EPA-460/3-76-017
May 1976
ASSESSMENT
OF AUTOMOTIVE SULFATE
EMISSION CONTROL
TECHNOLOGY
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Mobile Source Air Pollution Control
Emission Control Technology Division
Ann Arbor, Michigan 48105
-------
EPA-460/3-76-017
ASSESSMENT
OF AUTOMOTIVE SULFATE
EMISSION CONTROL
TECHNOLOGY
by
K.C. Bachman, E.L. Holt, W.R. Leppard
and E.E. Wigg
Exxon Research and Engineering Company
P.O. Box 51
Linden, New Jersey
Contract No. 68-03-0497
EPA Project Officer: Robert J. Garbe
'Prepared for
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Mobile Source Air Pollution Control
Emission Control Technology Division
Ann Arbor, Michigan 48105
May 1976
-------
This report is issued by the Environmental Protection Agency to report
technical data of interest to a limited number of readers. Copies are
available free of charge to Federal employees, current contractors and
grantees, and nonprofit organizations - in limited quantities - from the
Library Services Office (MD35) , Research Triangle Park, North Carolina
27711; or, for a fee, from the National Technical Information Service,
5285 Port Royal Road, Springfield, Virginia 22161.
. report was furnished to the Environmental Protection Agency by
Exxon Research and Engineering Company, Linden, New Jersey 07036,
in fulfillment of Contract No. 68-03-0497. The contents of this report
are reproduced herein as received from Exxon Research and Engineering
Company. The opinions, findings, and conclusions expressed are those
of the author and not necessarily those of the Environmental Protection
Agency. Mention of company or product names is not to be considered
as an endorsement by the Environmental Protection Agency.
Publication No. EPA-460/3-76-017
11
-------
TABLE OF CONTENTS
Page No.
I. Summary 1
II. Introduction 4
III. Task I - Literature Search 7
IV. Task II - Characterization of Sulfur Emissions from Non-
Catalyst Vehicles 9
IV.1 Experimental Procedures 9
IV.1.1 Vehicle Preparation 9
IV.1.2 Test Fuels 9
IV. 1.3 Test Procedure 10
IV. 2 Results 11
IV.2.1 Gasoline Fueled Vehicles 11
IV.2.2 Diesel Fueled Vehicles 13
V. Task III - Factors Affecting Sulfate Emissions from
Oxidation Catalyst-Equipped Vehicles 15
V.I Summary of Results 15
V.I.I Exhaust 02 Level 15
V.I. 2 Noble Metal Composition 16
V.I. 3 Catalyst Age 16
V.I.4 Catalyst Type 17
V.I.5 Other Factors 17
V.I.6 Graphic Representation of Results 17
V.I.7 Thermal Decomposition of A12(804)3 22
V.2 Experimental Procedures 22
V.2.1 Vehicle Preparation 22
V.2.2 Test Fuels 22
V.2.3 Test Procedure 23
t
V.3 Results 23
V.3.1 Base Case Emissions 23
V.3.1.1 Gaseous Emissions 24
V.3.1.2 Sulfate Emissions 24
V.3.1.2.1 Raw Exhaust 24
V.3.1.2.2 Post-Catalyst
Sulfate Emissions 24
V.3.1.2.3 Sulfur Dioxide Emissions -
Sulfur Balance 26
V.3.2 Effect of Limited Secondary Air 27
V.3.2.1 Pelleted Catalyst Vehicle 28
V.3.2.1.1 Gaseous Emissions 28
V.3.2.1.2 Sulfate Emissions 29
V.3.2.1.3 Sulfur Dioxide Emissions . . 29
-------
Page No.
V.3.2.2 Monolith Catalyst Vehicle 29
V. 3.2.2.1 Gaseous Emissions 29
V.3.2.2.2 Sulfate Emissions 30
V.3.2.2.3 Sulfur Dioxide Emissions . . 30
V.3.2.3 Three-Way Catalyst System - Results ... 30
V.3.3 Effect of Residence Time 31
V.3.3.1 Pelleted Catalyst Vehicle 31
V.3.3.1.1 Gaseous Emissions 31
V.3.3.1.2 Sulfate Emissions ...... 32
V.3.3.1.3 Sulfur Dioxide Emissions ... 32
V.3.3.2 Monolith Catalyst Vehicle . . 32
V.3.3.2.1 Gaseous Emissions 33
V.3.3.2.2 Sulfate Emissions 33
V.3.4 Effect of Catalyst Noble Metal Composition .... 33
V.3.4.1 Pelleted Catalyst Vehicle 34
V.3.4.1.1 Gaseous Emissions 34
V.3.4.1.2 Sulfate Emissions 35
V.3.4.1.3 Sulfur Dioxide Emissions ... 35
V.3.4.2 Monolith Catalyst Vehicle 35
V.3.4.2.1 Gaseous Emissions 35
V.3.4.2.2 Sulfate Emissions 36
V.3.4.2.3 Sulfur Dioxide Emissions ... 36
V.3.5 Effect of Changes in Catalyst Operating
Temperature 36
V.3.5.1 Pellet Catalyst Vehicle 36
V.3.5.1.1 Gaseous Emissions 36
V.3.5.1.2 Sulfate Emissions 37
V.3.5.1.3 Sulfur Dioxide Emissions ... 37
V.3.5.2 Monolith Catalyst Vehicle 38
V.3.5.2.1 Gaseous Emissions 38
V.3.5.2.2 Sulfate Emissions 38
V.3.5.2.3 Sulfur Dioxide Emissions ... 38
V.3.6 Effect of Higher Noble Metal Loading . 38
V.3.6.1 Gaseous Emissions 39
V.3.6.2 Sulfate Emissions 39
V.3.6.3 Sulfur Dioxide Emissions 39
V.3.7 Effect of Catalyst Aging 39
V.3.7.1 Pellet Catalyst Vehicle 40
V.3.7.1.1 Gaseous Emissions 40
V.3.7.1.2 Sulfate Emissions 40
V.3.7.1.3 Sulfur Dioxide Emissions ... 41
V.3.7.2 Monolith Catalyst Vehicle 41
V.3.7.2.1 Gaseous Emissions 41
V.3.7.2.2 Sulfate Emissions 41
V.3.7.2.3 Sulfur Dioxide Emissions ... 42
ii
-------
Page No.
V.3.8 Thermal Decomposition of ^2(804)3 42
VI. Task IV - Feasibility Studies of Sulfate Removal
from Exhaust Gas by Traps 45
VI. 1 Summary of Results 45
VI.1.1 Vehicle Tests 45
VI.1.2 Laboratory Screening 46
VI.1.2.1 Sorbent Activity 46
VI.1.2.2 Effect of Operating Conditions
on Sorption Efficiency 46
VI.2 Vehicle Durability Tests of Traps 47
VI.2.1 Experimental Procedures 47
VI.2.1.1 Vehicle Preparation 47
VI. 2.1.2 Test Fuels 47
VI.2.1.3 Test Procedure 47
VI.2.2 85 CaO/10 Si02/5 Na20 Benchmark Pellets 47
VI.2.2.1 Benchmark Preparation 48
VI.2.2.1.1 Powder Preparation 48
VI.2.2.1.2 Pellet Preparation 43
VI.2.2.2 Experimental Results 49
VI.2.2.3 Chemical Analysis of Used Pellets ... 51
VI.2.3 CaC03 Chips 55
VI.2.3.1 Test Conditions 55
VI.2.3.2 Experimental Results 55
VI.2.4 ZnO Pellets 56
VI.2.4.1 Experimental Results 56
VI.2.5 85 CaO/10 SiC>2/ 5 Na20 Rings 57
VI.2.5.1 Test Conditions 57
VI.2.5.2 Experimental Results 57
VI.3 Status of Sulfate Traps 58
VI.4 Laboratory Screening of Sulfate Sorbents 61
VI.4.1 Experimental Procedure 61
VI.4.2 Experimental Results 61
VI.4.2.1 Benchmark Sorbent 63
VI.4.2.2 Other Calcium Sorbents 63
VI.4.2.3 Magnesium Sorbents 64
VI.4.2.4 Alumina Sorbents 64
VI.4.2.5 Other Sorbents 65
VI.4.3 Conclusions from Laboratory Screening
Programs 65
VI.5 Effect of Space Velocity 65
*
VI.6 Activated Charcoal for Sulfate Removal from
Exhaust Gas 67
VII. References . . . 69
iii
-------
Page No,
Appendix A - Task I - Literature Search 70
A.I Introduction 70
A. 2 Summary and Conclusions 70
A.2.1 Thermodynamics of Sulfuric Acid Production .... 70
A.2.2 Reaction of Sulfur Dioxide and Trioxide
with Exhaust Gas Constituents and Exhaust
System Components 71
A.2.3 Automotive Catalysis of Sulfur Dioxide 71
A.2.4 Sulfate Traps 72
A.3 Thermodynamics of Sulfuric Acid Production 72
A.3.1 Thermodynamics of Sulfur Trioxide Production ... 72
A.3.2 Thermodynamics of Sulfur Trioxide Hydration ... 75
A.3.3 Thermodynamics of Sulfuric Acid Condensation ... 75
A.4 Reaction of Sulfur Dioxide and Trioxide with
Exhaust Gas Constituents and Exhaust System
Components 78
A.4.1 Reaction of Ammonia with Sulfur Trioxide 78
A.4.2 Reduction of Sulfur Trioxide by Ammonia 80
A.4.3 Reduction of Sulfur Trioxide by Carbon
Monoxide 80
A.4.4 Reduction of Sulfur Dioxide by Carbon
Monoxide 81
A.4.5 Reaction of Sulfur Oxides with Iron 81
A.4.6 Reactions of Sulfur Trioxide with
Aluminum Oxide 83
A.5 Automotive Catalysis of Sulfur Dioxide 84
A.5.1 Platinum Catalysis: Industrial Application ... 85
A.5.2 Platinum Catalysis: Automotive Application ... 93
A. 6 Sulfate Trap . 96
A.6.1 Particulate Trap 97
A.6.2 Sorbent Trap 97
A. 7 References 109
Appendix B - Measurements Techniques 112
B.I Gaseous Emissions 112
B.2 Measurement of Sulfate Emissions 112
B.2.1 Exhaust Particulate Sampling System 112
B.2.1.1 Sampling System Components 112
B.2.1.2 Diluent Air Preparation System 114
B.2.1.3 Flow Development Tunnel 114
B.2.1.4 The Exhaust Injection System 114
B.2.1.5 Isokinetic Probe 114
B.2.1.6 Particulate Collecting Stage 118
iv
-------
Page No,
B.2.2 Exhaust Particulate Sampling
System Performance 118
B.2.2.1 Rapid Mixing of Exhaust and
Diluent Air 118
B.2.2.2 Development of Uniform Flow in
Flow Development Tunnel 119
B.2.2.3 Tunnel Sampling Losses 121
B.2.2.4 Equivalent Emission Rates with
Parallel Filters 121
B.2.2.5 Temperature Maintenance of the
Particulate Collection Stage 122
B.2.3 The Goks^yr-Ross Technique 126
B.2.4 Analytical Determination of Sulfate 126
B.2.4.1 Reagents 129
B.2.4.2 Titration Apparatus 129
B.2.4.3 Standardization of 83(0104)2 Solution . . 129
B.2.4.4 Detailed Titration Procedure 130
B.2.4.5 Effect of Nitric Acid on Measurement
of Sulfate 130
B.2.4.6 Precautions about Titration Procedure . . 131
B.2.4.7 Sulfate Determinations on Glass Fiber
Filters Spiked with Known Quantities
of H2S04 132
B.2.4.8 Comparison of Titrimetric and
Gravimetric Procedures on Filters
from Vehicle Tests 133
B.3 Measurement of S02 Emissions 135
B.3.1 The TECO Sulfur Dioxide Analyzer 135
B.3.2 The Peroxide Bubbler Method 135
B.4 References Used in Appendix B 140
Appendix C - Practical Operating Considerations with
Sulfate and S02 Analytical Methods 141
C.I Cross-Checks of Analytical Techniques 141
C.I.I Cross-Checks of Sulfate Analytical Techniques . . 141
C.I.1.1 Comparison of Goks^yr-Ross and
Filter Methods 141
C.I.1.2 Comparison of Sulfate
Sampling Points 144
C.I.1.3 Comparison of Absolute Accuracy
of Analytical Methods 145
C.I.2 Cross-Checks of S02 Analytical Techniques .... 146
C.I.2.1 TECO Results 146
C.I.2.2 Peroxide Bubbler Results 150
C.I.2.3 Analytical Procedure Problems
and Solutions 150
-------
Page No.
Appendix D - Supplemental Information on Sulfate
Traps 155
D.I Analyses of Used CaO/Si02/Na20 Pellet Sorbant .... 155
D.2 Westvaco Corporation Report on the Use of
Activated Carbon to Remove 803 from Exhaust 162
Appendix E - Raw Data Tables 179
-------
LIST OF TABLES
Table No. Title Page No.
IV-1 Base Gasoline Characteristics 9
IV-2 Sulfur Oxide Emissions from Non-Catalyst,
Gasoline Fueled Vehicles 12
IV-3 Sulfur Oxide Emissions from a Diesel
Powered Vehicle 13
V-l Effect of Preconditioning on S04=
Emissions 25
V-2 Thermal Decomposition of Aluminum Sulfate .... 43
VI-1 Radial Distribution of Sulfur and Calcium
in Used Trap Pellets 54
A-l Equilibrium Constants for the Oxidation
of Sulfur Dioxide to Sulfur Trioxide 73
A-2 Free Energies and Equilibrium Constants
for the Reaction of Ammonia and Sulfur
Trioxide Forming Ammonium Sulfate 79
A-3 Free Energies and Equilibrium Constants
for the Reaction of Ammonia and Sulfur
Trioxide to Form Nitrogen, Water and
Sulfur Dioxide 79
A-4 Free Energies and Equilibrium Constants
for the Reduction of Sulfur Trioxide by
Carbon Monoxide 80
A-5 Free Energies and Equilibrium Constants
for the Reduction of Sulfur Dioxide by
Carbon Monoxide 81
A-6 Free Energies and Equilibrium Constants
for the Reaction of Sulfur Trioxide with
Ferric Oxide 82
A-7 Free Energies, Equilibrium Constants, and
Equilibrium Partial Pressures for the
Reaction of Sulfur Trioxide with Gamma
Aluminum Oxide 83
A-8 Summary of Rate Equations for Platinum
Catalysis of the Oxidation of Sulfur
Dioxide 90
vii
-------
Table No. Title Page No.
A-9 Decomposition Temperatures of Sulfated
Sorbents 102
A-10 Mass and Volume Requirements of Various
Sorbents 103
A-ll Solubilities of Various Fresh and
Sulfated Sorbents, Grams per 100 ml of
Cold Water 105
A-12 Estimated Costs of Various Sorbents 108
B-l Comparative Titrations of Sulfate of
Samples With and Without Nitric Acid 131
B-2 Experiments Demonstrating Influence of
Ion Exchanger on Sulfonazo III End
Points 132
B-3 Comparison of Titrimetric and Gravimetric
SC>4= Analyses on Parallel Filters 133
B-4 862 Measurements at Indicated Oxygen
Concentrations 138
B-5 Composite Effects of C02, ®2 and co on
TECO S02 Response 139
C-l Characteristics of Analytical Methods 141
C-2 Comparison of Goksf£yr-Ross and Filter
Sulfate Measurements (Pre-Trap or With
Trap Removed) 142
C-3 Simultaneous Sulfate Readings at Three
Sampling Points 144
C-4 Comparison of TECO and Bubbler S02
Results as a Function of S02 Concentration ... 148
C-5 Determination of TECO Accuracy in
Undiluted Exhaust Gas 149
C-6 Comparison of Titrimetric and Gravimetric
Analyses of Peroxide Bubbler Samples 151
E-l Emissions From The 1974 Chevrolet 180
E-2 Emissions From The 1974 Mazda RX-4 181
viii
-------
Table No. Title Page No.
E-3 Emissions from the 1974 Honda CVCC 182
E-4 Emissions from the 1974 Peugeot Diesel 183
E-5 Emissions from the Base Case Pelleted
Oxidation Catalyst System with Air Pump .... 184
E-6 Emissions from the Base Case Monolithic
Oxidation Catalyst System with Air Pump .... 185
E-7 Emissions from the Pelleted Oxidation
Catalyst System with Limited Secondary Air . . . 186
E-8 Emissions from the Monolithic Oxidation
Catalyst System with Limited Secondary Air . . . 187
E-9 Emissions from the Pelleted Oxidation
Catalyst System Operated at High Space
Velocity (with an Air Pump) 187
E-10 Emissions from the Monolithic Oxidation
Catalyst System Operated at Low Space
Velocity (with an Air Pump) 188
E-ll Emissions from the Pelleted Oxidation
Catalyst System Operated at High Space
Velocity with a Pt Catalyst (and an
Air Pump) 188
E-12 Emissions from the Monolithic Oxidation
Catalyst System with Pt Catalysts
(and an Air Pump) 189
E-13 Emissions from the Pelleted Oxidation
Catalyst System Operated at Higher than
Normal Temperature (and with an Air Pump).... 189
E-14 Emissions from the Monolithic Oxidation
Catalyst System Operated at Lower than
Normal Temperature (and with an Air Pump).... 190
E-15 Emissions from the Pelleted Oxidation
Catalyst System Operated with High Metal
Loading Catalyst (and with an Air Pump) 190
E-16 Emissions from the Pelleted Oxidation
Catalyst System with an Aged Catalyst
(and with an Air Pump) 191
E-17 Emissions from the Monolithic Oxidation
Catalyst System with Aged Pt/Pd
Catalysts (and an Air Pump) 192
ix
-------
Table No. _ Title _ Page No.
E-18 Vehicle Durability Test Results of
Pelleted 85 CaO/10 CaO/5 Na20 .......... 193
E-19 Vehicle Durability Test Results of CaC03
Chips ...................... 195
E-20 Vehicle Durability Test of Pelleted Zinc
Oxide Sorbent .................. 196
E-21 Vehicle Durability Test Results of Ringed
85 CaO/10 Si02/5 Na20 .............. 197
E-22 Laboratory Screening of Sorbent Materials .... 198
E-23 Laboratory Screening of Sorbent Materials -
Effect of Space Velocity ............ 200
-------
LIST OF FIGURES
Figure No. Title Page No.
V-l Effect of Catalyst and System Operating
Variables on Sulfate Emissions - Pelleted
Catalyst-FTP Cycle 18
V-2 Effect of Catalyst and System Operating
Variables on Sulfate Emissions - Pelleted
Catalyst-96 km/h Cruise 19
V-3 Effect of Catalyst and System Operating
Variables on Sulfate Emissions - Monolith
Catalyst-FTP Cycle 20
V-4 Effect of Catalyst and System Operating
Variables on Sulfate Emissions - Monolith
Catalyst-96 km/h Cruise 21
VI-1 Trap Efficiency at 64 km/h - Pellets 52
VI-2 Pressure Drop at 64 km/h - Pellets 53
VI-3 Trap Efficiency at 64 km/h - Rings 59
VI-4 Pressure Drop at 64 km/h - Rings 60
VI-5 Apparatus for Laboratory Screening of Sorbents . 62
A-l Equilibrium Conversion - SO-+1/2 0 * SO . . 74
A-2 Equilibrium Conversion - SO +H 0 * H SO, ... 76
J £- £ f
A-3 Dew Point of H SO 77
B-l Exhaust Particulate Samples 113
B-2 Schematic of Dehumidification Section 115
B-3 Counter Current Exhaust Injection System .... 117
B-4 Dew Point of Diluted Exhaust Versus Air/Exhaust
Dilution Ratio 123
B-5 Relative Humidity of Exhaust Dilution Air
Mixture at Vicinity of Sampling Probes During
the 1972 Federal Test Driving Cycle 124
B-6 Temperature Control System Performance -
Catalyst Equipped Car 125
xi
-------
Figure No. Title Page No.
B-7 Relative Humidity of Exhaust Dilution
Air Mixture at Vicinity of Sampling
Probes During 64 km/h Cruise Conditions
at 32°C 127
B-8 Finned Tube Cooling Setup 128
B-9 Recovery of Sulfate From Spiked Glass Fiber
Filter Samples 134
B-10 Principle of Operation - TECO S0_
Instrument 136
B-ll Permatube Drying System 137
xii
-------
I. Summary
The research carried out under this contract assessed the
potential for the on-vehicle control of sulfate emissions from oxidation
catalyst equipped vehicles. The program was divided into four tasks:
• a literature search on the fate of sulfur in automotive exhaust
systems and on possible methods of trapping sulfate in the
exhaust system,
• measurement of sulfate emissions from non-catalyst vehicles,
• an assessment of the effect of catalyst and engine operating
variables on the level of sulfate emissions, and
• an experimental evaluation of the feasibility of using sorbents
to remove sulfate from the exhaust prior to emission to the
atmosphere.
In addition, this report contains data on:
• an experimental study of the thermal decomposition of aluminum
sulfate which may shed light on the storage and release of
sulfate from catalyst surfaces,
• comparisons of different analytical techniques for determining
S0« and S0,~ in automotive exhaust, and
• a report by Westvaco Corp. on the potential for using activated
charcoal £
Literature Search
charcoal as a reductant for S0_ in auto exhaust.
The literature search developed data on the thermodynamics of
sulfate production, possible reactions of SCL and SO- in the exhaust
system, the oxidation of S0_ over automotive oxidation catalysts, and
the potential for sulfate traps. Thermodynamics indicate that conversions
of S02 to 863 of greater than 50% are possible. At temperatures above
350°C, oxygen partial pressure becomes an important variable in determin-
ing the extent of conversion of SO- to SO-. The reaction of SO- and SO,,
with both CO and the iron surfaces of the exhaust system are favorable.
Also, reaction of SO- with the alumina surface of the catalyst is
favorable below 425°C.
While the literature on the automotive catalysis of S0_ is
limited, extensive data on the oxidation of SO- over Pt catalysts were
found. These data indicate that the reaction is first order in SO™ and
that reducing oxygen partial pressure should reduce S0_ oxidation rate.
The functional relationship between oxygen partial pressure and SO™
oxidation rate is complex.
-------
- 2 -
Calcium oxide appears to be the most promising sorbent for S0_.
Other possible sorbents are oxides of Mg, Mn, and Al.
S0,= Emissions from Non-Catalysts Vehicles
As expected S0,= emissions 'from three gasoline powered vehicles
(conventional V-8, rotary engine, and CVCC engine) were low, equivalent
to less than 1% of the sulfur in the fuel. Conversion of fuel sulfur to
sulfate in a diesel vehicle was somewhat higher, averaging nearly 2%.
The presence of high amounts of carbonaceous particulate in the diesel
exhaust may have caused oxidation of SO-, either in the sampling or
analytical procedure. Further study of SQt~ emissions from diesels is
recommended.
Factors Affecting Sulfate Emissions From Catalyst-Equipped Vehicles
The factors investigated in this task included:
• exhaust 0_ level,
• catalyst age,
o catalyst type (pelleted vs. monolith),
• noble metal composition (Pt vs. Pt-Pd),
• noble metal loading,
a catalyst temperature, and
• residence time of the exhaust gas over the catalyst.
Of these factors, the first two, exhaust 0_ level and catalyst age, were
found to have significant effects on sulfate emission rates.
The effect of exhaust 0,, level was studied by comparing sulfate
emissions when an air pump was used at all times with sulfate emissions
when an air pump was used during cold engine operation only. Lowering
exhaust 0« level in this fashion lowered sulfate by a factor of 5 to 7
on the FTP, and by factors of 2 (for pelleted catalyst) and 10 (for mono-
lithic catalyst) at 96 km/h cruise. Optimization of carburetion should
allow these lower sulfate emissions to be obtained without large increases
in CO and HC emissions. In-house tests conducted by Exxon Research on the
three-way catalyst system, which operates with no excess 02 in the exhaust,
showed sulfate emissions only slightly above those from non-catalyst cars,
while CO and HC were controlled to below Clean Air Act Standards. While
the three-way catalyst system shows definite promise as a means of
simultaneously controlling CO, HC, NO , and S04= emissions, further work
is necessary to demonstrate the durability of the system.
Tests with catalysts which had been aged for 40 000 km or more on
the AMA cycle showed sulfate emissions which were lower than those from
fresh catalysts by a factor of 2 or more with no commensurate loss in CO
and HC emission control.
-------
- 3 -
Monolithic catalysts emit more sulfate under low speed conditions
than do pelleted catalysts, due to the greater sulfate storage capacity of the
pelleted catalysts. The sulfate stored by pelleted catalysts under low
speed conditions is released during the early stages of high speed opera-
tion. After longer periods of high speed operation both types of catalysts
emit similar amounts of sulfate.
The other factors examined: noble metal composition (Pt vs. Pt-Pd),
noble metal loading, catalyst temperature, and residence time had no significant
effect on sulfate emissions, except under conditions where CO and HC control
also suffered.
Sulfate Removal From Exhaust Gas By Traps
Four sorbents, 85 CaO/10 SiO_/5 Na_0 both as pellets and as
rings, CaCO- chips, and ZnO. pellets, were tested for S0,= removal
capability in vehicle tests. The 85 CaO/10 SiO /5 Na.O pellets removed
nearly all the S0,= in exhaust during a 42 000 ran test, but sulfation of
the CaO led to unacceptable pressure drop build-up in the trap. Pres-
sure drop increased from an initial value of 1 000 pascals to a final
value of 30 000 pascals. Using this sorbent in the form of rings de-
creased the pressure drop build-up, but also lowered S0,= removal effi-
ciency to unacceptable levels. Neither CaCO_ nor ZnO? showed good S0,=
removal efficiency.
A number of potential sorbents were tested under laboratory
conditions. Only calcium based materials showed the proper combination
of S0,= removal efficiency and physical stability.
Tests conducted by Westvaco Corp. indicate that the use of
activated charcoal as either a reductant or a sorbent for SO is
impractical.
-------
- 4 -
II. Introduction
The purpose of the research carried out under this contract
was to assess potential means for on-vehicle control of sulfate emissions
from oxidation catalyst-equipped vehicles. These emissions are of
concern because of the possible adverse health effects of ambient
sulfate. While many questions remain about the effect automotive
sulfate emissions will have on ambient sulfate concentrations, and about
the levels of ambient sulfate which constitute a health hazard, prudence
demands that a thorough investigation of the methods for controlling
automotive sulfate emissions be carried out.
All commercial gasolines contain small amounts of sulfur.
Recent data assembled by ERDA's Bartlesville Energy Research Center
showed a range of gasoline sulfur levels of 0.004 to 0.144 wt. %, with
an average sulfur content of 0.033 wt. %. This average was based on
nearly 1400 samples obtained nationwide in the summer of 1974. It
agrees well with similar surveys conducted in recent years.
In & conventional internal combustion engine, gasoline sulfur
is combusted to sulfur dioxide (SOz), the thermodynamically predicted
oxidation product at high temperature. In the absence of an oxidation
catalyst emission control system, essentially all of the S02 formed in
the combustion process is emitted to the atmosphere in that form.
Studies conducted by Exxon Research and others (2,3), prior to the
beginning of this contract research, indicated that less than 1% of the
S02 formed in combustion was converted to sulfate in conventional engine,
non-catalyst vehicles.
However, when oxidation catalysts are present in the emission
control system, some of the S02 formed in combustion is further oxidized
to sulfur trioxide (S03), which can combine with water in the exhaust to
form sulfuric acid (H2S04). Sulfuric acid can react with a variety of
materials present in the exhaust or the exhaust system to form sulfate
salts (MS04). Exxon Research's studies indicate that the hydration of
803 occurs to completion in the exhaust system, but relatively little
further reaction of H2S04 occurs. For the purposes of this report, and
to comply with the convention which has developed, 803, H2S04, and MS04
will be jointly referred to as sulfate.
The fact that some S02 is oxidized to 863 over noble metal
oxidation catalysts used in automotive emission control systems is not
surprising. Almost all of these catalysts contain platinum as their
major active component. Platinum has been used as a catalyst for the
oxidation of S02 to S03 in sulfuric acid production for many years. While
noble metal oxidation catalysts for automotive emission control were under
development since the late 1950's, it was not until Spring of 1973 that
concern was raised over the potential problems associated with automotive
sulfate emissions from oxidation catalyst-equipped cars. The fact that
S02 emissions from mobile sources contribute less than 1% of the national
inventory of S02 undoubtedly contributed to this lack of concern
-------
- 5 -
In Spring, 1974, when this contract was initiated, there was
general agreement on the following points:
• Sulfate was emitted from non-catalyst, conventional
engine vehicles at a level equivalent to less than
1% of the sulfur in the gasoline used by the vehicle.
• Sulfate emissions from oxidation catalyst-equipped
vehicles were significantly higher than from non-
catalyst vehicles, with conversion of gasoline
sulfur to sulfate apparently being a complex
function of a number of poorly defined variables.
• Sulfate was stored on oxidation catalyst surfaces
under low speed conditions and released under high
speed conditions.
There was no general agreement on the amount of sulfate emitted by non-
conventional engines, the effect of various catalyst system operating
parameters on sulfate emissions, or on the feasibility of using sorbents
to trap sulfate emissions in the exhaust system. This program was
designed to provide data in most of these areas.
The program was divided into four tasks. Task I involved a
literature search on the fate of sulfur in automotive exhaust systems
and on possible methods of trapping sulfate in the exhaust system.
Results of this literature search were published as an interim report,
which has been reproduced, with minor editorial changes, as Appendix A
of this report. A summary of the report also appears as Section III of
the body of this report.
Task II involved the measurement of sulfate emissions from
four non-catalyst vehicles powered by the following engines:
• conventional Otto cycle, V-8 configuration
• rotary
• prechamber stratified charge
• diesel» automotive configuration *
The purpose of this task was primarily to expand the data base from non-
catalyst conventional engine vehicles to other non-catalyst vehicles.
HDV Diesels were not studied in this program but are being studied
under EPA Contract 68-03-2116 with Southwest Research Institute in
San Antonio, Texas.
-------
- 6 -
Task III involved an assessment of the factors influencing 502
to S03 conversion over commercial noble metal oxidation catalysts used for
automotive emission control. Factors investigated included:
• catalyst type (pelleted vs. monolith),
• noble metal composition (Pt vs. Pt-Pd),
• noble metal loading,
• catalyst age,
• exhaust 02 level,
• catalyst temperature, and
• residence time of exhaust gas over the catalyst.
The purpose of this task was to identify those operating parameters of
catalytic automotive emission control systems which might be adjusted to
minimize sulfate emissions while maintaining CO and HC control.
Task IV involved a study of the feasibility of using sorbents
to remove sulfate from the exhaust prior to emission to the atmosphere.
Both laboratory screening tests to identify potential sorbents and
vehicle mileage accumulation tests to assess sorbent effectiveness were
carried out.
-------
- 7 -
III. Task I - Literature Search
Task I involved a literature search on the fate of sulfur in
automotive exhaust systems and on possible methods of trapping sulfate on
sorbent traps placed downstream of the oxidation catalyst. Results of
this literature search were published as an interim report of this
program (4). A copy of this interim report, with minor editorial changes,
appears as Appendix A of this report. The summary and conclusions of
the interim report are repeated below.
Thermodynamics of Sulfate Production
1. At typical oxidation catalyst temperatures, conversions of
sulfur dioxide to sulfur trioxide greater than 50% are
thermodynamically possible. At lower temperatures even
greater conversions are possible.
2. The equilibrium conversion is strongly dependent upon oxygen
concentration. At temperatures above 350°C, decreasing the
oxygen concentration decreases the equilibrium conversion,
suggesting a possible control strategy.
3. Thermodynamics and kinetics show that exhaust sulfur trioxide
may hydrate to gaseous sulfuric acid within the vehicle's
exhaust system, depending upon driving mode.
4. Thermodynamics show that the gaseous sulfuric acid will begin
to condense at about 150°C, which is below the temperature at
the tailpipe exit for all driving modes except startup.
Reaction of Sulfur Dioxide and Trioxide with Exhaust Gas Constituents
and Exhaust System Components
1. Thermodynamics shows that ammonia will reduce sulfur trioxide
to the dioxide. However, exhaust ammonia will be oxidized over
the oxidation catalyst before reaction can take place.
2. The formation of ammonium sulfate is favorable only below
225°C.
3. Thermodynamics shows that the reduction of both sulfur oxides
by carbon monoxide is favorable.
4. Reaction of both oxides with the iron oxide surfaces of the
exhaust system is favorable below 425°C.
5. Reaction of sulfur trioxide with the aluminum oxide catalyst
substrate is possible below 425°C. The presence of carbon
monoxide may lower this temperature by about 50°C.
-------
- 8 -
Automotive Catalysis of Sulfur Dioxide
1. The rate limiting step in the catalytic oxidation of sulfur
dioxide is the surface reaction between adsorbed oxygen and
adsorbed sulfur dioxide.
2. The following rate equation appears to represent best the
available experimental data for industrial catalysis and
should be valid for automotive catalysis;
1/2
- PS0
rate =
P
j/2
where ki is the rate constant, subscripted P's are the
partial pressures of the compounds in the subscripts,
Ke is the equilibrium constant of the oxidation reaction,
and the subscripted K's are the adsorption equilibrium
constants for the compounds in the subscripts.
This equation is in accord with the above rate limiting
mechanism. This equation also indicates a possible control
strategy of limiting the amount of oxygen over the catalyst.
The automotive catalysis literature is limited. In addition,
the data are confounded by many experimental problems,
notably the storage/release phenomenon. In general, this
literature says that more sulfur trioxide is formed over
catalysts than is formed in non-catalyst vehicles.
Sulfate Traps
1. The most promising means of removing sulfur trioxide from
the exhaust stream is to react it with a basic metal oxide.
2. Based on a selection criterion consisting of seven require-
ments, the most promising sorbent material is calcium oxide.
Other less promising but still potential sorbents are the
oxides of magnesium, manganese, and aluminum.
-------
- 9 -
IV. Task II - Characterization of Sulfur Emissions from
Non-Catalyst Vehicles
Task II involved measurement of S02, SC>4=, and I^S emissions
from four non-catalyst vehicles: 1) a 1974 350 CID Chevrolet V-8, 2) a
1974 Mazda RX-4 equipped with a rotary engine, 3) a 1974 Honda Civic
equipped with a prechamber stratified charge engine, and 4) a 1975 Peugeot
504 GL equipped with a prechamber diesel engine. All four vehicles showed
low conversion of fuel sulfur to sulfate and no detectible H2S emissions.
In the case of the three gasoline fueled vehicles, sulfate emissions
were less than 1% of the sulfur in the fuel. Conversion of fuel sulfur
to sulfate was somewhat higher in the diesel vehicle, averaging nearly
2%. The presence of high amounts of carbonaceous particulate in the
diesel exhaust may have caused oxidation of SC*2 in either the sampling
or analytical procedure, leading to erroneously high S04= values for the
diesel. This problem is discussed in greater detail below.
IV.1 Experimental Procedures
IV.1.1 Vehicle Preparation
The four vehicles had previously been used in other experimental
programs at Exxon Research. To minimize any effect of these prior test
programs on the results of this study, the program called for replacing
the exhaust systems on the four vehicles. We were able to do this on
all vehicles but the Honda CVCC, which was not commercially available in
the U.S. at the time the tests were carried out. Other than changing
the exhaust systems and assuring that the cars were in proper mechanical
operating condition, no special vehicle preparation steps were taken.
IV.1.2 Test Fuels
Two test gasolines, one containing 0.065 wt. % sulfur, the
other containing 0,032 wt. % sulfur, were used in this portion of the
program. Both fuels were blended from a base fuel, which contained
0.008 wt. % sulfur, by addition of equal quantities of thiophene and di-
t-butyldisulphide. Characteristics of the base gasoline are summarized
in Table IV-1.
Table IV-1
Base Gasoline Characteristics
RVP 91.5 kPa (13.3 psi)
% Evap. @ °C
70°C 33
100 51
150 83
Research Octane 93.0
Motor Octane 82.8
Gravity (g/cc) 0.75
FIA (vol.%)
Arom 33.9
Olef. 11.5
Sulfur (wt. %) 0.008
Lead (grac/1) <0.002
-------
- 10 -
The diesel was also tested on two fuels, one containing 0.17
wt. % sulfur, the other containing 0.35 wt. % sulfur. The lower sulfur
level fuel was a standard No. 2 diesel fuel. The higher level fuel was
prepared by doping the lower sulfur fuel with an appropriate amount of
di-t-butyldisulfide.
IV.1.3 Test Procedure
The vehicles were tested on the high sulfur fuel first. The
test sequence consisted of the following parts:
(1) 800 km (500 mile) accumulation - Federal Durability Driving
Cycle (AMA Cycle), on an automatic chassis dynamometer.
(2) 1975 FTP,
(3) 20 minute idle,
(4) 2 hrs. 96 km/h (60 mph) cruise,
(5) 1975 FTP (after overnight soak).
The two hour 96 km/h cruise was broken into four half hour periods and
separate 20 minute samples taken during each period.
Sulfate emissions were measured by two techniques: Exxon
Research's exhaust particulate sampler; and the Gokstfyr-Ross technique.
Both are described in detail in Appendix B. The Exxon Research exhaust
particulate sampler is a CVS compatible, air dilution type sampler which
collects particulate, including sulfate, from an isokinetic sample of
diluted exhaust. The particulate is collected on 0.2 micron glass fiber
filters. Sulfate is leached from the filter with dilute nitric acid,
then determined by colorimetric titration with barium perchlorate using
Sulfanazo III as an indicator. Details of this procedure are also
presented in Appendix B.
The Goks^yr-Ross technique was used on undiluted exhaust and
involves condensing and filtering 1^804 at between 60 and 90°C, conditions
under which t^O will not condense out of exhaust. The actual collector
used was a modified condenser coil. After completion of a test, H2S04
was washed from the coil and determined using the barium perchlorate-
Sulfanazo III method.
S02 was also determined by two methods, instrumentally using a
TECO UV pulsed fluorescence instrument, and wet-chemically using a hydrogen
peroxide bubbler technique. Both approaches are described in Appendix B.
-------
- 11 -
Hydrogen sulfide, H2S, was determined by the Jacobs-Braverman-
Hochheiser technique (5) which is capable of measurement of H?S at the ppb
level. H2S is absorbed in a bubbler containing cadmium hydroxide solution.
This solution is treated with N, N-dimethyl-p-phenylene-diamine in
sulfuric acid and ferric chloride to form methylene blue. Methylene
blue concentration is then measured spectrophotometrically at 670 my.
Gaseous emissions were measured using standard Federal Test
Procedure instrumentation, CO by non-dispersive infra-red (NDIR), HC by
flame ionization detection (FID), and oxides of nitrogen (NOx) by chem-
iluminescence.
IV.2 Results
IV.2.1 Gasoline Fueled Vehicles
As expected, sulfate emissions from the three gasoline fueled
vehicles were equivalent to less than 1% of the sulfur consumed with the
fuel. The remainder of sulfur in the fuel appeared as SC>2. In all
tests, H2& emissions were below the limit of detection of the Jacobs-
Braverman-Hochheiser method, about 0.006 g/km. S02 and S04= emissions
data for the three gasoline fueled vehicles are summarized in Table
IV-2. The 1975 FTP emission results reported in Table IV-2 are the
average of the two FTP's run; the 96 km/h results are for the average of
the two hour test. Detailed gaseous and sulfur oxide emissions data are
presented in Appendix E, Tables E-l - E-3.
An experimental problem was encountered in testing the Honda
CVCC which prevented us from obtaining a full set of data at the 96 km/h
cruise condition. The Honda CVCC is a front wheel drive vehicle and
cannot be operated normally on the chassis dynamometer. We tested by
rotating the car 180° from normal test position and placing the front
wheels on the dynamometer rolls. Operation in this fashion was possible
only if the dynamometer rolls rotated in the opposite direction from
that when a rear wheel drive car is used. With this mode of operation, we
were unable to obtain a good match between dynamometer and road load for
the Honda. As a result, the vehicle overheated during the 96 km/h,
cruise, and it was necessary to test at a lower speed, 80-88 km/hr.
The 804° emission values reported in Table IV-2 were obtained
using the exhaust particulate sampler. At these low levels the vehicle S04~
emissions caught on the filters approached the blank corrections for the
S04= content of the filters themselves, and of the solutions used in
their analysis. Thus the absolute emission values in Table IV-2 are
less important than the fact that they again verify the conclusion that
non-catalyst, gasoline powered vehicles emit little, if any, S04=.
-------
TABLE IV-2
SULFUR OXIDE EMISSIONS FROM NON-CATALYST,
GASOLINE FUELED VEHICLES
Vehicle
1974 Chevrolet V-8
1974 Mazda RX-4
1974 Honda CVCC
Fuel Sulfur
Level, Wt %
0.065
0.032
0.065
0.032
0.065
0.032
Test
Mode
1975 FTP
96 km/h
1975 FTP
96 km/h
1975 FTP
96 km/h
1975 FTP
96 km/h
1975 FTP
80-96 km/h
1975 FTP
88 km/h
SO 2 Emissions
g/km
0.40*
0.15**
0.20*
0.075**
0.50* (0.30)
0.13**
0.65*(0.15)
(0.10)
0.20* (0.08)
(0.08)
0.10*
(0.04)
% of Fuel S
185
124
210
123
225(150)
112
640(150)
(135)
180(65)
(105)
180
(120)
SO £_
g/km
0.0005
0.0002
0.0005
0.0002
0.0004
0.0000
0.0009
0.0002
0.0002
0.0000
0.0000
0.0000
Emissions
% of Fuel S
0.5
0.6
1.1
0.9
0.4
0
2.0
0.1
0.4
0
0
0
i
H
ts>
1
* Analysis by TECO - Dilute Exhaust
** Analysis by H202 Bubbler - Undiluted Exhaust
( ) Analysis by H202 Bubbler - Dilute Exhaust
-------
- 13 -
The S02 values in Table IV-2, collected by the TECO instrumental
or H2<32 bubbler methods, both of which are described in Appendix B, were
erratic and invariably high. As discussed in Appendix C, Section C.I.2.1,
a number of interferences turned up in using the TECO instrument, including
quenching of the S02 fluorescence by 02, CO, and C02, and spurious fluorescence
by hydrocarbons. Early attempts to correct for these factors were
unsuccessful, resulting in the very high'S02 values, reported in Table IV-2,
obtained with the TECO. The most out of line result was obtained with
the Mazda, whose rotary engine put out the highest level of hydrocarbon
emissions, thereby contributing a large amount of unwanted fluorescence
response. The bubbler samples were closer to 100% sulfur balance. How-
ever, the three FTP runs monitored by bubbler, from dilute exhaust, ranged
from 65 to 150% sulfur balance. As discussed in more detail in Appendix C,
Section C.I.2.2, this poor reproduclbility and inaccuracy is probably due
to the low levels of sulfur dioxide sampled in this test mode. As a
result, the sample analyses were not reliable because of the factors dis-
cussed in Section C.I.2.2. The cruise samples, on the other hand, gave
larger sulfur samples, hence more accurate analyses. They still averaged
high, however, about 120%. This is in line with a later study of bubbler
accuracy, Section C.I.2.3, which showed this technique to be subject to a
bias in this direction for an as yet undetermined reason.
IV.2.2 Diesel Powered Vehicle
Sulfur oxide emissions from the Peugeot Diesel are summarized
in Table IV-3. Detailed gaseous and sulfate emissions are reported in
Appendix E, Table E-4.
Table IV-3
Sulfur Oxide Emissions From a Diesel Powered Vehicle
Fuel Sulfur S02 Emissions* SOA° Emissions
Content. Wt. % Test Mode g/km% of Fue3~S g/km% of FueJ~S
0.17 1975 FTP 0.22 71 0.017 4.2
96 km/h 0.13 57 0.006 1.6
0.35 1975 FTP 0.29 47 0.012 1.4
96 km/h 0.24 48 0.010 1.7
* Analysis by 1^02 Bubbler - Dilute Exhaust
-------
- 14 -
Sulfate emissions from the Diesel were higher than from the
gasoline powered vehicles, as might be expected from the higher sulfur
level fuel used. However, measured conversions of fuel sulfur to sulfate
were also higher than expected with the Diesel. These conversions
averaged roughly 2% for the Diesel as compared with less than 1% for the
gasoline powered vehicles. The high level of carbonaceous particulate
present in Diesel exhaust, and trapped on the particulate filter, may
have promoted oxidation of S02 either in the sampling or analytical
procedure. Further study of Diesel sulfate emissions, preferably in
short term tests in which the build up of carbonaceous material on the
filters is minimized, is necessary before firm conclusions can be drawn.
The S02 measurements were all low. The reason for this is not
known. Possibly the H202 solutions were not able to completely oxidize
the high level of S02 entering the bubblers within the available residence
time. Another possibility may be the high level of carbonaceous particulate
material which was deposited throughout the collection system. This
deposition may also have led to retention of S02 by the particulate. As
seen in Appendix Table E-4, TECO readings were also low, although this
is believed due to particulate fouling of the sample chamber windows.
-------
- 15 -
V. Task III - Factors Affecting Sulfate Emissions From
Oxidation Catalyst-Equipped Vehicles
This task involved an assessment of the factors influencing
S02 to 803 conversion over commercial,noble metal, automotive oxidation
catalysts. The factors investigated included:
catalyst type (pelleted vs. monolithic),
noble metal composition (Pt vs. Pt-Pd),
noble metal loading,
catalyst age,
exhaust 02 level,
catalyst temperature, and
residence time of exhaust gas over the catalyst.
The purpose of this task was to identify those parameters of catalytic
automotive emission control systems which might be adjusted to minimize
sulfate emissions while maintaining CO and HC control.
V.I Summary of Results
V.I.I Exhaust 02 Level
Of the operating parameters studied, exhaust 02 level had the
greatest effect on sulfate emissions. As part of this program, the
effect of limiting exhaust 02 level was studied by comparing sulfate
emissions when an air pump was used at all times, with sulfate emissions
when an air pump was used only during cold engine operation, the first
two minutes of the FTP. Limiting exhaust 02 level in this fashion
resulted in sulfate emissions which were lower by a factor of 5 to 7
than the full air case, over the FTP. At 96 km/h, elimination of the
air pump resulted in sulfate emissions which were lower by a factor of 2
for the pelleted catalyst, and by a factor of more than 10 for the
monolithic catalyst. The pelleted catalyst was tested on a car designed
to operate lean, without an air pump, and as a result, limiting air pump
use had only a small effect on CO and HC emissions. The monolithic
catalyst, on the other hand, was tested on a car designed to use an air
pump. Limiting air pump use of this car resulted in a relatively large
rise in CO emissions, from 2.8 to 5.3 g/km, although HC emissions did
not change, showing 0.30 and 0.27 g/km for the two cases. Much of this
rise in CO emissions could probably be eliminated by optimization of carburetion.
The ultimate embodiment of the limited exhaust 02 approach is
operation with no excess 02 in the exhaust, at the stoichiometric point.
The exhaust can then be passed over a catalyst, referred to as a "three-
way catalyst", which will react the CO, H2, HC, NOX, and 02 in the
exhaust to form H20, C02, and N2- This approach requires very close
control of air-fuel ratio, as close to stoichiometric as possible. If
the exhaust is too rich, insufficient 02 will be available to oxidize
-------
- 16 -
CO and HC, and these emissions will rise. If the exhaust is too lean,
NOx reduction, which does not proceed in the presence of excess 02, will
cease. Since this level of control is beyond the capability of any current
open-loop fuel metering system, a closed loop control means is necessary.
