EPA-R2-72-106
Nnvember 1972 ENVIRONMENTAL PROTECTION TECHNOLOGY
^valuation of Monitoring Methods
and Instrumentation for Hydrocarbons
and Carbon Monoxide
in Stationary Source Emissions
Office of Research and Monitoring
U.S. Environmental Protection Agoncy
Washington, D.C. 20480
-------
EPA-R2-72-106
Evaluation of Monitoring
Methods and Instrumentation
for Hydrocarbons
and Carbon Monoxide
in Stationary Source Emission
by
J. Driscoll, J. Becker
J. McCoy, M. Young
and J. Ehrenfeld
Wai den Research Corporation
359 Allston Street
Cambridge, Massachusetts
Contract No. 68-02-0320
Program Element No. A11010
Project Officer: Fredric C. Jaye
Division of Chemistry and Physics
National Environmental Research Center
Research Triangle Park, N. C. 27711
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION. AGENCY
WASHINGTON D.C. 20460
November 1972
-------
This report has been reviewed by the Environmental Protection
Agency and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the Agency,
nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
TABLE OF CONTENTS
Section Title Page
1 INTRODUCTION 1-1
2 CHARACTERISTICS OF HYDROCARBON AND CARBON MONOXIDE
EMISSION SOURCES 2-1
2.1 Mechanisms Producing HC and CO 2-1
2.2 Emissions from Fuel Combustion in Boilers 2-5
2.3 Municipal Incinerators 2-6
2.4 Gray Iron Foundries 2-15
2.5 Refineries 2-18
2.6 Asphalt Batching Plants 2-19
3 MANUAL METHODS FOR THE DETERMINATION OF HYDROCARBONS
AND CARBON MONOXIDE 3-1
3.1 Introduction 3-1
3.2 Hydrocarbon Analysis 3-1
3.3 Carbon Monoxide Analysis 3-8
4 STATE-OF-THE-ART SURVEY FOR HYDROCARBON AND CARBON
MONOXIDE MONITORING INSTRUMENTATION 4-1
4.1 Introduction 4-1
4.2 Hydrocarbon Monitors 4-1
4.3 Carbon Monoxide Monitors 4-20
5 LABORATORY AND PILOT PLANT STUDIES 5-1
5.1 Introduction 5-1
5.2 Pilot Plant Description and Gas Doping System 5-1
5.3 Laboratory Determination of Instrument Response
Curves 5-8
5.4 Sample Conditioning System for Pilot Plant
Measurements 5-15
5.5 Pilot Plant Accuracy Studies for CO and HC
Analyzers 5-21
5.6 Pilot Plant Comparisons of CO and HC Levels 5-41
6 FIELD EVALUATIONS OF THE CONTINUOUS MONITORS 6-1
6.1 Sampling System and Trailer 6-2
6.2 Incinerator Tests 6-6
6.3 Gray Iron Foundry 6-41
6.4 Refinery Tests 6-99
6.5 Zero and Span Dri ft 6-129
WALDEN RESEARCH CORPORATION
-------
TABLE OF CONTENTS (Cont.)
Section
7
8
9
APPENDIX I
APPENDIX II
Title
CONCLUSIONS
RECOMMENDATIONS
REFERENCES
TRANSLATION OF CARBON MONOXIDE PAPER
COMPUTER PROGRAM FOR MONITOR CALCULATIONS AND
STATISTICAL ANALYSIS
Page
7-1
8-1
9-1
1-1
II-l
Table No.
LIST OF TABLES
Title
Page
2-1 Carbon Monoxide Emission Estimates by Source Category -
1968 2-2
2-2 Estimates of Hydrocarbon Emissions by Source Category -
1968 2-3
2-3 Size Categories for Stationary Combustion Sources 2-6
2-4 Emission Parameters for Fossil Fuel Combustion 2-7
2-5 Range of Concentration and Emission of Carbon Monoxide
and Hydrocarbons from Combustion Devices 2-8
2-6 Estimated Incinerator Emissions 2-10
2-7 Assumed Nominal Concentrations of Gaseous Pollutants at
Entrance to APC Device 2-12
2-8 Incinerator Emission Factor Summary 2-14
2-9 Emissions of Air Contaminants from Catalytic Cracking
Unit Regenerator Stacks 2-20
3-1 Columns and Conditions for Separation of Aromatic
Olefinic and Aliphatic Hydrocarbons 3-5
3-2 Columns and Conditions for the Separation of Organic
Acids by Gas Chromatography 3-6
3-3 Columns and Conditions for the Separation of Aldehydes
by Gas Chromatography 3-7
3-4 Hydrocarbon Analysis by Gas Chromatography 3-8
3-5 Preparation of Calibration Curve for Pd(o-phen)2Cl2
Method 3-16
iv
WALDEN RESEARCH CORPORATION
-------
LIST OF TABLES (Cent.)
Table No. Title Page
3-6 Sample Size as a Function of CO Concentrations 3-16
3-7 Analysis of CO in Flue Gases Using CO-Pd(o-phen)2Cl2
Method :..:... 3-19
4-1 Specifications of Some Commercially Available Hydro-
carbon Monitors 4-2
4-2 Specifications of Some Commercially Available Carbon
Monoxide Monitors 4-4
4-3 Characteristics of a Flame Ipnization Detector 4-7
4-4 Reported Hydrocarbon Sensitivities with a Dupont 400
Analyzer 4-15
5-1 Pilot Plant Specifications 5-6
5-2 Characteristics of the Pilot Plant on #2 Oil and Gas .. 5-7
5-3 Preliminary Evaluation of the Ethane Doping System 5-7
5-4 Instruments Included in the Study 5-14
5-5 Accuracy Studies for CO Analyzers on Pilot Plant Flue
Gas, 4/11/72 5-22
5-6 Accuracy Studies for CO Analyzers on Pilot Plant Flue
Gas, 4/12/72 5-24
5-7 Accuracy Studies of CO Analyzers (14 April 1972) 5-26
5-8 Calculation of CO Concentrations on Beckman 315 Using
the Non-Linear Calibration Curve 5-28
5-9 Accuracy Studies for Hydrocarbon Analyzers on Pilot
Plant Flue Gas, 4/11/72 5-30
5-10 Accuracy Studies for Hydrocarbon Analyzers on Pilot
Plant Flue Gas, 4/12/72 5-32
5-11 Accuracy Studies in Pilot Plant 5-35
5-12 Statistical Summary - Pilot Plant Data - Hydrocarbons -
12 April 1972 5-36
5-13 Statistical Summary - Calibration Gas Runs - Hydro-
carbons - 19 Apri 1 1972 5-37
5-14 Statistical Summary - Pilot Plant Data - Carbon
Monoxide - 12 April 1972 5-38
5-15 Statistical Summary - Calibration Gas Runs - Carbon
Monoxide - 19 April 1972 5-39
5-16 Compilation of Data Acquisition Channel Numbers of
Instruments 5-40
5-17 Statistical Summary - Carbon Monoxide - Doped Flue
Gas - 12 April 1972 5-42
WALDEN RESEARCH CORPORATION
-------
LIST OF TABLES (Cont.)
Table No. Title Page
5-18
5-19
5-20
5-21
5-22
5-23
5-24
5-25
5-26
5-27
5-28
5-29
5-30
5-31
5-32
5-33
5-34
5-35
5-36
5-37
5-38
6-1
Comparison of Carbon Monoxide Levels for #4 Fuel Oil
in Pilot Plants
Comparison of HC Levels for #4 Fuel Oil in Pilot
Plants
Pilot Plant Data - Carbon Monoxide - 24 February
1972
Statistical Summary - Pilot Plant Data - Carbon
Monoxide - 24 February 1972
Pilot Plant Data - Hydrocarbons - 24 February 1972
Statistical Analysis of Pilot Plant Data - Hydrocarbon
Monitors (24 February 1972)
Pilot Plant Data - Carbon Monoxide - 25 February
1 972
Statistical Summary - Pilot Plant Data - Carbon
Monoxide - 25 February 1972
Pilot Plant Data - Hydrocarbons - 2 March 1972
Pilot Plant Data - Carbon Monoxide - 2 March 1972
Statistical Summary - Pilot Plant Data - Hydrocarbons -
2 March 1972
Statistical Summary - Pilot Plant Data - Carbon
Monoxide - 2 March 1 972
Pilot Plant Data - Hydrocarbons - 28 March 1972
Statistical Summary - Pilot Plant Data - Hydrocarbons -
28 March 1972
Regression Equation Constants - Hydrocarbon Monitors -
28 March 1972
Pilot Plant Data - Carbon Monoxide - 28 March 1972
Statistical Summary - Pilot Plant Data - Carbon
Monoxide - 28 March 1972
Pilot Plant Data - Carbon Monoxide - 13 April 1972
Statistical Summary - Pilot Plant Data - Carbon
Monoxide - 13 April 1972
Pilot Plant Data - Carbon Monoxide - 14 April 1972
Statistical Summary - Pilot Plant Data - Carbon
Monoxide - 14 April 1972
Incinerator Calibrations for Monitors
5-44
5-44
5-45
5-47
5-48
5-49
5-51
5-53
5-54
5-55
5-57
5-58
5-59
5-60
5-61
5-63
5-64
5-65
5-66
5-67
5-69
6-9
.j WALDEN RESEARCH CORPORATION
-------
LIST OF TABLES (Cont.)
Table No. Title Page
6-2
6-3
6-4
6-5
6-6
6-7
6-8
6-9
6-10
6-11
6-12
6-13
6-14
6-15
6-16
6-17
6-18
6-19
6-20
6-21
Statistical Summary - Incinerator - Carbon Monoxide -
26 April 1972
Statistical Summary - Incinerator - Carbon Monoxide -
27 April 1972
Statistical Summary - Incinerator - Carbon Monoxide -
2 May 1972
Statistical Summary - Incinerator - Hydrocarbons -
3 May 1972
Statistical Summary - Incinerator - Carbon Monoxide -
3 May 1972
Incinerator - Hydrocarbons - 4 May 1972
Statistical Summary - Incinerator - Hydrocarbons -
4 May 1972
Incinerator - Carbon Monoxide - 4 May 1972
Statistical Summary - Incinerator - Carbon Monoxide -
4 May 1972
Statistical Summary - Incinerator - Hydrocarbons -
9 May 1972
Statistical Summary - Incinerator - Carbon Monoxide -
9 May 1972
Statistical Summary - Incinerator - Hydrocarbons -
10 May 1972
Statistical Summary - Incinerator - Carbon Monoxide -
10 May 1972
Statistical Summary - Incinerator - Hydrocarbons -
11 May 1972
Statistical Summary - Incinerator - Carbon Monoxide -
11 May 1972
Statistical Summary - Incinerator - Hydrocarbons -
17 May 1972
Statistical Summary - Incinerator - Carbon Monoxide -
17 May 1972
Comparison of CO Concentrations by Wet Chemical and
Instrumental Methods
Condensible Organic Matter Collected in Water Trap
Preliminary Results with a Selective Combustor Coupled
to a Flame lonization Detector
6-11
6-12
6-14
6-15
6-17
6-18
6-22
6-23
6-27
6-28
6-30
6-31
6-32
6-33
6-34
6-36
6-37
6-39
6-40
6-43
vi -j WALDEN RESEARCH CORPORATION
-------
LIST OF TABLES (Cont.)
Table No. Title Page
6-22
6-23
6-24
6-25
6-26
6-27
6-28
6-29
6-30
6-31
6-32
6-33
6-34
6-35
6-36
6-37
6-38
6-39
6-40
6-41
6-42
6-43
6-44
6-45
6-46
6-47
6-48
Foundry and Refinery (CO Boiler Inlet) Calibrations for
CO and HC Monitors
Foundry Data - Hydrocarbons - 25 May 1 972
Foundry Data - Carbon Monoxide - 25 May 1972
Statistical Summary - 25 May 1972
Foundry Data - Hydrocarbons - 26 May 1972
Statistical Summary - Foundry Data - Hydrocarbons -
26 May 1972 ,
Foundry Data - Carbon Monoxide - 26 May 1972
Statistical Summary - Foundry Data - Carbon Monoxide -
26 May 1972
Foundry - 30 May 1 972
Foundry Data - Hydrocarbons - 1 June 1972
Statistical Summary - Foundry Data - Hydrocarbons -
1 June 1972
Foundry Data - Carbon Monoxide - 1 June 1972
Statistical Summary - Foundry Data - Carbon Monoxide -
1 June 1972
Foundry Data - Hydrocarbons - 2 June 1972
Statistical Summary - Foundry Data - Hydrocarbons -
2 June 1972
Foundry Data - Carbon Monoxide - 2 June 1972
Statistical Summary - Foundry Data - Carbon Monoxide -
2 June 1972
Foundry - 5 June 1 972
Foundry - 6 June 1 972
Foundry - 7 June 1 972
Foundry - 8 June 1 972
Foundry - 9 June 1 972
Stati sti cal Summary - 5 June 1 972
Statistical Summary - 6 June 1972
Statistical Summary - 7 June 1972
Statistical Summary - 8 June 1972
Statistical Summary - 9 June 1972
6-50
6-51
6-55
6-58
6-59
6-62
6-63
6-65
6-67
6-68
6-70
6-71
6-73
6-74
6-76
6-77
6-79
6-81
6-83
6-84
6-86
6-88
6-89
6-90
6-91
6-92
6-93
viii
WALDEN RESEARCH CORPORATION
-------
LIST OF TABLES (Cont.)
Table No. Title Page
6-49 Comparison of Hydrocarbon Values by Manual Gas Chromat-
ographic Analysis and Continuous Hydrocarbon Monitors
for a Gray Iron Foundry 6-95
6-50 Comparison of CO Determinations by Wet Chemistry and
Continuous Monitors at a Gray Iron Foundry 6-96
6-51 Statistical Summary - Wet Chemical vs Instruments 6-97
6-52 Refinery (Inlet to CO Boiler) - 29 June 1972 6-101
6-53 Refinery (Inlet to CO Boiler) - 30 June 1972 6-103
6-54 Refinery (Inlet to CO Boiler) - 3 July 1972 6-104
6-55 Refinery (Inlet to CO Boiler) - 4 July 1972 6-106
6-56 Refinery (Inlet to CO Boiler) - 5 July 1972 6-108
6-57 Refinery (Inlet to CO Boiler) - 6 July 1972 6-109
6-58 Statistical Summary - Refinery HC Data 6-111
6-59 Statistical Summary - Refinery CO Data 6-113
6-60 Comparison of Hydrocarbon Values by Manual Gas Chro-
matographic Analysis and Continuous Hydrocarbon Moni-
tors for Petroleum Refinery (Catalytic Cracking Unit -
Inlet to CO Boiler) 6-115
6-61 Comparison of CO Determinations by Wet Chemistry and
Continuous Monitors at a Catalytic Cracking Unit (CO
Boiler Inlet) 6-116
6-62 Hydrocarbon Concentrations at the Outlet of the CO
Boiler 6-118
6-63 Calibration Equations for CO Monitors at Outlet of CO
Boiler 6-119
6-64 Refinery (Outlet of CO Boiler) - 7 July 1972 6-120
6-65 Refinery (Outlet of CO Boiler) - 10 July 1972 6-121
6-66 Refinery (Outlet of CO Boiler) - 11 July 1972 6-123
6-67 Refinery (Outlet of CO Boiler) - 12 July 1972 6-125
6-68 Refinery (Outlet of CO Boiler) - 13 July 1972 6-127
6-69 Statistical Summary - Refinery Outlet CO Data 6-130
6-70 Hydrocarbon Monitors - Drift Measurements 6-131
6-71 CO Monitor - Zero Drift 6-132
6-72 CO Monitors - Span Drift 6-133
6-73 Stability Measurements 6-134
ix WALDEN RESEARCH CORPORATION
-------
LIST OF FIGURES
Figure No. Caption Page
2-1 Projections of Future Active Incinerator Capacity 2-9
2-2 Cupola 2-16
2-3 The Course of Gas Composition and Temperature in the
Cupola Shaft (schematic) 2-17
2-4 Cylindrical, Water-Cooled, Carbon Monoxide Waste-Heat
Boiler 2-21
2-5 Asphalt Batch Mix Plant 2-22
3-1 Separation of Organic Acids by G.L.C 3-4
3-2 Absorption Spectrum for [Pd2(CO)2(o-phen)2] 4H20 3-13
3-3 CO Collection and Reaction Flask 3-14
3-4 Calibration Curve for CO-Pd(phen)Cl2 Method 3-15
3-5 Collection and Transfer of CO Samples x 3-18
4-1 Schematic of a Flame lonization Detector 4-6
4-2 CO, CH, THC Monitor 4-9
4-3 Materials and Spectral Regions for Semiconductor Lasers .. 4-19
4-4 Schematic Diagram of NDIR Analyzers 4-22
4-5 Andros Carbon Monoxide Analyzer 4-26
5-1 Walden Pilot Plant 5-2
5-2 Pilot Plant Test Section Assembly 5-3
5-3 Pilot Plant Schematic 5-4
5-4 Intertech NDIR 5-9
5-5 Leeds & Northrup NDIR 5-10
5-6 Beckman 6800 Chromatograph 5-11
5-7 Beckman 315 NDIR 5-12
5-8 MSA LIRA 300 (NDIR) 5-13
5-9 Beckman 400 (FID) 5-16
5-10 Intertech URAS II (NDIR) 5-17
5-11 Beckman 6800 THC (Chromatograph) 5-18
5-12 Pilot Plant Flue Gas Sample Conditioning System for CO
and Hydrocarbon Analyzers 5-19
6-1 Sampling System for CO and HC Monitors 6-4
6-2 Sample Manifold and Analysis System 6-5
6-3 Schematic of Data Acquisition System 6-7
WALDEN RESEARCH CORPORATION
-------
LIST OF FIGURES (Cont.)
Figure No. Caption Page
6-4 Views of the Mobile Laboratory 6-8
6-5 Modification of Beckman 400 6-42
6-6 Schematic of Gray Iron Foundry 6-44
6-7 Calibration Curve for Beckman 6800 at Foundry 6-46
6-8 Calibration Curve for Beckman 315 (NDIR) at Foundry 6-47
6-9 Calibration Curve for L&N NDIR at Foundry 6-48
6-10 Calibration Curve for MSA NDIR at Foundry 6-49
6-11 Comparison of Fast Response (30 sec) and Slow Response
(5-10 min) HC Monitors on a Gray Iron Foundry - May 30,
1972 6-98
6-12 Variation of CO and HC with Time, Gray Iron Foundry,
May 30, 1972 6-100
6-13 Comparison of Instrument Zero and Span as Function of
Time 6-135
Xi WALDEN RESEARCH CORPORATION
-------
1. INTRODUCTION
This report was prepared under Contract No. 68-02-0320 covering the
period from July 1, 1971 to November 30, 1972. The principal objectives
of this program were (1) to review the state-of-the-art methods on moni-
toring instruments for carbon monoxide and hydrocarbons in stationary
sources, and (2) to prepare recommendations for short- and long-term de-
velopment programs for hydrocarbon and carbon monoxide monitors.
During the initial phase of the program, the types of monitors for CO
and HC were reviewed, selected, and acquired. Two different techniques
for CO were considered for evaluation, e.g., Non-Dispersive Infrared (NDIR)
and Process Chromatography [reduction of CO to CH. and analysis with a
flame ionization detector (FID)]. The techniques for hydrocarbon analysis
included FID, NDIR (ethane sensitized), and process Chromatography (FID de-
tection) which provides CH. and total HC (THC). The advantage of this
latter instrument is that "reactive" (non-methane) HC can be determined by
difference.
Laboratory tests on the analyzers were conducted in the Walden 500,000
Btu/hr pilot plant. The fuels used were natural gas as well as #2 and #4
fuel oil. During these tests, the determination of the accuracy, drift,
interferences, precision, etc., was conducted for CO and HC monitors. The
laboratory tests also included the development of wet chemical methods for
CO and HC analysis. These techniques would be used to check the accuracy
of the instruments in the field. A sampling system for the CO and HC
analyzers was assembled and tested.
For the field evaluation of the analyzers, a 20-foot trailer was rein-
forced and equipped with a sample manifold, data acquisition system, instru-
ment mountings and adequate 120V lines. The instruments were installed in
the trailer and field tests were conducted at:
(1) Municipal Incinerator
(2) Gray Iron Foundry
(3) Petroleum Refinery
1-1
WALDEN RESEARCH CORPORATION
-------
A computer program (Appendix II) was written for the reduction of the
data from this study. This converted the data acquisition system MV read-
ings to ppm CO or HC and provided statistical analysis including regres-
sion equations, correlation coefficients, and "t" tests. The results ob-
tained in the pilot plant (accuracy, etc.) and in the field are given in
Sections 5 and 6 respectively.
For each source which was tested, a brief write-up of the process is
given to enable the reader to understand the sampling problems associated
with each industry. This is followed by the laboratory pilot plant and
field studies. A brief description of recommendations and conclusions then
finishes the discussion.
1-2 WALDEN RESEARCH CORPORATION
-------
2. CHARACTERISTICS OF HYDROCARBON AND CARBON MONOXIDE EMISSION SOURCES
Carbon monoxide and hydrocarbons are emitted from a large variety
of stationary sources. Although this source group produces considerably
less emissions than the transportation category, it is significant, as
shown in Tables 2-1 and 2-2 (NAPCA, 1970 a and b). In this section, the
nature of CO and HC emissions is discussed to provide information to
assist in the selection of monitoring equipment. Discussion of emissions
from fuel burning in boilers, municipal incinerators, foundry cupolas,
refinery catalyst regenerators and asphalt batching plants is included.
2.1 MECHANISMS PRODUCING HC AND CO
Several mechanisms are believed to produce HC and CO emissions
from the sources listed above, and also from other stationary sources as
well. Hydrocarbons can be produced by partial combustion (or pyrolysis)
and evaporation. While combustion processes are usually designed to con-
sume all of the hydrocarbons by reacting with the oxygen in air, it is
possible to generate hydrocarbon species during the very complex set of
reaction steps that occur during combustion.
Very reactive intermediate species (free radicals) consisting
of fragments of hydrocarbon molecules are formed in the primary attack
on the fuel. In the presence of sufficient air, these fragments even-
tually are consumed to form COp and water. However, if the air supply
is limited, possibly only on a local scale, then these fragments can
combine to produce stable hydrocarbon species. Thus, in the incomplete
combustion of the simplest hydrocarbon molecule, methane, it is possible
to find higher molecular weight species such as ethylene, acetylene.
This process is the basis of several important industrial processes for
acetylene production [Anon (1971)]. If the process is initiated by sub-
jecting the feed materials to high temperature without supplying air,
it is usually referred to as pyrolysis. The mechanisms in both are similar.
In a complex system such as an incinerator where parts of the
charge may be isolated from the air supply, both pyrolysis and partial
combustion processes can occur. Similarly in a foundry cupola furnace,
2-1
WALDEN RESEARCH CORPORATION
-------
TABLE 2-1
CARBON MONOXIDE EMISSION ESTIMATES BY SOURCE CATEGORY - 1968a
CO Emissions,
Source 10 tons/year Percent of Total
Transportation 63.8 62.8
Motor Vehicles 59.2 58.2
Gasoline 59.0 58.0
Diesel 0.2 0.2
Aircraft 2.4 2.4
Vessels 0.3 0.3
Railroads 0.1 0.1
Other Non-Highway Use 1.8 1.8
of Motor Fuels
Fuel Combustion in Stationary
Sources 1.9 1.9
Coal 0.8 0.8
Fuel Oil 0.1 0.1
Natural Gas N N
Wood 1.0 1.0
Industrial Processes 11.2 11.0
Solid Waste Disposal 7.8 7.7
Miscellaneous 16.9 16.6
Man Made 9.7 9.5
Forest Fires 7.2 7.1
Total 101.6 100.0
aThese emission estimates are subject to revision as more refined informa-
tion becomes available.
N = Negligible
Source: NAPCA (1970b)
2-2 WALDEN RESEARCH CORPORATION
-------
TABLE 2-2
ESTIMATES OF HYDROCARBON EMISSIONS BY SOURCE CATEGORY - 1968J
Source
Transportation
Motor Vehicles
Gasoline
Diesel
Aircraft
Railroads
Vessels
Non-Highway Use, Motor
Fuels
Fuel Combustion - Stationary
Coal
Fuel Oil
Natural Gas
Wood
Industrial Processes
Solid Waste Disposal
Miscellaneous
Forest Fires
Structural Fires
Coal Refuse
Organic Solvent Evaporation
Gasoline Marketing
Agricultural Burning
Total
HC Emissions,
10 tons/year
16.6
15.6
15.2
0.4
0.3
0.3
0.1
0.3
0.7
0.2
0.1
N
0.4
4.6
1.6
8.5
2.2
0.1
0.2
3.1
1.2
1.7
32.0
Percent of Total
51.9
48.7
47.5
1.2
1.0
1.0
0.2
1.0
2.2
0.7
0.3
N
1.2
14.4
5.0
26.5
6.9
0.2
0.6
9.7
3.8
5.3
100.0
These emission estimates are subject to revision as more refined informa-
tion becomes available.
N = Negligible
Source: NAPCA (1970a)
2-3
WALDEN RESEARCH CORPORATION
-------
portions of the charge reside in an oxygen deficient zone, where hydro-
carbons can be produced.
The other principal mechanism for production of hydrocarbon
emissions is evaporation of volatile species from a mixture. Solvent
emissions are all due to this process. Hydrocarbon emissions from as-
phalt batching plants can arise from the asphalt heating process which
volatilizes the more volatile components. The amount and chemical com-
position of the emission depend on the temperature of the process and
the range of boiling points of the compounds in the base material.
Carbon monoxide is produced in combustion processes by two
major reactions. First, during the primary attack by oxygen on hydro-
carbons, CO is formed by a fast reaction mechanism. The subsequent
reaction of CO to COg is slower and requires adequate residence time in
the presence of oxygen to achieve completion. Thus, in a poorly designed
or overloaded combustor, such as an incinerator, it is possible to pro-
duce significant CO emissions even in the presence of large excess of air.
In the solid fuel fed processes, incineration or smelting, gaseous hydro-
carbons, initially produced by pyrolytic mechanisms, form the combustion
reactants.
Second, CO is produced by the reduction of C02 over solid car-
bon in oxygen-deficient zones, according to the reaction [Engels and
Weber (1967)]:
C + C02 + 2CO
Thus, under proper conditions of temperature and oxygen content, the
gases formed in the primary combustion zone and containing high COp con-
centration, may equilibrate as the gases pass through regions containing
unburned fuel. The reduction reaction above is endothermic so that the
gases will cool as the reaction progresses. When the temperature falls
sufficiently that the rate of reaction becomes negligible, the equilibra-
tion process will stop and the concentration of CO will be frozen. This
occurs at temperatures of 600-700°C [Engels and Weber (1967)].
2-4 WALDEN RESEARCH CORPORATION
-------
In cupola furnaces, the design specifically provides a reduc-
tion zone where conditions for high CO production prevail. This is neces-
sary to prevent oxidation of the melt. In incinerators, similar zones
exist, not by design, but through poor distribution of both charge and
combustion air. It is also possible to produce CO in improperly-fired
stokers if the bed depth is allowed to build up excessively.
As a final note, the presence of hydrocarbons in combustion
processes is generally accompanied by CO, since the conditions support-
ing the partial combustion reactions necessary to form the HC's will
also restrict the complete oxidation of CO. Conversely, it is possible
(and observed) that CO emissions will not contain appreciable hydrocar-
bons, wherever the principal mechanism is C02 reduction.
2.2 EMISSIONS FROM FUEL COMBUSTION IN BOILERS
The levels of HC and CO emissions from fuel combustion depend
on the type of boiler and fuel. Combustion equipment was arbitrarily
divided by Driscoll and Berger (1971) into three size groups as shown
in Table 2-3. Each of these three groups was then subdivided by fuel
type, namely, coal, oil, and gas. For each subclass values considered
typical of well-operated equipment as reported in the literature have
been summarized by Driscoll and Berger (1971) in Table 2-4. Typical ef-
fluent temperature, flow velocity and other operating characteristics
are included.
Additional data for CO and hydrocarbons have been added to pro-
vide a representation of the ranges that might be found in equipment in
various conditions, has been summarized by Driscoll and Berger (1971) and
is shown in Table 2-5, expressed as parts per million (by volume). Whereas
emission factors are the most commonly reported set of units in the air
pollution control literature, parts per million is the unit used in monitor-
ing instrument specifications. The figures in parentheses indicate the
extremes of possible concentration values for grossly malfunctioning units.
2-5 WALDEN RESEARCH CORPORATION
-------
TABLE 2-3
SIZE CATEGORIES FOR STATIONARY COMBUSTION SOURCES
Large Intermediate Small
Btu per hour >_ 5 x 108 3xl05-5xl08 < 3 x 105
Boiler horsepower >_ 15,000 10-15,000 < 10
Pounds of steam
per hour ^500,000 350-500,000 < 350
Megawatts >^ 50 < 50 MW Not Applicable
In well-operated units the levels of both CO and HC will be close
to the lower end. Reduction of both CO and HC can be accomplished by in-
creasing excess air and improving mixing in the chamber.
2.3 MUNICIPAL INCINERATORS
2.3.1 General
Municipal incinerators are a major source of pollution
today and will increase in importance as the per capita output of solid
waste continues to grow, and as the amount of land available for fill
continues to decrease. For municipal incinerators, the national In-
stalled capacity now and forecast to the year 2000 by Niessen (1970) is
shown in Figure 2-1. The almost 100,000 tons per day (TPD) of present
capacity is found in about 300 plants. The number of plants is expected
to almost triple over the next thirty years but the number of units of
concern will still remain tractable.
An estimate of existing (1968) and projected (2000) emis-
sions from municipal incinerators is shown in Table 2-6. For some loca-
tions, incinerator emissions can constitute a large fraction of the total.
For example, in New York City, the 12,000 TPD of municipal incineration
capacity is reputed to contribute over 20 percent of the total particulate
emission in the area. The on-site systems (chiefly the multitude of apart-
ment house units) are said to contribute about 32 percent of the total.
2-6 WALDEN RESEARCH CORPORATION
-------
CO
in
o
to
TJ
c
3
o
Q.
s-
d)
Cn
0)
CQ
C.
to
o
o
to
I
CM
I O
CQ >i
< I
h- oo
UJ
:r>
CO
8
a:
o
GO
a:
LU
GO
oo
1
^
IB
E
GO
-u
U CD
^~ E
IB C
O 03
O -t->
C
II
CU
CT>
i-
tB
_J
r^ ^"^
i +j
IB to
E -r-
GO a
^ ^
-l->
"D to
0) -r-
JD E Q
^
a:
tt) "O
en-i-
i- c/>
ro
<~ c
U !-
O E
1 ^Vfc
<4-
1
1
1
1
s
^_
IB
C
'E
O
z
fc«
in
i
i
i
(
i
i
i
i~ i.
<<:
GO tO
X X
&« VJ
0 CD
CM
<0j
fc«
O &«
vw f^\
^^ Ul
I
1
1
1
I
1
I
i_
t- !-
r- <£
^& ^*&
to
r*
S
O
0
1
LO
rx
LO
rx
LO
CM
LO
CM
LO
1^
LO
r^
i
o
CM
O
CM
LO
r^
LO
r>>
i
LO
r
LO
"~
to
to &«
0)
0 i-
X !-
LLJ
i
CD
1 O
O O -x.
O LO i
^3- tx, O
O
Kfr
1 1
o o o
O O i
CO «* X
co
^~
o o o
LO O r~
rx en x
*±
^1-
1 1
0 O O
O LO i
«* rx >t
«3-
1
LO o en
rx o cu
CM ^ Z.
a. *-*
E en
s =
4- .C
^ O 0
0^-- CU
(B 0)
4-> U. (_ O
GO o CO <-J
LO CO
CO CD
CO
O
LO CO
,
co o
LO CO
CO' CD
CO
1
CD CO
co i o
X
o co
co
i 1 r
CO O LO
1*-"
ox o
x. CO \
LO x^ rx
LO m «*
n^
i O O
CD
LO CO CO
in o in
i O O
**
O i
^-
X 0
1^-
*J-
O CM
'x CD
«d-
»d-
1 LO
o co
1 .
X 0
«d-
CM CO CM
00 0 CD
GO GO Z. Z.
4->
C
o
CO 0
c
0 3
^ ^
+J
c
co o
LO 0
o &«
LO
in
v *
4->
C
in o
3- o
CD ^
o
0
LO
O
CD
O
r*
o
x^
LO
i
O
rx
o
o
rx
0
o
LO
o
CD
«d-
o
o
3
U
+J
z:
01
ai
z.
en
cu
z.
to
I c
0 0
i. .a
a i.
>, to
^. o
c
£
i-
cu
a.
>,
4?
f
u
IB
a.
lO
o
M-
0
&«
LO
C
r
S
s_
,
+J
O
IB
Q-
18
O
4
O
&S
O
c
SI
L.
CU
Q-
>.
+V
o
IB
Q_
IB
O
4-
O
^S
O
t/)
4J
C
cu
^-
to
c
to
i.
1
3
4->
CQ
O
o
o
*l
CO
to
ro
to to
r- CL)
T3 S-
GO GO
O
GO
CQ
O
O
ra
2-7
WALDEN RESEARCH CORPORATION
-------
TABLE 2-5
RANGE OF CONCENTRATION AND EMISSION OF CARBON MONOXIDE AND
HYDROCARBONS FROM COMBUSTION DEVICES (POSSIBLE EXTREME
VALUES OF CONCENTRATION GIVEN IN PARENTHESES)
of Fuel C0 HC (as nexane)
Concentration Range, ppm Concentration Range, ppm
Coal
Large
Medium
Small
Oil
Large
Medium
Small
Gas
Large
Medi urn
Small
0-110
0-500
0-2000 (10,000)
0-5
0-120 (1100)
0-1500 (10,000)
0-5
0-5
0-5
0-10
0-750
0-400
0-5
0-40 (100)
0-40 (100)
0-5
0-5
0-5
2-8 WALDEN RESEARCH CORPORATION
-------
TOTAL ACTIVE INCINERATOR CAPACITY
(thousands of TPD)
» » _* N) ro to C
_*>!-» (71 U> CO -3 -» C
(000 0 0 0 0 0 C
X
X
/
/
/
^
S
/
55 1970 1975 1980 1985 1990 1995 2000
year
Figure 2-1. Projections of Future Active Incinerator
Capacity (Niessen, 1970).
2-9
-------
TABLE 2-6
ESTIMATED INCINERATOR EMISSIONS
(thousands of tons per year)
Pollutant
Mineral Parti cul ate
Combustible Parti cul ate
Carbon Monoxide
Hydrocarbons
Sulfur Dioxide
Nitrogen Oxides
Hydrogen Chloride
Volatile Metals (lead)
Polynuclear Hydrocarbons
Subtotal (Particulates)
Total
1968 Emissions
Estimate
Furnace Stack
90
38
280
22
32
26
8
0.3
0.01
182
496
56
32
280
22
32
22
6
0.3
0.005
142
450
2000 Emissions
Estimate
Furnace Stack
708
131
829
64
161
147
219
0.055
0.03
1064
2259
118
49
829
64
160
114
147
0.025
0.009
391
1481
Source: Niessen (1970)
2-10
WALDEN RESEARCH CORPORATION
-------
2.3.2 Emissions
Carbon monoxide is the most significant pollutant (on a
weight basis) emitted from municipal incinerator furnaces. Existing air
pollution control (ARC) devices are installed for particulate control and
are ineffective for this pollutant. The main mechanism for the formation
of hydrocarbons and CO in municipal incinerators appears to be the pyroly-
sis or partial combustion of the charge in oxygen deficient zones in the
furnace [Niessen (1970)]. In well-mixed multiple chamber units, with large
excess air, much of these partially-burned species are oxidized all the way
to COp. When hydrocarbons appear in the effluent, CO is also found, although
the converse is not always true. The ratio of these two species was found to
be [Niessen (1970)]:
CO/HC = 17.0+12
Although the variability is quite large, this pairing was found to be
statistically significant. A typical set of emission factors and flue
gas'composition concentrations was determined by examining a composite
set of emissions data from many incinerators [Niessen (1970)]. The emis-
sion factors for CO and hydrocarbons were:
CO = 34.8 # CO/ton
HC = 2.7 # HC/ton
The typical concentration of the exit gas is given in Table 2-7.
The actual level of emissions from a given incinerator
can depart substantially from the typical values as given above. One
major determinant in the emission level is the type of incinerator. The
following eleven major concepts or types represent the usual construc-
tion of incinerators in use at the present time.
. Batch Feed
Continuous Feed - Rectangular Construction
Continuous Feed - Horizontal Cylindrical Construction
2-11
WALDEN RESEARCH CORPORATION
-------
TABLE 2-7
ASSUMED NOMINAL CONCENTRATIONS OF GASEOUS POLLUTANTS
AT ENTRANCE TO ARC DEVICE
Type
co2
CO
NO,
Concentration
ppm (volume)
45,000
1,500
15
Basis
ADL Estimated Average
ADL Estimated Average
Rose et al , JAPCA 8
NO
Total Hydrocarbons (as CH.)
Total Organic Acids (as CH3COOH)
Total Aldehydes (as HCHO)
so
HC1
HF
(4), 297-309 (1959T
35 Rose et al, JAPCA 8
(4), 297-309 (1959J
100 Recent Tests on 3 New
York City area incin-
erators
1 Recent Tests on 3 New
York City area incin-
erators
3 Recent Tests on 3 New
York City area incin-
erators
30 ADL Estimated Average
50 Recent Tests on 3 New
York City area incin-
erators
60 Recent Tests on 3 New
York City area incin-
erators
2-12
WALDEN RESEARCH CORPORATION
-------
Continuous Feed - Ignition Grate plus Burnout Kiln Construction
Continuous Feed - Volatilizing Kiln and Burnout Grate Construction
Continuous Feed - Three-Stage Rotary Kiln Construction
Continuous Feed - Waterwall Construction
Continuous Feed - Waterwall Construction/Suspension Burning
Semi-Continuous Feed - Package Incineration
Semi-Continuous Feed - Rotary Grate Construction
Continuous Feed - Residue Fusion
Emission levels of almost all pollutants are influenced
by the type of unit, as shown in Table 2-8 [Niessen (1970)]. In general,
emission levels will be lower and show less temporal fluctuations from
continuously fed units than from batch fed units. Multiple stage construc-
tion provides better contact between the partially combusted species in the
exit gases with air. In these and other well-mixed units, CO and hydro-
carbon emission levels may be down from the typical values by a factor of
10 or more. The time required in a well-mixed system to reduce the CO con-
centration from 1 percent to negligible values varies from less than a
millisecond at 300 percent excess to about 0.035 second at 50 percent excess
air.
The emission levels also vary markedly with the nature
of the charge. Little quantitative data exist relating these factors.
Not surprisingly, the levels are lowest for the most easily combustible
charges. It is customary (although not recommended) in many installations
to segregate the wastes and charge the better burning portions during the
daylight hours, leaving the wet and difficult-to-burn portions for the
night operators. This practice can cause the emission levels to jump
more than a factor of 10, for example from 150-200 ppm CO to 1500-2000
ppm CO, at night, and from 5-10 ppm hydrocarbon to 50-150 ppm at night.
Temperatures reported in incinerators vary and may be as
hot as 2150°F. The actual temperature to be found at a stack sampling
location can generally be expected to be in the 300 to 600°F range, de-
pending on furnace type, air pollution control equipment used and distance
from the combustion unit.
2-13
WALDEN RESEARCH CORPORATION
-------
c
^
S
*i
j
s
£
c
«
S
>-
o:
s:
^ =
o: *
P 5
o -5
<£ t
00 ^ 1
CM g a
S
II
l=5
t» H
li
a -
1?
i
?i
H O
.0 £sz "«* $ ^^
OO ^ * O CM CM 0
to «« ? ai«
eg 5g o MNO
10 0
n CM 32
.0 °?o °*o ^ **- .
OD epS - J r>o
O O) PI p
.0 "Pio ^» <". ">T
o JS o rt CM o no
n ^ PM CM
<^ r
ii
E (J
O O) n o
,cs -p. ^» "". "! *
O 2 «"*" » «- o no
n M in CM
1«
I. o "~:» "> "! *
"ft ng CM- R no
2 Nq S.2128
in ui ^IK U»«N *- r«o
^* T1* 00 " ^
1
*
r-p «0 ?oo 2 8S
". Z O)Z mcM no
n »~ ^ n w
ff
o|«g I
1 *
o en « o>
So °.m ^» " *1
S8 sa wS s no
c e
H|
ISm "w Ǥp ' 1 "
cnrM PMC) nK en no
N - CM CM
w i
ii
| 58 iS SR 1 SS
|{ » «* S?P 5 Sfi
|J RS =S r>S R "°
II
r^ u»r^ oq qr»
to *nr^ wn < no
^ c*» w r»
" Vi *R Ss
o ON n R o£
R S«t 5s $8
o b « <> g o 2
W *~
$ 2,. 3JS $0
53 »S S| °!
o 52 r)| §|
2 So, SR $g
-: CM o n g o 2
CO ^ t O>
us en CM o> o 9) o
J CM in ri g 62
^ 5f ^
S 9i C>( O) O O> Q
*" w ^ n S °
S ot r^ 5 Q o> o
J o. jn r> g o «
r-> r^ 32!
« CO V u) O CT) O
ri in o ri g o 2
* ^
f- r» 22?
« 00 f O> O CT> O
rt in o n fi o 2?
o ^ ^.
(N n j 2
r>. O O CD O Q * Q
~ ^^ n 8 ° *
to o\ 3 2 _
(N ^ f^ CT) Q mo
c, Wj; ng 02
u> en 5 2!
CN - r* at Q o
CM WR »g 02
to o> 2 2J «.
(s* ^ r*. d> o "1 Q
a «s "1 °=
rx O f** v> O Ot O
(N in JQ ri g o S2
8,
b o
o b
0 0
o o
B
o *
b o
o b
(N
ST.
0 O
B
o <»
o o
CM
3-r
o b
(N
S*
o o
B
O *t
O O
s.
o o
CM
K
0
o o
1,
o o
B
o «»
o o
n
o v
o o
!
II 5 I
i?« I
2 'o 1 I
i 5
o
3*3 S,
o o * R
° c S -5
III
ill ?ll ?3s sli
s-tiB T> a o « ? » 2
X ft ^-^ "S *"
d'S'S v^'S u'o'S 5'o'S
£5 J^
1)
I
5
2-14
WALDEN RESEARCH CORPORATION
-------
2.4 GRAY IRON FOUNDRIES
2.4.1 General
Foundries have often been singled out as a major indus-
trial source of particulate emissions. The metal melting operation can
also generate hydrocarbon and carbon monoxide emissions. There are sev-
eral types of melting furnaces employed in gray iron foundries including
cupola, reverberatory, electric arc, and electric induction furnaces.
The cupola is the most commonly utilized and is the most serious poten-
tial emitter of hydrocarbons and CO.
The cupola is very similar to a blast furnace except the
charge is iron as scrap or pig iron instead of iron ore. A typical cupola
is depicted in Figure 2-2 [Taylor (1959)]. In large installations, opera-
tion is semi-continuous with the charge added at intervals through the
charging door, and the melt periodically removed (or tapped off) through
the taphole. In many installations, the cupola is emptied every night by
dropping the bottom plate and allowing the residual charge to drop out.
2.4.2 Emissions
As noted above, the principle mechanism for the forma-
tion of CO is the reduction of COp by carbon in the superheat and melt-
ing zones. Figure 2-3 shows a theoretical profile of the composition and
temperature of the waste gas from a cupola [Engels and Weber (1967)]. The
CO concentration can range from 10-20%. Hydrocarbon emissions accompany
the CO resulting from pyrolysis and incomplete combustion of the coke fuel.
Data on typical hydrocarbon levels in uncontrolled cupola waste gas are lack-
ing. The temperatures of the gases above the charge range from 300-600°C.
