EPA-460/3-77-008
June 1977
ASSESSMENT
OF AUTOMOTIVE
SULFATE EMISSION
CONTROL TECHNOLOGY
U.S. ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
of Mobile Source Air Pollution Control
Emission Control Technology Division
Ann Arbor, Michigan 48105
-------
EPA-460/3-77-008
ASSESSMENT OF AUTOMOTIVE
SULFATE EMISSION CONTROL
TECHNOLOGY
by
M.G. Griffith, R.A. Bouffard, E.L. Holt,
M.W. Pepper, and M. Beltzer
Products Research Division
Exxon Research and Engineering
P.O. Box 51
Linden Avenue
Linden, New Jersey 07036
Contract No. 68-03-0497
EPA Project Officer: R.J. Garbe
Prepared for
ENVIRONMENTAL PROTECTION AGENCY
Office of Air and Waste Management
Office of Mobile Source Air Pollution Control
Emission Control Technology Division
Ann Arbor, Michigan 48105
June 1977
-------
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 (MD-35), Research Triangle Park, North Carolina
27711; or, for a fee, from the National Technical Information Service,
5285 Port Royal Road, Springfield, Virginia 22161.
This report was furnished to the Environmental Protection Agency
by Products Research Division, Exxon Research and Engineering, P.O.
Box 51, Linden Avenue, Linden, N. J. 07036, in fulfillment of Contract
No. 68-03-0497. The contents of this report are reproduced herein
as received from Exxon Research and Engineering. The opinions,
findings, and conclusions expressed are those of the author and not
n.ecessarily those of the Environmental Protection Agency. Mention
of gompany or product names is not to be considered as an endorsement
by the Environmental Protection Agency.
Publication No. EPA-460/3-77-008
-------
TABLE OF CONTENTS
Page No.
I. Summary 1
II. Introduction 4
III. Modification I - Feasibility Studies of Sulfate Removal
From Exhaust Gas By Traps 6
III.l Summary of Results 6
III. 2 Introduction 6
III. 3 Sorbent Composition and Pilling 8
III.3.1 Dry Mixtures 8
III.3.1.1 Physical Combinations - Pelleting. . 8
III.3.1.2 Chemical Combinations - Pelleting. . 10
III.3.2 Wet Mixtures 10
III.3.2.1 Chemical Combinations by
Co-Precipitation 10
III.3.2.2 Application of Cement Chemistry
to Sorbent Preparation 14
III.3.2.2.1 Wet Mixtures of Calcium
Hydroxide With Silica
and Alumina - A Type of
Pozzolanic Cement ... 16
III.3.2.2.2 Use of Portland and
Calcium Aluminate
Cements 19
III.3.3 Laboratory Reactor Tests of
Sorbent Particles . 21
III.3.3.1 Experimental Procedure 21
III.3.3.2 Experimental Results 22
III.3.3.2.1 Benchmark Sorbent ... 22
III.3.3.2.2 Dry Mixes 22
III.3.3.2.3 Wet Mixes 25
III.3.3.2.4 Cements 25
III.3.3.2.5 Discussion of
Experimental Results. . 25
III.3.4 Effect of Heat Treatment and
Compaction on Pellets 25
III.3.4.1 Temperature Effects on
Pellet Strength 25
III.3.4.2 Compaction Effects on Pellet
Strength and Sorbent Activity. ... 26
-------
Page No.
III.3.5 Increased Pellet Porosity 29
III.3.5.1 Pore Volume and Surface
Area Relationship 29
III.3.5.2 Heat Treating 32
III.3.5.3 Blowing Agents 35
III.3.6 Effect of Sodium on Sorbent Activity 39
III.3.6.1 Sodium Content of
Benchmark Sorbent 39
III.3.6.2 Sodium Content of Other Sorbents . . 42
III.3.7 Effect of Pore and Surface Area
Properties on Sorbent Efficiency 42
III.3.8 Chemical Analysis of Used Sorbents 42
III.3.8.1 Rings from Vehicle Testing 42
III.3.8.2 Pellets from Laboratory Testing. . . 45
III.3.9 Other Sorbents 47
III.3.9.1 Carbon Pellets 47
III.3.9.2 Magnesium Hydroxide 48
III.3.9.3 Ca(OH)2 and Colloidal Silica .... 48
III.4 Vehicle Durability Testing of 804" Trap Sorbents .... 48
III.4.1 Summary of Results 48
III.4.2 Experimental Procedures 50
III.4.2.1 Vehicle Preparation 50
III.4.2.2 Test Fuel 50
III.4.2.3 Test Procedure 50
III.4.3 CaO/Si02/Na20 (Benchmark) Rings ... 52
III.4.3.1 Experimental Results 52
III.4.4 CaO/S102/Na20 (Benchmark) Pellets 55
III.4.4.1 Experimental Results . 55
III.4.5 Ca(OH)2/Si02 (Dicalite) Pellets 57
III.4.5.1 Experimental Results 57
III.4.6 Conclusions from Vehicle Durability
Evaluation 59
IV. Modification II - Effect of Noble Metal
Composition on Catalyst Activity 60
IV. 1 Summary of Results 60
IV. 2 Introduction 60
IV. 3 Experimental 61
IV.3.1 Monolith Oxidation Catalysts 61
IV.3.2 Test Vehicle 63
IV. 3.3 Test Fuel . 63
IV.3.4 Test Sequence 63
11
-------
Page No.
IV.4 Results 65
IV.4.1 Sulfate Emissions 65
IV.4.2 Other Gaseous Emissions 65
IV.4.3 Fuel Consumption and Catalyst Outlet
Exhaust Temperatures 72
IV.4.4 Oxygen Level in Exhaust Gas 73
IV. 4.5 Conclusions 73
V. Modification III - Perovskite Catalysts 74
V.I Summary of Results 74
V.2 Introduction 74
V.3 Experimental Conditions 74
V.4 Emission Results 75
V.4.1 Vehicle Engine-Out Emissions 75
V.4.2 Emission Results with GM Palletized
Oxidation Catalyst 76
V.4.3 Emission Results with Pelleted
DuPont Perovskite Catalyst 76
V.4.4 Comparison of Perovskite and GM Pelleted
Catalysts at Normal Operating Temperatures 76
V.5 High Temperature Operation 80
V.5.1 Vehicle Modifications 80
V.5.2 Comparison of GM and Perovskite Catalyst Gaseous
Emissions in the High Temperature Configuration ... 80
V.5.3 Sulfate Emissions from GM Pelleted and Perovskite
Oxidation Catalysts in High Temperature
Configurations 83
V.5.4 Comparison of Emissions Between Standard
and High Temperature Configurations 83
V.6 Comparison of Perovskite Monolithic
Catalysts with Engelhard PTX-IIB 86
V.6.1 Test Procedure 86
V.6.2 Emission Results with Monolithic Catalysts 87
VI. References 89
Appendix A - Selection of Calcium, Silica, and Alumina Components
for Physical Combinations 91
Appendix B - Preparations and Properties of Sorbent Pellets 92
iii
-------
Page No.
Appendix C - Measurement Techniques 96
C.I Gaseous Emissions 96
C.2 Measurement of Sulfate Emissions 96
C.2.1 Exhaust Particulate Sampling System 96
C.2.1.1 Sampling System Components 96
C.2.1.2 Diluent Air Preparation System 98
C.2.1.3 Flow Development Tunnel 100
C.2.1.4 The Exhaust Injection System 100
C.2.1.5 Isokinetic Probe 100
C.2.1.6 Particulate Collecting Stage 102
C.2.2 Exhaust Particulate Sampling
System Performance 102
C.2.2.1 Rapid Mixing of Exhaust and Diluent Air . . 102
C.2.2.2 Development of Uniform Flow in Flow
Development Tunnel 103
C.2.2.3 Tunnel Sampling Losses 105
C.2.2.4 Equivalent Emission Rates with
Parallel Filters 105
C.2.2.5 Temperature Maintenance of the Particulate
Collection Stage 106
C.3 The Goks4yr-Ross Technique 110
C.4 The TECO Sulfur Dioxide Analyzer 110
C.5 Analytical Method for Determination of Sulfate on Glass
Fiber Filters Using a Recording Tltrator and
Probe Colorimeter 116
Appendix D - Test Results - Sulfate Trap Vehicles 125
Appendix E - Vehicle Emissions - Modification II 130
Appendix F - Vehicle Fuel Consumptions - Modification II 136
Appendix G - Vehicle Exhaust Gas Temperatures - Modification II ... 139
Appendix H - Test Results for Burned-Up Catalyst 142
iv
-------
LIST OF TABLES
Table No. Tltle Page No.
III-l Effect of Heat Treatment on Crush Strength 9
III-2 Chemical Combinations from Dry Mixes;
Pelleting Trials 11
III-3 Effect of Sintering at 1350°C Pellets
Prepared from Dry Mixes 12
III-4 Chemical Combinations by Coprecipitatlon 13
III-5 Chemical Combinations by Coprecipitation 15
III-6 Wet Mixtures of Calcium Hydroxide with Silicas. ... 17
III-7 Wet Mixtures of Calcium Hydroxide with
Aluminas and Aluminum Silicates 18
III-8 Wet Mixtures of Calcium Hydroxide and Calcium
Carbonate with Portland Cement and
Calcium Alumlnate Cement 20
III-9 Dry Mixtures of Calcium Hydroxide with
Silica and/or Alumina 24
111-10 The Effect of Heat Treatment Conditions
on Pellet Crush Strength 27
III-ll Preparation of 85 Ca(OH)2/15 Diatomite Pellets of
Different Crush Strengths with Heat Treating
at Two Conditions 28
111-12 85 Ca(OH)2/15 Diatomite Sorbent-Effects of Crush
Strength and Heat Treating on Density and Activity. . 30
III-13 Pore Volume, Surface Area and Density of Sorbents
vs. 804* Trapping Efficiency 31
111-14 Effect of Heat Treating at 315°C and 540°C on Crush
Strength and Density of Pelleted Sorbent
Compositions 33
111-15 504™ and S02 Trapping Efficiencies of Pellets Heat
Treated at 540°C vs. Efficiencies After Heat
Treating at 315°C 34
-------
Table No. Title Page No.
111-16 Crush Strength and Density of Pelleted Sorbents. ... 36
111-17 804" and S02 Trapping Efficiencies of
Pelleted Sorbents 37
111-18 Evaluation of Sorbents in the Laboratory
Screening Test 38
111-19 Sorbents with Cellulose and Starch 40
111-20 Reactor Tests of "Benchmark" Sorbent 41
111-21 (CaO)3 (S102 ^03)1/2 Sorbent Preparations; Effect
of Sodium Content on 804" Trapping Efficiency 43
111-22 Laboratory Test 44
111-23 Analysis of Benchmark Sorbent (85 CaO/10 Si02/5 Na20)
as Rings After Vehicle Test and as Pellets
After Laboratory Reactor Test. 46
111-24 Carbon Pellets, Sample C 47
111-25 Evaluation of Sorbents in the Laboratory
Screening Test 49
111-26 Test Fuel Blend 51
111-27 Surface Area and Void Space in Packed Beds
of Pellets and Rings 53
111-28 1974 Vehicle (351 CID, PTX-IIB) 804" Emissions,
SET Cycle @ 16 000 km 54
111-29 Vehicular Durability Testing of 804" Sorbent Traps . . 56
111-30 Vehicular Durability Testing of 804" Sorbent Traps . . 58
IV-1 Catalyst Characterization 62
IV-2 Test Fuel Blend 64
IV-3 Sulfate Emissions at Start and Finish
of Mileage Accumulation 66
IV-4 Average Gaseous Emissions 72
IV-5 Fuel Consumption, g/km 72
vl
-------
Table No. Title Page No.
IV-6 Maximum Exhaust Gas Temperatures, °C 73
V-l Test Vehicle FTP Emissions Without Catalyst 75
V-2 Emission Characteristics, Vehicle Equipped with GM
Catalyst in Stock Configuration 77
V-3 Emission Characteristics, Vehicle Equipped with
Pelletized DuPont Catalyst in Stock Configuration. . . 78
V-4 Comparison of GM and Perovsklte Oxidation Catalysts. . 79
V-5 Catalyst Conversions 79
V-6 Catalyst Temperatures Standard Vehicle
Configuration 80
V-7 Catalyst Temperatures Modified Vehicle
Configuration 81
V-8 Comparative Gaseous Emissions Under High Temperature
Configuration Conditions 82
V-9 Average FTP Gaseous Emissions. 84
V-10 Comparative Sulfate Emission Rates of the GM and
Perovskite Catalysts in Standard and High
Temperature Configurations 85
V-ll Emission Characteristics, Vehicle Equipped with
Monolithic Perovskite Catalyst in High Temperature
Configuration 87
V-12 Emission Characteristics, Vehicle Equipped with
Engelhard Monolithic Oxidation Catalyst in
High Temperature Configuration 88
B-l Physical Combinations - Dry Mixtures 92
B-2 Preparation of Chemical Combinations
by Coprecipitatlon 93
B-3 Pressure Drop Across 804" Trap 95
C-l 502 Measurements at Indicated Oxygen Concentrations. . 115
C-2 Composite Effects of C02, 02, and CO on
TECO S02 Response 116
vii
-------
Table No. Title Page No.
C-3 Sulfate Extractions Using Dilute Nitric
Acid and Water - Comparison Data . 124
D-l Test Car 124 - 804" Trap Vehicle Containing
Benchmark Rings (12.7 mm OD) 125
D-2 Test Car 124 - 804" Trap Vehicle Containing
Benchmark Pellets (4.76 mm OD) 127
D-3 Test Car 99 - 804" Trap Vehicle Containing Ca(OH>2/
Dicalite (4.76 mm Pellets) 129
E-l 1975 Vehicle (351 CID) Pt/Pd Oxidation
Catalysts (AEW 2/6C/50M) 130
E-2 1975 Vehicle (351 CID) Pt/Rh Oxidation
Catalysts (AEW 2/3C/40/M) 132
E-3 1975 Vehicle (351 CID) Pt Oxidation
Catalysts (AEW 2/12C/40/M) 134
F-l Fuel Consumption, (g/km) 1975 Car, Pt/Pd
(6-C) Oxidation Catalysts 136
F-2 Fuel Consumption, (g/km) 1975 Car, Pt/Rh
(3-C) Oxidation Catalysts 137
F-3 Fuel Consumption (g/km) 1975 Car, Pt
(12C) Oxidation Catalysts 138
G-l Maximum Exhaust Gas Temperature at Catalyst
Outlet Pt-Pd (AEW 2/6C/50M) 139
G-2 Maximum Exhaust Gas Temperature at Catalyst
Outlet Pt/Rh (AEW 2/3C/40M) 140
G-3 Maximum Exhaust Gas Temperature at Catalyst
Outlet Pt (AEW 2/12C/40M) 141
H-l 1975 Vehicle (351 CID) Pt/Pd Oxidation
Catalyst (AEW 2/6C/50/M) 142
H-2 1975 Vehicle (351 CID) Pt/Pd Oxidation
Catalyst (AEW 2/6C/50/M) 143
H-3 Fuel Consumption, (g/km) 1975 Car, Pt/Pd
(6-C) Oxidation Catalysts 144
H-4 Maximum Exhaust Gas Temperature at Catalyst Outlet
Pt/Pd (AEW 2/6C/50M) 145
viii
-------
LIST OF FIGURES
Figure No. Title Page No.
III-l Apparatus for Laboratory Screening of Sorbents. ... 23
IV-1 S04= Emissions - 1975 FTP 67
IV-2 804" Emissions - FET Cycle. 68
IV-3 804" Emissions - SET Cycle 69
IV-4 804= Emissions - 64 km/h Cruise 70
IV-5 S04= Emissions - 96 km/h Cruise 71
C-l Exhaust Particulate Sampler 97
C-2 Schematic of Dehumidification Section 99
C-3 Counter Current Exhaust Injection System 101
C-4 Dew Point of Diluted Exhaust vs.
Air/Exhaust Dilution Ratio 107
C-5 Relative Humidity of Exhaust Dilution Air
Mixture at Vicinity of Sampling Probes
During the 1972 Federal Test - Driving Cycle 108
C-6 Temperature Control System Performance -
Catalyst-Equipped Car 109
C-7 Relative Humidity of Exhaust Dilution Air Mixture
at Vicinity of Sampling Probes During 64 km/h
Cruise Conditions at 32°C Ill
C-8 Finned Tube Cooling Setup 112
C-9 Principal of Operation - TECO S02 Instrument 114
C-10 Titration of Sulfate Using Thorin Indicator 119
C-ll Block Diagram of Brinkman PC600 Probe Colorimeter . . 121
ix
-------
I. Summary
The research described in this report was conducted under an
extension of EPA Contract No. 68-03-0497, referred to as Phase II.
Results of Phase I of this contract were documented as EPA Report
460/3-76-017(1). Phase I consisted of work in four areas related to
sulfate emissions from vehicles:
• Literature search of relevant sources
• Measurement of sulfate emissions from non-
catalyst vehicles
• Factors affecting sulfate emissions from
oxidation catalyst-equipped cars
• Studies of the feasibility of sulfate traps
in vehicle exhaust systems
The current program concentrated on extending the fourth area,
sulfate traps. In addition, smaller programs were carried out to inves-
tigate the effect of noble metal composition of automotive catalysts on
their tendency to convert S02 to H2S04, and of the activity of several
perovskite-based oxidation catalysts developed by DuPont.
Modification I - Sulfate Traps
The trap work was aimed at developing Improved sorbents for
804s capable of efficiently removing this material from exhaust gas at
a low pressure drop over a long service life. The previous work had
shown that good efficiency was possible using small pellets containing
primarily Ca(OH)2, with lesser amounts of Na20, as the active components.
However, pressure drop was unrealistically high. On the other hand,
rings of the same composition provided lower pressure drops, but also
much reduced sorption efficiency. Accordingly, the focus of the current
work was on achieving a satisfactory compromise between sorption and
pressure drop.
Two principal paths were pursued; changes in chemical compo-
sition, and in geometry. The former approach involved retaining CaO
or Ca(OH)2 as the active Ingredient, but eliminating sodium. Sodium
had been identified as responsible for substantial amounts of S02
sorption, in addition to the desired reaction with 804*. Sorption of
S02 is undesirable since it cuts down on the sorbent capacity for 804*
and increases the rate of sorbent swelling, which leads to higher pressure
drops. The second approach involved the use of smaller rings and larger
pellets than had been used previously, in hopes of achieving a better
compromise between contact efficiency and pressure drop through the bed.
-------
- 2 -
The removal of N&20 from the sorbent composition was found
to increase the difficulty of compacting sorbent particles with adequate
mechanical strength. A number of inert binders, based on A1203 and/or
S102 were mixed with basic calcium compounds, including the hydroxide,
oxide and carbonate, in wet and dry preparations. A number of mixing
techniques, pelleting conditions, and post-pelleting heat treatments
were Investigated. In particular, use was made of cement chemistry
in preparing experimental sorbents, since the mixtures used were similar
to Portland and calcium-»aluminate cements.
It was found that a number of very strong pellet samples
could be made by these techniques. However, laboratory testing of
these pellets showed poor sorptlon efficiency for 804". It was deter-
mined that this was caused by the low porosity of the hard pellets, and
also by the absence of sodium. A number of attempts were made to increase
porosity, using calcination or burn-out agents. In addition, improved
preparation techniques were developed to allow the introduction of
controlled quantities of sodium. The goal was to add enough sodium to
enhance 804" sorption without greatly increasing sorbent reactivity
towards 802.
Based on these laboratory studies, two sorbent compositions
were identified as sufficiently promising to warrant further vehicle
testing. These were a mixture of 85 wt % CaO/10% 8102/5% Na2<>, and
85 wt % Ca(OH)2/15% 8102- The first sorbent was fabricated into rings
of 12.7 mm outside diameter and pellets of 4.8 mm diameter. The rings
were smaller than those tested in Phase I of this contract, while the
pellets were larger. The second sorbent was made up into the 4.8 mm
pellet configuration.
These sorbents were loaded into empty GM catalyst canisters,
of 4.2 litre capacity, which were mounted In the muffler position of
oxidation catalyst-equipped vehicles. Although 804™ sorption efficiency
was high initially, it fell off rapidly. None of these sorbents were
tested beyond 15 000 km of vehicle operation since sorption efficiency
had fallen below 90% in each case by this time.
Modification II - Noble Metal Effect on Catalyst Conversion of 802
Three pairs of monolithic oxidation catalysts were prepared
by Matthey-Bishop, Inc. The catalysts were made as identical as possible
except for their noble metal composition. One pair contained only Pt
as the active metal, another pair Pt-Pd, and the third Pt-Rh. The three
pairs were tested sequentially on the same vehicle for 12 800 km each
and the gaseous and 804" emissions measured. The following table shows
the average 804" emissions measured over the EPA Sulfate Emission Test,
using a fuel containing 320 ppm sulfur. Emissions of CO and HC were
similar for all three sets of catalysts. Both 804" and the regulated
emissions did not change measurably over the short duration of these tests.
-------
- 3 -
Catalyst Average 804°, g/km
Pt-Rh 0.010
Pt-Pd 0.025
Pt 0.030
Modification III - Perovskite Catalysts
Two types of experimental catalysts supplied by DuPont were
tested for activity for CO and HC removal from exhaust gas and for con-
version of 802 to 804°. Their activity was compared with that of
commercial catalysts. These comparisons were made under normal vehicle
operating conditions and also under high temperature operation conditions.
The latter were obtained by adding port liners and insulation to the
test vehicle and resulted in increases in catalyst temperature of
100-200°C.
The first comparison was between a sample of beaded perovskite
catalyst, containing Pt at a level roughly comparable to the total Pt-Pd
level in commercial GM beaded catalyst, and the GM catalyst. Initially,
the activity of these two catalysts was tested sequentially on a conven-
tional vehicle. The GM catalyst, in a 4.2 litre canister, was run for
12 800 km and the Pt-perovskite catalyst for 6 700 km. It was found that
CO and HC emissions were substantially lower for the GM catalyst. This
was attributed primarily to poor light-off performance of the perovskite
catalyst. Surprisingly, 804° emissions from the GM catalyst were also
lower.
After vehicle modification to provide higher temperature catalyst
operation, emissions tests were repeated on both catalysts. The GM
material still retained a substantial advantage, and testing of the beaded
samples was concluded.
Samples of a monolithic version of DuPonts's Pt-perovskite
catalyst were also tested. In this case, their activity was compared
with the commercial Engelhard PTX-IIB monolithic catalyst. Each catalyst
was aged for 800 km prior to emissions testing. The test vehicle was
kept in the high temperature configuration. The commercial catalyst
showed much lower CO and HC emission levels, although it generally
resulted in higher 804** emissions. However, the poor CO and HC activity
of the perovskite catalyst precluded further work in this program.
-------
- 4 -
II. Introduction
Modification I - Sulfate Traps
The work performed under Phase I of this contract had shown
that sulfate traps, consisting of sorbent filled canisters placed in
the exhaust system of catalyst-equipped vehicles, could sorb nearly all
of the 804" from the exhaust gas. The most efficient and long-lived
system had consisted of small (^3 mm) pellets pressed from a powder
whose composition was nominally 85 wt % CaO/10% Si02/5% Na20. However,
although this sorbent proved greater than 90Z effective for some 40 000
km of operation, pressure drop was a major problem, increasing from an
initial value of 1 kPa to a final value of 30 kPa. A typical exhaust
catalyst has about 4 kPa pressure drop. The pressure drop increase was
ascribed to several factors. As the CaO became sulfated, the pellets
tended to swell, since CaS04 is a larger molecule than CaO. In addition,
some powdering was evident, resulting in a filling in of the void volume
between pellets.
An attempt to overcome the pressure drop problem through the
use of sorbent rings in place of pellets was only partially successful.
Rings of about 1.6 cm outside diameter were fabricated from the same
material described above. The greater void volume in the bed provided
a lower pressure drop initially, 0.5 kPa, and an increase to only 4 kP4
after 20 000 km of testing. However, the poorer gas-solid contacting
produced a much lower 804" removal efficiency, 80% initially, but falling
off rapidly after only about 5 000 km.
Based on this experience with pellets and rings made up of
the CaO/Si02/Na20 composition, several further approaches to achieving
a better trade-off between pressure drop and sorptlon efficiency suggested
themselves:
• Change in chemical composition
The vehicle tests had shown that substantial quantities of
S02 were being picked up as well as 804'. Preferably, only 804" should
be sorted, since 802 is a normal and relatively innocuous component of
exhaust gas in all motor vehicles (catalyst and non-catalyst) and its
sorption lowers capacity for 804" pick-up. Laboratory studies had shown
that the sodium was primarily responsible for 802 sorption. Therefore,
it was postulated that its removal would result in pellets with greatly
increased capacity for 804". Furthermore, their sulfation rate, hence
swelling and resultant pressure drop Increase, should be minimized.
Thus, a major part of Phase II of this contract has been concerned with
the formulation and fabrication of sodium-free, calcium-based sorbent
preparations.
-------
- 5 -
• Change in sorbent geometry
Phase I had shown that 3 mm pellets were too small to allow
low pressure drop and 1.6 cm O.D. rings too large for good gas-solid
contacting. Therefore, It was decided to try larger pellets and smaller
rings, maintaining the original sodium-containing formulation, in an
effort to achieve a better balance between pressure drop and sorption
efficiency.
•Modification II - Noble Metal Effect on Catalyst Conversion of S02
Other work has indicated that catalyst type can influence the
extent of 502 conversion to 804" in exhaust gas(2»3»4)t These studies
showed that a Pt-Rh containing catalyst tended to produce less S04"
than Pt or Pt-Pd catalysts. However, these studies were confounded by
other properties in which these catalysts differed beyond just noble
metal composition. They were manufactured by different companies and
were dissimilar in washcoat properties and method of metal impregnation.
In an effort to separate out the effect of noble metal composition from
other factors, it was decided to request a single manufacturer, in this
case Matthey-Bishop, Inc., to prepare three pairs of monolithic catalysts.
These were to be as identical as possible in all respects, save their
noble metal composition. By then testing the three pairs on the same
vehicle under similar operating conditions, it should be possible to
determine if noble metal composition plays a role in oxidation catalyst
selectivity for SC-2 conversion. If no differences were observed, it
would have to be concluded that the reported effects in the literature
were due to other factors. If a significant difference were obtained,
then it could be concluded that noble metal composition does play a
role, although further work would still be necessary to determine if
it were the sole or even the dominant effect.
Modification III - Perovsklte Catalysts
DuPont has been working towards the development of lead-
tolerant automotive oxidation catalysts. Their approach has been to
use noble metal supports capable of anchoring the metals tightly enough
to prevent the high temperature sintering normally seen with conventional
alumina supports. By means of this technique, they hope that catalysts
can be operated at temperatures sufficiently high to prevent lead
deposition and reaction with the catalyst, while avoiding thermal de-
gradation. Such temperatures have been reported by DuPont to be in the
vicinity of 730°c(5). At the time Modification III was begun, there
appeared no practical means of consistently achieving such temperatures
without vehicle modifications which would lower fuel efficiency. In
view of the growing importance of good fuel economy, it was agreed that
the DuPont catalyst would be tested in this program only with unleaded
gasoline. Catalyst temperatures would be raised above usual levels,
but only through techniques such as port liners and insulation, which
should not affect fuel economy.
-------
- 6 -
III. Modification I - Feasibility Studies of
Sulfate Removal From Exhaust Gas By Traps
III.l Summary of Results
Initially a number of laboratory preparations were made of
sodium-free compositions containing basic calcium compounds and inert
binders. These preparations used both wet and dry mixing techniques.