The most advanced approach for providing such control uses an oxygen
sensor in the exhaust to provide feed back control to an electronic fuel
injection system.
Tests on a three-way catalyst system were not part of the
scope of work of this contract. However, since tests conducted by Exxon
Research, as part of its in-house program on control of sulfate emissions,
show that sulfate emissions from this system are similar to those from a
non-catalyst car, and significantly lower than from a typical oxidation
catalyst-equipped vehicle, a discussion of this system has been included
in this report.
Tests with a Volvo equipped with a three-way catalyst system
consisting of a Bosch oxygen sensor and an Engelhard TWC-9 catalyst
showed very low sulfate emissions under all test conditions. For the
FTP, this system's sulfate emissions, with gasoline containing 0.03 wt.
% sulfur, were 0.0012 g/km. This is only 5% of the sulfate emissions
from the monolithic catalyst-air pump case, and very nearly the 0.0006
g/km average usually found for sulfate emissions from a non-catalyst
vehicle.
While limited excess air is of definite interest in controlling
sulfate emissions, its use may involve some penalties. It is not certain
that production vehicles which use oxidation catalysts but not air pumps
could generally meet emission standards more stringent than 5.6 g/km CO
and 0.56 g/km HC over extended operation. Three-way catalyst systems
are still under development. Their durability has not yet been adequately
demonstrated. Also, they require replacement of conventional carburetors
with electronic fuel injection systems, which raises the price of emission
control.
V.I.2 Noble Metal Composition
The data obtained in this program indicated that platinum
catalysts might produce more sulfate than platinum-palladium catalysts.
This difference, if it does exist, is smaller than the difference found
by Beltzer, e± al. (6), of Exxon Research, between sulfate emissions for
platinum-rhodium and platinum-palladium catalysts.
V.I.3 Catalyst Age
The final factor found to have a significant effect on sulfate
emissions was catalyst age. The effect of catalyst age was studied by
-------
- 17 -
comparing relatively fresh monolithic and pelleted catalysts with
similar catalysts which had been aged 40 000 km or more on the AMA
cycle. While the aged catalysts were still adequate for the control of
CO and hydrocarbon emissions, their sulfate emissions were lower than
those from fresh catalysts by a factor of 50% or more.
The effect of catalyst age on sulfate emissions has been
confirmed in a 20 car fleet test underway as part of Exxon Research's
in-house program. In this test, 20 1975 California model vehicles
equipped with catalysts and, with two exceptions, air pumps, are accumulating
mileage over a prescribed road course at an average speed of about 55 km/h.
Emission measurements, made every 6 400 km, show relatively low FTP
sulfate emissions at 0 km, apparently due to storage of sulfate on the
catalysts, peak sulfate emissions, equivalent to about 10% conversion of
gasoline sulfur to sulfate, at 6 400 km, and a rapid decline in sulfate
emissions thereafter. At 25 600 km most vehicles were emitting about a
third of the sulfate emitted at 6 400 km. Detailed results of this test
which will continue to 80 000 km will be presented in a Feb. 1976 SAE
paper. (7)
These data on aged catalysts are important in two respects.
First, the emission factors measured at low mileages may be adjusted for
deterioration to reflect the average emissions over the useful life of
the catalyst. Second, the fact that sulfate formation activity can be
lost without a corresponding loss in CO and hydrocarbon control strongly
suggests that it should be possible to design a catalyst which can
control CO and hydrocarbon, but would minimize sulfate formation.
V.I.4 Catalyst Type
Monolithic oxidation catalysts emit more sulfate under low
speed operating conditions, such as the FTP, than do pelleted catalysts.
This is due to the greater sulfate storage capacity of pelleted catalysts.
The sulfate stored by pelleted catalysts under low speed operating
conditions is released during the early stages of high speed operation.
After longer periods of high speed operation, pelleted catalysts emit
about the same amount of sulfate as do monoliths. Theie findings may
be useful to the air quality models which would consider actual modes
of usage. It was not the purpose of this study to determine which type
of catalyst would provide the lower overall level of sulfate emissions
in actual use.
V.I.5 Other Factors
The other factors examined: noble metal loading, catalyst
temperature, and residence time of exhaust gas over the catalyst had no
significant effect on sulfate emissions, except under conditions where
CO and HC control also suffered.
V.I.6 Graphic Representation of Results
The information presented above on the effect of catalyst and
system operating variables on S04= emissions is summarized in Figures
V-l to V-4. Figure V-l shows S04= emissions from pelleted catalysts
during the first FTP of the test sequence. Figure V-2 shows average
S04= emissions from pelleted catalysts for the two hour, 96 km/h cruise.
Figure V-3 shows S04= emissions from monolithic catalysts during the
first FTP, and Figure V-4, average S04= for the 96 km/h cruise.
-------
FIGURE V-l
EFFECT OF CATALYST AND SYSTEM OPERATING
VARIABLES ON SULFATE EMISSIONS
PELLETED CATALYST — FTP CYCLE
0.04
co
o
CO
CO
^
LJ
LLl
H-
CO
0.02
LoJ -a
CO £.
LJ
CO
<
CO
Q -o
LJ Q.
I0-
H*
<
a -D
LJ CL
a.
S
LJ
oo
I
-------
FIGURE V-2
EFFECT OF CATALYST AND SYSTEM OPERATING
VARIABLES ON SULFATE EMISSIONS
PELLETED CATALYST -96 KM/H CRUISE
co
<
a:
o
N
CO
CO
CO
LJ
I-
<
U_
_J
CO
o.io
0.05
0.00
LJ -o
CO Q.
LJ
CO
LJ
Q -o
LU Q-
•a
CL
m
o_
<
SHORT
RESIDENCE
TIME
o z a.
xg£
o
LU
h-
vO
I
-------
FIGURE V-3
EFFECT OF CATALYST AND SYSTEM OPERATING
VARIABLES ON SULFATE EMISSIONS
MONOLITH CATALYST — FTP CYCLE
0.04
to
Qi
O
••»
CO
CO
CO
LU
LU
CO
0.02
0
1777*
LU -o
co a.
LU
CO
<
00
Q -o
LU Q-
< CO
LU <
LUl-
CtJ <
IO
LU Q-
O LU
LU Q.
a
— LU
CO «
NJ
O
a.
S
LU
-------
FIGURE V-/J
EFFECT OF CATALYST AND SYSTEM OPERATING
VARIABLES ON SULFATE EMISSIONS
MONOLITH CATALYST-96 KM/H CRUISE
CO
co
CO
CO
UJ
I-
<
u.
_l
z>
CO
0.10
0.05
UJ -a
co a.
ui a:
CO T?
<<
CO
Q -a
< CO
a: <
i o
H-
i
O LJ ~a
zoQ-
OZ>
-ILJ Q.
T3
a_
Q.
•
Q.
*88 KM/H
-------
- 22 -
V.I.7 Thermal Decomposition of Al2(804)3
The thermal decomposition of pure Al2(S04)3 must be understood
in order to clarify the role of catalyst sulfur storage in influencing
vehicular sulfate emissions. It was studied over a temperature range of
370 to 700°C. This is not a controllable phenomenon, like those just
discussed, but it does appear to play an important role in determining
the rate of desorption of sulfur oxides by catalysts. Synthetic exhaust
gas, containing no sulfur oxides, was flowed over a pelleted sample, and
the concentration of S02 and sulfate in the exiting gas determined. It
was found that the total rate of sulfur oxide release increased with
temperature, which was expected. However, the ratio of 803 to S02 also
increased with temperature, which is in the opposite direction from the
gas phase thermodynamic equilibrium predictions. Apparently, the de-
composition of solid Al2(S04)3 is controlled by the thermodynamics of
a different reaction, and the gaseous components were "frozen" in their
non-equilibrium ratio by the rapid sampling. In the presence of a noble
metal catalyst, such as occurs when Al2(S(>4)3 decomposes in the vehicle
catalyst substrate, the gas phase equilibrium is more rapidly established.
V.2 Experimental Procedures
V.2.1 Vehicle Preparation
Two 2050 kg (4500 Ib.) inertia weight, 1975 vehicles meeting the
Federal Interim Standards, were used in this study of the feasibility of
minimizing S04= emissions through modifications to the catalyst control
system. The first was a 5.7 litre (350 CID) V-8 Chevrolet equipped with
a pelleted oxidation catalyst. It was modified by the addition of an air
pump to reduce its CO and HC emissions to approximately 2.1 g/km (3.4 g/mi.)
and 0.26 g/km (0.41 g/mi.) respectively. The second was a 5.7 litre (351 CID)
V-8 Ford equipped with air pump and an Engelhard PTX-IIB (§) monolithic oxidation
catalyst which treated half of the exhaust. It was modified by addition of
a second monolith on the other side of the engine, lowering CO and HC
emissions to about the same level as the first vehicle.
V.2.2 Test Fuels
Two test gasolines were used in this program, one containing
0.032 wt. % sulfur, the other 0.012 wt. %. Both were blended from the
base stock described in Table IV-1 and brought up to the proper sulfur
level by addition of equal quantities of thiophene and di-t-butyldisulfide.
The lower sulfur content fuel was used only in the base case runs reported
in Appendix Tables E-5 and E-6. The 0.032 wt. % fuel, representative of the
average sulfur content of U.S. gasolines, was used in these and all other
subsequent tests.
-------
- 23 -
V.2.3 Test Procedure
Initially, both cars described in Section V.2.1 were broken in
for 3 200 km, using a simulated turnpike driving cycle, on our automatic
kilometer accumulation dynamometers. The 0.032 wt. % sulfur fuel was
used.
After break-in, both cars were tested in a variety of system
configurations, to be described subsequently. Initially, the test
procedure for each new configuration consisted of preconditioning with
the chosen test fuel for 800 km, using either the simulated turnpike
cycle or a simulated city cycle. Both driving modes were based on
control tapes available at Exxon Research and should not be confused
with the LA-4 and HFET driving cycles generated by the EPA. Midway in
the program, it was concluded that results obtained with both cycles
were sufficiently similar to warrant using only one. It was decided
to continue using only the turnplfce schedule for preconditioning. By
so doing, the number of test sequences for each configuration tested
after this decision was cut in half and the preconditioning step was
more rapid.
Following the 800 km preconditioning, each vehicle was put
through a test sequence as follows:
(1) 1975 FTP
(2) 20 minute idle
(3) 2 hrs. 96 km/h cruise
(4) 1975 FTP (after overnight soak)
The two hour 96 km/h cruise was broken in four half-hour periods, with
separate 20 minute samples taken during each period.
Sulfate and S02 emissions were measured by the same techniques
mentioned in Section IV.1.3 and described in detail in Appendix B;
sulfates by the exhaust particulate sampler from dilute exhaust and by
the Goksrfyr-Ross technique from undiluted exhaust; S02 was measured in
the dilute exhaust by the TECO instrument and by H202 bubblers from un-
diluted exhaust. Gaseous emissions were determined using standard FTP
instrumentation; CO by NDIR, HC by FID and NOX by chemiluminescence.
V.3 Results
V.3.1 Base Case Emissions
The two production vehicles, modified as described in Section
V.2.1, were each tested after break-in, under a total of four pre-
conditioning-fuel combinations. The primary aim of these tests was to
determine if the target levels of 2.1 g/km CO and 0^25 g/km HC emissions
had been achieved, and to provide a baseline of 804" emissions against
which the results of future system modifications could be compared.
-------
- 24 -
V.3.1.1 Gaseous Emissions
The four preconditioning-fuel combinations run with these
vehicles provide a total of four test sequences, or eight FTP runs, for
each vehicle, with which to determine average gaseous emissions. Each
car had one FTP run aborted by human error or mechanical failure. The
seven remaining FTP runs for each car gave average emission values of
2.3 g/km CO, 0.18 HC and 2.2 NOx for the pelleted catalyst car, and 2.7
CO, 0.29 HC and 1.9 NOX for the monolith catalyst car. In all cases
these results either satisfied the targets of 2.1 CO, 0.26 HC and 1.9
NOx, or were deemed close enough to continue the study. The detailed
gaseous emission results for these FTP runs, as well as idle and 96 km/h
cruise modes are given in Appendix Table E-5 for the pelleted catalyst
car and E-6 for the monolith catalyst car.
V.3.1.2 Sulfate Emissions
V.3.1.2.1 Raw Exhaust
Sulfate determinations of the raw exhaust, prior to the catalyst,
were made for both cars. This was done, using the Goks^yr-Ross method, to
verify the results obtained in Task II, which showed very low sulfate
emissions out of the engine. In this case the Goks^yr-Ross readings were
taken in one hour segments during the 96 km/h cruise modes, in contrast
to the half-hour segments in Task II. This provided larger samples, an
aid in achieving accurate analyses. The idle modes were sampled in their
entirety, as before. Appendix Tables E-5 and E-6 show these GoksjJyr-Ross
results. It is seen that the pelleted catalyst vehicle had raw exhaust
sulfate levels at or below the values shown in Appendix Table E-l for a
car with a similar engine, at comparable fuel sulfur levels. The mono-
lith equipped vehicle, with a different engine, yielded somewhat higher,
although still low, values for engine-out sulfate emissions.
V.3.1.2.2 Post-Catalyst Sulfate Emissions
The most important sulfate results, of course, are tailpipe
emissions, after the exhaust has passed over the oxidation catalyst.
These are the "filter" values, shown in Appendix Tables E-5 and E-6. During
cruises, the 20 minute filter samples were taken, as in Task II, at half-
hour intervals. A number of interesting observations can be made. First,
the effect of preconditioning mode, turnpike or city driving, appears not
to have had a major effect on S04= emissions. Thus, comparing average
FTP and 96 km/h cruise values for each fuel sulfur level, as shown in
Table V-l, no systematic difference is found between the two modes.
Primarily because of these results, it was decided to use only one pre-
conditioning mode, and the turnpike cycle was chosen since it allowed
more rapid preconditioning. This lack of preconditioning effect is
surprising since much work prior to and subsequent to these experiments
has clearly shown its importance in determining the storage capacity of
the catalyst and hence its SOtt" emission rates. The cycles used here
were sufficiently different, 80 and 32 km/h average speeds, to have led
one to expect dissimilar effects.
-------
- 25 -
Table V-l
Effect of Preconditioning on SO/= Emissions
S04~ Emissions, g/km
City Turnpike
0.012 wt. % 0.032 wt. % 0.012 wt. % 0.032 wt.
Pellet Catalyst
FTP* .003 .006 .002 .007
96 km/h** .029 .060 .037 .063
Monolith Catalyst
FTP* .007 .026 .007 .023
96 km/h** .020 .051 .024 .050
* Average of initial and final FTP
** Average of four half hour tests
Another point worth noting from Appendix Tables E-5 and E-6
is the comparison between the sulfate emissions from initial and final
FTP runs of each sequence. In the case of the pelleted catalyst car,
three of the four sequences showed higher sulfate values for the initial
FTP run. The other sequence showed no change. The monolith catalyst
car showed this effect to a lesser extent, and only with the low sulfur
fuel, not with the 0.032 wt % fuel.
The explanation for these results seems to stem from the fact
that a two hour period of 96 km/h cruise occurred between the initial
and final FTP runs of each sequence. During this high speed cruise,
as will be shown shortly, it appears that sulfur release from the cata-
lyst takes place. The catalyst was therefore "cleaner" for the final
FTP than it was for the initial FTP and tended to retain more sulfur
oxides, resulting in lowered tailpipe emissions. This effect is more
pronounced for the pelleted catalyst than for the monolith because of
the former's greater mass and surface area of alumina. This provides
greater storage capacity for sulfur oxides. Storage capacity may also
explain why the sulfate emissions during the FTP runs were higher for
the monolith than for the pelleted catalyst car. If it is assumed that
the relatively low speed FTP driving mode is a "storage" cycle, then
the pelleted catalyst should retain more sulfate while the lower capacity
monolith allows more to escape.
-------
- 26 -
On the other hand, the 96 km/h cruise sulfate emissions, aver-
aged over the full two hours of each run, were lower for the monolith
catalyst than for the pelleted catalyst. This was due primarily to the
smaller "spike" obtained during the first half hour of the cruise. Both
catalysts tended to give similar results after this period. These re-
sults indicate that less sulfur storage occurs during the preceding low
speed FTP and idle modes for the monolith catalyst than for the pelleted
catalyst, with consequently less release at the beginning of the high
speed cruise. Even after reaching a lined out value, the cruise sulfate
emissions for both catalysts were higher than during their FTP runs,
again indicating that the high speed cruise is not a sulfur storage mode,
compared to the FTP driving cycle. The sulfur balance results discussed
in the next section will show these effects even more clearly;
V.3.1.2.3 Sulfur Dioxide Emissions - Sulfur Balance
Sulfur dioxide emissions from cars need not be controlled,
because of their very small percentage contribution to the total atmos-
pheric sulfur burden coming from all sources, and the fact that this
material is only slowly oxidized to 804" in the atmosphere. It is
still useful however to monitor S02 emissions during S0,~ testing, even
though in the presence of catalysts, a total sulfur balance, comparing
SC>2 + S04= emissions with that predicted from the sulfur content and
amount of fuel consumed, should not be expected due to storage effects.
On the contrary, departures from a 100% sulfur balance indicate whether
a catalyst, during a given driving mode, is retaining or releasing
sulfur oxides. Therefore S02 measurements and their use to calculate
total sulfur balances, can help explain sulfate emission results
obtained with different catalysts, and driving and preconditioning
modes.
Two methods of S02 measurement were used. The TECO instrument
monitored the S02 content of diluted exhaust from the particulate sampler
which had been collected in sample bags, while peroxide bubblers were
used to directly sample post-catalyst diluted exhaust. These techniques
are described in Appendix B and additional tests aimed at comparing the
two are detailed in Appendix C. Each has certain advantages and disad-
vantages. The TECO procedure allows breaking down the two hour cruise
period into half hour periods, permitting a more detailed look at the
phenomenon of sulfur release from the catalyst. The same could have been
accomplished with the bubbler method, but much more chemical analysis
time would have been required. On the other hand, the TECO instrument,
in reading the contents of the sample bags, during all driving modes,
but especially the idle and FTP, was pushed to the limits of its sensi-
tivity. Therefore, the usefulness of its readings is limited primarily
to a qualitative evaluation of the change in S02 emissions during the
half hour periods of the 96 km/h cruise. The bubbler method provided
more sample to work with, hence should be more accurate than the TECO
method. In particular, the two hour sampling period during the 96 km/h
cruise insured a high level of sulfur collection for accurate analysis.
Less sample was collected during the FTP and idle periods, with a re-
sultant less accurate value.
-------
- 27 -
A total sulfur balance can be calculated, using the bubbler
S02 and the filter S04= values, for each driving mode. For the cruises,
only an integrated average is shown in Appendix Tables E-5 and E-6, using
the two hour bubbler results and the average of the four half-hour filter
values. However, the individual half hour TECO values, viewed in con-
junction with the half hour filter results, give a qualitative picture
of the sulfur balance change during the run. It is seen from Appendix
Table E-5 that the FTP values for the pelleted catalyst car were all
less than 100%, confirming that this is a sulfur storage mode. The
idle modes were also, with one unexplained exception, below 100%. On
the other hand, the cruises all averaged over 100%, showing this to be
a sulfur desorption mode. To the extent the TECO values are reliable,
it appears that S02 desorption follows a similar time pattern as does
S0^=, with the highest levels coming off at the beginning of each cruise.
If the monolith indeed shows less storage effect, the differ-
ences in sulfur balance between low speed, FTP runs and the high speed
cruise should be less pronounced than for the pellet catalyst. As shown
in Appendix Table E-6, the sulfur balances for the three test sequences
for which complete analyses are available do show a closer approach
between FTP and cruise than do the pellet balances of Appendix Table E-5.
Thus, the eight FTP runs average a 92% sulfur balance and the three
cruises 135%. In contrast, the eight pellet FTP runs average 23% and
the four cruises 158%. Finally, the four sets of TECO values show little,
if any, change in S02 desorption over the duration of the cruises. This
behavior is similar to the sulfate emission pattern reported in the pre-
vious section for the monoliths.
V.3.2 Effect of Limited Secondary Air
Current vehicle catalyst systems operate with an excess of 02
over the stoichiometric requirements, to promote oxidation of CO and HC.
This additional 02, supplied by lean carburetion and sometimes augmented
by air pump injection into the exhaust gas, also promotes the oxidation
of S02 to 803. In an effort to quantify this effect, and to determine
if the level of excess 02 could be lowered to provide minimum S02
reaction while providing satisfactory CO and HC oxidation, a series of
vehicle studies have been carried out as part of this contract. These
tests were conducted with the base vehicles and catalysts described
previously. In both cases, the air pump was operated only during the
first 135 seconds of the FTP, to provide more rapid and efficient CO and
HC control during this critical portion of the test. Secondary air from
the air pump was vented to the atmosphere during the remainder of the
FTP and during the idle and 96 km/h modes. In one additional test, a
three-way catalyst system with feedback oxygen sensor was also run.
-------
- 28 -
As described in the following sections, these low excess air
systems produced dramatic reductions in the level of sulfate emissions.
At the same time, control of CO and HC emissions was affected to a much
lesser extent. For example, the pelleted catalyst car showed, for the
initial FTP runs, a five-fold reduction in sulfate compared to the base
case, and only a slight increase in CO. The monolith car showed a doubling
of CO, but an eight-fold drop in sulfate emissions. Finally, a three-way
catalyst car gave very low CO and HC emissions over the FTP, and sulfate
emissions of the same order as for non-catalyst cars.
V.3.2.1 Pelleted Catalyst Vehicle
The amount by which excess Q£ levels were lowered was measured
during idle and 96 km/h .cruise modes. At idle, it dropped from a range
of 5.5-7.7% to about 1-3%. At 96 km/h cruise, the changes were more
dramatic; from 3-3.5% to about 0.9%.
V.3.2.1.1 Gaseous Emissions
Beginning with this phase of the program, it was decided to
use only the 0.032 wt. % sulfur content fuel, since this most closely
represented the national average. Appendix Table E-7 presents the gaseous
and sulfate emissions for both turnpike and city cycle preconditioning
modes. The two test sequences give an average FTP CO value of 3.38
g/km, HC of 0.20 g/km, and NOX of 1.96 g/km. These show a slight rise
in CO compared to the base pellet catalyst case, Section V.3.1.1.1, but
no significant change in HC or NOX emissions. The CO increase, in part
at least, may have been due to an enrichment of the carburetor calibration
during the city preconditioned test sequence. The first half hour of
the cruise gave CO levels typical of those from the previous cruise; the
last 1.5 hours were much higher and the subsequent FTP run yielded the
highest CO values of the four FTP runs.
V.3.2.1.2 Sulfate Emissions
The Goks^yr-Ross samples taken before the catalyst were typically
low, and not significantly different than for the base case car. Filter
samples after the catalyst are much lower than the base case results,
however. The two initial FTP runs averaged only 0.002 g/km and the two
final FTP runs 0.001 g/km. In contrast, the base case FTP emissions
with 0.032 wt. % sulfur fuel shown in Appendix Table E-5 averaged 0.010
g/km for the two initial FTP runs and 0.002 for the final two.
The cruise sulfate emissions were also much lower for the
limited secondary air case. The turnpike preconditioned cruise showed
the same initial spike in sulfate emissions, followed by a gradual
decline, but the absolute levels were about half those found for the
base case. The other cruise, following city preconditioning, presented
something of an anomaly. A spike was observed for the first half hour,
although somewhat lower than for the first cruise. After this, however,
sulfate emissions fell rapidly to very low levels. It is believed this
is related to the increase in CO emissions due to accidental carburetor
enrichment, discussed in the previous section. This is a very dramatic
example of the effect of air-fuel ratio on sulfate emissions in the
-------
- 29 -
absence of secondary air. Even ignoring the carburetor drift, and
looking only at the more consistent sequence, it is clear that removing
secondary air after the warm-up phase of the FTP significantly lowered
sulfate emissions, as did complete removal of secondary air during the
96 km/h cruise.
V.3.2.1.3 Sulfur Dioxide Emissions
Emissions of S02 during the initial FTP of each test sequence, Appendix
Table E-7, were higher than those from the final FTP. This meant, of course,
that total sulfur balances were higher also. This contrasts with the base
case pellet vehicle, where initial and final FTP S02 emissions and total
sulfur balances were generally similar for both FTP run's of cne same test
sequence. In addition, the SC>2 emissions and total sulfur balances of
the initial FTP runs were much higher for the limited secondary air case
than for the base case. This indicates that, when less oxygen is present,
the pellet catalyst has less tendency to store sulfur even during the
low speed FTP mode, and under certain conditions, it may actually re-
lease a small amount of previously stored sulfur.
The 96 km/h cruise results snow cnat the tendency to release
sulfur in the absence of secondary air is accentuated at high speeds.
The consistent sequence showed a sulfur balance of 199%, showing the by
now usual initial spike. When the carburetion began to richen out, this
rose to 267%, and the spike effect was washed out by the change in air-
fuel ratio. Both values are higher than those observed for the base
case at the same fuel sulfur content. This "cleaning" process was
apparently so effective that the final FTP runs, following the cruise
modes, were now capable of again storing sulfur, even with limited
secondary air. It would seem then that the sulfur content of the catalyst,
when sulfur sorption or release occurs* is a function of the level of
oxygen in the exhaust gas.
V.3.2.2 Monolith Catalyst Vehicle
The limited secondary air procedure was tested on this vehicle
using only the turnpike preconditioning mode, with the 0.032 wt. %
sulfur content fuel. Gaseous, sulfate and S02 emissions are given in
Appendix Table E-8. Excess 02 levels, which had been about 77 at idle
and 4-4.5% at 96 km/h cruise, when the air pump had been used, fell to
around 2 and 0.7% at idle and cruise respectively when the pump was vented.
V.3.2.2.1 Gaseous Emissions
The FTP CO emissions rose more with limited secondary air,
compared to the base case, than occurred with the pellet catalyst vehicle.
Thus the average CO value for the limited secondary air monolith catalyst
car was 5.3 g/km, compared to only 2.7 g/km for the base case monolith
catalyst car. Average HC and NOx emissions did not change, however, from
the base case.
-------
- 30 -
The increase in CO emissions may be due to the fact that the
car used for the monolith catalyst studies was originally equipped with
an air pump, and its carburetion Tnay have accordingly been set relatively
rich. The car equipped with the pellet catalyst was designed for use
without an air pump and may have had a leaner carburetor calibration
to compensate. Unfortunately, no measurements are available of average
air-fuel ratios over the FTP.
V.3.2.2.2 Sulfate Emissions
Sulfate emissions from this car were much lower than from the
same vehicle using an air pump full-time. ~The two FTP runs gave identical
sulfate values, 0.003 g/km, compared to an average of 0.023 at the
same fuel sulfur level for the full air case. Cruise results were only
between 0.003 and 0.005 g/km, compared to an average 0.050 for the full
air case. No pattern was observed for the four separate half-hour
cruise samples. These results are in agreement with those obtained on
the pellet catalyst car, showing that limiting excess air has a significant
effect in reducing sulfate emissions from catalyst-equipped cars.
The Goks^yr-Ross pre-catalyst samples showed unusually high
levels of sulfate emissions in the raw exhaust, about equivalent during
cruise, to the post-catalyst filter samples. The reason for this is not
clear. Some type of experimental problem is the most plausible explana-
tion at this time.
V.3.2.2.3 Sulfur Dioxide Emissions
Emissions of S0£ during the cruise were, as indicated by the
TECO values, constant. The more quantitative H202 bubbler results
yield, together with the average of the four filter sulfate values, a
sulfur balance of about 120%. The two FTP runs show a puzzling differ-
ence, with the first showing a sulfur balance of 32% and the last,
109%. In the full air pump case with this car, Appendix Table E-6,
sulfur balances were similar for initial and final FTP runs. No expla-
nation is available for this finding, and insufficient data exist to
confirm its reality.
V.3.2.3 Three-Way Catalyst Results
Since limited excess 02 in the exhaust has been found to be
effective in minimizing S04= emissions, a logical extension of this approach
is operation with no excess 02» at the stoichiometric point. The exhaust
can then be passed over a catalyst, referred to as a "three-way catalyst"
(TWC) which is designed to react the CO, H2, NOX and 02 in the exhaust to
form H20, C02, and N2-
-------
- 31 -
This approach, which is still in an early stage of development,
requires very close control of air-fuel ratio, as close to stoichiometric
as possible. If the exhaust is too rich, insufficient 02 will be available
to oxidize CO and HC, and these emissions will rise. If the exhaust is too
lean, NOx reduction, which does not proceed in the presence of excess 02,
will cease. Since this level of control is beyond the capability of any
production open-loop fuel metering system, a closed-loop control means is
necessary. The most advanced concept for providing closed-loop control
uses an 02 sensor in the exhaust.
To determine whether closed-loop control of exhaust 02 content
would provide the desired reduction in S04= emissions, a Volvo equipped
with a feedback control system using a Bosch 02 sensor and a three-way
catalyst developed by Engelhard was tested. All tests were run with fuel
containing about 0.03 wt. % sulfur.
This car was tested after the system had accumulated 6 400 and
25 000 km of aging. An FTP at the lower kilometer level gave results of
1.04 g/km CO, 0.14 HC, and 0.56 NOX. Sulfate emissions were only 0.8
mg/km. At the higher kilometer level, a similar FTP showed 2.12 g/km CO,
0.23 HC, 0.46 NOX, and 1.7 mg/km sulfate. An EPA Highway fuel economy
test at 25 000 km was also run, yielding only 0.4 mg/km of sulfate emissions.
These few tests indicate the potential of the three-way catalyst
system to simultaneously control gaseous emissions while producing only very
low amounts of sulfate. Much more work is needed, however, to demonstrate
the durability and practicality of this pystem for large scale use.
V.3.3 Effect of Residence Time
The effect of residence time, defined here in terms of space
velocity, was also investigated on vehicles using both pelleted and
monolithic Pt-Pd catalysts. For the pelleted catalyst, high space
velocity was created by replacing the standard 4.25 litre converter with
a smaller 2.6 litre unit. The smaller converter normally contains
higher noble metal concentration catalyst. For these tests, it was
filled with the lower noble metal concentration catalyst normally used
in the larger converter. For the monolithic catalyst, lower space
velocity was created by placing two monoliths in series on each side of
the engine.
V.3.3.1 Pelleted Catalyst Vehicle
V.3.3.1.1 Gaseous Emissions
The results for this test sequence, conducted with the 0.032
wt. % sulfur fuel and turnpike preconditioning, are shown in Appendix
Table E-9 and can be compared with those shown for the same vehicle, but
using a full size 4.25 litre catalyst converter, in Appendix Table E-5.
-------
- 32 -
Average FTP values for the small converter, high space velocity vehicle
are 2.0 g/km for CO and 0.16 for HC. These compare closely with the
average values of 2.3 and 0.18 obtained with the conventional car,
indicating little, if any, effect of space velocity, over the range
investigated here, on FTP gaseous emissions of CO and HC. At higher
overall flow rates, 96 km/h cruise, where a greater space velocity
effect on oxidation efficiency would be expected, there is a very slight
increase in CO but essentially no effect on HC. Thus, the CO emissions
with the smaller converter at cruise mode average 0.07 g/km, compared to
0.03 g/km for the larger converter. Hydrocarbon emissions averaged 0.02
g/km for both configurations.
V.3.3.1.2 Sulfate Emissions
Surprisingly, the initial and final FTP sulfate values were
higher than those measured for the base case with the same fuel. The
base car averaged 0.010 g/km for the two initial FTP runs made with
0.032 wt. % sulfur fuel, but the smaller converter gave a value of 0.018
g/km sulfate. Similarly, the two final FTP runs on the base car yielded
an average of 0.002 g/km sulfate, compared to 0.008 for this configuration.
Efficiency of S02 oxidation to sulfate would be expected to decrease, or
stay the same with increasing space velocity. The reason for this
anomalous finding, or whether it is real, is not known. Perhaps the rate
of removal of sulfate from the exhaust gas via reaction with the catalyst
wash coat is affected more by residence time than is the oxidation of S02
to 803. This would explain the escape of more sulfate at the shorter
residence time.
During the 96 km/h cruise, the typical initial spike of sulfate
was observed, and the two hour average was 0.068 g/km, comparable to the
0.062 found for the base car.
V.3.3.1.3 Sulfur Dioxide Emissions
The bubbler values show, for the initial and final FTP runs, a
drop in S02 emissions, leading to a fall-off in total sulfur balance
from 33% for the initial FTP to 15% for the final. This pattern differs
from the base configuration, where first and last FTP run sulfur balances
were similar. During the cruises, the combined bubbler value and the
average of the four filter samples shows a total average sulfur balance
of 158%, comparable to that found with the base car.
V.3.3.2 Monolith Catalyst Vehicle
In order to achieve a space velocity with the monolith catalyst
lower than that used for the base case, a second catalyst was added to
each side of the engine, in series with the original catalysts. This
-------
- 33 -
halved the space velocity, but also doubled the amount of noble metal
in the system.
V.3.3.2.1 Gaseous Emissions
Halving the space velocity and doubling the noble metal content
had a clearly beneficial effect on CO and HC emissions, lowering the
average FTP values to 1.5 and 0.13 g/km for CO and HC respectively, as
shown in Appendix Table E-10, from base case values of 2.7 and 0.29.
V.3.3.2.2 Sulfate Emissions
Sulfate emissions from the FTP runs were lowered, compared to
the base case, by this decrease in space velocity. This is all the more
surprising in view of the doubling of noble metal compared to the base
case. The average for the double catalyst configuration is 0.015 g/km,
while the base case was 0.025 for the 0.032 wt. % sulfur fuel. These
lower results are directionally consistent with the effect of increasing
space velocity, found with the pellet car to increase sulfate emissions.
One possible explanation for these seemingly anomalous findings has al-
ready been discussed in Section V.3.3.1.2, namely, that under FTP con-
ditions at least, sorption of sulfate on the catalyst substrate is the
determining factor for tailpipe emissions rather than the kinetics of
sulfate formation over the catalyst. Since sorption rate should increase
with decreasing space velocity, then, under this theory even though the
rate of formation is increasing at lower space velocities, the rate of
sorption is increasing even faster, resulting in a net decrease in tailpipe
emissions.
At 96 km/h cruise, the average sulfate emissions over the two
hour test interval was 0.055 g/km, compared to 0.050 for the base case
with the same fuel. There seemed to be a more pronounced sulfate spike
during the first half hour, due perhaps to the greater sulfur storage
capacity of the doubled catalyst beds compared to the base case. A drop
in S02 emissions, leading to a fall-off in total sulfur balance from an
initial 111% to a final FTP value of only 52%. This is at variance with
the base case result, where first and last FTP sulfur balances were
similar, and no explanation is available.
During the two hour 96 km/h cruise, the average sulfur balance
was 109%, in good agreement with the 113% shown in Appendix Table E-6
for the base case.
V.3.4 Effect of Catalyst Noble Metal Composition
This report has already discussed the gaseous and sulfur oxide
emission characteristics of vehicles equipped with monolith or pellet
catalysts containing Pt-Pd mixtures as their active metals. In this
-------
- 34 -
section, we will discuss the behavior of similar catalysts, but containing
only Pt. 6
The Pt-Pd monoliths were replaced with similarly sized experimental
monoliths, at the same metal loading but containing only Pt. After the
standard aging and preconditioning steps, the Pt catalyst was put through
the test sequence for gaseous and S04= emissions.
For the pellet comparisons, it was necessary to switch from
the standard 4.25 litre (260 in3) converter supplied with the production
car to a 2.6 litre (160 in3) converter. An experimental Pt pellet
catalyst was made available by a catalyst manufacturer, but there was
insufficient quantity to fill the larger reactor. Therefore, the Pt
versus Pt-Pd comparison was made with the smaller reactor used for both
catalysts. A discussion of residence time effects themselves was given
in Section V.3.3, and the shorter residence times here should not affect
the pellet comparisons discussed in their section.
V.3.4.1 Pelleted Catalyst Vehicle
V.3.4.1.1 Gaseous Emissions
This system must be compared with the pelleted car described
in Section V.3.3.1, since both use the small 2.6 litre catalyst canister.
The only difference between the two sets of runs is the noble metal
composition; Pt here and Pt-Pd in Section V.3.3.1. Noble metal loading
remains the same. The FTP gaseous emissions shown in Appendix Table fi-
ll are, on average, 2.4 g/km CO and 0.13 g/km HC. These are slightly
higher than the 2.0 for CO reported in Section V.3.3.1 for the Pt-Pd
catalyst, but slightly lower than the 0.16 g/km found for HC emissions.
Both quantities are subject to considerable uncertainty in the course of
FTP testing, and a more sensitive indicator of small activity differences
is probably presented by the steady state 96 km/h cruises. The Pt
catalyst averaged 0.05 g/km CO and 0.012 g/km HC over this driving mode.
The Pt-Pd showed 0.07 g/km CO and 0.019 g/km HC. There appears then to
be a slight advantage for the Pt, although not enough to greatly influence
the FTP runs.
V.3.4.1.2 Sulfate Emissions
As with nearly all pellet catalyst test sequences, here too
the initial FTP showed higher sulfate emissions than the final one, which
follows the 96 km/h cruise. In this case, the values were 0.034 g/km
and 0.007 respectively. The initial value was higher than the comparable
run with the Pt-Pd catalyst, but the final FTP runs were similar for the
two catalysts.
-------
- 35 -
At 96 km/h cruise conditions, a larger initial spike of sulfate
was seen with the Ft catalyst, but the subsequent emissions were comparable
to those from the Pt-Pd catalyst. This larger release of sulfate during
the beginning of the cruise could be indicative of more sulfate being
produced, and stored, during the preceding lower speed driving modes.
As will be seen below, the 502 release results tend to confirm this
hypothesis.
V.3.4.1.3 Sulfur Dioxide Emissions
The bubbler values for S02 emissions during the two FTP runs
are equivalent. However, the much greater sulfate release during the
first FTP yields a total sulfur balance of 40%, compared to only 22% for
the final FTP. The most striking feature is the high level of S02
release during the cruise mode. On an overall average, taking the
integrated bubbler value of 143% for S02, and the average of the four
filter sulfate samples, the total sulfur balance is 207%. It is clear
from the TECO reading for the first half hour, as well as the first
filter sample, that much of this SOX excess is released during the beginning
of the cruise.
The magnified storage-release phenomenon seen with this Pt
catalyst, compared to its Pt-Pd counterpart, may be due to several
factors. Platinum may be more active than Pt-Pd for conversion of S02
to sulfate, especially at low speed driving modes. This, in turn, could
lead to more storage on the catalyst, since sulfate should react with
the substrate more easily than does 802* Then under higher speed and
temperature driving modes, this extra sulfur is released. Alternatively,
the storage-release effects may be related to the catalyst substrate
itself. The two catalysts compared here were made by two manufacturers
and, in addition to their noble metal changes, may have different
quantities and types of substrate or manufacturing techniques. Not
enough information is available to confirm these speculations, nor even
to confirm if the observed storage-release behavior differences are real.
V.3.4.2 Monolith Catalyst Vehicle
V.3.4.2.1 Gaseous Emissions
The two FTP runs, shown in Appendix Table E-12, gave average
CO and HC emissions of 1.9 and 0.18 g/km respectively, considerably
lower than the values of 2.7 g/km for CO and 0.29 g/km for HC found with
the Pt-Pd catalyst in the base case tests reported in Section V.3.1.2.1.
Because of the uncertainties inherent in FTP gaseous emission results,
the steady state emissions at 96 km/h cruise were also compared to
determine if an activity difference between the two catalysts existed.
The Pt catalyst averaged 0.08 g/km for CO and 0.01 g/km for HC. In
contrast, the Pt-Pd catalyst averaged 0.21 g/km CO and 0.03 g/km HC.
There appears then to be a real activity advantage for the Pt catalyst.
-------
- 36 -
V.3.4.2.2 Sulfate Emissions
In contrast to the four Pt-Pd base case FTP runs with 0.032 wt. ',
sulfur fuel, the two FTP runs with the Pt catalyst were not similar.
Here the final FTP value of 0.042 g/km was significantly higher than the
initial value of 0.024. Together they averaged 0.033 g/km, compared to
the Pt-Pd catalyst average of 0.025. In addition, the 20 minute idle
period yielded 0.20 g for the test, compared to only 0.03 g for the idle
period with the Pt-Pd catalyst. On the other hand, at 96 km/h cruise,
the Pt catalyst averaged only 0.042 g/km, compared to 0.051 for the Pt-
Pd sample. It would appear from these results that the Pt catalyst may
produce slightly more sulfate at lower speed driving and slightly less
at high speed driving than does the Pt-Pd catalyst.
V.3.4.2.3 Sulfur Dioxide Emissions
The two FTP runs showed similar S02 release characteristics,
typical of monoliths. The overall sulfur balances were somewhat higher
for the final FTP, however, because of the larger sulfate release. The
bubbler value for the 96 km/h cruise, combined with the four half hour
filter sulfate samples, gave an average sulfur balance of 94%, with no
apparent trend with time, based on the TECO and filter values.
V.3.5 Effect of Changes in Catalyst Operating Temperature
The effect of catalyst operating temperature on S04= production
has been investigated on our two test vehicles. The car tests involved
raising the operating temperature of a pelleted catalyst by insulating
and heating the exhaust pipe upstream of the converter and retarding
spark timing, and lowering the monoliths' operating temperature by
moving them back about 100 cm from their normal post-manifold position
and externally cooling the exhaust lines. In all other respects, the
vehicles were similar to their basic configurations.
In the pellet test, the maximum catalyst outlet temperatures
were raised by 110°C, from 450°C to 560°C, for the FTP. At 96 km/h
cruise, an identical rise was observed, from 620°C to 730°C. The monolith
maximum temperatures were lowered from the normal 610°C to 510°C during
the FTP. For the cruise mode, the maximum temperature dropped from
645°C to 540°C. In all other respects, these tests were conducted in
the same way as were the base case runs, with Pt-Pd catalysts and 0.032
wt. % sulfur fuel.
V.3.5.1 Pellet Catalyst Vehicle
V.3.5.1.1 Gaseous Emissions
The two FTP runs differed considerably in CO emissions, due
probably to some undetected variations in the vehicle operation during
cold-starting. The only thing that can be said is that the average CO
-------
- 37 -
emissions, 2.0 g/km, were similar to the base case value of 2.3. It is
not possible from this to tell if the higher operating temperature sig-
nificantly improved CO conversion. The cruise results, averaging 0.024
g/km, were also slightly lower than the base case cruise average of
0.031. Emissions of HC, however, did show A .large decrease at higher
temperature. The two FTP results averaged 0.09 g/km, compared to 0.18
for the base case. At cruise, the higher temperature case yielded an
average of only 0.004 g/km, compared to the base case average of 0.018.
The greater effect of temperature on HC emissions than CO emissions is
not surprising, given the easier homogeneous oxidation of the former.
Since the spark was retarded for this test, NOX emissions also
came down, to an average of 1.5 g/km for the FTP, from the base case
average of 2.2.
V.3.5.1.2 Sulfate Emissions
The two FTP runs, as illustrated in Appendix Table E-13,
showed the usual pattern of higher sulfate emissions on the first, and
lower on the last FTP. Their average value of 0.007 g/km was not significantly
different than for the four base case FTP runs with the same fuel, 0.006
g/km. The cruise results also showed the usual pattern for pellet
catalysts during cruise, a high spike initially, which levels off thereafter.
Neither the height of the spike, nor the two hour average, 0.055 g/km,
differed from the base case, which had an average of 0.051.
V.3.5.1.3 Sulfur Dioxide Emissions
In contrast to the four base case FTP runs with the 0.032 wt.
% sulfur fuel, which showed similar S02 desorption properties, regardless
of whether they were first or last in the test sequence, at the higher
operating temperature condition the first FTP resulted in a much higher
rate of S02 release than did the final FTP. The initial high temperature
FTP released somewhat more S02 than did the lower temperature FTP runs,
while the final FTP was much lower than the base case results. Combined
with their respective sulfate filter results, these two FTP runs showed
sulfur balances of 47 and 7%.
At cruise, the bubbler results also showed a higher rate of
S02 release than for the lower temperature base case. Combining the S02
with the average sulfate for the four half hour filters gives an overall
average total sulfur balance of 163%. Although the total sulfur release
occurring during the first FTP and the cruise are slightly higher than
for the base case sequences, they don't seem sufficiently high to have
cleaned off the catalyst enough to account for the almost complete
removal of sulfur from the exhaust that occurred during the last FTP.
-------
- 38 -
V.3.5.2 Monolith Catalyst Vehicle
V.3.5.2.1 Gaseous Emissions
Lowering the operating temperature during the FTP had a harmful
effect on CO emissions, but not apparently on HC emissions. Values of
3.8 and 0.32 g/km respectively were obtained, as shown in Appendix Table
E-14, compared to 2.8 and 0.30 for the higher temperature base case FTP
runs.
Neither the CO nor HC emissions during cruise were affected,
for the monolith catalyst, by lowering temperature. The base case
values, at a temperature of about 645°C, averaged 0.21 g/km CO and 0.03
g/km HC. In this test, at a temperature of about 540°C, the emission
values were 0.23 and 0.03 g/km.
V.3.5.2.2 Sulfate Emissions
The FTP sulfate emissions at the lower temperature showed a
sharp drop compared to base case results, averaging only 0.010 g/km with
0.032 wt. % sulfur fuel, compared to 0.024. In addition, the initial
FTP run was only half the level of the final run, whereas the base case
sequences had shown uniformity between first and last FTP runs.
At cruise, however, temperature did not have a significant
effect on sulfate emissions, averaging 0.055 g/km compared to 0.051 at
the higher temperature base case.
V.3.5.2.3. Sulfur Dioxide Emissions
At the lower temperature, the two FTP runs showed S02 releases
comparable to the base case, as were the total sulfur balances of 77 and
83% for the first and last runs. At cruise, however, the S02 release
was only about half that produced at the higher temperature, with a
resulting total average sulfur balance of only 84%, compared to 113% for
the base case. Such a showing of comparable sulfur storage at low speed
driving modes, and less release at high speed, could indicate a higher
steady state catalyst sulfur content is achieved at lower operating
temperatures.
V.3.6 Effect of Higher Noble Metal Loading
The catalysts used for the studies discussed so far were, as
far as known, of similar metal loadings for each type. That is, the
various monoliths had the same loading, and so did the various pellet
samples. In this section, we will discuss a test aimed at determining
-------
- 39 -
if noble metal loading plays a role in S04~ formation over the catalyst.
The standard pelleted catalyst vehicle results shown in Section V.3.1.1
were compared with results obtained by emptying the 4.25 litre catalyst
container and refilling it with another batch of Pt-Pd catalyzed pellets,
which contained a 60% greater concentration of noble metal. It is
believed the two contents were about 1.5 and 2.4 grams per vehicle. The
higher loading catalyst was designed for use in the smaller, 2.6 litre
canister, it being desired in commerical applications to use the same
total amount of noble metal per car, regardless of catalyst canister
size. A similar experiment was not possible with monolith catalysts
because of the then unavailability of such catalysts at different metal
concentrations.
