IHESIS

This is to certify that the

thesis entitled

THE FLASH PYROLYSIS OF CELLULOSE
IN THE PRESENCE OF K2C03

presented by

JONATHAN E. TRAUTZ

has been accepted towards fulfillment
of the requirements for

Master of Science degree in Chemical Engineering

Major rofessor

Dennis J. Miller
Date 1"“,6‘ 85

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

 

 

 

MSU

LIBRARIES
m

\r

 

 

RETURNING MATERIALS:
Place in book drop to
remove this checkout from
your record. FINES will
be charged if book is
returned after the date
stamped below.

 

 

 

 

 

 

 

THE FLASH PYROLYSIS OF CELLULOSE
IN THE PRESENCE OF K2C03

By

JONATHAN E. TRAUTZ

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering

1985

ABSTRACT

THE FLASH PYROLYSIS 0F CELLULOSE
IN THE PRESENCE OF K2C03

By
JONATHAN E. TRAUTZ

Quantitative analysis of product yields from the high temper-
ature flash pyrolysis of cellulose in the presence of K2003have been
conducted to determine the effect of the salt on the product yields.
Experiments were run on a wire screen pyrolysis reactor.

At elevated temperatures and rapid heating rates the pyrolysis
product yields are significantly altered by the presence of K2003.
Similar product yield trends are found for slow and high temperature
flash pyrolysis when K2003 is present.

Neutron activation of the pyrolysis char reveal that the 1.0 weight
percent loaded samples contain all of the K2003 originally impregnated on
to them however, a majority of the K2003 has been lost from the 5.0 and
10.0 weight percent samples. Experimental results suggest that K2003
acts like a catalyst at lower concentrations (1.0 percent) and that the
excess K2C03 in the higher concentrations (5.0 and 10.0 percent) is lost

through a reaction with char.

I dedicate this work to Denise; my wife, and my friend.
For without her continuous support this endeavor could
never have been completed.

ii

ACKNOWLEDGMENTS

I would like to acknowledge the continuous and indispensable
educational support provided by Dr. Dennis J. Miller for the

completion of this degree.

LIST OF TABLES
LIST OF FIGURES

CHAPTER

I.

II.

111.

TABLE OF CONTENTS

Page
.......................... vi
......................... v-i'i

INTRODUCTION ...................... l
A. General Introduction ,,,,,,,,,,,,,,,, l
B. Pyrolysis Parameters ,,,,,,,,,,,,,,,, 3
C. The Influence of Additives on the Slow Pyrolysis

of Wood Materials ,,,,,,,,,,,,,,,, 6
0. Research Objective ,,,,,,,,,,,,,,,,, 7
EXPERIMENTAL APPARATUS ,,,,,,,,,,,,,,,,, 10
A- Pyrolysis Reactor ................. 10
3. Electrical System .................. l3
C. Gas Collection System ,,,,,,,,,,,,,,,, l8
0. Gas Plug Dispersion ,,,,,,,,,,,,,,,,, 28
E. Sample Preparation ,,,,,,,,,,,,,,,,, 32
EXPERIMENTAL TECHNIQUE ................. 35
A. Preparations for Gas Collection and Analysis , , 35
8- Sample Loading ................... 37
C. Purging and Flash Pyrolysis ,,,,,,,,,,,,, 38
0. Gas Collection and Analysis ,,,,,,,,,,,,, 39
E. Collection of Data' ................. 40
F. Neutron Activation of Char ,,,,,,,,,,,,, 42

iv

Page

CHAPTER
IV. SAMPLE CALCULATIONS AND EXPERIMENTAL RESULTS ...... 46
A. Data Manipulation ................. 46
B. Tabulated Data and Calculated Results ....... 58
C. Summary of Experimental Results .......... 73
V. CONCLUSION ....................... 77
A. Comparison of Slow Versus High Temperature Flash
Pyrolysis of Cellulose in the Presence of
K2C03 ....................... 77
B. K2C03 After the Pyrolysis of Cellulose ....... 79
C. K2C03: Catalyst or Reactant? ........... 81
0. Chemistry of Pyrolysis in the Presence of K2C03 . . 83
E. Conclusions .................... 85
F. Suggestions for Future Work ............ 86
REFERENCES ............................ 87
APPENDIX l ........................... 90

TABLE

10.

ll.

l2.

l3.

I4.
15.

LIST OF TABLES

Cellulose Sample Preparation Values ...........

Neutron Activation: Control Samples

Neutron Activation: Char Samples ............

Control Values: Integrated Counts and Areas

Control Values: Weight/Area (xl0'3 grams/cm2) ,,,,,,
Experimental Values: Integrated Counts ,,,,,,,,,

Experimental Values: Integrated Areas and Weight of
Pyrolysis Gases Produced ...............
Experimental Values: Weight of K2C03 and Cellulose
in the l4.5 Milligram Sample, Weight Percent (with
respect to cellulose) of Products ,,,,,,,,,,
Experimental Values: Char Weights and Weight
Percents .......................
Experimental Values: Weight of K2C03 Lost and C0
Produced from the K2C03 reaction ,,,,,,,,,,,
Experimental Values: Corrected Weight and Weight
Percent of C0 .....................

Experimental Values: Integrated Counts of C2H2 and
CzHu .........................

Experimental Values: Integrated Area, Weight and
Weight Percent of C2H2 and Cth ...........

Char Yields and Neutron Activation Study (“2K) .....

Average Gas Weight Percent Values ............

vi

Page
34

44

59
60

70

76

LIST OF FIGURES

FIGURE Page
1. Products of Pyrolysis .................. 2
2. Pyrolysis Reactor ..................... 12
3. Electrical System: Pyrolysis Reactor .......... l6
4. Electrical System: Gas Analysis ............. 20
5. Gas Collection System .................. 22

Flow Schematic of Model lS4L Perkin-Elmer Vapor
Fractometer (referenced from the vapor fracto-

meter manual ) ..................... 24
7. Experimental Apparatus .................. 31
8. Example of Chart Recorder Print-Out ........... 4l

9. Neutron Activation Graph: Rate Versus Weight of K2C03 . . 45

l0. Pyrolysis of Cellulose .................. 84

CHAPTER I

INTRODUCTION

A. General Introduction

"Flash pyrolsis“ are words which for the untrained mind conjure up
images (If unknown scientific frontiers. These words are in fact not as
overwhelming as they appear. "Flash" by definition refers to a process
which is occurring rapidly or for a very brief moment; instantaneous.
In terms of experimental parameters it indicates heating rates in
excess of 250°C per second. A definition of pyrolysis is given by T.
Milnelz "Pyrolysis of carbonaceous materials has been defined as
incomplete thermal degradation, resulting in char, condensible liquids,
or tars and gaseous products, generally in the absence of air." There-
fore, "Flash Pyrolysis" of cellulose is the rapid heating of cellulose
in a non-oxygen atmosphere.

Cellulose is a polymer composed of d-glucose units (six carbon
sugars) with molecular weights greater than 100,000 in wood.2 On a
weight percent basis cellulose is usually the major component of wood;
therefore, any analysis of cellulose can be used as a good represen-
tation of wood in general.

The three major product groups resulting from cellulose pyrolysis

are shown in Figure 1. The pyrolysis gases are a mixture of several

gas compounds: hydrogen, carbon monoxide, carbon dioxide, methane and

mwm»_ocxa we muuseoca .H mczmwd

m¢4flu

Ammoz<mwmoosno m-m Ammzwgom mmoozqolav
-ommwmz¢-w.: 24803859 EHfisHmm -239 T m8q3qmo
mHmwnommm

 

Amzommaoomowm .32 mmzoq .38 .moo .08 8mg

several lower molecular weight hydrocarbons. The second major product
group is made up of tars and condensible liquids. This group is
comprised of mainly levoglucosan (1,6-anhydro-B-D glucopyranose) as
well as smaller amounts of other anhydroglucose compounds and water.3’4’5
Char (charcoal), the third major product, is a material consisting mostly
of carbon. The actual lyyrolysis yields produced for each of the
different groups are dependent upon several parameters. These parameters
and their influence upon product yields are discussed in the next section.
The various by-products of wood pyrolysis have many applica-

tions.6’7’8

Biomass is a renewable resource; consequently, the poten-
tial for wood pyrolysis as a source of energy and chemical feedstocks
is in principle unlimited. Because biomass has lower sulfur and ash
content than coal, it is also a much cleaner form of energy to use.2
The tarry liquids can be used as boiler fuels and the char as a home
heating material. The gaseous products can be used as a low to medium
B.T.U. gas which can be pipelines for home and industrial use. The
product gases of pyrolysis contain ethylene and other low molecular
weight hydrocarbons. They can be used as starting materials in the
petro-chemicalanulpetroleum industries for synthetic rubber, plastics,

alcohols and liquid fuels production. One can, therefore, see the need

for research in an area which has so many applications.

8. Pyrolysis Parameters
Pyrolysis of wood materials has been covered extensively in the

1'20 Several recurring trends have appeared in regards to

literature.
the influence that the external parameters have upon the pyrolysis

product composition. The parameters of interest to this study are

heating rate, residence time, peak temperature, sample particle size
and reactor design. This section will cover each of these parameters
and their influence upon product composition.

Only a few studies were found which addressed the heating rate

9,10

parameter. Scientists at the Massachusetts Institute of Technology

found that at lower temperature ranges (<600-7000C) with decreasing

9 For

heating rate there is a corresponding decrease in char yields.
the same conditions there is an increase in gas and tar yields. Slow
pyrolysis Thermal Gravimetric Analysis (TGA) studies conducted by
another group support these same trends for the char production.10 At
higher temperatures (>600-7000C) the heating rate appears to have very
little influence on the product yields. The predominant factor at high
temperatures is peak temperature alone.

The residence time is the duration in which a wood material is in
contact with a heat source. The influence of this parameter upon
pyrolysis product yields is covered in several papers.1’7’9’11’12
It is found that as the residence time is increased the amount of tar
and total gaseous materials produced increase as well. At the same time
there is a corresponding decrease in the amount of char produced. A
simple explanation for this phenomena is presented by M. R. Hajaligolgz
"This behavior is believed to reflect the fact that at zero holding
time and lower temperatures, decomposition of the cellulose is in-
complete when the final temperature is attained. Holding the sample at
the final temperature allows continued decomposition which generates
additional tar and gas and less char." As was found for the heating

rate parameter, residence time appears 'to have very little influence

upon the product composition above the temperature range 600-7000C.

At these high temperatures the peak temperature becomes the dominating
parameter.

The final temperature that a sample is pyrolyzed at (peak temper-
ature) is very influential upon product yields. Several studies have
shown that as the temperature increases there is a corresponding
increase in the total amount of gas produced and decrease in the amount

of char produced.8’9’11’12’13’14

The influence of peak temperature on
the individual gases within the total gas yield is also covered in these
studies. Only two reports were found to contain information pertaining

to temperature effects on tar yields.9’11

They reported that the tar
yields increased with temperature, reaching a maximum between BSD-700°C
and then decreased with further temperature increases. The difference
in temperature values at which maximum tar yields are reported in these
studies is probably a result of dissimilar heating rates, residence
times and wood materials pyrolyzed.

The effect of sample particle size upon product composition was
covered in one study.11 The amount of tar produced for the three
different sample sizes used did not appear to be size dependent. The
authors felt that residence time of the wood material in the reactor
might be the deciding factor on tar production. In general the char
yields increased as the particle size decreased, and gas yields were
slightly higher for larger particles except at the highest temperature
(700°C).

The final parameter of pyrolysis is reactor design. There have
been many different types of pyrolysis reactors used, varying from
batch bench top to pilot plant continuous feed reactors. These reactor
15 wire-mesh‘,9’16

designs include micro-wave induced pyrolysis, solar

5 4,17

continuous feed with fluidized bed,11’13 fixed bed,
d,7,18 6.12.19 as

furnaces,

introduced sample be continuous feed entrained flow,

well as commercially available Thermal Gravimetric Analyzers (TGA).10’20
Because the design of these reactors is so varied they inherently
have different heating rates (slow versus flash pyrolysis), residence
times, peak temperatures, and wood particle sizes used in them. There-
fore the reactor design determines the types of parameters as well as
their settings. This in turn determines the pyrolysis product yields.
The trends are well established for the parameters discussed in

the preceding paragraphs. Another important parameter which has been
shown to significantly influence the pyrolysis product yields is the

presence of additives. These materials and their influence will be

covered in the next section.

C. The influence of additives on the slow pyrolysis of wood material.

The previous section covered the influence that various parameters
have on pyrolysis product composition. Completion of a literature survey
on wood pyrolysis revealed several studies which used additives when

4,17,18,20’21’22 In these studies the cellulose

pyrolyzing samples.
samples are impregnated by soaking in aqueous solutions containing
known concentrations of the additives. The samples are then removed
and dried. The type of additives used are varied, ranging from acids
to alkali salts. These materials are shown to significantly alter the
pyrolysis product yields.

One of the more frequently used additives is potassium carbonate

(K2C03). The affect that K2C03 has upon slow pyrolysis product yields

is well established.17920,21s22 Research experiments have shown that

the presence of K2C03 increases char yields while significantly decreas-
ing tar formation. Two reports demonstrated over a ten-fold decrease

by weight in tar formation for samples containing one weight percent

21,22 17

K2C03 when compared to unloaded samples. A study by D. L. Pyle
disclosed that char yields increase as the weight percent of K2C03 in the
samples is increased. This study also shows that for twenty weight
percent K2C03 samples there is a 2.4-fold weight increase in char
production. Gas yield were' shown ‘to increase by twenty-five percent
as a result of K2C03 being present.

Another major influence of additives is that they lower the initial

4,17,20,21

decomposition (pyrolysis) temperature of cellulose. Economi-

callyrthisnfight prove to be important in that it lowers the amount of
energy required to pyrolyze a given amount of cellulose.

It is obvious that the presence of additives such as K2C03 in
cellulose influence the product yield from pyrolysis. All the refer-
enced studies pertaining to additives are for slow pyrolysis only.
Pyrolysis of cellulose in the presence of additives has brought many
questions to the surface which should be addressed in more detail.
These questions and the means.to finding their solution are presented

in the next section.

0. Research Objective

The pyrolysis of wood materials in the presence of additives shows
a great deal of promise for selectively altering product yields. The
influence that K2C03 has on slow pyrolysis product yields is well
established; therefore, this would be a desirable additive to use when '

conducting further pyrolysis research. As is suggested in the previous

section, many questions remained unanswered after conducting a litera-
ture survey of pyrolysis. These questions will be addressed in the
following paragraphs.

