WATER POLLUTION CONTROL RESEARCH SERIES • 14010 DYI 02/71
a
Evaluation of
New Acid Mine Drainage
Treatment Proce;
ENVIRONMENTAL PROTECTION AGENCY • WATER QUALITY OFFICE
-------
WATER POLLUTION CONTROL RESEARCH SERIES
The Water Pollution Control Research Reports describe.
the results and progress in the control and abatement
of pollution in our Nation's waters. They provide a
central source of information on the research, develop-
ment, and demonstration activities in the Environmental
Protection Agency, Water Quality Office, through inhouse
research and grants and contracts with Federal, State,
and local agencies, research institutions, and industrial
organizations.
A tripilicate abstract card sheet is included in the
report to facilitate information retrieval. Space is
provided on the card for the user's accession number
and for additional uniterms.
Inquiries pertaining to Water Pollution Control Research
R.eports should be directed to the Head, Project Reports
System, Office of Research and Development, Environmental
Protection Agency, Water Quality Office, Washington, D. C.
20242.
-------
Evaluation of a
New Acid Mine Drainage
Treatment Process
by
Black, Sivalls & Bryson, Inc.
Applied Technology Division
135 Delta Drive
Pittsburgh, Pennsylvania 15238
for the
ENVIRONMENTAL PROTECTION AGENCY
WATER QUALITY OFFICE
Program No. 14010 DYI
Contract No. 14-12-529
February, 1971
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $1.50
Stock Number 5501-0075
-------
EPA Rev lev/ Notice
This report has been reviewed by the Environmental
Protection Agency and approved for publication.
Approval does not signify that the contents neces-
sarily reflect the views and policies of the Environ
mental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement
or recommendation for use.
-------
ABSTRACT
An economic and engineering evaluation of a submerged coal refuse
combustion process to convert acid mine water (AMW) to potable water
has been made,, In this process coal refuse is burned in molten iron to
supply energy for distillation or reverse osmosis, and the coal refuse
sulfur is trapped in a slag for eventual recovery of sulfur. Laboratory
experimentation was conducted on those areas which could profoundly
affect the process., These areas were: A laboratory demonstration of
slag desulfurization to produce sulfur, the evaluation of slag sulfur
retention characteristics, slag capability for neutralizing AMW and
determination of slag compositions having acceptable fluidities. Lab-
oratory results indicated that sulfur is obtained, high slag sulfur
partition ratios are achieved, fluid slags are produced, and that de-
sulfurized slags are not suitable for neutralization.
Engineering studies show that the process has potential for supplying
inexpensive energy for distillation and permits the recovery of sulfur
so that distilled water is economically produced. Depending upon the
AMW composition and sulfur selling price ($20 to $30/ton) the break-even
price of water for a 5 MM GPD plant varies between $.42 and $.16/1,000
gals when a 14 percent capital interest charge is used.
This report was submitted in fulfillment of Program No. 14010 DYI,
contract No. 14-12-529 under the sponsorship of the Environmental Pro-
tection Agency.
key words: Acid mine water, distillation, slag desulfurization, slag
characterization, submerged coal combustion, two-stage coal
combustion, coal refuse
-------
CONTENTS
Page
Abstract „ i
Conclusions „ xi
Recommendations . . . . xii
Introduction. .... 1
Process Description ..... 3
Process Engineering .... 7
Computer Simulation 7
Important Process Parameters 9
Process Design 12
Economic Evaluation .19
Determination of Equipment Costs 19
Determination of Capital Investment Requirement 26
Determination of Break-even Price of Water 28
Determination of Operating Revenue 29
Break-even Price of Water and Process and Economic Factors ... 29
Acknowledgements 44
References. „ „ 45
Glossary. . . 47
Appendix A - Laboratory Studies 49
Slag Fluidity. . „ 49
Introduction 49
Experimental Procedure - General. .... 49
Standardization of Experimental Techniques 50
Calibration of the Herty Fluidity Equipment 55
Discussion of Results - The Effect of Basicity on Fluidity. . 55
Effect of Calcium Sulfide Content on Fluidity 55
Significance of Fluidity and Apparent Viscosity Measurements. 60
Effect of MgO Additions on Slag Fluidity 61
Effect of Fluorspar on Slag Fluidity 63
Effect of Si02/Al203 Ratio on Fluidity 64
Engineering Design Recommendations 64
Sulfur Retention by Slag 67
Introduction. ° 67
Experimental Procedure. ....... 68
Results and Discussion 69
Engineering Design Specifications 72
Detailed Slag Characterization. . 73
Apparent Density of High Sulfur Bearing Slags 75
Experimental Equipment and Procedure 75
Discussion of Results. 75
11
-------
CONTENTS CONT'D
Page
Viscosity Measurements 77
Introduction 77
Experimental Procedure 77
Discussion of Results 79
Engineering Design Recommendations ... 81
External Surface Area „ . 83
Introduction . 83
Theory . „ . „ 83
Equipment and Experimental Procedure 83
Discussion of Results 84
Total Specific Surface Area 87
Introduction . . . „ 87
Discussion of Results. 87
Crushing Energy Requirements. ....... 89
Introduction 89
Experimental Equipment and Procedure 89
Calibration of Crushing Energy Equipment 91
Discussion of Results 91
Engineering Design Recommendations 94
Heat Capacity .......... 95
Introduction 95
Experimental Equipment and Procedure 95
Heat Capacity. 95
Discussion of Results 96
Engineering Design Recommendations 96
Acid Mine Water Neutralization With Slag 99
Introduction 99
Experimental „ .99
Effectiveness of Slag Alkali for Neutralization 101
Continuous Neutralization Tests. 103
AMW Neutralization 105
Batch Neutralization Studies 107
Engineering Design Recommendations ... .110
Kinetics of Sulfur Recovery from Slag Ill
Introduction Ill
Theory Ill
Experimental Procedure 116
Discussion of Results. 119
Engineering Reactor Design Recommendations 127
iii
-------
CONTENTS CONT'D
Page
Refractory Lining Life. ...... <,<,...« 131
Introduction „ „ . „ 131
Experimental Procedure 131
Results. ...... o 135
Design Recommendations 137
Appendix B The Process Working Area Diagram .139
Appendix C Equipment Cost ........ 145
Appendix D - The Theory of Carbon Solubility Rates 151
Introduction 151
Theory ...... „ 151
Discussion of Theoretical Calculations „ 153
IV
-------
FIGURES
Figure Page
1 Flow Chart Acid Mine Water Treatment Process 5
2 Effect of Heat Rate on Process Operability 14
3 Effect of AMW Concentration on Process Operability. . . .16
4 Coal Handling Complex . ... 0 .... o 20
5 Neutralization Complex. . „ 21
6 Distillation and Waste Heat Boiler Complexes 23
7 Direct Fired Furnace, Steam Turbine-Air Compressor and
Combustor Complexes 24
8 Desulfurization Complex 25
9 Rotary Kiln Dryer 27
10 Effect of Sulfur on Break-even Price of Water 34
11 Effect of Plant Capacity on Capital Investment 35
12 Effect of Plant Capacity on Break-even Price of Water . .38
13 Process Flow Chart of AMW Treatment Plant 39
14 Effect of Water Selling Price on Payback 40
15 Fluidity Test Apparatus „ „ . 51
16 Relation Between Fluidity and Residence Time of
Sample in Furnace . o 53
17 Effect of Temperature on Viscosity of a Synthetic Slag. .56
18 Effect of Temperature on Slag Fluidity 57
19 Relationship of Slag Fluidity and Slag Viscosity 58
20 Effect of Basicity on Apparent Viscosity of Sulfur-Free
Slags . 59
21 The Effect of CaS Addition on Apparent Viscosity 59
-------
FIGURES CONT'D
Figure Page
22 The Effect of Magnesia Content on Apparent Slag
Viscosity ....... . ................ 62
23 The Effect of Fluorspar on Apparent Slag Viscosity ... 62
24 The Effect of Silica to Alumina Weight Ratio on
Apparent Slag Viscosity ................. 65
25 The Effect of Time on the Approach to Partition Ration
Equilibrium ....................... 70
26 The Effect of Contact Time on the Sulfur Content
of Iron ............ . ............ 70
27 The Effect of Slag Basicity on Equilibrium Partition
Ratio. . „ ....................... 71
28 The Effect of Particle Diameter on Apparent Density. . . 76
29 Viscosimeter ..................... .78
30 Effect of Carbon Bob Residence Time in Furnace on
Measured Viscosity ................... 80
31 The Effect of Slag Basicity and Particle Size on
External Specific Surface ................ 85
32 The Effect of Basicity on Total Specific Surface Area. . 88
33 Crushing Energy Apparatus ................ 90
34 Calibration Curve for Aluminum Wire Used in Crushing
Energy Test. «, . . « .................. 92
35 The Effect of Slag Basicity on Grinding Energy
Requirements 0 ...... ............... 93
36 The Effect of Basicity and Temperature on Specific Heat
Capacity ..... ................... 97
37 Continuous Neutralization Apparatus ........ . . .100
38 Effect of Slag Particle Size and Basicity Ratio on Slag
Alkali Utilization in Neutralization .......... 102
39 Effect of Slag Particle Size and SiO /Al 0 Ratio on
the Slag Alkali Utilization Factor . . . ........ 102
VI
-------
FIGURES CONT'D
Figure Page
40 Effect of Operating Parameters on the Neutralization
of Acid Mine Water with Sulfur Free Slag 104
41 The Effect of Throughput on pH of Effluent Leaving
Flow Neutralizing Reactor 106
42 Effect of Slag Particle Size on pH as a Function
of Neutralization Time. 108
43 Variation of pH with Time as a Function of Slag
Basicity. 109
44 Variation of Equilibrium Constant with Temperature. . .113
45 Equilibrium Data for the System, CaS + xlUO + yO = CaO
+ aSO + bH0S + cS0 + dSQ ? . . .114
2 2. / o
46 Stainless-Steel Reactor Desulfurization Equipment . . .117
47 Ceramic-Tube Reactor Desulfurization Equipment 118
48 Temperature Effect on Steam-Slag Reaction 120
49 Effect of Time on Desulfurization, Slag Basicity of
0.82 122
50 Effect of Time on Desulfurization, Slag Basicity of
0.90 123
51 Effect of Time on Desulfurization, Slag Basicity of
1.01 124
52 Effect of Temperature on Specific Reaction Rate . . . .126
53 Correlation Between Actual and Predicated Desul-
furization Results. . 128
54 The Effect of Slag Contact on Refractory Erosion. . . .136
55 Slag Basicity vs. Spent Slag Recycle Fraction of
Various Flux Rates. 140
56 Required Preheated Air Temperature vs. Slag Recycle
Fraction of Various Flux Rates. 141
57 Percent Sulfur in Combustor Slag vs. Slag Recycle
Fraction at Various Flux Rates 143
vii
-------
FIGURES CONT'D
Figure Page
58 Process Working Area Diagram for a 2 MM GPD Plant
using Partially Neutralized Dilute AMW and a Heat
Rate of 2 MM BTUs/1,000 Gals AMW. . „ . „ 144
59 Effect of Particle Radius on Residence Time Needed
in Iron Melt to Dissolve Carbon 154
60 Effect of Particle Radius on Hot Metal Depth
Required to Dissolve Carbon 155
Vlll
-------
TABLES
Table
I
II
III
IV
V
VI
VII
VIII
IX
X
XI
XII
XIII
XIV
XV
XVI
XVII
XVIII
XIX
XX
XXI
Page
AMW Compositions Used in this Study 12
Ultimate Analysis of Coal Refuse (% by Weight) .... 12
Equipment Complex Cost 28
Determination of Break-even Price of Water 30
BEPW for Process Operation Under Various Conditions. . 31
Capital Investment for Various Plant Sizes 36
Stream Capacities and Temperatures for 5 MM GPD Plant. 41
Synthetic Slag Chemical Composition 52
The Effect of Remelting on Slag Fluidity 54
The Effect of Remelting on Slag Fluidity 54
The Variation of Slag Apparent Viscosity with
Basicity and CaS Content
60
Slag Compositions Used in Partition Ratio Studies. . . 68
Slag Compositions Used in Characterization Studies . . 75
Comparison of Surface Areas for Glass Beads Determined
by Air Permeability and Micrometer Methods 83
A Comparison of Crushing Energy Requirements for Silica
and Slags 94
Standard Acid Mine Water Composition 101
Refractories Used in Laboratory Tests 133
Slag Composition Used in Refractory Study 134
Summary of Test Results 134
Computer Results Used to Generate Process Working
Area Diagram 139
Coal Handling Complex Cost 146
IX
-------
TABLES CONT'D
Table Page
XXII Neutralization Complex Cost. . „ 146
XXIII Direct Fired Heater Complex Cost 147
XXIV Combustor Complex Cost 148
XXV Distillation Complex Cost. „ 148
XXVI Desulfurization Complex Cost 149
XXVII Rotary Kiln Dryer Cost 149
-------
CONCLUSIONS
This study has advanced the state of the art for using a two-stage coal
refuse combustion process for the treatment of acid mine water. An
engineering and laboratory study has shown that the process has the
technical capability for converting acid mine water into a potable water
product. An engineering and economic study of the process indicates that
the process can produce distilled water profitably. Some additional
work will be required before the process can be demonstrated on a large
scale.
Specific conclusions derived from this study are:
lo By combusting coal refuse in a molten bath of iron, low cost
energy is available to produce distilled water on a profitable
basis. For a plant processing five million gallons per day of
acid mine water, the break-even price of water will vary
between 16 and 42 cents per thousand gallons depending upon
the price of by-product elemental sulfur and the cost of coal
refuse.
20 The distilled water will find utility as an industrial or
municipal water supply.
3. Using coal refuse as a source of fuel for distillation will
eliminate this source of acid mine water production.
4. Slag formed in the combustor is not suitable for neutralization
of acid mine water. The process flow chart has been modified
to accept limestone neutralization or to distill as-received
acid mine water directly.
5. To maintain the slag and iron contained within the combustor in
a molten condition, a combustion air preheater will be required.
Preheat energy is derived from secondary combustion of the
carbon monoxide rich offgas produced in the combustor.
6. Fluid slags exhibiting better than anticipated sulfur retention
capacities (partition ratios) have been developed.
7. Slag desulfurization to produce elemental sulfur from combustor
slag has been demonstrated. However, data are lacking with
regard to optimum sulfur yield and quality.
8. Suitable commercial refractories have been found that can be
used in the combustor for operation with high sulfur slags.
9. The process has a high degree of flexibility and can readily
accommodate wide variations in acid mine water composition and
flow rate.
-------
RECOMMENDATIONS
Based on the results of this study, it is recommended that:
1. The kinetics of slag desulfurization be evaluated on a larger
scale in the laboratory using both molten and solid slags to
quantitatively determine sulfur yield and quality.
2« The reduction kinetics of calcium sulfate to calcium sulfide in
molten slags which come in contact with molten iron containing
carbon be evaluated in the laboratory using an induction furnace
and synthetic slag.
3o Kinetic data be obtained in the laboratory for carbon solubility
rate while pneumatically injecting coal refuse beneath the
surface of a molten iron bath.
4. Lances be designed and tested in the laboratory for total
immersion in molten iron using water and other liquid media as
cooling agents.
5. An evaluation of the refractories selected in this work be made
in the induction furnace employing calcium sulfate-bearing slags.
6, A new cost estimate be made for this process based on results of
above recommendations.
Xll
-------
INTRODUCTION
This report describes the study of a novel process for the elimination
of acid mine water (AMW) drainage as a source of water pollution. This
process utilizes coal refuse, a source of AMW, as fuel to generate steam
for the conversion of AMW to potable water. Energy for steam generation
to operate evaporators for distillation or to drive pumps for reverse
osmosis, is derived from a two-stage coal refuse combustion process. In
the first stage of combustion, high-sulfur coal refuse or similar low-
cost fuel is dissolved in a molten iron bath. In the second stage of
combustion the fuel carbon is burned with air at the surface of the iron
bath, generating hot carbon monoxide which can be further burned to
release additional heat in a boiler,.
Two-stage combustion makes it possible to use high sulfur bearing fuels
without polluting the air. Fuel sulfur is trapped in the iron from
which it is removed via a lime-bearing slag in the form of calcium
sulfide, without generating sulfur oxides. Sulfur is also recovered from
the reduction of the sulfate content of the acid mine water. Sulfates
contained in the sludge generated by distillation or reverse osmosis
units are dried and added to the combustor as part of the slag. Sulfur
is extracted from the calcium sulfide in the slag by treating the hot
slag with steam and air to recover elemental sulfur.
The recovery of sulfur from the acid mine water and the fuel, coupled
with the utilization of coal refuse as a fuel, provides the economic
incentive for treatment of acid mine water using this process.
Earlier preliminary technical and economic evaluations of the concept
described above showed that it warranted further study. As a first step
towards such a study, the Federal Water Quality Administration funded a
program for a limited bench scale study of process parameters that can
be readily evaluated in the laboratory. The results of the experimental
work were used to arrive at a more reliable technical and economic evalu-
ation of the process.
This report presents the up-dated evaluation of the process, along with
a discussion of the process engineering and cost estimating methods used
in the evaluation. Laboratory work, on which the engineering was based,
is presented as Appendix A of this report. To enable the reader to refer
to any part of the laboratory work, each experimental study is presented
as an autonomous section of Appendix A. In each section, objective of the
work, experimental methods, results, and engineering design recommendations
are presented.
-1-
-------
PROCESS DESCRIPTION
Figure 1 presents a flow chart of the process. The dotted lines on the
flow chart indicate that the acid mine water (AMW) may or may not be
partially neutralized. Partial neutralization will be required for
concentrated AMW to prevent excessive corrosion of the flash distillation
equipment, but for moderately concentrated AMW the process economics are
more attractive without neutralization. If neutralization is required,
AMW is introduced into a neutralizer (1) where it is contacted with finely
divided limestone to partially neutralize the AMW to a pH of three or
more. The limestone used for partial neutralization reduces the amount
of flux introduced into the dryer (4) for use in combustor (5). The
neutralized water which contains suspended solids is pumped to a flash
distillation unit (7) to produce potable water and a concentrated brine
slurry which is subsequently fed to a rotary kiln dryer (4). If acid
mine water is not neutralized, it is fed directly into the distillation
unit.
The rotary kiln dryer serves three functions: 1) to dry the concentrated
brine slurry from the distillation unit, 2) to calcine dolomitic lime-
stone to produce lime and magnesia for use as flux in the combustor,
and 3) to preheat the portion of the desulfurized spent slag from the
desulfurization unit. The contents of the dryer are fed to the combustor
(5) to minimize the quantity of dolomitic limestone required in the
process.
The combustor is a refractory-lined steel vessel that contains molten
iron,, Coal or coal refuse is pneumatically injected beneath the surface
of the iron bath where the carbon is dissolved to free its sulfur for
ultimate reaction with the flux floating on the molten iron surface. Air
is then injected slightly below the surface of the bath and reacts with
carbon to produce a carbon monoxide rich offgas. Heat generated during
the combustion of the coal provides the necessary heat of reaction to
reduce calcium sulfate contained in the dryer solids to calcium sulfide.
In addition, the combustor provides the energy required to produce iron
from iron compounds contained in the dryer solids and pyrites contained
in the coal,, Molten elemental iron is continuously removed from the
combustor. Slag containing calcium oxide, magnesium oxide, ash and
calcium sulfide is continuously removed from the combustor and sent to
the slag desulfurization unit (8) where it is contacted with steam and
air to produce a sulfur-rich gas. Elemental sulfur is condensed out of
this gas and sent to storage.
Desulfurized spent slag exiting the desulfurization unit is divided into
two streams which proceed to the dryer, and to a spent slag storage pile.
Spent slag consists of a dry mixture of silica, alumina, magnesium
hydroxide and calcium hydroxide.
Carbon monoxide rich offgas generated in the combustor is used to supply
energy for operation of auxiliary equipment. A large fraction of the
-3-
-------
combustor offgas is sent to the waste heat boiler (10) which provides
high pressure steam for the steam turbine-air compressors (15) and the
exiting low pressure steam for the flash distillation unit. Steam
generated in the waste heat boiler undergoes a pressure reduction through
the steam-turbine air compressors before entering the distillation unit.
In the discussion which follows, steam from the waste heat boiler is
assumed to enter the distillation unit directly. Steam turbine air
compressors are used to generate pressurized air for combustion and coal
pneumatic conveying. Combustor offgas is also used to provide the energy
requirements for air preheating (13), for drying and calcining the dryer
contents, and drying the incoming coal (14).
-4-
-------
i
Ul
Low Pressure Steam
Air
15
i Steam Turbine
Air Compressors
High Pressure
Steam
10
Boiler
Wet
Coal
14
Coal Dryer ;
13
i Air
Preheater
Coal
Air
Heated Air
Water
Kiln Dryer
Offgas
Combustor
Sulfur
Bearing Slag
T
Iron
Acid Mine
Water (AMW)
Limestone
Neutralization
Partially Neutralized
AMW
Distillation
Brine
-Flux
8
Desulfurization
Steam &
Air
Sulfur
FIGURE 1 - FLOW CHART ACID MINE WATER TREATMENT PROCESS
-Slag
-------
PROCESS ENGINEERING
Computer Simulation
To simulate the AMW Treatment Process, a computer program of the energy
and material balance equations for the process was prepared. The digital
computer simulation permits all process parameters to be varied and yields
quantities and temperatures of all process streams as output. A brief
description of the unit operations involved in the process and the
assumptions in their operation is now presented.
An important consideration in operation of the process is the composition
and concentration of the AMW. In this study it is assumed that the AMW
composition and concentration does not change with time. In reality the
acid mine water composition will vary with time. Therefore, values used
in the process simulation should be considered as yearly averages.
Partial neutralization of AMW is accomplished by using finely divided
limestone. Quantity of limestone required for neutralization depends
upon concentration of AMW and effectiveness of limestone in neutralization.
In the process it was assumed that the limestone is 80 percent efficient
in utilizing its lime content to produce a pH of about three in the
partially neutralized water. The partially neutralized water containing
suspended solids proceeds to the distillation unit.
As previously stated, if the AMW is moderately acidic it can be used
directly in the flash distillation unit. By doing this, the partial
neutralization unit operation is eliminated and the AMW proceeds to the
distillation unit. Distillation is accomplished by using conventional
flash distillation equipment in which the steam requirement for the
evaporation of acid mine water is supplied by a waste heat boiler. Com-
bustor offgas serves as the energy source to generate steam in the waste
heat boiler. The waste heat boiler is a standard item of equipment in
which the carbon monoxide rich offgas from the combustor is reacted with
air to produce carbon dioxide. The heat of combustion of carbon monoxide
to carbon dioxide and the sensible heat of the incoming combustor offgas
supply the energy to convert water to steam for use in the flash distil-
lation evaporators. The combustor offgas entering the boiler is assumed
to undergo a ten percent loss in temperature in transit from the combustor.
The waste heat boiler is assumed to operate at an efficiency of 90 percent
with a flue gas leaving at 280°F. The temperature of the combustor offgas
entering the waste heat boiler (and all other auxiliary equipment) will
depend upon the combustor operating temperature. In this simulation, the
combustor is assumed to operate at 2700°F.
The combustor offgas is used to supply energy for the rotary kiln dryer,
combustion air preheater and coal dryer in addition to the waste heat
boiler. In the simulated process, the combustor offgas requirements for
all auxiliary equipment except the waste heat boiler are determined first.
The remaining combustor offgas is then used in the waste heat boiler. The
reason for this is that the energy requirements for the auxiliary functions
-7-
-------
are fixed by the quantities of the process streams, however, some
latitude is possible in the design of a flash distillation unit to use
more or less steam. In a flash distillation plant, the heat transfer
surface area required to evaporate water is related to the economy
factor (defined as the pounds of water distilled per pound of steam
used). Consequently, within specified design limits, the heat transfer
area of the distillation plant can be made to accommodate available
steam. To minimize capital cost by minimizing required evaporator heat
transfer area, all of the remaining combustor offgas (after all auxiliary
functions have been satisfied1) is used in the boiler to generate steam.
Potable water and a concentrated brine slurry at 180°F are produced in
the distillation unit. Concentrated (60 percent water by weight) brine
slurry, containing all of the acid mine water constituents entering the
distillation unit, proceeds to a rotary kiln dryer where it is dewatered.
The rotary kiln dryer is also used to calcine the dolomitic limestone.
Calcination converts the dolomite to lime and magnesia which are
required as fluxing agents in the combustor. In this study., dolomitic
limestone was used as the flux. However, a calcitic limestone or a com-
bination of calcitic and dolomitic limestones could have been employed
to obtain any desired ratio of lime to magnesia in the combustor slag.
The dolomitic limestone was assumed to contain 60 percent calcium car-
bonate and 40 percent magnesium carbonate by weight. Also added to the
dryer is a recycle stream of desulfurized spent slag. The stream enters
the dryer to be heated to 2200°F for use in the combustor, because reuse
of the lime content of the spent (desulfurized) slag decreases the com-
bustor flux costs. In essence, then, combustor offgas is fired directly
in the rotary kiln to supply hot gases for heat of vaporization to dry
the concentrated brine slurry, the heat of reaction for calcination, and
the sensible heat necessary to raise the solids to a temperature of
2200°F.
The dry solids from the rotary kiln dryer are fed to the combustor,
which is assumed to operate at 2700°F and the air preheat temperature
is adjusted to maintain this temperature. In the combustor, coal refuse
is pneumatically injected beneath the molten iron bath where its
sulfur and carbon are dissolved in the molten iron. The volatile matter
of the coal is cracked to carbon monoxide and hydrogen. Air is also
added to the combustor slightly below the molten iron surface to com-
bust the carbon. Combustion of coal serves to supply the necessary heat
required to reduce the calcium sulfate contained in the kiln dryer
solids to calcium sulfide in the slag layer floating on the molten iron.
The ash content of the coal is also transferred to the slag layer.
Sulfur present in the coal dissolves in the molten iron and reacts
with the lime contained in the slag to produce calcium sulfide. Iron
contained in the coal as pyrities is reduced to elemental iron. In com-
bustor operation iron (containing one percent sulfur) and slag rich in
calcium sulfide are continuously removed from the unit. Even though
laboratory data show that the iron will contain less than 0.3 percent S,
iron sulfur content was conservatively established at one percent. A
complete energy balance encompassing all reactions occurring in the com-
bustor at a temperature of 2700°F, with kiln solids added at 2200°F and
coal pneumatically injected at ambient temperature is used
-------
to determine the air preheat temperature necessary to maintain the re-
quired temperature in the combustor,, Carbon monoxide rich combustor
offgas is continually removed from the combustor and, as previously
stated, serves as the energy source for the auxiliary equipment.
An important variable in the operation of the combustor is the basicity
of the slag. Basicity (defined as the weight percent ratio, CaO +
MgO/Al-O., + SiO~) is related to the operational characteristics of the
slag. It has been found experimentally that basicity should be in the
range of 008 to 1.2 to produce a sufficiently fluid slag which can be
easily hand led„ Of equal importance in producing suitable slags for the
combustor operation is the sulfur content of the slag. There is a
maximum sulfur content in slag above which the sulfur causes the flu-
idity of the slag to decrease to inoperable levels. Laboratory exper-
imentation has indicated that a maximum of ten weight percent sulfur in
the slag is usable.
Combustor offgas, which is rich in carbon monoxide, is combusted to car-
bon dioxide in a direct-fired heat exchanger to preheat the air entering
the combustor. Air heater heat transfer surface and materials of con-
struction are adjusted according to the air preheat temperature re-
quirements. The resulting flue gas, which is assumed to exit at a
temperature of 500°F, is used together with additional combustor offgas
in a coal dryer. In the coal dryer, coal surface moisture is removed
to facilitate pneumatic injection into the combustor. Wet coal enters
the coal dryer at ambient temperature and flue gases leave at 280°F.
Calcium sulfide bearing slag from the combustor is sent to the desul-
furization unit where it is contacted with steam and air at 2000°F to
produce a sulfur rich gas. The sulfur is condensed and sent to storage.
