close

Вход

Забыли?

вход по аккаунту

?

The influence of microwave radiation on sorption and the use of frequency response to study diffusion

код для вставкиСкачать
INFORMATION TO USERS
This manuscript h as been reproduced from the microfilm master. UMI films
the text directly from the original or copy submitted. Thus, some thesis and
dissertation copies a re in typewriter face, while others may be from any type of
computer printer.
The quality of th is reproduction is d e p e n d e n t u p o n th e quality of th e
co py subm itted. Broken or indistinct print, colored or poor quality illustrations
and photographs, print bleedthrough, substandard margins, and improper
alignment can adversely affect reproduction.
In the unlikely event that the author did not send UMI a complete manuscript
and there are missing pages, these will be noted.
Also, if unauthorized
copyright material had to be removed, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, charts) are reproduced by
sectioning the original, beginning at the upper left-hand com er and continuing
from left to right in equal sections with small overlaps.
Photographs included in the original manuscript have been reproduced
xerographically in this copy.
Higher quality 6” x 9’ black and white
photographic prints a re available for any photographs or illustrations appearing
in this copy for an additional charge. Contact UMI directly to order.
Bell & Howell Information and Learning
300 North Z eeb Road, Ann Arbor, Ml 48106-1346 USA
800-521-0600
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
THE INFLUENCE OF MICROWAVE RADIATION ON SORPTION AND THE USE
OF FREQUENCY RESPONSE TO STUDY DIFFUSION
A Dissertation Presented
by
MICHAEL D TURNER
Submitted to the Graduate School o f the
University o f Massachusetts Amherst in partial fulfillment
o f the requirements for the degree of
DOCTOR OF PHILOSOPHY
September 2000
Chemical Engineering
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number 9988849
Copyright 2000 by
Turner, Michael David
All rights reserved.
UMI*
UMI Microform 9988849
Copyright 2001 by Bell & Howell Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
Bell & Howell Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
© Copyright by Michael D. Turner 2000
All Rights Reserved
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
THE INFLUENCE OF MICROWAVE RADIATION ON SORPTION AND THE USE
OF FREQUENCY RESPONSE TO STUDY DIFFUSION
A Dissertation Presented
by
MICHAEL D. TURNER
Approved as to style and content by:
illiam Curtis Conner
RoBert L. Laurence, Member
to ) / .V
________
Dionisios G. Vlachos, Member
Member
Michael F. Malone, Department Head
Department of Chemical Engineering
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DEDICATION
To my parents and my wife Cara for all their love and support.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ACKNOWLEDGMENTS
I would like to begin by thanking my advisors, Professors W. C. Conner and R. L.
Laurence. They have not only taught me the knowledge needed for my research career,
but have treated me as a colleague and friend for which I am eternally grateful. Thank
you both.
Professors K. S. Yngvesson and D. G. Vlachos deserve special thanks for serving
as committee members in spite o f their busy schedules. I would like to especially thank
Professor Yngvesson for teaching me more about microwave engineering than could be
garnered from any textbook.
To my parents, it has been your love and support over the last 28 years that has
made me the person that I am today. I look back to the time dad and I spent in the garage
tinkering and all he taught me. To my mom, who has been an inspiration by her
unwavering strength over the past several years. You have taught me perseverance in the
face o f hardship. I love you both, thank you.
I am forever grateful to my caring and loving wife Cara. Without her support and
love, this work could never have been completed. I look back with fondness on those
long nights in the lab when you lent a helping hand and your companionship. I love you
very much.
To my extended family, thank you all for your support over the last five years. I
am especially thankful to my brother and sister for their help, both here and back home.
To those members o f the Conner and Laurence research group, Femao, Chin,
Monaca, and Laurent, I thank you for your help and support over the last five years. To
v
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
those friends in the department that I have had the privilege to meet during my stay in
.Amherst, thank you for everything.
To Paul, Gary, and Joe in the shop, I cannot thank you enough for all you have
done for me. I will miss our lunchtime discussions and your friendship. To Bea, Bobbie,
Chris, Ellen, Margie, Pat, and Sandy in the main office, without your help the piles of
paperwork from the university would be unintelligible, thank you.
Finally, to those friends with whom I have spent many a night at Rafter’s: Brenda
Capobianco, Abby Fuhrman, FCyla Fuhrman, Rich Fuhrman, Dolly Pedevillano, Laurie
Halbert, Carol Oulundsen, George Oulundsen, Sam Sabbagh, Joseph Schroer, Melissa
Schroer, Mike Schultz, Ray Smith, Barbara Tramonte, and Pam Weisenberg, you have all
made my stay in Amherst enjoyable.
vi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
ABSTRACT
THE INFLUENCE OF MICROWAVE RADIATION ON SORPTION AND THE USE
OF FREQUENCY RESPONSE TO STUDY DIFFUSION
SEPTEMBER 2000
MICHAEL D. TURNER, B.S.ChE, UNIVERSITY OF FLORIDA
Ph.D., UNIVERSITY OF MASSACHUSETTS AMHERST
Directed by: Professor W. Curtis Conner Jr.
The influence o f microwave radiation on adsorption selectivity was studied using
the sorption o f cyclohexane and methanol on high-silica zeolites. The amount o f
microwave energy absorbed depends on the specific adsorbent/adsorbate system. The
adsorbent, a high-silica zeolite, is effectively “transparent” to microwave radiation, while
the two adsorbates reflect high (methanol) and low (cyclohexane) absorption of
microwave energy. The measured system temperatures required for desorption by
microwave energy were lower than those required for conventional heating. Further,
microwave radiation can change the sorption selectivity; the adsorbate with the greater
microwave absorptivity is selectively desorbed. The conclusion is that the surface and
adsorbed species can be selectively heated because the rate o f microwave energy
absorption can be greater than the rate o f heat transfer from the surface.
Recent studies suggest that microwave energy can be employed in catalysis and that
the results differ from “conventional” heating. This effort studied the influence o f
microwave energy on automotive exhaust catalysis in the presence and absence o f a
catalyst poison (SO 2 ). The conclusion is that microwave energy can induce catalyst
vii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
lightoff (the temperature where 50% of final conversion is achieved) more efficiently
than conventional heating and can reverse the poisoning by SO 2 for a commercial threeway catalyst.
Zeolites have a variety o f industrial uses, from catalyst supports to membrane
separations. The ability to accurately measure the rate of diffusion in these materials is o f
great importance. A frequency response device has been designed and constructed to
measure diffusion in zeolitic systems. The apparatus was modeled after previous devices
constructed by Yasuda, Rees, Meunier, and Grenier (Yasuda 1976b; Rees and Shen 1993;
Grenier, Bourdin et al. 1995). A detailed description of the apparatus and its
improvements are presented. The device was tested using an n-hexane/silicalite and a
methanol/silicalite system. The model developed by Yasusda (Yasuda 1982) is used to
analyze the frequency response and estimate the diffusion coefficients. This model can
describe a system having multiple diffusivities and a surface resistance to diffusion. The
results for n-hexane/silicalite and methanol/silicalite agree with those o f van den Begin
(Begin and Rees 1989) and Nayak (Nayak and Moffat 1988), respectively.
viii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS......................................................................................................... v
ABSTRACT............................................................................................................................. vii
LIST OF TABLES..................................................................................................................xiii
LIST OF FIGURES.................................................................................................................xiv
NOMENCLATURE...............................................................................................................xvii
CHAPTER
1. GENERAL INTRODUCTION............................................................................................. I
2. MICROWAVE RADIATION S INFLUENCE ON SORPTION AND
COMPETITIVE SORPTION IN ZEOLITES.............................................................3
Background................................................................................................................. 3
2.1
2.1.1
2.1.2
2.1.3
2.2
How does microwave energy interact with matter?........................................ 3
Microwave Equipment....................................................................................... 7
Temperature Measurement.............................................................................. 10
Literature Review.....................................................................................................11
2.2.1
2.2.2
2.3
2.4
Sorption and Microwaves.................................................................................13
Chemical Reaction and Microwaves.............................................................. 16
O bjective...................................................................................................................17
Experimental Section............................................................................................... 19
2.4.1
2.4.2
2.4.3
2.4.3.1
2.4.3.2
2.4 3.3
2.4.3.4
M aterials............................................................................................................19
Apparatus.......................................................................................................... 21
Procedure.......................................................................................................... 24
Sample Preparation....................................................................................... 24
Microwave Heating o f Dry Zeolites........................................................... 25
Single Component........................................................................................ 25
Multi-Component......................................................................................... 26
ix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.5
2.5.1
2.5.2
Microwave Heating o f the Zeolites Themselves......................................... 28
Single Component.............................................................................................32
2.5.2.1
2.5.2.2
2.5.2.3
2.5 .2.4
2.5.3
Adsorption...................................................................................................32
Desorption....................................................................................................34
Microwave Interactions with M olecules..................................................38
Low Temperature Desorption....................................................................41
Competitive Adsorption...................................................................................42
2.5.3.1
2.5.3.2
2.6
28
Results and Discussion
Initial Loading for Competitive Adsorption Experiments...................... 43
Influence o f Microwave Radiation............................................................45
Conclusions................................................................................................................ 51
3 THE EFFECT OF MICROWAVE ENERGY ON THREE-WAY AUTOMOTIVE
CATALYSTS POISONED BY SOz ......................................................................... 55
3.1
3.2
Objectives................................................................................................................... 55
Experimental A spects............................................................................................... 57
3.2.1
3.2.2
3.2.3
Materials.............................................................................................................57
Apparatus............................................................................................................58
Procedure........................................................................................................... 60
3 .2.3 .1
3.2.3 2
3.2.3 3
3.3
Results and Discussion.............................................................................................. 62
3 .3 .1
3.3.2
3.4
Temperature Calibration............................................................................ 60
Sample Preparation..................................................................................... 60
Transient Experiments................................................................................61
Effect of Microwave Energy and S 0 2on CO O xidation...............................62
Effect of S 0 2 on Propane Oxidation...............................................................65
Conclusions................................................................................................................68
4 THE DESIGN AND CONSTRUCTION OF A FREQUENCY RESPONSE
APPARATUS TO INVESTIGATE DIFFSUION IN ZEO LITES......................... 70
4.1
Background................................................................................................................. 70
4.1.1
4.1.2
Elementary Principles o f Diffusion................................................................. 70
Techniques used to measure diffusion coefficients...................................... 71
X
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.2.1 Zero Length Chromatography (ZLC)........................................................ 72
4.1.2.2 Constant Pressure Sorption M ethod............................................................73
4.1.2.3 Constant Volume Sorption Method............................................................73
4.1.2.4 Step Response M ethod................................................................................74
4.1.3
4.1.4
Frequency Response M ethod............................................................................ 74
History o f the Frequency Response M ethod..................................................75
UJ
Is)
Objective....................................................................................................................76
Theoretical................................................................................................................. 77
4.3.1
4.3.2
4.3 .3
4.4
Previous Devices....................................................................................................... 85
4.4.1
4.4.2
4.4.3
4.4.4
4.5
4.5.1.1
4.5.1.2
4.5.1.3
4.5.1.4
4.5.1.5
4.5.1.6
4.5.1.7
4.5.1.8
4.5.2
4.5.3
Apparatus............................................................................................................87
Servomotor......................................................................................................89
Metal Bellows P um p.....................................................................................89
Position Sensor...............................................................................................90
Sorption Cham ber......................................................................................... 90
Pressure Transducers and Ballast V olum e................................................ 91
Dosing M anifold........................................................................................... 92
Vacuum System ............................................................................................ 92
Heating and Insulation.................................................................................. 93
Experimental M aterials....................................................................................93
Procedure........................................................................................................... 94
Results......................................................................................................................... 99
4.6.1
4.6.2
4.7
Yasuda’s D evice................................................................................................85
Rees’ Device...................................................................................................... 85
Grenier’s D evice................................................................................................86
General Improvement........................................................................................86
Experimental Setup....................................................................................................87
4.5.1
4.6
General M odel................................................................................................... 77
Multiple Diffusion Processes............................................................................81
Surface Resistance to Diffusion....................................................................... 83
Blank Experiments............................................................................................99
Sorption Experiments......................................................................................101
Conclusions............................................................................................................... 106
5. CONCLUSIONS.................................................................................................................107
xi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6 RECOMMENDATIONS FOR FUTURE RESEARCH............................................... 110
APPENDICES
A. MICROWAVE EFFICENCY CALCULATIONS....................................................... 112
B PARTS LIST FOR FREQUENCY RESPONSE APPARATUS.................................121
BIBLIOGRAPHY................................................................................................................... 124
xii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
LIST OF TABLES
Table
Page
2-1 - Mode frequencies o f microwave oven cavity(Buffler 1993).......................................9
2-2 - Loading summary for single component adsorption - moles adsorbate per
kilogram of adsorbent..................................................................................................... 33
2-3 - Summary of zeolite loadings after initial binary adsorption for all cases moles adsorbate per kilogram o f adsorbent.................................................................. 44
3-1 - Engelhard three-way catalyst composition as reported in the Material Safety
Data Sheet (MSDS)..........................................................................................................57
3-2 - Reactant gas cylinder concentrations............................................................................58
3-3 - Feed composition and m/e o f both the reactants and products monitored by
the mass spectrometer......................................................................................................61
4-1 - Operating conditions for the n-hexane/silicalite and methanol/silicalite
systems............................................................................................................................ 101
4-2 - Parameters used in the fitting o f the data presented in Figure 4-12 and
Figure 4-13......................................................................................................................105
A-l - Experimental data for the calibration o f the inputmicrowave pow er....................115
A-2 - Physical property data and individual components o f the energy balance
xiii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
118
LIST OF FIGURES
Figure
Page
2-1 - Four basic types o f dielectric polarization.......................................................................5
2-2 - Microwave field orientation and relative strength in the TEio mode o f a
cross section o f a WR284 waveguide. The position o f the zeolite adsorbate
bed takes advantage o f the maximum in microwave electric field parallel to
the short side........................................................................................................................11
2-3 - Real and imaginary parts o f the complex dielectric constant for water at
298K (Courtesy ofZ . Wang)............................................................................................21
2-4 - Inlet manifold for the microwave mediated sorption experiments. The on­
line mass spectrometer and mass flow controllers is each computer
controlled............................................................................................................................ 22
2-5 - Schematic o f the microwave waveguide sorption system. The 2.45 GHz
generator was operated from 0 to 300 Watts of continuous power. The
zeolite adsorbent bed passes through the short axis o f the 7.2cm x 3 .4cm
rectangular waveguide.......................................................................................................24
2-6 - Heating curves at two positions in the zeolite bed and in the gas effluent for
silicalite exposed to microwave radiation - temperature versus time......................... 30
2-7 - Heating curves at two positions in the zeolite bed and in the gas effluent for
DAY exposed to microwave radiation - temperature versus time...............................31
2-8 - Single component adsorption breakthrough curves —Effluent concentration
versus time after introduction o f the adsorbate (methanol or cyclohexane)
into the 30cm3-m in'1 He diluent stream.......................................................................... 34
2-9 - Desorption o f methanol from silicalite by microwave radiation Temperature (above) and effluent concentration (below) versus time with
changes in microwave energy from 0->40—>80—>120—>80—>40—>0 W atts............ 36
2-10 - Summary of the single component desorption experiments performed with
cyclohexane - Temperature and amount of cyclohexane desorbed from
silicalite or DAY zeolites versus microwave power...................................................... 37
2 -1 1 - Summary of the single component desorption experiments performed with
methanol - Temperature and amount of methanol desorbed from silicalite or
DAY zeolites versus microwave power.......................................................................... 38
xiv
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2 -1 2 - Initial adsorption, for Case 1 (cyclohexane adsorbed before methanol is
introduced at 290 minutes), o f the cyclohexane-methanol-silicalite system.
Temperatures in the bed (upstream and downstream) and in the gas effluent
are shown above while the effluent concentration of methanol and
cyclohexane (dotted) are shown below.......................................................................... 45
2-13 - Competitive sorption o f the cyclohexane-methanol-silicalite system
initially loaded using Case 1 (Figure 2-12). Temperatures in the bed
(upstream and downstream) and in the gas effluent are shown above while
the effluent concentration of methanol and cyclohexane (dotted) are shown
below.................................................................................................................................. 47
2 -1 4 - Summary o f the competitive adsorption experiments performed on the
zeolite DAY - ln(Selectivity Ratio: a or P) versus 1/T, see text for
definition.............................................................................................................................49
2 -1 5 - Summary o f the competitive adsorption experiments performed on the
zeolite silicalite - In(Selectivity Ratio: a or P) versus 1/T, see text for
definition............................................................................................................................50
3-1 - Inlet manifold for the three-way catalyst experiments. The on-line mass
spectrometer and mass flow controllers is each computer controlled....................... 59
3-2 - Carbon dioxide (m/e = 44) production versus bulk catalyst temperature for
conventional and microwave experiments....................................................................63
3-3 - Carbon dioxide (m/e = 44) production versus bulk catalyst temperature for
conventional and microwave experiments in the presence and absence of
sulfur dioxide, a catalyst poison.....................................................................................65
3-4 - Propane (m/e = 29) production versus bulk catalyst temperature for
conventional experiments in the presence and absence o f sulfur dioxide, a
catalyst poison.................................................................................................................. 67
4-1 - Sketch of the random walk theory confined by geometric constraints
(Karger and Ruthven 1992)............................................................................................. 72
4-2 - Theoretical in-phase and out-of-phase functions for a slab...................................... 80
4-3 - Theoretical in-phase and out-of-phase functions for a sphere..................................81
4-4 - Theoretical in-phase and out-of-phase functions for a process in which the
local Henry’s Law constants are the same but the characteristic time
constants differ by an order of m agnitude.....................................................................82
XV
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4-5 - Theoretical in-phase and out-of-phase functions with the surface resistance.......... 84
4-6 - Schematic o f the frequency response device, specific information on
manufacturer and model number can be found in Appendix B ...................................89
4-7 - Electron micrograph of the silicalite sample used in these experiments, the
crystal size is 50x10x10 (am (Photo courtesy o f Prof. Tsapatsis’ research
g ro u p )................................................................................................................................94
4-8 - Using an estimation of the diffusion coefficient, the particle size is
determined such that the intersection o f a horizontal line from the y-axis
intersects a vertical line from the x-axis in the region between the 0.005 Hz
and 5 Hz boundaries........................................................................................................ 95
4-9 - [P * ,™ , - Pe\ and [pBlank - P ^ ^ ] versus the mass o f the silicalite for
the n-hexane/silicalite system. The point at which they intersect is the
optimal sample mass.........................................................................................................97
4-10 - Amplitude ratio for the blank, n-hexane, and methanolexperiments.....................100
4-11 - Phase lag for the blank, n-hexane, and methanol experiments...............................101
4 -1 2 - In-phase and out-of-phase data and fit using the two-diffusion-coefficient
model with a surface barrier for methanol/silicalite....................................................102
4 -1 3 - In-phase and out-of-phase data and fit using the two-diffusion-coefficient
model with a surface barrier for n-hexane/silicalite....................................................103
A -1 - Net power, power absorbed by the water load, and corrected net power
versus the power setting for the calibration o f the input microwave power
experiments. The corrected net power and power absorbed by the water load
are in good agreement................................................................................................... 116
A-2 - Radial temperature profile o f silicalite exposed to 100W o f microwave
energy in the one-meter column for several at several different experimental
times................................................................................................................................119
xvi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
NOMENCLATURE
Symbols
A
x dimension of the microwave cavity (m)
a
long dimension of the waveguide [Chapter 2], (m), or radius o f the spherical
particle [Chapter 4], (m)
B
y dimension of the microwave cavity (m), or amount o f adsorbate sorbed in
adsorbent, (mol)
_.
,
(hlV\
U J
Bi
Biot Number —
b
short dimension o f the waveguide, (m)
C
z dimension of the microwave cavity, (m)
Cp
constant pressure heat capacity, ( J - g '- K 1)
c
speed of light in vacuum, (~3.0
d
depth at which 63 .2% o f the incident energy has been absorbed.
dV
volume o f the adsorbing species, (cm3)
D
diffusion coefficient, (m2-s ’)
E
electric field inside the material, (V -m 1)
f
frequency, (Hz)
fc
cutoff frequency, (Hz)
x
10
O
1
m-s' )
AHads isosteric heat o f adsorption, (kJ-m ole1)
h
measured overall heat transfer coefficient, (W-K '-m'2)
K
Henry’s Law constant
k
thermal conductivity, (W-m '- K 1)
xvu
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
L
length o f the plane sheet, (m)
/
the number o f maxima in a mode in the x dimension o f the microwave cavity
m
the number o f maxima in a mode in the y dimension o f the microwave cavity or
mode index for a waveguide
n
the number o f maxima in a mode in the z dimension o f the microwave cavity or
mode index for a waveguide
P
total system pressure, (Torr)
Pv
power absorbed per unit volume, (W-m'3)
p
amplitude ratio of the system pressure
q
loading o f zeolite with adsorbate, ((mole adsorbate)-(kg adsorbent)'1)
R
radius o f zeolite bed, (m), or universal gas constant, (8.314 J-mof'-K*1)
T
temperature, (K)
t
time, (s)
r
total system volume, (m3)
Pads
volume o f adsorbate adsorbed on the zeolite,
I"
amplitude ratio of the system volume
(m3)
Greek Symbols
a
( k '
thermal diffusivity ------\f* ~ p j
P
adsorption selectivity using Clausius-Clapeyron
5
dimensionless variable
e
complex dielectric permittivity
relationship
xviii
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
s'
real part o f the complex permittivity
s"
imaginary part o f the complex permittivity
S.,
permittivity o f free space (8 .85 x 10‘12 F-m'1)
<p
phase lag, (radians)
7
adsorption selectivity using microwaves
n
reduced angular frequency
K
length of the electromagnetic wave in a vacuum
p
density (g-cm'3)
CO
angular frequency, (radians-s1)
xix
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 1
GENERAL INTRODUCTION
The importance o f zeolites and catalysts in today’s industry is certain. For this
reason, the scientific community has gone to great measures to study the mass transport
and kinetic aspects of these materials. Over the past 20 years, the research literature has
suggested that for adsorption and chemical reactions microwave energy has an effect
different from that of conventional heating. However, these articles fail to describe the
basic physics behind the observed differences. The first portion o f this dissertation will
probe these fundamental questions and infer how microwave energy interacts with an
adsorbate/adsorbent system.
The first half of Chapter 2 reviews the essential concepts associated with the
interaction of microwave energy with materials. Included is a short discussion o f
temperature measurement and current microwave generating equipment available. The
second half of the chapter presents some of the existing experimental work on sorption
and catalysis using microwave energy and focuses on the interaction o f microwave
energy with zeolites, single component adsorption, and finally competitive adsorption.
Chapter 3 describes the application of the concepts learned in Chapter 2 to carbon
monoxide oxidation using a state-of-the-art three-way automotive catalyst in the presence
or absence of a poison.