Large quantities of air are drawn into the cupola by leak-
age through the charging doors, amounting to as much as the waste gas in
some older installations. The mixture resulting from the addition of air
containing free oxygen is potentially explosive. After-burners have been
traditionally used to consume the CO in a controlled manner and prevent
2-15 WALDEN RESEARCH CORPORATION
-------
Spirit arrester
IT
Metal charges
Limestone charges
Coke charges
Blast pipe
Preheating
Bottom door in
dropped position
Bottom prop
Melting zone
Superheating zone
1
Figure 2-2. Cupola (from Taylor, 1959).
2-16
WALDEN RESEARCH CORPORATION
-------
UPPER EDGE OF BURDEN COLUMN
w w
II
U D
§O
- S
u
1
REDUCTION
ZONE
OXIDAT ION
ZONE
TU YERE *
V
\
i
\CJ
V
I
V
, -
~-
1 zz
j
±S T
'"
^
fff
CMP
!
i
u
ERA
TUR
E
PRE-HEATING
:o ZONE
i)
SUP
TRI
^ i
1_
^T_rNG ZONE
ERHEAT
AND
CKLE ZONE
) 4 8 12 16 20' 24
0,-,CO,-and CO Content in %
0 400 XO 1200 16002000X00
GAS TEMPERATURE IN °C
Figure 2-3. The Course of Gas Composition and Temperature
in the Cupola Shaft (Schematic).
2-17
WALDEN RESEARCH CORPORATION
-------
explosions. Now this mode of control takes on an additional role, in
controlling the emissions of CO and hydrocarbons as pollutants. In the
simplest installations, without sophisticated parti oil ate control devices,
the afterburners usually consist of burners located just above the charging
doors firing into the cupola shaft. The heat released raises the tempera-
ture of the waste gas sufficiently to support smooth, controlled combustion
of the CO (1200-UOOeC) [AFS (1967)]. The temperature of the gases after
combustion rise to about 2000°F.
In more complex installations the waste gases are drawn
off below the charging doors to minimize air infiltration and thus the
load on the particulate control device, scrubber or baghouse. In these
cases, an external fume incinerator or afterburner is used to control CO
emissions.
In either case, the levels of CO in the effluent gases
will have been reduced from the high levels at the charge to the order
of 1000-5000 ppm in external afterburners, and 5000-20,000 ppm in in-
ternal afterburners with excursions up to perhaps 50,000 ppm (5%).
Emissions levels of both CO and hydrocarbons will fluctuate during the
various parts of the charging cycle, and in smaller operations may
change considerably during the day.
2.5 REFINERIES
2.5.1 General
Petroleum refining operations can release hydrocarbons
from many points in the overall process. The principal points of emis-
sion are storage tanks, vents, and catalytic cracker regnerator stacks.
The installation of closed cycle storage systems, as generally required
presently, reduces the emissions from that source to a very small mag-
nitude. The catalyst regenerator still represents a significant source.
Central to refining operations is the conversion of the
organic molecules as present in the crude feed stock to molecules of
2-18 WALDEN RESEARCH CORPORATION
-------
higher economic value, for example, the gasoline range. This conversion
is almost universally accomplished in catalytic reactors. In the process,
the catalyst becomes coated with a carbonaceous material and gradually
loses ifs efficacy. The activity of the catalyst is restored by burning
off the carbon in a regenerator under controlled conditions. Insufficient
air to oxidize all the carbon all the way to carbon monoxide is used. The
resultant gases contain as a result high levels of both CO and hydrocar-
bons. The reducing combustion condition is maintained in the regenerator
to prevent damage to the catalyst.
2.5.2 Emissions
Carbon monoxide emissions from uncontrolled regenerator
stacks are in the 6-10$ (by volume) range. The hydrocarbons generally
fall between 100 and 1500 ppm. A set of emission data, taken in the most
extensive study of refinery emissions, is shown in Table 2-9 [LAAPCD (1958)],
Regenerator stack temperatures are usually 800-1000°F [Cuffe (1967)].
The regenerator effluent gas contains a significant amount
of potential energy tied up in the heat of combustion of the CO and hydro-
carbons. Most refineries feed the regenerator stack gas to a waste heat
boiler (see Figure 2-4) to recover that energy for economic reasons. Now,
with additional emission controls, it is even more likely to find a waste
heat boiler in conjunction with a catalyst regenerator. Emissions from
the waste heat boiler will have similar characteristics to gas-fired con-
ventional combustion units, as described above.
2.6 ASPHALT BATCHING PLANTS
Asphalt batching plants represent a potential source of hydro-
carbon emissions, although most concern with this type of source has been
focused on the particulate emissions. The typical operations are shown
in Figure 2-5. The overall process involves heating and classifying the
gravel or crushed stone aggregate, mixing it with heated asphalt, weigh-
ing and transfer to vehicle.
2-19 WALDEN RESEARCH CORPORATION
-------
12
o
CJ3
UJ
at
cr>
CM
O
2
o
o
o
a:
u.
co
o
o
u_
o
00
co
CO
to
,,_
c-
^
r
^
«
"(8
O
h~
.,_>
r
C
13
01
c
^
u
ra
i.
O
u
r-
^
[ *
to
O
<4_
o
CO
a.
I
o
11-
o
^
CO
JC
^
a
-^
r-*
I i
in o o o
i O O O
o o o
lf> CO 00
CM !>» r-
CM VO
CT) O O O
000
O O O
> TJ U.
fO "x^ tJ
o to co
*^x. ^~
10 J3 "
i .a co
.Q 4->
-Q * to
CO C£
to CO IB Q)
-i-> 4-» a: o>
1- IB fc.
C OH TJ (B
ZD 0) -C
-0 CO 0
*<- 0) U- I/I
O CO -r-
u. ai a
j- i
CO JZ (J ^^
JD U1 >, 0
E ai o
z u_ an co
(B
TD
s-
co
o.
(/)
c
o
CO
c
IO
S-
E
Q.
Q.
(0
-a
J-
OJ
CL
to
c
o
43
O)
CD
c
(0
fc-
E
a.
a.
to
-a
t.
a>
Q.
t/)
c
o
4-»
O)
CD
C
(O
SM
£
cx
Q.
to
cz
o
i/)
in
^
E
UJ
a
c
to
to
c:
IB
C
^
(B
-M
O
C_J
r~-
CM *5j" ^^ ^i" Ol * >
f^ CO LO LO CM ^J" t~~ *~**
CM <
LO O «4-
vo ^ cr> rC P^ °l
i CM CO VO r- O
III III
00 VO CM O
o
CM
0
CO VO "d- CM «* O
COCMC3OCDOC>CJ
LO
VO «3" CM O r*s
r i VO r t O
III III
00 VO CM O
,
o
f*~
CT> 00 CO CM LO O
r^ f-~- ^~ ^t~ CM ^* ^ CD
^~ ^>
CO O ^>
i O> ^" LO CD CTl
CM f^ O> ^* CO
i CM CO VO i O
III III
oo oo oo r-. co CTV
en o vo i
CO
o
CM
o
a>
10 4->
(B 4->
(B
** ^*
c o
CO 0) 3C
CD 4J C_) z
O (B CO DC O
-r- «
J3 -f- M- t. j_ to «
J- O O , -o
23 CO i 3 C -C !-
«4- -O tO M- O CO c
Or -i- +J i 5"° "*
OCCOOI COeC«tO
00
a
CJ
a.
(U
u
i.
3
O
oo
2-20
WALDEN RESEARCH CORPORATION
-------
Figure 2-4. Cylindrical, Water-Cooled, Carbon Monoxide
Waste-Heat Boiler.
2-21
WALDEN RESEARCH CORPORATION
-------
SANO AND
AGGREGATE
8INS
7
CYCLONE
COID AGGREGATE
BUCKET ELEVATOR
STACK
HOT AGGREGATE-
BUCKET ELEVATOR
Figure 2-5. Asphalt Batch Mix Plant.
2-22
WALDEN RESEARCH CORPORATION
-------
Hydrocarbon emissions can arise in the asphalt heating step and
in the mixer, where the lower boiling components in the asphalt can be
volatized. Very little data are available on the levels of hydrocarbons
or the total emissions per unit processed (emission factor). The chemi-
cal composition of the hydrocarbons emitted in the volatilization process
will be markedly different from that of the other processes where pyrolysis
is the principal mechanism. Insignificant quantities of low molecular
weight paraffins (methane and ethane) and unsaturated (ethylene and acetylene)
species will be emitted. Rather the effluent will contain principally higher
weight paraffins and aromatic species.
2-23
WALDEN RESEARCH CORPORATION
-------
3. MANUAL METHODS FOR THE DETERMINATION OF HYDROCARBONS AND CARBON
MONOXIDE
3.1 INTRODUCTION
In this section, a brief review of manual methods* for CO and
hydrocarbon determinations is given. This is followed by a description
of the necessary modifications during laboratory evaluations of the
methods. The object of the work described in this section was to provide
reliable laboratory methods to check the validity of the continuous
monitors in the field at the field test sites. The actual results
(comparisons with continuous monitors) obtained with the manual methods
are discussed in detail in Section 6.
3.2 HYDROCARBON ANALYSIS
A number of techniques have been developed for measurement of
classes of hydrocarbons rather than one individual member of a class.
These are described in the following paragraphs. The G.C. methods appear
most useful but a review of other techniques is also given.
A number of Russian papers have described the use of polarog-
raphy for the determination of aldehydes in the atmosphere (McNevin and
Clone, 1953). One paper describes a procedure for determination of
mixtures of aldehydes and ketones via derivative (semicarbazone) forma-
tion. This technique would be useful for determination of classes of
aldehydes unsaturated, aromatic, etc., since their half-wave potentials
are considerably different. A polarographic method was used to determine
acrolein at approximately the 1 ppm level by Van Sandt in 1955, however,
the sampling and transfer procedure was quite elaborate and time consuming.
Polarography might have some usefulness for determining aldehydes but
it suffers from lack of sensitivity (colorimetric methods are sensitive)
in the ppm level) and is more of a laboratory tool than a routine
analytical technique.
*
Here "manual" connotates collection of the sample in the field and
analysis by colorimetric or instrumental methods in the laboratory.
3-1 WALDEN RESEARCH CORPORATION
-------
A Technicon Autoanalyzer method for aldehydes has been described
by Younghams, et al (1967). This colorlmetric method used Schiffs reagent
with an excess of sulfite ion. The amount of colored product produced by
reaction with pararosaniline hydrochloride is then dependent on the concen-
tration of the aldehyde absorbed in the solution. One major problem with
this technique is that the rate of color formation depends on the structure
of the aldehyde. This method has a very limited range of linearity.
Altshuller, et al (1961) have reviewed colorimetric methods for analysis
of aldehydes in the ambient air and in source effluents.
Colorimetric techniques are described which can distinguish
between:
(1) formaldehyde - chromatropic acid
(2) unsaturated aldehydes such as acrolein and crotonal -
dehyde-4-hexyl resourcing!'
(3) higher aldehydes - C3 or greater - 3 methyl benzo-
thiazole or p-nitrobenzene diazonium fluoroborate
These colorimetric methods appear very useful for the determination of
aldehydes since they are very sensitive and can give the desired selec-
tivity for acrolein and other aldehydes.
A photometric titration method is described for organic acids
by Lorch, et al (1966) which has been adapted to the Technicon auto-
analyzer. A major problem with the direct application of this technique
is interference from acidic species such as C02 and basic species which
could be present in combustion gases. A technique used for organic
acids by LA APCD involves collection in NaOH. The sample is neutralized,
extracted into an organic solvent and washed to remove traces of the
excess inorganic acids. Following the extraction, the sample is
titrated with NaOH using a standard acid-base indicator. This technique
suffers from lack of sensitivity and is not necessarily interference-
free.
Colorimetric methods are described by Ayers (1956) and
Iwayama (1959) who determined the copper soaps formed in chloroform
3-2
J WALDEN RESEARCH CORPORATION
-------
solution following the reaction of the potassium salt with copper
nitrate.
The sensitivity of the method was improved considerably by
Duncombe (1962) who complexed the copper with diethyldithiocarbonate.
A modification of this procedure was used by Dal ton and Kowalski (1967)
for analysis of organic acids with a Technicon.
The one drawback with this procedure is that it is for C^-
C-|8 organic acids. For lower organic acids, the solubility in aqueous
solution is too great to extract directly. A slight modification of
this procedure, however, might make it usable for C3-Cg organic acids.
If the sodium salt of the organic acid (less than £$) is treated with a
saturated solution of sodium chloride prior to extraction with chloro-
form, the distribution coefficient can be shifted enough so that the
soap - (03 or 64) will be preferentially extracted into the chloroform.
Thus, this modification should make the determination of butyric or
caprylic acids feasible.
The most useful technique for hydrocarbon analysis is gas
chromatography, since all hydrocarbons can be determined by the same
technique. Individual hydrocarbons and classes can be determined
quantitatively and with adequate sensitivity by G.C. with a flame ioni-
zation detector. An example of the separation of G£ to Cg organic
acids is shown in Figure 3-1. Some columns and conditions for separa-
tion of aliphatic, aromatic hydrocarbons, organic acids and aldehydes
are given in Tables 3-1 to 3-3.
For our referee hydrocarbon method, we selected the method of
Bellar and Sigsby (1967) which utilizes silica gel columns. The optimal
conditions were slightly different than those of Bellar and Sigsby but
good separation of CH^, ^6' anc' ^2H4 were obtained with the conditions
in Table 3-4. The data on the comparison of the manual hydrocarbon
analysis with the instruments are given in Section 6.
WALDEN RESEARCH CORPORATION
-------
2
- _ -- _. _ ii i_. t
1
\
I
L
M
1. Acetic
2. Propionic Instrument:
3. iw Butyric
4. n-Butyric _ .
1 Sample:
5. iio-Valeric
6. n-Veleric
4 Detector:
Sensiti-.ili'-
I
^u
> Column-
6 Tempersture
i
Flow Rates:
Chert Sp«*d
U L.
1 i 1 i i
0 2 4 6 8 10
Time (mm)
SP-1200
fiJtier/Victoreen Sonet 4400
GC
0.1% Aqueous Soln. of Fatty
Acids Cj - Cj
FIO
80 x 10-" aft
Single, 6 ft x 3 mm ID glaii.
packed with 10% SP-1200. 1%
HjPO, on 80/100 meih
Ch'omosorbW-HP
Inj: 300°C
Del. 300°C
Carrier He, 35 ml/mtn
Hydrogen: 28 ml/min
Air: 300 ml/mm
0.5 in/mm
Figure 3-1. Separation of Organic Acids by G.L.C.
3-4
WALDEN RESEARCH CORPORATION
-------
GO
O
CO
a:
^£
CJ
0
a:
Q
1C
o
1 C
I
T~
a.
r 1
1
^£
0
o
z.
r 1
U_
LU
i
O
o
r 1
r r
oo 5:
CD
LU OS
> > 3 ro
r r ' N 3 "
O O O O-i^
a. o o .a o
0 S--r-
co = l^» ro S-
3 00 r- 0 J3
O -^ 1 QJ
S- r Lf> ^9 S_
O ^f O <-
a. -x. - r M-
,
0 0
CD CD
CO
c
O CO
-a o
ro ^3
U rO
O E
^ O
TJ S-
>ij r^£
^T~
>1
-0 00
co co
cnoj
r- ^3-
oo OJ E
QJ
f* c*
LU r | 0
1-3 .
^: r
-DO fO
C QJ C
ro I <
** "O **
C
cc ro (/)
1 . LU
r
*t CJ *»
^^ i- CO^-^ CO
OJ ro r^ -O
vo . vo o
CT> i > VT> O
QJ E r rO
**-^ CO LU * ' *^
a
-o QJ J=
OJ CO C +->
CO ra -i -i-
3 (0 S
QJ -f->
CO ^ JD C
r 0 E
QJ -C 3
CT>-l-> CO i
r- C 0
rO 3 O U
O -i-
r- « +J S-
r CO (O
O
>> T3 C
4J - O QJ
QJ OJ O O.
r- r O) O
L. 1
ro VO " ~
> CO O
E ro LO
r
.
CJ
CJ
CD
CO
c~
O
J3
S-
co »
(Q l~~ "T"
0- O
QJ OO
r VO
^_ CT^
, f~~
^" ^r_**
* LO
VO S-
VO ro 4->
CT» D> CO
r CD E
- ro ii
r
CO O
^" ^* *o
QJ E 1
co E t.
(O 3 J
3 S- *
tj (O fO
- r C
r ^ *r
r- E
0) Q. 3
E ro
O O
r- 0 O "-
CO O -5
*3- O
Or- -a
O 1 X QJ
oj LO -*:
LO - O
CJ 1 O ra
^3 x- * ^* f~^
,
CJ
CD
CO
C"
O
s-
ro
O
O
S-
-o
^
r O
VO I
CJ 0
| |
i VO
CJ 0
LO ^" i
O O CJ
1 1 1
r OO i
CJ O O
LO
0
1
r
r^
CO
e
a
_c
(__>
^~
rO
C
ri
t
"^_
m
HH
#i
IrO
E
r
(
^~
r
3E
O
4_>
1
QJ
3
CT
r-
E
.E
'-^ O
O QJ
0 +J
O
00 QJ
*°*-^ >
r
CO J->
ra o
en ra
S-
S- -!->
QJ J3
- 3
S- 00
.
CJ
CD
CO
0
'^3
ro
E
O
S-
^^
T3
ro
CO
E
«l
<0
i.
ro
a_
^ ^
If)
vo
CT)
^^
CO
OJ
1
1
- E
QJ 1
-r-> CO
ro >-
ro -t->
-!-» -O
J= E E
O. (O <-
^CL OJ
O 1 -l->
QJO.ro
-0 S- -E 3
r- O O. I
-O CO CO .Q
O O S-
E -c O
CO O O. CO
E S- O
E -E E
3 O >> O
r r S-
O E O -E
O O »->
-------
^
0.
^f
fr
CO
o
1
J»~
o
or
in
CJ
co
s
DQ
CO
Q
t i
O
CJ
I-H
z.
to
QJ
r
0
Q.
CO
3
.^£
ra
Ol
ra
NX
1
QJ
01
o
c
o
a.
Q.
^
z
+
H-
»l
O)
ra
^s
ro
C
E
3
*O
0
»*>
CVJ
o
r__
v_^
O
CM
C
O)
Q)
2
4-i
|
CO
o
1
0
to
ro
i
a
,
0
CD
to
a
r-
O
O
r-
C
ro
CO
S-
o
S-
01
o
4->
r^ ro
CO E
VO O
CD S-
r f"
<-^ 0
LO (/>
1 ra
CVJ CO
o
CO
ra
»«
^-s «*
LO
<** rD
r^* *^C
CO
#1
** >«k
- ja
J= (U
i/) i
r- Q.
ro Q-
^
C <0
0.-I- Q.
HE E *i"~
OJ "^x. "O
4-> i (0
E
CO i
i CD O
0.**- 0
E >>
(0 S- r-
-a
CO
-3
ro
-^ E
VO -C
CD LO
o
" -^ t-
f*
*tf" "^
!«.
1 "
r~* "T"
CVJ
*
r\ *|
*~ "*+ "^
*& O
' C
3
LO >-
HH f
HH +J
^
**"^* S
0)
4-> -O
ro 0)
O. -C
i- 01
TD ro
rd 2
r ^
O N
0 4->
>> s-
i ro
Cf> 3
CO
IM
c
ro
^ -»->
^- to
VO O
CT> CO
r
~_~- .
^_
00 10
t 4->
CO I-H
CM
CO
ft »^
> or
^^
i >
co
CO U-
(
«s
i-
o
^ r"*
O ro
CO ^
,_
O M-
0 O
^.
!__
o> c
^
CO O
c m
i ro
^^
( r
4_) 4^
CO -r-
>» 2
lo -o
CX (U
i- CT 4J
* S- CO
OS- C
O ro Ol
1^ O r
f- >>
II LO -C
1 ro 4->
^
^-. OJ S-
E :n 4->
O v_>
O o~^ >>
O O i
CM 00 O
* 'CM O_
.
t ^
~~i
CD
00
-o
o
0
r
c
ro
CO
S-
0
CO
C
>-O C
4J S-
CO O 0
Q. «
^ifi ^D
i "O dJ
0 ro -G
Q. 10
T3 3
S- C S-
^5 n3 CJ
»
0
CD
to
-o
r-
0
O
^
C
ro
cn
O
vo vo ^
CJ
1
o
1
CJ
i
i CM r-
CJ
o
O
(U E
4-> 3
ro O) i
O 4-> O
O ro O
C.
to
O
ro
O
+J
ro O
§CM
^^ ^ **
vo -C
CD O cn
t r~~
*-- <1>
« CQ
CO
1 0
CJ E
CO S-
r O ro
r- XI
" M- O_
CO 4J
' C *-3
O)
r »f n
* O
co co
*1
O) S- »
to o> .c
rO to -i-
S- >- E
CD U_ CO
*§ C
T3 «* E
i- ~O 3
0 O -i- C
ro O E O
T3 ro 3 O
0 C i
r- ro O O CO
to T- O 4->
rO -^"^ C to
-Q -O O O- Q-
i- -r- Vt- eC T-
s- o i U. -a
-t-> ra 3 U. ra
to
^^ O °r~ * r
LO-i- T3 CO O
CJ 00 10 U
ro QJ ro >,
(O XI C -C r-
^i- QJ Q. co
*O f-^
T3 ra T3 QJ
- VOuC -r- C
O CO4-J 3 QJ
ra O -C CT r-
a. -r- >,
S- &« ra QJ r .C
OJ C3 C C 4->
E r I QJ CM QJ
r- r- 3 0 >,
S- >> 0 i
"i- >)4->CncO CM O
COO ro
00 ,
Z. 3 0
3: oo CD
CJ
CD
to
^3
^
o
<
o
1
c
ro
cn
c_
0
4-> O 4J i O.
S-S C C 1
0 O T- C Q-
i O T3 -i- CO LO
.
CJ
o
CD
CD
to to
o -o
0 O
0 0
^ »n-
c: c
ra ra
cn cn
S- S-
0 0
CD
co
«\
.
E
OJ
JZ
CJ
*-^
LO i
IO ro
£ <
Ct
CO
co i
i
LO «
CM 00
CM 0
O
S-
^^ i-
vo c
1 QJ
LO 4->
co
c^.
O
QJ 10
4-> QJ
to E
O
? O
O CM
1 1
M- o
LO
1*-^* r^
QJ CJ
-!-» 0 -O
r- f>«. C
r CD ro
QJ r
0 1 O
O O
.C i
I/) l~"~ 1
0) II O
61 oo
o
CM C O
T 'i .*
O ^ ro
CD Q.
r E ra
S-
C C3 0
o vo o-
o
CD
to
a
r-
O
^^
O
CD
co
ft
.
S-
cn
o
+J
ra
VO S-
CD Jd
r- CJ
*^*
«e^- *73
LO
1 »
CM
LO 31
VO
r «
CO
^" >l |"
CO QJ
r- O
*-> Q
x^
QJ
1^3 ^""x ^3
o- a.
coo. !-
n: to T3
OJ ra
^S S-
LO r
CO > O
0 0
4-0 >,
00 r-1
CM cn
»« CM
»* CO
a c
-O C QJ
C ra r
ra >»
CJ -C
CM 0 -M
CM CO
-c >»
+-> CM i
i- II O
21 Q.
*^^
a &s
OJ . LO
4-3 O- CM
ra oo i
O QJ
OS- CO
.
CJ
CD
to
-o
o
-------
*^
3:
Q.
o:
CD
o
I
^?
o
T-
CJ
oo
^
TJ
o
CJ
c
E
r*~
o
CJ
-a
o
-C
CO
2:
(/)
CO
o
CO
CX
oo
^^
co E
x x CO
(O -C
M to CJ
O
^^ * M
00 CO
CO - - "
E LO
^ to s:
-C C71
31 O r-
2*i 0^"' -S
i- r-. in
"i IO i.
CD 3 1 O
E u 2
o cr> -Q
to ^- LO co
S- II i-
if o ** ** *4
0
-a >>-D c
CO i -i- O
4-> CD 3
ra cr o
O CO T- O
0 C. <£>
CO
-I >> .
E >, s- o
E -C ro U
3 4-> C >,
. CO O
0 >>!- cn
CJ i~ | »
O ra CO
>> CX4J C
S- to CO
ra C i
i CO to >,
i x: ra -C
.r- 4J 4->
CX ^ QJ
ra "O >,
0 E O i
«d- o
(/)!-! CX
t/> 14- cn
ra CO *S
r cnn. o
CD eC - r-
CJ
o
1 9
CD
CD
to to
a> co
-a -o
>) >-J
_CT -C
0) CO
~o ~o
1 1
C
ra E
(/) I""
fO O
cn o
L o
co <:
r- U_
J- U_
y_
fO fcS
0 0
CM
to
rO ^"^ =
"S. CJ LO
0 CM
z o
CD O
iTT >,
CO O XI
*f^ f^
3 II =
O 1 CM
CJ
CD
t/)
QJ
T3
^)
-C
a>
T3
0
N
rO
CD
^
,
--. u_
to
to oo
CT>
ro
>^
^^ rO
CO -^
CM to
i O
O
oT rO
I
C
o
0
^ ^
c
p
LO
0 CM
O '
0 0
r-.
X
it
t J
OJ
^
E C
^ e
r 3
E i
O
0 U
to
>^
-!-> 5-
ra ro
to i
fO 'I-
cn ex
ra
i. U
QJ
- 3
S- CJ
CJ
CD
O
3:
CJ
(J
LU
ft
O
3:
CJ
0)
s:
^
0
3:
CJ
3:
-d
rO
N^
CO
iO _^
to
ro to
3 ro
CD
^^ ^ -s
0 S-
.«>o CO
CO IO !-
4J CO S-
(0 II S-
i 1 ro
ro O
f
4-> C tO
-C -r- rO
CX E 2
r i 3T
>> E
E
ra LO CO
o ^J- -a
C/l !-
---i-> E
r- to ro
T3 0
to s_
CD ra CX
c cn ra
r- O
E
r-
<7"
>|K^
O
O r-
CM ro
| C"
f*^ ^C
CT>
1^ M
f\ ft
i- X
r |
r
** " ^T"
r «i
CM >
QJ
Tco T
-i^ tO 4->
3 CD 'r-
rO r E
Z * Q
to
ra
CD
J- 3
ai i
si s-
S- 0
rO «/)
0 O
E
(/I O
rO S-
O
3:
c
0
CO CO
E !->
ro ro
-C i
I-) ro
CO -C
S- 4->
3 -C
>>. - cx
O i
O o >>
CXO !->
O 3
"CM XI
. » I -
O O T3
E S- >,0 CO
r- CO i O II VJ
ro -i- O
(-> S- Q-
O
CD
CO
CO
T3
>*
-C
CO
T3
^~
*t
O 1 LO
CM -1 CM
CJ
CD
r*
CJ
i
c\
0
i
^_
XI
o
.
«-x >
00 O
to z
CTi
i "
««^
LJ_
r~
00 00
1
cn *
f^^ (O
0 >,
i ro
' - O
r^" i
o
CO ro
CM >-
x:
4J
^
3
,__
0
o
^
£~
^-
X
E
o
o
o
+*^*f
^^
CO
CX
ra
to
i
.
t ^
i
CD
*
0
0
M
LU
J
3:
u
>r_-
*« * r^
LO = r
1^ >> CO
1 f~ <~
to cx o
tO ra OQ
i S-
cn .
» o -s
M
E ro LU
- EX:
x: o -a o.
^ S- C rO
x: ra i_
CJ CD
+-> S- 0
r-
n
S-
CO
^ x^^x
S- 0
S- o
ro O
CJ LO
II
CO H-
r
E c
O !-
!-> E
O ro
U
>> S-
r~ CO
cn->-
S-
QJ S-
E ro
CO U Z
^_
>,M ^
X: ro ro
4J 2 CX
CO ra
- s- s-
5- -i- O
*-> ro Q-
.
CJ
CD
1
ra 1
S- CO
rO "O
1 CXr S-
CO CO ro CO
cx to E x:
to s- +->
CO S- O O
O O <*_ to
3: -a 4- -a co
>>
S-
ra
CJ
A
O
3:
o
3;
3C E >> O ra >»
CJ CDI x:
T- r E CO CO
CM (/) ro O T3 "O
3: CO -r- -r- >,r-
CJ Q o +j x: ra
3-7
WALDEN RESEARCH CORPORATION
-------
TABLE 3-4
HYDROCARBON ANALYSIS BY GAS CHROMATOGRAPHY
Equipment - Gow Mac chromatograph equipped with gas sampling
valve and flame ionization detector
Air: 420-470 cc/min
H2: 35-45 cc/min
He: 17-25 cc/min @ line pressure of 27 Ibs.
Col. Temp: 60-70°C
Component Retention Time (Min)
CH4 3-4
C2H4 lol- 12
C2Hg 16-20
3.3 CARBON MONOXIDE ANALYSIS
3.3.1 Introduction
An improved manual method for CO was needed to check the
"accuracy" of the instruments at the various sources to be tested. The
approximate carbon monoxide concentration ranges are given below:
Fossil fuel combustion (pilot plant) 50-1000 ppm
Municipal Incinerator 100-1000 ppm
Gray iron foundry 1000 ppm - 10%
Catalytic cracker
CO boiler inlet 16%
CO boiler outlet 10-50 ppm
A single comparison method was needed which could be
used over this entire range of CO concentrations. A survey of manual
methods for the determination of carbon monoxide was conducted by
Driscoll and Berger (1971).
3-8 WALDEN RESEARCH CORPORATION
-------
The following was condensed from their report:
Manual methods for the analysis of carbon monoxide
depend on one of the following reactions:
Reduction of a Metal
CO + MX2 + H20 -* M + C02 + 2HX
Oxidation
°2
CO - C02
catalyst
Complexation
CO + MX + xH20 * CO MX xH20
The "reduction" methods are attractive because of their simplicity, and
freedom from interferences; however, the long reaction time (for the re-
duction process) and low solubility of CO in aqueous solutions restrict
these methods to grab sampling techniques. This is not considered a
serious disadvantage since a gas sampling bag can be used to obtain the
average CO concentration over a period of time. The grab sample of CO
can then be collected from the bag.
The oxidation methods appear interesting since a long
(integrated) sampling time is preferred to eliminate transients occurring
during flue gas sampling. The major difficulty is that C02 and other
acidic flue gas components have to be "quantitatively" scrubbed prior to
oxidation of CO. In view of the high and variable C02/C0 ratio, these
techniques do not appear very promising.
The complexation methods have been used in power plants
when combustion is poor and CO concentrations are, therefore, high (200-
2000 ppm). Gas volumetric apparatus, such as the Haldane and Orsat,
depend on selective scrubbing (absorption) of CO and measurement of the
difference in volume absorbed in the aqueous scrubbing solution. The
3-9 WALDEN RESEARCH CORPORATION
-------
basic limitation of these volumetric techniques is the measurement of very
small differential volumes. Since the carbon monoxide concentration in
combustion sources and incinerators is less than 1000 ppm, absorption
(complexation) methods lack adequate sensitivity for most sources.
The only class of methods which appeared promising for CO
determinations were the reduction methods. A description of these is given
below:
The gas sample containing CO is introduced into a flask
containing an excess of PdCl2 solution. CO is oxidized according to the
reaction, PdCl2 + CO + H20 * Pd + C02 + 2HC1. Reaction times of 6 to
24 hours have been employed. Allen and Root (1955) have shown that the
reaction time can be reduced to two hours.
The Pd "reduction" methods have determined the extent of
reaction by filtration and subsequent weighing of the free Pd, reacting
the Pd with excess of Br2 and titrating the excess Br2> and photometric
determination of the Pd solution.
Another approach is the determination of the unreacted
PdCl2- After removing free Pd, the unreacted PdCl2 may be converted to
Pdl.= and measured spectrophotometrically at 406-410 nm. Phosphomolybdic
acid reacts with PdClp to form a blue heteropoly acid (molybdenum blue)
which is determined spectrophotometrically in the 635-720 nm region.
The above methods are not specific for CO since any
species which can be oxidized by Pd is a potential interference. This
includes unsaturated organics, H2$, hydrogen, etc. Burianec and
Burianova (1963) have reported a CO method which depends on the formation
of a red-violet complex between CO, Pd and o-phenanthroline. The reaction
is reported to be:
2[Pd(phen)2] C12 + 4CO + 6H20-*-[Pd2(CO)2(phen)2] 4H20
+ 4HC1 + 2C02 + 2phen
3-10 WALDEN RESEARCH CORPORATION
-------
Burianec and Burianova (1963) indicate that the method is
relatively interference free. Only H« at levels greater than 15%, hLS and
CpHp in concentrations over 1% have an effect. The time required for
analysis for this method is less than 30 minutes compared to 2 to 24 hours
for the other Pd reaction methods.
In view of the obvious merits of this method, we decided
to investigate it in detail for use in the program. In the following sec-
tions, the method development and the test results are described.
3.3.2 Development of a Wet Chemical Method for Carbon Monoxide
A wet chemical method for carbon monoxide based upon the
work of Burianova and Burianec (1963) was investigated. Barianova and
Burianec (1963) characterized a new carbonyl compound of palladium and
developed it into a quantitative method for carbon monoxide (1963a). They
found that: (a) Pd (II) is reduced to free Pd, (b) one mole of carbon
monoxide is required for every mole of Pd (II) reduced (Pd + CO + FLO
2 Pd + C02 + 2H ), and (c) each mole of the bis(o-phenanthroline) palladium
(II) chloride complex in aqueous solution reacted with two moles of CO.
Results of elemental, polarographic, micromanometric, and infrared analysis
lead to the postulation of the following structure:
4H2O
The formation of this complex can be expressed by the reaction:
H20 } [Pd2(CO)2(phen)2] 4H20 + 4HC1
+ 2C02 + 2 phen
2[Pd(phen)2] C12 + 4CO
3-11
WALDEN RESEARCH CORPORATION
-------
Based upon these results, Burianec and Burianova (1963a) describe the quan-
titative detection of carbon monoxide using this reaction. The complex
which is formed is red violet and has an absorption maximum at 550 nm
(Figure 3-2). A translation of the paper of Burianec and Burianova is
given in Appendix I.
Their procedure involves the addition of a solution con-
taining bis(o-phenanthroline)-palladium (II) chloride and hydrazine hydrate
to a round bottom flask. The flask is evacuated and the gas mixture con-
taining CO is allowed to fill the flask. After the reaction has occurred,
the absorbance of the solution is determined versus distilled water at
550 nm. We found difficulties with this procedure, presumably, due to
the time lag between addition of reagents.
Hydrazine is a catalyst in the reaction when carbon monoxide
is present, but has a side reaction with the Pdfa-phenL CU solution when
no CO is present. The palladium (II) hydrazine complex which forms, tends
to polymerize. The addition of ammonia to the flask after the reaction
is complete, is reported to reverse this process. However, the time lag
in the original procedure is sufficiently long enough so that the palladium-
hydrazine reaction proceeds extensively. Then, the addition of ammonia does
not reverse the reaction and the precipitate formed interferes with spec-
trophotometric determination.
This procedure was modified by reversing the addition of
sample and reaction solutions. The carbon monoxide is added to a dry
evacuated flask. Then the reaction solutions are added. The modified
method is outlined in the following section. With this order of addition,
no difficulty is observed in obtaining reproducible blanks and consequently
reproducible CO results. Although this method was reported to be rela-
tively free of interferences by Burianec and Burianova, this aspect will
be investigated further.
A calibration curve was prepared using carbon monoxide-
air mixtures ranging from 0.01% - 0.1% CO by volume. The collection and
3-12 WALDEN RESEARCH CORPORATION
-------
0.6
0.5
UJ
u
i
0.4
0.3
0.2
0.1
I
I
400 425 450 475 500 525
WAVELENGTH (nm)
550 575
600
Figure 3-2. Absorption Spectrum for [Pd2(CO)2(0-phen)2] 4H20
reaction flask is shown in Figure 3-3. The calibration curve (Figure 3-4)
is logarithmic illustrating the departure of the complex from Beer's Law.
The data are given in Table 3-5. The regression equation for the line in
mgCO = e^4'03 abs~4-68'. The data were consistent with a regression yield-
ing a correlation coefficient of 0.966 between the absorbance and mg CO.
To obtain an estimate of the range (ppm), calculations were performed
based on the stoichiometry of the reaction and the known molar concentra-
tion of palladium in the reaction solution. The maximum amount of CO which
is able to react with the Pd (phen)2 C12 is 0.2278 mg. Since the mg CO.
depends on the volume of gas sampled as well as its concentration, adjust-
ment of the volume sampled can extend the useful range of the calibration
curve beyond 1000 ppm. Based on a 50 ml gas sample, the upper limit is
3000 ppm CO. If a smaller volume of gas is used, CO concentrations up
to 10-20% can be used with the same calibration curve. A list of the
sample volume size as a function of CO concentration is given in Table 3-6.
3-13
WALDEN RESEARCH CORPORATION
-------
TEFLON STOPCOCK
250 ml
ROUND BOTTOM
FLASK
1/4" SS UNION
SEPTUM
Figure 3-3. CO Collection and Reaction Flask.
3-14
-------
1.0
o>
0.1
0.01
0.20 0.40 0.60 0.80 1.00
ABSORBANCE
Figure 3-4. Calibration Curve for CO-Pd(phen)Cl2 Method,
3-15
-------
TABLE 3-5
PREPARATION OF CALIBRATION CURVE FOR Pd(0-phen)2 C12 METHOD
mgCO
0.0174
0.0180
0.0228
0.0380
0.0393
0.0430
0.0457
0.0598
0.0810
0.0910
Abs
0.198
0.230
0.230
0.260
0.389
0.447
0.383
0.423
0.496
0.565
mgCO
0.100
0.124
0.128
0.129
0.146
0.164
0.174
0.198
0.210
Abs
0.670
0.644
0.730
0.644
0.630
0.659
0.699
0.779
0.680
TABLE 3-6
SAMPLE SIZE AS A FUNCTION OF CO CONCENTRATIONS
CO Concentration
(ppm)
50
1,000
500
5,000
50,000
Technique Approximate Sample
Value (ml)
Evacuated Flask
Evacuated Flask
Syringe Injection
Syringe Injection
Syringe Injection
150-200**
50
100
10
1
*
Sample absorption flask = 250 ml round bottom
**
Close to the detection limit
3-16 WALDEN RESEARCH CORPORATION
-------
A 108 ppm CO calibration gas was used to study the
accuracy and precision of this method. The average of 113 ppm ob-
tained gives an accuracy of +4.4%. A precision of 12.4% was obtained
for 25 samples.
The applicability of the CO-Pd(o-phen)2 C12 method
to the analysis of stack gases was examined in the Walden pilot
plant.* The unit was fired on gas and doped with CO to levels of
250 to 750 ppm. Samples were collected in aluminized mylar bags
(Figure 3-5a) and analyzed using the wet chemical method for CO.
The results are given in Table 3-7 (#1-8). The CO levels found
experimentally differed considerably from the theoretical levels.
The reagents were checked by analyzing an approximately 0.08% CO
mixture in air; the analysis gave 864 ppm CO. Thus, no problem
existed with the reagents. This suggested that there was an inter-
ference present in the flue gas samples. Since the pilot plant
was fired on gas, there were few possibilities for interferences.
Since the C02 level in the bag was 6.4% (checked) by Orsat) and
C02 is a product of the reaction (below), its presence could shift
the equilibrium and drive the reaction in the reverse direction.
This could account for the low results obtained in the analysis
of the pilot samples.
2[Pd(phen)2]Cl2 + 4CO + 6H20 «-» [Pd2(phen)2] 4H20 + 4HC1
+ 2C02 + 2phen
To check this possibility, the C02 was removed by passing the bag
sample through a prior ascarite tube (Figure 3-5b). CO samples
*
This is described in Section 5.2.
3-17
WALDEN RESEARCH CORPORATION
-------
WATER
KNOCKOUT
TRAP
ALUMINIZED
MYLAR
SAMPLING BAG
FLUE
GAS
MB21 METAL
BELLOWS PUMP
a. Collection of Flue Gas Sample
ASCARITE CARTRIDGE
FOR C02 REMOVAL
STOPCOCK
GLASS
CO REACTION
FLASK
b. Transfer of Flue Sample to CO Reaction Flask
Figure 3-5. Collection and Transfer of CO Samples,
3-18
-------
TABLE 3-7
ANALYSIS OF CO IN FLUE GASES USING CO-Pd(0-phen)2 C12 METHOD
Sample
1
2
3
4
5
6
7
8
la
2a
3a
4a
5a
6a
7a
8a
9a
lOa
lla
12a
ppm CO
114
76
54
107
77
61
62
85
386
499
480
Aver. 455 +45
106
122
125
118 +8
349
276
351
329 +31
406
414
420
413 +5
Theoretical ppm CO
750
750
250
250
250
250
250
250
500
500
500
180
180
180
280
280
280
380
380
380
Remarks
Without Ascarite
Tube
With Ascarite
Tube
3-19 WALDEN RESEARCH CORPORATION
-------
collected from the pilot plant incorporating the ascarite tube
into the analytic sampling line are given in Table 3-7 (la - 12a).
Good agreement is now obtained between theoretical and measured
ppm CO. The losses of CO on ascarite were examined by analyzing
pure CO mixtures (no C02) with and without the sorbent tube in
the sampling line. No significant differences were found. Thus,
the ascarite did not affect the CO results but did remove the C02
interference from the analytical method.
Statistical analysis of the CO samples in Table
3-7 with the ascarite tube gives the following:
ppm CO[Pd(o-phen) C12] = 1.02 ppm (theor. CO) - 12.9
The modified procedure, used to check the
reliability of the instruments in the field, is given in the
following section.
3.3.3 Procedure for CO Analysis - Palladium Chloride-
o-Phenanthroline Method
3.3.3.1 Sample Collection
(a) Set up the equipment as shown in
Figure 3-5a.
(b) Before placing the 30 liter sampling
bag in position set the flow rate at
0.05 liters per minute.
(c) Start the pump and collect the sample
over two thirty to forty minute periods,
(d) Then shut off pump, seal the sampling
bag.
3-20 WALDEN RESEARCH CORPORATION
-------
3.3.3.2 Sample Analysis
(a) Evacuate a 250 ml round-bottomed flask fitted
with two-way stopcock (Figure 3-3); record barometric pressure and tempera-
ture. Record AHg.
(b) Partially fill flask (min AHg is 100-150 mm)
with gas mixture to be analyzed.* Record AHg.
(c) Carefully add 10 ml bis-(o-phenanthroline)-
palladium (Il)-chloride solution, allowing no air to enter flask. Add
0.1 ml 6.3M hydrazine hydrate solution. Shake vigorously for ten minutes
on a wrist action shaker.
(d) Evacuate flask and fill with air to atmos-
pheric pressure. Add 0.2 ml of concentrated ammonium hydroxide (d = 0.91).
(e) After ten minutes, read absorbance of the
sample at 550 nm with a colorimeter (H20 reference cell).
3.3.3.3 Blank Preparation
Follow steps (a) through (e) with the following
exception: in Step (b) partially fill flask with air.
3.3.3.4 Calculations
Vf(Pab-P1)(293.3)
Vc 760 (Tr)
nnm TO - 24.1 X VI X 1 Q
ppm CO -- ~
*
For high concentrations of CO, inject syringe filled with sample at
this point.
3-21 WALDEN RESEARCH CORPORATION
-------
V = volume of gas sampled
Vf = volume of flask = 289.3 ml and stopcock
P . = barometric pressure - AHgf
P. = barometric pressure - AHg.
W = weight of CO in mg from calibration curve
T = room temperature, °K
24.1 = standard molar volume, 760 mm Hg and 70°F
28 = molecular weight of CO
3.3.3.5 Calibration Curve Preparation
(a) Prepare grab bags in following way:
1. 0.1% - 10 ml CO in 10 liters air
2. 0.08% - 8 ml CO in 10 liters air
3. 0.06% - 6 ml CO in 10 liters air
4. 0.04% - 4 ml CO in 10 liters air
5. 0.02% - 4 ml CO in 20 liters air
6. 0.01% - 2 ml CO in 20 liters air
(b) Analyze gas mixtures as described in Section 2.
