United States
Environmental Protection
Agency
Hazardous Waste Engineering
Research Laboratory
Cincinnati OH 462%
Research and Development
EPA/600/S2-84/153 Mar. 1985
Project Summary
'/r,o
Auto-Oxidation Potential of
Raw and Retorted Oil Shale
D. A. Green
A study was conducted to assess the
potential spontaneous combustion
hazard of solid waste streams pro-
duced in the processing of oil shale. In
addition, the utility and precision of
various test methods to assess this haz-
ard were determined. The best meth-
ods found involved a differential scan-
ning calorimetry test and nonadiabatic
oxygen absorption test. In both cases,
the sample was slowly heated in air. In
the former case, exothermic activity
was monitored; in the latter, changes
in exhaust gas composition were moni-
tored.
The carbonaceous retorted oil shale
samples appear to present less hazard
of spontaneous combustion than
bituminous coal. None of the raw
shales were found to present as great a
hazard as Wyoming subbituminous
coal. However, the two raw Utah shale
samples tested were found to present
hazards exceeding those of less reac-
tive bituminous coals. These materials
represent a potential hazard, but if
stored and disposed of under condi-
tions suitable for long-term storage of
reactive coals, should not spon-
taneously combust.
The codisposal of raw and retorted
shale should be approached cautiously
and preferably be avoided if possible,
as the presence of raw shale increases
the energy contents of the mixture and
in some cases decreased the tempera-
ture at which exothermic activity was
observed. Codisposal of byproduct
sulfur with retorted shale appears to
cause no increase in risk of spon-
taneous combustion (but may lead to
leaching problems).
This Project Summary was devel-
oped by EPA's Hazardous Waste Engi-
neering Research Laboratory, Cincin-
nati, OH, 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
information at back).
Introduction
The main challenge in determining reac-
tivity of solids is to measure, in the labora-
tory, properties of the solids which can be
used to predict how full-scale storage piles
will behave. Factors that influence the
tendency of a storage pile to self heat and
eventually ignite can be grouped into two
main categories. The first category includes
those properties that are peculiar to the
solid itself: reactivity toward oxygen, and
heat release as a function of temperature.
The second group of factors relating to the
pile and its construction: overall dimen-
sions, particle size, degree of compaction,
homogeneity, ambient temperature, tem-
perature of placed material, precipitation,
wind speed, etc. It is quickly seen that pile
characteristics are going to be more dif-
ficult to measure and most likely subject to
more variation than the properties of the
solid itself.
In this study, an investigation of labora-
tory methods to determine spontaneous
combustion tendencies was conducted.
These methods were then applied to raw
and retorted oil shale samples to determine
the relative hazards presented by these ma-
terials. In addition, several coal samples
were tested by the same methods so that
the hazards of the oil shale materials could
be related to the better known phenome-
non of self heating in coal storage piles.
Materials and Methods
Retorted shale samples from the Lurgi,
Tosco II, Paraho direct-mode, Hytort and
Union B processes were tested. A raw
-------
shale sample from Federal lease tract C-a in
Colorado, two raw shale samples from
Federal lease tract Ua/Ub in Utah, Wyom-
ing subbituminous coal. Western Kentucky
#9 bituminous coal and Pocahontas #3 (Vir-
ginia) bituminous coal were also tested.
Sulfur and organic carbon contents and
heating values of these materials are given
in Table 1. Heat capacities of these materi-
als, as determined by differential scanning
calorimetry, are given in Table 2.
Differential Scanning
Calorimetry
Differential scanning calorimetry (DSC)
involves measuring the heat evolved or ab-
sorbed by a sample at a given temperature
relative to a known reference material at the
same temperature. When the temperature
is increased or decreased, heat effects arise
from differences in specific heats, phase
changes and chemical reactions. Low tem-
perature oxidation is an exothermic reac-
tion. When spontaneous heating occurs,
the heat produced by this reaction (and
other exothermic reactions including the
heat of wetting) is greater than the heat
which is rejected to the surroundings and
the material which is oxidizing increases in
temperature. As the temperature of the
material increases, the rate of oxidation in-
creases and, in some cases, a rapid temper-
ature increase occurs.
