembly is
mounted on a PVC bottom plate, which is equipped with a
solution inlet tube and a PVC screen (18 x 36 mesh). The
latter acts as a support and permits even distribution of
solution flow through the bed of small particulate electrodes
dispersed between the graphite and steel electrode plates.
A sidearm is provided on the solution inlet tube or a separate
tube is attached to the bottom section of the cell in order to
introduce gases such as nitrogen or air. These can be used
10
-------
EXIT PORT
TOP VIEWS
O
GRAPHITE
ANODE
STAINLESS STEEL
CATHODE
ELECTRODE BED
PARTICLES
PVC SCREEN BOTTOM
O
DIAPHRAGM
FIGURE 1. Rectangular Cells
-------
periodically to sparge the bed so as to remove any accumulated
solids which could clog the bed or insulate it electrically.
The cell is assembled by bolting various components together,
cementing with epoxy, or welding with a hot air gun (plastic
to plastic bonds).
Figure Ib shows the same cell equipped with a diaphragm. The
material used in this study is PVC coated fiberglass insect
screen (18 x 36 mesh) welded to a rigid PVC frame. The dia-
phragm can be placed directly against one of the major elec-
trodes, or can be positioned at some desired distance between
the two. These features permit one to operate the cell with
the electrode bed particles completely isolated from the
graphite anode or the steel cathode. Also, the bed can be
sectionalized into anodic and cathodic compartments.
A typical cylindrical cell with concentric major electrodes
is shown in Figure 2a. The bottom plenum chamber, made from
PVC or clear Lucite, is equipped with a solution and gas
inlet tube. The chamber is bolted to a top section which
holds a PVC flow distributor and the inner cylindrical electrode,
The top section, with built-in solution exit port, may be
metallic and thus act as the outer major electrode, or may be
constructed from plastic material. In the latter case, the
outer electrode would consist of a thin-wall metallic cylinder
which is slid into the top section and fits snugly against the
inner plastic wall. The bed of small particulate electrodes
is dispersed in the annular space between the inner and outer
major electrodes. The bottom of the inner electrode is sealed
to insure that all the solution flowing through the cell is in
contact with the bed in the electric field between the major
electrodes. Because of the difference in surface areas of the
concentric major electrodes, one possible modification in the
operation of the circular cell involves the polarity impressed
on these electrodes. The inner electrode could be the anode
and the outer the cathode as in Figure 2b, or vice versa
(Figure 2c). For this study stainless steel (#304) was em-
ployed as the cathode material. The anodes were constructed
from graphite and titanium coated with ruthenium dioxide
(Electrode Corporation). The latter material is more chemically
inert than graphite, and the oxide coating is known to catalyze
oxidation reactions. As is true for the rectangular cells, the
circular cells can be equipped with PVC screen diaphragms,
which in this case would be cylindrical and concentric with
the major electrodes.
12
-------
•
(a)
INNER ELECTRODE
EXIT PORT
OUTER ELECTRODE
•PVC SCREEN BOTTOM
INLET
TOP VIEWS
(b)
(c)
ELECTRODE BED PARTICLES
FIGURE 2. Circular Cells
-------
Ten different materials were evaluated as particulate elec-
trode bed media for AMD treatment in the Ecotrol cells. The
candidates are listed in Table 1. In this laboratory
Norit RB-II carbon had been used in cells for treating
various metal finishing waste waters. The Nuchar and
Pittsburg carbons have been studied as catalysts for the
oxidation of Fe(II) in AMD by aeration1^>25»*6. witco
carbons were chosen because of their relatively high surface
area (1200 m^/g) , low ash content (1%), and low acid soluble
iron content (0.1%). Their uniform spherical shape was also
a property of interest. In this regard, Columbia LCK carbon
was tested because of the platelet geometry of the particles.
The economic advantage of using ground-up bituminous and
anthracite coals prompted examination of these materials.
Finally, samples of graphite were tested. These have the
same rod-like shape of the Norit RB-II particles , but are
not activated or porous like the carbon. In addition, it
was thought that the graphite would exhibit greater abrasion
resistance upon continued use.
The cells described above were designed for continuous flow-
thru treatment of AMD with the liquid flowing from the bottom
to the top of the cell. Under this general mode of operation,
two variations are possible in regard to the physical dis-
position of the bed during fluid flow. In one case, the bed
is held in a static compact state, either by mounting a
constraining PVC screen on the top of the bed or simply by
pumping at relatively low volume flow rates so that the
apparent liquid velocity is less than required to expand the
bed over its normal backwash and drain condition. In the
second case, the bed is allowed to expand or become fluidized
by virtue of the high liquid velocities employed. Preliminary
work with waste water similar to AMD had demonstrated a need
for some form of bed agitation to free the precipitate that
may form on the particles and major electrodes. Beyond this,
the benefits of reduced chemical polarization, enhanced mass
transfer and gas bubble expulsion caused by agitation and
separation of the particulate electrodes warranted further
investigation in the area of AMD treatment.
For runs at low flow rates (less than 0.4 liter/min or
0.1 gal./min ) batches of synthetic AMD were made up and stored
in 114 to 189 liter (35 to 50 gallon) plastic tanks from which
the liquid was fed to the cells by means of bench-top peris-
taltic pumps (Masterflex Pump, Cole Palmer Co.). For studies
14
-------
TABLE 1. ELECTRODE BED MATERIALS
Material
Norit RB-II
Nuchar WV-W
Witco
Witco 256
Pittsburgh BPL
Columbia LCK
Bituminous Coal
Anthracite Coal
Graphite 9H
Graphite 6B
Description
Activated carbon from peat; rod-
shape (2x7 millimeters)
Coal base activated carbon;
granules (8x^0 mesh)
Petroleum base activated carbon;
spherical granules (6x16 mesh)
Spherical granules (4x10 mesh)
Bituminous base activated carbon;
granules (6x16 mesh)
Coke base activated carbon;
platelets (4x10 mesh)
Jagged granules (6x16 mesh)
Jagged granules (6x16 mesh)
Hard pencil lead; rod-shape
(2x10 millimeters)
Soft pencil lead; rod-shape
(2x7 millimeters)
Source
American Norit
West Virginia
Pulp & Paper
Witco Chemical
Witco Chemical
Pittsburgh Activated
Carbon
Union Carbide
Pittsburgh Bureau
of Mines
Blue Coal Corp.
Berol Corp.
Berol Corp.
-------
at high flow rates (up to 11 liter/min or 3 gal./min ) a
more complex system was used. This is illustrated by the
flow diagram in Figure 3. The symbols refer to the following
components:
FV-10 Flow Control Valve (PVC)
V-20,30,40 Quick-close Ball Valves (PVC)
FM-10 Rotameter (No.FP-l-35-G-10-83),
Fisher-Porter
FM-20 Orifice Flowmeter with Mercury
Manometer
P-10 Jabsco Centrifugal Pump
(39 liter/min or 10 gal./min )
P-20 BIF Metering Pump (1 liter/min
or 0.3 gal./min )
T-10 Plastic-Lined Fiberpak Drum
(208 liter or 55 gal.)
CM/R-10 Recording Conductivity Meter,
Balsbaugh 910
VC-10 D.C. Power Supply (15V - 6/30A),
R. 0. Hull
Plumbing between components consisted of 3.8 cm (1.5 in.)
rigid and flexible PVC tubing and fittings. In order to accom-
modate the need for high feed flow rates while avoiding the
inconvenience of storing large volumes of premixed feed,
synthetic AMD with the desired composition was continually
generated during a given run by in-line mixing of metered
tap water and concentrated acidic ferrous sulphate solution.
For all of the Phase I electrochemical runs, D. C. power was
applied to the cells with rectifiers manufactured by the
R. 0. Hull Co. (0-15 Volts, 0-6/0-30 Amps). When applied
voltages greater than 15 were required, the rectifiers were
ganged in series.
Synthetic AMD
Simulated AMD was used for the laboratory studies. Solutions
with the desired pH and metal composition were prepared by
dissolving appropriate amounts of reagent grade sulfuric acid,
sodium hydroxide, and the common hydrated sulfates of Fe(II) ,
Fe(III), Mn(II) , Ca(II), Mg(II) , and Al(III) in tap water.
During runs with AMD at pH 5, the surface of the feed solution
was blanketed with nitrogen to prevent oxidation of Fe(II) by
aeration.
16
-------
T-10
1. JLW
Vjn/ r
c
1
-U 1
1
1
>J
P-20
(
/
r ~\
£J
I
V-20 |
CITY / — \ *
•••••
V
H><] »
FV-10
[
FM 20
1
vc-io
i
f
•
s]
CELL S )
?V-30
FM ^
WATER \^_S
10
.
.
P-10 f
V-40
TO DRAIN
FIGURE 3. Flow System for Evaluation of AMD Treatment at High Flow Rates
-------
Operating Procedures
In preparation for a run with AMD of a given composition, an
ample supply of electrode bed material was pretreated by first
soaking the particles in 10 percent H2S04 for three days to
dissolve any acid soluble matter. The bed was then rinsed
continuously with tap water until the supernatant wash liquid
exhibited a pH approximately equal to that of the AMD to be
studied. After this rinse, the particles were soaked for
several days in the AMD solution. During this period, the
supernatant liquid was removed and fresh AMD added several
times in order to saturate the bed with the test solution.
This procedure was followed to minimize carbon absorption
and pH aging effects , which would otherwise have been encountered
during electrolysis of the AMD.
The pretreated bed was slurried into the cell with the aid of
the AMD feed solution. For the runs at low flow rates with
static beds , the bed was tamped slightly with a PVC screen
assembly. The latter was left resting on the top of the bed
for the duration of the experiment in an attempt to insure a
constant degree of compaction from run to run. Voltage was
applied to the cell soon after the flow of AMD was adjusted
to the desired rate.
For the runs at high flow rates, concentrated acidic ferrous
sulphate solution 8,000 ppm Fe(II) , stored in a plastic drum,
T-10 (Figure 3), was metered to the tap water line to provide
the required AMD feed. To insure constancy of AMD composition
the feed was monitored periodically by withdrawing samples for
analysis from the exit of the valve, V-30. Also, on-the-spot
checks for any drastic changes were made by reading the re-
corded output from a conductivity cell placed in the feed
line. Once the appropriate valve and pump settings were made,
the concentration of Fe(II) in the feed usually remained
within a few percent of the desired value for the duration of
the run.
Feed solution was continually pumped through the cells for
periods of a few to sixteen hours depending on the amount of
data required from the run. Samples of feed and effluent
were taken periodically for analysis to evaluate the degree
of treatment under a given set of operating conditions.
18
-------
Analytical Methods
The approved methods listed in Table 2 were used during the
Phase I study. Initial attempts to analyse for Fe(II) in
solutions with high Fe(III)/Fe(II) ratios by the bypyridine
method indicated that high concentrations of Fe(III) inter-
fered with this test. In order to obtain an accurate analysis,
therefore, Fe(II) was determined by titration with standard
dichromate solution. Total iron was measured in the usual
manner with bypyridine following the reduction of Fe(III)
with hydroxylamine.
Analyses were generally done within one hour after the
samples were taken. Aliquots of high pH solution were
fixed with acid to preserve Fe(II) values. All samples
were filtered before analysis to remove suspended solids.