It is anticipated that slag from the combustor will undergo a heat loss
and enter the desulfurization reactor at approximately 2300°F where it
will be contacted with steam and air which will result in the following
overall net reaction at 2000°F.
H20
CaS + %02 (air) = CaO + S
Desulfurized slag leaving the desulfurization unit contains lime,
magnesia, silica, and alumina. A portion of the desulfurized slag is
recirculated to the dryer to utilize its valuable lime content as a
fluxing agent in the combustor. The remaining slag is sent to storage for
sale,
Important Process Parameters
The important process parameters and alternatives are AMW composition,
coal composition (in particular its sulfur content), coal rate into the
combustor, quantity of dolomitic limestone added to the process, basicity
and sulfur content of combustor slag, economy factor of the distillation
unit, and air preheat temperature required to maintain the combustor at
2700°F0 Design and economic significance of these parameters will now
be discussed.
-9-
-------
The decision to partially neutralize the AMW depends upon its composition.
The AMW compositions used in this study do not require neutralization.
However, very acidic acid mine water will require partial neutralization
to prevent excessive corrosion of expensive distillation plant materials
of construction.
The AMW composition has a direct effect upon the basicity and sulfur
content of the corabustor slag and governs quantity of flux which must
be used in the process and amount of spent slag recycled to the dryer to
maintain the requisite slag basicity„
The higher the sulfate content, the more energy is required to convert
calcium sulfate to calcium sulfide in the combustor. Consequently,
additional coal must be used to supply additional energy requirements
for the sulfate to sulfide reaction, increased quantity of solids requir-
ing drying in the rotary kiln dryer, and increased energy requirements
for air preheating and coal handling. Thus the AMW composition has an
equally profound effect on process design and process economics which
will be shown later.
The coal refuse composition and heating value is an equally important
process variable. The heating value of the coal refuse determines the
quantity of coal required in the combustor to satisfy the energy
requirements of the entire process. The sulfur content of the coal refuse
has a profound effect upon the composition of the combustor slag.
Laboratory studies indicate the sulfur content of a usable fluid slag
to be ten percent or less by weight, therefore, more flux and recycled
spent slag will be required to maintain a sulfur content in the slag
below ten percent. In addition, the sulfur content of the coal refuse
has a profound effect upon the economics of the process because of by-
product credits, and will be discussed later. Coal refuse also contains
the ash constituents alumina and silica. The quantity of flux required
to maintain the basicity of the combustor slag in the range of 0.8 to 1.2,
depends in part on the quantity of the ash constituents in the coal refuse.
The quantity of coal refuse added to the combustor is directly related
to the quantity of combustor offgas produced. In the discussion which
follows, coal consumption will be measured by heat rate, which is defined
as millions of BTU's (coal refuse) consumed per thousand gallons of
AMW processed.
The combustor offgas is used to supply the energy requirements of the
rotary kiln dryer, the air preheater, the coal dryer, and the waste heat
boiler. As previously stated, the flash distillation unit can be designed
to accommodate various economy factors (defined as the pounds of water
distilled per pound of steam used). For conventional flash distillation
installations the economy factor varies between a ratio of five to ten.
Consequently, as long as there is sufficient combustor offgas available
to the waste heat boiler to generate an economy factor of ten, the process
-10-
-------
is operable. This also means that the process is operable over a
range of heat rates provided the economy factor is adjusted accordingly.
Operation at a lower economy factor requires less heat transfer surface
with a correspondingly lower capital investment cost. This can be
economically attractive, provided the added operating costs associated
with using larger quantities of coal do not offset the benefits of the
lower capital investment,,
As a process design consideration, the heat rate establishes the economy
factor in the distillation unit, the air preheat temperature required
to maintain the combustor at 2700°F (as more coal is combusted per
gallon of AMW, a lower air temperature is required), and the heat
requirements for coal drying which are directly related to the quantity
of coal used in the combustor. Increasing the heat rate adds more ash
to the combustor slag and increases the flux requirement to maintain
an operable slag basicity in the combustor.
Basicity of the combustor slag is affected by AMW composition, the
quantity and composition of the coal refuse and the quantity of flux
added to the combustor. Because fluid slags are obtained when the slag
basicity is in the range of 0.8 to 1.2 and contains less than ten
percent sulfur, the process is considered operable when the various
selected process parameters yield a combustor slag that meets these
criteria.
The desulfurized slag recycled to the rotary kiln dryer for eventual use
as a flux in the combustor is a variable which is inter-related with all
other parameters. The primary purpose of recycling spent slag is to
minimize the quantity of limestone required in the process. Desulfurized
slag is normally sent to storage for sale, however, a fraction of this
slag is directed to the rotary kiln dryer and is called the slag recycle
fraction. The slag recycle fraction affects the size of the kiln and
the composition of the combustor slag. The flux rate or the quantity of
dolomitic limestone added to the process affects the kiln operation and
the composition of the combustor slag. The dolomitic limestone entering
the kiln is calcined to provide lime and magnesia for the combustor oper-
ation0 Its primary purpose is to flux the ash constituent of the coal and
to provide a source of calcium to react with sulfur to form calcium
sulfide. The quantity of flux required is determined by combustor slag
basicity and is inter-related with the slag recycle fraction. Increased
flux requirements of the process increase the energy requirements of the
kiln to perform the calcination reaction. Higher energy requirements for
the kiln demand that more combustor offgas be produced which results in
a higher heat rate for the process.
In summary, the important process parameters are the acid mine water
composition, composition of the coal refuse, heat rate, desulfurized slag
recycle fraction, flux rate, slag basicity and percent sulfur in the
slag. Quantitative effects of these variables on process design and
economics will be discussed in the following sections.
-11-
-------
Process Design
The interrelationship and the operable ranges of the important process
variables and their effect on process design will be discussed in this
section. In this study, two concentrations of acid mine water are
considered. The process design and economic analysis were based on
dilute and moderately concentrated AMW of the compositions shown in
Table I,
TABLE I
AMW Compositions Used in this Study
Dilute ppm Moderately Concentrated, ppm
Acidity (as ppm CaCOo) 400 1200
Sulfate 1061 3183
Total Iron 200 600
Calcium (as Ca) 80 240
Aluminum (as Al) 5 15
Magnesium (as Mg) 24 72
The dilute acid mine water composition was suggested by the Environmental
Protection Agency as an average composition of all AMW generated.
Moderately concentrated acid mine water was selected to show the effect
of acid mine water concentration on process design and economics. A
coal refuse with a heating value of 6,000 BTUs per pound and the com-
position shown in Table II was selected as representative of a high-
sulfur coal refuse.
TABLE II
Ultimate Analysis of Coal Refuse
(% by weight)
Carbon 40.6
Hydrogen 2.9
Oxygen 3.7
Nitrogen .7
Sulfur 10.0
Moisture 3.0
Ash 39.3
Sulfur content of this coal refuse, in the form of organic sulfur and
pyrites, is ten percent by weight. Once the compositions of the AMW
and coal refuse are established, the remaining variables to be studied
are the heat rate, spent slag recycle fraction, air preheat temperature,
flux rate, slag basicity, and slag sulfur content. In regard to neutral-
ization alternative, this study will evaluate the effects of utilizing
partially neutralized and unneutralized AMW.
The energy and material balance computer program was used to generate
quantities and temperatures of all process streams as functions of
above variables. The operable range of the process is subject to the
following constraints: 0.8 to 1.2 basicity, ten percent or less slag
-12-
-------
sulfur content, and 2000°F maximum air preheat temperatures based on
the cost of materials of construction. These constraints are best
illustrated using a Process Working Area Diagram (PWAD). The PWAD is a
plot of slag basicity vs. the slag recycle fraction at various flux rates
into the kiln. A particular PWAD pertains to one selected heat rate,
acid mine water composition, and coal refuse composition. Located on
this diagram are the constraints of basicity, sulfur content in the slag,
and air preheat temperature.
The PWAD is generated by running the energy and material balance computer
program over a wide range of process parameters. For a selected heat
rate, AMW concentration, and coal refuse composition, the slag basicity
is generated as a function of the slag recycle fraction and the flux
rate into the combustor. These data are plotted as the slag basicity vs.
slag recycle fraction at various flux rates. A plot is then made of the
sulfur content in the slag as a function of slag recycle fraction at
various flux rates. A line is drawn through this plot at the ten percent
sulfur content level to pick out coordinates of the flux rate and the slag
recycle fraction which yield a ten percent sulfur content in the slag.
These points of ten percent sulfur content are then plotted on the PWAD
to show the constraint of sulfur composition in the slag. In the same
manner, slag recycle fraction--flux rate coordinates which yield a
2000°F air preheat temperature are determined and plotted on the PWAD.
Two additional constraints are shown on the process working area
diagram which are a minimum 0.8 and a maximum 1.2 slag basicity. In
this manner the process working area (PWA) is established for which any
point within the PWA yields an operable process. The technique used to
generate a typical process working area diagram is explained in detail in
Appendix B.
The PWAD's shown in Figures 2 and 3 were prepared to illustrate the effect
of heat rate and AMW composition on the PWA for a two million gallon/day
plant, using as fuel a coal refuse having a heating value of 6,000 BTU/lb
and a sulfur content of ten percent. Figure 2 pertains to treatment of
neutralized AMW, initially of the dilute composition shown in Table I.
Figure 2a is based on a heat rate of two million BTU/1,000 gallons of
AMW, while Figure 2b is based on a heat of 3.25 million BTU/1,000 gallons
of AMW. The PWAD's are constrained at the top and bottom by 1.2 and 0.8
basicity limits respectively, on the left by the maximum ten percent
sulfur in the slag and on the right by the maximum air preheat temperature
of 2000°F. Any point selected within this area will yield an operable
process. For example, in Figure 2a, if the point of intersection of the
minimum basicity and maximum sulfur content is selected, the process
operating parameters would be a flux rate of 230 tons per day and a slag
recycle fraction of 0.26, (fraction of desulfurized slag that is reused
in combustor). The same intersection in Figure 2b corresponds to a
flux rate of 400 tons per day at a slag recycle fraction of 0.22. At
the higher heat rate more flux is required because more coal is used.
This results in more ash in the slag which must be offset by additional
flux to obtain a basicity of 0.8. Also, more sulfur from the coal enters
the process, therefore, more flux is required to dilute the sulfur in the
slag to a level of ten percent.
-13-
-------
0}
TO
00
Maximum Basicity
Maximum Air
Temperature, 2000°F
\Minimum Basicity
Maximum
10% Sulfur
JL
J L
J
0 .1 .2 .3 .4 .5 .6 .7
Slag Recycle Fraction
.9 1.0
CNl
Figure 2 - a Conditions: Heat Rate 2 MM BTU/1,000 gal.
Partially Neutralized Dilute AMW
\ Maximum Basicity
en
to
PQ
oo
Maximum Air
Temperature, 2000°F
Maximum
. 10% Sulfur
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0
Slag Recycle Fraction
Figure 2 - b Conditions: Heat Rate 3.25 MM BTU/1,000 gal
Partially Neutralized Dilute AMW
FIGURE 2 - EFFECT OF HEAT RATE ON PROCESS OPERABILITY
14-
-------
Figures 2a and 2b which differ only in the heat rate represented, show
that the PWA is larger at the higher heat rate. Since more coal energy
is used at the higher heat rate, the required air preheat temperature
is lower at a given slag recycle fraction. Consequently, the maximum air
preheat temperature of 2000°F occurs at higher values of the slag
recycle fraction.
Figure 3 is similar to Figure 2, except that it pertains to treatment of
unneutralized AMW, of both the dilute and concentrated compositions shown
in Table I. Figures 2b and 3a will be used to illustrate the effect of
partial neutralization on the PWA. These figures were generated using a
heat rate of 3.25 and a dilute acid mine water concentration. There is
no significant difference in the PWA when partially neutralized or
unneutralized acid mine water is used. This results from the fact that
the limestone used to partially neutralized the AMW is eventually
deposited and recovered in the combustor. Some small changes do exist
because of redistribution of the heat requirements in the various unit
operations but they are of no practical consequence.
To compare the effect of acid mine water concentration on the PWA, Figure
3b was prepared using a heat rate of 3.25 and the moderately concentrated
acid mine water. Comparison of Figures 3b and 3a showed that the PWA is
considerably smaller when a more concentrated acid mine water is used.
This reduction in the PWA is due to the maximum air preheat temperature
occurring at lower values of the slag recycle fraction. This is expected
since the higher concentration acid mine water requires more energy in
the combustor to convert the sulfates to sulfides. If the coal rate is
held constant, then this additional energy must come from using higher
air preheat temperatures. Consequently, under the same conditions of
flux rate and slag recycle fraction the required air preheat temperature
will be higher for a more concentrated acid mine water and the maximum
air preheat constraint line will occur at lower values of the spent slag
recycle fraction. Using the minimum basicity-maximum sulfur content
point for comparison it is seen that for the moderately concentrated acid
mine water, 410 tons per day of flux is required at a slag recycle fraction
of 0.3 Both the dilute and moderately concentrated acid mine water
process operations, require about the same quantity of flux. This is due
to the effect of recycling more slag to the combustor for the more
concentrated acid mine water (0.3 as compared to 0.2). The reason for
recycling spent slag is to reduce the quantity of limestone required by
the process.
In some situations it is possible that the maximum air preheat temperature
constraint line will occur to the left of the maximum sulfur content
line. This means that at the chosen heat rate the process is inoperable--
there is insufficient energy to run the process. Making the process
operable will require a higher heat rate. Similarly there may be
situations where the maximum sulfur content constraint line will not
appear on the PWA. For example, if a low sulfur coal refuse and an
extremely dilute acid mine water were used, the slag composition within
the basicity range of 0.8 to 1.2 will not yield a slag of more than ten
-15-
-------
o
•H
CO
CO
cs
ON
CO
Maximum Basicity
Maximum \ Minimum Basicity
10% Sulfur
0 .1 ,2
Conditions:
Heat Rate 3.25 MM BTU/
1,000 Gal. Dilute AMW
Maximum Air
Temperature, 2000°F
.4 .5 .6 .7 .8
Slag Recycle Fraction
Figure 3 - a
.9 1.0
co
CO
CQ 00
v Maximum Basicity
\ Minimum
Maximum Basicity
10% Sulfur
Conditions:
Heat Rate 3.25 MM BTU/1,000 Gal.
Maximum Air
Temperature, 2000°F
0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0
Slag Recycle Fraction
Figure 3 - b
FIGURE 3 - EFFECT OF AMW CONCENTRATION ON PROCESS OPERABILITY
-16-
-------
percent sulfur. Under these conditions the heat rate should be decreased
to yield a ten percent sulfur bearing slag provided there is sufficient
energy to perform the distillation without using an economy factor
higher than ten. Using economy factors higher than ten would result in
excessive distillation capital investment costs.
It is of interest to note that the combustor can accommodate a wide
range of variations in AMW concentrations and is evidenced by the rather
broad PWA's shown on Figures 3a and 3b. Reasonably wide variations in
AMW concentration can be accommodated rather easily by simply varying
flux and/or slag recycle rates. These simple expedients provide sub-
stantial versatility to the process.
In summary, the interrelationship of heat rate, neutralization, slag
recycle fraction, and acid mine water concentration on the process
working area has been shown. Having established a means of determining
operable ranges of the various process parameters, attention is now
directed to determining the economic optimum point within the process
working area.
-17-
-------
ECONOMIC EVALUATION
Having established a technique (PWAD) to determine the operable ranges
of the process parameters, the next logical step is to determine the
point of operation within the process working area which maximizes
profit. The profit or loss for the process will be a function of the
process design parameters discussed previously, and process economic
considerations such as plant capacity, capital investment requirements,
raw material costs, product selling prices, capital interest charge,
plant maintenance and labor requirements„
Determination of Equipment Costs
For convenience in determining the capital investment, the principal
items of equipment have been grouped into nine complexes:
(1) Coal handling-for drying and pneumatic conveying of
coal into the combustor.
(2) Neutralization of AMW.
(3) Steam turbine driven air compressors-to convey the coal
and to supply compressed combustion air into the combustor.
(4) Direct fired heater-to preheat the air for the combustor.
(5) Waste heat boiler-to generate steam from combustor offgas
for use in the distillation and steam turbine complex.
(6) Rotary kiln dryer-to calcine flux and dry the concentrated
brine slurry from distillation.
(7) Combustor.
(8) Desulfurization-to recover sulfur from combustor slag.
(9) Distillation.
Figure 4 shows the major pieces of equipment comprising the Coal Handling
Complex. Wet coal is pneumatically conveyed from a storage pile to a
rotary dryer. Combustor offgas and offgas from the direct fired furnace
serve as the energy source for drying. Dried coal is crushed in a
hammermill and sent to a temporary storage tank. Coal from the storage
tank is fed to a pneumatic feeder tank where air from the steam turbine-
air compressors is used to transport the crushed coal into the combustor.
The Neutralization Complex is shown in Figure 5. A stainless steel
pump is used to pump the acid mine water from its source to a neutral-
ization tank. Limestone from a storage pile is pneumatically conveyed
into the neutralization tank, which is a closed vessel containing
agitators. Partially neutralized acid mine water plus suspended solids
are pumped to the Distillation Complex.
19-
-------
o
I
Coal to
- COAL r/r^Lr
-------
Acic< ''inr I.'ate.r
^artial ly
TO Ms till at; ion
Ueutrr. 1 n nation
m-ir 5 - KETlrTT/LI7ATTON
-21-
-------
Figure 6 shows the Distillation and Waste Heat Boiler Complexes. The
Distillation Complex is a standard flash distillation scheme. Using a
heat exchanger, the partially neutralized acid mine water is preheated
with product steam from an evaporator and then proceeds through a series
of evaporators where it is further heated by the steam produced in
each evaporator. The preheated acid mine water then enters a heat
exchanger where it is brought to its "flash" point. Low pressure steam
exiting from the steam turbine-air compressors is flashed in the
evaporators to produce steam. Concentrated brine slurry exiting the
last evaporator is fed to the rotary kiln dryer. The steam product from
the last evaporator is condensed as product water in the heat exchanger
used to preheat the neutralized acid mine water. The water and uncondensed
steam leaving the heat exchanger which brings the neutralized acid mine
water to the flash point is recycled to the waste heat boiler. The waste
heat boiler burns combustor offgas to generate high pressure steam
from the condensed steam exiting the Distillation Complex. The high
pressure steam is then used in the steam turbine air compressors to gen-
erate compressed air for pneumatically conveying coal into the combustor
and air combustion in the combustor.
The Direct Fired Heater, Steam Turbine Air Compressor and Combustor
Complexes are shown in Figure 7o Two air compressors are used — each
requiring high pressure steam from the waste heat boiler. One compressor
is used to generate compressed air (eight through ten psig) for
pneumatically conveying coal into the combustor. The second air com-
pressor generates compressed air (four through five psig) which is used
to combust the coal in the combustor.. The low pressure steam exiting
at each compressor is collected and directed to the Distillation Complex
where it is used to bring the partially neutralized acid mine water to
the flash point. The compressed combustion air is fed to the direct
fired heater for preheating prior to admittance into the combustor. The
direct fired heater burns combustor offgas to supply energy for
preheating.
The steel combustor is refractory lined and contains lances for admitting
coal and combustion air into the vessel. The dried coal is pneumatically
conveyed from the coal handling complex and the preheated combustion air
is produced in the direct-fired heater. Dry solids from the rotary
kiln dryer are metered and fed into the combustor using a gas tight star
valve. The combustor offgas produced in the combustion of coal is
collected and sent to various other equipment complexes. Slag is
continually removed from the combustor and proceeds to the desulfurization
complex. Iron is also continually removed from the combustor, granulated,
and sent to storage.
One possible desulfurization scheme is shown in Figure 8. The slag exits
the combustor via an enclosed conveyor and is immediately granulated by
a water spray, and is brought to a crusher. The crushed slag is then fed
to a refractory lined steel reaction vessel where it is contacted with
water and air to produce a sulfur rich offgas. The sulfur rich offgas is
-22-
-------
Product
rater to Vnsto I'cater Poilpr
^artially
Neutralized-*.
ATI-.'
Concentrated
Priiic Slurry
To rotary J'iln
Prvcr
Evaporators
Ptear ^
an "urMries
TOT'.; nF*-or_
nf f f>.TR
ft xr
_J
FirrPE 6 - PIFTIT.LATION AN? r/F-TF I'l'./T FPTLfr
-------
Direct Fired
Furnace
\J
Dry Coal
Air
Flue Gas
To Coal Dryer
_L
Air
To Kiln
To Coal
Dryer
Combustor
Offgas
To Waste
Heat Boiler
Compressed Air
(4-5 psig)
Dry Solids From
Rotary Kiln Dryer
Air
Steam Turbine
Air Compressors
High Pressure
Steam From
Waste Heat Boiler
Low Pressure Steam
to Distillation
\ Slag to
\Desulfurization
High Pressure
Steam
Compressed
Air (8 psig)
To Coal Feeder
Low Pressure
Steam to
Distillation
Iron
Granulator
Granulated
Iron
FIGURE 7 - DIRECT FIRED FURNACE, STEAM TURBINE-AIR COMPRESSOR AND COMBUSTOR COMPLEXES
-------
I
S3
Ol
Cor>buPtor
SI a 7
rCT'T T"TTT'7ArT"Tr" pn1-IT r1"
-------
sent to a condenser where sulfur is condensed, using water as the
cooling medium, and collected into a storage tank. The resulting con-
denser offgas is recycled for additional reaction in the desulfurizer.
Part of the desulfurized slag leaving the reaction vessel is conveyed,
via a belt conveyor, to the enclosed conveyor, to coat and protect the
bucket belt from the molten slag leaving the combustor. The remaining
slag is sent to storage, and to the rotary kiln dryer.
The rotary kiln dryer is shown in Figure 9. The dryer receives slag
from the Desulfurization Complex, concentrated brine slurry from
distillation, and flux from storage. The slag and flux are conveyed in-
to the dryer while the brine slurry is pumped in. Not shown in the
figure is the combustion equipment necessary to burn the combustor off-
gas to supply energy requirements of the dryer.
To facilitate the determination of the capital investment requirement
for various plant sizes and operating conditions, the equipment cost
associated with each of the equipment complexes was determined at one
set of operating conditions„ The equipment complexes were sized for
their respective capacity requirements at this set of operating con-
ditions. Having established the cost of an equipment complex at a
particular capacity, this cost was conveniently scaled up or down for
different capacity requirements resulting from changes in plant size and
operating conditions„ Table III presents current (mid 1970) costs of
the equipment complexes and capacity basis for these costs. Costs
presented in Table III are purchased equipment costs at the factory except
for the distillation complex which is on an installed turn-key basis.
The cost data were generated for a two million gallon/day plant utilizing
dilute partially neutralized acid mine water, ten percent sulfur coal
refuse with a heating value of 6,000 BTUs/lb. 230 tons per day flux, a
slag recycle fraction of 0.4, a heat rate of 2 MM BTUs/1,000 gallon AMW,
and a combustor slag basicity of 0.8. A complete breakdown of costs of
the equipment associated with each equipment complex is presented in
Appendix C.
Determination of Capital Investment Requirement
To determine capital investment requirement for a given plant capacity
and set of operating parameters, the process working area must be
established for a given AMW concentration, coal refuse composition, and
heat rate. An operable point is selected in the process working area
and the energy and material balance computer program is run to establish
the stream capacities associated with each of the equipment complexes.
Once these stream capacities are known, the equipment cost data presented
in Table III can be scaled.
The following procedure was used to determine the fixed capital invest-
ment requirement. Costs associated with the various equipment complexes,
excepting distillation, are assumed to yield a total purchase equipment
cost. Installation cost for this equipment was assumed to be 40 percent
-26-
-------
I
ro
Concentrate?^ Trine .c
P filiation
Dry Colic's to ^or^^u^;tor
-------
Complex
Coal Handling
Neutralization
Steam turbine Air
Compressors
Direct fired furnace
Waste Heat Boiler
Combustor
TABLE III
Equipment Complex Costs
Cost*
91,300***
49,700
28,500
33,000
38,000
42,700
61,700
98,000
108,000
90,000
46,600
Capacity Basis
333.3 tons/day coal refuse
2 MM GPD AMW
Distillation, neutralized
Acid mine water
Distillation, neutralized
Acid mine water
Distillation, dilute
Acid mine water
Distillation, moderately
Concentrated acid mine
water
Desulfurization
Rotary kiln dryer
2,500,000***
2,000,000***
2,500,000***
2,800,000***
73,000
315,000
333.3 tons/day coal refuse
C/S**, 700°F
C/S, 1000°F
C/S, 1300°F
C/M**, 1300°F
C/M, 1700°F
C/M, 2000°F
2 MM GPD AMW
805 ton /day combustor
offgas
Economy factor 10
Economy factor 5
Economy factor 10
Economy factor 10
406 tons/day slag
283.9 tons/day of solids
* Factory purchase cost in mid-1970, except for distillation plant
which is priced for turn-key installation.
** C/S carbon steel; C/M Chrome moly. Capacity: 636 tons/day air
*** Installed cost of 2 MM GPD AMW plant.
of the total purchased equipment cost. The piping, electrical and
utility costs associated with this equipment were each assumed to be ten
percent and process instrumentation 15 percent of the total purchased
equipment cost. Combining the total purchased equipment cost plus the
installation, piping, electrical, utilities, and instrumentation cost
yields the total physical plant cost. Engineering and construction costs
were set at 30 percent of the physical plant cost. Combining engineering
and construction costs with the physical plant costs results in the
direct plant cost. A contractor's fee of five percent and a con-
tingency of ten percent is applied to the direct plant cost yielding the
fixed capital investment excluding cost of distillation. Combining the
fixed capital investment and installed cost of the distillation plant
yields the total fixed capital investment requirement for the given plant.
Determination of the Break-even Price of Water
The process utilizes acid mine water, coal refuse, and flux as raw
materials and produces potable water, sulfur, iron and spent slag as
-28-
-------
products. In this study the acid mine water is considered to be available
at no cost, the coal refuse cost is that associated with transporting
the coal refuse to the plant site and assumed to be in the range of 0 to
$0.5/ton and the flux cost was set at $2 per ton. Rather than consider
the profit or loss of the process, the break-even price of water (BEPW)
was used. Althought the selling price of water can vary considerably
depending upon its marketability at the plant location, to simplify this
presentation the selling price of the potable water is the price of water
required to yield a no-profit-or-loss operation. The break-even price of
water has the units of $/l,000 gallons of potable water. Obviously, the
lower the BEPW for a given process, the more desirable is that process.
The selling price of sulfur was varied in the range of $25 to $40/ton to
show its effect on process economics. Although the iron produced will
contain less than 0.3 percent sulfur, the study assumes the iron contains
one percent. The iron was conservatively valued at $20 per ton. Unused
spent slag from desulfurization can be briquetted to form a uniform road
aggregate which was valued at $0.5/ton. The cost of the briquetting press
was not included in the overall plant cost because it was felt that
utilization of spent slag should not be tied down to one specific
application. Also, other slag desulfurization possibilities employing
molten slag are possible for desulfurization which would preclude the need
for an agglomeration operation (briquetting).
Determination of Operating Revenue
Once the fixed capital investment for a given plant capacity has been
established, the daily capital interest charge can be determined.
Knowing the costs and quantities of raw materials, assuming a three
percent maintenance charge and a labor charge of $300 per day enables
the daily production cost to be determined. The daily production credits
(not including the potable water credit) can be established since the
quantity and selling price of the products are known. The difference
between the daily production charge and the daily production credits is
set equal to the potable water credit so that the operating revenue is
zero. Knowing the potable water credit and the daily production of
potable water enables the break-even price of water to be determined.
A breakdown of the cost factors comprising the operating revenue and
break-even price of water is presented in Table IV for a five million
gallon per day plant, moderately concentrated AMW, eight percent sulfur
refuse and a heat rate of 3.25 million BTU per 1,000 gallons of AMW.