DSC data, obtained while the tempera-
ture of the sample and reference are in*
creased at a slow constant rate, indicate
the difference in heat flow between sample
and reference as a function of temperature.
When the reference is an empty sample pan
of equal mass and specific heat to the pan
in which the sample is held, the net heat ef-
fect is that produced by the sample. When
the temperature of the apparatus is re-
stricted to eliminate the possibility of phase
changes in the sample, endothermic effects
are associated with the heat capacity of the
sample, and with some materials, such ef-
fects as drying, desorption of gases, and
devolatilization. Exothermic effects in ex-
cess of the sample heat capacity are
associated with exothermic chemical reac-
tions.
Samples were tested in a duPont Instru-
ments Model 951* differential scanning
calorimeter equipped with a standard low
pressure cell. The calorimeter was coupled
to a duPont Instruments Model 1090/1092
thermal analyzer which permitted tempera-
Table 1. Description of Samples
Material (-325 mesh)
W. Kentucky #9 Bituminous Coal
Pocahontas H3 Bituminous Coal
Wyoming Subbituminous Coal
Utah Raw Shale (66 GPT)
Utah Raw Shale (28 GPT)
C-a Raw Shale
TOSCO II Retorted Shale
Hytort Retorted Shale
Paraho Retorted Shale
Union Shale Mixture
Lurgi Retorted Shale
Total
Sulfur
Weight %
Dry Basis
3.73
0.65
0.63
1.85
0.75
0.98
0.53
2.40
0.57
0.68
0.86
Organic
Carbon
Weight %
Dry Basis
61
69
64
28
11
5.2
3.5
3.8
3.1
4.1
0.13
Heating Value
Btu/lb
Dry Basis
13050
12120
11560
6240
2090
878
433
420
188
209
44
Table 2. Mean Heat Capacity of Coal and Shale Materials
Based on Initial Sample Weight {Jig, - 325 mesh)
Material
C-a Raw Shale
Utah Raw Shale (66 GPT)
Utah Shale (28 GPT)
Hytort Retorted Shale
Pocahontas tf3 Bituminous Coal
W. Kentucky #9 Bituminous Coal
Wyoming Subbituminous Coal
Lurgi Retorted Shale
TOSCO II Retorted Shale
Paraho Retorted Shale
Union Shale Mixture
25-200
0.76
2.18
1.17
1.08
1.19
0.69
1.28
0.82
0.39
0.73
2.57
Temperature Range (°
2&400
0.72
2.13
1.15
1.03
1.28
0.47
0.67
0.87
0.89
0.71
1.75
C)
25-600
0.73
2.56
1.23
1.17
1.40
0.44
0.63
0.92
0.86
0.66
1.60
'Mention of trade names or commercial products does
not constitute endorsement or recommendation for
use.
ture programming and data acquisition/
playback using magnetic disk storage.
Samples of between 17 and 54 mg were
used. Variations in sample mass for the dif-
ferent materials tested were primarily due
to differences in bulk density. Samples
were loaded into open aluminum pans; the
pans were completely filled to the top edge
and excess material was removed so that
the sample surface was level. An empty
pan of comparable mass (and heat capaci-
ty) was used as a reference.
With air flowing through the cell, a pre-
weighed sample in a tared aluminum pan
was placed on the sample thermocouple
disk and an empty sample pan was placed
on the reference thermocouple disk.
The test was conducted by programming
the heater to increase the temperature of
the thermoelectric disk to which sample
and reference thermocouples are attached.
A precisely controlled heating ramp of 2° C
per minute was followed from slightly
above ambient temperatures (25°-30° C).
Different samples were heated to different
final temperatures (380°-550° C) but in all
cases the test was continued to the end of
the exotherm. Exothermic reactions were
sensed by a slight lead in temperature of
the sample thermocouple over the refer-
ence thermocouple. The instrument was
calibrated to convert this lead into an actual
heat effect.
