Environmental Protection Technology Series
PYROLYSIS OF INDUSTRIAL WASTES FOR OIL
AND ACTIVATED CARBON RECOVERY
Industrial Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development. U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
6. Scientific and Technical Assessment Reports (STAR)
7. Interagency Energy-Environment Research and Development
8. "Special" Reports
9. Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series. This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/2-77-091
May 1977
PYROLYSIS OF INDUSTRIAL WASTES
FOR OIL AND ACTIVATED CARBON RECOVERY
by
F. B. Boucher
E. W. Knell
G. T. Preston
G. M. Mallan
Occidental Research Corporation
La Verne, California 91750
Project Number S-801202
Project Officer
H. Kirk Willard
Industrial Pollution Control Division
Industial Environmental Research Laboratory
Corvallis, Oregon 97330
INDUSTRIAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has "been reviewed by the Industrial Environmental
Research Laboratory - Cincinnati, U.S. Environmental Protection Agency,
and approved for publication. Approval does not signify that the
contents necessarily reflect the views and policies of the U.S. Environ-
mental Protection Agency, nor does mention of trade names or commercial
products constitute endorsement or recommendation for use.
n
-------
FOREWORD
When energy and material resources are extracted, processed, converted,
and used, the related pollutional impacts on our environment and even on our
health often require that new and increasingly more efficient pollution con-
trol methods be used. The Industrial Environmental Research Laboratory -
Cincinnati (lERL-Ci) assists in developing and demonstrating new and improved
methodologies that will meet these needs both efficiently and economically.
"Pyrolysis of Industrial Wastes for Oil and Activated Carbon Recovery"
presents the results of pilot plant operations that pyrolyzed four industrial
residues and produced fuel oil and usable charcoal carbon. An economic eval-
uation indicated that a 1200-ton-per-day tree bark conversion plant could be
built and operated with a profit of $10 per ton of .-dry bark. Rice hulls and
grass straw produced similar results. For further information, contact the
Food and Wood Products Branch, lERL-Ci.
David G. Stephan
Director
Industrial Environmental Research Laboratory
Cincinnati
m
-------
ABSTRACT
The Garrett Research and Development Company has developed a new
flash pyrolysis process that can produce up to two barrels of synthetic
fuel oil from a ton of dry cellulosic solids. This report presents the
results of a four-phase laboratory, pilot plant, product evaluation, and
engineering evaluation program to study the pyrolytic conversion of
Douglas fir bark, rice hulls, grass straw, and animal feedlot waste to
synthetic fuel oil and char for either briquettes or powdered activated
carbon. Using an existing U-ton-per-day pilot plant, good quality pro-
ducts were obtained from all feedstocks except animal waste, which has
objectionably high concentrations of nitrogen, sodium, and potassium.
An interesting wax byproduct was obtained from the pyrolysis of fir bark
and grass straw. Excellent pilot plant material balances were obtained
for oil production runs on Douglas fir bark and standard test boiler.
Semiquantitative pilot plant runs on grass straw indicated that similar
yields of oil and char can also be expected from this feedstock. The
pyrolytic chars from tree bark and rice hulls were evaluated as a low
cost source of activated carbon, but equipment limitations led to rather
poor results. However, tree bark char was satisfactorily compressed to
produce excellent quality charcoal briquettes.
The economic evaluation shows that a 1200-dry-ton-per-day tree
bark conversion plant could be built for $13.3 million (excluding land)
and operated for $2.9 million per year, including amortization, with a
profit of about $10 per ton of dry bark. The break-even point for this
process to produce synthetic fuel 0il and char for briquettes appears to
be 300 dry tons of bark per day.
This report was submitted in fulfillment of Grant S-801202 by
Occidental Research Corporation (formerly Garrett Research and Develop-
ment Company, Inc.) under the sponsorship of the U.S. Environmental
Protection Agency. The report covers the period August 1972 to March
1975, and work was completed as of March 1975-
iv
-------
CONTENTS
Foreword iii
Abstract iv
List of Figures vi
List of Tables vii
Acknowledgments ix
Sections
I Introduction 1
II Conclusions 2
III Recommendations 5
IV Phase I - Laboratory Pyrolysis Studies
Introduction 6
Laboratory Pyrolysis Procedures 6
Experimental Results 11
Interpretation of Results 21
V Phase II - Pilot Plant Studies
Introduction 38
Pilot Plant Procedures 39
Material Balance 44
Experimental Results and Discussion 48
Material Balance Results and Discussion 59
Product Treatment and Properties 73
VI Phase III - Product Evaluation
Introduction .. 90
Product Preparation 91
Oil Evaluation 95
Char Evaluation 102
VII Phase IV - Economic Feasibility and
Preliminary Process Design
Introduction 108
Design Basis 108
Process Flow Ill
Plot Plan 113
Equipment Discussion 113
Economic Evaluation 118
Environmental Considerations 120
VIII References 149
Ix Appendices 150
-------
FIGURES
Number Page
1-1 Product yields vs. temperature 22
1-2 Distribution of volatiles in feed and
products: Rice hulls 23
1-3 Char devolatilization vs temperature:
Rice hulls 25
1-4 Distribution of volatiles in feed and
products: Animal waste 27
1-5 Char devolatilization vs temperature:
Animal waste 28
1-6 Product yields vs temperature 30
1-7 Distribution of volatiles in feed and
products: Fir Bark 31
1-8 Char devolatilization vs temperature:
Fir bark 32
1-9 Distribution fo volatiles in feed and
products: Grass straw 34
1-10 Char devolatilization vs temperature:
Grass straw 35
1-11 Pyrolytic oil yields from various feed
stock vs temperature 36
II-l GR&D 4 TPD pilot plant flowsheet 40
II-2 GR&D 4 TPD pilot plant plot plan 41
II-3 Pyrolytic oil viscosity vs temperature 75
II-4 Bark char screen analysis 80
II-5 Particle size distribution, electrically
produced tree bark char 81
III-l Corrected nitric oxide concentration
vs burner air flow 98
IV-1 Tree bark pyrolysis process flow
diagram and material balance 169
IV-2 Tree bark pyrolysis - plot plan 114
IV-3 Project schedule/ tree bark pyrolysis
facility 119
IV-4 System environmental balance, tree bark
pyrolysis, 1200 T/D (dry) 121
IV-5 System environmental balance, tree bark
pyrolysis, 300 T/D (dry) 122
VI
-------
TABLES
Number pag(
1-1 Laboratory Pyrolysis Results -
Rice Hulls 12
1-2 Laboratory Pyrolysis Results -
Animal Waste 13
1-3 Laboratory Pyrolysis Results -
Fir Bark 14
1-4 Laboratory Pyrolysis Results -
Grass Straw 15
1-5 Comparison of Experimental vs.
Calculated Char Heating Values 18
II-l Feed Analysis Summary 49
II-2 Macroscopic Liquids Yield Data -
Oil Production Runs 60
II-3 Summary of Operations Data 62
II-4 Macroscopic Liquids Yield Data -
Comparison to Isokinetic Results 63
II-5 Summary of Isokinetic Oil Yield Tests -
Rice Hulls 64
11-6 Gas Yield and Composition -
Pyrolysis of Tree Bark 66
II-7 Gas Yield and Composition -
Pyrolysis of Rice Hulls 67
II-8 Gas Yield and Composition -
Pyrolysis of Grass Straw 68
II-9 Elemental Balance - Pyrolysis of
Tree Bark 70
11-10 Elemental Balance - Pyrolysis of
Rice Hulls 71
11-11 Elemental Balance - Pyrolysis of
Grass Straw 72
11-12 Pyrolytic Oil Properties 74
11-13 Char Analysis Summary 78
11-14 Summary of Char Properties 79
11-15 Summary of Liquid Yield - Tree Bark 82
11-16 Summary of Liquid Yield - Rice Hulls 83
H-17 Summary of Liquid Yield - Rice Hulls 84
11-18 Summary of Liquid Yield - Rice Hulls 85
11-19 Summary of Liquid Yield - Grass Straw 86
11-20 Summary of Composition Data -
Tree Bark 37
11-21 Summary of Composition Data -
Rice Hulls 88
11-22 Summary of Composition Data -
Grass Straw 89
vii
-------
TABLES (Continued)
Number Pag<
III-l Properties of Bark Oil Shipped to KVB
Enginerring for Combustion Testing 92
III-2 Comparison of Pyrolytic Oil Properties
Measured by Garrett and KVB 93
III-3 Properties of Char Sent to St. Regis 94
III-4 Typical Properties of No. 6 Fuel Oil
and Pyrolytic Oil 96
111-5 Typical Oil Combustion Test Data 100
III-6 St. Regis Preliminary Char Evaluation .... 104
IV-1 Tree Bark Pyrolysis, Design Basis 109
IV-2 Capital Cost Estimate, Tree Bark
Pyrolysis 1200 T/D (dry) 124
IV-3 Capital Costs by Area, 1200 T/D (dry) 125
IV-4 Equipment List, 1200 T/D 126
IV-5 Summary of Operating Costs, 1200 T/D 130
IV-6 Operating Cost Estimate, 1200 T/D 131
IV-7 Manpower Summary (1200 T/D.and 300 T/D) .. 136
IV-8 Capital Cost Estimate, 300 T/D (dry) 138
IV-9 Capital Costs by Area, 300 T/D 139
IV-10 Equipment List, 300 T/D 140
IV-11 Summary of Operating Costs, 300 T/D 143
IV-12 Operating Cost Estimate, 300 T/D 144
-------
ACKNOWLEDGMENTS
The financial and material support of eight indepen-
dent organizations contributed greatly to the success of
this program.
The Georgia Pacific Corporation paid for 40 tons of
shredded Douglas fir bark and the Crown Zellerbach Corp-
oration paid for its transportation from Northern Cali-
fornia to the Garrett pilot plant facilities in La Verne,
California. The Rice Growers Association of California
paid for the milling and transportation of 20 tons of rice
hulls and supplied valuable consultation regarding utiliza-
tion of this waste product. The Oregon Seed Council paid
for the transportation of 10 tons of rye grass straw. The
World Farm Foundation provided samples of dried animal feed-
lot wastes. The Northwest Natural Gas Company provided
consultation advice throughout the course of this program
regarding markets for synthetic fuels. Finally, the State
of Oregon and the City of Portland, Oregon provided supple-
mental financial assistance in return for a slight increase
in the scope of work in order to provide synergistic infor-
mation regarding the conversion of the organic portion of
municipal refuse to synthetic fuels. The assistance of
these organizations is greatly appreciated.
IX
-------
SECTION I
INTRODUCTION
The Garrett Research and Development Company has dev-
eloped a new flash pyrolysis process which can produce over
two barrels of synthetic fuel oil from a ton of dry cellu-
losic solids. Preliminary tests showed that the residual
char from pyrolysis of homogeneous feedstocks such as tree
bark and rice hulls may also be valuable as a source material
for powdered activated carbon. Garrett Research built and
has successfully operated a four ton per day pilot plant for
studying this pyrolysis process on coal and on solid wastes.
Certain industrial wastes, as well as coal and municipal waste,
have also been pyrolyzed to produce synthetic fuel oil and
char in a smaller bench scale laboratory reactor. Based on the
pilot plant and bench scale reactor results, a proposal was
submitted to the Environmental,-Protection Agency in March, 1972
to investigate the pyrolysis of tree bark, animal feedlot waste,
rice hulls and grass straw using the Garrett process. In June
1972, EPA awarded Grant S-801202 in order to conduct a four
phase program covering laboratory pyrolysis studies, pilot plant
studies, product evaluation studies, and finally, an economic
feasibility and preliminary process design study.
The first effort of this program was initiated in August
1972, and was based upon the proven premise that by employ-
ing a short residence time process such as the Garrett Flash
Pyrolysis system, substantially higher liquid fuel yields
can be obtained from carbonaceous solids than had heretofore
been observed. Additionally, due to very rapid heat trans-
fer conditions, the resulting byproduct char tended to
possess a much larger surface area than chars produced by
slower, conventional fixed-bed or fluid-bed processes. The
purpose of this EPA program was to demonstrate on a modestly
sized pilot plant basis, the economic and technical viability
of employing this process for converting large volumes of
waste materials to saleable products which would possess a
wide market potential.
-------
SECTION II
CONCLUSIONS
1. High yields of synthetic fuel oil were obtained
from rice hulls, animal feedlot waste, Douglas fir bark and
grass straw in a single, short residence time bench scale
pyrolysis reactor and confirmed in the four ton per day
pilot plant. The yields were:
a) Grass straw - 30 to 50%
b) Douglas fir bark - 35 to 50%
c) Rice hulls - 25 to 40%
d) Animal waste - 30%
The yields were not sensitive to pyrolysis temperatures
in the temperature range 420 to 540°C (785 to 1005°F) except
for Douglas fir bark, where increased temperatures resulted
in higher oil yields, an optimum being about 510°C (950°F).
2. The pyrolysis of Douglas fir bark yielded approxi-
mately 5 wt % of wax as a crude by-product of the oil collec-
tion system. It was found that this wax could be extracted
from the pyrolytic oil during the normal quench procedure
using a relatively inexpensive paraffinic quench oil that is
immiscible with the non-wax portion of the pyrolytic product
oil. In several respects, this crude wax product was similar
to that obtained by commercial Douglas fir bark solvent extrac-
tion operations. A similar waxy product was also found in the
grass straw oil.
3. Good quality oil was obtained from all feedstocks
except animal feedlot waste, which yielded oil containing
5 to 7% nitrogen — unacceptably high for a boiler fuel.
In addition, the char from animal waste contained excessive
amounts of sodium and potassium. The disposal of raw animal
waste using the Garrett Flash Pyrolysis process is therefore
not considered feasible without additional pretreatment.
For this reason animal feedlot waste was deleted from the
pilot plant phase of the contract program.
-------
4. Based upon the results of the laboratory phase,
operating conditions for the pilot plant phase employing an
existing four ton per day facility were established for Douglas
fir bark, rice hulls and grass straw.
5. Excellent and accurate pilot plant results were ob-
tained during oil production runs. Two measures of pilot plant
accuracy are available: material balance closure of carbon
and hydrogen, and comparison of char yields by two independent
methods. The elemental material balances strongly support the
overall validity of these results. Average closures of C, H,
and 0 balances for the tree bark runs were respectively 96%,
97% and 104%. Similarly, the closures for the rice hull runs
were 99%, 100% and 100%. The statistic char-yield-by-difference
compares favorably with the char yield from the char production
runs when only electrical heaters on the reactor were used for
heat input. On a moisture and ash free basis, the by-difference
yield for tree bark char was 22.4% compared to the direct meas-
urement value of 22.3%.
6. Semi-quantitative pilot plant runs on rye grass straw
indicate that good quality fuel oil and usable char could
probably be produced. However, operating anomalies were
encountered which suggest that either further tests, or a
modest redesign effort would be required to handle this par-
ticular feedstock. These problems were tentatively attributed
to the unusual and difficult flow characteristics of the
resulting pyrolytic char.
7. An engineering process design and economic evalua-
tion of the Garrett Flash Pyrolysis process was made for the
conversion of Douglas fir bark to synthetic fuel oil and
saleable char products. The design calculations and cost
estimates were prepared for two commercial size plants using
the pilot plant data obtained in Phase II and the evaluations
of the oil and char as marketable products described in Phase
III.
The pyrolysis plant economics were estimated for feed
rates of 300 and 1200 tons per day (oven dry basis). The
products considered are a low sulfur tree bark oil suitable
for use as a utility boiler fuel or as a blend with No. 6
fuel oil in a utility boiler; and char which is marketable
as either a solid fuel, as charcoal briquettes or as a source
of inexpensive, powdered, activated carbon.
-------
The economic evaluation shovs that a commercial plant sized to
process 300 tons per day (dry) of tree bark will cost $5.0 million and
have an annual operating cost of $1.7 million, i.e. about $21 per ton,
including plant amortization. This size plant is expected to "break
even when the char is sold as briquettes. A 1200-ton-per-day plant,
producing similar products, shows a good return on investment with a
profit of about $10 per ton of dry bark. The plant will cost $13.3
million and have an annual operating cost of $2.9 million, i.e. about
$10 per ton, including plant amortization.
-------
SECTION III
RECOMMENDATIONS
1. Based on the very favorable economics projected by
this program, immediate steps should be initiated to fund
the construction of a 300 to 1200 ton per day prototype
pyrolysis plant to convert wood wastes to synthetic fuel oil
and char for briquettes.
2. Since higher oil yields might be obtained from
Douglas fir bark at higher pyrolysis temperatures, this
should be investigated in a very brief bench scale program.
3. Because high oil yields can be expected from grass
straw, the operating anomalies experienced in handling this
light, fluffy material should be investigated further and
effective operating techniques and design should be sought
for this feedstock.
4. The animal feedlot waste feedstock yielded synthetic
oil and char of unacceptable quality. Because of the intense
interest in finding an economical solution for the animal
feedlot waste problem, serious considerations should be given
to combining a protein recycling process with the Garrett
pyrolysis operation. Such a combination, if successful, could
result in about half of the manure's dry weight being recycled
as a 30-35% protein feed, with the remaining half pyrolyzed to
produce an acceptable synthetic fuel oil and a relatively low-
ash char.
5. Owing to the high potential value of the wax ob-
tained in the pyrolysis of Douglas fir bark, further studies
should be made to determine the steps required to separate
and purify this wax to obtain a saleable non-fuel product.
-------
SECTION IV
PHASE I - LABORATORY PYROLYSIS STUDIES
INTRODUCTION
Phase I of the study of the pyrolysis of industrial
waste presents the results of a laboratory evaluation program.
A five pound per hour bench scale pyrolysis reactor was used
to study the applicability of the Garrett flash pyrolysis
process for the conversion of Douglas fir bark, rice hulls,
grass straw and animal feedlot waste to synthetic fuel oil
and activated carbon. The effects of pyrolysis temperature,
feed moisture content and particle size on product yields
and quality were determined in order to set optimum condi-
tions for subsequent pilot plant operations using a four ton
per day unit.
The results of the laboratory study indicate that over
80% of the volatile matter in each of the four feedstocks can
be converted to synthetic liquid and gaseous fuels in a sin-
gle, short residence time pyrolysis operation. Oil yields
based on dry reactor feed were 30% from animal waste, 40%
from rice hulls and 50% from grass straw. Oil yields from
fir bark were 35 to 50% depending upon pyrolysis conditions.
Good quality synthetic fuels were obtained from all feed-
stocks except animal waste, which yielded oil with 5 to 8%
nitrogen and a by-product char containing high concentrations
of sodium and potassium salts.
LABORATORY PYROLYSIS PROCEDURES
Feed Preparation
Feedstocks for the laboratory pyrolysis phase were provided
by several private organizations as contributions to the
program.
Size analyses of the four feedstocks were of significant
economic importance, but were often difficult to determine
accurately. Tree bark and grass straw were especially
troublesome since these materials tended to ball up during
-------
dry screening' operations due to their fibrous nature when
milled to fine particle sizes. In the latter part of the
laboratory pyrolysis program, sedimentometer analyses were
performed on the feeds and product chars, and the size
distributions thus obtained always indicated smaller particles
than the dry screening procedure.
Rice Hulls -
Ninety kg (200 Ib) of rice hulls was obtained from the milling
of Calrose medium grain rice grown in the Sacramento valley.
The material received had a moisture content of 6.4% and was
essentially all minus 0.3 cm (1/8 inch) in size. About 34 kg
(75 Ib) was shredded using a 40 hp Rietz RD-12 disintegrator
with a l/16th inch screen. The resulting material was 95%
minus 20 mesh and was then dried under nitrogen for 24 hours
at 110°C. This material was used in pyrolysis runs 372-08
and 372-10.
Another 34 kg. of the as-received rice hulls was similarly
shredded with the Rietz disintegrator and then further
reduced in size by Vortec, Inc. of Los Angeles using their
Model M-l Impact Mill and Model C-l Air Classifier in series.
The material was estimated by Vortec to be approximately 70%
minus 200 mesh and was then dried in the same manner as
described above. This finely divided material was used as
feed for pyrolysis runs 372-14, 372-16, 372-18 and 372-20.
A subsequent shipment was received from the same source and
shredded in the Rietz disintegrator followed by further shred-
ding in a MikroPulverizer using a #010 screen, to substantially
under 200 mesh. At this point the material contained 6.8%
moisture, and it was used as the feed in runs 372-34 and
372-48 without further drying.
Animal Waste -
Ninety kg (200 Ig) of animal feedlot waste was obtained from
a dairy farm in the Chino valley of southern California.
The material received had a moisture content of 11.8% and
was minus 0.6 cm (1/4 inch) in size. Thirty-four kg (75 Ib)
was shredded in the Rietz disintegrator and further processed
by Vortec, Inc. in the same manner as the rice hulls, result-
ing in material which was approximately 70% minus 200 mesh.
This material was dried at Garrett Research under nitrogen
for 10 hours at 90°C and was used as feed in runs 372-22,
372-24 and 372-28.
7
-------
The remaining animal waste was later milled in a MikroPulver-
izer with a #010 screen and further screened to produce about
29 kg (65 Ib) of minus 200 mesh material. This was dried
under nitrogen for 24 hours at 90°C and used as feed in runs
372-40 and 372-46.
Tree Bark -
Ninety kg (200 Ib) of minus 0.6 cm (1/4 inch) Douglas fir
bark was snipped to Garrett Research from northern California,
The material received had a moisture content of 18%. Thirty-
four kg (75 Ib) was shredded in the Rietz disintegrator and
processed by Vortec, Inc. as described above. It was then
dried under nitrogen and used as the feed for runs 372-26
and 372-30.
A subsequent larger shipment of bark from the same source
was dried in a fluid-bed unit and then shredded in the Rietz
disintegrator for pilot plant operations. Some of this
material was screened to obtain about 11 kg (24 Ib) of minus
24 mesh material. This was further dried under nitrogen for
runs 372-32 and 372-44.
Grass Straw -
One hundred twenty-five kg (275 Ib) of annual rye grass
straw from the Willamette Valley was received unshredded.
About 34 kg (75 Ib) was shredded in a MikroPulverizer with
a 1/4 inch screen and a #46 punch screen. The resulting
material, minus 6 mesh, was further processed by Vortec,
Inc. in their impact mill and air classifier system to
produce 23 kg (50 Ib) of material which dry screening
showed to be 50% minus 200 mesh. Vortec experienced
difficulty in milling the grass straw and could process
the material at only one-third the usual rate, due to (a)
low particle density diminishing the effectiveness of the
impact mill, (b) long stringy particle shape, with L/D up
to 20 persisting after several passes through the mill, and
(c) electrostatic charges on the particles. This material
was dried under nitrogen in the same manner as the other
feeds, and used in runs 372-36, 372-38 and 372-50.
Bench Scale Pyrolysis System
The apparatus used for pyrolysis experiments in the
laboratory phase consisted of a vibrating screw feeder, a
reactor, three cyclone separators, a glass condenser train,
and a gas sampling system. Several modifications were made
to the previously existing collection system to improve the
accuracy of the results. Some improvements which were made
during the laboratory phase are described below.
-------
Gas Sampling
The gas sampling train provides a time-averaged sample for
analysis by mass spectrometry and gas chromatography . In
the earlier runs this was accomplished by drawing several
(usually five) equal increments of reactor off-gas into a
240 cc glass sample bulb by draining a Na-SO./H^SO. buffer
solution from the bulb at equally spaced times during the
run. The system and its operating procedure were described
in detail in an unpublished progress report (GR&D 72-038-1,
October 1, 1972) .
Usually two 250 cc gas sample bulbs were filled using this
technique, and both were analyzed. Occasionally the two
analyses conflicted significantly. This could have been
caused by uneven feeding, since bridging and clumping of the
fibrous feed materials at the feeder exit had been observed
previously. In an effort to obtain a more accurate time-
averaged sample, the gas sampling train was modified after
run 372-30. For all subsequent runs a bleed stream of gas
was drawn continuously and at a constant rate into a pre-
viously evacuated #3 gas cylinder. Two cylinders were filled,
each continuously over half the duration of the run. This
modification resulted in improved consistency between the
first-half and second-half gas analyses, and the gas
compositions reported for run 372-32 and subsequent runs
are considered more reliable than those for the earlier runs.
S Sampling -
At the beginning of the program it was recognized that con-
centrations of H2O and H2S in a gas sample are not readily
determined by mass spectrometry. A separate sampling train
was therefore used for these two components, consisting of
an aerosol filter, a Drierite column and a CuSO./Drierite
column. This train and its operation were also described
in the monthly progress report GR&D 72-038-1.
The Drierite column typically absorbed 0.5 to 2 g H-O during
a three hour run, while the CuSO./Drierite column aBsorbed
0.05 to 1 g H2S. The concentration of water vapor was
usually substantially higher than the concentration of
hydrogen sulfide, so that a small amount of breakthrough of
water from the Drierite trap into the H2S trap could increase
the apparent H_O concentration dramatically without decreas-
ing the apparent H-O concentration enough to be noticeable.
That such an event was happening was suspected upon examina-
tion of results from some of the earlier runs. The remedy
was to add a second Drierite column after the first as a
guard trap. For runs where the second Drierite column
showed a very small weight increase compared to the first, it
-------
could be assumed that no H20 was reaching the H2S trap.
Even with this improvement, however, gas composition from
some of the later runs showed unreasonably high E^S levels
which were inconsistent with other runs using the same
feed, and for which no obvious trend with temperature or
other operating conditions could be determined.
Aerosol Sampling -
At the beginning of the program an electrostatic precipitator
(ESP) was used to catch aerosol droplets which form when the
hot oil vapor produced during pyrolysis is suddenly cooled.
These aerosol droplets are often too small to be collected
in the glass condenser train or in the five micron filter
bag after the condensers. In bench scale pyrolysis tests,
the amount of oil which forms as aerosol is often one-fourth
to one-third of the total oil yield. It is therefore quite
important to recover this aerosol in order to obtain meaning-
ful results.
In analyzing the results of early runs, it was seen that the
aerosol production indicated by the weight gain of the Pall
and Gelman filters in the H20/H2S sampling train usually
exceeded by about 65% the amount of aerosol collected by the
ESP. The conclusion was that the ESP was not collecting all
the aerosol. Therefore the H2O/H2S sampling train data were
used to calculate aerosol production in all runs, and the ESP
was operated only for air pollution control and as a qualita-
tive indication of aerosol production.
Pyrolysis Operations
Before the start of a pyrolysis run, the entire system was
purged with inert gas to exclude oxygen. After the reactor
system reached its operating temperatures, the vibrating
screw feeder was started at a rate of 2 to 7 Ib/hr, usually
about 5 Ib/hr. The reactor exit temperature was recorded by
a thermocouple in the gas stream. This is referred to as
the pyrolysis temperature hereafter.
The hot reactor effluent was passed through three cyclones
to separate pyrolytic char from the gases. Pyrolytic oil
and part of the pyrolytic water were then condensed in two
water-cooled glass condensers and one cooled by liquid from
a dry ice/methanol bath. Some of the oil aerosol was then
caught by a five micron filter bag. Portions of the gas
stream were then diverted to the gas sampling and H20/H2S
sampling trains, and the rest of the gas was passed through
the electrostatic precipitator and an incinerator before
venting to the atmosphere.
10
-------
EXPERIMENTAL RESULTS
The results obtained from 21 runs carried out during the
laboratory pyrolysis phase are reported in Tables 1-1 through
1-4,
Product Compositions
Pyrolysis of solid waste by the Garrett process yields four
products — char, oil, gas and water. However, these products
are not completely separated by the bench scale apparatus
collection train, as they are in the Garrett pilot plant and
subsequent commercial plants. The procedures by which product
yields and compositions given in Tables 1-1 through 1-4 were
determined are discussed below.
Feed -
In every pyrolysis run, the feed material was routinely sub-
mitted to the Garrett analytical laboratory for proximate and
ultimate analysis. These analyses revealed moisture contents
up to 7%, depending on atmospheric conditions, the material's
storage history, and intentional variations in moisture level.
The moisture contents of the reactor feeds are given separately
in Tables 1-1 through 1-4, and the analyses shown have been cal-
culated on a dry basis.
Char -
The major part of the pyrolytic char product in each run was
caught by the first of the three cyclones. Small amounts of
char were also recovered from the other two cyclones and
from the piping connecting the last cyclone and the glass
condenser train. The weights of char from these locations
were all reduced by their respective moisture contents as
determined by proximate analysis, to arrive at "dry char."
In addition, a small amount of char fines usually passed
through all the cyclones into the glass collection train.
The tarry liquids from the glass collection train were there-
fore analyzed for quinoline-insoluble contents as an indica-
tion of the char carryover. This contribution to the char
yield was usually on the order of 1% of the total char yield,
which was calculated as follows:
ErWt char in cyclones, dry basis + •,
Dry char yield ~ Wt quinoline-insoluble in glass train x 100%
Wt feed, dry basis
11
-------
Table 1-1. LABORATORY PYROLYSIS RESULTS - RICE HULLS
tan Number
Feed Size, mesh
Reactor Exit Ifeip, °C
•P
Peed Analysis,
Wt * dry
C
H
N
S
Cl
Ash
0 (difference)
Feed Moisture, Wt %
Pyrolytic Char, dry
% of dry feed wt
Gross Heating
Value, cal/g a
Btu/Ub
C, Wt.%
H
N
S
a
Ash
0 (difference)
Pyrolytic Oil, dry
% of dry feed wt
Gross Heating
Value, cal/g *
Btu/lba
C, Wt.%
H
H
S
ci
Adi
O (difference)
Pyrolytic Gas, dry
% of dry feed wt
Gross Heating
Value, kcal/son
Bta/scf
H-. Mol %
CO
CO,
on
/2?l*
S?6
5
CJ+
HjS
Pyrolytic Waterb,
* of dry feed wt
Ibtal Products,
% of dry feed wt
372-08
-20
504
940
40.00
5.23
0.67
0.13
0.03
20.85
33.09
100.00
2.20
35.4
3200
5800
35.45
2.12
0.38
0.09
0.13
52.01
9.82
100.00
26.0
5700
10,300
62.66
5.61
1.26
0.07
0.14
0.28
29.98
loO.oo
10.4
1890
201
22.88
35.00
40.71
1.41
—
—
•
-
166.66
30.3
102.1
372-10
-20
421
790
38.80
5.24
0.48
0.14
0.20
19.93
35.21
loo. oo
0.0
53.9
3500
6200
36.55
3.74
0.70
0.14
0.24
39.70
18.93
loO.oti
18.6
6300
11,400
67.16
5.77
1.25
0.07
0.09
0.11
25.55
100.00
0
NA
17.5
90.0
372-14
-200
502
935
39.62
5.37
0.5S
0.09
19.26
35.11
100.00
0.0
26.1
3300
6000
35.44
2.22
0.22
0.16
53.69
8.27
100.00
50.4
5500
10,000
64.52
4.56
1.28
0.00
-
0.36
29.28
100.00
32.2
2750
292
40.09
22.45
28.85
7.47
0.17
0.06
0.06
0.23
0.62
100.00
15.1
123.8
372-16
-200
471
880
39.41
5.51
0.47
0.16
0.21
18.18
36.06
100.00
0.44
35.9
3400
6100
36.03
2.64
0.38
0.12
0.16
49.22
11.45
100.00
44.2
5800
10,400
62.44
5.81
1.37
0.10
0.29
0.60
29.39
100.00
12.1
2020
215
7.38
22.98
60.69
4.57
0.91
0.65
0.91
0.34
1.57
100.66
11.2
103.4
372-18
-200
427
800
39.63
5.21
0.43
O.I'.
