EPA-670/2-74-051
June 1974
Environmental Protection Technology Series
                                    LIMESTONE  AND
               LIMESTONE-LIME  NEUTRALIZATION
                          OF  ACID  MINE  DRAINAGE
                                 National Environmental Research Center
                                   Office of Research and Development
                                  U.S. Environmental Protection Agency
                                           Cincinnati, Ohio 45268

-------
                                                             EPA 670/2-7^-051
                                                             June  1971*
                    LIMESTONE AND LIMESTONE-LIME NEUTRALIZATION

                                OP ACID MINE DRAINAGE
                                        By

                                 Roger C.  Wilmoth
                         Mining Pollution Control Branch
                  Industrial Waste Treatment Research Laboratory
                            Program Element  No.  1BBC40
                    NATIONAL ENVIRONMENTAL RESEARCH CENTER
                      OFFICE OF RESEARCH AND DEVELOPMENT
                     U.S. ENVIRONMENTAL PROTECTION AGENCY
                            CINCINNATI, OHIO  1*5268
For sale by the Superintendent of Documents, U.S. Government
      Printing Office, Washington, D.C. 20402

-------
            EPA REVIEW NOTICE

    This report has teen reviewed "by the
National Environmental Research Center,
Cincinnati, and approved for publication.
Mention of trade names or commercial prod-
ucts does not constitute endorsement or
recommendation for use.
                    11

-------
                                FOREWORD
     Man and his environment must be protected from the adverse effects
of pesticides, radiation, noise and other forms of pollution, and the
unwise management of solid waste.  Efforts to protect the environment
require a focus that recognizes the interplay "between the components of
our physical environment—air, water, and land.  The National Environ-
mental Research Centers provide this multidisciplinary focus through
programs engaged in

           •  studies on the effects of environmental
              contaminants on man and the biosphere, and

           •  a search for ways to prevent contamina-
              tion and to recycle valuable resources.

     This report defines the important parameters for the use of lime-
stone, hydrated lime, and combination limestone/lime treatment of acid
mine drainage discharges containing iron in predominately the ferric
state.
                                        A.  W.  Breidenbach, Ph.D.
                                        Director
                                        National Environmental
                                        Research Center,  Cincinnati
                                   iii

-------
                               ABSTRACT

    The critical parameters affecting neutralization of ferric-iron acid
mine waters were characterized by the U.S. Environmental Protection
Agency in comparative studies using hydrated lime, rock-dust limestone,
and a combination of the two as neutralizing agents.  The advantages
and disadvantages of each of these neutralizing agents were noted.   On
the ferric-iron test water, combination limestone-lime treatment pro-
vided a better than 25-percent reduction in materials cost as compared
to straight lime or limestone treatment.  Significant reduction in sludge
production was noted by the use of rock-dust limestone and by the use of
combination treatment as compared to hydrated-lime treatment.  Emphasis
on optimizing limestone utilization efficiencies resulted in an increase
from approximately 35-percent to 50-percent utilization.  Studies using
limestone that had been ground to pass a UOO-mesh screen resulted in
utilization efficiencies near 90 percent.
                                    iv

-------
                                 CONTENTS
                                                                    Page
Fore-word	    iii
Abstract	    iv
List of Figures	    vi
List of Tables	,	    vii
Acknowledgements	,	    ix
Section
I     Conclusions	    1
II    Recommendations	    3
III   Introduction	    h
IV    Procedures	    8
V     Results	    15
VI    Discussion	    6U
VII   References	    73
VIII  Glossary	*	    76
IX    Appendix	    78

-------
                                LIST OF FIGURES

  No.                                                                  Page
  1    Schematic flow diagram of the dual-neutralization system	   9
  2    Titration curves for limestone and lime	  16
  3    Effect of detention time on limestone  utilization efficiency...  20
  k    Effect of temperature on limestone utilization  efficiency	  22
  5    Effect of temperature on UoO-mesh limestone  utilization
         efficiency	  23
  6    Effect of temperature on limestone utilization  efficiency
         @ pH 6.5 and 5 gpm (18.9 1/m)	  2k
  1    Effect of temperature on limestone utilization  efficiency
         @ pH 5-0 and 15 gpm (56.8 1/m)	  25
  8    Wire-mesh reactor "baffle sketch	  29
  9    Cost of various combinations in two-stage limestone/lime
         treatment schemes	  U7
 10    Supernatant turbidity and sludge buildup during settling.   (All
         tests were conducted at pH 6.5.)	  58
^
 11    Effect of lime/limestone cost ratio on the economics of
         combination treatment	  6l
 12    Lime/limestone cost ratio and process  cost reduction at  pH  9
         (simulated)	  62
 13    Sulfate-conductivity relationship at Norton, W.  Va	  70
 lU    Sulfate-conductivity relationship at Mocanaq.ua,  Pa	  Jl
 15    Comparison between EPA and Lovell's sulfate-conductivity
         correlation	  72
                                     VI

-------
                             LIST OF TABLES

No.
 1   Typical Detention Times	  11
 2   Grassy Run Water Quality Data, 1971	  12
 3   Manufacturer1 s Chemical Analyses of Reagents	  13
 U   Spectrochemical Analysis of Germany Valley Limestone	  13
 5   Sieve Analyses (Dry) of As-Received Lime and Limestone	  lU
 6   Research Plan for Limestone Investigations	  15
 7   Effect of pH on Effluent Iron, Effluent Turbidity and
       Limestone Requirement	  17
 8   Effect of Detention Time	  19
 9   Limestone Neutralization of Acid Mine Drainage Logarithmic
       Probability Plots for Mean Reactor Detention Time	  27
10   Effect of Baffles at 15 GPM (56.8 Liters/Min)	  28
11   Wet-Wash Sieve Tests	'	  30
12   Batch Scale Neutralization Test	  30
13   Comparison of liOO-Mesh Limestone vs Regular Limestone
       6 15 GPM (56.8 Liters/Min)	  32
lU   Comparison of ^00-Mesh Limestone vs Regular Limestone
       § 5 GPM (18.9 Liters/Min)	  33
15   Comparison of Finely-Ground Germany Valley Limestone and
       Fine-Mesh Limestone from York, Pennsylvania	  3^
16   Limestone Slurry vs Dry Feed Neutralization Study Batch
       Scale Tests	  36
17   Limestone Split Treatment Feasibility Study	  38
18   Simulated Sludge-Recycling Tests	  37
19   Chemistry Analyses for-Simulated Sludge Recycle Batch Tests...  UO
20   Comparison of Sludge Recirculation vs No Recirculation
       g 10 GPM (37-9 Liters/Min)		  Ul
21   Summary of Limestone Utilization Efficiencies from Continuous
       Flow Tests	  k2
22   Limestone vs Lime § 10 GPM (37-9 Liters/Min)	  M
23   Limestone vs Lime @ 5 GPM (18.9 Liters/Min)	  U5
                                  vii

-------
No.                                                                page
2k   High pH Limestone Study Parameters	 U6
25   Limestone-Lime Batch Scale Study	 kQ
26   Limestone-Lime vs Limestone @ 5 GPM (18.9 Liters/Min)	 51
21   Limestone-Lime vs Limestone @ 10 GPM (37 = 9 Liters/Min)	 52
28   Chemistry Analyses:  Pilot Plant Studies	 51*
29   Limestone-Lime vs Lime § 5 GPM (18.9 Liters/Min)	 55
30   Limestone & Lime vs Lime g 10 GPM (37=9 Liters/Min)	 56
31   Material Cost Advantage of Combination Treatment	 53
32   Comparison of Sludge Percent Solids	 57
33   Comparison of Reagent Costs for Treating Mine Drainage	 67
3k   Estimates of Total Treatment Costs	 68
35   Chemical Analyses for Variable pH Study	 79
36   Effect of Detention Time - 15 GPM (56.8 1/m) vs
       10 GPM (37-9 1/m)	 80
37   Effect of.Detention Time - 15 GPM (56.8 1/m) vs
       5 GPM (18.9 1/m)	 81
38   Effect of Detention Time - Two Reactors in Series vs
       One 'Reactor	 82
39   Effect of Detention Time - Two Reactors in Series @ 5 GPM
       vs One Reactor g 15 GPM	 83
UO   Chemical Analyses for Detention Time Study	 8U
Ul   Effect of Reactor Baffles % 15 GPM (56.8 Liters/Min)	 85
k2   Comparison of Wood and Screen Baffles at 15 GPM
       (56.8 Liters/Min)	 86
U3   Effect on Screen-Baffle Mesh Size at 15 GPM
       (56.8 Liters/Min)	 87
kh   Comparison of Three vs Two Wire-Mesh Baffles at 15 GPM
       (56.8 Liters/Min)	 88
U5   Chemical Analyses for Baffle Study	 89
k6   Chemical Analyses for UOO-Mesh Limestone Study	 90
1*7   Chemical Analyses for Lime vs Limestone Study	 91
                                 Vlll

-------
                            ACKNOWLEDGEMENTS

    Thanks are extended to Ronald D. Hill, Robert B.  Scott,  James~L.
Kennedy, Alvin ¥. Irons, Roger A. Dean, Maxine Cooper,  Loretta Davis,
and Curtis Corley for their helpful suggestions and willing  assistance.
A special thanks is extended to J. Randolph Lipscomb for his invaluable
contributions to the success of this project.
                                    IX

-------
                                SECTION I
                               CONCLUSIONS

    Reaction time, effluent pH, mixing chamber design, reagent particle
size, and temperature were identified as critical parameters affecting
the limestone treatment process for neutralizing acid mine drainage (AMD),
Reaction times of 20 to 30 minutes appeared optimum.  Decreasing the
effluent pH increased limestone utilization efficiency.  Use of 1^-mesh
wire screen reactor baffles abraded reaction coatings from limestone
particles and increased both mixing efficiency and limestone utilization
efficiency.  Although UOO-mesh limestone was clearly superior in reac-
tivity to rock-dust limestone (approximately 100 mesh), the commercial
scarcity of UOO-mesh stone would make rock dust the logical choice.  The
economics of onsite grinding should be further investigated, however.
Use of a shallow holding pond may increase water temperature and thus
enhance limestone efficiency.  Use of such a pond would be discontinued
during cold months.
  v  Maximum utilization efficiency observed during continuous-flow,
pH 5-0 studies using rock dust was in the area of 60 percent (or 1.7
times the stoichiometric requirement).  Continuous-flow studies at
pH 6.5 resulted in efficiencies of about 50 percent (or twice the
stoichiometric requirement).
    Studies at pH 6.5 using UOO-mesh limestone yielded utilization
efficiencies above 90 percent (or 1.1 times the stoichiometric require-
ment ).
    Sludge recycling studies provided no quantitative conclusions since
continuous recycling facilities were not available and the attempted
arrangement was not entirely satisfactory.  However, limestone utiliza-
tion efficiency was improved sufficiently to warrant its further investi-
gation as a highly promising technique.

-------
    Hydrated-lime treatment appeared to be more economical than the
optimum case of limestone treatment, but the advantages of the more dense
limestone sludge may offset the initial cost advantage of lime treatment.
    Combination rock-dust limestone — hydrated-lime treatment, which must
be accomplished in two separate stages, worked most satisfactorily with a
first-step (limestone) pH near U.  Subsequent lime treatment of the par-
tially neutralized water could be taken to any desired pH, thus making
the process amenable to ferric and ferrous iron situations.  Reaction
times of 20 to 30 minutes were required for efficient utilization of
limestone in the first step of treatment.  For second-stage lime treatment,
10 to 15 minutes of reaction time were sufficient.
    Combination limestone-lime treatment of ferric acid mine drainage pro-
duced half the sludge volume of lime treatment alone although slightly
more than limestone treatment alone.  The solids content of the sludge
from combination treatment was five times as great as that produced by
lime treatment, though significantly less than that produced by limestone
treatment.
    The combination process has significant potential for reducing the
cost of AMD treatment.  On the ferric-iron test water, a better than 25-
percent reduction in material cost was achieved by using combination treat-
ment instead of straight lime or straight limestone alone.  Water quality
of the effluents from all three processes was comparable.
    A lime/limestone raw-material cost ratio of 1.8:1 was the break-even
point on the test water where no economic advantage would be achieved by
using limestone-lime rather than lime alone.  As the ratio increased, so
did the cost advantage of combination treatment.
    Increasing the final treatment pH slightly decreased combination
treatment's economic advantage but combination treatment should still be
economically beneficial in ferrous iron situations.

-------
                               SECTION II
                             RECOMMENDATIONS

    An economic study should be undertaken to determine the feasibility
of grinding limestone at the treatment plant to ^00-mesh size.
    Limestone sludge recycling should be fully investigated on equipment
designed specifically for recycling applications.
    The limestone-lime process should be demonstrated on ferrous-type
waters.
    Work should be done on sludge disposal techniques and related costs.

-------
                               SECTION III
                              INTRODUCTION

    With the advent and enforcement of stronger pollution laws, the coal
mining industry has "begun an extensive program of acid water treatment
plant design and construction.  To date, Pennsylvania, the state which
ranked third^ ' in coal production for 1970, has approximately 200 deep-
                                   (2)
mine treatment plants in operation.     Less than 50 of these are highly
sophisticated, large-scale plants.  The number one and two coal producers,
West Virginia^ ' and Kentucky    have 35 and 5 acid mine drainage (AMD)
treatment plants in operation respectively.  Virtually all of these plants
use lime as the neutralizing agent.
    Lime neutralization has several significant disadvantages:  high raw-
material cost, hazardous nature of the material itself, high potential for
accidental overtreatment, and production of a high-volume, low-density
sludge.  The state of the art of straight lime neutralization is fairly
well known at this time.
    The U.S. Environmental Protection Agency (EPA) and other researchers
have investigated other neutralizing agents such as limestone, soda ash,
sodium hydroxide, sodium sulfide, and combinations of these in attempts
to improve the AMD treatment process.
    In 1970, Wilmoth and Hill    investigated soda ash and found it too
expensive for general treatment-plant applications.  Care and Zawadzki
                             (7)
(1965) and Zawadzki and Glenn    (1968) studied sodium sulfide and hydro-
gen sulfide — limestone treatment of AMD; they indicated sodium sulfide
costs would generally prohibit its use.  Due to high raw-material costs,
sodium hydroxide would not only be too expensive but is also hazardous
to handle.
    Use of limestone has been touched on by a variety of investigators.
Earlier work by Braley et al. (1951)   ; Clifford and Snarley (195*0  J;

-------
Glover  (l967r10^; Zurbach (1963)    ; Wheat land and Borne (19.62)
Chamberlain  (l968)^13^; Hoak et al. (19^5)     ; Deul and Mihok
Calhoun (l968); Wilmoth and Hill (I970rl8); and Holland et al.
laid the groundwork for present investigation.  These researchers pointed
out that limestone has certain inherent advantages over lime for AMD treat-
ment; e.g., lower raw-material cost, less hazardous material to handle,
marginal problems resulting from accidental overt reatment , and dense, low-
volume sludge production.  The major disadvantages cited were the inability
of limestone to produce a pH above 7 (necessary for rapid, efficient iron
oxidation) and the low utilization efficiencies that resulted in excess
limestone usage.
    Ford, Boyer, and Glenn ^  ' of Bituminous Coal Research, Inc. (BCR)
approached the limestone — ferrous iron application problem by recycling
sludge in a 0.5-gpm (1.9 1/min), continuous- flow, lab-scale system.
Using a 100-percent excess of powdered limestone (50 percent utilization
efficiency) and a 1:1 ratio of sludge-slurry-recirculation flow to AMD
feed flow, they were able to reduce the effluent iron concentrations from
80 mg/1 to below 7 mg/1.  BCR was confident that their recirculation pro-
cess could effectively treat AMD waters with ferrous-iron concentrations
below 500 mg/1.  A major disadvantage to their process was that a 60-
minute reaction time was required to achieve the desired results.  This
reaction time would be expensive in terms of tank size and mixing costs
for large flow situations.  Another important contribution of Ford, Boyer,
and Glenn was characterizing the type of limestone most effective in AMD
treatment.  They designed a simple test for field use in evaluating lime-
stones for the user's particular application.      In general, the most
desirable stones have small particle size, are virtually pure calcium
carbonate, and/or are mostly calcite and have a large surface area.
                                                     \
Dolomitic limestones were found to be significantly ineffective in AMD
treatment .
    Dr. Lovell of Pennsylvania State University used a unique approach
to the ferrous iron situation in work at the Penn State - EPA Hollywood
Facility.  Using a rotary attrition-reaction mill and coarse rather than

-------
powdered limestone, Lovell claimed' limestone utilization effi-
ciencies in excess of 80 percent and observed across-the-board reductions
in ferrous content as high as 80 percent, using AMD sources with initial
ferrous concentrations near 100 mg/1.
    An 80-percent reduction in the 100 mg/1 ferrous concentration still
did not meet the State water quality standards of Pennsylvania,, however,
since the effluent iron was in excess of 7 mg/1.  Detention time in the
rotary reactor was approximately 2 minutes.  Lovell later used biological
oxidation as a pretreatment to limestone neutralization in this flow scheme
(at 18 C), and an initial ferrous concentration of 37^ mg/1 was reduced
to 26 mg/1 by a trickling filter-type, surface biological reactor with a
detention time of 5 minutes.  Subsequent limestone neutralization by the
rotary mill reduced the ferrous content to 20 mg/1 and increased the pH
to 6.2-6.6.  Additional oxidation occurred in the settling lagoon, and the
discharged effluent contained only 6 mg/1 of ferrous iron.  The rotary-mill
approach first tried on AMD by Zurbach    , and later by Mihok and
           (13)
Chamberlain     , was improved by Lovell, who demonstrated its promise in
AMD treatment schemes for ferric situations.  For most ferrous situations,
some method of preoxidation must be employed to improve economics.  The
biological oxidation scheme was demonstrated by Lovell to have significant
promise.  As in any biological system, upsets due to temperature or water
quality variations were inevitable but were felt to be solvable problems.
    The EPA approach to date has been to attempt to improve the utiliza-
tion efficiency of limestone in treatment of waters where the iron is al-
ready in the ferric state since it was felt that straight limestone treat-
ment of ferrous waters would have little chance of economically competing
with lime.  Also, EPA pursued the suggestion of Holland, et al.    , and
others and developed a two-stage, combination limestone-lime treatment
process that offers significant advantages over both lime treatment and
limestone treatment alone and that should have direct application to
ferrous iron situations.  Results of the EPA investigations (conducted at
EPA's Mine Drainage Treatment Field Site at Norton, West Virginia) are
presented in this report.

