United States
Environmental Protection
Agency
Industrial Environmental
Research Laboratory
Research Triangle Park NC 2771
Research and Development
EPA-600/S7-84-066 July 1984
<&ER& Project Summary
Correlation of Coal Properties
with Environmental Control
Technology Needs for Sulfur
and Trace Elements
D.M. White, L.O. Edwards, A.G. Eklund, D.A. DuBose, F.D. Skinner, D.L.
Richmann, and J.C. Dickerman
This report reviews existing reports
and data on the occurrence of sulfur and
trace elements in U.S. coals and on the
effect of coal properties on trace
element partitioning during coal utiliza-
tion. Areas of emphasis include 1) the
effect of depositional conditions on the
formation and composition of mineral
matter in coal, 2) the elemental concen-
tration of major and trace elements in
U.S. coals as a function of rank and geo-
graphic location. 3) analytical methods
used for evaluating the modes of
occurrence of these elements in coal, 4)
conceptual models for predicting sulfur
and trace element occurrence as a
function of depositional conditions and
chemical equilibrium, and 5) the fate of
major and trace elements during coal
cleaning, combustion, gasification, and
waste disposal.
Coal washability data for 44 U.S. coal
samples were used to statistically
estimate the trace element reduction
potential for a coal as a function of
sulfur and ash reduction. Data fits were
especially good for elements associated
with the clay minerals, and to a lesser
extent with the sulfides. Coal combus-
tion data from 15 previous studies at
commercial power plants were also
analyzed; but, due to differences in
technological processes at various
plants, possible analytical errors, and
limited data, statistical correlations are
uncertain. Areas identified for future
research into elemental partitioning as a
function of coal properties include
development of further information on
the effect of depositional conditions on
coal quality, extension of the coal
washability data base to additional
coals (including analysis of mineral
forms in these coals), and collection and
analysis of additional data on the
partitioning of trace elements during
combustion.
This Project Summary was developed
by EPA's Industrial Environmental
Research Laboratory. Research Triangle
Park. NC. to announce key findings of
the research project that is fully docu-
mented in a separate report of the same
title (see Project Report ordering infor-
mation at back).
Introduction
Of the 92 naturally occurring elements,
all but 16 have been detected in coal.
Except for the elements which form the
organic structure of coal (H, C, O, S, and
N) and the associated major minerals (Al,
Ca, Fe, Mg, K, Si, Na, and Ti), the
concentrations of other elements are
almost universally less than 0.1 percent
Eight of the trace elements—Sb, As, B,
Cd, Ge, Hg, Mo, and Se—are frequently
found in concentrations greater than
typically found in the earth's crust. For
certain locations and ranks of coal, Pband
Zn are also found in concentrations
greater than the earth's crustal average
All other trace elements are found in
smaller amounts than in typical crustal
rocks Table 1 summarizes the average
enrichment factors for US coals.
-------
Table 1. Enrichment Factors of Trace Elements in Coal Relative to the A verage Composition of
the Earth's Crust'
Limited
Enrichment (1 -31
Element
Pb
Hg
Mo
Zn
Enrichment
1.04
2.38
2.15
1.11
Moderate
Enrichment (3- 10}
Element
Sb
B
Cd
Ge
Enrichment
5.80
4.30
4.35
4.23
Significant
Enrichment OW)
Element
As
Se
Enrichment
10.2
61.4
"A verage enrichment factor - average element concentration in coal -j- average element concentra-
tion in earth's crust.
During coal mining, preparation,
transportation, and utilization, trace
elements may be released into the
environment and potentially affect occu-
pational and public health. The exact
pathways and magnitudes of human
exposure to coal-related trace element
releases are not known. The National
Research Council has identified seven
trace elements (As, B, Cd, Pb, Hg, Mo, and
Se) plus compounds of C, N, and S as
being of greatest concern.
Minerals in Coal
Origin and Characteristics
Over 125 separate minerals have been
identified in coal. General pathways for
accumulation of mineral matter in coal
include:
1. Accumulation in the plant during its
growth.
2. Detrital materials washed or blown
into the peat-forming environment
(including volcanic sources).
3. Sorption of soluble ions from
associated water sources on the
surface of peat particles.
4. Precipitation by chemical and/or
biochemical processes during peat
accumulation.
5. Precipitation by chemical processes
during subsequent stages of coalifi-
cation.
Minerals resulting from materials
introduced into the bog or swamp at the
time of peat accumulation are generally
referred to as "syngenetic" minerals and
include the first four processes of mineral
matter origin discussed above. Minerals
precipitated in cracks, fissures, and other
voids in the coal matrix after peat burial
are identified as "epigenetic" minerals.
The syngenetic minerals are further
subdivided into two groups: "authigenic"
minerals formed from materials originally
in the peat or formed in-place by
precipitation or other chemical/biological
processes, and "detrital" minerals con-
sisting of solids transported into the
peat by water and wind. An alternative
method of differentiating minerals is to
distinguish minerals easily separated
from the organic structure of coal
("extraneous" or "adventitious" minerals)
from those too closely associated with the
organic matrix to be readily separated
("inherent" minerals). While epigenetic
minerals are generally large and thus
readily separable (i.e., extraneous),
precise and unequivocal distinctions
between authigenic and detrital, or
between extraneous and inherent—
although potentially of substantial scien-
tific and technological importance—are
not always possible.
Major Mineral Groups
The four major mineral groups are
clays, carbonates, sulfides, and silicates.
Lesser mineral groups include sulfates,
phosphates, oxides, hydroxides, and
salts. The associations of trace elements
with the various mineral groups are
identified in Table 2.
T/ie clay minerals (also referred to as
"aluminosilicates") constitute the largest
single mineral group in most coals,
ranging from 40 to 80 percent of the total
inorganic fraction. The most common
clays are illite, kaolinite, found as finely
divided grains (grain diameters are
generally less than several microns)
mixed with the coal and as horizontally
banded kaolinite-rich claystones inter-
bedded with the coal. Illite and montmo-
rillonite are important in coal preparation
processes because of their swelling
tendencies in the presence of water (and
thus the lowering of their specific
gravity). Clay minerals are important to
trace element analyses because of their
cation exchange capacity, particularly in
lignite and subbituminous coal.
