Citation
Heterogeneous-phase reactions of nitrogen dioxide with vermiculite-supported magnesium oxide

Material Information

Title:
Heterogeneous-phase reactions of nitrogen dioxide with vermiculite-supported magnesium oxide
Creator:
Kimm, Larry Thomas, 1960-
Publication Date:
Language:
English
Physical Description:
xiii, 201 leaves : ill. ; 29 cm.

Subjects

Subjects / Keywords:
Combustion ( jstor )
Jet engines ( jstor )
Kinetics ( jstor )
Oxygen ( jstor )
Reactants ( jstor )
Reaction kinetics ( jstor )
Solids ( jstor )
Sorbents ( jstor )
Sorption ( jstor )
Surface areas ( jstor )
Dissertations, Academic -- Environmental Engineering Sciences -- UF
Environmental Engineering Sciences thesis, Ph. D
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1995.
Bibliography:
Includes bibliographical references (leaves 192-200).
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Lqarry Thomas Kimm.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
021933959 ( ALEPH )
33662103 ( OCLC )

Downloads

This item has the following downloads:


Full Text










HETEROGENEOUS-PHASE REACTIONS OF NITROGEN DIOXIDE WITH
VERMICULITE-SUPPORTED MAGNESIUM OXIDE












By

LARRY THOMAS KIMM


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


1995
























This work is dedicated to my family, especially my wife, Lisa, whose patience, love

and support during this effort made it all possible, and to my children, Meredyth and Wilson,

who always give me the best welcome home after a long day.













ACKNOWLEDGMENTS


I am extremely grateful for the support of my doctoral committee chairman, Dr. Eric

R. Allen, for his wealth of knowledge in academic research, for sharing in the excitement of

discovery, as well as for providing encouragement when problems arose. I will always fondly

remember our enlightening discussions about my research, as well as life in general, as high

points in my research. I also appreciate the sage advice of Dr. Dale A. Lundgren, whose

practical and applied approach to engineering goes a long way beyond what can be learned

in a textbook. I am thankful to both professors and to my fellow air pollution graduate

students, for their friendship and comradery during my time as a graduate student at UF.

Special thanks to Dr. Joseph D. Wander, whose belief in the value of academic

research is unparalleled, for providing the idea for this project, as well as for his help,

friendship and advice all along the way. I am also thankful to Dr. W. Emmett Bolch, Jr.,

whose letter years ago convinced me to come to the University of Florida, and to Dr. Robert

J. Hanrahan, to whom I credit the excellent performance of the experimental system.

I would like to acknowledge Colonel Robert A. Capell, USAF, BSC, who saw the

need for a bioenvironmental engineer with a Ph.D. in air pollution. Finally, I would like to

thank the Armstrong Laboratory/OL-EQS, Tyndall AFB, Florida for providing funding for

this research project, and the U.S. Air Force and the Air Force Institute of Technology, for

their confidence in selecting me for advanced education and for the financial support.













TABLE OF CONTENTS




ACKNOWLEDGMENTS .............................................. iii

LIST OF TABLES ................................................... vii

LIST OF FIGURES .................................................. viii

A B STR A C T ........................................................ xii

1. INTRODUCTION .................................................. 1
B background ........................ ........................... 1
Properties and Health Effects of Nitrogen Oxides .................. 1
Environmental Impacts of NOx ................................ 4
Sou rces ......... ........ ....... ... .... .. ... .... ....... .. 4
NOX Formation ............................................... 7
Fuel NOX .............................................. 7
Therm al NOx ............................................. 8
Regulation of NOx Emissions ...................................... 11
NOx Controls for Stationary Sources ................................ 13
Application of NOx Control Strategies ........................ 13
Modifications to Operating Conditions ......................... 14
Combustion M modifications .................................. 15
Exhaust Gas Treatment .................................... 16

2. CONTROL OF NOx THROUGH GAS-SOLID INTERACTION ............. 18
Exhaust Gas Treatment M ethods ................................... 18
Catalytic M ethods ....................................... 19
Non-Catalytic M ethods ..................................... 21
Jet Engine Test Cells ........................................... 22
U.S. Air Force (USAF) Sponsored Research .......................... 25
Vermiculite-based Catalyst .................................. 26
Field Studies ............................................. 30
R elated R research ............................................... 31
G as-Solid Interaction ........................................... 32








Generalized Non-Catalytic Gas-Solid Reactions .


M athematical M odels ...............................
Progressive-Conversion Model ..................
Shrinking Unreacted-Core Model ................
G rain M odel ................................
Analagous Gas-Solid Reaction System Research ..........
S02-CaO Reaction ...........................
S02-Calcium Hydroxide Reaction ...............
SO2-CalciumCaebonate Reaction ...............
Other Acid-gas-Calcium-based-solid Reactions .....
General Applicability of the Shrinking Unreacted-Core Model
Research Justification ...............................
Research Objectives ................................

3. EXPERIMENTAL METHODS AND MATERIALS ..........
General Research Approach ..........................
Experimental Variables ........................
MgO-Vermiculite Reactive Sorbent Material .


Temperature


Pressure ......................
Gas Composition ...............
Gas Flow Rate(s) ...............
Gas Reaction/Sorption/Desorption ..
Fixed-Bed Reactor System ....................
General Experimental Considerations ......
Experimental Arrangement ..............
Tubular Fixed-bed Reactors .............
Experimental Procedures .....................
G as A nalyses ..............................
Sorbent Surface Area Determination ............
Sorbent Particle Size Determination .............

4. RESULTS AND DISCUSSION ...................
Intrinsic Kinetic Studies ......................
Limitation of Gas-Film-Mass-Transfer Resistance .
Solid Sorbent Characterizations ................
Surface Area Analyses .................
Particle Size Distributions ....................
NOx Removal by Mg(N03)2-coated Vermiculite ....
NO, Removal by Sorbent Material ..............
NO Removal by Sorbent Material .........
NO2 Removal by Sorbent Material ........
Effects of Operational Variables ................


. . . . . . . . . . . 3 3


. . . . . . . . . .








Effects of NO2 Concentration ................................ 94
Effects of Oxygen ......................................... 99
Effects of Bed Temperature ................................ 105
Activation Energy Determination ...................... 112
Effects of W ater ....................................... 120
Effects of Residence (Reaction) Time ......................... 122
Sorbtech-Supplied Sample Results ................................. 123
Pressure Drop Characteristics .................................... 124
Magnesium Nitrate Surface Decomposition .......................... 127
Sorbent "Lifetime" and Regeneration ............................... 130
Proposed Reaction M echanism .................................... 132
Application of the Shrinking Unreacted Core/Grain Model ............... 133
Diffusion-Through-Gas-Film-Control ......................... 134
Chemical-Reaction-Control ................................ 137
Diffusion-Through-Inert-Product-Layer-Control................. 140
Derivations of the Shrinking Unreacted-Core Model ................... 144
Correlation Between Gas-Phase Concentration and Solid Conversion ....... 146
Local-Equilibrium Theory .................................. 146
Application of Shrinking Unreacted-Core/Grain Model to Experimental Data. 148
50 ppm NO2 at a Reaction Temperature of 423 K ................ 149
20 ppm NO2 at a Reaction Temperature of 373 K ................ 165
200 ppm NO2 at a Reaction Temperature of 473 K ............... 169
General Applicability of the Shrinking Unreacted-Core Model ............ 178

5. PRACTICAL CONSIDERATIONS .................................. 182
Constraints on MgO-Vermiculite Sorbent Usage ...................... 182

6. CONCLUSIONS AND RECOMMENDATIONS ......................... 185
Conclusions .................................................. 185
Recommendations for Further Research ............................. 188

LIST OF REFERENCES ............................................. 192

BIOGRAPHICAL SKETCH ........................................... 201













LIST OF TABLES


Table page

1-1. Physical properties of nitrogen oxides (NOy) ............................ 2

1-2. Emission factors for nitrogen oxides (NO, as NO2) ........................ 6

2-1. Physical properties of selected Mg-containing inorganic compounds .......... 28

3-1. USAF F110 turbine engine emissions data (as reported) ................... 54

4-1. BET surface areas of selected samples of laboratory-prepared
M gO-vermiculite sorbent .......................................... 82

4-2. Intrinsic first-order rate coefficients (s-1) for N02 removal by
MgO-vermiculite sorbent ([02]=0%) ................................. 100

4-3. Intrinsic first-order rate coefficients (s') for NO2 removal by
MgO-vermiculite sorbent ([02]=10%) ................................ 106













LIST OF FIGURES


Figure pge

2-1. Schematic representation of a jet engine test cell (JETC) ................. 24

2-2. Schematic representation of the shrinking unreacted-core model ............ 38

2-3. Schematic representation of the grain model ........................... 40

3-1. Experimental arrangement for packed-bed studies ....................... 59

3-2. Example low-flow-rate rotameter calibration curves
(Omega N042-15 ST Tube) ....................................... 61

3-3. Example high-flow-rate rotameter calibration curves
(Omega N092-04G Tube) ...................................... ... 62

3-4. Schematic representation of 316 stainless steel packed-bed reactor .......... 64

3-5. Schematic representation of internal sampling tube location and appearance ... 65

3-6. Schematic representation of 316 stainless steel mixing or sampling manifold ... 66

3-7. Typical chemiluminescent NO, analyzer calibration curve (for NO channel) ... 74

3-8. Typical IR-2100 oxygen analyzer calibration curve ...................... 75

4-1. Gas-film-mass-transfer resistance evaluation NO2 penetration (C,,/C.m)
versus bed exposure time ([N02]n=200 ppm, T=473 K, [02]=0%) .......... 81

4-2. Log-probability plot ofMgO-vermiculite sorbent particle size distribution .... 84

4-3. Log-probability plot of Akrochem Elastomag 170 MgO powder particle size
distribution (manufacturer-provided data) ............................. 85







4-4. NO removal by humidified MgO-vermicuilite sorbent ([NO]m=210 ppm,
T=423 K, [02]=11%, 3% H20 vapor) ............................... 88
4-5. NO2 removal versus residence time (T=473 K, [02]=9.9%) ................ 90

4-6. First-order kinetic plot ([N02],=100 ppm, T=473 K, [02]=10%) ........... 92
4-7. First-order kinetic plot ([N02]m=100 ppm, T=473 K, [02]=10.5%) .......... 93
4-8. NO2 penetration vs bed exposure time (T=473 K, [02]=0%) ............... 95
4-9. NO2 penetration vs bed exposure time (T=423 K, [02]=0%) ............... 96
4-10. NO2 penetration vs bed exposure time (T=373 K, [02]=0%) ............... 97
4-11. NO2 penetration vs bed exposure time (T=473 K, [02]=10%) ............. 102

4-12. NO2 penetration vs bed exposure time (T=423 K, [O2]=10%) ............. 103

4-13. NO2 penetration vs bed exposure time (T=373 K, [O2]=10%) ............. 104
4-14. NO2 penetration vs bed exposure time ([NO2]1in=200 ppm, [02]=0%) ....... 108
4-15. NO2 penetration vs bed exposure time ([NO2]i=100 ppm, [02]=0%) ....... 109
4-16. NO2 penetration vs bed exposure time ([NO2],=50 ppm, [02]=0%) ........ 110

4-17. NO2 penetration vs bed exposure time ([N02],=20 ppm, [02]=0%) ........ 111

4-18. NO2 penetration vs bed exposure time ([NQ2]m=100 ppm, [02]=10%) ...... 113

4-19. NO2 penetration vs bed exposure time ([NO2],=50 ppm, [02]=10%) ....... 114
4-20. NO2 penetration vs bed exposure time ([N02],M=20 ppm, [02]=10%) ....... 115

4-21. Arrhenius plot of In k versus 1000/T (T between 473 and 423 K,
([NO2]=100 ppm, [02]=0% ) .................................... 116
4-22. Arrhenius plot of In k versus 1000/T (T between 473 and 373 K,
([NO2]M=20 ppm, [021=0%) ..................................... 117
4-23. NO2 penetration versus bed exposure time (20 ml -120 added to 7 g
MgO-vermiculite sorbent, [N02]m=100 ppm, [02]=0% T=473 K) ......... 121








4-24. Predicted (Ergun equation) versus experimental pressure drop as a function
of superficial velocity for a 0.1 m long bed ofMgO-vermiculite sorbent
(T=473 K) ................................................... 126

4-25. Thermal decomposition of Mg(N03)2. 6H20 on vermiculite (temperature
increasing w ith tim e) ........................................... 128

4-26. Thermal decomposition of used MgO-vermiculite sorbent (T=523 K) ....... 129

4-27. Comparative bed outlet NO2 and NO concentrations versus bed exposure
time showing NO production ..................................... 131

4-28. Graphical representation of the shrinking unreacted-core model under diffusion-
through-gas-film control ........................................ 135

4-29. Graphical representation of the shrinking unreacted-core model under chemical-
reaction control ............................................... 138

4-30. Graphical representation of the shrinking unreacted-core model under diffusion-
through-inert-product-layer control ................................ 141

4-31. Shrinking unreacted-core model evaluation ([N02]m=50 ppm, T=423 K,
[o0 21= 0% ) .......................... ........................ 150

4-32. Shrinking unreacted-core chemical-reaction-control-equation evaluation, Tc=397
minutes, ([NO2].,=50 ppm, T=423 K, [02]=0%) ...................... .152

4-33. Shrinking unreacted-core product-layer-diffusion-equation evaluation, Td=820
minutes, ([NO2,=50O ppm, T=423 K, [02]=0%) ....................... 153

4-34. Shrinking unreacted-core chemical-reaction-control-equation evaluation, Tr=205
minutes, ([N02],=50 ppm, T=423 K, [02]=0%) ....................... 154

4-35. Shrinking unreacted-core product-layer-diffusion-equation evaluation, Td=957
minutes, ([N02]m=50 ppm, T=423 K, [02]=0%) ....................... 155

4-36. Cumulative mass NO2 removed versus bed exposure time ([NO2]j=50 ppm,
T=423 K [02]=0% ) ........................................... 157

4-37. Cumulative mass NO2 removed versus bed exposure time for. a long-term run
([NO2]in=50 ppm, T=423 K, [021]=0%, Gas samples collected from
center of bed) ......................... .. ......... ............ 158








4-38. Plot of inverse first-order rate coefficient (1/k) versus bed exposure time
([NO2].=47.5ppm, T=423 K, [021=0%) ............................ 163
4-39. Experimental versus predicted NO2 concentration ([NO2],=47.5 ppm,
T=423 K, [O2]=0%, Gas samples collected from middle of bed) ........... 164

4-40. Shrinking unreacted-core model evaluation ([N02]m=20 ppm, T=373 K,
[02]=0% ) .................................................... 166

4-41. Shrinking unreacted-core chemical-reaction-control-equation evaluation, T,=1023
minutes, ([N02,]m=20 ppm, T=373 K, [021=0%) ....................... 167

4-42. Shrinking unreacted-core product-layer-diffusion-equation evaluation, rd=1093
minutes, ([NO2]m=20 ppm, T=373 K, [02]=0%) ....................... 168

4-43. Experimental versus predicted NO2 concentration ([NO2]n=19.1 ppm, T=373 K,
[02]=0%, Gas samples collected from middle of bed) ................... 170

4-44. Shrinking unreacted-core model evaluation ([NO2],n=200 ppm, T=473 K,
[0 2]= 0% ) .................................................. 171

4-45. Shrinking unreacted-core chemical-reaction-control-equation evaluation, Tr=495
minutes, ([N02]n=200 ppm, T=473 K, [02]=0%) ..................... 172

4-46. Shrinking unreacted-core product-layer-diffusion-equation evaluation, Td= 1015
minutes, ([NO2],=200 ppm, T=473 K, [02]=0%) ...................... 173

4-47. Experimental versus predicted NO2 concentration ([NO2]m=170.0 ppm, T=473 K,
[02]=0%, Gas samples collected from middle of bed)................... 175

4-48. Cumulative mass NO2 removed versus bed exposure time for a long-term run
([NO21,n=170 ppm, T=473 K, [02]=0%, Gas samples collected from center
of bed) ...................................................... 176

4-49. Comparative bed NO2 and NO concentrations versus bed exposure time during
long-term run showing NO production ([NO2]i=170 ppm, T=473 K, [02]=0%,
Gas samples collected from center of bed) ........................... 177













Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

HETEROGENEOUS-PHASE REACTIONS OF NITROGEN DIOXIDE WITH
VERMICULITE-SUPPORTED MAGNESIUM OXIDE

By

Larry Thomas Kimm

August 1995


Chairperson: Eric R. Allen
Major Department: Environmental Engineering Sciences

Controlling nitrogen oxides (NO) from a non-steady-state stationary source like a jet

engine test cell (JETC) requires a method that is effective over a wide range of conditions.

A heterogeneous, porous, high surface area sorbent material comprised of magnesium oxide

powder attached to a vermiculite substrate has been commercially developed for this purpose.

Data from extensive laboratory testing of this material in a packed-bed flow system are

presented. NO2 removal efficiencies, kinetics, and proposed NO2 removal mechanisms over

a range of representative JETC exhaust gas characteristics are described. Exhaust gas

variables evaluated included: NO2 concentration, temperature, flow rate (retention time),

oxygen content and moisture content. Availability of water and oxygen were found to be

important variables. It is probable that water is necessary for the conversion of MgO to

Mg(OH)2, which is a more reactive compound having thermal stability over the range of








temperatures evaluated. Gaseous oxygen serves to oxidize NO to NO2, the latter being more

readily removed from the gas stream. The presence of oxygen also serves to offset thermal

decomposition of NO2 or surface nitrite/nitrate. Effective "lifetime" and regenerability of the

exposed sorbent material were also evaluated. NO2 removal efficiencies were found to

greatly exceed those for NO, with a maximum value greater than 90 percent. The effective

conversion of NO to NO2 is a crucial requirement for removal of the former. The reaction

between NO2 and MgO-vermiculite is first-order with respect to NO2. The temperature

dependence of the first-order rate coefficients provided evidence that data were collected in

the region of chemical-reaction control. Activation energies associated with this reaction

ranged from approximately 20 and 36 kJ/g-mol when oxygen was not present and 15 to 25

kJ/g-mol when oxygen was present in the system. The shrinking unreacted-core model was

used to describe the physical and chemical mechanisms occurring in the removal of NO2 from

a flowing gas stream. It appeared that the reaction begins with control by chemical reaction

and progressed to control by diffusion processes as reaction progressed. An empirical

relationship was also developed which allowed for the prediction of NO2 removal with

increasing bed exposure time. The range of temperatures and concentrations evaluated, while

valid for the representation of simulated jet engine test cell conditions, appeared to be a

transition region making the absolute determination of rate-limiting mechanisms) difficult.

With impending regulations aimed at controlling JETCs as stationary sources of NOR, results

from these studies will be useful to environmental engineers developing air pollution controls

for similar sources. Knowledge of parameters affecting the efficiency and capacity of NO,

control systems employing this medium is necessary to ensure compliance with regulations.













CHAPTER 1
INTRODUCTION



Background



The combustion of carbonaceous jet fuels during jet engine testing produces

significant quantities of nitrogen oxides (N0). Besides direct negative impacts on human

health, nitrogen oxides have been linked to other detrimental effects on air quality and the

environment. The latter include interactions with hydrocarbons to produce photochemical

oxidants and smog, and contribution to the phenomenon of acid precipitation through nitric

acid formation. Nitric oxide emissions may also contribute to the degradation of visibility and

aesthetics by direct emission of high NO levels in elevated plumes followed by atmospheric

oxidation to NO2 and fine particulate aerosols resulting from by-products of NO2 reactions

(Wark and Warner, 1981).

Properties and Health Effects of Nitrogen Oxides


There are seven possible forms of nitrogen oxides (NxOy). Two of the most abundant

of these gaseous oxides of nitrogen, nitric oxide (NO) and nitrogen dioxide (NO2), are the

predominant air pollutants of environmental and health concern. These two forms, however,

are rapidly interchangeable in the atmosphere and are often grouped together and collectively








2
called "NOx." Some of the physical properties associated with gaseous nitrogen oxides

(NxOy) are shown in Table 1-1. Nitrous oxide (N20) is another oxide of nitrogen that is also

present in the atmosphere at appreciable concentrations. Nitrous oxide is a gas with

recognized anesthetic properties, which has also been suggested to be a product of

combustion. Its ambient concentration is approximately 0.5 parts per million (ppm), well

below the threshold concentration for biological effects. Fortunately, in the troposphere, it

is balanced atmospherically by a cycle that is independent of the other oxides of nitrogen.

Nitric oxide (NO) is a colorless gas whose ambient concentration is normally well below 0.5

ppm. At this concentration it produces minimal effects on human health. However, nitric

oxide is a precursor to the formation of NO2 and is an active compound in the formation of

photochemical smog, initiating reactions that produce secondary air pollutants.

Table 1-1. Physical properties of nitrogen oxides (NxOy).


Species Mol Wt Solubility in Melting Point
(g/gmol) Water (C)
__________________(mL/100g)______
NO 30.01 7.34 -163.6

NO2 46.01 See Note 1 See Note 2
NO3 62.00 Soluble (ether) Decomp. 20.0
N20 44.02 130.0 -102.4
N2Oa3 76.01 Soluble (ether) -111.0
N204 92.02 130.52 -11.3
N205 108.01 Soluble 40.7

NOTES: 1. Reacts with H20 forming HONO2 and/or HONO
2. Liquid and solid forms are primarily N204
Sources: Neal et al., 1981.; Lange's Handbook of Chemistry, 1973.


Boiling Point
(0C)


-151.7

See Note 2
Unavailable
-89.0
2.0
21.2
Sublimes 32.3












The most harmful effects of NOx are primarily attributed to exposure to NO2, the most

toxic of these oxides. NO2 is a reddish-brown gas that is quite visible in sufficient

concentration (> 0.25 ppm). NO2 is not a primary pollutant in the sense that it directly affects

human health, unless the exposure concentration is high. NO2 exerts its main toxic effects on

the lungs via free-radical-mediated reactions and other mechanisms. Both NO and NO2 are

free radicals that may produce lipid peroxidation reactions within human cells. The

environmental hazard of NO2 is primarily associated with the pulmonary effects of the

pollutant. Exposure to concentrations of NO2 from 0.7 to 5.0 ppm for 10 to 15 minutes have

produced abnormalities in pulmonary airway resistance. Exposures to 15 ppm can cause eye

and nose irritation, and pulmonary discomfort is noted at 25 ppm for exposures of less than

one hour.

The greatest danger of exposure to NO2 is the delay in experiencing its full effect upon

the respiratory system. The delayed effect of NO2 injury is made potentially more dangerous

by two other factors. Human perception of the odor of NO2 is insufficient to warn against

injury or even death. Additionally, NO2 quickly desensitizes an individual to its odor through

olfactory fatigue, and if NO2 levels gradually rise, a person could unknowingly be exposed

to concentrations high enough to cause permanent injury or death. Because of the low

solubility of NO2 in water, it is only slowly removed from the lungs by circulating blood and

may remain in contact with lung cells for prolonged periods of time. Given the same total

dose, short-term exposure to high concentrations of NO2 is more injurious than long-term

exposure to lower concentrations. The toxic effects of NO2 are often synergistic with or










additive to those of other environmental contaminants (Wark and Warner, 1981; Neal et al.,

1981).

Environmental Impacts of NOX


In addition to the direct effects on health, nitrogen oxides may create other detrimental

effects on air quality and the environment. These effects include interactions with

hydrocarbons to produce photochemical oxidants and smog; contribution to acid precipitation

through nitric acid formation; and degradation of visibility and aesthetics by formation of

high NO2 levels in elevated plumes and fine particulate aerosols resulting from by-products

of NO. reactions (Organization for Economic Cooperation and Development, 1983).

Sources


Over 95 percent of all the man-made nitrogen oxides that enter the atmosphere in

the United States are produced by the combustion of fossil fuels. As of 1990, over 55 percent

of these emissions are attributed to stationary sources, such as utility and industrial boilers,

gas turbines, and stationary engines. Approximately 40 percent are from mobile sources

(transportation). The remaining fraction is from miscellaneous sources including industrial

processes and waste disposal. On a global scale, the anthropogenic NO, emission rate is

minor compared to natural emissions and formation in the environment. As a result, air

pollution associated with NO,, is mainly a local or regional problem (Cooper and Alley, 1994;

Wark and Warner, 1981).

The level of effect of nitrogen oxides on the local environment is highly dependent

upon the emission rates of sources and prevailing meteorological conditions and topological








5
features in the local area. At an emission source, the concentration of nitrogen oxides is much

higher than ambient background levels. Emission rates vary widely depending on the type of

source and type of fuel used, as well as on the type and quality of pollution control equipment

employed. This variation in NOx production rates for uncontrolled sources is quite evident

in Table 1-2.

From Table 1-2, it can be seen that an individual aircraft appears to emit negligible

amounts of nitrogen oxides when compared to much larger combustion sources. However,

recent studies by Johnson et al. (1992) indicate that aircraft emissions into the troposphere

at an altitude above 10 km produce increased concentrations of ozone, off-setting ozone

depletion, but contributing to increased surface temperatures ("global warming"). Modeling

suggests that the radiative forcing of surface temperature is about 30 times more sensitive to

aircraft emissions of nitrogen oxides than surface emissions. In the vicinity of an airport or

Air Force base, where ground operations (taxiing, take-offs, landings, and maintenance in

engine test cells) are combined with flight operations for a large number of aircraft, the

contribution of pollutants in the local area due to nitrogen oxides can be significant. NO.

standards have traditionally not been enforced for jet engine test cells, which qualify for

regulation as stationary sources. Gas turbine engines operating as mobile sources also have

not been regulated in the past. Evolving regulations including the 1990 Clean Air Act

Amendments, the failure of existing controls on stationary industrial processes to significantly

lower urban ozone concentrations, and a demonstrated correlation between ozone levels and










Table 1-2. Emission Factors for Nitrogen Oxides (NO. as NO2)


Source Average Emission Factor (as NO2)
Coal
Household and commercial 8 lb/ton of coal burned
Industry and utilities 20 lb/ton of coal burned
Fuel Oil

Household and commercial 12-72 lb/1000 gal of oil burned
Industry 72 lb/1000 gal of oil burned
Utility 104 lb/1000 gal of oil burned
Natural Gas

Household and commercial 116 lb/million ft3 of gas burned
Industry 214 lb/million ft3 of gas burned

Utilities 390 Ib/million ft3 of gas burned
Gas Turbines 200 lb/million ft3 of gas burned
Waste Disposal
Conical incinerator 0.65 lb/ton of waste burned
Municipal incinerator 2 lb/ton of waste burned
Mobile Source Combustion

Gasoline-powered vehicle 113 lb/1000 gal of gasoline burned

Diesel-powered vehicle 222 lb/1000 gal of oil burned
Aircraft

Conventional 23 lb/flight per engine
Fan-type jet 9.2 lb/flight per engine
Nitric Acid Manufacture 57 lb/ton of acid product


Source: Wark and Warner, 1981 (Based upon U.S. EPA AP-67, 1970).








7

NOx concentrations seem to point toward increased regulatory pressure to reduce NO.

emissions from jet engine test cell activities in the near future.


NO, Formation



If NOx standards for test cell emissions are to be considered, clearly those systems

available for controlling these emissions must be thoroughly evaluated. To reduce NOx

emissions, two primary actions can be taken. These are the direct control of the reactions)

producing the pollutant (combustion process), or control of the process effluent. To

understand the means of controlling NO, emissions, it is critical to examine the basic

chemistry, thermodynamics, and kinetics of the formation reactions. Oxides of nitrogen

formed by combustion processes are generally caused either by the conversion of chemically

bound nitrogen in the fuel ("fuel NO,"), or by thermal fixation of atmospheric nitrogen in the

combustion air ("thermal NOx").

The principal factors affecting the formation of nitrogen oxides in combustion

processes are the amount of fuel-bound nitrogen, peak combustion temperature, oxygen level

at peak temperature, and residence time in the combustion zone (Organization for Economic

Cooperation and Development, 1983).

Fuel NO.

When a fuel containing organically bound nitrogen is burned, the contribution of

the fuel-bound nitrogen to total NO, can be significant. The N-C bond is much weaker than

the N-N bond in molecular nitrogen, so fuel nitrogen can be more easily oxidized to NO by








8
oxygen in the combustion gas. The mechanism of the conversion of fuel nitrogen is believed

to proceed via a series of intermediates. One of the most important intermediates is hydrogen

cyanide (HCN), which is fully converted from bound nitrogen in the reaction zone of the

flame. Moving away from the reaction zone, the amount of HCN decreases as a result of its

conversion into N2 and NOR. Ammonia (NH3) may be detected along with N2, HCN, and NO.

in the products of combustion of nitrogen-containing fuels in the presence of a deficiency of

atmospheric nitrogen. The conversion of fuel nitrogen into HCN is rapid, and the rate-

limiting stage in the formation of NO. is the oxidation of HCN. The yield of NOx is

influenced by the content of nitrogen in the fuel, the amount of excess air, and, to a lesser

extent, combustion temperature (Ismagilov et al., 1990; Cooper and Alley, 1994; Wark and

Warner, 1981).

Thermal NO,

Due to the extremely high temperatures in the combustion zones of aircraft jet engines

(up to 2500 K), thermal nitrogen oxides are formed by the oxidation of atmospheric nitrogen.

The accepted model for the chemical reactions responsible for NO, formation in the post-

combustion zone was developed by Zeldovich in 1946. The reactions in Zeldovich's model

are as follows:

N2+O NO + N (1-1)


N +02 NO + O (1-2)


and N+OHNO+H


(1-3)








9
In presenting these reactions, it is assumed that the fuel combustion reactions

(between C, H, and 0) have reached equilibrium and that the concentrations of 0, H, and OH

can be described by equilibrium equations. Considering only the thermodynamics of NO,

formation, the two overall reactions of concern are those which produce NO and NO2. The

relevant equilibrium reactions are

N2 + 02 2NO (1-4)


and NO + 1/202 N02 (1-5)

The equilibrium constants for reactions (1-4) and (1-5) (Kp, and Kp2, respectively) are


(pVw)2 (y,,,
K,, ( = PNo) (-No)2 (1-6)
K -- --- = ---- (.1-6)
(PN2)PO2 (yN2)O2



S(PNo) (PT)(YN0) (1-7)
PNo2 (P o 2) YO(Y o2 )




where: Kp = equilibrium constant,

Pi = partial pressure of component i (atm),

Y, = mole fraction of component i,

and PT = total pressure (atm).

An analysis of values for Kp, and Kp2 at various temperatures indicates that

thermodynamically, the forward reaction in (1-5) is favored over the forward reaction in (1-4)

at low temperatures, with very little NO formed at temperatures below 1000 K. Hence, the








10

formation of NO2 is favored at low temperatures, but NO2 dissociates back to NO at higher

temperatures. The rate of equilibrium NO formation increases rapidly with increasing

temperature. As the combustion zone temperature rises above 1000 K, the formation of NO2

becomes less likely under equilibrium conditions. Kinetically, however, the NO formation

reaction is a slow one, which affects the availability of reactant to form NO2. NO formation

by reaction (1-1) has a very high activation energy (317 kilojoules/mole) and is most likely

the rate-controlling reaction (Ismagilov et al., 1990).

It has been observed experimentally that NO concentrations in the combustion flame

zone are significantly higher than those predicted by the Zeldovich mechanism. This "prompt"

NO formation is attributed to super-equilibrium radical concentrations that are likely in

hydrocarbon flames. It has been suggested that the intermediate, HCN, is formed when N2

reacts with a hydrocarbon radical. HCN combines with OH to form CN, and then CN is

oxidized to NO. Based upon experimental data, MacKinnon (1974) developed a model that

predicts the concentration of NO formed during combustion as a function of temperature, N2

and 02 concentrations, and time. At a total pressure of one atmosphere, the model equation

is:

CNO = 5.2 x 1017 (exp -72000/T) YN2(Yo2)12 t (1-8)

where: CNO = NO concentration (ppm),

T = absolute temperature (K),

Y = mole fraction of component i,

and t = time (seconds).








11
Consideration of equation (1-8) allows insight into the effect of temperature on NO

formation, particularly in the hot flame zone. Once exhaust gases move away from this zone,

they cool rapidly, reducing the reaction rates by orders of magnitude. If excess oxygen is

present as the gas cools, the conversion of NO to NO2 (reaction (1-5)) is favored.

Thermodynamics predicts that the cooled gas will consist primarily of NO2. In reality,

although the favored ambient form of NOX is NO2, flue gases from combustion contain

predominantly NO. Approximately 90 to 95 percent of NOx emitted from combustion

processes appears as NO, that is thermodynamically unstable in the environment as its

temperature drops. However, the decomposition of NO into N2 and 02 and the reaction of

NO with 02 to form NO2 are kinetically limited. Thus, the high concentrations of NO formed

at high temperature in the combustion zone are "frozen in" and are carried out with exhaust

gases into the atmosphere. From equation (1-8), therefore, it is seen that, with respect to the

reduction of thermal NOx, combustion control strategies could be developed that are aimed

at a reduction in peak temperature, a reduction in gas residence time at peak temperature, and

a reduction of oxygen concentration in the high-temperature zone (Cooper and Alley, 1994;

Wark and Warner, 1981; Organization for Economic Cooperation and Development, 1983).



Regulation of NOX Emissions



The federal primary air quality standard for NO2 is presently 0.05 ppm, based upon

an annual arithmetic mean. Emission and performance standards for various stationary

sources of NOx have also been established. Although the exhaust from jet engine test cells








12
(JETCs) is regulated for soot opacity, JETCs currently operate under implicit exemption from

NOx regulations, although they qualify as stationary sources. Ifjet engine testing operations

are conducted in an ozone non-attainment area or where photochemical smog is already a

significant problem, they become obvious targets for environmental regulation. Mobile

sources are regulated under Title II of the Clean Air Act Amendments of 1990. Part B of

these amendments addresses aircraft emission standards. A jet engine in a test cell could be

considered to be a stationary source under Title I, and may be included for regulation under

Title IV, Acid Deposition Control (Quarles and Lewis, 1990). In anticipation of these

possibilities, the Administrator of the EPA and the Secretary of Transportation, in concert

with the Secretary of Defense, have commissioned an investigation into the implications of

regulating jet engine test cells. Some of the issues studied included the impacts of not

controlling nitrogen oxides, the existence of appropriate control technologies, costs

associated with these technologies and their effects on safety, design, structure, and operation

or performance of aircraft engines and performance tests in test cells (EPA-453/R-94-068).

After the recommendations of this study are evaluated, NOx emission standards for jet engine

test cells may be promulgated and enforced in the near future. A review of the most common

NOx control methods for stationary sources is useful for the consideration of their

applicability to JETC exhaust control.












NO. Controls for Stationary Sources



Application of NO Control Strategies


Based upon well characterized chemical and thermodynamic principles that control

the formation of fuel and thermal NOx, several combustion modifications or changes to

operating conditions can be used for these purposes. In power generation and some waste

disposal applications, "low NOX" burners have been developed whose designs inhibit NO.

formation by controlling the mixing of fuel and air. While some research is being conducted

on jet engine designs to reduce NO production (Beard, 1990), this approach needs to be

evaluated for its applicability in light of strict and specialized military engine performance

requirements and standards. Combustion modifications can reduce NO formation by

lowering one or more of the parameters: peak temperature, gas residence time, or oxygen

concentration in the flame zone. It is difficult to employ these controls in a JETC application

due to the changing power settings associated with test and acceptance runs and associated

combustion exhaust fluctuations. Additionally, strict performance standards required during

engine testing limit the applicability of combustion modifications which may make test results

unrepresentative.










Modifications to Operating Conditions


Several modifications to stationary and mobile source operating conditions have been

developed that can be used to reduce NO, formation from combustion. These include 1) low-

excess-air firing; 2) off-stoichiometric combustion; 3) flue gas recirculation; 4) reduced air

preheat; 5) reduced firing rates; and 6) water injection. Excess air is the amount of air that

is in excess of the amount stoichiometrically required for 100 percent combustion of the fuel.

Due to imperfect mixing of air and fuel in the combustion zone, there must be some excess

air present at all times to reduce fuel waste and to prevent smoke formation. As excess air

decreases, NO. follows, while using less fuel. Off-stoichiometric combustion (also called

staged-combustion) burns the fuel in two or more steps. The initial flame zone is fuel-rich,

and the following zone(s) is (are) fuel-lean. Combustion with remaining air in the resulting

fuel-rich regions in the primary flame zone are controlled by heat transfer. Although the

overall air/fuel ratio is near stoichiometric, the primary NOx formation zone of the flame is

operated in a low-NOx condition.

Rerouting some of the flue gas back to the burner primary combustion zone is called

flue gas recirculation. This process not only reduces the peak flame temperature, but also

lowers the partial pressure of available oxygen at the burner, thereby decreasing NOQ

formation. Reducing the amount of combustion-air preheat lowers the peak temperature in

the primary combustion zone, decreasing thermal NO, production. Since JETCs operate

using ambient air as combustion air, this technique is obviously not applicable to these

operations. Likewise, reduced firing rates to reduce heat release per unit volume cannot be








15
implemented under the strict protocols ofjet engine testing. Water or steam injection into the

fuel, combustion air, or combustion zone is very effective at reducing NO. formation by

reducing combustion temperatures (Cooper and Alley, 1994). Introducing additives into jet

engine combustion zones or fuels, however, is a potential problem due to losses in output

(which is closely monitored during testing) or physical damage to the engine or turbine blades.

Some research has been conducted on other fuel or combustion zone additives as applied to

jet engine test cells, which will be discussed in the following sections.

Combustion Modifications


Combustion modifications can reduce NOX formation by lowering one or more of the

parameters peak temperature, gas residence time, or oxygen concentration in the flame zone.

Peak temperatures can be reduced by 1) using a fuel-rich primary flame zone; 2) increasing

the rate of flame cooling; or 3) decreasing the adiabatic flame temperature by dilution

(exhaust gas recirculation). Gas residence time in the primary flame zone can be reduced by

1) changing the shape of the flame zone; or 2) using the steps listed in the previous strategy.

Oxygen concentration in the flame zone can be reduced by 1) decreasing the overall excess

air rates; 2) controlling the mixing of fuel and air; and 3) using a fuel-rich primary flame zone

(Cooper and Alley, 1994). In general, it is difficult to employ these controls in a jet engine

test cell application due to the changing power settings associated with test and acceptance

runs and associated performance fluctuations. Additionally, the strict performance standards

required during military jet engine testing limit the applicability of combustion modifications.










Exhaust Gas Treatment


Exhaust gas treatment appears to be the most applicable method of controlling NOx

emissions due to the difficulty in using combustion/operations modifications with JETCs and

the priority placed on unperturbed engine performance during testing. The high volume of

exhaust from a military jet engine, combined with the need for higher levels of NOx reduction,

makes the selection of the optimal exhaust gas treatment a difficult task. There are two main

types of exhaust gas treatment processes, wet and dry processes. Wet processes typically

employ the absorption of NOx. The main disadvantage of wet absorption is the low solubility

of NO. Additionally, these processes often create liquid waste that may require proper

disposal at significant cost. For these reasons, dry processes are generally preferred. These

include selective or non-selective catalytic reduction, selective non-catalytic reduction,

irradiation, and dry sorption (Cooper and Alley, 1994).

In selective catalytic reduction, only the NO species are reduced (to N2 gas), leaving

free oxygen unreacted. With a suitable catalyst (typically a precious metal, although a variety

of other catalysts have been studied) ammonia (NH3), hydrogen (H2), carbon monoxide (CO),

or hydrogen sulfide (H2S) can be used as the reducing gas. NH3 is the most commonly used

gas. A number ofstoichiometries have been proposed for this reaction. For example, Davini

(1988) suggested

4NO + 4NH3 +02 4N2 + 6H20 (1-9)

Although the selective catalytic reduction process is very effective and is capable of

achieving NO reduction rates exceeding 90 percent in controlling steady-state processes, its








17

application to JETCs is expensive and involved compared to other treatments. At

temperatures from 900 to 1000 C, NH3 will reduce NO, to N2 without a catalyst; however,

NO. reduction efficiencies are only 40-60 percent. If the temperature is too low, unreacted

NH3 will be emitted ("ammonia slip"), and if the temperature is too high, NH3 can be oxidized

to NO. Thus both of these situations are obvious problems. Dry sorption techniques have

the advantage of simplicity of operation and minimization of waste effluent. Relatively high

NOX reduction rates have been documented (Nelson et al., 1989; Lyon, 1991).













CHAPTER 2
CONTROL OF NO, THROUGH GAS-SOLID INTERACTION



Exhaust-Gas Treatment Methods



A significant amount of research has been conducted over the past two decades to

determine a means of treating flue gases to reduce nitrogen oxide emissions. Some of the

most extensive research has been conducted in Japan, which was one of the first countries to

enact strict NOx emission regulations. In general, three classes of catalytic methods have been

evaluated. The first class involves mixing the waste gas with methane (CH4) or other gaseous

fuels/reducing agents before exposing the mixture to a catalyst; the second involves mixing

the effluent with ammonia before exposure to a catalyst; and the third involves exposure to

a catalyst or adsorbent with or without methane (or other fuel) or ammonia additions.

Numerous materials have been proposed as catalysts, and a number of these have been

patented, however, those receiving the most attention have been platinum or other precious

metals or metallic oxides from the alkaline earth group (Nelson et al., 1989; Klimisch and

Larson, 1975; Meubus, 1977).









Catalytic Methods


Cohn [U.S. Patent 3,118,727 (1964)] has been issued a patent that describes a

process for purifying waste gases containing nitrogen oxides by mixing them with a fuel (such

as CH4) and passing the mixture over a platinum- or rhodium-containing catalyst at an initial

reaction temperature of 690-780 F. Acres and Hutchings [U.S. Patent 3,806,582 (1974)]

describe a similar process. Childers, Ellis, and Ryan [U.S. Patent 2,910,343 (1959)] describe

a methane process involving two catalyst beds in series, one containing platinum or alumina

and the second containing nickel or alumina. Vanadium oxide, molybdenum oxide and/or

tungsten oxide catalysts on alumina or silicic acid substrates were used in combination with

CH4 as described by Nonnenmacher and Kartte [U.S. Patent 3,279,884 (1966)]. Such base

metal catalysts as iron, cobalt, nickel and copper dispersed on a refractory support were used

by Reitmeier [U.S. Patent 2,924,504 (1960)]. Other patents relating to the use of methane

and precious metal catalysts for the removal of NOx from waste or tail gases include those of

Andersen and Green [U.S. Patent 2,970,034 (1961)]; Romeo [U.S. Patent 3,425,803 (1969)];

Newman [U.S. Patent 3,467,492 (1969)]; Kandell and Nemes [U.S. Patent 3,567,367

(1971)]; Andersen, Romeo, and Green [U.S. Patent 3,098,712 (1963)]; and Hardison and

Barr [U.S. Patent 3,402,015 (1968)] (Nelson et al., 1989; Lewis, 1975).

The use of ammonia as a reducing gas in the presence of a catalyst was described

earlier. Baiker et al. (1987a, 1987b) and Chen et al. (1990) describe the selective catalytic

reduction of nitric oxide with ammonia upon catalysts comprised of mono- and multi-layers

of vanadia supported on titania, and mono-layers of vanadia immobilized on titania-silica








20

Davini (1988) studied the reduction of nitrogen oxide with ammonia in the presence of

carbonaceous soots from industrial boilers. Lee and Kline [U.S. Patent 3,864,451 (1975)]

describe a method for removing NO in the presence of sulfur dioxide by mixing ammonia with

the flue gas in the presence of a catalyst selected from the group of platinum and transition

metals and oxides and mixtures of these. Andersen and Keith [U.S. Patent 3,008,796 (1961)]

have described the use of NH3 in combination with cobalt, nickel, or iron supported on

alumina, silica, silica gel, or diatomaceous earth. Lyon (1987) describes an improved method

using ammonium injection in a tightly controlled noncatalytic process. A problem with this

method is the formation of ammonium salt particulate matter. Other patents relating to the

use of ammonia and precious metal or other catalysts for the removal of oxides of nitrogen

from a waste or tail gas include those of Cohn, Steele, and Andersen [U.S. Patent 2,975,025

(1961)] and Keith and Kenah [U.S. Patents 3,245,920 (1966); and 3,328,155 (1967)] (Nelson

et al., 1989).

Control of nitrogen oxides is also attainable by several processes that do not utilize

methane or ammonia, because the latter may be difficult to control. Additionally, the desired

ammonia reaction occurs only within a narrow range of temperatures. One process that

received great attention was the "RAPRENOx" selective reduction reaction first developed

by Perry and Siebers (1986). In this process, cyanuric acid (non-toxic) is mixed with NOx-

containing exhaust gases from which the NO, concentration (as NO) is significantly reduced

(NTIS Tech Notes, 1987). Heap et al. (1988) combined this technology with combustion

modification techniques to further improve the reduction of NO,. Unfortunately, it appears

that this reaction is negatively affected by the presence of free oxygen (Wicke et al., 1989).








21

It has been determined that various alkali metal oxides and silicate materials can reduce

nitrogen oxides without additional methane or ammonia. Harris, Morello, and Peters [U.S.

Patent 3,459,494 (1969)] have successfully evaluated CaO, SrO, BaO, K20, Na2O and others

supported on Alundum cement, porcelain, silica, extended alumina, or alumina beads. Heavy-

metal catalysts (copper, silver, nickel, molybdenum, palladium, and cobalt), supported on

alumina were patented by Ryason [U.S. Patent 3,454,355 (1969)]. The catalytic properties

of copper for the reduction of NOx without added methane or ammonia are described by

various investigators, including Gehri and Frevel [U.S. Patent 3,718,733 (1973)] and Kressley

[U.S. Patent 3,682,585 (1972)]. Kranc and Lutchko [U.S. Patent 3,576,596 (1971)]

employed a combination of copper and chromium impregnated on carbon supports to remove

both NO and CO from a waste gas.

Non-Catalytic Methods


Many problems associated with catalytic control methods often make the use of

non-catalytic methods more appealing. The most common problems with catalytic techniques

include catalyst fouling or poisoning, and a narrow range of operating temperatures. These

problems can be avoided by using non-catalytic controls. A number of dry sorption

techniques that do not employ catalysts have also been evaluated as they apply to the control

of nitrogen oxides. Ganz (1958) evaluated activated carbon, aluminosilicate, silica gel,

manganese dioxide, copper dioxide, and coke for the removal of high concentrations of NO2

and found aluminosilicate to be the most suitable sorbent. Kyollen [U.S. Patent 3,498,743

(1970)] described a process employing either sodium carbonate, calcium carbonate, or








22

calcium oxide, that was highly dependent upon temperature and humidity conditions.

Kitagawa [U.S. Patent 3,382,033 (1968)] developed a dry method that utilizes inorganic salts

such as PbSO4, KMn04, KC103, NaCIO, Na2MO4, K2S203, and Na2HPO4 and inorganic

oxides like AsOa and PbO2 as suitable adsorbents. Collins [U.S. Patent 3,674,429 (1972)]

derived a two-stage process involving silica gel and an activated crystalline zeolitic molecular

sieve. The use of activated carbon to adsorb nitrogen oxides has been studied extensively

(Badjai et al., 1958 (Wark and Warner, 1981)). Its high surface area, adsorption rate and

capacity, and regenerability are well known. James and Hughes (1977) used calcined

limestones and dolomites (MgO-containing) to remove NO at high concentrations with good

results. Kikkinides and Yang (1991) reported some success in the simultaneous sorption of

NO2 and SO2 using a weak-base macroreticular resin material.


Jet Engine Test Cells



With all of the previous research that has been conducted for application to industrial

processes and power generation equipment, relatively little has been done to specifically

address the control of NOX from jet engine test cells. In order to better understand the

problems associated with NO. control in jet engine test cells, it is useful to examine their

general structure, purpose and characteristics. The Department of Defense operates

approximately 180 cells, which are considered to be stationary sources by the U.S. EPA. The

purpose of a jet engine test cell is to provide a structure to evaluate engine performance

during controlled testing after maintenance to assure proper operation before returning the








23

jet engine to service. The engines are operated over their full range of thrust, representative

of typical operational modes. Total test times can range from less than an hour to two to

eight hours. The most common jet engine test cell design is shown in Figure 2-1. The engine

to be tested is mounted horizontally in the U- or L-shaped enclosure and combustion air is

drawn in through sound-deadening baffles. Unequal flow distributions are corrected by

turning vanes to provide an undistorted airflow at the engine outlet at a specified velocity

(often < 50 feet per second). Exhaust gases are blown into a large, long tube augmenterr

tube) with a convergent entrance section. The purposes of the augmenter are to draw air into

the test cell and engine with equal air pressure at the inlet and outlet of the engine (Note

venturi shape), to draw a portion of the air around the engine housing to provide cooling

similar to that experienced during flight, and to dilute and cool engine exhaust to prevent

damage to construction materials. Augmented gas temperatures can vary from about 200

to 2000 F (100-1100 C) depending upon engine firing rate and augmentation ratio.

Generally, the temperature of these gases is less than 500 F (260 C). The augmentation ratio

is varied by the placement of the engine relative to the augmenter throat. Some augmenter

tubes contain cooling water sprays to further quench the exhaust gases. Exhaust gases exit

the augmenter tube through a perforated basket, which helps to dissipate some jet engine

exhaust momentum as well as acoustical energy. In some cases, this basket can be adjusted

to reduce back pressure on the engine. Gases leaving the augmenter tube fill a blast room

before exiting the test cell through a stack. Cooled exhaust is vented to the atmosphere

through multiple channels within the stack to minimize noise (Johnson and Katz, 1990).

Depending upon the engine being tested, the volume flow rate of gas leaving the cell can be








Combustion and
., Augmentation Air



Intake Sound Treatment

Ol Exhaust Sound Treatment

Intake Baffles


Engine


Augmenter Tube


Figure 2-1. Schematic Representation of a Jet Engine Test Cell.


Exhaust








25

enormous. A unique feature of jet engine test cells is their highly variable test cycles (and

NOx production associated with varying peak flame temperature). Because of this variability,

a NOx control technology applicable to jet engine test cells must perform over a wide range

of rapidly varying operating conditions.


Research Sponsored by the U.S. Air Force (USAF)



The Air Force (through the USAF Engineering and Services Center and USAF Civil

Engineering Support Agency, now the Human Systems Center-Armstrong Laboratory,

Tyndall AFB, Florida) has recently sponsored seven independent projects that specifically

target the control ofNOx from JETCs. Johnson and Katz (1989) showed that returning of

exhaust with extra fuel was effective, but not economical. Ham et al. (1989) reported partial

success with non-catalytic reduction using additives (ammonia, amines, and hydrazines) to

simulated exhaust. Unfortunately, as temperature increased, NOx production increased. A

dual-bed mordenite-copper catalyst did offer some improvement in sensitivity to variable

conditions over some better known selective catalysts (Kittrell, 1991). Berman et al. (1991)

evaluated the photopromotion of NO thermal decomposition by various metallic oxides. This

decomposition was very effective, but only in the absence of oxygen. Initial work reported

by Petrik (1991) also found that NO, reduction by ceramic-supported electrocatalysts was

inhibited by even trace amounts of oxygen; however, continued development of these

materials has produced catalytic ceramics that may be effective inside the jet engine (Gordon,

1994). Lyon (1991) evaluated NO, reduction at the surface of heavier Group IIa oxides, but








26

the toxicity of these heavy metals is a detriment. The seventh study, which was selected for

development, involved vermiculite and magnesium oxide (MgO)-coated vermiculite. NO.

removal rates for this medium were appealing, and the technology has many other advantages

including economics, simplicity, regenerability, and particularly, insensitivity to changes in

exhaust gas conditions or composition (Nelson et al., 1989). The USAF has sponsored

numerous studies in order to quantify emissions from jet engines. Seitchek (1985), Fagin

(1988), and Spicer et al. (1990) have published jet engine emission values for many engines

in the USAF-inventory operating on JP-4 jet fuel. Similar studies are underway for emissions

associated with the combustion of JP-8 fuels. These documents describe NOX and other

combustion by-product emission rates correlated with power setting or operational

procedures. They are very useful for determining appropriate exhaust gas concentrations in

order to simulate jet engine exhaust gas composition in the laboratory. Many lab-scale studies

have used these data in their experimental design.

Vermiculite-based Catalyst


Vermiculite ((Mg,Fe(),Al)3(AI,Si)4010(OH)24H2O, a thermally expanded form of a

common aluminosilicate material) has been used as a carrier for other materials including

nitrates and phosphates, in which the combinations are used as slow-release fertilizers.

Evanshen [U.S. Patent 3,757,489 (1973)] describes the treatment of flue gases with

polyvinylpyrrolidone with or without the addition of a catalyst of a nitrate or sulfate of copper

or silver. He suggested suspending these materials on a carrier such as vermiculite. Sanitech,

Inc., (now Sorbent Technologies Corporation, Twinsburg, Ohio) developed a new class of








27

sorbents originally designed for the control of SO2, that were based upon MgO (an alkaline

earth [IHa] metal oxide) coated onto vermiculite or perlite [Nelson, U.S. Patents 4,721,582;

4,786,484, 1988; and 4,829,036 (1989)]. SO2 reduction is believed to occur primarily by

reaction with MgO, forming complex sulfites or sulfates. It was assumed that NOx could be

removed in a similar manner, forming nitrites or nitrates. Sanitech researchers began

evaluating expanded vermiculite as a catalyst in 1987. A patent describing the use of

vermiculite as a selective reduction catalyst [US Patent 4,806,320] was granted to Sanitech

in 1989.

In August 1988, the Air Force awarded Sanitech a Small Business Innovation

Research (SBIR) project to further develop the new catalyst for application to jet engine test

cells. Nelson et al. (1989) reported the results of a series of tests designed to establish the

applicability and technical feasibility of using vermiculite as a catalyst to reduce NOx from

simulated jet engine exhaust. These studies, beginning as a selective-catalytic approach,

evaluated how vermiculite variables (type and size), bed size, solid additives to vermiculite,

flue gas variables (temperature, composition, flow rate) and the addition of methane or

ammonia to the exhaust gases generally affected NO, reduction.

Some initial regeneration studies were also performed. These screening studies

indicated that a MgO-coated vermiculite material showed great promise as a NO1-reducing

agent for a wider variety of conditions than encompassed by commercial catalysts. Some of

the major conclusions of this study included data indicating that vermiculite alone, without

ammonia or methane additions, demonstrates the ability to reduce NO. by 50 to 73 percent

over the temperature range of 250-850 F. The use of gaseous ammonia only marginally








28

improved NO. reduction and increased the sensitivity of the process to temperature change.

Vermiculite coated with MgO removed approximately 70 percent of NOx and the process was

relatively insensitive to changes in waste gas composition, velocity and temperature.

It is believed that the NO. species were chemically adsorbed by the MgO/vermiculite

product and attached at the surface as complex magnesium nitrites and/or nitrates. Tanabe

and Fukuda (1974) found that MgO has a higher number of active (basic) sites compared to

other alkaline earth metal oxides (i.e., CaO, SrO and BaO). Zhang et al. (1994) further

attributed the activity of MgO surfaces to oxygen vacancies in the lattice structure. Some

selected physical properties of MgO, Mg(OH)2, and magnesium nitrite and nitrates are shown

in Table 2-1.

Table 2-1. Physical properties of selected Mg-containing inorganic compounds.

Species Mol Wt Density Mol Vol Melting Boiling
(g/g-mol) (g/cm3) (cmS/g- Point (C) Point (C)
mol)
MgO 40.30 3.58 11.26 2852 3600
Mg(OH)2 58.32 2.36 24.71 See Note -----
Mg(N02)2-3H20 170.36 ---------------------Decomp -----
100
Mg(N03)2-2H20 184.35 2.03 90.81 129 -----
Mg(N03)2-6H20 256.41 1.64 156.35 89 Decomp
330

Note: Loses water of hydration at 350 C.

Source: CRC Handbook of Chemistry and Physics, 73rd Edition, 1992.








29

Initial studies found that the sorbent material could be successfully regenerated with

minor reduction in adsorption capacity of the regenerated material. Ruch et al. (1990)

confirmed that the MgO-vermiculite/NOx reaction is most likely due to chemisorption and

that the vermiculite/NOx sorption is a result of physical adsorption based upon hysteresis

during limited adsorption/desorption tests. This observed hysteresis is most probably the

result of the forces of chemisorption, generally involving electron transfer between the solid

adsorbentt) and gas adsorbatee) molecules held at the surface. These forces are typically

much stronger than the intermolecular van der Waals forces associated with physical

adsorption (Kovach, 1978; Szekely et al., 1976). They also evaluated the surface areas of

these products using N2 and water adsorption and Brunauer-Emmett-Teller (BET) theory

and found the sorbent surface areas to be larger than would be expected from the collective

surface areas of the vermiculite and MgO. This suggests that the vermiculite support

effectively enhances the surface area development of the MgO. Reported nitrogen isotherm

surface areas ranged from 16 to 55 m2/g for sorbents prepared at 550 C. Sieved samples had

a maximum surface area of 36 m2/g. Fourier transform infrared (FTIR) solid spectra indicated

the presence of Mg(OH)2, which disappeared as conditioning temperature increased. Nitrite

was found in some samples exposed to low-concentration flue gases by FTIR solid and

solution spectra, but results were determined to be inconclusive. X-ray diffraction interplanar

spacings showed that Mg(N03)2 was present.










Field Studies


Pilot-scale tests of this MgO-vermiculite sorbent were conducted on a JETC facility

at Tyndall AFB, Panama City, Florida, between 9-11 June, 1992 and on 18 September, 1992.

A two-stage filter bed design, consisting of four inches of virgin vermiculite in front of eight

inches of MgO-coated vermiculite, was evaluated through four approximately 20 minute runs

of subscale drone engines. An additional run was conducted through four inches of activated

carbon, followed by the eight-inch MgO-vermiculite bed. NO, removal efficiencies for the

first four tests were approximately 50 percent. This increased to approximately 80 percent

with the addition of the activated carbon bed. Between June and September 1992,

approximately 150 unmonitored engine tests took place. When the September 1992 test was

conducted, NOX removal had decreased to 60 percent. Gas temperatures, velocities, and

pressure drop through the beds were also measured. The Air Force sponsor had imposed a

pressure drop limit of two inches of water on the bed, which was not exceeded. However,

approximately two-thirds of the volumetric flow was allowed to bypass the filter to prevent

overheating inside the JETC facility. Results showed that significant amounts of unburned

(non-methane) hydrocarbons (primarily ethylene) passed through the beds that did not contain

activated carbon (Nelson et al., 1992; Wander and Nelson, 1993).

More regeneration studies were conducted by Nelson et al. (1990) to attempt to

optimize the regeneration of MgO-vermiculite after "saturation" with NOx (and SO2). The

sorbents were alternately exposed to simulated jet engine exhaust and regenerated in reducing

atmospheres for a number of treatment cycles to evaluate NO, sorption capabilities.








31

Performance during sorption/desorption cycles was evaluated on the basis of weight gain and

loss. It appeared that there was a loss of surface area when the material was regenerated at

excessively high temperatures (> 800 C). Regeneration consisted of three steps: 1) drying;

2) heating to a temperature above 550 C (in the presence of N2, or CO or CH4 to reduce off-

gassing NO,); and 3) cooling. Of these, CO is the gas of choice. Heating during regeneration

was accomplished using a rotary-kiln calciner, a vertical tube apparatus, and a horizontal

conveyor kiln. Wickham and Koel (1988) have published an excellent review on the

reduction of NOX by CO.


Related Research



A review of current literature indicates that no research other than that already

described has been conducted on this MgO-vermiculite sorbent. As a result, no data on

intrinsic kinetics or reaction mechanisms) exist. Significant amounts of related research,

however, have been conducted on similar acid/acid anhydride gas-base/gas-solid systems.

The results of these studies have been used to design and construct appropriate experimental

systems to collect data for interpretation using proven and accepted sorption study

techniques. A short review of some of the fundamentals of gas-solid interactions and

chemisorption is necessary to better understand the complex phenomena involved in these

heterogeneous systems.











Gas-Solid Interaction



The smallest unit of a gas-solid system can be represented by the interaction between

a single solid particle and a flowing gas stream. This representation is convenient and simple,

and, in principle, may be generalized to more complex multi-particle systems, such as packed

beds. These interactions can be catalytic, in which the solid functions as a catalyst -- often

in the facilitation of a gas-phase reaction -- or non-catalytic, in which the surface acts to

physically or chemically remove reactants from the gas phase.

Adsorption occurs with a decrease in surface free energy and normally a decrease in

entropy; as a result, it is generally an exothermic process. In physical adsorption, sorbed

species are attracted and held at the solid surface by van der Waals dispersion forces. These

forces are much weaker than chemical bonding forces; the heat evolved is small, generally

between 1 and 10 kcal/g-mole (4-40 kJ/g-mole), approximately the heat of vapor

condensation for relatively non-polar substances. Chemisorption is primarily responsible for

gas-solid reactions and catalysis on solid surfaces. In chemisorption, the forces involved are

of the same order of magnitude as those for chemical bonding, and the heat of chemisorption

is often between 10 and 150 kcal/g-mole (40-600 kJ/g-mole). The process of physical

adsorption, like condensation, has very little or no activation energy and as a result is assumed

to occur rapidly. Chemisorption often displays a higher activation energy, implying a slow

rate; however, chemisorption reactions having small activation energies are known. Unlike

physisorption, chemisorption is specific to the gases and solids involved, just like chemical








33

reactions are. Chemisorption is also temperature dependent, occurring faster as temperature

increases; however, the amount chemisorbed at equilibrium decreases with temperature

(Hayward and Trapnell, 1964; Szekely et al., 1976).

Gas-solid reactions are inherently complex due to the number of simultaneous

processes that occur throughout the course of reaction. Any equilibrium conditions reached

are often dynamic rather than static in nature, since the system continues to change with time.

In addition to mass transfer and diffusion processes, these reactions often include heat transfer

and structural changes in the solid phase. Ideally, to properly describe a particular gas-solid

reaction, all of these simultaneous and consecutive processes should be taken into

consideration. In practicality, however, this would make for extremely complex mathematical

models. A good engineering model should be one that closely approximates the real situation

without too many mathematical complexities that will make it too cumbersome to use. The

use of simplifying assumptions and the careful design of experiments, to reduce or limit the

effects of variables on a given process so that it may be neglected, are often necessary to

produce useable data and models.


Generalized Non-Catalytic Gas-Solid Reactions



A generalized non-catalytic gas-solid reaction may be represented by the following

reaction scheme:


aA (gas) + bB (solid) -- cC (solid) + dD (gas)


(2-1)








34

where A and B are gaseous and solid reactants, respectively, C and D are products, and a,b,c,

and d are stoichiometric coefficients. This overall reaction may actually comprise several

sequential or simultaneous steps, of which one is rate-controlling. The three basic steps in

the generalized non-catalytic gas-solid diffusion-reaction process are

1) Gas-phase mass transfer via diffusion of the gaseous reactant from the bulk gas

stream to the external surface of the solid sorbent particle.

2) Gas-solid interaction on or within the solid sorbent particle through

a) Diffusion of the gaseous reactant into the pores of the solid complex,

which could consist of a combination of solid reactants and products.

b) Adsorption of the gaseous reactant on the surface of the solid complex.

c) Chemical reaction at the surface.

d) Desorption and diffusion of gaseous products, if any, from the surface and

out of the pores of the solid complex.

3) Gas-phase mass transfer of any gaseous products from the external solid surface

into the bulk gas stream

The number of diverse processes and steps involved makes the analysis of gas-solid

reactions a potentially unwieldy problem. In general, each step provides resistance to

chemical reaction, and these resistances are additive. When the processes occur in series, it

is necessary to determine which step provides the major resistance to reaction, and ultimately

controls the overall reaction rate. The resistances of the different steps can vary greatly from

each other, so the step with the highest resistance can be considered to be the rate-controlling








35

step. A simplified expression for the rate-determining step can then be used to describe the

overall process.

Three major processes are generally considered to be rate-controlling in non-catalytic

gas-solid reactions:

1) Mass transfer of the gaseous reactant from the bulk gas phase through the stagnant

film layer surrounding the particle to the surface of the solid sorbent. The rate of this

step is most dependent upon the fluid dynamics of the gas flow around the solid

particle.

2) Chemical kinetics of the reaction between the gaseous and solid reactants.

3) Mass transfer of the gaseous reactant or product by diffusion through any

product layer at the surface of the solid or in the internal pore structure of the solid.

The rate of this step is highly dependent on the extent of reaction and the physical

properties of the product layer or solid sorbent.


Mathematical Models



The classical Langmuir theory, which is based solely on available surface area, is often

insufficient to account for the formation of a solid product layer on the surface of the solid

reactant and associated surface changes. This theory assumes a uniform reactant surface,

with reaction ceasing after all surface area is covered (Laidler, 1987). Related non-linear

sorption models are not applicable to systems with porous sorbents where isotherms are

favorable and chemical reaction and pore diffusion resistances can be significant (Biyani and








36

Goochee, 1988). In general, gas-solid reactions can be described by an ideal Langmuir

isotherm only when there is exclusive and complete monolayer sorption, all active sites on the

sorbent are equivalent and there is no interaction between adjacent adsorbate molecules

(Comes et al., 1993). Solid structural changes resulting from the chemical change in the solid

can be quite complex, but are generally classified as being either sintering (and resultant pore

closure), swelling, softening, and cracking (Szekely et al., 1976). These changes may affect

diffusivities of gaseous reactants as well. A number of mathematical models based upon these

recognized physical and chemical phenomenon have been developed and described in the

literature.

Progressive-Conversion Model


This model is a simplified, idealized model for the non-catalytic reaction of particles

with a surrounding fluid. A basic assumption of this model is that the reactant gas enters and

reacts throughout the particle at all times, although probably at different rates at different

locations within the particle. This model obviously assumes a highly porous particle structure

to allow for easy reactant diffusion with little intraparticle resistance to diffusion. With

negligible diffusional (or mass transfer) resistance, the overall rate is controlled by chemical

kinetics, which is generally slow in this case.

Shrinking Unreacted-Core Model


Evidence from a wide variety of situations indicates that the progressive-conversion

model does not accurately approximate the behavior of real particles. Often, the reaction

produces an inert product layer of converted solid reactant which can produce significant








37

resistance to diffusion. A mathematically simple model, that was developed by Yagi and

Kunii in 1955, is often used to describe non-catalytic gas-solid reactions. This model, often

called the shrinking unreacted-core model, is based upon the visualization that the reaction

occurs first at the outer surface of an idealized spherical reactant particle (Yagi and Kunii,

1961). As reaction progresses, the reaction front (or zone) moves radially into the solid,

often leaving behind completely converted material in the form of an inert solid product called

"ash." A graphical representation of this model is shown in Figure 2-2.

The reaction steps in this model are the same as those previously described for a

generalized non-catalytic gas-solid reaction, occurring sequentially, although in any given

reaction, some steps may not occur. Effective diffusivity of the adsorbate through the solid

product is assumed to be a constant. Various equations relating the radial position of the

reaction zone (a measure of solid conversion) with reaction time have been developed. These

equations will be described in following sections before they are applied to the present study.

Although originally developed for application to non-porous solid reactants, the model has

been successfully used to describe systems utilizing porous solid sorbents. The main physical

difference between the two situations is that the reaction zone is diffuse in a porous solid as

opposed to a sharp interface in a non-porous one. A variable-diffusivity shrinking-core-model

was developed by Krishnan and Sotirchos (1993a). Their analysis revealed the strong effects

of reaction temperature, gas-phase concentration, and product layer thickness on effective

diffusivity of the reactant gas through the solid reactant and "inert" product layer matrix.










Stagnant Gas Film


Gaseous Reactant


Reaction


Gaseous Product


Progress of Reaction








Unreacted Sorbent Particle


Figure 2-2. Schematic representation of the shrinking unreacted-core model.










Grain Model

To account for sorbent porosity, Szekely and Evans (1970) first improved upon the

basic shrinking unreacted-core model by further defining the structure of the idealized

sorbent pellet or particle. The particle is assumed to be made up of compacting individual

grains of uniform size and shape. A schematic representation of the grain model is shown in

Figure 2-3. While in reality, this assumption may not be entirely accurate, the model presents

a means of accounting for the effects of particle and grain shapes on reaction rates. Each

grain still reacts following the shrinking unreacted-core model. A shape factor for one of

three idealized predominant shapes (either a sphere, cylinder, or flat plate) enters the relevant

shrinking unreacted-core model equations. This shape factor can be used to qualitatively

discern information about the sorbent surface physical structure as it affects sorption

behavior.

This model has been further modified by Hartman and Coughlin (1976), among others,

to account for increased diffusional resistance to chemical reaction in the product layer. This

effect is greatly compounded when the molar volume of the product is significantly larger than

that of the solid reactant, causing coverage or blockage of potentially available intergrain pore

surface area. This may cause a discrepancy between the theoretical solid-surface conversion-

versus-time behavior, as predicted by the model, and actual experimental results.

Ramachandran and Smith (1977) took the model a step further in modeling the single-pore

behavior in a porous pellet to predict the conversion versus time relationship for gas-solid

non-catalytic reactions. Their model accounts for the influence of pore diffusion, diffusion

through the product layer, and surface reaction. A key parameter useful in





















( --- I






Unreacted Sorbent Particle







Individual Grains





Figure 2-3. Schematic representation of the grain model.








41

describing the observed discrepancy between theoretical and observed solid conversion is the

effective diffusivity through the product layer. Incomplete conversion, which decreases with

an increase in intrapellet diffusional resistance, is predicted using the random pore model

developed by Bhatia and Perlmutter (1981a). A unique feature of this model is that solid

pores are not assumed to be of uniform size, so smaller pores are more easily blocked when

a dense or large-volume product is formed on the surface. Dam-Johansen et al. (1991) used

physical parameters of a chalk sorbent to modify the grain model down to the micrograin

level. While this provides a physical description of the reaction occurring at the sub-grain

level, the reaction taking place still follows the shrinking unreacted-core mechanism.



Analogous Gas-Solid Reaction System Research

The most similar gas-solid reaction systems discussed in current literature describe

the reactions between sulfur dioxide (SO2) and lime/calcium oxide (CaO), calcium hydroxide

(Ca(OH)2), or limestone (calcium carbonate (CaCO3)). Reactions of acidic gases with

calcined limestone have been closely studied because of their industrial importance. The high

porosities of various forms of CaO (nominally over 50 percent) makes them quite suitable for

tests of the numerous forms of the shrinking unreacted-core model. Although these models

have been used quite successfully, the fundamental rate-controlling processes are still not well

understood (Borgwardt et al., 1986). Data on reactions between other acid gas/acid

anhydride gases and other carbonate rocks are also available. In these processes, SO2-

containing gases are contacted with lime or limestone, either as a wet slurry or dry solid. SO2

is generally collected on the solid surface for reuse or disposal. Regeneration may produce








42

reuseable sulfuric acid or in some cases, elemental sulfur. A magnesium oxide process similar

to lime or limestone scrubbing is sometimes used where MgO is hydrated to Mg(OH)2 which

reacts with SO2 forming MgSO3/MgSO4 solids. These can be heated, generating SO2 and

MgO. The SO2 can normally be recovered as a sulfuric acid product (Cooper and Alley,

1994).

SO,-CaO Reaction


Borgwardt (1970) investigated the reaction of SO2 with calcined limestones and found

a first-order chemical reaction to be the predominant process with resistance limiting SO2

sorption by small particles. He found that the reaction rate decreased rapidly as CaO

conversion increased. Some of the solids contained significant quantities of MgO, which did

not readily react with SO2 at the elevated temperatures evaluated. Wen and Ishida (1973)

found, through an application of the grain model, that the overall reaction rate between SO2

and CaO was highly temperature dependent, particularly at lower temperatures where it was

controlled by chemical kinetics. They reported that various kinetic studies have indicated that

reaction rates may vary considerably (more than an order of magnitude) depending upon the

type of limestones used. This variability is a common problem inherent in the study of

heterogeneous sorbents.

The previously mentioned study by Hartman and Coughlin (1976), which improved

the applicability of the grain model, also described the SO2-CaO reaction. The strong

relationship between surface area and the reactivity of CaO toward SO2 was analyzed via the

shrinking unreacted-core model by Borgwardt and Bruce (1986). The surface area is








43

apparently very sensitive to the presence of water vapor, which reduces porosity (Borgwardt,

1989). A graphic representation of the shrinking unreacted-core model was depicted in a

series of energy dispersive X-ray analysis micrographs of calcined limestone chalk and SOz

(Dam-Johansen and Ostergaard, 1991a). The associated data were fitted with a grain model

of the shrinking unreacted-core mechanism (Dam-Johansen, 1991b). Bjerle et al. (1992)

reported useful experimental techniques for describing this reaction as well. They were able

to use the shrinking unreacted-core model to describe the various stages of reaction as

conversion progressed. Sotirchos and Zarkanitis (1992) found that the sulfation rates of MgO

were comparable to those for CaO. A simple grain model was also used to interpret data

describing a furnace sorbent slurry injection process, in which a limestone slurry was calcined

for use in SO2 removal (Damle, 1994).

SO,-Calcium Hydroxide Reaction


Flue-gas desufiurization through spray drying of a Ca(OH)2 slurry has recently become

an important alternative to the traditional wet lime or limestone scrubbing techniques. In

spray drying, SO2-containing flue gases are contacted in a dryer with finely atomized lime

slurry that absorbs the SO2. As water evaporates from the slurry droplets, reacted solids and

unreacted Ca(OH)2 remain and are collected on bag filters. This accumulated residual

Ca(OH)2 serves to remove considerable quantities of unreacted SOz from the inside of the

bags as well as in the ducts. Ruiz-Alsop and Rochelle (1988) used a bench-scale fixed-bed

reactor to study this reaction. The shrinking unreacted-core model was applied to their

results. Among the many factors evaluated, relative humidity of the gas was determined to








44
be the most important variable affecting sorption behavior. Chu and Rochelle (1989)

examined the simultaneous removal of SO2 and NO. (as NO) by Ca(OH)2 and in some cases

additives, including fly ash, CaSO3, and NaOH. These additives improved NO, removal.

Another bench-scale evaluation of the removal of SO2 and NO by Ca(OH)2 was conducted

by Jozewicz et al. (1990). NO was not very reactive toward Ca(OH)2 relative to SO2, but

adding Mg(OH)2 did improve NO removal somewhat.

SO2-Calcium Carbonate Reaction


Pigford and Sliger (1970) noted the effects of increased product layer diffuisional

resistance in the reaction between SO2 and CaCO3. The overall reaction rate was governed

by both diffusion of SO2 through a layer of solid reaction product, progressively formed on

the active solid surface, and diffusion of SO2 through pores in the solid. The direct sulfation

of CaCO3 was studied by Hajaligol et al. (1988) via thermogravimetric analyses and a bench-

scale fluidized bed setup. The product layer diffusion controlled shrinking unreacted-core

model was used to fitted to the experimental data, using a correction factor for product layer

volume, although chemical kinetics were still important to determining the overall reaction

rate. Snow et al. (1988) also evaluated the direct sulfation of limestone under conditions that

did not decompose CaCO3 to CaO. Their results were similar to those of Hajaligol et al.

(1988). As previously mentioned, a variable diffusivity shrinking unreacted-core model was

successfully applied to the direct sulfation of CaCO3 by Krishnan and Sotirchos (1993b).

Effective gas diffusivity through the solid product layer was found to be a strong function of

gas temperature and SO2 concentration in the bulk gas-phase.












Other Acid-gas-Calcium-based-solid Reactions


Other sorption studies employing the shrinking unreacted-core model and its

derivations have been conducted to study the control of other acid gases. The reaction of

hydrogen chloride (HC1) with CaO is of interest for the control of acid vapors emitted from

municipal and hazardous waste incineration operations. Gullett et al. (1992) used a

specialized fixed-bed reactor to study this reaction and determine the controlling mechanism

and kinetics. They examined the relative importance of chemical reaction and product layer

diffusion control using the combined resistance shrinking unreacted-core model. Their

research also provided data on how operating parameters like reaction temperature and gas-

phase concentration affect sorption. The strong resistance to chemical reaction provided by

solid-state diffusion was proved to follow a grain model by Weinell et al. (1992). Pakrasi

(1992) performed kinetic studies of HC1 removal by hydrated lime powder in a bench-scale

fixed-bed reactor. His studies included a thorough examination of the chemical kinetics and

possible mechanism associated with the HCl-CaO reaction. Relative humidity played a key

role in determining HCI removal rate and the extent of solid sorbent conversion. Other

relevant variables such as gas concentration and temperature were also examined. Incomplete

conversion (compared to that theoretically predicted stoichiometrically or by the shrinking

unreacted-core model), was the result of increased diffusional resistance in the product layer.

Both the shrinking unreacted-core model and an empirical model were used to interpret

observed conversion versus time data.












The process of coal gasification often produces significant quantities of hydrogen

sulfide (H2S). Limestones (either calcined or uncalcined) have been investigated as sorbents

for the removal of these gases through the direct sulfidation of the solids via

thermogravimetric analyses (Krishnan and Sotirchos, 1994). The shrinking unreacted-core

model was employed both quantitatively and qualitatively to derive kinetic parameters and a

to provide a description of the rate-limiting mass transfer processes. In this case, since the

product, calcium sulfide (CaS) is more porous and less voluminous than the CaCO3 reactant,

diffusional resistances were limited.


General Applicability of the Shrinking Unreacted-Core Model



From the previous discussion, it is readily seen that the shrinking unreacted-core

model and its derivations are useful tools for describing the complex chemical and physical

processes associated with gas-solid reactions. If the required physical parameters are known,

much detail about the physical rate-controlling mechanisms can be gleaned from a reaction

under study using these models. Combined with the traditional chemical kinetics principles

for determining intrinsic kinetic parameters, data can be collected which can ultimately be

used to improve the use of a gas-solid reaction to control gaseous pollutants.













Research Justification



No other research has been reported that would delineate the intrinsic chemical kinetic

parameters associated with the reactions of NOx with MgO-vermiculite sorbents. Likewise,

the probable reaction mechanism has not yet been determined, including the identification of

rate-controlling step(s). A chemico-physical model like the shrinking unreacted-core model

may be used to explain sorbent performance characteristics and provide useful information

related to the optimization of a process employing this novel sorbent for removing NOx from

jet engine test cell exhaust, or even other combustion sources. Reviewing the effects of other

operating parameters such as gas concentration, temperature, flow rate, moisture, and oxygen

can provide a more complete picture of the sorptive phenomena occurring. All of these will

be useful to the engineer or scientist who may attempt to employ this medium in an air

pollution control situation in the future.


Research Objectives



Based upon the justification for this basic research, the objectives of this study are to

1. Qualify and quantify N0O sorption, singly and in combination, by

MgO-vermiculite using appropriate isotherms. Determine appropriate kinetic parameters

associated with the sorption reactions. Determine the controlling mechanisms) and whether

these are chemical reaction, diffusion, or mass transfer limited.








48



2. Evaluate saturation characteristics (lifetime) of the material. Develop a model

to predict sorbent performance with time.

3. Evaluate the desorption and regeneration efficiencies of NOX on

MgO-vermiculite.

4. Evaluate the pressure-drop characteristics of the material over time as gas

collection progresses.













CHAPTER 3
EXPERIMENTAL METHODS AND MATERIALS



General Research Approach



A fixed-bed reactor system was designed, constructed and used to collect the relevant

data to meet the established research objectives. The present study systematically addressed

individual exhaust gas components, focusing primarily on NO,, under strictly controlled

conditions to delineate the nature of the sorption rates and mechanisms. Processes and rates

for removal of NO2 and NO by MgO-vermiculite sorbent were independently evaluated.

