EPA-600/2-78-026
February 1978
Environmental Protection Technology Series
CONVERSION OF CATTLE FEEDLOT MANURE
TO ETHYLENE AND AMMONIA SYNTHESIS GAS
Robert S. Kerr Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Ada, Oklahoma 74820
-------
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-600/2-78-026
February 1978
CONVERSION OF CATTLE FEEDLOT MANURE
TO ETHYLENE AND AMMONIA SYNTHESIS GAS
by
William J. Huffman
James E. Halligan
Roger L. Peterson
Texas Tech University
Lubbock, Texas 79409
Grant No. S-802934
Project Officer
R. Douglas Kreis
Source Management Branch
Robert S. Kerr Environmental Research Laboratory
Ada, Oklahoma 74820
ROBERT S. KERR ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ADA, OKLAHOMA 74820
-------
DISCLAIMER
This report has been reviewed by the Robert S. Kerr Environmental
Research Laboratory, 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. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
ii
-------
FOREWORD
The Environmental Protection Agency was established to coordinate
administration of the major Federal programs designed to protect the
quality of our environment.
An important part of the Agency's endeavors to fulfill its mission
involves the search for information about environmental problems, manage-
ment techniques and new technologies through which optimum use of the
nation's land and water resources can be assured. The primary and ulti-
mate goal of these efforts is to protect the nation from the scourge of
existing and potential pollution from all sources.
EPA's Office of Research and Development conducts this search through
a nationwide network of research facilities.
As one of these facilities, the Robert S. Kerr Environmental Research
Laboratory is responsible for the management of programs to: (a) investi-
gate the nature, transport, fate and management of pollutants in ground-
water; (b) develop and demonstrate methods for treating wastewaters with
soil and other natural systems; (c) develop and demonstrate pollution con-
trol technologies for irrigation return flows; (d) develop and demonstrate
pollution control technologies for animal production wastes; (e) develop
and demonstrate technologies to prevent, control or abate pollution from
the petroleum refining and petrochemical industries; and (f) develop and
demonstrate technologies to manage pollution resulting from combinations
of industrial wastewaters or industrial/municipal wastewaters.
This report is a contribution to the Agency's overall effort in ful-
filling its mission to improve and protect the nation's environment for
the benefit of the American public.
William C. Galegar, Director
Robert S. Kerr Environmental
Research Laboratory
m
-------
ABSTRACT
An investigation was undertaken to scale-up a partial oxidation
reactor to evaluate the potential for producing an ammonia synthesis gas
from cattle feedlot residue. Because conventional ammonia production
involves a commercial process which produces an intermediate synthesis
gas from natural gas, air, and water, the only part of the total process
that needed to be investigated was the reactor system. The synthesis
gas from the reactor can be further processed to produce ammonia using
existing, commercial technology.
During this phase of the study, a nominal 2.5 meter (m) long, falling-
bed reactor was designed, constructed, and operated along with supporting
feed and recovery equipment; manure feed rates to the reactor ranged from
7.7 to 23.6 kilograms per hour (kg/hr). The results show that significant
yields of an ammonia synthesis gas and ethylene can be produced in the
same reactor when the system is fed with a mixture of air, steam, and manure.
This study and earlier bench-scale experiments (EPA-660/2-74-090) show
that the product gases are compatible with both ethylene and ammonia syn-
thesis gas production. The maximum production rates were not encountered
during the investigation; however, the experimental data and earlier
results show that the ammonia equivalent of manure is in excess of one-half
metric ton of ammonia per metric ton of dry, ash-free manure. The production
of ethylene ranged from 21 to 70 grams per kilogram (g/kg) of dry, ash-free
manure; production rates up to 80-90 g/kg may be possible pending further
studies on reinjection of liquid products. The process has been developed
assuming that manure contains about 15 percent moisture and 25 percent ash.
The production of ammonia synthesis gas from cattle feedlot manure is
recommended for scale-up to a pilot demonstration unit using the technology
developed here and that used for general solid waste processing.
Additional studies are needed before a complete design can be speci-
fied for the production of ethylene and other petrochemical intermediates.
The development of this technology could have a significant impact on the
feedstock supply for the petrochemical industry and should, therefore, be
vigorously pursued at the current scale and during the operation of the
recommended pilot demonstration unit for production of ammonia synthesis
gas.
This report was submitted in fulfillment of Grant No. S-802934 by
Texas Tech University under the partial sponsorship of the U.S. Environ-
mental Protection Agency. This report covers the period January 1, 1974,
to June 30, 1977, and work was completed as of June 30, 1977.
iv
-------
CONTENTS
Page
Foreword iii
Abstract iv
List of Figures vi
List of Tables vii
Acknowledgments viii
Sections
I Introduction 1
II Conclusions 4
III Recommendations 5
IV SGFM Process Description 7
V Experimental Equipment, Procedures, and Materials 9
VI Discussion of Results 22
References 47
Appendices
A. Table A-l Operating Conditions for Individual Runs
and Samples 52
Table A-2 Gas Composition Data 54
Table A-3 A Model for Reactor Operating Characteristics. . 56
B. A Model for Reactor Operating Conditions 58
C. Patent Disclosure 63
-------
FIGURES
Number page
1 General flow sheet for the SGFM process 8
2 Schematic diagram of SGFM process 10
3 Reactor dimensions 11
4 Char hopper and distributor plate 13
5 Electrical system 16
6 Cyclone dimensions 18
7 Influence of reactor temperature on dry gas yield 26
8 Ultimate hydrogen yield as a function of average reactor
temperatures 27
9 Ethylene yield as a function of average reactor temperature . . 28
10 Temperature profiles in reactor 31
VI
-------
TABLES
Number Page
1 Summary of Operating Conditions and Product Gas Data 23
2 Calculated Gas Product Data 25
3 Selected Material Balance and Ash Data 39
4 Effect of Particle Size on Ash Content 43
5 Preliminary Results from Tar Pyrolysis 45
vi i
-------
ACKNOWLEDGMENT
The financial and moral support of the following individuals is
gratefully acknowledged.
Mr. J. J. Heitz
Mr. R. Douglas Kreis
Mr. Thurman Whitis
Mr. Charles E. Ball
Vice President, Pioneer Corporation
Project Officer, U. S. Environmental
Protection Agency
Senior Vice President, Pioneer
Corporation
Executive Vice President, Texas Cattle
Feeders Association
An academic fellowship sponsored by Phillips Petroleum Company is also
deeply appreciated.
We would also like to extend our gratitude to all the undergraduate
students at Texas Tech University who worked on this project. Their
enthusiasm was a constant source of encouragement. A special thanks goes
to Mr. H. D. Young, doctoral candidate at Texas Tech University, who pro-
vided several ideas during the course of the work and conducted basic
studies which contributed to a better understanding of reactor operation
and design.
viii
-------
SECTION I
INTRODUCTION
The cattle feedlot industry in the United States supplies a high-
volume, beef consumer market as indicated by the sales of nearly 25 million
head of feedlot cattle in 1972J This beef consumer market demands a high-
volume industry at all times despite the rather wide and severe fluc-
tuations that do occur, and seem to be the rule rather than the exception.
Both the high volume and fluctuations are illustrated by the production
cycles recorded within the service area of one utility company in West
Texas: approximately 4 million head were fed-out in 19742 compared to nearly
5 million head in 1973;3 actual production in 1975 may be less than 3.5
mil lion.^
The most negative aspect of feedlot operations is the solid by-product,
cattle manure, which can harm the environment by water run-off into water-
sheds and by noxious gas emitted into the atmosphere.5 When cattle are
fattened by grazing, the large area of land involved dilutes the potential
effect of pollution; but, a concentration of potential problems is, by
definition, associated with feedlots. Land spreading, still the most
used method of manure management,6 cannot be economically attractive when
large quantities of the by-product must be trucked any distance to achieve
only dilution ($2.25 per ton charge for a ten-mile haul has been estimated
by Sweeten?). As a result, huge piles of cattle feedlot manure accumulate.
This accumulation has been accentuated in recent years because there
is a trend toward larger feedlots, even with the severe cycles that the
industry experiences. For example, within the utility service area noted
above, 127 feedlots produced 1.2 million head of cattle in 1966; 188 feed-
lots produced 3.7 million head in 1970; and 150 feedlots produced 4.1 million
head in 1974. In the Hereford-Dilimit area on the High Plains of Texas, the
estimated production of 30 or fewer lots is 800,000 - 1,000,000 cattle.8
The concentration and accumulation of cattle feedlot manure is currently
an undesired by-product of the industry, but it may also be viewed as a
"residual resource," particularly in terms of the energy deficiencies
projected for Texas and the United States. Proceeding on the assumptions
that the environment must be protected and that a productive end-use can be
developed for this resource, one immediately asks, "What are the technical
options? How much do they cost? Can a profit be realized?".
Several technical options for using this by-product of the industry
have been considered by a number of investigators. These options include,
but are not necessarily limited to: a) utilization for land fertilizer;9'10
b) direct combustion (either singly or combined with solid fossil fuel)
1
-------
for energy value;11 c) feed ingredient in cattle rations;10 d) synthesis
gas generation by anaerobic digestion;1^ e) synthesis gas by pyrolysis or
partial oxidation;'-* and f) production of single cell protein."14
The economics of these options for cattle feedlot manure and their
possible use for general biomass utilization have been considered by
many groups and individuals as a means of supplementing food and energy
supplies. The evaluations include those of Boersma and Barlow,15 Elias,16
all inclusive; instead, references are cited only to show the breadth and
diversity of interest in virgin and residual biomass utilization. The
general results of these investigations and evaluations suggest that the
technical feasibility and economics of the options for cattle feedlot man-
ure are comparable to that for the utilization of municipal solid waste
(excluding direct combustion) as indicated by Halligan and Huffman.23
Using the incomplete, but more reliable results available from municipal
solid waste demonstration projects,11 one can project that similar processes
for cattle feedlot manure utilization are quite probably viable providing
the volume of waste processed within a radius of 50 miles exceeds 750 tons
per day, a viable, high volume, local market exists for the conversion
product, and long-term (10-20 years) investment financing can be obtained.
Without long-term financing, one must, however, conclude that the proces-
sing and utilization of huge volumes of cattle feedlot manure is, at best,
a marginal investment at this time in history. In other words, potential
investors in supplemental or alternate energy source projects soon learn
that oil and gas are still highly attractive energy sources. It is reasonable
to expect that oil-equivalent prices will probably have to exceed $15-18
(in 1975 dollars) before several of the aforementioned alternates become
truly profitable for private investments. These marginal alternates
include those which depend upon biological action or high solids residence
time (> 30 min).12
It is very important to note that none of the cited evaluations have
placed or would have placed a dollar value on the aesthetics of removing
manure and other solid residues (residual biomass) from the environment.
We doubt that one can assign a value; but, one must carefully consider
this qualitative aspect. For the short-term development of such processes,
the importance of pollution abatement may well be the crucial one instead
of the need for alternate energy supply. This concept has been summarized
by Whetstone'9 as follows: "... much of the cost in the conversion of
wastes to energy may be legitimately charged to pollution control; all of
the cost of biomass crops must be borne by the energy and other values
obtained ..." (emphasis by Whetstone). Alternately, as one of us (JEH)
has noted: solid residues are not an answer to the energy crisis; the
energy crisis may well be the answer to the solid residue disposal problem.
Thus, one may project that pollution abatement may provide a key motivation
factor for developing energy processes based on biomass. The improvements
evolved from such processing may then lead to highly competitive energy
processes based on biomass feedstocks.
-------
The preceding discussion strongly indicates that research on the utili-
zation of cattle feedlot manure is justified, particularly in terms of pollu-
tion abatement. The question is then what option. Certainly, investigations
on a number of options must be carried far enough to provide a sufficient
data base for realistic economic comparisons. This report presents the
results on one option which is consistent with the preceding economic
projections: the thermochemical processing or partial oxidation of cattle
feedlot manure for high-volume production of ammonia synthesis gas. The
technical basis for this research has been reviewed in a number of publi-
cations by Halligan and his col leagues^ >21» 4,25,26 ancj wjn no^ be re-
counted here; the reader is requested to refer to these publications,
particularly reference numbers 13, 24, and 25.
The rationale for choosing to investigate and develop thermochemical
processing and specifically the production of an ammonia synthesis gas
was and is as follows. Large numbers of cattle are concentrated in feedlots
and these animals require many thousands of acres of land in grain crops.
In turn, these crops demand significant quantities of fertilizers for
efficient production, such as anhydrous ammonia.24 As a result, large
ammonia plants (500-1000 tons per day) are not uncommon in the vicinity
of extensive, fed-cattle operations such as those in Texas, California,
Colorado, and Arizona.
Most of the ammonia produced in the United States is based on reforming
natural gas with steam to produce the hydrogen-nitrogen mixture required
for ammonia synthesis.27 Because of the projected, critical shortage
and/or higher price levels of natural gas, many ammonia producers are inves-
tigating the replacement of natural gas with other hydrogen sources.2°
Obviously, the utilization of cattle feedlot manure would not eliminate the
need for natural gas in ammonia production, but the interdependence of
ammonia production and feedlots via feed grain does strongly suggest that
some natural gas replacement may certainly be expected, upon development
of an economical process.
