EPA-600/2-76-206
September 1976
Environmental Protection Technology Series
            TREATMENT OF  ACID MINE DRAINAGE BY
                   THE  ALUMINA-LIME-SODA PROCESS
                                    I
                                    55
                                    \
           532:
\
 ID
                                 Industrial Environmental Research Laboratory
                                     Office of Research and Development
                                     U.S. Environmental Protection Agency
                                            Cincinnati, Ohio 45268

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency,  have been grouped  into five series. These five  broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:

     1.    Environmental Health Effects Research
     2.    Environmental Protection Technology
     3.    Ecological Research
     4.    Environmental Monitoring
     5.    Socioeconomic Environmental Studies

This report  has been  assigned to the  ENVIRONMENTAL PROTECTION
TECHNOLOGY series. This series describes research performed to develop and
demonstrate  instrumentation, equipment, and methodology to repair or prevent
environmental degradation from point and  non-point sources of pollution. This
work provides  the new  or improved technology required for the control and
treatment of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                                 EPA-600/2-76-206
                                                 September 1976
TREATMENT OF ACID MINE DRAINAGE BY THE ALUMINA-LIME-SODA PROCESS
                               by

  J. W. Nebgen, D. F. Weatherman, M. Valentine, and E. P. Shea

                   Midwest Research Institute
                  Kansas City, Missouri  64110
                       Grant No. S-802816
                         Project Officer

                        Roger C. Wilmoth
            Resource Extraction and Handling Division
             Crown Mine Drainage Control Field Site
                Rivesville, West Virginia  26588
              U.S. ENVIRONMENTAL PROTECTION AGENCY
               OFFICE OF RESEARCH AND DEVELOPMENT
          INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
                     CINCINNATI, OHIO  45268

-------
                             DISCLAIMER
     This report has been reviewed by the Industrial Environmental
Research Laboratory, U.S. Environmental Protection Agency, and ap-
proved for publication.  Approval does not signify that the con-
tents necessarily reflect the views and policies of the U.S. En-
vironmental Protection Agency, nor does mention of trade names or
commercial products constitute endorsement or recommendation for
use.
                                  ii

-------
                               FOREWORD
     When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on
our health often require that new and increasingly more efficient pollution
control methods be used.  The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and
improved methodologies that will meet these needs both efficiently and
economically.

     This report describes the adaptation of a treatment process, originally
developed for desalination of brackish water, for use in producing a high-
quality effluent from acid mine drainage.  The process involves lime
addition to pH 12 combined with sodium aluminate addition to precipitate
calcium and sulfate as calcium sulfoaluminate rather than calcium sulfate.
In this manner, the sulfate levels can be drastically reduced and the
effluent from the process can, in most cases, meet chemical and microbio-
logical standards for potable water.  The process provides an option
heretofore not available with neutralization processes treating acid mine
drainage—namely, the production of a product water meeting chemical and
microbiological standards for potability.  This is one of several projects
undertaken by lERL-Ci  to develop and demonstrate acid mine drainage
treatment and abatement processes to provide alternatives in the selection
of treatment facilities to meet the demands of the expanding extractive
industries.
                                        David G.  Stephan
                                            Director
                          Industrial Environmental Research Laboratory
                                           Cincinnati
                                    lii

-------
                              ABSTRACT
     The alumina-lime-soda process is a chemical desalination process
for waters in which the principal sources of salinity are sulfate
salts and has been field tested at the Commonwealth of Pennsylvania's
Acid Mine Drainage Research Facility, Hollywood, Pennsylvania, as a
method to recover potable water from acid mine drainage.  The alumina-
lime-soda process involves two treatment stages.  Raw water is reacted
with sodium aluminate and lime in the first stage to precipitate dis-
solved sulfate as calcium sulfoaluminate.  In the second stage, the
alkaline water (pH = 12.0) recovered from the first stage is carbon-
ated to precipitate excess hardness.  Following carbonation, product
water meets USPHS specifications for drinking water.

     The alumina-lime-soda process is attractive when compared to
other water recovery processes, e.g., ion exchange or reverse osmo-
sis.  Alumina-lime-soda desalting depends strictly upon chemical
processes and thus can be operated using conventional equipment and
procedures.  The constituents removed are contained in easily de-
watered solids.  There are no waste liquid streams needing treatment
or special handling.

     Alumina-lime-soda process economics are influenced most by the
cost of sodium aluminate.  Widespread application of the alumina-
lime- soda process will increase demand for sodium aluminate, and
should spur interest in alternate sources of this treatment chemical.

     Operating costs for recovering potable water from an acid mine
drainage having an acidity of 700 mg/liter and a sulfate level of
750 mg/liter are estimated to be in the range of $0.21 to $0.27/m3
($0.79 to $1.04 per 1,000 gal).
                                  iv

-------
                              CONTENTS

                                                                Page

Abstract	    iv

Figures	    vii

Tables	    viii

Acknowledgments	    x

Section

I        Introduction	     1

II       Conclusions 	     3

III      Recommendations 	     5

IV       Design of the Alumina-Lime-Soda Process for Acid
           Mine Drainage	     6

V        Experimental Development of the Process 	    10

           General Discussion	    10
           Stage I - Alumina-Lime-Soda Treatment	    14
           Stage II - Ratio Mixing of Raw AMD/Stage I
             Effluent and Carbonation	    25
           Stage I - Sludge Characterization	    29

VI       Data Interpretation and Discussion	    37

           Reduction of Experimental Data	    37
           Relation of Sludge Stoichiometry to Other
             Systems	    42

-------
                        CONTENTS (concluded)

                                                                Page

Section

VII      Alumina-Lime-Soda Process Chemistry for Recovering
           Water From Acid Mine Drainage	    45

           Stage I - AMD Reactions With Lime and Sodium
             Aluminate	    45
           Stage II - Carbonation	    48
           Chemical Requirements for the Alumina-Lime-Soda
             Process	    49
           Stage II - Ratio Mixing of Stage I Effluent/Raw
             AMD and Carbonation	    50

VIII     Chemical Costs for the Alumina-Lime-Soda Process
           for Several Acid Mine Drainages in Pennsylvania .  .    53

IX       Sodium Aluminate for the Alumina-Lime-Soda Process.  .    61

X        Alumina-Lime-Soda Process Economics 	    65

XI       Discussion of Results	    69

Appendix A - Specifications for the 190,000 Liter/Day
               (50,000 Gal/Day) Demonstration Plant	    74

Appendix B - Alumina-Lime-Soda Chemical Requirements for
               Recovering Water From Acid Mine Drainage. ...    85

References	'	    94
                                 vi

-------
                               FIGURES

No.                             Title                           page

 1     Recovery of Potable Water From Acid Mine Drainage
         Using the Alumina-Lime-Soda Process 	     7

 2     Alumina-Lime-Soda Bench-Scale Pilot Plant 	    11

 3     Treatment Chemical Addition in Stage I Reactor	    15

 4     Stage I Effluent/Sludge Interface in Settling Tank.  .  .    18

 5     Additional Reaction Vessel for Extended Residence
         Time Studies	    26

 6     Sand Filtration of Solids Phase From Settler	    30

 7     Vacuum Filtration Specific Resistance Plot.  	    34

 8     Stage I Reactor Effluent Settling Data (Initial Solids
         Concentration Equal to 0.6%)	    36

 9     Sulfate Removal Plotted Against Causticity in Stage  I
         Effluent	    38

10     Calcium in Sludge Plotted Against Causticity in
         Stage I Effluent	    39

11     Total Construction Costs	    67

12     Total Annual Operation and Maintenance Costs	    68

13     System Configuration for the 190,000 Liter/Day
         (50,000 Gal/Day) Demonstration Plant	    75

14     Spatial Layout of the 190,000 Liter/Day (50,000
         Gal/Day) Demonstration Plant	    76

15     Flow Diagram to Determine Material Balance	    77

                                 vii

-------
                               TABLES

No.                             Title                           Page

 1     Average Composition of Hollywood, Pennsylvania, Acid
         Mine Drainages, January to June 1975. ........    12

 2     Analytical Methods for Field Tests	    13

 3     Sulfate Reduction With and Without Caustic Soda
         Stabilization of Sodium Aluminate 	    17

 4     Experimental Data for Design of Stage I.  Runs With
         Proctor No. 2 Acid Mine Drainage	    20

 5     Experimental Data for Design of Stage I.  Runs With
         Proctor No. 1	    24

 6     Ratio Mixing of Stage I Effluent and Proctor No. 2
         Acid Mine Drainage	    27

 7     Experimental Data for Design of Stage II Runs With
         Proctor No. 2	    28

 8     Vacuum Filtration Data	    32

 9     Proctor No. 1, Hollywood, Pennsylvania, Raw AMD
         Treated by Alumina-Lime-Soda:  257. Treated in
         Stage I	    54

10     Commonwealth of Pennsylvania Hawk Run AMD Plant,
         Philipsburg, Pennsylvania, Raw AMD Treated by
         Alumina-Lime-Soda:   75% Treated in Stage I	    55

11     Proctor No. 2, Hollywood, Pennsylvania, Raw AMD
         Treated by Alumina-Lime-Soda:  807. Treated in
         Stage I	    56
                                viii

-------
                         TABLES (concluded)

No.                             Title                           page

12     Sawmill Run, Raw AMD Treated by Alumina-Lime-Soda:
         85% Treated in Stage 1	   57

13     Young and Son Coal Corporation, Parker's Landing,
         Pennsylvania Raw AMD Treated by Alumina-Lime-Soda:
         90% Treated in Stage 1	   58

14     Bethlehem Mines Corporation Mariana Mine No. 58,
         Marianna,  Pennsylvania,  Raw AMD Treated by
         Alumina-Lime-Soda:   1007. Treated in Stage I	   59

15     Cost Estimate:   20,000 Tons Per Year Sodium Alumlnate
         Manufacturing Plant (Calcine Process)  	 '.   63

16     Tank Capacities for 190,000 Liter Per Day (50,000
         Gal/Day) Demonstration Plant	   82

17     Bill of Materials for Demonstration Plant (50,000
         Gal/Day) Acid Mine  Drainage to Drinking Water ....   84
                                ix

-------
                           ACKNOWLEDGMENTS
     The principal investigator for this project was Dr. John W.
Nebgen, Principal Chemist.  He was assisted by Douglas F. Weatherman,
Assistant Chemist, who served as project leader, responsible for day-
to-day operation of the project, and by Evan P. Shea, Senior Chemical
Engineer, and Mark Valentine, Associate Chemical Engineer, who were
responsible for the engineering analysis.  Field personnel at the
Hollywood facility were Michael McGregor and Robert Bartley.

     The program was conducted in Midwest Research Institute's Physi-
cal Sciences Division, Treatment and Control Processes Section,
Dr. A.  D. McElroy, Head.

-------
                               SECTION I

                              INTRODUCTION
      The alumina-lime-soda process  for  removing calcium sulfate  hard-
 ness  from water was  conceived  at  Midwest  Research  Institute  in  1970
 and developed under  an  Office  of  Saline Water  program during 1971  to
 1972.—'   The  process involves  two stages  which are briefly described
 as  follows.

      The first step  of  the  alumina-lime-soda process  involves treat-
 ing brackish  water with lime and  sodium alaminate  at  a pH of 12.0.
 The sodium aluminate removes substantial  quantities of calcium and
 magnesium sulfate as insoluble  sulfoaluminates under  the high pH
 conditions.

      In  the second step, effluent from  the lime-sodium aluminate
 treatment, which is  highly  alkaline, is treated by addition  of car-
 bon dioxide to neutralize the excess causticity and to precipitate
 calcium  carbonate.   Thus, the process will yield a completely soft-
 ened  water which will meet  drinking water standards if salinity  is
 due to sulfates  of calcium  and magnesium.

      Acid mine  drainage  (AMD) arises from the  oxidation of residual
 pyrite and marcasite (FeS2> in active and abandoned coal mines.  It
 consists primarily of dilute sulfuric acid and iron sulfates.  Alu-
minum, calcium,  and magnesium sulfates are also present due to neu-
 tralization and  solubilization of soil materials by the acid.  The
alumina-lime-soda process is well suited  for dealing with the sev-
eral problems which are posed by acid mine drainage.  The process
provides for neutralization of acid, removal of iron and hardness
components, and  for desalination to potable water standards.   The
sole by-product is a sludge which is readily dewatered, and water
recoveries in the 95 to 98% range are possible.

-------
     The Commonwealth of Pennsylvania Department of Environmental Re-
sources and the U.S. Environmental Protection Agency funded the first
phase of a proposed two-phase study to field test the alumina-lime-
soda process for recovering water from acid mine drainage.  The Phase
I program, presented in the report, was conducted at the Commonwealth
of Pennsylvania's Mine Drainage Research Facility at Hollywood,
Pennsylvania.  A bench-scale unit was installed at the facility, and
two mine drainages were evaluated with respect to the alumina-lime-
soda process.  The primary objective of the program was to obtain
information necessary to design, build, and operate a 190,000 liter/
day (50,000 gal/day) demonstration plant.  This objective was met by
obtaining operating experience and experimental data which established
alumina-lime-soda chemical and mass balances for the acid mine drain-
age application and identified other operational and design param-
eters.  The program was particularly sensitive to operational prob-
lems which might arise with mine drainage, and to cost-sensitive
parameters.

     This report will discuss in detail the alumina-lime-soda process
in the context of recovering potable water from acid mine drainage.

-------
                             SECTION II

                             CONCLUSIONS
     Important conclusions drawn from study results are listed below.

     1.  The alumina-lime-soda process is capable of producing potable
water from acid mine drainages.  The process is best suited for acid
mine drainages having sulfate concentrations ranging between 400 and
1,200 mg/liter.

     2.  The process economics for recovering potable water using
alumina-lime-soda hinge upon costs of sodium aluminate.  The high
purity sodium aluminate currently marketed in the United States is
not an essential requirement for the process.  A lower grade sodium
aluminate could be manufactured by calcining soda ash and bauxite.
Such a product would be adequate and would reduce treatment costs
by 30 to 40%.

     3.  The alumina-lime-soda process can be integrated into con-
ventional waterworks operations and operates very much like conven-
tional lime-soda water softening processes.  There are no require-
ments for sophisticated equipment necessary for other processes to
recover potable water from acid mine drainage, e.g., ion exchange
or reverse osmosis.

     4.  Not all raw acid mine drainage needs to be treated by
alumina-lime-soda to recover a product water meeting potable water
standards.  The alumina-lime-soda process will remove sulfate to
concentrations of 100 mg/liter or less.  Hence, effluent from the
alumina-lime-soda reactor can be blended with raw acid mine drain-
age to obtain product water with a sulfate concentration of 250
mg/liter.  The blending also reduces the carbon dioxide require-
ments in the process second stage where pH is dropped to near 7.0.

-------
     5.  The alkaline conditions (pH «< 12.0) and residence time (ca.
90 min) for reacting raw acid mine drainage with alumina-lime-soda
chemicals, coupled with the fact that the process operates in the
presence of air, preclude the need for a separate unit operation to
oxidize iron (II) to iron (III).  Oxidation of reduced metallic ions
to insoluble trivalent oxides is rapid under the alkaline process
conditions.

     6.  Dissolved iron and aluminum in the raw acid mine drainage
are effective in removing a significant portion of the dissolved sul-
fate during alumina-lime-soda treatment.  The lime used in the pro-
cess to maintain pH 12.0 conditions creates a situation where  in situ
iron and aluminum will precipitate calcium sulfoferrite and sulfoalu-
minate salts.  Thus, the requirements for sodium aluminate to  remov^
sulfate to 250 mg/liter or less are significantly diminished.

     7.  The calcium sulfoaluminate and sulfoferrite salts, which con-
tain the salts removed by the alumina-lime-soda process, do not pose
serious scaling problems to reactors and pipes.  The insoluble cal-
cium sulfosalts act as seeds during the reaction.  Thus, the precipi-
tates formed during the reaction tend to find deposition sites on
the seed rather than on reaction vessel walls or pipe surfaces.

     8.  The sole by-product of the alumina-lime-soda process  is a
mixture of solids containing mainly calcium sulfoaluminate, sulfo-
ferrite, and carbonate.  The sludges are microcrystalline and  readily
dewatered to a cake which can be easily handled.  The process  does
not generate waste salt or acid streams such as those encountered in
ion exchange or reverse osmosis operations which require special
treatment for adequate disposal.

-------
                             SECTION III

                           RECOMMENDATIONS


      1.  A190m3/day  (50,000 gal/day) demonstration plant should be
constructed in order  to obtain the more detailed process information
required for development of full-scale plants designed to recover
potable water using the alumina-lime-soda process.  The demonstration
plant should be operated for 1 year to collect the following informa-
tion:

     *  Long-term operation data;
     *  Chemical requirements;
     *  Operating costs;
     *  Labor and power costs;
     *  Capital costs for full-scale plant; and
     *  Design and operating criteria.

     2.  Alternate sources of sodium aluminate should be explored in
order to reduce chemical costs of the alumina-lime-soda process.
Sodium aluminate obtained by calcining bauxite and soda ash should
be evaluated using the small pilot unit.  The small-scale experi-
ments for evaluating calcined bauxite/soda ash sodium aluminate can
be conducted concurrently with the demonstration study.

-------
                             SECTION IV

    DESIGN OF THE ALUMINA-LIME-SODA PROCESS FOR ACID MINE DRAINAGE
     A flow diagram for the recovery of potable water from acid mine
drainage (AMD) using the alumina-lime-soda process is presented in
Figure 1.  The process consists of two basic stages:  Stage I--
Alumina-Lime-Soda Treatment; and Stage II--Ratio Mixing of Raw AMD/
Stage I Effluent and Carbonation.  Both stages involve the precipi-
tation of solids which are separated from product water by filtra-
tion.  The final step of the process is a pH adjustment to produce
water meeting drinking water specifications.

     As can be seen from Figure 1, raw AMD is first split into two
streams.  The major fraction is treated by alumina-lime-soda in
Stage I, while the minor stream is mixed with Stage I effluent in
the second process stage.