Moistened sorbent and/or humidified gases were used to evaluated the NO. removal

characteristics ofMg(OH)2. A variety of test conditions were used to determine important

operating parameters and limitations. By combining gaseous components, their interactions

in the presence of the sorbent material could be evaluated. Intrinsic and overall kinetic

parameters were determined. A complete description of the kinetics of gas-solid reactions

is very complicated, especially for systems involving solid products, for which the processes

at both the gas-solid interface and the reaction interface between the reactant and product

solids must be considered. Although the rate expressions can be quite complex, the reactions

between adsorbate and adsorbent can often be described by first-order kinetics (Szekely et

al., 1976). Data were mathematically interpreted using the shrinking unreacted-core model








50

and its variations, as well as empirically. These models provided a chemico-physical basis for

explaining sorbent performance and probable reaction mechanisms.


Experimental Variables



MgO-Vermiculite Reactive Sorbent Material.

This material (approximately 45% MgO to 55% vermiculite substrate by weight) was

prepared according to a process patented by the inventor (U.S. Patent 4,721,582, Sorbent

Technologies Corporation, (Sorbtech), Twinsburg, Ohio, 1988). Commercial-grade

magnesium oxide was used (98% volatile-free MgO) with a manufacturer-reported surface

area of 170 m`/g (Elastomag 170, Akrochem Corporation, Akron, Ohio). While the

process of preparing the sorbent is described in detail in the patent, a more general descriptive

process is presented here for informative and comparative purposes. Sorbent was prepared

in small batches, by necessity, in stainless steel pans in the laboratory. A quantity of coarse

vermiculite was weighed out in the pan, to which a 4:1 mass ratio of deionized water was

added. This is the maximum ratio allowed in the patent, but is greater than the 2:1 ratio

reported by Sorbtech in the preparation of a sample provided by them for experimental use.

Trial-and-error practice in the laboratory indicated that sufficient water was necessary to

attach the majority of the MgO particles to the surface of the vermiculite. Since the

vermiculite is highly hydrophilic, it readily absorbs the extra water. The sorbent is

subsequently heat-treated, during which process any excess water will be driven off. The

water and vermiculite were mixed to ensure complete absorption.








51

Magnesium oxide powder was prepared for addition to the mixture by sieving with

a 28-mesh Tyler Standard sieve (600 urngm opening). Forty-five grams MgO per 55 grams dry

vermiculite was added to the wet vermiculite while the mixture was manually stirred using a

stirring tool. This ratio was the same as that reported by Sorbtech and is less than the 60:40

maximum mass ratio described in the patent. The same ratio was used so that the two

sorbents could be compared. Sorbtech is able to prepare large batches of material using

mechanical stirring, perhaps allowing them to use less water, while more evenly coating the

sorbent with the Mg(OH)2 slurry.

After thorough mixing, the sorbent was allowed to air dry overnight until the

individual particles were again "free-flowing." The pans of sorbent were then placed in a

preheated muffle furnace set at 550 C and "conditioned" for 30 minutes. It is believed that

this process increases sorbent surface area by expanding the particles as a result of driving out

interior water vapor via heating. This process also dehydrates the Mg(OH)2 formed by

previously hydrating the MgO powder.

Since only small batches of MgO-coated vermiculite could be made in the laboratory,

this process had to be repeated a number of times to produce a sufficient quantity of sorbent

for experimental purposes. All of these individual batches were combined in an air-tight

polyethylene container and mixed together. While the preparation procedures were carefully

followed in making each batch, some variation could be expected in the sorbent from batch

to batch. The sorbent itself is inherently heterogeneous, however, so it was hoped that these

variations could be minimized by mixing the various batches. Samples of the combined batch

mixture were randomly selected for comparative surface area analyses as a measure of this








52

variation, as well as for comparison with data from the Sorbtech samples to ensure that the

sorbent was properly prepared.

The sorbent material was incorporated into a packed-bed arrangement in a

non-reactive 316 stainless steel tubular reactor in a controlled temperature (tube furnace)

environment, which will be described later. A single batch of material was used for all

experiments. Limited surface area/composition analyses of the material were performed to

establish baseline values. Comparative experiments were conducted using sorbent samples

provided by Sorbtech.

Temperature

Reaction temperatures were controlled using a tube furnace (Model 421135,

Thermolyne, Dubuque, Iowa) to contain the reaction vessel. Sorption study temperatures of

373,423, and 473 K were used. These temperatures were chosen because they represent a

reasonable range expected for augmented/cooled exhaust gases. Gases were preheated

before entering the furnace to ensure that they were at or above reaction temperature (+ 20

C) when they entered the bed, so they would not need additional heating.

Pressure

All runs were conducted at atmospheric pressure. The bed outlet gases were

exhausted to the atmosphere through a laboratory ventilation system. Pressure drop versus

exposure time was evaluated through the bed, using a U-tube manometer connected to ports

at the bed inlet and outlet. Pressure drop versus gas velocity was evaluated via an empirical

correlation.










Gas Composition


To quantify emissions from jet engines, the USAF has sponsored numerous studies

and has published jet engine emission values for many engines in the USAF inventory. These

documents describe NO, and other combustion by-product emissions correlated with power

setting or operational procedures (Spicer et al., 1990). These studies are very useful for

determining appropriate exhaust gas concentrations to simulate jet engine exhaust gas

composition in the laboratory. Many lab-scale studies have used these data in their

experimental design. Gas composition was kept constant for each run. Major components

of JETC exhaust were evaluated individually (mixed in N2 or air): NO (5-200 ppm), NO2

(20-200 ppm), and 02 (0-10%). Gas concentration ranges relate to representative values at

four common engine settings (nominally, idle, 30%, 75% and 100% power) for the USAF

Fl 10 turbine engine. Actual engine emissions data are shown in Table 3-1. Water vapor (3-

5% by volume at temperature) was evaluated in cohort with other gaseous components.

Water vapor may hydrate the MgO, forming a reactive hydroxide.

It is seen from this table that jet engine exhaust contains significant quantities of

unburned hydrocarbons, carbon monoxide (CO) and carbon dioxide (COz). Particulate matter

in the form of soot or condensation aerosols will also most likely be present in real exhaust.

All of these exhaust components have the potential to affect the NOx sorption performance

of the sorbent. For the purposes of this study, these variables were not included to allow for

the evaluation of the basic NO,-sorbent reactionss.










Table 3-1. USAF Fl 10 Turbine Engine Emissions Data (as reported)


Power THC NO (I
Setting (ppmC)

Idle 7 13.8

30 Percent 6 30

63 Percent 3 97

Intermediate 3.5 243
(High
Mach)
105% 335 21.5
Afterburner
-Augmented

Source: Spicer et al., 1990, page 27.


ppmv) NO (ppmv) CO (ppmv) C02 (%)


11.2

28

92

227



3.7


85

23

13

15



178


0.98

1.25

2.35

3.17



0.42


Gas Flow Rate(s)


Gas flow rates) were kept constant for each run. However, flow rates/gas velocities

were varied to produce two contact times of 0.5 and 1.0 second, keeping pressure drop to

a minimum. Contact times were calculated by dividing bed length by superficial gas velocity.

This is only an approximation of the true distribution of contact times since the actual gas

velocity may be based upon a more tortuous pathway through the packed bed volume,

increasing or decreasing the contact time of an individual reactant gas molecule. These

superficial velocities correspond to nominal space velocities of 3,600 and 7,200 per hour.

Since mass transfer limitations from the gas phase to the solid surface can obscure

intrinsic chemical kinetics, it is important to limit this resistance to chemical reaction. A

sufficiently high gas velocity will effectively decrease the thickness of the stagnant gas film








55

layer surrounding the sorbent particle. Practical constraints on pressure drop will likely limit

gas velocity in the application of this control technology to jet engine test cells. Comparative

experiments and mathematical evaluations were performed to ensure that this source of error

was effectively minimized. Results will be presented in a subsequent section.

Gas Reaction/Sorption/Desorption

Gas reaction/sorption was measured by the change in concentration of the test

gas(es) with bed residence and exposure times. Gases were sampled at the inlet and outlet

of the sorbent bed, and also from three points within the bed (of nominal 0.1-m length). Both

short-term (to dynamic equilibrium) and long-term results (taking the bed to "saturation," a

function of bed exposure time) were collected. Concentration-versus-time data from within

the bed were used to evaluate intrinsic chemical kinetics. Overall kinetic data were derived

from progressive removal efficiencies and were used to model overall sorbent performance

with increasing bed exposure times.


Fixed-Bed Reactor System



General Experimental Considerations


Many gas-solid reactions are studied using individual sorbent particles or pellets.

These individual particles are placed in a controlled environment and exposed to the adsorbate

of interest. Sorption behavior is determined gravimetrically using a microbalance and mass

balance calculations. An increase in sorbent mass is directly related to solid sorbent

conversion via stoichiometry. Some fundamental gas-solid studies employ either thin films








56

of solid sorbent reactant, or fine particles dispersed within an inert material bed such as quartz

silica sand or glass beads. While these types of studies are valuable for collecting data on

fundamental reactions occurring between the phases, in this study, by necessity, the use of the

prepared sorbent was a fundamental limitation. Experimental results may or may not be

transferable to the basic reactions between NO. gases and MgO or Mg(OH)2 solids. It is

possible that the process of making the sorbent promotes sorption through the enhancement

of available surface area. It is also possible, however, that the inefficient transfer of Mg(OH)2

slurry to the vermiculite substrate before calcination may create dense agglomerates or

otherwise make MgO unavailable for reaction with NO.. The effects on ultimate sorbent

conversion and bed utilization will be discussed later.

Often, particularly in chemical or materials engineering studies of gas-solid reactions,

the emphasis is on the solid sorbent surface and its conversion as reaction progresses.

Experiments are designed for the purpose of optimizing the solid sorbent itself Solid

conversion is generally measured directly through chemical analyses, including spectroscopic

and chromatographic methods. Physical properties of the sorbent can be evaluated using X-

ray diffraction or electron microscopy, for example. While these previously discussed studies

are invaluable, in environmental engineering and particularly, in air pollution control, the

focus is often on the gas-phase concentration and the effective removal of the pollutant of

interest. This applied engineering approach is the result of the unavoidable fact that

environmental regulations mandate the reduction of emissions to specified limits and

compliance with regulations is determined through air sampling. It is still crucial, however,

that gas-phase concentrations be correlated with solid conversion through an appropriate








57

chemico-physical model to adequately explain the removal of gaseous contaminants. A means

for accomplishing this correlation using local equilibrium theory and assumptions regarding

constant-pattern behavior for mass transfer in a fixed-bed will be detailed later.

The basic principles that govern chemical kinetics and diffusion phenomena in single

particles are normally unaffected by the presence of other particles in an arrangement like a

packed bed. Changes in the flow field due to the packed bed arrangement likewise do not

invalidate these principles (Szekely et al., 1976). In this way, a fixed-bed experimental

system may be used to extend the information obtainable from single particles to multiple-

particle systems. Multiple-particle systems are obviously of more interest since practical

systems must be of this type. Since the sorbent material is highly heterogeneous, data

reproducibility between samples is very important. This was evaluated by comparing results

for the same experimental conditions between different samples of sorbent material.

While the basic principles of single gas-solid interaction will still apply, the physical

characteristics of any experimental fixed-bed reactor can markedly affect chemical kinetics

and overall sorption behavior. It is critical that such a system be carefully designed, to ensure

that any experimental data collected are scientifically valid and defendable. Often, practical

constraints such as temperature and, particularly, gas velocity and its effect on pressure drop

through the bed, must be incorporated into the experimental design. Such other factors as

particle fluidization velocity and the potential for temperature and concentration gradients

must be accounted for as well. In this research, MgO will be attached to its vermiculite

substrate. Results, however, may be different from those for individual MgO particles due

to this arrangement.











Experimental Arrangement




The experimental arrangement for sorption and desorption/regeneration employed in

this study is shown in Figure 3-1. Nitrogen oxides (mixed in N2) were supplied from separate

certified standard cylinders (Bi-Tec Southeast, Inc., Tampa, Florida) to ensure reliability of

concentrations and to minimize contaminant concentrations. Initially, NO2 mixtures were

generated using a NOx generator (Model 100, Thermo Environmental Instruments, Franklin,

Massachusetts) as a source of NOz. This generator produces NO2 by mixing ozone generated

from UV irradiation of air with NO. While the system could easily and efficiently convert

high concentrations of NO to NO2, two major problems resulted from its use. First, it is

extremely difficult to perfectly titrate ozone with NO using this system, without leaving

excess ozone. The flow controls in the unit are insufficient for this purpose. A related but

more important variation was the ultraviolet lamp output used to generate ozone appeared

to vary from one test to the next for a given voltage setting. The presence of ozone was

unwanted in the experimental study since it could complicate data interpretation by oxidizing

NOx to other species not readily measurable with the instrumentation used.

As a result, it was necessary to use standard cylinder gas as an NO2 source. Since this

is actually N204 under pressure, there is a small potential for the gas to decay with time.

While total NOX was the same as certified, some NO "contamination" is generally present in

the cylinder gas. The NO contamination in the NO2 cylinders was always less than two to five

percent. Similar impurity values were noted in the NO2/NO ratio in the NO cylinders. This









To Vent





Data Logger Analyzer(s)





PC ~~-- ----------


Rotameters

t -- Mixing ]
Chamber Heating





So
Stainless

I I . I .. . . .


IN H20
-N---------------------
Air NO-N2 N02-N2 N2 Heating

Figure 3-1. Experimental arrangement for packed-bed studies.


Coalescing
Filter


rbent in
Steel/Furnace








60

is a problem with attempting to use "pure" NO2 (or NO) in gas-solid sorption studies

employing cylinder gases. These gases were diluted with dry air or prepurified N2 to the

appropriate concentrationss. Cylinder gas flow rates were measured using precision

rotameters (Omega Engineering, Inc., Stamford, Connecticut) with only glass and/or stainless

steel wetted parts. While the manufacturer provided calibration data, all rotameters (both

high- and low-flow-rate tubes) were independently calibrated in the laboratory using a primary

airflow standard (Gilibrator, Gilian Instrument Corporation, Wayne, New Jersey). These

calibrations were extremely important for accurate gas flow rate measurements, particularly

when using low-flow-rate tubes, which have non-linear calibration curves. Good agreement

was reached between the calibration curves, as can be seen in the example comparative

calibration curves, Figures 3-2 and 3-3. Laboratory calibration curves were used to measure

gas flow rates in this study.

All system components were made of either 316 stainless steel or Pyrex TM glass to

reduce the potential for NOx removal by system reactivity. Inert gas transfer lines were made

from Teflon TM tubing of minimal lengths to minimize transfer losses. Most connections were

through 316 stainless steel compression fittings (Swagelok, Inc., Solon, Ohio). Where pipe

thread connections were necessary, Teflon TM tape was used to prevent leakage. Before

experiments were initiated, the entire system was leak-checked and NO, decay (reactivity)

measurements were made for the system. Each system component was isolated and evaluated

separately. Overall NO and NO2 removal as a result of system reactivity were estimated at

less than one percent. In this way, any changes in NOx concentration with time could be

attributed solely to the presence of the sorbent material in the fixed-bed reactor.










Air Flow (mL/min)
100


80 ... .


60


40


20 -


0 I I I I
0 20 40 60 80 100 120 140 160
Rotameter Reading

Fgr3- Omega Data l1 Lab Calibration
Figure 3-2. Example low-flow-rate rotameter calibration curves (Omega N042-15ST Tube).










Air Flow (mL/min)
2500


20 0 0 . .


1500


1 0 0 0 -.......


5 0 0 -................. ...


0 --I I I I I I
0 20 40 60 80 100 120 140 160
Rotameter Reading

-- Omega Data -- Lab Calibration
Figure 3-3. Example high-flow-rate rotameter calibration curves (Omega N092-04G Tube).










Tubular Fixed-bed Reactors

The reactors, and mixing and sampling manifolds were machined from 316 stainless

steel seamless pipe stock (2.54 cm ID) (T.M.R. Engineering, Micanopy, Florida). Schematic

representations of the reactors, sample lines, and mixing and sampling manifolds are shown

in Figures 3-4, 3-5, and 3-6, respectively. While it is often standard practice and may be

sufficient to use only inlet and outlet concentrations from a fixed-bed to determine overall

conversion efficiency, this is obviously inadequate for the determination of intrinsic chemical

kinetics. For this reason, the reactors were designed to allow for the collection of samples

from three points within the packed bed. Using data collected from these points, a more

accurate depiction of the concentration versus time profile within the bed can be discerned.

This leads to the calculation of valid intrinsic chemical kinetic parameters. An Arrhenius-type

expression was used to determine corresponding activation energies.

Sampling lines were connected to the reactor via permanently welded 316 stainless

steel compression fittings (Swagelok, Inc., Solon, Ohio) and could be easily removed between

runs. One-eighth inch diameter internal sampling lines were used to minimize the cross-

sectional area blocked by the lines which could disrupt flow through the bed. The fractional

area blocked by the cross-section of a sampling line was approximately 16 percent. A total

of 12 holes (0.1 cm diameter) were drilled in the sampling lines. As shown in Figure 3-5,

these holes were placed in three places, located at the centroids of two concentric circles of

equal area within the reactor diameter. Each sampling location was from three sets of two

pairs of holes, the latter offset by 90 from each other.












Packed Bed Region


2.5 cm 2.5 cm 2.5 cm 2.5 cm




Sampling Tube Connection Points
316SS Screel (D 16SS Screen

Thermocouple Attachment Screw


316 SS Fixed-bed Reactor Schematic (Not to Scale)
Total Length Approximately 47 cm.


Figure 3-4. Schematic representation of 316 stainless steel packed-bed reactor (end connections not shown).














Connection to Sample Valve/Line---h*


Reactor Cross Section


(2.54 cm ID)


Locations of 1 mm holes
Each set 4 holes
Spaced 11 mm apart



. 1.7 mm from reactor wall
(End of tube)


Figure 3-5. Schematic representation of internal sampling tube location and appearance.


















Mixing or Sampling Line Attachment Points


316 SS Mixing or Sampling Manifold
Total Length Approximately 47 cm.







Figure 3-6. Schematic representation of 316 stainless steel mixing or sampling manifold.


r!F








67

This arrangement was chosen to ensure that a representative sample of gas was collected

from any given location within the bed, avoiding the potential bias caused by any flow

disturbances through the packed bed, or by reactor wall effects. The number of holes and

perpendicular placement were intended to prevent inconsistencies due to hole blockage during

the course of a run, although this did not appear to be a problem. A thermocouple attachment

screw mounting was placed on the outside of the reactors) to allow for the independent

measurement of reaction temperature for comparison with the indicated furnace temperature.

The mixing and sampling manifolds were simply sections of 316 stainless steel pipe,

approximately 70 cm long (28 inches), with six permanently emplaced 316 stainless steel

compression fittings for the connection of gas transfer lines and/or valves, as appropriate.

Before use, machined parts were cleaned and degreased with methanol, followed by

acetone and a deionized water rinse. Cleaning all parts with deionized water between runs

was sufficient during the normal conduct of experiments. For a nominal bed length of 10 cm,

the bulk volume of a packed bed was approximately 51 cm3. Individual sorbent samples were

measured out using a graduated cylinder filled to a loosely packed volume of approximately

51 cm3. These samples were transferred to preweighed sealable polyethylene sample

containers and weighed (approximate mass 7-7.5 g/sample). Since MgO comprised

approximately 45% of the total sorbent by mass, each bed contained approximately 3 g MgO.

After sampling lines were inserted into the reactor, the bed was packed between circular 316

stainless steel screens (12x12 mesh) at the center of the clean tube using a marked steel ram

rod to ensure that the bed was properly located. End connections were replaced and a Type

K (Nickel-Chromium versus Nickel-Aluminum) thermocouple (Model XCIB-4-4-3, Omega












Engineering, Inc., Stamford, Connecticut) and meter were attached to the outside of the

reactor.


Experimental Procedures



The packed reactor was placed inside the tube furnace with the bed centered within

the furnace and the tube element mounted vertically on the furnace control base. The vertical

arrangement was chosen to optimize gas-solid contact by avoiding the potential for solid

settling and channeling of the gas stream through the bed volume. The middle 15 cm of the

furnace was reported to be completely temperature-controlled by the manufacturer, providing

a sufficient length for isothermal experimental conditions. Once the reactor was in place,

sampling lines were individually connected to two-way stainless steel valves (Model W- 188,

Whitey Co., Highland Heights, Ohio) at the sampling manifold. Gas lines entering the tube

furnace were wrapped in high-temperature heating tapes. These had been previously

calibrated using a Type K thermocouple to measure surface temperature at various settings

on the variable voltage controller. Depending on the reaction temperature of interest, the heat

tape voltage controller was set to produce a gas preheat temperature approximately 20 C

above reaction temperature. The tube furnace was turned on and allowed to heat up for

approximately one hour to achieve temperature control, while the bed was simultaneously

being flushed with approximately 2 sLpm dry N2. After this conditioning period,

thermocouple readings for the tube furnace and independent reactor thermocouple were








69

compared and averaged to determine actual experimental reaction temperature. This reaction

temperature was used to calculate the actual flow rate through the bed, which could then be

used to determine residence time at any given sampling point within the bed. It was assumed

that the sorbent bed was at the same temperature as the reactor surface after sufficient

equilibration time.

When gas humidification was used, a secondary preheating heat tape was used before

the gases entered the humidification vessel, which consisted of an impinger apparatus

mounted inside a modified 3-liter PyrexTM flask (Southern Scientific, Inc., Micanopy, Florida)

containing deionized water. The flask was placed in a heated flask mantle connected to a

separate voltage controller. Various temperature/flow rate combinations were evaluated to

allow for 3-5% water vapor to be picked up by the flowing gas stream. The humidified gases

were then reheated to ensure that they were approximately 20 C above reaction temperature.

The actual water vapor content of the gases was determined by weighing collected condensed

water from a stainless steel coalescing filter (Model 41S6, Balston, Inc., Haverhill,

Massachusetts) at the bed outlet. Water vapor was removed from the sample gas stream

before analysis through condensation in the sampling manifold. Gas flow rate was always

corrected for the presence of water vapor. The effect of moisture on reaction was sometimes

evaluated simply by adding sufficient deionized water to a virgin bed to convert all MgO into

Mg(OH)2. This was verified by the presence of water vapor at the bed outlet. Once

converted, Mg(OH)2 is stable over the range of temperatures evaluated (Sidgwick, 1952).











Gas Analyses



All reactions were monitored by selectively removing gas samples from the stainless

steel tubular reaction vessel through selected valves connected to the sampling manifold

system. NO, concentrations were analyzed using a chemiluminescent analyzer (Model 42H,

Thermo Environmental Instruments, Franklin, Massachusetts) based upon the reaction of NO

with 03, following reduction of NO2 to NO. The feasibility of the commercial

chemiluminescent analyzer was established in 1970 by Fontijn et al. and the first prototype

was developed in 1972 (Steffenson and Stedman, 1974). The method is now widely used,

in fact, the use of a chemiluminescent method is required for the measurement of NO2 in the

atmosphere (40 CFR, Part 50, Appendix F). The principle of operation of the

chemiluminescent analyzer is based upon the chemiluminescent reaction between NO and

ozone (03). This reaction is as follows:

NO+03- [ NO2*+02] N02+02+ hv (3-1)

where hv, which represents the quantity of light energy emitted by an electronically excited

intermediary (N02*), is simply Planck's constant (h) multiplied by the frequency of the

emitted radiation (v), in this case with a wavelength greater than 600 nm. A sample of gas

is drawn from a selected point in the system by an external pump. The sample is mixed with

ozone generated by an internal ozonator inside a reaction chamber. The resultant reaction

produces a characteristic luminescence with an intensity directly proportional to NO

concentration, since 03 is provided in excess, and its concentration can be considered a










constant in the reaction. This light energy is detected by a photomultiplier tube, which

generates an electronic signal that is then converted to a NO concentration reading.

Nitrogen dioxide is measured by difference through its thermal conversion into NO.

A solenoid valve is used to switch to the NOX mode, where all NO2 in the gas sample is

converted to NO, based upon the thermodynamic principles previously explained, and is

combined with NO originally in the sample. In the NO mode, gas does not pass through the

thermal converter and therefore contains both NO and NO2, with only the NO capable of

reacting with ozone to produce light energy. Thus, the NO2 concentration of the sample is

determined by subtracting the signal from the NO mode from the larger signal obtained in the

NOx mode since:

[NO.] [NO] = [NO2 ] (3-2)

The instrument specifications, as reported in the manufacturer's manual, indicate

several scales providing an overall measurement range from zero to 5,000 ppm, with a limit

of detection of 50 ppb (0.050 ppm). Background noise level is 25 ppb (0.025 ppm), well

below the detection limit. The instrument exhibits linearity in response of + 1% of full scale,

with a zero drift (24-hour maximum) of 50 ppb (0.05 ppm), and a span drift (24-hour

maximum) of+ 1% of fulil scale. The separate scales are linear over their entire ranges, which

was verified by multiple-point calibrations with certified gases. While the interference of

HNO3 and HONO with NO2 measurements in chemiluminescent analyzers is well documented

(Spicer et al., 1994; Joseph and Spicer, 1978), the concentrations of these interferents were

expected to be low in the system employed. When sufficient moisture was present to form








72

acids, they were ultimately collected in the coalescing filter or condensed out in the sampling

manifold prior to analysis.

Similarly, the presence of water vapor itself can interfere with NOx measurements

directly by NO2 absorption, as well as via third-body quenching of chemiluminescence

(Trdona et al., 1988; Campbell et al., 1982). Therefore, it is important to minimize contact-

time of the gas with water vapor in sampling lines. Any NO2 removed by gas humidification

is accounted for by the fact that bed inlet gases were sampled after this point in the system.

The analyzer reaction chamber is maintained at a vacuum of approximately 29 inches of

mercury (98 kPa) to minimize third-body quenching effects. Another common interference

with NOx measurements in the atmosphere is caused by the presence of such other nitrogen-

containing compounds as peroxyacetyl nitrate (PAN), which can be converted into NO in the

thermal converter. Again, these species are not suspected to be present in the experimental

procedures used in this study.

To prevent particulate matter contamination of the chemiluminescent analyzer, a 0.5

grm pore size TeflonTM filter (Gelman Sciences, Ann Arbor, Michigan) was placed in the

sample line at the analyzer sample inlet. This filter was found to be essentially non-reactive

to NOx. Without such a filter, it is quite easy to experience blockage of internal capillary

tubing. Deposits could also form on the glass filter that separates the reaction chamber from

the photomultiplier tube, causing a loss in instrument sensitivity (Klapheck and Winkler,

1985).

A continuous gas analyzer (Model IR-2100, Infrared Industries, Inc., Santa Barbara,

California) was used to electrochemically measure 02 concentrations. The principle of








73

operation upon which this instrument is based is a simple coulometric process in which

oxygen in the sample gas stream is reduced in an electrochemical cell. The gas stream enters

the cathode cavity, with any oxygen being metered to the cathode through a diffusion barrier.

At the cathode, oxygen is electrochemically reduced via the reaction

02+ 2H20zO + 4e 401H (3-3)

An electrolyte solution in the cathode cavity contains potassium hydroxide (KOH), which

facilitates migration of the generated hydroxyl ions to an anode, where they are oxidized back

to oxygen. This occurs following the reaction

40H" 02+ 21H20 + 4e (3-4)

The resulting cell current is directly proportional to the oxygen concentration in the sample

gas stream. This current is measured electronically and is converted into an indicated

concentration value. The instrument measures oxygen concentration between zero and 25%,

with a manufacturer-reported accuracy of+ 2% of full-scale on all ranges. When using this

instrument, one must keep the sample inlet pressure and flow rate within the specified limits

to ensure accuracy of the measurement.

Multiple-point calibrations were performed on all instruments following procedures

outlined in the appropriate instrument manuals. Certified standard gases were used as transfer

standards whose composition is traceable to the National Institute of Standards and

Technology, and appropriate calibration curves were determined by challenging the

instruments with known standard mixtures. Typical calibration curves for the N0O and 02

analyzers are shown in Figures 3-7 and 3-8, respectively. Daily zero and span checks were

performed on each analyzer, and quality control charts were produced. Running sample









Indicated [NO] (ppm)


0 100 200 300 400


Standard [NO]


(ppm)


-- Lab Conditions
Figure 3-7. Typical chemiluminescent NO. analyzer calibration curve (for NO channel).


500


400


300


200


100


0


500










Ln Instrument [02] (Percent)


1 1.5 2 2.5 3 3.5
Ln [1021 (Percent)


+ Mid Scale High Scale
Figure 3-8. Typical IR-2100 oxygen analyzer calibration curve.


3.5

3

2.5

2

1.5

1

0.5

0


0.5











means and standard deviations were shown on the control charts. Based upon control chart

results, instruments were recalibrated when necessary. Quality assurance was maintained by

showing that all quality control systems were in good order.

As the NO/NOx chemiluminescent analyzer has a flow bypass system installed to

ensure that samples are always collected at ambient pressure, all experimental measurements

were made at the same bypass flow rate used during the initial calibration. Failure to do so

can result in gas concentration measurement errors as great as 10 percent from one end of

the bypass flow scale to the other. Additionally, a NO, generator (Model 100, Thermo

Environmental Instruments, Franklin, Massachusetts) was used to determine the actual

analyzer NOx converter efficiency and ensure that the efficiency exceeded 95%.



Sorbent Surface Area Determination



In preparation for packed-bed studies, BET surface areas of selected samples of

laboratory-prepared MgO-vermiculite sorbent, Sorbtech-supplied sorbent, and raw coarse

vermiculite were determined using a Micromeretics ASAP 2000 automated surface-area and

pore size analyzer. Duplicate samples were run to verify the validity of the data. Small

samples of sorbent (approximately one gram) were first degassed at an absolute pressure of

10 to 15 mm Hg at 623 K for approximately eight hours. Nitrogen gas (in helium) was

sequentially added to the sample tube at a temperature of 77 K. The volumes of nitrogen

adsorbed at various partial pressures between approximately 0.1 and 0.3 atmosphere were








77

recorded by the analyzer. These data were plotted according to the BET model to determine

the total surface area of the sample. Initial screening results were used as an indication of the

best preparation method to use in making the sorbent. Based upon these results, a large

quantity of sorbent was prepared following the procedure already described. Five additional

samples were taken and analyzed using the BET method for precision estimates.


Determination of Sorbent Particle Size



The MgO-vermiculite sorbent is inherently a highly heterogeneous material. As a

result, the range of particle sizes in a given sorbent sample can be great. Samples of

MgO-vermiculite sorbent were sieved using a standard ASTM sieving apparatus. Cumulative

mass collected at each stage was plotted against particle diameter using a log-probability chart

to determine the mass median particle diameter. This median size was used in all calculations

for which sorbent particle diameter was required to determine flow characteristics through

the packed bed. It was noted that this value may or may not be the most accurate value, since

sieving the sorbent material likely causes a decrease in particle size through mechanical

abrasion. The true size is probably somewhat larger. Data for magnesium oxide powder

particle size was plotted similarly using a manufacturer-provided mass distribution. Once the

MgO powder has been applied to the surface of the vermiculite, any agglomeration would

tend to create larger size particles. It is not known, however, how the application of

Mg(OH)2 slurry affects particle size. Results of these determinations are presented in the

following chapter.













CHAPTER 4
RESULTS AND DISCUSSION



Intrinsic Kinetic Studies



Experiments were conducted to collect data that provide insights into relevant

parameters controlling NO. removal efficiencies, potential reaction mechanisms) and

reaction kinetics. Duplicate runs were conducted for every set of conditions to generate the

data necessary for the determination of kinetic parameters. A comparison of results between

duplicate runs under similar conditions showed remarkable reproducibility (statistical error

< 5%). To determine intrinsic reaction kinetics, NOx concentrations were followed with

increasing bed exposure time from a selected sampling valve, which represented a specific bed

residence time. This was necessary because of a system artifact that created a lag-time of

approximately 3-4 minutes to allow for line flushing until a stable instrument reading was

reached after a sampling valve was selected. To explain further, for a given set of conditions,

four separate runs would be necessary to collect data for all five sampling points; bed inlet,

three points within the bed, and bed outlet. Bed inlet values were checked on all runs, before

and after the other data were collected. Every 10 seconds, the three NO, analyzer channels

displayed 10-second average NO, NO2, and NO, concentration values ([NO2] calculated by

difference). Data were collected every minute throughout the run. From these data,








79

concentration versus bed residence time curves could be plotted. Values for the same bed

exposure time were used to plot concentration of interest versus residence time, which was

a function of sampling point location and, thus, reaction time.


Limitation of Gas Film Mass Transfer Resistance



In the conduct of intrinsic chemical kinetic studies of gas-solid reactions in packed-

bed reactor systems, it is critical that gas film mass transfer resistance to reaction be

minimized. The following description of the limitation of this resistance is based upon the

work of Fogler (1992) and Szekely et al. (1976). The mass flux from a fluid to a solid

surface is the product of a mass transfer coefficient and the concentration driving force.

When gas film mass transfer is rate-limiting, the extent of reaction or fractional solid

conversion is linear with respect to time. In a packed-bed system, the concentration driving

force is the difference between reactant concentration at the void space at a point in the bed

and the concentration at the surface of the sorbent particle adjacent to this point. Gas film

mass transfer resistance can generally be minimized by using small sorbent particles and high

superficial gas velocities. At low velocities, the stagnant gas film boundary layer thickness

is large -- at high velocities this thickness is decreased, reducing the mass transfer limitation

through increased convective mass transfer.

Operating at lower reaction temperatures also helps to reduce gas film mass transfer

limitations. When a reaction is gas film mass transfer limited, it will generally be insensitive

to increases in temperature. Since gas film mass transfer is not an activated process, while








80

chemical reactions can have substantial activation energies, operating at lower temperatures

should lead to control by chemical kinetics, making intrinsic kinetic data collected in this

range more valid. The effects of gas film mass transfer resistance were evaluated by varying

flow rate through the bed at the highest reaction temperature and graphically evaluating its

effect on initial reaction rate. Figure 4-1 is an example of this type of study. The definition

ofNO2 penetration is the fraction of inlet NO2 concentration that exits the bed at a given time,

i.e., Cou/Ci. While superficial gas velocity was varied by an order of magnitude, this did not

greatly affect the initial rates. The absolute variation in associated rate constants determined

during this type of evaluation was generally less than ten percent. Because increasing flow

rate through the packed-bed by an order of magnitude did not significantly affect the reaction

rate, the lowest flow rate was used for the majority of experimental runs to conserve cylinder

gases.

It will be shown later that, in most cases, the initial chemical kinetics were affected by

an increase in temperature, providing more evidence that gas film mass transfer was not rate

limiting. Using the mass-transfer correlation of Ruthven (1968), the ratio of superficial

velocity to sorbent particle diameter is sufficient to minimize mass-transfer resistance, within

the temperature range evaluated. Because flow through a packed-bed is a highly complex

phenomenon and the intricate details relative to the gas-solid reaction mechanism and the

physical characteristics of the solid sorbent are not fully understood, it is possible that some

gas-film resistance to reaction exists. However, based upon the preceding discussion, it does

appear that this resistance has been effectively minimized in this experimental research study.











NO2 Penetration (Fraction)


0.6


Bed Exposure Time (min)


_- 3.05 alpm -A 6.06 alpm -E 29.49 alpm

Figure 4-1. Gas film mass-transfer resistance evaluation NO2 penetration (Co/C.) versus bed exposure time ([NO]m=200 ppm,
T=473 K, [02]=0%).


0.8


0.4


- -

.. . .. .


0.2











Solid Sorbent Characterizations


Surface Area Analyses


Five samples of the composite batch of MgO-coated vermiculite were randomly

selected for BET-specific surface area analyses. The intent was to look at the variability in

surface area between smaller batches of material prepared in the laboratory. If the variation

between samples is small enough, the mean surface area and sample standard deviation can

be used to establish confidence limits for sorbent sample surface area. Results are shown in

Table 4-1.



Table 4-1. BET surface areas of selected samples of laboratory prepared
MgO-vermiculite sorbent.

Sample BET Surface Area (mn2/g)

A 41.92 + 0.19

B 36.08 + 0.15

C 37.93 + 0.11
D 39.20 + 0.17

E 42.10 +0.17

Mean 39.45

Sample Standard Deviation 2.59










Particle Size Distributions


A sample of MgO-vermiculite sorbent was sieved, as previously described, and

results were plotted on a log-probability graph, shown in Figure 4-2. Since many particle size

data are log-normally distributed, this type of plot should produce a linear representation from

which relevant statistical parameters -- most importantly, the mass median diameter and

geometric standard deviation -- can be determined. From Figure 4-2, it can be seen that the

distribution is fairly linear between the 10th and 90th percentiles, so these data were used for

this purpose. From the distribution, the mass median diameter (MMD) is approximately 1.7

mm, and the geometric standard deviation (ag) is approximately 1.5. Since the sorbent

particles were mechanically abraded during the sieving process, the true MMD is probably

somewhat larger, perhaps in the range of 2 to 3 mm. The median diameter determined here,

however, can be used for calculating flow characteristics for a packed bed filled with this type

of sorbent.

A log-probability graph for the manufacturer-supplied size-distribution data on the

MgO powder used to prepare the combination sorbent is shown at Figure 4-3. From this

graph, the MMD for the MgO powder used is approximately 1.3 gm, and Og is equal to

approximately 2.5. Since the chemical reaction occurs at the surface of the MgO particles,

this distribution can provide useful data for use in determining physical parameters via the

shrinking unreacted-core model. The MMD is probably not the most appropriate dimension

to use for these purposes, however, since the reactions occurring


















S 70 - ib__--


-4f 50 I
F0 I i |
CO I

-w 50 l i | i
4 30-- i- -

i l L
10
Qm I S I I
10 I I/ II

0.1 1 10
MgO-Vermiculite Particle Size, Dp (mm)


Figure 4-2. Log-probability plot ofMgO-vermiculite sorbent particle size distribution.














99.99 ,.- ,


99.9 --"---



99
^ iI1 / i l l i i l

-. i i i I|I / + I i !i




Q) 90



Q) 70
m i

50i i i i
SI ; I I / I | | | | |






100 101 102
MgO Particle Diameter, (um)


Figure 4-3. Log-probability plot of Akrochem Elastomag 170 MgO powder particle size distribution (manufacturer-provided
data). Il








86

on the particles are a function of particle surface area. The MgO particle surface median

diameter can be calculated from the log-probability distribution using the well-known Hatch-

Choate equations (Hinds, 1982). A calculated count median diameter of approximately 0.1

gm led to the determination of a value of approximately 0.6 im for the surface median

diameter using these equations.


NOX Removal by Mg(NO3)2-coated Vermiculite



Results for Mg(N03)2-coated vermiculite were similar to those for the plain heat-

treated vermiculite. Removals of NO2 were less than five percent over the range of

temperatures (373 to 473 K) and concentrations (20 to 200 ppm) evaluated. NO removal

efficiencies were much lower, and approached zero. The results of these tests were useful in

verifying that if magnesium nitrate is formed on the solid surface during the course of the

MgO (Mg(OH)2) reaction with NO2, and all the remaining MgO (or Mg(OH)2) is utilized or

is otherwise unavailable for reaction, then these results preclude the further removal of NOx

from the gases supplied to the bed, and reaction/removal should cease.


NO- Removal by Sorbent Material



It was found to be more difficult to remove NO than NO2 by reaction with MgO, the

former being generally less reactive than the latter. The removal of each species was

independently affected by the experimental variables evaluated. Specific results are described

in the following paragraphs.










NO Removal by Sorbent Material


NO removal by MgO-coated vermiculite sorbent appeared to depend on inlet NO

concentration and bed temperature. At low inlet NO concentrations (< 10 ppm), NO removal

efficiencies varied from initial values of approximately 60 percent to over 90 percent. In

general, initial NO removal efficiency was greater at 373 K than at 423 or 473 K, indicating

removal by physisorption rather than by chemisorption. The presence of water or water

vapor did improve the removal efficiency of NO, but overall rates of removal were still low.

As previously described, apparent NO removal was enhanced by the presence of oxygen,

through the formation of NO2. In their evaluation of NO removal by Ca(OH)2 solids, Chu and

Rochelle (1989) found that the gas- or solid-phase oxidation of NO to NO2 played an

important role in the removal of the former. At higher NO concentrations (> 100 ppm), the

overall removal efficiencies were lower (less than 30 percent). Jozewicz et al. (1990) found

no NO (from a 400 ppm NO gas stream concentration) removal by Ca(OH)2 solids without

the presence of oxygen at temperatures between 70 and 180 C. Carbon monoxide (CO) has

been widely used as a probe molecule in surface science studies because it is one of the

simplest adsorbate-adsorbent systems. Mejias et al. (1995) found the interaction of CO on

MgO to be weak and of electrostatic origin, without noticeable chemical contributions. Since

CO and NO are similarly non-reactive, this observation suggests that there may be a fixed NO

sorption capacity, possibly the result of limited surface area available for physisorption.

Figure 4-4 is a representative concentration profile for NO removal by a

MgO-vermiculite sorbent bed, without a significant quantity of NO2 present. As can be seen




Full Text
N02 Penetration (Fraction)
+- T=373K T=423K ~b~ T=473K
Figure 4-18. N02 penetration versus bed exposure time ([NO2]in=:100 ppm, [O2]=10%).


197
Mutasher, E. L; Khan, A. R.; and Bowen, J. H., "Communications: Reversible, Noncatalytic
Reactions Between Gases and Solids in Fixed Beds," Industrial Engineering Chemistry
Research. 28(10): 1550-1553, 1989.
Nelson, B.W.; Nelson, S.G.; and Higgins, M.O., A New Catalyst for NO.. Control.
ESL-TR-89-11, HQ AFESC, Tyndall AFB, Florida, 1989.
Nelson, B.W.; Van Stone, D.A.; and Nelson, S.G., Development and Demonstration of a
New Filter System to Control Emissions During Jet Engine Testing. CEL-TR-92-49, HQ
AFCESA, Tyndall AFB, Florida, 1992.
Nelson, S.G; Van Stone, D.A.; Little, R.C.; and Peterson, K.A., Laboratory Evaluation of a
Reactive Baffle Approach to NO.. Control. Final Report, Contract Number FO8635-90-C-
0053, HQ AFCESA, Tyndall AFB, Florida, 1993.
Nelson, S.G., (Sanitech, Inc.), U.S. Patent 4,721,582 (Jan. 26, 1988).
Nelson, S.G.; Nelson, B.W.; and Higgins, M.O., Study of the Regenerabilitv of a Unique New
Sorbent that Removes S02-N0.. from Flue Gases. Phase II Final Report, EPA SBIR
Contract 68-D80066, Sanitech, Inc., Twinsburg, Ohio, 1990.
Newton, G. H.; Chen, S. L.; and Kramlich, J. C., "Role of Porosity Loss in Limiting S02
Capture by Calcium Based Sorbents," AIChE Journal. 35(6):988-994, 1989.
NTIS Tech Notes, "Process Eliminates 99% of NOx from Exhaust in Lab Tests," DOE
Technology Application, July 1987.
Organization for Economic Cooperation and Development (OECD), Control Technology for
Nitrogen Oxide Emissions from Stationary Sources. Paris. France, 1983.
Pakrasi, A., "Kinetic Studies on the Removal of Hydrogen Chloride from Flue Gas by
Hydrated Lime Powders in a Bench Scale Fixed Bed Reactor," Ph.D. Dissertation; University
of Tennessee, Knoxville, Tennessee, 1992.
Perry, R. A. and Siebers, D. L., "Rapid Reduction of Nitrogen Oxides in Exhaust Gas
Streams," Nature. 324 (18):657-658, 1986.
Petrik, M.A., Use of Metal Oxide Electrocatalvsts to Control NO..Emissions from Fixed
Sources. ESL-TR-89-29, HQ AFESC, Tyndall AFB, Florida, 1991.
Pigford, R. L. and Sliger, G., "Rate of Diffusion-Controlled Reaction Between a Gas and a
Porous Solid Sphere," Industrial Engineering Chemistry Process Design and Development.
12(1):85-91, 1973.


T=373 K, [O2]=0%). [ y = O.OOlx + 0.06, rM).99, Fit Std Error = 0.001, F-724.0],
ON
OO


+ Experimental Predicted
Figure 4-47. Experimental versus predicted N02 concentration ([N02]in=170.0 ppm, T=473 K, [O2]=0%, Gas samples collected
from middle of bed).


189
provided from exhaust-gas augmentation may accomplish some of this oxidation.
2. More MgO-vermiculite surface studies are necessary to better understand the
nature of the surface and the role it plays in the removal of N02 from a flowing gas stream.
Pore size distributions and associated structural implications may be very useful in further
explaining the rate-limiting processes involved in this reaction. The measurement of
additional physical parameters is required for the strict mathematical interpretation of
shrinking unreacted-core model results. Real-time surface analyses would allow for following
the progress of reaction, which could be combined with post-sorption analyses for the
quantification of solid conversion and product formation. A modified reactor design that
would allow for the removal of all used sorbent after a run (without any loss) would be an
improvement in this regard.
3. More traditional sorption studies, where the progress of reaction is followed
using thermogravimetric methods and single sorbent pellets, may also assist in the application
of the shrinking unreacted-core model to this reaction. This research is actually a step ahead
in that regard, since the basic research on the reaction between N02 and MgO powder
(without vermiculite substrate) has not yet been conducted or reported. The use of pressed
pellets of MgO would also be useful in this regard. Dispersing small quantities of MgO in an
inert bed of quartz sand is another method that may allow for the collection of this type of
data.
4. The range of experimental conditions studied should be extended to evaluate
the application of this sorbent material to other combustion sources. The use of NOx
concentrations from one to two orders of magnitude greater than used here for the collection


99
surface is still being converted into an inert product layer after the same period of exposure
time.
From Figure 4-9, at a reaction temperature of 423 K, initial N02 penetration is
lowest for the two lower inlet concentrations. Since there is less energy being supplied via
heat energy to overcome the activation energy barrier, it may be that only a limited number
of molecules can be removed. After approximately 25 minutes of bed exposure time have
elapsed, the overall penetration numbers again converged toward a value between
approximately 30 and 40 percent. At a reaction temperature of 373 K (Figure 4-10), there
was no discernible pattern observe for the initial penetration values, although the value for an
inlet N02 concentration of 20 ppm remained the lowest. Toward the end of the bed exposure
time period, however, N02 penetrations were lowest for the two highest concentrations. This
may be an indication of the higher concentration gradient serving to facilitate N02 removal
in promoting diffusion of the N02 molecules toward available MgO interior reaction surfaces.
Intrinsic first-order rate constants for these data are shown in Table 4-2.
Effects of Oxygen
Like NO removal, initial N02 removal efficiencies were improved when oxygen was
present. Since some NO, present as impurity in the N02 gas supplied, is oxidized to N02,
the actual N02 removal rate may be higher than observed, particularly at lower temperatures.
It is possible that NO, which is physisorbed on the sorbent surface, is eventually oxidized to
N02 before being sorbed to form a nitrite or nitrate compound. Since 02 concentrations
were at the percent level, they are considered to be in great excess relative to N02


72
acids, they were ultimately collected in the coalescing filter or condensed out in the sampling
manifold prior to analysis.
Similarly, the presence of water vapor itself can interfere with NOx measurements
directly by N02 absorption, as well as via third-body quenching of chemiluminescence
(Trdona et al., 1988; Campbell et al., 1982). Therefore, it is important to minimize contact
time of the gas with water vapor in sampling lines. Any N02 removed by gas humidification
is accounted for by the fact that bed inlet gases were sampled after this point in the system.
The analyzer reaction chamber is maintained at a vacuum of approximately 29 inches of
mercury (98 kPa) to minimize third-body quenching effects. Another common interference
with NOx measurements in the atmosphere is caused by the presence of such other nitrogen-
containing compounds as peroxyacetyl nitrate (PAN), which can be converted into NO in the
thermal converter. Again, these species are not suspected to be present in the experimental
procedures used in this study.
To prevent particulate matter contamination of the chemiluminescent analyzer, a 0.5
pm pore size Teflon filter (Gelman Sciences, Ann Arbor, Michigan) was placed in the
sample line at the analyzer sample inlet. This filter was found to be essentially non-reactive
to NOx. Without such a filter, it is quite easy to experience blockage of internal capillary
tubing. Deposits could also form on the glass filter that separates the reaction chamber from
the photomultiplier tube, causing a loss in instrument sensitivity (Klapheck and Winkler,
1985).
A continuous gas analyzer (Model IR-2100, Infrared Industries, Inc., Santa Barbara,
California) was used to electrochemically measure 02 concentrations. The principle of


29
Initial studies found that the sorbent material could be successfully regenerated with
minor reduction in adsorption capacity of the regenerated material. Ruch et al. (1990)
confirmed that the MgOvermiculite/NOx reaction is most likely due to chemisorption and
that the vermiculite/NOx sorption is a result of physical adsorption based upon hysteresis
during limited adsorption/desorption tests. This observed hysteresis is most probably the
result of the forces of chemisorption, generally involving electron transfer between the solid
(adsorbent) and gas (adsorbate) molecules held at the surface. These forces are typically
much stronger than the intermolecular van der Waals forces associated with physical
adsorption (Kovach, 1978; Szekely et al., 1976). They also evaluated the surface areas of
these products using N2 and water adsorption and Brunauer-Emmett-Teller (BET) theory
and found the sorbent surface areas to be larger than would be expected from the collective
surface areas of the vermiculite and MgO. This suggests that the vermiculite support
effectively enhances the surface area development of the MgO. Reported nitrogen isotherm
surface areas ranged from 16 to 55 m2/g for sorbents prepared at 550 C. Sieved samples had
a maximum surface area of 36 m2/g. Fourier transform infrared (FTIR) solid spectra indicated
the presence of Mg(OH)2, which disappeared as conditioning temperature increased. Nitrite
was found in some samples exposed to low-concentration flue gases by FTIR solid and
solution spectra, but results were determined to be inconclusive. X-ray diffraction interplanar
spacings showed that Mg(N03)2 was present.


CHAPTER 2
CONTROL OF NOx THROUGH GAS-SOLID INTERACTION
Exhaust-Gas Treatment Methods
A significant amount of research has been conducted over the past two decades to
determine a means of treating flue gases to reduce nitrogen oxide emissions. Some of the
most extensive research has been conducted in Japan, which was one of the first countries to
enact strict NOx emission regulations. In general, three classes of catalytic methods have been
evaluated. The first class involves mixing the waste gas with methane (CH4) or other gaseous
fuels/reducing agents before exposing the mixture to a catalyst; the second involves mixing
the effluent with ammonia before exposure to a catalyst; and the third involves exposure to
a catalyst or adsorbent with or without methane (or other fuel) or ammonia additions.
Numerous materials have been proposed as catalysts, and a number of these have been
patented, however, those receiving the most attention have been platinum or other precious
metals or metallic oxides from the alkaline earth group (Nelson et al., 1989; Klimisch and
Larson, 1975;Meubus, 1977).
18


80
chemical reactions can have substantial activation energies, operating at lower temperatures
should lead to control by chemical kinetics, making intrinsic kinetic data collected in this
range more valid. The effects of gas film mass transfer resistance were evaluated by varying
flow rate through the bed at the highest reaction temperature and graphically evaluating its
effect on initial reaction rate. Figure 4-1 is an example of this type of study. The definition
ofN02 penetration is the fraction of inlet N02 concentration that exits the bed at a given time,
i.e., C(JCm. While superficial gas velocity was varied by an order of magnitude, this did not
greatly affect the initial rates. The absolute variation in associated rate constants determined
during this type of evaluation was generally less than ten percent. Because increasing flow
rate through the packed-bed by an order of magnitude did not significantly affect the reaction
rate, the lowest flow rate was used for the majority of experimental runs to conserve cylinder
gases.
It will be shown later that, in most cases, the initial chemical kinetics were affected by
an increase in temperature, providing more evidence that gas film mass transfer was not rate
limiting. Using the mass-transfer correlation of Ruthven (1968), the ratio of superficial
velocity to sorbent particle diameter is sufficient to minimize mass-transfer resistance, within
the temperature range evaluated. Because flow through a packed-bed is a highly complex
phenomenon and the intricate details relative to the gas-solid reaction mechanism and the
physical characteristics of the solid sorbent are not fully understood, it is possible that some
gas-film resistance to reaction exists. However, based upon the preceding discussion, it does
appear that this resistance has been effectively minimized in this experimental research study.


(x-x) -
Figure 4-34. Shrinking unreacted-core chemical-reaction-control-equation evaluation, xr=205 minutes ([NCy^SO ppm,
T=423 K, [OJ=0%). [ y = 0.005x + 0.0002, rMJ.98, Fit Std Error = 0.003, F=812.9].
4^


100ppm s 50ppm 20ppm
Figure 4-11. N02 penetration vs bed exposure time (T=473 K, [O2]=10%).


139
quantity of material reacting is proportional to the available surface area of the unreacted core
(which at the beginning of the reaction would be the exterior surface of the particle as there
is no inert product layer formed) and can be expressed in the same manner as equation 4-4:
a
4 ti r 2 dt
C
b
4 7i r] &
bks[A]g
(4-13)
Here ks represents the intrinsic first-order rate coefficient for the surface reaction (cm/s).
Writing Nb in terms of the shrinking radius by analogy to equation 4-7:
a
4717]
~dr
4 2 c
7t rc
dt
bkt[A\
(4-14)
This may be rearranged and integrated to derive an expression describing how the unreacted
core shrinks with time:
-apB f*c drc bks[A]g dt (445)
After integration, t may be isolated by rearrangement:
t =
B
bkJA]
-CR-r)
(4-16)
The time for complete conversion of the solid under chemical reaction control, defined as rn
occurs when r= 0:
r
aPBR
MS\A\
X
(4-17)


195
Heap, M. P.; Chen, S. L.; Kramlich, J. C.; McCarthy, J.M.; and Pershing, D. W., "An
Advanced Selective Reduction Process for NOx Control," Nature. 335:620-622, 1988.
Hinds, W.C., Aerosol Technology: Properties. Behavior and Measurement of Airborne
Particles. John Wiley and Sons; New York, pp 69-100, 1982.
Ismagilov, Z. R.; Kerzhentsev, T. L.; and Susharina, T. L., "Catalytic Methods for Lowering
the Amount of Nitrogen Oxides in Exhaust Gases on Combustion of Fuel," Russian Chemical
Reviews. 59 (10):973-988, 1990.
James, N. J. and Hughes, R., "Rates of NOx Absorption in Calcined Limestones and
Dolomites," Environmental Science and Technology. 11(13): 1191-1194, 1977.
Johnson, C.; Henshaw, J.; and Mclnnes, G. "Impact of Aircraft and Surface Emissions of
Nitrogen Oxides on Troposheric Ozone and Global Warming," Nature. 355:69-71, 1992.
Johnson, S.A. and Katz, C.B. Feasibility of Reburning for Controlling NOxEmissions from
Air Force Jet Engine Test Cells. ESL-TR-89-33, HQ AFESC, Tyndall AFB, Florida, 1989.
Joseph, D. W. and Spicer, C. W., "Chemiluminescence Method for Atmospheric Monitoring
of Nitric Acid and Nitrogen Oxides," Analytical Chemistry. 50(9): 1400-1403, 1978.
Jozewicz, W.; Chang, J. C. S.; and Sedman, C. B., "Bench-Scale Evaluation of Calcium
Sorbents for Acid Gas Emission Control," Environmental Progress. 9(3): 137-142, 1990.
Karlegard, A. and Bjerle, I., "Kinetic Studies on High Temperature Desulphurization of
Synthesis Gas with Zinc Ferrite," Chemical Engineering Technology. 17:21-29, 1994.
Kikkinides, E. S. and Yang, R. T. "Simultaneous S02/NOx Removal and S02Recovery from
Flue Gas by PressureSwing Adsorption," Industrial Engineering Chemistry Research.
30(8): 1981-1989, 1991.
Kittrell, R.J., High Temperature NO.. Control Process. ESL-TR-89-36, HQ AFESC, Tyndall
AFB, Florida, 1991.
Klapheck, K. and Winkler, P., "Sensitivity Loss of a NOx-Chemiluminescence Analyzer Due
to Deposit Formation," Atmospheric Environment. 19(9): 1545-1548, 1985.
Klimisch, R. L. and Larson, J. G., The Catalytic Chemistry of Nitrogen Oxides. Plenum Press;
New York, 1975.


73
operation upon which this instrument is based is a simple coulometric process in which
oxygen in the sample gas stream is reduced in an electrochemical cell. The gas stream enters
the cathode cavity, with any oxygen being metered to the cathode through a diffusion barrier.
At the cathode, oxygen is electrochemically reduced via the reaction
02 + 2H20 + 4e' 40H- (3-3)
An electrolyte solution in the cathode cavity contains potassium hydroxide (KOH), which
facilitates migration of the generated hydroxyl ions to an anode, where they are oxidized back
to oxygen. This occurs following the reaction
40H- 02 + 2H20 + 4e* (3-4)
The resulting cell current is directly proportional to the oxygen concentration in the sample
gas stream. This current is measured electronically and is converted into an indicated
concentration value. The instrument measures oxygen concentration between zero and 25%,
with a manufacturer-reported accuracy of 2% of full-scale on all ranges. When using this
instrument, one must keep the sample inlet pressure and flow rate within the specified limits
to ensure accuracy of the measurement.
Multiple-point calibrations were performed on all instruments following procedures
outlined in the appropriate instrument manuals. Certified standard gases were used as transfer
standards whose composition is traceable to the National Institute of Standards and
Technology, and appropriate calibration curves were determined by challenging the
instruments with known standard mixtures. Typical calibration curves for the NOx and 02
analyzers are shown in Figures 3-7 and 3-8, respectively. Daily zero and span checks were
performed on each analyzer, and quality control charts were produced. Running sample


174
There is some deviation outside the confidence intervals in the middle of the run, most
probably the result of a transition from one mechanism to another. Both the combination of
controlling-mechanism and increased-molar-volume equations predicted x values between
approximately 400 to 500 minutes. To evaluate the predictability of sorption behavior at this
(relatively) high concentration and temperature scenario, the empirical relationship between
1 Ik and bed exposure time was used to produce the plot shown in Figure 4-47. Since this
relationship is most predictable after an exposure time greater than approximately 30 to 40
minutes, it is likely that the rate coefficients predicted are for the slow-rate-diffusion process.
For this reason, the model does not predict early sorption behavior well, leaving a gap
between the experimental and predicted curves.
Figure 4-48 is a plot of cumulative N02 mass removed versus bed exposure time,
which clearly shows the decreased mass removal rate as reaction progressed and diftusional
resistance increased. While a larger mass of N02 was collected in this case, MgO utilization
was still less than one percent. An interesting finding related to this slowdown in mass
removal is depicted in Figure 4-49, which shows N02 and NO concentrations as a function
of time as measured at the middle of the bed. It can be seen that as the N02 concentration
increased toward the equilibrium product layer diffusion control situation at approximately
40 minutes, NO production was increasing toward a maximum. As N02 removal rate
decreased due to the diftusional resistance, the rate of product formation (including NO) also
decreased. After 300 minutes, N02 removal was approximately 5% and the NO
concentration had almost reached its initial value. It is expected that when N02 was no longer


160
conversion in gassolid reactions to the effective closure of a narrow size range of small
surface pores, leaving larger pores of smaller surface area available for reaction at a greatly
reduced rate. Bhatia and Perlmutter (1981b) presented a correlation that allows for the
calculation of maximum conversion based upon initial porosity (ea) and Z:
eo
= (Z-l)(l-e; (4-36)
Assuming an initial porosity value of 0.3, based upon values used for CaO by the authors, the
calculated value for Xmax is approximately 16%. High initial sorbent porosity (and associated
high surface area) can create the potential for internal diffusion limitations to chemical
reaction. Marsh and Ulrichson (1985) found that grains of CaO obtained by calcining
Ca(OH)2 contained significant quantities of micropores, many of them around 22 in radius.
Effective pore closure occurs when this radius decreases to approximately 5 A. If intrapellet
diffusional resistance is appreciable, pore closure may occur at the surface, significantly
limiting diffusion into the particle interior and reducing ultimate MgO conversion within the
sorbent. In micropore diffusion, the diffusing molecules are "trapped" by the attractive forces
near the pore walls in a process similar to surface diffusion, but steric effects are more
important. Small differences in molecular size or shape can produce large differences in
diffiisivity (Ruthven, 1988). The simple grain model cannot realistically simulate porosities
less than 26%, whereas initial porosities of calcined limestones can be greater than 50%
(Damle, 1994).


200
Yagi, S. and Kunii, D., "Fluidized-solids reactors with continuous feed III. Conversion in
experimental fluidized-solid reactors," Chemical Engineering Science. 16(3):364 -391, 1961.
Yang, R. T. and Chen, J. M., "Kinetics of Desulfurization of Hot Fuel Gas with Calcium
Oxide. Reaction between Carbonyl Sulfide and Calcium Oxide," Environmental Science and
Technology. 13(5):549-553, 1979.
Zhang, X.; Walters, A. B.; and Vannice, M. A., "Catalytic Reduction of NO by CH4 over
Li-Promoted MgO," Journal of Catalysis. 146(2):568-578, 1994.
Zhidomirov, N. U., "Active Centres of Magnesium Oxide Surface and Calculations of
Dissociative Chemisorption of Methane on Modified MgO," Catalysis Today. 13:517-522,
1992.


CHAPTER 4
RESULTS AND DISCUSSION
Intrinsic Kinetic Studies
Experiments were conducted to collect data that provide insights into relevant
parameters controlling NOx removal efficiencies, potential reaction mechanism(s) and
reaction kinetics. Duplicate runs were conducted for every set of conditions to generate the
data necessary for the determination of kinetic parameters. A comparison of results between
duplicate runs under similar conditions showed remarkable reproducibility (statistical error
< 5%). To determine intrinsic reaction kinetics, NOx concentrations were followed with
increasing bed exposure time from a selected sampling valve, which represented a specific bed
residence time. This was necessary because of a system artifact that created a lag-time of
approximately 3-4 minutes to allow for line flushing until a stable instrument reading was
reached after a sampling valve was selected. To explain further, for a given set of conditions,
four separate runs would be necessary to collect data for all five sampling points; bed inlet,
three points within the bed, and bed outlet. Bed inlet values were checked on all runs, before
and after the other data were collected. Every 10 seconds, the three NOx analyzer channels
displayed 10-second average NO, N02, and NOx concentration values ([N02] calculated by
difference). Data were collected every minute throughout the run. From these data,
78


1 (1-x)1/3 1 3(1-X)2/3 2(1-X)
+ 473K-R 473K-D
Figure 4-44. Shrinking unreacted-core model evaluation ([N02]in=200 ppm, T=473 K, [O2]=0%).


(x-0-
Os
-J


T=373K T=423K -b- T=473K
Figure 4-19. N02 penetration versus bed exposure time ([NO2]in=50 ppm, [O2]=10%).


179
forming a sharp front which moves inward with increasing exposure time. Larger samples
with varying (CaO) thickness may lead to significant intraparticle difrusional limitations,
which may intensify as the solid product forms particle agglomerations (Zarkanitis et al.,
1990). While dynamic equilibrium conditions are achieved at a time equal to x, and effective
conversion approaches a fairly constant value, N02 is still being removed from the gas stream
at a rate dependent upon its effective diffiisivity through the product layer and/or porous
structure of the sorbent. There remains a significant residual N02 removal capacity, which
continues at significant levels for extended periods of time.
The molar volume ratios of Mg(N03)2 products to Mg(OH)2 reactants are
approximately 50% smaller than the ratio to MgO reactants, which may partially account for
the improved performance of hydrated sorbents. This results in more available surface area
through unblocked pore-surface area on a normalized time basis. The shrinking unreacted-
core model was used to interpret results under these conditions and, in genera}, x values were
approximately twice as large. This means that hydrated sorbents can remove approximately
twice as much mass from a gas stream with a given mass flow rate. Due to the inherent
complications involved in the study of gas-solid reactions, chemical kinetics or the shrinking
unreacted-core model alone can not completely describe the molecular processes that are
occurring. A structural investigation of reactants and products may be helpful for this
purpose. The crystalline structure of MgO is the same as that of NaCl, while Mg(OH)2
crystallizes with the Cdl2 layer structure, in which every Mg atom is surrounded by six OH.
Every OH forms three bonds to Mg atoms in its own layer, while contacting three OH of an
adjacent layer (Wells, 1952). A larger bulk reactant would be comprised of relatively thin,


43
apparently very sensitive to the presence of water vapor, which reduces porosity (Borgwardt,
1989). A graphic representation of the shrinking unreacted-core model was depicted in a
series of energy dispersive X-ray analysis micrographs of calcined limestone chalk and S02
(Dam-Johansen and Ostergaard, 1991a). The associated data were fitted with a grain model
of the shrinking unreacted-core mechanism (Dam-Johansen, 1991b). Bjerle et al. (1992)
reported useful experimental techniques for describing this reaction as well. They were able
to use the shrinking unreacted-core model to describe the various stages of reaction as
conversion progressed. Sotirchos and Zarkanitis (1992) found that the sulfation rates of MgO
were comparable to those for CaO. A simple grain model was also used to interpret data
describing a furnace sorbent slurry injection process, in which a limestone slurry was calcined
for use in S02 removal (Damle, 1994).
S02-Calcium Hydroxide Reaction
Flue-gas desulfurization through spray drying of a Ca(OH)2 slurry has recently become
an important alternative to the traditional wet lime or limestone scrubbing techniques. In
spray drying, S02-containing flue gases are contacted in a dryer with finely atomized lime
slurry that absorbs the S02. As water evaporates from the slurry droplets, reacted solids and
unreacted Ca(OH)2 remain and are collected on bag filters. This accumulated residual
Ca(OH)2 serves to remove considerable quantities of unreacted S02 from the inside of the
bags as well as in the ducts. Ruiz-Alsop and Rochelle (1988) used a bench-scale fixed-bed
reactor to study this reaction. The shrinking unreacted-core model was applied to their
results. Among the many factors evaluated, relative humidity of the gas was determined to


199
Svensson, R.; Ljungstrom, E.; and Lindqvist, O., "Kinetics of the Reaction Between Nitrogen
Dioxide and Water Vapour," Atmospheric Environment. 21(7): 1529-1539, 1987.
Szekely, J.; Evans, J.W.; and Sohn, H.Y., Gas-Solid Reactions. Academic Press, New York,
1976, pp 257-258.
Tambe, S.; Gauri, K. L.; and Cobourn, W. G., "Kinetic Study of S02 Reaction with
Dolomite," Environmental Science and Technology. 25(12):2071-2075, 1991.
Tanabe, K. and Fukuda, Y., "Basic Properties of Alkaline Earth Metal Oxides and Their
Catalytic Activity in the Decomposition of Diacetone Alcohol," Reaction Kinetics and
Catalysis Letters. 1 (l):21-24, 1974.
Tidona, R. J.; Nizami, A. A.; and Cernansky, N. P., "Reducing Interference Effects in the
Chemiluminescent Measurement of Nitric Oxides from Combustion Systems," Journal of the
Air Pollution Control Association. 38(6):806-811, 1988.
Wander, J.D. and Nelson, S.G., "NOx Control for Jet Engine Test Cells," 93-RA-83C-01,
Paper presented at the 86th Annual Meeting and Exhibition of the Air and Waste
Management Association, June 13-18, 1993, Denver, Colorado.
Wark, K. and Warner C.F., Air Pollution: Its Origin and Control. 2nd ed.,
HarperCollinsPublishers; New York, 1981, pp 371-423.
Wells, A. F., Structural Inorganic Chemistry. 2nd ed., Oxford at the Clarendon Press; Oxford,
England, pp 362, 416-417, 1952.
Wen, C. Y. and Ishida, M., "Reaction Rate of Sulfur Dioxide with Particles Containing
Calcium Oxide," Environmental Science and Technology. 7(8):703-708, 1973.
Wicke, B. G.; Grady, K. A.; and Ratcliffe, J. W., "Limitations on the Rapid Reduction of
Nitrogen Oxides in Exhaust Gas Streams," Nature. 338:492-493, 1989.
Wickham, D. T. and Koel, B. E., "Steady-State Kinetics of the Catalytic Reduction of
Nitrogen Dioxide by Carbon Monoxide on Platinum," Journal of Catalysis. 114:207-216,
1988.
Wolff, E. H. P.; Gerritsen, A. W.; and Van Den Bleek, C. M., "Multiple Reactor Testing of
a Synthetic Sorbent for Regenerative Sulfur Capture in Fluidized Bed Combustion of Coal,"
The Canadian Journal of Chemical Engineering. 71:83-93, 1993.


LIST OF FIGURES
Figure gage
2-1. Schematic representation of a jet engine test cell (JETC) 24
2-2. Schematic representation of the shrinking unreacted-core model 38
2-3. Schematic representation of the grain model 40
3-1. Experimental arrangment for packed-bed studies 59
3-2. Example low-flow-rate rotameter calibration curves
(Omega N042-15ST Tube) 61
3-3. Example high-flow-rate rotameter calibration curves
(Omega NO92-04G Tube) 62
3-4. Schematic representation of 316 stainless steel packed-bed reactor 64
3-5. Schematic representation of internal sampling tube location and appearance ... 65
3-6. Schematic representation of 316 stainless steel mixing or sampling manifold ... 66
3-7. Typical chemiluminescent NOx analyzer calibration curve (for NO channel) ... 74
3-8. Typical IR-2100 oxygen analyzer calibration curve 75
4-1. Gas-film-mass-transfer resistance evaluation N02 penetration (C^/C^
versus bed exposure time ([N02]in=200 ppm, T=473 K, [O2]=0%) 81
4-2. Log-probability plot of MgOvermiculite sorbent particle size distribution .... 84
4-3. Log-probability plot of Akrochem Elastomag 170 MgO powder particle size
distribution (manufacturer-provided data) 85
Vlll


52
variation, as well as for comparison with data from the Sorbtech samples to ensure that the
sorbent was properly prepared.
The sorbent material was incorporated into a packed-bed arrangement in a
non-reactive 316 stainless steel tubular reactor in a controlled temperature (tube furnace)
environment, which will be described later. A single batch of material was used for all
experiments. Limited surface area/composition analyses of the material were performed to
establish baseline values. Comparative experiments were conducted using sorbent samples
provided by Sorbtech.
Temperature
Reaction temperatures were controlled using a tube furnace (Model 421135,
Thermolyne, Dubuque, Iowa) to contain the reaction vessel. Sorption study temperatures of
373, 423, and 473 K were used. These temperatures were chosen because they represent a
reasonable range expected for augmented/cooled exhaust gases. Gases were preheated
before entering the furnace to ensure that they were at or above reaction temperature (+ 20
0 C) when they entered the bed, so they would not need additional heating.
Pressure
All runs were conducted at atmospheric pressure. The bed outlet gases were
exhausted to the atmosphere through a laboratory ventilation system. Pressure drop versus
exposure time was evaluated through the bed, using a U-tube manometer connected to ports
at the bed inlet and outlet. Pressure drop versus gas velocity was evaluated via an empirical
correlation.


178
removed, NO would no longer be produced and would indeed reach its initial bed inlet
concentration.
General Applicability of the Shrinking Unreacted-Core Model
It appears from the cases just presented that the shrinking unreacted-core model
provides a basis for chemically and physically describing the sorption behavior occurring
during the removal of N02 by the MgOvermiculite sorbent. Combined with the classical
chemical kinetics evaluations performed, these results allow for the prediction of sorption
behavior with time under a variety of conditions. The use of the local-equilibrium theory to
describe the transition from packed-bed theories to the shrinking unreacted-core model was
a key step in the process. While the local-equilibrium theory allowed for the determination
of solid conversion as it relates to N02 penetration, it is apparent that this conversion is
relative rather than absolute, based upon the low sorbent utilization rates found. These low
utilization rates are certainly a complex function of many factors, however, such as the low
N02 concentrations evaluated, as well as probable low porosities, which may produce
intraparticle diffusional resistances, are key parameters. The high solid-product-to-reactant
molar volume ratio is another mitigating factor in this situation. Dam-Johansen and
Ostergaard (1991b) attributed the rapid decrease in reaction rate during the sulfation of
limestone particles to filling of micropores. The macroporous structure remaining may allow
some sulfation to continue at a very slow rate. The degree of conversion decreases linearly
from the particle surface inward to a certain distance from the surface where it goes to zero,


CHAPTER 1
INTRODUCTION
Background
The combustion of carbonaceous jet fuels during jet engine testing produces
significant quantities of nitrogen oxides (NOx). Besides direct negative impacts on human
health, nitrogen oxides have been linked to other detrimental effects on air quality and the
environment. The latter include interactions with hydrocarbons to produce photochemical
oxidants and smog, and contribution to the phenomenon of acid precipitation through nitric
acid formation. Nitric oxide emissions may also contribute to the degradation of visibility and
aesthetics by direct emission of high NO levels in elevated plumes followed by atmospheric
oxidation to N02 and fine particulate aerosols resulting from by-products of N02 reactions
(Wark and Warner, 1981).
Properties and Health Effects of Nitrogen Oxides
There are seven possible forms of nitrogen oxides (NxOy). Two of the most abundant
of these gaseous oxides of nitrogen, nitric oxide (NO) and nitrogen dioxide (N02), are the
predominant air pollutants of environmental and health concern. These two forms, however,
are rapidly interchangeable in the atmosphere and are often grouped together and collectively
1


2 2.1 2.2 2.3 2.4 2.5 2.6 2.
1000/T (K-1)
Bed Exposure Time
^ 5 minutes 10 minutes -b 15 minutes 20 minutes
Figure 4-22. Arrhenius plot of In k versus 1000/T (T between 473 and 373 K, [NO2]in=20 ppm, [O2]=0%).


4-4. NO removal by humidified MgOvermicuilite sorbent ([NOJ^^IO ppm,
T=423 K, [02]=11%, 3% H20 vapor) 88
4-5. N02 removal versus residence time (T=473 K, [02]=9.9%) 90
4-6. First-order kinetic plot ([NOJ^lOO ppm, T=473 K, [O2]=10%) 92
4-7. First-order kinetic plot ([NO2]fa=100 ppm, T=473 K, [O2]=10.5%) 93
4-8. N02 penetration vs bed exposure time (T=473 K, [O2]=0%) 95
4-9. N02 penetration vs bed exposure time (T=423 K, [O2]=0%) 96
4-10. N02 penetration vs bed exposure time (T=373 K, [O2]=0%) 97
4-11. NQ2 penetration vs bed exposure time (T=473 K, [O2]=10%) 102
4-12. N02 penetration vs bed exposure time (T=423 K, [O2]=10%) 103
4-13. N02 penetration vs bed exposure time (T=373 K, [O2]=10%) 104
4-14. N02 penetration vs bed exposure time ([N02]in=200 ppm, [O2]=0%) 108
4-15. N02 penetration vs bed exposure time ([N02]in:= 100 ppm, [O2]=0%) 109
4-16. N02 penetration vs bed exposure time ([NOJ^SO ppm, [O2]=0%) 110
4-17. N02 penetration vs bed exposure time ([NO2]in=20 ppm, [O2]=0%) Ill
4-18. N02 penetration vs bed exposure time ([N02]¡=100 ppm, [O2]=10%) 113
4-19. N02 penetration vs bed exposure time ([NOJ^SO ppm, [02]=10%) 114
4-20. N02 penetration vs bed exposure time ([NO2]i=20 ppm, [02]=10%) 115
4-21. Arrhenius plot of In k versus 1000/T (T between 473 and 423 K,
([NO2]m=100 ppm, [O2]=0%) 116
4-22. Arrhenius plot of In k versus 1000/T (T between 473 and 373 K,
([NO2L=20 ppm, [O2]=0%) 117
4-23. N02 penetration versus bed exposure time (20 ml H20 added to 7 g
MgOvermiculite sorbent, [NO2]in=100 ppm, [O2]=0% T=473 K) 121
IX


50
and its variations, as well as empirically. These models provided a chemico-physical basis for
explaining sorbent performance and probable reaction mechanisms.
Experimental Variables
MgOVermiculite Reactive Sorbent Material.
This material (approximately 45% MgO to 55% vermiculite substrate by weight) was
prepared according to a process patented by the inventor (U.S. Patent 4,721,582, Sorbent
Technologies Corporation, (Sorbtech), Twinsburg, Ohio, 1988). Commercial-grade
magnesium oxide was used (98% volatile-free MgO) with a manufacturer-reported surface
area of 170 m2/g (Elastomag 170, Akrochem Corporation, Akron, Ohio). While the
process of preparing the sorbent is described in detail in the patent, a more general descriptive
process is presented here for informative and comparative purposes. Sorbent was prepared
in small batches, by necessity, in stainless steel pans in the laboratory. A quantity of coarse
vermiculite was weighed out in the pan, to which a 4:1 mass ratio of deionized water was
added. This is the maximum ratio allowed in the patent, but is greater than the 2:1 ratio
reported by Sorbtech in the preparation of a sample provided by them for experimental use.
Trial-and-error practice in the laboratory indicated that sufficient water was necessary to
attach the majority of the MgO particles to the surface of the vermiculite. Since the
vermiculite is highly hydrophilic, it readily absorbs the extra water. The sorbent is
subsequently heat-treated, during which process any excess water will be driven off. The
water and vermiculite were mixed to ensure complete absorption.


69
compared and averaged to determine actual experimental reaction temperature. This reaction
temperature was used to calculate the actual flow rate through the bed, which could then be
used to determine residence time at any given sampling point within the bed. It was assumed
that the sorbent bed was at the same temperature as the reactor surface after sufficient
equilibration time.
When gas humidification was used, a secondary preheating heat tape was used before
the gases entered the humidification vessel, which consisted of an impinger apparatus
mounted inside a modified 3-liter Pyrex flask (Southern Scientific, Inc., Micanopy, Florida)
containing deionized water The flask was placed in a heated flask mantle connected to a
separate voltage controller. Various temperature/flow rate combinations were evaluated to
allow for 3-5% water vapor to be picked up by the flowing gas stream. The humidified gases
were then reheated to ensure that they were approximately 20 C above reaction temperature.
The actual water vapor content of the gases was determined by weighing collected condensed
water from a stainless steel coalescing filter (Model 41S6, Balston, Inc., Haverhill,
Massachusetts) at the bed outlet. Water vapor was removed from the sample gas stream
before analysis through condensation in the sampling manifold. Gas flow rate was always
corrected for the presence of water vapor. The effect of moisture on reaction was sometimes
evaluated simply by adding sufficient deionized water to a virgin bed to convert all MgO into
Mg(OH)2. This was verified by the presence of water vapor at the bed outlet. Once
converted, Mg(OH)2 is stable over the range of temperatures evaluated (Sidgwick, 1952).