A large scale continuous feed pyrolysis unit would more than likely
be designed to run at high temperatures and rapid heating rates. ’
Consequently, research conducted in the area of flash pyrolysis of
cellulose at high temperatures with additives present would help to
better define the product yields expected from full scale pyrolysis units.
The influence that heating rate or peak temperatures have upon product
yields is covered in Section B. It was shown that the peak temperature
becomes the predominant parameter at high temperatures; therefore, one
has to question whether K2C03 has any influence upon pyrolysis product
yields at elevated temperatures (>7000C). Restated: are product yields
from the flash pyrolysis of cellulose at high temperatures altered by
the presence of K2C03 or is the high peak temperature the predominant
factor influencing product yields?

Another question resulting from the previous paragraph is brought
to mind. If the yields are altered because K2C03 is present, do they
follow the same trends as were shown for slow pyrolysis?

Nowhere in the literature was it questioned what becomes of the
K2C03 after pyrolysis. Does the K2CO3 remain on the char? Also,
does the K2C03 act as a catalyst or a reactant when pyrolyzing cellulose?

In order to attempt to answer these questions an experimental
scale pyrolysis reactor was constructed which is capable of both rapid
heating rates and elevated peak temperatures. Neutron activation
studies of the char were also conducted to help clarify the question of

K2C03 location. The next three chapters will cover experimental

apparatus, experimental technique, and sample calculations/experimental

results.
Research Objective

Conduct experimental research to determine:

1. If product yields from the flash pyrolysis of cellulose at elevated
temperatures are altered by the presence of K2C03. I

2. 1f the yields are altered, do they follow the same trends as shown
for slow pyrolysis?

3. Does the K2C03 remain on the char?

4. Does K2C03 act as a catalyst or a reactant?

CHAPTER II

EXPERIMENTAL APPARATUS

A. Pyrolysis Reactor

The reactor used in the experiments is a screen type as shown in
Figure 2. By passing a high voltage and amperage between the copper
electrodes (denoted f) in Figure 2, the two 325 mesh stainless steel
screens (9) become red hot because of their resistance to electrical
conduction. In this way the cellulose sample placed between the
screens is pyrolyzed. -

The reactor consists of a 2" 0.0. stainless steel tube 6" in
length (d). One end is closed off except for a Swagelock fitting
connected to a %" copper tube (i). This is the gas outlet which is
in line with the gas collection system. The opposite end of the stain-
less steel tube is swaged to a 2" Swagelock fitting (j). The inner
lining of the reactor (h) is a hollow tube of 99.99 percent alumina
5" in length by 1 and 3/4" 0.0. This liner acts as a support for the
electrodes. Because the resistance to electrical conduction is large
in this ceramic material, all electrical current passes through the
electrodes and screens only. The calculation of resistance in alumina
is given in Appendix 1, Part A.

The power supply leads (b) are insulated 8 gauge capper wire which
are introduced into the reactor through air tight Conax fittings. These
power leads connect to the electrodes. Two thermocouples are run into

10

11

Figure 2. Pyrolysis Reactor

c1.
c2.

Helium Inlet

Electrical Wire Inlets
Thermocouple Inlets
Thermocouples in Reactor
Reactor Wall

Reactor Support

Copper Electrodes

325 Mesh Wire Screens
Alumina Inner Reactor Lining

Product Gas Outlet
2" Swagelock Fitting

Brass Cross Fitting

12

Louommm mmm>_ocza

.0

   

3 3.3..

.N mcamwn

13

the reactor through another Conax fitting (c1). One thermocouple
is placed between the screens and is connected to the temperature
controller. This thermocouple is an exposed junction Chromel-Alumel
(K type) thermocouple designed for fast response times. It provides
a signal to the temperature controller which controls the maximum
temperature that the screens reach. This will be discussed in more
detail in the electrical system section. The second thermocouple is an
ungrounded junction Chromel-Alumel (K type) which is placed above the
screens close to the inside wall of the reactor. This thermocouple
is used to indicate the wall temperature. During pyrolysis the levo-
glucosan is dispersed onto the reactor wall which reaches a maximum
temperature of 120°C. Because levoglucosan does not begin to decompose
until 280°C it is safe to assume further decomposition is minimal.
Helium is introduced through a Swagelock fitting connected to
a 1/8" copper tube (e). The Swagelock fitting (a) and the Conax
fittings (b) and C1) are connected to a brass cross fitting (k) which
is threaded into the 2" Swagelock fitting (j).
The approximate volume of the reactor which includes inlet and
outlet tubing is 197 cubic centimeters. Through the use of Swagelock
and Conax fittings an air tight reactor was constructed. This is an

essential criterion for conducting pyrolysis experiments.

8. Electrical System

The electrical systems used in the pyrolysis experiments are
outlined in Figures 3 and 4. Figure 3 is the electrical system for the
pyrolysis reactor. Figure 4 is the electrical system for gas analysis.

The pyrolysis reactor system will be discussed first.

14

Because of the rapid heating rate required for flash pyrolysis, in
excess of 250°C per second, it was predicted that a large power require-
ment would be needed. Before an available power supply was used, calcula-
tions were made to estimate the electrical power required. The extreme
case was analyzed in order to determine the maximum power requirement.
This was the case where the screens were almost instantly heated to
1,0000C and were then maintained at this temperature.

An energy balance around the screens gives

Power required to = Total energy lost to
heat screens surroundings (2-l)

As outlined in Part B of Appendix 1, there are three methods of energy

loss from the screens; conduction, radiation and free convection. Thus,

Total energy Energy lost Energy lost Energy lost
lost to = due to + due to + due to free
surroundings conduction radiation convection

" = (26.49 + 222.39 + 99.05) Joules/sec

“ = 347.93 Joules/sec (2—2)
The arithmetic average of resistance FNUli" the screens is 0.3292 ohms.
Therefore

)0.5

I (current) = (power/resistance = 32.51 Amps (2-3)

and Volts = I x RM: 10.70 (M)

The actual power needed when conducting the experiments was approximately
40 amps and 9 volts. The difference between the theoretical and actual
power requirements was possibly due to the inability to calculate an
accurate value for the resistance of the screens. The Electron Arc
Division power supply used is depicted as (a) in Figure 3.

Figure 3 shows an enlargement (i) of the copper electrodes (9)

and the wire mesh screens (h) located within the reactor. As described

15

Figure 3. Electrical System: Pyrolysis Reactor

a. Electron Arc Division Power Supply
b. Magnetic Contactor

c. Omega 4001 Single Set Point Proportional and
On-Off Controller

d. Omega Model 650 Thermocouple Thermometer

e. Exposed Junction Chromel-Alumel (K Type)
Thermocouple

f. Ungrounded Junction Chromel-Alumel (K Type)
Thermocouple

9. Copper Electrode
h. 325 Mesh Wire Screens

i. Enlargement of Electrical System Inside Reactor

 

 

16

 

H. . -. . . . . - . . . . -.. . . .. .. . .u Louummm meg—Pia “swam? 28.758; .m 853...
u
I“. '\ / \
__
.\.-, -
41 ill. 41
00000000000 r. O. U CI 0
.\

 

 

 

 

 

5i
.[l
0

®

6

G

 

 

 

17

in the previous section, the exposed junction Chromel-Alumel (k type)
thermocouple (e) is placed between the wire screens. This is connected
to an Omega 4001 Single Set Point Proportional and On-Off Controller
(c). This controller is interphased with a magnetic contactor (b)
which when activated opens the curcuit coming from the power supply.

The undergrounded junction Chromel-Alumel (K type) thermocouple (f)
is connected to an Omega Model 650 Thermocouple Thermometer (d) which
provides a digital read out of the temperature.

The operation of the electrical system in the pyrolysis experiment
is as follows:

1. The Omega 4001 controller is preset at 900°C so that above
this temperature the controller signals the magnetic contactor to open.

2. The power supply is then turned on resulting in current flowing
through the screens. The high resistance screens heat up and become
glowing hot. The heating rate obtained in running experiments is in
excess of 300°C per second.

3. Once the set point temperature is reached (as detected by the
thermocouple placed between the screens) the controller signals the
magnetic contactor to break the power supply curcuit. This results
in the screens cooling down until the temperature is again below the
controller set point.

4. Once this occurs the magnetic contactor controlled by the
Omega controller opens the circuit again to the screens. The heating
process then starts over again.

Because of the delay in response from the thermocouple, controller
and contactor, the temperature of the screens was found to oscillate

around the set point of 900°C. This resulted in temperature variations

18

of 50-60°c above the set point and 80-1200C below.

Figure 4 shows a schematic of the electrical system used for gas
analysis. The Model 154L Perkin-Elmer Vapor Fractometer gas chromato-
graph (b) has two separate electrical systems. The first system is
powered by a 110 volt AC output (c). This provides power to a blower
fan and a heating element.

The second system requires a 9 volt DC supply which is provided
by the Power/Mate Corporation power supply (a). This provides the
power required to run the bridge curcuit of the thermal conductivity
detector used to analyze the gas products. For a more in-depth under-
standing of this circuitry see the manual for the Model 154L Vapor
Fractometer or the gas collection system section.

The signal generated by the detector located in the gas chromato-
graph drives the Model XKR Sargent-Welch Recorder (d) in Figure 4. From
the recorder a qualitative analysis of the gas components is obtained.
The recorder is interphased with a Sargent-Welch Electronic Integrator
(e) which provides a quantitative analysis of the gases. By using this
electrical system a qualitative and quantitative analysis of the gaseous

products are obtained.

C. Gas Collection System

A schematic of the gas system used in the pyrolysis experiments
is shown in Figure 5. A description of the various equipment or
materials used as well as their corresponding function are discussed in
this section. Small letters surrounded by parentheses refer to

equipment shown in Figure 5.

19

Figure 4. Electrical System: Gas Analysis

w

Power/Mate Corporation Power Supply

b. Model 154L Perkin-Elmer Vapor Fractometer
c. 110 Volt AC Power Source

d. Model XKR Sargent-Welch Recorder

e. Sargent-Welch Electronic Integrator

20

mwm»_m:< mow

”Empmxm _mowcuum_m

.v mgzmwu

 

p.
q.

21

Figure 5. Gas Collection System

Helium Gas Tank

Liquid Nitrogen Trap

Whitey Sample Cylinder

T-Connector

3-Way Whitey Ball Valve

Atmospheric Vent in G.C. Line

Model 154L Perkin-Elmer Vapor Fractometer
Bubble Flow Meter

2-Way Whitey On-Off Ball Valve
T-Connector and Pressure Gauge
Pyrolysis Reactor

3-Way Whitey Ball Valve

Atmospheric Vent for Pyrolysis Reactor
Dry Ice/Acetone Trap

6-Port Valve

Atmospheric Outlet for Reactor Line

Sample Collection Loop

r1 and r2. 3-Way Whitey Ball Valves

S.

t.

Liquid Nitrogen Trap
Connecting Tube

6—Way Whitey Sample Valve
Sample Loop

Calibration Gas Tank

Calibration Gas Atmospheric Outlet

22

Emumxm compow—_ou mow .m beamed

/.l--.--llll-llllll llllillil.--l-

 

 

N. C

 

 

--—--.—----- ...--.,..-....

T
I
l
l
I

L--—-.---- .-.—-.-...-o-J

u

 

 

 

 

 

 

 

 

23

(a) Helium Gas Tank - A 99.999 percent pure source of helium was

 

used in the pyrolysis experiments. This helium was utilized to flush
the reactor, collect the gaseous products preceding pyrolysis of the
cellulose sample and as a carrier gas in the gas chromatograph.

(b) Liquid Nitrogen Trap - This trap consists of a 14" by a"

 

O.D. U-shaped stainless steel tube. Inside the tube is 0.5 grams of
silica gel desiccant (6-16 mesh) usecj to absorb any impurities which
might be present in the helium. The U-shaped tube is placed in a Dewar
flask of liquid nitrogen (-196°C).

(C) Whitey Sample Cylinder -This 500 cc sample cylinder is filled
with Linde 3A molecular sieves (1/8" pellets). These molecular sieves
further absorb any impurities missed by the upstream nitrogen cold
trap. By passing the helium through the liquid nitrogen trap and then
the sample cylinder, the gas will have had all moisture and other impu-
rities removed‘uaa level undetectable in the gas chromatograph.

(d) T-Connector - The purified helium is split into two lines at

 

this point. The first line runs to the gas chromatograph (G.C.);
the second to the reactor.

(e) 3-Wanghitey Ball Valve - This 3-way valve in the G.C. line

 

allows the helium gas to be directed to the G.C. or to be vented to
the atmosphere at (f).

(f) Atmospheric Vent in G.C. Line - This vent is used to vent

 

impurities to the atmosphere when the sample cylinder molecular sieves
are being regenerated by heating at 200°F.

(9) Model 154L Perkin-Elmer Vapor Fractometer - This gas chroma-

 

tograph is used to analyze the product gases from the pyrolysis exper-

iments. Figure 6 shows the flow schematic of the G.C. The difference

24

Lopmeouomsd Loam> cmepmicmxema Jemfi .mcoz mo ovumsmzom zo_n

A_m::ms gmumeopomgm goqm> an» Eocm umucmemwmgv

 

htOmmwuo<
02.4Q24m
mdoli

zo_m_uwma

IA

\

02.4midm

o 50.1.
«0... x0040
20:09.2.

 

f...
,~nvL

 

«Oputwo I

 

 

aw¢¢<<IU

ZiDdOU III

I
eL

 

Zw>0

 

 

«Obsqfivwa weammwum

.m mczmwu

 

//

l l 5&3
. .A-v. m<o
Anv L1 «euu<u

 

I
w8<0 wunmmwcm

.__._xw

 

 

 

25

between the thermal conductivities 0f the carrier gas (helium) and
components of interest are detected in the detector cell using thermis-
tors. This difference is converted to an electrical signal which
is relayed to the chart recorder described in the gas analysis elec-
trical system section. The injection block shown in Figure 6 could not
be made air tight subsequently, it was removed.

The column used in the G.C. was a 60/80 mesh Carbosieve S—II.
The dimensions of this stainless steel column are 5 feet in length by
1/8 inch 0.0. It was used successfully to separate carbon monoxide,
carbon dioxide, methane and the C2 hydrocarbons.

(h) Bubble Flow Meter - This meter is used to determine the flow

 

rate through the G.C. The flow meter is connected to the outlet of
the G.C. downstream from the column side of the detector.