It was found difficult to fabricate strong pellets from dry mixtures,
but wet methods, relying on cement chemistry, resulted in very strong
pellets. Laboratory activity screening showed poor sorption efficiency,
however, due In part to poor porosity of the very hard pellets. It was
also found that the presence of some sodium greatly enhanced sorption
efficiency.
Attention was, therefore, directed to methods for improving
pellet porosity, through the use of calcination and of burn-out agents,
and of controlled sodium addition. Improved porosity was achieved
through the burning out of cellulose and starch, with some Improvement
in sorption efficiency. A greater improvement was achieved by the
development of better washing techniques, which allowed controlled
addition of sodium to the sorbent preparations in amounts which enhanced
sorption efficiency towards S04 , without greatly Increasing sorption of
802.
Two preparations were chosen for vehicle testing; 85 wt % CaO/
10% Si02/5% Na20, and 85% CaO/15% S102. The former was fabricated into
rings of 12.7 mm outside diameter and 4.8 mm pellets. The latter was
made up as 4.8 mm pellets only. In no case was satisfactory 804 sorption
achieved during the vehicle runs, and no sorbent was tested for more than
15 000 km.
III.2 Introduction
In the Final Report of Phase I of this contract^, laboratory
preparation and vehicle testing of pellets and rings prepared from
"benchmark" material, nominally a ratio of 85/10/5 of Ca(OH)2/S102/Na20,
was described. Although the benchmark pellets exhibited excellent
activity for 804" removal in laboratory reactor and vehicle trap tests,
one major problem appeared. The pressure drop through the vehicle trap
increased over the course of the test xfrom an Initial value of 1 kPa
Ov4 Inches H2<3) to some 30 kPa. This was attributed to several factors:
swelling of the pellets as sulfatlon proceeded, since CaS04 is a larger
molecule than is Ca(OH)2; and collection of fine particles, attritted
from the pellets, in the relatively densely packed bed.
-------
- 7 -
A variety of approaches suggested themselves as possible
solutions to the pressure drop problem:
A. Use a trap bed with greater void volume
B. Prepare harder sorbent particles to reduce level
of fines in the bed
C. Reduce the SC>2 sorption ability of the sorbent,
thereby reducing sulfation rate and hence degree
of swelling. This would also provide the added
benefit of increased sorption capacity for 804**.
As a first step in attempting to improve trap performance,
sorbent particles in the form of rings were prepared from the benchmark
composition. Vehicle testing of these rings, also reported in Reference 1
showed greatly improved pressure drop, but also considerable reduction in
S04= sorption performance. Thus, the greater void volume afforded by
rings compared to pellets did reduce pressure drop, as predicted, but
the resulting decrease in gas-solid contacting efficiency was apparently
too great.
As a result of the above work done under Phase I of the contract,
it was concluded that Phase II should try the following approaches.
A. Retain the benchmark composition because of its known
ability to remove 804= from exhaust gas, but attempt
to optimize particle geometry to obtain a satisfactory
compromise between pressure drop and contacting
efficiency. This resulted in the preparation of
smaller rings and larger pellets than those reported
under Phase I.
B. Remove Na from the sorbent composition, since some
preliminary laboratory results obtained in Phase I
indicated that much of the 802 sorption effect was
due to its presence. This work in Phase I also
suggested that the removal of Na adversely affected
the strength of sorbent particles. Therefore, a
variety of binding agents, compounding techniques,
and heat treatments would have to be studied in
efforts to obtain sorbent particles of the requisite
hardness, which retained sufficient porosity for good
S04*a sorption activity.
Based on these considerations, a large number of Ca-based
sorbents have been prepared, with a number of binders, in physical and
chemical combinations. Both dry and wet mixing techniques have been
-------
- 8 -
employed, as have a variety of pre- and post-pelleting treatments. Those
preparations capable of compaction into strong pellets have been tested
in a laboratory reactor for 804" sorption activity. In addition, attempts
have been made to correlate physical properties of these pellets, such
as surface area, pore volume, and pore size distribution, with sorption
activity. Finally, vehicle tests of full size traps have been conducted
using rings and pellets made from benchmark material. The rings were
/ smaller and the pellets larger than their counterparts from Phase I. In
addition, pellets made from a Na-free composition have also been vehicle
tested.
III.3 Sorbent Composition and Pilling
The aim of these studies was to determine the suitability of
basic inorganic compounds of calcium, combined physically or chemically
with inorganic binders, for fabrication into compacted shapes, and the
efficiency of such particles for S04* removal.
These preparations differed from the benchmark sorbent, which
is of nominal composition 85 CaO/10 Si02/5 Na20, in containing little or
no sodium. The sodium component has previously been shown to be
responsible for sorption of S02 which is undesirable from the stand-
point of loss of trap capacity for 864" and too rapid swelling.
III.3.1 Dry Mixtures
III.3.1.1 Physical Combinations - Pelleting
A series of nine dry mixtures were prepared using calcium oxide,
calcium hydroxide, and calcium carbonate, each with three binder systems.
Selection of suitable ingredients for these mixtures followed a survey
of commercially available materials. For the dry mixtures, a requirement
for the Ingredients is that they flow readily in the pelletizing operation.
In the automatic pellet press, the dry material must flow without caking
or binding to fill 3 mm dies. From our experience, materials which are
granulated in the range of 14-35 mesh have suitable flow properties.
For silica and alumina binders, chemical reactivity is Important
in addition to flow properties. Forms of silica and alumina which are
"hydrated", rather than calcined, are believed to be the best choice,
providing chemical reactivity and good binding.
Alcoa Hydrated Alumina C-31 (coarse) (64.9 wt % equivalent
A1203) was selected as a reactive form of alumina with good flow properties.
Similarly, PPG Hi-Sil 210 hydrated amorphous silica (94 wt % equivalent
8102) was chosen as a suitable source of silica. A description of the
materials selected for the dry mixtures is given In Appendix A.
-------
- 9 -
These combinations of calcium compounds with silica and/or
alumina were then evaluated in pelleting trials using a sixteen-station
rotary tablet press having 3.2 mm dies and upper and lower punches.
The results of these trials are given in Appendix Table B-l.
All formulations flowed well during pelleting. However, those
mixtures containing calcium oxide or calcium carbonate presented severe
problems during attempts to make hard pellets. The machine was operated
beyond recommended limits in tests with these mixtures. Even with the
use of 3% Sterotex (as die lubricant), reasonable pellets were obtained
only with mixture No. 5 (85% CaO/15% A1203). However, these pellets
disintegrated after heating at 95°C.
For the mixtures containing calcium hydroxide, no problems
were encountered during pelleting. Hard pellets were obtained with each
of the three binder systems after heat treating at 315°C. The pellets
made from calcium hydroxide in combination with alumina, or with silica/
alumina, were slightly harder than those made with silica.
This lack of success in dry-compacting calcium oxide and calcium
carbonate was probably due to the hardness of these materials and difficulty
in obtaining cohesion between grain particles. An attempt was made to
wet-granulate finely powdered calcium carbonate, with only fair success.
The resulting material flowed poorly during pelleting and, even by hand-
filling the dies, no coherent pellets could be made. This is in contrast
with wet granulation of finely powdered calcium hydroxide. This material
flowed well during pelleting and very hard pellets were made from 100%
calcium hydroxide, as shown in Table III-l.
TABLE III-l
EFFECT OF HEAT TREATMENT ON CRUSH STRENGTH
Crush Strength of 100%
Heat Treatment Ca(OH)2 Pellets (kg)
initial 7.7
48 h @ 95°C ,
+ 6 h @ 150°C /'°
16 h @ 315°C 8.6
No further dry mix pelleting studies were performed with
calcium oxide or calcium carbonate. All the calcium hydroxide prepa-
rations, including the pellets from 100% calcium hydroxide were evaluated
for 804™ sorption efficiency using the laboratory reactor, as discussed
in Section III.3.3.
-------
- 10 -
III.3.1.2 Chemical Combinations - Pelleting
Table III-2 lists results using dry mixtures of calcium hydroxide
with silica and/or alumina to give the stoichiometric proportions: (CaO)3
Si02, (CaO)3 A1203, and (CaO)3 (Si02 Al203)]/2. For these mixtures, the
binder concentrations are higher than the 15% employed above. For (CaO)3
S102, 21.3% silica is used, and for (CaO)3 A1203, 31.4% alumina is required.
The mixture corresponding to (CaO)3 (Si02 Al203>i/2 contains 9.9% silica
and 16.8% alumina.
The crush strength of the unfired pellets from each of these
mixtures was lower than that of the corresponding preparation using less
binder. For example, 85 Ca(OH)2/15 Si02 sorbent had a crush strength
of 4.1 kg, whereas (CaO)3 8102 sorbent had a crush strength of only
2.4 kg. The hardest pellets were obtained using the (CaO)3 A1203 compo-
sition (3.8 kg). Even these were weaker than the corresponding pellets
from 15% alumina (4.5 kg).
These pellets were then fired for two hours at 1350°C, which
is representative of commercial conditions for making calcium silicates
and calcium aluminates, as in cement manufacture. The results are given
in Table III-3. As a result of this heat treatment, the pellets corre-
sponding to (CaO)3 Si02 and (CaO)3 (Si02 Al203)i/2 crumbled to powder.
The pellets of (CaO)3 A1203 remained intact; however, crush strength
decreased from 3.8 kg to 2.9 kg. Therefore, this approach to sorbent
fabrication was not pursued further. With sufficient effort, conditions
might be found where CaO and A1203 or Si02 particles could be effectively
sintered to give strong pellets; however, because the sorbent would be
in the form of CaO, rehydration or carbonatlon under vehicle exhaust
conditions holds the likelihood of breaking up the sintered structure.
III.3.2 Wet Mixtures
III.3.2.1 Chemical Combinations by Co-Precipitation
The three chemical combinations discussed in Section III.3.1.2
were also prepared by a co-precipitation technique. Appendix Table B-2
describes the materials and procedure used, while Table III-4 gives the
chemical and physical properties of the pellets obtained from the three
preparations. Very hard pellets were obtained from the (CaO)3 S102
preparation (9.4 kg). For the (CaO)3 ,Al2®3 preparation, the pellets
cracked and became weak after treatment at 315°C (crush strength 1.4 kg)
even though the initial appearance and hardness was satisfactory
(crush strength 4.4 kg). Hard pellets were obtained from the (CaO)3
(S102 Al203)i/2 combination (initial crush strength 7.2 kg, although
these became somewhat weaker after 315°C (5.0 kg).
-------
- 11 -
TABLE III-2
CHEMICAL COMBINATIONS FROM DRY MIXES;
PELLETING TRIALS
Nominal Composition:
(CaO)3SiQ2 (CaO)3Al203
Ingredients, Wt. % (1)
Calcium Hydroxide (2) 78.7
Silica (3) 21.3
Alumina (4)
Operation During Pelleting
Flow into Dies: good
Ease of Compaction/
Ejection: fair
Crush Strength of 3.2 mm Pellets
Initial Crush Strength (kg) 1.9
Crush Strength after
16 h @ 315°C (kg) 2.4
68.6
31.4
good
good
3.4
3.8
73.3
9.9
16.8
good
good
3.6
3.5
Notebook Reference 4349
50-1
50-2
50-3
(1) Percentages given are for pure ingredients, i.e., Ca(OH)2, Si02, and
A1203. In the actual formulations, the percentages are adjusted to
allow for water content of the ingredients.
(2) Granulated calcium hydroxide, 85 wt. % Ca(OH)2, balance water.
(3) Hi-Sil 210 Hydrated Amorphous Silica, 94 wt. % Si02, balance water.
(4) Alcoa Hydrated Alumina C-31 (coarse), 64.9 wt. X A1203, balance water.
-------
TABLE III-3
EFFECT OF SINTERING AT 1350°C PELLETS PREPARED FROM DRY MIXES
Nominal Composition;
(CaO)i S102
A1203
(Ca03 (S102 Al203)i/2
Ingredients. Wt. %
Calcium Hydroxide
Silica
Alumina (1)
78.7
21.3
68.6
31.4
73.3
9.9
16.8
Crush Strength of 3.2 mm Pellets (kg)
Initial 1.9
After 16 hrs. @ 315°C 2.4
After 2 hrs. @1350°C powdered
3.4
3.8
2.9
3.6
3.5
powdered
Notebook Reference 4349
50-1, 116
50-2, 116
50-3, 116
(1) Alumina as hydrated alumina, 64.9 wt. % A1203, balance water.
-------
- 13 -
TABLE III-4
CHEMICAL COMBINATIONS BY COPRECIPITATION :
PELLETING TRIALS
Nominal Composition: (CaO)j S102 (CaO)^ AljO-j
Sterotex 3% 3%
Analysis (1) Calculated Found Calculated Found
CaO 73.7 72.2 62.2 64.4
Si02 26.3 23.5
A1203 — — 37.8 32.5
Na20 — 3.5 -- 7.0
Operation During Pelleting
Flow into Dies: good excellent
Ease of Compaction/
Ejection: good good
Crush Strength of 3.2-mm Pellets
(CaO)i(SiO?Al?OOl
3%
Calculated Found
67.5 58.2
12.0 9.8
20.5 18.1
14.6
good
good
Initial Crush Strength (kg)
Crush Strength after
16 h @ 315°C (kg)
6.3
9.4
4.4
1.4 (2)
7.2
5.0
Notebook Reference 4349 -
28
30
41
(1) Analysis for elements is by emission spectroscopy on material which has
been calcined for 5 h at 815°C.
(2) Cracks formed in pellets
-------
- 14 -
Chemical analyses on these preparations after calcining at
815°C are given in Table III-4. These analyses indicated that the
water-wash procedure used for these preparations was not satisfactory
in removing excess sodium. Expressing sodium content as Na20, these
materials contained the equivalent of 3.5-14.6 wt % Na20. One of our
goals was to produce sodium-free sorbents which would be unreactive to
SC>2' Therefore, three new preparations of (CaO)3 A^Os, (CaO)3 S102
and (CaO>3 (Si02 Al203)i/2 were made with more thorough water-washing.
The new wash procedure consisted of reslurrying the filter cake with
distilled water in a separate vessel, stirring 30 minutes, and refilter-
ing. This sequence was repeated three times. Satisfactory low sodium
values were obtained using the revised procedure. Analyses of these
preparations and the results of crush strength tests on the pelleted
materials are given in Table III-5. Residual sodium contents were very
low (0.5-2.3% expressed as Na20) relative to levels obtained previously
(3.5-14.6%). Low sodium content is desirable because such sorbents may
be expected to be more selective in sorting S04* rather than S02- How-
ever, an unexpected result of the new procedure and lower sodium levels
was a reversal in crush strengths. Previously, (CaO)3 S102 pellets
were the strongest and (CaO)3 A1203 pellets were the weakest. With the
new procedure, the (CaO)3 A1203 pellets were the strongest, having a
crush strength of 10.5 kg, whereas a crush strength of only 1.4 kg was
obtained previously. Pellets from (CaO>3 S102 had a crush strength of
3.6 kg by the new procedure, whereas a strength of 9.4 kg was obtained
previously. No significant change in crush strength occurred in
preparations of the mixed sorbent, (CaO>3 (S102 Al203)i/2 (crush
strengths 5.0 and 4.8 kg).
Of the six preparations by coprecipitation, we selected
(CaO)3 A1203, made by the improved water-wash procedure, as the best
candidate for further evaluations. Pellets of this material were tested
for S04° trapping efficiency in a laboratory reactor, as described in
Section III.3.3.
III.3.2.2 Application of Cement Chemistry to Sorbent Preparation
Surprisingly, the addition of S102 or A1203 to Ca(OH)2 was
found, in dry mixes, to lower the crush strength of compacted pellets
rather than to provide binding action, as discussed in Sections III.3.1.1
and III.3.1.2 and shown below. However, as reported in Section III.3.1.1,
strong pellets could be made from Ca(OH)2 with no binder by a wet granu-
lation technique. Therefore, it was decided to look at this method in
more detail, along with the use of binders. A number of preparations
Calcium Hydroxide + Binder
15% 21.3% 15% 31*4*
Binder: None SiO? S102 A1203 A1203
Crush Strength (kg) 8.6 4.1 2.4 4.5 3.8
-------
TABLE III-5
CHEMICAL COMBINATIONS BY COPRECIPrrATION
Improved Water Wash Procedure
Nominal
Composition
Analysis
(3)
(CaO)
(CaO)
(CaO)3 (Si02
First
Calculated Preparation
(1)
Second
Preparation
(2)
First
Calculated Preparation
(1)
Second
Preparation
(2)
First
Calculated Preparation
(1)
Second
Preparation
(2)
CaO
Si02
A1203
Na20
73.7
26.3
72.2
23.5
3.5
74.5
24.0
2.3
62.2
37.8
64.4
32.5
7.0
64.4
30.2
1.1
67.5
12.0
20.5
58.2
9.8
18.1
14.6
64.4
10.5
20.8
0.5
Crush. Strength of
3.2-mm Pellets
Initial Crush
Strength, kg
6.3
2.1
4.4
6.2
7.2
6.1
en
I.
Crush Strength
After 16 hours
@ 315 °C, kg
9.4
3.6
1.4
10.5
5.0
4.8
Notebook Ref-
erence 4349
28
90
30
62
41
64
(1) Using original water wash procedure.
(2) Using improved water wash procedure. (The filter cake is reslurried three times with distilled water and reflltered.)
(3) Analysis on samples heated 5 hours at 815°C.
-------
- 16 -
were made using vet pastes of the mixtures, which were then dried and
granulated prior to pelleting. In this approach, we were guided by the
chemistry of calcium-derived cements. These are "hydraulic" cements,
hardening by the addition of water due to formation of hydrated phases
of calcium silicates and calcium aluminates. It was hoped to use hydrau-
lic cements to "cement together" particles of calcium hydroxide or
calcium carbonate to give a mechanically strong, although porous,
structure.
Hydraulic cements were incorporated into sorbent compositions
by either (1) combining ingredients which react with each other and with
water to harden, or (2) using commercially available hydraulic cements
which can bind calcium hydroxide or calcium carbonate particles analogous
to the way that aggregate stone is bound in making concrete. The first
approach is illustrated by "pozzolanic cements", described in Section
III.3.2.2.1. The latter by Portland and calcium aluminate cements, as
described in Section III.3.2.2.2.
III.3.2.2.1 Wet Mixtures of Calcium Hydroxide
With Silica and Alumina - A Type
of Pozzolanic Cement
The pozzolanic cements are mixtures of hydrated lime (calcium
hydroxide) with various sources of active aluminum silicates, known as
pozzolanas. In the presence of water, the pozzolana reacts with
calcium hydroxide to form hydrated calcium aluminates and silicates,
which constitute the binding matrix. We were seeking to agglomerate
powdered calcium hydroxide by such a mechanism.
A series of mixtures of calcium hydroxide with various sources
of silica and alumina, which are described in Tables III-6 and 111-7,
were prepared. Calcium hydroxide was mixed with sufficient binder to
give 5 or 15 wt % alumina and silica. The mixture was blended with
water to form a paste, and allowed to age 24 hours in a covered vessel.
The resulting material was dried at 65°C and granulated to 14-40 mesh
before pressing to 3.2 mm pellets.
Table III-6 lists the results of pelleting trials with com-
binations of calcium hydroxide and various sources of silica; These
consist of a hydrated amorphous silica, an anhydrous amorphous silica,
diatomites - both calcined and uncalcined, and a colloidal silica.
The hardest pellets of all, after heat treating at 3158C, were obtained
with 15 wt % uncalcined dlatomite. These have a crush strength of
12.8 kg, which is significantly harder than corresponding pellets
obtained from 100% calcium hydroxide (8.6 kg). Crush strength using
5% of uncalcined diatomite is only slightly lower (11.0 kg). The soft-
est pellets were obtained from calcined diatomite, having a crush
-------
TABLE III-6
WEI MIXTURES OF CALCIUM HYDROXIDE WITH SILICAS
Composition (1)
Benchmark SorbenC
"85 CaO/10 S102/5 NaaO"
85
15
95
5
95
15
Wt. % Calcium Hydroxide
We. Z Hydrated Amorphous
Silica (Lo-Vel 27)
Wt. % Calcium Hydroxide
Wt. Z Hydrated' Amorphous
Silica (Lo-Vel 27)
Wt. % Calcium Hydroxide
Wt. Z Diatomite (uncalcined)
(Dicalite SA3)
Crueh Strength Of
3.2 mm Pellets, kg
After Heat
Initial Treating (2)
6.6 8.8
5.0 7.4
10.9 9.6
7.6 12.8
Laboratory
Total Exposure SOV Trapping
To Feed, Hours Efficiency, Z
1.3 94
2.6 94
6.9 72
1.3 57
3.7 71
5.7 62
Reactor Test: 100.000 v/v/hr., 370°C
S02 Trapping Final Crush Notebook
Efficiency. Z Strength of Pellets. kg Reference 4349
60 44
48
33 6.6
87-1
87-2
20 68-1
12
12 11.5
95 Wt. Z Calcium Hydroxide 4.5 11.0
5 Wt. Z Diatomite (uncalcined)
(Dicalite SA3)
85 Wt. Z Calcium Hydroxide 2.3 3.2
15 Wt. Z Diatomite (calcined)
85 Wt. Z Calcium Hydroxide
15 Wt. Z Anhydrous Amorphous 7.8 9.6
Silica (Syloid 244)
85 Wt. Z Calcium Hydroxide
15 Wt. % Colloidal Silica 3.6 4.8
(Ludox HS 40)
(1) For pelleting, all mixtures are blended with 3.0 wt. % Sterotex die lubricant.
(2) 48 hours @ 95°C/6 hours <§ 150°C/16 hours @ 315°C
68-2
87-3
68-3
71
-------
TABLE III-7
WET MIXTURES OF CALCIUM HYDROXIDE WITH ALDMHAS AHB ALDMITOW SILICATES
Composition (1)
78.4 Wt. % Calcium Hydroxide
21.6 We. % Hydrated Alumina
(Alcoa C-330)(3)
92.8 Wt. Z Calcium Hydroxide
7.2 Wt. % Hydrated Alumina
(Alcoa C-330)
83 Wt. % Calcium Hydroxide
17 Wt. % Kaolinite (uncalcined)
(Hydrite PX)(4)
85 Wt. % Calcium Hydroxide
15 Wt. Z Kaolinite (calcined)
(Glomax PJD)(5)
85 Wt. % Calcium Hydroxide
15 Wt. % Bentonite
(Bentolite L-2) (6)
Crush Strength Of
3.2 mm Pellets, kg
After Heat
Initial Treating (2)
Laboratory Reactor Test; 100.000 v/v/hr.. 370*C
6.4
8.1
4.8
5.1
6.1
9.6
11.0
11.4
8.6
12.4
Total Exposure
To Feed. Hours
1.3
3.5
6.0
1.3
4.0
6.0
1.3
4.0
6.0
S04* Trapping
Efficiency. %
83
55
56
75
57
54
63
57
55
S02 Trapping
Efficiency. %
0
0
23
15
10
Final Crush Notebook
Strength of Pellets, kg Reference 4349
13
0
0
12.7
11.5
68-4-
68-5
68-7
68-8
68-6
H
a
11.7
(1) For pelleting, all mixtures are blended with 3.0 wt. % Sterotex die Lubricant.
(2) 48 hours @ 95eC/6 hours <§ 150°C/16 hours @ 315°C
(3) Alcoa C-330 Hydrated Alumina, 65.0 wt. Z A1203, balance water.
(4) Hydrite PX, 45.3% Si02, 38.4% Al2<>3.
(5) Glomax PJD, 53.8% Si02, 44.4% A1203-
(6) Bentolite L-2, 71.5% Si02, 15.4% Al2<>3.
-------
- 19 -
strength of 3.2 kg. The synthetic silicas produced somewhat softer
pellets than those obtained with the diatomites. Using 15 wt % of
hydrated amorphous silica, the crush strength is 7.4 kg, and with 5%
the crush strength is 10.9 kg. With 15 wt % of anhydrous amorphous
silica, crush strength is 9.6 kg. Relatively soft pellets (4.8 kg)
were obtained using 15 wt % silica derived from a silica colloidal
dispersion.
Table II1-7 lists test results with combinations of calcium
hydroxide and alumina or aluminum silicates, two formulations with
hydrated alumina were prepared corresponding to 5 wt % and 15 wt %
A1203, giving pellets of crush strength 11.0 and 9.6 kg, respectively.
These values are slightly higher than those obtained using 5 and 15
wt % hydrated amorphous silica (9.6 and 7.4 kg, respectively), and
comparable to values obtained with 5 and 15 wt % uncalcined diatomite
(11.0 and 12.8 kg respectively). Very hard pellets were obtained
with 17 wt % uncalcined kaolinite (11.4 kg) and 15 wt % bentonite
(12.4 kg). Pellets from 15 wt % of calcined kaolinite were softer than
those obtained using uncalcined kaolinite (8.6 vs 11.4 kg).
III.3.2.2.2 Use of Portland and Calcium Aluminate Cements
Portland cement and calcium aluminate cement are widely used
hydraulic cements consisting of calcium silicates and calcium aluminates.
These are fine powders of specified composition and properties, which
harden upon mixing with water. We agglomerated powdered calcium hydroxide
(or calcium carbonate) by mixing the calcium compound with a minor portion
of cement and then blending with water to make a paste.
We selected a Portland Type I cement which consists typically
of 44.0% tricalcium silicate, 22.3% dicalcium silicate, 14.3% tricalcium
aluminate, and 8.0% tetracalcium aluminoferrite. Calculated CaO content
is 61.8%, S102 is 19.4%, and A1203 is 7.1%. The calcium aluminate
cement consists largely of mixed calcium aluminates, mainly monocalcium
aluminate. Calculated CaO content is 35.8%, A1203 + T102 is 44.0% and
S102 is 8.6%.
Four wet mixtures were prepared with Ca(OH)2 and commercial
hydraulic cements, as shown in Table III-8. The cements consist mainly
of calcium silicates and calcium aluminates which may be represented as
compositions of CaO, S102 and A1203. In order to obtain correspondence
to other sorbents which contain 15 wt % S102 or A1203, we chose to base
the quantity of cement used in each formulation on the portion of the
cement which is not "CaO". For example, Portland Type I cement consists
of 61.8 wt % CaO and, by definition, 38.2 wt % "non-CaO". Therefore, a
combination consisting of 60.7 wt % calcium hydroxide and 39.3 wt %
Portland cement is equivalent to 60.7 wt % calcium hydroxide, 24.3 wt. %
CaO, and 15.0 wt % "non-CaO". These four preparations were made using
-------
TABLE III-8
i OF CALCIUM HYDROXIDE
Crush
3.2 mm
Initial
5.4
11.2
(3)
6.8
(3)
9.9
1.8
(4)
Strength of
Pellets, kg
After Heat
Treating (2)
11.9
12.7
(3)
11.5
(3)
11.5
4.2
(4)
AHD CALCIUM CARBONATE WITH PORTLAffl)
CDfOn AHD CALCIUM ALU81NATE LBULN'I
Laboratory Reactor Test: 100,000 v/v/hr. , 370°C
Total Exposure
To Feed, Hours
1.3
4.0
6.0
L.2
3.7
5.7
1.3
3.8
5.8
S04= Trapping
Efficiency, Z
75
76
70
85
84
76
64
30
72
S02 Trapping Final Crush Notebook
Efficiency, Z Strength of Pellets, kg Reference 4349
3 68-9
0
3 13.0
93
23 68-10
10
10 (3)
91
0 43
0
0 (3)
88-1
88-2
92
Composition (1)
68.0 Wt. Z Calcium Hydroxide
32.0 Wt. Z Portland Type I
Cement
60.7 Wt. % Calcium Hydroxide
39.3 Wt. % Portland Type I
Cement
79.0 Wt. Z Calcium Hydroxide
21.0 Wt. Z Calcium Aluminate
Cement (Altas Refcon)
81.0 Wt. % Calcium Hydroxide
19.0 Wt. % Calcium Aluminate
Cement (Alcoa CA-25)
73.2 Wt. % Calcium Carbonate
26.8 Wt. Z Portland Type I
Cement
60.7 Wt. Z Calcium Carbonate
39.3 Wt. Z Portland Type I
Cement
76.6 Wt. Z Calcium Carbonate
23.4 Wt. Z Calcium Aluminate
Cement (Atlas Refcon)
89.0 Wt. Z Calcium Carbonate
19.0 Wt. Z Calcium Aluminate
Cement (Alcoa CA-25)
(1) For Pelleting, all mixtures are blended with 3 wt. Z Sterotex, (2) 48 hours @ 95°C/6 hours Q 150°C/16 hours
-------
- 21 -
32 and 39.3 wt % of Portland Type I cement, 21.0 wt % of Atlas Refcon
calcium aluminate cement, and 19.0 wt % of Alcoa CA-25 calcium aluminate
cement.