The second largest mineral group in
most coals is the carbonates. Siderite and
dolomite formation are associated with
peatification: siderite forms in peats of
fresh-water origin and dolomite forms in
marine-influenced environments. Calcite
and ankerite are commonly found in
cracks and fissures and are therefore
considered to be primarily associated
with second-stage coalification.
Sulfide minerals are of major environ- I
mental importance because of their
conversion to SOa during combustion and
to H2S during reduction processes.
Principal among the sulfide minerals are
the two polymorphs of iron disulfide
(FeSz): pyrite and marcasite. Sphalerite
(ZnS), galena (PbS), and chalcopyrite
(CuFeS2) are also common. As and Hg are
found in coal primarily in association with
pyrite, while Cd is associated with
sphalerite. Fine-grained sulfides are
usually formed by syngenetic processes
associated with the bacterial conversion
of sulfates to sulfur. Coarse-grained
sulfide complexes of pyrite, sphalerite,
galena, and sometimes chalcopyrite are
associated with both syngenetic and
epigenetic precipitation processes. Re-
action of siderite [iron carbonate (FeCOs)]
with H2S can also produce pyrites.
Quartz (SiO2) is the dominant silicate
mineral in coal. Quartz grains found in
coal can result both from authigenic
precipitation of pore fluids and from fine-
grained detritus carried by water and
wind. Detrital quartz is a significant
constituent of clay tonsteins. With the
possible exception of feldspars, the other
silicates found in coal are generally too
rare to be of significance.
Occurrence of Trace Elements
in U.S. Coals
Elemental Concentrations
Trace element statistics were tabulated
by rank province for the 4,402 channel
samples in the National Coal Resource
Data System (NCRDS) as of November
1982. Statistics were also prepared for
bituminous coals from the Pottsville,
Monongahela, and Allegheny units of the
Appalachian Region, the eastern and
western regions of the Interior Province,
and the Rocky Mountain Province.
Significant variations are present in the
trace element contents of U.S. coals, both
by rank and by geographic area. The
variation in composition by rank is given
in Table 3 which shows the average value
of each element in the earth's crust
(known as "clarkes") in the far right
column.
Several trends in these data are
noteworthy. First, Ca, Mg, Na, B, Ge, and
Sr occur at increased concentrations in
low-rank coals; these elements are
believed to occur as exchangeable
cations bound to the organic matter and
clays in low-rank coals. Second, the
average concentration of several
philes (Fe, As, Cd, Ni, and Zn) and S
greater in the Interior Province and
-------
Table 2. Principal Minerals Found in Coal and Trace Constituents
Mineral Phases
Clay minerals
Kaolinite
////re
Montmorillonite
Chlorite
SuHides
Pyrite, Marcasite
Sphalerite
Galena
Chalcopvrite
Carbonates
Calcite
Siderite
Ankerite
Dolomite
Silicates
Quartz
Zircon
Tourmaline
Plagloclase feldspar
Alkali feldspar
Muscovite
Sulfates
Bariie
Gypsum
Phosphates
Apatite
Oxides and Hydroxides
Limonite. Goethite
Diaspora
Hematite. Magnetite
Ftutile
Salts
Halite
Sylvite
Bischofite
Major Constituents
Al. Si
Al, Si, K
Al. Si. Mg. Fe
Al. Si. Mg
Fe.S
Zn.S
Pb.S
Cu. Fe, S
Ca
Fe
Ca.Fe
Ca.Mg
Si
Si.Zr
Ca, Mg. Fe. B. Al. Si
Ca. Na. Al. Si
K, Na. Al. Si
K. Na. Al. Si
Ba.S
Co, S
Ca, P. F
Fe
Al
Fe
Ti
Na. Cl
K, Cl
Mg.CI
Abundance*
5-30%
30->60%
1-10%
1-5%
1-10%
1-5%
1-5%
<1%
5-30%
5-30%
5-30%
t-10%
1-10%
1-5%
-------
Table3. Sulfur. Ash, and Trace Elements in U.S. Coals by Rank'
Units Anthracite
Bituminous
Subbituminous
"Whole-coal basis; mean ± standard deviation.
* Aver age elemental concentration in crystal rocks.
Lignite
Clarkes*
Number of samples
Total S
Sulfate S
Pyritic S
Organic S
Ash
Si
At
Ca
Mg
Na
K
Fe
Mn
Ti
P
Sb
As
Ba
Be
B
Cd
Cr
Cu
Ge
Pb
Hg
Mo
Ni
Se
Sr
Th
U
V
Zn
%
%
%
%
%
%
%
%
%
%
%
%
%
%
%
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
ppm
52
0.86 ± 0.80
O.O2 ± 0.02
0.37 ±0.78
0.49 ± 0. 16
15.0 ± 8.94
6.56 ± 1.41
4.73 ±0.77
0.20 ± O.24
0. 15 ± O.21
0.07 ±O.1 3
0.30 ±O.1 3
0.87 ± O.85
0.01 ± 0.02
0.29 ±0.11
0.07 ±0.10
1.12± 1.84
7.67 ± 19.6
102 ±61.7
1.32 ± 0.85
11.8± 10.5
0.22 ± 0.30
47.2 ± 60.9
18.9 ± 16.4
1.71 ±2.17
7.32 ± 6.92
0.23 ± O.27
2.68 ± 2.44
28.5 ± 32.0
3.11 ±2.58
73.7 ± 77.0
6.09 ± 2.92
1.94 ± 3.38
37.8 ± 33. 1
12.6 ± 14.5
3527
2.31 ± 2.08
0.14 ±0.27
1.50 ± 1.82
0.92 ± 0.63
12.6 ±9.26
5.25 ± 1.51
2.87 ± 0.92
0.42 ±0.59
0.11 ±0.10
0.06 ± 0.07
0.22 ± 0.97
2.37 ± 1.86
0.01 ± 0.01
0. 14 ± 0.07
0.05 ± 0.09
1.22 ±2.27
20.3 ± 41.8
84.4 ± 132
2.22 ± 1.66
37.6 ±41.7
0.91 ± 7.30
20.5 ±27.5
17.8 ± 17.8
6.33 ± 10.2
14.3 ±27.5
0.21 ±0.42
3.28 ±4. 13
16.9 ±. 19.2
3.39 ±4.00
88. 1 ± 95.0
3.03 ± 3. 15
1.85 ±2.71
22.3 ± 19.7
92.4 ± 689
640
0.59 ± 0.67
O.O5 ± 0. 10
0.19 ± 0.41
0.37 ± 0.30
14.0 ±23.7
5.21 ±2.11
2.47 ± 0.85
1.72± 1.16
0.42 ± 0.33
0.25 ± 0.33
0.10 + 0.09
0.95 ± 0.92
0.01 ± 0.02
0. 13 ± 0.06
0. 13 ± 0. 14
0.85 ± 1.40
6.17 + 15.5
447 ± 924
1.30± 1.77
52.6 ± 38.5
0.38 ±0.47
14.9 + 25.6
14.1 ± 14.3
5.93 ± 6.55
6.35 + 5.47
0.1 0 + 0.11
2.1 9 ±2.82
7.O2 + 8.44
1.44 ± 1.42
171 ± 184
5.13 + 4.64
2.1 3 ±3.84
23.9 ±30.6
17.4 ±21.1
183
1.08 ± 1.01
0.22 ± 0.67
0.36 ±0.64
0.57 + 0.38
21. 5 ± 22.9
6.73 + 4.62
2.77+ 1.46
3.31 ± 2.28
1.01 +0.74
0.37 + 0.53
0.19 ±0.24
1.51 + 1.41
0.03 + 0.02
0.20 ± 0. 12