Break-even Price of Water and Process and Economics Factors
In the discussion which follows, the break-even price of water will be
determined as a function of plant size, acid mine water concentration,
sulfur content of coal refuse, selling price of sulfur, capital interest
charge, purchase cost of coal refuse, heat rate and whether neutralization
is used.
To determine the optimum economic operating conditions (optimum operable
point in the process working area), a series of process working areas
were generated and the BEPW determined. In Table V the BEPW is shown
-29-
-------
TABLE IV
Determination of Break-even Price of Water
Investment Cost $ 8,100,000
Potable Water Production 4,975,000 GPD
Daily Production Cost
Capital Interest Charge*, (14%)
Flux, 1105 Tons @ $2/ton
Coal Refuse, 1427 Tons @ $0.25/ton
Labor
Maintenance, 3% of Investment
$
Daily Production Credits (not including potable water credit)
Sulfur, 126 tons @ $25/ton $ 3,150
Iron, 60 tons @ $20/ton 1,200
Slag, 1082 tons @ $.5/ton 541
$ 4,891
Operating Revenue (not including potable water credit) (1,801)
Break-even Price of Water $1.801 x 1,000 = $Q 36/LOOO
4,975,000 Qf water
Potable Water Credit $ 1,801
Operating Revenue 0
* Capital Interest Charge = $8,100,000 x .14/360 days
as a function of various process parameters. Table V consists of six
sets of results, which show the BEPW as a function of slag basicity, slag
recycle fraction, heat rate, neutralization, AMW concentration, and
the sulfur content of the coal refuse. Data of Table V was determined
for a five million gallon per day plant, a 14 percent capital interest
charge, and the following selling prices or purchase costs for the
various raw materials or products; sulfur-$30 per ton, iron-$20 per ton,
spent slag-$0.50 per ton, coal refuse-$0.25 per ton, and dolomitic
limestone-$2 per ton.
Using the process working area generated for a heat rate of 3.25,
operating conditions were selected which yielded slag basicities of 0.8,
loO, and 1.2 at a maximum of ten percent sulfur in the slag. These
operable points occur on the maximum ten percent sulfur line of the
PWA0 The BEPW was then determined for each of these operating conditions.
As seen (Data Set 1) in Table V, the break-even price of water increased
as the basicity increased. This means that the process should be run
at the minimum operable basicity to yield the lowest break-even price of
water.
-30-
-------
TABLE V
BEPW for Process Operation Under Various Conditions
i
OJ
Data
Set
1
2
3
4
5
6
Heat Rate
MM BTU/
1,000 gal
3.25
3.25
3.25
2.00
2.00
2.00
1.75
2.00
2.25
2.50
3.00
3.25
3.25
3.25
3.25
3.25
3.25
3.25
3.25
Flux Rate
Tons /day
395
490
570
242
230
200
207
242
280
305
368
395
395
415
415
408
442
408
372
Slag
Recycle
Fraction
.24
.15
.05
.26
.40
.64
.31
.26
.22
.25
.25
.24
.24
.22
.22
.30
.08
.30
.45
Slag
Basicity
0.8
1.0
1.2
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
0.8
%S in
Slag
10.0
10.0
10.0
10.0
8.6
5.7
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
Economy
Factor
5.0
5.6
6.3
8.3
8.4
8.7
9.8
8.3
7.3
6.6
5.4
5.0
5.0
5.0
5.0
5.3
5.3
5.3
5.3
BEPW $/
1000 gal
.14
.21
.31
.35
.37
.41
.41
.35
.31
.28
.19
.14
.14
.09
.09
.05
.23
.05
-.14
AMW
cone
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Dilute
Med Cone
Med Cone
Med Cone
Med Cone
Neutral-
ization
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
Yes
No
No
No
No
No
No
% sulfur
in coal
refuse
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
10.0
8.0
10.0
12.0
-------
In Data Set 2 the effect of spent slag recycle fraction on the BEPW is
shown. These operating points were selected along the minimum basicity
line on the PWA; therefore, the results yield a slag basicity of 0.8.
The results indicate that the break-even price of water increases as the
spent slag recycle fraction increases. Consequently, the process should
be run using a minimum spent slag recycle fraction consistent with
maintaining a maximum of ten percent sulfur in the slag.
These conclusions indicate that the optimum economic operating point occurs
at the intersection of the minimum basicity line with the maximum (ten
percent) sulfur line of the process working area. With this in mind, Data
Set 3 was generated for heat rates of 1.75 to 3.25 and a slag basicity of
0.8 with ten percent sulfur in slag. Each of the selected heat rates yields
an operable process. However, at a heat rate of 1.75 the economy factory
is 9.5 whereas a heat rate of 3.25 yields an economy factor of five.
Distillation units are normally designed for a maximum economy factor of
ten and, since these economy factors are within the range of commercial
distillation equipment design, the data represent the range of heat rates
for which the process is operable. As shown in Table V the BEPW decreases
as the heat rate increases. This is the result of using an inexpensive
source of energy (coal refuse) for the process which allows a less efficient
distillation unit to be built at a correspondingly lower price. These results
show that optimum economic operation is achieved when using a distillation
unit with an economy factor of five and operating the combustor with a ten
percent sulfur slag of 0.8 basicity.
Data Set 4 presents a comparison of the BEPW for a process operating with
and without neutralization. For this comparison the optimum economic
point (0.8 basicity-ten percent sulfur in slag) was used. As seen, the BEPW
decreased when the acid mine water is not neutralized. Specifically, the
break-even price of water decreases from $0.14 to $0.09/1,000 gallons of
water when neutralization is not employed.
Data Set 5 is used to show the effects of acid mine water concentration on
the BEPW. The break-even price of water decreases from $0.09 to $0.05/1,000 «
gallons of water as the acid mine water concentration increases from dilute
to moderately concentrated. Therefore, it is evident that unneutralized,
concentrated (and still be compatible with the corrosion resistance of the
distillation unit) acid mine water should be used in the process.
The effect of coal refuse sulfur content on the BEPW is shown in Data Set 6.
The break-even price of water decreases as the sulfur content of the coal
refuse increases.
In summary, the BEPW is minimized at a slag basicity of 0.8, a ten percent
sulfur content in the slag, a heat rate which yields the minimum economy
factor (5) and neutralization is not employed in the process. In addition,
the BEPW decreases as the sulfur content of the coal refuse and the acid
mine water sulfate concentration increases.
-32-
-------
The effect of the selling price of sulfur and the purchase price of coal
refuse on the break-even price of water will be illustrated for a process
using moderately concentrated acid mine water, an eight percent sulfur coal
refuse, a heat rate of 3.25 (economy factor equal to five), and a set of
operating conditions which yield a basicity of 0.8 and ten percent sulfur
in the slag. A 14 percent capital interest charge was used in this study.
The BEPW, for a five million gallon per day plant operating at these
conditions, as a function of the selling price of sulfur and the purchase
price of coal refuse is shown in Figure 10. This figure shows that as
the selling price of sulfur increases, the BEPW decreases. Also, as the coal
refuse purchase price increases, the BEPW increases. Also shown in Figure
10 is an area of anticipated price ranges of sulfur and coal refuse.
The purchase price of coal refuse is difficult to ascertain. At present,
coal refuse does not have a market and, consequently, has no market price.
It is believed that the cost of coal refuse will be mainly the cost
associated with transporting it to the AMW plant site. The price of coal
refuse was assumed to range from 0 to $0.5 per ton. The zero cost number
will probably apply to a coal producer operating the Acid Mine Water
Process. Coal refuse produced from the coal washing operation can be
brought directly to the AMW plant site at the same cost as transporting it
to a coal refuse pile. The $0.25 per ton number is assumed to apply to a
noncoal producer whose AMW plant is in close proximity to a coal washing
facility. The $0.5 per ton number is believed to be a reasonable cost for
coal refuse when the acid mine water plant is some distance away from the
coal refuse source.
Selling price of sulfur is extremely variable. In the last four years,
it has ranged from a high of approximately $40 per ton to a low of less
than $20 per ton. Figure 10 shows that the selling price of sulfur is a
prime consideration in the economics of the process. At a $25 per ton
selling price, the BEPW is $0.29, whereas at $30 per ton the break-even
price of water drops to $.16/1,000 gallons of water. The purchase price
of coal refuse is an equally important economic consideration as can be
seen from Figure 10 which shows a $0.14/1,000 gallons of water variance in
the BEPW as the price of coal refuse varies from 0 to $0.5 per ton.
As of mid-1970 the selling price of sulfur has been approximately $25/ton.
Assuming coal refuse is available at $0.25/ton, the BEPW for a five million
gallon per day plant is $0.36/1,000 gallons of water.
The capital investment requirements and intermediate results necessary to
calculate the capital investment for various sized acid mine water treatment
plants are shown in Table VI. This, table was prepared using moderately
concentrated AMW, an eight percent sulfur coal refuse, a heat rate of 3.25,
a set of operating conditions which yield a basicity of 0.8 and ten percent
sulfur in the slag. The capital investment requirement increases as the
plant size increases. Figure 11 indicates that the capital investment is
not a linear function of plant capacity and economies can be realized by
using higher plant capacities. This infers that the plant should be located
-33-
-------
,60
o
co
ts
o
to
O
o
o
o
0)
4-J
CO
O
•H
PM
d
01
CO
cu
M
pq
.50
.40
.30
,20
.10
Basis: 14% Capital Interest Charge
15
20
25
30
Coal Refuse
$0.50/Ton
Coal Refuse,
$0.25/Ton
Coal Refuse,
$0/Ton
35
40
Selling Price of Sulfur, $/Ton
FIGURE 10 - EFFECT OF SULFUR PRICE ON BREAK-EVEN PRICE OF WATER
-34-
-------
12
in
p
c 3
c
o
H^
a
j_i
T 4
rar.ic3.ty, ?'ill:'.on
? rpr "\".v
FJ/'URE 11 -
0" "T/'.?'T r:yvr>^rTTv 01\ PA^TTAT T'.'Vl f
-35-
-------
TABLE VI
Capital Investment for Various Plant Sizes
Equipment Complex Cost 0.5 MM GPP 1 MM GPP 2 MM GPP 5 MM GPP 10 MM GPP
Coal Handling 53,000 81,000 122,000 213,000 321,000
Neutralization -0- -0- -0- -0- -0-
Steam Turbine Air Compressor 16,000 25,000 38,000 66,000 100,000
Pirect Fired Furnace 41,000 63,000 95,000 164,000 249,000
Waste Heat Boiler 39,000 59,000 90,000 156,000 236,000
Rotary Kiln Pryer 144,000 218,000 330,000 572,000 867,000
Pesulfurization 38,000 57,000 87,000 150,000 229,000
Combustor 27.000 40.000 61.000 106.000 161.000
1. Total Purchased Equipment 358,000 543,000 823,000 1,427,000 2,163,000
Installation, 40% of 1 143,000 218,000 330,000 571,000 865,000
Piping, 10% of 1 36,000 54,000 82,000 143,000 216,000
Electrical, 10% of 1 36,000 54,000 82,000 143,000 216,000
Instrumentation, 15% of 1 54,000 82,000 124,000 214,000 324,000
Utilities, 10% of 1 36,000 54.000 82.000 142.000 216.000
2. Physical Plant Cost 663,000 1,005,000 1,523,000 2,640,000 4,000,000
Engineering & Construction,
30% of 2 199.000 302.000 457.000 792.000 1.200.000
3. Pirect Plant Costs 862,000 1,307,000 1,980,000 3,432,000 5,200,000
Contractor's fee 5% of 3 43,000 65,000 99,000 172,000 260,000
Contingency, 10% of 3 86.000 131.000 198.000 343.000 520.000
4. Fixed Capital Investment 991,000 1,503,000 2,277,000 3,947,000 5,980,000
5. Installed Cost, Distillation 1.045.000 1.583.000 2.400.000 4,158.000 6.304,000
Total Capital Investment 2,036,000 3,086,000 4,677,000 8,105,000 12,284,000
-------
at a large source of acid mine water provided coal refuse is available in
the vicinity.
Figure 12 shows the break-even price of water as a function of plant
capacity and the capital interest rate, with coal refuse at $0.25/ton
and sulfur at $25/ton. This figure indicates that both plant size and
capital interest rate has a profound effect on the break-even price of
watero For example, at a capital interest rate of 14 percent the BEPW
decreases from $1.30 to $0.14/1,000 gallons of water when the plant size
is increased from one to ten million GPD of AMW. For a five million GPD
AMP plant, the BEPW decreases from $0.36 to $0/1,000 gallons of water
when the capital interest charge is reduced from 14 percent to six per-
cent.
Plant capacity has an important effect on the break-even price of water.
The BEPW is considerably reduced as the plant capacity is increased
from one to ten million GPD. Normally, the flow of most acid mine water
streams is less than two million GPD. In heavily mined regions, several
acid mine water streams can occur relatively close to each other; con-
sequently, several streams could be combined to yield a five million GPD
stream. Based on this reasoning, a five million gallon per day plant
will be used to illustrate the capacity and temperature of the individual
process streams and payback as a function of the selling and break-even
price of water. Table VII presents the capacity and temperature of all
process streams for a process utilizing moderately concentrated AMW,
eight percent sulfur coal refuse, a heat rate of 3.25, a flux rate of
1,105 ton/day and a slag recycle fraction of 0.084. These values yield
a slag basicity of 0.8 and a ten percent sulfur content in the slag. For
convenience, the process flow chart is shown in Figure 13 with all streams
labeled as in Table VII.
Figure 14 was prepared to illustrate the effect of water selling
price on capital recovery. This figure shows a plot of payback versus
the difference between selling price and break-even price of water. Pay-
back is total investment divided by annual profit. For Figure 14,
investment is $8.5 million for a five MM GPD plant, operating under the
conditions of Table VII. The capital is higher than that of Table VI,
because it includes cost of land, spare parts, and shakedown, plus working
capital requirements during shakedown. Because the break-even price of
water includes all direct and indirect costs, the difference between
selling (at the plant) and break-even prices is used as profit in
calculating payback. The selling price of water at the plant will depend
on demand for water in nearby markets and cost of transporting water to
these markets. The break-even price of water is a function of process
parameters and capital interest charges.
A six percent capital interest charge is not unrealistic if municipal
money in the form of tax-exempt six percent bonds is available. For a
plant with an estimated life of 20 years and a one year construction and
shakedown period, operating as indicated in Table VII the break-even
-37-
-------
1.2
1.0
u-i
o
c.
c
c-
CJ
4-J
rt
a;
u
I
"ni
n-
I
I
I
2 4 r
Plant Capacity, "illior °alions /-'V Per
12 -
VArpT
-38-
-------
AMW
43
Sulfur
Coal
Refuse
FIGURE 13 - PROCESS FLOW CHART OF AMW TREATMENT PLANT
-39-
-------
o
•H
a
l-
40
30
20
10
0
0.10
0.20
rifferencc I!ctveen rellir." T1ricn
of "atar, !-/]0^0 '
0.50 P.f.O
:-O'rn "ri cc
-40-
-------
TABLE VII
Stream Capacities and Temperatures for 5 MM GPD Plant
Stream
5
6
7
8
10
11
12
13
14
16
17
20
21
22
26
27
29
30
31
32
33
34
36
37
38
39
41
42
43
44
45
46
47
48
Description
Dryer solids
Combustor slag
Sulfur product
Spent slag
Capacity Tons/day
767
1245
126
1182
Terno. °F
Combustor offgas (assuming 10% heat loss) 3191
Preheated air 2484
Dry, crushed coal 1354
Flux 1105
Iron product 60
Combustor offgas to dryer 700
Potable water 22807
Acid mine water 20948
Concentrated brine slurry 214
Dryer flue gas 2307
Air 956
Water product 20734
Combustor offgas to air preheater 174
Recycled spent slag 99
Spent slag product 1082
Air 2484
Air preheater flue gas 412
Air for preheating 238
Wet Coal 1427
Coal drying kiln flue gas 513
Combustor offgas to coal drying kiln 31
Air 42
Combustor offgas to boiler 2286
Air 3122
Boiler flue gas 5409
Water to and steam from boiler 2103
Air into steam-turbine air compressors 2619
Air to convey coal into combustor 135
Water into desulfurizer 36
Air into desulfurizer 270
2200
2700
200
2000
2430
1332
150
70
2700
2430
180
70
180
250
60
212
2430
70
300
70
500
60
70
280
2430
60
2430
60
2280
300
60
60
60
60
price of water is $0/1,000 gallons. If the plant sells water at $0.50/
1,000 gallons (at the plant), Figure 14 shows that payback would be nine
years from the initial date of the bond issued (beginning of construction).
This is based on yearly profits of $900,000 which are set aside in a
sinking fund which conservatively earns five percent interest per year.
The sinking fund could, at the end of nine years, be used to retire the
bonds. If this is done, then the profits become $1.4 million per year
since the capital interest charge is now eliminated. If the profits
generated in the 10th to 20th year are accumulated yearly in a construction
fund earning five percent interest, the construction fund would be worth
$17.8 million at the end of 20 years when the plant life is exhausted.
-41-
-------
Assuming an average inflation rate of four percent, in 20 years the
same plant would require $17.7 million to construct. The construction
fund could then be used to totally finance a new plant and the profits
would be $1.4 million per year. It is impossible to predict the future
but the above figures indicate that the process can generate sufficient
profits to be self-perpetuating.
The important conclusion is that if low cost money is available, the
economics of the process are highly favorable and should stimulate
municipalities to rid their communities of acid mine water pollution.
An alternative manner in which to consider the economics of the process
is to determine the selling price of water required to capitalize the
plant over its life without any profit. Figure 14 shows that the
difference between selling price and break-even price of water must be
$0.235/1,000 gallons to yield a payback period of 20 years. This means
that for the above plant, the selling price of water must be $0.235/1,000
gallons to eliminate five MM GPD of acid mine water at no cost (provided
that capital is available at six percent).
As of this writing, (late 1970), the assumption of six percent tax-
exempt municipal bond capital appears quite conservative. Distilled
water prices in the range of $0.2/1,000 gallons to $0.5/1,000 gallons at
the plant appear quite realistic for plants that are not too remote from
consumers. Therefore, this process offers the incentive of profit or,
at least, no cost for eliminating acid mine drainage from our streams.
It is difficult to make a cost comparison between AMW treatment processes
because the literature data are based on AMW of different compositions
for plants of various capacities or are estimated at different capital
interest rates. In the following discussion, published cost data for
lime neutralization and ion-exchange treatment of AMW were adjusted to a
common basis for comparison with the new process. The common basis is a
plant constructed at a six percent capital interest rate to treat five
MM GPD of AMW of acidity of 500 (ppm CaCC>3) . The current study has shown
that the new process has a zero operating cost (no cost) when eight per-
cent sulfur coal refuse at $0.25/ton is used in the base case plant.
A study of an ion-exchange AMW treatment process was reported by J. L. Rose
at the Third Symposium on Coal Mine Drainage Research in May, 1970 (pages
267-278 of the Proceedings). Rose estimated costs of $0.73 and $0.45
per 1,000 gallons of AMW for plant capacities of one and ten MM GPD.
These data were used to determine an ion-exchange AMW treatment cost
of $0.55/1,000 gallons for the above-mentioned base case. Comparing the
ion-exchange and new process treatment cost, it is seen that the ion-
exchange process is substantially more costly.
At the Second Symposium on Coal Mine Drainage Research in May, 1962,
-42-
-------
(pages 274-290 of the Proceedings) Corsaro reported costs for lime
neutralization plants treating up to 8.1 MM GPD of AMW of various
concentrations. Using Corsaro's data, an operating cost of $0.20/1,000
gallons of AMW was obtained for the base case plant. In a report
entitled "Operation Yellowboy" for the Pennsylvania Coal Research
Board, Dorr-Oliver estimated lime neutralization treatment costs for
various actual AMW sources. They estimated an operating cost of
$0.22/1,000 gallons AMW at a four percent capital interest rate for the
Blue Coal Corporation Loomis No. 4 site with a 5.7 MM GPD AMW source of
560 acidity. Adjusting these data for a six percent capital interest
charge results in a $0.25/1,000 gallons operating cost. It is interesting
to note that the operating costs determined by Dorr-Oliver for various
AMW sources ranged from $0.10 to $1.23 per 1,000 gallons.
Even if it assumed that lime neutralization treatment costs only $0.20/
1,000 gallons of AMW, the new process is the least expensive means to
treat AMW because it has no operating cost for the base case.
There are other considerations which must be included in this discussion.
Although operating costs for lime neutralization are high, the capital
requirements are low enough to make lime neutralization attractive to
small mine operators. On the other hand, both the ion-exchange and lime
neutralization processes present by-product disposal problems. The new
AMW treatment process does not produce any pollution-producing by-products
and generates distilled water which can be sold to industry at a profit.
Also the availability of distilled water in an area can bring in new
industry and enhance the economic well-being of the community. In
contrast, the lime neutralization treatment process does not produce
water of any significant commercial value and the ion-exchange process
produces a water whose quality is available in numerous locales through-
out the nation and offers no particular inducement for industry.
Accordingly, a meaningful comparison must be based on treatment of a
specific stream, to include consideration of actual capital and operation
cost.
-43-
-------
ACKNOWLEDGEMENTS
The study described in this report was financed by a Federal Water
Quality Administration (FWQA) contract to Black, Sivalls & Bryson, Inc.,
(BS&B). The Applied Technology Division of BS&B acknowledges the
technical assistance of Messrs. R. D0 Hill, and R. B. Scott of the
EPA and of the Chemical Service Engineers, Inc., personnel who
conducted the bench-scale laboratory work.
-44-
-------
REFERENCES
1. Endell, K. and Ci Wens, Industrial Heating, 4, (2), 143, (1937).
2. Rait, J. R., Trans, Brit. Ceram. Soc., 40 157-204, 231-269, (1941).
3. Fehling, H. R., "Erosion of Refractories by Coal Slag", Institute of
Fuel J. Vol. II No. 59, pp. 451-458, June (1938).
4. Herty, C. H. etal, Min. Met. Invest. Coop. Bull., No, 46 1, (1930).
5. McGannon, H. E., "The Making, Shaping & Treating of Steel," United
States Steel Corporation, August (1964).
6. Herty, C. H., Blast Furnace & Steel Plant, 25, 1000 (1937).
7. Herty, C. H., Jr. Trans. Amer. Inst. Min & Met. Eng., Iron and
Steel Div., pp. 284-299 (1929).
8. Gul'tyai, I. I. Iz Akad Nauk SSSR, Otd. Tekhn, Nauk. Met. Toplivo,
No. 3, (1962).
9. F. Henen, G. R. etal, Acid Open Hearth Research Assn., Inc. Bull.
No, 1, (1945).
10. Kozakevitch, P., Rev. de Metallurgie, Vol. 51. 569-587, (1954).
11. Machin, J. S., Lee T. B., Hanna, D. L., J. Am. Ceram. Soc, Vol. 35
pp. 322-25, (1952).
12. Cameron, J. Gibbons, T. B., and Taylor, J., Journal of the Iron
and Steel Institute, pp. 223-28, December (1966).
13. Hatch, W. G., and Chipman, J., Metals Transactions, Vol. 49.
pp. 479, 86 (1946).
14. Sharma, R. A., and Richardson, F. D., Journal of the Iron and Steel
Institute, Vol. 198 p. 386, (1961).
15. Sharma, R. A., and Richardson, F. D., Journal of the Iron and Steel
Institute, Vol. 200, p. 373, (1962).
16. Kozakevitch, P. "Physico-chemical Measurements at High Temperatures,"
Butterworths Scientific Publications, pp. 212-221, (1959).
17. Dallavalle, J. M., "Micromeritics," Pitman Publishing Corp., New
York, pp. 272, 334, (1948).
18. Gross, J., "Crushing & Grinding," U. S. Bur. Mines Bull. 402, (1938).
-45-
-------
19. Brown, G. G., "Unit Operations," John Wiley & Sons, Inc. New York
pp. 42-45 (1950).
20. Perry, J. H., "Chemical Engineers Handbook," McGraw Hill Book
Company, Inc. p. 226 (1950).
21. Hougen, 0. A., Watson, K. M. , and Ragatz, R. A., "Chemical Process
Principles," Part I, John Wiley & Sons, Inc. New York, pp. 262,
(1954).
22. Boulanger, C. and Leroy, C. , Fr. Patent 873, 093, June 29, 1942.
23. Smyers, W. H. and Manny, E. H., U. S. Patent 3,249, 402, October 2,
1962.
24. Gunterman, W. Fischer, F. and Kraus, H., Ger. Patent 1,184,895,
January 7, 1965.
25. Franklin, R. L., Guseman, J., and Pelczarski, E. A., U. S. Patent
3,125,438, November 8, 1962.
26. Odeen G. A., etal, Norway Patent 83,374, March 22, 1954.
27. Rudweva A. V., and Panov, A. S., Iz., Akad, Nauk SSSR, Otd, Khim.
Nauk, pp. 553-8 (1962).
28, Steyn, J. G. D., Mineraby, pp. 108-17, May 35 (269), (1965).
29. Squires, A. M., "Reaction Which Permits Cyclic Use of Calcined
Dolomite to Desulfurized Fuels Undergoing Gasification," Am. Chetn.
Soc. Div., Fuel Chem. Vol. 10, No. 4, pp. 20-41, September 11-16,
(1966).
30. Woehlbier, F. H., and Rengstorff, G. W. P., "Preliminary Study of
Gas Formation During Blast Furnace Slag Granulation with Water",
Preprint of paper presented at the annual meeting of the Air
Pollution Control Association, June 26, 1968.
31. Wen, C. Y. Industrial and Engineering Chemistry, pp. 34-54,
September (1968).
32. Olsson, R. G., Koump, V., and Penzak, T., "Rate of Solution of
Carbon in Molten Iron-Carbon Alloys," Annual AIME meeting of
February, 1965.
33. Leary, R. J., and Ostrowski, E. J., "Pneumatic Lance Injection of
Carbonaceous Solids for Recarburizing Open-Hearth Melts," United
States Bureau of Mines, Open File Report, 1963.
-46-
-------
GLOSSARY
Basicity - Weight ratio, (CaO + MgO)/(Al203 + SiC^), of the components
in a slag.
Desulfurized Slag - Slag in which all or most of the calcium sulfide has
been removed.
Economy Factor - For a distillation plant, the weight ratio of the
water distilled to the steam used for distillation.
Flux - Any agent which preferentially combines with the impurities of a
molten metal and aids in the smelting operation.
Heat Rate - Millions of BTUs of fuel consumed per thousand gallons of
acid mine water processed.
Partition Ratio - Percent of sulfur in a slag in contact with iron
divided by percent of sulfur in the iron.
Process Working Area - The resulting operable range of process variables
when certain constraints on slag basicity, sulfur content of slag and
air preheat temperature are applied to the process working area diagram.
Process Working Area Diagram - Plot of slag basicity vs. slag recycle
fraction at various flux rates into the process.
Slag - Product formed by the action of a flux upon the gangue of an ore
or ash of a fuel. In this study, a mixture of silica, alumina, mag-
nesium oxide, calcium oxide and calcium sulfide,,
Slag Fluidity - Measure of the ability of a slag to flow. A qualitative
term used interchangeably with apparent viscosity.
Spent Slag - Same as desulfurized slag.
Spent Slag Recycle Fraction - The fraction of the desulfurized slag
exiting the desulfurization complex which is recycled back to the kiln
dryer.
-47-
-------
APPENDIX A
LABORATORY STUDIES
SLAG FLUIDITY
Introduction
Addition of limestone to control combustor slag properties contributes
heavily to process costs. Limestone is added to the slag to increase
its sulfur-retention capacity and to control its fluidity. A highly
fluid slag is undersirable. because it increases costs by rapidly attacking
combustor refractories^-'2' . On the other hand, the slag must be fluid
enough to flow from the combustor. Accordingly, a search was made for
viscous-but-pourable slags that could be derived from coal ash and AMW
constituents with a minimum of additives.