The molecular-scale pore dimensions of zeolites makes diffusion a key factor in
their uses as adsorbents and supports for catalysts. Therefore, an accurate measurement
of diffusion is desirable in the design of a chemical process. Chapter 4 describes the
l
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
design and construction o f a frequency response device that can measure the
characteristic times for diffusion processes over three orders o f magnitude.
Chapter 4 begins with a review o f the fundamental concepts related to the study o f
diffusion. Following this review, the typical methods used to determine diffusion
coefficients in zeolitic materials are described. The last portion o f chapter 4 is a brief
history o f the frequency response technique and the presentation o f the design and
construction of a frequency response device.
Chapter 5 is a summary o f the major conclusions o f this dissertation and chapter 6
presents the recommendations for future work. Appendix A discusses the unsuccessful
attempt to measure the efficiency o f a microwave system. Included in this appendix is a
summary of the work completed, the problems encountered, and the possible solutions to
these problems. Appendix B is a parts list for the frequency response device described in
chapter 4.
2
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 2
MICROWAVE RADIATION’S INFLUENCE ON SORPTION AND COMPETITIVE
SORPTION rN ZEOLITES
2 .1
Background
In a paper submitted at the 28th Microwave Symposium on Quality Enhancements
using Microwaves, Edward R. Peterson asked the following question (Peterson 1993):
“Do microwaves stimulate reactions other than through thermal means?”
His answer was:
“It is hard to tell.”
This statement summarizes the sentiment o f most researchers investigating the effects of
microwave energy on sorption and chemical reactions. In fact, it is this question, and the
search for its answer, in some small part, which is the impetus for this dissertation. The
following sections o f this chapter summarize the interaction o f microwave energy with
matter and with the equations that describe these interactions as they relate to heat
transfer. Subsequently, a brief description is presented about microwave equipment and
the measurement of temperature in a microwave field. Finally, the publications
discussing the application o f microwave energy to sorption and catalytic systems are
reviewed.
2.1.1
How does microwave energy interact with matter?
It is important to have a fundamental understanding o f how a material interacts with
microwave energy. Without this understanding, it would not be possible to formulate an
3
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
explanation o f the interaction o f microwave energy with the systems studied in this
dissertation:
•
The adsorption of methanol and cyclohexane, both individually and
simultaneously, on silicalite and de-aluminated-Y (DAY) zeolites; and
•
The reaction o f carbon monoxide (CO) to form carbon dioxide (CO 2 ) on a
three-way automotive catalyst in the presence and absence o f the catalyst
poison, sulfur dioxide (SO 2 ).
A dielectric substance is material that does not conduct electricity but can sustain an
electric field (Neufeldt and Guralnik 1988). The theory o f an electromagnetic field
interacting with a dielectric was developed in the first half o f the 20th century by Cole and
Debye (Debye 1935; Cole and Cole 1941). They drew the conclusion that there are four
basic types o f dielectric polarization (see Figure 2-1 (Zlotorzynski 1995)):
1. Electronic polarization, by realignment o f electrons around the nuclei.
2. Atomic polarization, by the relative displacement o f nuclei due to the unequal
charge distribution within the molecule.
3. Orientation polarization, by the reorientation o f permanent dipoles by the electric
field.
4. Space charge polarization, by the displacement o f free electrons restricted by
grain boundaries.
4
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
N o Field
® @
©0 ©O©
Q©0 ©©
©0 ©O©
®
®
@ ® ®
@ ®
®
Electric Field A p p lied
/^v
'S
“
/^\
'S
Electronic
Polarization
Atomic
Polarization
/^v
'S ”
'S
™ 'S '
/^N /^N
^S'
'S ' ' S
^S'
0 0 0 0 ©
o©@o©
ooo@ ©
Orientation
Polarization
Space Charge
Polarization
Figure 2-1 - Four basic types of dielectric polarization
The strength of these polarizations are characterized using the material property called
the complex permittivity (e), which is defined as follows:
|e = s '- / s"
Equation 2.1
The real part of the complex permittivity (s') is related to the degree o f polarization o f the
material. The greater e', the more electromagnetic energy is stored in the material when
placed in an electromagnetic field. Each of the four types o f polarization responds with a
different time scale. At very low frequencies, all the polarizations are in-phase with the
electric field. For an electric field oscillating at microwave frequencies (300MHz 300GHz (Pozar 1990)), the electronic and atomic polarizations are much too rapid to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
contribute to dielectric heating. However, the orientation and space-charge polarization
occur on the same time scale as the electric field oscillations at microwave frequencies.
For the orientation polarization, as the electric field switches direction, the permanent
dipole o f the molecule attempts to align itself with the electric field. It is the phase lag
between this type o f polarization and the electric field that leads to absorption o f energy
and Joule heating. The rate o f conversion o f this electrical energy to heat in the material
is proportional to the imaginary part o f the complex permittivity (e"). For frequencies
much greater than microwave frequencies, none o f the polarizations are able to align with
the electric field, and no interactions are observed (Zlotorzynski 1995). It is apparent that
the complex permittivity is a function o f the electromagnetic frequency. Thus, not only
the strength o f the electric field but also its frequency is quite important when
determining the interactions o f microwave energy with dielectric materials.
From the preceding explanation, it is clear why electromagnetic energy in the
microwave region interacts so differently from energy in the infrared region,
conventional heating. A second difference is that microwave radiation is a more
penetrating type of electromagnetic radiation than infrared radiation. Microwave
radiation has a wavelength of about 12cm at 2.45GHz. Equation 2.2 is used to calculate
the penetration depth o f microwave radiation into a sample based on the complex
permittivity(Buffler 1993).
77
= .^ .
itt
1
r (I ( „ \2
{ M £/ s-)
r
^
Equation 2.2
J
6
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Thus, microwave energy interacts with a fraction o f the material volume, whereas
infrared or conventional heating interacts only with the surface of the material.
With an understanding of the interactions o f materials with microwave energy, it
is important to comprehend the ramifications o f these interactions as they pertain to heat
transfer The unsteady-state form o f Fourier’s Law describes infrared heating as follows
(Bird, Stewart et at. 1960):
Equation 2.3
ct
However, due to the large penetration depth o f microwave energy, a source term must be
added to this equation (Buffler 1993).
5T_
Equation 2.4
ct
Equation 2.5
The power absorbed per unit volume contributes directly to the energy balance. This
means that the final temperature of the material is dependent only on the amount o f
energy absorbed and the rate at which energy is lost to the surroundings. This will prove
to be a key argument in the chapters to follow.
2.1.2
Microwave Equipment
A common mistake made by the layperson is to believe that a power setting on a
standard microwave oven actually reflects the applied power. This is incorrect. For
example, a 1000W microwave oven set at 50% power does not generate 500W. It
generates 1000W, but on a 50% duty cycle, the microwaves are on only half the time.
7
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Thus, a standard microwave produces an averaged power and does not emit radiation
continuously. It is important to recognize that the duty cycle differs for different makes
o f microwave ovens. The problem o f average power can be solved, but at a modest
expense. Microwave power sources are available that provide adjustable continuous
microwave power. These are quite expensive, even the low power models. Our research
group purchased a 2.45 GHz, 300W model from Sairem Corporation for approximately
$7000 in 1997. In addition to continuous microwave generators at fixed frequencies,
variable-frequency generators are available commercially. These generators are very
expensive (>$50,000) and beyond the budget for this project.
In addition to the problem o f averaged power, there is the added problem o f a
non-uniform electric field inside the microwave. This non-uniformity is caused by
standing waves inside the oven. A microwave oven is a microwave cavity. The
following equation defines the frequency o f the modes in a cavity related to the
dimensions of the cavity (Buffler 1993):
f ( G H z ) =c
2A
v
2B j
Equation 2.6
V2Cy
Table 2-1 shows the frequencies and /, m, n indices that are possible for frequencies
between 2.4GHz and 2.5GHz for a cavity o f dimension: A = 35.6cm, B = 25.4cm, and C
= 30.5cm. Note that it is impossible for a mode to exist if any two indices equal zero.
8
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1able 2-1 - Mode frequencies o f microwave oven cavity (Buffler 1993)
/
0
0
1
1
2
2
3
3
4
4
5
5
m
0
4
0
4
2
3
1
3
3
1
2
2
*_
n
Mode Freq (GHz)
5
2.461*
1
2.413
2.497
5
I
2.450
4
2.446
3
2.456
4
2.414
0
2.447
1
2.496
2.401
2
0
2.417
2.467
1
mot e cannot exist
It is evident from Table 2-1 that a number of standing waves are present inside a
microwave cavity. Thus, if a small sample were used, it could be possible that the
sample resides in a region where there was either constructive or destructive interference.
Thus, one could never know the actual amount o f power experienced by the sample.
To remedy this problem, a microwave waveguide could be used. A waveguide is
usually a hollow section o f either rectangular or cylindrical metal. Waveguides are
normally used as transmission lines in the transport o f microwave energy. Only the
transverse electric (TE) and transverse magnetic (TM) modes o f an electromagnetic wave
are possible in waveguides. In fact, depending on the dimensions o f the waveguide, there
are cut-off frequencies below which propagation o f either mode is not possible. The cut­
off frequency for a rectangular waveguide is determined by (Pozar 1990):
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
In order for a mode to propagate, the operating frequency must be greater than the cutoff
frequency. Therefore, for a standard S-band (2.45GHz) waveguide whose dimensions are
a = 7.214cm by b = 3.404cm the only mode that can propagate is the TEio mode, whose
cutoff frequency is 2.078GHz. For reasons not discussed here, the TMoi and TMio
modes have a cut-off frequency o f zero, and, therefore, do not exist (Pozar 1990). All
modes greater than the TEio, such as TEn or TMu, have cutoff frequencies greater than
2 .45GHz and will not propagate in this size guide. Therefore, the benefit o f using an Sband waveguide at 2.45GHz is that the system under study is exposed to a well-defined
electric field propagating in only the TEio mode (see Figure 2-2 (Pozar 1990)).
2.1.3
Temperature Measurement
Temperature measurement in the presence of a microwave field is problematic.
The use o f a typical thermocouple is not possible because the microwave energy would
use the metal shielding o f the probe as an antenna, and electrical arcing could occur. In
addition, from experience even if electrical arcing does not occur, the probe will give
very erratic readings in the presence o f an electric field. This is probably due to a build­
up o f electric charge on the thermocouple sheath. The solution to the problem of
temperature measurement in the presence o f a microwave field is a costly one. Fiber
optic temperature probes are available from few sources (Nortec and Luxtron are two).
10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Quartz Sample Tube
t - M icrow ave Electric Field
] - Z eolite Bed
M icrowave
Choke
o
m
M icrowave W aveguide
Drawing not to scale
7.214 cm
Figure 2-2 - Microwave field orientation and relative strength in the TEio mode o f a
cross section o f a WR284 waveguide. The position o f the zeolite adsorbate bed takes
advantage o f the maximum in microwave electric field parallel to the short side.
These probes are made o f a glass fiber coated with Teflon; both materials are essentially
transparent to microwave energy. The probes are accurate to 0.1K, which is adequate for
most temperature measurements. The probes are limited in temperature to about 500K at
which point the Teflon begins to degrade. In 1997, our research group purchased, for
about $8000, four probes, and a probe reader, which can measure the four probes
simultaneously.
2.2
Literature Review
Studies on the use o f microwave energy in sorption and catalytic processes were
initiated in the middle o f the 1980s. More recently, the National Science Foundation
11
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
(NSF) and the Electrical Power Research Institute (EPRI) began holding microwave
reaction workshops in 1993. The purpose o f these and subsequent workshops was to
“define those chemical and physical processes where interaction with microwave energy
has specific advantages compared to conventional heating methods.” (Krieger-Brockett,
Mingos et al. 1993). Five interactions were noted in the 1993 workshop (KriegerBrockett, Mingos et al. 1993):
1. A direct interaction of microwave energy with certain classes o f adsorbing
molecules. The ability o f a molecule or bulk substance to absorb microwave
radiation is quantified by the dielectric property o f the material. ... This direct
absorption can lead to localized introduction o f energy to a specific site or
region from a remote source o f microwaves.
2. Strong coupling between microwaves and the adsorbing material leads to
time-temperature profiles which are quite different from those found in
conventional heating.
3. Localized superheating and rapid bulk temperature increase. These depend on
the rate of energy transfer from the species absorbing the microwaves to the
non-absorbing materials present. If this rate is low, then localized highenergy reactivity sites may be generated. The achievement o f very high
temperatures in localized regions can have particular implications in catalysis
for example.
4. The ability to start, stop, or ramp the energy input rapidly. This can
selectively enhance quenching reactions both in solids and solutions and could
12
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for example, lead to the isolation o f new phases and exotic intermediates in
chemical reactions.
5. Microwaves may be transmitted over long distances with minimal loss.
This portion o f this dissertation probes directly the first three o f these interactions. The
first of these interactions is examined by investigating how strongly microwave energy
interacts with sorbates that have either a high or low dielectric constant. The second
interaction cannot be measured directly, but based on heat transfer arguments, a system
temperature profile is determined. The third interaction is examined by investigating
how microwave energy interacts with the reaction o f carbon monoxide on a three-way
automotive catalyst.
The applications o f microwave energy in the literature are numerous. Microwave
energy has been used in: wood drying, chemical reactions, adsorption, polymers, ceramic
heating, zeolite synthesis, catalyst synthesis, and many others. The next section discusses
the research that deals with adsorption and competitive adsorption, and heterogeneous
reactions.
2.2.1
Sorption and Microwaves
Few studies in the scientific and patent literature discuss the effect o f microwave
energy on the sorption process. However, the claims made in these studies are
encouraging since they show that microwave energy has an effect different from that o f
conventional heating.
U.S. Patent 4,322,394 (1982) (Mezey and Dinovo 1982) describes a process that
“uses microwave energy to affect dielectric heating o f saturated solid non-carbon
13
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
adsorbent to remove the adsorbed materials, which results in a more rapid, efficient, and
safe regeneration than conventional heating.” The process involves the separation o f CO 2
and H2S from natural gas using a 4A molecular sieve. The adsorption step reduces the
total concentration of C 0 2 and H2S from greater than 20% to less than 1%. The
desorption step is carried out under microwave energy at 2.45 GHz and 300 W. The
effluent stream contains a maximum total concentration of both C 0 2 and H2S at a 95%
level in the first minutes of the desorption and then decreases to 0% C 0 2 and 10% H2S.
U.S. Patent 4,421,651(1983) (Burkholder and Fanslow 1983) also describes the
use of microwave energy to desorb a saturated sorbent. The patent summarizes a process
that uses silicalite to remove ethanol from beer and then uses microwave energy to
regenerate the silicalite, thus producing a 60% ethanol stream. The drawings from the
patent show that microwave generators may be applied to the side o f large vessels by the
use of Teflon windows. The Teflon is transparent to microwave energy and also acts as a
seal.
In 1984, a paper by G. Windgasse and L. Dauerman, “Microwave Treatment of
Hazardous Waste: Removal o f Volatile and Semi-volatile Organic Contaminants from
Soil” (Windgasse and Dauerman 1992), a process is described in which steam distillation
is affected by microwave energy. The microwave energy at a frequency o f 2.45 GHz and
a maximum power of 616 W is used to treat sand, humus soil, and an industrial sample,
all of which were contaminated with jet fuel. In this process, the microwave energy
penetrates the soil and heats the water throughout the soil matrix. The developing steam
causes the volatile organic compounds to be stripped from the soil without
14
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
decomposition. This process demonstrates an alternative energy source to effect the
common unit operation of separations.
U.S. Patent 5,509,956 (1996) (Opperman and Arsenault 1996) describes a process
and apparatus where a polymeric adsorption media, saturated with a volatile organic
compound (VOC), is contained in a cylindrical vessel. A waveguide, with a microwave
source attached, extends along the longitudinal axis o f the cylindrical vessel. The
microwave energy is applied and the VOC is desorbed and drawn out of the vessel.
U.S. Patent 5,282,886 (1996) (Kobayashi, Mizuno et al. 1994) and a recent paper
“Control of Adsorption by Microwave Irradiation” (Kobayashi, Kim et al. 1996) describe
a process in which a mixed gas is contacted with a non-carbonaceous adsorbent while
irradiating the sorbent with microwave energy. The gas component with the smaller
dielectric coefficient is selectively adsorbed by the adsorbent, and the gas with the larger
dielectric coefficient is desorbed. The system that the patent uses is that of the waterCFC113-NaY in the presence o f 2.45 GHz microwave energy. A mixture o f 1,000 ppm
CFC113 and 8,000 ppm water in a 400 cm3-min'1 helium stream is passed over a 0.6 g
sample of NaY until equilibrium was achieved. The microwave energy is then turned on
and the water, which is the component with the higher dielectric coefficient, is desorbed
and the CFC113 is adsorbed. Conversely, when the microwaves are turned off, the
CFC113 is replaced by the re-adsorbing water. This example demonstrates that the
selectivity for CFC113 over water is greatly enhanced in the presence o f microwave
energy.
All the research presented above show clearly that there is a difference between a
system exposed to conventional heating and microwave energy. For the most part, these
15
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
studies give a general description o f the results but offer little insight into specific
microwave interactions occurring in the system. In the chapters that follow, an
explanation is presented o f how microwave energy interacts with an adsorbate/adsorbent
system to produce results different than those expected of the same system exposed to
conventional heating. In addition, these articles lack the quantitative determination o f the
effect that microwave energy has on the system. This dissertation will make this
quantitative determination by measuring the amount of energy needed to affect a change
in the sorption process (i.e., the efficiency o f the process).
2.2.2
Chemical Reaction and Microwaves
Most research on the effect o f microwave energy on chemical reactions has
studied liquid phase organic synthesis reactions. Although interesting, this research is not
relevant to this dissertation. A few reports discuss the effect o f microwave energy in
heterogeneous reactions. Two are selected for discussion below.
Cha has developed a process for the reduction o f NOx using char irradiated by
microwaves (Cha and Kong 1995). The technology employs a two-step process where
NOx is adsorbed onto a carbon material and then the NOx-laden carbon is irradiated with
microwaves producing N 2 and CO 2 that can be vented. However, if the char bed is
heated by conventional methods, the NOx is evolved as NO2 and NO, which does not
solve the problem o f NOx removal. Thus, microwave energy shows a distinct advantage
over conventional heat for this reaction.
Roussy et al. investigated the controlled oxidation of methane on doped catalysts
irradiated by microwaves (Roussy, Thiebaut et al. 1993). The result is an increase in
selectivity from 30% to 90%, depending on whether microwave or conventional energy
16
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
was used. The authors suggest that the enhanced selectivity for microwave energy is
based on selective excitation o f weakly-held oxygen atom sites by the electric field on
which the initial reaction step occurs.
These two examples show that microwave heating can produce quite different
products in heterogeneous reactions as compared to conventional heating. In order to
probe this notion further, the effect o f microwave energy on a three-way automotive
catalyst in the presence and absence o f a poison was studied. The results were compared
to the same system using only conventional heating.
2.3
Objective
Zeolites and other molecular sieves have the potential to remove volatile organic
compounds (VOCs) from process effluents to levels below parts per million. For
molecules o f the size o f the intracrystalline pores, the heats o f physical adsorption often
exceed 40kJ-mole'‘. However, in order for the VOCs to be removed from the adsorbent,
at least this amount o f energy must be supplied. Heating is normally affected by
conventional means such as electrical heaters or steam stripping. The entire system (the
vessel, the adsorbent, and the adsorbate) must be raised to the desorption temperature.
Intriguingly, bulk zeolites are essentially transparent to microwave radiation. On
the other hand, polar VOCs and the oxide surfaces readily absorb microwave energy.
Thus, desorption might be effected without the necessity o f heating the entire system to
the desorption temperature (Mezey and Dinovo 1982; Roussy, Zoulalian et al. 1984;
Stuerga and Gaillard 1996).
Many studies have investigated the effects o f microwave radiation on catalytic
reactions (Krieger-Brockett, Mingos et al. 1993; Cha and Kong 1995; Bond and Moyes
17
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1997). Relatively few studies in the scientific and patent literature discuss the effect of
microwave radiation on the processes o f physical sorption from the gas phase onto
zeolitic solids (Kobayashi, Mizuno et al. 1994; Kobayashi, Kim et al. 1996). These few
studies are interesting because they show that effects of microwave radiation and
conventional heating are different. However, these studies afford no explanation o f the
physical basis of the experimental results.
Since microwave radiation has such complex interactions with materials, three
approaches were used. The results are combined to yield a more complete picture o f how
microwave radiation interacts with our systems. The first study explores the interaction
of microwave radiation with the zeolite itself. Although it is often generalized that
zeolites are completely transparent to microwave radiation, it has been found that the
zeolite may play, in fact, an important role in the desorption process. With a baseline
established for the interaction o f the microwave radiation with the support, the second
study examines how a single sorbate adsorbed on the zeolite interacts with microwave
radiation. How does the dielectric permittivity o f the adsorbate affect the
adsorbate/adsorbent interaction with the microwave radiation? Finally, the third study
assesses the ability o f microwave radiation to influence the selectivity o f the sorption
process. The microwave selectivity results are compared to the conventional heating
method by means of the heats o f adsorption using the Clausius-CIapeyron equation (Vads
vs. Temperature).
18
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.4
2.4.1
Experimental Section
Materials
The experiments utilized two zeolites. The first was a hydrophobic silicalite
(Si/Al > 100) sample from United Catalysts with a “nominal” elliptical pore dimension of
5.3A x 5.6A for the straight channels, and 5. IA x 5.5A for the sinusoidal channels (Meier,
Olson et al. 1996; Cook and Conner 1999), and a packed bed density o f 0.49g/cm3.
These were shipped in the form o f 1/16th inch (0.16cm) extrudates. The second was a
hydrophobic de-aluminated Y zeolite (DAY) sample from Degussa (Si/Al > 100) with a
circular pore dimension o f 7.4A (Meier, Olson et al. 1992), a packed bed density o f
0.35g/cm3, and a pore volume o f 0.3cm3/g (Degussa 1992). These were shipped in the
form of 1/4 inch (0.64cm) Raschig rings. These rings were crushed and screened to a
Tyler Mesh Size o f 10, which is equivalent to the 1/16 inch extrudates.
The dielectric properties of the zeolites were measured using Hewlett-Packard
8753 and Hewlett-Packard 8510 automated network analyzers in conjunction with a
Hewlett-Packard dielectric probe. The real part o f the dielectric permittivity was
measured to be 1.6 for silicalite and 1.7 for DAY zeolite. Microwave absorption in a
material is proportional to the value o f the imaginary part o f the dielectric permittivitty,
s", as discussed in greater detail later. The dielectric probe method was not sufficiently
sensitive to measure e" for pure samples o f either zeolite; however, an upper limit o f 0.05
was estimated. The other experiments reported in this chapter clearly establish that the
microwave loss o f pure zeolites is appreciable, but not so large as to be inconsistent with
the above upper limit for e". When a component such as methanol was adsorbed in
silicalite, e" was easily measurable (-10 at 2.45 GHz). However, cyclohexane adsorbed
19
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
on silicalite showed the same dielectric permittivitty (e') as for pure silicalite, and e" was
not measurable.