(c) Make a calibration curve by plotting mg CO
vs absorbance.
(d) Calculations
1. Calculate V as in Section 4.
2. mg CO = e'4'03 abs-4'68>
Field Test Data
The Pd(o-phen) wet chemical method for carbon
monoxide was evaluated and tested at an incinerator, a gray iron foundry,
3-22 WALDEN RESEARCH CORPORATION
-------
and a catalytic cracking unit (inlet and outlet of CO boiler) in a pet-
roleum refinery. The results are discussed in Section 6.
3-23 WALDEN RESEARCH CORPORATION
-------
4. STATE-OF-THE-ART SURVEY FOR HYDROCARBON AND CARBON MONOXIDE MONITORING
INSTRUMENTATION
4.1 INTRODUCTION
A state-of-the-art survey was conducted by searching the current
literature. This was supplemented by a search of manufacturers of CO and
HC monitors in Science "Guide to Scientific Instruments" and the ACS Lab-
oratory Instrumentation Issue. The literature was requested from all man-
ufacturers of instruments. Those instruments applicable to continuous
monitoring were evaluated in detail and the specifications were recorded.
A summary of this data is given in Tables 4-1 and 4-2 for hydrocarbons
and carbon monoxide respectively. For those techniques, in common use,
we have discussed the theory, as well as the advantages and disadvantages
of each principle in Sections 4.2 and 4.3.
4.2 HYDROCARBON MONITORS
4.2.1 Flame lonization
4.2.1.1 Principle of Operation
When a volatile carbon compound is introduced
into a hydrogen flame, ions are produced. The process leading to ion pro-
duction is obscure but it is believed to be due to chemi-ionization. If
a positive potential (^lOOV) is applied to the burner housing, the ions
are accelerated toward the collector electrode which is directly above
the flame. The ion current that is produced is proportional to the con-
centration of the species added. The signal (ion current) is determined
with a conventional electrometer. A schematic of a flame ionization
detector is shown in Figure 4-1. The characteristics of an FID are given
in Table 4-3. The detector responds to all organic compounds except
formic acid. In general, the response is greatest with hydrocarbons and
diminishes with substitution of elements such as N, 0, and halogens.
Solids in the flame tend to cause noise and anomalous response. The
response factors of many of the hydrocarbons are within approximately
5% of each other [Dietz (1967)].
4-1 WALDEN RESEARCH CORPORATION
-------
UJ
1
II
21
i
«j^
CO
s
§
o
X
UJ
1 1
1
J
i^£
^H
g__^
^E5
^i
LU
u
UJ
i
CO
u.
o
CO
o
CJ
1
u_
1 1
o
UJ
f\
at
i- o
CO
ai i
«j
o
o
i/i
r-
E
c
o
^
+J
Q
0)
*Q.
U
c
t
i.
CL.
0)
1
>,
t
5E
r O
CT>CM CM
XXX
00 Cn O
r-r t
XXX
. in r-.
CM t r
LOO O
LO LO in
i CM
OO O
en o LO
CO O CO
i in r>»
«*«*
55
O. CL
ssi
oo o
1 1 1
i < O
.c
"«- x
o
*-> >>
u j=
o> 0.
M *
c5 cr
o
u 10
** £
ID O
CO i-
fc-* g *
O O 10 O
1-1 I-H IB >-i
U. U- CO LL.
§CM O
o oo
c
1
u
o
CM
X
*
X
0
CM
0
LO
CM
8
en
VO
z* *!
O 3C
o
H E
Q. S.
Q.
5 °
Q. O O
CL O O
0 0
o o o
i i CM
1 1 1
o o o
.c
+J
'x
>>
1^
Q.
n)
1-
o
f^%
N LO
o co vo
O »- i
CM t-
OO U.
x
1- IO
1 s
CO 0
LO
X
o
X
CM
"
o
CM
0
in
>^
^~
IO
5
1
IO
o <*-
1-1 C
U- >-.
O CM
0
LO CO
i «j o
CO CC J-^
CM =3 LL.
.C
o eli
1 1
CM O
x x
CM CO
X X
cyi r*v
1 VO
in o LO
CO 171 00
*J r
IO OI
C O
Z
o
i.
0-
1
o
»
O) (J
0 i-
S. 4J
^ S
"«9 Q. S-
3 O
O CO +>
r- W
of en «i
0>-i c 4J
1-1 O i- 01
Lu Z LO O
§
CM
3^
IO
v^> ^
Z t-
1 1
.
=1
IO
X CO
LU N
^)
4J *IO
«C«t
VO
00 0»r-
X XX
CT> CM CO
r CM
X XX
CM VO CO
CM
LO 00 O LO
CM VO CO CM
'
0 0
r^ LO
VO CM
r- CM
O) CO
C C
(0 (Q
X X
oi co
Z Z C
1 1 O
s_
IO IO IO
IO -> Q >-!
O 0 >-i O
Z Z U. Z
o
o o
CO *S. O
CM CM CO
LU * t I-H CO
<:u.o
J
_
10
J3
O
1
C
o
VI
0
^~
CM
X
oo
X
CO
"
VO
LO
0)
c
IO
X
0)
Q-
Q.
S
in
o
a
>^_
on
o
o
o
.
10
c
1-^
I/I
I/I
01
a.
4-2
WALDEN RESEARCH CORPORATION
-------
VI
C *"»
O X
r- 0
V» S.
C CX
9) 0.
E «J
r «*
O
O> VI
i- .0
0) i
to
O
O
V)
+J
^
£
««j
C
o
.4^
U
r^
X X
0^ 00
CM IT) CM
to PO ro
g 8
5 *""
OJ
c
(O
X
o z i/i
o >
* la
E -4^
«- o u c
^c ^^ o o
O U S t-
01 C 4^
o Q vi +> w w> a
i >-i « a> -c 3 >-<
U. U- C3 O t ^3 U.
O CM
O CM
CM CM U-
U9 1 1 O
u> o vo in
i CO CM CO
1 I
ai CM CM
XXX X
00 Cfi 0^ ^~
rti i ^r
SO LO
u) r--
"*
oo
85
LO
CM Si
x>«
O O E
r r CX
CX
O O
E E §
^L ex *
CX CX r-
1
I r O
^=
**
"5
>,
CX
->« * « US i«
U- U. U- UJ U.
CO O
LO IO LO CM
r- < i O
i i CM m
j:
o at
«- TS
to u s_
1U C !
Cc II tj
* C
«-> VI O
C/4_J =>
4-3
WALDEN RESEARCH CORPORATION
-------
_
c
^ *
I/I
O
tn
£
E
O
OO
OC.
o
z *
O 0)
'5
UJ *
o
£ *
?^ O *-*
O t. VI
2 o- o
1 *"
0 5"
GO "? '*""
2r* **"* >r"
§. vi
... 01 C
UJ rv QJ
_I CO
CO
1
r.
01 E
5*" 01 _i
_l 0
5!
1 1
o
OL
LU
S
o
uj r°-
§'C
c
1/0 ct
U_
O
i
» i
0
t <
Lu
HH QJ
O -o
UJ 0
a. ^
^
C
S.
g
o
as
a
z
c
01
o
CD
t-
o
3
It.
U
o.
s
+J
o
Ul
XI
a
5
c
o
i.
0
CM
X
o
X
r-.
O
in
CM
o
ro
a.
o
o
i *
o
o
°s
t-
c c
O i-
o at
-n 01
s-°
0?
r- ro
+J
ro O <
S
s
1
Q)
ca
IT
f * ^
CXi^
iW O
^"^~~*
00
X
00
X X
en o
o o
to
i.
M-
O
X
+ 1
i,
^-T3
"°.JS
01 'r-
'ia t.
3 0
ae 01
o 01
z-o
S
L.
Ol
i»I
"5.1^
E Ol
fl) U
^j ^.^*^^
r 00 3 LO CMOOO
CM O l/> Ct> r r r . r in
X X X XX XXXXX
CM in o o cocn CM CM *j- o r-.
k "~ F~ *~ *~ "~
X X _ X XX XXXXX
oo 10 en CM oicn mcnr^-mo
in
o o o o CM O to ro r-I
in us CM r CM r-. rv cxi
o o o co
§in in fo
en . r-« ro
r>- CM
CO
CO OO CO CO CO OO
M- . . .
Lu Ll-U.OU-Ll.Ll.
X &< X X* X V9 X
+ 1 +1 +I+I+I-H+I
o o
E E E 0 0 0 E
O.B.O. o t_) o O 3.
o. o. o. xxoo ooo.
in in o M
O O O CMCMXo'^OoS
r «» r r r r rOr r m
1 1 1 II 1 1 1 t 1 1
O O O OOOOOOOO
-pi ai v.
t- 01 i- ex r o
oc ai i r-~ cc en *J cu M- at
r- Ei-ro. oo c c v> C fli" Q. 0 m -t-> 0 ro
ai -o -a c « 0) « oi u o 01 >ai3iejcai
> Ocon-r- -oS-DS i-l- i E+J-Q-a-c
oc o. x r x o - cxoi-ooxci*.
UIO l»- 0) M- O-r-O-r- VI e Q 43 U O 1C O
o-aoac <*- n_ EOI
U4- -r- O O O r fc. i.r CUUCMC
r-I CXC Z -"-E'r-E 34->3 3MO -C 3 jS <-> +>U-r-OI
>, Q -r- -r- -.^ >, C >) C -OC4-> »-> -OO->> >> <->O
r Ol-l+JU** r -r- r -r- Ol U Or r O (J C
ro O X OI3->- a ro i ro r Q o Ol O Q-r-ra iqOOIra
O o ja etuE z u O.U a. z o -o z z E <-J O o -o J3
VI
So a oi
LO IO r
UJ O O i O i- i-
_l r i ro O +J QJ Q
§»» X CM r CM VI N
CM LU II r 3 >ir
O «f 1 CO r "t r» J3 r 01
O Z X «t r Ct +J E ""O
CM tO -^ O £ tO 1 O O C 0
oo => ex. t- r>m _it-u -o -M aico
Ol ro S ro i. c O)
ca m CD o LU - _ix
4-4
WALDEN RESEARCH CORPORATION
-------
*
M
8
cvj
i
LU
_l
§
1
C
^
§
I/I
C
i
r
a
^
Ol
o>
1/1
o
o
>,
lj;>
$- +J
at c
OL at
oo
c
o
r- VI
si
&
0)
a.
r-
O
C
F-
<\
0»
|
!
.
o in to
l~~ CM OO r
X X X X X
r*. r** O O ro
r CM
X X X X X
OCM OOCM *
r r CM
i.
01
r^*Q.
01 E
U 10
o co tom in
* CM r CM CM
^
in o o
co P»» in
oo in CM
r CM
.
OO 1/5 OO tO
U- LL! LL. u.
MM M M
+ tfi ₯l Ti
88 S S
CMX M O
r in in r
II 1 1
oo o o
ai ai a)
c c c
E-i 5 5
10 o. a. Q.
at o o o
.01- L. J_
u u u
m g t_ C L_ c c_
a o> *> ai *> a> +j
r Or Ui U
c£ o> v oi ai o> ai
g""*-!- 01 1- Ol-l^ "w
tO TJ IOT3 tO TO
S § o
ro ro O
CM i i ro
sl 1 i
2
^
U
f
c .
o
VI
o
^. ^
CM CM
X X
0009
X X
coro
""
10 VO
in in
f
to
U-
^
?l
in
o
i.
o
a!
c 4^
5 a>
ja
i_
S5
3 VI
at-
>E
^ ai
X -M
>-< "(O
33
§§
.
(A
c
» i
Peerless
Co.
>^
a.
0
2
Q
J=
U
VI
3
CM
CM
CM
VI
N
1
VI
VI
a*
^
X
rx.
r>
X
o>
CO
S
00
LU
O
^>
?l
0
g
o
g
ia
c
c o
O i-
) * o
(O 01
o *->
^ 01
X T3
O
Ol
O r-
X O
r E
lO 3L
O *>
CM
8
§
to
.
1
'
r Ol
X X
o at
X X
0 01
CM
in
co
o in
in 3
t
to
ul
M
T"i
g
a.
in
oS
t.
0
cv»-> *
SO 3C Q
ai o »
+J U.
00) O
+J T3 +» >j
8,5 8-g
<*- Q. if- C
vi C vi E
i- oj - Ol
>>4-> >. 01
'n jc "a
4J *» 4^*0
51- « C
X u 10
41 U.
a o
S in
O ro
Q.
5 ,
VI fc.
u «
tu S
C Of
0 i.
E U 4-> VI
0) « O J3
C k 0 10
r- »- |X_1
US
CM
X
CO
CM
X
r
*
S
CM
8
o
u
K
a.
o
8
^.
zo
U I
u.
0
f^
°"S
» Q|
VI *J
r^" *
^1
o a
O
CM
O
co
T3
a
|
"c
4-5
WALDEN RESEARCH CORPORATION
-------
Removable vent
Platinum
wire igniter
Iff ,J V '-- -~~Tr
^hLle
Air diffuser
inltt
collector
Polarizing
electrode
and igniter
Hydrogen
Air or
V oxygen inlet
Sample inlet
a. Detector
Electrode Gap
Ionizing Source
I AAA/WH
nun
Duckinc Voltage
Recorder
b. Schematic lonization Detector Circuit
Figure 4-1. Schematic of a Flame lonization Detector
4-6
WALDEN RESEARCH CORPORATION
-------
TABLE 4-3
CHARACTERISTICS OF A FLAME IONIZATION DETECTOR
-5
lonization Efficiency 10
Linear Dynamic Range 10
Noise Level (amps) 3 x 10"
-12
Minimum Detectable Quantity (g/sec) 3 x 10"
(as C3H8)
Minimum Detectable Quantity by Volume 10'11
Substances Detectable All Organics
4.2.1.2 Commercial Equipment
Beckman, Bendix, Scott, Intertech, etc. See
Table 4-1 for specifications of flame ionization monitors.
4.2.1.3 Discussion
Advantages
(a) Linear response over six decades
(b) Responds to low concentrations of HC
(c) No interference from hUO, air, or inorganic
gases
(d) Sensitive to a wide variety of hydrocarbon
classes
(e) Output linear with concentration
(f) Relatively inexpensive and reliable instru-
ments are available
Disadvantages
(a) Flow rate dependent
(b) Output is related to the number of carbon
atoms
4-7 WALDEN RESEARCH CORPORATION
-------
(c) Supply of hydrogen and air is necessary
(d) A good filter is required since particu-
lates may lead to capillary plugging or
noise in the flame
(e) It responds to methane as well as other
hydrocarbons
4.2.1.4 Conclusions
This principle is the most useful technique for
hydrocarbon monitoring, however, it is not useful for "reactive" (non-
methane) HC. It is possible to use a selective combustor cycled between
off (total HC) and combustion of all hydrocarbons except CH. in front of
an FID. This then gives total HC and CH^ with the "reactive" HC being
obtained by difference (see Section 6).
4.2.2 Gas Chromatography (Hydrocarbons)
4.2.2.1 Principle of Operation
A sample of air containing the hydrocarbon is
injected, via a carrier gas, directly into the flame ionization detector
to determine the total hydrocarbon concentration. A second sample is in-
jected (within M minute), passed through a stripper column (porapak) to
remove traces of C02> water, and high molecular weight hydrocarbons. The
methane which passes the stripper column is separated from CO on a 5A molec-
ular sieve column, and detected with a flame ionization detector (Fig. 4-2).
The non-methane hydrocarbons are determined by the difference between
CH. and THC. Due to the elution time of the methane (2 to 3 minutes),
this technique is a semi-continuous and not a real time monitor.
4.2.2.2 Commercial Availability
Beckman, Bendix, Tracer, etc.
4.2.2.3 Discussion
Advantages
(a) Non-methane hydrocarbon concentrations can
be determined
4-8 WALDEN RESEARCH CORPORATION
-------
UJ
LU
X
GO
D
o
<
m
CT>
UD
cn
ITJ
0)
»
c
0)
a;
1-
o
o
s:
o
o
o
CM
I
Ol
CD
4-9
WALDEN RESEARCH CORPORATION
-------
(b) Carbon monoxide can be determined with the
same instrument
(c) Broad dynamic range (3-4 orders of magnitude)
Disadvantages
(a) Not a real time monitor
(b) Instruments are not portable and are very
complex
4.2.2.4 Conclusions
At the present time, the process gas chromato-
graphs provide the only commercially available monitors which can deter-
mine non-methane hydrocarbons with any reliability, however, the semi-
continuous monitoring capability can lead to problems on a non-steady
state source (see Section 6).
4.2.3 Selective Combustion
A variety of approaches are described which utilize either
catalytic or high temperature oxidation of a HC sample to produce a change
in temperature, a combustion product (H20) which can be measured with a
simple detector or partial oxidation and, therefore, response to different
classes of HC species. These techniques with advantages and disadvantages
are described below.
4.2.3.1 Flame lonization Detection
(a) Principle of Operation - A sample is fed
through a combustor which consists of a coiled platinum wire in a quartz
tube, and into a flame ionization detector (FID). As the temperature of
the combustor is increased (by increasing current), one can quantitatively
burn the olefinic fraction (250-500°C), and the paraffinic fraction (1000°C)
with a small loss of CH4. Thus, by cycling on (T ^ 20°C), one can obtain
methane and total HC; "reactive" hydrocarbons can then be calculated from
the difference between the two readings (see Section 6.0).
4-10 WALDEN RESEARCH CORPORATION
-------
(b) Commercial Equipment - The Gow Mac selective
combustor (can be coupled to many commercial FID's).
(c) Discussion
Advantages
(1) Same as those of FID
(2) Selective combustor can be readily
adapted for use with many FID's
(3) Can determine reactive (non-methane)
hydrocarbons with this device
4.2.3.2 Piezoelectric Crystal
(a) Principle of Operation - The sample gas is
passed through a cartridge containing Dowex 50 WX8 (Na ) ion exchange
resin to remove water and quantitatively pass the hydrocarbon fraction.
The dry hydrocarbons are combusted in the selective combustor and the
water formed is detected with a piezoelectric crystal coated with a
hydroscopic polymer. In practice, the device consists of two piezo-
electric coated crystals. Uncombusted air was passed over one and the
combusted sample was passed over the second. The difference in fre-
quency is then proportional to concentration.
(b) Commercial Equipment - None available.
(c) Discussion
Advantages
(1) No H2 or air supply needed for
operation
(2) Not flow sensitive like FID
4_11 WALDEN RESEARCH CORPORATION
-------
(3) Both the device and the electronics are
very simple
(4) Can be calibrated using electrolytic
generation of hydrogen
Disadvantages
(1) Slow response ^5 minutes
(2) High MW hydrocarbons cause tailing;
therefore, not necessarily quantitative
(3) Gives total H^O; or other atmospheric
Hg containing species will be an inter-
ference. CM* would give twice the re-
sponse of C^H?. The paraffinic hydro-
carbons, which are less photochemically
reactive than olefinic HC, give more
response
(4) Not enough data on correlation with FID
4.2.3.3 C02 Detection
(a) Principle of Operation - A sample is passed
through a (XL trap (possibly ascarite), a selective combustor, and into a
C(L analyzer. Total hydrocarbons and methane can be determined by pro-
gramming the selective combustor temperature between room temperature (THC)
and 1000°C (CH^). The combusted HC would be determined with a NDIR or
perhaps a more sensitive CCu analyzer such as reduction to CH. and detec-
tion with a FID. It would be difficult, however, to scrub out flue gas
containing M0% CO,, then detect MO ppm C02 with an NDIR.
(b) Commercial Equipment - Not available.
(c) Discussion
Advantages
(1) Total weight % carbon is determined
(2) Non-methane hydrocarbons can be determined
4-12 WALDEN RESEARCH CORPORATION
-------
Disadvantages
(1) No sensitive continuous CO- analyzers
are available
(2) C0£ in the flue gas has to be quanti-
tatively removed
4.2.3.4 Temperature Rise
(a) Principle of Operation - The sample is passed
into a heated (275°C) vanadia catalyst where the hydrocarbons are combusted
to CCL. The carbon monoxide in the sample remains unchanged. The resultant
temperature rise is measured with a thermocouple and fed into a differential
amplifier. By varying the temperature of the catalyst, different classes
(olefins, paraffins, etc.) of compounds can be determined. The temperature
rise is a function of the amount oxidized, the molar heat of oxidation, flow
rate.
(b) Commercial Equipment - Available from Purad
Corporation.
(c) Discussion
Advantages
(1) Ability to differentiate between meth-
ane and non-methane hydrocarbons
(2) Simple and inexpensive equipment
(3) No hydrogen or air supply is necessary
(4) Unit can apparently be correlated either
with NDIR or FID by varying catalyst
temperature
Disadvantages
(.1) This is a batch type rather than a
continuous analyzer
(2) Would not be very useful with unknown
streams since the temperature rise is
sensitive to oxidation efficiency, etc.
4-13 WALDEN RESEARCH CORPORATION
-------
(3) It is necessary to supply a water sat-
urated air stream to the detector
4.2.3.5 Conclusions
Selective combustion appears to be a very promis-
ing approach for the determination of non-methane hydrocarbons. The only
other technique available at the present time involves a complex and ex-
pensive process gas chromatograph. The most useful process involves
coupling a selective combustor to a flame ionization monitor. The piezo-
electric and temperature rise techniques are of interest only because of
their simplicity and portability.
4.2.4 UV Absorption
4.2.4.1 Principle of Operation
Aromatic hydrocarbons absorb strongly in the
near ultraviolet (2000-40008) due to electronic transitions of their pi
electrons. Since these TT* *- TT transitions are highly allowed, by sym-
metry considerations, the absorption coefficients are extremely high
(e ^ 10-10 Jl/mole-cm). Aliphatic hydrocarbons (CH., C2Hg, C3Hg, etc.)
which have only sigma electrons do not absorb appreciably in this region.
This technique may be useful for monitoring cer-
tain aromatic hydrocarbons since sensitivities (Table 4-4) vary from
several ppm to ^300 ppm full scale for aromatic hydrocarbons. Aliphatic
hydrocarbons which absorb weakly can be detected from ^2000 ppm to M% FS.
4.2.4.2 Commercial Equipment
Dupont Model 400 photometric analyzer
Beckman UV analyzer
Honeywell hydrocarbon monitor
4.2.4.3 Discussion
Advantages
(1) No interference from COp, CO, water vapor
. . . WALDEN RESEARCH CORPORATION
-------
TABLE 4-4
REPORTED HYDROCARBON SENSITIVITIES WITH A
DUPONT 400 ANALYZER*
Compound
Aniline
Benzaldehyde
1 ,3 Butadiene
Furfural
Nitrobenzene
Styrene
Tetrachl oroethyl ene
Toluene
p-Xylene
Acetal
Acetone
Dioxane
Formic Acid
i-amyl alcohol
ppm Full Scale
33
3.9
67
1.6
3.0
8.2
270
185
150
100,000
2,200
30,000
40,000
5,000
*
Dupont Bulletin 5A
4-15 WALDEN RESEARCH CORPORATION
-------
(2) Some equipment is relatively simple and
inexpensive
(3) Detects many aromatic compounds which are
used as solvents
Disadvantages
(1) Does not give uniform response to all hydro-
carbons
(2) Interference from SCL, NCL, Hg - not very
specific
(3) Major use will be in process streams where
the composition is known
4.2.4.4 Conclusions
This principle may be useful for monitoring solvent
emissions from small processes where an inexpensive monitor for hydrocarbons
is required.
4.2.5 Non-Dispersive Infrared
4.2.5.1 Principle of Operation
Hydrocarbons absorb infrared radiation in the re-
gion of 3y. The non-dispersive infrared analyzer relies on the selective
absorption or transmission radiation at a particular frequency. The radia-
tion passing through the sample cell is attenuated by the presence of
hydrocarbons in the sample. The reference side of the detector thereby
receives more IR radiation than the sample side and the gas in that com-
partment absorbs more heat and expands. The pressure imbalance between
the sides of the detector is sensed by a metal diaphragm capacitance sen-
sor and is proportional to the amount of HC in the sample. There are a
number of variations on this technique in the devices of the various manu-
facturers.
4_15 WALDEN RESEARCH CORPORATION
-------
4.2.5.2 Commercial Availability
Intertech, etc. See Table 4-1.
4.2.5.3 Discussion
Advantages
(a) Simple instrumentation with few
moving parts
(b) Portable
Disadvantages
(a) Requires considerable sample
treatment - very susceptible to
water vapor interference
(b) Does not correlate with FID mon-
itors
(c) Responds differently to HC -
response depends on the IR ab-
sorption coefficient not on the
structure of the HC
4.2.5.4 Conclusions
This concept would be most useful for monitoring
a single known component in a process stream. It would not be very use-
ful with a varying mixture of hydrocarbons.
4.2.6 Laser Spectroscopy
4.2.6.1 Principle of Operation
A monochromatic tuned infrared diode laser has
been used to selectively determine the concentration of various hydrocar-
bons (Hinckley and Kelly, 1971). The lasers can be fine tuned by changing
the diode current which alters the dielectric constant of the mixed crystal
A diode laser can be tuned over 40 cm" or continuous bands up to 2 cm
wide. This broad tuning should prove adequate for pollution measurements.
4-17 WALDEN RESEARCH CORPORATION
-------
A variety of spectra regions can be used depending on the composition of
the crystal (see Figure 4-3).
4.2.6.2 Commercial Availability
Not available.
4.2.6.3 Discussion
Advantages
(a) Can selectively determine a single
hydrocarbon in a mixture
(b) Can determine the average concen-
tration across the stack
(c) A number of other pollutants (CO,
CO , SO , NO, etc.) can be mea-
sured although not simultaneously
Disadvantages
(a) Complex equipment and excessive
cooling requirements (to ^°K)
for detector
(b) More of a research tool than a
routine monitor
4.2.6.4 Conclusions
This technique should prove to be a useful re-
search tool but considerable engineering improvements will be necessary
to simplify the equipment.
4-18 WALDEN RESEARCH CORPORATION
-------
cvj
*°0
1
O
tr
1 i 1
LU
WAVENUMB
O4
tn
O
1
° o x \ o^"
Z CO Z I o
. . <
_ CO Z - \
_ cvj < ro I
-r O cvj cvj / ~.
o> A f ZOO/2
^x XT I 2 CO /?_
C |_ Q_
CO x
x c A T
' to I j
° f « T T -
.cT WK 1
°- i i ^ n- r~ "
~~ ^1 0 0 X iL~
£ 0 X 0 X
^ 2 1 T ~
x < 1
- r,** < «* i
tf) O ^ ^
^ j. <7" «r J.
c o < »
-1 x 0 «)
< ^> cox
T3
o
O
O
^r-
1
>-
X
* fN
1
WAVELENC
6
01
(/)
(O
S-
o
-M
o
3
T3
O
O
s_
o
I/)
c
o
CD
OJ
o:
to
s-
4->
O
0)
a.
to
a
CO
i
a>
3
Ol
4-19
WALDEN RESEARCH CORPORATION
-------
4.3 CARBON MONOXIDE MONITORS
4.3.1 Non-Dispersive Infrared
4.3.1.1 Principle of Operation
Carbon monoxide absorbs radiation at about 4.3y.
Measurement of this radiation absorbed can be made by wavelength disper-
sion or non-dispersive analyzers. The non-dispersive analyzer relies on
the selective absorption or transmission of radiation at a particular
frequency to make the measurement selection. Several types of units are
currently commercially available. One relatively recent unit uses a cir-
cularly available interference filter to select a particular wavelength
of IR radiation. Similarly fixed wavelength bandpass filters are also
used in commercial units. The original non-dispersive units and the
units which are generally thought of when "NDIR" is mentioned are based
on the original work of Luft in Germany in the late 1930's. These units
in several forms and levels of selectivity use the presence of the gas
molecule itself to provide the wavelength selectivity. One such unit
produced by Beckman Instruments uses a two chambered detector filled
with CO. One chamber receives unattenuated energy through a reference
cell while the other cell is filled with sample gas.
The radiation passing through this cell is at-
tenuated by the presence of CO in the sample. The reference side of the
detector thereby receives more IR radiation than the sample side and the
gas in that compartment absorbs more heat and expands. The pressure im-
balance between sides of the detector is sensed by a metal diaphragm
capacitance sensor and is proportional to the amount of CO in the sample
cell. Variations on this technique involve removal of IR radiation by
interfering gases.
A number of different non-dispersive infrared
analyzers are available, e.g., those based on the Luft (Intartech, Beck-
man) dual beam - dual detector chopped unit, the Lira (MSA) dual beam-
single detector, and negative filtration (L&N). The Luft type analyzers
WALDEN RESEARCH CORPORATION
-------
have been described above and are shown in Figure 4-4. The description
of the negative filtering analyzer is given below.
In the L&N instrument (negative filtration), a
single source and single cell. At the end of the sample compartment are
two cells which contain a non-infrared absorbing gas and carbon monoxide,
respectively. The amount of energy absorbed in the CO cell is measured
with a thermopile. As the concentration of CO increases, the (non-infrared
absorbing) reference signal decreases, thus negative filtration. This
concept is not useful for measuring very low concentrations of CO < 200
ppm.
4.3.1.2 Commercial Availability
See Table 4-2.
4.3.1.3 Discussion
Advantages
(a) Very simple instrumentation with few
moving parts
(b) Many of the NDIR's are portable
(c) Continuous monitoring capability is pro-
vided with rapid (30 sec) to slow
(3 minutes) response times
Disadvantages
(a) Requires considerable sample treatment
4.3.1.4 Conclusions
NDIR analyzers are inexpensive, reliable, and
easy to maintain monitors for carbon monoxide.
WALDEN RESEARCH CORPORATION
-------
I I /
IR radidtiun source^. ] /! I
"^W^J Q '~*>' -
c --^^'»-----< -
ce//
Hrlerence Lull
Detection chiirnbtir
Mvnihranu capacitor
T
't/i rriutor
IR radiation
' source
' Diaphragm disk
; Filter cell
Sample cell
Detection chamber
Amplifier
Indicator device
a. Intertech URAS
b. Leeds and Northrup
Figure 4-4. Schematic Diagram of NDIR Analyzers.
4-22
WALDEN RESEARCH CORPORATION
-------
4.3.2 Selective Combustion
4.3.2.1 Principle of Operation
The sample is passed through a charcoal column
(to remove traces of hydrocarbons) and over a hopcalite catalyst which
oxidizes the carbon monoxide to CCL- The resultant change in the temp-
erature of the catalyst is determined with a thermocouple and a differen-
tial amplifier. HC and CO can be detected in the same instrument by split-
ting the sample stream and reducing the flow to the CO sensor. Thus, the
same amplifier readout is used for both instruments by electrical switching.
4.3.2.2 Commercial Availability
Purad Corporation
4.3.2.3 Discussion
Advantages
(a) Inexpensive and portable
(b) HC can be done with the same instrument
Disadvantages
(a) HC have to be removed on a charcoal scrub-
ber, it would be difficult to determine
when the charcoal is used up
(b) Hopcalite catalyst has to be replaced
(c) Best suited for high CO concentrations
(d) Carrier gas has to be saturated with
water vapor
4.3.3 Gas Chromatography (CO)
4.3.3.1 Principle of Operation
A sample is passed through a precolumn (porapak)
to remove water, C0~, and high molecular weight hydrocarbons, and onto an
4-23
WALDEN RESEARCH CORPORATION
-------
analytical column of 5A molecular sieve column where it is separated from
CH.. The methane is eluted first in approximately two minutes. The CO
passes from the column (approximately 3 minutes) into a catalytic (nickel)
reduction chamber where it is reduced to CH,, and is detected with a
flame ionization detector (see Figure 4-2). A more detailed description
is given in Section 4.2.
4.3.3.2 Commercial Availability
Beckman, Bendix, Tracer, etc.
4.3.3.3 Discussion
Advantages
(a) Hydrocarbons can be determined with the same
instrument
(b) Broad dynamic range (3-4 orders of magnitude)
(c) Low levels of CO can be measured
Disadvantages
(a) Not portable (weighs approximately 250
pounds)
(b) Is a semi-continuous monitor for CO (5-10
minutes response time)
(c) Complex instrument requires trained personnel
to set up and make periodic adjustments
(d) A supply of "hydrocarbon free" compressed air
and hydrogen are needed
4.3.4 NDIR With Fluorescence Source
4.3.4.1 Principle of Operation
A blackbody source (heated filament) is used to
16 18
excite the fluorescence of a low pressure mixture of CO and CO con-
tained in a rotating cell. When the cell is rotated by 90°, the blackbody
source is blocked and the excited CO emission becomes the light source for
4_24 WALDEN RESEARCH CORPORATION
-------
16* 18*
the analyzer (see Figure 4-5). The CO and CO radiation source is
I C 1 Q
alternately chopped by cells containing either CO or CO . The detector
then sees CO radiation passed by the CO cell and CO radiation passed
18 16
by the CO cell. If CO is present in the sample cell, the beam atten-
16 18
uation is proportional to the CO concentration. Since the CO abundance
in the atmosphere is less than 0.5%, this signal acts as a zero and elimi-
nates the need for filter cells.
4.3.4.2 Commercial Availability
Available from Andros Scientific Co., California.
4.3.4.3 Discussion
Advantages
(a) Higher selectivity for CO is obtained.
(b) Filter cells and water knock traps are not
needed.
(c) Particulate matter will not be a major
problem because both CO^6 and CO^ inten-
sities will diminish at the same time.
Disadvantages
(a) PbSe detector requires cooling to liquid N£
temperatures for adequate sensitivity, how-
ever, for high CO concentrations, thermopiles
or bolometers may be available.
4.3.4.4 Conclusions
This analyzer should prove to be very useful for
source monitoring because it eliminates filter cells or interference fil-
ters, and the water knock out trap.
4-25 WALDEN RESEARCH CORPORATION
-------
U oo
S i
OL A
It
UJ
oc
o
&
UJ
Q
111
s-
O
cc
CD
4-26
-------
4,3.5 Laser Spectroscopy
See Section 4.2.6.
4.3.6 Mercury Replacement
4.3.6.1 Principle of Operation
Carbon monoxide reacts with mercuric oxide at
200°C according to:
C0(g) + HgO(s) *- CO, + Hg(g)
A e-
The mercury vapor can be determined by measuring the absorbance at 253.7
nm. The extinction coefficient of about 4.5 x 10 £/mole cm provides a
means of detecting very low CO concentrations, e.g., 50 ppm [Robbins, e_t
al_. (1968)]. Other reducing gases such as hydrocarbons will also react
with HgO and have to be removed [Palanos (1971)].
4.3.6.2 Commercial Availability
Was available from Bachrach Instrument but is no
longer being produced.
4.3.6.3 Discussion
Advantages
(a) Offers sensitivity not approached
by NDIR analyzers
Disadvantages
(a) HgO is consumed rapidly; there-
fore, this would not be very
useful for high levels of CO
(b) Hg which is formed tends to be-
come absorbed in the walls and
unless careful sample treatment
4-27 WALDEN RESEARCH CORPORATION
-------
is followed, the instrument will
not be very accurate
(c) Hydrocarbons have to be quantita-
tively scrubbed out to prevent
interference
4.3.6.4 Conclusions
This technique is not very useful for monitor-
ing high levels of carbon monoxide.
4_28 WALDEN RESEARCH CORPORATION
-------
5. LABORATORY AND PILOT PLANT STUDIES
5.1 INTRODUCTION
The objectives of the laboratory program were to investigate cal-
ibration procedures, including preparation of gaseous calibration stan-
dards for the monitors, obtain accurate calibration curves, build and test
a sample conditioning system for use in the field, and obtain preliminary
zero and span drift data. The most important studies, however, were the
determination of the accuracy of the CO and HC monitors in the pilot plant
since this could not be done in the field. In the following sections the
data from the laboratory program are given and explained in detail.
5.2 PILOT PLANT DESCRIPTION AND GAS DOPING SYSTEM
5.2.1 Description of Pilot Plant
A photograph of the Wai den combustion pilot plant is
shown in Figure 5-1. The unit consists of a 400,000 Btu/hour (Jackson
and Church) furnace with a combination gas/oil burner. The waste heat
is discharged and the exhaust gas from the burner is passed into a series
of carbon steel test sections three feet in length and eight inches in
diameter. The flue gas is cooled down to about 300°F by an air-cooled
heat exchanger and passed into a second series (3) of carbon steel test
sections (sampling areas). The gas is pulled out of these test sections
by a Westinghouse induced draft fan and exhausted through corrugated pipe
at roof level.
Detailed schematics of the sampling test sections and
pilot plant are shown in Figures 5-2 and 5-3, respectively. Each of the
test sections has four static ports which can be used for particulate or
gas sampling.
The original burner was designed to burn either natural
gas or #2 fuel oil. When firing #2 oil, the excess air was only 50% in
excess of theoretical. Units of this size normally run at 100 to 150%
*
This will be discussed in Section 6.
5-1 WALDEN RESEARCH CORPORATION
-------
to
Q_
4->
O
r
r-
a.
O)
TJ
(O
in
-------
u.
o
OJ
u
O
Q.
CM
I
tn
O)
s_
3
O)
5-3
-------
o
GAS
DOPING
SYSTEM
cc
UJ
HO
-------
excess air. The unit has been modified to attain better combustion. A
90 CFM fan was placed on the air intake to the fire box. Using this aux-
iliary fan, we are able to burn #4 fuel oil with good combustion efficiency
at the highest oil flow of 3.6 gal/hr. With these large nozzles, the heat
output is increased beyond the design capacity of the original unit. The
Btu output was increased to 650,000 Btu/hr with three 1.2 gas/hr nozzles.
The volumetric flow in the duct was increased to 160 CFM with the auxiliary
fan. The pilot plant specifications are given in Table 5-1. The concen-
trations of some gaseous and particulate pollutants are given in Table 5-2
for oil and gas firing.
5.2.2 Preliminary Hydrocarbon Doping System Runs
The doping system which was set up used a Fischer & Porter
0-3.5 £/min flowmeter, stainless steel valve, and a stainless steel toggle
switch downstream to prevent condenstation (from the flue gas) in the flow-
meter. Hydrocarbons were injected into the duct about 6 feet downstream
from the furnace entrance. The sampling location selected for the analysis
was the cold test (about 250-300°F) section (Figure 5-3) which is between
the heat exchanger and the ID fan. Using the Beckman Model 400 HC analyzer
and adding C2Hg in the doping section, the concentration in the duct was
profiled at the sampling point. Measurements were made at 1" intervals
across the duct. No variation in HC concentration was observed across the
duct. Some of the runs were made with the Beckman 400 analyzer to check
out the doping system. The analyzer was calibrated with a 102 ppm CpHg
mixture supplied by Scott Research. The flowmeter was set at various
levels and the tflf, concentration was measured in the duct. The data are
given in Table 5-3. The flowmeter appears linear. The furnace, firing
on gas, gives a background HC concentration of about 5 ppm. The flowmeter
setting of 2 corresponds to approximately 200 ppm of C2Hg. The total volume
of gas in the duct is about 2000 £/min. Thus the expected C^Hg concentration
in the duct is approximately 100 ppm and the measured value is 95 ppm. This
indicates that both the doping system and the analyzer are functioning well.
This flowmeter will give a maximum flow of approximately 1600-1700 cc/min.
Therefore, concentrations from about 25 ppm up to 700 ppm can be obtained
5-5 WALDEN RESEARCH CORPORATION
-------
TABLE 5-1
PILOT PLANT SPECIFICATIONS
1. FURNACE
200,000-650,000 Btu/hr output; gas, #2, #4, and possibly #6 fuel oil
Warm air space heater
Flue gas exit temperature 900°F
2. HEAT EXCHANGER
Duty - cool flue gases from 900 to 275°F
Heat duty 56,000 Btu/hr
3. TEST SECTIONS
See separate drawings for details
Hot test section: Temperature 600-900°F (depending on fuel flow
rate and composition)
Diameter = 8"
Cold test section: Temperature 300°F
Diameter = 8"
Gas velocity from about 3 to 10 ft/sec
4. Fl. F2. F3
Flow meters to measure fuel and combustion air flow rates
Fl range: Direct reading meter for measuring flows of #2 and #4
oil from 0.25 to 5 GPH
F2 range: 2000-10,000 SCFH (for flue gas)
F3 range: 500-190 SCFH (of natural gas)
5. F4
Flue gas flow rate - system will operate from 4900 to 7500 SCFH on
oil and from 2500 to 5000 SCFH on gas
6. Tl
Temperature control system; controls cooling air
Flow rate to maintain 900°F temperature in exit from the furnace
7. T2
Temperature control system; controls cooling medium flow rate to
heat exchanger to maintain flue gas temperature exit from 325 to
150°F
8. T,P
Temperature and pressure sensors
Temperature: Max. 900°F; min. 150°F
Pressure: Approximately atmospheric
5_6 WALDEN RESEARCH CORPORATION
-------
TABLE 5-2
CHARACTERISTICS OF THE PILOT PLANT ON #2 OIL AND GAS
Participate
Hydrocarbon (ppm)
NOX (ppm)
CO (ppm)
Maximum Temperature
(Hot Section)
Maximum Flow (CFM)
*
After steady-state
Fuel Oil
0.008 gr/SCF
^2-5
50
200-1000
(°F) ^900
150
conditions are attained
Gas
nil
MJ.5-5
40
50-500
^750
120
TABLE 5-3
PRELIMINARY EVALUATION OF THE ETHANE DOPING SYSTEM
Flowmeter
Setting
0
2
4
6
8
*
As ethane by
ppm THC
5
100
185
275
350
Beckman Model 400 THC
ppm Ethane Measured
95
180
270
345
analyzer
5-7
WALDEN RESEARCH CORPORATION
-------
with the doping system. This system was used to obtain the accuracy
tests for both HC and CO described in this section.
5.3 LABORATORY DETERMINATION OF INSTRUMENT RESPONSE CURVES
Calibration gases were prepared by injecting known volumes of
CO or CpHg via gas tight syringes into about 100 liters of air contained
in aluminized mylar bags. The volume of air was determined with a dry
gas meter. The instruments were run for 20 minutes with one reading taken
every five minutes. The gas mixtures were prepared at two different times.
Therefore, each point represents an average of 4 measurements. The addi-
tional point at about 100 ppm represents the tank calibration gas. For
both the CO and C2Hg, good agreement was obtained between the prepared
gas mixtures and the tank calibration mixtures. In Figures 5-4 and 5-5
the mV readings are plotted vs ppm CO for the Intertech and L&N instru-
ments. The line represents a linear regression analysis of the data.
The full scale range on these instruments is 2000 and 5000 ppm CO, respec-
tively. Both instruments are linear over this range. The original 25 mm
CO sample loop on the Beckman 6800 (process chromatograph) was replaced
with 9 mm of 1/16" tubing. The original loop would saturate at about
500 ppm CO. This loop should extend the range to about 1500-2000 ppm.
The CO data are given in Figure 5-6 for the Beckman 6800. The 100 ppm
points seem to be considerably off the calibration points, however, the
response is certainly linear over this range.
The ranges for the various instruments are given in Table 5-4.
The Beckman 315 NDIR curve is shown in Figure 5-7. This instrument is
actually designed for ambient air operation since the cells are about 1
meter long. The instrument response is non-linear at high concentrations
even when the data are plotted on semi-log paper. The response is fairly
linear for CO concentrations less than about 300 ppm. The instrument
response vs CO is given in Figure 5-8 for the MSA NDIR. Although response
is fairly linear over this range, the instrument shows considerable drift
over short periods of time. The response of this instrument is considerably
5-8
WALDEN RESEARCH CORPORATION
-------
1000
900
800
700
600
£
/
o
u
°-500
400
300
200
100
-1
1234
millivolts
Figure 5-4. Intertech NDIR.
5-9
-------
1000
900
800
700
7
600
8
I
a
500
400
300
200
100
02 4 6 8 10 12 14
millivolts
Figure 5-5. Leeds & Northrup NDIR.
5-10
-------
1000
750
O
o
500
250
3 4
millivolts
Figure 5-6. Beckman 6800 Chromatograph.