     The key process step occurs in Stage I where the raw AMD is
mixed with the treatment chemicals, i.e., sodium aluminate and hy-
drated lime.  The chemical reactions in this first stage will ac-
complish the following:

     *  Neutralize acid;
     *  Precipitate heavy metals and magnesium; and
     *  Remove calcium sulfate.

The first two effects are those normally encountered using lime
treatment of acid mine drainage; the third effect is unique to the
alumina-lime-soda process.

     Calcium sulfate i.s removed during treatment with sodium alumi-
nate and lime because insoluble calcium sulfoaluminates are formed.
Calcium sulfate produced by the neutralization of acid and precipi-
tation of heavy metals and magnesium will react at a pH of 12.0 with
aluminate ion (from sodium aluminate and in situ aluminum in the
mine drainage) to yield mixtures of calcium sulfoaluminates having
compositions of CaSO-AO-SCaO-xIO and

-------
                       pH 2.5-3.5
                       Fe 50-150  ma/I
                       Al 20-50 mg/i
                       SO4 750 ma/I
               Stage!
               (Alumina-Lime-Soda
               Treatment)
pH 11.9-12.1
                                        Solids

                                        Calcium Sulfoolumlnate
                                        Calcium Sulfoferrite
                                        Mg(OH)2
                                        MnO(OH)
                      Solid.

                      CoCO3
                      AI(OH)3
                      F«(OH)3
                      Mn(OH)2
                                                                                     Fe<0. 1 ma/I
                                                                                     Al<0.1 ma/1
                                                                                     SO4 250 mg/l
                                                                                     Coifaonote Alkolinity
                                                                                        35 mg CoCCyi
                                                              F«<0.1
                                                              AKO.l
                                                              SO4 250
                                                              BicoftxMOl* Aftnlinily
                                                                 35 my CoCOj/1
                                                           *
                                                         Disposal
Figure  1.   Recovery of potable  water  from acid mine  drainage  using the  alumina-lime-soda  process.

-------
In addition, calcium sulfate will also react with in situ iron in the
ruw watur to produce calcium sulfoferrite, CaSO^-Fe203'3CaO-xH20.
'Liu: purpose of the sodium aluminatc is to provide the system with
sufficient aluminum to reduce sulfate levels to acceptable values,
while the lime acts to maintain the pH in the 11.9 to 12.1 range ana
to stabilize the calcium sulfoaluminate and sulfoferrite precipi-
tates.  A sulfate concentration of 100 rag/liter is a practical lower
limit for sulfate removal in Stage I.

     Following Stage I reaction, the treated mine drainage is allowed
to settle, at which time the solids are separated from  the effluent.
Effluent is sent to Stage II of the process, and settled solids are
filtered to recover additional treated water.  The alumina-lime-soda
solids will settle to a concentration of only about 2%, but can be
concentrated to about 10 to 127. by gravity or vacuum filtration.
The filtrate recovered represents about 25% of the total Stage I
effluent.

     Water treated by alumina-lime-soda in Stage I contains excess
lime which must be removed in Stage II.  The clear effluent from the
solids settling and the filtrate removed by solids filtration are
mixed with raw acid mine drainage in Stage II to neutralize a part
of the excess lime.  The proportion of volume between Stage 1 ef-
fluent and raw mine drainage is determined by the sulfate wanted in
product water.  Effluent from Stage I contains about 100 rag/liter
of sulfate.   If product water is to contain 250 mg/liter of sulfate,
it can be blended with raw acid mine drainage to produce this con-
centration.   This ratio mixing of raw acid mine drainage and Stage
I effluent will precipitate the heavy metals in the raw AMD stream.

     The mixing of the two streams will not neutralize all the ex-
cess lime.  Thus, carbon dioxide is added in Stage II to complete
neutralization of the lime by precipitating calcium carbonate.  Care
must be taken in this step of the process to avoid overcarbonating
the system.   Too much carbon dioxide will redissolve the calcium
carbonate precipitate as calcium bicarbonate and result in unneeded
hardness in the product water, as well as an unnecessary consumption
of carbon dioxide.  Therefore, the Stage II carbonation is controlled
at a pH of 10.3, the point of minimum solubility of calcium carbonate.

-------
     The solids generated in Stage II are separated from the product
water by filtration using a sand filter.

     The filtrate will have a pH of 10.3 and will contain about 35 rag/
liter of dissolved calcium carbonate.  The pH of the product is dropped
to the 7.0 to 8.0 range by addition of a small amount of carbon dioxide
to convert the dissolved carbonate to bicarbonate.

     Following this final pH adjustment step, the water recovered from
acid mine drainage using the alumina-lime-soda process will meet pota-
ble water specifications.

-------
                               SECTION V

               EXPERIMENTAL DEVELOPMENT OF THE PROCESS
GENERAL DISCUSSION

     The scheme for using the alumina-lime-soda process to recover
water from acid mine drainage described in Section IV has been de-
veloped through laboratory studies in a 190 liter/day (50 gal/day)
bench-scale pilot unit.  A photograph of the assembled unit is pre-
sented in Figure 2, together with identification of essential parts.
The bench-scale pilot plant consists of:

     Stage I

     190 liter (50 gal) raw water storage tank;
     20 liter (5 gal) Stage I reaction vessel;
     40 liter (10 gal) lime storage vessel;
     1 liter Erlenmeyer Flask as the sodium/aluminate storage vessel;
     70 liter (18 gal) square prism settler; and
     20 liter (5 gal) sludge-filtrate collection vessel.

     Stage II

     20 liter (5 gal) Stage I effluent feed tank;
     Size 1A carbon dioxide cylinder; and
     20 liter (5 gal) reactor/carbonator vessel.

     The system also contains various pumps, stirrers, and flowmeters,

     The unit was installed at the Commonwealth of Pennsylvania's
Mine Drainage Research Facility, Hollywood, Pennsylvania, in January
1975.  Two acid mine drainages available at the Hollywood site were
used to develop operating parameters for the alumina-lime-soda pro-
cess.  Average compositions for these mine drainages for the project
period are presented in Table 1.
                                   10

-------
Stage I   Alumina-Li me-Soda Treatment
  1  Raw AMD Storage Tank
     Reactor (Stage  I)
     Lime Storage Tank
     Sodium Aluminate  Storage Vessel
     Settler
     Stage I  Sludge-Filtrate Collection Vessel
           Figure  2.   Alumina-lime-soda  bench-scale  pilot  plant.
2
3
4
5
o
Stage II  Carbonation
  7  Stage I  Effluent Storage Tank
  8  OC>2 Tank
  9  Reactor/Carbonator  (Stage II)

-------
       Table 1.   AVERAGE COMPOSITION OF HOLLYWOOD,  PENNSYLVANIA,
                 ACID MINE DRAINAGES, JANUARY TO JUNE 1975
                                    Proctor  No.  1
                    Proctor No. 2
 PH
 Acidity,  mg/j6  as
 Sulfate,  mg/je  as  SO,
 Calcium,  mg/£  as  Ca
 Magnesium, mg/l as  Mg
 Sodium, mg// as Na
 Iron, mg/4 as  Fe  total
 Iron, mg/j(L as  Fe^
 Aluminum, mg/£ as Al
       3.8
     180
     300
       6
      15
       8
      30
Not determined
      20
  2.8
700
750
  8
 25
  2
100
 19
 40
     The bulk of  the effort was  spent  in  establishing  Stage I param-
eters for the stronger mine drainage,  Proctor  No.  2.   This drainage
is more representative of  those  which  are likely candidates for
alumina-lime-soda treatment.  Experiments with Proctor No. 1 were
significant, however, since the  different water yielded data which
greatly facilitated the interpretation of results, particularly  the
identification of chemical processes.

     After Stage  I was established for both Proctors Nos. 2 and  1,
the Stage II processes were evaluated.  Since  the  carbonation reac-
tion of Stage II  involves straightforward chemical reactions in  con-
trast to Stage I, less emphasis  was placed there.

     A majority of the experimental data  for process development con-
sisted of analyses of aqueous phases at .various stages in the process
for important constituents.  The constituents  sought and the analyti-
cal methods used  are shown in Table 2, together with expected accura-
cies.  The bulk of the constituents were  analyzed using Hach Chemical
Company Colorimetric Techniques,  These methods were chosen because
of their rapidity.  As a result  of the simplified analysis, we could
know within an hour what was happening in  a particular experiment,
and make appropriate changes.
                                 12

-------
               Table 2.  ANALYTICAL METHODS FOR FIELD TESTS
     Constituent

 Total and calcium
   hardness

 Phenolphthalein and
   methyl orange
   alkalinity^'

 Acidity
 Sulfate
 Total  iron
 Aluminum
PH
            Method
 EDTA titration
 Titration  with 0.1 N HC1  to
  phenolphthalein and
  methyl orange end points

 Titration  with 0.1 N NaOH
  to phenolphthalein end
  point

 Hach Chemical Company
  turbidimetric method

 Hach Chemical Company
  colorimetric method

Hach Chemical Company
  colorimetric method

Beckman pH meter
    Accuracy

 1,000  + 50  mg/4


 1,000  + 30  mg/jj



 1,000  + 50  mg/4



 1,000  + '100 mg/A


    10 + 0.5


  100 + 10 mg/j&


+ 0.05 pH units
J/  Results used to establish the different forms of alkalinity:
      Causticity - Principal constituent when most of the total alkalinity
                     is phenolphthalein alkalinity.
      Carbonate - Principal constituent when the phenolphthalein alkalinity
                    is about half the total alkalinity.
      Bicarbonate - Principal constituent when most of the total alkalinity
                      is methyl orange alkalinity.
                                   13

-------
     In subsequent discussion in this section, we shall present the
pertinent data obtained during the experiments.  Data presented rep-
resent a daily average for a particular run.  The average is that
obtained from hourly analyses of the effluents from the runs.

     The discussion will be organized around particular aspects of
the process, i.e., Stage I, Stage II, and Solids Separation.

STAGE I - ALUMINA-LIME-SODA TREATMENT

Experimental Procedures and Observations

     The Stage I experiments consisted of the addition of lime and
sodium aluminate to the raw acid mine drainage in the 20 liter (5
gal) reactor and allowing the mixture to react for a specified
length of time.  Reaction time was determined by the flows of the
various process streams, and by the position of the overflow port
on the Stage I reactor.  Following reaction, the sludge was allowed
to settle, and the effluent sampled for analysis.  In later stages
of the program, the effluent was sent to the Stage II reactor "where
it served as feed in carbonation experiments.  Figure 3 shows the
lime and sodium aluminate being added to the Stage I reactor.

     The lime needed for the reaction in Stage I was added in the
following manner:  calcium hydroxide (5 g/liter) was added to the
raw acid mine drainage in the 40 liter (10 gal) lime storage vessel.
This mixture was constantly stirred and pumped into the system by
means.of a small metering pump.  Raw mine drainage was used to carry
the lime solution in order to simplify interpretation of results.
In most cases, th« lime solution constituted about 35 to 407, of the
raw water flow through the system.

     The sodium aluminate was also added as a solution.  Technical
grade sodium aluminate (50 g/liter) was dissolved in distilled water
and pumped to the Stage I reactor at rates from 1.0 to 1.8 ml/min.
During initial stages of the program, we found that sodium aluminate
solutions prepared in this manner were unstable; approximately half
of the available aluminum in the sodium aluminate precipitated out
in the storage flask after a day or two and was not being added into
the Stage I reactor.   To overcome the instability problem, we sta-
bilized the sodium aluminate solution by adding excess caustic soda.
                                  14

-------
                                           SODIUM ALUMINATE
                                                I
Figure 3.   Treatment chemical addition in Stage I  reactor.

-------
The NU..O/AI..O.. molar  ratio  In  liquid  (indium  a luiiiLimlo  LH  !.!>,  whurcmn
the NU20/A120-1 ratio  in commercial dry  sodium  alumlnate is  1.13  to
1.15.  For  the process development studies,  the  sodium aluminate so-
lution was  stabilized by adding  12 g  of sodium hydroxide  pellets for
every 100 g of dry sodium aluminate to  achieve the  1.5:1  Na90/Al_0,
ratio found in the liquid sodium aluminate.

     The instability  of the sodium aluminate was detected by wide
variances in sulfate  removal.  Table  3  compares  sulfate removal  be-
fore and after caustic soda stabilization.   The  sodium aluminate
solution in both cases was a 57,  solution at  a  flow  rate of  1.5 ml/
min.  Before stabilization, the  sulfate reduction was  quite erratic,
whereas the results after stabilization showed a marked increase in
removal rate and are  relatively  nonvariant.

     The effluent from the Stage I reactor flowed to the  70-liter
(18-gal) settler where the slurry of  microcrystalline  calcium  sulfo-
aluminate and calcium sulfoferrite formed during the Stage  I reac-
tion of raw AMD with  lime, and sodium aluminate  was allowed to set-
tle out.  The effluent from the  settler was  collected  for use  in
Stage II of the process.  The slurry  settled rapidly and  was easily
vacuum-filtered.  Figure 4 shows the  sludge  blanket which forms  in
the bottom  of the settler.  For  testing purposes only, the  slurry
was collected and batch-filtered using  an Eimco  93  cm2 (0.1 ft2)
filter leaf with a nylon filament cloth filter.

     The sludge generated in the process is  readily filtered either
by vacuum or sand filtration.  (Filtration experimental data appear
on p. 29.)  The individual precipitate  particles seem  to  form  in
themselves, i.e., the precipitation reactions  are self-seeding.   We
noticed very little scale buildup on  stirrers  or reaction vessels
during the  course of  the project.  For  example,  it  was never neces-
sary to suspend experiments to clean  scale or  to unclog piping.

     This experience  contrasts with conventional lime  treatment  of
AMD in which gelatinous metal hydroxides are troublesome  with  re-
spect to scaling and  clogging.   The difference between sludges gen-
erated in the alumina-lime-soda process and  those in conventional
lime treatment is due to the pH conditions at which the sludges  are
formed.  In alumina-lime-soda, the sludges are formed  at  a  pH  of
12.0 where  they form  microcrystals of calcium  sulfoaluminate or
sulfoferrite.  In conventional lime treatment, the  sludges  are
formed at pH's near or below neutrality, and are amorphous  since
no calcium  sulfate and lime are incorporated into them to form the
microcrystals.
                                 16

-------
       Table 3.  SULFATE REDUCTION WITH AND WITHOUT CAUSTIC SODA
                   STABILIZATION OF SODIUM ALUMLNATE
(a)  With unstabilized sodium aluminate

                 Test No.            Percent  sulfate  removal

                   2-19                        43

                   2-20                        48

                   2-21                        38

                   2-24                        54

                   2-25                        19

                   2-26                        69

                   2-27                        78

                   2-28                        87


(b)  With stabilized  sodium aluminate

                   4-3                          89

                   4-7                          88

                   4-8                          91

                   4-10                        85

                   4-11                        89

                   4-16                        90
                                17

-------
                                                         SLUDGE
                                                         INTERFACf
Figure 4.   Stage I effluent/sludge interface in settling tank.
                             L8

-------
     Discussion of analytical data obtained for Stage I reactions
using Proctors Nos.  2 and 1 mine drainages in continuous runs follows,

Proctor No. 2 Continuous Runs

     Data obtained for establishing design parameters for Stage I are
presented in Table 4.  These data represent those obtained after pro-
cess consistency was achieved through caustic soda stabilization of
sodium aluminate solutions.

     The concentrations listed for the addition of lime and sodium
aluminate are the calculated concentrations before reaction in Stage
I.  Calculated concentrations are used so that dosages of the treat-
ment chemicals can be compared directly to concentrations before and
after reaction.

     The data contained in Table 4 were reduced further so that im-
portant process parameters could be identified and quantified.  The
evaluation of the raw data is discussed in more detail in Section VI
of this report.

     The data in Table 4 show further that sulfate residuals in Stage
I effluent can be reduced to below 100 mg/liter when raw water is
treated with sodium aluminate and lime.  When the pH of the system
drops below 12.0, sulfate concentrations in the treated effluent
rise together with aluminum residuals.  This observation confirms
that the formation of calcium sulfoaluminate is pH-sensitive--a pH
of 12.0 is the minimum value for effective sulfate removal.