The cyclohexane and methanol adsorbates used were obtained from Fisher
Chemical Company with purity greater than 99%. Methanol has a kinetic diameter o f
3.85A (Harrison, Leach et al. 1984), and a dielectric constant o f 33.62 as a liquid at 20°C
(Lide 1991). Cyclohexane has a kinetic diameter o f -6 .3 A (Harrison, Leach et al. 1984),
and a dielectric constant o f 2.023 as a liquid at 20°C (Lide 1991). It should be
remembered that the dielectric constant is the real part o f the permittivity (s') and
indicates the ability o f the material to be polarized by an electric field, for example, a
microwave field. The imaginary part o f the dielectric permittivity (e") is directly related
to the real part and has a maximum equal to e72 for a particular frequency in the Debye
model. Data on e" are scarce, but the proportionality o f e" (max) to e' makes e' an
approximate guide to the strength of interaction o f a material with microwaves Figure 23). Microwave absorption in an adsorbate/adsorbent system is more complex, as there
are several forms o f energy absorption and energy transfer as will be discussed below.
20
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
80
70
60
50
-
-
-
-
35
-
30
-
25
-
20
-
15
-
10
-
5
40
0
2
4
6
10
8
12
14
16
18
Frequency (GHz)
Figure 2-3 - Real and imaginary parts o f the complex dielectric constant for water at
298K (Courtesy o f Z. Wang).
2.4.2
Apparatus
The experiments were executed with an apparatus consisting of an inlet manifold
connected to a reactor designed for exposure to microwave radiation. Either one or two
adsorbates can be introduced into the system through the inlet manifold (Figure 2-4), by
means o f bubble saturators. The concentration o f each adsorbate in the gas stream can be
adjusted by varying the temperature o f the saturator and by adding a diluent stream.
21
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
D iluent
MFC 1
Spectrometer
Helium
MFC 2
>4
Legend
MFC - Mass F low
Controller
®
M ethanol
Saturator
MFC 3
C yclohexane
Saturator
- Valve
Figure 2-4 - Inlet manifold for the microwave mediated sorption experiments. The on­
line mass spectrometer and mass flow controllers is each computer controlled.
A Nortech NoEMI-TS Series fiber-optic temperature measuring system and an
on-line Balzers QMG112A Quadrupole Mass Spectrometer were connected to the
apparatus. The fiber-optic temperature probes were used to measure the bulk zeolite
temperature at the entrance and exit o f the zeolite bed. In addition, a probe was placed in
the effluent stream to measure the temperature of the gas leaving the zeolite bed. The
mass spectrometer was used to measure the concentration o f the sorbates in the effluent.
The microwave reactor was designed and built for these experiments. The reactor
system consists o f a Sairem Model GMP 03 K/SM, 300W maximum, continuous variable
power microwave source operating at 2.45GHz, a Lectronic Corporation coaxial-toWR284 wave guide adapter, a modified section of a Lectronic Corporation WR284 wave
22
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
guide, and a Raytheon Corporation lkW water dummy load. The reactor system is
shown in Figure 2-5. The microwave source measures both the forward and reflected
power through its coaxial output cable. The calibration o f the ratio o f the reflected power
to the forward power was checked with an Hewlett Packard Model 8753 automatic
network analyzer. The main body o f the reactor is the length of modified rectangular
WR284 waveguide. The internal dimensions o f the waveguide are 7.214cm x 3.404cm
(Pozar 1990). At 2.45GHz, the electric field inside the waveguide is in the TEio mode.
This mode has uniform field strength along the short axis o f the waveguide, as seen in
Figure 2-2 (Pozar 1990). As a result of this field alignment, ports were drilled through
the long sidewalls o f the waveguide to admit a quartz sample tube, 1 80cm in diameter,
through the microwave waveguide. This modification allows the entire sample to be
exposed to a uniform amount o f microwave radiation. Microwave chokes (1.91 cm in
diameter and 5.08cm in length) were installed over the ports in order to prevent the
operator from being exposed to any spurious microwave radiation emitted from the ports.
For a given diameter port, the length of the chokes are calculated from standard equations
(Pozar 1990). The water load was used to remove any radiation transmitted through the
waveguide. In addition, the water load may be used to perform an energy balance on the
system. The water load is a sink, which minimizes reflection of the microwave radiation
inside the reactor. The power supply never showed more than 3W o f power reflected to
the microwave source.
23
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Fiber Optic
Temperature
Probes
Wav
Guide
M icrowave
Chokes
Inlet
Gas
M anifold
C ooling
Water In
M icrow ave
Water
Load
Reactor
C ooling
Water Out
Online
M ass
Spectrometer
300 W, 2.45G H z
M icrow ave
Generator
Figure 2-5 - Schematic of the microwave waveguide sorption system. The 2.45 GHz
generator was operated from 0 to 300 Watts o f continuous power. The zeolite adsorbent
bed passes through the short axis o f the 7.2cm x 3.4cm rectangular waveguide.
2.4.3
2.4.3.1
Procedure
Sample Preparation
The zeolites used were calcined at 773 K for 24h under a flow o f dry air. This
removes any pre-adsorbed moisture, hydrocarbons, and templates. In order to avoid
thermal shock to the zeolites, heating rates were never larger than IOK-min'1. The
zeolites were subsequently heated to 423K in a microwave field under a dry helium
stream for 2h to remove residual moisture prior to each experiment.
24
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The length of the bed was constrained to 3.404cm, the length o f the short axis side
of the waveguide; thus, the bed volume was 8.65cm3. The bed contained either 4.26g of
silicalite or 3 .01 g of DAY. The total flow rate through the bed for all experiments was
30cm3-min'‘ of ultra-high-purity helium.
2.4.32
Microwave Heating of Dry Zeolites
At the start of the experiment, a 30cm3-min'’ flow of ultra-high-purity helium was
introduced into the system. Because the temperature of the laboratory changes from day
to day, the temperature of the zeolite at the current room temperature was used as the
baseline for any temperature increases. With the flow rate unchanged, the microwave
power was then increased from 0W to 40W and the temperature history recorded.
Similarly, the power was increased from 40W to 80W and then from 80W to 120W with
the temperature history recorded at each new power level. The procedure was then
reversed and the temperatures were recorded at each power level as the power was
dropped from 120W to 80W then from 80W to 40W and finally from 40W back to 0W.
The microwave power levels described here are the forward power levels as indicated by
the microwave source. Reflected power never exceeded 1% o f the forward power.
2.4.3 3 Single Component
For the single component adsorption, the adsorbate entered the system through a
saturator in the inlet manifold. The concentration o f the adsorbate in the gas phase was
determined by adjusting the temperature of the saturator and by adding a diluent stream.
The concentration o f the effluent from the reactor was measured by an on-line mass
spectrometer. Fiber-optic temperature probes were placed inside the bed at locations near
25
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the entrance and exit. Also, a probe was placed immediately after the end o f the bed and
suspended away from the tube wall in order to measure the temperature o f the gas leaving
the bed.
At time zero, a stream o f I5cm3-min'1 o f helium saturated at 293K with one
adsorbate was mixed with a diluent stream o f 15cm3-min‘l of helium. This gas stream
was introduced into the system through the inlet manifold. Initially, the bed contained no
adsorbates. Adsorption commenced with the introduction o f the saturated helium stream.
The system was determined to be at steady state when the concentration o f the effluent
equaled the inlet concentration. Desorption was induced by means o f microwave
radiation. With no changes made to the inlet manifold flow rates, the microwave power
source was raised to 40W o f continuous microwave power. The mass spectrometer
measured the changes in the effluent concentration, and the fiber-optic probes measured
the system temperatures. A new steady state having been established, the power level
was increased by 40W to 80W, and the responses were measured. Again, the power level
was increased by 40W to 120W, and the responses were measured. Once steady state
was established at the highest power setting, the power was decreased in the reverse
sequence in which it was raised, waiting for steady state at each power level.
2.4.3.4
Multi-Component
The competitive adsorption o f two sorbing species was studied on each zeolite.
Since single file diffusion is known to occur in zeolites, the order in which the sorption
process occurs may influence desorption. There are three possible permutations for the
binary adsorption process. Two o f the three permutations are sequential adsorption o f the
adsorbates on the zeolite. The third is the simultaneous adsorption o f both components.
26
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Case 1). Cyclohexane adsorption steady state is achieved. Methanol is then introduced
and the system is allowed to reach a new steady state with both adsorbates
present.
Case 2). Methanol adsorption steady state is achieved. Cyclohexane is then introduced
and the system is allowed to reach a new steady state with both adsorbates
present.
Case 3). Both methanol and cyclohexane are introduced simultaneously over the zeolite
and the system is allowed to reach a new steady state with both adsorbates
present.
For Case 1 and Case 2: At time zero, a stream o f 15cm3-min'' o f helium
saturated at 293K with one adsorbate was mixed with a diluent stream o f 15cm3-min‘' o f
helium. This gas stream was introduced into the system through the inlet manifold.
Initially, the bed contained no adsorbates. Adsorption commenced with the introduction
o f the saturated helium stream. Steady state was established when the concentration o f
the effluent equaled the inlet concentration. Once steady state was achieved, the diluent
stream was diverted through a second saturator containing the second adsorbate. With
the first adsorbate stream unchanged, the second sorbate was adsorbed on the zeolite until
a new steady state was established which might result in some desorption o f the first
adsorbate.
For Case 3: At time zero, a stream o f lScnr’-min'1 of helium saturated at 293K
with cyclohexane was mixed with a stream o f 15cm3-min'1 of helium saturated at 293 K
with methanol. This gas stream was introduced into the system through the inlet
27
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
manifold. Initially, the bed contained no adsorbates. Adsorption commenced with the
introduction o f the saturated helium stream. Steady state was established when the
concentrations o f the effluent equaled the inlet concentrations.
With steady state established in all cases, the competitive adsorption proceeded
using microwave radiation. With no changes made to the inlet manifold flow rates, the
microwave power source was raised to 40W of microwave power. The mass
spectrometer measured the changes in the effluent concentrations, and the fiber-optic
probes measured the system temperatures. The power level was increased by 40W to
80W, and the responses were measured. Again, the power level was increased by 40W to
120W, and the responses were measured. Once steady state was established at the
highest power setting, the power was decreased in the reverse sequence by which it was
raised, back to zero, waiting for steady state at each power level.
2.5
2.5 .1
Results and Discussion
Microwave Heating of the Zeolites Themselves
The zeolites used in these experiments are comprised o f silica. Pure silica has a
dielectric loss approaching zero (e"~lx KT* to I x 10'3) and does not heat appreciably
when exposed to microwave radiation. However, zeolites are high surface area materials,
and because silica bonds end in silanols at a surface or defect, the concentration o f
silanols in the zeolite may be large. Values in the literature range from about 5 OH-nm'2
to 16 OH-nm'2 (Guermeur and Jacolin 1994). The larger value is calculated as the total
amount o f OH, including the internal OH, divided by the BET surface area measured by
nitrogen adsorption.
28
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Silanols differ from bulk silica in that they do have a significant dielectric loss.
Thus, when the silanols are exposed to microwave radiation, the OH groups can become
rotationally excited producing heat, thus leading to an overall temperature increase
(Guermeur and Jacolin 1994). There are two principal mechanisms by which the heat
may be transferred, from the silanol-surface to the solid, and from the surface to gas
phase. The rates o f heat transfer by each o f these mechanisms are not the same. The Biot
number was estimated as 0.01, confirming that the transfer to the solid is much more
rapid than heat transfer to the gas. The amount of heat transferred depends on the masses
and heat capacities o f the different phases. In this system, the mass o f silanol-surface is
much lower than either the gas or bulk phases present.
Figures 2-6 and 2-7 show the heating o f silicalite and DAY, respectively, in the
presence of microwave radiation. The heating takes place in 40W power increments
from 0W to 120W under a 30cm3-m inI flow o f dry helium. The zeolites were pretreated
to remove any adsorbed species. Consequently, the increase in temperature of the zeolite
may be attributed to the interactions of the microwaves with the silanol groups and the
surface. This provides a basis for comparison with those cases when microwave
radiation interacts with adsorbed species.
The average temperature increases above ambient for the silicalite sample were
9K, 2 IK, and 32K at 40W, 80W, and 120W respectively, whereas the temperature
increases for the DAY sample were only 5K, 12K, and 17K at the same power levels.
These differences may be attributed to two factors. Firstly, the densities o f the zeolites
differ, 0.49g-cm'3 for silicalite and 0.35g-cm'3 for DAY; thus, there is more silicalite per
29
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
unit volume with which the microwaves interact. Secondly, there may be differing
hydroxyl content in each o f the samples. However, this quantity was not measured.
Measured in these experiments are the temperature o f the bulk zeolites at the
beginning and end of the bed, and o f the gas phase passing over the sample. The
instantaneous temperature o f the silanol-surface is not measured. The obvious
temperature difference across the bed is due to convection. There is also a temperature
difference between the bed and the gas passing over the bed.
330
320
Downstream Bed Temperature
Upstream Bed Temperature
Effluent Gas Temperature
310
300
290
280
120
60
180
Time (min)
Figure 2-6 - Heating curves at two positions in the zeolite bed and in the gas effluent for
silicalite exposed to microwave radiation —temperature versus time.
30
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
310
305
Downstream Bed Temperature
Upstream Bed Temperature
Effluent Gas Temperature
300
295
2 90
285
120
60
180
Time (min)
Figure 2-7 - Heating curves at two positions in the zeolite bed and in the gas effluent for
DAY exposed to microwave radiation - temperature versus time.
At 40 W, 80W and 120W o f applied microwave power, the temperature increases between
the end of the bed and the gas were 6.9K, 15.3K, 23.4K for silicalite (Figure 2-6) and
2.9K, 6.8K, 10. IK for DAY (Figure 2-7). This suggests that heat transfer is not
sufficiently rapid to achieve a thermal steady state in these experiments. The rate by
which microwave energy is selectively absorbed by the silanols is greater than the rate by
which heat is transferred to the bulk phase. The zeolite bed has a measured overall heat
transfer coefficient o f
1.5 W -K ''-m * 2.
The conclusion is that
T siianois
(not measured) >
Tbuik > Tgas phases. The energy absorbed is localized on the surface and transfer is too
slow to reach thermal equilibrium.
31
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
If adsorbed species are present, they can modify the amount o f energy absorbed
because the adsorbates are concentrated on the surface. In addition, the adsorbed species
can have dielectric properties, which differ substantially from the other phases present.
Further, adsorbed species can modify the dielectric properties at the surface where they
adsorb (Guermeur and Jacolin 1994). These interactions were studied individually and
then in concert in the experiments described below.
2.5.2
Single Component
Four single-component desorption experiments in the presence o f microwave
radiation were executed to investigate the following:
1. How microwave radiation interacts with the adsorbate;
2. How microwave radiation can affect the desorption of the adsorbate.
The experiments were performed with DAY and silicalite zeolites using a
component with a low dielectric loss, cyclohexane, and a component with a high
dielectric loss, methanol. A detailed explanation and comparison o f each o f the four
possible combinations is presented below.
2.5.2.1
Adsorption
Each of the four experimental breakthrough curves is shown in Figure 2-8. The
loadings of each o f the four systems, shown in Table 2-2, were obtained by integration of
the breakthrough curves. The loadings are given as moles o f adsorbate per gram of
adsorbent. This data shows that the loading o f the DAY zeolite is two to three times that
of the silicalite for either adsorbate.
32
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 2-2 - Loading summary for single component adsorption - moles adsorbate per
kilogram o f adsorbent
System
Loading (moles adsorbate
/ kg adsorbent)
Cyclohexane/DAY
Cyclohexane/Silicalite
Methanol/DAY
Methanol/Silicalite
2.10
0.95
4.50
1.70
In addition to loading, a qualitative assessment o f the diffusion o f each adsorbate
in the different adsorbents may be made based on the characteristics o f the breakthrough
curve. An estimate of the rate o f diffusion is based on the elapsed time from the initial
breakthrough to the time at which adsorption equilibrium is achieved. The smaller this
time, the more rapid the diffusion o f the adsorbate in the adsorbent. Figure 2-8 compares
the breakthrough curves o f the cyclohexane/silicalite system with the cyclohexane/DAY
system and the methanol/silicalite system with the methanol/DAY system. It is apparent
that both cyclohexane and methanol diffuse more rapidly in DAY than in silicalite. This
is expected since the pore diameter o f DAY, 7.4A, is larger than that of the silicalite,
5.3A x 5.6A straight channels and 5.1A x 5.5A sinusoidal channels (Meier, Olson et al.
1992). In addition, methanol diffuses much more rapidly than cyclohexane in both
zeolites due to the smaller kinetic diameter o f methanol, 3.85 A, as compared to the
kinetic diameter of cyclohexane, 6.3 A (Harrison, Leach et al. 1984).
33
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0
60
120
180
240
300
Time (min)
Methanol Adsorbed on DAY
Cyclohexane Adsorbed on DAY
—■— Cyclohexane Adsorbed on Silicalite
—•— Methanol Adsorbed on Silicalite
Figure 2-8 - Single component adsorption breakthrough curves - Effluent concentration
versus time after introduction o f the adsorbate (methanol or cyclohexane) into the 30cm3min'1 He diluent stream.
2.5.2.2
Desorption
Once an adsorption steady state had been established, the single component
desorption experiments using microwave radiation were performed. Figure 2-9
represents the raw data collected for desorption of the methanol-silicalite system. With
no changes made to the inlet manifold flow rates, the microwave power source was
increased to 40W o f microwave power at 180 minutes. The mass spectrometer measured
34
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the changes in the effluent concentration, and the fiber-optic probes measured the system
temperatures. Desorption occurred rapidly after the introduction o f microwave radiation.
This may be observed in Figure 2-9 as the effluent concentration o f methanol and the
bulk zeolite temperature increased simultaneously. The amount o f methanol desorbed
from the zeolite was calculated by integrating the area enclosed by the effluent
concentration and the baseline value o f methanol at the start o f the experiment. Once
steady state had been established at 230 minutes, the power level was increased by 40W
to 80W and the responses were measured. Again, the power level was increased by 40W
to 120W and the responses measured. Once steady state was established at the highest
power setting, the power was decreased in the reverse sequence, from 120W to zero in
40W increments, waiting for steady state at each power level.
The strong interaction o f methanol with the microwave radiation can be seen in
the temperature history o f the zeolite. At 230 minutes, the microwave power is increased
from 40W to 80W. As the temperature o f the zeolite bulk increases, it appears to
“overshoot” its equilibrium value at that power level. However, this “overshoot” is
related to the concentration history o f the adsorbed methanol, in that, as more methanol is
desorbed from the zeolite, the temperature decreases. Thus, as the amount o f methanol
decreases, the microwave radiation interactions decrease and less heat is absorbed. This
leads to a decrease in the bulk zeolite temperature. However, this is not observed in the
cyclohexane systems. Thus, the conclusion is that cyclohexane adsorbed on the zeolite
has weak interactions with the microwave radiation. Also, the net desorption ceases as a
steady state is reached. Therefore, the temperature is established as a result o f a balance
between the microwave power absorbed and the total heat loss.
35
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
370
350
----
330
310
0.008
290
0.006 m
0.004
0.002
0.000
180
240
300
360
420
480
540
Time (min)
- • -
Downstream Bed T em perature
Upstream Bed T em perature
Effluent Methanol C oncentration
Figure 2-9 - Desorption o f methanol from silicalite by microwave radiation Temperature (above) and effluent concentration (below) versus time with changes in
microwave energy from 0-> 40-> 80-> l20—>80—>40—>0 Watts.
Figures 2-10 and 2-11 are a graphical summary o f all the raw data gathered in the
four single-component adsorption experiments. Figure 2-10 summarizes the desorption
of cyclohexane from silicalite and DAY zeolites, and Figure 2-11 summarizes the
desorption o f methanol from silicalite and DAY. These figures show the temperature of
the zeolite (shown in bar graph format) and the amount o f adsorbate desorbed from the
zeolite (shown in linear format) as a function of the applied microwave power. Figures
2-10 and 2-11 show both the temperature of the zeolite loaded with adsorbate and zeolite
without adsorbate. This information was used to ascertain what portion o f the microwave
36
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
heating may be attributed to microwave interactions with the adsorbate and what portion
attributed to the adsorbent. The microwave interactions with the adsorbed species are
described below.
0.0003 2.
0.0000
a3
0
40
80
120
80
40
0
Applied Microwave Power (W)
•••! Average Bed Temperature of DAY Blank
i/w a Average Bed Temperature of DAY with Cyclohexane
Average Bed Temperature of Silicalite Blank
i
i Average Bed Temperature of Silicalite with Cyclohexane
—a— Total Moles of Cyclohexane D esorbed / gram DAY
—• — Total Moles of Cyclohexane D esorbed / gram Silicalite
Figure 2-10 - Summary o f the single component desorption experiments performed with
cyclohexane - Temperature and amount of cyclohexane desorbed from silicalite or DAY
zeolites versus microwave power.
37
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0.0025
0.0020
to ^
<U
E1
0.0015
g
0.0010
a-
0.0005
£ I
« <
- 0.0000
co « 60
O) o
Q) O
2 < 40
g
- -0.0005
to
20
in
0
40
80
120
80
40
0
Applied Microwave Power (W)
Average Bed Temperature of DAY Blank
\ / / / i Average Bed Temperature of DAY with Methanol
iH M Average Bed Temperature of Silicalite Blank
i
I Average Bed Temperature of Silicalite with Methanol
- Total Moles of Methanol Desorbed / gram DAY
—• — Total Moles of Methanol Desorbed / gram Silicalite
Figure 2 -1 1 - Summary o f the single component desorption experiments performed with
methanol - Temperature and amount of methanol desorbed from silicalite or DAY
zeolites versus microwave power.
2.5.2.3
Microwave Interactions with Molecules
If a polar molecule is exposed to an alternating electric field, a torque is applied to
the molecule. This torque causes the molecule to rotate in an attempt to align its dipole
moment in the direction o f the electric field. However, molecules cannot always orient
themselves completely at microwave frequencies since the electric field changes
orientation approximately every 5x1 O'10 seconds. This phase lag between the dipole
38
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
moment of the molecule and the electric field produces a molecule’s dielectric loss. The
dielectric loss is a measure o f a molecule’s ability to convert microwave radiation into
heat. Since each molecule present interacts with the microwave radiation, it converts a
portion of the radiation to heat. Thus, the interaction of microwave radiation with the
adsorbed species is dependent on the magnitude o f the dielectric loss and the amount of
the species adsorbed (Buffler 1993; Zlotorzynski 1995).