0.21
18.04
36.34
100.00
3.2
50.1
3400
6200
37.68
3.21
0.29
0.13
0.20
40.86
17.63
100.00
30.5
5500
9900
59.83
5.87
0.92
0.08
0.31
0.52
34.47
100.00
U.I
2140
228
19.85
28.17
45.95
2.81
0.29
0.29
0.52
0.23
1.89
100.00
1.6
93.3
372-20
-200
518
965
38.47
5.16
0.44
0.13
0.19
20.47
35.14
loo. oo
2.97
31.3
3100
5700
33.47
2.34
0.30
O.U
0.20
53.94
9.64
100.03
49.8
5500
10,000
60.13
5.93
1.34
0.07
0.25
0.-32
31.96
160.00
20.3
2350
250
23.26
28.45
40.75
3.49
0.78
0.21
0.43
0.21
2.42
100.00
2.9
104.3
372-34
-200
521
970
39.74
5.04
0.37
0.01
0.13
18.81
35.90
166.00
6.52
43.5
3300
5900
36.99
3.17
0.16
0.00
0.08
39.09
20.51
100.00
29.5
5200
9400
56.65
6.15
0.74
0.04
0.09
0.56
35.77
100.00
4.0
3110
331
28.72
50.33
10.81
1.86
1.01
0.34
0.00
0,34
6.59
100.00
22.1
99.1
372-48
-200
510
950
37.29
5.04
0.82
0.18
0.12
23.77
32.78
loo. 'oo
5.45
34.5
2800
5100
29.14
2.44
0.86
0.09
0.07
48.14
9.26
100.00
28.2
5400
9700
57.18
6.04
1.64
0.03
0.11
3.94
31.06
loo. 06
7.6
2590
275
0.00
47.14
41.64
3.78
1.26
0.57
1.38
0.22
4.01
100.00
20.3
90.6
Calculated
' Excluding feed moisture
12
-------
Table 1-2. IAIWKATOKV
MKIJl.'lS - ANIMAL WAS'lT-i
Km
FocVl R.17.0, *t>«h
R \\ctor Kxit 'l\mp. , "C
°r
Feed Analysis, Wt." Dry
C
H
N
S
Cl
Ash
O (Difference)
Feed f-feisture, Wt.%
Pyrolytic Char, Dry
% of dry feed wt.
Gross Heating
Value, cal/g
Btu/lb
C, Wt.%
H
N
S
Cl
Ash
O (Difference)
Pyrolytic Oil, Dry
% of dry feed wt.
Gross Heating
Value, cal/g
Btu/lb
C, Wt.%
H
N
S
Cl
Ash
O (Difference)
Pyrolytic Gas, Dry
% of Dry Feed Wt.
Gross Heating
Value, kcal/sctn
Btu/scf
H-, Mol %
CO
%
c2f
&
Pyrolytic Water, * of
dry food wt.
Total prcxiucts, '' of
dry feerl wt.
372-22
-200
'.04
940
39.29
4.70
2.31
0.56
1.73
23.29
28.12
100.00
2.99
29.6
3027
5449
34.54
2.16
1.88
0.93
3.74
48.82
7.93
100. 00
16.8
6600
11,800
64.83
6.92
6.99
0.23
0.21
1.06
19.76
100.00
15.1
1900
202
5.89
19.63
62.19
5.57
0.58
0.80
0.58
0.35
4.41
100.00
12.2
73.7
372-24
-200
493
920
39.10
4.72
2.27
0.52
1.75
23.11
28.53
100.00
1.02
39.4
2990
5382
33.46
2.18
1.57
0.56
3.11
51.57
7.55
100.00
20.3.
6700
12,100
64.86
,7.04
6.48
0.25
0.40
3.05
17.92
100.00
19.4
2750
292
9.10
18.76
43.24
8.55
0.76
0.83
0.87
0.32
17.57
100.00
7.1
86.2
372-28
-200
421
805
39.01
4.79
2.36
0.50
1.78
23.33
28.23
100.00
1.22
48.1
3324
5983
36.29
2.65
1.99
0.68
3.30
50.28
4.81
100.00
20.0
6200
11,100
61.97
6.76
5.25
0.26
0.17
1.58
24.01
100. 00
10.8
2130
226
6.08
21.90
55.90
6.08
0.46
0.75
0.80
0.23
7.80
100.00
10.1
89.0
372-40
-200
477
890
36.34
'..55
2.W
.76
1.48
29.44
24.75
100.00
1.82
52.9
2500a
4400
28.23
2.25
1.78
0.78
2.10
51.12
13.74
100.00
14.4
6400
11,500
63.96
6.37
6.91
0.42
0.52
4.20
17.62
100.00
3.6
3170
337
11.51
31.89
46.52
4.08
0.96
0.96
0.24
1.92
0.96
100.00
12.6
83.5
372-46
-200
'jlO
950
39.81
5.05
2.09
.74
1.70
19.80
30.81
100.00
0.77
44.0
2600a
4600
29.77
1.93
1.76
0.68
2.22
51.95
11.69
100. 00
17.0
6300
11,400
61.80
7.23
6.52
0.39
0.13
1.85
22.08
100. 00
9.5
1530
163
7.63
21.58
63.33
2.28
0.40
0.40
0.51
0.23
3.64
100.00
10.7
81.2
Cilojlntod
Kxcludincj
13
-------
Table 1-3. 1V\U01
C, Wt %
H
N
S
Cl
Ash
O (difference)
Pyrolytic Oil, dry
% of dry feed wt
Gross Heating
Value, cal/g
Btu/lb
C, Wt %
H
N
S
Cl
Ash
O (difference)
lyrolytic Gas, dry
% of dry feed wt
Gross Heating
Value, Real/son
Btu/scf
Ho, Mol %
372-26
-200
510
950
49.50
5.56
0.14
0.04
0.10
8.69
35.97
100.00
0.79
27.8
4211
7579
48.15
2.38
0.23
0.09
0.37
33.74
15.04
100.00
33.5
6200
11,200
62.57
6.84
0.63
0.04
0.07
0.35
29.50
100.00
14.2
3820
406
14.05
14.43
39.23
15.25
1.65
1.05
1.05
1.65
11.64
100.00
372-30
-200
432
810.
48.30
5.31
0.17
0.04
0.20
11.72
34.26
100. 00
1.54
41.6
4589
8260
49.86
4.02
0.14
0.09
0.17
21.38
24.34
100.00
28.7
5700
10,300
60.45
6.04
0.48
0.07
0.16
2.08
30.72
100.00
8.2
2090
222
6.26
14.66
64.70
10.46
1.03
0.37
0.84
0.00
1.68
100. 00
372-32
" -200
471
SBO
49.09
5.47
0.19
0.05
0.04
9.92
35.24
100.00
4.20
55.5
4500a
8100a
49.93
4.35
0.17
0.14
0.04
19.67
25.70
100.00
28.0
5400
9700
57.73
6.35
0.29
0.02
0.07
0.13
35.41
100.00
4.3
4800
510
22.37
22.37
32.18
6.13
1.43
14.30
0.71
0.31
0.20
100.00
372-42
-24
510
950
52.20
4.21
0.32
0.16
0.02
5.65
37.44
100.00
0.0
39.3
5300a
9500a
58.33
4.34
0.28
0.10
0.02
13.16
23.77
100.00
24.7
6400
11,600
62.98
7.25
0.38
0.13
0.04
0.58
28.64
100.00
5.7
2700
287
1.89
39.84
49.45
3.94
1.57
0.63
0.79
1.42
0.47
100.00
372-44
-200
538
1000
49.03
5.37
0.27
0.09
0.03
10.72
34.49
100.00
2.68
31.7
4400a
8000a
51.32
2.92
0.22
0.07
0.03
27.42
18.02
100.00
27.8
7000
12,700
69.12
6.80
0.46
0.00
0.02
1.23
22.37
100.00
47.8
1610
171
2.32
28.42
63.80
2.53
0.63
0.23
0.44
0.48
1.15
100.00
Pyrolvtic Kator ,
t of dry food wt 1.9 15.2 9.2 14.6 15.8
Tbtal Products,
» of dry food wt 77.4 93.7 97.0 84.3 123.1
' Fbrcliviirv; Ctxxl timstuir-
14
-------
Table 1-4. lAISORATOTO PYR3LYSIS RESULTS - GRASS STRflW
Run Number
Feed Size, mesh
Reactor Exit Temp., °C
T
Fteed Analysis, Wt % dry
C
H
N
S
Cl
Ash
O (difference*
Peed Moisture, Wt %
Pyrolytic char, dry
% of dry feed wt
Gross Heating Value, cal/g
Btu/lba
C, Wt %
H
N
S
Cl
Ash
O (difference)
372-36
-200
482
900
44.96
5.97
0.50
0.53
0.35
5.70
41.99
100.00
0.0
23.2
4600
8300
50.98
3.69
0.45
0.84
0.53
24.28
19.23
100.00
372-38
-200
510
950
45.16
70
60
62
0.26
5.92
41.74
100.00
1.36
20.5
4600
8300
372-50
-200
538
1000
44.76
5.72
0.59
0.61
0.26
6.75
41.31
100.00
1.44
15.0
4500
8100
48.43
3.01
0.56
0.94
0.61
34.03
12.42
100700
Pyrolytic Oil, dry
% of dry feed wt
Gross Heating Value, cal/g
Btu/lb
C, Wt %
H
N
S
Cl
Ash
O (difference)
35.7
5200
9400
58.55
5.57
1.33
0.08
. 0.10
0.50
33.87
100.00
34.9
5500
9800
58.22
6.21
1.11
0.19
0.08
1.03
33.16
100.00
31.7
5700
10,200
59.58
6.29
1.20
0.08
0.12
1.16
31.57
100.00
Pyrolytic Gas, dry
% of dry feed wt
Gross Heating Value, kcal/scm
Btu/scf^
H,, Mai %
Pyrolytic Water , % of dry feed wt
Total Products, % of -dry feed wt
8.1
3090
329
7.43
51.07
28.14
6.05
0.96
0.84
0.72
0.48
4.31
100.00
19.5
83.0
13.9
2710
288
1.15
44.11
41.72
5.52
1.07
0.69
0.99
0.69
4.06
100.00
23.9
84.5
1 Calculated
' Excluding feed moisture
15
-------
The composition of the dry char reported in Tables 1-1 through
1-4 is the mass average of the compositions of the dry-basis
chars removed from the three cyclones. For example/
Per cent carbon zWt carbon from each cyclone 100%
dry basis sTotal dry wt from each cyclone x
Oil -
The major part of the pyrolytic oil in each run was recovered
from the glass condenser train and the five micron filter
bag. As much product as possible was recovered by decanting
from the various vessels, and the total amount of material
was determined by weighing the decanted portions and weigh-
ing each piece of the glass train both before and after the
run. The decanted portions were submitted for analysis, and
their moisture contents and quinoline-insoluble contents
were subtracted to determine the amount of dry, char-free
pyrolytic oil collected from each section of the glass
collection train and the filter bag. In addition, the
amount of aerosol oil product was calculated from the
weight of aerosol caught on the Pall filter and the Gelman
filter in the H_O/H_S sampling train, and the measured
volume of the portion of reactor off gas passing through
the I^O/KLS train. The total oil yield was calculated as
follows:
Wt oil in glass vessels, dry, char-free +
{Wt oil in filter bag, dry, char-free + }
Dry oil yield = Wt aerosol oil calculated, dry basis x 100%
Wt feed, dry basis
The composition of the dry pyrolytic oil reported in Tables
1-1 through 1-4 is the mass average of the compositions of the
dry-basis material decanted from the three collection flasks
in the glass condenser train. The implicit assumption that
this mass average composition is representative of all the
pyrolytic oil, including the undecanted residue, the filter
bag catch, and the aerosol, is believed to be quite accurate.
Gas -
The composition of the pyrolytic product gas from each run
shown in Tables 1-1 through 1-4 was determined as follows.
For each component (H2, CO, etc.) the ratio (moles component/
moles carrier gas) was calculated from the analyses of both
gas samples and the results averaged. From these average
component ratios, as well as from the weight of H2S caught
16
-------
by the CuSO./Drierite trap in the E^O/E^S sampling train,
the carrier-gas-free pyrolytic gas composition was determined.
The pyrolytic gas yield was then calculated from the composi-
tion and the measured volume of carrier gas used during
the run.
In spite of the improvements made to the gas sampling train
during the program, a number of factors adversely affect the
reliability of the reported gas yields and compositions.
1. The low feed rate sometimes obtained with the vibrating
screw feeder caused the gas sample to be over 99% carrier
gas in some cases. This made an accurate gas analysis
difficult.
2. The concentrations of C02 and H_ varied from run to
run in a way apparently unrelated to temperature or other
operating conditions. These variations can be seen in
Tables 1-1 through 1-4. However, this problem may be due to
the mass spectrometry/gas chromatography analytical
technique, since frequent inconsistencies were noted
concentrations in the two gas samples
same run.
between H- and CO
collected from th
Since the gas yield is calculated on the basis of the gas
composition, whatever unreliability is attached to the
composition extends to the yield also.
Water -
Even though pyrolysis occurs in the absence of oxygen in
the Garrett process, a significant amount of water is
produced due to the fairly high oxygen content in the feed
materials. In the bench scale pyrolysis tests, part of the
pyrolytic water was recovered from the glass condenser
train and the five micron filter bag, as moisture in the
oil and occasionally as a separate aqueous phase. The
additional water produced was calculated from the weight
gain in the Drierite traps and the measured volume of
reactor gas passed through the H20/H2S train. The moisture
content of the recovered char also aaded a small amount to
the water yield. The total water yield was calculated as
follows:
"Dry" water
yield
Wt moisture in cyclones +
r Wt moisture in glass vessels and filter +-,
Wt moisture calculated from trap -
Wt moisture in feed
Wt feed, dry basis
x 100%
17
-------
While pyrolytic water and oil are collected together in
the laboratory equipment, they will be collected separetely
in the pilot plant using procedures similar to those en-
visioned for a commercial plant. However, based on past
experience on pyrolysis of municipal solid waste, the
separated pyrolytic water is still expected to retain a
fairly high concentration of organics which, while completely
biodegradable, could cause a secondary pollution problem.
In the design of a commercial plant, therefore, careful
attention must be given to reducing such pollution to a
minimum.
Product Heating Values
Gross heating values, or higher heats of combustion, for the
pyrolytic oils and chars are reported in Tables 1-1 through 1-4
Some of the char heating values were determined by an out-
side testing laboratory, but all others were calculated from
the empirical Dulong-Berthelot correlation:
cal/g = 81.37(%C) + 345 (%H - %N * g "-) + 22.2(%S)
Empirical correlations of this type were originally developed
for coal-like fuels. However, this correlation does provide
a good approximation for the heating values of these solid
waste chars, as indicated by comparing bomb calorimeter
values with calculated results shown in Table 1-5. The Dulong
correlation is known to be valid for pyrolytic liquid fuels
also, as evidenced by calorimeter data obtained from earlier
Garrett studies on municipal solid waste.
Table 1-5. COMPARISON OF EXPERIMENTAL VS.
CALCULATED CHAR HEATING VALUES
Bomb Calorimeter Dulong-Berthelot
Run Number Results, cal/g Calculated Results, cal/g
372-22 3027 3200
372-24 2990 3100
372-26 4211 4100
372-28 3324 3600
372-30 4589 4400
18
-------
Material Balance
The raw product yields reported in Tables 1-1 through 1-4 do
not always add up to 100%. However, the conversions of dry
feed to products must be made to total 100% in some reason-
able way in order to draw rational conclusions from the
laboratory data. The manner in which these "correct" product
yields were determined is summarized below.
The material and elemental balances implicit in the forma-
tion in Tables 1-1 through 1-4 can be regarded as a system of
linear equations. In order to choose which quantities should
be unknowns and which information (equations) should be used
to solve for these unknowns, the following guidelines were
observed.
1. A high degree of reliance can be put on the reported
pyrolytic char yield, since the recovery of this product
is quite straightforward, i.e. weighing the material.
2. A fairly high degree of reliance can be put on the
reported ultimate analyses of the feed and the pyrolytic
char.
3. The reported gas composition should not be relied upon
too heavily due to known sampling variations.
4. The reported pyrolytic oil, gas, and water yields should
not be heavily relied on.
For each pyrolysis run, two calculations were made (by a
simple computer program).
1. The four product yields — per cent char, oil, gas and
water — were considered unknown, and the four linear
equations used to solve for them were:
a) Total products = total feed
b) Ash in char and oil = ash in feed
c) Carbon in char, oil and gas = carbon in feed
d) Hydrogen in all four products = hydrogen in feed
2. The reported per cent char yield was accepted as known,
and the other three product yields were considered unknown.
The three equations used to solve for them were the over-
all, carbon and hydrogen balances, i.e. (a), (c), and (d)
above.
19
-------
In 17 cases (out of 21 runs) the char yield calculated by
the first procedure agreed closely with the yield reported
in Tables 1-1 through 1-4. In these cases the calculation by
the second procedure was adopted, i.e. the reported char
yield was accepted as being accurate, and only the oil, gas
and water yields were calculated. In four cases the char
yield calculated by the first procedure was significantly
different (by up to 17% yield) from the reported char yield,
and in these cases, the calculation from the first procedure
was adopted.
Several spot checks were made on the sensitivity of the cal-
culation procedures to variations in input data, and the
following observations are noted.
1. Variations in the reported ash, carbon and hydrogen
contents of the feed had a significant effect on all cal-
culated yields. This is disturbing because ultimate
analyses of the feed materials occasionally did show un-
expected changes, due perhaps to failure to obtain a
representative sample for analysis.
2. Variation of the ash content in the pyrolytic oil had
a negligible effect on the calculated product yields. .
3. Variation of the CO/H_ levels in the pyrolytic gas
composition had a negligible effect on the calculated
product yields.
All subsequent interpretation in this report of the results
of the laboratory pyrolysis phase is based on the "corrected"
yields of pyrolytic char, oil, gas and water determined by
the procedures discussed above. The justification for this
is as follows:
1. Meaningful interpretation and comparison of yield
results cannot be made if yields do not total 100%.
2. The procedure used to determine the "corrected" yields
is simple and a minimum of subjective judgment was in-
volved .
3. In most cases, the pyrolytic oil yield was the only
value substantially changed from the raw data yield. The
implication is that the "adjustments" are thus not arti-
ficial, but are corrections for experimental limitations
in the product collection system and/or procedure.
20
-------
INTERPRETATION OF RESULTS
Rice Hulls
Yields of dry pyrolytic char, oil, gas and pyrolytic water
are shown in Figure 1-1. The char yield clearly decreases
as the pyrolysis temperatures are increased. However, the
oil yield of about 40% is apparently independent of
pyrolysis temperature over the range investigated.
The effect of increased feed particle size is seen in the
char yields plotted at 421 and 504°C. These two runs were
made with minus 20 mesh feed rather than the minus 200 mesh
used for the other six runs. The two char yields are higher
than would be expected from the trend with temperature,
implying that the larger particles are not completely pyro-
lyzed in the short residence time provided in the bench
scale reactor. However, the oil yield at 504°C was
apparently not adversely affected by the larger particle
size.
The effect of feed moisture content is seen in several of
the rice hull runs. The runs conducted at 510 and 521°C
were made with feed containing over 5% moisture, and the
respective char yields are high compared to the observed
trend with respect to temperature. The runs at 427,504
and 518°C were made with feed containing only around 2.5%
moisture, and the char yields of the latter two also seem
high. On the other hand, moisture content does not appear
to have much effect on oil yields. A reasonable explanation
of the effect of feed moisture is that additional residence
time is required in order to achieve a given extent of
pyrolysis, much as larger feed particles require additional
time.
One of the economic attractions of flash pyrolysis is the
high conversion to volatile products which is obtainable
compared to longer residence time processes. In fact,
numerous studies are reported in the literature (e.g. Badzioch
and Hawksley ) where the dry weight lost during flash
pyrolysis of coal is often greater than the volatile content
as determined by ASTM test methods. It is of interest,
therefore, to see what happens to the moisture and ash free
portion of the industrial wastes investigated during flash
pyrolysis. Figure 1-2 shows the volatile components in the
feed and in the products from pyrolysis of rice hulls.
Each pyrolysis run is represented by a group of three
columns. The first column shows the per cent of the
moisture and ash free rice hulls feed which reports as gas
and as tar plus light oil, by a Fischer assay to 500°C.
21
-------
FIGURE I—I. Product yields vs. temperature,
NJ
CO
a
H
W
><
Q
O
OP
*
D
W
H
>H
60
50
40
30
10
REACTOR EXIT TEMPERATURE, °F
800 850 900 950 1000
I
I
T
W/fTER
I -
60
50
40
30
20
10
400 450 500
REACTOR EXIT TEMPERATURE, °C
REACTOR EXIT TEMPERATURE, °F
800 850 900 950 1000
I
T
GAS
550 400 450 500
REACTOR EXIT TEMPERATURE,
550
-------
FIGURE 1-2. Distribution of volatiles in feed and products: Rice hulls
ro
CO
100
80
60
40
20
0
RUN NO
E-i
Z
w
w
(X
ac
o
H
I
\
372-08
372-10
372-14
372-16
372-18
372-20
372-34
372-48
FEED
(moisture and
ash free basis)
Volatile matter
Fischer assay:
Gas
Tar + light oil
Gas
Water of reaction
Oil
Char volatile content
PRODUCTS
(moisture and
ash free basis)
-------
Qualitatively, this column indicates the gas and oil yields
which could be obtained by long residence time pyrolysis.
The second column shows the per cent volatile matter in the
moisture and ash free feed. This represents the potential
yield (mostly gas) at 950°C for an "infinite" residence time.
The third column shows the pyrolytic oil, water and gas
yields from Figure 1-1 as well as the volatile matter
remaining in the pyrolytic char. All four quantities are
expressed as a per cent of the feed on a moisture and ash
free basis. This third column represents the total volatile
matter recovered from flash pyrolysis.
Comparison of the heights of the second and third columns of
each group demonstrates the advantage of flash pyrolysis
over long residence time operations. In every case (except
run 372-10 for which no gas analysis was made) the total
volatile content of the products exceeds the volatile
matter of the feed. Furthermore, in many cases the weight
loss from the feed (gas, water and oil yields, third
column) is greater than the volatile matter in the feed.
This means that during flash pyrolysis operations there
occurs a substantial amount of gasification of the fixed
carbon in the feed. This result is of great economic
significance, since it demonstrates that the advantage of
flash pyrolysis over long residence time processing, pre-
viously documented for coal conversion, can be exploited
for the conversion of industrial wastes as well.
It is also useful to compare the first column, Fischer
assay tar, light oil and gas, with the pyrolytic oil and
gas yield portions of the third column. In almost every
case the yield of usable synthetic fuels from flash
pyrolysis is greater than the Fischer assay tar, light oil
and gas of the feed material.
The quality of the pyrolytic char from rice hulls is indi-
cated in Figure 1-3. The presumption is that the more
completely the char has been devolatilized, the higher its
potential value for upgrading to an activated charcoal
substitute. Per cent volatile matter remaining in the char
is plotted versus temperature, and per cent Fischer assay
tar is also shown for comparison and for qualitative con-
firmation of trends observed in the volatile matter content.
The expected temperature effect is observed, that is, at
higher pyrolysis temperature the product char has been mare
completely devolatilized. The volatile matter contents cor-
relate strongly with the char yields shown in Figure 1-1,
suggesting that temperature, particle size and feed moisture
content all have qualitatively the same effects on extent
of devolatilization as they do on char yield. This is
reasonable, of course.
24
-------
FIGURE 1-3. Char devolatilization vs. temperature: Rice hulls
70
800
REACTOR EXIT TEMPERATURE, °F
850 900 950
60
H
PQ
ffi
40
CO
H
O
30
$ 2°
tiC
10
T
T
VOLATILE MATTER
O -20 MESH
• - 200 MESH
o
o
410
FISCHER ASSAY TAR
I
I
450 500
REACTOR EXIT TEMPERATURE, °C
1000
540
25
-------
Since rice hulls have a relatively high ash content, the
product char is commensurately high in ash. The properties
of this ash could have an important effect upon the design
of a full scale pyrolysis plant. Fusion tests on the ash
from the pyrolytic char showed its softening temperature to
be about 1500°C (2700°F) in an oxidizing or reducing atmos-
phere. Such a high temperature will not be approached in
any commercial plant, but the abrasion characteristics of
the ash may have to be considered. The ash in rice hulls
consists of approximately 35% silicon and 11% sodium and
potassium as determined by emission spectrographic analysis.
Animal Waste
Yields of dry pyrolytic char, oil, gas and pyrolytic water
are shown in Figure 1-1. Again the char yield clearly de-
creases with increased pyrolysis temperature while the oil
yield, at about 30%, appears to be independent of temperature
in the range investigated.
The effect of feed moisture content on char yield appears to
be quite strong. The two highest char yields, from the runs
at 477 and 504°C, correspond to the highest feed moisture
contents, which, however, were both less than 3%.
The distribution of volatiles in the feed and products from
these runs is shown in Figure 1-4. Again the total volatile
content in the products (third column) exceeds the volatile
matter of the feed (second column). However, in contrast
to rice hull pyrolysis, the amount of usable devolatilization
products, i.e. the pyrolytic oil and gas in the third column,
is equal to or only slightly greater than the Fischer assay
tar, light oil and gas in the first column. Also to be
noted are the high ash and nitrogen contents of the pyrolytic
oil from animal waste (see Table 1-2), compared to the oil
from the other feed materials.
Devolatilization of the pyrolytic char from animal waste
is shown in Figure 1-5. Here again the volatile matter
remaining in the pyrolytic char correlates well with the
char yield shown in Figure 1-1. The one exception is the low
volatile matter plotted at 477°C; this is believed to be an
error in analysis, particularly in view of the relatively
high Fisher assay tar in the same char.
The pyrolytic char from animal feedlot wastes was quite
high in ash content. This ash had a softening temperature
of about 1200°C (2200°F), which would not be expected to
cause problems in the full scale process. However, the
animal waste ash contained about 28% sodium and potassium,
and such a high concentration of water soluble constituents
destroys the char's potential as an activated carbon
substitute for water treatment applications.
26
-------
FIGURE I-*4. Distribution of volatiles in feed and products: Animal waste
100
80
I 60
OS
w
40
w
B
20
0
RUN NO
1
1
1
1
372-22
372-24
372-28
372-40
372-46
FEED
(moisture and
ash free basis)
Volatile matter
Fischer assay:
Tar + light oil*
-»••
7/,
^
y//.
* S
O§
Gas
Water of reaction
Oil
Char volatile content
PRODUCTS
h (moisture and
ash free basis)
-------
FIGURj t-5u Char devolatilization vs. temperature: Animal waste
CQ
W
8
fc
to
o
w
H
i
o
H-
dP
70
800
60
50
40
30
20
10
0
420
REACTOR EXIT TEMPERATURE, "F
650 900
950
1
1
VOLATILE MATTER
FISCHER ASSAY TAR
450 500
REACTOR EXIT TEMPERATURE,°C
520
-------
Fir Bark
Yields of dry pyrolytic char, oil, gas and pyrolytic water
from Douglas fir bark are shown in Figure 1-6. The char yield
appears to decrease generally with increasing pyrolysis
temperature, as is reasonable. The oil yield improves with
increasing pyrolysis temperature. The effect of minus 24
mesh feed rather than minus 200 mesh is consistent with the
results discussed earlier for rice hulls. The char yield
from 372-42 (minus 24 mesh) at 510°C is much higher (39%)
than from run 372-26 (minus 200 mesh) at the same temperature.
Correspondingly, the oil yield is much lower (28%) from the
larger particle size feed. Moisture content of the feed has
the same effect on char yield as noted previously. The two
runs showing relatively high char yields in Figure 1-6, at 471
and 538°C, correspond to the highest feed moisture contents,
although both were less than 5% moisture. In the bark
pyrolysis runs, however, in contrast to rice hulls and
animal waste, the oil yield also was affected by feed
moisture, in just the opposite way from the char yield;
that is, moisture in the fir bark feed had a detrimental
effect on the pyrolytic oil yield.
The moisture and ash free volatile components in bark and
its pyrolysis products are shown in Figure 1-7. The amount
of usable devolatilization products, indicated by the
pyrolytic gas and oil parts of the third column in each
set, is substantially greater than the Fischer assay tar,
light oil and gas shown in the first column of each set,
for the higher temperature runs (372-26, 372-42, 372-44).
For the other two runs the total yield of devolatilization
products from flash pyrolysis is about the same as the Fischer
assay tar, light oil and gas. In all cases the total
volatile content of the products is greater than the volatile
matter in the feed.