     For process design purposes, it appears that target sulfate
residuals should be about 100 mg/liter.  If lower concentrations
are sought, then the system will require higher pH's.  The added
costs for lime needed to obtain the higher pH are not worth the
minimal sulfate removal benefit.   Thus, we have chosen the sulfate
level of 100 mg/liter as the target concentration for Stage I
effluent.
                                  19

-------
                         Table 4.  EXPERIMENTAL DATA FOR DESIGN OF STAGE I.  RUNS WITH PROCTOR NO.  2 ACID MINE DRAINAGE
N>
O
Treatment chemicals (calculated values)

Test
No.
4-2
4-3"
4-7
4-8
4-10
4-11
4-16
4-17
4-22
4-23
4-24
4-25
4-28
4-29
4-30
5-5
5-28
5-29
5-30
5-31
6-4
6-5
6-6
6-8

Acidity
(mg/2 as CaC03)
730
730
720
730
730
690
790
810
850
860
860
700
670
650
650
630
620
560
590
640
690
660
680
690
Raw AMD
Fe
(mg/x)
90
95
100
95
95
90
90
100
120
120
110
110
120
115
115
110
105
100
100
100
105
105
100
110
Sodium aluminate
Al
(mg/Q
40
40
45
60
50
60
40
40
45
45
50
60
65
60
60
60
40
40
40
ND
40
40
40
55
S04
(Bg/x)
650
680
680
680
600
700
700
680
750
780
800
720
780
750
750
700
700
700
670
690
720
690
700
800
Causticity
(mg/*. as CaC03>
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
430
Al
(mg/x)
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
130
Lime
Ca
(mg/4 as CaC03)
2,600
2,600
2,600
2,600
2,600
2,600
2,600
2,600
2,340
2,470
2,600
2,470
2,600
2,660
2,660
2,660
2,470
2,470
2,470
2,470
2,470
2,470
2,470
2,470
Causticity
(ng/i as CaC03
2,300
2,300
2,300
2,300
2,300
2,310
2,270
2,260
2,030
2,130
2,240
2,200
2,320
2,390
2,390
2,390
2,230
2,250
2,240
2,220
2,200
2,210
2,200
2,200

-------
                                                              Table 4.  (Concluded;
ro
Test
No.
4-2
4-3
4-7
4-8
4-10
4-11
4-16
4-17
4-22
4-23
4-24
4-25
4-28
4-29
4-30
5-5
5-28
5-29
5-30
5-31
6-4
6-5
6-6
6-8

Fe
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil

Al
(me/;)
1.0
0.5
0.2
0.3
0.2
0.2
0.3
0.4
2.7
1.1
1.2
0.2
1.7
2.6
1.3
1.4
ND
ND
ND
ND
1.3
0.3
1.1
0.4

504
(mg/ji )
90
70
80
60
90
80
70
90
100
190
170
170
160
180
170
140
80
50
130
140
130
160
160
190
Stage 1 effluent
Hardness
(rag/i as CaC03>
1,130
930
870
1,140
1,120
1,120
1,090
1,060
900
1.000
890
1,000
830
1,010
1,350
1,070
1,250
1,030
1,050
1,240
760
1,180
760
940

Causticity
(mg/i as CaCC>3)
920
790
810
980
950
940
970
880
720
900
610
620
610
650
750
700
970
910
740
840
580
730
600
630
— — • — —
£H
12.3
12.2
12.2
12.4
12.3
ND
XD
ND
ND
11.9
11.9
12.0
11.8
11.7
11.0
11.8
12.0
12.0
12.0
11.9
11.8
11.8
ND
ND

Raw AMD
(ml/min)
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
100
Flow rates
Lime
(ml/min)
70
70
70
70
70
70
70
70
60
65
70
65
70
75
75
75
65
65
65
65
65
65
65
65

NaAl02
(ml/min)
1.70
1.50
1.50
1.65
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
1.50
Residence
(min)
87
37
87
87
87
87
87
87
93
90
87
90
87
85
85
85
90
9O
90
90
90
90
90
90
             ND = Not determined.
             Nil =   0.1 mg/

-------
     The hardness and causticity data  for Stage 1 effluent as shown
in Table 4 are indicative of reactions with  the lime.  The difference
between hardness and causticity in the effluent is a measure of how
much noncarbonate hardness is  left in  the treated water.  Noncarbon-
ate hardness in treated acid mine drainage is due to calcium sulfate.
Thus, the smaller the difference between hardness and causticity,
the greater has been the removal of calcium  sulfate.  We note in the
data that differences between  hardness and causticity are relatively
constant at pH's above 12.0.   However, absolute levels of hardness
and causticity increase with increasing pH,  a phenomenon which is
expected.  This observation implies that the minimum quantity of
lime needed for the alumina-lime-soda process is that which main-
tains the pH at 12.0.  If more lime is present than the minimum,
it serves no useful purpose in the process chemistry.

     As can be seen from the data in Table 4, the hardness and
causticity levels in Stage I effluent vary considerably.  However,
the minimum causticity level where sulfate is effectively removed
is on the order of 600 mg/liter as CaCC^.  This causticity value
will maintain the Stage I reaction pH at 12.0.

     Thus, it can be concluded that sufficient lime must be added
to maintain minimum pH and causticity  levels; excess lime is unnec-
essary and will increase processing costs.   The costs are increased
in two ways.  First is the cost of excess lime and second is the
cost of the carbon dioxide needed to neutralize the excess lime.
Therefore, it is important to control the Stage I reactions by
means of maintaining the pH at 12.0 and the  causticity level of 600
mg/liter by the use of lime.

     Residence times were varied during shake-down experiments early
in the program.  These experiments indicated that the minimum resi-
dence in Stage I needed  for maximum sulfate removal was in the 80-
to 100-mln range.  Thus, a residence time of approximately 90 min
was chosen for the design experiments reported in Table 4.

     The relatively long residence times needed for the Stage I re-
actions and the fact that the reactions were conducted while exposed
to air resulted in the complete oxidation of ferrous iron in the raw
AMD to ferric oxide.  The ferric oxide in turn appeared to react
with calcium sulfate to precipitate a calcium sulfoferrite.   The
oxidation of ferrous iron to ferric iron could be observed in the
lime storage vessel.  When lime was first added to the raw AMD, a
dark blue-green precipitate formed due to ferrous iron.  As the
mixture was stirred, the color of the precipitate changed from
                                  22

-------
 blue-green to brown and finally to a rust color normally associated
 with ferric hydroxide.  This color change was virtually complete
 within 1 hr.

 Proctor No. 1 Continuous Runs

      Table 5 lists the analyses for tests conducted using Proctor No.
 1 acid mine drainage.   This water is much more dilute than Proctor
 No.  2.   The data were  obtained and analyzed in the same manner as
 described for Proctor  No.  2.

      Again the lime, sodium aluminate,  and raw acid mine drainage
 were combined in the Stage I  reactor.   The reaction products were
 discharged to the settler  where the sludge was removed.  The ef-
 fluent  was analyzed for constituents of interest.   The concentra-  •
 tions of the treatment chemicals were  reduced for  the Proctor No.
 1 experiments,  so that flow rates and  residence times would be
 similar to those used  for  the Proctor  No.  2 studies.   Sodium alu-
 minate  was added as a  1.6% solution stabilized with 1.9 g/liter
 of caustic soda.   Lime was added as a   0.2% solution.

      The  most  significant  fact disclosed by the tests with Proctor
 No.  1 was the  rate for lime in the  system.   Causticity of the Stage
 I effluent was  consistently below 600 mg/liter and sulfate levels
 never fell below 100 mg/liter.   Further, pH of Stage  I effluent
 rarely  was 12.0 and larger aluminum residuals  were noted when pH
 was  below 12.0.   These observations suggest  that sulfate was  not
 being removed because  insufficient  lime  was  present to form stable
 calcium sulfoaluminate.  The  results imply  that the lime in alumina-
 lime-soda prefers  to react  with  calcium  sulfate and aluminate  ion
 to form the calcium sulfoaluminate  sludge.   Sludge  formation  con-
 sumes lime, thus making  it  unavailable for pH  control.   Closer
 examination of  the  Proctor  No. 1 raw water indicated  that magnesium
was proportionally  much  higher in the Proctor  No.  1 than  in Proctor
No.  2.  Thus, a greater  amount of the lime was  reacting with mag-
nesium  in  the case  of Proctor No. 1 than in Proctor No.  2,  and  in-
sufficient lime was being added to  form the sludge and to maintain
pH and  causticity requirements.
                                  23

-------
Table 5.  EXPERIMENTAL DATA FOR DESIGN OF STAGE I.   RUNS WITH PROCTOR NO. 1
Raw AMD
Test
No.
5-12
5-13
5-14
5-15
5-16
5-17
5-19
5-20
5- ISA
5-16A
5-17A
5-19A
5-20A
Acidity
(mg/i as CaCO-j)
150
170
160
180
170
180
170
180
180
170
180
170
180
Fe
(ms/i)
15
25
25
35
30
20
20
35
35
30
20
20
35

Sodium

aluminate
Al S04 Causticity
Pug//) Png/i) (ng/i a« CaCO3)
20 300 140
15 320 140
40 280 110
25 300 110
20 350 140
30 300 140
30 280 140
25 320 140
25 300 110
20 350 140
30 300 140
30 280 140
25 320 140
Treatment

chemicals


Lime


Al Ca Causticity
Lag/I) (ag/ 1 as CaC03) (3)
530
550
480
540
480
240
400
610
480
450
250
290
560
£§
11.5
11.7
11.9
12.0
11.9
11.4
11.5
12.0
12.0
11.9
11.5
11.4
12.0
Saw AMD
(ml/min)
100
100
75
75
75
75
75
75
75
75
75
75
75
1,060
1,060
1,060
1,060
1,060
1,110
1,110
1,110
1,060
1,060
1,110
1,110
1,110
Flow rates
Lime
(ml/mln)
75
75
56
56
56
60
60
60
56
56
60
60
60
1
1
1
1
1
1
1

NaA102
(ml/min)
1.50
1.50
1.20
1.20
1.50
1.50
1.50
1.50
1.20
1.50
1.50
1.50
1.50
,000
990
990
990
990
,030
,040
,030
990
990
,030
,040
,030
Residence
ti«
(•in)
85
85
113
113
113
110
110
110
189
189
183
183
183

-------
      During the testing of Proctor No. 1, extended residence times
 were tried to see if insufficient reaction time was responsible for
 the poor sulfate removal.  These studies (5-ISA through 5-20A in
 Table 5) were conducted by placing two of the 15-liter (5-gal) re-
 actors in series as shown in Figure 5.  Data from the experiments
 indicate that extended residence time has little effect on the sul-
 fate removal in Stage I of the alumina-lime-soda process.  It is
 concluded therefore that sulfate removal is more dependent on suf-
 ficient lime to maintain the pH at 12.0 and causticity at 600 mg/
 liter than it is on residence time.

 STAGE II - RATIO MIXING OF RAW AMD/STAGE I  EFFLUENT AND CARBONATION

      The tests on Stage II were conducted with Proctor No. 2 acid
 mine drainage.  The treated,  filtered water from Stage I was col- •
 lected in a holding tank, and the water was then mixed with raw
 AMD in the proper ratio before carbonation.   The treated and raw
 AMD are mixed in such a ratio that the final sulfate level is 250
 rag/liter or less,  which meets the USPHS limits.   The mixing of the
 two waters also allows the acid in the raw  AMD to partially neu-
 tralize the hydroxide alkalinity in  the effluent water.   By mixing
 with the raw AMD we also reduce the  amount  of C02 needed in the
 process.

      The effluent  from Stage  I for Proctor  No.  2 AMD has a sulfate
 level of approximately 100 mg/liter,  the  raw AMD has a sulfate
 level of approximately 700 mg/liter.   Based  upon this,  the calcu-
 lated ratio to yield 250 mg/liter sulfate would  be  5 parts of
 treated AMD to 1 part  raw AMD for Proctor No.  2.

      A series  of different ratios were  tried  to  establish the prin-
 ciple of  ratio mixing.   The results are presented in Table 6.   The
 alkalinity  levels are  lower in the effluent  than calculated,  prob-
 ably  due  to  the reaction  of Stage  I effluent with magnesium in  the
 raw AMD.  The  differences  between actual and calculated sulfate  lev-
 els are within reasonable  limits  for the accuracy of  the  analytical
method.   Based on these results,  the optimum ratio  for Proctor No. 2
would be about 3 parts of  raw AMD to 1 part of treated effluent.
                                  25

-------
Stage I Alumina-Lime-Soda Treatment

 1)  Sodium aluminate storage vessel
 2)  Reactor (Stage I)
 3)  Reactor (Stage I) for extended residence time
 4)  Settler

 Figure 5.   Additional reaction vessel for extended
              residence time studies.
                          26

-------
     Table 6.   RATIO MIXING OF  STAGE I EFFLUENT AND PROCTOR NO.
                          ACID MINE DRAINAGE
                              Alkalinity. mg/i        804,
                   Acidity   Actual  Calculated   Actual  Calculated

Raw AMD               730       00        -         680
Stage I effluent      00      930        -          90
Parts mixed
   Stage I:   raw AMD
      1:1             00       30       200        440       385
      2:1             00      280       377        400       288
      3:1             00      430       515        260       240
      5:1             00      550       653        200       191
     The two streams were mixed as they entered the carbonator, which
has a  10-liter capacity.  The C02 was then added to the reactor
through a bubbler tube at the bottom of the vessel.  The acidity of
the raw AMD neutralizes some of the hydroxide alkalinity, which re-
duces  the amount of C02 required in the first carbonation.  The C02
must be added to reduce the pH to 10.3, the point of minimum calcium
carbonate solubility (about 35 mg/liter).  Lowering the pH below 10.3
will result in the redissolution of calcium carbonate, manifested by
an increase in the bicarbonate alkalinity and a decrease in the car-
bonate alkalinity.  If the pH is above 10.3, the system will contain
free hydroxide alkalinity (causticity).  The chemistry described
above  is analogous to that encountered in conventional lime-soda
ash softening—'

     Table 7 shows the concentrations of constituents in effluents
from the Stage II (carbonator) reactor.  The small size of the
bench-scale unit made carbonation very difficult to control.  The
only way to achieve this pH control was by turning the C02 on and
off for various lengths of time.  In all the tests, some excess
C02 was introduced, resulting in pH's less than 10.3 and in un-
desired bicarbonate alkalinity.  However, the tests numbered 6-5
and 6-6, where the pH approached 10.0, show that as we approach
the pH of 10.3, we will be able to achieve the maximum removal of
CaC03.   By achieving minimum values of carbonate alkalinity, we
will remove the maximum amount of CaCO-j, and the amount of C02
needed in the pH adjustment step will be minimized.
                                 27

-------
                                     Table 7.   EXPERIMENTAL DATA FOR DESIGN OF STAGE II RUNS WITH PROCTOR NO. 2
ro
co
Concentration C02 effluent

Test
No.
4-23
4-24
4-25
4-28
4-29
4-30
5-5
5-28
5-29
5-30
5-31
6-4
6-5
6-6
6-8

Fe
(mg/i)
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil
Nil

Al
(ag/Z)
2.3
0.9
0.1
1.7
2.6
2.8
1.8
0.5
0.8
0.9
1.1
1.4
4.2
2.8
2.7

S04
(mgAO
260
340
260
250
230
240
240
160
150
160
170
230
180
260
250

Hardness
(mg/i as CaC03)
330
600
280
250
290
250
290
270
260
160
220
300
160
330
240
Carbonate
alkalinity
(mg/i as CaC03)
70
55
00
45
30
30
20
00
00
00
00
00
25
25
45
Bicarbonate
alkalinity
(rag/2, as CaC03)
25
155
335
55
55
00
145
140
210
210
120
100
00
00
55


PH
9.0
8.7
7 .4
9.1
9.4
9.6
8.2
8.2
7.7
8.6
8.3
8.1
9.9
9.8
9.2
Flow
Raw AMD
(ml/min)
33
33
29
33
33
33
33
33
33
34
34
33
33
33
33
rates
Treated AMD
(ml/min)
165
165
140
165
170
170
175
165
165
165
165
165
165
165
165
              Nil
0.1 mg/,2.

-------
 STAGE  I  -  SLUDGE CHARACTERIZATION

     The calcium sulfoaluminate  and  sulfoferrite  solids  generated  in
 Stage  I  of the  alumina-lime-soda process produce  a  sludge which  is
 filtered to recover water.   The  following discussion contains data
 obtained from experiments dealing with  sludge  filtration and settling.

 Filtration Analysis

     The filtration analysis was necessary to  develop an effective
 method of  treating the bulk phase solids from  the settler to recover
 liquids  and produce a solid material suitable  for handling and dis-
 posal.   Two filtration methods were evaluated  for use in the 190,000
 liter/day  (50,000 gal/day) pilot plant—sand and vacuum.  Vacuum fil-
 tration  is  expensive in both capital requirements and operating  tech-
 niques.  For small plants, i.e.,  the 190,000 liter/day (50,000 gal/
 day) demonstration plant described in Appendix A, the sand filter  is
 preferred,  despite the fact that filtration rates are much slower.
 The following discussion presents findings of  the sand and vacuum
 filtration  experiments.

 Sand Filtration  Experiments - Sand filter tests using a  graded sand
 and gravel medium were conducted in a 6-in. diameter plexiglas column
 to determine the  filtration rate, rate of cake buildup on the sand,
 and solids  separation efficiency.  The solids  concentration in the
 feed solution was approximately  1.9%.  The rates of filtration and
 cake buildup are  shown in Figure 6 as they vary with the volume of
 solution filtered.  An average filtration rate of 1,600  liters/day/
m2 (40 gal/day/ft2) of filter area was obtained, which corresponds
 to a solids  filtration rate, or  solids loading, of 31 kg/m2/day  (6.3
 Ib/ft2/day) at a  1.9% solid concentration in the feed slurry.  The
 solids include water of hydration associated with the chemical com-
position of  the  calcium sulfoaluminate/sulfoferrite sludge.   The
 cake buildup on  the filter was determined to be about 2.5 cm/hr
 (1.0 in./hr) based upon cake buildup and flow rate curves in Figure
6.
                                  29

-------
2
                                       345



                                      VOLUME FILTERED (1000ml)
                                                                                           -20.0
                                                                                                  
-------
      The  solids  collected  on  the  surface  of the  sand  filter were  of
 a  thick-paste  consistency.  The dry  solids  content  of the  filter  cake
 was  about 10%.   When  removed  from the  filter,  the solids did  not  flow
 due  to  their microcrystalline nature.   Thus,  the solids should  be
 readily handled  for disposal,  since  they  do not  exhibit characteris-
 tics normally  associated with amorphous sludges  and slimes.

      The  suspended solids  content in the  clear filtered effluent  was
 nearly  zero.

 Vacuum  Filter  Experiments  - Vacuum filter experiments were conducted
 using an  Eimco 9.3 cm2  (0.1 ft )  filter leaf  to  determine vacuum  fil-
 tration rates.   Table 8 shows  the vacuum  filtration data obtained.
 from the  experiments.

      The  filter  yield,  B  , is the rate of  solids filtered per  unit
 area, which is given by the following  expression.
                           B = p » P » W * a 1/2
                               LU x R x tc J
                                   ry
where     P = vacuum pressure, g/cnr

          W = mass of solids filtered per unit volume of filtrate,
                g/ml

          a = tf/tc, form time per cycle time, sec/sec - dimensionless

          p = filtrate viscosity, g/cm-sec

                                      r\
          R = specific resistance, sec^/g

         t  = cycle time, sec

         tf = form time, sec
                                  31

-------
          Table 8.  VACUUM FILTRATION DATA
                 Filtration measurements
      Time (sec)                    Volume  (ml)


          60                            1,200


         120                            1,800


         165                            2,000



                Filtration parameters


     Vacuum pressure, nun Jig                      597

                          r\
 P = Vacuum pressure, g/cnr                      811


Cj_ = Initial solids concentration, 7.                1.9


Cf = Final solids concentration in cake, %         11.7


 A = Filtration area, cm^                         93


tc = Cycle time, sec                             165


tf = Form time, sec                              120


 p = Viscosity, g/cm-sec                           0.009


     Total filtrate volume, ml                 2,000
                         32

-------
      The mass of solids filtered per unit volume of filtrate,  W   is
 calculated by the following equation.
                         W =
                             (100 - Ci) - (100 - Cf)            (5"2)
                                  CJL           Cf
 where      P = density of filtrate,  g/ml

          C^ = initial solids concentration,  %

          Cf = final  solids  concentration in  cake,  %

     The  specific  resistance is  calculated by plotting  the  time  per
 unit volume of filtrate  versus the  filtrate  volume.   This plot of
 experimental  data  is  shown  in Figure  7.   The slope of the line,   b
 must be used  in the  following equation  to calculate  the  specific
 resistance.