A quantitative formulation o f the above concepts introduces real and imaginary
parts of the complex (relative) dielectric permittivity:
je = s '- / s"
Equation 2-1
The out-of-phase component, e", gives rise to the absorption o f microwave power, and
the absorbed power per unit volume material is as follows:
Pv = 2;r f s ns E~
Equation 2-5
These measurements were performed at a constant frequency o f 2.45 GHz. The absorbed
power is thus proportional to three quantities: e", the volume o f the adsorbing species,
dV, and E2. The latter is, in turn, proportional to the microwave power density in the
sample, which was controlled by varying the output power o f the microwave source.
Of the various media investigated, the zeolite had the largest volume, and a
relatively low value of e", as shown by our dielectric probe measurements. The adsorbed
cyclohexane had a small volume, and also a very small e". The adsorbed methanol had a
small volume, but a large value for e". The conclusions drawn from our single­
component desorption experiments are consistent with the relative magnitudes o f these
39
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
quantities. This should be kept in mind as the results o f the measurements related to
competitive sorption are discussed.
Figure 2-10 shows that virtually all o f the heating came from the microwave’s
interactions with the adsorbent when cyclohexane, a material with a low dielectric loss,
was adsorbed onto either the silicalite or the DAY zeolite. The adsorbent temperature
was the same, within experimental error, whether the cyclohexane was or was not
adsorbed. Therefore, the dielectric loss o f cyclohexane is sufficiently small such that
little microwave radiation is converted to heat. Thus, it fails to raise the temperature of
the adsorbent. This is consistent with the very low value for e" obtained in our dielectric
probe measurements.
However, Figure 2-11 indicates that, when methanol was the adsorbate on either
zeolite, a large portion o f the microwave heating was due to the microwave interactions
with the adsorbed methanol. This heating effect was manifested as temperature rises as
much as 19K above the blank DAY sample and 37K above the blank silicalite sample.
These interactions were dependent on the amount of methanol adsorbed on the zeolite. In
the case o f methanol adsorbed on DAY, the difference in temperature between the
methanol/DAY system and the DAY blank was relatively constant. A value o f 17K-19K
occurred over a range o f microwave power. The total amount o f methanol desorbed from
the DAY increases as the incident microwave power increases. Thus, as the power is
increased, the interaction o f the microwaves is stronger, but there is less methanol
adsorbed on the DAY for the microwaves to interact. These two processes offset each
other, and the temperature difference remains constant.
40
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
These series o f experiments show that a steady state between the amount of
methanol or cyclohexane that remains adsorbed is achieved within several minutes as the
amount of microwave energy is changed from OW to 120W. However, the temperatures
measured in the gas phase or within the bed differ at each level o f microwave energy
above zero. It should be possible to estimate the heats o f adsorption from a ClausiusClapeyron (Vads vs. Temperature) plot of these data. Different heats o f adsorption,
however, would be calculated from the use of either the gaseous or the bed temperatures.
The belief is that microwave radiation does not modify the heat o f adsorption. Rather,
the belief is that the “effective” temperature in the presence o f microwave radiation has
not been measured. Further, these systems are saturated systems where the amount o f
sorbate adsorbed is essentially independent o f pressure. Thus, isosteric heats o f
adsorption cannot be estimated from these data.
2.5.2.4
Low Temperature Desorption
In the systems where cyclohexane was the adsorbate, it was observed that
cyclohexane desorbed at temperatures lower than expected. At the 40W power level in
the DAY/cycIohexane system, the temperature of the DAY increased approximately 5K
above ambient due to the DAY self-heating. At this temperature, it was not expected that
desorption would occur; however, a significant amount o f cyclohexane was desorbed at a
system temperature o f only 298K. The temperature has been measured in the zeolite bed
and in the gas phase after the bed. These are not the local temperatures at the surface
where microwave energy is primarily absorbed. If the desorption rate is greater than the
rate at which heat is transferred to achieve thermal equilibrium, the measured
temperatures do not reflect the effective surface temperatures. Thus, the desorption is
41
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
found to occur without the necessity o f heating the system to the same temperatures
required if thermal equilibrium is achieved.
The process o f desorption is somewhat different from the simple heating o f the
system by microwave radiation; the process o f desorption is significantly endothermic.
Thus, the system is locally cooled as the desorption takes place. These findings suggest
that the microwave-induced desorption processes are quite complex. Most importantly,
desorption is found to occur without heating the whole system to the same temperature as
for a conventional process. What are the effective temperatures in this system: the gas
phase temperature, the bulk temperature, the instantaneous surface temperature (with
species adsorbed) or the “effective” surface temperature after desorption? The
conclusion is that these differences in temperature are inherent to the evident “microwave
effect” found for desorption induced by microwave radiation.
2.5 .3
Competitive Adsorption
It has been shown that there are inherent differences in the desorption of
cyclohexane and methanol. These differences show that the adsorbate has a direct
influence on the absorption of microwave energy. It has also been argued that absorption
o f microwave energy by a high-surface area, partially transparent adsorbent is a local
surface phenomenon. Six multicomponent competitive adsorption experiments in the
presence o f microwave radiation investigated:
1. What happens when two adsorbing species are present?
2. Is the selectivity of adsorption influenced by microwave radiation?
42
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The experiments were performed with DAY and silicalite zeolites using a binary
adsorption system consisting of one component with a low dielectric loss, cyclohexane,
and a second component with a high dielectric loss, methanol. A detailed explanation
and comparison o f each o f the six possible permutations is presented below. These
experiments show that microwave radiation can have a unique influence in competitive
adsorption, which is independent of the order in which the components are adsorbed.
2.5 .3.1
Initial Loading for Competitive Adsorption Experiments
There are three protocols for loading the zeolites with two components o f the
competitive adsorption experiments: sequential or simultaneous adsorption. Figure 2-12
represents the raw data collected for the initial loading, based on Case I (see section
2 .4 .3 .4), o f the cyclohexane-methanol-silicalite competitive adsorption experiment. At
time zero, a stream o f 15cm3-min'1 o f helium saturated at 293K with cyclohexane was
mixed with a diluent stream of 15cm3-min'1 of helium. This gas stream was introduced
into the system through the inlet manifold. Initially, the bed contained no adsorbates.
Adsorption commenced with the introduction of the saturated helium stream. A typical
breakthrough curve was observed between time zero and 240 minutes with breakthrough
occurring at approximately 60 minutes. The exotherm associated with the adsorption of
the cyclohexane was observed as an increase in the bulk temperature o f the zeolite
between zero and 60 minutes. With steady state established at 240 minutes and the
cyclohexane stream unchanged, the diluent stream was diverted through a second
saturator containing the methanol at 293K. As the methanol began to adsorb, it displaced
some o f the cyclohexane already adsorbed. This is evident as the cyclohexane
43
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
concentration increases above its steady state value, between time 240 minutes and 360
minutes. Also, between 240 minutes and 360 minutes the exotherm associated with the
adsorption of methanol is observed and due to the fact that the sorption capacity for
methanol was larger than for cyclohexane. At 360 minutes, breakthrough o f the
methanol occured as marked by the rapid increase in the concentration o f methanol in the
effluent. By 480 minutes, both cyclohexane and methanol concentrations had reached
new steady states. Similar results were obtained in the other five experiments.
The adsorption loadings o f each o f the three cases for each zeolite, shown in
Table 2-3, were obtained by integration o f the breakthrough curves, similar to the curve
in Figure 2-12. The loadings are based on moles o f adsorbate per gram o f adsorbent.
Table 2-3 - Summary o f zeolite loadings after initial binary adsorption for all cases moles adsorbate per kilogram of adsorbent
Zeolite
Silicalite
Silicalite
Silicalite
DAY
DAY
DAY
Adsorption Method
Case
Case
Case
Case
Case
Case
moles C6H12 /
kg adsorbent
0.35
0.15
0.21
0.80
1.53
1.06
1
2
3
1
2
3
moles C H 30H /
kg adsorbent
1.74
1.74
2.14
1.40
1.20
1.50
44
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
294
291
288
0.004 o
- 0.002 m
0.000
60
120
180
240
300
360
420
480
Time (min)
Downstream Bed T em perature
Upstream Bed Tem perature
—•— Effluent G as Tem perature
—■— Effluent Methanol Concentration
• • • Effluent Cyclohexane Concentration
Figure 2 -1 2 - Initial adsorption, for Case 1 (cyclohexane adsorbed before methanol is
introduced at 290 minutes), o f the cyclohexane-methanol-silicalite system. Temperatures
in the bed (upstream and downstream) and in the gas effluent are shown above while the
effluent concentration o f methanol and cyclohexane (dotted) are shown below.
2.5 .3 .2 Influence o f Microwave Radiation
Once the initial loading o f the zeolites was complete, the competitive adsorption
experiments using microwave radiation were performed. Figure 2-13 represents the raw
data collected for the cyclohexane-methanol-silicalite competitive adsorption experiment
that was initially loaded using Case 1. At 480 minutes, the adsorbate/adsorbent were
exposed to microwave radiation by a step change in the incident power from zero to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40 W. Since the zeolite is saturated with equilibrium amounts o f both cyclohexane and
methanol, any increase in the zeolite temperature will cause desorption to occur. Thus,
both the concentrations o f the cyclohexane and o f the methanol begin to rise immediately
after the microwave radiation was applied and the bulk temperature o f the zeolite rose.
However, microwave radiation had a greater interaction with methanol than with
cyclohexane, as observed earlier. Therefore, as the methanol was desorbed, the
cyclohexane, having weak interactions with microwaves, can re-adsorb preferentially into
the zeolite in lieu of the methanol, thus changing the adsorption selectivity. This is
evident in Figure 2-13 as the concentration o f cyclohexane in the effluent drops below its
initial steady state value and as the concentration o f methanol in the effluent rises above
its initial steady state value.
Because the microwave interactions with the adsorbed species cause the
temperature o f the zeolite to increase, it could be argued that pure heating effects are the
cause o f the differences in the species sorbed at steady state and not the microwave
radiation. However, the heat of sorption for methanol in silicalite, 43kJ-mole'' (Thamm
1989), or methanol in DAY, 45 kJ-mole"1 (Izmailova, Karetina et al. 1994), is lower than
the heat of sorption for cyclohexane in silicalite, SSkJ-mole'1(Cavalcante and Ruthven
1995) or cyclohexane in DAY, 50 kJ-mole'1 (Barthomeuf and Ha 1973). Thus, the
species with the larger heat of adsorption (cyclohexane) should be the component that
desorbs more with increasing temperature. Yet, experiments show that the amount of
cyclohexane adsorbed actually increases. This is due to the ability o f methanol to absorb
microwave energy more efficiently than cyclohexane.
46
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
350
330
*
310
3
5 290 ^
aQ>E
a)
l—
- 0.006
—
- 0.004
,
-
0.002
r~
480
600
720
840
960
1080
1200
1320
1440
Time (min)
Downstream Bed T em perature
U pstream Bed T em perature
—•— Effluent G as T em perature
— • Effluent M ethanol C oncentration
—•— Effluent C yclohexane C oncentration
Figure 2 -1 3 - Competitive sorption of the cyclohexane-methanol-silicalite system
initially loaded using Case 1 (Figure 2-12). Temperatures in the bed (upstream and
downstream) and in the gas effluent are shown above while the effluent concentration of
methanol and cyclohexane (dotted) are shown below.
After steady state is established at the 40W power level, at approximately 635
minutes, the power level was changed to 80W in a stepwise manner. Results similar to
those at the 40W power level were obtained. The same was true for the 120W power
level. Once steady state was established at the highest power setting, the power was
decreased in the reverse sequence, from 120W to zero in 40W increments, waiting for
steady state at each power level.
47
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figures 2-14 and 2-15 are a graphical summary o f raw data collected during these
six competitive adsorption experiments. Figure 2-14 is the summary o f the three
competitive adsorption experiments performed on the DAY sample, and Figure 2-15 is
the summary of the three competitive adsorption experiments performed on the silicalite
sample. The plots are o f the form In y or In P vs. 1/T where y and p are defined as
follows:
M icro w a ve Power
^ Cyclohexare
,\,ficrowave Power
y
t
_______________________t M ethanol __________
Microwave Power=0
Equation 2.8
V Cyclohexare
\-ficrowave Power=0
t M ethanol
Q c vclohexate
Q M ethanol
Qcvclolie:xate
Q M ethanol
exp
-A H c+ AHm
R
Equation 2.9
T
oj
The ratios y and P may be viewed as adsorption selectivities based on the relative
amounts of methanol and cyclohexane adsorbed in the zeolite at conditions produced by
microwave radiation and conventional heating respectively. Each o f the ratios is
normalized to ambient conditions, room temperature with no applied source o f radiation.
Figures 2-14 and 2-15 show how microwave radiation changes the adsorption selectivity
o f cyclohexane over methanol versus conventional heating in both DAY and silicalite.
As microwave radiation is applied to the system, the selectivity for adsorption o f
48
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cyclohexane over methanol increases. This increase in the adsorption selectivity is seen
in both Figures 2-14 and 2-15 as a rise in (In y) with increasing microwave power and
ultimately rising bulk zeolite temperature. However, if this experiment were conducted
using conventional heating the resulting adsorption selectivity (P) would be based on the
heats o f adsorption o f methanol and cyclohexane on the zeolite.
■■s
cr
1
&
>
"o
a>
20W
■
♦
•
80V\l
>
4i
40' V
♦
It*
(w
a>
OT
1r
▼
-1
3.05
3.10
3.15
3.20
3.25
3.30
3.35
3.40
3.45
1/T*103 (1/K)
•
■
♦
▼
In a vs
In a vs
In a. vs
In p vs
1/T based
1/T based
1/T based
1/T based
on Case 1
on Case 2
on Case 3
on heats of adsorption
Figure 2-14 - Summary o f the competitive adsorption experiments performed on the
zeolite DAY - ln(Selectivity Ratio: a or P) versus 1/T, see text for definition.
49
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
f
120W
is
a.
&
>
o
0)
0)
w,
^
80W
40W
o
ovT
-1
▼
-2
2.90
3.00
3.20
3.10
3.30
3.40
3.50
1/T*10 (1/K)
•
■
♦
▼
In a vs
In a vs
In a vs
In p vs
1/T
1/T
1/T
1/T
based on Case 1
based on Case 2
based on Case 3
based on heats of adsorption
Figure 2 -1 5 - Summary o f the competitive adsorption experiments performed on the
zeolite silicalite - ln(SeIectivity Ratio: a or p) versus 1/T, see text for definition.
Figures 2-14 and 2-15 show that the adsorption selectivity o f cyclohexane over methanol
decreases with 1/T. These results clearly show that microwave radiation changes the
adsorption selectivity o f both the cyclohexane-methanol-silicalite and cyclohexanemethanol-DAY systems. However, the magnitude of the changes differs with the zeolite.
There is no significant influence o f the order o f adsorbate introduction.
50
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2.6
Conclusions
The interaction o f microwave radiation with materials is complex. To de­
convolve these interactions, three types o f experiments were performed. The results
from these studies were combined to yield a more complete picture o f the interactions of
microwave radiation with our sorbing systems. The first study examined the interaction
of microwave radiation with the zeolite itself. These results show that zeolites absorb
microwave radiation only quite weakly. The conclusion is that interaction between the
microwaves and zeolites is due to the surface and silanol groups on the surface o f the
zeolite. Silanols have a significant dielectric permittivity; therefore, they can selectively
absorb microwave energy. The fiber-optic temperature probes are not able to measure
the temperature o f these surface silanols. However, the temperature o f the zeolite and of
the gas passing over the zeolite was measured. These temperatures suggest that the heat
transfer is not rapid enough to achieve thermal equilibrium. Thus, the rate by which the
microwave energy is absorbed by the surface silanols is greater than the rate by which the
heat is transferred to zeolite and finally to the gas phase. The conclusion is that the
temperature profile in the system is TSiianoi (not measured) >TBuik > T gos- Note that these
measurements are steady-state values and not equilibrium values
With a baseline established for the interaction of the microwave radiation with the
support, the second study was to understand how a single adsorbate on the zeolite
interacts with microwave radiation. Based on the temperatures measured during the
desorption experiments, cyclohexane has weak interactions with the microwave radiation.
This is manifested by temperature rises above ambient that were the same, within
experimental error, as the zeolite without sorption. Conversely, during the methanol
51
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
desorption the temperature increase was significant, as much as 20K above the blank
zeolite temperature. This leads to the conclusion that methanol has a rather strong
interaction with the microwave radiation. These results are not surprising in that the
dielectric constants for these adsorbates suggest that cyclohexane (dielectric constant of
2.023) should have weak interactions with microwave radiation, whereas methanol
(dielectric constant of 33.62) should have strong interactions. This was also confirmed
with the microwave probe measurements.
In addition to measuring the strength o f the interactions of the microwave
radiation with the zeolite, it was observed that cyclohexane desorbed at temperatures
between -2 9 8 K and -333K, this is lower than would be expected in conventional heating
(normal boiling point of cyclohexane is -353K (Lide 1991)). The temperatures measured
in the zeolite bed and in the gas phase after the bed are average bulk zeolite temperatures
and not the local temperatures at the surface where microwave energy is primarily
absorbed. Thus, the desorption occurs without the necessity of heating the system to the
same temperature required if thermal equilibrium were achieved. The desorption process
induced by microwaves is complex. The effective temperatures in this system are the gas
phase temperature, the solid temperature, the instantaneous surface temperature (with
species adsorbed) or the “effective” surface temperature after desorption. The conclusion
is that differences between these temperatures are inherent to the evident “microwave
effect” found for desorption induced by microwave radiation.
Finally, the third study examined the ability of microwave radiation to influence
the selectivity of the sorption process. The results o f these experiments show that with an
increase in the microwave power, and thus the temperature of the system, that methanol
52
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
is desorbed and cyclohexane adsorbed in its place. This result is the obverse to what
would occur in conventional heating. Because the heat o f adsorption o f methanol (43 kJmol'1 in silicalite and 45kJ-m of1 in DAY) is lower than the heat o f adsorption o f
cyclohexane (63kJ-mor' in silicalite and 50kJ-mofl in DAY), cyclohexane should have
been desorbed and methanol should have been adsorbed in its place or desorbed to a
lesser extent. Thus, these results show that independent o f how the zeolite was initially
loaded, the microwave radiation changes the adsorption selectivity o f both the
cyclohexane-methanol-silicalite and cyclohexane-methanoI-DAY systems. It is
hypothesized that this change in selectivity will occur for other competitive sorption
processes in the presence o f microwave radiation. The changes in selectivity will depend
on the dielectric properties o f the adsorbate/adsorbent system.
Microwave radiation would be expected to have a similar influence on sorption
for other systems where the bulk is essentially transparent to microwaves. This would
include oxides and mixed metal oxides which have low permittivity, (e.g., silicas,
aluminas, silica-aluminas, zeolites, etc., but not activated carbons). However, the ability
of the surfaces o f these adsorbents to absorb microwave radiation would depend on the
surface chemistry (e.g., hydroxylation) and the permittivity o f the adsorbed species. It is
suggested that all surfaces possess higher permittivities than the bulk due to the potential
polarizability of the terminal bonds.
There are three advantages of the use of microwave radiation to influence
sorption. Firstly, the overall energy demands for desorption (regeneration) induced by
microwave energy are less than those induced by conventional resistive heating or steam
stripping. The conventional approaches of regenerating an adsorbent require that the
53
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
whole system (adsorbent and gas phases as well as the container, etc.) must be raised to
the desorption temperature. Microwave-induced desorption need only directly heat the
surface/adsorbed phases. Secondly, the time required for regeneration by microwave
heating can be significantly less than the time required by conventional approaches.
Microwave energy is transferred more rapidly and can be more selectively absorbed by
the adsorbed phase. Finally, the selectivity o f desorption may be controlled. This control
is possible if the permittivities o f the adsorbates differ.
54
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 3
THE EFFECT OF MICROWAVE ENERGY ON THREE-WAY AUTOMOTIVE
CATALYSTS POISONED BY S 0 2
3.1
Objectives
The study o f sulfur poisoning on three-way automotive catalysis is o f increasing
importance in light o f the recent amendments to the Clean Air Act o f 1990. Numerous
studies have demonstrated that the sulfur present in gasoline decreases the effectiveness
o f catalysts for the control o f contaminants in auto exhaust (Engler, Lindner et al. 1995).
All automotive fuels produced today have some level o f sulfur-containing organic
compounds present. When burned in an internal combustion engine, these compounds
are converted primarily into sulfur dioxide (S 0 2). When the S 0 2 enters the catalytic
converter, it may adsorb on the noble metals o f the catalyst. The adsorbed sulfur can
react to form sulfides with the noble metals under reducing conditions and sulfites or
sulfates under oxidizing conditions preventing the adsorption o f carbon monoxide (CO)
for further oxidation reactions. In addition, the S 0 2 forms sulfates with the base metal
oxides, such as ceria (CeOx) (Lundgren, Spiess et al. 1995). When these sulfates form,
the base metal oxides may lose their oxygen storage capabilities. This results in poor
performance for feed streams where the air to fuel ratio is less than stoichiometry (rich
conditions). The cost o f completely removing these sulfur compounds is not
economically practical with the current fuel processing technologies. For this reason,
alternative methods must be developed to counter the detrimental effects o f S 0 2 on the
three-way catalyst.
55
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Several studies have investigated the effects of microwave radiation on catalytic
reactions (Krieger-Brockett, Mingos et al. 1993; Cha and Kong 1995; Bond and Moyes
1997). The previous chapter has shown that microwave energy can be selectively
absorbed by the surface o f oxide supports (Turner, Conner et al. 1998; Turner, Laurence
et al. 2000). The result was a rapid, efficient, and selective heating o f the surface. It was
inferred that a steady-state temperature gradient existed in a flowing gas with the
temperature of the surface being the greatest, followed by the bulk solid temperature, and
finally the gas phase temperature. Consequently, desorption was induced at a lower bulk
solid temperature than would be required for conventional desorption. This does not
mean desorption occurred at a lower temperature, since the temperature o f the surface
could be greater than the normal desorption temperature. The importance is that the
microwave energy is directed toward the surface of the support, and in some cases toward
the adsorbate itself (if the adsorbate has a high dielectric constant). This focusing o f the
microwave energy allows the desorption to occur without heating the entire system to the
desorption temperature. Therefore, microwave energy could provide a two-fold
advantage in the SO2 /TWC system. Firstly, the microwave energy may be able to desorb
the surface adsorbed sulfur since sulfates and sulfides are among the most polarizable
bonds in the system and would be most receptive to absorb microwave energy. Secondly,
if microwave energy can selectively heat the surface of the TWC, lightofF (defined here
as the temperature when 50% of final conversion is achieved) may be induced at a bulk
temperature lower than is normally required, although the effective surface temperature
could be much greater. This would greatly reduce overall emissions o f CO since a large
56
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
portion o f the CO in the exhaust o f an automobile is emitted prior to lightoff o f the
catalyst.