An abundance of data is available for the effect of slag composition on
the viscosity of blast furnace and steel making slags . However, to
minimize process operating costs, very low basicity high sulfur containing
slags will be required. For these types of slags, literature data for
the effect of high sulfur concentration on slag viscosity are scarce-*.
Since a large number of slag compositions are possible, a quick screening
test was required to rank the various types of slags according to their
relative fluidities so that more detailed measurements could be completed
on those deemed suitable.
A simple test used in steelmaking operations has been developed by Herty
and has proven to be a valuable guide for a rough evaluation of slag
fluidity. For this reason, the Herty method was modified and adopted
for use in our screening experiments. The principle of the Herty test is
quite simple: A known volume of molten slag is poured down an inclined
plane. Contact with the cold inclined plane causes the slag to solidify.
The thickness of the slag layer at any arbitrary specified point is a
function of the fluidity (viscosity) of the material. Obviously a test
as simple as this is influenced by a number of slag properties including
heat transfer characteristics. However, as a relative guide for comparing
differences in fluidity, the test has proved it can provide valuable
information2'6'^'7.
Experimental Procedure - General
Master batches of slag of a given composition were prepared in two-
kilogram lots,, Slags were prepared by adding reagent grade ingredients
into a closed pyrolytic graphite crucible that was inserted within a
high temperature furnace maintained at 2700°F. An argon purge was used
in the interior of the furnace to eliminate air from the system. After
a soaking time of six hours, the carbon crucible containing the molten
slag was removed from the furnace, the slag was then quenched in argon
and examined to determine if a homogeneous melt had been formed. It was
found that to obtain a uniform slag, it was necessary to heat, melt, cool,
and crush the slag three times before homogenity was observed.
-49-
-------
The equipment used in the Herty Test is shown in Figure 15. With the
exception of the orifice cup which was made of pyrolitic graphite, all
other components were of carbon steel construction. The inclined plane
(30 inches long) was maintained at an angle of 14 degrees from the hor-
izontal. A departure was made from the recommended £ angle of inclination
of 30 degrees because the slag, upon solidifying, tended to break and
slide down the inclined plane. Consequently, a 14 degree inclination
angle was adopted to render the slag immobile after solidification.
The experimental procedure consisted of heating a 100 gram slag sample
in the high temperature furnace until temperature equilibration was
achieved. Variations in the test results during experimentation soon
showed that a residence time at temperature of about three hours was
required to assure temperature uniformity. Sulfur analysis of the slag
before and after melting indicated no significant sulfur losses. The
hot sample was withdrawn from the furnace and transferred to the orifice
cup as quickly as possible and immediately poured down the inclined plane.
Generally, the transfer of the slag from the furnace to the orifice cup
was completed in 20 seconds or less. Measurements were made of the length,
width, and thickness of the solidified slag lying on the inclined plane.
These were used to correlate slag relative fluidity and, indirectly,
viscosity with changes in composition.
Standardization of Experimental Techniques
A short study was conducted to determine the effect of slag time at
temperature and position in the furnace on reproducibility of slag
fluidity tests. Results of this work using a commercial blast furnace
slag (slag No. 1, Table VIII) are presented in Figure 16. A slag
residence time of at least three hours in the furnace is required to
maintain reproducibility. Although not severe, a temperature gradient
(as indicated by fluidity differences) appears to exist in the furnace
since one sample position continually yielded lower fluidity values. In
all subsequent work, a minimum time at temperature of three hours was
established as a standard„
To evaluate the effect of the number of slag remelts required to obtain
a uniform slag composition and to determine the effect of number of
remelts on slag fluidity, a second series of tests was completed. In
these tests, two slags (Table VIII, slags No. 2 and 3) were used in the
fluidity tests, crushed to -20 mesh after cooling, remelted and used
again in the fluidity test. This procedure was repeated three times for
slag No. 2 and four times for slag No. 3. Results of this work are
shown in Table IX and X. Data indicate fluidity increased with increasing
number of remelts. Data in Tables IX and X indicate that three remelts
per slag sample should be sufficient to obtain consistently reproducible
fluidity test results. The fluidity increase can be attributed to
increased compositional uniformity of the slag sample, which was observed
upon remelting. This is consistent with the literature where similar
results were obtained.
The time element involved in moving the slag sample from the heating furnace
-50-
-------
nrn'fice
I
Ui
Inclined rianc
"inirr i ^ -
-------
Ln
M
I
TABLE VIII
Synthetic Slag Chemical Composition
Chemical Composition, Percent by Weight
SLAG
1.
2.
3o
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14 o
15.
CaO
43.3
38.0
30.0
49.6
42.9
46.9
34.8
34.3
51.7
47.9
28.8
23.6
19.4
42.9
42.9
MgO
7.0
10.0
12.0
3.1
12.0
13.1
9.7
12.0
3.2
7.0
11.5
11.0
10.8
12.0
12.0
CaS
3.8
--
12.9
--
--
--
--
8.6
--
--
12.4
15 „ 9
19.3
--
--
S
1.7
--
5.7
--
--
--
--
3.8
--
--
--
--
--
--
--
Al?0?
10.4
--
12.7
12.2
12.7
11.3
15.7
12.7
12.7
12.7
12.2
11.7
11.4
5.0
Si02
35.6
52.0
32.4
31.1
32.4
28.7
39.8
32.4
32.4
32.4
31.0
29.8
29.2
40.1
CaF2
4.0
4.0
8.0
10
Basicity* Si09/Al90^
22.5
22.6
1.1
0.9
0.9
1.2
1.2
1.5
0.8
1.02
1.2
1.2
0.9
0.8
0.7
1.2
1.2
3.4
2/6
2.5
2.6
2.5
2.5
2.6
2.6
2.6
2.5
2.5
2.6
8.0
1.0
CaO/MgO
672
3.8
2.5
16
3.6
3.6
3.6
2.9
16.2
6.8
2.5
2.1
1.8
3.6
3.6
* Basicity is expressed as the weight ratio of basic oxides (CaO + MgO) to acid oxides (SiO? + A1-0-)
-------
o
20
K
c/
o
t-
J.-1
t 16
c:
r;.
O
•^r.
"O"
o
,®-
Tosition :'n ""urr.ace
0 ^ront Tenter
• i I I
I t
20 40 60 "0 ion ]2° 140 1 f-0 1 ?o
Residence ^ire of Sarp.lc in ^r.r^rr.r, T'i-nter,
Fin:ri: ifi - p.n,ATTo?: j'rn'i'r:* "LTT^TT' AT-T TTi'T^7"]'rr
^r r/.?'"LE I)* rTTry».pr
-53-
-------
into the orifice cup was also evaluated. Within the limits of experi-
mental error, data reproducibility was unaffected provided elapsed
time did not exceed 20 seconds.
TABLE IX
The Effect of Remelting on Slag Fluidity
Slag Composition (Weight Percent): CaO--38.0, MgO--10.0, S.,02--52.0
Fluidity Stringer
Number of Remelts Length (Inches) Furnace Temp. °p
1 17.13 2640
1 19.75 2640
1 17.10 2640
1 18.75 2640
Average 18.18
2 19.75 2640
2 20.25 2640
Average 20.00
3 20.50 2640
3 21.50 2640
3 22.50 2640
Average 21.50
TABLE X
The Effect of Remelting on Slag Fluidity
Slag Composition (Weight Percent): CaO 30_0, CaS 12.9,
MgO 12.0, A1203, 12.7, S..02, 32.4
Fluidity Stringer
Number of Remelts Length (Inches) Furnace Temp.Op
212.42600
2 10.5 2600
2 11.0 2600
Average 11.3
3 12.8 2600
3 10.3 2600
3 14.8 2600
Average 12.7
4 13.3 2600
4 13.4 2600
4 11.8 2600
Average 12.8
-54-
-------
Calibration of the Herty Fluidity Equipment
To establish a correlation between viscosity and the Herty fluidity of
a given slag, the equipment was calibrated by use of a slag of known
viscosity (see Table VIII, slag No. 4)8. A plot of the literature data
for viscosity as a function of temperature is presented in Figure 17.
Using the absolute viscosity shown in Figure 17 and Herty fluidity at the
same temperature (Figure 18) it is possible to cross-correlate the two
measurements so that an indication of viscosity can be obtained from the
Herty fluidity results . Such a correlation is shown in Figure 19.
Because of the complexity of the physical phenomena involved during the
cooling and flow of molten slag down the inclined plane it is not
possible to claim great accuracy for the correlation. Nevertheless, it
has a practical use in that a correlation of this nature permits a
rough estimate of the viscosity of the slag.
Discussion of Results - The Effect of Basicity on Fluidity
To explore the lime requirements of the process, fluidity tests were
completed on slags No. 5, 6, 7 (Table VIII) to determine the effect of
basicity ratio on slag fluidity. Basicity ratio is the weight ratio of
basic oxides (CaO and MgO) to acid oxides (SiC>2 and A^C^) in the slag.
In these tests, the weight ratio of silica to alumina was maintained
constant at 2.5. This ratio was chosen to simulate the contribution of
silica and alumina from the coal ash and limestone that will be used in
operation of the combustor. Fluidity measurements were completed at
a furnace temperature of 2600°F and related to apparent viscosities
through the correlation of Figure 19. Figure 20 shows that the apparent
viscosity exhibits a minimum at a basicity ratio of about 1.1. Viscosity
at this ratio was approximately six poise.
Increasing the basicity ratio to 1.5 resulted in a viscous slag with an
apparent viscosity greater than 22 poise. Decreasing the basicity to
0.8, also increased the viscosity but not at the same rate. At this
basicity ratio, an apparent viscosity of approximately ten poise was
observed.
A comparison of measured apparent viscosity and literature" values is
shown in Figure 20. The comparison indicates a reasonable approximation
to the actual viscosity.
Effect of Calcium Sulfide Content on Fluidity
Because the process operates with slags containing approximately ten
times the sulfur content of the iron-making slags discussed in the
literature, the effect of calcium sulfide content on slag fluidity was
evaluated. In these tests, calcium sulfide replaced calcium oxide
starting with the slag composition given in Table VIII for slag No. 5
(see Table VIII, slags No. 3, 5, 8 for slag composition used in this work)
The results of this work are shown in Figure 21 which presents the effect
of calcium sulfide addition on the apparent viscosity of the slag having
an initial basicity of 1.2. Figure 21 shows the apparent viscosity of
the slag was relatively unchanged provided that calcium sulfide addition
-55-
-------
40
30
0)
CO
•H
O
CO
O
o
CO
•H
20
10
o
2,
1.5
6-
(Slag No. 4 Table VIII)
2400 2500 2600 2700
Temperature, °F
2800
FIGURE 17 - EFFECT OF TEMPERATURE ON VISCOSITY
OF A SYNTHETIC SLAG
-56-
-------
28
26
24
22
20
18
16
12
in
8
2400
2500 2600
Temper at ure,
To.
2700
2HOO
UPE .18 - ET^l-TT OF TE'PFP-ATTTr OK fLA^ '"LI"I^TTV
-57-
-------
0)
co
•H
O
PL,
•r-l
co
O
O
CO
•r-t
22
20
18
16
14
12
10
8
6
(Slag No. 4 Table VIII)
O Average of Data Points
8 10 12 14 16 18 20 22 24 26
Slag Fluidity
FIGURE 19 - RELATIONSHIP OF SLAG FLUIDITY AND SLAG VISCOSITY
28
-58-
-------
-------
did not exceed approximately eight percent. This agrees reasonably
well with the statement found in the literature10 for blast furnace
slagso At high calcium sulfide contents, apparent viscosity increased
significantly. At 13 percent calcium sulfide an apparent viscosity
of about 16 poise was observed. At this calcium sulfide level, basicity
of the slag (based on residual lime and magnesia content) was 0.9. A
comparison of the apparent viscosity obtained at this sulfide level with
those of Figure 20 for 0.9 basicity slag shows that addition of calcium
sulfide essentially doubled slag viscosity. Thus the data indicate
that adding calcium sulfide to the slag has a pronounced effect on
increasing its apparent viscosity.
Data of Figure 21 were recalculated to determine the percentage
replacement of the initial calcium oxide by calcium sulfide and to
determine its effect on basicity and apparent viscosity of the slag.
These results are presented in Table XI and show that as CaS replaces
calcium oxide, apparent viscosity increases and basicity decreases. At
a replacement level of 52.4 percent (slag containing 22.5 percent CaS)
basicity of the slag decreases to 0.6 and an apparent viscosity of 56
poise is expected. The latter value is based on an extrapolation of
Figure 20. Interpolation between data comprising Figures 19 and 20
indicates that a slag containing 22.5 percent calcium sulfide (ten per-
cent sulfur) and having a basicity of 0.8 will yield an apparent
viscosity of 44 poise. It is believed (discussed in a later section)
that an apparent viscosity of this magnitude will result in an operable
combustor.
TABLE XI
The Variation of Slag Apparent Viscosity With
Basicity and CaS Content
Initial Composition (Weight Percent): CaO 42.9, MgO 12.0, Al_0 12.7,
Si02 32.4
Furnace Temperature 2600°F
Percent Replacement of Apparent Viscosity
Initial CaO by CaS Gas Basicity Poise (Average)
0 0 1.2 6
20.0 8.6 1.0 9
30.1 12.9 0.9 16
47.7 17.9 0.8 36*
52.4 22.5 0.6 56*
* Extrapolated Value
Significance of Fluidity and Apparent Viscosity Measurements
The term fluidity as used in this report should not be confused with the
scientist's fluidity which is defined as the reciprocal of the absolute
viscosity-^. Because of the substantial number of factors that can
influence the Herty relative fluidity test, it is not intended that the
data given here define an absolute viscosity measurement. Rather they
should be a guide to relatively rank the flowability of each of the slags
-60-
-------
studied. For expediency's sake, fluidity has been converted via the
correlations presented earlier to an apparent viscosity which is more
generally understood.
In operation of the combustor, it is extremely important that the slag
be maintained in a fluid condition to facilitate removal from the
combustor and to minimize slag expansion due to entrainment of rising
gas bubbles. At the same time, sufficiently high viscosity must be
maintained in the slag layer to minimize corrosive attack by the slag
material on the refractory lining of the vessel. Consequently, an
optimum slag viscosity should be between the extreme low and high.
A blast furnace can operate over a wide range of slag viscosities up
to 30 poise. The fluid slag percolates down through a packed bed of
solids which implies that flow at these high viscosity levels is not
a severe problem. Also this slag desulfurizes pig iron located in the
hearth of the furnace. In the combustor, slag need not pass through
the interstices of a bed of solid particles, as in the blast furnace.
Therefore, it is reasonable to assume that viscosities higher than the
limits used in the blast furnace operation can be effectively employed
provided that the slag can be poured from the combustor.
In the open hearth, slag viscosities are maintained at a level not much
in excess of two poise. The reason for that is that removal of sulfur
from steel is normally controlled by diffusion of sulfur through the
slag layer. Consequently, to minimize residence time in the open
hearth and increase production, low viscosity slags are employed. In
the combustor, the kinetics of desulfurizing the iron contained in the
vessel are not as critical since (1) a much longer slag residence time
can be employed in the combustor, and (2) the iron sulfur content will
be approximately 50 times that found in steel making operations.
Accordingly, sulfur recovery by the slag does not have the critical
dependence on viscosity found in open hearth practice.
For these reasons, the combustor probably will be operable with slags
with apparent viscosities of 50 poise or less. The upper limit is on
the apparent viscosity for an operable combustor is not precisely
known. Effect of high-viscosity fluid slags on slag expansion, corrosion
rate and capacity for entrapment of solid particles can only be
positively established during operation of the combustor. However,
we believe that if the combustor is operated under the guidelines
specified in the Engineering Design Recommendations, satisfactory
combustor operation will ensue,
Effect of MgO Additions on Slag Fluidity
To determine effect of magnesia on apparent slag viscosity, a series of
fluidity measurements were completed at a furnace temperature of 2600°F.
In these experiments, basicity and silica to alumina weight ratio were
maintained at 1.2 and 2.6 respectively. Magnesium oxide content was
varied over the range of 3.2 percent to 12 percent (see slags 5, 9, 10,
Table VIII). Results of this work are shown in Figure 22 .
-61-
-------
0)
co
•H
O
CM
CO
O
O
CO
•H
d
cu
!-l
CD
P.
P.
0)
TD
•r-1
3
CO
S
CO
O
d
CD
o
n
0)
PM
24
20
16
12
Furnade Temperature 2600 F
4 8 12
Percent Magnesia in Slag
FIGURE 22 - THE EFFECT OF MAGNESIA CONTENT ON APPARENT
SLAG VISCOSITY
24
20
16
12
Apparent Viscosity
6-11 poise
Furnace Temperature 2600°F
4 8 12
Percent Fluorspar in Slag
16
FIGURE 23 - THE EFFECT OF FLUORSPAR ON APPARENT
SLAG VISCOSITY
-62-
-------
Data reveal that apparent viscosity of the slag decreases with increasing
magnesia content. This is consistent with the literature5, and indicates
that magnesium oxide or dolomitic limestones can be used to decrease
the overall viscosity of the slag. At an MgO content of 3.2 percent by
weight, a fluid slag was not obtained at a furnace temperature of 2600°F.
This point, although not shown in Figure 22, is represented by the
asymptote for the curve drawn through the liquid slag points. References
to the literature 11 indicates that this slag composition borderlines the
liquidatus point for the slag.
The results thus indicate that slag apparent viscosity can be decreased
by increasing magnesium oxide content. From an economic viewpoint, the
MgO concentration should be maintained as low as possible and still
maintain good slag fluidity characteristics because MgO is not effective
for reaction with sulfur. Selection of the optimum MgO content will be
discussed later in the conclusions portion of this section.
Effect of Fluorspar on Slag Fluidity
It is well knownlU that calcium floride (fluorspar, CaF2) has a dramatic
effect on slag viscosity. In general, as the CaF2 content increases
slag viscosity greatly decreases. Because it is desirable to have
available a strong fluidity control additive, a series of experiments
were completed to determine the effect of CaF2 on the apparent viscosity
of slags containing both high magnesia and high sulfur.
In this work, the percentage of CaS and CaF£ was varied, but the silica
to alumina weight ratio was maintained at 2,5 and the calcium oxide to
magnesium oxide weight ratio was maintained at 1.8. Slag basicities
covered the range of 0.7 to 1.2 (see slags 5, 11, 12 and 13, Table VIII).
Fluidity data were obtained at a furnace temperature 2600°F.
Figure 23 presents results of this work in the form of a plot of percent
CaS versus percent CaF2 with a correlating iso-viscosity curve covering
the range of experimental apparent viscosities of six to eleven poises.
The pronounced effect of calcium fluoride addition in lowering apparent
viscosity of high sulfur bearing slags is readily observed. A comparison
of Figures 21 and 23 verifies this,, For example, at a 16 percent calcium
sulfide content in the slag, apparent viscosity (Figure 20) is approxi-
mately 27 poises. By the addition of eight percent calcium fluoride to
the same slag, apparent viscosity was decreased to the range of six to
eleven poise (Figure 23).
In general, the data show that if a more fluid slag is required, calcium
fluoride additions can easily compensate for the deleterious effect of
calcium sulfide on apparent slag viscosity. Calcium fluoride will
probably not be required to decrease slag viscosity in the operation of
the combustor. However, its effect is encouraging as it offers a means
of control when required.
-63-
-------
Effect of SiQ2/Al203 Ratio on Fluidity
Effect of coal ash chemistry (particularly the major constituents,
silica and alumina, on apparent slag viscosity was investigated. In this
study 1.2 basicity slag (Table VIII, slag Nos. 5, 14, 15) having a
constant CaO/MgO weight ratio of 3.6 was tested at a furnace temperature
of 2600°F<, The silica to alumina ratio was varied over the range of one
to eight„ Results are presented in Figure 24 and show that apparent
viscosity of the slag decreases as the silica to alumina ratio increases
and asymptotically approaches a minimum value of about four poise. This
effect is in agreement with general trends reported by Machin et al.
The data are of particular interest since they indicate that variations
in coal ash chemistry will not significantly affect the overall
viscosity of the slag. It is expected that the typical coal ash will
have a silica to alumina ratio varying in the range of about two to
three. Consequently, a change of this magnitude in the silica to
alumina ratio should present no more than a one poise variation in
slag viscosity,,
Engineering Design Recommendations
Experimental work completed on the fluidity of various slags for use in
the combustor has yielded significant operating guideline parameters.
Conclusions and recommendations which should serve as the design
criteria for an engineering and economic analysis of the process are as
follows:
lo Slag used in operation of the combustor should contain ten
to twelve percent MgO and have a lime to magnesia weight
ratio of about 1.8 to 2<,10
2, Silica to alumina weight ratio should be maintained as high
as possible, preferably above three.
3. Using these ratios, a calcium sulfide content of 22.5 percent
(ten percent sulfur) will result in a slag having an apparent
viscosity of 44 poise at 2600°F.
4. A typical slag composition meeting these requirements is:
CaO 21.5 percent, MgO 12.0 percent, CaS 22.5 percent, A1203 12.7
percent, Si02 31.3 percent.
5. Effects of slag composition on slag expansion due to gas
entrapment and ability of the slag to entrap solid particles
resulting from the coal and coal ash are not known. These
effects must be determined during actual operation of the
experimental combustor.
-64-
-------
16
0)
CO
o
PM
12
CO
o
o
CO
•r-l
>
4-1
d
CO
P.
Furnace Temp. 2600°F
Basicity 1.2
2468
Weight Ratio of Silica to Alumina
FIGURE 24 - THE EFFECT OF SILICA TO ALUMINA WEIGHT RATIO ON
APPARENT SLAG VISCOSITY
-65-
-------
SULFUR RETENTION BY SLAG
Introduction
Sulfur is introduced into the combustor via two streams of solids. The
first is the stream of sulfates recovered from the acid mine water
through the distillation and drying units. These sulfate-bearing solids
will be deposited into the combustor slag layer. The second source of
sulfur is coal that is injected beneath the surface of the molten iron
bath. The coal sulfur exists as pyrite and as organic sulfur bound
within high carbon content molecules. From an economic point of view,
it is desirable that all of the input sulfur be recovered as calcium
sulfide in the molten slag. Of particular importance, both from a
technical and economic viewpoint, is the distribution of sulfur between
the liquid slag and molten iron at steady state or equilibrium. This
distribution, termed the partition ratio, is defined as the percentage of
sulfur in the slag divided by the percentage of sulfur in the iron. To
maximize the operating credits associated with the combustor, it is
necessary that the partition ratio be as high as possible. This is to
say that the bulk of the sulfur should exist in the slag with very little
in the iron phase. If high partition ratios (20 or more) can be obtained
at relatively low excess lime, lime consumption will be low and operating
costs will be reduced. Additionally, as the partition ratio increases,
sulfur content in the iron will decrease which should improve the by-
product value of the iron produced in the process.
A considerable amount of literature is available concerning the
partition ratio and desulfurizing power of blast furnace and steelmaking
slags^» , in general, high partition ratios are associated with
increasing temperature, reducing atmospheres, low slag viscosities and
high slag basicities. However, the literature data are generally confined
to slags containing less than two percent sulfur. Unlike steelmaking
practice, the combustor will operate with a slag containing large
amounts (about ten percent) of sulfur. For such slags, no literature
data are available to estimate expected partition ratios.
Calcium sulfide solubility for a CaO-Al203~Si02 system has been evaluated
over a fairly broad range of compositions-^-^. Depending upon the region
of the slag system, sulfur solubility varied from about two to six
percent. For a CaO-Si02 system, saturation sulfur concentration increased
from about three to six percent with increasing concentration of silica
in melts liquid at 1550°C ^J15. No data were available for determining
the effect of undissolved CaS on partition ratio. Accordingly, a
laboratory investigation was completed to determine the partition ratio
and expected sulfur contents for both iron and slag when a high sulfur
bearing slag (CaO-MgO-Si02-Al203 system) was placed in contact with
molten iron. No attempt was made to measure the maximum solubility of
CaS in these slags. Rather, the work was directed to an applied research
effort to evaluate the overall effect of high sulfur slag (regardless
of solution state) on combustor operating parameters.
-67-
-------
Experimental Procedure
Two separate experimental procedures were used in the sulfur partition
ratio studies. The first procedure consisted of measuring equilibrium
partition ratios for a system employing sulfur-free slags and high-
sulfur molten iron. In these studies, a CaO-MgO-Si02-Al203 slag of the
composition given for slag G in Table XII was employed. Master batches
of this slag were prepared from reagent grade materials. The well
blended mixture was placed in graphite crucibles and heated to 2600 F
for five hours. The slag was then removed from the furnace, cooled,
crushed, and reheated to 2600°F. This procedure was repeated three
times ' to provide a slag of uniform composition. After three remelts,
the slags were ready for use in the partition ratio test. In these tests,
the molten iron bath was prepared by melting reagent grade iron powder
which was then saturated with graphite. Once the iron was saturated,
iron sulfide was added to the melt in varying proportions to produce
high sulfur pig iron. In later experimentation, the iron powder was
replaced with a four percent carbon pig iron to avoid the necessity of
saturating the melt with carbon.
TABLE XII
Slag Compositions Used in Partition Ratio Studies
Slag Weight Percent of Component
Identification CaO MgO SiOp Al 0 CaS Basicity
A 29.0 8.0 26.3 10.2 26.6 1.01
B 27.2 7.8 27.5 10.7 26.8 0.91
C 26.0 7.2 29.0 11.2 26.6 0.82
D 31.7 8.8 23.7 9.4 26.5 1.22
E 37.3 10.4 28.1 11.0 13.2 1.22
F 25.9 7.2 19.4 7.7 39.8 1.22
G 42.9 12.0 32.4 12.7 0.0 1.20
To obtain partition ratio data, a known weight of molten slag (previously
melted in a graphite crucible) was added to a graphite crucible con-
taining the molten iron. The ratio of slag weight to iron weight was
maintained at 0.3. The crucible containing the molten iron-slag mixture
was then covered with a graphite lid and maintained at a furnace temp-
erature of 2600°F. Furnace temperature was controlled to + 25°F of the
set point temperature. The furnace chamber was purged with argon for
duration of the run. At the end of the designated test time, the graphite
crucible was removed from the furnace and cooled to room temperature in
an argon atmosphere. The iron and slag samples were then removed from
the crucible and prepared for analysis. The slag was crushed to -100
mesh to free entrained particles of iron which were removed by dry
magnetic separation. Representative samples of both the slag and iron
were taken for carbon and sulfur analysis.
The second series of partition ratio tests differed from the first in
that sulfur-free iron was used. In these tests, sulfur in the form of
calcium sulfide was added to the slag. In this manner, the effect of
sulfur transfer from the slag to the iron could be established. Otherwise,
-68-
-------
the experimental procedure was identical to that outlined above.
Results and Discussions
Results of this work are presented in Figure 25 which shows the effect
of time-at-temperature on the approach to partition ratio equilibrium.
A partition ratio of about 35 was achieved after 50 hours in the furnace.
In general, the equilibrium partition ratio obtained from melts in which
the sulfur was initially contained within the iron are about ten percent
lower than that obtained from the reverse or high sulfur slag condition.
This is due to the difficulty of maintaining a four percent carbon
level in the molten iron. Sulfur removal from the iron is accompanied
by carbon consumption which in turn decreases the equilibrium partition
ratio that may be achieved. At time zero, when all of the sulfur is in
the slag, the partition ratio is infinite. Consequently, as soon as
sulfur is transferred from the slag to the sulfur-free metal, the ratio
drops extremely rapidly as shown by the upper curve of Figure 25. On
the other hand, when all of the sulfur is initially contained in the metal,
the partition ratio is zero and gradually builds as sulfur is transferred
to the slag (see lower curve of Figure 25). In either case, both
methods yield essentially the same equilibrium results.