0.18 ±0.28
1.01 + 2.23
22.8 ± 138
415 +531
1.98 ±2.71
1 14 + 75.2
0.55 + 0.61
13.5 + 18.2
17.2 + 21.2
10.7+11.1
6.90 + 7.88
0. 15 + 0. 14
5.99 ± 26.3
8.35+ 19.7
2.70 ± 3.67
309 ±258
7.13 + 5.70
3.39 ± 10.3
29. 1 + 37.3
22.5 + 79.8
0.026
—
—
—
—
28.15
8.23
4.15
2.33
2.36
2.09
5.63
0 10
0.57
0.11
02
1.8
425
2.8
10
O.2
100
55
1.5
12.5
O.08
1.5
75
0.05
375
9.6
2.7
135
70
separation of fine-grained minerals
depends on both the actual modes of
elemental occurrence and the size to
which the sample is ground.
The initial efforts to examine the
distribution of trace elements in coal by
use of organic affinity methods were
those of L Morton and K.V. Aubrey, and of
P. Zubovic and co-workers at the U.S.
Geological Survey (USGS). During the
1970s, studies conducted by the Illinois
State Geologial Survey (ISGS), U.S.
Department of Energy (DOE), and Bitu-
minous Coal Research, Inc. (BCR) provided
additional information concerning the
mineral and organic associations of
sulfur and trace elements in coal.
The ISGS work consisted of a series of
interrelated studies on the modes of
sulfur and trace element occurrence. The
three major components of this work
addressed the level and variability of
trace elements in various coals, statistical
correlations between various trace
elements and the cause of such relation-
ships, and the organic affinity of individual
trace elements. To express an element's
organic affinity, ISGS developed an index
to estimate the distribution of a specific
trace element relative to the organic and
inorganic fractions of an individual coal.
To determine the index value, washability
tests are conducted at several specific
gravities and trace element contents of
each specific gravity fraction are mea-
sured. For a trace element predominantly
associated with the inorganic fraction of
the coal, the organic affinity index will be
near zero. For a trace element found
predominantly in the organic fraction, the
index value can be greater than 1.00.
Gravity separation studies conducted by
DOE and BCR focused primarily on the
potential for sulfur and trace element
removal during coal cleaning. J.A.
Cavallaro, M.T. Johnston, and A.W.
Deurbrouck tested 455 coal channel
samples to determine the potential for
ash and sulfur reduction from coal
cleaning. More recent studies of the ash
and sulfur reduction potential of Pennsyl-
vania, Ohio, West Virginia, and the wes-
tern U.S. have been initiated. Cavallaro,
G.A. Gibbon, and Deurbrouck conducted
gravity separation tests on 10 coals to
determine the potential for removal of
eight trace elements. Sulfur and trace
element removal studies conducted by
BCR analyzed 26 different coal samples.
While significant variations in trace
element reduction potential exist among
the coals tested in the studies, general
agreement exists between these values
and the ISGS organic affinity index.
As indicated above, gravity separation
methods provide only circumstantial
evidence as to the mode of trace element
occurrence in coal. To directly investigate
modes of occurrence and organic/
inorganic associations, photomicrograph-
ic methods, using optical and scanning
electron microscopes and microprobes,
are extremely valuable. The findings of
such studies are of specific importance to
understanding the environmental fate of
trace elements in coal, particularly as
they relate to trace element partitioning!
during coal cleaning. For example, a
-------
number of trace elements were found in
micronsized minerals throughout the
organic matrix: Zn and Cd in sphalerite,
Cu in chalcopyrite, Zr and Hf in zircon, and
rare earth elements Y and Th in phos-
phates. Because of their small grain size
and close association with the organic
matrix, these elements may have a high
apparent organic affinity based on
float/sink tests. On the other hand. As
and Hg were found to be in solid solution
in coarser-grained pyrites: they are,
therefore, more likely to be removed with
the sink material. In most coals, Pb was
found as a sulfide and as a substitute for
Ba in various barium-bearing minerals. In
some of the Appalachian Basin coal
samples, lead selenide was found more
frequently than lead sulfide. Significant
amounts of Se, Br, Be, B, Ge, Ti, and U
have also been found in association with
organic matter.
Conceptual models for predicting the
modes of occurrence for various trace
elements and sulfur in coal have been
developed by several different researchers.
The initial model, developed by P. Zubovic
et al. at the USGS during the 1960s,
focused on chemical bonding of trace
elements to coal's organic matrix. Models
developed by C.B. Cecil, R.W. Stanton,
and FT. Dulong and by R.B. Finkelman,
both at the USGS, have deemphasized
the importance of organic bonding for
most trace elements, and have focused
on the modes of occurrence and origin of
inorganic minerals in coal. The model of
Cecil et al. attributes the mineral matter
in most coals to authigenic materials
found in coal-forming plants, while the
Finkelman model postulates that detrital
materials washed or blown into the
ancient swamp system account for many
of the minerals present in coal. Cecil et al.
and Finkelman used some of the same
coal samples for obtaining support for
these two models, but drew different
interpretations on the origin of minerals
found in these common samples.