-------
    Yocum and Grandt of Peabody Coal Company are planning to demonstrate
the latest limestone, lime, and combination limestone-lime treatment
techniques on a large scale (2000 gpm or 7600 1/min) under EPA sponsor-
ship   (Grant 1^010 DAX).  As of this report, operational difficulties
have delayed initiation of research on their ferric-type water.
    EPA is continuing work on the combination process using a highly-
polluted ferrous AMD discharge at the Crown Mine Drainage Control Field
Site near Morgantown, West Virginia.

-------
                               SECTION IV
                               PROCEDURES

DEFINITIONS
    Definitions and an explanation of the calculations used in this report
are presented under Section VIII, Glossary.
TEST EQUIPMENT
    A schematic drawing of the dual neutralization system used is presented
in Figure '1.  Water was pumped from Grassy Run, a stream heavily polluted
with AMD, through a pressurized sand filter.   The filter removed large
particulate matter that in past experiences have tended to clog the flow
control valves and thus vary the water flow.   In a full-scale operation,
this filter would not be required.
    From the filter, the water went to a 500-gallon (1890-liter) supply
tank.  By maintaining a continuous overflow,  liquid level in the tank was
held constant, and a constant discharge rate  could be maintained.  Globe
valves were used to adjust flows into two separate 150-gallon (568-liter),
stainless-steel reaction tanks equipped with flash mixers and two lU-
mesh, stainless-steel screen baffles.  The baffles, in addition to decreas-
ing reactor short-circuiting, served to increase limestone reaction effi-
ciency by abrading the particles of limestone as they passed through the
screen.  It was assumed that the screen did two things:  first, limestone
particle size was reduced somewhat by collision with the screen; and
second, reaction products such as ferric hydroxide and calcium sulfate
were abraded from the particle surface thus exposing fresh reactive areas.
    By having a dual system such as this, it  was possible to use one side
as a control and the other side as the variable and thus expedite the re-
search efforts.  One side of the system was referred to as Process A and
the other as Process B.  The dual system allowed immediate evaluation of
the isolated variable under study.

                                    8

-------
AMD STREAM
                  0
  500
GALLON
J90 LITER)
                  CONSTAN
                     HEAD
                    SUPPLY
                     TANK
                                                  PROCESS  A
                                              SIDE ONE
                                FLASH MIXER
             GLOBE
             VALVE
                                                             150
                                                             ALLON
                                                             LITER)
                            fphfjCONTROLLER
LIMESTO
                              LIME OR
                             LIMESTONE
             SIDE TWO
            PROCESS B
                                             TREATED
                                             WATER
                                                                                     TREATED
                                                                                     WATER
      Figure 1. Schematic flow  diagram of the dual  neutralization system.

-------
    BIF helix-type dry feeders were used to add hydrated lime and rock-
dust limestone to the reactors.  These feeders were controlled automatically
by pH controllers whose probes were located in the settling pools.
    The neutralized waters from Process A and Process B were pumped into
identical pools for sludge separation.  These pools held approximately
13,000 gallons (1*9,000 liters) each and were U feet (1.36 meters) deep.
Two baffles were installed in each pool to decrease short-circuiting
and improve settling.  Canvas covers, sidewall insulation, and infrared
heat lamps kept the pools operational during winter.
    Continuous sludge removal from the pools was not possible because of
pool design, so periodic pumping was required.  This task was performed
at the end of each test run (generally 7 days in length).
    Maximum flow capacity of each side of the system was 15 gpm (21,600 gpd)
or 56.8 1/min (82,000 I/day).  Flow rates were varied from 5 to 15 gpm
(18.9 to 56.8 1/min) -to study the effects of reaction time on system
efficiency.
PHYSICAL MEASUREMENTS
    Water flow rates were determined volumetrically in a 15-gallon (56.8-
liter ) container.  Reagent usage was calculated by collecting and weighing
five one-minute samples from the dry feeder, determining the average weight
per minute, and multiplying by the number of minutes of operation.  Dual
elapsed time meters reading in minutes and tenths and equipped with dynamic
braking circuitry recorded the operating time for each dry feeder.  Though
pH was automatically controlled, frequent probe cleaning and operator at-
tention was required for precise process control.
    Reactor and clarifier detention times were determined by instantane-
ously injecting salt and continuously monitoring effluent conductivity
until it returned to base level.  Mean probable detention times were deter-
mined by logarithmic probability plots of the resulting conductivity data.
A summary of typical detention times is given in Table 1.
                                    10

-------
                            Table 1.  TYPICAL DETENTION TIMES
Flow
rate,
qpm
5
10
15

Mean probable
detention time,
minutes
19.1
9.1
5.7
Reactors , »
Theoretical^"'
detention time,
minutes
21.0
10.5
7.0
Settling pools, v
Effi- ,.x
ciancy
percent
91
87
82
Mean probable
detention time,
hours
23. k
10.5
5.2
Theoretical^*"
detention time
hours
39
19.5
13
Effi- ,
ciency
percent
60
5k
kO
(a)
(b)
Theoretical detention time « volume r flow rate.
Efficiency = mean probable time f theoretical X 100.
Note:  The effective capacity of each reactor is 105 gallons.
       The effective capacity of each settling pool is 11,700 gallons,
       To convert from gpm to liters/minute, multiply by 3.785.

-------
    Sludge settling rates were difficult to determine, since a definite
supernatant/sludge interface never clearly developed during settling.
During the early part of this research, the position of the interface was
estimated to determine the settling curves.  Later, it was decided to
measure supernatant turbidity as a function of time in lieu of estimating
the interface position.  Sludge was therefore allowed to settle undisturbed
in 1000-ml graduated cylinders, and supernatant turbidity was periodically
measured.  It was possible to read the sludge buildup in the bottom of the
graduate even though the settling interface never clearly developed.
PROCEDURES FOR CHEMICAL ANALYSES
    Conductivity and pH were measured potentiometrically.  Total iron,
aluminum, magnesium, and calcium were determined by atomic absorption
                  (o-3\             (23)
spectrophotometry.      EPA methods     were used for sulfate, total
                                                              (2U)
solids, alkalinity, and turbidity determinations.   The Salotto     acidity
method was used in which hydrogen peroxide is added to oxidize the metals;
a cold titration is then made to pH 7-3.
CHARACTERISTICS OF REACTANTS
    AMD from Grassy Run was treated in these studies (Table 2).
              Table 2.   GRASSY RUN WATER QUALITY DATA, 1971

Parameter
PH
Specific conductance
Acidity as CaCO
Calcium
Magnesium
Aluminum
Sulfate
Iron (Total)
Temperature
Flow

Unit
PH
Mmhos/cm
mg/1
mg/1
mg/1
mg/1
mg/1
mg/1
Op
CFS^ '

Mean/ x
2.8U)
1100
^30
106
35
33
590
92
U8
10.8

Maximum
3.U
22liO
6^0
170
120
69
1200
170
57
uo

Minimum
2.5
3)4.0
130
18
21
18
76
Ik
39
1.8
Standard
Deviation
___
Ul5
125

12
12
260
39
5.U
9.U
/, \Median value.
V " / ITI — Anvt-rrs-otA-4- -P-wmvn /TTPO J-«. ntt wi / r^ *"irt TV1 1 1 T 4- "1 T\l ir VviT f^ ^OM
                                   12

-------
    Grassy Run is a small, heavily polluted stream in which an estimated
90 percent of the flow emanates from abandoned coal mines.  Approximately
70 percent of this pollutant flowed directly out of underground mines.
Some sewage is also present in the creek.  Virtually all iron was in the
ferric state.
    The limestone and hydrated lime used in this study were obtained from
Germany Valley Limestone Company, Riverton, West Virginia.  In order to
obtain the smallest particle size commercially available, the rock-dust
form of limestone was used.  All tests in this study were made using the
hydrated form of lime.
    Table 3 presents the manufacturer's chemical analyses and cost for
lime and limestone.  A spectrochemical analysis of the limestone was made
by BCR and is reported in Table U.
         Table 3-  MANUFACTURER'S CHEMICAL ANALYSES OF REAGENTS
                             Hydrated lime,a       Rock-dust limestone,0
Parameter	min. percent comp.	min.  percent comp.
CaO
MgO
CaCO- equivalent
Si02
72.00
o.Uo
130
—
53.0
0.38
98.8
0.^9
      cost $l8.00/ton ($19.85/metric ton) in bags or 0.90 cents/lb
(1.985 cents/kg).
 Limestone cost $6.00/ton ($6.6l/metric ton) in bags or 0.30 cents/lb
(0.66 cents/kg).
  Table k.  SPECTROCHEMICAL ANALYSISa OF GERMANY VALLEY LIMESTONE
                                (percent)
Loss on
ignition
1*3.0
Si02 M2°3 Fe2°3 Mg° Ca° Ti°2 Na2° K2° Mn°2
1.0 0.43 0.15 1.16 97.0 O.OU 0.02 0.1 0.03
       analysis identified only one compound —CaCO_ (96.
                                   13

-------
    Sieve analyses of as-received lime  and limestone  are presented  in

Table 5-


    Table 5.   SIEVE ANALYSES  (DRY) OF AS-RECEIVED LIME AND LIMESTONE
Screen size
                            Hydrated lime,
                            percent passing
Rock-dust limestone,
  percent passing
50 mesh
100 mesh
200 mesh
UOO mesh
Percent not passing sieve
96.2
24.3
2.8
0.9
3.8
99-1
58.7
28.3
9-7
0.9

-------
                                SECTION V
                                 RESULTS

OPTIMIZING LIMESTONE TREATMENT
    The limestone investigations basically followed the research plan
outlined in Table 6.  Each of the variables listed was investigated
independently first, and then cross-checked against other variables to
insure that the parameter under study exhibited consistent behavior
under a variety of conditions.

          Table 6.  RESEARCH PLAN FOR LIMESTONE INVESTIGATIONS
Order
of
study
1
2
3
1*
5
6
7
8
Variable
pH
Detention time
Temperature
Reactor baffles
Particle size
Limestone feed: dry vs slurry
Limestone feed: one vs two steps
Sludge recycle
Action
investigated
Effect on utili-
zation efficiency
ir
11
11
11
n
11
ti
    Some of the initial studies on the variables were performed on batch
scale equipment.  When warranted, results of batch studies were also dem-
onstrated under continuous-flow conditions.
Effect of pH
    Titration curves for limestone and lime on the Grassy Run water had
been developed earlier    and are presented in Figure 2.  Limestone's

                                   15

-------
                      pHll
                     pHIO
                      pH9
                      pH8
                    X
                    a
ON
                      PH7
                      pH6
                      pH5
                      pH4
                      PH3(
                                  I
 I
                I
  I
                                 157
315
               472      629      786     944


                REAGENT REQUIREMENT, mg/l


Figure 2. Titration curves for limestone and lime.
                                                                                   1100
1258

-------
titration curve indicated approximately 1.8 times as much limestone was
required to neutralize to pH 6 than to pH 5 and k times as much was re-
quired to neutralize to pH 7 than to pH 5.  To treat to pH 7 would thus
require excessive limestone.  Earlier tests    had shown that the iron
concentration could be reduced to less than 7 ing/1 at pH's above 5.  Most
of the limestone testing was done at pH 5 for two reasons:
                             /
    1.  The effluent quality was generally acceptable with the exception
        of pH; and
  s
    2.  The limestone reaction was quite sensitive in the pH 5 area of
        the titration curve.  This sensitivity made the reaction more
        amenable to precise control techniques, thus resulting in increased
        reliability and precision.
    Several continuous-flow tests were made at 15 gpm (56.8 1/min) at
various pH levels.  This flow rate corresponded to 5.7-min reaction time
and 5-2-hr settling time.  No reactor baffles were used.  Results of the
pH runs (Table 7) confirmed that iron concentrations meeting the Pennsyl-
vania requirement of 7 mg/1 could be obtained at pH 5-
      Table 7.  EFFECT OF pH OK EFFLUENT IRON, EFFLUENT TURBIDITY
                AND LIMESTONE REQUIREMENT8-
PH
5.1
5.6
6.1
7.0
Limestone usage ,
(lbs/1000 Gal)b
6.59
9-06
9-70
11.11
Effluent Quality
Iron,
(mg/1)
2.3
l+.O
2.3
0.7
Turbidity ,
(JTU)
53
kh
29
10 (est)
Alkalinity,
(mg/1)
1
2
2U
78
*Data recorded V28/70.  Flov rate:  15 gpm (56.8 1/min) or 5.7-min de-
tention time; initial acidity range:  khO-kjO mg/1.  No reactor baffles
were used.
 To convert lbs/1000 gal to kg/cu m, multiply by 0.120.
                                   17

-------
Neutralizing to pH 6 instead of to pH 5 required 1.5 times as much lime-
stone (compared to 1.8 in the titration curves).  Only 1.7 times as much
was needed to neutralize to pH 7 instead of to pH 5 in the continuous-
flow tests (compared to four times as much in the titration tests).  This
difference was attributed to the longer detention times provided in the
continuous-flow settling pool.   Chemistry data for this study are pre-
sented in Table 35 of the Appendix.
    In summary, optimum limestone usage was observed in the pH range of
5-0 to 5-5; at higher pH's, the usage was progressively less efficient.
Effect of Detention Time
    Detention time was evaluated by operating each side (Process A and
Process B) of the dual system at different flow rates and then by putting
two reactors in series as opposed to a single reactor.   Table 8 summarizes
the results of the detention tests.  Since the tests were made at different
times, the raw water quality characteristics varied; therefore, the varia-
tion in limestone usage from one test to another was due to these raw water
quality changes.
    Results shown in Table 8 indicate that increasing the detention time
decreased limestone usage.  The 19.^ percent reduction indicated in the
extreme case of two reactors in series at 5 gpm (l8.9 1/min) vs.  one re-
actor at 15 gpm (56.8 1/min) — detention times of ^3 minutes vs.  5 min-
utes — illustrates the strong time dependence of the limestone reaction.
    Data from these and later tests are included in Figure 3.   Detention
times of 20 to 25 minutes appear to be optimum, since large increases in
reaction time beyond this point produce only minimal gains in efficiency.
    Chemistry data summaries for this study are presented in Table 1*0 of
the Appendix.
Effect of Temperature
    A series of batch tests was used to determine the effect of tempera-
ture on the utilization efficiency of limestone.  The tests were made
during a very cold period to establish the lower points on the graph

                                   18

-------
                                     Table 8.   EFFECT OF DETENTION  TIME
H
VO
Test Conditions
Process A vs. Process B
15 gpm vs. 10 gpm
(5 min. vs. 8 min.)*
15 gpm vs. 5 gpm
(5 min. vs. 17 min.)*
Two reactors in series
@ 15 gpm vs. one @ 15 gpm
(Ik min. vs. 5 min.)*
Tuto reactors in series
§ 5 gpm vs. one § 15 gpm
(43 min. vs. 5 min.)*
Process A
Limestone
Usage, #/lQOQ
Gallons
7.50
7.01
<*.B3
6.77
Process B
Limestone
Usage, #/1000
Gallons
6.82
6.11
5.23
8.40
'Percent
Reduction
in Usage
9.1
12.8
7.6
I9.k
bias observed
difference
statistically
significant?
Yes
Yes
Yes
Yes
Appendix
Table
Number
36
37
38
39
             *Mean  Probable Detention Time

            Note:  To convert from gpm to liters/minute, multiply by  3.785.
                   To convert from Ibs/lOQO gal to kg/cu m, multiply  by 0.120.