                         „ _  2 x b x  P  x A2
                         R       iTTw(5'3>

where     A =  filter area

     Other factors are the same as those  defined above.

     Substituting the values in Table 8 into the various equations,
and using specific resistance,  R ,  as determined by Figure 7  it is
found that:

     B = 3.03 x 10~3 g/cm2/sec

       = 2,600 kg/m2/day

       = 535 Ib/ft2/day
                                33

-------
   0.10,—
   0.08
^  0.06
>
   0.04
   0.02
      0
       0
                   b =2.83x 10"5sec/m|2
                                         >*y
             I	I
                   500
                                 I      I      I      I      I
  1000
Vol (ml)
1500        2000
Figure 7.  Vacuum filtration specific resistance plot.
                           34

-------
Settling Analysis

     The purpose of the settling study was to determine if the solids
could be separated effectively by settling.  If so, what would be the
residence time and solids contents of the supernatant and sludge?
The solids produced in the reactor settle as a bulk phase at the high
concentrations obtained (0.6%); whereas discrete particle settling is
obtained at low solids concentrations.  Bulk phase settling can be
characterized by measuring the settling velocity of the solid-liquid
interface.  The solids settle freely until the solids concentration
in the bulk phase becomes high enough that compression begins.  At
this point, the settling velocity decreases sharply.

     Figure 8 shows a plot of the height of the interface versus set-
tling time.  The settling occurred in a cylindrical container having
a cross-section of 0.06 nr (0.66 ft ), which contained reactor efflu-
ent with an initial solids concentration of 0.6%.  A constant settling
rate of 1.8 cm/min (0.7 in/min) occurred for the first 5 min.  Solids
compression began after 10 min, diminishing the settling rate
accordingly.

     The solids concentrations in the bulk solid and liquid phases
were measured following settling.  The supernatant liquid concentra-
tion was 0.02%, and the bulk solid phase was 1.89% solids.  The bulk
solids phase contains material which is filtered to recover Stage I
effluent.
                                 35

-------
                    10
15
20
25
30
                       SETTLING  TIME (min.)
Figure 8.  Stage I reactor effluent  settling data  (initial solids
                  concentration equal to 0.6%).

-------
                              SECTION VI

                 DATA INTERPRETATION AND DISCUSSION
REDUCTION OF EXPERIMENTAL DATA

     The raw data presented in Tables 4 and 5 have been interpreted
in terms of sulfate and calcium removed in Stage I of the alumina-
lime-soda process as functions of total iron and aluminum inputs and
of causticity of the Stage I effluent.  Molar ratios of sulfate/
(iron + aluminum) and calcium/(iron + aluminum) have been calculated,
in which metal used is the sum of the in situ iron and aluminum, and
aluminum added as sodium aluminate.  These ratios have been plotted
against the causticity of Stage I effluent and are shown in Figures
9 and 10.

     The shape of the curves in Figures 9 and 10 indicate that both
sulfate removal and calcium removal is stabilized when causticity of
Stage I effluent reaches about 600 mg/liter.  Thus, causticities
(from lime addition) in Stage I effluent above 600 mg/liter repre-
sent unnecessary use of lime, which in turn will increase carbon
dioxide requirements in Stage II.  For this reason, the process has
been designed around a causticity of 600 mg/liter (as CaC03) in
Stage I effluent.

     The 600-mg/liter breakpoint corresponds with a sulfate/(iron -f
aluminum) ratio of about 0.8, and a calcium/(iron + aluminum) ratio
of about 2.2.  These ratios suggest a stoichiometry in the Stage I
sludge of:
                           0.8S03-0.5R203-2.2 Ca               (6-1)

where    S03 = sulfate (anhydride)

        R203 - iron + aluminum

          Ca = calcium

                                  37

-------
              1-Or-
oo
          -   0.9
          UJ
          O


          £   0.8

          z

          o   0.7

          +

          £   0.6
O  0.5

I
>  0.4
O
          UJ
          UJ
          —I

          O
              0.3
              0.2
              0.1
                200
                      I	I
                                                     I
                 300
400        500         600        700


          CAUSTICITY (mg/literasCaCOa)
800
900
1000
                  Figure 9.   Sulfate removal plotted against  causticity in Stage I effluent.

-------
0,6 -
0.5
  200
300
                   400
 500       600

CAUSTICITY
                                                              1000
       Figure  10.   Calcium in  sludge plotted against causticity
                         in Stage  I effluent.
                                    39

-------
The coefficient of R20o is 0.5, since each mole of RoOo contains two
moles of metal.  If the calcium is distributed between sulfate and
oxide (and coefficients multiplied by 10 to clear decimal points),
the following stoichiometry of Stage I sludge is established.

                                                               (6-2)

     Closer examination of the data indicated that some sodium oxide
was probably present in the sludge.  As will be shown later, the high-
sulfate and low-sulfate forms of calcium sulfoaluminate and sulfofer-
rite consistently have CaO/^O-j ratios of 3.0.  Thus, it is likely
that a more proper representative of the Stage I sludge stoichiometry
is:

                                             0                 (6-3)

In this case, the extra mole of base needed to obtain the CaO/Al2Oo
ratio of 3.0 is found as Na20  originally present as sodium aluminate.

     The amount of sodium oxide in the sludge appears to be quite var-
iable.  A plot of sodium oxide/trivalent metal ratios yielded a set of
points too scattered to indicate a clear trend.  Thus, it is likely
that the composition of the Stage I sludge is more accurately
represented:

                     8CaS04-5R203'15(CaO, Na20)                (6-4)

     Conclusions concerning the effect of the two sources of triva-
lent metals on the sludge stoichiometry are discussed below.

     In the case of Stage I reactions, the iron and aluminum can be
obtained from two sources:  that contained in the raw AMD (in situ
iron and aluminum), and that added as sodium aluminate.  In order to
separate the effect of j.n situ iron and aluminum from that of the
sodium aluminate, experiments were conducted using solutions of iron
sulfate and aluminum sulfate in the MRI Kansas City laboratories.
These experiments indicated that when iron sulfate (either ferrous
or ferric) was treated with lime, sulfate was removed in approxi-
mately an 0.5 sulfate-iron ratio.  In a pure aluminum sulfate solu-
tion, sulfate was removed in a 1.5 sulfate-aluminum ratio.  In mixed
iron-aluminum systems, however, the sulfate/(iron + aluminum) ratio
of the sludge was about 0.6, indicating that in situ aluminum, when
                                   40

-------
mixed with in situ iron, is less effective for sulfate removal than is
pure aluminum sulfate.   Thus,  we conclude that each mole of in situ
iron and aluminum will  remove  about 0.5 mole  of sulfate.  The remain-
ing sulfate is removed  by aluminum added as sodium aluminate.

      With this information,  it was possible to reduce the empirical
formula (6-1) of the  sludge  to its components,  one component being
that generated by in  situ metals,  and  a second component generated by
sodium aluminate.   In the case of  Proctor No.  2 data,  the in situ
iron and aluminum constituted  about 407. of the total trivalent metals.
The constituent ratios  in (6-4) were adjusted by assuming that each
mole of JLn situ metals  would remove 0.5 mole  of sulfate, and subtract-
ing the appropriate constituents to isolate the effect of aluminum in
sodium aluminate (60% of  the total trivalent  metal content of the sys-
tem).   This exercise  yielded a composition of:

                      2CaS04'Al203'3(CaO,  Na20)                (6-5)

This  empirical  formula  can be  reduced  further to equal amounts  of the
so-called  "high-sulfate form"  and  "low-sulfate  form"  calcium sulfoalu-
minates.l/   Discussion  of  these two forms  of  sulfoaluminates  follows
on  p.  42.
                    2 |2CaS04-Al^-S^aO, Na20)|  	^

              CaS04-Al203-3(CaO, Na20)(low-sulfate  form)  +     (6-6)

              3CaS04-Al203-3(CaO,  Na20)(high-sulfate form)

     The preceding  interpretation  is based on results  obtained with
Proctor No. 2 acid mine drainage.   It is likely  that the  proportion
of high-sulfate and low-sulfate  forms of calcium sulfoaluminate pro-
duced by sodium aluminate will vary from one mine drainage to another.
In general, we would expect that the less iron in raw acid mine drain-
age, the greater the sulfate removal by sodium aluminate.  As will be
shown in the following discussion,   sodium aluminate will remove more
sulfate per mole of aluminum in other systems  where iron is absent in
the raw water.

-------
RELATION OF SLUDGE STOICHIOMKTRY TO OTHER SYSTEMS

     The alumina-lime-soda process has the ability to remove calcium
sulfate from the water through the precipitation of calcium sulfo-
aluminate materials having general compositions.

                       xCaS04"A12°3'3Ca°'yH2°                  <6~ 7)

The calcium sulfoaluminates have been known for many years; they are
insoluble materials which prevent "flash set" of portland cement con-
crete.  Indeed, gypsum is added to portland cement for the specific
purpose of combining with calcium aluminate in the hydration process
to retard the concrete set.  As a result, practically all data in the
technical literature pertaining to the calcium sulfoaluminate are as-
sociated with concrete hardening reactions.

     Two forms of calcium sulfoaluminate are recognized as being very
important in the set of concrete, and have been well characterized.—'
The two forms differ with respect to their sulfate-aluminum molar ra-
tios.  The so-called "high-sulfate" form has the composition:

                       3CaS04-Al203-3CaO-3lH20                 (6-8)

In this case, the sulfate-nluminum molar ratio is 1.5.  This material,
also known as ettringite, is the stable species when aluminate ion
reacts with calcium sulfate in a large excess of lime, such as would
be found in an ordinary mixture of portland cement and water.  The
second calcium sulfoaluminate is referred to as the "low-sulfate"
f«snm and has the composition:

                       CaS04'Al203'3CaO-12H20                  (6-9)

In tliis second case, the sulfate-aluminum ratio is only 0.5, since
1 mole (of calcium sulfate is associated with 1 mole of A1203 contain-
ing two atoms of aluminum.  Normally, this material is not found
after cement has hardened.
                                   42

-------
     In addition to the calcium sulfoaluminates, calcium sulfofer-
rites have also been reported.  However, these materials are less
well characterized.  The principal calcium sulfoferrites are the
ferric oxide (Fe20o) analogues of the high-form and low-form cal-
cium sulfoaluminates.-'  The principal reason for the lack of char-
acterization of the sulfoferrites appears to be due to the relative
unimportance of the iron constituents in portland cement, and to
the relative insolubility of iron oxide compared to aluminum oxide.

     Contrary to the portland cement system where calcium sulfate
and aluminate must be dissolved in limited amounts of water, the
alumina-lime-soda system involves the removal of dissolved sulfate
from water very much in excess.  Thus, the virtual instantaneous
formation of calcium sulfoaluminate in the portland cement-water
system does not occur in the alumina-lime-soda process.  As a re-
sult, about a 90 min residence time is needed in the alumina-lime-
soda process to complete formation of calcium sulfoaluminate.

     In the case of acid mine drainage, the water to be treated
contains significant quantities of dissolved iron and aluminum.
These materials will react as calcium sulfate scavengers if the
raw water is treated with lime in the pH 11.9 to 12.0 range.
Thus, acid mine drainage contains materials which can be used to
remove sulfate from the water without addition of sodium alumi-
nate.  In general, however, the dissolved iron and aluminum
contents of the raw AMD are insufficient to reduce sulfate levels
to 250 mg/liter, the maximum concentration permitted under drink-
ing water standards.

     The form of calcium sulfoaluminates and sulfoferrites pro-
duced when the alumina-lime-soda process is used for treating acid
mine drainage appears to be a mixture of the high-sulfate and low-
sulfate forms, with the latter species predominating.  This obser-
vation is contrary to those observed when the alumina-lime-soda
process is applied to brackish waters having calcium/magnesium
sulfate salinity,-/ and which contain no dissolved iron or alumi-
num.  In the OSW study, the calcium sulfoaluminate composition cor-
responded almost totally with the high-sulfate form, i.e., 3 moles
of calcium sulfate removed per mole of A1203.  The principal dif-
ference between the OSW work and the present study dealing with
acid mine drainage lies in the iron content in raw water.  Virtually
no iron was present in raw water used in the former study, while
iron concentrations were in the 20 to 100 ppm range for the present
study.
                                   43

-------
     We believe that the reason for the low-sulfate form is due to
part of the aluminum (from in situ aluminum or sodium aluminate)
becoming "encapsulated" by calcium sulfoferrite from the in situ
iron.  This encapsulation would make some of the aluminum unavail-
able for reaction with calcium sulfate to form ettringite.
                                  44

-------
                              SECTION VII
         ALUMINA-LIME-SODA PROCESS  CHEMISTRY FOR RECOVERING
                    WATER FROM ACID MINE DRAINAGE
      The  alumina-lime-soda  process  involves  two  stages  as  described
 earlier in  Section  IV.   This  section will describe  in detail  the
 several chemical  reactions  which occur  in the  two process  stages.
 The description is  based upon experimental results  attained with
 the Proctor No. 2 acid mine drainage.   In addition,  formulae  for
 estimating  quantities of the  treatment  chemicals will be presented.

 STAGE I - AMD REACTIONS  WITH  LIME AND SODIUM ALUMINATE

      There  are four basic reactions which occur in  Stage I:   neutrali-
 zation of acid; precipitation of metals; removal of  sulfate by in situ
 iron  and aluminum;  and removal of sulfate by sodium  aluminate.

 Neutralization of Acid

      The first reaction  that  occurs when lime  and sodium aluminate
 are added to raw  AMD is  the neutralization of  acid.  This neutraliza-
 tion  is simply:
                  H+ + OH" - ^ H20                        (7-1)
Some of the acid is due to the hydrolysis of dissolved iron and alumi-
num sulfates in the AMD.
                            3H20 - >Fe(OH)3 + 3H+          (7-2)

                                 - >A1(OH)3 + 3H+          (7-3)
     When lime is added, the acid from the hydrolysis is neutralized,
and iron and aluminum precipitate as their respective hydroxides.

                                        Fe(OH)3                (7-4)

                                        A1(OH)3                (7-5)

-------
Precipitation of Metals

     In general, only  the  trivalent metals will be precipitated during
the neutralization reactions,  i.e., raising  the pH to  the neutral
(pH =7.0) range.  When the pH is  raised  to  12.0, all  metals  in the
raw AMD, with the exception of the alkali metals  (sodium and  potas-
sium) and calcium, will precipitate.  Common metals  in mine drainage
include magnesium, manganese,  and  ferrous iron.   The precipitation
reactions are:

                        Mg"1"1" + 20H"	>Mg(OH)2                (7-6)

                                          Mn(OH)2                (7-7)

                                          Fe(OH)2                (7-8)

     As the pH is raised,  the  ferrous iron and manganous hydroxides
will rapidly oxidize to their  respective  trivalent hydroxides.

                   2Fe(OH)2 +  1/202 + H20	>2Fe(OH)3       (7-9)

                   2Mn(OH)2 +  1/202	>2MnO(OH) + 1^0       (7-10)

     Thus, the alumina-lime-soda process  overcomes problems asso-
ciated with oxidation of ferrous and manganous components normally
encountered in conventional neutralization processes where pH is
raised only near neutrality.

Removal of Sulfate with ±n situ Iron and  Aluminum

     The key reaction in the alumina-lime-soda process involves the
removal of calcium sulfate by  aluminate ion  at a system  pH of 12.0.
Results of the present study indicate that the iron  and  aluminum com-
ponents of the raw AMD will remove some calcium sulfate  in the Stage
I reactions.   Products generated by the in situ iron and aluminum ap-
pear to approximate the low-sulfate form  of  calcium  sulfoaluminate
and calcium sulfoferrite.   (A  discussion  of  the low-sulfate and high-
sulfate sulfoaluminates and sulfoferrites is presented in Section
VI.)  Reactions of in situ iron and aluminum thus can be represented:
  2Fe(OH)3 + bCa   + SO   + 60H" 	^ CaSO^'fe203'3CaO'xll20   (7-11)

  2A1(OH)3 + toa""" + S04= + 60H"	>CaS0'Al203'3CaO-xH20   (7-12)
                                  46

-------
     Methods for estimating the amount of sulfate that is removed by
the in s_itu iron and aluminum are discussed  later in this section and
in Appendix B.

Removal of Sulfate by Sodium Aluminate

     Normally, there is insufficient iron and aluminum content in acid
mine drainage to reduce sulfate to acceptable levels.  In order to in-
crease the sulfate removal of the process, sodium aluminate is added
to the system to further reduce the sulfate  concentration.

     When sodium aluminate is added, each mole of alumina will remove
 . mole of sulfate.  The reaction may be generalized as:
4(Na+Al02") + 903   + 4S04~ + 80H" —

                                         2Na+                   (7-13)

     The experimental results indicate that calcium oxide is removed
along with calcium sulfate in the calcium sulfoaluminates (and  sulfo-
ferrites) formed by sodium aluminate and by in situ iron and alumi-
num.  This lime is an essential part of the calcium sulfosalts, and
is needed to keep the removed calcium sulfate from resolubilizing in
the process.  In the sodium aluminate experiments, we found that so-
dium oxide, Na20, is also being removed.  Thus, we show 1 mole  of
Na20 removed with each 4 moles of CaSO^, 2 moles of A1203, and  5
moles of CaO in Eq. (7-13).  However, the data further indicate that
the exact amount of Na20 removed is highly variable.  Therefore, a
more proper representation of the composition of solids removed by
sodium aluminate is:
                   4CaS04-2Al203-6(CaO, Na20)-xH20             (7-14)
     In this representation, the variability of sodium oxide to
aluminum ratio can be accounted for.