5-11
-------
1000
8
100
10
20 30
millivolts
40
Figure 5-7. Beckman 315 NDIR.
50
5-12
-------
1000
900
800
'CO ONLY
700
rCO + AN EQUAL
/ CONCENTRATION OF C2Hg
600
to
x
0s" 500
+
8
400
300
A CO CALIBRATION
CO+ C2H6 CALIBRATION
200
100
10
20
millivolts
Figure 5-8. MSA LIRA 300 (NDIR)
5-13
-------
TABLE 5-4
INSTRUMENTS INCLUDED IN THE STUDY
Instrument
Beckman 315
Leeds & Northrup
Intertech
Beckman 6800
MSA 300
Beckman 6800 THC
Beckman 400
Intertech
Species
Detected
CO
CO
CO
CO
CO
THC
THC
C2Hg
Principle
NDIR
NDIR
NDIR
GC with FID
FID
Process GC
FID
NDIR
ppm Full Scale
M50
5000
2000
Variable with size of
loop and attenuation
Cset for M500 ppm for
these tests)
4000
2000 (as C2H6)
1000 (as C2H6)
200 (as C2H6)
5-14
WALDEN RESEARCH CORPORATION
-------
damped. One important observation which we made here is that the MSA in-
strument responded to both CO and hydrocarbons (C2Hg) quantitatively. Note
curves a and b in Figure 5-8. These were reproducible.
The hydrocarbon instruments in Figures 5-9 and 5-10 are linear
over the 0-1000 ppm range. The size of the THC sample loop in the Beck-
man 6800 was decreased since the original loop saturates at about 150 ppm.
This channel still appears non-linear over this range. This response is
quite unusual in that the instrument is not saturating but has an upward
knee in the curve (Figure 5-11). This feature was not observed in later
calibrations.
Several interesting features can be derived from these data.
If we use a two point extrapolation [e.g., zero and 100 ppm (10% of final
value)] to the 1000 ppm level obtained from linear regression, errors
from 10 to 25% can result depending on the analyzer used. The extrapola-
tion of the 500 ppm (50% of final value) reduces the error to <5% which
is certainly acceptable. Thus, the calibration gas should be at least
50% of the value being measured if accurate results are to be obtained.
Although the zero and single point calibration points (span)
changed slightly from day to day, these calibration curves were repeated
4 weeks later (in the trailer at the incinerator) and the values (regres-
sion equations) were within about 5 percent, verifying the long term
stability of the calibration curves.
5.4 SAMPLE CONDITIONING SYSTEM FOR PILOT PLANT MEASUREMENTS
The sample conditioning system for the CO and hydrocarbon
analyzers is given in Figure 5-12. None of the instruments to be evalu-
ated could be heated, therefore, the water was removed in a condensate
trap. The stream was then filtered through a fine particulate filter,
passed into the metal bellows (MB 21) pump and into the sampling manifold.
The pressure, and therefore the flow rate, in the manifold can be main-
tained at <1.0 to 5 psi via a Moore adjustable low pressure relief valve.
The gas flow from the stack was varied by the control valve on the mani-
fold. (With the manifold pressure at 1.6 psi, the flow through the
5-15 WALDEN RESEARCH CORPORATION
-------
1000
25
50
millivolts
75
Figure 5-9. Beckman 400 (FID)
5-16
-------
1000
900
800
700
CO
ffi
-------
1000
900
800
700
CO
§ 600
o
o
DC
Q 500
X
400
300
200
100
123 456
millivolts
Figure 5-11. Beckman 6800 THC (Chromatograph).
5-18
-------
01
S-
o>
fsl
8
m
X
JO
o
_J
o
u.
iti
_i
Q.
1
-CO
z
111
2
D
-cc
i
UJ>tU_l
CC-lCCO
flL>
n:
TD
c
tt>
o
o
OJ
c
o
-o
c.
o
c_>
QJ
Q.
E
to
0)
c
(O
IT)
OJ
i.
5-19
-------
system was approximately 2.5 £/min). The instruments were fed directly
off the manifold. The internal vacuum sources (pumps) supplied with the
instruments were not utilized since the AP above ambient in the manifold
provided adequate sample flow. The system is shown in Figure 5-12. The
following section describes the filtration system used in the pilot plant
and field.
5.4.1 Filtering Systems
Initially, a 47 mm glass fiber filter was tested with the
pilot plant firing on #2 oil. The filter plugged within five minutes.
Following this, a series of tests was conducted using a cylinder (12" x 2")
of loosely packed glass wool as a filtering media. The glass wool did not
get plugged with the #2 fuel oil and appeared to remove the majority of
the particulate. Therefore, we placed a glass wool filter into the sample
conditioning system.
In order to give the filtration system a more rugged test,
the pilot plant was fired with #4 oil. After 30 minutes of operation, the
glass wool filter plugged and the system had to be shut down. A filter
holder (8" x 10" constructed of plexiglass) was designed to provide a
larger filtering area. A 0.45y glass fiber filter was successfully run
at the normal sampling rate (2.5 £/min) with #2 and #4 fuel oil for 6-8
hours continuously. An 8" x 10" Gelman type A glass fiber filter was
used in the Hi-Vol type filter holder. No problems were encountered with
the filter plugging. This redesigned filter might be adequate for the
sources to be tested since the particulate matter from #4 fuel oil is
very fine and the particulate loading was high. This filter would run
for 8-10 hours/day for about four days without plugging. The major dif-
ficulty with this filter was that it took about 1/2 man day to break the
filter out of the system,' change it, check for leaks and return it to the
system. For this reason, it was considered unacceptable for use in the
field. Therefore, a better fine particulate filter was needed.
5-20 WALDEN RESEARCH CORPORATION
-------
A Balston tubular glass fiber filter (Model 95S)
with a type A cartridge was tried in the system. This filter could be
changed in a few minutes and thus provided a considerable advance over
the 8" x 10" plexiglass holder. We ran this Balston filter in the pilot
plant buring #2 and #4 fuel oil for over three weeks (8 hours/day) at
2.5 Ji/rnin and it did not plug.
The Balston cartridge filter was then designed into both
the pilot and field sampling systems after the excellent response of the
filter under the severe testing conditions. This filter was used in all
the field and most of the pilot plant tests.
5.5 PILOT PLANT ACCURACY STUDIES FOR CO AND HC ANALYZERS
Accuracy studies were run in the Wai den pilot plant with natural
gas as the fuel. With a firing rate of about 375,000 Btu/hr, the back-
ground levels of hydrocarbons and carbon monoxide were about 0 and 70 ppm,
respectively. The instruments were corrected for the background levels by
averaging the concentrations before and after a series of tests. The
volumetric flow rate of the flue gas was determined by stoichiometry
(e.g., fuel feed rate plus excess air determined by Orsat analysis).
The rate of addition of CO and C2Hg was measured with Fisher and Porter
rotameters. These were calibrated with the appropriate gases using a
soap film flow meter. The accuracy of the CO and C2Hg addition experi-
ments is estimated as +10% as a result of errors in volumetric flow,
pollutant addition and excess air determinations. The instruments were
calibrated prior to the accuracy studies each day. A comparison of the
analyzers is given in Table 5-4.
The data for the CO analyzers are given in Tables 5-5 and 5-6
for the data taken on 4/11 and 4/12/72 and in Table 5-7 for 4/14/72. The
L&N and Intertech NDIR give good agreement, about 90%, with the theoretical
CO values over the 200-1000 ppm range. The L&N yields low values, 50-60%
of theoretical, at the 100 ppm level but since this is only 1% of the full
scale sensitivity, this is not considered a major problem.
5-21 WALDEN RESEARCH CORPORATION
-------
00
r--.
"v^
r~.
co
0
UJ
U.
^
*t
0_
I.,
O
i
i i
LO CL.
i
LO Z
o
UJ
_i co
03 CtL
^C ' '
1 M
gi
^
^^
O
C-J
o:
o
u.
co
LU
l i
Q
1
co
o
^^
rv*
»
l ^
(_5
*£.
OJ
-a
r
X
o
o
s
c
o
-Q
rO
o
fz
<"*
Q,
"U
-o
TJ
ra
O
Q.
O
O
n
^^
CO
Q.
S-
S-
o
^
oO
to
-o
O)
1
*
LO
r«
CO
c
rtS
J
o
(L)
O3
JZ
O
OJ
4-)
s-
(1)
c:
» i
o
o
CO
<£>
c
f^ Q^ CO CO CO LO ^* CD CsJ **~ ^O *^~ OJ CO CM CVJ CD CD r*-. LO CO ^^- CD CD
co co r"*^ r*^ r^» r^* co co LO LO r*^ r^* r*^. r^* r^^ r*^ r*^ i*^ co c\j c\j c\j c\j co co
i i CMCMCOCOCOCOCOCOCOCOLOLOLOLOLOLOLO
co co LO co r^» co co LO ^~ LO ^~ LO cvj r^ co CD ^~ 0*1 ^~ co LO LO co
coLOLOcor--cocrio>io>icnocf>cnocriOcfio>icr>LO«*LOLo i i
i r i i coc\jc\JLO^-LO«3-^3-*3-cr>tT>oicr> i i
I 1 lr r r-CTiiCriCncr>O>>CriCOCOr-LOCOCOcri«^-COLO
1 1 1 LOLOLOCOLOLOVOCftOCMCOOCVJCOCVJCVlCVJLOLO^l- 1 1
i i r i CMCMCVJCMLOLOLOLOLOLOOOO 1 1
1 1 1
OLOOLOOLOOLOOLOCMCOOLOOLOCOLOOLOOLOOLOO
CMCSJCOCO«*«d-LOLOOOCMOJCOCO*d-^l-LOLOOOi i CMCMCO
CO <* LO
1 1 1
I/)
c
o
4_>
(O
i.
-Q
'-
(0
o
E
Q.
Q.
0
LO
CVJ
-o
ra
0
0
dJ
en
ra
E
o
s-
o
cu
I ^
fO
o
ra
0
*
5-22
WALDEN RESEARCH CORPORATION
-------
QJ
T3
«3
CL
CL
O
O
CO
«c
co
Q-
S-
o
CO
"O
0)
O)
OOOOOOOOOOOOOOOO
OOOOOOOOOOOOOOOO
c.
8
CO
-------
oo
00
~^
^J-
IT)
D.
O
_l
ii
O.
o
LU
M
a:
o
co
LU
HH
a
^3
co
o
««
a:
=>
o
X
o
c
o
_O
s_
CQ
o
(U
(U
o
o
00
vo
03
J
O
0)
CO
EBEEEEEEEEEEEEEEEEEEEEEEEEE
0-CLCLQ.CLCiQ.Q.Q.Q.CS.Q.Q.O-D.CLCLQ.Q.a.Q.Q.Q.CLQ-Q.Q.
CiC^Q.CXQ.Q.Q.C^C^Cl.Q.Q.Q.D-Q-Q.Q.Q.Q.Q.Q.Q.Q.CLa-CLQ-
ooooooooooooooooooooooooooo
oooooooooooooooooooOLOLDLOLOLni-nLn
i i i i c\je\icMif>Lnir>LnLnoooOLf)LnLnmcMcMc\jcMc\jc\jo>j
i i i conroi
r- OOOOCVJCO(Tt'*COOi
r CMVOCTlCTl
(O
T3
cvj i cororocounLnmi-ouf)
co
CMC\J
^ovor^voi^r^i vor--c\jfr>cMi
c\jcMCMc\jCMLnLn muncorocoroi
o 10 o in oin
ro «3-
ro«*
-------
o
o
CO
*f.
-a
01
-o
-^
<:
o
0
Q.
Q.
CD
T3
X
O
c
i
c
o
-Q
(O
O
E
D.
O
o
oo
1
Q.
13
S-
i-
o
o3
to
-o
Ol
_l
LO
oo
c:
CO
J
U
Ol
CO
JC
o
O)
4_3
s-
Ol
1 1
o
o
CO
vo
c:
(0
u
Ol
00
O)
E
1
p: p: p: ^ gi
Q. Q. Q. Q. Q.
Q. D. Q. Q. Q.
O O O O O
O O CD O O
CO CM «3- O CTi
LD CT> LO V£> CO
CO i i i
i i n o oo
CM c^ co co r-~
CM
LD CO CO CO CO
CO r i i r
CM r i r r
n to LO 01 en
CM vo en co co
CM
CO i O CM CO LO
^j- r*"- r^, vo ^o r
i LO O LO O LO
O O i r- CM CM
£
5-25
WALDEN RESEARCH CORPORATION
-------
TABLE 5-7
ACCURACY STUDIES OF CO ANALYZERS
(14 April 1972)
13050
13055
14000
14005
14022
14025
14030
14035
14040
14045
14053
14055
15000
15005
15010
15015
15020
15025
15030
15035
15040
15045
15050
15055
16000
16005
16010
16015
16020
16025
16030
16035
Note: No
CH 2
Intertech
253.58
251.22
246.49
251.22
355.28
348.18
343.45
550.39
537.38
544.48
544.48
538.56
543.29
966.63
985.55
985.55
973.72
' 984.37
986.73
979.64
976.09
985.55
546.84
546.84
543.29
370.65
369.47
358.82
367.10
264.22
273.68
265.41
HC Doping
Pilot Plant -
CH 7
Beckman 315
197.59
197.59
195.40
197.59
354.58
357.34
357.34
635.99
639.28
642.58
635.99
635.99
632.69
1085.03
1078.15
1074.70
1081.59
1078.15
1088.47
1 081 . 59
1 081 . 59
1085.03
649.20
642.58
645.89
362.90
362.90
362.90
362.90
199.78
199.78
201.99
Carbon Monoxide
CH 11
L&N
238.95
218.61
213.18
226.06
327.76
331.83
309.46
507.43
499.30
491.84
489.13
483.70
491.16
870.16
878.30
859.32
861.35
850.50
846.43
864.06
853.21
858.64
481.67
495.23
492.52
294.54
291.15
304.03
313.53
185.38
190.13
202 . 33
CH 16
MSA
242.52
240.35
213.31
214.05
294.01
354.38
357.79
479.16
517.78
547.58
514.09
487.97
507.30
900.09
860.26
870.32
835.16
890.03
874.22
855.39
874.22
889.19
492.36
531.85
552.98
327.57
328.95
321.32
343.43
241.07
213.31
204.47
Theor. CO
(ppm)
250
250
250
250
300
300
500
500
500
500
500
500
500
1000
1000
1000
1000
1000
1000
1000
1000
1000
500
500
500
300
300
300
300
250
250
250
5-26
WALDEN RESEARCH CORPORATION
-------
The Beckman 315 which was designed to be an ambient air monitor
has a 1-meter absorption cell. This long absorption path length yields
a non-linear calibration curve above about 100 ppm. For the values cal-
culated in Tables 5-5 and 5-6, we used a linear regression of the 100
and 250 ppm point calibration data. This overestimates the values at
100, underestimates the values at 300 ppm and above. In order to verify
this, we went back to the original printout, obtained the MV values and
estimated the CO concentrations from the actual calibration curve (Figure
5-7). The data are given in Table 5-8. The values, derived in this man-
ner, are a considerably better approximation than those in Tables 5-5
and 5-6. The computer program was modified to use a curve which fits
the calibration data in Figure 5-7. The range of this instrument can be
extended considerably beyond 100 ppm, with reasonable accuracy, if the
calibration curve is utilized, or if a cubic calibration curve which
fits the data is used. This is shown in Table 5-7. Note that the instru-
ment follows much better although some problems are still evident.
The MSA 300 values for CO were nearly twice the CO concentrations
added to the duct. Since the MSA is sensitized with isobutylene, a hydro-
carbon with an IR absorption spectrum similar to CO, rather than CO, we
investigated the effect of C^Hg on the calibration of the instrument.
The results are given in Figure 5-8 a and b. Note that the instrument
responds quantitatively to both CO and C2Hg. The response is about
8 MV/1000 ppm for CO only and about 18 MV/1000 ppm for CO + C2H6. This
is clearly a serious drawback with the indirect sensitization technique.
For the high levels of CO found in most sources, this is not a problem
since the hydrocarbons are only 2-5% of the CO concentration. In Table 5-7
where no HC were added, the instrument shows good agreement with the other
CO analyzers.
The Beckman 6800 chromatograph for CO was modified by replacing
the original loop (36 cm x 1/8" o.d.) with a smaller one (9 cm x 1/8" o.d.).
The CO channel was found to saturate at about 500 ppm with the longer loop.
With the new loop, the range could be extended beyond 1500 ppm CO. This
instrument was calibrated each day. Although the calibration curves were
5-27 WALDEN RESEARCH CORPORATION
-------
TABLE 5-8
CALCULATION OF CO CONCENTRATIONS ON BECKMAN 315 USING
THE NON-LINEAR CALIBRATION CURVE
4/11/72
pprn CO
4/12/72
ppm CO
*
See Figure
Added to Duct
100
200
300
500
1000
500
200
100
Added to Duct
100
200
500
1000
500
250
100
5-7.
MV
7.0
15.7
22.0
31.2
44.0
31.6
15.8
6.8
MV
9.9
13.3
30.1
43.4
33.4
19.9
10.1
ppm CO from *
Calibration Curve
58
190
310
560
1050
560
190
55
ppm CO from
Calibration Curve
95
140
550
1050
580
265
100
5-28
WALDEN RESEARCH CORPORATION
-------
the same, considerably better accuracy data were obtained on 4/11/72. The
average of the data on 4/11 and 4/12 (Tables 5-5 and 5-6), however, is in
close agreement with the results obtained with the other instruments.
The data for the hydrocarbon accuracy studies are given in
Tables 5-9 and 5-10. The Intertech (ethane sensitized) NDIR shows good
agreement, about 95%, with the theoretical HC concentration over the 100
to 1000 ppm range. This indicates no losses of C^Hg in the sampling sys-
tem. The Beckman 400 FID values, however, are only about 75% of theoret-
ical. The Beckman 6800 (THC) values vary from about 50% at high concen-
trations to 70% at the lower concentrations of C0HC. If these values
£. D
for the FID's are compared with the Intertech NDIR data, the Beckman 400
gives about 90% of the Intertech values on both 4/11 and 4/12/72 but the
Beckman 6800 yields a constant 67-70% on 4/12 and a variable 50-130% of
the Intertech values on 4/11. Clearly, the 6800 was not operating prop-
erly on 4/11. The low values (percent theoretical) cannot be explained
at this time. Some of the variability at low concentrations is due to
the high zero background and low millivolt readings. This cannot, how-
ever, explain the low hydrocarbon values. The original sample loop in
the Beckman 6800 was replaced with a smaller one to prevent saturation
above 200 ppm. The data were calculated from the calibration curve. The
linear regression data were not used. The actual calibration points were
found to be more accurate.
The data from Tables 5-5 to 5-10 were used to determine the
accuracy of the CO and HC analyzers. The percent theoretical values were
calculated and any which were greater than 20 were not used in the calcu-
lations. Both HC monitors displayed low values (for no apparent reason)
on 4/11. Since all NDIR's gave accuracies of >90%, the URAS II (C2Hg).
values were adjusted to 92%. The Beckman 400 was adjusted by a similar
amount. Note that the data are then similar to that of 4/12. The accur-
acy of the C-Hg sensitized NDIR is about the same as a CO NDIR but the
accuracy of the Beckman 400 FID is nearly 10% lower.
The accuracy of the NDIR's for CO are all about 90% or greater
with an overall average of 95%. There is a considerable variation in
5-29 WALDEN RESEARCH CORPORATION
-------
CM
-o
CO
o 10
35
ID (U
31 3
CM i
O U_
CD
O.
O
I
t4
(X
CT> CO
I C£
Lf> LU
M
00
t-t CM
ii O
to .c
(O !->
o a)
CD N
s- '-13
0) -i-
c: c
»i CD
co
o
o
00
<£>
C O
O
a>
CQ
o
o
-V
o
0)
co
oooooooooooooooooooooooooo
oooooooooooooooooooooooooo
i iir ir-CMCMCMCMCOCOfOLOLOLOLOLnLOOOOOOOO
c\j LO oo csj f^ i " i QQ Lf^ ^^ Q^ c\j c\j vj o^ ^^ co c\j i.o v,o oo f**^ r*1^ r**^ ^o
r~.r^cTiCTicnoocococM^j-^-cT>cncoooocnr ocricncriocrioo
i i i i cMCMCMCMcMrocMrocococooo^d-nro
^3-«d-«3-r^r^r^r^.r^.r».r--
cMe\jroro^-«d-LnLnoocMcMnro«=j-^}-LnLnoOricMCMroro
ro «!i- LO
5-30
WALDEN RESEARCH CORPORATION
-------
d) in
T3
S- -r-
O QJ
O) N
S-
O)
(-1 CU
CO
o
o
00
UD
o
O
O)
CQ
O
o
fO
J
o
O)
co
OOOOOOOOOOOOOOO
ooooooooooooooo
^J-O>i LOCOOOOOCMOVOOOCOCOOCXJ
CTiCOC\J«3OCT>'!i-C\J^3-^J-CO«^-COCnOCT>
COCO«d-C\JCOC\JC\JC\JOJC\li r i i
Lf5Lf)i
OOOOOdl
un
5-31
WALDEN RESEARCH CORPORATION
-------
CM
CM
to
$- -i-
-C TD
U CU
O) N
-(-> -r-
$- !->
O) -I
+J tO
C C
I-H CU
o
o
CO
10
c: <->
00 VD CM *d- 00 CT> CM I
i cooii
un
5-32
WALDEN RESEARCH CORPORATION
-------
T3
to
CMr
<_> LU
Q-
Q.
O
O
I
Lf)
CM
o
(O
I/)
c:
O
(O
o
o
S-
T3
i I CM
i i O
to -C
ID 4->
S- !-
u o>
0) N
M -i-
S- -t-5
QJ -
(-> CO
c: c.
HH Q;
CO
o
o
CO
VO
(O
o
OJ
CO
o
o
c:
to
o
O)
CO
o o o o o o
o o o o o o
CM i r r- i r
Lf) O O> CO CT> CTi
CM «=f co ro ro oo
CM 1 I 1 i r
CT> ro in ro LO o
LO r~. CT> oo en oo
o LO i LO o LD
Lf> Lf) O O r r
Lf) IO
5-33
WALDEN RESEARCH CORPORATION
-------
the accuracy values from day to day but this is expected in view of the
many factors involved in the calculation of the accuracy. In fact, our
estimates of the accuracy are +10%. We see that this is the variation in
the instruments. The standard deviations for a single instrument accuracy
for any one day over 100-1000 ppm were 10-12%, again in good agreement
with our calculations. The MSA "accuracy" values were only calculated
for 4/14 when no HC were added due to the HC influence on this instrument.
The value is in good agreement with those of the other analyzers.
In summary, the accuracy of an NDIR analyzer appears to be about
95% for CO or C2Hg while that of an FID appears to be about 10% lower or
86% (but see below).
Statistical data are given in Tables 5-12 through 5-15 for CpHg
and CO from 100-1000 ppm. See Table 5-16 for a listing of the instru-
ments versus the channel on the data acquisition system. The data in
Tables 5-12 and 5-14 were obtained from the pilot plant doping while the
data from 5-13 and 5-15 represent HC and CO values from calibration bags.
Any major differences between these data should be due to interference
from the gas components. All the data were calculated from the calibra-
tion curves generated in Section 5.3. The calibration values in Table
5-13 should be within +2-3% due to the estimated accuracy of calibration
gas preparation.
Data obtained on April 19 for HC indicated that the Beckman 400,
6800 and Intertech NDIR were within +3% of the expected values over the
100 to 1000 ppm range. The accuracy data for the HC analyzers in Table
5-10 shows that the Beckman 6800 was not working properly and that the
data points at 1000 ppm appear low. The interesting feature is that the
Beckman 400 C2Hg results are about 9-10% lower than the Intertech (C2Hg
sensitized) NDIR results for this data. Note the difference in the regres-
sion equations for Ch 0 and 15 in Tables 5-12 and 5-13. The Beckman 400
gives lower HC results than the NDIR on flue gas. It might be due to
species which retarded or quenched ions, e.g., C02 or H20. This type of
interference (C02 or H20) would be constant, since the firing conditions
5-34 WALDEN RESEARCH CORPORATION
-------
TABLE 5-11
ACCURACY STUDIES IN PILOT PLANT
Accuracy (%)
Instrument
HC
Uras II
Beck 400
CO
Uras II
Beckman 6800
Beckman 315
MSA 200
L&N
*
Corrected val
**
No HC doping
^Points at 100
4/11/72 4/12/72 4/14/72
81 (92*) 95 **
76 (86*) 85 **
100 90 105
92
92 100 109
100
(86+) (95+) 93
ues
this day
ppm have been dropped
Average
94
86
98
92
100
91
5-35
WALDEN RESEARCH CORPORATION
-------
K
«
UI
Q
I/I
» z
K UI
X U. -»
a u.
V ul X
r 3 o
u
O -O a (M
1-1 0 O (M
h- IT f* O «
»- « a <* a »
Z -I
« ui x -"
_l X
T
U
o
in x
o
a i
O
X U4
a a.
> O X
T -I U
Lft »
-« t>
I
o in
Z M ...
^ UI X "* ** ^^
_l K O
t\ $£
UI
>- X
O o m *
_i a «
a. a a r~ -i
...
r
u
3
c>
c r
o o
Ot
in
O
a
a '
o z
CO
t>
e
9
o
t
u
t
u
c
u
cxi
I
Lf)
r a.
* O
*
« VI
_J UJ
a. -
»
*
X
<
o
_j
»^
a.
o m
O (\J
o »
T
(J
o
i^
*
.
^«
l
a.
ui
o
O
u
tr
o
X
>- in
X ul
a.
x
U)
z o
- UI
r »-
Z
r~ e- o
n * o
-« -< o
X
o
in
in
in in
in
z m o
o in o
ft IT * O
k- i/i o rt o
Z
< hi X tn o
_IX U I
a. o
ui
>- X
o
O
a
o
in
i
u
O
IB
(M
(M
00
*
in in
X £ X
U U U
O
o
m m
o -
c
u
X
o
5-36
WALDEN RESEARCH CORPORATION
-------
z
o >-
It Z
a iu
« »-
W «,
o «->
a u.
Ob.
> bi
I O
* z
o
IM
o»
o-
x
O
(V
o>
O
o
o
oo
(M
>
O.
O
I/)OC
z
o
a> i
it
« m «>
o IM e
o tfi ^ in
acui - o 9
oo. .
o
o
o
l
GO
ui
cc
a
o
c
X
u
o »
o iv
o
a
IT
I
* Z
O
l/l
I/I
o
in o
c o
UJ I
a u
e
bl
a
o o
z
o
1C
QC IT
X U,
b.
M
* o
*
z
o
V)
Itl
S S
e IT IT
III
X
U
(J
oo
o
cc
a
«
o
o
ex
a
* z
b.
010
a
u
o
in
o
in cc
z
o
a> i
a
I
U
I
u
z
u
a
in
X
u
Ub) u
(XI IV
m m
z in in in
u
oa
o u in
cc t>
oac
Vbl X o (B
It- O (M
Z I
^>
*
* z -co
O (MO
>- in o> o
I/) O O
in
ui x a o
ac o rn
<9 I
UJ
a
O IT
I O "
00
o »
X O O
u
(VI
rw
c in m
x x
o u
XXX
U LI U
5-37
WALDEN RESEARCH CORPORATION
-------
r
u
r
u
o
a
£ r
x in u
o »-
z z
O u.
1
J
Z ~> "
3 j. >
UJ B >^
00
«X T
X U
o
Z I
o
T
O il 3
T 1.
10 T
» J U
O I/I
C
a
^ ~ -S
o
o
O
o
_» X
1 ^
o
3
_J
T>
(\l
C3 .»
X
u
T -
O
c
(M
0
r
* z
* o
tv
Z UJ X
< T U
3. U.'
1
3
_i
a.
/)
«
r
*
T
, Q CM
O. t-H
.-£
«- 3«
I HI CSJ
I O. r
in
UJ
i >- UJ
CD cc o
\j
^ o
t c
c
u
X
C
z
O l/l
T U.
I UJ
* U.
«
-i (V
T
O
n
M
n
vj
e
a
o
a
in
v
o
a
a
a
*
-" X
O 3
t t
r
u
Z3 z
to Q
s§
PS
1/1
a.
UJ
O
«n
^
UJ ~*
Q O
- I
X
O
Z 1
o
I »-
3.
Z LU
o u
S X
x u>
4 »
U 2
* Z
* O
t-»
-n
i- i«
Z UJ
< IT
-J '3
a uj
r
H-
3
_i
~»
a
r
u
^^
*^
T
l_>
^
a
I
u
rvi
a
i
u
^
a
a*
in
K
in
a
-»
IT
T>
I
M
*
VI
h-
*
in
«
(V
tn
i
a
a
a
3
IT
a
r-
»
f*>
^^
^«
<
O
m
M
O
T
o
o
O
O
o
o
*
o
M
a
>
fV
"0
r\j
c
3
O
O
*
0
^
(U
T
*
r*.
o
I
o
n
c
M
C
3
O
O
o
-
-------
6
02
97
m o> «
K «*»»
I « M Ixx M
> » «
O
O
kl
O
u
CO
CO C\J
CARBON MONOXID
ION COEFFICIENTS
CH 7 CH
36 .7864
00 .9793
hi X xx
a u
IX
o
u
o «
o p>
xx o O
O 1C
I
u
kl V xx
00
60
2
69
NOXI
RO
Okl
a
ac o
« _i
U VI
u
X
u
IM
c
m
o
e
z moo-
to
oo
X I X X X
u u u u u
Ul
O
X
o
OV)
X kl
X
u
X
U
X X
o o
X
O
X
U
moo
W O* ID O
Z Z f- CM f- O
Okl * (u o
in
.
i
« n r> IM PI
x * in m m in
u
Uk.
a
* o o
z r> o «
o IM «a o -
m o IM
V) x I e
kl u
I
kl
O
u
0> o>
to o
m »
i i
o
o
o
X
o
o
z
KM
U J
a.
x
u
« tf>
* tr>
u
oo
r 5
Z I
X
X
S£
uz
CD o
IM O
e o
0
(W
o
a-
IM
xx
IM
in
o
en o
U kl
«
in
in n
IV
IV
I
x ;
u
O I-
» O 4D
xx e «
X e- *'
0
o
IM
I
X X
u u
£ 5
U
X X
U U
5-39
WALDEN RESEARCH CORPORATION
-------
TABLE 5-16
COMPILATION OF DATA ACQUISITION CHANNEL NUMBERS OF INSTRUMENTS
Channel No.
Ch 0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Instrument Manufacturer
Beckman 400
Beckman 6800
Intertech URAS II
Beckman 6800
Beckman 6800
Bendix 8200
Beckman 315
Bendix 8200
L&N
Intertech URAS II
MSA Model 200
Species
HC
CO
CO
CH4
THC
CO
CO
HC
CO
C2H6
CO
Type of
Instrument
FID
PGC
NDIR
PGC
PGC
PGC
NDIR
PGC
NDIR
NDIR
NDIR
5-40
WALDEN RESEARCH CORPORATION
-------
were kept the same. One would expect that the percentage of interference
would change, e.g., decrease as the HC level was increased. This was not
seen with the present data, however.
The calibration curves for the HC analyzers were prepared with
air containing 21% (L while the flue gas contained only 10% Q^. This re-
duced level of 0~ could affect the ionization efficiency. In fact,
Fegraus (1968) notes a similar effect with FID analyzers. He found devi-
ations of 10% in HC response with variation of 02 from 21%. The effect
was variable and depended on the instrument. The range for different
instruments was from as little as 2% to as high as 30%. This effect,
e.g., variation in 0^, probably accounts for the lower values obtained
with the FID.
A comparison of the CO analyzers is given in Tables 5-14 and
5-15. The MSA has a low correlation with the other analyzers because
of the HC interference. We were doping with HC as well as CO in these
studies. Ch 11 appeared to be working on 19 April but we had difficulties
with the EA data acquisition system for this channel.
The major comparisons are then between Ch 1, 2, and 7. Ch 1
did not appear to be functioning as well on calibration gas as the flue
gas, however, Ch 2 and 7 show no detectable difference between the cali-
bration gas and flue gas measurements.
The statistical data for comparison of CO measured to theoretical
is given in Table 5-17.
5.6 PILOT PLANT COMPARISONS OF CO AND HC LEVELS
Flue gas from the Wai den Combustion Facility firing #4 fuel oil
was pulled into the sampling manifold (Figure 5-12). The instruments
which were run include non-dispersive infrared analyzers such as the
MSA (CO), Leeds & Northrup (CO), and the Intertech (C2Hg), the Beckman
6800 chromatograph for CO and HC, and the Beckman 400 total hydrocarbon
analyzer. The analyzers were zeroed, and appropriate span gas
5-41 WALDEN RESEARCH CORPORATION
-------
TABLE 5-17
STATISTICAL SUMMARY - CARBON MONOXIDE
DOPED FLUE GAS - 12 APRIL 1972
Channel
1
2
7
11
16
Reading
Slope (m)*
0.48308
0.93056
0.43668
0.96427
1.82106
= m(standard) + b
Intercept (b)*
38.93016
4.34086
109.49732
-6.85270
-58.82708
Correlation
Coefficient (r)
0.85297
0.99203
0.94988
0.99028
0.98709
5-42
WALDEN RESEARCH CORPORATION
-------
calibrations were made prior to the testing. The outputs of the instru-
ments were read out every five minutes on an Esterline Angus digital data
acquisition system. The CO and HC data were calculated and are given in
Tables 5-18 and 5-19.
The data (January 1972) in Table 5-18 indicate that the analyzers
are tracking the CO fairly well. This is quite encouraging in view of
the differences in techniques, e.g., NDIR vs. chromatography. The data
were adjusted to line up the difference in sampling times between the
NDIR's and the GC. The NDIR analyzers have short response times but the
GC samples only once every five minutes. Thus, the GC lags behind the
continuous analyzers by at least one cycle (5 minutes).
There are two types of HC analyzers in Table 5-19. The Beckman
400 and Beckman 6800 (gas chromatograph) have flame ionization detectors.
The agreement between these analyzers is quite good. Note that there is
no detectable methane (<100 ppm) in the exhaust gases from fuel oil com-
bustion. The URAS II (Intertech) analyzer is a NDIR sensitized with
ethane. This detector is very sensitive to water vapor. The high read-
ings correspond to interference rather than an HC signal since a change
of an order of magnitude (2.0 to 20 ppm) had no apparent effect on this
analyzer (see Table 5-19).
The pilot plant carbon monoxide data for February 24, 1972,
are shown in Tables 5-20 and 5-21. The CO levels vary from 300-600 ppm
on #4 fuel oil. The instruments have correlation coefficients close to
0.9 and slopes which do not deviate much from unity (Table 5-21). Thus
the agreement is quite good.
Data for the hydrocarbon channels are given in Table 5-22. The
hydrocarbon monitors which have FID (Ch 0, 5, 10) show reasonable agree-
ment with each other but the ethane sensitized NDIR yields a HC level
about five times higher. This is due to the water vapor interference
for this monitor. Statistical analysis of the data in Table 5-23 is
given below.
WALDEN RESEARCH CORPORATION
-------
TABLE 5-18
COMPARISON OF CARBON MONOXIDE LEVELS FOR #4 FUEL OIL IN PILOT PLANTS
Beckman 6800 (Ch 1)
ppm CO
L&N - IR (Ch 11)
MSA - IR (Ch 16)
._
189
192
178
167
167
402
403
935
87
140
-
219
256
216
155
175
297
423
477
1030
108
184
144
142
124
175
129
139
139
293
435
541
930
62
184
113
113
TABLE 5-19
COMPARISON OF HC LEVELS FOR #4 FUEL OIL IN PILOT PLANTS
ppm HC*
Beckman 400 (Ch 0) Beckman 6800 (THC) (Ch 5) URAS II - IR (Ch 15)
28.2
10.0
17.6
7.2
5.7
1.6
1.3
2.2
9.7
21.6
23.5
5.6
1.5
1.5
1.2
0.9
as C2H6
27.8
9.4
16.1
--
--
1.7
1.5
1.7
8.0
26.9
13.3
--
--
--
--
0.5
34.1
22.9
23.7
26.4
22.5
22.5
20.9
21.0
21.6
22.8
24.8
25.4
20.8
23.1
24.8
24.1
5-44
WALDEN RESEARCH CORPORATION
-------
I
UI
o
O
00
LO
CM
00 r
Of
'S
CM
0.
o
hI
Q-
UJ
C
X
o
c
z
z
-------
o
o
a.
z *CD « e o »- n>«D * -<« » o ^ o f\» r~ «c « » et ^« » if> » *> rw ^ o j»> o in (> ^ (\j in »>- t\j
at x * o«w « (M m o» « w ©IT # * eo r- cc <*»-« CD in - -*M oa> «» ^ <> « (*
-------
a.
4
a
UJ
a
C t-
Z 7
C U
C U.
X U.
3c LU r
< o u
O U
IT
C
--
1/1 I
uj d
_l
a
~
o-
or
C
Z I
c
O «J
a: a:
X UJ
o
K
<
^
IT
o
o
<; a.
a.«
K
C
_J
>-
Q
IT
CJ-
T
(_
f\J
>-
(V
c
c
a
rv.
1
U
T
«J
I
o
5-47
WALDEN RESEARCH CORPORATION
-------
n i> s> -~ a 9> »*»-«-«fvi
x in r- » « f- w-« v> «] IM ni « o 9 rv -» «i o .» * * r- o «\AI » 01 « - ^- »> IT o »« « ai iv r- nr- ru » in « **r- » oooe<*» o oe oa - a: rv c ma. cr c c- (
S C
_ . a i
ca « o i
§5? "
CM Qi g
CVJ > g
I Z ^» >. ji i m fn-*) 0 »*> /> ^- x. « « * > M ^ t> o ^* > > ^ Ji m"*- -^ * -- "^ f 'X>
If) QC 1 c- j » » T ! -v o >x-t' mT4 «-nl>C"-^t» _i x
_| rf^ « u
CO h-OQ o
< r> -»* iTin oo^-«-» iMWif^i^* tfiriro o -» rv f\< *^> n -* * tr* tf oc ~ *~ pg
» o c o » o o -30 cooo^o o o o ^ o o jco^ os o o ioow>^j«c:i?^
. A0 WALDEN RESEARCH CORPORATION
D-HO
-------
TABLE 5-23
.YSIS* OF P]
HYDROCARBON MONITORS - (Feb. 24, 1972)
STATISTICAL ANALYSIS* OF PILOT PLANT DATA
Y
Ch 5
Ch 10
Ch 15
Y = M X + b,
X
Ch 0
Ch 0
Ch 0
r = correl
M b
0.287 4.74
00.171 5.29
-0.0637 7.87
ation coefficient
r
0.365
0.489
0.499
WALDEN RESEARCH CORPORATION
-------
No HC data were obtained for February 25, 1972, but the CO data
are given in Table 5-24. Ch 11 appears to be reading lower CO levels than
the other analyzers. The statistical data in Table 5-25 bears this out.
Ch 1 and 16 are in excellent agreement with each other. Each analyzer
seems to behave poorly occasionally although the calibration checked out
that day. The HC and CO data for March 2 are shown in Tables 5-26 and
5-27. The HC levels in the duct were increased by addition of
of CpHg to the hot test section. CO levels were kept high and constant
by modifying furnace conditions. The statistical data in Table 5-28
indicate better agreement between Ch 0 and 15 (NDIR) than between Ch 0
and 5 (both FID's). The statistical data for the CO analyzers in Table
5-29 is considerably better than the HC data. The NDIR analyzers (Ch 11
and 16) agree closely but Ch 1 (PGC) has a high intercept and slightly
low slope which may indicate some problem.
The Bendix 8200 which functioned only for a short period of
time was down for repair and no more pilot plant data was taken with
this analyzer.
On March 28, 1972, the pilot plant effluent was spiked with
CpH, ranging from 0-5 ppm. No CO was added to the duct. The data for
hydrocarbons and CO are given in Table 5-30. The statistical data for
the HC in Table 5-31 indicate that Ch 15 (NDIR) follows Ch 0 (FID) bet-
ter than Ch 5 (PGC with FID). This may be due to the large discrepancies
between 11 and 12:30 where large differences are observed between these
instruments.
It is clear from the data in Table 5-30 that Ch 0 and 5 are in
better agreement than 0 and 15. Note 13:15, 14:10, 14:30, etc. where
Ch 0 and 5 give excellent agreement but Ch 15 is 20 ppm higher due to
the water vapor interference. This is further evidenced by the correla-
tion of selected data (12:30-15:00) from Table 5-30 as summarized in
Table 5-32. This can be compared to the statistical data in Table 5-31
for the entire set.
5-50 WALDEN RESEARCH CORPORATION
-------
I
LU
Z
?
*
Z <
o -^
flc
-------
UJ
c
»
X
c
c
S
2 4) c\»^if>«^ir'*'*if><^iflif>i^ir<*''i-o»^-*^o^»o^occ*rri^o
sc
ir x M in <* * (\i M m -* «o o> *> ** in 2 x
CM a
J> »_
rsj tr> o IT o IT- ec oiriroiroir cir oir cir ciooir
kntnoo<-4^^«(\i(\i(vmm<^-*ininooi-4^4Oj(V(*)n
ooceoccceoccc c- cccocccccc
*-*ir>k/).r>tr)kf>iiiiniS)if>tnif>if>Lrtr>
-------
a
iti
o
«r
Ut
C
-
X IT
c »-
z z
C U
z - *
O U. r
X U.
tr uj z
.c
o
o
o
M
Z
O
OD
at
c
tt
«» i-
(\t in
« IT
' ' o
» o *-
- t-
t- - IT
a IT
t- _
2 U Z
< ft <_'
-i tr
a c
u
- c
O
T
o
C
C
<7>
a
»
x
* Z o o
O 4-0
- - tf o
m « o
>- t/>
z u. z --
«i a «j
-1(9
a li-
ar
c
c
c
or
o \r>
UJ
c
I
u
z
u
IT)
If)
CO
<
c
z
C V
3 U
U
Z Z -0
O tu x
x a
tr ui z
«»u. <_>
c u.
* (VI
* f »
C t- (V
f" C
z --
C' I
o
o
- c
o
c
o
rv.
o
o
o
*D
-*
cr
r
o
r
o
a
UJ
a
o
OJ
1-1 o
h-CQ
<^ Oi
U-
c
X
c
z
c
3
o
X
a: in
< u.
«-' _i
a
x
* «
* w
t- C
z
«> li.
_l IS1
a M
O
_l
*->4
a
M «
z z z
u u u
c c
-a
M
Z I
c
3 »-
a.
Z UI
it ac.
or uj z
at- c
o z
* z
* o
(V
I
o
o
p- * C.
cr
c
l/t . rt
>- I/I «
Z U. Z c
«a u IT
o
c
- c
o
o
_J
»-l
a
z
(J
rv
r-
ir
IV'
e
o
o
o
o
n
IT
ft
-------
IT IT IT « IT o r*
CM > cn
I X i
in
uj '5
_l < Of
s §£
CM
3
QL
O
ii
Q.
r
L
V
c
f a
»- c
o
cc-er-ornacrv.
i \t -a j < ir
ce«ceir or «.^
'
crecec-
scoccecoccoeeececcocoeoccoeeeccccececccs^e-cc
*-»«*-***«*«*i«inir^ir>/iin*i''''ixiJiD c*« o« DC, *> c c
5-54
WALDEN RESEARCH CORPORATION
-------
><
CM a: o>
J> S-
'I
I
?
z «
o r*
a>
at I
Z T
< iT|
ci-'C f*f\ * OT^-«C
P3O^^O-^S^^'V«*O/^O-^30 -*^ 3OO^r> -.-^3
ereceoe
oc-cecocroooc-cc-oc coooc-oococoeo
" '
5-55
WALDEN RESEARCH CORPORATION
-------
UJ
c
-*
x
c
z
c
X
z o» &o> oerccoocco'iraceraDartrocceccrcrccoc ococ
_
^^
*
*
> r*i 9> m P~ r*> tf> tr> f*- f~ in
no
in
ir
IT cc^-^cvcvrr>rr<-*irircc ^-fvivrr rr -*xfir
coooooooecoooooooooooooo
*
WALDEN RESEARCH CORPORATION
-------
a
u
a
Ct-
* z
ft U.