-------
   45
                                                     TWO IN SERIES
                                                         @ 5 gpm
   40
   35
   30
 3
 C
 E  25
 •»
UJ
5
Z20
   15
   10
    LEGEND
   pH 6.5, 14-mesh baffles
 pH 5.0, no baffles
 pH 5.0 , 14-mesh  baffles
 NOTE:
 5 gpm = 18.9 l/min
 10 gpm = 37.9 l/min
 15 gpm = 56.8 l/min
                      10 gpm
                                                       5 gpm
                                                            • 15 gpm
                                               10 gpm
                                       15 gpm
               I
         I
                                  I
I
    30
        40                 50                 60
LIMESTONE UTILIZATION  EFFICIENCY, percent
 Figure  3. Effect of detention time on limestone utilization efficiency.
                                   20

-------
(Figure U).  Electric immersion heaters were used to elevate the temperature
to define the remainder of the graph.  Each batch test had !+0-minutes reac-
tion time and all effluent pH's were in the pH 6.2 to 6.6 range.
    The effect of temperature was slight until it exceeded approximately
13 C (55 F).  Between 13°C and 20°C, a noticeable increase in limestone
utilization efficiency occurred.  The efficiency at 20°C was 1.7 times
that at 11 C.  Ho further increase in efficiency was observed at temper-
atures above 25 C.
    The above series of tests was repeated using limestone that had been
ground to 99-6 percent passing ^00-mesh and the results are shown in
Figure 5-  Efficiencies of the ItOO-mesh limestone were roughly l.U times
(28 percent) higher than the regular rock-dust limestone.  The general
trend of the curves was the same, however.
    Since mine drainage is discharged from underground mines at 10 to 12 C,
heating the water prior to limestone treatment would be advantageous.
Such a step would probably be uneconomical, however, except during the
summer months when a shallow holding pond could be used for the influent
water.
    The 90-percent plus efficiencies seen in the UOO-mesh limestone at
20 C (Figure 5) were the highest efficiencies ever obtained with batch-
scale limestone treatment in these studies.
    Data collected from continuous-flow tests made later in the study are
presented in Figures 6 and 7 to illustrate the observed effects of temper-
ature on limestone utilization efficiency.  The effect of temperature was
not as pronounced under continuous-flow conditions as it was in the pre-
vious batch-scale studies.  Possibly the longer reaction times for the
batch tests account for this difference.  The effect of temperature was
still quite strong in both cases, however.
Effect of Baffles
    The use of baffles was initially investigated as a means of improving
mixing and thus increasing effective detention times.  Three baffle sys-
tems were tested:  wooden, small-mesh screen (30-mesh), and large-mesh
                                    21

-------
ro
         80
       c
       0)
       « 70
       0
       a
         60
         50
       3  40
          30
                                             1
I
I
I
                    5       10       15      20       25      30      35       40
                  (41°F)   (50°F)    (59°F)   (68°F)   (77°F)    (86°F)   (95°F)   (104° F)

                                          TEMPERATURE, °C
             Figure 4. Effect of temperature on limestone utilization efficiency (batch tests).

-------
ro
u>
                100
              * 90

              c
              0)
              v

              0>
              a


              vT 80
              u

              It 70
              <60
                50
                40
                          1
           I
1
I
I
I
I
  5       10       15       20      25

(41°F)   (50° f)  (59°F)    (68° F)   (77°F)

                  TEMPERATURE, °C
                                                                      30      35       40

                                                                    (86°F)  (95°F)   (104°F)
                Figure  5. Effect of temperature on 400-mesh limestone utilization efficiency

                         (batch tests).

-------
IX)
-ft-
         70
       O
       a
      >: 60
      z
      UJ
Z
o


<  50
N
         45
         40
                               I
1
I
                                     9       12      15

                                       TEMPERATURE, °C
                                                       18
                         21
                        24
27
           Figure 6. Effect of temperature on  limestone utilization efficiency @ pH 6.5 and

                     5 gpm (18 .9 l/m) during continuous-flow studies.

-------
VJI
         60
        o
        a
       Z 50
       Ul

       U
       N

       5 40
       i—
       3
         30
                                      I
         I
I
9       12       15

    TEMPERATURE, °C
                                                               18
                 21
24
27
            Figure 7. Effect of temperature on limestone utilization efficiency @ pHS.O and 15 gpm

                    (56.8 l/m) during continuous-flow studies.

-------
 screen (lA-mesh).   A summary of'the  effects  of the  baffles  on mean probable
 detention times is given in Table  9.  Wooden baffles  increased mixing
 efficiency more effectively than either  of the wire-mesh baffles.
     The effectiveness of various baffle  configurations  on limestone usage
 is summarized in Table 10.   Wooden baffles were more  effective than no
 baffles because of an increase  in  detention  time.   Small-mesh wire baffles
 were slightly better than wooden baffles,  even though their mixing effi-
 ciency was somewhat poorer.   Large-mesh  wire screen baffles improved lime-
                                                                \
 stone usage more than the small-mesh screen.   Use of  three  rather  than two
 large-mesh baffles showed little or  no advantage.
     Figure 8 shows the wire-mesh baffle  arrangement used.   It was  concluded
 that wire-mesh baffles increased limestone utilization  efficiency'by:
         1.  Removing coatings (probably  iron hydroxide)  from  the surface
             of the limestone particles by  abrasion against  the wire  screen;
         2.  Improving mixing efficiency  and  thus increasing detention  time
             and decreasing  short circuiting; and
         3.  Possibly breaking the  limestone  into smaller particles  (no
             tests  were made  to  confirm this, however).
                                                             «-.
     At a flow rate of 15  gpm (56.8 1/min), typical limestone  utilization
 efficiencies for the study were  as follows:
                                                          Percent
              No  baffles	;	••	  52
              Two wooden baffles	  56
              Two small-mesh  wire baffles	  58
              Two large-mesh  wire baffles	  60
     Chemistry summaries for  this study are presented  in  Table  U5 of  the
Appendix.
Effect of  Particle Size
                / 1 Q ^
    Ford et  al.      of BCR has emphasized the  importance of particle size
to the effectiveness of limestone  utilization.   BCR recommended use  of
minus-325-mesh stone for best results.
                                   26

-------
     Table 9.   LIMESTONE NEUTRALIZATION OF ACID MINE DRAINAGE LOGARITHMIC PROBABILITY PLOTS FOR
                MEAN REACTOR  DETENTION TIME
rv>
Flow
Rate*
GPM
15 GPM
15.1
15.0
14.85
Ik. 56
1ft. 85

14.85

14.85

14.85

10 GPM
10.2
5 GPM
5.20
5.00
4.90

5.00

Condition
in Reactor
Reactor No.

Standard*
Standard
2 Wood Baffles
2 Uood Baffles
2 Small 400-mesh
Screen Baffles
2 Large 200-mesh
Screen Baffles
3 Large 200-mesh
Screen Baffles
2 Reactors
in Series

Standard

Standard
Standard
2 Large 200-mesh
Screen Baffles
2 Reactors
in Series

A
B
A
A

B

A

A

A

A

A
B

A

A
Theoretical
Detention
Time (To)

6.95 Min.
7.00 "
7.07 "
7.21 "

7.07 "

7.07 "

7.07 "

14.48 "

10.3 Min.

20.2 Min.
21.0 "

21.4 «

43.0 "
Mean Probable Log-
arithmic Detention
Time (T)

5.84 Min.
4.76 "
7.00 "
7.07 "

5.51 "

5.79 "

6.29 "

13.68 »

7.62 Min.

16.6 Min.
17.7 "

19.5 "

42.9 "
Mixing
Efficiency,
Percent (T/To)

58
68
99
98

78

82

89
'
95

74

82
84

91

99

     *Standard = No baffles
    Note:   To convert  from gpm  to liters/minute, multiply by 3.785.

-------
                        Table 10.   EFFECT OF BAFFLES AT 15 BPM (56.8  LITERS/HIM)
to
CO
Test Conditions
Process A vs. Process B
Process A
Limestone
Usage, #71000
Gallons
Tuo blood Baffles vs.
no Baffles 7.66
Tuo Ulood Baffles vs. Tuo
Fine Mesh Screen Baffles 5.90
Tuo Large Mesh Screen Baffles
vs. Tuo Small Mesh Screen
Baffles 8.34
Three Large Mesh Screen
Baffles vs. Tuo Large Mesh
Screen Baffles 9.U2
Process B
Limestone
Usage, #71000
Gallons
B.5U
5.55
9.UO
8.97
Percent
Reduction
in Usage
10.3
5.9 ,
11.3
t».8
Ulas observed
difference
statistically
significant?
No
Yes
No*
No
Appendix
Table
Number
kZ

            "Too feu samples
            Note:  To convert from lbs/1000 gal to kg/cu m, multiply by 0.120.

-------
Note: To convert inches
to cm, multiply by 2.54.
                  INFLUENT /
                                              37.5 in.
                       HOSE
2X4 S4S
          -13"-
        PLYWOOD
                               TOP VIEW
                      (REACTOR SHOWING INFLUENT HOSE,
                          MIXER, AND BAFFLE LOCATIONS)
                    3V
   TYPICAL BAFFLE DETAIL
                                            BOLT
                                         PLYWOOD
SCREEN
     Figure 8. Wire-mesh reactor baffle sketch.

-------
    To evaluate the use of finely ground limestone on the Grassy Run water,
2000 pounds (907 kg) of rock-dust limestone were ground in a jet mill until
100 percent passed a 325-mesh screen.
    A dry sieve test on the finely ground stone was not successful due to
static electricity, and the particles would not pass through a 50-mesh
screen.  To alleviate the problem, the stone was washed through the sieves
with cold water.  Amounts retained on each screen were dried, weighed, and
compared to the original weight of the sample.
    For comparison, the wet-wash sieve test was also performed on standard
rock-dust limestone.
    Results of both tests are presented in Table 11.

                     Table 11.  WET-WASH SIEVE  TESTS
                                Limestone passing through sieve,
Mesh size                   	(percent)	
                            Standard rock dust      Jet-milled rock dust
   200                             T6.6                    100.0
                                   59-9                     99.6
    Initial batch tests were made to compare the reaction efficiency of
the finely ground stone to the regular stone.  Both these tests were made
using identical 100-gallon (378.5-liter)  batches of AMD and a UO-min
reaction time for each.  Results of these tests are given in Table 12.
             •  Table 12.  BATCH SCALE NEUTRALIZATION TEST

Date
Volume of water, gal.
Reaction time, min.
Initial pH
Final pH
Quantity limestone added, grams
Utilization efficiency, percent
Temperature, °C
Rock- dust
limestone
2/2/71
100 (378.5 £)
1*0
2.9
6.5
6U3.1
36.6
6
UOO-mesh
limestone
2/2/71
100 (378.5
UO
2.9
6.5
^57- 7
51.2
7


t}






                                   30

-------
    It required 28.8 percent less finely ground limestone (UOO mesh) than
regular limestone to do the same job.
    Following batch-scale testing, the UOO-mesh stone was compared to reg-
ular limestone under 15 gpm (56.8 1/min), 7-min detention time, continuous-
flow conditions.  Table 13 shows results of this test.  Limestone usage was
reduced even further under continuous-flow conditions; kk percent less kQQ-
mesh limestone was required than regular limestone.  Continuous-flow utili-
zation efficiencies of about 80 percent were obtained with the ^00-mesh
limestone.
    Continuous-flow comparisons (Table lU) were then made at a 5 gpm
(18.9 1/min) flow rate (21-min detention time).  Utilization efficiency
averaged better than 90 percent during the test run.  These efficiencies
were the highest ever recorded for continuous-flow limestone treatment on
Grassy Run AMD.
    Summaries of chemical analyses for these studies are presented in
Table k6 of the Appendix.
    BCR has also emphasized that not all limestones are effective in AMD
treatment.  They characterized the ideal stone as having a large surface
area, a high calcium content, and a low impurity level.
    A high-calcium, small particle-size limestone (98 percent passing 325-
mesh) sample was obtained from the National Gypsum Company, York, Penna.,
to illustrate BCR's results on the Grassy Run water.
    A batch neutralization comparison was made between the York stone and
the high-calcium, finely-ground 325-mesh stone (Table 15) used at Norton.
The limestone used at Norton was significantly superior in reactivity with
AMD, thus indicating, as BCR had observed, that considerably more was
involved than particle size.
                                    31

-------
                                Table 13.   COMPARISON OF UOO-MESH LIMESTONE  VS. REGULAR
                                LIMESTONE § 15 GPM (56.8 LITERS/MIN)
U)
K>
                            Process A
                   400-Mesh Limestone @ 15 GPM
             Process 8
Renular 'Rock Dust1 Limestone @ 15 GPM
Date
2/09/71
2/10/71
2/11/71
2/12/71
Aug.
Efflu-
ent
PH
5.3
5.0
5.1
5.0
5.1
Limestone
Usage,
#/1000 aal.
3.87
3.57
U.10
k.25
3.95
Utili-
zation
Effcv.
84.0%
89.0%
79.7%
77.0%
82.<*%
Stoichi
ometric
Factor
1.16
1.12
1.23
1.27
1.20
•Efflu-
ent
pH
5.1
5.1
5.2
5.2
5.2
Limestone
Usage,
#/1000 aal.
5.38
6.95
7.89
7.95
7.04
Utili-
zation
Effcy.
59.7%
45.9%
42.1%
41.7%
47.4%
Stoichi-
ometric
Factor
1.62
2.12
2.36
2. 38
2.12
Percent Re-
duction in
UsaQe
28.1
48.6
l+B.O
46.5
43.9
           Note:  To convert from lbs/1000 gal to kg/cu m, multiply  by 0.120.
                                              CHEMISTRY ANALYSES
                                                EFFLUENT WATER
                                          Process A
         Process  B
Date
2/09/71
2/10/71
2/11/71
2/12/71
Iron
10 mg/1
10
10
1.8
Turbidity
63 JTU
58
43
79
Iron
5 mg/1
15
15
1.3
Turbidity
Bl JTU
79
56
92
           NOTE:  Observed difference between  Process A and Process B uas  statistically significant at the  95-
                 percent confidence level.

-------
                Table Ik.   COMPARISON OF ^00-MESH LIMESTONE  VS. REGULAR LIMESTONE
                             @ 5 GPM (18.9 LITERS/MIN)
4QO-Mesh
Efflu-
Oate ent
pH
2/24/71 5.0
(Temp. 7°C)
2/25/71 5.1
(Temp. 6°C)
2/25/71 5.1
(Temp. 14°C)
2/26/71 5.0
(Temp. 11°C)
Averages 5.1
Process A
Limestone @
Limestone
Usage,
#/1000 qal.
2.72
3.27
3.57
3.24
3.20
15 GPM
Utili-
zation
Effcy.
98.1%
90.1*
81. 3%
94.7%
91. 2%
Regular
Stoichi-
ometric
Factor
1.00
1.07
1.17
1.02
1.07
Efflu-
ent
pH
5.1
5.1
5.0
5.0
5.0
Process
'Rock Dust1
Limestone
Usage,
#/1000 qal.
5.62
5.32
5.38
5.85
5.54
B
Limestone § 15 G
Utili-
zation
Effcv.
49.6%
47.7%
55.2%
53.9%
51.6%
Stoichi-
ometric
Factor
2.00
2.07
1.79
1.85
1.93
?M
Percent Re-
duction in
Usage
51.6
48.3
33.6
44.6
42.2%

LO
U)
          Note:   To  convert from lbs/1000 gal  to kg/cu m, multiply by  0.120.
                                             CHEMISTRY ANALYSES
                                         Process  A
                         Process  B
                            Date
Iron	Turbidity
            Iron
          Turbidity
                            2/24/71
                            2/25/71
                            2/25/71
                            2/26/71
 6.4 mg/1
 5.6
 5.5
 6.0
45 JTU
34
36
26
3.1 mg/1
5.0
7.0
4.0
45 JTU
28
60
35
          NOTE:  Observed difference between Process  A and Process B uas statistically significant at the  95-
                 percent confidence level.