Relationships Between Coal
Properties and Utilization
Technologies
Coal Preparation
Several distinct processes can be used
to clean raw coal. The most commonly
used approach is mechanical separation
of inorganic minerals based on the
difference in specific gravity of various
minerals (s.g.=2.0-4.0) and pyrite(s.g.=5.0)
versus the organic coal fraction (s.g.=
.15-1.5). The raw coal is divided into a
'float" fraction (primarily containing the
lighter organic materials and fine-grained
"inherent ash" in coal) and a "sink"
fraction (consisting mainly of clays,
shales, pyrites, and other "extraneous"
minerals).
Sulfur and Ash Removal
DOE has analyzed the removal of sulfur
from roughly 750 raw coal samples from
throughout the U.S. The DOE study,
published in 1976, analyzed455 samples,
roughly half of which were from producing
mines in the Northern Appalachian
Region, and the balance from the other
major U.S. coal regions. More detailed
studies on the coals of Pennsylvania have
also been completed, and studies of Ohio,
West Virginia, and western U.S. coal are
underway. These studies have included
standard washability tests to determine
sulfur and ash removal potential for
various coals as a function of the specif ic
gravity of separation and size of the raw
coal feed.
The general conclusion of these
studies is that ash and sulfur contents
can be reduced for almost all coals by
physical coal cleaning (PCC). Figure 1
shows significant similarities in coal
washability for coal from the same seams
and regions. For example, excluding the
Allegheny Formation coals, most eastern
and midwestern coals show a greater
improvement in cleanability by grinding
from 3/8-in. (9.5 mm) to 14 mesh (1.18
mm) than from 1-1/2 in. (38.1 mm) to
3/8-in. (9.5 mm). The high sulfur
reduction potential for the Allegheny
coals may be explained in part by the
higher percentage of pyritic sulfur found
in the Allegheny coals, and may also
indicate that the size of pyrites in these
coals is larger—and thus more readily
separable than in other coals. Genetically,
this suggests that the Allegheny coals
may have formed under geologic condi-
tions which resulted in relatively low
levels of syngenetic sulfur fixation,
followed by precipitation of significant
amounts of coarse-grained epigenetic
pyrite.
Trace Element Removal
The removal of trace elements from
coal during coal cleaning has been
examined both in laboratory washability
studies and in analyses of operating
preparation plants. A total of 45 coal
samples have been evaluated in three
separate studies—23 from Appalachia,
13 from the Illinois Basin, and 9 from the
western U.S.
Element removal data for major and
trace species from these coals are
summarized in Table 4. The table is
divided into six regions—the Mononga-
hela, Allegheny, and Pottsville units of
the Appalachia Region, the Illinois Basin,
the Northern Great Plains, and the Rocky
Mountains. Dividing the Appalachian
Province into three units is based on the
differences in depositional conditions
believed to have existed during each of
these geologic periods and the resulting
impact on the coals. These geologic units
are also roughly equivalent to dividing the
Appalachian Province into a southern
region consisting of southern West
Virginia and states to the south (which
produce coals from the Pottsville Group)
and a northern region consisting of the
remainder of the province (which produce
primarily from the Monongahela and
Allegheny Formations).
These data indicate that Allegheny
Formation coals have the greatest
removal potential for most trace elements
and that Rocky Mountain coals have the
least potential. Due to the limited number
of samples available from each region
(especially from the Northern Great
Plains Province), the validity of the
computed values is subject to considerable
statistical uncertainty. However, the
favorable comparison of these results
with the sulfur reduction data obtained
by others suggests that the computed
values for trace element removal may
reasonably reflect the regional values.
The data on gravity separation of trace
elements from these earlier studies were
analyzed by Radian Corporation during
this study, using several statistical
techniques, including linear correlations,
cluster analysis, discriminant analysis,
principal components, step-wise regres-
sion, and regression of the logit transform*
of the data. Mineral suites identified from
correlation analysis of the feed coal (i.e.,
elements with correlation coefficients
>0.70) were the trace element sulfides
(As, Cd, Pb, Sb, and Zn), the clays (Al, K,
Si, Ti, plus the trace elements Co, Cr, and
V), and pyrite (Fe and S). Flourine
correlated with elements in several
mineral suites—the clays (Al, K, Si, Ti,
and Cr), the chalcophiles (Cu and Ni), and
Be.
Correlation tests of the trace element
removal data found that every element
examined (except sulfur and the chalco-
philes As, Cd, Hg, and Sb) correlated
strongly with total ash removal. The
chalcophiles As, Hg, and Fe, were
moderately correlated with pyritic sulfur,
while As, Co, Hg, Pb, Se, and Fe
correlated with total sulfur. Efforts to
*logit(p) = ln[p/(100-p)].
S
-------
GO i Note: Seam/Region <%S);
50-
40 •
10'
f //
<*»
!*»'
X
<<
Legendi
Appalachia
• Monongahela Fm.
— Allegheny Fm.
- Pottsville Gp.
Interior
——— — Western
i >
1'/,//? 3/e-//7 14 mesh
Feed, top size
Figure 1. Coal cleaning SO ? emissions reduction percentages for selected coals and feed sizes.
correlate percent removal of an element
with the quantity of that element in the
feed coal found that, except for elements
occurring in the clays, no strong correla-
tions existed. A logical interpretation of
this finding is that the abundance of
clays, rather than the abundance of the
trace element itself, determines the
extent of removal possible.
Regressions using the logit transfor-
mation were run on the data using
element removal as the dependent
variable and ash removal and sulfur
removal as the independent variables.*
Ash and sulfur removal are parameters
commonly obtained during standard coal
washability tests and, therefore, are
attractive as independent variables for
predicting trace element partitioning. The
strong statistical correlation of trace
element removal with ash and sulfur
removal in the existing data also supports
use of these two independent variables.
Table 5 gives results Of regressions
obtained for trace element removal
versus ash removal alone. Corresponding
to the results obtained using simple linear
correlation analysis, R-square values for
a number of elements are relatively good.
As expected, the data for Al, Si, V, and
other clay-associated elements fit the
predicted value line well. For As and other
elements more closely associated with
the sulfides in the coal than with the total
ash content, the scatter of data and the
size of the confidence intervals are large.