     The composition (7-14) represents a mixture of the high-sulfate
and low- sulfate forms of calcium sulfoaluminate in a 50:50 mixture,
4CaS04-2Al203-6(CaO, Na20)'xH20

                            3CaS04 '&l203-3(CaQ, Na20)-xH20     (7-15)
                                  47

-------
     Note  that  the  first product has  the  same composition of  that
generated  from  the  in  situ aluminum,  the  low-sulfate calcium  sulfo-
aluminate.  The second product contains 3 moles of  sulfate  for every
mole of Al20q (high-sulfate  form), and has a composition similar to
that of ettringite,  the calcium sulfoaluminate responsible  for con-
trolling set in portland cement hardening.

STAGE II - CARBONATION

     The second stage  of the alumina-lime-soda process involves the
addition of carbon  dioxide to reduce  pH to acceptable levels.  The
carbonation step also  removes hardness since calcium carbonate is
precipitated in the  reaction.  The basic  chemical reaction  in Stage
II is:

                 20H"  + Ca"*" + C02	>CaC03 + H20           (7-16)

     In the actual process, effluent  from Stage I is mixed  with raw
mine drainage.   As has been discussed in  Section V, the sulfate lev-
els from Stage  I are about 100 mg/liter,  and causticity levels are
about 600 mg/liter as  CaC03.  Thus, sulfate concentrations  in prod-
uct water can be increased to higher  levels.  (Sulfate concentrations
of 250 mg/liter are permitted under U.S.  Public Health Service Stan-
dards.)  As a result,  raw AMD can be used to neutralize a part of the
excess causticity and  increase sulfate levels to 250 mg/liter.  The
partial neutralization of Stage I effluent involves the neutraliza-
tion of reactions, (7-1), (7-2), and (7-3), discussed earlier  and les-
sens the amount of carbon dioxide needed  for complete neutralization.

     The optimum pH for precipitating the most- hardness from  an alka-
line solution is 10.3.  At this pH, hardness is in the form of cal-
cium carbonate  at a 35 mg/liter concentration.  If excess C02 is ad-
ded to drop the  pH to  lower values, the precipitated calcium  carbonate
will redissolve  forming calcium bicarbonate.

                 CaC03 + C02 + H20	>Ca"H" + 2HC03"          (7-17)

The presence of  calcium bicarbonate will  result in a harder product
water than need  be, as well as  unnecessary consumption of  carbon
dioxide .•
                                  48

-------
      However,  if  the  precipitated  calcium  carbonate  is  removed while
 the  pH is  at  10.3,  the  subsequent  addition of C02 will  drop  the pH
 to the 7.0 to  7.5 range.   Since  only 35 rag/liter of  calcium  carbon-
 ate  is present in solution after separation, the amount of carbon
 dioxide needed to convert  dissolved carbonate to bicarbonate for
 dropping the pH is  minimized.
                    C03  + C02 + H20 - 2HC03"               (7-18)

     Thus, we have included in the process design provisions for car-
bonating Stage II effluent after its separation from calcium carbon-
ate in order to obtain product water meeting pH specifications.

CHEMICAL REQUIREMENTS FOR THE ALUMINA-LIME-SODA PROCESS

     The preceding discussion of alumina-lime-soda chemistry has
served as the basis for the development of formulae for estimating
chemical requirements for the process.  Details of the mathematical
development of the formulae are presented in Appendix B.  The for-
mulae have been developed on the basis of two sodium aluminates.
The first is referred to as "Dry-NaAl02" and is solid sodium alumi-
nate now commercially available.  The second is "Calcined" (Cal-
NaA102) and represents sodium aluminate produced by calcining a
mixture of soda ash and bauxite.  The production of Cal-NaAlOo is
discussed in Section IX of the report.

     The formulae for estimating chemical requirements for Stage I
of the alumina- lime -soda process are:
     Sodium Aluminate;
                                RS04-ES04-0 . 86Fe- 1 . 78A1
                gDry-NaAl02/j& -- - - -~g -

                                RS04-ES04-0.86Fe-1.78Al
                gCal-NaAl02/£ -- ~ -


where     Rso, = sulfate concentration in raw AMD,  mg/X

          ESO  ~ 8ul-fate concentration in effluent from Stage I,
             4
                                                               (7-19)


                                                               (7-20)
                                 49

-------
            Fe =  total  iron  concentration in raw AMD (in situ iron) ,
            Al = aluminum concentration in raw AMD (in situ aluminum),
                   mg/jfc

     Hydrated Lime. 937,;

     gCaO/j&:(Dry-NaAl02) =   0.00079  acid + 0.00332 Mg + 0.00145 Mn +

       0.00213 Fe + 0.00442 Al + 0.477J   + 0.621  (gDry-NaAl02/jO (7-21)

     gCaO/j&:(Cal-NaAl02) =   0.00079  acid + 0.00332 Mg + 0.00145 Mn +

       0.00213 Fe + 0.00442 Al + 0.477J   + 0.739  (gCal-Nal02/jO   (7-22)

where        acid = acidity of raw AMD,  mg/jfc  as CaCO?

               Mg = magnesium concentration in raw AMD, mg/j£
               Mn - manganese concentration  in  raw AMD,

               Fe = total jLn situ iron in raw AMD, mg/j£

               Al - total in situ aluminum in raw  AMD, mg/jj

       Dry-NaAl02 = dosage of dry sodium aluminate,  g/jfc

       Cal-NaAl02 = dosage of calcined sodium aluminate,  g/Jfc

     The choice of the formulae used depends upon  the kind of  sodium
aluminate used in Stage 1.  The formulae will yield  dosages in grams
per liter which is equivalent to kilograms per  1,000 liters for full-
scale systems.  Dosages calculated with the  formulae can  be converted
to English units (pounds per 1,000 gal) by multiplying results by  8.34.

STAGE II - RATIO MIXING OF STAGE I EFFLUENT/RAW AMD  AND CARBONATION

     The formula for estimating carbon dioxide  requirements in Stage
II of the process does not depend upon sodium aluminate type.   Rather,
                                  50

-------
it depends upon the causticity of the blend of Stage I effluent and
raw AMD for produce water having the desired sulfate concentration.
In establishing carbon dioxide requirements, one must also determine
the proportions of Stage I effluent and raw AMD in the blend.

     Carbon Dioxide;

                  gC02/jfc = 0.00046    CAUS + 35|                (7-23)

where     CAUS = causticity of the blend of Stage I effluent and raw
                   AMD, mg/4 as CaCO-j

     The causticity (CAUS) is the difference between the hydroxide
alkalinity of Stage I effluent and the acidity of raw AMD.  The ad-
ditional 35 rag/liter requirement in the formula represents that C02
which is used to drop the pH from 10.3 to the 7.0 to 8.0 range.

                    CAUS = EFF x 600 - AMD x acid               (7-24)

where     EEF = fraction of the blend that is Stage I effluent

          AMD = 1 - KKV = fraction of blend that is raw AMD

         acid = acidity of raw AMD, mg/4 as CaCO-j

          600 = hydroxide alkalinity of Stage I effluent, mg/jfc as
                  CaC03

     The fraction of the blend which is Stage I effluent is deter-
mined on the basis of sulfate values in the raw AMD, in the Stage I
effluent, and in the finished product water.  This term is calcu-
lated as follows:
                                  J4 -  S04
                          EFF =           *
                                RE
                                 SO, -  SO,
                                  51

-------
where     R__  = sulfate concentration in raw AMD,
R__
       sulfate concentration in finished product water,
         usually 250 m/A
ESO  = sulfate concentration in Stage I effluent,
   4     2 100 rag/A
                        52

-------
                             SECTION VIII

         CHEMICAL COSTS FOR THE ALUMINA-LIME-SODA PROCESS FOR
              SEVERAL ACID MINE DRAINAGES IN PENNSYLVANIA
      Principles of the alumina-lime-soda process have been applied
 to several acid mine drainages  in Pennsylvania.   Chemical  composi-
 tions of the raw AMD were used  in conjunction with formulae devel-
 oped for estimating chemical requirements discussed previously.
 Results of this exercise  are presented  in Tables 9 through 14.

      Also included in Tables 9  through  14 are estimated  chemical
 costs for treating the various raw mine  drainages to sulfate levels
 of 250 mg/liter.   The costs  are based on two  kinds of sodium
 aluminate--"dry"  sodium aluminate which  is available commercially,
 and "calcined"  sodium aluminate which is manufactured by heating a
 soda ash-bauxite  mixture  to  1000°C (1832°F).   Dry sodium aluminate
 contains  water  of crystallization and hence contains less  active
 alumina  than does calcined.   A discussion of  the  two kinds  of'so-
 dium aluminate  is presented  in Section IX.  As can be  seen  from the
 tables,  the  use of the  calcined sodium aluminate  is  preferred be-
 cause of  its  lower cost.

      The  chemical costs have  been  calculated  using  the following bases:
                Chemical                Cost ($/kfi)      Cost ($/lb)

           Dry sodium aluminate            0.298           0.135
           Calcined sodium aluminate       0.176           0.080
           Hydrated lime, 93%              0.044           0.020
           Carbon dioxide                  0.110           0.050
     The cost for dry sodium aluminate is that quoted by Reynolds
Metals Company for bulk purchases.  The cost for calcined sodium
aluminate is discussed in detail in Section IX.   The costs of
lime and carbon dioxide are estimated based upon current price
ranges of the two chemicals.

                                 53

-------
     Table 9.  PROCTOR NO. 1, HOLLYWOOD, PENNSYLVANIA, RAW AMD TREATED
             BY ALUMINA-LIME-SODA:  257.  TREATED IN STACE 1



PH
Acidity, mg/4 as CaC03
Sulfate, mg/£ as SO^
Chloride, rag/A as Cl
Bicarbonate, mg/£ as CaCOo
Calcium, mg/i as Ca
Magnesium, mg/A as Mg
Sodium, mg/£ as Na
Iron, mg/£ as Fe
Aluminum, rag /A as Al
Manganese, mg/£ as Mn

Before
treatments'
3.8
180
300
Unknown
0
6
15
8
30
20
Unknown
Commercial
After

treatment
7.0-8.0
0
250
Unknown
35
58
14
20
nil
nil
nil











dry Calcined
sodium aluminate sodium aluminate

Chemical requirements (kg/cu m) :
Carbon dioxide
Sodium aluminate
Hydrated lime, 93% Ca(OH)2
Total chemical costs ($/l,000 cu m)
Chemical requirements (lb/1,000 gal):
Carbon dioxide
Sodium aluminate
Hydrated lime 93% Ca(OH)~
Quantity
& j
0.02
0.04
0.24

Lb $/l
0.20
0.37
2.03
Cost Quantity
?/cu m Kg J
0.003 0.02
0.012 0.04
0.011 0.24
0.026
,000 gal Lb $1,
0.010 0.20
0.050 0.30
0.041 2.02
Cost
/cu m
0.003
0.006
0.011
0.020
000 gal
0.010
0.024
0.040
  Total chemical costs ($/l,000 gal)
0.101
0.074
a/  Raw AMD data from present study.
                                  54

-------
        Table 10.  COMMONWEALTH OF PENNSYLVANIA HAWK RUN AMD PLANT
                        PHILIPSBURG, PENNSYLVANIA
                RAW AMD TREATED BY ALUMINA-LIME-SODA:  757.
                            TREATED IN STAGE I



pH
Acidity, mg/* as CaC03
Sulfate, mg/o as SO^
Chloride, mg/« as Cl
Bicarbonate, mg/jf, as CaCOj
Calcium, mg/jfc as Ca
Magnesium, rog/£ as Mg
Sodium, mg/jK as Na
Iron, mg/£ as Fe
Aluminum, mg/j as Al
Manganese, tag/ 't as Mn
Before
treatment-'
3.7
384
648
Unknown
0.00
118
24
Unknown
101
Unknown
Unknown
After
treatment
7.0-8.0
0.00
250
Unknown
35
20
10
150
nil
nil
nil
                                        Commercial dry
                                       sodium aluminate
                                       Quantity    Coat
Chemical requirements (kg/cu m):

Carbon dioxide
Sodium aluminate
Hydrated lime, 93% Ca(OH)2

  Total chemical cost ($/cu m)

Chemical requirements (lb/1,000 gal):

Carbon dioxide
Sodium aluminate
Hydrated lime, 93% Ca(OH)2

  Total chemical cost ($/l,000 gal)
0.18
0.45
1.13
 Lb

1.54
3.74
9.38
0.02
0.13
0.05
                     Cnlcinwl
                 sodium aluminate
                 Quantity    Cost
0.18
0.36
1.12
 cu m

0.02
0.06
0.05
          0.20               0.13
0.08
0.50
0.19
1.54
3.03
9.30
0.08
0.24
0.19
          0.77
                   0.51
a/  Raw water data from:  R. Kunin and J.  J.  Demchalk,  "The Use of Am-
      berlite Ion Exchange Resins in Treating Acid Mine Waters  at Philips-
      burg, Pennsylvania," Paper presented at fifth  Symposium in Coal
      Drainage Research, Louisville, Kentucky,  October 1974.
                                    55

-------
           Table 11.  PROCTOR NO. 2, HOLLYWOOD, PENNSYLVANIA
                RAW AMD TREATED BY ALUMINA-LIME-SODA:  807.
                          TREATED IN STAGE I



pH
Acidity, mg/jfc as CaCO,
Sulfate, mg/£ as SO^
Chloride, mg/£ as Cl
Bicarbonate , mg/£ as CaCOj
Calcium, mg/j?, as Ca
Magnesium, mg/£ as Mg
Sodium, mg/£ as Na
Iron, mg/£ as Fe
Aluminum, mg/£ as Al
Manganese, mg/fl as Mn

Before
treatment*.'
2.8
700
750
Unknown
0
8
25
2
100
40
Unknown
Commercial













dry
sodium aluminate

Chemical requirements (kg/cu m) :
Carbon dioxide
Sodium aluminate
Hydrated lime, 93% Ca(OH)2
Total chemical costs ($/cu m)
Chemical requirements (lb/1,000 gal):
Carbon dioxide
Sodium aluminate
Ifydrated lime, 93% Ca(OH)2
Quantity
Cost
After
treatment
7.0-8.0
0
250
Unknown
35
20
10
145
nil
nil
nil
Calcined
sodium aluminate
Quantity Cost
Kg $/cu m K£ £/cu m
0.18
0.51
1.58

Lb $/l
1.48
4.26
13.17
0.02
0.15
0.07
0.24
,000
0.07
0.58
0.26
0.18 0.02
0.41 0.07
1.56 0.07
0.16
gal Lb $/ 1,000 gal
1.48 0.07
3.46 0.28
13.03 0.26
  Total chemical costs ($/l,000 gal)
0.91
0.61
a/  Raw AMD data from present study.
                                   56

-------
               Table  12.  SAWMILL RUN, RAW AKD TREATED BY
               ALUMINA-LIME-SODA:  85% TREATED IN  STAGE  I


PH
Acidity, mg/A as CaC03
Sulfate, fflgAe as SO^
Chloride, mg/A as Cl
Bicarbonate, tag/ A as CaCO-j
Calcium, mg/4 as Ca
Magnesium, mg/jfc as Mg
Sodium, mg/4 as Na
Iron, mg/jl as Fe
Aluminum, mg/jj as Al
Manganese, mg/A as Mn

Before
treatment^
3.2
690
1,020
40
0
109
81
Unknown
28
8
5
Commercial













dry
sodium aluminate

Chemical requirements (kg/cu m):
Carbon dioxide
Sodium aluminate
Hydrated lime, 937. Ca(OH)2
Total chemical cost ($/cu m)
Chemical requirements (lb/1,000 gal):
Carbon dioxide
Sodium aluminate
Hydrated lime, 93% Ca(OH),
Quantity
M
0.21
0.97
1.85

Lb $/l
1.74
8.11
15.40
Cost
$/cu m
0.02
0.29
0.08
0.39
,000 «al
0.09
1.09
0.31
After
treatment
7.0-8.0
0
250
40
35
30
15
270
nil
nil
nil
Calcined














sodium aluminate
Quantity
SS. j
0.21
0.79
1.83

Lb •$/!,
1,74
6.57
15.22
Cost
t/cu m
0.02
0.14
0.08
0.24
000 gal
0.09
0.53
0.30
  Total chemical cost ($/l,000 gal)
1.49
                   0.92
a/Raw AMD data from Commonwealth of Pennsylvania Department of Environ-
      mental Resources.
                                   57

-------
Table 13.  YOUNG AND SON COAL CORPORATION, PARKER'S LANDING, PENNSYLVANIA
               RAW AMD TREATED BY ALUMINA-LIME-SODA:  907.
                          TREATED IN STAGE I



PH
Acidity, mg/fl as CaCO-j
Sulfatc, mg/£ as SO^
Chloride, mg/l as Cl
Bicarbonate, mg/£ as CaC03
Calcium, mg/jfc as Ca
Magnesium, mg/£ as Mg
Sodium, rag/A as Na
Iron, mg/jfc as Fe
Aluminum, mg/4 as Al
Manganese, mg/jj as Mn



Chemical requirements (kg/cu m) :
Carbon dioxide
Sodium aluminate
Hydra ted lime, 93% Ca(OH)2
Total chemical costs ($/cu m)
Chemical requirements (lb/1,000 gal):
Carbon dioxide
Sodium aluminate
Hydrated lime, 93% Ca(OH)2
Before
treatment^/
2.6
785
1,360
Unknown
0
138
58
Unknown
225
38
4
Commercial dry
sodium aluminate
Quantity Cost
Kg $/cu in
0.24 0.03
1.17 0.35
0.24 0.03
0.49
Lb $/l,000 gal
1.96 0.10
9.72 1.31
21.35 0.43
After
treatment
7.0-8.0
0
250
Unknown
35
28
10
300
nil
nil
nil
Calcined
sodium aluminate
Quantity Cost
Kg $/cu m
0.24 0.03
0.95 0.17
0.24 0.03
0.31
Lb $/ 1,000 gal
1.96 0.10
7.89 0.63
21.13 0.42
  Total chemical costs ($/l,000 gal)
1.84
1.15
&f  Raw AMD data from Commonwealth of Pennsylvania Department of Environ-
      mental Resources.
                                    58