It is important to note that, during the course of this project, the
investigators have received serious contacts from more than six private
firms relative to commercialization; two of the firms are continuing their
analyses and discussing a joint venture. Several of these contacts have
developed from the joint public announcement issued by the U. S. Environmental
Protection Agency, Pioneer Corporation, Texas Cattle Feeders Association,
and Texas Tech University on July 29, 1975.
-------
SECTION II
CONCLUSIONS
1. The 2.5 m long reactor used in this study demonstrated that a partial
oxidation reactor for producing an ammonia synthesis gas can be
scaled-up from 0.1-0.4 kg/hr to 7.7-23.6 kg/hr of as-received feed
using reaction temperature as the correlation parameter. Heat transfer
to particles appears to be the controlling factor for scale-up con-
siderations.
2. The investigation has shown that 21-70 g/kg of ethylene can be pro-
duced using the reactor design and operation that evolved during the
course of the investigation. The production of this valuable olefin
depends upon reaction temperature and the time-temperature history
of the solids-gas in the reactor design as operated.
3. Limited energy/material balances were insufficient to demonstrate
that the reactor can be operated adiabatically. The results were,
however, sufficient to show that the large, 2.5 m long reactor was
similar to the small bench-scale unit. Thus, reactor char (approxi-
mately 35 percent of the total dry mass containing 50-55 percent
ash; heat value approximately 3,300 calories per gram (cal/g) would
have to be consumed within the process to generate steam and/or
compensate for sensible heat effects.
-------
SECTION III
RECOMMENDATIONS
1. A reactor process for the production of an ammonia synthesis gas
should be scaled-up to a pilot demonstration unit which is capable
of processing 200-300 ton per day of as-received cattle feedlot
manure (15-30 percent moisture, 20-25 percent ash by weight). Be-
cause a final design will require approximately one year to prepare,
from date of acceptance of this report, it is anticipated that the
additional data needed to confirm energy balances could be available
from the current reactor for critical design considerations. These
data are currently being obtained as part of an United States Energy
Research and Development Administration Contract with the Fuels From
Biomass Program, Division of Solar Energy (Contract No. E(29-2)-3779).
This recommendation is based upon the generally positive results of
this study. It is strongly urged that this recommendation be given
serious consideration to help demonstrate to the important agricul-
tural sector of this country that positive, technical advances can
and are being made to help alleviate their rising, energy-related
costs and investments through the use of a neglected solid residue
resource.
2. Investigations at the current reactor scale should be continued to
specify energy and material balances and help define the factors that
control the production of ethylene and other valuable petrochemical
intermediates. Studies should also be initiated to evaluate the
feasibility and desirability of reinjecting product tar and liquid
into the reactor to increase hydrocarbon yields.
3. Basic studies should be initiated to develop a firm understanding
of the residence time-history of the reactor design-operation
that evolved during the investigation. These studies could provide
valuable information relative to the use of other solid wastes and
feedstocks.
4. A development program should also be undertaken to assess the effect
of reaction pressure on product yield. This study could be initiated
on a small , bench-scale unit because the effect of reaction tempera-
ture on scale-up has been established for the production of an ammonia
synthesis gas. The effect of pressure up to 30 atmospheres (atm)
should be investigated to provide a crucial data base for selecting
economic optimums. This study could also be incorporated into the
first recommendation.
-------
Because the volume of cattle feedlot residue fluctuates rapidly from
year to year, immediate studies should be initiated to evaluate the
use of other agricultural waste products as feedstocks for the
reactor process. This would stabilize the supply of feedstock and
reduce the investment risk.
Studies should be initiated to define better the end-uses of the
final, ash product from the process; to determine the pollution
potential of the final wastewater stream and the potential recovery
of organics contained therein; and to determine the stack emissions
during combustion of the product tar.
-------
SECTION IV
SGFM PROCESS DESCRIPTION
The current scope of the Synthesis Gas From Manure (SGFM) process is
one which is projected to produce an ammonia synthesis gas (Nitrogen +
Hydrogen) and ethylene (a key chemical feedstock) from the solid by-pro-
duct of commercial feedlots. A block diagram of the overall process that
is projected for a full-scale plant is shown in Figure 1. In Figure 1,
it must be appreciated that complex chemical processing is involved in
each block, except for Feed Preparation and Energy Recovery where standard
equipment and combustion units would be employed.
The heart of the process is the reactor section which utilizes high
temperature to decompose the solid feed into useful gas products for
sales. A solid char (approximately 3300 cal/g) and an organic liquid are
also produced but these will be used internally. The solid char will be
needed to provide certain utility needs, e.g., generation of steam. The
liquid product will be recycled to the reactor to assure that the hydro-
carbon content of the gas can be achieved and/or combusted to recover its
energy content.
After production in the SGFM process, the ammonia synthesis gas could
be fed to an existing ammonia plant as a replacement for part of the natural
gas feed, or other feedstock that might be used. The ethylene could be
transported and sold to a petrochemical company using an existing pipeline;
the pipeline transport may require that the ethylene be blended with another
gas. Thus, in terms of this scope, the SGFM process would be located near
an existing ammonia plant with close access to an existing pipeline. Our
current information indicates that these criteria can be met in the
Hereford-Dimmit area of West Texas. Other locations could be chosen on
the basis of local market needs for ammonia synthesis gas, low BTU gas
and/or ethylene. Other product options are also being considered by the
investigators.
As indicated above, the heart of the process is the reactor design and
operation. Other items in the process are already commercially available
and do not require additional development. Thus, this project has concerned
itself with reactor development.
-------
LIQUID
RECOVERY+
TREATMENT
WASTE
WATER
00
STEAM
MANURE
AIR
t
w
FEED
PREPARATION
^
P
™
SGFM
RFAPTHR
SYSTEM
It
ENERGY
RECOVERY
^ GAS
* RECOVERY
AMMONIA
SYNTHESIS
GAS
ETHYLENE
ASH
Figure 1. General flow sheet for the SGFM process.
-------
SECTION V
EXPERIMENTAL EQUIPMENT, PROCEDURES AND MATERIALS
As with most developmental projects, the equipment and procedures
were in a constant state of change throughout the course of the project.
As more data became available on the operating characteristics of the
system, various modifications were initiated to increase reliability,
improve the accuracy, remove limitations, and enhance understanding of
the data. The material presented in this section incorporates the original
and latest revisions to illustrate difficulties in working with hot solids
and gases. Characterization data on cattle feedlot manure are given in
reference 24.
FLOW OF FEED AND PRODUCT COMPONENTS
Solid manure was fed to the top of the hot reactor where it fell
countercurrent to a gas mixture of steam and air which were introduced
through a distribution plate at the bottom. Gases resulting from the
pyrolysis and partial oxidation of the manure exited with the input gas
stream out the top of the reactor. The gas was stripped of entrained
solids, tars, and condensables in stages. Char was removed from the bottom
of the reactor through an opening in the input gas distribution plate.
This flow path scheme is illustrated in Figure 2.
REACTOR AND FEED HOPPER
The reactor was made from schedule 40 stainless steel pipe, Figure
3. The main body of the reactor consisted of a 15.24 cm diameter, 1.524 m
long section at the top for separation of the solids and gas. In order to
allow for approximately 2 cm vertical expansion due to the high tempera-
ture operation, the reactor was welded to horizontal, 20 cm I-beams at the
bottom and guided at the top by triangular ears which moved in slots welded
to another set of I-beams.
A star feeder, Model R2-653 manufactured by Beaumont Birch Co., was
mounted between the manure hopper and the reactor using a Reliance VSD
variable speed motor, Model B56G3102, with a 12 to 48 sprocket gear ratio.
The inlet gas distribution plate at the bottom of the reactor consisted
of a 0.95 cm stainless steel plate with concentric circles of 0.1588 cm
holes drilled into it aroung a 2.54 cm hole in the center. This distribution
plate was welded to a thin collar which was welded to the bottom flange of
the reactor to provide an even distribution of gas to the bottom of the
-------
Solid
Holder
Heaters
Pressure
01 Control
Valve
Condensate
Removal
Turbine
Meter Product Gas
to Stack
3 Figure 2. Schematic diagram of SGFM process.
10
-------
(All Dimensions in Centimeters)
45.7
31.8 - 900#RFWNFLN6-
STAINLESS STEEL
31.8 - SCHEDULE 40 PIPE
STAINLESS STEEL
20.3 - STD. WT. WELD CAP-
STAINLESS STEEL
20.3 - SCHEDULE 40 PIPE-
STAINLESS STEEL
20 x 15 CONCENTRIC REDUCER-
STAINLESS STEEL
15 - SCHEDULE 40 PIPE
STAINLESS STEEL (316-L)
31.1 - COLLAR-
STAINLESS STEEL
15.2 - 900# RAISED FACED
WELD NECK FLN6 W/BLIND
STAINLESS STEEL
Figure 3. Reactor dimensions
11
-------
reactor. A 2.54 cm OD stainless steel tube passed through the bottom flange
and was welded to the central 5.08 cm hole to permit char to exit the reactor.
This tube extended about 15 cm through the outside of the bottom flange
where it was connected with Swagelok fittings to the char hopper. The
whole distribution plate system and flange was fitted with very close
tolerance into the bottom of the reactor.
CHAR HOPPER AND RAM
The char hopper was constructed to allow solids flow to be diverted
into one of two sections to isolate the char produced during start-up
and steady state conditions (Figure 4). The steady state accumulation was
used for material balance estimates. The original design had a large flange
at the top of the hopper to provide easy access to the interior. Because
of the awkwardness or removing the whole hopper after each run to remove
solids, the difficulty in removing char from each side without crossover,
and gas leaks in the flange, the flange was cut off and a plate was perma-
nently welded on. Access to each of the two sections was then provided
using 10 cm, flanged nipples which could be easily opened to obtain samples.
A pipe was installed through the center of the hopper to act as a guide
for an air-actuated ram. The ram was a polished rod which extended through
the distribution plate and effectively plugged the exit port of the plate
when it was in the up position. In the down position, the top of the ram
was about 15 cm below the entrance port to the char hopper which allowed
solids to flow from the reactor. The ram cylinder was pressed against a
lead gasket at the bottom of the char hopper by chains and turnbuckles
to provide an airtight seal.
REACTOR HEATERS
The heating section of the reactor consisted of two, 60 cm long sections.
Each section was constructed of 4 sections of ceramic heaters, manufactured
by the Lindberg Co., Model 50752, Type 77-KSD. The sections were fitted
around the pipe and secured with external bands. Each section was rated at
40 amps, 206 volts and each was protected with circuit breakers. During
the course of the research, several failures were experienced with these
heaters, but the failures were due to one case of improper electrical
installation and excessive temperature during one of the early runs. Due
to these incidents, several of the elements in a given section (2 elements
to a section) progressively failed. By the end of Run 10, only one element
in the lower section, or 12.5 percent of the original capacity, was operating.
Only about 50 percent of the upper bank was still operating during the
final runs. About 25 percent of these failures were obviously due to mech-
anical breaks in the nichrome lead-in wires in places where they could not
be repaired. Some of the other failures were apparently due to oxidation
and melting of the nichrome material. The entire bank of heaters have
been replaced.
12
-------
Reactor
Flapper
Ram Guide
Ram
Divider Plate
Access Port
Air Piston
TUT
TUT
Air-Steam
Distributor
Plate
Inlet Tube
For Air-Steam
Mixture
Char Hopper
Flange
Figure 4. Char hopper and distributor plate,
13
-------
PREHEATER
The preheater consisted of a U-shaped loop of stainless steel tubing.
The first section was 4.9 m of 1.90 m tubing and the remaining 5.2 m was
1.59 cm tubing. This loop was connected by 1.90 cm braided copper welding
leads to a Hobart Model T-500-5422, 500 amp welding generator. To prevent
melting of the preheater tubing, the contactor was controlled by a Thermo
Electric, Model 32042-02-005, Mini-Monitor Analog Latching, High-Temperature
Alarm triggered by a thermocouple clamped to the top of the preheater.
The tubing of the preheater was encased in insulation. Natural convec-
tion from the large surface area of the preheater tubing and a limited
power supply minimized any chance of melting except in the case of no
air flow through the tubing.
INSTRUMENTATION AND CONTROLS
Type K thermocouples were used to monitor the temperature profile in
the reactor and were attached to a 24-point, Leeds and Northrup recorder,
Model 547. A Flexapulse pneumatic timer, manufactured by Eagle Signal
Corp., was used to control the motion of the ram. This unit could be set
in control independently both the time the ram remained in the up-position
and the down-position.
To control the input air stream, a Brooks rotameter, Model 1110-06F1A1A,
was installed and calibrated using a wet test meter. A Brooks 1110-01F1A1A
rotameter was used to measure helium flowrate to the feed hopper. The
final helium flowrate required for the hopper was beyond the capability
of this rotameter, but it was not replaced because the flow was not criti-
cal to the operation. The cold, product gas rate was measured using a
Rockwell, Model TP-4, 5.08 cm turbometer.
An Eli son Eagle Eye flowmeter (mechanical manometer) with an Annubar
semi-pitot type sensor was purchased to measure the steam flowrate. This
unit proved to be inadequate due to its slow response time (measured in
hours) and the relatively high air content of the steam. A calibrated
needle valve was finally installed and the steam flowrate was measured by
diverting the steam flow through a hose placed in a bucket of cold water.
The increase in weight of the water bucket resulting from condensed steam
was used to calculate the steam flowrate both before and after a run.