3.2
Experimental Aspects
3.2.1 Materials
This study used Engelhard Corporation’s state-of-the-art three-way automotive
catalyst. This catalyst was supplied in the form o f a 1.25cm diameter by 3.25cm long
ceramic monolith. A summary o f the catalyst composition is presented in Table 3-1.
Table 3-1 - Engelhard three-way catalyst composition as reported in the Material Safety
Data Sheet (MSDS).
<2w/o neodymium oxide
<2w/o lanthanum oxide
<2w /o rhodium
< lw /o palladium
< lw /o platinum
>50w /o ceramic
<20w /o aluminum oxide
<20w/o zirconium oxide
<10w/o cerium oxide
<5w/o strontium oxide
<2w/o nickel oxide
The reactants were purchased as premixed gases in four separate cylinders from
Matheson Gas Products Incorporated. This reduces the overall operating cost while
maintaining the ability to change the feed composition. The four gas mixtures are shown
in Table 3-2. Helium was chosen over nitrogen as the carrier gas for two reasons.
Firstly, nitrogen and carbon monoxide have the same mass (m/e = 28), thus if nitrogen
were used as the carrier gas, it would complicate the carbon monoxide in the mass
spectrometer detector. Secondly, helium has a higher thermal conductivity than nitrogen,
which allows for better heat transfer to and from the catalyst.
57
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 3-2 - Reactant gas cylinder concentrations.
3.2.2
Cylinder
Composition
1
2
3
4
Ultra-high-purity helium (99.9999%)
1.38% C 3 H 8 and 5.25% CO in helium
47% O 2 in helium
0.283% SO 2 in helium
Apparatus
The experiments were performed with an apparatus consisting o f an inlet
manifold connected to a reactor designed for exposure to microwave radiation, Figure 31. The inlet manifold consists o f four computer-operated mass flow controllers. These
controllers can mix the reactants over a wide range o f ratios for inlet into the reactor. All
tubing in the manifold is nickel-free stainless steel for two reasons. Firstly, it is resistant
to corrosion by sulfiirous and sulfuric acids, which can be formed as byproducts.
Secondly, if stainless steel containing nickel were used, the carbon monoxide in the feed
stream could react with the nickel, forming nickel carbonyl, a nerve toxin.
58
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Legend
MFC - M ass F low Controller
0
- V alve
Microwave
Catalyst
Waveguide
Helium
MFC I
MFC 2
MFC 3
MFC 4
Figure 3-1 - Inlet manifold for the three-way catalyst experiments. The on-line mass
spectrometer and mass flow controllers is each computer controlled.
The reactor is similar to that presented in section 2.4.2 with the addition of four330W bar heaters clamped to the exterior o f the waveguide. These heaters are used to
heat the catalyst up to 5 13K. A K-type thermocouple and an on-line Balzers QMG112A
Quadrupole Mass Spectrometer were connected to the apparatus. The fiber optic
temperature probes were not used in this experiment because the operating temperature of
the reactor is beyond the design limit o f the probes (473 K). The thermocouple used to
measure the gas effluent was placed at the end o f the microwave chokes to avoid the
microwave radiation from using the metal as an antenna, which could cause electrical
arcing inside the reactor. The mass spectrometer was used to measure the concentration
o f the products in the effluent or reactants in the inlet by bypassing the reactor.
59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
3.2.3
3.2.3.1
Procedure
Temperature Calibration
As mentioned above, the thermocouple cannot be in physical contact with the
catalyst when microwave energy is present. Therefore, in order to estimate the bulk
catalyst temperature in the presence o f microwave energy, a calibration was performed.
Using only conventional heating, the thermocouple was placed in direct contact with the
heated catalyst and the temperature was recorded. Next, a short length o f tubing was
added so that the thermocouple was just outside the microwave choke and this
temperature was recorded for the same heating power. This calibration is used to correct
the effluent gas temperature to reflect the bulk catalyst temperature in the presence of
microwave radiation.
3 .2.3 2 Sample Preparation
Prior to all experiments, the catalyst was heated in a tube furnace to 773 K in the
presence o f oxygen to remove any adsorbed poisons or impurities. The catalyst was then
allowed to stabilize at 373K for Ih under a flow o f the reactant mixture. The reactor feed
composition was chosen to simulate the exhaust o f a late model 3.8L V-6 engine (Beck
and Sommers 1995). The feed concentrations and m/e o f both the reactants and products
monitored by the mass spectrometer are reported in Table 3-3. The stoichiometry was
cycled between net reducing and net oxidizing conditions at 0.5Hz to simulate closed
loop operation o f a typical engine. This was accomplished by varying the oxygen
concentration in the feed by ±3% centered about stoichiometry. The total flow rate of
reactants through the catalyst for these experiments was 1.67 L-min'1, a space velocity o f
60
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
approximately 20,000 h'1. Sulfur dioxide was included in the feed at 30ppm to study the
effects of poisoning on the CO lightoff temperature.
Table 3-3 - Feed composition and m/e o f both the reactants and products monitored by
the mass spectrometer.
Compound
Feed Concentration (ppm)
m/e
He - Helium
CO - Carbon Monoxide
C 3 H8 - Propane
O 2 - Oxygen
CO 2 - Carbon Dioxide
SO 2 - Sulfur Dioxide
Balance
1150
300
1988-2122 (±3% step change at 0.5 Hz)
0 (product, not present in feed)
30 (used only during sulfur studies)
4
28
29
32
44
64
Sets o f transient experiments were performed on the three-way automotive
catalyst using conventional heating and heating assisted by microwave energy. These
tests were used to determine the following:
1.
The effect o f microwave energy on the CO lightoff temperature o f the
catalyst both in the presence and absence o f SO 2 ; and
2.
Whether microwave energy could reverse the poisoning of the CO
conversion by SO2 .
3.2.33
Transient Experiments
With the temperature of the catalyst stable at 373K, the microwave energy was
introduced at 300W. Simultaneously, the bar heaters clamped on the exterior o f the
waveguide were turned on. This resulted in the bulk catalyst temperature being increased
from 3 73K to 513K. After the catalyst reaches 513K, the heaters were turned off and the
catalyst allowed to cool back to 373K. During the experiment the effluent gas was
analyzed for the following m/e (species): 28(CO), 29(C3Hs fragment), 32(02), and
61
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44(CC>2). The experiment was repeated in the absence of microwave energy to measure
the CO lightoff temperature for conventional heating.
It is known that SO 2 shifts the CO lightoff temperature higher in temperature
(Beck and Sommers 1995). Therefore, these experiments were repeated with SO 2 present
in the feed stream to determine whether microwave energy can shift the CO lightoff
temperature back to its un-poisoned value.
3 .3
3 .3 .1
Results and Discussion
Effect of Microwave Energy and SO 2 on CO Oxidation
For the purpose o f this discussion, the CO lightoff temperature is attained when
50% of the final CO conversion is achieved. The effect of microwave energy on the CO
lightoff temperature in the absence o f SO2 is shown in Figure 3-2. The lightoff
temperature using only conventional heating was 485K. However, a combination o f
conventional heating and microwave energy shifts the lightoff temperature approximately
9K lower to 476K.
The final bulk temperature o f the catalyst was 5 13K regardless o f the presence o f
microwave radiation. This is not remarkable since the monolith is mostly ceramic, a
substance o f low permittivity. Thus, only a fraction o f the microwave energy was
adsorbed in this experiment; most o f the energy passed through the waveguide and was
absorbed by the water load. Therefore, it is concluded that the small fraction o f
microwave energy was selectively adsorbed by the surface/adsorbed-layer o f the threeway catalyst. Although the effective surface temperature cannot be measured, it is
believed that the effective surface temperature is greater than the bulk catalyst
62
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
temperature which is greater than the gas phase temperature in these steady-state flow
experiments.
0.45
a
0.40 -
□
Conventional Heat
Conventional Heat w/MW
0.35 ra
c
0.30 -
O)
Sj
0.25 -
0)
^
0.20 -
0.10
-
0.05
453
463
483
473
493
503
513
Bulk Catalyst Temperature (K)
Figure 3-2 - Carbon dioxide (m/e = 44) production versus bulk catalyst temperature for
conventional and microwave experiments
Therefore, the actual CO lightoff temperature may not have decreased, because the
effective surface temperature not known. However, the bulk catalyst lightoff temperature
has been decreased by 9K. This is significant because a large fraction o f the pollution
from automobiles is emitted prior to the lightoff o f the catalyst. Therefore, if the catalyst
were made to lightoff at a lower temperature, the amount o f pollution generated by
automobiles would be drastically reduced.
63
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The transient experiments were performed a second time with SO 2 in the feed
stream. These results, in addition to the data presented in Figure 3-2, are reported in
Figure 3-3 for comparison. This data confirms previous reports that SO 2 shifts the CO
lightoff temperature higher, from 485K for the un-poisoned catalyst to 494K for the
poison catalyst when only conventional heating is used. However, if a combination o f
conventional heating and microwave energy is used. Figure 3-3 shows that the lightoff
temperature is reduced from 494K to 482K. Thus, microwave energy has lowered the
lightoff temperature 12K. This is 3K lower than the lightoff temperature o f the un­
poisoned catalyst using only conventional heating, 485K.
The S 0 2 poisoning not only shifts the CO lightoff temperature higher but also
decreases the overall conversion o f CO. Figure 3-3 shows that for a poisoned catalyst the
overall CO conversion is 85% at the maximum temperature o f 5 13K when using only
conventional heating. When microwave energy is used to augment the conventional
heating, the conversion is 100% at a temperature o f only 500K.
The most probable explanation for the results presented above is that the sulfides
and sulfates formed during poisoning by the SO2 are the most polarizable bonds in the
system and are most likely to absorb microwave energy. This absorption o f microwave
energy causes these sulfur compounds to decompose and desorb with sulfur probably
being removed as SO 2 . This leaves the noble metals and base metal oxides available for
CO oxidation, leading to an increase in conversion. Any further microwave energy
absorbed at the surface could then increase the effective surface temperature or interact
with adsorbed species to promote lightoff at a lower bulk temperature.
64
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
^
□
■
0.45
0.40 -
a
Conventional Heat
Conventional Heat w/ MW
Conventional Heat w/ MW and S 0 2
Convnetional Heat w/ SO,
0.35 aj
. I 5 0.30 ^
C/5
© 0.25
0.20
-
0.15
0.10
453
463
473
483
493
503
513
Bulk C atalyst Tem perature (K)
Figure 3-3 - Carbon dioxide (m/e = 44) production versus bulk catalyst temperature for
conventional and microwave experiments in the presence and absence o f sulfur dioxide, a
catalyst poison
3.3.2
Effect of SO2 on Propane Oxidation
During the study on the effect of microwave energy on the three-way automotive
catalyst, a result was discovered that has not been seen in the literature to date. All the
previous research suggests that SO 2 increases the lightoff temperature o f the catalyst and
reduces the conversion (Monroe, Krueger et al. 1991; Beck and Sommers 1995).
Although a reduction in the hydrocarbon conversion was observed, it was noted that the
SO2 reduced the hydrocarbon lightoff temperature o f Englehard Corporation’s catalyst by
approximately 100K.
65
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
These results came about while using a tube furnace to perform the conventional
experiments; this was prior to the construction o f the waveguide reactor, which is limited
to 5 13K. Other than the method o f heating the overall system is identical to the one used
in the microwave experiments. Three types o f experiments were performed using the
tube fUmace setup. In all experiments, the temperature was cycled from 373K to 773K
back 373K. The first two experiments, the catalyst was cycled in temperature both in the
presence and absence o f SO 2 . The third experiment the catalyst was poisoned at low
temperature, the poison removed from the gas stream, and the temperature cycle
performed. The results o f these experiments are presented in Figure 3-4.
For the un-poisoned experiment, the results show that the conversion increases
monotonically up to the final temperature o f 773K. It is possible that if the temperature
were allowed to increase the conversion would increase toward 100% from 70% at 773K.
The lightoff temperature for this experiment is approximately 650K. For the poisoned
experiment, the conversion o f the hydrocarbon increases rather rapidly to its final value
o f 70% at 773 K with a lightoff temperature close to 540K. The fact that the conversion
has reached its maximum value at only 70% is typical, however, the decrease in the
hydrocarbon lightoff temperature from 650K. for the un-poisoned experiment to only
540K is rather surprising.
66
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
0.16
Conventional Heat w/ S 0 2
Conventional Heat
0.14
8
QO
g 0.12
C/3
05
CNI
-5
°o
0 .1 0
0.08
0.06
400
600
500
700
800
T em perature (K)
Figure 3-4 - Propane (m/e = 29) production versus bulk catalyst temperature for
conventional experiments in the presence and absence o f sulfur dioxide, a catalyst poison.
In the third type of experiment, as the temperature is increased from 3 73 K to
773 K the results are similar to those for the poisoned experiment, however, as the
temperature is decreased the results are similar to un-poisoned experiment. Thus, as the
catalyst begins to heat the poison is still adsorbed and the results correspond to those o f
the poisoned experiment, as the temperature increases further the poison desorbs.
Finally, as the temperature is decreased, the poison is no longer present and the results
follow that o f an un-poisoned catalyst.
67
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
At this time, there is no reasonable explanation as to the cause o f these results.
This effect is mentioned here so that future members o f our research group can further
investigate this phenomenon.
3.4
Conclusions
In this work, the effect o f microwave energy on a state-of-the-art three-way
automotive catalyst both in the presence and absence o f SO2 has been studied. The
results obtained provide four general conclusions.
1.
Microwave energy can shift the lightoff temperature o f the bulk catalyst 9K
lower when compared to the lightoff temperature using only conventional
heating.
2.
Results obtained by other research groups that SO 2 shifts the CO lightoff
temperature o f the bulk catalyst higher when compared to an un-poisoned
catalyst was confirmed.
3.
Microwave energy can shift the lightoff temperature o f the bulk catalyst
poison by SO 2 lower than the lightoff temperature o f the un-poisoned catalyst
using only conventional heating.
4.
For a catalyst poisoned by SO 2 , microwave energy can increase the overall
conversion o f CO from 85% at 5 13K using conventional heating, to 100% at
500K using microwave energy.
These studies demonstrate the ability o f microwave energy to influence
heterogeneous catalytic reactions. This influence will be greatest if the energy can be
directed to the surface, i.e., the bulk does not interact strongly with the microwave
68
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
energy. The concentration o f microwave energy on the surface will depend on the nature
o f the surface (hydroxyls, etc.), active sites (metals), and the specific adsorbing (reacting
or desorbing) species. It is also important that the reaction or desorption can occur prior
to thermal equilibration; otherwise, there will be little difference for microwave induced
sorption/reaction compared to conventional heating.
In addition, it should be noted that these experiments did not investigate the
influence o f microwave power on the reaction. However, using higher microwave power
could cause an even more pronounced shift in the lightoff temperature. Finally, the
unique result that SO2 has lowered the propane lightoff temperature was observed. This
result has not previously been published to our knowledge.
69
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 4
THE DESIGN AND CONSTRUCTION OF A FREQUENCY RESPONSE
APPARATUS TO INVESTIGATE DIFFUSION IN ZEOLITES
4.1
4.1.1
Background
Elementary Principles o f Diffusion
There are two major diffusion processes that occur in nature. The first type is
mass transfer diffusion that results from a concentration gradient. The second type is
Brownian motion, which is molecular motion in the absence o f a concentration gradient.
Fick’s First Law describes mass transfer diffusion o f single species (Fick 1855):
Equation 4.1
J = -DW c
where J is the net flux of molecules, D is the diffusion coefficient, and Vc is the
concentration gradient.
In the case of Brownian motion, tagging a fraction of the molecules, radioactively
or otherwise, and measuring their mean displacement determines the rate o f diffusion.
The random walk theory is used to correlate the mean displacement of the tagged
molecules to the diffusion coefficient. In 1905 Einstein showed that the diffusivity and
the mean displacement are related as follows (Einstein 1905):
Equation 4.2
for one dimensional diffusion,
( r 2(t
= 6 Dt
Equation 4.3
70
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
for isotropic diffusion in three dimensions. These equations are known as Einstein’s
relations.
By the addition o f geometric constraints to the relations developed in the random
walk theory, it is possible to describe diffusion in zeolites. If a zeolite is modeled as a
series o f interconnecting cages, then an individual jump can be described in three parts.
The first is a jump from the wall o f the cage to the center o f the cage, the second is a
jump from the center o f one cage to the center o f another, and third is a jump from the
center of the second cage to the wall o f the cage (Figure 4-1 (Karger and Ruthven 1992)).
This constraint results in the following relation between the mean displacement and the
diffiisivity (Karger and Ruthven 1992):
Equation 4.4
which reduces to equation 4.3 for random walk in an unconstrained system.
4.1.2
Techniques used to measure diffusion coefficients
Currently, there are quite a number o f experimental methods used to measure
diffusion coefficients. These methods generally fall under the categories o f sorption rate
measurements or chromatographic techniques. The remainder o f this section is devoted
to describing these methods.
71
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Ri2
R il
Figure 4-1 - Sketch o f the random walk theory confined by geometric constraints (Karger
and Ruthven 1992)
4.1.2.1
Zero Length Chromatography (ZLC)
The principle o f ZLC is simple. A small, thin sample is placed in-line in a gas
stream. The gas stream is connected to a flame ionization detector (FID) or thermal
conductivity detector (TCD) o f a gas chromatograph. The sample is loaded with an
equilibrium amount o f sorbate. At t = 0, the gas stream is switched to an inert purge gas
at a sufficient flow rate to maintain a zero concentration of sorbate on the exterior o f the
crystals.
The technique o f ZLC has major advantages and disadvantages over the static
sorption methods that follow. The advantage is that if the carrier flow rate is kept
sufficiently large, mass and heat transfer limitations can be greatly reduced. The
disadvantage is that the response can be distorted by mass transfer limitations and axial
dispersion in the sample bed (Karger and Ruthven 1992).
72
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.2.2 Constant Pressure Sorption Method
Consider a small sample in a volume maintained at constant pressure. At t = 0, a
step change in the sorption pressure is made. The adsorption or desorption is then
measured as a function of time either volumetrically or gravimetrically. For the
volumetric measurement, the system pressure is maintained constant by the use o f gas
burette. Therefore, at constant pressure the amount adsorbed or desorbed is proportional
to the volume change. In the gravimetric case, the sample mass is placed on a very
sensitive balance and the weight gain or loss is the measure o f the amount sorbed or
desorbed. In order to satisfy the constant pressure condition, a closed system volume is
chosen to be sufficiently large and the sample size sufficiently small so that the change in
volume associated with sorption is negligible. A benefit o f using a small sample is that
heat and mass transfer limitations are reduced (Karger and Ruthven 1992).
4.1.2.3
Constant Volume Sorption Method
In this method, the sorption is measured by the change in pressure o f a closed
system in which a step change is made to sorption pressure. In this case, the volume o f
the system is best-kept small and the sample large so that a large pressure signal may be
measured, increasing the accuracy. This technique is able to measure faster diffusion
coefficients than the constant pressure method because the pressure transducer has a
faster response time than the gravimetric balance. The disadvantage of this system is for
low pressures or strongly adsorbing species; the time required to introduce or remove the
sorbate becomes the limiting step. This problem is reduced by the technique that follows
(Karger and Ruthven 1992).
73
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.2.4 Step Response Method
In order to reduce the time required to introduce or remove the sorbate, the step
response technique was developed. In this method, instead o f introduction of the sorbate,
the sorbent contains an equilibrium amount o f sorbate. At time t = 0, the volume is step
changed using a piston or metal bellows, and the pressure response is measured as the
system re-equilibrates (Karger and Ruthven 1992).
4.1.3
Frequency Response Method
The frequency response method is closely related the step response method
mentioned above. In this case, instead o f using a piston or bellows to compress the
system, the piston or bellows is driven by a periodic function. This periodic function can
be o f any form; however, a sinusoidal function is widely used. A fast response pressure
transducer measures the pressure response o f this volume perturbation. This response is
then fit using a set o f equations derived by Yasuda (see section 4.3 .1). From this fit, the
diffusion coefficient is estimated. The model developed by Yasuda is linear, and
therefore, one must be cautious and operate in a regime where the adsorption isotherm is
linear over the conditions of the experiment. This approximation is usually valid for
volume changes o f up to ±5%. A benefit of this linearity is that if two independent
processes are occurring with different time constants, it is possible to represent the
contributions o f each process using Yasuda’s model. Thus, frequency response is able to
discern more than one time constant occurring simultaneously in a system.
74
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.1.4
History o f the Frequency Response Method
The first use of frequency response (PR) to measure transport properties was in
1861 when Angstrom applied a sinusoidal temperature to the end o f a metal rod and
measure the resulting temperature a distance down the rod (Angstrom 1861). In 1963
Naphtali and Polinski used the FR method to study the adsorption o f hydrogen on a
nickel catalyst (Naphtali and Poinski 1963). This method proved useful because
hydrogen adsorption on nickel occurs at two distinct time scales. Naphtali and Polinski
were able to characterize this difference using the FR method.
The first use of FR to measure mass diffusion was in 1970 when Evnochides and
Henley measured the diffusion o f ethane in polyethylene (Evnochides and Henley 1970).
This was accomplished by varying the gas pressure sinusoidally at low frequencies, and
the mass o f the polyethylene sample was measured using a microbalance.
Although Yasuda was not the first person to the FR method, he is responsible for
making the method a useful tool in the measurement o f diffusion coefficients. In 1976,
Yasuda determined that the technique formulated in the Naphtali and Polinski paper
could be extended to gas-solid systems (Yasuda 1976a; Yasuda 1976b). In 1982 Yasuda
studied the diffusion of krypton in sodium mordenite using frequency response (Yasuda
1982). Based on Fick’s law o f diffusion in zeolites, Yasuda derived the characteristic
functions for the in-phase and out-of-phase components o f the pressure response to a
sinusoidal volumetric variation. The equations were derived for a plane sheet and an
isotropic sphere, which represent one-dimensional and three-dimension diffusion in a
zeolite. The functions that Yasuda derived are linear functions and, therefore, can be
75
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
added together to account for processes occurring at more than one characteristic time
scale.
In 1992, the group o f Oprescu, Rees, and Shen derived the in-phase and out-of­
phase functions for the parallelepiped crystal geometry. This allows for the measurement
o f the diffusion coefficient in all three dimensions; thus, one can determine the anisotropy
o f diffusion in the crystal.
Finally, in 1996, Grenier, Meunier, Bourdin, and Sun developed the thermal
frequency-response technique. In this method, the phase lag between the temperature and
pressure response determines the diffusion coefficient. This new method permits the
investigation of thermal effects that occur during sorption.