V.3.6.1 Gaseous Emissions
No significant effect was found on FTP emissions of CO and HC.
Although the initial and final FTP runs varied, due to run-to-run variation,
their average values of 2.0 and 0.16 g/km for CO and HC respectively, as
shown in Appendix Table E-15, were similar to the base case values of
2.3 and 0.18. Under cruise conditions, however, the CO and HC emissions
were somewhat lower than the base case results, averaging 0.024 g/km CO
and 0.012 HC compared to 0.031 and 0.018.
V.3.6.2 Sulfate Emissions
Sulfate emissions were also similar to the base case results.
The initial FTP gave higher values than the final, and together they
averaged 0.008 g/km, compared to a base case average of 0.006 g/km. At
96 km/h cruise, the initial spike was seen, and the average of 0.066
g/km was not significantly different than the base case average of 0.061
g/km with the 0.032 wt. % sulfur fuel.
V.3.6.3 Sulfur Dioxide Emissions
As with gaseous and sulfate emissions, not much difference was
noted from the base case. Sulfur dioxide emissions were slightly lower
during the FTP runs, with a resultant lower overall sulfur balance, but
the effect Is not deemed large. During the 96 km/h cruise, almost
identical sulfur balances were achieved for this run and the base case
runs.
V.3.7 Effect of Catalyst Aging
It is well known that automotive oxidation catalyst activity
for CO and HC tends to decrease as the catalyst ages. A series of runs
were made to determine if the activity of such catalysts changes with
age for 862 oxidation. In addition, the sulfur oxides storage properties
of aged catalysts were compared with those of fresh catalysts.
-------
- 40 -
The standard vehicles were modified for these tests by replacing
their low mileage monolithic and pelleted catalysts with similar catalysts
which had seen previous service for at least 40,000 km. The aged pair
of monoliths had seen respectively 41,400 and 49,000 km. The former had
been run using fuel containing 0.025 g/litre of lead over the AMA cycle.
The latter was operated for 45,000 km, also on the AMA cycle, with lead
sterile fuel, then 480 km with fuel at 0.125 g/litre of lead, and finally
3500 km on lead sterile fuel again, which restored its CO and HC activity
substantially to the pre-lead level. The aged pelleted catalyst had
been mounted on a car driven 40,000 km on the AMA durability cycle,
using fuel containing 0.0075 g/litre lead.
The pelleted catalyst, after aging, showed no increase in FTP
gaseous emissions, but did have a 33% drop in sulfate emissions. The
monolith catalyst also showed no change in FTP gaseous emissions with
aging, but a large decrease in sulfate emissions, about 15%. It appears
then that predictions of sulfate emissions from the total vehicle population,
which have been based on measurements from fresh prototype or certification
cars, will overestimate the levels from a population of normal age distribution.
V.3.7.1 Pellet Catalyst Vehicle
V.3.7.1.1 Gaseous Emissions
As shown in Appendix Table E-16, the aged pellet catalyst was
tested under both turnpike and city driving preconditioning modes since
these sequences were run early in the program before the decision was
made to concentrate on the former preconditioning mode. The average CO
emissions for the four FTP runs was 1.9 g/km and for HC, 0.18. These
are very similar to the base case values of 2.3 and 0.18. Although not
directly comparable, since the catalysts used here and that for the base
case were different, these results indicate that 40,000 km of aging
under the conditions employed here had little or no effect on FTP emissions
of CO and HC. A more sensitive test, steady state activity at 96 km/h
cruise, showed average CO emissions of 0.09 g/km, compared to the base
case average of 0.03. Emissions of HC were 0.03 g/km, against the base
case average of 0.02. These results show that some activity loss did
occur upon aging, although not enough to influence FTP results.
V.3.7.1.2 Sulfate Emissions
Although FTP gaseous emissions were equivalent for the fresh
and aged catalyst samples, sulfate emissions showed a. decrease due to
aging. The familiar pattern of higher initial FTP results was seen for
each test sequence, but the absolute values were generally lower.
Therefore, the overall average was only 0.004 g/km, compared to the base
case of 0.006. Under cruise conditions, the sharp initial spike was
observed, but again absolute levels tended lower. The overall averages
were 0.049 g/km for the aged catalyst and 0.061 for the fresh.
-------
- 41 -
V.3.7.1.3 Sulfur Dioxide Emissions
The two initial FTP runs show S02 emissions comparable to the
base case, as are their overall sulfur balances. However, for some
reason not understood, the final FTP runs gave significantly lower
sulfur balances than their fresh catalyst counterparts. It would seem
that the aged catalyst, after "cleaning" by the high speed cruise, has a
greater tendency to remove sulfur oxides from the exhaust gas than does
a fresh catalyst. Perhaps the catalyst surface has been roughened,
providing more area for sulfur oxide sorption.
During the 96 km/h cruise, however, no more S(>2 is released
from the aged catalyst than from the base case fresh catalyst, averaging
99% from the former and 91% from the latter. Overall sulfur balances
are also similar, averaging 134% for the two aged catalyst test sequences
and 139% for the two fresh catalyst sequences run with the 0.032 wt. %
sulfur fuel.
V.3.7.2 Monolith Catalyst Vehicle
V.3.7.2.1 Gaseous Emissions
The FTP emissions of CO and HC from the aged catalysts, shown
in Appendix Table E-17, were similar to those from the fresh monoliths.
The four runs gave an average of 2.9 g/km CO and 0.28 g/km HC, compared
to 2.7 and 0.29 for the base case. One of these runs, the initial FTP
in the test sequence which followed city driving preconditioning, appeared
to go unusually rich, as evidenced by its unusually high CO and HC and
low NOx emissions. This appeared to affect S02 and sulfate emissions
also, as will be discussed below.
Although no difference was observed in FTP gaseous emissions
between fresh and aged catalysts, as discussed previously, the FTP is
not as sensitive an indicator of catalyst activity as is steady state
operation. The two 96 km/h cruises with the aged catalysts showed
average CO and HC emissions of 0.64 and 0.066 g/km respectively. In
contrast, the four cruise modes run with the fresh catalysts had averages
of only 0.21 and 0.033 g/km respectively, indicating some loss in activity
for the aged catalysts.
V.3.7.2.2 Sulfate Emissions
The aged catalysts showed a significant drop in sulfate emissions,
for both the FTP and cruise driving modes. The average of the four FTP
runs was 0.009 g/km, compared to 0.025 for the fresh catalysts. At 96
km/h cruise, the comparison of the averages was 0.033 versus 0.051 g/km.
The highest FTP value, 0.014 g/km, corresponded to the one
which seemed to be unusually rich based on its gaseous emissions, as
discussed in the previous section. The next section will discuss a
-------
- 42 -
possible reason for this connection. It is also interesting that the
first half hour cruise segment following this rich FTP run was unusually
high in sulfate emissions, 0.053 g/km.
V.3.7.2.3 Sulfur Dioxide Emissions
Of the four FTP runs, again the rich run gave anomalous
results, in this case unusually high S02 emissions, corresponding to
144% of the fuel sulfur consumed. In contrast, the other three runs
ranged from 72 to 97%. Again, excepting the anomalous run, total sulfur
balances ranged from 75 to 102%. The anomaly was 152%.
During the two cruise modes, the total sulfur balances,
obtained from the bubbler values for S02 and the averaged filter samples
for sulfate, were 112 and 125%. The latter includes the one high half-
hour segment immediately following the rich FTP run. These values, both
for the cruises and FTP runs are in generally good agreement with the
results from the fresh catalysts, indicating aging had little effect on
the relatively small storage properties of monolith catalysts.
The enriched carburetion during the anomalous FTP run may have
contributed to the unusually high level of S02 release observed. To the
extent the carburetor abnormality persisted into the cruise mode, it may
have caused additional S02 release then also. The TECO results do
indicate that the first half hour of this cruise released twice as much
S02 as the subsequent intervals. The effect of enrichment on sulfur
release from catalysts is not surprising in itself. It was observed
previously in the tests run with pelleted catalyst and limited excess
air, Section V.3.2.1. However, in that instance, while S02 and total
sulfur release was going up, sulfate decreased sharply. In the present
case, sulfate increased along with S02« Of course, the major difference
between the two runs is the full time use of an air pump in this monolith
run. It may be that in the presence of sufficient excess oxygen, the
additional CO can still trigger catalyst release of sulfur, but more of
the freed S02 can now be converted to sulfate during its passage through
the catalyst bed.
V.3.8 Thermal Decomposition of Al2(S04)-^
Because of the importance of sulfur release from oxidation
catalysts, presumably through the decomposition of Al2(S04)3, several
runs were made in a laboratory set-up, described in Section VI.4.1, in
an effort to learn more about the mechanism of this reaction. In one
run, a 13 cm^ charge of pellets, made from pure Al2(S04)3, was placed in
the tube reactor. A synthetic exhaust gas, identical to that detailed
in Section IV.4.1 was used, at a space velocity of 100,000 v/v/hr.,
except that no S02 or sulfate was added. Water of hydration was driven
-------
- 43 -
off, and the material then heated progressively to 370, 480, 590, and
700°C. At each temperature, the S02 was monitored, with the TECO unit
described in Appendix B, until it reached a constant level. Measurements
of H2S04 were then made for a one hour period with the Goks^yr-Ross method,
also described in Appendix B, Section B.2.3. Table V-2 shows the results.
It is seen that the quantity of sulfur released increases with temperature,
which is not surprising. What is unexpected is the rapid increase in the
ratio of sulfate to S02- Given the equilibrium relationship between S03
and S02» which shows that 863 decomposes with increasing temperature
(Appendix Figure A-l) , just the opposite ratio change would be predicted.
Table V-2
Thermal Decomposition of Aluminum Sulfate
3.2 mm Al2 (804)3 pellets Space Velocity = 100,000 v/v/hr.
Synthetic Exhaust Gas Composition: 12% C02; 12% H20; 3% 02; Balance N2
S02
Temperature, Concentration, Concentration, _ S03/S02 _
_ ^C _ _ ppm _ _ ppm _ Found Equilibrium
370 1.0 1.3 1.3
480 2.0 5.3 - 2.6 ^20
590 2.5 23.7 9.5 ^ 2.3
700 10.0 133 13.3 ^ 0.5
These results can probably best be explained by assuming that
the Al2(S04)3 is not decomposing by the straight thermal mode;
Al2(S04)3 •* A1203 + 3S03
but rather by a hydrolysis mechanism, given the high concentration of
in the synthetic exhaust gas;
Al2 (804)3 + 3H2° •* A1203 + 3H2S04
Thus the initial release of sulfate from the Al2(S04)3 is not regulated
by the S03/S02 equilibrium relationship, but rather the equilibrium for
steam hydrolysis, which can explain the relatively high concentrations of
sulfate.
Of course, once produced at high temperatures, there is a
thermodynamic driving force for the decomposition of H2S04;
-------
- 44 -
H2S04 •*• H20 + 80s
The equilibrium curve in Appendix Figure A-2 shows that at a temperature
as low as 400°C, only about 10% of the H2S04 will remain undissociated.
With H2S04 decomposition yielding free 803, it would now be expected that
the S03/S02 equilibrium would take over and lead to a rapid rise in S02
concentration, especially at the higher temperatures. The fact that this
was not observed leads to the conclusion that the hydrolysis reaction is
controlling and that probably the decomposition of 803 to S02 does not
occur rapidly enough to establish the equilibrium ratio in the short
residence time available before the sampling train.
It should be borne in mind, of course, that these results were
obtained with pure Al2(S04)3, in the absence of any catalyst. In the
case of a vehicle oxidation catalyst, where the ^2(864)3 would be in
close contact with active metal catalyst surfaces, it is likely that the
803 •*• S02 reaction rate would be enhanced during the reaction products
release from the catalyst bed, and a 803/802 ratio closer to the equilibrium
value would be measured.
-------
- 45 -
VI. Task IV - Feasibility Studies of Sulfate Removal
From Exhaust Gas By Traps
This task consisted of a feasibility study of the use of a
solid sorbent trap, located in the exhaust system downstream of the
oxidation catalyst, for removal of sulfate from the exhaust gas. The
program consisted of both vehicle studies of the durability and performance
of such traps and laboratory screening of potential sorbent materials.
Four vehicle tests were run for intervals up to 42 000 km. The sorbent
particles used were the following:
• 85% CaO/10% Si02/5% Na20 pellets
• 85% CaO/10% Si02/5% NazO rings
• CaCOs chips
• ZnO pellets
These and a number of other candidate sorbent materials were
also screened for sulfate removal activity, under several operating
conditions, in a laboratory reactor. The materials tested, in addition
to the above, were:
CaO
CaC03
Mn02
MgO
MgC03
85% MgO/10% Si02/5%
A1203
BaO
80% CaO/20% Si02
ZnO
CaSi03
Ca(OH)2
VI.1 Summary of Results
VI.1.1 Vehicle Tests
A total of four vehicle durability runs were made, using
oxidation catalyst-equipped vehicles with trap canisters containing
85 CaO/10 Si02/ 5 Na20 (benchmark), ZnO and CaC03. The benchmark material*
was used in the form of both pellets and rings. The benchmark pellet
test was run for a total of 42 000 km. It was found that nearly all of
the sulfate was removed from the exhaust during this test, as well as a
significant fraction of the S02. Total particulate and calcium emissions
were also low, indicating little attrition of the sorbent. However,
pressure drop build-up in the trap was impracticably large, rising from
an initial value of about 1 000 pascals to a final value of 30 000
pascals.
* This material was found, during prior company-sponsored work, to be
very active for sulfate sorption and has since been used as the
standard against which other sorbents are evaluated.
-------
- 46 -
The benchmark material was also fabricated into rings, offering
a greater void volume than the pellets. As a result, pressure drop
through the trap increased at a slower rate than in the case of the
pellets, 500 pascals initially and 4 000 after 20 000 km. However, the
sorption efficiency for sulfate was adversely affected. Initial values
were only 80%, falling to 30% after 15 000 km. The CaCOs chips and ZnO
pellets were tested for distances of 10-17 000 km and had little or no
sulfate pick-up.
VI.1.2 Laboratory Screening
VI.1.2.1 Sorbent Activity
A number of sorbent candidates, which appeared to meet all or
most of the criteria for such materials, as described in Task I, were
tested for S02 and sulfate pick-up efficiency in a laboratory reactor
using synthetic exhaust gas.
It was found that several forms of A1203, as well as the
80% CaO/20% Si02 preparation, were much inferior to the benchmark sorbent;
the former because of poor sorption efficiency, the latter because of a
lack of physical strength. The CaO-Si02 was prepared by calcination of
a dry mix prior to pilling. Barium oxide was a complete failure, due to
hydration and subsequent dissolution of the hydroxide in its own water
of hydration. Another material giving poorer results than the benchmark
sorbent was CaC03. A number of materials were found to have activity
equivalent to the benchmark, at least for short periods of time. In ad-
dition, they showed generally lower activity for S02 removal, a desirable
feature. These sorbents included Zr02> CaS±0^, CaO, Mn02, and several
MgO-based samples. Only the calcium-based materials, however, are free
from the disadvantage of forming water soluble sulfates.
VI.1.2.2 Effect of Operating Conditions on Sorption Efficiency
In addition to the temperature used for the bulk of the sorbent
screening runs, 480°C, several runs were also made at 370°C. Magnesia,
ZnO, and the benchmark sorbent did not show a significant activity
change, but the Zr02 sample did.
Screening runs were also carried out at space velocities other
than the standard of 100,000 v/v/hr. Values of 150,000 and 50,000 were
used, and in some runs the 370°C temperature was also used. At 150,000
v/v/hr and 480°C, the benchmark sorbent showed no apparent loss of
activity, but MgO and ZnO both showed some loss due to decreased contact
time. At this higher space velocity and 370°C, the benchmark sample
still showed no loss in sorption efficiency. Magnesia showed some loss.
-------
- 47 -
Finally, at 50,000 v/v/hr and 480°C, results for benchmark
sorbent, MgO and CaO were similar to those at 100,000 v/v/hr, indicating
little or no effect of increasing residence time. One exception was
greater 502 pick-up by the CaO sample.
VI.2 Vehicle Durability Tests of Traps
VI.2.1 Experimental Procedures
VI.2.1.1 Vehicle Preparation
Vehicle tests were conducted on two 351 CID V-8 Fords, a 1973,^
and a 1974 model. They were equipped by us with two Engelhard PTX-IIBVy
Pt-Pd monolith oxidation catalysts, one on each bank of the engine in the
post-manifold position. Air pumps were also added. The traps themselves
consisted of conventional 4.25 litre canisters, designed for use with
pellet catalysts. They were emptied of their original catalyst charge
and refilled by us with sorbent particles. The traps were placed in the
exhaust system under the rear seat and the standard mufflers removed.
VI.2.1.2 Test Fuels
Essentially all of the kilometer accumulation and testing done
in this task were conducted with the 0.032 wt. % sulfur content fuel
described in Section V.2.2.
VI.2.1.3 Test Procedure
Kilometer accumulation was accomplished on automatic dynamometers,
using the AMA durability cycle. Periodically, the vehicles were emissions
tested using the Exxon Research exhaust particulate sampler described in
Appendix B. Initially, measurements were made at 64 kph cruise conditions
only, using two successive one hour intervals, but at the 17 600 km
point of the first test (benchmark pellets), a hot start 1975 FTP was
added after the two one hour cruises. Measurements were made of total
particulate, total sulfate.and total calcium (or zinc) emissions, as
well as pressure drop across the trap for each cruise mode. During the
hot start FTP runs, these three particulate measurements, as well as
gaseous emissions of CO, HC, and NOX were recorded. Fuel economy and
802 emissions were also measured during each test mode. The results for
each of the four durability runs are recorded in Appendix Tables E-18 to
E-21.
VI.2.2 85 CaO/10 S102/5 NazO (Benchmark) Pellets
The initial vehicle durability run was made using 3.2 mm
cylindrical pellets made from this material, which had been identified
during an earlier Exxon Research program as an active sorbent for sulfate.(8)
It has been designated benchmark sorbent, since other sorbent's activity
is evaluated against this material.
-------
- 48 -
VI.2.2.1 Benchmark Preparation
VI.2.2.1.1 Powder Preparation
The starting material for the benchmark pellets was made by
the following procedure, using these materials:
• Calcium nitrate Ca(N03>2'4H20, reagent
• Kieselguhr, calcined
• Sodium metasilicate Na2S103*9H20, reagent
• Sodium hydroxide, reagent
1. Add 1700 gms of calcium nitrate to 6 litres of distilled water
while stirring.
2. Slowly add 25 gms of calcined Kieselguhr.
3. Slowly add 118 gms of sodium metasilicate and stir constantly.
4. Dissolve 718 gms of sodium hydroxide in 750 ml of distilled water.
5. Add sodium hydroxide solution to 6 litre solution dropwise over
^2 hours using a dropping funnel. Stir solution constantly.
6. After sodium hydroxide has been added, continue stirring for one
additional hour.
7. Vacuum filter the mixture using two large, ^16 cm diameter, Buchner
funnels. Wash the filter cakes twice, each with 500 ml of distilled
water.
8. Place cakes in large evaporating dishes and dry overnight in an
oven @ 250°F.
9. Powder solid filter cakes to pass through 14 Tyler mesh screen.
VI. 2.2.1.2 Pellet Preparation
The screened, dried powder described in the previous section
was compacted into pellets using a Stokes compacting machine, Model 900-
512-1. This is a 16 station, rotary motion unit which allows upper and
lower compaction. Punches and dies of 3.2 mm were used.
It should be noted that all machine settings cannot be quantified,
and that the "best" settings for forming acceptable pills are obtained on
a trial and error basis. There are three variables which the operator
can control: speed of pilling, depth of die fill, and compaction pressure.
-------
- 49 -
Qualitatively, the settings which resulted in the "best" pills were:
speed of pilling at maximum (this had little or no effect on quality of
pilling), depth of die fill near maximum (this powder was quite fluffy
and could be greatly compacted), and compaction pressure set just below
the point at which the over-pressure safety releases (1/8 ton/face).
After pilling, the pills were calcined for four hours at 1200°F.
VI.2.2.2 Experimental Results
The results obtained with the experimental vehicle equipped
with a sulfate trap containing benchmark pellets are shown in Appendix
Table E-18. Prior to testing the sulfate trap, the vehicle without trap
was run for 3,218 km on 440 ppm S fuel and tested for particulate emissions.
The total particulate emission rate averaged 0.171 gm/km for two tests
while the sulfate emission rate averaged 0.0858 gm/km. The test procedure
was a one hour cruise at 64 km/h. For both runs, parallel filters
showed very good weight correspondence.
The trap was then installed and the above test repeated with
no accumulated mileage on the trap. The total particulate emissions for
duplicate runs averaged 0.012 gm/km with sulfate emissions of 0.002
gm/km. This represents a reduction of 98% in sulfate emissions. Another
duplicate set of 64 km/h tests was then made followed by different
test cycles of one hour at idle and two hours at 96 km/h. For all
test cycles, sulfate emissions were reduced by greater than 95%. During
these tests a third parallel filter was collected and analyzed for calcium.
In all cases, the calcium emission rates were very low indicating little
or no sorbent attrition. Testing of a similar oxidation catalyst vehicle
but without a CaO trap resulted in calcium emissions of 4 x 10~3 gm/km.
Similar oxidation catalyst vehicles without traps and with sulfur free
fuel produced total particulate emission rates higher than those observed
with the trap vehicle. Therefore, it appears that the sulfate trap is
removing other exhaust particulate in addition to sulfates.
After these tests, trap mileage was extended to 1,609 km by
running on the AMA durability cycle. Duplicate one hour tests at 64
km/h were made at 1609 and 1786 trap kilometers. The results show low
total particulate and sulfate emissions. At this point, the fuel was
switched from the 480 ppm sulfur fuel to the 320 ppm S fuel. Results at
3,218, 4,827, and 6,436 trap kilometers showed, in all cases, that
total particulate and sulfate emissions were very low, indicating the
trap was active and not attriting.
While CaO is a very active sorbent, its volume Increases sig-
nificantly as it sulfates. Based on crystalline densities, the complete
sulfation of CaO to CaS04 would produce a three-fold increase in volume.
While the pellets are somewhat porous, they cannot accommodate such an
-------
- 50 -
expansion internally and must expand into the void volume of the bed.
This expansion will cause the pressure drop across the bed to increase
as sulfation increases. To examine this problem, pressure drops have
been measured during the 42 km/h testing. A GM reactor filled with
pellets of the same geometry as the CaO pellets typically exhibits a
pressure drop of about 1,000 Pa (4 inches of water) at 64 km/h. The
sulfate trap at 3 218 km showed a pressure drop of 9 950 Pa (40 inches of
water) indicating expansion of the sorbent material. At 4 827 and
6 436 trap kilometers, the pressure drop remained relatively constant.
As trap testing continued at 64 km/h through 12 870 km accumu-
lation, the sulfate, total particulate and calcium emissions remained
low. Sulfate was 0.001 g/km and calcium at or near the detection limit
of 3 x 10~5 g/km. Total particulate was 0.003 g/km at 9 650 km. The
12 870 km value was inadvertently lost due to experimental error. The
pressure drop over this interval ranged between about 15 000 and 23 000
Pa (60-92 inches of water).
At 17 600 km, it was noted that the two successive 64 km/h
cruise tests had reversed the order of total particulate emissions.
Previously, the first test had shown lower levels, but at this point,
and for the remainder of the test, the first cruise began showing higher
levels. Sulfate and calcium emissions continued to remain about the
same for the two cruises, however. To examine this problem, the metals
filter used previously only for calcium analysis was analyzed for the
following metals: lead, iron, copper, aluminum, zinc, chromium, and
nickel. In all cases, the total metal emissions were higher for the
first test. The iron emissions were the largest fraction. In most cases,
the first cruise gave much higher iron values than the second. The lead
emissions are also somewhat higher during the first test. Therefore, it
appears that the metal emissions are responsible for the increased particulate
emissions during the first test cycle with the iron emissions making up
the largest fraction.
At the 17 600 km point, a hot start FTP was added to the test
sequence to assess the effects of cyclic operation on emissions. It is
seen in Appendix Table £-18 that at this and all subsequent test periods,
the cyclic values for total particulate, sulfate, and calcium emissions
were generally higher than for the cruise modes.
From 17 600 km to the final test interval at 42 400 km, the
values for these three emissions did not increase. The only possible
increase was recorded for calcium emissions at the final measurement
point. Pressure drops, however, continued to rise, reaching as high as
29,000 Pa (116 inches of water) at the end. It should be noted that the
pressure drop tended to change during the course of a cruise, gradually
falling as water of hydration was driven from the sorbent. For example,
at 42 400 km, the initial pressure drop for the first cruise was 37 500
Pa (150 inches of water), lowering to 29 000 Pa after about fifteen
minutes.
-------
- 51 -
Gaseous emissions of CO, HC, and NOX showed no trend over the
interval from 17 600 to 42 400 kilometers, taken during the hot start
FTP runs. Fuel economy did decrease with increasing trap age, however,
presumably because of the harmful effect of increasing back pressure.
Initial fuel economy for the trap car at 64 km/h was about 8.6 km/litre,
but fell to 6.0 km/litre at the end of the run. Similarly, the first
FTP, at 17 600 km, gave an average fuel economy value of 4.2 km/litre,
which fell to 3.9 at the conclusion of the run. Finally, S02 readings,
taken before and after the trap with the TECO instrument described in
Appendix B, showed approximately a 50% removal efficiency for this gas,
holding steady throughout the run. It should be pointed out that the
TECO had not been correctly calibrated as yet, so the absolute numbers
in Appendix Table E-18 are not as reliable as are the before and after
trap comparisons.
This durability test was terminated after 42 400 km, primarily
because of the high pressure drop through the trap. Examination of the
sorbent bed showed that the pellets, for the most part, had remained
intact. They were, however, adhering one to another with some fines in
the open areas. The volume expansion of the pellets was then determined
by measuring ten random pellets of the fresh sorbent and ten random
used pellets. The expansion was found to be 13% which would theoretically
increase the pressure drop by a factor of 5. The fines would also plug
the flow passages and increase this factor significantly. A number of
analyses were performed on the pellets themselves and are described in
the next section. A summation of the sulfate sorption efficiency and
pressure drop build-up of the trap is presented in Figures VI-1 and VI-2.
VI.2.2.3 Chemical Analysis Of Used Pellets
Samples of the fresh and used sorbent material were analyzed
to determine the mechanism of sulfation. A detailed report of the results
of these tests appears in Appendix Section D.I. The first analysis was a
determination of density and pore volume. The density and pore volume of
the fresh sorbent show that the initial pellets had an internal void volume
of 49%. This allowed deep penetration of the sulfur oxides into the pel-
lets and also accommodated a large part of the sulfation volume expansion
as shown by the decrease in void volume of the used pellets to 35%.
The bulk sulfate content of the fresh and used material was
determined. The fresh material was less than 0.1% sulfate. The used
material was ^0% sulfate by weight. The sulfate gradient across the
pellet was determined by selectively scraping pills at different diameters
and analyzing the scrapings. The gradient was also determined by scanning
electron microscopy - X-ray energy spectrum. The X-ray energy spectrum
showed the following gradient:
-------
FI3URE VI - 1
TRAP EFFICIENCY AT 64 KM/H-PELLETS
100
80
60
UJ
o
LL
U.
UJ
Q_
Qi
20
0
10
I
20 30
KILOMETERS x 10'3
40
50
Ul
to
-------
FIGURE VI - 2
PRESSURE DROP AT 64 KM/H-PELLETS
CO
O
<
Q.
Q.
O
LU
o:
ai
a:
a.
50
40
on
30
20
10
TYPICAL MUFFLER PRESSURE DROP
- AT 64 KM/H = IxlO3
Oi
OJ
10 20 30
KILOMETERS x 10'3
40
50
-------
- 54 -
Table VI-1
Radial Distribution of Sulfur and Calcium in Used Trap Pellets
Sulfur Calcium
Radial Position Relative Peak Height Relative Peak Height
1.0 (edge) 53 50
0.76 56 50
0.65 54 49
0.50 36 77
0 (center) 35 70
The relative peak heights are very nearly proportional to
concentration. Both methods show the same gradient within experimental
accuracy. These results are very important since they show that sulfation
is taking place throughout the pellets.
Fresh and used samples were analyzed by X-ray diffraction to
determine the crystalline compounds present. The fresh sample showed
calcium oxide, calcium silicate hydrate (Ca2 Si04*H20) and sodium carbonate
hydrate (Na2 (X>3'10H20). The calcium silicate is due to reaction of the
calcium oxide and silicon oxide during calcining. The sodium carbonate
hydrate is due to the reaction of the sodium oxide with carbon dioxide
either during calcining or in the interim between calcination and analysis
followed by hydration.
The used samples showed a very large amount of calcium sulfate,
some calcium oxide, calcium silicate hydrate, and sodium carbonate
hydrate. These compounds would be expected given the compounds found in
the fresh material with the exception of possibly the sodium carbonate.
Thermodynamics favors the conversion of the carbonate to the sulfate.
Due to the initial low concentration of sodium oxide, 5%, and the large
amount of calcium sulfate, it would be very difficult to see sodium
sulfate in the X-ray diffraction patterns. However, sodium carbonate
was identified, Indicating, at most, only partial formation of sodium
sulfate.
-------
- 55 -
VI.2.3 CaC03 Chips
Although the durability test of Benchmark pellets showed favorable
sulfate pick-up efficiency and capacity characteristics, the build-up of
pressure drop was unfavorable. One possible method of overcoming this is
through the use of a sorbent which expands less upon sulfation. Calcium
carbonate increases in volume by only a factor of 1.4 upon sulfation, com-
pared to an almost three-fold increase in the case of CaO. A price must be
paid, of course, for this improvement in volume expansion. Because the
starting molecule is now larger, the sorbent will have less capacity for
sulfate sorption per unit of volume or weight than would CaO. Since our
interest was basically in minimizing pressure drop, the same type of vehicle
test described in Section VI.2.2 was repeated anyhow, this time using pure
CaC03 as the sorbent in place of the benchmark material, which was
primarily CaO.
VI.2.3.1 Test Conditions
Because of the hardness of CaC03, the sorbent particles were
not pelleted from powder. Rather, marble chips were granulated and the
4/7 mesh fraction was used to simulate as closely as possible the 3.2 mm
pellets used previously. The same car was used, but fresh monolith
catalysts were mounted to provide maximum levels of sulfate in the
exhaust. Surprisingly, however, the new set of catalysts gave only
about one half the level of sulfate emissions as had the original catalysts
when they were fresh (corrected to the 0.032 wt. % sulfur level fuel).
Because of this discrepancy, and also because of the now realized effect
of catalyst aging on sulfate production, it was decided to measure the
sulfate emissions entering the trap, and compare these with the after
trap filter results in order to calculate removal efficiency. The
Goks^yr-Ross technique, described in Appendix B, was used to follow the
before trap sulfate level of the exhaust gas. Because of its lack of
constant volume sampling, it can be used only for the steady state
cruise modes.
VI.2.3.2 Experimental Results
As shown in Appendix Table E-19, the sulfate sorption efficiency
of this trap was quite poor. If the filter values of sulfate after the
trap are compared with the baseline emissions, only about 16-55% pick-up
is calculated, which varied randomly over the 17 400 km duration of the
run. If the actual pre-trap sulfate values, obtained from the raw
exhaust during each 64 km/h test by the Goksrfyr-Ross method are used as
the basis for comparison, essentially no sulfate sorption appears to be
occurring. For reasons to be discussed in Appendix C, the Gokstfyr-
Ross method may tend to give low values. Therefore, we will assume the
post-trap filter samples, compared to the base case sulfate emissions,
-------
- 56 -
as providing the correct value for sulfate sorption efficiency. In
either case, though, the results are poor.
Over the course of the run, no trend was observed in total
particulate emissions during cruise modes. The last FTP run did give
the highest value, however. Calcium emissions themselves were at or
near the detection limit for most of the tests, indicating little or no
sorbent attrition.
The pressure drop showed no upward trend during this run,
indicating that in the absence of significant sulfation, which would
result in sorbent swelling, there is little or no tendency for pressure
drop to increase.
Contrary to the benchmark pellets, the CaCOs chips did not
cause a drop in vehicle fuel economy with aging, measured both at cruise
and on the FTP. Again, this is undoubtedly because of the lack of
pressure drop build-up. Actually, there was an improvement with time of
fuel economy, but this was probably due to other vehicle changes unrelated
to the trap. Gaseous emissions showed no significant trend with time.
No measurements were made of S02 sorption, but in view of the low level
of activity for sulfate, this would not be expected to be large.
Because of the disappointing sulfate sorption performance, the
test was discontinued at the 17 400 km mark. It is not known if the
poor reactivity is inherent in CaC03, or was due to the very dense, non-
porous nature of the chips used. As will be seen in Section VI.4.2.2,
similar chips did show some activity in laboratory screening runs.
VI.2.4 ZnO Pellets
The next durability test was conducted with the same vehicle
used for the CaCC>3 chips, with fresh catalysts installed. The trap was
filled with 5 mm extrudates of ZnO, a catalyst material manufactured by
the Harshaw Chemical Company.
VI.2.4.1 Experimental Results
The car was first baselined, with the trap empty of sorbent.
These results are shown in Appendix Table E-20. At 64 km/h cruise, the
total sulfate emissions averaged 0.031 g/km, and total particulate 0.078
g/km. The subsequent cruise tests, after aging periods of 5 300, 9 700,
and 15 900 km, averaged about a 30% removal of sulfate. Removal rates
for total particulate averaged just over 50%. During the hot start FTP
runs, however, no sulfate and little total particulate removal was
found.
-------
- 57 -
Readings of the S02 concentrations in the raw exhaust, before
and after the trap, showed that no S(>2 sorption was occurring, not
surprising in view of the poor sulfate activity. It was also found that
the level of zinc particulate emissions was significantly higher than
that observed with any of the base case tests or tests with other sorbents.
This indicates a high level of attrition for the ZnO sorbent. The
pressure drop was lower initially than for either the benchmark pellets
or CaCC>3 chips, which may be a reflection of the larger size of the ZnO
starting material. The failure of the pressure drop to increase with
time is again probably due to the lack of sulfation and consequent
swelling.
VI.2.5 85 CaO/10 SJ02/5 Na20 Rings
As was discussed in Section VI.2.2, the benchmark pellets were
very effective for sulfate sorption. However, they also swelled during
sulfation which caused an unacceptably large increase in pressure drop.
One possible method of minimizing this effect is through the use of
particle shapes which offer inherently lower pressure drops than do
pellets. Rings were selected for initial testing of this concept, using
the benchmark powder as the starting material.
VI.2.5.1 Test Conditions
The benchmark powder was fabricated into rings of 1.58 cm O.D.
x 0.48 cm I.D. and 0.79 cm height by Girdler Chemical, Inc. These
dimensions were not the optimum choice for compromising pressure drop
and surface area/volume parameters, but happened to be the size of dies
available to Girdler at the time of our fabrication request.
A new vehicle was used for this run, but like the car used for
the three previous durability runs, it was a full size, 351 CID -^
V-8, with air pump and was equipped by us with two Engelhard PTX-IIB ®
monolith catalysts, one on each bank of the engine. As before, a 4.25
litre catalyst canister was emptied of its original charge, refilled with
sorbent and placed in the exhaust system under the rear seat.
VI.2.5.2 Experimental Results
As shown in Appendix Table E-21, the benchmark rings werp
somewhat poorer than the benchmark pellets, Section VI.2.2, for sorption of
sulfate. For the 64 km/h cruise tests, the absolute levels of sulfate
emissions, over the first 7 660 km, averaged about 0.005 g/km, compared
to 0.001 g/km for the pellets. On the basis of the base case emissions
of 0.033 g/km, this equalled about 85% efficiency. However, with increasing
kilometer accumulation, sorption efficiency began falling off sharply, as
shown in Figure VI-3, with sulfate emissions rising to an average of 0.014
g/km at the end of the run at 20 400 km. Under the hot start FTP test con-
-------
- 58 -
dition, sulfate emissions did not show such a trend, and averaged about
0.007 g/km or 72% removal compared to the base case. Total particulate
emissions for both types of tests followed the same general trend as did
sulfate emissions. Calcium emissions were at or near the detection limit
for the entire run.
Surprisingly, even during the early part of the run, when
fairly high sulfate pick-up rates occurred, there was no pick-up of S02,
as evidenced by comparing TECO readings of the undiluted exhaust before
and after the trap. This indicates that the inherently good activity of
the benchmark material for S02, at least in the pelleted form, is negated
by the poorer gas-solid contacting efficiency of a ring bed. Since
sulfate is a more active species than is S02, it is less affected by the
drop in contacting efficiency. Directionally, this widening of the
pick-up efficiency gap between sulfate and S02 is desirable, since S02
sorption is unnecessary and merely increases the rate of sulfation of
the sorbent. In turn, this decreases its effective capacity for sulfate
and increases the rate of swelling.
The trap pressure drop, as expected, was lower than for the
pellets initially and also rose at a slower rate, as shown in Figure VI-4.
Thus, at 3 000 km, the pressure drop was only about 600 Pa (2.4 inches of
water) at 64 km/h cruise, and rose only to 4 200 Pa (M.7 inches of water)
at 20 400 km. In contrast, the pellet bed had reached about 20 000 Pa
(80 inches of water) by this time.
Gaseous emissions of CO and HC showed a sharp rise toward the
end of the run, but it is not clear if this was due to a problem with
the carburetor or the catalysts. If enrichment of the carburet ion
occurred, it was not reflected in the fuel economy, which remained
constant over the entire run. On the other hand, a loss in catalyst
activity for CO and HC should have been accompanied by an even greater
loss in S02 oxidation activity, which, as was shown under Task III of
this contract, is even more sensitive to catalyst activity. The evidence
here is ambiguous. At 14 400 km, when the greatest CO value was recorded,
there was no drop in sulfate emissions on the filter, nor even in pre-
trap sulfate values measured by the Goks^yr-Ross method. At 20 400 km,
where the highest HC value occurred, the Goksrfyr-Ross measurement did
show a drop in pre-trap exhaust sulfate concentration.
VI.3 Status of Sulfate Traps
Although good 504° trapping efficiency and capacity, and
relatively low pressure drop have been demonstrated, it has not been
possible yet to combine all of these necessary features into one structure.
Pellets of the CaO/Si02/Na20 composition have met the efficiency and
capacity criteria, but not the pressure drop requirements. Several
-------
FIGURE VI - 3
TRAP EFFICIENCY AT 64 KM/H-RINGS
O
100
80
60
UJ
o 40
UJ
< 20
QC
Ln
vo
I
I
10 15
KILOMETERS x 10'3
20
25
-------
FIGURE VI -
PRESSURE DROP AT 64 KM/H-RINGS
CO
o
rH
X 4
CO
_l
<
O
Q.
o 2
CO
CO
Ul
0
o
I
TYPICAL MUFFLER
10 15
KILOMETERS x 10'3
20
25
-------
- 61 -
areas of improvement with pellets should be available, however. If the
composition is changed to lower the reactivity towards 502, then the
rate of sulfation will be limited to that due to S04= pick-up only.
This will result in a slower build-up of pressure drop and increased
capacity for S04=. Laboratory tests have indicated that the omission of
Na20 can reduce S02 pick-up, and this lead will be followed up with
vehicle tests. In addition, to the extent accumulation of fine particles
in the bed due to powdering of the pellets has increased the pressure
drop, this can be minimized by learning how to make harder pellets.
The ring test showed that pressure drop can be lowered sig-
nificantly, but initial efficiency and also capacity suffered. However,
the first ring test reported here was carried out with a particle geometry
chosen on the basis of die availability, rather than optimization of
such factors as flow characteristics and surface to volume ratio.
Therefore, further work with rings and other shapes, such as saddles,
should result in further improvements. Improved canister geometry may
also prove useful.
VI.4 Laboratory Screening of Sulfate Sorbents
VI.4.1 Experimental Procedure
Laboratory testing of sulfate sorbent candidates was carried out
in a bench top reactor, illustrated in Figure VI-5. The sorbent, generally
in the form of granules or approximately 3 mm pellets, was contained in a
Vycor reactor, about 25 mm in diameter, mounted vertically in a tube furnace.
Test temperatures were mostly 480°C, although some tests were run at 370°C.
The constituents of the synthetic exhaust gas were blended through a bank
of rotometers, to give the composition shown in Appendix Table E-22. The
only exceptions were H20 and 803 vapor, which were added by passing the
heated gas through a 0.005 normal solution of H2S04. No H2, CO or HC was
added to the gas stream, even though provision was made for them, since
these components would be present in very small quantities in a real,
post-catalyst exhaust gas, and should have no effect on sorbent activity.
Thermocouples were mounted adjacent to the sorbent bed, whose volume was
about 13 cm3, to allow control and recording of the bed temperature.
Analysis of the gas stream for S02 and S03 was made, before or
after the sorbent bed, by means of a TECO UV fluorescence instrument and
the Goksrfyr-Ross coil. Both methods are described in Appendix B.
VI.4.2 Experimental Results
In addition to the benchmark material, a number of other possible
sorbent materials have been screened. The primary purpose of this search
was to find sorbents as active as the benchmark for sulfate pick-up, but
-------
FIGURE VI - 5
APPARATUS FOR LABORATORY SCREENING OF SORBENTS
TC
TRC
V
AIR N2 / H2 502
CO
C3
C02
FEED BLENDING
WATER_
VAPOR"
S03
"•8AJULS1
|l
SORBENT
BED
. VYCOR
•Z/REACTOR
FURNACE
BYPASS
HEATED LINE
H2S04
(GOKS0YR-
ROSS COIL)
ANALYZERS
to
I
(THE BYPASS is LOCATED AFTER THE WATER VAPOR + $03 ARE ADDED,)
-------
- 63 -
less active towards S02> Samples based on Ca, Al, Ba, Zr, Mg, Zn, and
Mn have been studied. In some cases, commercial samples have been
available in the proper size range and were tested as received. In
other cases, the sorbents were prepared in house. In some cases, this
required preparation of the powders, followed by pelleting, or if not
suitable for pelleting, granulation and sieving. In others, suitable
compounds were available for pelleting, if possible, or granulation and
sieving if not. Appendix Table E-22 shows all of the materials tested,
in what form and weight, for how long, and their activity towards sulfate
and S02.
VI.4.2.1 Benchmark Sorbent
All of these experimental sorbents were compared to the bench-
mark material. The latter was tested at intervals in the laboratory
reactor to insure that the equipment was functioning properly. As shown
in Appendix Table E-22, all tests with this material showed complete or
nearly complete sulfate pick-up, even out to 8.4 hours of exposure and
a lower temperature of 370°C. Sorption of S02 was as high as 93% after
1-2 hours exposure, but fell to 79% after about 4 hours, at 370°C, and
35% after eight hours, at 370°C. The pellets used for these tests were
made by the procedures described in Sections VI.2.2.1.1 and VI.2.2.1.2.
VI.4.2.2 Other Calcium Sorbents
Other calcium-based sorbents tested in this program included
an 80 CaO/20 Si02 material, prepared in an analogous manner to the
benchmark; a commercially available calcium silicate powder known as
Micro-Gel, used for removing acidic components from organic liquids;
commercial calcium carbonate chips, identical to those used in the
vehicle durability test described in Section VI.2.3; and pellets formed
from Ca(OH)2» using an aluminum stearate binder, which dehydrated to CaO
during calcination of the pellets. Except for the CaCOs chips, all of
these calcium-based sorbents initially showed 100% efficiency for sulfate
removal. The CaCOa had only about 75% efficiency after 1.3 hours, which
fell to 50% after 3.7 hours. Of the other materials, the 80 CaO/20 S102
and Micro-Cel pellets were tested for only 2.5 and 3.5 hours respectively,
showing little or no activity loss. The CaO-blnder sample, tested only
at 370°C, fell from 100% removal at 3.2 hours to only 67% at 4.6 hours.
The CaO-binder and CaC03 samples, therefore, have less capacity than the
benchmark, but no conclusions can be drawn about the 80 CaO/20 Si02 and
Micro-Cel materials. The S02 pick-up efficiency of all four of these
calcium-based sorbents was much lower than for the benchmark, with the
best, the CaO-binder combination, showing only 33% at 1 hour and falling
to 12% after 4.6 hours.
The CaC03 material was tested in the form of chips, because
CaCOS powder was too hard to form into pellets. The other three materials
-------
- 64 -
could be pelleted, but the resulting particles, while strong enough for
laboratory reactor screening, were much weaker than the benchmark and
would not have been suitable for vehicle testing. Thus, although good
sulfate pick-up and proper selectivity towards S02 has been shown by
calcium-based materials, none, except for the benchmark, are yet suitable
for vehicle testing.
Several interesting points are worth further discussion in
connection with calcium systems. First, the physical properties of the
sorbent seem to affect the sorption efficiency. Of the four systems
just described, the hardest, CaC03, was the least active. It also had
the least surface roughness. Chemical composition would not be expected
to play a role (H2S04 is a strong acid and should react equally well
with CaO, Ca(OH)2 and CaCOs). It should also be noted that although
most of the calcium sorbents retained good sulfate pick-up efficiency,
none were very active towards S02, except for the benchmark, which was
the only one to also contain Na2<3. It would appear then that in the
absence of highly basic sodium, good selectivity between sulfate and SO2
pick-up can be achieved. This effect will also be seen with the magnesium
sorbents described in the next section.
VI.4.2.3 Magnesium Sorbents
Four versions of magnesium-based sorbents were tested in the
laboratory. The first was pure Mg(OH)2, which was pelleted and then
calcined. The resulting pellets were soft, but strong enough for laboratory
evaluation. The second material consisted of spheres of magnesite, a
naturally occurring HgC03. This was calcined to MgO, retaining its
physical strength in the process. Third, commercially available pellets
of MgO (Harshaw Chemical Company), were tested as received. Finally, a
magnesium analog to the calcium benchmark material described in Section
VI.4.2.1 was prepared in a similar manner.
All four of these magnesium-based sorbents initially showed
complete removal of sulfate. Only the one containing Na20 picked up any
SOg, again confirming the effect of basic sodium compounds. The Harshaw
MgO pellets began to lose activity between the first and second hours,
falling to about 75% at 2.7 hours and 17% at 4.2 hours. The other three
did not show any loss, over total test periods ranging from 2.2 to 7.4
hours. The calcined magnesite maintained this activity even when the
temperature was lowered to 370°C.
VI.4.2.4 Alumina Sorbents
Three different alumina samples were also tested. Two commercially
available pellets, from Norton and Matheson, Coleman and Bell, were run.
The latter was a high surface area, activated form. In addition, a
-------
- 65 -
section of honeycomb substrate, with a high surface area alumina washcoat,
was also run. All three showed very poor activity for sulfate pick-up,
and essentially none for SO-.