All combinations of calcium hydroxide with Portland cement,
and with calcium aluminate cements, produced hard pellets, except for
the combination containing 21 wt % Atlas calcium aluminate cement.
This combination set to a very hard material, which was judged too hard
to be pelleted, so Instead was broken into 6-meah particles for reactor
testing. The combinations employing 32.0 and 39.3 wt % Portland cement
produced pellets of 11.9 and 12.7 kg crush strength, respectively.
Similar pellets were obtained with 19.0 wt % Alcoa calcium aluminate
cement (crush strength 11.5 kg).
Calcium carbonate is an alternative to calcium hydroxide for
combination with cements. Calcium carbonate has an advantage over
calcium hydroxide of undergoing less volume increase upon sulfation
(sulfated volume expansion ratio =1.39 for CaC03, vs. 1.56 for Ca(OH)2)»
One mixture of powdered calcium carbonate with 26.8 wt % Portland Type I
cement was prepared and made into a wet paste as above. This mixture
contained the equivalent of 73.2 wt % CaC03, 16.6 wt % CaO and 10.2
wt % "Si02/Al203". The paste solidified to a very hard material which
was broken up and reactor-tested directly, without being made into pellets.
Another mixture with 39.3 wt % Portland cement was also made, as were two
mixtures with calcium aluminate cement, at the 23.4 and 19.0 wt % levels.
The higher Portland cement level resulted in very hard pellets. The two
calcium aluminate mixtures gave unsatisfactory pellets.
III.3.3 Laboratory Reactor Tests of Sorbent Particles
III.3.3.1 Experimental Procedure
Using a laboratory reactor, the experimental sorbents described
in Sections III.3.1 and III.3.2 were evaluated for efficiency in trapping
804" and S02> Thirteen ml of sorbent (3.2 mm pellets unless otherwise
noted) were placed in a 25.4 mm diameter quartz tube reactor. A stream
of synthetic exhaust gas containing 12% each of C02 and H20, 20 ppm S02,
5 ppm 803, 3% Q£ and balance N2, was blended through a bank of rotometers,
and passed over the sorbent at 100,000 v/v/hr and 370°C. 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 to allow control and recording of
the bed temperature. Trapping efficiencies were calculated from 804" and
S02 concentrations of the gas stream entering and exiting the reactor bed.
-------
- 22 -
The experimental apparatus is shown schematically in Figure III-l.
Analysis of the gas stream for 802 and 803 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 C.
III.3.3.2 Experimental Results
Tables III-6 to III-9 list the sorption activities of the 10
candidate materials chosen from the preparations of Sections III.3.1 and
III.3.2 on the basis of pelleting properties. These Include samples
of wet mixtures of Ca(OH)2 with various silicas, aluminas, and aluminum
silicates, cement-containing combinations, and dry mixtures of Ca(OH)2
with S102 and/or A1203. In addition, a new benchmark preparation, of
nominal 85 Ca(OH)2/10 S102/5 NazO composition was also tested.
III.3.3.2.1 Benchmark Sorbent
As indicated in Table III-6, the benchmark sorbent of nominal
composition 85 CaO/10 S102/5 Na20 trapped 94% of the 804" in the feed
gas during the first 2.6 hours of the test. The trapping efficiency
fell to 72% after 6.9 hours of the test. In previous tests^1), the
804° trapping efficiency of a similar material, at 480°C and 100,000
v/v/hr was 100% after 8.4 hours. Also, at 480°C but 150,000 v/v/hr,
efficiency was 97% after 4.2 hours. Previous to the present test, the
benchmark sorbent was run only once at 370°C, and this was at 150,000
v/v/hr, rather than 100,000 v/v/hr, the space velocity now being used.
Under those conditions, 804*" trapping efficiency was 100% after 2.8 hours
and 36% after 4,7 hours. The current 804* trapping efficiencies of 94%
after 2.6 hours and 72% after 6.9 hours are roughly comparable to the
earlier results at the higher space velocity. The present reactor
testing conditions, employing the lower reactor temperature (370°C vs.
480°C), which is more typical of vehicle condition, constitutes a more
severe test for experimental sorbents, since the rates for chemical
reaction and diffusion through the sorbent may be expected to be lower
at the lower temperature.
Trapping efficiency for S02» as found in the present test, is
33-60% over a period of 6.9 hours, which is similar to sorption levels
which have previously been obtained with benchmark sorbent.
III.3.3.2.2 Dry Mixes
As indicated in Table III-9, pellets made from 100% calcium
hydroxide are Ineffective as a sorbent, giving essentially nil 804"
trapping efficiency. Two dry mixtures of calcium hydroxide with silica
and with alumina, have also been tested as described in Table III-9.
Trapping efficiencies for 804" are 46-68%. As expected for the lack
of sodium content in these mixtures, 802 trapping efficiency is near zero.
-------
- 23 -
FIGURE III-l
AfPAlATOS FOt UBOUTOtT 6CAKEKVK OF SOMEHT8
T
TR(
iiiii
AIR N2 / H2 S02
/ C°
C02
PEED BLENDING
WATER
%
1
A
/\
/i
,/ 1
'
VAPOR UJA.U -
50 HEATED LINE
i y
\s
\ /
/
1 /
.SORBENT
•^ BED
x VYCOR
t/REACTOR
xFURNACE
k^i
BYPASS
i
J~
s
S02 H2S04
TECO) (GOKS0YR-
ROSS COIL)
ANALYZERS
(THE BYPASS is LOCATED AFTER THE WATER VAPOR + $03 ARE ADDED,)
-------
TABLE III-9
Composition (1)
85 Ca(OH)2/15 S102
85 Ca(OH)2/15 A1203 (3)
1002 Calcium Hydroxide
Crush
3.2 mm
Initial
2.7
3.5
6.0
DRY MIXTURES
Strength Of
Pellets^ kg
After Heat
Treating (2)
4.1
4.5
8.6
OF CALCIUM HYDROXIDE WITH SILICA
Total Exposure
To Feed, Hours
1.3
3.3
1.2
3.6
5.6
3.3
4.5
7.0
9.0
Laboratory
S04™ Trapping
Efficiency, %
68
46
65
60
54
Q
19
4
0
AND/OR ALUMTHA
Reactor Test: 100^000 v/v/hr., 370°C
S02 Trapping Final Crush Notebook
Efficiency. Z Strength of Pellets, kg Reference 4349
8 33-1
5 5.3
5 33-2
0
3 4.4
5 24-B
5
5 6.6
8
(1) For pelleting all mixtures were blended with 3.0 wt. % Sterotex die lubricant.
(2) 48 hours @ 95°C/6 hours @ 150°C/16 hours @ 315°C
(3) A1203 from hydrated alumina (64.9 wt. % A1203 equivalent)
-------
- 25 -
III.3.3.2.3 Wet Mixes
Only one wet mixture with silica has been reactor tested.
This is the combination of calcium hydroxide with 15 wt % uncalcined
diatomite which made very hard pills (12.8 kg crush strength). As
indicated in Table III-6, S04= trapping efficiency varied from 57-71%
over the 5.7 hour test period. Trapping efficiency for 802 was only
12-20%. Of the combinations with alumina tested, (see Table III-7)
the two mixtures with 7.2 wt % hydrated alumina and 21.6 wt % hydrated
alumina were essentially equivalent, giving 804" trapping efficiencies
of 83-54% over a six hour period. One aluminum silicate (15% bentonite)
was also tested and had similar characteristics (trapping efficiency
63-55% over 6 hours).
III.3.3.2.4 Cements
Only one pelleted cement composition was tested - calcium
hydroxide with 32% Portland Type I cement (see Table III-8). The trapping
efficiency of this material for S04= was 70-76% over the 6 hour test
period. The S02 trapping efficiency was near zero. Two cement compositions
were tested as 6-mesh chips rather than as pellets. One of these was
formulated from calcium hydroxide and 21 wt % Atlas Refcon calcium
aluminate cement. Trapping efficiency for 804" of this material was
85-76%. Trapping efficiency for 802 was only 23-10%. The other compo-
sition tested as chips was made from calcium carbonate with 26.8 wt %
Portland cement. Trapping efficiency for S04= of this material was
30-72% and 802 trapping efficiency was near zero.
III.3.3.2.5 Discussion of Experimental Results
None of the experimental sorbents described above were as
efficient in trapping 804° as the benchmark sorbent. This may in part
be due to a beneficial effect due to the sodium component in the benchmark
sorbent. Also, because many of the materials produce such hard pellets,
the resultant porosity may be low. Later sections will describe attempts
to increase porosity through such techniques as calcining, use of burn-out
agents and pellet compaction to lower density levels.
III.3.4 Effect of Heat Treatment and Compaction on Pellets
III.3.4.1 Temperature Effects on Pellet Strength
A brief study was undertaken to determine the optimum temperature
for heat treating pallatized trap material. These studies were carried
out with the benchmark sorbent 85 CaO/10 8102/5 Na20, which had been
made into 3 mm pellets after drying at 65°C.
-------
- 26 -
Heat treatment drives off excess water and conditions the
pellets to temperatures that will be encountered in service. These
temperatures may be expected to range from 260-480°C. We also wanted
to know if greater pill strength can be achieved by calcining at much
higher temperatures.
In three test series, pellet samples were exposed to pro-
gressively greater temperatures up as high as 1100°C. After each
temperature interval, a sample of pellets was removed and tested for
crush strength. The data from these tests is given in Table III-10.
Over the temperature range studied for this particular compo-
sition, there was an optimum temperature range for pellet strength at
260-315°C. In this temperature range, crush strength reached a maximum
of 7.3-7.7 kg. Crush strength decreased to below 3 kg above 425°C,
and to 1.5 kg at 565°C.
It was found that calcining 5 hours at 540°C removes essentially
all water of hydration and converts calcium hydroxide to calcium oxide.
However, at 315° most of the calcium remains in the form of calcium
hydroxide. The equilibrium vapor pressure for water with calcium hydroxide
is 100 mm at 300°C and 760 mm at 512°C. It appears that dehydration of
calcium hydroxide above 315°C in the oven test is one reason for loss of
crush strength.
Another possible explanation is loss of water from calcium
silicate hydrates, which decompose between 300-800°C. Hydrated calcium
silicates form a gel-like structure which is a primary binding component
of materials such as Portland Cement. Probably such a structure con-
tributes to the strength of pellets of this composition.
Although loss of crush strength above 315°C has been demon-
strated under dry oven conditions, it is expected that in a vehicle,
where exhaust gas contains M.2% water, higher temperatures are allowed
without deterioration because of the equilibrium of water vapor with
hydrated sorbent.
III.3.4.2 Compaction Effects on Pellet Strength
and Sorption Activity
In order to investigate the effects of pelleting to lower
crush strengths, and heat treating at -650°C, three batches of pellets
were made from one preparation of 85 Ca(OH)2/15 Dlatomite, having
different crush strengths, by varying the compacting conditions. As
listed in Table III-ll, the weakest pellets, Batch 1, were made to an
initial crush strength of 2.9 kg. After heat treating at 315°C. these
pellets had a crush strength of 5.0 kg and density of 1.25 g/cnH. A
-------
TABLE 111-10
The Effect of Heat Treatment Conditions
on Pellet Crush Strength (1)
Series 1 Series
Temperature Hrs. Crush Strength Hrs. Crush
°C @ Temp. kg @ Temp.
65 64 6.6 (2) 64
.1.
95
150
205
260
315
370
425
455
480
«
' \
v
66
•V
8
v
(
16
\
f
8
540 5 2.2
\
t
565 16
4
1100 - 8
2 Series 3
Strength Hrs. Crush Strength
kg @ Temp. . kg
6.6 (2) 64 6.6 (2)
5.6 64 (3)
1 I 1
5.5 8 (3)
\
v \<
16 7.7
/
7.3 i
\
, i°
8 52"^
16 . 2.8
f
3.7
N
v \f
16 2.3
« _
^
1.5
^
1.1
(1) The above data are for 3-mm pellets produced from 85 CaOAOSi02/5 NaoO sorbent which had been dried 64 hours
@ 65°C prior to granulating and pelletizing. Notebook references 4349-4, 12, 13.
(2) These are initial crush strengths for the sorbent after pelletizing.
(3) No sample was taken for crush strength."
-------
TABLE III-ll
PREPARATION OF 85 Ca(OH)2/15 DIATOMITE PELLETS 07 DIFFERENT
CRUSH STRENGTHS WITH HEAT TREATING AT TWO CONDITIONS
Heat Treatment
After Pelleting
Batch 1
Crush Density
Strength, kg g/cm3
3.2 mm Pellets (1)
Batch 2
Crush
Strength, kg
Density
g/cm3
Batch 3
Crush Density
Strength, kg g/cm3
None
2.9
5.0
7.0
16 h @ 95°C
6 h @ 150°C
16 h @ 315°C
5.0
1.25
8.3
1.34
11.6
1.49
KJ
00
16 h @ 95°C
6 h @ 150°C
16 h @ 315°C
2 h @ 540°C
4 h @ 650°C
3.3
1.19
6.3
1.29
8.5
1.35
(1) Notebook Reference 4349-138. The different original crush strengths were obtained froa one
preparation of sorbent by using three different fill settings during pelleting, i.e., less
fill produces softer pellets.
-------
- 29 -
portion of these pellets was further heated at 650°C, giving pellets
of crush strength 3.3 kg, and density 1.19 g/cm3. Of the heat treated
pellets, these had the lowest crush strength and the lowest density.
As shown in Table 111-12, these also had the highest S04= trapping
efficiency yet obtained for sorbents of this composition (81-73% over
11.7 hours). The pellets heat treated at 315°C were only slightly lower
in efficiency (84-60% over 9.4 hours). Previous pellets of this sorbent
made at higher crush strength (12.5-12.8 kg) and higher density (1.58-
1.71 g/cm3) had lower 804™ trapping efficiencies (71-12% over 11.3 hours).
The lower activity of the pellets having higher crush strength and higher
density was probably a consequence of lower porosity and smaller pores.
It is apparent that a compromise has to be made between crush strength
and porosity, so as to obtain pellets having both adequate mechanical
strength and activity.
III.3.5 Increased Pellet Porosity
III.3.5.1 Pore Volume and Surface Area Relationship
Table 111-13 lists pore volume, surface area, and density measure-
ments on pellets of benchmark sorbent and pellets made from 100% calcium
hydroxide. In the laboratory reactor screening test, benchmark sorbent
was much more effective at sorbing S04= than was 100% Ca(OH)2 sorbent
(94-72% trapping efficiency vs. 0-20%). Referring to Table 111-13,
the superior trapping efficiency of benchmark sorbent is associated
with higher pore volume (0.238 vs. 0.142 ml/g) and lower density (1.407
vs. 1.744 g/cm3) relative to 100% Ca(OH)2 sorbent. However, total
surface area is less for benchmark sorbent (8.5 vs 16.3 m2/g). The
surface area is less even though total pore volume is higher, because
most of the pore volume occurs in large pores. This is apparent from
a distribution analysis of pore volume and surface area versus pore
size. Considering pore sizes in the range of 25-1200QA, benchmark
sorbent has 99% of its porosity in pores of 200-1200 A, and 75% in pores
of 600-1200 A. However, 100% Ca(OH)2 sorbent has 38% of0its porosity
in pores of 25-200 A, and only 19% in pores of 600-1200 A. Consequently,
93% of the surface area in benchmark sorbent occurs in pores of 200-1200
A, with 55% in the 600-1200 A range. This analysis only covers pores up
to 1200 A. Because in the case of benchmark sorbent there is a sizable
contribution to pore volume by the largest pores in this range, we may
expect that significantopore volume and surface area also exists due to
pores larger than 1200 A. In contrast, 100% Ca(OH)2 sorbent has 74%
of its surface area in pores of 25-200 A,045% in the smallest pores of
25-100 1, and only 4% in pores above 600 A. The reactivity of benchmark
sorbent is apparently a result of substantial porosity in the 600 + A
size range. As discussed in the following sections, this information
was applied to the design of potentially more efficient sorbents.
Additional studies on the relationship of sorbent physical properties
and sorption activity may be found in Section III.7.
-------
TABLE 111-12
85 Ca(OH)2 /15 DIATOMITE SORBENT
EFFECTS OF CRUSH STRENGTH AND HEAT TREATING ON DENSITY AND ACTIVITY
at Treating Conditions
48 h @ 95°C
6 h @ 150°C
16 h @ 315°C
48 h @ 95CC
6 h @ 150°C
16 h @ 315°C
5 h <§ 540°C
16 h @ 95°C
6 h @ 150°C
16 h @ 315°C
16 h @ 95°C
6 h @ 150°C
16 h @ 315°C
2 h @ 540°C
4 h
-------
- 31 -
TABLE 111-13
PORE VOLUME, SURFACE AREA AMD DENSITY OF SORBENTS
VS. SPA" TRAPPIHC EFFICIENCY
Pore Volume, ml/g
Surface Area (BET) sq. m./g
(2)
Pore Volume Distribution, Z '
25- 100 A
100- 200 A
200- 600 A
600-1200 A
Surface Area Distribution, !T
25- 100 1
100- 200 1
200- 600 A
600-1200 1
Density, g/cm^
804 Trapping Efficiency, %
Benchmark Sorbent ,..,
"85 CaO/10 Si02/5 Na?0"U'
0.238
8.5
0.2
1.2
23.6
75.0
2.3
5.3
37.6
54.7
1.407
94-72
1 100* Ca(OH)9(1)
0.142
16.3
15.1
22.8
42.9
19.1
45.4
28.2
22.7
3.7
1.744
0-20
(1) 3.2 mm pellets heat treated 16 hours at 315°C.
(2) Pore volume and surface area distributions are for pore sizes
in th| range of 25 - 1200 A. Contributions from pores larger than
1200 A are not included in these analyses, although contributions
from these may be appreciable for the benchmark sorbent.
-------
- 32 -
III.3.5.2 Heat Treating
In an initial attempt to increase 804" trapping efficiency by
increasing porosity, nine samples of pelleted sorbents were heated at
540°C, with the intent to dehydrate Ca(OH)2 to CaO. The vapor pressure
of water due to dissociation of Ca(OH)2 at 540°C is approximately 180
kPa (1.75 atmospheres). At 512°C the pressure is 101 kPa (1 atmosphere).
We had previously decided to limit heat treatment of pelleted sorbents
to 315°C so as to avoid dehydration of Ca(OH)2 while obtaining maximum
crush strength. As discussed in Section III.3.4.1 crush strength of
benchmark sorbent reaches a maximum at 315°C and decreases at higher
temperatures. Consequently, we were concerned with the effect of heat
treating on crush strength as well as the effect on porosity and reactivity.
Table 111-14 lists inspections on six samples of pellets which
were heat treated at 540°C after initial treatment at 315°C. As expected,
the crush strength of benchmark sorbent was much reduced - from 8.8 kg
to 2.6 kg. Less severe reductions in crush strength occurred with the
other sorbents. Pellets made from a wet mixture of 15 wt % uncalcined
diatomite in calcium hydroxide experienced the least loss in crush
strength (12.8 kg vs. 12.5 kg). All pellet samples decreased in density
by as much as 11% as a result of weight loss. Essentially no change
occurred in pellet dimensions. The weight loss was only half (or less)
that which would be expected if all the Ca(OH)2 were converted to CaO.
Apparently, much of the water bound to calcium in pelleted sorbents is
more tightly bound than is water in pure Ca(OH)2- An alternate explanation
is that the pellets are insufficiently porous but sufficiently strong to
contain water vapor in excess of one atmosphere internal pressure.
Three of these sorbents were tested for 804° trapping efficiency
in the laboratory reactor. In Table III-15 are listed 804° and 802
trapping efficiencies for benchmark sorbent, 85 Ca(OH)2/15 diatomite
sorbent, and 78.4 Ca(OH)2/21.6 hydrated alumina sorbent. Results are
given for pellets tested after heating at 315°C (our previous heat
treating temperature) and after heating at 540°C. There is np_ increase
in S04= trapping efficiency of any of these sorbents as a result of
heating at the higher temperature. In the case of the diatomite and
hydrated alumina preparations, S04= trapping efficiencies appear to have
decreased slightly, and S02 trapping efficiencies have increased. The
802 trapping efficiency of the alumina-containing sorbent has increased
from 0 to 30-49%. Based on these results, and the pore volume/surface
area measurements discussed in the preceding section, it is concluded that
heating at 540°C does not produce the large pores which are apparently
required for adequate S04= trapping efficiency.
-------
TABLE 111-14
EFFECT OF HEAT TREATING AT 315'C AND 540'C ON
CRUSH STRENGTH AND DENSITY OF PELLETED SORBENT COMPOSITIONS
Composition (3.2-mm pellets)
Benchmark 85 CaO/10 Si02/5 NaaO
100% Ca(OH)2
85 wt. Z Ca(om2
15 wt Z Diatomite (uncalcined)
78.4 wt. Z Ca(OH)2
21.6 wt. Z Hydrated Alumina
83 wt. Z Ca(OH)2
17 wt. Z Kaolinite (uncalcined)
68 wt. Z Ca(OH)2
32 wt. Z Portland Type I Cement
Crush
315«c(l)
8.8
9.9
12.8
9.6
11.4
11.9
Strength (kg)
540°C(2)
2.6
7.3
12.5
6.9
10.0
8.5
Density,
315°C
1.407
1.744
1.714
1.683
1.709
1.735
B/cm3
540°C Notebook Reference 4349
1.279 68-11
1.648 68-11
1.577 68-1
1.531 68-4
1.540 68-7
1.680 68-9
U)
u>
(1) After 16 hours at 315*C.
(2) Pellets which had been heated 16 hours at 315*C were further heated 5 hours at 540*C.
-------
TABLE 111-15
804" AND S02 TRAPPING EFFICIENCIES OF PELLETS HEAT TREATED AT 540°C
VS. EFFICIENCIES AFTER HEAT TREATING AT 315°C
Composition
Benchmark Sorbent
"85 CaO/10 S102/5 Na20"
85 Wt:. % Calcium Hydroxide
15 We. % Diatomite (uncal-
cined) (Dicalite SA3)
78.4 Wt. % Calcium Hydroxide
21.6 Wt. % Hydrated Alumina
(Alcoa C-330)
Laboratory Reactor Screening Test:
Pellets After Heat
% Change
In Density
After Heating
5 Hours @
540°C CD
-9.1%
i
-8.0%
de
a -9.0%
Total
Exposure
To Feed
Gas, Hours
0.9
1.3
2.6
3.1
5.3
6.6
6.9
8.9
10.9
12.2
15.4
17.4
1.3
3.4
3.7
5.4
5.7
6.7
9.3
11.3
1.2
1.3
3.2
3.5
4.5
6.0
7.4
9.4
Treating
804=
Trapping
Efficiency,
%
94
94
—
—
—
72
—
—
— —
—
—
57
— —
71
—
62
—
—
— —
-
83
—
55
~
56
— -
_—
@ 315°C
S02
Trapping
Efficiency,
%
60
48
—
—
—
33
—
—
—
—
—
20
—
12
—
12
—
—
— —
,-, ,
0
—
—
—
0
—
—
100,000 v/v/hr., 370°C
Pellets After Heat
Treating
SO^"
Trapping
Efficiency,
%
95
__
—
88
81
73
—
67
63
50
45
54
70
34
—
28
—
23
14
12
67
—
51
—
45
—
33
30
@ 540°C
SO 2
Trapping
Efficiency,
%
79
— —
—
64
58
57
—
43
43
42
42
39
31
25
—
28
—
16
26
26
49
— .
43
—
30
—
38
38
Notebook
Reference
4349
44
68-1
68-4
(!•)•• The density of 3.2'mm pellets which were-initially heat-treated at—315°C was remeasured-after heat treating at 540°C.
-------
- 35 -
III.3.5.3 Blowing Agents
Sorbent preparations discussed in previous sections have had
inadequate reactivity for 804"°, apparently because pores are too small.
In an initial attempt to generate large pores, compositions of 85 Ca(OH)2/15
diatomite were prepared using 5 and 10 wt % of added cellulose fibers, of
three different sizes (50-60 ym, 100-140 ym, and 290 ym in length). Of
these compositions, only those containing 5 and 10 wt % of the smallest
cellulose fibers (50-60 ym) were successfully pelleted, although the 10%
composition had to be pelleted in two passes, i.e., soft pellets were
made which were reground and repressed to make satisfactory pellets.
The resulting pellets were heated at 540°C for 70 hours to burn off the
cellulose fibers. In Table 111-16 are listed crush strengths and
densities of these pellets vs. those properties obtained with sorbent
without cellulose, and with benchmark sorbent. As a result of incorpo-
ration of cellulose fibers, crush strengths were reduced from 12.5 kg
(without cellulose) to 8.6 and 6.5 kg for 5 and 10% cellulose, respectively.
These crush strengths were still satisfactory. Densities decreased as a
result of weight loss from 1.714 g/cm^ (without cellulose) to 1.379 and
1.407 with 5 and 10% cellulose, respectively. These are the lowest
densities which we have measured for pellets of other than benchmark
composition. The decrease in density due to 5% cellulose is 20%. It
was surprising that the higher level of cellulose did not provide even
lower density; however, this was probably a result of the different
compacting conditions.
Laboratory reactor tests for the sorbent made with 5% cellulose,
vs. that without cellulose, are given in Table III-17. After ^3.5 hours
on test, the difference in 804 sorption efficiencies of the two sorbents
is 20-25%, with the cellulose composition having the higher efficiency.
However, even these efficiencies are 20-30% lower than those obtained
with benchmark sorbent.
A wet mixture preparation of CaO/SiO/Na20 plus 10% cellulose
after burn-out of cellulose showed S04= trapping efficiency almost as
good as benchmark sorbent prepared by the coprecipitation method (Table
111-18). The sorbent's trapping efficiency during a 10 hour evaluation
ranged from 90 to 77%. However, its S02 trapping efficiency was high,
69-33%. The Inclusion of cellulose and then burn-out apparently helped
improve the sorbent's porosity and activity. A previous wet mixture of
CaO/SiO/Na20 had shown much poorer S04= trapping efficiency: 72-55%.
The sorbent was prepared by mixing the dry Ingredients together
and then sufficient water was added to make a paste. The material was
dried at 65°C and granulated to 14-40 mesh and pressed into 3.2 mm pellets.