(i) 2-Way Whitey On-Off Ball Valve - This valve regulates the

 

helium flow rate through the pyrolysis reactor. It may also be put in
an off position while batch pyrolysis of the sample is taking place or
to check for air leaks in the reactor.

(j) T-Connector and Pressure Gauge - This gauge is used to detect

 

pressure changes from leakage in the reactor before running each sample.

(k) Pyrolysis Reactor - Discussed in detail in the two preceding

 

sections.

(1) 3-Way Whitey Ball Valve - This valve has three different

 

positions. The first is the off position and is used for the same
reasons as outlined for the 2-way valve (i). The second position allows
gases passing through the reactor to be vented to the atmosphere at (m).
After the cellulose sample is in place and the reactor sealed, this

position is used to initially purge the reactor with helium, thus

26

removing a majority of the air present in the reactor. The final
setting allows product gases from the reactor to be diverted to
collection traps whose description will be forthcoming.

(m) Atmospheric Vent for the Pyrolysis Reactor - Discussed in

 

the preceding paragraph.

(n) Dry Ice/Acetone Trap - This trap consists of a 22" length

 

of k" 0.0. U-shaped stainless steel tube with fiberglass packed inside.
The tube is placed in a Dewar flash containing dry ice and acetone
(-77°C). The purpose of this trap is to collect any water or particulate
matter which is a product of the pyrolysis reaction.

(0) 6-Port Valve - This valve has two positions. The first

 

position (down) allows the reactor gases coming from the dry ice/acetone
trap to go through the sample-collection loop (q) and eventually be
vented to the atmosphere at (p). In this position gaseous products

from the reactor are collected in the sample-collection loop. Simulta-
neously, helium from the reference side of the G.C. passes through the
valve and goes directly to the column where it is eventually vented to
the atmosphere (h). See Figure 6 for more detail. There is at no time
any direct connection between the G.C. and reactor gas lines.

In the second position (up) the reference side helium from the G.C.
goes through the sample-collection loop (q), then returns to the column
as previously discussed. In this way the gaseous products from the
pyrolysis reactor which were collected in the sample-collection loop
are flushed into the G.C. for qualitative and quantitative analysis.

The reactor line in this position is diverted directly to the atmosphere

at (p)-

27

(p) Atmospheric Outlet for Reactor Line - Discussed in the
preceding paragraph.
(q) Sample-Collection Loop - This loop has two important sections.
l. Liquid Nitrogen Trap (5)
2. 6-Way Whitey Sample Injection Valve (u)
(r1 and r2) 3-Woy Whitey Ball Valve - These two 3-way valves are

 

used in combination for two purposes. In the first position gases
flowing through the sample-collection loop pass through the liquid
nitrogen trap (s). The second position allows the same gases to by-pass
the liquid nitrogen trap through a connecting tube (t) located between
the valves.

(5) Liquid Nitrogen Trap - This trap consists of a 22" length by

 

%" 0.0. U-shaped stainless steel tube placed in a Dewar flask filled
with liquid nitrogen. As the product gases from the reactor pass
through this cooled tubing they condense and are retained. The trap
contains 0.l grams of silica gel, added to the inside of the tube to
insure complete collection of carbon monoxide which has a boiling point
close to liquid nitrogen. Once all the gases are collected, the trap
is placed in boiling water so the liquids become gases again. These
gases are then flushed into the G.C. for analysis.

(t) Connecting Tube - Tube used to by-pass the liquid nitrogen

 

trap (s).

(u) 6-way Whitey Sample Valve - This six port valve is used to
inject known volumes of calibration gas used in calibrating the
experimental results. Like the six port valve (0), this valve has two
positions. The first position allows the gases coming from the

calibration gas tank (w) to flow through the 2 cc sample loop (v)

28

and exit to the atmosphere at the calibration gas atmospheric outlet
(x). While this is occurring the gases flowing through the sample-
collection loop (q) are entering the valve and exiting without coming
in contact with the calibration gases. When the valve is switched

to the second position the sample-collection loop flows into the sample
loop (v). In this way a 2 cc sample of the calibration gas is injected
into the sample-collection loop (q) which will eventually be analyzed
in the gas chromatograph (9). Also in this position the gas coming
from the calibration gas tank (w) is vented directly to the atmosphere
at (x).

(v) Sample Loop - This 36.8" length by l/B“ 0.0. copper tube has

 

a volume of 2 cc. Its use is described in the preceding paragraph.

(w) Calibration Gas Tank - This tank contains a gas of known

 

composition. This compositionis.5.33 percent carbon monoxide, 5.26
percent carbon dioxide, 5.20 percent methane, and the balance Helium.
All percents are by volume.

(x) Calibration Gas Atmo§pheric Outlet - Discussed in the preceding

 

paragraphs.

0. Gas Plug Dispersion
Because the distance from the liquid nitrogen trap (s) to the G.C.
(g) in Figure 5 is so long (107" of l/8" copper tubing) it was suspected
that a plug of gas might disperse before reaching the G.C. column. If
this occurred then good separation in the column would not be achieved.
By modeling the concentration of a plug of gas as it flows through
the tubing as a function of both time and distance, the following

equation is obtained.

29

3C _ D 2
~31- - AB 3 C2 (2-5)
8X

As outlined in Part C of Appendix 1, given the appropriate boundary

conditions the solution to equation (2-5) is

c = C0 (1-erf (x/JhtoAB)) (2-6)
where

C0 = original concentration of the plug of gas

x = distance from the original plug boundary

DAB== diffusion coefficient = 1.397 (cm)2/second

If the most extreme dilution of the plug of gas which can be tolerated
is defined as C = 0.OlCo at x = 11" (one-half the original plug length
when it leaves the liquid nitrogen trap). then equation (2-6) can

be rearranged to solved for the time t.

 

(2-7)

122.9 seconds

('l’
II

The actual time required for the plug to reach the column at a
flow rate of 30 (cm)3/minute was calculated to be 11.64 seconds.
Because the difference between these two values was so large it was
safe to assume that the plug of gas would reach the G.C. column intact
with only minimal dilution occurring.

Figure 7 shows a schematic of the three main components of the
experimental apparatus combined; PYrolysis reactor, electrical system,

gas collection system.

30

Figure 7. Experimental Apparatus

a.

b.

Electron Arc Division Power Supply
Magnetic Contactor

Omega 4001 Single Set Point Proportional
and On-Off Controller

Omega Model 650 Thermocouple Thermometer
Helium Gas Tank

Pyrolysis Reactor

Gas Collection System (simplified)

Model 154L Perkin-Elmer Vapor Fractometer
Model XKR Sargent—Welch Recorder

Sargent-Welch Electronic Integrator

31

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mzumsmaa< Faucmswgmaxm

 

.n mesowu

 

 

 

 

 

 

 

 

 

 

 

 

IE]:
El

 

 

 

 

 

 

GI
ell

 

CM
CH
\

3|

 

 

 

 

 

32

E. Sample Preparation

The cellulose used in the experiments was #4 Whatman filter paper.
It was chosen because of its purity and thinness. It has been shown”
that ash can act as a catalyst26; therefore this low ash content filter
paper, less than 0.06 percent by weight, is an excellent source of pure
cellulose. The advantage of using these thin samples (0.008 inch) is
that the formation of a thick layer of char during the pyrolysis
reaction is avoided. Therefore it is less likely that the product
materials formed inside the sample will react appreciably with the hot
char surfaces as they travel to the surface of the sample.

In running the experiments it was desired to make 0.0, 1.0, 5.0
and 10.0 weight percent K2C03 in cellulose samples. Weight percent

is defined as

Weight _ (gms 0f K2C03 in the cellulose) x 100 (2-8)
Percent (gms of cellulose + gms of K2C03 in cellulose)

 

The samples were impregnated with K2C03 by placing the cellulose paper
in distilled water which contained a known concentration of alkali salt.
After the water evaporates the K2C03 remains on the sample. The method
for preparing the samples is as follows.

Samples of Whatman filter paper (17mm by 35mm) were prepared with
an average weight of 0.05376 grams. Knowing the desired weight percent
and the average weight of each cellulose sample, the average weight of
K2C03 required per sample could be calculated using equation (2-8).
These values are presented in Table 1.

Ten of the precut/preweighed cellulose samples were placed in
distilled water then weighed immediately to determine their water

up-take. The average water up-take was determined to be 0.12575 grams/

33

cellulose sample. Knowing the density of water at 25°C to be 0.996077

grams/ml the average water up-take per sample was converted to 0.126245
ml/cellulose sample. Since it was desired to use 80 ml of solution to

place the samples in for impregnation, the following equation was used

to calculate the weight of K2C03 required in 80 ml of water.

(grams of K2C03/sample) x 80 ml water

Grams Of K2CO3 = TTmT'odeistilled water/sampleTE (2-9)

 

The calculated values are presented in Table 1.

Once the sample were prepared, they were dryed in a furnace for
24 hours at 80°C. After this they were placed in a glass desicator
for an additional 30 days before experiments were conducted. As a
control the 0.0 weight percent samples were run through the procedure
of impregnation and drying, except of course no K2C03 was present in
the distilled water. By knowing the original weight and the weight
after impregnating each sample, the difference between these values
would give the grams of K2C03 absorbed in that particular cellulose
sample. Equation (2-8) could then be used to calculate the actual weight
percent for each sample. Ten samples were prepared for each desired

weight percent. Their average values are given in Table 1.

34

Table 1. Cellulose Sample Preparation Values

 

 

 

 

Desired Average grams of £2903 Gms of 52C93 Actual Average
Weight % Cellulose Sample 80 ml water Weight Percent
0.0 0.000 0.000 0.00
1.0 5.4307 x 10"l+ 0.344 1.38
5.0 2.8295 x 10'3 1.793 6.49

10.0 5.9733 x 10-3 3.785 13.05

 

CHAPTER III

EXPERIMENTAL TECHNIQUE

A. Preparation for Gas Collection and Analysis.

This section will cover three topics of experimental preparation.
They are preparing the G.C. and other recording instruments for use,
purging the collection traps and gas lines, and running a daily control
sample gas.

Before the pyrolysis experiments can be conducted the G.C. and
recording instruments must be adjusted. The flow rate in the G.C.
column is established at 30 cc per minute by regulating the column
pressure control to a setting of 14 psig. This flow rate can be
confirmed by using the bubble flow meter (h) of Figure 5. Both the
Power/Mate Power Supply (a) in Figure 4 and the detector voltage switch
on the G.C. are turned on and adjusted to 9 volts. The G.C. temperature
control is turned to the blower setting which activates the oven fan.
Other settings which should be preset on the G.C. are power control
at 100 percent power and recorder range control at an attenuation of 32.
The temperature control is now adjusted to 175°C. The heating of the
column removes any impurities which may have collected while the
column was not being used. Subsequently the temperature control is
readjusted t0 the blower setting and the column cooled to room

temperature. The G.C. is now ready for gas analysis.

35

36

During the initial heating of the column to 175°C the cold traps
and gas lines are also purged of impurities. While the 6-port valve
(0) of Figure 5 is in the up position, the liquid nitrogen traps (b)
and (s) are placed in boiling water. This removes any impurities
which may have collected on the silica gel inside the traps. After
five minutes the boiling water is removed and the first trap (b) is
placed in liquid nitrogen. The second trap (s) is removed from the
flowing helium of the sample collection loop (0) by adjusting the
3-way valves (r1 and r2). The gas collection system is now ready to be
used.

The power setting on the Model XKR Sargent-Welch Recorder is
adjusted to the "on" position. Other controls which should be preset
are chart speed at two centimeters per minute, variable span in the "off"
position, and span set at 50 millivolts.

The Sargent-Welch Electronic Integrator (e) in Figure 4 is placed
on the stand-by power setting. The span should be preset at 50
millivolts.

A control sample gas is used each day an experiment is run to
allow evaluation of the pyrolysis gaseous products both qualitatively
and quantitatively. The volume percent of CO, CH“, and C02 in the
calibration gas tank is 5.33, 5.20, and 5.26, respectively. With the
6-port valve (0) of Figure 5 in the up position and the 6-way valve (u)
in the first position, the trap (s) is opened and placed in liquid
nitrogen. After purging the sample loop (v) for 1-2 minutes with gas
from the calibration gas tank (w), the 6—way valve (u) is switched to
the second position. This injects a two cc sample of the calibration

gas into the sample-collection loop (0). Because the sample loop

37

is upstream of the liquid nitrogen trap (s) the sample gas is collected
in this trap. The recorder and integrator are now set on the record
position. After five minutes the liquid nitrogen is removed and the
trap is placed in boiling water. The sample gas is now flushed into the
G.C. for analysis. For details on gas analysis see Chapter III,

Section C.

8. Sample Loading.

This section covers the details of preparation for sample loading.
The chronological order in which they are presented coincide with the
actual experimental approach used.

The electrodes (f) and thermocouples (c2) in Figure 2 are first
cleaned with acetone. This is done in order to remove any levoglucosan
which may still be present from the previous experiment. All handling
of electrodes, thermocouples, screens and samples are done with plastic
gloves in order to prevent contamination with body oils. The screens
are put in place and heated in air for 10 seconds at 900°C. This burns
off any residue from screens that have been previously used. It also
prevents further oxidation of the metal screens during experimental
runs.

A 14.5 milligram cellulose sample is cut and weighted from the
previously prepared samples outlined in Chapter II, Section E. The
approximate size of these 14.5 milligram samples is one centimeter
by one centimeter. The prepared sample and the exposed junction
Chromel-Alumel (K type) thermocouple are then placed between the screens

using acetone cleaned forcepts.

38

One of the alumina tube liners (h) in Figure 2 is slipped into
place around the electrodes and screens. A clean alumina tube,
prepared by heating in a furnace at 700°C for one hour, is used for
each experiment. The outer reactor wall (d) is put in place and swaged
with the 2" Swagelock fitting (j) to form an air tight seal. Finally
the product gas outlet tube (i) is swaged to its appropriate fitting.

The pyrolysis reactor is now completely sealed.

C. Purging and Flash Pyrolysis.

This section covers the details of preparation for purging the
reactor and flash pyrolysis of the cellulose sample. Once again the
chronological order given coincides with the actual experimental approach
used.

By placing the 3-way valve (1) of Figure 5 in the vent position (m)
and 2-way valve (i) open at a flow rate of 300 cc per minute, the reactor
is flushed with helium for 3 minutes to remove air. After this time
the 3-way valve (1) is put in the off position and the reactor is
pressurized to 40 psig. The 2-way valve (i) is then closed and the
reactor is checked for pressure leaks using the pressure gauge (j).
If a leak is detected then a soap solution is used to locate it and
appropriate measures are taken to eliminate the problem. If there is
no change in pressure for a period of 30 minutes then the reactor is
considered air tight and the pressure is released through the atmospher-
ic vent (m).