VI.A.2.5 Other Sorbents
Several other sorbent materials were also tested. These were
the oxides of Ba, Zr, Zn and Mn. Granules of BaO were a total failure,
hydrating rapidly to Ba(OH>2 and dissolving. Pellets of Zr02 were
initially active for sulfate, but showed lower capacity than the bench-
mark. Starting from 100% efficiency after 1.3 hours only 84%
pick-up was obtained at 4 hours. The temperature was lowered to 370°C,
and activity at 5.1 hours was 73% and 12% at 8.8 hours. No S02 pick-up
was observed. Zinc oxide extrudates showed good activity for 4.5 hours,
initially at 480 *C, dropped to 82% at 3.9 hours, recovered to 100% at
4.5 hours and 370°C, but fell to 58% at 8.7 hours and 370°C. Little, if
any. S02 sorption was found. Finally, small Mn02 granules were completely
effective for sulfate removal for 2.7 hours, but fell to 82% at 3.9
hours. A relatively large amount of SC>2 was also picked up, 55% at 1.2
hours and 21% at 3.9 hours.
VI.4.3 Conclusions From Laboratory Screening Program
In addition to the previously identified benchmark sorbent, a
number of other materials have been identified which show laboratory
effectiveness for sulfate removal. Of these, some also offer an advantage
over the benchmark sorbent in terms of a lower reactivity towards S02.
These Include the pelleted Ca-contalning sorbents, the two calcined MgO
samples, Zr02, ZnO, and Mn02. Of these, only the Ca-based sorbents
satisfy the criterion of having a water insoluble sulfate. Although a
vehicle test would be necessary to determine with certainty whether or
not a water soluble sulfate can be tolerated in a trap, we believe the
next phase of testing should be restricted to calcium sorbents.
The prime difficulty encountered with the promising sodium-
free calcium sorbents is an inability to form pellets strong enough to
withstand vehicle testing. The ability of the benchmark material to be
so fabricated indicates that Na20 is acting as a binding agent. The
problem then is to find a combination of other binding agents and preparation
techniques which will allow the fabrication of strong pellets. In turn,
this would allow vehicle testing of calcium sorbents which offer the
promise of good selectivity between sulfate and S02-
VI.5 Effect of Space Velocity
In addition to the screening runs conducted at a space velocity
of 100 000 v/v/hr, described in Section VI.4, several of the more active
-------
- 66 -
materials were also run at 150 000 v/v/hr, all at 480 °C. This was done
to provide some feel for the possibility of using smaller vehicle reactors,
should underfloor space considerations require such a step. Samples of
the benchmark, MgO (calcined magnesite) and ZnO which had been previously
tested at 100 000 v/v/hr were kept on stream at the higher gas flow
rate. All other conditions of gas composition and bed volume were
maintained constant. A fresh charge of benchmark sorbent was also
tested.
As shown in Appendix Table E-23, the previously used benchmark
sorbent, which had seen 8.4 hours at the lower space velocity, maintained
its high efficiency for sulfate trapping even after an additional 6.1
hours at the higher space velocity. At the end of the total of 14.5
hours, sulfate pick-up was still 100% and S02 pick-up 20%. The MgO
sample, which had been tested for 7.4 hours at 100 000 v/v/hr, showed a
sharp fall in sulfate efficiency after an additional 1.3 hours at the
higher space velocity and was down to 40% after a total of 11.6 hours.
Surprisingly, a small amount of S02 pick-up was seen at the higher space
velocity, where none had been noted previously. The ZnO sample showed
an anomalous result. After being exposed to feed gas for 8.2 hours at
100 000 v/v/hr, it showed only 50% sulfate sorption after about one hour
at the higher space velocity. However, 2.5 hours later, still at
150 000 v/v/hr, sorption efficiency rose back to 100%. We have no
explanation for this result. Finally, a fresh charge of benchmark
sorbent was run for 4.2 hours at 150 000 v/v/hr, and showed an initial
efficiency for sulfate of 100%, and 97% at the end. Pick-up of S02 was
in the 70-80% range initially, and about 50% at 4.2 hours. The sulfate
values are comparable to those at 100 000 v/v/hr, but the S02 pick-up
rates are considerably lower.
Several sorbents were also tested at 50 000 v/v/hr. These
were the benchmark, calcined magnesite and CaO. The benchmark material,
after 4.2 hours at 150 000 v/v/hr. was exposed for an additional 10
hours at 50 000 v/v/hr. It showed complete sulfate pick-up over this
interval, and an S02 sorption rate somewhat better than that seen at the
highest space velocity. The calcined magnesite, after 11.6 hours at the
two higher space velocities, was run for an additional 4.4 hours at
50 000 v/v/hr. Improved sulfate sorption was obtained. For example, at
the 11.6 hour mark and 150 000 v/v/hr, only 40% trapping efficiency was
noted, but at 16 hours and the lowest space velocity, this had risen to
64%. No improvement in S02 sorption with decreased space velocity was
noted. The CaO sample had previously been run only at 370°C and 100 000
v/v/hr, falling to 67% sulfate efficiency at 4.6 hours. At the more
favorable 480°C and 50 000 v/v/hr, however, 88% efficiency was obtained
even after 8.9 hours total exposure. A higher level of S02 pick-up was
also found.
-------
- 67 -
Finally, fresh samples of the benchmark, calcined magnesite
and MnC>2 sorbents were tested under the least favorable combination of
conditions, high space velocity, 150 000, and low temperature, 370°C.
As seen in Appendix Table E-23, all showed good initial activity for
sulfate sorption, but lost this more rapidly than at higher temperatures
and/or lower space velocities. The benchmark and Mn02 samples also
showed reduced 862 pick-up activity. The calcined magnesite exhibited
the same zero efficiency for SO- as at the other operating conditions.
VI.6 Activated Charcoal for Sulfate Removal from Exhaust Gas
During the course of this contract, discussions were held with
the Westvaco Corporation to determine if activated charcoal would be
useful in a sulfate trap. The primary mechanism by which sulfate would
be removed was expected to be reduction of 863 to 862 by reaction with
carbon. A secondary mode could be the sorption of sulfate by the carbon.
As a result of these discussions, a brief experimental program was con-
ducted by Westvaco at their Charleston Research Center, North Charleston,
S.C. Their report of this work is presented in full in Appendix Section D.2.
Briefly, Westvaco tested, in a laboratory reactor, the ability
of activated carbon granules to remove sulfate from various simulated
exhaust gas mixtures as a function of temperature, space velocity and
granule size. It was found that at temperatures of about 315°C, almost
complete conversion of 80^ to 802 occurred using a fine (12 x 30 mesh)
carbon bed and a space velocity as high as 60 000 V/V/ hr. However, con-
version fell off rapidly as granule size increased or temperature was
lowered. At the lower temperature the non-converted sulfate appeared
to be sorbed in the carbon, and could be released as S02 upon heating
to 315°C. Sorption efficiency also was adversely affected by increased
mesh size and space velocity.
In addition to reactivity towards SOo, the activated carbon
could be expected to react with other oxidants in the exhaust, such as
02, H20, and NO . A series of runs was also made therefore to determine
the rates of these undesirable side reactions. It was found that, in
addition to SO-j, 02 and H^O do react with the carbon to a significant
extent at the 315°C temperature necessary to achieve good sulfate con-
version. Various combinations of the three gaseous reactants, 02, 1^0
and 803 showed carbon burn-off rates 2-5 times greater than that due to
803 alone. Nitric oxide did not appear to react. At 370°C the burn-off
rate increased by a factor of 7-10. Several runs were made with a
carbon specially modified to minimize burn-off, and it was found to cut
burn-off rate in half without apparently harming the 803 reaction rate.
As a result of this work, it has been concluded that it would
be extremely difficult, if not impossible, to utilize activated carbon in
a practical sulfate trap. The principal drawback is the rapid reaction
which will occur at the usual vehicle-mounted trap temperatures, around
370°C, between oxidants such as 02 and 1^0 and the carbon. This burn-off
-------
- 68 -
will result in rapid depletion of the bed material, as well as possibly
leading to CO formation and a fire hazard. Although there may be a
narrow temperature region in which acceptable rates of sulfate reaction
and burn-off can be compromised, a complex temperature control system
would be necessary to maintain it over the full range of driving con-
ditions .
-------
- 69 -
VII. References
1. "Motor Gasolines, Summer, 1974," Petroleum Products Survey No. 88,
Bureau of Mines, U. S. Department of the Interior, Jan. 1975.
2. M. Beltzer, R. J. Campion, and W. L. Petersen, "Measurement of
Vehicle Particulate Emissions," SAE Paper 740286, February, 1974.
3. W. R. Pierson, &± al_., "Sulfuric Acid Aerosol Emissions From Catalyst
Equipped Engines", SAE Paper 740287, February, 1974.
4. W. R. Leppard, Interim Report on Sulfate Control Technology Assessment,
Contract No. 68-03-0497, November 13, 1974.
5. W. B. Jacobs, M. M. Braverman and S. Hochheiser, Anal. Chem.,
29:1349 (1957).
6. M. Beltzer, e_t^ aJL., "The Conversion of SO? Over Automotive Oxidation
Catalysts" SAE Paper 750095, February, 1975.
7. B. J. Kraus, et^ al^. "Critical Factors Affecting Sulfate Emissions"
SAE Paper to be presented February, 1976.
8. Letter to Senator Jennings Randolph from D. G. Levine, Exxon Research
and Engineering Company, November 16, 1973.
-------
- 70 -
Appendix A
Task I - Literature Search
A.I Introduction
This literature search was conducted to bring together and
examine the literature pertaining to the fate of sulfur oxides in
automotive exhaust systems. Currently available gasoline averages about
300 ppm sulfur in the form of organic sulfur compounds. During the
combustion process, gasoline sulfur is oxidized to sulfur dioxide. In
non-catalyst vehicles, this sulfur dioxide is emitted to the atmosphere.
In vehicles equipped with oxidation catalysts for control of carbon
monoxide and hydrocarbon emissions, further oxidation to sulfur trioxide , SO.,
takes place. This SO^ can then combine with water, forming sulfuric acid.
The literature was reviewed to investigate the thermodynamic potential
and kinetics of forming the trioxide and to examine the fate of both
oxides in the exhaust system. Since sulfuric acid emissions may be
deleterious, the literature pertaining to removal of sulfur oxides from
gaseous streams was reviewed. Stress was placed on the use of metal-
oxide sorbents for this purpose.
To cover these subjects, the body of this report is divided
into four sections. The first section details the thermodynamics of
sulfur trioxide formation, reaction with water, and condensation. The
second section examines possible reaction with materials in the exhaust
gas or system. The third system reviews the catalytic oxidation of
sulfur dioxide on platinum catalysts. The last section examines possible
means of removing sulfur trioxide from the exhaust stream.
A.2 Summary and Conclusions
A.2.1 Thermodynamics of Sulfuric Acid Production
1. At typical oxidation catalyst temperatures, conversions of
sulfur dioxide to sulfur trioxide greater than 50% are
thermodynamically possible.
2. The equilibrium conversion is strongly dependent upon oxygen
concentration. At temperatures above 400°C, decreasing the
oxygen concentration decreases the equilibrium conversion,
suggesting a possible control strategy.
3. Thermodynamics and kinetics show that exhaust sulfur trioxide
may hydrate to gaseous sulfuric acid within the vehicle's
exhaust system, depending upon driving mode.
4. Thermodynamics show that the gaseous sulfuric acid will begin
to condense at about 150°C which is below the temperature
at the tailpipe exit for all driving modes except startup.
-------
- 71 -
A.2.2 Reaction of Sulfur Dioxide and Trioxide with Exhaust Gas
Constituents and Exhaust System Components
1. Thermodynamics shows that ammonia will reduce sulfur trioxide
to the dioxide. However, exhaust ammonia will be oxidized
over the oxidation catalyst before reaction can take place.
2. The formation of ammonium sulfate is favorable only below
225°C.
3. Thermodynamics shows that the reduction of both sulfur oxides
by carbon monoxide is favorable.
4. Reaction of both oxides with the iron oxide surfaces of the
exhaust system is favorable below 425°C.
5. Reaction of sulfur trioxide with the aluminum oxide catalyst
substrate is possible below 425°C. The presence of carbon
monoxide may lower this temperature by about 50°C.
A.2.3 Automotive Catalysis of Sulfur Dioxide
1. The rate limiting step in the catalytic oxidation of sulfur
dioxide is the surface reaction between adsorbed oxygen and
adsorbed sulfur dioxide.
2. The following rate equation appears to represent best the
available experimental data for industrial catalysis and
should be valid for automotive catalysis
1/2
kl (pS02 P02 - PS03 >
rate
1/21/2
(1 + Ko2 Po2
Where k^ is the rate constant, subscripted P's are the
partial pressures of the compounds in the subscripts,
Ke is the equilibrium constant of the oxidation reaction,
and the subscripted K's are the absorption equilibrium
constants for the compounds in the subscripts.
This equation is in accord with the above rate limiting
mechanism. This equation also indicates a possible
control strategy of limiting the amount of oxygen over
the catalyst.
-------
- 72 -
3. The automotive catalysis literature is limited. In addition,
the data are confounded by many experimental problems,
notably the storage/release phenomenon. In general, this
literature says that more sulfur trioxide is formed over
catalysts than with non-catalyst vehicles but good
quantitative data are lacking.
A.2.4 Sulfate Traps
1. The most promising means of removing sulfur trioxide from the
exhaust stream is to react it with a basic metal oxide.
2. Based on a selection criterion consisting of seven requirements,
the most promising sorbent material is calcium oxide. Other
less promising but still attractive sorbents are the oxides
of magnesium, manganese, and aluminum.
A. 3 Thermodynamics of Automotive Sulfuric Acid Production
In assessing the production and fate of sulfuric acid within a
vehicle's engine or exhaust system, there are three basic reactions
which must be considered. These reactions are: the reaction of sulfur
dioxide with oxygen to form sulfur trioxide, the hydration of the sulfur
trioxide to gaseous sulfuric acid, and the condensation of the gaseous
acid to produce a finely dispersed aerosol. The first reaction is
important as it shows that the formation of substantial amounts of
sulfur trioxide is thermodynamically possible at typical automotive
catalyst temperatures. The second and third reactions are important
since they define the state, sulfur trioxide, gaseous sulfuric acid, or
condensed sulfuric acid aerosol, in which the sulfur trioxide will exist
in the exhaust system. This is important from the point of view of
controlling the potential emissions within the exhaust system. Hydration
of the trioxide is possible during some driving modes. Condensation,
however, is unfavorable except possible at start-up and idle.
A.3.1 Thermodynamics of Sulfur Trioxide Production
The reaction of sulfur dioxide with oxygen can proceed either
homogeneously or catalytically. The homogeneous reaction is extremely
slow (14) and would be negligible considering the very small residence
times within the vehicle, typically on the order of a few seconds. In
the presence of oxides of nitrogen, the rate of sulfur dioxide oxidation
is increased markedly. A calculation using the rate-constant data from
Duecker and West (14) shows that under high NO conditions, such as high
speed or load, the conversion of sulfur dioxide to the trioxide would be
M.,0%. Since the reaction rate depends on the square of the NO partial
pressure, the conversion would be much less under low NO conditions.
The heterogeneous catalysis of this reaction will be discussed later.
-------
- 73 -
No matter how rapid the reaction rate, the maximum conversion
will be limited by thermodynamic equilibrium. Table A-l (48) lists the
free energies and equilibrium constants for the reaction as written
S02(g) + 1/2 02(g) ^=^ S03(g). (!)
The equilibrium constant for this reaction is
Ke. /-l!g!4—^- ,
(2)
where P indicates component partial pressure in atmospheres. Figure A-l
shows the effect of temperature on the conversion for typical exhaust
gas compositions. Lines of different oxygen concentration are shown to
illustrate the effect of differing air injection rates upstream of an
automotive oxidation catalyst. As this figure illustrates, increasing
the oxygen content of the exhaust increases the maximum possible conversion
to sulfur trioxlde. In the region of typical oxidation catalyst temperatures,
500 to 650°C, the maximum possible conversion is between 90 and 50% when
using an air pump supplying 5% oxygen. Assuming that equilibrium controls
the oxidation of sulfur dioxide, operation of the vehicle at stoichiometric
or slightly lean air-to-fuel ratios, either by careful carburetion or by
using a three-way catalyst system, the production of sulfur trioxide
TABLE A-l
Equilibrium Constants for the Oxidation
of Sulfur Dioxide to Sulfur Trioxide
Temperature
(°K)
300
350
400
450
500
550
600
650
700
750
800
850
900
950
1000
1050
1100
Free Energy
of Reaction
(K cal/mole)
-16906.50
-15775.99
-14649.01
-13520.27
-12389.19
-11257.38
-10127.01
- 8999.82
- 7876.73
- 6757.71
- 5642.04
- 4528.59
- 3416.23
- 2304.12
- 1192.02
80.38
+ 1029.68
Equilibrium
Constant
(atm -1/2)
2.069 x IQl2
7.088 x 10*
1.008 x 10°
3.682 x 10°
2.600 x 10^
2.972 x 10*
4.892 x 10;J
1.061 x 10;
2.878 x 102
9.313 x 101
3.477 x 101
1.460 x 101
6.754
3.389
1.822
1.039
6.244 x ID'1
-------
Figure A-l
Equilibrium Conversion
S02(g)+l/2 02(g) = S03(g)
§
M
co
O
o
§
H
H
CJ
CO
O
CO
O
H
1.0
0.8
0.6
0.4
0.2
0.0
0.2% 0
300
400
500
600
700
800
Temp., °C.
-------
- 75 -
could be reduced. For instance, operation at 0.2% oxygen Instead of 5%
could reduce the conversion by 25 to 65% over the typical oxidation-
catalyst temperature range. However, at the lower temperatures, absolute
conversions could still be as high as 70% at the 0.2% level. Therefore,
from an equilibrium standpoint, reduction of the oxygen content of the
exhaust could result in lowering sulfur trioxide production. However,
absolute control would not be feasible. The kinetics of this reaction
will be discussed in the Automotive Catalysis section.
A.3.2 Thermodynfuntcs of Sulfur Trioxide Hydration
As the S03 produced by the engine or over an oxidation catalyst
proceeds through the exhaust system, it is cooled significantly. As the
temperature drops, the hydration of sulfur trioxide to gaseous sulfuric
acid becomes more favorable as shown in Figure A-2, where the fraction
of sulfur trioxide hydrated to gaseous sulfuric acid is given as a
function of temperature. This plot was generated using equilibrium
constants calculated from Eqn. (17) of Gmetro and Vermeulen (20).
Hydration begins at about 450°C and is complete at about 200°C. Typical
tailpipe exit temperatures range from ^200 to over 500°C (our measurements,
24) for conditions from idle to extended high speed driving. Thus, it
is possible for sulfur trioxide to either survive throughout the exhaust
system or to be completely hydrated depending upon the driving mode.
Gillespie and Johnstone (19) in studying sulfuric acid formation found
that dry sulfur trioxide would form an aerosol instantaneously upon
contact with moist air. These results imply that the hydration reaction
is extremely rapid and would be controlled by equilibrium in an exhaust
gas environment.
A.3.3 Thermodynamics of Sulfuric Acid Condensation
The dew point of gaseous sulfuric was experimentally determined
and compared with calculated thermodynamic values by Lisle and Sensenbaugh
(39). They found excellent agreement. The calculated dew point as a
function of temperature is shown in Figure A-3. More recently, Verhoff
and Banchero (56) have reviewed the literature concerning sulfuric acid
dew points and correlated the data which they felt were accurate. The
resulting least squares correlation is shown in Figure A-3 for a water
content of 12%. As this figure shows, both curves predict dew points
within a few degrees of each other in the range applicable to exhaust
sulfuric acid levels.
A typical vehicle operating on an average fuel of 300 ppm
sulfur would produce exhaust containing approximately 10 ppm H2S04
assuming 50% conversion of sulfur dioxide to trioxide. At this con-
centration, condensation would begin at M.33°C which is below the exhaust
temperature at the tailpipe exit. The rate of condensation, as pointed
out previously, is very rapid, therefore, should the exhaust temperature
-------
Figure A-2
Equilibrium Conversion
S03(g) + H20(g) == H2S0
oo
t\l
a
ra
C
o
O
n)
1.0
0.2
0.0
o\
I
100
200
300
500
Temp., °C.
-------
- 77 -
Figure A-3
Dew Point of H0S
100
10
o>
CO
CO
CO
cd
o
CM
PC
s
PL,
PL,
1.0
0.1
Lisle & Sensenbaugh
Verhoff & Banchero [56]
(for 12%,water)
o.or
90
100
110
120
130
140
150
160
Temp., C.
-------
- 78 -
drop below 133°C, condensation of the gaseous sulfuric acid would immediately
occur in the tailpipe. However, for the majority of driving situations,
the sulfuric acid would exit the tailpipe in the vapor rather than the
condensed state.
A. 4 Reaction of Sulfur Dioxide and Trioxide with Exhaust
Gas Constituents and Exhaust System Components _
There are several exhaust gas constituents and system materials
with which the sulfur oxides can react. The two main exhaust gas constituents
capable of reaction with sulfur dioxide and sulfur trioxide are ammonia
and carbon monoxide, both of which are more basic than the sulfur oxides.
The pertinent reactions are:
2NH3 + 80s + H20 5=* • (NH4) 2 S(>4 (3)
2NH3 + 3S03 v=i N2 + 3H20 + 3S(>2 (4)
S03 + CO ?—* C02 + S02 (5)
2S02 + 4CO ^-* 82 + 4C02 (6)
The first reaction is thermodynamically unfavorable at exhaust
temperatures and, further, any ammonia would be oxidized over the oxidation
catalyst before reaction with the trioxide could take place. The second
reaction is also very unlikely due to the oxidation of the ammonia. The
reduction reactions by carbon monoxide are both favorable but most likely
limited due to kinetics. The reactions of both sulfur oxides with the
iron or aluminum oxide catalyst support are favorable. The aluminum
oxide can alternately form the sulfate and decompose as the catalyst
temperature cycles during transient driving modes producing a sulfate
storage/release phenomena.
A. 4.1 Reaction of Ammonia with Sulfur Trioxide
The free energies and equilibrium constants for Reaction 3 are
given as functions of temperature in Table A-3. The thermodynamic
equilibrium constant, given by the relationship
P ~
for an exhaust gas containing 10 ppm NH3, 12% H20, and 10 ppm S03 is
8.3 x 1Q15 atm~^. Comparison of this value with the equilibrium constants
in Table A-2 shows that the reaction is favorable only at temperatures
below 225°C. Since the equilibrium constants are such a strong function
-------
- 79 -
of temperature, order of magnitude changes in the concentration of
either ammonia or sulfur trioxide will not appreciably alter the temperature
at which the reaction becomes thermodynamically favorable. Temperatures
below 225°C occur in the exhaust only at the tailpipe exit and only
during start up and some extended idle periods. Thus, the production of
ammonium sulfate in the exhaust system would be possible only during
these modes. Ammonium sulfate may, however, be formed after exiting
the tailpipe in either the atmosphere or a particulate sampling apparatus.
The thermodynamic favorability of this occurrence would depend
simultaneously upon the cooling and dilution rates.
TABLE A-2
Free Energies and Equilibrium Constants for the Reaction of
Ammonia and Sulfur Trioxide Forming Ammonium Sulfate (33)
Equilibrium
Temp Free Energy Constant
(°K) (cal/mole) (atm"4)
298 -63,830 6.18 x 1046
400 -49,100 6.75 x 102f
500 -34,900 1.80 x 1015
600 -21,000 4.47 x 107
700 - 7,400 5.32
800 + 5,800 2.60 x 10~2
The possibility of ammonium sulfate production is most likely
in catalyst-equipped vehicles where the potential of forming sulfur
trioxide is greatest. However, in these vehicles, the oxidation of
ammonia over the catalyst must be considered. The oxidation of ammonia
over platinum to produce nitric oxide is a very important and well-known
industrial reaction (15). The reaction is very rapid with almost
complete oxidation.
TABLE A-3
Free Energies and Equilibrium Constants for the Reaction of Ammonia
and Sulfur Trioxide to form Nitrogen, Water, and Sulfur Dioxide (33)
Temp
(°K)
298
400
500
600
700
800
900
1000
Free Energy
(cal/mole)
-105,510
-114,100
-122,500
-131,000
-139,400
-147,700
-155,900
-164,100
Equilibrium
Constant
(atnT4)
2.21 x 1077
2.22 x 1062
3.54 x 1053
5.26 x 1047
3.36 x 1043
2.25 x 1040
7.26 x 103?
7.36 x ID35
-------
- 80 -
A.4.2 Reduction of Sulfur Trloxlde by Ammonia
Reaction 4 illustrates a reaction mechanism whereby exhaust
ammonia can be consumed. The equilibrium thermodynamics for this reaction
are given in Table A-3. The extremely large equilibrium constants show
that this reaction has the potential of almost complete removal of
either the ammonia or the sulfur trioxide, whichever is Initially present
at lower concentration. Assuming that the reaction kinetics are sufficiently
rapid, this reaction would then preclude the formation of ammonium
sulfate.
Typical automotive exhaust without catalysts contains from one
to six ppm ammonia depending upon driving mode with an average of 2.2
for a typical driving cycle (23). Our results with the dual-catalyst
system show that an active automotive oxidation catalyst will readily
oxidize ammonia resulting in tailpipe concentrations typically less than
1 ppm. Therefore, even at low exhaust temperatures where ammonium
sulfate formation is thermodynamlcally favorable, only small amounts
could be made. This agrees with Ford Motor Company's finding (17) that
little or no ammonium sulfate or bisulfate is found in automotive particulate.
A.4.3 Reduction of Sulfur Trioxide by Carbon Monoxide
Reaction 5 is a second possible reaction which would lead to a
reduction in sulfur trioxide emissions. The equilibrium thermodynamics
for this reaction, as given in Table IV, were calculated from data
presented in the JANAF Thermochemical Tables (20). The equilibrium
constants indicate that, even for carbon monoxide concentrations in the
ppm range, the reaction is favorable at exhaust gas concentrations. The
extent of the reaction would, however, probably be limited by the reaction
rate.
TABLE A-4
Free Energies and Equilibrium Constants for the
Reduction of Sulfur Trioxide by Carbon Monoxide
Equilibrium
Temp Free Energy Constant
(°K) (Cal/mole) (dimenslonless)
500 -44,870 4.11 x 1019
600 -45,020 2.51 x 1016
700 -45,160 1.26 x 1014
800 -45,310 2.39 x 1012
900 -45,450 1.09 x 1QH
1000 -46,670 1.59 x 1010
-------
- 81 -
A.4.4 Reduction of Sulfur Dioxide by Carbon Monoxide
The last possible reaction is the reduction of sulfur dioxide
to elemental sulfur by reaction with carbon monoxide. The thermodynamic
data calculated from the JANAF Tables (30) are presented in Table A-5.
The equilibrium constant for this reaction is written as,
4
PC02
Kp - § =•
PCO PS02 (8)
In a pre-catalyst vehicle, typical exhaust concentrations would be
PCQO =0.12 atm, PgQo = 2 x 10~5 atm, and PCO =0.01 atm. At these levels,
the reduction of sulfur dioxide by carbon monoxide would become favorable
at temperatures below 600°C. In a catalyst vehicle the partial pressure
of carbon monoxide would be typically less than 10"^ atm in which case
the reaction would become favorable at temperatures less than 375°C.
Hence, for both catalyst and non-catalyst vehicles, the reduction of
sulfur dioxide to elemental sulfur is thermodynamically feasible within
the exhaust system.
TABLE A-5
Free Energies and Equilibrium Constants for the
Reduction of Sulfur Dioxide by Carbon Monoxide
Equilibrium
Temp Free Energy Constant
(°K) (cal/mole) (atm~2)
500 -73,390 1.21 x 1032
600 -68,440 8.54 x lO2*
700 -63,560 7.01 x 1019
800 -58,660 1.06 x 10?-°
900 -53,800 1.16 x 1013
1000 -48,540 4.07 x 1010
A. 4. 5 Reaction of Sulfur Oxides With Iron
The most prevalent solid material with which sulfur oxides can
react is the iron of the engine and exhaust system. Since the newer
model cars and all future oxidation catalyst vehicles are operated net
lean, either by carburetion or by air injection, the internal iron
surfaces of the exhaust system will be oxidized. A typical iron-sulfur
oxide reaction is the sulfation of ferric oxide by sulfur trioxide,
Fe203 + 3S03 f="*Fe2 (804)3
-------
- 82 -
A similar type of reaction can also be written for sulfur dioxide where
iron sulfite is the product. These types of reactions are industrially
important in the corrosion of iron or steel surfaces in contact with
flue or stack gases containing oxides of sulfur. In this vein, the
sulfation of iron has been examined by several researchers (35, 36, 24).
Warner and Ingraham (58, 59) have also investigated the reverse reaction
in^their studies on the processing of metallic ores.
These studies as well as thermodynamic calculations, show that
the sulfation of ferric oxide by sulfur trioxide in the concentration
range of catalyst vehicles becomes favorable at temperatures below
450°C. Table A-6 contains the pertinent thermodynamic data. Therefore,
sulfation is possible over a large fraction of the exhaust system.
Furthermore, there will be a zone in which the temperature will oscillate,
as the driving mode changes, around the temperature at which the reaction
is favorable. It is, therefore, possible for this zone to either pick
up or release sulfur oxides depending upon the temperature and sulfur
oxide concentration.
TABLE A-6
Free Energies and Equilibrium Constants for the
Reaction of Sulfur Trioxide with Ferric Oxide
Temp
(°K)
500
600
700
800
900
1000
Free Energy
(cal/mole)
-70,730
-57,790
-49,940
-32,170
-19,420
- 6,720
Equilibrium
Constant
(atnT3)
8.29 x 1030
1.13 x 102J
3.92 x 10~
6.15 x 108
5.20 x 10?
2.94 x 101
Warner and Ingraham (59) allude to the mechanism of ferric
sulfate decomposition. Their work with decomposition under atmospheres
of varying sulfur dioxide and trioxide compositions indicates that the
decomposition products are more likely to be sulfur dioxide and oxygen
than sulfur trioxide.
Therefore, it is possible to store sulfate in the exhaust
system by the reaction of sulfur trioxide with iron and then to release
this stored sulfate as sulfur dioxide and oxygen.
-------
- 83 -
The relative reaction rates of iron sulfation by sulfur dioxide
and sulfur trioxide can be assessed from the work of Pechkovsky (44) and
Krause, et_ al . (35). Pechkovsky studied the reaction of sulfur dioxide
with various metal oxides in an oxidizing environment. He showed that
sulfation of some metals, notably magnesium oxide, was very slow with
only sulfur dioxide present. When a small amount of catalyst, such as
ferric oxide, was added, the reaction proceeded much more rapidly,
presumably due to the oxidation of sulfur dioxide to the trioxide which
then reacted more readily. Krause, et^ al. (35), using a radioactive
tracer technique, found sulfur trioxide to be on the order of 10^ times
more reactive on a molar basis than the dioxide.
A. 4. 6 Reactions of Sulfur Trioxide with Aluminum Oxide
Another material in the exhaust system of a catalyst-equipped
vehicle with which sulfur trioxide can react is the aluminum oxide
catalyst substrate via the reaction
A1203 + 3S03 *^ A12(S04)3. (10)
The thermodynamic data for this reaction (33) are given in Table A- 7 for
gamma form of aluminum oxide. The equilibrium partial pressures for
sulfur trioxide are also presented in this table. These data show that
this reaction is favorable for typical sulfur trioxide exhaust gas
compositions when the temperature is below 425°C. While no quantitive
data are available on the kinetics of this reaction, qualitatively they
are rapid enough to be significant as witnessed by the sulfate storage
noted by several investigators (2, 18, 45).
TABLE A-7
Free Energies, Equilibrium Constants, and Equilibrium
Partial Pressures for the Reaction of
Sulfur Trioxide with Gamma Aluminum Oxide
Temp
(°K)
298.16
400
500
600
700
800
900
1000
Free
Energy
(cal/mole)
111,000
96,900
83,200
69,900
56,700
43,800
30,900
18,200
Equilibrium
Constant
(atm~3)
2.34 x 1081
8.87 x 1052
2.34 x 103°
2.91 x 1025
5.05 x 1017
9.25 x 1011
3.19 x 107
9.50 x 103
Equilibrium
Partial Pressure
of Sulfur Trioxide
(atm)
7.53 x 10~28
2.24 x 10-18
7.53 x 10~13
3.25 x 10~9
1.26 x 10"°
1.03 x 10~4
3.15 x 10~3
4.72 x 10~ 2
-------
- 84 -
The release of this stored sulfur can also play an important
part in automobile sulfate emissions. As the above reaction indicates,
the aluminum sulfate can decompose back to the oxide with the release of
sulfur trioxide. Such a release would become thermodynamically favorable
above 420°C. Warner and Ingraham have examined the thermodynamics (58)
and kinetics (59) of the decomposition of aluminum and ferric sulfates.
As discussed above, they found that the rate of ferric sulfate decomposition
was proportional to the sulfur dioxide driving force. That is to say
that the decomposition favors the formation of sulfur dioxide rather
than the trioxide. Although they did not investigate the mechanism of
aluminum sulfate decomposition, by analogy with ferric sulfate decomposition,
aluminum sulfate could decompose into oxygen and sulfur dioxide. Thus,
it could be possible for sulfur trioxide formed on the catalyst to react
with the substrate only to be released later at a higher temperature as
sulfur dioxide. With this possibility, care must be exercised in designing
and interpreting experiments examing the fate of gasoline sulfur over an
oxidation catalyst.
Kelly (33) presents an alternate route for aluminum sulfate
decomposition in the presence of carbon monoxide. The reaction is
Al2(S04)3 + 3CO^ A1203 + 3S02 + 3C02. (11)
The equilibrium constant for this reaction, K1/^ • pgo ^cOo^CD is on
the order of 3 x 10^ atm for all temperatures. Since 2his equilibrium
constant is very large, the partial pressure of carbon monoxide must be
extremely small before this reaction becomes unfavorable. Thus this
reaction may provide a route for sulfate release. Kelley feels that the
actual mechanism of this reaction is the decomposition of aluminum
sulfate to sulfur trioxide which is subsequently reduced to sulfur
dioxide by carbon monoxide. He feels that the addition of carbon monoxide
may reduce the decomposition temperature of aluminum sulfate by about
50°C.
A.5 Automotive Catalysis of Sulfur Dioxide
The only oxidation catalysts presently envisioned for automotive
application are platinum and platinum-alloy catalysts. This section
will, therefore, be limited to these catalysts. The bulk of the literature
covering platinum oxidation of sulfur dioxide pertains mainly to use in
sulfuric acid plants and, to a much lesser extent, in flue gas sulfur
dioxide control. In these applications, the sulfur dioxide concentrations
are much higher, 5-8% and 1% respectively, than found in typical automotive
exhaust, 20 ppm. Even though the exhaust concentration is orders of
magnitude below the percent level, the kinetic mechanism of catalysis
over platinum should be the same. This suggests that the rate equations
would also be valid. The first part of this section will review this
literature. The second part will review the more limited, recent literature
dealing directly with the sulfur dioxide oxidation over automotive
catalysts.
-------
- 85 -
A.5.1 Platinum Catalysis; Industrial Application
The use of platinum as a catalyst in the oxidation of sulfur
dioxide was first disclosed in a patent by Peregrine Phillips in 1831
(41). This discovery is considered the foundation of the contact process
for manufacturing sulfuric acid.
It wasn't until the 1900's that quantitative data became
available concerning the rate and the mechanism of the oxidation reaction
over various forms of platinum. This began with the classic work of
Knietsch (34) in 1901. Knietsch examined the effect of space velocity
and inlet concentrations on the conversion of SO2 as a function of
catalyst temperature. The catalyst consisted of 0.5 gm. of platinum
supported on 5 to 10 gms. of asbestos. The oxygen concentration was
held constant at 10% and 7 - 20% S02 concentrations were examined over
the temperature range of 300 to 900°C. For this range of variables,
conversions of from 0 to 99% were obtained. At a given space velocity,
the experimental conversion was low at 300°C. As the temperature was
increased, the conversion increased, rapidly reaching a maximum in the
temperature range of 400-500°C depending upon space velocity. Maximum
conversions were in the range of 70 to 99% again depending upon space
velocity. The lower space velocities peaked at lower temperatures and
at greater conversions. As the temperature was increased further, the
conversions decreased paralleling equilibrium with the lower space
velocities nearer equilibrium. The limited data at the higher 802 level
(20% instead of 7%) showed a decrease in conversion at identical temperature
and space velocity. From these data, Knietsch deduced that the reaction
rate was proportional to the concentrations of sulfur dioxide and oxygen
(61) and is independent of the concentration of sulfur trioxide. The
accuracy of these data have since been questioned due to Knietsch1s poor
temperature control and large conversions.
In another classical study of heterogeneous catalysis, Boden-
stein and Fink (6,7) investigated the oxidation of SO2 over a platinum
gauze of 0.06 mm diameter wire. This study, conducted over the experimental
temperature range of 150-250°C, produced a reaction rate which is proportional
to the first power of the sulfur dioxide concentration and inversely
proportional to the square root of the sulfur trioxide concentration for
the case where Cs02/C()2 ** 3. Where this ratio is greater than 3, the
reaction rate is proportional to the first power of the oxygen concentration
and inversely proportional to the square root of the sulfur trioxide
concentration. The usefulness of these rate equations is, however,
somewhat limited due to the very low temperatures investigated.
Lewis and Ries (37, 38), objecting to the poor control of
experimental conditions by Knietsch and to Bodensteln's low temperature
range, obtained catalytic data under carefully controlled conditions
approximating actual contact plant operations. Their kinetic experiments
-------
- 86 -
were designed and conducted so as to approximate a differential reactor,
that is, a reactor in which the conversion is allowed to change only to
a small extent. Thus, the reaction is taking place throughout the
reactor with approximately the same reactant and product concentrations.
In addition, the temperature does not change due to reaction exo- or
endothermicity since conversions are differential. This is the classically
accepted method for obtaining accurate kinetic data.
Lewis and Ries performed three separate series of tests. In
the first series the effect of inlet sulfur dioxide concentration was
examined by varying the amount of sulfur dioxide in an air-sulfur dioxide
feed. In the second series, the amount of oxygen was varied by dilution
of the air-sulfur dioxide feed with nitrogen. The final series examined
the effect of sulfur trioxide on the reaction kinetics. For this series,
a platinum preconvertor was used to generate feeds with varying sulfur
dioxide and trioxide concentrations.
An attempt was then made to fit these data with an equation
obtained directly from the law of mass action
r -k (S02)2 02 - 3 . (12)
1x6
This equation, which they found to fit Knietsch's data, would not fit
their data. They then tried Bodenstein's equation,
r = k S02/S03 1/2 (13)
which was somewhat better than the law of mass action but still incapable
of interpreting the data. Several other forms of rate equations were
tried with the best results being obtained with the form
, „ In Pesoi P<;OO
r - k ps°2 p-^pe^ ; U4)
PS03 * S02
where the superscript e denotes the equilibrium partial pressure. This
rate equation also correlates the data of Bodenstein and Fink better
than the rate equation proposed by Bodenstein and Fink.
Unfortunately, as pointed out by Uyehara and Watson (55), the
total amount of platinum used by Lewis and Ries was not determined. The
amount used was constant for all of the experiments making the rate data
consistent. However, it is impossible to derive absolute reaction rate
constants since the amount of catalyst is unknown.
-------
- 87 -
Taylor and Lenher (54) used a static platinum hot-wire technique
to examine the approach to equilibrium from both sides at a temperature
of 665 °C. They found that the following rate equation best represented
their data
PS03
In 1937, Salsas Serra (50) examined the experimental work of
Knietsch and Bodenstein and Fink in light of the law of mass action. In
opposition to the findings of Lewis and Ries discussed above, he found
that the law of mass action adequately represented these data. The law
of mass action for the oxidation of sulfur dioxide yields the following
rate equation,
r " kl p2S02 P02 " k2 p2S03- <16>
Hougen and Watson (26) developed general rate equations for
heterogeneous catalysis where one of the elementary reaction steps in
the overall series of steps is assumed to be rate controlling. In
general, the elementary reaction steps, neglecting mass transfer steps,
of the general reaction of A + B to produce C are:
1. The adsorption of either or both reactants
on the catalyst surface.
2. The reaction of A + B either both as
absorbed species or as one absorbed specie
and one gaseous specie.
3. The desorption of the produce from the
catalyst surface.
Assuming that one of these steps is rate controlling, i.e., much slower
than the other steps while the remaining steps are at equilibrium, a
general rate equation can be written for each case. Uyehara and Watson,
in a companion paper (55) , applied this procedure to the catalytic
oxidation of sulfur dioxide. The data of Lewis and Ries were selected
for analysis since they were obtained in an apparatus reasonably resembling
a differential reactor. The data of Knietsch, Bodenstein and Fink, and
Taylor and Lenher were unsatisfactory since they were all obtained in
static systems In which concentration and temperature gradients were
present. Uyehara and Watson examined each possible rate limiting case
-------
- 88 -
and determined that the limiting step is
1/2
kl
r _
y * 1/2 \o
(1 + K02 P02 + KS03 PS03' (I?)
where KQ? and KSQO are the absorption equilibrium constants for oxygen
and sulfur trioxide respectively. It should be noted that Uyehara and
Watson have dropped the sulfur dioxide absorption term from the denominator
of the general rate equation. Based on Lewis and Ries' first experimental
series looking at the effect of sulfur dioxide concentration, Uyehara
and Watson concluded that this absorption term is negligible. Expressions
for the coefficients appearing in this equation were updated by Hougen
and Watson (25) by the inclusion of Hurt's data (28). Since the mass of
Hurt's catalyst was known, an absolute rate constant could be determined.
These coefficients are:
8.000
k.. = exp - ' + 14.154, . (18)
(19)
K
KS03 - exp ±E*°ifii _ !L51 ^ ( }
RT R
Boreskov (8) found that the data of Bodenstein and Fink,
Taylor and Lenher and Pligunov (unpublished) agreed very well with the
rate equation
r - k, PS02 P02 - k.
0.5 0.25 (2D
PS03
He further states that this equation can be accounted for by assuming
that the rate limiting step is the absorption of S02 on the catalyst.
This equation was also successfully used by Cheslova and Boreskov (11,
12) in the study of catalytic activity of various platinum catalysts.
Roiter, ejt al/» (47) examined the oxidation near equilibrium
of sulfur dioxide over a platinum screen by means of a tracer technique
using a radioactive sulfur isotope. The kinetics were also examined
under non-equilibrium conditions using a standard static measurement
technique. The experimental temperature ranged from 600 to 674°C while
the initial sulfur dioxide level ranged from 2.7 to 6.3%. The data were
then fit to the Boreskov equation with large discrepancies being noted.
-------
- 89 -
A rate equation was then developed assuming that the rate determining
step is the surface reaction of adsorbed oxygen and sulfur dioxide
™
r=kPS02P02 - (22)
This rate equation agreed quite well with the data under equilibrium and
non-equilibrium conditions.
At this point in historical review, it is instructive to
examine some of the common features of the studies and rate equations
reported. One Important experimental aspect, which is particularly
germane to actual large-scale reactors such as on vehicles, is the
neglect of mass transfer effects in the analysis of the data. This
neglect of mass transfer may be responsible for the difference in rate
equations. For convenience, Table A-8 summarizes the rate equations and
pertinent experimental conditions. The effect of mass transfer has been
examined in more recent literature and will be discussed later.
All of the rate equations, with the exceptions of Bodenstein
and Fink's second equation and Salsas Serra's equation, predict that the
forward reaction rate is proportional to the first power of the sulfur
dioxide partial pressure. The functional dependency on the oxygen
partial pressure is somewhat less clear. Several of the early equations
are independent of oxygen partial pressure, however, in these cases the
experimental oxygen levels were in large excess. The experimentally
small concentration changes due to reaction were almost insignificant in
comparison with the large unreacted excess. This and other experimental
inaccuracies yielded equations which were independent of oxygen effects.
Such was the case of Lewis and Ries' data and analysis. Later analysis
of these data by Uyehara and Watson did, however, show an oxygen dependency
proportional to the square root of the oxygen partial pressure. This
same oxygen functionality was also determined by Roiter where various
oxygen concentrations were examined.
In essentially all of the experimental work, a marked reduction
was noted with the presence of even very small amounts of sulfur trioxide.
Two methods of accounting for this retardation were used in the develop-
ment rate equation. The first method was to place the sulfur trioxide
partial pressure in the denominator of the forward reaction rate term.
The second method was to include a reverse-reaction term in the rate
equation such as exemplified in the Uyehara and Watson and Roiter equations.
The second method also provides mathematical tractability in that the
rate can go to zero as equilibrium is approached. The first method, see
for instance the equations of Bodenstein and Fink or Lewis and Ries, does
not account for equilibrium conditions and, therefore, would be unacceptable.
-------
TABLE A-8
Investigator
Knietsch [ 34 ]
Bodenstein and Fink [6,7 ]
Bodenstein and Fink [ 6,7]
Lewis and Ries [ 37.38J
Taylor and Lenher [54}
Salsas Serra [ 5tt]
Uyehara and Watson [ 55 ]
Iloreskov [ 8 ]
Roiter e£ al . [ 47 ]
Olsen et al. [ 42 ]
SUMMARY OF RATE EQUATIONS FOR PLATINUM CATALYSIS OF THE
Rate Equation
r = kj^ PSQ2 PQ2
-0.5
r = ki PS02 PS03
-0.5
r = ki POZ Pso3
PeS03 PS02
r = ki PS02 !»
peso2 pso3
-0 5
r = ki (PSQ2 - PeS02) PS03 ' - k2 (Peso.j ~ PS03>
2 2
r = ki P S02 P02 ~ k2 Pso3
kl n s £
-n p —
r ,, „ 0.5 0.5 „ ,2 P"2 r°u2 K
(1 .+ PQ2 KOZ + PSQ.J Kgo3;
r ° kl PSOZ P62 P02 ' PS03~ ' ~ k2 PQZ
kP °'5P PS°3
r = (Pen P °'5 PS03^ (B + D Pen r2
r " (.Pso2 "02 ~ Ke "S03'
OXIDATION OF SULFUR DIOXIDE
Catalyst S02 02 Conversion
Description Range Range Range
Pt on asbestos 7% + 2-% 10% 0-99%
Pt mesh, 0.06mm .
diameter PS02'P02 '3
7% Pt on asbestos <0.5% air
Pt wire
Analysis of data of Knietsch
[34] and Bodenstein [6,7]
'03
• (Analysis of Lewis and Ries data [L— 4,5] with
inclusion of Hurt data [H-8]
Pt wire
Pt screen 1-3% ^19 0-100%
0.2% Pt on 1/8 6.45% air 4-70
alumina pellets
(50,000 v/v/hr.
Temp.
Range
300-900 °C
150-250°C
150-250°C
400-450°C
525-700°C
530-850°C
420-700°C
350-480°C
-------
- 91 -
Of the equations discussed, the best with respect to data
representation, mathematical formulation and generality is that of
Uyehara and Watson. This equation has the general form of the law of
mass action and represents the reliable differential reactor data of
Lewis and Ries quite well. In addition, the absolute rate constants are
available where they are not for several of the other equations.
The effect of diffusion on the overall sulfur dioxide conversion
of a large-scale reactor can be determined from the work of Hurt (28).
Hurt used the oxidation of sulfur dioxide over platinum as an example of
his procedure for correlating the performance of small scale with large
scale reactors. An estimate of the influence of mass transfer on the
overall sulfur dioxide conversion in a Of reactor can be made using
Hurt's kinetic data and mass transfer correlations.