105
While initial N02 penetration was lowest for an inlet N02 concentration of 20 ppm, it rapidly
increased and reached a value of approximately 50%. This short initial period of efficient
removal may represent the small quantity of N02 molecules initially physically adsorbed on
the surface of the sorbent, or a chemical reaction induction period which occurs before
dynamic equilibrium conditions are achieved.
From all of these data, it seems apparent that the range of reaction temperatures
evaluated encompass a transition region where sorbent N02 removal characteristics in the
presence of oxygen are changing. The significance of the effects of temperature variation on
N02 penetration and the difficulty in distinguishing between these effects at the two lower
temperatures will be examined in a later section. Table 4-3 contains intrinsic first-order rate
coefficient data obtained with 02 present, which can be compared to the values displayed in
Table 4-2 as a means of estimating the significance of oxygen in its effect on N02 removal by
the sorbent.
Effects of Bed Temperature
There appeared to be a N02 reaction rate dependence on bed temperature when
temperatures exceeded 423 K. In the range from 373 to 423 K, initial first-order N02
removal rate coefficients were often approximately the same. When temperatures greater
than 423 K were used, there was a much greater difference in first-order rate coefficient
values. As expected, with all other variables kept constant, higher bed temperatures produced
higher rate coefficient values. Additionally, as temperature increases, theoretically, more
adsorbate molecules with sufficient activation energy for reaction to occur will strike the


136
a a dNb
S dt 4nR2 dt
b dNA
4nR1 dt
bksmf[A\) bks[A]g
(4-4)
in which kg is the mass transfer coefficient between the gas phase and the sorbent particle.
Letting pB represent the molar density of B in the solid, and Fbe the particle volume,
the number of moles of B present in the particle is:
N pBr (4-5)
The decrease in the volume or radius of the shrinking unreacted core as a result of the
reaction of adNB moles of solid and (or with) bdNA moles of gas is given by the equation
-adNB = -bdNA = -apBdV = -apBd(4/3 nr)3 = -a\npJf]drc (4-6)
Substituting equation 4-6 in equation 4-4 produces an expression for the rate of reaction in
terms of the shrinking radius or unreacted core:
a dNB
S dt
a?/c dr c
R2 dt
- bkMh
(4-7)
By rearranging and integrating, an expression for how the unreacted core shrinks with time
is derived:
(4-8)


188
sorption behavior over a variety of conditions, in particular showing a transition from
chemical reaction control to product layer diffusion control. In some cases it appeared that
both processes were simultaneously rate-limiting. It is evident that the range of
concentrations and temperatures evaluated represented a transition region, which made it
difficult to conclusively determine rate-limiting mechanisms. The possibility of simultaneous
and sequential surface reactions makes modeling the N02 removal a challenging problem.
12. An empirical equation was developed that allows for the prediction of N02
removal over extended periods of time. MgO utilization rates are far below stoichiometrically
predicted values, most probably the result of some combination of lost surface area due to
pore blockage or reduced surface area due to coverage of available reactant by a higher-
molar-volume product, and an insufficient concentration-gradient driving force for continued
reaction.
Recommendations for Further Research
Based upon the conclusions presented above and the experiences gained during the
conduct of this experimental study, a number of recommendations for further research in the
area have become evident:
1. The oxidation of NO to N02 is critical to the effective employment of this
sorbent medium in the control of total NOx. Options for the accomplishment of this oxidation
should be investigated. This oxidation could be accomplished prior to entering the bed or
within the bed, although residence times there may be too short. Contact with oxidizing
agents like peroxide solutions, for example, may be a starting point. In practicality, oxygen


Air Flow (mL/min)
Rotameter Reading
Omega Data I Lab Calibration
Figure 3-2. Example low-flow-rate rotameter calibration curves (Omega N042-15ST Tube).


CHAPTER 6
CONCLUSIONS AND RECOMMENDATIONS
Conclusions
The gassolid reactions between N02 and NO and MgO-vermicuite sorbent were
studied using a bench-scale packed-bed reactor and simulated operating conditions
representative of jet engine test cells. The effects of primary variables including gas-phase
concentrations, reaction temperature, moisture, and gas velocity were evaluated. The design
of the packed-bed reactor allowed for the collection of gas samples from various points within
the bed, which is an improvement over the common use of inlet and outlet measurements in
the determination of intrinsic chemical kinetic parameters in packed-bed studies. First-order
rate coefficients were determined for use in the determination of activation energies
associated with this reaction using the Arrhenius relationship. Sorption behavior over time
was interpreted using the shrinking unreacted-core model, as well as an empirical model,
which described the chemical and physical processes occurring during the course of the
reaction. It should be emphasized that the conclusions reached from this research study are
specific to the sorbent studied and the operating conditions used. Based upon the results of
the experiments conducted and the interpretation of data collected during the course of this
study, the following conclusions can be drawn:
185


Cum Mass NO 2 Removed (mg)
Figure 4-36. Cumulative mass N02 removed
collected from center of bed).
I NOx Analyzer Data
versus bed exposure time ([NO2]in=50
ppm, T=423 K, [O2]=0%, Gas samples


20
Davini (1988) studied the reduction of nitrogen oxide with ammonia in the presence of
carbonaceous soots from industrial boilers. Lee and Kline [U.S. Patent 3,864,451 (1975)]
describe a method for removing NO in the presence of sulfur dioxide by mixing ammonia with
the flue gas in the presence of a catalyst selected from the group of platinum and transition
metals and oxides and mixtures of these. Andersen and Keith [U.S. Patent 3,008,796 (1961)]
have described the use of NH3 in combination with cobalt, nickel, or iron supported on
alumina, silica, silica gel, or diatomaceous earth. Lyon (1987) describes an improved method
using ammonium injection in a tightly controlled noncatalytic process. A problem with this
method is the formation of ammonium salt particulate matter. Other patents relating to the
use of ammonia and precious metal or other catalysts for the removal of oxides of nitrogen
from a waste or tail gas include those of Cohn, Steele, and Andersen [U.S. Patent 2,975,025
(1961)] and Keith and Kenah [U.S. Patents 3,245,920 (1966); and 3,328,155 (1967)] (Nelson
et al., 1989).
Control of nitrogen oxides is also attainable by several processes that do not utilize
methane or ammonia, because the latter may be difficult to control. Additionally, the desired
ammonia reaction occurs only within a narrow range of temperatures. One process that
received great attention was the "RAPRENOx" selective reduction reaction first developed
by Perry and Siebers (1986). In this process, cyanuric acid (non-toxic) is mixed with Ma
intaining exhaust gases from which the NOx concentration (as NO) is significantly reduced
(NTIS Tech Notes, 1987). Heap et al. (1988) combined this technology with combustion
modification techniques to further improve the reduction of NOx. Unfortunately, it appears
that this reaction is negatively affected by the presence of free oxygen (Wicke et al., 1989).


143
Before integrating equation 4-23, one of the variables must be eliminated. As in gas-film
diffusion, Na can be eliminated by writing it in terms of rc. Substituting again from equation
4-6, and rearranging produces an integrable form:
- apB /r'(- Wcdre bDt[A] f dt (4-24)
Jr.-R r R Jo
From this,
t =
6bD'[A]g
rc >
[1 3(-f)2
R
2 ^
R
(4-25)
Defining rd as the time to complete conversion (r= 0) of the solid particle under product-
layer-diffusion control:
*PbK2
Xj 6bDe[A]g
(4-26)
This leads to an expression for the progress of the reaction in terms of rd\
= 1 3(f 2()3
x, R R
(4-27)
This may also be expressed in terms of fractional solid conversion from equation 4-11:
p(X) = 1 3(1-Z)m + 2(1-X)
x .
(4-28)


12
(JETCs) is regulated for soot opacity, JETCs currently operate under implicit exemption from
NOx regulations, although they qualify as stationary sources. If jet engine testing operations
are conducted in an ozone non-attainment area or where photochemical smog is already a
significant problem, they become obvious targets for environmental regulation. Mobile
sources are regulated under Title II of the Clean Air Act Amendments of 1990. Part B of
these amendments addresses aircraft emission standards. A jet engine in a test cell could be
considered to be a stationary source under Title I, and may be included for regulation under
Title IV, Acid Deposition Control (Quarles and Lewis, 1990). In anticipation of these
possibilities, the Administrator of the EPA and the Secretary of Transportation, in concert
with the Secretary of Defense, have commissioned an investigation into the implications of
regulating jet engine test cells. Some of the issues studied included the impacts of not
controlling nitrogen oxides, the existence of appropriate control technologies, costs
associated with these technologies and their effects on safety, design, structure, and operation
or performance of aircraft engines and performance tests in test cells (EPA-453/R-94-068).
After the recommendations of this study are evaluated, NOx emission standards for jet engine
test cells may be promulgated and enforced in the near future. A review of the most common
NOx control methods for stationary sources is useful for the consideration of their
applicability to JETC exhaust control.


140
An expression for the decrease in the radius of the shrinking core or the increase in fractional
conversion of the solid particle surface then follows, including a substitution from equation
4-11:
*(*)' 7'1 R '1 (4-18)
r
A linear plot of the right-hand side of the equation versus bed exposure time would allow for
the prediction of rr, and identify chemical-reaction kinetics as the rate-controlling process.
Diffusion-through-Inert-Product-Laver Control
This analysis is somewhat more complicated since it requires a two-step process, first
considering diffusion flux in a partially reacted particle, then letting the size of the unreacted
core change with time. From Figure 4-30, both reactant A and the unreacted core radius
move inward with time; however, the unreacted core shrinkage is slower than the flow rate
of A toward the unreacted core by approximately three orders of magnitude (the order of the
ratio of solid density to gas density), so a steady-state assumption regarding the dimension
of the unreacted core is valid. This assumption greatly simplifies the mathematics, because
it can now be assumed that the rate of reaction of A is equal to its rate of diffusion to the
surface of the unreacted core:
- = 4tir2QA = A%R 2Q^ = 4tir^QAC = Constant
dt
(4-19)


120
associated resistance to reaction or is simply a result of cumulative experimental error. Since
the rate of reaction becomes very slow toward the end of the run, small errors in the
determination of rate constants would become exaggerated in this range. It is also possible
that there is more than one process controlling the reaction, or multiple reactions could be
simultaneously taking place, obscuring the apparent activation energies measured.
Effects of Water
Water vapor carried by humidified gases greatly improved N02 removal efficiency and
overall sorption capacity. It is possible that this effect is due to the conversion of MgO into
the more-reactive Mg(OH)2, which is thermally stable over the range of temperatures
evaluated. Additionally, since N02 is somewhat soluble in water, a small quantity may be
removed in this way. This may cause the formation of nitrous and/or nitric acids, which are
probably more reactive than N02 and should be immediately neutralized by the basic sorbent,
preventing any potential for corrosion problems.
Intrinsic first-order rate coefficients (and N02 removal rates) for these experiments
were generally comparable in the early stages of reaction, however, N02 removal efficiency
remained higher for an extended period of time when water was present in the system. Figure
4-23 is a typical example of an N02 concentration versus bed exposure time profile utilizing
moistened sorbent. The possible reasons for this effect include the increased reactivity of
Mg(OH)2 compared to MgO, the solubility of N02 in water, and the lower molar volume ratio
of Mg(N03)2 to Mg(OH)2 compared to the molar volume ratio of Mg(N03)2 to MgO. N02
solubility was avoided in experiments where the sorbent was wetted prior to exposure to
flowing gases to convert the MgO on the surface into Mg(OH)2, leaving little free water.


79
concentration versus bed residence time curves could be plotted. Values for the same bed
exposure time were used to plot concentration of interest versus residence time, which was
a function of sampling point location and, thus, reaction time.
Limitation of Gas Film Mass Transfer Resistance
In the conduct of intrinsic chemical kinetic studies of gassolid reactions in packed-
bed reactor systems, it is critical that gas film mass transfer resistance to reaction be
minimized. The following description of the limitation of this resistance is based upon the
work of Fogler (1992) and Szekely et al. (1976). The mass flux from a fluid to a solid
surface is the product of a mass transfer coefficient and the concentration driving force.
When gas film mass transfer is rate-limiting, the extent of reaction or fractional solid
conversion is linear with respect to time. In a packed-bed system, the concentration driving
force is the difference between reactant concentration at the void space at a point in the bed
and the concentration at the surface of the sorbent particle adjacent to this point. Gas film
mass transfer resistance can generally be minimized by using small sorbent particles and high
superficial gas velocities. At low velocities, the stagnant gas film boundary layer thickness
is large at high velocities this thickness is decreased, reducing the mass transfer limitation
through increased convective mass transfer.
Operating at lower reaction temperatures also helps to reduce gas film mass transfer
limitations. When a reaction is gas film mass transfer limited, it will generally be insensitive
to increases in temperature. Since gas film mass transfer is not an activated process, while


14
Modifications to Operating Conditions
Several modifications to stationary and mobile source operating conditions have been
developed that can be used to reduce NOx formation from combustion. These include 1) low-
excess-air firing; 2) off-stoichiometric combustion; 3) flue gas recirculation; 4) reduced air
preheat; 5) reduced firing rates; and 6) water injection. Excess air is the amount of air that
is in excess of the amount stoichiometrically required for 100 percent combustion of the fuel.
Due to imperfect mixing of air and fuel in the combustion zone, there must be some excess
air present at all times to reduce fuel waste and to prevent smoke formation. As excess air
decreases, NOx follows, while using less fuel. Off-stoichiometric combustion (also called
staged-combustion) burns the fuel in two or more steps. The initial flame zone is fuel-rich,
and the following zone(s) is (are) fuel-lean. Combustion with remaining air in the resulting
fuel-rich regions in the primary flame zone are controlled by heat transfer. Although the
overall air/fuel ratio is near stoichiometric, the primary NOx formation zone of the flame is
operated in a low-NOx condition.
Rerouting some of the flue gas back to the burner primary combustion zone is called
flue gas recirculation. This process not only reduces the peak flame temperature, but also
lowers the partial pressure of available oxygen at the burner, thereby decreasing NOx
formation. Reducing the amount of combustion-air preheat lowers the peak temperature in
the primary combustion zone, decreasing thermal NOx production. Since JETCs operate
using ambient air as combustion air, this technique is obviously not applicable to these
operations. Likewise, reduced firing rates to reduce heat release per unit volume cannot be


186
1. NO is not readily removed by MgOvermiculite. Any removal probably
occurs through limited physisorption and/or conversion into nitrogen dioxide by oxygen
present in the augmented exhaust gas. The oxidation of NO to N02 is crucial to the
successful employment of this medium.
2. Nitrogen dioxide removal efficiencies greatly exceed nitric oxide removal
efficiencies, with maximum values greater than 90 percent. N02 continues to be removed by
the sorbent at a greatly reduced rate for extended periods of time.
3. High concentrations of N02 were readily removed from the gas stream, rapidly
converting MgO into Mg(N03)2. The formation of a solid product layer slowed the reaction
down as it progressed from control by chemical-reaction kinetics to diffusion control, a much
slower process. The higher concentration gradients provided the larger driving forces for the
continued removal of N02 through diffusion.
4. The reaction between N02 and MgOvermiculite is first-order with respect
to N02. The temperature dependence of the first-order rate coefficients is evidence that data
were collected in the region of chemical reaction control. Activation energies associated with
this reaction were in the range from approximately 20 to 36 kJ/g-mol in the absence of
oxygen and 15 to 25 kJ/g-mol when oxygen was present in the system. These values are
indicative of a low-energy chemisorption process.
5. Wetted sorbent or humidified gases promote N02 removal, possibly by the
conversion of MgO into Mg(OH)2, which appears to be more reactive and is thermally stable
over the range of temperatures studied. Since the molar volume ratio of magnesium nitrate
to magnesium hydroxide is lower than when compared to MgO, there is less potential for


132
It is also possible that the transformed surface catalyzes the continued oxidation (or
reduction) of N02. A more descriptive discussion of sorbent lifetime within the context of
the shrinking unreacted core model will be presented later.
Beds regenerated through heating to approximately 550 K can be reused with little
loss of performance. Obviously, regenerating solely through heating serves to release NOx
from the solid surface into the gas phase. This might defeat the purpose of using the sorbent
to control NOx. Introduction of a reducing gas (CO appears to be the reductant of choice)
when regenerating the sorbent can alleviate this problem (Nelson et al., 1990).
Proposed Reaction Mechanism
While the exact reaction mechanism in this gas-phase interaction is unknown, the
experimental data provide information that may be used to describe a possible mechanism.
While the NQ2 removal greatly exceeded NO removal, some NO was removed initially,
presumably via a physisorption mechanism, as previously discussed. The finding of NO
production during N02 sorption led to an investigation of possible reaction mechanisms that
would ultimately produce NO. A stepwise mechanism was envisioned, in which N02
molecules are sequentially incorporated into intermediate products, ultimately forming
Mg(N03)2.
A simplified mechanism is proposed as follows:
1) N02 + MgO MgN03
2) N02 + MgN03 MgN2Q5


21
It has been determined that various alkali metal oxides and silicate materials can reduce
nitrogen oxides without additional methane or ammonia. Harris, Morello, and Peters [U.S.
Patent 3,459,494 (1969)] have successfully evaluated CaO, SrO, BaO, K20, Na20 and others
supported on Alundum cement, porcelain, silica, extended alumina, or alumina beads. Heavy-
metal catalysts (copper, silver, nickel, molybdenum, palladium, and cobalt), supported on
alumina were patented by Ryason [U.S. Patent 3,454,355 (1969)]. The catalytic properties
of copper for the reduction of NOx without added methane or ammonia are described by
various investigators, including Gehri and Frevel [U.S. Patent 3,718,733 (1973)] and Kressley
[U.S. Patent 3,682,585 (1972)]. Kranc and Lutchko [U.S. Patent 3,576,596 (1971)]
employed a combination of copper and chromium impregnated on carbon supports to remove
both NO and CO from a waste gas.
Non-Catalvtic Methods
Many problems associated with catalytic control methods often make the use of
non-catalytic methods more appealing. The most common problems with catalytic techniques
include catalyst fouling or poisoning, and a narrow range of operating temperatures. These
problems can be avoided by using non-catalytic controls. A number of dry sorption
techniques that do not employ catalysts have also been evaluated as they apply to the control
of nitrogen oxides. Ganz (1958) evaluated activated carbon, aluminosilicate, silica gel,
manganese dioxide, copper dioxide, and coke for the removal of high concentrations of N02
and found aluminosilicate to be the most suitable sorbent. Kyollen [U.S. Patent 3,498,743
(1970)] described a process employing either sodium carbonate, calcium carbonate, or


[N02] (ppm)
1 min
I 5 min
10 min
B
15 min
20 min
30 min
-A-
40 min
X
1 hour
Bed Exposure Time
Figure 4-5. N02 removal versus residence time (T=473 K, [02]=9.9%).
VO
o


5
features in the local area. At an emission source, the concentration of nitrogen oxides is much
higher than ambient background levels. Emission rates vary widely depending on the type of
source and type of fiiel used, as well as on the type and quality of pollution control equipment
employed. This variation in NOx production rates for uncontrolled sources is quite evident
in Table 1-2.
From Table 1-2, it can be seen that an individual aircraft appears to emit negligible
amounts of nitrogen oxides when compared to much larger combustion sources. However,
recent studies by Johnson et al. (1992) indicate that aircraft emissions into the troposphere
at an altitude above 10 km produce increased concentrations of ozone, off-setting ozone
depletion, but contributing to increased surface temperatures ("global warming"). Modeling
suggests that the radiative forcing of surface temperature is about 30 times more sensitive to
aircraft emissions of nitrogen oxides than surface emissions. In the vicinity of an airport or
Air Force base, where ground operations (taxiing, take-offs, landings, and maintenance in
engine test cells) are combined with flight operations for a large number of aircraft, the
contribution of pollutants in the local area due to nitrogen oxides can be significant. NOx
standards have traditionally not been enforced for jet engine test cells, which qualify for
regulation as stationary sources. Gas turbine engines operating as mobile sources also have
not been regulated in the past. Evolving regulations including the 1990 Clean Air Act
Amendments, the failure of existing controls on stationary industrial processes to significantly
lower urban ozone concentrations, and a demonstrated correlation between ozone levels and


86
on the particles are a function of particle surface area. The MgO particle surface median
diameter can be calculated from the log-probability distribution using the well-known Hatch-
Choate equations (Hinds, 1982). A calculated count median diameter of approximately 0.1
pm led to the determination of a value of approximately 0.6 pm for the surface median
diameter using these equations.
NO.. Removal by MgfNO,V,-coated Vermiculite
Results for Mg(N03)2-coated vermiculite were similar to those for the plain heat-
treated vermiculite. Removals of N02 were less than five percent over the range of
temperatures (373 to 473 K) and concentrations (20 to 200 ppm) evaluated. NO removal
efficiencies were much lower, and approached zero. The results of these tests were useful in
verifying that if magnesium nitrate is formed on the solid surface during the course of the
MgO (Mg(OH)2) reaction with N02, and all the remaining MgO (or Mg(OH)2) is utilized or
is otherwise unavailable for reaction, then these results preclude the further removal of NOx
from the gases supplied to the bed, and reaction/removal should cease.
NO.. Removal by Sorbent Material
It was found to be more difficult to remove NO than N02 by reaction with MgO, the
former being generally less reactive than the latter. The removal of each species was
independently affected by the experimental variables evaluated. Specific results are described
in the following paragraphs.


123
Sorbtech-Supplied Sample Results
Experimental results for samples of sorbent provided by Sorbtech were generally
comparable to those for laboratory-prepared sorbent. Initial removal rates were greater than
99% for all cases evaluated; however, N02 penetration remained lower at the end of the run
for these samples. Intrinsic first-order rate constants were approximately the same early in
the reaction, but the decrease with extended bed exposure time was not as great. As
discussed earlier, the Sorbtech samples actually showed lower average BET surface areas
than the laboratory prepared sorbent. This difference in available surface area would
generally indicate less activity for the Sorbtech samples in the removal of N02, which
apparently was not the case. It may be that the laboratory-prepared sorbent had higher
porosity, which would contribute to a higher total surface area. The physical appearance of
the Sorbtech sorbent was darker, with a more brown to tan color than the cleaner white
appearance of the laboratory-prepared sorbent.
One distinct difference between the samples was that the Sorbtech sorbent was
significantly more dense than the laboratory prepared sorbent, a bed volume of the Sorbtech
sorbent weighed approximately 40% more than the laboratory sorbent. It was thought that
the higher mass might be the result of the presence of water in the Sorbtech sorbent, which
would also help to explain the improved N02 removal efficiency with time. Preweighed
samples of both sorbents were heated in a muffle furnace to a temperature of 623 K for one
hour to evaluate weight loss from dehydration. This proved to be insignificant, with the


T=373K T=423K -a- T=473K
Figure 4-20. N02 penetration versus bed exposure time ([NO2]in=20 ppm, [O2]=10%).


ACKNOWLEDGMENTS
I am extremely grateful for the support of my doctoral committee chairman, Dr. Eric
R. Allen, for his wealth of knowledge in academic research, for sharing in the excitement of
discovery, as well as for providing encouragement when problems arose. I will always fondly
remember our enlightening discussions about my research, as well as life in general, as high
points in my research. I also appreciate the sage advice of Dr. Dale A. Lundgren, whose
practical and applied approach to engineering goes a long way beyond what can be learned
in a textbook. I am thankful to both professors and to my fellow air pollution graduate
students, for their friendship and comradery during my time as a graduate student at UF.
Special thanks to Dr. Joseph D. Wander, whose belief in the value of academic
research is unparalleled, for providing the idea for this project, as well as for his help,
friendship and advice all along the way. I am also thankful to Dr. W. Emmett Bolch, Jr.,
whose letter years ago convinced me to come to the University of Florida, and to Dr. Robert
J. Hanrahan, to whom I credit the excellent performance of the experimental system.
I would like to acknowledge Colonel Robert A. Capell, USAF, BSC, who saw the
need for a bioenvironmental engineer with a Ph.D. in air pollution. Finally, I would like to
thank the Armstrong Laboratory/OL-EQS, Tyndall AFB, Florida for providing funding for
this research project, and the U.S. Air Force and the Air Force Institute of Technology, for
their confidence in selecting me for advanced education and for the financial support.


0.6
N02 Penetration (Fraction)
0.5
0.4
0.3 -
0.2
o.-i
o
0 10 20 30 40 50 60
Bed Exposure Time (min)
100ppm ~13 50ppm x_ 20ppm
Figure 4-13. N02 penetration vs bed exposure time (T=373 K, [O2]=10%).
o


[NO] (ppm)
Residence time (sec)
Exposure Time
1 hour 1 1.5 hours 2 hours
Figure 4-4. NO removal by humidified MgOvermiculite sorbent ([NO]in=210 ppm, T=423 K, [02]=11%, 3% H20 vapor).
00
oo


I [N02] [NO]
Figure 4-49. Comparative bed N02 and NO concentrations versus bed exposure time during long-term run showing NO
production ([NO2]in=170 ppm, T=473 K, [O2]=0%, Gas samples collected from middle of bed)).


48
2. Evaluate saturation characteristics (lifetime) of the material. Develop a model
to predict sorbent performance with time.
3. Evaluate the desorption and regeneration efficiencies of NOx on
MgOvermiculite.
4. Evaluate the pressure-drop characteristics of the material over time as gas
collection progresses.


N02 Penetration (Fraction)
+- T=373K T=423K -b- T=473K
Figure 4-16. N02 penetration versus bed exposure time ([NO2]in=50 ppm, [O2]=0%).


HETEROGENEOUS-PHASE REACTIONS OF NITROGEN DIOXIDE WITH
VERM3CULITE-SUPPORTED MAGNESIUM OXIDE
By
LARRY THOMAS KIMM
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1995


37
resistance to diffusion. A mathematically simple model, that was developed by Yagi and
Kunii in 1955, is often used to describe non-catalytic gas-solid reactions. This model, often
called the shrinking unreacted-core model, is based upon the visualization that the reaction
occurs first at the outer surface of an idealized spherical reactant particle (Yagi and Kunii,
1961). As reaction progresses, the reaction front (or zone) moves radially into the solid,
often leaving behind completely converted material in the form of an inert solid product called
"ash." A graphical representation of this model is shown in Figure 2-2.
The reaction steps in this model are the same as those previously described for a
generalized non-catalytic gas-solid reaction, occurring sequentially, although in any given
reaction, some steps may not occur. Effective diffusivity of the adsorbate through the solid
product is assumed to be a constant. Various equations relating the radial position of the
reaction zone (a measure of solid conversion) with reaction time have been developed. These
equations will be described in following sections before they are applied to the present study.
Although originally developed for application to non-porous solid reactants, the model has
been successfully used to describe systems utilizing porous solid sorbents. The main physical
difference between the two situations is that the reaction zone is diffuse in a porous solid as
opposed to a sharp interface in a non-porous one. A variable-diffiisivity shrinking-core-model
was developed by Krishnan and Sotirchos (1993 a). Their analysis revealed the strong effects
of reaction temperature, gas-phase concentration, and product layer thickness on effective
diffusivity of the reactant gas through the solid reactant and "inert" product layer matrix.


10
formation of N02 is favored at low temperatures, but N02 dissociates back to NO at higher
temperatures. The rate of equilibrium NO formation increases rapidly with increasing
temperature. As the combustion zone temperature rises above 1000 K, the formation of N02
becomes less likely under equilibrium conditions. Kinetically, however, the NO formation
reaction is a slow one, which affects the availability of reactant to form N02. NO formation
by reaction (1-1) has a very high activation energy (317 kilojoules/mole) and is most likely
the rate-controlling reaction (Ismagilov et al., 1990).
It has been observed experimentally that NO concentrations in the combustion flame
zone are significantly higher than those predicted by the Zeldovich mechanism. This "prompt"
NO formation is attributed to super-equilibrium radical concentrations that are likely in
hydrocarbon flames. It has been suggested that the intermediate, HCN, is formed when N2
reacts with a hydrocarbon radical. HCN combines with OH to form CN, and then CN is
oxidized to NO. Based upon experimental data, MacKinnon (1974) developed a model that
predicts the concentration of NO formed during combustion as a function of temperature, N2
and 02 concentrations, and time. At a total pressure of one atmosphere, the model equation
is:
CNO 5.2 x 1017 (exp -72000/T) YN2(YG2)1/21 (1-8)
where: CNO = NO concentration (ppm),
T = absolute temperature (K),
Y¡ = mole fraction of component i,
and t = time (seconds).


142
The flux of A within the inert layer is expressed by Fick's law for equimolar
counterdifliision: (Note that 0A is positive.)
Qa D^-1 (4-20)
dr
where De represents the effective diffusion coefficient of the gaseous reactant through the
product layer (cm2/s). Owing to variations in the properties of the inert layer, the value of
this coefficient is difficult to estimate and is generally determined experimentally. The
effective diffusion coefficient can be obtained by adjusting De in the shrinking unreacted-core
or grain model until the model predictions fit experimental data (Marsh and Ulrichson, 1985).
Equations 4-19 and 4-20 may be combined:
. 4=Z> m. C <4-2
dt e dt
Integrating through the inert layer :
- f' = 4rcD f[AV d[A] (422>
dt Jr r2 e J[A]g4A]s
therefore,
dN.
A
dt
(- h 4nDJAl
r R
(4-23)
For a shrinking unreacted-core of given size, dN/dt is constant, but the inert layer
becomes thicker as the core shrinks, which decreases the diffusion rate of A through the layer.


134
subsequent product layer divisional limitation of reaction rate) may further improve the utility
i
of the model to describing reaction mechanism(s).
It is appropriate now to examine the equations for the three potential rate-limiting
resistances affecting heterogeneous gas-solid reactions. The detailed derivations can be
found in most standard texts on gas-solid reactions. The following is based on the
descriptions reported by Szekely et al. (1976) and Levenspiel (1972). Using the generalized
gas-solid reaction represented by equation (2-1)
aA (gas) + bB (solid) cC (solid) + dD (gas)
to identify the only process acting, we can describe the rate of reaction with respect to
conversion of the solid reactant for three idealized or limiting resistance situations: diffusion-
through-gas-film control, chemical-reaction control, and diffiision-through-inert-product-layer
control.
Diffiision-Through-Gas-Film Control
Whenever the stagnant gas film layer surrounding the sorbent particle produces
significant mass transfer resistance to reaction, the gas-phase concentration profile for
reactant A is as shown in Figure 4-28. From the figure, it is evident that no gaseous reactant
is present at the solid surface, therefore, the concentration driving force [the quantity ([A]&1S
- [A]^ = ([A]g [A]J] is a constant, having the value [A]g. It is convenient to derive kinetic
expressions for gassolid reactions based on the external surface area of the sorbent particle,
S = 4kR2. From equation 2-1, we know that a(dNB) = b(dNA), where NA and NB represent
the number of moles of the reactants. Therefore, we can write


32
Gas-Solid Interaction
The smallest unit of a gas-solid system can be represented by the interaction between
a single solid particle and a flowing gas stream. This representation is convenient and simple,
and, in principle, may be generalized to more complex multi-particle systems, such as packed
beds. These interactions can be catalytic, in which the solid functions as a catalyst -- often
in the facilitation of a gas-phase reaction or non-catalytic, in which the surface acts to
physically or chemically remove reactants from the gas phase.
Adsorption occurs with a decrease in surface free energy and normally a decrease in
entropy; as a result, it is generally an exothermic process. In physical adsorption, sorbed
species are attracted and held at the solid surface by van der Waals dispersion forces. These
forces are much weaker than chemical bonding forces; the heat evolved is small, generally
between 1 and 10 kcal/g-mole (4-40 kJ/g-mole), approximately the heat of vapor
condensation for relatively non-polar substances. Chemisorption is primarily responsible for
gassolid reactions and catalysis on solid surfaces. In chemisorption, the forces involved are
of the same order of magnitude as those for chemical bonding, and the heat of chemisorption
is often between 10 and 150 kcal/g-mole (40-600 kJ/g-mole). The process of physical
adsorption, like condensation, has very little or no activation energy and as a result is assumed
to occur rapidly. Chemisorption often displays a higher activation energy, implying a slow
rate; however, chemisorption reactions having small activation energies are known. Unlike
physisorption, chemisorption is specific to the gases and solids involved, just like chemical


Bed Exposure Time (min)
1 3.05 alpm 6.06 alpm 29.49 alpm
Figure 4-1. Gas film mass-transfer resistance evaluation N02 penetration (C^/C^) versus bed exposure time ([NO]in=200 ppm,
T=473 K, [O2]=0%).
00


144
When product-layer diffusion is rate-controlling, a plot of the right-hand-side of equation 4-
28 versus bed exposure time allows for the determination of rd.
Derivations of the Shrinking Unreacted-Core Model
As reaction proceeds and a product layer is formed, the decreasing product layer
diffusion rate can affect the overall rate. Equation 4-19 assumes a constant value of De,
which may not be a valid assumption. Variation of the (decreasing) diffusion coefficient in
the product layer may have to be taken into account to describe a reaction in which a more
voluminous solid product forms (Krishnan and Sotirchos, 1993). More-elaborate models can
be constructed allowing for changes in the controlling mechanism with time or solid-surface
conversion. Often, a simple modification to the basic equations allows for an improved
representation of the experimental data. When product-layer-diffusion is truly the rate-
limiting process, the inclusion of the molar volume ratio (represented by Z) in equation 4-28,
as follows, may improve the fit (Szekely et al., 1976):
J_ 3[Z (Z + (1 Z)(l X))]2* 3(1 ^y/3 (4-29)
Z 1
This is especially true when the product layer molar volume is significantly larger (or
smaller) than that of the solid reactant. Referring back to Table 2-1, four molar volume ratios
are of interest in this research:


58
Experimental Arrangement
The experimental arrangement for sorption and desorption/regeneration employed in
this study is shown in Figure 3-1. Nitrogen oxides (mixed in N2) were supplied from separate
certified standard cylinders (Bi-Tec Southeast, Inc., Tampa, Florida) to ensure reliability of
concentrations and to minimize contaminant concentrations. Initially, N02 mixtures were
generated using a NOx generator (Model 100, Thermo Environmental Instruments, Franklin,
Massachusetts) as a source of N02. This generator produces N02 by mixing ozone generated
from UV irradiation of air with NO. While the system could easily and efficiently convert
high concentrations of NO to N02, two major problems resulted from its use. First, it is
extremely difficult to perfectly titrate ozone with NO using this system, without leaving
excess ozone. The flow controls in the unit are insufficient for this purpose. A related but
more important variation was the ultraviolet lamp output used to generate ozone appeared
to vary from one test to the next for a given voltage setting. The presence of ozone was
unwanted in the experimental study since it could complicate data interpretation by oxidizing
NOx to other species not readily measurable with the instrumentation used.
As a result, it was necessary to use standard cylinder gas as an N02 source. Since this
is actually N204 under pressure, there is a small potential for the gas to decay with time.
While total NOx was the same as certified, some NO "contamination" is generally present in
the cylinder gas. The NO contamination in the N02 cylinders was always less than two to five
percent. Similar impurity values were noted in the NO/NO ratio in the NO cylinders. This


This work is dedicated to my family, especially my wife, Lisa, whose patience, love
and support during this effort made it all possible, and to my children, Meredyth and Wilson,
who always give me the best welcome home after a long day.


2
called "N0X." Some of the physical properties associated with gaseous nitrogen oxides
(NxOy) are shown in Table 1-1. Nitrous oxide (N20) is another oxide of nitrogen that is also
present in the atmosphere at appreciable concentrations. Nitrous oxide is a gas with
recognized anesthetic properties, which has also been suggested to be a product of
combustion. Its ambient concentration is approximately 0.5 parts per million (ppm), well
below the threshold concentration for biological effects. Fortunately, in the troposphere, it
is balanced atmospherically by a cycle that is independent of the other oxides of nitrogen.
Nitric oxide (NO) is a colorless gas whose ambient concentration is normally well below 0.5
ppm. At this concentration it produces minimal effects on human health. However, nitric
oxide is a precursor to the formation of N02 and is an active compound in the formation of
photochemical smog, initiating reactions that produce secondary air pollutants.
Table 1-1. Physical properties of nitrogen oxides (NxOy).
Species
Mol Wt
(g/gmol)
Solubility in
Water
(mL/lOOg)
Melting Point
(C)
Boiling Point
(C)
NO
30.01
7.34
-163.6
-151.7
no2
46.01
See Note 1
See Note 2
See Note 2
no3
62.00
Soluble (ether)
Decomp. 20.0
Unavailable
n2o
44.02
130.0
-102.4
-89.0
n2o3
76.01
Soluble (ether)
-111.0
2.0
n2o4
92.02
130.52
-11.3
21.2
NA
108.01
Soluble
40.7
Sublimes 32.3
NOTES: 1. Reacts with H20 forming H0N02 and/or HONO
2. Liquid and solid forms are primarily N204
Sources: Nealetal., 1981.; Lange's Handbook of Chemistry, 1973.


82
Solid Sorbent Characterizations
Surface Area Analyses
Five samples of the composite batch of MgO-coated vermiculite were randomly
selected for BET-specific surface area analyses. The intent was to look at the variability in
surface area between smaller batches of material prepared in the laboratory. If the variation
between samples is small enough, the mean surface area and sample standard deviation can
be used to establish confidence limits for sorbent sample surface area. Results are shown in
Table 4-1.
Table 4-1. BET surface areas of selected samples of laboratory prepared
MgO-vermiculite sorbent.
Sample
BET Surface Area (m2/g)
A
41.92 + 0.19
B
36.08 + 0.15
C
37.93 + 0.11
D
39.20 + 0.17
E
42.10 + 0.17
Mean
39.45
Sample Standard Deviation
2.59


Stagnant Gas Film
Figure 4-29. Graphical representation of the shrinking unreacted-core model under chemical-reaction control.
u>
00


26
the toxicity of these heavy metals is a detriment. The seventh study, which was selected for
development, involved vermiculite and magnesium oxide (MgO)-coated vermiculite. NOx
removal rates for this medium were appealing, and the technology has many other advantages
including economics, simplicity, regenerability, and particularly, insensitivity to changes in
exhaust gas conditions or composition (Nelson et al., 1989). The USAF has sponsored
numerous studies in order to quantify emissions from jet engines. Seitchek (1985), Fagin
(1988), and Spicer et al. (1990) have published jet engine emission values for many engines
in the USAF-inventory operating on JP-4 jet fuel. Similar studies are underway for emissions
associated with the combustion of JP-8 fuels. These documents describe NOx and other
combustion by-product emission rates correlated with power setting or operational
procedures. They are very useful for determining appropriate exhaust gas concentrations in
order to simulate jet engine exhaust gas composition in the laboratory. Many lab-scale studies
have used these data in their experimental design.
Vermiculite-based Catalyst
Vermiculite ((Mg,Fe(n),Al)3(Al,Si)4O10(OH)2-4H2O, a thermally expanded form of a
common aluminosilicate material) has been used as a carrier for other materials including
nitrates and phosphates, in which the combinations are used as slow-release fertilizers.
Evanshen [U.S. Patent 3,757,489 (1973)] describes the treatment of flue gases with
polyvinylpyrrolidone with or without the addition of a catalyst of a nitrate or sulfate of copper
or silver. He suggested suspending these materials on a carrier such as vermiculite. Sanitech,
Inc., (now Sorbent Technologies Corporation, Twinsburg, Ohio) developed a new class of


+ Experimental Predicted
Figure 4-43. Experimental versus predicted N02 concentration ([N02]in=19.1 ppm, T=373 K, [O2]=0%, Gas samples collected
from middle of bed).
o


124
laboratory samples losing approximately two percent mass, and the Sorbtech samples losing
about three percent. The higher density material did not produce a higher surface area on a
per bed basis, because there was little difference between the surface areas of laboratory-
prepared samples and Sorbtech samples. In fact, laboratory prepared samples had, on
average, 10% more surface area, which may be attributable to more microporosity, and may
help in explaining the decreased performance since the high-surface-area micropores are
more prone to blockage or plugging as reaction progresses and the nature of the sorbent
surface changes.
Pressure Drop Characteristics
The arrangement of particles in packed beds produces a complex geometry that makes
the flow field difficult to define. For most practical problems, empirical relationships between
pressure drop and the flow rate and characteristics of the fluid are used. When an empirical
equation can be validated with experimental data, this is the best situation. Of numerous
correlations proposed for relating pressure drop to flow rate, the Ergun equation is most
widely accepted (Szekely et al., 1976):
M 150^
c3 cp d
^ s p
L
Op
(4-3)


191
presence of water are truly the result of the Mg(OH)2 formed and the change in molar
volume ratio, or are a result of an absorption process.
8. Studies similar to those reported here should be conducted using S02 as the
adsorbate gas as it is also effectively removed by this sorbent. The basic kinetic and
mechanistic data on the reaction between S02 and MgO have not yet been reported. The
simultaneous exposure of the sorbent to S02 and NOx should also be evaluated since there
is a reported synergistic effect on the combined removal when the two species are present
together. The presence of S02 may also help in the removal of NO by its reduction to
nitrogen. Steciak et al. (1995) reported that the reaction: MgO + S02 + NO MgS04 + !4N2
is thermodynamically feasible below 1030 C, but no kinetic data have been reported. An
important point to be stressed here is that there are many other potential uses for this sorbent
that could be investigated.
9. The use of high-purity N02 from a permeation tube apparatus would avoid the
problem of NO contamination in bed feed gas. This would allow for an improved evaluation
of the NO formation seen during the course of reaction in this research. If there were no NO
present in the samples that entered the bed, formation by chemical reaction on the sorbent
surface would be easier to follow.
10. The application of this sorbent medium to ambient N02 sampling should be
investigated. Since low concentrations of N02 can be effectively removed using this sorbent
it may be feasible to attempt to collect low-level concentrations of N02 from urban
atmospheres for analysis if conditions were optimized.


36
Goochee, 1988). In general, gas-solid reactions can be described by an ideal Langmuir
isotherm only when there is exclusive and complete monolayer sorption, all active sites on the
sorbent are equivalent and there is no interaction between adjacent adsorbate molecules
(Comes et al., 1993). Solid structural changes resulting from the chemical change in the solid
can be quite complex, but are generally classified as being either sintering (and resultant pore
closure), swelling, softening, and cracking (Szekely et al., 1976). These changes may affect
diffusivities of gaseous reactants as well. A number of mathematical models based upon these
recognized physical and chemical phenomenon have been developed and described in the
literature.
Progressive-Conversion Model
This model is a simplified, idealized model for the non-catalytic reaction of particles
with a surrounding fluid. A basic assumption of this model is that the reactant gas enters and
reacts throughout the particle at all times, although probably at different rates at different
locations within the particle. This model obviously assumes a highly porous particle structure
to allow for easy reactant diffusion with little intraparticle resistance to diffusion. With
negligible diffiisional (or mass transfer) resistance, the overall rate is controlled by chemical
kinetics, which is generally slow in this case.
Shrinking Unreacted-Core Model
Evidence from a wide variety of situations indicates that the progressive-conversion
model does not accurately approximate the behavior of real particles. Often, the reaction
produces an inert product layer of converted solid reactant which can produce significant


Figure 2-3. Schematic representation of the grain model.
o


169
Results from a long-term run are presented at Figure 4-43. It appears from this figure
that full "effective" utilization and achievement of dynamic equilibrium occurred closer to 500
minutes than to 250 minutes, indicating that the combination of resistances equation may be
best-suited for describing sorption behavior in this situation. Chemical kinetics may be
partially controlling due to the low N02 concentration and associated limited reactant
availability, and diffusion may be controlling due to the lack of concentration driving force
to promote reaction. The ratio of N02 removed to NO produced was approximately 3:1 after
7 hours exposure time. The total mass of N02 removed at this time was approximately 10
mg, resulting in a very low utilization rate.
200 ppm NO? at a Reaction Temperature of 473 K
The other extreme condition evaluated was the highest N02 concentration at the
highest reaction temperature. The shrinking unreacted-core model was used to fit these data
as well. Figure 4-44 is a comparison of the two primary equations over the course of one
hour. While the product-layer-difiusion-control equation is fairly linear over this entire
period, the chemical-reaction-control equation shows a rapid increase in equation value,
corresponding to a rapid increase in N02 penetration early in the reaction. As expected, the
product-layer-diftusion-control equation seems more appropriate for describing the entire
course of reaction over the first hour. Figures 4-45 and 4-46 depict linear least-squares fits
to the chemical-reaction-control equation and product-layer-diffusion-control equation,
respectively. From these figures, it is evident that statistically, the product-layer-difiusion-
control-equation best represents the data, based upon the statistical parameters evaluated.


Bed Exposure Time (min)
NO Analyzer Data
x
Figure 4-37. Cumulative mass N02 removed versus bed exposure time for a long-term run ([NO2]in=50 ppm, T=423 K, [O2]=0%,
Gas samples collected from center of bed).


107
sorbent surface. Figures 4-14 through 4-17 are typical plots, similar to those presented in the
previous sections, in which the effect of reaction temperature can more readily be seen.
These figures depict sorbent performance when oxygen was not present in the system.
Similar plots for those experiments where oxygen was present in the system are shown at
Figures 4-18 through 4-20.
When the N02 inlet concentration was 200 ppm (without oxygen), initial N02
penetration was lowest at a bed reaction temperature of 473 K (Figure 4-14). Penetrations
for the lower temperatures were somewhat higher initially. This separation in penetration
remained until a bed exposure time between 20 and 25 minutes had elapsed, when N02
penetration at 473 K exceeded the values at the other temperatures. If diffusion through the
inert product layer controls the reaction at high conversion levels, an increase in temperature
would increase the diffusion rate since effective diffusivity can be temperature dependent and
improved diffusivity would be expected. This would cause higher reaction rates and overall
conversion levels (Marsh and Ulrichson, 1985). While the initial rate of change in N02
penetration versus time was low, it rapidly increased before beginning to taper off toward the
end of the run. This coincides with a relative increase in N02 penetration rate for the two
lower temperatures evaluated. This delay in increasing N02 penetration most likely represents
the added time required to complete formation of the outermost product layer due to the
decreased temperatures employed.
Similar curves are seen for an N02 inlet concentration of 100 ppm (Figure 4-15). For
inlet N02 concentrations of 20 and 50 ppm (Figures 4-16 and 4-17, respectively), initial
penetrations were low for all temperatures, but it appears that a higher temperature was


17
application to JETCs is expensive and involved compared to other treatments. At
temperatures from 900 to 1000 C, NH3 will reduce NOx to N2 without a catalyst; however,
NOx reduction efficiencies are only 40-60 percent. If the temperature is too low, unreacted
NH3 will be emitted ("ammonia slip"), and if the temperature is too high, NH3 can be oxidized
to NO. Thus both of these situations are obvious problems. Dry sorption techniques have
the advantage of simplicity of operation and minimization of waste effluent. Relatively high
NOx reduction rates have been documented (Nelson et al., 1989; Lyon, 1991).


25
enormous. A unique feature of jet engine test cells is their highly variable test cycles (and
NOx production associated with varying peak flame temperature). Because of this variability,
aNOx control technology applicable to jet engine test cells must perform over a wide range
of rapidly varying operating conditions.
Research Sponsored bv the U.S. Air Force (USAF)
The Air Force (through the US AF Engineering and Services Center and USAF Civil
Engineering Support Agency, now the Human Systems Center-Armstrong Laboratory,
Tyndall AFB, Florida) has recently sponsored seven independent projects that specifically
target the control of NOx from JETCs. Johnson and Katz (1989) showed that reburning of
exhaust with extra fuel was effective, but not economical. Ham et al. (1989) reported partial
success with non-catalytic reduction using additives (ammonia, amines, and hydrazines) to
simulated exhaust. Unfortunately, as temperature increased, NOx production increased. A
dual-bed mordenite-copper catalyst did offer some improvement in sensitivity to variable
conditions over some better known selective catalysts (Kittrell, 1991). Berman et al. (1991)
evaluated the photopromotion of NO thermal decomposition by various metallic oxides. This
decomposition was very effective, but only in the absence of oxygen. Initial work reported
by Petrik (1991) also found that NOx reduction by ceramic-supported electrocatalysts was
inhibited by even trace amounts of oxygen; however, continued development of these
materials has produced catalytic ceramics that may be effective inside the jet engine (Gordon,
1994). Lyon (1991) evaluated NOx reduction at the surface of heavier Group Ila oxides, but


19
Catalytic Methods
Cohn [U.S. Patent 3,118,727 (1964)] has been issued a patent that describes a
process for purifying waste gases containing nitrogen oxides by mixing them with a fuel (such
as CH4) and passing the mixture over a platinum- or rhodium-containing catalyst at an initial
reaction temperature of 690-780 F. Acres and Hutchings [U.S. Patent 3,806,582 (1974)]
describe a similar process. Childers, Ellis, and Ryan [U.S. Patent 2,910,343 (1959)] describe
a methane process involving two catalyst beds in series, one containing platinum or alumina
and the second containing nickel or alumina. Vanadium oxide, molybdenum oxide and/or
tungsten oxide catalysts on alumina or silicic acid substrates were used in combination with
CH4 as described by Nonnenmacher and Kartte [U.S. Patent 3,279,884 (1966)]. Such base
metal catalysts as iron, cobalt, nickel and copper dispersed on a refractory support were used
by Reitmeier [U.S. Patent 2,924,504 (I960)]. Other patents relating to the use of methane
and precious metal catalysts for the removal of NOx from waste or tail gases include those of
Andersen and Green [U.S. Patent 2,970,034 (1961)]; Romeo [U.S. Patent 3,425,803 (1969)];
Newman [U.S. Patent 3,467,492 (1969)]; Kandell and Nemes [U.S. Patent 3,567,367
(1971)]; Andersen, Romeo, and Green [U.S. Patent 3,098,712 (1963)]; and Hardison and
Barr [U.S. Patent 3,402,015 (1968)] (Nelson et al 1989; Lewis, 1975).
The use of ammonia as a reducing gas in the presence of a catalyst was described
earlier. Baiker et al. (1987a, 1987b) and Chen et al. (1990) describe the selective catalytic
reduction of nitric oxide with ammonia upon catalysts comprised of mono- and multi-layers
of vanadia supported on titania, and mono-layers of vanadia immobilized on titania-silica


54
Table 3-1. USAF FI 10 Turbine Engine Emissions Data (as reported)
Power
Setting
THC
(ppmC)
NOx (ppmv)
NO (ppmv)
CO (ppmv)
C02 (%)
Idle
7
13.8
11.2
85
0.98
30 Percent
6
30
28
23
1.25
63 Percent
3
97
92
13
2.35
Intermediate
(High
Mach)
3.5
243
227
15
3.17
105%
Afterburner
-Augmented
335
21.5
3.7
178
0.42
Source: Spicer et al., 1990, page 27.
Gas Flow Ratef s')
Gas flow rate(s) were kept constant for each run. However, flow rates/gas velocities
were varied to produce two contact times of 0.5 and 1.0 second, keeping pressure drop to
a minimum. Contact times were calculated by dividing bed length by superficial gas velocity.
This is only an approximation of the true distribution of contact times since the actual gas
velocity may be based upon a more tortuous pathway through the packed bed volume,
increasing or decreasing the contact time of an individual reactant gas molecule. These
superficial velocities correspond to nominal space velocities of 3,600 and 7,200 per hour.
Since mass transfer limitations from the gas phase to the solid surface can obscure
intrinsic chemical kinetics, it is important to limit this resistance to chemical reaction. A
sufficiently high gas velocity will effectively decrease the thickness of the stagnant gas film



PAGE 1

+(7(52*(1(2863+$6( 5($&7,216 2) 1,752*(1 ',2;,'( :,7+ 9(50&8/,7(6833257(' 0$*1(6,80 2;,'( %\ /$55< 7+20$6 .,00 $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( 6&+22/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$

PAGE 2

7KLV ZRUN LV GHGLFDWHG WR P\ IDPLO\ HVSHFLDOO\ P\ ZLIH /LVD ZKRVH SDWLHQFH ORYH DQG VXSSRUW GXULQJ WKLV HIIRUW PDGH LW DOO SRVVLEOH DQG WR P\ FKLOGUHQ 0HUHG\WK DQG :LOVRQ ZKR DOZD\V JLYH PH WKH EHVW ZHOFRPH KRPH DIWHU D ORQJ GD\

PAGE 3

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

PAGE 4

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f 6SRQVRUHG 5HVHDUFK 9HUPLFXOLWHEDVHG &DWDO\VW )LHOG 6WXGLHV 5HODWHG 5HVHDUFK *DVf§6ROLG ,QWHUDFWLRQ LY

PAGE 5

*HQHUDOL]HG 1RQ&DWDO\WLF *DV6ROLG 5HDFWLRQV 0DWKHPDWLFDO 0RGHOV 3URJUHVVLYH&RQYHUVLRQ 0RGHO 6KULQNLQJ 8QUHDFWHG&RUH 0RGHO *UDLQ 0RGHO $QDODJRXV *DV6ROLG 5HDFWLRQ 6\VWHP 5HVHDUFK 6f§&D2 5HDFWLRQ 6&DOFLXP +\GUR[LGH 5HDFWLRQ 6&DOFLXP&DIERQDWH 5HDFWLRQ 2WKHU $FLGJDV&DOFLXPEDVHGVROLG 5HDFWLRQV *HQHUDO $SSOLFDELOLW\ RI WKH 6KULQNLQJ 8QUHDFWHG&RUH 0RGHO 5HVHDUFK -XVWLILFDWLRQ 5HVHDUFK 2EMHFWLYHV (;3(5,0(17$/ 0(7+2'6 $1' 0$7(5,$/6 *HQHUDO 5HVHDUFK $SSURDFK ([SHULPHQWDO 9DULDEOHV 0J2f§9HUPLFXOLWH 5HDFWLYH 6RUEHQW 0DWHULDO 7HPSHUDWXUH 3UHVVXUH *DV &RPSRVLWLRQ *DV )ORZ 5DWHVf *DV 5HDFWLRQ6RUSWLRQ'HVRUSWLRQ )L[HG%HG 5HDFWRU 6\VWHP *HQHUDO ([SHULPHQWDO &RQVLGHUDWLRQV ([SHULPHQWDO $UUDQJHPHQW 7XEXODU )L[HGEHG 5HDFWRUV ([SHULPHQWDO 3URFHGXUHV *DV $QDO\VHV 6RUEHQW 6XUIDFH $UHD 'HWHUPLQDWLRQ 6RUEHQW 3DUWLFOH 6L]H 'HWHUPLQDWLRQ 5(68/76 $1' ',6&866,21 ,QWULQVLF .LQHWLF 6WXGLHV /LPLWDWLRQ RI *DV)LOP0DVV7UDQVIHU 5HVLVWDQFH 6ROLG 6RUEHQW &KDUDFWHUL]DWLRQV 6XUIDFH $UHD $QDO\VHV 3DUWLFOH 6L]H 'LVWULEXWLRQV 12[ 5HPRYDO E\ 0J1fFRDWHG 9HUPLFXOLWH 14; 5HPRYDO E\ 6RUEHQW 0DWHULDO 12 5HPRYDO E\ 6RUEHQW 0DWHULDO 1 5HPRYDO E\ 6RUEHQW 0DWHULDO (IIHFWV RI 2SHUDWLRQDO 9DULDEOHV Y

PAGE 6

(IIHFWV RI 1 &RQFHQWUDWLRQ (IIHFWV RI 2[\JHQ (IIHFWV RI %HG 7HPSHUDWXUH $FWLYDWLRQ (QHUJ\ 'HWHUPLQDWLRQ (IIHFWV RI :DWHU (IIHFWV RI 5HVLGHQFH 5HDFWLRQf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

PAGE 7

/,67 2) 7$%/(6 7DEOH SDJH 3K\VLFDO SURSHUWLHV RI QLWURJHQ R[LGHV 1[2\f (PLVVLRQ IDFWRUV IRU QLWURJHQ R[LGHV 12[ DV 1f 3K\VLFDO SURSHUWLHV RI VHOHFWHG 0JFRQWDLQLQJ LQRUJDQLF FRPSRXQGV 86$) ), WXUELQH HQJLQH HPLVVLRQV GDWD DV UHSRUWHGf %(7 VXUIDFH DUHDV RI VHOHFWHG VDPSOHV RI ODERUDWRU\SUHSDUHG 0J2YHUPLFXOLWH VRUEHQW ,QWULQVLF ILUVWRUGHU UDWH FRHIILFLHQWV Vnf IRU 1 UHPRYDO E\ 0J2YHUPLFXOLWH VRUEHQW >2@ bf ,QWULQVLF ILUVWRUGHU UDWH FRHIILFLHQWV Vnf IRU 12] UHPRYDO E\ 0J2YHUPLFXOLWH VRUEHQW >2@ bf 9OO

PAGE 8

/,67 2) ),*85(6 )LJXUH JDJH 6FKHPDWLF UHSUHVHQWDWLRQ RI D MHW HQJLQH WHVW FHOO -(7&f 6FKHPDWLF UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO 6FKHPDWLF UHSUHVHQWDWLRQ RI WKH JUDLQ PRGHO ([SHULPHQWDO DUUDQJPHQW IRU SDFNHGEHG VWXGLHV ([DPSOH ORZIORZUDWH URWDPHWHU FDOLEUDWLRQ FXUYHV 2PHJD 167 7XEHf ([DPSOH KLJKIORZUDWH URWDPHWHU FDOLEUDWLRQ FXUYHV 2PHJD 12* 7XEHf 6FKHPDWLF UHSUHVHQWDWLRQ RI VWDLQOHVV VWHHO SDFNHGEHG UHDFWRU 6FKHPDWLF UHSUHVHQWDWLRQ RI LQWHUQDO VDPSOLQJ WXEH ORFDWLRQ DQG DSSHDUDQFH 6FKHPDWLF UHSUHVHQWDWLRQ RI VWDLQOHVV VWHHO PL[LQJ RU VDPSOLQJ PDQLIROG 7\SLFDO FKHPLOXPLQHVFHQW 12[ DQDO\]HU FDOLEUDWLRQ FXUYH IRU 12 FKDQQHOf 7\SLFDO ,5 R[\JHQ DQDO\]HU FDOLEUDWLRQ FXUYH *DVILOPPDVVWUDQVIHU UHVLVWDQFH HYDOXDWLRQ 1 SHQHWUDWLRQ &A&A YHUVXV EHG H[SRVXUH WLPH >1@LQ SSP 7 >2@ bf /RJSUREDELOLW\ SORW RI 0J2f§YHUPLFXOLWH VRUEHQW SDUWLFOH VL]H GLVWULEXWLRQ /RJSUREDELOLW\ SORW RI $NURFKHP (ODVWRPDJ p 0J2 SRZGHU SDUWLFOH VL]H GLVWULEXWLRQ PDQXIDFWXUHUSURYLGHG GDWDf 9OOO

PAGE 9

12 UHPRYDO E\ KXPLGLILHG 0J2f§YHUPLFXLOLWH VRUEHQW >12-AA,2 SSP 7 >@ b b + YDSRUf 1 UHPRYDO YHUVXV UHVLGHQFH WLPH 7 >@ bf )LUVWRUGHU NLQHWLF SORW >12-AO22 SSP 7 >2@ bf )LUVWRUGHU NLQHWLF SORW >12@ID SSP 7 >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 14 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH >1@LQ SSP >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH >1@LQ SSP >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH >12-A62 SSP >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH >12@LQ SSP >2@ bf ,OO 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH >1@cf SSP >2@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH >12-A62 SSP >@ bf 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH >12@Lf SSP >@ bf $UUKHQLXV SORW RI ,Q N YHUVXV 7 7 EHWZHHQ DQG >12@P SSP >2@ bf $UUKHQLXV SORW RI ,Q N YHUVXV 7 7 EHWZHHQ DQG >12/ SSP >2@ bf 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH PO + DGGHG WR J 0J2f§YHUPLFXOLWH VRUEHQW >12@LQ SSP >2@ b 7 .f ,;

PAGE 10

3UHGLFWHG (UJXQ HTXDWLRQf YHUVXV H[SHULPHQWDO SUHVVXUH GURS DV D IXQFWLRQ RI VXSHUILFLDO YHORFLW\ IRU D P ORQJ EHG RI 0J2f§YHUPLFXOLWH VRUEHQW 7 .f 7KHUPDO GHFRPSRVLWLRQ RI 0J1f + RQ YHUPLFXOLWH WHPSHUDWXUH LQFUHDVLQJ ZLWK WLPHf 7KHUPDO GHFRPSRVLWLRQ RI XVHG 0J2f§YHUPLFXOLWH VRUEHQW 7 .f &RPSDUDWLYH EHG RXWOHW 1 DQG 12 FRQFHQWUDWLRQV YHUVXV EHG H[SRVXUH WLPH VKRZLQJ 12 SURGXFWLRQ *UDSKLFDO UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO XQGHU GLIIXVLRQ WKURXJKJDVILOP FRQWURO *UDSKLFDO UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO XQGHU FKHPLFDO UHDFWLRQ FRQWURO *UDSKLFDO UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO XQGHU GLIIXVLRQ WKURXJKLQHUWSURGXFWOD\HU FRQWURO 6KULQNLQJ XQUHDFWHGFRUH PRGHO HYDOXDWLRQ >12-A2 SSP 7 >2bf 6KULQNLQJ XQUHDFWHGFRUH FKHPLFDOUHDFWLRQFRQWUROHTXDWLRQ HYDOXDWLRQ [U PLQXWHV >12@LQ SSP 7 >2@ bf 6KULQNLQJ XQUHDFWHGFRUH SURGXFWOD\HUGLIIXVLRQHTXDWLRQ HYDOXDWLRQ WG PLQXWHV >12-A2 SSP 7 >2@ bf 6KULQNLQJ XQUHDFWHGFRUH FKHPLFDOUHDFWLRQFRQWUROHTXDWLRQ HYDOXDWLRQ W PLQXWHV >12-A62 SSP 7 >2@ bf 6KULQNLQJ XQUHDFWHGFRUH SURGXFWOD\HUGLIIXVLRQHTXDWLRQ HYDOXDWLRQ WG PLQXWHV >12-A62 SSP 7 >2@ bf &XPXODWLYH PDVV 1 UHPRYHG YHUVXV EHG H[SRVXUH WLPH >12-A62 SSP 7 >2@ bf &XPXODWLYH PDVV 1 UHPRYHG YHUVXV EHG H[SRVXUH WLPH IRU D ORQJWHUP UXQ >12@LQ SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP FHQWHU RI EHGf [

PAGE 11

3ORW RI LQYHUVH ILUVWRUGHU UDWH FRHIILFLHQW Nf YHUVXV EHG H[SRVXUH WLPH >1@LQ SSP 7 >2@ bf ([SHULPHQWDO YHUVXV SUHGLFWHG 1 FRQFHQWUDWLRQ >12-MA76 SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP PLGGOH RI EHGf 6KULQNLQJ XQUHDFWHGFRUH PRGHO HYDOXDWLRQ >12@LQ SSP 7 >2@ bf 6KULQNLQJ XQUHDFWHGFRUH FKHPLFDOUHDFWLRQFRQWUROHTXDWLRQ HYDOXDWLRQ [U PLQXWHV >12@LQ SSP 7 >2bf 6KULQNLQJ XQUHDFWHGFRUH SURGXFWOD\HUGLIILLVLRQHTXDWLRQ HYDOXDWLRQ UG PLQXWHV >12-A2 SSP 7 >2bf ([SHULPHQWDO YHUVXV SUHGLFWHG 1 FRQFHQWUDWLRQ >12-AA SSP 7 >2b *DV VDPSOHV FROOHFWHG IURP PLGGOH RI EHGf 6KULQNLQJ XQUHDFWHGFRUH PRGHO HYDOXDWLRQ >1@LQ SSP 7 >2@bf 6KULQNLQJ XQUHDFWHGFRUH FKHPLFDOUHDFWLRQFRQWUROHTXDWLRQ HYDOXDWLRQ [U PLQXWHV >1@LQ SSP 7 >2@ bf 6KULQNLQJ XQUHDFWHGFRUH SURGXFWOD\HUGLIILLVLRQHTXDWLRQ HYDOXDWLRQ [G PLQXWHV >1@LQ SSP 7 >2@ bf ([SHULPHQWDO YHUVXV SUHGLFWHG 1 FRQFHQWUDWLRQ >1@LQ SSP 7 >2b *DV VDPSOHV FROOHFWHG IURP PLGGOH RI EHGf &XPXODWLYH PDVV 1 UHPRYHG YHUVXV EHG H[SRVXUH WLPH IRU D ORQJWHUP UXQ >12/ SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP FHQWHU RI EHGf &RPSDUDWLYH EHG 1 DQG 12 FRQFHQWUDWLRQV YHUVXV EHG H[SRVXUH WLPH GXULQJ ORQJWHUP UXQ VKRZLQJ 12 SURGXFWLRQ >12-A+2 SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP FHQWHU RI EHGf [L

PAGE 12

$EVWUDFW RI 'LVVHUWDWLRQ 3UHVHQWHG WR WKH *UDGXDWH 6FKRRO RI WKH 8QLYHUVLW\ RI )ORULGD LQ 3DUWLDO )XOILOOPHQW RI WKH 5HTXLUHPHQWV IRU WKH 'HJUHH RI 'RFWRU RI 3KLORVRSK\ +(7(52*(1(2863+$6( 5($&7,216 2) 1,752*(1 ',2;,'( :,7+ 9(50,&8/,7(6833257(' 0$*1(6,80 2;,'( %\ /DUU\ 7KRPDV .LPP $XJXVW &KDLUSHUVRQ (ULF 5 $OOHQ 0DMRU 'HSDUWPHQW (QYLURQPHQWDO (QJLQHHULQJ 6FLHQFHV &RQWUROOLQJ QLWURJHQ R[LGHV 12rf IURP D QRQVWHDG\VWDWH VWDWLRQDU\ VRXUFH OLNH D MHW HQJLQH WHVW FHOO -(7&f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f R[\JHQ FRQWHQW DQG PRLVWXUH FRQWHQW $YDLODELOLW\ RI ZDWHU DQG R[\JHQ ZHUH IRXQG WR EH LPSRUWDQW YDULDEOHV ,W LV SUREDEOH WKDW ZDWHU LV QHFHVVDU\ IRU WKH FRQYHUVLRQ RI 0J2 WR 0J2+f ZKLFK LV D PRUH UHDFWLYH FRPSRXQG KDYLQJ WKHUPDO VWDELOLW\ RYHU WKH UDQJH RI [LL

PAGE 13

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f§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f GLIILFXOW :LWK LPSHQGLQJ UHJXODWLRQV DLPHG DW FRQWUROOLQJ -(7&V DV VWDWLRQDU\ VRXUFHV RI 12[ UHVXOWV IURP WKHVH VWXGLHV ZLOO EH XVHIXO WR HQYLURQPHQWDO HQJLQHHUV GHYHORSLQJ DLU SROOXWLRQ FRQWUROV IRU VLPLODU VRXUFHV .QRZOHGJH RI SDUDPHWHUV DIIHFWLQJ WKH HIILFLHQF\ DQG FDSDFLW\ RI 12[ FRQWURO V\VWHPV HPSOR\LQJ WKLV PHGLXP LV QHFHVVDU\ WR HQVXUH FRPSOLDQFH ZLWK UHJXODWLRQV ;OOO

PAGE 14

&+$37(5 ,1752'8&7,21 %DFNJURXQG 7KH FRPEXVWLRQ RI FDUERQDFHRXV MHW IXHOV GXULQJ MHW HQJLQH WHVWLQJ SURGXFHV VLJQLILFDQW TXDQWLWLHV RI QLWURJHQ R[LGHV 12[f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f 3URSHUWLHV DQG +HDOWK (IIHFWV RI 1LWURJHQ 2[LGHV 7KHUH DUH VHYHQ SRVVLEOH IRUPV RI QLWURJHQ R[LGHV 1[2\f 7ZR RI WKH PRVW DEXQGDQW RI WKHVH JDVHRXV R[LGHV RI QLWURJHQ QLWULF R[LGH 12f DQG QLWURJHQ GLR[LGH 1f DUH WKH SUHGRPLQDQW DLU SROOXWDQWV RI HQYLURQPHQWDO DQG KHDOWK FRQFHUQ 7KHVH WZR IRUPV KRZHYHU DUH UDSLGO\ LQWHUFKDQJHDEOH LQ WKH DWPRVSKHUH DQG DUH RIWHQ JURXSHG WRJHWKHU DQG FROOHFWLYHO\

PAGE 15

FDOOHG 1; 6RPH RI WKH SK\VLFDO SURSHUWLHV DVVRFLDWHG ZLWK JDVHRXV QLWURJHQ R[LGHV 1[2\f DUH VKRZQ LQ 7DEOH 1LWURXV R[LGH 1f LV DQRWKHU R[LGH RI QLWURJHQ WKDW LV DOVR SUHVHQW LQ WKH DWPRVSKHUH DW DSSUHFLDEOH FRQFHQWUDWLRQV 1LWURXV R[LGH LV D JDV ZLWK UHFRJQL]HG DQHVWKHWLF SURSHUWLHV ZKLFK KDV DOVR EHHQ VXJJHVWHG WR EH D SURGXFW RI FRPEXVWLRQ ,WV DPELHQW FRQFHQWUDWLRQ LV DSSUR[LPDWHO\ SDUWV SHU PLOOLRQ SSPf ZHOO EHORZ WKH WKUHVKROG FRQFHQWUDWLRQ IRU ELRORJLFDO HIIHFWV )RUWXQDWHO\ LQ WKH WURSRVSKHUH LW LV EDODQFHG DWPRVSKHULFDOO\ E\ D F\FOH WKDW LV LQGHSHQGHQW RI WKH RWKHU R[LGHV RI QLWURJHQ 1LWULF R[LGH 12f LV D FRORUOHVV JDV ZKRVH DPELHQW FRQFHQWUDWLRQ LV QRUPDOO\ ZHOO EHORZ SSP $W WKLV FRQFHQWUDWLRQ LW SURGXFHV PLQLPDO HIIHFWV RQ KXPDQ KHDOWK +RZHYHU QLWULF R[LGH LV D SUHFXUVRU WR WKH IRUPDWLRQ RI 1 DQG LV DQ DFWLYH FRPSRXQG LQ WKH IRUPDWLRQ RI SKRWRFKHPLFDO VPRJ LQLWLDWLQJ UHDFWLRQV WKDW SURGXFH VHFRQGDU\ DLU SROOXWDQWV 7DEOH 3K\VLFDO SURSHUWLHV RI QLWURJHQ R[LGHV 1[2\f 6SHFLHV 0RO :W JJPROf 6ROXELOLW\ LQ :DWHU P/O22Jf 0HOWLQJ 3RLQW r&f %RLOLQJ 3RLQW r&f 12 QR 6HH 1RWH 6HH 1RWH 6HH 1RWH QR 6ROXEOH HWKHUf 'HFRPS 8QDYDLODEOH QR QR 6ROXEOH HWKHUf QR 1$ 6ROXEOH 6XEOLPHV 127(6 5HDFWV ZLWK + IRUPLQJ +1 DQGRU +212 /LTXLG DQG VROLG IRUPV DUH SULPDULO\ 1 6RXUFHV 1HDOHWDO /DQJHnV +DQGERRN RI &KHPLVWU\

PAGE 16

7KH PRVW KDUPIXO HIIHFWV RI 12[ DUH SULPDULO\ DWWULEXWHG WR H[SRVXUH WR 1 WKH PRVW WR[LF RI WKHVH R[LGHV 1 LV D UHGGLVKEURZQ JDV WKDW LV TXLWH YLVLEOH LQ VXIILFLHQW FRQFHQWUDWLRQ SSPf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

PAGE 17

DGGLWLYH WR WKRVH RI RWKHU HQYLURQPHQWDO FRQWDPLQDQWV :DUN DQG :DUQHU 1HDO HW DO f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f 6RXUFHV 2YHU SHUFHQW RI DOO WKH PDQPDGH QLWURJHQ R[LGHV WKDW HQWHU WKH DWPRVSKHUH LQ WKH 8QLWHG 6WDWHV DUH SURGXFHG E\ WKH FRPEXVWLRQ RI IRVVLO IXHOV $V RI RYHU SHUFHQW RI WKHVH HPLVVLRQV DUH DWWULEXWHG WR VWDWLRQDU\ VRXUFHV VXFK DV XWLOLW\ DQG LQGXVWULDO ERLOHUV JDV WXUELQHV DQG VWDWLRQDU\ HQJLQHV $SSUR[LPDWHO\ SHUFHQW DUH IURP PRELOH VRXUFHV WUDQVSRUWDWLRQf 7KH UHPDLQLQJ IUDFWLRQ LV IURP PLVFHOODQHRXV VRXUFHV LQFOXGLQJ LQGXVWULDO SURFHVVHV DQG ZDVWH GLVSRVDO 2Q D JOREDO VFDOH WKH DQWKURSRJHQLF 12[ HPLVVLRQ UDWH LV PLQRU FRPSDUHG WR QDWXUDO HPLVVLRQV DQG IRUPDWLRQ LQ WKH HQYLURQPHQW $V D UHVXOW DLU SROOXWLRQ DVVRFLDWHG ZLWK 12[ LV PDLQO\ D ORFDO RU UHJLRQDO SUREOHP &RRSHU DQG $OOH\ :DUN DQG :DUQHU f 7KH OHYHO RI HIIHFW RI QLWURJHQ R[LGHV RQ WKH ORFDO HQYLURQPHQW LV KLJKO\ GHSHQGHQW XSRQ WKH HPLVVLRQ UDWHV RI VRXUFHV DQG SUHYDLOLQJ PHWHRURORJLFDO FRQGLWLRQV DQG WRSRORJLFDO

PAGE 18

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f LQGLFDWH WKDW DLUFUDIW HPLVVLRQV LQWR WKH WURSRVSKHUH DW DQ DOWLWXGH DERYH NP SURGXFH LQFUHDVHG FRQFHQWUDWLRQV RI R]RQH RIIVHWWLQJ R]RQH GHSOHWLRQ EXW FRQWULEXWLQJ WR LQFUHDVHG VXUIDFH WHPSHUDWXUHV JOREDO ZDUPLQJf 0RGHOLQJ VXJJHVWV WKDW WKH UDGLDWLYH IRUFLQJ RI VXUIDFH WHPSHUDWXUH LV DERXW WLPHV PRUH VHQVLWLYH WR DLUFUDIW HPLVVLRQV RI QLWURJHQ R[LGHV WKDQ VXUIDFH HPLVVLRQV ,Q WKH YLFLQLW\ RI DQ DLUSRUW RU $LU )RUFH EDVH ZKHUH JURXQG RSHUDWLRQV WD[LLQJ WDNHRIIV ODQGLQJV DQG PDLQWHQDQFH LQ HQJLQH WHVW FHOOVf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

PAGE 19

7DEOH (PLVVLRQ )DFWRUV IRU 1LWURJHQ 2[LGHV 12[ DV 1f 6RXUFH $YHUDJH (PLVVLRQ )DFWRU DV 1f &RDO +RXVHKROG DQG FRPPHUFLDO OEWRQ RI FRDO EXUQHG ,QGXVWU\ DQG XWLOLWLHV OEWRQ RI FRDO EXUQHG )XHO 2LO +RXVHKROG DQG FRPPHUFLDO OE JDO RI RLO EXUQHG ,QGXVWU\ OE JDO RI RLO EXUQHG 8WLOLW\ OE JDO RI RLO EXUQHG 1DWXUDO *DV +RXVHKROG DQG FRPPHUFLDO OEPLOOLRQ IW RI JDV EXUQHG ,QGXVWU\ OEPLOOLRQ IW RI JDV EXUQHG 8WLOLWLHV OEPLOOLRQ IW RI JDV EXUQHG *DV 7XUELQHV OEPLOOLRQ IW RI JDV EXUQHG :DVWH 'LVSRVDO &RQLFDO LQFLQHUDWRU OEWRQ RI ZDVWH EXUQHG 0XQLFLSDO LQFLQHUDWRU OEWRQ RI ZDVWH EXUQHG 0RELOH 6RXUFH &RPEXVWLRQ *DVROLQHSRZHUHG YHKLFOH OE JDO RI JDVROLQH EXUQHG 'LHVHOSRZHUHG YHKLFOH OE JDO RI RLO EXUQHG $LUFUDIW &RQYHQWLRQDO OEIOLJKW SHU HQJLQH )DQW\SH MHW OEIOLJKW SHU HQJLQH 1LWULF $FLG 0DQXIDFWXUH OEWRQ RI DFLG SURGXFW 6RXUFH :DUN DQG :DUQHU %DVHG XSRQ 86 (3$ $3 f

PAGE 20

1; FRQFHQWUDWLRQV VHHP WR SRLQW WRZDUG LQFUHDVHG UHJXODWRU\ SUHVVXUH WR UHGXFH 12[ HPLVVLRQV IURP MHW HQJLQH WHVW FHOO DFWLYLWLHV LQ WKH QHDU IXWXUH 12 )RUPDWLRQ ,I 12[ VWDQGDUGV IRU WHVW FHOO HPLVVLRQV DUH WR EH FRQVLGHUHG FOHDUO\ WKRVH V\VWHPV DYDLODEOH IRU FRQWUROOLQJ WKHVH HPLVVLRQV PXVW EH WKRURXJKO\ HYDOXDWHG 7R UHGXFH 12[ HPLVVLRQV WZR SULPDU\ DFWLRQV FDQ EH WDNHQ 7KHVH DUH WKH GLUHFW FRQWURO RI WKH UHDFWLRQVf SURGXFLQJ WKH SROOXWDQW FRPEXVWLRQ SURFHVVf RU FRQWURO RI WKH SURFHVV HIIOXHQW 7R XQGHUVWDQG WKH PHDQV RI FRQWUROOLQJ 12[ HPLVVLRQV LW LV FULWLFDO WR H[DPLQH WKH EDVLF FKHPLVWU\ WKHUPRG\QDPLFV DQG NLQHWLFV RI WKH IRUPDWLRQ UHDFWLRQV 2[LGHV RI QLWURJHQ IRUPHG E\ FRPEXVWLRQ SURFHVVHV DUH JHQHUDOO\ FDXVHG HLWKHU E\ WKH FRQYHUVLRQ RI FKHPLFDOO\ ERXQG QLWURJHQ LQ WKH IXHO IXHO 14;f RU E\ WKHUPDO IL[DWLRQ RI DWPRVSKHULF QLWURJHQ LQ WKH FRPEXVWLRQ DLU WKHUPDO 12[f 7KH SULQFLSDO IDFWRUV DIIHFWLQJ WKH IRUPDWLRQ RI QLWURJHQ R[LGHV LQ FRPEXVWLRQ SURFHVVHV DUH WKH DPRXQW RI ILLHOERXQG QLWURJHQ SHDN FRPEXVWLRQ WHPSHUDWXUH R[\JHQ OHYHO DW SHDN WHPSHUDWXUH DQG UHVLGHQFH WLPH LQ WKH FRPEXVWLRQ ]RQH 2UJDQL]DWLRQ IRU (FRQRPLF &RRSHUDWLRQ DQG 'HYHORSPHQW f )XHO 12 :KHQ D IXHO FRQWDLQLQJ RUJDQLFDOO\ ERXQG QLWURJHQ LV EXUQHG WKH FRQWULEXWLRQ RI WKH IXHOERXQG QLWURJHQ WR WRWDO 12[ FDQ EH VLJQLILFDQW 7KH 1& ERQG LV PXFK ZHDNHU WKDQ WKH 11 ERQG LQ PROHFXODU QLWURJHQ VR IXHO QLWURJHQ FDQ EH PRUH HDVLO\ R[LGL]HG WR 12 E\

PAGE 21

R[\JHQ LQ WKH FRPEXVWLRQ JDV 7KH PHFKDQLVP RI WKH FRQYHUVLRQ RI IXHO QLWURJHQ LV EHOLHYHG WR SURFHHG YLD D VHULHV RI LQWHUPHGLDWHV 2QH RI WKH PRVW LPSRUWDQW LQWHUPHGLDWHV LV K\GURJHQ F\DQLGH +&1f ZKLFK LV IXOO\ FRQYHUWHG IURP ERXQG QLWURJHQ LQ WKH UHDFWLRQ ]RQH RI WKH IODPH 0RYLQJ DZD\ IURP WKH UHDFWLRQ ]RQH WKH DPRXQW RI +&1 GHFUHDVHV DV D UHVXOW RI LWV FRQYHUVLRQ LQWR 1 DQG 12[ $PPRQLD 1+f PD\ EH GHWHFWHG DORQJ ZLWK 1 +&1 DQG 12[ LQ WKH SURGXFWV RI FRPEXVWLRQ RI QLWURJHQFRQWDLQLQJ IXHOV LQ WKH SUHVHQFH RI D GHILFLHQF\ RI DWPRVSKHULF QLWURJHQ 7KH FRQYHUVLRQ RI IXHO QLWURJHQ LQWR +&1 LV UDSLG DQG WKH UDWH OLPLWLQJ VWDJH LQ WKH IRUPDWLRQ RI 12[ LV WKH R[LGDWLRQ RI +&1 7KH \LHOG RI 12[ LV LQIOXHQFHG E\ WKH FRQWHQW RI QLWURJHQ LQ WKH IXHO WKH DPRXQW RI H[FHVV DLU DQG WR D OHVVHU H[WHQW FRPEXVWLRQ WHPSHUDWXUH ,VPDJLORY HW DO &RRSHU DQG $OOH\ :DUN DQG :DUQHU f 7KHUPDO 12 'XH WR WKH H[WUHPHO\ KLJK WHPSHUDWXUHV LQ WKH FRPEXVWLRQ ]RQHV RI DLUFUDIW MHW HQJLQHV XS WR .f WKHUPDO QLWURJHQ R[LGHV DUH IRUPHG E\ WKH R[LGDWLRQ RI DWPRVSKHULF QLWURJHQ 7KH DFFHSWHG PRGHO IRU WKH FKHPLFDO UHDFWLRQV UHVSRQVLEOH IRU 12[ IRUPDWLRQ LQ WKH SRVWn FRPEXVWLRQ ]RQH ZDV GHYHORSHG E\ =HOGRYLFK LQ 7KH UHDFWLRQV LQ =HOGRYLFKnV PRGHO DUH DV IROORZV 1 A 12 1 f 1 rr 12 f DQG 1 2+A12 + f