The CaO/SiO/Na20 + 10% cellulose was heat treated in a Vycor reactor in
a tube furnace, initially at 340°C under nitrogen, then in air and the
-------
TABLE 111-16
CRUSH STRENGTH AND DENSITY OF PELLETED SORBENTS
Composition (3.2-mm pellets)
Benchiiark 85 CaO/10 S102/5 Na20
Hours At Average Crush Average Density
540°C Strength (kg) g/cm3 Notebook Reference 4349
2.6
1.279
68-11
85 wt., % Ca(OH)2
15 wt, % Diatomite (uncalcined)
12.5
1.577
68-1
85 wt, % Ca(OH)2
15 wt,, % Diatomite (uncalcined)
pluH 5 wt. Z Solka-Floc BW-40
" (Cellulose Fibers)
70
8.6
1.379
122-2
OJ
ON
85 wt,, % Ca(OH)2
15 wt. Z Dlatoalte (uncalcined)
pluu 10 wt. Z Solka-Floc BW-40
" (Cellulose Fiber*)
70
6.5
1.407
122-1
-------
TABLE 111-17
S04" AND S02 TRAPPING EFFICIENCIES OF PELLETED SORBENTS
Crush Strength (kg)
Composition (3.2-mm Pellets)
Benchmark Sotbent
V85 CaO/10 S102/5 Na20"
NB-4349-44
Actual Na20 content • 4.5%
Heat treated 5 hrs. @ 540°C
Initial
(After 540°C)
2.6
After Reactor
Tcjt (370°C)
6.8
85 wt. % Calcium Hydroxide
15 wt. % Diatomite (uncalcined)
(Dicalite SA3)
NB-4349-68-1
Heat treated 5 hrs. @ 540°C
85 wt. % Calcium Hydroxide
15 wt. % Diatomite (uncalcined)
(Dicalite SA3)
plufl 5 wt. % added
50-60 ym Cellulose Fibers
(Solka-Floc BW-40)
NB-4349-122-2
Heat treated 70 hrs. @ 540°C
12.5
11.6
8.6
7.7
Laboratory Reactor Screening Test:
100.000 v/v/hr.. 370°C
Total
Exposure
To Feed
Gas, Hours
0.9
3.1
5.3
6.6
8.9
10.9
12.2
15.4
17.4
1.3
3.4
5.4
6.7
9.3
11.3
1.3
3.5
4.8
9.3
10.5
15.3
S04"
Trapping
Efficiency, %
95
88
81
73
67
63
50
45
54
70
34
28
23
14
12
80
57
53
44
35
31
S02
Trapping
Efficiency, %
53
38
32
31
17
17
16
16
13
5
0
2
0
0
0
23
17
15
20
12
15
I
CO
-------
TABLE 111-18
EVALUATION OF SORBENTS IN THE LABORATORY SCREENING TEST
Composition
CaO/S10/Na20 + 10% Cellulose
(83/10/8)
(KB 4349-152A)
Ca(OH)2 + Ludox HS
(70/30)
(NB 4349-71)
Carbon (Calgon)
Staple A
Maple B
Final Heat
Treating Density
Conditions g/cm3
3 hrs @ 1.28
540°C
16 hrs @ 1.40
315°C
Laboratory
100,
Total
Crush Strength, kg Exposure
Reactor
000 v/v/h
SOA =
After Reactor To Feed Trapping
Initial Test Gas, Hours Efficiency
4.2 8.2 1.0
3.0
5.0
6.0
8.0
10.0
4.8 6.1 1.0
3.8
6.8
8.8
11.6
>13.5 11.9 1.0
3.0
5.5
6.5
8.4
11.6
13.5 11.0 1.3
3.0
4.8
5.8
9.1
90
78
87
78
73
77
54
49 •
46
45
24
62
70
65
59
68
65
55
50
73
63
76
Screening Test
, 370°C
S02
Trapping
, % Efficiency, %
69
47
47
39
33
33
^
26
18
. 31
25
25
3
14
3
11
8
11
0
3
3
3
0
-------
- 39 -
temperature was gradually raised to 540°C and maintained at 540°C for
three hours. The burn-out of the cellulose was judged complete when
all the grey coloration of the pellets disappeared.
In Table 111-19 are listed data obtained with four preparations
of 85 Ca(OH)2/15 dlatomite made with the following cellulose or starch
additions: (1) no addition, (2) 5 wt % of 50-60 pm cellulose fibers,
(3) 10 wt % of 25 pm cellulose fibers, (4) 10 wt % of starch, and (5)
20 wt % of starch. Although in general, densities were lower as a
result of burning out the starch or cellulose components, SQ^~ trapping
efficiencies were not improved as much as desired. Further, inspection
of density data before and after burn-out indicated that overall shrinkage
of pellets could be significant over the extended burn-out times employed
(70-120 hours). In one case, pellet shrinkage was sufficient to result
in a net density increase after burning out a cellulose or starch component.
The last preparation in Table 111-19, which was made with 20% starch, was
heat treated for 16 hours at 315°C and then for 71 hours at 540°C. The
average pellet height decreased from 3.21 mm (after 315°C) to 3.08 mm,
(after 540°C), diameter decreased from 3.16 to 3.09 mm, and weight de-
creased from 35.3 to 32.6 mg. As a result of combined shrinkage and
weight loss, the density after burn-out at 540°C was actually slightly
greater (1.41 g/cm3) than the density after 315°C (1.40 g/cnH). It is
likely that such shrinkage is closing the large pores generated by the
burn-out procedure, although no actual pore size analyses are available.
III.3.6 Effect of Sodium on Sorbent Activity
One of our intentions in preparing sorbent formulations has
been to avoid use of sodium so as to minimize sorption of S02- It
appears, however, that some sodium may be required In order to obtain
high S04= trapping efficiency. Moreover, we have found that elimination
of sodium does not entirely eliminate S02 sorption, and that Na20 content
on the order of 5-7% may be tolerable in this respect.
III.3.6.1 Sodium Content of Benchmark Sorbent
In Table 111-20 are listed reactor test results for two prepa-
rations of "85 CaO/10 S102/5 Na20" benchmark sorbent, each tested two
times. These two preparations differed in actual sodium content (as a
result of different wash procedures) and also in 804= trapping efficiency.
The first preparation contained 4.5 wt % equivalent Na20 content. Its
S04= trapping efficiency was 95-54% over 17.4 hours. The second prepa-
ration, which contained 7.5 wt % equivalent Na20 had trapping efficiencies
of 97-85% over 18.2 hours. After ^8 hours on test, S02 trapping
efficiencies were below 30%.
-------
TABLE 111-19
SORBENTS WITH CELLULOSE AHD STARCH
position
85 Ca(OH)2/15 Dtatomite
Plus Starch or Cellulose as Indicated
No Starch or Cellulose
(UB-4349-68-1)
Final Heat
Treating
Conditions
5 h @ 540°C
5 wt. Z Solka-Floc
50-60 urn Cellulose Fibers
(NB-4349-122-2)
10 wt. Z Solka-Floc BW-300
25 pm Cellulose Fibers
(BB-4349-148-1)
10 wt. Z Soluble Starch (MC&B)(2)
(NB-4349-125)
20 wt. t CFC Buffalo Starch 3401
Corn Starch
(10-4349-148-3)
Density
g/cm3
1.58
Crush Strength, kg
After Reactor
Initial Test
70 h @ 540°C 1.38
71 h @ 540°C 1.39
120 h § 5408C 1.47
71 h 8 540*C 1.41
12.5
11.6
8.6
7.7
7.1
a.a
8.4
6.5
10.2
Laboratory Reactor Screening Test
100.000 v/v/h. 370"C
Total
Exposure
To Feed
Gas, Hours
1.3
3.4
5.4
6.7
9.3
11.3
1.3
3.5
4.8
9.3
10.5
15.3
1.3
3.7
5.7
7.0
9.5
11.5
1.3
4.3
6.3
so4-
Trapping
Efficiency, Z
70
34
28
23
' 14
12
80
57
53
44
35
31
19
40
30
30
31
45
86
61
56
S02
Trapping
Efficiency, X
31
25
28
16
26
26
23
17
15
20
12
15
31
19
19
14
14
11
25
23
23
NOT EVALUATED IN SCREENING TEST
o
I.
(i. fee also Table 111-17
(2) See also Table 111-22
-------
TABLE II1-20
REACTOR TESTS OF "BENCHMARK" SOHBEHT
Notebook Reference
4349-44
4349-44
Equivalent
Na20
Content
4.5
4.5
4349-120
7.5
4349-120
7.5
Heat
Treating
Conditions
16 h @ 315 8C
16 h
-------
- 42 -
III.3.6.2 Sodium Content of Other Sorbents
Another example of the effect of sodium content on trapping
efficiency is given in Table 111-21. Two preparations of "(CaO)3
(Si02 Al203)i/2" contained, respectively, 14.6 and 0.5 wt % equivalent
Na20. Again, the sorbent having the higher level of sodium had the
higher 804° trapping efficiency (94-90% vs. 84-45% over 11 hours).
Trapping efficiencies for 862 were 32-21% for the sorbent containing
0.5% Na20, and 61-42? for the sorbent containing 14.6% Na20.
III.3.7 Effect of Pore and Surface Area
Properties on Sorption Efficiency
A variety of pelleted sorbents, covering a range of activities
for sorption of 804"°, were analyzed for pore volume, pore size distribu-
tion and surface area. These analyses were obtained by the mercury
penetration method run to 24,000 psi. A summary of the data are reported
in Table 111-22 and also included for comparison are other laboratory test
results: heat treating conditions; crush strength; and S04= trapping
efficiency.
The results showed that as the crush strength of the pellets
increased, directionally the pore volume above 500 A and the surface area
decreased. The data also showed that an increase in porosity probably
helps S04= sorption, but is not the only controlling factor. Apparently,
the pore structure affects the selectivity, however, the results suggest
that the activity of various components of the sorbent is an important
variable. The Na20 in many of the sorbents prepared improved the 804°*
sorption. Unfortunately, it also improved 802 sorption.
III.3.8 Chemical Analysis of Used Sorbents
III.3.8.1 Rings from Vehicle Testing
As part of Modification I of this contract, a vehicle durability
test had been performed on benchmark sorbent fabricated into rings having
the dimensions 15.9 mm O.D., 5.6 mm I.D., and 7.9 mm height'^'. The test
was discontinued in March, 1975, after 20 400 kilometers accumulation,
because of unsatisfactory 804° trapping efficiency. Subsequently, in
June, the trap was removed from the vehicle and cut open. The rings
were still intact, although highly compacted and seemingly "cemented"
together. The rings were hard, and there was very little fines. Volume
expansion of individual rings was estimated to be ^30%. The opened trap
was stored loosely covered until samples were to be taken for analysis.
After standing five weeks, the hard rings had become very soft, which
made sampling of individual rings very difficult. Samples were taken
from the inlet area of the trap and from the exit area. A few intact
rings were obtained from the middle of the trap. Samples of fresh rings
and rings from the trap were analyzed for Ca(OH)2» 804"*, and C03*.
-------
TABLE 111-21
Nominal Composition (1)
"(CaO)3 (Si02 Al203)1/2"
NB 4349-64
"(CaO)3 (S102 Al203)1/2"
NB 4349-41
(CaO)3 (Si02 Al203)i/2 SORBENT PRETARATIONS ;
EFFECT OF SODIUM CONTENT ON SO^" TRAPPING EFFICIENCY
Laboratory
100,
Equivalent Total
Na?0 Crush Strength (kg) Exposure
Content Initial After Reactor To Feed
Wt. % (2) (After 315°C) Test (370°C) Gas, Hours
0.5 4.8 7.0 1.3
4.8
6.1
9.0
11.0
14.6 5.0 7.8 2.4
4.4
5.8
8.5
10.5
Reactor Screening
000 v/v/h, 370°C
504=
Trapping
Efficiency, Z
84
33
75
56
45
90
93
94
94
94
Test
S02
Trapping
Efficiency, X
32
21
32
32
»
30 5
I
61
56
53
42
42
(1) 3.2 mm pellets heat treated 16 h @ 315°C
(2) Analysis after calcining 5 h @ 815°C by emission spectroscopy.
-------
TABLE 111-22
Sorbent
Ca(OH)2
Ca(OH)2(Dicalite)
(85/15)
Ca(OH)2(Dicalite)
(85/15)
Ca(OH)2(Dicalite) + Starch
(77/14/9)
Ca(OH)2(Dicalite) + 10*
, Solka Floe BW-AO
Ca(OH)2(Dicallte) + 5%
Solka Floe BW-40
Ca(OH)2(S102)
(85/15)
CaO/SiO/Na20
(85/10/5)
(CaO)3(S102-AL203)jj Na20(0.5Z)
Na20(14.6Z)
LABORATORY TEST
Heat
Treating
Conditions
°C
315
315
540
650
315
650
315
650
315
540
540
540
' 315
315
315
315
315
315
Crush
Strength
kg
9.9
12.5
9.9
3.3
5.0
6.3
8.3
8.5
11.6
8.4
6.5
8.6
4.1.
12.2
6.6
8.8
6.3
.5.0
Cum. Z of
Fore Volume
Above
500 X
20
12
20
60
40
50
54
48
30
60
50
50
16
40
76
76
54
92
Above
100 X
80
76
80
96
86
94
88
96
80
94
94
94
70
100
100
100
94
98
Cum. Surface Area
Above
500 X
0.74
0.65
0.69
5.5
2.1
4.9
3.1
4.1
1.4
3.5
2.8
3.7
0.34
2.1
4.5
.5-2
4.5
2.2
Above
100 A
8.7
6.6
22.6
23.6
13.0
21.5
14.1
17.3
12.1.
12.9
19.2
20.1
8.0
8.9
8.8
10.0
18.0
3.1
Laboratory Reactor
Screening Test
100,000 v/v/h, 370°C
Test
Hours
9.0
6
11
12
9
6
15
3
10
7
11
11
Trapping Efficiency. Z
20-0 10-5
71 -
70 -
78 -
83 -
86 -
80 -
68 -
94 -
94 -
84 -
94 -
62
12
73
63
56
31
46
93
72
45
93
20 -
31 -
32 -
24 -
25 -
23 -
8 -
41 -
60 -
32 -
61 -
12
26
0
8
23
15
5
29
33
30
42
Notebook
Reference
4349- 68-11
4349-68-1
It
4349-138-1A
4349-138-1
4349-138-2A
4349-138-2
4349-138- 3A
4349-138-3
4349-125
4349-122-1
4349-122-2
4349-33-1
4349-120
4349-10
4349-44
4349-64
4349-41
-------
- 45 -
The results of these analyses are given in Table 111-23. The analyses
for 804= and C(>3= are given in terms of wt % CaS04 and wt % CaC03,
although Na2S04 and Na2C03 are certainly also present. The fresh rings
contained 58 wt % Ca(OH)2 and 11 wt % CaCOs, the latter arising by
sorption of C02 from the air during preparation and storage. The used
rings from the inlet and exit sections of the trap contained only 2%
Ca(OH)2- The concentrations of CaSCty in the inlet and exit samples
were 17% and 12%, respectively. Because of the very low Ca(OH)2
concentrations versus the much higher CaS04 concentrations, the softening
of the rings during storage was probably due to hydration (and con-
sequent expansion) of CaSCty (or Na2S04) rather than to hydration of
any CaO which may have been present in the trap. Equivalent CaC03
contents for the inlet and outlet samples were found to be 36 and 57
wt %. The lower analysis for these three Ca compounds for the inlet
rings may be due to the presence of unanalyzed CaO and/or CaSi03. The
higher inlet temperature would make these more likely than at the trap
outlet.
Several intact rings were scraped to obtain samples from the
outer surface, the inner surface, and the interior 1/3 of the rings.
The 804" concentrations (expressed as CaS04) of samples from the outer
and inner surfaces were 13.8% and 12.6%, respectively. The concentration
in the interior_of the rings was 9.7%. These data demonstrate that
significant 864 collection occurred in the interior portion of the
ring, as was found previously with 3.2 mm pellets^).
The bulk concentration of CaSC>4 in the rings, 12-17% after
20 400 km, was much less than the concentration (55%) previously found
in the 3.2 mm pellets after 42 400 km^1'. Although the degree of
sulfation was lower for the rings, S04= sorption efficiency was much
inferior, for example, ^30% for the rings at 15 000 km, vs. >95% for
the pelleted material. The low trapping efficiency was most likely
associated with the much lower geometric surface area of the ring bed
relative to 3.2 mm pellets.
III.3.8.2 Pellets from Laboratory Testing
The data obtained with the used rings were very similar to
those obtained with 3.2 mm pellets after laboratory reactor tests using
synthetic exhaust gas. The fresh pellets were found to contain 53%
Ca(OH)2, nil CaS04, and 23 wt % CaCO,,. After the reactor test, the
pellets contained only 5% Ca(OH)2, 2% CaS04, and 72% CaCC-3. These data
demonstrate that the calcium hydroxide component of such sorbent compositions
reacted rapidly and essentially completely with C02 from the exhaust gas,
and was converted to CaCC>3, which became the active component for reacting
with 804".
For the pelleted sorbent, 76-79% of the composition was accounted
for as Ca(OH)2, CaSCty, and CaCQ$. For a nominal 85 CaO/10 Si(>2/5 Na20
composition, approximately 87% Ca(OH)2 (or 89% CaCOs) is expected. This
left approximately 10% of Ca(OH)2 (or CaCOs) unaccounted for, which
presumably was present as calcium silicates and/or CaO.
-------
TABLE 111-23
Analysis
Wt. % Ca(OH>2
(3)
(2)
Wt.
Wt. % CaC03
(A)
ANALYSIS OF BENCHMARK SORBENT (85 CaO/10 S102/5 Na20)
AS RINGS AFTER VEHICLE TEST AND AS PELLETS
AFTER LABORATORY REACTOR TEST
Total
(1)
Fresh Rings
(Heat Treated 675°C)
58
—
11
69
Used Rings
(20 400 km)
Inlet Sample
2
17
36
55
Used Rings
(20 400 km)
Exit Sample
2
12
57
71
Fresh
3.2-mm Pellets
(Heat Treated 315°C)
53
0
23
76
Used
3.2-mm Pellets
(After 6.9 hrs.
in Reactor Test)
2
72
79
(1) The dimensions of the rings are: outside diameter = 15.9 mm, Inside Diameter =5.6 mm, height
(2) Determined by thermal gravimetric analysis from weight loss between 450 and 650°C.
(3) Gravimetric analysis of total S04 expressed as CaS04.
(4) Analysis of C02 by tltratlon expressed as CaC03.
7.9 mm.
-------
- 47 -
III.3.9 Other Sorbents
III.3.9.1 Carbon Pellets
Three samples of specially prepared carbon pellets obtained
from Calgon Corporation were evaluated in a laboratory screening test.
The porosity of these samples were: Sample A, 0.1 ml/g; Samples B and
C 0.25 ml/g. Sample C was a modification of Sample B with a reduction
in ash content. The pellets were 4.19 mm O.D. and ranged in length
from about 4 to 10 mm. The S04= results on Samples A and B are
summarized in Table 111-18. They could trap some S04=, about 65-74%,
with little reaction with S(>2, but overall were not as effective as
the benchmark sorbent. Sample C showed a 70% reduction in S04=, as
shown in Table 111-25, during a ten hour laboratory test which was run
at 370°C and a space velocity of 100,000 v/v/h. These carbon pellets
also showed very little S02 sorption. The synthetic exhaust gas con-
tained 12% each of C02 and H20, 20 ppm S02, 5 ppm 803, 3% 02, and the
balance N2«
The carbon pellets, Sample C, were further evaluated to deter-
mine their practical application. One major potential problem with the
use of carbon as a sorbent-reductant is its capacity for oxidation. The
burn-off rate on the specially prepared Sample C was measured at 316°C
(600°F) and 371°C (700°F). The feed gas contained 12% C02 and H20,
3% 02, and the balance nitrogen. A summary of the burn-off rates follows:
Table 111-24
Carbon Pellets, Sample C
(Space Velocity:100,000 v/v/h)
Temperature^ °C Burn-Off Rate, %/hO-)
316 0.006
371 0.05
,,, _ Wt. loss x 100
(1) Rate -
Avg. wt. x time
The oxidation rate was high at 371°C (700°F) and may be unacceptable for
long service life. The rate of carbon oxidation for a viable system should
probably be less than 0.01%/h. No further laboratory work was carried out
because of the high burn-off rate at the realistic vehicle operating tempera-
ture of 371°C.
-------
- 48 -
III.3.9.2 Magnesium Hydroxide
A sample of magnesium hydroxide pellets, R-987, specially
prepared by Merck and Company, showed lower 804* removal rate than the
benchmark sorbent (Table 111-25). In addition, Mg(OH)2 pellets showed
about 16% loss in weight, possibly due to exfoliation of MgS(>4 from the
surface. The 804* trapping efficiency averaged 75% for the ten hour
test, however 802 sorption was 30%. No further work was conducted
with this material.
III.3.9.3 Ca(OH)2 and Colloidal Silica
A single wet mixture of Ca(OH)2 was made using Ludex colloidal
silica. To a wet paste of Ca(OH)2 with water was added sufficient Ludex
HS 40 colloidal suspension to make up 15% by weight of Si02- The re-
sulting mix was dried, granulated and pelleted in the usual manner.
Physical properties and laboratory activity results are shown in Table
111-18. Sulfate trapping efficiency was poor, and no further work was
done with this approach.
III.A Vehicle Durability Testing of 804° Trap Sorbents
In the work described in Section III.3, the aim was to identify
at least two compositions which could be fabricated into strong and durable
sorbents. These sorbents were to provide good S04= removal efficiency and
have a minimum effect on back pressure through the trap. In the present
section, tests of the most promising sorbents are described, in vehicle
mounted traps for 15 000 km. The sorbent compositions used were the
following:
• 85% CaO/10% Si02/5% Na20 rings (12.7 mm O.D.)
• 85% CaO/10% Si02/5% Na20 pellets (4.76 mm O.D.)
• 85% Ca(OH)2/15% Si02 (Dicalite) pellets (4.76 mm O.D.)
III.4.1 Summary of Results
Three vehicle durability runs were made, using oxidation catalyst-
equipped vehicles with trap canisters containing the 804"* trap sorbent
candidates. The benchmark material (CaO/S102/Na20) was used in the form
of rings and also pellets, and a sodium-free calcium sorbent (Ca(OH)2/Si02)
was evaluated in the form of pellets.
None of the durability tests were extended beyond 15 000 km
because of the significant loss in 804° removal efficiency of the sorbents
with kilometer accumulation. The sorbents showed above 90% 804" trapping,
and also high 802 sorption at 0 km. As the test progressed, the 802
sorption decreased, but unfortunately, 804" removal also decreased, to
-------
TABLE 111-25
EVALUATION OF SORBENTS IN THE LABORATORY SCREENING TEST
Laboratory Reactor Screening Test
100.000 v/v/h. 370°C
Composition
Final Heat
Treating
Conditions
Density
g/cm3
Crush Strength, kg
After Reactor
Test
Carbon (Calgon)
Sample C
Initial
13.5
11.4
Mg(OH>2 (Merck and Co.
Sample R-987
Inc.)
9.4
4.1
CaO/SiO/Na20 (85/15/5)
(NB 4798-1A) (3.2 mm OD pellets)
16 hours
@ 315°C
1.308
8.6
13.0
Total
Exposure
To Feed
Gas, Hours
1.0
4.0
5.3
6.3
9.1
1.4
2.3
4.3
5.7
7.9
9.9
1.4
2.6
4.6
6.0
7.6
9.6
Trapping
Efficiency, %
61
76 .
63
75
74
89
79
74
76
64
66
<
86
89
89
90
87
87
S02
Trapping
Efficiency, %
0
5
8
11
11
43
32
30
26
24
24
52
35
37
26
29
29
49-
VO
-------
- 50 -
below 90%. The benchmark pellets and the sodium-free sorbent showed
essentially equivalent trapping efficiency after accumulating 15 000 km.
these two preparations were more effective than the benchmark rings.
The back pressure through the S04= traps during the durability
runs remained negligible. Physical inspection of the sorbents at the
end of the test showed no deterioration and no build-up of fines in the
trap. The fabricated sorbents were hard and strong and withstood
vehicular testing.
III.4.2 Exper imental Procedures:
111.4.2*1 Vehicle Preparation
Vehicle tests were conducted using two cars with 351 C1D V-3
engines; a 1973 and a 1974 model. Both vehicles were equipped by us with
two Engelhard PTX-IIB® Pt-Pd monolith oxidation catalysts, one on each
bank of the engine in the post-manifold position. The vehicles were
also each equipped with an air pump. The traps consisted of the con-
ventional 4.25 litre canisters normally used by General Motors for pellet
oxidation catalysts. The traps containing the experimental sorbent
particles were placed in the exhaust system under the rear seat, and
the standard mufflers were removed to accommodate the S04= traps.
III.4.2.2 Test Fuel
All the kilometer accumulation and testing done in this program
were conducted using a 0.032 wt % sulfur content gasoline. Several
typical inspections on the test fuel are presented in Table 111-26*
III.4.2.3 Test Procedure
All the kilometer accumulations were carried out on automatic
mileage accumulation dynamometers using the modified AHA durability cycle.
Periodically, the vehicles were emission tested using the Exxon Research
exhaust particulate sampler described in Appendix C. Each sorbent was
evaluated for 15 000 km on a vehicle including emission testing at the
following points: 0, 5 000, arid 15 000 kilometers. Initially, measure-
ments were obtained at 0 km with and without the 804° trap installed.
The test sequence used for each emission test was the following: FTP,
BET, SET, FET, SET, SET, and 96 km/h (2 hours). Measurements were made
of total sulfates, as well as gaseous emissions of CO, HC, and NOX.
Also, pressure drops across the trap at idle and at 64 km/h cruise
mode were obtained. The results for each of the durability runs are
recorded in Appendix D, Tables D-l through D-3.
-------
- 51 -
TABLE 111-26
TEST FUEL BLEND
RVP, kPa 88.6 (12.9 psl)
% Evap.
@ °C % Off
40 5
55 20
98 50
171 90
Breakdown Time, minutes 960
Gravity, g/cc 0.7599
Sulfur, ppm 320
Lead, g/1 <0.002
FIA. Vol %
Aromatic 25.3
Olefin 10.4
Saturates 64.3
Research Octane 93
Motor Octane 84
-------
- 52 -
III.4.3 CaO/Si09/Na20 (Benchmark) Rings
In the two vehicle durability tests previously performed with 85%
CaO/10% Si02/5% Na20 sorbent, two particle geometries were employed. The
first test used 3.2 mm cylindrical pellets. In the second test, cylin-
drical rings of 15.9 mm O.D., 5.6 mm I.D., and 7.9 mm height were used.
Pressure drop across the trap bed was a major problem with the pellets,
which, after 20 400 km, was 20 kPa at 64 km/h cruise. At similar kilo-
meters, using the rings, the pressure drop was only 4.2 kPa. Although
pressure drop with the rings was lower, S04= sorption efficiency was
only 59% after 11 000 km compared to more than 95% for the pelleted
material. These results are presented in detail in the Final Report
for Phase I of this contract(D.
In considering another sorbent geometry, our targets were a
better combination of low pressure drop and high S04= sorption efficiency,
especially after extended operation. The pressure drop in a fixed bed
can be decreased by increasing the void fraction in the bed, and also by
increasing effective particle size. The sorption efficiency can be
raised by increasing the total surface area of the rings in the trap
bed. In Table 111-27 are listed calculated void fractions and surface
areas for several geometries of trap particles, assuming 0.35 as the
void space for packed beds of right cylinders.
In the case of the rings previously used (geometry D, Table
111-27), the void space was increased by 23%, relative to the 3.2 mm
pellets. However, this was at the expense of the total surface area,
which for the rings was only 34% that of the pellets. These effects
probably account for the reduced pressure drop with the rings, and the
loss in trapping efficiency. It was decided to fabricate rings of
geometry E, using the benchmark sorbent, for vehicle testing, since the
total surface area for E is greater than for rings D or F and the void
fraction is also high.