With the 6-p0rt valve (0) in the second position, the 3-way valve

(1) is switched to the third position. This allows gases from the

reactor passing through the dry ice/acetone trap (n) to be vented to the

39

atmosphere at (p). The 2-way valve (i) is adjusted to allow a flow
rate of 30 cc per minute for 25 minutes. This further helps purge both
the reactor and the gas line of any air. Two minutes before purging is
complete, trap (s) is opened and placed in liquid nitrogen. Trap (n)
at this time is placed in dry ice/acetone. The 6-port valve (0) is
moved to the first position (down) so that gases coming from the reactor
will now pass through the sample-collection loop (q).

After the 25 minute purge is complete the reactor is sealed at both
ends by putting valves (i) and (l) in the closed position. The sample
is then pyrolyzed for l5 seconds at atmospheric pressure as outlined

in Chapter II, Section B.

0. Gas Collection and Analysis.

After pyrolysis the reactor is allowed to cool for five minutes.
The 3-way valve (l) is opened to the third position again and the flow
rate through the reactor is established at 30 cc per minute using the
2-way valve (i). In this way the product gases from pyrolysis are
collected in the liquid nitrogen trap (s) for 25 minutes.

When this time period has expired the liquid nitrogen trap (s)
is closed off and the Dewar flask containing liquid nitrogen removed.
The 6-port valve (0) is placed in the up position so that helium from
the reference side of the G.C. is now flowing through the sample-
collection loop (q) and into the G.C. column. The liquid nitrogen
trap is opened again and placed in boiling water. The product gases
collected in the trap are now flushed into the G.C. for analysis. The
dry ice/acetone trap is also placed in boiling water to remove any

moisture which may have collected.

40

The preceding techniques are used to prepare the reactor for
pyrolysis, to pyrolyze the sample, to collect the product gases, and to

flush the gases into the G.C. for qualitative and quantitative analysis.

E. Collection of Data.
l. Gas
In the preceding sections the methods used to collect the
pyrolysis product gases were covered. This section will cover how these
gases are analyzed both qualitatively and quantitatively.

The control sample gas or pyrolysis product gases are released from
the liquid nitrogen trap (s) and flushed into the G.C.‘s column. After
two minutes from the time of release, the temperature control on the
G.C. is adjusted to 175°C. The ramping of the temperature to 175°C
allows better separation of the gaseous components in the Carbosieve
S-II column. As is described in the manual for the 154L Vapor Fracto-
meter, "Qualitative analysis is based on the times at which components
emerge from the column. For a particular column and a given set of
operating conditions, the retention time for each component is character-
istic of the substance and serves to identify it." As each component
elutes from the column, its thermal conductivity is measured against
the reference gas (helium). This difference in thermal conductivities
is converted into an electrical signal which drives the Model XKR
Sargent-Welch Recorder and produces a graphic display as shown in
Figure 8. This readout shows a peak for each component as a function
of time (qualitative) and concentration (quantitative).

The amount of each component is directly proportional to the area

under its curve. This curve is electronically integrated using the

pzoipcwca Lmvcoomm peace we m_asmxm .m weaned

 

 

 

 

 

 

 

41

 

 

 

 

 

_ n
L :5
l l- E;
m
6.
“u l¢*111
m 2.
.w.
1. a.
m 3.
9
4n. Co..+.U.S<9++< w» C
034...}: to} $333 6. cm.
$.38 .‘Ttoer..$Lvm w 600

u 9460

 

 

 

42

Sargent-Welch Electronic Integrator. The integrator records the
results in the form of lines on the margin of the recorder chart
as shown in Figure 8.

Each day an experiment is conducted a control sample gas is run
as outlined in Chapter III, Section A. Since this sample gas is of
a known volume, it can be used to calibrate the pyrolysis product gases
of that day both qualitatively and quantitatively. For details on data
manipulation, see Chapter IV, Section A. The actual integrated values
obtained for the daily controls and experimental runs are presented in
Tables 4 and 6, respectively.

2. Char

After the gaseous products are analyzed, both ends of the

reactor are closed off using valves (i) and (l) in Figure 5. The
product gas outlet tube (1) of Figure 2 is removed, followed by the
outer reactor wall (d) and alumina liner (h). Using acetone-cleaned
forceps the char sample found between the screens is removed and weighed.
The weighed values for the various experimental runs are given in
Table 9.

3. Tar

Although actual quantitative weight measurements for the tar

were not conducted, visual observation of the amount of tar condensed

on the ceramic tube allowed the trend in tar yield to be determined.

F. Neutron Activation of Char
As the weight percent of K2C03 was increased in the prepared
samples, it was observed that the weight of the remaining char

increased as well. Neutron activation studies were conducted to

43

determine whether the increase in weight was a result of an increased
amount of K2C03 remaining in the char or whether K2C03 actually produced
more char.

Neutron activation was used to determine the amount of potassium
(K) remaining in the char samples. An assumption made in this study
is that the potassium remaining in the char is still present in the
form of K2C03.

One char sample from each of the weight percent categories along
with control samples were irradiated in the nuclear reactor at Michigan
State University. The rate (counts per second) were determined for

42K. The control

the gama-ray emission from the radioactive decay of
samples consisted of weighed amounts of pure K2C03 in quantities com-
parable to those found in the 1.0, 5.0 and 10.0 weight percent categor-
ies. Table 2 shows the control sample weights and their corresponding
rates. Table 3 gives the rates for the various char samples used.
Figure 9 is a graph constructed by plotting the rate versus weight
of K2C03 using control sample values. Knowing the rate of the various

char samples, this graph can be used to find the grams of K2C03 for

that corresponding sample. These values are presented in Table 3.

44

Table 2. Neutron Activation: Control Samples

 

 

Weight of K2003 Approximate Weight Rate
(grams) Percent (Counts per second)
1.45 x 10‘“ 1.0 0.7
7.25 x 10"+ 5.0 3.2
1.45 x 10'3 10.0 5.3

 

Table 3. Neutron Activation: Char Samples

 

Experimental Run # Weight Percent Rate Weight of K2C03 (grams)

 

9 1.04 0.9 2.00 x 10'”+
15 6.17 0.9 2.00 x 10'“
22 12.80 0.5 1.05 x 10-“

 

oo o.m oi ohm o.~ o.”

.moumx ..8 £3.82 mamsm> 3mm ”conga 8.3.2352 .5552 .m 8:6:

35:31:. 18: z :5:

R ’

 

45

> c

 

 

It

 

 

11

'0;

RNA meouunv

a.¢..u
o..-

vo.m

as

vo.m

 

CHAPTER IV

SAMPLE CALCULATIONS AND EXPERIMENTAL RESULTS

A. Data Manipulation

The methods used to collect the experimental data were covered in
Chapter III, Section C. This section will show how the rough data is
converted to meaningful quantitative values. Experimental Run #14 will
be used as an example for all data manipulations. All tabular data
referred to in this section are presented in the second section of this
chapter.

1. Weight of CO, CH”, and CO2 in the control sample gas.

 

The volume percentage of CO, CH“, and 002 in the calibration gas
tank is 5.33, 5.20 and 5.26, respectively. There is a dead space of
0.045 cc's in the 6-way Whitey sample valve; therefore a 2 cc injected
sample will contain 2.045 cc. Each of the components are treated as an
ideal gas at atmospheric pressure. By correcting to room temperature

the weight of each component is calculated in the following manner:

3
Weight = 0.0533 x 2.045 cm x l mole x 28 grams x 273°K

(4-1)
of CO 22,400 on3 l mole 298°K

1.2482 x 10'1+ grams

46

47

 

Weight _ 0.0520 x 2.045 cm3 x 1 mole x 16 grams x 273°K z (4_2)
Of CHA 22,400 cm3 1 mole 298°K

6.9585 x 10'5 grams

Weight
Of C02

 

0.0526 x 2.045 cm3 x l mole x 44 grams x 273°K = (4_3)
22,400 cm3 l mole 2980K

1.9357 x 10'“ grams.

2. Integrated area (A) for the control sample components.

 

The integrated area is given by the following relationship:

number of counts x recorder chart speed.
A = 750 counts / centimeter - minute (4-4)

 

A recorder chart Speed of two centimeters per minute were used for all
control samples and experimental runs. Run #14 was conducted on
10-25-84, therefore the sample control counts from this same date will
be used for calibration purposes. These values are given in Table 4.
The integrated areas are calculated using equation (4-4) as follows:

= 26.25 counts x 2 centimeters/minute = -2 2 _
A<°°) 750 counts / centimeters - minute 7'000 X 1° cm (4 5)

 

A(CH ) = 23.8 counts x 2 centimeters/minute 6.347 x 10"2 cm2 (4-6)
1,

750 counts / centimeters - minute

 

= 28.83 counts x 2 centimeters/minute = -2 2 _
A(°°2) 750 counts / centimeters - minute 7°°88 x 1° cm (4 7)

The calculated values for the sample control integrated areas are

given in Table 4.

48

3. Weight per integrated area for each sample component.
Both the weight and integrated area are known for each component
of the sample gas for the day Run #14 was conducted. The weight per

integrated area can now easily be calculated.

Weight of C0 = 1.2482 x 10‘“ grams
”A(C°) 7.000 x 10'2 cm2

 

 

= 1.7831 x 10"3 grams (4-8)
cm

-5
Weight of CHn = 6.9585 x 10. grams
A(CH§)' 6.347 x 10'2 cm

 

 

1.0963 x 10"3 grams (4-9)
cm

 

-u
Weight of C09 1.9357 x 10 ¥_grams
A(C02) 7.688 x 10’2 cm2

 

= 2.5178 x 10'3 grams (4-10)
cm '

The calculated values for the various sample runs are presented in Table
5. These values are important in that they given a relationship between
the weight of each component and a particular chart area. These values
can then be used to calibrate the experimental data which is collected
the same day.

4. Calculate the integrated area (Aex ) for pyrolysis Experimental
Run #14. p

 

The integrated areas of the experimental runs can be calculated
using equation (4-4). The integrated counts presented in Table 6 for
Experimental Run #14 are 625.85 for CD, 88.25 for CH“, and 191.15 for

C02. Therefore

 

_ 625.85 counts x 2 cm./min. = 2 _
A(°°)exp ' 750 counts/ cm.-mih. 1'6689 cm (4 11)

= 88.25 counts x 2 cm./min. = 2 _
A(CHL‘)exp 750 counts/ cm.-min. °°2353 cm (4 12)

49

= 191.15 counts x 2 cm./min. _ 2
A(C02)exp 750 counts/ cm.-min. ' 0‘5097 cm ' (4-13)

The integrated areas for all experimental runs are given in Table 7.
5. Calculate the weight of each component produced for pyrolysis
Experimental Run #14.
Knowing the area (Aexp) of each experimental component as calcula-
ted in equations (4-11) to (4-13) and its corresponding weight per area
of the sample component as presented in equations (4-8) to (4-10), the

weight of each component produced from the pyrolysis of cellulose can

be found using the following relationship:

Weight of each . .
= A x grams calibration sample.
component produced exp area of calibration sample (4'14)

 

For the three components of interest from Run #14 one finds the

Weight of C0 = (1.6689 cmz) x (1.7831 x 10'3 grams/cmz) = (4-15)
2.9759 x 10"3 grams

Weight of 0H.+ = (0.2353 cm?) x (1.0963 x 10'3 grams/cmz) = (4-16)
2.5801 x 10'“ grams

Weight of C02 = (0.5097 cmz) x (2.5178 x 10"3 grams/cmz) = (4-17)

1.2834 x 10'3 grams~

The various weights for each of the experimental runs are given in
Table 7. Before the weight percent of product (with respect to
cellulose) can be calculated for each component, the weights of both
K2C03 and cellulose present in the 14.5 milligram sample must be deter-

mined. The next two sections will cover how these weights are calculated.

50

6. Weight of K9C01present in the 14.5 milligram sample for Run
#14.

Equation (2-8) of Chapter II, Section E can be rearranged to solve
for the weight of K2C03 since both weight percent and original weight

of the sample are known for each experimental run.

 

Weight of = (weight percent) x (gms of cellulose + gms of K2C03)
chos 100 (4-18)
since

(gms of cellulose + gms of K2C03) = 14.5 milligrams

therefore
weight of K2C03 = (14.5 x 10'5 grams) x (weight percent). (4-19)

Run #14 has a weight percent of 6.3 therefore, using equation (4-19)

the calculated value is:

weight of K2CO3 = (14.5 x 10’5) x 6.3 = 9.1350 x 10“+ grams.
(4-20)

See Table 8 for calculated values.

7. Weight of cellulose in the 14.5 milligram sample for Run #14.

This value is easily computed using the following relationship:

Weight of = weight of original _ weight of K2C03
cellulose sample (14.5 mg) present in sample (4-21)

For Run #14 this value is:

SSITBIOEZ = (14.5 x 10'3 grams) - (9.135 x 10"+ grams) = (4-22)

1.35865 x 10'2 grams.

51

The calculated values for the different experimental runs are presented

in Table 8.

8. Calculate the weight percent (with respect to cellulose)
of each of the pyrolysis components.

 

 

It is important to recognize that this weight percent is with
respect to cellulose only, where as the weight percent defined by
equation (2-8) is with respect to cellulose and K2C03. As the amount
of K2C03 in the 14.5 milligram sample is increased, the amount of
cellulose present decreases. This new weight percent takes into consid-
eration the changing weight of cellulose present for the various weight
percent samples. In this way the quantity of each component produced
for a given amount of cellulose and the effect that K2C03 has on it is

better defined.

 

Weight percent with = weight of pyrolysis component x 100,
respect to cellulose weight of cellUTosegih the sample (4-23)

For Experimental Run #14 the weight percent for each of the components

calculated using equation (4-23) are:

 

 

 

Weight percent = 2.97589 x 10'3 grams x 100 = _
co 1.35865 x 10-2 grams 21°90 (4 24)
Weight percent = 2.580l4 x 10'“ grams x 100 = _
CH1+ 1.35865 x 10'2 grams 1'90 (4 25)
Weight percent = 1.28339 x 10'3 grams x 100 = 9 45 (4-26)
C02 1.35865 x 10‘2 grams '

See Table 8 for the calculated values.

52

9. Weight percent of char with respect to cellulose.

 

This can be easily computed using equation (4-23).