The most important measurement relat-
ing DCS data to spontaneous heating
behavior is the onset temperature of the ex-
othermic oxidation reaction. When an ex-
otherm is observed while a sample is being
heated, this temperature is characterized by
extrapolating the slope of the leading edge
of the exothermic peak to the baseline. It is
assumed that the lower this temperature is,
the greater the tendency of the material to
spontaneously heat. This provides a means
for empirically ranking materials of
unknown self-heating potential by com-
parison to materials of known heating po-
tential. The Wyoming subbituminous and
Western Kentucky bituminous coals ex-
hibited the earliest exothermic onset tern-
-------
peratures followed by the 66 gallon/ton raw
Utah shale. The low volatility bituminous
Pocahontas #3 coal was less reactive than
any of the raw western shales that were
tested. The carbonaceous retorted shale
samples exhibited considerably higher exo-
thermic onset temperatures than the raw
shale, indicating that they present less of a
spontaneous heating hazard than the raw
shale and much less than the coal. Decar-
bonized shale from the Lurgi process was
also tested but absolutely no exothermic
reaction was observed at temperatures up
to 550 ± °C. The Hytort retorted shale was
extremely unreactive, exhibiting an exo-
thermic onset temperature higher than that
of any of the retorted shale samples except
for the Lurgi retorted shale.
Nonadiabatic Oxygen
Absorption Test
The materials described above were also
subjected to a nonadiabatic oxygen absorp-
tion test.1 In this test, 100 g samples of
material were placed in a glass cell.
Humidified air at 60 cc/min was passed
through the cell and the temperature of the
cell and contents were increased at a rate of
25° C/hour. The exhaust gas from the cell
was analyzed chromatographically; the de-
pletion of oxygen and increased levels of
carbon dioxide were used as an indicator of
reactivity.
Based on the gas analysis, a spon-
taneous combustion liability index (the S in-
dex) was calculated. The formula for this
index was devised by Schmeling1 and is
based on the conversion of oxygen in the
inlet gas to C02.
where S = the spontaneous com-
bustion index
h, = 21 - oxygen concen-
tration at 125° C
h2 = 21 - oxygen concen-
tration at 150° C
ht = 21 — oxygen concen-
tration at 175° C
x, = carbon dioxide concen-
tration at 150° C
- carbon dioxide con-
centration 125° C
x2 = carbon dioxide concen-
tration at 175° C
- carbon dioxide con-
centration at 150° C
All concentrations are expressed in volume
percent.
The test indicated that the Wyoming
?ubbituminous coal was the most reactive
material with an S index of 108 in the - 48
+ 100 particle size test and an S index of
165 when tested in the -325 particle size.
This is consistent with the generally ob-
served phenomena of spontaneous heating
in low rank coals. The criterion, adopted by
Schmeling, that materials with S indices
greater than 30 (for the. -48 + 100 mesh
size) are dangerous, also puts the Western
Kentucky No. 9 bituminous coal (S = 37.5)
in this category. As would be expected
from historical observations of spon-
taneous combustion, the bituminous coals
are much less dangerous than the sub-
bituminous coal.
The tests indicated, as expected, that the
retorted shales are less likely to spon-
taneously combust than any of the coals or
raw shales. The three coals rank in the
order expected from past experience and
larger scale testing. The raw shales exhibit
oxygen absorption behavior which ranks
with their energy content. (The richer Utah
shale has a higher S index than the leaner
Utah shale and.the three western shales fol-
low the same order in S index as in higher
heating value.)
While none of the raw shales rank as high
as the subbituminous coal, the richer (66
GPT) Utah shale ranks intermediate be-
tween it and the high volatility Western
Kentucky #9 bituminous coal and the leaner
(28 GPT) Utah shale ranks intermediate be-
tween the Western Kentucky #9 coal and
the low volatility Pocahontas #3 bituminous
coal. On the basis of this test, the raw
shales present a hazard of spontaneous
combustion on the order of bituminous
coals, and the greater the energy content of
the shale, the greater the tendency for
spontaneous combustion. All of the re-
torted shales tested rank well below the
bituminous coals in spontaneous combus-
tion risk based on Schmeling's index.
In the course of the study, four other ex-
perimental methods were tried to determine
their suitability as spontaneous combustion
hazard indicators.
These tests were:
(a) a method based on weight loss data
from thermogravimetric analysis of
samples in air using a 2° C/min tem-
perature ramp.
(b) a method based on exothermic oxi-
dation of samples by hydrogen
peroxide.
(c) a method based on isothermal high
pressure differential scanning
calorimetry.