At 40 mph cruise conditions, a typical superficial mass
velocity would be 248 Ib/hr-ft^. This represents a particle Reynolds
number of 30 for 1/8" pellets. Using Hurt's kinetic correlation and
mass transfer correlation, the overall conversion is Influenced approximately
25% by mass transfer and approximately 75% by kinetics. Although these
values may be in error due to kinetic differences in Hurt's catalyst and
automotive catalysts, it does illustrate that mass transfer effects can
play a significant role in the formation of automotive sulfates. It
also points out that care should be exercised in designing and analyzing
kinetic experiments particularly when the goal of the experiment is to
predict large-scale reactor behavior from small-scale reactor experiments.
Smith, et_ al, (1, 22, 29, 42, 52) completed a very comprehensive
program involving the design of fixed bed catalytic reactors taking into
account heat and mass transfer. The program Involved measurement of the
chemical kinetics in a differential reactor, examination of diffusional
effects, examination of heat transfer effects, and mathematical modeling
of an integral reactor incorporating kinetics and heat and mass transfer.
The results are of particular interest since the reaction used was
oxidation of sulfur over a catalyst of 0.2 weight percent platinum on
1/8" alumina pellets. The platinum was applied such that it penetrated
only the outer 1/32" of the pellet. This platinum loading and geometry
approximates that of the GM catalyst. In addition, the range of temperature
and space velocities spans the range of interest in automotive applications.
However, the sulfur dioxide levels were constant and typical of acid
plant operations, 6%, and the oxygen levels were those of air. Thus,
the only variable Investigated was conversion. Overall conversions up
to 70% were examined in one paper by using a preconverter ahead of the
differential reactor (42). The maximum conversion over the differential
reactor was in all but two cases less than 30%.
-------
- 92 -
Olsen, et al., (42) chose to model their results with the rate
equation of Uyehara and Watson (55). Since the inlet concentrations to
the preconverter were always the same, there existed a definite relationship
between Ps02» PS(>3» and P02' Therefore, the Uyehara and Watson equation
could be simplified, using the relationship, to
1/2
PS02 P02 - -- (23)
where B and D are constants dependent only upon temperature. Olsen, et
al., found that this equation was an excellent representation of their
data when the component partial pressures were taken at their interfacial
values and not their bulk values, that is to say that diffusion is taken
into account.
In a latter paper of this series, Argo and Smith (1) obtained
differential reaction rates using larger catalyst pellets than Olson,
1/4" in place of 1/8". These data were then correlated with the Olson
data by an activity factor in conjunction with the rate equation of
Olson. The activity factor and rate equation correlated both sets of
data very well again illustrating that this general equation is capable
of describing the catalytic oxidation of sulfur dioxide by platinum.
Olson also examined five other rate controlling mechanisms. The data
were correlated by rate equations derived from the following rate
limiting elementary steps:
1. reaction of adsorbed sulfur dioxide and
unadsorbed oxygen (this mechanism yields
the Uyehara and Watson rate equation),
2. reaction of adsorbed sulfur dioxide with
gas-phase molecular oxygen, or
3. reaction of adsorbed oxygen and gas phase
sulfur dioxide.
Unfortunately, the data could not distinguish among these three steps.
The data were, however, not correlated by any mechanism which assumes
that adsorption or desorption is the rate limiting step.
The differential reaction data were also analyzed by Hurt's
(28) method which also accounts for the effect of diffusion. This
method failed to correlate the data with the primary objection being
that the resistance due to reaction at the catalytic surface was not
independent of mass velocity. Olson found that the diffusional effects
-------
- 93 -
were better accounted for by using the mass-transfer correlations of
Hougen and Wilkie (27). Using these correlations, differences in the
sulfur dioxide and trioxide partial pressures between the bulk gas and
the catalytic surface were as great as 40% depending upon gas mass
velocity, degree of conversion, and temperature. For a GM reactor
running at 40 mph, partial pressure differences on the order of M.0%
would be predicted for both sulfur dioxide and trioxide. Therefore,
this treatment in qualitative agreement with Hart's method Indicates
that mass transfer can be playing a role in the production of sulfurlc
acid over the GM type of automotive catalyst.
The mechanism of the oxidation of sulfur dioxide over platinum
was examined by Kaneko and Okanaka (31, 32) using a radioactive tracer
technique. The experiments were conducted near 400°C in a reclrculating
reactor using a platinum gauze catalyst of 0.1 mm diameter wire. The
kinetic mechanism was examined on both sides of equilibrium, i.e.
oxidation of sulfur dioxide and decomposition of sulfur trioxide. The
reaction mixture contained only oxygen, sulfur dioxide, and sulfur
trioxide with the oxygen and sulfur dioxide always in stoichiometric
amounts. In one series of experiments, radioactive sulfur was used to
follow the reaction. In the second series, an isotope labelling of
oxygen was employed. The combined results of both series show that the
rate-determining step is the surface reaction of adsorbed oxygen and
adsorbed sulfur dioxide.
In summary, the literature covering the mechanism of sulfur
dioxide catalysis by platinum indicates the rate limiting step is the
surface reaction between adsorbed oxygen and adsorbed sulfur dioxide.
The rate equation which produced the best representation of experi-
mental data was that of Uyehara and Watson (55) which is in agreement
with the above rate controlling mechanism. Most of the other rate
equations are special cases of the more general Uyehara-Watson equation
which further substantiates its validity.
The rate equation indicates two factors which should be con-
sidered with respect to automotive sulfate emissions. First, since the
rate equation is first order with respect to sulfur dioxide partial
pressure, sulfuric acid production will be linearly proportional to fuel
sulfur level. Second, since the reaction is half order in oxygen,
reducing excess oxygen partial pressure to zero, as is done with the
three-way catalyst system, should reduce sulfate emissions to very low
levels.
A.5.2 Platinum Catalysis: Automotive Application
Although the literature is void of reaction rate and reaction
mechanism studies which relate directly to exhaust oxidation of sulfur
dioxide, there are several references covering phenomenological observation
-------
- 94 -
of the catalysis of exhaust sulfur dioxide. Chrysler (13), Ethyl (16),
Ford (17) , and General Motors (18) have made public their data on sulfate
emissions as requested in the March 8, 1974 Federal Register. In addition,
Pierson, et_ al. , (45) and Beltzer, ejt al., (2) have presented SAE
papers dealing with automotive sulfate emissions. All of these references,
with the exception of General Motors, have examined platinum catalysis
only from the standpoint of tailpipe emissions of sulfuric acid.
General Motors has conducted a limited amount of laboratory work aimed
directly at elucidating the kinetics.
In reviewing these references, there is, for the most part, a
common problem in not being able to obtain a balance when catalysts are
employed between the sulfur burned in the fuel and the sulfur emitted
from the tailpipe. When catalysts are not used, balances can be obtained.
Since the majority of the tailpipe sulfur in a non-catalyst vehicle is
sulfur dioxide and a sulfur balance can be made, the analytical and test
procedures for sulfur dioxide can be assumed to be valid and accurate.
The ability then not to be able to close a sulfur balance on a catalyst
vehicle is due either to an experimental or analytical problem or to the
storage phenomena discussed previously. This problem makes it difficult
to assess quantitative catalytic effects from observed tailpipe emissions.
Ford (17, 45) has been able to make sulfur balances for Engelhard
IIB-catalyst and non-catalyst vehicles operating under steady cruise
conditions. They found that at 60 mph both the Engelhard IIB and GM
catalysts converted 44% of the sulfur dioxide to trioxide, which is
close to the equilibrium conversion at the test conditions. The GM
catalyst at 30 mph converted 84% of the sulfur dioxide which is again
very close to the equilibrium value. These conversions were all calculated
on the basis of total sulfur out of the tailpipe, thus eliminating the
storage problem noted with the GM catalysts.
Ethyl Corp. (16) found that under 40 mph cruise testing,
monolithic platinum catalysts converted 43% of the fuel sulfur to exhaust
sulfate where conversion is based on total sulfur emitted. Their tests
showed that approximately 58% of the fuel sulfur was emitted with the
rest being stored. This conversion cannot be compared with equilibrium
since the catalyst temperature was not given. The magnitude of these
results are in line with Ford's findings.
Beltzer, et al., (2) have also measured sulfate emissions for
steady cruise conditions. Sulfur dioxide measurements were not made
precluding making a sulfur balance. The sulfate emissions at 60 mph for
the pelletized oxidation catalyst agree fairly well with those obtained
by Ford using approximately the same fuel sulfur level. If the storage
effects are similar then by implication, this conversion is also close
to equilibrium.
-------
- 95 -
The GM (18) extended 60 mph cruise tests on their catalysts
show a large spread both in sulfate conversion and in sulfur balances.
In general, average conversions based on sulfur emitted were in the
range of 20 to 40%. This is somewhat lower than the conversions
discussed above, but still represents a significant approach to
equilibrium.
The Chrysler reference (13) did not present conversion data
for steady cruise conditions. A consensus of these references shows
that the reaction rate is of sufficient magnitude that the oxidation
of the exhaust sulfur dioxide approaches equilibrium in actual vehicle
operation. Due to the storage effect together with possible analytical
problems, a more quantitative assessment of the reaction rate is
impossible.
In addition to their vehicle work, GM (18) also reported some
preliminary parametric studies using a laboratory-scale reactor. They
also caution the reader that the analytical methods caused significant
uncertainties in the sulfur trioxide measurements. The storage problem
was also a large factor. In one instance, less than 30% of the inlet
sulfur could be accounted for.
GM examined the conversion both as a function of temperature
and space velocity using actual vehicle exhaust. They found that the
conversions, based on measured sulfur trioxide and average sulfur input,
were between 5 and 15% with only a weak temperature dependency. From
analysis of these data, with respect to temperature and space-velocity
dependency, it was concluded that the reaction is kinetically controlled.
The effect of platinum loading was also investigated by increasing the
loading from 0.1% to 1.0% platinum. The observed effect was only a very
slight increase in measured sulfur trioxide. This finding is in contra-
diction to a kinetically controlled reaction. If the reaction were
kinetically controlled for both loadings, then an order of magnitude
change in platinum loading would change the rate by an order of magnitude.
Invariance to loading conforms with either a diffusion or equilibrium
controlled reaction.
It is informative to examine these data assuming that the
outlet sulfur trioxide is the difference between the outlet sulfur
dioxide and the total inlet sulfur. This assumption could be valid if
the sulfur trioxide analyses were in error and no storage was taking
place. Under this assumption, conversions are all very near equilibrium
indicating that the reaction is equilibrium controlled. Some credibility
must be assigned this probability since GM's platinum loading data also
indicate an equilibrium controlled reaction. However, since the storage
problem is real and the data severely limited, this assumption may not
be entirely valid. Nonetheless, the possibility exists that more sulfur
dioxide is being produced than the GM data indicate.
-------
- 96 -
The form of the rate equation describing the catalysis reaction
can, to a limited extent, be deduced from the above studies. The results
of Beltzer, Ford, and GM show that sulfate emissions increase with
gasoline sulfur level. Beltzer and Ford find that conversion appears to
be invariant with respect to fuel sulfur level which means the reaction
rate is proportional to the first power of the sulfur dioxide partial
pressure in agreement with the rate equations previously discussed. The
quantitative effect of oxygen partial pressure can be assessed from GM's
vehicle and laboratory studies. These studies show that reducing the
partial pressure of oxygen over the catalyst reduces sulfate emissions
and the reaction rate. The quantitative dependency on the reaction rate
cannot be determined due to the limited data of dubious accuracy.
In summary, the use of a platinum automotive-exhaust catalyst
does result in oxidation of fuel sulfur to yield sulfur trioxide. The
extent of this oxidation is clouded by the problems of sulfur storage
and inaccurate analytical techniques. Results showing anywhere from
10% to complete approach to equilibrium have been observed. The variability
of these data precluded any quantitative analysis of the reaction rate,
however, the data did indicate the form for the reaction-rate equation.
The rate equation should be linear with respect to sulfur dioxide partial
pressure and should be proportional to some positive power of the oxygen
partial pressure. This form is in agreement with the rate equations
obtained in the industrial catalysis section.
A.6 Sulfate Trap
There are two conceivable means of removing the sulfuric acid
from a vehicle's exhaust system. The first is by the use of particulate
traps in which the condensed acid is removed as droplets. The second
is by reaction of the acid either as sulfur trioxide, gaseous sulfuric
acid, or condensed, liquid sulfuric acid with a suitable sorbent.
The capacity of either type of trap can be conservatively estimated for
50,000 miles by assuming an overall fuel economy of 10 miles per gallon.
If the fuel sulfur level is taken as the industry average of 0.03 weight
percent, and it is assumed that the worst possible case of total conversion
exists, then 4.2 kgm or 131 gm. moles, of sulfur will be consumed and
must be trapped.
The particulate trapping technique has several serious problems
associated with it which preclude its use in automotive applications.
The sorbent trap method is more attractive. Consideration of several
important properties which a sorbent material must process shows that the
most promising sorbent material is calcium oxide. Other attractive
sorbents are the oxides of magnesium, manganese, and aluminum.
-------
- 97 -
V.6.1 Particulate Trap
There are several very serious problems associated with
collecting these sulfuric acid emissions with a particulate trap. The
largest problem is that the sulfuric acid exits the tailpipe in the
gas phase for the majority of driving conditions. This has been discussed
previously. Therefore, for a particulate trap to be feasible, an exhaust
heat exchanger would be required. Assuming condensation is possible,
then the condensed acid must be separated from the exhaust gas by some
means. This separation will be exceedingly difficult due to the extremely
small particle size of the condensed acid. Typically, the particle
sizes for the effluent of acid plants is less than 2 microns for 85 to
90% of the acid by weight (53). Ford (17) found that >90% of the exhaust
sulfate mass was less than 0.25 microns. The problems associated with
this separation would be almost unsurmountable given the space limitations
and low back-pressure requirement. Assuming these problems can be
overcome, the problem then becomes one of containing the liquid acid.
Assuming the acid would be diluted by 50 weight percent water, the
volume of acid solution collected over 50,000 miles would be 18.5
liters. Not only must this be contained, it must be protected from
further dilution with water during cold starts and shut downs. Based
on all of these problems, it is not feasible to use particulate traps
to remove automotive sulfate particulate.
A.6.2 Sorbent Trap
The second method of reducing exhaust sulfate particulate is
to trap the sulfur trioxide or sulfuric acid by chemical reaction with
a solid sorbent material. In this manner, the potential sulfate
particulate is trapped and stored in the exhaust system as a solid
material. The obvious sorbents to consider for this application are
solids which are chemically basic. While there are only a few references
in the literature concerning the reaction of basic materials with either
sulfur trioxide or gaseous sulfuric acid, there are numerous references
to reactions with sulfur dioxide. The majority of these references are
concerned with the selection of potential sorbent material for the
removal of sulfur dioxide from various stack gases. Since the general
reaction of either oxide is a gaseous acidic component reacting with a
basic solid component and since the trioxide or gaseous acid is more
acidic than the dioxide, then it would be expected that suitable sulfur
dioxide sorbents would also serve as suitable trioxide sorbents. In
fact, several researchers (3, 44) have found that the reactivity of the
trioxide is orders of magnitude greater than that of the dioxide. Thus,
the literature of sulfur dioxide removal will be used as a basis for
selecting potential sulfate trap sorbents.
-------
- 98 -
In selecting potential sorbents for automotive exhaust
application, there are several important factors which must be con-
sidered. These factors are 1 - activity for sulfur trioxide or sulfuric
acid removal, 2 - thermal stability, 3 - volume and weight restrictions,
4 - potential side reaction, 5 - water solubility, 6 - cost and avail-
ability of sorbent, and 7 - toxicity of fresh or sulfated sorbent. The
importance of the first factor goes without saying, a sorbent must
be able to sorb over the temperature range encountered in automotive
exhaust. The thermal stability of the sulfated sorbent is also vital
to potential sorbents. The sulfated sorbent must not begin to decompose
at temperatures below the maximum expected operating temperature of the
catalyst. With this restriction, sulfated sorbents which decompose at
temperatures below 800°C will not be considered.
Since the sulfate trap must be located in the exhaust system
on the underside of the vehicle where space is limited, volume and
weight restrictions are important. Another important factor to consider
in the selection of sorbent materials is the possibility of side reactions
with exhaust gas constituents such as water or carbon dioxide to form
stable compounds. Such reactions would use potential sulfate sorbent
and decrease the sulfation capacity of the trap.
The water solubility of the fresh and sulfate sorbent material
must be considered. There are several conditions in which part or all
of the sulfate trap could be exposed to liquid water. If the sorbent
is appreciably soluble, then the potential sulfate capacity can be
decreased. This leaching of material can also lead to problems in
maintaining the structural integrity of some possible trap configurations.
Since a successful sulfate trap has the possibility of being
installed on millions of vehicles, the cost and availability of the
sorbent must be considered. Thus, expensive metals such as gold or
silver or materials available in limited quantities such as various
rare earth elements cannot be considered.
Since the trap represents a potential source of particulate
emissions due to fresh or sulfated sorbent attrition, the toxicity of
the fresh and sulfated sorbent as well as other possible sorbent
compounds must be considered. For instance, beryllium would be ideal
from the volume-weight aspect due to its low molecular weight and
divalency. However, it and its compounds are extremely toxic (51) and,
hence, cannot be considered as potential sorbents.
With these factors in mind, the literature has been reviewed
to select possible candidate sorbents to be used in a vehicle sulfate
trap. There are several comprehensive references (8, 4, 21, 40, 44,
59, 60) in the literature in which a wide variety of potential sulfur
-------
- 99 -
dioxide sorbents have been examined. These studies were all aimed at
removing sulfur dioxide in the concentration range of 1 to 5 percent
from a stack gas. Although these studies are for a higher concentration
and a different oxide of sulfur, the activity results can be readily
extrapolated to the sorption of sulfur trioxide from automotive exhaust.
The major chemical classification of the sorbent compounds
investigated are the metal oxides. The sorbent material is envisioned
as being installed in the system either as the oxide or as the metal
which will readily oxidize in the oxidizing exhaust atmosphere. The
general sorption reaction can be written as
MeO + S03 (g) -> Me S04 (24)
where Me is a general symbol representing any sorbent metal. Lowell, et al,
(40) evaluated the oxides of 47 elements for potential use as sulfur
dioxide sorbents. Their thermodynamic calculations show that the formation
of sulfates for nearly all of the elements considered is favorable.
Pechkovsky (44) examined the rate of sulfation of several metal oxides
in powder form over a temperature range of 400-1000°C. He found that
calcium oxide was more active than magnesium oxide which was more active
than zinc oxide.
The Bureau of Mines (4) conducted a program looking at the
activity of several bulk oxides with respect to their ability to sorb
sulfur dioxide. Experimentally, a bed of sorbent, 8-24 mesh particles,
was exposed to a synthetic flue gas with 0.3 volume percent sulfur dioxide.
The space velocity was maintained at 1,050 v/v/hr and temperatures of 130
and 330°C were examined. Typical results showed that active sorption
materials would remove essentially all of the sulfur dioxide for a period
of time. As sulfation of the sorbent proceeded, a point was reached where
breakthrough would occur. From this point, the fraction of sulfur dioxide
removed was found to decrease linearly with time.
The sorbent materials were rank ordered with respect to the
amount of sulfur dioxide removed per unit mass of sorbent at the point
of 90% removal of sulfur dioxide. The most active sorbents in order of
activity were the oxides of manganese, cobalt, and copper.
The temperature study showed that for most sorbents approximately
twice as much sulfur dioxide had been sorbed up to the 90% breakthrough
point at 330°C than at 130°C. At the higher temperature, it was also
noted that the sulfur dioxide had made a significant penetration into the
particles of the more active sorbent materials.
-------
- 100 -
Vogel, €£ al_. , evaluated the activity of several metal oxides
supported on alumina. All test samples were made with the same metal
equivalents so the various materials could be directly rank ordered
as to activity. Samples of each supported sorbent were exposed to a
synthetic stack gas and the outlet was continuously analyzed for sulfur
dioxide. From these data, the sulfate loading of each sorbent was
determined for the conditions where the outlet sulfur dioxide concentration
was five percent of the inlet concentration. In addition, the maximum
loading was determined by extrapolation of the data.
The materials were then rank ordered as to the percent of
sorbent reacted at the five percent breakthrough point. This ranking
also agreed with the ranking based on percent of sorbent reacted at
maximum sulfation capacity with one exception. The one exception was
the Bureau of Mines alkalized alumina sorbent which was included in
this study as a benchmark. This material has been extensively studied
(4, 5, 43, 40) as a sulfur dioxide sorbent but would not be acceptable
as a vehicle trioxide sorbent due to its low sulfation capacity per unit
volume.
The activity of the sorbents in order of decreasing activity is:
the oxides of sodium, strontium, copper, calcium, and chromium. All of
these materials had in excess of 50% of the sorbent reacted at the 5%
breakthrough point. These were followed by the oxides of : barium, lead,
cadmium, manganese, magnesium, iron, cobalt, nickel, and zinc. Tin
and vanadium oxides showed no apparent activity. The alumina substrate
was also tested and found to be completely inactive. In general, these
results show that the alkali and alkaline earth metals exhibit highest
reactivities. In addition, copper and chromium showed good reactivity.
Welty (60) conducted a theoretical study of the reactivity of
potential sorbent cations based on a characterization factor consisting
of the cation radius, electronegativity, and valence. Using this factor,
the alkali metals are the most promising sorbents followed by the alkaline
earth metals, then by various transition metals. This reactivity scale
is in general accord with the experimental and theoretical works described
previously.
The second necessary property a sorbent must have is thermal
stability of the sulfated material. Although the temperature regime
the sorbent sees in vehicle use can to some extent be controlled by
location in the exhaust system, the sulfated material must have a
higher decomposition temperature than it is expected to experience.
Temperatures at or near the exit of the exhaust pipe can be as high as
800°C under sustained high speed driving of vehicles equipped with
oxidation catalysts. Therefore, the sulfated sorbent must have a
decomposition temperature at or above 800°C.
-------
- 101 -
Table A-9 lists the decomposition temperatures of several sulfated
sorbents. The temperature ranges and slight disagreement between refer-
ences are due to the difficulty in experimentally determining decomposition
temperature and in the different experimental methods employed. These
results show that the sulfated alkali and alkaline earth metals all show
acceptable decomposition temperatures. The transition metals, however,
show borderline temperatures, particularly aluminum oxide. If these ma-
terials are to be considered, their installation would have to be limited
to points as far from the oxidation catalysts as practical.
As estimated above, the sulfate trap must have the capacity to
react with 131 g moles of sulfur. Table A-10 lists the mass of typical
sorbents which would fulfill this requirement. In addition, the volume
of sorbent as calculated using the crystalline density is included. In
the cases where more than one crystal structure exists, an average density
was used. There are several important general conclusions which can be
arrived at from an examination of this table. Within a given group of the
periodic table, the required amount of the lower molecular weight sorbents
are lighter and of lower volume. For instance, the required mass and
volume of sodium oxide are 8.12 kg and 3.58 litres whereas the higher
molecular weight cesium oxide requires 36.9 kg and 8.69 litres.
Another important consideration is the valence of the sorbent
cation. For instance, two univalent cations are required for each sulfate
anion whereas only one divalent cation is. If these cations have approxi-
mately the same molecular weight, then the divalent one would be preferable.
An excellent example is the third period of the periodic table. The first
three members of this period, sodium, magnesium, and aluminum, have nearly
the same atomic weight but they are, respectively, univalent, divalent,
and trivalent. The required amount of sodium oxide is 8.12 kgm, of mag-
nesium oxide is 5.28 kgm, and of aluminum oxide is 4.45 kgm. The volumes
likewise decrease. Therefore, to minimize trap mass and volume, lower
cation molecular weight sorbents with higher valence states are preferred.
There is another problem which must be considered in the selection
of a sorbent material and in the engineering of the sorbent structure to
be installed in a vehicle. This problem is the increase in volume of the
sorbent as sulfur is picked up. The magnitude of this problem is shown
in Table A-10 where the volumes of the totally sulfated sorbents are given
along with the ratio of the sulfated volume to the fresh sorbent volume.
This volume increase is a result of two compounding factors. First, the
mass of the trap is continuously increasing due to the pickup of sulfur
trioxide and dioxide. Second, the crystalline density of the sulfate is
always less than the corresponding oxide.
-------
- 102 -
TABLE A-9
DECOMPOSITION TEMPERATURES OF SULFATED SORBENTS
Compound
Decomposition
Temperature
Determination
Method
Reference
Sodium
Potassium
Cesium
Magnesium
Calcium
Barium
Aluminum
Manganese
Iron
>800°C
>800°C
>800°C
750°C
890-972°C
1180
>1200°C
>800°C
f 652
/ 590-639°C
^650-950°C
699-790°C
880-1100°C
630
781-810°C
702-736°C
700-840°C
Thermodynamic Calculation
In air flow
Thermodynamic Calculation
In air flow
Vacuum
In air flow
Inert gas flow
In air flow
In air
Vacuum
Inert gas flow
In air flow
In air
40
40
40
60
3
60
3
40
58
3
40
3
40
58
40
3
40
-------
- 103 -
TABLE A-10
MASS AND VOLUME REQUIREMENTS OF
Sorb en t
Material
Na 0
V
Cs20
MgO
CaO
BaO
A12°3
MnO
Fe203
CuO
ZnO
MgC03
CaCO
BaCO
Na2C°3
K CO
Mass
Req'd., Kg
8.12
12.34
36.9
5.28
7.35
20.1
4.45
9.29
6.97
10.4
10.7
11.0
13.1
25.9
13.9
18.1
Vol.
Req'd., 1
3.58
5.32
8.69
1.48
2.21
3.51
1.12
1.70
1.33
1.63
1.90
3.73
4.65
5.84
5.48
7.46
Sulfated
Mass, Kg
18.6
22.8
47.4
15.8
17.8
30.6
14.9
19.8
17.5
20.9
21.1
15.8
17.8
30.6
18.6
22.8
VARIOUS SO
Sulfated
Vol., 1
6.94
8.58
11.2
5.93
6.46
6.79
5.51
6.09
5.64
5.80
5.97
5.93
6.46
6.79
6.94
8.58
Sulfated Volume,
Expansion Ratio
1.94
1.61
1.29
4.00
2.92
1.94
4.92
3.58
4.24
3.56
3.14
1.59
1.39
1.16
1.27
1.15
*Assuming 131 g moles of sulfur to be trapped.
-------
- 104 -
Unfortunately, the magnitude of this volume increase problem
increases for the more desirable (from initial volume and weight con-
siderations) sorbents. For instance, aluminum oxide has the lowest
volume required, yet has the highest volume increase ratio. On the other
extreme, cesium oxide has the largest volume requirement, yet the lowest
expansion ratio.
One means of minimizing this problem is to provide for this
expansion chemically by using a sorbent which gives up some mass as it
picks up the sulfur oxides. This exchange reaction must release a com-
pound which itself is not a deleterious emission. One class of ideal
exchange sorbents are the metal carbonates which release carbon dioxide
in exchange for sulfur oxide. Several potential carbonates are listed
in Table A-10 where their volume and mass requirements are shown. It is
immediately obvious that the volume and mass of the carbonate sorbents
is much greater than for the oxide, however, the volume expansion is dras-
tically reduced. Use of the carbonates would simplify the engineering
of the physical sorbent structure by eliminating the need for large volume
expansions. This simplification more than outweighs the initial increased
volume of the sorbent carbonates over the oxides.
There are two side reactions of the sorbent with exhaust gas
constituents which need to be considered in selecting potential sorbent
materials. The first side reaction is the reaction of the sorbent oxide
with carbon dioxide to form the sorbent carbonate. This reaction is
important for several reasons. If the carbonate is unreactive towards
sulfur trioxide, then formation of carbonate decreases the potential
sulfation capacity of the trap. If the sorbent carbonate is reactive,
it is possible that the reactivity will be lower than for the sorbent
oxide which would lower the reactivity of the trap. Another problem
is that the formation of carbonate causes a volume expansion. This
expansion within a sorbent particle causes the pore volume to decrease
which, in turn, can hinder diffusion of the sulfur trioxide into the
particle. Therefore, the potential amount of internal sulfation will
be decreased.
The best example of carbonation is the reaction of calcium
oxide. Thermodynamic calculations show (see for instance Ref. [46])
that the carbonation of calcium oxide is favorable in an exhaust gas
environment at temperatures below 760°C. Thermodynamic calculations
also show that the calcium carbonate formed will react with the oxides
of sulfur. Thus, the possibility of carbonate formation exists and will
be competitive with the sulfation. If the carbonate is formed, it can
also react with the sulfur oxides. However, the possibility exists that
this reaction will be significantly slower than the reaction with the
oxide which would lower the overall activity of the trap. There are
no rate data available which are applicable to exhaust gas concentra-
tions and temperatures. Therefore, the possible activity reduction
cannot be determined or estimated.
-------
- 105 -
TABLE A-ll
Sorbent
Material
Sodium
Potassium
Cesium
Magnesium
Calcium
Barium
Aluminum
Manganese
Iron
Copper
Zinc
:ES OF VARIOUS FRESH AND SULF.
•.TIES IN GRAMS PER 100 ML. OF
Oxide
Solubility
d.*
v.s.
v.s.
0.00062
0.131
3.48
i
i
i
i
0.00016
Carbonate
Solubility
7.1
112
260.5
0.0106
0.0015
0.002
0.0067
i
0.001
Sulfate
Solubility
S.
12
167
26
0.209
0.0002
31.3
52
si. s.
14.3
S
* d. - dissolves
v.s. - very soluble
S. - soluble
sl.s - slightly soluble
i - insoluble
-------
- 106 -
The amount of pore volume reduction due to volume increase
upon carbonation can be estimated from Table A-10. For the same equivalent
of calcium, the carbonate requires 2.10 times the volume of the oxide.
Therefore, if the initial particle of calcium oxide had a porosity of
50%, complete carbonation would fill all of the pore volume assuming all
expansion is internal and no reaction occurs with sulfur oxides.
These same problems are also present with essentially all
potential sorbent oxides. Since the desired type of sorbent reaction,
a basic sorbent with the acidic sulfur oxides, is the same as for
carbonation, potential sulfur oxide sorbents will also react with
carbon dioxide.
The second detrimental side reaction is that of either the
fresh or sulfated sorbent with water. The main problem associated
with the reaction of water is the volume expansion of the bed at low
temperatures such as during startup. Several of the sorbent oxides
can react with water to form hydroxides and most of the sulfated sorbents
are highly hydroscopic and form hydrates with as many as 18 molecules
of water per molecule of sulfated sorbent (i.e., Al2(S04)3'18 H20).
These added water molecules cause a volume increase which can plug pores
and even increase pressure drop across the catalyst bed. Since the water
of hydration is driven off even at very low temperatures, the problem
of volume increase would be applicable only at startup.
Both possible side reactions apply to almost all potential
sorbents. The degree to which these reactions would decrease the activity
and/or capacity of a sulfate sorbent is a very complex question which
cannot be answered without direct experimentation under exhaust gas
conditions. Since all potential sorbents could exhibit these problems
to some unknown degree, a selection of potential sorbents using the
criterion of deleterious side reactions is impossible.
Since the sulfate trap must operate in an environment where
contact with liquid water is highly probable at many times during the
lifetime of the trap, the fresh and sulfated sorbent should be insoluble.
Dissolution of the fresh sorbent would decrease the total sulfation
capacity of the trap. Dissolution of either or both the fresh and
sulfated sorbent would decrease the structural strength of the sorbent
particles possibly leading to increased attrition and/or channelling.
The solubilities of several potential sorbent oxides, carbonates
and sulfates are given in Table A-ll. These values were all obtained from
reference (1). The table shows that the oxides, carbonates, and sulfates
of the alkali metals are all quite soluble with the solubility increasing
with increasing molecular weight. The alkaline earth oxides show
more favorable solubilities. Magnesium and calcium oxides have
acceptable solubilities. Barium oxide is more soluble and could present
-------
- 107 -
some problems. The carbonates of the alkaline earth series all have
acceptable solubilities. The solubilities of the sulfates decreases
with increasing molecular weight. The solubility for magnesium sulfate
would be borderline while the higher molecular weight members would be
acceptable.
The oxides and carbonates of the transition metals are insol-
uble or, at worst, only slightly soluble. The sulfates are, however,
quite soluble with the exception of iron which is only slightly soluble.
Based on solubility restrictions, none of the alkali metals
would be acceptable. Calcium of the alkaline earth metals would be
acceptable, and iron of the transition metals listed would be acceptable.
Another consideration of somewhat lesser importance in select-
ing potential sorbents is the cost and availability of materials. The
real importance of cost and availability would come if two sorbents were
to show equal potential. However, for screening purposes, only those
materials which are in short supply or are very costly should be eliminated.
For reference, the costs of several potential sorbent materials
are listed in Table A-12. The cost listed is for enough material to trap
131 g moles of sulfur. All material costs were obtained from Reference
(9) and pertain to bulk quantities f.o.b. New York. The first cost
column assumes that the trap material can be prepared by some simple
mechanical treatment, e.g. pelletization. The costs for the transition
metals are based on the bulk metal rather than the oxides. The second
cost column was developed assuming that a soluble salt, in this case the
nitrate, would be required as the raw material in a precipitation procedure.
The first cost column shows that the least expensive sorbents
would be the oxides of the alkaline earth metals followed by the transition
metals. In all cases, the cost of the soluble nitrate salts were higher
than for the oxides. Also, the cost of preparing the sorbent from the
salt would be greater than from the oxide.
Ideally, the construction of the trap would preclude fresh or
sulfated sorbent attrition. However, in practice attrition and emission
of the trap material will occur. Therefore, emissions of the fresh and/or
sulfated sorbent cannot be harmful. Two examples immediately stand out:
beryllium and lead. Beryllium, the lowest molecular weight alkaline earth
metal, would be an ideal sorbent cation from many of the considerations
discussed above. However, beryllium oxide and sulfate are extremely toxic
(57) and, therefore, must be eliminated from consideration. The same is
true for lead. The lead oxide candle procedure for quantitative sulfur
dioxide measurement shows an excellent sorptive activity but lead oxide
and sulfate emissions are harmful. Hence, lead must also be eliminated
as a potential sorbent.
-------
- 108 -
Table A-12
ESTIMATED COSTS OF VARIOUS SORBENTS
Molecular Qf Sorbent Estimated Sorbent Cost as
Sorbent Weight per trap Sorbent Cost. $ Nitrate Salt. $
MgO 40 5.2 0.50 14.00
CaO 56 7.3 0.20*
BaO 153 20.0 6.80 11.00
Ni 59 7.7 27.00 23.00
Cu 64 8.4 10.00 23.00
Zn 65 8.5 6.50 7.70
Mn 55 7.2 5.20
* from limestone
In selecting potential sorbents based upon these considerations,
there is no single material which stands out in all categories. The one
cation which is consistently near the top in all categories is calcium.
This cation could be employed either as the oxide or as the carbonate.
Other potential sorbents which show promise are magnesium, manganese, and
aluminum. Again, the most probable sorbent would be the oxide of these
materials. The remaining potential sorbents have deficiencies in one or
more categories.
-------
- 109 -
A.7 References
1. Argo, W. B., and J. M. Smith, Ind. Eng. Chem., 45, 298 (1953)
2. Beltzer, M., R. J. Campion, and W. L. Peterson, SAE Paper 740286,
Detroit, Michigan, 1974
3. Bienstock, D., L. W. Brunn, E. M. Murphy, and H. E. Brown, U.S.
Bur. Mines Inform. Circ. 7836, 1958
4. Bienstock, D., J. M. Fields, and J. G. Myers, U.S. Bur. Mines Rep.
Invest., 5735, 1961
5. Bienstock, D., J. H. Fields, and J. G. Myers, J. Eng. Power, 86
353 (1964)
6. Bodenstein, M. and C. G. Fink, Z. Physic. Chem., 60, 1 (1907)
(C.A. 1-2849)
7. Bodenstein, M. and C. G. Fink, Z. Physik. Chem., 60, 46 (1907)
(C.A. 1-2850)
8. Boreskov, G. K., Zk. Fiz. Khim., 19, 535 (1945)
9. "Chemical Marketing Reporter", Schnell Publishing Company, Inc.,
July 15, 1974
10. The Chemical Rubber Co., "Handbook of Chemistry and Physics", 48th
Ed., 1967-68.
11. Cheslova, U. S. and G. K. Boreskov, Zhur. Fiz. Khim., 30, 2560
(1956) (C.A. 51-9276)
12. Cheslova, U.S. and G. K. Boreskov, Doklady Akad. Nauk 5.S.S.R., 85,
377 (1952) (C.A. 46-9961)
13. Chrysler Corp., Response to EPA Request for Sulfate Data, May 6, 1974
14. Duecker, W. W. and J. R. West, "The Manufacture of Sulfuric Acid",
Reinhold Publishing Corporation, New York, 1959
15. Emmet, P. H., ed. "Catalysis" Reinhold Publishing Corporation, New York,
1960
16. Ethyl Corporation, Automotive Sulfate Emissions, May 3, 1974
17. Ford Motor Company, Ford Response to EPA Request for Data on Automotive
Sulfate Emissions, May 7, 1974
18. General Motors, Response to the March 8, 1974 Federal Register
Regarding Automotive Sulfate Emissions: A Status Report, May 7, 1974
-------
- 110 -
19. Gillespie, G. R. and H. F. Johnstone, Chem. Eng. Prog., 51, 74-F
(1955)
20. Gmetro, J. D. and T. Vermuelen, A I Ch E J., 10, 940 (1964)
21. Haas, L. A., U.S. Bur. Mines Inform. Circ. I.C. 8608 (1973)
22. Hall, R. E. and J. M. Smith, Chem. Eng. Prog., 45, 459 (1949)
23. Harkins, J. H. and S. W. Nicksic, Environ. Sci. Techno1. ,
1, 751 (1967)
24. Heinen, C. M., SAE Paper 486J, Detroit, Michigan (1962)
25. Hougen, 0. A., and K. W. Watson, "Chemical Process Principles Part III:
Kinetics and Catalysis", John Wiley and Sons., Inc. New York (1957)
26. Hougen, 0. A. and K. W. Watson, Irid. Eng. Chem., 35, 529 (1943)
27. Hougen, 0. A. and C. R. Wilkie, Trans. Am. Inst. Chem. Engrs.,
45, 445 (1945)
28. Hurt, D. M., Ind. Eng. Chem., 35, 522 (1943)
29. Irvin, H. B., R. W. Olson, and J. M. Smith, Chem. Eng. Prog., 47,
287 (1951)
30. JANAF Thermochemical Data, Dow Chemical Co., Midland, Mich.
31. Kaneko, Y. and H. Odanaka, J. Res. Inst. Catalysis, Hokkaido Univ.,
13, 29 (1965)
32. Kaneko, Y. and H. Odanaka, J. Res. Inst. Catalysis, Hokkaido Univ., 14,
213 (1966)
33. Kelley, K. K., C. H. Shomate, F. E. Young, B. F. Naybor,
A. E. Salo, and E. H. Hoffman, U.S. Bur. Mines Tech. Paper, 685,
(1949)
34. Knietsch, R., Chem. Ber., 34, 4069 (1901)
35. Krause, H. H., A. Levy, and W. T. Reid, J. Eng. Pow., 90, 38 (1968)
36. Levy, A., E. L. Merryman, and W. T. Reid, Environ. Sci. Technol., 4,
653 (1970)
37. Lewis, W. K. and E. D. Ries, Ind. Eng. Chem., 17, 593 (1925)
38. Lewis, W. K. and E. D. Ries, Ind. Eng. Chem., 19, 830 (1927)
39. Lisle, E. S. and J. D. Sensenbaugh, Combustion, 36, 12 (1965)
40. Lowell, P.S., K. Schwitzgebel, T. B. Parsons, and. K. J. Sladek,
Ind. Eng. Chem. Process. Des. Develop., 10, 384 (1971)
41. Lange, G. , A. C. Cummings, and F. M. Miles, "The Manufacture of Acids
.and Alkalies, Vol. IV. The Manufacture of Sulfuric Acid (Contact
Process) D. Van Nostrand Co., Inc., New York, 1925.
-------
- Ill -
42. Olson, R. W., R. W. Schuler, and J. M. Smith, Chem. Eng. Prog., 46,
614 (1950)
43. Paige, J. I., J. W. Town, J. H. Russell, and H. J. Kelly,
Bur. Mines Rep. Invest. 7414 (1970)
44. Pechkovsky, V. V., J. Appl. Chem. (Russ.), 30, 1643 (1950)
45. Pierson, W. P., R. H. Hammerle, and J. T. Rummer, SAE Paper 740287,
Detroit, Michigan, Feb. 1974
46. Reid, W. T., J. Eng. Power, 92, 11 (1970)
47. Roiter, V. A., N. A. Stukanovskaya, G. P. Korneichuk, N. S. Volikovskaya,
and G. I. Golodets, Kin. i Katal., 1, 408 (1960)
48. Ross, L. W., Sulfur, 65, 37 (1966)
49. Russell, J. H., J. W. Town, and H. J. Kelly, U.S. Bur. Mines Rept.
Invest. 7415 (1970)
50. Salsas - Serra, F., Chem. et Ind., 37, 1056 (1937)
51. Sax, N. I., "Dangerous Properties of Industrial Materials" Reinhold
Book Corporation, New York, 1968
52. Schuler, R. W., V. P. Stallings, and J. M. Smith, Chem. Eng. Prog.
Symp. No. 4, 48, 19 (1952)
53. Sittig, M. , "Sulfuric Acid Manufacture and Effluent Control", Noyes
Data Corporation, 1971
54. Taylor, G. B. and S. Lenher, Z. Physic. Chem. Bodenstein-Festbend, 30
(1931) (C.A. 25-5341)
55. Uyehara, 0. A. and K. M. Watson, Ind. Eng. Chem., 35, 541 (1943)
56. Verhoff, F. H. and J. T. Banchero, Chem. Eng. Prog., 70, 71 (1974)
57. Vogel, R. F., B. R. Mitchell, and F. E. Massoth, Environ. Sci.
Technology 8, 432 (1974)
58. Warner, N. A. and T. R. Ingraham, Can. J. Chem., 38, 2196 (1960)
59. Warner, N. A. and T. R. Ingraham, Can. J. Chem. Eng., 40 , 263 (1962)
60. Welty, A. B., Hydrocarbon Processing, 104 (1971)
61. Weychert, S. and A. Urbanek, Inter. Chem. Eng., 9, 396 (1969)
-------
- 112 -
APPENDIX B
MEASUREMENT TECHNIQUES
B.I Gaseous Emissions
Gaseous emissions were measured using standard instrumentation
for Federal Emission Test Procedures. CO and CC-2 were analyzed by NDIR,
hydrocarbons by FID, and NOX by chemiluminescence.
B.2 Measurement of Sulfate Emissions
Sulfate emission samples were collected using Exxon Research's
exhaust particulate sampling system, and also by the Goks^yr-Ross tech-
nique. The sulfate content of samples collected by both methods was
determined by colorometric titration using Sulfanazo (III) as an indi-
cator. Each of these techniques is described in detail below.
B.2.1 Exhaust Particulate Sampling System
The exhaust particulate sampling system has been designed to
collect particulate matter at constant temperature during the 1972 and
1975 Federal Test Procedure, and 64 km/h (40 mph) cruise conditions.
This system is capable of frequent and convenient operation, and is
compatible with constant volume sampling (CVS) of auto exhaust. Com-
patibility is obtained because the particulate sampler requires only
a small portion of the diluted exhaust, the major portion of the sample
is available to the CVS system for the measurement of gaseous emissions.
Conditions used in the measurement of exhaust particulate conform to
those mandated by the Federal Test Procedures for gaseous emissions.
This sampling system uses a small tunnel which means that low
dilution ratios are used, allowing gaseous emissions such as CO, hydro-
carbons, NOX and S02 to be measured accurately. While low dilution
ratios are desirable from the standpoint of CVS gaseous emission meas-
urements, the collection of a proportional sample of particulate matter
at constant temperature 32°C (90°F) from a sample stream having a high
dew point without causing condensation of water requires an advanced
temperature control system.
B.2.1.1 Sampling System Components
The particulate sampler which has been discussed previously
(1,2) is shown schematically in Figure B-1. This system has five major
components:
1. A diluent air preparation system
2. A flow development tunnel
3. An exhaust injector system
4. An isokinetic sampling probe
5. A particulate measuring device, which in the case shown
is a 0.2 micron glass fiber filter
-------
FIGURE B-l
EXHAUST PARTICULATt SAMPLER
INTAKE
DILUENT AIR
DEHUMIDIFIER
PRE-COOLER
FILTER
BOX
COUPLED
MIXING-
BAFFLES
HEAT
EXCHANGE
AIR COOLED
COMPRESSOR
FLOW
DEVELOPMENT
TUNNEL
MIXING
TURBULATORS
TO CVS
PUMP
U)
I
ISOKINETIC
SAMPLING
PROBE
EXHAUST
INJECTOR
FILTER
HOUSING
-------
- 114 -
The overall function of this system is to allow the collection
of particulate matter from an isokinetically sampled portion of diluted
exhaust which has been cooled to 32°C by dilution with chilled, dehumid-
ified, filtered air. The function of each of the components in accom-
plishing this objective is described below.
B.2.1.2 Diluent Air Preparation System
This system consists of a dehumidifier, filter, coupled mixing
baffles, a cooling system, and mixing turbulators.
The dehumidif ier shown schematically in Figure B.2 minimizes
the possibility of condensation occurring in the sampling system during
a run, and is an integral part of the temperature control system. Dil-
uent air is dried by passage through a filter and a slowly rotating
desiccant wheel containing laminated flat and corrugated asbestos,
impregnated with a regenerable desiccant, LiCl. Dehumidification of
diluent air and desiccant reactivation are concurrent processes, so that
dehumidif ication can be carried out on a continuous basis. The dehumid-
if ier, a Honeycombe Model HC 750-EA is manufactured by Cargocaire Engin-
eering Corporation is described in their Bulletin No. 07169(3).
Dehumidified inlet air passes from the dehumidifier to a filter
box containing a paper filter, a bed of activated charcoal, and a second
paper filter. This assembly is the standard filter box assembly for the
Scott Research Constant Volume Sampler (CVS) unit. The filter assembly
removes the particulate matter present in the diluent air and reduces
and stabilizes the background hydrocarbon content of the diluent air.
Because regeneration of the desiccant is accomplished by
heating, the dehumidified air emerging from the drum is above ambient
temperature. A pre-cooler situated between the dehumidifier and the
CVS filter cools the dehumidified air stream down to ambient temperature
to take the additional cooling load imposed by the dehumidif ication step
away from the final cooling system. The pre-cooler consists of several
rows of coils through which chilled city water is passed.
The coupled mixing baffles continuously divide the dehumidi-
fied, filtered air into two portions, one which passes through the
cooling system, and a second portion which bypasses the cooling system.
The position of the mixing baffles is controlled by a rapid response,
deviation-type controller operating on an input signal from a thermo-
couple in the filter housing. The system is designed to maintain 32°C
at the filter housing, during the 1972 or 1975 Federal Test Procedures,
and 64 km/h cruise.