Figure 26 presents the effect of contact time on the sulfur content of
iron (initially containing no sulfur) exposed to slags (Table XII, slags
D, E, F) containing about six to eighteen percent sulfur. At equilibrium
(about 50 hours) the small sulfur loss by the slag results in relatively
low iron sulfur content. In general, the sulfur content in the iron tends
to increase as the initial^ slag sulfur increases. However, even at an
initial slag sulfur content of 17.7 percent, the equilibrium percent
sulfur in the metal was only 0.25 percent.
Consequently, it can be conservatively estimated that in the commercial
operation of the combustor the iron produced will contain less than 0.3
percent sulfur.
Data on Figures 25 and 26 also present some interesting insight into the
slag and metal sulfur reactions that may be expected to occur in the
commercial combustor. Because the steady state sulfur content in the
metal is relatively low, it is unlikely that any of the sulfur introduced
into the slag by the solids recovered from the distillation unit will
enter the molten metal phase. Under commercial conditions there should
be sufficient sulfur and pyrites in the coal so that the bulk of the
transfer will occur from the metal to the slag. This would imply that
reduction of sulfate to sulfur will occur primarily in the slag phase
or at the slag-iron interface.
Because the combustor will operate commercially with a slag having a
basicity of about 0.8 to 1.0, a series of tests were completed to
determine effect of slag basicity on equilibrium partition ratio. For
this work, slags A, B, C, D, whose compositions are given in Table XII
were used. Results of this work are presented in Figure 27. Data
indicate that the partition ratio decreases slightly with decreasing
basicity. This is consistent with the literature5 which indicates that
-69-
-------
o
•rl
J-J
ID
d
o
•H
4-1
200
100
Conditions 2600°F
11.77. S in slag (1.2 B/A)50
or metal initially
O Sulfur in Slag
• Sulfur in Metal
25 50
Time, (Hours)
FIGURE 25 - THE EFFECT OF TIME ON THE APPROACH TO
PARTITION RATIO EQUILIBRIUM
0.30
0.20 -
0.10
n-
Initial Slag Sulfur:
O 5.9%
• 11.8%
O17.77.
25 50
Time, (Hours)
75
FIGURE 26 - THE EFFECT OF CONTACT TIME ON THE
SULFUR CONTENT OF IRON
-70-
-------
50
o
•1-1
4J
§ 25
•r-j
4-1
•i-l
P-t
Conditions: 2600°F
11.77o in slag initially
0.6 0.7 0.8 0.9 1.0 1.1 1.2
Slag Basicity
FIGURE 27 - THE EFFECT OF SLAG BASICITY ON EQUILIBRIUM
PARTITION RATIO
-71-
-------
better partition ratios are obtained at high slag basicities. For a
1.2 basicity slag, a partition ratio of about 45 was obtained. The
partition ratio decreased linearly with decreasing slag basicity until
at 008 basicity a partition ratio of about 38 was achieved. The
decrease in partition ratio was rather slight.
Engineering Designs Specifications
Based on the results of the partition ratio study, the following
recommendations are given:
1. At a slag to iron weight ratio of 0.3 partition ratios of
about 50 can be achieved,,
2o The partition ratio decreases with decreasing slag basicity
but the effect is rather slight. For a 0.8 basicity slag,
a partition ratio of about 35 to 40 will be obtained.
3. In the basicity range of 0.8 to 1.2, residual sulfur content
in the hot metal will be less than 0.30 percent. Slag sulfur
content will exceed ten percent.
-72-
-------
DETAILED SLAG CHARACTERIZATION
Based on the fluidity measurements of various types of slags, three
slags were chosen covering the range of basicities of 0.83 to 1.01 as
suitable for use in the operation of the combustor. In the slags,
the silica to alumina ratio and the lime to magnesia ratio were main-
tained constant at about two. The slags contained sulfur in the form of
calcium sulfide (16.9 to 18.0 weight percent). Composition of these
slags is presented in Table XIII.
These three slags were chosen as the basic slags for detailed study and
were used throughout the course of this work as described in the following
sections„
-73-
-------
APPARENT DENSITY OF HIGH SULFUR BEARING SLAGS
Experimental Equipment and Procedure
Two techniques were employed to measure apparent density of slags as a
function of particle diameter. Standard pychnometer and burette methods
were employed. Of the two, the bulk of the measurements were conducted
using the burette method. Within the limits of sample composition var-
iations, the latter method was considered as reliable as the pychnometer
technique, but much faster.
The experimental procedure consisted of filling a 25 ml burette to a
known level with carbon tetrachloride. The burette was graduated in
0.1 ml increments and was capable of being read to + 0.05 ml. Carbon
tetrachloride was used as the displacement medium to prevent any reaction
with the sulfur constituents in the slag. Once the burette was filled
to some level with liquid, a known weight of solids was introduced into
the fluid. From the solids' weight and volume change read from the
burette, the apparent solids' density was calculated.
Discussion of Results
The effect of particle diameter on apparent density of 0.82, 0.90, and
1.01 basicity slags shown in Table XIII is presented in Figure 28. All
slags show the same effect in that particle apparent density increases
as the average diameter decreases. Within the limits of experimental
error, the two lower basicity slags (0.82 and 0.90) yield a common curve.
However, the higher 1.01 basicity slags yield a significantly lower
apparent density. Although the reason for such behavior is not known
conclusively, the difference is attributed to variations in the crystal
structure upon solidification „ In the relatively narrow range of
basicities 0.82 and 0.90, the crystal structure of the solid phases would
tend to resemble each other. However, as the basicity was increased to
1.0 and beyond, the crystal characteristics would shift as estimated from
phase diagrams for analogous systems-'.
TABLE XIII
Slag Compositions Used in Characterization Studies
Composition, weight percent Ratios
Basicity Si02 A1203 MgO CaO CaS Si02/Al203 CaO/MgO
0.82 30.0 15.0 12.3 24.7 18.0 2.00 2.01
0.90 28.6 14.4 13.0 26.0 18.0 1.99 2.00
1.01 27.6 13.8 13.9 27.8 16.9 2.00 2.00
-75-
-------
3.2
3.0
o
o
2 2.8
4-1
a
2.6
0 1.01
© .90
• ,8?
I
I
1
.01 .1 1.0
Average Particle Diameter, mm
FIGURE 28 - THE EFFECT OF PARTICLE DIAMETER ON APPARENT DENSITY
-------
VISCOSITY MEASUREMENTS
Introduction
Surface tension and viscosity data are not available for the high sulfur
bearing slags that will be used in the combustor. To reasonably estimate
the heat and mass transfer that occurs within the combustor, an
experimental program was initiated to measure these physical properties.
For these measurements, an oscillating-bob viscometer^'H was constructed
and a maximum bubble-pressure device-*-" for surface tension measurements
was designed.
The theory and technique for determining viscosity and surface tension
values will not be discussed in this report, as detailed explanations
can be found in the literature » » . Briefly, the viscosity measure-
ment is obtained by the logarithmic decrement method^, in this
technique, an initial torque is given to a wire from which a platinum
bob is suspended and immersed in. the fluid whose viscosity is to be
measured. Depending upon the magnitude of the viscous forces resisting
the rotation of the bob, a characteristic dampened oscillation frequency
is obtained. The decrement, or the ratio of two successive dampened
oscillation amplitudes, is a function of the fluid viscosity.
In the maximum bubble pressure surface tension measurement, a platinum
tube is immersed vertically into the molten slag. Argon is forced
down the tube and the maximum pressure achieved prior to release of a
gas bubble is recorded. The surface tension of the fluid is a function
of the gas pressure reading and the depth of immersion of the tube in the
fluid16.
Experimental Procedure
Equipment (see Figure 29) used for the viscosity measurements was
similar to that employed by Machin and Hanna-'--'-. It was of the oscil-
lating or torsion pendulum type. One major modification was made to
the assembly which related to the method by which the angular dis-
placement or oscillation was measured. Machin and Hanna used a mirror
that reflected light to measure the angular displacement. In this
study, angular displacement was measured by the interruption of a
photocell whose output was made to be proportional to the radial de-
flection of the suspension torque wire. Except for this and some minor
modifications with regard to the length of the torsion wires and more
convenient means for setting the vibrating system in motion, other
features were comparable.
The oscillating bob and crucible were made from platinum. The equipment
dampening constant which is a function of the shape and dimensions of
the bob and crucible was evaluated by calibrating the equipment in an
oil of known viscosity, which covered the range of 0.6 to 81 poises as a
function of temperature. The calibration work showed that the assembled
apparatus used in this work could effectively cover the range of eight
to 40 poise.
-77-
-------
Initial Torque
Solenoid
Torsion Wire
Leveling
Arm
Bob in Slag
ample
O 00
Furnace Heating Elements
Light Source
Photocell
Photocell
Recorder
FIGURE 29 VISCOSIMETER
-78-
-------
Discussion of Results
During the first experiments to determine viscosity of high sulfur con-
tent slags (slag No. 3 Table VIII) a severe operating problem was en-
countered. The platinum crucible and bob were severely attacked by the
molten high sulfur bearing slag and resulted in total failure of the
equipment. The severe corrosive attack of the slag on the platinum
caused holes in the crucible walls. The platinum bob was equally
damaged and was of no further use in the experiments.
New crucibles and osciallating bobs were constructed from pyrolitic
graphite rods. Viscosity tests depend on the liquid wetting the walls
of the container and the bob to cause the liquid to shear upon itself.
Since slag does not wet graphite, the internal surface of the crucibles
and the external surface of the oscillating bob were corrugated. The
corrugations were made parallel to the longitudinal axis of both the
crucibles and the bob in the form of deep grooves, approximately %-inch
deep by 3/16-inch wide. When slag filled the grooves of the bob and
crucible, the rotation of the bob caused the slag to shear upon itself.
To determine whether this method was feasible, the equipment was recal-
ibrated using the oil mentioned previously. Tests were then completed
on molten slags of known viscosity^ (slag No. 7, Table VIII). These
measurements were conducted at a furnace temperature of 2650°F. Actual
slag temperatures, as measured by the immersion of the platinum rhodium
thermocouple into the slag, indicated a slag temperature of 2570°F,
approximately 80 degrees lower than the furnace temperature. A number
of viscosity measurements were made at this temperature to determine the
effect of time on slag viscosity. It was known that significant air
infiltration existed in the furnace and was oxidizing the graphite rods.
Consequently, the time measurement was taken to determine the effective
length of time that a graphite rod would be employed without severely
influencing the viscosity measurements.
These results are shown in Figure 30. The initial viscosity measurement
was the highest (approximately 16) when the graphite bob was first in-
serted into the furnace. This agrees reasonably well with the literature^
for sulfur free slags of this composition, where the reported viscosity
at 2552°F is 14 poise. As time increased, viscosity decreased rather
rapidly. This was primarily attributed to the oxidation of the graphite
rod which had the effect of lowering the equipment constant^ It should
be mentioned that the furnace was continually purged with argon to
minimize air infiltration but air entry could not be prevented entirely.
Because of the expense and work required to machine the rods and
crucibles, the argon purge was increased in an effort to minimize oxidation.
High sulfur bearing slags were then evaluated to determine the effect of
sulfur content on slag viscosity. In these measurements, and at the
argon flow employed, a maximum furnace temperature of 2650°F was achieved.
Attempts to measure the actual slag temperature were of no avail as the
thermocouple was immediately attacked and destroyed by the high sulfur
bearing slags. Based on work with sulfur-free slags, it is estimated that
the slag temperatures were approximately 100°F lower than the indicated
-79-
-------
20
15
o
a.
4-)
•r-l
CO
O
o
to
•r-l
10
50 100 150 200
Time from first viscosity measurement, Minutes
FIGURE 30 - EFFECT OF CARBON BOB RESIDENCE TIME IN
FURNACE ON MEASURED VISCOSITY
-80-
-------
furnace temperature. At these low temperatures, the bulk of the
viscosity measurements was outside the range of the instrument.
However, measurements of 0.82 basicity slag show that viscosity
increases with increasing calcium sulfide content. At an estimated slag
temperature 2550°, viscosities of 16, 22 and 40 poise were obtained at
calcium sulfide concentrations of zero, six and twelve percent respectively.
At higher calcium sulfide contents, viscosity was outside the range of
the viscometer.
At the higher argon gas flow rates employed in this latter work, a crust
was observed on the surface of ths slag in the crucible. When the
oscillating bob was introduced in the crucible, force had to be exerted
to break through the top of the slag. Consequently, it is highly probable
that the surface of the slag was receiving some cooling from argon
introduction. After considerable stirring, a fluid surface was main-
tained and viscosity measurements were completed. However, it is
probable that some solid particles did exist in the slag. Since this
method of measurement is subject to severe perturbations in a two-phase
system, viscosity measurements can be only estimates. Based upon freezing
point data, an estimate of the actual slag temperature can be made.
Freezing points of 0.82, 0.9 and 1.01 basicity slag were found to be
2450, 2340, and 2447°F respectively. Based on this knowledge and the
temperature difference measured by thermocouples inserted in sulfur-free
slag, it is probably true that the slag temperature was in the neighbor-
hood of 2550°Fo Because slags of this nature will experience a halving
of the viscosity for 100°F rise in temperature^ it is reasonable to
expect that slag viscosity at an actual temperature of 2600°F will be
about 40 to 50 poise, at a calcium sulfide content of about 20 percent.
Because of the sulfur attack on the platinum equipment, no surface tension
measurements were attempted.
Engineering Design Recommendations
Because of the experimental difficulties encountered in measuring the
viscosity and surface tension of high sulfur bearing slags, no reliable
absolute measurements were possible. However, the viscosity data tend to
support conclusions and recommendations derived from fluidity measure-
ments reported in an earlier section.
-81-
-------
EXTERNAL SURFACE AREA
Introduction
As a part of the detailed slag characterization study, external specific
surface area was determined for each of the three slags of Table XIV as
a function of particle size. The external surface area measurements
were a useful correlating factor in kinetic studies for desulfurization
and neutralization of acid mine water. This section presents the results
obtained from the experimental investigation of the external surface
areas of high-sulfur bearing slags.
TABLE XIV
Comparison of Surface Areas for Glass Beads
Determined by Air Permeability and Micrometer Methods
Glass Bead Specific Surface, sq cm/gram
Diameter, mm Air Permeability Micrometer
3.15 7.9, 7.7, 7.4 7.6
4.06 5.5, 5.8, 6.1 5.9
5.95 3.9, 4.1, 4.4 4.1
Theory
The theory relating the specific surface of solids to the pressure drop
obtained in the laminar flow of fluids through packed beds of granular
materials was first formulated by Kozeny . His equation relates the
volumetric flow of fluid to the pressure drop established per unit
length of bed, the specific surface of the solids, the porosity of the
bed, the viscosity of the fluid, and the cross-sectional area of the
packed column. Since all quantities are easily measured, the specific
surface of the materials is readily obtained by measuring the appropriate
variables and calculating the specific surface from the equation:
s = 14 ( E3 )^
d (Pv (1 - E)2)
where, d = apparent density of the solids, grams/cu cm
E = porosity
v = kinematic viscosity of fluid, cirr/sec sq cm/sec
S = specific surface, sq cm/gram
P = permeability of packed bed, cm/(sec)
The above equation is valid provided laminar flow exists through the
packed bed.
Equipment and Experimental Procedure
Equipment used to measure the external specific surface area by air
permeability methods is well known and will not be presented here1'.
Essentially, the apparatus consists of a flow meter, an inclined mano-
meter and a graduated (volumetrically) sample holder of known cross-
sectional area.
-83-
-------
The experimental procedure was as follows. A known weight of solids
(whose apparent density is known) was poured into the graduated glass
sample tube holder and slightly vibrated to maximize packing of the
solids. Sample volume and height was determined from the graduation
and cross-sectional area of the sample tube holder. Bed porosity
was calculated from known weight, density and volume of the packed bed.
A known volume of air was then caused to flow downward through the
packed bed. Pressure drop was measured across the bed once flow reached
a steady state. In general, pressure drop measurements were obtained for
five different flow rates„ For the equipment used in this work, laminar
flow occurred when the air flow rate was maintained at less than five cc
per second„ From the pressure drop and flow rate data, the permeability
constant for the system was calculated. Using this value along with the
kinematic viscosity of air, porosity and apparent density of the solids
the specific surface was calculated using the equation presented earlier.
Discussion of Results
Before conducting surface measurements on the synthetic slag, equipment
was standardized by measuring the surface areas of glass beads. In this
work, glass beads having nominal diameters of 3, 4 and 6 were employed.
Specific surface of the glass beads was determined initially by micrometer
measurements of the diameter and by obtaining a particle count per unit
weight for each size of glass bead. In this manner, it was possible to
calculate their specific surfaces in terms of square centimeters surface
area per gram of sample. A comparison of the specific surface obtained
by micrometer measurement and air permeability methods is presented in
Table XIV„ Agreement between the two methods was excellent.
The effect of slag basicity and particle size on external specific surface
for the synthetic slags is presented in Figure 31. The data show that a
difference exists in the external surface areas for each of the slags
provided that the particle diameter is greater than 0.2 mm. At the
smaller particle sizes (within the limits of the experimental error) the
data are adequately defined by one common curve.
It is not clear why there should be a divergence in the specific surface
area measurement at the larger particle sizes. Based on the apparent
density data presented earlier, it would be expected that the lower
basicity slags would have the least specific surface since they were
heavier. However, this density variation is not sufficient by itself to
account for the spread in the data. Differences are probably the result
of significant variations in fracture characteristics of the materials.
-84-
-------
S
co
n
bO
B
o
cr1
en
0)
O
to
o
T-l
4-1
•d o
OJ
CD
Cl
w
1.01
0.9
0.82
10
iiii
0.04
0.1 1.0
Particle Diameter, millimeters
FIGURE 31 - THE EFFECT OF SLAG BASICITY AND PARTICLE SIZE
ON EXTERNAL SPECIFIC SURFACE
-85-
-------
TOTAL SPECIFIC SURFACE AREA
Introduction
As part of the detailed characterization work, total specific surface
area for each of the synthetic slags shown in Table XIII was determined.
Surface area measurements were obtained with a Perkin Elmer Shell Model
212 D Sorptometer.
The principle of the surface area measurement is based on measurement of
an amount of gas adsorbed by the solid sample. In this method a known
mixture of nitrogen and helium is passed over the sample in a sample
tube and the effluent is monitored by a thermal conductivity detector.
While the gas is flowing to the sample, the sample tube is cooled by
immersion in a bath of liquid nitrogen. The cooled sample adsorbs a
certain amount of nitrogen from the gas stream which is indicated on a
recorder chart as a peak. The area of this peak is proportional to the
volume of nitrogen adsorbed. After equilibrium is established, the
recorder pen returns to its original position and the liquid nitrogen gas
is removed from the sample tube. As the sample tube warms, the adsorbed
gas is released and enriches the effluent gas passing through the sample
tube,, A desorption peak is then obtained which is in the reverse direction
of the adsorption peak. When desorption is complete, a known volume of
nitrogen is added to the nitrogen helium stream and the resulting
(calibration) peak is recorded. By comparing the areas of the desorption
and calibration peaks, the volume of nitrogen adsorbed by the sample can
be calculated.
Discussion of Results
Effect of slag basicity on the total specific surface area is., presented
in Figure 32. The specific surface decreases from about 5.4 square
meters per gram for the 0.82 basicity slag to a low of about 1.5 square
meters per gram for the 1.01 basicity slag. It appears that the higher
basicity slags, because they are more fluid in the liquid conditions,
tend to yield more glass-like solids with low porosity. Attempts
were made to measure pore size distribution of the slags but to no avail.
Discussions with the manufacturer of the Sorptometer indicated that at
these low surface areas and porosities, the equipment was being operated
at its minimum reliability levels. To overcome this problem, different
cells would have to be purchased to accommodate our materials. Inasmuch
as the total surface area measurements did not appear to be a good
correlating variable for other work, purchase of additional equipment was
not made. Consequently, pore size distributions were not made.
-87-
-------
00
CO
co
6
CO
M
60
TO
dJ
J-l
01
O
CO
CO
O
H
1.0
Basicity
FIGURE 32 - THE EFFECT OF BASICITY ON TOTAL
SPECIFIC SURFACE AREA
-88-
-------
CRUSHING ENERGY REQUIREMENTS
Introduction
Sulfur recovery from slag and acid mine water neutralization with slag
can both be accelerated by crushing the slag to a smaller size. There-
fore, the economies of these operations can be affected by cost of
crushing or grinding slag. Because both the capital cost and the oper-
ating cost for crushing equipment depend on the crushing energy required,
a study of crushing energy requirements for slag was undertaken.
Experimental Equipment and Procedure
A modified version of the drop weight machine used by Gross25 was adopted
for this work. A schematic diagram of the apparatus is presented in
Figure 33. The crushing chamber was housed in a cylindrical die made of
stainless steel construction. The crushing chamber consisted of a
cavity located within the die and was 6.1 centimeters in diameter. A
smooth-fit stainless steel plunger was used to transmit the energy
derived from the falling 2.62 kg drop weight to the sample contained
between the plunger and the base plate. Walls between the plunger and
the die were lubricated with graphite to minimize plunger friction.
The test procedure was relatively straightforward Before beginning a
crushing energy test the external specific surface area of the test solids
was determined by the air permeability method. From this sample, a
known weight of material was taken and introduced into the die. The
plunger was inserted within the die and rested on top of the bed of
solids. Once the die was assembled, the entire apparatus was placed on
a circular piece of aluminum wire (0.0640 inches thick) that had been
previously centered directly beneath the drop weight. The aluminum wire
had previously been calibrated to determine the energy absorbed by the
wire as function of wire diameter after deformation. In this manner, it
was possible to determine the energy absorbed by the solids and the energy
transmitted to the wire. In practice, the drop weight was raised by
means of a string mounted on an overhead pulley to a known height above
the plunger resting on the test solids. Once the elevated drop weight
had finished oscillating and was stationary, the string was cut to permit
the hemispherical drop weight to impact on the center of the plunger.
After impact, the aluminum wire was removed from beneath the die and
measured at seven positions to determine deformation of the wire. Knowing
the kilogram-centimeters of energy input by the drop weight and amount
of energy absorbed by the wire, it was then possible to calculate the
energy absorbed by the impacted solids. If additional work was required
to crush the solids still further, a new aluminum ring was positioned
under the die and the drop weight was once again raised and released.
Crushed solids were removed from the die and weighed to determine weight
loss if any (usually less than 0.5 percent). A new external specific
surface measurement was then conducted to determine increase in surface
area after crushing. Knowing the new surface area generated and the energy
- 89-
-------
Drop Weight
Plunger
Base
V
Die
.Slag Sample
-Aluminum Wire
FIGURE 33 - CRUSHING ENERGY APPARATUS
-90-
-------
input, it is then possible to calculate the crushing energy or Rittinger's
number.
Calibration of Crushing Energy Equipment
The thickness of the aluminum wire used to measure the transmission of
energy from the die to the floor upon which the die rested was measured
at seven different equally spaced positions before and after each test.
Consequently, hundreds of measurements were obtained on the average
diameter of the wire before deformation. The average diameter of the
wire was 0.0640 plus or minus 0.0002. Less than one half of one percent
of the data fell outside of this range (within plus or minus 0.0005).
Uniformity of the initial starting diameter of the wire resulted in
excellent reproducibility of the calibration curve data. In these
calibration experiments, no solids were introduced into the die. To
obtain deformation data, the drop weight was elevated to varying heights
above the plunger and released. The height from which the drop weight
was released was varied over the interval of two to twenty centimeters.
In the event that the aluminum wire was not evenly deformed after
releasing the drop weight, the test was discarded with the assumption
that a direct on-center high of the drop weight was not achieved.
Consequently, it was likely that some binding of the plunger occurred
with a resulting non-measureable energy loss attributed to friction.
Results of this work are presented in Figure 34.
The figure shows a linear relationship between deformation and energy
input up to about 40 kilograms-centimeters. Because of alignment
difficulties between the plunger and drop weight at higher energy inputs,
data reproducibility became poorer. Accordingly, in all subsequent
experimentation the weight was not raised to an elevation greater than
15 centimeters.
Discussion of Results
Effect of slag basicity on grinding energy requirements is presented in
Figure 35. In this work, slags having basicities of 0.82, 0.90 and
1.01 were used. In each of the crushing tests initial particle size of
the slag was -10, + 20 mesh. Results show that the new surface area
produced per unit of energy absorbed increases with increasing slag
basicity. This would indicate that higher basicity slags are more
readily crushed and can be accommodated in smaller sized equipment.
To obtain a relative ranking and a reference point on the grindability of
the slag as compared to a standard material for which equipment has
already been sized, the grinding energy requirements were determined for
a silica sand (Ottawa sand). The results of this work along with the
comparative values obtained from the slags is presented in Table XV.
It is evident from the higher new surface area generated per unit energy
absorbed for the silica sand, the slags produced in this work are
slightly more difficult to grind. However, as the synthetic slags
produced in this work were very dense and since it is expected that the
slag produced during the actual operation of the combustor will be more
porous, the sizing of crushing and grinding equipment can be safely
-91-
-------
co
a)
J3
o
d
H
a
o
•H
0)
Q
i
B3
•i-l
P
01
!-l
•r-l
.0640
.0630
.0620
.0610
.0600
.0590
.0580
.0570
0 10 20 30 40 50
Energy Input, kilograms - centimeter
FIGURE 34 - CALIBRATION CURVE FOR ALUMINUM WIRE USED IN
CRUSHING ENERGY TEST
-92-
-------
O
CO
M Cl
!-J 0)
cu u
•H M
C O
-------
based on silica sand characteristics. This assumption should facilitate
the selection of suitable commercial grinding equipment.
TABLE XV
A Comparison of Crushing Energy Requirements for Silica and Slags
Average Crushing Energy
Material sq cm/Kg-cm
Silica 3.32
1.01 Basicity Slag 2.99
0.90 Basicity Slag 2.60
0.82 Basicity Slag 2.38
It should be pointed out that the new surface area generated per unit
of energy absorbed by the solids for silica sand are at some variance
with the literature. Work done by Gross^-° indicates a value of 17.56 as
compared to the 3.32 square centimeters per kilogram centimeter obtained
in this work. This variation is attributed to the difference in the
method by which the new surface area generated was measured. In the
literature work, a rate-of-solution method rather than air permeability
techniques was used to determine surface areas. The rate of solution
method will yield higher surface areas because of penetration of the
solvent into the interstices of the particle. Air permeability techniques
will result in establishing only external surface area variations and
afford little opportunity for any significant penetration within
the particle. Consequently, surface area measurements by the latter
technique will tend to be low by comparison.
It is of interest to point out that the new surface area generated per
unit of energy absorbed by the solid as determined in this work agree
reasonably will with literature values for measurements obtained on
commercial unitsl9o por example, values reported for grinding quartz in
various size ball mills range from 2.6 to 6.8 square centimeters per
kilogram centimeter as compared with the value of 3.32 obtained in this
work.
Engineering Design Recommendations
Based on results of the crushing energy requirement study, slag produced
in this work can be assumed to have grinding characteristics comparable
to silica. This criterion should facilitate the selection of commercial
crushing and grinding equipment for use in the process.
-94-
-------
HEAT CAPACITY
Introduction
To adequately define the heat balances in the process, it is necessary to
have reliable data on the heat capacity (specific heat) of the slag
at the temperature at which slag is used. This section presents the
experimental work conducted on the three slags (Table XIII) selected as
typical of those that may be employed in the operation of the combustor.
Experimental Equipment and Procedure
Standard calorimetry equipment and procedures were used to determine
specific heat of the solids. The calorimeter was a one-liter capacity
vacuum bottle,, To preclude the possibility of slag reaction with water,
ethylene glycol was used as the heat absorption liquid. A laboratory C.P.
grade of glycol was used. Specific heat data on the ethylene glycol were
obtained from the literature . A Beckmann differential thermometer
graduated to 0001°C and capable of interpolation to 0.005°C was used to
measure the temperature rise of glycol. A standard laboratory thermometer
used to measure the end point temperature of glycol and the cooled solids
was accurate to within 0.5°C.