4.2
Objective
Diffiisional mass transport in zeolites is of significant importance in shape-
selective catalysis and separation technology. Mass transport plays an essential role in
determining the ability o f reactants to interact with the catalytic sites in a zeolitic
structure and the rate at which products diffuse in the zeolite. Often there are multiple
diffusivities in a single system, for example, the diffusion of p-xylene in silicalite. This
system has two diffusion coefficients, one for diffusion in the straight channels of
silicalite, another for diffusion in the sinusoidal channels (Shen and Rees 1993). The
diffusion coefficients for these two processes differ by approximately an order of
magnitude. An apparatus that is able to measure diffusion coefficients that span several
orders of magnitude has been designed and constructed. The device is based on the
method of frequency response (FR) o f a closed system. This method is best described as
76
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the ultimate response to an input function over a spectrum o f frequencies (Ogunnaike and
Ray 1994).
This chapter describes the design, assembly, and automation o f a FR device using
a servomotor, a metal bellows, a rapid-response differential pressure transducer, standard
high-vacuum components, and a personal computer equipped with analog to digital dataacquisition hardware and Lab VIEW® software. The pressure data collected by this
instrument are fit to models previously developed by Yasuda to estimate the diffusion
coefficients of the n-hexane/silicalite and methanol/silicalite systems.
4.3
4.3 .1
Theoretical
General Model
The FR method measures the pressure response of a closed sorption system to a
periodic volume change. Using the model derived by Yasuda, the pressure response data
may be used to describe diffusion of an adsorbate in a zeolitic adsorbent (Yasuda 1982).
Consider a sorption system at equilibrium. If this system is perturbed by a
sinusoidal change in volume, Eqn. (4.5), the vapor pressure o f the adsorbate diffusing in
to and out o f the zeolite will reach a periodic steady state, Eqn. (4.6). In Eqns. (4.5) and
(4.6) the subscript e denotes the equilibrium value o f the variable, the lower case variable
preceding the transcendental function, either p or v, is the amplitude ratio, the Greek
letter co is the angular frequency o f the perturbation, and the Greek letter p is the phase
lag o f the pressure response to the variation in volume.
F = Ve( 1- vcos(cot))
Equation 4.5
P = P e(\ + pcos((ot + <p))
Equation 4.6
77
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
By combining the Fickian diffusion model, the mass balance, and the locally
linear adsorption isotherm o f the adsorbate/adsorbent system, Yasuda derived the inphase and out-of-phase equations for both a plane sheet and a sphere. For a plane sheet,
the in-phase and out-of-phase equations are as follows:
f
V
~\
cos <p- 1 = KSXc
Equation 4.7
sin (p = KSXs
Equation 4.8
I Pj
M
Kp j
where SXc and SXs are defined as:
1 sinh tjx + sin tjx
d\c = —
V, cosh rjx + cos rjx
Equation 4.9
1 sinh 7 , - sin rjx )
= — cosh 7 , + c o s 7 , j
fix
Equation 4.10
where the reduced angular frequency, //,, is based on the angular frequency, <u, the
thickness o f the plane sheet, Z,, and o f the diffusion coefficient D :
Equation 4.11
and K is related to the local slope o f the adsorption isotherm at Pe and Te :
K =
RT
V.
dB
dP.
Equation 4.12
where B is the amount o f adsorbate sorbed.
78
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A similar treatment for a sphere yields the same in-phase and out-of-phase
equations, Eqns. 4.7 and 4.8 respectively. However,Slc and Su are replaced withe>3c. and
S3s defined as:
.
J
_ 3 sinh rj3 - sin rj~
rj3 [ c o s h - c o s /^3
6 1 sinh rj3 + sin tj3
<>3, = —
73 2 [cosh//3 - cos t\3
Equation 4.13
11
73
Equation 4.14
J
Now //, becomes rj3 and is a based on the radius o f the sphere, a, in place of the
thickness of the plane sheet, L :
Equation 4.15
Figures 4-2 and 4-3 show the theoretical in-phase and out-of-phase functions for a slab
and sphere, respectively. In both figures, the frequency at which both the inflection point
o f the in-phase response and the maximum o f the out-of-phase response is defined as the
comer frequency. This frequency is the inverse o f the characteristic time o f the process.
The comer frequency for both a slab and a sphere are defined as follows:
J com er
/
j 2
/«,
D
for a sphere o f radius a
a'
Equation 4.16
for a slab of thickness L
Equation 4.17
79
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
Slab Geometry
D = 1x10‘11 m / s
a = 10(j.m
<
D
CO
CO
.C
In-phase
Out-of-phase
Q.
Oi
3
o
a)
CO
CD
.C
Q.
■
C
0
2
1
0
1
log (frequency)
Figure 4-2 - Theoretical in-phase and out-of-phase functions for a slab
80
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
Sphere G eom etry
D = 1x1 O'11 m2/s
a = 10nm_______
0)
CO
(0
.c
CL
In-phase
O ut-of-phase
3
0
<
D
VJ
(0
-C
a.
1
c
0
2
0
1
1
log (frequency)
Figure 4-3 - Theoretical in-phase and out-of-phase functions for a sphere
4.3.2
Multiple Diffusion Processes
The model derived by Yasuda is linear. Therefore, the ultimate response o f the
system will be the sum o f the individual responses present in the system. This result is
shown in Eqn. 4.18 and 4.19:
Equation 4.18
Equation 4.19
81
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Equation 4.20
For a process in which the local Henry’s Law constants are the same but the
characteristic time constants differ by an order o f magnitude, an ultimate response similar
to Figure 4-4 results.
2
<
u
C/5
CO
JZ
In -p h ase
Q.
+L
O
1
0)
C/5
CD
.C
CL
I
c
0
2
1
0
1
log (frequency)
Figure 4-4 - Theoretical in-phase and out-of-phase functions for a process in which the
local Henry’s Law constants are the same but the characteristic time constants differ by
an order o f magnitude
82
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.3.3
Surface Resistance to Diffusion
If a surface resistance is present, the equations above can be modified. This
modification, also developed by Yasuda (Yasuda, Suzuki et al. 1991), incorporates a rate
constant for the surface resistance in the in-phase and out-of-phase components as
described below:
Equation 4.21
P
G>9
II
a K -4
(ok
a >
1 -
k
<o
J
a K “
1
«
J
S
CO
Equation 4.22
0
where:
/ = 1 or 3
a «0.01 for commercial zeolites
c = 1 for commercial zeolites
0=
CO
k
A
- + C*u
(a + cSIC)2
J
is the surface resistance rate constant [s' ]
Figure 4-5 shows the theoretical in-phase and out-of-phase functions with the surface
resistance modification. This figure shows that the in-phase and out-of-phase functions
cross, if a surface resistance is present.
83
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
0)
l/J
In-phase
Out-of-phase
CO
-C
a.
oI.
3
O
a>
CO
CD
0
2
0
1
1
log(frequency)
Figure 4-5 - Theoretical in-phase and out-of-phase functions with the surface resistance.
In order to use these equations to estimate a diffusion coefficient, the pressure
response history must be measured and recorded. A frequency response device capable
of measuring and recording pressure data reliably over three decades of frequency, from
0.005Hz to 5Hz has been designed, built, and automated. A description o f the device and
a comparison to previous devices is presented in the following section.
84
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.4
4.4.1
Previous Devices
Yasuda’s Device
Three previous FR devices have been documented in the literature. Yasuke
Yasuda designed the first apparatus in 1976, which was derived from a standard
adsorption apparatus (Yasuda 1976a; Yasuda 1976b; Yasuda and Saeki 1978). Yasuda’s
first instrument used a Pirani gauge to measure the pressure in the system. This type of
gauge has a rather slow response time and is not useful for measurement of
multicomponent gases because it measures pressure based on thermal conductivity o f the
gas phase. In later versions o f Yasuda’s device, a capacitance manometer gauge replaced
the Pirani gauge. The capacitance manometer measures the pressure by sensing the net
force exerted on a diaphragm in the gauge. This gauge has a faster response (>10ms)
than the Pirani gauge and allows experiments using multicomponent gases. The volume
perturbation is performed using a bellows attached to a gearbox and variable speed
motor. The connection o f the crankshaft of the bellows to the gearbox produces a
sinusoidal volume variation. An event marker is placed to record the volume maxima;
this is used to measure the phase lag o f the resulting pressure response. Using this setup,
the device can measure frequencies from 0.001 Hz to 0.25Hz.
4.4.2
Rees’ Device
The second device built by Lovat C. Rees uses magnets to perform the volume
perturbation (Rees and Shen 1993). Thus Rees’ device is limited to step changes and
square wave frequency response. Rees’ magnets can produce a very close approximation
to a square wave; the rise time is less than 20ms. However, with a total rise time of 20ms
85
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the ideality o f the square wave worsens as the frequency exceeds I OHz. Considering this,
the device can measure frequencies from 0.01 Hz to 10Hz. Rees’ device also uses the
capacitance type o f pressure transducer.
4.4.3
Grenier’s Device
The third device built by Ph. Grenier measures the thermal frequency response to
sorption (Grenier, Bourdin et al. 1995; Bourdin, Grenier et al. 1996; Bourdin, Gray et al.
1998). The temperature measurement is extremely accurate with a standard error on the
order of 10"4 K. However, the temperature sensing apparatus is quite complex and
components are quite expensive. In addition to the temperature measurement, the
pressure is measured using a capacitance-type pressure transducer. The volume
perturbation is performed using a cam to drive a bellows. The cam may be o f any shape
and thus any input function can be used to drive the volume. One potential problem in
using a cam is that at high frequencies the bellows may not have a spring constant
sufficient to return to its equilibrium position before the cam meets the bellows again.
This is known as “floating the bellows.” The data provided in papers by this group
indicate that they achieve frequencies from 0.001 Hz to 30Hz without floating occurring.
4.4.4
General Improvement
This system can measure a range over three orders o f magnitude in frequencies,
from 0.005Hz to 5Hz. The upper limit o f this range is fixed by the time constant o f the
pressure transducer. This range of frequencies is comparable to the three previous
devices The system also makes use of the capacitance-type pressure transducer for
measurement o f the system pressure. The volume change is performed using a metal
86
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
bellows driven by a servomotor. This combination produces a very accurate sine wave
approximation because the servomotor has a feedback loop and a 1000-line indexer
allowing very accurate control o f the motion. Therefore, this system, like Yasuda’s, uses
a push-pull setup removing the possibility of floating. The ability to control the position
o f the servomotor allows the device to perform step changes with rise times of 50-80ms.
The total cost o f this system was approximately $30,000 including computer, data
acquisition, vacuum components, and pressure transducer. Although any device of this
nature is complex, this device was built with all off-the-shelf pieces save two, and those
were quite simple modifications.
Although the devices described above have not been seen first hand, it appears as
though the pressure measurement devices are located at a distance o f at least 30cm to
50cm from the sample. For a reasonable operating pressure o f 1 Torr, it would take
approximately 30ms for the pressure wave to reach the measurement device. Therefore,
at frequencies greater than 30Hz there would be an averaging o f the pressure data prior to
measurement by the transducer. The distance of the pressure transducer from the sample
in our system is only 8cm. This translates to a delay o f 7ms at an operating pressure o f 1
Torr. At these conditions, the system would have to operate at 145Hz before it would
experience this averaging o f the pressure signal.
4.5
4.5.1
Experimental Setup
Apparatus
Figure 4-6 presents a schematic of the FR device. The essential components of
the device are the servomotor, bellows pump, position sensor, sorption chamber,
differential pressure transducer, ballast volume, dosing manifold, vacuum system, and
87
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
heaters and insulation. As mentioned earlier, the servomotor is feedback-controlled and
has 1000-line indexer, that allows accurate control o f the position and speed o f the motor
shaft. A metal bellows pump is connected to the servomotor and is used to drive the
volume perturbation. A position sensor is placed on the shaft of the bellows pump to
index the position of minimum volume. The bellows is attached to the sorption chamber
where the sample is placed. Connected to the sorption chamber are the pressure
transducer and ballast, the dosing manifold, and the vacuum system. The pressure
transducer is a 10-Torr differential capacitance manometer. The measurement side o f the
transducer is attached to the sorption chamber and the reference side is connected to the
ballast volume. The ballast volume is placed in an isothermal water bath to prevent the
reference pressure from changing during an experiment. The dosing manifold is used to
introduce sorbaies into the sorption chamber. The vacuum system consists o f a roughing
pump and turbo pump. The entire system can be pumped down to <10^ Torr to aid in
degassing the sample. Finally, the sorption chamber is heated using band heaters and is
covered with an insulating blanket to reduce heat loss. This allows the sorption
experiments to run at temperatures up to 473K. A detailed discussion o f each component
is provided bellow. The letters in parentheses correspond to the component in Figure 4-6.
Specific information on manufacturer and model number can be found in Appendix B.
88
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
DO.
EE.
W
FF
EE
GG
HH
B8
Figure 4-6 - Schematic o f the frequency response device, specific information on
manufacturer and model number can be found in Appendix B
4.5.1.1
Servomotor
A servomotor (A) is coupled to a metal bellows pump (C) by a motor coupling
(B). The servomotor is capable o f speeds ranging from 0.001 Hz to ~ 40 Hz. The 40 Hz
limitation is imposed when the torque produced by the servomotor is no longer sufficient
to rotate the metal bellows pump. The motor coupling is used to relieve any mechanical
strain if the shafts o f the servomotor and the metal bellows pump are not perfectly
aligned. If the coupling were not present, the bearings on both the servomotor and the
metal bellows pump would fail rapidly.
4.5.1.2
Metal Bellows Pump
The metal bellows pump contained a reed valve assembly when purchased. This
assembly was removed since the bellows is used to change the volume o f the system and
89
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
not to pump gas. When removed the pump head, which held the assembly in place, was
not of the correct dimension to seal the system with a Teflon® encapsulated silicone Oring (D). In addition to the pump head being o f the wrong dimension, a half-nipple (F)
needed to be tungsten-inert-gas-welded (TIG) to the top o f the pump head in order to
mate the remaining high-vacuum components to the device. Thus, it was decided to
reproduce the pump head with two modifications to correct the problems encountered.
This piece is the pump head-replacement (E).
4.5.1.3
Position Sensor
The position sensor is made o f an optical switch (HH) and a chopper wheel (GG).
The chopper wheel is made from a piece o f 2mm thick aluminum plate and is 8cm in
diameter. A slot was cut into the wheel 1mm wide by 5mm deep. The wheel was
attached to the shaft o f the bellows pump. The optical switch is positioned such that the
wheel is between the emitter and the sensor. The wheel is positioned so that when the
bellows is fully compressed the notch cut into the wheel aligns with the emitter and
sensor of the switch. In this position, the sensor is triggered and the signal is transmitted
by the data acquisition system to be recorded by the computer.
4.5.1.4
Sorption Chamber
The sorption chamber consists of the metal bellows pump, the pump head-
replacement, the half-nipple, the conical reducer (Gi), the rapid response 10-Torr
differential pressure transducer (attached on the sample side) (M), the sample holder (I),
the reducing tee (J), the view port (K), and the high vacuum metal bellows valve (Li).
The high vacuum metal bellows valve is used to isolate the sorption chamber from the
90
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
rest of the device after the adsorbate is introduced. Once isolated, the total volume o f the
system is 584 cm3. All surfaces exposed to the adsorbate are 316-stainless steel with the
exception of the sample holder, which has been nickel-plated. The sample holder was
made from a piece o f schedule-80 2-inch (5.08cm) diameter mild steel pipe. Mild steel
was chosen over stainless steel because o f its ease in drilling and tapping holes o f small
dimension. Oxidation of the mild steel pipe would certainly occur so the piece was
nickel-plated.
4.5.1.5
Pressure Transducers and Ballast Volume
The reference side o f the 10 Torr differential pressure transducer is connected a
150cc ballast volume (H) by a double-sided high vacuum flange (N) and a short length of
0.25-inch (0.64cm) stainless steel tubing. This ballast volume is connected to the dosing
side of the device by a combination o f two rotatable blank high vacuum flanges (Pi, P 2 )
connected by a length of 0.25-inch (0.64cm) stainless steel tubing (Q). The ballast
volume is isolated from the dosing side of the device by a second high vacuum metal
bellows valve (L 2 ). The ballast volume is submerged in a constant temperature bath (O).
The pressure on the reference side o f the pressure transducer will not vary greatly with
changes in room temperature because the temperature-controlled ballast volume contains
more than 90% o f the total volume on the reference side of the pressure transducer.
The operator can adjust the pressure on the reference side o f the pressure
transducer. By maintaining a differential pressure of ±0.0004 Torr, the range o f
operating pressures is increased from 0.3-10 Torr for an absolute pressure transducer to
0.3-300 Torr for the differential model. This design provides a 30-fold increase in
operating pressure without loss o f precision (±0.0004 Torr) and provides a great cost
91
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
savings since each o f the rapid response differential pressure transducers costs in excess
o f $3500.
A dosing manifold is attached to the sorption chamber and ballast volume either
to admit an adsorbate or to adjust the pressure of the ballast volume. The main body of
the dosing manifold is made from a four-way cross (U) and a three-way tee (Ri).
Attached to the dosing manifold are a leak/shut-off valve (CC), a high vacuum metal
bellows valve
(L 3 ),
a dosing manifold pressure transducer (T) and the associated
connecting hardware (Si, Vj, V 2 , V 3 , W, X).
4.5 .1.6
Dosing Manifold
An inlet manifold (R2, V4) is connected to the dosing manifold, which is capable
o f introducing an adsorbate through either a gaseous sample port (L4, DD) or a liquid
sample port
(L 5 ,
S2). Once the adsorbate is admitted into the inlet manifold, by opening
either high vacuum metal bellows valve (L4,
L 5 ),
the adsorbate is leaked into the dosing
manifold through the leak/shut-off valve.
4.5.1.7
Vacuum System
Also connected to the dosing manifold, by a conical reducer (G2) and reducing
coupling (Y), is a high vacuum pumping system consisting o f a turbo molecular drag
pump (Z), connected to a rotary-vane roughing pump (BB) by a flexible metal hose (AA).
This vacuum pumping system allows the entire FR device to be evacuated to below 10-6
Torr. At this pressure, the sample held in the sample holder could be outgassed to
remove any adsorbates that were adsorbed by contact with atmosphere after calcinations
or after a previous experiment.
92
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.5.1.8 Heating and Insulation
Mica band heaters (EEi, EE 2 , and FF) are mated to the two 4.5-inch (1 1.4cm)
flanges of the sorption chamber and to the 2.75-inch (7cm) flange o f the half-nipple.
This allows the sample to be heated to a maximum temperature of473K ; the melting
temperature o f the Teflon-encapsulated O-ring imposes this limitation.
Insulation is required to minimize temperature excursions caused by the
constantly changing laboratory temperature. The insulation is made like a down
comforter for a bed. The outer cover is made o f Kevlar® cloth. The packing material is
8pm boro-silicate glass fiber. The stitching is NOMEX® fiber. These materials are all
capable of tolerating well in excess o f the maximum system temperature (473K).
4.5.2
Experimental Materials
The adsorbent used in these experiments was a siiicalite sample prepared in
Professor Michael Tsapatsis’ laboratory at the University of Massachusetts. The siiicalite
crystal size was confirmed as 50x10x10 pm by electron microscopy (Figure 4-7). The
siiicalite was calcined at 773K for 24 hours in dry air. The adsorbate used was n-hexane.
The reagent was purchased from Aldrich with a purity of greater than 95%.
93
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Figure 4-7 - Electron micrograph o f the siiicalite sample used in these experiments, the
crystal size is 50x10x10 pm (Photo courtesy o f Prof. Tsapatsis’ research group)
4.5.3
Procedure
The choice of crystal size is of importance because the estimated comer frequency
(Eqn. 4.16 or 4.17) must be within the measurable frequency range o f the device. With
an estimate of the diffusion coefficient, Figure 4-8 can be used to determine the
appropriate crystal size. For example, a crystal size o f 10pm would be suitable for a
diffusion coefficient in the range o f 10'12 - 10'9 m2 s'1.
The amount of zeolite to be studied must be chosen carefully. If too much zeolite
is used, the amplitude ratio o f the pressure response may not be sufficiently large to be
measured by the pressure transducer. If the amount o f zeolite is too small, the amplitude
ratio may not be distinguishable from the blank experiment. The optimum amount of
zeolite is determined by a simple mass balance over the amount o f adsorbate in the
sorption chamber gas phase and adsorbed in the zeolite. For the closed system o f the
sorption chamber gas phase and the zeolite, the total number of moles in the system is
constant, Eqn 4.23.
94
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
f = 5 Hz
= 0.005 Hz
0.1
1
10
100
1000
Particle size [|im]
Figure 4-8 - Using an estimation of the diffusion coefficient, the particle size is
determined such that the intersection o f a horizontal line from the y-axis intersects a
vertical line from the x-axis in the region between the 0.005 Hz and 5 Hz boundaries
nr ~ nsas + nads = Cflst
Equation 4.23
where n is the number of moles and subscript T, gas, and ads denotes total, gas phase and
adsorbed, respectively. At low operating pressures, the ideal gas law is used to
approximate the number of moles of gas in the sorption chamber, Eqn. 4.24.
gas
=
PV
RT
Equation 4.24
where R is the universal gas constant, P is the pressure, V is the volume, and T is the
temperature o f the sorption chamber. The number o f moles adsorbed in the zeolite is
95
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
approximated by the use o f the Henry’s Law constant as determined from the adsorption
isotherm, Eqn. 4.25.
\njJs=mHP
Equation 4.25
where m is the mass of the zeolite, H is the Henry’s Law constant.
The pressure of the sorption chamber at very high frequency approaches the
pressure of the blank experiment, because the characteristic time for diffusion is much
larger than the frequency o f the volume perturbation. Therefore, the pressure change in
the sorption chamber is attributed to the volume change o f the metal bellows.
PBiunt = pe0 ± v)
Equation 4.26
For frequencies in the range o f the diffusion time constant, the pressure is calculated as
follows:
p E x p e r im e n t —
f mH + ---v \
RT)
Equation 4.27
1
Figure 4-9 shows a plot o f [PExpmmnt - p J and [pBlank - P&pcnm,„J versus the mass
o f the siiicalite for the n-hexane/silicalite system. As the mass o f the zeolite sample
increases the difference between the pressure at high frequency and the equilibrium
pressure approaches zero. This means that if too much zeolite is used, the experimental
pressure will not be distinguishable from the mean pressure. For small amounts o f
zeolite, the difference between the pressure at high frequencies and the pressure at
frequencies comparable to the diffiisional time constant approaches zero. Therefore, if
too little zeolite is used, the signal will be indistinguishable from the blank experiment.
96
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The optimal sample mass is that mass where the two curves o f Figure 4-9 intersect, in
these experiments about 90mg. At this mass both pressure differentials are equidistant
from their maximum or minimum values. This determines the mass o f the sample to be
used in the experiments.