The controller operates by comparing an input signal from a
thermocouple in the filter housing with a set point signal, and takes
corrective action to either raise or lower the output signal until the
set point and thermocouple input are equal. The controller used was an
Electronic Control System Model 6700 Controller(4) . The output signal
-------
FIGURE B-2
SCHEMATIC OF DEHUMIDIFICATION SECTION
REACTIVATION AIR
INLET FILTER
WET AIR OUTLET
TO FILTER
BOX
DRYAiR
T AIR HEATER
REACTIVATION
FAN
DESICCANT
•*
REACTIVATION
SECTOR
DRY AIR FAN
HUMID AIR
INLET
FILTER
Ln
I
-------
- 116 -
from the controller is fed to an electric to pneumatic transducer (5)
which in turn activates a pneumatic controller (6) which operates the
coupled baffles.
The cooling system is an air cooled condensing evaporator which
has a cooling capacity of 8,300 kcal/hr. The evaporator is a Dunham-Bush,
Model SCO-50C unit (7) containing ten rows of custom-made cooling coils
(8).
The mixing turbulators insure that chilled air is thoroughly
mixed with the portion of air bypassing the cooling system before the
stream is used to dilute the vehicle exhaust. The turbulators consist
of six semi-circular perforated plates attached to a 1/2" diameter
wall tube at their centers, arranged in a helical series sequence along
the tunnel axis. This arrangement allows both longitudinal and lati-
tudinal mixing.
Maximum flow through the diluent air preparation system is
determined by the cooling capacity of the chiller. Presently, this
limitation is about 13 m^/min.
B.2.1.3 Flow Development Tunnel
The exhaust and diluent air are mixed and a uniform velocity
profile is developed in the flow development tunnel. The flow develop-
ment tunnel is a 2.23 m long section of a 4-inch diameter Schedule 5
(actual I.D. = 11.0 cm) stainless steel pipe. Neither the length nor
the diameter of the flow development tunnel have been optimized, but
as will be shown in another section, a dilution tunnel of these dimen-
sions is satisfactory for this purpose.
B.2.1.4 The Exhaust Injection System
The raw exhaust is mixed with the diluent air normally used
in the CVS in such a way as to completely mix the two in as short a
time interval as possible. This is accomplished by injecting the ex-
haust in a countercurrent direction to the diluent air stream. Previous
experiments (1) have shown this to be the most efficient way of obtain-
ing a rapidly mixed, uniformly distributed diluted exhaust sample stream.
Figure B-3 shows a schematic of the exhaust injector in the counter-
current position.
B.2.1.5 Isokinetic Probe
Isokinetic sampling is required to insure that the particulate
sampled is representative of the particulate in the main stream; that
is, the particulate concentration and size distribution in the probe
sample should correspond to that of the main stream. The probes are
designed so that the sample stream is divided into two parts with a
volume ratio equal to the ratio of the cross-sectional areas of the
openings of the sample probes and the tunnel cross-sectional area, that is
-------
FIGURE B-3
COUNTER CURRENT EXHAUST INJECTION SYSTEM
TRICLOVER
CONNECTIONS
RAW EXHAUST
DILUTION
AIR
TAILPIPE CONNECTOR
DILUTED EXHAUST
4.3" ID S.S. FLOW
DEVELOPMENT
SECTION
COUNTERCURRENT
INJECTOR
-------
- 118 -
Area (probe) _ Flow Rate in SCFM (probe)
Area (tunnel) Flow Rate in SCFM (tunnel)
Another problem to be considered in probe designs is minimizing
sample deposition in the probes. When suspended particulate matter
leaves the tunnel and enters the sampling probe, it is leaving a low
surface to volume region and entering a high surface to volume region.
Relative sample losses by impaction should be greater in the probe than
in the tunnel. Therefore, the probe should be as short and direct as
possible to minimize the residence time of the particulate matter in
the probe. The filter housing connected to the probe is flared out as
soon as physically possible to minimize the surface to volume ratio of
the housing and thereby reduce sampling losses by impaction in this
portion of the sampling system.
B.2.1.6 Particulate Collecting Stage
At present, particulates are collected by filtering the sample
through pre-weighed filters. In principle, other particulate collectors
such as impactors and other devices, could be utilized with the particulate
sampling system. In this paper, total particulates are determined
gravimetrically using Gelman Type E glass fiber filters which have an
effective porosity of 0.2 microns.
B.2.2 Exhaust Particulate Sampling System Performance
In order to function properly for sulfate emissions, the
exhaust particulate sampling system should have the following capabilities:
(1) mix exhaust and diluent air rapidly»
(2) Allow development of a uniform velocity profile in the
flow development tunnel,
(3) minimize sampling losses in the tunnel,
(4) give equivalent emission rates with parallel filters, and
(5) maintain constant temperature at the particulate
collecting stage.
All of the above have been adequately documented (1,2) and will be
reviewed in this section.
B.2.2.1 Rapid Mixing of Exhaust and Diluent Air
Three methods of injecting exhaust into diluent air were
tested: co-current flow, perpendicular flow, and countercurrent flow.
In each case, the exhaust was injected through a 5.1 cm O.D. x 0.089 cm
wall stainless steel tube into the flow development tunnel. The efficiency
of the three injection methods was tested by measuring hydrocarbon
-------
- 119 -
concentrations in the diluted exhaust at a point approximately 2.23 m
downstream of the injection point. Hydrocarbons were chosen as the
tracer because they are easier to measure than particulates. If the gaseous
components of the exhaust are not evenly distributed over the flow
cross-section, there is no reason to believe that the particulates will
be well distributed. The ultimate test of uniformity of particulate
distribution in the tunnel is the consistent attainment of equivalent
particulate emission rates with parallel filters. The results showed
that uniform distribution was obtained only by countercurrent injection.
B.2.2.2 Development of Uniform Flow
in Flow Development Tunnel
To insure that samples taken at any point in the tunnel cross-
section will contain the same amount of particulate material, a uniform
radial distribution of particulate material in the tunnel must be obtained.
The small size of the tunnel would make it difficult to obtain reliable
measurements of velocity profiles. However, it is well known that the
higher the Reynolds Number of turbulent flow, the flatter the velocity
profile (9). However, over the range of interest for this system, the
effect of this flattening of the velocity profile is negligible. Con-
sider the system as having a flow of air at 32°C through a 11 cm
diameter pipe.
%e = ^ (1)
where NRe = Reynolds Number
D = pipe diameter = 11.0 cm = 0.110 m
U = average fluid velocity = ^ ^^g 1370 m/min
8.21 x 10A m/hr.
p = density = 11.14 kg/m3
y = fluid viscosity = 0.186 cp = 0.670 kg/m-hr.
- „ ^ m/hr) (1.14 kg/m3)
0.670 kg/m-hr = "^O00
N =(0.110 m) (8.2 x 10 m/hr) (1.14 kg/m)
- -
Equation (1) shows that the Reynolds Number varies inversely with diameter
for constant volumetric flow. Therefore, decreasing pipe diameter to 2.5 cm
would increase N^e to 615,000 while increasing pipe diameter to 40 cm
would decrease NRe to 40,000.
-------
- 120 -
One measure of the flatness of the velocity profile is the
ratio of the mean gas velocity to the maximum gas velocity. It has been
shown experimentally that for turbulent flow in smooth pipes (9)
( \
if -(l)
ao,
where u = point velocity
U = maximum velocity at center
Y = distance from the wall
R = pipe radius, and
N = a constant depending on Reynolds Number
Schlicting (9) shows that average velocity u is,
- 2N2
(N+l) (2N+1) (3)
The following table shows that the effect of changing pipe diameter over
a large range would be negligible.
Effect of Reynolds Number
on Velocity Profile
N u/U
23,000 6.6 .807
110,000 7.0 .816
500,000 8.0 .837
Another important factor in choosing the diameter of the
flow development tunnel is its effect on the length of the tunnel and
the diameter of the probes. As a general rule, ten pipe diameters are
usually sufficient to develop a fully turbulent velocity profile. The
larger the diameter, the longer the tunnel required and the longer the
residence time in the flow development section. Longer .residence time
leads to higher particulate settling and greater inaccuracy in the
measurement. Therefore, the tunnel diameter should be minimized. How-
ever, as tunnel diameter decreases, the pressure drop through the tunnel
increases and the size of the probes needed for isokinetic sampling
decreases. The problems caused by high pressure drop are obvious.
Smaller diameter probes should be avoided since they provide higher
surface to volume ratios and result in more loss of particulate by
impaction. The 11 cm diameter pipe in use offers a reasonable compromise
between these various factors.
-------
- 121 -
B.2.2.3 Tunnel Sampling Losses
Particulate deposition in the flow development section was
measured by introducing an artifically produced mono-disperse (3.5 micron
diameter) methylene blue aerosol into the exhaust injector in the same
manner as for auto exhaust. The system was disassembled after the run,
the tunnel surface washed with methanol and the washings analyzed
spectrophotometrically. The sensitivity of the method for methylene
blue is in ppb range. Analysis showed that tunnel losses are small,
amounting to less than 1% of the total aerosol introduced. No dye was
detected in the tunnel section housing the exhaust injector. About
0.1% of the aerosol was deposited in the tunnel mid-section, and about
0.3% was deposited in the tunnel section housing the probes. Independent
tests by U. S. Environmental Protection Agency workers with a tunnel of
similar dimensions have confirmed our results regarding tunnel sampling
losses (10).
B.2.2.4 Equivalent Emission Rates with Parallel Filters ,
Since only a small fraction of the diluted exhaust is sampled
for the particulate analysis, at least two parallel probes coupled to
the appropriate filters are needed to serve as internal checks on the
sampling system. One method of determining whether proper sampling is
achieved relies on the ratio of the weight of particulate collected
(W/0 by filter A, and the volume flow rate (FA) through probe A. This
ratio should equal the corresponding ratio of these parameters for
filter B and probe B, that is:
^ = ^ = ^ = 'etc.
FA FB FC
The particulate emission rates in grams/kilometer (g/km)
should be the same for all filters in a given run since
WA FP WB Fp
— = — = etc.
km FA Akm FB Akm
where Fp is the volume flow rate through the tunnel and Akm the distance
in kilometers accumulated on the particular test procedure.
Excellent agreement between parallel filters has been obtained
using this sampling system with conventional and catalyst-equipped
vehicles operating on a variety of unleaded fuels under cyclic and state
test conditions. Partial documentation of this agreement has been
previously described (1,2).
-------
- 122 -
8.2.2.5 Temperature Maintenance of the
Particulate Collection Stage
The dehumidifier is a key component of the temperature control
system, particularly since the sampling system is one in which the
air/exhaust dilution ratios are low, unlike other particulate sampling
systems (11, 12). This means that the relative humidity of the diluent
air is a key parameter. For example, during the steep acceleration
portion of the Federal Test Procedure, the exhaust volume flow rate from
a vehicle equipped with a 350 CID V-8 engine may be as high as 3.4 m^/min.
This means that dilution ratio would drop below 4 in the sampling tunnel.
If the relative humidity of the diluent air was high, attempts to control
the filter temperature at 32°C would result in condensation of water
vapor, with the associated loss of particulate matter. This is shown in
Figure B-4 which depicts mixture dew point-dilution ratio dependence as
a function of the relative humidity of the diluent (24°C) air.
The key role of the dehumidifier can readily be seen. If the
relative humidity of the diluent air is low, e.g., below 50%, it should
be possible to maintain a 32°C filter temperature without condensation
occurring. It can also be readily seen that in the absence of the de-
humidifier, on humid days, the dew point of the mixture would exceed
32°C at dilution ratios as high as four, so that condensation would
invariably occur during the acceleration portions of the driving cycle.
Attempts to control filter temperatures by omitting the
dehumidification step and chilling the diluent air would be difficult
to accomplish since water condensing on the coils would feed back latent
heat, decreasing the efficiency of the cooler. Continued running would
probably result in the condensed water freezing on the heat exchangers,
rendering them inoperative.
Figure B-5 shows a typical trace of the relative humidity of
the diluted exhaust in the vicinity of the sampling probes during FTP
operation with a catalyst-equipped vehicle. This trace is obtained
by withdrawing a sample just above the sampling probes and filtering
it prior to monitoring the humidity. Filtration is necessary in order
to protect the rapidly responding humidity sensor (13) .
It is evident that the relative humidity in the region of the
probe closely parallels the changes in the cycle driving patterns. At
no point in the driving cycle does the relative humidity at the probes
reach the saturation level at 32°C. As the relative humidity is lower
upstream, condensation in the tunnel upstream of the probe does not occur.
Figure B-6 shows the temperature-time trace at the filter during
the above run which is a typical case. A total flow rate of about
13 m3/min. was used. A 1.2 m long x 5 cm I.D. finned tube between the
tailpipe and the exhaust injector was needed to suppress temperature spikes
-------
FIGURE B-4
60 -
o° 50
I
£ 40
3
30 -
o
P-.
o 20 -
10 -
1.0
2.0
DEW POINT OF DILUTED EXHAUST VS.
AIR/EXHAUST DILUTION RATIO
At Indicated Relative Humidities
of 24*C Dilution Air
1 3.4 m /min @ Accel to 93 km/h
(100% RH)
( 80% RH)
3.0 4.0 5.0 6.0 7.0
Dilution Ratio (Dilution Air/Raw Exhaust)
8.0
9.0
10.0
-------
FIGURE B-5
RELATIVE HUMIDITY OF EXHAUST DILUTION AIR
MIXTURE AT VICINITY OF SAMPLING PROBES DURING
THE 1972 FEDERAL TEST-DRIVING CYCLE
100
.c
a
6
0
LI!
LU
0.
60
30
I
M
NJ
I
4 5 6 7 8 9 10
TIME. HUNDREDS OF SECONDS
11 12 13
-------
- 125 -
FIGURE B-6
TEMPERATURE CONTROL SYSTEM PERFORMANCE
CATALYST-EQUIPPED CAR
o
o
ta
J3
S
i
H
H
iJ
M
to
50
40 _
30 _
20 —
10 —
— 1
13 m3/MIN CHILLED AIR
SYSTEM WITH FINNED TUBE
r-v^*-^-— --'•*-— >^,
^^^^^r
r~^
' TIME DURING THE 1972 FTP
-------
- 126 -
above 32°C during the steep acceleration portion of the driving cycle.
It should be noted that the system is designed to prevent temperature
excursions above 32°C, not to maintain that temperature during the
course of the entire run.
Figure B-7 is a typical relative humidity-time trace for a
64 km/h steady state cruise experiment. The relative humidity surges
to about 25% on start up and slowly decreases with running time.
Complete temperature control can be obtained at 64 km/h
encapsulating the finned tube in a 10 cm diameter metal cylinder through
which ambient air is pumped in a countercurrent direction to the flow
of raw exhaust. Figure 6-8 shows a schematic of this additional
temperature control feature. Encapsulating the finned tube is not necessary
for the driving cycle. At the 64 km/h cruise, however, the temperature
would slowly rise above 32°C after about 20 minutes if the finned tube
was not encapsulated.
B.2.3 The Goks^yr-Ross Technique
This technique (14) involves passing a sample of exhaust through
a condenser maintained between 60 and 90°C. This is above the dew point
of the H20 in the exhaust, but below the dew point of the H2S04 in the
exhaust. As a result, H2S04 is condensed in the coil, but H20 which might
catch S02, is not condensed.
In our configuration, a. modified 300 mm Graham Condenser, equipped
with a 60 mm diameter medium porosity fritted glass filter, is used.
This unit has been found to give better than 95% collection efficiency at
flow rates up to 4 liters/min. Precautions must be taken, however, with
regard to the temperature of the incoming gas. If it is too hot, the coil
will not provide sufficient cooling to condense all the acid. If the gas
is too cool, some may condense out upstream of the coil. The sample lines
leading to the condenser must also be chosen with care, to avoid possible
reaction with sulfate. Glass or quartz lines are best, but if this is
impossible, and mechanical considerations demand the us>e of metal lines,
only tubing which has been passivated by prior exposure to 1^804 containing
gases should be used.
After sampling, the condenser coil and glass frit are emptied
and rinsed thoroughly. The total liquid is diluted to a standard volume
and analyzed for sulfate by the Sulfanazo III method.
B.2.4 Analytical Determination of Sulfate
Sulfate in samples collected by either the exhaust particulate
sampler or the Goks^yr-Ross technique is determined by a titrimetric pro-
cedure using a color indicator, Sulfanazo III [4,5 di-hydroxy-3, 6-bis
(0-sulphaphenylazo)-27, naphthalene-disulfonic acid]. The procedure
-------
FIGUBE B-7
RELATIVE HUMIDITY OF EXHAUST DILUTION
AIR MIXTURE AT VICINITY OF SAMPLING
PROBES DURING 64 Wh CRUISE CONDITIONS
at 32°C
100,
ULJ
,w
^00
ui
cc
10
15
TIME (MINUTES)
-------
- 128 -
FIGURE B-8
FINNED TUBE COOLING SETUP
Raw Exhaust
Outlet
nun 11111 I I I ill
MINI! I M || I M ||
To Exhaust
Injector
Cooling Fins
Air Inlet
-------
- 129 -
has been adapted from that of Budesinsky and Krumlova (15).
In the case of the samples from the exhaust particulate sampler,
soluble sulfate is leached from the filter with dilute nitric acid. The
leach solution is heated to boiling to drive off excess nitric acid,
filtered to remove insoluble material, passed through an ion-exchange
mliimn to remove interfering cations, and then buffered with methenamine
to a pH of 3-4. The resulting solution is titrated with barium
perchlorate using Sulfonazo (III) as the indicator. Washings from the
Goks^yr-Ross coil are treated in similar fashion starting with the ion
exchange step.
B.2.4.1 Reagents
The reagents used are as follows:
1. Barium perchlorate standard solution 0.01N
2. Barium perchlorate standard solution 0.001N
3. Hexamethylenetetramine (5% aqueous)
4. Sulfonazo III indicator (0.1 g/100 ml H20)
5. Ethyl Alcohol, absolute
6. Acetone
7. Nitric acid (2% aqueous)
8. Dowex 50 W-X8 cation exchange resin (50-100 mesh)
B.2.4.2 Titration Apparatus
The following apparatus was used in the titrations:
1. Ion exchange column, 1 cm x 25 cm.
2. Burettes (at least 0.05 ml div)
3. Low range pH paper - J. T. Baker Dual-Tint pH 1.0-4.3.
B.2.4.3 Standardization of Ba(0104)2 Solution
The barium perchlorate solution is standardized by titration
against previously standardized 0.01N sulfuric acid as follows:
(1) 5 mis of 0.01N sulfuric acid is pipetted into a 125-ml
Erlenmeyer flask.
(2) 45 mis of deionized water and 2.5 mis 2% nitric: acid
are added.
(3) Adjust the pH as described in the Procedure and titrate
with barium perchlorate solution.
(4) Calculate normality.
N Ba (C104)2 = mis H2S04 x N H2S04
mis Ba (€104)2
-------
- 130 -
B.2.4.4 Detailed Titration Procedure
(1) Cut and place 1/4 section of filter in 100-ml beaker.
(1/2 filter can be used for lower levels of 804).
(2) Add 3.0 mis of 2% nitric acid, wetting the filter section
completely.
(3) Add 20 mis of deionized water.
(4) Place small glass stirring rod in beaker and cover with
watch glass.
(5) Digest samples on hot plate and allow to boil for 5 minutes.
(6) Cool beakers and filter liquid through glass wool into
the ion-exchange column containing ^7.5 g of Dowex 50 W-X8 cation exchange
resin, collecting sample in 125 ml Erlenmeyer flasks.
(7) Wash beaker with 3 x 5 ml portions of deionized water. Add
washings to column. After each washing, squeeze liquid from the filter
by pressing it against the wall of the beaker with the stirring rod.
(8) Place the flask on a hot plate and evaporate to approximately
5 mis.
(9) Adjust the pH of the remaining solution to pH 4 using 5%
hexamethylenetetramine.
(10) Add 10 mis ethyl alcohol and 25 mis acetone to the flask.
(11) Add 3 drops Sulfonazo III indicator.
(12) Titrate with 0.01N barium perchlorate solution. (If
sulfate content is low, use 0.001N barium perchlorate.) If the sample
requires more than 10 mis of barium perchlorate to reach the end point,
the results are unreliable and should be discarded. Another portion
of filter should be treated as in steps 1-7. Solution should be collected
in 100 ml volumetric flask. An appropriate size aliquot is taken and the
test is continued with step 8.
B.2.4.5 Effect of Nitric Acid on Measurement of Sulfate
Because nitric acid is used as the leaching agent, several
experiments to ascertain the effect, if any, of nitric acid on the deter-
mination of sulfate were carried out. The test matrix, shown in the
-------
- 131 -
following table, shows that there is no effect of nitric acid on the
titration.
Table B-l
Comparative Titrations of Sulfate of
Samples With and Without Nitric Acid
ml 0.012 N Ba(C104>2
Contains to Titrate
Sample HN03 Theory Actual Ami
Na2SC>4, H2S04 Yes 5.0 4.9 -0.1
Na2S04, H2S04 No 5.0 4.7 -0.3
H2S04 Yes 5.0 4.7 -0.2
H2S04 No 5.0 4.8 -0.2
B.2.4.6 Precautions About Titration Procedure
The above procedure was arrived at after experimental work
identifying .sources of error and the means to minimize or eliminate
these errors was worked out. Reliable results can be obtained if the
following precautionary measures are taken:
(1) Keep HNO-j at minimum
(2) Keep water in sample at minimum
(3) pH is very critical (^4.0)
(4) Use blank for color comparison of end point
(5) Change Dowex frequently (approx. 3 samples)
(6) If sample is basic - adjust with nitric acid
For example, it was shown that not only is the ion-exchange
step necessary to eliminate interference, but that it is necessary to
insure that the capacity of the ion-exchange bed is not close to
exhaustion. It was also shown that it is necessary to use a new batch
of ion-exchange resin rather than a regenerated batch. Positive devi-
ations were obtained when titrating the effluent from a regenerated ion-
exchange bed. The results of the tests described above are shown in
Table B-2.
-------
- 132 -
Mis. H?S04
Blank
0.5
0.75
.00
,50
2.00
2.00
1.50
1.00
Blank
0.2
0.5
0.1
0.8
1.0
2.0
Blank
.1
.1
0.
0.
0.2
0.2
2.0
2.0
TABLE B-2
Experiments Demonstrating Influence of
Ion Exchanger on Sulfonazo III End Points
Mis. Titration
Theory
0.0
0.5
0.75
00
50
2.00
2.00
1.50
1.00
0
.0
.0
.0
0.8
1.0
2.0
0
,0
,0
,0
2.0
2.0
2.0
Actual
0.
0.
0.65
1.0
1.25
1.45
2.2
1.75
1.1
0.15
2.
4.
1.
0.
1.
10
95
45
80
00
1.90
0.25
,10
,00
,90
,00
,30
1.40
Ami
+0.1
0
-0.1
0
-0.25
-0.55
+0.20
+0.25
+0.10
+0.15
+0.10
-0.05
+0.45
0
0
-0.10
+0.25
+1.10
+1.00
+0.90
+1.00
-0.7
-0.6
Same column for
all samples
0.012 N Ba(C104)2
New Dowex each
sample
0.012 N Ba(C104)2
0.0012 N Ba(C104>2
New Dowex
0.012 N Ba(C104>2
Regenerate Dowex with
1:5 HC1
0.0012 N Ba(C104>2
0.012 N Ba (C104)2
As shown above, the Dowex ion exchanger has to be changed
frequently, and that large errors result if one attempts to work with a
presumably regenerated ion exchanger. Satisfactory results can be obtained
if the Dowex resin is changed after every two to three determinations.
B.2.4.7 Sulfate Determinations on Glass Fiber Filters
Spiked with Known Quantities of H2S04
A series of filters were spiked with known quantities of H^SO^
using 1.000 N H2S04 and a 5 yl syringe and with 0.0112 N H2S04 solutions.
The filters were leached with water, the leach solution worked up for
-------
- 133 -
analysis as discussed in Section B.2.4.4. Figure B-9 shows a plot of
added sulfate versus sulfate recovered by titration. It can readily
be seen that analysis of sulfate on the filters is as reliable as
analysis of solutions containing known quantities of sulfate. The next
step was to determine if this method is workable with filters containing
actual auto exhaust since the presence of organic exhaust particulate
matter on these filters could possibly introduce substances which
interfere with the color change of the titrimetric procedure. To deter-
mine if such interferences existed, the procedure was then checked against
the gravimetric determination of sulfate on parallel filters from
actual vehicle test runs since the presence of organic materials would
have no effect on the latter determination.
B.2.4.8 Comparison of Titrimetric and Gravimetric
Procedures on Filters from Vehicle Tests
Comparisons were then made between the titrimetric and
gravimetric analytical procedures on particulate filters from actual
vehicle runs. The titrimetric analysis was carried out on quarter sec-
tions of the filters, the gravimetric on an entire parallel filter.
The particulate filters were generated from vehicles equipped with
palletized and monolithic catalysts, operating on fuels of sulfur
levels ranging from 0.004 to 0.14 wt %. Both cyclic and steady state
test modes were used. The results are shown in Table B-3.
Table B-3
Comparison of Titrimetric and Gravimetric
SO^ = Analyses on Parallel Filters
(Actual Vehicle Runs, Sample Data)
Fuel Catalyst S0,= Emissions, g/mi.
Run Type Sulfur. % Type Titrimetric Gravimetric
40 mph, 1/2 hr. 0.14 Pelletized 0.036 0.040
1975 FTP
1975 FTP
0.012 0.011
0.0099 0.0081
40 mph, 1 hr. 0.004 V 0.004 0.004
1975 FTP 0.14 Monolith 0.306 0.287
0.296 0.288
0.176 0.173
.004 0.053 0.061
0.023 0.020
60 mph, 20 min. 0.14
0.004
0.029 0.020
0.259 0.239
0.294 0.253
0.265 0.247
0.009 0.007
0.011 0.008
0.011 0.008
Regression analysis on 26 comparison sets showed:
Titrimetric S04= = 1.029 Gravimetric S0^~ + 0.00085
Standard Deviation =0.01
-------
- 134 -
Figure. B-9
Recovery of Sulfate From
Spiked Glass Fiber Filter Samples
10.0
9.0
8.0
"S 7.0
% 6.0
4J
rH
•rl
0)
4J
CD
CO
00
5.0
4.0
3.0
2.0
1.0
0
Ideal Sulfate
Recovery Line
2345678
y gms Sulfate/Filter by Titration
10
-------
- 135 -
B.3 Measurement of SO2 Emissions
SO2 in diluted exhaust was measured using a Thermo Electron
Corporation flECO Model 40) Sulfur Dioxide Analyzer (16). SC>2 in raw
exhaust was measured using a hydrogen peroxide bubbler method. Each
of these techniques is described below.
B. 3.1 The TECO Sulfur Dioxide Analyzer
This device operates on a pulsed-fluorescence UV absorption
principle as follows. A gas sample is submitted to a source of pulsed
ultraviolet light through a monochromatic filter. Sulfur dioxide
molecules energized to an excited state by the high intensity light
source, return to the ground state by emitting a monochromatic light,
which passes through a narrow-band filter, and impinges upon the light
sensitive surface of a photomultiplier tube. The intensity of this
radiation is directly proportional to the SO2 concentration.
This measurement method was chosen because of the following
reasons:
(1) It is more convenient than wet chemical, conductimetric, or
coulometric methods.
(2) Both continuous and integrated measurements of SOo in diluted
exhaust can be made.
(3) Measurement of SO2 emission rates could be incorporated as part
of the gaseous emission measurements routinely obtained using
diluted exhaust collected by the CVS system.
The operating principle of the TECO analyzer is depicted schematically
in Figure B-10.
To eliminate the possibility of water adsorbing and condensing
in the sample chamber on the walls and optical filters, the water in
the sample stream was removed upstream to the TECO analyzer. Initially,
Drierite was used to remove the water but it was found that at the 5
to 10 ppm level of S02, the Drierite removed all the S0_.
The water removal problem was solved by using the Permatube
Drying System (17) shown in Figure B-ll. This system dries the sample
stream by passing it through a bundle of tubes which are permeable
to water but essentially impermeable to S02- Water is purged by
countercurrent flow of dry air or nitrogen. The effectiveness of the
permeable system in reducing the water level of a humid sample stream
below 10 ppm lUO while retaining the SO- in the sample has been
established by our Analytical Division. The Model PD-500-72 Perma Pure
Dryer according to the manufacturer (17) has the capability of taking a
feed having a 50°C dew point and extracting sufficient water at 150 litres/hr.
feed rate to reduce the dew point of the effluent to -30°C.
-------
Figure B-10 Principle of Operation
TECO S02 Instrument
\_7
PULSATING
ULTRAVIOLET
LIGHT
SAMPLE
GAS OUT
SAMPLE GAS CONTAINING S02 IN
CO
cr>
PHOTOMULTIPLIER
TUBE
ELECTRONICS
-------
Figure B-ll
Perinatube Drying System
WET FEED
INLET
LOW PRESSURE WET PURGE GAS OUTLET
SHELL HEADER
DRY PRODUCT
OUTLET
HEADER L PERMEABLE TUBE BUNDLE LOW PRESSURE
DRY PURGE GAS
INLET
i
H
LO
I
DRY
N2
-------
- 138 -
Our own tests with dry S02 in N2 and in air have shown
that no differences in TECO readings were obtained when the sample is
introduced directly into the analyzer, or when it passes through
the dryer prior to entry into the analyzer.
A millipore filter is used upstream to the dryer to prevent
any particulate matter from entering and eventually clogging the
dryer, and for that matter, from possibly entering and contaminating
the analyzer.
It has been found that CC>2, CO, and 02 are strong quenching
agents, while N2 exhibits a negligible quenching effect. The
Instrument reponse is therefore sensitive to background gas composition.
Absolute values of S09 concentrations necessitate calibration of the
instrument in a background representative of the sample to be analyzed.
For example, prior to a laboratory study on S02 conversion as a function
of oxygen concentration, it was necessary to assess the effect of oxygen
quenching (18). Various samples were made by preparing bell jar mixtures
containing 30 ppm S02, 12% C02, varied amounts of oxygen, and nitrogen as
the balance gas. Measurements of the 802 concentration of these mixtures
indicate an approximate 1 ppm reduction in instrument 802 response for
each 2% increment in oxygen concentration, as shown below.
Table B-4
S0« Measurements at Indicated
Oxygen Concentrations(a)
Oxygen TECO S02
Concentration Response
(%) (ppm)
0 29.7
1 28.8
2 28.4
4 26.8
6 26.2
(a) Basic Mixture, 30 ppm S02, 12% C02, balance N2.
The quenching effects of CO and C0« was also measured using
mixtures of 30 ppm S02 in pure CO and C02. In C02, the response to
30 ppm of S02 was 1.7 ppm, that in CO only 8.5. Measurements were then
carried out on bell jar mixtures of 30 ppm of S02 in background air
containing different concentrations of CO, 0., C-jHg, and C02. The
results given below show that instrument response is sensitive to overall
concentration of quenching species. Therefore data obtained where the
total background concentration of quenching species changes significantly
-------
- 139 -
Table B-5
Composite Effects of C02, 02 and
CO on TECO SOp Response
Total Quencher
Mixture Species Concentration (%) Instrument
Composition CC02] + (&2\ + (CQ3 Response (ppm)
30 ppm SO2
1.42% 0.
0.09% CO
0.051% H2 14.0 28
445 ppm C3Hg
12.5% C02
30 ppm S02
4.78% 02
14.3% C02 23.4 23
4.33% CO
348 ppm C3Hg
from the calibration gas quencher level should be corrected for the
inherent changes in instrument response.
For work with dilute exhaust, quenching effects are minor.
Quenching by exhaust CO- and CO is negligible for a CVS air diluted
sample due to dilution. Although the oxygen concentration increases with
dilution, it presents a reasonably constant quench background and can
be taken into account by calibration of dilute S02 in air mixtures.
Properly used, the precision of this instrument is about 0.5 ppm S02.
B.3.2 The Peroxide Bubbler Method
In this technique the gas sample is first passed throueh an
appropriate filter to remove any sulfate, then into a bubbler containing
80 ml. of a 3% H202 solution. The H202 oxidizes S02 to H2S04, which
is retained in the bubbler. The bubbler solution and washings are
analyzed for sulfate using the Sulfanazo III technique.
-------
- 140 -
B.4 References Used Appendix B
1. M. Beltzer, R. J. Campion, and W. L. Petersen, "Measurement of
Vehicle Particulate Emissions," SAE Paper 740286, February-March,
1974, Detroit, Michigan.
2. M. Beltzer, National Institute of Environmental Health Sciences.
3. Bulletin No. 07169, "HoneyCombe Industrial Dehumidifiers,"
HoneyCombe Industrial Division, Cargocaire Engineering Corporation,
Amesbury, Massachusetts.
4. Electronic Control Systems, Fairmont, West Virginia.
5. Electric Pneumatic Transducer, Model No. T5129, Fairchild
Industrial Products Division, Winston-Salem, North Carolina.
6. Pneumatic Controller, Model and Size B-51XC4, Conoflow Corporation,
Blackwood, New Jersey.
7. Dunham-Bush Corporation, West Hartford, Connecticut.
8. Coolenheat Incorporated, Linden, New Jersey.
9. H. Schlicting, "Boundary Layer Theory," New York, McGraw Hill Book
Co., Inc., pp. 504-5 (1960).
10. R. L. Bradow and J. B. Moran, "Sulfate Emissions from Catalyst
Cars-A Review," SAE Paper 750090, February 1975.
11. K. Habibi, Env. Sci. and Technol, 4, 239 (1970).
12. J. B. Moran and 0. J. Manary, Interim Report PB 196783, "Effect
of Fuel Additives on the Chemical and Physical Characteristics of
Particle Emissions in Automotive Exhaust," NAPCA, July 1970.
13. Instruments for Measurement and Control of Relative -Humidity,
Brochure B-ll and Form D-ll, Phys-Chemical Research Corporation,
New York.
14. H. Goksyr and K. Ross, J. Inst. Fuel, 35:177 (1962).
15. B. Budesinsky and L. Krumlova, Analytica Chimica Acta, 39, 375
(1967.)
16. Sulfur Dioxide Pulsed Fluorescent Gas Analyzer Model 40, Thermo
Electron Corporation, Waltham, Massachusetts.
17. Instruction and Operation Manual PD 101, Perma Pure Dryer, Perma
Pure Products, Inc., Oceanport, New Jersey.
-------
- 141 -
APPENDIX C
PRACTICAL OPERATING CONSIDERATIONS WITH SULFATE
AND S02 ANALYTICAL METHODS
C.I Cross-Checks of Analytical Techniques
A number of analytical techniques, described in Appendix B,
have been used in connection with this contract. Four basic methods are
involved, as shown in Table C-l. Since there are two methods each for
Table C-l
Characteristics of Analytical Methods
Exhaust
Species Method Gas Driving Modes Sample Source
Filter diluted cyclic, constant speed flowing gas
Sulfate
Goks^yr-Ross undiluted constant speed flowing gas
TECO diluted cyclic, constant speed flowing gas, bag
S02
Bubbler undiluted constant speed flowing gas
sulfate and for S02 measurement, a number of attempts were made to cross-
check between the two. Some of these attempts yielded satisfactory results,
while others did not. In attempting to understand the reasons behind the
failures, limitations in the capabilities of some of these techniques
under certain operating conditions were uncovered. These will be discussed
in the following sections.
C.I.I Cross-Checks of Sulfate Analytical Techniques
C.I.1.1 Comparison of Goks^yr-Ross and Filter Methods
During the durability testing of sulfate traps, it was decided
that a more accurate measure of trap pick-up efficiency should be obtainable
by directly measuring the concentration of sulfate in the exhaust gas
entering the trap during a test and comparing this to the outlet concen-
tration determined by the particulate filter. Prior to this we had relied
on periodic checks of the test vehicle, using the particulate filter, with
the sorbent removed, to establish a base line for sulfate emissions from
the catalyst. As our sophistication grew in the area of sulfate production
-------
- 142 -
over catalysts, the magnitudes of run-to-run variations and catalyst
aging effects on sulfate production became apparent, prompting the desire
for simultaneous measurement of trap inlet and outlet sulfate concentrations
during efficiency tests.
The immediate stimulus for this type of check occurred at the
beginning of the CaCOS sorbent vehicle durability test, Section VI.2.3.
Prior to mounting the trap, the vehicle, with fresh monolithic catalysts,
was run for 3 200 km to age the catalysts. At this point, two one hour
64 km/h cruise runs were made, as shown in Table C-2, and values of 0.026
and 0.032 g/km of sulfate obtained on the particulate filter. This pointed
up the difficulty of obtaining accurate percentage pick-up values for any
given run, unless simultaneous pre-trap readings were also obtained. There-
fore, subsequent readings, with the trap in place, were made of the pre-trap
Table C-2
Comparison of Goks^yr-Ross and Filter Sulfate Measurements
(Pre-Trap or With Trap Removed)
64 km/h Runs
Sulfate Emissions, g/km
Catalyst Age, km Goks^yr-Ross* Filter
3 200** 0.026
.032
7 200 0.019
.020
14 500 .015
.023
20 600 .018
.022
20 600** .046 .036
.045 .042
.037 .031
.055 .031
* pre-trap readings
** trap removed from vehicle
-------
- 143 -
sulfate concentration by the Goks^yr-Ross technique, in addition to the
post-trap filter measurements.
It was found that the Goks«5yr-Ross readings, out to 20 600 km,
were lower than the values obtained at 3 200 km by the filter method.
While catalyst aging might be the reason, the Goks^yr-Ross readings
showed no downward trend over this interval. At 20 600 km (corresponding
to 17 400 km of trap aging) the durability test was ended, the sorbent
removed from the trap canister and some mechanical repairs made to the
test vehicle's engine. At this point, the vehicle, minus trap, was now
used to run some direct comparisons between the Goks^yr-Ross and Filter
techniques. The former samples the raw exhaust downstream of the
catalyst, the latter, of course, the diluted exhaust from the particulate
sampler tunnel.
Several inconsistencies were noted. First, the Goks^yr-Ross
values jumped sharply from their previous levels; the four one hour cruises
in this configuration averaging 0.046 g/km compared to the first six runs
beginning at 7 200 km, which averaged only 0.020 g/km. It is difficult to
see how removal of the trap, located downstream of the sampling point, or
minor mechanical repairs, could have produced a real change in sulfate
production over the catalyst of this magnitude. Experimental error involving
the Goks^yr-Ross method itself may have been involved, and this will be
discussed in more detail shortly.
This jump in Goks^yr-Ross readings also brought them to a level
exceeding the simultaneous values determined by the filter method. The
four filter values at 20 600 km averaged 0.035 g/km. In contrast, the
six Goks^yr-Ross readings in the interval from 7 200 to 20 600 km had been
lower than the 3 200 km values recorded by the filter, 0.020 g/km compared
to 0.029.
The average filter readings at 3 200 and 20 600 km are fairly
close to each other and can probably be accounted for by the known
variability of catalyst vehicle sulfate emissions. However, the more
than doubling of the Goks^yr-Ross readings suggests there may be some
inherent sources of experimental error in the method. Our experience with
this technique indicates several possible problem areas. In our runs, it
was necessary to use a stainless steel tube to connect the exhaust system
and the condenser itself. This line was externally heated in an effort to
prevent acid condensation. However, it is difficult to balance this heating
so that line condensation is completely eliminated on the one hand and the
gas is not heated so high that H2S04 cannot be completely condensed in
the coil on the other hand. Compounding this problem is the possibility
of chemical reaction on the walls of the metal sample line. Depending on
the degree of sulfation of the metal surface, there may be times when
-------
- 144 -
reaction occurs, leading to a lower collection of acid in the coil. At
other times, there may be a transfer of sulfated material from the sample
line to the coil, leading to erroneously high values. The transferred
material may be either entrained or revaporized acid, or else actual
chunks of metal sulfate flaked off the wall.
Because of the foregoing uncertainties, additional work will
be necessary to pin down and eliminate potential trouble spots in using
the Goks^yr-Ross method for sampling vehicular exhaust and to ensure that
it can be used reliably for this purpose. Laboratory use of this technique
is much simpler and should not be subject to the same degree of uncertainty.
C.I.1.2 Comparison of Sulfate Sampling Points
The previous section described how, in the only direct simultaneous
sulfate readings taken (at 20 600 km), the Goks^yr-Ross method, sampled
from raw gas in the vehicle exhaust system itself, gave higher readings
than the filter, sampled many feet further downstream, from the dilution
tunnel. In order to ensure that this result was not due to loss of sulfate
during its passage through the exhaust system and dilution tunnel, an
experiment was run using a vehicle equipped with monolith oxidation
catalysts but no sulfate trap. Two samples of undiluted exhaust were
withdrawn simultaneously from a position equivalent to where the trap
inlet would have been located and from a point just before the entrance to
the dilution tunnel. These two samples were passed through the Goks^yr-Ross
coil. A third sample was taken at the same time in the conventional manner
by filter from the dilution tunnel. Table C-3 shows the results obtained
for two successive hour readings, running the car at 64 km/hr cruise.
Table C-3
Simultaneous Sulfate Readings at Three Sampling Points
Sulfate Emissions, g/km
Method Sample Location 1st Hour 2nd Hour
Goksfiyr-Ross inlet to empty 0.027 0.034
trap
Goksrfyr-Ross inlet to .010 .014
dilution tunnel
Particulate Filter dilution tunnel .026 .035
-------
- 145 -
Several points are of interest. First, the Goks^yr-Ross readings from
the two different sampling points are not in agreement. The readings from
further downstream are lower. However, it cannot be inferred from this
that sulfate was lost to the walls during its passage down the exhaust
system, since the filter values, still further downstream, are in excellent
agreement with the first set of Goks^yr-Ross readings. A more likely
explanation for the low Goks^yr-Ross readings at the dilution tunnel
inlet is related to the connection to the condenser. Closer coupling of
the condenser coil and the exhaust line were possible at this point.
Therefore, the sample of exhaust gas may have entered the condenser at
too high a temperature to permit complete capture of the acid. Unfortunately,
a second condenser was not placed in series with the first to determine if
this was actually occurring.
A second interesting aspect is the showing that in each of the
three sampling regions, the second hour of running produced higher levels
of sulfate. Referring back to Table C-2 also, it can be seen that in
many cases the second of two successive one hour cruises gives higher
sulfate values, and in no case is the second hour significantly lower
than the first. It is possible that during the first hour of operation,
the sulfate storage capacity of the catalysts were partially or completely
saturated so that during the second hour emissions were more nearly equal
to S02 oxidation.
C.I.1.3 Comparison of Absolute Accuracy of Analytical Methods
As has been seen in the two preceding sections, a number of
discrepancies have been observed between simultaneous Goks^yr-Ross and
filter samples, and even between simultaneous GoksflSyr-Ross samples taken
from different portions of the exhaust system. Although a number of
possible difficulties have been pointed out which make one suspicious
of the Goks^yr-Ross results, it would be desirable to have known sulfate
concentrations in the exhaust gas, so that the analytical results can be
compared on an absolute basis. Unfortunately, this cannot be done when
the catalyst is used as the source of sulfate. Even though the S02 inlet
concentration to the catalyst is known from the fuel sulfur content and
consumption rate, the conversion rate to sulfate is not.
In an effort to provide a known level of sulfate in the exhaust
gas, a test vehicle was fitted with fresh catalysts to prevent the
possibility of sulfate release from the catalyst. The car was run with
essentially sulfur free fuel, <20 ppm. Sulfate was added to the system
by continuously metering into the exhaust stream, just downstream of the
catalysts, a known amount of IN H2S04. However, it appears that the
acid was not evaporated by the hot exhaust gas rapidly enough to prevent
reaction with the walls of the exhaust system. In all cases, both the
Goks^yr-Ross and filter samples gave values much lower than the amount of
sulfate injected. Only one result was as high as 75%, the rest ranged
from about 20 to 50% recovery. Even more indicative of wall reaction was
-------
- 146 -
the collapse of the exhaust pipe in a cloud of orange colored smoke after
only about eight hours of acid injection. Clearly then, in order to
conduct a meaningful experiment of this type, an aerosol generator will
have to be used rather than simple acid injection.
C.I.2 Cross-Checks of S02 Analytical Techniques
Two methods of measuring S02 have been used; the TECO pulsed
UV fluorescence instrument and the H202 bubbler technique. Both are
described in Appendix B. The TECO was used to sample gases from the
dilution tunnel, while the bubbler could be used for both raw and diluted
gases. In the former mode it was useful only for steady speed runs,
with proportionally diluted gases it was possible to measure S02 during
cyclic and steady speed runs. In the following sections, we will discuss
the agreement or lack thereof between the two methods, and probable
reasons for those cases where agreement was not obtained. In the case
of Task 2, where non-catalyst vehicles were run, problems associated with
obtaining good sulfur balances will also be discussed. Sulfur balances,
of course, are not to be expected with catalyst cars because of the sulfur
storage phenomenon.
C.I.2.1 TECO Results
In Appendix B, Section B.3.I., the TECO instrument and certain
calibration problems were described. These problems, which basically are
due to the quenching effect of CO, C02 and 02 on the fluorescence of UV-
excited S02» and to the fluorescence of hydrocarbon compounds, were not
known at the beginning of our work on Task 2. The literature available
on the TECO did not mention these phenomena. As a result, the TECO S02
values taken for the 1974 Chevrolet, Mazda, and Honda, shown in Appendix
Tables E-l through E-3, were made with an instrument calibrated using an
S02 in nitrogen gas. This, of course, did not correct for the quenching
and fluorescence interferences. An attempt was made to develop a correction
factor to account for the difference in composition between the calibration
gas and the air-diluted exhaust gas from which actual samples were taken,
but this was not successful, as indicated by the "corrected" values for
S02 which are given in Appendix Tables E-l through E-3. They are uniformly
high, and the Mazda gave the highest readings of the three cars. This
may have been due to the high hydrocarbon emissions from this vehicle,
much higher than from the other two cars. As mentioned, hydrocarbons are
able to fluoresce and give a positive signal to the TECO instrument. The
diesel tests showed a less than complete S02 balance, in contrast to the
other three cars. In this instance, it is believed that fouling of the
sample cell by the very high level of particulate matter in the exhaust led
to the low S02 readings.
-------
- 147 -
Another problem encountered with the TECO was that of sensitivity.
Although the unit had a low range of 0-10 ppm, in practice it was found
that this scale was too noisy to be used. It was necessary, therefore, to
use the 0-50 ppm range. As a consequence, the unit could not be used under
conditions of very low S02 concentration, as when low sulfur fuels were
burned or with low speed driving modes, when the exhaust gas was diluted
to a greater extent in the constant volume sampler.