Weight percent = grams of char x 100
of char grams of cellUlose in the sample

 

(4-27)

The weights of the char samples are given in Table 9. For Run #14

the char weight is 1.7 milligrams, thus

1.7 x 10'3 grams x 100

Weight percent = .;2
of char 1.35865 x 10 grams

= 12.51 (4-28)

 

Table 9 gives the calculated values of weight percent of char.

10. Weight of K2C03 lost from the cellulose sample during pyrolysis.
In conducting the neutron activation studies it was observed that
not all of the original K2C03 was still present in the 5.0 and 10.0
weight percent char samples. The estimated weight of K2C03 remaining
in these char samples is given in Table 3. If these weights can be
used as a rough estimate for the amount of K2C03 remaining in the other
char samples then the grams of K2C03 lost as a result of pyrolysis can

easily be calculated as follows:

Weight of K2C03 grams of K2C03 originally grams of K2C03 .
lost during = present in 14.5 mg - remaining in
pyrolysis sample char (Table 3)

(4-29)

For Experimental Run #14 (approximately 5.0 weight percent) the

”Elght.°f K2C°3 ‘95t = (9 135 x 10"“) - (2.0 x 10'“) =
during pyrolySls

7.135 x 10'“+ grams. (4-30)

53

If calculating a 10.0 percent sample then the value used from Table 3
would be 1.05 x 10J grams. The computed values are shown in Table 10.
The amount of K2C03 remaining in the 1.0 weight percent char

samples are approximately equal to the amount originally present.
Therefore these calculations do not apply for the 0.0 percent and 1.0

percent samples.

11. Weight of CO produced from the reaction of K2C03 with char.

 

It has been shown that the following reaction takes place

between K2C03and carbon (char)27.28.29

K2C03 + 2C(char) = 2K + 3C0 (4-31)

In the event that the K2C03 lost during the pyrolysis of cellulose
reacts in this way, the amount of C0 produced from this reaction can
easily be calculated. Since the weight of K2C03 lost during pyrolysis
has been determined by the neutron activation study and for every mole
of K2C03 lost there are three moles of C0 produced, the following

relationship can be used to calculate the desired quantity.

 

Weight of CO produced from K2C03 reaction = (4-32)
grams of K C03lost 1 mole K2C03 x 3 moles C0 x 28 grams CO
from cellu ose x 138.18 gms K2C03 1 mole K2003 1 mole CO

For Run #14 this value is calculated to be:

Weight of CO produced from K2C03 reaction =

7.l35 x 10':ng K2C03 x l mole logo, x 3 moles co x 28 grams co
138.18 gms K2C03 l mole K2003 l mole CO

4.3374 x 10‘“ grams.

Table 10 gives calculated values for the experiments performed.

54

12. Calculate the actual weight of Cinroduced from cellulose.

 

The actual weight of C0 produced from cellulose, excluding that
which is a result of the K2C03 reaction of equation (4-31) is calculated

using the following equation:

grams of CO from the grams of C0 produced
Actual weight = pyrolysis experiment - from K2C03 reaction
of CO (equation 4-15) (equation 4-33)
(4-34)

The calculated value for Experimental Run #14 is:

Actgglcgeight = (2.9759 x 10’3 grams) — (4.3374 x 10’“ grams) =

2.5422 x 10'3 grams. (4-35)
See Table 11 for the calculated values of the various experimental runs.

13. Calculate the corrected weightgpercent of CD.
This weight percent refers to the actual grams of C0 calculated
in the preceding section. This excludes the CD which was a result of

the reaction with K2C03. Therefore:

 

 

Corrected weight = actual grams of Cngroduced x l00 (4-36)
percent of CO weight of cellulose in the sample
For Run #14 the corrected weight percent of C0 is:
Corrected weight = 2.5422 x 10'3grams x 100 =
percent of C0 1.35865 x 10'2 grams 18'71 (4.37)

Table 11 gives the calculated values for the experiments performed.

55

14. Weight of 02H,+ in the control sample.

 

At a later time it was decided to evaluate the CZH.+ (ethylene)
and CZH2 (acetylene) produced in the pyrolysis experimental runs.
A new 0.56 cc sample loop was put in place of the 2.0 cc loop. A
tank of pure CZH, was used to purge the sample loop. The weight of
CzHu can be calculated in the following manner using the same constraints
as found in Part 1 of this section.

Weight of C2H4 = 0.56 cm3 x l mole x 28 grams x 273°K = (4-38)
22,400 cm3 l mole 298°K

 

6.4l28 x lo-hgrams.

15. Integrated area (A) for the CzHu control sample.

 

Only one CZH“ control sample was run to use as a standard for all
experimental runs. The number of integrated counts procured from this
run using the 0.56 cc sample was 122.25. Using equation (4-4) the
integrated area is calculated to be:

122.25 counts x 2 cm./min.

2
750 counts/cm.-min. = 0:3250 cm (4'39)

 

A(C2Hn) =

16. Weight per integrated area for the CZH1+ control sample.

 

Both the weight and integrated area for CZHl+ are known so the

desired value can easily be calculated.

Weight of CZHQ

6.4l28 x 10-”grams = -3 rams
0.3260 cm2 1'957‘ X ‘0 '%£FT‘

 

 

A(C2Hu)
(4-40)

56

17. Calculate the integrated areas for CQHA and CQHQ from
Experimental Run #3. ' T ' '

 

 

Because of technical difficulties (bases line drift on the
recorder, unstable G.C. electrical system at high temperatures) only
a limited number of integrated counts were obtained for Csz and C2H2.
The integrated counts and their corresponding K2CO3 weight percents
are presented in Table 12. The integrated counts from Run #3 are
59.25 for C2H1+ and 23.80 for C2H2' Equation (4-4) is used to calculate

the desired values.

= 59.25 counts x 2 cm./min. = -1 2 _
A(C2Hu) 750 counts / cm.-min. 1.580 x 10 cm (4 41)

 

23.80 counts x 2 cmglmin. = -2 2 _
750 counts / cm.-mih. 6'347 X 10 cm (4 42)

A(C2H2)

The integrated areas for the selected experimental runs are given in

Table 13.

18. Calculate the weights of Can and C2H2 for Experimental Run #3.

 

Since the thermal conductivities of Csz and C2H2 are very
similar, the weight per integrated area calculated for CzHu can be used

for C2H2 with little resulting error. Utilizing equation (4-14) one

finds:
”glatt 0f = (1,530 x 10"1 cm2) x (l.967l x lo"3 grams/cmZ) =
3.1081 x 10'” grams (4'43)
NELAht °f (6.347 x lO'Z an2 ) x(l.967l x 10'3 grams/cmz) =
2

1.2485 x 10'” grams (4-44)

57

Table 13 shows the calculated weights of Cth and C2H2 for the selected

experimental runs.

19. Calculate the weight percent (with respect to cellulose) of
CzHu and CZHZ for Experimental Run #3.

 

For Run #3 the weight of cellulose in the sample was 14.5 milligrams

therefore using equation (4-23) the desired values can easily be

 

 

calculated.

Weight Percent = 3.108l x lo.“ grams x 100 = 2.14 (4-45)
of CZHA 1.45 x 10'[ grams

Weight Percent = l.2485 x 10'7g3ams x 100 -

of 02H. 1.45 x 10—2 grams ‘ 0'86 (4'46)

The calculated values for the various experimental runs are presented

in Table 13.

58

B. Tabulated Data and Calculated Results

Table 4. Control Values: Integrated Counts and Areas.

 

 

Integrated Counts Integrated Areas (x10'2 cmz)
Date CO CH“ C02 A(CO) A(CHt) A(C02)

9-25-84 29.00 25.40 30.00 7.733 6.773 8.000
9-25-84 29.40 25.00 29.60 7.840 6.667 7.893
9-25-84 29.00 25.60 32.00 7.733 6.827 8.533
10-10-84 27.00 24.45 26.90 7.200 6.520 7.173
10-11-84 28.10 25.50 29.30 7.493 6.800 7.813
10-15-84 31.00 26.90 26.0* 8.267 7.173 ---
10-16—84 28.53 24.80 29.45 7.608 6.613 7.853
10-17-84 28.35 25.30 29.65 7.560 6.747 7.907
10-23-84 26.50 26.00 32.00 7.067 6.933 8.533
10-25-84 26.25 23.80 28.83 7.000 6.347 7.688
10-30-84 26.65 24.75 29.73 7.107 6.600 7.928
10-31-84 26.00 24.25 29.55 6.933 6.467 7.880
11-05-84 27.75 25.10 30.00 7.400 6.693 8.000
11-07-84 27.93 25.08 30.24 7.493 6.680 8.128
11-08-84 28.10 25.05 30.48 7.448 6.688 8.064

 

*Recorder pen drifted below base line, data not valid.

59

Table 5. Control Values: Weight/Area (x 10'3 grams/cmz)

 

 

 

 

Grams C0 Grams CH“ Grams C02

Date WIS—UF— A(CH.,) AT 005
9-25—84 1.6141 1.0273 2.4196
9-25-84 1.5921 1.0438 2.4523
9-25-84 1.6141 1.0193 2.2684
10-10-84 1.7336 1.0673 2.6984
10-11-84 1.6657 1.0233 2.4774

10-15-84 1.5099 0.9701 ---

10-16-84 1.6406 1.0522 2.4649
10-17-84 1.6510 1.0314 2.4481
10-23-84 1.7663 1.0036 2.2684
10-25-84 1.7831 1.0964 2.5178
10-30-84 1.7563 1.0543 2.4416
10-31-84 1.8003 1.0761 2.4564
11-05-84 1.6867 1.0396 2.4196
11-07-84 1.6657 1.0417 2.3815

11-08-84 1.6756 1.0404 2.4004

 

60

 

 

Table 6. Experimental Values: Integrated Counts.
Experimental Date of Weight % Integrated Counts
Run # Experiment of K2003 CO CHn 002
1 9-26-84 0.00 843.40 196.60 111.5*
2 9-28-84 0.00 769.00 188.15 123.95
3 10-02-84 0.00 777.00 162.50 109.10
4 10-02-84 0.00 753.50 187.40 119.30
5 l0-04-84 0.00 825.00 193.00 107.50
6 10-04-84 0.00 640.25 125.50 92.0*
7 11-05-84 l.74 615.50 119.50 187.50
8 11-05-84 1.43 546.45 106.75 184.45
9 11-07-84 1.04 490.00 89.00 179.00
10 11-08-84 1.57 694.60 150.30 200.10
11 11-08-84 1.43 575.45 125.00 186.95
12 11-08-84 1.45 569.25 118.25 180.50
13 10-25-84 6.30 574.55 110.18 225.85
14 10-25-84 6.30 625.85 88.25 191.15
15 10-30-84 6.17 621.15 100.50 232.50
16** 10-30-84 6.37 723.65 74.25 377.25
17 10-30-84 6.37 689.00 105.00 203.00
18 10-31-84 6.36 651.90 91.75 252.00
19 11-05-84 6.80 682.20 74.40 405.90
20 10-10-84 13.30 711.25 102.60 194.30
21*** 10-11-84 13.65 540.10 94.75 285.80
22 10—15-84 12.80 710.50 96.00 201.00
23*** 10-16-84 12.80 587.38 9l.43 272.65

Table 6 (continued)

61

 

 

Experimental Date of Wei ht % Integrated Counts
Run # Experiment 0f 2C03 C0 CH8 C02
***
24 10-16-84 13.40 456.75 68.00 270.00
25 10-16-84 13.40 758.25 101.48 189.05
26 10-17-84 13.30 754.38 83.65 289.99
27 10-23-84 12.60 660.75 108.00 195.25
28**** 10-23-84 12 60 341.00 49.00 303.00

 

* Recorder pen drifted below the base line, data not valid.

** Possible air lead in reactor, data not valid.

*** Sample fused to screen, data not valid.

**** Sample fell out of screen, data not valid.

62

 

 

Table 7. Experimental Values: Integrated Areas and Weight of
Pyrolysis Gases Produced.

Exgfigimental Integrated Areas (cmz) ”Piggficgg Pygglysgsaggses

CO CHu 002 CO CH.+ 002

1 2.2491 0.5243 -- 3.6137 0.5401 --
2 2.0507 0.5017 0.3305 3.2949 0.5169 0.7867
3 2.0720 0.4333 0.2909 3.3292 0.4464 0.6924
4 2.0093 0.4997 0.3181 3.2285 0.5148 0.5772
5 2.2000 0.5147 0.2867 3.5349 0.5302 0.6823

6 1.7073 0.3347 -- 2.7433 0.3448 --
7 1.6413 0.3187 0.5000 2.7685 0.3313 1.2098
8 1.4572 0.2847 0.4919 2.4579 0.2959 1.109l
9 1.3067 0.2373 0.4773 2.1765 0.2472 1.1368
10 1.8523 0.4008 0.5336 3.1036 0.4170 1.2808
11 1.5345 0.3333 0.4985 2.5712 0.3468 1.1967
12 1.5180 0.3l53 0.4813 2.5435 0.3281 1.1554
13 1.5321 0.2938 0.6023 2.7320 0.3221 1.5164
14 1.6689 0.2353 0.5097 2.9759 0.2580 1.2834
15 1.6564 0.2680 0.6200 2.9092 0.2826 1.5138

16 -- -- -- -- -- --
17 1.8373 0.2800 0.5413 3.2270 0.2952 1.3217
18 1.7384 0.2447 0.6720 3.1297 0.2633 1.6507
19 1.8192 0.1984 1.0824 3.0685 0.2063 2.6190
20 1.8967 0.2736 0.5181 3.2880 0.2920 1.3981

21 -— —- -- -- -- --
22 1.8947 0.2560 0.5360 2.8607 0.2483 1.3245

Table 7 (continued)

63

 

Weight of Pyrolysis Gases

 

Expgglmgntal Integrated Areas (cm?) Produced (x10.8 grams)
00 CH, 002 CO CH“ 002
23 -- -- -- -- -- --
24 -- -- -- -- -- --
25 .0220 .2706 .5041 3.3173 .2847 .2426
26 .0468 .2231 .7733 3.3213 .2301 .8932
27 .7620 .2880 .5207 3.1122 .2890 .1811

28

 

64

 

 

Table 8. Experimental Values: Weight of K2003 and Cellulose in the
14.5 Milligram Sample, Weight Percent (with respect to
cellulose) of Products.