« «
U «-
O-
T U.
ou.
VUJ
10
0
* ?
* o
a U>
-J1
a i
t-t.
c
l/l
T
U
«
^
m
c
o
o
O
C
co
T 1 T
l_> O O
t/1
z
c
T
« u, (V a o
o o >c i\j o
O Z if> - » O
or uJ ^-» 9 o o
C 2 . . .
> UJ T I O
r u- o
C
M
C if) m => «
K- IT O C
7 -
a f
_) UJ O
a -
c
a
c cr (v
X
o a uj T i
x ^ cj
2
*-H
*
* z r-
o o*
*- tP o
.
_JQT
O. a
M
rvj
O
o
o
C
o
o
o
C
_/
O
c
o
c,
c
o
c
c
T or
o
rvi
if
e
2
I I
U O
Ifi IT
1
o
T
U
5-57
WALDEN RESEARCH CORPORATION
-------
a
tu
o
X (/)
c »-
2 Z
C U.
7 -
O U.
X U,
T ill T
< C O
t! <_'
o
oc
UJ»
00
-a
o
z i
z «
O uj «-
xa
a: o i
« _i u
Ul/l
a
in
e
o
o
* O
-I
«r or
_l T
a o
o
O
O
c
if
a
M
»
a
O
-* o
IT cr c
t-i/1
Z lu X
< at o
_ a
a. u.
a
C IT
e- - o
< UJ
_l *
a -
*-
c
o
T
O
I
u
T
U
i
u
I
U
u. »'
c. o
-(t
X
c
Z I
c
x »- a ir
a. -» <
Z UJ
m
-------
IT
I
c «c co
& O* O
j «o» IT
> CC *^ OD
. . in tn o oo
IT
I
-c r- c
» « a <\j«
c IT * fM rn *
m -4 in c^c ir<
ft.' e if or
I
00
O
CO
cc
C IT C IT O IT O
c r < -» if ir c
c c. c c o c o
* ».*»«* IT
a: csj
o o r-.
co >- on
UJ
O
< Q£
I <
00
CM
d.
O
I-H
Q.
z
c
O ifi
C "
> r
i o
2
4
_J
a
^>
c
a
rv r\.rv. (v r-
*- ft, ft. !**'* <ifiroc«« ~eft.'fn
ooocoeocooooococoo
occccooooccococc occcc c
-
5-59
WALDEN RESEARCH CORPORATION
-------
a
UJ
o
Z lf>
o »-
I Z
a u-
« -
o o
o >-
a u.
* ?
* o
Z _)
«r UJ
_j cc
a. TC
o
- i.'
c
_i
.
a
-f
a
i
t->
rvi
cc
e
c
o
o
ir
O-
ru
ct
o
o
o
-o
cc
»
a
c.
IT
*
c
in M
t~ \fi ifi
Z I/) »
«UJ I f-
_ja o
ad
UJ
K a
oc
(VI (VI
t- IT
OC
i
O
o
o
c
O
o
CO i-t
e IT if
J I I
U U «->
z
c
CC
or a>
< uj *-
c. u ir
ITT
c
o
o
o
* C
z
o
«(/I
-JUu
o
o
IT
in
X
u
ct
a
O £
c cr
o o>
o r>
X
o
c
CC
a
o in
cc
c r- * h-
> I/I X - o x
x _) o
*
* Z
O
^ 1/1
t- in
z
4 U. X
_ICC U
a. o
»
cc
in
»
o
o
o
c
_l
»-N
&
>- o
x «c
U"
I
5-60
WALDEN RESEARCH CORPORATION
-------
TABLE 5-32
REGRESSION EQUATION CONSTANTS*
HYDROCARBON MONITORS
March 28, 1972 - (1230-1500 only)
Y
Ch 5
Ch 15
Ch 15
Ch 0
Ch 0
Ch 5
Y = m X + b,
X
Ch 0
Ch 0
Ch 5
Ch 5
Ch 15
Ch 15
r = correl
m
1.25
1.13
0.880
0.789
0.702
0.868
ation coefficient
b
-2.93
14.9
18.1
2.68
-6.09
-10.3
r
0.991
0.888
0.874
0.991
0.888
0.874
5-61
WALDEN RESEARCH CORPORATION
-------
An Intertech NDIR (Ch 2) was added to the evaluation program
about March 15, 1972. The CO data for March 28 are given in Table 5-33.
All of the CO monitors show good agreement, but note the low slope and
high intercept for Ch 1 and the high slope and negative intercept for
Ch 16, the statistical analyzer in Table 5-34.
Early in April 1972, a Beckman 315 NDIR analyzer was also
included in the evaluation. This instrument was designed for ambient air
(1 meter cell) but we planned to evaluate it for low levels of CO in flue
gas. The pilot plant data are given in Table 5-35. No C2Hg was added to
the duct, therefore the HC levels were so low that no usable data were
obtained. The statistical data for CO are shown in Table 5-36. The cor-
relation coefficients for the CO NDIR analyzers are very high, about 0.98,
but that of the CO-GC (Ch 1) is slightly lower, about 0.95. Both the CO
by GC and CO by Beckman 315 (Ch 7) have large intercepts indicating the
possibility of some bias.
No HC data were obtained on April 14, 1972. The CO data, how-
ever, are given in Table 5-37. The calibration curves fed into the com-
puter were wrong for both Ch 7 and 16. The linear, rather than the third
order calibration equation, was used for calculations for Ch 7 and the
calibration curve for Ch 16 is for CO and HC combined; it is, therefore,
off by a factor of two. The statistical analysis is shown in Table 5-38.
Note that Ch 1 shows a large intercept and low slope compared to Ch 11.
This feature has been noticed throughout all the pilot plant testing.
5-62 WALDEN RESEARCH CORPORATION
-------
CD
r- £> f- »- CD » (M
o* ~« r*- o* i*- o *o
K a, a- o »
> » .
-O 3> » -
I
X - StiiS
*" i c r- (V * *
X o o- * r\j o oc
^5 a a c- c ^*
O
C If O U" C If i
Q JSJ O C O O O O <
CO Qi CX»
in
S1"
LU | O
_j {
^Q ^f i
5 gra E
.CM -
3
°- I
3 1
ii
a.
< o -*iriiriroorrittif>(Ei'OGO«c*-*mC''*flj'*(y'*f^cciri(\jc\joja''r>'-«comec~*
a
t~
c
_l ^. ,e_-^«.f>-.«.C
>- riirirO' *<^»
Q ........
* r^ r-aao 0 ^ ~ »»-.- e e c caccr CCTCOIJ cocotr
ircirciroircir>c;iricirciroirciroiroirociroi/'c:«<*ir^r-(y circir ciroireircirc ir
c-«u*ir c.c^-»-fvi\,frir''*'*ifircc.»-^.cfvr>r'-*tfircc»-^«-^^.^-^f\. f> ^) 1^1 f/ 1*1 ^ ^ fn ^ fn f^ f^ (^ fo f*> (^ ^ ^ -^ ^ -^ ^
g_g3 WALDEN RESEARCH CORPORATION
-------
U.
c
X 1/1
c >-
z z
c u.
3 -
CJ
Z "
Ok.
X U.
Z UJ
« o
I-' O
3
0
9 fc
H-
a
^*>
(VI (VI
o-
i
o
e
c
o
c
I *-
o
o
o
f-
o>
r^
^M
cc
9~
o
o
o
o
^1
o>
rr
(VI
o»
0
IBH
o»
a
c
o
0
o
^
r^
x
Ql
o
o
^
(r
r^
IT
(VI
C
(/>
1-4
UJ »
C 0
-or
»r
o
Z 1
c
s »-
a.
Z UJ
0 0
cc ac
ac uj
« i-
0 Z
H»
z
0
ir
^ (A.
Z UJ
«* a
oc
ac
c
IT
1
c
o
o
0
o
-»
~-t
»
<3>
-C
0
o
o
o
a
o
z z -« o KJ o «
ouj - * co o (Vi
CO tt
acuj xi o ~
c
X
c a
o n
o (vi
o cr
oc
UJ 3
C C
» a
Z '-»
O UJ O
a a
a: o x
< _j u
t_> 1/1
-C
(V
ct
CO
o
o
o
\i
X
X
u
ir
Z UJ X
« or <_>
_i a
a uj
a
K O M
O o c
_l o c
x
u
c
-*
in
c
»
ir
C
CO
IT
i- fVJ --
M «
X
(J
X
«J
<
1
I
u
X
(J
X
u
X
u
5-64
WALDEN RESEARCH CORPORATION
-------
o >--o.o-*'»>'»*'
l»> O r- O t*
t
i ^ o
X
o
o
O CM
uo OQ r*»
ro as: o>
I < r
LO O
_i a:
OQ c£ Q.
CX.
/ > x u
i c
D-
r
J
O ?
O t! --
T
iM»
OT
oojionom3'Y)oi/iia/>oinomoinojioi/ioin3jioinoji^>ir
-
*
*
5-65
WALDEN RESEARCH CORPORATION
-------
to
CO
in
LU
_i
CO
<
* OJ I C 0
,-u ,~i > _ o
e in » *» > o
*+ » » » a o
u
* c * c c
-» m o -.
IT » X 0 >
UJ ~4 q> o> » o a-
x «/) o
e »-
z z
C UJ
y - * c * tr
J » -n 3 ~< >
? -« » » -O O -XJK-
30. O> 3> 3 > »
i MJ r
a 3 _>
o c
7
* o *> 3 -« o r\j
* «- * o i rw »>
<\! IT 0 C > T
« a> 3 » j> i>
^ uj r
« T (J
-J X
a. o
j
1- 0 fl < ,* « ~. ?. 01 ~ 0
3 O^^-,^ »«^*0
_j c, .P » '.' m « > a> ni ^- o
" "5>>>j> _ ^ i/i i/i m o
r E _ _, i i t,
1 U U 1
<
rf "i or « c
Q Od /> > * 0 *
K. -VJ >- « m o> (^ o N-
|_ O> -- UJxOD«4-oro
S crrcc i _ _. i o
<, UOOUO XO 1
1 1 O
o- -« ₯
c£ c u-
h Q. TUjir»0*CC
Otf J/133*^-
I .zzr^f-somrw
MCO 3^jji*3*in
""' X ul X x 1 0
II * U. U
II (JU.
>- LU a
QJQ #!\)O»»^-
«
^X OMToao^X
^^ in o * * in
loo r?^??? i^S^0 - "
, z: j j h-
^ =
I-HQ3 333/IJ1->
I n? _| o » »- n a»
* n i: x rccrc
X'J>JT^XX'13B LJO»JOL>
-1 jj U
0 -1
T
* <
rvj
H- o h- f- f- < f~-
z x op x or OD OE
< UJ U
_J>4
1
l/>
3
_l "
1 « r- o> X 00
r x * x * OB
* o rv. fn o
x « m ~i M o
4* <9 o ^- o **> o
» -< O ~ It o O
at u
u>
0
in (» c- ir c rr
"* m * -< 3 -4]
~> >c rvi m = m
UJ » ^-i o> ^ (V, 0 (j,
-4 X T M
X O
o
Z 1
o
T ar or er c or
3 M 3 > K.
Z r* JT> *- o ^ **>
O UJ K. ao 3 »- *-
x o r -i
» J U
U 1/1
Z 49 3 »> .n 3
« O ^ 3 r-. IT «
-« rv ir' o - p- ^
l/> H 0 . X C
z uj r -i
< X O
-J 0
a. uj
i
^ o x ^ r^ ^^
O o > ji h- _
_j o * x -n -
« 3 3 -. > >
T -
O
M ^ 0
tree;
U _) 0 0 0
* ir c. r^ o
X C1 -4 * * 3
< o '>n "> - T o
> r* 3 »! O- 0 0
i r *- t> an =
a. u -4 .n x -4
UJ _ | 1
a i i
-« »» y- m 3 «n
-^ < >n m c -4
ujji_4xo«ot)
>-iXl«4»r-oin
x o Cf- **> in cv
0 II-
Z 1 1
o
T >- IV C C C C
X -^ 3 » 1
ZOJ^*^* 3O 3
O " 3 3- f>
XulIO^- 3O">
4 t- O 4t "VI to
O Z -> " -
»«
z « o e - «
* O -I O OD T »
i/) m o ^ o -^
H- (/>
zujxino4- 3 J} -i in J>
I 0 TO X (M *
cj fn c*) ^ r^
c
u
c
u
5-66
WALDEN RESEARCH CORPORATION
-------
OCVJ
LU I I l
>4
to
H «'' "^'7*
*-
o ».»."it«|Ni.*.»'......»..^..:j...^.^'7?'^"i.*'^*ri'^T'C'^*"r"^*"CT
O ** »* ! j,^ ,_ j,^ ^^ t ^^ ^^ ^^ 4
* **************************************************
5-67
WALDEN RESEARCH CORPORATION
-------
r
u
o o
O
^ xi
,3
oiTiojioir oj>oji oji
o«-^f\irvjfT>f)'*-»if»i/>oo«--
*
*
c ,0 WALDEN RESEARCH CORPORATION
D-OO
-------
If
o
a»
t
u
OP
a
o
o
a
UJ
o
o
o
l/^ in l/^ l/^
r c c t c o
u
(/* l/1 if1 If lf^
T >T -T >T -T -T
IT. IT IT
* 10 ifi
* (\ I*" < C
A ^ rt CD o
^C *^ ^ o ^ o
» a- ,- o
r -< < M .AJ o
rf > f- c *
M 0 * 0 00
* IT rv ^ o m
i i
q
u
r* fV f- ^^ ^
-^ r^
x x i r JT
u u u u u
a
O
_j
tr
> z
«f O
UJ i/J
O ^j
- a
X O
C U.'
2 I
I
Z
I
DC
4
at
5
m
a
UJ
o
(/
4
3
Q
T
,
,_
a.
i^'
o
UI
2
hH
«J
>- -z
< O
~ »>
I/)
UJ I/I
c u;
T
X O
O U)
Z i
o
I
^
o
T
a
u
X
o
fV
z
0
*~*
T
(J
<£>
r
o
"
T
r^
I
l_j
*V
I
(
I
U
m
«
nj
,_
o
o
s
^,
^
u
^
c.
a
fcfc
^
r-
J>
f^.
C
^
1
*/
T
ji
^
O
T
t
1
(*}
(X*
T
^
'*
4
1
C
0
0
3
o
f^ "^ f^ ifl
O f»- -C >
0 T * fV
cr r- ^ -<*
o rvj o >
,_
c. * rr ^»
r\ - -
rvj r^ c
O i_> O O
«_,*»,
nO r*- c. o
^ T ^- O
M ^ D
CT1 *> "y
~* fV
O i£ ^"* ^
O **" t *\
o 10 x tn
*
o fv u a
-T ir st
1
a* o T
*V f*1 r"1 ^~
^ 4 - £
"*£)"*)'*)
-^ C ^ IT
fv ^ * j:
""
r
u
T X
-------
6. FIELD EVALUATIONS OF THE CONTINUOUS MONITORS
A discussion of the mobile van is given in Section 6.1 and the field
test results for the three sources are discussed in detail in Sections 6.2-
6.4.
The general procedure followed at the three field sites was similar
to that employed during the Wai den pilot plant tests. The van was moved
to an accessible point near the sampling location to minimize the amount
of sampling lines needed. The major objective of the field program was
to determine which types of monitors could be used, under the varying
conditions at the several sources. Data on instrument drift were also
obtained. Work at the three field sites was carried out during the spring
and summer of 1972 as shown below.
Municipal Incinerator 26 April - 17 May
Grey Iron Foundry 23 May - 9 June
Petroleum Refinery 28 June - 13 July
The general characteristics of these types of sources were described in
Section 2.
The individual instruments were calibrated in the field prior to the
tests. Appropriate regression equations (up to third order polynomials)
were fitted to the calibration curves for use in a computer-based data
reduction and statistical analysis program. The span and zero for each
channel were checked daily, and for most of the days were readjusted to
conform with the calibration values. During the series at the refinery
the zero and span levels were not reset each day but were allowed to
drift in order to obtain a measure of the stability of the monitors.
The data for these tests were hand reduced. The results of these tests
and other stability runs are given in Section 6.5.
The computer generated statistical summaries contain a set of param-
eters characteristic of the pair-wise correlations between the various
channels (monitors). The slope and intercept represent a linear
6-1 WALDEN RESEARCH CORPORATION
-------
regression equation relating the variable (dependent) to the column vari-
ables. The correlation coefficient is the commonly computed Pearson value,
r. The square of the printed value corresponds to the fraction of the
total variance of the ordinate variable (dependent) that has been accounted
for by the regression relationship with the independent variable.
The value of the "t" statistic printed relates to the statistical sig-
nificance of the difference between paired sets of data. A significant
value of "t" means that the mean difference between two sets is not zero.
In general for the number of samples in most of the sets examined, sig-
nificance at the 95% confidence levels occurs at a value of "t" of about
2. Note that this test is based on the mean of the differences, and is
not the same as a test of the difference of two means.
In the data that were reduced by hand, the same regression statistics
were computed. In addition, the significance of the slope relative to
unity was examined also by computing "t" based on the calculated regres-
sion line slope. A significant difference would tend to indicate that
each of the instrument pair was not producing corresponding outputs.
6.1 SAMPLING SYSTEM AND TRAILER
The field program involved evaluation of the CO and HC monitors
at three sites: municipal incinerator, grey iron foundry, and petroleum
refinery. The program was to involve a substantial amount of equipment
including eight instruments, a data acquisition system, compressed gas
tanks for calibrations and fuel as well as other equipment. It was
determined that the best way to handle this equipment was to install
the instruments in a mobile van. This would also provide a protected
work area.
A conventional twenty-foot office trailer was acquired and rein-
forced to accommodate an additional 1 to 1.5 tons of added weight. The
floor was covered with sheets of 1/2" plywood. The instruments were
secured by mounting in frames built from 2" x 4" 's. The sampling main-
fold was bolted directly to the floor and the tanks were secured to the
6-2 WALDEN RESEARCH CORPORATION
-------
wall by heavy chains. Two 120V electrical lines (20 amp and 30 amp) were
run from the site. This 6 kW allowed a 30% safety factor when all the
instruments were operating. An additional 2 kW (220V, 7.5 amps) was
needed to run the probe.
The sampling system considerations for this program may be dif-
ferent than those for many other applications and reflect the peculiar
characteristics of the sources tested. Scrubbers were ahead of the
sampling point in the incinerator and foundry. The refinery stream was
at high temperature. In these cases, high molecular weight hydrocarbon
species are not expected to be present in the duct. Thus, we were able
to place the condenser near the probe and cool the sample to ice temper-
atures (see Figure 6-1). This allowed us to use unheated lines from
the condenser to the trailer. Since C,-C. HC and CO are not soluble
in water to any large extent, very little, if any, of these species
would be lost in the trap. In studies conducted in the field, e.g., HC
analysis of the condensate, showed only a trace of these species present,
as expected (see Table 6-20).
The sample was withdrawn from the stack at a flow rate of about
3 liters/min through a heated sintered stainless steel filter (Intertech)
to remove particulate matter (>10y), and then through a water-knockout
trap (Figure 6-1). The sample then passed through 50-100 feet of 1/4"
tubing to the trailer. Inside the trailer the sample was filtered again
through a fine (<0.3y) Balston particulate filter and pumped into a 30-
liter mixing chamber (Figure 6-2). The sample was pressurized to about
1 psi with a Metal Bellows (MB21) pump. The pressure was controlled in
the system by the Moore low-pressure regulator in the Beckman 400. This
unit dumps about 2.5 liters/min to maintain the pressure at 1 psi and
passes about 0.5 liter/min through the infrared analyzers. The flame
ionization detector in the Beckman 6800, Bendix 8200, and Beckman 400
were operated with tank hydrogen. Air was supplied by a Metal Bellows
(MB51) compressor (note than 35 psi is required to operate the air-
actuated sampling values in the 6800 and 8200). Any hydrocarbons in
7
A step-down transformer was also necessary at the refinery and foundry.
6-3 WALDEN RESEARCH CORPORATION
-------
o
1/1
s-
o
c
o
a
c
(O
o
S_
o
to
>>
oo
CD
Q.
6-4
-------
Ul
in
>>
GO
10
r-
10
>>
IT3
(U
'o.
E
«
oo
C\J
VO
Ol
t_
3
O)
6-5
-------
the air for these units was removed by passing the intake air through a
packed bed (18" x 2 " i.d.) of Fisher 6-14 mesh activated charcoal. A
plug of glass wool was used to remove particulate matter.
The output from the instruments was fed into an Esterline Angus
D 2020 digital data acquisition system. Four-conductor cable was used
for the chromatographs and two-conductor cable was utilized for all other
instruments. The output from the EA is coupled to an ASR33 teletypewriter
which punches out the data (millivolt readings) on paper tape (see Figure
6-3). The computer program for transforming the data into ppm is given
in Appendix B. Some pictures of the instruments set up in the mobile
laboratory are shown in Figure 6-4.
6.2 INCINERATOR TESTS
The tests at the municipal incinerator were conducted from
26 April to 17 May 1972. Data were taken every five minutes and the sys-
tem was left running continuously (about 23 hours/day) except for calibra-
tion. Since a considerable amount of data were taken, e.g., a typical
computer printout of concentrations for CO and HC is about eight pages,
we have included only the summarized statistical data for each day. A
compilation of all the data has been sent to the project officer. One
complete day of data for CO and HC has been included in this section to
show the concentration fluctuations typical of the process.
The instruments were recalibrated in the field for the same
range of operation as for the pilot plant. The resulting curves were
essentially identical to those previously shown in Section 5. Differ-
ences between the new set and the older data, obtained about four weeks
earlier, were in general less than +5%. The equations which were fitted
to the data for use in the computer program, are shown in Table 6-1.
All of the previous hydrocarbon readings have been expressed as
ppm ethane reflecting the use of that species on the calibrating gas.
In the field, methane was used instead. The two instruments with the
FID, i.e., the Beckman 400 and 6800 thus read as ppm methane. The NDIR
readings (Intertech URAS II) are not equivalent to ppm methane unless
6_6 WALDEN RESEARCH CORPORATION
-------
to
UJ
00
O
CC
o
Q
Ol
00
c:
o
to
D
Zui
-------
Ester!ine-Angus D-2020 and ASR-33 Teletype
Sampling Manifold
Of):
Instrument Set Up in Mobile Laboratory Air and H2 Manifold - L&N NDIR Analyzer
(on right)
Figure 6-4
6-8
WALDEN RESEARCH CORPORATION
-------
TABLE 6-1
INCINERATOR CALIBRATIONS FOR MONITORS
£h Equation
0 12.78X - 9.42
5 -23.55X2 + 308.55X + 14.48
15 44.48X - 5.27
1 179.78X - 20.96
2 118.25X - 1.05
7 0.41X2 + 7.23X - 5.76
11 67.SIX - 17.79
16 55.76X - 15.38
6-9 WALDEN RESEARCH CORPORATION
-------
the gas sampled is essentially entirely composed of that species. The
reference cell in that monitor is sensitized with ethane.
No useful hydrocarbon data were obtained on 26 April 1972,
although a considerable amount was taken for CO. Ch 1, Ch 7 and Ch 11
vary from about 100 to 350 ppm CO. The values for Ch 16 are consider-
ably higher with concentrations ranging from about 450 to 600 ppm. The
statistical data for the 26 April are shown in Table 6-2. The agree-
ment between Ch 1, Ch 7, and Ch 11 is only moderate (correlation coef-
ficients of about 0.6) but is not too bad ("t" tests for 1, 7 and 11
are within two standard deviations of each other) considering that no
data were dropped. Note the high intercepts for Ch 16 compared to these
instruments and "t" tests which indicate that the data from this instru-
ment are statistically different from that of the others.
On 27 April, the CO concentrations range from 100 to 500 ppm
with an average concentration of 300 ppm for Ch 1, Ch 7, and Ch 11.
Ch 16, at the same time, ranges from 400 to nearly 700 ppm CO. The
average CO concentration was approximately 550 ppm for Ch 16. The
statistical data for 27 April are given in Table 6-3. The agreement
between Ch 1, Ch 7, and Ch 11 is quite good. This is due to the large
number of samples as well as larger range of CO concentrations. The
"t" test indicates again that Ch 16 ls_ statistically different from the
other CO analyzers.
On May 2, hydrocarbon data were obtained on Ch 0, Ch 3, and
Ch 15. The HC levels for Ch 0 (total hydrocarbons) ranged from 0.2 to
6 ppm with an average of about 1.5 ppm. Ch 3 (methane) ranged from 0.2
to 2 ppm with an average concentration of 0.9 pp. Ch 15 (NDIR) gives
HC (as CpHg) of 16 to 24 ppm and does not correlate with total hydro-
carbons or CH4>
The Intertech NDIR for CO (Ch 2) had been working in the pilot
plant but during the trip to the incinerator, the light source filament
became loose. This was fixed on May 2, 1972 and the instrument was
brought back on line. The statistical data for the CO analyzers for
6-10 WALDEN RESEARCH CORPORATION
-------
Q.
UJ
O
UJ
c
-CO
CD & h-
*
f
cc
o>
cc
(M
*
*-*
C
0
O
o
r*4
O
O
l~
o
(V'
o>
f!
* o
3- _l
c u. i
>- 1 O
o
>c » o «
(V) » »\j
cr
f
ir
C
C V
3
U
I
«J
c ^
o
o *
U ft
z z i o r-
o uj MO
ac ct
a ui z i o
«u. o
OU |
<£
Z
^ S (\J
-.=>-.
r-- s ^
.- c c
T
<-> I
Z
^~
(J
C f' *
O »- f
o »- o
e o
T c
c
c
C
2
7
or i/>
< UJ
O _J
a
5
* «
* n
U. (^
z
c
>c
o
o
-J-
rv
I
u
I
C!
a
UJ
a
-4
UJ3T
0 0
- or
X
c
Z 1
c
I t-
a
Z UJ
0 U
CD a
ac uj
« t-
U 2
!>
(V
*
r~
«
T
U
* 2
* O
N-»
I/"
a: ui
O Ul
H- or
4 O
X UJ
u> a
z
^<
ac
»
N. IT
f
T -.
0 -»
O
C
0
3
O
0
^
£
cr
c
\j
r^
-»
rv
(VI
^^
*
IT
£
-/i
^
c
f~,
tv
J1
c
c
c.
f
T
U
(\
x
c
t
I
u
-------
u,
o
o> -« ao
<> o
« c* *
C * oo h- o
uj > r- f^ * o
c « .
x (vi (vi nj o
X U
c
z
C V
I U, (V F e 4
o r- m e -*
z z IT en o *
o uj a- oo o *
§(K »
UJ X O (VI
« U. O I
u u.
c
» * C Cl IT
* Z »=»)(>
o ^ ^ o ^n cc
(V O SC ^
at - ...»
o *
7 c- »v 9 r-
* z
o
7
Cta,
»-a
CE at
u. a
z
>-
(_J
?
*-
l»>
^
X
o
c
^ ^.
c
1
T
U
«
3
C
^
^4
OC
VI
er
«
i
'Vl
a
*
IT
ir
_
T
T
O
I O
(O CO
1-1 ce.
c
z
c
1
U
a
T
<
T
U
*
(VJ
ar
u1
z
c
ni
c
C
(V
fVl ("
-c
M IT >C
O « (VI
UJ X « O>
c c
-"(t X £ "
x - o *
uz
in
IT
c
-»
i
o
o
* z
o
a: o>
c uj
(Vi c
* o
» o
o
u
x t^
o ^
e
. o
T
U
IT
-------
2 May are given in Table 6-4. CO concentrations for Ch 1, Ch 2, Ch 7,
and Ch 11 vary from 100 to 400 ppm with an average concentration of
200 ppm. Ch 16, again, displays considerably higher concentrations,
e.g., 500 to 700 ppm with an average close to 600 ppm. The correlation
coefficients between Ch 1, Ch 2, and Ch 7 are good, but Ch 11 did not
track the other monitors on this particular day. The reason for this
is not known. Ch 16 has very low correlation coefficients with the
other instruments as expected from the CO concentrations noted above.
The high regression intercepts (Table 6-4) for the analyzers may be the
results of some bias. The "t" tests indicate that the CO levels for
Ch 16 are greater than two standard deviations from the other analyzers.
On 3 May, the HC levels (THC) on Ch 0 and Ch 5 vary from 0.2
to 6 ppm from 8 AM up until 1 AM the following morning. During this
time, the CH. levels (Ch 3) are about 20 to 50% of the THC values.
Ch 15 (NDIR) does not correspond to the other HC channels since the
values are usually about 40 to 50 ppm. This signal was probably due
to residual water vapor interference that we had demonstrated in our
laboratory studies. At 1 AM, some wet trash which had been collected
the previous day was burned. The hydrocarbon levels on Ch 0 went from
5 to 75 ppm in a ten minute period. The Beckman 6800 THC (Ch 5) sat-
urated at about 40 ppm although the Beckman 400 went to 100 ppm. The
saturation of the 6800 could have been prevented by changing the atten-
uation manually but an operator would have to be presented. This prob-
lem represents a distinct disadvantage of the Beckman 6800 chromatograph
in situations where the concentration can fluctuate over a wide range.
The HC levels finally stabilized about 50 ppm and the CH, levels stayed
at about 20% of this value. Below 50 ppm, Ch 0 and Ch 5 THC track each
other very well. The statistical data are given in Table 6-5. The cor-
relation coefficients between THC and CH. are quite high.
The CO concentrations on 3 May generally ranged from 200 to 400
ppm for Ch 1, Ch 2, Ch 7, and Ch 11. Soon after the wet trash was charged,
the CO values went up to 600 to 700 ppm and then flattened out about 500
ppm. The CO and HC signals followed each other during the cycling. Ch 16
6-13 WALDEN RESEARCH CORPORATION
-------
*-> cvj
o
^
C'
u
4
UJ -I
<- T
X t/> VJ
o »-
z z
O UJ
T -
J
Z * ^
3 *.
? ^
r >oj r
TO 'J
* 3
* "-
-vi
I
Y i
- T 'J
< T
T T>
J -)
*
-4
_> -
T
t
O
o
IT1
**
uj 4
O
x o
c
o
T
1^1
1 «J 'J
L
T
« 4
* n
vi
«&.
I 0
o r
i- ij u
» M
iJ /I
-^
O
3*
r
0
« « m
0 * »
'A JO O
!* (W *H
IV' ! OC
O > 3
I/I l/> O>
« in 3
N- P- 3
^
/> 3 -n
n o n
T 3 ^
N- 3
-
3 /> *
^ f\ D
3 T J1
3 ^ >-
_
-4 -VI -
L r c
_> j j
3 O 9
000
e (vi (Vi
(VI (VI (VI
3 ~i "VI
\J-VI-VJ
3 «\l I\J
t> C D
r\j <\j fVI
O 3 3
45 O O
vi (vi rvi
*>!*
C C C
(J U CJ
ll^
^^
*.
*
0
3
O
0
**
a-
3
^
O
K.
>
in
n
M
o
/j
0
^
t
J
0
o
(VI
(VI
M
M
(VI
*o
(VI
o
o
!V1
-*
r,
O
O
o
0
o
^
IT
»^
r^
4»
r"
*
n
o
4»
O
rvj
t>
3
f)
^
O
*
c
o
3
O
o
(VI
3
M
0
43
(VI
-5
-O
VI
-^
c
u>
*9 (VI ro o
A * \n « o
O (M -» > * O
"» 3D * f~ in O
r M- M v- M >
U
^ W O O t*l
J» « > O «9
3 r- r^ o *
Ul -4 (VI (V) IM O i/»
O . . . . .
-> r -^ -> o (\i
X U I
o
O I/I
T ui o «v. o e (vi
J «r n 3 » n
s t
x uj r i -
a u. a i
u u.
-t fVI
I I
M 3
X>-
O l/t X -« 3
I- jj o
» r> o
3 O -VI
VI 'VI t>
T
(J
C'
o
c
u
X
<
1
a.
a
i/)
!-
Ul «
o o
I
X
0
Z 1
0
T
£
0 J
Or\
.*.
u in
« Z
* 3
«4
/>
X k
t
a.
UJ
a
in
»^
Ul *
o o
- T
X
o
Z 1
o
T t-
1.
Z -d
O ^
Z UJ
4 <^
O Z
-^
* z
* 0
»-4
-"
O Ul
* 3C
< '3
X 'iJ
ii r
z.
»4
(J
-r
£
C'
(_>
^H
^^
T
O
*.
r
VI
r
o
-^
T
O
O
r-*
f. <
U
««4
^
T
U
r*-
T
t_>
VI
r
u
_H
T
(J
4D
0
C
~l
r~
«9
*
fc
o
O
^
£
IV)
-<
3
O
O
3
.^
-
n
*^
n
*
^
>
"n
-\l
O
(VI
o
o
».
VI
>
(VI
^.
3
^1
Jl
0
(VI
Iw
1
3
^
3
3
-,
^4
C
«o
^
o
(V
n
o
*
(V1
o
4^
*
_,
3
-5
3
-^
^
IW
^
n
vi
c
B
*
r^
o-
4^
O
M
fl
(VI
*^
^
n
TO
n
-r,
4/
~*
3
O
3
3
O
n
4
>
*
M
rvi
-^
(VI
r,
o
3*
V*
V*
g>
(VI
30
03
O
3
3
3
*
3
H
»
*
-^
-)
.^
4^
>
-
r
*
VI
0
r^
^
i/>
*
_
+
1
o
3
3
3
3
,,
2O
M
£
m
-4*
-*
1
0
n
M
-*
r»
r,
U
(VI
4*
^^
c
3
0
O
-»
-»
3
*
T
JD
hh.
*»
A>
^
-n
o
-^
.-
a
n
5
*
r»
vi
o
3
O
O
3
<^
n
^
~*
VI
»
^
3
vi
ri
»
4
O
J*
IM
>
>
t_«
-*
-H
r^
C
0
O
O
-i
^
fta.
3
or
^
^
t>
c
NB
_H
3
^,
,
r>
o
c
-4
C
o
o
3
0
>
_
*
+
*
rf«
B
(VI
^
n
3
*
»)
O
*
rt
«
*>
(VI
^
m
vi
n
^.
o
"VI
^4
»
A
'^
C
U
6-14
WALDEN RESEARCH CORPORATION
-------
0 T » 0
f*l ^* f** O
m i/i (T * o
* -4 * 0
t/> X f
Z (/I U1
C K
X Z
tt U.
<" IT T O 4
(_)<_! f f O f-
O «« iT * C O ~>
X u. - «O IT O <£
v i) x ~
I O 0
u
* 2
* c me>x- j- ^. o o - L. T
' T: :_
t "t
^- (_
+- C. f. If C
S ? T CT ^ » ru m o
t cfsiJ *o*o
I t . . n .e » ft o
1 7 _ * m IT o
9; c V) r 1 1* ru o
P Z 0 i
t o
2 CM T
is» a if
£2 c * ir r 5^; £ = § ;
., ,, O2 tf> IT* ^ o rn
y>: i j j 5 is - °. '. °. ?
Z« L, I. 0 C j.^ X (VJ -< 0 ftl
in I"H x u. o
1 1 CO -
vo * c
>. l * ru o o ru
1 1 1 r** 7^ 3* O cr ^
S3 Is CK " c. 1 ".
^Z.^ t-iriic
h 3 OQ « aj c i
C/> O£ x H-
§dj
r K
S>- c. rv. o CT
o< c e a o <
. . /±r z o o 3D 1/1 ^
P° -
»*
m
»
in a: x
z u
X 1
2
U
o in
c a
>- 0 X
I _l O
V*
*
* Z
3
* -J~
a a
C l/l
>- U, T
« a u
i iJ
2" U
*-
C
Z c
H-
T
<-
»
>
a.
UJ
o
i/i
o
9
^
VI ft X
Z 0
c
X t
a
< >-
oa.
o uj in
a: u
o a
>-UJ X
It- 0
z
*
* z
Q
- ^
a if
C t/)
- U. X
c a: o
1 a
UJ Jj
~Z 1
^-
C
?. c
3
C,
O O
ru or
(M »
ivi a
c rv
cp if^
f- O
r- r.
c f-
c f^
c C
a- 5
IT
1 t
rvj o
f~i O
^ o
a c
K O
c -
c a.
O '4
c -
c y
I I
u c
T
X
«
^_
c
cr
o
<*>
f
^fc
a
a
rv
ir
'
n
*~
i
m
n
o
rvj
c
c
o
o
9*
AJ
^
^
X
^
IT
IT
-------
does not appear to respond to the increased level of CO. The statistical
data are given in Table 6-6. Four of the analyzers, Ch 1, Ch 2, Ch 7, and
Ch 11, show good correlation with each other and the regression equations
have slopes near unity with small intercepts. The "t" tests indicate that
these four instruments, but not Ch 16, are not statistically different
from each other.
The HC data for May 4 are given in Table 6-7. This table has
been included to illustrate the variation in pollutant concentrations for
a municipal incinerator. The statistical data for HC are given in Table
6-8. Note the high correlation between THC and CH.. The data indicate
that the methane was about 25% of the THC (Ch 0).
The CO data for 4 May 1972 are given in Table 6-9. Ch 16 does
not respond to CO. Of the data in Table 6-9, only Ch 2 and Ch 11 appear
to be close. Both Ch 7 and Ch 1 are reading CO concentrations which are
nearly 100 ppm lower. The statistical data in Table 6-10 bear this out.
Ch 16 as well as Ch 1 are greater than two standard deviations from Ch 2
or Ch 11, e.g., statistically different than Ch 2 or Ch 11. Ch 7, how-
ever, does correlate with both Ch 2 and Ch 11 and is within two standard
deviations. Therefore, only Ch 2, Ch 7, and Ch 11 give valid results on
4 May 1972.
On 9 May 1972, the HC levels ranged from 5 to 39 ppm CH^ (aver-
age is about 20 ppm) for the first 14 hours. During the next 10 hours,
the values are less than 2 ppm on Ch 1 while Ch 3 and Ch 5 are reading
0 ppm. This feature of the data leads to the low correlation coefficient
in Table 6-11 between Ch 0, 3, and 5. The "apparently" higher correlation
between Ch 3 and Ch 5 is an "artifact" of the large number of zero's
during the last 10 hours of operation. The basic data show, however, that
Ch 0 and Ch 5 do follow each other well during the first 14 hours and the
correlation coefficient computed only for that time would be higher than
shown in Table 6-11. On the other hand, Ch 15 does not correspond to the
HC level but again appears to respond to residual water vapor. This
channel (15), the NDIR monitor, will not be mentioned again in Section 6.2
because the results continued to be similarly inaccurate.
6-16 WALDEN RESEARCH CORPORATION
-------
«
« r»
IM
«»
Ot
X
u
oo
*
O
o
o
in
IM
c
X VI U
o >-
Z Z
C u
I -
C u.
CC U.
(T UJ X
< o . o
o o
o
-1
a.
ui
o
UJ »
co
f at
cc
r* O
ec
x
c
-
1-1 ^
Z
I-H CO
rv
T
r-
I
I
(_
UJ
c
K
o
C (A
1 Uj
u
Z Z
O UJ
cc ac
i ru
X IM
U
o
o
I
l^
JT
« in
in r»i
o o>
I
10
I
(J
in
IV1
X
C
Z
O
X
X V>
t UJ
U I
a
* <
* in
u.
a o
^.' i/"
z
u
ro
in
rv
1*1
»
(VI
X
o
T
.
&
UJ
o
I/I
M
UJ X
o c
i-, a
x
o
Z 1
c
T t-
a
Z UJ
o o
cc a
K UJ
4 ^
o z
t~>
* z
* o
» »
ir
x in
C U.'
t- a
<3 3
X U.
U. or
Z
»
O
*c
"
X
u
v^
-<
X
^J
K
I
o
rvi
T
o
^*
T
l_
O
T
^)
oc
_«
ru
fw
i
rt
i^.
r^
o
^j
i
c^
ru
r*
IT
e~
»
ru
M
h-
c
.
IT
IT
1
C
c
o
c
ru
^
m
m
*
rv'
^
i
^-
*c
(M
m
_-
cr
c^
rvi
.
o
F~(
^*
O
o
o
c.
.
o
in
o-
-o
cc
cc
r-
Cf1
rv
>
-c
X
1
J"l
(^
c
f~
^
c^
1
c:
<2
o
o
o
cc
-3
IT
.
r»~
(J-
1
if
c
I
rv
rv
^t
rvi
£
,0
»-t
r^
I
o
o
o
o
e
^-
w~
f**
^
.
^
ru
~*
(VI
>
X
f
^
^~
l/
(V
in
*
c
o
o
0
o
-o
_4
rr
c;
_
d
^
*»
^
IT.
'^
.
cc
f*>
in
m
r>
3
cy
*
ir
^
^
IT
(V
*
-r
rv
X
(J
X
o
I
o
I 1
t
I
6-17
WALDEN RESEARCH CORPORATION
-------
"* ««« K»«*^MH««» ** *«? »»* 0«> Mm<0
»M««ir>««»i*iu>«3«««n»M«fc«nA4«>*«K«M*««».»>«
»»»» * » »*» »» »«
& ^ M i '
"**m * "* * * *
z a
u
aaaaaa«aaa«aa
a 9 a 9 09
aaaa aa
999939
u
M (VJ M (V
u
*
*»
MS fl)tnin*9vtni/iininiftin9in(f>iniftiniftm'^ifttft'^t) tftrr^-z^j^o
M ^J ^ f^ (VJ (Vi
im9ift9inaift9inamai/i am am o IP o in
* *
(MM IM MM
O
O CVJ
r^** Q£ r*^
I O O>
1O >- t
Ul Z
? °
9
a
o in
1 ;
*
1999999
X
I- X
< o
X
9aoa«*am««fninr«ii
>»»»»
a 9» »
* »
ssasz:
,
IT * 1/1
i if -3 m
I 4 ^ 1/1
u
, in aMin in
->MMN -«
MMMMM.
«
in «
IV l>
« ««
am 9»t
-* * lAM
"M'l
in am am «
S3'
e oo a o<
«l««3^'
1'1'T
Samoinamc ma
-------
X .* .Mf* MM «!*« -»» »
U
in 00090000900090009000900,9000000 o oooogooooaoruoo-ao oo-\j
3O93999eO9a3OOO9aO
T '/»in ui '
tftam ainamam 9in0mau>au« ainaifiainainaiflainantainoin oin oinoistoiroin e>ir> oina
*mi/)09^-«NMin««*i*>i«ao^«wMin(n«*inii»««^^(y«iinW'#*in>fla9^^(MM!»i(^ * 9 3 o
o
r 5
-r
>- z
< O
*
r * M M * in iv M IM
u ivicg (vi
$ . .** «.!«
6-19 WALDEN RESEARCH CORPORATION
-------
M ' T T** ' ?**! *"**f""* M * mm n««ir to «N
.
u
i/itnui tniuni Mm mm lAft Mrvni
tfi n§ iw tfl in m >M M in » m j> T M n T>
««o_i.«Nn"n<*t«««)in
,4 <«««
i
10
'/> T
2 U
U
o in
IT M
S 5
T
« U
z
r
u
«ow*». oi««»i«t«S»"»!3fT
2*'* .-. i'*'|*t5K 23 2t*-
i«in ii»m - o
>^*4*>*
inm-** nirtnnjNx * M
1*1
. I ' i ! i ' I i i '
M«»e M 4"1 «*«*»omoii>o«t« iA«in«iwi 'i»-9«4«fl
K««S-M«a«(«>«'««ininaio-<-«« MM n < « il V9"«*
oo.*aoc»o «mi*tn (*(»i ni r\ir\( >«> in m ni
inomamomoinoinomo
Mto«HMMMnm««iAino
OOOOOOOOOO33OO
oooa
Ilioii
6-20
WALDEN RESEARCH CORPORATION
-------
2
C
cj
3 !T>
r -H
*
*
o
O
T
O
-^ eo
o in o in
fn-* -*
OO3
vf> JV lf» i
*
6-21 WALDEN RESEARCH CORPORATION
-------
in o
*y
in u
z
IU
»»
u
m
0. -<
Ik
Ul X
3 U
U
2
O
>- n
«
UJ X
X O
ac
j
1 U
o;
o
i
-
Z. *£.
H- i y
CO
± '*
>- i
1 i i *v*
tiJ Q^
1 <; CO
«c al o
I- =>oa
(/> oi
-iS
<0
(-> Q£
. . ^^
F"H ^J
1 >-
co u:
> i
g
co
in
2
a
QE
«
a
3
>- M
f ttl
bfttJ
.%
«
«n
x u.
o o
H-
« 111
X IM
w
Z (/)
M
u
»4
» t» I n
s
^~
^
t
K. « e «
» « « «
m « in ...