The extent of char devolatilization is shown in Figure 1-8.
The volatile matter content remaining in the pyrolytic char
correlates closely with the char yields shown in Figure 1-6,
even including the effect of the larger minus 24 mesh
particle size run at 510°C. The Fischer assay tar contents
are consistent with the volatile matter analyses. The ash
content of the pyrolytic char from fir bark was relatively
low and the ash had a softening temperature of about 1200°C
(2200°F). Emission spectrographic analysis of the tree
bark ash showed only about 9% sodium and potassium.
Grass Straw
The yields of dry pyrolytic char, oil, gas and pyrolytic
water on a dry feed basis are shown in Figure 1-6. In the
grass straw pyrolysis runs, only minus 200 mesh feed was
29
-------
CO
o
a
w
M
Cu
a
60
50
40
30
D
U
M
X 20
10
FIGURE 1—6. Product yields vs. temperature
REACTOR EXIT TEMPERATURE, °F REACTOR EXIT TEMPERATURE, °F
800 850 900 950 1000 800 850 900 950 1000
I
WATER
1
60
50
40
30
10
400 450 500
REACTOR EXIT TEMPERATURE, °C
I
T
OIL
GRASS STRAW
CHAR
WATER
550 400 450 500
REACTOR EXIT TEMPERATURE, °C
550
-------
FIGURE 1-7. Distribution of volatiles in feed and products; Fir bark
CO
100
80
2 60
W
U
as 40
a
H
W
5-
20
0
RUN NO
372-26
372-30
372-32
372-42
372-44
FEED
(moisture and
ash free basis)
Fischer assay:
Tar + light oil »-
\
^
ss^
^
##
Gas
Water of reaction
Char volatile content
PRODUCTS
(moisture and
ash free basis)
-------
FIGURE 1-8. Char devolatilization vs. temperature: Fir bark
800
en
60
OQ
50
a
w
Q 40
w
w 30
H
o
gj 20
u
H
df>
0
420
— X
REACTOR EXIT TEMPERATURE, °F
850 900 950
1
T
T
VOLATILE MATTER
*•
X
-24 MESH
- 200 MESH
1000
"T"
450
FISCHER ASSAY TAR _
500
550
REACTOR EXIT TEMPERATURE, °C
32
-------
used, and its moisture content was always less than 1.5%;
so the only variable was pyrolysis temperature.
The yield of dry pyrolytic char is very low compared to the
yield from the other three feed materials. It decreases
with increasing pyrolysis temperature. The oil yield is
consistently 50% or more, and does not appear to be much
affected by the pyrolysis temperature over the 60°C range
investigated.
Pyrolysis of the moisture and ash free volatile portion of
the grass straw is illustrated in Figure 1-9. Comparison of
the pyrolytic gas and oil in the third column of each set
with the Fischer assay tar, light oil and gas in the first
column illustrates dramatically that grass straw is a high
grade feedstock for flash pyrolysis. It is also notable
that pyrolytic oil accounts for most of the volatile product.
As with the other feedstocks, the total volatile content of
the products exceeds the volatile matter in the grass straw
feed. Even further, the feed weight loss (gas, oil and
water in the third column) is greater than the volatile
content of the feed, in all three runs.
The quality of the devolatilized pyrolytic char is shown in
Figure 1-10. The effect of temperature on volatile matter
remaining in the char is very similar to its effect on char
yield. (The inconsistently high Fischer assay tar of almost
6% at a pyrolysis temperature of 510°C may be incorrect.)
f*
Comparison of Pyrolytic Oil Yields from All Feedstocks
The yields of pyrolytic oil versus temperature for all four
feed materials studied are summarized in Figure. 1-11. The
left-hand plot shows the same oil yields given in Figures 1-1
and 1-6. In general, grass straw yielded 50% oil, rice hulls
and fir bark gave 40%, and animal waste 30%. (The differ-
ences among the four materials are largely due to their
different ash contents; almost all of the ash goes to
pyrolytic char.) It is apparent from this plot especially,
that fir bark is the only feedstock of the four for
which the yield of pyrolytic oil can be greatly affected
by moderate variation in pyrolysis temperature, feed
particle size, feed moisture content or (by hypothesis)
residence time.
For purposes of economic evaluation, the important oil yield
to consider will be the total heating value obtained in the
form of oil, from a given weight of feed. This comparison
is shown in the right-hand plot of Figure 1-11, which was
obtained from the oil yields in the left-handplot and the
heating values tabulated in Tables 1-1 through 1-4. Generally,
33
-------
FIGURE 1-9.
Distribution of volatiles in feed and products: Grass straw
100
80
§ 60
w
O
H
W
40
20
0
RUN NO.
372-36
372-38
FEED
(moisture and
ash free basis)
I
E£
Volatile matter
Fischer assay:
Tar + light
%
XV
V/
f\S\C
372-50
Gas
Water of reaction
Oil
Char volatile content.
PRODUCTS
(moisture and
ash free basis)
-------
FIGURE I-10. char devolatilization vs. temperature: Grass straw
REACTOR EXIT TEMPERATURE, °F
850 900 950 1000
w
K
to
W
H
i
u
50
40
30
20
10
T
0
450
VOLATILE MATTER
FISCHER ASSAY TAR
500
REACTOR EXIT TEMPERATURE, °C
550
35
-------
FIGURE 1-11. Pyrolytic'" oil yields from various feedstocks vs. temperature
REACTOR EXIT TEMPERATURE, °F
800 850 900 950 1000
REACTOR EXIT TEMPERATURE, °F
800 850 900 950 1000
ow
0 50
W
w
t;
&
Q
ti 40
O
*>
•t
O
w
H
* 30
J
H
O
20
1 1 I 1 1
CONVERSION |
+ *
— + _
X
•4 x
^b -i
• •
•*
*
. A A
— A x* —
X
• RICE HULLS
A ANIMAL WASTE
X FIR BARK
+ GRASS STRAW
1 1
3000
2800
Q
W
h 2600
JH
p
C5 2400
rf
u
Q- 2200
w
H
* 2000
H
0
1800
1600
_ 1 1 1 * j.
HEAT CONTENT 1 X
•f ~~
_ A
—
— 9
•
• *•
^^
_ X A
A _
9
— X
•
_ x
1 1
1 1
10 g
CM
£
n
9 r
g
D
ffl
8 g
^
Q
J
7 H
^i
H
O
6
400 450 500 550
REACTOR EXIT TEMPERATURE, °C
400 450 500 550
REACTOR EXIT TEMPERATURE, °C
-------
this plot confirms the impression to be gained from the
left-hand plot. Grass straw gives the greatest heating
value in the form of pyrolytic oil. It is confirmed that
variations in temperature, particle size and moisture
content can drastically affect the oil yields from fir bark,
perhaps even so as to exceed yields from grass straw in
terms of heating value recovered.
The quality of the pyrolytic oil is not defined by its heat-
ing value alone. For example, oil from any of the four
feedstocks has a low sulfur content, an advantage in consider-
ing the oil as a boiler fuel. However, the nitrogen contents
of the oil from animal waste were in the range 5 to 7%. This
means the animal waste would require pretreatment before
pyrolysis in order to produce a marketable pyrolytic fuel.
Comparison of Pyrolytic Char Yields from All Feedstocks
In consideration of the objective of obtaining activated char
from waste materials, it is important to note that over 80%
of the volatile material was removed from the feed during
most of the bench scale tests. This can be seen by reviewing
Figure 1-2, 1-4, 1-7 and 1-9. In these figures the double
cross-hatched segment of each third column represents volatile
matter remaining in the char, while the second column
represents the volatile matter in the feed. In most cases
the char volatile matter is less than 20% of the original
(feed) volatile matter. Exceptions occurred in runs where
the pyrolysis temperature was low, or the feed moisture
content was high, or a larger feed particle size was used.
In a commercial plant using the proprietary Garrett flash
pyrolysis process, an even greater degree of devolatilization
can be expected, since the char undergoes partial com-
bustion (to supply process heat) before its removal as a
product.
37
-------
SECTION V
PHASE II - PILOT PLANT STUDIES
INTRODUCTION
This section presents the results of the Phase II stud-
ies (pilot plant runs) on feed stocks of Douglas fir bark,
rice hulls and rye grass straw. The primary objective was
to obtain process engineering data to design commercial scale
demonstration plants for the flash pyrolysis of these indus-
trial wastes. Yields were obtained by overall, total run,
material balances and confirmed by spot stream samples and
flow rates when the system had reached steady state. During
the pilot plant studies, sufficient quantities of the pyrol-
ysis products (char, oil, and gas) were obtained for charac-
terization and evaluation during Phase III (Product Evalua-
tion) of this program.
The Phase II studies included (1) preparation of feed-
stock for flash pyrolysis; (2) production of start-up char
by direct electrical heating from the reactor walls; (3)
production of oil and char for product evaluation; (4) con-
firmation of operating conditions for optimum yields; and
(5) determination of product char properties. The most im-
portant parameter, pyrolysis temperature, was established
during the Phase I laboratory studies. Based on these data,
the expected optimum temperature was selected for all oil
production runs. Oil yields obtained at these conditions
confirmed that the laboratory results could be duplicated
at the pilot plant scale.
In addition to a discussion of the procedures used in
operation of the pilot plant, this section presents detailed
summaries of the material balance and yield results for all
runs and lists the properties of the char, oil and gas pro-
ducts obtained for all runs. Evaluations of the products
obtained are discussed in the following section of this
report, Phase III - Product Evaluation.
3ft
-------
PILOT PLANT PROCEDURES
Feed Preparation
About 450-900 kilograms per hour (1000-2000 pounds per
hour) of feedstock is transferred pneumatically from storage
bins into a stirred fluid bed drier. A propane fired jet
delivers 120°C (250°F) gas to the bed at 125 cycles per second
resulting in sonic energy to increase drying efficiency. The
discharge from the drier is pneumatically transferred at 450-
900 kg/hr (1000-2000 Ib/hr) either to storage or to a vertical
hammer mill for secondary shredding. Milled and dried feed-
stock can be returned to storage or transferred pneumatically
to the pyrolysis feed bin.
Pyrolysis System
Figure II-l presents a simplified flow diagram of the
pyrolysis pilot plant. A plot plan of the equipment is shown
in Figure II-2.
The GR&D four TPD pilot plant was built in 1972 as a test
facility to evaluate the flash pyrolysis process (which had
been successfully proven on the laboratory scale), and to
obtain the heat and material balance data needed for the
design of larger units. Operations after shakedown testing
in 1971 were devoted primarily to processing of solid waste
until early in 1974. A summary of the pilot plant operations
on solid wastes through 1974 is presented below:
Feed Stock Feed Weight Operating Time
Ibs Hours
Municipal Solid Waste 32,603 1,160.2
Bark 28,193 453.8
Rice Hulls 24,228 316.0
Straw 3,761 69.6
Char Loop -
Char is dropped from a heat traced bin into a screw
feeder that controls the rate at 90-900 kg/hr (200-2000
Ib/hr) into a pneumatic transport line. The transport gas
is developed by a 135,000 kcal/hr (530,000 Btu/hr) gas
fired generator. The generator produces gas at 985°C (1800°F),
which heats the char to an ignition temperature of approxi-
mately 540°C (1000°F) and transports it to the heater where
39
-------
FIGURE H- 1
GARRETT RESEARCH a DEVELOPMENT 4 -TON /DAY PILOT PLANT FLOWSHEET
CHAR LOOP
' VENT {
rr '
I ' P5AQTOR UOQP
t v i
f »(W>Nt BAG 4
I H°USE ' * A^nur
Q STORAGE DRUM j T ~V
1 , „ .rtCYCLONE 1
1 i " " v r^
jnr.vn nNFj Y
IV ! ^£v%-.. xLROTARY\/
i ^-•ovy ^LOCK Y
AROTARYI \>.',;itEEO * i
THAR YLOCK | Y BIN PRODUCT
CHAR | REACTOR ± CHAR
HEATER 1 \~J \ /CHAR
1, \ / Y BIN
1^ A|p I \ /FEEDER ~4~ ' — ' ' '
L_J | i— E*/ \ /CHAR .
if \ /FEEDER
N^J rtrtC »' ^T ^__^_ — , 1
2 PREHEATER 1 ^
INERT
, GAS
! GENERATOR
J.,
IcRUBBER OEMISTERS
lUENCHll n^^^ \n
TANK X J\ ,, NO
U / \ f~^n QUENCH
I 1 -TT— COOI FR
DtCANTtR "~ *ff p *W 1 ii w
-D- n I
'1 V,,.,., .* _n_ *n..,,,_^ ,. ._•
* FILTLRS
A
VENT
, , f
•*" i i I ^ ^ IpRnntirT
L ITAN*
AFTERBURNER
Tl
O
30
m
B
-------
FIGURE n-2
GRSD 4 TPD PILOT PLANT PLOT PLAN
15
30
APPROX. SCALE: | =15
O
AFTERBURNER
N
LIQUID
COLLECTION
TRAIN
CONTROL
ROOM
STORAGE BINS
' BLOWER
SECONDARY
SHREDDER
O°
PYROLYSIS
FEED BINS
CHAR LOOP
BAGHOUSE
BED PULSE-JET
DRIER
BINS
-------
partial combustion raises the temperature to 705-930°C
(1300-1700°F). The hot char is separated from the transport
and flue gases by a primary cyclone and the cyclone discharge
is fed by rotary valve to the reactor loop or returned to the
char bin depending on the desired mode of operation. Fines
from attrition and combustion are removed from the gases by
a secondary cyclone. If H2S is in the flue gas, it can be
removed by reaction with iron oxide in a contactor vessel.
The gases then pass through a baghouse containing eight Nomex
bags before venting to the atmosphere.
Reactor Loop -
The feedstock from the pyrolysis feed bin is delivered
by a rotary valve into a live bin hopper where a screw feeder
controls the feed rate at 13.5-135 kg/hr (30-300 Ib/hr) into
a pneumatic transport line. Nitrogen or recycle gas trans-
ports the feed into the pyrolysis reactor. The pyrolysis
reactor is equipped with electrical wall heaters, and start-up
char can be produced by direct heating. However, in normal
operations, pyrolysis is conducted by circulating hot char
from the char loop, and in this mode of operation, the hot
particulate char is transported into the pyrolysis reactor
for intimate mixing with the feed, by nitrogen or recycle gas
that has been preheated by a separate tubular heater. The
circulating char is removed from the pyrolysis gas stream
by a primary cyclone and returned to the char bin by a rotary
valve. A second cyclone removes fine product char from the
gas stream and discharges it out of the system to a water
jacketed char receiver. The condensibles-laden gas stream
from the second cyclone is directed to the oil collection
system, or to a direct gas fired afterburner if only reactor
parameters are under study.
Oil Collection System -
The condensibles-laden gas stream is quenched by con-
tacting with a recirculating immiscible oil fluid. Several
types can be used and were studied at length. The one used
most often was a C,2""ci4 branch-chained aliphatic compound
trade named AMSCO 550 By the Union Oil Co. The quenched gas
passes through a venturi scrubber and a series of demister
pads for removal of aerosols formed during the rapid pyrolysis
reaction. The scrubbed gas can be directed through a fiber-
glass packed pad for final mist elimination or through a heat
exchanger to condense water of reaction prior to the mist
eliminator. The gas can then be exhausted to the atmosphere
through the afterburner or can be recycled.
42
-------
The quenching fluid is pumped by an inline centrifugal
pump through basket filters and a heat exchanger from a
decanter tank to a surge tank. The liquid is pumped from the
surge tank by a gear pump to the venturi scrubber that drains
back into the decanter tank. The liquid from the surge tank
is also pumped to the quench tank by an inline centrifugal
pump to control quench temperature and also drains back into
the decanter tank.
Product Oil Treatment
If product oil treatment is needed, a variable rate
positive displacement pump transfers product oil from the
bottom of the decanter tank through basket filters to a cen-
trifuge. The centrifuge separates the product oil from the
quench fluid. The quench fluid is returned to the decanter
tank and the product oil can be processed through a thin film
evaporator if adjustment of the final moisture content is
required.
Analyses
The feed and product streams were subjected to the ana-
lytical procedures summarized below.
Char and Feed -
These materials were subjected to ash and elemental
analyses, bulk density, particle density (by Fluid displace-
ment) , ash composition (by atomic absorption), feed moisture
(oven dry for four hours), and particle size distribution
(by sieve, and micromerograph).
Condensibles -
These were analyzed for water (azeotropic distillation
and Karl Fischer), quench fluid (by centrifuging and separ-
ating phases), char (filter residue from an acetone plus 5%
caustic wash), ash and elemental analyses, and for COD, (if
the water content was greater than 50%).
Oil Properties -
Viscosity (Saybolt Univ.), density (pycnometer), flash
point (Pensky-Martens) and pourpoint were determined.
Gas -
Combined gas chromatography and mass spectrometry.
43
-------
MATERIAL BALANCE
Scope and General Methods
A major objective of the pilot plant program was to measure
directly the oil yields obtained during operation of the
pyrolysis process at the pilot plant scale while in the char
heating mode. Process conditions were held at steady state
for all runs with the exception of a few short tests. The
most important parameter, pyrolysis temperature, was studied
in depth in the laboratory phase of this program. Based on
those data, an expected optimum temperature was selected for
all oil production runs, and the oil yield at these conditions
was compared with the laboratory pyrolysis results. The
gathering of other yield data was subordinated to the above
objective to a greater or lesser degree depending upon the
relative importance of the other data for corroboration of
oil yield results. Accordingly, emphasis was placed on ac-
curate measurements and analysis of the feed and liquids
recovered from the oil collection system, including water
byproducts.
Material balances were calculated for the pyrolysis
step for each feedstock. Oil and water yields were measured
directly and the gas was determined indirectly by sampling
and compositional analysis. Gas yields and rates were com-
puted from the known nitrogen carrier gas rate and the meas-
ured gas composition. Combining those results gave the char
yield by difference since some of this product is consumed
in the process and cannot be measured directly when operating
in the char heating mode. However, the yield obtained by
difference can be confirmed by comparison with the pyrolysis
yields from operating in the direct heating mode which gener-
ally gives excellent mass balance closures.
Material balance data were obtained only for the
pyrolysis step of the process, as calculation of an over-
all process material balance closure is beyond the scope
of this program. Complete material balance analysis of
the char heating operation was impractical for several
reasons. First, the product char composition cannot be
accurately determined in a mixture with startup petroleum
coke. Also, the amount of char combusted cannot be
accurately calculated, since inert gas generated from
natural gas combustion was used for char heater solids
transport, which masks the products of char combustion.
Finally, operation was limited to 16 hours per day owing
to manpower limitations. The daily heatup period, an
unsteady state, also adds an additional complication.
44
-------
An elemental balance for the pyrolysis step was made
based on the mass balance calculated for the char heating
runs described above. Pyrolytic char composition was ob-
tained from the direct heating operation data. All other
compositions were directly measured. This procedure per-
mits calculation of closure for carbon, hydrogen and oxygen,
based on feed, giving an independent check on the accuracy
of the mass balance and, particularly, the oil yield.
Oil yield was also measured by isokinetic sampling.
This technique involves taking a slipstream sample of the
hot pyrolysis off-gases and condensing the oil and water in
cold traps. By this procedure a measurement of oil and water
yields was obtained which is completely independent of the
macroscopic result.
Macroscopic Measurements
Feedstock -
Several tons of feedstock was prepared in advance,
kept in storage bins, and transferred pneumatically to
the feed bin in loads of about 900 kg (2000 Ib). Typically,
two or three loads were charged in a run. The material
was sampled in the storage bin compartment just prior to
transfer. Several increments were taken using a slotted
tubular "thief" to assure a representative sample. Each
sample was analyzed for moisture (the feedstocks were
hydroscopic at low moisture contents), and the weight of
moisture and dry feed was calculated for each load. Then,
by careful physical blending, a composite was formed for
each run, and moisture, ash, and elemental analyses were
made on the composite. Feed weight was recorded continuously.
Condensibles -
Typically, 70-80% of the oil yield was recovered in the
primary collection vessel, the decanter. (Refer to the flow
diagram, Figure II-l). The product oil is considerably
denser than the quench oil (about 1.2 vs 0.8 g/cc) and settled
readily to the bottom. Product oil was drawn off daily at
the end of the operations and stored in drums. This batch
removal procedure assured correspondence between feed weight
and oil recovery for material balance calculations. Since
some soluble, as well as entrained quench oil was also
removed with the oil products, each drum was sampled at sev-
eral levels to provide a representative averaged sample of
the drum. Each sample was analyzed for moisture, quench
oil, and char contents. A mathematical composite was then
formed for the complete run in proportion to the drum weights.
These composite analyses were then used to calculate the
material and elemental balances described above.
45
-------
By careful control of the quench oil flow and tempera-
ture, the gas leaving the primary scrubber contained most of
the byproduct water, along with some entrained product oil
and quench oil. Most of the entrained oils were then removed
by the total condenser and drained to a drum which was
weighed, sampled, and analyzed daily by the same procedure
used for the product oil from the decanter. The gas leaving
this condenser then passes through a mist eliminator of
packed fiberglass. The net weight gain of the fiberglass
element was recorded. Typically 1-2% of the oil produced
was collected at this point. Oil from the fiberglass elements
was analyzed, and its composition was accounted for in report-
ing product oil composition.
Condensibles in the gas vented to the afterburner were
measured by isokinetic slipstream sampling. Product oil loss
was typically about 5% of the total oil. The measured quench
oil and water losses agreed with their known vapor pressures.
The inventory of quench oil was carefully weighed in and out
to account for any solubility of product oil.
Pyrolysis of two feedstocks (tree bark and grass straw)
produced from 5 to 15% wax (based upon feed) which was sol-
uble in the paraffin quench oil in all proportions at the
operating temperature. This wax yield could not be measured
accurately by net change in quench inventory owing to the
relatively large amount of quench oil compared to the quan-
tity of feed processed during any typical run. Since the
production of this wax fraction was not foreseen, suitable
dewaxing equipment was not provided and a direct measurement
could not be made. However, the wax yield was satisfactorily
determined indirectly by measuring the wax concentration
changes in the quench oil by vacuum distillation.
Char -
The term pyrolytic char refers to the thermal decompo-
sition product, as distinguished from product char which has
been partially burned in the char heating operation. Pyroly-
tic char cannot be measured directly during operation in the
char heating mode. It can, however, be measured from the
direct heating runs where the char is neither burned nor
mixed with substitute char* Two methods were used. The pre-
ferred method was to simply remove all product char from the
receiver bin, weigh, and sample it. This was done in only
a few cases. Normally, the char bin weight transmitter was
used. An absolute weight could not be obtained, however,
due to changes in stress because of thermal expansion effects
of the connecting pipes. Thus only an approximate yield
result could be obtained by this measurement. For long runs
46
-------
(when steady state temperatures were reached) the yield of
pyrolytic char was estimated by comparison of rate of weight
gain of the char bin to feed rate.
Microscopic Measurements
Isokinetic Oil Yield Test -
Isokinetic sampling was developed as a means of quickly
measuring oil yield. It is a useful technique for scanning
the effects of operating variables and to confirm the macro-
scopic results. In essence, the method involves a represen-
tative sampling of the gas stream containing solids and pos-
sibly condensed products. This is done by matching the
gas velocity in the sample tube to that of the process line
to assure representative capture of the high momentum (larger
than 1 micron) particles. Sample gas flow rate is therefore
proportional to process flow via the area ratio of process
pipe and sample tube areas. A collection train of four 4-liter
kettles in series was used to condense and collect the vapors:
one at ambient temperature, one at ice water, and two in a
dry ice-trichloroethylene bath. The gradual temperature
reduction is necessary to minimize plugging problems. The
kettles were followed by a submicron filter to remove final
traces of oil aerosol and then a positive displacement meter
calibrated against a wet test meter. While process pressure
was adequate for flow, a needle valve followed by a vacuum
pump was found to be more effective configuration, since a
high pressure drop results in better flow control. Very
accurate measurements of feed weight and composition are
required for this test. Feed material was weighed to within
0.05 kg (0.1 Ib) and loaded into the feeder hopper. Gener-
ally the feed weight exceeded 45 kg (100 Ibs). A limit to
accuracy in this method is the process gas flow which was
measured to within 2%, relative, which is equivalent to 1%
absolute accuracy of the reported yield.
Gas Yield -
The gas yield was calculated from gas composition, total
gas flow, and feed rate measurements. Integrated gas samples
of 15-30 minutes duration were collected. This was done
simultaneously with the isokinetic oil yield tests for which
the feed rate and composition during the sampling period were
known accurately. The integrated sampling method averages
gas composition variations due to normally occurring small
fluctuations of feed rate and gas flow.
47
-------
EXPERIMENTAL RESULTS AND DISCUSSION
The 4-ton-per-day pilot plant was used to obtain pro-
cess engineering data to design commercial scale demonstration
plants for the flash pyrolysis of industrial solid wastes.
The feedstocks processed were Douglas Fir tree bark, rice
hulls and rye grass straw. The primary objectives were to
determine yields both by macroscopic balance and by isokinetic
stream sampling, and the production of sufficient quantities
of pyrolysis products (char, oil and gas) for characterization
and evaluation.
The program outline was (1) preparation of feed stock
for flash pyrolysis (particle size, etc.); (2) production of
start-up char by electrical heating; (3) determination of
product char properties; (4) confirmation of operating condi-
tions for optimum yields; and (5) production of oil for study
and evaluation.
Feed Handling Procedures
Table II-l presents a summary of the feed analyses used
for pilot plant runs. Douglas fir tree bark required drying
and milling before it was fed to the plant. Tree bark char
was produced for start-up and the oil collection system was
operated periodically during char production to evaluate the
collection system and train personnel. Circulation of hot
tree bark char for reactor heating was not found to be feasi-
ble in the existing system which originally had been designed
for coal conversion. Consequently, petroleum coke was used
as the circulating solid heat carrier in subsequent pilot
plant runs to determine operating conditions and oil yields.
Rice hulls were processed as received without any feed
preparation. The char produced appeared to behave satisfac-
torily in the pilot plant during char circulation studies, but
its recovery proved to be somewhat insufficient with the
existing cyclone design. Thus, operations for the production
of rice hull oil also utilized petroleum coke as the circula-
ting heat transfer media.
The only feed preparation for rye grass straw was milling.
A quantity of oil was produced utilizing petroleum coke. The
resulting composite char confirmed expectations that the
existing equipment designed for coal could not have circulated
the straw char effectively.
48
-------
Table II-l. FEED ANAItfSIS SUMflMCf
<£>
Run Feed-
Number stock 28
49-73
50-73
52-73
57-73
Mesh Analysis
Cumulative % Retained
60
100 150 200 325 -325
37-73 Bark
38-73 Bark
Bark
Bark
8.0 43.6 63.4 73.4 80.6 90.6 9.4
5.8 35.0 54.5 66.7 76.0 87.7 12.3
6.5 30.2 46.4 57.8 67.1 80.8 19.2
4.6 20.4 33.4 43.8 53.3 70.8 29.2
4.2 27.8 45.5 55.9 63.1 72.1 27.9
15.7 42.2 55.9 68.2 76.6 83.5 16.5
9.8 29.2 49.0 64.6 73.5 83.5 16.5
26.5 53.4 65.9 73.5 78.9 86.2 13.8
20.4 45.2 59.5 68.1 74.1 82.8 17.2
18.1 41.6 56.2 66.4 73.2 82.3 17.7
Bark 26.7 52.2 67.8 74.9 80.0 86.& 13.1
81.4 92.2 95.0 96.5 97.5 2.5
Rice
Hulls 41.7
63-73
64-73
66-74 Straw 5.3 35.5 61.4 75.9 84.2 89.8 10.2
Ultimate Analysis
H.,0 V.M. Ash
H
N
Cl
O
5.5 - 11.1 - - - - -
5.1 70.55 4.16 55.90 6.13 0.30 0.09 - 37.6
5.4 - - ____ _-
6.4 69.52 9.75 54.07 6.22 0.37 0.09 - 39.25
4.2- - _--- --
4.9- - _--- --
6.32 59.32 17.05 41.53 5.56 0.35 0.16 0.05
63.32 18.20 44.33 5.19 0.37 0.17 0.05 31.69
18.91 40.00 5.40 0.41 0.05 0.13 35.10
19.39 39.75 5.30 0.33 0.11 0.09 35.03
5.9
5.9
6.6
8.56 42.84 6.57 0.50 0.33 0.35
9.11 45.53 6.28 0.53 0.35 0.37
8.50 42.26 6.22 0.53 0.35 0.35
9.62 44.91 5.92 0.56 0.38 0.37 38.61
6.01 42.46 6.21 0.53 0.36 0.32
6.44 45.48 5.86 0.57 0.39 0.34 40.92
-------
Summary of Operations
Initial operation to produce tree bark char was con-
ducted over a period of 6 weeks, pyrolyzing 5069 kg (11,175 Ibs)
of tree bark by direct heating during 8 runs totalling 156 hrs
of feeding time, and producing about 1360 kg (3000 Ibs) of
tree bark char. Over 900 kg (2000 Ibs) of tree bark pyroly-
sis oil was collected during these runs as a secondary objec-
tive to evaluate the oil collection system and train the
operating personnel. The presence of a wax fraction in the
tree bark pyrolysis oil was noted and the operating problems
created by the wax phase in the operation of the oil collec-
tion system were resolved during these char production runs.
Tree bark char combustion and cyclone efficiencies were
studied over a two week period consisting of four runs total-
ing 23 hours of char circulation. The equipment designed for
coal proved to be unsatisfactory for the lighter density tree
bark char. Rather than spend considerable time and money
redesigning and modifying existing equipment, higher density
petroleum coke was used as the circulating solid. However,
the reactor feed rate then had to be decreased, owing to the
lower reactivity of this char and consequent lower temperature
upon combustion in the char heater.
The operating conditions for the production of pyrolytic
oil from tree bark were established during four runs over a
four week period feeding 5450 kg (12,000 Ibs) of tree bark
for 140 hrs. Wax associated problems were resolved during
these runs and a technique of controlling the moisture con-
tent of the product oil was established. The subsequent tree
bark pyrolysis oil production run covered a two week period
feeding 2270 kg (5000 Ibs) of tree bark for 105 hrs and
collecting 681 kg (1500 Ibs) of pyrolytic oil.