Pressure taps were provided every 15.24 cm along the reactor to
measure the pressure drop. These taps were manifolded through three,
5-way valves to a Honeywell Pneumatic Differential Pressure Transmitter.
The output of the transmitter was connected to two pressure gauges. One
gauge was located on the main control panel and one was installed on a
pressure panel at the star feeder.
To control the reactor heaters, two Victory Engineering, Model 1501AG-
58-IK, proportional temperature controllers were used to drive Victory,
Model Pak 10-0-1X triac contactors. Due to the severe voltage load required
14
-------
for the controllers at startup and variable power supply, the original
triac contactors were replaced and a capacitive bypass for surge protection
was installed. New silicon-controlled rectifiers (SCR's), Vectrol , Model
VSSC-1020-240-1FAC, were also installed which were capable of handling
63 amps at 240 volts with full surge protection. In addition, the original
proportional controllers were modified using 2500 ohm variable center
tap resistors on the output stages. The center tap was adjusted to provide
voltage through a rectifying bridge to the 15 volt D.C. control circuit of
the Vectrol SCR's (Figure 5).
A Fisher, Model 657-ES, pressure control valve with an integral Fisher
Wizard, Model 4100ZR, controller was installed to regulate the pressure in
the system. The pressure sensor for the control was located just after
the reactor exit port in the 5.08 cm line leading to the cyclone
MODIFICATIONS
Although the reactor system operated successfully on the first run,
extensive modifications were required to allow a wider range of operating
conditions and to solve various problems that developed.
Feed Hopper
From the very first run, it was apparent that any steam entering the
star feeder would cause the manure flow to stop almost immediately. Moist
manure would be packed into a stiff paste by the mechanical action of the
star feeder or would simply agglomerate to form a plug in the hopper above
the star feeder. Initially, it was found that steam flow to the feeder
could be prevented by the excessive use of helium flowrates (one cylinder
per hour). A concentrated effort was undertaken to eliminate leaks. The
ultimate solution required to minimize leaks required new, heavier seals
on the 10 cm, feed inlet section at the top of the hopper and a new cone
for the bottom of the hopper. It was also necessary to redrill the flange
separating the cone and the hopper, reweld all the joints on the hopper,
and install silicon seals on required flanges. These modifications cut
the required helium flow by an order of magnitude.
Unfortunately, it was soon apparent that this was not enough. As the
pressure rose in the reactor during startup, backflow (steam and gas)
into the feed hopper occurred to equalize pressure in the hopper. If back-
flow occurred faster than adjustments to the helium purge flow rate,
plugging resulted. To prevent this startup backflow from occurring, a 0.95
cm bypass line was installed between the top of the manure hopper and a
point just below the star feeder. This line incorporated a 5 cm by 38
cm water-cooled condenser with a small reservoir and blowdown valve to
condense any steam in the bypass stream.
To monitor any backflow that still inadvertently occurred, another
Thermo Electric Mini-Monitor, Latching High-Temperature Alarm was installed
with a warning buzzer. The alarm was actuated when the temperature below
15
-------
Semiconductor
Fuse
Thermo
couple
Figure 5. Electrical system.
16
-------
the star feeder exceeded approximately 50°C. The alarm signal was used to
set the helium flowrate after the startup was accomplished and this permitted
more conservative helium flowrates to be used. All gas rate data have
been corrected for the added helium rate.
Cyclone and Tar Trap
By the end of Run 4, it was apparent that the original filter arrange-
ment for removing solids from the product gas was not acceptable. After
a few runs, the filter became clogged with condensed tar and small particles.
The clogging was apparently caused by the small surface area available, the
dense packing of particles, the unheated design, and a high amount of solid
entrainment (greater than expected).
A sample of the filtered material was obtained and its size distribution
was estimated. This particle size distribution was used to design a heated
cyclone (Figure 6). The cyclone was installed as close to the reactor as
possible because heat transfer calculations showed that it was impractical
to insulate a long length of pipe to the degree required. The filter
was removed from service. The cyclone was heated with natural gas burners
during the investigation. For a full-scale or commercial operation, the
cyclone would be installed within the reactor as is common in petroleum
industry catalytic crackers.
After solving the particle carryover problem, the low efficiency of
the trap was investigated. It was found that only small quantities of the
tar were condensing at this point. Almost all of the tar was being collected
in the impinger located immediately after the tar trap. To increase the
surface area available for condensation, a rolled piece of aluminum screen
was dimpled and inserted into 76 cm of the 5 cm, air-cooled line immediately
upstream of the trap to act as an impinger. This screen was not very
effective because the impinger still collected most of the tar. However,
a small amount of a higher melting point tar did accumulate in the tar trap.
Both sections of the tar collection system were later steam-traced and a
10 cm diameter, 30.5 cm long reservoir was attached to the impinger to
collect tar. The steam tracing was installed because it was found that tar
would drain from the impinger at an adequate rate when the temperature
was greater than 75°C.
Mater-Gas Separation
The water-gas separation system was originally constructed with a 5
cm knock-out drum, a modified, inverted-bucket steam trap and another
impinger of the same type as used in the tar trap system. The steam trap
did not have the necessary throughput and was used only as a backup for
overflow. A 1.27 cm needle valve was modified into a float arrangement
and installed as the primary valve. This valve was operational but would
frequently become stuck with entrained, gummy tar at room temperature.
Taps with a hammer were used to achieve a stable operation. After the
last run of this investigation, a fire started in the tar trap due to an
improper shutdown. This excessive heat vaporized or degraded the
17
-------
(All Dimensions in Centimeters)
20 to 5 reducer
Blind flange with 20 bore
10.2
SIDE VIEW
Figure 6. Cyclone dimensions.
18
12.7
-------
collected tars. These tars were carried over and plugged up the water-
cooled condenser. Steam was used to clear the condenser, but further
entrainment of the tars accumulated in the needle-valve, float assembly
noted above. Rather than clean this assembly, a large, commercial float
valve was installed.
Flow Measurement System
A very dilute mist of oil was entrained with product gas despite the
cooling and condensation equipment that was installed. Some of this mist
accumulated on the gas turbometer and would prevent rotation. To prevent
fouling, another water-cooled condenser and a 3.18 cm diameter cyclone was
added. A bypass of the turbometer was also installed because most tar
was formed at startup. The small cyclone was observed to remove very little
of the entrained liquid in the gas and a cloth filter was installed for
final cleanup. The filter was effective, but the filter media had to be
replaced every run because of internal by-passing and available surface
area. A larger, final filter has been installed.
OPERATIONAL PROCEDURE
1. The reactor was heated to operating temperature using an approxi-
mate air purge rate of 5100 Standard Liter Per Hour (SLH).
2. After setting the helium purge rate, steam flow was initiated.
The steam and air rates were then adjusted to those for a given
run.
3. Once the system was thermally stabilized and the water removal
system was functioning, the manure feed was started. A ram
cycle of 3 seconds down and 10 seconds up was employed.
4. After the system reached a new thermal steady-state, the cyclone,
tar trap and water containers were purged of accumulated material.
The char hopper flapper was switched to the steady-state section
and a run to collect gas, tar, and solids for a material balance
was initiated. Gas samples for analysis were collected every
15-20 minutes.
5. The mass balance or run period was ended by switching the flapper
in the char hopper back to the startup side. Gas, tar, water,
and solid were then collected.
6. Gas samples were taken by connecting a gas sampling bottle across
the sample loop and adjusting the pressure drop across the loop
to 0.06 - 0.10 atmospheres.
7- The system was shut down by reversing steps 1, 2, and 3.
19
-------
8. Char was removed from the hopper and weighed.
ANALYTICAL PROCEDURES
Gas samples were collected in 250 ml gas collection bottles and ana-
lyzed using a Varian Aerograph Gas Chromatograph with two columns. One
column was a 3.65 m long by 318 cm diameter tube packed with 50/80 mesh
Poropak Q. A 1.2 m precolumn of 0.318 cm tubing packed with 50/80 Poropak
R was used in conjunction with the Poropak Q column. These columns were
used to quantitatively measure hydrogen, methane, carbon dioxide, ethylene,
and ethane. A 3.65 m long by 0.318 cm diameter column packed with 45/60
mesh, 5 A molecular sieve was used to measure hydrogen, oxygen, nitrogen,
methane, and carbon monoxide. Runs 1 through 5 were analyzed through the
courtesy of Young29 who was studying the theoretical aspects of the pyro-
lysis process using a different chromatograph with similar columns.
A series of one ml volumes of air were injected before and after each
samples series. The air peaks were averaged and a correction factor was
estimated, which was used to adjust the sample peak heights. At least
two separate injections on each column were made for each unknown sample.
The peak heights were then converted to moles using calibration curves.
Any oxygen present in the sample was estimated to be due to air contamina-
tion and was subtracted out along with a rationed amount of nitrogen.
The two column were related by the amount of methane measured on each side.
This procedure was refined, and a computer program was written to calculate
the various run results from the chromatographs and other input data.
Sieve studies on the manure feedstock and char were done by shaking
samples through standard sieves with an automatic sieve shaker for 10
minutes. The fractions were then weighed. Moisture content of the manure
found by heating samples in a weighed crucible at 104°C overnight, desi-
ccating, and then reweighing. Ash content was obtained by heating the sample
in an air environment at 950°C overnight and then finding the weight
differential. The bulk density of the solids was measured using a calibrated
volumetric bottle. Because of the small total volume of the volumetric
bottle, the sample density of large particles may be lower than the true
value.
PREPARATION OF FEEDSTOCK
:ained from a local, commercial feedlot (Lubbock Feedlots)
Based on the insensitivity of the process to the actual manure feedstock,^>
Manure was obtained from
UUO wU Wl I wi iW tiiw^ii^f ** » » • uj ** • »• ••— |- .•—•—•———• —— — •— — — — — r
the commercial manure pile was surveyed by visual examination and to find
a reasonably dry, fresh manure for the feedstock. The commercial feed-
stock was collected and transported to the Texas Tech Feedlot where a
standard, farm hamnermill was used to grind the manure. A screen of 2.5 cm
squares was used in the hanmermill. The manure used in Runs 1 through
20
-------
3 was sun-dried for a few hours on a large plastic sheet. The manure used
for the remaining runs did not require drying. In Runs 1 through 8, the
manure was only sieved to remove gross particles greater than about 1 cm
in diameter. For Runs 9 and 10, the manure was sieved to eliminate par-
ticles greater than 1/3 cm in diameter.
21
-------
SECTION VI
DISCUSSION OF RESULTS
OPERATION OF SYSTEM
A run was initiated by heating the reactor to the desired, nominal
temperature and holding it at this level until the temperature of auxiliary
equipment reached a steady level. This heat-up time was normally 4 to 7
hours during which an air purge of 5100 liters per hour (L/H) was passed
through the reactor. Once a thermal steady-state had been achieved, a
helium purge of the feed hopper was initiated to prevent steam from conden-
sing in the inlet feedstock lines and starfeeder (Note: In a full-scale
operation, a part of the product or waste gas could be diverted for this
purpose; helium was chosen for this purpose because it was the carrier gas
in the gas chromatograph). The desired steam and air flow rates for an
acutal run were then established and stabilized. After stable operation
of the water-gas separation system was achieved, manure feed was introduced.
The system was permitted to reach a new steady-state at the controlled
reactor temperature. This new steady-state was generally achieved in 15
to 30 minutes. Gas, tar and char samples were then collected for analysis
at 15 to 30 minute intervals. Reactor operation with manure feed ranged
from 0.5 to 3 hours, excluding cool down.
Only data from Runs 1, 6, 7, 8, 9, and 10 are discussed in this section
because others were too short to provide reliable results. These short
runs (2, 3, 4, and 5) were due to electrial heater failures and plugged,
solids removal filters. This filter pluggage was the major reason for
installing the cyclone shown in Figure 2.
EXPERIMENTAL RESULTS
The data from the operation of the large-scale reactor have been arranged
and summarized in Table 1. A complete listing of results is given in
Appendix 1. In general, these results indicated that dry gas yield and gas
composition per unit of manure feed were largely independent of the air and
steam feedrates, over the ranges studied. Reactor temperature was the only
consistent variable which would order the data. This is in agreement with
the results reported by Halligan, et.al. for a bench scale apparatus.24,25
These correlations and interpretations are discussed below, using an average
reactor temperature calculated by the trapezoidal rule. The value given
for the bench-scale unit was the arithmetic average of two measurements.
While it did not appear as a significant correlation parameter, oxygen
in the air feed is a key reactor parameter because the partial oxidation of
22
-------
Table 1. SUMMARY OF OPERATING CONDITIONS AND PRODUCT GAS DATA
PO
00
Operating Conditions
Manure Feed Rate, kg daf/hra
Manure Feed Rate, kg ar/hr
Air Feed Rate, kl/hr
Steam Feed Rate, kg/hr
Particle Size, in
Average Temperature, °C
Product Gas Data
Total Dry Gas, 1/g dafc
Heat Value (HHV), cal/1
Gas Composition, vol %
H2
N2
CH4
CO
co2
C2H4
C2H6
Run Number
5.
7.
1.
5.
<0.
71
1.
1
22
76
149
44
95
1
19
2855
25.
14.
12.
11.
30.
4.
0.
2
6
8
6
8
7
3
6a
7.21
10.60
1.700
4.54
<0.95
695
(0.667)
2918
22.2
27.8
7.7
15.3
20.7
6.4
0.5
6b
16.15
23.61
1.487
4.54
<0.95
641
0.580
3790
20.0
15.1
12.6
21.3
22.1
8.5
0.4
7
12.