PAGE 22

,Q SUHVHQWLQJ WKHVH UHDFWLRQV LW LV DVVXPHG WKDW WKH IXHO FRPEXVWLRQ UHDFWLRQV EHWZHHQ & + DQG 2f KDYH UHDFKHG HTXLOLEULXP DQG WKDW WKH FRQFHQWUDWLRQV RI 2 + DQG 2+ FDQ EH GHVFULEHG E\ HTXLOLEULXP HTXDWLRQV &RQVLGHULQJ RQO\ WKH WKHUPRG\QDPLFV RI 12[ IRUPDWLRQ WKH WZR RYHUDOO UHDFWLRQV RI FRQFHUQ DUH WKRVH ZKLFK SURGXFH 12 DQG 1 7KH UHOHYDQW HTXLOLEULXP UHDFWLRQV DUH 1 rr 12 f DQG 12 rr 1 f 7KH HTXLOLEULXP FRQVWDQWV IRU UHDFWLRQV f DQG f .SO DQG .S UHVSHFWLYHO\f DUH ZKHUH .S HTXLOLEULXP FRQVWDQW 3 SDUWLDO SUHVVXUH RI FRPSRQHQW L DWPf
PAGE 23

IRUPDWLRQ RI 1 LV IDYRUHG DW ORZ WHPSHUDWXUHV EXW 1 GLVVRFLDWHV EDFN WR 12 DW KLJKHU WHPSHUDWXUHV 7KH UDWH RI HTXLOLEULXP 12 IRUPDWLRQ LQFUHDVHV UDSLGO\ ZLWK LQFUHDVLQJ WHPSHUDWXUH $V WKH FRPEXVWLRQ ]RQH WHPSHUDWXUH ULVHV DERYH WKH IRUPDWLRQ RI 1 EHFRPHV OHVV OLNHO\ XQGHU HTXLOLEULXP FRQGLWLRQV .LQHWLFDOO\ KRZHYHU WKH 12 IRUPDWLRQ UHDFWLRQ LV D VORZ RQH ZKLFK DIIHFWV WKH DYDLODELOLW\ RI UHDFWDQW WR IRUP 1 12 IRUPDWLRQ E\ UHDFWLRQ f KDV D YHU\ KLJK DFWLYDWLRQ HQHUJ\ NLORMRXOHVPROHf DQG LV PRVW OLNHO\ WKH UDWHFRQWUROOLQJ UHDFWLRQ ,VPDJLORY HW DO f ,W KDV EHHQ REVHUYHG H[SHULPHQWDOO\ WKDW 12 FRQFHQWUDWLRQV LQ WKH FRPEXVWLRQ IODPH ]RQH DUH VLJQLILFDQWO\ KLJKHU WKDQ WKRVH SUHGLFWHG E\ WKH =HOGRYLFK PHFKDQLVP 7KLV SURPSW 12 IRUPDWLRQ LV DWWULEXWHG WR VXSHUHTXLOLEULXP UDGLFDO FRQFHQWUDWLRQV WKDW DUH OLNHO\ LQ K\GURFDUERQ IODPHV ,W KDV EHHQ VXJJHVWHG WKDW WKH LQWHUPHGLDWH +&1 LV IRUPHG ZKHQ 1 UHDFWV ZLWK D K\GURFDUERQ UDGLFDO +&1 FRPELQHV ZLWK 2+ WR IRUP &1 DQG WKHQ &1 LV R[LGL]HG WR 12 %DVHG XSRQ H[SHULPHQWDO GDWD 0DF.LQQRQ f GHYHORSHG D PRGHO WKDW SUHGLFWV WKH FRQFHQWUDWLRQ RI 12 IRUPHG GXULQJ FRPEXVWLRQ DV D IXQFWLRQ RI WHPSHUDWXUH 1 DQG FRQFHQWUDWLRQV DQG WLPH $W D WRWDO SUHVVXUH RI RQH DWPRVSKHUH WKH PRGHO HTXDWLRQ LV &12 [ H[S 7f <1<*f f ZKHUH &12 12 FRQFHQWUDWLRQ SSPf 7 DEVROXWH WHPSHUDWXUH r.f
PAGE 24

&RQVLGHUDWLRQ RI HTXDWLRQ f DOORZV LQVLJKW LQWR WKH HIIHFW RI WHPSHUDWXUH RQ 12 IRUPDWLRQ SDUWLFXODUO\ LQ WKH KRW IODPH ]RQH 2QFH H[KDXVW JDVHV PRYH DZD\ IURP WKLV ]RQH WKH\ FRRO UDSLGO\ UHGXFLQJ WKH UHDFWLRQ UDWHV E\ RUGHUV RI PDJQLWXGH ,I H[FHVV R[\JHQ LV SUHVHQW DV WKH JDV FRROV WKH FRQYHUVLRQ RI 12 WR 1 UHDFWLRQ ff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f WKHUHIRUH LW LV VHHQ WKDW ZLWK UHVSHFW WR WKH UHGXFWLRQ RI WKHUPDO 12[ FRPEXVWLRQ FRQWURO VWUDWHJLHV FRXOG EH GHYHORSHG WKDW DUH DLPHG DW D UHGXFWLRQ LQ SHDN WHPSHUDWXUH D UHGXFWLRQ LQ JDV UHVLGHQFH WLPH DW SHDN WHPSHUDWXUH DQG D UHGXFWLRQ RI R[\JHQ FRQFHQWUDWLRQ LQ WKH KLJKWHPSHUDWXUH ]RQH &RRSHU DQG $OOH\ :DUN DQG :DUQHU 2UJDQL]DWLRQ IRU (FRQRPLF &RRSHUDWLRQ DQG 'HYHORSPHQW f 5HJXODWLRQ RI 12 (PLVVLRQV 7KH IHGHUDO SULPDU\ DLU TXDOLW\ VWDQGDUG IRU 1 LV SUHVHQWO\ SSP EDVHG XSRQ DQ DQQXDO DULWKPHWLF PHDQ (PLVVLRQ DQG SHUIRUPDQFH VWDQGDUGV IRU YDULRXV VWDWLRQDU\ VRXUFHV RI 12[KDYH DOVR EHHQ HVWDEOLVKHG $OWKRXJK WKH H[KDXVW IURP MHW HQJLQH WHVW FHOOV

PAGE 25

-(7&Vf LV UHJXODWHG IRU VRRW RSDFLW\ -(7&V FXUUHQWO\ RSHUDWH XQGHU LPSOLFLW H[HPSWLRQ IURP 12[ UHJXODWLRQV DOWKRXJK WKH\ TXDOLI\ DV VWDWLRQDU\ VRXUFHV ,I MHW HQJLQH WHVWLQJ RSHUDWLRQV DUH FRQGXFWHG LQ DQ R]RQH QRQDWWDLQPHQW DUHD RU ZKHUH SKRWRFKHPLFDO VPRJ LV DOUHDG\ D VLJQLILFDQW SUREOHP WKH\ EHFRPH REYLRXV WDUJHWV IRU HQYLURQPHQWDO UHJXODWLRQ 0RELOH VRXUFHV DUH UHJXODWHG XQGHU 7LWOH ,, RI WKH &OHDQ $LU $FW $PHQGPHQWV RI 3DUW % RI WKHVH DPHQGPHQWV DGGUHVVHV DLUFUDIW HPLVVLRQ VWDQGDUGV $ MHW HQJLQH LQ D WHVW FHOO FRXOG EH FRQVLGHUHG WR EH D VWDWLRQDU\ VRXUFH XQGHU 7LWOH DQG PD\ EH LQFOXGHG IRU UHJXODWLRQ XQGHU 7LWOH ,9 $FLG 'HSRVLWLRQ &RQWURO 4XDUOHV DQG /HZLV f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f $IWHU WKH UHFRPPHQGDWLRQV RI WKLV VWXG\ DUH HYDOXDWHG 12[ HPLVVLRQ VWDQGDUGV IRU MHW HQJLQH WHVW FHOOV PD\ EH SURPXOJDWHG DQG HQIRUFHG LQ WKH QHDU IXWXUH $ UHYLHZ RI WKH PRVW FRPPRQ 12[ FRQWURO PHWKRGV IRU VWDWLRQDU\ VRXUFHV LV XVHIXO IRU WKH FRQVLGHUDWLRQ RI WKHLU DSSOLFDELOLW\ WR -(7& H[KDXVW FRQWURO

PAGE 26

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f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

PAGE 27

0RGLILFDWLRQV WR 2SHUDWLQJ &RQGLWLRQV 6HYHUDO PRGLILFDWLRQV WR VWDWLRQDU\ DQG PRELOH VRXUFH RSHUDWLQJ FRQGLWLRQV KDYH EHHQ GHYHORSHG WKDW FDQ EH XVHG WR UHGXFH 12[ IRUPDWLRQ IURP FRPEXVWLRQ 7KHVH LQFOXGH f ORZ H[FHVVDLU ILULQJ f RIIVWRLFKLRPHWULF FRPEXVWLRQ f IOXH JDV UHFLUFXODWLRQ f UHGXFHG DLU SUHKHDW f UHGXFHG ILULQJ UDWHV DQG f ZDWHU LQMHFWLRQ ([FHVV DLU LV WKH DPRXQW RI DLU WKDW LV LQ H[FHVV RI WKH DPRXQW VWRLFKLRPHWULFDOO\ UHTXLUHG IRU SHUFHQW FRPEXVWLRQ RI WKH IXHO 'XH WR LPSHUIHFW PL[LQJ RI DLU DQG IXHO LQ WKH FRPEXVWLRQ ]RQH WKHUH PXVW EH VRPH H[FHVV DLU SUHVHQW DW DOO WLPHV WR UHGXFH IXHO ZDVWH DQG WR SUHYHQW VPRNH IRUPDWLRQ $V H[FHVV DLU GHFUHDVHV 12[ IROORZV ZKLOH XVLQJ OHVV IXHO 2IIVWRLFKLRPHWULF FRPEXVWLRQ DOVR FDOOHG VWDJHGFRPEXVWLRQf EXUQV WKH IXHO LQ WZR RU PRUH VWHSV 7KH LQLWLDO IODPH ]RQH LV IXHOULFK DQG WKH IROORZLQJ ]RQHVf LV DUHf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

PAGE 28

LPSOHPHQWHG XQGHU WKH VWULFW SURWRFROV RI MHW HQJLQH WHVWLQJ :DWHU RU VWHDP LQMHFWLRQ LQWR WKH IXHO FRPEXVWLRQ DLU RU FRPEXVWLRQ ]RQH LV YHU\ HIIHFWLYH DW UHGXFLQJ 12[ IRUPDWLRQ E\ UHGXFLQJ FRPEXVWLRQ WHPSHUDWXUHV &RRSHU DQG $OOH\ f ,QWURGXFLQJ DGGLWLYHV LQWR MHW HQJLQH FRPEXVWLRQ ]RQHV RU IXHOV KRZHYHU LV D SRWHQWLDO SUREOHP GXH WR ORVVHV LQ RXWSXW ZKLFK LV FORVHO\ PRQLWRUHG GXULQJ WHVWLQJf RU SK\VLFDO GDPDJH WR WKH HQJLQH RU WXUELQH EODGHV 6RPH UHVHDUFK KDV EHHQ FRQGXFWHG RQ RWKHU IXHO RU FRPEXVWLRQ ]RQH DGGLWLYHV DV DSSOLHG WR MHW HQJLQH WHVW FHOOV ZKLFK ZLOO EH GLVFXVVHG LQ WKH IROORZLQJ VHFWLRQV &RPEXVWLRQ 0RGLILFDWLRQV &RPEXVWLRQ PRGLILFDWLRQV FDQ UHGXFH 12[ IRUPDWLRQ E\ ORZHULQJ RQH RU PRUH RI WKH SDUDPHWHUV SHDN WHPSHUDWXUH JDV UHVLGHQFH WLPH RU R[\JHQ FRQFHQWUDWLRQ LQ WKH IODPH ]RQH 3HDN WHPSHUDWXUHV FDQ EH UHGXFHG E\ f XVLQJ D IXHOULFK SULPDU\ IODPH ]RQH f LQFUHDVLQJ WKH UDWH RI IODPH FRROLQJ RU f GHFUHDVLQJ WKH DGLDEDWLF IODPH WHPSHUDWXUH E\ GLOXWLRQ H[KDXVW JDV UHFLUFXODWLRQf *DV UHVLGHQFH WLPH LQ WKH SULPDU\ IODPH ]RQH FDQ EH UHGXFHG E\ f FKDQJLQJ WKH VKDSH RI WKH IODPH ]RQH RU f XVLQJ WKH VWHSV OLVWHG LQ WKH SUHYLRXV VWUDWHJ\ 2[\JHQ FRQFHQWUDWLRQ LQ WKH IODPH ]RQH FDQ EH UHGXFHG E\ f GHFUHDVLQJ WKH RYHUDOO H[FHVV DLU UDWHV f FRQWUROOLQJ WKH PL[LQJ RI IXHO DQG DLU DQG f XVLQJ D IXHOULFK SULPDU\ IODPH ]RQH &RRSHU DQG $OOH\ f ,Q JHQHUDO LW LV GLIILFXOW WR HPSOR\ WKHVH FRQWUROV LQ D MHW HQJLQH WHVW FHOO DSSOLFDWLRQ GXH WR WKH FKDQJLQJ SRZHU VHWWLQJV DVVRFLDWHG ZLWK WHVW DQG DFFHSWDQFH UXQV DQG DVVRFLDWHG SHUIRUPDQFH IOXFWXDWLRQV $GGLWLRQDOO\ WKH VWULFW SHUIRUPDQFH VWDQGDUGV UHTXLUHG GXULQJ PLOLWDU\ MHW HQJLQH WHVWLQJ OLPLW WKH DSSOLFDELOLW\ RI FRPEXVWLRQ PRGLILFDWLRQV

PAGE 29

([KDXVW *DV 7UHDWPHQW ([KDXVW JDV WUHDWPHQW DSSHDUV WR EH WKH PRVW DSSOLFDEOH PHWKRG RI FRQWUROOLQJ 12[ HPLVVLRQV GXH WR WKH GLIILFXOW\ LQ XVLQJ FRPEXVWLRQRSHUDWLRQV PRGLILFDWLRQV ZLWK -(7&V DQG WKH SULRULW\ SODFHG RQ XQSHUWXUEHG HQJLQH SHUIRUPDQFH GXULQJ WHVWLQJ 7KH KLJK YROXPH RI H[KDXVW IURP D PLOLWDU\ MHW HQJLQH FRPELQHG ZLWK WKH QHHG IRU KLJKHU OHYHOV RI 12[ UHGXFWLRQ PDNHV WKH VHOHFWLRQ RI WKH RSWLPDO H[KDXVW JDV WUHDWPHQW D GLIILFXOW WDVN 7KHUH DUH WZR PDLQ W\SHV RI H[KDXVW JDV WUHDWPHQW SURFHVVHV ZHW DQG GU\ SURFHVVHV :HW SURFHVVHV W\SLFDOO\ HPSOR\ WKH DEVRUSWLRQ RI 12[ 7KH PDLQ GLVDGYDQWDJH RI ZHW DEVRUSWLRQ LV WKH ORZ VROXELOLW\ RI 12 $GGLWLRQDOO\ WKHVH SURFHVVHV RIWHQ FUHDWH OLTXLG ZDVWH WKDW PD\ UHTXLUH SURSHU GLVSRVDO DW VLJQLILFDQW FRVW )RU WKHVH UHDVRQV GU\ SURFHVVHV DUH JHQHUDOO\ SUHIHUUHG 7KHVH LQFOXGH VHOHFWLYH RU QRQVHOHFWLYH FDWDO\WLF UHGXFWLRQ VHOHFWLYH QRQFDWDO\WLF UHGXFWLRQ LUUDGLDWLRQ DQG GU\ VRUSWLRQ &RRSHU DQG $OOH\ f ,Q VHOHFWLYH FDWDO\WLF UHGXFWLRQ RQO\ WKH 12[ VSHFLHV DUH UHGXFHG WR 1 JDVf OHDYLQJ IUHH R[\JHQ XQUHDFWHG :LWK D VXLWDEOH FDWDO\VW W\SLFDOO\ D SUHFLRXV PHWDO DOWKRXJK D YDULHW\ RI RWKHU FDWDO\VWV KDYH EHHQ VWXGLHGf DPPRQLD 1+f K\GURJHQ +f FDUERQ PRQR[LGH &2f RU K\GURJHQ VXOILGH +6f FDQ EH XVHG DV WKH UHGXFLQJ JDV 1+ LV WKH PRVW FRPPRQO\ XVHG JDV $ QXPEHU RI VWRLFKLRPHWULHV KDYH EHHQ SURSRVHG IRU WKLV UHDFWLRQ )RU H[DPSOH 'DYLQL f VXJJHVWHG 12 1+ r‘ 1 + f $OWKRXJK WKH VHOHFWLYH FDWDO\WLF UHGXFWLRQ SURFHVV LV YHU\ HIIHFWLYH DQG LV FDSDEOH RI DFKLHYLQJ 12 UHGXFWLRQ UDWHV H[FHHGLQJ SHUFHQW LQ FRQWUROOLQJ VWHDG\VWDWH SURFHVVHV LWV

PAGE 30

DSSOLFDWLRQ WR -(7&V LV H[SHQVLYH DQG LQYROYHG FRPSDUHG WR RWKHU WUHDWPHQWV $W WHPSHUDWXUHV IURP WR & 1+ ZLOO UHGXFH 12[ WR 1 ZLWKRXW D FDWDO\VW KRZHYHU 12[ UHGXFWLRQ HIILFLHQFLHV DUH RQO\ SHUFHQW ,I WKH WHPSHUDWXUH LV WRR ORZ XQUHDFWHG 1+ ZLOO EH HPLWWHG DPPRQLD VOLSf DQG LI WKH WHPSHUDWXUH LV WRR KLJK 1+ FDQ EH R[LGL]HG WR 12 7KXV ERWK RI WKHVH VLWXDWLRQV DUH REYLRXV SUREOHPV 'U\ VRUSWLRQ WHFKQLTXHV KDYH WKH DGYDQWDJH RI VLPSOLFLW\ RI RSHUDWLRQ DQG PLQLPL]DWLRQ RI ZDVWH HIIOXHQW 5HODWLYHO\ KLJK 12[ UHGXFWLRQ UDWHV KDYH EHHQ GRFXPHQWHG 1HOVRQ HW DO /\RQ f

PAGE 31

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f RU RWKHU JDVHRXV IXHOVUHGXFLQJ DJHQWV EHIRUH H[SRVLQJ WKH PL[WXUH WR D FDWDO\VW WKH VHFRQG LQYROYHV PL[LQJ WKH HIIOXHQW ZLWK DPPRQLD EHIRUH H[SRVXUH WR D FDWDO\VW DQG WKH WKLUG LQYROYHV H[SRVXUH WR D FDWDO\VW RU DGVRUEHQW ZLWK RU ZLWKRXW PHWKDQH RU RWKHU IXHOf RU DPPRQLD DGGLWLRQV 1XPHURXV PDWHULDOV KDYH EHHQ SURSRVHG DV FDWDO\VWV DQG D QXPEHU RI WKHVH KDYH EHHQ SDWHQWHG KRZHYHU WKRVH UHFHLYLQJ WKH PRVW DWWHQWLRQ KDYH EHHQ SODWLQXP RU RWKHU SUHFLRXV PHWDOV RU PHWDOOLF R[LGHV IURP WKH DONDOLQH HDUWK JURXS 1HOVRQ HW DO .OLPLVFK DQG /DUVRQ 0HXEXV f

PAGE 32

&DWDO\WLF 0HWKRGV &RKQ >86 3DWHQW f@ KDV EHHQ LVVXHG D SDWHQW WKDW GHVFULEHV D SURFHVV IRU SXULI\LQJ ZDVWH JDVHV FRQWDLQLQJ QLWURJHQ R[LGHV E\ PL[LQJ WKHP ZLWK D IXHO VXFK DV &+f DQG SDVVLQJ WKH PL[WXUH RYHU D SODWLQXP RU UKRGLXPFRQWDLQLQJ FDWDO\VW DW DQ LQLWLDO UHDFWLRQ WHPSHUDWXUH RI ) $FUHV DQG +XWFKLQJV >86 3DWHQW f@ GHVFULEH D VLPLODU SURFHVV &KLOGHUV (OOLV DQG 5\DQ >86 3DWHQW f@ GHVFULEH D PHWKDQH SURFHVV LQYROYLQJ WZR FDWDO\VW EHGV LQ VHULHV RQH FRQWDLQLQJ SODWLQXP RU DOXPLQD DQG WKH VHFRQG FRQWDLQLQJ QLFNHO RU DOXPLQD 9DQDGLXP R[LGH PRO\EGHQXP R[LGH DQGRU WXQJVWHQ R[LGH FDWDO\VWV RQ DOXPLQD RU VLOLFLF DFLG VXEVWUDWHV ZHUH XVHG LQ FRPELQDWLRQ ZLWK &+ DV GHVFULEHG E\ 1RQQHQPDFKHU DQG .DUWWH >86 3DWHQW f@ 6XFK EDVH PHWDO FDWDO\VWV DV LURQ FREDOW QLFNHO DQG FRSSHU GLVSHUVHG RQ D UHIUDFWRU\ VXSSRUW ZHUH XVHG E\ 5HLWPHLHU >86 3DWHQW ,f@ 2WKHU SDWHQWV UHODWLQJ WR WKH XVH RI PHWKDQH DQG SUHFLRXV PHWDO FDWDO\VWV IRU WKH UHPRYDO RI 12[ IURP ZDVWH RU WDLO JDVHV LQFOXGH WKRVH RI $QGHUVHQ DQG *UHHQ >86 3DWHQW f@ 5RPHR >86 3DWHQW f@ 1HZPDQ >86 3DWHQW f@ .DQGHOO DQG 1HPHV >86 3DWHQW f@ $QGHUVHQ 5RPHR DQG *UHHQ >86 3DWHQW f@ DQG +DUGLVRQ DQG %DUU >86 3DWHQW f@ 1HOVRQ HW DOf /HZLV f 7KH XVH RI DPPRQLD DV D UHGXFLQJ JDV LQ WKH SUHVHQFH RI D FDWDO\VW ZDV GHVFULEHG HDUOLHU %DLNHU HW DO D Ef DQG &KHQ HW DO f GHVFULEH WKH VHOHFWLYH FDWDO\WLF UHGXFWLRQ RI QLWULF R[LGH ZLWK DPPRQLD XSRQ FDWDO\VWV FRPSULVHG RI PRQR DQG PXOWLOD\HUV RI YDQDGLD VXSSRUWHG RQ WLWDQLD DQG PRQROD\HUV RI YDQDGLD LPPRELOL]HG RQ WLWDQLDVLOLFD

PAGE 33

'DYLQL f VWXGLHG WKH UHGXFWLRQ RI QLWURJHQ R[LGH ZLWK DPPRQLD LQ WKH SUHVHQFH RI FDUERQDFHRXV VRRWV IURP LQGXVWULDO ERLOHUV /HH DQG .OLQH >86 3DWHQW f@ GHVFULEH D PHWKRG IRU UHPRYLQJ 12 LQ WKH SUHVHQFH RI VXOIXU GLR[LGH E\ PL[LQJ DPPRQLD ZLWK WKH IOXH JDV LQ WKH SUHVHQFH RI D FDWDO\VW VHOHFWHG IURP WKH JURXS RI SODWLQXP DQG WUDQVLWLRQ PHWDOV DQG R[LGHV DQG PL[WXUHV RI WKHVH $QGHUVHQ DQG .HLWK >86 3DWHQW f@ KDYH GHVFULEHG WKH XVH RI 1+ LQ FRPELQDWLRQ ZLWK FREDOW QLFNHO RU LURQ VXSSRUWHG RQ DOXPLQD VLOLFD VLOLFD JHO RU GLDWRPDFHRXV HDUWK /\RQ f GHVFULEHV DQ LPSURYHG PHWKRG XVLQJ DPPRQLXP LQMHFWLRQ LQ D WLJKWO\ FRQWUROOHG QRQFDWDO\WLF SURFHVV $ SUREOHP ZLWK WKLV PHWKRG LV WKH IRUPDWLRQ RI DPPRQLXP VDOW SDUWLFXODWH PDWWHU 2WKHU SDWHQWV UHODWLQJ WR WKH XVH RI DPPRQLD DQG SUHFLRXV PHWDO RU RWKHU FDWDO\VWV IRU WKH UHPRYDO RI R[LGHV RI QLWURJHQ IURP D ZDVWH RU WDLO JDV LQFOXGH WKRVH RI &RKQ 6WHHOH DQG $QGHUVHQ >86 3DWHQW f@ DQG .HLWK DQG .HQDK >86 3DWHQWV f DQG f@ 1HOVRQ HW DO f &RQWURO RI QLWURJHQ R[LGHV LV DOVR DWWDLQDEOH E\ VHYHUDO SURFHVVHV WKDW GR QRW XWLOL]H PHWKDQH RU DPPRQLD EHFDXVH WKH ODWWHU PD\ EH GLIILFXOW WR FRQWURO $GGLWLRQDOO\ WKH GHVLUHG DPPRQLD UHDFWLRQ RFFXUV RQO\ ZLWKLQ D QDUURZ UDQJH RI WHPSHUDWXUHV 2QH SURFHVV WKDW UHFHLYHG JUHDW DWWHQWLRQ ZDV WKH 5$35(12[ VHOHFWLYH UHGXFWLRQ UHDFWLRQ ILUVW GHYHORSHG E\ 3HUU\ DQG 6LHEHUV f ,Q WKLV SURFHVV F\DQXULF DFLG QRQWR[LFf LV PL[HG ZLWK 0Dn LQWDLQLQJ H[KDXVW JDVHV IURP ZKLFK WKH 12[ FRQFHQWUDWLRQ DV 12f LV VLJQLILFDQWO\ UHGXFHG 17,6 7HFK 1RWHV f +HDS HW DO f FRPELQHG WKLV WHFKQRORJ\ ZLWK FRPEXVWLRQ PRGLILFDWLRQ WHFKQLTXHV WR IXUWKHU LPSURYH WKH UHGXFWLRQ RI 12[ 8QIRUWXQDWHO\ LW DSSHDUV WKDW WKLV UHDFWLRQ LV QHJDWLYHO\ DIIHFWHG E\ WKH SUHVHQFH RI IUHH R[\JHQ :LFNH HW DO f

PAGE 34

,W KDV EHHQ GHWHUPLQHG WKDW YDULRXV DONDOL PHWDO R[LGHV DQG VLOLFDWH PDWHULDOV FDQ UHGXFH QLWURJHQ R[LGHV ZLWKRXW DGGLWLRQDO PHWKDQH RU DPPRQLD +DUULV 0RUHOOR DQG 3HWHUV >86 3DWHQW f@ KDYH VXFFHVVIXOO\ HYDOXDWHG &D2 6U2 %D2 1D DQG RWKHUV VXSSRUWHG RQ $OXQGXP FHPHQW SRUFHODLQ VLOLFD H[WHQGHG DOXPLQD RU DOXPLQD EHDGV +HDY\ PHWDO FDWDO\VWV FRSSHU VLOYHU QLFNHO PRO\EGHQXP SDOODGLXP DQG FREDOWf VXSSRUWHG RQ DOXPLQD ZHUH SDWHQWHG E\ 5\DVRQ >86 3DWHQW f@ 7KH FDWDO\WLF SURSHUWLHV RI FRSSHU IRU WKH UHGXFWLRQ RI 12[ ZLWKRXW DGGHG PHWKDQH RU DPPRQLD DUH GHVFULEHG E\ YDULRXV LQYHVWLJDWRUV LQFOXGLQJ *HKUL DQG )UHYHO >86 3DWHQW f@ DQG .UHVVOH\ >86 3DWHQW f@ .UDQF DQG /XWFKNR >86 3DWHQW f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f HYDOXDWHG DFWLYDWHG FDUERQ DOXPLQRVLOLFDWH VLOLFD JHO PDQJDQHVH GLR[LGH FRSSHU GLR[LGH DQG FRNH IRU WKH UHPRYDO RI KLJK FRQFHQWUDWLRQV RI 1 DQG IRXQG DOXPLQRVLOLFDWH WR EH WKH PRVW VXLWDEOH VRUEHQW .\ROOHQ >86 3DWHQW f@ GHVFULEHG D SURFHVV HPSOR\LQJ HLWKHU VRGLXP FDUERQDWH FDOFLXP FDUERQDWH RU

PAGE 35

FDOFLXP R[LGH WKDW ZDV KLJKO\ GHSHQGHQW XSRQ WHPSHUDWXUH DQG KXPLGLW\ FRQGLWLRQV .LWDJDZD >86 3DWHQW f@ GHYHORSHG D GU\ PHWKRG WKDW XWLOL]HV LQRUJDQLF VDOWV VXFK DV 3E6 .0Q .& 1D&,2 1DA2 .6 DQG 1D+3 DQG LQRUJDQLF R[LGHV OLNH $V DQG 3E DV VXLWDEOH DGVRUEHQWV &ROOLQV >86 3DWHQW f@ GHULYHG D WZRVWDJH SURFHVV LQYROYLQJ VLOLFD JHO DQG DQ DFWLYDWHG FU\VWDOOLQH ]HROLWLF PROHFXODU VLHYH 7KH XVH RI DFWLYDWHG FDUERQ WR DGVRUE QLWURJHQ R[LGHV KDV EHHQ VWXGLHG H[WHQVLYHO\ %DGMDL HW DO :DUN DQG :DUQHU ff ,WV KLJK VXUIDFH DUHD DGVRUSWLRQ UDWH DQG FDSDFLW\ DQG UHJHQHUDELOLW\ DUH ZHOO NQRZQ -DPHV DQG +XJKHV f XVHG FDOFLQHG OLPHVWRQHV DQG GRORPLWHV 0J2FRQWDLQLQJf WR UHPRYH 12 DW KLJK FRQFHQWUDWLRQV ZLWK JRRG UHVXOWV .LNNLQLGHV DQG
PAGE 36

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f ([KDXVW JDVHV DUH EORZQ LQWR D ODUJH ORQJ WXEH DXJPHQWHU WXEHf ZLWK D FRQYHUJHQW HQWUDQFH VHFWLRQ 7KH SXUSRVHV RI WKH DXJPHQWHU DUH WR GUDZ DLU LQWR WKH WHVW FHOO DQG HQJLQH ZLWK HTXDO DLU SUHVVXUH DW WKH LQOHW DQG RXWOHW RI WKH HQJLQH 1RWH YHQWXUL VKDSHf WR GUDZ D SRUWLRQ RI WKH DLU DURXQG WKH HQJLQH KRXVLQJ WR SURYLGH FRROLQJ VLPLODU WR WKDW H[SHULHQFHG GXULQJ IOLJKW DQG WR GLOXWH DQG FRRO HQJLQH H[KDXVW WR SUHYHQW GDPDJH WR FRQVWUXFWLRQ PDWHULDOV $XJPHQWHG JDV WHPSHUDWXUHV FDQ YDU\ IURP DERXW WR ) &f GHSHQGLQJ XSRQ HQJLQH ILULQJ UDWH DQG DXJPHQWDWLRQ UDWLR *HQHUDOO\ WKH WHPSHUDWXUH RI WKHVH JDVHV LV OHVV WKDQ ) &f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f 'HSHQGLQJ XSRQ WKH HQJLQH EHLQJ WHVWHG WKH YROXPH IORZ UDWH RI JDV OHDYLQJ WKH FHOO FDQ EH

PAGE 37

&RPEXVWLRQ DQG $XJPHQWDWLRQ $LU ([KDXVW )LJXUH 6FKHPDWLF 5HSUHVHQWDWLRQ RI D -HW (QJLQH 7HVW &HOO

PAGE 38

HQRUPRXV $ XQLTXH IHDWXUH RI MHW HQJLQH WHVW FHOOV LV WKHLU KLJKO\ YDULDEOH WHVW F\FOHV DQG 12[ SURGXFWLRQ DVVRFLDWHG ZLWK YDU\LQJ SHDN IODPH WHPSHUDWXUHf %HFDXVH RI WKLV YDULDELOLW\ D12[ FRQWURO WHFKQRORJ\ DSSOLFDEOH WR MHW HQJLQH WHVW FHOOV PXVW SHUIRUP RYHU D ZLGH UDQJH RI UDSLGO\ YDU\LQJ RSHUDWLQJ FRQGLWLRQV 5HVHDUFK 6SRQVRUHG EY WKH 86 $LU )RUFH 86$)f 7KH $LU )RUFH WKURXJK WKH 86 $) (QJLQHHULQJ DQG 6HUYLFHV &HQWHU DQG 86$) &LYLO (QJLQHHULQJ 6XSSRUW $JHQF\ QRZ WKH +XPDQ 6\VWHPV &HQWHU$UPVWURQJ /DERUDWRU\ 7\QGDOO $)% )ORULGDf KDV UHFHQWO\ VSRQVRUHG VHYHQ LQGHSHQGHQW SURMHFWV WKDW VSHFLILFDOO\ WDUJHW WKH FRQWURO RI 12[ IURP -(7&V -RKQVRQ DQG .DW] f VKRZHG WKDW UHEXUQLQJ RI H[KDXVW ZLWK H[WUD IXHO ZDV HIIHFWLYH EXW QRW HFRQRPLFDO +DP HW DO f UHSRUWHG SDUWLDO VXFFHVV ZLWK QRQFDWDO\WLF UHGXFWLRQ XVLQJ DGGLWLYHV DPPRQLD DPLQHV DQG K\GUD]LQHVf WR VLPXODWHG H[KDXVW 8QIRUWXQDWHO\ DV WHPSHUDWXUH LQFUHDVHG 12[ SURGXFWLRQ LQFUHDVHG $ GXDOEHG PRUGHQLWHFRSSHU FDWDO\VW GLG RIIHU VRPH LPSURYHPHQW LQ VHQVLWLYLW\ WR YDULDEOH FRQGLWLRQV RYHU VRPH EHWWHU NQRZQ VHOHFWLYH FDWDO\VWV .LWWUHOO f %HUPDQ HW DO f HYDOXDWHG WKH SKRWRSURPRWLRQ RI 12 WKHUPDO GHFRPSRVLWLRQ E\ YDULRXV PHWDOOLF R[LGHV 7KLV GHFRPSRVLWLRQ ZDV YHU\ HIIHFWLYH EXW RQO\ LQ WKH DEVHQFH RI R[\JHQ ,QLWLDO ZRUN UHSRUWHG E\ 3HWULN f DOVR IRXQG WKDW 12[ UHGXFWLRQ E\ FHUDPLFVXSSRUWHG HOHFWURFDWDO\VWV ZDV LQKLELWHG E\ HYHQ WUDFH DPRXQWV RI R[\JHQ KRZHYHU FRQWLQXHG GHYHORSPHQW RI WKHVH PDWHULDOV KDV SURGXFHG FDWDO\WLF FHUDPLFV WKDW PD\ EH HIIHFWLYH LQVLGH WKH MHW HQJLQH *RUGRQ f /\RQ f HYDOXDWHG 12[ UHGXFWLRQ DW WKH VXUIDFH RI KHDYLHU *URXS ,OD R[LGHV EXW

PAGE 39

WKH WR[LFLW\ RI WKHVH KHDY\ PHWDOV LV D GHWULPHQW 7KH VHYHQWK VWXG\ ZKLFK ZDV VHOHFWHG IRU GHYHORSPHQW LQYROYHG YHUPLFXOLWH DQG PDJQHVLXP R[LGH 0J2fFRDWHG YHUPLFXOLWH 12[ UHPRYDO UDWHV IRU WKLV PHGLXP ZHUH DSSHDOLQJ DQG WKH WHFKQRORJ\ KDV PDQ\ RWKHU DGYDQWDJHV LQFOXGLQJ HFRQRPLFV VLPSOLFLW\ UHJHQHUDELOLW\ DQG SDUWLFXODUO\ LQVHQVLWLYLW\ WR FKDQJHV LQ H[KDXVW JDV FRQGLWLRQV RU FRPSRVLWLRQ 1HOVRQ HW DO f 7KH 86$) KDV VSRQVRUHG QXPHURXV VWXGLHV LQ RUGHU WR TXDQWLI\ HPLVVLRQV IURP MHW HQJLQHV 6HLWFKHN f )DJLQ f DQG 6SLFHU HW DO f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f$Of$O6Lf22+f+2 D WKHUPDOO\ H[SDQGHG IRUP RI D FRPPRQ DOXPLQRVLOLFDWH PDWHULDOf KDV EHHQ XVHG DV D FDUULHU IRU RWKHU PDWHULDOV LQFOXGLQJ QLWUDWHV DQG SKRVSKDWHV LQ ZKLFK WKH FRPELQDWLRQV DUH XVHG DV VORZUHOHDVH IHUWLOL]HUV (YDQVKHQ >86 3DWHQW f@ GHVFULEHV WKH WUHDWPHQW RI IOXH JDVHV ZLWK SRO\YLQ\OS\UUROLGRQH ZLWK RU ZLWKRXW WKH DGGLWLRQ RI D FDWDO\VW RI D QLWUDWH RU VXOIDWH RI FRSSHU RU VLOYHU +H VXJJHVWHG VXVSHQGLQJ WKHVH PDWHULDOV RQ D FDUULHU VXFK DV YHUPLFXOLWH 6DQLWHFK ,QF QRZ 6RUEHQW 7HFKQRORJLHV &RUSRUDWLRQ 7ZLQVEXUJ 2KLRf GHYHORSHG D QHZ FODVV RI

PAGE 40

VRUEHQWV RULJLQDOO\ GHVLJQHG IRU WKH FRQWURO RI 6 WKDW ZHUH EDVHG XSRQ 0J2 DQ DONDOLQH HDUWK >+D@ PHWDO R[LGHf FRDWHG RQWR YHUPLFXOLWH RU SHUOLWH >1HOVRQ 86 3DWHQWV DQG f@ 6 UHGXFWLRQ LV EHOLHYHG WR RFFXU SULPDULO\ E\ UHDFWLRQ ZLWK 0J2 IRUPLQJ FRPSOH[ VXOILWHV RU VXOIDWHV ,W ZDV DVVXPHG WKDW 12[ FRXOG EH UHPRYHG LQ D VLPLODU PDQQHU IRUPLQJ QLWULWHV RU QLWUDWHV 6DQLWHFK UHVHDUFKHUV EHJDQ HYDOXDWLQJ H[SDQGHG YHUPLFXOLWH DV D FDWDO\VW LQ $ SDWHQW GHVFULELQJ WKH XVH RI YHUPLFXOLWH DV D VHOHFWLYH UHGXFWLRQ FDWDO\VW >86 3DWHQW @ ZDV JUDQWHG WR 6DQLWHFK LQ ,Q $XJXVW WKH $LU )RUFH DZDUGHG 6DQLWHFK D 6PDOO %XVLQHVV ,QQRYDWLRQ 5HVHDUFK 6%,5f SURMHFW WR IXUWKHU GHYHORS WKH QHZ FDWDO\VW IRU DSSOLFDWLRQ WR MHW HQJLQH WHVW FHOOV 1HOVRQ HW DO f UHSRUWHG WKH UHVXOWV RI D VHULHV RI WHVWV GHVLJQHG WR HVWDEOLVK WKH DSSOLFDELOLW\ DQG WHFKQLFDO IHDVLELOLW\ RI XVLQJ YHUPLFXOLWH DV D FDWDO\VW WR UHGXFH 12[ IURP VLPXODWHG MHW HQJLQH H[KDXVW 7KHVH VWXGLHV EHJLQQLQJ DV D VHOHFWLYHFDWDO\WLF DSSURDFK HYDOXDWHG KRZ YHUPLFXOLWH YDULDEOHV W\SH DQG VL]Hf EHG VL]H VROLG DGGLWLYHV WR YHUPLFXOLWH IOXH JDV YDULDEOHV WHPSHUDWXUH FRPSRVLWLRQ IORZ UDWHf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

PAGE 41

LPSURYHG 12[ UHGXFWLRQ DQG LQFUHDVHG WKH VHQVLWLYLW\ RI WKH SURFHVV WR WHPSHUDWXUH FKDQJH 9HUPLFXOLWH FRDWHG ZLWK 0J2 UHPRYHG DSSUR[LPDWHO\ SHUFHQW RI 12[ DQG WKH SURFHVV ZDV UHODWLYHO\ LQVHQVLWLYH WR FKDQJHV LQ ZDVWH JDV FRPSRVLWLRQ YHORFLW\ DQG WHPSHUDWXUH ,W LV EHOLHYHG WKDW WKH 12[ VSHFLHV ZHUH FKHPLFDOO\ DGVRUEHG E\ WKH 0J2YHUPLFXOLWH SURGXFW DQG DWWDFKHG DW WKH VXUIDFH DV FRPSOH[ PDJQHVLXP QLWULWHV DQGRU QLWUDWHV 7DQDEH DQG )XNXGD f IRXQG WKDW 0J2 KDV D KLJKHU QXPEHU RI DFWLYH EDVLFf VLWHV FRPSDUHG WR RWKHU DONDOLQH HDUWK PHWDO R[LGHV LH &D2 6U2 DQG %D2f =KDQJ HW DO f IXUWKHU DWWULEXWHG WKH DFWLYLW\ RI 0J2 VXUIDFHV WR R[\JHQ YDFDQFLHV LQ WKH ODWWLFH VWUXFWXUH 6RPH VHOHFWHG SK\VLFDO SURSHUWLHV RI 0J2 0J2+f DQG PDJQHVLXP QLWULWH DQG QLWUDWHV DUH VKRZQ LQ 7DEOH 7DEOH 3K\VLFDO SURSHUWLHV RI VHOHFWHG 0JFRQWDLQLQJ LQRUJDQLF FRPSRXQGV 6SHFLHV 0RO :W JJPROf 'HQVLW\ JFPf 0RO 9RO FPJ PROf 0HOWLQJ 3RLQW r&f %RLOLQJ 3RLQW r&f 0J2 0J2+f 6HH 1RWH 0RI1I} 77 If 0J1f+ 0J1f+ 'HFRPS 1RWH /RVHV ZDWHU RI K\GUDWLRQ DW & 6RXUFH &5& +DQGERRN RI &KHPLVWU\ DQG 3K\VLFV UG (GLWLRQ

PAGE 42

,QLWLDO VWXGLHV IRXQG WKDW WKH VRUEHQW PDWHULDO FRXOG EH VXFFHVVIXOO\ UHJHQHUDWHG ZLWK PLQRU UHGXFWLRQ LQ DGVRUSWLRQ FDSDFLW\ RI WKH UHJHQHUDWHG PDWHULDO 5XFK HW DO f FRQILUPHG WKDW WKH 0J2f§YHUPLFXOLWH12[ UHDFWLRQ LV PRVW OLNHO\ GXH WR FKHPLVRUSWLRQ DQG WKDW WKH YHUPLFXOLWH12[ VRUSWLRQ LV D UHVXOW RI SK\VLFDO DGVRUSWLRQ EDVHG XSRQ K\VWHUHVLV GXULQJ OLPLWHG DGVRUSWLRQGHVRUSWLRQ WHVWV 7KLV REVHUYHG K\VWHUHVLV LV PRVW SUREDEO\ WKH UHVXOW RI WKH IRUFHV RI FKHPLVRUSWLRQ JHQHUDOO\ LQYROYLQJ HOHFWURQ WUDQVIHU EHWZHHQ WKH VROLG DGVRUEHQWf DQG JDV DGVRUEDWHf PROHFXOHV KHOG DW WKH VXUIDFH 7KHVH IRUFHV DUH W\SLFDOO\ PXFK VWURQJHU WKDQ WKH LQWHUPROHFXODU YDQ GHU :DDOV IRUFHV DVVRFLDWHG ZLWK SK\VLFDO DGVRUSWLRQ .RYDFK 6]HNHO\ HW DO f 7KH\ DOVR HYDOXDWHG WKH VXUIDFH DUHDV RI WKHVH SURGXFWV XVLQJ 1 DQG ZDWHU DGVRUSWLRQ DQG %UXQDXHU(PPHWW7HOOHU %(7f WKHRU\ DQG IRXQG WKH VRUEHQW VXUIDFH DUHDV WR EH ODUJHU WKDQ ZRXOG EH H[SHFWHG IURP WKH FROOHFWLYH VXUIDFH DUHDV RI WKH YHUPLFXOLWH DQG 0J2 7KLV VXJJHVWV WKDW WKH YHUPLFXOLWH VXSSRUW HIIHFWLYHO\ HQKDQFHV WKH VXUIDFH DUHD GHYHORSPHQW RI WKH 0J2 5HSRUWHG QLWURJHQ LVRWKHUP VXUIDFH DUHDV UDQJHG IURP WR PJ IRU VRUEHQWV SUHSDUHG DW & 6LHYHG VDPSOHV KDG D PD[LPXP VXUIDFH DUHD RI PJ )RXULHU WUDQVIRUP LQIUDUHG )7,5f VROLG VSHFWUD LQGLFDWHG WKH SUHVHQFH RI 0J2+f ZKLFK GLVDSSHDUHG DV FRQGLWLRQLQJ WHPSHUDWXUH LQFUHDVHG 1LWULWH ZDV IRXQG LQ VRPH VDPSOHV H[SRVHG WR ORZFRQFHQWUDWLRQ IOXH JDVHV E\ )7,5 VROLG DQG VROXWLRQ VSHFWUD EXW UHVXOWV ZHUH GHWHUPLQHG WR EH LQFRQFOXVLYH ;UD\ GLIIUDFWLRQ LQWHUSODQDU VSDFLQJV VKRZHG WKDW 0J1f ZDV SUHVHQW

PAGE 43

)LHOG 6WXGLHV 3LORWVFDOH WHVWV RI WKLV 0J2YHUPLFXOLWH VRUEHQW ZHUH FRQGXFWHG RQ D -(7& IDFLOLW\ DW 7\QGDOO $)% 3DQDPD &LW\ )ORULGD EHWZHHQ -XQH DQG RQ 6HSWHPEHU $ WZRVWDJH ILOWHU EHG GHVLJQ FRQVLVWLQJ RI IRXU LQFKHV RI YLUJLQ YHUPLFXOLWH LQ IURQW RI HLJKW LQFKHV RI 0J2FRDWHG YHUPLFXOLWH ZDV HYDOXDWHG WKURXJK IRXU DSSUR[LPDWHO\ PLQXWH UXQV RI VXEVFDOH GURQH HQJLQHV $Q DGGLWLRQDO UXQ ZDV FRQGXFWHG WKURXJK IRXU LQFKHV RI DFWLYDWHG FDUERQ IROORZHG E\ WKH HLJKWLQFK 0J2f§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f K\GURFDUERQV SULPDULO\ HWK\OHQHf SDVVHG WKURXJK WKH EHGV WKDW GLG QRW FRQWDLQ DFWLYDWHG FDUERQ 1HOVRQ HW DO :DQGHU DQG 1HOVRQ f 0RUH UHJHQHUDWLRQ VWXGLHV ZHUH FRQGXFWHG E\ 1HOVRQ HW DO f WR DWWHPSW WR RSWLPL]H WKH UHJHQHUDWLRQ RI 0J2f§YHUPLFXOLWH DIWHU VDWXUDWLRQ ZLWK 12[ DQG 6f 7KH VRUEHQWV ZHUH DOWHUQDWHO\ H[SRVHG WR VLPXODWHG MHW HQJLQH H[KDXVW DQG UHJHQHUDWHG LQ UHGXFLQJ DWPRVSKHUHV IRU D QXPEHU RI WUHDWPHQW F\FOHV WR HYDOXDWH 12[ VRUSWLRQ FDSDELOLWLHV

PAGE 44

3HUIRUPDQFH GXULQJ VRUSWLRQGHVRUSWLRQ F\FOHV ZDV HYDOXDWHG RQ WKH EDVLV RI ZHLJKW JDLQ DQG ORVV ,W DSSHDUHG WKDW WKHUH ZDV D ORVV RI VXUIDFH DUHD ZKHQ WKH PDWHULDO ZDV UHJHQHUDWHG DW H[FHVVLYHO\ KLJK WHPSHUDWXUHV &f 5HJHQHUDWLRQ FRQVLVWHG RI WKUHH VWHSV f GU\LQJ f KHDWLQJ WR D WHPSHUDWXUH DERYH & LQ WKH SUHVHQFH RI 1 RU &2 RU &+ WR UHGXFH RIIn JDVVLQJ 12[f DQG f FRROLQJ 2I WKHVH &2 LV WKH JDV RI FKRLFH +HDWLQJ GXULQJ UHJHQHUDWLRQ ZDV DFFRPSOLVKHG XVLQJ D URWDU\NLOQ FDOFLQHU D YHUWLFDO WXEH DSSDUDWXV DQG D KRUL]RQWDO FRQYH\RU NLOQ :LFNKDP DQG .RHO f KDYH SXEOLVKHG DQ H[FHOOHQW UHYLHZ RQ WKH UHGXFWLRQ RI 12[ E\ &2 5HODWHG 5HVHDUFK $ UHYLHZ RI FXUUHQW OLWHUDWXUH LQGLFDWHV WKDW QR UHVHDUFK RWKHU WKDQ WKDW DOUHDG\ GHVFULEHG KDV EHHQ FRQGXFWHG RQ WKLV 0J2YHUPLFXOLWH VRUEHQW $V D UHVXOW QR GDWD RQ LQWULQVLF NLQHWLFV RU UHDFWLRQ PHFKDQLVPVf H[LVW 6LJQLILFDQW DPRXQWV RI UHODWHG UHVHDUFK KRZHYHU KDYH EHHQ FRQGXFWHG RQ VLPLODU DFLGDFLG DQK\GULGH JDVEDVHJDVVROLG V\VWHPV 7KH UHVXOWV RI WKHVH VWXGLHV KDYH EHHQ XVHG WR GHVLJQ DQG FRQVWUXFW DSSURSULDWH H[SHULPHQWDO V\VWHPV WR FROOHFW GDWD IRU LQWHUSUHWDWLRQ XVLQJ SURYHQ DQG DFFHSWHG VRUSWLRQ VWXG\ WHFKQLTXHV $ VKRUW UHYLHZ RI VRPH RI WKH IXQGDPHQWDOV RI JDVf§VROLG LQWHUDFWLRQV DQG FKHPLVRUSWLRQ LV QHFHVVDU\ WR EHWWHU XQGHUVWDQG WKH FRPSOH[ SKHQRPHQD LQYROYHG LQ WKHVH KHWHURJHQHRXV V\VWHPV

PAGE 45

*DV6ROLG ,QWHUDFWLRQ 7KH VPDOOHVW XQLW RI D JDVVROLG V\VWHP FDQ EH UHSUHVHQWHG E\ WKH LQWHUDFWLRQ EHWZHHQ D VLQJOH VROLG SDUWLFOH DQG D IORZLQJ JDV VWUHDP 7KLV UHSUHVHQWDWLRQ LV FRQYHQLHQW DQG VLPSOH DQG LQ SULQFLSOH PD\ EH JHQHUDOL]HG WR PRUH FRPSOH[ PXOWLSDUWLFOH V\VWHPV VXFK DV SDFNHG EHGV 7KHVH LQWHUDFWLRQV FDQ EH FDWDO\WLF LQ ZKLFK WKH VROLG IXQFWLRQV DV D FDWDO\VW RIWHQ LQ WKH IDFLOLWDWLRQ RI D JDVSKDVH UHDFWLRQ f§ RU QRQFDWDO\WLF LQ ZKLFK WKH VXUIDFH DFWV WR SK\VLFDOO\ RU FKHPLFDOO\ UHPRYH UHDFWDQWV IURP WKH JDV SKDVH $GVRUSWLRQ RFFXUV ZLWK D GHFUHDVH LQ VXUIDFH IUHH HQHUJ\ DQG QRUPDOO\ D GHFUHDVH LQ HQWURS\ DV D UHVXOW LW LV JHQHUDOO\ DQ H[RWKHUPLF SURFHVV ,Q SK\VLFDO DGVRUSWLRQ VRUEHG VSHFLHV DUH DWWUDFWHG DQG KHOG DW WKH VROLG VXUIDFH E\ YDQ GHU :DDOV GLVSHUVLRQ IRUFHV 7KHVH IRUFHV DUH PXFK ZHDNHU WKDQ FKHPLFDO ERQGLQJ IRUFHV WKH KHDW HYROYHG LV VPDOO JHQHUDOO\ EHWZHHQ DQG NFDOJPROH N-JPROHf DSSUR[LPDWHO\ WKH KHDW RI YDSRU FRQGHQVDWLRQ IRU UHODWLYHO\ QRQSRODU VXEVWDQFHV &KHPLVRUSWLRQ LV SULPDULO\ UHVSRQVLEOH IRU JDVf§VROLG UHDFWLRQV DQG FDWDO\VLV RQ VROLG VXUIDFHV ,Q FKHPLVRUSWLRQ WKH IRUFHV LQYROYHG DUH RI WKH VDPH RUGHU RI PDJQLWXGH DV WKRVH IRU FKHPLFDO ERQGLQJ DQG WKH KHDW RI FKHPLVRUSWLRQ LV RIWHQ EHWZHHQ DQG NFDOJPROH N-JPROHf 7KH SURFHVV RI SK\VLFDO DGVRUSWLRQ OLNH FRQGHQVDWLRQ KDV YHU\ OLWWOH RU QR DFWLYDWLRQ HQHUJ\ DQG DV D UHVXOW LV DVVXPHG WR RFFXU UDSLGO\ &KHPLVRUSWLRQ RIWHQ GLVSOD\V D KLJKHU DFWLYDWLRQ HQHUJ\ LPSO\LQJ D VORZ UDWH KRZHYHU FKHPLVRUSWLRQ UHDFWLRQV KDYLQJ VPDOO DFWLYDWLRQ HQHUJLHV DUH NQRZQ 8QOLNH SK\VLVRUSWLRQ FKHPLVRUSWLRQ LV VSHFLILF WR WKH JDVHV DQG VROLGV LQYROYHG MXVW OLNH FKHPLFDO

PAGE 46

UHDFWLRQV DUH &KHPLVRUSWLRQ LV DOVR WHPSHUDWXUH GHSHQGHQW RFFXUULQJ IDVWHU DV WHPSHUDWXUH LQFUHDVHV KRZHYHU WKH DPRXQW FKHPLVRUEHG DW HTXLOLEULXP GHFUHDVHV ZLWK WHPSHUDWXUH +D\ZDUG DQG 7UDSQHOO 6]HNHO\ HW DO f *DVf§VROLG UHDFWLRQV DUH LQKHUHQWO\ FRPSOH[ GXH WR WKH QXPEHU RI VLPXOWDQHRXV SURFHVVHV WKDW RFFXU WKURXJKRXW WKH FRXUVH RI UHDFWLRQ $Q\ HTXLOLEULXP FRQGLWLRQV UHDFKHG DUH RIWHQ G\QDPLF UDWKHU WKDQ VWDWLF LQ QDWXUH VLQFH WKH V\VWHP FRQWLQXHV WR FKDQJH ZLWK WLPH ,Q DGGLWLRQ WR PDVV WUDQVIHU DQG GLILXVLRQ SURFHVVHV WKHVH UHDFWLRQV RIWHQ LQFOXGH KHDW WUDQVIHU DQG VWUXFWXUDO FKDQJHV LQ WKH VROLG SKDVH ,GHDOO\ WR SURSHUO\ GHVFULEH D SDUWLFXODU JDVf§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f§VROLG UHDFWLRQ PD\ EH UHSUHVHQWHG E\ WKH IROORZLQJ UHDFWLRQ VFKHPH D$ JDVf E% VROLGf F& VROLGf G' JDVf f

PAGE 47

ZKHUH $ DQG % DUH JDVHRXV DQG VROLG UHDFWDQWV UHVSHFWLYHO\ & DQG DUH SURGXFWV DQG DEF DQG G DUH VWRLFKLRPHWULF FRHIILFLHQWV 7KLV RYHUDOO UHDFWLRQ PD\ DFWXDOO\ FRPSULVH VHYHUDO VHTXHQWLDO RU VLPXOWDQHRXV VWHSV RI ZKLFK RQH LV UDWHFRQWUROOLQJ 7KH WKUHH EDVLF VWHSV LQ WKH JHQHUDOL]HG QRQFDWDO\WLF JDVVROLG GLIIXVLRQUHDFWLRQ SURFHVV DUH f *DVSKDVH PDVV WUDQVIHU YLD GLIIXVLRQ RI WKH JDVHRXV UHDFWDQW IURP WKH EXON JDV VWUHDP WR WKH H[WHUQDO VXUIDFH RI WKH VROLG VRUEHQW SDUWLFOH f *DVVROLG LQWHUDFWLRQ RQ RU ZLWKLQ WKH VROLG VRUEHQW SDUWLFOH WKURXJK Df 'LIIXVLRQ RI WKH JDVHRXV UHDFWDQW LQWR WKH SRUHV RI WKH VROLG FRPSOH[ ZKLFK FRXOG FRQVLVW RI D FRPELQDWLRQ RI VROLG UHDFWDQWV DQG SURGXFWV Ef $GVRUSWLRQ RI WKH JDVHRXV UHDFWDQW RQ WKH VXUIDFH RI WKH VROLG FRPSOH[ Ff &KHPLFDO UHDFWLRQ DW WKH VXUIDFH Gf 'HVRUSWLRQ DQG GLIIXVLRQ RI JDVHRXV SURGXFWV LI DQ\ IURP WKH VXUIDFH DQG RXW RI WKH SRUHV RI WKH VROLG FRPSOH[ f *DVSKDVH PDVV WUDQVIHU RI DQ\ JDVHRXV SURGXFWV IURP WKH H[WHUQDO VROLG VXUIDFH LQWR WKH EXON JDV VWUHDP 7KH QXPEHU RI GLYHUVH SURFHVVHV DQG VWHSV LQYROYHG PDNHV WKH DQDO\VLV RI JDVf§VROLG UHDFWLRQV D SRWHQWLDOO\ XQZLHOG\ SUREOHP ,Q JHQHUDO HDFK VWHS SURYLGHV UHVLVWDQFH WR FKHPLFDO UHDFWLRQ DQG WKHVH UHVLVWDQFHV DUH DGGLWLYH :KHQ WKH SURFHVVHV RFFXU LQ VHULHV LW LV QHFHVVDU\ WR GHWHUPLQH ZKLFK VWHS SURYLGHV WKH PDMRU UHVLVWDQFH WR UHDFWLRQ DQG XOWLPDWHO\ FRQWUROV WKH RYHUDOO UHDFWLRQ UDWH 7KH UHVLVWDQFHV RI WKH GLIIHUHQW VWHSV FDQ YDU\ JUHDWO\ IURP HDFK RWKHU VR WKH VWHS ZLWK WKH KLJKHVW UHVLVWDQFH FDQ EH FRQVLGHUHG WR EH WKH UDWHFRQWUROOLQJ

PAGE 48

VWHS $ VLPSOLILHG H[SUHVVLRQ IRU WKH UDWHGHWHUPLQLQJ VWHS FDQ WKHQ EH XVHG WR GHVFULEH WKH RYHUDOO SURFHVV 7KUHH PDMRU SURFHVVHV DUH JHQHUDOO\ FRQVLGHUHG WR EH UDWHFRQWUROOLQJ LQ QRQFDWDO\WLF JDVVROLG UHDFWLRQV f 0DVV WUDQVIHU RI WKH JDVHRXV UHDFWDQW IURP WKH EXON JDV SKDVH WKURXJK WKH VWDJQDQW ILOP OD\HU VXUURXQGLQJ WKH SDUWLFOH WR WKH VXUIDFH RI WKH VROLG VRUEHQW 7KH UDWH RI WKLV VWHS LV PRVW GHSHQGHQW XSRQ WKH IOXLG G\QDPLFV RI WKH JDV IORZ DURXQG WKH VROLG SDUWLFOH f &KHPLFDO NLQHWLFV RI WKH UHDFWLRQ EHWZHHQ WKH JDVHRXV DQG VROLG UHDFWDQWV f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f 5HODWHG QRQOLQHDU VRUSWLRQ PRGHOV DUH QRW DSSOLFDEOH WR V\VWHPV ZLWK SRURXV VRUEHQWV ZKHUH LVRWKHUPV DUH IDYRUDEOH DQG FKHPLFDO UHDFWLRQ DQG SRUH GLIIXVLRQ UHVLVWDQFHV FDQ EH VLJQLILFDQW %L\DQL DQG

PAGE 49

*RRFKHH f ,Q JHQHUDO JDVVROLG UHDFWLRQV FDQ EH GHVFULEHG E\ DQ LGHDO /DQJPXLU LVRWKHUP RQO\ ZKHQ WKHUH LV H[FOXVLYH DQG FRPSOHWH PRQROD\HU VRUSWLRQ DOO DFWLYH VLWHV RQ WKH VRUEHQW DUH HTXLYDOHQW DQG WKHUH LV QR LQWHUDFWLRQ EHWZHHQ DGMDFHQW DGVRUEDWH PROHFXOHV &RPHV HW DO f 6ROLG VWUXFWXUDO FKDQJHV UHVXOWLQJ IURP WKH FKHPLFDO FKDQJH LQ WKH VROLG FDQ EH TXLWH FRPSOH[ EXW DUH JHQHUDOO\ FODVVLILHG DV EHLQJ HLWKHU VLQWHULQJ DQG UHVXOWDQW SRUH FORVXUHf VZHOOLQJ VRIWHQLQJ DQG FUDFNLQJ 6]HNHO\ HW DO f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f UHVLVWDQFH WKH RYHUDOO UDWH LV FRQWUROOHG E\ FKHPLFDO NLQHWLFV ZKLFK LV JHQHUDOO\ VORZ LQ WKLV FDVH 6KULQNLQJ 8QUHDFWHG&RUH 0RGHO (YLGHQFH IURP D ZLGH YDULHW\ RI VLWXDWLRQV LQGLFDWHV WKDW WKH SURJUHVVLYHFRQYHUVLRQ PRGHO GRHV QRW DFFXUDWHO\ DSSUR[LPDWH WKH EHKDYLRU RI UHDO SDUWLFOHV 2IWHQ WKH UHDFWLRQ SURGXFHV DQ LQHUW SURGXFW OD\HU RI FRQYHUWHG VROLG UHDFWDQW ZKLFK FDQ SURGXFH VLJQLILFDQW

PAGE 50

UHVLVWDQFH WR GLIIXVLRQ $ PDWKHPDWLFDOO\ VLPSOH PRGHO WKDW ZDV GHYHORSHG E\
PAGE 51

6WDJQDQW *DV )LOP 3URJUHVV RI 5HDFWLRQ 8QUHDFWHG 6RUEHQW 3DUWLFOH )LJXUH 6FKHPDWLF UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO

PAGE 52

*UDLQ 0RGHO 7R DFFRXQW IRU VRUEHQW SRURVLW\ 6]HNHO\ DQG (YDQV f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f HQWHUV WKH UHOHYDQW VKULQNLQJ XQUHDFWHGFRUH PRGHO HTXDWLRQV 7KLV VKDSH IDFWRU FDQ EH XVHG WR TXDOLWDWLYHO\ GLVFHUQ LQIRUPDWLRQ DERXW WKH VRUEHQW VXUIDFH SK\VLFDO VWUXFWXUH DV LW DIIHFWV VRUSWLRQ EHKDYLRU 7KLV PRGHO KDV EHHQ IXUWKHU PRGLILHG E\ +DUWPDQ DQG &RXJKOLQ f DPRQJ RWKHUV WR DFFRXQW IRU LQFUHDVHG GLIIXVLRQDO UHVLVWDQFH WR FKHPLFDO UHDFWLRQ LQ WKH SURGXFW OD\HU 7KLV HIIHFW LV JUHDWO\ FRPSRXQGHG ZKHQ WKH PRODU YROXPH RI WKH SURGXFW LV VLJQLILFDQWO\ ODUJHU WKDQ WKDW RI WKH VROLG UHDFWDQW FDXVLQJ FRYHUDJH RU EORFNDJH RI SRWHQWLDOO\ DYDLODEOH LQWHUJUDLQ SRUH VXUIDFH DUHD 7KLV PD\ FDXVH D GLVFUHSDQF\ EHWZHHQ WKH WKHRUHWLFDO VROLGVXUIDFH FRQYHUVLRQ YHUVXVWLPH EHKDYLRU DV SUHGLFWHG E\ WKH PRGHO DQG DFWXDO H[SHULPHQWDO UHVXOWV 5DPDFKDQGUDQ DQG 6PLWK f WRRN WKH PRGHO D VWHS IXUWKHU LQ PRGHOLQJ WKH VLQJOHSRUH EHKDYLRU LQ D SRURXV SHOOHW WR SUHGLFW WKH FRQYHUVLRQ YHUVXV WLPH UHODWLRQVKLS IRU JDVVROLG QRQFDWDO\WLF UHDFWLRQV 7KHLU PRGHO DFFRXQWV IRU WKH LQIOXHQFH RI SRUH GLIIXVLRQ GLIIXVLRQ WKURXJK WKH SURGXFW OD\HU DQG VXUIDFH UHDFWLRQ $ NH\ SDUDPHWHU XVHIXO LQ

PAGE 53

)LJXUH 6FKHPDWLF UHSUHVHQWDWLRQ RI WKH JUDLQ PRGHO R

PAGE 54

GHVFULELQJ WKH REVHUYHG GLVFUHSDQF\ EHWZHHQ WKHRUHWLFDO DQG REVHUYHG VROLG FRQYHUVLRQ LV WKH HIIHFWLYH GLILLVLYLW\ WKURXJK WKH SURGXFW OD\HU ,QFRPSOHWH FRQYHUVLRQ ZKLFK GHFUHDVHV ZLWK DQ LQFUHDVH LQ LQWUDSHOOHW GLIIXVLRQDO UHVLVWDQFH LV SUHGLFWHG XVLQJ WKH UDQGRP SRUH PRGHO GHYHORSHG E\ %KDWLD DQG 3HUOPXWWHU Df $ XQLTXH IHDWXUH RI WKLV PRGHO LV WKDW VROLG SRUHV DUH QRW DVVXPHG WR EH RI XQLIRUP VL]H VR VPDOOHU SRUHV DUH PRUH HDVLO\ EORFNHG ZKHQ D GHQVH RU ODUJHYROXPH SURGXFW LV IRUPHG RQ WKH VXUIDFH 'DP-RKDQVHQ HW DO f XVHG SK\VLFDO SDUDPHWHUV RI D FKDON VRUEHQW WR PRGLI\ WKH JUDLQ PRGHO GRZQ WR WKH PLFURJUDLQ OHYHO :KLOH WKLV SURYLGHV D SK\VLFDO GHVFULSWLRQ RI WKH UHDFWLRQ RFFXUULQJ DW WKH VXEJUDLQ OHYHO WKH UHDFWLRQ WDNLQJ SODFH VWLOO IROORZV WKH VKULQNLQJ XQUHDFWHGFRUH PHFKDQLVP $QDORJRXV *DV6ROLG 5HDFWLRQ 6\VWHP 5HVHDUFK 7KH PRVW VLPLODU JDVf§VROLG UHDFWLRQ V\VWHPV GLVFXVVHG LQ FXUUHQW OLWHUDWXUH GHVFULEH WKH UHDFWLRQV EHWZHHQ VXOIXU GLR[LGH 6f DQG OLPHFDOFLXP R[LGH &D2f FDOFLXP K\GUR[LGH &D2+ff RU OLPHVWRQH FDOFLXP FDUERQDWH &D&ff 5HDFWLRQV RI DFLGLF JDVHV ZLWK FDOFLQHG OLPHVWRQH KDYH EHHQ FORVHO\ VWXGLHG EHFDXVH RI WKHLU LQGXVWULDO LPSRUWDQFH 7KH KLJK SRURVLWLHV RI YDULRXV IRUPV RI &D2 QRPLQDOO\ RYHU SHUFHQWf PDNHV WKHP TXLWH VXLWDEOH IRU WHVWV RI WKH QXPHURXV IRUPV RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO $OWKRXJK WKHVH PRGHOV KDYH EHHQ XVHG TXLWH VXFFHVVIXOO\ WKH IXQGDPHQWDO UDWHFRQWUROOLQJ SURFHVVHV DUH VWLOO QRW ZHOO XQGHUVWRRG %RUJZDUGW HW DO f 'DWD RQ UHDFWLRQV EHWZHHQ RWKHU DFLG JDVDFLG DQK\GULGH JDVHV DQG RWKHU FDUERQDWH URFNV DUH DOVR DYDLODEOH ,Q WKHVH SURFHVVHV 6 FRQWDLQLQJ JDVHV DUH FRQWDFWHG ZLWK OLPH RU OLPHVWRQH HLWKHU DV D ZHW VOXUU\ RU GU\ VROLG 6 LV JHQHUDOO\ FROOHFWHG RQ WKH VROLG VXUIDFH IRU UHXVH RU GLVSRVDO 5HJHQHUDWLRQ PD\ SURGXFH

PAGE 55

UHXVHDEOH VXOIXULF DFLG RU LQ VRPH FDVHV HOHPHQWDO VXOIXU $ PDJQHVLXP R[LGH SURFHVV VLPLODU WR OLPH RU OLPHVWRQH VFUXEELQJ LV VRPHWLPHV XVHG ZKHUH 0J2 LV K\GUDWHG WR 0J2+f ZKLFK UHDFWV ZLWK 6 IRUPLQJ 0J60J6 VROLGV 7KHVH FDQ EH KHDWHG JHQHUDWLQJ 6 DQG 0J2 7KH 6 FDQ QRUPDOO\ EH UHFRYHUHG DV D VXOIXULF DFLG SURGXFW &RRSHU DQG $OOH\ f 62&D2 5HDFWLRQ %RUJZDUGW f LQYHVWLJDWHG WKH UHDFWLRQ RI 6 ZLWK FDOFLQHG OLPHVWRQHV DQG IRXQG D ILUVWRUGHU FKHPLFDO UHDFWLRQ WR EH WKH SUHGRPLQDQW SURFHVV ZLWK UHVLVWDQFH OLPLWLQJ 6 VRUSWLRQ E\ VPDOO SDUWLFOHV +H IRXQG WKDW WKH UHDFWLRQ UDWH GHFUHDVHG UDSLGO\ DV &D2 FRQYHUVLRQ LQFUHDVHG 6RPH RI WKH VROLGV FRQWDLQHG VLJQLILFDQW TXDQWLWLHV RI 0J2 ZKLFK GLG QRW UHDGLO\ UHDFW ZLWK 6 DW WKH HOHYDWHG WHPSHUDWXUHV HYDOXDWHG :HQ DQG ,VKLGD f IRXQG WKURXJK DQ DSSOLFDWLRQ RI WKH JUDLQ PRGHO WKDW WKH RYHUDOO UHDFWLRQ UDWH EHWZHHQ 6 DQG &D2 ZDV KLJKO\ WHPSHUDWXUH GHSHQGHQW SDUWLFXODUO\ DW ORZHU WHPSHUDWXUHV ZKHUH LW ZDV FRQWUROOHG E\ FKHPLFDO NLQHWLFV 7KH\ UHSRUWHG WKDW YDULRXV NLQHWLF VWXGLHV KDYH LQGLFDWHG WKDW UHDFWLRQ UDWHV PD\ YDU\ FRQVLGHUDEO\ PRUH WKDQ DQ RUGHU RI PDJQLWXGHf GHSHQGLQJ XSRQ WKH W\SH RI OLPHVWRQHV XVHG 7KLV YDULDELOLW\ LV D FRPPRQ SUREOHP LQKHUHQW LQ WKH VWXG\ RI KHWHURJHQHRXV VRUEHQWV 7KH SUHYLRXVO\ PHQWLRQHG VWXG\ E\ +DUWPDQ DQG &RXJKOLQ f ZKLFK LPSURYHG WKH DSSOLFDELOLW\ RI WKH JUDLQ PRGHO DOVR GHVFULEHG WKH 6&D2 UHDFWLRQ 7KH VWURQJ UHODWLRQVKLS EHWZHHQ VXUIDFH DUHD DQG WKH UHDFWLYLW\ RI &D2 WRZDUG 6 ZDV DQDO\]HG YLD WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO E\ %RUJZDUGW DQG %UXFH f 7KH VXUIDFH DUHD LV

PAGE 56

DSSDUHQWO\ YHU\ VHQVLWLYH WR WKH SUHVHQFH RI ZDWHU YDSRU ZKLFK UHGXFHV SRURVLW\ %RUJZDUGW f $ JUDSKLF UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO ZDV GHSLFWHG LQ D VHULHV RI HQHUJ\ GLVSHUVLYH ;UD\ DQDO\VLV PLFURJUDSKV RI FDOFLQHG OLPHVWRQH FKDON DQG 6 'DP-RKDQVHQ DQG 2VWHUJDDUG Df 7KH DVVRFLDWHG GDWD ZHUH ILWWHG ZLWK D JUDLQ PRGHO RI WKH VKULQNLQJ XQUHDFWHGFRUH PHFKDQLVP 'DP-RKDQVHQ Ef %MHUOH HW DO f UHSRUWHG XVHIXO H[SHULPHQWDO WHFKQLTXHV IRU GHVFULELQJ WKLV UHDFWLRQ DV ZHOO 7KH\ ZHUH DEOH WR XVH WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO WR GHVFULEH WKH YDULRXV VWDJHV RI UHDFWLRQ DV FRQYHUVLRQ SURJUHVVHG 6RWLUFKRV DQG =DUNDQLWLV f IRXQG WKDW WKH VXOIDWLRQ UDWHV RI 0J2 ZHUH FRPSDUDEOH WR WKRVH IRU &D2 $ VLPSOH JUDLQ PRGHO ZDV DOVR XVHG WR LQWHUSUHW GDWD GHVFULELQJ D IXUQDFH VRUEHQW VOXUU\ LQMHFWLRQ SURFHVV LQ ZKLFK D OLPHVWRQH VOXUU\ ZDV FDOFLQHG IRU XVH LQ 6 UHPRYDO 'DPOH f 6&DOFLXP +\GUR[LGH 5HDFWLRQ )OXHJDV GHVXOIXUL]DWLRQ WKURXJK VSUD\ GU\LQJ RI D &D2+f VOXUU\ KDV UHFHQWO\ EHFRPH DQ LPSRUWDQW DOWHUQDWLYH WR WKH WUDGLWLRQDO ZHW OLPH RU OLPHVWRQH VFUXEELQJ WHFKQLTXHV ,Q VSUD\ GU\LQJ 6FRQWDLQLQJ IOXH JDVHV DUH FRQWDFWHG LQ D GU\HU ZLWK ILQHO\ DWRPL]HG OLPH VOXUU\ WKDW DEVRUEV WKH 6 $V ZDWHU HYDSRUDWHV IURP WKH VOXUU\ GURSOHWV UHDFWHG VROLGV DQG XQUHDFWHG &D2+f UHPDLQ DQG DUH FROOHFWHG RQ EDJ ILOWHUV 7KLV DFFXPXODWHG UHVLGXDO &D2+f VHUYHV WR UHPRYH FRQVLGHUDEOH TXDQWLWLHV RI XQUHDFWHG 6 IURP WKH LQVLGH RI WKH EDJV DV ZHOO DV LQ WKH GXFWV 5XL]$OVRS DQG 5RFKHOOH f XVHG D EHQFKVFDOH IL[HGEHG UHDFWRU WR VWXG\ WKLV UHDFWLRQ 7KH VKULQNLQJ XQUHDFWHGFRUH PRGHO ZDV DSSOLHG WR WKHLU UHVXOWV $PRQJ WKH PDQ\ IDFWRUV HYDOXDWHG UHODWLYH KXPLGLW\ RI WKH JDV ZDV GHWHUPLQHG WR

PAGE 57

EH WKH PRVW LPSRUWDQW YDULDEOH DIIHFWLQJ VRUSWLRQ EHKDYLRU &KX DQG 5RFKHOOH f H[DPLQHG WKH VLPXOWDQHRXV UHPRYDO RI 6 DQG 12[ DV 12f E\ &D2+f DQG LQ VRPH FDVHV DGGLWLYHV LQFOXGLQJ IO\ DVK &D6 DQG 1D2+ 7KHVH DGGLWLYHV LPSURYHG 12[ UHPRYDO $QRWKHU EHQFKVFDOH HYDOXDWLRQ RI WKH UHPRYDO RI 6 DQG 12 E\ &D2+f ZDV FRQGXFWHG E\ -R]HZLF] HW DO f 12 ZDV QRW YHU\ UHDFWLYH WRZDUG &D2+f UHODWLYH WR 6 EXW DGGLQJ 0J2+f GLG LPSURYH 12 UHPRYDO VRPHZKDW 62f§&DOFLXP &DUERQDWH 5HDFWLRQ 3LJIRUG DQG 6OLJHU f QRWHG WKH HIIHFWV RI LQFUHDVHG SURGXFW OD\HU GLIOLLVLRQDO UHVLVWDQFH LQ WKH UHDFWLRQ EHWZHHQ 6 DQG &D& 7KH RYHUDOO UHDFWLRQ UDWH ZDV JRYHUQHG E\ ERWK GLIIXVLRQ RI 6 WKURXJK D OD\HU RI VROLG UHDFWLRQ SURGXFW SURJUHVVLYHO\ IRUPHG RQ WKH DFWLYH VROLG VXUIDFH DQG GLIIXVLRQ RI 6 WKURXJK SRUHV LQ WKH VROLG 7KH GLUHFW VXOIDWLRQ RI &D& ZDV VWXGLHG E\ +DMDOLJRO HW DO f YLD WKHUPRJUDYLPHWULF DQDO\VHV DQG D EHQFK VFDOH IOXLGL]HG EHG VHWXS 7KH SURGXFW OD\HU GLIIXVLRQ FRQWUROOHG VKULQNLQJ XQUHDFWHGFRUH PRGHO ZDV XVHG WR ILWWHG WR WKH H[SHULPHQWDO GDWD XVLQJ D FRUUHFWLRQ IDFWRU IRU SURGXFW OD\HU YROXPH DOWKRXJK FKHPLFDO NLQHWLFV ZHUH VWLOO LPSRUWDQW WR GHWHUPLQLQJ WKH RYHUDOO UHDFWLRQ UDWH 6QRZ HW DO f DOVR HYDOXDWHG WKH GLUHFW VXOIDWLRQ RI OLPHVWRQH XQGHU FRQGLWLRQV WKDW GLG QRW GHFRPSRVH &D& WR &D2 7KHLU UHVXOWV ZHUH VLPLODU WR WKRVH RI +DMDOLJRO HW DO f $V SUHYLRXVO\ PHQWLRQHG D YDULDEOH GLIIXVLYLW\ VKULQNLQJ XQUHDFWHGFRUH PRGHO ZDV VXFFHVVIXOO\ DSSOLHG WR WKH GLUHFW VXOIDWLRQ RI &D& E\ .ULVKQDQ DQG 6RWLUFKRV Ef (IIHFWLYH JDV GLIIXVLYLW\ WKURXJK WKH VROLG SURGXFW OD\HU ZDV IRXQG WR EH D VWURQJ IXQFWLRQ RI JDV WHPSHUDWXUH DQG 6 FRQFHQWUDWLRQ LQ WKH EXON JDVSKDVH