TABLE 2. ANALYTICAL METHODS
Parameter Method
Aluminum Colorimetric-Phenanthroline
Calcium and EDTA Titration for Total
Magnesium Hardness
Iron, Total Colorimetric-Bipyridine
Ferrous Colorimetric-Bipyridine for
Low Fe(III) Samples
Dichromate Titration for
High Fe(III) Samples
Manganese Colorimetric-Persulfate
Results and Discussion
It should be noted at this point that no attempt was made to
effect 100 percent conversion of Fe(II) or reduction in total
iron (Fe-p) concentration during the Phase I studies, as the
19
-------
cells used were too small to provide the required residence
time. The main objective was to cause enough change in the
concentration of these species by various cell and operating
adjustments so that the effects of these adjustments could
be clearly distinguished and the data applied to the design
of the Phase II system.
Screening Runs with Rectangular Cell
Initial experiments consisted of evaluating the effects of
power and the addition of a PVC screen diaphragm to the cell.
High ferrous AMD with pH 3 and Fe(II) concentration of about
500 ppm was processed at the rate of 83 ml/min (0.022 gal./
min ). Data obtained with a static Norit bed in a cell
containing major electrodes (15.2 cm wide by 17.8 cm high)
spaced 5 cm (2 in.) apart are listed in Table 3. The first
column on the left shows the cumulative running time and the
hour at which a sample of the effluent from the cell was taken
for analysis. The remaining columns list the applied potential,
current drawn, exact composition of the feed and effluent,
and the percent conversion of Fe(II) relative to the feed
value. All iron concentrations are expressed in parts per
million (mg/liter).
Since it is known that the oxidation of Fe(II) by oxygen is
catalyzed by activated carbon, AMD was initially pumped through
the cell while the major electrodes were disconnected from the
rectifier, in order to determine the degree of Fe(II) oxidation
by any oxygen dissolved in the feed or absorbed on the Norit
bed. After five hours of such operation, analysis of the
effluent indicated only 2% conversion of Fe(II). A small
amount of ferric hydroxide precipitate was formed. Shortly
after the fifth hour sample was collected, a potential of
10 volts was applied to the cell and maintained for an addi-
tional nine hours. During this period solids were continually
precipitated throughout the bed while a steady-state Fe(II)
conversion of about 23% was recorded.
A second run was conducted using the same cell fitted with a
PVC diaphragm against the major cathode. In this condition
the unit drew about one-third the current observed without
the diaphragm at the same applied potential. However, the
average percent conversion of Fe(II) was almost twice the
value measured during the previous run. After the fifth hour
effluent was sampled, ferric hydroxide precipitate was
20
-------
TABLE 3. DATA FOR AMD TREATMENT IN TWO INCH RECTANGULAR CELL
WITH STATIC NORIT BED
AMD FEED3 _EFFLUENTa
TIME
(hr)
Volts
Amps
F
PH
"r
Fe(ll)
1
Fe(lll)
t
PH
F6T
Fe(ll)
1
Fe(lll)
%CONV
(No Diaphragm)
5.0
6.5
12.0
12.5
14.0
0
10.0
10.0
10.0
10.0
0
6.0
7.8
7.9
8.3
3.0
3.0
3.0
3.0
3.0
495
495
495
495
495
487
487
487
487
487
8
8
8
8
8
2.8
2.7
2.7
2.7
2.7
475
412
410
432
410
475
350
368
388
368
0
62
42
45
42
2
28
24
20
24
(Diaphragm Against Cathode)
1.0
5.0
6.0
10.0
10.0
10.0
10.0
10.0
2.2
2.4
2.6
2.6
3.1
3.1
3.1
3.1
475
475
475
475
462
462
462
462
13
13
13
13
2.9
2.8
2.8
2.8
322
388
322
345
272
325
265
288
50
63
57
57
41
30
43
38
alron concentrations measured as ppm or mg/liter.
-------
removed from the bed by a combination of nitrogen purge and
fast flow of feed solution applied for several minutes. As
a result, the current immediately increased to 2.8 amps. An
hour later, when there was a small amount of fresh solids
generated on the bed, the effluent showed an improvement in
Fe(II) conversion. However, the improvement decreased as
the run was continued into the tenth hour and more solids
were deposited on the bed.
The results of these initial runs indicated that for efficient
Fe(II) conversion the Ecotrol cell should be modified with the
inclusion of a diaphragm against the major cathode to increase
the anodic character of the bed and that means must be provided
for periodic purging of the bed to remove non-conductive
solids precipitated during the treatment of AMD at pH 3 or
higher.
Evaluation of Electrode Bed Media
The materials described in Table 1 were evaluated as static
beds in a rectangular cell with major electrodes (15.2 cm
wide by 17.8 cm high) spaced 10.2 cm (4 in.) apart and PVC
screen placed against the cathode. Electrolysis was conducted
for seven hours while AMD (pH 3, 500 ppm Fe(II)) was continu-
ously pumped through the bed at a rate of 0.16 liter/min
(0.043 gal./min ). Except for the runs with the coal samples,
the cell current was maintained at 3 amps in each case.
The results of the evaluation are shown in Table 4. The
applied potential, current and percentage figures are
averages of values measured on an hourly basis after the
cell attained equilibrium in regard to Fe(II) and Fex re-
moval. The pH values of the effluents were generally the
same as the feed (3.0) or slightly less (2.8). Solids were
generated in every run, the most voluminous being observed
when anthracite and bituminous coals were used.
The bed materials are listed in Table 4 in the order of de-
creasing effectiveness for AMD treatment. The percent
reduction in Fe(II) concentration observed with Graphite 9H
was almost as high as that for the Norit bed with similar
shaped particles; however, the graphite required a much
higher applied voltage. Apparently the high surface area
and porous structure of activated carbon are not essential
properties for an effective electrode bed material in view
22
-------
TABLE 4. EVALUATION OF ELECTRODE BED MATERIALS
PERCENT
REDUCTION IN
CONCENTRATION OF
BED
Norit RB-II
Graphite 9H
Pittsburgh BPL
Witco 357
Graphite 6B
Nuchar WV-W
Columbia LCK
Anthracite
Witco 256
Bituminous
Volts
6.4
16.0
8.8
10.9
8.0
9.7
11.4
25.0
8.4
25.0
Amps
3.0
3.0
3.0
3.0
3.0
3.0
3.0
0.25
3.0
0.20
Fe(ll)
32
23
18
17
17
15
13
10
10
6
Fe
j
18
11
12
12
8
10
9
10
6
5
-------
of the results with the smooth, non-porous particles of
Graphite 9H.
It was impossible to pass a current greater than 0.25 amps
through the anthracite and bituminous beds even at an
applied potential of 25 volts. Consequently, the degree
of Fe(II) oxidation was poor for these systems. The high
resistance of the coal-filled cells is probably due to the
high resistivity of the particles, 105 to lO1^ ohm-cm.
Another factor is the jagged shape of the ground coal which
would influence electrical contact between the particles and
solution flow patterns.
From a consideration of the data in Table 4 and costs of
the materials, it was concluded that, of the solids tested,
Norit RB-II would be the best electrode media for demonstration
of the treatment of AMD. Further work may identify other
bed materials which could be considered from a cost stand-
point for commercial units. However, to meet the objectives
of this project, Norit was used in all subsequent tests.
Effects of AMD pH
The data in Table 5 illustrate the effects of varying the
pH of AMD solutions containing Fe(II) in essentially two
different initial concentrations. Runs 13 to 16 were con-
ducted with 500 ppm Fe(II) solutions, while 50-60 ppm Fe(II)
AMD was used for Runs 17 and 18. A rectangular cell with
10.2 cm (4 in.) spacing of the major electrodes (12.7 cm
wide by 17.8 cm high) and static anode bed was used. The
flow rate was 0.15 liter/min (0.040 gal./min ) and running
time generally about 12 hours during which samples were
taken from the cell exit port and analyzed on an hourly basis.
All of the data in Table 5 are average values of measurements
made after the cell exhibited a steady-state level of treat-
ment.
Consider first the runs with high ferrous AMD. A comparison
of the values for applied voltage necessary to maintain a
current of approximately 3 amps indicates a solution conduc-
tivity effect. As the sulfuric acid concentration was
decreased in going from a pH 2 to a pH 3 solution, the appliec
voltage had to be increased by a factor of almost 2. Further
increase in pH to 4 and higher by the addition of sodium
hydroxide resulted in a decrease of the applied voltage
24
-------
TABLE 5, DATA ON EFFECTS OF AMD pH
AMD FEED3
PERCENT
ro
Ui
RUN
13
14
15
16
17
18
i
_£H
2.0
3.0
4.0
5.3
2.0
5.3
FeT
500
490
515
488
51
64
Fe(ll)
495
485
503
480
48
61
1
Fe(lll)
5
5
12
8
3
3
Volts
7.3
17.2
13.5
13.5
8.0
14.6
Amps
3.1
3.1
3.0
3.0
3.0
3.1
DECREASE
IN
Fe(ll)
195
115
120
114
40
44
REDUCTION 11
CONCENTRATION
Fe(l I) Fa
39
24
24
24
83
72
2
5
15
15
6
52
All iron concentrations measured as ppm or ing/liter.
-------
relative to the value for the pH 3 run.
The effect of pH on the degree of AMD treatment is indicated
by the Fe(II) data (Decrease in Fe(II)) and the values of
percent reduction in Fe(II) and Fe^ concentrations. The
last two variables are plotted in Figure 4. The degree of
Fe(II) conversion drops from its value of 39 percent for the
pH 2 solution to a plateau value of 24 percent for the AMD
with pH 3 to 5. At pH 2 some elemental iron was plated on
the cathode, and only a small amount of hydrous oxides was
produced. An increase in solids production at the high pH
levels is reflected by curve in Figure 4 for percent reduction
in total dissolved iron. This value increases and levels off
with increasing pH of AMD feed.
The influence of pH variation on the applied potential for
the two runs with 50-60 ppm AMD is the same as observed with
500 ppm Fe(II). Considering the Fe(II) values, however,
there appears to be little change in the degree of Fe(II)
oxidation with pH in the low ferrous solutions. The percent
reduction in Fe-p increased significantly with pH, as would be
expected on the basis of the results with the 500 ppm Fe(II)
solutions.
It was concluded that the effect of pH on the Fe(II) conversion
lies mainly in the extent of ferric hydroxide formation.
Conductivity differences can be compensated for by appropriate
adjustment of the applied potential. While a decrease in
oxidation effeciency with increased pH was observed with the
high ferrous AMD, the pH effect was negligible for the 50-60
ppm solutions since much less solid was deposited during the
pH 5 run by virtue of the low iron concentration.
Variation of Fe(II) and Fe(III) Concentrations
The same cell and operating conditions as used for the work
on pH effects were applied in a study of Fe(III)/Fe(II)
variations. The results are shown in Table 6. For the first
three runs, the concentration of Fe(II) was varied in pH 2
water containing very little Fe(III). To examine the effect
of the Fe(III)/Fe(II) ratio, the same pH and Fe(II) concentra-
tions were used for the next three runs. However, Fex was
kept as close as possible to 500 ppm by the addition of Fe(III)
in Fe2(S04)3.xH20.
26
-------
a
o
cd
I*
0)
OS
50-
40.
0)
o
c
o
a
§ 30
c!