»
X X C X L> 1
U U U U
M
fe ' « e w
U ' (M 9 | «
5 lit M » 0 *
Mk
TO' Z>
_ « i
1 *
f* W f> <*l O
x ^-i »4 t -M et
o w w w w
! o m m <*> |
. i - - 5 £ 5 S £
_ u
'« 1 ».*»'« «
X , : '» « M
C» flft ^ > ^B Mk 1^
Vf ^ . ^ ' ^ ^ "
M«
i i
|A ^*
10 . «
« «» m it "i
I M i - M i M W
u m n m n 9E
*
W l*> !>!>
X | ** * "^ ^*
u 1 *i « i n n
o m in n
u
-3
X
u
r
u
'C
9
e
A.
HP
in
r»
M
»
"1
ni
9
3
»»
t
m
£
U
I I
6-22
WALDEN RESEARCH CORPORATION
-------
HrM*»MI-*
SS3S
OinoO>4)«-*MIV
M« *» M
4 »!« {
o
M«tfi0«MoM<* M« ».
*«««» l~ «l»l>- tfltfl " '
*
n n » +!»«
u »
»3S3*o4 00*55
n tn '/i - » > n
x * oni*«oi»<^4)M*iin
« '
<« Mon4« nr»M*<4n
^* o 5* ** ^ ^' ^t ^ ^ & o ^ '
- o-orc
X
O
CM
I O Ol
VO 03 r-
QC
* !'' '<' f « !'
out oin 9in ai/i oin oir o
!
CO
<:
a:
o
r
u
->«MOI» M»» eamt
- «« 4* *>-
C_3
£ T
X O
s -
s .
« u
(J
a:
o x
t- u
t
O ^f W £Jw»
^^ W| ^^ 1^ W
ii*** *** i
-«a MMMOIT
1 ^ Wl ^» ^« O O* O ]
-J
I I
w«M«Bjo*«0»3'8*.ss*.d|»- «
iMinviintni/i«'«««-«*in««'t-*--«-r>«r>f"n«
3093OM3IM* * 3 O 30 IWOO 39»{
O vC
o ooeiMMM
«« «WM
« o - o-H >
* t- K ^
m
-------
u
"
z
u
r t i *
z
o
f\i * - T -1
»->
c
o
en
vo
Ul ««
o
X
c
i
z r-
o
5
X
< o
u
>aDN9>Bm < ^ _»« in « ><
it « tu <
MMN « Mmi
«
^-
m
UJ
^^nini runjnj ninini
-------
^ "
»»>* "MM M Mli^-ii
» «»₯ « »
r-!n * ** »# «« «»- * <»«* ii» -*--«'« ^»-»n rt««i »»in*>/»o-^o>
M-* **« MMM _..
o
o
I
1X3
' **$ **' « <- =
_ .._.. _ _ ._ _ _ '-"O
** »-_--»_ -~r ~- V ^ ^' - » ^ -^1 ^_ -- ^ - '_ -^. -.^ '.^ -_^ "_1 ^."_^.^_ i"^-i^__ » 'MAX -. M n n x
3 31 S 3 S 1>
=i^"i**Ji**^'^*'i*'^"^*~'^1'^*'^'11i~'>i"i'7Ta. "I^^"I"^"
. rrrrrvrrrrrrrrrrr,rrrrr3rrr.rjr;rrrrrrrrrrrrrrrrrr
>Min*»»»mmmfunjr.j<\immmr7mmmininmmmmmmmnj«f*3««««««oo»4>«-'><*£
>
--- - -_- _ -._ _ --._.. _-.__.*-t>iJtior**
I
xi *. . . ^ . . . .*?*?. .....^r^?hl..
ac x
i u
o
(M
0:
o z
t- o
T
UJ
T <\J^*«'\iAja)!O-o4)aj*o-vjcD«o-«o*3ai!*^)Oo»-vj*«ooorvjo»ooo3DD »oo o ry o »
oaaooooaoaooooooooooeoAooooaoooaooo'oooeooootaooaooo
* !!»***.
WALDEN RESEARCH CORPORATION
-------
in o in o o ivi M o o o
ru .* »>
r- o* +
r
O
(VI ()
c -
-c c
T
U
03000000OO
j» in in in .n /> n
*
*
6-26
WALDEN RESEARCH CORPORATION
-------
i i
= i $ i! i
i/)
K
u
>- « ' » »
O (M O »>
3 > » »
r
a:
o CM
2 >-
"- - LU
o: a
ii
(/> O
<: z
o o
p§
J5S
£
00
hJ IV * r- »
u o »»»»
Z ><- » ft « «
Ul ll» P- O - «
X »
Ul Z -. o _ oj
U. U I It
-«»*-»
7 » e * e M
C (VI (M > '«» T O>
T
O
«
VI
u.
K O
o
- UJ
««SI
X «
.J I/I
z
u
> M x
o ' ' fw in
o M » in o
a o in , o ».
T O Cl I i(M (VI
U I ; I |
r r T r r
u o u u u
r T r z
u u u u
X
<
»
1
*
VI
»rt
X
o
»
1
Ul
o
_l
I/I
I
**
(/t
(A
Ul
X
a
Ul
X
X
^
>
1
a
VI
*4
*
«
i
».
CH
u
K
Ul
z
2
O
h*
I/I
I/I
Ul
x
a
Ul
i
«
x
2
T
r-
Z
u
M
I
^)
Ju
T
O
^
»"
Z
°
«
<
T
U
T
U
(VJ
X
u
f*
T
O
K
(VI
Cl
M
K
r-
f*^
^
*
n
(M
IT
fa
I
o
tr
^
D
-i
r
u
o
fV)
*
^
^
O
IV
n
^
o
o
3
^
T
_
.^
r
_t
~4
_,
M
r
u
X
o
in
0
(VI
<\T
(VI
o
J.
rvj
^.
T
r\J
r
u
^
-*
IT
m
f>-
h-
c
-»
o
0
^
o
^
t
h>
^;
^
r
r
^-
r
o
o
*>
o
,
*
e
^
in
c
OJ
n
>
T
n
r
v
_j
_<
^
^i
n
"3
_,
IJi
4
**
C
U
o
o
o
«
«
N.
m
n
*4
O
0
^-
«
~*
-
**
J
r
^
M
_ .
-c
r
u
o
o
o
e
3
^
^
o>
or
>n
(T-
m
o
M
,,
,f
.^
1
9
pr>
O
^
^
-VI
^
*
^
^
ru
«
E
(j
6-27
WALDEN RESEARCH CORPORATION
-------
.
PI
v» x
o fc»
!i
It 2
xo u
. H J -
S SI 5
p» S
« «
Jd:
M ZZ Z
FW * O
* 1 *
2 -
B <
* o
M »-«"
e cv 4
i a
r a
MI «XpTrT"«
Z l_»
J I-
* sv
> si ,
9 ** * c
PI O C
1 «-4
I . 5.5, c
U i «"
* r *
*o
»-' in
o «
»-U X
^ flC U
flr nr
So
zt>
3C o
1 (J
1
o
odd
i«.i o m
K 0 M
« 4
o e
1 »>
Ol f^
1
o m
j ,
S £^ a 01
,,
1
in
x
S '5
a «5? *
Pi e » «
M o * in
«« O M *
C 0 0 (M
1
»- UJ X -< <
55 °
"i i n
M «
a UJ ui
r a
>- o o CP o
o u c pi » PI
' z o o \n OD c»
fl i & v * °
B
w£ ? » S
! = 5 S S
58 5 * * J
x C
u ;« S
> o
uj en * w o ?>
5 "~~ 0 M * *
J^ .^ * * *
<-) >- OK '
Z «t »-tfi x 1 o 1
i i i ;jr «ui o
^^ 3C ^
i i cn tf_
K> >-. 0 M l\>
> I o e * f-
LU OH S o, 0 - (D
__J ^£ £/} ^* Or |W 0
x
O M K
^«
o
u
M
**
*
* *
^
CQSZ -r*!*!*"1 a
? Is 5 *
CO O£
^^ '
1 O ;
«t O PI o in in
o ae -
loZ g5S*M*| OUU
fc I ' 1
fT 4
H= '8 «
^ 52 I
la s s 22
!3 j i a s
< 1U U (V» AJ M
'W *
»*
- d !
Ol Ai 0
t
J *
5!
<
*
m
x
u
tft
1 ^«
w
0
*>« X
X § c
"I * *
1 otj
'^
^ 111
I SS
i 'tri 5
S £S
iz a
ii °
0 I/' I/I Pi
I I I I
o o c
0 (M *
e «D PI
<\l * O
i
j « o
a> * o
M * 0
1
* 0 »
in o
« O h-
M 0 <
IM o m
0 * -
1 0 PJ
*
IM
*
C
*
in
a
o
n
J x o in m M
J 0 -.
K
p
X
c
0 IT IT
x; x x
u <-> u
- i i
P,
x
6-28
WALDEN RESEARCH CORPORATION
-------
The CO levels for 9 May ranged from 200 to 450 ppm. The average
concentration was about 300 ppm. All the instruments, with the exception
of Ch 16, are in good agreement with each other. The statistical data in
Table 6-12 show the excellent agreement between Ch 1, Ch 2, Ch 7, and
Ch 11.
The THC levels (Ch 0 and Ch 5) for 10 May were generally in the
range of 2 to 8 ppm with an average value of about 3 ppm. The CH. levels
(Ch 3) were 0.5 to 4 ppm with an average level of about 1.2 ppm. Ch 0
and Ch 5 track each other somewhat as noted by the correlation coefficient
of 0.69 in Table 6-13. CH4 (Ch 3) also tracks the total HC (Ch 5) level
on this day.
For May 10, Ch 1, Ch 7, and Ch 11 read essentially the same CO
levels while the Ch 2 data are about 100 ppm higher. Ch 16 results are
200 to 300 ppm higher than Ch 1, Ch 7, or Ch 11. The range of concentra-
tions was from 150 to 500 ppm for Ch 1, Ch 7, and Ch 11. The statistical
data are given in Table 6-14.
The THC values for 11 May varied from 0.2 to 25 ppm. The methane
varied from several tenths to 5 or 6 ppm. The methane was about 25% of
the total HC. The statistical data in Table 6-15 indicate that the two
THC channels (Ch 0 and Ch 5) correlate well with each other but the cor-
relation coefficient with Ch 3 was very low. For some reason CH. does not
correlate with total HC on this day.
On May 11, the CO values for Ch 2 and Ch 11 correspond but the
values for Ch 7 and Ch 1 are about 100 ppm lower. It is interesting to
note that Ch 1 and Ch 7 track each other with regard to the CO level. The
statistical data in Table 6-16 are quite interesting. Ch 1 and Ch 2 have
a high correlation, however, the "t" test indicates that the data for 1 and
2 are statistically different as noted above in the apparent CO concentra-
tions for the analyzers. Ch 11, which has a lower correlation coefficient
with Ch 2, however, is not statistically different from Ch 2 as noted by
the "t" test. Another interesting feature is that although Ch 7 is sta-
tistically different (>2a) from Ch 2 and Ch 11, it is not significantly
5-29 WALDEN RESEARCH CORPORATION
-------
-* e O
z z
c u.
z «- o » e o CD O CD -" -«
i- ru ru o » r-
I
I
o
I
u
2
U
I
U
in
C
(\.
in
c
(V
c
(VI
o
-C >C
rvi (\j
-c
^
(V
f
f
tvi
in
c
IT
vf
(V)
UJ
c
!*
X
c
z
O I/I
Z UJ
jg ^
O UJ
CD ft
If UJ
< u.
0 U.
c
z
0
a^_
^
O l/>
>~ UJ
X
UJ »-
z
H"
7
*~
K * K e
CD » e 0 e
« IM K * ** 0 .
*«)>«< K O
r - ru .- 3
u
o -« 00 o N-
in co « o co
i o rvi o o .
1-1 o m * o f-
T ~ I - 0 -i
U 1
1
(VI K 0 «D »
M B e
0> O CD CD ~*
IV * O K (VI t-
(Vj o in m «
Z -« 0
O 1
e « iv. o r-
o o> M if ec
~» o *c (vi o (vi
c (V * e o
u i i T
t '
i (V' ^j
a.
UJ
O
O O
^* or
X
o
Z 1
c
z
z
ouj
00 0*
oro
z
* o
I/I
cri/i
ou,
t-a
T UJ
u. a
z
»-
u
X
*~
a
a
UJ
0
I/)
*4
U- X
c c
X
o
Z 1
c
Z 1-
a
Z UJ
O (_>
x a
X UJ
« l-
u r
* z
* o
to*
I/I
X I/)
C UJ
»- or
^ t5
I UJ
U. QL
Z
>-
z
*
in
i
a
in
pi.
f*
*
^
(VI
ru
(V
c
c
o
0
c
o
*
0
X
m
c
fl
o
0
o
o
0
=>
c
f-^
c
~
rv
~1
i_
ru
IVI
in
o
in
rvi
in
*>
i
(V
a
(VI
*»*
o
o
0
^
^
a
a
or
f
4
X
^>
in
m
CD
c
o
0
o
m
o
or
ir
cr
r^
3
u
21
-
O
^
in
o
rv.
i
c
o
o
o
o
X
**
m
ir
\c
i
IT
ir
rvi
0
(T
rv.
in
»
*^
«
o
o
0
cr
X
(VI
r-
;
IT
'
^~
,
O
rvi
o
C
3-
C
O
0
0
G
O
l/l
-------
bJ
o
*- o
m
in -
tt z
a u,
« «- IV * f IT
o «J » oc e »
o *- in >o o o ci
a u. . o o o o
c u. «
> uj x i i -«
x o u
u
>c o x o
". -3 « »>
Ot J C
*: c c o
i
U
m (vt «
> ec o
§
I m
Z U
C
I I
ft
« at sr
o -t O
a a o
01/1
»-IU T
« a <_
x<»
UlUJ
za
-C
*
uJ
X u.
u.
»*
» c
*
z
c
Z H-
^«
o
ir
-c
in »
IM
I
o
o
o
»
(vi
«
I
u
I
u
I
c.
O
O
fV'
a.
in
(V
tr
*?
c.
X
a
u
o
a:
c
V t/1
xui
«J
a
* z
* <
i u.
C C
< UJ
I
u
IT
*_l
X
J
IT
T
O
C
ir
(V.
in
(VI
m
IT
rvi
*c
in
(V
IT
in
(V/
C
c
ui r m
in
>fj
^ ~ r
e
t- u.1 I
< 2 U
z cr
e
J
If
ev
C _ r-
IT
(\
X X
u u
in m
X X
XXI
u u u
6-31
WALDEN RESEARCH CORPORATION
-------
« OD M *> »»
»*»*»
X
* in IM K o
Of o « e > o
« -» M IM fM MO
X IT O
C >-
z z
c u.
I »-
(J
OE U.
at u x
« o u
o u
z
* o
*
»- (V.
IT -I
c uj z
- at u
|
o
o
c
e
a
ev
*
a
at
if
i\
x
*
T -
(J
i-t O
£ 5
ec
in
IM
0>
IT
Mt
M M IM
c
z
c
|
at i
in
ni
03
in
IT
IV'
m
IM
IT
IV,
«c
in
IM
^
**
X
u
*
0
X
K U
Otf)
to1
ll '
^ B
a u x
ou.
o
*
z
O (M
nr ^f.
C I/I T
KUJ U
K
Z
^»
(_> _
J
" j
°
M $
» «
m m
* «
M i"
x IM
1
« O
IM O
o r-
IM
1
» O
» o
in e
c. c
m e
e 9
e ac
o 4n
e r
e r
1
x IV
X X
o o
2
It
*
^
Ml
O
e
e
o
e
0
*
0
"
(U
4
^M
IM
c
1
r
x
u
o
IM
e
0
e
e
e
f^
in
«
i
m
Is
N.
*
(V)
m
^
O1
fv
,_
1
-
X
u
o
o
o
e
o
m
o
(M
^>
1
C
O>
1
^
o*
fta
n
I
IT
IV
a
i
*>
x
Ui » e r- o
a> a.
at o T
«-i u
uin
Z « o u- * m
3 « o »vi j->
IM * o f r
If
a i/)
C bJ
- at
«(S
z uj
iu a
u
7
*£
c
f" t^J
a -
z
J
(M
T
v
I
u
(*
T
U
|W
IM
OD
*
O
fib
r«-
*+
«
«
I
»
o
CM
m
e
^
>
i»>
rvi
c
rvi
ev
I
c
o
^
c
c
^
OD
»
^
in
IM
r*
in
en
rv.
»
^^
«
IM
"
0
O
3
c
o
r-
rv
>
*
IT
*
ac
(M
m
^
in
^*
-
-------
l/> X
z «n u
O h-
CD Z
a ui
u u
o - in
a u.
o u.
>- UI X
I O «J
u
z
o
i- in
a «
o _i
t- ui x
< a: u
z ac
ui o
z o
o
z o
IT *
f- ao
o o «
004-0
o
ac uJ *^ o o o o
a a. «...
>- o x -.
Z o o o o
o « o in « o o o
ac ff*
u o m PO *
z o o - -> r-
- o « - ni
r -
o
: ^
1 en
ir> ii
i
vo
a:
£ o
ZD CO
(/> O£
SO
a:
HH Q
I >-
00 31
oo
in in ft
XXX
U O (J
o
CD
ac
o z m oo m o o>
ac ui < * in o 4>
aac .
> ui x o
x u. u
£ £ £ r
o u o o
o
z
o m
o in o
o m *
t- co I
« UJ U
ac -
ui
u
z
>
^^ (VI (VI (VI
1/1 x *0 *o ^ ^4?
z u -»-"-«
o
o
ac
o « O> »
o ni r- *
o o o> at i
ac
« h-
o a.
o ui
tr o
o cc
o
CD
(Vi
o
«ac u
ac a
ui ui
z a:
u o»
z o o o>
»* O 00
X O
u
o
*
(VI
(VI
in
o
(U
-------
_ (VI ~ _
<0 * M O
000
0
o
m « «
rg -.
*
IM
O
" I
xm u
ot-
z z
O ui
O U.
3> U.
at uj x
« o u
u u
a _i
o uj
>- et
« a
ct o
Ul O
9
O>
IM
f*l
« M
O 9
0 «
m
r- 0
oe
p CM
vo
i
LU
i
i
>-
Q£ UJ
§O
II
X
=> O
00 Z
o
3"
PS
23
I U
«c
CO
X
u
E
U
r x
o u
X
o
r
u
IM
O
(M
O
IM
.0
Ul
X
o
SW1
Ul
u
z z
OUI
at ui
« u.
uu.
0
z
o
NCINERATOB
T TES1
£
X
u
X
u
X
u
IM
U
X
o
o-
9
«
IW
IV
IM
0
(VI
1
«
9
(VI
O
9
9
9
9
»
IM
1
m
i
m
IM
t
0
O
9
9
9
9
IM
1
0
IM
fW
IM
e
o
a
o
o
in
IM
IM
IM
0
IM
ft
*
0
1
o ui x
>- ac <->
o
o
o
o
u
Z
o
o
9
IM
o-
co
in
»
m
co
a)
o
X -*
<_>
* IM r- ~*
z
u
r
o
r
o
(M
-o
(VI
c
X X Z K
U O O U
I
u
z
o
to
or i
X
u
» IM IM
m « «
(M
O
a o
o x
t- Ul O
at -
ui in
x
o
IM IM
a>
m
at
>
.t
a.
Ul
0
in
»«
Ul Tt
O O
x
f,
o
I K-
a.
z ui
00
oat
at ui
« -
u z
IHl
z
o
t-»
i
IV
f^
If)
r*-
*
o
^».
0
N-
f>
^
*
i
9
9
9
9
9
OC
0
in
P*
ru
P.
in
^-
o>
o
f^
1
c
fd»
r^
in
.
o^
o
l-t
9
9
9
9
.
9
m
9*
0
r»i
o
m
oc
0
9
*
IM
^
t
0
in
_
^
i
e
9
9
9
«
9
IM
0
4>
.
in
^
1
^
(U,,
f*>
in
*
^
>~
o
r4
ru
o
9
9
0
0
^-
IM
in
-o
^.
*
*
o
^
IM
9
.
^
-C
^
ru
^
m
c
9
*^
0
O
o
9
3
^
,1^
^
IM
O)
,-rf
IT
1
r>.
in
o
*
.^
^^
^H
1
O
0
*
in
^
^
i^
I
0
^
i*.
«
IM
m
IM
i
^
-------
different than Ch 1 ("t" test = 0.28 for Ch 1 and Ch 7). Examination of
the raw data bear out these statistical results and indicate that the
computer data reduction and statistical analysis program was working.
This set of results points to the importance of considering
significance as well as degree of correlation. In intercomparison of
two channels, high correlation coefficient may be interpreted as a
measure of good agreement. It is, however, only a measure of how the
two sets correspond. The significance test in the computer routine
determines whether it is likely that on the average there is a difference
between the two. Alternatively, as was done in the manual reduction of
data for certain of the days, a check of the significance of the depar-
ture of the slope from unity can provide similar evidence of distinctly
different functioning of two channels. Thus, in the above set of data,
one would suspect either Ch 1 or Ch 2, although the correlation coef-
ficient was quite high (0.918).
Many of the HC values on 17 May were too low to be measurable
on any of the instruments. There are several excursions up to 20 or
30 ppm where Ch 0 and Ch 5 track exceptionally well. Ch 3 (CHj is
usually about 20% of the THC concentration. The statistical data for
these results are given in Table 6-17.
The statistical data for CO is given in Table 6-18 for 17 May.
The data are of limited usefulness since the incinerator went down about
19:00 hours and the data (0 ppm) were recorded for another 10 hours beyond
this time, greatly biasing the statistical results. The data records show,
however, that Ch 1 and Ch 2 track each other well. The values for Ch 7
and Ch 11 are close to each other but about 100 ppm higher than Ch 1 or
Ch 2.
6.2.1 Wet Chemical Results
The PD-(O-phen) method (described in Section 3.
was used to provide a check on the reliability of the continuous monitors
at the municipal incinerator. On 11 May, six different samples were
6-35
WALDEN RESEARCH CORPORATION
-------
» o * ec e
K « * o
CD cc in o
Z I/I
o t-
X Z
X U.
O "-* 4fl
ac u, .
o u.
> U) X
X O U
IT
(VI
^ o *
C O CO
o> o »
<> e m
a
ui
o
z
o
X I
DC
o in
acu)
oa
xo x
o
o
(VI
(VI
o
(VI
o
o
o
o
* z
* o
X
T
r-
o
O
o
c (V 9
c a IT
c r- M
o c »
o
»
«
cc
O
r
*
* Z (VI O
O r\i o
M ifl IT O
am c. c.
oi/i
»-U. I -" *-
02 U
XO
UJUJ
za
(vi o
o
m
o
o
o
o
c
I
(J
rv
o
ff
a
cc
>r
e ir ir
X I T
r
X
x
or i/*
fl U. f *= O
O «J « (V. O f»l
o z in in m o o
(c ui ^- «o Ul X " ** O -
X U. U
I
o
i
o
c
I
o
a
c
a: »-
ui
z *-
if.
X
o
o
-=
c
o
o
o
o
o
iC
tn
(VI
in
-c
in
^
X
o
in a
z
o
X 1
a
« t-
o o.
O Ul
flc o
0 (t
>UI
X t-
z
*
* z
o
»_l
a v,
C I/)
KUJ
IX (9
Ul Ul
z a
0
z
*-
«>
X
tj
in
^
x
u
fcf^
X
°
o
X
o
>o
(*)
»
o
m
*->
1
If
K
^
r-
(VI
i
c
rw
M
^^
O
0
o
c
c.
in
^
in
*c
^4
1
1^
">
CO
(VI
(VI
1
o
o
0
c
c
IT
in
c
1
OD
(VI
in
m
t-t
c
o
o
o
o
(VI
1*1
£>
K
If'
"~
'c
C!
^*
If
If
o
o
o
0
o
(V,
*~
cc
m
1
cc
z>
rv,
^
"
^
9
C
cy
in in
X X X X
o o o w
o
X
XXX
u u u
6-36
WALDEN RESEARCH CORPORATION
-------
M
n
*
IT O
m o
~ co
m ftj
e
o
a.
u
o
I
u
IV
« »s
o
OD
IT
*
o
at
99-0
9 r- »
" *- 9 iTl
f- * OC
(V C I/
- ffl O
» X O t
(V
1*1
»
c
z
c
IT
O1
.»
» 3
» -i
C u.
- X
C I
x c
a:
o CM
I r-.
I
vo
£
T
4/1
IM
(M
o
« -C
<\J
c-
c
- I 2
ft, IV
»
a »
1 O
UJ
C
I
u
a; o
^* ~* o
c
(^
I/I
* D
"- X
u. a
2
o ,f
O J
C.
C a
2 U.
»_
^ ? r-
O UJ
X Ct
a uj z
< u. u
u u.
c
*
* ?
M
O o
-* h-
o
c
rv
a K
i
(_
o
a
a i"
« uJ
(-> -J
a
3
u. I/
(V,
r\j
IVJ
IV
£
(\
(V
a
M
o»
IV
(M
O
l\
rvi
IM
O
ft.
(Vi
o
IM
IM
IM
IM
ft.
ft.
C ^
c ir
(M
rv
O r^
IV *V'
IM
IM
!V
ft.
»
^
I
(V
a.
IT
ft.
-
e
a
UJ
o
1/1
UJX
cc
^-cr i
X C>
o
r i
c
00
cca
truj i
* z
» c
C U.
t- X
< 3
-C iTl « o
a-
a.
c
3 ~
f (\
ft.
rv
i
iv
e\
rv.
I
u
x
V
6-37
WALDEN RESEARCH CORPORATION
-------
collected in aluminized mylar bags and brought back to the laboratory for
wet chemical analysis. The CO concentrations by the instrumental methods
were the average values for the 15 to 20 minute sample collected at each
time. Five individual samples were collected for analysis. The CO samples
were analyzed in duplicate and the averages are reported in Table 6-19.
The results on the comparison of wet chemistry and instruments are shown
in Table 6-19. Statistical analysis of the data in Table 6-19 yields the
following results:
y(L&N) = 0.707 [X(Pd-O-Phen)] + 72.9; corr. coeff. = 0.93
y(Intertech) = 1.06 [X(Pd-O-Phen)] - 8.7; corr. coeff. = 0.93
y(Intertech) = 1.49 [X(L&N)] - 11.4; corr. coeff. = 0.99
6.2.2 Losses in Sampling Line
The sampling system described in Section 6.1 cools the
sample in an ice trap prior to analysis by the continuous monitors. This
could lead to a loss of any hydrocarbons which are condensable at that
temperature. In order to determine whether this could be a problem, the
condensate in the water knockout trap was extracted with (3) 20-ml por-
tions of chloroform and then three portions of benzene. These extracts
were air dried and weighed. The weights of HC found are given in
Table 6-20 for both the incinerator and foundry. The equivalent conden-
sate concentration in the sampled stream for 4 May calculated from the
data in Table 6-20 and assuming a sampling time of 23 hours at a rate of
3 liters/min is about 0.13 ppm (converted to a gas phase concentration).
The average value measured by the Beckman 400 is 15 ppm. Thus, the error
(HC lost in the water trap) is less than 1%. The values for the 9 and
10 May are 0.36 and 0.49 in the water trap and 18 and 4 ppm by the Beck-
man 400, respectively. We see that, only on the 10 May, is the HC in the
water trap a significant portion (about 10%) of the HC measured by the
analyzer. The average
-------
TABLE 6-19
COMPARISON OF CO CONCENTRATIONS BY WET CHEMICAL
AND INSTRUMENTAL METHODS
ppm CO
Intertech
11 May 72 389
325
340
342
327
327
12 May 72 268
255
266
257
220
L&N
337
294
306
298
294
294
262
242
265
246
222
Pd (0-Phen)
369
310
330
316
290
298
264
286
275
252
200
6-39 WALDEN RESEARCH CORPORATION
-------
TABLE 6-20
CONDENSIBLE ORGANIC MATTER COLLECTED IN WATER TRAP
Date Site Water (g) Organic (mg)
5/4 Incinerator 173 12.4
5/9 Incinerator 510 34.7
5/10 Incinerator 509 46.3
6/1 Foundry 372 21.6
6/2 Foundry 386 21.3
6/5 Foundry 645 35.7
6/6 Foundry 305 7.1
6/7 Foundry 500 26.9
6-40 WALDEN RESEARCH CORPORATION
-------
6.2.3 Selective Combustor Results
The Beckman 400 analyzer was modified by coupling a Gow
Mac Selective Combustor in series with the Beckman burner (Figure 6-5).
We hoped that this modification would allow the determination of non-
methane hydrocarbons with a flame ionization detector. The selective
combustor was set up with known mixtures of methane and ethane. The
combustor current (temperature) was set so that all of the ethane was
combusted but none of the methane was burned. Thus, with the current
off, total hydrocarbons are measured and with the current on, only CH.
is determined. The non-methane HC is the difference between these two
values.
The data obtained with the selective combustor are
given in Table 6-21. This instrument is compared with a process chroma-
tograph. The statistical data on the comparison of the incinerator re-
sults are quite good in view of the differences between the two instru-
ments. The results obtained at the foundry (also in Table 6-21) are
somewhat better indicating the versatility of the technique.
6.3 GRAY IRON FOUNDRY
The gray iron foundry used for the test program, recycles
engine blocks and other automobile parts. The melted down charge is
cast into manhole covers. A diagram of the foundry is shown in Figure
6-6. The sampling site selected was at the outlet of the wet scrubber
because of the high temperatures (>1500°F) at the top of the cupola in
the early afternoon. The foundry was only operated for about 8 to 10
hours/day. The sampling was usually started about 8:00 AM and continued
until the foundry closed down at 3 to 4:00 in the afternoon. Tests were
conducted at this site between 23 May and 9 June 1972.
Since both the CO and HC levels were considerably higher at the
foundry than the incinerator, the instruments were calibrated with higher
levels of gases. On the instruments set up for high concentrations (MSA,
L&N), the gain was turned down. With the Beckman 6800, the volume of the
6-41 WALDEN RESEARCH CORPORATION
-------
UJ
"ztn
00
oc
\l
UJ <
tt O
cc
o
5 UJ
^§
23
i00
I
Q.
3
UJ
CO
i
o.
l
UD
01 cQ
'o
X
T
Ul
_J
Q.
<
V)
UJ iT I
>£&>
O
o
O
O)
CO
c
o
ro
O
o
O
to
I
UD
O)
S-
3
en
6-42
-------
TABLE 6-21
PRELIMINARY RESULTS WITH A SELECTIVE COMBUSTOR COUPLED TO
A FLAME IONIZATION DETECTOR
5/17
5/4
5/9
CH4
THC
6/6
6/9
THC
Beckman 400
9.8
8.7
3.0
8.4
13.8
12.5
24.8
(Beck 6800) = 0.
(Beck 6800) = 6.
B400
Foundry
160
93
93
5480
3200
(ppm)
Beckman 6800
10.8
8.1
6.3
9.2
11.5
15.9
14.7
82 + 0.69 (B400
3 + 0.40 (Beck
Manual Anal.
143
105
102
5150
3450
Sel. Comb.
2.9
1.9
1.9
0.3
0.8
2.8
1.4
+ S.C.); corr
400); corr
B400 + S.
81
46
27
2640
1038
CH, (ppm)
Chromatograph
+ B400 Beckman 6800
3.7
3.0
2.5
0.8
1.4
1.2
1.4
. coeff. = 0.62
. coeff. = 0.77
C. Manual
103
51
33
2480
1092
6-43
WALDEN RESEARCH CORPORATION
-------
T3
C
o
s-
o
(J
Ol
.c
o
I
UD
OJ
3
0>
CD
c
CL4->
e s-
ra o
00 D-
6-44
WALDEN RESEARCH CORPORATlOt«
-------
sample loops were reduced considerably. The loops selected were about 4
to 5 cm of 1/32" o.d. stainless steel tubing. With adjustments of the
amplifier in the 6800, high levels of HC and CO could be brought on
scale.
Some of the calibration curves under these conditions (Figures
6-7 through 6-10) were quite different in appearance from those shown
previously (see Figures 5-4 through 5-11) for lower concentrations.
The curve for the Beckman 6800 does not have the strange shape previously
observed, but, instead, is well fitted by a linear equation. The NDIR
monitors begin to saturate as shown. It was necessary to use higher
order equations to fit these curves.
The calibration curve shown in Figure 6-7 was based on ethane.
A similar curve was prepared using methane, and was the basis for the
data reduction. Methane was also used to calibrate the Beckman 400
(Ch 0). Ethane was used to calibrate the NDIR. Thus, in the data
tables in this and the next section (refinery), ppm hydrocarbons is ex-
pressed as methane for Ch 0, Ch 3, and Ch 5, and as ethane for Ch 15
(NDIR).
The calibration equations for the instruments are given in
Table 6-22. These data were obtained using a curve fitting program on
the computer. At low concentrations, most of these instruments were
quite linear. The deviation from linearity is due to non-linear ab-
sorption of the infrared radiation, e.g., deviations from Beer's Law.
The hydrocarbon data for the foundry on 25 May are given in
Table 6-23. Ch 3 on Beckman 6800 chromatograph worked only briefly
yielding high CH^ values. Ch 15 does appear to be responding to some
HC's but the response certainly does not correlate with the flame ioni-
zation detectors. The chromatograph (Ch 5) does not track the FID
(Ch 0) at the foundry unless the HC concentration remains at a reason-
ably constant level for a long period of time. The FID response appears
to be following the changes in the actual levels in the exhaust over the
WALDEN RESEARCH CORPORATION
-------
2000
1500
~cb
z
CM
o
9
1000
i
500
468
millivolts
10
Figure 6-7. Calibration Curve for Beckman 6800 at Foundry.
6-46
-------
5000
4000
3000
8
2000
1000
62
y = IrP-rr - 1200
66 68 70
millivolts
72 74
76
Figure 6-8. Calibration Curve for Beckman 315 (NDIR) at Foundry.
6-47
-------
50,000
40,000
o
u
30,000
20,000
J-.
10,000
8 12
16 20 24
millivolts
28 32 36
40
Figure 6-9. Calibration Curve for L&N NDIR at Foundry.
6-48
-------
40,000
HIGH RANGE (ATTENUATOR at X2)
30,000
20,000
10,000
LOW RANGE (ATTENUATOR at X3)
8 12 16 20 24 28 32 36 40 44
Figure 6-10. Calibration Curve for MSA NDIR at Foundry.
6-49
-------
TABLE 6-22
FOUNDRY AND REFINERY (CO BOILER INLET) CALIBRATIONS
FOR CO AND HC MONITORS
Ch
189X - 10.5 (<1200 ppm)
476.3
PPm)
3
5
457.19X + 22.96
439.5X + 38.1
15
7
11
16
147.35X - 26.5 (<1000 ppm)
5.5X2 - 90.5X + 954 (>1000 ppm)
v = 89318
y 77.25-X
0.933X3 - 3.39X2 + 596. 87X - 129.
4.44X3 + 162. 67X2 + 2563. 7X
+ 119.54X (>1X)
6-50
WALDEN RESEARCH CORPORATION
-------
I
oo
O
CQ
on
CO O CVJ
cvj o: r>.
I O (Ti
vo >-1
CO
>-
i oO'»o »mi« «« «)!
z-~--«
00 MOM
M« « #'* I
i oo 001
00 O
»«j«*«l>
OOMMfW
9
J
T-I
«
MMMlS
o -« M
« o«
5SS2S
waflt^sfoCSJ
^^«(««« *< »
* **^ 4410 ^'^(^^M^n *^ W«^
«*^* » **» do»»00» »
|p llm< Ui Mi>^<(p«»<«^l^^^OOOoaooo ooo oo oo
o oooao aoooooo
Kooooaoooooooooo
I i
l*aoooa
I ^MMMM
So: MOM a
MM'MOl
ai ooo a
ia:
Mai
oooo o
MM MM MMMMlM MMM MM
«*
6-51
WALDEN RESEARCH CORPORATION
-------
T
o
«ntr> <" trip o ^- in » - «vi -
o X
T *J
a:
i
fvj
oc
; r
^ oc <> oc
-------
c
o
O
oo
CVJ
i
vo
DQ
«=C
: i
in 'o*i« 0'flD +\**+-o
M O » «<* H«** Ml*
X
0 -- - »-.
f»«l
I I , I
! ' I
I
m OD ru * r» « OD * in r-* OD nj (M oo M r-
z jl'J'.'C^L'ti'*^*
O X OO^flOO^^mfn^iB**!'
QB ^J ' ^* ^^ 1^ ^^t O^ OD OD O^ ^^ Q) ^d ^9 ^Om ^0
Ol
o ^
V O O O O O OO OO O O OlOOO Ol
X . . V
X OOOOOOOOOOlOOOOOO
I U !
* , § i
s I : ^ ' 1
o cG^eotroin >o IT o >o m 01 o o m'
> (MM m * O» I\IM ^*^ ^'HWI|WO> O> M
o !
ta.
«OOO OIM * «OOD OM 4"
-------
daily cycle. For example, the HC concentration (Ch 0) goes from 200 to
500 ppm at about 1430 (25 May), but Ch 5 does not reach this level until
1510. Also, at 1455, the HC level on Ch 0 changed from 500 to 1000 ppm
and Ch 5 does not reach this concentration until 1525. The data on this
day indicate that at the foundry, the response time of the Beckman 6800
is over 30 minutes. This excessively long time may be due to adsorption
at high HC concentrations which would tend to be more of a problem with
a "batch type monitor," e.g., one injection every five minutes as in the
Beckman 6800 analyzer. Note that after a half hour, the readings on the
Beckman 6800 (Ch 5) reach the same levels as the Beckman 400 (Ch 0). The
HC levels at the foundry are considerably higher in the afternoon than
in the morning.
The CO data are given in Table 6-24. Ch 2 and Ch 16 were not
working properly. Ch 1 (chromatograph) appeared to be saturating, e.g.,
the sample loop is too large for the CO concentration. Ch 7 also appears
to be saturating above 1% carbon monoxide (10,000 ppm). The only CO
analyzer which appears to be functioning properly is the L&N (Ch 11).
A statistical summary of the May 25 data is given in Table 6-25.
Since only one CO monitor was working properly, all intercorrelations are
not meaningful. Comparison of the good CO channel (Ch 11) and the HC FID
channel (Ch 0) shows a high correlation.
The correlation of the two hydrocarbon monitors (Ch 0 and Ch 5)
is poor, although a visual scan of the data shows the instrument readings
more or less corresponded. Only the data from 1524 hours were examined.
The very high values of t imply that the two channels do not correspond.
The HC data for May 26 are shown in Table 6-26. Ch 15, again,
does not show much agreement with the HC levels determined by flame ioni-
zation analyzer. The HC concentrations very from several hundred to sev-
eral thousand with considerably higher values in the afternoon. The slow
response of the Beckman 6800 (Ch 3, 5) leads to inaccuracies in HC con-
centrations compared to the fast response HC analyzer, e.g., Ch 0. The
6-54
WALDEN RESEARCH CORPORATION
-------
.o eo in m o in» »co
in >o (urn m* o o-« I
m mo t>-in or- ominin«r>-«moini'nrMcD^in mm oa> cc in ni- om or-
»a>»»* * o o> » *
g * x !v^;
rr w o Hf m ID * n w
rj CU * OD O1 9" OC ' IT
O ^.r^-.^-^-t^.^. ^. _<,_ .. _ *^^.<^. ^* cv cv *-i~ r-»- cvijv rv" ^- »- tv (Viv *v(v
Z (>
^^^ ^^^^ "?*^j^ (j*1l«^» »^* * *~ 3°*^3** J^1^1^ "I^^."! ^^^^ ^.^^.^j ^
I QQl O^ ^ ^" ^ ^- ^ ^ "^ f'' * f*l fH (V (\< ^.' *" ft (U ^t^^<£^f,'^CC^^^tC '*>»'*»/> t/i Ji */>»/> o o o *^^-« f\j (\j ro m * ^ ** *i/>iniPi/>uiooo o o » ~t i
""
6_55 WALDEN RESEARCH CORPORATION
-------
o oiv in
>«0Booo 00 »oo«>i«>o«J
**»»«»» * f>*ir>Kf-
T rvi d i/> in ^ (v M « « a- T M r- * -c « f^ » m » K r- f^ M o o«r c o c * * * * o r^ * <\jt-«*--»i*i-« «o>f- IM-*^ irir >r « o« »>MO>« ^ 7" tO ^ * '^ ^ ^ -^ ^ ^l *^ "*> w% «r ^
CM § -«-----. --.---.---.----.-.--.«-- «-«-------- -..
I T
i c^eco^r".
1C1 O- " -f IT
.O O> ~> r\j
« iv IT ^ *
r T -*oo'»-'r'-«*i
z u ^oo^^mflc*^
D (fl t> C J3O-^»'*
O C-^^ ^ O O
-nc o
*
* *
6-56 WALDEN RESEARCH CORPORATION
-------
V < ^ -- O O »
* C> ro » -ar
tj CC'M i(^^^»C^h*flC^-«-~« f C < !*>
ccaaaccaococ aota o a a a CT
3
Si O
I T
^O 1 rv -'- r- - r- *\ '.-.-*
LxJ ^^ *"
e£ | «J * ~ ? IT -f !T
I < £ fVJ -< !\J
r
T
O
UL.
ro*^-^ -^* & & 'f\ in^^ri j^oooo
*
*
6-57 WALDEN RESEARCH CORPORATION
-------
CM
CT>
^~
un
CM
^
(O
s:
un
CM 1
i
OL
UJ «t
cQ ^"
I co
i
^
0
1 1
I
co
i i
I
CO
s_
"^
V.
f s^^^
o
-M C
(13 (1)
^ *r~
O) U
s- <-
c^ t^.
O >4-
0 OJ
o
o
rx
*
4-J
Q.
a;
o
j-
O)
-M
C
r
^
>HH^
O)
a.
o
^~
CO
oo r^
LT) CO
- in
^* r^ ^~
1 i r
>* r^ CM
o LO co
o o o
1 1
£
^c
O LO O
^: ^: x:
C_) O O
o1 -?
ii i
E
^ O LO U
jc ^ ^: >>
o o o
6-58
WALDEN RESEARCH CORPORATION
-------
in
z
o
CM
I
UJ
_ I
m
O
OCVJ
Q£ t^
Q CT>
>- i
-------
£ -
Cu i*>
i/> i
Z u
O
CD
a
<
o
o IT
(X ocai(^<^(\)'*»«Moor)»oin«'*i'<-«ifl^« o(^»iP'*on^ofv)^-co * a- o
o
oo o>.-<
§ :
i o i
VO z o
o
LU O
_l u-
OQ
o-ofvcre
.
OCOOOOOOOCOOOOOCCCOCOOCOCCCCCCCOOCOOcCOCCOCC
'
*
g_go WALDEN RESEARCH CORPORATION
-------
methane concentration (Ch 3) appears to be a significant (>50%) portion of
the total hydrocarbon level at the foundry. The HC statistics for 26 May
are given in Table 6-27.