Heat Capacity
Before beginning the experiments, the heat absorption value for the
calorimeter, Beckmann thermometer and stirer were determined. This was
done by use of the method of mixtures whereby hot and cold water were
mixed in the calorimeter to determine the steady state temperature. From
the results, heat absorption and losses were determined. Once the
calorimeter constant was evaluated, the method was standardized by
determining the specific heat of alumina of temperatures up to 1100°C.
A comparison of the experimental and literature values for the specific
heat of alumina over the temperature range studied permitted an estimate
to be made of the heat losses experienced during the transfer of the hot
alumina sample from the furnace of the calorimeter. The alumina was a
thin disc of 1 inch diameter by approximately 1/8 inch thick. Since the
slag samples were available in the form of thin discs of comparable size
and shape, a reasonable estimate of transfer heat losses for the slag
could be obtained from the alumina studies. Alumina data show that
radiant heat transfer losses and vaporization of glycol will introduce no
more than a seven percent error in the specific heat measurement. At a
temperature of 1100°C the measured heat content of the alumina disc was
seven percent lower than that calculated from literature values. The
major proportion of heat losses were associated with vaporization of the
ethylene glycol.
Experimental procedure for the slag samples consisted of briquetting a
known weight of slag into a cylindrical disc of one inch diameter and about
1/8 inch thickness. The slag disc was then inserted within a ceramic
combustion thimble. The weight of the thimble varied between 60 and 70
grams and the ratio of thimble weight to slag sample weight ranged from
four to eight. Relatively massive thimbles were employed to minimize
heat losses during the transfer of the disc from the muffle furnace to
-95-
-------
the calorimeter. Before the heated solids were placed in the calorimeter,
a known volume of ethylene glycol (500 grams) was introduced into the
vacuum bottle and allowed to temperature equilibrate. The Beckmann
differential thermometer was then adjusted to read near the bottom of
the scale so that a maximum temperature increase in the ethylene glycol
of 5.2°C could be obtained. In the event the temperature rise of the
glycol exceeded the range of the thermometer, the test was discarded.
Once the solids were transferred from the thimble to the calorimeter,
fluid agitation was maintained by means of a glass stirrer. To establish
an absolute value for the final temperature of the fluid and solids, a
standard laboratory thermometer was employed.
The heat content per gram of slag as the function of furnace temperature
was calculated from the initial and final temperatures of the slag, the
calorimeter constant, and the temperature increase of the ethylene glycol
along with the weights of materials involved. Heat content was measured
above the final slag temperature in the calorimeter. A polynomial fit
was then developed for the experimental data and the resulting equation
was differentiated with respect to temperature to obtain a second equation
defining specific heat of the slag as a function of temperature.
Discussion of Results
Effect of slag basicity and temperature on the heat content (calories
per gram above 28°C) of the solid materials is presented in Figure 36A.
Variations in heat content between slags of different basicities were not
observed. Within the limits of experimental error all three slags yield
the same values. Data were then correlated by means of the polynomial curve
shown in Figure 36B. The data correlate reasonably well. The resulting
polynominal was differentiated to arrive at the specific heat capacity
of the slags as a function of furnace temperature. This result is
presented in Figure 36B. For comparative purposes, literature values^,
two types of compounds comparable to those that may be formed in the slags
are also presented. The literature values effectively bracket the range
of specific heat capacities obtained in this work. At temperatures of
900°C, the slope of the experimental curve tends to approach zero. This
is attributed to radiation and glycol vaporization losses experienced
at the higher temperature levels which would tend to offset the heat
capacity increase with increasing temperatures. The experimental heat
capacity data should prove to be adequate for design purposes.
Engineering Design Recommendations
The specific heat capacity of high-sulfur bearing slags covering a range
of basicities of 0.8 to 1.0 is adequately defined by the curve of Figure 36A.
These data cover the temperature range of 200 to 1000°C. Values obtained
at temperatures in excess of 600°C may be on the conservative (low) side.
The equation correlating the heat capacity as a function of temperature
is as follows:
Cp = 0.196 + 2.02 x 10"4T - 9.33 x 10'8T2
where Cp is in calories/degree C/gram, T is in degrees C.
-96-
-------
300
e
to
I-l
bO
200
C
O
O
100
300 600 900 1200
Furnace Temp. °C
FIGURE 36A - THE EFFECT OF TEMPERATURE ON HEAT CONTENT OF SLAG
o
o
S
to
ji
to
0)
•H
J3
to
o
«
01
o
0)
3.0
2.8
2.6
2.4
2.2
CaAl2'S.,02 (Ref. 28)
Ca Mg (8^)2 (Ref. 28)
250 500 750
Furnace Temp. C
1000
FIGURE 36B - THE EFFECT OF BASICITY AND TEMPERATURE
ON SPECIFIC HEAT CAPACITY
-97-
-------
ACID MINE WATER NEUTRALIZATION WITH SLAG
Introduction
Neutralization of acid mine water is usually not recommended prior to
distillation or reverse osmosis treatment. However, the process generates
a neutralizer at no additional cost. This neutralizer is the slag
generated in the combustor, which has a high lime content after processing
for sulfur recovery.
Depending upon neutralization characteristics of desulfurized slags, it
is possible that neutralization of acid mine water prior to distillation
could lower the overall process capital and operating costs. For this
reason, continuous and batch neutralization tests were carried out on
various types of slags to determine their effectiveness in neutralizing
acid mine water. This section presents results of this work.
Experimental
Experimental equipment used in the continuous neutralization studies is
presented in Figure 37. The packed bed reactor containing slag particles
was made of 3.06 cm I.D0 plexiglass tube flanged on both ends. The
flanges contained fritted glass discs. The packed bed rested on the bottom
disc and the top disc was used to prevent the elutriation of solids
with the effluent stream,, Fresh acid mine water was caused to flow
upward through the bed of solids by means of a variable speed pump.
Samples of the liquid leaving the fixed bed were obtained by inserting
five ml pipette through the top flange of the reactor. Samples were
obtained at two minute intervals over the entire run and back-titrated with
sodium hydroxide to determine equivalent calcium oxide consumption in
the slag. Samples were also obtained from the effluent stream to
determine pH as a function of time.
In addition to the continuous neutralization test, batch neutralization
studies were also completed. In the batch test program, a standard
procedure using 200 ml of acid mine water and approximately 0.40 grams
of slag was adopted. In these tests a slag sample of known weight and
particle size was placed within a 400 ml beaker along with 200 cc of
acid mine water and agitated by means of a magnetic stirer. Measurements
of pH were obtained as a function of time over the entire test.
In this work a standard acid mine water composition was adopted as
proposed by the Environmental Protection Agency. This composition is
presented in Table XVI.
To determine the percentage of slag alkali content that is effectively
utilized for neutralizing acid mine water, the following test procedure
has been employed to rank the various slags. A slag sample (about 2.5
grams) was weighed on an analytical balance and transferred to a 250 ml
beaker containing 15 ml of water. The beaker was covered and the water
heated to boiling. The hot mixture was allowed to stand for several
minutes to slake any free lime that may be contained within the slag.
-99-
-------
Sampling Port
CO
cu
too
a
CO
Slag Sample
5 gallon
Spent AMW
Reservoir
3.06 cm Dia
Glass Column
Rotameter
A Variable Speed Pump
5 gallon
Fresh AMW
Reservoir
FIGURE 37 - CONTINUOUS NEUTRALIZATION APPARATUS
-100-
-------
A burette was then used to add a quantity of 0.5 N sulfuric acid to
react with the lime. Once the end-point had been reached, 100 ml of
sulfuric acid was added in excess. The mixture was then washed into an
erlenmeyer flask, boiled for 15 minutes, cooled to room temperature and
titrated with 0.5 N sodium hydroxide to the phenolphthalein end point.
Knowing the amount of sulfuric acid, sodium hydroxide and the weight of
the sample used in the test the weight of equivalent CaO per gram of
sample was calculated. Both CaO and MgO contained within the slag were
expressed on a CaO basis.
TABLE XVI
Standard Acid Mine Water Composition
Component Concentration (Mg/1)
Ca (as Ca) 80
S04 1061
Mn 2
Al 5
Fe (Total) 200
Mg (as Mg) 24
Acidity (Ca C03) 400
Effectiveness of Slag Alkali for Neutralization
Using the alkali effectiveness test procedure described above, an alkali
utilization factor was defined. This factor (percent alkali utilization)
measures the percentage of total slag alkali that is effective for
neutralizing sulfuric acid in the specified test procedure. Thus a
100 percent alkali utilization indicates that all of the calcium oxide
and magnesium oxide contained within the slag is effective for neutraliz-
ation.
The effect of basicity ratio and slag particle size on the percent alkali
utilization factor is presented in Figure 38. The data indicate that,
as the basicity ratio increases for any given particle size of slag, the
efficiency of alkali utilization increases. For a 1.5 basicity slag,
100 percent of the alkali (CaO + MgO) contained within the slag enters
into the neutralization reaction provided the particle size of the slag
is smaller than 60 mesh. As the particle size increases, the utilization
factor decreases rapidly. For a given particle size, the alkali contained
in the lowest basicity slags (0.8) is least effective for neutralization
Rather strong neutralization test conditions were employed in these
experiments. Consequently, the effect of basicity ratio on the alkali
utilization factor tends to be somewhat diminished. Under milder
neutralization conditions it can be expected that the effect of basicity
and particle size on alkali utilization factor will be more pronounced.
Nevertheless, the test is indicative of the relative ease with which
different slags can react with acids.
In the commercial operation of the combustor, it is probable that the
slag will contain silica and alumina in the ratio of two to one. However,
-101-
-------
= 2.5
•H 100
CO
80
•i-i
i—i
•1-1
4-J
P
•i-l
"3 60
60
co 40
20
Slag Size
O 140/200 Mesh
X 60/100 Mesh
D 10/20 Mesh
/ 6/10 Mesh
0.8
1.0
Basicity
1.2
1.4
FIGURE 38 - THE EFFECT OF SLAG PARTICLE SIZE AND BASICITY RATIO
ON SLAG ALKALI UTILIZATION IN NEUTRALIZATION
a
o
w 100
TO
•H
JJ
80
•r4
•-<
JS 60
t—i
<
60
^ 40
20
1.2 Basicity
O 140/200 Mesh
X 60/100 Mesh
O 10/20 Mesh
/ 6/10 Mesh
Si°2/A12°3
FIGURE 39 - EFFECT OF SLAG PARTICLE SIZE AND S
ON THE SLAG ALKALI UTILIZATION FACTOR
-102-
-------
because of the flux characteristics and variations in coal ash, this
ratio may vary substantially. Accordingly, a brief study was completed
using a 1.2 basicity slag to determine effect of particle size and silica-
to-alumina ratio on the alkali utilization factor. Results of this work
are presented in Figure 39. Higher alkali utilization is obtained as
particle size decreases. The silica to alumina effect is somewhat analagous
in that alkali utilization increases with decreasing Si02/Alo03 ratio.
This is probably due to an occlusion factor resulting from the formation
of more glass at the higher Si02 concentrations. Inasmuch as commercial
slag will contain a ratio of about two, alkali utilization will not be
severely impeded.
Continuous Neutralization Tests
For the continuous neutralization studies, synthetic sulfur-free slags
were employed in the equipment shown in Figure 37. Because of the
relatively high capital and operating costs associated with grinding
hard, abrasive materials to fine sizes, comparatively coarse particles
were used in this work. It was reasoned that since a wide particle size
distribution would be obtained in crushing the slag, it is only necessary
to study the neutralization rate for the top size, or coarse particles.
In this manner, if the neutralization reactor design were based on the
top size, a safe and conservative approach to the engineering design
would be possible because smaller particles will be consumed quite
rapidly.
In this work, two particle sizes were employed: -% + 10 mesh and -10
+ 20 mesh. A slag of 1.2 basicity (slag No. 5 Table VIII) was used for
the reacting material. Results of the neutralization study are presented
in Figure 40 which shows the effect of particle size, acid mine water
flow rate, and throughput on slag consumption. Slag consumption is
expressed in terms of equivalent moles of calcium oxide consumed per
second per square centimeter of slag external surface. The data show
that depending upon flow rate, the neutralization reaction is controlled
by diffusion either through the liquid phase surrounding the slag
particle or by diffusion of the acidic constituents through the slag
particle itself. In Figure 40, the rate of equivalent lime consumed per
unit time per unit of external surface area is independent of flow rate
provided that the acid mine water flow was maintained in excess of 370cc
per minute. This implies that the liquid diffusion of acidic constituents
through the liquid medium surrounding the slag particle is not the con-
trolling mechanism., Instead, the data show that neutralization rate is
controlled by diffusion through the interior of the slag particle. It
is evident from the decreasing rate of equivalent lime consumption with
increasing throughput, the resistance due to internal diffusion within
the slag particle effectively diminishes the neutralization rate.
Observation of large, partially neutralized, slag particles revealed
that a clear topochemical reaction had occurred. Consequently, equivalent
lime consumption rates were normalized by basing the reaction on the
external surface area of the slag particle. If the equivalent lime
consumution were expressed as a function of time only, distinctly different
curves are obtained for each of the different acid mine water flow rates
-103-
-------
oo
o
60
CO
CM
B
O
I
o
cu
CO
TJ
0)
o
u
O
to
o
co
-------
employed. As the acid mine water flow rate was decreased to about 145cc
per minute, a much lower neutralization rate was obtained, indicating
that diffusion within the liquid phase surrounding the slag particles
becomes the primary resistance at lower flow rates. If the liquid phase
diffusional resistance were the only controlling step in the neutralization
reaction, it would be expected that a constant neutralization rate
independent of throughput would be obtained. As can be seen from the data,
neutralization rate decreased slightly with increasing throughput.
Such a phenomenon indicates that diffusion within the slag particle
contributes to the overall neutralization reaction rate at this low flow
rate.
At the lowest acid mine water flow rate (145cc per minute) the space
velocity (volume of acid mine water per volume of packed bed) was
relatively low at 2.19. At his space velocity, the apparent residence
time of the acid mine water while in contact with slag was approximately
27 seconds. Even at this relatively long contact time the pH of the
effluent leaving the slag bed was not raised appreciably. This result
is presented in Figure 41 which shows the effect of acid mine water
throughput on pH of about 4.7 was obtained at the onset of flow but the
effluent pH rapidly decreased as throughput increased. This is the
result of some precipitation and/or the increase in the diffusion path
to the interior of the particle as lime is consumed topochemically.
Erratic behavior of the pH at throughputs in excess of 40 was due to
poor control on the feed rate as the feed reservoir was depleted of acid
mine water.
The packed bed contained approximately 100 times the equivalent of lime
required to neutralize all acid mine water in the feed reservoir.
Consequently, a series of experiments were completed whereby the acid mine
water was continuously recirculated through the packed bed in a closed
loop system to determine the ultimate or steady state pH that could be
achieved. Based on the alkali utilization factor test it was expected
that only about 30 percent of the slag alkali would be effectivt in the
neutralization. Since the latter tests were obtained under rather strong
neutralization conditions, continuous recirculation of the acid mine
water would establish effectiveness of this slag in neutralizing rela-
tively mild acid liquids. Results of these tests showed that steady state
pH values of 4.6 and 4.1 were obtained for the -% + 10 and -10+20 mesh
materials respectively. These data tend to support the conclusions ob-
tained from the alkaline utilization factor i.e., slag alkali utilization
decreases with increasing slag particle size.
AMW Neutralization
After the completion of the recycle tests, the -10 + 20 mesh slag
particles were screened to determine their extent of degradation. Screen
analysis showed that less than 1.6 percent passed through -20 mesh with
0.04 percent passing 100 mesh. Thus it can be expected that slag
degradation will not occur during neutralization causing an increase in
the overall reaction rate.
-105-
-------
T3
0)
PQ
CD
1-1
CO
60
cl
CD
0)
C
CD
3
M-l
w
Space Velocity 2.19 Min.
Slag Particle Size -10 + 20
Flow Rate 145 cc/min
Feed pH=2.0
20 40 60 80
Throughput (Volume of AMW/Volume of Slag Bed)
FIGURE 41 - EFFECT OF THROUGHPUT ON pH OF EFFLUENT LEAVING
FLOW NEUTRALIZING REACTOR
100
-106-
-------
Because the alkali utilization factor and continuous neutralization tests
indicate that very fine grinding of the slag will be required to
neutralize the acid mine water to a pH of seven in reasonably sized
equipment, the continuous tests were terminated in favor of batch
neutralization studies. Calculations based on the continuous rate study,
which assume that the major resistance is liquid diffusion, indicate
that the slag particles must be ground to finer than 200 mesh. At this
size consist, a pH of seven can be achieved. However, capital equipment
and operating costs would be substantial.
Batch Neutralization Studies
Because of the slow neutralization rates and poor alkali utilization
factors achieved in the earlier work, batch neutralization tests were
completed on a 1.2 basicity slag over a wide range of particle sizes.
Figure 42 presents results of this work and shows effect of slag particle
size and time on resultant pH of the neutralized acid mine water.
Neutralization rate increases with decreasing slag particle size. How-
ever, neutral (pH = 7) water was obtained only with slag particle sizes
finer than 270 mesh. Extremely long neutralization times were required
for all slags having larger particle size. Although all of the data are
not shown in Figure 42, residence times in the order of days were required
to bring the pH to seven for the larger slag particles. Based on these
results, it does not appear that the synthetic slags produced in this work
are commercially attractive for neutralization of acid mine water.
Because all of the neutralization tests were completed with sulfur free
slags, a second series of batch neutralization tests were completed on
slag obtained from the desulfurization unit. For these tests, slags
containing about eight percent sulfur initially were desulfurized by the
air water reaction to sulfur levels of one percent or less. It was
anticipated that the removal of sulfur from the slag might increase the
internal porosity of the slag particles and high neutralization rates could
be achieved. The results of this work are presented in Figure 43 which
shows the effect of contact time, basicity, and particle size on acid
mine water pH. Only typical data are presented in this figure. In
particular, representative data were selected to show the effect of
decreasing particle size with decreasing basicity. For example, for a
0.92 basicity slag ground to -100 mesh, the neutralization rate is com-
parable to the lower basicity 0,82 basicity slag ground to -200 mesh.
This would be expected because the alkali utilization factor increases
with increasing basicity and decreasing particle size.
Extensively long residence times are required to obtain neutral water.
As a point of reference, a commercial grade of limestone ground to -200
mesh is shown for comparison. As is evident limestone is more effective
for neutralization than desulfurized slag.
If the desulfurized slag contained residual sulfur considerable hydrogen
sulfide was generated during the neutralization step. Inasmuch as it is
unlikely that desulfurization will be 100 percent effective in recovery
of sulfur, use of desulfurized slag for neutralization is not recommended.
Production of hydrogen sulfide will necessitate use of an enclosed neu-
tralization tank or a pond to prevent an air pollution problem. This
-107-
-------
o
00
I
Batch Neutralization Tests, Initial pH of Acid Mine Water = 2.0
Slag No. 5 (Table I) 1.2 Basicity
10 hours
4 hours
1 hour
0.2
1.2
0.4 0.6 0.8 1.0
Average Slag Particle Size (Millimeters)
FIGURE 42 - EFFECT OF SLAG PARTICLE SIZE ON pH AS A FUNCTION OF NEUTRALIZATION TIME
1.4
-------
I
I—1
o
I
8
Batch Neutralization Tests, Initial pH of Acid Mine Water = 2.0
pH 6.4 @ 2.5 hrs
O
D Limestone -200M
• 1.0. Basicity
60 Mesh
X 0.92 Basicity
100 Mesh
O 0.82 Basicity
200 Mesh
X pH 8.1 @ 48 hours
pH 6.9 @ 30 hrs
20
40
60 80
Time Minutes
100
120
140
FIGURE 43 - VARIATION OF pH WITH TIME AS A FUNCTION OF SLAG BASICITY
-------
becomes prohibitively expensive. Consequently, it is recommended that
neutralization with desulfurized slag be deleted from consideration.
Engineering Design Recommendations
Based on results of continuous batch neutralization studies with sulfur-
free and desulfurized synthetic slags it is not recommended that the
slags produced in the combustor be used for neutralization of acid mine
water. These slags tend to exhibit a strong occlusion factor which pre-
vents all of the alkali from being effectively used for neutralization
unless the slag is ground to finer than 200 mesh. This will necessitate
the use of expensive (ball mill) grinding equipment. Additionally, if
the slag leaving the desulfurization or sulfur recovery unit contains
residual sulfur it will generate a serious air pollution problem, as
hydrogen sulfide is evolved from the neutralization system. This will
necessitate a closed neutralization tank along with hydrogen sulfide
removal equipment so that foul odors will not be released into the
atmosphere. It is expected that the hydrogen sulfide concentrations will
be too low to afford an economical recovery system which can generate
elemental sulfur. Consequently, the additional capital outlay will only
increase the cost of the process„ Based on this work, it is recommended
that the acid mine water be fed directly into the distillation unit
without benefit of neutralization.
-110-
-------
KINETICS OF SULFUR RECOVERY FROM SLAG
Introduction.
The economics of the process depend upon the kinetics and efficiency of
sulfur recovery from slag. Sulfur can be recovered from the slag in
two ways. One is to react the slag with carbon dioxide in the presence
of moisture to produce calcium carbonate and hydrogen sulfide.
(1) CaS + H20 + C02 = CaCO + H2S
The hydrogen sulfide can be treated by the Glaus process to produce
elemental sulfur according to the simultaneous reactions.
2H2S + 302 = 2H20 + 2S02
(2)
2H2S + S02 = 2H20 + 3S
These reactions are well known^~^". in general, the hydrogen sulfide
production step is carried out at low temperatures with water in the
liquid phase. The Glaus reaction for conversion of hydrogen sulfide to
elemental sulfur is completed at elevated temperatures (about 800°F
depending upon the process). Reaction (1) is a preferred route for the
recovery of sulfur when calcium sulfide bearing slags are available at
ambient temperatures,, In this manner, heat requirements can be minimized,
However, when slag is available at high temperatures (2000°F or more)
the following reaction
(CaS) 4- H20 = H2S + (CaO)
slag slag
affords a more economically attractive approach to sulfur recovery. The
mechanism of this reaction is not well known and as yet cannot be
explained „ However, it is generally believed that an oxygen deficiency
exists in the slag which permits the decomposition reaction (3) to occur.
Inasmuch as high temperature hydrogen sulfide is available, the addition
of air to the gas stream should generate S02 and permit the thermal con-
version of H2S to elemental sulfur according to equation (2).
Because carrying out hydrogen sulfide generation and reduction to sulfur
simultaneously offers a substantial capital and operating cost reduction,
the C02 treatment cost emphasis was placed on this approach. No attempts
were made to evaluate the mechanism of reaction (1). Instead, efforts
were concentrated on evaluating the yield of H2S and elemental sulfur in
one single step.
Theory 5
It is generally assumed that the sulfur that chemically combines with
-111-
-------
slag during the desulfurization of steel exists in the form of calcium
sulfide „ However, the form and nature of the CaS found in solid slags
is subject to speculation. For example, various authors 27-28s in studies
on CaO-MgO-SiC>2 slags, have shown that calcium sulfide exists in
crystalline form as well as a solid solution possessing complex formula
structure. Consequently, equilibrium data based on theoretical and
experimental work associated with pure calcium sulfide as given by
(4) CaS + H20 = CaO + H2S
(5) CaS + %02 = CaO + hS^ (gas)
(6) CaS + xH20 + y02 = CaO + aH2S + bS02 + cS
are not necessarily valid for desulfurization of sulfur containing slag.
Nevertheless, the theoretical considerations do provide valuable
insight into the desulfurization reaction and are presented in Figure 44.
As is evident from Figure 44, reaction (5) has a favorable thermodynamic
effect for completion because of its high negative free energy. The
high equilibrium constant explaining why the desulfurization reaction in
steel refining requires that the oxygen level in the steel be maintained
as low as possible to prevent the spontaneous decomposition of calcium
sulfide. For this reason, the process operates with a high carbon
level in the iron bath during carbon combustion,, By maintaining a high
carbon level in the iron bath, oxygen content in the pig iron is main-
tained at an extremely low level „ However, in the desulfurization
reaction for recovery of sulfur from the spent slag, the reverse situation
must be maintained to obtain high sulfur yields.
Equation (4) is another reaction tending to promote the desulfurization
of spent slag. Desulfurization of calcium sulfide is favored by increas-
ing temperatures. It is interesting to note that even though the
equilibrium constant is low, generation of hydrogen sulfide by this
technique compares favorably with the commercial reaction given by
Equation (1).
To determine tie probable gas phase compositions for the reactant species,
defined in Reaction (6), a free energy minimization program was developed
in a computer assessment of the reaction products that may be obtained
under equilibrium conditions as a function of temperature and molar
water-to-oxygen ratios. Typical results for these equilibrium calculations
at a temperature of 1400°K are presented j_n Figure 45. At low water-
to-oxygen ratios conversion proceeds almost entirely to elemental sulfur.
This is attributed mainly to the high negative energy free change associated
with reaction (5). As the water-to-oxygen ratio increases to ten and
higher, equilibrium sulfur conversion tends to decrease. Although not
shown, the effect of temperature is 'to increase the conversion to sulfur
at all water-to-oxygen ratios.
-112-
-------
10
JJ
CO
fl
O
§
•H
•H
3
10
-2
-3
-4
o Data Reference (11)
Equilibrium Calculation
— Equilibrium Calculation
e?
7 9
(104/°K)
11
13
FIGURE 44 - VARIATION OF EQUILIBRIUM CONSTANT WITH TEMPERATURE
-113-
-------
d
0)
d
o
•u
0)
4-1
03
•H 0.1
1
M
d
o
•r-i
jj
o
CO
.00
Sulfur
.01
0.1
Moles HO/Mole
1.0
10
FIGURE 45
EQUILIBRIUM DATA FOR THE SYSTEM, CaS 4- xH20 4- y02
at 1400°K
= CaO 4- aS02 4- bH2S 4- cS2 + dS8
-------
o rv
Literature data are available which indicate that the water granulation
of liquid slag under oxidizing conditions tends to promote the formation
of a high concentration of hydrogen sulfide as compared to sulfur dioxide.
Contrary to equilibrium calculations, ratios of hydrogen sulfide to
sulfur dioxide as high as 100 to one have been obtained. Additionally,
during the quenching reaction, a large amount of hydrogen was also formed.
Although it was stated that the nature of the reaction mechanism leading
to hydrogen formation during granulation was not known nor could it be
explained, a tentative hypothesis for the desulfurization of slag was
put forth which indicated that kinetics, rather than equilibria, dictated
the overall yield of products.
Inasmuch as literature on desulfurization of blast furnace slags (and
those comparable to the type used in our process) was rather scant, an
experimental program was undertaken to determine effect of temperature,
particle size, and reactant flow rates on the yield of sulfur and/or
hydrogen sulfide, A theoretical kinetic model based on the concepts
outlined by Wen for non-catalytic heterogenous (solid-fluid) reactions
was adopted. The model assumes that a shrinking spherical unreacted
core of radius r exists within each spherical slag particle of radius
R, Using a pseudo steady-state solution of the equation describing
the change in reactant concentration with time at various locations
within the spherical particle, a solution was obtained relating reaction
time, t, to reactant concentrations. This solution is shown as equation
2)
(7) t = "'su V . . X - c_ + I (l _ _c ) + R (x fc)
3 ( k D ) ( R3 ) k ( R ) 2D(
m s
where:
a = interfacial area, sq ft/cu ft of solids
C = weight fraction of sulfur in solids at any time
C = concentration of reactant fluid in bulk phase, mole/ft
SO
C = weight fraction of sulfur in solids at time zero
Co^ = initial concentration of solid reactant, mole/cu ft
oO
D = diffusivity, sq ft/min
k = mass transfer coefficient across fluid film, ft/min
m
ks = reaction rate constant based on surface, ft/min
r = distance from center of sphere to reaction surface ft
R = radius of particle ft
t = time, minutes
-115-
-------
The first product term within the brackets of Equation (7) defines the
fluid film resistance, that is, the transport of reactant from the bulk
gas phase across the boundary layer surrounding the particle. The second
term defines the contribution of chemical resistance and the last term
relates to the resistance afforded by diffusion through the ash or reacted
layer surrounding the unreacted core. Depending upon which resistance
controls the overall reaction rate, it is possible to simplify Equation
(7) substantially. Assuming that chemical reaction controls (it will be
shown later that this indeed is true) Equation (7) can be rewritten as
where X is the fraction of calcium sulfide converted to the sulfur-free
form. If chemical reaction controls the overall desulfurization rate, a
logarithmic plot of l-(l-X)^-'^ versus t should have a slope of unity.