III.4.3.1 Experimental Results
A Komage T-5 single station tablet press was used to fabricate
the rings from powder prepared as described in Reference (1), p. 48.
Because of greater void volume and more surface area, the new geometry
was expected to afford less pressure drop and greater S04= trapping
efficiency in comparison to the rings previously tested. After heat
treating at 315°C, the new rings had an average crush strength of
3.2 kg and a density of 1.263 g/cm3. Packing characteristics were deter-
mined using a graduated cylinder having an inside diameter of 14.4 cm.
A 3.8 litre packed bed weighed 2.12 kg, corresponding to a void fraction
of 0.56. This experimental void fraction is close to the predicted value
of 0.51.
-------
TABLE 111-27
Outside Diameter, mm
Inside Diameter, mm
Height, mm
Void Fraction in Packed Bed*(6)
Surface Area Per Particle, mm
Volume per particle**
-1
Total Surface Area, mm
Total bed volume***
-1
SURFACE AREA AND VOID SPACE IN
PACKED BEDS OF PELLETS AND RINGS
Cylindrical Pellets
A
3.20
3.2
0.35
1.89
_B
4.76
4.76
0.35
1.26
C
6.35
6.35
0.35
0.94
Cylindrical
D
15.9
5.6
7.9
0.43
0.64
E
12.7
6.4
6.4
0.51
0.94
Rings
F
19.0
9.5
6.4
0.51
0.74
1.23
0.82
0.61
0.42
0.61
0.48
«•« (l)
* Void Fraction = 0.35 +
The first term, 0.35, is the void space found for right cylinders, having height = outside diameter
** Surface Area per particle
Volume per particle Height (OD - ID)
*** Total Surface Area Q g5 /Surface Area per particle
Total Bed Volume
Volume per particle
-------
- 54 -
A standard 4.26 litre GM oxidation catalyst reactor was charged
with 2.21 kg of the benchmark rings. The trap was installed on a 1974
car, equipped with V-8 engine, air pump, and two PTX-IIB® oxidation
catalysts (the test vehicle described in Section III. 4. 2.1).
The results obtained with this vehicle are shown in Appendix B,
Table B-l. Prior to installing the trap, S04= emissions were obtained.
All testing and kilometer accumulation was done using the 320 ppm sulfur
test fuel. The baseline emission tests at 0 km, with and without the 864 =
traps installed, showed that the 804° trapping efficiency for these rings
was less than obtained with the rings previously tested (O.D. 15.9 mm).
Since the new sorbent geometry provided greater geometric surface area,
it was surprising that the sorption efficiency was less. However, the
fabrication of stronger rings with controlled sodium content may have
resulted in less efficient sorbent. Trapping efficiency at 0 km was
51.7% during the EPA Sulfate Emission Test cycle (SET). The back pressure
through the trap was negligible under idle and 64 km/h cruise (Appendix D,
Table D-4) . The pressure drop through the trap was only 0.5 kPa.
It was decided to continue this test, since it was of interest
to know how S04= trapping efficiency is maintained as a function of test
kilometers. The test vehicle accumulated kilometers using the modified
AMA durability cycle on the automatic mileage accumulation dynamometer.
After accumulating 6 400 km, the S04= removal efficiency had significantly
decreased to 26.7% under the SET. The back pressure through the trap
continued low at 0.5 kPa at idle and 0.9 kPa at 64 km/h. The test
vehicle accumulated an additional 10 000 km to obtain an additional check
on the performance of the benchmark rings. The test car was evaluated at
16 000 km with and without the trap installed. The 804° trapping efficiency
was higher and similar to that observed at 0 km under SET, but showed no
804= removal under the FTP testing mode. The S04= measurements with and
without the trap on the vehicle at 16 000 km are summarized as follows:
Table 111-28
1974 Vehicle (351 CID, PTX-IIB)
504° Emissions. SET Cycle @ 16 OOP km
. g/km
With trap installed 0.0051
Without trap 0.0105
% Efficiency* 51.4
S04- tr.ppln, efficiency - (1 -
„
-------
- 55 -
The pressure drop through the trap remained low; 0.5 kPa
idle and 1.0 kPa at 64 km/h. The 16 000 km results showed that the
12.7 mm O.D. rings provided better trapping durability than the rings
(15.9 mm O.D.) previously tested. The new sorbent provided a better
geometric surface and harder rings. However, the rings were not as
efficient as the previously tested benchmark pellets, and the vehicle
testing was terminated after reviewing the results with the EPA.
III.4.4 CaO/Si02/Na20 (Benchmark) Pellets
A large batch of CaO/S102/Na20 (85/10/5) was prepared and
fabricated into 4.76 mm O.D. pellets for vehicular evaluation. The
powder was different than the original benchmark material described in
Reference 1, being prepared with better control of Na£0 content, and
with improved crush strength, using the techniques described in Sections
III.3.2.1 and III.3.4.1. It was our intention that these pellets would
provide better control of back pressure than did the originally tested
3.2 mm pellets; in addition, they should provide better S04= sorption than
did the cylindrical rings of 12.7 mm O.D. Also, the 4.76 mm pellets have
a higher surface area to volume ratio, 1.26 mm~l, versus 0.94 mm~l for the
rings (Table 111-27).
III.4.4.1 Experimental Results
A standard 4.26 litre GM oxidation catalyst canister was charged
with the benchmark pellets. The trap was installed on a 1974 car, equipped
with a V-8 engine, air pump, and two PTX-IIB\£)oxidation catalysts. The
vehicle was set to manufacturer's specifications, and no attempt was made
to meet any particular emission standards. The NOX emission levels are
higher than the standards and was possibly due to a malfunctioning E6R
valve. The results obtained with the vehicle equipped with the 4.76 mm
benchmark pellets are reported in Appendix D, Table D-2.
Prior to kilometer accumulation, emission levels were measured
at 0 km, with and without the trap installed. The S04= trapping efficiency
under the SET, FET, and 96 km/h testing modes was greater than 95%, while
under FTP conditions, only 34%. For example, under the SET testing mode,
the average S04= emission was 8.8 rag/km without the trap and 0.4 mg/km
with the benchmark sorbent in the trap. Similarly, under 96 km/h cruise,
804= level was 25.9 mg/km without the trap and 0.4 mg/km with the trap.
The back pressure through the trap was negligible. After 5 000 km, the
cyclic 804*" sorption efficiency had fallen slightly.
The test vehicle accumulated an additional 10 000 km, and it
was again tested under the different testing modes. In addition, baseline
emissions were obtained on the vehicle with the S04= trap removed. At
15 000 km, the results showed a significant decrease in the sorbent's
efficiency. In Table 111-29, the S04= emission levels and trapping
efficiencies are summarized.
-------
- 56 -
TABLE 111-29
VEHICULAR DURABILITY TESTING OF 504° SORBENT TRAPS
CaO/Si02/Na20 Pellets (4.76 mm P.P.)
S0&= Emissions, me/km
FTP
SET
FET
96 km/h
0 km
4.1
8.8
10.6
25.9
No Trap
15 000 km
3.0
2.1
5.6
18.7
0 km
2.7
0.4
0.3
0.4
With Trap
5 000 km
2.9
0.8
0.8
0.5
15 000 km
4.1
1.8
1.5
3.0
Trapping Efficiency. %
FTP
SET
FET
96 km/h
0 km*
34
95.5
97.2
98.5
5 000 km*
29.0
90.9
92.5
98.1
15 000 km*
0
79.5
85.8
88.4
15 000 km**
0
14.3
73.2
83.9
* S04- trapping efficacy - (1 - gff gg gg , „ _ )
« S04= trapping efficiency = (1 - ff g* gg , „ „„„
-------
- 57 -
The S04= emission levels without the trap installed showed a
decrease at 15 000 km from those obtained at 0 km, indicating some loss
in catalyst activity during the evaluation period. However, the 804=
emission results reconfirm that the CaO/S102/Na20 pellets showed a
significant loss in 804= removal efficiency.
During the 15 000 km test, the back pressure through the trap
was negligible (Appendix D, Table D-4), and inspection of the pellets and
the trap at the end of the test showed no accumulation of fines. This
suggests that the pellets had improved crush strength, compared to those
described in Reference 1. However, because of increased hardness, the
804° trapping, as measured under the different testing modes, was poorer
than observed with the originally tested benchmark pellets. After re-
viewing the results with the EPA, it was agreed that no further testing
of the benchmark sorbent was warranted.
III.4.5 Ca(OH)2/SiO? (Dicalite) Pellets
A large batch of Ca(OH)2/Si02 (Dicalite) (85/15) pellets was
prepared using the wet mixture preparation described in Section
and fabricated into 4.76 mm O.D. pellets. This is a sodium free prepa-
ration, which is intended to minimize sorbent pick-up of 802 •
III.4.5.1 Experimental Results
A GM oxidation catalyst canister was charged with the pellets
and installed on a 1973 V-8 car, equipped with an air pump and two
PTX-IIB catalysts in the post manifold position. The detailed emission
results obtained during the 15 000 km durability test are reported in
Appendix D, Table D. A summary of the S04= emissions and trapping
efficiencies Is presented in Table 111-30.
Sulfate emission levels were measured at 0 km, with and without
the S04= trap Installed. The S04= emission levels without the trap were
higher than S04= levels observed with the 1974 car used to evaluate the
benchmark sorbent. The differences in S04= activity between the two
catalyst pairs is partially due to differences in catalyst activity. The
PTX-IIB on the 1973 car had accumulated approximately 20 000 km, while
the PTX-IIB on the 1974 car had been aged for 50 000 km.
The 804= removal at 0 km was equivalent to the benchmark sorbent,
but then decreased significantly. The S04= removal from 5 000 km to the
end of the test remained at about the same level, with 50-60% 804° removal
under the testing modes, except FTP, where trapping was ineffective. In
fact, the operating conditions of the car apparently changed, so=that the
S04= production over the catalysts during the FTP, and hence S04= emissions,
increased from the 0 km no trap test. This made a quantitative measure
of trap efficiency at 5 000 and 15 000 km impossible for the FTP. However,
the significant decrease in activity during the other test modes confirmed
the decision to cease testing this sorbent.
-------
- 58 -
TABLE 111-30
VEHICULAR DURABILITY TESTING OF S04 SORBENT TRAPS
Ca(OH)£/Dicalite
No Trap
0 km
FTP 4.5
SET (4 tests) 38.4
FET 35.4
96 km/h (2 hours) 51.7
S0*=
0 km
FTP 66.6
SET (4 tests) 96.4
FET 97.7
96 km/h (2 hours) 59.0
(4.76 mm O.D.)
S04= Emissions, mg/km
With Trap
0 km 5 000 km 15
1.5 10.8
1.1 16.4
0.8 12.3
5.7 20.6
Trapping Efficiency, .%*
5 000 km 15 000 km
57.3 58.4
65.3 62.1
60.2 . 62.5
000 km
12.7
16.0
13.4
19.4
. ... = . ,,. . ,- S04~ with trap
* S04 trapping efficiency = (1 - SQ^ without trap @ 0 km
-------
- 59 -
The back pressure through the trap at idle was 0 kPa, and at
64 km/h, 0.75 kPa. Physical inspection of the pellets showed no visible
deterioration. The sodium free sorbent formed hard pellets and withstood
vehicular testing. Unfortunately, the sorbent was not effective. The
removal of sodium from the sorbent composition adversely affected its
sulfate pick-up properties.
III.4.6 Conclusions from Vehicle Durability Evaluation
None of these durability tests were extended beyond 15 000 km
because of the significant loss in SC>4= removal efficiency of the sorbents.
Both the pellet sorbents showed about 90% or better S04= removal at 0 km
for the SET, but the sorbents also showed high S02 removal. As the tests
progressed, S02 removal dropped rapidly. The benchmark pellets and the
sodium free pellets showed overall better SQ^** trapping than the benchmark
rings, probably due to differences in the total surface area of the pellets
versus the ring geometry. The three sorbents at 15 000 km were not effective
in removing S04= under the FTP testing mode, but showed significant differences
under the other testing modes.
The back pressure through the S04= traps was negligible during
the three durability tests. Physical inspection of the rings and pellets
at 15 000 km showed no visible deterioration and no build-up of fines in
the trap. The fabricated sorbents formed strong and hard rings and pellets
which withstood vehicular testing. Unfortunately, these stronger and harder
sorbents were not effective on extended testing.
-------
- 60 -
IV. Modification II - Effect of Noble Metal
Composition on Catalyst Activity
IV.1 Summary of Results
Three pairs of monolith oxidation catalysts were prepared by
Matthey-Bishop, Inc., differing primarily in noble metal composition.
The noble metal compositions were Pt-Pd, Pt-Rh, and Pt. These catalysts
were evaluated sequentially on the same car for 12 800 km each. The
Pt-Rh catalyst was found to give the lowest sulfate emissions under
every testing mode. The Pt-Pd and Pt catalysts gave higher sulfate
emissions and were roughly comparable to each other. For example,
the average sulfate emission levels measured using the EPA Sulfate
Emission Test were the following, with the test fuel containing 320
ppm sulfur.
Catalyst Average 504°, g/km
Pt-Rh 0.010
Pt-Pd 0.025
Pt 0.030
During the program, the other gaseous emissions were similar for all
three catalysts at these low mileages.
The results confirm a growing body of data that Pt-Rh containing
catalysts are lower sulfate emitters than other commercial catalysts, and
the noble metal composition itself, in addition to or instead of other
factors, such as support type, is playing a major role. The data offer
strong support for the possibility of developing selective catalysts
which have high activity for CO and HC oxidation and low activity for
S02 oxidation.
IV.2 Introduction
This program, Modification II, is an extension to Contract No.
68-03-0497, "An Assessment of Sulfate Emission Control Technology". The
purpose of this work is to help determine whether or not noble metal
composition changes are the major factors causing the observed differences
in sulfate emissions from various commercial catalysts.
Since the oxidation of S02 in automotive exhaust is a catalytic
process, it is to be expected that different catalysts will show different
emission rates. One of the most promising approaches to 804™ control Is
the use of catalysts which are selectively more active for CO and HC
oxidation than they are for S02 oxidation. Beltzer, et al.,(2) reported
-------
- 61 -
that Pt-Rh monoliths showed significantly less SC-2 oxidation than
nominally similar Pt-Pd monoliths. This finding was subsequently
verified by Bradow and MoranO). Some of the data obtained by Holt,
et al.,(7) indicated that Pt catalysts might have higher S04= emission
rates than Pt-Pd catalysts. In a recently completed API sponsored
program(4), similar trends among catalyst systems on 804° emissions
were observed.
The above studies were confounded by the fact that many
other catalyst properties, besides noble metal composition, were differ-
ent. Thus, catalysts from different manufacturers were used, and it
is probable that, in addition to noble metal composition, such para-
meters as the amount and composition of washcoat, surface area, and
pore volume, noble metal loading, and cell geometry, varied between
catalysts. To reduce ambiguity in this program, one manufacturer,
Matthey-Bishop, Inc., prepared the three catalyst pairs, holding all
other properties, except noble metal composition, constant as far as
possible.
The program was originally divided into two tasks. Task 1
involved the testing of three fresh catalysts, identical in all respects
as far as possible, except for their noble metal composition. Task 2
required the testing of two aged catalysts (40 000 km), which were also
as similar as possible, except for noble metal composition. However,
the latter pair of catalysts did not become available, as had been
expected, so this part of the program was not done.
The results from Task 1 evaluated whether the noble metal
composition changes are the major factors affecting the observed dif-
ferences in S04= emissions. Task 2 would have provided data on whether
such differences are maintained over the useful life of the catalysts.
IV.3 Experimental
IV.3.1 Monolith Oxidation Catalysts
Three pairs of monolithic oxidation catalysts were prepared
by Matthey-Bishop, Inc. for use in this program. They were as identical
as possible, except for their noble metal composition, which consisted
of Pt, Pt-Pd, and Pt-Rh. A description and characterization of the
catalysts are summarized in Table IV-1. The catalysts supplied by
Matthey-Bishop were made from Corning ceramic monoliths, extruded
honeycomb structures containing approximately 32 channels per car of
cross sectional area. The catalysts contained an alumina washcoat-and
a noble metal loading of approximately 1.6 x 10"3 g/cm3. The Pt
catalyst's H.S.A. coating surface area was about 2 times higher than
the other two prepared catalysts, but the washcoat loading for all
three catalyst pairs was equivalent.
-------
TABLE IV-1
CATALYST CHARACTERIZATION*
Pt-Pd** Pt-Rh Pt
AEW 2/62/50/M AEW 2/3C/40/M AEW 2/12C/40/M
I II
H.S.A. Coating Surface Area (m2/cm3) 11.1 10.2 10.7 22.5
Catalyst Loading (g/cm3) 1.8 x 10~3 1.6 x 10~3 1.6 x 10~3 1.5 x 10~3
9 3
Catalyst Surface Area (nr/cm) 0.12 0.11 0.16 0.11
Washcoat Loading (g/cm3) 0.10 0.10 0.11 0.12
Substrate 32 cell/cm2 S
i
Size -------- 12.5 cm diameter x 7.5 cm length- -------
* Inspections provided by Matthey-Blshop, Inc.
** An additional pair of Pt-Pd catalysts was prepared because one of the catalysts
failed during mileage accumulation.
-------
- 63 -
IV.3.2 Test Vehicle
The monolith catalysts were evaluated on a 1975 Ford LTD
with a 5.85 litre V-8 engine. The car was equipped with an air pump.
This vehicle exhaust system had been modified to accept two monolith
catalysts in the post manifold position. The test vehicle had been
used previously in Contract No. 68-03-0497 and described in the previous
final report of Phase I, Section V.2.l(1). Before starting the emission
test sequence with each set of catalysts, the vehicle was tuned-up and
set to manufacturer's specifications.
IV.3.3 Test Fuel
The test fuel contained 0.032 wt % sulfur. The fuel was
blended from a base fuel by the addition of equal amounts of thiophene
and di-t-butyl disulfide. Typical inspections on the test fuel are
summarized in Table IV-2.
IV.3.4 Test Sequence
The catalyst pairs were tested sequentially on the same vehicle
using the same test equipment in order to maintain uniformity in aging
and test conditions. S04= and other gaseous emissions were measured
using the following sequence:
1975 Federal Test Procedure (FTP)
Federal Economy Test (FET)
EPA Sulfate Emission Test (SET)
Idle (20 minutes)
64 km/h one hour cruise
96 km/h one hour cruise
After the initial emission test sequence, the test vehicle was placed on
the automatic mileage accumulation dynamometer (MAD), and accumulation was
carried out using the ERE turnpike cycle having an average speed of
91 km/h. At intervals of 3 200, 6 400, 9 600, and 12 800 km, the vehicle
was put through the above emission testing sequence. The preconditioning
between FTP tests was 3 200 km of MAD accumulation, while the precondition-
ing for the other portions of the test sequence was the preceding SO,0
emission test.
Standard instrumentation was used for CO, HC, and NOX measure-
ments. S04= emissions were measured using ERE's CVS exhaust partlculate
sampler^) and S02 emissions were measured using an H202 bubbler technique.
Measurements for 804= were made for the entire FTP, FET, SET, and idle
tests, but were broken into half hour segments for the two cruise modes.
A more detailed description can be found in the final report on Phase I
of Contract No. 68-03-0497, Section IV.1.3(D.
-------
- 64 -
TABLE IV-2
TEST FUEL BLEND
RVP, kPa 68.8 (10.0 psi)
% Evap. @ °C
45 5
64 20
112 50
179 90
Breakdown tine, minutes 960
Gravity, g/cc 0.7596
Sulfur, ppm 320
Lead, g/1 <0.002
FIA, Vol. %
Aromatic 35.5
Olefin 6.9
Saturates 56.6
-------
- 65 -
IV. 4 Results
This program provided data on the effect of noble metal
composition changes on automotive sulfate emissions, at least with
relatively fresh systems. The total kilometer accumulation was too
short to see any appreciable effect of catalyst aging.
The test sequence run with the Pt-Pd catalysts was repeated
with a new pair of Pt-Pd catalysts because inspection of the catalysts
at the end of the first 12 800 km test showed that one catalyst had
partially melted. The catalyst's failure was probably due to an
overtemperature while accumulating kilometers on the automatic mileage
accumulation dynamometer. As the test progressed, an increase in CO
and HC emissions were observed. Initially, we thought the Increase
in emissions was caused by a malfunctioning carburetor, however, the
partial loss of one converter would account for the increase. The
catalyst failure was probably due to ignition system failure. Apparently,
the dump value which is operated by a thermal switch was not able to
protect the catalysts from overtemperature. A new ignition system was
installed in the test vehicle before installing the next pair of catalysts.
The individual S04= emissions and other gaseous emissions on
each catalyst pair are included in Appendix E (Tables E-l to E-3) .
The fuel consumption and exhaust temperatures are reported in Appendices
F and G, respectively. All the data on the initial runs on the Pt-Pd
catalyst that burned up are reported in Appendix H.
IV. 4.1 Sulfate Emissions
Analysis of the SOf test results on the three pairs of oxid-
ation catalysts showed that the Pt-Rh catalysts provided the lowest
S04= emission levels at each test interval and under each testing mode.
These results are graphically illustrated in Figures IV-1 to IV-5.
It is interesting to note that at 0 km, the Pt catalysts showed lower
S04= levels than the Pt-Pd, but as the test progressed, the Pt catalyst
directionally showed higher S0^= levels. This low mileage behavior of
the Pt catalyst may have been caused by its higher washcoat surface
area, which could provide greater initial storage capacity for SO^**.
No attempt was made at drawing the best line through all the data points,
since each data point is one test/mode/test Interval. In general, SO,"
emission levels were the highest using the steady state cruise testing
modes. The S0^= emissions at 0 and 12 800 km under each testing mode
are summarized in Table IV- 3. Individual SO^B emission results at each
test interval are reported in Appendix E.
IV. 4. 2 Other Gaseous Emissions
All three catalysts gave similar CO and HC emission levels
during the 12 800 km test. Individual gaseous emission results are
-------
TABLE IV-3
SULFATE EMISSIONS AT START AND FINISH OF MILEAGE ACCUMULATION
Test Sequence
FTP
FET
SET
*
Idle (20 min.)
**
64 km/h (1 hr) I
II
**
96 km/h (1 hr) I
SO, Emissions, g/km
Pt-Rh
0.004
0.005
0.004
- - -
0.008
0.021
0.035
0.027
0 km
Pt-Pd
0.009
0.024
0.028
0.009
0.064
0.057
0.063
0.053
Pt
0.006
0.009
0.010
0.019
0.036
0.043
0.038
0.031
Pt-Rh
0.007
0.020
0.009
0.017
0.024
0.023
0.039
0.022
12 800 km
Pt-Pd
0.014
0.055
0.026
0.018
0.062
0.059
0.062
0.052
Pt
0.021
0.064
0.031
0.018
0.059
0.055
0.065
0.047
SO
**
i " results at idle reported as g/test.
The two cruise modes were broken up into two half-hour segments.
-------
.070
- 67 -
FIGURE IV-1
804= EMISSIONS - 1975 FTP
.060
.050
A Pt (12-C)
O Pt-Pd (6-C)
D Pt-Rh (3-C)
in
zs
o
CO
CO
-a-
O
CO
.040
.030
.020
,010
3 200
6 400 9 600
TEST KILOMETRES
12 800
-------
- 68 -
FIGURE IV-2
804° EMISSIONS - FET CYCLE
.070
.060
.050 .
.040 .
CO
§
M
CO
CO
O
CO
.030 .
.020-
.010
3 200
6 400
TEST KILOMETRES
9 600
12 800
-------
- 69 -
FIGURE IV-3
EMISSIONS - SET CYCLE
.060
.050
A Pt (12-C)
O Pt-Pd (6-C)
D Pt-Rh (3-C)
4
atr
§
M
u
•a-
o
CO
.040
.030
.020 -
3 200
6 400 9 600
TEST KILOMETRES
12 800
-------
- 70 -
.070
§
M
CO
OT
o
OT
.020 -
.010
FIGURE IV-4
804" EMISSIONS - 64 km/h CRUISE
A Pt (12-C)
O Pt-Pd (6-C)
D Pt-Rh (3-C)
3 200
6 400 9 600
TEST KILOMETRES
12 800
-------
- 71 -
FIGURE IV-5
S04= EMISSIONS - 96 km/h CRUISE
.070
V)
a
o
CO
M
O
CO
A Pt (12-C)
O Pt-Pd (6-C)
D Pt-Rh (3-C)
.010 _
3 200
6 AOO
TEST KILOMETRES
9 600
12 800
-------
- 72 -
tabulated In Appendix E. A summary of the average gaseous emissions
measured during the program under FTP, FET, and SET testing modes
follows (Table IV-4):
Table IV-4
Average Gaseous Emissions*
FTP, g/km FET, g/km SET, a/km
Catalyst
Pt-Pd
Pt-Rh
Pt
CO
2.
3.
2.
77
27
28
HC
0.43
0.37
0.33
NO*
1.69
1.68
1.78
CO
0.21
0.23
0.52
HC
0.10
0.05
0.08
NQx
2
2
2
.41
.73
.58
CO
0.67
1.18
0.05
HC
0.10
0.05
0.10
NO*
2.24
2.50
2.43
* The gaseous emissions averaged over the 12 800 km test period.
The Pt-Rh catalyst pair gave higher CO emissions under the FTP
and SET modes, but was low on the FET. Its HC emissions were low over
all three test modes. The Pt catalyst pair was lowest in CO for the FTP
and SET, but highest for the FET. HC emissions were roughly comparable
to the other catalysts for all test modes. No explanation for this
sometimes contradictory results are available.
IV.4.3 Fuel Consumption and Catalyst Outlet Exhaust Temperatures
The fuel consumption and catalyst outlet exhaust gas temperatures
on each test run are tabulated in Appendices F and G. The fuel consumption
was determined in two ways: the direct fuel weight and calculated from
the emission data using the carbon balance method. The agreement In fuel
consumption by the two methods was good. The fuel consumption for each
testing mode remained relatively constant with minor fluctuations between
runs for each catalyst system evaluated. Differences between catalysts
are probably due to the fact that the test car was tuned-up each time
catalysts were changed, resulting in slight variations in engine calibrations.
A summary of calculated fuel economy under the SET modes follows:
Table IV-5
Fuel Consumption, g/km
Pt-Pd
Pt-Rh
Pt
SET
118.6
133.7
126.2
Standard
Deviation
6.2
5.8
2.7
Coefficient of
Variation, %
5.2
4.3
2.1
-------
- 73 -
Thermocouples were placed in the exhaust system at the outlet
of the catalysts to measure exhaust gas temperatures. Some variation
between catalyst pairs was observed, but as in the case of fuel con-
sumption results, was probably due to slight changes in engine cali-
brations following tune-ups. Examples of the exhaust gas temperature
ranges observed under the SET follow:
Table IV-6
Maximum Exhaust Gas Temperatures, °C
SET Cycle
Kilometers Pt-Pd Pt-Rh Pt
0 607 671 654
3 200 613 671 582
6 400 571 671 616
9 600 577 616 517
12 800 604 649 616
IV.4.4 Oxygen Level in Exhaust Gas
The oxygen concentration in the raw exhaust ranged from 5.0-
5.5% under the different testing modes. In this program, the air pump
on the test vehicle was used at all times, so the three catalyst pairs
were exposed to the same 02 levels during SO^8* emission testing. Under
each testing mode, excess 62 was present in the exhaust gases.
IV.4.5 Conclusions
The Pt-Rh catalyst gave the lowest sulfate emissions under
every testing mode at every test interval. The Pt-Pd and Pt catalysts
gave higher S0^° emissions and were comparable to each other. With the
testing sequence used, sulfate emissions were highest under the steady
state cruise conditions.