-------
       T*ble 14.  BETHLEHEM MINES CORPORATION MAR1ANNA MINE NO. 58
                         MARIANNA, PENNSYLVANIA
               RAW AMD TREATED BY ALUMINA-LIME-SODA:  100%
                           TREATED IN STAGE I

Before

treatments/
PH
Acidity, mgAe as CaC03
Sulfate, rag/ i as SO^
Chloride, mg/ji as Cl
Bicarbonate, mg/4 as CaC03
Calcium, mg/£ as Ca
Magnesium, mg/£ as Mg
Sodium, mg/£ as Na
Iron, mg/4 as Fe
Aluminum, mg/jfc as Al
Manganese, mg/A as Mn

2.64
4,080
10,000
1,830
0
444
404
2,420
815
475
38
Commercial











dry
sodium alumlnate

Chemical requirements (kg/cu m):
Carbon dioxide
Sodium aluminate
Hydrated lime, 93% Ca(OH>2
Total chemical costs ($/cu m)
Chemical requirements (lb/1,000 gal):
Carbon dioxide
Sodium aluminate
Hydrated lime, 93% Ca(OH),
Quantity
£S
0.3
10.1
15.2

Cost
$/cu m
0.03
3.01
0.67
3.71
Lb $/l,000 gal
2.5
84.5
127.0
0.13
11.41
2.54
mBV_9»^H
After
treatment
7.0-8.0
0
250
1,830
35
14
3
4,300
nil
nil
nil
Calcined














sodium aluminate
Quantity
K£ j
0.3
8.2
15.0

Lb $/l,
2.5
68.5
125.1
Cost
5/cu m
0.03
1.45
0.66
2.14
000 gal
0.13
5.48
2.50
  Total chemical costs ($/!.,000 gal)
14.08
                    8.11
a/  Raw AMD data from Commonwealth of Pennsylvania, Department of Environ-
      mental Resources.
                                    59

-------
      The keys  to  the  costs  of  the alumina-lime-soda  lie  in  two  fac-
 tors:   the  sulfate content  of  the raw mine drainage, which  determines
 the  fraction of mine  drainage  treated by alumina-lime-soda; and the
 amount  of in situ iron and  aluminum contained within the mine drain-
 age.  These two factors determine the quantity of  sodium aluminate
 which is required for the process;  and  since sodium aluminate is the
 most  expensive treatment chemical of the process,  its  requirement
 has  the greatest  effect on  chemical costs.

      It would  appear  from the  data  that sulfate concentrations  in the
 1,100-  to 1,200-mg/liter range are  a likely upper  limit  for utilizing
 the alumina-lime-soda process.  In  this range, about 857. of the raw
 AMD is  treated by alumina-lime-soda, and overall chemical costs lie
 in the  range of $0.25 to $0.30/cu m ($0.95 to $1.15/1,000 gal)  of
 product water.

      The chemical cost estimates in Tables 9 through 14  were made in
 the following  manner.  First,  the fraction of Stage I  effluent  needed
 for the final  blend was computed based on sulfate  concentrations in
 raw AMD, sulfate  in the blend  at 250 mg/liter, and sulfate  in the
 Stage I effluent  of 100 mg/liter.

     Second, sodium aluminate  and lime dosages were determined  using
 formulas as developed in Appendix B.  These dosages were multiplied by
 the fraction treated and by the factor 1.05 to account for water los-
 ses in  the  Stage I sludges.

     Third, the causticity of  the blended Stage I  effluent and  raw
AMD was calculated for the carbonation step.  From this value,  the
requirement for carbon dioxide was  estimated.  The carbon dioxide
requirement was multiplied by  1.03  to account for water  losses  in
the Stage II CaC03 precipitate.

     These estimates of treatment chemical quantities were then mul-
tiplied by the appropriate chemical cost and the results summed  to
yield the chemical cost values presented in Tables 9 through 14.
                                  60

-------
                              SECTION IX

          SODIUM ALUMINATE FOR THE ALUMINA-LIME-SODA PROCESS
     The key  to alumina-lime-soda process economics  lies  in  the cost
of  sodium aluminate.   Sodium aluminate is a chemical which has lim-
ited application as a  flocculating agent, and as a source of spe-
cialty alumina chemicals, e.g., catalysts.  As a result,  the demand
for sodium aluminate has been somewhat limited.  In  1972, the pro-
duction of sodium aluminate was estimated to be about  16,400 metric
tons (18,000  tons)/year.

     The most recent price of dry sodium aluminate is  $250.00/metric
ton ($275.00/ton).  This quote from Reynolds Metals  Company  on July 1,
1975, represents a bagged cost, and they will discount the price
$4.55/metric  ton ($5.00/ton) if shipped on a bulk basis.  Thus, the
current sodium aluminate price is $0.176/kg ($0,135/lb) on a bulk
basis.  Liquid sodium aluminate costs $0.194/kg ($0.088/lb)  on an
18,000-kg (40,000-lb) basis (Nalco Chemical Company).  When  con-
verted to a dry basis, the cost equivalent today of  sodium aluminate
is  about $0.44/kg ($0.20/lb), rendering liquid sodium  aluminate non-
competitive with the solid in full-scale plant operations.   Thus,
liquid sodium aluminate was not considered in cost estimations.

     If a number of alumina-lime-soda plants using mine drainage
with sulfate concentrations in the 1,100- to 1,200-mg/liter  range
totaling 19 million liters (5 million gallons) per day were  built,
one would need approximately 5,500 to 7,300 metric tons (6,000 to
8,000 tons) per year of dry sodium aluminate.   If the alumina-lime-
soda process was  applied to areas having other kinds of problems,
e.g.,  high calcium sulfate waters in the western states, the demand
for sodium aluminate would increase further.   This quantity  repre-
sents about a 50% increase in the demand for the chemical.   In ad-
dition, high grade sodium aluminate would not be required.   Thus,
it is likely that the price would drop substantially by the  time
the process is developed to full scale.
                                  61

-------
     An alternative approach to sodium aluminate would be its manu-
facture by calcining bauxite and soda ash,,—' rather than by its
crystallization from sodium aluminate solution.  Process principles
are almost identical with those associated with portland cement
manufacture, except that the temperature of the rotary kiln is in
the 900 to 1100°C  (1650 to 2000°F) range rather than in the 1300 to
1500°C (2350 to 2700°F) range.  Approximate composition range of so-
dium aluminate produced from calcining bauxite-soda ash mixtures will
be:  Na2°» 37 to 40%; A12<>3, 52 to 54%; and Si02, 5 to 77,.

     Sodium aluminate produced by the calcination of bauxite and soda
ash would be less pure than that presently produced commercially.
Bauxite contains silicious impurities which result in the formation
of sodium silicate as well as sodium aluminate.  However, the soda
fraction of the sodium silicate is useful in the alumina-lime-soda
process so its presence is not detrimental.  The silica will be
precipitated in the sludge and not appear in the product water.
The important constituent in the product is the sodium aluminate,
since the alumina fraction is all important for sulfate removal.

     The calcining of sodium carbonate and bauxite is a straight line
process closely resembling the manufacturing process for cement.  The
same general types of equipment used in cement manufacturing are
needed for calcining soda ash and bauxite to yield sodium aluminate,
carbon dioxide, and water.  The general types of equipment are:
crushing, grinding, rotary kiln, and conveying and handling.

     A cost estimate for a new 18,200 metric ton (20,000 ton)/year
sodium aluminate plant based on calcining of soda ash and bauxite
was made using cost data from previous years converted to 1975
dollars by the use of Marshall and Stevens Equipment Cost Index
and the Construction Cost Index.

     The selling price of sodium aluminate (857. purity) is calcu-
lated to be $0.176/kg ($0.08/lb).   This price compares to sodium
aluminate produced as a specialty product (72% purity) of $0.298/
kg ($0.135/lb).

     The cost estimate is given in Table 15.
                                 62

-------
    Table 15.  COST ESTIMATE:  20,000 TONS PER YEAR SODIUM ALUMINATE
                  MANUFACTURING PLANT (CALCINE PROCESS)
Capital costs

Equipment delivery cost
  Installation (1.05)
Buildings
Land [16.2 hectares (40 acres)]

  Subtotal
Contig. at 20%
Operating capital

  Total capital cost

18,200 metric tons (20,000 tons)/year, 350-

Operating costs per day

Raw materials and fuel
Utilities
Labor - 5 men at $3.75/hr
Supervision
Maintenance at 20% equipment cost
Capital cost at 7%
Depreciation at 6% equipment
             at 2% building
Overhead and payroll taxes at 11%
Selling and advertising
Taxes and insurance

  Total operating cost

  Total sales [at $0.176/kg ($0.08/lb)]

  Gross profit

  Gross profit
          $  500,000
             525,000
              80,000
             200.000

           .1,305,000
             261.000

           1,566,000
             434.000

          $2,000,000

day operation, 59,687 kg/day
            (132,638 Ib/day)


          $    4,933.38
                 500.00
                 462.85
                  35.71
                 277.77
                 400.00
                  85.71
                   4.57
                 509.14
                 663.19
                 636.66

          $    8,568.98/day

          $   10,611.04/day

          $    2,042.06/day

          $  714,721.00/year
                                  63

-------
      Calcined sodium aluminate can be manufactured by utilizing ex-
 cess capacity of a portland cement plant.  The equipment and person-
 nel need for the production of sodium aluminate from calcination of
 soda ash and bauxite are the standard pieces of equipment used to
 manufacture cement.   A cement manufacturer producing sodium alumi-
 nate in his plant could cut the cost below the calculated cost for
 a new plant.

      Raw material and fuel costs are based on the following assump-
 tions.   Hie bauxite  raw material is assumed to be high grade having
 an alumina content of 55% and a silica content of 7% or less.   Suf-
 ficient soda ash is  mixed with the bauxite to form both sodium alu-
 minate  and sodium silicate.   We have estimated bauxite costs to be
 $33/metric ton ($30/ton) for the low-silica material;  the cost of
 soda ash is estimated to be  $55/metric ton ($50/ton).   These cost
 estimates have been  made from recent data from the U.S.  Bureau of
 Mines and the  Chemical Marketing Reporter.   After the  bauxite  and
 soda ash are mixed,  the feed is fired at  about 1000°C  in a rotary
 kiln.   The carbon dioxide  evolved during  the reaction  is collected
 for  its use in the alumina-lime-soda process,  or  for its purifica-
 tion and resale.

      Thus,  we  conclude  that  sodium aluminate produced  by calcining
 bauxite  and soda  ash  is  the  product  which  should  be  sought  for  the
 full-scale  use  of the alumina-lime-soda water  treatment  process.
 Impurities  in  the product will  not  interfere with  the  process,  and
 it will  be  anhydrous.   The anhydrous nature  of the bauxite-soda ash
 process  is  an advantage  since "dry"  sodium aluminate currently  dis-
 tributed contains about  25 to 30% water of crystallization.  The
 added alumina in  the calcined product  (52 to 54% compared to 42 to
43% in the crystallized product) means that  less calcined sodium
aluminate is needed compared to the dry sodium aluminate.
                                 64

-------
                               SECTION X

                  ALUMINA-LIME-SODA PROCESS  ECONOMICS
      This section presents  estimates  of  construction  and  operational
 costs for treatment  of AMD  water  using the  alumina-lime-soda  process.
 Raw AMD considered for  this economic  analysis was assumed to  be  equiv-
 alent to Proctor  No.  2.   Plant  capacities selected  for the development
 of  the cost  estimates are 0.5 MGD,  1.0 MGD, and  5.0 MGD.*  Almost all
 treatment plants  needed  for AMD water treatment  should fall within  the
 0.5 to 5.0 MGD range; therefore,  the  cost estimates should be repre-
 sentative.

      The  costs  developed  are based  on incorporating the following unit
 operations in  the  alumina-lime-soda process:

      *  Pump station
      * Chemical mix basin
      *  Reacting basin
      *  Settling basin
      * Carbonating basin
      * Final storage basin
      * Chemical feed system (storage tanks and feeders)
      * Vacuum  filter

All basins were assumed to be constructed of reinforced concrete.
Most  mixers,  scrapers, baffling and structural supports would con-
sist  of corrosion-resistant  materials due to alkaline conditions  ex-
pected throughout the treatment process.   One of the major equipment
costs  is the vacuum filter to separate Stage I effluent from the  cal-
cium  sulfoaluminate/sulfoferrite sludges.  With these assumptions,
the total estimated construction costs (1975)  for three plants of
*  1 MGD = 3,780.000 liter/day = 3.780 cu m/day.
                                 65

-------
differing capacity utilizing  the alumina-lime-soda process  are  shown
below.
       Plant capacity  (MGD)          Total construction costs

               0.5                           $  352,000
               1.0                           $  516,000
               5.0                           $1,382,000


      These costs  include contingencies  (10%), but  exclude engineering
and  legal costs.  These costs are displayed graphically in Figure 11.

      The operation and maintenance costs  (excluding depreciation) for
the  three plant sizes are estimated below.  Costs  are given in  $/l,000
gal.  The chemical costs shown are based  on treatment of Proctor No. 2
AMD.
                         0.5 MGD         1.0 MOD         5.0 MGD

Power
Labor
Chemical
Sludge disposal

Total ($/l,000 gal)           1.04           0.92             0.79
      ($/cu m)                0.27           0.24             0.21

Annual costs           $189,800       $335,800       $1,441,800
     Figure 12 gives the graphical representation of the operation
and maintenance costs.

     Sodium aluminatc manufactured by calcining bauxite and soda ash
is assumed in the chemical cost estimates.  The cost of solids
disposal is assumed to be $5.00/ton of material having a dry solids
content of 10%.
                                  66

-------
   2.Or
                      TOTAL CONSTRUCTION COSTS
_o

5
   1.0
   0.5
 o
u

   0.2
   0.1
       0
                              2
3
                          Plant Capacity (MGD)

                     (1 MGD~3800 cu. meters/day)
               Figure 11.   Total construction costs.
5
                                67

-------
     2.0
 I/I
 o
U
 u
 c
 O
 C
 8.
o
 c
     1.0
     0.5
     0.2
     0.1
        0
                        TOTAL ANNUAL OPERATION

                        AND  MAINTENANCE COSTS
                             Plant Capacity (MGD)

                        (1 MGD  3800 cu. meters/day)
      Figure 12.  Total annual operation and maintenance costs,
                                 68

-------
                               SECTION XI

                         DISCUSSION OF RESULTS
      Hie field studies conducted at the  Hollywood facility resulted in
 an accumulation of about 800 hr of operating experience with the field
 unit.   This experience amply demonstrated the basic operability of the
 alumina-lime-soda process in acid mine drainage,  yielded the informa-
 tion and data needed to fully describe the process in terms of mine
 drainage,  and firmed up various operating parameters needed to design
 a Phase II demonstration plant.   The study also yielded information
 which permitted the assessment of costs  of treating acid mine drainage
 with the alumina-lime-soda process and generally  compared this process
 with other processes which are either in use or under investigation
 for mine drainage treatment.

      The data accumulated during the program have demonstrated that
 the alumina-lime-soda process will produce water  meeting drinking
 water standards from acid mine drainage.   It is particularly useful
 for mine drainages  having sulfate concentrations  in the 400 mg/liter
 to 1,200 mg/liter range.   Above  1,200 mg/liter, alumina-liroe-soda
 costs  become  excessive because of increased treatment-chemical con-
 sumption and  because  of the  larger quantity of  water which must be
 treated in the  first  stage  of the process.   Below 400 mg/liter,  raw
 mine drainage can be  treated  with lime to  produce water suitable for
 many uses,  and  almost suitable as drinking water.

     The alumina-lime-soda process  is particularly attractive  when com-
 pared  to other water  recovery  processes,  e.g.,  ion exchange or reverse
 osmosis.   Alumina-lime-soda desalting depends strictly upon chemical
 processes  and thus can  be operated using conventional  equipment and
 procedures.   There are  no resins  to backwash or regenerate, nor are
 there chemicals to be  recovered and recycled as is  found in ion ex-
 change.  There are no membranes to foul,  nor is pretreatment of raw
water required as is  the case with reverse osmosis.  In the alumina-
 lime-soda process, removed constituents are found  in easily dewatered
 solids and are in a readily disposable form.  There are no waste li-
quid streams needing treatment for disposal.

                                  69

-------
     Scaling was not a serious problem in the alumina-lime-soda process
during the Hollywood tests.  During the course of the study (about 800
hr of operating experience), we did not observe any significant scaling
of stirrcrs or reaction vessels, nor clogging of pipes or pumps.  The
precipitates formed during the process act as seeds for the reaction
products, thus discouraging scale formation.

     Since Stage I of the process operates under alkaline conditions
(pH of 12.0), the reduced forms of dissolved iron and manganese are
rapidly oxidized in the process.  This feature ensures essentially com-
plete removal of heavy metals which may be present in the raw acid mine
drainage.

     Analysis and interpretation of experimental data pertaining to
process reactions have resulted in development of a good scientific
rationale for reactions and products.  This rationale will be gener-
ally applicable to mine drainage and has been organized to facilitate
calculations of chemical and mass balances in mine drainage in general.
The iron and aluminum present in raw mine drainage participate in
alumina-lime-soda reactions to precipitate calcium sulfate and there-
fore serve a useful purpose in the process.  Each mole of iron or alu-
minum present in the raw water will remove about one-half mole of sul-
fate as calcium sulfoferrite or sulfoaluminate.   The rate of removal
by aluminum is somewhat smaller than expected, possibly due to iron
encapsulation of the aluminum.  Aluminum added to the process as sodium
aluminate, on the other hand, removes sulfate in a 1:1 sulfate/aluminum
molar ratio.  Hence, sodium aluminate is more effective than the in situ
aluminum in raw acid mine drainage towards the removal of sulfate.