•H
o 20'
•H
4J
0
10-
Initial Fe(II) s=r 500 ppm
Fe(II)
-G-
•O
•O
I
3
pH of AMD
FIGURE 4. Effect of pH on Percent Reduction in Iron Concentration
-------
TABLE 6. DATA ON EFFECTS OF Fe(II) AND Fe(III) CONCENTRATIONS
ro
oo
1
£H_
2.0
2.1
2.0
2.0
2.0
2.0
5.3
5.3
5.3
AMD
f Cm
500
266
51
500
495
448
488
250
64
FEED a
Fe(II)
495
260
48
495
285
55
480
250
61
1
Fe(III)
5
6
3
5
210
393
8
0
3
Volts
7.3
7.6
8.0
7.3
8.2
7.9
13.5
14.5
14.6
Amps
3.1
3.1
3.0
3.1
3.1
3.1
3.0
3.1
3.1
Effluen
FeClI)
300
125
8
300
130
30
366
150
17
All iron concentrations measured as ppm or mg/liter.
-------
The last three runs were performed to observe the effect of
Fe(II) concentration in AMD with pH 5. Attempts were not made
to maintain a Per value of 500 ppm in this series, as the
excess Fe(III) would merely have precipitated when added to
the pH 5 feeds.
Figure 5 contains plots of effluent Fe(II) concentrations
versus the corresponding feed values. A comparison of the
curves for the pH 2 solutions with and without initial
Fe(III) addition indicates some degree of back reduction of
Fe(III) to Fe(II) in the AMD with high Fe(III)/Fe(II) ratio.
Although the electrode bed was made predominately anodic,
the probability for Fe(III) reduction at the surface of
the major cathode would be higher in solutions with a
high Fe(III)/Fe(II) ratio than in the high Fe(II) water. Some
chemical reduction of Fe(III) by hydrogen generated at the
cathode could also occur.
Although a lower conversion of Fe(II) takes place in the
high ferrous AMD at pH 5 than at pH 2, the precipitation of
Fe(III) from the high pH solution would result in less back
reduction to Fe(II). Also, a lower concentration of hydrogen
would be produced in the pH 5 water. The difference in results
between the high and low pH AMD therefore, approaches zero
at low Fe(II) feed concentrations.
Attempts to fit the data to simple first and second order
rate equations were unsuccessful. However, the plots in
Figure 5 can be used as working curves to determine the
minimum length of cell needed to effect a desired degree
of Fe(II) oxidation. Consider, for example, the pH 2
solution with 500 ppm Fe(II) and no Fe(III). According to
the corresponding curve in Figure 5, as this solution is
passed through the unit cell used in this case, the effluent
will have a Fe(II) concentration of 305 ppm. The concentra-
tion of the effluent that would be obtained by flowing the
305 ppm solution through another unit length of cell can then
be approximated by reading from the curve the effluent value
corresponding to a 305 ppm feed. This procedure is continued
until the required cell length is determined.
Treatment of AMD Containing Other Metals Besides Iron
A series of runs were conducted with synthetic AMD containing
aluminum, calcium, magnesium and manganese. It was of interest
29
-------
1000
0)
PH
8.
a
c 100
o
•i-i
4J
cd
M
4J
C
0)
Cl
d
o
•u
C
w
10
Feed
2, FeT^Fe(II)
O pH 5
A pH 2, Fe(III) Added
J.
J_
100 200 300 400 500
Feed Concentration, ppm Fe(II)
FIGURE 5. Effluent Concentration versus feed
Concentration of Fe(II) for Treatment
in Small Rectangular Cell
30
-------
to determine the effect of electrolysis on these metals and
the effect of these metals on the degree of Fe(II) oxidation.
Cells and operating conditions were the same as used for the
pH and iron concentration studies. The results are shown in
Table 7, wherein Me is a general symbol for the metal present
besides iron. For the run with calcium and magnesium, these
were added in a weight ratio of 1.4 Ca/Mg. The magnesium
concentration was converted to a calcium equivalent and the
feed value of Me expressed as total calcium.
As for the effect of electrolytic treatment on Mn(II) in the
electrode bed cell, only a small percent reduction in this
metal's concentration was observed. The analytical procedure
used for determining manganese in solution does not differ-
entiate oxidation states , but converts all lower valence ions
to the permanganate form (Mn04~) by oxidation with persulfate.
The presence of Mn04~ in the effluent samples before the
addition of persulfate was precluded by the absence of the
familiar pink color characteristics of MnO/~ which is detect-
able even in 1 ppm Mn solutions containing excess Fe(II).
The small reduction in Mn(II) concentration could have re-
sulted from oxidation to insoluble MnO£; however, it was
impossible to detect a small amount of this solid in the mix-
ture with excess iron oxides, which were also formed.
In regard to an effect of Mn(II) on the conversion of Fe(II),
the data show a reduction of about 25% in the fraction of
initial Fe(II) oxidized in the presence of 50 or 90 ppm Mn(II).
The same data were obtained from replicate runs. These results
present an anomaly as one might expect the opposite effect when
known redox potentials are considered; i.e., oxidation of
Mn(II) to MnO^' which in turn would oxidize Fe(Il). No
explanation can be offered at this time.
It was concluded from this series of runs that, in view of the
small deleterious effect of manganese on the Fe(II) conversion,
the treatment of actual AMD containing 50 ppm or more of this
metal would require a somewhat longer residence time than
drainage waters having little or no manganese. It was hoped
that more definitive data would be obtained during the field
study. The results of the runs with calcium, magnesium and
aluminum indicate essentially no effect on the degree of Fe(II)
oxidation and a small percentage removal of these metal ions
in the high pH AMD.
31
-------
TABLE 7. DATA ON TREATMENT OF AMD CONTAINING OTHER METALS BESIDES IRON
Decrease Percent
FEED In Reduction in
1
£H
5.1
5.1
4.9
4.8
4.5
Fe(II)
413
395
352
395
390
Fe(III)
0
0
5
0
4
i
Me Volts
0 13.4
[MANGANESE]
48 15.6
87 12.3
Amps
3.1
3.1
3.1
[CALCIUM/MAGNESIUM (1.4)]
326b 10.6 3.1
[ALUMINUM]
61 12.0
3.1
i i
Fe(II)
113
80
74
114
105
1
Fe(II)
27
20
21
29
27
i
Me
-
4
10
3
3
aAll metal concentrations measured as ppm or mg/iiter.
^Total calcium (actual plus magnesium equivalent).
-------
Effect of AMD Temperature
All of the tests previously described were conducted with the
feed solution at room temperature. In anticipation of the
lower temperatures which would be encountered at the field
site, runs were carried out with feeds kept at 1.1°C (34°F) and
8.8°C (48°F) (deep well temperature). Synthetic AMD with a
pH of 5 and Fe(II) concentration of approximately 400 ppm was
treated in the same cell as used for pH and iron concentration
studies. Table 8 lists the average voltage, current and
percent reduction in Fe(II) and Fe-p concentrations observed
after steady-state treatment was attained during the 7-hour
runs. At a constant cell current, the degree of AMD treatment
was approximately the same in each case. The only difference
among the three runs is the voltage which had to be applied
to maintain a current of 3 amps. As was expected, the
necessary applied voltage increased with decreasing tempera-
ture. Figure 6 is a plot of the cell voltage versus tempera-
ture of the feed. The best straight line drawn through the
points yields an average temperature coefficient of -0.12
volts/degree, which indicated that temperature variations
to be encountered in the field could be compensated for by
small adjustments in applied voltage.
Solids Generated by Electrolysis
Throughout the course of the laboratory work, it was re-
peatedly observed that during electrolysis of synthetic AMD
at a pH of 3 or greater, a small fraction of the total
solids generated in the cell consisted of what appeared to
be a dark reddish-brown crystalline material that settled
at a rapid rate compared to the usual gelatinous Fe(OH>3
precipitate. Some effort was directed in Phase I to a
closer examination of the properties of this solid and of
the sludge formed after lime neutralization of the electro-
lyzed mine water. Characteristics of the electrolysis
solids were compared with those observed for AMD treated
by aeration.
Synthetic AMD with a pH of 4.7 and Fe(II) concentration of
283 ppm was pumped through a small rectangular cell with
2-inch electrode spacing. The effluent generated during
this treatment had a pH of 3.1 and contained some of the
desired reddish-brown crystalline flakes mixed with hydrous
ferric oxide. Analysis of a portion of the effluent after
33
-------
TABLE 8. EFFECT OF TEMPERATURE
ON AMD TREATMENT
Feed: pH=5.0
FeT=Fe(II)=400 ppm
TEMPERATURE
PERCENT
REDUCTION IN
CONCENTRATION OF
131
34
48
72
131
1.1
8.8
22.2
Volts
18.0
16.9
13.4
Amps
3.1
3.0
3.1
Fe(II)
29
23
27
Fe-j,
-•^_j_
15
16
20
34
-------
18-
17-
cn
4J
o
PL,
0)
O
(U
a
a
< 14-
13-
30
10
15
Slope = -0.12 volts/degree
40 50 60
Temperature, °F
20
70
FIGURE 6. Applied Cell Potential vs Temperature of AMD
35
-------
filtration showed 170 ppm Fe^ and 145 ppm Fe(II). While the
bulk of the collected effluent was stirred to distribute the
solid uniformly throughout the solution, a sample of the
mixture was taken for the usual measurements of settling rate
and sludge volume in a one-liter graduate cylinder. Sludge
density (weight of solids per final volume of sludge) was
determined by weighing the solid after drying at 110°C. A
second sample of the effluent was neutralized with lime and
the aforementioned sludge characteristics were measured.
In order to compare these results with the properties of
the solids generated by air oxidation, AMD with the same
Fe(II) concentration and pH as used for the electrolytic
treatment was aerated for approximately 30 hours until the
composition of the effluent approximated that from the
Ecotrol cell — pH 3.1, 156 ppm FeT and 125 ppm Fe(II). No
crystalline solid, characteristic of the electrolyzed
effluent, was observed in the aerated solution. Samples of
the latter were taken for measurement of sludge properties
before and after lime addition.
Because of the small scale of the laboratory runs, the
amount of total suspended solids per volume of effluent
was quite small. Consequently, in determining settling
rates it was not possible to time the drop in sludge volume
starting from the top of the graduate cylinder, since no
discernible sludge boundary was formed at this height after
stirring was stopped. What could be followed, however, was
the change in volume of a distinct sludge blanket formed at
the bottom of the cylinder. The level of this blanket rose
at first then fell in the normal manner. Sludge volume
versus time curves measured in this way for the electrolyzed
and aerated effluents before lime addition are shown in
Figure 7. Figure 8 contains similar curves for the lime-
treated effluents.
It is quite evident from Figure 7 that the solids generated
in the cell settled more rapidly than those formed during
aeration of the AMD. Although the difference is not as
great in Figure 8 for the lime-treated waters, the settling
rate of the solid in the electrolyzed effluent still appears
to be higher than that for the aerated AMD.
Table 9 contains values of the final sludge volumes (24
hours) and sludge densities measured for the above systems.
36
-------
6.0-
5.0-
e
3
CO / n
4J 4.0-
14-1
O
« 3.0-4
oo
•o
3
r-i
CO
o
a; 2.0-
O
>
1.0.
A
o-o-o.