0.014
0.012
o
-
0 .0 1 0 -
a)
c 0.008 ca
.c
Blank
Experiment
Experiment
g 0.006 13
C
fl
0.004 CL
0.002
-
0.000
0.00
0.05
0.15
0.10
0.20
0.25
0.30
Mass (g)
Figure 4-9 - [p ^ ^ , - Pe\ and [pghrdc - P ^ ^ , ] versus the mass o f the siiicalite for the
n-hexane/silicalite system. The point at which they intersect is the optimal sample mass.
After determining the sample mass, the calcined zeolite - in these experiments the
zeolite is siiicalite - is loaded into the sorption chamber by removing the view port (K)
and removing the sample holder (I). The siiicalite is placed in glass wool to reduce any
mass transfer limitations and to aid heat transfer. The glass wool is 8pm borosilicate
glass fiber purchased from Wale Apparatus Company in Hellertown, PA. The siiicalite
and glass wool are placed in the holder. The holder is returned to sorption chamber and
97
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
the view port reinstalled. Valves Lj, L2 , L3 , and CC are opened to evacuate the device
and degas the sample. The mica band heaters (EEi, EE 2 , and FF) may be used to heat the
sample to accelerate the degassing process. After degassing is complete, valves Li, L2 ,
L3, and CC are closed.
Prior to admitting the adsorbate, the metal bellows is brought into the bottomdead-center position (maximum volume). Once the bellows is positioned, the adsorbate
is admitted into the inlet manifold through either the liquid (S 2 ) or gas (DD) sample port.
The adsorbate is leaked through valve (CC) into the dosing manifold until the desired
pressure is attained as measured by the dosing manifold pressure transducer (T). Valve
Li is opened allowing the adsorbate to enter the sorption chamber. Valve Li is closed
and the pressure in the sorption chamber is allowed to equilibrate.
After equilibrium is achieved, if the sorption chamber pressure is less than 10
Torr, then valve L2 and L3 are opened so the ballast volume (H) is evacuated and the
differential pressure transducer acts as an absolute pressure transducer. However, if the
sorption chamber pressure is above 10 Torr, then valve L2 is opened and the adsorbate is
allowed to leak into the ballast volume until the pressure transducer reads a differential
pressure of 0.0000 Torr. When this pressure is attained valve L2 is closed. Since the
metal bellows is in the bottom-dead-center position, the deviation from zero differential
pressure is always positive. This allows for a maximum operating pressure o f 300 Torr in
the differential mode.
The Lab VIEW® program developed in our laboratory is used to vary the speed of
the servomotor and to log the pressure response of the differential pressure transducer.
98
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
These data are then fitted to the model previously described and a diffusion coefficient is
determined as described below.
4.6
Results
4.6.1
Blank Experiments
Blank experiments were performed to reference any apparent changes in p and
<o caused by delays in the responses of the pressure transducer and o f the data acquisition
system. The blank experiments were carried out at temperatures and pressures similar to
the temperatures and pressures o f the actual experiments but in the absence o f the
sorbate. The amplitude ratio o f the volume variation, v from Eqn. (4.5), is equal to the
amplitude ratio of the pressure response of the blank experiments, p b.
v = Pb
Equation 4.28
The actual value o f p h is constant at all frequencies for the blank experiments, since the
volume variation is a constant ±1.5 % due to the mechanical volume change o f the metal
bellows. However, if the response time o f the pressure transducer or the software is not
sufficiently rapid, then the measured value o f p h will be attenuated, most notably at
higher frequencies. Figure 4-10 shows the amplitude ratio for the blank experiments as
well as the amplitude ratios for both the n-hexane and methanol experiments. Once
measured, this attenuation can be easily removed from the pressure response o f the actual
experiment. The phase lag associated with delays in the equipment is obtained by fitting
the pressure response data o f the blank experiment to Eqn. 4.6. Figure 4-11 shows
typical phase lag data measured during a blank experiment as well as the n-hexane and
99
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
methanol experiments. This blank phase lag, <pb, will be subtracted from the actual
experiments phase lag, <p, yielding the corrected phase lag o f the sorption experiment,
A<p, as shown in Eqn. 4.29.
A(p = (p-(pb
Equation 4.29
0.020
0.018
_ 0.016
Q.
o 0.014
CD
CD
T3
0.012
i - 0.010
<
p (n-hexane)
p (blank)
p (methanol)
0.008
0.006
0.004
•2
1
0
1
log (frequency)
Figure 4-10 - Amplitude ratio for the blank, n-hexane, and methanol experiments
100
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
6.2
-e-
cn
(0
0
)
<t>(n-hexane)
4>(blank)
<t>(methanol)
CO
(0
.c
CL
5.2
•2
1
0
1
log (frequency)
Figure 4-11 - Phase lag for the blank, n-hexane, and methanol experiments
4.6.2
Sorption Experiments
Sorption experiments using n-hexane and methanol sorbed in siiicalite were
performed. The system conditions for the two systems are presented in Table 4-1.
Table 4-1 - Operating conditions for the n-hexane/silicahte and methanol/silicalite
systems
Temperature
Pressure
n-hexane/silicalite
325 K
0.9 Torr
methanol/silicalite
308K
0.8 Torr
The in-phase and out-of-phase data as well as the fits using the two-diffusion-coefficient
model with a surface barrier are plotted in Figures 4-12 and 4-13 for the
101
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
methanol/silicalite and n-hexane/silicalite systems. Both figures show the following two
characteristics:
1
.
Two inflection points in the in-phase function and two maxima in the outof-phase function.
2.
The in-phase and out-of-phase functions cross.
o
*
In-phase data
Out-of-phase data
In-phase fit for D, w/ surface barrier
Out-of-phase fit for D, w/ surface barrier
In-phase fit for D2 w/ surface barrier
Out-of-phase fit for D2 w/ surface barrier
In-phase fit for
w/ surface barrier
Out-of-phase fit for D,+D2 w / surface barrier
1.0
0.8
03
<
/)
CO
.C
Q. 0.6
O
O
0.4
<
u
on
CD
-C 0.2
Q■
--
c
0.0
-2
-1
log (frequency)
Figure 4 -1 2 - In-phase and out-of-phase data and fit using the two-diffusion-coefficient
model with a surface barrier for methanol/silicalite
102
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
°
*
In-phase data
Out-of-phase data
In-phase fit for D, w/ surface barrier
Out-of-phase fit for D, w/ surface barrier
In-phase fit for D2 w/ surface barrier
Out-of-phase fit for D2 w/ surface barrier
In-phase fit for D,+D2 w/ surface barrier
Out-of-phase fit for D,+D2 w/ surface barrier
O
0-3 -
-2
-1
0
log (frequency)
Figure 4 - 1 3 - In-phase and out-of-phase data and fit using the two-diffusion-coefficient
model with a surface barrier for n-hexane/silicalite
The first characteristic is representative o f there being two distinct diffusional
time constants. The explanation for these two diffusional processes is that siiicalite
contains both straight and sinusoidal pores. The straight pores are elliptical in shape with
a dimension o f 5.3 A x 5.6 A, and the sinusoidal pores are elliptical with a dimension o f
5.1A x 5.5A (Meier, Olson et al. 1992; Cook and Conner 1999). The kinetic diameters o f
n-hexane and methanol are 4.3 A (Breck 1974) and 3.85 A (Harrison, Leach et al. 1984),
respectively, and can therefore diffuse in either pore type. Thus, the two characteristic
time constants measured by the device are the diffusion o f the adsorbates in both the
straight and sinusoidal pores o f siiicalite.
103
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The second characteristic is suggestive o f a surface resistance to diffusion. Any
one or combination of the following can cause the surface resistance to diffusion:
•
Conformational barrier - In order to diffuse into the crystal the molecule
may have to undergo a change in conformation. In the case o f n-hexane, the
molecule would have to change from its more stable coiled conformation to
a less favored linear conformation (Biilow, Struve et al. 1980).
•
Thermal barrier - For systems with heats o f adsorption, if the heat is not
dissipated rapidly, the local temperature could increase. This would cause
the less to be sorbed until the local temperature was returned to ambient
(Biilow and Struve 1984).
•
Nonequilibrium barrier - Adsorption at a crystal surface may not be
sufficiently rapid to be considered an equilibrium process. Thus, the
kinetics of the adsorption process may introduce a surface barrier (Micke,
Biilow et al. 1994).
The analysis of Figures 4-12 and 4-13 consists o f determining the parameters of
the model presented that most accurately describe the recorded data. In these
experiments, the slab model was used to determine the system parameters. After
combining Eqns. 4.7-4.12 and 4.18-4.22 with the available physical data, there remained
five parameters that needed to be determined to correctly explain the data; these were £>/,
Dj, AT,, K z, and
k _ai .
Using a spreadsheet program, the parameters were adjusted by trial
and error until the model equations best described the experimental results. Using this
procedure, the following results were obtained:
104
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table 4-2 - Parameters used in the fitting of the data presented in Figure 4-12 and Figure
4-13
System
T (K)
325
n-hexane/silicalite
methanol/silicalite
308
In this notation, the subscript
P (Torr)
D , [m V ]
D2 [mV]
0.9
FC,
0.25
[s'1]
0.40
300
lxlO*10
3xl 0 *12
0 .8
0.62 0 . 2 1
1000
2 x l 0 ‘ 10
7x 10"12
refers to the straight pores and 2 to the sinusoidal pores.
These results were compared to the frequency response results of van den Begin
and Rees (Begin and Rees 1989) for the n-hexane/silicalite system and the sorption
uptake results of Nayak (Nayak and Moffat 1988) for the methanol/silicalite system. The
diffiisivities for the n-hexane/silicalite system presented in Table 4-2 were in good
agreement with the value o f 2 x 10' 10 m2 -s'‘ at 300K as reported by van den Begin in 1989.
Although van den Begin and Rees did not encounter a surface resistance, the kinetic
constant for the resistance was measured with little or no detriment to the diffusion
coefficient measurement.
For the methanol/silicalite system, Nayak used a sorption uptake method for
determining the diffiisivity. Ruthven (Ruthven 1995) compared various techniques used
to measure diffusion coefficients and found that frequency response generally yields
more rapid diffusion estimates than uptake experiments, in some cases by four orders o f
magnitude. For this reason, it was not unexpected that the results were two orders of
magnitude faster than Nayak’s report o f 8.28x 10' 14 m2 -s‘' at 308 K. Although this might
appear to be a large discrepancy, taking into account that different techniques can yield
widely varied results, and that the trend agrees with that reported by Ruthven, there is
confidence that these results are accurate.
105
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4.7
Conclusions
A new frequency response device has been design and built based on previous
efforts by Yasuda, Rees, and Grenier. This design uses the best attributes o f the three
previous devices and adds one overall improvement.
1. The device makes use o f the fast response capacitance manometer.
2. The device uses a servomotor to push-and-pull a metal bellows pump, which
drives the sinusoidal input function, removing the chance o f “floating” the
bellows.
3. The overall cost o f the device was kept below $30,000.
4. The pressure transducer is mounted only 8 cm from the sample, instead o f
~30cm, thus, removing any averaging o f the pressure signal above 30Hz.
The device is capable o f measuring frequencies between 0.005Hz and 5Hz, a
range of three full orders o f magnitude. The system can operate at temperatures between
room temperature and 473 K and pressures between 0.3 Torr and 300 Torr.
The apparatus has shown that it has the ability to distinguish at least two different
rate processes occurring simultaneously, the diffusion of n-hexane in the straight and
sinusoidal pores of silicalite. The diffusivities measured using this device are comparable
to those measured by Rees using the FR method in 1997. In addition to measuring the
diffusion rate processes, the kinetic parameters associated with a surface resistance to
diffusion were also measured.
106
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 5
CONCLUSIONS
The essential conclusions o f this dissertation are summarized in this section.
•
Based on the interactions o f microwave energy with dry zeolites, a temperature
gradient exists in the zeolite crystal. The measured temperatures in the system are
that the temperature of the bulk zeolite is greater than the temperature o f the gas
phase passing over the zeolite. Therefore, the surface siianols must be absorbing
energy at a rate that is greater than the rate at which the energy is being lost due to
conduction to the solid phase and convection to the gas phase.
•
The dielectric loss constant o f an adsorbate sorbed in a zeolite is a good indicator
o f how strongly microwave energy will interact with the adsorbate. For example,
a zeolite sample adsorbed with cyclohexane, a low dielectric substance, was
exposed to microwave radiation. The temperature rise comparable to the
microwave heating o f the zeolite with no adsorbate present. However, when
methanol, a high dielectric substance, adsorbed on a sample o f zeolite and
exposed to microwave energy, the temperature increase was as much as
2 0
K
greater the temperature rise o f the zeolite with no adsorbate present.
•
Microwave energy was found to cause desorption o f cyclohexane at bulk
temperatures lower than normally expected for cyclohexane. Although the bulk
solid temperature is lower than expected, the temperature o f the surface can be at
107
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
or above the normal desorption temperature. Thus, desorption occurs without
heating the system to the temperature required if thermal equilibrium were
achieved.
•
Microwave energy was found to change the adsorption selectivity for the
methanol/cyclohexane system adsorbed on silicalite or DAY zeolite. This result
was independent of the manner in which the adsorbates were sorbed.
•
Microwave energy can shift the CO bulk lightoff temperature 9K lower compared
to the lightoff temperature using only conventional heating. Although the bulk
lightoff temperature may be lower using microwave energy, the effective surface
temperature using microwave energy may still be greater than the conventional
bulk lightoff temperature.
•
Microwave energy can shift the lightoff temperature o f the bulk catalyst poisoned
by SO 2 lower than the lightoff temperature o f the unpoisoned catalyst using only
conventional heating. In addition, microwave energy can increase the overall
conversion o f CO from 85% at 5 13K using conventional heating, to 100% at
500K using microwave energy. The reason is that the sulfides and sulfates are the
most polarizable bonds in the system and are most likely to absorb microwave
energy. This will cause these compounds to decompose or desorb. This frees the
catalytic sites to carryout the CO oxidation reaction, restoring conversion.
•
Using the best attributes o f three previous frequency response devices designed
by Yasuda, Rees, and Grenier, a new frequency response device has been
108
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
designed and constructed with the addition o f one overall improvement. The
instrument is able to measure characteristic frequencies from 0.005Hz to 5Hz. In
addition, this research has shown the instrument can measure two characteristic
time scales occurring simultaneously.
109
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
CHAPTER 6
RECOMMENDATIONS FOR FUTURE RESEARCH
For as many questions as this dissertation has answered, it has posed many new
ones. This final chapter discusses the future directions recommended for this work.
The cyclohexane/methanol competitive adsorption system was used because there
was substantial data available on adsorption in zeolites for each compound. The
drawback to this system is that the system has little industrial applicability. Thus, it
would be interesting to find an industrially important system in which microwave energy
could be beneficial. In order to find this more suitable system, greater interaction with
industry is needed. In a short time, this interaction could yield a list several possible
systems to study. From this list, the use o f the network analyzer and the Hewlett Packard
probe could be used to narrow this list. For the systems remaining on the list, a literature
search would find those systems for which the most data are available. Missing data,
such as heats of adsorption or diffusivities, could be determined using the Simultaneous
Thermal Analyzer (STA) and the frequency response sorption apparatus.
A variable frequency microwave source should be purchased. The possible
number o f systems that could be studied increases with the combination o f this source
and the o f the network analyzer. The analyzer can be used to measure the microwave
spectrum o f the substances o f interest. Using these data, the appropriate frequency can be
determined to produce the most benefit.
The question o f the effective surface temperature needs to be addressed. The
experiments performed using the BET (Brunauer, Emmett, and Teller) system to measure
110
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
adsorption isotherms in the presence and absence of microwave energy should be
continued. A more accurate measurement of the actual surface temperature can be
determined by considering where the isotherm measured with microwave energy lies
compared to the isotherms measured without microwave energy.
Microwave energy could provide significant improvement in a reaction system
where the reactants are susceptible to microwave energy, but the product has a low
absorption o f microwave energy. Thus, the reactants can interact with the microwave
energy and promote the reaction whereas the product will not interact with the
microwave energy and thus not undergo any further reaction to unwanted byproducts.
In the area of zeolite synthesis, further collaboration with Michael Tsapatsis’
group would be interesting. A combination o f the tools and knowledge gathered by our
group in the area o f microwave energy, in conjunction with the vast knowledge o f zeolite
synthesis in their group could yield impressive results.
As discussed briefly in chapter 3, it was observed that SO 2 caused a lowering of
the hydrocarbon lightoff temperature for the Engelhard catalyst while all other literature
observes an increase in lightoff temperature in the presence o f SO 2 . A more detailed
study as to the mechanism for this reaction would be o f interest.
Finally, with most o f the bugs worked out of the frequency response system,
diffusion data should be able to be gathered very rapidly. At this time, there is no
obvious barrier to measurement o f diffusion coefficients for any number o f systems as
long as the characteristic time constant is within the range o f the device. An addition to
the device that could extend the range of the frequency span is a temperature control
system that could maintain the system temperature within 0. IK.
ill
Reproduced with permission o f the copyright owner. Further reproduction prohibited without permission.
APPENDIX A
MICROWAVE EFFICIENCY CALCULATIONS
Experimental work is not always successful, even when performing a simple
energy balance. The difficulty arises in the measurement o f system temperatures and
flow rates. This appendix discusses the efforts to complete the energy balance on a onemeter long column filled with silicalite, which was exposed to microwave energy. This
summary is written to provide future researchers a description o f what steps have been
taken, what problems exist, and what the possible solutions to these problems might be in
order to close the energy balance o f a microwave system.
A.
1 Apparatus and Instrumentation
The system vessel is a one-meter long stainless steel tube with an inside diameter
o f 7.62cm (Sin). This diameter was chosen because it is the correct dimension for the
propagation of microwave energy at 2.45GHz (S-band). At the top and bottom o f this
tube are S-band circular to rectangular waveguide transitions. Attached to the top
transition is a waveguide to coaxial cable adapter. This adapter allows the system to be
connected to the 300W continuous microwave power source with a coaxial cable.
Connected to the bottom transition is a lkW water load, which is used to absorb any
excess microwave energy that is not absorbed by the system. The vessel has a gas inlet at
the top of the tube and an exit at the base of the tube to allow either a purge gas or an
adsorbate laden carrier gas to pass over the sample contained inside the tube.
112
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
The instrumentation attached to the vessel consists o f five temperaturemeasurement ports equally spaced from the top to the bottom o f the column. Fiber optic
temperature probes are fed through these ports inside of glass capillary tubes, which are
sealed on one end. The advantage of using the capillary tubes as sleeves for the probes is
that the probes can be moved inside the sleeves to measure the radial temperature profile
of the bed as the experiment is running. The fiber optic probes are used because metal
thermocouples could cause electrical arcing when exposed to the microwave energy. In
addition to these five temperature-ports, there are temperature probes that measure the
inlet and outlet gas temperatures as well as the inlet and outlet temperatures o f the water
flowing through the water load. The water load temperatures are measured using signal
amplifiers manufactured by Omega® Corporation. These amplifiers boost the milli-volt
signal produced by the thermocouple to a 0 - 1 V output, which is easily measured by the
data acquisition software.
The gas flow rate is measured and controlled by a calibrated Datametrics® mass
flow controller. The water is gravity feed from a reservoir through the water load into a
catch basin. The reservoir is continually overflowing and thus a constant pressure head is
maintained. This constant pressure head ensures that the flow rate is constant throughout
the experiment. The flow rate is measured by flowing a specific volume o f liquid over a
measured period o f time.
In addition to temperature and flow rate measurements, the microwave power
source provides forward and reflected power measurements. There is no specific
information on how these power is measured so the values need to be accepted with some
uncertainty, as will be seen later in this appendix.
113
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
A. 2
Experimental Results
A number o f experiments have been performed in an effort to close the energy
balance on the one-meter column exposed to microwave energy. This information is of
great importance for the calculation o f how efficient microwave energy is in desorbing an
adsorbate from an adsorbent.
The first step in closing the energy balance for this system was determining how
much energy was actually entering the system. There was no way to verify how the
forward and reflected power was measured by the microwave power source and,
therefore, needed an independent measure o f these values. In order to make this
measurement, the waveguide to coaxial cable adapter were connected to the water load
using a short (3 1cm) length of S-band rectangular waveguide. Therefore, with nothing to
interfere with or absorb the microwave energy, the total amount of microwave energy
absorbed by the water load should be the net microwave power experienced by the
system. The net forward power is defined as the forward power minus the reflected
power as measured by the microwave power source. The power absorbed by the water is
calculated by multiplying the water mass flow rate by the change in temperature o f the
water by the heat capacity of water. The results are presented in Table A - l .
114
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Table A-l - Experimental data for the calibration of the input microwave power
Power
Setting (W)
Forward
Power (W)
Reflected
Power (W)
Net Power
(W)
50
75
50.75
75.68
100.65
125.77
150.71
175.69
200.78
225.72
250.70
275.80
300.77
0.76
50.0
74.6
99.3
124.1
148.8
173.5
198.1
222.5
246.4
270.4
294.6
100
125
150
175
200
225
250
275
300
1 .1 2
1.38
1.72
1.96
2 .2 0
2.67
3.22
4.31
5.36
6.13
Water Power
(W) mCpAT
36.5
57.0
72.6
88.5
Ratio
0.73
0.77
0.73
0.71
0.67
1 0 0 .1
118.6
136.5
155.8
179.5
191.9
209.1
0 .6 8
0.69
0.70
0.73
0.71
0.71
The data presented in Table A-l indicate that, on average, only 72% o f the net
power measured by the microwave power source actually reached the water load. Where
was the other 28% going? The answer lay in the coaxial cable that connects the
microwave power source to the waveguide to coaxial cable adapter. Using a portable
network analyzer from Prof. Yngvesson’s research group in the Electrical and Computer
Engineering Department, the cable was tested and found to have an attenuation of
approximately 1 decibel (dB) at 2.45GHz. A decibel is defined in equation A. 1.
1
attenuation = 1 0 log ( P\ dB
U
Equation A. 1
j
Thus for a ldB attenuation, approximately 74% of the power was transmitted with
the remaining 26% o f the power lost in the cable as heat. This attenuation occured in
both the forward and reverse directions. Figure A -1 presents the net power and the water
power data of Table A-l with the addition o f the corrected net power. The corrected net
power takes into account the attenuation o f both the forward and reflected power by the
coaxial cable and is defined as follows:
115
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
corr.net power = (0.74* forw ard pow er) —{[ .26* reflected pow er)
Equation A.2
The data in Figure A -1 shows that the corrected net power had an average loss o f
approximately 25%, which was a close match for the losses o f 28% observed in Table A1. Therefore, although the method by which the microwave power source measured the
forward and reflected power was unknown the values presented on the front panel o f the
microwave power source appeared to be accurate when corrected for by IdB attenuation
in the coaxial transmission cable.