Rice hull char production was conducted over a period
of two .weeks, pyrolyzing 3360 kg (7400 Ibs) of rice hulls
using direct heating during one run of 105 hrs and producing
approximately 1090 kg (2400 Ibs) of rice hull char. Rice
hull char combustion and cyclone efficiencies were studied
over a two day period consisting of one run of 14 hrs. The
equipment designed for coal appeared to operate satisfactorily
with rice hull char.
Circulation of rice .hull char was thus attempted during
initial operation for the production of rice hull oil. How-
ever, during a one week period, 1415 kg (3115 Ibs) of rice
hulls was fed over a total 24 hours operating time, and the
high char content of the pyrolytic oil indicated that the
cyclone efficiency was again unacceptably poor. Therefore,
petroleum coke was again used as the circulating heat trans-
fer solid. The production of pyrolytic rice hull oil was
50
-------
then conducted over a period of three weeks, feeding 7270 kg
(16,000 Ibs) of rice hulls during three runs totaling 144 hrs
of feeding time, and collecting 1270 kg (2800 Ibs) of oil.
Pyrolytic oil from rye grass straw was produced during
one run over a two week period, feeding 1820 kg (4000 Ibs) of
straw for 46 hrs with 70 hrs of char circulation and collec-
ting 500 kg (1100 Ibs) of oil.
Bark Char Production Runs -
Run 37-73 - The first tree bark char production run (37-73)
was 1680 kg (3700 Ibs) of tree bark processed over 51.8 hrs
for an average rate of 32.2 kg/hr (71 Ibs/hr). The feed rate
was limited by the electrical heat flux capacity of the
reactor wall heaters which was less than 3.3 kw/m (1 kw per
foot). The time-average pyrolysis temperature in the elec-
trically heated reactor was 513°C (957°F). The char produced
was 479 kg (1055 Ibs) for a yield of 29 wt %. The oil product
containing 28% water was successfully pumped and filtered at
60-66°C (140-150°F). The wax fraction remained in solution
in the quench liquid .above 49°C (120°F).
Run 38-73 - During the second tree bark char production run
(38-73), 986 kg (2170 Ibs) was processed over 33.9 hrs for
an average rate of 29 kg/hr (64 Ibs/hr). The reactor was
heated electrically to a time-average pyrolysis temperature
of 502°C (936°F). A total of 332 kg (731 Ibs) of char was
produced for a yield of 34%.
The oil collection system was then insulated and quench
fluid rates were reduced during this run to achieve a higher
quench system temperature and avoid the problem of wax depo-
sits. The higher collection system temperature resulted in
a low water content oil (about 8% water) that was very vis-
cous, resembling pitch. No effort was made to further treat
this oil.
Runs 39-73, 40-73, 41-73 - The third, fourth and fifth bark
char production runs (39-73, 40-73, and 41-73) used 291 kg
(620 Ibs), 422 kg (930 Ibs) and 590 kg (1300 Ibs) of tree bark,
processed over 8.6 hrs, 11.2 hrs and 17.2 hrs for average rates
of 32.9 kg/hr (72.5 Ibs/hr), 37.6 kg/hr (82.9 Ibs/hr) and
34.2 kg/hr (75.3 Ibs/hr) respectively. The reactor was elec-
trically heated to time-average pyrolysis temperatures of
493°C (921°F), 494°C (922°F) and 499°C (930°F). Total
char collected for the three runs was 267 kg (587 Ibs) for a
yield of 21%.
51
-------
The oil collection system was carefully studied to
optimize the oil moisture control methods. The bark oil
collected during these runs varied in moisture content
from 13% to 20%. Blends of the oil were made for viscosity
and density measurements as functions of water content and
temperature. Filtration of oils heated to 160-180°F revealed
that successful filtration depends not only on temperature
but also on water content; this is probably an effect of dis-
persion of the particles as well as an effect of viscosity.
Operation of a simple dewaxing unit was also tested on a
small quench fluid stream during these runs, but temperatures
below 120°F were not achieved and the wax remained in solution,
The temperature control of the entire liquid collection
system was again reviewed. The addition of a glycol heater
to raise the temperature of the circulating quench fluid
greatly improved control from this point forward.
Runs 42-73,43-73, 44-73 - The last three bark char produc-
tion runs (42-73, 43-73 and 44-73) used 218 kg (480 Ibs) of
tree bark over 6.2 hrs for an average rate of 35.4 kg/hr
(77.9 Ibs/hr), 254 kg (560 Ibs) over 6.9 hrs for an average
rate of 36.8 kg/hr (81.1 Ibs/hr), and 643 kg (1415 Ibs) over
20.2 hrs for an average rate of 31.8 kg/hr (70.0 Ibs/hr)
respectively. The time-average pyrolysis temperatures were
505°C (941°F), 491°C (916°F), and 495°C (924°F) respectively.
The product char from these runs was left in the char
bin for subsequent combustion -tests in the char heater.
However, char weights and therefore char yields were not
determined for these runs.
Test operations of the oil collection system indicated
that pyrolytic oil could be recovered without using the
centrifuge to remove quench liquid and without employing
the thin film evaporator for further water removal. Only
a cursory investigation of wax removal from the quench fluid
was made during these runs. The affect of .'increasing concen-
trations of soluble oil and wax in the quench fluid was left
for study during the later oil production runs.
Rice Hull Char Production.Run 57-73 -
Rice hull char was produced during one run by feeding
3440 kg (7360 Ibs) of rice hulls over 104.6 hours for an
average rate of 32.0 kg/hr (70.4 Ibs/hr) and producing
approximately 1090 kg (2400 Ibs) of char. The reactor was
heated electrically for an average pyrolysis temperature of
504°C (940°F). The char produced remained in the bin for
subsequent combustion tests; the change in the bin weight
recorder indicated a yield of approximately 40%. While the
52
-------
oil collection system was not operated, integrated isokinetic
gas samples indicated gas yields of 11.0 and 9.3 wt % were
achieved on a dry feed basis.
Bark Oil Production Runs -
Run 49-73 - The first Douglas fir tree bark production run
(49-73) used 1660 kg (3660 Ibs) of tree bark for 34.7 hrs
for an average rate of 57.5 kg/hr (126.4 Ibs/hr). The reactor
was heated by circulating petroleum coke at 527°C (980°F),
heated to an average temperature of 698°C (1288°F) by partial
char combustion. The low reactivity of the petroleum coke
limited the char circulation rate to 182 kg/hr (400 Ibs/hr)
since the combustion of petroleum coke was not rapid enough
to maintain the desired temperature at higher circulation
rates. The combustion of petroleum coke and product char was
also limited by the excess air that could be obtained from
the inert gas generator system. Increases in excess air
resulted in decreases in the carrier gas temperature and the
increase in char combustion did not compensate for the loss
in sensible heat. The average pyrolysis temperature for this
run was 515°C (959°F).
The collection system was operated during 31 hrs at 44°C
(112°F), and the collected oil was left in the system to
combine with the oil collected during the next run. Thus no
separate oil yield was determined for this run. Successful
removal of wax was achieved using -10°C (20°F) glycol as
a coolant through a coil in a 208 liter (55 gallon) drum.
The operation was a simple batch procedure performed as time
permitted and providing data sufficient for design of a
continuous operation.
Run 50-73 - The second tree bark oil production run (50-73)
used 1060 kg (2330 Ibs) of tree bark for 13.9 hrs for an
average rate of 76 kg/hr (167 Ibs/hr). The reactor was 491°C
(916°F), heated by circulating petroleum coke at an average
temperature of 681°C (1258°F).
The collection system was operated at 48°C (118°F) and
the overall yield for the 1st and 2nd production runs was
22% collected in the main receiver on a dry oil/dry feed
basis. (This yield does not include wax, quench oil solubles,
collections at demister pads, or vent losses). The integrated
isokinetic sample gave an estimate of the liquid yields that
would have been obtained if all sources had been collected.
On a dry feed basis the total oil yields from these runs
were:
Bark oil (dry) 39.0% (including wax)
Pyrolytic water 17.8%
53
-------
The wax concentrate from the batch dewaxing operation
was analyzed and found to be only 13% wax and 87% quench
liquid. Apparently the wax traps a significant amount of
quench liquid as it is cooling and the resulting solid
actually has a relatively low wax concentration.
Run 51-73 - The third tree bark oil production run (51-73)
used 1785 kg (3930 Ibs) of tree bark over a 30.8 hr period
for an average rate of 58 kg/hr (127.7 Ibs/hr). The reactor
was maintained at 490°C (915°F) by circulating petroleum
coke at 204 kg/hr (450 Ibs/hr) at a temperature of 747°C
(1375°F). The collection system was operated at 48°C (118°F),
and 342 kg (753 Ibs) of oil was drained from the primary
receiver for a 20.4% primary collection on a dry oil/dry
feed basis. The remaining oil yield was not collected.
Run 52-73 - The fourth bark oil production run (52-73) used
847 kg (1865 Ibs) of tree bark over 13.5 hrs for an average
rate of 62.7 kg/hr (138 Ibs/hr). The reactor was maintained
at 476°C (888°F) by circulating 182 kg/hr (400 Ibs/hr) of
petroleum coke at 808°C (1487°F). The collection train was
operated at from the primary receiver for a recovery of 24.6
wt % dry bark oil/dry feed basis. The remaining oil yield was
not collected.
Run 53-73 - The final Douglas fir tree bark pyrolytic oil
production run (53-73) was made with emphasis on a complete
macroscopic material balance. A total of 2326 kg (5129 Ibs)
of tree bark was processed over a 47 hr period for an average
rate of 49.5 kg/h4 (109 Ibs/hr). The reactor was heated to
an average pyrolysis temperature of 474°C (886°F) by circu-
lating 204 kg/hr (450 Ibs/hr) of petroleum coke heated to
828°C (1522°F). The collection system was operated at 52°C
(126°F) with improvements in the primary receiver that raised
the primary collection to 30% (compared to 21% during earlier
runs).
The final macro-balance gave 39.7 wt % oil yield vs
40 wt % by the integrated isokinetic samples of the product
stream. Integrated isokinetic samples were also taken from
the vent line to determine oil losses and gas yields. The
break-down of the overall yields on a dry feed basis was:
54
-------
Pyrolytic Oil
39.7
Wax
Water
Gas
Char (by diff.)
An attempt to filter the bark oil was not successful.
A worm-type positive displacement pump operated satisfactor-
ily but adequate filtration and filtration rates could not be
achieved. The small particle size of the char (50% less than
12 micron) probably results in blinding of the filter bags
by filling the interstices, preventing buildup of a filter
cake. Larger filter area would result in excessive oil loss.
(The pilot plant centrifuge is designed for liquids and would
not operate for any reasonable length of time while attempting
to remove solids).
Rice Hull Oil Production Runs -
The objectives of the rice hull pyrolytic oil production
runs were:
1. Produce rice hull pyrolytic oil for combustion
tests.
2. Test solubility limits of rice hull oil in the
quench fluid.
3. Obtain macroscopic yield data.
4. Obtain gas yield data.
5. Obtain isokinetic yield data.
6. Obtain vent gas aerosol loss data.
7. Establish techniques to control water content of
the product oil.
Run 59-73 - The first rice hull oil production run (59-73)
used 1413 kg (3115 Ibs) of rice hulls over 23.6 hrs for
an average rate of 52.7 kg/hr (116 Igs/hr). The reactor was
heated to 489°C (912°F) by circulating rice hull char heated
to an average of 708°C (1307°F) by partial combustion in the
char heater.
55
-------
The collection system was operated but due to the short
duration of the run, oil production was too limited for an
adequate material balance or conditioning of the quench
liquid.
Excessive char carry-over was observed and attributed to
attrition and combustion of the rice hull char resulting in
particle sizes too small for efficient collection by the
pilot plant cyclones. The high ash content of rice hull char
is not only undesirable in the pyrolytic oil product, but the
heat capacity of the circulating char steadily decreased as
the ash content of the char built up during the run.
Inspection of char loop piping revealed high ash content
and related high bulk density deposits in horizontal piping
sections of the pilot plant. Therefore, subsequent rice hull
pyrolytic oil runs were made using the less reactive, but
lower ash content petroleum coke as the circulating solid.
Run 63-73 - The second rice hull oil production run (63-73)
used 1716 kg (3784 Ibs) of rice hulls over 33.7 hrs for
an average rate of 50.9 kg/hr (112.3 Ibs/hr) . The reactor
was maintained at 481°C (897°F) by circulating petroleum
coke at 204 kg/hr (450 Ibs/hr) heated to 618°C (1144°F) by
partial combustion. The collection system was operated at
58°C (136°F) utilizing the 4.88m (16 ft) heat exchanger for
improved oil moisture control. The vent loss of oil was
measured to be 1.8 wt % on a dry feed basis.
Comparative yields on a dry feed basis were:
Macro Isokinetic
Rice Hull Oil 26.0 21.2
Pyrolytic Water 20.4 22.3
Run 64-73 - The third rice hull oil production run (64-73)
was a continuation of Run 63-73 to monitor the change in
macroscopic yields during dissolution into the quench liquid
of the soluble fraction of rice hull oil ("conditioning").
A total of 1943 kg (4284 Ibs) of rice hulls was fed over
35 hrs for an average rate of 55 kg/hr (120.9 Ibs/hr). The
reactor was maintained at 484°C (903°F) by circulating
petroleum coke at 295 kg/hr (650 Ibs/hr) at 675°C (1246°F).
The collection system was operated at 57°C (134°F) and com-
parative yields were:
56
-------
455°C(850°F) 516°C(960°F)
Macro Xsokinetic Isokinetic
Rice Hull Oil 25.3 25.5 25.0
Pyrolysis Water 21.0 20.4 21.3
Gas - 5.4 5.8
Char (by dif-
ference) - 48.7 48.9
Run 65-73 - The final rice hull oil production run (65-73)
was for continued "conditioning" of the quench fluid by solu-
ble rice hull oil components using 2479 kg (5465 Ibs) of
rice hulls over 50.8 hrs for an average rate of 48.8 kg/hr
(107.6 Ibs/hr). The reactor was maintained at 488°C (910°F)
by circulating petroleum coke at 295 kg/hr (650 Ibs/hr) at
683°C (1262°F). The collection system was operated at 57°C
(137°F) and comparative yields were:
516°C 488°C 460°C
(960°F) (910°F) (860°F)
Macro Isokinetic Isokinetic Isokinetic
Pyrolytic
Oil 22.1 23.7, 20.0^ 22.2, 21.3 22.0, 22.7
Pyrolysis
Water 22.4 20.9, 18.8 21.0, 18.4 19.4, 19.2
An expected increase in oil recovery, as the quench
liquid became "conditioned", did not occur; in fact, macro-
scopic yields decreased. These lower oil yields were thought
to result.from a larger feed particle size and a higher
moisture content, as indicated by a comparison with bench
scale results. However, subsequent isokinetic tests with
milled and dried rice hulls did not show substantially higher
oil yields.
Grass Straw Oil Production, Run 66-74 -
The objectives of the rye grass straw pyrolytic oil
production run were; (1) determine the macroscopic yield, (2)
determine properties of grass straw pyrolytic oil, (3) determine
an integrated isokinetic yield, (4) determine the gas yield.
The straw pyrolytic oil production run (66-74) used
1708 kg (3761 Ibs) of straw over 45.8 hrs for an average rate
of 35.6 kg/hr (78.4 Ibs/hr). The reactor was maintained at
483°C (9020F) by circulating petroleum coke at 318 kg/hr
57
-------
(700 Ibs/hr) at 658°C (1219°F). The collection system was
operated at 57° (135°F). Liquid losses in the vent were
determined to be 1.1 wt % oil and 3.7 wt % water, while the
gas yield was calculated at 5.5 wt % on a dry feed basis.
Comparative yields were:
Macro Isokinetic
Pyrolytic Oil
(including wax) 32.2 28.0, 26.8
Pyrolytic Water 24.9 26.6, 24.7
As inspection of the piping at the completion of this
run revealed a composite char with an angle of repose of
80-90° vs the 25-30° exhibited by the initial char. The
inability of this char to flow would account for high char
carry-over and the erratic feed rates experienced during
this run.
Other Pilot Plant Operations
Douglas fir tree bark char was conditioned for subse-
quent studies on activation by devolatilization and partial
combustion, and cyclone efficiencies were studied, during
runs 45-73, 46-73, 47-73, and 48-73. The overall cyclone
efficiency was only 83%, and this is attributed to the low
bulk density of the char and the high length-to-diameter
ratio of the char particle shape. The char captured by the
second stage cyclone was 50% minus 325 mesh. Experiments
revealed that tree bark char offers little resistance to
flow and appreciable amounts of gas leaked upward through
the char hopper into the cyclones which could have contri-
buted to low cyclone efficiency. Gas leakage upward through
the char could also result in fluidization which makes the
actual char circulation rate indeterminate and therefore
possibly higher than the capacity of the rotary valves.
Rice hull char was conditioned for activated charcoal
studies by devolatilization and partial combustion, and
cyclone efficiencies were studied during run 58-73. The
rice hull char was devolatilized at 704°C (1300°F) at an
estimated circulation rate of 91 kg/hr (200 Ibs/hr) for
10.5 hrs, reducing the char inventory from about 556 kg
(1226 Ibs) to approximately 454 kg (1000 Ibs). The combus-
tion of rice hull char was performed for three hours at an
estimated circulation rate of 136-159 kg/hr (300-350 Ibs/hr),
further reducing the char inventory from about 454 kg (1000
Ibs) to approximately 250 kg (550 Ibs).
58
-------
Cyclone efficiencies appeared to be 99-100% during char
devolatilization and 96-98% during rice hull char combustion,
but a sudden system upset prevented accurate measurements.
MATERIAL BALANCE RESULTS AND DISCUSSION
Oil and Water Yields
The pilot plant results confirm that a moisture free oil
yield of 42.8% can be obtained from the pyrolysis of Douglas
fir bark on a moisture-and-ash-free (MAP) feed basis. This
yield includes the bark wax and compares favorably with the
laboratory reactor result. Pyrolytic oil yields from rice
hulls and straw are somewhat lower at 30.2% and 33.3% respec-
tively. The cause of this deviation from the earlier labora-
tory reactor phase was not explored experimentally. However,
in all cases, the feedstock for pilot plant operation was of
considerably coarser size than that used for the laboratory
phase. Tree bark particle size for the pilot plant was in
the order of 70% above 200 mesh, and rice hulls and straw 95%.
Feedstock for the laboratory reactor was generally ground
through 200 mesh except for a few tests to check the effect
of coarse feed on oil yield. In those tests coarser particles
were shown to be detrimental to oil yield. In setting the
conditions for pilot plant runs it was anticipated that the
longer residence time in this unit of approximately one
second, compared to 0.3 in the laboratory would compensate
for the larger particle size. Although no firm experimental
verification of this could be obtained, the pilot plant
results on tree bark are consistent with this theory. The
results do suggest however, that rapid heat transfer to very
small feed particles is the key factor in flash pyrolysis
employing very short residence times. Although the oil yields
from rice hulls and straw fell somewhat short of the laboratory
reactor results, the yields from tree bark and straw are approx-
imately twice the Fischer Assay value and the yields from rice
hulls are 50% greater. In this regard the results confirm the
advantage of rapid heating and rapid quenching of the product
vapor.
The yields of oil and water were obtained during the oil
production runs (operating in the char heating mode). The
methods used were described in a preceeding section. A summary
of the macroscopic data is given in Table II-2. The data shown
correspond to all feed and product collections for the entire
run. They are not instantaneous (or rate) values. With the
exception of the vent loss (which was always small), the pro-
duct weights and analyses are all direct measurements.
59
-------
Table II-2. MACROSCOPIC LIQUID YIELD DATA, OIL PRODUCTION RUNS
Operating Conditions:
Char heating mode
Reactor ter.perature 482CC (900°F)
Reactor carrier gas - nitrogen
Run Number
FEEDSTOCK
Type
Moisture, wt % ,
Ash, wt % (db) '
Weight, Ib (db)
53-73
63-73
64-73
65-73
66-74
Tree
Bark
4.7
7.3
4890
Rice
Hulls
8.9
18.9
3446
Rice
Hulls
7.5
19.4
3963
Rice
Hulls
7.0
20.1
5080
Grass
Straw
5.6
9.1
3550
YIELD, wt % (db)
Oil
Primary Decanter
Condensers
Vent Loss
TOTAL
Wax
Water
Oil Moisture
Condensers
Vent Loss
Feed Moisture c
TOTAL
YIELD, wt % (MAF b )
Oil
Wax
Water
FISCHER ASSAY
Tar, wt % (MAF b )
30.5
4.9
1.5
1679"
19.2
5.4
1.4
26TO
18.3
5.6
1.4
2573
15.7
5.0
1.4
2271
2.8
39.8
3.0
16.7
18
nil
32.1
nil
25.2
nil
31.4
nil
26.0
20
nil
5.1
12.2
3.1
ifrf
3.9
22.1
4.2
(9.8)
2^77
4.0
22.8
2.3
(8.1)
2170"
3.3
23.4
3.3
(7.6)
22.4
28.0
nil
28.0
20.6
8.6
1.1
30.3
1.9
33.3
2.1
27.4
19
a db = dry basis feed and products
b MAF = moisture-and-ash-free basis
c Moisture in feed, kgs/100 kg dry feed
60
-------
The oil production operations consisted of eight runs:
four on tree bark, three on rice hulls, and one on grass
straw. Complete yield data were obtained for five runs: one
tree bark run and all the rice hull and grass straw runs.
Only partial yield data were obtained during the first three
tree bark runs since during this period the product quench
system was refined, improving its oil collection efficiency in
the primary decanter tank from about 50% in the first run
(49-73) to 80% in the last tree bark run (53-73). Measure-
ment, sampling and analysis techniques were also refined
during the first tree bark runs. Procedures for control of
the product oil moisture were also refined. Details of the
macroscopic recovery data are given at the end of this sec-
tion (Tables 11-15 through 11-19). The operating data are
summarized in Table II-3. Conditions were generally the
same for all runs, except for variations in feedstock prepar-
ation described previously. Several isokinetic oil yield
tests were made to corroborate the macroscopic oil yield
results. Although considerable care was exercised in recover-
ing or accounting for all oil produced, the isokinetic tests
were made to assure that no oversights or consistent bias
occurred. A description of the tests was given in the pro-
cedure section. It is reemphasized here that the isokinetic
measurements are wholly independent of the macroscopic
result. Further, the results are directly comparable,
because the procedures used to analyze for water, char and
oil were identical for all .determinations. The isokinetic
oil yield results are summarized in Table II-4 and compared
to the macroscopic results.
Eleven isokinetic tests were made during the rice hull
runs. The results of these tests are given in Table II-5.
The original purpose for all these tests was to determine oil
yield as a function of temperature. When the initial results
indicated the oil yield to be lower than expected, replica-
tions were made. These and the macro yields were all in good
agreement.
Because the rice hull oil yields were below the expected
results, a review of the laboratory reactor data was made
which revealed that yields were generally lower when proces-
sing coarser, wet feedstocks. The rice hull feed for the
pilot plant was processed as received, 95% +200 mesh and
8-9% moisture. Two additional tests were therefore made
processing material which had been oven-dried (1-2% moisture)
and screened through 20 mesh to reduce the +200 mesh to 85%.
The latter was an expedient in lieu of custom grinding due to
program time constraints. However, some improvement could be
expected if the hypothesis were true, and indeed the oil yield
was found to increase by 4%.
61
-------
TSbla II-3.
CP CPEKOTCNS DMA, OIL PRODUCTION CHAR HEKTEC
Temperatures, °C_
r\j
Run
Number
49-73
50-73
51-73
52-73
53-73
59-73
63-73
64-73
65-73
66-74
Feed-
Stock
Bark
Bark
Bark
Bark
Bark
Rice
Hulls
Rice
Hulls
Rice
Hulls
Rice
Hulls
Grass
Straw
Feed
Time
hrs
25.25
13.95
30.77
13.49
46.94
23.57
33.69
35.03
50.80
47.99
Feed
Rate
kgs/hr
58.1
75.7
57.9
62.7
49.6
60.0
50.9
55.5
48.8
35.6
Char
Rate
kgs/hr
181
318
204
181
204
340
204
295
295
318
Bin
Outlet
527
491
-
-
480
467
466
456
447
356
Char
Heater
Outlet
698
681
747
809
828
708
618
674
687
659
Char
to
Reactor
608
623
637
671
669
618
558
592
581
564
Reactor
Inlet
478
518
488
479
468
512
451
479
473
451
Reactor
Outlet
515
491
476
470
474
439
481
484
488
483
Collection
Train
44
48
48
52
52
45
58
57
58
57
-------
Table II-4. MfiCRDSCCPIC AND ISCKINETIC LIQUID YIELD DATA
(MOISTUKE-AND-ASH FREE BASIS)
Tree Bark Rice Hulls Grass Straw
Macroscopic Data
Yield, wt %
Oil and Wax 42.8 30.2 27.4
Water 16.7 26.6 27.4
a
Isokinetic Data
Number of Tests 2 11 2
Yield, wt %
Oil and Wax 42.8 28.7 30.8
Water 20.0 25.3 29.3
Fischer Assay
Tar Yield, wt % 18 20 19
a Slipstream sampling of product vapors and collection
in cold traps (See text).
63
-------
Table II-5. SUMMARY OF ISCKENflTIC OIL YIELD TESTS - RICE HULLS
Rice Hulls as-received (95% +200 mesh, 8-9% moisture)
63-73-B
64-73-A
64-73-B
65-73-B
65-73-C
65-73-D
65-73-E
65-73-F
65-73-G
Reactor
Temperature
°C °F
454
454
516
516
488
460
460
488
516
850
850
960
960
910
860
860
910
960
Yield, wt % (basis: dry feed)
Oil Water
21.1
25.5
24.0
23.7
22.2
22.0
22.7
21.3
20.0
,4
,4
22,
20,
21.3
20.9
21.0
19.4
19.2
18.4
18.8
Average 22.2 20.2
Dried & Screened Rice Hulls (85% +200 mesh, 1-2% moisture)
67-74-A
67-74-B
482
482
900
900
25.8
26.2
21.9
20.5
64
-------
The effect of temperature on oil yield was negligible
over the range tested, 455-516°C (850-960°F). However, this
may not always be generally true. In recognition of the
effect of coarse feedstock, particle heatup rate may have
been the controlling factor in these tests and the variation
of the final temperature within a narrow range, may have been
of lesser importance. For smaller particles with faster
heat-up rates and hypothetically higher oil yield, the yield
could show more sensitivity to reactor temperatures.
Gas Yield
Gas yield and composition determinations were made as
described in the procedural section. Again, these were
conducted simultaneously with the isokinetic tests, because
calculated yields depend upon the ratio of gas and feed rates,
and the latter was determined accurately during all isokine-
tic tests. These results are summarized in Tables II-6, II-7,
and II-8. The data are fairly reproducible and show some
sensitivity with regard to the process variables of feed-
stock and temperature.
Char Yield
Accurate determination of char yield was outside the
scope of this program since little information could be
obtained when operating in the char heating mode. However,
some information was obtained from char production runs when
operating in the direct heating mode. The general approach,
and some of the limitations, were discussed in the procedural
section. An additional consideration is that the final tem-
perature for the electrical heated runs was 50°F higher than
for the char heated runs. The magnitude of this effect may
be small in light of the relative insensitivity of the oil,
water and gas yields with respect to final temperature. With
these limitations in mind, the char composition data can
nevertheless be used to close the elemental balance. In
addition, the char yield from direct heated runs can be com-
pared to the yields from the char heated runs by accounting
for the weight difference. These results are presented below.
One direct measurement of char yield, and several char
compositions, were obtained during the tree bark char produc-
tion runs 37-73 and 38-73 when the receiver bin was cooled
and the entire contents transferred to storage, which allowed
accurate weighing and sampling.
Rice hull char production was conducted in one long run
during which the receiver bin was filled and only partially
emptied. Yield estimates in this case were made indirectly
via the rate of weight gain indicated by the bin scale.
Composition data were obtained on two drums of char removed
from the receiver bin.
65
-------
Table II-6. TREE BARK GAS YIEU3S AND CQMPQSITICN
Run
Date
Time
Reactor Temperature, °C
Feed Rate, kg/hr
Feed Rate, Ib/hr
Feed Moisture, wt %
Gas Yield
Wt %, dry feed
Nm /kg dry feed
(SCF/lb dry feed)
Gas Composition, mol %
H2
CO
co2
CH4
C2H4
C2H6
C3H6
C,H0
3 ,8
V
Molecular Weight
Gross Heating Value, kcal/Nm
(Btu/SCF)
Elemental Analysis
Wt %, C
H
O (by difference)
53-73
6/15
1045
477 (890°F)
51.7
(114)
4.7
15.8
0.114
(1.83)
9.96
35.96
45.05
1.15
5.43
0.62
0.91
0.00
1.24
32.5
1781
(200)
36.35
1.71
61.9.4
53-73
6/15
1347
477 (890°F)
52.6
(116)
4.7
17.8
0.122
(1.96)
5.56
36.28
50.03
1.09
4.77
0.51
0.86
0.04
1.08
34.4
1915
(215)
35.54
1.24
63.22
66
-------
Table II-7. RICE HUU. GAS YIEU3S AND OCMPOSITICN
Run
Test
Date
Time
Reactor Temperature, °C
Feed Rate, kg/hr
(Ib/hr)
Feed Moisture, Wt %
Gas Yield
Wt %, dry feed
Hm /kg dry feed
(SCF/lb dry feed)
Gas Composition, mol %
H2
CO
co2
CH4
C2»4
C2H6
C3!I6
C3H8
C4 +
Molecular Weight
Gross Heating Value, ,
kcal/rtnT
(Btu/SCF)
Elemental Analysis
Wt %, C
H
0(by
difference)
64-73
A
11/20
1402
453
(848)
86
(190)
5.38
0.035
(0.56)
0.58
45.77
51.67
0.97
0.23
0.14
0.086
0.029
0.344
36.16
1532
(172)
33.36
0.25
66.39
65-73
D
12/12
1420
460
(860)
65
(143)
6.8
5.09
0.036
(0.58)
2.14
55.93
38.73
2.07
0.34
0.26
0.17
0.085
0.00
33.48
1986
(223)
35.53
0.52
63.95
65-73
11/30
482
(900)
57
(126)
7.0
6.24
0.046
(0.74)
2.39
60.63
31.99
3.41
0.85
0.40
0.26
0.073
~\
0.00
32.16
2360
(265)
37.17
0.83
62.00
65-73
F
12/14
1024
488
(910)
77
(169)
7.3
5.48
0.038
(0.61)
1.46
57.38
37.32
2.16
0.48
0.20
0.14
0.034
0.679
33.88
2306
(259)
36.88
0.67
62.45
65-73
B
12/07
1455
513
(955)
63
(139)
6.7
8
0.