A
20
0 Electrolysis
Aeration
40 60
Time,t min
80
FIGURE 7
Comparison of Sludge Settling Rates
for Electrolyzed and Aerated AMD
Before Lime Addition
37
-------
16-
12-
-U
fi
Ol
00
1-J
o
8-
4 -
O Electrolysis
A Aeration
Time, min
FIGURE 8. Comparison of Sludge Settling Rates
for Electrolyzed and Aerated AMD
After Lime Addition
38
-------
The data for the sludges present before lime addition con-
cur with the difference in settling rates illustrated by
the curves in Figure 7. The opposite is true for the solids
after lime addition. However, the beneficial effect of the
small amount of crystalline solid in the effluent right after
electrolysis might have been diminished by dilution with
excess Fe(OH)3 formed during the neutralization. Also, it is
possible that upon standing over a long period in neutral or
high pH water, the crystalline solid formed by electrolysis
undergoes transition to an amorphous state.
TABLE 9. SLUDGE CHARACTERISTICS
METHOD OF FINAL VOLUME DENSITY
GENERATION (ml) (mg/ml)
(Before Lime Addition)
Electrolysis 8 16
Aeration 12 12
(After Lime Addition)
Electrolysis 35 20
Aeration 30 24
An air-dried sample of the mixture of solids from the
electrolyzed water was examined by x-ray analysis. The
powder photograph showed that the material was essentially
amorphous. Some scattering occurred, but the lines could
not be clearly identified. In view of the large excess of
Fe(OH)3 in the mixture, the volume fraction of the crystalline
solid might have been below the detectable limit of the instru-
ment (minimum sensitivity, 2 to 3%).
Further work in this area was scheduled for the Phase II
study in anticipation of essentially complete conversions of
Fe(II) and generation of large concentrations of the desired
solids.
39
-------
Cell Evaluation and Sizing
Rectangular Cells
A series of runs were conducted to evaluate the degree of
Fe(II) conversion at various flow rates in a cell with con-
strained bed and with the bed free to expand at relatively
high feed velocities. For this series the rectangular cell
with. 15.2 cm (6 in.) wide major electrodes spaced 10.2 cm
(4 in.) apart and PVC screen against the cathode was
connected into the flow system illustrated in Figure 3.
First, measurements of bed expansion versus fluid velocity
were taken with no DC potential applied to the cell. The
height of the bed after backwash and drain was taken as
the reference point, h0g. After recording the bed height
at flow rates in the range of 3.8 to 11.3 liter/min (1 to
3 gal./min), a plot of percent bed expansion (100A.hB/hjj°)
versus apparent velocity was made. The curve is shown in
Figure 9. It was impossible to obtain precise readings of
bed height, since expansion was not uniform throughout the
bed. This is most likely due to the shape and variation in
size of the Norit particles. Detectable expansion occurred
at a velocity of approximately 42.7 cm/min (1.4 ft/min) As
the velocity was increased above 76.2 cm/min (2.5 ft/min),
the bottom corners of the bed started to compact while the
center started a vortex motion, rising to the top center of
the column, falling along the middle of the walls, and flowing
back into the center in the bottom half of the column.
After extended application of high feed velocities, a slight
degree particle attrition and fines production was noticeable.
Percent expansion was independent of hg° for values of the
latter in the range of 17.8 cm (7 in.) to 45.7 cm (18 in.).
No hysteresis effect was detectable.
Table 10 contains the data obtained during steady-state
treatment of high ferrous AMD at pH 5. The feed velocities
applied in the first three runs were below the threshold
value for bed expansion, and the bed was held in a fairly
compact state by means of a retaining screen as described
earlier. Under these conditions, the percent conversion of
Fe(II) at constant current decreased as the flow rate was
increased from 0.15 to 0.45 liter/min (0.04 to 0.12 gal./min).
The second set of data in Table 10 was obtained by allowing
40
-------
I
45
75
cm/min
8-
CO
G
(0
a
pa
f»p
2 -
2 3
Apparent Velocity, ft/tnin
FIGURE 9. Percent Expansion of Norit Bed in
Rectangular Column
41
-------
TABLE 10. DATA ON AMD TREATMENT AT HIGH FLOW RATES
IN RECTANGULAR CELL
SYNTHETIC AMD FEED a
1
pH
*^
5.3
5.0
5.2
5.6
5.4
4.2
5.5
5.3
Fe
Fe(ll)
i
Fe(lll)
Flow
Rate
(liter/min)
Volts
Amps
% CONV
Fe(ll)
(Compact Bed h ° =17.8 cm)
488
505
515
438
470
-
458
438
480
500
503
438
455
319
453
438
8
5
12
(Free Bed
0
15
-
5
0
0.15
0.23
0.45
hB° =45.7 cm)
6.6
7.6
8.5
8.7
10.6
13.5
10.4
10.6
25.0
25.0
25.0
25.0
25.0
3.0
3.1
3.1
2.7
3.0
1.0
1.7
1.5
24
17
8
3
7
23
21
5
a All iron concentrations measured as ppm or mg/liter.
-------
the bed to expand by about one percent while the feed flow
rate was increased with each run up to 10.6 liter/min (2.8
gal./min ). Separation of the electrode particles by flows
greater than 7.6 liter/min (2.0 gal./min ) is reflected by
the drop in current drawn by the cell at a constant potential
of 25 volts. In Figure 10, the values of percent Fe(II) con-
version are plotted versus flow rate. The bell-shaped curve
shows an optimum flow rate of about 8.7 liter/min (velocity
of 51.8 cm/min ) for Fe(II) oxidation. The appearance of
the maximum in the curve is probably due to a combination of
factors, such as efficient solids removal, optimum separation
of bed particles and optimum residence time. At feed
velocities corresponding to flow rates greater than 8.7 liter/
min (2.3 gal./min ), either the residence time is not sufficient
and/or the separation of the particles, though visibly small,
is great enough so that the behavior of the system approaches
that of a conventional cell without the electrode bed.
The increase in efficiency derived from the operation with
the slightly expanded bed is exemplified by comparing results
from this system in Table 10 with the data obtained using the
compact bed. Although a larger bed volume and higher potential
were applied in the former case, approximately the same per-
cent conversion of Fe(II) was attained at one-third the current
drawn by the compact bed and at about 60 times the mass loading
(8.5 liter/min or 2.25 gal./min for expanded bed versus
0.15 liter/min or 0.04 gal./min for compact bed).
Circular Cells
For the evaluation of the cylindrical configuration, measure-
ments were made first with a small cell containing a static
packed bed of Norit carbon. The height of the bed was 10.2 cm
(4 in.). The diameter of the inner electrode was 1.9 cm
(0.75 in.), while that of the outer electrode was 8.6 cm
(3.4 in.) I.D. Stainless steel was used as the cathode
material and titanium coated with ruthenium dioxide as the
anode. Test solution with pH 3 and 500 ppm Fe(II) was pumped
through the cell at the rate of 57 ml./min (0.015 gal./min ).
Because of the difference in surface areas of the concentric
major electrodes, the main purpose of the runs with the small
cell was to determine any difference in power characteristics
and degree of AMD treatment between the configuration having
inner cathode -- outer anode and vice-versa. Runs were also
43
-------
30-
T
6.0
25-
20-
c
o
•H
co
tl
I
S 15-
0)
10-1
5-
"I—
8.0
2'.0
10.0
Liter/Min
Flow Rate, gal./min
FIGURE 10. Percent Fe(II) Conversion vs Flow Rate in
Rectangular Cell
44
-------
carried out with a PVC diaphragm placed against the cathode
wall.
Table 11 contains data taken during electrolysis in the small
circular cells. Consider first the cells without diaphragm.
A comparison of the current-voltage values indicates that the
cell behaves to some degree like a diode due to the difference
in surface areas of the inner and outer electrodes, i.e. , the
configuration with the small anode conducts better than that
with the small cathode. At the same cell current, the degree
of Fe(II) oxidation was less in the cell with the inner anode.
In addition, a pronounced odor of hydrogen sulfide was gener-
ated from this cell which was not noticed during electrolysis
with the small cathode configuration nor during treatment in
any of the rectangular cells.
With the screens placed against the cathodes , the same
diode effect as described above was observed. The potential
drop across the cells with diaphragm is higher than that with-
out diaphragm, since the screen prevents the semi-conductive
bed from coming in direct contact with the cathode. Thus the
current is lower in the diaphragm cell at a given applied
voltage. Introduction of the diaphragm to the small cathode
cell caused only a slight increase in Fe(II) removal. With
the small anode cell, however, the presence of the screen
doubled the percent reduction in Fe(II) concentration. In
addition, no sulfide odor was detectable, as was the case
without the screen. On the basis of these results, further
laboratory studies with circular cells were performed using
the inner anode-outer cathode configuration with diaphragm
against the cathode.
As was done with the rectangular cells, a series of runs
were conducted to evaluate the degree of Fe(II) conversion
at relatively high flow rates in large cylindrical cells
containing beds without constraints. In order to determine
the percent bed expansion with feed velocity, measurements
were taken using simulated cells made from clear Lucite
tubes. The curve for bed expansion in an annular volume
bounded by an inner tube with 5.1 cm (2 in.) diameter and
outer tube with 20.3 cm (8 in.) I.D. is plotted in Figure 11.
Points for a system with 25.4 cm (10 in.) outer and 10.2 cm
(4 in.) inner tubes fell within the same region as shown in
this figure. Bed expansion was more uniform than in the
rectangular cells. A change from uniform expansion to
45
-------
TABLE 11. DATA ON AMD TREATMENT IN SMALL CIRCULAR CELLS
AMD FEED3
ELECTRODE
CONFIGURATION
Volts
s,
Amps pH Fe-j
(NO DIAPHRAGM)
0.8 2.8 500
Fe(II)
495
6.6 0.8
2.8
503 495
1 % CONV
Fe(III) Fe(II)
8 24
8 16
/ ^
JL 3 * \J
(WITH DIAPHRAGM - PVC SCREEN)
•
0.64 3.0 500 488
0.64 3.0 500 498
12 28
32
iron concentrations measured in ppm or mg/liter.
-------
112-
CO
C
a
x
45
cm/min
i i
60 75
90
7
2 3
Apparent Velocity, ft/min
FIGURE 11.
Percent Expansion of Norit Bed
in Cylindrical Columns
47
-------
turbulent wave motion occurred at a velocity of approximately
83.8 cm/min (2.75 ft./min ).
Electrochemical measurements were made with one cell con-
sisting of an outer stainless steel cathode (20.3 cm or
8 in. I.D.) and inner coated titanium anode (5.1 cm or
2 in. O.D.). A graphite rod was also substituted for the
titanium to compare results. A second cell with outer steel
cathode (25.4 cm or 10 in. I.D.) and inner graphite anode
(10.2 cm or 4 in. D.) was also used. The results with AMD
at pH 2 and 5 are given in Table 12.
All runs except the first one with pH 2 AMD were conducted
with beds free to expand. Percent Fe(II) conversion on the
free beds (hg°=20.3 cm) is plotted versus flow rate in
Figure 12. In view of the results obtained for the rectangular
configuration, a bell-shaped curve was drawn through the
three points. From this curve it appeared that maximum
oxidation occurred at a feed flow rate of about 7.2 liter/min
or 1.9 gal./min (velocity of 24.4 cm/min or 0.8 ft./min ).
At this velocity, the bed was barely expanded. A comparison
of these results with the data for the first run in Table 12
indicates that conversion was higher in the compact bed.