Similarly, it can be shown that, if the slope of the log- log plot is 1/2,
then ash diffusion controls the overall reaction rate. It will be shown
in the discussion of the experimental results, that the mathematical
model presented by Wen adequately represents the experimental data and
affords the opportunity for scale up to commercial equipment.
Experimental Procedure
Two types of experimental equipment were used in this work. The first
employing stainless steel and glass equipment is shown in Figure 46.
Air and water were introduced by rotameters into a preheater maintained
at reaction temperature so that vaporization of the water occurred out-
side of the reactor proper. The heated gases were introduced into a
stainless steel reactor (1.5 inches I.D. by eight inches long) flanged on
both ends and heated by means of a muffle furnace. The muffle furnace
was controlled at the desired reaction temperature. The outlet of the
reactor was connected to an air condenser heated to a temperature of
approximately 400°F by means of an electrical resistance tape. This
condenser was used to collect sulfur. The outlet of the sulfur condenser
was connected to a water cooled glass condenser for removal of water
vapor which was collected in a graduated flask. A gas sampling port for
chromatographic analysis was located between the steam condenser and a
wet test flow meter. Gases exiting the wet test meter were collected
in a water displacement vessel for cumulative gas analysis.
Difficulties were experienced in operation of this equipment because of
the reaction of sulfur compounds with the vessel. Consequently, an
alternative experimental apparatus was constructed to obtain meaningful
kinetic data as shown in Figure 47. A variable speed positive displacement
pump was used to maintain a constant flow of water into an inconel pre-
heater (straight pipe) that was inserted within a high temperature fur-
nace set at the desired reaction temperature. Sufficient surface area
was allowed for heat transfer from the furnace to preheat and vaporize
the water to the approximate reaction temperature. The steam was directed
onto a ceramic combustor boat maintained within a ceramic combustion
tube located within an electrical resistance heated furnace. One half to
-116-
-------
Thermocouple
Temperature
Indicator-Contr'll
Water
Air
Preheater
Furnace
Reactor
•—
Bt
3 Thermocouples
in Reactor
400°F
Sulfur
Trap
Heating Tape
Condenser
Thermometer
Cooling
Water
Gas Sample
Port
5-liter
Water Trap
FIGURE 46 - STAINLESS-STEEL REACTOR DESULFURIZATION EQUIPMENT
-------
00
ilnconel Steam Tube
Variable-Speed
Pump
Air -
FURNACE
\
^Sample Boat XjQ
Ceramic Tube
Temperature
Indicator
Controller
FIGURE 47 - CERAMIC-TUBE REACTOR DESULFURIZATION EQUIPMENT
Gas Sampling
Port
Air-Cooled
Condenser
-------
one gram samples of slag were spread in a single layer of particles
across the ceramic combustion boat and the steam was directed across
the particles. Air or argon was introduced into the combustion tube
and across the ceramic boat by means of a rotameter. Effluent gases
from the combustion tube were air cooled and periodic spot samples were
obtained for gas chromatographic analysis,,
The experimental procedure consisted of introducing a known amount of
slag material into the combustion boat and reacting it with air and
water for a known period of time. The entire combustion boat sample
was then analyzed for residual sulfur content. Reaction time was varied
for a given sample to determine rate of desulfurization.
Discussion of Results
A considerable number of experiments employing pure calcium sulfide and
slags of various basicity and particle size was completed over the temp-
erature range of 1000 to 1700°F in the stainless steel reactor shown in
Figure 32. However, because of experimental difficulties and side
reactions only qualitative data were obtained for this system. Three
problems proved to be insurmountable in terms of obtaining qualitative
data. The first was inability to collect all of the sulfur formed in the
steam-calcium sulfide or steam-slag reaction. In the pure calcium sul-
fide experiments, reagent grade powdered (-100 mesh) calcium sulfide was
employed. At the steam flow rates employed, sufficient plugging of dust
carryover and sulfur condensate was observed. Additionally, the sulfur
condenser was unable to effectively trap all of the sulfur vapors
generated during reaction and substantial quantities were carried over
with steam to the steam condenser. Consequently, a cloudy steam con-
densate was collected containing substantial quantities of sulfur. For
these reasons it was not possible to quantitatively define the amount of
elemental sulfur obtained as a function of time. The third and final
anomaly was substantial corrosion of the stainless steel reactor and
deposits of scale in the desulfurized calcium sulfide slag samples. A
magnetic separation of the partially desulfurized slag and scale materials
was made. Sulfur analysis on the magnetic fraction indicated sub-
stantial quantities of sulfur present in the scale material. All of the
scale deposits could not be magnetically separated from the desulfurized
slag. Consequently, it was not possible to determine the extent of de-
sulfurization by a sulfur analysis of the spent slag. However, estimates
have shown that approximately 60 percent of the sulfur was removed from
the slag.
Even though the data could not be used for a quantitative assessment of
the reaction kinetics, several pertinent conclusions were drawn. Typical
data to show the effect of temperature obtained in these experiments are
presented in Figure 48. These data are presented only as a relative
guide and show that the rate of evolution of hydrogen sulfide increases
with increasing temperature. In addition to H^S, elemental sulfur was
produced in substantial quantities. In general, the results of these
qualitative experiments indicate that desulfurization rate increases
with decreasing particle size and increasing temperature.
119-
-------
(1) -60 + 100 Mesh
1200°F .82 Basicity
Slag 19% CaS
(2) -60 + 100 Mesh
1500°F .82 Basicity
Slag 19% CaS
CO
CM
co
CD
.010
-(2)
.005
(1)
100 200 300
Time, (Minutes)
FIGURE 48 - TEMPERATURE EFFECT ON STEAM-SLAG REACTION
-120-
-------
Even though the data are relatively inaccurate with regard to rate of
desulfurization, the fact that only approximately 60 percent of the
sulfur was removed in a reaction time of about five hours invalidates
the feasibility of desulfurization at these low temperatures (less than
1700 F) . For commercial utility, to obtain a reasonable size reactor,
residence times of the solid particles should be maintained at less than
one hour,,
An important consideration derived from these experiments was the fact
that the slag desulfurization reaction was topochemical in nature.
Coarse slag particles of approximately \ inch in diameter were de-
sulfurized using steam at a temperature of about 1700°F. An examination
(both visually and microscopically) of the desulfurized slag showed a
white outer layer surrounding a gray unreacted core. This tended to sub-
stantiate the hypothesis that experimental results could be treated
using a model based on a particle with a shrinking unreacted core. An
additional factor observed during these desulfurization experiments was
that the particle size did not change significantly during desulfur-
izationc A screen analysis on the desulfurized slag initially sized at
-6 +10 U.S. mesh showed that about 1.5 percent passed through the ten
mesh screen after desulfurization (approximately 60 percent sulfur
removal) „
Gas analysis of samples withdrawn during the course of these experiments
indicated the existence of nitrogen, oxygen, carbon dioxide, carbon
monoxide, hydrogen and hydrogen sulfide. In general, hydrogen sulfide
concentration in cumulative gas samples approached 12 percent by volume
on a water-free basis. Formation of hydrogen sulfide and hydrogen along
with the absence of sulfur dioxide tends to support the hypothesis put
forward in the literature with regard to slag desulfurization by water or
Experimental difficulties inherent in the steel reactor resulted in
expenditure of considerable quantities of slag. Because of limited
availability of slags, the all-ceramic desulfurization apparatus shown
in Figure 47 was designed to use only 0.5 to one gram slag samples. As
small quantities of slag were involved, it was not possible to quanti-
tatively collect sulfur produced. However, sulfur deposition was
evidenced in the cooler regions of the reaction tube. Periodic spot
samples obtained for gas analysis (particularly during the early portions
of the run) indicated no sulfur loss as gaseous S02 or H2S. Consequently,
it is assumed that all sulfur loss from the desulfurized slag sample
occurred by formation of elemental sulfur. The latter assumption requires
further experimentation on larger quantities of material using more
refined experimental equipment. Kinetics of sulfur removal were
followed by determining residual sulfur content in the slag after a
known period of reaction time. Results of this work are presented in
Figures 49, 50 and 51.
Figure 49 shows the desulfurization that occurs as a function of time
when correlated according to the topochemical reaction model proposed by
Wen (Equation 8). The data correlate for a model that assumes chemical
-121-
-------
1.0
o
u
0.1
0.0
0.82 Basicity Slag
Particle Size: -20 + 10
X
O
2600°F
2200°F
1800°F
1500°F
C = Wt. of sulfur
at anytime
C = Wt. of
sulfur
at ti
zer
50
100
Time, Minutes
FIGURE 49 - EFFECT OF TIME ON DESULFURIZATION, SLAG BASICITY OF 0.82
-122-
-------
O
U
Slag
Size: -20 + 40
Size:
2600°F
2000°F
O 1800°F
C = Wt. of sulfur at any-
time
C = Wt. of sulfur at time
o
zero
100
Time, minutes
FIGURE 50 - EFFECT OF TIME ON DESULFURIZATION
SLAG BASICITY OF 0.90
-123-
-------
1.0
u
,10
D
1.01 Basicity Slag
Particle Size: -20 + 40
• 2600°F
X 2200°F
O 2000°F
D 1800°F
C = Wt. of sulfur at
anytime
C = Wt. of sulfur at
time zero
,01
100
Time, Minutes
FIGURE 51 - EFFECT OF TIME ON DESULFURIZATION
SLAG BASICITY OF 1.01
-124-
-------
reaction controls, inasmuch as a slope of one prevails. The reaction
temperature shown is the furnace temperature and not the temperature
of the slag. Since steam and air were introduced into this system as
soon as the slag was placed in the hot zone of the furnace, desulfur-
ization occurs over a range of temperatures varying from somewhat above
room temperature up to the final reaction temperature for a given run.
Since the overall reaction rate thus determined is less than the true
reaction rate for the stated temperature, subsequent reactor design
based on these data should be conservative in nature.
Desulfurization data for the 0.82 basicity slag followed^ pattern
that was consistent with that obtained earlier in the stainless steel
reactor experiments. Figure 49 shows that the rate of desulfurization
increases with increasing reaction temperature. At temperatures below
1800°F chemical reaction controls desulfurization rate and the data are
correlated by a straight line of slope one. A change in slope occurred
for the 2200°F test at a desulfurization level to approximately 94 per-
cent. The departure from chemical reaction as exemplified by the change
of slope from one to 0.5 indicates that ash diffusion is controlling in
this region. This implies that the topochemical nature of the reaction
observed earlier with coarse particles in the steel reactor experiments
persists for the 0.82 basicity slags. As will be seen later, this
phenomenon did not exhibit itself in the higher basicity slags. It is
not known at this time why the 0.82 basicity slag shows this effect
whereas the other slags do not. As the furnace temperature was increased
to 2600 F, the desulfurization data exhibit chemical reaction control
over the entire sulfur removal range studied (residual slag sulfur con-
tent at the end of 20 minutes was 0.01 percent which corresponds to 95
percent sulfur removal). The reason for the apparent anomaly of re-
version to chemical reaction control lies in the fact that the slag
particles became liquid after 20 minutes in the furnace at 2600°F. In
this manner the unreacted core containing the residual sulfur was exposed
to the reaction gases. Consequently, diffusion through the solid ash
(reacted layer) region of these solid particles no longer existed.
Desulfurization data for 0.9 and 1.01 basicity slags are presented in
Figures 50 and 51 respectively. As is evident, they all follow a chemical
reaction controlled regime with the data lying along a straight line of
slope one. A comparison of the desulfurization data tends to indicate
that the ease of desulfurization increases with decreasing basicity.
However, subsequent calculations show that the apparent desulfurization
rate when corrected for the differences in external specific surface for
each of the slags, reduces the desulfurization data to a common denominator.
Using Equation (8), and the curves of Figures 49, 50 and 51, specific
reaction rates were calculated from the intercepts of the straight lines
shown in these figures. Correction of the specific reaction rate for
differences in external specific surface permits a correlation of
reaction rate as a function of temperature as shown in Figure 52. It
-125-
-------
50
10
a
CO
^
ra
.5
O
O
O 0.82 Basicity Slag
D 0.90 Basicity Slag
X 1.01 Basicity Slag
a = interfacial area, sq ft per cu
ft of solids
k = reaction rate constant based on
surface, ft per min
T = temperature, °R
(1/°T) x 104
FIGURE 52 - EFFECT OF TEMPERATURE ON SPECIFIC REACTION RATE
-126-
-------
can be seen that the data are correlated reasonably well by the Arrhenius
plot. For these data, the correlating equation is given by
(9) ak =3.036 x 104 Exp(-21,442/T)
s
where:
a = interfacial area, sq ft per cu ft of solids
k = reaction rate constant based on surface, ft per min
s
T = temperature, °R
Substitution of Equation (9) into Equation (8) and rearranging results
in
c = Co / L 3.86 x 10"8 t exp (-21.442/T)^
Cso
where:
C = weight fraction of sulfur in solids at any time
CQ = weight fraction of sulfur in solids at time zero
C = initial concentration of solid reactant^ made per cu ft
t = time, minutes
This is the correlating equation of all of the data presented in Figures
49, 50 and 51. By use of this equation, it is possible to determine the
residual sulfur content in a slag (over the basicity range studied) as a
function of time, temperature, specific surface and initial sulfur content.
A comparison of the calculated percent sulfur in the slag versus the
experimental sulfur actually obtained is presented in Figure 53. The
data correlated reasonably well and it should be pointed out that
Equation (10) tends to be conservative for the lower basicity slag. By
and large, the bulk of the data lying above the line on Figure 53
associated with 0.82 basicity slag. This is to say that Equation (10)
yields a higher residual sulfur content than was actually achieved in
the experimental program. Inasmuch as the lower basicity slags tend to
favor the overall economics of the process, an engineering reactor design
based on Equation (9) will prove to be conservative.
Engineering Reactor Design Recommendations
The experimental effort was completed using synthetic slags in an attempt
to anticipate results that may be achieved from slags produced under
commercial conditions. The synthetic slags contained no iron or alkali
compounds that may be expected to combine with a slag under actual operating
conditions. Additionally, slags produced in this work were of low surface
-127-
-------
3
M-l
T3
J-l
CO
CO
O
2468
Experimental % Sulfur
FIGURE 53 - CORRELATION BETWEEN ACTUAL & PREDICATED
DESULFURIZATION RESULTS
-128-
-------
area and high density unlike those that may be anticipated in commercial
production. Consequently, it is assumed that the desulfurization rate
data derived from this work will be conservative with regard to the
actual rate experienced in commercial equipment. Although conservative
desulfurization rate data were obtained, it is not possible to quanti-
tatively state the degree of conversion of sulfide materials to elemental
sulfur. Data would tend to indicate that conversion of elemental sulfur
is high; however, absolute conversion levels were not established. Al-
though material balances could not be obtained, the qualitative data
available indicate that conversion to elemental sulfur can be assumed to
be in excess of 90 percent. Until more detailed experiments can be
completed on desulfurization of the slag, it is recommended that 90
percent conversion be assumed for further engineering and economic feasi-
bility studies.
Using this assumption, and the design Equation (9) it is recommended that
the engineering feasibility study assume the following:
(1) A slag of approximately 0.8 basicity should be employed in the
combustor.
(2) Either a liquid or solid slag can be used for desulfurization.
When using a liquid slag, reaction temperature should be main-
tained as high as possible above the melting point. Reaction
times under this condition will be approximately 30 minutes.
(3) When using a solid slag, a reaction temperature in excess of
2000°F, and a slag residence time of one hour should be employed.
(4) Solid slags should be crushed to minus 20 U.S. mesh prior to
desulfurization.
(5) Desulfurization reaction is controlled by chemical reaction.
Consequently, fluid mass flow rates through the reactor can
be set at a minimum level so that gas phase mass transfer
balances the reaction rate for a given set of conditions.
(6) Further degradation of the slag particles due to desulfurization
will not occur.
-129-
-------
REFRACTORY LINING LIFE
Introduction
The combustor is the most critical piece of equipment in the acid mine
water procgss because it must contain molten iron and slag at temperatures
up to 2700 F. The frequency of maintenance shutdowns to re line the
combustor will not only determine the technical and operating feasibility
of the process but will also establish the economic potential of the
operation.
The main factor which determines the time interval between maintenance
shutdowns is the combustor refractory lining's resistance to slag attack.
Molten high-carbon iron does not reduce refractory life to a serious
degree. However, slags through mechanical erosion and chemical reaction
bring about a significant decrease in refractory life. Mechanical erosion
of the refractory occurs at the slag-refractory interface, and is often
intensified by a softening of the refractory brought on by reactions
between the slag and refractory. Chemical attack results from reaction
between the refractory and slag that has penetrated deeply into the
refractory. Such reactions cause refractory failure either by softening
due to formation of low melting compounds or by crumbling of the
refractory due to recrystalization.
The life expectancy of refractory linings in contact with low-sulfur (less
than three percent) slags found in commercial processes is well known.
However, a search of the literature and discussions with vendors failed
to uncover any information on life of refractories subjected to slags
of high-sulfur content (to minimize the amount of slag processed and to
facilitate sulfur recovery, combustor slags will contain approximately
ten percent sulfur). Accordingly, a laboratory study was conducted to
determine effect of high-sulfur slags on various refractories.
Experimental Procedure
The experimental method used in this study is best described as a
compromise between two commonly used refractory tests: the slag button
test and the slag drip test. In the commonly used version of the slag
button test, a hole drilled in the sample brick is filled with slag.
After the two materials have been in contact at high temperature for a
predetermined length of time, the brick is cooled and sectioned to measure
the slag penetration into the brick. Because this test requires a
minimum of furnace space, a large number of samples can be held at temp-
erature for prolonged periods of time. However, the slag in the sample
hole is stationary. Consequently, this test does not provide information
on mechanical erosion. The button test imposes a severe refractory temp-
erate evaluation. In commercial operations, the refractories are
generally exposed to high temperatures on only one face of the brick.
The other face is usually at a much lower temperature. In this manner,
a temperature gradient is established which is sufficient to solidify any
slag or molten material that has penetrated the brick. The solidified
-131-
-------
material thus acts as a barrier between the brick and molten slag and
tends to prevent further chemical attack. In the button test, temperature
gradients within the brick are very small or nonexistent, which precludes
the possibility of preventing chemical corrosion by solidification of
the molten corrosive materials. To some degree, this temperature
gradient effect is off-set by a compensating slag-dilution effect. The
slag dilution is the result of slag-refractory reactions that may alter
the properties of the small amounts of slag used in the test to such an
extent that reactions between slag and refractory are inhibited.
With exception of the temperature gradient effect, these limitations
do not exist in the slag drip test. In this test, molten slag is caused
to drip slowly onto a refractory sample to bring about some mechanical
erosion along with a strong chemical attack. However, this test requires
considerable space in a high temperature furnace and equipment for
feeding the slag into the furnace. As a compromise between the two types
of tests, the slag button test was modified to include periodic additions
of fresh slag to the test hole of the refractory sample. This modi-
fication tends to minimize the neutralization of slag by reaction with
the refractory. The test still does not induce mechanical erosion but
it does provide a valid comparison between various refractories.
For this study, five conventional types of bricks and three new more
expensive high-density bricks were selected, as shown in Table XVII.
The super duty fireclay was chosen as reference material, because it is
a widely used low-cost refractory. The other conventional refractories
consisted of two alumina and two magnesia bricks. The alumina can be
expected to resist attack from acid slags. A slag is classified as acid
if its acid oxide (silica and alumina) content by weight exceeds the basic
oxide (lime and magnesia) content. Since it was not known whether cal-
cium sulfide acts as a neutral compound, weak acid or weak base, alumina
as well as basic-slag-resistant magnesia were selected for testing.
Both the alumina and magnesia bricks were also tested in tar-impregnated
versions. Impregnating the brick with tar results in a deposit of
graphite in the pores of the refractory after the brick is heated to
operating temperature. The graphite fills and coats the refractory pores,
thereby minimizing slag penetration and attack of refractory.
In addition to conventional refractories, three new types of refractories
were studied. One of these is an isostatically pressed alumina of
extremely low porosity (0.1 percent). In the future, a wider variety
of refractory materials is expected to be available as isostatically
pressed very low porosity bricks. The remaining two samples were both
melted and cast to shape alumina produced by two different manufacturers.
The fused cast alumina designated as Type I was a grade that contained
air bubbles (blow holes) that were not removed during solidification. The
second alumina, Type II was a better quality brick free of blow holes.
No fused basic refractories were tested because their tendency to crack
under thermal shock would have severely complicated the test procedure.
No sudden temperature changes are anticipated during combustor operation,
including start-up and shut-down. Accordingly, fused basic refractories
-132-
-------
should not be excluded from consideration.
TABLE XVII
Refractories Used in Laboratory Tests
Conventional
Superduty Fireclay
High Alumina
Tar Impregnated Alumina
Magnesite-Chrome
Tar Impregnated Magnesia
New
Isostatically Pressed (High Density) Alumina
Fused & Cast Alumina Type I (with "Blow Holes")
Fused & Cast Alumina Type II (without "Blow Holes")
Although it is anticipated that a commercial combustor will contain
castable refractories as well as bricks, all samples were cut from brick
to insure that each refractory sample was prepared under controlled
conditions of mixing, forming and firing. The brick samples used were
approximately three inch cubes containing a one-inch diameter hole 1-%-
inch deep to hold the slag.
Slag used in this study contained 18.6 percent calcium sulfide and had a
basicity ratio of 0.8 where basicity is defined as the weight percent
ratio (CaO + MgO) /SiC>2 + A^Oo) . Chemical composition for the slag is
shown in Table XVIII. Preparation of the slag consisted of heating the
pure components to 2600°F, cooling, crushing and reheating the mixture
three times to insure a uniform composition. The resulting slag was
crushed to pass through 100 mesh U.S. sieve, and 12 gram portions were
briquetted to yield a 7/8-inch diameter by %-inch high briquette. Slag
briquettes were sized for easy placement in the sample brick test holes
while the samples were in the high-temperature furnace.
Twenty-six refractory samples were prepared for testing by preheating to
the 2640°F test temperature at a temperature rise of less than 50°F per
hour to avoid refractory fracture du • to thermal shock. The 2640°F test
temperature was chosen as the highest temperature at which the test
furnace could be expected to operate contiguously for the two-week
minimum test duration. Slag briquettes were inserted in the test holes
of the brick samples. The samples were periodically inspected and slag
was added to those that could hold more (as determin ' by the slag level
being lower than %-inch from the top of the test hole). Table XIX shows
total grams of slag added to each refractory sample. This table also shows
the duration of contact of each sample with slag at 2640°F. After each
refractory sample completed its scheduled time of contact with slag in
the furnace, it was removed and air cooled.
-133-
-------
TABLE XVIII
Slag Composition Used in Refractory Study
Component
CaS
CaO
MgO
A1203
SiCL
Percent by Weight
18.6
19.3
17.7
17.7
26.7
TABLE XIX
Summary of Test Results
Refractory Type
Superduty Fireclay
Superduty Fireclay
Superduty Fireclay
High Alumina
High Alumina
High Alumina
High Alumina
Tar-Impregnated Alumina
Tar-Impregnated Alumina
Magnesite Chrome
Magnesite Chrome
Magnesite Chrome
Tar-Impregnated Magnesia
Tar-Impregnated Magnesia
Tar-Impregnated Magnesia
Tar-Impregnated Magnesia
Isostatically Pressed Alumina
Isostatically Pressed Alumina
Fused & Cast Alumina I
Fused & Cast Alumina I
Fused & Cast Alumina I
Fused & Cast Alumina II
Fused & Cast Alumina II
Hours of
Exposure
to Slag
7.5
25
176
46
49
176
384
176
384
46
176
384
46
49
176
384
7.5
384
7.5
176
384
176
384
Grams
of Slag
Added
12
60
60
24
60
72
60
60
72
24
72
72
24
72
84
72
12
72
12
72
72
60
84
Change
in Hole
Radius, In.
0
0.3
0.45
0
0.15
0.2
0.3
0.2
0.2
0
0
0
0
0
0
0
0
0
0
0.05
0.05
0
0.05
Change In
Hole Depth,
Inches
0
0.5
0.7
0
0.1
0.3
0.3
0.1
0.1
0
0
0
0
0
0
0
0
0.1
0
0.1
0.2
0
0.1
-134-
-------
To determine the effect of slag attack on the refractory material, the
cooled brick sample was cut in a vertical plane as nearly through the
center of the hole as possible. The cut sample was first photographed
for a permanent record, and then was placed on an office copier when an
image was recorded on paper. The image was then used to determine depth
of the test hole. The cut sample was then cut on the horizontal plane
and the cut fraction was placed on the office copier and an image
recorded. This image was then used to measure length of the chord
generated by the cut and distance from the chord to the eroded brick hole
surface. From these measurements, the diameter of the test hole after
exposure to the slag was determined. This procedure was used because of
the difficulty in cutting the brick sample exactly on a diameter of the
test hole. Amount of erosion was determined by subtracting the original
radius from the radius after exposure of the test hole to slag.
Results
Results of this study are shown in Table XIX and Figure 54 where it is
seen that there is no measurable erosion of the magnesia samples
Figure 54 A, and very little erosion of the new high-density refractories
Figure 54 B. There is some erosion of the alumina samples (Figure 54 C)
and considerable erosion of the superduty fireclay Figure 54 D a
standard low-cost refractory used as reference material. The latter
suffered 0.45 inches of erosion after 176 hours of exposure. The high
alumina brick showed approximately one-half the erosion of the super-
duty fireclay for the same 176 hour exposure to slag. The attack on
tar-impregnated high alumina was even less than that of the plain high
alumina.
The superior erosion resistance of tar-impregnated high alumina brick
compared to high alumina brick is attributed to the carbon coating formed
within the tar-impregnated brick after firing. Samples of fused and
cast alumina (Types I and II) showed very little erosion after 384 hours
of exposure to slag. These materials have relatively low porosities
(one percent to four percent) but the isostatically pressed alumina has
even a lower pososity (0.1 percent) and tends to show a slightly greater
resistance to slag erosion. After 384 hours of contact to slag, this
material exhibited no significant slag penetration. Thus the alumina
data indicate that the fused and cast materials will be acceptable
refractories.
The magnesia samples (including the magnesite-chrome) indicated no change
in test hole dimensions after being in contact with slag for up to
384 hours. However, observation of the brick samples after testing
indicated that a large percentage of the slag effected an erosion-
free penetration of the brick. This implies a high degree of wettability
of the brick by the slag so that the slag freely entered the pores of the
brick. The slag-refractory reaction appeared to be minor with no
evidence of refractory softening or crumbling. For these reasons, it can
be expected that, in a commercial operation where the bricks have a
temperature gradient imposed on them, the slag will freeze after pene-
trating the pores of the brick. In this manner, refractory life will be
enhanced.
-135-
-------
Figure 54 A
Magnesite-Chrome
Brick
384 Hours
Figure 54 B
Fused Alumina-Type II
384 Hours
Figure 54 C
High Alumina
384 Hours
Figure 54 D
Fireclay
176 Hours
FIGURE 54 - THE EFFECT OF SLAG CONTACT ON REFRACTORY EROSION
Time Parameters: Hours of Slag-Refractory Contact
-136-
-------
Above results show that calcium sulfide in the slag acted as a mild base;
and, in conjunction with the lime and magnesia content of the slag (see
Table XVIII), caused it to act as a base. This basic slag attacked the
alumina bricks, but not the magnesia bricks. As is the case with most
slags, this was better contained by the less porous refractories.