18.
1.
3.
<0.
61
0.
97
96
904
63
95
7
406
3380
28.
23.
9.
16.
15.
4.
2.
2
2
2
4
4
9
7
8
12.34
18.05
1.402
3.08
<0.95
629
0.455
3523
17.4
26.7
14.1
21.2
14.1
5.8
0.7
9
12
18
4
2
<0
10
.70
.01
.249
.72
.32
668
(0
.718)
2624
15
36
8
20
14
4
0
.1
.8
.9
.3
.2
.2
.5
8
11
0
3
<0
.26
.75
.765
.72
.32
628
0
.318
3345
20
24
11
22
14
5
0
.9
.2
.7
.4
.8
.5
.5
a daf - dry, ash-free; ar = as received
b All data are average values from at least two samples. Individual gas samples were analyzed on the
gas chromatograph using at least two injections.
c Values in parentheses are back-calculated values using a nitrogen balance.
-------
the manure supplies part of the energy necessary to heat the manure feed-
stock to reaction temperature. This exothermic reaction is indicated by
comparing the average temperature of Runs 8 and 9 of Table 1. Run 9 was
operated with three times the air flow of Run 8 with the same electric
heat input. The apparent oxidation is indicated by a 39°C higher average
reactor temperature. The significance of this temperature increase is in-
dicated by noting that the total range of average temperature in Table 1 is
93°C.
CORRELATION OF DATA
The yields of total gas, ultimate hydrogen (Hg + CO; see discussion
below) and ethylene are given in Tables 1 and 2 and presented as a function
of average temperature in Figures 7, 8 and 9, respectively (the correlation
line through each data set is a least-squares fit). Comparison data from
the bench-scale experiments of Halligan, et.al.24>25 are also shown in these
plots. Because the slopes of these 3 data sets agree well with the bench-
scale data, the general interpretation is that reactor temperature must
be considered a key scale-up variable for the partial oxidation of manure
as developed in these studies.
A preliminary analysis of the operating characteristics of the reactor
has shown, however, that another parameter, volume fraction of solids, may
be important because this parameter could affect heat transfer as well as
alter the reaction history of particles. Such a factor may account for
part of the scatter in the data, but sufficient detail has not yet been
developed to identify the influence of this effect. A preliminary analysis
on the effects of the volume fraction of solids on reactor operation is
given as Appendix 2 of this report.
The apparent absence of oxygen or combustion effects on the product
gas rate or composition may be explained in part by the high temperature of
the reaction and the relatively low rate of oxygen feed. At the temperatures
in the reactor, complete reaction of oxygen within 60 cm of the air inlet
may be expected because the rate of feed is much less than stoichiometric.
Confirmation of this is indicated by a detailed comparison of runs 8 and 9
discussed above.* In addition, the volume of combustion products is small
because calculations have shown that combustion products only contribute
15-40% of the CO and 0)3 composition. The combustion or oxygen effect is
also confused with electric heater input energy. Either of these two sources
of energy can dominate, but the correlation variable is still temperature
in terms of manure pyrolysis.
PRODUCTION OF AMMONIA SYNTHESIS GAS
Figure 8 and Table 2 ill
an ammonia synthesis gas from
Figure 8 and Table 2 illustrate the potential for the production of .
rom manure. As discussed by Halligan, et.al./ »
24
-------
Table 2. CALCULATED GAS PRODUCT DATA
ro
in
Run Number
Ultimate H2 (H2 + CO), 1/g daf
Ultimate H2: (N2) Ratio
(Reformed H2): (N2) Ratio
(Reformed H2 without C2's): N2 Ratio
Ethyl ene Yield, 1/g daf
% C0H. without combustion products
1
0.44
2.5
8.0
6.0
0.056
6.8
6a
0.25
1.3
3.8
2.5
0.042
11.0
6b
0.24
2.7
9.5
6.1
0.047
10.3
7
0.18
1.9
4.7
3.5
0.020
8.7
8
0.17
1.4
4.9
3.6
0.026
9.9
9
0.25
0.95
2.6
1.9
0.030
9.5
10
0.14
1.8
5.1
3.7
0.017
8.7
-------
1.5
1.2
c/3
<
015
0.9
0.6
0.3
p
Bench-Scale Data,
Hailigan, et. al,
(1975)
_ / O
I I I
600 680 760 820
AVERAGE TEMPERATURE, °C
Figure 7. Influence of reactor temperature on dry gas yield.
26
-------
o
o
0.5
0.4
X
Q
_l
^ 0.3
il
0-£
§5 0.2
>- -I
X
LLJ
? °'1
560
Bench-Scale Data
Halligan, et. al,
(1975)
640 720 800
AVERAGE TEMPERATURE, °C
880
Figure 8. Ultimate hydrogen yield as a function of
average reactor temperatures.
27
-------
0.06
0.05
0.04
0.03
0.02
0.01
Bench-Scale Data
Halligan, et. al,
(1975)
I
I I
600 680 760 820
AVERAGE REACTION TEMPERATURE, °C
Figure 9. Ethylene yield as a function of average
reactor temperature.
28
-------
ultimate hydrogen is defined as the sum of H2 + CO, assuming the CO is
shifted as now accomplished in commercial ammonia plants (CO + 1^0 =
C02 + H2). Reformed hydrogen is defined as the ultimate hydrogen plus
the hydrogen that would result from the complete steam reforming of CH4
and other hydrocarbons (e.g., C2H4 + 2H20 = 2CO + 4H2): the steam reforming
step is also accomplished commercially using methane as a feedstock. These
results (Table 2) show that the necessary molar ratio of 3:1 can be realized
from the partial oxidation of manure over the range of conditions explored.
It is recognized that the proper H2:N2 ratio can always be achieved by
simply adjusting air feed at an imposed temperature. The objective of the
reactor process is to achieve maximum air input to minimize external heat
duty with the constraint of a 3H2:N2 ratio. Another objective of ammonia
synthesis gas production is to exploit the N2 content of air, as opposed
to viewing it as a diluent which must be balanced against the cost of
manufacturing oxygen. Conversely, effective utilization of N^ in the air
from the process minimizes carbon; but, nitrogen utilization in agricul-
tural sectors is of equal importance, if not more so, than carbon utili-
zation which is a common goal of many coal conversion processes.
The results on ammonia synthesis gas production verify the small-scale
reactor data reported by Halligan, et.al.24 and the reader is referred to
this reference for a more thorough discussion on the use and processing of
this product. It is obvious from the data in Table 1 and Table 2 that the
production of ammonia synthesis gas has not been optimized and that addi-
tional studies remain to be completed to develop the optimum. However, the
data from both the large and small reactors do show that ammonia synthesis
gas can be produced from cattle feedlot manure.
PRODUCTION OF ETHYLENE
The ethylene yield of Figure 9 illustrates one possible key difference
between the large-scale and bench-scale reactors. As noted earlier, the
large-scale reactor operates with a nearly complete countercurrent flow of
solid and gas, with 5-20 percent of the total product char (as fine particles)
exiting with the gas. This flow pattern was not achieved in the bench
reactor and 67-87 percent of the solid char product was entrained. The
higher fraction of hot, entrained solids (heat capacity about twice that
of the gas) plus a review of the bench-scale operation indicates that
product gas from the smaller reactor probably remained at temperature for
a longer period of time. Thus, a higher degree of olefin cracking or pyro-
lysis would be expected for the bench-scale unit. This interpretation is
supported by a 40°C temperature shift in ethylene data (Figure 9) compared
to a 20°C shift in the total gas (Figure 7) and ultimate H2 (Figure 8).
The total gas or ultimate hydrogen rates would not be affected much (< 10%)
for a 3-4 percent change in ethylene composition. The net ethylene pro-
duction would be a strong function of time and temperature because the rate
of the three-halves, decomposition kinetics at 6 percent (vol) concentration
and 600°C is only 5.5 percent the rate of decomposition at 700°C. Alter-
nately, the half-life at 600°C is 1746 sec (6 percent vol) whereas it is
only 67.7 sec at 700°C using the data of Kunugi, et. al.30
29
-------
The production of ethylene (hence its time-temperature history) is of
significant practical interest because the current sales value is 10-12
cents per pound and one major manufacturer indicates that a 20 cent per
pound price may evolve by the early 1980's.31 Using a value of 0.05 L/g
daf manure and the above current price, the ethylene in the synthesis gas
has a sales value of $11-$14 per metric ton daf manure. The current data
show that up to 70 kg C2H4/metric ton daf manure (approximately 0.56 L/g
daf in Figure 9) can be produced; our studies indicate that reinjection
of tars and oils produced in the reactor (10-30 percent by weight of manure
feed) could potentially raise the ethylene yield to 80-90 kg/ton daf.
Because the economics of ethylene production are attractive, it is impor-
tant to note that several of the runs in Table 2 suggest that significant
quantities of both ethylene and an ammonia synthesis gas could be produced
at the same reaction conditions. Our current research goals are to colla-
borate the reported results and maximize the production rate of both an
ammonia synthesis gas and ethylene.
EFFECT OF REACTOR OPERATION ON ETHYLENE PRODUCTION
During run one, 4.7 percent of ethylene was noted in the produced gas
stream. This high concentration at a high maximum temperature (920-930°C)
was attributed to a "quenching" effect in the upper half of the reactor,
which had been established by the lack of adequate insulation. It was
decided not to add additional insulation and investigate the possibility
of increasing the production of ethylene because this olefin has a high
sales value.
As a result of the light insulation, tremendous heat losses were sus-
tained which made it difficult to achieve an accurate heat balance on the
system and to develop critical interpretations on the reactor operation.
However, detailed reviews of the temperature profiles of the hot reactor
before adding manure and during operation provided some insight into the
actual operating characteristics. These temperature profiles indicated
that most of the combustion was occurring just above the distribution
plate, perhaps confined to a height not more than 30 cm from the plate.
This can be shown by comparing the temperature profiles of Run 8 and 9 as
shown in Figure 10. In these two runs, only one heating element was
operational and was at maximum capacity at all times. Run 9 had an air feed
ratio about 3 times that of Run 8. All other feed rates were the same.
The rapid temperature rise in Run 9 could only be accounted for by the
release of heat through combustion. The possibility that a true fluidized
bed could exist and distribute the heat within the lower section was
judged to be unlikely because of the noticeable dip in the temperature
profiles between the heaters where the reactor was very poorly insulated.
The differences in the temperature profiles above the heaters could be due
to a higher heat transfer coefficient inside the reactor in Run 9 at the
higher total gas flowrate. The temperature difference beyond the heaters
is, however, estimated to be of small importance compared to the differences
in the lower part of the reactor.
30
-------
900 -r-
800 _
700
O
o
CC
LU
D.
2
111
600
500
400
300
Lower
Heater
Upper
Heater
Section
No Feed Startup
Run 9
I I I I I I
20 1 2345678 9 10 11 12 13
THERMOCOUPLE
Figure 10. Temperature profiles in reactor.
31
-------
As noted previously, temperature was the only correlating variable and
there appeared to be no clear correlation between air feed rate and exit gas
composition except for slight shifts in the C02-CO ratios, provided the
dilution effect of the air was removed mathematically. The strong temperature
correlation and absence of other factors indicated that the desired synthesis
gas reaction was basically a true pyrolysis.48 This is supported by the
basic work of Young29 which indicates that most of the pyrolysis reaction
should be over by the time a manure particle reaches the highest temperature
zone in the reactor, provided heat transfer to a particle is not limiting.
Young's work29 indicated that pyrolysis after the maximum temperature is
unlikely for most of the particles of the size (< 0.3 cm) used in this
study.
This quick reaction concept is also supported by the fact that the ash
content of fines collected in the cyclone was comparable to that of the
char collected from the bottom of the reactor (see NON-GAS PRODUCTS AND
OPERATION DOWNSTREAM OF REACTOR, Reactor Char, Particle Size and Mass Reduc-
tion). In other words, on the basis of the ash content, the entrained fines
appear to have reacted to nearly the same degree as the char collected from
the bottom of the reactor. Because most of the entrained fines should have
a lower residence time with little opportunity to react with oxygen, this
observation strongly suggested that most of the solid-phase reaction occurs
in the upper half of the reactor.
Thus, the reactor was apparently divided into two major zones of
reaction. In the top zone, pyrolysis is expected with combustion occurring
in the bottom zone. The middle zone could sustain a mixture of these two
major reactions as well as several secondary ones. These secondary reactions
might include:
C + H20 *- H2 -i- CO (1)
CO + H20 »~ H2 + C02 (2)
Intermediate •- CH4 + C2H4 + H£ + C2Hg (3)
P u ^ C H + H (5)
The carbon monoxide for reaction 2 could be present because oxygen concentra-
tion in the bottom of the reactor is less than stoichiometric and the gas
influent is rich in steam. Also-carbon monoxide is a product of the pyrolysis
reaction of the solid itself32'33'34 and ash is known to exhibit some
catalytic behavior35. Thus, reactions 1 and 2 have the potential of
shifting the exit gas composition. The gas analysis data did not provide
any conclusive evidence regarding the influence of these first two reactions.