-------
                 Table 15.  COMPARISON OF FINELY-GROUND GERMANY VALLEY LIMESTONE
                            AND FINE-MESH LIMESTONE FROM YORK,  PENNSYLVANIA

                                                    Germany Valley stone         York,  Pa.  stone
    	(325-meah)	(325-raesh)
    Raw water  pH                                          3.0                       3.0
    Ram water temperature, °C                              7                       13
    Amount of water, gallons                             100 (378.5 liters)       100 (378.5 Itr)
    Amount of limestone required, gpm                    243.4                     464.9
    Effluent pH                                            6.5                       6.5
    Reaction time, minutes                                40                       40
*"   Utilization efficiency, percent                       76.2                      39.2
    Percent passing 325-mesh                             100.0                      99.8

-------
Effect of Dry Feed vs. Slurry Feed
    Previous studies by EPA(5) indicated that slurry feeding had no ad-
vantage over dry feeding.  The use of a slurry feed system with wire-
mesh baffles in the slurry mix tank was suggested, however, since smaller
particles of limestone might be produced from abrasion against the wire
mesh.
    Three series of batch studies were conducted to test this hypothesis.
Limestone slurry was prepared by adding Uo pounds (18.1 kg) of limestone
to 100 gallons (378.5 liters) of water in a slurry tank equipped with two
wire-mesh baffles and a flash mixer.  Since the tank inevitably had dead
spots in it, the slurry could not be assumed to be uniform.  Aliquots of
the slurry were used to neutralize the AMD.  In order to determine the
quantity of limestone actually added, duplicate amounts of slurry were
taken.  The first portion of slurry was added to the AMD in the reaction
tank; the duplicate was filtered, dried, and weighed.
    A comparison of the effectiveness of dry feeding vs. slurry feeding
is presented in Table 16.  In the first test, slurry feeding was more
efficient.  In the second series, slurry feeding was only slightly more
efficient.  In the final tests, dry feeding was more efficient.
    Thus, no significant difference was observed between dry and slurry
feeding, but dry feeding is more desirable because of its simplicity.
    These studies may clarify the role of the wire-mesh baffles in neu-
tralization reaction.  If the "baffles do decrease particle size, then slurry
feeding should have been significantly superior to dry feeding because of
the limestone's extended exposure to the screens in the slurry mix tank.
Since it wasn't, the baffles must have little effect upon particle size.
Effect of Two-Step Limestone Treatment
    A feasibility test was made to determine if any advantage would be
obtained by adding limestone to the AMD in two separate steps.  Four
identical 125-gallon (U73-liter) batches of AMD were used for the study.
Each received a total of 600 grams of limestone.  Reaction time for all
tests was UO minutes.
                                    35

-------
               Table 16.   LIMESTONE SLURRY VS. DRY FEED NEUTRALIZATION STUDY BATCH SCALE TESTS
U)
ON
Date
12/30/70
12/30/70
12/30/70
1/19/71
1/19/71
1/29/71
1/29/71
Type of
Limestone Initial
Used pH
Dry Rock
Slurried
Slurried
Dry Rock
Slurried
Dry Rock
Slurried
Dust
Rock Dust
Crusher Dust
Dust
Rock Dust
Dust
Rock Dust
2.9
2.9
2.9
3.1
3.1
2.9
2.9
Amount of
Limestone Final
Required pH
578.
U65.
of
5UO.
of.
682.
6B
-------
    In Test No. 1,  the  entire quantity of limestone was added initially and
the pH was recorded every 5  minutes.   Half of the  limestone was added ini-
tially in Test No.  2, and the remaining half was added after 20 minutes.
In Test No. 3, 75 percent was added initially,  and the remaining 25 per-
cent was added after 20 minutes.   Test No.  fc was the  same as Test No. 1
to check the  reproducibility of results.
    Results of the  study (Table 17) indicate that  there is no advantage
to two-step treatment.   In fact,  it was more advantageous to add all the
limestone initially due to the more rapid and complete reaction at lower
pH's.  The limestone added in this study was 2.76  times the stoichiometric
requirement.
Simulated Sludge-Recycling Tests
    Unfortunately,  the  Norton facility did not  have the capability to con-
tinuously withdraw  and  recycle sludge.   To  investigate the feasibility of
sludge recycling, it was necessary to  store the sludge in a reservoir and
use this to simulate continuous withdrawal  conditions.
    All tests were  done on a batch basis.   As shown in Table 18, no sludge
was added to  the first  test.   Sludge was  added  to  the following tests in
progressively larger quantities.   For  the final test, the sludge that had
precipitated  from Test  No. U's neutralization was  used.
               Table 18.   SIMULATED SLUDGE-RECYCLING TESTS

Test
No.

1
2
3
h
j>

Initial
pH

2.9
2.9
2.9
2.9
2.9
Quantity of
sludge added,
(gal)a

0
2
h
8
12.5
pH 5-min
after sludge
addition

-
U.8
5.2
5-8
5.6
Amount of
limestone
addedj
(grams)
500.0
289.2
155- i*
5^.2
325.0

Final
PH

6.8
6.8
6.8
6.8
6.8
A 100-gal (378.5-liter) batch of AMD was used for each test.  Total reac-
tion time for each test was kO minutes.  The majority of the limestone
was added 5 minutes after sludge addition.
^To convert gallons to liters, multiply by 3-785-
bThe sludge resulting from the Test No. ^ neutralization was used for
Test No. 5-
                                   37

-------
                        Table 17.  LIMESTONE SPLIT TREATMENT  FEASIBILITY STUDY
                   TEST NO. 1
TEST NO. 2
TEST NO. 3
TEST NO. 4
Time
0
5 Min.
10 Min.
15 Min.
20 Min.
25 Min.
30 Min.
35 Min.
40 Min.
Limestone
Added pH
Grams
600 2.9
5.70
6.00
6.18
6.30
6.38
6.45
6.50
6.50
Limestone
Added pH
Grams
300 2.9
4.90
5.10
5.30
300 5.40
6.15
6.25
6.38
6.45
Limestone
Added pH
Grams
450 2.9
5.50
5.80
6.00
150 6.18
6.30
6.38
6.42
6.50
Limestone
Added pH
Grams
600 2.9
5.70
6.00
6.20
6.35
6.40
6.45
6.50
6.50

U)
OD
    Note:  In each test, quantities of limestone were added at various intervals  to a reaction  cham-
           ber containing 125 gallons of acid mine drainage and stirred vigorously for  the  specified
           time as indicated.  (125 gal = 473 liters).

-------
    As more sludge was added, less limestone was required (until Test No.
5).  The increased limestone requirement of Test Wo. 5 is more represent-
ative of continuous-flow recycling as Test U's neutralization did not
supply as much alkalinity in 12.5 gallons (Vf.3 liters) of sludge as was
available in 2 gallons (7.6 liters) of the original sludge.  This con-
clusion is confirmed "by the chemical analyses shown in Table 19.
    The original sludge was significantly higher in percent solids, alka-
linity, and calcium than were the sludges from Tests 1 through 5.  Sludge
iron and sulfate were not dramatically affected by the recycling.
    Increasing the amount of recycled sludge increased the effluent sul-
fate content.  In Test No. 5S for example, the effluent was near satura-
tion for calcium sulfate.  No chemical parameters other than sulfate were
adversely affected by sludge addition.
    Conclusions drawn from these tests were:
        1.  Sludge recycling should increase reaction efficiency and
            sludge solids content, though not to the extreme extent shown
            in these tests.
        2.  With the exception of increased sulfate content, product
            water quality was not degraded by sludge recycling.
        3.  Partial utilization of the large residual alkalinity in the
            sludge can be accomplished by recycling.
    Following the batch-scale study, a continuous-flow test was attempted.
Small electrical oscillating pumps were used to transfer sludge from the
settling pool back into the reaction chamber.  The recycle flow rate was
regulated by throttling the discharge from the pumps to attain a 1 gpm
(3.8 1/min) flow rate.  A 10:1 AMD to recycled-sludge ratio was used.
Results of the continuous-flow study are presented in Table 20.  A sum-
mary of chemical analyses for this study is presented in Table U6 of the
Appendix.
    The steady decrease in efficiency in Process B was attributed to the
failure of the sludge pool to achieve equilibrium conditions.  Every

                                   39

-------
                Table 19.  CHEMISTRY  ANALYSES  FOR SIMULATED SLUDGE  RECYCLE  BATCH TESTS
•p-
o
Sample
Designation
Ram Feed
Supernatant #1
Supernatant #2
Supernatant #3
Supernatant #4
Supernatant #5


Sludge
Sludge HI
Sludge #2
Sludge #3
Sludge #4
Sludge #5
pH
2.8
6.2
6.2
6.6
6.5
6.5
Percent
Solids
21.0
5.72
12.30
1U. 10
15.85
15,76
Alka-
linity
—
29
35
29
33
35


65,000
22,000
20,000
20,000
27,500
32,500
Cond.
Mmhos/cm
1300
1180
1220
1200
1200
1200






*

Acidity
471
4
4
4
4
3








Cal-
cium
110
300
340
360
400
365


5200
2500
1500
1400
2300
3000
Mg.
36
35
36
36
38
36








Iron
132
2.5
2.0
1.0
1.0
3.0


4300
3800
4300
4200
3600
4000
Al. SO^
44 655
2.0 696
2.0 655
2.0 819
2.0 955
2.0 1470


2500
2200
3000
2900
2900
2900
Sludge
Added,
Percent
by Vol.

0
2
4
8
12.5








Reaction
Efficiency,
Percent

37.7
66.1
121.5
351.2
58.8*









           •Sludge added  to  Test No. 5 was left  from Test No.  4's neutralization.


           NOTE:  Test data  taken December 10-11, 1970.

                  All units  expressed as mg/1  unless otherwise noted.

-------
        Table  20.  COMPARISON  OF SLUDGE RECIRCULATION  V/S NO RECIRCULATION
                    @ 10 GPM  (37.9 LITERS/MIN)
Process A - No Sludge Recirculation
10 GPM Flow (Raw Water)
Date
3/31/71
4/01/71
4/02/71
4/03/71
Efflu-
ent
pH
5.0
5.0
5.0
5.0
Limestone
Usage,
#/1000 qal.
5.674
6.696
6.784
6.246
Utili-
zation
Effcy.
52.0%
48.2%
50.1%
55.0%
Stoicni-
ometric
Factor
1.88
2.06
1.94
1.78
Process 8 - Recirculating 1 GPM of Sludge
10 BPM Flow (Raw Water)
Efflu-
ent
PH
5.1
5.1
5.1
5.2
Limestone
Usage,
#/1000 qal.
2.426
1.799
3.814
4.859
Utili-
zation
Effcv.
113.1%
179.5%
90.8%
70.9%
Stoichi-
ometric
Factor
0.88
0.55
1.09
1.39

Sludge
Volume After One Hour:
Process A = 1.0 Percent
Process B = 3.8 Percent
(Mote:   To convert from lbs/1000 gal to kg/cu m,  multiply by 0.120.
                                     EFFLUENT WATER
                                  Process A
Process  B
'• Date
3/31/71
4/01/71
4/02/71
4/03/71
Iron
2.2
10
2
5
Turbidity
mg/1 63
61
28
57
JTU



Iron
7 mg/1
18
2
2
TOrbiditv
85 JTU
98
34
58
 NOTE:  Observed difference between Process A and Process B was statistically significant at the  95-
       percent confidence level.

-------
attempt was made to move the recycle pump inlets as often as possible to
provide representative sludge for recycling, -but this was obviously not an
ideal mechanism for continuous removal.   Basically, the results of the con-
tinuous-flow study served to reinforce the findings of batch scale studies;
i.e., that recycling does have potential for improving limestone utiliza-
tion.  Further studies are in order on equipment more suited to the task.
Comparison of Limestone Utilization Efficiencies from Various Continuous-
Flow Tests
    At this point, the major variables that could be studied at Norton
had been investigated.  The important parameters in the limestone reaction
had been characterized, and those that could be controlled were reaction time
(20 minutes or more), particle size (rock dust  most feasible economically).
baffle configuration (two lU-mesh wire screen baffles), and effluent pH
(the minimum possible to meet water quality discharge standards).  Table
21 presents a summary of the limestone utilization efficiencies observed
for the various continuous-flow tests.   At pH 5-0, an increase in effi-
ciency of 1^ percentage points was observed between the most efficient
and least efficient studies.
        Table 21.   SUMMARY OF LIMESTONE  UTILIZATION EFFICIENCIES
                         FROM CONTINUOUS FLOW TESTS
Test Condition
pH 5.0
15-gpm (56.8 liter/min) flow rate
10-gpm (37-9 liter/min) flow rate
5-gpm (18.9 liter/min) flow rate
Two reactors in series § 15 gpm
Two wooden baffles @ 15 gpm
Two small-mesh screen baffles @ 15 gpm
Two large-mesh screen baffles § 15 gpm
Two reactors in series with large-mesh
baffles § 5 gpm
pH 6.5
Two large-mesh baffles @ 5 gpm
Two large-mesh baffles @ 10 gpm
Efficiencies, percent
Observed Mean

1*6, 1*7, 1*7, 52
51
55
59
55, 56
55, 58
55, 60

62

1*0, 60
36, 1*7

1*8
51
55
59
56
57
58

62

50
1*2
                                   1*2

-------
LIMESTONE VS LIME

    After defining the important parameters in the limestone reaction, it
was time to test the optimized limestone configuration against its com-
petitor, hydrated lime.  Two continuous-flow studies were made:  one at
10 gpm (37 = 9 1/min) at pH 6.U, and one at 5 gpm  (18.9 1/min) at pH 6.5.
Results of these tests are given in Tables 22 and 23.  Chemical analyses
summaries are presented in Table U6 of the Appendix.  The comparison was
necessarily evaluated on the basis of cost, using lime raw-material cost
of 0.90 cents/lb and 0.30 cents/lb for limestone.  At 10 gpm (37 = 9 1/min),
lime was 36 percent cheaper than limestone for the cost of reagent usage.
However, at 5 gpm (18-9 1/min), the costs were closer, with lime only 20
percent cheaper than limestone.  The improved sludge solids content of
limestone (15-6 vs 2.6 percent) could possibly more than compensate for
the initial cost advantage of lime.
    Raising the treatment pH to 6.5 (as opposed to the optimization test
pH of 5-0) decreased utilization efficiencies of limestone from approxi-
mately 60 percent at pH 5 to ^0 percent at pH 6.5 and 5 gpm (18.9 liter/
min).  An example of 10-gpm (37^9 1/min) efficiencies obtained using
limestone at pH's above 6 is given in Table 2k along with the chemical
analyses.
                                    fn£.\
COMBINATION LIMESTONE-LIME TREATMENTv  '
    Limestone treatment offered several advantages over lime treatment;
namely, higher density — lower volume sludges, cheaper raw materials,
easier handling of materials because of the less toxic nature of limestone,
and freedom from the pollution potentials of accidental overtreatment.
Conversely, the biggest disadvantage to limestone was its relatively in-
efficient reaction rate, which in most cases caused lime to be more eco-
nomical.  Limestone was also unable to produce pH's in excess of 7-0,
which are necessary for rapid ferrous iron oxidation.
    In most cases, limestone treatment could effectively compete with
lime only on ferric iron waters or waters in which the iron could be
cheaply oxidized to ferric before neutralization (such as by biological
oxidation).

-------
                              Table  22.   LIMESTONE US  LIME § 10 GPM  (37.9  LITERS/MIN)
-t-
-1=-
Process A - Limestone @ 10 gpm
Date
ID/19/71
10/21/71
10/27/71
10/28/71
10/29/71
Mean
#/1000
gal.
Limestone
19.V70
22.84.5
13.599
18.839
18.575
18.666
Final
pH
6.U
6. it
6.3
6.U
6. k
6.<»
Cost
Limestone/
1000 gal.
5.8M*
6.85*1*
<».0800
5.6520
5.573*
5.600*
Process B -
rf/lobo
gal.
Lime
k.MQ
U.866
3.160
3.536
3.865
3.973
Final
pH
&.k
6. <»
6.3
6.3
6.t,
6.k
Lime @ 10 gpm
Cost
Lime/
1000 gal.
3.996$
«..379«
2.8M»*
3.182*
3.V79*
3.576*

Material
Cost
Reduction
31.6%
36.1%
30.3%
<»3.7%
M.1X
36.1%
                             Percent Solids  (Sludge) 15.6%
Percent Solids (Sludge)  2.6%
                                                                   Process A
                               Process  8
             Typical Utilization Efficiencies
            Process A = 35.5%  Process B = 100%
                 Raui Material Costs;

            Limestone:  0.300/lb @ $  6.00/ton
            Lime     :  0.9Q«/lb @ $18.DO/ton
            NOTE:  Observed difference between  Process  A  and Process B was statistically significant at the 95-
                   percent confidence level.  To convert from lbs/1000 gal to kg/cu m, multiply by 0.120.  To
                   convert from cents/1000 gal  to  cents/cu m, multiply by 0.264.
Effluent Effluent
Date Iron Turbiditv
10/19/71
10/21/71
10/27/71
10/28/71
10/29/71
2.2 mg/1
2.0
l.i»
2.6
13 JTU
11
12
12
10
Effluent Effluent
Iron Turbiditv
1-5 mg/1
3.0
1.6
1.4
6 JTU
6
a
5

-------
                Table  23.   LIMESTONE US  LIME @  5 GPM  (18.9 LITERS/MIN)


Date

11/16/71
11/17/71
11/18/71
11/19/71
11/20/71
11/21/71
Mean
Typical
Process
Process A •
#/1000
gal.
Limestone
15.851
16.282
15.471
15.603
17.796
16.241
16.207
- Limestone (£? 5 gpm

Final
pH
6.6
6.6
6.5
6.5
6.5
6.5
6.5
Cost
Limestone/
1000 qal.
4.755*
4.885*
4.641*
4.681*
5.339*
4.872*
4.862*
Process B - Lime
#/1000
gal.
Lime
4.675
4.610
4.426
4.382
4.076
3.658
4.305
Final
pH
6.6
6.6
6.5
6.5
6.3
6.4
6.5
@ 5 gpm
Cost
Lime/
1000 gal.
4.208*
4.149*
3.983*
3.944*
3.668*
3.292*
3.875*

Material
Cost
Reduction
11.596
15.1%
14.2%
15.736
31.3%
32.4%
20.3%
Utilization Efficiencies
A = 39.5% Process
B = 100%




Process A
Raw

Material Costs:
r-. • n •*("!*-. /i H Ct,\ ft c
Lime: 0.90c/lb. @ $18










nn/tnn
.00/ton





Date
11/16/71
11/17/71
11/18/71
11/19/71
11/20/71
11/21/71
Effluent
Iron
.5 mg/i
1.5
.2
.7
1.3
.9
Effluent
Turbidity
11 JTU
8
8
10
13
19

Process 8
Effluent Effluent
Iron' Turbidity
1.7
1.4
2.6
.6
.8
1.2
mg/1 5 JTU
s 4
4
4
5
6
NOTE:   Observed difference between Process A and Process B was  statistically significant at  the  95-
       percent confidence level.   To convert from cents/1000 gal  to cents/cu m, multiply by  0.264.
       To convert  from lbs/1000 gal to kg/cu m, multiply by 0.120.