(0
2
58
32
4
1
0
0
0
0
32
.36
061
.97)
.07
.24
.46
.30
.21
.44
.44
.070
.603
.65
2707
(304)
38
1
60
.36
.17
.47
64-73
B
11/21
1100
514
(958)
56
(124)
7.0
5.75
0/041
(0.65)
3.54
46.50
40.79
4.24
0.93
0.40
0.40
0.081
0.122
33.73
2057
(231)
35.73
1.07
63.20
57-73
09/13
1200
510 a
(950)
29
(64)
9.7
10.95
0.082
(1.32)
5.49
43.89
35.45
9.07
2.66
1.18
1:82
0.00
0.042
31.53
3188
(358)
39.01
2.45
58.54
57-73
09/14
1200
510 a
(950)
36
(80)
9.7
9.30
0.072
(1.15)
3.10
68.99
22.48
2.98
1.32
0.078
0.66
0.00
0.00
30.59
2663
(299)
39.16
0.92
59.91
67-74
A
482
(900)
42
(92) „
1.8
5.25
0.041
(0.66)
2.12
75.40
19.12
3.00
0.16
0.21
0.00
0.00
0.00
30.16
2538
(285)
39.12
0.61
60.27
67-74
B
482 ,
(900) 8
39
" W»
5.00
0.041
(0.66)
3.18
73.66
14.31
6.84
1.22
0.58
0.16
0.00
0.05
28.71
3126
(351)
41.46
1.53
57.01
a Direct Heating'
b Dried & Screened -20 mesh
-------
Table II-8. GRASS STRAW GAS YIELDS AND COMPOSITION
Run
Date
Time
Reactor Temperature, °C 482
Feed Rate , kg/hr
Feed Rate, Ib/hr
Feed Moisture, wt %
Gas Yield
Wt %, dry feed
Nm /kg dry feed
(SCF/lb dry feed)
Gas Composition, mol %
H2
CO
co2
CH4
C2H4
C2H6
C3H6
C3H8
V
Molecular Weight
Gross Heating Value, kcal/Nm
(Btu/SCF)
Elemental Analysis
Wt % C
H
O (by difference)
66-74
2/05
1155
(900°F)
39
(90)
5.7
5.54
0.047
(0.75)
2.17
80.68
7.86
4.87
3.43
0.71
0.10
0.00
0.00
28.17
3420
(384)
43.59
1.52
54.89
66-74
2/11
1126
482 (900°F)
38
(88)
5.7
5.46
0.045
(0.72)
3.20
79.01
12.66
3.65
0.41
0.36
0.05
0.00
0.00
28.86
2796
(314)
40.74
0.88
58.38
68
-------
Char was not produced from grass seed straw during this
phase of the program.
Elemental Balances
From all of the foregoing data, elemental balances for
each feedstock were made. For the cases of bark and rice
hulls, closures on carbon, hydrogen and oxygen were calcula-
ted. This was done using the direct measurements of oil,
water and gas yield, together with char and gas compositions.
The char yield was obtained by difference. Two measures of
the accuracy of these results are available: closure of the
carbon and hydrogen balance, and comparison of the char yield
by difference to the yield obtained from the direct heated
char production runs. To the extent these statistics compare
favorably, the overall validity of all the preceding results
can be judged to fairly represent the reported yields and
conversions.
For the case of straw pyrolysis, char compositions were
not available as noted above. Here both the yield and the
composition are generated by difference. The balances
are given on a moisture-and-ash-free (MAP) feed basis. Only
C, H, and 0 are reported since N & S in the gas were not meas-
ured. Also, oxygen is determined by difference and actually
includes all other elements (including N & S). Since closure
obviously could not be obtained for each run, all of the
available yield and composition data for each feedstock have
been averaged to smooth out the random variability. The table
values can be traced item for item to the source data refer-
enced previously. Details of the elemental composition data
are presented at the end of this section (Tables 11-20, 11-21,
and 11-22) .
The elemental balances support the overall validity of
the results of this work. Closure of C, H, and O balances
for tree bark are respectively 96%, 97% and 104%. Similarly,
the closure for the rice hull runs is 99%, 100% and 100%.
The statistic char-yield-by-difference compares favor-
ably to the results of the electrically heated char produc-
tion runs. On an MAF basis, the by-difference char yield for
bark is 22.4% vs the direct measured value of 22.3%.
Oil Moisture Analysis
Three methods were used in this work: azeotropic dis-
tillation using benzene or toluene and the titrimetric Karl
Fischer (KF) method. The azeotropic method was chosen and
used for all of the analyses so that the data would be inter-
nally consistent and comparable. This was done because the
KF method could not be used to analyze samples obtained from
isokinetic tests, the Brinks demister pad, or high moisture
69
-------
Table II-9. ELEMENTAL BALANCE PYH3LYSIS OF TREE BARK
(BASED CN MOISTORE-AND-flSH-FFEE FEED)
Yield, %
Oil
Wax
Water
Gas
a
Char
TOTAL b
Feed
% Closure
39.8
3.0
16.7
18.1
22.4
100.0
100.0
Composition, %
C
60.1
78.5
35.9
87.0
52.3
54.3
96
H
6.
13.
11.
1.
4.
5.
6.
97
0
1
1
5
3
9
1
0
33.9
8.4
88.9
62.6
8.7
41.2
39.6
104
a Difference
b (E Yield x Composition) / 100
70
-------
Table 11-10. ELEMENTAL BALANCE PYRQLYSIS OF RICE HULLS
(BASED CN MDISTOKE-AND-ASH-FKEE FEED)
Yield, %
Oil
Wax
Water
Gas
a
Char
b
TOTAL
Feed
% Closure
30.2
0.0
26.6
7.3
35.9
100.0
100.0
Composition, %
c
62.24
37.72
75.84
48.77
49.22
99.1
H
6.11
11.1
0.70
4.74
6.55
6.58
99.5
0
31.65
88.9
61.57
19.42
44.68
44.2
101.1
a Difference
b (E Yield x Composition) / 100
71
-------
Table II-ll. EtEMEMEAL BALANCE PYROLTSIS CF GRASS STRRW
IBASED CM MDISTORE-AND--ASH-FREE
Yield, %
Composition, %
Weight (per 100 kg of feed)
ro
Oil
Wax
Water
Gas
a
Char
TOTAL
Feed
33.3
2.1
27.4
5.5
31.7
100.0
100.0
C
61.9
78.5
42.2
79.6
49.8
49.8
H
6.4
13.1
11.1
1.2
3.7-
6.7
6.7
O
31.7
8.4
88.9
56.6
16.7
43.4
43.4
C
20.6
1.6
0.0
2.3
25.22
49.8
H
2.1
0.2
3.0
0.07
1.18
6.7
O
10.5
0.1
24.3
3.1
5.30
43.5
a Char yield and composition by difference.
-------
condensate samples. The KF method was thus useful only
for primary oil samples obtained from the decanter tank
(which represents 80% of the total oil yield). Also, the
effect of moisture on oil yields is small, about 1-2%
absolute on the oil yield result. The foregoing are all
based on the benzene azeotropic method with the exception
of the rice hull and straw isokinetic tests where toluene
was used as the azeotropic solvent.
The overall effect of choosing the azeotropic method is
to understate the oil yield somewhat since the product
decomposes slightly to CO- and HLO during the distillation
process. Taking the KF method as being the more accurate,
the benzene azeotropic method gives a high moisture result
by a factor of 1.2, and the toluene azeotropic method a
higher result by a factor of 1.6 (at an oil moisture content
of 15%). The net result of this is to understate the oil
yield by 1% absolute using benzene and 2-1/2% absolute using
toluene. This has no effect on the C and H closures, as
these in effect are based on the liquids (oil plus water)
taken together.
PRODUCT TREATMENT AND PROPERTIES
Product Oil Properties
Table 11-12 summarizes the pyrolytic oil properties for
the product oils obtained from all three feedstocks (tree
bark, rice hulls and grass straw). Figure II-3 shows the
viscosity vs temperature behavior of the oils. Other proper-
ties and yield data are presented in previous tables in this
section of the report.
Oil Treatment -
Filtration of the oil to remove char was generally not
successful due primarily to the high viscosity of these oils.
Tree bark oil from Run 53-73 was filtered more easily than oil
from Runs 49-73 through 52-73. However, 100 micron pore size
cloth was used and little char removal was accomplished in
either case. Only one drum was completely filtered through
25 micron cloth, but it had a low char content initially.
Typical results were:
Tree Bark Oil
Drum No. Filter size Char Content, wt %
(microns) Before After
167 100 1.64 1.42
170 100 1.47 1.30
107 25 0.68 0.23
(167) (25-Lab Test) (1.42) (0.48)
73
-------
Table 11-12. PYEOLYTIC OIL PROPERTIES
Feedstock Bark
Lot I
Composition, wt %
Quench Fluid 2.4
Char 1.1
Moisture
(Karl Fischer) 13.49
Moisture
(Benzene Azeo) 15.6
Bark
Lot II
0.0
1.2
9.58
11.8
Rice
Hulls
2.0
3.3
11.96
11.9
Rice Straw
Hulls
0.0 0.0
1.9 3.5
—
10.6 14.9
Specific Gravity, T/16°C (60°F)
50°C (122°F) 1.254
64°C (148°F) 1.243
79°C (175°F) 1.232
Viscosity, centi stokes
50°C (122'F) 600
(284SSF)
64°C (148°F) 200
( 99SSF)
79»C (175°F) 82
( 42SSF)
Pour Point, °C 21
(CF) (70)
Flash Point,
(PMcc) °C 110+
<°F) (230+)
Heat of Combustion,
cal
gm 5002
(Btu/lb) (8998)
Ultimate Analysis, Wt %
Moisture 13.49
Ash 0.66
C 50.38
H (dry basis) 6.90
N 0.27
S 0.03
Cl 0.03
O (by difference) 28.24
1.266
1.255
1.244
2000
(940SSF)
630
(300SSF)
160
( 77SSF)
24
(75)
110+
(230+)
5193
(9341)
9.58
0.74
52.57
6.67
0.32
0.04
0.02
30.06
1.230
___
1.200
660
(317SSF)
—
66
( 34SSF)
7
(45)
84
(184)
5257
(9457)
11.96
2.32
53.43
6.38
0.69
0.06
0.04
25.12
1.243
1.200
2100
(1001SSF)
- —
— — _
150
( 72SSF)
...... ___
— -
___ - —
— — — — — —
5529
(9945)
7.0
1.91
55.46
6.65
1.24
0.09
0.07
27.58
74
-------
5000
FIGURE II-3
PYROLYTIC OIL
VISCOSITY VS TEMPERATURE
1000
500
o
Q.
bJ
O
CO
100
50
A RICE HULL OIL AT 11.9% MOISTURE
O BARK OIL AT 15.6% MOISTURE
X STRAW OIL AT 14.9% MOISTURE
E BARK OIL AT 11.8% MOISTURE
(BENZENE AZEOTROPE MOISTURES)
IOL
I
1.5
1.55
1.6
I/T XIO-3.
1.65
1.7
1.75
75
-------
A small quantity of the Drum 167 oil which had been
passed through the 100 micron cloth was filtered through
25 micron cloth in the laboratory, but the cloth soon
became blinded. Drum 167 could not be filtered through
25 micron cloth in the pilot plant under any circumstances.
Filtration of rice hull oil was even less successful
than the tree bark product. Oil from Drum 525 was filtered
first through 25 micron cloth and then 5 micron cloth.
Typical results were:
Rice Hull Oil
Test Filter Size Char Content/ wt %
(microns) BeforeAfter
1 25 3.24 3.16
2 5 3.16 3.23
3 5 3.23 3.19
Although this oil readily passed through even the small-
pore 5 micron filter cloth, no char was removed. The general
conclusion must therefore be that filtration is not an
effective means of removing char from this type of oil.
An attempt was made to remove char from the oil using
the pilot plant centrifuge, but a limiting value was found
of about 1.5% char in the oil. The density of the small
particles is apparently very close to that of the oil, so
further reduction of char by centrifuging does not seem
practical.
Decantation of the paraffinic quench fluid from the
Lot I tree bark oil was only partially successful. After
two hours of settling at 60°C (140°F) some of the quench
fluid could be decanted, thus lowering its content in the
oil from 3.7 to 2.4%. However, Lot II tree bark oil con-
tained no quench fluid whatsoever.
Product Char
A summary of the char analyses is presented on Table
11-13. Table 11-14 summarizes the product char properties.
Figures II-4 and II-5 present the measured screen analyses
and particle size distribution, respectively. As expected,
the char produced by direct heating has higher volatile
matter content and lower ash than the char product after
combustion tests. Bulk densities of all chars measured are
about 0.19 gm/cc (12 Ib/ft ).
76
-------
Water
A detailed analysis was made of the byproduct water
collected in the total condensers in tree bark Run 53-73.
About 14.5% of the sample collected was oil recoverable by
distillation. The COD of this water phase was about 205
grams per liter, and pH was 3.6. A typical analysis of this
water is:
Component ppm g/1
Na 7
K 86
Cl 0.04
Fe 6080
Ca 8
Mg 2
Suspended
Solids 0.11
COD 204.8
77
-------
Table 11-13. CHAR ANALYSIS SUMMARY
Tyler Mesh Screen Analysis
(Cumulative % Retained)
Composition, vt
Rice Hulls
Run 57 (Original Char) a
Run 58 (After combustion tests)
00 7r.?o Park
RUM -II (o» igiiMl Char) '*
pun -47 (After combustion tests)
Run 53 Composite (Char and
Petroleum Coke)
[Particle density, 32]
11
8.81
10.30
5.3
3.9
2.9
[0.20]
£2
51.
45.
27.
15.
20.
[0.
85
70
8
3
1
62]
1.00
83.
gg-
67.80
48.
27.
53.
[1.
7
5
4
40]
150
92.
75.
65.
39.
70.
[1.
200
06
20
1
4
2
45]
95
79
76
50
80
[1
.01
.40
.4
.8
.5
.41]
325
97.
87.
89.
67.
89.
[1.
97
20
0
7
7
45]
Volatile
H2O Matter Ash C H N i 2
15.75 51.17 37.35 2.43 0.31 °'15 8'72
0.04 6.88 58.52 34.68 1.61 0.27 °'02 4-90
1.77 23.54 17.09 65.43 3.13 0.29 °-10 13.95
1.03 10.76 25.79 65.47 1.77 0.20 °-51 6'12
0.36 7.82 19.15 72.48 1.85 1.33 °-63 4-04
3Produced by direct heating
-------
Table 11-14. SS»ftHS,aP-<3aR PROPERTIES
HISTORY OF MATERIAL
Analysis, wt %
Volatile Matter
Ash
£
H
:i
s
0 (by difference)
Gross Keating Value, eal/g
Measured
Calculated a
Run 41 - Original
Tree Bark Char
(electrically
produced
23.54
17.09
65.43
3.13
0.29
0.10
13.96
5977 (10751
Btu/lb)
5770 (10379
Btu/lb}
Run 47 - Tree
Bark Char after
combustion tests
10.76
25.79
65.47
1.77
0.26
0.59
6.12
Run 53 - Tree
Bark Char plus
Petroleum Coke
7,. 8 2
19.15
72.48
.85
.83
1.
1.
0.65
4.04
Run 57 - Original
Rice Hull Char
(electrically
produced)
5.17
52.68
36.56
2.20
0.34
0.03
8.19
Run 5tf - Rice
Hull char after
combustion tests
6.88
58.52
34.68
1.61
0.27
0.02
4.90
5787 (10409
Btu/lb)
5653 (10169
Btu/lb)
6392 (11498
Btu/lb)
6338 (11401
Btu/lb)
3357 (6039
Btu/lb)
3362 (6048
Btu/lb)
3149 (5664
Btu/lb)
Specific Heat
T,°F
50
100
200
300
400
500
600
122
212
392
572
752
932
1112
Apparent Bulk Density kg/m
0.283
0.405
0.157
0.176
0.213
0.207
0.233
0.163
0.244
0.146
0.159
0.185
0.214
0.213
186 (11.6 lb/ftj) 190 (11.89 lb/ftJ)
0.142
0.194
0.185
0.186
0.206
0.216
0.228
199 (12.4 lb/ftj)
Calculated from Dulong Formula: He (Btu/lb) = 14544 (C)
62028 (H~)
0
+ 4050 (S)
-------
FIGURE H-4
BARK CHAR SCREEN ANALYSIS
28
oo
o
V)
UJ
2
-------
00
ft
tu
tu
a
z
-------
Table 11-15. SUMMARY OF LIQUID YIELD
TREE BARK - RUN 53-73
Weight of Feed;
Dry
Moisture
As fed
Product Oil Yield
Source
KP-101
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (Yield)
Water Yield
Source
Product Oil Moisture
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (recovery)
Less: Feed Moisture
Pyrolytic Water
Yield (basis: dry feed)
Quench Oil Inventory
Jsa.
lb
2218
109
232T~
(basis: dry
Weicht,
kg lb
662 1460
94 208
30 66
26 58
33 73
845 1865
4890
240
5130
feed)
% of feed %
29.9
4.2
1.3
1.2
1.5
38.1
Weight,
kg lb
114 251
252 556
15 33
68 150
449 990
(109) (240)
340 750
of yield
78.3
11.1
3.5
3.1
4.0
100.0
Charge,
Recovery,
Net Change,
kc
1196
1222
+26
lb
2637
2695
•••58
15.3%
82
-------
Table 11-16. SUMMARY OF LIQUID YIELD
RICE HULLS - RUN 63-73
Weight of Feed:
Dry
Moisture
As fed
Product Oil Yield
Source
KP-101
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (yield)
Water Yield
Source
kg_
1563
153
1716
3446
338
3784
(basis: dry feed)
Weight, % of feed % of yield
kg_
300
75
10
-
Ib
661
166
22
-
19.2
4.8
0.6
1.4
73.8
18.5
2.3
5.4
385
Product Oil Moisture
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (recovery)
Less: Feed Moisture
Pyrolytic Water
Yield (basis: dry feed)
Quench Oil Inventory
Charge
Recovery
Net Change,
648
707
+59
849
26.0
Weight,
100.0
61
343
3
65
Ib
135
755
6
144
472
(153)
1040
(338)
319
702
20.4%
Ib
1428
1559
+131
83
-------
Table 11-17. SUMMARY OF LIQUID YIELD
RICE HULLS - RUN 64-73
Weight of Feed:
Dry
Moisture
As fed
Product Oil Yield
Source
KP-101
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (yield)
Water Yield
Source
Product Oil Moisture
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (recovery)
Less: Feed Moisture
Pyrolytic Water
Yield (basis: dry feed)
Quench Oil Inventory
Charge,
Recovery,
Net Change,
643
594
-49
yg Ib
1797
146
1943
3963
321
4284
(basis: dry feed)
Weight,
% of feed % of yield
329
96
5
25
Ib
724
212
11
55
18.3
5.3
0.3
1.4
72.3
21.1
1.1
5.5
455 1002
25.3
Weight,
Ib
73 160
404 890
6 14
41
91
524
(146)
378
1155
(321)
834
21.0%
Ib
1418
1309
-109
100.0
84
-------
Table 11-18. SUMMARY OF LIQUID YIELD
RICE HULLS BUN 65-73
Weight of Feed:
Dry
Moisture
As fed
Product Oil Yield
Source
KP-101
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (yield)
Water Yield
Source
Product Oil Moisture
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (recovery)
Less: Feed Moisture
Pyrolytic Water
Yield (basis: dry feed)
Quench Oil Inventory
Charge,
Recovery,
Net Change,
1S2.
640
596
-44
kg Ib
2304
175
2479
(basis:
Weight,
kg Ib
361
116
5080
385
5465
dry feed)
% of feed % of yield
33
510
Ib.
1410
1314
-96
795
255
73
15.7
5.0
1.4
1123 22.1
Weight,
Ib
76 168
539 1188
75 166
690
(175)
515
1522
(385)
1137
22.4%
70.8
22.7
6.5
100.0
85
-------
Table 11-19. SUMMARY OF LIQUID YIELD
GRASS STRAW - RUN 66-74
Weight of Feed:
Dry
Moisture
As fed
Product Oil Yield
Source
KP-101
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (yield)
AMSCO solubles
Water Yield
Source
Product Oil Moisture
T-104 & Brinks Shell
Brinks Pads
Other
Vent Loss
TOTAL (recovery)
Less: Feed Moisture
Pyrolytic Water
Yield (basis: dry feed)
Quench Oil Inventory
Charge,
Recovery,
Net Change,
513
489
-24
kg lb
1610 3550
95 210
1705 3760
(basis: dry feed)
Weight, % of Feed
331 731
132 290
7 16
18
488
31
579
lb
1130
1078
-52
39
1075
68
1143
20.6
8.2
0.4
1.1
30.3
1.9
32.2
Weight,
lb
58 129
367 809
11 24
60
132
496 1094
(95) (210)
401 884
24.9%
86
-------
Table 11-20. SUMMARY OF COMPOSITION DATA
PYROLYSIS OF TREE BARK
(MOISTURE-AND-ASH-FREE)
Feed
Oil
Run
AR 10/16/72
AR 5/17/73
53
53
Average
Average
53.71
54.07
53.42
55.90
54.28
60.14
H
6.06
6.22
6.01
6.13
6.10
49
49
&
&
50
51
52
53
52
53
60
59
60
60
60
60
•
•
•
*
•
•
94
04
17
44
16
10
5
5
5
6
6
6
.95
.82
.80
.10
.18
.13
6.10
O(diff)
39.62
33.76
Gas
53
53
Average
Char
36.35
35.54
35.94
1.48
62.58
37
37
37
37
38
41
Average
92.83
93.98
86.66
84.18
85.25
78.92
86.97
4.33
8.70
87
-------
Table 11-21. SUMMARY OF COMPOSITION DATA
PYROLYSIS OF RICE HULLS
(MOISTURE-AND-ASH-FREE)
FEED
Run Ash
63 18.91
64 19.39
65 20.11
AVG 19.47
57 18.20
CHAR
Spl Source Ash
856 B101 T103 50.22
857 " 50.96
MAF
C
49.
49.
49.
49.
54.
33
31
03
22
19
H
6.
6.
6.
6.
6.
66
57
52
58
34
N
0.
0.
0.
0.
0.
51
51
27
40
43
S
0.06
0.14
0.04
0.08
0.23
O
43.
43.
44.
43.
38.
28
46
04
60
81
MAF
C
74.
72.
75
70
H
4.
4.
86
34
N
0.
0.
62
57
S
0.04
0.04
O
19.
22.
73
35
858, 859, 60 not analyzed
477 B101 grab
478
AVG 856, 857
AVG all
Oil Analyses (MAF
Run Spl HjO Ash
63 722 15.1 0.80
63 732 10.7 1.53
64 748 10.7 1.64
64 770 7.8 1.75
64 778 9.2 1.93
65 794 18.7 1.19
65 799 14.2 1.33
65 813 8.7 2.99
65 841 11.3 2.24
65 845 13.1 2.71
AVG all
Test Temp
*F
64-73-A 848
64-73-B 860
AVG (char heat) 850
65-73 900
65-73-F 910
67-74-A . 900
67-74-B a 900
AVG (char heat) 900
AVG all 900
65-73-B 955
64-73-B 958
57-73-A a 950
57-73-B 950
AVG (char heat) 960
AVG all 950
NOTES; 8
79.
76.
73.
75.
43
49
73
84
- Azeo H,0,
C
61.
61.
60.
61.
61.
61.
64.
62.
63.
63.
62.
Gas
19
06
30
77
99
17
73
78
65
76
24
4.
4.
4.
4.
78
98
60
74
not corr
H
5.
6.
5.
6.
6.
5.
6.
6.
6.
6.
6.
Analyses
Yield
5
5
5
6
5
5
5
5
5
8
5
10
.38
.09
.24
.24
.48
.25
.00
.86
.49
.36
.75
.95
9.30
7
8
Direct
.06
.59
79
05
94
32
27
81
05
35
29
20
11
0.
0.
0.
64
64
59
for char
N
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
79
84
88
91
95
90
93
92
96
95
90
comp)
0.69
0.31
0.04
S
0.07
0.05
0.07
0.07
0.08
0.10
0.05
0.10
0.08
0.08
0.08
14.
17.
21.
19.
0
32.
32.
32.
30.
30.
31.
28.
29.
29.
29.
30.
46
58
04
42
16
00
81
93
79
91
24
85
02
01
67
& Yield Summary
AHc
172
223
198
265
259
285
351
262
290
304
231
358
299
267
364
C
—
33.
35.
34.
37.
36.
39.
41.
37.
38.
38.
35.
39.
39.
37.
38.
36
53
44
17
88
12
46
72
66
36
73
01
16
04
0'6
H
_
0.25
0.52
0.38
0.83
0.67
0.61
1.53
0.70
0.91
1.17
1.07
2.45
0.92
1.12
1.40
O
_
66.
63.
65.
62.
62.
60.
57.
61.
60.
60.
63.
58.
59.
61.
60.
39
95
17
00
45
27
01
57
43
47
20
54
91
84
53
Heating
-------
Table 11-22. SUMMARY OF COMPOSITION DATA
PYROLYSIS OF GRASS STRAW
(MOISTURE-AND-ASH-FREE)
H
Feed
Run Composite
(replicate)
Average
Lab reactor data (Avg. of 3)
Oil
Run Composite
Ga^s
Run 66
Run 66
Average
50.1
49.4
49.8
47.9
6.91
6.50
6.7
6.2
43.0
44.1
43.5
45.9
61.9
43.59
40.74
42.16
6.4
1.52
0.88
1.20
31.7
54.89
58.38
56.64
89
-------
SECTION VI
PHASE III - PRODUCT EVALUATION
INTRODUCTION
This section presents the results of the product evalu-
ation studies (Phase III) using the pyrolytic char and oil
products generated during the pilot plant studies (Phase II) .
Converting solid waste materials to liquid fuel is
intended to make it possible to conveniently store and trans-
port the combustible portion of the waste while reducing its
ash content, and to be able to fire it using conventional
fuel-oil-firing equipment. The KVB Company, Inc. of Tustin,
California was hired under subcontract to conduct combustion
tests to determine flame stability and pollutant emissions of
the pyrolytic oils from tree bark and rice hulls relevant to
their use as fuels in industrial and utility boilers. The
results of the tests were favorable and show that pyrolytic
oils can be successfully used as a substitute or supplement
to the residual petroleum-based oils typically used in
commercial fuel oil firing equipment.
The byproduct char samples were evaluated for use as
barbecue briquettes and as a raw material for the manufacture
of activated carbon. The Royal Oak Charcoal Division of
Georgia Pacific Corporation contributed studies in associa-
tion with this program on two* lots of Douglas fir bark char
and concluded that this material is suitable for the manu-
facture of good quality barbecue briquettes provided the
volatile matter content is kept below about 15%.
The St. Regis Paper Company was hired under subcontract
to evaluate the suitability of bark and rice hull char for
making activated carbon. The St. Regis equipment was un-
fortunatley not capable of handling and pelletizing the fine
char at commercially feasible levels of binder concentrations,
90
-------
Nevertheless, the results indicate that char could be activa-
ted with some difficulty even under these conditions and that
a modestly attractive throw-away, powdered activated carbon
could probably be produced. It was also found that the rice
product was too high in ash to produce an acceptable activa-
ted product. In order to verify these conclusions, however,
further tests would have to be conducted employing equipment
designed to handle this finely divided material.
PRODUCT PREPARATION .
Pyrolytic Oil
Preparation of the bark and rice oils shipped to KVB
for combustion testing was discussed in some detail in
Section V of this report (Pilot Plant Studies). These steps
consisted of filtering and centrifuging the oil for char
removal and decanting the oil to separate any carry-.over
quench fluid; A summary of the properties of the tree bark
oil shipped to KVB is presented on Table III-l.
Table III-2 shows a comparison of the properties of
rice hull oil - Lot 1 as measured by both Garrett and KVB.
The only substantial difference is in the measured moisture
content of the oil. The differences among moisture contents
obtained by using different analytical techniques were dis-
cussed in the previous section. The Toluene Azeotropic
Distillation Method used by KVB suffers from systematic errors
when applied to pyrolytic oils as discussed earlier. The
oil analyzed by Garrett shows about 12% water by the Karl
Fischer Method and this is within the range to be expected
compared to KVB's results.
Conditioning of Pyrolytic Char
The starting material for activation studies by the
St. Regis Paper Company was the tree bark char produced by
direct heating in Run 37-73. The material was originally
to have been sent directly to St. Regis, but because of the
high volatile matter (VM) content of 23.03%, further devol-
atilization in the pilot plant was necessary. Two batches
were processed, with reactor gas temperatures being approx-
imately 93°C (200°F) higher for the second batch. Product
weights of 4.6 kg (10.14 lb) (9.62% VM) and 6.6 kg (14.48 Ib)
(3.87% VM) resulted. Samples of 0.3 kg each of original
char (23.03% VM), Batch 1 product C9.62% VM), and Batch 2
product (3.87% VM) were shipped to St. Regis.
Table III-3 presents an analysis of the Run 37-73 char
before screening or devolatilization, and includes a summary
of devolatilization conditions and results.