350
o
a
*
300
g
Forward Pow er - Reflected Pow er (Net Power)
W ater Pow er
Corrected Net Pow er
250
CD
I 200
Q_
£
150
§
100
uU
>
to
50
°
-\
ft
°
o
*
0
50
100
150
200
250
300
350
Forward Power Setting (W)
Figure A -1 - Net power, power absorbed by the water load, and corrected net power
versus the power setting for the calibration o f the input microwave power experiments.
The corrected net power and power absorbed by the water load are in good agreement.
116
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
With confidence in the input power measurements, the waveguide to coaxial cable
adapter and water load were mated to the waveguide transitions. The final set of
experiments attempted to close the energy balance with the column loaded with 293Og o f
silicalite. The most important step that must be followed was to remove any adsorbates
from the silicalite; this included water that adsorbed from the ambient air. It was not a
simple task to dry almost 3kg o f silicalite in a tube furnace; therefore, the silicalite was
dried in the column. In order to assure that the carrier gas contained no water, the vent
nitrogen from a liquid nitrogen tank was used as the carrier gas source. Because this gas
is the vapor off boiling liquid nitrogen the gas must be absolutely dry. A purge of
2000cm 3 -min' 1 was used and the silicalite heated to 383K for two days and remained
under a purge at all times.
The experiment used to determine the energy balance was quite simple. Three o f
the five probes in the bed were recorded, one at the top position, one at the middle
position, and one at the bottom position. All three probes were placed at the midline of
the column. Only three probes were used because the analog to digital board only had
three empty channels on which to record the data. The microwave energy was set at
150W of forward power as measured on the front panel o f the microwave power source.
The microwave power was continued until the system reached a steady state temperature
profile. At steady state the amount of power absorbed by the silicalite was equal to the
amount o f power being lost to the carrier gas and to the room by conduction and
convection.
To calculate the amount o f power being absorbed by the silicalite at steady state,
the microwave energy was turned off and the temperature profile was recorded. The
117
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
cooling curve was fit using a polynomial and the derivative taken with respect to time
(dT/dt). Then the amount of power absorbed by the silicalite was the slope o f the cooling
curve versus time (dT/dt) times the mass of silicalite times its heat capacity. The amount
o f power transferred to the carrier gas was the product of the mass flow rate o f the
nitrogen, the change in temperature, and its heat capacity. Table A-2 shows the
individual components o f the energy balance and the physical property data used in the
calculations. From this table, the amount of power absorbed by the components o f the
system is greater than the corrected net input power. This is a thermodynamic
impossibility; therefore, an error in measuring the system variables has been made.
Table A-2 - Physical property data and individual components of the energy balance
mass or m
Carrier gas (N2)
Water load
Silicalite
Total Power Absorbed
0.04 g-s ' 1
9.2 g-s' 1
2930 g
1.03 J-g'-K ' 1
4.184 J-g'-K ' 1
0.82 J-g '-K ' 1
dT
A T or —
dt
44 K
0.42 K
0.04 K-s' 1
Power (W)
1.8 W
16.1 W
96.1 W
II4W
111 W
52 W
59 W
Forward Power * 0.74
R eflected Power * 1.26
Corrected Net Power
If the assumption is made that the corrected net power was accurate, then there
must be an error in the measurement o f at least one o f the system variables. It is believed
that the temperature measurement o f the silicalite contained the greatest error. Previous
work done on this column when it was first constructed showed that there was a
significant radial temperature profile (Figure A-2). This profile exists because, for this
size waveguide (7.62cm) and frequency o f microwave energy (2.45GHz), there are three
118
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
modes that can propagate, meaning that they have a cutoff frequency ( f c) lower than
2.45GHz. These modes are the TE u { f c = 1.459GHz), the TM0i ( f c= 1.906GHz), and
the TE 21 ( / c = 2.421GHz). The superposition o f these three modes causes a spatial non­
uniformity in the microwave electric field leading to a unique temperature profile.
Therefore, a rigorous mapping o f the transient temperature profile during cooling in both
the axial and radial directions is needed to close the energy balance.
360
350
340
-
330
CJ)
co 3 2 0
La>
Q_
E 310
o
I-
c
300
290
280
1
0
1
Adimensional Radius
Figure A-2 - Radial temperature profile o f silicalite exposed to 100W o f microwave
energy in the one-meter column for several at several different experimental times
In order to perform these experiments, a significant amount o f new equipment
will need to be purchased. First, several temperature probes and another probe reader are
119
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
needed to measure this large number o f temperatures. Second, more analog to digital
channels are needed to monitor and record this growing number o f system variables.
Third, instead o f using the glass capillaries as sleeves, several probes can be inserted
through each port at different radial positions. By making these three changes a more
accurate estimate can be made o f how much energy is being absorbed by the silicalite.
If this temperature mapping fails to provide closure to the energy balance, the
corrected net power measurement would then be the next value to be questioned. In
order to make an accurate measure o f the forward and reflected power from the source, a
high power directional coupler would need to be purchased. A directional coupler has
the ability to split a certain percentage o f power off o f the main transmission line. By
routing this portion o f the microwave power through a known attenuator, one o f the
power meters currently in use in Prof. Yngvesson’s research group would yield a more
accurate measurement of the forward and reflected power levels.
If both o f the previous methods fail to resolve the energy balance, a different type
or size o f load would need to be tried. The current water load is rated at lkW, which is
three-times too large for the current microwave power source. Thus, slow flow rates are
required to attain measurable temperature differences between the inlet and outlet.
Although, these problems exist in the current system, the change would be useful in finetuning the energy balance and not remedying the larger problem that currently exists.
The implementation o f this information in combination with the new equipment closure
to the energy balance can be made rather rapidly.
120
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
APPENDIX B
PARTS LIST FOR FREQUENCY RESPONSE APPARATUS
This appendix contains the manufactures part numbers for the equipment drawn in
Figure 4-6.
A
Servomotor with 1000-line encoder (HIS-4525-55-E-0-2-028), controller
(TDC330/08), and connection cable (CK-HIS-35/45-15) manufactured by
Warner Electric and purchased from Electro Sales Company, Incorporated.
B
Sure flex flange (03J5/8 and 03 J 1/2) motor coupling and sure flex sleeve (03 JE)
from Flyn-Mar Incorporated
C
Model 118HT metal bellows pump (30916) from Senior Flexonics Incorporated,
Metal Bellows Division
D
Teflon encapsulated O-ring (7855-873) from Ace Glass Corporation
E
Replacement cap for metal bellows pump. Custom, no manufacturer.
F
2.75-inch half-nipple (401002) from MDC Vacuum Components Incorporated
G
Conical reducer 4.5-inch - 2.75-inch (402032) from MDC Vacuum Components
Incorporated
H
150cc 316-stainless steel cylinder (316L-HDF4-150) manufactured by Swagelok
and purchased from Albany Valve and Fitting Company
I
Sample holder, Custom, no manufacturer
J
Special reducing four-way cross (405015-XXXX) from MDC Vacuum
Components Incorporated
121
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
K
4.5-inch glass view port (450004) from MDC Vacuum Components Incorporated
L
1.33-inch ultra-high vacuum pneumatic valves (323004) from MDC Vacuum
Components Incorporated
M
Rapid response capacitance manometer (120AD-21461) connected to a Dual
Channel Digital Power Supply/Readout (PR4000AP2V2N) with cable (C B 120-110) from MKS Instruments Incorporated
N
1
.33-inch double-sided flange (100881044) from MKS Instruments
O
Constant temperature bath
P
1 33-inch rotatable blank flange (100881014) from MKS Instruments
Q
0.25-inch 316-stainless steel tubing
R
1.33-inch three-way cross (404000) from MDC Vacuum Components
Incorporated
S
1 33-inch quick disconnect adapter (413006) from MDC Vacuum Components
Incorporated
T
10-Torr capacitance manometer (122 AA-00010AC) from MKS Instruments
U
1.33-inch four-way cross (405000) from MDC Vacuum Components Incorporated
V
1.3 3-inch flange to KF16 adapter (730000) from MDC Vacuum Components
Incorporated
W
1
0-inch flexible metal vacuum line (722020) from MDC Vacuum Components
Incorporated
X
KF16 clamps (701000), and KF16 center ring, O-ring (710000) from MDC
Vacuum Components Incorporated
Y
2.75-inch flange to 1.33-inch flange (402011) from Pfeiffer Vacuum
122
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Z
TMU071 turbo pump (PM P02 801 A), DCU100 control unit (PM C01 694), 3M
DC connection cable (PM 0 5 1 103-T), and splinter screen (PM 006 376-X) from
Pfeiffer Vacuum
AA
Metal flex hose (PF 131 225-X) from Pfeiffer Vacuum
BB
DU 02.5A roughing pump (PK D41 703 A) from Pfeiffer Vacuum
CC
EVN 116 leak valve (BPV50500) from Pfeiffer Vacuum
DD
1 33-inch Del-Seal to Swagelok adapter (414000) from MDC Vacuum
Components Incorporated
EE
4.5-inch diameter by 1.5-inch wide 650W /I20V mica band heater (MB 4.51.5/650A) from Hotwatt Incorporated
FF
2.75-inch diameter by 1.5-inch wide 400W/120V mica band heater (A-13738)
from Hotwatt Incorporated
GG
Aluminum chopper wheel 8 cm in diameter and 2mm thick with a slot cut 1mm
wide by 5mm deep. No manufacturer, custom.
HH
Optical switch type VTL1ID7-20 purchased from Newark Electronics, stock
number 52F8761
* (not shown) - PCI-MIO-16XE-50 personal computer card (777385-01), SCB - 6 8
connector block (776844-01), and SH 6 8 -6 8 -EP cable (184749-01) from National
Instruments
123
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
BIBLIOGRAPHY
Angstrom, A. J.. Annalen der Physik und Chemie, 114, 513-30 (1861).
Barthomeuf, D. and B. H. Ha, “Adsorption of Benzene and Cyclohexane on Faujasitetype Zeolites”, J. Chem. Soc., Faraday Trans I, 69, 2158-65 (1973).
Beck, D. D. and J. W. Sommers, “Impact of Sulfur on Three-way Catalysts: Comparison
of Commercially Produced Pd and Pt-Rh Monoliths”, in Catalysis and
Automotive Pollution Control III. Studies in Surface Science and Catalysis, 96,
721-48, A. Frenner and J. M. Bastin Eds. (Amsterdam, Elsevier Science B. V.)
(1995).
Bird. R. B., W. E. Stewart, and E. N. Lightfoot, Transport Phenomena (New York, John
Wiley and Sons) (1960).
Bond, G. and R. B. Moyes, “Applications of Microwaves in Catalytic Chemistry”, in
M icrowave-Enhanced Chemistry, Kingston, H. M. and S. J. Haswell Eds.
(Washington, D. C., American Chemical Society) (1997).
Bourdin, V., P. G. Gray, Ph. Grenier, and M. F. Terrier, “An Apparatus for Adsorption
Dynamics Studies Using Infrared Measurement o f the Adsorbent Temperature”,
Rev. Sci. Inst. 69, 2130-6 (1998).
Bourdin, V., P. Grenier, F. Meunier, and L. M. Sun, “Thermal Frequency Response
Method for the Study o f Mass-Transfer Kinetics in Adsorbents”, A IC hE J., 42,
700-12(1996).
Breck, D. W., Zeolite M olecular Sieves (New York, John Wiley and Sons) (1974).
Buffler, C. R., Microwave Cooking and Processing: Engineering Fundamentals fo r the
Food Scientist (New York, Van Nostrand Rienhold) (1993).
Biilow, M. and P. Struve, “Experimental Evidence of the Influence o f Sorption-heat
Release Processes on the Sorption Kinetics o f Benzene in NaX Zeolite Crystals”,
./. Chem. Soc.. Faraday Trans. /. 80, 813-822 (1984).
Biilow, M., P. Struve, G. Finger, and C. Redszus, “Sorption Kinetics o f n-Hexane on
MeA Zeolites o f Different Size Crystals.” J. Chem. Soc., Faraday Trans. /, 76,
597-615 (1980).
Burkholder, H. R. and G. E. Fanslow, “Method o f Recovering Adsorbed Liquid
Compounds from Molecular Sieve Columns”, U. S. Patent, 4421651, Iowa State
University Research Foundation, Inc. (1983).
124
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Cavalcante, C. L. and D. M. Ruthven, “Adsorption of Branched and Cyclic Paraffins in
Silicalite. 1. Equilibrium”, Ind. Eng. Chem. Res., 34, 177-184 (1995).
Cha, C. Y. and Y. Kong, “Enhancement o f NOx Adsorption Capacity and Rate o f Char by
Microwaves”, Carbon, 33, 1141-6 (1995).
Cole, K. S. and R. H. Cole, J. Phys. Chem., 9, 341 (1941).
Cook, M. and W. C. Conner, “How Big are the Pores o f Zeolites”, in Proceedings o f the
12th International Zeolite Conference, M. M. J. Treacy, B. K. Marcus, M. E.
Bisher, and J. B. Higgins Eds. (Baltimore, MD, Materials Research Society)
(1999).
Debye, P.. “Hochfrequenzverluste und Molekulstruktur.”, Phys. Z., 35,
101
(1935).
Degussa, AG., “Wessalith DAY: a hydrophobic zeolite for gas purification” in Technical
Bulletin, 4307.1, (Frankfort, Degussa AG) (1992).
Einstein, A., Ann. Phys., 17, 349 (1905).
Engler, B. H., D. Lindner, E. S. Lox, A. Schafer-Sindlinger, and K. Ostgathe,
“Development o f Improved Pd-Only and Pd/Rh Three-way Catalysts”in Catalysis
and Automotive Pollution Control III. Studies in Surface Science an d Catalysis,
96, 441-60, A. Frenner and J. M. Bastin Eds. (Amsterdam, Elsevier Science B.
V.) (1995).
Evnochides, S. K. and E. J. Henley, “Simultaneous Measuremnet o f Vapor Diffusion and
Solubility Coefficients in Polymers by Frequency Response Techniques”, Journal
o f Polymer Science: Part A-2, 8, 1987-97 (1970).
Fick, A. E„ Annalen der Physik und Chemie, 94, 59 (1855).
Grenier, Ph., V. Bourdin, L. M. Sun, and F. Meunier, “Single-Step Thermal Method to
Measure Intracrystalline Mass Diffusion in Adsorbents”, A IC hE J., 41, 2047-57
(1995).
Guermeur, R. and C. Jacolin, “Influence o f Surface Silanols on the Dielectric Properties
of Nitrogen Adsorbed on Activated Silica”, Surface Science, 315, 323-336 (1994).
Peterson, E. R., “Microwave Chemistry: A Conceptual View of the Literature”, in
Oualitity Enhancements Using Microwaves; 28th Microwave Symposium
Proceedings Quality Enhancements Using Microwaves (Montreal, International
Microwave Power Institute) (1993).
125
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Harrison, I. D., H. F. Leach, and D. A. Whan, Proceeding o f the 6 th International Zeolite
Conference. Proceeding o f the 6 th International Zeolite Conference, Butterworths,
Guildford (1984).
Izmailova, S. G., I. V. Karetina, S. S. Khvoshchev, and M. A. Shubaeva, “Calorimetric
and IR-Spectroscopic Study of Methanol Adsorption on Zeolites”, Journal o f
Colloid and Interface Science, 165, 318-324 (1994).
Karger, J. and D. M. Ruthven, Diffusion in Zeolites an d other M icroporous Solids (John
Wiley & Sons, New York) (1992).
Kobayashi, S., Y. K. Kim, C. Kenmizaki, S. Kushiyama, and K. Mizuno, “Control o f
Adsorption by Microwave Irradiation”, Chemistry Letters, 769-770 (1996).
Kobayashi, S., K. Mizuno, S. Kushiyama, R. Aizawa, Y. Koinuma, and H. Ohuchi, “Gas
Adsorption and Desorption Method”, U. S. Patent, 5282886, Director-General o f
Agency o f Industrial Science and Technology (1994).
Krieger-Brockett, B., M. Mingos, and J. Wan, in Proceedings: M icrowave-Induced
Reactions Workshop, M. Burka and K. R. Amamath Eds. (Pacific Grove, Meeting
Planning Associates) (1993).
Lide, D. R., Ed., CRC Handbook o f Chemistry and Physics (CRC Press, Boston) (1991).
Lundgren, S., G. Spiess, O. Hjortsberg, E. Jobson, I. Gottberg, and G. Smedler, “Sulfur
Adsorption and Desorption on Fresh and Aged Ce Containing Catalysts”, in
Catalysis and Automotive Pollution Control III. Studies in Surface Science and
Catalysis, 96, 763-74, A. Frenner and J. M. Bastin Eds. (Amsterdam, Elsevier
Science B. V.)(1995).
Meier, W. M. and D. H. Olson, Atlas o f Zeolite Structure Types (ButterworthHeinemann) (1992).
Mezey, E. J. and S. T. Dinovo, “Adsorbent Regeneration and Gas Separation Utilizing
Microwave Heating”, U. S. Patent, 4322394, Batteile Memorial Institute (1982).
Micke, A., M. Biilow, and M. Kocirik, “A Nonequilibrium Suface Barrier for Sorption
Kinetics o f p-Ethyltoluene on ZSM-5 Molecular Sieves” J. Phys. Chem., 98, 924929 (1994).
Monroe, D. R„ M. H. Krueger, D. D. Beck, and M. J. D'Aniello, “The Effect o f Sulfur on
Three-way Catalysts”, in Catalysis and Automotive Pollution Control II. Studies
in Surface Science and Catalysis, 71, 593-616, A. Crucq Ed. (Amsterdam,
Elsevier Science B. V.) (1991).
126
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Naphtali, L. M. and L. M. Poinski, “A Novel Technique for Characterization of
Adsorption Rates on Heterogenous Surfaces”, J. Phys. Chem., 67, 369-75 (1963).
Nayak, V. S. and J. B. Moffat (1988). “Sorption and Diffusion o f Alcohols in Heteropoly
Oxometalate and ZSM-5 Zeolite”, J. Phys. Chem., 92, 7097-7102 (1988).
Neufeldt, V. and D. B. Guralnik, Eds., W ebster’s New World Dictionary, 3rd. Edition
(Simon and Schuster, Inc., New York) (1988).
Ogunnaike, B. A. and W. H. Ray, Process Dynamics, Modeling, and Control (Oxford
University Press, New York) (1994).
Opperman, S. H. and M. S. Arsenault, “Regeneration Apparatus for Recovery o f
Volatiles”, U. S. Patent, 5509956, Horizon Holdings, Inc. (1996).
Peterson, E. R„ “Microwave Chemistry: A Conceptual View o f the Literature”, in
Qua/itity Enhancements Using Microwaves; 28th Microwave Symposium
Proceedings Quality Enhancements Using M icrowaves (Montreal, International
Microwave Power Institute) (1993).
Pozar, D. M., Microwave Engineering (Addison-Wesley Publishing Company' Reading)
(1990).
Rees, L. C. V. and D. Shen, “Characterization o f Microporous Sorbents by Frequency
Response Methods”, Journal o f Gas Separations and Purification, 7, 83-9 (1993).
Roussy, G., J. M. Thiebaut, M. Souri, A. Kiennemann, and G. Maire, “Controlled
Oxidation o f Methane on Doped Catalysts Irradiated by Microwaves”, in
Oualitity Enhancements Using Microwaves; 28th Microwave Symposium
Proceedings Quality Enhancements Using Microwaves (Montreal, International
Microwave Power Institute) (1993).
Roussy, G., A. Zoulalian, M. Charreyre, and J. Thiebaut, “How Microwaves Dehydrate
Zeolites”, J. Phys. Chem., 88 , 5702-5708 (1984).
Ruthven, D. M„ “Diffusion in Zeolites”, in Zeolites: A Refined Toolfo r D esigning
Catalytic Sites. Studies in Surface Science and Catalysis, 97, 223-234, A. Crucq
Ed. (Amsterdam, Elsevier Science B. V.) (1995).
Shen, D. and L. C. V. Rees, “Frequency Response Measuremnets o f p-Xylene Diffusion
in Silicalite-1 and -2”, J. Chem. Soc., Faraday Trans. I, 89, 1063-5 (1993).
Stuerga, D. A. C. and P. Gaillard, “Microwave Athermal Effects in Chemistry: A Myth's
Autopsy.” Journal o f Microwave Power and Electromagnetic Energy, 31, 87-113
(1996).
127
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Thamm, H., “Calorimetric Study on the State o f C1-C4 Alcohols sorbed on Silicalite”, J.
Chem. Soc., Faraday Trans. I, 85, 1-9 (1989).
Turner, M. D., R. L. Laurence, W. C. Conner and K. S. Yngvesson “Microwave
Radiation's Influence on Sorption and Competitive Sorption in Zeolites”, AIChE
Journal, 46, 758-68 (2000).
Turner, M. D., W. C. Conner, R. L. Laurence, S. Yngvesson, “The Influence o f
Microwave Energy on Adsorption in Zeolites”, in Proceedings o f the 12th
International Zeolite Conference, M. M. J. Treacy, B. K. Marcus, M. E. Bisher,
and J. B. Higgins Eds. (Baltimore, MD, Materials Research Society) (1999).
van den Begin, N. G. and L. V. C. Rees, in Proceedings o f the Eighth International
Zeolite Conference, Amsterdam, Elsevier (1989).
Windgasse, G. and L. Dauerman “Microwave Treatment o f Hazardous Wastes: Removal
of Volatile and Semi-Volatile Organic Contaminates from Soil”, J. o f Microwave
Power and Electromagnetic Energy, 27, 23 (1992).
Yasuda, Y., “Determination of Vapor Diffusion Coefficients in Zeolite by the Frequency
Response Method”, J. Phys. Chem., 8 6 , 1913-7 (1982).
Yasuda, Y„ “Frequency Response Method for Study o f the Kinetic Behavior o f a GasSurface System. 1. Theoretical Treatment”, J. Phys. Chem., 80, 1867-9 (1976a).
Yasuda, Y., “Frequency Response Method for Study o f the Kinetic Behavior o f a GasSurface System. 2. An Ethylene-on-Zinc Oxide System”, J. Phys. Chem., 80,
1870-5 (1976b).
Yasuda, Y. and M. Saeki, “Kinetic Details o f a Gas-Surface System by the Frequency
Response Method”, J. Phys. Chem., 82, 74-80 (1978).
Yasuda, Y., Y. Suzuki, and H. Fukada, “Kinetic Details o f a Gas-Porous Adsorbent
System by the Frequency Response Method”, J. Phys. Chem., 95, 2486-92 (1991).
Zlotorzynski, A., “The Application o f Microwave Radiation to Analytical and
Environmental Chemistry”, Critical Reviews in Analytical Chemistry, 25, 43-76
(1995).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Документ
Категория
Без категории
Просмотров
0
Размер файла
5 333 Кб
Теги
sdewsdweddes
1/--страниц
Пожаловаться на содержимое документа