However, it should be noted that 2.5 times the current was
drawn in the latter case. The expanded bed with hg° equal to
50.8 cm (20 in.) yielded a conversion equal to that on the
compact 20.3 cm (8 in.) bed for the same current draw. There
was no difference in the results with graphite and coated
titanium anodes.
The first three runs with pH 5 water showed essentially no
change in the amount of Fe(II) oxidized as the flow was in-
creased from 6.8 to 9.5 liter/min (1.8 to 2.5 gal./min ).
The last three runs with the high pH AMD were conducted with
different initial Fe(II) concentrations to generate a working
curve for scale-up.
Choice of Configuration, Sizing & Materials
for Phase II Pilot Cells
It was difficult to make a choice of configurations based
solely on the raw data obtained during the Phase I study.
For example, consider the numbers in Table 13. For the
treatment of pH 5 AMD at approximately the same flow rate
48
-------
VO
TABLE 12. DATA ON AMD TREATMENT AT HIGH FLOW RATES IN CIRCULAR CELLS3
AMD FEEDb
1
El
2.3
2.3
2.3
2.2
1.9
5.3
5.3
5.7
5.3
5.4
5.8
FeT
450
455
525
388
493
463
480
475
480
265
61
Fe(II)
450
455
525
383
488
455
475
475
475
265
59
1
I Fe(III)
(CELL WITH
0
0
0
5
5
(CELL WITH
8
5
0
5
0
2
Flow Rate
(liter/min)
20.3 cm. I.D.
6.
6.
7.
9.
6.
25.4 cm
6.
8.
9.
8.
8.
8.
8
8
6
5
8
. I.D.
8
5
5
5
5
5
HB"
(cm)
CATHODE)
20.
20.
20.
20.
50.
3
3
3
3
8
Volts
14.0
16.0
15.5
12.5
13.0
Amps
10
4
4
4
10
.0
.0
.0
.0
.0
% CONV
Fe(II)
30
17
18
4
26
CATHODE)
45.
45.
45.
45.
45.
45.
7
7
7
7
7
7
32.0
33.0
32.0
33.0
28.0
44.0
10
8
8
8
6
4
.5
.0
.5
.0
.5
.0
11
11
10
11
27
32
aAll runs except the first were conducted with free beds.
All iron concentrations measured as ppm or mg /liter.
-------
25-
20-
c
o
•H
03
§15
CJ
M
IH
(U
10 -
5 -
6.0
1.5
8.0
Liter/Min
hB° = 20.3 cm(8in.)
2.0
Flow Rate, gal/min
2.5
FIGURE 12. Percent Fe(II) Conversion vs
Flow Rate in Large Circular Cell
50
-------
TABLE 13. COMPARISON OF RECTANGULAR AND CIRCULAR
CELL PERFORMANCE
CELL
Rec.
Cir.
Rec.
Cir.
Rec.
Cir.
LENGTH
OF
CELI
(cm)
45.7
45.7
45.7
45.7
17.8
20.3
FLOW
RATE
(liter/min)
(AMD - pH 5, 500
8.7
8.5
6.8
6.8
(AMD - pH 2, 500
0.45
6.8
RESIDENCE
TIME
(min)
ppm Fe(II)
0.867
2.54
1.11
3.18
ppm Fe(II)
5.05
0.845
APPARENT
VOLUME
(liter)
Bed-Expanded)
7.53
21.7
7.53
21.7
Bed-Compact)
2.29
5.77
AMPS
1.7
8.0
2.7
10.5
3.1
10.0
%
CONV
Fe(II)
23
11
3
11
24
30
-------
(8.7 liter/rain or 2.3 gal./min ) in rectangular and circular
cells of equal length and with approximately the same major
electrode spacing, the percent reduction in Fe(II) concen-
tration in the rectangular cell was twice that observed for
the circular cell despite the higher current drawn by the
latter and longer residence time in this higher volume cell.
When the same AMD was pumped through the same two cells at a
flow rate of 6.8 liter/min (1.8 gal./min ), the difference
in performance was not as great as in the first case. The
percent reduction in Fe(II) concentration by the circular
cell was 3.7 times that in the rectangular system. However,
the circular cell drew 3.9 times the current. A consideration
of residence time would still let the rectangular cell appear
to be somewhat better, but the exact degree of improvement
cannot be estimated. For runs with the pH 2 AMD, the reverse
was found to be true. The degree of Fe(II) oxidation was
slightly higher in the circular cell. However, although
this cell drew 3.2 times the current in the rectangular cell,
a consideration of the higher mass loading in the circular
system and the higher residence time in the rectangular cell
would make the circular configuration with compact bed appear
to be superior for treating low pH AMD.
After examining the raw data, calculations were made for
scale-up to pilot size (18.9 liter/min or 5 gal./min
average, 500 ppm initial iron, 10 ppm or lower final iron)
to compare dimensions and gain some idea of the design of the
Phase II system depending on cell configuration. The total
residence time or length of cell required for the desired
treatment was determined from working curves such as those
in Figure 5. Figure 13 shows a comparable working curve
plotted from the data in Table 12 on treatment of pH 5 AMD
with different initial iron concentrations in the large
circular cell. In those cases where the data was not
sufficient to permit plotting of actual working curves, an
average percent removal of Fe(II) per unit length cell based
on available numbers was taken as a guiding factor. After
the required total length of cell was determined for each
configuration and flow rate employed in the laboratory studies,
the dimensions required for a 18.9 liter/min (5 gal./min )
loading were calculated.
Based on the results of such calculations and a considera-
tion of factors such as availability of components with the
desired size, ease of construction, and timing, the circular
52
-------
1000
H
H
s^
-------
cell was found to be the preferred configuration.
For sizing purposes, the laboratory data on treatment of
pH 5 AMD was used, since the rate of Fe(II) oxidation was
lower in this solution than in the pH 2 water. From cal-
culations based on the working curve in Figure 13 and in-
formation on the diameter of commercially available major
electrode materials, a basic unit cell was designed which
is 45.7 cm (18 in.) long and has a 10.2 cm (4 in.) diameter
inner anode enclosed by an outer cathode cylinder with
25.4 cm (10 in.) I.D. (7.6 cm or 3 in. spacing of major
electrodes). Initial plans called for six of these unit
cells to be used for preliminary testing in the field. Then,
12 additional units were to be added to provide the cal-
culated total contact time required for the desired degree
of treatment.
The total required length of cell was broken down into
45.7 cm (18 in.) segments to provide flexibility in testing
and operation. Also, it was thought that advantage could be
taken of a known multiple entry effect, whereby a greater
net conversion of Fe(II) could occur if the AMD is passed
through a number of unit cells than if the solution were
passed through a single cell with a length equal to the
total of the unit cells. The unit cells were provided with
the means for treating the AMD on a compact and expanded bed
and for adding the PVC-screen diaphragm described throughout
this report.
As for the choice of material for the major electrodes, it
was decided that graphite would be used for both cathode and
anode. Since no data was obtained which suggested that
coated titanium would yield superior results, graphite was
the immediate anode candidate in view of the lower cost of
this material. All of the laboratory data were measured with
cells having stainless steel cathodes. However, plans were
made to elaborate on the previously discussed polarity
reversal tests during initial systems testing in the field.
Use of the all graphite cell would enable this to be
accomplished without having to dismantle the cells to ex-
change an outer steel electrode, for example, with an outer
graphite electrode. The latter procedure would be a required
operation in a mixed electrode cell in order to eliminate
anodic dissolution of iron from the steel surface. On the
basis of the laboratory evaluation, Norit RB-II pellets were
54
-------
chosen as the electrode bed media in the pilot cells.
From a comparison of the size of the Phase II cells and the
voltage-current values applied to the large circular cell
during the laboratory work, a D.C. power supply providing
a maximum of 50 volts and 500 amps was considered to be
ample for the field study.
55
-------
SECTION V
PHASE II - FIELD STUDIES
The Delaware Pumping Station located in Wilkes-Barre,
Pennsylvania was selected by the D.E.R. as the field site
for pilot plant operation. Prior to the June 1972 flood
generally four 9462 liter/min (2500 gal./min ) pumps had
been continuously operated to prevent surface subsidence
and basin flooding. Three of the pumps were used to draw
water from a depth of about 152 m (500 ft ). The fourth
was operated from a 76.2 m (250 ft ) depth in order to
discharge AMD with a lower iron content. Table 14 lists
the composition of deep and shallow well samples taken in
January 1972.
As indicated by this data, the Delaware Pumping Station had
provided the opportunity for evaluating treatment of high
and low ferrous AMD. Since the pH values of the waters were
relatively high and did not differ greatly between the two
levels, provision was also made during the planning stage of
Phase II for incorporating an acid feed - pH control unit
into the pilot plant to permit observation of the treatment
of pH 2 AMD.
As soon as the restriction on operating the pumps was lifted
by the D.E.R. in August after the flood, samples of the deep
well water were analyzed for Fe(II) and Fe^. The total iron
concentration was a few parts per million higher than the
ferrous content which averaged about 52 ppm. This value is
approximately an order of magnitude lower than the usual
numbers observed before the flood occurred. It was thought
that if the deep pump were kept running over a sufficient
period, the Fe(II) content would increase to the normal value.
However, as seen in Figure 14, the Fe(II) content of the raw
AMD in shaft #2 did not recover to pre-flood values over the
period of the Phase II work. In fact, during project down
time (November to May) when the pumps were not operated, the
Fe(II) concentration fell below 50 ppm.
In order to evaluate treatment of high ferrous AMD, therefore,
the initial plant was modified by adding a system for metering
concentrated ferrous sulfate into the raw AMD, similar to what
was done during the laboratory studies.
56
-------
TABLE 14. ANALYSIS OF AMD FROM
DELAWARE PUMPING STATION
pH
Alkalinity
Acidity, pH4
PH8
FeT
Fe(II)
Mn
Ca
Mg
Al
Na
so4
Turbidity
Total Solids
Suspended Solids
Settleable Solids
Specific Conductance
(micromhos/cm)
PUMP NOS.
1, 2 & 3
4.9
11
0
765
355
360
49
259
68
3.5
115
2700
50
5085
120
0.025
3750
P P NO. 1 DH
RAISED
5.9
101
0
40
60
60
14
109
18
0.1
32
1200
30
1990
30
0.025
2150
Samples taken January 25, 1972, and analyzed by
Department of Health - Division of Sanitary Engineering,
Commonwealth of Pennsylvania; values in units of ppm
or mg/liter unless specified otherwise.
57
-------
00
I so
d
o
•H
8 60
•u
d
u
d
o
40-
20-
i i
8/21
—o.
I I
1972
Delaware Pumping Station
.Q Q
.0-
.0^-0—0
^ I I P_J I I I I I I I I
11/r J 5/1' 1973 6/11
Time , weeks
FIGURE 14. Variation in Deep Well AMD Composition with Time
-------
In addition to electrochemical processing, the AMD was
treated by aeration in order to develop comparative economics.
Aeration runs were carried out in the Operation Yellowboy Van,
a mobile pilot plant built for this purpose by Dorr-Oliver
under a grant from the Commonwealth of Pennsylvania31.
Description of Pilot Plant
Figure 15 shows a schematic diagram of the Phase II system.