91
-------
Table III-l. PROPERTIES OF BARK OIL SHIPPED
TO KVB ENGINEERING FOR COMBUSTION TESTING
LQT I II
Pilot Plant Runs 49, 50, 51, 52 53
New weight, Kg 821 602
Composition, wt %
Moisture 17.0 13.3
Quench Fluid 2.4 0.0
Char 1.1 1.2
Specific gravity, T/16°C (60°F)
50°C (122°F) 1.254 1.266
64°C (148°F) 1.243 1.255
79°C (175'F) 1.232 1.244
Viscosity SFS Cp SFS Cp
50°C (122aF) 284 753 940 2500
64°C (148°F) 99 260 300 800
79°C (175°F) 42 105 77 200
Pour point 21°C (70°F) 24°C (75°F)
Plash Point (P.M.c.c.) 110+°C (230+°F) 110+°C (230+°F)
92
-------
Table II1-2.
COMPARISON OF PYROLYTIC OIL PROPERTIES
MEASURED BY GARRETT AND KVB
Ultimate Analysis, wt %
Moisture
Ash
C
H fexcl. H~0)
N Z
S
Viscosity, Centistokes
77°C (170°F)
79°C (175°F)
Flash Point
Heat of Combustion
RICE HULL OIL - LOT 1
Garrett Analysis
11.96
2.32
53.43
6.38
0.69
0.06
90 (420SSU)
84°C (184°F)
5250 cal/gm
(9457 Btu/lb)
KVB Analysis
19.56^
2.36
55.27
4.73
0.78
0.08
140 (650 SSU)
104°C (223°F)
5400 cal/gm
(9723 Btu/lb)
Karl Fischer titration
Toluene azeotrope distillation
93
-------
Analv«i« of Bark Char 37-73
Component
Kt t (dry ba«is)
DovolatiHtntion Conditions
Wt feed est, :kg
Wt product-, kg
Devolatilization Losses, Wt »
Run Duration, minutes
Feed
-20 mesh Bark Char from Run 37-73
Init. VK, before screening
Transport gas
Gas Flow Rate (Preheater/Transport)
Gas Residence Time per cycle
(Reactor Inlet to Cyclone #1 exit)
Temperatures, °C
Preheater Exit Gas
Reactor Inlet Gas
Reactor Exit Gas
Cyclone il inlet Gas
Feed Rate, kg/hr
Cyclone Efficiencies (Basis: products)
Cyclone *1
Combined (three cyclones)
Pressure, Reactor Inlet, atmg
Char volatile matter (cyclone II), wt.%
Table IZX-3. PROPERTIES AND DEVOLATILIZATION CONDITIONS
FOR BARK CHAR SENT TO ST. REGIS PAPER COMPANY
c
68.00
H N 9
3.48 0.27 0.08
1
t.t
ASH
26.75
2
8.5 "
0
(Dlff.)
1.42
Fischor
Assay
Tar
2.6
Volatile
Matter
23.03
4.6(10.14 lb)
28.1
167.
23.03%
2.5/1.0 SCFM
0.5 sec
482 ( 9000F)
760(1400°F)
704(1300"?)
2.30(5.07 Ib/hr)
67.4
92.6
<0.1(<1.5 peig)
9.62
6-6(14.48 lb)
26.1
240.
23.03%
2.5/1.0 SCFM
0.5 sec
871(UOO«F)
482 ( 900°F)
871-899(1600-16SO"F)
816(1500'F)
2.23(4.91 Ib/hr)
64.1
92.5
<0.1(<1.5 psig)
3.87
-------
OIL EVALUATION
Combustion tests were conducted by KVB, Inc. on two lots
of tree bark oil and two lots of rice hull oil described in
Tables III-l and III-2 respectively. A detailed description
of the test conditions and results of these tests is given in
a report from KVB to Garrett which is included in the Appendix
(Section IX of this report). The objective of the KVB tests
was to determine flame stability and pollutant emissions of
the pyrolytic oils relevant to their use as fuels in indus-
trial and utility boilers. The KVB test results are summar-
ized below and other details from earlier tests are given in
Reference 2.
Pyrolytic oils were found to be suitable fuels for a
large power station boiler with a properly designed storage
and handling system. Some of the differences of pyrolytic
oils from the heavy residual oils commonly used in utility
boilers are:
1. They are water-based and therefore are immiscible
with petroleum-derived oils. (They can, however, be
blended as a binary dispersion with most types of
residual fuel oils).
2. They are slightly corrosive because they contain
acetic acid as well as other organic acids; therefore,
they require additional consideration in shipping,
storage, pumping, and plumbing. -•
3. They are heat-sensitive. Although moisture content
can be controlled in the production process to give a
viscosity-temperature relationship similar to that of
residual oil, the preheating which is necessary to bring
the fuels to firing viscosity (100-200 SSU), if not done
carefully, can cause irreversible changes in physical
properties. In addition, the viscosity is not a function
of temperature alone, but of the temperature-time history
of the pyrolytic oil. Thus the liquid's properties can
change over a long storage period at elevated temperature,
4. A lower heat content, 25 to 30% less than that of a
similar volume of residual oil, requires greater fuel
pumping and storage capacity for a given heating rate.
Typical properties of No. 6 fuel oil and pyrolytic oils are
shown in Table III-4.
95
-------
Table III-4. TYPICAL PROPERTIES OF NO. 6 FUEL OIL
AND PYROLYTIC OIL
No. 6 Pyrolytic Oil
Carbon, wt % (dry basis) 85.7 57.5
Hydrogen 10.5 7.6
Sulfur 0.5 - 3.5 0.1 - 0.3
Chlorine - 0.3
Ash 0.5 0.2-0.4
Nitrogen 2 Q 0.9
Oxygen •"*" 33.4
Btu/pound 18,200 10,500
Sp.Gr. 0.98 1.30
Lb/gallon 8.18 10.85
Btu/gallon 148,840 113,910
Pour Point °F 65 - 85 90
Flash Point °F 150 133 &
Viscosity SSU @ 190°F 90 - 250 1,000&
a
jumping temperature °F 115 160
a
Atomization temperature °F 220 240
Oil contains 14 wt % H_0
96
-------
Tree Bark Oil
Petroleum-derived residual oil was used to fire the KVB
test boiler for cold starts and to purge the boiler for the
last ten to fifteen minutes of firing on each day's testing
on tree bark oil. Lot 1 was burned satisfactorily for sever-
al minutes at a time, but trouble was encountered due to
repeated plugging of the atomizers by char particles present
in the oil. While larger diameter atomizer offices in a
commercial scale boiler might not have plugged, these small-
er scale tests reveal the need for a complete range of stan-
dardized inspection tests for pyrolytic oils similar to the
common tests used on petroleum-derived oils.
The remainder of the tree bark oil tests were made using
the Lot 2 oil. With the ignitor on, the tree bark oil burned
stably over a wide range of conditions. Operation without
sustained use of the ignitor might have been possible if
there had been enough pyrolytic fuel to enable variations in
burner settings. Ignitor fuel consumption was small:'less
than two percent of the total heat input.
After a shutdown of several seconds to five minutes, the
burner could be restarted on tree bark oil without using
residual oil. No attempt was made to start up on tree bark
oil with a cold furnace, but this was satisfactorily accom-
plished later with pyrolytic rice hull oil.
Changes from one fuel to the other were made gradually
over a period of several minutes, so that a wide range of
mixtures of tree bark oil and residual oil were burned. No
difficulty was encountered in burning these mixtures.
For all the test points on this batch of tree bark oil,
the firing rate was approximately 380,000 Kcal/hr (1,500,000
Btu/hr). A total of seven hours of running time was accumu-
lated firing 100% pyrolytic oil. Carbon monoxide concentra-
tions on a dry basis were below 50 ppm (volume) for all points,
and at normal excess air levels, there was no visible smoke
except for a very light plume (about 10% density) which
occurred at a few conditions. However, moving the diffuser
and gun 1.3cm (1/2 in.) farther into the fire box eliminated
the smoke. Combustion was stable and smokeless as excess
oxygen was increased from 2.5% to 9.4% with a tree bark oil
temperature of 91°C(196°F).
The tree bark oil burned well with two-stage combustion,
a widely-used technique for reducing nitric oxide formation
in large boilers. Figure III-l shows nitric oxide concentra-
tion in the flue gases vs Ag, the percent of stoichoimetric
air flow through the burner. Each curve has two sections.
97
-------
FIGURE III-I
CORRECTED NITRIC OXIDE CONCENTRATION VS. BURNER AIR FLOW
600 i-
ro
vo
00
200
O,D « SINGLE STAGE COMBUSTION
•,• > TWO-STAGE COMBUSTION
BARKOIL (0.32% N)
FEDERAL LtMIT (NEW BOILERS)
HAWAIIAN RESIDUAL (0.092%N)
j_
I
I
40
80 120 160
PERCENT STOICHIOMETRIC AIR AT BURNER
X
m
200
-------
The open symbols represent single-stage combustion; that is,
all the air flows through the burner. The solid symbols
represent two-stage combustion wherein part of the air flows
through the burner and the remainder enters the furnace
farther downstream through the second-stage air torus. The
overall combustion air flow was as low as 117% of stoichio-
metric for several of these points.
For the Lot 2 tests, a low-nitrogen No. 5 residual oil
from a Hawaiian refinery was used for startup and post-run
fuel system purges. Nitric oxide data for this oil are also
shown in Figure III-l. An analysis of the oil is given in
the KVB report in the Appendix.
For both oils, two-stage combustion results in less
nitric oxide emission. The original data sheets are also
reproduced in the KVB report in the Appendix. Typical test
data for tree bark oil and the Hawaiian residual oil are
given in Table III-5. No changes in burner configuration
were made in changing from one fuel to the other.
Rice Hull Oil
The fuel derived from rice hulls was also supplied in
two batches. Complete analyses are given in Table III-2.
The water analyses are believed inaccurate as discussed above
due to the Toluene Azeotrope Method used to determine moisture
content. Corrected values of C, H, and O were used in deter-
mining flue gas O- vs excess air for thi,s oil. This particular
calculation is not very sensitive to fuel composition.
Lot 1 -
Initially, attempts were made to fire Lot 1 under exactly
the same conditions which were successful with the Lot 2
tree bark oil. However, these were unsuccessful due to
repeated internal blockage of the atomizer. Subsequent test-
ing was carried out using Lot 2 with very little difficulty.
Lot 2 -
There was a visible difference between Lots 1 and 2 for
both the tree bark oil and the rice hull oil. In both cases
Lot 2 was more viscous and had a creamy texture, while Lot 1
was more watery in texture. Rice hull oil Lot 2 seemed to
have much less tendency toward atomizer blockage and thus was
fired successfully despite its higher viscosity.
99
-------
Table III-5. TYPICAL OIL COMBUSTION TEST DATA
Hawaiian
Hawaiian Tree Bark Oil Tree Bark Oil
o
o
Test Description Single-stage
Duration/ minutes
Pressures
Windbox , mmHg
Fuel, atmg
Steam at gun, atmg
Temperature
Fuel, °C
Steam at rotameter, °C
Windbox airT °C
Combustion chamber, exit, *C
Stack gas at inlet, °C
Flows
Atomizing steam, Kg/hr
Fuel, Kg/hr (from 02)
Total air, Kg/hr
Total air, percent of stoichio-
metric (from O2>
Burner air, percent of stoichio-
metric
Concentrations
Excess oxygen (dry) , percent
NO (dry) corrected to 3% 0_
Smoke density (visual) , %
SO-, ppm
SO,, ppm
CO , ppm
45
1.12
3.4
3.4
93
174
159
593
190
6.8
725
130
•
130
5.0
133
0
211
1.7
15
Two-stage
6
0.28
3.4
3.9
72
174
160
560
182
8.2
665
117
73
3.2
90
0
a
a
25'
Single-stage
45
1.03
4.9
4.8
89
174
161
643
199
12
75
723
140
140
6.1
441
0
18
0.9
30
Two- stage
4
0.28
3.6
3.7
93
174
164
626
188
8.2
69
664
140
75
6.1
165
0
a
a
6
Not measured
-------
In order to minimize thermal alteration of the rice hull
oil, the interim storage drum temperature was kept low, 54PC
(130°F) and atomizing steam boiler pressure was kept as low
as was practical, 6.4 atm C80 psi at the rotameter). The
steam mass flow rates varied from 10 to 20% of the fuel flow
rate. Steam pressure at the gun was approximately 4.7 atm
(55 psi) for most points and oil pressure was within 0.7 atm
CIO psi) of steam pressure.
The first runs using Lot 2 were made using combustion
air at 104°C (220°F) and 168°C (335°F), with the ignitor on.
Successful cold-furnace startups were also made using only
rice hull oil fuel. The first firing lasted 1 hour 27 minutes
before shutdown became necessary due to marked deterioration
of the spray.
In the next firing, clinker buildup on the diffuser
occurred. This was eliminated by moving the oil gun tip
1 cm (1/2 in.) farther into the furnace so that the tip
was flush with the downstream face of the diffuser. Because
of this change the rest of the rice hull oil runs are not
strictly comparable to the tree bark oil runs.
In all, the rice hull oil runs covered a range of
176,000 Kcal/hr to 503,000 Kcal/hr (700,000 to 2,000,000
Btu/hr) in heat input rate. Carbon monoxide emissions were
below 100 ppm at excess oxygen levels from 1% to 10%. There
appeared to be no unburned carbon emissions in this range.
Staged combustion at 3% excess oxygen resulted in nitric
oxide levels of 200 to 400 ppm. Use of the ignitor was not
necessary to sustain the flame.
Conclusions
1. In general it can be said that the pyrolytic oils
gave satisfactory performance in the combustion tests. It
must be stressed that the pyrolytic oil is not interchangeable
with distillate or residual oils in most boiler installations
without modifications to storage and pumping facilities. This
is due to several properties of the pyrolytic oils which were
mentioned previously: they are acidic; their properties are
fairly easily changed by overheating; their properties change
if the oils are allowed to evaporate at firing temperature
for appreciable lengths of time; and they can form deposits
on valves and heating surfaces if the system is not properly
designed. However, the system changes necessary to accomo-
date pyrolytic oils may not be more difficult than, say, the
changes needed to accomodate No. 6 heating oil in a system
designed for No. 2 oil.
101
-------
2. The pyrolytic oils appeared to be compatible (i.e.
could be blended) with most residual fuel oils representing
different geographical sources. However, the residual oils
did not completely purge the fuel system of pyrolytic oils.
- 3. The most likely initial problems encountered by
pyrolytic oil users would be in quality control, storage,
pumping, and atomizer blockage. Users should run thorough
pumping and atomizing tests before attempting to fire these
oils. Fuel specifications should restrict suspended solids,
gum formation, heating value changes, viscosity changes, and
other variations which are possible in any fuel oil, or which
may in some cases be peculiar to pyrolytic oils.
4. It appears possible that when firing pyrolytic fuels
the Federal Nox limitations on new units can be met by using
staged combustion, although perhaps without much margin.
5. The pyrolytic oils tested had such low sulfur
content that no problems due to sulfur oxide corrosion or
sulfur oxide emissions are anticipated from this oil.
6. Stack gas cleanup would be required to meet EPA
particulate regulations in the oils tested.
CHAR EVALUATION
Tests were made on several lots of Douglas fir bark char
to determine its suitability as charcoal briquettes, or as a
raw material for activated carbon. Char preparation was
discussed above.
Charcoal Briquettes
Two lots of 4.5 kg (10 Ibs) each of tree bark char were
sent to the Royal Oak Charcoal Company ( a Division of Georgia-
Pacific Corporation) having the following characteristics:
Lot 1 Lot 2
Volatile Matter, wt % 24.0 7.4
Ash, wt % 9.4 9.2
Briquettes were made on a laboratory press and the test
results are summarized below.
102
-------
Lot 1 -
The high volatile matter content of this lot caused the
char to be very spongy. The material required a large per-
centage of starch to bind it into briquettes, and the resul-
ting briquettes developed large cracks upon drying with a
generally bad appearance. In addition, the burning briquettes
smoked vigorously and emitted a disagreeable odor. Lot 1
char was thus judged not to be suitable for barbecue briquettes,
Lot 2 -
This char briquetted very well with substantially less
starch binder and dried firmly without cracking. The bri-
quettes ignited and burned well without any disagreeable
odor and produced a typical grayish-brown ash.
The Lot 2 char is considered to be suitable for manufac-
turing good quality barbecue briquettes. Royal Oak estimated
that the volatile matter content might be as high as 15% and
still have the desired properties to make a good quality
barbecue briquette.
The report from Royal Oak Charcoal Company is included
in the Appendix (Section IX) of this report.
Tree Bark Char Activation
Three samples of Douglas fir bark char were sent to the
St. Regis Paper Company for evaluation as a potential raw
material for activated carbon. The properties of these
samples are given on Table III-3, and the preparation of
the samples was described above.
Preliminary Evaluation -
The three samples were shipped for prescreening and the
most suitable for activation was chosen for further study.
Of these three materials, the best prospect for activated
carbon was that which, by a Garrett determination, contained
9.62% volatiles. Following St. Regis' procedures for volatile
determination, the data did not correspond to that supplied
by Garrett. However, duplication of Garrett1s technique by
St. Regis yielded very similar results. The ASTM procedure
followed by Garrett can yield low values for volatile matter
if these volatiles are driven off very slowly. Both Garrett
and St. Regis agreed that the method used by St. Regis gives
a value which is more meaningful in evaluating the chars'
potential for activation. The moisture, volatile matter, ash
content, reaction rate in an activating medium, nitrogen and
carbon dioxide surface areas were measured, and the results
are given in Table II1-6.
103
-------
Table III-6. ST. REGIS PRELIMINARY CHAR EVALUATION
Material
St. Regis Analysis
Reactivity
Surface Area
H
O
*»
(Garrett Volatile
Designation) %
3.87
9.62
'* H,O
3.0
3.0
a
C Voletiles
20
20
% Ash
28
28
Activation Rate i
Gas
12% CO,
Bal. Nj
12% CO.
Bal. N,
i 850°C
mg/mq-min.
.0075
.0072
S, Area
nr/g
20
20
CO. Area
V/g
90
180
23.03
3.0
25
30
12% CO,
Bal. N,
.0073
40
Determined with 850°C as a final temperature.
-------
The data indicated the three chars had similar volatile
matter, ash content, and reactivity. Inspection of the lab-
oratory data showed a larger amount of volatile matter given
off below 600°C for the 9.62% sample over the 3.87% sample,
although the total volatile matter was nearly the same at
850°C. As some preliminary pore development takes place
by volatilization of material below 600°C, this factor indi-
cated the 9.62% material might be more suitable for activated
carbon than the 3.87% material. The 23.03% volatile material
appeared under microscopic study to be in the form of only
slightly charred wood or bark. There would probably be little,
if any, advantage in working with this material in relation to
beginning with totally untreated raw tree bark. The determin-
ation of nitrogen and carbon dioxide surface areas finally
demonstrated that the 9.62% material was the most desirable.
The differences between the nitrogen and carbon dioxide areas
gives a qualitative indication of the micropore structure
which may2be utilized in further activation. This difference
was 160 m /gm for the 9.62% material, the highest of the group.
Upon being informed of St. Regis' decision to use the
so-called 9.62% material for further work, Garrett Research
provided two additional samples of this material, and these
were blended by St. Regis with the original to ensure unifor-
mity.
Compaction Studies -
Preliminary compaction work using coal tar pitch as a
binder and a pellet press at various temperatures, pressures,
and binder concentrations proved unsuccessful. Consultation
with and experimentation by K.G. Industries in Chicago
substantiated these findings. A liquid phase, binder, parti-
cularly lignosulfonate pitch, was suggested. Work at St.
Regis showed that the lignosulfonate pitch solution would
pelletize, cure to a strong pellet, and withstand carboniza-
tion without deterioration. However, the carbonized binder
is much more reactive than the base char and under mild acti-
vation conditions would burn off leaving the char powder
unsupported by a binder matrix.
The above results pointed to the necessity of develop-
ing a low reactivity binder system with high strength.
Toward this end, two coal tar pitch dispersions were made.
The first was a blend of octanol and coal tar pitch and the
other was an octanol-acetone solution with coal tar pitch
added. Pellets produced from octanol and pitch alone proved
to be the strongest after carbonization and these were acti-
vated under the same conditions as the protein bound pellets.
There was some deterioration of the pellets, but considerably
less than experienced when using the protein colloid or the
lignosulfonate.
105
-------
These studies are described in more detail in a report
to Garrett from the St. Regis Paper Company which is included
in the Appendix (Section IX).
Conclusions -
Nitrogen surface area measurements were made on the
octanol-pitch bound char and the lignosulfonate bound char,
both activated at 11% C02 at 750°C for, two hours. These
surface areas were 254 m /g and 336 m /g for the octanol-
pitch and lignosulfonate ijound material respectively. It
appears that the octanol may hinder pore development. This
is possible, as the octanol may be able to be trapped by the
char's pores and, if unable to volatilize rapidly out of the
system, will carbonize in place, plugging the pores and thus
hindering further surface area development.
Coal tar pitch and lignosulfonates are both inexpensive
materials, i.e., approximately 8 and 5 cents/lb respectively.
A protein colloid binder would be economically feasible,
commercially, only at very low binder levels, as the cost is
approximately 40 cents/lb. Organic dispersants such as
octanol are quite expensive and would probably require a
solvent recovery system if used in large quantities.
Even the unactivated char has a high ash content. Thus,
activation at even a moderate burn off would yield a carbon
with a very high ash. Acid washing of the char to lessen ash
content could be beneficial for activation. The fact that ash
content can affect binder efficiency and thus final product
strength also points toward the need for study in this area.
Through activation of palletized Garrett material, nitrogen
surface areas of 250 and 340 m /gm were obtained, based on
total material. On an ash free basis the area is 360 to 480
mvgm carbon.
No optimization to obtain the minimal binder level for
tree bark char was performed. Binder levels of 22% in general
are too high for commercialization. A typical commercial bin-
der concentration is 6% but attempts to pelletize at this
level were unsuccessful. If the Garrett char production sys-
tem were revamped to permit granular material to be formed,
the necessity for compaction could possibly be eliminated.
Unless a suitable binder system is developed, Garrett tree
bark char as supplied may be utilized only as a source of raw
material for powdered activated carbon. Its low density would
make activation as a powder difficult.
106
-------
Rice Hull Char Activation
Activation tests on rice hull char were conducted subse-
quent to the work on the tree bark product. While the rice
hull material was considerably easier to compact, the very
high ash content of approximately 60% is generally believed
to preclude its acceptance in its delivered state as a source
material for activation. St. Regis reports that due to this
high ash content, quantitative atomic absorption analyses on
this material were not conducted. However, the high surface
area of the available carbon in the raw rice hull char was
believed to be quite suitable for activation, if further pro-
cessing steps could be employed to lower the ash level. Some
acid leaching may look promising, but further work, outside
the scope of this program, would have to be undertaken before
this product could be considered as a suitable source of
activated charcoal.
The St. Regis report, covering details of their work on
rice hull char, is included in the Appendix (Section IV ) to
this report.
107
-------
SECTION VII
PHASE IV - ECONOMIC FEASIBILITY STUDY AND
PRELIMINARY PROCESS DESIGN FOR TREE BARK PYROLYSIS
INTRODUCTION
The final section of this report presents the results
of Phase IV - an engineering process design and economic
evaluation of the Garrett Flash Pyrolysis process for the
conversion of Douglas fir bark to synthetic fuel oil and
saleable char products. The design calculations and cost
estimates are prepared for two commercial size plants using
the pilot plant data described above in Phase II, and the
evaluation of the oil and char as marketable products
described in Phase III.
The two sizes considered are 300 and 1200 Tons per
day (272 and 1089 metric tons per day) pyrolysis plants
(on an oven dry basis). Products are a low sulfur barkoil
suitable for use as a utility boiler fuel or as a blend with
NO. 6 fuel oil in a utility boiler; and char which is mar-
ketable as a fuel, as charcoal briquettes or as a source of
inexpensive, powdered, throw-away activated carbon.
The economic evaluation indicates that a commercial
plant sized to process 300 Tons per day (dry) of tree
bark will essentially break even assuming the char is
sold as briquettes. A 1200 Ton per day plant, however,
shows a fair return on investment, even if only the oil
is sold.
DESIGN BASIS
Table IV-1 presents the design basis and operating
conditions for the plant.
108
-------
Table IV-1. DESIGN BASIS
Feed: Tree Bark
Primary Shredder
Air Classifier
Secondary Shredder - Drier
Bark Consumed as Drier Fuel
Pyrolysis Yields at 510 *C
(910»F)
= 45 wt. % Moisture
= 16 hrs/day - 7 days/wk
80 -8cm (-3")
= -8cm (-3") Overhead,
45% Moisture
+8cm (+3") Recycle to
primary shredder.
=24 hrs/day - 7 days/wk
80% -24 mesh, 5% Moisture
= (5 wt % of feed, maximum
(dry basis))
Wt % of dry
Pyrolysis feed
Oil
Wax
Gas
Water
Char
37
3
17
15
28
100
Gas
CO
CO^
(Mole %)
10.0
36.0
46.2
5.4
0.6
0.9
H. 0.9
100.0
HHV = 1.04 x 107 J/std m"
C2H6
C3H6
MW
(280 Btu/SCF)
32.5
109
-------
Table IV-1. DESIGN BASIS (cont.)
Pyrolysis (cont.)
Liquid Collection Train
Barkoil
Sp Gr @ 80°C(175°F) =
1.24
VISC, CP @ 80°C(175°F) =
200
HHV (DRY) = 5830 cal/g
(10,500 Btu/lb)
15 wt % H_0 in Barkoil
5 wt % Barkoil in Quench
10 wt % Wax in Quench
1 wt % Quench in Barkoil
Decanter Res. Time=30 min.
Liquid Storage
Cost Basis
Tank
Rundown
Off Spec
Product Storage
Quench Storage
Storage
Capacity
8 hours
1 day
3 days
Volume of
Quench
System
February, 1974 costs
20—year plant life
Site is clear and level
110
-------
An economic evaluation is given for two feed rates;
300 Tons per day and 1200 tons per day, on an oven dry
basis. These rates correspond to 545 tons per day and
2182 tons per day wet tree bark, respectively, having an
average moisture content of 45 weight percent water. The
plant is designed to process the bark at these rates,
seven days per week, 350 days per year. Primary shredders
are sized to operate 16 hours per day followed by suffi-
cient storage capacity to permit the rest of the plant
to operate 24 hours per day at the design rates.
Marketable products are barkoil and pyrolytic char.
The wax produced is of a low quality, and for purposes of
this feasibility study, it is assumed that the wax cannot
be economically recovered at the purity required by the
market. All the gas produced is consumed in the process
as transport recycle and fuel gas. In addition, a small
amount of the dry shredded bark may be needed as fuel in
the Secondary Shredder - Drier (less than 5%).
The pyrolysis reactor operates at about 1 atmg
(15 psig) and 510°C (950°F). The liquid collection train
decanter is designed to operate at 0.34 atmg (5 psig) and
80°C (175°F). Pressure balances on the fluidized char
beds have not been calculated, but the 0.68 atm (10 psi)
drop allowed in the design is sufficient to meet the re-
quirements in the pyrolysis area.
No special flexibility or redundancy is included in
the design. Spare equipment is provided only where dic-
tated by process considerations (e.g. spare pumps for the
quench and barkoil from the decanter in the liquid collec-
tion train to guard against upset conditions in the
pyrolysis area). Plant capacity is basically limited by
the capacity and operation of the primary and secondary
shredders.
The economic analysis is based on February 1974,
equipment costs. Operating labor and utilities costs
are also based on February 1974 levels. It is assumed
that the site is clear and level, and that the plant
would have a useful life of 20 years. The cost of the
land is not included.
PROCESS FLOW
Figure IV-1 presents the process flow diagram and
material balance for the plant.
Ill
-------
Tree bark from the receiving area is loaded onto a
system of conveyors which carry the material to the Primary
Shredder. Primary shredding reduces the size of the bark
to less than 8 cm (3 inches!. The shredded bark is conveyed
to interim storage providing 8 hours surge capacity. The
shredded bark, still at 45 weight percent moisture, is
essentially dust free.
Shredded bark from the Storage Bin is conveyed to the
Air Classifier which separates the bark less than 8 cm
(3 inches) into an overhead stream, passing through the
Air Classifier Cyclone to the Packaged Secondary Shredder-
Drier System from the bark greater than 8 cm (3 inches)
into a bottoms stream which is screened to remove inorgan-
ics and recycled to the Primary Shredder.
The Packaged Secondary Shredder-Drier System receives
the 8 cm (3 inch) and smaller bark from the Air Classifier.
This system introduces the raw material into a grinding
chamber along with hot inert gas. Particle size is reduced
to 80% less than 24 mesh and simultaneously dried to
5 weight percent moisture. The finely shredded, dried
bark is fed to a secondary surge system, which has 8 hours
capacity, so that upsets or maintenance requirements in
the shredding system will not interrupt operation in the
pyrolysis area.
A screw feeder conveys a measured amount of dried,
shredded bark to the Reactor, lifted by transport gas and
combined with hot 760°C (1400°F) circulating char. Flash
pyrolysis occurs in the reactor at 510°C (950°F) to con-
vert the bark into bark oil, char, wax, and pyrolysis
gases. Char is separated from the reaction mix in the
Reactor Cyclones and stored at 510°C (950°F) in the Char
Surge Hopper. The 510°C (950°F) char flows to the Char
Heater where a portion of the char is burned to raise the
temperature of the circulating char to 760°C (1400°F).
Combustion gases are separated from the char in the Char
Heater Cyclones and the heated char is stored at 760°C
(1400°F) in the Heated Char Seal Hopper. The 760°C
(1400°F) char flows from the hopper to join the shredded
bark from the Reactor Screw Feeder ahead of the Reactor.
The net char make is withdrawn at 510°C (950°F) from the
Char Surge Hopper and quenched with water to 121°C (250°F)
in the Char Make Hopper. The Hot Char Conveyor transports
the product char to char storage.