PAGE 58

2WKHU $FLGDDV&DOFLXPEDVHGVROLG 5HDFWLRQV 2WKHU VRUSWLRQ VWXGLHV HPSOR\LQJ WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO DQG LWV GHULYDWLRQV KDYH EHHQ FRQGXFWHG WR VWXG\ WKH FRQWURO RI RWKHU DFLG JDVHV 7KH UHDFWLRQ RI K\GURJHQ FKORULGH +&f ZLWK &D2 LV RI LQWHUHVW IRU WKH FRQWURO RI DFLG YDSRUV HPLWWHG IURP PXQLFLSDO DQG KD]DUGRXV ZDVWH LQFLQHUDWLRQ RSHUDWLRQV *XOOHWW HW DO f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f 3DNUDVL f SHUIRUPHG NLQHWLF VWXGLHV RI +& UHPRYDO E\ K\GUDWHG OLPH SRZGHU LQ D EHQFKVFDOH IL[HGEHG UHDFWRU +LV VWXGLHV LQFOXGHG D WKRURXJK H[DPLQDWLRQ RI WKH FKHPLFDO NLQHWLFV DQG SRVVLEOH PHFKDQLVP DVVRFLDWHG ZLWK WKH +&f§&D2 UHDFWLRQ 5HODWLYH KXPLGLW\ SOD\HG D NH\ UROH LQ GHWHUPLQLQJ +& UHPRYDO UDWH DQG WKH H[WHQW RI VROLG VRUEHQW FRQYHUVLRQ 2WKHU UHOHYDQW YDULDEOHV VXFK DV JDV FRQFHQWUDWLRQ DQG WHPSHUDWXUH ZHUH DOVR H[DPLQHG ,QFRPSOHWH FRQYHUVLRQ FRPSDUHG WR WKDW WKHRUHWLFDOO\ SUHGLFWHG VWRLFKLRPHWULFDOO\ RU E\ WKH VKULQNLQJ XQUHDFWHGFRUH PRGHOf ZDV WKH UHVXOW RI LQFUHDVHG GLIIXVLRQDO UHVLVWDQFH LQ WKH SURGXFW OD\HU %RWK WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO DQG DQ HPSLULFDO PRGHO ZHUH XVHG WR LQWHUSUHW REVHUYHG FRQYHUVLRQ YHUVXV WLPH GDWD

PAGE 59

7KH SURFHVV RI FRDO JDVLILFDWLRQ RIWHQ SURGXFHV VLJQLILFDQW TXDQWLWLHV RI K\GURJHQ VXOILGH +6f /LPHVWRQHV HLWKHU FDOFLQHG RU XQFDOFLQHGf KDYH EHHQ LQYHVWLJDWHG DV VRUEHQWV IRU WKH UHPRYDO RI WKHVH JDVHV WKURXJK WKH GLUHFW VXOILGDWLRQ RI WKH VROLGV YLD WKHUPRJUDYLPHWULF DQDO\VHV .ULVKQDQ DQG 6RWLUFKRV f 7KH VKULQNLQJ XQUHDFWHGFRUH PRGHO ZDV HPSOR\HG ERWK TXDQWLWDWLYHO\ DQG TXDOLWDWLYHO\ WR GHULYH NLQHWLF SDUDPHWHUV DQG D WR SURYLGH D GHVFULSWLRQ RI WKH UDWHOLPLWLQJ PDVV WUDQVIHU SURFHVVHV ,Q WKLV FDVH VLQFH WKH SURGXFW FDOFLXP VXOILGH &D6f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f§VROLG UHDFWLRQ WR FRQWURO JDVHRXV SROOXWDQWV

PAGE 60

5HVHDUFK -XVWLILFDWLRQ 1R RWKHU UHVHDUFK KDV EHHQ UHSRUWHG WKDW ZRXOG GHOLQHDWH WKH LQWULQVLF FKHPLFDO NLQHWLF SDUDPHWHUV DVVRFLDWHG ZLWK WKH UHDFWLRQV RI 12[ ZLWK 0J2YHUPLFXOLWH VRUEHQWV /LNHZLVH WKH SUREDEOH UHDFWLRQ PHFKDQLVP KDV QRW \HW EHHQ GHWHUPLQHG LQFOXGLQJ WKH LGHQWLILFDWLRQ RI UDWHFRQWUROOLQJ VWHSVf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f DQG ZKHWKHU WKHVH DUH FKHPLFDO UHDFWLRQ GLIIXVLRQ RU PDVV WUDQVIHU OLPLWHG

PAGE 61

(YDOXDWH VDWXUDWLRQ FKDUDFWHULVWLFV OLIHWLPHf RI WKH PDWHULDO 'HYHORS D PRGHO WR SUHGLFW VRUEHQW SHUIRUPDQFH ZLWK WLPH (YDOXDWH WKH GHVRUSWLRQ DQG UHJHQHUDWLRQ HIILFLHQFLHV RI 12[ RQ 0J2f§YHUPLFXOLWH (YDOXDWH WKH SUHVVXUHGURS FKDUDFWHULVWLFV RI WKH PDWHULDO RYHU WLPH DV JDV FROOHFWLRQ SURJUHVVHV

PAGE 62

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f $ YDULHW\ RI WHVW FRQGLWLRQV ZHUH XVHG WR GHWHUPLQH LPSRUWDQW RSHUDWLQJ SDUDPHWHUV DQG OLPLWDWLRQV %\ FRPELQLQJ JDVHRXV FRPSRQHQWV WKHLU LQWHUDFWLRQV LQ WKH SUHVHQFH RI WKH VRUEHQW PDWHULDO FRXOG EH HYDOXDWHG ,QWULQVLF DQG RYHUDOO NLQHWLF SDUDPHWHUV ZHUH GHWHUPLQHG $ FRPSOHWH GHVFULSWLRQ RI WKH NLQHWLFV RI JDVf§VROLG UHDFWLRQV LV YHU\ FRPSOLFDWHG HVSHFLDOO\ IRU V\VWHPV LQYROYLQJ VROLG SURGXFWV IRU ZKLFK WKH SURFHVVHV DW ERWK WKH JDVVROLG LQWHUIDFH DQG WKH UHDFWLRQ LQWHUIDFH EHWZHHQ WKH UHDFWDQW DQG SURGXFW VROLGV PXVW EH FRQVLGHUHG $OWKRXJK WKH UDWH H[SUHVVLRQV FDQ EH TXLWH FRPSOH[ WKH UHDFWLRQV EHWZHHQ DGVRUEDWH DQG DGVRUEHQW FDQ RIWHQ EH GHVFULEHG E\ ILUVWRUGHU NLQHWLFV 6]HNHO\ HW DO f 'DWD ZHUH PDWKHPDWLFDOO\ LQWHUSUHWHG XVLQJ WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO

PAGE 63

DQG LWV YDULDWLRQV DV ZHOO DV HPSLULFDOO\ 7KHVH PRGHOV SURYLGHG D FKHPLFRSK\VLFDO EDVLV IRU H[SODLQLQJ VRUEHQW SHUIRUPDQFH DQG SUREDEOH UHDFWLRQ PHFKDQLVPV ([SHULPHQWDO 9DULDEOHV 0J2f§9HUPLFXOLWH 5HDFWLYH 6RUEHQW 0DWHULDO 7KLV PDWHULDO DSSUR[LPDWHO\ b 0J2 WR b YHUPLFXOLWH VXEVWUDWH E\ ZHLJKWf ZDV SUHSDUHG DFFRUGLQJ WR D SURFHVV SDWHQWHG E\ WKH LQYHQWRU 86 3DWHQW 6RUEHQW 7HFKQRORJLHV &RUSRUDWLRQ 6RUEWHFKf 7ZLQVEXUJ 2KLR f &RPPHUFLDOJUDGH PDJQHVLXP R[LGH ZDV XVHG b YRODWLOHIUHH 0J2f ZLWK D PDQXIDFWXUHUUHSRUWHG VXUIDFH DUHD RI PJ (ODVWRPDJ p $NURFKHP &RUSRUDWLRQ $NURQ 2KLRf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

PAGE 64

0DJQHVLXP R[LGH SRZGHU ZDV SUHSDUHG IRU DGGLWLRQ WR WKH PL[WXUH E\ VLHYLQJ ZLWK D PHVK 7\OHU 6WDQGDUG VLHYH SP RSHQLQJf )RUW\ILYH JUDPV 0J2 SHU JUDPV GU\ YHUPLFXOLWH ZDV DGGHG WR WKH ZHW YHUPLFXOLWH ZKLOH WKH PL[WXUH ZDV PDQXDOO\ VWLUUHG XVLQJ D VWLUULQJ WRRO 7KLV UDWLR ZDV WKH VDPH DV WKDW UHSRUWHG E\ 6RUEWHFK DQG LV OHVV WKDQ WKH PD[LPXP PDVV UDWLR GHVFULEHG LQ WKH SDWHQW 7KH VDPH UDWLR ZDV XVHG VR WKDW WKH WZR VRUEHQWV FRXOG EH FRPSDUHG 6RUEWHFK LV DEOH WR SUHSDUH ODUJH EDWFKHV RI PDWHULDO XVLQJ PHFKDQLFDO VWLUULQJ SHUKDSV DOORZLQJ WKHP WR XVH OHVV ZDWHU ZKLOH PRUH HYHQO\ FRDWLQJ WKH VRUEHQW ZLWK WKH 0J2+f VOXUU\ $IWHU WKRURXJK PL[LQJ WKH VRUEHQW ZDV DOORZHG WR DLU GU\ RYHUQLJKW XQWLO WKH LQGLYLGXDO SDUWLFOHV ZHUH DJDLQ IUHHIORZLQJ 7KH SDQV RI VRUEHQW ZHUH WKHQ SODFHG LQ D SUHKHDWHG PXIIOH IXUQDFH VHW DW & DQG FRQGLWLRQHG IRU PLQXWHV ,W LV EHOLHYHG WKDW WKLV SURFHVV LQFUHDVHV VRUEHQW VXUIDFH DUHD E\ H[SDQGLQJ WKH SDUWLFOHV DV D UHVXOW RI GULYLQJ RXW LQWHULRU ZDWHU YDSRU YLD KHDWLQJ 7KLV SURFHVV DOVR GHK\GUDWHV WKH 0J2+f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

PAGE 65

YDULDWLRQ DV ZHOO DV IRU FRPSDULVRQ ZLWK GDWD IURP WKH 6RUEWHFK VDPSOHV WR HQVXUH WKDW WKH VRUEHQW ZDV SURSHUO\ SUHSDUHG 7KH VRUEHQW PDWHULDO ZDV LQFRUSRUDWHG LQWR D SDFNHGEHG DUUDQJHPHQW LQ D QRQUHDFWLYH VWDLQOHVV VWHHO WXEXODU UHDFWRU LQ D FRQWUROOHG WHPSHUDWXUH WXEH IXUQDFHf HQYLURQPHQW ZKLFK ZLOO EH GHVFULEHG ODWHU $ VLQJOH EDWFK RI PDWHULDO ZDV XVHG IRU DOO H[SHULPHQWV /LPLWHG VXUIDFH DUHDFRPSRVLWLRQ DQDO\VHV RI WKH PDWHULDO ZHUH SHUIRUPHG WR HVWDEOLVK EDVHOLQH YDOXHV &RPSDUDWLYH H[SHULPHQWV ZHUH FRQGXFWHG XVLQJ VRUEHQW VDPSOHV SURYLGHG E\ 6RUEWHFK 7HPSHUDWXUH 5HDFWLRQ WHPSHUDWXUHV ZHUH FRQWUROOHG XVLQJ D WXEH IXUQDFH 0RGHO 7KHUPRO\QH 'XEXTXH ,RZDf WR FRQWDLQ WKH UHDFWLRQ YHVVHO 6RUSWLRQ VWXG\ WHPSHUDWXUHV RI DQG ZHUH XVHG 7KHVH WHPSHUDWXUHV ZHUH FKRVHQ EHFDXVH WKH\ UHSUHVHQW D UHDVRQDEOH UDQJH H[SHFWHG IRU DXJPHQWHGFRROHG H[KDXVW JDVHV *DVHV ZHUH SUHKHDWHG EHIRUH HQWHULQJ WKH IXUQDFH WR HQVXUH WKDW WKH\ ZHUH DW RU DERYH UHDFWLRQ WHPSHUDWXUH &f ZKHQ WKH\ HQWHUHG WKH EHG VR WKH\ ZRXOG QRW QHHG DGGLWLRQDO KHDWLQJ 3UHVVXUH $OO UXQV ZHUH FRQGXFWHG DW DWPRVSKHULF SUHVVXUH 7KH EHG RXWOHW JDVHV ZHUH H[KDXVWHG WR WKH DWPRVSKHUH WKURXJK D ODERUDWRU\ YHQWLODWLRQ V\VWHP 3UHVVXUH GURS YHUVXV H[SRVXUH WLPH ZDV HYDOXDWHG WKURXJK WKH EHG XVLQJ D 8WXEH PDQRPHWHU FRQQHFWHG WR SRUWV DW WKH EHG LQOHW DQG RXWOHW 3UHVVXUH GURS YHUVXV JDV YHORFLW\ ZDV HYDOXDWHG YLD DQ HPSLULFDO FRUUHODWLRQ

PAGE 66

*DV &RPSRVLWLRQ 7R TXDQWLI\ HPLVVLRQV IURP MHW HQJLQHV WKH 86$) KDV VSRQVRUHG QXPHURXV VWXGLHV DQG KDV SXEOLVKHG MHW HQJLQH HPLVVLRQ YDOXHV IRU PDQ\ HQJLQHV LQ WKH 86$) LQYHQWRU\ 7KHVH GRFXPHQWV GHVFULEH 12[ DQG RWKHU FRPEXVWLRQ E\SURGXFW HPLVVLRQV FRUUHODWHG ZLWK SRZHU VHWWLQJ RU RSHUDWLRQDO SURFHGXUHV 6SLFHU HW DO f 7KHVH VWXGLHV DUH YHU\ XVHIXO IRU GHWHUPLQLQJ DSSURSULDWH H[KDXVW JDV FRQFHQWUDWLRQV WR VLPXODWH MHW HQJLQH H[KDXVW JDV FRPSRVLWLRQ LQ WKH ODERUDWRU\ 0DQ\ ODEVFDOH VWXGLHV KDYH XVHG WKHVH GDWD LQ WKHLU H[SHULPHQWDO GHVLJQ *DV FRPSRVLWLRQ ZDV NHSW FRQVWDQW IRU HDFK UXQ 0DMRU FRPSRQHQWV RI -(7& H[KDXVW ZHUH HYDOXDWHG LQGLYLGXDOO\ PL[HG LQ 1 RU DLUf 12 SSPf 1 SSPf DQG bf *DV FRQFHQWUDWLRQ UDQJHV UHODWH WR UHSUHVHQWDWLYH YDOXHV DW IRXU FRPPRQ HQJLQH VHWWLQJV QRPLQDOO\ LGOH b b DQG b SRZHUf IRU WKH 86$) ), WXUELQH HQJLQH $FWXDO HQJLQH HPLVVLRQV GDWD DUH VKRZQ LQ 7DEOH :DWHU YDSRU b E\ YROXPH DW WHPSHUDWXUHf ZDV HYDOXDWHG LQ FRKRUW ZLWK RWKHU JDVHRXV FRPSRQHQWV :DWHU YDSRU PD\ K\GUDWH WKH 0J2 IRUPLQJ D UHDFWLYH K\GUR[LGH ,W LV VHHQ IURP WKLV WDEOH WKDW MHW HQJLQH H[KDXVW FRQWDLQV VLJQLILFDQW TXDQWLWLHV RI XQEXPHG K\GURFDUERQV FDUERQ PRQR[LGH &2f DQG FDUERQ GLR[LGH &2A 3DUWLFXODWH PDWWHU LQ WKH IRUP RI VRRW RU FRQGHQVDWLRQ DHURVROV ZLOO DOVR PRVW OLNHO\ EH SUHVHQW LQ UHDO H[KDXVW $OO RI WKHVH H[KDXVW FRPSRQHQWV KDYH WKH SRWHQWLDO WR DIIHFW WKH 12[ VRUSWLRQ SHUIRUPDQFH RI WKH VRUEHQW )RU WKH SXUSRVHV RI WKLV VWXG\ WKHVH YDULDEOHV ZHUH QRW LQFOXGHG WR DOORZ IRU WKH HYDOXDWLRQ RI WKH EDVLF 12[VRUEHQW UHDFWLRQVf

PAGE 67

7DEOH 86$) ), 7XUELQH (QJLQH (PLVVLRQV 'DWD DV UHSRUWHGf 3RZHU 6HWWLQJ 7+& SSP&f 12[ SSPYf 12 SSPYf &2 SSPYf & bf ,GOH 3HUFHQW 3HUFHQW ,QWHUPHGLDWH +LJK 0DFKf b $IWHUEXUQHU $XJPHQWHG 6RXUFH 6SLFHU HW DO SDJH *DV )ORZ 5DWHI Vnf *DV IORZ UDWHVf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

PAGE 68

OD\HU VXUURXQGLQJ WKH VRUEHQW SDUWLFOH 3UDFWLFDO FRQVWUDLQWV RQ SUHVVXUH GURS ZLOO OLNHO\ OLPLW JDV YHORFLW\ LQ WKH DSSOLFDWLRQ RI WKLV FRQWURO WHFKQRORJ\ WR MHW HQJLQH WHVW FHOOV &RPSDUDWLYH H[SHULPHQWV DQG PDWKHPDWLFDO HYDOXDWLRQV ZHUH SHUIRUPHG WR HQVXUH WKDW WKLV VRXUFH RI HUURU ZDV HIIHFWLYHO\ PLQLPL]HG 5HVXOWV ZLOO EH SUHVHQWHG LQ D VXEVHTXHQW VHFWLRQ *DV 5HDFWLRQ6RUSWLRQ'HVRUSWLRQ *DV UHDFWLRQVRUSWLRQ ZDV PHDVXUHG E\ WKH FKDQJH LQ FRQFHQWUDWLRQ RI WKH WHVW JDVHVf ZLWK EHG UHVLGHQFH DQG H[SRVXUH WLPHV *DVHV ZHUH VDPSOHG DW WKH LQOHW DQG RXWOHW RI WKH VRUEHQW EHG DQG DOVR IURP WKUHH SRLQWV ZLWKLQ WKH EHG RI QRPLQDO P OHQJWKf %RWK VKRUWWHUP WR G\QDPLF HTXLOLEULXPf DQG ORQJWHUP UHVXOWV WDNLQJ WKH EHG WR VDWXUDWLRQ D IXQFWLRQ RI EHG H[SRVXUH WLPHf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

PAGE 69

RI VROLG VRUEHQW UHDFWDQW RU ILQH SDUWLFOHV GLVSHUVHG ZLWKLQ DQ LQHUW PDWHULDO EHG VXFK DV TXDUW] VLOLFD VDQG RU JODVV EHDGV :KLOH WKHVH W\SHV RI VWXGLHV DUH YDOXDEOH IRU FROOHFWLQJ GDWD RQ IXQGDPHQWDO UHDFWLRQV RFFXUULQJ EHWZHHQ WKH SKDVHV LQ WKLV VWXG\ E\ QHFHVVLW\ WKH XVH RI WKH SUHSDUHG VRUEHQW ZDV D IXQGDPHQWDO OLPLWDWLRQ ([SHULPHQWDO UHVXOWV PD\ RU PD\ QRW EH WUDQVIHUDEOH WR WKH EDVLF UHDFWLRQV EHWZHHQ 12[ JDVHV DQG 0J2 RU 0J2+f VROLGV ,W LV SRVVLEOH WKDW WKH SURFHVV RI PDNLQJ WKH VRUEHQW SURPRWHV VRUSWLRQ WKURXJK WKH HQKDQFHPHQW RI DYDLODEOH VXUIDFH DUHD ,W LV DOVR SRVVLEOH KRZHYHU WKDW WKH LQHIILFLHQW WUDQVIHU RI 0J2+f VOXUU\ WR WKH YHUPLFXOLWH VXEVWUDWH EHIRUH FDOFLQDWLRQ PD\ FUHDWH GHQVH DJJORPHUDWHV RU RWKHUZLVH PDNH 0J2 XQDYDLODEOH IRU UHDFWLRQ ZLWK 12[ 7KH HIIHFWV RQ XOWLPDWH VRUEHQW FRQYHUVLRQ DQG EHG XWLOL]DWLRQ ZLOO EH GLVFXVVHG ODWHU 2IWHQ SDUWLFXODUO\ LQ FKHPLFDO RU PDWHULDOV HQJLQHHULQJ VWXGLHV RI JDVf§VROLG UHDFWLRQV WKH HPSKDVLV LV RQ WKH VROLG VRUEHQW VXUIDFH DQG LWV FRQYHUVLRQ DV UHDFWLRQ SURJUHVVHV ([SHULPHQWV DUH GHVLJQHG IRU WKH SXUSRVH RI RSWLPL]LQJ WKH VROLG VRUEHQW LWVHOI 6ROLG FRQYHUVLRQ LV JHQHUDOO\ PHDVXUHG GLUHFWO\ WKURXJK FKHPLFDO DQDO\VHV LQFOXGLQJ VSHFWURVFRSLF DQG FKURPDWRJUDSKLF PHWKRGV 3K\VLFDO SURSHUWLHV RI WKH VRUEHQW FDQ EH HYDOXDWHG XVLQJ ; UD\ GLIIUDFWLRQ RU HOHFWURQ PLFURVFRS\ IRU H[DPSOH :KLOH WKHVH SUHYLRXVO\ GLVFXVVHG VWXGLHV DUH LQYDOXDEOH LQ HQYLURQPHQWDO HQJLQHHULQJ DQG SDUWLFXODUO\ LQ DLU SROOXWLRQ FRQWURO WKH IRFXV LV RIWHQ RQ WKH JDVSKDVH FRQFHQWUDWLRQ DQG WKH HIIHFWLYH UHPRYDO RI WKH SROOXWDQW RI LQWHUHVW 7KLV DSSOLHG HQJLQHHULQJ DSSURDFK LV WKH UHVXOW RI WKH XQDYRLGDEOH IDFW WKDW HQYLURQPHQWDO UHJXODWLRQV PDQGDWH WKH UHGXFWLRQ RI HPLVVLRQV WR VSHFLILHG OLPLWV DQG FRPSOLDQFH ZLWK UHJXODWLRQV LV GHWHUPLQHG WKURXJK DLU VDPSOLQJ ,W LV VWLOO FUXFLDO KRZHYHU WKDW JDVSKDVH FRQFHQWUDWLRQV EH FRUUHODWHG ZLWK VROLG FRQYHUVLRQ WKURXJK DQ DSSURSULDWH

PAGE 70

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f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

PAGE 71

([SHULPHQWDO $UUDQJHPHQW 7KH H[SHULPHQWDO DUUDQJHPHQW IRU VRUSWLRQ DQG GHVRUSWLRQUHJHQHUDWLRQ HPSOR\HG LQ WKLV VWXG\ LV VKRZQ LQ )LJXUH 1LWURJHQ R[LGHV PL[HG LQ 1f ZHUH VXSSOLHG IURP VHSDUDWH FHUWLILHG VWDQGDUG F\OLQGHUV %L7HF 6RXWKHDVW ,QF 7DPSD )ORULGDf WR HQVXUH UHOLDELOLW\ RI FRQFHQWUDWLRQV DQG WR PLQLPL]H FRQWDPLQDQW FRQFHQWUDWLRQV ,QLWLDOO\ 1 PL[WXUHV ZHUH JHQHUDWHG XVLQJ D 12[ JHQHUDWRU 0RGHO 7KHUPR (QYLURQPHQWDO ,QVWUXPHQWV )UDQNOLQ 0DVVDFKXVHWWVf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

PAGE 72

7R 9HQW )LJXUH ([SHULPHQWDO DUUDQJHPHQW IRU SDFNHGEHG VWXGLHV 92

PAGE 73

LV D SUREOHP ZLWK DWWHPSWLQJ WR XVH SXUH 1 RU 12f LQ JDVVROLG VRUSWLRQ VWXGLHV HPSOR\LQJ F\OLQGHU JDVHV 7KHVH JDVHV ZHUH GLOXWHG ZLWK GU\ DLU RU SUHSXULILHG 1 WR WKH DSSURSULDWH FRQFHQWUDWLRQVf &\OLQGHU JDV IORZ UDWHV ZHUH PHDVXUHG XVLQJ SUHFLVLRQ URWDPHWHUV 2PHJD (QJLQHHULQJ ,QF 6WDPIRUG &RQQHFWLFXWf ZLWK RQO\ JODVV DQGRU VWDLQOHVV VWHHO ZHWWHG SDUWV :KLOH WKH PDQXIDFWXUHU SURYLGHG FDOLEUDWLRQ GDWD DOO URWDPHWHUV ERWK KLJK DQG ORZIORZUDWH WXEHVf ZHUH LQGHSHQGHQWO\ FDOLEUDWHG LQ WKH ODERUDWRU\ XVLQJ D SULPDU\ DLUIORZ VWDQGDUG *LOLEUDWRU *LOLDQ ,QVWUXPHQW &RUSRUDWLRQ :D\QH 1HZ -HUVH\f 7KHVH FDOLEUDWLRQV ZHUH H[WUHPHO\ LPSRUWDQW IRU DFFXUDWH JDV IORZ UDWH PHDVXUHPHQWV SDUWLFXODUO\ ZKHQ XVLQJ ORZIORZUDWH WXEHV ZKLFK KDYH QRQOLQHDU FDOLEUDWLRQ FXUYHV *RRG DJUHHPHQW ZDV UHDFKHG EHWZHHQ WKH FDOLEUDWLRQ FXUYHV DV FDQ EH VHHQ LQ WKH H[DPSOH FRPSDUDWLYH FDOLEUDWLRQ FXUYHV )LJXUHV DQG /DERUDWRU\ FDOLEUDWLRQ FXUYHV ZHUH XVHG WR PHDVXUH JDV IORZ UDWHV LQ WKLV VWXG\ $OO V\VWHP FRPSRQHQWV ZHUH PDGH RI HLWKHU VWDLQOHVV VWHHO RU 3\UH[ r1 JODVV WR UHGXFH WKH SRWHQWLDO IRU 12[ UHPRYDO E\ V\VWHP UHDFWLYLW\ ,QHUW JDV WUDQVIHU OLQHV ZHUH PDGH IURP 7HIORQ r1 WXELQJ RI PLQLPDO OHQJWKV WR PLQLPL]H WUDQVIHU ORVVHV 0RVW FRQQHFWLRQV ZHUH WKURXJK VWDLQOHVV VWHHO FRPSUHVVLRQ ILWWLQJV 6ZDJHORN ,QF 6RORQ 2KLRf :KHUH SLSH WKUHDG FRQQHFWLRQV ZHUH QHFHVVDU\ 7HIORQ r1 WDSH ZDV XVHG WR SUHYHQW OHDNDJH %HIRUH H[SHULPHQWV ZHUH LQLWLDWHG WKH HQWLUH V\VWHP ZDV OHDNFKHFNHG DQG 12[ GHFD\ UHDFWLYLW\f PHDVXUHPHQWV ZHUH PDGH IRU WKH V\VWHP (DFK V\VWHP FRPSRQHQW ZDV LVRODWHG DQG HYDOXDWHG VHSDUDWHO\ 2YHUDOO 12 DQG 14 UHPRYDO DV D UHVXOW RI V\VWHP UHDFWLYLW\ ZHUH HVWLPDWHG DW OHVV WKDQ RQH SHUFHQW ,Q WKLV ZD\ DQ\ FKDQJHV LQ 12[ FRQFHQWUDWLRQ ZLWK WLPH FRXOG EH DWWULEXWHG VROHO\ WR WKH SUHVHQFH RI WKH VRUEHQW PDWHULDO LQ WKH IL[HGEHG UHDFWRU

PAGE 74

$LU )ORZ P/PLQf 5RWDPHWHU 5HDGLQJ 2PHJD 'DWD f§,f§ /DE &DOLEUDWLRQ )LJXUH ([DPSOH ORZIORZUDWH URWDPHWHU FDOLEUDWLRQ FXUYHV 2PHJD 167 7XEHf

PAGE 75

$LU )ORZ P/PLQf 5RWDPHWHU 5HDGLQJ 2PHJD 'DWD f§,f§ /DE &DOLEUDWLRQ )LJXUH ([DPSOH KLJKIORZUDWH URWDPHWHU FDOLEUDWLRQ FXUYHV 2PHJD 12* 7XEHf

PAGE 76

7XEXODU )L[HGEHG 5HDFWRUV 7KH UHDFWRUV DQG PL[LQJ DQG VDPSOLQJ PDQLIROGV ZHUH PDFKLQHG IURP VWDLQOHVV VWHHO VHDPOHVV SLSH VWRFN FP ,'f 705 (QJLQHHULQJ 0LFDQRS\ )ORULGDf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f DQG FRXOG EH HDVLO\ UHPRYHG EHWZHHQ UXQV 2QHHLJKWK LQFK GLDPHWHU LQWHUQDO VDPSOLQJ OLQHV ZHUH XVHG WR PLQLPL]H WKH FURVV VHFWLRQDO DUHD EORFNHG E\ WKH OLQHV ZKLFK FRXOG GLVUXSW IORZ WKURXJK WKH EHG 7KH IUDFWLRQDO DUHD EORFNHG E\ WKH FURVVVHFWLRQ RI D VDPSOLQJ OLQH ZDV DSSUR[LPDWHO\ SHUFHQW $ WRWDO RI KROHV FP GLDPHWHUf ZHUH GULOOHG LQ WKH VDPSOLQJ OLQHV $V VKRZQ LQ )LJXUH WKHVH KROHV ZHUH SODFHG LQ WKUHH SODFHV ORFDWHG DW WKH FHQWURLGV RI WZR FRQFHQWULF FLUFOHV RI HTXDO DUHD ZLWKLQ WKH UHDFWRU GLDPHWHU (DFK VDPSOLQJ ORFDWLRQ ZDV IURP WKUHH VHWV RI WZR SDLUV RI KROHV WKH ODWWHU RIIVHW E\ r IURP HDFK RWKHU

PAGE 77

3DFNHG %HG 5HJLRQ FP FP FP FP 66 )L[HGEHG 5HDFWRU 6FKHPDWLF 1RW WR 6FDOHf 7RWDO /HQJWK $SSUR[LPDWHO\ FP )LJXUH 6FKHPDWLF UHSUHVHQWDWLRQ RI VWDLQOHVV VWHHO SDFNHGEHG UHDFWRU HQG FRQQHFWLRQV QRW VKRZQf R? A

PAGE 78

5HDFWRU &URVV 6HFWLRQ FP ,'f )LJXUH 6FKHPDWLF UHSUHVHQWDWLRQ RI LQWHUQDO VDPSOLQJ WXEH ORFDWLRQ DQG DSSHDUDQFH R?

PAGE 79

0L[LQJ RU 6DPSOLQJ /LQH $WWDFKPHQW 3RLQWV 66 0L[LQJ RU 6DPSOLQJ 0DQLIROG 7RWDO /HQJWK $SSUR[LPDWHO\ FP )LJXUH 6FKHPDWLF UHSUHVHQWDWLRQ RI VWDLQOHVV VWHHO PL[LQJ RU VDPSOLQJ PDQLIROG

PAGE 80

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f WR DOORZ IRU WKH LQGHSHQGHQW PHDVXUHPHQW RI UHDFWLRQ WHPSHUDWXUH IRU FRPSDULVRQ ZLWK WKH LQGLFDWHG IXUQDFH WHPSHUDWXUH 7KH PL[LQJ DQG VDPSOLQJ PDQLIROGV ZHUH VLPSO\ VHFWLRQV RI VWDLQOHVV VWHHO SLSH DSSUR[LPDWHO\ FP ORQJ LQFKHVf ZLWK VL[ SHUPDQHQWO\ HPSODFHG VWDLQOHVV VWHHO FRPSUHVVLRQ ILWWLQJV IRU WKH FRQQHFWLRQ RI JDV WUDQVIHU OLQHV DQGRU YDOYHV DV DSSURSULDWH %HIRUH XVH PDFKLQHG SDUWV ZHUH FOHDQHG DQG GHJUHDVHG ZLWK PHWKDQRO IROORZHG E\ DFHWRQH DQG D GHLRQL]HG ZDWHU ULQVH &OHDQLQJ DOO SDUWV ZLWK GHLRQL]HG ZDWHU EHWZHHQ UXQV ZDV VXIILFLHQW GXULQJ WKH QRUPDO FRQGXFW RI H[SHULPHQWV )RU D QRPLQDO EHG OHQJWK RI FP WKH EXON YROXPH RI D SDFNHG EHG ZDV DSSUR[LPDWHO\ FP ,QGLYLGXDO VRUEHQW VDPSOHV ZHUH PHDVXUHG RXW XVLQJ D JUDGXDWHG F\OLQGHU ILOOHG WR D ORRVHO\ SDFNHG YROXPH RI DSSUR[LPDWHO\ FP 7KHVH VDPSOHV ZHUH WUDQVIHUUHG WR SUHZHLJKHG VHDODEOH SRO\HWK\OHQH VDPSOH FRQWDLQHUV DQG ZHLJKHG DSSUR[LPDWH PDVV JVDPSOHf 6LQFH 0J2 FRPSULVHG DSSUR[LPDWHO\ b RI WKH WRWDO VRUEHQW E\ PDVV HDFK EHG FRQWDLQHG DSSUR[LPDWHO\ J 0J2 $IWHU VDPSOLQJ OLQHV ZHUH LQVHUWHG LQWR WKH UHDFWRU WKH EHG ZDV SDFNHG EHWZHHQ FLUFXODU VWDLQOHVV VWHHO VFUHHQV [ PHVKf DW WKH FHQWHU RI WKH FOHDQ WXEH XVLQJ D PDUNHG VWHHO UDP URG WR HQVXUH WKDW WKH EHG ZDV SURSHUO\ ORFDWHG (QG FRQQHFWLRQV ZHUH UHSODFHG DQG D 7\SH 1LFNHO&KURPLXP YHUVXV 1LFNHO$OXPLQXPf WKHUPRFRXSOH 0RGHO ;&,% 2PHJD

PAGE 81

(QJLQHHULQJ ,QF 6WDPIRUG &RQQHFWLFXWf DQG PHWHU ZHUH DWWDFKHG WR WKH RXWVLGH RI WKH UHDFWRU ([SHULPHQWDO 3URFHGXUHV 7KH SDFNHG UHDFWRU ZDV SODFHG LQVLGH WKH WXEH IXUQDFH ZLWK WKH EHG FHQWHUHG ZLWKLQ WKH IXUQDFH DQG WKH WXEH HOHPHQW PRXQWHG YHUWLFDOO\ RQ WKH IXUQDFH FRQWURO EDVH 7KH YHUWLFDO DUUDQJHPHQW ZDV FKRVHQ WR RSWLPL]H JDVf§VROLG FRQWDFW E\ DYRLGLQJ WKH SRWHQWLDO IRU VROLG VHWWOLQJ DQG FKDQQHOLQJ RI WKH JDV VWUHDP WKURXJK WKH EHG YROXPH 7KH PLGGOH FP RI WKH IXUQDFH ZDV UHSRUWHG WR EH FRPSOHWHO\ WHPSHUDWXUHFRQWUROOHG E\ WKH PDQXIDFWXUHU SURYLGLQJ D VXIILFLHQW OHQJWK IRU LVRWKHUPDO H[SHULPHQWDO FRQGLWLRQV 2QFH WKH UHDFWRU ZDV LQ SODFH VDPSOLQJ OLQHV ZHUH LQGLYLGXDOO\ FRQQHFWHG WR WZRZD\ VWDLQOHVV VWHHO YDOYHV 0RGHO :O :KLWH\ &R +LJKODQG +HLJKWV 2KLRf DW WKH VDPSOLQJ PDQLIROG *DV OLQHV HQWHULQJ WKH WXEH IXUQDFH ZHUH ZUDSSHG LQ KLJKWHPSHUDWXUH KHDWLQJ WDSHV 7KHVH KDG EHHQ SUHYLRXVO\ FDOLEUDWHG XVLQJ D 7\SH WKHUPRFRXSOH WR PHDVXUH VXUIDFH WHPSHUDWXUH DW YDULRXV VHWWLQJV RQ WKH YDULDEOH YROWDJH FRQWUROOHU 'HSHQGLQJ RQ WKH UHDFWLRQ WHPSHUDWXUH RI LQWHUHVW WKH KHDW WDSH YROWDJH FRQWUROOHU ZDV VHW WR SURGXFH D JDV SUHKHDW WHPSHUDWXUH DSSUR[LPDWHO\ r& DERYH UHDFWLRQ WHPSHUDWXUH 7KH WXEH IXUQDFH ZDV WXUQHG RQ DQG DOORZHG WR KHDW XS IRU DSSUR[LPDWHO\ RQH KRXU WR DFKLHYH WHPSHUDWXUH FRQWURO ZKLOH WKH EHG ZDV VLPXOWDQHRXVO\ EHLQJ IOXVKHG ZLWK DSSUR[LPDWHO\ V/SP GU\ 1 $IWHU WKLV FRQGLWLRQLQJ SHULRG WKHUPRFRXSOH UHDGLQJV IRU WKH WXEH IXUQDFH DQG LQGHSHQGHQW UHDFWRU WKHUPRFRXSOH ZHUH

PAGE 82

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r1 IODVN 6RXWKHUQ 6FLHQWLILF ,QF 0LFDQRS\ )ORULGDf FRQWDLQLQJ GHLRQL]HG ZDWHU 7KH IODVN ZDV SODFHG LQ D KHDWHG IODVN PDQWOH FRQQHFWHG WR D VHSDUDWH YROWDJH FRQWUROOHU 9DULRXV WHPSHUDWXUHIORZ UDWH FRPELQDWLRQV ZHUH HYDOXDWHG WR DOORZ IRU b ZDWHU YDSRU WR EH SLFNHG XS E\ WKH IORZLQJ JDV VWUHDP 7KH KXPLGLILHG JDVHV ZHUH WKHQ UHKHDWHG WR HQVXUH WKDW WKH\ ZHUH DSSUR[LPDWHO\ r& DERYH UHDFWLRQ WHPSHUDWXUH 7KH DFWXDO ZDWHU YDSRU FRQWHQW RI WKH JDVHV ZDV GHWHUPLQHG E\ ZHLJKLQJ FROOHFWHG FRQGHQVHG ZDWHU IURP D VWDLQOHVV VWHHO FRDOHVFLQJ ILOWHU 0RGHO 6 %DOVWRQ ,QF +DYHUKLOO 0DVVDFKXVHWWVf DW WKH EHG RXWOHW :DWHU YDSRU ZDV UHPRYHG IURP WKH VDPSOH JDV VWUHDP EHIRUH DQDO\VLV WKURXJK FRQGHQVDWLRQ LQ WKH VDPSOLQJ PDQLIROG *DV IORZ UDWH ZDV DOZD\V FRUUHFWHG IRU WKH SUHVHQFH RI ZDWHU YDSRU 7KH HIIHFW RI PRLVWXUH RQ UHDFWLRQ ZDV VRPHWLPHV HYDOXDWHG VLPSO\ E\ DGGLQJ VXIILFLHQW GHLRQL]HG ZDWHU WR D YLUJLQ EHG WR FRQYHUW DOO 0J2 LQWR 0J2+f 7KLV ZDV YHULILHG E\ WKH SUHVHQFH RI ZDWHU YDSRU DW WKH EHG RXWOHW 2QFH FRQYHUWHG 0J2+f LV VWDEOH RYHU WKH UDQJH RI WHPSHUDWXUHV HYDOXDWHG 6LGJZLFN f

PAGE 83

*DV $QDO\VHV $OO UHDFWLRQV ZHUH PRQLWRUHG E\ VHOHFWLYHO\ UHPRYLQJ JDV VDPSOHV IURP WKH VWDLQOHVV VWHHO WXEXODU UHDFWLRQ YHVVHO WKURXJK VHOHFWHG YDOYHV FRQQHFWHG WR WKH VDPSOLQJ PDQLIROG V\VWHP 12[ FRQFHQWUDWLRQV ZHUH DQDO\]HG XVLQJ D FKHPLOXPLQHVFHQW DQDO\]HU 0RGHO + 7KHUPR (QYLURQPHQWDO ,QVWUXPHQWV )UDQNOLQ 0DVVDFKXVHWWVf EDVHG XSRQ WKH UHDFWLRQ RI 12 ZLWK IROORZLQJ UHGXFWLRQ RI 1 WR 12 7KH IHDVLELOLW\ RI WKH FRPPHUFLDO FKHPLOXPLQHVFHQW DQDO\]HU ZDV HVWDEOLVKHG LQ E\ )RQWLMQ HW DO DQG WKH ILUVW SURWRW\SH ZDV GHYHORSHG LQ 6WHIIHQVRQ DQG 6WHGPDQ f 7KH PHWKRG LV QRZ ZLGHO\ XVHG LQ IDFW WKH XVH RI D FKHPLOXPLQHVFHQW PHWKRG LV UHTXLUHG IRU WKH PHDVXUHPHQW RI 1 LQ WKH DWPRVSKHUH &)5 3DUW $SSHQGL[ )f 7KH SULQFLSOH RI RSHUDWLRQ RI WKH FKHPLOXPLQHVFHQW DQDO\]HU LV EDVHG XSRQ WKH FKHPLOXPLQHVFHQW UHDFWLRQ EHWZHHQ 12 DQG R]RQH f 7KLV UHDFWLRQ LV DV IROORZV 12 2 >1r @ 1 KY f ZKHUH KY ZKLFK UHSUHVHQWV WKH TXDQWLW\ RI OLJKW HQHUJ\ HPLWWHG E\ DQ HOHFWURQLFDOO\ H[FLWHG LQWHUPHGLDU\ 1ff LV VLPSO\ 3ODQFNnV FRQVWDQW Kf PXOWLSOLHG E\ WKH IUHTXHQF\ RI WKH HPLWWHG UDGLDWLRQ Yf LQ WKLV FDVH ZLWK D ZDYHOHQJWK JUHDWHU WKDQ QP $ VDPSOH RI JDV LV GUDZQ IURP D VHOHFWHG SRLQW LQ WKH V\VWHP E\ DQ H[WHUQDO SXPS 7KH VDPSOH LV PL[HG ZLWK R]RQH JHQHUDWHG E\ DQ LQWHUQDO R]RQDWRU LQVLGH D UHDFWLRQ FKDPEHU 7KH UHVXOWDQW UHDFWLRQ SURGXFHV D FKDUDFWHULVWLF OXPLQHVFHQFH ZLWK DQ LQWHQVLW\ GLUHFWO\ SURSRUWLRQDO WR 12 FRQFHQWUDWLRQ VLQFH LV SURYLGHG LQ H[FHVV DQG LWV FRQFHQWUDWLRQ FDQ EH FRQVLGHUHG D

PAGE 84

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f 7KH LQVWUXPHQW VSHFLILFDWLRQV DV UHSRUWHG LQ WKH PDQXIDFWXUHUnV PDQXDO LQGLFDWH VHYHUDO VFDOHV SURYLGLQJ DQ RYHUDOO PHDVXUHPHQW UDQJH IURP ]HUR WR SSP ZLWK D OLPLW RI GHWHFWLRQ RI SSE SSPf %DFNJURXQG QRLVH OHYHO LV SSE SSPf ZHOO EHORZ WKH GHWHFWLRQ OLPLW 7KH LQVWUXPHQW H[KLELWV OLQHDULW\ LQ UHVSRQVH RI s b RI IXOO VFDOH ZLWK D ]HUR GULIW KRXU PD[LPXPf RI SSE SSPf DQG D VSDQ GULIW KRXU PD[LPXPf RI s b RI IRLO VFDOH 7KH VHSDUDWH VFDOHV DUH OLQHDU RYHU WKHLU HQWLUH UDQJHV ZKLFK ZDV YHULILHG E\ PXOWLSOHSRLQW FDOLEUDWLRQV ZLWK FHUWLILHG JDVHV :KLOH WKH LQWHUIHUHQFH RI +1 DQG +212 ZLWK 1 PHDVXUHPHQWV LQ FKHPLOXPLQHVFHQW DQDO\]HUV LV ZHOO GRFXPHQWHG 6SLFHU HW DO -RVHSK DQG 6SLFHU f WKH FRQFHQWUDWLRQV RI WKHVH LQWHUIHUHQWV ZHUH H[SHFWHG WR EH ORZ LQ WKH V\VWHP HPSOR\HG :KHQ VXIILFLHQW PRLVWXUH ZDV SUHVHQW WR IRUP

PAGE 85

DFLGV WKH\ ZHUH XOWLPDWHO\ FROOHFWHG LQ WKH FRDOHVFLQJ ILOWHU RU FRQGHQVHG RXW LQ WKH VDPSOLQJ PDQLIROG SULRU WR DQDO\VLV 6LPLODUO\ WKH SUHVHQFH RI ZDWHU YDSRU LWVHOI FDQ LQWHUIHUH ZLWK 12[ PHDVXUHPHQWV GLUHFWO\ E\ 1 DEVRUSWLRQ DV ZHOO DV YLD WKLUGERG\ TXHQFKLQJ RI FKHPLOXPLQHVFHQFH 7UGRQD HW DO &DPSEHOO HW DO f 7KHUHIRUH LW LV LPSRUWDQW WR PLQLPL]H FRQWDFWn WLPH RI WKH JDV ZLWK ZDWHU YDSRU LQ VDPSOLQJ OLQHV $Q\ 1 UHPRYHG E\ JDV KXPLGLILFDWLRQ LV DFFRXQWHG IRU E\ WKH IDFW WKDW EHG LQOHW JDVHV ZHUH VDPSOHG DIWHU WKLV SRLQW LQ WKH V\VWHP 7KH DQDO\]HU UHDFWLRQ FKDPEHU LV PDLQWDLQHG DW D YDFXXP RI DSSUR[LPDWHO\ LQFKHV RI PHUFXU\ N3Df WR PLQLPL]H WKLUGERG\ TXHQFKLQJ HIIHFWV $QRWKHU FRPPRQ LQWHUIHUHQFH ZLWK 12[ PHDVXUHPHQWV LQ WKH DWPRVSKHUH LV FDXVHG E\ WKH SUHVHQFH RI VXFK RWKHU QLWURJHQ FRQWDLQLQJ FRPSRXQGV DV SHUR[\DFHW\O QLWUDWH 3$1f ZKLFK FDQ EH FRQYHUWHG LQWR 12 LQ WKH WKHUPDO FRQYHUWHU $JDLQ WKHVH VSHFLHV DUH QRW VXVSHFWHG WR EH SUHVHQW LQ WKH H[SHULPHQWDO SURFHGXUHV XVHG LQ WKLV VWXG\ 7R SUHYHQW SDUWLFXODWH PDWWHU FRQWDPLQDWLRQ RI WKH FKHPLOXPLQHVFHQW DQDO\]HU D SP SRUH VL]H 7HIORQr1 ILOWHU *HOPDQ 6FLHQFHV $QQ $UERU 0LFKLJDQf ZDV SODFHG LQ WKH VDPSOH OLQH DW WKH DQDO\]HU VDPSOH LQOHW 7KLV ILOWHU ZDV IRXQG WR EH HVVHQWLDOO\ QRQUHDFWLYH WR 12[ :LWKRXW VXFK D ILOWHU LW LV TXLWH HDV\ WR H[SHULHQFH EORFNDJH RI LQWHUQDO FDSLOODU\ WXELQJ 'HSRVLWV FRXOG DOVR IRUP RQ WKH JODVV ILOWHU WKDW VHSDUDWHV WKH UHDFWLRQ FKDPEHU IURP WKH SKRWRPXOWLSOLHU WXEH FDXVLQJ D ORVV LQ LQVWUXPHQW VHQVLWLYLW\ .ODSKHFN DQG :LQNOHU f $ FRQWLQXRXV JDV DQDO\]HU 0RGHO ,5 ,QIUDUHG ,QGXVWULHV ,QF 6DQWD %DUEDUD &DOLIRUQLDf ZDV XVHG WR HOHFWURFKHPLFDOO\ PHDVXUH FRQFHQWUDWLRQV 7KH SULQFLSOH RI

PAGE 86

RSHUDWLRQ XSRQ ZKLFK WKLV LQVWUXPHQW LV EDVHG LV D VLPSOH FRXORPHWULF SURFHVV LQ ZKLFK R[\JHQ LQ WKH VDPSOH JDV VWUHDP LV UHGXFHG LQ DQ HOHFWURFKHPLFDO FHOO 7KH JDV VWUHDP HQWHUV WKH FDWKRGH FDYLW\ ZLWK DQ\ R[\JHQ EHLQJ PHWHUHG WR WKH FDWKRGH WKURXJK D GLIIXVLRQ EDUULHU $W WKH FDWKRGH R[\JHQ LV HOHFWURFKHPLFDOO\ UHGXFHG YLD WKH UHDFWLRQ + Hn + f $Q HOHFWURO\WH VROXWLRQ LQ WKH FDWKRGH FDYLW\ FRQWDLQV SRWDVVLXP K\GUR[LGH .2+f ZKLFK IDFLOLWDWHV PLJUDWLRQ RI WKH JHQHUDWHG K\GUR[\O LRQV WR DQ DQRGH ZKHUH WKH\ DUH R[LGL]HG EDFN WR R[\JHQ 7KLV RFFXUV IROORZLQJ WKH UHDFWLRQ + + Hr f 7KH UHVXOWLQJ FHOO FXUUHQW LV GLUHFWO\ SURSRUWLRQDO WR WKH R[\JHQ FRQFHQWUDWLRQ LQ WKH VDPSOH JDV VWUHDP 7KLV FXUUHQW LV PHDVXUHG HOHFWURQLFDOO\ DQG LV FRQYHUWHG LQWR DQ LQGLFDWHG FRQFHQWUDWLRQ YDOXH 7KH LQVWUXPHQW PHDVXUHV R[\JHQ FRQFHQWUDWLRQ EHWZHHQ ]HUR DQG b ZLWK D PDQXIDFWXUHUUHSRUWHG DFFXUDF\ RI s b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

PAGE 87

,QGLFDWHG >12@ SSPf 6WDQGDUG >12@ SSPf /DE &RQGLWLRQV )LJXUH 7\SLFDO FKHPLOXPLQHVFHQW 12[ DQDO\]HU FDOLEUDWLRQ FXUYH IRU 12 FKDQQHOf

PAGE 88

/Q ,QVWUXPHQW >@ 3HUFHQWf 0LG 6FDOH r +LJK 6FDOH )LJXUH 7\SLFDO ,5 R[\JHQ DQDO\]HU FDOLEUDWLRQ FXUYH

PAGE 89

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f ZDV XVHG WR GHWHUPLQH WKH DFWXDO DQDO\]HU 12[ FRQYHUWHU HIILFLHQF\ DQG HQVXUH WKDW WKH HIILFLHQF\ H[FHHGHG b 6RUEHQW 6XUIDFH $UHD 'HWHUPLQDWLRQ ,Q SUHSDUDWLRQ IRU SDFNHGEHG VWXGLHV %(7 VXUIDFH DUHDV RI VHOHFWHG VDPSOHV RI ODERUDWRU\SUHSDUHG 0J2YHUPLFXOLWH VRUEHQW 6RUEWHFKVXSSOLHG VRUEHQW DQG UDZ FRDUVH YHUPLFXOLWH ZHUH GHWHUPLQHG XVLQJ D 0LFURPHUHWLFV $6$3 DXWRPDWHG VXUIDFHDUHD DQG SRUH VL]H DQDO\]HU 'XSOLFDWH VDPSOHV ZHUH UXQ WR YHULI\ WKH YDOLGLW\ RI WKH GDWD 6PDOO VDPSOHV RI VRUEHQW DSSUR[LPDWHO\ RQH JUDPf ZHUH ILUVW GHJDVVHG DW DQ DEVROXWH SUHVVXUH RI WR PP +J DW IRU DSSUR[LPDWHO\ HLJKW KRXUV 1LWURJHQ JDV LQ KHOLXPf ZDV VHTXHQWLDOO\ DGGHG WR WKH VDPSOH WXEH DW D WHPSHUDWXUH RI 7KH YROXPHV RI QLWURJHQ DGVRUEHG DW YDULRXV SDUWLDO SUHVVXUHV EHWZHHQ DSSUR[LPDWHO\ DQG DWPRVSKHUH ZHUH

PAGE 90

UHFRUGHG E\ WKH DQDO\]HU 7KHVH GDWD ZHUH SORWWHG DFFRUGLQJ WR WKH %(7 PRGHO WR GHWHUPLQH WKH WRWDO VXUIDFH DUHD RI WKH VDPSOH ,QLWLDO VFUHHQLQJ UHVXOWV ZHUH XVHG DV DQ LQGLFDWLRQ RI WKH EHVW SUHSDUDWLRQ PHWKRG WR XVH LQ PDNLQJ WKH VRUEHQW %DVHG XSRQ WKHVH UHVXOWV D ODUJH TXDQWLW\ RI VRUEHQW ZDV SUHSDUHG IROORZLQJ WKH SURFHGXUH DOUHDG\ GHVFULEHG )LYH DGGLWLRQDO VDPSOHV ZHUH WDNHQ DQG DQDO\]HG XVLQJ WKH %(7 PHWKRG IRU SUHFLVLRQ HVWLPDWHV 'HWHUPLQDWLRQ RI 6RUEHQW 3DUWLFOH 6L]H 7KH 0J2f§YHUPLFXOLWH VRUEHQW LV LQKHUHQWO\ D KLJKO\ KHWHURJHQHRXV PDWHULDO $V D UHVXOW WKH UDQJH RI SDUWLFOH VL]HV LQ D JLYHQ VRUEHQW VDPSOH FDQ EH JUHDW 6DPSOHV RI 0J2f§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f VOXUU\ DIIHFWV SDUWLFOH VL]H 5HVXOWV RI WKHVH GHWHUPLQDWLRQV DUH SUHVHQWHG LQ WKH IROORZLQJ FKDSWHU

PAGE 91

&+$37(5 5(68/76 $1' ',6&866,21 ,QWULQVLF .LQHWLF 6WXGLHV ([SHULPHQWV ZHUH FRQGXFWHG WR FROOHFW GDWD WKDW SURYLGH LQVLJKWV LQWR UHOHYDQW SDUDPHWHUV FRQWUROOLQJ 12[ UHPRYDO HIILFLHQFLHV SRWHQWLDO UHDFWLRQ PHFKDQLVPVf DQG UHDFWLRQ NLQHWLFV 'XSOLFDWH UXQV ZHUH FRQGXFWHG IRU HYHU\ VHW RI FRQGLWLRQV WR JHQHUDWH WKH GDWD QHFHVVDU\ IRU WKH GHWHUPLQDWLRQ RI NLQHWLF SDUDPHWHUV $ FRPSDULVRQ RI UHVXOWV EHWZHHQ GXSOLFDWH UXQV XQGHU VLPLODU FRQGLWLRQV VKRZHG UHPDUNDEOH UHSURGXFLELOLW\ VWDWLVWLFDO HUURU bf 7R GHWHUPLQH LQWULQVLF UHDFWLRQ NLQHWLFV 12[ FRQFHQWUDWLRQV ZHUH IROORZHG ZLWK LQFUHDVLQJ EHG H[SRVXUH WLPH IURP D VHOHFWHG VDPSOLQJ YDOYH ZKLFK UHSUHVHQWHG D VSHFLILF EHG UHVLGHQFH WLPH 7KLV ZDV QHFHVVDU\ EHFDXVH RI D V\VWHP DUWLIDFW WKDW FUHDWHG D ODJWLPH RI DSSUR[LPDWHO\ PLQXWHV WR DOORZ IRU OLQH IOXVKLQJ XQWLO D VWDEOH LQVWUXPHQW UHDGLQJ ZDV UHDFKHG DIWHU D VDPSOLQJ YDOYH ZDV VHOHFWHG 7R H[SODLQ IXUWKHU IRU D JLYHQ VHW RI FRQGLWLRQV IRXU VHSDUDWH UXQV ZRXOG EH QHFHVVDU\ WR FROOHFW GDWD IRU DOO ILYH VDPSOLQJ SRLQWV EHG LQOHW WKUHH SRLQWV ZLWKLQ WKH EHG DQG EHG RXWOHW %HG LQOHW YDOXHV ZHUH FKHFNHG RQ DOO UXQV EHIRUH DQG DIWHU WKH RWKHU GDWD ZHUH FROOHFWHG (YHU\ VHFRQGV WKH WKUHH 12[ DQDO\]HU FKDQQHOV GLVSOD\HG VHFRQG DYHUDJH 12 1 DQG 12[ FRQFHQWUDWLRQ YDOXHV >1@ FDOFXODWHG E\ GLIIHUHQFHf 'DWD ZHUH FROOHFWHG HYHU\ PLQXWH WKURXJKRXW WKH UXQ )URP WKHVH GDWD

PAGE 92

FRQFHQWUDWLRQ YHUVXV EHG UHVLGHQFH WLPH FXUYHV FRXOG EH SORWWHG 9DOXHV IRU WKH VDPH EHG H[SRVXUH WLPH ZHUH XVHG WR SORW FRQFHQWUDWLRQ RI LQWHUHVW YHUVXV UHVLGHQFH WLPH ZKLFK ZDV D IXQFWLRQ RI VDPSOLQJ SRLQW ORFDWLRQ DQG WKXV UHDFWLRQ WLPH /LPLWDWLRQ RI *DV )LOP 0DVV 7UDQVIHU 5HVLVWDQFH ,Q WKH FRQGXFW RI LQWULQVLF FKHPLFDO NLQHWLF VWXGLHV RI JDVf§VROLG UHDFWLRQV LQ SDFNHG EHG UHDFWRU V\VWHPV LW LV FULWLFDO WKDW JDV ILOP PDVV WUDQVIHU UHVLVWDQFH WR UHDFWLRQ EH PLQLPL]HG 7KH IROORZLQJ GHVFULSWLRQ RI WKH OLPLWDWLRQ RI WKLV UHVLVWDQFH LV EDVHG XSRQ WKH ZRUN RI )RJOHU f DQG 6]HNHO\ HW DO f 7KH PDVV IOX[ IURP D IOXLG WR D VROLG VXUIDFH LV WKH SURGXFW RI D PDVV WUDQVIHU FRHIILFLHQW DQG WKH FRQFHQWUDWLRQ GULYLQJ IRUFH :KHQ JDV ILOP PDVV WUDQVIHU LV UDWHOLPLWLQJ WKH H[WHQW RI UHDFWLRQ RU IUDFWLRQDO VROLG FRQYHUVLRQ LV OLQHDU ZLWK UHVSHFW WR WLPH ,Q D SDFNHGEHG V\VWHP WKH FRQFHQWUDWLRQ GULYLQJ IRUFH LV WKH GLIIHUHQFH EHWZHHQ UHDFWDQW FRQFHQWUDWLRQ DW WKH YRLG VSDFH DW D SRLQW LQ WKH EHG DQG WKH FRQFHQWUDWLRQ DW WKH VXUIDFH RI WKH VRUEHQW SDUWLFOH DGMDFHQW WR WKLV SRLQW *DV ILOP PDVV WUDQVIHU UHVLVWDQFH FDQ JHQHUDOO\ EH PLQLPL]HG E\ XVLQJ VPDOO VRUEHQW SDUWLFOHV DQG KLJK VXSHUILFLDO JDV YHORFLWLHV $W ORZ YHORFLWLHV WKH VWDJQDQW JDV ILOP ERXQGDU\ OD\HU WKLFNQHVV LV ODUJH f§ DW KLJK YHORFLWLHV WKLV WKLFNQHVV LV GHFUHDVHG UHGXFLQJ WKH PDVV WUDQVIHU OLPLWDWLRQ WKURXJK LQFUHDVHG FRQYHFWLYH PDVV WUDQVIHU 2SHUDWLQJ DW ORZHU UHDFWLRQ WHPSHUDWXUHV DOVR KHOSV WR UHGXFH JDV ILOP PDVV WUDQVIHU OLPLWDWLRQV :KHQ D UHDFWLRQ LV JDV ILOP PDVV WUDQVIHU OLPLWHG LW ZLOO JHQHUDOO\ EH LQVHQVLWLYH WR LQFUHDVHV LQ WHPSHUDWXUH 6LQFH JDV ILOP PDVV WUDQVIHU LV QRW DQ DFWLYDWHG SURFHVV ZKLOH

PAGE 93

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f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

PAGE 94

%HG ([SRVXUH 7LPH PLQf DOSP DOSP DOSP )LJXUH *DV ILOP PDVVWUDQVIHU UHVLVWDQFH HYDOXDWLRQ 1 SHQHWUDWLRQ &A&Af YHUVXV EHG H[SRVXUH WLPH >12@LQ SSP 7 >2@ bf

PAGE 95

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f $ % & ( 0HDQ 6DPSOH 6WDQGDUG 'HYLDWLRQ

PAGE 96

3DUWLFOH 6L]H 'LVWULEXWLRQV $ VDPSOH RI 0J2YHUPLFXOLWH VRUEHQW ZDV VLHYHG DV SUHYLRXVO\ GHVFULEHG DQG UHVXOWV ZHUH SORWWHG RQ D ORJSUREDELOLW\ JUDSK VKRZQ LQ )LJXUH 6LQFH PDQ\ SDUWLFOH VL]H GDWD DUH ORJQRUPDOO\ GLVWULEXWHG WKLV W\SH RI SORW VKRXOG SURGXFH D OLQHDU UHSUHVHQWDWLRQ IURP ZKLFK UHOHYDQW VWDWLVWLFDO SDUDPHWHUV f§ PRVW LPSRUWDQWO\ WKH PDVV PHGLDQ GLDPHWHU DQG JHRPHWULF VWDQGDUG GHYLDWLRQ f§ FDQ EH GHWHUPLQHG )URP )LJXUH LW FDQ EH VHHQ WKDW WKH GLVWULEXWLRQ LV IDLUO\ OLQHDU EHWZHHQ WKH WK DQG WK SHUFHQWLOHV VR WKHVH GDWD ZHUH XVHG IRU WKLV SXUSRVH )URP WKH GLVWULEXWLRQ WKH PDVV PHGLDQ GLDPHWHU 00'f LV DSSUR[LPDWHO\ PP DQG WKH JHRPHWULF VWDQGDUG GHYLDWLRQ 2Jf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

PAGE 97

0J2f§9HUPLFXOLWH 3DUWLFOH 6L]H 'S PPf cSUREDELOLW\ SORW RI 0J2YHUPLFXOLWH VRUEHQW SDUWLFOH VL]H GLVWULEXWLRQ

PAGE 98

)LJXUH /RJSUREDELOLW\ SORW RI $NURFKHP (ODVWRPDJ p 0J2 SRZGHU SDUWLFOH VL]H GLVWULEXWLRQ PDQXIDFWXUHUSURYLGHG GDWDf

PAGE 99

RQ WKH SDUWLFOHV DUH D IXQFWLRQ RI SDUWLFOH VXUIDFH DUHD 7KH 0J2 SDUWLFOH VXUIDFH PHGLDQ GLDPHWHU FDQ EH FDOFXODWHG IURP WKH ORJSUREDELOLW\ GLVWULEXWLRQ XVLQJ WKH ZHOONQRZQ +DWFK &KRDWH HTXDWLRQV +LQGV f $ FDOFXODWHG FRXQW PHGLDQ GLDPHWHU RI DSSUR[LPDWHO\ SP OHG WR WKH GHWHUPLQDWLRQ RI D YDOXH RI DSSUR[LPDWHO\ SP IRU WKH VXUIDFH PHGLDQ GLDPHWHU XVLQJ WKHVH HTXDWLRQV 12 5HPRYDO E\ 0JI129FRDWHG 9HUPLFXOLWH 5HVXOWV IRU 0J1fFRDWHG YHUPLFXOLWH ZHUH VLPLODU WR WKRVH IRU WKH SODLQ KHDW WUHDWHG YHUPLFXOLWH 5HPRYDOV RI 1 ZHUH OHVV WKDQ ILYH SHUFHQW RYHU WKH UDQJH RI WHPSHUDWXUHV WR .f DQG FRQFHQWUDWLRQV WR SSPf HYDOXDWHG 12 UHPRYDO HIILFLHQFLHV ZHUH PXFK ORZHU DQG DSSURDFKHG ]HUR 7KH UHVXOWV RI WKHVH WHVWV ZHUH XVHIXO LQ YHULI\LQJ WKDW LI PDJQHVLXP QLWUDWH LV IRUPHG RQ WKH VROLG VXUIDFH GXULQJ WKH FRXUVH RI WKH 0J2 0J2+ff UHDFWLRQ ZLWK 1 DQG DOO WKH UHPDLQLQJ 0J2 RU 0J2+ff LV XWLOL]HG RU LV RWKHUZLVH XQDYDLODEOH IRU UHDFWLRQ WKHQ WKHVH UHVXOWV SUHFOXGH WKH IXUWKHU UHPRYDO RI 12[ IURP WKH JDVHV VXSSOLHG WR WKH EHG DQG UHDFWLRQUHPRYDO VKRXOG FHDVH 12 5HPRYDO E\ 6RUEHQW 0DWHULDO ,W ZDV IRXQG WR EH PRUH GLIILFXOW WR UHPRYH 12 WKDQ 1 E\ UHDFWLRQ ZLWK 0J2 WKH IRUPHU EHLQJ JHQHUDOO\ OHVV UHDFWLYH WKDQ WKH ODWWHU 7KH UHPRYDO RI HDFK VSHFLHV ZDV LQGHSHQGHQWO\ DIIHFWHG E\ WKH H[SHULPHQWDO YDULDEOHV HYDOXDWHG 6SHFLILF UHVXOWV DUH GHVFULEHG LQ WKH IROORZLQJ SDUDJUDSKV

PAGE 100

12 5HPRYDO EY 6RUEHQW 0DWHULDO 12 UHPRYDO E\ 0J2FRDWHG YHUPLFXOLWH VRUEHQW DSSHDUHG WR GHSHQG RQ LQOHW 12 FRQFHQWUDWLRQ DQG EHG WHPSHUDWXUH $W ORZ LQOHW 12 FRQFHQWUDWLRQV SSPf 12 UHPRYDO HIILFLHQFLHV YDULHG IURP LQLWLDO YDOXHV RI DSSUR[LPDWHO\ SHUFHQW WR RYHU SHUFHQW ,Q JHQHUDO LQLWLDO 12 UHPRYDO HIILFLHQF\ ZDV JUHDWHU DW WKDQ DW RU LQGLFDWLQJ UHPRYDO E\ SK\VLVRUSWLRQ UDWKHU WKDQ E\ FKHPLVRUSWLRQ 7KH SUHVHQFH RI ZDWHU RU ZDWHU YDSRU GLG LPSURYH WKH UHPRYDO HIILFLHQF\ RI 12 EXW RYHUDOO UDWHV RI UHPRYDO ZHUH VWLOO ORZ $V SUHYLRXVO\ GHVFULEHG DSSDUHQW 12 UHPRYDO ZDV HQKDQFHG E\ WKH SUHVHQFH RI R[\JHQ WKURXJK WKH IRUPDWLRQ RI 1 ,Q WKHLU HYDOXDWLRQ RI 12 UHPRYDO E\ &D2+f VROLGV &KX DQG 5RFKHOOH f IRXQG WKDW WKH JDV RU VROLGSKDVH R[LGDWLRQ RI 12 WR 1 SOD\HG DQ LPSRUWDQW UROH LQ WKH UHPRYDO RI WKH IRUPHU $W KLJKHU 12 FRQFHQWUDWLRQV SSPf WKH RYHUDOO UHPRYDO HIILFLHQFLHV ZHUH ORZHU OHVV WKDQ SHUFHQWf -R]HZLF] HW DO f IRXQG QR 12 IURP D SSP 12 JDV VWUHDP FRQFHQWUDWLRQf UHPRYDO E\ &D2+f VROLGV ZLWKRXW WKH SUHVHQFH RI R[\JHQ DW WHPSHUDWXUHV EHWZHHQ DQG & &DUERQ PRQR[LGH &2f KDV EHHQ ZLGHO\ XVHG DV D SUREH PROHFXOH LQ VXUIDFH VFLHQFH VWXGLHV EHFDXVH LW LV RQH RI WKH VLPSOHVW DGVRUEDWHDGVRUEHQW V\VWHPV 0HMLDV HW DO f IRXQG WKH LQWHUDFWLRQ RI &2 RQ 0J2 WR EH ZHDN DQG RI HOHFWURVWDWLF RULJLQ ZLWKRXW QRWLFHDEOH FKHPLFDO FRQWULEXWLRQV 6LQFH &2 DQG 12 DUH VLPLODUO\ QRQUHDFWLYH WKLV REVHUYDWLRQ VXJJHVWV WKDW WKHUH PD\ EH D IL[HG 12 VRUSWLRQ FDSDFLW\ SRVVLEO\ WKH UHVXOW RI OLPLWHG VXUIDFH DUHD DYDLODEOH IRU SK\VLVRUSWLRQ )LJXUH LV D UHSUHVHQWDWLYH FRQFHQWUDWLRQ SURILOH IRU 12 UHPRYDO E\ D 0J2YHUPLFXOLWH VRUEHQW EHG ZLWKRXW D VLJQLILFDQW TXDQWLW\ RI 1 SUHVHQW $V FDQ EH VHHQ

PAGE 101

>12@ SSPf 5HVLGHQFH WLPH VHFf ([SRVXUH 7LPH KRXU KRXUV KRXUV )LJXUH 12 UHPRYDO E\ KXPLGLILHG 0J2f§YHUPLFXOLWH VRUEHQW >12@LQ SSP 7 >@ b b + YDSRUf RR

PAGE 102

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f ZKHUH & 1 FRQFHQWUDWLRQ LQ SSP W UHVLGHQFH WLPH LQ VHFRQGV DQG N ILUVWRUGHU UDWH FRQVWDQW LQ VHFn ,QWHJUDWLQJ HTXDWLRQ f SURGXFHV

PAGE 103

>1@ SSPf PLQ f§,f§ PLQ PLQ f§%f§ PLQ PLQ PLQ $ PLQ f§;f§ KRXU %HG ([SRVXUH 7LPH )LJXUH 1 UHPRYDO YHUVXV UHVLGHQFH WLPH 7 >@ bf 92 R

PAGE 104

& & H[SNWf f ZKHUH & ,QOHW 1 FRQFHQWUDWLRQ LQ SSP DQG W EHG UHVLGHQFH WLPH LQ VHFRQGV )URP (TXDWLRQ f LW FDQ EH VHHQ WKDW D SORW RI OQ&&f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f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f WR D GLIIXVLRQFRQWUROOHG SURFHVV ZKLFK ZRXOG EH D VORZ SURFHVV ,W ZDV IRXQG WKDW DQ G\QDPLFf HTXLOLEULXP

PAGE 105

)LJXUH )LUVWRUGHU NLQHWLF SORW >12@LQ SSP 7 >2@ bf 'DVKHG OLQHV UHSUHVHQW b FRQILGHQFH OLPLWV 0' .!

PAGE 106

,Q &&Rf 5HVLGHQFH 7LPH VHFf %HG ([SRVXUH 7LPH PLQXWHV PLQXWHV PLQXWHV f§ PLQXWHV )LJXUH )LUVWRUGHU NLQHWLF SORW >12@LQ SSP 7 >2@ bf 92 X!

PAGE 107

FRQGLWLRQ RFFXUUHG DIWHU DSSUR[LPDWHO\ PLQXWHV RI EHG H[SRVXUH WLPH SURYLGLQJ WKH PRVW UHOLDEOH UHVXOWV IRU VXEVHTXHQW H[SHULPHQWV DQG GDWD DQDO\VHV :KLOH WKHVH GDWD ZHUH PRVW XVHIXO IRU WKH SUHGLFWLRQ RI VRUEHQW EHKDYLRU ZLWK WLPH LW PDGH WKH DFFXUDWH GHWHUPLQDWLRQ RI LQWULQVLF RU DSSDUHQWf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f RI LQOHW 1 H[LWLQJ WKH EHG DW DQ\ JLYHQ EHG H[SRVXUH RU UHDFWLRQ WLPH RU &RXW&LQ 2EYLRXVO\ WKLV YDOXH LV LQ SHUFHQWf b PLQXV WKH 1 UHPRYDO HIILFLHQF\ SHUFHQWDJH ,Q

PAGE 108

SSP SSP SSP SSP )LJXUH 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 92

PAGE 109

1 3HQHWUDWLRQ )UDFWLRQf )LJXUH 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 92 2Y

PAGE 110

)LJXUH 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf 92

PAGE 111

JHQHUDO KLJKHU 1 FRQFHQWUDWLRQV SSPf ZHUH UHPRYHG PRUH HIILFLHQWO\ DQG PRUH UDSLGO\ WKDQ ORZHU 1 FRQFHQWUDWLRQV SDUWLFXODUO\ DW HOHYDWHG WHPSHUDWXUH 7KLV ILQGLQJ YDOLGDWHV WKH XVH RI D FRQFHQWUDWLRQGHSHQGHQW FKHPLFDO UHDFWLRQ PRGHO 7KH KLJKHU FRQFHQWUDWLRQV RI 1 PD\ SURYLGH PRUH RSSRUWXQLW\ IRU FRQWDFW ZLWK DFWLYH VLWHV RQ WKH VRUEHQW VXUIDFH DUHD DW ZKLFK WKH FKHPLFDO UHDFWLRQVf WDNH SODFH ,Q VRPH FDVHV WKLV PD\ PHDQ WKDW WKH H[WHUQDO VRUEHQW SDUWLFOH VXUIDFH EHFRPHV FRYHUHG E\ DQ LQHUW 0J1ff SURGXFW OD\HU OHDGLQJ WR UDWH OLPLWDWLRQ E\ D GLIIXVLRQFRQWUROOHG SURFHVV IROORZLQJ WKH VKULQNLQJ XQUHDFWHG FRUH PHFKDQLVP $ KLJKHU FRQFHQWUDWLRQ JUDGLHQW ZRXOG SURYLGH D GULYLQJ IRUFH WR IDFLOLWDWH SRUH GLIIXVLRQ DQG GLIIXVLRQ WKURXJK WKH LQHUW 0J1ff SURGXFW OD\HU WR DYDLODEOH UHPDLQLQJ UHDFWLYH 0J2 VXUIDFH DUHD LQ WKH LQWHULRU SRUWLRQV RI WKH VRUEHQW SDUWLFOHV 5HSUHVHQWDWLYH 1 SHQHWUDWLRQ GDWD SUHVHQWHG LQ )LJXUH VKRZV WKDW DW D UHDFWLRQ WHPSHUDWXUH RI LQLWLDO 1 SHQHWUDWLRQ ZDV ORZ OHVV WKDQ RQH SHUFHQWf VXJJHVWLQJ WKDW WKH 1 UHPRYDO HIILFLHQF\ ZDV JUHDWHU WKDQ b IRU DOO FRQFHQWUDWLRQV HYDOXDWHG $IWHU DQ LQLWLDO FRQGLWLRQLQJ SHULRG RI DSSUR[LPDWHO\ PLQXWHV WKHUH LV D GLVWLQFW GLYHUVLRQ EHWZHHQ WKH FXUYHV UHSUHVHQWLQJ DQG SSP DQG WKH FXUYHV UHSUHVHQWLQJ DQG SSP 1 SHQHWUDWLRQ LQFUHDVHG WR DSSUR[LPDWHO\ b IRU WKH ODWWHU ZKLOH LW UHPDLQHG DW DSSUR[LPDWHO\ b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

PAGE 112

VXUIDFH LV VWLOO EHLQJ FRQYHUWHG LQWR DQ LQHUW SURGXFW OD\HU DIWHU WKH VDPH SHULRG RI H[SRVXUH WLPH )URP )LJXUH DW D UHDFWLRQ WHPSHUDWXUH RI LQLWLDO 1 SHQHWUDWLRQ LV ORZHVW IRU WKH WZR ORZHU LQOHW FRQFHQWUDWLRQV 6LQFH WKHUH LV OHVV HQHUJ\ EHLQJ VXSSOLHG YLD KHDW HQHUJ\ WR RYHUFRPH WKH DFWLYDWLRQ HQHUJ\ EDUULHU LW PD\ EH WKDW RQO\ D OLPLWHG QXPEHU RI PROHFXOHV FDQ EH UHPRYHG $IWHU DSSUR[LPDWHO\ PLQXWHV RI EHG H[SRVXUH WLPH KDYH HODSVHG WKH RYHUDOO SHQHWUDWLRQ QXPEHUV DJDLQ FRQYHUJHG WRZDUG D YDOXH EHWZHHQ DSSUR[LPDWHO\ DQG SHUFHQW $W D UHDFWLRQ WHPSHUDWXUH RI )LJXUH f WKHUH ZDV QR GLVFHUQLEOH SDWWHUQ REVHUYH IRU WKH LQLWLDO SHQHWUDWLRQ YDOXHV DOWKRXJK WKH YDOXH IRU DQ LQOHW 1 FRQFHQWUDWLRQ RI SSP UHPDLQHG WKH ORZHVW 7RZDUG WKH HQG RI WKH EHG H[SRVXUH WLPH SHULRG KRZHYHU 1 SHQHWUDWLRQV ZHUH ORZHVW IRU WKH WZR KLJKHVW FRQFHQWUDWLRQV 7KLV PD\ EH DQ LQGLFDWLRQ RI WKH KLJKHU FRQFHQWUDWLRQ JUDGLHQW VHUYLQJ WR IDFLOLWDWH 1 UHPRYDO LQ SURPRWLQJ GLIIXVLRQ RI WKH 1 PROHFXOHV WRZDUG DYDLODEOH 0J2 LQWHULRU UHDFWLRQ VXUIDFHV ,QWULQVLF ILUVWRUGHU UDWH FRQVWDQWV IRU WKHVH GDWD DUH VKRZQ LQ 7DEOH (IIHFWV RI 2[\JHQ /LNH 12 UHPRYDO LQLWLDO 1 UHPRYDO HIILFLHQFLHV ZHUH LPSURYHG ZKHQ R[\JHQ ZDV SUHVHQW 6LQFH VRPH 12 SUHVHQW DV LPSXULW\ LQ WKH 1 JDV VXSSOLHG LV R[LGL]HG WR 1 WKH DFWXDO 1 UHPRYDO UDWH PD\ EH KLJKHU WKDQ REVHUYHG SDUWLFXODUO\ DW ORZHU WHPSHUDWXUHV ,W LV SRVVLEOH WKDW 12 ZKLFK LV SK\VLVRUEHG RQ WKH VRUEHQW VXUIDFH LV HYHQWXDOO\ R[LGL]HG WR 1 EHIRUH EHLQJ VRUEHG WR IRUP D QLWULWH RU QLWUDWH FRPSRXQG 6LQFH FRQFHQWUDWLRQV ZHUH DW WKH SHUFHQW OHYHO WKH\ DUH FRQVLGHUHG WR EH LQ JUHDW H[FHVV UHODWLYH WR 1

PAGE 113

7DEOH ,QWULQVLF ILUVWRUGHU UDWH FRHIILFLHQWV Vnf IRU 1 UHPRYDO E\ 0J2YHUPLFXOLWH VRUEHQW >2@ bf 1 &RQFHQWUDWLRQ 5HDFWLRQ 7HPS r .f SSP SSP SSP SSP PLQXWH PLQXWHV PLQXWHV PLQXWHV PLQXWHV PLQXWHV PLQXWHV PLQXWHV PLQXWHV R R

PAGE 114

FRQFHQWUDWLRQV )RU WKLV UHDVRQ WKH DSSDUHQW ILUVWRUGHU NLQHWLFV ZDV QRW DIIHFWHG E\ WKH SUHVHQFH RI LQ H[KDXVW JDVHV 5HSUHVHQWDWLYH SORWV RI 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH VLPLODU WR WKRVH VKRZQ IRU H[SHULPHQWDO UXQV ZKHUH R[\JHQ ZDV SUHVHQW DUH VKRZQ DW )LJXUHV WKURXJK ,Q )LJXUH DW D UHDFWLRQ WHPSHUDWXUH RI LQLWLDO 1 SHQHWUDWLRQ YDOXHV ZHUH OHVV WKDQ b IRU DOO WKUHH 1 FRQFHQWUDWLRQV HYDOXDWHG ,Q WKLV FDVH DQ LQOHW 1 FRQFHQWUDWLRQ RI SSP ZDV QRW HYDOXDWHG VLQFH WKH 1 F\OLQGHU FRQWDLQHG RQO\ SSP JDV DQG WKHUHIRUH GLOXWLRQ ZLWK DLU ZRXOG KDYH GHFUHDVHG WKLV FRQFHQWUDWLRQ 7KH VKDSHV RI WKH SHQHWUDWLRQ FXUYHV DUH VLPLODU WR WKRVH VHHQ LQ )LJXUH ZLWK 1 SHQHWUDWLRQ UHPDLQLQJ ORZ DQG UHPRYDO HIILFLHQF\ UHPDLQLQJ KLJKf IRU DQ LQOHW 1 FRQFHQWUDWLRQ RI SSP 2YHUDOO 1 SHQHWUDWLRQV IRU UXQV ZLWK R[\JHQ SUHVHQW UHPDLQHG ORZHU DW D UHDFWLRQ WHPSHUDWXUH RI ZKHQ FRPSDUHG WR UXQV FRQGXFWHG ZLWKRXW R[\JHQ 7KH SUHVHQFH RI R[\JHQ PD\ IDFLOLWDWH WKH FRQWLQXHG 1 UHPRYDO E\ WKH VRUEHQW 7KLV GLVWLQFWLRQ EHWZHHQ 1 SHQHWUDWLRQ LQ WKH SUHVHQFH DQG DEVHQFH RI GLG QRW DSSO\ IRU DOO 1 FRQFHQWUDWLRQV DW D UHDFWLRQ WHPSHUDWXUH RI )LJXUH f 1 SHQHWUDWLRQ DIWHU D EHG H[SRVXUH WLPH RI RQH KRXU ZDV DFWXDOO\ KLJKHU IRU DQ LQOHW 1 FRQFHQWUDWLRQ RI SSP ZKHQ R[\JHQ ZDV SUHVHQW FRPSDUHG WR ZKHQ R[\JHQ ZDV DEVHQW $W SSP 12A WKH SHQHWUDWLRQV ZHUH DSSUR[LPDWHO\ WKH VDPH DQG ZHUH OHVV DW SSP $W WKLV WHPSHUDWXUH .f R[\JHQ PD\ DLG LQ WKH UHPRYDO RI ORZ FRQFHQWUDWLRQV RI 1 $W D UHDFWLRQ WHPSHUDWXUH RI )LJXUH f WKH HIIHFW WUHQG ZDV UHYHUVHG 6LPLODU WR WKH HIIHFW VHHQ LQ )LJXUH DW WKH ORZHVW WHPSHUDWXUH HYDOXDWHG LW DSSHDUV WKDW D VXIILFLHQW JDVf§VROLG FRQFHQWUDWLRQ JUDGLHQW PXVW H[LVW WR SURYLGH WKH GULYLQJ IRUFH IRU UHDFWLRQ WR RFFXU

PAGE 115

SSP f§Vf§ SSP f§SSP )LJXUH 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf

PAGE 116

1 3HQHWUDWLRQ )UDFWLRQf )LJXUH 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf R X!