Raw AMD was pumped into a 568 liter (150 gallon) stainless
steel tank for mixing either with 10% sulfuric acid and/or
1.57o ferrous sulphate solution if a change in pH or composition
were required. The pH was controlled by means of a Universal
Interloc, Inc. Controller (Model 1000) in conjunction with a
Precision Products metering pump, (Model 9701-21). The con-
centrated ferrous sulphate solution was added with an identical
metering pump adjusted as required.
For evaluation of electrolytic treatment the AMD was fed to
the bank of cells and the effluent subsequently treated with
lime in a flash mixer housed in the Yellowboy Van. During the
aeration runs, AMD feed was pumped directly to the flash
mixer for pH adjustment with lime prior to treatment in the
aeration tank also housed in the van. Samples for analysis
were taken from the overflow of the feed pH adjust tank, the
cells effluent sump, and from the overflow of the aeration
tank.
The raw AMD was delivered to the system by means of a 37.8
liter/min (10 gal./min ) Jabsco Positive Displacement Pump.
The cells feed and effluent pumps were of the centrifugal
type, 37.8 liter/min (10 gal./min ), A.O. Smith Corporation.
Plumbing consisted mainly of one inch PVC pipe, fittings and
ball valves. Calibrated Fischer & Porter Rotameters were
used for monitoring the flow.
Figure 16 shows a cutaway view of a unit cell. The outer
electrode consists of electrolytic graphite pipe (Union
Carbide, 45.7 cm or 18 in. high, 25.4 cm or 10 in. I.D. ,
2.54 cm or 1 in. wall) coated on the outside surface with
phenolic resin to prevent seepage. A four inch diameter
graphite rod served as the central electrode. To provide
electrical contact, a copper strap was bolted to the outer
shell, and a copper rod was threaded into the central electrode,
Not shown in Figure 16 are the PVC screen diaphragm which was
59
-------
LINE FROM
SHAFT PUMP
ON
O
EFF.
SUMP
ELECTROLYTIC CELLS
AERATION
TANK
SURGE TANK
WITH LEVEL
CONTROLS
PH AND
Fe H ADJUST
LIME
SCREW
FEEDER
FLASH MIXING
TANK
BLOWER
TO
DISPOSAL
CONC.
FeS04
FIGURE 15. Flow Diagram of Pilot Plant
-------
EFFLUENT
PLENUM \
COPPER
BUSSINGX
EXIT
GRAPHITE
CYLINDER
OUTER
ELECTRODE
COPPER
BUSS STRAP
GRAPHITE ROD
INNER ELECTRODE
BED PARTICLES
PERFORATED PVC
DISTRIBUTOR PLATE
RUBBER
GASKET
INLET TO
PVC PLENUM
CHAMBER
FIGURE 16.
Unit Cell Used in Pilot Plant
61
-------
fitted against the cathode and a top restraining screen which
was used when runs with a compact bed were conducted. For
convenience in initial systems testing, six of the unit cells
were connected for series hydraulic operation with the flow
upward in each cell. This system is shown in Figure 17.
For final systems operation six three-stage units were used
in series (Figure 18). Staging was accomplished by replacing
the top plenum assembly shown in Figure 16 with PVC flanges
and using one central graphite rod common to all three
sections. The intervening flanges could be fitted with
distributor plates for bed support in each section in order
to attain the multiple entry effect, or the effect of one
long compact bed could be studied by removing these distribu-
tors.
For all runs the cells were connected electrically in parallel.
Power was supplied by means of a Rapid Electric Rectifier,
0-48DCV/500 A. The rectifier and central control panel are
shown in Figure 19.
Operational Data
Problems were encountered when attempts were made to dissolve
technical grade ferrous sulphate in water for making up the
concentrated Fe(II) solution to be metered into the raw AMD.
Vigorous mixing, addition of acid, and heating would not pro-
vide the required Fe(II) concentrate for high ppm feed. It
was decided to use the more costly reagent grade material
which dissolved readily. Because of the prohibitive expense
involved in the addition of large quantities of this salt,
the decision was made to conduct all Phase II runs using the
AMD at its natural composition and with Fe(II) concentration
adjusted to 250 ppm instead of 500 ppm as originally planned.
Initial Systems Testing
Initial runs were conducted using the 50 ppm Fe(II) mine
water with pH adjusted to 2. Feed was pumped through indi-
vidual cells shown in Figure 16 while observations of per-
cent Fe(II) conversion versus cell current were made. With
this system a reevaluation of the effect of major electrode
polarity was conducted. The resulting data plotted in Figure
20 confirmed observations made during Phase I that conversion
is higher when the central graphite electrode is anodic and
the PVC screen is placed against the cathode wall.
62
-------
FIGURE 17. Operation with Six Unit Cells
-------
FIGURE 18. Complete Electrochemical System
-------
"
•• n
FIGURE 19. Power Section of Electrochemical Plant
-------
Feed
c
o
•I-<
en
H
C
o
0)
40-
30-
20-
10-
pH 2
Fex & Fe(II) = 50 ppm
18.9 liter/min (5 gal./min)
Unit Cell
i
10
i
15
Current, amps
FIGURE 20. Effect of Electrode Polarity on Fe(II) Conversion
-------
The next series of runs were performed to further investigate
the difference between a compact and expanded bed in the
circular cells. For these tests four 3-stage cells were used,
two with compact beds and two with sectionalized beds with
space provided for bed expansion. Feed with 250 ppm Fe(II)
and pH 2 was processed at 3.8 and 18.9 liter/min (zero and
one percent expansion of the free beds). As shown by curves
in Figure 21, at the limiting current values the percent
conversions of Fe(II) on the compact and loose beds were
approximately equal at 3.8 liter/min (1 gal./min ), a flow
rate below the threshold value for bed expansion. For the
18.9 liter/min (5 gal./min ) flow (Figure 22) the expanded
bed effected about 25% higher conversion than the compact
system. Also, the optimum conversion at 18.9 liter/min
(5 gal./min ) with the expanded bed was approximately equal
to the maximum conversion at 3.8 liter/min (1 gal./min ).
The latter result further emphasized the beneficial effect
of bed expansion in view of the higher mass loading with
the 18.9 liter/min (5 gal./min ) flow.
Finally, the complete system shown in Figure 18 was assembled.
In view of the favorable results obtained with the previous
runs, provision was made for bed expansion in each of the 18
unit cells. With all six columns in operation tests were
conducted with 250 ppm (Fe(II)-pH 2 AMD'at flow rates of
3.8 liter/min (1 gal/min ), 11.4 liter/min (3 gal./min ),
18.9 liter/min (5 gal./min ), and 26.5 liter/min (7 gal./min )
The data obtained for the first three flow rates are plotted
in Figure 23 as percent Fe(II) conversion versus the applied
current, expressed as a percentage of the theoretical current
required for complete conversion of 250 ppm ferrous iron™.
Maximum conversion at 3.8 liter/min (1 gal./min.) was slightly
higher than at 18.9 liter/min. (5 gal./min ) as expected
based on a consideration of contact time. However, conversion
at 11.4 liter/min (3 gal./min ) was less than at 18.9 liter/
min. The optimum combination of residence time and degree of
bed expansion most likely occurred at 18.9 liter/min , the
designed loading for this system, since at 26.5 liter/min
(7 gal./min) an Fe(II) conversion higher than 10% could not
be attained regardless of the power applied.
Although conversion was maximized at 5 GPM as per scale-up,
the highest degree attained with the 250 ppm Fe(II) feed was
76%. One possible reason for the lower than expected con-
version is the occurrence of back reduction of Fe(IlI) by
67
-------
Feed
pH 2
90 H FeT«Fe(II) = 250 ppm
3.8 liter/min (1 gal./min)
c
o
•H
« 70-
-------
o\
90 -
d
•3 70
03
0)
c
o
H
M
50 -
30 -
Feed
pH 2
FeTfc* Fe(II) - 250 ppm
18.9 liter/min (5 gal./min)
100
Six Unit Cells - Series Flow
Bed
• Compact
O Expanded
200
300
400
FIGURE 22
Current, amps
Effect of Bed Compaction for Nineteen Liter Per Minute Feed
-------
90-
Feed
pH 2
FeT & Fe(II) = 250 ppm
ti
o
•H
CO
V4
(U
70-
0)
50-
30-
0-
€
18 Unit Cells - Series Flow
Bed - Unrestrained
Flow Rate
• 3.8 liter/min (1 gal./min)
D 11.4 liter/min (3 gal./min)
Q 18.9 liter/min (5 gal./min)
—T~
300
1
100
1 r^
200
70 Theoretical Current
FIGURE 23. Effect of Flow Rate on Fe(II) Conversion
-------
hyrdrogen generated in the pH 2 water. In order to decrease
the volume of hydrogen accumulated as the liquid flowed
through the cells, vent pipes were installed in the exit
line of each cell. Repetition of the 18.9 liter/min (5 gal./
min ) run with this modification resulted in an increase in
percent Fe(II) conversion to 85. This figure represents the
highest degree of Fe(II) oxidation attainable by electrolysis
alone during this study.
Final Electrolysis and Aeration Runs
In order to generate cost figures for electrochemical treat-
ment, pH 2 and pH 5 AMD with Fe(II) concentrations of 40 and
250 ppm were processed at the rate of 18.9 liter/min. (5 gal./
min ). Comparative 18.9 liter/min aeration runs were con-
ducted with the high ferrous AMD at pH 2 and 5.
The results of electrochemical treatment followed by neutrali-
zation are shown in Table 15. Table 16 contains the data
obtained by aeration after neutralization. Except for the
Al, Mn, Ca, and Mg values, the numbers represent averages of
many measurements taken over a forty-hour period after an
equilibrium condition had been reached during electrolysis
treatment, and over a twenty-hour period in the case of the
aeration series. Toward the end of each run the feed, cell
effluent, neutralization and neutralization-aeration overflows
were samples for Al, Mn, Ca, and Mg analyses. These were
performed by personnel in the Water Quality Laboratory of
Wilkes College in Wilkes-Barre. All samples including the
neutralization overflows were filtered through medium porosity
paper before analysis.
Considering first the results of electrolytic processing
alone (feed and cell effluent values in Table 15) , it is
seen that for treatment of the low ferrous AMD, the cell
system as designed produced an expected effluent containing
less than 10 ppm Fe(II). The results for the high ferrous
waters are short of the design value especially in the case
of the pH 5 AMD. As discussed previously, the lower con-
version in the high pH system is believed to be due to the
deleterious coating of ferric hydroxide on the electrode
bed. The 18.9 liter/min (5 gal./min ) liquid flow apparent-
ly was not vigorous enough to aid in expelling the solids from
the cells.