For the elements which exhibited
relatively strong correlations with sulfur
removal as well as for selected other
elements, fits of the element removal
were made as a function of both ash and
sulfur removal. These results are shown
in Table 6. The R-square values for Pb,
Co, As, Fe, Se, and Hg are significantly
improved. This result is to be expected,
given the mode of occurrence of these
elements in association with pyrite. Also
as expected, the R-square for the chalco-
philesCd, Cu, andZn — occurring as fine-
grained sphalerite and chalcopyrite
distributed throughout the coal matrix-
is only slightly influenced by the addition
of sulfur reduction as a variable. This
also reflects the fact that, although these
three trace elements occur in coal as
sulfides, only a small percentage of the
total sulfur in the coal is present as trace
element sulfides.
Combustion
Direct combustion is currently the
primary use of coal in the U.S. Most coal
is pulverized and fired in large boilers,
although a significant percentage is also
stoker-fired in smaller industrial and
commercial burners. During combustion,
trace elements undergo a variety of
transformations. Some are melted and
included in the aluminosilicate matrix
which forms the major structure of coal
fly ash; others fall to the base of the boiler
as bottom ash or slag; and still others are
'logit (ER) = A + (B x logit (AE)) + (Cxlogit (SR)) where
ER= percent element removal, AR = percent ash
removal, SR = percent sulfur removal, and A, B,
and C = fitted regress coefficients
-------
Table 4. Regional Summary for Removal of Major and Trace Element Species by Gra
Appalachian Province
Number of samples
Total S
Pyritic S
Organic S
Ash
Si
Al
Ca
Mg
Na
K
Fe
Mn
Ti
P"
Sb"
As"
Be"
B"
Cd
Cr
Co"
Cu
F°
Pb
Hg
Ni
Se"
v*
Zn»
Monongahela
9
29 ±15
40 ±20
4 ± 4
52+23
57 ±27
52 ±28
55+26
57 ±29
48+29
62 ±27
45 ±20
54 ±24
58 ±25
56 ±23
34 ±19
53 ±22
37+23
20 ±21
43 ±20
41 ±24
46 ±30
46 ±22
51 ±25
49 ±25
36 ±14
43 ±20
39 ±18
41 ±23
50 ±28
Allegheny
5
53 ±17
66 ±15
14 ± 9
73 ±12
78 ±12
72 ±13
51 ±f4
80 ±11
72 ±11
82+12
76+5
78 ±13
71 ±14
57 ±25
32 ±2
77 ±h
44 ±12
62 ±-
48+12
57 ±17
69 ±15
61 ±14
66 ±16
74 ±10
58+8
52 ± 7
65 ±13
59 ±16
65 ±10
Pottsville
9
27 ±18
52 ±22
9± 8
74 ±20
79 ±15
71 +21
56+26
79+17
70 ±25
81 ±19
65 ±24
75 ±24
70 ±23
64 ±25
28 ±13
75 ±10
43 ±20
30 ± 6
54 ±20
65 ±23
39 ±25
48 ±20
80 ±12
60 ±23
51 +20
46+23
43 ±17
60 ±28
69 ±24
vity Separation of 4
Interior
Province
13
38 ± 9
57 ±15
0± 5
48 ±15
47+22
42+21
55 ±34
53 ±25
44 ±28
47 ±27
61 ±19
57 ±25
42 ±24
67 ±44
17 ±16
61 ±20
22+10
10 ± 6
71 ±22
37 ±18
39 ±20
35+20
46 ±20
48 ±19
34 +22
31 ±18
35+16
29+15
57+27
14 U.S. Coals'
N. Great Plains
Province
2
50+8
53 ± 9
15+3
41 + 8
52+4
36+2
13+7
16 + 1
18+9
59 ±10
68 ±15
31 ±16
42+6
24+6
43+22
69+11
12+6
—
33 ±13
34 ±13
—
28+3
28 ±11
41 ± 4
36 ±21
5+2
46+13
16 ±13
47 +22
Rocky Mountain
Province
6
11+8
32 ±23
-9 ±13
35 ±23
22 ±30
14 +20
8 ±20
12 ±20
12 ±19
34+46
28+23
37+38
15 ±23
11+8
24 ±27
32 ±23
-5 ±31
1 ±-
35 ±26
24 ±15
3±-
14 ±11
24 ±15
34 ±18
25 ±20
15 ±15
15 +15
8 ±21
34 +30
"Whole-coal basis; mean + standard deviation.
"Mean and standard deviation based on less than total number of samples due to lack of data.
Table 5. Estimated Trace Element Removal During Coal Washing Based on Regression ofLogit
Transform of Ash Removal Data
Coefficients
Element
Al
Si
V
K
Na
Ti
Cr
Be
Pb"
Mn
Cu
Co"
Ni
Zn
P
Ca
Fe"
Se"
As"
"ff"
Cd
N
33
31
33
33
34
31
43
33
43
44
42
25
42
34
33
33
34
34
33
44
43
700% x/?2
55
94
91
90
86
85
84
75
74
74
70
68
67
66
56
49
48
47
42
38
20
Intercept"
-0.2107 (0.054)
0.2585 (0.056)
-0.9691 (0.077)
0.3398 (0.087)
-0.5754(0.116)
-0.0357 (0.089)
-0.6115(0.077)
-1.3232(0.106)
-0.2583 (O.O95)
-0.0376 (0. 150)
-0.6770 (0.092)
-0.8937 (0. 168)
-0.9964(0.110)
-0.025O (0. 145)
-0.3041 (0215)
-0.4688 (0.227)
0.1355 (0.157)
-0.6969(0.133)
0.3095 (0. 135)
-0.7126 (0.134)
-0.0358 (0. 183)
% Ash Removed*
1.0451 (0.042)
0.9285 (0.045)
1.1124 (0.063)
1.1668(0.068)
1.3246 (0.092)
0.9O40 (0.071)
1.0051 (0.069)
0.85O8 (0.087)
O.9227 (O.O85)
1.4249 (0. 13O)
0.7979 (0.082)
0.87O3 (0.126)
0.88O5 (0.097)
0.9127 (0.115)
1.0525 (0. 169)
1.0123 (O 185)
0. 6725 (0. 124)
0.5590 (0. 105)
0.4992 (0. 105)
0.5796 (0.115)
0.5024 (0. 156)
N= The number of coal samples used in the regression.