-------
                       Table 2<*-A.   HIGH pH  LIMESTONE RUN @ 10 GPM (18.9 LITERS/MIN)
CT\
Date
VI 3/71
VI 3/71
V1V71
VI 5/71
V16/71
Effluent
PH
6.2
6.2
6.5
6.k
S.U
Limestone Stoichio- Utili-
Requirement, metric zation
#/1000 Gal. Factor Efficiency
13.181
12.87*4
11.913
13.183
It*. 019
2.92 37.5%
2.86 37.U%
3.01 36.*t%
3.36 32.*t%
3.36 33.8%
Effluent Effluent
Iron Turbidity
1.6 50
1.1 58
2.3 55
U.5 58
..6 5.
Process
B
A
A
A
A

Note: To

Sample
Raw Feed
Pools A &
convert
Table
from lbs/1000 gal
2U-B. CHEMICAL
pH at
Time of Acidity
Analysis as CaC03 Cond.
2.7
B 6.3
^90 1180
3.0 920
to kg/cu m, multiply by
0.120.

ANALYSES FOR THE HIGH pH LIMESTONE STUDY
Total
Ca Mg Al Iron
96 31 32 88
2^50 250 1.1 3.1
Sul- Alka-
fate linity Turbidity
700
570 62 55
pH UJhen
Sample
Was Taken
2.7
6.3

           All units  are mg/1 except for  turbidity CJTU;,  conductivity (micromnos/cm) and pH.

-------
    In an effort to cover a broader spectrum of application, the logical
extension of the research was to combine! the limestone and lime processes.
Since limestone was cheap and highly reactive at lov pH's (less than 5),
as shown in Figure 2, it could be added first to the AMD.  Lime, being more
expensive but highly reactive to pH 9 and above, might be used to "polish"
the water previously treated with limestone.  In this manner, combination
treatment should enable both limestone and lime to be used in their re-
spective most efficient ranges of reactivity.  Utilization of limestone's
lower raw-material cost should therefore result in an overall cost reduc-
tion of the .combination as compared to either reagent alone.  An improve-
ment in sludge characteristics should also result from the use of lime-
stone.  Use of the lime should enable the application of the technique to
all ferrous iron situations where a high pH would be required for rapid,
efficient iron oxidation.
Batch Scale Tests
    Investigation of combination limestone-lime treatment began with a
series of neutralization studies made on 125-gallon (l*73-liter) batches
of Grassy Run water.  The neutralization reactor was equipped with a flash
mixer and two lU-mesh wire screen baffles.  Use of a 1500-gallon (5680
liter) storage reservoir insured consistent water quality for all tests.
    In each test, the AMD was neutralized to pH 6.5-  One test was made
using limestone alone to neutralize to pH 6.5; and another used lime
alone, the rest of the tests used limestone first to neutralize to a given
pH (3.5, k.Q, U.5, 5.0, 5-5, and 6.0) and then added lime to increase the
j?H to 6.5.  Reaction time for each neutralizing agent was 20 minutes.
Where combination treatment was tested, 20 minutes were allowed for the
limestone reaction and an additional 20 minutes for lime.  Since limestone
and lime involve independent reactions that must take place in that re-
spective .order, the reaction times used simulated a 5-gpm (l8.9 1/min)
flow rate through the Norton pilot plant.
    Results of these tests are shown in Table 25 and Figure 9-  Reagent
cost was the important parameter under consideration.  Combination

-------
                                Table 25.   LIMESTONE-LIME  BATCH  SCALE STUDY
4=-
Co
Initial
Limestone
PH
3.5
4.0
4.5
4.7
5.0
5.5
6.0
6.5


Amt. Lime-
stone Re-
quired ,lba/
1000 qal.
3.784
4.215
4.905
6.160
6.685
8.483
9.654
11.644

All lime
Cost of
Limestone
/1000 gal,
Cents
1.135
1.265
1.472
1.848
2.067
2.545
2.896
3.493


Final
pH
6.5
6.5
6.5
6.5
6.5
6.5
6.8
6.5

6.5
Amt. Lime
Required,
IDS/
1000 qal.
1.549
1.331
1.118
0.672
0.304
0.118
0.016
All lime-
stone
3.379
Cost of
Lime per
1000 gal,
Cents
1.394
1.198
1.006
0.605
0.274
0.106
0.014
All lime-
stone
3.041
Total Cost,
Cents per
1000 qal.
2.529
2.463
2.590
2.453
2.341
2.651
2.910
3.493

3.041
Total Neu-
tralizer
Added, mq/1
705.64
721.93
770.34
848.98
876.46
1038.19
1162.07
1398.62

5<»7.95
Stoi-
chio-
metric
Factor
1.17
1.19
1.27
1.40
1.45
1.72
1.92
2.31

0.91
          Reaction Time
20 minutes Limestone
20 minutes Lime
          Reactor Capacity = 125 gallons (473 Itr)  Lime costs 0.90 cents/lb; Limestone costs  0.3  cents/lb

          Note:  To convert from lbs/1000 gal to  kg/cu m, multiply by 0.120.  To convert from cents/1000 gal
                 to cents/cu m, multiply by 0.264.

-------
-e-
vo
                3.50< r-
                3.30c
              -  3.10c
              o
              O)
              o
              8
              | 2.90c
IS)
O
u
*. 2.70e
Z
UJ
o
IS
oc

  2.50c
                 2.30c
              (NOTE: All water was treated to a final pH of 6.5; where two-step

                     treatment was used, limestone was added first to various

                     pH's, then  lime was added to increase the pH to 6.5. Raw

                     water pH was .2.6. To convert cents/1000 gal. to cents/cu m,

                     multiply by 0.264.)
                   pH2.5
                                                  O
                                                                                     ©LIME ALONE
                                                                                     O LIMESTONE-LIME
                                                                                     OLIMESTONE  ALONE
                                I
                                     I
1
I
                                                                                  I
                 3.0
                   6.0
        6.5
                                 4.0                 5.0

                      FIRST STAGE (LIMESTONE) TREATMENT pH


Figure 9. Cost of various  combinations in two-stage limestone/lime treatment  schemes.

-------
treatment to limestone  (first stage) pH 5-0 proved to "be the most economical.
This combination was 33 percent cheaper than limestone alone and 23 percent
cheaper than lime alone.  As seen in Figure 9, virtually all combinations
of two-stage treatment were more economical than either lime or limestone
alone.
Continuous-Flow Tests
    The success of the batch scale tests led directly into continuous-
flow testing of the combination process.  For the pilot testing, Process
A consisted of two reactors in series.  Limestone was continuously added
to the first reactor to partially neutralize the AMD.  The partially neu-
tralized water was then pumped to a second reactor where lime was added.
Control of the pH of the first reactor effluent was accomplished by man-
ually adjusting a variable speed drive on the limestone dry feeder.
Lime addition to the second reactor was controlled automatically by a pH
controller to maintain a desired pH in the clarifier.
    Process B consisted of a single reactor also equipped with two wire-
mesh baffles.  Neutralizer addition to Process B was controlled automa"t-
ically by clarifier effluent pH.
    Combination limestone-lime treatment was first tested against lime-
stone treatment to pH 6.5 at flow rates of 5 and 10 gpm (l8.9 and 37.9
1/min) and then against lime treatment to pH 6.5, also at 5 and 10 gpm.
    Slight differences were observed when batch-scale results were ex-
panded to pilot scale.  A first-step (limestone) pH of 5.0, which was
optimum in batch studies, required virtually no second-stage (lime) treat-
ment in the pilot plant to reach a clarifier effluent pH of 6.5.  This
difference was attributed to the longer detention time provided by the
pilot plant clarifier.  Lowering the first-stage pilot plant pH near it pro-
duced results comparable to the batch scale studies at a first-stage pH
of 5.
    Tables 26 and 2? present the physical and cost data for combination
limestone-lime treatment versus limestone treatment at 5 and 10 gpm
                                   50

-------
                     Table 26.   LIMESTONE-LIME US LIMESTONE @  5 GPM  (18.9  LITERS/MIN)
vn

Date
8/17/71
9/21/71
9/29/71
9/30/71
10/1/71
10/2/71
10/5/71
Average
Process A
rf/lQQQ
Gal.
Limestone
5.453
2.569
3.955
4.348
3.867
4.495
5.905
4.370
- Combination Limestone & Lime
Lime-
stone
PH
4.7
4.5
3.8
4.1
3.9
4.4
4.2
4.2
#/1000
Gal.
Lime
.849
.973
1.253
1.339
1.388
1.252
1.562
1.231
Final
PH
6.7
6.4
6.4
6.4
6.4
6.4
6.4
6.4
Cost
Lime-
stone,
Cents
1.636
.771
1.187
1.304
1.160
1.349
1.772
1.311
& 5 qpm
per 1000
Lime ,
Cents
.764
.876
1.128
1.205
1.249
1.127
1.406
1.108
Process B - Limestone
Gal.*
Total,
Cents
2.400
1.647
2.315
2.509
2.409
2.476
3.178
2.419
#/1000
Gal.
Limestone
10.550
6.749
10.078
12.125
12.455
10.936
16.234
11.304
Final
PH
6.6
6.4
6.4
6.4
6.4
6.5
6.5
6.5
Cost*
/1000
Gal.
Cents
3.165
2.025
3.023
3.638
3.737
3.281
4.87D
3.391
Material
Cost Re-
duction
24.256
18. 7%
23.4%
31.0%
35.5%
24. 5%
34.7X
28.756
*Chemical cost only.
Raw Material Costs:
           Limestone = 0.30  cents/lb. @ $ 6.00/ton
           Lime      = 0.90  cents/lb. @ $18.00/ton

           NOTE:  Observed difference between Process A and Process B was statistically significant at the 95-
           percent confidence  level.  To convert from lbs/1000 gal to kg/cu m,  multiply by 0.120.  To convert
           from cents/1000 gal to cents/cu m, multiply by 0.264.

-------
                     Table 27.   LIMESTONE-LIME US  LIMESTONE @ 10  GPM (37.9 LITERS/MIN)
ro

Date
6/15/71
6/17/71
6/18/71
6/24/71
Average
Process A
#/1000
Gallons
Limestone
5.153
4.120
3.967
5.556
4.699
- Combination Limestone & Lime
Lime-
stone
PH
i*. 3
3.7
3.65
4.5
4.0
#/1000
Gallons
Lime
0.110
1.68
1.643
0.648
1.020
i 10 gpm
Cost /1000 Gal.
Final
pH
6.1
6.7
6.5
6.5
6.5
Lime-
stone,
Cents
1.546
1.236
1.190
1.667
1.410
Lime,
Cents
0.099
1.512
1.479
.583
.918
Total,
Cents
1.645
2.748
2.669
2.250
2.328
Process B
0/1000
Gallons
Limestone
10.524
12.44
12.80
10.846
11.653
- Limestone @
Final
pH
6.2
6.5
6.5
6.4
6.4
Cost
/1000
Gal.
Cents
3.157
3.732
3.840
3.254
3.496
10 gpm
Material
Cost Re-
ductions
Percent
47.9
26.4
30.5
30.9
33.4

Raui Material Costs:
          Limestone = 0.30  cents/lb. @ $ 6.00/ton
          Lime      = 0.90  cents/lb. @ $18.00/ton
          NOTE:   Observed  difference between Process A and Process B was statistically significant  at .the 95-
                 percent confidence level.  To convert from lbs/1000 gal to kg/cu m, multiply  by 0.120.  To
                 convert from cents/1000 gal to cents/cu m, multiply by 0.264.

-------
respectively (18.9 and 37-9 1/min).  Chemical analyses for all pilot plant
studies are presented in Table 28.
    For combination treatment versus lime treatment, the physical and
cost data are given in Tables 29 and 30.
Reducing Material Cost
    An overall summary of the effectiveness of combination treatment in
reducing material costs is given in Table 31, which shows that increasing
the reaction time had very little effect on the cost advantage of the com-
bination treatment as compared to lime treatment alone (25.2 vs. 25.7 per-
cent).  In the case of limestone, however, the difference in reaction time
was significant (33.^ vs. 28-7 percent), as the limestone reaction is
strongly time dependent.  Increasing reaction time significantly increased
limestone utilization efficiency and, in the above case, was obviously
a greater benefit to the limestone reaction than to the combination pro-
cess, since a reduction in cost advantage occurred at the lower flow rate.
       Table 31.  MATERIAL COST ADVANTAGE OF COMBINATION TREATMENT
Combination treatment
                                   Cost advantage of combination,
   as compared to: _ Flow; 10 gpm (37«9 1/min )a  5 gpm (l8.9 1/min )
      Lime                           25-2                   25-7
      Limestone                      33. ^                   28.7
Reaction time @ 10 gpm, 9-1 min.
 Reaction time @ 5 gpm, 19-1
Effluent Quality
    No advantage of one process over another was observed in terms of ef-
fluent water quality (Table 28).  As expected, all three processes pro-
duced effluents of similar quality since the solubilities of such elements
as iron and aluminum were more critically responsive to pH than to any
other parameter.
                                    53

-------
                           Table 28.   CHEMISTRY  ANALYSES:   PILOT PLANT STUDIES
V/l

5 GPM
Raw Feed (Grassy Run)
Limestone-Lime Effluent
Limestone Effluent
10 GPM
Raw Feed (Grassy Run)
Limestone-Lime Effluent
Limestone Effluent
5 GPM
Raw Feed (Grassy Run)
Limestone-Lime Effluent
Lime Effluent
10 GPM
Raw Feed (Grassy Run)
Limestone-Lime Effluent
Lime Effluent
PH

2.7
6.4
6.5

2.5
6.5
6.4

2.7
6.4
6.3

2.8
6.5
6.5
Alk.
as CaCO.,

0
16
18

0
20
55

0
15
10

0
18
10
Cond.
Mmhos/cm

1580
1320
1350

2240
1570
1640

1900
1520
1520

1480
1150
1150
Acidity
as CaCO, Ca Mg

560
0
0

600
0
0

570
2
1

370
1
1

120 45
360 38
360 k5

130 44
350 42
370 46

135 46
340 47
350 41

170 120
200 120
200 120
Total
Iron

130
0.8
1.4

110
1.8
1.6

120
0.6
0.5

32
1.0
1.0
Al.

43
2
3

69
1
1

41
1
1

23
2
2
so,

930
850
830

650
760
740

630
630
650

480
380
410
Turbid-
ity, JTU

_
9
12

_
15
18

-
3
5

_
a
3
         (VOTE:  All units expressed as mg/1 unless  otherwise noted.
                multiply by 3.785.
To convert  from gpm to liters/minute,

-------
                       Table 29.   LIMESTONE-LIME VS LIME  § 5 GPM (18.9 LITERS/MIN)
VJl
VJ1
Process A - Combination Limestone

Date


7/2U/71
7/25/71
7/29/71
7/30/71
Average
0/1000 Lime-
Gallons stone
Limestone pH

3.3.5 3.6
U.W7 ..3
..297 ..1
..093 3.8
..0.6 ..0
#/1000
Gallons
Lime

1.352
1.308
1.3.9
1.393
1.351

Final
pH

6.2
6.5
6.2
6.5
6.,
& Lime
Cost
Lime-
stone,
Cents
1.00.
1.33.
1.289
1.228
i.m
§ 5 gpm
per 1000

Lime,
Cents
1.217
1.177
1.21k
1.25k
1.216
Process B - Lime
Gal.

Total,
Cents
2.221
2.511
2.503
2. .8 2
*.*»
#/1000
Gallons
Lime

3.^62
3.U08
3.9U
3.751
3.63«.

Final
PH

6.2
6.3
6.2
6.3
6.3
@ 5 gpm
Cost
/1000
Gal,
Cents
3.116
3.067
3.523
3.376
3.271
Material
Cost Re-
duction,
Percent
28.7
18.1
29.0
26.5
25.7
Raw Material Costs:
          Limestone = 0.30 cents/lb. @  $ 6.DO/ton
          Lime      * 0.90 cents/lb. @'|l8.00/ton

          NOTE:  Observed difference between Process A and Process B bias  statistically significant at the 95-
                 percent confidence level.  To convert from lba/1000 gal  to kg/cu m, multiply by 0.120.   To
                 convert from cents/1000 gal to cents/cu m, multiply by 0.264.