PAGE 117

1 3HQHWUDWLRQ )UDFWLRQf RL R %HG ([SRVXUH 7LPH PLQf SSP af§ SSP f§[B SSP )LJXUH 1 SHQHWUDWLRQ YV EHG H[SRVXUH WLPH 7 >2@ bf R

PAGE 118

:KLOH LQLWLDO 1 SHQHWUDWLRQ ZDV ORZHVW IRU DQ LQOHW 1 FRQFHQWUDWLRQ RI SSP LW UDSLGO\ LQFUHDVHG DQG UHDFKHG D YDOXH RI DSSUR[LPDWHO\ b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

PAGE 119

7DEOH ,QWULQVLF ILUVWRUGHU UDWH FRHIILFLHQWV Vnf IRU 1 UHPRYDO E\ 0J2f§YHUPLFXOLWH VRUEHQW >2@ bf 1 &RQFHQWUDWLRQ 5HDFWLRQ 7HPS r .f SSP SSP SSP PLQXWH PLQXWHV PLQXWHV PLQXWHV PLQXWHV PLQXWHV PLQXWHV PLQXWHV PLQXWHV

PAGE 120

VRUEHQW VXUIDFH )LJXUHV WKURXJK DUH W\SLFDO SORWV VLPLODU WR WKRVH SUHVHQWHG LQ WKH SUHYLRXV VHFWLRQV LQ ZKLFK WKH HIIHFW RI UHDFWLRQ WHPSHUDWXUH FDQ PRUH UHDGLO\ EH VHHQ 7KHVH ILJXUHV GHSLFW VRUEHQW SHUIRUPDQFH ZKHQ R[\JHQ ZDV QRW SUHVHQW LQ WKH V\VWHP 6LPLODU SORWV IRU WKRVH H[SHULPHQWV ZKHUH R[\JHQ ZDV SUHVHQW LQ WKH V\VWHP DUH VKRZQ DW )LJXUHV WKURXJK :KHQ WKH 1 LQOHW FRQFHQWUDWLRQ ZDV SSP ZLWKRXW R[\JHQf LQLWLDO 1 SHQHWUDWLRQ ZDV ORZHVW DW D EHG UHDFWLRQ WHPSHUDWXUH RI )LJXUH f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f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f )RU LQOHW 1 FRQFHQWUDWLRQV RI DQG SSP )LJXUHV DQG UHVSHFWLYHO\f LQLWLDO SHQHWUDWLRQV ZHUH ORZ IRU DOO WHPSHUDWXUHV EXW LW DSSHDUV WKDW D KLJKHU WHPSHUDWXUH ZDV

PAGE 121

1 3HQHWUDWLRQ )UDFWLRQf 7. 7 7 )LJXUH 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH >1@LQ SSP >2@ bf R RR

PAGE 122

7 7 E 7 )LJXUH 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH >12@LQ U SSP >2@ bf R

PAGE 123

1 3HQHWUDWLRQ )UDFWLRQf 7 7 E 7 )LJXUH 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH >12@LQ SSP >2@ bf

PAGE 124

7 7 aDa 7 )LJXUH 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH >12@ID SSP >2@ bf

PAGE 125

UHTXLUHG WR PDLQWDLQ D UHDVRQDEOH DPRXQW RI 1 UHPRYDO IURP WKH JDV VWUHDP :KHQ R[\JHQ ZDV SUHVHQW LQ WKH V\VWHP LW KDG D OHVV QRWLFHDEOH HIIHFW ZKHQ FRPSDULQJ WKH WHPSHUDWXUH HIIHFWV IRU D JLYHQ LQOHW 1 FRQFHQWUDWLRQ DV FDQ EH VHHQ LQ )LJXUHV WKURXJK $JDLQ LQLWLDO 1 SHQHWUDWLRQ ZDV ORZHVW DW WKH KLJKHVW WHPSHUDWXUH IRU DOO FRQFHQWUDWLRQV EXW LW LQFUHDVHG PRUH QRWLFHDEO\ IRU WKH KLJKHU FRQFHQWUDWLRQV )RU DQ LQOHW 1 FRQFHQWUDWLRQ RI SSP SHQHWUDWLRQ VWD\HG EHORZ b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f ,QLWLDO DFWLYDWLRQ HQHUJLHV ZHUH RIWHQ GLIILFXOW WR GHWHUPLQH GXH WR V\VWHP IOXFWXDWLRQV GXULQJ WKH ILUVW WR PLQXWHV RI RSHUDWLRQ 7KLV LQVWDELOLW\ SUREDEO\ FDXVHG VRPH RI WKH

PAGE 126

1 3HQHWUDWLRQ )UDFWLRQf 7 7 aEa 7 )LJXUH 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH >12@LQ SSP >2@ bf

PAGE 127

7 7 E 7 )LJXUH 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH >12@LQ SSP >2@ bf

PAGE 128

‘}f§ 7 7 D 7 )LJXUH 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH >12@LQ SSP >2@ bf

PAGE 129

,Q N ,Q Vf %HG ([SRVXUH 7LPH % PLQXWHV PLQXWHV aE PLQXWHV PLQXWHV )LJXUH $UUKHQLXV SORW RI ,Q N YHUVXV 7 7 EHWZHHQ DQG >12@P SSP >2@ bf R?

PAGE 130

7 .f %HG ([SRVXUH 7LPH A PLQXWHV PLQXWHV Ef§ PLQXWHV f§ PLQXWHV )LJXUH $UUKHQLXV SORW RI ,Q N YHUVXV 7 7 EHWZHHQ DQG >12@LQ SSP >2@ bf

PAGE 131

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f +DMDOLJRO HW DO f GHWHUPLQHG WKH DFWLYDWLRQ HQHUJ\ DVVRFLDWHG ZLWK WKH VXOIDWLRQ RI &D2 E\ 6WR EH DSSUR[LPDWHO\ N-JPRO 7KH DFWLYDWLRQ HQHUJ\ FDOFXODWHG IURP LQLWLDO UDWH GDWD IRU WKH VXOIDWLRQ RI &D2 E\ 0DUVK DQG 8OULFKVRQ f ZDV N-JPROH =DUNDQLWLV f GHWHUPLQHG DFWLYDWLRQ HQHUJLHV RI WKH UHDFWLRQ RI &D2 ZLWK 6 WR EH DSSUR[LPDWHO\ N-JPROH 7KH ODWWHU LQYHVWLJDWRU GHWHUPLQHG WKLV UHDFWLRQ WR EH OHVV VHQVLWLYH WR UHDFWLRQ WHPSHUDWXUH WKDQ WKH UHDFWLRQ RI 6 ZLWK FDOFLXP FDUERQDWH +H DOVR IRXQG WKDW WKH VXOIDWLRQ RI UHDJHQWJUDGH 0J2 SURFHHGV DW UDWHV FRPSDUDEOH WR WKRVH IRU &D2 6RWLUFKRV DQG =DUNDQLWLV f FRQILUPHG WKDW WKH DFWLYDWLRQ HQHUJ\ DVVRFLDWHG ZLWK WKH VXOIDWLRQ RI &D2 ZDV N-JPRO XVLQJ DQ 6 FRQFHQWUDWLRQ RI SSP DW UHDFWLRQ WHPSHUDWXUHV EHWZHHQ DQG & %RUJZDUGW f KDG SUHYLRXVO\ UHSRUWHG WKDW WKH

PAGE 132

DFWLYDWLRQ HQHUJ\ DVVRFLDWHG ZLWK WKH VXOIDWLRQ RI &D2 E\ D SSP 6 JDV VWUHDP DW WHPSHUDWXUHV EHWZHHQ DQG & UDQJHG IURP DSSUR[LPDWHO\ WR N-JPRO 5XL] $OVRS DQG 5RFKHOOH f HVWLPDWHG WKH DFWLYDWLRQ HQHUJ\ RI WKH UHDFWLRQ EHWZHHQ 6DQG &D2+f WR EH DSSUR[LPDWHO\ N-JPRO 2EYLRXVO\ WKHUH FDQ EH VLJQLILFDQW YDULDWLRQV LQ WKH PHDVXUHPHQW RI DFWLYDWLRQ HQHUJLHV GHSHQGHQW XSRQ H[SHULPHQWDO FRQGLWLRQV 2WKHU VWXGLHV RI DFLG JDVVROLG UHDFWLRQV KDYH SURGXFHG UHVXOWV YDU\LQJ E\ D VLPLODU PDJQLWXGH 12! SSPf WKHVH HQHUJLHV ZHUH UHGXFHG WR ]HUR DQG VRPHWLPHV EHFDPH QHJDWLYH DIWHU WR PLQXWHV 7KLV ILQGLQJ ZRXOG VXSSRUW D PHFKDQLVWLF WUDQVLWLRQ IURP FKHPLFDO NLQHWLF FRQWURO WR GLIIXVLRQ FRQWURO WKURXJK D SURGXFW OD\HU RU LQWUDSDUWLFOH GLIIXVLRQ FRQWURO ,W LV QRW NQRZQ ZKHWKHU WKLV LV D IXQFWLRQ RI LQFUHDVLQJ SURGXFW OD\HU WKLFNQHVV DQG

PAGE 133

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f ZKLFK LV WKHUPDOO\ VWDEOH RYHU WKH UDQJH RI WHPSHUDWXUHV HYDOXDWHG $GGLWLRQDOO\ VLQFH 1 LV VRPHZKDW VROXEOH LQ ZDWHU D VPDOO TXDQWLW\ PD\ EH UHPRYHG LQ WKLV ZD\ 7KLV PD\ FDXVH WKH IRUPDWLRQ RI QLWURXV DQGRU QLWULF DFLGV ZKLFK DUH SUREDEO\ PRUH UHDFWLYH WKDQ 1 DQG VKRXOG EH LPPHGLDWHO\ QHXWUDOL]HG E\ WKH EDVLF VRUEHQW SUHYHQWLQJ DQ\ SRWHQWLDO IRU FRUURVLRQ SUREOHPV ,QWULQVLF ILUVWRUGHU UDWH FRHIILFLHQWV DQG 1 UHPRYDO UDWHVf IRU WKHVH H[SHULPHQWV ZHUH JHQHUDOO\ FRPSDUDEOH LQ WKH HDUO\ VWDJHV RI UHDFWLRQ KRZHYHU 1 UHPRYDO HIILFLHQF\ UHPDLQHG KLJKHU IRU DQ H[WHQGHG SHULRG RI WLPH ZKHQ ZDWHU ZDV SUHVHQW LQ WKH V\VWHP )LJXUH LV D W\SLFDO H[DPSOH RI DQ 1 FRQFHQWUDWLRQ YHUVXV EHG H[SRVXUH WLPH SURILOH XWLOL]LQJ PRLVWHQHG VRUEHQW 7KH SRVVLEOH UHDVRQV IRU WKLV HIIHFW LQFOXGH WKH LQFUHDVHG UHDFWLYLW\ RI 0J2+f FRPSDUHG WR 0J2 WKH VROXELOLW\ RI 1 LQ ZDWHU DQG WKH ORZHU PRODU YROXPH UDWLR RI 0J1f WR 0J2+f FRPSDUHG WR WKH PRODU YROXPH UDWLR RI 0J1f WR 0J2 1 VROXELOLW\ ZDV DYRLGHG LQ H[SHULPHQWV ZKHUH WKH VRUEHQW ZDV ZHWWHG SULRU WR H[SRVXUH WR IORZLQJ JDVHV WR FRQYHUW WKH 0J2 RQ WKH VXUIDFH LQWR 0J2+f OHDYLQJ OLWWOH IUHH ZDWHU

PAGE 134

1 3HQHWUDWLRQ )UDFWLRQf %HG ([SRVXUH 7LPH PLQf V 7 )LJXUH 1 SHQHWUDWLRQ YHUVXV EHG H[SRVXUH WLPH PO + DGGHG WR J 0J2YHUPLFXOLWH VRUEHQW >12@LQ SSP >2@ b 7 .f

PAGE 135

:KHQ JDV KXPLGLILFDWLRQ ZDV XVHG VLQFH 1 FRQFHQWUDWLRQV ZHUH PRQLWRUHG GLUHFWO\ DW WKH EHG LQOHW FKHPLFDO NLQHWLFV ZDV QRW D IXQFWLRQ RI 1 DEVRUEHG LQ WKH KXPLGLILFDWLRQ YHVVHO 6LQFH WKH PRODU YROXPH UDWLR LV VPDOOHU ZKHQ VWDUWLQJ ZLWK WKH K\GUR[LGH PRUH 1 FDQ EH UHPRYHG EHIRUH WKH H[WHULRU VXUIDFH EHFRPHV FRYHUHG DQG 0J2+f LV QR ORQJHU UHDGLO\ DYDLODEOH IRU UHDFWLRQ 7KLV PD\ PHDQ WKDW LW WDNHV ORQJHU WR IRUP D FRPSOHWH LQHUW SURGXFW OD\HU VKHOO ZKHQ PRLVWXUH LV DYDLODEOH LQ WKH V\VWHP $IWHU WKLV OD\HU LV IRUPHG KRZHYHU WKH W\SLFDO VKULQNLQJ FRUH EHKDYLRU ZLWK SURGXFW OD\HU GLIIXVLRQ FRQWURO ZRXOG VWLOO EH H[SHFWHG (IIHFWV RI 5HVLGHQFH 5HDFWLRQf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b 7KLV VLWXDWLRQ GLG QRW DSSHDU WR EH JDVILOPPDVVWUDQVIHUOLPLWHG DV DOUHDG\ GLVFXVVHG $GGLWLRQDOO\ VLQFH ORQJHU UHVLGHQFH WLPHV FRUUHVSRQG WR ORZHU JDV YHORFLWLHV PHDVXUHG EHG SUHVVXUH GURSV ZHUH ORZHU ZKLFK LV SUHIHUDEOH IRU -(7& DSSOLFDWLRQV 3UHVVXUH GURS UHVXOWV ZLOO EH SUHVHQWHG ODWHU

PAGE 136

6RUEWHFK6XSSOLHG 6DPSOH 5HVXOWV ([SHULPHQWDO UHVXOWV IRU VDPSOHV RI VRUEHQW SURYLGHG E\ 6RUEWHFK ZHUH JHQHUDOO\ FRPSDUDEOH WR WKRVH IRU ODERUDWRU\SUHSDUHG VRUEHQW ,QLWLDO UHPRYDO UDWHV ZHUH JUHDWHU WKDQ b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b PRUH WKDQ WKH ODERUDWRU\ VRUEHQW ,W ZDV WKRXJKW WKDW WKH KLJKHU PDVV PLJKW EH WKH UHVXOW RI WKH SUHVHQFH RI ZDWHU LQ WKH 6RUEWHFK VRUEHQW ZKLFK ZRXOG DOVR KHOS WR H[SODLQ WKH LPSURYHG 1 UHPRYDO HIILFLHQF\ ZLWK WLPH 3UHZHLJKHG VDPSOHV RI ERWK VRUEHQWV ZHUH KHDWHG LQ D PXIIOH IXUQDFH WR D WHPSHUDWXUH RI IRU RQH KRXU WR HYDOXDWH ZHLJKW ORVV IURP GHK\GUDWLRQ 7KLV SURYHG WR EH LQVLJQLILFDQW ZLWK WKH

PAGE 137

ODERUDWRU\ VDPSOHV ORVLQJ DSSUR[LPDWHO\ WZR SHUFHQW PDVV DQG WKH 6RUEWHFK VDPSOHV ORVLQJ DERXW WKUHH SHUFHQW 7KH KLJKHU GHQVLW\ PDWHULDO GLG QRW SURGXFH D KLJKHU VXUIDFH DUHD RQ D SHU EHG EDVLV EHFDXVH WKHUH ZDV OLWWOH GLIIHUHQFH EHWZHHQ WKH VXUIDFH DUHDV RI ODERUDWRU\ SUHSDUHG VDPSOHV DQG 6RUEWHFK VDPSOHV ,Q IDFW ODERUDWRU\ SUHSDUHG VDPSOHV KDG RQ DYHUDJH b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f 0 A F FS G A V S / 2S f

PAGE 138

ZKHUH -3 SUHVVXUH GURS / EHG GHSWK IY WKH YRLG IUDFWLRQ MX JDV YLVFRVLW\ S D VKDSH IDFWRU DEVRUEHQW VXUIDFH DUHD PHDQ SDUWLFOH GLDPHWHU PDVV PHGLDQ GLDPHWHUf S\ JDV GHQVLW\ DQG e VXSHUILFLDO JDV YHORFLW\ DOO LQ DSSURSULDWH DEVROXWH XQLWV ,W FDQ EH VHHQ IURP HTXDWLRQ f WKDW DW ORZ IOXLG YHORFLWLHV WKH ILUVW YLVFRXVf WHUP SUHGRPLQDWHV DQG DW KLJK JDV YHORFLWLHV WKDW WKH VHFRQG LQHUWLDOf WHUP LV GRPLQDQW DQG WKH SUHVVXUH GURS LV SURSRUWLRQDO WR WKH VTXDUH RI JDV YHORFLW\ 7KH VKDSH IDFWRU WHUP SV LV GHILQHG DV WKH UDWLR RI WKH VXUIDFH DUHD RI D VSKHUH RI HTXDO YROXPH WR WKH SDUWLFOH WR WKH VXUIDFH DUHD RI WKH SDUWLFOH 7KH YDOXH RI WKLV WHUP LV REYLRXVO\ IRU VSKHULFDO SDUWLFOHV DQG QRUPDOO\ OLHV LQ WKH UDQJH IURP WR IRU QDWXUDOO\ RFFXUULQJ PDWHULDOV 7KH PHGLDQ VRUEHQW SDUWLFOH GLDPHWHU ZDV DSSUR[LPDWHO\ PP 7KH YDOXH IRU eY ZDV GHWHUPLQHG E\ WKH PHFKDQLFDO FRPSUHVVLRQ RI WKH VRUEHQW WR D SRLQW ZKHUH JDV ZRXOG QRW IORZ WKURXJK WKH EHG UHGXFLQJ WKH YROXPH E\ DSSUR[LPDWHO\ SHUFHQW IY f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

PAGE 139

3UHVVXUH 'URS 3Df 6XSHUILFLDO 9HORFLW\ PVf 3UHGLFWHG ([SHULPHQWDO )LJXUH 3UHGLFWHG (UJXQ HTXDWLRQf YHUVXV H[SHULPHQWDO SUHVVXUH GURS DV D IXQFWLRQ RI VXSHUILFLDO YHORFLW\ IRU D P ORQJ EHG RI 0J2f§YHUPLFXOLWH VRUEHQW 7 .f

PAGE 140

0DJQHVLXP 1LWUDWH 6XUIDFH 'HFRPSRVLWLRQ 6ROLG 0J1fr+ &HUWLILHG $&6 JUDGH )LVKHU 6FLHQWLILF 3LWWVEXUJK 3HQQV\OYDQLDf ZDV GLVVROYHG LQ GHLRQL]HG ZDWHU DQG DSSOLHG WR KHDWWUHDWHG YHUPLFXOLWH DQG WKH PL[WXUH ZDV DOORZHG WR DLU GU\ 7KH PDVV UDWLR DSSOLHG ZDV DSSUR[LPDWHO\ b 0J1f 7KLV PDWHULDO ZDV SDFNHG LQVLGH D FOHDQ UHDFWRU DQG V/SP GU\ 1 ZDV GHOLYHUHG WR WKH EHG %HG RXWOHW FRQFHQWUDWLRQV RI 12 DQG 1 ZHUH IROORZHG RYHU WLPH DV EHG WHPSHUDWXUH ZDV LQFUHDVHG IURP DPELHQW WHPSHUDWXUH WR WKURXJK LQ r. VWHSV 5HVXOWV DUH SORWWHG LQ )LJXUH )URP WKLV ILJXUH LW FDQ EH VHHQ WKDW DW WHPSHUDWXUHV XVHG IRU VRUSWLRQ OLWWOH 12 DQG 1 ZHUH UHOHDVHG FRPSDUHG WR WKRVH HPLVVLRQV REVHUYHG DW PXFK KLJKHU WHPSHUDWXUHV DQG HVSHFLDOO\ ,Q HLWKHU FDVH WKH UDWLR RI112 UHOHDVHG ZDV DSSUR[LPDWHO\ EHWZHHQ DQG +HDWLQJ VDPSOHV RI XVHG VRUEHQW WR D WHPSHUDWXUH RI SURGXFHG VLPLODU UHVXOWV 6HH )LJXUH f 7KLV LPSOLHV WKDW 1 LV WKH SUHGRPLQDQW JDVHRXV UHDFWDQW LQ WKH IRUZDUG UHDFWLRQ $OVR LW LV LPSRUWDQW WR UHFRJQL]H WKDW WKH GHFRPSRVLWLRQ RI 0J1f LV QRW D VLJQLILFDQW VRXUFH RI 12 $OWKRXJK LW LV REYLRXVO\ QHFHVVDU\ WR FRQWURO 12 HPLVVLRQV WKHVH VWXGLHV LPSO\ WKDW 12 SURGXFWLRQ REVHUYHG LQ 1 VRUEHQW UHPRYDO VWXGLHV PXVW RFFXU WKURXJK D PHFKDQLVP RWKHU WKDQ GHFRPSRVLWLRQ RI 0J1f

PAGE 141

f§f§ >12 @ >12@ )LJXUH 7KHUPDO GHFRPSRVLWLRQ RI 0J1f f + RQ YHUPLFXOLWH WHPSHUDWXUH LQFUHDVLQJ ZLWK WLPHf WR

PAGE 142

>12 @ >12@ )LJXUH 7KHUPDO GHFRPSRVLWLRQ RI XVHG 0J2f§YHUPLFXOLWH VRUEHQW 7 .f WR 92

PAGE 143

6RUEHQW /LIHWLPH DQG 5HJHQHUDWLRQ :KHUHDV LQLWLDO VRUEHQW 1 UHPRYDO UDWHV ZHUH RIWHQ JUHDWHU WKDQ SHUFHQW IUDFWLRQDO 1 UHPRYDO GHFUHDVHG ZKLOH 1 FRQFHQWUDWLRQV OHDYLQJ WKH EHG LQFUHDVHG ZLWK EHG H[SRVXUH WLPH )LJXUH f ,Q WKLV H[DPSOH G\QDPLF HTXLOLEULXP ZDV DSSURDFKHG DIWHU DSSUR[LPDWHO\ PLQXWHV RI EHG H[SRVXUH $W DSSUR[LPDWHO\ WKH VDPH WLPH 12 RXWOHW FRQFHQWUDWLRQV LQFUHDVHG WR YDOXHV JUHDWHU WKDQ 12 LQOHW FRQFHQWUDWLRQ 7KH 12 SURGXFHG ZDV PXFK JUHDWHU WKDQ WKDW ZKLFK ZRXOG UHVXOW IURP QLWUDWH GHFRPSRVLWLRQ DV SUHYLRXVO\ GHVFULEHG $OVR 12 SURGXFWLRQ WKURXJK WKHUPDO 1 GHFRPSRVLWLRQ DW WHPSHUDWXUHV EHORZ LV PLQLPDO 7KHVH REVHUYDWLRQV LQGLFDWH WKDW WKHUH PXVW EH DQRWKHU PHFKDQLVP IRU 12 SURGXFWLRQ $ FRQVHFXWLYH PXOWLVWHS FKHPLFDO UHDFWLRQ RI WZR RU PRUH PROHFXOHV RI 1 ZLWK 0J2 OHDGLQJ WR WKH EHG VXUIDFH EHFRPLQJ VDWXUDWHG ZLWK 0J1f PD\ EH SURSRVHG 2QH VWHS ZKLFK LW LV FRQYHQLHQW WR SURSRVH DV WKH ODVWf PXVW EH DQ R[LGDWLRQ SURFHVV LQ ZKLFK WKH HTXLYDOHQW RI DQ R[\JHQ DWRP LV GHOLYHUHG WR VRPH LQWHUPHGLDWH SURGXFW KDYLQJ WKH HPSLULFDO IRUPXOD 0J1 0J2 1f 2Q D PDVVUDWLR EDVLV WKLV SURFHVV DSSHDUV WR RFFXU RQ DYHUDJH ZKHQ DSSUR[LPDWHO\ PJ 1 SHU JUDP RI 0J2 DUH VRUEHG :KLOH WKLV FRUUHVSRQGV WR PLQLPDO XWLOL]DWLRQ RI WKH EHG VLJQLILFDQW FDSDFLW\ IRU 1 UHPRYDO SHUVLVWV ORQJ DIWHU LQLWLDWLRQ RI WKLV ILQDO R[LGDWLRQ SURFHVV DV 1 UHPRYDO HIILFLHQFLHV RI b FDQ FRQWLQXH IRU KRXUV ,W PD\ EH WKDW 0J2 RU 0J2+f DW WKH H[WHUQDO SDUWLFOH VXUIDFH LV FRYHUHG E\ D QLWUDWHG PRQROD\HU VR RQO\ LQWHUQDO R[LGH RU K\GUR[LGH LV DYDLODEOH IRU UHDFWLRQ

PAGE 144

f§ >1@ >12@ )LJXUH &RPSDUDWLYH EHG RXWOHW 1 DQG 12 FRQFHQWUDWLRQV YHUVXV EHG H[SRVXUH WLPH VKRZLQJ 12 SURGXFWLRQ

PAGE 145

,W LV DOVR SRVVLEOH WKDW WKH WUDQVIRUPHG VXUIDFH FDWDO\]HV WKH FRQWLQXHG R[LGDWLRQ RU UHGXFWLRQf RI 1 $ PRUH GHVFULSWLYH GLVFXVVLRQ RI VRUEHQW OLIHWLPH ZLWKLQ WKH FRQWH[W RI WKH VKULQNLQJ XQUHDFWHG FRUH PRGHO ZLOO EH SUHVHQWHG ODWHU %HGV UHJHQHUDWHG WKURXJK KHDWLQJ WR DSSUR[LPDWHO\ FDQ EH UHXVHG ZLWK OLWWOH ORVV RI SHUIRUPDQFH 2EYLRXVO\ UHJHQHUDWLQJ VROHO\ WKURXJK KHDWLQJ VHUYHV WR UHOHDVH 12[ IURP WKH VROLG VXUIDFH LQWR WKH JDV SKDVH 7KLV PLJKW GHIHDW WKH SXUSRVH RI XVLQJ WKH VRUEHQW WR FRQWURO 12[ ,QWURGXFWLRQ RI D UHGXFLQJ JDV &2 DSSHDUV WR EH WKH UHGXFWDQW RI FKRLFHf ZKHQ UHJHQHUDWLQJ WKH VRUEHQW FDQ DOOHYLDWH WKLV SUREOHP 1HOVRQ HW DO f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f $ VLPSOLILHG PHFKDQLVP LV SURSRVHG DV IROORZV f 1 0J2 0J1 f 1 0J1 0J14

PAGE 146

f 1 0J1 0J1f 12 ,W VKRXOG EH QRWHG WKDW 1 LV IXQFWLRQLQJ DV DQ R[LGL]LQJ DJHQW LQ WKH ILQDO VWHS WR FRPSOHWH WKH IRUPDWLRQ RI 0J1f LQ WKH DEVHQFH RI R[\JHQ ,I R[\JHQ ZHUH SUHVHQW WKHQ LW PD\ FRPSHWLWLYHO\ FRPSOHWH WKLV ODVW VWHS UHGXFLQJ 12 IRUPDWLRQ :KHQ 0J2+f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f§VROLG UHDFWLRQV ,QFRUSRUDWLQJ VXFK SK\VLFDO SDUDPHWHUV DV WKH PRODU YROXPH UDWLR RI SURGXFWV WR UHDFWDQWV ZKLFK FDQ DIIHFW SURGXFW OD\HU IRUPDWLRQ DQG

PAGE 147

VXEVHTXHQW SURGXFW OD\HU GLYLVLRQDO OLPLWDWLRQ RI UHDFWLRQ UDWHf PD\ IXUWKHU LPSURYH WKH XWLOLW\ L RI WKH PRGHO WR GHVFULELQJ UHDFWLRQ PHFKDQLVPVf ,W LV DSSURSULDWH QRZ WR H[DPLQH WKH HTXDWLRQV IRU WKH WKUHH SRWHQWLDO UDWHOLPLWLQJ UHVLVWDQFHV DIIHFWLQJ KHWHURJHQHRXV JDVVROLG UHDFWLRQV 7KH GHWDLOHG GHULYDWLRQV FDQ EH IRXQG LQ PRVW VWDQGDUG WH[WV RQ JDVVROLG UHDFWLRQV 7KH IROORZLQJ LV EDVHG RQ WKH GHVFULSWLRQV UHSRUWHG E\ 6]HNHO\ HW DO f DQG /HYHQVSLHO f 8VLQJ WKH JHQHUDOL]HG JDVVROLG UHDFWLRQ UHSUHVHQWHG E\ HTXDWLRQ f D$ JDVf E% VROLGf F& VROLGf G' JDVf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t6 >$@A >$@J >$@-@ LV D FRQVWDQW KDYLQJ WKH YDOXH >$@J ,W LV FRQYHQLHQW WR GHULYH NLQHWLF H[SUHVVLRQV IRU JDVf§VROLG UHDFWLRQV EDVHG RQ WKH H[WHUQDO VXUIDFH DUHD RI WKH VRUEHQW SDUWLFOH 6 N5 )URP HTXDWLRQ ZH NQRZ WKDW DG1%f EG1$f ZKHUH 1$ DQG 1% UHSUHVHQW WKH QXPEHU RI PROHV RI WKH UHDFWDQWV 7KHUHIRUH ZH FDQ ZULWH

PAGE 148

6WDJQDQW *DV )LOP )LJXUH *UDSKLFDO UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO XQGHU GLXVLRQWKURXJKJDVILOP FRQWURO X!

PAGE 149

D D G1E 6 GW Q5 GW E G1$ Q5 GW ENVPI>$?f ENV>$@J f LQ ZKLFK NJ LV WKH PDVV WUDQVIHU FRHIILFLHQW EHWZHHQ WKH JDV SKDVH DQG WKH VRUEHQW SDUWLFOH /HWWLQJ S% UHSUHVHQW WKH PRODU GHQVLW\ RI % LQ WKH VROLG DQG )EH WKH SDUWLFOH YROXPH WKH QXPEHU RI PROHV RI % SUHVHQW LQ WKH SDUWLFOH LV 1f S%U f 7KH GHFUHDVH LQ WKH YROXPH RU UDGLXV RI WKH VKULQNLQJ XQUHDFWHG FRUH DV D UHVXOW RI WKH UHDFWLRQ RI DG1% PROHV RI VROLG DQG RU ZLWKf EG1$ PROHV RI JDV LV JLYHQ E\ WKH HTXDWLRQ DG1% EG1$ DS%G9 DS%G QUf D?QS-I@GUF f 6XEVWLWXWLQJ HTXDWLRQ LQ HTXDWLRQ SURGXFHV DQ H[SUHVVLRQ IRU WKH UDWH RI UHDFWLRQ LQ WHUPV RI WKH VKULQNLQJ UDGLXV RU XQUHDFWHG FRUH D G1% 6 GW D"F GU F 5 GW EN0K f %\ UHDUUDQJLQJ DQG LQWHJUDWLQJ DQ H[SUHVVLRQ IRU KRZ WKH XQUHDFWHG FRUH VKULQNV ZLWK WLPH LV GHULYHG f

PAGE 150

,QWHJUDWLRQ DQG UHDUUDQJHPHQW LVRODWHV WKH WLPH RI UHDFWLRQ W f§n; >OAf-@ f WNM$@W 5 7KHUHIRUH XQGHU JDVILOPGLIILLVLRQ FRQWURO UF DQG ZH ZLOO GHILQH UI DV WKH WLPH IRU FRPSOHWH UHDFWLRQ RI WKH SDUWLFOH D3%5 n 8-$n f :H FDQ UHZULWH WKLV HTXDWLRQ LQ WHUPV RI IUDFWLRQDO VROLG FRQYHUVLRQ ;f E\ DOORZLQJ ; YRAXPH XQUHDFWHG FRUH WRWDO SDUWLFOH YROXPH f LW U Lf 7L5 5 f :KHQFH ZH FDQ GHILQH WKH GHSHQGHQFH RI ; XSRQ W W I f§ Lf ; 9 5 f 3ORWWLQJ VROLG FRQYHUVLRQ YHUVXV EHG H[SRVXUH RU UHDFWLRQ WLPH DV SUHYLRXVO\ GHVFULEHG VKRXOG UHYHDO D OLQHDU UHODWLRQVKLS KDYLQJ D VORSH RI US LI JDVILOPGLIILLVLRQ LV UDWHOLPLWLQJ &KHPLFDO5HDFWLRQ &RQWURO )URP )LJXUH LW FDQ EH VHHQ WKDW WKH FRQFHQWUDWLRQ JUDGLHQW DQG WKH UHODWHG SURJUHVV RI UHDFWLRQ DUH XQDIIHFWHG E\ WKH SUHVHQFH RI DQ\ LQHUW OD\HUVf LQ WKLV VLWXDWLRQ 7KH

PAGE 151

6WDJQDQW *DV )LOP )LJXUH *UDSKLFDO UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO XQGHU FKHPLFDOUHDFWLRQ FRQWURO X!

PAGE 152

TXDQWLW\ RI PDWHULDO UHDFWLQJ LV SURSRUWLRQDO WR WKH DYDLODEOH VXUIDFH DUHD RI WKH XQUHDFWHG FRUH ZKLFK DW WKH EHJLQQLQJ RI WKH UHDFWLRQ ZRXOG EH WKH H[WHULRU VXUIDFH RI WKH SDUWLFOH DV WKHUH LV QR LQHUW SURGXFW OD\HU IRUPHGf DQG FDQ EH H[SUHVVHG LQ WKH VDPH PDQQHU DV HTXDWLRQ D WL U GW & E L U@ t ENV>$@J f +HUH NV UHSUHVHQWV WKH LQWULQVLF ILUVWRUGHU UDWH FRHIILFLHQW IRU WKH VXUIDFH UHDFWLRQ FPVf :ULWLQJ 1E LQ WHUPV RI WKH VKULQNLQJ UDGLXV E\ DQDORJ\ WR HTXDWLRQ D ‘@ aGU F W UF f GW ENW>$? f 7KLV PD\ EH UHDUUDQJHG DQG LQWHJUDWHG WR GHULYH DQ H[SUHVVLRQ GHVFULELQJ KRZ WKH XQUHDFWHG FRUH VKULQNV ZLWK WLPH DS% IrF GUF ENV>$@J GW f $IWHU LQWHJUDWLRQ W PD\ EH LVRODWHG E\ UHDUUDQJHPHQW W % EN-$@ &5Uf f 7KH WLPH IRU FRPSOHWH FRQYHUVLRQ RI WKH VROLG XQGHU FKHPLFDO UHDFWLRQ FRQWURO GHILQHG DV UQ RFFXUV ZKHQ U U D3%5 06?$? ; f

PAGE 153

$Q H[SUHVVLRQ IRU WKH GHFUHDVH LQ WKH UDGLXV RI WKH VKULQNLQJ FRUH RU WKH LQFUHDVH LQ IUDFWLRQDO FRQYHUVLRQ RI WKH VROLG SDUWLFOH VXUIDFH WKHQ IROORZV LQFOXGLQJ D VXEVWLWXWLRQ IURP HTXDWLRQ r‘rfn n n 5 n ‘ f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f VR D VWHDG\VWDWH DVVXPSWLRQ UHJDUGLQJ WKH GLPHQVLRQ RI WKH XQUHDFWHG FRUH LV YDOLG 7KLV DVVXPSWLRQ JUHDWO\ VLPSOLILHV WKH PDWKHPDWLFV EHFDXVH LW FDQ QRZ EH DVVXPHG WKDW WKH UDWH RI UHDFWLRQ RI $ LV HTXDO WR LWV UDWH RI GLIIXVLRQ WR WKH VXUIDFH RI WKH XQUHDFWHG FRUH WLU4$ $b5 4A WLUA4$& &RQVWDQW GW f

PAGE 154

6WDJQDQW *DV )LOP )LJXUH *UDSKLFDO UHSUHVHQWDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO XQGHU GLIILLVLRQWKURXJKLQHUWSURGXFWOD\HU FRQWURO

PAGE 155

7KH IOX[ RI $ ZLWKLQ WKH LQHUW OD\HU LV H[SUHVVHG E\ )LFNnV ODZ IRU HTXLPRODU FRXQWHUGLIOLLVLRQ 1RWH WKDW $ LV SRVLWLYHf 4D 'A f GU ZKHUH 'H UHSUHVHQWV WKH HIIHFWLYH GLIIXVLRQ FRHIILFLHQW RI WKH JDVHRXV UHDFWDQW WKURXJK WKH SURGXFW OD\HU FPVf 2ZLQJ WR YDULDWLRQV LQ WKH SURSHUWLHV RI WKH LQHUW OD\HU WKH YDOXH RI WKLV FRHIILFLHQW LV GLIILFXOW WR HVWLPDWH DQG LV JHQHUDOO\ GHWHUPLQHG H[SHULPHQWDOO\ 7KH HIIHFWLYH GLIIXVLRQ FRHIILFLHQW FDQ EH REWDLQHG E\ DGMXVWLQJ 'H LQ WKH VKULQNLQJ XQUHDFWHGFRUH RU JUDLQ PRGHO XQWLO WKH PRGHO SUHGLFWLRQV ILW H[SHULPHQWDO GDWD 0DUVK DQG 8OULFKVRQ f (TXDWLRQV DQG PD\ EH FRPELQHG f =! P &f§ } GW H GW ,QWHJUDWLQJ WKURXJK WKH LQHUW OD\HU Inf§ UF' I>$9 rG>$@ f! GW -U U H ->$@J$@V WKHUHIRUH G1 $ GW K Q'-$O U 5 f )RU D VKULQNLQJ XQUHDFWHGFRUH RI JLYHQ VL]H G1GW LV FRQVWDQW EXW WKH LQHUW OD\HU EHFRPHV WKLFNHU DV WKH FRUH VKULQNV ZKLFK GHFUHDVHV WKH GLIIXVLRQ UDWH RI $ WKURXJK WKH OD\HU

PAGE 156

%HIRUH LQWHJUDWLQJ HTXDWLRQ RQH RI WKH YDULDEOHV PXVW EH HOLPLQDWHG $V LQ JDVILOP GLIIXVLRQ 1D FDQ EH HOLPLQDWHG E\ ZULWLQJ LW LQ WHUPV RI UF 6XEVWLWXWLQJ DJDLQ IURP HTXDWLRQ DQG UHDUUDQJLQJ SURGXFHV DQ LQWHJUDEOH IRUP DS% fUn :FGUH E'W>$@ I fGW f -U5 U 5 -R )URP WKLV W E'n>$@J UF f! > If 5 A 5 f 'HILQLQJ UG DV WKH WLPH WR FRPSOHWH FRQYHUVLRQ U f RI WKH VROLG SDUWLFOH XQGHU SURGXFW OD\HUGLIIXVLRQ FRQWURO r3E. ;M E'H>$@J f 7KLV OHDGV WR DQ H[SUHVVLRQ IRU WKH SURJUHVV RI WKH UHDFWLRQ LQ WHUPV RI UG? f§ f§I r f§f [ 5 5 f 7KLV PD\ DOVR EH H[SUHVVHG LQ WHUPV RI IUDFWLRQDO VROLG FRQYHUVLRQ IURP HTXDWLRQ S;f f§ =fP ;f [ f

PAGE 157

:KHQ SURGXFWOD\HU GLIIXVLRQ LV UDWHFRQWUROOLQJ D SORW RI WKH ULJKWKDQGVLGH RI HTXDWLRQ YHUVXV EHG H[SRVXUH WLPH DOORZV IRU WKH GHWHUPLQDWLRQ RI UG 'HULYDWLRQV RI WKH 6KULQNLQJ 8QUHDFWHG&RUH 0RGHO $V UHDFWLRQ SURFHHGV DQG D SURGXFW OD\HU LV IRUPHG WKH GHFUHDVLQJ SURGXFW OD\HU GLIIXVLRQ UDWH FDQ DIIHFW WKH RYHUDOO UDWH (TXDWLRQ DVVXPHV D FRQVWDQW YDOXH RI 'H ZKLFK PD\ QRW EH D YDOLG DVVXPSWLRQ 9DULDWLRQ RI WKH GHFUHDVLQJf GLIIXVLRQ FRHIILFLHQW LQ WKH SURGXFW OD\HU PD\ KDYH WR EH WDNHQ LQWR DFFRXQW WR GHVFULEH D UHDFWLRQ LQ ZKLFK D PRUH YROXPLQRXV VROLG SURGXFW IRUPV .ULVKQDQ DQG 6RWLUFKRV f 0RUHHODERUDWH PRGHOV FDQ EH FRQVWUXFWHG DOORZLQJ IRU FKDQJHV LQ WKH FRQWUROOLQJ PHFKDQLVP ZLWK WLPH RU VROLGVXUIDFH FRQYHUVLRQ 2IWHQ D VLPSOH PRGLILFDWLRQ WR WKH EDVLF HTXDWLRQV DOORZV IRU DQ LPSURYHG UHSUHVHQWDWLRQ RI WKH H[SHULPHQWDO GDWD :KHQ SURGXFWOD\HUGLIIXVLRQ LV WUXO\ WKH UDWH OLPLWLQJ SURFHVV WKH LQFOXVLRQ RI WKH PRODU YROXPH UDWLR UHSUHVHQWHG E\ =f LQ HTXDWLRQ DV IROORZV PD\ LPSURYH WKH ILW 6]HNHO\ HW DO f -B >= = =fO ;ff@r B A\ f = 7KLV LV HVSHFLDOO\ WUXH ZKHQ WKH SURGXFW OD\HU PRODU YROXPH LV VLJQLILFDQWO\ ODUJHU RU VPDOOHUf WKDQ WKDW RI WKH VROLG UHDFWDQW 5HIHUULQJ EDFN WR 7DEOH IRXU PRODU YROXPH UDWLRV DUH RI LQWHUHVW LQ WKLV UHVHDUFK

PAGE 158

f 0J1f+ 0J2 f 0J1f+ 0J2 f 0J1f+ 0J2+f f 0J1f+ 0J2+f $V DOO RI WKHVH YDOXHV DUH QRWLFHDEO\ JUHDWHU WKDQ RQH WKH XVH RI WKH =WHUP LV H[SHFWHG WR DLG LQ WKH LQWHUSUHWDWLRQ RI WKH H[SHULPHQWDO GDWD FROOHFWHG LQ WKLV VWXG\ 6]HNHO\ HW DO f DOVR SRLQWHG RXW WKDW WKH XVH RI RQH RU WKH RWKHU RI WKH JRYHUQLQJ HTXDWLRQV PD\ QRW EH VXIILFLHQW WR EHVW GHVFULEH WKH H[SHULPHQWDO GDWD 6LQFH WKH WZR HTXDWLRQV DUH VLPLODU LW LV RIWHQ GLIILFXOW WR GLVWLQJXLVK EHWZHHQ FKHPLFDOUHDFWLRQ FRQWURO DQG SURGXFWOD\HUGLIIXVLRQ FRQWURO ,Q VRPH FDVHV DVVXPLQJ JDVILOP GLIIXVLRQ LV QRW UDWH OLPLWLQJ FKHPLFDOUHDFWLRQ DQG SURGXFWOD\HUGLIIXVLRQ PD\ FROOHFWLYHO\ FRQWURO WKH UHDFWLRQ UDWH ,Q VXFK FDVHV D WHUP Rf UHSUHVHQWLQJ WKH UDWLR RI GLIIXVLRQDO UHVLVWDQFH WR FKHPLFDO UHDFWLRQ UHVLVWDQFH FDQ EH GHILQHG WR DSSO\ WKH UHODWLYH LPSRUWDQFH RI HDFK UHVLVWDQFH WR WKH UHDFWLRQ UDWH ,F5 f H 7KLV WHUP LV DSSOLHG YLD WKH HTXDWLRQ f§ ‘ m: r RS;f [ f

PAGE 159

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f DQG 5XWKYHQ f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f $FFRUGLQJ WR 6X]XNL f PRVW

PAGE 160

DGVRUSWLRQ LVRWKHUPV DUH IDYRUDEOH DQG H[KLELW WKH QRQOLQHDU /DQJPXLU IRUP $ IDYRUDEOH LVRWKHUP FDQ EH DVVXPHG IRU D VLQJOH DGVRUEDWH ZLWK D KLJK DGVRUSWLRQ FDSDFLW\ ZKRVH FRQFHQWUDWLRQ GHFUHDVHV LQ WKH GRZQVWUHDP IORZ GLUHFWLRQ )UH\ f 7KH DV\PSWRWLF OLPLW LV UHDFKHG ZKHQ WKH IOXLGSKDVH IURQW PDWFKHV WKH DGVRUEHGSKDVH IURQW DQG s e D & f r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f ZKHUH HY UHSUHVHQWV WKH YRLG IUDFWLRQ RI WKH EHG &RUUHODWLRQ RI WKLV PRGHO ZLWK WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO XVLQJ WKH DVVXPSWLRQ RI FRQVWDQWSDWWHUQ EHKDYLRU SURGXFHV 3 E GJ A G;B H GW GW 9 UDWH f

PAGE 161

7KH UDWH HTXDWLRQV REWDLQHG IURP WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO FDQ EH XVHG WRJHWKHU ZLWK WKH IL[HGEHG PRGHO EXW VLPSOH DQDO\WLFDO VROXWLRQV DUH KDUG WR ILQG DQG WKH GHWHUPLQDWLRQ RI UDWH FRQVWDQWV E\ WKH PHWKRG RI OHDVW VTXDUHV LV QRW SRVVLEOH 6RPH W\SH RI QXPHULFDO SURFHGXUH KDV WR EH XVHG LQVWHDG %MHUOH HW DO f 7KH FRQVWDQWSDWWHUQ DVVXPSWLRQ DOORZV WKH XVH RI WKH UHODWLRQVKLS ; f 6XEVWLWXWLQJ WKH TXDQWLW\ &&Df IRU ; LQ WKH VKULQNLQJ FRUH HTXDWLRQV DOORZHG IRU WKH HYDOXDWLRQ RI WKHLU DSSOLFDELOLW\ WR WKH H[SHULPHQWDO GDWD FROOHFWHG .DUOHJDUG DQG %MHUOH f XVHG WKLV DSSURDFK LQ WKHLU VWXG\ RI WKH UHPRYDO RI +6 E\ ]LQF IHUULWH 2WKHU QRWDEOH H[DPSOHV ZKHUH WKH TXDQWLW\ &&4f ZDV XVHG DV DQ LQGLFDWLRQ RI VROLG FRQYHUVLRQ DUH WKH ZRUN RI :ROII HW DO f ZKR XVHG <$OVXSSRUWHG &D2 WR UHPRYH JDVHRXV 6 DQG 'DP-RKDQVHQ DQG 2VWHUJDDUG Df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

PAGE 162

WKH HYDOXDWLRQ RI WKH H[WUHPH H[SHULPHQWDO FRQGLWLRQV WKH UHPRYDO RI SSP 1 DW DQG WKH UHPRYDO RI SSP 1 DW ZHUH HYDOXDWHG 7KHVH WKUHH FDVHV FRQWDLQ GDWD FROOHFWHG LQ WKH DEVHQFH RI R[\JHQ XVLQJ GU\ VRUEHQW QR 0J2+ff 7KH UHOHYDQFH RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO WR WKHVH VLWXDWLRQV ZLOO DOVR EH JHQHUDOO\ GLVFXVVHG (DFK FDVH ZLOO EH SUHVHQWHG VHSDUDWHO\ ZLWK D GLVFXVVLRQ RI WKH VSHFLILF DSSOLFDELOLW\ DQG FRQVLGHUDWLRQV SSP 12 DW D 5HDFWLRQ 7HPSHUDWXUH RI )LJXUH UHSUHVHQWV WKH PDWKHPDWLFDO WUDQVIRUPDWLRQ RI VROLG VRUEHQW FRQYHUVLRQ GDWD ; RU 1 SHQHWUDWLRQf LQWR WKH YDOXHV REWDLQHG IURP WKH VKULQNLQJ XQUHDFWHGFRUH HTXDWLRQV 7KH OHIW RUGLQDWH VKRZV WKH FKHPLFDOUHDFWLRQFRQWUROHTXDWLRQ YDOXHV (TXDWLRQ f DQG WKH ULJKW RUGLQDWH LQGLFDWHV WKH SURGXFW OD\HU GLIIXVLRQ FRQWURO HTXDWLRQ YDOXHV (TXDWLRQ f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b FRQILGHQFH LQWHUYDOV 7KH WZR HTXDWLRQV DUH VHSDUDWHO\ HYDOXDWHG DV VKRZQ LQ )LJXUHV DQG )URP )LJXUH WKH HTXDWLRQ RI WKH OHDVWVTXDUHV OLQHDU UHJUHVVLRQ OLQH IRU WKH FKHPLFDOUHDFWLRQ FRQWUROHTXDWLRQ LV \ W ZLWK DQ U YDOXH RI )URP WKLV

PAGE 163

;f ;f f§;f r .5 r .' )LJXUH 6KULQNLQJ XQUHDFWHGFRUH PRGHO HYDOXDWLRQ >12@LQ SSP 7 >2@ bf R

PAGE 164

HTXDWLRQ ]U WKH LQYHUVH RI WKH VORSH RI WKH UHJUHVVLRQ OLQHf LV HTXDO WR PLQXWHV :KLOH WKH FRHIILFLHQW RI GHWHUPLQDWLRQnV YDOXH LV UHODWLYHO\ KLJK WKHUH LV REYLRXVO\ VRPH FXUYDWXUH $ FRQVLGHUDEO\ EHWWHU ILW ZDV REWDLQHG IRU WKH SURGXFWOD\HUGLIILLVLRQFRQWURO HTXDWLRQ ZKHUH ERWK WKH U DQG )VWDWLVWLF YDOXHV DUH KLJKHU DV VKRZQ LQ )LJXUH 2QO\ WKH LQLWLDO SRLQW IDOOV RXWVLGH WKH b FRQILGHQFH LQWHUYDO )URP WKLV ILW UG LV GHWHUPLQHG WR EH HTXDO WR PLQXWHV DSSUR[LPDWHO\ WZLFH DV ODUJH DV Uf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f DQK\GURXVf WR 0J2 = LV HVWLPDWHG WR EH DSSUR[LPDWHO\ 8VLQJ (TXDWLRQ D YDOXH RI UG HTXDO WR PLQXWHV ZDV FDOFXODWHG D b VPDOOHU QXPEHU WKDQ WKDW FDOFXODWHG XVLQJ (TXDWLRQ ,W DSSHDUV WKDW WKH PRGHO SUHGLFWV D VLJQLILFDQW HIIHFW IURP WKH PRODU YROXPH UDWLR GLIIHUHQFH

PAGE 165

;f Q L L L %HG ([SRVXUH 7LPH PLQf )LJXUH 6KULQNLQJ XQUHDFWHGFRUH FKHPLFDOUHDFWLRQFRQWUROHTXDWLRQ HYDOXDWLRQ W PLQXWHV >1@ SSP 7 >2@ bf > \ [ UA2 )LW 6WG (UURU ) @ .f

PAGE 166

)LJXUH 6KULQNLQJ XQUHDFWHGFRUH SURGXFWOD\HUGLIIXVLRQFRQWUROHTXDWLRQ HYDOXDWLRQ 7G PLQXWHV >12@LQ SSP 7 >2@ bf > \ 222O[ U0L )LW 6WG (UURU ) @ +f§r

PAGE 167

[[f )LJXUH 6KULQNLQJ XQUHDFWHGFRUH FKHPLFDOUHDFWLRQFRQWUROHTXDWLRQ HYDOXDWLRQ [U PLQXWHV >1&\A62 SSP 7 >2bf > \ [ U0)LW 6WG (UURU ) @ A

PAGE 168

;f ;f )LJXUH 6KULQNLQJ XQUHDFWHGFRUH SURGXFWOD\HUGLIIXVLRQFRQWUROHTXDWLRQ HYDOXDWLRQ [G PLQXWHV >12@LQ SSP 7 >2@ bf > \ 222O[ UA2 )LW 6WG (UURU ) @ NL

PAGE 169

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f )URP WKHVH ILJXUHV WKH PDVV UHPRYDO UDWH GRHV QRW DSSHDU WR DSSURDFK ]HUR HYHQ DIWHU DQ H[WHQGHG H[SRVXUH SHULRG 7KH WRWDO PDVV RI 1 UHPRYHG DIWHU KRXUV ZDV RQO\ PJ RU [ n PROHV 7KLV LV VLJQLILFDQWO\ OHVV WKDQ WKH SUHGLFWHG VWRLFKLRPHWULF FDSDFLW\ RI WKH PDVV RI 0J2 LQ WKH EHG FRUUHVSRQGLQJ WR D XWLOL]DWLRQ UDWLR RI DSSUR[LPDWHO\

PAGE 170

&XP 0DVV 12 5HPRYHG PJf )LJXUH &XPXODWLYH PDVV 1 UHPRYHG FROOHFWHG IURP FHQWHU RI EHGf f§,f§ 12[ $QDO\]HU 'DWD YHUVXV EHG H[SRVXUH WLPH >12@LQ SSP 7 >2@ b *DV VDPSOHV

PAGE 171

%HG ([SRVXUH 7LPH PLQf 12 $QDO\]HU 'DWD [ )LJXUH &XPXODWLYH PDVV 1 UHPRYHG YHUVXV EHG H[SRVXUH WLPH IRU D ORQJWHUP UXQ >12@LQ SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP FHQWHU RI EHGf

PAGE 172

b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f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b 7KH PDLQ UHDVRQV IRU WKLV SKHQRPHQRQ DUH VWUXFWXUDO FKDQJHV RI WKH SRUHV 0D[LPXP FRQYHUVLRQ LV UHDFKHG ZKHQ SRUH FORVXUH RFFXUV 5DPDFKDQGUDQ DQG 6PLWK f 7KH UHDFWLRQ RI 6 ZLWK &D2 RU 0J2 IRUPV D VROLG SURGXFW &D6 RU 0J6 ZKLFK RFFXSLHV PRUH VSDFH WKDQ WKH VROLG UHDFWDQW ,I WKHUH LV QR LQFUHDVH LQ WKH RYHUDOO YROXPH RI WKH SRURXV SDUWLFOHV GXULQJ UHDFWLRQ WKH SRUH VSDFH PD\ EHFRPH FRPSOHWHO\ ILOOHG ZLWK RU FRYHUHG E\ VROLG SURGXFW 2QFH SRUHV EHFRPH SOXJJHG WKH LQWHULRU RI WKH SRURXV SDUWLFOHV DUH SUDFWLFDOO\ LQDFFHVVLEOH WR JDVHRXV UHDFWDQWV VLQFH GLIIXVLRQ WKURXJK WKH VROLG SURGXFW LV DQ H[WUHPHO\ VORZ SURFHVV 6RWLUFKRV DQG =DUNDQLWLV f %KDWLD DQG 3HUOPXWWHU f DWWULEXWHG WKH FRQWLQXHG VORZ VROLGVXUIDFH

PAGE 173

FRQYHUVLRQ LQ JDVf§VROLG UHDFWLRQV WR WKH HIIHFWLYH FORVXUH RI D QDUURZ VL]H UDQJH RI VPDOO VXUIDFH SRUHV OHDYLQJ ODUJHU SRUHV RI VPDOOHU VXUIDFH DUHD DYDLODEOH IRU UHDFWLRQ DW D JUHDWO\ UHGXFHG UDWH %KDWLD DQG 3HUOPXWWHU Ef SUHVHQWHG D FRUUHODWLRQ WKDW DOORZV IRU WKH FDOFXODWLRQ RI PD[LPXP FRQYHUVLRQ EDVHG XSRQ LQLWLDO SRURVLW\ HDf DQG = HR =OfOH f $VVXPLQJ DQ LQLWLDO SRURVLW\ YDOXH RI EDVHG XSRQ YDOXHV XVHG IRU &D2 E\ WKH DXWKRUV WKH FDOFXODWHG YDOXH IRU ;PD[ LV DSSUR[LPDWHO\ b +LJK LQLWLDO VRUEHQW SRURVLW\ DQG DVVRFLDWHG KLJK VXUIDFH DUHDf FDQ FUHDWH WKH SRWHQWLDO IRU LQWHUQDO GLIIXVLRQ OLPLWDWLRQV WR FKHPLFDO UHDFWLRQ 0DUVK DQG 8OULFKVRQ f IRXQG WKDW JUDLQV RI &D2 REWDLQHG E\ FDOFLQLQJ &D2+f FRQWDLQHG VLJQLILFDQW TXDQWLWLHV RI PLFURSRUHV PDQ\ RI WKHP DURXQG ƒ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f 7KH VLPSOH JUDLQ PRGHO FDQQRW UHDOLVWLFDOO\ VLPXODWH SRURVLWLHV OHVV WKDQ b ZKHUHDV LQLWLDO SRURVLWLHV RI FDOFLQHG OLPHVWRQHV FDQ EH JUHDWHU WKDQ b 'DPOH f

PAGE 174

0RVW H[SHULPHQWDO ZRUN SXEOLVKHG RQ &D26 UHDFWLRQV ZDV FRQGXFWHG XVLQJ 6 FRQFHQWUDWLRQV LQ WKH WHQWKV RI D SHUFHQW RQH WR WZR RUGHUV RI PDJQLWXGH JUHDWHU WKDQ 1 FRQFHQWUDWLRQV HYDOXDWHG LQ WKLV UHVHDUFK ,W LV QRWHG WKDW (TXDWLRQ GRHV QRW KDYH D JDV SKDVH FRQFHQWUDWLRQ WHUP $W D JLYHQ FRQYHUVLRQ OHYHO RU SURGXFW OD\HU WKLFNQHVVf 0DUVK DQG 8OULFKVRQ f IRXQG WKDW GRXEOLQJ WKH EXON JDVSKDVH DGVRUEDWH FRQFHQWUDWLRQ VKRXOG GRXEOH WKH RYHUDOO UHDFWLRQ UDWH FRQVLVWHQW ZLWK WKH LGHD RI D SURGXFW OD\HU VXUURXQGLQJ WKH DFWLYH VXUIDFH DUHD RI WKH R[LGH $QRWKHU H[DPSOH LV WKH GHJUHH RI ]LQF IHUULWH XWLOL]DWLRQ LQ WKH UHPRYDO RI SSP +6 DW WR & DV UHSRUWHG E\ .DUOHJDUG DQG %MHUOH f ZKLFK YDULHG IURP WR b $ VLJQLILFDQW FRUUHODWLRQ EHWZHHQ DGVRUEDWH FRQFHQWUDWLRQ DQG VROLG FRQYHUVLRQ LQ WKH UHDFWLRQ EHWZHHQ 6 DQG &D2 ZDV UHSRUWHG E\ 6LPRQV f 6LPRQV UHSRUWHG D XWLOL]DWLRQ IDFWRU RI DSSUR[LPDWHO\ IRU WKLV HIIHFW PHDQLQJ D WHQIROG LQFUHDVH LQ FRQFHQWUDWLRQ SURGXFHG D VL[IROG LQFUHDVH LQ HIIHFWLYH XWLOL]DWLRQ %DVHG XSRQ WKLV FRUUHODWLRQ D WHQIROG ORZHU FRQFHQWUDWLRQ ZRXOG SUHGLFW DQ ;PD[ YDOXH RI DSSUR[LPDWHO\ b DQG D KXQGUHGIROG ORZHU FRQFHQWUDWLRQ ZRXOG SUHGLFW DQ ;PD[ YDOXH RI DSSUR[LPDWHO\ b 7KLV PD\ SDUWLDOO\ H[SODLQ WKH ORZ 1 XWLOL]DWLRQ UDWH E\ 0J2 6LQFH OLWWOH LV NQRZQ DERXW WKH SRUH VL]H GLVWULEXWLRQ RI WKH 0J2YHUPLFXOLWH VRUEHQW LW LV GLIILFXOW WR VD\ ZKDW HIIHFW LW PD\ KDYH RQ VRUEHQW XWLOL]DWLRQ +RZHYHU 6LPRQV DOVR SRLQWHG RXW WKDW VPDOOSRUH SOXJJLQJ LV GRPLQDQW IRU WR SP SDUWLFOHV ZKLOH SRUHPRXWK SOXJJLQJ LV GRPLQDQW IRU SP SDUWLFOHV 5DWH OLPLWDWLRQV UHVXOWLQJ IURP SURGXFWOD\HU GLIIXVLRQ SULRU WR SOXJJLQJ RU ILOOLQJf DUH PRVW HYLGHQW IRU WKLFN SURGXFW OD\HUV RFFXUULQJ ZLWKLQ WKH ODUJHU SRUHV RI ODUJHU SDUWLFOHV ,I SRUH SOXJJLQJ ZLWKLQ WKH JUDLQV RI WKH SDUWLFOH ZHUH RFFXUULQJ ORZHU FRQYHUVLRQ OHYHOV DW

PAGE 175

KLJKHU WHPSHUDWXUHV ZRXOG EH H[SHFWHG 1 SHQHWUDWLRQ FXUYHV ZLWK D VKDUS EUHDN WR D IODW FXUYH ZRXOG EH H[SHFWHG LI WRWDO SRUH FORVXUH RFFXUV +RZHYHU SRVLWLYH VORSHV DUH VHHQ HYHQ DIWHU D VKDUS GURS LQ UHDFWLRQ UDWH 0DUVK DQG 8OULFKVRQ f 7KH FDOFLQDWLRQ RI &D2+f WR &D2 UHVXOWV LQ KLJKO\ SRURXV SDUWLFOHV ZLWK D WKHRUHWLFDO SRURVLW\ RI DSSUR[LPDWHO\ b 1HZWRQ HW DO f 'HSHQGLQJ XSRQ WKH WHPSHUDWXUH XVHG WKH DFWXDO SRURVLW\ FDQ EH ORZHU SDUWLFXODUO\ IRU WKH FDOFLQDWLRQ RI &D2+f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f YHUVXV WLPH FXUYHV DUH XVHIXO %DVHG XSRQ WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO FRPSOHWH FRQYHUVLRQ ZRXOG EH SUHGLFWHG WR RFFXU DW DSSUR[LPDWHO\ PLQXWHV RQHKDOI RI DSSUR[LPDWHO\ PLQXWHVf /RRNLQJ DW )LJXUH

PAGE 176

N VHFf )LJXUH 3ORW RI LQYHUVH ILUVWRUGHU UDWH FRHIILFLHQW Nf YHUVXV EHG H[SRVXUH WLPH >1@LQ SSP 7 >2@ bf &? 8f

PAGE 177

([SHULPHQWDO 3UHGLFWHG )LJXUH ([SHULPHQWDO YHUVXV SUHGLFWHG 1 FRQFHQWUDWLRQ >1@LQ SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP PLGGOH RI EHGf

PAGE 178

HIIHFWLYH 1 UHPRYDO KDV VWDELOL]HG DW D YDOXH RI DSSUR[LPDWHO\ b DW WKLV H[SRVXUH WLPH $OORZLQJ WKH UHDFWLRQ WR SURJUHVV IRU IRXU PRUH KRXUV 1 UHPRYDO KDG RQO\ FKDQJHG WR b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

PAGE 179

;f ;f ;f $ SSP5 [ SSP' )LJXUH 6KULQNLQJ XQUHDFWHGFRUH PRGHO HYDOXDWLRQ >12@LQ SSP 7 >2@ bf R? 2Q

PAGE 180

[ 2V -

PAGE 181

7 >2@ bf > \ 222O[ U0f )LW 6WG (UURU )@ 21 22

PAGE 182

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

PAGE 183

([SHULPHQWDO 3UHGLFWHG )LJXUH ([SHULPHQWDO YHUVXV SUHGLFWHG 1 FRQFHQWUDWLRQ >1@LQ SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP PLGGOH RI EHGf R

PAGE 184

[f ;f r ;f .5 r .' )LJXUH 6KULQNLQJ XQUHDFWHGFRUH PRGHO HYDOXDWLRQ >1@LQ SSP 7 >2@ bf

PAGE 185

[ )LJXUH 6KULQNLQJ XQUHDFWHGFRUH FKHPLFDOUHDFWLRQFRQWUROHTXDWLRQ HYDOXDWLRQ W PLQXWHV >1@LQ SSP 7 >2@ bf > \ [ A )LW 6WG (UURU ) @ R WR

PAGE 186

%HG ([SRVXUH 7LPH PLQf )LJXUH 6KULQNLQJ XQUHDFWHGFRUH SURGXFWOD\HUGLIIXVLRQFRQWURO HTXDWLRQ HYDOXDWLRQ 7G PLQXWHV >12AMA22 SSP 7 >2@ bf > \ 222O[ U0f )LW 6WG (UURU ) @ M 8f

PAGE 187

7KHUH LV VRPH GHYLDWLRQ RXWVLGH WKH FRQILGHQFH LQWHUYDOV LQ WKH PLGGOH RI WKH UXQ PRVW SUREDEO\ WKH UHVXOW RI D WUDQVLWLRQ IURP RQH PHFKDQLVP WR DQRWKHU %RWK WKH FRPELQDWLRQ RI FRQWUROOLQJPHFKDQLVP DQG LQFUHDVHGPRODUYROXPH HTXDWLRQV SUHGLFWHG [ YDOXHV EHWZHHQ DSSUR[LPDWHO\ WR PLQXWHV 7R HYDOXDWH WKH SUHGLFWDELOLW\ RI VRUSWLRQ EHKDYLRU DW WKLV UHODWLYHO\f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f DOVR GHFUHDVHG $IWHU PLQXWHV 1 UHPRYDO ZDV DSSUR[LPDWHO\ b DQG WKH 12 FRQFHQWUDWLRQ KDG DOPRVW UHDFKHG LWV LQLWLDO YDOXH ,W LV H[SHFWHG WKDW ZKHQ 1 ZDV QR ORQJHU

PAGE 188

([SHULPHQWDO 3UHGLFWHG )LJXUH ([SHULPHQWDO YHUVXV SUHGLFWHG 1 FRQFHQWUDWLRQ >1@LQ SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP PLGGOH RI EHGf

PAGE 189

&XP 0DVV 14 5HPRYHG PJf %HG ([SRVXUH 7LPH PLQf 12[ $QDO\]HU 'DWD )LJXUH &XPXODWLYH PDVV 1 UHPRYHG YHUVXV EHG H[SRVXUH WLPH IRU D ORQJWHUP UXQ >12@LQ SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP FHQWHU RI EHGf

PAGE 190

f§,f§ >1@ >12@ )LJXUH &RPSDUDWLYH EHG 1 DQG 12 FRQFHQWUDWLRQV YHUVXV EHG H[SRVXUH WLPH GXULQJ ORQJWHUP UXQ VKRZLQJ 12 SURGXFWLRQ >12@LQ SSP 7 >2@ b *DV VDPSOHV FROOHFWHG IURP PLGGOH RI EHGff

PAGE 191

UHPRYHG 12 ZRXOG QR ORQJHU EH SURGXFHG DQG ZRXOG LQGHHG UHDFK LWV LQLWLDO EHG LQOHW FRQFHQWUDWLRQ *HQHUDO $SSOLFDELOLW\ RI WKH 6KULQNLQJ 8QUHDFWHG&RUH 0RGHO ,W DSSHDUV IURP WKH FDVHV MXVW SUHVHQWHG WKDW WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO SURYLGHV D EDVLV IRU FKHPLFDOO\ DQG SK\VLFDOO\ GHVFULELQJ WKH VRUSWLRQ EHKDYLRU RFFXUULQJ GXULQJ WKH UHPRYDO RI 1 E\ WKH 0J2f§YHUPLFXOLWH VRUEHQW &RPELQHG ZLWK WKH FODVVLFDO FKHPLFDO NLQHWLFV HYDOXDWLRQV SHUIRUPHG WKHVH UHVXOWV DOORZ IRU WKH SUHGLFWLRQ RI VRUSWLRQ EHKDYLRU ZLWK WLPH XQGHU D YDULHW\ RI FRQGLWLRQV 7KH XVH RI WKH ORFDOHTXLOLEULXP WKHRU\ WR GHVFULEH WKH WUDQVLWLRQ IURP SDFNHGEHG WKHRULHV WR WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO ZDV D NH\ VWHS LQ WKH SURFHVV :KLOH WKH ORFDOHTXLOLEULXP WKHRU\ DOORZHG IRU WKH GHWHUPLQDWLRQ RI VROLG FRQYHUVLRQ DV LW UHODWHV WR 1 SHQHWUDWLRQ LW LV DSSDUHQW WKDW WKLV FRQYHUVLRQ LV UHODWLYH UDWKHU WKDQ DEVROXWH EDVHG XSRQ WKH ORZ VRUEHQW XWLOL]DWLRQ UDWHV IRXQG 7KHVH ORZ XWLOL]DWLRQ UDWHV DUH FHUWDLQO\ D FRPSOH[ IXQFWLRQ RI PDQ\ IDFWRUV KRZHYHU VXFK DV WKH ORZ 1 FRQFHQWUDWLRQV HYDOXDWHG DV ZHOO DV SUREDEOH ORZ SRURVLWLHV ZKLFK PD\ SURGXFH LQWUDSDUWLFOH GLIIXVLRQDO UHVLVWDQFHV DUH NH\ SDUDPHWHUV 7KH KLJK VROLGSURGXFWWRUHDFWDQW PRODU YROXPH UDWLR LV DQRWKHU PLWLJDWLQJ IDFWRU LQ WKLV VLWXDWLRQ 'DP-RKDQVHQ DQG 2VWHUJDDUG Ef DWWULEXWHG WKH UDSLG GHFUHDVH LQ UHDFWLRQ UDWH GXULQJ WKH VXOIDWLRQ RI OLPHVWRQH SDUWLFOHV WR ILOOLQJ RI PLFURSRUHV 7KH PDFURSRURXV VWUXFWXUH UHPDLQLQJ PD\ DOORZ VRPH VXOIDWLRQ WR FRQWLQXH DW D YHU\ VORZ UDWH 7KH GHJUHH RI FRQYHUVLRQ GHFUHDVHV OLQHDUO\ IURP WKH SDUWLFOH VXUIDFH LQZDUG WR D FHUWDLQ GLVWDQFH IURP WKH VXUIDFH ZKHUH LW JRHV WR ]HUR

PAGE 192

IRUPLQJ D VKDUS IURQW ZKLFK PRYHV LQZDUG ZLWK LQFUHDVLQJ H[SRVXUH WLPH /DUJHU VDPSOHV ZLWK YDU\LQJ &D2f WKLFNQHVV PD\ OHDG WR VLJQLILFDQW LQWUDSDUWLFOH GLIUXVLRQDO OLPLWDWLRQV ZKLFK PD\ LQWHQVLI\ DV WKH VROLG SURGXFW IRUPV SDUWLFOH DJJORPHUDWLRQV =DUNDQLWLV HW DO f :KLOH G\QDPLF HTXLOLEULXP FRQGLWLRQV DUH DFKLHYHG DW D WLPH HTXDO WR [ DQG HIIHFWLYH FRQYHUVLRQ DSSURDFKHV D IDLUO\ FRQVWDQW YDOXH 1 LV VWLOO EHLQJ UHPRYHG IURP WKH JDV VWUHDP DW D UDWH GHSHQGHQW XSRQ LWV HIIHFWLYH GLIILLVLYLW\ WKURXJK WKH SURGXFW OD\HU DQGRU SRURXV VWUXFWXUH RI WKH VRUEHQW 7KHUH UHPDLQV D VLJQLILFDQW UHVLGXDO 1 UHPRYDO FDSDFLW\ ZKLFK FRQWLQXHV DW VLJQLILFDQW OHYHOV IRU H[WHQGHG SHULRGV RI WLPH 7KH PRODU YROXPH UDWLRV RI 0J1f SURGXFWV WR 0J2+f UHDFWDQWV DUH DSSUR[LPDWHO\ b VPDOOHU WKDQ WKH UDWLR WR 0J2 UHDFWDQWV ZKLFK PD\ SDUWLDOO\ DFFRXQW IRU WKH LPSURYHG SHUIRUPDQFH RI K\GUDWHG VRUEHQWV 7KLV UHVXOWV LQ PRUH DYDLODEOH VXUIDFH DUHD WKURXJK XQEORFNHG SRUHVXUIDFH DUHD RQ D QRUPDOL]HG WLPH EDVLV 7KH VKULQNLQJ XQUHDFWHG FRUH PRGHO ZDV XVHG WR LQWHUSUHW UHVXOWV XQGHU WKHVH FRQGLWLRQV DQG LQ JHQHUD`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f FU\VWDOOL]HV ZLWK WKH &GO OD\HU VWUXFWXUH LQ ZKLFK HYHU\ 0J DWRP LV VXUURXQGHG E\ VL[ 2+ (YHU\ 2+ IRUPV WKUHH ERQGV WR 0J DWRPV LQ LWV RZQ OD\HU ZKLOH FRQWDFWLQJ WKUHH 2+ RI DQ DGMDFHQW OD\HU :HOOV f $ ODUJHU EXON UHDFWDQW ZRXOG EH FRPSULVHG RI UHODWLYHO\ WKLQ

PAGE 193

LUUHJXODUO\ VKDSHG OD\HUV DQG PD\ VKRZ PRUH RU OHVV GLIIXVLRQDO UHVLVWDQFH DV WKH ODWWLFH H[SDQGV ZLWK UHDFWLRQ IRUPLQJ D FRPSOH[ VWUXFWXUH %DPIRUG HW DO f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f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f %HFDXVH RI WKH KLJK GLIIXVLRQDO UHVLVWDQFH RI WKH SURGXFW OD\HU WKH FRQFHQWUDWLRQ RI JDVHRXV UHDFWDQW DW WKH UHDFWLRQ LQWHUIDFH DQG FRQVHTXHQWO\ WKH UHDFWLRQ UDWH PD\ QRW EH XQLIRUP HYHQ XQGHU WKH

PAGE 194

L XQLIRUP WHPSHUDWXUHV DQG FRQFHQWUDWLRQV LQ WKH SRURXV QHWZRUN 6WUXFWXUDO SURSHUWLHV GHSHQG QRW RQO\ RQ FRQYHUVLRQ EXW DOVR WKH UHODWLYH UDWHV RI UHDFWLRQ LQWULQVLF NLQHWLFVf DQG PDVV WUDQVSRUW LQ WKH SURGXFW OD\HU 6RWLUFKRV DQG
PAGE 195

&+$37(5 35$&7,&$/ &216,'(5$7,216 &RQVWUDLQWV RQ 0J29HUPLFXOLWH 6RUEHQW 8VDJH %DVHG XSRQ WKH H[SHULPHQWDO UHVXOWV FROOHFWHG LQ WKLV VWXG\ LW LV DSSDUHQW WKDW WKHUH PD\ EH VRPH SUDFWLFDO FRQVWUDLQWV RQ WKH HIIHFWLYH HPSOR\PHQW RI WKH 0J2f§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b UHGXFWLRQ LQ 12[ HPLVVLRQV PD\ EH VXIILFLHQW 7KLV UHVHDUFK KDV VKRZQ WKDW WKH VRUEHQW FDQ FRQWLQXH WR UHPRYH DQ DYHUDJH RI DSSUR[LPDWHO\ b RI WKH WRWDO TXDQWLW\ RI 1 IRU D VLJQLILFDQW SHULRG RI WLPH SDUWLFXODUO\ IURP ORZHU FRQFHQWUDWLRQ JDV VWUHDPV SSPf 7R SXW VRPH RI WKH UHVXOWV LQWR SHUVSHFWLYH WKH PDVV RI VRUEHQW UHTXLUHG IRU WKH UHPRYDO RI 1 IURP D ODUJHU YROXPH VRXUFH FDQ EH FDOFXODWHG 8VLQJ WKH UHVXOWV IURP WKH FRQGLWLRQV XQGHU ZKLFK WKH EHG ZDV QHDUO\