(d) an adiabatic test similar to the non-
adiabatic test described earlier.
These tests were found to be unsuitable for
both technical and practical reasons.
Conclusions
It must be noted that the results reported
here are based on a very limited number of
samples of raw and retorted shale.
Substantial variations in composition of
raw shales occur with geographic and
stratagraphic location. The retorted shale
samples came from pilot plant operations
which may not be completely represen-
tative of commercial operations. Hence, it
is strongly recommended that actual
samples of waste materials proposed for
field disposal be tested to determine their
actual spontaneous combustion hazard.
Retorted shales investigated in this study
are unlikely to present a spontaneous com-
bustion hazard. These include retorted
shales from the Paraho direct, TOSCO II,
Hytort, and Lurgi processes and a mixture
of retorted shale, raw shale "fines" and
sulfur from the Union B process. These
materials proved to be far less reactive than
Pocahontas #3 low volatility bituminous
coal which is generally regarded at the low
end of the spectrum of coals susceptible to
spontaneous heating. This conclusion was
reached on the basis of both the exother-
mic onset temperature, as determined by
differential scanning calorimetry, and the
nonadiabatic oxygen absorption test, and
supported by TGA weight loss data. The
Lurgi retorted shale is noncombustible and
could not burn even if an attempt was
made to ignite it.
The raw western shales, while not as
liable to ignite as the Wyoming Smith-
Roland subbituminous coal (which is
generally placed at the higher end of the
spectrum of coals susceptible to spon-
taneous heating) present a potential
hazard. The richer of the Utah shale
samples (66 GPT) is particularly reactive,
falling between the Wyoming sub-
bituminous and the Western Kentucky high
volatility bituminous coal (of intermediate
reactivity with regard to coal) in the
nonadiabatic oxygen absorption test. In the
DSC test, it falls below the Western Ken-
tucky coal but above the relatively unreac-
tive Pocahontas 13 coal. In the TGA weight
loss test (based on mass remaining after
heating to 300° C) a greater weight loss is
observed with this sample than with the
Western Kentucky bituminous coal.
The leaner Utah shale (28 GPT) is in-
termediate in tendency to autoignite be-
tween the Western Kentucky coal and the
Pocahontas #3 coal in both the non-
adiabatic oxygen absorption test and the
DSC test (ranked by exothermic onset tem-
perature). This material falls below the Coal
samples in the TGA weight loss tests. By
Schmeling's criteria, (S index < 30), this is
-------
also a potential hazard. The C-a shale has
about the same exothermic onset tempera-
ture as the leaner Utah shale (i.e., inter-
mediate between the Western Kentucky
and Pocahontas bituminous coals, but
ranks considerably lower than the
Pocahontas coal in the nonadiabatic oxy-
gen absorption test with an S index of
about 6). This is probably the least reactive
of the three western raw shales tested but
should still be regarded as posing potential
problems.
The results of the study suggest that car-
bonaceous retorted oil shale poses less
hazard of spontaneous combustion than
bituminous coal and it is logical that the
more severe the retorting process and the
more complete the removal of the organic
matter, the less reactive the resulting waste
product will be. This does not imply that
the risk of spontaneous combustion can be
ignored but rather that the risk should be
low if proper disposal practices are fol-
lowed. Good practices would likely include
cooling before disposal, compaction in lifts,
and excluding air flow into the pile. Decar-
bonized shales, such as the Lurgi decar-
bonized retorted shale, are essentially inert,
since substantially all of the energy content
has been removed. Such shales should pre-
sent no hazard of spontaneous combus-
tion.
On the basis of DSC testing, codisposal
of byproduct elemental sulfur with retorted
oil shale will not increase the hazard of
spontaneous combustion of the mixture.
The mixing of raw shale fines with retorted
oil shale for disposal should be approached
very cautiously and preferably be avoided.
The addition of as little as 5% raw shale fines,
appears to lower the exothermic onset tem-
perature of the mixture to approximately
that of the fines themselves. Although the
energy available in the mixture is much less
than that of the raw shale fines, the poten-
tial for spontaneous combustion may be
significantly increased as compared to the
retorted shale alone and may in fact be as
great as that of the raw shale fines alone.