-------
            Table  30.  LIMESTONE & LIME WS LIME @ 10 GPM (37.9 LITERS/MIW)

Date
8/5/71
8/6/71
8/7/71
8/8/71
8/9/71
8/10/71
Average
Process A
#/1000
Gallons
Limestone
1.213
.980
2.728
..062
2.609
..851
2.7.1
- Combination Limestone & Lime
Lime-
stone
pH
..2
3.6
..8
..5
..3
...
,.3
#/1000
Gallons
Lime
.283
..79
.303
.696
1.161
1.295
.703
Final
pH
6.5
6.5
6.5
6.5
6.5
6.3
6.5
Cost
Lime-
stone,
Cents
.36.
.29.
.818
1.219
.783
1..55
.822
M ID gpm
per 1000
Lime,
Cents
.255
..31
.273
.626
1.0.5
1.166
.633
Process B - Lime
Gal.
Total,
Cents
.619
.725
1.091
1.8.5
1.828
2.621
1.W5
#1000
Gallons
Lime
1.116
1.181
2.133
2.269
3.006
3.269
2.162
Final
pH
6.5
6.5
6.6
6.5
6.3
6.5
6.5
& 10 gpm
Cost
/1000
Gal,
Cents
1.00.
1.063
1.920
2.0.2
2.705
2.9.2
l.«S
Material
Cost Re-
duction,
Percent
38.3
31.8
.3.2
9.6
32..
10.9
25.2
Ram Material Costs:
Limestone = 0.30  cents/lb @ $ 6.00/ton
Lime      =0.90  cents/lb @ 818.00/ton
NOTE:   Observed  difference between Process A and Process B was statistically significant  at  the 95-
       percent confidence level.  To convert from lbs/1000 gal to kg/cu m,  multiply by 0.120.  To
       convert from cents/1000 gal to cents/cu m, multiply by 0.26^.

-------
Sludge Characteristics

    It was difficult to determine sludge settling rates, since a definite
supernatant/sludge interface never clearly developed during settling.  In-
stead of attempting to estimate the position of the interface, it was de-
cided to measure supernatant turbidity as a function of time.  Sludge from
all three processes was allowed to settle undisturbed in 1000-ml graduates,
and supernatant turbidity was measured periodically.  It was also possible
to read the sludge buildup in the bottom of the graduates, even though the
settling interface never developed.  Figure 10 presents graphs of both sets
of data.  Water treated with lime clarified most rapidly, followed closely
by limestone-lime and limestone treatment.  All three supernatants achieved
similar clarity at the end of 1 hour.
    Sludge produced by lime was the most voluminous of the three and com-
pacted gradually.  Limestone and limestone-lime treatment produced signif-
icantly smaller volumes of sludge than did lime treatment.  After 2 hours
of settling, lime sludge occupied 10 percent of its original volume, while
limestone and limestone-lime occupied less than h percent of their initial
volumes.
    Sludge samples taken during the individual pilot-plant test runs were
analyzed to determine the percent solids; results are presented in Table 32.
             Table 32.  COMPARISON OF SLUDGE PERCENT SOLIDS
- — — — 	 — — — — — —
Sludge solids,
percent
Flow rate,
gpm
5 (18.9 1/min)
10 (37-9 1/min)
Treatment method
Limestone-lime
10.7
6.0
Lime Limestone
2.3 13A
l.U 15.6
Theoretical
settling
time,
hr
39
19.5
    Longer detention time, which significantly increased sludge percent
solids for limestone-lime and for lime sludges, had little comparative
effect on limestone sludge.

                                    57

-------
             10
             10
20
  30       40
TIME, minutes
        [•] LIME
                                                   50
                                      ^ LIMESTONE
                                      0 LIMESTONE-LIME
20
  30       40
   TIME, minutes
50
60
Figure 10. Supernatent turbidity and sludge  buildup during  settling.
          (All tests were conducted at pH 6.5.)

-------
Advantages and Potentials of Combination Treatment
    Compared to straight lime or limestone treatment, combination lime-
stone-lime treatment provided greater than 25-percent reduction in material
cost for treatment of the Norton AMD to pH 6.5.  In addition to the cost
advantage, combination treatment produced a  sludge whose solids content
was up to five times higher than sludge produced by lime neutralization,
though not as high as sludge from  limestone  neutralization.  Combination
treatment was nearly as effective  in clarification as lime and better than
limestone.  The volume of sludge produced by combination treatment was
roughly one-third that of lime  treatment and slightly less than limestone
sludge volume.  All three processes  produced effluent waters of comparable
quality»
    Although this study was performed on a ferric iron water, the enormous
potential of combination limestone-lime treatment of ferrous iron AMD is
obvious.
    In theory, combination limestone-lime treatment should be applicable
to virtually all AMD situations.   Whether or not an economic advantage
can be realized is a matter that can only be determined by evaluating
each individual site and its required process parameters. The variables that
must be considered are:
        1.  Raw-material cost of rock-dust limestone and hydrated lime.
        2.  Effectiveness of limestone in AMD treatment (composition of
            stone).
        3.  Reaction time.
        U.  Treatment pH.
    The effects of reaction time have already been discussed.  Ford,
Young, and Glenn(20) reported in detail on the  effectiveness of various
limestones in AMD treatment and suggested a  simple test for stone evalu-
ation.  Raw material costs and  treatment pH  will be covered in the fol-
lowing discussion.
                                     59

-------
    To determine the effects of raw-material costs for lime and limestone
on process economics, usage data from Table 29 was chosen as an example.
Both the lime and limestone raw-material costs were varied and the cost
reduction calculated.  A plot of the resulting data is given in Figure
11.  When the lime/limestone raw-material cost ratio was less than 1-78:1,
the cost advantage of combination treatment over lime treatment no  longer
existed.  As the lime/limestone ratio increased, so did the advantage of
combination treatment.  At ratios below 1.78:1, lime was more economical.
    Assuming the water treated in Table 26 were all ferrous iron AMD and
a pH in the range of 9.0 were required for efficient oxidation by aeration,
roughly 20 percent more lime would be required (verified by titration
tests) to effect the pH increase from 6.k to 9.0.  Thus, Process B would
require an additional 0.727 lbs/1000 gallons of lime (20 percent of 3.63*0
for a total of k.36l.  The amount of limestone required by Process A
would not change but the amount of lime would be increased by the same
0.727 pounds to 2.078 lbs/1000 gallons of water.   In this example, cost
reduction because of combination treatment would decrease to 21.^ percent
at limestone = 0.30 cents/lb and lime = 0.90 cents/lb.
    As in the previous case, raw-material costs for lime and for limestone
were varied, and the resulting effect on combination-treatment cost ad-
vantage was calculated.  Figure 12, Case I, presents a graph of this data.
In this simulated ferrous iron situation, a lime/limestone raw-material
cost ratio of 1.80:1 was the breakeven point.   For ratios less than
1.80:1, lime was more economical and combination treatment gained rapid
advantage as the ratio increased.
    If 70 percent more lime were required to increase the pH from 6.5 to
9 (as was the case on a severely polluted ferrous water recently studied
by the Norton staff), then the cost advantage of combination treatment
over lime would decrease to 15.1 percent @ LS = 0.30 and L = 0.90 cents/lb
and the breakeven lime/limestone raw-material cost ratio would again equal
1.80:1.  These data are also presented in Figure 12, Case II.
                                   60

-------
CT\

H
                   2.000
               =   1.500
               o
               u
               O
               u
                    1.000
                   0.500
                                             BREAKEVEN POINT
                                         I
                                           I
1
I
I
I
I
                                           234567


                                         LIME/LIMESTONE RAW-MATERIAL COST RATIO
                                               45



                                               38



                                               30

                                               25



                                               13
                                                                                                         c

                                                                                                         0
                                                                                                         u
                                                                                                         tm

                                                                                                         9

                                                                                                         a.
                                                                                                         UI
                                               .«    I
                                           z

                                           oo
                                           s
                                           o
                                           u
o     E?

(-11)   °

(-24)  Z

(-50)  2

      U


      O
      ui
      oc

      i—
      to

      O
                       Figure 11. Effect of lime/limestone cost ratio on the economics of combination treatment.

-------
z
LU
t/»
O
U
to
O
u
  O
  _j
  <
      2.000
      1.500
   _j  1.000
      0.500
     2.000
      1.500
      1.000
     0.500
                                        CASE I (20% MORE LIME)
                             'BREAKEVEN POINT
                  1
                  LIME
                       LIMESTONE RAW MATERIAL COST REDUCTION
CASEH (70% MORE LIME)
                  I
                             BREAKEVEN POINT
                                 I
                I
I
                                 42
                                 37
                                 29
                                 25
                                 11

                                 -10) |
                                 -20) «
                                 -40 )•-*
                                     ui
                                                                            8
         28  £
         20  g
         17   *

         1-14) 8
                                                                      -(-28)
         0        1234567
                     LIME/UMESTONE RAW MATERIAL COST REDUCTION
     Figure 12. Lime/limestone cost ratio and process cost reduction  at
                pH 9 (simulated).
                                                                       8
                                      62

-------
    In summary, combination treatment appears to offer nearly as great an
economical advantage in ferrous iron situations as in ferric ones.   This
conclusion will be tested in future studies by EPA.  Although combination
treatment would require a higher initial investment in equipment, it is
felt that the advantage realized in reagent cost reduction would quickly
offset this increased initial expenditure.  Simple batch tests could be
performed on prospective waters to determine the feasibility of limestone-
lime treatment in each specific situation.
                                     63

-------
                               SECTION VI
                               DISCUSSION

    The effort to increase limestone utilization efficiency was moderately
successful.  Utilization efficiencies in the range of 50 to 60 percent
were not as high as had been hoped for but were significant improvements
over the early work at Norton/5' where 32 percent was a typical value.
                                                       (27)
    The results reported here correlate well with BCR's'    findings in
respect to efficiency, since BCR used twice the stoichiometric require-
ment for a utilization efficiency near 50 percent.  BCR, however, used a
smaller particle size (325 mesh) to enhance their efficiency somewhat to
the 50-percent value.
    Efficiencies as high as 65 percent were reported by Calhoun,
who used a rotary mill reactor oh a predominately ferric iron water.
                        (13)
The U.S. Bureau of Mines     (USBM) made no mention of limestone usage
during their rotary mill study, but they did observe that the rotary
reactor produced stone of extremely small particle size (89 percent
passed 200 mesh) which may account for Calhoun's reported efficiency.
Johns-Manville     used USBM's rotary mill reactor in their study at
Osceola Mills, Pa., and reported utilization efficiencies for limestone
                      (21)
of 3^ percent.  Lovell     performed a significant amount of work on
rotary mill limestone neutralization and reported maximum utilization
efficiencies as high as 80 percent.  Lovell readily points out that
accurate measurements of limestone usage in the rotary reactor configura-
tion are very difficult to make and quite unreliable for precise calcu-
lations.  Considering all the factors that surround the limestone reaction,
it is felt that the high observed efficiencies credited to rotary mill
application must be almost entirely due to particle size.  Reaction time
                                                        (13)
in Lovell's mill was 2 minutes.  Mihok and Chamberlain's     rotary mill
had 5 minutes, and Calhoun's had 3.5 minutes.  Reaction times this short
would be vastly inefficient if the particle size were not extremely small.

-------
Lovell performed particle  size  analyses,  not  on the  immediate reactor
discharge (as did USBM) but  on  the  settled sludge  in the lagoon; he re-
ported approximately 20 percent passing 325 mesh.  Approximately 30 per-
cent would not pass a  200-mesh  screen (therefore,  70 percent did pass).
In this report, only 30 percent of  rock-dust  limestone would pass a 200-
mesh screen.  It is apparent, therefore,  that the  high efficiencies ob-
served by rotary mill  reactors  were indeed due  to  the rotary reactor
grinding the stone to  extremely fine particle size and thus enhancing
reactivity.  Of course, the  rotary  action constantly removed reaction
coatings from the stones to  expose  fresh  reaction  surfaces.  Rotary re-
actors appear to be a  promising method for limestone treatment of AMD.
                   (19)
    Holland et al.,     used rock dust to treat  a  ferrous discharge and
reported that a 30-percent excess of limestone was required.  This works
out to a utilization efficiency near 80 percent.   A  sieve analysis of
Holland's rock dust indicated 67-3  percent would pass a 325-mesh screen—
hence the high efficiency.
    Holland went on to postulate that the apparent limestone sludge density
and volume advantage over  lime  sludge would be "more imaginative than real"
since lime sludge, when continually decanted, will vastly decrease in vol-
ume over extended periods  of time.   Although  Holland had no figures on
limestone sludge, he felt  it would  not react  the same way.  There has been
a great deal of emphasis from EPA—as well as from independent researchers—
on diminishing the sludge  volume (even by highly sophisticated techniques
such as vacuum filtration  and centrifugation) in order to increase land
use efficiency and reduce  costs of  disposal.  In most instances, it seems
impractical to store and decant sludge for extended  periods, as Holland
suggests.  It is felt  that the  immediate  advantages  of limestone sludge
are indeed real and important characteristics of the limestone process.
    Both Calhoun^16^ and Holland^19'  suggested the use of combination
limestone/hydrated-lime treatment.   Holland performed some studies on
the process but failed to  cite  any  particular advantages (although he
recommended its use on ferrous  iron waters).  Combination limestone/
hydrated-lime treatment(28)(32)  has been  used since  1957 in Japan in AMD
                                   65

-------
treatment of unusually high concentrations of iron (as high as 6000 mg/1).
Ikegami     reported first-step limestone requirements of 0.8 times the
stoichiometric requirement followed by 2.0 times the stoichiometric re-
quirement for lime addition.  He further recommended a limestone reaction
time of 20 to 30 minutes.  Ikegami points out that combination treatment
has been used since 1957 instead of lime treatment, since sludge volume
to be stored was reduced to a minimum.  His treatment scheme is tailored
toward byproduct removal as well as pollution abatement, since he selec-
tively precipitates iron, calcium sulfate, and aluminum during the treat-
ment process.
    Various researchers have published cost figures for neutralization of
AMD.  Table 33 contains a summary of selected results on various types of
waters, as well as various treatment techniques.  These reported costs
represent chemical costs for the neutralizing reagents used—not total
cost of treatment—and have been converted to cost per mg/1 of acidity
per million gallons of water for direct comparison.  No corrections were
made for variations in raw-material costs between the individual reports.
Soda ash costs were inordinately high at Uo to 50 cents per mg/1 acidity
per million gallons.  Good correlation was observed between the results
of independent researchers on soda ash.  Reported lime costs ranged from
5-3 to 8.5 cents.  Rock-dust limestone costs were on the same order as
lime, with a range of 5-8 to 9-8 cents per mg/1 acidity per million gal-
lons.  Agreement between researchers in both of these areas was also
good.  Combination limestone/hydrated-lime costs of ^.3 cents ranked
second only to rotary mill limestone treatment, whose costs ranged from
1.8 to 2.8 cents per mg/1 acidity per million gallons treated.  Agreement
between Lovell     and Calhoun     on rotary mill results was also good.
    Estimates of total treatment costs for neutralization (given in
Table 31*) ranged from 8.7 cents per 1000 gallons for lime treatment at
Morea Strip Pit to $1.28 per 1000 gallons for lime treatment at Bethlehem
Mines Marianna No. 58 Mine-  Limestone costs ranged from 13.6 to 76 cents
per 1000 gallons.  Johns-Manville     showed a significant cost advantage
for combination limestone/hydrated-lime treatment as compared to lime
                                   66

-------
           Table 33.   COMPARISON OF  REAGENT  COSTS FOR  TREATING MINE  DRAINAGE
~ Chemical Costs
Source
Corsaro, et al
Corsaro, et al
Uilmoth & Hill(5)
Uilmoth & Hill(5)
Uilmoth & Hill(5)
(yj\
Ford^27;
Lovell(21>
Lovell(21)
Lovell(21)
( i fil
CalhounUb;
This Report
This Report

This Report
Type Treatment Acidity
mg/1
Lime - Ferrous ' Iron
Lime - Ferrous Iron
Limestone (Rock Oust) - Ferric
Lime - Ferric Iron
Soda Ash - Ferric Iron
2+
Limestone (Recycle Sludge) - Fe
Lime - Ferrous Iron
Soda Ash - Ferrous Iron
Limestone - Rotary Mill - FeZ*
3+
Limestone - Rotary Mill - Fe
Limestone (Rock Oust) - Fe
Lime

Combination Limestone - Lime
1400
650
600
600
550

190
513
567
1500

360
560
570

570
Cents Cents per
/1000 mg/1 acid
Gal. itv/106 G
11.0
5.5
5.9
3.2
26.8

1.1
3.4*
25.1*
2.7

1.0
3.4
3.3

2.4
7.9
8.5
9.8
5.3
48.7

5.8
6.6*
44.2*
1.8

2.8
6.1
5.7

4.3
Raw
1- Material
i. Cost
@ $22.00 ton bulk
@ $22.00 ton bulk
& $6.00 ton bags
§ $17.00 ton bags
@ $100.00 ton bags

N. A.
@ $20.00 ton bulk*
@ $100.00 ton bags*
§ $3.28 ton bulk

§ $4.60 ton bulk
@ $6.00 ton bags
@ $18.00 ton bags
Ls-$6.00 ton bags
@ L-418.QO ton bags
•Assuming:  Lime cost  of $20.00 per  ton bulk
           Soda Ash cost of $100.00 per ton bags

Note:   To convert from cents/1000 gal  to cents/cu  m, multiply by 0*264.