71
-------
TABLE 15. AMD TREATMENT BY ELECTROLYSIS AND NEUTRALIZATION
AMD*
CELL EFFLUENT*
NEUTRALIZATION OVERFLOW*
N>
' ' , 1 7. CCUJV '
£• {!& B& Fe(II) F«(III) Al Mn Ca Jig, Fedl) 2H
2fl Fe(II) Fe(III) Al Mi
(18.9 liter/rain , 7.5 Volts, 220 Amps)b
2.0 38.8 2.2 0.92 4.1 140.0 97.9 2.0 5.0 36.0 0.34 3.9 (300) 79.9
(18.9 liter/rain , 10 Volts, 350 Amps)
2.0 250.0 5.0 1.0 4.2 95.0 85.0 2.0 37.0 218.0 0.54 3.8 (130) 80.0
(18.9 Htee/mtft , 12.5 Volts, 350 Amps)
5.4 35.0 0 0.35 3.5 100.0 94.9 5.4 7.5 27.5 0.20 3.3 (110) 95.0
(18.9 llter/mln , 28 Volts, 380 Amps)
5.4 245.0 10.0 1.0 4.6 100.0 102.0 4.8 110.0 28.0 0.96 4.3 84 104.0
Fe(III) Al
Ca
(Lime Feed Kate, 1.3 g /liter)
87.5 7.6 0 0 1.12 1.9 500 87.6
(Lime Feed Rate, 2.2 g /liter)
84.8 7.3 0 0 0.54 1.6 1,000 125.0
(Llm« Feed R*te, 0.38 g /liter)
78.0 7.1 0 0 0.52 1.5 330 98.3
(Lime Feed Rste, 0.66 g /liter)
55,2 7.1 0 0 0.52 2.3 240 148.0
*A11 metal concentrations measured as ppm or mg /liter.
bConver«ion factors:
liters x 0.264 ! gallons
grams/liter x 0,00834 I pounds/gallon
-------
TABLE 16. AMD TREATMENT BY NEUTRALIZATION AND AERATION
AMD8
NEUTRALIZATION - AERATION OVERFLOW3
»' Lime Aerator *
Flow Feed Blower
(liter/min ) £H Fe(II> Fe(III) Al Jfa Ca Mg (g /liter)
-------
After the runs with the pH 5 waters were completed, attempts
were made to clean the beds by pumping dilute sulfuric acid
through the cells while periodically reversing the polarity
of the major electrodes. Analysis of the effluent acid
solution showed mainly ferric and very little ferrous iron
in the wash. Following this treatment, pH 5 high ferrous
AMD was passed through the cells while no D.C. potential was
applied to the electrodes. When the effluent pH and Fe(II)
concentration equaled the feed values, power was applied.
Analysis of the effluent during a few hours' operation at
18.9 liter/min (5 gal./min ) showed a significant improvement
in Fe(II) conversion over the value indicated by the data in
Table 15. These observations suggest that for electrolytic
treatment of high pH AMD, a periodic bed maintenance procedure
as described above might have to be included in the operation
of a large-scale plant in order to maintain the required
degree of Fe(II) oxidation.
It is difficult to draw definite conclusions regarding the
effect of electrolysis on aluminum and manganese in view of
the low concentrations of these metals in the raw AMD.
However, it appears from the analytical data that reductions
in aluminum concentration of 40 to 60% might have been attained
except during treatment of pH 5 - high ferrous AMD when,
along with the relatively low Fe(II) conversion, the reduction
in aluminum value dropped to 4%. Manganese concentrations
were reduced by 5 to 107o in accordance with the laboratory
results. No explanation other than analytical error can be
offered for the apparent anomaly of higher calcium concentra-
tions in the cell effluent than in the feed. In view of the
fact that lime was used for neutralization, this anomaly was
not further investigated.
Natural aeration of the cell effluent during lime addition and
neutralization resulted in complete oxidation of residual
ferrous iron, as this species was not detected in the
neutralization overflow (see Table 15). Also, essentially
all of the iron was removed from the overflow by filtration.
Manganese concentrations were further reduced during
neutralization. The increase in aluminum values is believed
to be due to impurities in the lime used.
Approximately 10 times the stoichiometric amount of air was
fed to the mine water during the neutralization-aeration
runs (Table 16). Both ferrous and ferric iron were detected
74
-------
in the filtered overflow. Manganese concentrations were
reduced by 50%, while the aluminum values increased with lime
addition.
A comparison of the lime feed rates for high ferrous AMD
in Tables 15 and 16 shows that the lime demand for neutraliza-
tion after electrolysis was higher than for aeration, more so
for the pH 2 than for the pH 5 runs. In this regard, it
should be noted that two different batches of lime were used
for these tests. An older lot which had been stored in the
Yellowboy Van from previous projects was depleted during the
electrochemical runs. A new batch obtained from the D.E.R.
was used for the aeration studies. It is possible that the
older lot had deteriorated on standing. Thus, the apparent
higher lime demand of the electrolyzed water. In addition to
this factor, the significantly higher lime uptake of the
electrolyzed versus aerated pH 2 AMD most likely occurred
as a consequence of pH adjustment of the raw mine water.
This was done essentially by pH titration. Since the re-
quired pH of 2 lies on the standard acid-base titration curve
in the region where significant changes in volume of added
acid have only a small effect on pH, more acid might have
been used in adjusting the pH of AMD treated in the cells
than added to the mine water treated by aeration, which was
done on a different day.
During electrochemical treatment of the pH 5 AMD, the dark
red and apparently crystalline solid observed during the
laboratory runs was formed with the usual ferric hydroxide
precipitate. However, the volume fraction of the desired
solid generated in the actual AMD was no greater than found
in the laboratory. X-ray patterns of the isolated solid
again showed no evidence for a crystalline material.
Characteristics of the sludges formed after lime treatment of
the electrolysed and aerated AMD were compared. Although
final volumes and percent solids (9.7) of the settled sludge
were approximately the same for each case, the curves in
Figure 24 indicate a definite improvement in sludge settling
rate when the AMD was electrolysed before lime treatment.
75
-------
80 -
0)
CO
•u
O
H
o
6>e
\~S
(!)
00
3
CO
O
OJ
o
>
60 -
O Electrolysis-Neutralization
(pH 7.1)
Aeration-Neutralization (pH 7.5)
40 -
20-
Fe(II)JS;250 ppm
•o-
I
5
I I
10 15
Time, min
20
FIGURE 24. Sludge Settling Rates
76
-------
SECTION VI
COST ANALYSIS AND DISCUSSION
Based on the data in Tables 15 and 16, estimates were made
of the capital and operating costs for 473 liter/min (125
gal./min ) plants treating pH 2 AMD (95% oxidation of Fe(II) )
by electrochemical oxidation and by aeration. It is realized
that some degree of lime tratement must precede aeration in
order to provide the appropriate pH for rapid oxidation of
Fe(II). However, in view of the lime demand discrepancy
between the pH 2 runs discussed earlier and since costs of
lime neutralization for both treatment systems should be
approximately the same, these costs were not included in the
present analysis. Further bases for scale-up are that the
cells system be constructed as one piece of equipment and
that power could be scaled up linearly from the 18.9 liter/
min (5 gal./min ) data. Tables 17 and 18 show the capital
cost estimates for electrochemical treatment and aeration
respectively. Operating costs are compared in Table 19. As
indicated by these figures, the electrochemical system is
more expensive in fixed capital requirements and in operating
costs.
The difference in operating costs between the two processes
can be lessened if total treatment including neutralization
is taken into account. It is known that whereas high ferric
AMD can be satisfactorily neutralized with limestone,
neutralization-aeration of high ferrous drainage requires
the addition of the more expensive lime in order to attain
the appropriate pH for rapid oxidation of ferrous iron^O.
Limestone could not be used for this project because of the
nature of the apparatus in the Yellowboy Van (lime screw
feeder). However, limestone is the preferred material for
neutralizing the effluent from the electrochemical cells.
Although hydrogen production was not studied during the
present project, further cost savings for operation of a
large-scale electrochemical plant might ensue from the
recovery and sale of by-product hydrogen.
It should be mentioned here that based on scale-up calculations
from laboratory data obtained with small packed bed electrode
cells, engineers at Tyco Laboratories have estimated the cost
of a 378 liter/min (100 gal/min ) plant to be about 1/5 the
figure shown in Table 17 ^0. However, in view of the large
77
-------
TABLE 17. CAPITAL COST ESTIMATE FOR
ELECTROCHEMICAL SYSTEM
(473 liter/min, 125 gal./min)
Purchased Equipment
Cells & Pump $32 ,500
Installation @ 22% 7,150
Piping @ 40% 13,000
Instrumentation @ 5% 1,625
$ 54,275
Electrical
DC Power (100 KW
Installed) 2,600
Buss Bars Installed 1,300
Plant Electrical @ 10% 3.250
7,150
Buildings @ 35% 11,375
Land and Improvements @ 10% 3 ,200
Carbon Bed @ 50$/lb 5,400
Physical Plant Cost 81,400
Engineering and Construction @ 25% 20,350
Direct Plant Cost 101,750
Contractor's Fee @ 8% 8,140
Contingency @ 15% 15,260
Fixed Capital $125,150
BASIS: pH 2, 250 ppm Fe(II) , 95% Conversion.
Cells system designed as one piece of equipment.
Linear power scale-up.
Neutralization equipment not included.
78
-------
TABLE 18. CAPITAL COST ESTIMATE FOR
AERATION SYSTEM
(473 liter/min, 125 gal./min)
Purchased Equipment
4,000 gal. Capacity Tank $ 5,000
Pump 1,000
Aerator 1,000
$ 7,000
Installation (3 22% 1,540
Piping @ 40% 2,800
Instrumentation @ 5% 350
Electrical @ 10% 700
Building @ 35% 2,450
Land and Improvements @ 10% 700
Physical Plant Cost 15,540
Engineering and Construction @ 25% , 3.885
Direct Plant Cost 19,425
Contractor's Fee @ 8% 1,555
Contingency @ 15% 2,915
Fixed Capital $23,895
BASIS: pH 2, 250 ppm Fe(II) , 95% Conversion.
Neutralization equipment not included.
79
-------
TABLE 19. ESTIMATES OF OPERATING COSTS
FOR ELECTROCHEMICAL AND
AERATION SYSTEMS
(473 liter/min, 125 gal./min)
Raw Materials (Annual Bed
Replacement)
Power
Maintenance and Supplies
@ 5% Fixed Capital
Labor
Indirect Labor @ 40%
Electrochemical
Aeration
Oxidation
$ 2,000
13,480
6,260
12 ,000
4,800
Aeration
_
5,200
1,200
12 ,000
4,800
Total Annual Costs
$38,540
$23,200
c/1,000 gal.
71.4
43.0
BASIS: 95% Conversion, pH 2 AMD
24-hour day, 300 days/year.
Labor, 4 men at 407» time to both systems.
Neutralization costs not included.
80
-------
scale-up factor involved (1.25 liter/min or less to 378 liter/
min ), pilot plant tests with a 378 liter/min system should
be conducted.
The foregoing discussion pertains to the removal of iron from
AMD. There are currently no regional requirements for manganese
and aluminum concentrations. However, it is expected that most
states experiencing the AMD problem will adopt the proposed EPA
standard of less than one ppm total heavy metals. The effluents
generated during total electrochemical-lime treatment of the
Delaware Pump water showed a reduction in manganese levels
from 4 to 2 ppm, but most of this occurred in the neutraliza-
tion step. The apparent higher reduction in aluminum concentra-
tion must be further substantiated in view of the low feed
values encountered in this case. Studies should be conducted
with AMD containing these metals in much higher concentration
than found in the Delaware water in order to obtain more
definite data on the electrolytic removal mechanism. The
deleterious effect of manganese on Fe(II) oxidation observed
during the laboratory study should also be investigated further
in regard to the treatment of AMD containing high manganese
levels.
The elusive "crystalline" solid generated during electrolysis
of AMD at pH 3 or higher continues to be merely a laboratory
phenomenon. Formation of small quantities of this solid in
a laboratory cell was clearly demonstrated before the EPA and
DER project officers during an early meeting on the progress
of Phase I activity. However, all attempts to generate large
amounts of the material and to identify it were not successful.