R* x 100 =R-square expressed as a percentage. This is the percent of total variation around a hori-
zontal (average) line that is reduced by tilting the line to the indicated regression line (R2 is
computed in the transformed scale).
"Coefficients are for the logit equation (see previous footnotes); the standard error for each
coefficient (a measure of the goodness of fit) is shown in parentheses.
"Elements showing significant improvement in R2 when fit includes both ash and sulfur removal;
see Table 6.
vaporized and oxidized. The volatile
elements may subsequently condense
onto the available surfaces (mostly fly ash
particles) or be vented as vapors. Nearly
all trace elements become enriched in the
ash since most of the mass (i.e., the
carbon) has burned away. For sulfur and
most trace elements, the nature of the
combustion and cooling processes (i.e.,
temperature profiles, residence time,
amount of excess oxygen) is more
important in determining the final form of
the element than its structure or form in
the original coal.
Table 7 presents the elemental distri-
bution patterns for the more common
major and trace elements based on data
from 15 studies where a mass balance for
trace elements was calculated. The
Emissions column refers to measurements
of the flue gas/fly ash mixture exiting the
plant via atmospheric (stack) emissions.
All the plants included in the averages
controlled particulate emissions with at
least 98 percent control efficiency.
The mass flow of the major ash-form-
ing elements, which generally are not
enriched or depleted in any waste stream,
give an idea of the total mass partitioning.
For example, for Al, Ca, Fe, Mg, and Ti, the
percentage into the bottom ash averaged
-------
Table 6. Estimated Trace Element Removal During Coal Washing Based on Regression ofLogit
Transform of Ash and Sulfur Removal Data
Estimated Coefficients
Element
Pb
Co
As
Fe
Se
Hg
Cd
Cu
Zn
N
43
25
33
34
34
44
43
42
34
100% xR*
74
51
88
68
52
79
42
58
70
48
53
67
47
47
63
38
36
50
20
17
25
70
19
72
66
24
67
Intercept*
-0.2583 (0.95)
0.7117(0.166)
0.2024 (0.096)
-0.8937 (0. 168)
0.2525 (0.247)
-0.3696 (0.202)
0.3095 (0. 135)
1.0303(0.147)
0 7483 (0. 147)
0. 1355 (0. 157)
1.0441 (0.187)
0.7004(0.183)
-0.6969 (0. 133)
0.0280 (0. 167)
-0.2741 (0. 164)
-0.7126(0.134)
-0.0126(0.175)
-0.2954(0.179)
-0.0358 (0. 183)
0.5608 (0.245)
0.2798 (0.273)
-0.6770 (0.092)
-0.0397 (0. 196)
-0.5295 (0. 129)
-0.0250 (0. 145)
0.8042 (0.273)
0.1211 (0.211)
Ash Removal*
0.9227 (0.085)
0.7268 (0.067)
0.8703 (0. 126)
0.6491 (0. 120)
0.4992 (0. 105)
0.3145(0.090)
0.6725(0.124)
0.4273(0.115)
0.5590 (0. 105)
0.3755 (0 103)
0.5796 <0. 1 15)
0.3958 (0. 120)
0.5024 (0. 156)
0.3633(0.179)
0.7979(0.082)
0.7477 (0.086)
0.9127(0.115)
0.8493 (0. 132)
Sulfu" RemovaP
0.7978(0.123)
0.4615(0.069)
0.8294 (0. 166)
0.4615(0.131)
0.6870(0.107)
O.4922 (0. 107)
0.7810(0.131)
0.5444 (0. 128)
0.6155(0.116)
0.408 (0.115)
0.6192(0.126)
0.4111 (0.130)
0.5024(0.175)
0.2999(0.196)
0.4826(0.159)
0. 1615 (0 102)
0.6110/0.191)
0. 1407 (0. 147)
N=The number of coal samples used in the regression.
R2 x / 00=R-square expressed as a percentage. This is the percent of total variation around a hori-
zontal (average) line that is reduced by tilting the line to the indicated regression line (R2 is
computed in the transformed scale).
a=Standard error in parentheses.
18.0 ± 0.8 and the hopper ash averaged
80.5 ± 0.8; this would leave about 1.5
percent for the emissions. Using these
values as indicative of total mass flow,
the enrichment/depletion of an element
in a particular output stream can be
assessed.
These data reveal:
• Ag, As, Cd, Cl, F, Hg, Pb, S, Sb, Se and
Zn are depleted in the bottom ash.
Only U appears to be enriched at all,
and the standard deviation for U is
large.
• Elements significantly depleted in
the hopper ash are Cl, F, Hg, Mo, and
S. Only Ag appears to be significantly
enriched in the hopper ash.
• Elements enriched in the fly ash/flue
gas emissions are As, B, Cd, Cl, Cr, F,
Hg, Mo, Ni, Pb, S, Sb, and Se.
• Note should be taken of the standard
deviation of each number; a large
deviation indicates a very broad
range of values or that one data point
was far from the average.
• Closure of from 70 to 130 percent is
within the expected limits of sampling
and analytical procedures. For values
outside this range, the partitioning
data are uncertain.
• Low closure values indicate that
some of the element was escaping
detection in the waste streams (e.g.,
Hg, F). High closure indicates that
either coal analysis was low or the
waste stream samples are easily
contaminated (e.g., Cl, Cr). Elements
with poor closure (or large uncertain-
ty) are Ba, Cl, Cr, Cu, F, Hg, Mo, Pb, S,
Sb, Se, U, and Zn.
Further analyses of these data were
carried out, but the results are uncertain.
For example, analysis of partitioning and
enrichment as a function of coal rank
indicated that several chalcophile ele-
ments (As, Fe, Pb, Zn) may be enriched
in the emissions stream of subbituminous
coals, but may be erroneous due to the
limited amount of data available.
Evidence, while sparse, suggests that
partitioning of the elements is largely
independent of the properties of the coal
and that elemental partitioning during
combustion is determined by the process
conditions (e.g., temperature and com-
bustor design). However, data are insuffi-
cient to detail these relationships in any
well-defined manner.
Synfue/s
A survey of published synfuel charac-
terization studies shows that most of the
sulfur originally in the coal is found in
the product gas and tar/oil by-products.