     The cost of the alumina-lime-soda process is most sensitive to
the cost of sodium aluminate.  In addition, the operating characteris-
tics of the process are dependent on the form 'and stability of sodium
aluminate supplied.  Pure sodium aluminate solutions have been found to
be unstable  and will rapidly precipitate aluminum hydroxide upon
standing.  Sodium aluminate solutions can be stabilized, however, by
the addition of excess caustic soda.  This stabilization is expensive
and adds sodium ions to the product water.  For these reasons, dry
sodium aluminate is preferred for use in full-scale plants.  Our search
for a source of dry sodium aluminate led us to an old process that
meets the alumina-lime-soda process requirements quite well; it can
                                   70

-------
 furnish  sodium aluminate at a  price considerably better  than  that cur-
 rently quoted  by manufacturers of  Bayer-process alumina  products.

      The process consists of calcining high grade bauxite and soda ash.
 Experimental results  suggest that  a high purity sodium aluminate is
 unnecessary for the successful utilization of the alumina-lime-soda
 process.  About 50 years ago,  an accepted method for manufacturing
 sodium aluminate was  the calcination of bauxite-soda ash mixtures.  This
 method fell from favor because of  silica impurities in the product.  For
 the alumina-lime-soda process, however, the silica impurities are not
 deleterious, and sodium aluminate  manufactured by the calcination pro-
 cess  is  recommended.  The cost of  sodium aluminate manufactured by cal-
 cining bauxite and soda ash is estimated to be about $0.176/kg ($0.08/
 Ib),  compared  to $0.298/kg ($0.135/lb) for the dry sodium aluminate cur-
 rently being marketed.

      The process which evolved out of the field studies  at the Hollywood
 mine  drainage  facility involves two stages,  in Stage I, a majority of
 the input mine drainage is reacted with lime/sodium aluminate in a
 stirred  reactor at a  pH of 12.0.   This system yields a precipitate of
 calcium  sulfoaluminates and calcium sulfoferrites which  is permitted to
 settle.  The supernatant liquid is transferred to Stage  II of the pro-
 cess; the settled sludge is filtered on a sand filter, and the filtrate
 recombined with the supernatant liquid.  In Stage II of  the process,
 the filtrate plus supernatant  liquid is combined with the remaining
 fraction of the mine  drainage  in a reactor which completes the overall
 process as follows:   the acid  in the mine drainage is neutralized with
 excess causticity in  the Stage 1 effluent; metal oxides  and hydroxides
 are precipitated from the mine drainage; and carbon dioxide is added in
 sufficient quantities to reduce the pH to 10.3 and precipitate calcium
 carbonate.   The liquid is then filtered to yield a second sludge plus
 an effluent which is  lightly carbonated to a pH of 7 to 8.   The relative
 proportions of mine drainage admitted to Stages I and II of the process
 are determined by the desired  sulfate levels in the product water.   For
 Proctor No.  2 mine drainage at the Hollywood facility, approximately
 three-fourths is treated in Stage I and the remaining one-fourth is
added to Stage II to yield product water having sulfate concentrations
of 250 ppm or less.  This method of operating the process minimizes the
requirements for all the treatment chemicals--lime,  sodium aluminate,
                                  71

-------
and carbon dioxide, and thus minimizes the costs for operating the
process.

     The alumina-lime-soda process uses the same principles as conven-
tional waterworks  processes using lime-soda ash softening.  The princi-
pal difference between the two processes is that alumina-lime-soda op-
erates at a pH of  12.0 rather than the 9.5 to 10.3 range in lime-soda
ash softening.  The similarity in the water treatment processes will
result in utilization of standard water treatment-type equipment for
the alumina-lime-soda process.  As a consequence, capital and operating
costs for alumina-lime-soda will be similar to those associated with
conventional lime-soda ash treatment.

     Costs of chemicals for treating various acid mine wastes by the
alumina-lime-soda  process have been calculated.  Comparison of these
costs plus estimated operating costs indicate that the alumina-lime-
soda process is quite competitive with other processes which are being
considered for mine drainage treatment.  The Proctor No. 2 water (acid-
ity of 700 mg/liter, sulfate of 750 mg/liter) is representative of mine
drainages which are candidates for treatment by the alumina-lime-soda
process.  The cost for chemicals for Proctor No. 2 water is estimated
to be $0.16/cm3 ($0.61/1,000 gal).

     The operation and maintenance costs for recovering water from
Proctor No. 2 AMD  have been estimated for plants having capacity of
1,900 cm3/day (0.5 MGD); 3,800 cra3/day (1.0 MGD) and 19,000 cm3/day
(5.0 MGD).   The respective water recovery costs are:  $0.27/cm3 ($1.04/
1,000 gal), $0.24/cm3 ($0.92;1,000 gal), and $0.21/cm3 ($0.79/1,000
gal).

     A design for a 190,000 liter per day (50,000 gal/day) demonstra-
tion plant has been developed and is presented in this report.   The
estimated cost for procuring and constructing this plant is $95,000 to
$100,000.  Operation of such a plant for a period of 1 year is recom-
mended to further demonstrate the process and obtain operating and cost
information suitable for design of full-scale plants.
                                  72

-------
     In summary:

     The alumina-lime-soda process operated well on two mine drainages
of moderate acidity and will yield water of drinking water quality.

     The process requires only conventional water treatment equipment
and thus will be easy to adapt to large scale.  Operation will require
no special skills not presently found in municipal water treatment
practice.

     Water yields are essentially quantitative, with no brines for dis-
posal or recycle streams needing recovery.  By-product sludges are sim-
ilar to water treatment sludges.

     Detailed engineering design data for bulk scale plants (up to about
5 MGD for acid mine drainage) are presently lacking.  A pilot plant
sized at about 190,000 liter per day (50,000 gal/day) is suitable for
obtaining detailed engineering specifications and firming up cost data
for construction and operation of bulk scale plants.
                                  73

-------
                             APPENDIX A

              SPECIFICATIONS FOR THE 190,000 LITER/DAY
                 (50,000 GAL/DAY) DEMONSTRATION PLANT
SYSTEMS DESIGN

     The demonstration plant design is based on analytical results from
laboratory tests and material balances based on the production of
190,000 liters/day  (50,000 gal/day) of potable water.  A flow sheet
showing the system  configuration is presented in Figure 13.  The spa-
tial layout of the  plant is shown in Figure 14.  The proposed continu-
ous system features a chemical reactor, settler, carbonator, and sand
filters.  The chemical reactor Is a continuous stirred tank reactor
in which the lime and sodium aluminate are mixed and reacted with the
AMD water.  The pH  in the reactor is maintained between 11.9 and 12.1
by lime addition.   The settler and primary sand filter are solids
separation units for removal of the suspended solids formed in the
reactor.  The carbonator is also a continuously stirred tank where
the pH and sulfate  concentration are adjusted by carbon dioxide addi-
tion and blending untreated AMD water with clarified settler--primary
filter effluent.  A final sand filter is needed for removal of resid-
ual solids formed in the carbonator.

     The material balances were solved to determine the flow rates
through each segment of the system.  The following material balances
were used to determine component flow rates,  Xj , in the system de-
picted in Figure 15.

where     X. - water flow rate

           a = solids weight fraction

           b = sulfate weight fraction

Water Balances

     Overall:   Xx = X? + Xu + 50,000

                                  74

-------
          Row

          Wot*
X*l	^
01
                              Sodium-Alum i.iofi




                           Uf^Uul
                           ft Lim.
                                 i
                                 I
                                                                                       (*)  Soli* Iron Sood F:|I«. Scraoed of! 
-------
                                     40'
40'
                         FINAL STORAGE
                            TANK
                           (4'O.D.)
                                                                          40'
                                                              SCAlh )/?'• 2'
                                   40'
          Figure 14.   Spatial layout of the  190,000  liter/day
                 (50,000 gal/day)  demonstration plant.
                                     76

-------
                                                       X7,o7
     50,000gpd,
                 XI  = Water Flow Rate (gpd)
                 ai = Solids Weight Fraction
                 bi = Sulfate Weight Fraction
Figure 15.  Flow diagram to determine material balance.
                          77

-------
     Junction  1:  X^  = X2  + X3

     Settler:  X^ = X5 + Xfi

     Primary filter:  X5 = Xy + Xg

     Junction  2:  Xg  = X6  + Xg

     Carbonator:  X^Q = X   + X^

     Final filter:  X  =  Xu + 50,000

Solids Balances

     Settler:  a^X^ = a5X5 + a&X&

     Primary filter:  a_X_ = a?X  + ag

     Final filter:  a1QX[Q = a^u +

Sulfate Balance

     Carbonator:  b1QXin = b X.j + bgXg

     The equations can be  combined  algebraically to yield the follow-
ing expressions for calculation of  flow rates,   X.^ , in the system.

                X2 =  50,0007 Q(l -  K3)(l -  KjK2)(l + K4)j

where     ^ - (a4 -  a6)/(a5 - a6)

          K2 = (a5 -  afl)/(a7 - aQ)


          K3 = 
-------
The final filtration values for  a^  and  a-^2  were estimated as 0.05
and 0.0, respectively.

          a.  = 0.0060               a0 = 0.0
           4                         o

          a  = 0.0190               a   = 0.0005

          a, = 0.0002               a.,., = 0.05
           o                         11

          a? = 0.10                 a12 = 0.0

     The  b-j  value used for design was 0.00075  (750 ppm), which is
the sulfate concentration of Proctor No. 2 AMD water.

          X3 = K4*9

          X10 = X3 + X9

          Xll = K3X10

     The  K^  ,  K2  and  K-  factors were calculated using  solids
concentration data from settling and  filtration  tests  conducted in
the laboratory.  It is notable  that the  K   factors are  dependent
on concentration only and independent  of flow.   K^  can  be  calcu-
lated from experimental settling data, while Ko and  £3  are
calculated from filtration data.

     K/   is  the ratio of  the  flow  rates  of  streams  X-j  to   X9  as
they enter  the carbonator.  The two  streams are  mixed  to adjust  the
pH and  sulfate concentration.   For design purposes  the drinking water
standard  of  250 ppm  (wt.  fraction  =  0.00025) was selected as the  sul-
fate concentration in the carbonator  effluent.   Laboratory  results
indicate  that  the  reactor should be  operated to  maintain a  100 ppm
(wt. fraction  = 0.00010)  residual  sulfate concentration. This con-
centration will not change in  the  settling  and  filtration steps;
therefore,  the values  of  bg   and  b^Q may be  set  at  0.00010 and
0.00025,  respectively.  The expression for   K4   may now  be  expressed
as a function of the  sulfate concentration  of the untreated AMD water,
b3   , as  follows:
                                  79

-------
                         v  =     0.00015
                         K4   	
                              (b3 - 0.00025)

     The system may now be solved by inputting the values  a^   ,  a^
a^  ,  a-j  ,  a.*  ,  a,Q  ,  a-Q  ,  a^2  and  b3  .  The "a" values
were extracted from the settling and filtration  (Section V)  analy-
sis sections and are listed below.  The value of  a^  was measured
in the laboratory as 0.0005 following carbonation.

     Inputting the "a" and "b" concentrations into the "K" equations
yields the following numerical values.

     K! = 0.31

     K2 = 0.19

     K3 = 0.01

     K4 = 0.31

     The equation for component flow rates,  X^  , can now be solved
systematically using the calculated "K" values.

      XL " 200,366 liters per day (52,937 gal/day)

      X2 = 156,252 liters per day (41,282 gal/day)

      X3 = 44,144 liters per day (11,655 gal/day)

      X4 = 156,252 liters per day (41^282 gal/day)

      X5 = 48,437 liters per day (12,797 gal/day)

      X6 = 107,816 liters per day (28,485 gal/day)

      X7 = 9,201 liters per day (2,431 gal/day)

      Xg = 39,235 liters per day (10,366 gal/day)
                                   80

-------
      Xg = 147,051 liters per day (38,851 gal/day)

     X10 = 191,165 liters per day (50,506 gal/day)

     Xu = 1,915 liters per day (506 gal/day)

     Product water = 189,250 liters per day (50,000 gal/day)

     The quantities of waste solids generated in streams  Xy  and
X    are calculated below:

(Solids)? = 8.33a7X? = 8.33(0.10)(2,431) = 918 kg/day (2,025 Ib/day)

(Solids)u = 8.33a11X11 = 8.33(0.05) (506) = 96 kg/day (211 Ib/day)

  Total solids for disposal - (Solids)y + (Solids)-^ - 1,014 kg/day

                                                       (2,236 Ib/day)

     The individual tank capacities necessary for effective perfor-
mance were determined by calculating the product of the experimental
residence time and the hydraulic loading on the tank as determined by
the material balances.  Residence times have been discussed in the
previous analytical sections.  These results indicate the residence
times for the reactor and carbonator are 90 and 60 min, respectively.
Settling tests (Section V, Figure 8) revealed that only 30 min were
necessary for efficient separation.  The settling time for the dem-
onstration plant has been increased to 60 min to allow for turbulence
due to injection and discharge of materials.  The feed tank was sized
for a residence time of 90 min for flow equalization as well as to
minimize fluctuations in pollutant concentrations.  A final storage
and recarbonation tank was provided for product water storage and pH
adjustment, if necessary.  An arbitrary 15 min residence time was
selected for this tank.  Table 16 summarizes residence times, hy-
draulic loadings, and calculated capacities for each tank.
                                  81

-------
        Table  16.  TANK CAPACITIES FOR  190,000  LITER PER DAY
                   (50,000 GAL/DAY)  DEMONSTRATION  PLANT
    Tank

Feed
Reactor
Settler
Carbonator
Final storage
Re^_dence time
    (days)

     0.063
     0.063
     0.042
     0.042
     0.010
Hydraulic loading
 I/day (gal/day)

 200,101 (52,937)
 156,046 (41,282)
 156,046 (41,282)
 191,913 (50,506)
   Capacity
    l(gal)

12,606 (3,335)
 9,832 (2,601)
 6,517 (1,734)
 8,017 (2,121)
 189,000 (50,000)    1,890   (500)
     The ssnd filters were designed on  the basis of experimental re-
sults from Section V  and  solids  loading calculations.   The  solids
produced in the reactor and carbonator  are similar in their dewater-
ing, or filtering characteristics; therefore, the filtration study
conducted on the reactor effluent is applicable to that of the carbo-
nator.  The design criterion, which was developed from  the experi-
mental study, is the solids filtration  capacity, 31 kg/day/m2  (6.3
lb/day/ft2).  This value is independent of the solids concentration
in the feed to the filter, therefore, it is an adequate value  for
both primary and final filters.  When the solids loading rate,
(Solids^  , is divided by the solids filtration capacity, the filter
area is calculated.  The solids  loading rates on the primary and
final filters were previously calculated as 919 kg/day  (2,025  Ib/day)
and 96 kg/day (211 Ib/day), respectively.

     Primary filter:   Area = 919/31 = 30 m2 (323 ft2)

     Final filter:  Area = 96/31 = 3.1 m2 (33 ft2)

     The solids represent the "dry weight" of sludge material which
is about 50% by weight water of hydration not removed at 105°C.  The
solids collected on the sand  filter are about 107. dry and 907. free
water.
                                  82

-------
     Solids collected on the sand filter will be scraped off and dis-
posed of on land.  The solids have a density of approximately 1.1 g/
cm3 (70 lb/ft3), yielding about 9 m3 (320 ft3) of sludge per day for
disposal.

DEMONSTRATION PLANT EQUIPMENT SELECTION

     Preliminary corrosion tests were conducted during the operation
of the bench pilot scale AMD plant to assist in selection of materials
of construction for the 190,000 liter per day (50,000 gal/day) demon-
stration plant.  The corrosion tests showed that ferrous metals would
corrode in contact with the acid mine water and also in the highly
alkaline (pH 11.9 to 12.1) reaction mixture.  Therefore, plastic
materials of construction were chosen for all equipment in contact
with the acid and alkaline solutions.  Fiberglass was chosen for the
storage tank reactor, the settler and carbonator.  In addition to
corrosion control, product purity was also a determining factor.  Cor-
rosion products could have an adverse impact on product water quality.

     The bill of materials with costs is given in Table 17.  The costs
in this table were obtained by written quotations from manufacturers
and vendors.  The total equipment cost for the 190,000 liter per day
(50,000 gal/day) plant is $46,290 with freight estimated at $1,200.
A quotation was received for a chlorinator, which may or may not be
added later.  The chlorinator cost is $2,715.

SCHEDULING AND MANPOWER REQUIREMENTS FOR THE DEMONSTRATION PLANT

     The plant construction and installation costs are estimated to
be $48,600, which will include all subcontract work such as electri-
cal, concrete and rigging.  The sand filters will be constructed on-
site and the associated costs are included in those estimated for
construction and installation.

     The startup of the demonstration plant will require the services
of three people for a one-shift, 5-day-week operation; two operators
and a supervisor, who will be responsible for the analytical work and
data keeping as well as the personnel scheduling.