The 510°C (950°F) reaction mix from the Reactor
Cyclones (containing the net bark oil, wax, and pyrolysis
gas products) is contacted with circulating quench oil
in the Quench Venturi and cooled to 80°C (175°F). At these
112
-------
conditions the bark oil and wax is condensed and separated
from the remaining pyrolysis gases in the Decanter. The
Decanter also provides 30 minutes residence time for the
quench oil which is sufficient to give good separation of
the bark oil from the circulating quench oil. The net bark
oil product is withdrawn from the decanter and pumped to
product storage. Off spec and Rundown Tanks are also pro-
vided if further processing is required before the bark oil
is pumped to storage. A slip stream of the circulating quench
is fed to Wax Removal Facilities where the net wax make is
removed from the system.
A small portion of the condensed bark oil is entrained
as a fine mist with the pyrolysis gases from the Decanter.
This entrained bark oil is scrubbed from the gas stream by
washing the stream with quench oil through the Demister Ven-
turi. The quench and bark oil are separated from the gas in
the Demister Pot and returned to the Decanter to separate the
bark oil from the quench. The gas from the Demister Pot is
compressed to 1 atmg (15 psig) and recycled to the pyrolysis
area as transport and fuel gas.
PLOT PLAN
A typical plot layout for a 1200 Ton per day Tree Bark
Pyrolysis plant is shown on Figure IV-2. Total area require-
ment for a plant this size is about 22,000-24,000 m (5-1/2 to
6 acres) including all storage, shops, offices, parking,
etc. A plot layout for a 300 Ton per day plant is not
shown, but, using a similar arrangement, the area require-
ment would be about 16,000 m (4 acres).
EQUIPMENT DISCUSSION
Figure IV-1 presents the Process Flow Diagram and a
material balance for a feed rate of 1200 Tons per day (dry)
to the primary shredder. Tables IV-4 and IV-10 present a
detailed list of the equipment required. Sufficient
receiving and intermediate storage is provided to permit
the processing plant to operate at design rates, 24 hours
per day, 350 days per year.
Primary Shredder
The Primary Shredder reduces the size of the raw bark
to 80% less than 8 cm (3 inch) particles. This is a nec-
essary first step to permit effective separation of rocks,
earth, and other inorganics from the raw bark and to pro-
vide an optimum feed to the Secondary Shredder-Drier unit.
113
-------
FIGURE CT-2
o
o
<$>
•: -
*$ K.O. for
fjtf/nr/fs
sroire
I &/ c»AMj£/?^;sr, *.o(/rfX frcimi
H«. TB«l* T
PLOT PLAN
G;,I':
-------
Criteria for primary shredding equipment at a commer-
cial plant include:
1. Mill capacity, when producing a 8 cm (3 inch)
product, should be great enough so that only two units,
one operating and one spare, are needed.
2. On-size material should be removed from the mill
as quickly as possible to minimize the production of
inorganic fines which are difficult to separate from
the shredded bark.
3. No pretreatment or preselection of the feed should
be required.
Air Classifier
The Air Classifier serves to separate the oversize
bark greater than 8 cm (3 inch) for recycle to the
Primary Shredder and to also separate the denser inorganic
substances from the lighter 8 cm (3 inch) and less bark
which is to be further shredded, dried and pyrolyzed. The
rough separation is made according to density, size, and
aerodynamic characteristics of the shredded mixture.
A classifier may be built as a zig-zag, straight
vertical, straight horizontal, or other suitable unit.
Studies have shown that the zig-zag principle is the most
efficient in that it allows the separation of materials
with closely similar densities and with other properties
(e.g. size) that are almost identical. A typical unit con-
sists of a vertical zig-zag column with multiple stages and
upward air circulation. The feed is introduced part way up
the column. Lighter components are carried out the top of
the unit with the air stream, and are then removed by an
external cyclone. Heavy inorganic substances such as glass,
metals, stones, etc., which cannot be transported up through
the column at the chosen air velocity, fall out at the
bottom of the unit along with the heavier oversize chunks
of bark. Simple screening of the underflow from the air
classifier removes the inorganics from the larger bark
fraction.
It is important that the inorganic fraction in the
stream elutriated from the top of the classifier be as
small as possible. Otherwise, the hard, abrasive inerts
will cause substantial wear in downstream equipment such as
the secondary grinding mills, transfer lines, and cyclones.
The inorganics also increase the ash content of the pyroly-
tic oil and char, and may result in a loss of revenue,
owing to off-specification products.
115
-------
Secondary Shredder—Drier
The Secondary Shredder-Drier unit proposed is an appar-
atus in which the 8 cm (3 inch) and smaller bark from the air
classifier overhead is simultaneously shredded to 80% less
than 24 mesh particles and dried to 5 wt % moisture. High
yields of oil from the pyrolysis of organic solids are obtain-
ed by the extremely rapid transfer of heat to small particles,
which requires that the feed be shredded to a nominal minus
24 mesh.
Raw bark is fed to the grinding chamber along with hot,
inert gas. In the grinding chamber, rotating hammers reduce
the particle size to minus 24 mesh while the hot gas simul-
taneously removes moisture. Drying in the system is greatly
accelerated by the increased surface area as the particle size
is reduced.
The flow of hot gas selectively air-conveys particles
smaller than 24 mesh up and out of the grinding chamber. The
air suspended, ground product passes through an adjustable
velocity separator where a size classification is made.
This separator returns any oversize material to the grind-
ing chamber while product sized material is air conveyed
away. The minus 24 mesh bark is separated from the air
stream in a cyclone separator and the fines from the cyclone
are removed from the air stream by a fabric dust collector.
A small amount of the fines may need to be burned as fuel in
the drier to supplement the pyrolysis gases. The integrated
Secondary Shredder-Drier system is commercially proven and
readily available.
Pyrolysis
The pyrolysis process is based on the rapid heating of
organic materials, using a proprietary heat-exchange system.
Dry, finely shredded waste is delivered by a screw feeder
into the transport line, where it is picked up by recycled
gas and carried to the pyrolysis reactor. Rapid and com-
plete pyrolysis takes place at an elevated temperature (510°C,
950°F) as the feed travels upwards under turbulent flow con-
ditions. The only pressure required is that which must be
provided to move material through the system. The maximum
system pressure is generally less than 1 atmg (15 psig).
The technology involved in the char hoppers and circulation
system is well known through such application as oil refinery
fluid catalytic cracking (FCCl units, etc.
Char particle size and cyclone design are closely
related and critical to the successful operation of the
circulating char heat exchange system. This aspect of the
design and operation must be carefully considered and close-
ly controlled.
116
-------
Residence time in the reactor is relatively short - a
few seconds or less - to obtain the product yield structure
indicated. The reaction mix must be removed from the reaction
zone quickly before cracking into carbon, lighter fractions,
and gas can occur. The quenching step that follows must there-
fore be closely coupled to the reactor.
Liquid Collection and Storage
The gases leaving the reactor contain char, pyrolysis
gases, condensible vapors, and recycle transport gas. After
the char is removed by the hot cyclones in the reactor unit,
the reaction mix is quickly quenched in a venturi quench
system to about 80°C (175°F) to recover the pyrolysis oil
before thermal cracking can take place. The Quench Venturi
is a highly efficient contacting device which operates on
the principle of heat exchange by direct contact of the hot
reaction mix with finely dispersed droplets of cold quench.
The technology involved is well known and commercial units
are readily available. Pressure drop is low ( 13-15 cm of
water (5-6 inches WG)) but heat transfer is quite rapid.
The quenched reaction mix and quench fluid flow to the
Decanter vessel. This vessel is designed for high liquid
residence time and good separation between outlet vapor and
liquid streams. The bark oil is effectively separated from
the circulating quench fluid by gravity settling, but exper-
ience indicates that a significant amount of finely dispersed
bark oil remains entrained in the outlet vapor stream. This
vapor stream is therefore further cooled and scrubbed using
the Demister Venturi.
The Demister Venturi operates with a relatively high
pressure drop ( 75-100 cm of water (30-40 inches WG)) and
effectively scrubs the entrained oil droplets from the
vapor stream. Efficiencies are in the order of 98-99%
droplet recovery. Close control of the temperature in the
Demister indirectly provides control of the water content
of the product bark oil by controlling the amount of water
condensed from the vapor stream.
Vapor from the Demister is compressed to 1 atmg
(15 psig), using a liquid seal compressor such as a Nash
Pump. This type of unit is specified to provide reliability
and to minimize down time and maintenance time required by
the presence of corrosive and abrasive materials, e.g.
residual fine particles of char in the gas stream.
117
-------
ECONOMIC EVALUATION
The bases for the design and the equipment requirements
for the plant are described above. For the 1200-ton-per-day
plant, Tables IV-2 and IV-3 summarize the capital costs and
Table IV-4 presents a detailed breakdown of the capital cost
estimate. The operating cost estimate is shown in Table
IV-6 and is summarized on Table IV-5. A manpower summary
is presented in Table IV-8 and Figure IV-3 shows the pro-
posed project schedule.
Similarly, for the 3 00-'bon-per-day plant, the capital
costs breakdown and summaries are presented in Tables IV-10,
IV-8 and IV-9, respectively. Table IV-12 shows the detailed
operating cost estimate and the operating cost summary is
presented on Table 'IV-11. All costs are based on February,
1974 data.
Estimated product revenues are:
Product
Value
($/ton)
Revenue
($/ton of dry bark)
Bark Oil
Char - Fuel
42
- Charcoal
Briquettes
- Activated
Carbon
TOTALS
20
40
15.50 15.50
1.15
4.60
15.50
9.20
$16.25 $20.10 $24.70
The yield of oil obtained is approximately 0.28 m /
dry ton (1.75 bbl/ton) with a heating value averaging
5.8 Kcal/g (10,500 Btu/lb). The oil is a low-sulfur fuel
with some of the characteristics of No. 6 heating oil, the
standard fuel for utility plants.
The economics presented below for the 300-lon-per-day
and the 1200-ton-per-day commercial pyrolysis plants are
based on the data obtained from the continuous pilot plant
processing tee bark as discussed in Phase II of this report.
As shown below, the 300-ton-per-day plant will essentially
break even when selling heating oil and charcoal briquettes,
but the 1200-ton-per-day plant produces a fair return on
investment on the basis of bark oil sales alone.
118
-------
FIGURE BT-3
PROJECT SCHEDULE
TREE BARK PYROLYSIS FACILITY
NUMBER OF MONTHS
PROJECT COORDINATION
PROCESS ENGINEERING
DESIGN ENGINEERING
BID EVALUATION S PURCHASE
EQUIPMENT
SITE PREPARATION
CONSTRUCTION
START - UP
1
2
3
4
5
6
7
8
9
10
1 1
12
13
14
15
16
17
18
19
20
21
22
23
24
-------
Nominal; Feed Rate (dry basis)
Item 300 Ton/Pay 1200 Ton/Pay
Capital Cost
(Excluding land) $4,960,000 $13,310,000
Annual Operating Cost 1,717,000 2,857,000
Cost per Ton (including
Plant Amortization) 21.18 10.03
Revenue from Sale
of Oil and Char:
Barkoil Only 15.50 15.50
plus Char as Fuel 16.65 16.65
Char as Briquettes 20.10 20.10
Char as Activated
Carbon 24.70 24.70
At a 300 t^n/day scale, the operating cost is pro-
jected at $21.18 and revenues of $16.65 to 24.70 per dry
ton of bark, depending upon the value of the char produced.
At a 1200 ton/day scale, revenues remain the same, but
projected operating costs (including plant amortization
but excluding land costs) drop to $10.03 per Ton.
ENVIRONMENTAL CONSIDERATIONS
The pyrolysis process will greatly reduce the diffi-
culties of meeting not only the pollution control regula-
tions of today, but the stricter standards of the future.
An environmental balance for the 300 Ton per day plant is
shown in Figure IV-5 and a balance for the 1200 Ton per day
plant is shown in Figure IV-4. It can be seen that the
quantity of sterile solids going to landfill consists only
of rocks and dirt carried in with the bark. Much thought
has been given to minimizing air pollution. Because it is a
totally contained operation from which air is excluded,
The Flash Pyrolysis process has intrinsic advantages in
this respect and the gaseous emissions will be significantly
reduced. Air pollution problems are also minimized by the
care taken in handling air or gas streams from the classi-
fier and secondary shredder-drier.
120
-------
FIGURE Tg-4
SYSTEM ENVIRONMENTAL BALANCE
TREE BARK PYROLYSIS
1200 T/D (DRY)
VENT
AIR 2600 T/D
COMBUSTION GASES 2000 T/D
INPUT
TREEBARK
2I82T/D'5)45WT%H20
ELECTRICITY
6500 KW
PROCESS WATER
110 T/D
AIR
4000 T/D
OUTPUT
BARKOIL
445 T/D
CHAR
270 T/D
WAX
35 T/D
UNRECOVERED
SOLIDS TO LANDFILL
(INORGANICS ENTRAINED
IN TREEBARK FEED ONLY)
WASTE WATER
180 T/D
121
-------
FIGURE ET-5
SYSTEM ENVIRONMENTAL BALANCE
TREE BARK PYROLYSIS
300 T/D (DRY)
VENT
AIR 650 T/D
COMBUSTION GASES 500 T/D
INPUT
TREEBARK
545T/D© 45WT%H20
ELECTRICITY
1625 KW
PROCESS WATER
27 T/D
AIR
1000 T/D
OUTPUT
BARKOIL
110 T/D
CHAR
67 T/D
WAX
9 T/D
UNRECOVERED
SOLIDS TO LANDFILL
(INORGANICS ENTRAINED
IN TREEBARK FEED ONLY)
WASTE WATER
45 T/D
122
-------
The secondary shredder-drier is blanketed by an inert
gas stream operating on total recycle in order to eliminate
any fire hazard. All conveying gases are thus eventually
passed through the process heater where odors are destroyed
and carbonaceous particles are burned at a temperature of
982°C (1800°F). This single gaseous effluent from the
entire process is then cooled to 121°C (250°F) and filtered
to remove any traces of ash and inert contaminants. The
gases vented to atmosphere consist primarily of air and
products of combustion containing traces of SO and NO .
The projected environmental impace as a result of this
project is expected to be as follows:
Noise
The process utilizes a variety of machinery to trans-
port, shred, classify and pulverize the as-received mate-
rial. All of this equipment will be installed in such a
manner that the sound intensities at the plant boundary are
expected to be less than 55 decibels (dB).
Stack Gas Emissions
The total quantity of gaseous emmissions from the
proposed 1200 on per day plant is expected to be 2000
Tons per day. This entire stream is filtered and discharged
through a single stack. The stack gases will have the
following characteristics based upon 12.5 volume percent
C02:
SO 0.02 vol %
NO 60 ppm
x
3
Particulates 0.07 g/m (0.03 gr/scf)
Liquid Effluents
The quantity of water produced by pyrolysis at the
1200 Ton per day unit and discharged from the system as
liquid is 180 ons per day. This water contains only small
quantities of dissolved solids, but could have a COD con-
tent as high as 100,000 ppm. However, there are no water
disposal problems, provided secondary sewage treatment
facilities are available.
123
-------
Table IV-2. CAPITAL COST ESTIMATE - 1200 T/D
ro
Plant
Primary Shredding
Secondary Shredding
Pyrolysis
Liquid Collection & Storage
Offsite
Total Equipment Cost
Equipment Co8t,$
788,00.0
1,160,000
600,000
1,021,000
3,565,000
Installation Factor
1.8
2.0
3.0
2.5
Installed Cost
A/E (7 1/2%)
GR&D Cost (4 1/2%)
TOTAL
Contingency (20%)
Total Cost Estimate
Installed Cost,$
1,420,000
2,320,000
1,800,000
2,550.000
1,810,000
9,900,000
744,000
447,000
11,091,000
2,219,000
13,310,000
-------
Table IV-3. CAPITAL COSTS BY AREA - 1200 T/D
Primary Shredding
Installed Cost
GR&D + A/E (12%)
Contingency (20%)
Secondary Shredding
Installed Cost
GR&D + A/E (12%)
Contingency (20%)
Pyrolysis
Installed Cost
GR&D + A/E (12%)
Contingency (20%)
Liquid Collection
Installed Cost
GR&D + A/E (12%)
Contingency (20%)
Offsite
$1,420,000
171,000
$1,591,000
318,000
$1,909,000
2,320,000
279,000
$2,599,000
520,000
$3,119,000
1,800,000
216,000
$2,016,000
403,000
$2,419,000
2,550,000
307,000
$2,857,000
572,000
$3,429,000
$ 1,909,000
$ 3,119,000
$ 2,419,000
$ 3,429,000
Installed Cost
GR&D + A/E (12%)
Contingency (20%)
1,810,000
218,000
$2,028,000
406,000
$2,434,000
$ 2,434,000
$13,310,000
125
-------
Table IV-U. EQUIPMENT LIST - 1200 T/D
Item Service Estimated Cost
PRIMARY SHREDDING
Materials Handling
Raw Bark Conveyor $ 35,000
Pan Feed Conveyor 30,000
Primary Shredder 160,000
Discharge Conveyor 35>000
Air Classifier Recycle
Conveyor 25,000
Rotary Valve 2,000
Distribution Conveyor 10,000
Shredded Bark Conveyor 25,000
Air Classifier Feed
Conveyor 73,000
Blowers/Fans
Air Classifier Circula-
tion & Vent Fans 35,000
Separation Equipment
Air Classifier 80,000
Air Classifier Cyclone 2,000
Screen 25,000
Tanks/Bins
Landfill Bin 1,000
Shredded Storage Bin 250,000
(TOTAL) $788,000
SECONDARY SHREDDING
Materials Handling
Secondary Shredder- 1,030,000
Drier
Reactor Screw Feeder 50,000
Delumper 20,000
Tanks/Bins
Reactor Surge Storage 60,000
(TOTAL) $1,160,000
126
-------
Table IV-U. EQUIPMENT LIST - 1200 T/D (cont.)
PYROLYSIS SECTION
Item
Materials Handling
Reactors
Separation Equipment
Tanks/Bins
Miscellaneous
Service
Hot Char Conveyor
Reactor
Char Heater
Reactor Cyclones
Char Heated Cyclones
Reactor Char Surge
Hopper-
Char Heater Surge
Hopper
Char Make Hopper
Slide Valves
Flapper Valves
Pyrolysis Structure
TOTAL
Estimated Cost,S
10,000
15,000
15,000
35,000
35,000
70,000
30,000
25,000
80,000
15,000
170,000
600,000
127
-------
Table IV-4. EQUIPMENT LIST - 1200 T/D (cont.)
LIQUID COLLECTION AND STORAGE
Item
Heat Exchangers
Compressors
Service
Quench Cooler
Demister Cooler, Sealwater
Cooler & Recycle
Cooler
Fuel Gas Compressor
Estimated Cost,$
150,000
30,000
350,000
Separation Equipment
Tanks/Bins
Vessels
Miscellaneous
Rundown Tank Pump 2,200
Decanter Product Pump &
Spare 7,000
Decanter Quench Pump &
Spare 17,800
Demister Venturi Pump 4,000
Decanter Feed Pump 3,000
Product Oil Pump & Spare 7,000
Seal Water Pump 4,000
Rundown Centrifuge 75,000
Line Filters 6,000
Rundown Tanks 40,000
Off - Spec. Tank 24,000
Product Tanks 80,000
Quench Oil Storage 35,000
Demister Pot 14,000
Compressor KO Pot 14,000
Fuel Gas KO Pot 14,000
Decanter 68,000
Compressor Water Surge Pot 4,000
Quench Venturi 42,000
Demister Venturi 30,000
TOTAL 1,021,000
128
-------
Table IV-4. EQUIPMENT LIST - 1200 T/D (cont.)
OFFSITE
Item Service Estimated Cost,$
Equipment
Flare
Bag Filter House
Steam Generator
Liquid Nitroaer Storage
Power Substation
Instrument Air Compressor
Char Storage
Laboratory Equipment
Tools
Spare Parts
Plant Vehicles
Installed Cost @ 15 % of Onsite Plant Installed Cost 1,210,000
Buildings
Operations/ Maintenance/Lab 400,000
Administration 200,000
TOTAL 1,810,000
129
-------
Table IV-5- SUMMARY OF OPERATING COSTS - 1200 T/D
Costs by Area $/Year $/Ton Dry Feed
Primary Shredding 655,000 1.56
Secondary Shredding 535,000 1.27
Pyrolysis 443/000 1.05
Liquid Collection &
Storage 773,000 1.85
Offsites & Miscella-
neous 451,000 1.0-7
TOTAL 2,857,000 6.80
Costs by Category
Labor (63 total) 660,000 1.57
Supervision 165,00-0 0.39
Maintenance 705,000 1.68
Utilities 512,000 1.22
Taxes & Insurance 399,000 0.95.
Overhead G & A 349,000 0.83
Other Costs 67,000 Q.16
TOTAL 2,857,000 6.80
Capital Charge (20
years 88%) 1,358,000 3.23
4,215/000 10.03*
* $10.03/Ton (DRY) is equivalent to $5.51/Ton "As Received"
wet tree bark.
130
-------
Table IV-6. OPERATING COSTS ESTIMATE - 1200 T/D
1200 T/d for 350 operating days/year
Capital Cost Estimate
Labor Estimate
Operating 11
Maintenance 11
Supervision 2
Adnin. & Other 11
420,000 T/year
5 13,310,000
(All Shifts)
(Day Shift)
(All Shifts)
(Day Shift)
74 TOTAL
Primary Shredding
Capital Cost Estimate
Operating Costs
Variable
Electric power
$ 1,909,000
hr.
1300 KW x 16 ^- x
2^ aay
350 X KKH
Fixed
Labor
4 shift x $60,000
Supervision
(25% of .op. labor)
Maintenance
Labor (1.3% of capital)
Mat'l (4.0% of capital)
Operating Supplies
(0.5% of capital)
Overhead G&A
(35% of labor total)
Taxes & Insurance
(3% of capital)
TOTAL FIXED
TOTAL OPERATING COSTS
$/Year
73,000
$/Ton Dry Feed
0.18
240,000
60,000
25,000
76,000
10,000
114,000
57,000
582,000
655,000
0.57
0.14
0.06
0.18
0.02
0.28
0.14
1.38
1.56
131
-------
Table IV-6. OPERATING COST ESTIMATE - 1200 T/D (coat.)
Secondary Shredding
Capital Cost Estimate $3,119,000
Operating Costs
Variable $/Year $/Ton Dry Feed
Electric power
1700 KW x 24 hrs x
350 days x |~ 143,000 0.34
Fixed
Labor
1 shift x $60,000 60,000 0.14
Supervision
(25% of op. labor) 15,000 0.04
Maintenance
Labor (1.3% of capital) 41,000 0.10
Mat'l (4.0% of capital) 125,000 0.30
Operating Supplies
(0.5% of capital) 16,000 0.04
Overhead G&A
(35% of total labor) 41,000 0.10
Taxes & Insurance
(3% of capital) 94,000 0.22
TOTAL FIXED 392,000 0.93
TOTAL OPERATING COSTS 535,000 1.27
132
-------
Table IV-6. OPERATING COST ESTIMATE - 1200 T/D (cont.1
Pyrplysis
Capital Cost Estimate 52,419,000
Operating Costs
Variable $/Year $/Ton Dry Feed
Electric power
200 KW x 24 hrs x
350 days x ^77 17,000 0.04
T^»•*•
Water
8GPM 5 20^/1000 gal. 1,000 —
TOTAL VARIABLE 18,000 0.04
Fixed
Labor
2 shift x $60,000 120,000 0.29
Supervision
(25% of op. labor) 30,000 0.07
Maintenance
Labor (1.3% of capital) 31,000 0.07
Mat'l (4.0% of capital) 97,000 0.23
Operating Supplies
(0.5% of capital) 12,000 0.03
Overhead G&A
(35% of total labor) 63,000 0.15
Taxes & Insurance
(3% of capital) 72,000 0.17
TOTAL FIXED 425,000 1.01
TOTAL OPERATING COSTS 443,000 1.05
133
-------
Table IV- 6. OPERATING COST ESTIMATE - 1200 T/D (coat.)
Liquid Collection & Storage
Capital Cost Estimate § 3,429,000
Operating Costs
Variable $/Year $/Ton Dry Feed
Electric Power , e
3000 KW x 24 x 252,000 0.60
Water
10 GP« @ IMS-gal. 1'000
Total Variable 253,000 0.60
Fixed
Labor 2/shift x 60,000 120/000 0.26
Supervision 30,000 0.07
Maintenance
Labor (1.3 % of Capital) 45,000 0.11
Material (4 % of Capital) 137,000 0.33
Operating Supplies
(0.5 % of Capital) 17,000 0.05
Overhead G & A
(3.5 % of Total Labor) 68,000 0.16
Taxes & Insurance
(3 % of Capital) 103,000 0.25
Total Fixed 520,000 1.25
Total of Costs 773,000 1.85
134
-------
Table IV-6. OPERATING COST ESTIMATE - 1200 T/D (cont.)
Offsites & Miscellaneous
Capital Cost Estimate $ 2,434,000
Operating Costs
Variable $/Year $/Ton Dry Feed
Electric Power
300 KW x 24 x 350 x
1C/KWH 25,000 0.06
Fixed
Labor 2/shift x 60,000 120,000 0.29
Supervision 30,000 0.07
Maintenance Labor
(1.3 % Of Capital) 31,000 0.07
Maintenance Material
(4 % of Capital) 97,000 0.23
Operating Supplies
(0.5 % of Capital) 12,000 0.03
Overhead G & A
(35 % of Total Labor) 63,000 0.15
Taxes & Insurance
(3 % of Capital) 73,000 0.17
Total Fixed 426,000 1.01
Total of Costs 451,000 1.07
135
-------
Table IV-7 MANPOWER SUMMARY - 300 & 1200 T/D PLANTS
Primary Shredding
2 Shredder Operations
1 Storage & Bin Handling
1 Conveyors & Air Classifier
Secondary Shredding.
1 Shredder Operator
Pyrolysis
1 Reactor Control Room
1 Char Heater, Char Handling
Liquid Collection & Storage
1 Compressor House
1 Quench / Collection
Offsites
1 Storage - Product Shipping
1 Utilities
Supervision
1 Recieving & Feed Preparation
1 Utilities/ Offsites, Maintenance
Engineering
1 Mechanical Engineer (1 shift Only)
136
-------
Table IV-7. MANPOWER SUMMARY - 300 & 1200 T/D PLANTS (cont.)
Maintenance
1200 T/D Plant
5.3 % of Capital = (.053)(S 13,310,000)
* $ 705,000
Labor = 1.3 % 173,000 (11 men)
Materials = 4.0 % 532,000
$ 705,000
300 T/D Plant
5.3 % of Capital - (,053)($ 4,960,000)
$ 263,000
Labor = 1.3 % 65,000 (4 men)
Materials = 4.0 % «= 198,000
$ 263,000
137
-------
Table IV-8» CAPITAL COST ESTIMATE - 300 T/D
GO
00
Plant
Primary Shredding
Secondary Shredding
Pyrolysis
Liquid Collection & Storage
Offsite
Total Equipment Cost
Equipment Cost,$
355,000
380,000
200,000
380,000
1,260,000
Installation Factor
1.8
2.0
3.0
2.5
Installed Cost
A/E Cost (7 1/2%)
GR&O Cost (4 1/2 %)
Total
Contingency (20 %)
Total Cost Estimate
Installed Cost,$
640,000
760,000
600,000
950,000
740,000
3,690,000
277,000
166,000
4,133,000
827,000
4,960,000
-------
Table IV-9. CAPITAL COSTS BY AREA - 300T/D
Primary Shredding
Installed Cost $640,000
GR&D + A/E (12 %) 77,000
717,000
Contingency (20 %} 143,000
860,000 $ 860,000
Secondary Shredding
Installed Cost $760,000
GR&D + A/E (12 %) 91,000
851,000
Contingency (20 %) 170,000
1,021,000 $1,021,000
Pyrolysis
Installed Cost $600,000
GR&D + A/E'(12 %) 72,000
672,000
Contingency (20 %) 135,000
807,000 $ 807,000
_Liquid Collection
Installed Cost $950,000
GR&D + A/E (12 %} 114,000
1,064,000
Contingency (20 %) 213,000
1,277,000 $1,277,000
Offsite
Installed Cost $740,000
GR&D + A/E (12 %) 89,000
829,000
Contingency (20 %) 166,000
995,000 $ 995,000
$4,960,000
139
-------
Item
Table IV-10. EQUIPMENT LIST - 300 T/D
Service
Estimated Cost
PYROLYSIS SECTION
Materials Handling
Separation Equipment
Tanks/Bins
Miscellaneous
Hot Char Conveyor
Reactor Cyclones
Char Heater Cyclones
Reactor Char Surge Hopper
Char Heater Surge Hopper
Char Make Hopper
Slide Valves
Flapper Valves
$ 5,000
25,000
20,000
30,000
20,000
15,000
25,000
5,000
(TOTAL) $1^5,000
PRIMARY SHREDDING
Materials Handling
Blowers/Fans
Separation Equipment
Tanks/Bins
Rav Bark Conveyor $ 23,000
Pan Feed Conveyor 20,000
Primary Shredder 75,000
Discharge Conveyor 15,000
Air Classifier Recycle
Conveyor 11,000
Rotary Valve 2,000
Distribution Conveyor 8,000
Shredded Bark Conveyor 15,000
Air Classifier Feed
Conveyor 25,000
Air Classifier Circulation
& Vent Fans
Air Classifier
Air Classifier Cyclone
Screen
Landfill Bin
Shredded Storage Bin
13,000
32,000
2,000
13,000
1,000
100,000
(TOTAL) $355,000
140
-------
Table IV-ia EQUIPMENT LIST - 300 T/D (cant.)