In order to determine just how low in S02 concentration the
instrument could be used, a tabulation was made of results from a number
of runs from Task 3. In these cases, a proper instrument calibration was
made, using S02 in air, which approximates the composition of diluted
exhaust gas. These included fuels of 0.012 and 0.032 wt. % sulfur content,
and FTP and 96 km/h cruise driving modes. Preconditioning of the catalysts
prior to running was with either the turnpike or city driving cycle. As
shown in Table C-4, the FTP runs generally yielded very low S02 levels in
the diluted exhaust gas. During the driving cycle S02 levels frequently
dipped well below 1 ppm. On the other hand, at 96 km/h cruise, the values
were generally at least 2-3 ppm. Since there is no way of knowing the
"correct" value of S02 which should be emitted from a given test, due to
the unknown factors of catalyst storage and release or oxidation of exhaust
gas 802, the reliability of each measurement was judged by its agreement
with a simultaneous H202 bubbler reading. Although there can be no guarantee
that either reading is correct for a given run, a close agreement is good
evidence that both are correct.
It was found that the FTP runs, with their low S02 exhaust gas
concentrations, gave very poor agreement between the TECO and bubbler
methods. Ratios of bubbler to TECO values ranged from 0.06 to 1.55. On
the other hand, the cruise values showed ratios ranging only from 0.88
to 1.24, with an average of 1.11. Thus, it appears likely that the TECO,
when properly calibrated, yields reliable S02 readings at concentrations
of several ppm or more. However, at somewhat lower levels the instrument
does not appear useful. Of course, this conclusion is subject to the
objection that at such low levels, the bubbler method may be subject to
erratic results, along with or instead of the TECO. In the absence of
any independent check on the accuracy and reproducibility of the bubbler
method at such low values we can only rely on the difficulty of reading
values as low as 1 ppm or below on a 0-50 ppm scale to conclude that the
TECO, at least, should not be used in this concentration range.
Another test was made, at 96 km/h, to determine the absolute
accuracy of the TECO unit. A non-catalyst vehicle was run for one hour,
using a fuel containing 0.032 wt. % sulfur. The fuel consumption was
determined and the total amount of sulfur burned was calculated as 4.46
grams. The TECO, calibrated with an S02 in air gas mixture, sampled the
diluted exhaust gas continuously, and showed a sulfur usage of 4.60 grams.
-------
TABLE C-4
COMPARISON OF TECO AND BUBBLER S02 RESULTS AS A FUNCTION OF
Fuel Sulfur,
Wt. %
0.032
0.032
0.012
0.012
0.032
0.032
0.032
S02 CONCENTRATION
Task 3 Results - Diluted Exhaust Gas
Driving Mode
Test Condition
Base Case
(turnpike)
Base Case
(city)
Base Case
(turnpike)
Base Case
(city)
Restricted Air
(turnpike)
Restricted Air
(city)
Aged Catalyst
(turnpike)
ppm range
0.2-1.0*
0 - 0.8
0.3-1.4
0.2-0.7
0.3-0.5
0.1-2.0
0.2-1.3
1.1-4.0
0 -1.5
1.3-5.3
0.3-1.3
0 -0.8
0 -0.6
FTP
g Sulf
TECO
0.18
0.11
0.43
0.18
0.11
0.30
0.18
0.82
0.14
1.05
0.22
0.11
0.08
ur/test
H202
0.19
0.17
0.26
0.25
0.08
0.11
0.02
0.00
0.67
0.01
0.87
0.00
0.14
0.02
H202
TECO
1.06
1.55
0.60
1.39
1.00
0.06
0.82
0.07
0.83
1.27
0.25
ppm range
3-4
3-5
2-3
1-3
5-10
3-10
3-4
96 km/h
g Sulfur/test H202
TECO H202 TECO
7.5 8.2 1.09
6.4 7.0 1.09
4.0 4.4 1.10
i
3.3 2.9 0.88 £
i
13.9 15.7 1.13
13.8 16.9 1.22
7.1 8.8 1.24
* Two FTP runs are initial and final in test sequence used in Task 3.
-------
- 149 -
A final check was made of the absolute accuracy of the TECO,
this time when sampling raw rather than diluted exhaust. In this case,
a catalyst-and trap-equipped car was used, since sulfate analytical
methods were also being examined, as described in Section C.I.1.1 of this
Appendix. In two of the 64 km/h cruise runs, each consisting of two
one hour intervals, TECO readings were taken before the catalyst, and
between the catalyst and trap. The instrument had been calibrated with
502 in air. However, with an air pump supplying secondary air, the
oxygen content of the sampled exhaust was about 6%. Our earlier work
on oxygen quenching effects, described in Appendix B, Section B.3.1,
had indicated that the correction factor between 6% and 20% oxygen should
be very small.
As seen in Table C-5, the TECO values before the catalyst
agreed very well with the amount of sulfur calculated to be liberated
during fuel combustion, with the first two hours averaging only 1%
high and the last two hours about 8% high. There was, of course, no way
of judging the absolute accuracy of the TECO readings after the catalyst,
so again reliance was placed on whether or not the results agreed with
concurrent bubbler readings. Table C-5 shows ambiguous results were
obtained. During the first two hours, the TECO readings did not agree
with each other. This might have been attributed to changes in the
catalyst storage characteristics, but the bubbler results do not show
this hour-to-hour change. In addition, both TECO readings are substantially
below their bubbler counterparts. The second set of readings, however,
are in very good agreement with each other, both hour to hour and between
the TECO and bubbler values. It should also be mentioned that an attempt
Table C-5
Determination of TECO Accuracy in Undiluted Exhaust Gas
64 km/h Cruise 0.032 Wt. % Sulfur Fuel
Grams of Sulfur
Consumed TECO Bubbler
Run Hour In Fuel Before Catalyst After Catalyst After Catalyst
1 1 1.88 2.04 0.72 0.38
2 1.89 1.78 .50 .36
2 1 1.83 2.02 .56 .55
2 1.79 1.91 .54 .56
-------
- 150 -
was made to monitor S02 in the dilution tunnel, but under this driving
condition the concentration was under 1 ppm and erratic results were
obtained.
C.I.2.2 Peroxide Bubbler Results
The second method for monitoring S02 emissions, besides the
TECO instrument, is the use of peroxide-containing bubblers. This
technique was described in Appendix B, Section 6.3.2. Several comparisons
involving the bubbler and the TECO were described in Section C.I.2.1.
It was found that the bubbler method may have some problems when the
total quantity of S02 absorbed is relatively small, as in the short
duration, low concentration FTP runs shown in Table C-4. On the other
hand, longer duration, higher concentration runs, such as the 96 km/h
cruises of Table C-4 are, judging by the good cross-check with the TECO
results, reliable. Other bubbler tests discussed in Section C.I.2.1
showed that the bubbler accurately measured total fuel sulfur content in
a one hour, 96 km/h cruise with 0.032 wt. % sulfur fuel, but ambiguous
cross-checks were not obtained during four one hour 64 km/h cruises,
Table C-5.
Invariably, in those cases where the bubbler results could be
compared directly to the amount of fuel sulfur consumed, the results
have come out higher than expected. Such values were obtained with the
non-catalyst cars used in Task 2 and with before catalyst samples from
Tasks 3 and 4. A number of samples were rechecked, after titrimetric
determinations, by analyzing retained portions of the sample solutions
gravimetrically. These samples were taken during 96 km/h cruise tests
of two non-catalyst cars, the 1974 Mazda and Honda used in Task 2, and
from the 1975 350 CID vehicle used in Task 3. In this instance, the
sample was taken before the catalyst. Table C-6 shows that both the
titrimetric and gravimetric procedures gave higher sulfur values than
calculated from the fuel consumption and sulfur content. Each result shown
represents the average of 2-4 separate cruises. The gravimetric gave
somewhat closer agreement, however, averaging 22% higher than the expected
value. The titrimetric procedure averaged 37% higher. Some possible
reasons for these high results, and steps taken to correct them are described
in the next section. It should be pointed out here that although the
gravimetric technique gave results in closer agreement with the calculated
sulfur values, this method is not suitable for general use since its
sensitivity is less than that of the titrimetric procedure, and it is
much more time consuming.
C.I.2.3 Analytical Procedure Problems and Solutions
In the previous section, a comparison of titrimetric and
gravimetric analyses of perioxide bubbler samples showed that both were
-------
- 151 -
Table C-6
COMPARISON OF TITRIMETRIC AND GRAVIMETRIC
ANALYSES OF PEROXIDE BUBBLER SAMPLES
Vehicle
Mazda
Mazda
Honda
Honda
1975-
350 CID
Average
96 km/h - Averages of one hour runs
Fuel Sulfur, grams /test
Sulfur,
Wt. %
0.065
.032
.065
.032
.032
.032
.012
.012
.032
Calculated
1.85
1.15
1.10
0.50
4.20
4.25
1.65
1.55
4.30
Titri-
metric
2.1
1.6
1.4
0.7
5.6
5.0
2.6
2.6
5.3
Gravi-
metric
1.9
1.4
1.4
0.7
4.4
4.7
2.3
2.1
4.9
Titrimetric
Calculated
113
153
127
140
133
118
158
168
123
137
Gravimetric
Calculated
103
121
127
140
105
110
140
136
114
122
higher than the expected results, with the titrimetric even higher than
the gravimetric, by an average of about 12%. Thus, it appears that there
is a systematic error in these techniques, leading to somewhat high results.
This may have been related to uncontrolled blank corrections, which will
be discussed shortly. First, the possible reasons for the difference
between the titrimetric and gravimetric procedures will be considered.
It will be remembered that titrimetric analyses of peroxide
bubbler samples also were high in comparison to properly calibrated TECO
readings, Table C-4, by about 11% on average. Still another comparison
-------
- 152 -
of the tltrimetric procedure, this time against the gravimetric, using
filter sulfate samples, can also be made. For this purpose, the results
in Appendix B, Table B-3 were used. There the titrimetric values average
14% higher than the gravimetric. Thus we have three independent checks
in which the titrimetric procedure averaged 11-14% higher than TECO or
gravimetric results. The most likely explanation for this lies in the
difficulty of detecting the end point of the barium perchlorate-Sulfanazo III
titration described in Appendix B, Section B.2.4. A rather subtle change
in a shade of blue occurs, and when monitored by eye, a natural tendency
would be to exceed the end point.
The problems leading to overly high sulfur values for both
the gravimetric and the titrimetric methods may be related to blank
corrections. Blank corrections may be required to account for any sulfate
content in the deionized water used to prepare the peroxide bubbler
solutions, to rinse out the bubblers after use (the rinsings are added
to the sample), and to prepare the barium perchlorate titrant solution.
Prior to the comparisons reported in the previous section, these blanks
were not determined and used in a systematic manner. Subsequently, the
blank correction was determined as 0.0023 mg/ml of bubbler solution. Since
a total of 100 ml is used in each bubbler for the peroxide and rinse solutions,
the correction to be applied is 0.00023 grams per bubbler determination.
An illustrative calculation can be made to demonstrate the likelihood that
a blank correction of this magnitude can affect the accuracy of a given
test. If we assume that the equivalent of 2 grams of sulfur is emitted
during the test, and the bubbler samples one part in 2800 of the exhaust
gas (dilution tunnel flow rate of 400 SCF/min. and sample flow rate through
the bubbler of 0.14 SCF/min.) then the equivalent of 0.0007 grams of
sulfur is trapped in the bubbler, or about 0.0021 grams of sulfate. Thus
the blank correction amounts to approximately 10% of the total determination,
and if not subtracted out will lead to a result too high by this amount.
The importance of the blank correction will depend, of course, on the
total amount of sulfur trapped in the bubbler. In turn, this will depend
on the fuel sulfur content, the driving mode, the length of sampling time,
and the fraction of exhaust gas sampled. An examination of Tables C-4 through
C-6 shows that a wide range of trapped sulfur is encountered in practice,
leading to situations where the blank correction may be as large or larger
than the sample itself to those where the blank correction is negligible.
Since the titrimetric and gravimetric methods have consistently overestimated
sulfur values, regardless of the amount of trapped sulfur, it appears that
blank corrections, although they should be made systematically, do not
provide the explanation for this overestimation.
After a systematic procedure was adopted for dealing with the
blank correction, a series of 17 standard samples were submitted for
analysis, scattered randomly among the working samples. These were designed
to determine the accuracy and reproducibility of the titration procedure
itself, divorced from any sampling considerations. A standard H2S04 solution
-------
- 153 -
was used to prepare 100 ml samples, equivalent to bubbler samples, each
containing 0.0048 grams of sulfate. Thus the blank correction amounts
to about 5% of the true value. With this blank correction applied, the
average value of the 14 usable results was 0.0051 grams of sulfate,
about 0.0003 grams high.* However, the standard deviation was 0.0004
grams of sulfate, so the difference does not appear to be significant.
It would seem then that some step in the sampling procedure may have led
to the high results, or possibly systematic error in measuring fuel sulfur
content or consumption. However, none of these possibilities seems likely.
The flow rates during sampling were carefully monitored with wet test meters.
Any sorption efficiencies of the bubblers would have led to low rather
than high results. The fuel sulfur contents were checked in two ways.
The fuels were blended carefully from low sulfur base stocks by addition
of sulfur compounds. Then the sulfur levels were measured by X-ray
fluorescence or micro-coulometry. In addition, round-robin studies with
the laboratories have shown our analytical techniques for fuel sulfur
content to be reliable. Finally, the fuel consumption values were deter-
mined simultaneously by weight loss of the fuel container and by the
carbon balance method. Generally, the two techniques agreed within
several percent. The reason for the overestimations of S02 values, there-
fore, remains unclear. Fortunately, this parameter is not of direct
quantitative interest, since it serves generally only to provide a picture
of the trends of sulfur storage on catalysts and on traps. As such, the
accuracy already achieved seems satisfactory for this purpose. As will
be discussed shortly, the principal measurement required, that of filter
measurements of sulfate emissions, is not affected by the considerations
discussed here in regard to peroxide bubbler samples for S02.
Before turning to the filter samples, a final word is in order
regarding the blank correction. Subsequent to the completion of this
contract, we have replaced our manual titration, visible end-point
procedure with a fully automated system, including spectrophotometric
end-point determination. Blank corrections recently determined by this
improved procedure have been at least an order of magnitude smaller than
that previously reported. It would appear, therefore, that the manual
procedure blank was not due so much to actual sulfate contents of the
solutions employed as to the ability of the human eye to detect an end
point. That is, even in the absense of any sulfate, a finite amount
of titrant was required to produce a visible change in indicator color.
With the more sensitive instrumental method, this amount is much less.
The other 3 samples gave results too high or low by factors of 2 or
more, due apparently to operator error, and were not used in the
calculation of the average.
-------
- 154 -
In the filter procedure for measuring sulfate emissions,
Appendix B, Section B.2.1, an additional source of spurious sulfate can
arise from the filter itself. Blank corrections from the solutions are
generally negligible, since the sampling rate of 15 SCF/min. is over 100
times greater than that used for the bubbler. Therefore, even if only
1% conversion of S02 to sulfate is occurring, about the level encountered
with non-catalyst vehicles, the blank correction would still be no greater
a fraction of the filter sulfate sample than for the corresponding bubbler
S02 sample. Usually, however, our primary analytical concern is with
catalyst vehicles which convert substantially more than 1% of the S02 to
sulfate, further reducing the importance of the solution blank correction.
The blank correction due to sulfate leached off the filter itself varies
from batch to batch. However, the value of this correction has ranged from
the equivalent of 0.5 mg/km to 1 mg/km for the FTP, and thus is important
again only for very low emittiiig cars. However, once the blank for a
given batch of filters is established, even this small source of error
is eliminated.
-------
- 155 -
APPENDIX D
SUPPLEMENTAL INFORMATION ON SULFATE TRAPS
D.1 Analyses of Used CaP/SiO^/NaP Pellet Sorbant
Scanning electron microscopy, x-ray energy dispersive analysis,
x-ray diffraction analysis, and chemical analyses for sulfate and
carbonate have been carried out on a (CaO) sulfate "trap" from a catalytic
exhaust operated vehicle. This characterization was made to determine the
fate of sulfur in the automotive fuel and evaluate the effectiveness of the
H2SP/ removal from the exhaust gas.
X-ray diffraction analysis provided the crystalline composition
of the fresh and used material shown in Table I.
Table I
X-Ray Diffraction Analysis
(Crystalline Compounds Present)
Fresh
Calcium Oxide (CaO)
Calcium Silicate, hydrate
(Ca2Si04.H20)
Sodium Carbonate, hydrate
(Na2C03.10H20)
Used. 26.634 Miles
Calcium Oxide (CaO)
Calcium Sulfate (CaS04)
Calcium Silicate, hydrate
Sodium Carbonate, hydrate
Density, pore volume, and carbonate content were also obtained
on the samples with the following results:
Density, grams/ml
Pore Volume, ml/gram
Carbonate, (calc. as % CP2)
Fresh
2.70
0.35
1.29
Used
2.77
0.20
10.7
Infra-red examination (H. T. White - AID) of a fresh powdered
sample showed calcium oxide, calcium silicate and a calcium-nitrogen
compound (fresh material was prepared from a nitrate before calcination).
Similar analysis for a used sample showed calcium sulfate, calcium
carbonate and calcium oxide. A chloroform extract of the used sample
indicated only a trace amount of hydrocarbon to be present.
Several samples of the used material and calcium sulfate was
analyzed by both the titrimetric and gravimetric methods for sulfate ion.
The titrimetric technique is applicable to low levels and/or small
(milligram) size samples. The gravimetric method is best applied for
0.1 to 0.3 gram size samples where a more homogeneous sample is assured.
Reasonable agreement was obtained and values obtained are found in
Table II. Differences probably result from sample inhomogeneity.
-------
- 156 -
Table II
Sulfate Analyses, % 50,°
Sample Titrimetric SO," Gravimetric S0.
CA-4 Used, 26,600 miles
CA-6 CaSO^
CA-7 Fines 26,600 miles
CA-8 CaSO,
wt. % wt. %
43.83 39.79
75.23 68.41
59.77 51.04
71.23 65.62
To determine how the sulfuric acid mist reacted with and affected
the lime in the usual 1/8 inch (3 mm) pills, two means of analysis were
used. The first involved a careful scraping of two pills. The outside
(about 1 mm) was separated for analysis. Similarly, a section about 3/4
of the way in was isolated. The remainder and the two other portions
were analyzed for sulfate ion by the gravimetric method. Analyses are
shown below:
Sample Gravimetric SO."
K 4—
wt. %
Fresh (CaO +SiOo + Na20) 0.054
Used, 26,634 mites, ground homogenized pills 38.98
Outside of pills, 26,663 miles 47.84
3/4 way in pills, 26,634 miles 34.61
Center of pills, 26,634 miles 25.01
For the second method, x-ray energy analysis was used. A pill
was cut in cross section and the surface was traversed in a line from one
edge through the center to the other edge. This was done in the scanning
electron microscope. X-ray energy analyses for sulfur and calcium were
determined at selected locations across the catalyst pellet. The relative
peak intensities for sulfur correlate with the gravimetric sulfate
determinations. The relative sulfur content decreases going into the
pellet, while calcium increases. This is shown in Table III.
-------
- 157 -
Table III
Scanning Electron Microscopy: X-Ray Energy Spectrum(s)
(Relative Peak Intensities)
Sample in Cross Section, 3 mm Pill Sulfur (SO^?) Calcium
mm. in from outside, left or right Rel. PeaTT"Ht. Rel. Peak Ht.
0.01 left (edge) 53 50
0.35 left 56 50
0.52 left 53 49
0.52 right 55 48
0.75 right 36 77
0.75 left 35 77
1.50 center 35 70
A scanning electron micrograph at 1,OOOX of a fresh CaO pill is
compared with that of a used pill in Figure(s) 1 and 2. These show the
granular, porous structure of the surface. In cross section they were
similar.
Two glass fiber filters for total exhaust particulate collec-
tion from tests run with Ford 99 were examined. Filter L-478 was from
a run, where no CaO trap was in operation. The H2S04 emission was
0.127 gms. H-SO^ per mile. A total of 55.9 mgs. H2SOA was collected.
The second filter, 99-43, was from a trap operation (26,500 miles);
0.0007 grams t^SO^ per mile emission and only 0.87 mgs I^SO^ was
collected. Figures 3 and 4 at 200X compare the no trap versus trap
operation. The no trap filter, L-478, shows particulates in the 1-10
ym size range. Filter 99-43 has a small number of larger, 20-50 ym
particles, probably from attrited trap material. At l.OOOX in Figures 5
and 6 the glass fiber filter from the no trap run shows the glubular
wetting of the fibers by the sulfuric acid. From the trap run (Figure 6)
only a few ^10 ym particles are observed. Figures 7 and 8 are x-ray
energy spectrums obtained from the respective filters. The major peak,
silicon, at 1.85 KEV is from the glass fibers. The sulfur peak at 2.3 KEY,
in the no trap run has a relative intensity of 25. The very low sulfur
from the "trap" run has a relative intensity of 6.
These micrographs and x-ray spectrums indicate the effectiveness
of the sulfuric acid reduction using the trap material.
-------
SCANNING ELECTRON MICROGRAPHS -- 1, OOOX
Fig. 1 - Fresh CaO
Fig. 2 - Used CaO
\
.
•
• '•»;
- -; .:.-•',
1
'S
Nr
- ;*
*"*
It
'
^
' "
---- *
' '
CC
I
-------
m
SCANNING ELECTRON MICROGRAPHS -- 200X ' ' 50
Fig. 3 Fig. 4
L-478 No Trap 99-43 CaO Trap
v***
.
i
i-1
in
^3
I
55.9 mgs. H2S04 ^ ^s'
0.127 gms H2S04/mile °- °007 gms-
-------
SCANNING ELECTRON MICROGRAPHS -- 1, OOOX
Fig. 5
L-478 No Trap
Glass Fiber Filters Ford 99
Fig. 6
99-43 CaO Trap
'\
f. *
,-
.
'4;» .• -.>i
^>
*%h>
fjfiHh
*. k*.. 'Tj <
'
o
i
55.9 mgs. H2SO4
0.127 gms H2SO4/mile
0.87 mgs. H2SO4
0.0007 gms H2SO4/mile
-------
X-RAY ENERGY SPECTRUM(S)
GLASS FIBER FILTERS FORD 99
FIG. 7
L-478 No Trap
55.9 mgs H2SO4
0.127 gms H2SO4/mile
FIG. 8
99-43 CaO Trap
0.87 mgs H2SO4
0.0007 gms H2SO4/mile
Silicon, Major Peak at 1. 85 KEV
Sulfur Peak at 2.3 KEV
-------
- 162 -
D.2 Westvaco Corporation Report on the Use of
Activated Carbon to Remove SO, From Exhaust
INTRODUCTION
The sulfur content of automotive fuels is converted to sulfur dioxide
during combustion and is emitted through conventional engine exhaust
systems in this form. However, catalytic converters which are to be
installed to control hydrocarbon and carbon monoxide emissions from
future automobiles have been found to cause partial oxidation of the S02
to $03. This sulfur tri oxide combines with water vapor present to form
H£S04 which is emitted with the exhaust stream in the form of a mist or
aerosol. The acid emission constitutes not only a health hazard but
also could cause severe corrosion of exhaust system and body components.
Experimental evidence developed by Esso Research and Engineering indi-
cates that the degree of S02 conversion to $03 is on the order of 30-70%.
Since the industry average sulfur content of gasoline is 300 ppm (by
weight) this translates to an exhaust stream S03 concentration level of
about 6-14 ppm (by volume), or an average of 10 ppm. On this basis,
acid emissions would be about 3.5 gm. 33% H2S04 per gallon of fuel con-
sumed assuming ambient conditions of 80°F and 60% relative humidity.
One possible means of control of S03/H2S04 emissions from vehicles is the
use of a reactor containing activated carbon to chemically reduce SOs
formed in the ronvprtor b^rk to SO'' '.vhich is pov/ normally emitted snd
!• K v*r\t-\vs r- ** ^ *~ "• T^»***» ** r* \ i f* «•• *•. «-*TT. -L ~ •* .•* «».-— LT* _._t r> _ T ^ • _ . _-T "I -
tin i v, ii i v_pi ._ov- 1 1 1,0 u i coo oc v ci c pu IIUIIUM pi uu I CHI . ricllllllllcltj' Ua I L.U I O~
tions indicated that carbon consumption due to the SOs reaction would be
only about 0.6 pound over the course of 50,000 miles under the S03 produc-
tion conditions noted above so that such a system might well prove
feasible.
The laboratory investigation discussed here was intended as a first step
in establishing possible applicability of activated carbon in $03 con-
trol for automotive systems. Principal areas of study involved the
effects of temperature and reactor space velocity on the reduction and
sorption of $03, and the rate of side reactions with other exhaust gas
constituents resulting in carbon burn-off.
WESTVACO CORPORATION
Charleston Research Center N. Charleston, S. C.
-------
- 163 -
EXPERIMENTAL APPARATUS
The experimental apparatus is shown schematically in Figure 1 and con-
sisted basically of a gas mixing and S03 generation section, a reactor
section for contacting the gases with activated carbon, a condenser and
filter section for removing and collecting water and $63 from the
effluent gas, and an S02 analysis section.
$03 and water vapor were generated simultaneously by introducing a sul-
furic acid solution into a Vycor tube maintained at 1000-1200°F by an
electric furnace. The proper liquid flow rate was obtained using a Sage
Model 220 syringe pump. At the low liquid flow rate used (approximately
0.37 cc/min.) a constant rate of vapor generation was produced by admit-
ting the liquid through a drawn glass capillary which, at its exit and
in the heated zone, was in contact with a small piece of fritted Vycor
glass. This arrangement was found to greatly attenuate the tendency
toward vapor surges produced by the buildup and flash evaporation of
individual liquid drops in the hot tube.
Other mixture gases were admitted through a concentric tube surrounding
the liquid discharge capillary. Thus, flow of diluting gases through
this tube prevented condensation of H2S04 in the cold inlet section by
preventing back diffusion of the vapor. A more detailed diagram of
this inlet section is shown in Figure 2.
After pabSciyt: liiruuyh liie 22 inch heated generator section the gases
entered a reactor tube (0.864" ID) held in a separate tube furnace. The
carbon sample was supported in this tube between a glass frit and a glass
wool plug. A thermocouple inserted into the carbon bed provided tempera-
ture measurement while temperature control was performed manually using
an autotransformer.
Exit gases from the reactor were conducted into a 30 inch condenser
cooled by tap water. A stopcock at the bottom allowed draining of the
condensate. At the top of this condenser, $03 not absorbed into the
condensate water was trapped as H2S04 mist by a 60 mm diameter medium
porosity glass frit (10-15 micron nominal pore size).
Following this filter, SO;? produced in the reaction with carbon was .
analyzed by an Envirometrics S02 Analyzer (Model S-364) calibrated to
give full scale deflection for 10 ppm S02-
WESTVACO CORPORATION
Charleston Research Center N. Charleston, S. C.
-------
N2 02 NO
J
\
H2S04 Reservoir
Syringe
Pump
FIGURE 1
8.03 REACTION APPARATUS
$03 Generator
\
60 mm Medium
Frit Filter
Condenser
Thermocouple
Tube Furnaces
S02
Analyzer
Drain
Carbon Sample
Vent
I
(-•
Oi
I
Charleston Research Center
WESTVACO CORPORATION
N. Charleston, S. C.
-------
Gas Mixture
Inlet
H2S04 Liquid
Inlet
^
I
/
'/.
't
-^
^
Rubber Stoppers
V«jrt Hole
\A
Teflon Heat
Sh-eld
Heated Zone
Vycor Tube
Capillary Tube
\
Sintered Vycor
Chip
Ui
\
FIGURE 2. DETAIL OF GAS-LIQUID INLET TO S03 GENERATOR
WESTVACO CORPORATION
Charleston Research Center
N. Charleston, S. C.
-------
- 166 -
EXPERIMENTAL PROCEDURE
Initial tests of the apparatus were conducted without carbon in the
reactor tube in order to check the S03 generator and test the $03 trap
(condenser and fritted filter) for efficiency. In these "dry runs" it
was indicated that the $03 pickup in the condenser was about equal to
that filtered out by the frit, the combination giving about 85% overall
recovery. Tests using an isopropanol scrubber solution following the
frit indicated some leakage through this filter but a complete mass
balance still could not be obtained. There were visual indications that
acid mist also passed through the scrubbers without being captured.
Because of these problems of incomplete $03 capture, the method used in
later runs to establish the amount of $03 passing through the carbon
bed was to compare the amount of $03 captured by the frit in the carbon
reaction runs with that picked up by the frit in the dry runs (corrected
for time) where there was zero conversion. Analysis of the gas stream
for S02 showed that in fact there was no conversion in the absence of
carbon in the reactor. Continuous SC>2 analyses was made during the
carbon reaction studies to establish the degree conversion of $03.
The general procedure for carbon reaction experiments involved preheating
the $03 generator tube to 1000°F and the carbon reactor to the desired
reaction temperature under a flow of nitrogen equal to the total experi-
mental gas flow of 10 scfm. The N2 was then reduced to 9 cfm and the
syringe pump started to inject 0.37 cc/min. of 0.01 N H2S04. The SC-2
effluent concentration was monitored continuously and luldl SGj break-
through was determined at the end of each run by washing the filter frit
and titrating with 0.001 N KOH to a methyl red endpoint. Because of the
very small quantities of acid involved, this titration was performed with
a syringe microburet. As noted previously the ratio of the amount of
picked up on the frit to that picked up in dry runs (corrected to the
same time interval) was taken as the fractional breakthrough.
WESTVACO CORPORATION
Charleston Research Center N. Charleston, S. C.
-------
- 167 -
EXPERIMENTAL RESULTS
Two sets of experiments were performed to investigate the $03 reaction
with activated carbon and one series of tests was made to determine the
rate of carbon burn-off in the presence of exhaust gas components.
S03 Reactions with 12x30 Mesh Carbon
Initial tests of $03 conversion were made with a 12x30 mesh (0,0661" -
0.0234") granular Westvaco carbon under the conditions outlined in
Table I. The SO-j concentration used approximates gas from catalytic con-
verters, while the space velocity range is representative of that in
reactors of practical size for automotive applications. The variables in
these tests were temperature and space velocity.
TABLE I
EXPERIMENTAL TEST CONDITIONS - 12x30 MESH
Carbon Particle Size - 12x30 Mesh
Carbon Bed Depth - 1", 1/2"
cr\
\j\j 9w\j iu« y o i r /
Volumetric Flow Rate - 10 SCFH
S03 Concentration - 9 ppm
H20 Concentration - 10%
N2 Concentration - Balance
Reaction Temperature - 450°, 500° , 600°F
In each of the runs in this series it was observed that the initial
level of S02 production was low, followed by a gradual increase toward a
steady state level of conversion which was generally established within
about 40 minutes.
At a reaction temperature of 600°F and a space velocity of 30,000 hr.-1
essentially all of the influent SOs was reduced to S02 according to gas
analysis. Subsequent determination of acid collected by the filter frit
during the 5 hour run indicated a quantity of $03 equivalent to 1.4% of
the influent. These results at 600°F are compared to the data obtained
at 500° and 450°F in Table II.
WESTVACO CORPORATION
Charleston Research Center N. Charleston, S. C.
-------
- 168 -
TABLE II
RESULTS OF $03 REDUCTION BY 12x30 MESH CARBON
Temperature
(°F)
Conversion to S02
(By^ Gas Analysis)
Total
(By Fi
% S03 Removed
Iter Analysis)
S.V. = 30,000 hr.-l
600
500
450
100%
80%
45%
>98%
>99%
>99%
S.V. = 60,000 hr.-l
600
94%
>98%
From these data it appears that at 600°F, the influent S03 was almost
completely reduced to S02 by the carbon. At the lower temperatures,
less conversion was found but since the filter titrations showed essenti-
al 1 ~\y coiViplcts 303 rerfiGVal vt appears that S03 sorptiori or stordye took
place on the carbon, bince the same carbon sample had been used in all
of these runs and the experiments had been performed in the order of
decreasing temperature, confirmation of the $03 sorption phenomenon was
sought by heating the sample in a stream of nitrogen to 600°F and
analyzing the effluent for S02- An integration of the resultant S02
concentration/purge volume curve accounted for about 70% of the
apparently unconverted $03 from the 450° and 500° runs. Thus the total
mass balance between $03 generated and S02 produced is close to 90% and
probably within the limits of experimental error.
These data demonstrate that a high degree of $03 conversion to S02 can
be obtained at a temperature of 600°F while at lower temperatures
partial conversion is accompanied by almost total removal of the balance
by sorption. Further, the sorbed $03 is shown to be reduced to S02 on
heating to 600°F.
One additional test was made with the 12x30 mesh carbon at 600°F at a
space velocity of 60,000 hr.-l (STP) by employing one-half the previous
carbon charge in the reactor. The results indicated 94% conversion to
S02 by 9as analysis and greater than 98% removal of S03 as measured by
filter analysis.
Charleston Research Center
WESTVACO CORPORATION
N. Charleston, S. C.
-------
- 169 -
S03 Reactions with 6x8 Mesh Carbon
A second series of experiments was performed to assess the effects of
temperature and space velocity on conversion using a larger mesh size
carbon which may be more practical from a pressure drop standpoint in
actual reactor applications.
In these runs changes in space velocity were produced by changes in bed
depth at a constant feed rate. The experimental conditions are outlined
as follows in Table III.
TABLE III
EXPERIMENTAL TEST CONDITIONS - 6x8 MESH
Carbon Mesh Size - 6x8 Mesh (0.132 - 0.0937")
Total Gas Flow Rate - 10 scfh
$03 Concentration - 9.75 ppm
H20 Concentration - 10%
N2 Concentration - Balance
Temperature - 600°, 550°, 500°F
Space Velocity - 23,600 to 67,500 hr. ^ (STP)
Carbon Bed Depth - 0,33" to 1,33"
The results of these experiments are presented in Figure 3 which shows
the conversion obtained as a function of space velocity and the term
V/FQ with units of
Bed Volume (ft.3)/SOs Feed Rate (gm mole/hr.).
At a space velocity comparable to that used in the first runs with 12x30
mesh carbon (i.e. 30,000 hr.~l) it was found that at 600°F complete
conversion was again attained. At the lower temperatures, conversion
was significantly less than that obtained with the smaller particle size
material. As indicated by the percentage notations in this figure, total
removal of SOs as determined by filtration was 99% at 600° and 98% for
the combined runs at the two lower temperatures.
At the higher space velocities it is clear that decreases occurred in
both conversion and capture of $03. As indicated, capture dropped to 92%
for the combined runs at 44,200 hr."1 and 81% at 87,500 hr."1. It is
WESTVACO CORPORATION
Charleston Research Center N. Charleston, S. C.
-------
FIGURE 3
REDUCTION OF S03 TO SOe OVER 6x8 MUSH WESTVACO GRANULAR ACTIVATED CARBON
ro
O c
o
•r~
CO
i.
0)
>
C
o
o
1.0 -
0.8 -
0.6 -
0.4 -
0.2-
500°F
8.75
Space Velocity, VsxlO'4 (hr.'1)
4.42 2.95
550° F
99%
o
I
1.0
2.0
V/F0 (ft.3/gm mol-hr.
I
3.0
NOTE: % Refers to Total $03
Removal
For Conditions See Table III
2.36
i
4.0
WESTVACO CORPORATION
Charleston Research Center
N. Charleston, S. C.
-------
- 171 -
possible that S03 bypassing was promoted by the use of a very small
reactor bed size compared to the particle size. The bed to particle
diameter ratio was less than 8:1 while the bed depth to particle diameter
ratio ranged down to 2:1. It is, therefore, likely that improved per-
formance would be obtained in full size reactors.
Burn-off Tests
As noted previously the rate of carbon burn-off due to direct reaction
with SOs is relatively small. However, reaction with other exhaust
stream constituents such as 02» H20 and NO is also expected to contrib-
ute to carbon loss. A series of experiments was, therefore, conducted
to evaluate the carbon burn-off rate due to these components.
After initial unsuccessful efforts to measure C02 and CO produced in the
oxidation reactions, a gravimetric method was developed to determine
burn-off rates. Using the same apparatus as that employed in SOs con-
version studies, a carbon sample was simply contacted with selected gas
mixtures at a space velocity.of 30,000 hr.~l for a period of time long
enough to produce a measurable weight loss. The general procedure was
as follows:
1. Heat sample under N2 purge to 700°F for 1 hour in reactor
tube.
2. Cool and transfer ssinple to weiyiviny Lube.
3. Heat sample in weighing tube to 220°F under N£ purge, cool
under N£.
4. Seal tube and weigh sample.
5. Transfer sample to reactor; heat to desired reaction tempera-
ture under N2; react carbon with desired gas mixture.
6. Heat to 700° under N2 purge.
7. Cool and repeat weighing procedure.
Post-heating at 700°F and sample conditioning prior to weighing were
employed to prevent sorption effects from interfering since total weight
changes due to reaction were very small.
The experimental conditions used are listed in Table IV. Various reac-
tion gas mixtures were used and each component, when present, was at the
concentration noted.
WESTVACO CORPORATION
Charleston Research Center N. Charleston, S. C.
-------
- 172 -
TABLE IV
EXPERIMENTAL CONDITIONS
FOR BURN-OFF TESTS
Carbon Mesh Size
Gas Concentrations, 02
H20
S03
NO
N2
Space Velocity
Total Gas Flow Rate
Temperatures
- 6x8 Mesh
- 3%
- 10%
- 10 ppm
- 2,000 ppm
- Balance
- 30,000
- 10 scfm
- 600°, 700°F
Data obtained in these experiments is compiled in Table V. Listed here
is the reactant mixture, temperature, average weight of sample in the
reactor during each run, the weight loss due to reaction, the reaction
time in hours, and a burn-off rate (-r%/hr.) as calculated from:
Wt. Loss x 100%
~r Avg. Wt. x Time
Initial runs (1
mately 5 hours.
subsequent runs
in an effort to
obtained in the
ments of carbon
runs.
-4) were made with a reaction time interval of approxi-
Because of the relatively small weight change involved,
were made over longer intervals (approximately 20 hours)
improve accuracy. Agreement between burn-off rates
long and short runs was poor and conservative assess-
performance should be made using data from the longer
Several observations may be made concerning the results obtained:
1. Reactant Species
SOs, 02» and H20 are seen to react with carbon individually
and in combination to a significant extent at 600°F. Compare
Runs 7, 8, 9 and 10. In every case the burn-off rate due to
these reactants was 2 to 5 times greater than that calculated
for S03 alone. Compare with Run (*), Nitric oxide was not
tested alone, but in mixture with other gases it did not
increase the burn-off rate. Compare Run 5 with 7, and Run 2
with 3.
Charleston Research Center
WESTVACO CORPORATION
N. Charleston, S. C.
-------
- 173 -
TABLE V
DATA FROM BURN-OFF MEASUREMENTS
Run
No.
1
2
3
4
5
?
8
9
10
*
i i
12
Reaction Gas
02
SOs, HgO, 02
S03, H20, 02, NO
S03, H20, 02, NO
S03, H20, 02, NO
SOs, H20, 02, NO
SOs, H20, 02
SOs, H20
02
H20
•en-. rr,i/~ii
WVX^J WIA t W 14
iUS, H2U, \JZ
S03, H20, 02
Temp.
(°F)
700
700
700
600
600
700
600
600
600
600
1 7!-|-oH f^
1 IA WWW. I w
DUU
600
Avg. Wt.
(gm)
SHORT RUN
5.4854
5.4293
5.3427
5.2879
LONG RUN
5.2216
4.7607
4.2974
4.1973
4.1418
4.0795
r 4 . 2 GM S
4..5UUU
5.8239
Wt, Loss
(gm)
S
0.0148
0.0984
0.0999
0.0098
S
0.1228
0.7991
0.1274
0.0729
0.0380
0.0302
=mr.T n 1 nfW
uittfcs » ^- 3 i \s\si\j
u.uny/
0.0625
Time
(hrs.)
5,2
4.8
5.0
5.0
25.1
25.0
22.5
22.0
17.5
20,0
Pr»rn/o v*c n
\X VI 1 V V. 1 ^ •
^^.u
19.7
Burn-off Rate
(-r%/hr.)
0.052
0.38
0.37
0.037
0.0938
0.671
0.132
0.0789
0.0524
0.037
r\r\ O niQ
u 1 1 \-» • w i i_*
U.UD^D
0.055
WESTVACO CORPORATION
Charleston Research Center
N. Charleston, S. C.
-------
- 174 -
2. Reaction Temperature
An increase in reaction temperature from 600° to 700°F increased
the burn-off rate by a factor of 7 to 10. Compare Run 3 with 4,
and Run 5 with 6.
3. Carbon Properties
Carbons used in Runs 11 and 12 were specially modified to
improve burn-off resistance. These carbons exhibited about half
the burn-off rate produced in comparable runs with untreated
materials. Compare Runs 11 and 12 with Runs 6 and 8. Gas
analysis made in Run 11 showed complete conversion of SOs to S02-
WESTVACO CORPORATION
Charleston Research Center N. Charleston, S. C.
-------
- 175 -
CONCLUSIONS
The bench scale data that have been obtained indicate that activated
carbon can be employed to chemically reduce SOs back to S02 with essenti-
ally 100% efficiency under concentration and space velocity conditions
compatible with automotive applications.
At low temperatures, $03 is completely picked up by carbon by a sorption
mechanism, with conversion to S02 increasing with temperature and
approaching 100% at about 600°F. SOs sorbed at lower temperatures is
evolved as S02 at temperatures near 600°F.
At this time the major problem area in the practical application of
carbon reactors to $03 control appears to be that actual exhaust system
temperatures typically exceed 600°F and that at temperatures much above
this level, side reactions of carbon with other exhaust constituents may
lead to an unacceptable rate of burn-off. Figure 4, for example, shows
the predicted decrease in weight with time of an initial 4 pound reactor
charge, which is representative of that estimated for full scale
applications. Burn-off rates were assumed to be those found experi-
mentally at 600° and 700°F and it was also assumed that the rate of reac-
tion depends upon the carbon weight remaining, a point which has not
actually been confirmed. The time scale in this figure is not readily
converted to vehicle mileage since the reaction is so slow and a large
excess of reactants is present at any flow rate or engine speed. In
t^rms <~>f i">r>pra^i nn hn||rs then Tsble VI comn2res rsactor life assumin0
ultimate burn-off"levels'of 35% and 50%,
TABLE VI
PREDICTED REACTOR LIFE
3% 02, 10% H20, 10 ppm
Carbon
Standard 6x8 M
Standard 6x8 M
Modified 6x8 M
Temperature
(°F)
700
600
600
Rate
(%/hr.)
0.671
0.0938
0.0550
Time to
35% Burn-off
(hours)
60
460
815
50% Burn-off
(hours)
100
750
1,260
From these data it is quite clear that reactor operation must be
restricted to temperatures near 600°F and is likely to require special
oxidation resistant activated carbon.
WESTVACO CORPORATION
Charleston Research Center
N. Charleston, S. C.
-------
FIGURE 4
PREDICTED BURN-OFF OF 4 POUND CHARGE
o
+J
(J
ra
HI
en
c
Ol
c
o
JD
i-
•o
c
O
D.
4.0-1
3.0 -
2.0-
1.0-
Gas Composition: 10 ppm
10% H20
2000 ppm NO
Balance N2
700°F, Standard Carbon
100
.200
Theoretical for $03 Only
/ at 50 MPH
600°F, Modified Carbon
X
600°F, Standard
Carbon (without NO)
400
600
800
1000
Hours
WESTVACO CORPORATION
Charleston Research Center
N. Charleston, S. C.
-------
- 177 -
RECOMMENDATIONS
Laboratory results suggest that chemical reduction by activated carbon
is a promising candidate technique for S03 emission control. Full scale
prototype tests would, therefore, appear warranted unless projected
reactor system costs can clearly be shown to disqualify this approach.
The reactor cost alone is expected to be low since materials of construc-
tion and fabrication would be similar to an ordinary muffler. It is
further possible that, this reactor might provide sufficient sound attenu-
.ation (in combination with the catalytic converter) to replace the
muffler altogether, with consequent cost savings.
In terms of cost it is likely that carbon reactor feasibility may hinge
on meeting the temperature control requirements which are indicated by
present data. A relatively low cost control system is suggested in
Figure 5. Temperature above the desired control point, acting perhaps on
a bimetallic coil, would close a butterfly valve in the direct exhaust
pipe and divert gases through a finned heat exchanger pipe to decrease
gas temperature. Calculations could be made to estimate the required
length of exchanger to maintain 600°F at the reactor.
Should prototype tests be performed Table VII lists recommended reactor
configuration parameters based on conversion and pressure drop
*! vr •* 4- -J «. »-. f
.I UC.IUIU)
TABLE VII
RECOMMENDED REACTOR PARAMETERS
Reactor Bed Volume - 225-300 in.2
Wt. Carbon Charge - 4-5 pounds
Bed Depth (D) - 1" < D< 2.5"
Average Bed Face Area - >120 in.2
Granular C Mesh Size - 6x8 to 8x14
WESTVACO CORPORATION
Charleston Research Center N. Charleston, S. C.
-------
- 178 -
FIGURE 5
EXHAUST GAS TEMPERATURE CONTROLLER
Engine
Heat
Exchanger
Loop
I I I I I 1 I
I 1 I I I I I
Catalytic
Converter
Thermostatic Valve
I
Control Point
— Carbon Reactor
Charleston Research Center
WESTVACO CORPORATION
N. Charleston, S. C.
-------
- 179 -
APPENDIX E
RAW DATA TABLES
-------
TABLE E-l
EMISSIONS FROM THE 1974 CHEVROLET
Gaseous Emissions,
Fuel Sulfur
Level. Wt. X Test Mode
0.065
1975
20 mln. Idle
FTP
(8/t
96 km/h
0.032
1
V
1975
1975
20 mln. idle
FTP
FTP
est)
(D*
(2)
(3)
(4)
(g/test)
96 km/h
1
1975
FTP
(D*
(2)
(3)
(4)
g/km
CO
7.
37.
1.
4
7.
7.
84.
1.
!
7.
5
3
4
1
6
6
3
3
HC
0
4
0
0
0
4
0
0
.89
.71
.05
1
.88
.78
.40
.08
i
.81
SOx
1
4
1
1
0
4
0
1
.31
.02
.19
1
4-
.31
.93
.10
.79
i
.12
S02 Emissions
TECO,
Dilute
Exhaust
0.5
2.6
0.2
I
0.3
0.2
1.6
0.1
1
n.?
g/km
H202
Bubbler,
Raw Exhaust
0.8
0.15
I
0.4
0.075
1
% of
TECO,
Dilute
Exhaust
220
220
150
i
150
200
270
160
I
99(1
Fuel S
H202
Bubbler ,
Raw Exhaust
68
124
i
68
123
I
Filter ,
Dilute
Exhaust
0.002
0.000
0.001
0.002
0.001
0.001
0.001
0.002
0.033
0.001
0.001
0.001
0.001
n nn9
S04° Emissions
g/km
Goks^yr-Ross ,
Raw Exhaust
<0.02
<0.002
<0.002
<0.002
<0.002
0.05
<0.002
<0.002
<0.002
<0.002
% of Fuel S
Filter ,
Dilute
Exhaust
0.6
0
0.5
1
0.5
0.5
0.3
1
4
1
1
1
1
1
Goksrfyr-Ross ,
Raw Exhaust
<1
<1
<1
<1
<1
5
<2
<2
<2
<2
__
00
o
* CO, HC, NOx, and S02 values given represent an average over the two hour test.