One would expect hydrogen concentration to be dependent upon these reactions
and hence steam feed rate; but, hydrogen concentration did not correlate
with steam feed rate. Again, temperature was the dominant factor. As also
32
-------
discussed by Halligan, et.al.,24>25 the steam-carbon and steam-hydrocarbon
reactions are estimated to have a minor role at the reaction conditions
based on the results of other investigations. In particular, the rate of
steam reactions are thought to be low because reactor pressure is 2
atm or less and average reactor temperature is generally less than 927°C
(1700°F; 36, 37). Briefly, steam was added in the bench-scale experiments
of Halligan et.al.24»25 as a "fluidizing" medium which could be easily
separated from the gas products by condensation and a similar inert role
was assumed for the larger unit. For example, without the addition of steam,
the gas velocities in the larger unit would range from approximately 3
cm/sec, to 40 cm/sec instead of the actual 23-97 cm/sec.
In terms of the production of ethylene, the chemistry suggested that
reactions (3), (4), and (5) above are probably the critical factors. Most
reported results32>34,35 suggest that reaction (3) is preceded by reaction
(6)
Heat
Cellulosic Material •- Intermediate + CO + (XL (6)
where the intermediate has been found to be largely composed of levoglucosan
when the initial feed is pure cellulose and wood32»34; a very diverse
mixture with no major components is obtained when the cellulosic material
contains high amounts of ash and other inerts.34 it is crucial to note,
however, that the final gas products and their relative amounts are not
significantly affected, regardless of the actual type of cellulosic feed-
stock, as long as the same amount of total gas is produced29. Thus, ash
and inerts, roughly speaking, may catalyze to the degree of decomposition
of the initial intermediate without significantly altering the desired,
final step (the mixture of gas products).
Accordingly, the rate of decomposition of gases or the secondary
pyrolysis reactions thereof appear to be the crucial factors in generating
C2 and higher hydrocarbons. A low rate of decomposition of C£H4 is desired
and this appears to have been achieved in the S6FM reactor with a simple,
practical design. It is known that ethylene can be produced from coal38
and cellulosic^2 feedstocks at significant concentrations (> 5%), but a
sophisticated or high-vacuum apparatus has been generally required. Only
one other practical reactor design has been reported to achieve such concen-
trations to our knowledge and that is the process developed by Garrett
Research and Development for a sub-bituminous coal..39 Most other processes
have attempted only to maximize CO and H2 to provide the usual, versatile
synthesis gas. The SGFM process and the Garrett research offer a potentially
significant advance in the state-of-the-art because the results suggest that
valuable organics can be directly produced. Processes which produce only
CO and H2 obviously require additional reactors in order to provide organics.
A detailed analysis of the production of ethylene in the SGFM reaction is
presented below.
Based on the temperature profiles given in Figure 10 and the half-life
data of ethylene given in the previous section, one may then estimate,
33
-------
roughly, that the residence time of ethyl ene at temperatures greater than
800 C is on the order of 10 seconds or less, e.g., the half-life at 800°C
is 9.0 seconds at a 6 percent concentration. This estimate of residence time
does not, however, give a complete picture.
The maximum reported gas phase concentration of ethyl ene from a cellu-
losic feedstock or a derivative thereof is on the order of 20-25 percent^^O.
Allowances for dilution in the SGFM reactor indicate that the higher con-
centration may be 12 percent ^4. When this value is combined with an
isothermal, tubular reactor situation for the three-halves order reaction,
calculations show that 6 percent ethyl ene is obtained in 6.38 seconds at
800°C. It is conceivable that the true half -life is much less because the
decomposition of ethyl ene is still quite rapid down to 700°C.
To quantify the time-temperature history in the SGFM reactor more
distinctly, assume that the maximum, total gas generation occurs at the
maximum temperature and remains constant. Then using the upper limit of
gas velocities observed, the approximate, minimum residence time of gas
between 900°C and 600°C is
velocity = 21234 1 1000 cm 1 hr 1 atm 1123° K 4
hr
3600 sec 1 .36 atm 273°K 3.14 232cm
= 76.9 cm/sec
residence time =
°»
= 0.62 sec
This residence time is of the same order of magnitude of the calculated
ha If -life at 900°C at a concentration of 12 percent, 0.9 seconds. If the
local concentration near a particle is assumed to be the critical factor,
the ethylene concentration could approach 25 percent. For this case, the
calculated half-life is 0.623 seconds at 900°C.
These estimates have been based upon idealized estimates. When com-
bined with non-ideal features such as non-constant gas generation, backmixing
of gas, potential catalytic surface effects and diffusion from a particle
or droplet, then, these estimates on the time- temperature history of ethy-
lene help explain why other processes (except for those of Garrett) have
not yielded significant quantities of ethylene. The calculations also
indicate that the important feature of the SGFM reactor design/operation may
be the reasonably high velocity coupled with the parabolic temperature.
The previous estimates and discussion suggest the following physical
description. Cold manure at room temperature enters the reactor and is
heated by product gas to the initial reaction temperature of 250-300°C;
initial decomposition of the solid then gives some CO and 0)2; further heating
of the solid generates intermediates which decompose to hydrocarbon gases
34
-------
, C2H4, CgHs) and hydrogen. This valuable gas appears to be continuously
produced as solid enters the high temperature zone and is immediately swept
towards the feed section by the other generated gases, nitrogen, and steam.
The continual sweep action into the low temperature environment minimizes
further decomposition of the hydrocarbons to reduce the effective residence
time below the value indicated above (0.6-0.9 seconds). The remainder of
the solid then falls into the combustion part of the reactor to undergo
further reaction to a char product (it is possible that heavy droplets of
intermediate also enter this combustion zone).
The preceding calculations and description have been presented to
illustrate the critical time-temperature history that may have been developed
in the operation of the SGFM reactor. The calculations are, at best, only
approximate because total gas velocities at all points of the reactor are
unknown. The physical description must be considered preliminary. Never-
theless, these interpretations are thought to incorporate consistent reasoning
for explaining the observed production of ethylene (see also section
ETHYLENE FROM TAR/OIL PRODUCT) in the falling-bed, countercurrent design
of the SGFM reactor. The ideal case is an infinite or near infinite tem-
perature to achieve maximum manure pyrolysis to intermediates coupled with
zero residence time to minimize hydrocarbon (014, C2H4) decomposition.
In summation, it is thought that the combination of the high counter-
current velocity (23-97 cm/sec) with the parabolic temperature profile
(illustrated by Figure 10) has lead to the generation of the potentially
valuable product gases. Neither the operating velocity nor temperature
effects are novel in themselves nor do they appear to be particularly use-
ful by themselves. It is the operating combination that may be important
and permits: (a) a high conversion to be achieved to reduce the residue
volume; (b) a high rate of total gas generation to provide valuable gas
mixtures even with the dilution of nitrogen from air; and (c) a high yield
of ethylene to maximize the potential product sales value. Obviously,
additional studies are needed to confirm this hypothetical interpretation
of the data.
A patent disclosure on the ethylene production was filed with the
Environmental Protection Agency on July 17, 1975. A copy of this dis-
closure is given in Appendix 3.
SCALE-UP OF REACTOR DESIGN
The previous material has presented the results of the investigations
which strongly suggest that ammonia synthesis gas and ethylene can be pro-
duced simultaneously from cattle feedlot manure using the SGFM reactor design
and operation. This product mix indicates a promising process and one may
legitimately ask "How can the reactor be scaled-up to a higher throughput,
say, 200-500 tons per day for a demonstration plant?"
The question can be answered by the direct statement that the reported
results do not provide a sufficient data base for a sound, engineering design
or a reactor for the production of both products. To design a larger
35
-------
reactor on the reported data would be an extreme extrapolation of the ethylene
production technology. That is, the reactor was not originally designed to
produce ethylene. The production of significant quantities of ethylene is
a result of unanticipated operating conditions for a specific (or individual)
reactor designed to produce ammonia synthesis gas. The data needed to specify
residence time-temperature history in the SGFM reactor are currently unknown.
Design for Only Ammonia Synthesis Gas Production
The production of ammonia synthesis gas by itself appears to be relatively
simple and direct without any evidence of subtle factors. This statement is
based upon the results from the bench-scale reactor,24,25 the current pilot-
scale reactor, and fundamental microreactor studies.29 All of these studies
show that reaction temperature is the only critical variable for the pro-
duction of ammonia synthesis gas, providing feedstock particles are less
than approximately 1.5 mm with approximately 80 percent being less than
1 mm (0.0394-inch or approximately 1/32-inch). The crucial design factors
for the production of ammonia synthesis gas appear to be:
- Sufficient heat transfer to particles to assure that the entire
volume of individual particles reach a temperature of 850-900°C;
the time to heat particles to this temperature is sufficient for
reaction.
- Sufficient fluid phase residence time to "crack" or decompose
as much of the generated gases to hydrogen (H2) and/or carbon
monoxide (CO); again, the time required to heat the particles
appears to be sufficient.
- Adequate amount of nitrogen (N2) and oxygen (03) to achieve the
desired SHg:^ ratio and an authothermal or near-autothermal
operation.
These are the crucial factors and it is projected that the current results
have demonstrated these factors to a degree that existing engineering tech-
nology for solid wastes39,41,42,43 and general petrochemical processing
could be used to design and build a larger reactor. Some practical design
information and improvements are still needed, but a larger, working reactor
could probably be built and demonstrated after a reasonable shake-down or
start-up period. This conclusion should, however, be verified by an
independent engineering design review. The additional information required
includes:
- Detailed data on the sulfur content of product gas to design a
sulfur removal system.
- Reactor operation at 10-30 atmospheres to reduce compressor
costs in supplying product to an ammonia synthesis plant.
This should be accomplished in the larger plant with back-up
on bench-scale unit because such operation exceeds our
current capabilities; some high pressure investigations are
planned in the near future.
- Detailed heat and material balance investigations to support the
results obtained to date. This work is planned within the next
36
-------
6 months and would be important in establishing a larger plant
without overdesign.
- An analysis and treatment method for the waste water that is pro-
duced within the process.
Additional information that would be useful, but not necessary for design,
should include data on alternate or mixed-manure feedstocks. These feed-
stocks could include other residual cellulosics or biomass from the agri-
cultural or forestry sector such as cotton burrs and gin trash; sawdust;
field stubble; and corn and other stalks. It is important to note that
while information on these new feedstocks is not necessary for the reactor
design and scale up, the data may be needed to assure that a practical,
full-scale plant can be economically justified. That is, the cattle feedlot
industry is a highly volatile one and other cellulosics may be needed to
assure that feed to sustain a high volume operation (1000-2000 ton per day)
is always available within a 30-50 mile radius of the plant site. The use
of alternate or mixed feedstocks would also expand the realm of application
to other agricultural locales besides high-volume feedlots and is highly
recommended as a subject for further investigation.
Design for Ethylene and Ammonia Synthesis Gas Production
As noted, the information for the design of a reactor to produce ethylene
along with synthesis gas is not available: the data that specify the
existing ethylene residence time-temperature history. Part of these data
should be obtained within the near future and will include information on
the complete gas composition within the reactor. These internal, gas com-
position data will provide a direct knowledge of where the ethylene is formed.
Analysis of the data will also help specify the ethylene decomposition rate
by comparing the data to pure gas reaction rate data and velocity of the
total gas stream along the reactor using a nitrogen balance. This information
will be used to develop new operating conditions and techniques for improved
ethylene production. The new techniques could include re-injecting the liquid
tar product that is currently produced (see ETHYLENE FROM TAR/OIL PRODUCT,
page 44 ). To establish a firm data base, studies should also investigate
(1) the degree of solid and gas phase backmixing within the reactor;
(2) the hold-up volume of solid in the reactor; and (3) the operation of
a 2 or 3-stage set of true fluidized beds as opposed to the falling bed,
countercurrent operation.
Action Taken on Needed Design Data
A program has been funded by the U. S. Energy Research and Development
Administration to obtain detailed heat and balance data; internal gas com-
position data; and bench-scale data on several alternate feedstocks with
scale-up of one. This new program began in September 1976.
37
-------
MATERIAL AND HEAT BALANCES
The major effort over the past year was to develop a working reactor
system. During the course of the investigation, several mechanical problems
were encountered and solved; but, unfortunately, some of the problems
destroyed any possible chance of obtaining a complete material balance on
all runs. The best results are given below. The original equipment train
was designed to provide a reliable heat balance; but, again, to develop a
working system, much of the original design was modified (see EXPERIMENTAL
EQUIPMENT, PROCEDURES, AND MATERIALS) which nullified any accuracy and pre-
cision within the available time frame. Some calibration for heat balances
has been completed and the results are reported below.
Material Balance on Reactor and Gas Analysis
The two, most reliable material balance run data are given in Table 3.
The results show that 11-14 percent of the total mass input or output was
not measured at designated sample or collection junctions. There are
several possibilities for this discrepancy. After reviewing all data,
operating characteristics, and observing the tar build-up inside transfer
lines, it was concluded that a significant part of this discrepancy was due
to uncollected tar and solid fines in transfer lines, and, probably, solid
hold-up in the reactor and cyclone. Because the reactor and cyclone were
allowed to discharge solid 3-4 hours after stopping the manure feed, it is
believed that a majority of the unaccounted mass is tar deposited on the
inside of transfer lines. Preliminary estimates indicate that the tar
product has a density somewhat greater than water. Thus, for 5.08 cm ID
transfer lines, the estimated thickness over 914 cm of pipe for 1500 g of
unaccounted material is:
Tar Film = ^r- frnr^ m ,"£ L = 0.98mm
914 cm 3.14 5.08 cm
= approx. 1/32 in,
This coating thickness is well within the visual inspection of transfer line
tubing and supports the projection that a significant fraction of tar was
deposited on pipe walls. Further evidence of this observation is that a
fire started in the main transfer line between the cyclone and tar trap
during improper shutdown of Run 10 and was self-sustaining for several
minutes before being extinguished.