Item
Service
Estimated Cost
Tanks/Bins
Heat Exchangers
Compressors
Pumps
SECONDARY SHREDDING
Materials Handling
Secondary Shredder - Drier $300,000
Reactor Screw Feeder 1*0,000
Delumper 15,000
Reactor Surge Storage 25,000
(TOTAL) $380,000
LIQUID COLLECTION & STORAGE
Quench Cooler $ 70,000
Demister Cooler, Sealwater
Cooler & Recycle
Cooler 8,000
Fuel Gas Compressor 100,000
Rundown Tank Pump 800
Decanter Product Pump &
Spare 1,600
Decanter Quench Pump &
Spare 5,^00
Demister Venturi Pump 1,200
Decanter Feed Pump 900
Product Oil Pump & Spare 1,100
Seal Water Pump 1,000
Rundown Centrifuge 30,000
Rundown Tanks 10,000
Off-Spec. Tank 8,000
Product Tanks 20,000
Quench Oil Storage 15,000
Demister Pot 6,000
Compressor KO Pot 7,000
Fuel Gas KO Pot 5,000
Decanter 60,000
Compressor Water Surge Pot 3,000
Separation Equipment
Tanks/Bins
Vessels
141
-------
Item
Table IV-10. EQUIPMENT LIST - 300 T/D (cont.)
Service Estimated Cost
Miscellaneous
LIQUID COLLECTION & STORAGE
Quench Venturi
Demister Venturi
OFFSITE
$ 15,000
11,000
(TOTAL) $380,000
Equipment
Flare
Bag Filter House
Steam Generator
Liquid Nitrogen Storage
Power Substation
Instrument Air Compressor
Char Storage
Laboratory Equipment
Tools
Spare Parts
Plant Vehicles
Installed Cost § 15 % of Onsite Plant Installed Cost
Building
Operations/Maintenance/Lab
Administration
200,000
100,000
(TOTAL) $7^0,000
142
-------
Table IV-11. SUMMARY OF OPERATING COSTS - 300 T/D
$,Year $,Ton Dry Feed
Costs By Area
Prinary Shredding 502,000 4.78
Secondary Shredding 232,000 2.21
Pyrolysis 280,000 2.67
Liquid Collection & Storage 383,000 3.65
Offsites & Miscellaneous 320,000 1.05
TOTAL 1,717,000 16.36
Costs By Category
Labor (63 total) 660,000 6.29
Supervision 165,000 1.57
Maintenance 262,000 2.50
Utilities 146,000 1.39
Taxes & Insurance 149,000 1.42
Overhead G&A 311,000 2.96
Other Costs 24,000 0.23
TOTAL 1,717,000 16.36
Capital Charge (20 years @ 8 %) 506,000 4.82
TOTAL COSTS 2,223,000 21.18
*$21.18/Ton (DRY) is equivalent to $11.65/Ton "As Received'
Wet tree bark
143
-------
Table IV-12. OPERATING COST ESTIMATE - 300 T/D
300 T/D for 350 Operating Days Per Year
Capital Cost Estir.ate 5 4,960,000
Labor Estir.ate
Operating 11
Maintenance 4
Supervision 2
105,000 T/Year
Administrative & Other 11
(All Shifts)
(Day Shift)
(All Shifts)
(Day Shift)
67 Total
Primary Shredding
Capital Cost Estimate $ 860,000
Operating Costs
Variable
Electric Power
325 KW x 16 hr/day x 350
x 1C/KWH
Fixed
$/Year $/Ton Dry Feed
18,000
Labor 4/Shift x $ 60,000
Supervision 25 % of Operating
Labor
Maintenance Labor (1.3 % of
Capital)
Maintenance Materials (4 %
of Capital)
Operating Supplies (0.5 % of
Capital)
Overhead G&A (35 % of Labor
Total)
Taxes & Insurance (3 % of
Capital)
Total Operating Costs 502,000
0.17
240,000
60,000
11,000
34,000
4,000
109,000
26,000
464,000
502,000
2.29
0.57
0.10
0.32
0.04
1.04
0.25
4.61
4.78
144
-------
Table IV-12. OPERATING COST ESTIMATE - 300 T/D (cont.)
Secondary^ Shredding
Capital Cost Estir.ata ? 1,022,000
Operating Costs
$/Year $/Ton Dry Feed
.Variable
Electric Power
425 KW x 24 hours x 350 davs
x 1C/KWH " 36,000 0.34
Fixed
Labor I/Shift x 60,000 60,000 0.57
Supervision (25 % of Operating
Labor) 15,000 0.14
Maintenance Labor (1.3 % of Capital) 13,000 0.12
Maintenance Material (4 % of
Capital) 41,000 0.39
Operating Supplies (0.5 % of Capital) 5,000 0.05
Overhead G & A (35 % of Labor) 31,OO.Q 0.30
Taxes & Insurance (3 % of Capital) 31,000 0.29
Total Fixed 196,000 1.87
Total Operating Costs 232,000 2.21
145
-------
Table IV-12. OPERATING COST ESTIMATE- 300 T/D (cent.)
Pyrolysis
Capital Cost Estimate S 806,000
Operating Costs
$/Year $/Ton Dry Feed
Variable
Electric Power
50 KW x 24 hours x 350 days
X 1C/KWH 4rOOO 0.04
Water
2 GPM @ 20C/100 gal. - _i
Total Variable 4,000 0.04
Fixed
Labor 2/Shift x 60,000 120,000 1.14
Supervision (25 % of Operating
Labor) 30,000 0.29
Maintenance Labor (1.3 % of
Capital) 10,000 0.10
Maintenance Materials (4 % of
Capital) 32,000 0.30
Operating Supplies (0.5 % Of
Capital) 4,000 0.04
Overhead G & A (35 % of Total
Labor) 56,000 0.53
Taxes & Insurance (3 % of Capital) 24,000 0.23
Total Fixed 276,000 2.63
Total Operating Costs 280,QOO 2.67
146
-------
Table IV-12. OPERATING COST ESTIMATE - 300 T/D (cont.)
Liquid Collection & Storage
Capital Cost Estimate S 1,277,000
Operating Costs
$/Year $/7on Dry Feed
Variable
Electric Power
750 KW x 24 x 350 x 1C/KWH 63,000 0.60
Water
3 GPM @ 20C/1000 gal. - _r
Total Variable 63,000 0.06
Fixed
Labor 2/Shift x 60,000 120,000 1.14
Supervision 30,000 0.29
Maintenance Labor (1.3 % of
Capital) 17,000 0.16
Maintenance Material (4 % of
Capital) 51,000 0.49
Operating Supplies (0.5 % of
Capital) 6,000 0.06
Overhead G & A (35 % of Total
Labor) 58,000 0.55
Taxes & Insurance £3 % of Capital) 38,000 0.36
Total Fixed 320,000 3.05
Total Operating Costs 383,000 3.65
147
-------
Table IV-12. OPERATING COST ESTIMATE. 300 T/D (cont.)
Offsite
Capital Cost Estimate $ 995,000
Operating Costs
$/Year $/Ton Dry Feed
Variable
Electric Power
300 KW x 24 x 3590 x 1C/KWH 25,000 0.24
Fixed
Labor 2/Shift x 60,000 120,000 1.14
Supervision 30,000 0.29
Maintenance Labor (1.3 % of
Capital) 13,000 0.12
Maintenance Material (4 % of
Capital) 40,000 0.38
Operating Supplies (0.5 % of Capital) 5,000 0.05
Overhead G'& A (35 % of Total
Labor) 57,000 0.54
Taxes & Insurance (3 % of Capital) 30,000 0.29
Total Fixed 295,000 2.81
Total Operating Costs 320,000 3.05
148
-------
SECTION VIII
REFERENCES
1. Badzioch, S. and P.G.W. Hawksley. Kinetics of Thermal
Decomposition of Pulverized Coal Particles, Ind Eng Chem
Process Des Develop. 9_ (4): 521-530, 1970.
2. Finney, C.S.f and Sotter, J.G., "Pyrolytic Oil From Tree
Bark: Its Production and Combustion Properties",
Presented to A.I. Ch. E. 77th Annual Meeting, Pittsburgh,
Pa., June 3, 1974.
149
-------
SECTION IX
APPENDICES
A. Royal Oak Charcoal Company evaluation of
bark char as barbecue briquettes 151
B. St. Regis Paper Company "Evaluation of
Douglas Fir Bark Char as a Raw Material
for Activated Carbon" 152
C. St. Regis Paper Company evaluation of
rice hull char as a raw material for
activated carbon 164
D. KVB, Inc. "Combustion Tests of Pyrolytic
Oils made from Solid Waste Materials" 168
150
-------
APPENDIX A
ROYAL OAK CHARCOAL COMPANY
PROCESS CONTROL REPORT
Copies To: George M. Kalian, Garrett research
C.S. Finney, .CarreCC ?-esearch
Katt Could, C.'P. Portland, Oregon
M.M. Powers, Royal Oak Charcoal
Number: 61
Location: Cookeville, Tn.
Date: 3-19-74
By: John Kllnk
Experimental briquets made with Pyrolytic Char fron Garrett
Research and Development Conpany
Carrett Research and Development Company, Inc.
1S55 Carrion 3d.
La Verne, California 91750
Phone: 714-593-7421
Two lots consisting of 10 Ibs. each of pyrolytic charcoal froa wood bark was
sent to Royal Oak Charcoal Company for evaluation as a raw nateriai for making
barbecue briquets. Briquets were cade on a laboratory press and th% following
observations and conclusions were obtained.
Lot No. 1
24.OT.
9.4X
66.61
Volatiles
Ash
Fixed Carbon
Ihe extremely high volatile caused the char to be very spongy. Material required
up to 12", starch.to bind into briquets. The briquets developed large cracks and
splits wben drying. Their appearance was bad.
The burning briquets smoked vigorously and had a disagreeable odor. The ash formed
by the burning briquets was reddish-brown in color. This hi^h volatile char is not
suitable for barbecue briquets. Briquets were sent to Carrett Research for their
evaluation.
Lot No. 2
7.42
9.21
83.4X
Volatile
Ash
Fixed Carbon
This material briquettod very well with 7Z starch binder. The briquets were firn
and dried without developin£ cracks or splitting. The briquets ignited and burned
veil without any disagreeable odor, the ash is a grayish-brovn which is normal for
barbecue briquets. This char is suitable for nanufacturins good quality barbecue
briquets. Samples of these briquets have been sent to Carrect Research. Perhaps,
the volatilcs could be Increased up to 152 ind scill have the desired properties
to naie a good charcoal briquet for barbecuing.
u
John Klink
Process Control Kana-ccr
151
-------
APPENDIX B
SPECIAL REPORT FOR GARRETT RESEARCH AND DEVELOPMENT CO., INC.
Evaluation of Douglas Fir Bark Char as a Raw Material for Activated Carbon
SUMMARY
St. Regis Technical Center
West Nyack, New York
Three chars were supplied to St. Regis Paper Company by Garrett Research
and Development Company, Inc. Through a preliminary evaluation, the
best prospect of the three, as a raw material for activated carbon,
was chosen for further study. This material was Douglas fir bark
char. The reactivity in an activating gas, moisture, volatiles and
ash contents of the three samples were very close. The Douglas fir
sample had a significant difference between its measured nitrogen
carbon dioxide surface areas and thus, probably has the largest micropore
structure suitable for further activation.
Bulk density for the Gariett Douglas fir char was 0.271 g/cc untapped.
Analyses indicated that the char contained impurities of essentially
calcium, magnesium, silicon, aluminum and potassium.
Two additional batches of the material were supplied by Garrett Research
and then blended. The chars being too fine for use in the equipment
available necessitated compaction. Preliminary compaction attempts
using coal tar pitch in a hot pellet press proved unsuccessful. Further
work was performed following suggestions made by K. G. Industries.
Lignosulfonate, protein colloid, coal tar pitch dispersed in octanol and
coal tar pitch dispersed in octanol with acetone added to dissolve some
additional pitch were used as binders. These materials were subjected
to an activation step and most deteriorated significantly. Surface
areas on the most promising materials indicated the pitch octanol dis-
persion though the best binder may have hindered pore development. The
char has no coking properties, thus thermally induced agglomeration is
not feasible.
An attempt was made to utilize the char in its supplied state but proved
unsuccessful. Further work can be performed only if granular char is
made available or a compaction procedure suitable for the char is devel-
oped.
October, 1973
V. Del Bagno
Approved:
Author:
152
-------
CONTENTS
INTRODUCTION.
PRELIMINARY MATERIAL EVALUATION .
CHAR CATEGORIZATION
COMPACTION AND ACTIVATION STUDIES
CONCLUSIONS
154
155
157
160
163
TABLES
Table I Garrett Char Preliminary Evaluation
Table II Sieve Analysis of Douglas Fir Bark Char
Table III Reactivity of Garrett Douglas Fir Bark Char in
Activation Media
Table IV Metal Analysis of Garrett Char
Table V Evaluation of Lignosulfonate Pitch Strength
During Activation
Table VI Lignosulfonate Binder Evaluation
156
158
158
159
161
162
153
-------
Widely varied carbonaceous materials have historically been used for
commercial production of activated carbons. These activated materials
are characterized by having high void volumes and high surface areas as
a consequence of which ttey adsorb large amounts of gases and vapors as
compared to their volume. The same property also enables them to adsorb
liquids and solids from solution. Depending on the porous structure and
surface functional groups, activated carbons are classified into gas phase
carbons, decolorizing caibons, water treatment carbons, etc. Today, most
commercial carbons are produced from bituminous or lignite coals, petro-
leum sludges, coconut and nut shells, and paper mill wastes. By a suit-
able choice of production technique, an activated carbon can be made
from any of these raw mal.erials. The quality of the product, hoxvever,
has been found to be greatly dependent upon the nature of the raw
material when activated by currently used commercial processes.
Garrett Research and Development Co. Inc., developed a process vhich
yields chars possibly suitable for the production of activated carbons.
St. Regis evaluated a sample supplied by Garrett as a possible raw
material for activated carbon.
This report contains the data developed by St. Regis. Four sections are
contained herein. The first deals with a preliminary screening of three
materials. Three samples were supplied by Garrett so that the most
promising could be chosen for further expended work. The second section
discusses the categorization or the char chosen as, potentially, the best
raw material for activated carbon. The chemical reactivity in various
activating gases was measured. An X-ray scan and metal analysis and
sizing of the char were performed. Compaction and preliminary activation
studies were made and the developed information is presented in section
three. Conclusion and suggestions for further work are listed in
section four.
154
-------
PRELIMINARY MATERIAL EVALUATION
On May 3, 1973, St. Regis Paper Company contracted with Garrett Research
and Development Company, Inc. to provide support to Garrett1s work in
the area of pyrolysis of industrial solid wastes under 11PA Grant S-801202.
The support was specifically to be in the evaluation, as a potential raw
material for activated carbon, of char produced by Garrt-tt.
Three samples of raw material manufactured from Douglas fir bark were
shipped for prescreening and the most suitable as an activated carbon
was chosen for further study. Of these three materials the best prospect
as a raw material for activated carbon was that which, by a Garrett deter-
mination, contained 9.62% volatiles. Following St. Regis' procedures for
volatile determination, the data did not correspond to chat supplied by
Garrett. However, duplication of Garrett1s technique yielded very similar
results. The procedure followed by Garrett (ASTil) can yield low values
for volatile matter if these volatiles are driven off very slowly. The
method used by St. Regis is thought to give a value which is more meaning-
ful in evaluating the chars potential for activation.
The moisture, volatile matter, ash content, reaction rate in an activating
medium, nitrogen and carbon dioxide surface areas were measured and the
results are given in Table 'I.
The data indicated the three chars had similar volatile matter, ash content,
and reactivity. Inspection of the labors Lory uaLa showed a larger amount
of volatile mattsr given off below 600°C for the 9.62% sample over the
3.87% sample, although the total volatile matter was nearly the same at
850°C. As some preliminary pore development takes place by volatilization
of material below 600°C, this factor indicated the 9.62% material might
be more suitable for activated carbon than the 3.87%. The 23.03% volatile
material appeared under microscopic study to be in the form of only slightly
charred wood or bark. There would probably be minor, if any, advantage
working with this over beginning with totally untreated raw material. The
determination of nitrogen and carbon dioxide surface areas finally demon-
strated that the 9.62% material was the most desirable. The difference
between the nitrogen and carbon dioxide areas gives a qualitative indi-
cation of the micropore structure which may be utilized in ftirther acti-
vation. This difference was 160 m^/gm for the 9.62% material - the
highest of the group.
Upon being informed of the decision to use 9.62% material for further
work, Garrett Research provided two additional samples of this material,
and the three samples were blended to ensure uniformity and supply a
large batch for compaction.
155
-------
TABLE I
GARRETT CHAR PRELIMINARY EVALUATION
Ui
Macerial
(Garretc Volatile
Designation)
3.87
9.62
23.03
Analysis
7. tt>0 % Vo la tiles1 1. Ash
Reactivity
Activation Rate 0 850°C
Gas mg/mR-min._
Surface Area
N2 Area C02 Area
nr/g III-/K
3.0
3.0
3.0
20
20
25
28
28
30
127. C02
Bal. N2
127. C02
127. CO,
.0075
.0072
.0073
20
20
0
90
180
40
Bal.
Determined with 850°C. as a final temperature.
-------
CHAR CATEGORIZATION
Sizing of the bark char supplied showed the material to be predominantly
less than 120 mesh. Results of the sieve analysis are shown in Table II.
Bulk density was found to be 0.271 g/cc untapped and 0.298 g/cc tapped.
In addition to the reactivity analysis performed through preliminary
screening of the material, additional determinations were made and this
data is listed in Table III.
Following ASTM procedure D-388, Garrett material was found not to have
any agglomerating character. An X-ray scan and atomic adsorption metal
analysis of the char was performed and these results are shoxm in
Table IV. High concentrations of calcium, magnesium and potassium are
expected from barks. Phosphorus (P) level which was not measured should
be close to magnesium. Traces of manganese, copper and zinc are also ex-
pected. The amounts of aluminum and iron along with some of the trace
materials like nickel, chromium and strontium are unusual. These
materials may represent some contamination from process equipment or
chemicals used in the char production process.
157
-------
TABLE II
SIEVE ANALYSIS OF DOUGLAS FIR BARK CHAR
SIEVE SIZE % OF TOTAL VEIGHT REMAINING ON SCREEN
20 negligible
40 1.90
60 8.92
120 14.12
200 20.82
smaller than 200 54.24
TABLE III
REACTIVITY OF CARRETT DOUGLAS FIR BARK CllAR IN ACTIVATION MEDIA
Temperature
<°c)
750
750
850
850
950
Gas Composition
(Balance is N2)
12% C02
53% C02
12% C02
53% C02
12% C02
Reactivity
(mg/mg - tnin.)
.0018
.0027
.0073
.0090
.0125
158
-------
TABLE IV
Metal
Al
Si
P
S
Cl
K
Ca
Ti
Cr
Mn
Fe
Ni
Cu
Zn
Sr
Ba
Na
Li
Mg
METAL ANALYSIS OF
Detected
Scan
Raw Char
X
X
X
X
X
X
X
X
Not measured
X
X
X
X
X
X
X
Not measured
Not measured
Not measured
GARRETT CHAR
by X -ray
M
Ash
X
X
X
X
Not Detected
X
X
X
Not measured
X
X
X
X
X
X
X
Not measured
Not measured
Not measured
Atomic Adsorption
Analysis
y g/g (as is basis)
9240
3670
Not measured
<300
Not measured
4030
17200
340
37
913
8760
80
42
150
800
<200
629
4
3280
159
-------
COMPACTION A'H) ACTIVATION STUDIES
The Garrctt material supplied was fine and of low density. An attempt
was made to utilize the fine powder. Gas dispersion proved difficult
and stabilization of proper flow patterns was impossible x
-------
Nitrogen surface area measurements were made on the octanol-pitch bound
char and the lignosulfonate bound char, both activated at 117, C09 at 7503C
for two hours. These surface areas were 254 m2/g and 336 m2/g for the
octanol-pitch and iignosulfonato bound material respectively. It appears
that the octanol may hinder pore development. This is possible as the
octanol may be able to be trapped by the char's pores and if unable to
volatize rapidly out of the system will carbonize in place plugging the
pores and thus hindering further surface area development.
Coal tar pitch and lignosulfonates are both inexpensive materials, i.e.,
approximately 8 and 5 cents/'Ib. respectively. A protein colloid binder
would be economically feasible, comnercially, only at very low binder
levels as the cost is approximately 40 cents/lb. Organic dispersants such
as octanol are quite expensive and would probably requite a solvent
recovery system if used in large quantities.
TABLE V
EVALUATION OF LIGNOSULFONATE PITCH STRENGTH DURING ACTIVATION
ACTIVATION CONDITIONS
Temperature Time Activating Gas
850°C 3 hrs. Flue Gas
750°C 2 hrs. 11% C02, 897. N2
750°C
1 hr. 77, C02, 937. N2
PRODUCT -CONDITION
Ashed
Predominantly powder;
the few large particles
remaining were quite weak.
Similar to original
material.
161
-------
TABLE VI
LIGNOSULFONATE BINDER EVALUATION
Binder
Water only
Lignosulfonate
Solution
67^ Binder
227,1 Binder
227. Binder
227o Binder
Carbon
Temperature
60° C
140° C
140° C
140° C
25°G
Cure
Technicjue
None
None
100°C Preliminary
Cure
177°C Final Cure
None
100°C Preliminary
Cure
177°C Final Cure
After Carbonization
No Pelletization
Obtainable
No Pelletization
Obtainable
Weak Pellet
Very weak pellet
Strong pellet
Percentage based on char only. Water in binder solution not included.
162
-------
CONCLUSIONS
1. Unless a suitable binder system is developed, Garrettt char as supplied
may be utilized only as a source of raw material for powdered acti-
vated carbon. Its very low density xvould make activation as a
powder very difficult at best.
2. The ash content of the char is high. Thus activation at even a
moderate burn off would yield a carbon with a very high ash. Acid
washing of the char to lessen ash content would be beneficial for
activation. The fact that ash content can affect binder efficiency
and thus final product strength also points toward the need for study
in this area.
3. Through activation of pelletized Garrett material nitrogen surface
areas of 250 and 340 m-/gm were obtained. This value is per gram of
total material. On an ash free basis this area is 360 to 480 m.-/gm
carbon.
4. No optimization to obtain the minimum binder level vas performed.
Binder levels of 22% in general are too high for commercialization.
A typical commercial binder concentration is 6% but attempts to
palletize at this level were unsuccessful.
5. If the Garrett char production system was ravampcd to permit granular
material to be formed, the necessity for compaction could possibly
be eliminated.
163
-------
APPENDIX C
PAPER COMPANY West Nyack Road. West Nyack, NewYork 10994
December 11, 1973
Mr. C. S. Finney
Assistant Manager
Solid .Pollution Program
Garrett Research and Development Company, Inc.
1855 Carrion Road
La Verna, California 91750
Dear Mr. Finney:
On October 8, 1973, a sample of rice hull char was received. In
response to your request to R. L. Miller, I performed a prelim-
inary evaluation to determine if the char has potential for use
as a raw material for activated carbon. The following are the
results of the work accomplished.
Char Categorization
The physical and chemical properties of the char are shown in
Table I. The volatile content is low and the ash very high.
This ash level severely limits the chars' usefulness as a raw
material for activated carbon. There is no agglomerating char-
acter inherent in the material and the nitrogen surface area
though low, when taken on a per gram of carbon basis, is high
for a raw char, namely 150 m^/gm carbon. The bulk density of
the rice hulls was greater than the Douglas Fir char but still
rather low. Sieving of the material showed it to be predomin-
antly less than 120 mesh. However, there was a more significant
fraction above 40 mesh than was present with Douglas Fir char.
This data is shown in Table II.
The reactivity of the rice hulls in activating gases is shown in
Table III. This material was less reactive than the bark char.
The activation rates are not normalized to an equal carbon basis.
If normalization is performed, the rates for the rice hulls are
still significantly lower. The lower reactivity may not be due
to a difference in the carbon substrate but, rather to a blockage
of the activating gas as a result of the high ash content. Two
suitable materials for activated carbon are coal and petroleum
cokes. They commonly have activation rates of .0011 and .0050
mg/mg-min respectively in an atmosphere of 53% C0_ and at 850°C.
164
-------
Page 2
12-11-73
Under microscopic study, the rice hull char appeared to contain
three distinct fractions, namely:
1. Semi-carbonized material consisting of large, flat
thin particles.
2. Fully-carbonized material, consisting of short
fibers of intermediate particle size.
3. Very small carbonized particles of spherical
shape. The particles of this fraction were very friable.
These three fractions were separated in an attempt to isolate a low
ash, volatile containing, fraction suitable for activation. The
analysis of these are shown on Table IV. The 20 x 100 mesh cut had
the highest volatile content, highest reactivity and lowest ash level.
Unfortunately, the difference in ash contents were not as great as
hoped; the ash concentration still being too high to make acid
washing feasible even for the most promising cut.
At this point, it was apparent that the rice hull char as supplied
was not a suitable raw material for activated carbon. Thus, quan-
titative atomic absorption analysis of the char and attempts to
compact and activate the carbon were not made.
Conclusion
The high surface area of the carbon in the raw char from rice hulls
and the suitable reactivity indicate it may be useful as a raw
material for activated carbon if processing steps can be employed
during char generation to minimize the ash level in the final
product. However, in its present form, it is not suitable for this
end use.
Vincent Del Bagno
VDB:dk:lp
Attachments
165
-------
TABLE I
Rice Hull Char Properties
Moisture: 2.07.
Volatile Content : 7.67.
Ash: 737.
Fe,
P, Si, Ni, Cu, Zn
2
Impurities : Fe, Mn, Ti, Ca, K, Cl, S,
Agglomerating
Character: None
Nitrogen Surface,
Area (m^/gm)
Bulk Density
Tapped: 0.50gm/cc
Untapped: 0.34gm/cc
1. Determined with 850°C as a final temperature.
2. Given by qualitative X-ray scan. This technique cannot detect
Na, Cr, Mg, and Li. Elements obtained were identical for both
the raw char and ashed material.
166
-------
TABLE II
Sieve Analysis of Rice Hull Char
Sieve Size % of Total Weight Remaining On Screen
20 0
40 8.0
60 5.3
120 12.2
200 22.8
Less than 200 51.7
TABLE III
Reactivity of Garrett Rice Hull Char in Activation Media
Conditions
Temperature Gas Composition Reactivity
°C (Balance Is Nitrogen) (mg/mg-min)
750 12% CO, .0002
750 53% C0« .0010
850 12% CO, .0009
850 53% CO, .0013
950 12% COg .0041
TABLE IV
Properties of Three Rice HullL_C_h_ar.
Sieve Fractions
Reactivity
Volatile @ 850°C & 53% C02
Sieve Fraction Content (mg/mg-min.) Ash
20 x 100 12.4% .0024 55%
100 x 200 7.2% .0016 69%
Less than 200 6.1% .0015 75%
167
-------
APPENDIX D
Copies of Appendix D entitled "Combustion Tests of Pyro-
lytic Oils made from Solid Waste Materials" by KVB, Tustin,
California, can be obtained upon request from Industrial Envi-
ronmental Research Laboratory, Cincinnati, Ohio 45268.
168
-------
PAGE NOT
AVAILABLE
DIGITALLY
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
REPORT NO.
EPA-600/2-77-091
3. RECIPIENT'S ACCESSIOI*NO.
4. TITLE AND SUBTITLE
PYROLYSIS OF INDUSTRIAL WASTES FOR OIL AND
ACTIVATED CARBON RECOVERY
5. REPORT DATE
_May .1.917, issuing date_
6. PE~RFbRMir3G ORGANIZATION CODE
. AUTHOR(S)
F. B. Boucher, E. W. Knell, G. T. Preston, G. M. Mallan
8. PERFORMING ORGANIZATION REPORT NO.
. PERFORMING ORGANIZATION NAME AND ADDRESS
Occidental Research Corporation (Formerly Garrett
Research and Development Company, Inc.) 1855 Carrion
Road, La Verne, California 91750
10. PROGRAM ELEMENT NO.
1BB610
11. CONTRACT/GRANT NO.
Grant S - 801202
12. SPONSORING AGENCY NAME AND ADDRESS
Industrial Environmental Research Laboratory - Gin., OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final Report
14. SPONSORING AGENCY CODE
EPA/600/12
IS. SUPPLEMENTARY NOTES
is. ABSTRACT
Occidental Research Corporation (formerly Garrett Research and Develop-
ment Company, Inc.) has developed a new Flash Pyrolysis process which can produce up
to two barrels of synthetic fuel oil from a ton of dry cellulosic solids. This re-
port presents the results of a four-phase laboratory, pilot plant, product evaluation
and engineering evaluation program to study the pyrolytic conversion of Douglas fir
bark, rice hulls, grass straw and animal feedlot waste to synthetic fuel oil and char
With the use of an existing U ton/day pilot plant, good quality products were obtain-
ed from all feedstocks except animal waste. A wax by-product was obtained from the
pyrolysis of fir bark and grass straw. Excellent pilot plant material balances were
obtained for oil production runs on Douglas fir bark and rice hulls, and these were
satisfactorily combusted in a standard test boiler. Similar yields were obtained
from semi-quantitative runs using grass straw. The pyrolytic chars from tree bark
and rice hulls were evaluated as a source of activated carbon, and tree bark char was
satisfactorily compressed to produce excellent quality charcoal briquettes.
The economic evaluation shows that a 1200 dry ton/day tree bark conversion
plant could "be "built and operated with a profit of about $10/ton of dry bark. The
breakeven point for this process to produce synthetic fuel oil and char for briquettes
appears to be 300 dry tons of bark/day.
17.
KEY WORDS AND DOCUMENT ANALYSIS
a.
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COS AT I Field/Group
Wastes, bark, straw, rice plants,
wood wastes, by-products, pilot
plants, disposal, cost estimates,
activated carbon
fuel oil generation
solid waste pyrolysi
by-product evaluatio
feasibility analysis
energy conversion
13 B
18. DISTRIBUTION STATEMENT
Release Unlimited
19. SECURITY CLASS (ThisReport}
Unclassified
21. NO. OF PAGES
181
20. SECURITY CLASS (Thispage)
Unclassified
22. PRICE
EPA Form 2220-1 (9-73)
171 £-U.S. GOVERNMENT PRINTING OFFICE: 1977-757-056M16 Region No. 5-11
------- |