-------
                              Table 3k.  ESTIMATES OF TOTAL TREATMENT COSTS
CO
Source
Draper (31)
( 17)
Corsarg. et al
Tyoout(29'
(29)
Tybout^'
( ?9)
Tybout^3'
(VQ)
Tyboutv"'
(29)
Tyboutv^ J
{
-------
treatment on Hollywood's Proctor No.  2 water (limestone/hydrated-lime
cost 77 cents per 1000 gallons  and lime cost $1.16 per 1000  gallons) for
a 3^-percent cost reduction.  This was slightly better than  the 25-percent
cost advantage noted in this  report.
SULFATE-CONDUCTIVITY CORRELATION
    During the course of studies at Norton and at  a temporary test site
in Pennsylvania, a  significant  volume of sulfate data  was collected under
a variety of conditions.   An  attempt  was made to correlate the sulfate and
conductivity parameters.   The resulting data plots and fourth-order
regression analysis lines  of  best fit are given in Figures 13 and 1^.
Considerable scatter is apparent in the data points for both sites.  Dr.
       (21)
Lovell,    of Pennsylvania State University, considered the  same correla-
tion on four test waters at Hollywood, Pa.,  and found  a surprisingly good
correlation.  There was considerably  less scatter  in Lovell's data for
two reasons:  First, Lovell used a gravimetric sulfate technique that was
superior in accuracy and precision to the turbidimetric method used at
Norton; and second, the variation in  Lovellfs test conditions was probably
not as great as  those used by EPA.
    A  direct comparison of Lovell's and EPA's correlation is shown in
Figure 15.  The  general trend of both relationships was the  same.  Con-
ductivity may be useful as an indirect sulfate measurement technique.
                                    69

-------
  4375
  3750
E  3125
 *.
Z
O
H-
^2500
u
Z
O  1875
  1250
   625
     EQUATION:
Sulfate (mg/l) = 59.89488247-
      0.02795421776(0) +
      8.887542104 x 10^ (C)2-
      3.141224943 x 10-'(C)3 +
      3.894139494 x 10'" (C)«
      where C = Conductivity
      (from 386 Data Points)
       0                1000              2000             3000
                             CONDUCTIVITY, micromhos/em
    Figure 13. Sulfate-conductivity relationship at Norton, West Virginia.
            4000

-------
  4375
  3750
o»
E
 k.
z
o
3125
of  2500
Z
ui
 4th ORDER REGRESSION ANALYSIS EQUATION:
  SULFATE (mg/l) =-56.13828615  + 1.161732301 (C)-
• 4.046974728 x 10'4 (C)2 +1.469000499 x KT7(C)3-
  1.637894642 x 1Q-"(C)4
  where  C = Conductivity
  (from 95  data points)
Z
o
u
1875
   1250
    625
                        1000             2000             3000
                               CONDUCTIVITY, micromhos/cm
                                                                          4000
             Figure 14. Sulfate-conductivity relationship  at  Mocanaqua, Pa.

-------
ro
            4375
            3750
            3125
          <2500
          at
u

O 1875
u
            1250
             625
                0             1000            2000           3000           4000

                                    CONDUCTIVITY ,  micromhos/cm

                Figure 15. Comparison between EPA and Lovell's sulfate-conductivity correlation.

-------
                                 SECTION VII

                                 REFERENCES
 1.  U. S. Bureau of Mines,  Advanced Data on Coal-Bituminous  and Lignite—
     1970, U.S. Bureau of Mines,  Pittsburgh, Pa.,  February  1972

 2.  Hoffman, R. H. , Pennsylvania Department of Environmental Resources,
     Division of Mine  Drainage  &  Erosion, Personal Correspondence,
     June 12, 1972

 3.  Cal dwell, Don E.,  West  Virginia Department of Natural  Resources,
     Division of Water Resources, Personal Correspondence,  June 13, 1972

 h.  Regan, Herman,  Chief Sanitation Engineer,  Kentucky Water Pollution
     Control Commission,  Personal Correspondence,  June 13,  1972

 5-  Wilmoth, Roger  C.  and Ronald D.  Hill, Neutralization of  High Ferric
     Iron Acid Mine  Drainage, Federal Water Quality Administration, Re-
     port 14010ETV 08/70,  Washington, D.C., August 1970

 6.  Care, R. R. and E. A. Zawadzki,  The  Evaluation of the Use of Acid
     Retardation Resins for  Control  of Acid Mine Drainage. Report BCR1-182,
     Bituminous Coal Research,  Inc.,  Monroeville,  Pa., December
 7.  Zawadzki, E. A. and R. A.  Glenn,  Sulf ide Treatment of Acid Mine Drain-
     age, BCR Report L-290, Appalachian Regional Commission, Bituminous
     Coal Research, Inc., Monroeville, Pa., July 1968

 8.  Braley, S. A., G. A. Brady and R. S. Levy, A Pilot Plant Study of
     the Neutralization of Acid Drainage from Bituminous Coal Mines,
     Sanitary Water Board, Pennsylvania Department of Health, Harrisburg,
     Pa., April 1951

 9.  Clifford, J. E. and C. A.  Snarley, Studies of Acid Mine Waters with
     Particular Reference to the Racoon Creek Watershed, Battelle Memorial
     Institute, Columbus, Ohio,
10.  Glover, H. G., The Control of Acid Mine Drainage Pollution by Bio-
     chemical Oxidation and Limestone Neutralization Treatment, 22nd
     Industrial Waste Conference, Purdue University, May 1967

11.  Zurbach, P. E., Dissolving Limestone from Revolving Drums in Flowing
     Water. Trans. American Fisheries Society, Vol. 92, 2, April 19o3
                                     73

-------
12.  Wheatland, A. B., and B. J. Borne, "Neutralization of Acidic Waste
     Waters in Beds of Calcined Magnesite," Waste and Waste Treatment ,
     January 1962

13.  Mihok, E. A. and C. E. Chamberlain, Factors in neutralizing Acid Mine
     Waters with Limestone, Second Symposium on Coal Mine Drainage Research,
     Mellon Institute, Pittsburgh, Pa., May 1968

lU.  Hoak, R. D., C. J. Lewis, and W. W. Hodge, "Treatment of Spent Pickling
     Liquors with Limestone and Lime," Industrial and Engineering Chemistry,
     Vol. 37, June
15-  Deul, M. and E. A. Mihok, Mine Water Neutralization Research, U. S.
     Bureau of Mines, Report Number 6987, Pittsburgh, Pa., 1967

16.  Calhoun, F. P., Treatment of Mine Drainage with Limestone, Second
     Symposium on Coal Mine Drainage Research, Mellon Institute, May 1968

17.  Corsaro, J. L. , C. T. Holland, and D. J. Ladish, Factors in the Design
     of an Acid Mine Drainage Treatment Plant , Second Symposium on Coal Mine
     Drainage Research, Mellon Institute, May 1968

18.  Ford, C., R. Young, and R. Glenn, Optimization and Development of Im-
     proved Chemical Techniques for the Treatment of Coal Mine Drainage,
     Bituminous Coal Research, Inc., Federal Water Quality Administration
     Report 1U010E12 01/70, Washington, D. C., January 1970

19.  Holland, C. T., R. C. Berkshire, and D. F. Golden, An Experimental
     Investigation of the Treatment of Acid Mine Water Containing High
     Concentrations of Ferrous Iron with Limestone, Third Symposium on
     Coal Mine Drainage Research, Mellon Institute, Pittsburgh, Pa., May
     1970

20.  Ford, C. , J. Boyer, and R. Glenn, Studies on Limestone Treatment of
     Acid Mine Drainage - Part II, Bituminous Coal Research, Environmental
     Protection Agency Report 1U010E12 12/71, Washington, D.c". , Dec. 1971

21.  Lovell, Harold L. , An Appraisal of Neutralization Processes to Treat
     Coal Mine Drainage, Pennsylvania State University, Environmental
     Protection Agency Report 1^010 EFN (Date not certain yet)

22.  Lovell, Harold L. , Experience with Biochemical-Iron-Oxidation Limestone-
     neutralization Process, Fourth Symposium on Coal Mine Drainage Research,
     Mellon Institute, Pittsburgh, Pa., April 1972

23.  Environmental Protection Agency, Methods for Chemical Analysis of Water
     and Wastes, 1971, Cincinnati, Ohio

2k.  Salotto, B. V., et al., Procedure for Determination of Mine Waste
     Acidity, Paper given at 15Uth National Meeting of the American
     Chemical Society, Chicago, Illinois, 1966

-------
25,  Bituminous Coal Research, Inc.,  "Studies  on  Limestone Treatment of
     Acid Mine Drainage," Federal Water  Quality Administration, Research
     Series 1^010 EIZ 01/70, Washington, D.  C., 1970

26.  Wilmoth, Roger C., Robert B. Scott, and Ronald D. Hill, Combination
     Limestone-Lime Treatment of Acid Mine Drainage. Fourth Symposium on
     Coal Mine Drainage Research, Mellon Institute, Pittsburgh, Pa.,
     April 1972

27.  Ford, Charles T., Development of a  Limestone Treatment Process for
     Acid Mine Drainage, Fourth Symposium on Coal Mine Drainage Research,
     Mellon Institute, Pittsburgh, Pennsylvania,  April 1972

28.  Ikegami, T., Recent Practice of  Waste Water  Treatment at Yanahara
     Mine, Joint Meeting of Mining and Metallurgical Institute of Japan
     and American Institute of Mining, Metallurgical, and Petroleum Engi-
     neers, Tokyo, Japan, May 1972

29.  Tybout, Richard A., A Cost-Benefit  Analysis  of Mine Drainage, Second
     Symposium on Coal Mine Drainage  Research,  Mellon Institute, Pittsburgh,
     Pennsylvania, May 1968

30.  Johns-Manville Products Corp., Rotary Precoat Filtration of Sludge
     from Acid Mine Drainage Neutralization, Environmental Protection
     Agency Report lUOlO DII 05/71, Washington, D. C., May 1971

31.  Draper, John C., Mine Drainage Treatment Experience, Fourth Symposium
     on Coal Mine Drainage Research,  Mellon  Institute, Pittsburgh, Pa.,
     April 1972

32.  Shimoiizaka, Junzo, S. Hasebe, H. Sato, and  T. Takahashi, The Utili-
     zation of Calcium Carbonate and  Calcium Hydroxide as Neutralizing
     Agent, in Acidic Mine Water Disposal. "The  Technology Reports of
     Iwate University," Volume 5, March 1971
                                     75

-------
                              SECTION VIII
                                GLOSSARY

CALCULATIONS AND DEFINITIONS
    Mean Probable Detention Time:  A mathematical approximation of the
average actual time a flowing liquid is detained in a vessel.  The value
is determined by tracer studies in which the tracer concentration exiting
from the tank is monitored in respect to time.  Two values are plotted on
probability paper for each measurement point; i.e., cumulative tracer
quantities to each point f total cumulative tracer quantity measured
(percent passing) and the ratio of elapsed time of each measurement 7
theoretical detention time of the vessel.  The theoretical time equals
the tank volume divided by the flow rate.  The resulting probability
plot is interpolated to derive the time for 50 percent passing.
    Detention Efficiency:  The ratio of mean probable detention time to
theoretical detention time expressed as a percentage.
    Utilization Efficiency;  A measure of the proportion of a neutralizer
that reacts with the acid water as compared to the amount originally
added.  Since alkalinity imparted to the water is considered a benefit,
the formula for utilization efficiency is:
                                 Alkalinity Used
        Utilization Efficiency = Alkalinity Added
            therefore,           Influent Acidity - Effluent Acidity +
        Utilization Efficiency = Effluent Alkalinity (as CaCOo)    X 100
                                 Alkalinity Added (as CaCO )
    Stoichiometric Factor;  The ratio of amount of neutralizer required
to treat original amount of acid present:
                                Alkalinity Added (as CaCO,.)
        Stoichiometric Factor = Influent Acidity (as CaCO_)

-------
    Statistically Significant Differences;  It was desirable to determine
mathematically if the apparent differences between Process A and Process
B were truly due to the variable under  study or just to random variation.
A statistical test was chosen to assist in data evaluation.  The test
used was the paired observation t test, which calculated a test statistic
t from the differences between paired measurements of Processes A and B
of the neutralization system.  The  95-percent confidence level was used.
When the calculated statistic t, exceeded the value from t tables, a sig-
nificant difference between  the processes was indicated.  If the calcu-
lated value was less than the table value, it was not possible to say
that the means of the two processes differed  (hence was not significant
at the specified confidence  level).
                                    77

-------
SECTION IX




 APPENDIX
    78

-------
                    Table 35.  CHEMICAL ANALYSES FOR VARIABLE pH STUDY
Sample
April 28.
Ram Feed
Pool B
April 29,
Raw Feed
Pool A
Pool B
April 30,
Ram Feed
Pool A
pH at
Time of
Analysis
1970
2.3
7.1
1970
2.4
6.4
4.8
1970
2.4
5.3
Acidity
as CaCO,
440
0
450
2
30
470
25
Cond.
1300
1000
1300
1000
1000
1400
1100
Ca
120
380
105
300
280
130
380
Mg
33
33
33
33
33
33
33
Al
34
0.3
30
1.0
6
36
2
Total
Iron
88
0.7
92
2.3
2.3
120
4
SO, Alk. Turb.
790 0
750 78 10*
760
820 -24 29
820 1.0 53
870
870 2 44
pH When
Sample
was Taken
2.3
7.0
2.4
6.1
5.1
2.4
5.6
*Estimated
                                                                                 •?

NOTE:  All units are mg/1 except for pH,  conductivity  (micromhos/cm),  and  turbidity (JTU).

-------
              Table  3S.   EFFECT OF  DETENTION  TIME - 15 GPM  (56.8 L/M) US  10 BPM (37.9  L/M)
CO
o
PROCESS
A = 15 GPM
Limestone
Effluent Usage*
Date pH #/1000 qal.
6/2/70 5.1
6/2/70 5.0
6/3/70 4.9
6/17/70 5.1
6/17/70 5.1
6/18/70 5.1
6/19/70 5.1
Means 5.1
8.03
7.89
8.00
6.55
6.11
6.35
9.59
7.50
Typical Utilization Efficiencies:
Process A = 46% Process B = 51%

Statistical Analysis
t limits from tables _+ 2.
t calculated from data -4.
Idas difference significant?

447
06
Yes X No
PROCESS B = 10
Effluent
PH
4.9
4.9
5.0
4.8
5.0
4.8
4.8
4.9

Date
6/2/70
6/2/70
6/3/70
6/17/70
6/17/70
6/18/70
6/19/70
Limestone
Usage,
#/1000 qal.
6.59
7.54
7.61
5.55
5.56
5.45
9.42
6.82
Process
GPM


Percent Reduction
in Limestone Usaqe








A
Iron Turbidity
8.0 mg/1
14
2.4
1.6
1.4
2.6
5.2
48 JTU
79
43
57
39
36
42
17.9
4.4
4.9
15.3
9.0
14.2
1.8
9.1
Process








B
Iron Turbidity
4.0 mg/1
4.8
2.0
1.0
1.6
1.2
2.5
18 JTU
26
22
41
21
28
29
          Note:  To convert from lbs/1000 gal to kg/cu mt multiply by 0.120.

-------
           Table 37.   EFFECT OF DETENTION TIME - 15 GPM  (56.8 L/M) US 5 GPM (18.9  L/M)
CD

Date
5/7/7D
5/8/70
5/25/70
5/26/70
5/27/70
5/26/70
Means

PROCESS A
Effluent
pH
5.0
5.3
6.0
«*. 8
4.6
4.9
5.1

= 15 GPM
Limestone
Usage,
#/1000 pal.
7.12
5.34
8.71
6.62
7.56
6.66
7.01

Typical Utilization Efficiencies:
Process A = 47% Process B = 55%
Statistical Analysis
t limits from tables +_ 2.571
+ r>al n il at.crl Frnm rlata —4.63
PROCESS 3
Effluent
PH
5.0
5.0
5.9
5.1
4.7
4.9
5.1


Date
5/7/70
5/25/70
5/26/70
5/27/70
= 10 GPM
Limestone
Usage,
#/100Q qal.
5.77
5.02
7.41
5.56
6.51
6.39
6.11
EFFLUENT
Process A
Iron Turbidity
7 mg/1 42 JTU
5 21
6 65
15 68

Percent Reduction
in Limestone Usaqe
18.9
6.0
14.9
16.0
13.9
4.3
12.8
QUALITY
Process 8
Iron Turbidity
1-6 mg/l 10 JTU
3 12
6 16
2.8 13
         Was difference  significant?  Yes  X   No 	
         Note:   To convert from Ibs/lODO gal to kg/cu  m, multiply by 0.120.