For example, the addition of as little as 5
percent raw shale fines (Utah 28 GPT) low-
ered the exothermic onset temperature to
that of the raw shale fines alone. However,
it must be emphasized that the data on this
point are somewhat conflicting. The Union
B shale mixture which contained 5.47
weight percent raw shale fines behaved in a
manner similar to other carbonaceous
retorted oil shales which did not include ad-
mixture of raw shale fines. Hence, it is rec-
ommended that anyone proposing disposal
of retorted and raw shale mixtures should
evaluate their particular mixture to define
its properties. It is considered likely that for
each specific retorted/raw shale mixture,
there is a unique value of raw to retorted
shale ratio which, if exceeded, will cause
the mixture to assume the properties of the
raw shale.
The raw western shales present a poten-
tial hazard, but if stored and disposed of
under conditions suitable for long term
storage of reactive coals, then they should
pose no greater risk than such coals. Ideal-
ly, however, raw shales should be pro-
cessed for energy recovery to the greatest
extent possible. Material produced during
crushing operations that is too small for
retorting would be preferably subjected to
combustion for production of process heat,
or if feasible, agglomerated with a binder to
a size where it could be retorted for maxi-
mum energy recovery.
The differential scanning calorimetry test
and the nonadiabatic oxygen absorption
test appear to be the most meaningful test
methods used in this study. These tests
were reproducible and, when the tests were
applied to coal samples, the results con-
formed to generally observed rankings of
spontaneous combustion potential. Of the
other tests considered, the peroxide test
was technically unsound; pressure differen-
tial scanning calorimetry and adiabatic oxy-
gen absorption tests did not produce
useful, reproducible results in the course of
this particular study. These methods may,
however, after further development, be
made useful. The thermogravimetric
analysis weight loss test, while potentially
useful in characterizing samples was found
unsuitable as results for coal samples did
not conform to generally observed spon-
taneous combustion rankings and the
retorted shale samples produced an insuf-
ficient response to evaluate.
A summary of results obtained in the dif-
ferential scanning calorimetry and non-
adiabatic oxygen absorption testing is given
in Table 3. As no reference standards are
available for the test parameters which
were determined, the accuracy of these
tests cannot be determined. However, the
ranking of the materials can be used as an
indicator of relative spontaneous combus-
tion hazard.
Reference
1. Schmeling, W.A., etal., "Spontaneous
Combustion Liability of Subbituminous
Coals: Development of a Simplified Test
Method for Field Lab/Mine Applica-
tion," In: Analytic Chemistry of Liquid
Fuel Sources, Uden, P.C., et al. eds,
ACS 1978.
Table 3. Summary of Results From Differential Scanning Calorimetry and
Nonadiabatic Oxygen Absorption Testing
Material
Wyoming Subbituminous Coal
Western Kentucky #5 Bituminous Coal
Utah Raw Shale (66 GPT)
C-a Raw Shale
Utah Raw Shale (28 GPT)
Pocahontas #3 Bituminous Coal
Paraho Retorted Shale
TOSCO II Retorted Shale
Union Shale Mixture
Hytort Retorted Shale
Lurgi Retorted Shale
onset
°C
130
193
211
226
227
230
300
306
321
357
*«
DSC*
exotherm
Jig
10900
13800
8320
920
2990
15700
480
560
860
1340
~0
Nonadiabatic
test
S index
165
60
86
5.6
44
42
0.27
1.4
4.6
3.8
0.00
*Tested in dry air, particle size: - 325 mesh
**/Vo exotherm observed to 550° C.
AUSGPO: 1985 — 559-111/10790
-------
-------
D. A. Green is with Research Triangle, Institute, Research Triangle Park, NC
27709.
E. R, Bates is the EPA Project Officer (see below).
The complete report, entitled "A uto-Oxidation Potential of Raw and Retorted Oil,"
(Order No. PB 85-156 248/AS; Cost: $16.00, 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:
Hazardous Waste Engineering Research Laboratory
U.S. Environmental Protection Agency
Cincinnati, OH 45268
United States
Environmental Protection
Agency
Center for Environmental Research
Information
Cincinnati OH 45268
Official Business
Penalty for Private Use $300
OOOC329 PS
U S ENVIR PROTECTION AGENCY
CHICAGO
------- |