Formation in significant quantities might require closer
control of operating conditions. The solid may be a coagulated
form of the hydroxide, which by a mechanism involving, for
example, some appropriate value of zeta potential, can only
exist in an electric field.
It is not known whether the presence of small amounts of the
crystalline solid or some other effect is responsible for the
noted higher settling rate of the sludge generated by lime
neutralization of the cell effluent. Whatever the explanation,
further investigation in this area is warranted, as anyone
familiar with the usual AMD neutralization sludges can well
appreciate a process involving simultaneous oxidation of Fe(II)
and conversion to a rapid settling and readily filterable form.
The application of operating conditions which promote formation
81
-------
of the desired solid should effect an improvement in the
electrochemical oxidation of Fe(II) in high pH AMD by
eliminating deleterious coating of the electrode bed by the
usual gelatinous hydroxide precipitate.
In conclusion, from scale-up estimates for a 473 liter/min
(125 gal./min ) loading based on data obtained from the pilot
tests, this electrochemical system does not appear to offer
a more economical approach to AMD treatment than conventional
neutralization-aeration. In connection with other applica-
tions related to AMD treatment, work is currently being done
in this laboratory on re-evaluation of scale-up methods and
on the engineering of improved high-flow cells for efficient
bed distribution, gas expulsion, and precipitate removal.
Sectionalized power application is also being considered.
During the AMD pilot runs, all of the cells were connected
in parallel to the same rectifier. However, as the solution
flows through the cells, voltage and especially current
requirements decrease because of a decrease in the Fe(II)
concentration. The use of multi-output rectifiers to provide
the appropriate power to different sections of the cell
system should result in operating costs savings which would
more than offset the added capital expense associated with
these rectifiers.
Assuming that developments currently under study will show
significant improvements over the economics that were
generated during this project, a 378 liter/min (100 gal./min )
portable plant should be built and operated for several
months at sites providing high and low acidity AMD with high
and low ferrous, manganese and aluminum concentrations. In
addition to the bed regeneration technique described in a
previous section of this report, removal of high pH precipi-
tate by purging with an inert gas should be evaluated as a
periodic maintenance procedure. Because of the complexities
in scaling up an electrochemical system of this type, the
data generated during operation of the 378 liter/min (100
gal./min ) plant will provide a firmer base than the 18.9
liter/min (5 gal./min ) figures for economic predictions
involving a full-scale plant, e.g., a 22,710,000 liter/day
(6,000,000 gal./day) system.
82
-------
SECTION VII
REFERENCES
1. Bituminous Coal Research, Inc. , Mine Drainage Abstracts-
A Bibliography, 1964-1970.
2. Applied Science Laboratories, Inc., Purification of Mine
Water by Freezing, Environmental Protection Agency,
Research Series 14010 DRZ 02/71, Washington, D.C. , Feb-
ruary, 1971.
3. Black, Sivalls, and Bryson, Inc., Evaluation of New Acid
Mine Drainage Treatment Process . Evironmental Protection
Agency, Research Series 14010 DYI 02/71, Washington, D.C.,
February, 1971.
4. Culligan International Co., Acid Mine Drainage Treatment
By Ion Exchange, Environmental Protection Technology
Series, EPA-R2-72-056, Washington, D.C. , November, 1972.
5. Rex Chain Belt, Inc., Treatment of Acid Mine Drainage by
Reverse Osmosis , Federal Water Quality Administration,
Research Series 14010 DRY 03/70, Washington, D.C., March,
1970.
6. Gulf Environmental Systems Co., Acid Mine Treatment Using
Reverse Osmosis, Environmental Protection Agency, Research
Series 14010 DYG 08/71, Washington, D.C. , August, 1971.
7. Hill, R. D. , Wilmoth, R. C., Scott, R. B. , Neutrolosis
Treatment of Acid Mine Drainage, 14010 05/71, Paper
presented at the 26th Annual Purdue Industrial Waste
Conference, LaFayette, Indiana, May, 1971.
8. Rex Chainbelt, Inc., Reverse Osmosis Demineralization of
Acid Mine Drainage. Environmental Protection Agency,
Research Series 14010 FQR 03/72, Washington, D.C., March,
1972.
9. Bituminous Coal Research, Inc. , Sulfide Treatment of Acid
Mine Drainage. Federal Water Pollution Control Administra-
tion, Research Series 14010 DLC (PB 187 866), Washington,
D.C., November, 1969.
83
-------
10. Horizons Inc., Foam Separation of Acid Mine Drainage«
Federal Water Quality Administration, Research Series
14010, FUI 10/71, Washington, B.C., October, 1971.
11. Catalytic Inc., Neutradesulfating Treatment Process for
Acid Mine Drainage. Environmental Protection Agency
Research Series 14010 DYH 12/71, Washington, D.C.,
December, 1971.
12. University of Utah, Removal of Dissolved Contaminants
from Mine Drainage, Environmental Protection Technology
Series, EPA-R2-72-130, Washington, D.C. , December, 1972.
13. Mihok, E. A., Deul, M. , Chamberlain, D. E., and Selmeczi,
J. G., Mine Water Research-The Limestone Neutralization
Process.. Report of Investigators 7191, U. S. Department
Of the Interior, Bureau of Mines, September, 1968.
14. Bituminous Coal Research, Inc. , Studies on Limestone
Treatment of Acid Mine Drainage. Federal Water Pollution
Control Administration, Research Series 14010 EIZ 01/70
DAST-33, Washington, D.C., January, 1970.
15. Wilmoth, R. C. and Hill, R. D., Neutralization of High
Ferric Iron Acid Mine Drainage. Federal Water Pollution
Control Administration, Research Series 14010 ETV 08/70,
Washington, D.C. , August, 1970.
16. Wilmoth, R. C., et.al. , Limestone Treatment of Acid
Mine Drainage. 14010 10/70, Paper presented at 1970
Society of Mining Engineers Meeting, St. Louis, Missouri,
October, 1970.
17. Bituminous Coal Research, Inc., Studies of Limestone
Treatment of Acid Mine Drainage Part II. Environmental
Protection Agency, Research Series, 14010 EIZ 12/71,
Washington, D.C., December, 1971.
18. Johns-Manville Products Corp., Rotary Precoat Filtration
of Sludge from Acid Mine Drainage Neutralization,
Environmental Protection Agency, Research Series 14010
DII 05/71, Washington, D.C. , May, 1971.
19. Bituminous Coal Research, Inc., Studies of Densification
of Coal Mine Drainage Sludge, Environmental Protection
Agency, Research Series 14010 EJT 09/71, Washington, D.C.
September, 1971.
84
-------
20. West Virginia University, Dewatering of Mine Drainage
Sludge. Environmental Protection Agency, Research Series
14010 FJX 12/71, Washington, D. C., December, 1971.
21. West Virginia University, Dewatering of Mine Drainage
Sludge. Environmental Protection Technology Series,
EPA-R2-73-169, Washington, D.C. , February, 1973.
22. Rozelle, R.B., et.al., Studies on the Removal of Iron
from Acid Mine Drainage Water. Wilkes College Research
and Graduate Center, Submitted to Coal Research Board,
Commonwealth of Pennsylvania, June, 1968.
23. Stumm, W., Lee, G.F., "Industrial and Engineering
Chemistry" 53, 143-146 (1961).
24. Harvard University, Oxygenation of Ferrous Iron. Federal
Water Pollution Control Administration, Research Series
14010 06/69, Washington, D.C. , June, 1969.
25. Mihok, E.A. , Mine Water Research-Catalytic Oxidation of
Ferrous Iron in Acid Mine Water by Activated Carbon,
Report of Investigators 7337, U.S. Department of the
Interior, Bureau of Mines, December, 1969.
26. Bituminous Coal Research, Inc. , Treatment of Ferrous Acid
Mine Drainage with Activated Carbon, Environmental
Protection Technology Series, EPA-R2-73-150, Washington,
D.C. , January, 1973.
27. Brookhaven National Laboratory, Department of Applied
Sciences , Treatment of Acid Mine Drainage by Ozone
Oxidation. Environmental Protection Agency, Research
Series 14010 FMH 12/70, Washington, D.C. , December, 1970.
28. Continental Oil Co. , Microbiological Treatment of Acid
Mine Drainage Waters . Environmental Protection Agency
Research Series, 14010 ENV 09/71, Washington, D.C.,
September, 1971.
29. Lovell, H.L. , Experience with Biochemical Iron-Oxidation
Limestone Neutralization Process, Paper presented at
Fourth Symposium on Coal Mine Drainage Research, Pitts-
burgh, Pennsylvania, April, 1972.
85
-------
30. Tyco Laboratories, Inc., Electrochemical Treatment of
Acid Mine Waters, Environmental Protection Agency
Research Series, 14010 FNQ 02/72, Washington, D.C.,
February, 1972.
31. Dorr-Oliver, Inc., Operation Yellowboy, Report submitted
to Pennsylvania Coal Research Board of the Department of
Mines and Mineral Industries, June 1966.
86
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE ANDSUBTITLE
ELECTROCHEMECAL REM3VAL OF HEAVY METALS FROM ACID
MINE DRAINAGE
5. REPORT DATE
May 197*1;
Issuing Date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
Nicholas B. Franco, Robert A. Balouskus
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORG '\NIZATION NAME AND ADDRESS
ECOTROL, Inc.
7379 Route 32
Columbia, Maryland 2101&
10. PROGRAM ELEMENT NO.
, 21 AFY/lU
11. CONTRACT/GRANT NO.
1U010 GAO (old number)
12. SPONSORING AGENCY NAME AND ADDRESS
National Environmental Research Center, USEPA
Cincinnati, Ohio 45268, and
Pennsylvania Dept. of Environmental Resources
Harrisburg, Pennsylvania
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
Laboratory and field studies were conducted to determine the economics of ferrous
iron oxidation in a cell containing a bed of conductive particles in the space
between the cathode and the anode. The effects of the process on other heavy
metals present in acid mine drainage (AMD) and on the character of solids pre-
cipitated during treatment of low acidity water were also observed.
A 18.9 liter/min (5 gal./min) pilot plant was operated at an actual mine site to
evaluate treatment of IfO and 250 ppm ferrous iron AMD at pH levels of 2 and 5* A
conventional aeration system was also included to generate comparative data.
Approximately 86 percent of the ferrous iron was oxidized during electrolysis of
the low pH water. The conversion rate was less in the pH 5 AMD due to coating
of electrode sites with ferric hydroxide. A if to 10$ decrease from k ppm occurred
in manganese concentration, while aluminum feed values of approximately 1 ppm were
reduced by ho to 60$ especially in the pH 2 water.
Estimates for a 473 liter/min (125 gal./min) plant based on the pilot data for
oxidation only indicate that capital and operating costs for electrochemical
treatment would be higher than those for aeration by factors of 5 and 1.7 respect-
ively.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
*Electrochemistry
*Iron
Electrolysis
Oxidation
Sludge
Cost Analysis
Electrodes
*Acid mine drainage
Heavy metals
Pennsylvania
81
13B
18. DISTRIBUTION STATEMENT
Release to Public
19. SECURITY CLASS (ThisReport)
UNCLASSIFIED
21. NO. Of PAGES
97
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
87
•frO.S. GOVERNMENT PRINTINu OFFICE: 1974 - 757-582/5319
------- |