PAGE 196

VDWXUDWHG SSP 1 .f WKH VRUSWLRQ FDSDFLW\ RI WKH VRUEHQW DIWHU KRXUVf ZDV DSSUR[LPDWHO\ PJ 1 J VRUEHQW EHG 7KLV FRUUHVSRQGV WR D PDVV ORDGLQJ UDWLR RI DSSUR[LPDWHO\ [ n J 1 SHU JUDP VRUEHQW DW DQ RYHUDOO UDWH RI [ n J 1 J VRUEHQWKRXU :KLOH LQLWLDO 1 UHPRYDO ZDV JUHDWHU WKDQ b WKH DYHUDJH UHPRYDO RYHU WKH ILYHKRXU SHULRG ZDV DSSUR[LPDWHO\ b 7DNLQJ DV DQ H[DPSOH D DFIP VRXUFH RI SSP1 DW WKH YROXPH RI EHG UHTXLUHG WR WUHDW WKLV VRXUFH ZLWK DOO RWKHU FRQGLWLRQV UHPDLQLQJ WKH VDPH DQG WDNLQJ WKH EHG WR VDWXUDWLRQ DIWHU RQH KRXU LV DSSUR[LPDWHO\ P QRW DQ LQRUGLQDWHO\ ODUJH VL]H 2EYLRXVO\ LW LV QRW SUDFWLFDO WR UHJHQHUDWH RU UHSODFH WKH EHG HYHU\ KRXU DQG WKH YROXPH RI JDV WR EH WUHDWHG PD\ EH ODUJHU VR D ODUJHU EHG ZRXOG EH XVHG $V ZDV DOUHDG\ SRLQWHG RXW MHW HQJLQH WHVW FHOOV DUH QRW VWHDG\ VWDWH VRXUFHV DQG WKH DYHUDJH WHVW UXQ PD\ EH PXFK VKRUWHU WKDQ DQ KRXU LQ GXUDWLRQ $IWHU MHW HQJLQH H[KDXVW LV DXJPHQWHG WKH FRQFHQWUDWLRQ RI 1 ZRXOG DOVR OLNHO\ EH OHVV WKDQ SSP VR WKH EHG ZRXOG ODVW ORQJHU EHFDXVH LW WDNHV ORQJHU WR IRUP WKH SURGXFW OD\HU ZKHQ UHDFWLQJ ZLWK ORZHU FRQFHQWUDWLRQV RI JDVHV /RZHU 1 FRQFHQWUDWLRQV ZLOO SURGXFH ORZHU UHPRYDO UDWHV EXW WKH EHG ZLOO ODVW ORQJHU :KHQ ORZHUFRQFHQWUDWLRQ JDVHV DUH H[SHFWHG HOHYDWHG WHPSHUDWXUHV FDQ KHOS WR LPSURYH GLIIWLVLYLW\ WKURXJK WKH SURGXFW OD\HU LQ WKH DEVHQFH RI D ODUJH FRQFHQWUDWLRQ JUDGLHQW GULYLQJ IRUFH 7KH HIIHFWLYH FRQYHUVLRQ RI 12 WR 1 LV D FUXFLDO HOHPHQW LQ WKH HYHQWXDO VXFFHVVIXO HPSOR\PHQW RI WKLV 12[ FRQWURO PHGLXP 6SLFHU HW DO f QRWHG WKDW H[KDXVW DXJPHQWDWLRQ FDQ LQFUHDVH WKH 12]WR12 UDWLR IURP RU WR RU :KLOH WKH\ DWWULEXWHG VRPH RI WKLV R[LGDWLRQ WR WKH SUHVHQFH RI SHUR[\ UDGLFDOV LQ IXHOULFK FRPEXVWLRQ R[\JHQ R[LGDWLRQ FDQ EH VLJQLILFDQW DV ZHOO 2EYLRXVO\ WKHUH DUH D QXPEHU RI RWKHU YDULDEOHV

PAGE 197

LQFOXGLQJ WKH SUHVHQFH RI RWKHU H[KDXVW JDV FRPSRQHQWV WKDW QHHG LQYHVWLJDWLRQ EHIRUH RQH FDQ WUXO\ SUHGLFW ZKDW PD\ KDSSHQ RQ WKH VRUEHQW VXUIDFH $V ZDV QRWHG E\ :DQGHU DQG 1HOVRQ f EHG SUHVVXUH GURS LV D SRWHQWLDOO\ VLJQLILFDQW OLPLWLQJ IDFWRU LQ WKH XVH RI WKH 0J2YHUPLFXOLWH VRUEHQW LQ D SDFNHGEHG VFHQDULR 1HOVRQ HW DO f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

PAGE 198

&+$37(5 &21&/86,216 $1' 5(&200(1'$7,216 &RQFOXVLRQV 7KH JDVf§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

PAGE 199

12 LV QRW UHDGLO\ UHPRYHG E\ 0J2f§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f 7KH IRUPDWLRQ RI D VROLG SURGXFW OD\HU VORZHG WKH UHDFWLRQ GRZQ DV LW SURJUHVVHG IURP FRQWURO E\ FKHPLFDOUHDFWLRQ NLQHWLFV WR GLIIXVLRQ FRQWURO D PXFK VORZHU SURFHVV 7KH KLJKHU FRQFHQWUDWLRQ JUDGLHQWV SURYLGHG WKH ODUJHU GULYLQJ IRUFHV IRU WKH FRQWLQXHG UHPRYDO RI 1 WKURXJK GLIIXVLRQ 7KH UHDFWLRQ EHWZHHQ 1 DQG 0J2f§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f ZKLFK DSSHDUV WR EH PRUH UHDFWLYH DQG LV WKHUPDOO\ VWDEOH RYHU WKH UDQJH RI WHPSHUDWXUHV VWXGLHG 6LQFH WKH PRODU YROXPH UDWLR RI PDJQHVLXP QLWUDWH WR PDJQHVLXP K\GUR[LGH LV ORZHU WKDQ ZKHQ FRPSDUHG WR 0J2 WKHUH LV OHVV SRWHQWLDO IRU

PAGE 200

DFWLYH VXUIDFH DUHD FRYHUDJH E\ WKH SURGXFW ZKLFK LPSURYHG VRUSWLRQ SHUIRUPDQFH ZLWK WLPH 7KH SUHVHQFH RI R[\JHQ LQ WKH H[KDXVW JDV DLGHG LQ WKH UHPRYDO RI 12 E\ WKH R[LGDWLRQ RI 12 WR 1 ,W DSSHDUHG WKDW R[\JHQ PD\ KDYH SURPRWHG WKH FRQWLQXHG UHPRYDO RI 1 E\ WKH VRUEHQW 7HPSHUDWXUH DIIHFWHG WKH UHDFWLRQ E\ LQFUHDVLQJ UHDFWLRQ UDWHV ERWK LQLWLDOO\ FKHPLFDOUHDFWLRQ FRQWUROf DQG E\ LPSURYLQJHIIHFWLYH GLIIXVLYLW\ 5HDFWLRQ WHPSHUDWXUH PXVW EH NHSW EHORZ VLQFH WKLV LV WKH WHPSHUDWXUH DW ZKLFK 0J1f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f $ UHDFWLRQ PHFKDQLVP KDV EHHQ SURSRVHG WR H[SODLQ WKH SURGXFWLRQ RI 12 GXULQJ WKH UHPRYDO RI 1 12 PD\ EH SURGXFHG GXULQJ WKH ILQDO R[LGDWLRQ VWHS RI PL[HG QLWULWHf§QLWUDWH SURGXFW WR LWV ILQDO QLWUDWH IRUP $IWHU WKH 1 UHPRYDO UDWH GHFUHDVHV 12 SURGXFWLRQ GHFUHDVHV DFFRUGLQJO\ 7KH VKULQNLQJ XQUHDFWHGFRUH PRGHO ZDV DSSOLHG WR 1 SHQHWUDWLRQ GDWD XVLQJ WKH ORFDOHTXLOLEULXP WKHRU\ WR HVWLPDWH VROLGVRUEHQW FRQYHUVLRQ 7KH PRGHO GHVFULEHG WKH

PAGE 201

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

PAGE 202

SURYLGHG IURP H[KDXVWJDV DXJPHQWDWLRQ PD\ DFFRPSOLVK VRPH RI WKLV R[LGDWLRQ 0RUH 0J2YHUPLFXOLWH VXUIDFH VWXGLHV DUH QHFHVVDU\ WR EHWWHU XQGHUVWDQG WKH QDWXUH RI WKH VXUIDFH DQG WKH UROH LW SOD\V LQ WKH UHPRYDO RI 1 IURP D IORZLQJ JDV VWUHDP 3RUH VL]H GLVWULEXWLRQV DQG DVVRFLDWHG VWUXFWXUDO LPSOLFDWLRQV PD\ EH YHU\ XVHIXO LQ IXUWKHU H[SODLQLQJ WKH UDWHOLPLWLQJ SURFHVVHV LQYROYHG LQ WKLV UHDFWLRQ 7KH PHDVXUHPHQW RI DGGLWLRQDO SK\VLFDO SDUDPHWHUV LV UHTXLUHG IRU WKH VWULFW PDWKHPDWLFDO LQWHUSUHWDWLRQ RI VKULQNLQJ XQUHDFWHGFRUH PRGHO UHVXOWV 5HDOWLPH VXUIDFH DQDO\VHV ZRXOG DOORZ IRU IROORZLQJ WKH SURJUHVV RI UHDFWLRQ ZKLFK FRXOG EH FRPELQHG ZLWK SRVWVRUSWLRQ DQDO\VHV IRU WKH TXDQWLILFDWLRQ RI VROLG FRQYHUVLRQ DQG SURGXFW IRUPDWLRQ $ PRGLILHG UHDFWRU GHVLJQ WKDW ZRXOG DOORZ IRU WKH UHPRYDO RI DOO XVHG VRUEHQW DIWHU D UXQ ZLWKRXW DQ\ ORVVf ZRXOG EH DQ LPSURYHPHQW LQ WKLV UHJDUG 0RUH WUDGLWLRQDO VRUSWLRQ VWXGLHV ZKHUH WKH SURJUHVV RI UHDFWLRQ LV IROORZHG XVLQJ WKHUPRJUDYLPHWULF PHWKRGV DQG VLQJOH VRUEHQW SHOOHWV PD\ DOVR DVVLVW LQ WKH DSSOLFDWLRQ RI WKH VKULQNLQJ XQUHDFWHGFRUH PRGHO WR WKLV UHDFWLRQ 7KLV UHVHDUFK LV DFWXDOO\ D VWHS DKHDG LQ WKDW UHJDUG VLQFH WKH EDVLF UHVHDUFK RQ WKH UHDFWLRQ EHWZHHQ 1 DQG 0J2 SRZGHU ZLWKRXW YHUPLFXOLWH VXEVWUDWHf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

PAGE 203

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

PAGE 204

SUHVHQFH RI ZDWHU DUH WUXO\ WKH UHVXOW RI WKH 0J2+f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f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

PAGE 205

/,67 2) 5()(5(1&(6 &)5 $SSHQGL[ ) 0HDVXUHPHQW 3ULQFLSOH DQG &DOLEUDWLRQ 3URFHGXUH IRU WKH 0HDVXUHPHQW RI 1LWURJHQ 'LR[LGH LQ WKH $WPRVSKHUH *DV 3KDVH &KHPLOXPLQHVHQFHf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f D %KDWLD 6 DQG 3HUOPXWWHU ' 7KH (IIHFW RI 3RUH 6WUXFWXUH RQ )OXLG6ROLG 5HDFWLRQV $SSOLFDWLRQ WR WKH 6/LPH 5HDFWLRQ $,&K( -RXUQDO f E %KDWLD 6 DQG 3HUOPXWWHU ' (IIHFW RI WKH 3URGXFW /D\HU RQ WKH .LQHWLFV RI WKH &/LPH 5HDFWLRQ $,&K( -RXUQDO Of %L\DQL 3 DQG *RRFKHH & ) 1RQOLQHDU )L[HG%HG 6RUSWLRQ :KHQ 0DVV 7UDQVIHU DQG 6RUSWLRQ $UH &RQWUROOLQJ $,&K( -RXUQDO f

PAGE 206

%MHUOH ;X ) DQG
PAGE 207

'DP-RKDQVHQ DQG 2VWHUJDDUG +LJK7HPSHUDWXUH 5HDFWLRQ EHWZHHQ 6XOSKXU 'LR[LGH DQG /LPHVWRQH ,, $Q ,PSURYHG ([SHULPHQWDO %DVLV IRU D 0DWKHPDWLFDO 0RGHO &KHPLFDO (QJLQHHULQJ 6FLHQFH f E 'DP-RKDQVHQ +DQVHQ 3)% DQG 2VWHUJDDUG +LJK7HPSHUDWXUH 5HDFWLRQV %HWZHHQ 6XOSKXU 'LR[LGH DQG /LPHVWRQH ,,, $ *UDLQ0LFURJUDLQ 0RGHO DQG LWV 9HULILFDWLRQ &KHPLFDO (QJLQHHULQJ 6FLHQFH f 'DPOH $ 6 0RGHOLQJ D )XUQDFH 6RUEHQW 6OXUU\ ,QMHFWLRQ 3URFHVV $LU DQG :DVWH 'DYLQL 3 5HGXFWLRQ RI 1LWURJHQ 2[LGHV ZLWK $PPRQLD 7KH $FWLYLW\ RI &HUWDLQ 6RRWV )XHO )RJOHU +6 (OHPHQWV RI &KHPLFDO 5HDFWLRQ (QJLQHHULQJ 3 7 5 3UHQWLFH +DOO (QJOHZRRG &OLIIV 1HZ -HUVH\ )UH\ ' 0L[HG*DV $GVRUSWLRQ '\QDPLFV RI +LJK&RQFHQWUDWLRQ &RPSRQHQWV LQ D 3DUWLFXODWH %HG $,*L( -RXUQDO f *DQ] 6 1 6RUSWLRQ RI 1LWURJHQ 2[LGHV E\ 6ROLG 6RUEHQWV -RXUQDO RI $SSOLHG &KHPLVWU\ RI WKH 8665 f *XOOHWW % -R]HZLF] : DQG 6WHIDQVNL / $ 5HDFWLRQ .LQHWLFV RI &D%DVHG 6RUEHQWV ZLWK +& ,QGXVWULDO (QJLQHHULQJ &KHPLVWU\ 5HVHDUFK f +DMDOLJRO 05 /RQJZHOO -3 DQG 6DURILP $) $QDO\VLV DQG 0RGHOLQJ RI WKH 'LUHFW 6XOIDWLRQ RI &D& ,QGXVWULDO (QJLQHHULQJ &KHPLVWU\ 5HVHDUFK f +DP '2 0RQL] DQG *RXYHLD $GGLWLYHV IRU 12> (PLVVLRQV &RQWURO IURP )L[HG 6RXUFHV (6/75 +4 $)(6& 7\QGDOO $)% )ORULGD +DQVHQ 3)% 'DP-RKDQVHQ -RKQVVRQ -( DQG +XOJDDUG 7 &DWDO\WLF 5HGXFWLRQ RI 12 DQG 1 RQ /LPHVWRQH 'XULQJ 6XOIXU &DSWXUH XQGHU )OXLGL]HG %HG &RPEXVWLRQ &RQGLWLRQV &KHPLFDO (QJLQHHULQJ 6FLHQFH f +DUWPDQ 0 DQG &RXJKOLQ 5 : 5HDFWLRQ RI 6XOIXU 'LR[LGH ZLWK /LPHVWRQH DQG WKH *UDLQ 0RGHO $,&K( -RXUQDO f +D\ZDUG '2 DQG 7UDSQHOO %0: &KHPLVRUSWLRQ QG HG %XWWHUZRUWK DQG &R /RQGRQ

PAGE 208

+HDS 0 3 &KHQ 6 / .UDPOLFK & 0F&DUWK\ -0 DQG 3HUVKLQJ : $Q $GYDQFHG 6HOHFWLYH 5HGXFWLRQ 3URFHVV IRU 12[ &RQWURO 1DWXUH +LQGV :& $HURVRO 7HFKQRORJ\ 3URSHUWLHV %HKDYLRU DQG 0HDVXUHPHQW RI $LUERUQH 3DUWLFOHV -RKQ :LOH\ DQG 6RQV 1HZ
PAGE 209

.ULVKQDQ 6 9 DQG 6RWLUFKRV 6 9 $ 9DULDEOH 'LIIXVLYLW\ 6KULQNLQJ&RUH 0RGHO DQG LWV $SSOLFDWLRQ WR WKH 'LUHFW 6XOIDWLRQ RI /LPHVWRQH 7KH &DQDGLDQ -RXUQDO RI &KHPLFDO (QJLQHHULQJ D .ULVKQDQ 6 9 DQG 6RWLUFKRV 6 9 6XOIDWLRQ RI +LJK 3XULW\ /LPHVWRQHV 8QGHU 6LPXODWHG 3)%& &RQGLWLRQV 7KH &DQDGLDQ -RXUQDO RI &KHPLFDO (QJLQHHULQJ E .ULVKQDQ 6 9 DQG 6RWLUFKRV 6 9 ([SHULPHQWDO DQG 7KHRUHWLFDO ,QYHVWLJDWLRQ RI )DFWRUV $IIHFWLQJ WKH'LUHFW /LPHVWRQH+6 5HDFWLRQ ,QGXVWULDO (QJLQHHULQJ &KHPLVWU\ 5HVHDUFK f /DQJHnV +DQGERRN RI &KHPLVWU\ WK HG 0F*UDZ+LOO %RRN &RPSDQ\ 1HZ
PAGE 210

0XWDVKHU ( / .KDQ $ 5 DQG %RZHQ + &RPPXQLFDWLRQV 5HYHUVLEOH 1RQFDWDO\WLF 5HDFWLRQV %HWZHHQ *DVHV DQG 6ROLGV LQ )L[HG %HGV ,QGXVWULDO (QJLQHHULQJ &KHPLVWU\ 5HVHDUFK f 1HOVRQ %: 1HOVRQ 6* DQG +LJJLQV 02 $ 1HZ &DWDO\VW IRU 12 &RQWURO (6/75 +4 $)(6& 7\QGDOO $)% )ORULGD 1HOVRQ %: 9DQ 6WRQH '$ DQG 1HOVRQ 6* 'HYHORSPHQW DQG 'HPRQVWUDWLRQ RI D 1HZ )LOWHU 6\VWHP WR &RQWURO (PLVVLRQV 'XULQJ -HW (QJLQH 7HVWLQJ &(/75 +4 $)&(6$ 7\QGDOO $)% )ORULGD 1HOVRQ 6* 9DQ 6WRQH '$ /LWWOH 5& DQG 3HWHUVRQ .$ /DERUDWRU\ (YDOXDWLRQ RI D 5HDFWLYH %DIIOH $SSURDFK WR 12 &RQWURO )LQDO 5HSRUW &RQWUDFW 1XPEHU )2& +4 $)&(6$ 7\QGDOO $)% )ORULGD 1HOVRQ 6* 6DQLWHFK ,QFf 86 3DWHQW -DQ f 1HOVRQ 6* 1HOVRQ %: DQG +LJJLQV 02 6WXG\ RI WKH 5HJHQHUDELOLWY RI D 8QLTXH 1HZ 6RUEHQW WKDW 5HPRYHV 61 IURP )OXH *DVHV 3KDVH ,, )LQDO 5HSRUW (3$ 6%,5 &RQWUDFW 6DQLWHFK ,QF 7ZLQVEXUJ 2KLR 1HZWRQ + &KHQ 6 / DQG .UDPOLFK & 5ROH RI 3RURVLW\ /RVV LQ /LPLWLQJ 6 &DSWXUH E\ &DOFLXP %DVHG 6RUEHQWV $,&K( -RXUQDO f 17,6 7HFK 1RWHV 3URFHVV (OLPLQDWHV b RI 12[ IURP ([KDXVW LQ /DE 7HVWV '2( 7HFKQRORJ\ $SSOLFDWLRQ -XO\ 2UJDQL]DWLRQ IRU (FRQRPLF &RRSHUDWLRQ DQG 'HYHORSPHQW 2(&'f &RQWURO 7HFKQRORJ\ IRU 1LWURJHQ 2[LGH (PLVVLRQV IURP 6WDWLRQDU\ 6RXUFHV 3DULV )UDQFH 3DNUDVL $ .LQHWLF 6WXGLHV RQ WKH 5HPRYDO RI +\GURJHQ &KORULGH IURP )OXH *DV E\ +\GUDWHG /LPH 3RZGHUV LQ D %HQFK 6FDOH )L[HG %HG 5HDFWRU 3K' 'LVVHUWDWLRQ 8QLYHUVLW\ RI 7HQQHVVHH .QR[YLOOH 7HQQHVVHH 3HUU\ 5 $ DQG 6LHEHUV / 5DSLG 5HGXFWLRQ RI 1LWURJHQ 2[LGHV LQ ([KDXVW *DV 6WUHDPV 1DWXUH f 3HWULN 0$ 8VH RI 0HWDO 2[LGH (OHFWURFDWDOYVWV WR &RQWURO 12(PLVVLRQV IURP )L[HG 6RXUFHV (6/75 +4 $)(6& 7\QGDOO $)% )ORULGD 3LJIRUG 5 / DQG 6OLJHU 5DWH RI 'LIIXVLRQ&RQWUROOHG 5HDFWLRQ %HWZHHQ D *DV DQG D 3RURXV 6ROLG 6SKHUH ,QGXVWULDO (QJLQHHULQJ &KHPLVWU\ 3URFHVV 'HVLJQ DQG 'HYHORSPHQW f

PAGE 211

4XDUOHV DQG /HZLV :+ 7KH 1HZ &OHDQ $LU $FW $ *XLGH WR WKH &OHDQ $LU 3URJUDP DV $PHQGHG LQ 0RUJDQ /HZLV DQG %RFNLXV :DVKLQJWRQ '& 5DPDFKDQGUDQ 3 $ DQG 6PLWK 0 $ 6LQJOH3RUH 0RGHO IRU *DV6ROLG 1RQFDWDO\WLF 5HDFWLRQV $,&K( -RXUQDO f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f 6QRZ 0 + /RQJZHOO 3 DQG 6DURILP $ ) 'LUHFW 6XOIDWLRQ RI &DOFLXP &DUERQDWH ,QGXVWULDO (QJLQHHULQJ &KHPLVWU\ 5HVHDUFK f 6RWLUFKRV 6 9 DQG
PAGE 212

6YHQVVRQ 5 /MXQJVWURP ( DQG /LQGTYLVW 2 .LQHWLFV RI WKH 5HDFWLRQ %HWZHHQ 1LWURJHQ 'LR[LGH DQG :DWHU 9DSRXU $WPRVSKHULF (QYLURQPHQW f 6]HNHO\ (YDQV -: DQG 6RKQ +< *DV6ROLG 5HDFWLRQV $FDGHPLF 3UHVV 1HZ
PAGE 213


PAGE 214

%,2*5$3+,&$/ 6.(7&+ 0DMRU /DUU\ 7KRPDV .LPP ZDV ERUQ LQ 0RXQW &OHPHQV 0LFKLJDQ RQ 1RYHPEHU $IWHU JUDGXDWLRQ IURP )UDVHU +LJK 6FKRRO LQ )UDVHU 0LFKLJDQ LQ KH DWWHQGHG WKH 86 $LU )RUFH $FDGHP\ DW &RORUDGR 6SULQJV &RORUDGR XQWLO KLV JUDGXDWLRQ DQG FRPPLVVLRQLQJ DV D VHFRQG OLHXWHQDQW LQ WKH 86 $LU )RUFH LQ +H UHFHLYHG D 0DVWHU RI 6FLHQFH GHJUHH LQ ELRORJ\ LQGXVWULDO K\JLHQHf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

PAGE 215

, FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWDEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 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` \ 5REHUW +DU 3URIHVVRU RI DQ KHPLVWU\ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWDEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ -RVHSK :DQGHU $GMXQFW $VVRFLDWH 3URIHVVRU RI (QYLURQPHQWDO (QJLQHHULQJ 6FLHQFHV

PAGE 216

7KLV GLVVHUWDWLRQ ZDV VXEPLWWHG WR WKH *UDGXDWH )DFXOW\ RI WKH &ROOHJH RI (QJLQHHULQJ DQG WR WKH *UDGXDWH 6FKRRO DQG ZDV DFFHSWHG DV SDUWLDO IXOILOOPHQW RI WKH UHTXLUHPHQWV IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ $XJXVW :LQIUHG 0 3KLOOLSV 'HDQ &ROOHJH RI (QJLQHHULQJ .DUHQ $ +ROEURRN 'HDQ *UDGXDWH 6FKRRO


Figure 4-10. N02 penetration vs bed exposure time (T=373 K, [O2]=0%).
VO


77
recorded by the analyzer. These data were plotted according to the BET model to determine
the total surface area of the sample. Initial screening results were used as an indication of the
best preparation method to use in making the sorbent. Based upon these results, a large
quantity of sorbent was prepared following the procedure already described. Five additional
samples were taken and analyzed using the BET method for precision estimates.
Determination of Sorbent Particle Size
The MgOvermiculite sorbent is inherently a highly heterogeneous material. As a
result, the range of particle sizes in a given sorbent sample can be great. Samples of
MgOvermiculite sorbent were sieved using a standard ASTM sieving apparatus. Cumulative
mass collected at each stage was plotted agaihst particle diameter using a log-probability chart
to determine the mass median particle diameter. This median size was used in all calculations
for which sorbent particle diameter was required to determine flow characteristics through
the packed bed. It was noted that this value may or may not be the most accurate value, since
sieving the sorbent material likely causes a decrease in particle size through mechanical
abrasion. The true size is probably somewhat larger. Data for magnesium oxide powder
particle size was plotted similarly using a manufacturer-provided mass distribution. Once the
MgO powder has been applied to the surface of the vermiculite, any agglomeration would
tend to create larger size particles. It is not known, however, how the application of
Mg(OH)2 slurry affects particle size. Results of these determinations are presented in the
following chapter.


44
be the most important variable affecting sorption behavior. Chu and Rochelle (1989)
examined the simultaneous removal of S02 and NOx (as NO) by Ca(OH)2 and in some cases
additives, including fly ash, CaS03, and NaOH. These additives improved NOx removal.
Another bench-scale evaluation of the removal of S02 and NO by Ca(OH)2 was conducted
by Jozewicz et al. (1990). NO was not very reactive toward Ca(OH)2 relative to S02, but
adding Mg(OH)2 did improve NO removal somewhat.
SO-,Calcium Carbonate Reaction
Pigford and Sliger (1970) noted the effects of increased product layer difliisional
resistance in the reaction between S02 and CaC03. The overall reaction rate was governed
by both diffusion of S02 through a layer of solid reaction product, progressively formed on
the active solid surface, and diffusion of S02 through pores in the solid. The direct sulfation
of CaC03 was studied by Hajaligol et al. (1988) via thermogravimetric analyses and a bench-
scale fluidized bed setup. The product layer diffusion controlled shrinking unreacted-core
model was used to fitted to the experimental data, using a correction factor for product layer
volume, although chemical kinetics were still important to determining the overall reaction
rate. Snow et al. (1988) also evaluated the direct sulfation of limestone under conditions that
did not decompose CaC03 to CaO. Their results were similar to those of Hajaligol et al.
(1988). As previously mentioned, a variable diffusivity shrinking unreacted-core model was
successfully applied to the direct sulfation of CaC03 by Krishnan and Sotirchos (1993b).
Effective gas diffusivity through the solid product layer was found to be a strong function of
gas temperature and S02 concentration in the bulk gas-phase.


122
When gas humidification was used, since N02 concentrations were monitored directly at the
bed inlet, chemical kinetics was not a function of N02 absorbed in the humidification vessel.
Since the molar volume ratio is smaller when starting with the hydroxide, more N02 can be
removed before the exterior surface becomes covered and Mg(OH)2 is no longer readily
available for reaction. This may mean that it takes longer to form a complete inert product
layer shell when moisture is available in the system. After this layer is formed, however, the
typical shrinking core behavior with product layer diffusion control would still be expected.
Effects of Residence (Reaction) Time
Studies varying gas flow rate and the resulting NOx bed residence time showed an
effect on N02 removal efficiency, as expected. When whole-bed half-second residence time
data were compared to data from the mid-point of the bed for a one-second residence time,
they were found to be directly comparable. These results suggest that the chemical reaction
is truly first-order with respect to NO^ and is essentially zero-order with respect to the solid,
MgO "concentration." The total quantity of N02 removed, however, would be a factor that
would affect overall bed performance, particularly as surface "saturation" is approached.
Better removal efficiencies were obtained with longer residence times. A residence time of
one second produced very high initial removal efficiencies, often greater than 99%. This
situation did not appear to be gas-film-mass-transfer-limited, as already discussed.
Additionally, since longer residence times correspond to lower gas velocities, measured bed
pressure drops were lower, which is preferable for JETC applications. Pressure drop results
will be presented later.


CHAPTER 5
PRACTICAL CONSIDERATIONS
Constraints on MgO-Vermiculite Sorbent Usage
Based upon the experimental results collected in this study, it is apparent that there
may be some practical constraints on the effective employment of the MgOvermiculite
sorbent in the control of nitrogen oxides from jet engine test cells. The widely varying
conditions produced by jet engine testing require a control medium that is effective over the
entire range. While the sorbent effectively removed N02 from a flowing gas stream, the
ability of the sorbent to remove NO was minimal. The low concentrations of total NOx
emitted from a jet engine test cell make their removal by the sorbent more difficult, especially
after a product layer has formed on the exterior surface of the sorbent particles. Depending
upon the focus of any regulations related to jet engine test cells promulgated in the future,
a 50% reduction in NOx emissions may be sufficient.
This research has shown that the sorbent can continue to remove an average of
approximately 50% of the total quantity of N02 for a significant period of time, particularly
from lower concentration gas streams (< 100 ppm). To put some of the results into
perspective, the mass of sorbent required for the removal of N02 from a larger volume source
can be calculated. Using the results from the conditions under which the bed was nearly
182


Cum Mass NQ2 Removed (mg)
Bed Exposure Time (min)
1 NOx Analyzer Data
Figure 4-48 Cumulative mass N02 removed versus bed exposure time for a long-term run ([NO2]in=170 ppm, T=473 K,
[O2]=0%, Gas samples collected from center of bed).


193
Bjerle, I.; Xu, F.; and Ye, Z., "Useful Experimental Technique for the Study of
Heterogeneous Reactions," Chemical Engineering Technology. 15:151-161, 1992.
Borgwardt, R. H., "Kinetics of the Reaction of S02 with Calcined Limestone," Environmental
Science and Technology. 4(l):59-63, 1970.
Borwardt, R. H., "Calcium Oxide Sintering in Atmospheres Containing Water and Carbon
Dioxide," Industrial Engineering Chemistry Research. 28(4):493-500, 1989.
Borgwardt, R. H. and Bruce K R, "Effect of Specific Surface Area on the Reactivity of CaO
with S02," AIChE Journal. 32(2):239-246, 1986.
Borgwardt, R. H.; Roache, N. F.; and Bruce, K. R., "Surface Area of Calcium Oxide and
Kinetics of Calcium Sulfide Formation," Environmental Progress. 3(2): 129-135, 1984.
Campbell, N.T.; Beres, G.A.; Blasko, T.J.; and Groth, R.H., "Effect of Water and Carbon
Dioxide in Chemiluminescent Measurement of Oxides of Nitrogen," Journal of the Air
Pollution Control Association. 32(5):533-535, 1982.
Chen, J. P.; Buzanowski, M. A.; Yang, R. T.; and Cichanowicz, J. E., "Deactivation of the
Vanadia Catalyst in the Selective Catalytic Reduction Process," Journal of the Air and Waste
Management Association. 40(10): 1403-1409, 1990.
Chu, P. and Rochelle, G. T., "Removal of S02 and NOx from Stack Gas by Reaction with
Calcium Hydroxide Solids," Journal of the Air Pollution Control Association. 39(2): 175-179,
1989.
Clark, A., The Theory of Adsorption and Catalysis. Academic Press, New York, 1970, pp.
239-271.
Comes, P.; Gonzalez-Flesca, N.; Menard, T.; and Grimalt, J.O., "Langmuir-Derived
Equations for the Prediction of Solid Adsorbent Breakthrough Volumes of Volatile Organic
Compounds in Atmospheric Emissions Effluents," Analytical Chemistry. 65(8): 1048-1053,
1993.
Cooper, C.D. and Alley, F.C. Air Pollution Control: A Design Approach. 2nd ed., Waveland
Press, Inc.; Prospect Heights, Illinois, pp 485-513, 1994.
CRC Handbook of Chemistry and Physics, 73rd ed., CRC Press; Boca Raton, Florida, pp 4-
72-4-73, 1993.
Dam-Johansen, K. and Ostergaard, K., "High-Temperature Reaction Between Sulphur
Dioxide and Limestone I. Comparison of Limestones in Two Laboratory Reactors and a
Pilot Plant," Chemical Engineering Science. 46(3):827-837, 1991a.


8
oxygen in the combustion gas. The mechanism of the conversion of fuel nitrogen is believed
to proceed via a series of intermediates. One of the most important intermediates is hydrogen
cyanide (HCN), which is fully converted from bound nitrogen in the reaction zone of the
flame. Moving away from the reaction zone, the amount of HCN decreases as a result of its
conversion into N2 and NOx. Ammonia (NH3) may be detected along with N2, HCN, and NOx
in the products of combustion of nitrogen-containing fuels in the presence of a deficiency of
atmospheric nitrogen. The conversion of fuel nitrogen into HCN is rapid, and the rate-
limiting stage in the formation of NOx is the oxidation of HCN. The yield of NOx is
influenced by the content of nitrogen in the fuel, the amount of excess air, and, to a lesser
extent, combustion temperature (Ismagilov et al., 1990; Cooper and Alley, 1994; Wark and
Warner, 1981).
Thermal NO..
Due to the extremely high temperatures in the combustion zones of aircraft jet engines
(up to 2500 K), thermal nitrogen oxides are formed by the oxidation of atmospheric nitrogen.
The accepted model for the chemical reactions responsible for NOx formation in the post
combustion zone was developed by Zeldovich in 1946. The reactions in Zeldovich's model
are as follows:
N2 + 0 ^ NO + N
(1-1)
N + 02 ** NO + 0
(1-2)
and N + OH^NO + H
(1-3)


200ppm 1 100ppm 50ppm 20ppm
Figure 4-8. N02 penetration vs bed exposure time (T=473 K, [O2]=0%)
VO


63
Tubular Fixed-bed Reactors
The reactors, and mixing and sampling manifolds were machined from 316 stainless
steel seamless pipe stock (2.54 cm ID) (T.M.R. Engineering, Micanopy, Florida). Schematic
representations of the reactors, sample lines, and mixing and sampling manifolds are shown
in Figures 3-4, 3-5, and 3-6, respectively. While it is often standard practice and may be
sufficient to use only inlet and outlet concentrations from a fixed-bed to determine overall
conversion efficiency, this is obviously inadequate for the determination of intrinsic chemical
kinetics. For this reason, the reactors were designed to allow for the collection of samples
from three points within the packed bed. Using data collected from these points, a more
accurate depiction of the concentration versus time profile within the bed can be discerned.
This leads to the calculation of valid intrinsic chemical kinetic parameters. An Arrhenius-type
expression was used to determine corresponding activation energies.
Sampling lines were connected to the reactor via permanently welded 316 stainless
steel compression fittings (Swagelok, Inc., Solon, Ohio) and could be easily removed between
runs. One-eighth inch diameter internal sampling lines were used to minimize the cross-
sectional area blocked by the lines which could disrupt flow through the bed. The fractional
area blocked by the cross-section of a sampling line was approximately 16 percent. A total
of 12 holes (0.1 cm diameter) were drilled in the sampling lines. As shown in Figure 3-5,
these holes were placed in three places, located at the centroids of two concentric circles of
equal area within the reactor diameter. Each sampling location was from three sets of two
pairs of holes, the latter offset by 90 from each other.


98
general, higher N02 concentrations (> 100 ppm) were removed more efficiently and more
rapidly than lower N02 concentrations, particularly at elevated temperature. This finding
validates the use of a concentration-dependent chemical reaction model. The higher
concentrations of N02 may provide more opportunity for contact with active sites on the
sorbent surface area at which the chemical reaction(s) take place. In some cases, this may
mean that the external sorbent particle surface becomes covered by an "inert" (Mg(N03)2)
product layer, leading to rate limitation by a diffusion-controlled process following the
shrinking unreacted core mechanism. A higher concentration gradient would provide a driving
force to facilitate pore diffusion and diffusion through the inert (Mg(N03)2) product layer to
available remaining reactive MgO surface area in the interior portions of the sorbent particles.
Representative N02 penetration data presented in Figure 4-8 shows that at a reaction
temperature of473 K, initial N02 penetration was low (less than one percent), suggesting that
the N02 removal efficiency was greater than 99%, for all concentrations evaluated. After an
initial conditioning period of approximately 10 minutes, there is a distinct diversion between
the curves representing 20 and 50 ppm and the curves representing 100 and 200 ppm. N02
penetration increased to approximately 45-50% for the latter, while it remained at
approximately 25% for the former. Also, the rate of change of penetration was significantly
higher for the higher N02 concentrations. Kinetically, the fast initial reaction suggested that
significantly more N02 was removed at higher concentrations in a shorter period of time than
at lower concentrations. It is possible that the decreases in N02 removal rates at
approximately 25 minutes for the two higher concentrations correspond to the completion of
the "inert" surface-product-layer shell, whereas for the two lower concentrations, the exterior


Stagnant Gas Film
Figure 4-30. Graphical representation of the shrinking unreacted-core model under diffiision-through-inert-product-layer control.


162
higher temperatures would be expected. N02 penetration curves with a sharp break to a flat
curve would be expected if total pore closure occurs. However, positive slopes are seen even
after a sharp drop in reaction rate (Marsh and Ulrichson, 1985). The calcination of Ca(OH)2
to CaO results in highly porous particles with a theoretical porosity of approximately 50%
(Newton et al., 1989). Depending upon the temperature used, the actual porosity can be
lower, particularly for the calcination of Ca(OH)2. Virtually all of the surface area is in the
internal pore structure. The increase in molar volume from solid reactant to product results
in the filling of the pore structure and due to rapid reaction rates and pore diffusional
resistances, may lead to plugging of the outer CaO layers, leaving the interior unreacted.
As was previously noted, the first-order kinetic rate coefficients decreased with time
as reaction progressed. An empirical relationship between the inverse values of these
coefficients and exposure time was noticed. After an initial period of adjustment to dynamic
equilibrium conditions, a linear relationship between \/k and time appeared. Figure 4-38
displays an example of this relationship. This relationship allowed for the prediction of k-
values at an extended bed exposure time, based upon the values obtained from the first hour
of the experimental run. These predicted values were used in the first-order kinetic equation
to predict N02 concentration at any given time. Predicted N02 outlet concentration values
are compared to experimental values for a long-term run, as shown in Figure 4-39.
To test the validity of t values predicted by the shrinking unreacted-core equations,
the N02 penetration (or fractional outlet concentration) versus time curves are useful. Based
upon the shrinking unreacted-core model, "complete" conversion would be predicted to occur
at approximately 180 minutes (one-half of approximately 360 minutes). Looking at Figure


N02 Penetration (Fraction)
Figure 4-12. N02 penetration vs bed exposure time (T=423 K, [O2]=10%).
o
u>


148
The rate equations obtained from the shrinking unreacted-core model can be used
together with the fixed-bed model, but simple analytical solutions are hard to find and the
determination of rate constants by the method of least squares is not possible. Some type of
numerical procedure has to be used instead (Bjerle et al., 1992). The constant-pattern
assumption allows the use of the relationship
X =
(4-35)
Substituting the quantity (C/Ca) for X in the shrinking core equations allowed for the
evaluation of their applicability to the experimental data collected. Karlegard and Bjerle
(1994) used this approach in their study of the removal of H2S by zinc ferrite. Other notable
examples where the quantity (C/CQ) was used as an indication of solid conversion are the
work of Wolff et al. (1993) who used Y-Al203-supported CaO to remove gaseous S02, and
Dam-Johansen and Ostergaard (1991a), who examined the S02-limestone reaction.
Application of Shrinking Unreacted-Core/Grain Model to Experimental Data
To determine the applicability of the shrinking unreacted-core model to the
experimental data collected in this study, the respective governing equations were evaluated
for various stages of the reaction. Three examples were chosen for presentation to show the
range of applicability of the model to the reaction between N02 and MgO. As an intermediate
situation, the removal of 50 ppm N02 at a reaction temperature of 423 K was chosen. For


1 (1-X) 1/3 1 3(1-X) 2/3 + 2(1-X)
A 20ppm-R x 20ppm-D
Figure 4-40. Shrinking unreacted-core model evaluation ([NO2]in=20 ppm, T=373 K, [O2]=0%).
o\
On


90
70
50
30
10
0
1
MgOVermiculite Particle Size, Dp (mm)
¡-probability plot of MgO-vermiculite sorbent particle size distribution.


147
adsorption isotherms are "favorable and exhibit the non-linear Langmuir form. A favorable
isotherm can be assumed for a single adsorbate with a high adsorption capacity, whose
concentration decreases in the downstream flow direction (Frey, 1992). The asymptotic limit
is reached when the fluid-phase front matches the adsorbed-phase front, and
£.
a C (4-32)
* o o
in which q represents moles adsorbate sorbed per mole of reactant in the sorbent, qa
represents the largest number of moles of adsorbate that can be sorbed per unit mole of
sorbent, C represents outlet gas concentration, and Ca represents inlet gas concentration.
Constant-pattern behavior is the opposite limit to "unfavorable" or "proportional-pattern"
behavior, in which the concentration front becomes diffuse and extended by mass transfer
with the solid.
The differential-concentration driving force allows for mass transfer between phases.
In the fixed-bed model, the simplified differential mass balance is represented by
-rate
P b dq
e dt
V
(4-33)
where ev represents the void fraction of the bed. Correlation of this model with the shrinking
unreacted-core model, using the assumption of constant-pattern behavior, produces
P b dg ^ dX_
e dt 0 dt
V
-rate
(4-34)


34
where A and B are gaseous and solid reactants, respectively, C and D are products, and a,b,c,
and d are stoichiometric coefficients. This overall reaction may actually comprise several
sequential or simultaneous steps, of which one is rate-controlling. The three basic steps in
the generalized non-catalytic gas-solid diffusion-reaction process are
1) Gas-phase mass transfer via diffusion of the gaseous reactant from the bulk gas
stream to the external surface of the solid sorbent particle.
2) Gas-solid interaction on or within the solid sorbent particle through
a) Diffusion of the gaseous reactant into the pores of the solid complex,
which could consist of a combination of solid reactants and products.
b) Adsorption of the gaseous reactant on the surface of the solid complex.
c) Chemical reaction at the surface.
d) Desorption and diffusion of gaseous products, if any, from the surface and
out of the pores of the solid complex.
3) Gas-phase mass transfer of any gaseous products from the external solid surface
into the bulk gas stream
The number of diverse processes and steps involved makes the analysis of gassolid
reactions a potentially unwieldy problem. In general, each step provides resistance to
chemical reaction, and these resistances are additive. When the processes occur in series, it
is necessary to determine which step provides the major resistance to reaction, and ultimately
controls the overall reaction rate. The resistances of the different steps can vary greatly from
each other, so the step with the highest resistance can be considered to be the rate-controlling


190
of intrinsic kinetic data would be extremely useful in this regard. These data would also be
useful in determining the true limits of MgO utilization by N02 in the sorbent, and would most
likely make the shrinking unreacted-core model results easier to interpret. The high
temperature limit appears to be between 523 and 573 K, at which magnesium nitrate
decomposition is significant.
5. Obviously, there are other exhaust gas components whose presence may affect
the N02 removal characteristics of this sorbent. These other components should be evaluated
to determine their respective roles in the N02 removal process. One major component
includes the combustion aerosols, which would probably accumulate on the surface of the
sorbent, possibly making MgO on the exterior surface of the particle unavailable for reaction.
Some of the carbonaceous aerosols may themselves remove N02. The effect of this added
layer on the effective diffusivity of N02 inward toward unreacted MgO inside the particle
would also need investigation.
6. More tests at all operating conditions, but at longer reaction times, should be
conducted to provide more information on reaction rates and final conversions. Fully
automated valve-switching and data-collection systems would be very useful when
performing long-term runs. The predictability of N02 penetration as it relates to MgO
conversion is useful information in the design of any application that might employ this
sorbent medium.
7. The role of water and water vapor needs further study to validate the findings
of this research. It is essential to know if the improved N02 removal capabilities found in the


- 3(1-X)2/3 + 2(1-X)
Figure 4-35. Shrinking unreacted-core product-layer-diffusion-control-equation evaluation, xd=957 minutes ([NO2]in=50 ppm,
T=423 K, [O2]=0%). [ y = O.OOlx + 0.003, r^O.98, Fit Std Error = 0.0006, F=814.8],
k/i


+ Experimental Predicted
Figure 4-39. Experimental versus predicted N02 concentration ([N02]in=47.5 ppm, T=423 K, [O2]=0%, Gas samples collected
from middle of bed).


180
irregularly shaped layers and may show more or less diffusional resistance as the lattice
expands with reaction, forming a complex structure. Bamford et al. (1984) noted that lattice
structure changes often associated with gas-solid reactions sometimes effectively eliminate
oxygen vacancies, which have been reported to be relevant to the chemical reactivity of MgO.
It must be emphasized that while these models seem to have some utility in providing
a chemico-physical basis for describing N02 removal by the MgO-vermiculite sorbent under
the range of experimental conditions studied, this applicability is limited by the experimental
constraints placed upon these variables. Referring to the sulfation of CaO, Wen and Ishida
(1973) stated that various kinetic studies indicate that reaction rates can vary significantly
depending upon the type of limestones used, sometimes more than an order of magnitude.
It is likely that MgO powder dispersed within a packed bed, or individual pure MgO pellets
would behave quite differently from the combined MgO-vermiculite sorbent particles studied.
The low N02 concentrations used in this study probably have obscured the distinction
between the controlling mechanisms, often making it difficult to absolutely distinguish one
rate-determining mechanism from another.
It is apparent that the processes of pore closure, product-layer diffusion, and product
sintering are important factors in this study, whose combination makes data analysis very
difficult. To obtain a better understanding of how these mechanisms ultimately affect
conversion-versus-time behavior, more information is required on the solid structure and how
it changes during the course of reaction (Marsh and Ulrichson, 1985). Because of the high
diffusional resistance of the product layer, the concentration of gaseous reactant at the
reaction interface, and consequently the reaction rate, may not be uniform even under the


CHAPTER 3
EXPERIMENTAL METHODS AND MATERIALS
General Research Approach
A fixed-bed reactor system was designed, constructed and used to collect the relevant
data to meet the established research objectives. The present study systematically addressed
individual exhaust gas components, focusing primarily on NOx, under strictly controlled
conditions to delineate the nature of the sorption rates and mechanisms. Processes and rates
for removal of N02 and NO by MgO-vermiculite sorbent were independently evaluated.
Moistened sorbent and/or humidified gases were used to evaluated the NOx removal
characteristics of Mg(OH)2. A variety of test conditions were used to determine important
operating parameters and limitations. By combining gaseous components, their interactions
in the presence of the sorbent material could be evaluated. Intrinsic and overall kinetic
parameters were determined. A complete description of the kinetics of gassolid reactions
is very complicated, especially for systems involving solid products, for which the processes
at both the gas-solid interface and the reaction interface between the reactant and product
solids must be considered. Although the rate expressions can be quite complex, the reactions
between adsorbate and adsorbent can often be described by first-order kinetics (Szekely et
al., 1976). Data were mathematically interpreted using the shrinking unreacted-core model
49


68
Engineering, Inc., Stamford, Connecticut) and meter were attached to the outside of the
reactor.
Experimental Procedures
The packed reactor was placed inside the tube furnace with the bed centered within
the furnace and the tube element mounted vertically on the furnace control base. The vertical
arrangement was chosen to optimize gassolid contact by avoiding the potential for solid
settling and channeling of the gas stream through the bed volume. The middle 15 cm of the
furnace was reported to be completely temperature-controlled by the manufacturer, providing
a sufficient length for isothermal experimental conditions. Once the reactor was in place,
sampling lines were individually connected to two-way stainless steel valves (Model W-l 188,
Whitey Co., Highland Heights, Ohio) at the sampling manifold. Gas lines entering the tube
furnace were wrapped in high-temperature heating tapes. These had been previously
calibrated using a Type K thermocouple to measure surface temperature at various settings
on the variable voltage controller. Depending on the reaction temperature of interest, the heat
tape voltage controller was set to produce a gas preheat temperature approximately 20 C
above reaction temperature. The tube furnace was turned on and allowed to heat up for
approximately one hour to achieve temperature control, while the bed was simultaneously
being flushed with approximately 2 sLpm dry N2. After this conditioning period,
thermocouple readings for the tube furnace and independent reactor thermocouple were


196
Krishnan, S. V. and Sotirchos, S. V., "A Variable Diffusivity Shrinking-Core Model and its
Application to the Direct Sulfation of Limestone," The Canadian Journal of Chemical
Engineering. 71:734-745, 1993a.
Krishnan, S. V. and Sotirchos, S. V., "Sulfation of High Purity Limestones Under Simulated
PFBC Conditions," The Canadian Journal of Chemical Engineering. 71:244-255, 1993b.
Krishnan, S. V. and Sotirchos, S. V., "Experimental and Theoretical Investigation of Factors
Affecting theDirect Limestone-H2S Reaction," Industrial Engineering Chemistry Research.
33(6): 1444-1453, 1994.
Lange's Handbook of Chemistry, 11th ed., McGraw-Hill Book Company; New York, p 4-88,
1973.
LeVan, M. D., "Fixed-Bed Adsorption of Gases: Effect of Velocity Variations on Transition
Types," AIChE Journal. 34(6):996-1005, 1988.
Levenspiel, O., Chemical Reaction Engineering. 2nd ed., John Wiley and Sons, Inc., New
York, 1972.
Lewis, W. H., Nitrogen Oxides Removal. Noyes Data Corporation; Park Ridge, New Jersey,
1975.
Lyon, R. K., "Thermal DeNOx: Controlling Nitrogen Oxides Emissions by a Noncatalytic
Process," Environmental Science and Technology. 21 (3):231-236, 1987.
Lyon, R.K., New Technology for Controlling NO., from Jet Engine Test Cells.
ESL-TR-89-16, HQ AFESC, Tyndall AFB, Florida, 1991.
Marsh, D. W. and Ulrichson, D. L., "Rate and Diusional Study of the Reaction of Calcium
Oxide with Sulfur Dioxide," Chemical Engineering Science. 40(3):423-433, 1985.
Mejias, J. A.; Marquez, A. M.; Fernandez Sanz, J.; Fernandez-Garcia, M.; Ricart, J. A.;
Sousa, C.; and Illas, F., "On modelling the interaction of CO on the MgO(lOO) Surface,"
Surface Science. 327:59-73, 1995.
Meubus, P., "Catalytic Decomposition of Nitric Oxide in the Presence of Alkaline Earth
Oxides," J. Electrochem. Soc.: Electrochemical Science and Technology. 124(l):49-58, 1977.
Neal, D. D.; Hoke, S. H; and Spencer, W. P., NO.c Sources. Properties and Analytical
Procedures. Technical Report 8907, U.S. Army Biomedical Research and Development
Laboratory, Fort Detrick, Maryland, 1981.


N02 Penetration (Fraction)
Figure 4-9. N02 penetration vs bed exposure time (T=423 K, [O2]=0%)
VO
Ov


70
Gas Analyses
All reactions were monitored by selectively removing gas samples from the stainless
steel tubular reaction vessel through selected valves connected to the sampling manifold
system. NOx concentrations were analyzed using a chemiluminescent analyzer (Model 42H,
Thermo Environmental Instruments, Franklin, Massachusetts) based upon the reaction of NO
with 03, following reduction of N02 to NO. The feasibility of the commercial
chemiluminescent analyzer was established in 1970 by Fontijn et al. and the first prototype
was developed in 1972 (Steffenson and Stedman, 1974). The method is now widely used,
in fact, the use of a chemiluminescent method is required for the measurement of N02 in the
atmosphere (40 CFR, Part 50, Appendix F). The principle of operation of the
chemiluminescent analyzer is based upon the chemiluminescent reaction between NO and
ozone (03). This reaction is as follows:
NO + O3- [N02* + 02] -N02 + 02 + hv (3-1)
where hv, which represents the quantity of light energy emitted by an electronically excited
intermediary (N02), is simply Planck's constant (h) multiplied by the frequency of the
emitted radiation (v), in this case with a wavelength greater than 600 nm. A sample of gas
is drawn from a selected point in the system by an external pump. The sample is mixed with
ozone generated by an internal ozonator inside a reaction chamber. The resultant reaction
produces a characteristic luminescence with an intensity directly proportional to NO
concentration, since 03 is provided in excess, and its concentration can be considered a


187
active surface area coverage by the product, which improved sorption performance with
time.
6. The presence of oxygen in the exhaust gas aided in the removal of NO by the
oxidation of NO to N02. It appeared that oxygen may have promoted the continued removal
of N02 by the sorbent.
7. Temperature affected the reaction by increasing reaction rates both initially
(chemical-reaction control) and by improvingeffective diffusivity. Reaction temperature must
be kept below 523 K, since this is the temperature at which Mg(N03)2 begins to decompose.
8. The pressure drop versus superficial velocity relationship can be described
using the Ergun equation. This equation should be used to design any system in which this
sorbent will be employed since pressure drop is a limiting factor in a jet engine test cell
application.
9. It is possible to regenerate the sorbent by heating for reuse. The initial N02
removal efficiency of regenerated sorbent is similar to fresh sorbent, but the extended
performance decreases with time, probably the result of structural changes (i.e., reduced
porosity/surface area, sintering, etc.).
10. A reaction mechanism has been proposed to explain the production of NO
during the removal of N02. NO may be produced during the final oxidation step of mixed
nitritenitrate product to its final nitrate form. After the N02 removal rate decreases, NO
production decreases accordingly.
11 The shrinking unreacted-core model was applied to N02 penetration data using
the local-equilibrium theory to estimate solid-sorbent conversion. The model described the


Generalized Non-Catalytic Gas-Solid Reactions 33
Mathematical Models 35
Progressive-Conversion Model 36
Shrinking Unreacted-Core Model 36
Grain Model 39
Analagous Gas-Solid Reaction System Research 41
S02CaO Reaction 42
S02-Calcium Hydroxide Reaction 43
S02-CalciumCafbonate Reaction 44
Other Acid-gas-Calcium-based-solid Reactions 45
General Applicability of the Shrinking Unreacted-Core Model 46
Research Justification 47
Research Objectives 47
3. EXPERIMENTAL METHODS AND MATERIALS 49
General Research Approach 49
Experimental Variables 50
MgOVermiculite Reactive Sorbent Material 50
Temperature 52
Pressure 52
Gas Composition 53
Gas Flow Rate(s) 54
Gas Reaction/Sorption/Desorption 55
Fixed-Bed Reactor System 55
General Experimental Considerations 55
Experimental Arrangement 58
Tubular Fixed-bed Reactors 58
Experimental Procedures 63
Gas Analyses 70
Sorbent Surface Area Determination 76
Sorbent Particle Size Determination 77
4. RESULTS AND DISCUSSION 78
Intrinsic Kinetic Studies 78
Limitation of Gas-Film-Mass-Transfer Resistance 79
Solid Sorbent Characterizations 82
Surface Area Analyses 82
Particle Size Distributions 83
NOx Removal by Mg(N03)2-coated Vermiculite 86
NQX Removal by Sorbent Material 86
NO Removal by Sorbent Material 87
N02 Removal by Sorbent Material 89
Effects of Operational Variables 94
v


183
saturated (170 ppm N02, 473 K) the sorption capacity of the sorbent (after 5 hours) was
approximately 40 mg N02/3.5 g sorbent bed. This corresponds to a mass loading ratio of
approximately 1.14 x 10'2 g N02 per gram sorbent, at an overall rate of 2.3 x 10'3 g N02 /g
sorbent-hour. While initial N02 removal was greater than 90%, the average removal over the
five-hour period was approximately 40%. Taking as an example a 10,000-acfm source of 200
ppmN02 at 473 K, the volume of bed required to treat this source, with all other conditions
remaining the same, and taking the bed to saturation after one hour, is approximately 3 m3,
not an inordinately large size. Obviously, it is not practical to regenerate or replace the bed
every hour, and the volume of gas to be treated may be larger, so a larger bed would be used.
As was already pointed out, jet engine test cells are not steady state sources and the average
test run may be much shorter than an hour in duration. After jet engine exhaust is augmented,
the concentration of N02 would also likely be less than 200 ppm, so the bed would last longer
because it takes longer to form the product layer when reacting with lower concentrations of
gases. Lower N02 concentrations will produce lower removal rates, but the bed will last
longer. When lower-concentration gases are expected, elevated temperatures can help to
improve difftisivity through the product layer in the absence of a large concentration gradient
driving force.
The effective conversion of NO to N02 is a crucial element in the eventual successful
employment of this NOx control medium. Spicer et al. (1990) noted that exhaust
augmentation can increase the NOz-to-NO ratio from 1 or 2:10 to 7 or 8:10. While they
attributed some of this oxidation to the presence of peroxy radicals in fuel-rich combustion,
oxygen oxidation can be significant as well. Obviously there are a number of other variables,


Figure 4-3. Log-probability plot of Akrochem Elastomag 170 MgO powder particle size distribution (manufacturer-provided
data).
00


112
required to maintain a reasonable amount of N02 removal from the gas stream. When oxygen
was present in the system, it had a less noticeable effect when comparing the temperature
effects for a given inlet N02 concentration as can be seen in Figures 4-18 through 4-20.
Again, initial N02 penetration was lowest at the highest temperature for all concentrations,
but it increased more noticeably for the higher concentrations. For an inlet N02 concentration
of 20 ppm, penetration stayed below 10% for an exposure time of nearly an hour.
Activation Energy Determination
To determine experimental activation energies, Arrhenius plots of the natural
logarithm of the bed-exposure-time-specific first-order rate coefficients versus the reciprocal
of the absolute reaction temperature were prepared. The rate of removal of N02 was
sometimes less temperature-dependent between 373 and 423 K, as seen in Figure 4-21.
When this was the case, data were collected at an intermediate temperature of approximately
450 K, since the plots were often more linear over the higher temperature range. A good
example of linearity in the plot between 473 K and 373 K can be seen at Figure 4-22. From
the slopes of these Arrhenius plots, the apparent activation energies can be determined. These
were approximately 33 kJ/g-mol at 5 minutes bed exposure time and about 8 kJ/g-mol at 20
minutes bed exposure time for the example shown in Figure 4-21. The decrease in activation
energy with bed exposure time suggests that a change in N02 removal mechanism occurs with
increasing bed exposure time. It is common, however, for Arrhenius plots of heterogeneous
reaction parameters to show curvature (Clark, 1970).
Initial activation energies were often difficult to determine due to system fluctuations
during the first 1 to 2 minutes of operation. This instability probably caused some of the


57
chemico-physical model to adequately explain the removal of gaseous contaminants. A means
for accomplishing this correlation using local equilibrium theory and assumptions regarding
constant-pattern behavior for mass transfer in a fixed-bed will be detailed later.
The basic principles that govern chemical kinetics and diffusion phenomena in single
particles are normally unaffected by the presence of other particles in an arrangement like a
packed bed. Changes in the flow field due to the packed bed arrangement likewise do not
invalidate these principles (Szekely et al., 1976). In this way, a fixed-bed experimental
system may be used to extend the information obtainable from single particles to multiple-
particle systems. Multiple-particle systems are obviously of more interest since practical
systems must be of this type. Since the sorbent material is highly heterogeneous, data
reproducibility between samples is very important. This was evaluated by comparing results
for the same experimental conditions between different samples of sorbent material.
While the basic principles of single gas-solid interaction will still apply, the physical
characteristics of any experimental fixed-bed reactor can markedly affect chemical kinetics
and overall sorption behavior. It is critical that such a system be carefully designed, to ensure
that any experimental data collected are scientifically valid and defendable. Often, practical
constraints such as temperature and, particularly, gas velocity and its effect on pressure drop
through the bed, must be incorporated into the experimental design. Such other factors as
particle fluidization velocity and the potential for temperature and concentration gradients
must be accounted for as well. In this research, MgO will be attached to its vermiculite
substrate. Results, however, may be different from those for individual MgO particles due
to this arrangement.


3
The most harmful effects of NOx are primarily attributed to exposure to N02 the most
toxic of these oxides. N02 is a reddish-brown gas that is quite visible in sufficient
concentration (> 0.25 ppm). N02 is not a primary pollutant in the sense that it directly affects
human health, unless the exposure concentration is high. N02 exerts its main toxic effects on
the lungs via free-radical-mediated reactions and other mechanisms. Both NO and N02 are
free radicals that may produce lipid peroxidation reactions within human cells. The
environmental hazard of N02 is primarily associated with the pulmonary effects of the
pollutant. Exposure to concentrations ofN02 from 0.7 to 5.0 ppm for 10 to 15 minutes have
produced abnormalities in pulmonary airway resistance. Exposures to 15 ppm can cause eye
and nose irritation, and pulmonary discomfort is noted at 25 ppm for exposures of less than
one hour.
The greatest danger of exposure to N02 is the delay in experiencing its full effect upon
the respiratory system. The delayed effect of N02 injury is made potentially more dangerous
by two other factors. Human perception of the odor of N02 is insufficient to warn against
injury or even death. Additionally, N02 quickly desensitizes an individual to its odor through
olfactory fatigue, and if N02 levels gradually rise, a person could unknowingly be exposed
to concentrations high enough to cause permanent injury or death. Because of the low
solubility of N02 in water, it is only slowly removed from the lungs by circulating blood and
may remain in contact with lung cells for prolonged periods of time. Given the same total
dose, short-term exposure to high concentrations of N02 is more injurious than long-term
exposure to lower concentrations. The toxic effects of N02 are often synergistic with or


133
3) N02 + MgN205 Mg(N03)2 + NO
It should be noted that N02 is functioning as an oxidizing agent in the final step, to complete
the formation of Mg(N03)2, in the absence of oxygen. If oxygen were present, then it may
competitively complete this last step, reducing NO formation. When Mg(OH)2 is the solid
reactant, the mechanism remains essentially the same, except that water may be liberated
during the first step in forming MgN03. This first step is presumed to be the rate-limiting step
since the other two steps are expected to be much faster.
From a comparison of the ratio of the quantity of N02 removed to NO produced as
reaction progressed, some support for this mechanism was obtained. Initially, the absolute
value of this ratio was very high, with much more N02 being removed than NO being
produced. As reaction progressed, however, this ratio often decreased to a point where 2 to
3 moles of N02 were removed for every mole of NO produced. This value is appropriate
given the proposed mechanism just shown. The time when this ratio was achieved was
normally toward the end of the one hour experimental period. It is likely that even after the
reaction becomes product-layer-diffusion-controlled, the other sequential steps are still
occurring so that N02 removal by the overall process is still apparent.
Application of the Shrinking Unreacted Core/Grain Model
The shrinking unreacted core model contains parameters that are useful for predicting
the behavior of gassolid reactions. Incorporating such physical parameters as the molar
volume ratio of products to reactants (which can affect product layer formation and


41
describing the observed discrepancy between theoretical and observed solid conversion is the
effective difiisivity through the product layer. Incomplete conversion, which decreases with
an increase in intrapellet diffusional resistance, is predicted using the random pore model
developed by Bhatia and Perlmutter (1981a). A unique feature of this model is that solid
pores are not assumed to be of uniform size, so smaller pores are more easily blocked when
a dense or large-volume product is formed on the surface. Dam-Johansen et al. (1991) used
physical parameters of a chalk sorbent to modify the grain model down to the micrograin
level. While this provides a physical description of the reaction occurring at the sub-grain
level, the reaction taking place still follows the shrinking unreacted-core mechanism.
Analogous Gas-Solid Reaction System Research
The most similar gassolid reaction systems discussed in current literature describe
the reactions between sulfur dioxide (S02) and lime/calcium oxide (CaO), calcium hydroxide
(Ca(OH)2), or limestone (calcium carbonate (CaC03)). Reactions of acidic gases with
calcined limestone have been closely studied because of their industrial importance. The high
porosities of various forms of CaO (nominally over 50 percent) makes them quite suitable for
tests of the numerous forms of the shrinking unreacted-core model. Although these models
have been used quite successfully, the fundamental rate-controlling processes are still not well
understood (Borgwardt et al., 1986). Data on reactions between other acid gas/acid
anhydride gases and other carbonate rocks are also available. In these processes, S02-
containing gases are contacted with lime or limestone, either as a wet slurry or dry solid. S02
is generally collected on the solid surface for reuse or disposal. Regeneration may produce


60
is a problem with attempting to use "pure" N02 (or NO) in gas-solid sorption studies
employing cylinder gases. These gases were diluted with dry air or prepurified N2 to the
appropriate concentration(s). Cylinder gas flow rates were measured using precision
rotameters (Omega Engineering, Inc., Stamford, Connecticut) with only glass and/or stainless
steel wetted parts. While the manufacturer provided calibration data, all rotameters (both
high- and low-flow-rate tubes) were independently calibrated in the laboratory using a primary
airflow standard (Gilibrator, Gilian Instrument Corporation, Wayne, New Jersey). These
calibrations were extremely important for accurate gas flow rate measurements, particularly
when using low-flow-rate tubes, which have non-linear calibration curves. Good agreement
was reached between the calibration curves, as can be seen in the example comparative
calibration curves, Figures 3-2 and 3-3. Laboratory calibration curves were used to measure
gas flow rates in this study.
All system components were made of either 316 stainless steel or Pyrex glass to
reduce the potential for NOx removal by system reactivity. Inert gas transfer lines were made
from Teflon tubing of minimal lengths to minimize transfer losses. Most connections were
through 316 stainless steel compression fittings (Swagelok, Inc., Solon, Ohio). Where pipe
thread connections were necessary, Teflon tape was used to prevent leakage. Before
experiments were initiated, the entire system was leak-checked and NOx decay (reactivity)
measurements were made for the system. Each system component was isolated and evaluated
separately. Overall NO and NQ2 removal as a result of system reactivity were estimated at
less than one percent. In this way, any changes in NOx concentration with time could be
attributed solely to the presence of the sorbent material in the fixed-bed reactor.


(x-0-
Figure 4-45. Shrinking unreacted-core chemical-reaction-control-equation evaluation, t=495 minutes ([N02]in=200 ppm,
T=473 K, [O2]=0%). [ y = 0.002x + 0.035, ^=0.94, Fit Std Error = 0.011, F=159.7],
-o
to


30
Field Studies
Pilot-scale tests of this MgO-vermiculite sorbent were conducted on a JETC facility
at Tyndall AFB, Panama City, Florida, between 9-11 June, 1992 and on 18 September, 1992.
A two-stage filter bed design, consisting of four inches of virgin vermiculite in front of eight
inches of MgO-coated vermiculite, was evaluated through four approximately 20 minute runs
of subscale drone engines. An additional run was conducted through four inches of activated
carbon, followed by the eight-inch MgOvermiculite bed. NOx removal efficiencies for the
first four tests were approximately 50 percent. This increased to approximately 80 percent
with the addition of the activated carbon bed. Between June and September 1992,
approximately 150 unmonitored engine tests took place. When the September 1992 test was
conducted, NOx removal had decreased to 60 percent. Gas temperatures, velocities, and
pressure drop through the beds were also measured. The Air Force sponsor had imposed a
pressure drop limit of two inches of water on the bed, which was not exceeded. However,
approximately two-thirds of the volumetric flow was allowed to bypass the filter to prevent
overheating inside the JETC facility. Results showed that significant amounts of unburned
(non-methane) hydrocarbons (primarily ethylene) passed through the beds that did not contain
activated carbon (Nelson et al., 1992; Wander and Nelson, 1993).
More regeneration studies were conducted by Nelson et al. (1990) to attempt to
optimize the regeneration of MgOvermiculite after "saturation" with NOx (and S02). The
sorbents were alternately exposed to simulated jet engine exhaust and regenerated in reducing
atmospheres for a number of treatment cycles to evaluate NOx sorption capabilities.


181
i
uniform temperatures and concentrations in the porous network. Structural properties
depend not only on conversion, but also the relative rates of reaction (intrinsic kinetics) and
mass transport in the product layer (Sotirchos and Yu, 1985). Gullett et al. (1992) compared
the magnitude of the sum of squared errors as well as the coefficient of determination in
defining the best-fitting shrinking-core-model equation in the reaction between HC1 and CaO.
They found that often either the chemical-reaction-control model or the product-layer-
diftiision-control-model is sufficient to describe sorption, without the inclusion of both models
simultaneously. The two main governing equations for chemical-reaction-control and
product-layer-diffusion-control are themselves not uniquely distinct, particularly when the
true solid sorbent conversion level is low. Therefore, care should be taken in using these
modeling results to quantitatively derive physical parameters of the sorbent under the
conditions evaluated, or in their extrapolation to different conditions under which MgO solids
may be used to remove N02 from a flowing gas stream.


Effects of N02 Concentration 94
Effects of Oxygen 99
Effects of Bed Temperature 105
Activation Energy Determination 112
Effects of Water 120
Effects of Residence (Reaction) Time 122
Sorbtech-Supplied Sample Results 123
Pressure Drop Characteristics 124
Magnesium Nitrate Surface Decomposition 127
Sorbent "Lifetime" and Regeneration 130
Proposed Reaction Mechanism 132
Application of the Shrinking Unreacted Core/Grain Model 133
Diffusion-Through-Gas-Film-Control 134
Chemical-Reaction-Control 137
Diffiision-Through-Inert-Product-Layer-Control 140
Derivations of the Shrinking Unreacted-Core Model 144
Correlation Between Gas-Phase Concentration and Solid Conversion 146
Local-Equilibrium Theory 146
Application of Shrinking Unreacted-Core/Grain Model to Experimental Data 148
50 ppm N02 at a Reaction Temperature of 423 K 149
20 ppm N02 at a Reaction Temperature of 373 K 165
200 ppm N02 at a Reaction Temperature of 473 K 169
General Applicability of the Shrinking Unreacted-Core Model 178
5. PRACTICAL CONSIDERATIONS 182
Constraints on MgO-Vermiculite Sorbent Usage 182
6. CONCLUSIONS AND RECOMMENDATIONS 185
Conclusions 185
Recommendations for Further Research 188
LIST OF REFERENCES 192
BIOGRAPHICAL SKETCH 201
vi


149
the evaluation of the extreme experimental conditions, the removal of 200 ppm N02 at 473
K and the removal of 20 ppm N02 at 373 K were evaluated. These three cases contain data
collected in the absence of oxygen, using dry sorbent (no Mg(OH)2). The relevance of the
shrinking unreacted-core model to these situations will also be generally discussed. Each case
will be presented separately, with a discussion of the specific applicability and considerations.
50 ppm NO-, at a Reaction Temperature of 423 K
Figure 4-31 represents the mathematical transformation of solid sorbent conversion
data (X or N02 penetration) into the values obtained from the shrinking unreacted-core
equations. The left ordinate shows the chemical-reaction-control-equation values (Equation
4-18) and the right ordinate indicates the product layer diffusion control equation values
(Equation 4-28). If an equation is representative of the mechanism occurring, then the plot
will be linear with a slope of the inverse time to complete reaction, tr or rd, as previously
discussed. From this figure, it visually appears that over the entire course of reaction
evaluated, the reaction-control-equation line shows noticeably greater curvature than the
product-layer-diftusion-control-equation. To apply some statistical significance to the fit of
these equations at various stages of the reaction, linear least-squares regressions of the data
were performed, including the determination of 95% confidence intervals.
The two equations are separately evaluated as shown in Figures 4-32 and 4-33. From
Figure 4-32, the equation of the least-squares linear regression line for the chemical-reaction-
control-equation is y = 1 = 0.0025t + 0.0223, with an r2 value of 0.95. From this


Ln Instrument [02] (Percent)
+ Mid Scale High Scale
Figure 3-8. Typical IR-2100 oxygen analyzer calibration curve.


BIOGRAPHICAL SKETCH
Major Larry Thomas Kimm was born in Mount Clemens, Michigan, on November 6,
1960. After graduation from Fraser High School, in Fraser, Michigan, in 1978, he attended
the U.S. Air Force Academy at Colorado Springs, Colorado, until his graduation and
commissioning as a second lieutenant in the U.S. Air Force in 1982. He received a Master
of Science degree in biology (industrial hygiene) from Old Dominion University, Norfolk,
Virginia, in 1988. He was competitively selected to continue his education under the USAF
Biomedical Sciences Corps advanced education program, and came to the University of
Florida and the Department of Environmental Engineering Sciences to begin his Ph.D. studies
in August 1991. Serving as abioenvironmental engineer in the U.S. Air Force, Major Kimm
returns to his career after receiving his degree. He has been married to the former Lisa Anne
Wilson since 1987, and is the father of two children, a daughter, Meredyth Grace, and a son,
Wilson Carl.
201


184
including the presence of other exhaust gas components, that need investigation before one
can truly predict what may happen on the sorbent surface.
As was noted by Wander and Nelson (1993), bed pressure drop is a potentially
significant limiting factor in the use of the MgO-vermiculite sorbent in a packed-bed scenario.
Nelson et al. (1993) evaluated the potential for NOx removal in a variety of flow-through bed
arrangements in which the gases would make indirect contact with the sorbent at the wall of
the container rather than physically passing through it. This did not work well, so it appears
that direct contact is necessary for the removal of N02 from the gas stream. The Ergun
equation describes the pressure-drop-versus-superficial-velocity relationship, so it can be
used in the design of a packed bed in keeping pressure drop below established limits. An
empty-bed residence time of approximately one second is recommended for efficient N02
removal, as well as in the minimization of pressure drop through the bed, although reduced
times still allow for some removal of N02. The physical dimensions of a unit designed for the
control of nitrogen oxides from a jet engine test cell must be based upon these constraints in
order to effectively remove the nitrogen oxides while minimizing the negative impacts on the
jet engine testing operations that are producing the exhaust to be treated.


161
Most experimental work published on CaO-S02 reactions was conducted using S02
concentrations in the tenths of a percent, one to two orders of magnitude greater than N02
concentrations evaluated in this research. It is noted that Equation 4-36 does not have a gas-
phase concentration term. At a given conversion level (or product layer thickness), Marsh
and Ulrichson (1985) found that doubling the bulk gas-phase adsorbate concentration should
double the overall reaction rate, consistent with the idea of a product layer surrounding the
active surface area of the oxide. Another example is the degree of zinc ferrite utilization in
the removal of2,600 ppm H2S at 500 to 600 C, as reported by Karlegard and Bjerle (1994),
which varied from 5 to 11%.
A significant correlation between adsorbate concentration and solid conversion in the
reaction between S02 and CaO was reported by Simons (1988). Simons reported a
utilization factor of approximately 0.6 for this effect, meaning a ten-fold increase in
concentration produced a six-fold increase in effective utilization. Based upon this
correlation, a ten-fold lower concentration would predict an Xmax value of approximately 3%,
and a hundred-fold lower concentration would predict an Xmax value of approximately 0.4%.
This may partially explain the low N02 utilization rate by MgO. Since little is known about
the pore size distribution of the MgO-vermiculite sorbent, it is difficult to say what effect it
may have on sorbent utilization. However, Simons also pointed out that small-pore plugging
is dominant for 1- to 10-pm particles, while pore-mouth plugging is dominant for 100-pm
particles. Rate limitations resulting from product-layer diffusion (prior to plugging or filling)
are most evident for thick product layers occurring within the larger pores of larger particles.
If pore plugging within the grains of the particle were occurring, lower conversion levels at


27
sorbents originally designed for the control of S02, that were based upon MgO (an alkaline
earth [Ha] metal oxide) coated onto vermiculite or perlite [Nelson, U.S. Patents 4,721,582;
4,786,484, 1988; and 4,829,036 (1989)]. S02 reduction is believed to occur primarily by
reaction with MgO, forming complex sulfites or sulfates. It was assumed that NOx could be
removed in a similar manner, forming nitrites or nitrates. Sanitech researchers began
evaluating expanded vermiculite as a catalyst in 1987. A patent describing the use of
vermiculite as a selective reduction catalyst [US Patent 4,806,320] was granted to Sanitech
in 1989.
In August 1988, the Air Force awarded Sanitech a Small Business Innovation
Research (SBIR) project to further develop the new catalyst for application to jet engine test
cells. Nelson et al. (1989) reported the results of a series of tests designed to establish the
applicability and technical feasibility of using vermiculite as a catalyst to reduce NOx from
simulated jet engine exhaust. These studies, beginning as a selective-catalytic approach,
evaluated how vermiculite variables (type and size), bed size, solid additives to vermiculite,
flue gas variables (temperature, composition, flow rate) and the addition of methane or
ammonia to the exhaust gases generally affected NOx reduction.
Some initial regeneration studies were also performed. These screening studies
indicated that a MgO-coated vermiculite material showed great promise as a NOx-reducing
agent for a wider variety of conditions than encompassed by commercial catalysts. Some of
the major conclusions of this study included data indicating that vermiculite alone, without
ammonia or methane additions, demonstrates the ability to reduce NOx by 50 to 73 percent
over the temperature range of 250-850 F. The use of gaseous ammonia only marginally


145
1) Mg(N03)2-6H20 : MgO = 13.9 : 1
2) Mg(N03)2-2H20 : MgO = 8.1 : 1
3) Mg(N03)2-6H20 : Mg(OH)2 = 6.3 : 1
4) Mg(N03)2-2H20 : Mg(OH)2 = 3.7 : 1
As all of these values are noticeably greater than one, the use of the Z-term is expected to aid
in the interpretation of the experimental data collected in this study.
Szekely et al. (1976) also pointed out that the use of one or the other of the governing
equations may not be sufficient to best describe the experimental data. Since the two
equations are similar, it is often difficult to distinguish between chemical-reaction control and
product-layer-diffusion control. In some cases, assuming gas-film diffusion is not rate-
limiting, chemical-reaction and product-layer-diffusion may collectively control the reaction
rate. In such cases, a term (o2) representing the ratio of diffusional resistance to chemical-
reaction resistance can be defined to apply the relative importance of each resistance to the
reaction rate:
IcR
6D (4-30)
e
This term is applied via the equation:
W o2p(X)
x
(4-31)


146
Correlation Between Gas-Phase Concentration and Solid Conversion
To use the shrinking unreacted-core model equations it is necessary to collect data
describing solid conversion with time. This is normally done using thermogravimetric
methods on single pellets, by way of real-time surface analyses, or post-sorption chemical
analyses. As was previously explained, these data were not collected since they are less
relevant to data directly indicative of the control efficiency of gaseous NOx pollutants. It is
necessary to be able to correlate these data so that they can be used in the model equations
to interpret sorption behavior. By invocation of assumptions related to constant-pattern
behavior associated with the local-equilibrium theory of fixed-bed reactors, the previously
displayed gas-penetration-versus-time data are made more amenable for this purpose.
Local-Equilibrium Theory
While a detailed discussion of the local-equilibrium theory can be found in any
standard text on adsorption in fixed beds, the following brief presentation is based upon
descriptions by Sherwood et al. (1975) and Ruthven (1984). When there is efficient mass
transfer between the gas and solid phases, the reactive gas travels through the packed bed as
a wave without a significant change in concentration, other than that caused by reaction with
the sorbent surface. The wave front reaches a steady state and its width, as it moves through
the bed, attains a "constant pattern." The constant-pattern approach is based on the
assumption that the width of the concentration front tends toward a constant value as the
concentration wave moves through a packed bed. The mass-transfer zone reaches a stable
asymptotic form in this case (LeVan et al.., 1988). According to Suzuki (1990), most


Magnesium Nitrate Surface Decomposition
Solid Mg(N03)2*6H20 (Certified A.C.S. grade, Fisher Scientific, Pittsburgh,
Pennsylvania) was dissolved in deionized water and applied to heat-treated vermiculite, and
the mixture was allowed to air dry. The mass ratio applied was approximately 15%
Mg(N03)2. This material was packed inside a clean reactor and 2 sLpm dry N2 was delivered
to the bed. Bed outlet concentrations of NO and N02 were followed over time as bed
temperature was increased from ambient temperature to 373 K through 573 K in 50 K steps.
Results are plotted in Figure 4-25. From this figure, it can be seen that at temperatures used
for sorption, little NO and N02 were released compared to those emissions observed at much
higher temperatures, 523 and especially 573 K. In either case, the ratio ofN02/NO released
was approximately between 10:1 and 16:1. Heating samples of used sorbent to a temperature
of 523 K produced similar results (See Figure 4-26). This implies that N02 is the
predominant gaseous reactant in the forward reaction. Also, it is important to recognize that
the decomposition of Mg(N03)2 is not a significant source of NO. Although it is obviously
necessary to control NO emissions, these studies imply that NO production observed in N02
sorbent removal studies must occur through a mechanism other than decomposition of
Mg(N03)2.