Experimental Weight (grams) Weight % of Products
Run # K2003(x10'“) cellulose (x10‘2) co 0H.+ 002
1 0.0000 1.4500 24.92 3.72 --
2 0.0000 1.4500 22.72 3.56 5.43
3 0.0000 1.4500 22.96 3.08 4.78
4 0.0000 1.4500 22.27 3.55 5.22
5 0.0000 1.4500 24.38 3.66 4.71
6 0.0000 1.4500 18.92 2.38 -—
7 2.5230 1.4248 19.43 2.33 8.49
8 2.0735 1.4293 17.20 2.07 8-33
9 1.5080 1.4349 15.17 1.72 7.92
10 2.2765 1.4272 21.75 2.92 8.97
11 2.0735 1.4293 17.99 2.43 8.37
12 2.1025 1.4290 17.80 2.30 8.09
13 9.1350 1.3587 20.11 2.37 11.16
14 9.1350 1.3587 21.90 1.90 9.45
15 8.9465 1.3605 21.38 2.08 11.13
16 -- -- -- -- --
17 9.2365 1.3576 23.77 2.17 9.74
18 9.2220 1.3578 23.05 1.94 12.16
19 9.8600 1.3514 22.71 1.53 19.38
20 19.2850 1.2572 26.15 2.32 11.12
21 -- -- -- -- --
22 18.5600 1.2644 22.63 1.96 10.48

Table 8 (continued)

65

 

 

Experimental Weight (grams) _2 Weight % of Products
Run # K2003(x10 ) cellulose (x10 ) CO CH4 002
23 -- -- -- -- --
24 -- -- -- -- --
25 19.4300 1.2557 26.42 .27 9.90
26 19.2850 1.2572 26.42 .83 15.06
27 18.2700 1.2673 24.56 .28 9.32

28

 

 

66

Table 9. Experimental Values: Char Weights and Weight Percents.

 

 

Experimental
Run # Milligrams of Char Weight % of Char
1* -- --
2* -_ --
3* __ __
4* _, __
5 0 10 0.689
6 0 30 2.068
7 1.10 7.721
8 1.10 7.696
9 1.10 7.666
10 1 20 8.408
11 1.10 7 696
12 1.10 7.698
13 2.40 17.665
14 1.70 12.512
15 1.80 13.230
16 0.80 --
17 1.70 12.522
18 1.60 11.783
19 1.00 7.400
20 2.40 19.091
21 3.80 --

22 1.80 14.236

Table 9 (continued)

67

 

 

Experimental
Run # Milligrams of Char Weight % of Char
23 -- --
24 -- --
25 1.60 12.742
26 1.40 11.136
27 2.20 17.360
28 —- --

 

*Sample not collected.

68

Table 10. Experimental Values: Weight of K2003 Lost and CO
Produced from the K2003 Reaction.

 

 

Weight of K2003 LOSt Weight of CO Produced
Experimental DuringBPyrolySis From KgCO3 Reaction
Run # (x10 grams) (x10 “ grams)
1 -- --
2 -- --
3 -- --
4 -- --
5 -- --
6 -- --
7 NC NC
8 NC NC
9 NC NC
10 NC NC
11 NC NC
12 NC NC
13 0.7l35 4.3374
14 0.7l35 4.3374
15 0.6947 4.2228
16 -- --
17 0.7237 4.3991
18 0.7222 4.3903
19 0.7860 4.7781
20 1.8235 11.0851

21 -- --

69

Table 10 (continued)

 

 

Weight of K2003 Lost Weight of CO Produced
Experimental During Pyrolysis From K2003 Reaction
Run # (x10’3 grams) (x10’4 grams)

22 1.7510 10.6444

23 -- --

24 -- --

25 1.8380 11.1733

26 1.8235 11.0851

27 1.7220 10.4681

28 -- --

 

NC - No Change

70

Table 11. Experimental Values: Corrected Weight and Weight Percent

 

 

of 00.
Experimental Actual Weight of CO Corrected Weight
Run # Produced (x10“3 grams) Percent of CO
1 -- --
2 -- _-
3 -_ --
4 -_ --
5 -- _-
5 -- --
7 NC NC
8 NC NC
9 NC NC
10 NC NC
11 NC NC
12 NC NC
13 2.2982 16.92
14 2.5422 18,71
15 2.4869 18.28
16 -- --
17 2.7871 20.53
18 2.6907 19.82
19 2.5907 19.17
20 2.1795 . 17.34
21 --

22 1.7963 14,21

71

Table 11 (continued)

 

 

 

Experimental Actual Weighs of CO Corrected Weight
Run # Produced (x10“ grams) Percent of CO
23 -- __
24 -- __
25 2.1999 17.52
26 2.2128 17.60
27 2.0654 16.30
28 -- __
NC - No Change

72

 

 

 

 

 

Table 12. Experimental Values: Integrated Counts of C2H2 and 02H“
Experimental Date of Weight % Integrated Counts
Run # Experiment of K2003 02H4 CZHZ
3 10-02-84 0.00 59.25 23.80
4 10-02-84 0.00 48.00 28.75
7 11-05-84 1.74 28.18 5.00
8 11-05-84 1.43 32.13 14.75
15 10-30-84 6.17 28.65 5.45
19 11-05-84 6.80 18.68 6.90
22 10-15-84 12.80 13.25 23.50
26 10-17-84 13.30 18.70 21.30
Table 13. Experimental Values: Integrated Area, Weight and Weight
Percent of C2H2 and C2H4.
Expgrimental Integrated Area Weight of Pyrolysis Weight %
un # (X10"2 cm2 Gases (x10 grams) of Products
02H“ 02H2 02 H 02H2 02H“ 02H2
3 15 8000 .3467 3.1081 1.2485 2.14 0.85
4 12 8000 .6667 2.5179 1,5081 1,74 1.04
7 7 5147 .3330 1.4782 0,2523 1.04 0.18
8 8 5680 .9333 1.6854 0 7737 1.13 0.54
15 7 6400 .4533 1.5029 0 2359 1.10 0.21
19 4 9813 .8400 0.9799 0 3520 0.73 0.27
22 3 5333 .2667 0.6950 1,2330 0,55 0.93
25 4 9867 .6800 0.9809 1.1173 0.73 0.89

 

73

0. Summary of Experimental Results.

Tables 14 and 15 give the averaged product weight percent values
for the flash pyrolysis of cellulose at different K2003 weight percents.
It is quite evident that the presence of K2003 in the cellulose samples
influences the product yields.

Table 14 shows the char yields for various K2003 weight percents.
It is observed that as the amount of K2003 is increased in the cellulose
sample, the amount of char remaining after pyrolysis increases as well.
Neutron activation studies of the char show that the 1.0 percent samples
contain approximately the same amount of K2003 as was originally
impregnated in the samples. This results in a 0.0 weight fraction loss
of K2003 for the 1.0 percent samples. The weight fraction of K2003
lost for the 5.0 percent and 10.0 percent char samples are 0.78 and 0.95,
respectively. Therefore a majority of the K2003 originally present in
these samples is in some way lost.

An important conclusion was reached from the neutron activation
study. The amount of K2003 remaining in the 1.0, 5.0 and 10.0 weight
percent char samples is comparably equal to or less than that originally
impregnated in the 1.0 percent samples. This indicates that the
increase in the weight of the char samples with increased loading of
K2003 in some way increases the amount of char produced just by its
presence when pyrolysis of cellulose takes place.

Table 15 gives the gas yields for various K2003 weight percents.
The weight percent of 00 declines then increases to a value slightly
above that of the unloaded sample for the 10.0 weight percent sample.
Ifthe K2003reaction with char (equation 4-31) is taken into considera-

tion then the 00 yield declines and reaches a minimum at the 10.0

74

percent sample. There is a continuous decline in methane (CH“) and
ethylene (C2H4) as the loading of K2003 is increased. Methane yields
decline by nearly 50 percent and ethylene by 67 percent when comparing
the unloaded and 10.0 percent samples. Acetylene (02H2) yields
decline as K2003 loading increases until the 10.0 percent sample. At
this weight percent the yield is approximately equal to the unloaded
samples.

The total gas yields are given in the two farthest columns
to the right in Table 15. The following trends are observed when
comparing the loaded to the unloaded samples for the uncorrected total
gas yield. Thereisa 13.0 percent decline in gas yields for the 1.0
percent samples. The 5.0 percent samples show a 3.3 percent increase
in gas yields. The most significant change is found in the 10.0
percent samples, where a 12.3 percent increase in gas yields is observed
when compared to the unloaded samples. For the corrected total gas
weight precents one finds a decline in yield for all K2003 catagories in
comparison to the unloaded samples. This shows the significant influence
that the K2003 reaction can have upon the total product gas yields.

Some of the experimental runs were not used in analysis of gas
yields for various physical reasons: recorder pen drifting below the
base line, possible air leak in the reactor, and sample fused to or
falling out of the screens. Other than samples rejected for these
reasons, a majority of the integrated values were found to be very
close to one another (within 15 percent of the averaged value) for a
given K2003 loading category. Occasionally a sample run or even a
particular gas within a run would be noticeably different when compared

to other values in the same weight percent category. Such is the case

75

for Experimental Run #6 and CO2 values from Run #19 and #26. Because
of the large discrepancy these values were not used in calculating
the averaged numbers given in Table 15.

When a sample is pyrolyzed in the reactor the tar materials
collect on the copper electrodes and the inside wall of the alumina
liner. Although actual quantitative analysis of tar yields were not
conducted, the following trend was observed. It was visually observed
that as K2003 loading of the samples increased, there was a proportion-
ate decline in tar yields. At 10.0 percent loading there was very
little tar produced compared to the unloaded samples.

In conclusion the tar and char yields of pyrolysis appear to be
the most influenced by K2003 loading of cellulose. The gas yields are

shown to vary depending upon the weight percent of K2003.

76

 

 

Table 14. Char Yields and Neutron Activation Study (”ZK).*
Weight % Weight % Weight of K2003 (grams) Weight Fraction**
of K2003 Char Initial Final of K2003 Lost
0.0 1.38 0.00 0.00 0.00
***
1.0 7.81 2.09 x IO'L+ 2.00 x 10"+ 0.00
5.0 12.52 9.26 x 10"+ 2.00 x 10'“ 0.78
10.0 14.91 18.97 x 10’1+ 1.05 x 10-“ 0.95

 

* All table data are averaged values from six experimental runs
in each K2003 weight percent category.

** Weight fraction of K2003 lost = (initial-final)/initial.

*** 0.00 within experimental error.

 

 

Table 15. Average Gas Weight Percent Values.
Weight % * Average Weight Percent Values Corrected #
of K2003 CO CO CH4 002 02H4 02H2 Total Total
** ** **
0.0 23.45 23.45 3.51 5.04 1.94 0.95 34.89 34.89
1.0 18.22 18.22 2.30 8.36 1.11 0.36 30.35 30.35
5.0 22.15 18.91 2.00 10.73*** 0.92 0.24 36.04 32.80
10.0 25.24 16.59 2.13 10.2l"a 0.67 0.94 39.19 30.54

 

* Corrected weight percent of CO.

** Experimental Run #6 not used in averaging values.

*** Experimental Run #19 not used in averaging values.

@ Experimental Run #26 not used in averaging values.
# Corrected Weight Percent of 00 used in summing values.

CHAPTER V

CONCLUSION

The pyrolysis of K2003-loaded cellulose samples at rapid heating
rates (>300°C per second) and high temperatures (900°C) was conducted
in order to answer the questions outlined in Chapter I, Section D. This
chapter deals with the experimental data collected and the interpretation
of the this data so as to attempt to answer the before mentioned ques-
tions. The data collected from the pyrolysis experiments, as presented
in Chapter IV, was found to be very consistant within the different K2003
weight percent categories. A summary of the experimental results is
given in Chapter IV, Section C.
A. Comparison of Slow Versus High Temperature Flash Pyrolysis of

Cellulose in the Presence of K2003.

It has been shown that the product yields of pyrolysis are influ-

enced more by elevated peak temperatures (>600-700°C) than either resi-

9’10 A number of slow pyrolysis studies

4,17,18,

dence time or heating rates.
have been conducted which use additives to alter product yields,
20’21’22 Slow pyrolysis of cellulose in the presence of K2003 has been
shown to increase gas and char yields while at the same time decreasing

17’ 20’ 21’ 22 The studies with K2003 were conducted at low

tar yields.
temperatures and slow heating rates; therefore, it is questioned whether
this particular additive has any influence upon product yields at

elevated temperatures and rapid heating rates.

77

78

Tables 14 and 15 give the char and gas yields from the flash
pyrolysis of cellulose at 900°C for varying K2003 weight percent samples.
When comparing samples loaded with K2003 to those which were not impreg-
nated (unloaded), the influence is obvious. The presence of K2003
results in increased char yields. The total gas yields appear to vary
depending on the weight percent of K2003 in the sample. Visual inspec-
tion of the tar condensing on the inside of the reactor shows a decline
in yield as a result of K2003 being present. Thus all three of the
major pyrolysis product groups are influenced by K2003 at an elevated
temperature of 900°C and heating rates greater than 300°C per second.

As shown in Table 14, the presence of K2003 increases the yield
(weight) of char in comparison to unloaded samples. This same increase
in char yield has been observed in low temperature slow pyrolysis
studies in which K2003 or other similar inorganic salts were used.17’22’23
Table 14 reveals that with increased loading of K2003 there is a corre-
sponding increase in the char production. This same trend has been

17 When comparing the 0.0 and

demonstrated in a slow pyrolysis study.
10.0 percent char sample yields, there is an eleven-fold increase as a
result of K2003 being present.

The uncorrected total gas yields (see Table 15) are found to vary
with K2003 loading. The increased loading of K2003 favors CO and CO2
production, at the same time resulting in a decline in hydrocarbon gas
yields (CH4, Csza CZHH). These same gas yield trends have been

22,23 The 1.0 percent samples show a

reported in slow pyrolysis studies.
decline in total gas yield in comparison to the unloaded samples. The
5.0 and 10.0 percent samples give an increased total gas yield for the

same comparison . A general increase in total gas yields have been

79

reported in low temperature slow pyrolysis studies.22’23

A possible
explanation for the decline in total gas yields observed in the 1.0
percent samples is that this study did not take into consideration the
water yield as was done in the slow pyrolysis studies.

Table 15 shows that there is a decline in the corrected 00 gas
yields when compared to the unloaded samples. This results in a corrected
total gas yield for loaded samples less than the unloaded samples.