     The continuous operation of the demonstration plant on three
shifts, 7 days a week, will require eight operators plus a supervisor.
                                  83

-------
                       Table 17.  IUU. OF MATERIALS FOR DEMONSTRATION  PLANT
                       (50,000 GAL/DAY) ACID MINE DRAINAGE TO  DRINKING WATER
1 - Raw water 152 ipm 30 m  (40 gpm  100  ft)  head  -  plastic  lined
    Centrifugal pump with bypass  system
1 - Raw water storage tank  15,000 £  (3,950  gal)  -  fiberglass
1 - Stage I reactor 12,600  i  (3,300  gal) dished
    Bottom tank with baffles  - fiberglass
1 - Turbine blade mixer
1 - Centrifugal pump - 114  ipm 7.5 m (30 gpm  25  ft)  head -  plastic  lined
1 - Settler 6,500 t (1,690  gal)   dished bottom tank  -  fiberglass
1 - Centrifugal pump 76 ipm 7.5 m (20 gpm 25  ft) head  - plastic  lined
1 - Slurry pump - positive  displacement 57  &pm (15 gpm) - diaphragm
1 - Sand filter - 18 sq m (200 sq ft) x 0.3 m (1 ft) deep with sludge  distributor,
    repair and recondition  present trickle  filter
1 - Centrifugal pump - 57 Ipm 7.5m  (15 gpm 25 ft) head - plastic  lined
1 - Carhonator 15,850 i (4,170 n;»l)  dished bottom tank with  baffles  -  fiberglass
1 - Turbine blade mixer
1 - Cent rlfiiHul pump - 152  > pm 30 in  (40 Kpm 100  ft)  head open Impeller  - plastic  lined
1 - Saiul III tor - 7 sq in (78  s<| ft)  x 0.3 m (1 ft) deep with  distributor
1 - CimtrlliiKul pump - 152  ipm 7.5m (40 gpm  25  ft)  head -  plastic  lined
1 - lU'fluont storage and recarbonation  tank - 1,900  t  (500  R;I I )  -  f'Vf)  I «nk __
1 - KlowmrliT  0-122 ipm (0-32 spin)  water
1 - KlowmcttT  0-46 jjpm (0-12 gpm) water
1 - PVC-limo slurry tank complete with  mixer, slurry pump,  and flowmeter
1 - PVC-sodium aluminate feed tank with mixer, pump  and flowmeter
1 - Flowmeter  0-114 4pm (0-30 gpm)  water
1 - Flowmeter  0-57 ipm (0-15 gpm)
1 - Carbon dioxide delivery and mixing  system
2 - pH flow-through electrodes
1 - two-pen recorder
      Subtotal
                                        Piping and valves
150 m (500 ft) 2.54 cm (1 in.) PVC pipe
30 m (100 ft) 1.91 cm (3/4 in.) PVC pipe
50 - 2.54 cm (1 in.) L's PVC
10 - 1.91 era (3/4 in.) L'S PVC
20 - 2.54 cm (1 in.) T's PVC
5 - 1.91 cm (3/4 in.) T's PVC
25 - 2.54 cm (1 in.) ball valvus PVC
10 - 1.91 cm (3/4 in.) ball valves PVC
10 - 2.54 cm (1 in.) diaphragm valves PVC
5 - 1.91 cm (3/4 in.) diaphragm valves PVC

   Subtotal

   Total

Freight

1 - Chlorinator for effluent (possible add on at later date)
S  1,065

   2,256
   4,215

   2,852
   1,016
   2,384
   1,013
   3,142
   1.000

   1,013
   4,684
   2.852
   1.412
   3,500
   1.065
    975
$44,961
    195
     30
     16.50
      1.90
      8.20
      1.05
    536.20
    178.00
    255.00
    107.75
$ 1,319.66

$46,280.66

$ 1,200

$ 2,715
                                                 84

-------
                             APPENDIX B

             ALUMINA-LIME-SODA CHEMICAL REQUIREMENTS FOR
              RECOVERING WATER FROM ACID MINE DRAINAGE
     This appendix presents the derivation of formulae presented in
Section VII needed to estimate chemical requirements for sodium alu-
minate and lime used in Stage I of the alumina- lime -soda process,
and for carbon dioxide needed in Stage II.  The formula for carbon
dioxide dosage includes the procedure for proportioning Stage I ef-
fluent and raw acid mine drainage in a blend to produce water with a
specific sulfate concentration.

SODIUM A1.UM1NATE (STACK 1)

     The addition of sodium aluminate to acid mine drainage will re-
move sulfate.  Sodium aluminate will be available in two forms.  Dry
sodium aluminate (Dry-NaAl02) is now commercially available and con-
tains 43% A1203 and 30% Na20.  Calcined sodium aluminate (Cal-NaAl02)
is that produced by firing bauxite and soda ash and contains 53%
A1203 and 397. Na20.  In addition, iron and alumina in the raw AMD will
also remove sulfate.  Experimental data indicate that sulfate concen-
trations can be reduced to 100 mg/ liter.

General Equation and Definition of Terms

                  RS04 - ES04 = A(NaA102) + B(Fe) + C(Al)         (B-l)

where     RSO, = sulfate concentration of raw water, mg/i

          ESQ. = sulfate concentration of Stage I effluent, a 100
    Dry-NaA102 = dry sodium aluminate required,

    Cal-NaAl02 = calcined sodium aluminate required,

            Fe = total in situ iron concentration of raw water,

            Al = jLn situ aluminum concentration of raw water, mg/4

                                   85

-------
     A, B, and C = coefficients needed to convert concentration into

                   sulfate equivalents


Sulfate Removed by Sodium Alumlnate


                   A(NaAl02) = A x (2CaS04'Al203'3CaO)             (B-2)


1 g Dry-NaAl02 = 430 mg A1203


1 g Cal-NaA102 = 530 mg A120


1 mmole A1203 = 102 mg A120


1 mmole Al 0., = 2 mtnoles SO,


1 mmole SO  = 96 mg/SO,


A = coefficient


Calculation of A;
A(Dry-NaAl07) =     x 2.0 x 96 = 809
           ^    102
A(Cal-NaAl09) =    . x 2.0 x 96 = 998
           2    102


SuLfate Removed by iji situ Iron



                     B(Fc) = B x (CaSO, -Fe203-3CaO)                (B-3)


56 mg PC = 1 mmole Fe


1 mmole Fe = 0.5 mmole SO,


1 mmole SO, = 96 mg SO^


B = coefficient


Calculation ofB:
        56


Sulfate Removal by in situ Aluminum
                     C(A1) = C x (CaS04-Al203-3CaO)                (B-4)


                                   86

-------
27 mg Al = 1 mmole Al

1 nmiole AJ =0.5 mmole SO/

1 mmole SO^  = 96 mg SO,

C= coefficient

Calculation of C:

C=°-5x96 = 1.78
        27

Equation  Solution and Substitution

                g          =      - ES04 - B(Fe) - C(A1)
            g Dry-NaAl02/£ -      "      "  '       "  '          (B-6)
                                            809
            g Cal-NaAi02/£ =  S04 -  S04 - 0^86(Fe) - 1.78(A1)
LIME  (STAGE  I)

     Addition of  lime to raw AMD will  (a)  neutralize  acid,  (b)  pre-
cipitate metals,  (c) stabilize  sulfoaluminate  and  sulfoferrite  sludge,
and  (d) maintain  process pH at  12.0.   In addition,  the  caustic  soda
(Na20) sodium aluminate provides hydroxide which can  substitute for
part of the  lime.  Lime to be used  is  hydrated lime having  93%  purity.

General Equation  and Definition of  Terms
                     h
               CaO = ECaOn = A (acid) + JJ(Mg) + C(Mn)  +  D(Fe) +
                     n=a
                                E(A1) + (F-H)(NaAl02)  +  G          (B-8)

where     CaO = hydrated lime,  93%  Ca(OH)2> requirement,  g/4

         CaOn = hydrated lime,  93%  CaO, required for  specific chemical
                reaction, g/1
                                   87

-------
         acid = acidity of raw water, mg/4  as  CaC03

           Mg = magnesium concentration  of  raw water,

           Mn = manganese concentration  of  raw water,  m

           I'e — total iron concentration of raw  water,  mg/j&

           Al = alumina concentration of raw water,  mg/4

   Dry-NaAl02 = dry sodium aluminate requirement,  g/£

   Cal-NaAl02 = calcined sodium aluminate requirement,  g/A

A, B, C, D, E,
     F, and H = coefficients

            G = constant to maintain causticity  at 600 mg/4  (as
                to maintain process pH at 12.0.

NeuLralizaLion

                              Ca()a = A(acid)                       (13-9)

1 g lime = 930 nig Ca(01l)2

74 mg Ca(OH)- = 1 mmole CaO

100 mg acid = 1.0 mmole acid

1.0 mmole acid = 1.0 mmole CaO

Calculation of A:
Precipitation of Magnesium and Manganese

                              CaOb = B(Mg)                         (B

                              CaOc = C(Mn)                         (B-ll)

24 mg Mg = 1 mmole Mg

1 mmole Mg = 1 mmole CaO


                                   88

-------
55 mg Mn = 1 iranole Mn



1 mmole Mn = 1 mmole CaO


1 g CaO = 930 mg Ca(OH)



1 mmole CaO = 74 mg Ca(OH)2


B, C = coefficients


Calculation of B:


B = — x — = 0.00332
    930    55



Calculation of C:


     1    74
C = -±- x — = 0.00145
    930   55


Stabilization of Sludge Formed From in situ Iron and Alumina



                  CaOd = D(Fe) = D x (CaS04.Fe203-3CaO)           (B-12)


                  CaOe = E(A1) = E x (CaS04-Al203*3CaO)           (B-13)


56 mg Fe - 1 mmole Fe


27 mg Al = 1 mmole Al



1 mmole Fe = 1.5 mmole CaO


1 mmole Al = 1.5 mmole CaO


1 mmole CaO - 74 mg Ca(OH)2


1 g lime = 930 mg Ca(OH)2



D, E = coefficients


Calculation of D:
                                   89

-------
Calculation of E:


K = -J- x 2^. x 1.5 = 0.00442
    930   27


Stabilization of Sludge Formed From Sodium Aluminate


                CaOf = F(NaAl02) = F x (2CaS04-Al203'3CaO)        (B-14)


1 g Ury-NaAlO  = 430 mg A1203


1 g Cal-NaA102 = 530 mg A1203


102 mg A1203 = 1 mmole A1203


1 mmole A1203 =3.0 mmole CaO


1 mmole CaO - 74 mg Ca(OH)2


1 g lime = 930 mg Ca(OH)2


F = coefficient


Calculation of F:


F (Dry-NaAl02) = -i- x -^ x 3.0 x 430 = 1.006
F (Cal-NaAl02) = — x -~ x 3.0 x 530 = 1.240



Excess Lime for pH Control
                                 CaO   =  G                           (B-15)
                                   g
600 mg/4 as CaC03 causticity = pH 12


600 mg = 6.0 mmole Ca(OH)2


1 g lime = 930 mg Ca(OH)2


74 mg Ca(OH)  = 1 mmole Ca(OH>2


G = constant
                                   90

-------
Calculation of G:


G-.6-°* 74 = 0.477
       930


Mint! Equivalent From Sodium A laminate


                           CaO,  = - H(NaAlO.,)                      (B-16)
                              h            *•                            '

1 g Dry-NaAl02 = 300 rag Na20


1 g Cal-NaAlO  = 1 mmole Na20


1 mmole Na^O = 1 mmole CaO


1 g lime = 930 mg CatOH>2


1 mmole CaO = 74 mg Ca(OH)


H = coefficient


Calculation of H:


H (Dry-NaAl02) = -~ x ~ x 300 = 0.385



H (Cal-NaAl02) = -~ x || x 390 = 0.501


Summation  and  Comb ina t ion

                     h
           g CaO/4 = ZCaOn = A(acid)  + B(Mg)  + C(Mn)  + D(Fe)  +
                     n=a

                               E(A1)  + G + (F-H)NaA102           (B-17)


       g CaOAe(Dry-NaAl02) = [0.00079 acid + 0.00332  Mg + 0.00145 Mn +


                               0.00213 Fe + 0.00442 Al + 0.477] +


                               0.621  g Dry-NaAl02                (B-18)


       g CaOAe (Cal-NaAlO  ) = [0.00079 acid + 0.00332  Mg + 0.00145 Mn +


                               0.00213 Fe + 0.00442 Al + 0.477] +


                               0.739  g Cal-NaAl02                (B-19)
                                  91

-------
CARBON DIOXIDE (STAGE II)

     Effluent from Stage 1 of the alumina-lime-soda process contains
600 rag/liter (as CaC03> of causticity and must be neutralized.  In
addition, the effluent will have a sulfate level of about 100 ing/
liter.  This effluent can be blended with raw AMD to bring the sul-
fate level in product water to 250 ing/liter, the maximum permissible
level for drinking water.  This blending procedure reduces the car-
bon dioxide requirements for complete neutralization.

     Carbon dioxide is added to drop pH to 10.3 and also precipitates
the maximum amount of CaCO.,; the carbonate concentration at pH 10.3
will be about 35 rag/liter as CaCOj.  After separating the CaCO-j from
the product water, it is carbonated further to reduce pH to the 7.0
to 8.0 range.  The 35 mg/liter carbonate will be converted to 35 mg/
liter bicarbonate.  The conversion of carbonate to bicarbonate will
require additional carbon dioxide.

Fraction of  Stage  I  Effluent  in  Blend with  Raw AMD

                            Rso4 - Pso4
                            Rso4 - Eso4
                                        = EFF                      (B-20)
where      EFF = fraction of Stage I effluent in the blend

          RS04 = sulfate concentration of raw AMD,
          r>
           S04 = sulfate concentration of Stage I effluent, z. 100

          PS04 - sulfate concentration of product water, ^ 250

thus,                          1 - EFF = AMD                       (B-21)

where     AMD = fraction of raw AMD in the blend
                                    92

-------
Causticity of Blend

                     (EFF x 600) -  (AMD x acid) - CAUS              (B-22)

where     CAUS = causticity of blend, in mg/ji as CaCO_

Carbon Dioxide

                         C02 = A x  (CAUS + 35)                     (B-23)

where     CO- = carbon dioxide requirement, g/£

 1 mmole
   causticity = 100 mg as CaC03

 1 iranole
   causticity = 1 mmole C02

  35 mg
    carbonate = 35 mg bicarbonate = 35 mg as CaCO

    44 mg C02 = 1 mmole C02

     1 mg C02 • 0.001 g C02

 carbonation
   efficiency =0.95

            A = coefficient

Calculation of A:

A = — x —— x 0.001 = 0.00046
    100   0.95

Substitution Into C02 Requirement Equation

                     g C07/je » 0.00046 x (CAUS + 35)               (B-24)
                                  93

-------
                                REFERENCES
1.  J. W. Nebgen, E. P. Shea, and S. Y. Chiu, "The Alumina-Lime-Soda
      Water Treatment Process," Office of Saline Water, Research and
      Development Progress Report No. 820, Contract No. 14-20-2935,
      PB 218 326, January 1973.

2.  E. Nordell, Water Treatment for Industrial and Other Uses.  Reinhold
      Publishing Corporation, New York, pp. 489-526 (1961).

3.  W. Lerch, F. W.  Ashton, and R. H. Bogue,  "The Sulphoaluminates of
      Calcium," J. Research Nat. Bur. Standards. .2:715-731 (1929).

4.  F. E. Jones, "The Formation of the Sulfoaluminates and Sulfofer-
      rites of Calcium in the Portland Cement-Water System," J. Phys.
      Chem., 49,:344-357 (1945).

5.  J. W. Ryzner, and A,  C. Thompson, "Sodium Aluminate," in Kirk-Othmer
      Encyclopedia of Chemical Technology. Second Edition, Vol. 2, pp.
      6-11 (1963).
                                   94

-------
                                   TECHNICAL REPORT DATA
                            (Please read Instructions on the reverse before completing)
 1. REPORT NO.
   EPA-600/2-76-206
                              2.
             3. RECIPIENT'S ACCESSION-NO.
 4. TITLE AND SUBTITLE
  Treatment of Acid Mine Drainage  by the Alumina-Lime-
  Soda Process
                                                           5. REPORT DATE
              September_1976  (Issuing Date)
             G. PERFORMING ORGANIZATION CODE
 7. AUTHOR(S)
  J.  W.  Nebgen, D. F. Weatherman,  M.  Valentine, and
  E.  P.  Shea
             8. PERFORMING ORGANIZATION REPORT NO
 9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Midwest Research Institute
  Kansas City, Missouri 6^110
             10. PROGRAM ELEMENT NO.

                 1BB610
             11. CONTRACT/GRANT NO.
                                                              S-802816
 12. SPONSORING AGENCY NAME AND ADDRESS
  Industrial  Environmental Research  Laboratory
  Office  of Research and Development
  U.  S. Environmental Protection Agency
 _. Cincinnati. Ohio  U5268	
              13. TYPE OF REPORT AND PERIOD COVERED
                  Fina.1  fi/7l - 12/7S
             14. SPONSORING AGENCY CODE
               EPA-ORD
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
         The alumina-lime-soda process  is a chemical desalination process for waters
  in which  the principal sources  of  salinity are sulfate salts a.nd  has  "been field
  tested at the Commonwealth of Pennsylvania's Acid Mine Drainage Research Facility,
  Hollywood, Pennsylvania, as a method to recover potable water from  acid mine drain-
  age.   The alumina-lime-soda process  involves two treatment  stages.  Raw water is
  reacted with sodium aluminate a,nd  lime in the first stage to precipitate dissolved
  sulfate as calcium sulfoaluminate.   In the second stage, the alkaline water
  (pH =  12.0) recovered from the  first stage is carbonated to precipitate excess hard-
  ness.   Following carbonation, product water meets USPHS specifications for drinking
  water.
       Alumina-lime-soda process  economics are influenced most by the cost of sodium
  aluminate.  Widespread application of the alumina-lime-soda process will increase
  demand for sodium aluminate, and should spur interest in alternate  sources of this
  treatment chemical.
       Operating costs for recovering  potable water from an acid mine drainage having
  an acidity of 700 mg/liter and  a_sulfate level of 750 mg/liter are  estimated to be
  in the range of $0.21 to $0.27/m   ($0.79 to $1.0^ per 1,000 gal).
                                KEY WORDS AND DOCUMENT ANALYSIS
                  DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS  C. COSATI Field/Group
  *Calcium hydroxides
  Neutralizing
  *Drainage - mine (excavations)
  Cost  comparison
  pH control
  Aluminum oxide
  Demineralizing	^^
 Acid Mine Drainage
 Coal Mine Drainage
 Ferric iron
 Pennsylvania
 Ferrous iron
 Softening - mine  drainage;
 13B
08/G
08/H
 8. DISTRIBUTION STATEMENT
  Release to  public
19, SECURITY CLASS (ThisReport)
 Unclassified
                                                                        21. NO. OF PAGES
105
                                              20. SECURITY CLASS (TMspage)
                                               Unclassified
                                                                        22. PRICE
EPA Form 2220-1 (9-73)
                                            95
                                                               ftUSGPO: 1976 — 657-695/6111 Region 5-11

-------