The results indicated that the noble metal composition is
playing a major role in determining the rate of oxidation of S02 to
803. The program did not provide any data on whether the observed
differences among catalysts on SQf emissions are maintained over the
useful life of the catalysts. The data, however, offer strong support
for the possible development of selective catalysts which have high
activity for CO and HC oxidation and low activity for S02 oxidation.
-------
- 74 -
V. Modification III - Perovskite Catalysts
V.I Summary of Results
In both standard and high temperature vehicle configurations
a palletized perovskite catalyst showed poorer gaseous emission control
than the GM pelletized catalyst. Sulfate emissions were also substantially
higher for the perovskite catalyst in both vehicle configurations.
In tests with the vehicle in the high temperature configuration,
it was found that a monolithic perovskite catalyst also exhibited poorer
gaseous emission control than an Engelhard monolithic system. However,
sulfate emissions from the perovskite monolith were lower than from the
Engelhard monolith.
V.2 Introduction
This program was designed to evaluate the durability, activity,
and sulfate formation activity of DuPont's perovskite Ft oxidation
catalysts when used with unleaded fuel. According to DuPont, these
catalysts possess a high degree of thermal stability. This stability
presumably allows the catalyst to be operated at temperatures of about
700°C, with the vehicle using a leaded fuel, without incurring losses
in catalyst activity due to poisoning by lead.
At the time this program was initiated, DuPont had begun vehicle
testing with leaded fuel. However, they had not completed their program
to optimize vehicle operating conditions for achieving the necessary
high catalyst temperatures without incurring fuel economy penalties.
Consequently, this program was restricted to testing DuPont's catalyst
system in an unleaded environment at temperatures which would not
entail a fuel debit. The objective was to provide information on the
activity and durability of DuPont's catalysts in this type of environment,
and thereby obtain a baseline case against which the effects of lead on
activity and durability could be eventually assessed.
V.3 Experimental Conditions
Emission measurements were made using the following test
sequence:
1975 FTP
2 SET
HET
2 SET
64 km/h cruise (1 hour)
-------
- 75 -
The test vehicle was a 1976 Ford LTD equipped with a 6550 cc
engine and an air pump. Ford Motor Company provided cylinder heads
with port liners to fit this engine. The DuPont catalyst pellets were
packed into a standard GM catalyst canister. The size of the canister,
4.25 litres, precluded close placement of the catalytic reactor in the
post manifold position. Each catalyst, when used, was placed in the
underfloor position. The standard vehicle catalysts, of course, were
removed. The sulfur content of both fuels, the unleaded kilometer
accumulation fuel, and the unleaded indolene used for emission testing,
was 300 ppm.
Kilometer accumulation was carried out on an automatic kilo-
meter accumulation dynamometer using the modified AMA durability driving
schedule. The modified AMA consisted of 11 laps, no wide open throttle
accelerations, and 89 kph maximum speed.
Prior to Initial emission testing, 1 050 km were accumulated
on the test vehicle in consumer service to provide engine break-in and
some combustion chamber deposits.
V.4 Emission Results
V.4.1 Vehicle Engine-Out Emissions
The vehicle was run through an FTP with the catalyst removed
before initiating the kilometer accumulation. This procedure was repeated
part-way through the test program. These steps were taken to provide a
check on vehicle calibrations, to Insure that major changes in carburetion
or spark timing, which could significantly affect vehicle emissions, had
not occurred. The FTP emissions, shown in Table V-l, Indicate that the
only significant change was a reduction in engine-out CO emissions.
Table V-l
Test Vehicle FTP Emissions Without Catalyst
g/km Fuel Economy
Vehicle km CO HC NO* (km/1)
1 230 20.6 2.1 2.3 4.92
14 780 12.6 2.0 2.6 5.07
-------
- 76 -
V.4.2 Emission Results with GM Pelletized Oxidation Catalyst
Table V-2 shows the emission results, starting with the fresh
GM catalyst, up to the point where the catalyst had accumulated 13 300
km on the AMA driving cycle. The data for the four SET's has been averaged
for each test sequence, as little variation in emissions was found over
the four SET's in each test sequence. The 64 km/h cruise data represents
the average of two bag samples and filters taken for two twenty minute
sampling periods.
Gaseous emission values over the FTP are lower for the test
sequence run after 6 640 catalyst km than initial measurements. This
may have been due, at least in part, to a small leak that was discovered
in the diaphragm of the sampling pump in the CVS system after the test
sequence. The leak was repaired at this point. Sulfate measurements
are unaffected by this leak. The 13 300 km results show the expected
upward trend in HC emissions over the FTP and SET cycles, compared to
the fresh catalyst values. The HET comparison is observed by an abnornally
high value in the fresh catalyst value. CO emissions are down at 13 300
km, but this is probably due to the lower engine-out emissions shown in
Table V-l.
The rise in sulfate emissions by 6 640 km is in line with
previous results obtained at Exxon in a study involving a twenty car
fleet test(21), which showed that sulfate emissions peaked in this
region. In the present test, the car was not run for a sufficient
number of kilometers to begin showing the expected downward drift in
sulfate emissions.
V.4.3 Emission Results with Pelleted DuPont Perovskite Catalyst
The perovsklte catalyst tested contained 0.066 wt % Pf, or
approximately 1.8 g Pt per 4.25 litre GM reactor. This is comparable
to the loading of the GM pelletized oxidation catalyst, which is 0.055
wt % noble metal or 1.5 g Pt-Pd per reactor. Full emission test sequences
were run after 3 300 km and 6 700 km. A duplicate FTP was performed at
3 300 km to confirm the gaseous emission results. The results of these
test sequences are shown in Table V-3.
V.4.4 Comparison of Perovskite and GM Pelleted
Catalysts at Normal Operating Temperatures
The perovskite catalyst showed poorer control of HC and CO
emissions than the GM pelletized catalyst. The CO and HC emissions
averaged over the kilometer accumulation for each catalyst are shown
in Table V-4. The data In Table V-4 are the averages of 3 FTP'a for
each catalyst system. CO and HC emissions are higher for the perovskite
catalyst by factors of 2.2 and 1.6, respectively. Overall conversions
are also shown. These have been calculated by averaging the emissions
from the two FTP runs made without the catalyst, which were shown in
Table V-l.
-------
TABLE V-2
EMISSION CHARACTERISTICS, VEHICLE EQUIPPED WITH GM CATALYST IN STOCK CONFIGURATION
GM
Vehicle Catalyst
km km Test
1 270 NONE FTP
SET*
HET
64 km/h Cruise
7 910 6 640 FTP
SET*
HET
64 km/h Cruise
14 560 13 300 FTP
SET*
HET
64 km/h Cruise
Emission Rates
Gaseous
CO
3.4
4.
4.
0.
2.
3.
1.
0.
2.
1.
0.
0.
6
3
04
4
5
2
04
8
2
9
03
of Indicated Exhaust Components
Emissions, g/km
HC NOy.
0.21 2.2
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
0.
20
52
08
20
17
10
03
33
25
25
14
1.
1.
1.
1.
1.
1.
1.
2.
2.
2.
1.
8
5
3
7
7
6
0
2
0
0
3
H2S04**
mg/km
3.4 (2.
1.
1.
1.
9.
3.
7.
4.
6.
5.
5.
8.
1
5
0
4
0
0
8
7
6
6
4
(1.
(1.
(1.
(7.
(3.
(8.
(9.
(4.
(5.
(5.
6)
2)
8)
4)
0)
2)
3)
2)
9)
8)
8)
(10.6)
Fuel
Economy
km/1
4.
6.
7.
9.
4.
6.
7.
8.
4.
6.
6.
8.
87
76
69
25
79
63
50
72 ,
-vl
69 7
64
64
55
* Average of 4 tests.
** Bracketed numbers are % conversion of fuel sulfur to sulfate, based
on emitted sulfate.
-------
TABLE V-3
EMISSION CHARACTERISTICS, VEHICLE EQUIPPED WITH PELLETIZED
DuPONT CATALYST IN STOCK CONFIGURATION
Vehicle
km
18 120
18 330
21 510
DuPont Emission
Catalyst Gaseous
km Test CO
3 330 FTP 6.1
SET* 4.5
HET 0.3
64 km/h Cruise 0.04
3 540 FTP* 10.4
6 720 FTP 5.6
SET* 3.6
HET 0.2
64 km/h Cruise 0.05
Rates of Indicated
Emissions,
HC
0.41
0.48
0.14
0.12
0.40
0.42
0.34
0.13
0.12
g/km
NOx
2.3
2.3
1.8
1.5
2.5
2.7
2.3
1.7
1.3
Exhaust Components
mg/km
3.5 (6.2)
11.9 (12.6)
12.8 (15.5)
13.2 (18.8)
6.6 (4.8)
17.2 (18.0)
21.7 (25.0)
17.8 (22.4)
Fuel
Economy
km/1
4.71
6.73
7.74
9.03
4.71
4.66
7.65
7.36
8.00
00
I
* Average of 4 tests
** Bracketed Numbers are % Conversion, Fuel sulfur to sulfate
* Duplicate FTP
-------
- 79 -
Table V-4
Comparison of GM and Perovskite Oxidation Catalysts
Average FTP
Gaseous Emission Rates Overall %
Pelletized (g/km) Conversion
Catalyst System CO HC C£ HC
GM 3.3 0.25 80 88
Perovskite 7.3 0.41 56 80
The poorer control of the perovskite catalyst is primarily due
to a slower rate of catalyst warm-up. Typical conversions are shown in
Table V-5 for Bag 1 and Bag 2 of the FTP. On the cold start, the
perovskite catalyst converts only 35% CO and 50% HC, compared to 71%
CO conversion and 68% HC conversion for the GM catalyst. Although
warmed-up conversions are slightly lower for the perovskite catalyst,
as evidenced by Bag 2 data, this had little effect on overall vehicle
emissions.
Table V-5
Catalyst Conversions
Cold Start Warmed-Up
Conversion Conversion
Bag 1 (%) Bag 2 (%)
Catalyst CO HC CO HC
GM 71 68 99 95
Perovskite 35 50 94 91
Sulfate emissions were also higher for the perovskite catalyst.
At approximately 6 400 km, the GM catalyst averaged 3.0 mg 804"/km for
the SET, compared to 17.2 mg S04=I/km for the perovskite catalyst. Only
on the FTP did the perovskite catalyst emit less sulfate than the com*-
merclal GM catalyst. During the SET, HET, and cruise, the perovskite
catalyst produced 3-5 times as much sulfate as the GM catalyst.
-------
- 80 -
The operating temperatures of both catalyst systems were
quite similar. This would be expected, since the vehicle A/F ratio
was running lean, approximately 15 to 15.5, so that catalytic con-
version of CO would not be a factor Influencing operating temperatures.
The operating temperatures for both catalyst systems is shown in
Table V-6.
Table V-6
Catalyst Temperatures
Standard Vehicle Configuration
Catalyst
Temperature, *C
Test Average Peak
AMA Durability Cycle 455 550
FTP 425 550
SET 505 540
HET 515
64 km/h cruise 480
V.5 High Temperature Operation
V.5.1 Vehicle Modifications
Cylinder heads with port liners were installed, and insulation
applied to the exhaust manifold and piping to increase the catalyst
operating temperatures. A comparison of the catalyst operating temper-
atures in the standard and high temperature configurations is shown In
Table V-7. As seen in this table, operating temperatures were about
200°C higher during the AMA durability cycle, and about 100°C higher
during the test sequence.
V.5.2 Comparison of GM and Perovskite Catalyst Gaseous
Emissions in the High Temperature Configuration
The gaseous emission data from the higher temperature con-
figuration tests are shown in Table V-8. As in the case of the standard
configuration, emissions with the perovskite catalyst were substantially
higher than those with the GM catalyst, particularly for the FTP. CO
and HC emissions were 17.3 and 0.62 g/km, respectively, for the perovskite
catalyst, and 3.7 and 0.36 g/km for the GM catalyst on the FTP. Emissions
-------
TABLE V-7
CATALYST TEMPERATURES
MODIFIED VEHICLE CONFIGURATION
Catalyst Temperature, °C
Average Peak
High Temperature Standard High Temperature Standard
Test System System System System
AMA Durability Cycle 650 455 740 550
FTP 515 425 550 500
SET 615 505 650 540 o>
!-•
HET 620 515 '
64 km/h cruise 600 480
-------
TABLE V-8
COMPARATIVE GASEOUS EMISSIONS UNDER HIGH TEMPERATURE CONFIGURATION CONDITIONS
Gaseous Emissions
Vehicle
km
22 600
23 580
Catalyst
Catalyst km Test
GM 14 350 FTP
SET*
HET
64 km/hr. cruise
Perovskite 7 780 FTP
*
SET
HET
64 km/hr. cruise
CO
3.7
2.9
0.4
0.03
17.3
3.1
0.5
0.04
g/km
HC
0.36
0.28
0.21
0.09
0.62
0.29
0.26
0.16
NOx
1.3
1.2
1.1
0.9
1.5
1.5
1.3
0.9
Fuel
Economy
km/1
4.69
6.57
7.50
8.45
oo
4.50 M
i
6.52
7.26
8.84
21 760
None
FTP
20.6
2.67
1.4
4.57
Average of four tests.
-------
- 83 -
on the warmed-up cycles were only slightly higher for the perovsklte
catalyst, pointing to catalyst warm-up as the major factor in the poorer
control for the perovskite catalyst. This is confirmed in Table V-9,
which shows the contribution of the cold start, Bag 1, and warmed-up
cycles, Bags 2 and 3, to the overall FTP emissions. A comparison
between the GM and perovskite catalysts shows the higher cold start
emissions with the perovskite catalyst. Since both catalysts see the
same nominal exhaust conditions during the cold start, it appears
that the lite-off temperature of the perovskite catalyst must be higher
than that of the GM catalyst.
V.5.3 Sulfate Emissions from GM Pallatized and Perovskite
Oxidation Catalysts in High Temperature Configuration
As expected from thermodynamic considerations, sulfate emissions
from both the GM and perovskite catalysts were substantially decreased
at the higher catalyst temperatures. Sulfate emissions are shown in
Table V-10 for both the test sequences in the high temperature con-
figuration and in the standard configuration that preceeded the install-
ation of the port liners and Insulation. Sulfate emissions decreased
by factors of approximately 2-4 on the GM and perovskite catalysts due
to increasing the catalyst temperature 100°C. Significant reductions in
sulfate emissions occurred for all test modes. However, the reductions
were higher for the warmed-up test modes, i.e., the SET, HET, and 64
km/h cruise. Sulfate emissions from the perovskite catalyst were con-
siderably higher than the GM catalyst. This is consistent with results
obtained in the standard configuration.
V.5.4 Comparison of Emissions Between Standard
and High Temperature Configurations
It appears that NOX emissions have decreased due to the addition
of port liners. Table V-8 showed the NOX emissions over the test
sequence with port liners and insulation. Compared to the tests without
port liners, Tables V-2 and V-3, NOX emissions have dropped 30-40%. Raw
exhaust traces during testing do not indicate any significant changes in
carburetion which would account for this NOX reduction. It is possible
that, although the engine is calibrated to run at an A/F of 15-16 during
cruise conditions, some cylinders may, at times, receive a rich mixture.
This is particularly true during accelerations where power enrichment
can occur. Under rich conditions, the port liners may act as a reduction
catalyst. NOX reduction with stainless steel reactors has been reported
in the literature. This may account for the reduced NOX emissions with
port liners.
-------
TABLE V-9
AVERAGE FTP GASEOUS EMISSIONS
Gaseous Emissions, g/km
Overall
Configuration
Standard
High
Temperature
Catalyst
None
GM
Perovskite
None
GM
Perovskite
CO
16.6
3.3
7.3
20.6
3.7
17.3
HC
2.07
0.25
0.41
2.67
0.36
0.62
Bag
CO
7.2
2.3
4.9
13.6
2.8
15.9
1
HC
0.44
0.15
0.23
0.71
0.21
0.43
Bag 2
CO
9.4
1.0
2.4
7.0
0.9
1.4
& 3
HC
1.63
0.10
0.18 ,
oo
1.95 *•
0.15
0.19
-------
- 85 -
TABLE V-10
COMPARATIVE SULFATE EMISSION RATES OF THE
GM AND PEROVSKITE CATALYSTS IN
STANDARD AND HIGH TEMPERATURE CONFIGURATIONS
Test
FTP
SET*
HET
64 km/h Cruise
Standard
SO -
(mg/Rm)
6.7
5.6
6.8
8.4
GM Catalyst
Configuration
Conversion
(%)
4.9
5.8
8.0
10.6
High Temperature
Configuration
S0,= Conversion
(mg/Rm) (%)
4.6 3.4
1.8 1.9
1.6 1.9
1.9 2.5
Perovsklte Catalyst
Test
FTP
SET*
HET
64 km/h Cruise
Standard
SO =
(mg/kml
6.6
17.2
21.7
17.8
Configuration
Conversion
(%)
4.8
18.0
25.0
22.4
High Temperature
Configuration
S0,= Conversion
(mg/Rm) (%)
3.1 2.2
7.4 7.6
8.5 9.7
8.8 12.2
*Average 4 tests
-------
- 86 -
A comparison of the vehicle FTP emissions, both with and
without catalysts, in the standard and high temperature configurations
is shown in Table V-9. HC emissions have been Increased significantly
by the addition of port liners and insulation. This applies both with
and without catalysts. The increase in HC emissions is probably too
great to be attributed to test variability. CO emissions increased
significantly only for the perovskite catalyst. Although the CO
emissions are higher with no catalyst and the GM catalyst in the
high temperature configuration, these differences are within test
variability. This may indicate that the perovskite catalyst lost
activity during kilometer accumulation in the high temperature con-
figuration. This loss in activity appears to be only during the cold
start, since Table V-9 indicates that warmed-up CO emissions did not
increase for the perovskite catalyst.
The increase in HC emissions in the high temperature config-
uration appears to be due to a slower rate of engine warm-up. Table
V-9 shows that HC emissions are considerably higher only during the
cold start. A slower rate of engine warm-up would increase HC emissions
and is consistent with the addition of port liners. The port liners
reduce the heat transfer from the exhaust to the engine coolant in the
area of the exhaust port. This reduction in heat transfer results in
the engine requiring more time to reach normal operating temperatures.
HC emissions are known to be increased at operating temperatures
significantly below normal. Even though catalyst warm-up should be
increased by the addition of the port liners, the increased warm-up
rate did not compensate for the increased HC emissions in the high
temperature configuration. An alternative explanation for the increased
HC in the high temperature configuration could be a longer choke action.
However, raw exhaust traces indicate that the choke action was not
appreciably affected by the addition of the port liners and insulation.
It should be pointed out that the increased HC found with port
liners in this case does not appear to be universal. Other workers have
reported reductions in HC emissions with port liners in thermal reactor
systems. Possibly the lack of a thermal reactor in our vehicle accounts
for the difference in results.
V.6 Comparison of Perovskite Monolithic
Catalysts with Engelhard PTX-IIB
V.6.1 Test Procedure
The test vehicle used for the pelletized catalysts was also
used for the test sequence with the monolithic catalysts. All tests
with the monolithic catalysts were with the vehicle in its high temp-
erature configuration. To establish a basis for comparison, a fresh
-------
- 87 -
Engelhard monolithic catalyst was tested, along with the perovskite
monolith. In both cases, the catalysts were not conditioned, but were
tested at zero kilometers.
V.6.2 Emission Results with Monolith Catalysts
The DuPont monolith provided poorer control of CO and HC than
the Engelhard catalyst. CO and HC FTP emissions for the DuPont catalyst
were 6.83 and 0.61 g/km, as opposed to 1.08 and 0.15 g/km, respectively,
for the Engelhard monolith. On the SET's, the DuPont catalyst averaged
4.15 g/km CO and 0.5 g/km HC, the Engelhard catalyst 0.7 g/km CO and
0.16 g/km HC. The emissions data for these two catalyst systems are
given in Tables V-ll and V-12.
Table V-ll
Emission Characteristics, Vehicle Equipped with
Monolithic Perovskite Catalyst in High Temperature Configuration
Emission Rate, g/km
Test
6
4
4
1
4
3
0
CO
.83
.94
.27
.61
.11
.14
.09
HC
0
0
0
0
0
0
0
.61
.55
.44
.43
.48
.56
.22
JNOx
3.
3.
3.
2.
3.
3.
1.
31
05
02
63
03
15
73
0
0
0
0
0
0
0
H2SOA*
.034
.024
.003
.003
.003
.002
.003
(22)
(22)
(3)
(4)
(3)
(2)
(4)
FTP
SET
SET
FET
SET
SET
64 kph, 1 hr
Bracketed numbers are % conversion of fuel sulfur to sulfate,
based on emitted sulfate.
Initial sulfate emissions were higher with the DuPont monolith;
but S02 conversions averaged about 3% on the last five runs of the test
sequence. Conversions with the Engelhard system averaged (first six
runs) about 7.5%. The Engelhard system exhibited considerably higher
sulfate emissions on the 64 km/h cruise, 0.026 g/km of H2S04, corres-
ponding to 30% conversion. The sulfate emission results are shown in
Tables V-ll and V-12.
-------
- 88 -
Table V-12
Emission Characteristics, Vehicle Equipped with
Engelhard Monolithic Oxidation Catalyst in
High Temperature Configuration
Test
FTP
SET
SET
FET
SET
SET
64 kph, 1 hr
CO
1.08
1.63
0.83
0.60
0.46
0.89
0.08
Emission
HC
0.15
0.09
0.22
0.24
0.08
0.25
0.04
Rate,
NOx
3.03
1.00
2.99
2.66
2.82
2.68
1.60
g/km
H2S04*
0.009 (6)
0.009 (8)
0.004 (4)
0.009 (9)
0.010 (11)
0.008 (7)
0.026 (30)
Bracketed numbers are % conversion of fuel sulfur to sulfate,
based on emitted sulfate.
These results Indicate that the monolithic DuPont catalysts
are a lower emitter of sulfate than their palletized counterpart. However,
due to its poorer control of CO and HC compared to commercial catalysts,
no further work is planned with the DuPont monolith.
-------
- 89 -
VI. References
1. K. C. Bachman, et al., "An Assessment of Automotive Sulfate Emission
Control Technology," EPA Report No. 460/3-76-017.
2. M. Beltzer, et al., "The Conversion of S02 Over Automotive Oxidation
Catalysts," SAE Paper 750095, February, 1975.
3. R. L. Bradow and J. B. Moran, "Sulfate Emissions from Catalyst Cars:
A Review," SAE Paper 750090, February, 1975.
4. R. A. Bouffard, "Fuel Sulfur Effect on Automotive Sulfate Emissions,"
American Petroleum Institute Publication No. 4277, September, 1976.
5. A. Lauder, DuPont, Private Communication.
6. R. England and D. J. Gunn, Trans. Inst. Chem. Eng., 48, T265 (1970).
7. E. L. Holt, et al., "Control of Automotive Sulfate Emissions," SAE
Paper 750683, June, 1975.
8. M. Beltzer, et al., "Measurement of Vehicle Particulate Emissions,"
SAE Paper 740286, February, 1974.
9. Bulletin No. 07169, "Honeycomb Industrial Dehumidifiers," Honeycombe
Industrial Division, Cargocaire Engineering Corporation, Amesbury,
Massachusetts.
10. Electronic Control Systems, Fairmont, West Virginia.
11. Electric Pneumatic Transducer, Model No. T5129, Fairchild Industrial
Products Division, Winston-Salem, North Carolina.
12. Pneumatic Controller, Model and Size B-51XC4, Conoflow Corporation,
Blackwood, New Jersey.
13. Dunham-Bush Corporation, West Hartford, Connecticut.
14. Coolenheat Incorporated, Linden, New Jersey.
15. H. Schlicting, "Boundary Layer Theory," New York, McGraw Hill Book
Co., Inc., pp. 504-5 (1960).
16. K. Habibi, Env. Sci. and Technol., 4, 239 (1970).
17. J. B. Moran and 0. J. Manary, Interim Report PB 196783, "Effect of
Fuel Additives on the Chemical and Physical Characteristics of Par-
ticle Emissions in Automotive Exhaust," NAPCA, July 1970.
-------
- 90 -
VI. References (continued)
18. Instruments for Measurement and Control of Relative Humidity,
Brochure B-ll and Form D-ll, Phys-Chemical Research Corporation,
New York.
19. Sulfur Dioxide Pulsed Fluorescent Gas Analyzer Model 40, Thermo
Electron Corporation, Waltham, Massachusetts.
20. Instruction and Operation Manual PD 101, Perma Pure Dryer, Perma
Pure Products, Inc., Oceanport, New Jersey.
21. H. Goksrfyr and K. Ross, J. Inst. Fuel, 35:177 (1962).
22. S. Krause, et al., "Critical Factors Affecting Automotive Sulfate
Emissions," SAE Paper 760091, February, 1976.
-------
- 91 -
APPENDIX A
Selection of Calcium, Silica, and Alumina
Components for Physical Combinations
1. Calcium Hydroxide
Calcium hydroxide (technical grade) is agglomerated by slowly
adding v/ater to powdered calcium hydroxide while stirring slowly in a
mixing apparatus. A quantity of 400 g of calcium hydroxide is mixed with
approximately 200 g of distilled water for the agglomeration. The product
is dried in an oven at 65°C to a moisture content of approximately 18 wt Z
and passed through a 14 mesh screen. Final moisture content is determined
by drying a sample for four hours at 110°C and calculating the percent of
weight loss.
2. Calcium Oxide
Source: Pfizer Incorporated
Mineral, Figments, and Metals Division
Description: "High Calcium Granular Quicklime"
10-100 mesh
94.0-96.0% CaO, balance inert
Before use this material is screened to remove particles larger
than 14 mesh (0.0469 in., 1.19 mm).
3. Calcium Carbonate
Source: Pfizer Incorporated
Mineral, Pigments, and Metals Division
Description: "Nelco High Calcium Limestone, ATF-20"
20+ mesh
94.0-96.0% CaC03, balance inert
4. Silica
Source: PPG Industries, Incorporated
Description: "Hi-Sil 210 Hydrated Amorphous Silica"
94 wt % Si02> balance inert
5, Alumina
Source: Aluminum Company of America
Description: Alcoa Hydrated Alumina C-31 (coarse)
64.9 wt % A1203 equivalent
(calculated for Al203'3H20 or A1(OH)3,
A1203 = 65.4 wt. 2)
-------
Composition, Wt %
Calcium Hydroxide (1)
Calcium Oxide (2)
Calcium Carbonate (3)
Silica (4)
Alumina (5)
Sterotex (6)
Operation During Pelleting
Flow into dies:
Ease of compaction/
ejection:
Crush Strength of 3.2 mm Pellets
Initial Crush Strength (kg)
Crush Strength after
16 hrs @ 315°C (kg) (7)
,
Notebook Reference 4349 -
1
85
-
-
15
-
-
fair
fair
2.7
4.1
33-1
2
85
-
-
-
15
-
fair
fair
3.5
4.5
33-2
3
85
-
-
7.5
7.5
-
fair
fair
2.9
4.4
33-3
4
-
85
—
15
-
+3%
good
poor
too soft
powdered
32-1
5
-
85
—
-
15
+3%
good
difficult
2.6
powdered
32-2
6
-
85
—
7.5
7.5
+3%
good
poor
0.6
powdered
32-3
7
-
—
85
15
-
+3%
good
poor
0.2
0.2
34-1
8
-
-
85
-
15
+3%
good
poor
0.6
0.2
34-2
9
-
—
85
7.5
7.5
+3%
good
poor
0.1
0.1
34-3
i
VO
1
(1) Calcium hydroxide is wet granulated, dried at 65°C, and passed through a 14 mesh screen. Final moisture
content = 15 wt %. In formulating, the weight of granular calcium hydroxide is adjusted to allow for
the moisture content.