Because the gas product is of major interest, it is important to note
that nitrogen balances using measured gas flow and gas composition agreed
within approximately 10 percent after the first fifteen minutes of operation
with manure with a clean turbometer (Run 6) or after installation of the
final filter prior to the turbometer (Run 10). Some of the individual
measurements did, however, deviate by as much as 34 percent and for this
reason, the average of the directly measured data was used in the correlations
38
-------
Table 3. SELECTED MATERIAL BALANCE AND ASH DATA
CO
UD
Item
A. Material Balance
1. Input, kg
Wet Manure (10% moisture)
Air
Steam
Total Input, kg
2. Output, kg
Gas
H20 + Organics
Tar
Solid Fines (Cyclone)
Char
Total Output, kg
% Output/Input
% Unaccounted Output
B. Ash Content, wt. %
Manure Feed
Cyclone, Solid Fines
Char
8
18.05
1.72
3.06
22.83
5.74
6.44
0.23
0.63
7.03
20.07
87.91
12.09
21.54
55.50
50.24
Run Number
9
12.94
3.86
9.95
18.75
7.08
4.88
0.23
0.74
3.74
16.67
89.91
11.09
19.59
49.45
46.50
-------
discussed previously. Thus, the volume of total gas output does have some
uncertainty and this does affect the material balance in Table 3. The gas
rate uncertainty could be 0.5-1.0 kg in both of the data sets cited.
With regard to the reliability of gas composition analysis, it is
important to note, as discussed in the section Analytical Procedures (page
20), that 2 or more readings were obtained per gas sample and each reading
was the average of 2 or more injections. Also, the gas composition data of
Run 1 were analyzed on a different chromatograph by a different operator
and the calibration used for the reported gas composition was confirmed by
different operators using the SGFM instrument after 6 months operation.
An infrared scan from 400-4000 Kaysers (approximately 2.5-18 microns) on
a sample of condensed gas (N2 boiling point) from Run 1 showed that there
were no major components in the gas other than those listed in Table 1
except for a small amount of ^2 (< 0.5%). Thus, in general, it was
concluded that the reported gas composition is accurate; precision is not
estimated to be better than 2 significant figures.
Heat Balance on Reactor
As noted earlier, it was not possible to obtain reliable heat balances
during the runs due to extensive modifications. A recent calibration of the
reactor using only an air input has shown that the heat loss from the reac-
tor without feed is approximately 95 percent. This heat loss is largely
due to the relatively poor insulation at the top of the reactor which is
thought to provide the quenching effect noted earlier in discussing ethylene
production. Other data and calculations show that the convective (natural)
heat transfer coefficient from the reactor is 4-5 Btu/hr ft2 °F, which is
the expected range of values for natural convection from a vertical cylinder.
These results suggest that the reaction as operated is potentially
near the desired autothermal case. Additional data are required before
this can be ascertained.
NON-GAS PRODUCTS AND OPERATION DOWNSTREAM OF REACTOR
Reactor Gas and Cyclone Tar
The gas phase exiting the reactor contained 5-21 percent of the total
collected char, by weight. This solid consisted of particles with diameters
ranging from 250 ym-20 ym as removed in the cyclone with smaller sizes
being entrained further downstream.
The appearance of the exit gas (before treatment) was that of a dense
smoke with a greenish-yellow tinge and a very unpleasant odor, something
approaching that of wet manure itself. This description of the gas is
apparently the "standard" product because we soon learned to associate the
odor with good operation, until most of the leaks were eliminated.
The gas also contains tar which apparently makes up the major fraction
of the aerosol or smoke. This tar begins to condense at approximately 320°C
40
-------
and this was the basis for operating with a heated cyclone. With the cyclone
maintained at approximately 300°C, no significant condensation of the tar
occurred and a free-flowing solid was discharged.
Tar and Condensation
As noted, tar began to condense at approximately 300°C and product was
obtained down to room temperature. Most of the collected tar was obtained
at 100-120°C but it would appear (Material Balance) that a significant
fraction condenses above this temperature and that a new condenser is
needed to collect all tar. One has been designed and constructed, but
not tested.
The tar collected at different temperatures has a noticeably different
appearance. The high temperature fraction tends to form a grainy, discrete
deposit at room temperature; its viscosity is greater than molasses at 100°C.
The initial visual observation suggests that this material is simply entrained
fines from the cyclone with some condensed tar. However, the ash content of
this fraction is less than 1 percent and the material is volatile.
The low temperature fraction (the fraction collected at 100-120°C has
the consistency of taffy at room temperature. At the collection temperature,
the fraction flows easily with a viscosity somewhere between water and
molasses. The fraction has an odor which approximates that of crude oil,
without hydrogen sulfide.
The tar impinger used to collect tar at 100-120°C worked very well and
removed approximately one pound of tar per hour. This design could be used
for a final tar collection in future reactors following collection units
which condense the tars with apparent boiling points greater than 120°C.
Water Condenser and Product
The product stream from the water-cooled condenser was a brown to yellow,
murky, foul-smelling liquid with a high sediment content. Over a period of
several weeks, the liquid separated into three phases: a foamy (organic?)
layer, a dark, watery layer, and a layer of sediment. This "waste" water
could be a severe pollutant, but current projections are that the organics
could be recovered by passing the stream through a bed of the char product
(similar to an activated charcoal bed). The bed solid plus adsorbed matter
could then be burned to produce steam or mixed with fresh manure feedstock
to recover valuable carbon and hydrogen. A gradient-temperature condensation
system below 100°C might permit a large fraction of the organic layer to
be collected separately. Either of these alternatives would minimize the
pollution potential. Further investigations on this procedure are needed
to develop a complete process design.
In general, the operating problems caused by condensable products (tar,
aqueous mixture, etc.) from manure pyrolysis must be experienced to be
appreciated fully by any investigator. This conclusion has been confirmed
by other investigators, and specifically during manure pyrolysis.4°
41
-------
Reactor Char, Particle Size, and Mass Reduction
The main fraction of the produced char (80-95 percent) exited from the
reactor as a free-flowing solid material with an approximate average particle
size of 70 urn. The solid burns quite easily and, in some cases, the problem
is how to keep the hot solid (200-300°C) from burning when exposed to the
atmosphere. For example, hot char was once dumped into a bucket with a
plastic bag. A low-level surface fire developed. Part of the solid was
dumped on a cold steel plate in a 5-8 cm pile. This material, under con-
trolled conditions, continued to burn for several hours, without a noticeable
smoke or odor.
A supplemental, combustion test of the char was run in a two-stage,
commercial furnace used to burn waste cotton.47,48 The brief test indi-
cated that there were no problems except for the expected high ash frac-
tion (50 percent).
The distinction between particle size of the various runs in Table 2
and Figures 7, 8, and 9 has been made because product char particles with
effective diameters greater than 0.318 cm were visually observed to be
grossly underreacted. A representative sample was sieved and the ash content
of the various fractions was obtained to confirm the visual observation.
These results (Table 4) indicate that approximately one-third or more of
the original volatile mass remained for those particles greater than 2 mm
(approximately 1/12-inch) compared to 15-20 percent for particles less than
1.5 mm. The calculation for these estimates in Table 4 is as follows:
A = mass of ash = constant a = original ash fraction
V = original volatile mass 3 = final ash fraction
o
Vp = final volatile mass
A
- a VF + A
= 3
V VF
TV= ! -a V + A = ] ' 3
T f\ • C **
F A 3 _ 1 - B
V Vr + A V^aA 1-a
vo F
VF . a(l - 3
VQ 3(1 - a
42
-------
Table 4. EFFECT OF PARTICLE SIZE ON ASH CONTENT
Particle Size
A. Feed
> 2 mm
1.41-2.0 mm
1.05-1.4 mm
0.71-1.05 mm
< 0.70 mm
B. Char Product
> 2 mm
1.41-2.0 mm
1.05-1.4 mm
0.71-1.05 mm
< 0.70 mm
Cyclone (< 0.70 mm)
(a)
% Ash
19.6-21.5
41
56
55
54
59
60-61
Approximate
% Volatiles
Remaining
100
36
20
20
21
17
16
(b)
(b)
8
100 x Vp/V ; a = 0.20 was used in these calculations
43
-------
iii ho w«iH-5"995SJ that Partlcles less than 1.5 mm (approximately 1/16-inch)
will be volatilized to a reasonable degree (80 percent) and that excessive
grinding of manure is not required for the SGFM process. The use of large
particle sizes is actually desired to minimize carry-over. The apparent lack
of effect of the 0.318 cm sieving noted in Table 1 is due to the fact that
80 percent of the particles as run are less than 1.5 mm.
These same results show that the reduction in total mass of the
original solid waste is = 100 - ((V> + A) * (V0 + A) x (100)) = 100 - a * 3
x 100 = 51-68 percent. The average reduction is approximately 60 percent
for Run 8 and a range of 55-65 percent was observed throughout the investi-
gation. The results also show that 60-85 percent, with an average value of
80 percent of the volatile mass that can be treated has been accomplished.
It is anticipated that the proposed process would use the char product for
steam regeneration or other uses; hence, 100 percent of the volatile mass
would be eliminated with only ash to be disposed. Mackenzie^ an
-------
Table 5. PRELIMINARY RESULTS FROM TAR PYROLYSIS
Maximum Reaction3
Temperature
°C
700
800
900
Products Obtained, % of Sample Fed
Total Gas Residue Ethyl ene Methane
9.2-14.8 52.7-63.9 1.5-3.3 0.9-1.6
18.4-24.7 40.8-56.9 4.4-7.13 1.8-2.9
27.2-33.2 33.4-46.3 7.6-10.6 3.2-4.8
aThis temperature is thought to be indicative of the operation of the large-
scale S6FM reactor.13'29
45
-------
gas residence times. Individual runs have been duplicated within ± 7 percent
relative error for the results cited. These initial data show that tar
reinjection may be very attractive because ethylene yield from the tar at
900°C is comparable to that for manure itself, 0/075-0.106g/g of daf feed.
The data also show that similar levels of total tar conversion may be
obtained or approximately 50-65 percent. The ethylene concentration in
the gas phase itself is in the neighborhood of 25 percent molar and signi-
ficant concentrations (3-5 percent) of another valuable olefin, propylene,
have also been observed. Current correlations among the data indicate
that the tar pyrolysis/reinjection is compatible with the large-scale SGFM
reactor, i.e., maximize operating temperature (see Table 5 where conversion
at 900°C > 600°C > 700°C).
In addition to the practical implications of these preliminary data,
the results also strongly suggest that manure pyrolysis actually proceeds
through a low molecular weight intermediate (tar) step which is in agreement
with basic studies on pure cellulose, wood, and municipal solid waste2^>32»
34,35,50,51,52. The results which support (not confirm] this mechanism are
(a) ethylene yield is comparable to manure based yield13; (b} product gases
are similar to solid pyrolysis data, including CO and C02"'3 ' , and
(c) maximum reaction temperature is comparable to that for solid manure
samples at comparable residence times for similar conversion levels
13,50,51,52. However, the production of combined propylene and butenes
(5-10 percent in gas phase) exceeds that previously measured in basic
solid pyrolysis studies or from the large-scale reactor. These compounds
will decompose at reaction temperature if the residence time is long enough.
Thus, their absence in the gas from the large-scale SGFM suggests the
importance of gas phase, time-temperature history or possibly the importance
of decomposition within a solid particle. Thus, one may project that
further optimization of time-time-temperature history in the large-scale
reactor would be required to recover, say, propylene.
Most of the work on the tar pyrolysis project is directed to esti-
mating the effects of tar and gas phase residence time on gas yield and
composition. The results will be used to estimate the point(s) of tar injec-
tion in the large-scale SGFM reactor and as supporting data in developing a
consistent model of the manure reactor kinetics.
46
-------
REFERENCES
1. Uvacek, E., "Economics of Feedlots and Financing," pp 11-27, The Feed-
lot, I. A. Dyer and C. C. O'Mary, Ed., Lea and Febiger, Philadelphia
1972.
2. Agricultural Development Department, Southwestern Public Service Com-
pany, "Cattle Industry Served By Southwestern Public Service Company,"
1974.
3. ibid., 1973.
4. Bickel, Glenn D., Agricultural Consultant, Southwestern Public Service
Company, Personal Communication, May, 1976.
5. Hart, S. A., "Manure Management," pp 162-175, The Feedlot, I. A. Dyer
and C. C. O'Mary, Ed., Lea and Febiger, Philadelphia, 1972.
6. Whetstone, 6. A., Parker, H. W., and Wells, D. M., "Study of Current and
Proposed Practices In Animal Waste Management," pp 7-17, Report EPA
Contract No. 68-01-0875, Water Resources Center, Lubbock, Texas, 1973.
7. Sweeten, John M., "Energy From Feedlot Manure," Agricultural Engineer-
ing, Texas Agricultural Extension Service, College Station, Texas, AENG
6, October 21, 1974.