-------
TABLE E-2
EMISSIONS FROM THE 1974 MAZDA RX-4
S02 Emissions
Gaseous Emissions. TECO,
Fuel Sulfur
Level, Wt. % Test Mode
0.065 1975 FTP
20 min. idle (g/test)
96 km/h (1)*
1 (2)
(3)
i (4)
1975 FTP
0.032 1975 FTP
20 min. Idle (g/test)
96 km/h (1)*
1 (2)
(3)
4- (4)
1975 FTP
CO
18.6
1.2
36.5
j
19.6
18.5
0.6
15.7
I
1
17.6
g/km
HC
2.47
0.06
3.32
1
2.43
2.51
0.18
7.68
1
1
2.05
Dilute
NOx
0.65
0.42
0.32
1
0.74
0.97
0.59
0.5
1
1
0.84
Exhaust
0.4
1.0
0.4
I
0.6
0.7
1.0
0.5
\t
0.6
R/km
H2
°2
Bubbler
Dilute
Exhaust
0.3
1.6
0.14
!
0*3
0.1
0.8
1
1
0.2
Raw
Exhaust
1.0
0.13
I
0.6
0.097
1
1
%
TECO,
Dilute
Exhaust
200
130
360
i
310
690
250
660
1
vl
590
of Fuel
S
SOA Emissions
H202 ,
Bubbler
Dilute
Exhaust
150
200
119
1
150
100
200
I
xl'
200
Raw
Exhaust
130
112
1
150
135
1
4?
Filter,
Dilute
Exhaust
0.000
0.000
0.000
0.000
0.000
0.000
0.002
0.003
0.002
0.000
0.000
0.000
0.000
0.002
?/km
Goks^yr-Ross ,
Raw Exhaust
<0.02
<0.002
<0.002
<0.002
<0.002
__«.___
<0.02
<0.002
<0.002
<0.002
<0.002
X of Fuel S
Filter,
Dilute
Exhaust
0.0
0.0
0.0
0.0
0.0
0.0
0.7
2
0.2
0.0
0.0
0.0
0.0
1
Goksrfyr-Ross
Raw Exhaust
..
<2
<1
-------
TABLE E-3
EMISSIONS FROM THE 1974 HONDA CVCC
S02 Emissions
g/km
% of
H207
Fuel Sulfur
Level. Ut. % Test Mode
0.065 1975 FTP
20 min. idle (g/test)
96 km/h (1)*
4- (2)
80 tan/h (1)*
^ (2)
1975 FTP
0.032 1975 FTP
20 min. idle (g/test)
88 km/h (1)*
4- (2)
1975 FTP
Gaseous Emissions,
CO
2.2
5.4
0.1
1
0.1
*
2.6
2.3
6.4
0.1
4
2.2
g/km
HC
0.53
2.82
0.01
4-
0.02
4-
0.66
0.60
2.62
0.01
4-
0.47
NOx
0.99
0.67
0.69
4
1.21
1.41
1.22
0.49
4-
1.38
TECO,
Dilute
Exhaust
0.2
0.7
0.1
J,
Oil
t
0.2
0.1
0.7
0.1
Jf
0.1
Bubbler
Dilute
Exhaus t
0.07
0.26
0.08
^L
0.08
0.05
0.04
^
0.02
Raw
Exhaust
0.60
0.10
1
0*09
_L
0.4
0.04
A
TECO,
Dilute
Exhaust
180
180
130
1
150
180
180
350
290
i
180
Fuel S
S04™ Emissions
H202
Bubbler
Dilute
Exhaust
60
65
105
1
V
70
90
120
*
36
Raw
Exhaust
150
130
±
125
1
200
120
A.
g/km
Filter,
Dilute
Exhaust
0.001
0.000
0.000
0.000
0.000
0.000
o.boo
0.000
0.000
0.000
0.000
0.000
**
Goksrfyr-Ross
Raw Exhaust
0.063
0.003
0.003
0.003
0.006
0.033
<0.001
<0.001
% of Fuel S
Filter,
Dilute
Exhaust
0.6
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
**
Goksrfyr-Ross
Raw Exhaust
10
3
3
3
6
11
<2
<2
—
**
CO, HC, NOX and S02 values given represent averages of the one hour tests conducted.
Spurious H20 condition in lines.
-------
TABLE E-4
Fuel Sulfur
Level. Wt. %
OU7
0.35
EMISSIONS FROM THE 1974 PEUGEOT DIESEL
S02 Emissions
g/tan
Gaseous Emissions,
Test Mode
1975 FTP
20 mln. Idle (g/test)
96 tan/h (1)
1 (2)
(3)
I (4)
1975 FTP
1975 FTP
20 min. idle (g/test)
96 km/h (1)
1 (2)
(3)
I (4)
1975 FTP
CO
2.13
26.9
0.35
0.30
0.29
0.27
1.93
1.93
25.9
0.45
0.54
0.55
0.57
2.21
g/km
HC
0.39 0.90
5.68 1.77
0.12 0.79
0.12 1.11
0.10 1.15
0.10 1.14
0.36 0.94
0.35 0.98
4.21 1.75
0.07 1.01
0.09 1.17
0.10 1.18
0.09 1.19
0.38 0.91
NOX
0
0
0
0
0
0
0
0
1
0
0
0
0
0
TECO,
Dilute
Exhaust
.076
.18
.047
.081
.080
.066
.039
.39
.12
.39
.37
.36
.35
.33
0
1
0
0
0
1
0
0
H202
Bubbler,
Dilute
Exhaust
.26
.24
.13
I
4
.17
.35
.10
.24
1
1
.24
% of
TECO,
Dilute
Exhaust
25
20
28
33
34
28
12
60
55
92
71
68
67
55
Fuel S
H202
Bubbler,
Dilute
Exhaust
85
135
57
1
*
56
54
54
48
1
1
40
Filter,
Dilute
Exhaust
0.011
0.047
0.005
0.005
0.006
0.006
0.023
0.012
0.070
0.012
0.009
0.009
0.011
0.013
SOu" Emissions
g/km
Goks^yr-Ross ,
Raw Exhaust
2
— — 3
— « 1
___ 1
— — — 1
»— 1
5
1
2
— — . 1
1
1
1
1
% of Fuel S
Filter, „
Dilute Goks^yr-Ross,
Exhaust Raw Exhaust
.4
.4
.9
.4
.8
.7
.1
.2
o _
8 —
^ -^
.1
• 3 —
(4
00
u
Goks^yr-Ross technique could not be used with diesel because particulates in the exhaust blocked the frit.
-------
TABLE E-5
EMISSIONS FROM THE BASE CASE PELLETED OXIDATION CATALYST SYSTEM WITH AN AIR PUMP
SO 2 Emissions
g/km
Fuel Sulfur
Level, Wt. % Test Mode
TURNPIKE DRIVING PRECONDITIONING.
0.032 1975 FTP
20 min. idle (g/test)
96 km/h (1)
(2)
(3)
(4)
1975 FTP
0.012 1975 FTP
20 min. idle (g/test)
96 km/h (1)
(2)
(3)
(4)
1975 FTP
CITY DRIVING PRECONDITIONING
0.032 1975 FTP
20 min. idle (g/test)
96 km/h (1)
(2)
(3)
(4).
1975 FTP
0.012 1975 FTP
20 min. idle (g/test)
96 km/h (1)
(2)
(3)
(4)
1975 FTP
Gaseous Emissions,
g/km
CO
2.27
0.89
0.036
0.027
0.022
0.029
2.24
0.02
0.042
0.045
0.029
0.042
2.68
2.20
0.15
0.029
0.031
0.036
0.021
1.49
3.17
2.72
0.039
0.023
0.021
0.028
2.11
HC
0.19
1.24
0.024
0.020
0.025
0.019
0.19
0.56
0.019
0.015
0.017
0.019
0.17
0.17
0.58
0.015
0.012
0.018
0.019
0.14
0.22
0.10
0.018
0.014
0.020
0.019
0.16
NOx
1.96
3.15
5.30
5.72
6.00
6.06
2.26
5.10
5.31
5.59
5.74
5.71
2.27
2.26
5.40
5.02
5.42
5.42
5.58
2.36
2.35
5.43
5.44
4.88
4.85
4.85
2.08
TECO,
Dilute
Exhaust
0.032
O.O(a)
0.12
0.08
0.07
0.05
0.021
__ _
O.O(a)
0.041
0.037
0.037
0.051
0.012
0.048
O.O(a)
0.091
0.057
0.062
0.050
0.020
0.044
0.26
0.057
0.026
0.026
0.026
0.023
H202
Bubbler,
Dilute
Exhaust
0.031
0.34
0.085
1
4
0.028
0.009
0.24
0.046
1
4
0.012
0.028
0.32
0.072
1
1
0.028
0.0033
0.11
0.030
1
4
0.0
% of Fuel S
TECO,
Dilute
Exhaust
29
0
144
94
80
60
20
___
0
122
. 107
105
145
30
47
0
106
64
70
55
20
105
108
164
78
78
78
58
H202
Bubbler,
Dilute
Exhaust
28
63
100
1
+
26
23
100
133
1
4
30
28
34
82
1
1
27
8
47
91
I
1
0.0
S04°
g/km
Goksrfyr-
Filter, Ross,
Dilute Raw
Exhaust Exhaust
0.011
0.026 0.008
0.104-^. ...
0.062— >0>0004
0.047-—-^...
0.038
0.0027
0.0015
0.021 0.008
0.052-^
0.033-— ^0>0009
o.ost-—.^
0.033-—^
0.0016
0.0094
0.029 0.0
0.093
0.050-— >
0.050-~_ . „-.,
0.047>°-0006
0.0017
0.0044
0.015 0.008
0.050-—-^.
0.021— ^^
0.023— --....
0.020- — "O'O011
0.0015
Emissions
% of Fuel S
Filter,
Dilute
Exhaust
6.7
3.3
84'
47- ~~
37-
30-
1.7
4.6
5.8
103-
65^
59-
61 '
2.7
6.1
2.1
72
38-—-—
38
34——"
1.1
7.0
4.0
96
41- — —
46-
41 — '
2.2
Goks^yr-
Ross,
Raw
Exhaust
—
1.1
OQ
. J
"=^""
___
2.5"
~^=— 7 1
—- t, * JL
—
—
0.0
On
. £.
--^O.S
___
2.5
1.1
^__
2.2
Sulfur
Balance ,
Bubbler +
Filter
35
66
150
I
4
28
28
106
205
1
4
33
34
36
128
1
4
28
15
51
147
1
4
2
(a) Below detection limit.
-------
TABLE E-6
EMISSIONS
FROM THE BASE CASE MONOLITHIC OXIDATION CATALYST SYSTEM WITH AN AIR PUMP
SO? Emissions
8/km
Gaseous Emissions,
Fuel Sulfur
Level, Wt. %
TURNPIKE DRIVING
On*io
. \j jt,
20
0.012
20
Test Mode
PRECONDITIONING
1975 FTP
rain, idle (g/test)
96 km/h (1)
1 (2)
1 8)
1975 FTP
1Q7C FTP
X7 / J r Lc
min. Idle (g/test)
96 km/h (1)
I (2)
(3)
J- (4)
1975 FTP
CO
2.16
0.69
0.22
0.22
0.20
0.20
2.13
1 SS
J . J J
0.36
0.18
0.16
0.16
0.16
4.08
K/km
HC
0.19
o'.36
0.030
0.031
0.032
0.031
0.19
0.24
0.39
0.033
0.035
0.031
0.025
0.31
NOy
1 75
j. . / j
5.27
5.49
5.50
5.49
5.64
2.11
1 79
j. • t y
6.22
4.85
5.21
5.34
5.13
1.94
TECO,
Dilute
Exhaust
0 049
0.19
0.042
0.040
0.040
0.040
0.017
0 016
0.13
0.018
0.010
0.010
0.014
0.015
H202
Bubbler,
Dilute
Exhaust
0 078
0.23
SAMPLE
LOST
1
0.068
Oft/0
• UHO
0.054
0.032
1
1
*
0.035
Z of Fuel S
TECO,
Dilute
Exhaust
42
25
50
48
48
47
14
OC
•Jj
43
58
32
32
45
35
H202
Bubbler,
Dilute
Exhaust
fiL
OH
30
SAMPLE
LOST
I
55
1ft/
.LUH
19
103
1
1
V
75
S04"
ft/km
Goksdyr-
Fllter, Ross,
Dilute Raw
Exhaust Exhaust
0.032 0.024
0.060^—^.
0 . 047-~^> 0 • OWO
S;°,;5> 0.0023
0.025
0.027 0.0
0.040~~—
O.oi9^^>" 0-0031
0.019 — ~>n
0.019" — -^u.OOZZ
0.0038
Emissions
X of
Filter,
Dilute
Exhaust
12
2.5
47
38
37
36
14
1 7
J./
6.4
83
40
40
41
5.8
Fuel S
Goksrfyr-
.Ross ,
Raw
Exhaust
2.0
21
.4
1.8
0.0
6.5
4.7
Sulfur
Balance ,
Bubbler +
Filter
7fi
/ o
32
1
1
69
121
25
154
\ '
81
CITY DRIVING PRECONDITIONING
0.032
20
0.012
20
1975 FTP
mln. Idle (g/test)
96 km/h (1)
I (2)
(3)
1 (4)
1975 FTP
1975 FTP
min. idle (g/test)
96 km/h (1)
1 (2)
(3)
I (4)
1975 FTP
3.42
0.11
0.23
0.22
0.20
0.20
2.81
1.16
0.69
0.24
0.24
0.24
0.22
0.60
0.37
0.039
0.039
0.038
0.035
0.31
0.17
0.37
0.035
0.035
0.035
0.030
2.11
5.71
6.07
6.15
5.93
5.44
1.93
1.91
6.19
5.37
5.37
5.51
5.36
0.015
0.13
0.044
0.046
0.044
0.046
0.014
0.028
0.0
0.024
0.024
0.016
0.020
0.014
0.098
0.054
0.062
1
1
+
0.110
0.043
0.022
0.032
1
I
0.034
14
16
53
54
53
54
12
60
0
72
72
49
60
34
82
7
73
1
i
88
93
7
97
1
1
78
0.025
0.037 0.027
0.057--
0.050 — -^".OOll
0.048- — =>n
0 . 050 — ^^ ' "
0.027
OftrtQQ _„_ __
0.010 0.044
0.031-- — __
0 . 018- ^-->~Q -0013
S;JJJ>°'°o«
0.0044
15
2.9
45
39
38
39
15
13
2.4
62
36
32
32
6.5
--_
2.2
0.9
0.5
___
10.0
2.5
2.5
97
10
113
I
1
103
106
9
138
I
1
85
-------
TABLE E-7
EMISSIONS FROM THE PELLETED OXIDATION CATALYST SYSTEM WITH LIMITED SECONDARY AIR
All Tests Conducted With 0,
SO? Emissions
8/km
Gaseous Emissions,
TURNPIKE
20 min.
Test Mode
CO
8/km
HC
DRIVING PRECONDITIONING
1975 FTP
idle (g/test)
96 km/h (1)
i (2)
(3)
4. w
1975 FTP
2.37
0.34
0.045
0.086
0.016
0.052
3.42
0.17
0.43
0.022
0.019
0.020
0.016
0.21
NOX
2.22
2.91
5.05
5.33
5.77
5.77
2.21
TECO
Dilute
Exhaust
0.14
O.O(a)
0.21
0.16
0.10
0.099
0.040
H202
Bubbler,
Dilute
Exhaust
0.11
0.0
0.16
1
I
0.038
% of
TECO,
Dilute
Exhaust
130
0
244
179
109
105
37
,032 Wt. % Sulfur Fuel
SOi"
Fuel S
H202
Bubbler
Dilute
Exhaust
105
0.0
178
1
35
8/km
Goksrfyr-
Filter, Ross,
Dilute Raw
Exhaust Exhaust
0.0019
0.0003 0.022
?:!£>«•"«>
u * u J.D" ^^ « nnn7
o.oie^0-0007 .
0.0006
Emissions
% of Fuel S
Goksrfyr-
Filter, Ross,
Dilute Raw
Exhaust Exhaust
1.2
0.0 2.5
16 J^>0'2
11- — — -« A
11 _>"•*
0.4
Sulfur
Balance,
Bubbler +
Filter
106
2.5
199
I
1
35
CITY DRIVING PRECONDITIONING
20 min.
1975 FTP
idle (g/test)
96 km/h (1)
1 (2)
(3)
J- W
1975 FTP
3.44
1.38
0.030
0.37
6.34
3.73
4.3
0.19
0.69
0.010
0.007
0.039
0.019
0.21
1.72
1.89
2.14
1.57
1.14
1.35
1.70
0.19
0.13
0.15
0.14
0.21
0.066
0.045
0.14
0.046
0.175
. 1
4
0.012
197
30
213
212
312
99
45
145
10
257
1
1
11
.0.0028
0.002 0.006
0.038^--...,,. OQ05
0.006---"^*^'
0.001-———.. nnriA
0 . OOl-^"^*^ '
0.0010
1.8
0.3 0.9
32 ~~~~x)
5 . 8-- —
1.0 -— _ ..
1.0 --^O^*
0.7
147
11
267
1
1
12
I
H
00
(a) Below detection limit.
-------
TABLE E-8
EMISSIONS FROM THE MONOLITHIC OXIDATION CATALYST SYSTEM WITH LIMITED SECONDARY AIR
All Tests Conducted With 0.032 Wt. %
S02
g/km
Gaseous Emissions,
8/km
Test
Mode
CO
HC
1975 FTP 4.90 0.22
20 min. idle (g/test) 0.26 0.45
96 km/h (1) 0.71 0.066
1 (2) 0.30 0.055
(3) 0.29 0.054
4 (4) 0.29 0.053
1975 FTP 5.73 0.33
NOx
TECO,
Dilute
Exhaust
1.99 0.028
5.24 0.12
5.06 0.066
5.06 0.054
5.06 0.060
5.20 0.060
1.92 0.125
Sulfur Fuel and Turnpike Driving Preconditioning
Emissions S04° Emissions
% of Fuel S
H202
Bubbler, TECO,
Dilute Dilute
Exhaust Exhaust
0.036 24
0.22 17
0.088 86
1 86
80
v- 81
0.133 106
TABLE E-9
EMISSIONS FROM THE PELLETED OXIDATION CATALYST
1975
20 min. idle
96 ki
1975
FTP
5 (g/test)
p/h (1)
(2)
(3)
(4)
FTP
All
2.21
0.28
0.068
0.056
0.056
0.081
1.78
Tests Conducted
0.18 2.18
0.75 4.23
0.017 6.21
0.031 5.86
0.012 6.21
0.017 6.97
0.13 2.19
With 0.032
0.047
0.0
0.099
0.050
0.040
0.072
0.036
Wt. %
0.020
0.24
0.092
1
0.009
H202
Bubbler,
Dilute
Exhaust
30
31
118
I
107
SYSTEM OPERATED AT HIGH
Sulfur fuel
49
0
116
58
47
79
37
and Turnpike
20
44
106
1,
g/km Z of Fuel S
Filter
Dilute
Exhaust
0.0028
0.028
0.005IT
0 *0025~
0 .0043"
0.0028
SPACE VEI
Driving
0.018
0.000
0.050-
0.055"
0.008
Goksjiyr- Goksrfyr-
, Ross, Filter, Ross,
Raw Dilute Raw
Exhaust Exhaust Exhaust
T c
0.019 2.7
>0.007« J;J>
><0.0024 2l6
If.
.OCITY (WITH AN AIR PUMP)
Preconditioning
i ^
0.0 0.0
~>0 0036 89 — ~~^-—
— "^ ' 41
>0.0045 4?!^==
5.6
1.8
•6.6
'2.2
0.0.
.2.9
-3.5
Sulfur
Balance,
Bubbler +
Filter
32
33
121
109
i
t-i
00
^j
I
33
44
158
15
-------
TABLE E-10
EMISSIONS FROM THE MONOLITHIC OXIDATION CATALYST SYSTEM OPERATED AT LOW SPACE VELOCITY (AND WITH AN AIR PUMP)
Test
Mode
20 min. idle (g/test)
96 km/h (1)
I
*
1975
(2)
(3)
(4)
FTP
EMISSIONS FROM
1975
FTP
20 min. idle (g/test)
96 km/h (1)
xl
1975
(2)
(3)
(4)
FTP
All
Gaseous
CO
0.24
0.018
0.041
0.018
0.018
1.70
Tests
Conducted
Emissions ,
g/km
HC
0.19
0.011
0.006
0.007
0.006
0.16
NOx
15.1
5.39
5.69
5.84
5.62
2.14
THE PELLETED OXIDATION
2.20
0.27
0.050
0.056
0.044
0.050
2.60
0.11
0.34
0.011
0.011
0.012
0.015
0.15
1.93
3.23
5.89
6.03
5.89
5.89
1.79
With 0.032 Wt. %
TECO,
Dilute
Exhaust
0.00
0.017
0.008
0.010
0.017
0.014
CATALYST
0.048
0.12
0.14
0.076
0.070
0.070
0.025
S02
g/km
H202
Bubbler
Sulfur Fuel
Emissions
% of
, TECO
Dilute Dilute
Exhaust
0.06
0.040
1
1
0.048
Exhaust
0
24
12
14
23
13
TABLE E-ll
and Turnpike
Fuel S
H202
Bubbler,
Dilute
Exhaust
7
57
1
4
44
Driving
Filter
Dilute
Exhaust
0.20
0.080~~
0.050"
0.047~~
0.044--
0.013
Preconditioning
S04 Emissions
_g/km % of
Goksrfyr-
, Ross, Filter,
Raw Dilute
Exhaust Exhaust
0.032 15
->Q (joe, 77
-"^9.0055 4^ .
--~> 44
-^0.0065 41__
8.1
SYSTEM OPERATED AT HIGH SPACE VELOCITY WITH A Pt CATALYST (AND
0.016
0.0
0.114
1
^
0.018
49
22
166
96
88
88
25
17
0.0
143
V
18
0.034
0.0018
0.150~-
0.059-
0.049-
0.049"
0.007
______ 9i
0.0 0.2
>0 0016 125
-^°-°°lb 50 '
>0 0008 4r~~~
,--=•0.0008 41^_-
4.4
Fuel S
Goks •Syr-
Ross,
Raw
Exhaust
2.4
-_=-fi
6.1
AN AIR PUMP)
0.0
]>-l 3
'
^-=*"0 6
Sulfur
Balance,
Bubbler +
Filter
111
22
109
52
40
0.2
207
22
00
00
-------
TABLE E-12
EMISSIONS FROM THE MONOLITHIC OXIDATION CATALYST SYSTEM WITH Pt CATALYSTS (AND AN AIR PUMP)
Test Mode
1975 FTP
20 min. Idle (g/test)
96 km/h (1)
(2)
(3)
(4)
1975 FTP
All Tests Conducted
With 0.032 Wt. % Sulfur Fuel and Turnpike Driving Preconditioning
S02 Emissions SO& Emissions
8/km
Gaseous Emissions,
g/km
CO
1.60
0.43
0.080
0.080
0.075
0.068
9 no
HC
0.19
0.13
0.010
0.009
0.015
0.004
n 17
NOx
5.27
3.63
3.98
3.84
3.63
1 Afi
TECO,
Dilute
Exhaust
0.030
0.12
0.016
0.010
0.016
0.002
n niA
H202
Bubbler,
Dilute
Exhaust
0.040
0.23
0.043
1
n OAA
% of
TECO,
Dilute
Exhaust
27
15
21
13
21
28
1 9
Fuel S g/km
H202 Goksrfyr-
Bubbler, Filter, Ross,
Dilute Dilute Raw
Exhaust Exhaust Exhaust
QA n AOA
29 0.20 0.023
1
1 " ^— -^"'*0 . 0019
•?Q n nA9
% of Fuel S
Goks^yr-
Filter, Ross,
Dilute Raw
Exhaus t Exhaus t
14
16 2.0
a*-—---"2*1-7
33__H>1.7
Sulfur
Balance ,
Bubbler +
Filter
48
45
94
1
TABLE E-13
EMISSIONS FROM THE PELLETED OXIDATION CATALYST
1975 FTP
20 min. idle (g/test)
96 km/h (1)
1 (2)
(3)
4- (4)
1975 FTP
2.93
0.27
0.025
0.030
0.022
0.019
0.98
0.11
0.34
0.006
0.000
0.006
0.002
0.068
1.53
2.87
3.76
3.90
4.11
3.90
1.44
0.083
0.0
0.099
0.078
0.076
0.058
0.037
SYSTEM OPERATED Al
0.045
0.011
0.102
1
i
0.004
79
0
119
91
90
67
35
41
1.3
120
3.7
0.010
0.000
0.0047
0.0
47
1.3
163
6.6
-------
TABLE E-14
EMISSIONS FROM THE MONOLITHIC OXIDATION CATALYST SYSTEM OPERATED AT LOWER THAN NORMAL TEMPERATURE (AND WITH AN AIR PUMP)
All Tests Conducted With
0.032 Wt. % Sulfur Fuel and Turnpike Driving Preconditioning
S02 Emissions S04° Emissions
8/km
Gaseous Emissions,
g/km
Test Mode
1975 FTP
20 min. idle (g/test)
96 km/h (1)
1 (2)
(3)
4 (4)
1975 FTP
EMISSIONS
1975 FTP
20 min. idle (g/test)
96 km/h (1)
1 (2)
(3)
\ (4)
1075 TTT^
CO
4.57
3.45
0.24
0.24
0.21
0.24
3.11
FROM THE
2.34
0.04
0.031
0.025
0.019
0.019
i «;•>
HC
0.36
0.72
0.036
0.036
0.033
0.033
0.28
PELLETED
0.20
0.24
0.010
0.020
0.006
0.011
nil
NO*
H202
TECO, Bubbler,
Dilute Dilute
Exhaust Exhaust
1.89 0
11.4 0
5.44 0
5.38 0
5.51 0
5.79 0
2.08 0
OXIDATION
2.36
4.74
6.05
5.90
6.05
5.90
9 1O
0
0
0
0
0
0
n
.099
.70
.032
.030
.030
.036
.077
CATALYST
.045
.0
.090
.076
.060
.040
non
0.089
0.44
0.029
J
0.090
TABLE
% of Fuel S
8/km
% of Fuel S
H202 Goksrfyr- Goks^yr-
TECO, Bubbler, Filter, Ross, Filter, Ross,
Dilute Dilute Dilute Raw Dilute Raw
Exhaust Exhaust Exhaust Exhaust Exhaust Exhaust
85
81
40
39
38
46
67
E-15
SYSTEM OPERATED
0.015
0.26
0.072
i
n m i
47
0
114
95
75
50
10
73
51
I
75
WITH HIGH
15
46
91
1
Ifi
0
0
0
0
0
0
0
METAL L
0
0
0
0
0
n
.0068
.037 0.014
:S>o.oo68
:S£>0.0039
m A — — —
OADING CATALYST
ni9 ____
.012 0.0
:;°p-o.oo2i
;°JP>o.oo3o
nnA _
3.8
2.8 0.9
48 — --==^5'9
8.3
(AND AN AIR PUMP)
12.0
1.4 0.0
t n —
Sulfur
Balance,
Bubbler +
Filter
77
54
84
83
27
47
146
VO
o
19
-------
TABLE E-16
EMISSIONS FROM THE PELLETED OXIDATION CATALYST SYSTEM WITH AN AGED CATALYST (AND AN AIR PUMP)
All Tests Conducted With 0.032 Wt. % Sulfur Fuel
S02 Emissions
R/km
Test Mode
Gaseous
CO
TURNPIKE DRIVING PRECONDITIONING
1975 FTP
20 min. idle (g/test)
96 km/h (1)
1 (2)
(3)
j (4)
1975 FTP
1.84
0.02
0.081
0.068
0.10
0.081
1.47
Emissions,
g/km
HC NO,
0.18 2.09
0.76 1.85
0.031 6.03
0.031 6.16
0.025 6.40
0.031 6.46
0.17 2.06
TECO,
Dilute
Exhaust
0.025
O.O(a)
0.088
0.072
0.080
0.053
0.009
H202
Bubbler,
Dilute
Exhaust
0;023
0.0
0.091
V
0.0025
% of Fuel S
TECO,
Dilute
Exhaust
24
0
97
79
87
61
9
H202
Bubbler,
Dilute
Exhaust
22
0.0
101
1
1
V
2.4
S04° Emissions
g/km
Filter,
Dilute
Exhaust
0.0036
0.0003
0.102 —
0.039- —
0.030-—.
0.022- —
0.0007
Goksrfyr-
Ross,
Raw
Exhaust
0.014
>-Q 0015
>n nni^
_«__.._
% of Fuel S
Filter,
Dilute
Exhaust
.2.3
0.04
75 —
29
22
16
0.4
Goks^yr-
Ross,
Raw
Exhaust
1.7
==-1 1
n Q
— — —
Sulfur
Balance,
Bubbler +
Filter
24
0.04
136
f
1
2.8
CITY DRIVING PRECONDITIONING
1975 FTP
20 min. idle (g/test)
96 km/h (1)
1 (2)
(3)
4- (4)
107"; TTTP
2.93
0.08
0.12
0.11
0.087
0.087
1 -Ai
0.22 2.49
0.73 2.97
0.022 5.99
0.018 5.57
0.027 6.12
0.027 6.05
n.ifi Lin
0.030
O.O(a)
0.004
0.004
0.004
0.0
n.n?a
0.032
0.058
0.090
4
n.nifi
29
0
3i9(b)
4. Kb)
0(a,b)
7Q
29
11
97
1
1
IS
0.0066
0.003
0.039--"
0.03& — —
0.032-'"
n nnA
0.024
^O.OOOS
--"0.0005
4.0
0.4
61 — - —
27—
27" — — —
24— — —
? A
3.0
=>0.4
>-0.3
33
11
132
I
1
1R
(a) Below detection limit.
(b) Low values may have been due to air leak in TECO sampling line.
-------
TABLE E-17
EMISSIONS FROM THE MONOLITHIC OXIDATION CATALYST SYSTEM WITH AGED Pt/Pd CATALYSTS (AND AN AIR PUMP)
All
Tests Conducted With 0.032 Wt.
S02 Emissions
R/km
Gaseous Emissions,
Test Mode
CO
8/km
HC
TURNPIKE DRIVING PRECONDITIONING
20
1975 FTP
min. idle (g/test)
96 km/h (1)
1 (2)
(3)
* (4)
1Q-TC pTO
2.80
1.08
0.70
0.64
0.68
0.70
0 71
CITY DRIVING PRECONDITIONING
20
1975 FTP
min. idle (g/test)
96 km/h (1)
1 (2)
(3)
I (4)
1975 FTP
3.46
0.83
0.63
0.64
0.55
0.56
2.65
0.27
0.65
0.078
0.076
0.075
0.077
079
. tfi
0.44
0.67
0.061
0.053
0.053
0.055
0.20
_NOx
1.90
6.01
4.50
4.16
4.16
4.01
1 QA
i . 7H
1.41
5.19
3.93
4.43
4.57
4.71
1.62
TECO,
Dilute
Exhaust
0.091
0.38
0.057
0.045
0.045
0.049
Ons R
. uoo
0.121
0.32
0.060
0.030
0.025
0.205
0.064
H202
Bubbler,
Dilute
Exhaust
0.104
0.24
0.064
1
1
0117
. J.i.7
0.18
0.17
0.078
1
1
0.088
% of
TECO,
Dilute
Exhaust
81
51
74
60
60
66
77
/ /
103
39
75
36
32
31
53
% Sulfur Fuel
SOA Emissions
Fuel S
H202
Bubbler ,
Dilute
Exhaust
90
32
85
1
*
07
y I
144
20
96
1
1
72
g/km
Goksrfyr-
Fllter, Ross,
Dilute Raw
Exhaust Exhaust
0.0068
0.0033 0.024
0.039~~^^.
0.030""'--5"'0''''026
0.025^-^.
0 . 028 — ~^~® ' "010
0.014
0.0009 0.025
0.053 0.0063
0.031
Q.02S~~~-~>.
0.028^"^^ 0.0032
0.006
% of Fuel S
Goksrfyr-
Filter, Ross,
Dilute Raw
Exhaust Exhaust
4.1
0.3 2.2
34~H^>2.2
23~— ~--=^
25 — — -=""0-3
C A
8.0
0.1 2.0
45 5.3
25
24~-^-~^>
22-—- --->"2-6
3.4
Sulfur
Balance,
Bubbler +
Filter
94
32
112
1
1
102
152
20
125
1
*
75
s>
I
-------
TABLE E-18
VEHICLE DURABILITY TEST RESULTS OF PELLETED 85 CaO/10
Filter,
Trap Total
Distance Test Farticulate
(km) Description. (gin/km)
Base 64 km/h
Car*
0* 64 km/h
64 km/h
1 hr. @ Idle
2 hrs.@ 97 km/h
1 610* 1 hr. @ 64 km/h
1 790** 64 km/h
3 220 64 km/h
64 km/h
4 830 64 km/h
64 km/h
6 440 64 km/h
64 km/h
0.171
0.012
0.030
0.023 gm
0.007
0.003
0.004
0.003
0.004
0.000
0.004
0.003
0.007
Diluted Exhaust
Total
804-
(gm/km)
0.086
0.002
0.003
0.001 gm
0.001
0.006
0.001
0.001
0.001
0.001
0.001
0.004
0.002
Total Trap
Ca AP @ 64 km/h
(gm/km) (Pa)
____ ____
2.0xlO~5
2.5xlO~5
3.6xlO~4 gm
6.2xlO~5
1.7xlO~4
2xl0^5 9 950
IxlO"5 9 950
2xlO~5 1 320
IxlO"5 1 320
<3xlO~5 9 700
<3xlO"5 9 700
504 Fuel
Goksrfyr-Ross Economy
(gm/km) (mpg)
21.0
20.0
20.5
19.5
18.1
20.8
21.3
20.9
20.4
20.4
20.8
5102/5% Na20
Hot 1975 FTP
Fuel Economy
CO HC NOX Wt. Carbon Balance
(Km/km) (gin/km) (gm/km) (mpg) (mpg)
____ ____
__..
.._„
_____ _ _— _ _ __«—
-------
TABLE E-18 (cont.)
Filter, Diluted Exhaust
Hot 1975 FIP
Trap
Distance
(km)
9 650
12 870
17 600
24 000
30 720
35 890
42 400
Test
Description
64 km/h
64 km/h
64 km/h
64 km/h
64 km/h
64 km/h
75 FTP
64 km/h
64 km/h
75 FTP
64 km/h
64 km/h
75 FTP
64 km/h
64 km/h
75 FTP
64 km/h
64 km/h
75 FTP
Total
Particulate
(Km/km)
0.002
0.003
0.013
0.005
0.024
0.007
0.003
0.016
0.005
0.003
0.014
0.008
0.004
0.019
0.008
0.004
0.017
Total
(gm/km)
0.001
0.001
0.001
0.001
0.001
0.002
0.003
0.001
0.001
0.003
0.001
0.001
0.001
0.001
0.001
0.002
0.005
0.002
0.002
Total
Ca
(gm/km)
4xlO~5
4x10-5
<3xlO~5
<3xlO~5
<3xlO-5
2xlO-4
3xlO~5
<3xlO~5
2xlO~4
3x10" 5
<3xlO~5
IxlO"4
<3xlO~5
<3xlO-5
2xlO~4
7x10-5
7xlO-5
4xlO~4
Trap S0^~
AP @ 64 km/h Goks^yr-Ross
(Pa) (gm/km)
14 900
22 500
14 900
22 500
14 900
14 900
22 500
22 500
22 500
22 500
22 500
22 500
37 500-29 000
29 000
Fuel
Economy CO
14.5
16.5
16.3
15.7
16.3
16.9
1.32
14.6
14.6
0.73
14.5
15.6
1.12
14.5
14.5
0.80
14.0
14.4
1.31
Fuel Economy
HC NOX Wt. Carbon Balance
(gm/km) (gm/km) (mpg) (mpg)
0.22 2.59 10.0 9.81
0.36 1.91 9.79
0.15 2.54 9.76
0.33 2.00 9.49
0.33 2.76 9.23
-------
TABLE E-19
Trap
Distance
(km)
Base
Car
3980
11200
17400
Test
Description
64 km/h
64 km/h
'75 FTP
'75 FTP
64 km/h
64 km/h
'75 FTP
64 km/h
64 km/h
I 7 1 T?TP
/ j r ir
64 km/h
64 km/h
' 7 •; 1?TP
Filter,
Total
Particulate
(gm/km)
0.044
0.061
0.029
0.018
0.034
0.057
0.027 •
0.023
0.036
0 024
0.039
0.057
n nil
VEHICLE
DURABILITY TEST RESULTS OF CaO>3 CHIPS
, Diluted Exhaust
Total
Total
Trap
S04° Ca AP @ 64 km/h
(gm/km)
0.026
0.032
0.013
0.011
0.017
0.027
0.005
0.015
0.020
Onnc
. UU J
0.014
0.022
n mi
(gm/km)
<3xlO~5
<3xlO~5
-------
TABLE E-20
VEHICLE DURABILITY TEST OF PELLETED ZINC OXIDE SORBENT
Trap
Distance
(km)
Base
Case
5 300
9 700
15 900
Test
Description
64 km/h
64 km/h
FTP
64 km/h
64 km/h
FTP
64 km/h
64 km/h
FTP
64 km/h
64 km/h
FTP
Filter,
Total
Particulate
(gm/km)
0.066
0.090
0.080
0.035
0.044
0.071
0.028
0.039
0.049
0.029
0.042
0.076
Diluted Exhaust
Total
S04=
(gm/km)
0.026
0.035
0.036
0.020
0.024
0.036
0.018
0.023
0.036
0.017
0.022
0.042
Total
Zinc AP
(gm/km)
< 2. 8x10-5
<2.8xlO"5
<1.0xlO"4
8.71x10-5
4.04
30.2
4.04
<2.80
20.2
3.1
<2.8
17.9
Trap
@ 64 km/h
(Pa)
750
750
500
500
500
500
S04=
Goks«Syr-Ross
(gm/km)
0.027
0.034
—
0.029
0.023
0.033
0.034
0.044
0.046
Fuel
Economy CO
20.73
20.67
1.14
20.29
20.03
1.54
19.73
20.00
0.83
21.56
22.02
0.60
Hot 1975 FTP
Fuel Economy
HC NOX Wt. Carbon Balance
(gm/km) (gm/km) (mpg) (mpg)
__ _ ___
0.27 2.78 11.63 12.01
0.?.9 3.07 11.58 11.75
0.13 2.50 11.74 12.12
0.13 3.87 12.48 12.77
-------
TABLE E-21
VEHICLE DURABILITY TEST RESULTS OF RINGED 85 CaO/10 SiO?/5% Na,0
Filter, Diluted Exhaust
Trap
Distance Test
(km) Description
Base 64 km/h
Case 64 km/h
75 FTP
3 028 64 km/h
64 km/h
75 FTP
4 105 64 km/h
64 km/h
75 FTP
7 662 64 km/h
64 km/h
75 FTP
11 024 64 km/h
64 km/h
75 FTP
14 400 64 km/h
64 km/h
75 FTP
20 400 64 km/h
64 km/h
75 FTP
Total
Particulate
(gm/km)
0.065
0.084
0.038
0.008
0.009
0.020
0.007
0.008
0.013
0.008
0.011
0.012
0.019
0.024
0.018
0.033
0.038
0.017
0.019
0.026
0.021
Total
804"
0.033
0.032
0.025
0.004
0.005
0.011
0.004
0.006
0.006
0.004
0.006
0.005
0.009
0.014
0.004
0.016
0.022
0.006
0.009
0.013
0.008
Total
Ca
(gm/km)
<3xlO~5
<3xlO~5
Not
Submitted
< 3x10-5
<3xlO~5
<3xlO~5
<3xlO~5
<3xKT5
< 3x10" 5
<3xlO-5
<3xlO"5
< 1x10" 4
3xlO~5
3x10-5
2.6xlO~4
Trap
AP @ 64 km/h
(Pa)
500
500
622
622
500
500
1 000
1 000
3 000
3 000
3 700
3 700
4 200
4 200
Hot 1975 FTP
S04= Fuel Fuel Economy
GoksfSyr-Ross Economy CO HC NOX Wt. Carbon Balance
(gm/km) (mpg) (gm/km) (gm/km) (gm/km) (mpg) (mpg)
0.028 21.3
0.035 22.1
0.62 0.07 2.96 11.7 11.6
0.029 21.5
0.031 21.5
3.18 0.24 2.86 11.5 11.4
0.031 22.1
0.022 22.8
2.17 0.08 3.09 11.6 11.4
0.034 22.2
0.022 22.2
1<99 o.09 3.83 11.7 11.7
0.018 21.8
0.024 21.8
2.30 0.26 3.42 11.4 11.2
0.025 21.7
0.025 21.5
7.82 0.29 3.36 11.5 10.9
0.010 21.3
0.017 21.5
4.83 0.75 3.33 u.4 u.0
-------
TABLE E-22
3
Sorbent Volume = 13 cm
Feed
Sorbent
85 CaO/10 Si02/5 Na20
(Benchmark)
A1203
(Norton 4102)
BaO**
80 CaO/20 Si02
(Harshaw)
Benchmark
Ca 3103
(Micro-Cel)
MgO
(Calcined Mg(OH)2)
MgO
(Calcined Magnesite)
MgO
(Harshaw)
Gas Composition: C02 =
Weight,
Shape grams
pellets 11.0
3.2 mm
pellets 14.5
3.2 mm
granules 38 . 1
^3 mm
pellets 11.3
3.2 mm
pellets 24.7
3.2 mm
pellets 12.7
3.2 mm
pellets 5.7
3.2 mm
pellets 20.2
3.2 mm
spheres 11.7
3-5 mm
pellets 16.8
3.2 mm
LABORATORY
SCREENING OF SORBENT MATERIALS
Temperature = 480°C (370°C where indicated)
12%; H20 = 12%;
Time On
Feed, Hrs.
1
1.3
2.6
2.5
1.3
4.0
5.1*
8.8*
2.3
1.3
3.5
1.3
4.5
1.3
2.7
4.0*
7.4
1.3
2.7
4.2
S02 = 15 ppm;
H2S04
Out, ppm
0
1.9
0
0
0.6
0.9
2.9
0.1
0
0.1
0
0
0
0.3
0
0
0
1.0
3.2
H2S04 = M ppm;
Sulfate
Trapping
Efficiency, %
100
55
100
100
84
73
12
97
100
97
100
100
100
92
100
100
100
74
17
Space Velocity - 100 000 v/v/h
NOX = 500 ppm;
S02 Out,
1
14
13
15
15
15
15
1
13
14
15
15
15
15
15
15
15
15
02 = 3%; balance N2
S02
Trapping Relative Ease of
Efficiency, % Pilling
93 good
7
15 fair
0
0
0
0
93 good
15 poor
7
0 poor
0
0
0
0
0
0
0
00
I
-------
TABLE E-22(CONT.)
Sorbent
85 MgO/10 S102/5 Na20
CaC03
(Marble)
ZnO
(Harshaw)
Benchmark
CaO
(3% aluminum stearate
binder)
A1203
(MCB activated)
Mn02
A1203
(washcoat, Corning)
Weight,
Shape grams
granules 5.7
6/10 mesh
chips 20
7/10 mesh
extrudate 17.7
5 mm
pellets 13.3
3.2 mm
pellets 10.6
3.2 mm
pellets 13.1
3.2 mm
granules 26.9
10/16 mesh
coated 6.4
honeycomb
Time On
Feed, Hrs.
2.2
1.3
3.7
1.2
3.9
4.5*
8.2*
3.9*
8.4*
1.0*
3.2*
4.6*
1.2
3.6
4.9
1.2
2.7
3.9
1.3
4.1
4.4
H2S04
Out , ppm
0
1.0
1.9
0
0.7
0
1.4
0
0
0
0
1.1
0.2
2.6
3.4
0
0
0.6
2.0
4.1
4.7
Sulfate
Trapping
Efficiency, %
100
74
50
100
82
100
58
100
100
100
100
67
94
21
0
100
100
82
64
25
14
S02 Out,
ppm
7.9
15
15
14
15
15
3.1
9.7
10
13.2
15
15
15
6.7
11.9
11.9
15
S02
Trapping Relative Ease of
Efficiency, % Pilling
\
.47 unsuccessful
0 ' unsuccessful
0
j
0
0
79 good
35
33 fair
12
0
0
0
55
21
21
0
•
* 370°C
** Rapidly hydrated to Ba(OH)2 and dissolved
-------
TABLE E-23
LABORATORY SCREENING OF SORBENT MATERIALS EFFECT OF SPACE VELOCITY
3
Sorbent Volume = 13 cm Temperature = 480°C (370°C where indicated)
Sorbent
Benchmark
MgO
(Calcined Magnesite)
ZnO
(Harshaw)
Benchmark
Benchmark*
MgO*
(Calcined Magnesite)
Mn02*
Benchmark
Weight,
Shape grams
pellets 13.5
3.2 mm
spheres 11.9
3-5 mm
Extrudate 17.7
5 mm
pellets 7.9
3.2 mm
pellets 8.2
3.2 mm
spheres 11.4
3-5 mm
granules 25.5
10/16 mesh
pellets 7.9
3.2 mm
Time On
Feed. Hrs.
9.6
12.6
14.5
8.7
11.6
9.3
11.8
1.2
3.2
4.2
1.2
2.8
4.7
1.2
3.1
5.0
1.2
3.8
6.7
10.6
12.6
14.1
H2S04
Out , ppm
0.6
0
0
0.7
2.0
1.9
0
0
0
0.1 •
0
0
3.5
0.2
1.2
2.4
0
2.2
4.9
0
0
0
Sulfate
Trapping
Efficiency, %
82
100
100
79
40
50
100
100
100
97
100
100
36
96
78
56
100
60
11
100
100
100
Feed Gas Composition • See Table E-22
S02 Out,
ppm
12
13.7
14.3
15
3.3-4.4
6.3
7.2
4.2
7.0
15
15
15
14
6.6
7.1
7.0
S02
Trapping
Efficiency. %
20
9
5
0
78-71
58
52
72
53
0
0
0 ^ '
,-7
/ „___
-^
60
57
58
Space Velocity,
v/v/hr.
150 000
150 000
150 000
150 000
150 000
150 000
150 000
50 000
i
S
o
1
-------
TABLE E-23 (CONT.)
Sorbent
MgO
(Calcined Magnesite)
CaO
(3% Aluminum Stearate
Binder)
Shape
spheres
3-5 mm
pellets
3.2 mm
Weight,
grams
11.9
13.9
Time On
Feed, Hrs.
12.8
14.5
16.0
5.9
7.4
80
. y
H2S04
0.8
1.0
1.2
0.4
Oc
• J
0.4
Sulfate
Trapping
Efficiency, %
75
69
63
88
O /
84
88
S02 Out, '
15
— —
12.3
13.2
13.2
S02
Trapping
Efficiency, %
0
25
20
20
Space Velocity,
v/v/hr.
50 000
50 000
* 370°C
i
ho
O
K*
I
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
'EPA-460/3-76-017
2.
3. RECIPIENT'S ACCESSION-NO.
4. TITLE AND SUBTITLE
Assessment of Automotive Sulfate
Emission Control Technology
5. REPORT DATE
May 1976
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
K.C. Bachman,
------- |