The statistical results do not show any clear trend. The various
pairs all show relatively high correlations. This is probably a conse-
quence of the few high readings about 1400 hours. The slopes of the re-
gression curves vary substantially, but, on the other hand, the t-test on
the differences does not indicate a significance between any of the sets
of data. The computer run shown in Table 6-27 incorporated the readings
from 1140 to 1200 hours, which were incorrect. The correlations were re-
run by hand, but no changes of consequence were found. The hand calculator
routine provides a different measure of statistical significance as noted
earlier, a t-test on the slope (t ) of the regression line. It is in-
teresting to note that this set of data, t , shows that the slope for
each pair of HC channels (Ch 0, Ch 5, Ch 15) except for Ch 0 vs Ch 5, is
significantly different from unity, although the t-tests on the differ-
ences between all pairs shows that none are significantly different from
zero.
On 26 May, neither Chi nor Ch 2 were in operation. Ch 11 and
Ch 16 follow each other throughout the whole range of CO concentrations
0.5-5% on the cupola effluent (Table 6-28). Ch 7 (Beckman 315) appears to
be saturated when the CO concentration approaches 1%. This is not unex-
pected since this analyzer has a long (50 cm) cell. The statistics for
the CO analyzers are given in Table 6-29. In this day, the CO statistical
summaries are a good indicator of the functional interrelationships be-
tween the instruments. For the pair, Ch 11-Ch 16, the correlation coeffi-
cient is high, the slope is near unity, and the t-test shows no significant
difference between the sets of readings. For the pairs involving Ch 7, con-
versely, the slopes depart considerably from unity, and the t statistic,
although not significant at the 95% level, is fairly large.
Hand calculated summaries were also calculated for this day,
as a check on the routine. It is interesting to note that, for the Ch 11-
Ch 16 pair, the tr was significant, indicating that the departure from a
6-61
WALDEN RESEARCH CORPORATION
-------
a
ui
o
£
o> go CD -0
oc op a
m cc o
* o> o
o> o OD
p- o OD
a o oo
a
e
o
o
m P- M
o x m M
X _l U
v>
*
Z 00
o mo
» m
o
o
O
o
(C
^
o
(V
<9
O
(VI
o
OD
m
§ui x
a u
30
o ui
u. a
x f
(Vj
CO
o x
O P-
o o cr
o «r
I r*
i i
Z UJ U
O »
O (VI
o (vi
o «
cc
a
4
U
o
X
o
> I/)
x uj
_l
a
* T
* «
i/>
: o
i
' UJ
o ->
U. l/t
r>
£
r
u
in
x
u
X
u
X X
u o
M
m
p-
in
«
*
*
a
in
x
(VI
*
OD
m
P-
p*
*
(VI
op
ir IT r-
1
ill
u o u
a
ui
I/)
p- (VI
Z
o
*
o»
m m
^ *
o
«/> or i m
z o *
CC I
or
c t- it
o a x
o uj in *c
a. o « «
a a <
>- UJ I (VI *
it- u o» (vi
z m (*i
- i «
* z * o
O 3D O
in o o
>-1/1 . o
cc i/>
o ui x mo
Z Z <_> CC
O UJ
u a
o o-
o a
c o P-
IVJ
(VI
(VI
cc
in
in
cc
cc
n
in in o
x
u
in in m
x x
u u
XXX
u u u
1
u
6-62
WALDEN RESEARCH CORPORATION
-------
2
r^
X
O
00 Z CM
esj o i»^
. .
UJ (_> >-
5-
Q£
a
a
§
.
irifirccccc >~>->-rv(\.(\.(\.-iviv(rf-«-««.tf«iririririro
oooooooocooooooooocococcccooccoococccoccoooccccooo
Cfacroc-oocoooooocooooccccooccc ^ rv
g_g3
WALDEN RESEARCH CORPORATION
-------
UJ
a
ui
o
M
X
o
z - o « t^ « M -< oc o m rvj « x a> t-- CD ODCD m * fn in in
O .
o
* x
o
o
LlJ O ^*CfM*^*»*^«^*^*>*
I u. ...««.««.«.«.«.....«..«.««......«.........
OQ Z ^-»«(\l^^^«(\j(XJ(V
-------
a
uj
a
a
i
i c
CE
c z
«- u.
X
o L-
z <-> o
O U.
z u.
UJ I
zo o
ou
a
a z
« o
o c
o o
in o
» o
x c
O
O
a.
ct
« Z in
> o
a <->
c
o
(V
(V,
c
* U-
a
O
>- UJ
at a
X
-
a c
«i
o z
c
O
o
-.00
o o ry
o
I
U
a
c
O » ^^
a
UJ
o
UJ i/)
a. »-
i c
tr
I
o
on I o-
^ -u o
C _1
3 a
a i
4 <
OU1
u.
* C 4
* T a-
or i
c
»
o-
c a
Z UJ
0 U
2 a
UJ
z >-
c z
or -
a
a Z
o o
p-4
U-
3 a:
c
z
3
C
b.
r-
c >
^H ^
r
u
ji
in
. in
IT
T
U
C
0
r^ o
c
(T
o
o
1
o
o
o
c
o
r^
c
M
f
1^ 0
TIT
6-65
WALDEN RESEARCH CORPORATION
-------
slope of unity was not due to randomness alone. Thus, it would appear
that Channel 11 reads systematically about 10% higher than Channel 16.
The HC and CO data for 30 May are given in Table 6-30. The
Beckman 6800 (Ch 3 and Ch 5) does not appear to be functioning well be-
cause of the very rapid change of HC levels in short periods of time.
Ch 0, for example, goes from 56 to 1018 ppm in a 5-minute period. At
the same time, the CO concentration on Ch 11 increases from an estimated
0.2% to 2% during the same 5-minute period, and the other CO channels
also showed sharp increases, although of less magnitude. The HC NDIR
(Ch 15) seems to follow the fluctuating level well, as it is quite
highly correlated with Ch 0, with a coefficient of 0.91. The statistics
for the three CO channels indicate about equivalent degrees of correla-
tion, but show a low slope for the curves involving Ch 11 as follows:
CO (Ch 16) = 0.94 CO (Ch 7) + 0.2; r = 0.94
CO (Ch 7) = 0.23 CO (Ch 11) + 0.25; r = 0.92
CO (Ch 16) = 0.20 CO (Ch 11) - 0.48; r = 0.86
The results for 1 June are quite similar. The HC and CO data
are shown in Tables 6-31 and 6-33, respectively, with the statistical sum-
maries in Tables 6-32 and 6-34. The Beckman 6800 appears to be more highly
correlated with the FID than on the previous day. The slopes for the CO
relationships are quite different than those of the previous day.
The HC data for 2 June are given in Table 6-35. The HC concen-
trations are extremely high (Ch 0) and range from 50 ppm to 2.3%. Ch 5
behaves more erratically on this day than previously, probably as a result
of the exceptionally high HC levels. The statistical data are tabulated
in Table 6-36. Note that all the HC analyzers have low correlation coef-
ficients compared to Ch 0. The CO data for 2 June are in Table 6-37, and
the statistical summary is given in Table 6-38. The CO values range from
0.5 to nearly 16%. Both Ch 7 and Ch 16 are not functioning properly. The
correlation between CO and HC emissions on this day is not very good, with
a correlation coefficient relating Ch 0 and Ch 11 of 0.51, i.e., only about
25% of the variability can be explained by the regression.
6-66
WALDEN RESEARCH CORPORATION
-------
TABLE 6-30
FOUNDRY - May 30, 1972
if
Hydrocarbons (ppm)
Time
10:14AM
10:16
10:20
10:22
10:30
10:35
10:45
10:55
11:10
11:20
11:30
11:45
12:OOPM
12:15
12:30
12:52
12:56
13:10
13:30
14:00
14:31
15:15
15:30
16:00
16:30
*
Expressed
Ch 0
28
30
56
56
56
1018
732
339
339
339
310
255
256
506
338
284
312
537
762
1101
508
537
987
791
570
as CH4,
Ch 3
_-
--
--
--
83
--
85
75
73
84
88
58
98
99
104
250
250
103
86
250
250
except
Ch 5
67
75
31
48
53
79
210
228
261
235
262
:
210
171
267
1178
245
272
533
237
382
408
131
Ch 1
100
134
107
125
118
368
246
146
128
149
133
183
131
175
181
255
281
333
230
230
320
284
143
Ch 15 which is
5 Ch 7
0.020
0.046
0.090
0.095
0.132
0.502
0.499
0.502
0.498
0.493
0.493
0.493
0.502
0.495
0.499
0.900
1.005
1.300
1.050
1.300
1.300
1.060
0.900
expressed as
CO Data (%)
Ch 11
0.045
0.110
--
--
--
1.970
1.620
0.640
0.735
0.735
0.840
0.735
0.650
0.950
1.900
1.965
2.150
4.001
3.668
4.322
5.072
3.986
2.200
ppm C2Hg.
Ch 16
0.055
0.055
0.070
0.200
0.250
0.980
0.920
0.640
0.765
0.765
0.780
0.655
0.655
0.795
0.855
0.965
1.175
1.240
1.240
1.265
1.330
1.240
1.050
6-67
WALDEN RESEARCH CORPORATION
-------
£
fib
- n ocff« r f\
oo a
z <
O o
rr» C IT (\)(\jcirM '''
rv a < ruMiCQcm
^ c .....
rr > i «""«ni/i
~ ~l u (v M -»-»- »-c»(y»crf'«'»rr'''(V't\.- r *
LLJ ijj ir t-./r»f>r»(7ar*-!Te*-f-^' rr^-^^jf ^ocir"
%"
5 5 £ j ,cir»<*»^^iv.^*ir*ir-iriiV<>V--J^(vVi-r^(v^ rv*
I H- 3 * ,»,*('--- »*
S r~ c
O u.
T rrvr\ cocccccecT'-^c*f^'^r\T' n^. ^cfva"« r«T'*"^xcrrv^>r
C) ^
-------
u-.-ir«ririr>o»-e(»'.--*ev.ec^«nor'«raa
2 o * * e « a r - oc » » in (r s^-roir«o>mirir«r- (vino x^fvcv^jin* & & t*\ m
i a .»»««»«»«« .......
IO C. I ^c*r-a
2 C! craira ir^> TIM ix
*
* *
6-69 WALDEN RESEARCH CORPORATION
-------
f
O K
a u. *
> ui x
1C l_>
if
in
o.
o
V
C
a
if i
a
O
O
U
»
fV
O>
O IT
a u>
ca. ...
VOX tvj «
c
o
c
o
CV
"VI
(T
f
* z
O
U _J
C U.
? a
31
C C
U. <-
O
C
O
X
t
o
I/
X
IT
r
r
o
V) T
2 O I
C
a v
< u ic
O O IP
O Z IP ^-
at ui x
a i «
v ui x i
i u u
a: c
c a
2 UJ l-
D >W
C
U.
a
ui
c
f o a
* -f \r
c «
IT a:
z
a i
o
* t- >c f
u a » T
a u -* o ^
C or . .
> ui X « *
T >- C
z rv
~ o
» e
f c
c c
a is
c u. i
? a «.
c u.
U. 2
T
I.
a
(V
o
r
z
o
IP f
I I
o u
e IT tr f
r t r r
6-70
WALDEN RESEARCH CORPORATION
-------
HH O
x 2 »c
o o -«
X ^(y^^^^^^^4^ -4 i\j r>> m m PI m fv* (vi (M
o
n z CM
co o r-.
i CQ at
...
l i * »
?5 » I ^ii- (v;.- >^- ^'^- ^.^i^. < ^ iv r> r- f. rn f iv (v tv
£r
g ^
O
>- 2
ac 3 h> «vh-^r-r~(viv»vi^o-p'^rvcr'c:»^^-o»^ «r a-fvh- oKr'^^rr>n-*-"ofviccoa3X.cr*>.*otr>.-
-------
u
a
4->
C
O
CO
CO
I
if)
LU
o
2 *
o
T
I
? u
T
a
4
O
I
o
(t
O
c. >-rv(v*r-rv4
-------
u
O.
a
u
O
U V
a. ->
c z
- u
y - * *
00 T »
2 - -C * O-
C b, ^ C
I <_>
C
c: 1
-
a: uj
< a
3£ ii
z: x
=> o
"g
r r
o i-
i
r
b
a
O
2
.
c u.
T -
a o
(V.
*
o
o
IT
(T
I
(J
o
T C
a
bJ
D
b U
a
to
b.
C
CO
I/I
Z b
C _
T a
a s
< <
» c
* i
u <-
M
V -
I If:
O
T r
u o
3 a *
Z b.' X IT IV
O U c <=
S3
bj r
~* *- <-
C Z
I
(I
< Z oc o
t; c
O £
c:
-------
o
1-1
£ ___^
ro (\j-- in«n -ir^ « r^ incvj o>
, 5
o
to o in ^^-HCDOD^ooinrvjin'CinfMr>-r-«9-*9orri9-*o-*-r-«oDoor^9cD
z tx in99m»»if\iincocD^'O-*^)in-*-*<»>m-*r^---«-fvjoojin«o
o o
c£ x o 9oo<*)in4f^-GDM9(Mh»*«ino(Mr')rr)co^inrr)in'*<£(\iinoinini
o ^ "*^"^
m o CM *
co o£ r^ £
i Q en
I
OOOOOOOOOC O O O O O O O O O OOCOOOOOOOOO OO O
9990000000 oocoo^- i ^-«^« -"
*
* * ************
6_74 WALDEN RESEARCH CORPORATION
-------
I/) X
Z CJ
o
a.
o
O
> I
M 10
o
*
O *
in
a
c i
z <_>
D
O
c
I
iToiroiroiroircir oirct (v IT oir^cir oir circtfic.ir oirc-ircircu" ctrtc IT c ir
dirircc ^-<\iiv(rp'. «««irtrcc^-^-i\.(\.r'irr««irtrco>~>-fvi\.r'rr««irircc.<
cooooecceoccooocoooccoocoooooocooooooooooo
6-75 WALDEN RESEARCH CORPORATION
-------
i/> x
2 wi o
o -
It 2
a u.
« i
u o
O «-> 1/1
IX U.
0 U.
> UJ Z
1C U
u
* 2
O
.
O X
c o
u. O
I
U
a
a-
f
M
o- o
- o
r» o
f e
o
f^
r
*
»
»
a
c.
o
o
o
o
o
a.
UJ
a
in
*
o
e
*
o
z
c
O
cc uj
oa
* z in o
o f^ o
l/l < O
> «1 « C.
IX I/I
O UJ Z ~
2 (X U
u a
c.
o
o
t
n
a
ru
»
o
r-
r*
>t
c if if m
z a i i
vo
W" Z
z u
c
X
Q |^
< u.
u c
o z in
oa "
z u. u
u.
* c
*
^
o in
X
a i-
C I/) Z
Z UJ U
0
u. t-
If
1
<",
^
fs.
in
i
p-
^
gp
-C
1
0
1
G
0
f-
IM
1
O
o
o
o
o
c
o
o
e
o
o
o
r~
c\,
C If If C.
Z I Z I
-c
(V
X
If
or
-c
Z
U
O
a
a
o
3
* <
I/)
a o
C I
Z UJ - aJ
z i-
2
*
2
O
> l/i
(X Ifl
c uj
Z It
3 IS
O UJ
u. a
«
»
a
tv,
OD
I
(M
O
IT O
a- 4?
-« IV
c
-c
Cl
o
Ifl
(V
f
J1 f-
a
c
a
>£
I
u
\r
n
o
o
-J
c
X
-o
o
o
z z z
u u o
z
o
z z
o u
6-76
WALDEN RESEARCH CORPORATION
-------
UJ
o
X
c
.
o -
o x
X
o
O
a.
a
ro
i
vo
UJ
_i
CO
-------
o
^ -o
o -<
x
z u
a
5
CO
vi
*~ r- IT
6-78
WALDEN RESEARCH CORPORATION
-------
a
>
a
ui
o
o.
I
u. »-
c z
- u
X - \l
o «. a
z - c
c u. o
z u.
ui I I
z o o
o o
a
3T Z
< O 0»
(j ~ jn
t- r
* _l
* U
a
2
> O
a »_
i
O
(V)
un
i c
2
UJ
C I
x c a
C O IT
Z *>**
o UJ ~* *c ^
za
o i un i
z _i u i
Cl/1
a
3
< Z -C c
UJ
c
u
UJ
a.
i
u
T
O
00
CO
VO
UJ
o
^1/1
X UJ
C c
i z -o
O UJ ~
z a
ui z
Z u. U
c u.
a -
a c
<
02
c ~«
t-
vi I
UI U
a -
c
z
D r-
C
u
I
u
i ^
e; o
c c
o
o
o
O
o»
»
r*
tt
oc
r-
a
UJ
Cl
UJ I/I
a. -.
i c
at
c
z
o
2
I/)
ZuJ
C _1
2 a
3 J
o a a> i>-
x ^ r^ f^
ui u
'NJ
c
u a a r-
X h- h- r-
u
X X
o u
w t~
c a
Z UJ
0 0
z a
UJ
Z l-
0 Z
a -
3
« Z
O C
1-1
I/
I/)
UJ
tir
a
.V1 W
or or
c
x
C
«
r^
c -*
< r\j
X *
u *
-c
l/l
cy
f~» fr
I r-
o ^-
c
c
^- c
c
3
*n
OJ
c
>-
^
a
r*i
CJ
r
m
fV
*
X X
(J U
r-
-------
At about this time, considerable difficulty was encountered with
the Esterline Angus (Model D 2020) data acquisition system, as evidenced
by missing points or spurious values. Since the computer reduction pro-
gram did not have a data screening capability, the remaining data were
treated by hand to avoid errors in the computer mode. The scrubber fan
speed was increased to improve particulate collection efficiency. This
resulted in the lower HC and CO values obtained during the last week of
testing at the foundry.
The HC and CO data for June 5, 6, 7, 8, and 9 are given in
Tables 6-39 to 6-43, respectively. On June 6, CO Channel 1 was brought
back on line. A smaller sample loop was used in the chromatograph for
these tests. During these tests, the Beckman 6800 was not functioning
properly because it had been set up for the higher HC levels encountered
initially at the foundry. The instrument was reading <10% of full scale
on most of the tests. We had found previously that the instrument did
not give very accurate results unless the reading was >10% of full scale.
The sensitivity could not be increased because of the high background.
We did not want to set the instrument up again (because of the inaccura-
cies in the results), therefore, the results on CH. and THC (Ch 3 and
Ch 5) are not very useful in Tables 6-39 to 6-43.
The statistical data for June 5-9 are tabulated in Tables 6-44
to 6-48. Channels 0 and 15 were the only HC channels providing data re-
liably during these five days. The correlations between the two are in-
cnnsistent. The correlation coefficient ranges from 0.18 to 0.81. The
relationship between CO and HC, as manifest by the correlation between
Channels 0 and 11, was also quite inconsistent. It is interesting to note
that the degree of correlation between Channel 0 and 15 was about the same
as that between Ch 0 and Ch 11, and varied in the same manner.
As in earlier data, the interconnections between the pairs of
CO monitors was quite good, often greater than 0.95. The slopes and inter-
cepts varied from day to day, however.
6-80 WALDEN RESEARCH CORPORATION
-------
TABLE 6-39
FOUNDRY - June 5, 1972
Hydrocarbons (ppm)*
Time
10:OOAM
10:20
10:30
10:40
10:50
11:00
11:15
11:30
11:45
12: 00PM
12:15
12:30
12:45
13:00
13:15
13:30
13:45
14:00
14:15
14:30
14:38
15:00
15:20
15:28
Ch 0 Ch 3
44
18
14
22
26
18
12
8
7
7
6
5
12
89 2
80 >250
44 16
31
177 55
290 98
291 100
78
36
298
55
Ch 5
58
78
68
67
69
58
49
43
37
35
34
31
30
52
50
45
37
78
114
119
369
55
55
360
Ch 15
321
150
138
143
134
131
138
133
131
134
130
136
137
140
175
187
166
172
225
275
625
315
147
135
Ch 7
0.632
0.079
0.075
0.265
0.317
0.175
0.095
0.058
0.028
0.037
0.095
0.037
0.095
0.330
0.318
0.200
0.228
0.423
0.600
0.900
1.025
1.340
CO (%)
Ch 11
0.365
0.080
0.080
0.270
0.270
0.190
0.105
0.080
0.030
0.050
0.110
0.040
0.110
0.420
0.420
0.245
0.255
0.600
0.970
1.210
1.480
1.735
Ch 16
0.050
0.015
0.015
0.030
0.040
0.030
0.015
0.088
0.028
0.031
0.110
0.032
0.110
0.315
0.280
0.180
0.195
0.475
0.900
1.200
1.590
1.800
6-81 WALDEN RESEARCH CORPORATION
-------
TABLE 6-39 (Cont.)
Time
15:30
15:38
16:15
Ch 0
188
296
17
*
Expressed as
Hydrocarbons (ppm)
Ch 3 Ch 5
360
86 340
42
CH4, except Ch 15
Ch 15
149
149
250
which is
CO (%)
Ch 7 Ch 11
0.775 1.230
expressed as ppm C«l
Ch 16
1.520
"6-
6-82 WALDEN RESEARCH CORPORATION
-------
TABLE 6-40
FOUNDRY - June 6, 1972
*
Hydrocarbons (ppm)
Time
10: 00AM
10:10
10:40
10:50
11:30
11:40
11:50
12:00
12:10
12:20
12:30
12:40
12:50
13:00
13:10
13:50
14:30
14:46
14:50
14:54
15:20
15:30
15:42
15:46
15:48
16:10
#
Expressed
Ch 0
10
26
184
112
56
334
228
99
38
31
15
10
35
51
74
5
19
160
81
148
21
39
94
46
92
27
Ch 3 Ch 5
78
81 6
98 93
62 94
104 74
22
45
45
31
34
25
22
24
25
19
28
6
-
-
-
12
13
-
-
-
136
as CH., except Ch
Ch 15
108
158
143
124
100
103
100
93
91
94
103
100
91
93
93
85
127
-
-
-
174
214
-
-
-
187
15 which
CO Data (%)
Ch 1 Ch 7
0.315
0.700
0.700
0.650
0.440
0.322
0.223 0.240
0.134 0.118
0.054 0.060
0.048 0.050
0.048 0.032
0.041 0.042
0.061 0.060
0.080 0.098
0.074 0.158
0.310
0.825
1.050
1.650
1.175
is expressed as p
Ch 11
0.320
0.800
0.870
0.800
0.585
0.380
0.270
0.155
0.090
0.090
0.065
0.070
0.095
0.145
0.200
0.385
1.435
1.540
2.480
2.580
.p. C2H6.
Ch 16
_
0.830
0.700
0.640
0.380
0.200
0.140
0.065
0.050
0.040
0.045
0.045
0.055
0.062
0.095
-
1.350
1.340
1.740
2.380
6-83
WALDEN RESEARCH CORPORATION
-------
TABLE 6-41
FOUNDRY - June 7, 1972
Hydrocarbons (ppm)
Time
9: 30AM
9:40
9:50
10:00
10:10
10:20
10:30
10:40
10:50
11:00
11:10
11:20
11:30
11:40
11:50
12: 00PM
12:10
12:20
12:30
12:40
12:50
13:00
13:10
13:20
13:40
13:50
Ch 0
21
31
50
22
14
9
17
9
25
23
46
15
133
178
34
18
320
63
38
58
160
236
114
91
43
84
Ch 3
2
-
11
-
-
-
-
-
-
-
-
-
44
30
3
-
-
15
6
-
69
87
56
37
-
-
Ch 5
4
40
69
63
49
39
36
34
32
35
32
54
66
59
42
31
34
41
40
32
61
75
71
54
-
Ch 15
102
110
110
93
88
88
100
127
146
157
189
225
221
200
178
171
163
183
165
183
215
230
189
178
-
Ch 1
0.117
0.335
0.140
0.180
0.186
0.250
0.174
0.280
0.260
0.380
0.690
0.600
0.430
0.240
0.210
0.310
0.410
0.400
0.510
0.630
CO Data (%)
Ch 7
0.135
0.300
0.675
0.382
0.380
0.210
0.282
0.198
0.322
0.340
0.405
0.500
0.600
0.420
0.310
0.240
0.365
0.498
0.430
0. 0
0.600
0.675
0.500
0.425
Ch 11
0.140
0.335
0.880
0.455
0.310
0.255
0.335
0.265
0.400
0.400
0.575
0.825
0.840
0.610
0.360
0.280
0.480
0.585
0.585
0.585
0.835
0.975
0.675
0.615
Ch 16
0.065
0.145
0.700
0.255
0.140
0.090
0.140
0.095
0.165
0.180
(*
0.310
0.560
0.570
0.390
0.165
0.100
0.225
0.335
0.345
0.320
0.585
0.755
0.430
0.385
6-84
WALDEN RESEARCH CORPORATION
-------
TABLE 6-41 (Cont.)
Hydrocarbons (ppm)
Time
14
14
14
14
15
15
15
15
16
16
:00
:15
:30
:45
:00
:15
:30
:45
:00
:08
*
Expressed
Ch 0
43
48
169
100
67
130
151
46
13
5
as CH4,
Ch 3 Ch 5
3
15
66
50
28
-
70
5
-
except
41
46
62
65
57
33
87
51
39
39
Ch 15
Ch 15
152
159
218
56
41
219
203
174
171
163
which
CO Data (%)
Ch 1 Ch 7
0
0
0
0
0
0
0
0
0
0
is expressed
.350
.525
.800
.625
.618
.745
.725
.800
.950
.6750
as ppm
Ch 11
0.
0.
1.
0.
0.
1.
0.
1.
1.
0.
c
490
675
190
850
970
060
980
160
380
8950
2H6"
Ch 16
0.
0.
0.
0.
0.
0.
0.
0.
1.
0.
210
420
950
600
740
720
875
940
180
700
6-85 WALDEN RESEARCH CORPORATION
-------
TABLE 6-42
FOUNDRY - June 8, 1972
Hydrocarbons (ppm)
Time
9: 40AM
9:50
10:00
10:10
10:20
10:30
10:40
10:50
11:00
11:10
11 : 20
11:30
11:40
11:50
12: 00PM
12:10
12:20
12:30
12:40
12:50
13:00
13:10
13:20
13:30
13:40
Ch 0
63
171
94
141
79
113
141
103
178
202
137
76
65
136
147
94
42
25
36
17
8
-
-
-
Ch 3
-
5
29
46
2
50
62
55
10
77
40
30
12
44
72
45
4
-
2
-
Ch 5
16
24
27
28
24
25
25
25
22
25
23
21
19
20
23
20
18
15
16
15
14
15
14
12
14
Ch 15
80
76
74
75
72
78
81
85
99
97
105
97
96
106
108
102
93
105
88
93
87
85
93
81
91
Ch 1
0.066
0.139
0.194
0.181
0.134
0.123
0.123
0.178
0.110
0.124
0.128
0.126
0.084
0.109
0.186
0.180
0.134
0.064
0.030
0.030
0.018
0.010
0.016
0.012
0.015
CO Data (35)
Ch 7
0.148
0.335
0.265
0.372
0.212
0.215
0.182
0.225
0.210
0.272
0.212
0.140
0.160
0.280
0.265
0.222
0.120
0.058
0.055
0.032
0.021
0.030
0.021
0.035
0.013
Ch 11
0.062
0.385
0.300
0.310
0.280
0.250
0.245
0.080
0.270
0.300
0.270
0.180
0.210
0.345
0.345
0.280
0.155
0.090
0.095
0.058
0.045
0.115
0.020
0.040
0.035
Ch 16
0.066
0.190
0.130
0.130
0.105
0.095
0.090
0.118
0.098
0.125
0.095
0.060
0.065
0.140
0.120
0.115
0.050
0.020
0.025
0.018
0.015
0.015
0.012
0.010
0.012
6-86
WALDEN RESEARCH CORPORATION
-------
TABLE 6-42 (Cont.)
Hydrocarbons (ppm)
Time
13:50
14:00
14:10
14:15
14:17
14:30 1
14:45
15:00 1
15:15
15:30 1
15:45
16:00
*
Expressed
Ch 0
232
661
109
-
-
335
349
754
122
324
617
301
as CH4,
Ch 3
95
-
102
-
100
45
104
87
-
102
83
except
Ch 5
16
107
140
-
119
94
184
89
405
220
110
Ch 15
Ch 15
109
155
407
-
454
234
426
205
396
203
199
which is
CO Data (%}
Ch 1 Ch 7
0.065 0.335
1.400 0.750
6.500 2.000
4.100
2.200
4.100
0.850
2.950
3.650
3.650
expressed as ppm
Ch 11
0.630
1.720
5.300
5.500
4.100
3.800
2.100
4.200
3.200
3.575
C2H6.
Ch 16
0.165
0.760
2.800
3.400
2.350
2.700
1.330
2.950
2.550
2.650
6-87 WALDEN RESEARCH CORPORATION
-------
TABLE 6-43
FOUNDRY - June 9, 1972
*
Hydrocarbons (ppm)
9: 30AM
9:40
9:50
10:40
10:50
11:00
11:10
11:15
11:18
11:30
11:40
11:50
12: 00PM
12:10
12:20
12:30
12:45
13:00
13:15
13:25
13:30
13:45
14:00
14:30
15:01
16:00
*
Expressed
Ch 0
49
25
16
16
4
14
31
61
19
79
130
108
89
120
116
60
271
271
271
54
500
1015
461
888
119
as CH4,
Ch 3
84
92
39
-
-
-
-
-
-
33
34
45
28
16
56
20
86
101
103
89
103
except
Ch 5
79
38
53
46
29
25
23
-
-
29
45
58
56
50
56
60
29
172
167
159
158
251
182
271
116
Ch 15
Ch 15
109
81
74
75
88
109
121
-
-
171
189
40
178
187
177
156
370
430
367
334
361
456
330
430
218
which is
CO Data (%)
Ch 7
0.925
0.550
0.395
0.245
0.270
0.325
0.365
0.620
0.675
0.550
0.625
0.700
0.950
0.950
3.400
1.925
1.450
1.500
2.075
2.450
2.825
2.300
1.175
expressed as
Ch 11
1.480
0.760
0.510
0.300
0.210
0.410
0.485
0.970
1.110
1.050
1.180
1.390
1.460
2.040
5.000
4.250
3.950
3.875
5.400
7.600
7.100
5.150
3.400
ppm C2Hg
Ch 16
1.460
0.540
0.300
0.120
0.080
0.190
0.255
0.720
0.890
0.800
0.940
1.200
1.245
1.640
4.025
3.565
3.215
2.350
3.200
3.500
4.450
3.050
1.800
6-i
WALDEN RESEARCH CORPORATION
-------
CVI
r^
CM
in
3
TJ
S-
o
I- (->
-!-» C
(O 0)
O) U
J- -r-
S- M-
O M-
O O)
O
O
QQ
<
OO
OO
II
fee
00
Q.
O)
O)
CO LT> r
CVJ ID i COOr CO <£>
C\J
^" CT> CO
CVJ i
I I
O CVJ CO
cr>CT>Lf>Lr)cocooo
CDCDCDCDCDCDCDCDCDO
(£>
IO
co
CO CM
o o
O CD
I
CO i O i
r t r r
CD O CD C3
CO
OO CM
CO
CT>
CM r
ID r
CVJ O r O
O i
LOOi
r-.
CO
r . O
6-89
WALDEN RESEARCH CORPORATION
-------
un
UJ
CO
CM
cr»
C
as a)
S-T-
S. 4-
O 4
O O)
o
o
to
o
-p
Q.
O
O)
a.
o
00
UD
CQ
LO 00 1^ ID CM
IO O Lf) r- LO
OCM
ID r- CM VO CO
r- I I
CM
O r
COCOrOfOvoVDMCOcTiCnCOCOOO
^ r~ ^* r^ CTi CTl CTi Cn CTi CT* CO CO O^ CTl
OOOOOOOOOOCDOOO
en in LO
-------
CVJ
O)
c
3
t-> C
ro o>
1- -r-
O M-
O O)
O
I
VO
CO
oo
<
tI
I
oo
ii
fee
oo
Q.
1)
s_
0)
(U
o
r
00
F CO C\J I
un LO «^- i c\j
VX3
i
cooocoocMOi^-ouncoo
II I I I i r I
«d-«*OO^J-^3-VDlOOO.CVJC\jOOCOCT>CriOO
ooooo'ooooo'oooooooo
r *d-i O
(XI
o
*3" CTl
o ^
LD LT>
C\J O
CO
CM
VO 00
co c\j
ooo
ooooooooo
I I I I
o
O LT>
CO VO
0000
OOOi O
OOO
o i tn in o
Or LOr OLDr f^-r^
6-91
WALDEN RESEARCH CORPORATION
-------
«3-
1
vo
UJ
co
t
CVJ
r-.
en
00
O)
c
-3
\
^H
QL
§
CO
_l
O
H- 1
CO
t 1
co
i_
4-)
S-
o
l~ 1 *
-t-> c
. in
CM
1
^_ ^_
00 CO
O CD
r>- o
co m
CO f~>
T
CM «*
CM O
O CO
in o
o in
t
6-92
WALDEN RESEARCH CORPORATION
-------
CM
cn
*~
cn
OJ
c
3
00
^^ 1
1
VO >-
at
LLJ C
to
Q.
U
i^
O)
C
^
(U
o
O
-
X
VOCMt^CMOO ^1-^J-CMI
cncncn«a-vooo r i^» oo CM o o
cn ^* ^" cn ^o cn CM ^* cn CM *^ LO
III 1 1 CM
1 1
LO LO ^o ^o oo oo r-~ ^^ ^j* <""> c^ CM CM CM cn cn LO i^o
cncncncy>cncncn^ovDvDV£)OOcooor-~r--.ooco
oooooooooooooooooo
CVJ
CM cn cn CM r ^~ cn
CMOooi CMCM cn cn CM vor^
oooooo i if> in r^. CM oo i cn m uo
III 1 OO P^ CM OO 1 ^" ^J" "!^ OO
r r \
1
oo^J"OOc\junr~- cnoi i£> r^-tocn^-ooo
Lf) LO ^* VO OO OO OO OO CM * ^" ^" OO VO CM r
O i i O CM O iOCMO r O O i O 00
10 f"* vo r^ r r^ i^ LO oo LO oo ^ LO LO LO co LO c^
r lOr^lOr^r-LOOOLOOOLOOLOLOOLOOLO
6-93
WALDEN RESEARCH CORPORATION
-------
To provide a check on the instruments at the foundry, a number of
samples were collected and analyzed by the chromatographic technique de-
scribed in Section 3.2. The data are given in Table 6-49. The total HC
on the manual analysis is simply the sum of CH., C^H. and C2Hg. The agree-
ment between the manual GC analysis for THC and the FID (Channel 0) was
excellent, as shown by the regression equation as follows:
THC (Ch 0) = 0.96 THC (Manual) + 4.6; r = 0.94
The slope is close to unity and according to the t-test, not significantly
different from unity. This result supports the assumption that the only
hydrocarbon species present in significant amounts are methane (CH.),
ethane (CpHo), and ethylene (C^H.). The manual analysis indicates that
the methane content of THC remains at a constant percent over the range
measured, with very high correlation, as shown below.
CH4 = 0.77 THC + 6.35; r = 0.99
Agreement between CH. as measured by the Beckman 6800 (Ch 3)
and the manual analyses was poor. The THC channel (Ch 5) also did not
correlate well with the manual analyses, as shown by the regression equa-
tion following:
THC (Ch 5) = 0.35 THC (Manual) + 46.5; r = 0.69
A comparison of wet chemical and instrumental methods is given
in Table 6-50. Examination of the data appears to indicate that Channel 7
saturates at levels above 1% CO. The statistical summary as shown in
Table 6-51, does not bear this out. The correlation coefficient for the
relationship between all three CO channels and the wet chemical analysis
exceeds 0.96. The slopes are quite different, however, and tend to indi-
cate different response characteristics of the various monitors.
A comparison of the THC by the Beckman 400 and Beckman 6800 is
shown in Figure 6-11. The figure depicts the different characteristics of
slow and fast response analyzers. The slower chromatograph is not suitable
for this type of highly fluctuating source. The average HC concentrations
measured by the two are considerably different being approximately 550 and
6-94 WALDEN RESEARCH CORPORATION
-------
TABLE 6-49
COMPARISON OF HYDROCARBON VALUES BY MANUAL GAS CHROMATOGRAPHIC ANALYSIS
AND CONTINUOUS HYDROCARBON MONITORS FOR A GRAY IRON FOUNDRY
Date
5/30/72
5/31/72
6/6/72
6/9/72
*
Expressed
Time
11:00-11:20
11:25-11:45
9:50-10:05
10:40-10:55
11:30-11:40
13:45-13:55
14:26-14:35
15:20-15:30
9:30- 9:55
10:30-10:50
11:30-12:00
13:30-13:50
14:30-14:45
15:45-16:00
as ppm C
*
ppm HC
(Continuous Monitors)
Ch.O Ch.3 Ch.5
339
604
18
148
56
5
160
30
32
16
105
275
461
119
75
>250
80
52
39
33
101
89
261
--
94
74
28
15
46
38
45
172
182
116
ppm
with
THC
155
584
95
129
70
22
113
24
30
36
102
331
519
195
HC (Manual Analysis
Gow-Mac Chromatograph)
ru c u r u
LH4 L2 4 24
132
424
70
112
50
22
103
17
30
36
102
271
437
130
33
160
25
17
20
10
7
60
82
65
6-95 WALDEN RESEARCH CORPORATION
-------
TABLE 6-50
COMPARISON OF CO DETERMINATIONS BY WET CHEMISTRY AND
CONTINUOUS MONITORS AT A GRAY IRON FOUNDRY
Date
5/25/72
6/6/72
6/9/72
CO (%)
Pd (0-phen)
15:22-15:46
14:34-15:20
11:00-11:20
11:25-11:45
9:50-10:05
10:40-10:55
11:30-11:40
13:45-13:55
14:26-14:35
15:20-15:30
9:40- 9:50
10:30-10:55
11:30-12:00
13:30-13:50
14:30-14:45
15:45-16:00
2.
1.
0.
1.
0.
0.
1.
2.
0.
0.
0.
5.
11.
2.
07
47
35
10
35
36
43
43
44
12
92
77
54
54
co2
12.
3.
15.
8.
5.
14.
15.
15.
15.
13.
7.
9.
15.
15.
14.
15.
U)
5
2
3
7
6
4
5
0
6
8
5
7
2
0
3
8
Or sat
o2 (%)
5.
18.
5.
12.
15.
6.
4.
5.
4.
7.
12.
11.
5.
4.
4.
4.
8
3
2
3
0
2
8
9
6
1
6
2
0
0
0
6
CO (%)
5.10
0.75
0.93
1.05
0.35
0.60
0.30
0.20
1.05
1.30
0.55
0.30
0.75
3.25
4.90
1.00
L&N
3.42
2.80
1.50
1.25
0.32
0.84
0.48
0.39
1.44
2.01
0.63
0.25
1.07
4.63
7.10
3.4
CO (%)
MSA
2.00
2.30
0.67
0.29
1.35
1.54
0.42
0.20
0.94
2.86
4.45
1.80
Ch.7
0.32
0.68
0.38
0.31
0.83
1.35
0.47
0.26
0.62
1.80
2.83
1.18
6-96
WALDEN RESEARCH CORPORATION
-------
,
LO
1
VO
UJ
I
t
CO
f
oo
1
UJ
s:
Z3
1
oo
zn
H- 1
l/l
>
<:
t i
^~
jt^,
UJ
re
i
UJ
^£
>-
<
y^
s»*
rs
|
OO
t-H
I
00
S-
4-*
S-
£ZZ ^- ~*
O
r~ 1 *
+-> c
fO O)
CD CJ
S- T-
S- if-
0 M-
o
0
^~^
X}
'
4J
o.
cu
o
S-
O)
+J
c
HH
'E'
Ol
o
00
^
X
CTl
LO LO r^ vo
co r~- r~N CTI o LO LO
i CM r I
1 1 1
CTl CTl 00 CO 00 00 VO
CTl CTl CTl CTt CTl CTl 1^
O O O O O O O
00
00 00 LO i CM 00 LO
< CM CM IO CM CM LO
O O O O O O O
1 1 1
CTl tO LO r 00 ^1" CM
LO LO co r^ LO LO ^~
O r O CM O ' O
to
r*^ vo r^ r to ^~ i-
i i O
o
^
3
I n r r
O
0
LO
n-
10
r~.
o
to
o
CTl
CO
r
J^1
3
O
O
(0
to
s_
o
o
C_)
O LO VO i CTl
CO «* LO O CTl LO
i ^a- i i
1 1
r-~ r r^ r»- LO vo
CTl CTl CTi CTi CTl CTt
O O O O O O
r^. oo o oo ** ta-x f*^ ^m^f ^Q V-^X
r r
o o o
o o o
o" o" o"
333
i r
O O O
O C_3 0
6-97
WALDEN RESEARCH CORPORATION
-------
I
IT)
o
Q.
in
OJ
cc
C\J
2 r--
O CT
CO
t
-o o
c n
o
o»
o >,
CO S-
'-O
>>
t/) eo
(O $-
Li_ CD
c o
o
in
- t.
J- O
(O 4->
Q.-r-
§c
o
I
vo
(U
S-
OH uidd
6-98
-------
200 ppm for the fast and slow response analyzers, respectively. Curves for
a rapid response CO and HC analyzer are shown in Figure 6-12. The CO and
HC levels follow each other consistently. These curves and the data pre-
sented earlier support the requirement for a rapid response monitor, i.e.,
less than 5 minutes for non-steady state sources.
6.4 REFINERY TESTS
The tests at the refinery were initiated on 28 June and continued
through 13 July 1972. The instruments were run around the clock and the
data were taken every five minutes with the data acquisition system. Since
we were reducing the data manually, we only calculated the values every
half hour to reduce the analysis time. This approach provided an excellent
picture of the concentration fluctuations at the refinery (which were very
small) and sufficient data for statistical analysis. The calibration
curves for the instruments were the same as those for the foundry (see
Table 6-22).
The sampling system was 'the same as that used at the foundry
and incinerator. The sampling site selected for testing at the refinery
was the catalytic cracking unit. Most of the refinery stack emissions of
CO and HC come from this source if it is uncontrolled. Since a waste heat
or CO boiler is the most common method for control of this source, we used
sample sites at the inlet and outlet of the CO boiler for our studies.
Thus, the flue gas is representative of uncontrolled and controlled sources,
respectively.