-------
               Table  38.   EFFECT OF DETENTION TIME  - TWO REACTORS IN SERIES US ONE REACTOR
oo
Two reactors in
Process A = Series © 15 qpm
Effluent
Limestone
Usage,
Date pH #/10DO qal.
10/23/70 5.1
10/28/70 5.0
11/10/70 5.1
11/10/70 5.0
11/11/70 5.0
11/12/70 5.0
Means 5.0

5.53
5.74
3.66
3.88
4.91
5.26
4.83

One reactor
Process B = @ 15 opm
Effluent
Limestone
Usage,
Percent Reduction
pH #/100Q qal. in Limestone Usaqe
5.0
5.0
5.1
5.1
5.1
5.2
5.1

Typical Utilization Efficiencies:
Process A = 59% Process

-
Statistical Analysis
t limits from tables + 2
t calculated from data +3

8 = 54%



.571
.673

Date
10/23/70
10/26/70
11/10/70
11/10/70
11/11/70
11/12/70
5.78
6.32
3.82
4.47
4.99
5.97
5.23
EFFLUENT
Process A
Iron Turbidity
3.0 mg/1 58 JTU
4.0 50
14 76
5.0 76
10 68
59
4.3
9.2
4.2
13.2
1.6
11.9
7.6
QUALITY
Process








B
Iron Turbidity
3.0 mg/1
4.0
20
5.0
15
-
57 JTU
52
78
80
80
36
           Was difference significant?  Yes X  No	

           Note:  To convert from lbs/1000 gal to kg/cu m, multiply by 0.120.  To convert from gpm to liters/
                  minute, multiply by 3.785.

-------
                   Table 39.   EFFECT OF DETENTION TIME  - TWO REACTORS  IN SERIES @ 5 GPM
                                V/S ONE REACTOR §  15 GPM
CO
Two reactors in
Process A = Series @ 5 GPM
Limestone
Effluent Usage,
Date pH #/100Q qal.
12/2/70 5.2 8.53
12/3/70 5.2 6.71
12A/70 5.2 6.79
12/16/70 5.U 5.06
Means 5.3 6.77

Process B =

Effluent
PH
5.2
5.2
5.3
5.k
5.3
One reactor
@ 15 GPM
Limestone
Usage,
#/100Q gal.
9.09
9.58
8.73
6.19
a.«



Percent Reduction
in Limestone Usage
6.2
30.0
22.2
18.3
19. k

Typical Utilization Efficiencies:
Process A = 62% Process B = kl%

Statistical Analysis
t limits from tables V 3.182
t calculated from data +3.23


Date
12/2/70
12/3/70
12A/70
Effluent
Process A
Iron Turbidity
8 mg/1 30 JTU
8 18
k kl
12/16/70 8 35
Quality
Process B
Iron Turbidity
12 mg/1 kB jtl
12 43
Ik 56
8 76
           Ulas difference significant?  YES  X  No	

           Note:  To convert  from lbs/1000 gal to kg/cu m, multiply by 0.120.  To convert from gpm to  liters/
                  minute, multiply by 3.785.

-------
                          Table 40.   CHEMICAL ANALYSES FOR DETENTION TIME STUDY
CO
fr-
Sample
pH at
Time of Acidity
Analysis as CaCO,
Cond.
Ca
Mg
Al
Total
Iron
cn AT if
OU j Mils. *
Turb.
pH when
Sample
was Taken
6/2/70 thru 6/19/70
Raw Feed
Pool A
Pool B
2.8
4.9
4.9
380
25
28
1700
1350
1350
130
320
320
39
39
40
34
6.0
6.9
105
5.7
2.8
1050
940
940
-
1
1
-
56
26

5.0
4.9
5/7/70 thru 5/27/70
Raw Feed
Pool A
Pool B
10/23/70
Raw Feed
Pool A.
Pool B
12/2/70
Raw Feed
Pool A
Pool B
2.7
5.0
5.0
thru 11/12/70
2.9
5.0
5.0
thru 12/16/70
2.8
5.0
5.0
440
20
20

380
15
16

520
17
13
1700
1400
1350

1000
870
850

1300
1200
1150
120
330
320

190
450
470

160
430
430
37
38
37

97
85
85

57
54
53
37
9
6.4

205
6.5
8

190
7
12
110
10.0
3.9

67
13
14

72
7
12
1050
950
950

580
540
530

850
660
640
-
8
5

-
1
1

-
2
1
-
37
25

-
65
64

-
31
56

5.1
5.1


5.0
5.1


5.3
5.3

     NOTE:   All units are mg/1 except for pH, conductivity  (micromhos/cm)f and turbidity (JTLJ).
            All values are means from individual data  sets.

-------
                   Table  41.   EFFECT OF REACTOR BAFFLES @ 15 GPM (56.8 LITERS/MIN)
co
Two Wood Baffles
PROCESS A = @ 15 apm

Limestone
Effluent Usage ,
Date pH
7/1/70 4.7
7/2/70 4.9
7/8/70 5.1
Means 4.9
#/1000 qal.
7.63
7.18
8.17
7.66
PROCESS B *» No Baffles - 15 gpm
Limestone
Effluent Usage ,
oH #/1000 aal.
4.8 7.96
5.0 8.32
5.2 9.35
5.0 8.54


Percent Reduction
in Limestone Usage
4.1
13.7
12.6
10.3




Typical Utilization Efficiencies:
Process A = 56%


Statistical
Process B = 52%


Analysis
t limits from Cables + 4.303
t calculated from

data +3.19

Effluent
Process A
Date Iron Turbiditi
7/1/70 2.0 mg/1 35 JTI
7/2/70 2.0 33
7/3/70 2.0 32
Quality
Process

B
/ Iron Turbidity
U 2.5 mg/1
2.5
2.4
50 JTQ
43
33
     Idas difference significant?  Yes 	  No  X
     Note:  To convert from lbs/1000 gal to kg/cu m, multiply by 0.120.

-------
              Table *»2.  COMPARISON OF WOOD AND SCREEN BAFFLES AT 15 GPM (56.8 LITERS/MIN)
co
CF\
Two uiood baf-
Pracess A = fles @ 15 qpm
Limestone
Effluent Usage,
Date pH #71000 qal.
7/21/70 5.1
7/22/70 5.2
7/22/70 5.2
7/23/70 5.2
Means 5.2
Typical Utilization Effici
PTnnoQQ A s R'S'K Ppnppcst?

Statistical Analysis
t limits from tables j£ 3.
t calculated from data -3.
tilnr-i f-( 4 -P -P »-» T* nn r-» m cs T i-in T ^ T r*rarr4"
5.31
5.82
5.82
6.66
5.90
encies:
B = 58%
182
288
9 V=c Y ftln
Two fine mesh screen
Process B = baffles @ 15 qpm
Limestone
Effluent Usage, Percent Reduction
pH #/100Q gal. in Limestone Usaqe
5.2 4
5.2 5
5.0 5
5.0 6
5.1 5

Date Iron
7/21/70 U.5
7/22/70 ^. 2
7/22/70 1.6
7/23/70 2.8
.98 6.2
.V7 6.0
.71 1.9
.02 9.6
.55 5.9
Effluent quality
Process A Process B
Turbidity Iron Turbidity
mg/1 68 JTU **.8 mg/1 65 JTU
28 3.7 28
40 2.0 90
40 2.8 80
     Note:  To convert from lbs/1000  gal  to kg/cu m,  multiply by 0.120.

-------
               Table <»3.  EFFECT ON SCREEN-BAFFLE  MESH SIZE AT 15 GPM (56.8 LITERS/MIN)
CO


Date
B/27/70
8/28/70
Means
Tuo large mesh
Process A = screen baffles
@ 15 gpm
Limestone
Effluent Usage ,
pH #/1000 pal.
4.9 7.87
4.8 8.80
4.9 8.34
Two small mesh
Process B = screen baffles
@ 15 gpm
Limestone
Effluent Usage,
pH #/1000 pal.
5.0 8.38
5.0 10. 41
5.0 9.40

Percent Reduction
in Limestone Usage
6.1
15.5
11.3

Typical
Utilization Efficiencies:
Process A = 60% Process B = 55%
Statistical Analysis
t limits from tables +_ 12.706
t calculated from data + 1.927
Effluent
Process A
Date Iron Turbidity
8/27/70 2.1 mg/1 30 Ji
8/28/70 2.4 35
Quality
Process B
Iron Turbidity
U l.O mg/l 30 JTU
1.2 35
     Was difference significant?  Yes 	  IMo  X*

     *Too feu samples.
     Note:  To convert  from lbs/1000 gal to kg/cu m, multiply by 0.120,

-------
          Table M*.   COMPARISON  OF  THREE  VS  TUO WIRE-MESH BAFFLES AT 15 GPM  (56.8  LITERS/MIN)
oo
CD
Three large mesh
Process A = screen baffles

@ 15 gpm
Two large mesh
Process B = screen baffles

Limestone
Effluent
Usage
Date DH #710.00 gal.
9/16/70 5.0
9/17/70 5.1
9/17/70 5.0
10/7/70 5.0
10/9/7Q 5.0
10/9/70 5.1
Means 5.0
11.27
9.29
9.60
8.64
8.89
8.81
9.42
Effluent
pH
4.9
5.0
4.9
5.1
5.2
5.1
5.0
@ 15 5pm
Limestone
Usage ,
#/1000 qal.
9.22
9.39
9.27
8.00
9.07
8.89
8.97


Percent Reduction
in Limestone Usage
18.2
1.1
3.4
7.4
2.0
0.9
4.8
Typical Utilization Efficiencies^
Process A = 55% Process B

Statistical Analysis
t limits from tables _+ 2
t calculated from data - 1
Was difference significant?

= 56%

.571
.281
Yes 	 No X



Date
9/16/70
9/17/70
9/17/70
10/7/70
10/9/70
10/9/70
Effluent
Process A
Iron Turbidity
27 JTU
5.2mg/l 31
9.2 38
12 52
10 34
9.0 48
Duality
Process B
Iron Turbidity
36 JTU
5.0 mg/1 30
8.1 38
14 58
12 39
10 52
     Note:  To convert from Ibs/lQQQ gal to kg/cu m, multiply by 0.120.

-------
                             Table U5.   CHEMICAL ANALYSES FOR BAFFLE STUDY
oo
vo
pH at
Time of
Sample Analysis
7/1-8/70
Raui Feed
Pool A
Pool B
7/21-23/7D
Raw Feed
Pool A
Pool B
8/27-28/70
Raw Feed
Pool A
Pool B

2.7
4.8
5.0

2.9
4.9
4.8

2.8
5.0
5.6
Acidity
as CaC03

610
k3
41

450
39
50

610
3.5
1.5
Cond.

1870
1400
1450

1400
1100 -
1100

1650
1250
1250
Ca

180
430
440

120
300
300

170
410
430
Mg

56
57
59

40
40
41

57
60
59
Al

42
9.7
9.7

35
4.3
8.4

50
3.8
1.0
Total
Iron

130
2.0
2.5

95
3.2
3.3

120
2.3
1.1
SD4

1000
860
860

780
720
720

1050
1050
1050
Alk.

_
2
2

_
2
2

_
2
2
Turb.

_
33
42

_
44
66

_
32
32
pH when
sample
was taken

2.7
4.9
5.0

2.9
5.2
5.1

2.8
4.9
5.0
9/16-10/9/70
Ran Feed
Pool A
Pool B
2.7
if. 9
5.0
630
26
26 "
1670
1400
1380
230
590
610
78
74
75
59
8.6
8.4
170
9.1
10
1270
910
780
—
2
2
_
38
42
2.7
5.0
5.0
     MOTE:  All units are mg/1 except for pH and conductivity (micromhos/cm).

            All values are means from individual data sets.

-------
                  Table 46.  CHEMICAL ANALYSES FOR 400-MESH LIMESTONE STUDY
pH at
time of
Sample analysis
2/9-12/71
Raw Feed
Pool A
Pool 8
2/24- 26/71
Raw Feed
Pool A
Pool B
3/31-4/3/71
Ram Feed
Pool A
Pool B

2.9
5.1
5.1

2.9
4.9
4.9
- Sludqe
2.9
4.9
5.2
Acidity
as CaCCL

400
19
17

360
19
11
Cond.

1200
900
900

900
700
730
Ca

100
300
330

150
460
450
Mg

33
36
35

50
48
48
Al

28
7.5
5.0

52
7,3
7.8
Total
Iron

130
8.0
9.1

ISO
5.9
4.8
s<\

575
520
540

640
530
510
Alk.

-
10
10

-
1
2
pH when
sample
Turb. was taken

-
61
77

nun
35
42

2.9
5.1
5.2

2.9
5.1
5.0
Recirculation Study
400
9.0
5.2
950
720
730
yb
240
260
28
27
27
34
4.3
4.0
3
4.8
7.3
570
430
450
-
2
4
-
52
69
2.9
5.0
5.1

MOTE:  All units are mg/1 except for ph, conductivity  (micromhos/cm), and  turbidity  (JTU).
       All values are means from individual data sets.

-------
                  Table V7.  CHEMICAL ANALYSES  FOR  LIME VS LIMESTONE STUDY
pH at
time of
Sample analysis
10/19-29/71
Raw Feed
Pool A
Pool B
11/16-21/71
Rau Feed
Pool A
Pool B

2.6
6.5
6.2

2.5
6.5
6.5
Acidity
as CaCD,

720
5.0
9.0

680
4
6
Cond.

1790
1570
1550

1840
1510
1470
Ca

110
300
290

150
420
410
Mg

40
39
30

52
53
43
Al

41
o.a
1.3

48
0.5
1.2
Total
Iron

120
2.1
1.9

130
0.9
1.4
SO^ Alk.

1200
1100 SO
1120 18

1080
1010 74
1030 15
pH when
sample
Turb. was taken

2.6
12 6.4
7 6.4

2.5
12 6.5
5 6.5
NOTE:  All units are mg/1 except for pH,  conductivity (micromhos/cm),  and turbidity (JTU).
       All values are means from individual data sets.

-------
                              TECHNICAL REPORT DATA
                        (Please read Instructions on the reverse before completing)
  REPORT NO.
    EPA-670/2-74-051
2.
                                                  3. RECIPIENT'S ACCESSIOWNO.
 . TITLE AND SUBTITLE

 LIMESTONE AND LIMESTONE-LIME NEUTRALIZATION
 OF  ACID  MINE DRAINAGE
                         5. REPORT DATE
                          June 1974;  Issuing Date
                         6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
                                                  8. PERFORMING ORGANIZATION REPORT NO.
 Roger  C.  Wilmoth
9. PERFORMING ORGANIZATION NAME AND ADDRESS
  Industrial Waste Treatment Research  Laboratory
 U.S.  Environmental Protection Agency
  Crown Mine Drainage  Control Field Site
  Box  555
  Rivesville, West Virginia             	
                         10. PROGRAM ELEMENT NO.
                         1BB040/ROAP  21AFY/TASK  03
                         11. CONTRACT/GRANT NO.
 12. SPONSORING AGENCY NAME AND ADDRESS
 National Environmental  Research Center
 Office of Research  and  Development
 U.S.  Environmental  Protection Agency
 Cincinnati. Ohio  45268
                         13. TYPE OF REPORT AND PERIOD COVERED
                                     Final
                         14. SPONSORING AGENCY CODE
 15. SUPPLEMENTARY NOTES
 16. ABSTRACT
 The  critical parameters  affecting neutralization of ferric-iron acid
 mine waters were  characterized by the U.S.  Environmental  Protection
 Agency in comparative  studies using hydrated lime, rock-dust limestone,
 and  a combination  of  the two as neutralizing agents.   The advantages  and
 disadvantages of  each  of these neutralizing agents were noted.  On the
 ferric-iron test water,  combination limestone-lime treatment provided a
 better than 25-percent reduction in materials cost as  compared to
 straight lime or  limestone treatment.   Significant reduction in sludge
 production was noted  by  the use of rock-dust limestone and by the use of
 combination treatment  as compared to hydrated-lime treatment.  Emphasis
 on optimizing limestone  utilization efficiencies resulted in an increase
 from approximately  35- to 50-percent utilization.  Studies using lime-
 stone that had been ground to pass a 400-mesh screen resulted in utili-
 zation efficiencies near 90 percent.
 7.
                           KEY WORDS AND DOCUMENT ANALYSIS
               DESCRIPTORS
              b.IDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
 *Limestone
 *Calcium hydroxides
 Neutralizing
 *Drainage-mine  (excavations)
 Cost  comparison
 Surface  drainage
              *Acid  mine drainage
              *C)oal  mine drainage
              *Ferric iron
              West Virginia
     13B
 8. DISTRIBUTION STATEMENT
              19. SECURITY CLASS (ThisReport)
                   UNCLASSIFIED
                                                              21. NO. OF PAGES
                                                                   102
        RELEASE TO PUBLIC
              20. SECURITY CLASS (Thispage)
                   UNCLASSIFIED
                                                              22. PRICE
EPA Form 2220-1 (9-73)
                                     92  £. U. S. GOVERNMENT PRINTING OFFICE.-197*-757-5M/5328 Region No. 5-11,

-------