Roughly 90 percent of the sulfur found in
the gas phase is usually in the form of
hydrogen sulfide, with the remainder
consisting of carbonyl sulfide, methyl and
ethyl mercaptan, carbon disulfide, and
other organic sulfur species. Most of the
trace elements appear to remain in the
gasifier ash/slag and solids removed
from the product gas. General exceptions
are F, Hg, Se, Zn, and Pb. However,
closure on the trace element mass
balances is generally poor due to problems
inherent in analyzing for species present
in low concentrations and in the collection
efficiency of the sampling methods,
especially for gas-phase trace elements.
The available data on sulfur and trace
element partitioning in coal gasification
systems have not generally been corre-
lated with coal properties. However, as
would be expected, coals with higher
concentrations of sulfur and various trace
elements produce higher concentrations
of these species in the process streams.
Leaching of Coal Wastes
Coal utilization by the electric power
industry generates more than 70 million
tons of solid waste per year. These
wastes include refuse from coal prepara-
tion plants, bottom ash and fly ash from
coal combustion, and flue gas desulfuri-
zation wastes. Most of these utility
wastes are nonhazardous and are dis-
posed of in landfills. However, because of
their large volumes, constituents released
from coal-fired utility wastes through
leaching may constitute a potential
hazard.
A substantial amount of work has been
performed on the leaching of coal wastes.
However, the relationship of coal proper-
ties to the leachate from coal wastes, the .
principal point of this study, is not directly (
addressed in the open literature. Although
-------
Table 7. A verage Elemental Distribution of Fly Ash
Element
Ag
Al
As
B
Ba
Be
Ca
Cd
Cl
Co
Cr
Cu
F
Fe
Hg
K
Mg
Mn
Mo
Na
Ni
Pb
S
Sb
Se
Ti
U
V
Zn
No. of
Points
3
8
10
7
7
9
6
7
4
6
9
7
8
8
10
3
7
7
7
3
9
10
5
8
11
5
4
8
8
rercen
Bottom Ash
5.4 ± 3.5
17.1 ±5.4
5.8 ± 8.0
1 1.2 ± 8.8
15.8 ± 5.0
15.3 ± 6.0
17.8 ±5.3
9. 1 ± 5.4
6.9 ± 7.0
15.4 ±3.7
12.7 ±5.2
11. 9 ±3.2
2.5 ±2.6
19.5 ± 4.9
2.7 ±3.9
19.9 ± 1.2
18.0 ±4.3
16.8 ± 6. 1
12.9 ± 14.9
15.5 ±5.7
14.4 ± 6.4
7.0 ± 4.8
3. 1 ± 5.2
7.1 ±6.7
3.0 ± 3.9
17.6 ±5.6
26.2 ± 16.0
12.5 ± 6.5
10.0 ±8.7
i uistnouiion or c/<
Hopper Ash
91.8 ±5.2
81.6 ±5.4
84.4 ± 16.4
81. 3 ± 11.8
82.8 ± 6.4
82.5 ± 8.6
80.6 ±5.2
76.3 ± 20.9
7.8 ± 7.4
82.3 ± 2.5
80.5 ± 9.6
83.4 ± 8.0
60.2 ± 37.3
79.3 ± 5.0
31.3 ± 35.2
79. 1 ± 2.8
79.9 ± 5.4
80.8 ± 8.9
63.9 ± 29.8
83.4 ± 3.9
74.3 ± 13.9
83.5 ± 18.0
28.2 ± 36.4
85.6 ± 13.9
71. 1 ± 22.4
81.2 ± 5.2
71.8 ± 16.0
83.8 ± 7.9
85.1 ± 10.7
smem
Emissions
2.8 ±1.7
1.2 ± 1.6
9.7 ± 14.0
7.6 ±9.5
1.4 ± 3.0
2.2 ± 3.9
1.6 ± 2.0
14.6 ± 17.6
85.3 ± 10.8
2.3 ± 1.6
6.8 ± 8.8
4.7 ±7.3
37.4 ± 38.2
1.3 ± 1.5
66.0 ± 38
1.0± 1.7
2. 1 ± 2.4
2.4 ±4.4
23.2 ± 34.0
1.1 ± 1.8
11.2± 15.4
9.5 ± 18.1
68. 6 ±40.9
7.3 ± 12.9
25.9 ± 22. 1
1.2 ± 1.3
2.0 ± 1.9
3.7 ±3.5
4.9 ± 5.4
Closure
Percent
112 ±4
96 ± 12
114 ±32
106 ±36
128 ± 35
92 ±22
99 ± 16
102 ± 34
167 ± 140
98 ± 17
121 ± 65
83 ±32
73 ±37
115 ±10
62 ±51
104 ±24
103 ± 16
122 ± 16
141 ±85
98 ±23
113 ±32
100 ±60
77 ±50
115 ±83
64 ±30
92 ±23
65 ±25
100 ±41
127 ± 70
"Normalized to 100 percent by dividing measured values by closure value. Measured values equal
normalized distribution times closure value.
general ranges of coal ash leachate
quality can be approximated from infor-
mation on what was originally in the
coal, specific or quantitative correlations
between coal properties and those of coal
ash leachate are not available. The most
useful correlation of coal properties to
leachate properties may be accomplished
for coal preparation wastes which are
not affected by the combustion process.
Coal preparation wastes are produced
when mineral matter is removed from
coal prior to combustion, conversion to
synthetic fuels or coke, or other final
uses. The resultant waste contains a
variety of inorganic materials ranging
from fine-grained clays to sands, shales,
and macroscopic mineral nodules. Ele-
ments generally found in coal refuse at or
above the 1 percent level include Fe, S, Al,
Si, K, Ca, Mg, Ti, and Na. As, B, Cd, Cr,
Pb, Mn, Hg, Ni, Se, and Zn may also be
present in significant amounts.
Boiler fly ash is generally removed from
the flue gas by electrostatic precipitators
or fabric filters. The alkali, alkaline earth,
and rare-earth elements plus As, Se, Cd,
Pb, Hg, and U are concentrated in the
glass phase of fly ash, while Cr, V, and Ga
appear to concentrate in the crystalline
mullite phase. Due to their silicate
structure these materials are only
moderately teachable. Of greater concern
are the elements found in the more
teachable spinel oxide—V, Cr, Mn, Co, Ni,
Cu, and Zn.