8. Agricultural Development Department, ibid., 1974.
9. Heichel, G. H., "Agricultural Production and Energy Resources," Amer.
Scientist, 64_, 64-72 (1976).
10. Sweeten, J. M., Reddell, D. L., and Stewart, B. R., "Feedlot Manure As
An Energy Source," 1974 Texas Section Meeting, Amer. Soc. Agric. Engi-
neers, Abilene, Texas, October 4-5, 1974.
11. Shultz, H. W., "Energy From Municipal Refuse: A Comparison of Ten
Processes," Prof. Engineer, 45(11), 20-24 (1975).
12. Schmid, L. A., "Feedlot Wastes to Useful Energy - Fact or Fiction," J.
Environ. Engineering, Proc. Amer. Soc. Civil Eng., 101(EE5). 787-793
(1975).
47
-------
13. Huffman, W. J., Peterson, R. L., and Halligan, 0. E., "Ammonia Synthesis
Gas Generation From Cattle Feedlot Manure," Texas Tech University,
Lubbock, Texas, Centennial ACS Meet., Div. Fertil. Soil Chem., April
6-7, 1976.
14. Han, Y. W., Dunlar, C. E., and Callihan, C. D., "Single Cell Protein
From Cellulosic Wastes," Food Technology, 25(2), 32-35 and 56 (1971).
15. Boersma, L. and Barlow, E. W. R., "Animal Waste Conversion Systems
Based On Thermal Discharges," Agric. Exp. Station, Oregon State Univ.,
Corvallis, Oregon, Special Report 416, September, 1974.
16. Elias, H. G., "Raw Materials For the Polymer Industry, Part II, The
Unconventional Situation," Chem. Tech, 1976.
17. Anonymous, "IGT Weighs Potential Of Fuels From Biomass," Chem. Eng. News,
24-26, Feb. 23, 1976.
18. Chancellor, W. J. and Goss, J. R., "Balancing Energy and Food Produc-
tions, 1975-2000," Science, 192, 213-218 (1976).
19. Whetstone, G. A., "A Viable Path Towards Biomass Utilization," Texas
Tech University, Lubbock, Texas, Fourth Ann. Meet., Biomass Energy
Institute, Inc., Winnipeg, Manitoba, Canada, Sept 29, 1975.
20. Parker, H. W. and Whetstone, G. A., "Fuels and Petrochemicals From Agri-
cultural Wastes," Texas Tech University, Lubbock, Texas, Paper 5e, 76th
Nat. AIChE Meet., Tulsa, Okla., March 10-13, 1974.
21. Halligan, J. E. and R. M. Sweazy, "Thermochemical Evaluation of Animal
Waste Conversion Processes," Texas Tech University, Lubbock, Texas,
72nd Nat. Meet., AIChE, St. Louis, May 21-24, 1972.
22. Engler, C. R., Walawender, W. P., and Fan, L. T., "Synthesis Gas From
Manure, Conceptual Design Study and Economic Analysis," Environ. Sci.
& Tech., 13(9), 1152-1157 (1975).
23. Halligan, J. E. and Huffman, W. J., "Potential For Solid Waste As An
Energy Source in Texas," Texas Tech University, Lubbock, Texas, Pro-
ject N/T 10, Governor's Energy Advisory Council, Final Report, November
1974.
24. Halligan, J. E., Herzog, K. L., Parker, H. W., and Sweazy, R. M.,
"Conversion of Cattle Feedlot Wastes To Ammonia Synthesis Gas," EPA-
660/2-74-090, U.S. Environmental Protection Agency, Corvallis, Oregon,
97330, December, 1974.
25. Halligan, J. E., Herzog, K. L., and Parker, H. W., "Synthesis Gas From
Bovine Wastes," Ind. Eng. Chem., Proc. Des. Dev., 14(1), 64-69 (1975).
48
-------
26. Wideman, W. S., Halligan, J. E., and Parker, H. W., "Preliminary Flow
Sheet And Economics For Production of Ammonia Synthesis Gas From Manure,"
Texas Tech University, Lubbock, Texas, 76th Nat. Meet., AIChE, Tulsa,
Okla., March 10-13, 1974.
27. Mark, H. P. , McKetta, J. J., and Othmer, D. F., Ed., "Manufactured Gas,"
pp 359-372, Kirk and Othmer Encyclopedia of Chemical Technology, 2nd
Ed., 10, John Wiley and Sons, N.Y. , N.Y., 1963.
28. Sloan, C. R. and McHone, A. A., "The Effect of the Energy Crisis On Ammonia
Producers," pp 91-95, Ammonia Plant Safety, 15, Chem. Eng. Prog., AIChE,
N.Y., N.Y., 1973.
29. Young, H., Unfinished Ph.D. Dissertation, Department of Chemical Engi-
neering, Texas Tech University, Lubbock, Texas, 79409, January 1976.
30. Kungui, T., Sakac, T., Soma, K., and Sasai, Y., "Kinetics and Mechanism
of the Thermal Reaction of Ethylene," Ind. Eng. Chem., Fund., 8(3),
374-383, 1969.
31. Wishart, R. S., "Petrochemical-Feed Outlook Seen Shaky," The Oil and
Gas Journal, pp. 48-50, May 19, 1975.
32. Nikitin, N. I., pp. 585-594 in Chemistry of Cellulose and Wood, N. I.
Nikitin, Ed., (Trans. By J. Schmorak), Israel Program For Scientific
Translations, Jerusalem, Israel, 1966.
33. Garner, W. and Smith, I. C., "The Disposal of Cattle Feedlot Wastes By
Pyrolysis," E.P.A. Contract 14-12-850, Midwest Research Institute,
Kansas City, Mo., 1973.
34. Shafizadeh, F., "Pyrolysis and Combustion of Cellulosic Materials," Ad-
vances in Carbohydrate Chemistry, 23, Wolfrom, M. L. and Tipson, R. S.
(Ed.), Academic Press, N.Y., N.Y., pp 419-474, 1968.
35. Shafizadeh, F., Mclntyre, C., Lundstrom, H., and Fu, Y-L., "Chemical
Conversion of Wood and Cellulosic Wastes," Proc. Montana Acad. Sci.,
33, 65-96 (1973).
36. McFarland, J. M., Glassey, C. R., McGauhey, P. H., Brink, S. A., and
Golucke, C. G., "Comprehensive Studies of Solid Waste Management,"
EPA 2RO-1-EC-00260-05, Sanitary Eng. Res. Lab., College of Eng. and
School Pub. Health, Univ. California, Berkeley, Final Report, 1972.
37. Feldkirchner, H. W. and Huebler, J., "Reaction of Coal With Steam-Hydrogen
Mixtures at High Temperatures and Pressure," Ind. Eng. Chem., Proc. Des.
Dev., 4(2), 134-142 (1965).
38. Coates, R. L., Chen, C. L., and Pope, B. J., "Coal Devolatilization in a
Low Pressure, Low Residence Time Entrained Flow Reactor," pp 92-107,
Coal Gasification, L. G. Massey (Ed.), Adv. Chem., Series, No. 131,
Amer. Chem. Soc., Washington, D.C., 1974.
49
-------
39. McMath, H. G., Lumpkin, R. E., Longanbach, J. R. and Sass, A., "A Pyrolysis
Reactor For Coal Gasification," Chem. Eng. Prog., 70(6). 72-73, June, 1974.
40. De La Garza, E. M., "Pyrolysis Products from Manure Tars," M. S. Thesis,
Department of Chemical Engineering, Texas Tech University, Lubbock, Texas,
79409, December, 1976.
41. Burton, R. S. and R. C. Bailie, "Fluid Bed Pyrolysis Solid Waste
Materials," Combustion, 13-18, Feb. 1974; also AIChE Sym. Ser. No. 122,
68, 140(1972).
42. Weinstein, N. J. and Toro, R. F., Thermal Processing of Municipal Solid
Waste For Resource and Energy Recovery, Ann Arbor Science, Ann Arbor,
Mich., 48106, 1976.
43. National Center For Resource Recovery, Inc., Resource Recovery From
Municipal Solid Waste, A State of the Art Study. Lexington Books, D. C.
Heath and Company, Lexington, Mass., 1974.
44. Wen, C. Y., Bailie, R. C., Lin, C. Y., and O'Brien, W. J., "Production
of Low BTU Gas Involving Coal Pyrolysis and Gasification," pp 9-28,
Coal Gasification, Op. Cit.
45. Levenspiel, 0., Chemical Reaction Engineering, 2nd Edition, John Wiley
and Sons, Inc., N.Y., N.Y., 1972.
46. MacKenzie, J. D., "Conversion of Cattle Manure Into Useful Products,"
Third Quarterly Report, EPA Grant No. R802933-01, September-December,
1974, Environmental Protection Agency, Nat. Animal Wastes Res. Program,
Ada, Okla.
47. Arnett, D. B., "The Disposal of Oil-Cotton Mixtures," M.S. Thesis,
Library, Texas Tech University, Lubbock, Texas 79409, 1976.
48. Peterson, R. L., "A Pilot Plant Study of Synthesis Intermediates From
Feedlot Wastes," M.S. Thesis, Library, Texas Tech University, Lubbock,
Texas, 79409, December, 1975.
49. Sweazy, R. M., Committee Chairman, M. S. Thesis, Khara, B. H.,
"Utilization and Disposal of Residue From the Partial Oxidation of
Cattle Manure," M. S. Thesis, Library, Texas Tech University, Lubbock,
Texas, 79409, 1975.
50. Maa, P. S. and Bailie, R. C., "Influence of Particle Sizes and Environ-
mental Conditions On High Temperature Pyrolysis of Cellulosic Material-I",
Comb. Sci. Tech., 7_, 257-269 (1973).
51. Kanury, A. M., "Thermal Decomposition of Wood Pyrolysis," Comb, and
Flame, 18, 75-83 (1972).
52. Kung, H. G., "A Mathematical Model of Wood Pyrolysis," Comb, and Flame,
1_8, 185-195 (1972).
50
-------
APPENDIX A
51
-------
TABLE A-l
OPERATING CONDITIONS FOR INDIVIDUAL RUNS AND SAMPLES
OPERATING CONDITIONS RUN AND.GAS SAMPLE NUMBER
1-4 1-5 6a-l 6a-2 6a-3 6b-6 6b-7 7-2 7-3 8-3
Manure Feed Rate, As Received
kg/hr
Ib/hr
Manure Feed Rate, Dry-Ash Free
kg/hr
Ib/hr
Air Feed Rate
kl/hr
SCF/hr
Steam Feed Rate
kg/hr
Ib/hr
Particle Size, less than
cm
inch
Raw Gas Production Rate
1/g daf
SCF/lb daf
Run Time With Manure, hr
7.76
17.1
5.22
11.5
1.15
40.5
5.45
12
0.95
3/8
1.19
19.1
0.6
7.76
17.1
5.22
11.5
1.15
40.5
5.45
12
0.95
3/8
1.19
19.1
10.
23.
7.
15.
1.
60
4.
10
0.
6
3
22
9
70
54
95
3/8
0.
(1
1.
693
1.1
13
10.6
23.3
7.22
15.9
1.70
60
4.54
10
0.95
3/8
0.645
) (10.4)
10.6
23.3
7.22
15.9
1.70
60
4.54
10
0.95
3/8
0.562
(9.0)
23
52
16
35
1
52
4
10
0
.6
.0
.2
.6
.49
.5
.54
.95
3/8
0
8
1
.544
.71
.0
23.6
52.0
16.2
35.6
1.49
52.5
4.54
10
0.95
3/8
0.613
9.83
19.0
41.8
13.0
28.6
1.90
67.2
3.63
8
0.95
3/8
0.480
7.69
1.6
19.0
41.8
13.0
28.6
1.90
67.2
3.63
8
0.95
3/8
0.327
5.24
18.1
39.8
12.4
27.2
1.40
49.5
3.09
6.8
0.95
3/8
0.467
7.49
2.3
-------
TABLE A-l (continued)
en
CO
OPERATING CONDITIONS
Manure Feed Rate, As Received
kg/hr
Ib/hr
Manure Feed Rate, Dry-Ash Free
kg/hr
Ib/hr
Air Feed Rate
kl/hr
SCF/hr
Steam Feed Rate
kg/hr
Ib/hr
Particle Size, less than
cm
inch
Raw Gas Production Rate
I/kg daf
SCF/lb daf
Run Time With Manure, hr
8-4 8-5
9-1
RUN AND GAS SAMPLE NUMBER
9-2 9-3 9-4 10-1
10-2 10-3 10-4
18.1
39.8
12.4
27.2
1.40
49.5
3.09
6.8
0.95
3/8
0.536
8.59
18
39
12
27
1
49
3
6
0
.1
.8
.4
.2
.40
.5 1
.09
.8
.95
3/8
0
8
.527
.44
18.0
39.7
12.7
28.0
4.25
18.0
39.7
12.7
28.0
4.25
18.0
39.7
12.7
28.0
4.25
50 150 150 1
2.72
6
0.32
1/8
0.680
(10.9)
2.72
6
0.32
1/8
0.705
(11.3)
2.72
6
0.32
1/8
0.705
(11.3)
18.0
25.9
12.7
28.0
4.25
50
2.72
6
0.32
1/8
0.761
(12.2)
11
25
8
18
0
27
3
8
0
.8
.9
.26
.2
.765
.72
.2
.32
1/8
0
7
.468
.5
11.8
25.9
8.26
1.8.2
0.765
27
3.72
8.2
0.32
1/8
0.268
4.3
11.8
25.9
8.26
18.2
0.765
27
3.72
8.2
0.32
1/8
0.237
3.8
11.8
25.9
8.26
18.2
0.765
27
3.72
8.2
0.32
1/8
0.306
4.9
2.1 2.1
-------
en
.TABLE A-2
GAS COMPOSITION DATA
(All Compositions Are Mole Percent Rounded To First Decimal)
RUN AND SAMPLE NUMBER
ponent
H2
N2
CH4
CO
co2
C2H4
C2H6
i. Mol. Wt.