Air Flow (mL/min)
Rotameter Reading
Omega Data I Lab Calibration
Figure 3-3. Example high-flow-rate rotameter calibration curves (Omega NO92-04G Tube).


Table 4-2. Intrinsic first-order rate coefficients (s'1) for N02 removal by MgO-vermiculite sorbent ([O2]=0%)
N02 Concentration:
Reaction Temp ( K):
473
200 ppm
423
373
473
100 ppm
423
373
473
50 ppm
423
373
473
20 ppm
423
373
1 minute
3.78
1.98
2.18
7.33
2.45
2.42
7.25
6.27
1.20
7.63
6.05
3.89
2 minutes
2.68
1.73
2.07
6.12
1.73
2.37
6.14
4.11
1.15
7.62
6.01
2.78
5 minutes
1.96
1.57
1.81
3.83
1.26
2.20
4.85
2.61
1.02
5.81
4.47
1.55
10 minutes
1.69
1.51
1.81
2.53
1.07
1.87
3.79
1.98
0.98
4.70
2.65
1.10
15 minutes
1.58
1.48
1.85
1.93
1.00
1.53
3.19
1.68
0.95
4.06
2.16
0.96
20 minutes
1.52
1.46
1.94
1.58
0.94
1.52
2.78
1.49
0.94
3.65
1.90
0.89
30 minutes
1.37
1.44
1.95
1.17
0.86
1.54
2.23
1.26
0.92
2.98
1.59
0.82
40 minutes
1.19
1.38
1.70
0.94
0.79
1.57
1.87
1.11
0.90
2.70
1.41
0.78
60 minutes
0.98
1.15
1.38
0.71
0.70
0.95
1.44
0.92
0.89
2.32
1.16
0.71
o
o


137
Integration and rearrangement isolates the time of reaction:
t 'X [l-(^)J] (4-9)
3 tkjA]t R
Therefore, under gas-film-diffiision control, rc = 0, and we will define rf as the time for
complete reaction of the particle:
aPBR
7' 3UJA'
(4-10)
We can rewrite this equation in terms of fractional solid conversion (X) by allowing:
1 X (vo^ume unreacted core
total particle volume
) =
4/3 it/-3 r
7 = (i)3
4/3 TiR 3 R
(4-11)
Whence we can define the dependence of X upon t:
t f
i-(-)3 = X
V R
(4-12)
Plotting solid conversion versus bed exposure or reaction time, as previously described,
should reveal a linear relationship having a slope of 1/ rp if gas-film-diffiision is rate-limiting.
Chemical-Reaction Control
From Figure 4-29, it can be seen that the concentration gradient, and the related
progress of reaction, are unaffected by the presence of any inert layer(s) in this situation. The


Packed Bed Region
2.5 cm 2.5 cm
2.5 cm
2.5 cm
316 SS Fixed-bed Reactor Schematic (Not to Scale)
Total Length Approximately 47 cm.
Figure 3-4. Schematic representation of 316 stainless steel packed-bed reactor (end connections not shown).
o\
4^


1/k (sec)
Figure 4-38. Plot of inverse first-order rate coefficient (1/k) versus bed exposure time ([N02]in=47.5 ppm, T=423 K, [O2]=0%).
C\
U)


42
reuseable sulfuric acid or in some cases, elemental sulfur. A magnesium oxide process similar
to lime or limestone scrubbing is sometimes used where MgO is hydrated to Mg(OH)2 which
reacts with S02 forming MgS03/MgS04 solids. These can be heated, generating S02 and
MgO. The S02 can normally be recovered as a sulfuric acid product (Cooper and Alley,
1994).
SO-.-CaO Reaction
Borgwardt (1970) investigated the reaction of S02 with calcined limestones and found
a first-order chemical reaction to be the predominant process with resistance limiting S02
sorption by small particles. He found that the reaction rate decreased rapidly as CaO
conversion increased. Some of the solids contained significant quantities of MgO, which did
not readily react with S02 at the elevated temperatures evaluated. Wen and Ishida (1973)
found, through an application of the grain model, that the overall reaction rate between S02
and CaO was highly temperature dependent, particularly at lower temperatures where it was
controlled by chemical kinetics. They reported that various kinetic studies have indicated that
reaction rates may vary considerably (more than an order of magnitude) depending upon the
type of limestones used. This variability is a common problem inherent in the study of
heterogeneous sorbents.
The previously mentioned study by Hartman and Coughlin (1976), which improved
the applicability of the grain model, also described the S02-CaO reaction. The strong
relationship between surface area and the reactivity of CaO toward S02 was analyzed via the
shrinking unreacted-core model by Borgwardt and Bruce (1986). The surface area is


165
4-39, effective N02 removal has stabilized at a value of approximately 22% at this exposure
time. Allowing the reaction to progress for four more hours, N02 removal had only changed
to 20%. The shrinking unreacted-core model can be used to reasonably predict "effective"
complete conversion of the solid, and the time at which this will occur. This can be combined
with the empirical model to predict what the N02 removal rate will be when this occurs. The
sorbent appears to continue to remove N02 at a rate that is probably a function of the
effective diffusivity through the product layer for subsequent removal by the abundant
quantity of MgO available inside the particle.
20 ppm N02 at a Reaction Temperature of 373 K
The controlling mechanism for this low-concentration, low-temperature extreme was
difficult to discern. From Figure 4-40, there appears to be little difference between the
chemical-reaction-control equation and the product layer diffusion equation. Progress of the
reaction is visually more non-linear for both equations during the first 20 minutes than for the
remainder of the curve. Fitting the equations to the longer-exposure-time data between 20
and 60 minutes produced a better fit. The results are shown in Figures 4-41 and 4-42. From
these curves, zr is calculated to be 1023 minutes, and zd is equal to 1093 minutes. Using the
combination-of-resistances equation for the data between 20 and 60 minutes, a zr value of
952 minutes was calculated. Using the differential molar volume equation, a zd value of 515
minutes was calculated. At the mid-point of the bed, again, relative z values at this point
would be approximately half of these values.


7
N0X concentrations seem to point toward increased regulatory pressure to reduce NOx
emissions from jet engine test cell activities in the near future.
NO.. Formation
If NOx standards for test cell emissions are to be considered, clearly those systems
available for controlling these emissions must be thoroughly evaluated. To reduce NOx
emissions, two primary actions can be taken. These are the direct control of the reaction(s)
producing the pollutant (combustion process), or control of the process effluent. To
understand the means of controlling NOx emissions, it is critical to examine the basic
chemistry, thermodynamics, and kinetics of the formation reactions. Oxides of nitrogen
formed by combustion processes are generally caused either by the conversion of chemically
bound nitrogen in the fuel ("fuel NQX"), or by thermal fixation of atmospheric nitrogen in the
combustion air ("thermal NOx").
The principal factors affecting the formation of nitrogen oxides in combustion
processes are the amount of fiiel-bound nitrogen, peak combustion temperature, oxygen level
at peak temperature, and residence time in the combustion zone (Organization for Economic
Cooperation and Development, 1983).
Fuel NO..
When a fuel containing organically bound nitrogen is burned, the contribution of
the fuel-bound nitrogen to total NOx can be significant. The N-C bond is much weaker than
the N-N bond in molecular nitrogen, so fuel nitrogen can be more easily oxidized to NO by


55
layer surrounding the sorbent particle. Practical constraints on pressure drop will likely limit
gas velocity in the application of this control technology to jet engine test cells. Comparative
experiments and mathematical evaluations were performed to ensure that this source of error
was effectively minimized. Results will be presented in a subsequent section.
Gas Reaction/Sorption/Desorption
Gas reaction/sorption was measured by the change in concentration of the test
gas(es) with bed residence and exposure times. Gases were sampled at the inlet and outlet
of the sorbent bed, and also from three points within the bed (of nominal 0.1-m length). Both
short-term (to dynamic equilibrium) and long-term results (taking the bed to "saturation," a
function of bed exposure time) were collected. Concentration-versus-time data from within
the bed were used to evaluate intrinsic chemical kinetics. Overall kinetic data were derived
from progressive removal efficiencies and were used to model overall sorbent performance
with increasing bed exposure times.
Fixed-Bed Reactor System
General Experimental Considerations
Many gas-solid reactions are studied using individual sorbent particles or pellets.
These individual particles are placed in a controlled environment and exposed to the adsorbate
of interest. Sorption behavior is determined gravimetrically using a microbalance and mass
balance calculations. An increase in sorbent mass is directly related to solid sorbent
conversion via stoichiometry. Some fundamental gas-solid studies employ either thin films


67
This arrangement was chosen to ensure that a representative sample of gas was collected
from any given location within the bed, avoiding the potential bias caused by any flow
disturbances through the packed bed, or by reactor wall effects. The number of holes and
perpendicular placement were intended to prevent inconsistencies due to hole blockage during
the course of a run, although this did not appear to be a problem. A thermocouple attachment
screw mounting was placed on the outside of the reactor(s) to allow for the independent
measurement of reaction temperature for comparison with the indicated furnace temperature.
The mixing and sampling manifolds were simply sections of 316 stainless steel pipe,
approximately 70 cm long (28 inches), with six permanently emplaced 316 stainless steel
compression fittings for the connection of gas transfer lines and/or valves, as appropriate.
Before use, machined parts were cleaned and degreased with methanol, followed by
acetone and a deionized water rinse. Cleaning all parts with deionized water between runs
was sufficient during the normal conduct of experiments. For a nominal bed length of 10 cm,
the bulk volume of a packed bed was approximately 51 cm3. Individual sorbent samples were
measured out using a graduated cylinder filled to a loosely packed volume of approximately
51 cm3. These samples were transferred to preweighed sealable polyethylene sample
containers and weighed (approximate mass 7-7.5 g/sample). Since MgO comprised
approximately 45% of the total sorbent by mass, each bed contained approximately 3 g MgO.
After sampling lines were inserted into the reactor, the bed was packed between circular 316
stainless steel screens (12x12 mesh) at the center of the clean tube using a marked steel ram
rod to ensure that the bed was properly located. End connections were replaced and a Type
K (Nickel-Chromium versus Nickel-Aluminum) thermocouple (Model XCIB-4-4-3, Omega


11
Consideration of equation (1-8) allows insight into the effect of temperature on NO
formation, particularly in the hot flame zone. Once exhaust gases move away from this zone,
they cool rapidly, reducing the reaction rates by orders of magnitude. If excess oxygen is
present as the gas cools, the conversion of NO to N02 (reaction (1-5)) is favored.
Thermodynamics predicts that the cooled gas will consist primarily of N02. In reality,
although the favored ambient form of NOx is N02, flue gases from combustion contain
predominantly NO. Approximately 90 to 95 percent of NOx emitted from combustion
processes appears as NO, that is thermodynamically unstable in the environment as its
temperature drops. However, the decomposition of NO into N2 and 02 and the reaction of
NO with 02 to form N02 are kinetically limited. Thus, the high concentrations of NO formed
at high temperature in the combustion zone are "frozen in" and are carried out with exhaust
gases into the atmosphere. From equation (1-8), therefore, it is seen that, with respect to the
reduction of thermal NOx, combustion control strategies could be developed that are aimed
at a reduction in peak temperature, a reduction in gas residence time at peak temperature, and
a reduction of oxygen concentration in the high-temperature zone (Cooper and Alley, 1994;
Wark and Warner, 1981; Organization for Economic Cooperation and Development, 1983).
Regulation of NO.. Emissions
The federal primary air quality standard for N02 is presently 0.05 ppm, based upon
an annual arithmetic mean. Emission and performance standards for various stationary
sources of NOxhave also been established. Although the exhaust from jet engine test cells


130
Sorbent "Lifetime" and Regeneration
Whereas initial sorbent N02 removal rates were often greater than 90 percent
fractional N02 removal decreased while N02 concentrations leaving the bed increased with
bed exposure time (Figure 4-27). In this example, dynamic equilibrium was approached after
approximately 20 minutes of bed exposure. At approximately the same time, NO outlet
concentrations increased to values greater than NO inlet concentration. The NO produced
was much greater than that which would result from nitrate decomposition, as previously
described. Also, NO production through thermal N02 decomposition at temperatures below
500 K is minimal. These observations indicate that there must be another mechanism for NO
production.
A consecutive multi-step chemical reaction of two or more molecules of N02 with
MgO leading to the bed surface becoming saturated with Mg(N03)2 may be proposed. One
step (which it is convenient to propose as the last) must be an oxidation process in which the
equivalent of an oxygen atom is delivered to some intermediate product having the empirical
formula MgN205 (MgO + 2N02). On a mass-ratio basis, this process appears to occur, on
average, when approximately 3 mg N02 per gram of MgO are sorbed. While this
corresponds to minimal utilization of the bed, significant capacity for N02 removal persists
long after initiation of this final oxidation process, as N02 removal efficiencies of 50-60% can
continue for hours. It may be that MgO or Mg(OH)2 at the external particle surface is
covered by a nitrated monolayer, so only internal oxide or hydroxide is available for reaction.


28
improved NOx reduction and increased the sensitivity of the process to temperature change.
Vermiculite coated with MgO removed approximately 70 percent of NOx and the process was
relatively insensitive to changes in waste gas composition, velocity and temperature.
It is believed that the NOx species were chemically adsorbed by the MgO/vermiculite
product and attached at the surface as complex magnesium nitrites and/or nitrates. Tanabe
and Fukuda (1974) found that MgO has a higher number of active (basic) sites compared to
other alkaline earth metal oxides (i.e., CaO, SrO and BaO). Zhang et al. (1994) further
attributed the activity of MgO surfaces to oxygen vacancies in the lattice structure. Some
selected physical properties of MgO, Mg(OH)2, and magnesium nitrite and nitrates are shown
in Table 2-1.
Table 2-1, Physical properties of selected Mg-containing inorganic compounds.
Species
Mol Wt
(g/g-mol)
Density
(g/cm3)
Mol Vol
(cm3/g-
mol)
Melting
Point (C)
Boiling
Point (C)
MgO
40.30
3.58
11.26
2852
3600
Mg(OH)2
58.32
2.36
24.71
See Note
MofNf 1 -3T-T f)
170.36
100
Mg(N03)2-2H20
184.35
2.03
90.81
129
Mg(N03)2-6H20
256.41
1.64
156.35
89
Decomp
330
Note: Loses water of hydration at 350 C.
Source: CRC Handbook of Chemistry and Physics, 73rd Edition, 1992.


Reactor Cross Section
(2.54 cm ID)
Figure 3-5. Schematic representation of internal sampling tube location and appearance.
o\


89
from the figure, the bed did not remove a large amount of NO from the flowing gas stream,
even in this case, with both water vapor and oxygen present. The observed NO removal may
be the result of its oxidation to form N02, which is removed by the bed. Since the gases were
humidified, it is reasonable to assume that the MgO on the sorbent surface was hydrated to
the hydroxide form. In general, when N02 was present, NO was no longer removed after a
specific exposure time and was actually produced after sorption of sufficient N02. The latter
observed effect will be discussed later.
NO-, Removal by Sorbent Material
Overall N02 removal efficiencies were found to greatly exceed those for NO. Initial
N02 removal efficiencies were always greater than 90 percent. As bed exposure time to N02
progressed, N02 removal efficiency decreased, suggesting a "saturation" phenomenon. After
approximately one hour of exposure time, N02 removal efficiency approached a stable value
that was much greater than zero. Typical plots of N02 concentration versus residence time
data are shown in Figure 4-5. In all of the cases evaluated, an intrinsic first-order kinetic
dependence was observed:
-dC/dt = kC (4-1)
where C= N02 concentration in ppm, t=residence time in seconds, and k=first-order rate
constant in sec'1. Integrating equation (4-1) produces:


156
To put these numbers into perspective, the cumulative mass of N02 removed by the
sorbent bed was calculated using an iterative approach by multiplying the one-minute average
N02 removal rates by the N02 mass flow rate for each minute of the run. These were added
together to produce a cumulative mass of N02 reacted at any given exposure time. It was
assumed that all NQ2 removed from the gas stream was taken up by the MgO-vermiculite
sorbent bed. An example of these data collected from the middle of the bed is plotted in
Figure 4-36. The middle of the bed was selected as the sampling point because it was
assumed that the front half of the bed would become "saturated" more rapidly than the back
half of the bed in a long-term run. It can be seen from this figure that the rate of cumulative
mass removal is initially greater than later in the run. This trend is more evident in Figure 4-
37, which shows data collected over a 7-hour period. This long-term run was conducted to
evaluate the x values predicted by the model equations. Since samples were collected at the
center of the bed, only half of the total bed mass was used for sorption up to this point, so x
values should be divided by two for application to this reduced-sorbent-mass and associated
surface area. Plots of the amount of N02 mass adsorbed versus time were concave toward
the time axis, as in most diffusion-controlled processes in which the buildup of concentration
inside the particle progressively reduces the driving force or the accumulation of product
increases resistance (Pigford and Sliger, 1973).
From these figures, the mass removal rate does not appear to approach zero, even
after an extended exposure period. The total mass of N02 removed after 7 hours was only
21.6 mg or 4.7x 10'4 moles. This is significantly less than the predicted stoichiometric
capacity of the mass of MgO in the bed, corresponding to a utilization ratio of approximately


39
Grain Model
To account for sorbent porosity, Szekely and Evans (1970) first improved upon the
basic shrinking unreacted-core model by further defining the structure of the idealized
sorbent pellet or particle. The particle is assumed to be made up of compacting individual
grains of uniform size and shape. A schematic representation of the grain model is shown in
Figure 2-3. While in reality, this assumption may not be entirely accurate, the model presents
a means of accounting for the effects of particle and grain shapes on reaction rates. Each
grain still reacts following the shrinking unreacted-core model. A shape factor for one of
three idealized predominant shapes (either a sphere, cylinder, or flat plate) enters the relevant
shrinking unreacted-core model equations. This shape factor can be used to qualitatively
discern information about the sorbent surface physical structure as it affects sorption
behavior.
This model has been further modified by Hartman and Coughlin (1976), among others,
to account for increased diffusional resistance to chemical reaction in the product layer. This
effect is greatly compounded when the molar volume of the product is significantly larger than
that of the solid reactant, causing coverage or blockage of potentially available intergrain pore
surface area. This may cause a discrepancy between the theoretical solid-surface conversion-
versus-time behavior, as predicted by the model, and actual experimental results.
Ramachandran and Smith (1977) took the model a step further in modeling the single-pore
behavior in a porous pellet to predict the conversion versus time relationship for gas-solid
non-catalytic reactions. Their model accounts for the influence of pore diffusion, diffusion
through the product layer, and surface reaction. A key parameter useful in


4
additive to those of other environmental contaminants (Wark and Warner, 1981; Neal et al.,
1981).
Environmental Impacts of NO.c
In addition to the direct effects on health, nitrogen oxides may create other detrimental
effects on air quality and the environment. These effects include interactions with
hydrocarbons to produce photochemical oxidants and smog; contribution to acid precipitation
through nitric acid formation; and degradation of visibility and aesthetics by formation of
high N02 levels in elevated plumes and fine particulate aerosols resulting from by-products
ofNOx reactions (Organization for Economic Cooperation and Development, 1983).
Sources
Over 95 percent of all the man-made nitrogen oxides that enter the atmosphere in
the United States are produced by the combustion of fossil fuels. As of 1990, over 55 percent
of these emissions are attributed to stationary sources, such as utility and industrial boilers,
gas turbines, and stationary engines. Approximately 40 percent are from mobile sources
(transportation). The remaining fraction is from miscellaneous sources including industrial
processes and waste disposal. On a global scale, the anthropogenic NOx emission rate is
minor compared to natural emissions and formation in the environment. As a result, air
pollution associated with NOx is mainly a local or regional problem (Cooper and Alley, 1994;
Wark and Warner, 1981).
The level of effect of nitrogen oxides on the local environment is highly dependent
upon the emission rates of sources and prevailing meteorological conditions and topological


83
Particle Size Distributions
A sample of MgO-vermiculite sorbent was sieved, as previously described, and
results were plotted on a log-probability graph, shown in Figure 4-2. Since many particle size
data are log-normally distributed, this type of plot should produce a linear representation from
which relevant statistical parameters most importantly, the mass median diameter and
geometric standard deviation can be determined. From Figure 4-2, it can be seen that the
distribution is fairly linear between the 10th and 90th percentiles, so these data were used for
this purpose. From the distribution, the mass median diameter (MMD) is approximately 1.7
mm, and the geometric standard deviation (Og) is approximately 1.5. Since the sorbent
particles were mechanically abraded during the sieving process, the true MMD is probably
somewhat larger, perhaps in the range of 2 to 3 mm. The median diameter determined here,
however, can be used for calculating flow characteristics for a packed bed filled with this type
of sorbent.
A log-probability graph for the manufacturer-supplied size-distribution data on the
MgO powder used to prepare the combination sorbent is shown at Figure 4-3. From this
graph, the MMD for the MgO powder used is approximately 1.3 pm, and og is equal to
approximately 2.5. Since the chemical reaction occurs at the surface of the MgO particles,
this distribution can provide useful data for use in determining physical parameters via the
shrinking unreacted-core model. The MMD is probably not the most appropriate dimension
to use for these purposes, however, since the reactions occurring


151
equation, zr (the inverse of the slope of the regression line) is equal to 397 minutes. While
the coefficient of determination's value is relatively high, there is obviously some curvature.
A considerably better fit was obtained for the product-layer-diffiision-control equation, where
both the r2 and F-statistic values are higher, as shown in Figure 4-33. Only the initial point
falls outside the 95% confidence interval. From this fit, rd is determined to be equal to 820
minutes, approximately twice as large as r which would necessarily be smaller since
chemical-reaction control assumes a more-rapid rate of reaction and, relatedly, a more-rapid
solid conversion rate.
It is quite likely that both processes are simultaneously playing a role in the limitation
of the reaction rate. It would be expected that chemical reaction control would be more
important early in the reaction process. To test this, the first 15 minutes of data are replotted
as shown in Figures 4-34 and 4-35. From these figures, the two mechanisms appear to be
equally important in describing initial sorption behavior under these conditions. To test the
combined resistance equation, the value of o2 was calculated to be 2.5, which when used in
Equation 4-31, produced a value of rr equal to 357 minutes. As was previously discussed,
when the molar volume ratio of product to reactant is large, this may play a role in the
ultimate determination of the overall rate as well as solid sorbent conversion efficiency. From
the data in Table 2-1, the molar volume ratio of Mg(N03)2 (anhydrous) to MgO, Z, is
estimated to be approximately 3.7. Using Equation 4-29, a value of rd equal to 330 minutes
was calculated, a 60% smaller number than that calculated using Equation 4-28. It appears
that the model predicts a significant effect from the molar volume ratio difference.


35
step. A simplified expression for the rate-determining step can then be used to describe the
overall process.
Three major processes are generally considered to be rate-controlling in non-catalytic
gas-solid reactions:
1) Mass transfer of the gaseous reactant from the bulk gas phase through the stagnant
film layer surrounding the particle to the surface of the solid sorbent. The rate of this
step is most dependent upon the fluid dynamics of the gas flow around the solid
particle.
2) Chemical kinetics of the reaction between the gaseous and solid reactants.
3) Mass transfer of the gaseous reactant or product by diffusion through any
product layer at the surface of the solid or in the internal pore structure of the solid.
The rate of this step is highly dependent on the extent of reaction and the physical
properties of the product layer or solid sorbent.
Mathematical Models
The classical Langmuir theory, which is based solely on available surface area, is often
insufficient to account for the formation of a solid product layer on the surface of the solid
reactant and associated surface changes. This theory assumes a uniform reactant surface,
with reaction ceasing after all surface area is covered (Laidler, 1987). Related non-linear
sorption models are not applicable to systems with porous sorbents where isotherms are
favorable and chemical reaction and pore diffusion resistances can be significant (Biyani and


125
where: JP=pressure drop, L=bed depth, factor, absorbent surface area mean particle diameter (mass median diameter), py=gas
density, and £/0=superficial gas velocity, all in appropriate absolute units.
It can be seen from equation (4-3), that at low fluid velocities the first (viscous) term
predominates, and at high gas velocities that the second (inertial) term is dominant and the
pressure drop is proportional to the square of gas velocity. The shape factor term, defined as the ratio of the surface area of a sphere of equal volume to the particle to the
surface area of the particle. The value of this term is obviously 1 for spherical particles and
normally lies in the range from 0.5 to 0.9 for naturally occurring materials. The median
sorbent particle diameter was approximately 2 mm. The value for £v was determined by the
mechanical compression of the sorbent to a point where gas would not flow through the bed,
reducing the volume by approximately 30 percent ( an average bed load produced an hydrophilic, and bulk gas flow does not occur through the internal porosity of the particles,
a value of 0.4 was used for 75 Pa; and when U= 0.2 m/s, AP is equal to 163 Pa. These estimated values correspond well
with average experimental AP values of 75 and 125 Pa, respectively. Figure 4-24 depicts
a comparison between experimentally measured pressure drop and the pressure drop
predicted by the Ergun equation.


53
Gas Composition
To quantify emissions from jet engines, the USAF has sponsored numerous studies
and has published jet engine emission values for many engines in the USAF inventory. These
documents describe NOx and other combustion by-product emissions correlated with power
setting or operational procedures (Spicer et al., 1990). These studies are very useful for
determining appropriate exhaust gas concentrations to simulate jet engine exhaust gas
composition in the laboratory. Many lab-scale studies have used these data in their
experimental design. Gas composition was kept constant for each run. Major components
of JETC exhaust were evaluated individually (mixed in N2 or air): NO (5-200 ppm), N02
(20-200 ppm), and 02 (0-10%). Gas concentration ranges relate to representative values at
four common engine settings (nominally, idle, 30%, 75% and 100% power) for the USAF
FI 10 turbine engine. Actual engine emissions data are shown in Table 3-1. Water vapor (3-
5% by volume at temperature) was evaluated in cohort with other gaseous components.
Water vapor may hydrate the MgO, forming a reactive hydroxide.
It is seen from this table that jet engine exhaust contains significant quantities of
unbumed hydrocarbons, carbon monoxide (CO) and carbon dioxide (CO^. Particulate matter
in the form of soot or condensation aerosols will also most likely be present in real exhaust.
All of these exhaust components have the potential to affect the NOx sorption performance
of the sorbent. For the purposes of this study, these variables were not included to allow for
the evaluation of the basic NOx-sorbent reaction(s).


23
jet engine to service. The engines are operated over their full range of thrust, representative
of typical operational modes. Total test times can range from less than an hour to two to
eight hours. The most common jet engine test cell design is shown in Figure 2-1. The engine
to be tested is mounted horizontally in the U- or L-shaped enclosure and combustion air is
drawn in through sound-deadening baffles. Unequal flow distributions are corrected by
turning vanes to provide an undistorted airflow at the engine outlet at a specified velocity
(often < 50 feet per second). Exhaust gases are blown into a large, long tube (augmenter
tube) with a convergent entrance section. The purposes of the augmenter are to draw air into
the test cell and engine with equal air pressure at the inlet and outlet of the engine (Note
venturi shape), to draw a portion of the air around the engine housing to provide cooling
similar to that experienced during flight, and to dilute and cool engine exhaust to prevent
damage to construction materials. Augmented gas temperatures can vary from about 200
to 2000 F (100-1100 C) depending upon engine firing rate and augmentation ratio.
Generally, the temperature of these gases is less than 500 F (260 C). The augmentation ratio
is varied by the placement of the engine relative to the augmenter throat. Some augmenter
tubes contain cooling water sprays to further quench the exhaust gases. Exhaust gases exit
the augmenter tube through a perforated basket, which helps to dissipate some jet engine
exhaust momentum as well as acoustical energy. In some cases, this basket can be adjusted
to reduce back pressure on the engine. Gases leaving the augmenter tube fill a blast room
before exiting the test cell through a stack. Cooled exhaust is vented to the atmosphere
through multiple channels within the stack to minimize noise (Johnson and Katz, 1990).
Depending upon the engine being tested, the volume flow rate of gas leaving the cell can be


-(1-X)
0.175
n i i i 1 1
0 20 40 60
Bed Exposure Time (min)
Figure 4-32. Shrinking unreacted-core chemical-reaction-control-equation evaluation, t=397 minutes ([N02] =50 ppm
T=423 K, [O2]=0%). [ y = 0.003x + 0.023, r^O.95, Fit Std Error = 0.012, F=208.1],
K)


Pressure Drop (Pa)
Superficial Velocity (m/s)
Predicted + Experimental
Figure 4-24. Predicted (Ergun equation) versus experimental pressure drop as a function of superficial velocity for a 0.1 m long
bed of MgOvermiculite sorbent (T=473 K).


33
reactions are. Chemisorption is also temperature dependent, occurring faster as temperature
increases; however, the amount chemisorbed at equilibrium decreases with temperature
(Hayward and Trapnell, 1964; Szekely et al., 1976 ).
Gassolid reactions are inherently complex due to the number of simultaneous
processes that occur throughout the course of reaction. Any equilibrium conditions reached
are often dynamic rather than static in nature, since the system continues to change with time.
In addition to mass transfer and difiusion processes, these reactions often include heat transfer
and structural changes in the solid phase. Ideally, to properly describe a particular gassolid
reaction, all of these simultaneous and consecutive processes should be taken into
consideration. In practicality, however, this would make for extremely complex mathematical
models. A good engineering model should be one that closely approximates the real situation
without too many mathematical complexities that will make it too cumbersome to use. The
use of simplifying assumptions and the careful design of experiments, to reduce or limit the
effects of variables on a given process so that it may be neglected, are often necessary to
produce useable data and models.
Generalized Non-Catalytic Gas-Solid Reactions
A generalized non-catalytic gassolid reaction may be represented by the following
reaction scheme:
aA (gas) + bB (solid) cC (solid) + dD (gas)
(2-1)


22
calcium oxide, that was highly dependent upon temperature and humidity conditions.
Kitagawa [U.S. Patent 3,382,033 (1968)] developed a dry method that utilizes inorganic salts
such as PbS04, KMn04, KC103, NaCIO, Na^O,,, K2S203, and Na2HP04 and inorganic
oxides like As03 and Pb02 as suitable adsorbents. Collins [U.S. Patent 3,674,429 (1972)]
derived a two-stage process involving silica gel and an activated crystalline zeolitic molecular
sieve. The use of activated carbon to adsorb nitrogen oxides has been studied extensively
(Badjai et al., 1958 (Wark and Warner, 1981)). Its high surface area, adsorption rate and
capacity, and regenerability are well known. James and Hughes (1977) used calcined
limestones and dolomites (MgO-containing) to remove NO at high concentrations with good
results. Kikkinides and Yang (1991) reported some success in the simultaneous sorption of
N02 and S02 using a weak-base macroreticular resin material.
Jet Engine Test Cells
With all of the previous research that has been conducted for application to industrial
processes and power generation equipment, relatively little has been done to specifically
address the control of NOx from jet engine test cells. In order to better understand the
problems associated with NOx control in jet engine test cells, it is useful to examine their
general structure, purpose and characteristics. The Department of Defense operates
approximately 180 cells, which are considered to be stationary sources by the U.S. EPA. The
purpose of a jet engine test cell is to provide a structure to evaluate engine performance
during controlled testing after maintenance to assure proper operation before returning the


118
discrepancies in the intrinsic first-order rate coefficients seen in the preceding tables. Best
estimates of the Arrhenius activation energy associated with this sorption process are between
20 and 36 kJ/g-mol during the first few minutes of sorption without oxygen, and between 15
and 25 kJ/g-mol with oxygen present, for the range of N02 concentrations and temperatures
evaluated. These values are within the range of activation energies generally associated with
chemisorption processes, as previously discussed.
When comparing activation energies for similar acid gas-alkaline solid reactions
reported by different investigators, it must be noted that the values depend on the
experimental arrangements used, the methods used in the analysis of experimental data, and
the accuracy by which initial reaction rates can be measured. The chemical reaction rate
coefficient usually influences only the initial stages of reaction because once a product layer
is formed, the local rate of reaction becomes diffusion controlled (Sotirchos and Zarkanitis,
1992). Hajaligol et al. (1988) determined the activation energy associated with the sulfation
of CaO by S02to be approximately 146 kJ/g-mol. The activation energy calculated from
initial rate data for the sulfation of CaO by Marsh and Ulrichson (1985) was 80 kJ/g-mole.
Zarkanitis (1991) determined activation energies of the reaction of CaO with S02 to be
approximately 17 kJ/g-mole. The latter investigator determined this reaction to be less
sensitive to reaction temperature than the reaction of S02 with calcium carbonate. He also
found that the sulfation of reagent-grade MgO proceeds at rates comparable to those for
CaO. Sotirchos and Zarkanitis (1992) confirmed that the activation energy associated with
the sulfation of CaO was 17 kJ/g-mol using an S02 concentration of 2,000 ppm at reaction
temperatures between 700 and 850 C. Borgwardt (1970) had previously reported that the


Figure 4-33. Shrinking unreacted-core product-layer-diffusion-control-equation evaluation, Td=820 minutes ([NO2]in=50 ppm,
T=423 K, [O2]=0%). [ y = O.OOlx + 0.003,rMi.999, Fit Std Error = 0.001, F=7125.9],
H*


9
In presenting these reactions, it is assumed that the fuel combustion reactions
(between C, H, and O) have reached equilibrium and that the concentrations of O, H, and OH
can be described by equilibrium equations. Considering only the thermodynamics of NOx
formation, the two overall reactions of concern are those which produce NO and N02. The
relevant equilibrium reactions are
N2 + 02 ** 2NO
(1-4)
and NO + 1/2 02 ** N02
(1-5)
The equilibrium constants for reactions (1-4) and (1-5) (Kpl and Kp2, respectively) are
where: Kp = equilibrium constant,
P, = partial pressure of component i (atm),
Yj = mole fraction of component i,
and PT = total pressure (atm).
An analysis of values for Kpl and K 2 at various temperatures indicates that
thermodynamically, the forward reaction in (1-5) is favored over the forward reaction in (1-4)
at low temperatures, with very little NO formed at temperatures below 1000 K. Hence, the


16
Exhaust Gas Treatment
Exhaust gas treatment appears to be the most applicable method of controlling NOx
emissions due to the difficulty in using combustion/operations modifications with JETCs and
the priority placed on unperturbed engine performance during testing. The high volume of
exhaust from a military jet engine, combined with the need for higher levels of NOx reduction,
makes the selection of the optimal exhaust gas treatment a difficult task. There are two main
types of exhaust gas treatment processes, wet and dry processes. Wet processes typically
employ the absorption of NOx. The main disadvantage of wet absorption is the low solubility
of NO. Additionally, these processes often create liquid waste that may require proper
disposal at significant cost. For these reasons, dry processes are generally preferred. These
include selective or non-selective catalytic reduction, selective non-catalytic reduction,
irradiation, and dry sorption (Cooper and Alley, 1994).
In selective catalytic reduction, only the NOx species are reduced (to N2 gas), leaving
free oxygen unreacted. With a suitable catalyst (typically a precious metal, although a variety
of other catalysts have been studied) ammonia (NH3), hydrogen (H2), carbon monoxide (CO),
or hydrogen sulfide (H2S) can be used as the reducing gas. NH3 is the most commonly used
gas. A number of stoichiometries have been proposed for this reaction. For example, Davini
(1988) suggested
4NO + 4NH3 + 02 -* 4N2 + 6H20 (1-9)
Although the selective catalytic reduction process is very effective and is capable of
achieving NO reduction rates exceeding 90 percent in controlling steady-state processes, its


Mixing or Sampling Line Attachment Points
316 SS Mixing or Sampling Manifold
Total Length Approximately 47 cm.
Figure 3-6. Schematic representation of 316 stainless steel mixing or sampling manifold.


N02 Penetration (Fraction)
Bed Exposure Time (min)
s- T=473K
Figure 4-23. N02 penetration versus bed exposure time (20 ml H20 added to 7 g MgO-vermiculite sorbent, [NO2]in=100 ppm,
[O2]=0%, T=473 K).


159
0.2%. It is interesting to note that the N02 removal to NO production ration was
approximately 4:1 after 3 hours, and was approximately 3:1 after 7 hours, supporting the
previously proposed mechanism of removal. The low utilization ratio is believed to be a
complex function of a number of important variables, including the low N02 concentration
in the bulk gas phase, the high product-to-reactant molar volume ratio, and reduced sorbent
porosity.
Without a significant concentration-gradient driving force, formation of a layer of
Mg(N03)2 on the exterior surface area of the sorbent will render the remaining available MgO
relatively inaccessible for reaction with N02. The high product-to-reactant molar volume
ratio means that a significant amount of surface area can become blocked from reaction as
product is formed. For reactions accompanied by an increase in the volume of the solid
phase, incomplete conversion may be expected with ultimate conversion decreasing with an
increase in intrapellet diffusional resistance. In many gas-solid reactions, maximum solid
conversion approaches a constant value less than 100%. The main reasons for this
phenomenon are structural changes of the pores. Maximum conversion is reached when pore
closure occurs (Ramachandran and Smith, 1977). The reaction of S02 with CaO or MgO
forms a solid product, CaS04 or MgS04, which occupies more space than the solid reactant.
If there is no increase in the overall volume of the porous particles during reaction, the pore
space may become completely filled with or covered by solid product. Once pores become
plugged, the interior of the porous particles are practically inaccessible to gaseous reactants
since diffusion through the solid product is an extremely slow process (Sotirchos and
Zarkanitis, 1992). Bhatia and Perlmutter (1983) attributed the continued slow solid-surface


56
of solid sorbent reactant, or fine particles dispersed within an inert material bed such as quartz
silica sand or glass beads. While these types of studies are valuable for collecting data on
fundamental reactions occurring between the phases, in this study, by necessity, the use of the
prepared sorbent was a fundamental limitation. Experimental results may or may not be
transferable to the basic reactions between NOx gases and MgO or Mg(OH)2 solids. It is
possible that the process of making the sorbent promotes sorption through the enhancement
of available surface area. It is also possible, however, that the inefficient transfer of Mg(OH)2
slurry to the vermiculite substrate before calcination may create dense agglomerates or
otherwise make MgO unavailable for reaction with NOx. The effects on ultimate sorbent
conversion and bed utilization will be discussed later.
Often, particularly in chemical or materials engineering studies of gassolid reactions,
the emphasis is on the solid sorbent surface and its conversion as reaction progresses.
Experiments are designed for the purpose of optimizing the solid sorbent itself. Solid
conversion is generally measured directly through chemical analyses, including spectroscopic
and chromatographic methods. Physical properties of the sorbent can be evaluated using X-
ray diffraction or electron microscopy, for example. While these previously discussed studies
are invaluable, in environmental engineering and particularly, in air pollution control, the
focus is often on the gas-phase concentration and the effective removal of the pollutant of
interest. This applied engineering approach is the result of the unavoidable fact that
environmental regulations mandate the reduction of emissions to specified limits and
compliance with regulations is determined through air sampling. It is still crucial, however,
that gas-phase concentrations be correlated with solid conversion through an appropriate


Combustion and
Augmentation Air
Exhaust
Figure 2-1. Schematic Representation of a Jet Engine Test Cell.


4-38. Plot of inverse first-order rate coefficient (1/k) versus bed exposure time
([N02]in=47.5ppm, T=423 K, [O2]=0%) . 163
4-39. Experimental versus predicted N02 concentration ([NOJj^T.S ppm,
T=423 K, [O2]=0%, Gas samples collected from middle of bed) 164
4-40. Shrinking unreacted-core model evaluation ([NO2]in=20 ppm, T=373 K,
[O2]=0%) 166
4-41. Shrinking unreacted-core chemical-reaction-control-equation evaluation, xr=1023
minutes, ([NO2]in=20 ppm, T=373 K, [OJ=0%) 167
4-42. Shrinking unreacted-core product-layer-diffiision-equation evaluation, -rd=1093
minutes, ([NOJ^O ppm, T=373 K, [OJ=0%) 168
4-43. Experimental versus predicted N02 concentration ([NOJ^^. 1 ppm, T=373 K,
[OJ=0%, Gas samples collected from middle of bed) 170
4-44. Shrinking unreacted-core model evaluation ([N02]in=200 ppm, T=473 K,
[O2]-0%) 171
4-45. Shrinking unreacted-core chemical-reaction-control-equation evaluation, xr=495
minutes, ([N02]in=200 ppm, T=473 K, [O2]=0%) 172
4-46. Shrinking unreacted-core product-layer-diffiision-equation evaluation, xd=1015
minutes, ([N02]in=200 ppm, T=473 K, [O2]=0%) 173
4-47. Experimental versus predicted N02 concentration ([N02]in:=170.0 ppm, T=473 K,
[OJ =0%, Gas samples collected from middle of bed) 175
4-48. Cumulative mass N02 removed versus bed exposure time for a long-term run
([NO2L=170 ppm, T=473 K, [O2]=0%, Gas samples collected from center
of bed) 176
4-49. Comparative bed N02 and NO concentrations versus bed exposure time during
long-term run showing NO production ([NOJ^HO ppm, T=473 K, [O2]=0%,
Gas samples collected from center of bed) 177
xi


15
implemented under the strict protocols of jet engine testing. Water or steam injection into the
fuel, combustion air, or combustion zone is very effective at reducing NOx formation by
reducing combustion temperatures (Cooper and Alley, 1994). Introducing additives into jet
engine combustion zones or fuels, however, is a potential problem due to losses in output
(which is closely monitored during testing) or physical damage to the engine or turbine blades.
Some research has been conducted on other fuel or combustion zone additives as applied to
jet engine test cells, which will be discussed in the following sections.
Combustion Modifications
Combustion modifications can reduce NOx formation by lowering one or more of the
parameters peak temperature, gas residence time, or oxygen concentration in the flame zone.
Peak temperatures can be reduced by 1) using a fuel-rich primary flame zone; 2) increasing
the rate of flame cooling; or 3) decreasing the adiabatic flame temperature by dilution
(exhaust gas recirculation). Gas residence time in the primary flame zone can be reduced by
1) changing the shape of the flame zone; or 2) using the steps listed in the previous strategy.
Oxygen concentration in the flame zone can be reduced by 1) decreasing the overall excess
air rates; 2) controlling the mixing of fuel and air; and 3) using a fuel-rich primary flame zone
(Cooper and Alley, 1994). In general, it is difficult to employ these controls in a jet engine
test cell application due to the changing power settings associated with test and acceptance
runs and associated performance fluctuations. Additionally, the strict performance standards
required during military jet engine testing limit the applicability of combustion modifications.


N02 Penetration (Fraction)
T-373K T=423K T=473K
Figure 4-14. N02 penetration versus bed exposure time ([N02]in=200 ppm, [O2]=0%).
o
oo


Indicated [NO] (ppm)
Standard [NO] (ppm)
Lab Conditions
Figure 3-7. Typical chemiluminescent NOx analyzer calibration curve (for NO channel).


51
Magnesium oxide powder was prepared for addition to the mixture by sieving with
a 28-mesh Tyler Standard sieve (600 pm opening). Forty-five grams MgO per 55 grams dry
vermiculite was added to the wet vermiculite while the mixture was manually stirred using a
stirring tool. This ratio was the same as that reported by Sorbtech and is less than the 60:40
maximum mass ratio described in the patent. The same ratio was used so that the two
sorbents could be compared. Sorbtech is able to prepare large batches of material using
mechanical stirring, perhaps allowing them to use less water, while more evenly coating the
sorbent with the Mg(OH)2 slurry.
After thorough mixing, the sorbent was allowed to air dry overnight until the
individual particles were again "free-flowing." The pans of sorbent were then placed in a
preheated muffle furnace set at 550 C and "conditioned" for 30 minutes. It is believed that
this process increases sorbent surface area by expanding the particles as a result of driving out
interior water vapor via heating. This process also dehydrates the Mg(OH)2 formed by
previously hydrating the MgO powder.
Since only small batches of MgO-coated vermiculite could be made in the laboratory,
this process had to be repeated a number of times to produce a sufficient quantity of sorbent
for experimental purposes. All of these individual batches were combined in an air-tight
polyethylene container and mixed together. While the preparation procedures were carefully
followed in making each batch, some variation could be expected in the sorbent from batch
to batch. The sorbent itself is inherently heterogeneous, however, so it was hoped that these
variations could be minimized by mixing the various batches. Samples of the combined batch
mixture were randomly selected for comparative surface area analyses as a measure of this


47
Research Justification
No other research has been reported that would delineate the intrinsic chemical kinetic
parameters associated with the reactions of NOx with MgO-vermiculite sorbents. Likewise,
the probable reaction mechanism has not yet been determined, including the identification of
rate-controlling step(s). A chemico-physical model like the shrinking unreacted-core model
may be used to explain sorbent performance characteristics and provide useful information
related to the optimization of a process employing this novel sorbent for removing NOxfrom
jet engine test cell exhaust, or even other combustion sources. Reviewing the effects of other
operating parameters such as gas concentration, temperature, flow rate, moisture, and oxygen
can provide a more complete picture of the sorptive phenomena occurring. All of these will
be useful to the engineer or scientist who may attempt to employ this medium in an air
pollution control situation in the future.
Research Objectives
Based upon the justification for this basic research, the objectives of this study are to
1. Qualify and quantify NOx sorption, singly and in combination, by
MgO-vermiculite using appropriate isotherms. Determine appropriate kinetic parameters
associated with the sorption reactions. Determine the controlling mechanism(s) and whether
these are chemical reaction, diffusion, or mass transfer limited.


This dissertation was submitted to the Graduate Faculty of the College of
Engineering and to the Graduate School and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.
August, 1995
/
Winfred M. Phillips
Dean, College of
Engineering
Karen A. Holbrook
Dean, Graduate School


87
NO Removal bv Sorbent Material
NO removal by MgO-coated vermiculite sorbent appeared to depend on inlet NO
concentration and bed temperature. At low inlet NO concentrations (< 10 ppm), NO removal
efficiencies varied from initial values of approximately 60 percent to over 90 percent. In
general, initial NO removal efficiency was greater at 373 K than at 423 or 473 K, indicating
removal by physisorption rather than by chemisorption. The presence of water or water
vapor did improve the removal efficiency of NO, but overall rates of removal were still low.
As previously described, apparent NO removal was enhanced by the presence of oxygen,
through the formation of N02. In their evaluation of NO removal by Ca(OH)2 solids, Chu and
Rochelle (1989) found that the gas- or solid-phase oxidation of NO to N02 played an
important role in the removal of the former. At higher NO concentrations (> 100 ppm), the
overall removal efficiencies were lower (less than 30 percent). Jozewicz et al. (1990) found
no NO (from a 400 ppm NO gas stream concentration) removal by Ca(OH)2 solids without
the presence of oxygen at temperatures between 70 and 180 C. Carbon monoxide (CO) has
been widely used as a probe molecule in surface science studies because it is one of the
simplest adsorbate-adsorbent systems. Mejias et al. (1995) found the interaction of CO on
MgO to be weak and of electrostatic origin, without noticeable chemical contributions. Since
CO and NO are similarly non-reactive, this observation suggests that there may be a fixed NO
sorption capacity, possibly the result of limited surface area available for physisorption.
Figure 4-4 is a representative concentration profile for NO removal by a
MgO-vermiculite sorbent bed, without a significant quantity of N02 present. As can be seen


13
NO.. Controls for Stationary Sources
Application of NO.c Control Strategies
Based upon well characterized chemical and thermodynamic principles that control
the formation of fuel and thermal NOx, several combustion modifications or changes to
operating conditions can be used for these purposes. In power generation and some waste
disposal applications, "low NOx" burners have been developed whose designs inhibit NOx
formation by controlling the mixing of fuel and air. While some research is being conducted
on jet engine designs to reduce NOx production (Beard, 1990), this approach needs to be
evaluated for its applicability in light of strict and specialized military engine performance
requirements and standards. Combustion modifications can reduce NOx formation by
lowering one or more of the parameters: peak temperature, gas residence time, or oxygen
concentration in the flame zone. It is difficult to employ these controls in a JETC application
due to the changing power settings associated with test and acceptance runs and associated
combustion exhaust fluctuations. Additionally, strict performance standards required during
engine testing limit the applicability of combustion modifications which may make test results
unrepresentative.


Figure 4-6. First-order kinetic plot ([NO2]in=100 ppm, T=473 K, [O2]=10%). Dashed lines represent 95% confidence limits.
MD
K>


To Vent
Figure 3-1. Experimental arrangement for packed-bed studies.
VO


TABLE OF CONTENTS
ACKNOWLEDGMENTS iii
LIST OF TABLES vii
LIST OF FIGURES viii
ABSTRACT xii
1. INTRODUCTION 1
Background 1
Properties and Health Effects of Nitrogen Oxides 1
Environmental Impacts of NOx 4
Sources 4
NOx Formation 7
Fuel NOx 7
Thermal NOx 8
Regulation of NOx Emissions 11
NOx Controls for Stationary Sources 13
Application of NOx Control Strategies 13
Modifications to Operating Conditions 14
Combustion Modifications 15
Exhaust Gas Treatment 16
2. CONTROL OF NOx THROUGH GAS-SOLID INTERACTION 18
Exhaust Gas Treatment Methods 18
Catalytic Methods 19
Non-Catalytic Methods 21
Jet Engine Test Cells 22
U.S. Air Force (USAF) Sponsored Research 25
Vermiculite-based Catalyst 26
Field Studies 30
Related Research 31
GasSolid Interaction 32
iv


71
constant in the reaction. This light energy is detected by a photomultiplier tube, which
generates an electronic signal that is then converted to a NO concentration reading.
Nitrogen dioxide is measured by difference through its thermal conversion into NO.
A solenoid valve is used to switch to the NOx mode, where all N02 in the gas sample is
converted to NO, based upon the thermodynamic principles previously explained, and is
combined with NO originally in the sample. In the NO mode, gas does not pass through the
thermal converter and therefore contains both NO and N02, with only the NO capable of
reacting with ozone to produce light energy. Thus, the N02 concentration of the sample is
determined by subtracting the signal from the NO mode from the larger signal obtained in the
NOx mode since:
[NOJ [NO] = [N02 ] (3-2)
The instrument specifications, as reported in the manufacturer's manual, indicate
several scales providing an overall measurement range from zero to 5,000 ppm, with a limit
of detection of 50 ppb (0.050 ppm). Background noise level is 25 ppb (0.025 ppm), well
below the detection limit. The instrument exhibits linearity in response of 1% of full scale,
with a zero drift (24-hour maximum) of 50 ppb (0.05 ppm), and a span drift (24-hour
maximum) of 1% of foil scale. The separate scales are linear over their entire ranges, which
was verified by multiple-point calibrations with certified gases. While the interference of
HN03 and HONO with N02 measurements in chemiluminescent analyzers is well documented
(Spicer et al., 1994; Joseph and Spicer, 1978), the concentrations of these interferents were
expected to be low in the system employed. When sufficient moisture was present to form


46
The process of coal gasification often produces significant quantities of hydrogen
sulfide (H2S). Limestones (either calcined or uncalcined) have been investigated as sorbents
for the removal of these gases through the direct sulfidation of the solids via
thermogravimetric analyses (Krishnan and Sotirchos, 1994). The shrinking unreacted-core
model was employed both quantitatively and qualitatively to derive kinetic parameters and a
to provide a description of the rate-limiting mass transfer processes. In this case, since the
product, calcium sulfide (CaS) is more porous and less voluminous than the CaC03 reactant,
diflusional resistances were limited.
General Applicability of the Shrinking Unreacted-Core Model
From the previous discussion, it is readily seen that the shrinking unreacted-core
model and its derivations are useful tools for describing the complex chemical and physical
processes associated with gas-solid reactions. If the required physical parameters are known,
much detail about the physical rate-controlling mechanisms can be gleaned from a reaction
under study using these models. Combined with the traditional chemical kinetics principles
for determining intrinsic kinetic parameters, data can be collected which can ultimately be
used to improve the use of a gassolid reaction to control gaseous pollutants.


119
activation energy associated with the sulfation of CaO by a 3,000-ppm S02 gas stream at
temperatures between 540 and 1100 C ranged from approximately 34 to 75 kJ/g-mol. Ruiz-
Alsop and Rochelle (1988) estimated the activation energy of the reaction between S02and
Ca(OH)2 to be approximately 12 kJ/g-mol. Obviously, there can be significant variations in
the measurement of activation energies dependent upon experimental conditions.
Other studies of acid gas-solid reactions have produced results varying by a similar
magnitude. Yang and Chen (1979) used a grain model to describe the reaction between
1.87% COS and CaO in the temperature range between 600 and 900 C. They determined an
activation energy of approximately 18 kJ/g-mol associated with this reaction. The apparent
activation energy for the reaction between 5,000 ppm HC1 and CaO at temperatures between
150 and 350 C determined by Gullett et al. (1992) was approximately 28 kJ/g-mol.
Activation energies determined by Krishnan and Sotirchos (1994) for the reaction between
H2S and CaC03 ranged from 14 to 22 kJ/g-mol. The activation energy associated with the
removal of H2S by zinc ferrite was 54 kJ/g-mol, which corresponded to a strong temperature
dependence of the diffusion-controlled reaction (Karlegard and Bjerle, 1994).
These activation energies associated with initial sorption kinetics that are reported
here varied inversely with N02 concentration. Runs conducted with higher inlet N02
concentration had increasingly lower associated activation energies; in fact, it appeared that
in some cases ([NOJ > 100 ppm) these energies were reduced to zero and sometimes became
negative after 30 to 40 minutes. This finding would support a mechanistic transition from
chemical kinetic control to diffusion control through a product layer, or intraparticle diffusion
control. It is not known whether this is a function of increasing product layer thickness and


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
HETEROGENEOUS-PHASE REACTIONS OF NITROGEN DIOXIDE WITH
VERMICULITE-SUPPORTED MAGNESIUM OXIDE
By
Larry Thomas Kimm
August 1995
Chairperson: Eric R. Allen
Major Department: Environmental Engineering Sciences
Controlling nitrogen oxides (NO*) from a non-steady-state stationary source like a jet
engine test cell (JETC) requires a method that is effective over a wide range of conditions.
A heterogeneous, porous, high surface area sorbent material comprised of magnesium oxide
powder attached to a vermiculite substrate has been commercially developed for this purpose.
Data from extensive laboratory testing of this material in a packed-bed flow system are
presented. N02 removal efficiencies, kinetics, and proposed N02 removal mechanisms over
a range of representative JETC exhaust gas characteristics are described. Exhaust gas
variables evaluated included: N02 concentration, temperature, flow rate (retention time),
oxygen content and moisture content. Availability of water and oxygen were found to be
important variables. It is probable that water is necessary for the conversion of MgO to
Mg(OH)2, which is a more reactive compound having thermal stability over the range of
xii


T=373K T=423K ~a~ T=473K
Figure 4-17. N02 penetration versus bed exposure time ([NO2]fa=20 ppm, [O2]=0%).


0.07
Bed Exposure Time (min)
Figure 4-46. Shrinking unreacted-core product-layer-diffusion-control equation evaluation, Td=1015 minutes ([NO^j^OO ppm,
T=473 K, [O2]=0%). [ y = O.OOlx + 0.0003, rM).99, Fit Std Error = 0.002, F=822.3],
-j
U)


4-24. Predicted (Ergun equation) versus experimental pressure drop as a function
of superficial velocity for a 0.1 m long bed of MgOvermiculite sorbent
(T=473 K) 126
4-25. Thermal decomposition of Mg(N03)2- 6H20 on vermiculite (temperature
increasing with time) 128
4-26. Thermal decomposition of used MgOvermiculite sorbent (T=523 K) 129
4-27. Comparative bed outlet N02 and NO concentrations versus bed exposure
time showing NO production 131
4-28. Graphical representation of the shrinking unreacted-core model under diffusion-
through-gas-film control 135
4-29. Graphical representation of the shrinking unreacted-core model under chemical-
reaction control 138
4-30. Graphical representation of the shrinking unreacted-core model under diffusion-
through-inert-product-layer control 141
4-31. Shrinking unreacted-core model evaluation ([NOJ^O ppm, T=423 K,
[OJ=0%) 150
4-32. Shrinking unreacted-core chemical-reaction-control-equation evaluation, xr=397
minutes, ([NO2]in=50 ppm, T=423 K, [O2]=0%) 152
4-33. Shrinking unreacted-core product-layer-diffusion-equation evaluation, td=820
minutes, ([NOJ^O ppm, T=423 K, [O2]=0%) 153
4-34. Shrinking unreacted-core chemical-reaction-control-equation evaluation, t =205
minutes, ([NOJ^SO ppm, T=423 K, [O2]=0%) 154
4-35. Shrinking unreacted-core product-layer-diffusion-equation evaluation, td=957
minutes, ([NOJ^SO ppm, T=423 K, [O2]=0%) 155
4-36. Cumulative mass N02 removed versus bed exposure time ([NOJ^SO ppm,
T=423 K, [O2]=0%) 157
4-37. Cumulative mass N02 removed versus bed exposure time for a long-term run
([NO2]in=50 ppm, T=423 K, [O2]=0%, Gas samples collected from
center of bed) 158
x


T=373K T=423K -b- T=473K
Figure 4-15. N02 penetration versus bed exposure time ([NO2]in=r00 ppm, [O2]=0%).
o


Stagnant Gas Film
Progress of Reaction
Unreacted Sorbent Particle
Figure 2-2. Schematic representation of the shrinking unreacted-core model.


6
Table 1-2. Emission Factors for Nitrogen Oxides (NOx as N02)
Source
Average Emission Factor (as N02)
Coal
Household and commercial
8 lb/ton of coal burned
Industry and utilities
20 lb/ton of coal burned
Fuel Oil
Household and commercial
12-72 lb/1000 gal of oil burned
Industry
72 lb/1000 gal of oil burned
Utility
104 lb/1000 gal of oil burned
Natural Gas
Household and commercial
116 lb/million ft3 of gas burned
Industry
214 lb/million ft3 of gas burned
Utilities
390 lb/million ft3 of gas burned
Gas Turbines
200 lb/million ft3 of gas burned
Waste Disposal
Conical incinerator
0.65 lb/ton of waste burned
Municipal incinerator
2 lb/ton of waste burned
Mobile Source Combustion
Gasoline-powered vehicle
113 lb/1000 gal of gasoline burned
Diesel-powered vehicle
222 lb/1000 gal of oil burned
Aircraft
Conventional
23 lb/flight per engine
Fan-type jet
9.2 lb/flight per engine
Nitric Acid Manufacture
57 lb/ton of acid product
Source: Wark and Warner, 1981 (Based upon U.S. EPA AP-67, 1970).


Stagnant Gas Film
Figure 4-28. Graphical representation of the shrinking unreacted-core model under diusion-through-gas-film control.
u>


45
Other Acid-aas-Calcium-based-solid Reactions
Other sorption studies employing the shrinking unreacted-core model and its
derivations have been conducted to study the control of other acid gases. The reaction of
hydrogen chloride (HC1) with CaO is of interest for the control of acid vapors emitted from
municipal and hazardous waste incineration operations. Gullett et al. (1992) used a
specialized fixed-bed reactor to study this reaction and determine the controlling mechanism
and kinetics. They examined the relative importance of chemical reaction and product layer
diffusion control using the combined resistance shrinking unreacted-core model. Their
research also provided data on how operating parameters like reaction temperature and gas-
phase concentration affect sorption. The strong resistance to chemical reaction provided by
solid-state diffusion was proved to follow a grain model by Weinell et al. (1992). Pakrasi
(1992) performed kinetic studies of HC1 removal by hydrated lime powder in a bench-scale
fixed-bed reactor. His studies included a thorough examination of the chemical kinetics and
possible mechanism associated with the HC1CaO reaction. Relative humidity played a key
role in determining HC1 removal rate and the extent of solid sorbent conversion. Other
relevant variables such as gas concentration and temperature were also examined. Incomplete
conversion (compared to that theoretically predicted stoichiometrically or by the shrinking
unreacted-core model), was the result of increased diffusional resistance in the product layer.
Both the shrinking unreacted-core model and an empirical model were used to interpret
observed conversion versus time data.


94
condition occurred after approximately 5-10 minutes of bed exposure time, providing the
most reliable results for subsequent experiments and data analyses. While these data were
most useful for the prediction of sorbent behavior with time, it made the accurate
determination of intrinsic (or apparent) activation energies more difficult. Data on the
activation energies associated with NOx removal by surface reaction with the
MgO-vermiculite sorbent will be presented in the following section, in which the effects of
reaction temperature are described.
Effects of Operational Variables
Numerous experiments were conducted to thoroughly evaluate the effects of
operational variables of primary importance, such as reactive gas concentrations, humidity,
reaction temperature, and gas flow rate. The combination of these results provided a clear
picture, from which conclusions regarding probable reaction mechanisms and the dominant
factors governing the NOx removal capabilities of the sorbent material were drawn.
Effects of NO-. Concentration
The effects of N02 concentration on overall removal rates were dependent upon the
bed temperature at which data were collected. Figures 4-8 through 4-10 show representative
N02 penetration versus bed exposure time data as a function of inlet N02 concentration for
the three reaction temperatures evaluated. N02 penetration is defined as the percentage
(fraction) of inlet N02 exiting the bed at any given bed exposure or reaction time, or Cout/Cin.
Obviously, this value is (in percent) 100% minus the N02 removal efficiency percentage. In


1 (1-X)1/3 1 3(1-X) 2/3+ 2(1X)
* 423K-R 423K-D
Figure 4-31. Shrinking unreacted-core model evaluation ([NO2]in=50 ppm, T=423 K, [O2]=0%).
o


LIST OF REFERENCES
40 CFR 50, Appendix F., "Measurement Principle and Calibration Procedure for the
Measurement of Nitrogen Dioxide in the Atmosphere (Gas Phase Chemiluminesence)."
Baiker, A.; Dollenmeier, P.; Glinksi, M.; and Reller, A., "Selective Catalytic Reduction of
Nitric Oxide with Ammonia: Part I, Monolayer and Multilayers of Vanadia Supported on
Titania," Applied Catalysis. 35:351-364, 1987a.
Baiker, A; Dollenmeier, P.; Glinski, M.; and Reller, A., "Selective Catalytic Reduction of
Nitric Oxide with Ammonia: Part II, Monolayers of Vanadia Immobilized on Titania Silica
Mixed Gels," Applied Catalysis. 35:365-380, 1987b.
Bamford, C.H.; Tipper, C.F.H.; and Compton, R.G., Comprehensive Chemical Kinetics.
Volume 21, Reactions of Solids with Gases, Elsevier Publishers; Amsterdam, pp 131-134,
1984.
Beard, J., "Computer Modelling Makes Jet Engines Less Noxious," New Scientist. 30, 1990.
Berman, E.; Dong, J.; and Lichtin, N.N., Photopromoted and Thermal Decomposition of
Nitric Oxide by Metal Oxides. ESL-TR-91-32, HQ AFESC, Tyndall AFB, Florida, 1991.
Bhatia, S. K. and Perlmutter, D. D., "A Random Pore Model for Fluid-Solid Reactions: II.
Diffusion and Transport Effects," AIChE Journal. 27(2):247-254, 1981a.
Bhatia, S. K. and Perlmutter, D. D., "The Effect of Pore Structure on Fluid-Solid Reactions:
Application to the S02-Lime Reaction," AIChE Journal. 27(2):226-234, 1981b.
Bhatia, S. K. and Perlmutter, D. D., "Effect of the Product Layer on the Kinetics of the
C02-Lime Reaction," AIChE Journal. 29(l):79-86, 1983.
Biyani, P. and Goochee, C. F. "Nonlinear Fixed-Bed Sorption When Mass Transfer and
Sorption Are Controlling," AIChE Journal. 34(10): 1747-1751, 1988.
192


temperatures evaluated. Gaseous oxygen serves to oxidize NO to N02, the latter being more
readily removed from the gas stream. The presence of oxygen also serves to offset thermal
decomposition of N02 or surface nitrite/nitrate. Effective "lifetime" and regenerability of the
exposed sorbent material were also evaluated. N02 removal efficiencies were found to
greatly exceed those for NO, with a maximum value greater than 90 percent. The effective
conversion of NO to N02 is a crucial requirement for removal of the former. The reaction
between N02 and MgOvermiculite is first-order with respect to N02. The temperature
dependence of the first-order rate coefficients provided evidence that data were collected in
the region of chemical-reaction control. Activation energies associated with this reaction
ranged from approximately 20 and 36 kJ/g-mol when oxygen was not present and 15 to 25
kJ/g-mol when oxygen was present in the system. The shrinking unreacted-core model was
used to describe the physical and chemical mechanisms occurring in the removal of N02 from
a flowing gas stream. It appeared that the reaction begins with control by chemical reaction
and progressed to control by diffusion processes as reaction progressed. An empirical
relationship was also developed which allowed for the prediction of N02 removal with
increasing bed exposure time. The range of temperatures and concentrations evaluated, while
valid for the representation of simulated jet engine test cell conditions, appeared to be a
transition region making the absolute determination of rate-limiting mechanism(s) difficult.
With impending regulations aimed at controlling JETCs as stationary sources of NOx, results
from these studies will be useful to environmental engineers developing air pollution controls
for similar sources. Knowledge of parameters affecting the efficiency and capacity of NOx
control systems employing this medium is necessary to ensure compliance with regulations.
Xlll


194
Dam-Johansen, K. and Ostergaard, K., "High-Temperature Reaction between Sulphur
Dioxide and Limestone II. An Improved Experimental Basis for a Mathematical Model,"
Chemical Engineering Science. 46(3):839-845, 1991b.
Dam-Johansen, K.; Hansen, P.F.B.; and Ostergaard, K., "High-Temperature Reactions
Between Sulphur Dioxide and Limestone III. A Grain-Micrograin Model and its
Verification," Chemical Engineering Science. 46(3):847-853, 1991.
Damle, A. S. "Modeling a Furnace Sorbent Slurry Injection Process," Air and Waste.
44:21-30, 1994.
Davini, P., "Reduction of Nitrogen Oxides with Ammonia: The Activity of Certain Soots,"
Fuel. 67:24-26, 1988.
Fogler, H.S., Elements of Chemical Reaction Engineering. P T R Prentice Hall; Englewood
Cliffs, New Jersey, 1992.
Frey, D. D., "Mixed-Gas Adsorption Dynamics of High-Concentration Components in a
Particulate Bed," AIGiE Journal. 38(10): 1649-1655, 1992.
Ganz, S. N., "Sorption of Nitrogen Oxides by Solid Sorbents," Journal of Applied Chemistry
of the USSR. 31 (1):128-130, 1958.
Gullett, B. K.; Jozewicz, W.; and Stefanski, L. A., "Reaction Kinetics of Ca-Based Sorbents
with HC1," Industrial Engineering Chemistry Research. 31(11):2437-2446, 1992.
Hajaligol, M.R.; Longwell, J.P.; and Sarofim, A.F., "Analysis and Modeling of the Direct
Sulfation of CaC03," Industrial Engineering Chemistry Research. 27(12):2203-2210, 1988.
Ham, D.O.; Moniz, G.; and Gouveia, Additives for NO;[ Emissions Control from Fixed
Sources. ESL-TR-89-24, HQ AFESC, Tyndall AFB, Florida, 1991.
Hansen, P.F.B.; Dam-Johansen, K.; Johnsson, J.E.; and Hulgaard, T., "Catalytic Reduction
of NO and N20 on Limestone During Sulfur Capture under Fluidized Bed Combustion
Conditions," Chemical Engineering Science. 47(9):2419-2427, 1992.
Hartman, M. and Coughlin, R. W., "Reaction of Sulfur Dioxide with Limestone and the Grain
Model," AIChE Journal. 22(3):490-498, 1976.
Hayward, D.O. and Trapnell, B.M.W., Chemisorption. 2nd ed., Butterworth and Co.;
London, 1964.


LIST OF TABLES
Table page
1-1. Physical properties of nitrogen oxides (NxOy) 2
1-2. Emission factors for nitrogen oxides (NOx as N02) 6
2-1. Physical properties of selected Mg-containing inorganic compounds 28
3-1. USAF FI 10 turbine engine emissions data (as reported) 54
4-1. BET surface areas of selected samples of laboratory-prepared
MgO-vermiculite sorbent 82
4-2. Intrinsic first-order rate coefficients (s'1) for N02 removal by
MgO-vermiculite sorbent ([O2]=0%) 100
4-3. Intrinsic first-order rate coefficients (s'1) for NOz removal by
MgO-vermiculite sorbent ([O2]=10%) 106
Vll


Table 4-3. Intrinsic first-order rate coefficients (s'1) for N02 removal by MgOvermiculite sorbent ([O2]:=10%)
N02 Concentration :
Reaction Temp ( K):
473
100 ppm
423
373
473
50 ppm
423
373
473
20 ppm
423
373
1 minute
4.74
2.81
1.74
4.94
3.25
1.55
6.94
7.46
2.39
2 minutes
4.87
2.18
1.64
3.98
2.28
1.25
6.55
4.58
1.45
5 minutes
3.78
1.57
1.55
2.84
1.50
1.09
5.71
3.01
1.01
10 minutes
2.42
1.38
1.53
2.16
1.21
1.05
5.02
2.45
0.88
15 minutes
1.88
1.29
1.52
1.81
1.11
1.03
4.33
2.15
0.83
20 minutes
1.54
1.22
1.53
1.59
1.06
1.02
3.92
1.94
0.80
30 minutes
1.09
1.13
1.57
1.31
1.00
1.00
3.24
1.65
0.76
40 minutes
0.87
1.04
1.58
1.13
0.96
1.00
2.77
1.50
0.75
60 minutes
0.65
0.88
1.43
0.90
0.88
1.00
2.11
1.27
0.72


91
C=C0 exp(-kt) (4-2)
where C0= Inlet N02 concentration in ppm and t=bed residence time in seconds.
From Equation (4-2) it can be seen that a plot of ln(C/C0) versus t should be linear,
with a slope of -k, if the reaction is first-order. Data from Figure 4-5 are plotted in this way
in Figure 4-6. A first-order kinetic expression appeared to describe N02 removal and was
used as a basis to model N02 decay. Plots similar to that shown in Figure 4-6 were used to
determine rate coefficients for different experimental conditions. Having established in all
earlier runs evaluated that data from five sampling points produced sufficiently linear first-
order plots, only three points were used to determine rate constants in subsequent runs,
concentrations at the inlet and outlet and at a point within the bed.
Since the bed surface nature is changing with time during runs, values for k also
change with time. Figure 4-7 is a representative figure showing the decrease in k-values as
reaction (exposure) progresses. Also, as reaction progresses, the rate of change in rate
coefficient values between readings decreases. This decrease in first-order rate coefficients
with increasing bed exposure time may be associated with a change in activation energy,
possibly indicative of a transition from one rate-limiting mechanism to another. Noticing this
trend allowed for the derivation of a predictive empirical relationship between the first-order
rate constant value and bed exposure time, which will be presented later. In the context of
the shrinking unreacted core model, this may represent the transition from chemical reaction
kinetics control (with significant associated activation energy) to a diffusion-controlled
process, which would be a slow process. It was found that an (dynamic) equilibrium


198
Quarles, J. and Lewis, W.H. The New Clean Air Act: A Guide to the Clean Air Program as
Amended in 1990. Morgan, Lewis and Bockius; Washington, D.C., 1990.
Ramachandran, P. A. and Smith, J. M., "A Single-Pore Model for Gas-Solid Noncatalytic
Reactions," AIChE Journal. 23 (3):353-361, 1977.
Ruch, R.J.; Krauss, C.J.; Bacon, W.E., Nelson, S.G.; and Nelson, B.W., The Physical Nature
and the Chemical Reactivity of a Heterogeneous MeO/Vermiculite Flue-Gas Sorbent. Project
Report. Kent State University, Kent, Ohio, and Sanitech, Inc., Twinsburg, Ohio, 1990.
Ruthven, D. M., "A Simple Method of Calculating Mass Transfer Factors for Heterogeneous
Catalytic Reactions," Chemical Engineering Science. 23:759-764, 1968.
Ruthven, D. M., "Adsorption Kinetics," Adsorption: Science and Technology. NATO
Advanced Science Institutes Series. Series E: Applied Sciences. 158:87-114, 1989.
Sidgwick, N. V., The Chemical Elements and Their Compounds. Volume I. Oxford at the
Clarendon Press; Oxford, England, pp 235-239, 1952.
Simons, G. A., "Parameters Limiting Sulfation by CaO," AIChE Journal. 34(1): 167-170,
1988.
Snow, M. J. H.; Longwell, J. P.; and Sarofim, A. F., "Direct Sulfation of Calcium Carbonate,"
Industrial Engineering Chemistry Research. 27(2):268-273, 1988.
Sotirchos, S. V. and Yu, H. C., "Mathematical Modelling of Gas-Solid Reactions with Solid
Product," Chemical Engineering Science. 40(ll):2039-2052, 1985.
Sotirchos, S. V. and Zarkanitis, S., "Inaccessible Pore Volume Formation During Sulfation
of Calcined Limestones," AIChE Journal. 38(10): 1536-1550, 1992.
Spicer, C.W.; Holdren, M.W.; and Smith, D.L., Aircraft Emissions Characterization: FI01
and FI 10 Engines. ESL-TR-89-13, HQ AFESC, Tyndall AFB, Florida, 1990.
Spicer, C. W.; Kenny, D. V.; Ward, G. F.; Billick, I.H.; and Leslie, N. P. "Evaluation of NOz
Measurement Methods for Indoor Air Quality Applications," Air and Waste. 44:163-168,
1994.
Steciak, J.; Levendis, Y. A.; and Wise, D. L., "Effectiveness of Calcium Magnesium Acetate
as Dual S02-NOx Emission Control Agent," AIChE Journal. 41(3):712-722, 1995.
Steffenson, D. M. and Stedman, D. H., "Optimization of the Operating Parameters of
Chemiluminescent Nitric Oxide Detectors," Analytical Chemistry. 46 (12):1704-1708, 1974.


[N02] [NO]
Figure 4-27. Comparative bed outlet N02 and NO concentrations versus bed exposure time showing NO production.


76
means and standard deviations were shown on the control charts. Based upon control chart
results, instruments were recalibrated when necessary. Quality assurance was maintained by
showing that all quality control systems were in good order.
As the NO/NOx chemiluminescent analyzer has a flow bypass system installed to
ensure that samples are always collected at ambient pressure, all experimental measurements
were made at the same bypass flow rate used during the initial calibration. Failure to do so
can result in gas concentration measurement errors as great as 10 percent from one end of
the bypass flow scale to the other. Additionally, a NOx generator (Model 100, Thermo
Environmental Instruments, Franklin, Massachusetts) was used to determine the actual
analyzer NOx converter efficiency and ensure that the efficiency exceeded 95%.
Sorbent Surface Area Determination
In preparation for packed-bed studies, BET surface areas of selected samples of
laboratory-prepared MgO-vermiculite sorbent, Sorbtech-supplied sorbent, and raw coarse
vermiculite were determined using a Micromeretics ASAP 2000 automated surface-area and
pore size analyzer. Duplicate samples were run to verify the validity of the data. Small
samples of sorbent (approximately one gram) were first degassed at an absolute pressure of
10 to 15 mm Hg at 623 K for approximately eight hours. Nitrogen gas (in helium) was
sequentially added to the sample tube at a temperature of 77 K. The volumes of nitrogen
adsorbed at various partial pressures between approximately 0.1 and 0.3 atmosphere were


In k (In s"1)
Bed Exposure Time
B- 5 minutes 10 minutes ~b- 15 minutes -20 minutes
Figure 4-21. Arrhenius plot of In k versus 1000/T (T between 473 and 423 K, [NO2]m=100 ppm, [O2]=0%).
o\


In (C/Co)
0 0.2 0.4 0.6 0.8 1 1.2
Residence Time (sec)
Bed Exposure Time
5 minutes 10 minutes 15 minutes 20 minutes
Figure 4-7. First-order kinetic plot ([NO2]in=100 ppm, T=473 K, [O2]=10.5%).
VO
u>


31
Performance during sorption/desorption cycles was evaluated on the basis of weight gain and
loss. It appeared that there was a loss of surface area when the material was regenerated at
excessively high temperatures (> 800 C). Regeneration consisted of three steps: 1) drying;
2) heating to a temperature above 550 C (in the presence of N2, or CO or CH4 to reduce off
gassing NOx); and 3) cooling. Of these, CO is the gas of choice. Heating during regeneration
was accomplished using a rotary-kiln calciner, a vertical tube apparatus, and a horizontal
conveyor kiln. Wickham and Koel (1988) have published an excellent review on the
reduction of NOx by CO.
Related Research
A review of current literature indicates that no research other than that already
described has been conducted on this MgO-vermiculite sorbent. As a result, no data on
intrinsic kinetics or reaction mechanism(s) exist. Significant amounts of related research,
however, have been conducted on similar acid/acid anhydride gas-base/gas-solid systems.
The results of these studies have been used to design and construct appropriate experimental
systems to collect data for interpretation using proven and accepted sorption study
techniques. A short review of some of the fundamentals of gassolid interactions and
chemisorption is necessary to better understand the complex phenomena involved in these
heterogeneous systems.


101
concentrations. For this reason, the apparent first-order kinetics was not affected by the
presence of 02 in exhaust gases.
Representative plots of N02 penetration versus bed exposure time, similar to those
shown for experimental runs where oxygen was present, are shown at Figures 4-11 through
4-13. In Figure 4-11, at a reaction temperature of473 K, initial N02 penetration values were
less than 1% for all three N02 concentrations evaluated. In this case, an inlet N02
concentration of200 ppm was not evaluated, since the N02 cylinder contained only 200 ppm
gas, and therefore, dilution with air would have decreased this concentration. The shapes of
the penetration curves are similar to those seen in Figure 4-8, with N02 penetration remaining
low (and removal efficiency remaining high) for an inlet N02 concentration of 20 ppm.
Overall N02 penetrations for runs with oxygen present remained lower at a reaction
temperature of473 K when compared to runs conducted without oxygen. The presence of
oxygen may facilitate the continued N02 removal by the sorbent.
This distinction between N02 penetration in the presence and absence of 02 did not
apply for all N02 concentrations at a reaction temperature of 423 K (Figure 4-12). N02
penetration after a bed exposure time of one hour was actually higher for an inlet N02
concentration of 100 ppm when oxygen was present compared to when oxygen was absent.
At 50 ppm NO^ the penetrations were approximately the same and were less at 20 ppm. At
this temperature (423 K), oxygen may aid in the removal of low concentrations of N02. At
a reaction temperature of 373 K (Figure 4-13) the 02 effect trend was reversed. Similar to
the effect seen in Figure 4-10, at the lowest temperature evaluated, it appears that a sufficient
gassolid concentration gradient must exist to provide the driving force for reaction to occur.


1 [NO ]
[NO]
2
Figure 4-25. Thermal decomposition of Mg(N03)2. 6H20 on vermiculite (temperature increasing with time).
to
00


1 [NO ]
[NO]
2
Figure 4-26. Thermal decomposition of used MgOvermiculite sorbent (T=523 K).
to
VO


I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
ft
Eric R. Allen, Chair
Professor of Environmental
Engineering Sciences
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Dale A. Lundgren
Professor Emeritus of Environmental
Engineering Sciences
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
W. Emmett Bolch, Jr.
Professor of Environmental
Engineering Sciences
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy. /} y /
Robert J. Har
Professor of <
an
hemistry
I certify that I have read this study and that in my opinion it conforms to
acceptable standards of scholarly presentation and is fully adequate, in scope and quality,
as a dissertation for the degree of Doctor of Philosophy.
Joseph D. Wander
Adjunct Associate Professor of
Environmental Engineering Sciences