Boiler bottom ash is a vitrified, relatively
inert residue which generally contains
lower concentrations of volatile trace
elements (As, Se, Cd, F, Mo, Pb, S, and Zn)
than fly ashes. While solid-phase fly ash
and bottom ash contain similar major
constituents, significant variations occur
in trace element composition. Leachates
from fly ash and bottom ash ponds in
general are also markedly different with
respect to trace element concentrations.
Flue gas desulfurization (FGD) can in-
volve several gas scrubbing processes.
The scrubber waste will contain small
amounts of entrained fly ash and volatile
trace elements absorbed from the flue
gas. Laboratory column leaching studies
of Ca-based FGD sludges have shown that
the first few pore volumes of leachate
contain elevated concentrations from the
entrained scrubber liquor. Subsequent
leachates level off to a dissolved solids
concentration controlled by the solubility
of the calcium sulfite, sulfate, or carbon-
ate.
Recommendations
Coal Quality
• Extension of Depositional Models
to Include Coal Quality. Depositional
models have been developed to
interpret the relationship between
ancient peat-forming environments
and coal deposit geometry. These
models can aid coal geologists in
interpreting the lateral continuity,
thickness, and overall tonnages in a
seam; and the association of coals
with other sedimentary facies. Cor-
relations have also been made
between depositional environments
and the ash and sulfur contents of
coals. The next logical step is the
extension of these models to help
predict additional coal quality infor-
mation based on depositional chem-
istry and other relevant parameters.
• Additional Study of the Origin of
Mineral Matter in Coal. Knowledge
of the origin of mineral matter in coal
is potentially of significant techno-
logical and environmental benefit by
assisting in prediction of coal quality
variations and in process selection
and optimization. Significant disa-
greements still exist among coal
geoscientists over the origin of
various minerals in coal. Currently
two significantly different models
exist. One model assumes that
detrital materials account for much
of the mineral matter in most coals.
The other model assumes that
detrital minerals are of minor impor-
tance. Using existing analytical
methods, support for both models
has been demonstrated using the
same coal samples. These models
need to be refined, and existing data
need to be reevaluated in order to
reconcile these differences. Where
appropriate, additional analytical
procedures and coal samples should
also be examined.
• Factors Influencing the Distribution
of Mineral Matter in Coal. Directly
related to the above recommenda-
tions is the analysis of genetic factors
which influence the extraneous
versus inherent nature of ash found
in coal. Extraneous minerals can be
readily removed from coals by con-
ventional washing, while inherent
ash cannot be removed. Coals such
as those from the Allegheny Forma-
tion have good washability, while
those from the Monongahela are
more difficult to clean. Determination
of the major difference in the coals as
a function of mineralogy, depositional
-------
environments, and other related
factors could be beneficial in evalua-
ting the use of future coal resources.
• Additional Analysis of the Mode of
Trace Element Occurrence. Existing
trace element data bases, such as
the USCHEM file currently maintained
by the USGS, are focused on providing
information on the quantity of trace
elements in U.S. coals, but do not
provide information on the modes of
occurrence for these elements.
Before compiling extensive data on
mineralogy or other characteristics,
however, standardized procedures
for evaluating samples need to be
developed. Existing methods are
expensive and imprecise.
Coal Utilization
• Analysis of Coal Washability for
Additional Coals. The 44 coals
included in this study include samples
from the major coal provinces in the
U.S., but provide little or no informa-
tion on a number of the major coal-
producing seams and regions in the
U.S. By comparison, sulfur removal
data is available for over 750 samples.
Specific attention in the additional
studies should be given to major coal
seams and regions which are pre-
sently washing coals or for which
coal washing is likely in the future.
• Impact of Mineralogy on Cleanability.
During this study the potential for
trace element removal from coals
was examined relative to the major
ash-forming elements and as a
function of the level of ash and sulfur
removable from the coal. Neither
approach directly evaluates the
minerals in the coals. Cleanability
needs to be examined as a function of
the minerals in the coals and the
modes of trace element occurrence.
• Additional Analysis of Existing Data
on Partitioning During Combustion.
The current study limited analysis of
the partitioning of trace elements
during combustion to studies for
which material balances were also
accomplished. Many additional stu-
dies (30-50) have been completed
which give elemental concentrations
in various combustion waste streams
without measuring flow rates. While
estimation of absolute elemental
distributions between process streams
is not possible for these studies, it is
possible to estimate enrichment
ratios. Using these studies, the data
base on partitioning of trace elements
during combustion could be expanded
10
significantly. Such a data base
should be large enough to allow a
more detailed analysis of statistical
relationships associated with com-
bustion partitioning.
Catalytic Effects of Trace Elements
on Sulfur Retention. Va, Fe, and
certain other elements can serve as a
catalyst for the conversion of S02 to
SOa. Because of the increased
efficiency by which SOs is absorbed
onto the surface of fly ash particles,
this catalytic effect may be important,
but has not been extensively investi-
gated.
Correlation of Leaching Characteris-
tics with Coal Properties. Little
research has been done on the
effects of coal properties on the
composition of combustion ashes
and their associated teachabilities.
The leaching potential of fly ash is
related to the abundance and compo-
sition of spinel (ferrite oxide) versus
aluminosilicates in the ash which in
turn is believed to be associated with
the composition of the coal. However,
these relationships have not been
well defined. When sampling coal
and coal ash in the same time periods
and under well known operating
conditions, the coal and the leacha-
bility of the ash need to be analyzed in
in detail, and correlations between
analyses examined.
D. M. White. L O. Edwards. A. G. Eklund. D. A. DuBose. F. D. Skinner. D. L
Richmann, and J. C. Dickerman are with Radian Corp., Austin, TX 78766.
James D. Kilgroe is the EPA Project Officer (see below).
The complete report, entitled "Correlation of Coal Properties with Environmental
Control Technology Needs for Sulfur and Trace Elements," (Order No. PB
84-200 666; Cost: $28.00. subject to change) will be available only from:
National Technical Information Service
5285 Port Royal Road
Springfield. VA 22161
Telephone: 703-487-4650
The EPA Project Officer can be contacted at:
Industrial Environmental Research Laboratory
U.S. Environmental Protection Agency
Research Triangle Park. NC 27711
"USGPO: 1984-759-102-10624
-------
'
2 = o
= -*, (D
n O 3
" -^ ^*
= 35
I", ff
s. o "
r^1
w 3
M =
£ 3
00 (D
5.
3D
------- |