Although quantitative analysis of tar yields were not conducted in
this study, visual inspection of the copper electrodes and inside wall of
the reactor revealed that as the K2003 loading increased, there was a
continued decline in tar production. This same decline in tar production
resulting from K2003 being present has been reported in several slow
pyrolysis studies.17’22’23

It is therefore concluded that the additive, K2003, continues to
alter pyrolysis product yields even at high temperatures and rapid heat-
ing rates. This study also shows that the same trends for product
yields from slow pyrolysis can be expected for high temperature flash
pyrolysis when K2003 is used as the additive. For both slow and high
temperature flash pyrolysis one finds that with increased K2003 loading

there is an increase in both char and total gas yields. At the same

time there is a corresponding decline in tar production.

B. K2003 After the Pyrolysis of Cellulose

A limited number of studies have been conducted which refer to
the state of or location of additives after pyrolysis. A slow pyrolysis
study of wood bark by Ross and Fong 2° used x-ray analysis to detect the

presence of K2003 on the surface of char samples. Unfortunately this

80

method did not allow a comparison of the initial and final weight of
K2C03 remaining on the char therefore, quantitative loss of K2003 due to
pyrolysis was not determined. Another slow pyrolysis study of cellulose

by Byrne et al.24

used several different flame-retardants. This study
showed that for a majority of the flame-retardants used the amount of
retardant originally impregnated in the cellulose was still remaining
on the char. Smaller amounts of some of the retardants were however,
found in the tars. Unfortunately, K2003 was not among the additives
used in this study.

One can therefore see the need for experimentation which better
defines the location of K2003 after pyrolysis. Experimentation would
also answer the question of whether the observed increases in char
yield with increased loading of K2003 is a result of the additional
weight of K2003 remaining on the char or whether there is an actual
altering of the product distribution.

In order to answer the above questions neutron activation of char
samples were carried out. For details on the method and results from
this activation study see Chapter III, Section F.

Table 14 gives the initial weight of K2003 impregnated for the
various sample weight percents as well as the corresponding final weight
of K2003 remaining in the char determined by neutron activation. Within
experimental error it has been determined that the 1.0 percent char
samples contain all of the K2003 originally loaded onto the cellulose.
Both the 5.0 and 10.0 percent char samples, however, contain an amount
of K2003 less than or equal to that loaded onto a 1.0 percent sample.
This means that for the 5.0 percent samples 78 percent of the K2003

originally impregnated on the cellulose no longer remained on the char.

81

Forifln210.0 percent samples 95 percent of the K2003 is lost from the
char. This proves that the increase in char weight observed for
increased loading of K2003 is not a result of K2003 remaining on the
char and that the presence of K2003 actually altered the product yields
of char.

Through neutron activation of char samples it is shown that the
5.0 and 10.0 percent char samples contain less K2003 then was originally
impregnated in the cellulose. A possible explanation for the K2003 not

present on the char is covered in the next section.

0. K2003 : Catalyst or Reactant?

Is K2003 a catalyst or a reactant? By definition a catalyst is a
substance which affects the rate of the reaction, but is recovered from
the reaction unchanged. A reactant is a substance which actually changes
in the reaction.

In referring to Table 14 it is observed that the presence of 1.0
weight percent K2003 results in a 5.7-fold increase in char formation in
comparison to the 0.0 percent samples. Table 15 shows that there is
also an appreciable difference between gas yields for these two weight
percents. However, as discussed in the preceding section and as
presented in Table 14, there appears to be no loss of K2003 from the
char for the 1.0 percent samples. This suggests that for the 1.0 percent
samples K2003 is not reacting during pyrolysis and is in actuality act-
ing as a catalyst.

Tables 14 and 15 show that the 5.0 and 10.0 percent samples sig-
nificantly influenCElXHfllthe char and gas yields. However, values from

Table 14 show that for these weight percent samples an appreciable

82

amount of K2003 originally present in the cellulose is no longer present
in the char. Ah equation for the reaction of K2003 with char was
presented in Chapter IV, Section A. This reaction is found to be

favored at high temperatures (>800°C).
K2003 + 20(char) = 2K + 300 (4-31)

As observed in Table 15, all of the indivdual gas yields either contin-
ually decline or increase as the loading of K2003 is increased. A
slight variation is found for the 10.0 percent samples of acetylene. The
obvious exception to this case are the 00 yields which first decline at
1.0 percent loading then increase for the 5.0 and 10.0 percent samples.
If it is assumed that the K2003 lost from the char in the 5.0 and 10.0
percent samples is reacting according to equation (4-31), then the 00
produced from this reaction can be subtracted from that which is actually
produced in pyrolysis. When this is done, as shown in Table 15, there
is in actuality a continual decline in 00 production. Therefore, one
finds the same consistent trends which are observed for the other gas
products of pyrolysis. The irregular product yields of 00 suggest that
K2003 might in fact be reacting with the char to produce 00. If this is
the case then the excess K2003 is acting as a reactant for the 5.0 and
10.0 weight percent samples.

When comparing the char yields for the 1.0 and 10.0 percent
samples there is less than a two-fold increase for a ten-fold increase in
K2003. However, comparison of the 0.0 and 1.0 percent samples show a
5.7-fold increase in char formation. The corrected total gas yields
presented in Table 15 show the same trends as the char yields. That is.

the corrected total gas yields are initially more affected by the

83

1.0 percent samples then the higher K2003 loaded samples. This further
suggests that the 1.0 percent samples have a catalytic influence upon
product yields where as the higher K2003 loaded samples have an excess
of K2003 which most likely acts as a reactant in a reaction such as

that given by equation (4-31).

0. Chemistry of Pyrolysis in the Presence of K2003.

As discussed in the preceding sections of this chapter, the
presence of K2003 is found to alter pyrolysis product yields. Referring
to Figure 1, the presence of K2003 favors the upper and lower pathways
which leads to increased yields of char and gas.

When a cellulose sample is flash pyrolyzed in the screen design
reactor used in these experiments, the tar materials produced are
almost instantly removed from the area of the hot screens and collect on
the inside wall of the reactor. Because of the rapid heating rate and
rapid removal of the tarry products, their further chemical breakdown
during pyrolysis is unlikely.

Under normal conditions (without additives present) the pyrolysis
of cellulose favors the breakage of the bonds labeled (d) in Figure ICL23
This results in the formation of levoglucosan. Madorsky et al.25
proposed that additives catalyze the breakdown of cellulose by cleavage
of the 0-0 bond (bonds a,b, and c in Figure 10) and that this results in
the destruction of the hexose units. This in turn facilitates the
formation of more char and gases. Since the removal of levoglucosan
from the area of the hot screens is so rapid, thus removing any likely-

hood of secondary reactions, this pr0posal appears to be a very valid

and plausible explanation for the influence of additives on pyrolysis

84

 

zcweuadueamd

 

=.=g sea ”.ml

w:

mmo_:__mu mo mwm>Pocxa

 

mwed=dduu

 

.02 623822

85

product yields.

E. Conclusions

Studies onthe high temperature flash pyrolysis of cellulose in
the presence of K2003 have been conducted in order to answer the questions
outlined in Chapter 1, Section D. The experimental research conducted
helped answer the questions as is outlined in the following paragraphs.

At elevated temperatures and rapid heating rates the presence of
K2003 when pyrolyzing cellulose has been shown to significantly alter
product yields. The same product yield trends observed for slow pyroly-
sis can be applied to high temperature flash pyrolysis when K2003 is
used. That is, as the loading of K2003 is increased there is a corre-
sponding increase in char and total gas yields and a continual decline in
tar production.

Neutron activation studies of the char products show that the
weight increase in char, as K2003 is increased, is not a result of
K2003 remaining on the char. The majority of the K2003 originally impreg-
nated on the 5.0 and 10.0 percent samples was lost during pyrolysis.
A plausible explanation for the loss of this K2003 is that at low
loading (1.0 weight percent) K2003 acts as a catalyst; however, at
increased loading (5.0 andIILO percent samples) experimental data suggest
that the excess K2003 acts as a reactant.

A previously proposed theory is used to help explain on a molecular
level the influence of K2003 on pyrolysis product yields.

A commercial application of these results is possible. A full-
scale wood pyrolysis unit could be run at lower heating rates and

temperatures when evaluating different additives. Running a large system

86

at these conditions would result in substantial energy saving as compared
to high temperature and heating rates. Once an additive is found which
gives a favorable product yield, the system could then be run at elevated
temperatures and heating rates where similar product trends could be

expected.

F. Suggestions for Future Work.

There are many areas both within this study as well as pyrolysis in
general in which further research can be conducted. Most importantly,
a more complete material balance on pyrolysis products (including H20,
H2, and tar weight yields) for the ongoing research at Michigan State
University would better define a molecular explanation for the influence
of K2003 on pyrolysis product yields. The use of different additives to
determine those which produce increased yields of materials for use in
the chemical or petro-chemical industries is another area in which
high temperature flash pyrolysis studies might be conducted. The area
of biomass pyrolysis has many unanswered questions which may be only

resolved through further theoretical and experimental research.

LIST OF REFERENCES

LIST OF REFERENCES

0 1Milne, T., "Pyrolysis--the thermal behavior of biomass below
600 C." ”A Survey of Biomass Gasification," 2, Chapter 5, Prepared
by the Solar Ehergy Research Institute/TR-33-239 July, 1979.

 

2Graboski, M. and Bain, R., "Properties of biomass relevant to
gasification," "A Survey of Biomass Gasification," 2, Chapter 5,
Prepared by the Solar Energy Research Institute/TR-33-239 July, l979.

 

3Shafizadeh, F., "Chemistry of pyrolysis and combustion of wood,"
"Progress in biomass conversion,"(Academic Press. Inc-s 1982). PP.
5l-76.

 

4Hixson, A.N. and Hsu, 0.0., "01 to 04 oxygenated compounds by
promoted pyrolysis of cellulose," Ind. Eng. Chem. Prod. Res. Dev.,
29, 1981, pp. l09-ll4.

5DeJenga, C.I., Antal Jr., M.J., and Jones Jr., M., "Yields
and composition of sirups resulting from the flash pyrolysis of
cellulosic materials using radiant energy," Journal of Applied Polymer
Science (l982), 27, pp. 4313—4322.

6Diebold, J. and Scahill, 0., "Ablative fast pyrolysis of biomass
in the entrained-flow cyclonic reactor at S.E.R.I.," Presented at the
14th Biomass Thermochemical Conversion Contractors Meeting, Arlington,
Virginia June 23-24, l982 U.S. Department of Energy contract # DE-
A006-76RLO 1830.

7Thurner, F., Mann, 0., and Beck, 5., "Kinetic investigation of
wood pyrolysis," Prepared for Department of Energy, Division of Solar
Energy. contract # DE-ASO4-79ET2004l, June l980.

8Caubet, 5., Corte, P., Fahim, C. and Traverse, J.P., "Thermo-
chemical conversion of biomass: Gasification by flash pyrolysis study,"
Solar Energy (T982), 29, No. 6, pp. 565-572.

 

9Hajaligol, M.R., Howard, J.B.,Longwell, J.P. and Peters. W.A.,
"Product compositions and kinetics for rapid pyrolysis of cellulose,"
Ind. Eng. Chem. Prod. Res. Dev. (l982),‘21, pp. 457-465.

 

87

88

1OAntal Jr., M.J., Friedman, H.L., and Rogers, F.C., "Kinetics
of cellulose pyrolysis in nitrogen and stream," Combusion Science
and Technology (1980), 21, pp. 141-152.

 

 

1‘Scott, 0.5. and Piskorz, 3., "The flash pyrolysis of aspen-
poplar wood," The Canadian Journal of Chemical Engineering (1982),
.Qg, pp. 666-674.

 

12Brink, D.L. and Massoudi, M.S., "A flow reactor technique for
the studycf'wood pyrolysis 1. Experimental," Journal of Fire and
Flammability, 9, April 1978, pp. 176-188.

13Caubet, S., Corte, P., Fahim, C. and Traverse, J.P., "Gaseous
fuel from biomass by flash pyrolysis,"'Energy from Biomass"(1981),
Applied Science Publishers, pp. 542-547.

14Deglise, X., Richard, 0., Rolin, A., and Francois, H., "Fast
pyrolysis/gasification of lignocellulosic materials at short residence
time,"'Energy from Biomass"(1981), Applied Science Publishers, pp.
548-553.

 

 

 

 

15Graef, M., Allen, G.G. and Krieger, 8.8., "Product distribution
in the rapid pyrolysis of biomass/ligin for production of acetylene,"
"American Chemical Society Symposium Series," 144, 1981, pp. 293-311.

16Niksa, 5.0., Russel, W.B. and Saville, D.A., "Captive sample
reactor for kinetic studies of coal pyrolysis and hydropyrolysis
on short time scales," Fuel (1982), o1, pp. 1207-1212.

17Pyle, 0.1. and Zaror, 0.4., "The effect of alkali salts on
low temperature pyrolysis," Department of Chemical Engineering and
Chemical Technology, Imperial College, London SW7 2AZ, England.

18Hilado, 0.0., and Brandt, D.L., "Char yield and flash-fire
propensity of pyrolysis gases from materials," Journal of Fire and
Flammability, 2, Oct. 1978, pp. 553-557.

19Steinberg, M. and Fallon, P.T., "Flash pyrolysis of biomass
with reactive and non-reactive gases," work performed for biomass
energy technology division, U.S. Department of Energy, Contract # DE-
A002-76CH00016.

20Ross, R.A., and Fong, P., "Catalytic conversion of wood barks to
fuel gases," Ind. Eng. Chem. Prod. Res. Dev. (1981), 29, pp. l97-203.

21Fung, D.P.C., Tsuchiya, Y., and Sumi, K., "Thermal degradation
of cellulose and levoglucosan -the effect of inorganic salts," Wood
Science, 5, No. 1, July 1972.

 

 

 

89

22Tsuchiya, Y., and Sumi, K., "Thermal decomposition products
of cellulose," Journal of Applied Polymer Science (1970), 14, pp.
2003-2013.

23Shafizadeh, F., "Pyrolysis and combusion of cellulosic materials,"
Carbonydrate Chemistry (1968), 23, pg. 419.

24Byrne, G.A., Gardiner, D. and Holmes, F.H., "The pyrolysis of
cellulose and the action of flame-retardants. 11. Further analysis
and identification of products,“ J. Appl. Chem., (1966), 16,

25Madorsky, S.L., Hart, V.E., and Straus, S., "Pyrolysis of
cellulose in a vacuum," J. Res. Natl. Bur. Std. (1956), 56, p. 343.

26Brunner, F.H., and Roberts, P.V., "The significance of heating
rate on char yield and char properties in the pyrolysis of cellulose,"
Carbon, 18, pp. 217-224.