Design Recommendations
The refractory study shows that the high sulfur content of the slag being
considered for the combustor does not significantly effect reactivity
of these slags with refractories.
For economic evaluation of the process use of the ma gne site-chrome brick
that performed so well in these tests should be assumed.
Prior to final design of the commercial installation, price and
availability of fused or other low-porosity basic bricks should be
investigated.
-137-
-------
APPENDIX B
THE PROCESS WORKING AREA DIAGRAM
To generate a process working area diagram (PWAD) requires that the
energy and material balance computer program be run using ranges of slag
recycle fraction and flux rate at the selected heat rate. For illustration
purposes, a two million gallon per day acid mine water plant utilizing
partially neutralized dilute AMW will be used. The PWAD will be generated
for a process operating at a heat rate of two million BTU/1,000 gallon
AMW using a ten percent sulfur coal refuse having a heating value of
6,000 BTUs/lb. The computer program was run using flux rates of 100,
200, 300 and 400 tons per day with spent slag recycle fractions of 0,
0.4, and 0.8. Results from these runs are shown in Table XX where the
flux rate, spent slag recycle fraction, resulting slag basicity, air
preheat temperature, and percent sulfur in slag are presented.
TABLE XX
Computer Results Used to Generate Process Working Area Diagram*
Percent
Flux Spent Slag Slag Air Sulfur
Tons/Day Recycle Fract. Basicity Temperature,°F In Slag
100 0 0.00 1507 18.44
100 o4 0.19 1610 12.17
100 .8 0.42 2051 4.10
200 0 0.47 1562 14.41
200 .4 0.67 1701 9.19
200 .8 0.90 2314 3.27
300 0 0.94 1616 11.83
300 .4 1.15 1786 7.43
300 .8 1.37 2556 2.60
400 0 1.42 1669 10.03
.4 1.63 1869 6.25
* Partially neutralized dilute acid mine water heat rate-2 MM BTU/1,000
gallon AMW, ten percent S coal refuse, combustor temperature 270QOF,
2,000,000 gallon AMW plant.
The first step in generating a PWAD is to prepare a plot of basicity
versus spent slag recycle fraction at various flux rates into the kiln.
Figure 55 is such a plot and was prepared using the data of Table XX.
In Figure 56 is presented a plot of air preheat temperature versus slag
recycle fraction at various flux rates. From this figure it is seen that
at a given flux rate the required air preheat temperature increases as
the slag recycle fraction increases. Also shown is the fact that at a
given slag recycle fraction the air preheat temperature increases as the
flux rate increases. Constructing a line at the 2000°F air preheat
-139-
-------
375
350 Tons/Day Flux
325
300
125
100
.2
.4
.6
1.0
Slag Recycle Fraction
FIGURE 55 - SLAG BASICITY VS. SPENT SLAG RECYCLE FRACTION OF
VARIOUS FLUX RATES
-140-
-------
4)
J-l
3
4-J
CO
H
(1)
-------
temperature through the curves shown in the figure yields points of flux
rate and slag recycle fraction which require a 2000°F air preheat
temperature. At 100, 200, 300 and 400 tons per day of flux, the air
preheat temperature is 2000°F when the slag recycle fraction is 0.77,
0.64, 0.57 and 0.51 respectively.
The next step is to plot the percent sulfur in the combustor slag as a
function of the slag recycle fraction at various flux rates. This plot
is shown as Figure 57<, In the same manner as before, a maximum ten
percent sulfur line is drawn through the curves to determine points of
spent slag recycle fraction and flux rate for which the sulfur content
of the slag is ten percent. At flux rates of 100, 200 300 and 400 the
spent slag recycle fraction which yields a ten percent sulfur content in
the slag are 0053, 0.43, 0.26 and 0 respectively.
The PWAD is constructed using results from the two previous figures and
applying them to Figure 58 coupled with minimum and maximum basicity
constraints. This results in the process working area diagram shown in
Figure 58. As seen, the PWA is bounded on the bottom and top by the
minimum and maximum basicity constraints which require that basicity be
in the range of 0.8 to 1.2. The PWA is bounded on the left by the
maximum ten percent sulfur in slag constraint and on the right by the
maximum air preheat temperature of 2000°F. The process working area is
a convenient representation of the operable ranges of the various process
parameters. Size and shape of the process working area will depend
primarily on the acid mine water concentration, heat rate, and the coal
refuse composition.
-142-
-------
00
CO
O
JJ
CO
1
O
O
C/3
fl
(1)
O
O
CN
VD
QO
Maximum 10% Sulfur
in Combustor Slag
100 tons/day flux
200
300
400
.2 .4 .6
Slag Recycle Fraction
.8
1.0
FIGURE 57 - PERCENT SULFUR IN COMBUSTOR SLAG VS. SLAG RECYCLE
FRACTION AT VARIOUS FLUX RATES
-------
f*>
4-1
•^
O
•H
CO
TO
pq
oo
Maximum Basicity -1.2
Process Working Area
Maximum
10% Sulfur
Maximum Air
Temperature 2000°F
Minimum Basicity - .8
.2
.4
.6
.8
1.0
Slag Recycle Fraction
FIGURE 58 - PROCESS WORKING AREA DIAGRAM FOR A 2 MM GAL/DAY
PLANT USING PARTIALLY NEUTRALIZED DILUTE AMW AND
A HEAT RATE OF 2 MM BTU'S/1,000 GALS AMW
-144-
-------
APPENDIX C
EQUIPMENT COSTS
To determine the capital investment requirement for the process, the
process equipment was grouped into nine equipment complexes. These
equipment complexes were designated as (1) coal handling system, (2)
neutralization, (3) steam turbine—air compressors, (4) direct-fired
air heater, (5) waste heat boiler, (6) combustor, (7) distillation,
(8) desulfurization, and (9) rotary kiln dryer. Individual costs of
these complexes were determined for a two MM GPD AMW plant utilizing
partially neutralized dilute AMW, a ten percent sulfur coal refuse
with a heating value of 6,000 BTU/lb., a heat rate of two MM BTUs/1,000
gallon AMW, a flux rate of 230 tons/day and a spent slag recycle
fraction of 004. The energy and material balance computer program
was run, using the above values to determine the capacity and/or
temperature of the process streams which control the size and conse-
quently the cost of the equipment complexes. These costs and
capacities are then used to scale up or down the equipment complex
costs for any plant size and operating conditions.
The process streams which control the size and cost of the various
equipment complex are as follows:
Equipment Complex Process Stream(s) Determining Capacity
Coal Handling System Coal Rate, Tons/Day
Neutralization Acid Mine Water, Gallons/Day
Steam Turbine—Air Compressors Coal Rate, Tons/Day
Direct Fired Furnace Air Preheat Temperature, °F,
and Air Rate, Tons/Day
Waste Heat Boiler Acid Mine Water, Gallons/Day
Combustor Combustor Offgas, Tons/Day
Distillation Acid Mine Water, Gallons/Day
Desulfurization Combustor Slag, Tons/Day
Rotary Kiln Dryer Dry Solids Leaving Kiln,
Tons/Day
Based on 333.3 tons/day of coal refuse, the coal handling system will
cost $91,290 (mid - 1970, purchase price at factory). The coal handling
system consists of a belt conveyor to transport coal from a storage pile
to a hammer mill, a rotary dryer to dry the crushed coal, pneumatic
conveying equipment to transport the crushed coal to process storage, a
storage tank, a hopper, and pneumatic conveying equipment to transport
the coal to the combustor lances. Individual costs for this equipment
plus relevant sizes and/or capacities are presented in Table XXI.
Based on partially neutralizing two MM GPD of AMW, the neutralization
complex will cost $50,795. The neutralization complex consists of a
stainless steel pump to pump AMW into a neutralization tank, a
pneumatic conveying system to transport limestone into the neutralization
145-
-------
TABLE XXI
Coal Handling Complex Cost
Item of Equipment Size/Capacity Cost
Coal Refuse Belt Conveyor 14" Belt x 22 Ft. $ 2,910
with Motor & Drive Equip. @ 150 Ft/Min.
Hammer Mill 14 tons/hr 5,620
Rotary Kiln Coal Dryer 7.5 ft. dia. x 55 ft. 69,000
Complete Pneumatic Conveying
System for Crushed Coal to
Process Storage 13.5 tons/hr 7,870
Process Storage Tank
Carbon Steel 1,000 Gallons 3,050
Pneumatic Conveying
System for Crushed Coal Into
Combustor Lances 13.5 tons/hr 2,050
Hopper, Carbon Steel 8 hr. hold-up 790
Total Cost, Coal Handling System $91,290
tank, two agitators to stir the AMW in the neutralization tank, and a
bronze pump to pump partially neutralized AMW to the distillation unit.
The individual equipment costs for the neutralization complex are
presented in Table XXII.
TABLE XXII
Neutralization Complex Cost
Stainless Steel and
Bronze Pump 2 MM GPD AMW $11,650
Neutralization Tank
Stainless Steel 42,000 Gallons 30,500
Limestone Pneumatic
Conveying System 2 tons/hr 5,925
Agitators (2) -- 2,720
Total Cost, Neutralization Complex $50,795
-146-
-------
High pressure steam exiting the waste heat boiler will be used in two
steam turbine-air compressors to generate compressed air at eight to
ten psig to transport coal refuse into the combustor and air at four
to five psig for combustion in the combustor. The low pressure steam
exiting the compressors will be used as the steam source for distillation.
Based on transport and combustion air requirements the steam turbine--
air compressor complex will cost $28,500. The air compressor required to
supply one ton/hr of air at eight to ten psig for pneumatic injection
of coal will cost $5,5000 The air compressor required to supply 25 tons/
hr of combustion air at four to five psig will cost $23,000.
The cost of a direct fired heater which is used to preheat the combustion
air entering the combustor depends on air preheat temperature and
combustion air rate into the combustor. The air preheat temperature
determines the material of construction used for the direct fired fur-
nace and the heat transfer surface required. Based on heating 636 tons/
day of air, the cost of the direct fired furnace at various air preheat
temperatures is shown in Table XXIII.
TABLE XXIII
Direct Fired Heater Complex Cost
Air Preheat Temp.,°F Material of Construction Cost
700 Carbon steel $ 33,000
1000 Carbon steel 38,000
1300 , Carbon steel 42,700
1300 Chrome/moly. 61,700
1700 Chrome/moly. 98,000
2000 Chrome/moly. 121,000
The waste heat boiler uses combustor offgas to generate steam for dis-
tillation. Based on generating 830 tons/day of steam at 100F superheat
and 250 psig, the waste heat boiler costs $90,000.
The combustor is a refractory lined steel vessel containing air and coal
lances, a slag-iron separation box, and a chrome/moly. star valve for
admitting dryer solids. Based on 805 tons/day of combustor offgas, the
combustor will cost $46,600. The individual costs associated with the
combustor complex are presented in Table XXIV. The distillation complex
consists of a complete flash distillation plant without equipment to
generate steam and provide AMW to the distillation plant. In the process
the waste heat boiler is used to supply the steam requirements for
distillation. Cost of a distillation plant depends on the capacity of
the plant, the economy factor used in the design and whether partially
neutralized AMW is used. The economy factor, which is a function of the
steam available for distillation, determines the heat transfer surface
required to distill the acid mine water. Consequently, the economy factor
affects the cost of the plant. Concentration of the acid mine water (if
neutralization is not used) affects plant cost because it determines the
materials of construction. When partial neutralization is used the AMW
-147-
-------
TABLE XXIV
Combustor Complex Cost
Item of Equipment
Carbon steel combustor
shell and slag-iron
separator
Refractory lining
Coal lances (2)
Air lances (6)
Size/Capacity
14 ft dia. x 15 ft
Magnesite-Chrome Brick
2" I.D. x 12 ft Copper
6" I.D. x 12 ft Copper
Total cost, Combustor complex
Cost
$ 9,000
21,600
15,000
1,000
$46,600
composition does not affect the cost of the distillation plant, In
Table XXV are presented the installed costs of distillation plants
operating with and without partial neutralization at economy factors of
five to ten for dilute and moderately concentrated AMW. These costs
were provided by the manufacturer for a two MM GPD plant.
TABLE XXV
Distillation Complex Cost
Distillation Conditions
Distillation, neutralized
Dilute acid mine water
Distillation, neutralized
Dilute acid mine water
Distillation, unneutralized
Dilute acid mine water
Distillation, unneutralized
Moderately concentrated
of acid mine water
Size/Capacity
2 MM GPD AMW, Economy
factor of 10
2 MM GPD AMW, Economy
factor of 5
Cost
$2,500,000
2,000,000
2 MM GPD AMW, Economy
factor of 10 2,500,000
2 MM GPD AMW, Economy
factor of 10 2,800,000
The desulfurization complex is used to remove sulfur from the combustor
slag. Based on the desulfurization of 406 tons/day of slag the desul-
furization complex is $73,000. The desulfurization complex consists of
an enclosed belt conveyor to transport hot slag to a crusher, a cold slag
belt conveyor to transport desulfurized slag to the hot slag conveyor to
provide a protective covering for the belt, a desulfurized slag quencher
and hood, a refractory lined steel reaction vessel where a sulfur rich
gas is produced, a sulfur condenser to condense sulfur from the gas, and
sulfur storage facilities0 The costs for this equipment plus the
relevant sizes and/or capacities are presented in Table XXVI.
-148-
-------
TABLE XXVI
Desulfurization Complex Cost
Item of Equipment Size/Capacity Cost
Enclosed hot slag belt conveyor 18" Belt, 75 ft $21,200
Cold slag conveyor 12" Belt, 30 ft 3,460
Crusher 20 tons/hr 7,850
Reaction Vessel 8 ft dia x 20 ft
Refractory Lined 26,500
Sulfur Condenser 35 tons/day sulfur 5,620
Sulfur storage tank 150 ft^ 3,370
Desulfurized slag quencher
& hood 388 tons/day spent slag 5.000
Total cost, Desulfurization Complex $73,000
Equipment associated with the rotary kiln dryer are: kiln structure,
combustion equipment to burn the combustor offgas and kiln refractory
lining. Based on 283.9 tons/day of solids leaving the kiln the cost
is $315,000 as shown in Table XXVII.
TABLE XXVII
Rotary Kiln Dryer Cost
Item of Equipment Size/Capacity Cost
Kiln structure 9 ft dia x 270 ft $265,000
Combustion equipment 275 tons/day combustor
offgas 30,000
Refractory lining High alumina brick 20,000
Total Cost, Rotary Kiln Dryer complex $315,000
Equipment costs for the various equipment complexes were determined
using standard capital cost estimating procedures or obtained from
manufacturer quotes.
-149-
-------
APPENDIX D
THE THEORY OF CARBON SOLUBILITY RATES
Introduction
The steady state operation of the combustor requires that a balance exist
between the rate of carbon removal by oxidation and the rate of carbon
solution in the molten iron bath. Literature data are available which
indicate that carbon removal rates in excess of one percent per minute
can be achieved. In general, the rate of carbon removal is limited by
the rate at which oxygen or air is introduced into the molten bath.
Carbon solution rates, on the other hand, present a more difficult problem.
Olsson, et. al. ' , in laboratory work pertaining to the rate of
solution of solid carbon in molten iron, have shown that solubility is
mass transfer controlled. Carburization of open hearth melts
using anthracite, coke, graphite, gas carbon and charcoal has been
reported by Leary et. al. . In their work on commercial open hearth
melts containing low carbon, they have reported carburization rates as
high as two percent carbon per minute. Another significant aspect of
this work was the conclusion that sulfur contained in the carbon was
distributed between the slag and metal with essentially no evolution
in the gas phase.
Because the literature data on recarburization were relatively scant and
generally done in shallow metal baths, a theoretical study was completed
to determine the effect of carbon particle size on residence time in the
hot bath and the depth of hot metal required to essentially affect a
complete dissolution of the carbon before it reaches the surface of the
molten iron. This section presents results of this work.
Theory
Since the solution of solid carbon into an iron melt is mass transfer
controlled^2, the rate of solution can be defined as
-(dn/dt) = KA (Ci-Cb) (1)
where n is the weight of carbon in the particle, K is the mass transfer
coefficient (cm/sec), A is the surface area of the particle (cm2), ci
and Cb are the carbon concentration at the reaction interface and in the
liquid phase respectively (g/cm). The mass transfer coefficient is
dependent on the -relative velocity between the carbon and the liquid
interface, on the diffusion coefficient and on the shape of the solid
particle.' A generalized mass transfer correlation is given by
L 1/3
(Kr/D) = 1.0 + 0.3 (Nr)2 (Ng) (2)
where r is the radius of the particle (cm), D is the diffusivity of
carbon in iron (cm2/sec), and NrNg are the Reynolds and Schmidt numbers
respectively (dimensionless).
151-
-------
Combining these two equations with a carbon material balance and a
particle force balance completely defines the physical system of rising
carbon particles dissolving in a molten iron bath. The material
balance for each particle (assumed .spherical) is
(dn/dt) = Ps d (4 r3/3)/dt (3)
where t is the time (sec), Ps is the solid density (g/cm3) and d
represents the differential operator.
The force balance about the particle rising in the molten iron bath at
its terminal velocity indicates that the buoyancy force is equal to the
sum of the drag and gravitational forces. Mathematically
4 (r3Pfg/3)= 4 (frV Pf)+ 4 (r3psg) (4)
where p is the density of the fluid (g/cm3), f is the drag coefficient
(dimenslonless), g is the gravitational constant (cm/sec2) and V is the
terminal velocity of the particle (cm/sec). The drag coefficient is a
function of the Reynold's number ^ and is given by
f = 18.5/(Nr)°'6 (5)
Combining equations (1-5) and solving for the change of particle radius
with residence time in the molten iron bath yields
-(dr/dt) = ai(l/r + a^O.071) (Ci-Cb) (6)
where
a
i = D/P
s
1/3 0.4 1/4 1/4 0.65
a2 = 0.31 Ns Pf (Ps-Pf) g /v
and v = viscosity (g/cm sec) and the other terms are as defined previously.
By maintaining the carbon composition and temperature in the bath constant
at steady state, Equation (6) can be integrated to find the residence
time of the carbon particle in the bath at any fraction of its initial
weight.
To compute the metal bath depth required to dissolve a particle, the
force balance (Eq. 4) can be solved for the terminal rising velocity and
combined with Equation (6) to yield
-(dh/dr) = l/(a3r"2'142 + a^'1-0?1) (7)
-152-
-------
where h is the bath depth required to dissolve a carbon particle to any
desired fraction of its initial weight and
a5 = (0.289 (p -i
Specification of the carbon feed rate into the combustor and a desired
carbon solution rate determines the amount of hot metal required in the
combustor. This results because the amount of hot metal is directly
proportional to the carbon feed rate and inversely proportional to the
carbon solution rate at steady state. Consequently, knowing the weight
of hot metal required and the depth of hot metal necessary to dissolve
the carbon particle the diameter of the combustor is readily established.
Discussion of Theoretical Calculations^
To determine the effect of particle radius on the particle residence
time required to dissolve 99.9 percent of the carbon particle Equation (6)
was integrated numerically on a digital computer. The results of this
work are presented in Figure 59. The parameters on the figure refer
to a surface factor which relates the actual surface area of the particle
to the surface area of a smooth sphere having the same diameter as the
particle in question. As is evident, the residence time requirement for
a given surface factor decreases with decreasing particle radius.
Numerical integration of Equation (7) shows a similar effect (Figure 60)
for the molten metal bath depth requirement as a function of carbon
particle radius.
The actual specific surface of the coal refuse after injection into the
molten iron bath is not known. However, it is believed that a
considerable expansion (somewhat like a popcorn effect) will occur due to
the evolution of volatile matter and thermal shock. It is believed that
the surface factor under these conditions will lie in the range of 10
to 100.
For the economic and engineering evaluation, the combustor was sized
conservatively using a surface factor of one. A closely sized coal
fraction was injected into the combustor having an average particle
diameter of 1,6 mm (-10 + 20 U.S0 mesh). As would be expected, in
actual operation of the combustor the coal will be ground to a wide size
distribution such as -10 mesh rather than cut to a close particle size.
However for a conservative estimate on the combustor dimensions,
closely sized coal was used.
-153-
-------
100.
o
01
CO
cu
•r-l
H
O)
o
•H
tn
0)
i— I
O
en
CM
10
•"•"•
1.0
ol
Surface Factor = 1
Surface Factor = 2
Surface Factor = 100
t
.2 .3 .4 .5 .6 .7 .8 .9 1.0
Particle Radius, TTo(cm.)
FIGURE 59 - EFFECT OF PARTICLE RADIUS ON RESIDENCE TIME NEEDED
IN IRON MELT TO DISSOLVE CARBON
-154-
-------
100.
.1
Factor = 10
• 2 .3 .4 .5 .6 .7 .8 .9 1.0
Particle Radius , ¥ o(ctn.)
FIGURE 60 - EFFECT OF PARTICLE RADIUS ON HOT METAL DEPTH
REQUIRED TO DISSOLVE CARBON
155-
-------
BIBLIOGRAPHIC:
Applied Technology Division, Black, Siv-
alls & Bryson, Inc., Evaluation of Acid
Mine Water Drainage Treatment Process,
Final Report, FWQA Contract No. 14-12-
529
ABSTRACT
An economic and engineering evaluation
of a submerged coal refuse combustion
process to convert acid mine water (AMW)
to potable water has been made. In
this process coal refuse is burned in
molten iron to supply energy for dis-
tillation or reverse osmosis, and the
coal refuse sulfur is trapped in a slag
for eventual recovery of sulfur. Lab-
oratory experimentation was conducted
ACCESSION NO.
KEY WORDS
Acid Mine Water,
Distillation,
Slag Desulfuriza-
tion,
Slag Characteriza-
tion,
Submerged Coal
Combustion,
Two-Stage Coal
Combustion,
Coal Refuse
BIBLIOGRAPHIC:
Applied Technology Division, Black, Siv-
alls & Bryson, Inc., Evaluation of Acid
Mine Water Drainage Treatment Process,
Final Report, FWQA Contract No. 14-12-
529
ABSTRACT
An economic and engineering evaluation
of a submerged coal refuse combustion
process to convert acid mine water (AMW)
to potable water has been made. In
this process coal refuse is burned in
molten iron to supply energy for dis-
tillation or reverse osmosis, and the
coal refuse sulfur is trapped in a slag
for eventual recovery of sulfur. Lab-
oratory experimentation was conducted
BIBLIOGRAPHIC:
Applied Technology Division, Black, Siv-
alls & Bryson, Inc., Evaluation of Acid
Mine Water Drainage Treatment Process,
Final Report, FWQA Contract No. 14-12-
529
ABSTRACT
An economic and engineering evaluation
of a submerged coal refuse combustion
process to convert acid mine water (AMW)
to potable water has been made. In
this process coal refuse is burned in
molten iron to supply energy for dis-
tillation or reverse osmosis, and the
coal refuse sulfur is trapped in a slag
for eventual recovery of sulfur. Lab-
oratory experimentation was conducted
ACCESSION NO.
KEY WORDS
Acid Mine Water,
Distillation,
Slag Desulfuriza-
tion,
Slag Characteriza-
tion,
Submerged Coal
Combustion,
Two-Stage Coal
Combustion,
Coal Refuse
ACCESSION NO.
KEY WORDS
Acid Mine Water
Distillation,
Slag Desulfuriza-
tion,
Slag Characteriza-
tion,
Submerged Coal
Combustion,
Two-Stage
Coal
Combustion,
Coal Refuse
-------
on those areas which could profoundly affect the process.
These areas were: A laboratory demonstration of slag
desulfurization to produce sulfur, the evaluation of slag
sulfur retention characteristics, slag capability for
neutralizing AMW and the determination of slag compositions
having acceptable fluidities. Laboratory results indi-
cated that sulfur is obtained, high slag sulfur partition
ratios were achieved, fluid slags are produced and de-
sulfurized slags are not suitable for neutralization.
Engineering studies show that the process has potential
for supplying inexpensive energy for distillation and per-
mits the recovery of sulfur so that distilled water is
economically produced. Depending upon the AMW composition
and sulfur selling price ($20 to $30/ton) the break-even
price of water for a 5 MM GPD plant varies between $.42
and $.16/1000 gals when a 14 percent capital interest charge
is used.
on those areas which could profoundly affect the process.
These areas were: A laboratory demonstration of slag
desulfurization to produce sulfur, the evaluation of slag
sulfur retention characteristics, slag capability for
neutralizing AMW and the determination of slag compositions
having acceptable fluidities. Laboratory results indi-
cated that sulfur is obtained, high slag sulfur partition
ratios were achieved, fluid slags are produced and de-
sulfurized slags are not suitable for neutralization.
Engineering studies show that the process has potential
for supplying inexpensive energy for distillation and per-
mits the recovery of sulfur so that distilled water is
economically produced. Depending upon the AMW composition
and sulfur selling price ($20 to $30/ton) the break-even
price of water for a 5 MM GPD plant varies between $.42
and $.16/1000 gals when a 14 percent capital interest charge
is used.
on those areas which could profoundly affect the process.
These areas were: A laboratory demonstration of slag
desulfurization to produce sulfur, the evaluation of slag
sulfur retention characteristics, slag capability for
neutralizing AMW and the determination of slag compositions
having acceptable fluidities. Laboratory results indi-
cated that sulfur is obtained, high slag sulfur partition
ratios were achieved, fluid slags are produced and de-
sulfurized slags are not suitable for neutralization.
Engineering studies show that the process has potential
for supplying inexpensive energy for distillation and per-
mits the recovery of sulfur so that distilled water is
economically produced. Depending upon the AMW composition
and sulfur selling price ($20 to $30/ton) the break-even
price of water for a 5 MM GPD plant varies between $.42
and $.16/1000 gals when a 14 percent capital interest charge
is used.
-------
Number
Subject Field &, Group
05D
SELECTED WATER RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
Organization
Environmental Protection Agency
Washington, D. C.
Evaluation of a New Acid Mine Drainage Treatment Process
iQ Authors)
James A. Karnavas
Eugene A. Pelczarski
16
21
Project Designation
14010 DYI
Note
22
Citation
Environmental Control Research Service
14010 DYI 2/71
Environmental Protection Agency
Washington, D. C.
23
Descriptors (Starred First)
Acid mine water*, distillation*, slag desulfurization*, slag characterization*,
coal refuse*
25
Identifiers (Starred First)
Submerged coal combustion*, two stage coal combustion*, acid mine water
treatment*
27
Abstract
^n ecOnomic and engineering evaluation of a submerged coal refuse combustion
process to convert acid mine water (AMW) to potable water has been made. In this
process coal refuse is burned in molten iron to supply energy for distillation or
reverse osmosis, and the coal refuse sulfur is trapped in a slag for eventual
recovery of sulfur. Laboratory experimentation was conducted on those areas which
could profoundly affect the process. These areas were: A laboratory demonstration
of slag desulfurization to produce sulfur, the evaluation of slag sulfur retention
characteristics, slag capability for neutralizing AMW and determination of slag
compositions having acceptable fluidities. Laboratory results indicated that sulfur
is obtained, high slag sulfur partition ratios are achieved, fluid slags are produced,
and that desulfurized slags are not suitable for neutralization.
Engineering studies show that the process has potential for supplying
inexpensive energy for distillation and permits the recovery of sulfur so that
distilled water is economically produced. Depending upon the AMW composition and
sulfur selling price ($20 to $30/ton) the break-even price of water for a 5 MM GPD
plant varies between $.42 and $.16/1,000 gals when a 14 percent capital interest
charge is used.
This report was submitted in fulfillment of Program No. 14010 DYI, contract
No. 14-12-529 under the sponsorship of the Environmental Protection Agency.
Abstractor
Institution
WR: 102
WR SI C
(REV JULY IS
U S DEPARTMENT OF THE INTERIOR
WASHINGTON. D. C 20240
• GPO: 1969-359-339
------- |