1-1
23.1
22.2
11.4
10.4
29.3
3.4
0.2
25.8
1-2
27.4
7.1
14.2
12.8
32.3
5.9
0.4
24.4
6a-l
28.7
15.2
10.6
17.9
20.8
5.1
1.8
22.7
6a-2
21.6
28.7
8.0
16.5
18.9
6.4
0.7
24.6
6a-3
23.9
33.0
9.3
15.3
16.2
2.4
-
23.3
6b-6
19.0
17.2
12.3
20.6
22.6
8.2
-
25.2
6b-7
21.0
12.9
13.0
22.0
21.7
8.8
0.7
24.5
7-2
29.0
20.6
8.9
17.9
16.0
4.7
2.9
22.0
7-3
27.4
25.7
9.4
14.9
14.9
5.2
2.6
23.2
8-3
19.2
29.9
11.9
21.3
11.1
6.0
0.6
23.3
Total Energy, HHV
kcal/1 2.64 3.51 3.48 2.99 2.43 3.59 3.97 3.45 3.34 3.42
BTU/SCF 297 395 391 336 273 404 447 388 376 384
-------
TABLE A-2 (continued)
(All Compositions Are Mole Percent Rounded To First Decimal)
RUN AND SAMPLE NUMBER
Component 8-4 8-5 9-1 9-2 9-3 9-4 10-1 10-2 10-3 10-4
H2 17.6 15.4 14.7 11.5 16.3 17.7 22.1 20.1 20.4 21.0
N2 22.6 27.8 38.7 37.5 37.6 34.8 26.0 24.3 23.9 22.4
CH4 15.3 15.1 8.6 9.3 8.5 8.9 11.0 11.5 11.6 12.7
CO 22.4 19.7 19.5 20.6 19.7 20.5 21.1 23.2 23.5 21.9
S C02 15.7 15.4 13.7 15.8 13.6 13.6 14.3 14.9 14.6 15.5
C2H4 5.7 5.8 4.2 4.7 3.9 4.0 5.2 5.4 5.5 6.0
C2Hg 0.8 0.7 0.5 0.6 0.4 0.5 0.3 0.6 0.6 0.5
Avg. Mol. Wt. 24.1 23.8 25.4 26.5 24.9 24.5 23.2 23.8 23.7 23.5
Total Energy, HHV
kcal/1 3.67 3.50 2.57 2.66 2.56 2.68 3.18 3.41 3.35 3.50
BTU/SCF 412 394 289 299 288 301 358 373 377 394
-------
AVERAGE USING
TRAPEZOIDA RULE
TABLE A-3
TEMPERATURE PROFILE OF INDIVIDUAL RUNS
(All Values Are °C And Are Average Values)
REACTOR LOCATION/RUN
0 GAS INLET
1
2
3
4
5
6
7
8
9
10
11
12
13 GAS OUT
1
300
565
740
825
875
860
930
930
855
720
590
460
425
325
6a
300
720
815
845
845
890
915
895
820
670
560
405
365
295
6b
300
625
690
735
780
840
865
860
765
630
520
370
355
295
7
300
660
690
690
730
785
825
815
730
595
490
365
355
315
8
300
570
685
740
770
835
865
840
765
625
515
355
325
285
9
300
845
860
840
825
845
845
845
690
-
485
380
370
320
10
300
595
680
735
765
815
815
775
775
635
525
385
365
300
711
694
641
617
629
668
628
-------
APPENDIX B
57
-------
A MODEL FOR REACTOR OPERATING CHARACTERISTICS
To highlight some of the operational features (not design) of the re-
actor discussed in the paper, a preliminary model is developed below. The
model is based on several key assumptions or approximations which may not be
entirely correct for design purposes of analysis of reactor dynamics, but
will be sufficient for an assessment of operating features. The assumptions
are noted where appropriate.
The basis for the model is the kinetic rate equation that has been used
to describe the strict pyrolysis of manure particles, as developed by Young29
and several other investigators for other cellulosics.'1'* In the actual re-
actor, both pyrolysis and oxidation occur, but it is assumed that oxidation
is complete within 30 to 60 cm of the air inlet (reactor bottom) because the
feed air rate is much less than the stoichiometric quantity.45 Thus, most
of the reaction products of the top fed manure are assumed to be generated
from pyrolysis.
For the assumed pyrolysis reaction, the rate equation is given by equa-
tions (la) or (Ib), using nomenclature similar to the above references:
dt
(la)
or
where
M
k
k
E
R
T
X
$ = k (1-X) (Ib)
mass of volatile or reactive manure, daf basis
rate constant = kQe -E/RT
pre-exponential factor, time"
energy of activation, cal/mole
gas-law constant, cal/mole, °K
absolute temperature, °K
mass conversion, mass of daf manure converted per unit mass
of daf manure feed
58
-------
t = time.
It is desirable, however, to express equation (la) in terms of individual
particles where Np is the total number and Mp is the mass of an individual
particle. Equation (la) may then be written as equation (2).
d NM
-£* • -" w <2>
If Np is assumed to be constant (i.e., no break-up or agglomeration of parti-
cles), then using the volume of a particle, Vp, and assuming all particles
are the same size, equation (2) may be written
1 d Mp
V" dt = " k Cm ^3'
where C = concentration of reactive mass per unit volume of parti-
cles; Cm = Mp/Vp.
Equation (3) or (4) represents a definition of reaction rate of a manure par-
ticle using the particle volume as the reaction volume*5; that is,
1 d M
rm = r- dt (4)
where r = reaction rate based on particle volume.
It is noted, at this point, that it is not necessary to assume that the par-
ticle volume, Vp, is a constant.
To describe the pyrolysis reaction occurring, a differential cylinder
at any point in the reactor may be chpsen. Then a steady-state material
balance may be written as follows:
("V • FmodX
1
where R = rate of reaction based on reactor volume
V = reactor volume
r
Fm« = feed rate of daf manure, Ib/hr.
For a constant number of particles, Np, the relationship between t^ and
is:
59
-------
R - 2 ,fi ,
Km " Vr dt ' (6a)
• Vp/Vv <6b>
• TV <6c>
The term y = Npyp/Vr is the volume fraction of particles existing in the re-
actor at any point. Combining equations (6c) and (5), one can develop
(Cmo Y/Fmo) < Vr = -(1/k) d (in Cj (7)
where CmQ = initial (or feed) concentration of reactive manure per unit
particle volume.
The development of equation (7) requires that Vp be considered a constant.
This assumption is probably not a good assumption for conversion levels ex-
ceeding 80 percent, but does appear to be a reasonable approximation up to
this level, especially considering the high fraction of ash in manure (20-
25 percent; 29). The assumption of constant Vp permits equations (8) and
(9) to be used in developing equation (7).* The term, Fm, is the feed rate
which corresponds to the particle concentration, Cm,
x • i
cn/cmo = Wm ' o (9)
in the reactor at any given point. Equation (7) may be integrated directly
to obtain a result which is descriptive of a general tubular reactor con-
taining a first order reaction. This result is given by equation (10), as-
suming that an average value of the product of the rate constant and the
volume fraction can be defined.
(& Cmo VFmo - - ln <10)
where (yT) = [ yk d Vr]/Vr-
When y is constant and the reaction is isothermal, the left-hand side of
equation (10) can be rearranged to yield, after dividing by k:
*The last equality in equation (9) does not require the assumption of con-
stant volume; Np must, however, be constant.
60
-------
T' - * C«, VFmo ' NpVFmo"
where T' is the residence time for the ideal case of constant y and k as
noted. However, equation (10) demonstrates that T' is only an apparent
value for the current reactor design because temperature was not constant
and y is not expected to remain constant.
A closer approximation to_the actual T would be to define a term (k)
and assume that (Icy) = (F) (y) so that:
In other words, the actual residence time will depend upon the flow regime(s)
in the reactor, through y. This result is not unexpected, but it must be
appreciated^ because the product distribution could also be altered by a
change in y. In equation (11), the rigorous average value also demonstrates
the close association between reaction temperature and particle distribution
within a reactor using a solid feedstock with these kinetics. It also sug-
gests a reason for part of the differences in product distribution, at con-
stant CmoVr/Fmo> between the tubular reactor developed here and that of a
fluidized bed or that of a similar tubular reactor with co-current flow.39,'"'
It is also important to realize that no attempt has yet beer^ made to alter
the flow regime in the reported reactor which would change y, and poten-
tially alter the product rate and/or distribution.
The volume fraction term (or 1-e, where e is the void fraction) can be
used to illustrate a probable limiting factor of the reactor operation: heat
transfer to the particle. At the temperatures involved, radiation to a par-
ticle is probably the dominant mechanism.'"' Thus, there probably exists a
critical y, yc, beyond which the expected or desired X can_no longer be ac-
hieved, as the reactor is now operated. Or, at values of y greater than
Yc> all particles will not be exposed to the same radiant energy transfer
from the wall. In this case, interparticle radiant energy transfer (or
perhaps conduction/convection, if the volume fraction approaches that of a
packed bed) would become a key factor. It may be argued that the effect of
y would simply be reflected in the measured temperature. However, the point
being discussed here is that an average temperature can probably be imposed
(or held constant over a range of Y) and that X may still be altered due to
the change in heat transfer from the wall surface case to interparticle
effects. Over the three-fold range of feed rates explored to date, this
potential limitation does not appear to have been approached.
The y effect also suggests that the reactor performance may depend upon
the manner in which the reactor is started up. For example, if a large in-
ventory of solids is allowed to accumulate in the reactor before solids dis-
charge is matched to solids feed, a significant bed (large y) could exist at
steady-state which would reduce heat transfer to particles as discussed
above and thereby affect X. Again, this limit does not appear to have oc-
cured as the reactor is currently operated, over the range of data explor-
ed.
61
-------
Young29 has also shown that at a given level of conversion over a wide
range of heating rates (50°C/min to 1000°C/min), a certain fraction (6) of
the mass loss will always exist as a low molecular weight gas (boiling point
less than 25°C at 1 atm) when the fluid phase residence time is generally
less than 2 seconds. Because the operation of the pilot-scale reactor ap-
proximates these conditions, a similar effect may be expected. The gas frac-
tion obtained from the reactor may then be expressed in terms of the 6-
fraction as follows:
M
po
SX = 6 M.
po
= M
po
= F,
mo
(13)
where M,, = mass (or SCF) of gas produced
FG = flow rate of gas produced.
This result serves to illustrate that FQ/FTO is a correct correlation vari-
able which is independent of particle volume assumptions. For this reason,
the gas yield has been expressed as standard I per kg daf manure feed in
Figure 7 of the text. A similar analysis may be used to demonstrate that
any fraction of the produced gas should be expressed on a similar basis, as
done in Figures 8 and 9.
62
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-78-026
2.
3. RECIPIENT'S ACCESSIOr*NO.
4. TITLE AND SUBTITLE
CONVERSION OF CATTLE FEEDLOT MANURE TO ETHYLENE
AND AMMONIA SYNTHESIS GAS
5. REPORT DATE
February 1978 issuing date
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
William J. Huffman, James E. Halligan, and
Roger L. Peterson
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Texas Tech University
Lubbock, Texas 79409
10. PROGRAM ELEMENT NO.
1HB617
11. CONTRACT/GRANT NO.
S-802934
12. SPONSORING AGENCY NAME AND ADDRESS
Robert S. Kerr Environmental Research Laboratory--
Office of Research and Development
U.S. Environmental Protection Agency
Ada, Oklahoma 74820
Ac
13. TYPE OF REPORT AND PERIOD COVERED
Final (1/1/74 - 6/30/77)
14. SPONSORING AGENCY CODE
EPA/600/15
15. SUPPLEMENTARY NOTES
16. ABSTRACT
An investigation was undertaken to scale-up a partial oxidation reactor to
evaluate the potential for producing an ammonia synthesis gas from cattle feedlot
residue. The synthesis gas from the reactor can be further processed to produce
ammonia using existing, commercial technology. During this phase of the study,
a 2.5 meter long, falling-bed reactor was designed, constructed, and operated
along with supporting feed and recovery equipment; manure feed rates to the
reactor ranged from 7.7 to 23.6 kilograms per hour. The results show that signifi-
cant yields of an ammonia synthesis gas and ethylene can be produced in the same
reactor when the system is fed with a mixture of air,, steam, and manure. This
study and earlier bench-scale experiments (EPA-660/2-74-090) show that the product
gases are compatible with both ethylene and ammonia synthesis gas production.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Agricultural Wastes, Organic Wastes,
Livestock, Cattle, Ammonia, Ethylene,
Fluidized Bed Processors, Chemical
Reactors
Anhydrous ammonia,
Feedlot wastes,
Energy recovery,
Wastes conversion
43 F
68 D
97 D
97 F
99 B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
71
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
63
------- |