27McKee, D.W. and Chatterji, D., "The catalyzed reaction of
graphite with water vapor," Carbon (1978), 16, pp. 53-57.

2°Huhh, F., Klein, J. and Juntgen, H., "Investigation on the
alkali-catalyzed steam gasification of coal: kinetics and inter-
actions of alkali-catalyst with carbon," Proc. Int. Symp. on
Catalyzed Carbon and Coal Gasification, Amsterdam, 1982.

29Wen-Yang, Wen, "Mechanisms of alkali metal catalysis in the
gasification of coal, char, or graphite," Catal. Rev.-Sci. Eng.,

22(1), 1980, pp. 1-28.

 

 

APPENDIX I

90

PART A - Calculating the resistance in 99.0 percent alumina.

Resistance = p L/A
L = Length = 1/4 inch
A = area = l in. x 1/2 in. = 0.5 in2

p resistivity

from Alumina as a Ceramic Material compiled and edited by Walter Gitzen,
American Ceramic Society (1970) page 79 for 99.0 percent alumina
a = 8x106 ohms-cm at l,000°c

thUS 6
’(8x10 ohm-cm) x (0.25in.) x31 in.

Resistance = (0.5 in.2) 2.54 cm

Resistance = 1,574,804 ohm

PART B - Calculating the power requirements for heating the screens
to 1,000 0.

An energy balance around the screens gives

Power required to _ Total energy lost
heat screens to surroundings

This is assuming the system is nonadiabatic. The energy lost to the
surroundings is in three forms; conduction, radiation, and free convec-

tion. Therefore,

Total energy lost = Energy lost due .+ Energy lost due
to surroundings to conduction to radiation

.1.

Energy lost due to
free convection

91

A. Calculate the energy lost due to conduction -

For unsteady-state the energy equation is:

A

pCp §%-= (v.l<VT) (See page 352 from Transport Phenomena by
Bird, Stewart and Lightfoot)

Assuming the thermal conductivity is independent of temperature

8T 2 .52. Y
(A) 53- = av T where o. = p p

()3:

screen

at t = 0 the surface of the screen is raised to T1 and maintained

at that temperature for t>0. In dimensionless form equation (A)

becomes
(B) 80 820 h (T 'To)
— = a —'2— W ere 9 = TT—T-T'

The boundary conditions are

1.0. t<0 T = To = 20°C 0 = 0 for all y
8.C. 1 y = 0 o = l T = T1 = l,000°C for all t>0
8.C. 2 y = w o = 0 T = T for all t>0

O

8.0. 2 can be applied to the problem since at y = 3/8 in.,

T = To thus y = «>. T = T

o .
The solution to the problem (equation (8)) is

y// 4011'.
e'nzdn

 

‘0
O“.

(C) -;—:;Q- 1 - erf (y/ J55? )

92

since (0) ql = -2A -——-
y = 0 y = 0

91
y 0 Mnat 1 0

 

integrating this equation with respect to time yields

9

2 l

99 or qu=f AK (T-T)dt
ql 7581'

01"

ZAnl/Zfll - Io)

(E) .sq = q conduction = /;;-

Solving this equation gives the energy lost due to conduction in a
helium atmosphere.

In heating the screens from 20°C to 1,000°C the properties of
density p, viscosity u, thermal heat transfer coefficient K and Cp
will most likely change. Therefore the values used are average values
between 20°C and 1,000°C or as close to 510°C as possible.

From Heat [nagsfer by J.P. Holman, fourth edition, McGraw
Hill Book Company, for helium at 800°F (527°C) the following values

were found
0 = 0.06023 Kg/m3 p = 38l.7-x 10’7 Kg/m - s
Cp = 5.2 KJ/KgoC K = 0.275 w/m°C
a = 8.774 x 10'4 mz/sec

93

Since t = 1 second, To = 20°C, T = l,000°C

l

A = area of one side 2

0f the screen = 2 1n'

Because both sides of the screen will conduct heat the screen area is
taken as twice as large, so
A = 4 in.2

The solution to equation (E) is

 

 

. 2 l/2
qconduction = /E_i*4 1n° 1 0.575w l J/sec ((1 SEC) -i
n m C W
(1000-20)°C l(2.54 cm)2 . lm2
7 I i
(8.774 x 10-4 mz/sec)1/2 l in.2 (100 cm)2
qconduction = 26.49 Joules/sec

. Calculating the Energy lost due to Radiation -

Radiation between two non-black surfaces is given by

(F) 4)

0A 2

= 4
ql,2 1 51,2 (T1 ' T

For one gray surface completely surrounded by another (concentric

spheres or cylinders)

 

5 = (overall interchange =
1,2
factors) 1 A 1
t if (5")
l 2 2
A1 = area of enclosed surface (wire mesh) =
A1 = 2 in2 x 2 (for both surfaces) = 4 in.2
A2 = area of ceramic liner = anL = 2n(0.75 in) (3.0 in)
A = 7.64159 in2

94

 

E1 = emissivity of metal = 0.6
E2 = emissivity of ceramic = 0.9
T1 = temperature of screen = l,000°C = 1,273°K
T2 = temperature of ceramic = 20°C = 293°K
thus
E 1
1,2 ‘ .
__l__ + 4171.7 2(1 _-|)
O 6 . 1 Tn 079
51,2 = 0.579768

0 = Stefan - Boltzman Constant = 5.676x10'8 W/mz 0K4

using equation (F).

5.676x10'8w J4 in2

qradiation
.r‘
62'0K4

10.5797681[(1,273)4 - (29314] °K
—l —I

2

 

J/sec 1 (2.54m)2 . 1m
1 . 2 I 2
W lln. (100cm)

qradiation 222.39 Joules/sec

. Calculate the Energy lost due to Free Convection -

 

2 3
Gr = Grashof Number = [ p B 9 D2 (Tl ’ To) ]
LI
3 _ -7 .
p = 0.05023 Kg/m p — 381.7x10 Kg/m - sec
9 = gravity = 9.8 m/secz B = l/T0° = 1/293°K
D = average distance from screen Pr = Prandtl Number = 0-72
to wall of the cylinder = 3/8 in.
T1 = l,000°C = 1,273°K
T = 20°C = 293°K

95

 

 

 

SO
, 0.06023Kg\? , 9.8 m l[11.27341 - (2934]°K,(3/8 in.)3

GI ' 3 I 0 TT 1

nl )F 293 K sec

(2.54 cm)3 , l m3
T

1 in.3 (100 cm)3

or = 70.58

therefore Pr - Gr = 50.78

Page 245 in Heat Transfer by J.P. Holman the average free-convection

 

heat transfer coefficients can be found through the following relationship

(9) Nu (Nusselt Number) = C . (Gr - Pr)m = S- - x
For Gr - Pr in the range 2x104 - 8 x 106 for a horizontal plate with
upper surface heated

C = 0.54, m = 1/4

Even though I am not in the range mentioned above, to get a rough estimate
the C and m values will be used. It is important to recognize that
these values are for a isothermal surface, in this case at l,000°C. The
screens are assumed to reach l,000°C almost instantly.

The equation for free convection can be written

- T )A

qfree convection = h (T1 0

from equation (9) the heat transfer coefficient h can be calculated to be

 

h = CK (Gr - Pr)m x = average distance from screen to
x reactor wall
x = 3/8 in.

$0

96

 

 

q o m -

Egg' = C K(Gr Pr) A (T1 TO) A = area of upper side of screen

x
A = 2 in'.2
9F C 2 in 2 0 54 0 275 ° 1/4
. . = . 1 . 1 . MM“ 3000 - 20) C L (50.78) J J/sec 1
top . l l o 1 l is w
3/8 Tn. m C

1 m _1 2.54 cm
T’
100 cm 1 in.

 

qF.C. = 52.6 J/sec

top

In the same manner the heat lost due to free convection can be calculated
for the bottom of the screen. From the same source for the lower surface

of a heated plate with Gr . Pr in the range l05 - 1011

C = 0.58 and m = 1/5

 

 

therefore
q - 2 0 1/5
F.C. _ 2 ln 1 (0.58) L0.275,,+(l,000 - 20) C J[(50.78) 1
- —l l
°°tt°m 3/8 in. m°C
J/secgg 1 m 12.54 cm
1 T
w 100 cm 1 in.
= 46.45 J/sec
Total Energy
lost due to = °F.C. + °F.C. = (52.6 + 46.45) J/sec = 99.05 J/sec

free convection top bottom

0) Calculate the total energy lost to the surroundings -

97

qradiation + qfree convection

qTotal qconduction +

qTotal

 

(26 49 + 222.39 + 99.05) J/sec = 347.93 Jogles

%gculate theo arithmetic average of the resistance over the range
0,0t01000°C-

The resistance for an individual wire is given by

 

R = p L

" 20°C ‘ A
p 0 = resistivity of stainless steel (304) at 20°C
20 C

°20°C = 7.2 x 10'5 ohm - cm

L = length of the wire = 2 inches
A = area of the individual wire = irr2

r = 0.0007 inches

 

thus
-5 ,
Rn = 7.2 x 10 ohm - cm+_2 in. % 1 1n. = 36.828 ohm
ir(0.0007 in) 2.54 cm
The total resistance R2000 is given by
l l l
= -— + —- l _ l
-°20°C R1 R2 + """ l ‘n ‘ " X in

where h = 325 because there are 325 wires in a one inch wide screen.
Solving for R2000 you get

= 39. = 35:828 °hm = 0.113318 ohm

R20°C n 325

 

98

Because the resistance in stainless steel is a function of temperature

the following relationship was obtained from Principles of Physics

 

by F. Bueche (Section Edition, McGraw-Hill Book Company, p. 374).

 

.. 0
o o 3 9 x 10"3
20 C '
°K

Since the temperature changes from 20°C to l,000°C I desire to calculate
a resistance which will give me an average for this temperature range.
1 b
Arithmetic Average = E:—- 7 f(x) dx
a

$0

Arithmetic Average of the resistance = Raa

1 1273°K

R = f (R o + R o . a o -
AA 1moK _ 293oK 2930K 20 C 20 C 20 C

 

(T - 293°K)) dT

thus

(R
- o a .

20°C) (°ZO°C)
980° K

 

[(1273°l<)2 - (293302]
2 2

0.32920241 ohm

99

F) Calculate the I (current) and V (voltage) required to heat the
screens at l,000°C-

H
II

)0.5 )0.5

(Power/Resistance = (qTotal/RAA

H
II

32.51 Amps

V = I x RAA = 10.70 volts

PART 0 Calculate the Gas Plug Dispersion

Because the distance from the liquid nitrogen trap to the G.C.
column is so long (107 inches of 1/8 inch copper tubing) it is feared
that the plug of gas might disperse before reaching the G.C. column.
If this occurred than good separation in the column would not be
achieved.

The plug of gas can be modeled as it flows through the tubing as a

function of both time and distance. At time equals zero, just before

the plug is released from the liquid nitrogen trap, it would look like

 

this. I """" ‘C = 0
I
t=0 I C
Peak Input 1 X

 

 

 

 

At some later time which is unknown at this time I determine that the most

extreme case which can be tolerated for dispersion of the plug is

x = L/2 C = (0.01)CO

100

This plug of gas would graphically look like
...-........:,,.--- C = C

      
   

O

r

 

 

D‘s.-- ---O~”

AK
9'

 

X
ll 10
C)

x = L/2 X = L/2

I am calculating the half width of the peak. The equation which describes

this change is

with Boundary Conditions

1. t50, C = 0, for all x
2. t>0, C = 00/2, x = 0
3. t>0, C = 0, x-+ w

The solution to this equation with the corresponding boundary conditions

is given on pages 353-354 of Transport Phenomena by Bird, Stewart and

 

Lightfoot (copyright 1960).

= (l-erf(-—x—))

/ D t
o 4-AB

filo

once again at x = L/2, C = 0.01 00 the above equation can be written

0.01 = (l - erf (—i/—2— ))

740 t
A8

for 0/00 = 0.01 from Figure 4.1-2 page 127 of Bird, Stewart and Lightfoot

the (erf) can be solved

101

74 DABt

Calculate the diffusion coefficient °AB for the diffusion of 002 (B)

through helium (A) at the following conditions

T = 293°K Pt = 101.3 kw/m2 = 101.3 x 103 w/m2
= - = 4 gm = = 44 gm
MA M.W. of hellum mole MB M.W. of CO2 mole

From Table 2.2 page 33 of Mass Transfer Operations by Robert E.

 

Treybal (third edition) I obtained the following data

 

EA/K = 10.22 EB/K = 195.2
rA = 0.2551 rB = 0.3941
1" + Y‘ E E
= ---———--—A B = ’- ———_AB = ._A i - ‘-
O

51- = 233—5 = 6 5599

EA8 44.665
from Figure 2.5 for %I_. 6.5599 f(EAJLJ = 0.4

A8 A8
= 0.52223

/ T/MA +Tl/MB

For mixtures of nonpolar gases or of a polar with a nonpolar gas

3/2
-4
l0 . (1.084 - 0.249 71mA + l/MB). T , /l/MA + l/MB

 

 

 

D =
AB .1“ 2.
Pt ( AB) f(KT/EAB)

equation (2.37)
Treybal

102

 

 

 

0 = 10'4 l(1.084 - 0.249 (0.5223) (293)3/24 0.52223 % ,(100 2cm)2
AB ' (101.3 x 103) (0.3941)2 0.4 1m
0AB = 0.397 cmz/sec
Solving equation (A) for time

t = x since X = L/2 = 22 inches/2 = 11 inches

16 0AB
(11 in)2 sec (2.54 cm)2
t = F l 2 T‘ 2
16 0.397 cm (11 in)

122.9 seconds

(.4.
ll

50 this is the time it takes to get to

C = 0.01 CD at x = L/2

Calculate the actual time for the plug of gas to travel from the liquid

nitrogen trap to the G.C. column.

area of 1/8 in. copper tubing = 3.318 x 10'3 in.2

length of tubing is 107 inches

3 2

thus volume = (107 in. ) x (3.318 x 10‘ in. ) = 0.355059 in.3

since the flow rate through the tube is 30 cm3/minute

 

Time required to . 3 . . 3
flow through the = 0.355059 ln. +mln.3%60 sec {(2.54 cm)_
tube 30 cm 1 min. in3

11.64 seconds

103

Conclusion:

Since there is such a large difference in time between the
modeled time of 122.9 seconds and the actual time of 11.64 seconds
it is safe to assume that the plug of gas will remain intact up to

the time it reaches the column.