(2) Pfizer High Calcium Granular Quicklime, 14-100 mesh. After mixing with binder these formulations were
reground and screened to 20+ mesh.
(3) Pfizer Nelco High Calcium Limestone, ATF-20, 20+ mesh. After mixing with binder these formulations were
reground and screened to 20+ mesh.
(4) Hi-Sil 210 Hydrated Amorphous Silica, 94 wt % Si02.
(5) Alcoa Hydrated Alumina C-31 (coarse), 64.9 wt % A1203 equivalent. In formulating .weight of Alcoa C-31
is adjusted to give the required concentration of
(6) Sterotex, lubricant for tablet making, Capital City Products Company, Columbus, Ohio.
(7) The pellets are heated in an oven 48 h at 95°C, then 6 h at 150°C, followed by 16 K at 315°C.
-------
- 93 -
APPENDIX B
TABLE B-2
Preparation of Chemical Combinations
_ By Coprecipitation _
1. "Tricalcium Silicate." (CaO)i SiC-2
Calculated for (CaO)3 Si02:
CaO = 73.68 wt %
Si02 = 26.32 wt %
Materials; Ca(N03)2'4H20, reagent grade, 1018 g = 4.31 moles
Na2Si03-9H20, reagent grade, 408 g = 1.44 moles
NaOH, 50 wt % solution, 460 g = 5.75 moles
Procedure ;
The calcium nitrate is dissolved in 3 litres of distilled water
in a 10 litre jar. A solution is made of the sodium metasilicate in 700 ml
of distilled water to which is added the sodium hydroxide solution. The
combined solution is added to the calcium nitrate solution via an addition
funnel over one hour. The resultant mixture is stirred for an additional 30
minutes and filtered by vacuum using a large Biichner funnel. The filter
cake is washed four times with 1000 ml of distilled water, dried in an oven
at 65 °C for 20 hours, and screened to 14-35 mesh.
NBR-4349-28
2. "Tricalcium Aluminate." (CaO)^ AlzP.q
Calculated for (CaO)3 A1203:
CaO = 62.25 wt %
A1203 = 37.75 wt %
Materials: Ca(N03)2'4H20, reagent grade, 859 g = 3.64 moles
Nalco No. 5, solution of sodium aluminate,
(equivalent wt % A1203 = 26.4, wt % NaOH = 24.4) use
468 g (=1.21 moles A1203 +2.85 moles NaOH)
NaOH, 50 wt % solution, 353 g = 4.41 moles
-------
- 94 -
Procedure;
The calcium nitrate is dissolved in 4 litres of distilled water
in a 10 litre jar. The Nalco No. 5 and sodium hydroxide solutions are
mixed in an addition funnel and added to the calcium nitrate solution
with rapid stirring over 30 minutes. The mixture is stirred for an addi-
tional 30 minutes, filtered by vacuum through a large Buchner funnel. The
filter cake is washed four times with 1000 ml of distilled water, dried
in an oven at 65°C for 20 hours, and screened to 14-35 mesh.
NBR-4349-30
3. "Tricalcium Aluminosilicate." (CaO)o (AloOs'
Calculated for (CaO)3 (Al203'Si02) /2
CaO = 67.48 wt %
Si02 = 12.05 wt %
A1203 = 20.47 wt %
Materials; Ca(N03)2'4H20, reagent grade, 1423 g = 6.03 mole
>, reagent grade, 285 g m 1.00 mole
Nalco No. 5, solution of sodium aluminate, 388 g -
1.00 mole A1203 +2.37 mole NaOH
NaOH, 50% solution, 614 g = 7.68 mole
Procedure;
The calcium nitrate is dissolved in 4 litres of distilled water
in a 10 litre jar. The sodium metasilicate is dissolved in 800 ml of
distilled water and mixed with 307 g of 50% sodium hydroxide solution in
an addition funnel. In a second addition funnel the Nalco No. 5 is mixed
with the remaining 307 g of 50% sodium hydroxide, plus 250 ml of distilled
water to make the liquid volumes equal in the two addition funnels. The
contents of the two addition funnels are simultaneously added to the cal-
cium nitrate solution aver a period of 30 minutes with rapid stirring.
Stirring is continued for 30 minutes.. The mixture is filtered by vacuum
using a large Buchner funnel. The filter cake is washed four times with
1 litre of distilled water, dried in an oven at 65°C for 20 hours, and
screened to 14-35 mesh.
NBR-4349-41
-------
- 95 -
APPENDIX B
TABLE B-3
PRESSURE DROP ACROSS S04= TRAP
Benchmark Rings (12.7 mm OP)
AP. kPa
Kilometre Idle 64 km/h
0 0 0.5
6 400 0.5 0.9
16 000 0.5 1.0
Benchmark Pellets (4.76 mm OP)
AP. kPa
Kilometre Idle 64 km/h
0 0 1.0
5 000 0.5 1.2
15 000 0.7 1.0
Ca(OH)2/Picalite Pellets (4.76 mm OP)
AP, kPa
Kilometre Idle64 km/h
0 0.2 0.75
5 000 0.2 0.5
15 000 0 0.75
-------
- 96 -
APPENDIX C
MEASUREMENT TECHNIQUES
C.I Gaseous Emissions
Gaseous emissions were measured using standard instrumentation
for Federal Emission Test Procedures. CO and C0£ were analyzed by NDIR,
hydrocarbons by FID, and NOX by chemiluminescence.
C.2 Measurement of Sulfate Emissions
Sulfate emission samples were collected using Exxon Research's
exhaust particulate sampling system. The sulfate content of samples
collected was determined by titration of sulfate to the thorln end point
using an automatic recording titrator and a probe colorimeter.
C.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.
C.2.1.1 Sampling System Components
The particulate sampler which has been discussed previously
(8 ) is shown schematically in Figure C-l._ 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 C-l
EXHAUST PARTlCULATt SAMPLER
INTAKE
DILUENT AIR
1
n
DEHUMIDIFIER
P RE- COOLER
\
HEAT
EXCH;
T
-/
^
L-^
(\NGE
J FILTER
~| BOX
^ COUPLED
MIXING
BAFFLES
i/4
f » » T
^ ilil
, i
X
: 10
AIR COOLED
COMPRESSOR
FLOW
DEVELOPMENT
TUNNEL
MIXING
TURBULATORS
TO CVS
PUMP
i
\o
ISOKINETIC
SAMPLING
PROBE
EXHAUST
INJECTOR
FILTER
HOUSING
-------
- 98 -
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.
C.2.1.2 Diluent Air Preparation System
This system consists of a dehumidifier, filter, coupled mixing
baffles, a cooling system, and mixing turbulators.
The dehumidifier 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
dehumidification 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 and described in their Bulletin No. 07169(9).
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 dehumidification 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 (!0). The output signal
-------
FIGURE C-2
SCHEMATIC OF DEHUMIDIFICATION SECTION
REACTIVATION AIR
INLET FILTER
WET AIR OUTLET
TO FILTER
BOX
REACTIVATION
FAN
DESICCANT
REACTIVATION WHEEL
AIR HEATER
REACTIVATION
SECTOR
DRY AIR FAN
HUMID AIR
INLET
FILTER
VO
VO
-------
- 100 -
from the controller is fed to an electric to pneumatic transducer (11)
which in turn activates a pneumatic controller(12) 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(13)'containing ten rows of custom-made cooling coils
(14).
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-vmin.
C.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.
C.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 acconplished by injecting the ex-
haust in a countercurrent direction to the diluent air stream. Previous
experiments (&) 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.
d . 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 0-3
COUNTER CURRENT EXHAUST INJECTION SYSTEM
TRICLOVER
CONNECTIONS
RAW EXHAUST
DILUTION
AIR
TAILPIPE CONNECTOR
DILUTED EXHAUST
V
COUNTERCURRENT
INJECTOR
i
M
O
M
I
4.3" ID S.S. FLOW
DEVELOPMENT
SECTION
-------
- 102 -
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.
C.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.
C.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 (8) and will be
reviewed in this section.
C.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
-------
- 103 -
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.
C.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 (15). 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.
"Re - (!)
where N^e = Reynolds Number
D = pipe diameter = 11.0 cm = 0.110 m
U = average fluid velocity = * (*•!$* 1370 m/min
* DZ
8.21 x 10A n/hr.
p = density = 11.14 kg/n3
y = fluid viscosity = 0.186 cp = 0.670 kg/m-hr.
N =(0.110 m) (8.2 x 10* m/hr) (1.14 kg/m3)
NRe - 0.670 kg/m-hr - -
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 %e to 615,000 while increasing pipe diameter to 40 cm
would decrease NRe to 40,000.
-------
- 104 -
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 (15)
00)
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 tine
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 comproaise
between these various factors.
-------
- 105 -
C.2.2.3 Tunnel Sampling Losses
Particulate deposition in the flow development section was
measured by introducing an artifically produced nono-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 (3).
C.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
(WA) 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:
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 Wg Fp
_a_ = — = — = etc.
km F^ Akm Fg 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 (8).
-------
- 106 -
C.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 (16,17). 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 C-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 C-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 (18).
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 C-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 m^/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 C-4
0)
M
3
4-1
X
m 'tU
n)
w
M
•H
-------
FIGURE C-5
RELATIVE HUMIDITY OF EXHAUST DILUTION AIR
MIXTURE AT VICINITY OF SAMPLING PROBES DURING
THE 1972 FEDERAL TEST-DRIVING CYCLE
100
Q.
E
in
Q.
in
60
30
o
00
234 5 6 7 8 9 10 11 12 13
TIME, HUNDREDS OF SECONDS
-------
- 109 -
o
o
g
W
w
E-H
FIGURE C-6
TEMPERATURE CONTROL SYSTEM PERFORMANCE
CATALYST-EQUIPPED CAR
50
40 —
30 _
20 —
10 _
I
13 m3/MIN CHILLED AIR |
SYSTEM WITH FINNED TUBE
TIME
THE 1972 FTP
-------
- 110 -
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 C-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 C-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.
C.3 The Goks^yr-Ross Technique
This technique (21) 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 I^SO^ in the
exhaust. As a result, H2S04 is condensed in the coll, 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 litres/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 use of metal lines, only tubing which has been passivated by
prior exposure to H2S04 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 Thorln method described in Section C.5.
C.4 The TECO Sulfur Dioxide Analyzer
This device (19) 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
-------
FIGURE C-7
RELATIVE HUMIDITY OF EXHAUST DILUTION
AIR MIXTURE AT VIC/N1TY OF SAMPLING
PROBES DURING 64 Wh CRUISE CONDITIONS
at 32°C
100
> ST 75
LU
£<^ 25
5 10
TIME (MINUTES)
15
-------
- 112 -
FIGURE C-8
FINNED TUBE COOLING SETUP
Raw Exhaust
Outlet
nun it 1 1 1 1 1 I 1 1 1
1111 II 1 1 I Ml
u
To Exhaust
Injector
Cooling Fins
Air Inlet
-------
- 113 -
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 S02 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 S02 in diluted
exhaust can be made.
(3) Measurement of 502 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 C-9.
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 S02.
The water removal problem was solved by using the Permatube
Drying System (2°) shown in Figure C-10. 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 counter-
current 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 H20 while retaining the S02 in the sample has been established by
our Analytical Division. The Model PD-500-72 Perma Pure Dryer according
to the manufacturer (20) 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.
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 intro-
duced directly into the analyzer, or what 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.
-------
FIGURE C-9
PRINCIPAL OF OPERATION
TECO SO? INSTRUMENT
FILTER
PULSATING
ULTRAVIOLET
LIGHT
SAMPLE
GAS OUT
SAMPLE GAS CONTAINING S02 IN
PHOTOMULTIPLIER
TUBE
ELECTRONICS
-------
- 115 -
It has been found that C02> CO, and (>2 are strong quenching
agents, while N£ 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 SC>2 conversion as a function
of oxygen concentration, it was necessary to assess the effect of oxygen
quenching (2). Various samples were made by preparing bell jar mixtures
containing 30 ppm SC>2, 12% C02, varied amounts of oxygen, and nitrogen as
the balance gas. Measurements of the S02 concentration of these mixtures
indicate an approximate 1 ppm reduction in instrument S02 response for
each 2% increment in oxygen concentration, as shown below.
Table C-l
S02 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 SO- in background air
containing different concentrations of CO, 02, 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
from the calibration gas quencher level should be corrected for the
inherent changes in instrument response.
-------
- 116 -
Table C-2
Composite Effects of C02, 02 and
CO on TECO SO,, Response
Total Quencher
Mixture Species Concentration (%) Instrument
Composition fcc^ + Ct^l + CCQ] Response (ppm)
30 ppm SO-
1.42% 0,
0.09% CO
0.051% H2 14.0 28
445 ppm
12.5% C02
30 ppm SO2
4.78% 02
14.3% C02 23.4 23
4.33% CO
348 ppm C3H8
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 SO--
C.5 Analytical Method for Determination of Sulfate on Glass Fiber
Filters Using a. Recording Titrator and Probe Colorimeter
The filters are extracted with water, and the resulting solution
is passed through an ion exchange resin to remove cations. Sulfate is
then titrated to the thorin endpoint using a recording titrator and probe
colorimeter. Sulfate can be measured if it is present in solution in
amounts of 10 tig or greater. The use of a recording titrator interfaced
with a probe colorimeter to sense colorimetric endpoints is better suited
to determination of low levels of sulfate than titration to a visual
endpoint. The titration system automatically plots the volume of
titrant dispensed from a driven burette against the change in color of
the indicator.
-------
- 117 -
Reagents
1. Barium perchlorate solution, 0.00521 M, available from Each
Chemical Company, Ames, Iowa.
2. Barium perchlorate solution, 0.000521 M in 80% 2-propanol pre-
pared from the above solution. (Propanol, distilled in glass,
is available from Burdick and Jackson Labs, Inc., Muskegan,
Michigan.)
3. Thorin indicator, 0.2% in water. This indicator is available
as a powder from the Hach Chemical Company,
4. Dowex 50W-X8 Cation Exchange Resin, 50-100 mesh,
5. Sulfuric acid solution 0.1000 N for standardizing the barium
perchlorate solution. This solution is available from the
Harleco Division of the American Hospital Supply Corporation,
Gibbstown, New Jersey. As received, this acid solution is
standardized to contain 4.90 g 22804 per litre at 25°C. The
acid is standardized at Harleco against tris hydroxymethyl
amlnomethane using methyl red as the indicator.
Standardization of the Barium Perchlorate Titrant
The burette reservoir and burette are filled with 0.000521 M
barium perchlorate solution. The electronic agreement of the PC/600 and
the E536 titrator is checked as described in the manufacturer's instruct
tion manual of operation. The instrumental parameters are set as indi-
cated below:
o Burette speed - 10 min/100% volume
• Mode switch - pH/mv position
• Chart range - 1 pH full scale
• Chart speed - 400 mm/100% volume
• Temperature compensator - set at room temperature
• Colorimeter filter - 520 nm
The position of the probe in the titration vessel is situated so that
the solution can be stirred through the probe without bubble formation,
since this would result in an unstable baseline.
-------
- 118 -
1. Pipet 5 ml of previously standardized 0.001 N H2S04 into a 100 ml
beaker equipped with a Teflon-coated magnetic stirring bar. This
contains 0.005 meq H2S04, (0.0025 m mol S04= 0.24 mg 804*).
2. Add 5 ml of water and 40 ml of 2-propanol to the beaker.
3. Add 3 drops thorin indicator to the beaker.
4. Degas the solution by stirring for 5 min. on a magnetic stirrer,
5. Place the beaker on the tltration stand and put the colorimeter
probe and burette tip into the beaker.
6. Titrate the solution and determine the endpoint, as shown in
Figure I, record the volume, VA>
7. Calculate the sulfate equivalent of the barium perchlorate titrant,
C, as follows:
mg 804* 0.24 mg 804*
ml Ba(C104>2 sol'n VA ml 0.000521 M Ba(C104>2
Ins trumentation
Sulfate titrations are performed using a Brinkman Metrohm
Herisau Potentiograph Recording Titrator. This system automatically
plots the volume of titrant dispensed versus the color of the indicator.
This titration system consists of the following components:
• E356 titrator
• E538/4 titration stand
• E522-10B, 10 ml burette
• Brinkman Probe Colorimeter (PC600 colorimeter)
The recorder output of the PC600, 100 mv, is directly related to the 100
mv range of the E536 titrator. The chart drive of the recorder is
synchronized with the burette drive, . Burette run-out times can be
varied from 5 to 200 minutes corresponding to chart speeds of 80 to
per minute.
-------
- 119 -
FIGURE C-10
TITRATION OF SULFATE USING THORIN INDICATOR
40
50
w
o
OT
60
70
80
90
ml 0.000521 M Ba(0104)2 Tltrant
-------
- 120 -
The Brinkman Probe Colorimeter consists of a PC600 colorimeter,
a 2 cm light path stainless steel fiber optics probe and 520 or 650 nm
interference filters. The filter colorimeter employs a fiber optics
probe which is placed in the solution to be titrated* Phase shifted AC
amplified light is transmitted to the probe to avoid interference from
ambient light.
The entire titrator assembly is available from Brinknan Instru-
ments, Westbury, New York.. A detailed description of the titrator-
colorimeter system is contained in the manufacturers' instruction manuals
of operation (Instructions for Use E536 Metrohm AG CH-9100 Herisau and
Operating Instructions for the Brinkman PC600 Probe Colorimeter). Figure
C-12 shows a block diagram of the colorimeter and the photometric probe.
Stepwise Procedure of Work-Up of Sulfate-
Containing Glass Fiber Participate Filters
1. The filter to be analyzed is cut into quarters with an exacto
knife using a. right angle lucite template placed on the filter
as a guide. No part of the template actually touches the part
of the filter to be analyzed.
2. The filter quarter section is folded, put into a 100 ml beaker,
and covered with 25 ml of doubly distilled deionized water.
Care is taken to ascertain that the whole filter is wet and
covered with water.
3. The beaker is covered with a watch glass and glass stirring rod
inserted between the beaker and watch glass.
4. The sample is digested by placing the beaker on a hot plate
(set low) and allowed to boil for 5 minutes. Following di-
gestion, the sample is allowed to cool down to room temperature.
5. The supernatant is filtered either through fine glass wool
(Pyrex Brand Wool, Corning Glass Works) or a paper filter
(9 cm diameter, Schleicher and Schuell, Inc.) into an ion-
exchange column (1 cm x 25 cm) containing about 7.5 ± 0.5
inches of Dowex 50 WX8 cation exchange resin, collecting the
sample in a 50 ml volumetric flask.
6. The beaker is washed three times with 5 ml portions of distilled
deionized water, the washings added to the column. After each
washing, liquid is squeezed from the filter by pressing it
against the wall of the beaker with the stirring rod and added
to the column with the wash solution.
-------
FIGURE C-ll
BLOCK DIAGRAM OF BRINKMAN PC600 PROBE COLORIMETER
Constant Voltage
Transformer
O-
Output
Meter
Regulator Lamp
-f 1-0
Reference
Cell
Amplifiers
Synchronouj
Detector
^ i { Filter
Signal
Coll
Pvobs
i
M
Nl
M
I
-------
- 122 -
7. The volume of liquid in the volumetric flask is made up to
the mark by the addition of distilled, deionized water. This
solution is the basic filter leach solution from which the
sulfate content of the filter and ultimately the vehicle
sulfate emission rate is determined.
Titration of Filter Leach Solution for Sulfate
1. A suitable aliquot, Vg, is added to a 100 ml beaker, contain-
ing a magnetic stirring bar.
2. Sufficient water is added to the beaker so that the total
volume (aliquot plus added water) equals 10 ml.
3. 40 ml of 2-propanol and 3 drops of the thorln solution are
added to the beaker.
4. The solution is degassed by stirring for 5 minutes with a
magnetic stirrer.
5. The solution is titrated using the Brinkman titrator to deter-
mine the endpoint with Vc ml of titrant* recording the titrant
volume. The endpoint is determined from the absorbance versus
volume of titrant chart as shown in Figure C-ll.
In the process of developing this analytical procedure, the
following precautionary steps were found to be necessary;
1. Insure that sufficient column capacity remains to remove the
metals in the sample. Use a new Dowex column on each high
level sample. Dowex can't be regenerated with EC1 since
results show a carry-over.
2. Insure that at least 1 ml of the titrant Is consumed in the
titration. If less than 1 ml is consumed, place a 20 ml
aliquot of the sample in a 150 ml beaker, add 80 ml of
2-propanol and add a known quantity of sulfuric acid suffi-
cient to give a titration of 1 to 3 ml. Titrate the solution.
Calculate as outlined below and subtract the known mg 804°
added to give the mg 804" in the filter.
mg 804** C mg 804" 50 ml 4
ml Ba(C104)2 soln) - m Ba (C104)2 soln * vT*T * 1 filter
aliquot filter
factor factor
-------
- 123 -
Additional Procedural Changes
To demonstrate that water and nitric acid remove equivalent
amounts of sulfate from the filters, comparis.on extraction data was ob-
tained. One quadrant of fiber glass filters from vehicle tests was
extracted with dilute nitric acid, the metals were removed by a cation
exchange resin and the sulfate content was determined by the Sulfanazo III
procedure. Another quadrant of the filters was extracted with water,
the metals were removed by a cation exchange resin and the sulfate con-
tent was determined by the thorin procedure.
The results which were obtained are given in table C-3 and in-
dicate the extraction procedures are equivalent. On the basis of these
results, it was decided to utilize the instrumental technique because
of its inherent greater sensitivity and precision, coupled with thorin
as the indicator.
When the visual titrimetric procedure for sulfate was supplanted
by the automatic titrimetric method, the indicator used was also changed
from Sulfanazo III to Thorin, to simplify the procedure. This indicator
substitution necessitated the use of a water extraction of the filters
in place of the original nitric acid extraction.
-------
- 124 -
TABLE C-3
SULFATE EXTRACTIONS USING DILUTE NITRIC
ACID AND WATER - COMPARISON DATA
Filter g. AL
708631
32
33
34
35
36
37
38
39
40
708617
18
19
20
21
22
23
24
25
26
27
28
meSOA"/filter
H20 Extraction
27.21
26.13
1.09
0.84
1.06
1.25
1.25
1.19
0.81
0.95
3.68
2.59
2.66
3.62
4.31
2.30
2,21
0.20
3.96
3.49
3.14
2.15
HN03 Extraction
28.28
29.09
0.90
1.11
1.04
1.25
1.49
1.65
0.99
1.00
3.38
2.51
2.73
3.50
4.22
2.36
2.30
0.15
3,19
3.03
2.69
2.33
ZA (HN03 - H20)
+ 3.8
+10.2
-22.2
4-23.4
- 1.9-
0.0
-16.1
+27.9
+18.2
+ 4,0
Average - 2.3
-------
Kilometres
0, No Trap
(Base Case)
0, No Trap
(Base Case)
0, With Trap
0, With Trap
Test Sequence
FTP
SET
SET
FET
SET
SET
FTP
SET
SET
FET
SET
SET
FTP
SET
SET
FET
SET
SET
FTP
SET
SET
FET
SET
SET
APPENDIX D
TABLE D-l
TEST CAR 124 - SC^1
3 TRAP VEHICLE(1)
CONTAINING BENCHMARK RINGS (12.7
(Fuel
Sulfur,
320 ppm)
ram OD)
Exhaust Gas Emissions
CO
_____
1.99
0.87
1.05
0.26
0.95
0.87
3.39
1.24
0.71
0.25
0.87
0.50
2.92
0.57
0.90
0.38
0.48
0.99
2.76
0.53
0.26
0.65
_— _
HC
g/km -
0.18
0.14
0.15
0.12
0.12
0.12
0.28
0.11
0.10
0.12
0.27
0.10
0.31
0.22
0.11
0.11
0.12
0.13
0.28
0.11
— — _
0.27
0.21
— —
NOX
4.26
7.82
7.78
9.46
8.13
7.94
5.53
8.09
7.94
10.19
7.41
8.34
5.73
7.71
7.80
9.68
8.45
7.53
5.86
8.25
_
9.95
8.31
____
S02
- - mg/km
49.6
91.5
80.2
44.8
90.5
72.1
42.7
99.3
64.9
47.2
72.6
93.8
43.1
26.6
32.6
40.6
27.8
42.2
21.9
34.1
38.5
29.0
51.0
31.4
S04=
_ _ _
7.7
10.8
5.1
10.9
6.2
3.7
5.5
4.9
4.1
10.7
4.8
8.5
4.3
3.6
2.7
1.8
2.4
2.1
5.1
3.4
2.8
4.5
3.1
3.2
S04= Trapping
Efficiency, %
34.8
40.0
40.0
83.3
60.0
65.0
22.7
43.3
53.3
58.3
48.3
46.7
NJ
Cn
(1) 1974 car equipped with PTX-IIB oxidation catalysts and air pump.
-------
TABLE D-l (CONTINUED)
Exhaust Gas Emissions
Kilometres
6 400, With Trap
16 000, With Trap
16 000, No Trap
Test Sequence
FTP
SET
SET
FET
SET
SET
96 km/h
64 km/h
FTP
SET
FTP
SET
CO
2.82
0.31
0.55
0.11
0.80
0.78
0.04
0.02
1.39
0.46
3.03
0.36
HC
0.12
0.08
0.09
0.04
0.28
0.21
0.04
0.02
0.25
0.18
0.41
0.14
NOX
3.27
5.87
5.23
4.78
4.83
4.51
8.01
3.18
4.38
5.07
5.27
5.55
SO 2
41.8
50.3
50.5
14.3
50.4
50.6
37.7
6.1
20.4
58.5
85.5
72.3
SQ4=
12.0
5.4
3.6
8.0
4.4
4.2
15.8
19.5
7.3
5.1
6.4
10.5
Trapping
Efficiency, %
0
10.0
40.0
25.9
26.7
30.0
0
51.4
-------
Kilometres
0, No Trap
(Base Case)
0, With Trap
5 000, With Trap
APPENDIX D
TABLE D-2
TEST CAR 124 (D
CONTAINING
BENCHMARK
- SOA= TRAP
PELLETS (4
VEHICLE
.76 mm OD)
*(2)
(Fuel Sulfur, 320 ppm)
Test Sequence
96
96
96
FTP
SET
SET
FET
SET
SET
km/h (2 hours)
FTP
SET
SET
FET
SET
SET
km/h (2 hours)
FTP
SET
SET
FET
SET
SET
km/h (2 hours)
CO
1.97
0.49
0.37
0.49
0.35
0.49
0.07
1.95
0.36
0.43
0.40
0.46
0.44
0.07
2.11
0.22
0.43
0.27
0.81
0.45
0.06
Exhaust
HC
— o /Irm — —
0.29
0.14
0.13
0.18
0.12
0.15
0.08
0.20
0.15
0.13
0.20
0.12
0.15
0.08
0.27
0.09
0.10
0.11
0.11
0.15
0.07
Gas Emissions
NOX
7.26
6.68
3.78
4.61
5.91
6.35
10.67
5.90
5.15
5.50
6.37
4.91
5.05
8.91
4.04
4.95
5.12
5.80
4.38
4.72
5.76
S02
- - mg/km
63.0
5.1
5.1
51.9
5.1
60.8
46.6
6.0
5.0
5.0
6.5
5.2
5.2
2.3
40.1
41.7
45.6
20.7
49.3
47.0
10.1
S04~
- - -
4.1
6.6
8.0
10.6
9.6
11.1
25.9
2.7
0.5
0.5
0.3
0.4
0.2
0.4
2.9
0.9
1.1
0.8
0.8
0.7
0.5
804- Trapping
Efficiency, %
34,
94,
94,
97,
95,
97,
98.5
29.0
89.8
87,
92.
90.
92.0
98.1
,5
.5
.9
I
M
to
I
(1) Test Car 124 - 1974 car equipped with two PTX-IIB catalysts and air pump.
(2) No trap where noted.
-------
TABLE D-2 (CONTINUED)
Kilometres
15 000, with trap
15 000, no trap
(Base Case)
Test Sequence
FTP
SET
SET
FET
SET
SET
96 km/h (2 hours)
FTP
SET
FET
96 km/h (1 hour)
S04= Trapping
CO
2.03
0.46
0.42
1.74
0.57
0.88
0.29
HC
.
0.27
0.11
0.13
0.17
0.11
0.11
0.12
NOX
3.77
3.87
3.78
3.57
3.33
3.26
4.80
SO 2
SO 4=
Efficiency, %
- — — mg/km- — - l' '
72.5
68.0
54.0
63.2
64.5
68.0
30.0
4.1
2.0
1.8
1.5
1.5
2.0
3.0
0
77.3
79.5
85.8
83.0
77.3
88.4
U<2>
0
0
14.3
28.6
73.2
4.8
83.9
2.31
0.71
0.51
0.20
0.37
0.14
0.16
0.08
3.45
4.36
5.93
7.29
6.5
83.1
37.4
59.0
3.0
2.1
5.6
18.7
OO
(1) % efficiency calculated using 0 km no trap S04= levels.
(2) % efficiency calculated using 15 000 km no trap S04= levels.
-------
Kilometres
0, No Trap
(Base Case)
0, With Trap
5 000, With Trap
15 000, With Trap
APPENDIX D
TABLE" D-3
TEST
CONTAINING
CAR 99 (D - S04= TRAP VEHICLE
Ca(OH)2/DICALITE (4.76 mm PELLETS)
(Fuel Sulfur,
Test Sequence
96
96
96
96
FTP
SET
SET
FET
SET
SET
km/h (2 hours)
FTP
SET
SET
FET
SET
SET
km/h (2 hours)
FTP
SET
SET
FET
SET
SET
km/h (2 hours)
FTP
SET
SET
FET
SET
SET
km/h (2 hours)
CO
-
1
0
0
0
0
0
0
0
0
0
1
0
0
0
2
0
0
2
0
1
0
2
0
0
2
0
1
0
_ _ _
.11
.34
.23
.33
.55
.45
.16
.91
.28
.42
.54
.82
.53
.21
.19
.62
.59
.91
.69
.53
.36
.50
.75
.70
.95
.98
.11
.58
320 ppm)
Exhaust Gas Emissions
HC NOy
- g
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
/Vm
.14
.08
.06
.06
.10
.10
.05
.14
.06
.11
.11
.09
.14
.03
.25
.09
.12
.09
.08
.09
.04
.49
.17
.14
.16
.14
.18
.05
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
2
3
2
3
3
3
2
2
3
.30
.58
.69
.80
.68
.64
.25
.34
.39
.34
.50
.49
.55
.30
.41
.67
.67
.64
.27
.38
.03
.71
.30
.21
.72
.70
.86
.30
S02 S04~
- - mg/km - - -
10
35
33
30
34
38
25
15
9
6
7
17
10
27
48
31
50
27
50
21
26
61
46
20
33
44
46
45
.9
.0
.2
.9
.8
.8
.8
.9
.5
.0
.9
.8
.6
.8
.9
.1
.0
.9
.4
.9
.8
.4
.5
.3
.3
.9
.6
.3
4
29
31
35
42
50
51
1
0
0
0
1
1
5
10
19
15
12
15
15
20
12
13
15
13
18
17
19
.5
.5
.6
.4
.0
.4
.7
.5
.6
.6
.8
.3
.8
.7
.8
.3
.7
.3
.7
.0
.6
.7
.0
.0
.4
.1
.8
.4
S04= Trapping
Efficiency, %
66.6
98.4
98.4
97.7
96.6
95.3
89.0
0
49.7
59.1
65.3
59.1
i
M
N>
1
60.9
60.2
0
66.
60.
62,
52.8
53.6
62.5
.1
.9
,1
(1) Test Car 99 - 1973 car with two PTX-IIB oxidation catalysts and an air pump.
-------
- 130 -
APPENDIX E
TABLE E-l
1975 VEHICLE (351 CID)
Pt/Pd OXIDATION CATALYSTS (AEW 2/6C/50M)
(Fuel Sulfur - 0.032 ttt %)
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
1975 FTP, g/km
CO
2.62
2.45
3.39
2.93
2.48
CO
0.12
0.11
0.36
0.19
0.29
CO
0.40
0.77
0.94
0.26
0.98
HC
0.78
0.29
0.39
0.43
0.28
HC
0.06
0.14
0.14
0.07
0.09
HC
0.11
0.09
0.12
0.07
0.12
NOX
1.67
1.64
1.61
1.70
1.84
FET,
NOX
2.66
2.16
2.06
2.46
2.73
SET,
NOX
2.37
2.23
2.00
2.30
2.29
S02W
0.023
0.059
0.085
0.071
0.086
g/km
S02W
0.019
0.034
0.026
0.035
0.034
g/km
S02^
0.033
0.068
0.046
0.045
0.056
S04!avb)
0.009
0.013
0.009
0.013
0.014
S04=(b)
0.024
0.028
0.034
0.042
0.055
S04=(b)
0.028
0.022
0.033
0.018
0.026
% Fuel
Sulfur As
20.6
54.9
79.
67,
«6:4
% Fuel
Sulfur As
% Fuel
Sulfur As
(a) SC>2 - bubbler method
(b) S04= - particulate filter
-------
- 131 -
APPENDIX E
TABLE E-l (Cont.)
Idle, g/Test
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
I
II
I
II
I
II
I
II
I
II
I
II
I
II
I
II
I
II
I
II
CO
0.21
0.22
0.94
0.46
0.20
CO
0.04
0.04
0.07
0.06
0.05
0.05
0.08
0.07
0.06
0.06
CO
0.06
0.05
0.08
0.11
0.09
0.09
0.09
0.10
0.09
0.09
HC
0.48
0.77
0.12
0.99
0.61
HC
0.02
0.02
0.02
0.02
0.02
0.02
0.03
0.03
0.02
0.02
HC
0.02
0.02
0.02
0.02
0.03
0.02
0.03
0.03
0.04
0.04
NOX
2.63
2.17
2.00
2.24
2.12
64 km/h,
NOx
0.58
0.52
0.65
0.58
0.66
0.62
0.70
0.73
0.67
0.69
96 km/h,
NOX
3.34
3.40
2.65
2.53
3.37
3.29
4.16
4.42
3.76
3.81
S09W
0.112
0.115
0.113
0.117
0.114
g/km
S02(a?
0.021
-------
- 132 -
APPENDIX E
TABLE E-2
1975 VEHICLE (351 CID)
Pt/Rh OXIDATION CATALYSTS (AEW 2/3C/40/M)
(Fuel Sulfur - 0.032 Wt %)
1975 FTP, g/km
% Fuel Sulfur As
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12^800
CO
3.98(c)
2.38(c)
3.46(c)
2.61(c)
2.90(c)
CO
0.13
0.26
0.24
0.21
0.32
CO
1.33
1.61
1.13
0.85
0.98
HC
0.62
0.28
0.28
0.26
0.42
HC
0.04
0.05
0.04
0.05
0.05
HC
0.04
0.05
0.04
0.05
0.08
NOX
1.70
1.66
1.70
1.77
1.57
FET, g/km
NOX
3.03
2.73
2.36
2.82
2.73
SET, g/km
NOX
2.27
2.42
2.52
2.63
2.67
S0,
0.066
0.050
0.114
0.058
0.118
S02(S)
0.038
0.028
0.040
0.036
0.056
S02<*>
0.081
0.094
0.083
0.053
0.085
SOA-0>)
0.004
0.005
0.005
0.009
0.007
so4-(b)
0.005
0.023
0.028
0.029
0.020
so*-(b)
0.004
0.008
0.012
0.017
0.009
S02
51.3
40.3
92.2.
48.5
93.7
% Fuel
S02
45.6
31.8
51.3
45.0
68.5
% Fuel
S02
86.5
102.2
92.2
61.2
88.5
SO*"
1.9
2.4
2.7
5.0
3.7
Sulfur As
S04"
4.3
17.7
23.9
24.0
16.2
Sulfur AA
SOA"
2.7
6.1
8.8
13.2
6.2
(a) S02 - bubbler method
(b) S04* - particulate-filter
(c) One to two stalls on cold start
-------
- 133 -
APPENDIX E
TABLE 2 (Cont.)
Idle, g/test
% Fuel Sulfur As
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
3
6
9
12
0
200
400
600
800
I
II
I
II
I
II
I
II
I
0.10
0.06
0.09
0.07
0.10
0.11
0.12
0.12
0.17
II
Kilometres
3
6
9
12
0
200
400
600
800
I
II
I
II
I
II
I
II
I
II
0.10
0.09
0.13
0.12
0.16
0.16
0.19
0.19
0.26
0.26
CO
0.81
0.06
0.27
0.06
0.19
SO
0.10
0.06
0.09
0.07
0.10
0.11
0.12
0.12
0.17
0.14
SO
0.10
0.09
0.13
0.12
0.16
0.16
0.19
0.19
0.26
0.26
HC
0.57
0.29
0.30
0.43
0.51
HC
0.03
0.02
::
0.02
0.01
0.01
0.01
0.02
0.02
HC
0.04
0.04
0.02
0.02
0.02
0.02
0.03
0.03
0.03
0.03
NO*
0.61
2.38
2.16
2.33
1.88
64 km/h,
NOX
0.96
0.87
0.84
0.78
0.77
0.74
0.71
0.63
0.83
0.76
96 km/h,
NOjt;
5.08
5.26
6.49
6.77
5.79
6.05
6.49
6.58
7.76
8.12
S02(a)
0
0
0.107
0
0
8/km
S02(a)
0.0174(c)
0.0267
0.038
0.038
0.044
g/km
fa)
S02W
0.048(c)
0.060
0.056
0.056
0.071
S04°(b)
0
0.013
0.008
0.009
0.017
S04=(b)
0.008(d)
0.021
0.026
0.036
0.025
0.028
0.026
0.026
0.024
0.023
fb}
S04=
0.035(d)
0.027
0.047
0.032
0.035
0.027
0.039
0.027
0.039
0.022
SO 2
0
0
16.6
0
0
% Fuel
S02
21.8
34.2
48.1
48.0
55.3
% Fuel
S02
56.8
69.5
66.2
68.8
77.6
S04=
0
1.3
0.8
0.9
1.7
Sulfur As
SPA"
6.6
17.8
21.9
30.7
21.2
23.7
22.2
22.8
20.0
19.4
Sulfur As
504°
27.4
20.7
36.3
24.1
27.5
20.7
31.7
21.9
28.2
1C O
(a) S02 - bubbler method
(b) S04= - particulate-filter
(c) One sample taken over the 1 hour test
(d) One sample taken for 20 minutes during each half hour segment
-------
- 134 -
APPENDIX E
TABLE E-3
1975 VEHICLE (351 CID)
Pt OXIDATION CATALYSTS (AEW 2/12C/40/M)
(Fuel Sulfur = 0.032 Wt %)
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9. 600
12 800
1975 FTP, g/km
% Fuel
Sulfur As
CO
2.38
1.79
1.95
2.33
2.96
CO
0.16
0.72
0.38
0.82
0.51
CO
0.46
0.56
0.53
0.27
0.70
HC
0.36
0.26
0.26
0.34
0.43
HC
0.03
0.10
0.05
0.11
0.10
HC
0.05
0.12
0.11
0.11
0.08
NOX
1.89
1.59
1.76
1.87
1.81
FET,
NOX
2.62
2.81
2.29
2.50
2.70
SET,
NOX
2.39
2.57
2.36
2.40
2.45
S02^
0.087
0.056
0.058
0.042
0.069
g/km
S02(a)
0.067
0.033
0.033
0.028
0.045
g/km
S02('a)
0.064
0.032
0.060
0.043
0.074
S0£=(b)
0.006
0.011
0.016
0.017
0.021
SO£=W
0.009
0.033
0.038
0.040
0.064
SOA=^
0.010
0.037
0.029
0.043
0.031
% Fuel
Sulfur As
% Fuel
Sulfur As
26.0
(a) S02 - bubbler method
(b) S04= - particulate filter
-------
- 135 -
APPENDIX E
TABLE E-3 (cont.)
Idle, g/test
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
I
II
I
II
I
II
I
II
I
II
Kilometres
0
3 200
6 400
9 600
12 800
I
II
I
II
I
II
I
II
I
II
CO
0.41
0.21
0.20
0.21
0.38
CO
0.06
0.06
0.05
0.05
0.04
0.05
0.06
0.07
0.06
0.06
CO
0.06
0.06
0.06
0.06
0.05
0.07
0.05
0.05
0.08
0.11
HC
0.19
0.70
0.55
1.10
0.76
HC
0.01
0.01
0.02
0.01
0.02
0.03
0.03
0.03
0.02
0.02
HC
0.01
0.02
0.03
0.03
0.03
0.03
0.03
0.02
0.02
0.02
NOX
1.94
2.47
2.52
2.38
1.68
64 km/h,
NOX
0.68
0.68
0.59
0.45
0.61
0.57
0.72
0.66
o:.67
0.66
96 km/h,
NOX
4.74
4.74
4.72
4.28
4.21
4.46
3.65
3.59
3.91
3.66
S02(a)
0.371
0.117
0.112
0.110
0.113
g/km
S02(a)
0.030(c)
0.034
0.022
0.022
0.028
g/km
S02(a)
.064(c)
.038
.036
.029
.051
S04=(b)
0.019
0.028
0.021
0.016
0.018
S04=(b)
0.036(d)
0.043
0.061
0.043
0.51
0.52
0.059
0.060
0.059
0.055
SOA=
-------
APPENDIX F
TABLE F-l
FUEL CONSUMPTION, (g/km)
1975 CAR, Pt-Pd (6-C) OXIDATION CATALYSTS
Kilometres
0 Weight
Carbon Balance
3 200 Weight
Carbon Balance
6 400 Weight
Carbon Balance
9 600 Weight
Carbon Balance
12 800 Weight
Carbon Balance
FTP
178.0
171.3
172.9
167.8
169.8
167.2
169.8
164.7
176.0
155.4
FET
104.1
110.4
109.0
104.7
104.4
105.2
107.9
104.9
111.7
100.0
SET
119.6
121.9
125.4
124.6
123.6
121.3
116.7
116.5
124.0
108.8
, 64 km/h 96 km/h
IdleU} I II I II
885.3 104.2 101.6
1026.0 107.0 104.9 102.1 102.5
858.1 111.5 106.0
1021.5 108.7 108.1 105.4 107.0 ,
t-»
OJ
921.6 109.1 108.3 7
976,1 109.8 109.8 106.8 108.3
858.1 109.5 108.9
948.9 104.9 104.9 103.0 110.7
844.4 110.7 110.3
858.1 97.1 98.2 94.6 93.7
(a) Fuel consumption at idle is reported in grams/test.
-------
Kilometres
3 200
6 400
9 600
12 800
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
APPENDIX F
TABLE F-2
FUEL CONSUMPTION, (g/km)
1975 CAR, Pt-Rh (3-C) OXIDATION CATALYSTS
FTP
177.5
185.4
178.5
181.9
185.4
178.3
171.3
174.7
193.7
184.7
FET
119.9
121.0
124.8
126.4
115.3
111.7
115.3
115.3
115.3
118.5
SET
130.0
136.4
132.5
133.9
129.1
130.8
127.3
126.0
137.0
141.4
Idle'1'
549.3
626.5
908.0
1017.0
858.1
998.8
830.8
1021.5
876.2
998.8
64 km/h 96 km/h
I II I II
113.7 120.1
116.6 117.4 124.1 129.8
112.7 126.6
114.0 114.7 127.8 130.5
111.5 125.2
113.8 115.1 123.7 127.8
111.1 126.8
114.0 111.7 120.9 127.2
114.6 130.6
117.6 115.9 135.7 141.2
(1) Fuel consumption at idle is reported in grams/test.
-------
Kilometres
3 200
6 400
9 600
12 800
Fuel Weight
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
APPENDIX F
TABLE F-3
1975
FTP
188.8
181.9
165.4
167.5
164.2
173.7
175.4
168.3
178.8
171.1
FUEL CONSUMPTION (g/km)
CAR, Pt (12C) OXIDATION CATALYSTS
FET
115.0
112.8
113.4
117.4
108.7
108.2
111.2
106.6
113.6
107.7
SET
129.1
126.9
126,0
130.4
123.1
124.2
122.9
123.8
125.4
125.6
m 64 km/h 96 km/h
IdleU; I II I II
871.7 110.4 120.0
1007.9 110.6 111.3 118.5 117.5
926.2 109.2 120.0
1019.0 110.0 105.8 117.8 115.7
826.3 108.7 113.9
1007.9 109.0 108.7 123.7 126.5
830.8 108.2 104.5
967.0 110.4 109.2 104.8 107.5
840.0 110.0 108.5
935.2 78.9 110.0 107.8 111.1
i
oo
I
(1) Fuel consumption at idle is reported in grams/test.
-------
Kilometres
APPENDIX G
MAXIMUM
TABLE
G-l
EXHAUST GAS TEMPERATURE AT CATALYST OUTLET
Pt-Pd (AEW 2/6C/50M)
Temperature, °C
FTP
632
610
604
596
602
FET SET
554 607
582 613
554 571
568 577
582 604
Idle 64 km/h
382 552
391 554
346 541
374 532
377 532
96 km/h
571
607
593
577
577
I
3 200
Co
6 400
9 600
12 800
-------
APPENDIX G
TABLE G-2
MAXIMUM EXHAUST GAS TEMPERATURE AT
Pt-Rh (AEW 2/3C/40M)
CATALYST OUTLET
Temperature, °C
Kilometres FTP FET SET
0 682 638 671
3 200 649 593 671
6 400 638 621 671
9 600 627 616 616
12 800 649 . 604 649
Idle 64 km/h 96 km/h
293 582 660
377 582 660
382 582 654
371 560 649
360 571 660
I
M
O
I
-------
Kilometres
0
3 200
6 400
9 600
12 800
MAXIMUM
APPENDIX G
TABLE G-3
EXHAUST GAS TEMPERATURE AT
Pt (AEW 2/12C/40M)
CATALYST OUTLET
Temperature, °C
FTP
646
635
602
593
604
FET SET
627 654
638 582
571 616
577 577
582 616
Idle 64 km/h
343 582
388 571
371 554
377 537
396 551
96 km/h
652
593
593
577
627
-------
- 142 -
APPENDIX H
TABLE H-l
1975 VEHICLE (351 CID)
Pt/Pd OXIDATION CATALYST* (AEW 2/6C/50/M)
1975 FTP, g/km
% Fuel Sulfur As
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
0
3 200
6 400
9 600
12 800
CO
2.01
8.45(c)
9.08(c)
ll.Ol(c)
5.87(c)
CO
0.93
1.71
2.32
3.03
1.81
CO
2.82
3.10
2.91
3.74
1.91
HC
0.11
0.44
0.37
0.73
1.45
HC
0.07
0.12
0.15
0.19
0.11
HC
0.095
0.14
0.18
0.17
0.21
NOX
y
1.70
1.50
1.60
1.78
1.02
FET, g/km
NOx
2.82
2.29
2.30
1.86
2.06
SET, g/km
NOx
1.91
2.81
2.23
1.96
1.81
S02(a)
0.110
0.098
0.110
0.093
0.084
(a)
S02
0.133
0.071
0.079
0.073
0.057
S02(a)
0.119
0.091
0.088
0.091
0.068
so^
0.012
0.003
0.004
0.004
0.006
_/i-\
SQ4
0.013
0.006
0.004
0.005
0.007
SOA-M
0.007
0.004
0.003
0.004
0.007
S02
98.2
89.0
100.0
81.6
77.8
% Fuel
S02
180.0
98.6
109.7
117.7
79.2
% Fuel
SO 2
141.6
116.7
110.0
126.4
89.5
S04=
7.1
1.6
2.4
2.3
3.4
Sulfur As
SO 4°
11.6
5.2
3.7
5.4
6.5
Sulfur As
S0&=
5.6
3.2
2.5
3.7
5.9
(a) S02 - bubbler method
(b) S04= - particulate-filter
(c) 2 stalls on cold start
Visual inspection of catalysts at 12 800 km showed that one converter had
been burned up.
-------
- 143 -
APPENDIX H
TABLE H-2
1975 VEHICLE (351 CID)
Pt/Pd OXIDATION CATALYST* (AEW 2/6C/50/M)
Idle, g/test
% Fuel Sulfur As
Kilometres
0
3 200
6 400
9 600
12 800
Kilometres
3 200
6 400
9 600
12 800
I
II
I
II
I
II
I
II
I
II
Kilometres
3 200
6 400
9 600
12 800
I
II
I
II
I
II
I
II
I
II
CO
0.17
5.79
7.66
25.82
42.29
CO
0.75
0.06
0.80
0.79
1.43
1.42
2.64
3.74
1.44
1.60
CO
0.08
0.08
1.11
1.02
2.17
2.27
3.65
4.04
1.74
1.87
HC
0.60
1.97
2.71
4.75
9.96
HC
0.02
0.01
0.04
0.04
0.06
0.05
0.07
0.07
0.06
0.03
HC
0.05
0.01
0.05
0.04
0.08
0.07
0.05
0.05
0.14
0.11
NOX
0.80
2.61
3.02
0.66
0.04
64 km/h,
NOX
0.46
0.66
0.69
0.68
0.74
0.72
0.84
0.79
0.86
0.58
96 km/h,
NOx
2.81
2.76
5.17
5.11
4.18
4.61
2.89
3.13
2.20
2.31
S0,(a)
0.666
0.491
0.502
0.260
0.179
g/km
/ \
S02
0.034(c)
0.067
0.080
0.067
0.059
g/km
S02(a)
0.056(c)
0.084
0.092
0.068
0.053
so4=(b)
0.046
0.010
0.012
0.032
0.022
-(b)
S04~^ '
0.013(d)
0.014
0.006
0.010
0.003
0.005
0.005
0.005
0.005
0.007
S04=(b)
0.040(d)
0.037
0.018
0.013
0.008
0.007
0.010
0.007
0.008
0.007
S02
155.6
73.5
74.7
73.0
52.0
% Fuel
S02
48.6
98.5
117.6
104.7
84.3
% Fuel
S02
80.0
110.5
131.4
113.3
85.5
S04=
7.2
1.0
1.1
5.9
4.2
Sulfur As
S04=
12.4
11.9
5.9
9.8
2.9
4.9
5.2
5.2
4.8
6.0
Sulfur As
S04=
38.3
35.0
15.8
12.0
7.6
6.7
10.5
7.6
8.9
7.3
(a) S02 - bubbler method
(b) S04= - particulate-filter
(c) One sample taken over 1 hour test
(d) One sample taken for 20 minutes during each half hour segment
* Visual inspection of catalysts at 12 800 km showed one converter had been burned up,
-------
Kilometres
3 200
6 400
9 600
12 800
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
Fuel Weight
Carbon Balance
APPENDIX H
TABLE H-3
FUEL CONSUMPTION, (g/km) ...
1975 CAR, Pt-Pd (6-C) OXIDATION CATALYSTS U'
FTP
177.2
176.2
181.3
171.6
179.0
171.9
189.0
188.8
181.3
178.3
FET
122.9
116.4
123.2
112.0
119.6
111.7
113.1
101.9
116.4
119.1
SET
141.6
131.0
137.0
123.8
133.5
126.5
130.4
121.5
128.1
128.1
Idle<2>
599.3
667.4
935.2
1044.2
908.0
1048.7
553.9
594.7
558.4
572.0
64 km/h 96 km/h
I II I . II
118.0 118.3
109.8 109.2 108.2 111.0
112.7 128.56
104.9 104.5 119.5 115.9
115.4 120.2
106.4 105.6 111.0 109.0
119.8 119.1
106.0 109.0 99.7 105.2
116.5 122.0
118.5 121.4 102.3 107.9
I
I
I
(1) Visual inspection of oxidation catalysts at 12 800 km showed that one converter had been burned up.
(2) Fuel consumption at idle is reported in grams/test.
-------
Kilometres
MAXIMUM
APPENDIX H
TABLE H-4
EXHAUST GAS TEMPERATURE AT
Pt-Pd (AEW 2/6C/50M)
CATALYST OUTLET
A
Temperature, °C
FTP
638
560
549
549
549
FET SET
604 660
571 571
571 571
571 582
527 549
Idle 64 km/h
449 571
316 510
327 516
232 549
216 527
96 km/h
632
604
582
604
560
3 200 560 571 571 316 510 604 *;
6 400
9 600
12 800
*Visual inspection of catalysts at 12 800 km showed that one converter had been burned up.
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
2.
3. RECIPIENT'S ACCESSION>NO.
4. TITLE AND SUBTITLE
An Assessment of (Automotive) Sulfate Emission
Control Technology
5. REPORT DATE
6/77
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
M. G. Griffith, R. A. Bouffard, E. L. Holt,
M. W. Pepper, M. Beltzer
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORG •\NIZATION NAME AND ADDRESS
Exxon Research and Engineering Co.
Linden Ave.
Linden, New Jersey
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-03-0497
12. SPONSORING AGENCY NAME AND ADDRESS
EPA-OAWM-MSAPC
2565 Plymouth Road
Ann Arbor, MI 48105
13. TYPE OF REPORT AND PERIOD COVERED
Final report
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This contfact consists of three parts. Modification I was aimed at develop-
ing sorbents for 804* capable of efficiently removing this material from exhaust gas at
low pressure drop over a long service life. Modification II studied three pairs of
monolithic automotive oxidation catalysts, nominally identical except for their noble
metal components, to determine the effect of composition on the level of 804** emissions.
Modification III tested two types of experimental perovskite-based noble metal auto-
motive oxidation catalysts for CO, HC, and 802 conversion activity.
In I, a mixture of 85 wt % CaO/10% Si02/5% Na20 was fabricated into 12.7 mm O.D.
rings and 4.8 mm pellets. A mixture of 85 wt% Ca(OH)2/15% Si02 was prepared as 4.8 mm
pellets. These three preparations were tested in canisters mounted in the muffler posi-
tiqn of oxidation catalyst-equipped vehicles. All three had high initial efficiency for
804° sorption, but fell off rapidly within 15 000 km of operation to less than 90%.
Modification II showed the following levels of 804** tailpipe emissions when the
catalysts were tested on a vehicle over the EPA Sulfate Emission Test using fuel of 320
ppm sulfur: Pt, 30 mg/km; Pt-Pd, 25 mg/km; Pt-Rh, 10 mg/km. In Modification III, a
Pt-Perovskite beaded catalyst was compared to a commercial General Motors Pt-Pd oxidatio
catalyst under normal and elevated temperature operating conditions. In both cases, the
commercial catalyst emitted lower levels of CO and HC, as well as S04=. A monolithic
version of the Pt-Perovskite was also compared to a commercial monolithic catalyst and
found to be inferior .in CO and HC activity.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS
c. COSATI Field/Group
Sulfates
Sulfuric Acid
Catalytic Converters
Exhaust Oxygen Level
8. DISTRIBUTION STATEMENT
Not Restricted
19. SECURITY CLASS (This Report)
Unclassified
21. NO. OF PAGES
20. SECURITY CLASS (Thispage)
TTnrlaccHf if A
22. PRICE
EPA Form 2220-1 (9-73)
------- |