The data obtained at the inlet to the CO boiler on June 29, 1972
is given in Table 6-52. Hydrocarbon concentrations are expressed as ppm
methane for Ch 0, Ch 3, and Ch 5, and as ppm ethane for Channel 15, through-
out this section. Ch 15, the Intertech NDIR (CpHg) is responding to a com-
bination of water vapor and CH4 in the flue gas. The water vapor response,
from previous experience with the analyzer is about 30-45 ppm. The device
also responds to CH. at a ratio of about 10:1, e.g., 300 ppm CH. in flue
gas is read as about 30 ppm on Ch 15. Since the device is not reading CH.
fi-QQ
D " WALDEN RESEARCH CORPORATION
-------
*HO SB SNOaaVOOUQAH
-------
TABLE 6-52
REFINERY (INLET TO CO BOILER)
June 29, 1972
Hydrocarbons (ppm)
Time
14: 00PM
14:25
14:30
15:00
15:30
15:40
15:50
16:00
16:05
16:15
16:20
16:30
17:00
17:30
18:00
18:30
19:00
19:30
20:00
20:30
21:00
21:30
22:00
22:30
Ch 0
296
321
324
312
314
314
303
307
315
318
315
318
301
310
325
311
311
325
325
341
355
355
355
355
Ch 3
296
306
314
290
272
288
268
264
256
292
292
280
238
288
264
246
238
242
246
280
288
322
322
338
Ch 5
419
228
188
249
284
256
265
276
329
276
249
339
212
213
216
169
196
Ch 15
76
81
85
76
76
72
72
68
76
81
76
81
72
72
78
64
72
64
68
76
85
76
81
Ch 1
18+
18+
18+
18+
18+
18+
18+
18+
18+
17.60
16.00
14.65
17.05
12.95
11.40
12.05
12.05
13.20
15.20
17.00
16.50
15.50
16.25
CO Data (%)
Ch 2
15.10
15.50
14.35
15.10
15.10
15.50
14.75
14.35
14.35
14.75
14.35
13.15
14.30
14.30
14.30
13.95
13.95
13.95
14.30
14.70
14.30
14.70
14.70
Ch 11
12.30
12.95
12.95
12.90
12.70
12.70
12.65
12.70
12.80
12.70
12.80
12.10
12.70
12.70
12.50
12.45
12.45
12.65
12.90
13.05
13.10
13.15
13.20
Ch 16
11.70
11.65-
11.70
12.10
11.30
12.05
11.90
11.90
12.55
12.45
12.10
11.10
11.55
12.45
12.10
12.00
11.90
12.15
12.35
11.90
11.65
12.80
11.90
WALDEN RESEARCH CORPORATION
-------
only (water vapor interferes), we will not discuss it further. Ch 0, Ch 5
(THC) and Ch 3 (CH.) should be reading approximately the same HC level since
only CH^ is present in this flue gas. Ch 5 tends to drift around because
the instrument was reading only about 10% of full scale. The attenuation
was changed in later tests and much better results were obtained. Ch 1
(CO) had been saturated with a high concentration of CO (leading to ampli-
fier "hang up"), and it took several hours before the "apparent" CO level
came back down. Ch 1 then gave values which are close to Ch 2 but note
that Ch 2 values are higher than Ch 11 or Ch 16. If we take the average
of Ch 2, Ch 11 and Ch 16, we note that the median value is just about the
concentration which Ch 11 is reading and that Ch 2 is 10% high while Ch 16
is 10% low. Most of the values come within O% of the mean CO concentration.
On 30 June, we see the same trend in the resulted observed on
29 June. These data are given in Table 6-53. Ch 0 and Ch 3 follow each
other but Ch 5 is drifting and malfunctioning. For CO, Ch 2, Ch 11 and
Ch 16 again are within ;HO% of each other and fairly constant while Ch 1
(chromatograph) displays considerable variation in the CO level. The data
for 3 July are given in Table 6-54. On this date, the HC channels (3 and
5) on the chromatograph were down and are not reported until 2AM. The data
acquisition system appeared to be putting out bad data for these channels.
Ch 1 and Ch 2 read CO concentrations which are one half the levels measured
by Ch 11 and Ch 16. The reason for this is not known but the original data
were rechecked and no mistakes were made in calculations. A check on the
process controls including the Op levels measured by the refinery personnel
indicate that the CO levels should be approximately the same on 3 July as
on 29 June. It appears that Ch 1 and Ch 2 are not measuring the correct CO
concentration while Ch 11 and Ch 16 are.
The HC and CO results for 4 July (Table 6-55) are similar to those
obtained on 3 July in that Ch 1 and Ch 2 are still reading low. On 5 July
(Table 6-56), Ch 0 went down and the attenuation of Ch 5 was changed to a
more sensitive position. Note that Ch 3 and Ch 5 show considerably better
agreement now. CO levels measured by Ch 11 and Ch 16 are within about +5%
of each other. On 6 July (Table 6-57), the Beckman 6800 broke down. This
6-102 WALDEN RESEARCH CORPORATION
-------
TABLE 6-53
REFINERY (INLET TO CO BOILER)
June 30, 1972
Hydrocarbons (ppm)
Time
23: 00PM
23:30
0:00
0:30
1:00
1:30
2:00
2:30
3:00
3:30
4:00
4:30
5:00
5:30
6:00
6:30
7:00
7:30
8:00
8:30
9:00
9:10
9:30
10:00
10:30
11:00
11:30
11:56
Ch 0
322
355
341
325
325
296
296
296
281
266
266
252
236
236
222
208
192
192
178
178
283
244
264
321
334
281
257
173
Ch 3
356
352
332
322
318
322
322
314
310
310
310
298
256
250
242
220
206
218
218
206
242
246
246
292
298
254
-
142
Ch 5
230
245
233
199
220
210
206
221
221
274
157
183
145
162
219
257
260
266
292
243
Ch 15
85
81
81
81
76
76
76
81
76
76
81
76
72
68
68
59
64
68
59
55
76
59
64
80
76
72
64
54
Ch 1
17.80
18+
18+
16.50
17.80
16.70
13.65
12.95
14.60
13.15
17.05
18+
14.90
15.70
16.50
16.70
11.20
12.90
10.80
12.25
16.70
18+
18+
17.50
11.05
9.40
CO Data (%)
Ch 2
15.10
15.10
15.10
15:10
14.70
14.70
14.30
14.70
14.30
13.95
14.30
13.95
13.60
14.30
13.95
13.10
13.50
12.65
13.10
13.10
13.10
14.35
14.35
14.35
12.70
13.10
Ch 11
13.45
13.25
13.20
13.20
13.15
13.10
13.05
12.95
12.90
12.75
12.75
12.45
12.25
12.15
12.10
11.95
11.70
12.00
11.95
11.95
12.95
12.85
13.45
13.50
13.05
12.60
Ch 16
12.80
12.35
12.65
13.00
13.00
11.60
12.60
12.10
11.90
12.45
12.45
12.35
12.40
11.25
12.30
11.90
11.50
11.90
11.50
12.45
12.70
13.10
12.55
13.00
11.70
11.60
6-103
WALDEN RESEARCH CORPORATION
-------
TABLE 6-54
REFINERY (INLET TO CO BOILER)
July 3, 1972
Hydrocarbons (ppm)
Time
11:29AM
12:00
12:30
13:00
14:15
14:40
15:00
15:30
16:00
16:30
17:00
17:30
18:00
18:30
19:00
20:00
20:30
21:00
21:30
22:00
22:30
11:15
11:20
11:26
11:27
Ch 0
162
183
186
192
182
137
131
168
157
140
153
171
165
147
143
198
162
143
139
175
176
143
161
149
163
Ch 3 Ch 5 Ch 15
56
61
61
61
51
41
41
51
46
46
41
-
51
46
51
61
51
51
61
61
61
Ch 1
5.09
4.96
4.88
5.08
0.55
5.05
5.06
5.10
5.11
5.09
5.10
5.10
5.10
5.08
5.11
2.50
5.09
5.09
5.07
4.70
5.09
CO Data (35)
Ch 2
7.95
7.95
7.25
7.25
6.60
6.20
6.00
6.40
6.15
5.85
6.20
6.20
6.20
5.80
6.20
6.80
6.40
6.40
6.00
6.60
6.60
Ch 11
12.50
13.30
12.50
12.80
12.70
11.80
11.90
13.20
13.20
12.80
12.95
13.20
12.95
12.50
12.20
13.50
12.90
12.50
12.10
11.80
12.70
Ch 16
9.60
10.85
10.10
10.10
10.85
9.80
9.35
10.10
10.10
9.70
10.25
10.10
10.50
10.10
9.80
10.15
10.60
9.80
9.90
9.90
10.65
6-104
WALDEN RESEARCH CORPORATION
-------
TABLE 6-54 (Cont.)
Hydrocarbons (ppm)
Time
23:00
23:30
0:00
Midnight
4 July
0:30
1:00
1:30
2:00
2:30
3:00
3:30
4:00
4:30
5:00
5:30
6:00
6:30
7:00
7:30
8:00
8:20
Ch 0
188
178
159
155
142
214
277
227
146
156
171
175
263
331
260
234
247
297
279
266
Ch 3
412
394
312
-
138
141
148
141
228
330
252
196
224
266
252
232
Ch 5
94
152
214
113
188
214
235
226
290
Ch 15
66
61
61
56
51
66
87
71
51
56
56
61
76
92
76
71
76
82
76
76
Ch 1
5.08
5.10
5.08
5.06
5.05
5.09
6.30
5.05
6.22
5.07
6.22
6.47
6.30
6.13
6.30
6.13
6.22
6.22
6.22
6.22
CO Data (%)
Ch 2
6.60
6.80
6.40
6.40
6.20
6.20
6.55
6.60
6.15
6.15
6.40
6.40
6.60
7.00
7.00
6.60
6.60
7.00
6.60
7.00
Ch 11
12.90
12.80
12.50
12.20
12.00
12.70
13.80
13.20
11.90
11.80
11.90
11.65
12.50
13.30
12.75
12.15
12.35
13.00
12.75
12.47
Ch 16
10.75
9.70
10.55
10.10
10.40
11.20
12.60
11.10
10.35
10.60
10.75
9.90
11.75
10.60
12.65
11.90
11.65
11.15
12.45
12.35
WALDEN RESEARCH CORPORATION
-------
TABLE 6-55
REFINERY (INLET TO CO BOILER)
July 4, 1972
Hydrocarbons (ppm)
Time
11:00 AM
11:22
11:30
12:00
Noon
12:15
13:00
13:30
13:45
14:00
14:10
14:30
15:00
15:30
15:45
16:00
16:30
17:00
17:30
18:00
18:30
19:00
19:30
19:50
20:20
Ch 0
317
332
359
363
318
310
421
391
449
397
450
468
415
315
379 '
344
296
289
312
333
301
265
293
285
Ch 3
270
286
272
264
258
198
326
348
360
348
364
404
348
320
-
-
-
-
-
-
-
-
216
197
Ch 5
406
397
369
381
317
273
241
259
186
196
197
242
183
150
130
-
230
188
253
176
-
143
Ch 15
75
75
75
71
71
62
80
84
96
92
95
99
86
82
81
72
66
65
70
73
66
60
65
63
Ch 1
6.22
6.13
6.30
6.30
6.22
6.40
6.40
6.30
-
4.70
5.05
5.05
5.06
5.08
5.06
5.07
5.06
CO Data (%)
Ch 2
7.00
7.50
7.50
7.00
7.25
7.52
7.25
8.40
9.00
7.50
7.20
7.20
6.80
6.80
6.80
5.80
6.60
Ch 11
12.00
13.30
13.30
12.45
13.80
14.25
14.55
15.30
15.20
14.85
14.65
13.30
13.90
14.30
14.65
14.25
13.60
Ch 16
11.95
12.65
12.80
12.55
13.95
14.25
13.90
13.60
13.25
13.00
12.20
12.15
11.30
11.70
11.75
11.75
11.45
6-106
WALDEN RESEARCH CORPORATION
-------
TABLE 6-55 (Cont.)
Hydrocarbons (ppm) CO Data (%).
Time Ch 0 Ch 3 Ch 5 Ch 15 Ch 1 Ch 2 Ch 11 Ch 16
20:50 330 240 74
21:20 381 296 78
21:50 433 350 88
22:20 409 334 86
22:50 372 295 74
6-107 WALDEN RESEARCH CORPORATION
-------
TABLE 6-56
REFINERY (INLET TO CO BOILER)
July 5, 1972
Time
15: 35PM
15:40
15:45
15:50
15:55
16:00
16:05
16:10
16:15
16:20
16:25
16:30
16:35
16:40
16:45
16.50
16:55
17:00
Hydrocarbons
Ch 3
94
94
95
94
98
104
104
107
-
112
107
106
106
109
111
115
118
112
(ppm)
Ch 5
104
104
105
111
103
100
100
98
101
103
101
104
105
104
107
108
113
116
CO (%)
Ch 11
12.80
12.80
12.80
12.90
13.15
12.95
13.20
13.20
13.20
13.20
13.20
13.20
13.20
13.25
13.20
13.40
13.50
13.40
Ch 16
12.70
12.60
12.70
12.70
12.35
13.10
13.10
12.60
13.35
12.75
12.40
12.95
13.20
13.20
13.40
12.95
13.35
11.90
6-108 WALDEN RESEARCH CORPORATION
-------
TABLE 6-57
REFINERY (INLET TO CO BOILER)
July 6, 1972
Time
9:01AM
9:15
9:30
9:45
10:00
10:15
10:30
10:45
11:00
11:15
11:30
11:45
12:00 Noon
CO (%)
Ch 11
10.80
11:55
12.05
12.20
11.80
11.50
11.80
12.10
12.50
12.60
12.50
12.45
12.90
Ch 16
10.70
11:20
11.60
12.20
11.90
11.60
11.45
11.60
12.65
12.70
11.90
12.15
13.35
6-109 WALDEN RESEARCH CORPORATION
-------
was later traced to the thermistor in the FID. The instrument could be oper-
ated but the hydrogen flow rate had to be increased. This led to such high
background currents that the signal due to the sample could not be determined.
Since less than one week of testing was left in the program, we decided that
it was not worthwhile to bring the instrument back on-stream. Thus, no more
HC results will be reported for HC determined with the 6800 chromatograph.
The constant levels of CO and HC at the refinery present some
rather interesting problems with statistical analysis. Correlation coef-
ficients measure how closely "b" changes when "a" is changed. If the change
for b is the same as the change for a, then a correlation coefficient close
to unity is observed. However, if a and b are the same all the time, the
statistical analysis breaks down. This is the case for the refinery where
CO and HC are essentially constant. Here the most effective statistics
are the mean concentration, the standard deviation (a) and the "t" test
between means. The statistical data for hydrocarbons is given in Table
6-58. On 29 June, Ch 0 and Ch 3 show excellent agreement with each other.
The mean HC concentrations (Ch 0 and Ch 3) are within 132. of each other.
On Ch 5, the attenuation was not set sensitive enough, e.g., HC reading
<10% full scale, therefore the accuracy of the results is quite poor and
the variability is high. Note that many of the results in Table 6-58
for Ch 5 are in agreement with Ch 0 and Ch 3. The variability of the
Beckman 400 (Ch 0) is considerably better than that of the CH. channel
(Ch 3) on the process chromatograph, e.g., 5.6% vs 12%. The attenuation
on Ch 5 was made 10 times more sensitive on 5 July. As a result the
variability in the output for Ch 3 and Ch 5 appears to decrease as shown
by the smaller values of the coefficient of variation on that day (4.5%
for Ch 5 and 8.1% for Ch 3).
In general Ch 0 and Ch 3 agree well. Tests of statistical sig-
nificance show a significant difference at the 95% confidence levels on
June 29, but not on the other day. Channels 3 and 5 are significantly
different on June 30 and July 5, The FID (Ch 0) and the THC channel on
the Beckman 6800 (Ch 5) were significantly different for all days. It is
WALDEN RESEARCH CORPORATION
-------
TABLE 6-58
STATISTICAL SUMMARY - REFINERY HC DATA
Date
6/29/72
6/30/72
7/3/72
7/4/72
7/5/72
*
Values
Ch 15.
**
C.V. =
Channel
0
3
5
15
0
3
5
15
0
15
0
3
5
15
3
5
in ppm methane for Ch 0,
coefficient of variation
*
Mean
323
286
259
75
260
265
222
72
163
53
295
279
245
73
106
89
Ch 5, and
= a/mean x
a
18
34
61
6
52
66
39
8
19
8
90
82
86
12
9
4
Ch 10 and
100
**
C.V.
5.6
1.9
23.5
8.0
20.0
24.9
17.6
11.1
11.7
15.1
30.5
29.4
35.0
16.4
8.5
4.5
ppm ethane for
6-m
WALDEN RESEARCH CORPORATION
-------
important to distinguish between statistical tests of differences between
sets of data and heuristic or practical measures. Here, compared with the
results at the foundry, for example, correspondence between the FID and
PGC is generally quite good.
The CO statistics are given in Table 6-59. Ch 1 (process chroma-
tograph with FID) gives higher CO values than any of the other analyzers on
29 June. The analyzer was set up with a very small sample loop and is
very near the maximum value which can be measured without dilution of the
sample to reduce the concentration. However, the average CO concentration
measured by the four analyzers is 13.9% and all the values are within +15%
of this number. On 30 June, the analyzer values for CO are considerably
closer with a mean concentration of 12.5% (for the four analyzers) and
deviation of +6% from the mean.
On 3 and 4 July the CO concentrations measured by Ch 1 and Ch 2
were 1/3 to 1/2 the values of Ch 11 and Ch 16. Since the conditions at
the refinery were exactly the same and these two analyzers broke down
completely two days later, we believe that Ch 11 and Ch 16 were giving
the most accurate results. On 29 June to 5 July the mean value of CO
and deviations for Ch 11 and Ch 16 are given below.
Mean of Ch 11 and Ch 16 Avg. Dev. (%)
June 29 12.4 +3
June 30 12.5 +2.5
July 3 11.4 +11
July 4 12.6 +5
July 5 12.6 +1
Note that the mean CO value for four of the five days is 12.5%. On 3 July,
the average CO value for Ch 11 was 12.7% and Ch 16 read very low. The
deviation between these two instruments was <+_5% except on one day where
Ch 16 read very low. Two CO analyzers working properly and set up for
the correct range should agree within +10%.
6-112 WALDEN RESEARCH CORPORATION
-------
TABLE 6-59
STATISTICAL SUMMARY - REFINERY CO DATA
Date
6/29/72
6/30/72
7/3/72
7/4/72
7/5/72
7/6/72
*
Values
**
C.V. =
Channel
1
2
11
16
1
2
11
16
1
2
11
16
1
2
11
16
11
16
11
16
in % CO
coefficient of variation
Mean
16.2
14.6
12.79
12.01
13.5
13.6
12.67
12.25
4.75
6.52
12.68
10.12
5.72
6.88
13.22
11.89
12.69
12.46
12.06
11.92
= a/mean x 100
a
2.2
6.5
0.31
0.42
2.4
2.0
0.54
0.54
1.06
0.60
0.48
0.41
0.62
0.63
1.06
1.14
0.67
0.71
0.56
0.70
**
C.V.
13.6
3.4
2.4
3.5
17.7
14.7
4.3
4.4
22.0
9.2
.3.8
4.1
10.8
9.2
8.0
9.6
5.3
5.7
4.6
5.9
6-113
WALDEN RESEARCH CORPORATION
-------
Again it is important to note the difference between statistical
and practical measures of comparison. In spite of the very similar appear-
ing data for Ch 11 and Ch 16, the mean values shown in Table 6-59 are sig-
nificantly different for the first four days.
A comparison of hydrocarbon values by the manual method and instru-
ments is given in Table 6-60. Tests with the selective combustor coupled
to Ch 0 (see Section 6.2), indicated that only methane was presented in the
effluent since the reading remained the same with the combustor current on
or off. The manual analyses were performed on samples collected in the
field in aluminized mylar bags and returned to the laboratory for analysis.
Ch 5 reads low (see explanation given earlier in this section) but Ch 0,
Ch 3 and manual CH. analyses are in reasonable agreement, e.g., within
jOO% of the mean value. With the manual method, only methane was found
in the sample, confirming our selective combustor and process chromato-
graphic results.
The relationships between the monitors and the laboratory analysis
are given by the following regression equations:
HC (CH 0) = 0.57 HC (Lab) + 166; r = 0.89
HC (Ch 3) = 0.35 HC (Lab) + 208; r = 0.79
HC (Ch 5) = -0.45 HC (Lab) + 353; r = 0.40
The results of grab samples and the continuous monitor readings
for the inlet to the CO boiler are given in Table 6-61. The values for
the instruments represent the average value over the period that the bag
was collected. The "grab samples" were collected in aluminized mylar bags
and returned for laboratory analyses by Orsat and the Pd(O-phen) methods.
The first two samples leaked as determined by the very low C02 and high 02
values. The CO concentrations were corrected by comparison to the average
COp results. The value was checked by calculating the oxygen level and
comparing it to the measured value. The wet chemical results for these
samples are still lower, even after correction, than those measured by
the instruments. The Pd(O-phen) method displays good agreement with the
instrumental methods (except for 6/30 #2) and verifies that the CO
6-H4 WALDEN RESEARCH CORPORATION
-------
TABLE 6-60
COMPARISON OF HYDROCARBON VALUES BY MANUAL GAS CHROMATOGRAPHIC ANALYSIS
AND CONTINUOUS HYDROCARBON MONITORS FOR PETROLEUM REFINERY
(CATALYTIC CRACKING UNIT - INLET TO CO BOILER)
ppm HC ppm HC (Manual Analysis
(Continuous Monitors) with Gow-Mac Chromatograph)
Date Time Ch.O Ch. 3 Ch.5 CH^ C2H2 C2H4
6/29/72 15:45-15:55 323 310 208 320
16:05-16:15 316 274 329 240
7/4/72 15:30-15:40 415 348 183 390
15:45-15:55 315 320 150 250
16:00-16:10 379 380
16:30-16:40 344 330
6-115 WALDEN RESEARCH CORPORATION
-------
TABLE 6-61
COMPARISON OF CO DETERMINATIONS BY WET CHEMISTRY AND CONTINUOUS
MONITORS AT A CATALYTIC CRACKING UNIT (CO BOILER INLET)
Date Sample No
6/29/72
6/30/72
7/5/72
*
Sample bag
**
Corrected
1*
2*
3
4
1
2
1
2
3
4
leaked
for air
CO (%)
. Pd (0-phen) 02 (%)
3.5
10.5**
4.4
8.8**
11.5
10.8
13.6
20. 6(?)
12.7
11.2
14.2
, note high oxygen
in leakage.
13.
10.
2.
2.
7.
8.
0.
3.
0.
1.
4
5
7
0
6
5
9
4
9
4
Orsat
co2 (%)
3.
6.
10.
11.
10.
11.
12.
10.
11.
11.
6
5
6
1
5
8
0
6
8
3
CO
3
10
4
8
7
7
7
8
8
7
8
8
CO (%)
(%) L&N MSA
.5
.5**
.2
.4**
.5
.6
.6
.5
.7
.4
.6
.5
12
12
12
12
13
15
12
12
13
13
.9
.7
.7
.7
.8
.3
.8
.8
.2
.2
11.7
11.9
11.7
11.9
14.0
13.6
12.7
12.7
13.1
13.3
value.
6-116 WALDEN RESEARCH CORPORATION
-------
concentrations measured by the continuous analyzers is correct. Statistical
analysis gives the following results:
CO (MSA) = 0.46 CO (W.C.) + 7.1; r = 0.72
CO (L&N) = 0.19 CO (W.C.) - 10.7; r = 0.61
The Orsat values are all much lower (ca. 30%) than the instrumental
methods or the Pd(O-phen) method. The reason for this is not known but the
preparation of the absorbing reagent may account for the low results by in-
efficient scrubbing of the CO.
During the last week of testing the CO boiler outlet, the Beckman
400 (THC) and the Beckman 6800 chromatograph were not functioning. The
only HC concentrations obtained, therefore, were the manual laboratory
analyses. These are given in Table 6-62. The measured HC concentrations
are low, e.g., in the 1 to 5 ppm range, at the outlet of the waste heat
(CO) boiler.
The calibration equations for the CO analyzers are given in
Table 6-63. Note that at these low CO concentrations, alj^ of the analyzers
have linear calibrations.
The data for the CO analyzers at the outlet of the CO boiler
are given in Tables 6-64 to 6-68 for 7 to 13 July, respectively. The
CO concentrations measured by the monitors vary over a factor of 20, e.g.,
Ch 7 vs Ch 11. Ch 1, Ch 2, and Ch 16 displayed considerable (>50%) drift-
ing over both short (hours) and long (days) periods. Ch 7 and Ch 11 do
show some variations in the apparent CO levels but most of these are
within about 10%. In addition to the marked difference in CO values,
Ch 1, Ch 2 and Ch 16 show unusually large drifts during measurement.
In order to determine which of the CO monitors were correct,
bag samples were collected and returned for laboratory analysis by the
Pd(O-phen) method. The initial tests run with the colorimetric method
using a 250 ml flask indicated "no CO". This means that the CO level was
less than 50 ppm which was the detection limit of the method. We reanalyzed
6-117 WALDEN RESEARCH CORPORATION
-------
TABLE 6-62
HYDROCARBON CONCENTRATIONS* AT THE OUTLET OF THE CO BOILER
Date
7/6/72
7/10/72
7/11/72
Time
19:57
20:09
21:27
13:13
15:49
9:30
12:57
*
Collected in alunrinized mylar bag
in the laboratory with a Gow-Mac
ppm HC
(as CH4)
0.6
1.0
1.1
5.3
4.0
5.7
3.8
and analyzed
Chromatograph.
6-118 WALDEN RESEARCH CORPORATION
-------
TABLE 6-63
CALIBRATION EQUATIONS FOR CO MONITORS AT
OUTLET OF CO BOILER
Channel Equation*
2 y = 220x
7 y = 2.64x
11 y = 54.5x
16 y = 50.3x + 9.03
*
y = ppm CO
x = millivolts
WALDEN RESEARCH CORPORATION
-------
TABLE 6-64
REFINERY (OUTLET OF CO BOILER)
July 7, 1972
Time
11: 50AM
12:00
12:10
12:20
12:30
12:40
12:50
13:00
13:10
13:20
13:30
Ch 1
54
74
63
80
94
107
120
133
160
153
173
Ch 2
65
59
70
61
44
79
53
35
44
63
42
CO (ppm)
Ch 7
9.9
10.3
10.9
11.0
11.4
11.1
10.9
10.0
9.7
9.6
9.6
Ch 16
65
65
56
55
53
30
30
64
37
46
49
Ch 11
195
191
200
193
195
181
189
187
182
174
177
WALDEN RESEARCH CORPORATION
-------
TABLE 6-65
REFINERY (OUTLET OF CO BOILER)
July 10, 1972
CO (ppm)
Time
12:10PM
12:20
12:30
12:40
12:50
13:00
13:14
13:30
13:44
14:00
14:14
14:30
14:44
15:00
15:16
15:30
15:45
15:50
16:00
16:30
17:00
17:30
18:00
18:30
19:00
Ch 2
66
142
173
149
100
166
85
90
79
47
37
50
59
68
46
46
42
50
74
83
50
59
4
11
20
Ch 7
9.8
11.9
13.6
13.1
13.8
15.1
17.1
4.6
14.7
14.4
13.8
14.9
14.5
15.2
13.3
13.2
9.2
10.7
11.2
12.5
14.3
13.6
13.7
16.8
23.0
Ch 11
145
164
182
183
188
204
202
199
199
210
207
203
202
212
214
105
134
180
211
236
224
236
258
268
Ch 16
109
129
114
126
133
113
106
117
120
121
121
119
118
105
119
114
116
117
112
109
119
120
116
122
120
6-121 WALDEN RESEARCH CORPORATION
-------
TABLE 6-65 (Cont.)
CO (ppm)
Time
19:30
20:00
20:30
21:00
21 : 30
22:00
22:30
23:00
23:30
0:00 Midnight
11 July
0:30
1:00
1:30
2:00
2:30
3:00
3:30
Ch 2
40
-
110
52
154
83
67
55
120
79
48
63
120
85
66
30
100
Ch 7
13.6
14.0
14.5
12.6
15.7
14.3
11.8
13.0
14.7
12.2
12.7
14.0
13.4
15.6
17.9
18.0
17.7
Ch 11
230
233
236
215
236
220
211
206
211
206
204
199
206
198
205
210
209
Ch 16
117
115
122
104
101
114
109
111
105
109
112
no
111
107
no
no
114
6-122 WALDEN RESEARCH CORPORATION
-------
TABLE 6-66
REFINERY (OUTLET OF CO BOILER)
July 11, 1972
CO (ppm)
Time
4: 00AM
4:30
5:00
5:30
6:00
6:30
7:00
7:30
8:00
9:15
9:40
9:50
10:00
10:30
11:00
11:30
12:00 Noon
12:30
13:00
13:10
13:30
14:30
15:00
15:30
16:00
Ch 2
15
57
-
-
79
101
-
59
29
131
116
61
53
22
118
44
-
-
-
-
-
-
57
98
72
Ch 7
14.6
11.8
12.5
12.6
12.9
12.4
13.8
15.3
12.8
10.5
12.4
13.5
12.7
11.4
11.3
11.6
12.4
12.2
11.6
10.8
11.1
11.6
11.6
11.6
15.3
Ch 11
202
202
208
213
212
215
230
232
228
204
224
214
215
209
217
199
207
220
211
205
207
224
215
196
192
Ch 16
116
109
101
114
118
113
125
113
116
120
112
82
102
62
72
72
112
62
72
112
62
46
51
112
82
6-123 WALDEN RESEARCH CORPORATION
-------
TABLE 6-66 (Cont.)
CO (ppm)
Time
16:30
17:00
19:00
19:30
20:00
20:30
21:00
21:30
22:00
22:30
23:00
23:30
0:00 Midnight
12 July
0:30
1:00
1:30
2:00
2:30
Ch 2
100
75
58
83
in
-
54
111
134
40
165
-
118
118
162
77
79
85
Ch 7
14.5
14.5
10.6
14.0
15.3
14.8
12.4
12.4
12.2
14.0
13.5
12.1
11.3
13.4
14.5
16.6
12.7
9.7
Ch 11
191
172
191
201
206
194
191
180
187
194
181
183
187
192
200
214
189
168
Ch 16
92
122
69
84
32
29
28
49
90
87
64
74
59
69
64
69
64
69
6-124 WALDEN RESEARCH CORPORATION
-------
TABLE 6-67
REFINERY (OUTLET OF CO BOILER)
July 12, 1972
CO (ppm)
Time
3: 00 AM
3:30
4:00
4:30
5:00
5:30
6:00
6:30
7:00
7:30
8:00
8:30
9:00
11:00
12:30PM
13:00
13:30
14:00
14:30
15:00
15:30
16:00
16:10
16:30
16:50
17:10
Ch 2
138
83
68
129
86
79
72
46
94
66
50
24
61
147
162
179
262
157
195
134
131
160
134
155
182
65
Ch 7
16.9
17.7
16.1
14.5
12.9
10.3
15.6
22.4
17.2
14.0
12.1
12.4
15.1
14.5
13.7
14.3
15.3
16.6
12.2
15.3
14.8
12.7
11.9
15.6
16.6
16.6
Ch 11
220
228
196
203
196
190
208
248
204
203
209
212
252
234
230
228
236
230
216
238
226
200
95
182
200
213
Ch 16
79
74
84
90
79
89
109
109
104
85
49
99
125
89
71
117
62
97
120
67
82
82
43
97
87
120
6-125 WALDEN RESEARCH CORPORATION
-------
TABLE 6-67 (Cont.)
CO (ppm)
Time
17:30
18:00
18:30
19:00
19:30
20:00
20:30
21:00
21:30
22:00
22:30
23:00
23:30
0:00 Midnight
13 July
0:30
Ch 2
166
85
199
186
202
173
195
220
165
156
202
170
258
114
220
Ch 7
16.4
14.1
14.8
15.8
16.6
14.8
13.2
14.8
16.4
12.2
16.1
18.5
14.0
12.4
10.3
Ch 11
206
188
204
176
186
164
169
193
198
207
192
201
199
201
191
Ch 16
107
112
107
117
87
72
97
87
87
26
62
77
82
66
66
6-126 WALDEN RESEARCH CORPORATION
-------
TABLE 6-68
REFINERY (OUTLET OF CO BOILER)
July 13, 1972
CO (ppm)
Time
1:OOAM
1:30
2:00
2:30
3:00
3:30
4:00
4:30
5:00
5:30
6:00
6:30
7:00
7:30
8:00
8:30
9:00
Ch 2
151
165
158
131
140
98
151
127
186
171
124
131
177
173
138
130
143
Ch 7
10.8
10.0
8.7
7.9
7.9
9.2
10.8
10.3
9.5
10.0
9.6
10.6
9.0
9.0
13.4
3.5
-
Ch 11
207
192
181
167
158
171
185
200
193
180
182
183
196
202
209
-
-
Ch 16
77
46
87
62
77
62
82
87
82
97
92
72
92
87
77
62
41
6-127 WALDEN RESEARCH CORPORATION
-------
six different bag samples, this time using a one liter flask. The results
indicated that the CO concentration at the outlet of the CO boiler was less
than 30 ppm. If we examined the data in Tables 6-64 through 6-68, we find
that only Ch 7, the Beckman 315 NDIR, gives measured CO concentration in
this range.
The outlet of the CO boiler represents a very unusual situation
in that a low concentration of CO (<50 ppm) is being measured in the
presence of a very high 12-16% concentration of CO,,. We felt that the
rejection of COp by the gas filters might be inadequate for this situa-
tion. Therefore, when the NDIR analyzers were returned to the laboratory
at the end of the field testing, we ran additional interference studies
with C02- This gave the following results:
CJ1 Analyzer 14% COp Expressed as ppm CO
2 Intertech 121
7 Beckman 1.9
11 L&N 169
16 MSA 70
The results were obtained by taking the average of three individual measure-
ments made at four levels of CQ2 with no CO added. The value of 14% COp
was selected since this represents the measured concentration at this site.
Note that the Beckman analyzer is the only device which utilizes an inter-
ference filter. All the other analyzers rely on gaseous filters. These
do not appear to be very useful for this situation.
If we go back to Tables 6-64 through 6-68, we see that Ch 11 with
the C02 interference subtracted from the CO value is about 30 ppm. Similarly,
Ch 2 data taken from 12 and 13 July (since the C02 interference studies
were run shortly after this) gives an average concentration of 20 and 25 ppm,
respectively, with COp interference subtracted out. Ch 16 yields 5 and 16
ppm CO, respectively, for 12 and 13 July with C02 subtracted.
6"128 WALDEN RESEARCH CORPORATION
-------
The statistics for the CO analyzers are given in Table 6-69. The
data are not particularly meaningful except for Ch 7 since much of the
response and, therefore, variation for the other analyzers could be due
to changes in the CCL concentration.
6.5 ZERO AND SPAN DRIFT
Determination of span and zero drift was made during both the
laboratory tests and also in the field. Short-term drift was measured in
terms of the standard duration of the output over a period of from one to
three hours during which time either zero or span gas was being passed
through the monitor. The standard deviation in ppm is shown in Tables
6-70 to 6-72. Unless noted in parentheses, the span gas contained 120
ppm of CO or 102 ppm hydrocarbon (ethane). The sampling interval was
generally about every five minutes. In addition, stability at several
fixed levels was determined during the pilot plant runs. Hydrocarbon
and CO levels were maintained at a constant level over a three-hour
period, with the data sampled every five minutes. Table 6-73 shows the
results of the experiments, performed on April 13, 1972. The hydro-
carbon levels were obtained by adding ethane to the duct. The CO levels
represent a combination of CO generated in the furnace and CO added in
the duct.
Another set of drift measurements was obtained over a series of
days, during which period the instrument zero and span controls were not
adjusted. The results are depicted in Figure 6-13. The L&N instrument
shows considerable variations in zero, but the span remains constant.
On the other hand, the Beckman FID shows very little variation in the
zero, but the span shows measurable day to day fluctuations.
6-129 WALDEN RESEARCH CORPORATION
-------
TABLE 6-69
STATISTICAL SUMMARY - REFINERY OUTLET CO DATA
Date
7/7/72
7/10/72
7/11/72
7/12/72
7/13/72
Channel
2
7
1
16
11
11
2
7
16
2
7
16
11
2
7
11
16
2
7
11
16
Mean
5.6
10.4
110
50
188
199
73
13.9
115
84
12.8
83
203
141
14.5
204
86
146
10
187
75
a
13
0.7
41
13
8
39
42
2.8
7
41
1.5
27
15
60
3
26
22
23
1.5
14.7
16
C.V. (%)
23.2
6.7
37.3
38.5
4.3
5.1
57.8
20.8
61.6
48.6
11.7
32.6
7.4
42.5
20.7
12.7
25.6
15.7
15.0
7.9
21.3
6-130
WALDEN RESEARCH CORPORATION
-------
co
1
LU
^"
LU
co
o
o
o
CU
-t-> a:
cu o
c
C 0
re a:
E o i
-* 0 1
O CO Lf>
cu ^o <~
CQ O
c
E 0
\x ^^
o ^a-
at
CQ
CO
cu
a_
;
re
co
_c
0
cu
-t-» a:
S- i i
CU ^"^
^_3 ^^
14
C
re
E o
-* 0
O CO
CU UD
CQ
c
re
EO
-* O
o «*
cu
CQ
CU
+J
re
OJCMCMCMCMOCMOOOOCMCM
OOOOOOOOOOOOO
r i CM i LO
CMi i i !* CO i UDCMOLO
co co co r-~ ii
1 1 1 .... II..
OVOU3OCOIO r i
co
CO IT)
r^«* co^-voi^covo^r^o^uo
C^l (^^ f^l f*^ Lp ^^ ^^ C,^ ^^" ^^ t^^ t"l
<* i C\J ^t
«* LO CO O CO CM VO
1 . 1 'Illl
O < i LO i i r
CT» CO CO CO CO LO
..I *! 'III! .1
o co o .i ..| .(iii
O O O O O O i
r*» i^. CM CM CM
c^ i~^ r^ "*^ ^"^ *^^ r^» r^* i^^- ^^ r*^ ^^ r*^
COCO^-CMOO^^^-.^,^.1^, .
^^L^L^'Lo'Lo'Io^io'yD'vo'^Io
a.
a.
o o o o
LO O O O
o o
O I
CO CO
V£>
-C
o
OJ LO 00 CO
co co co
i CO
O O CO CM
CM
CM CM
r^ r->
CM
LO ir>
U3
6-131
WALDEN RESEARCH CORPORATION
-------
TABLE 6-71
CO MONITOR - ZERO DRIFT
Date
12/30/71
12/31/71
1/3/72
5/8/72
5/22/72
5/30/72
6/5/72
6/8/72
6/9/72
Beckman
6800
1.5
1.0
-
4.8
-
4.7
-
-
1.5
Intertech Beckman
315
-
-
-
-
4.3 0.5
5.7 1.5
0.3
0.7
1.1
L&N MSA
3.6
3.1
2.6
-
30.3
60.3
4.7
6.5 52
5.6
Samples
35
26
25
48
24
12
10
9
19
6-132
WALDEN RESEARCH CORPORATION
-------
Q.
Q.
CO C
(T3 O
C3 -r-
c to
|*4 <_
co c
CU
o
C
o
o
CO
O)
o 'o.
^^ P~
(O
oo
^
to to
Q Q
O O O 0 O O O
CM CM CM CM CM CM CM
LO IO LO CO 00 CT CM
CM CM ^T
00 CO ^J" LO
1 I 1 VO VO 00 r
CM CM CM LO
VO O LO O ^" LO
CM %
CO CO \ CO ^ CM O
^ -^ co i co CM co
CM CM *^^ *^. ^^ ^^. "^v.
r i r «d- LO LO LO
oooooooo
CMOOOOOCMCM
r O O O O O r r
ft A * ft f*
O O O O LO
CO LO CM LO
OOOCOCMCOOICTi
CO 00 00 0 LO
O CM 1^ ^J" LO
i co «* co LO i i
~
P~» O CM ^" O LO CT>
CO CM CO i 00 LO *3"
f^ &\ r^ o co CT» CM
n~ CO 1^ CO " 1
O CM CM i *± i CM
LO
1 1 1 1 1 1 1 1
o o co
Ii 1 0 1 1 1 1
i U / 1 1 1 1
CM CO C\J
OJOOOJOJOJOOOJOO
***"*>. ^**^ "^"X, "^X^. "">»,, "s**v "*«s» ^^s.
LOLOLO^OVOI^OOCT^
*^**» ^"v *****. ^*x. "^^v. "^*fc "^x* ^^N.
^O CO LO ^O LO UD VO LO
6-133
WALDEN RESEARCH CORPORATION
-------
TABLE 6-73
STABILITY MEASUREMENTS
Inst. Mean Cone. Std. Dev.
Type Ch. (ppm) (ppm) S.D.
HC (FLO) 0 117.69 2.20 1.87
HC (FLO) 5 111.39 5.34 4.79
HC (NDIR) 15 128.61 3.95 3.07
CO (PGC) 1 804.23 62.54 7.78
CO (NDIR) 2 1003.20 68.31 6.81
CO (NDIR) 7 991.70 42.16 4.25
CO (NDIR) 11 914.70 58.74 6.42
CO (NDIR) 16 899.43 56.27 6.26
6-134
WALDEN RESEARCH CORPORATION
-------
I
0
>-i
-2
~ 3
LEEDS-NORTHRUP IR
\
108 ppm CO
\
\
ZERO GAS
2 34 5 6 78 910 1112
DAY
t
5 (-
0
INTERTECH URAS H IR
Z E RO GAS
2 3456789IOIII2
DAY
100
60-
20-
0 -
102 ppm C2Hfe
BECKMAN 400 FID
ZERO GAS
0 "
I 2 3 4 5 fe 7 8 9 10 I I 12
Figure 6-13. Comparison of Instrument Zero and Span as Function of Time
6-135
WALDEN RESEARCH CORPORATION
-------
7. CONCLUSIONS
The Pd(o-phen) colon metric method for carbon monoxide has been modi-
fied and can be used for CO concentrations from 50 ppm to 10%. This method
was used to provide a check on the instruments at all the test sites.
Accuracy tests for CO and hydrocarbons were run in a well controlled
gas fired pilot plant by doping the flue gas with 100 to 1000 ppm CO and
100 to 1000 ppm C2Hg. The NDIR analyzers for CO give an average accuracy
of 95% while the accuracy of the chromatograph is slightly lower, about
92%. This is quite encouraging in view of the difference in the tech-
niques. Doping with ethane yielded an accuracy of 94% for the ethane
sensitized NDIR and 86% for the flame ionization detector. The MSA NDIR
(indirect sensitization) for CO responded quantitatively to ethane. The
accuracy of negative filtering NDIR's (such as the L&N) drop off at CO
concentrations below 200 ppm. This is in agreement with the manufacturer's
recommendations that the instrument be used at CO concentrations greater
than 200 ppm. The range of a 0-50 ppm CO analyzer (Beckman 315) was ex-
tended to 1000 ppm (with no loss in accuracy) by using a cubic (non-linear)
calibration curve.
The CO concentrations at the incinerator ranged from 200-500 ppm when
dry refuse was burned, while the concentration could rise as high as 1000
ppm for wet or difficult to burn refuse. The NDIR's and the chromatographic
methods provided comparable results. The only exception was the MSA instru-
ment which was statistically different than the other instruments.
The hydrocarbon concentrations ranged from 2 to 20 ppm normally but
values as high as 100 ppm were recorded. The ethane sensitized NDIR was
not useful for this source due to the interference from water vapor and the
variable composition of hydrocarbons. The flame ionization analyzer and
the THC signal from the chromatograph were comparable for most of the data.
These two instruments appear to correspond well even in view of the longer
response time of the chromatograph. The CH. concentrations measured by
the chromatograph used from 10-40% of the total hydrocarbon concentrations.
7-1 WALDEN RESEARCH CORPORATION
-------
The CO concentrations in the foundry ranged from several thousand to
10-15%. Above 1% CO, the instruments which were set up for the high con-
centrations, e.g., L&N and MSA, gave the best results. The chromatograph
was set up for this range by using an extremely small sample loop and by
tuning the amplifier down. At the high concentrations above 1%, however,
even with these modifications, problems with saturation were encountered.
The instrument with the long (1 meter) cell (Beckman 315) did not produce
useable results at CO concentrations above 1%. The CO concentrations
measured by the instruments (MSA & L&N) were in good agreement with Orsat
and wet chemical CO determinations.
The hydrocarbon concentrations at the foundry varied quite rapidly
from 120 to several thousand ppm. The concentration could vary as much
as two decades over a thirty-minute period. The chromatograph data
appeared to indicate that the instrument had an extremely long response
time (45 minutes), presumably due to adsorption-desorption on the inlet
tubing. The HC values by chromatography would approach those of the flame
ionization monitor only when the concentration remained constant for a
long period of time. Only the flame ionization detector gave reasonable
hydrocarbon concentrations according to comparison with manual (laboratory)
HC analysis and correlation with the CO results. The chromatograph was
not useful for this non-steady state source. The NDIR hydrocarbon values
did not show good agreement with the flame ionization analyzer. The
selective combustor-coupled to the flame ionization analyzer provided some
promising preliminary non-methane hydrocarbon results when compared to the
manual method. The methane content of the foundry varies considerably
and was, in some cases, as high as 50% of the total hydrocarbon concentration.
At the refinery, CO concentrations at the inlet of the CO boiler were
10-12%. Here again, the instruments set up for the high range gave the
best results and were in agreement with the wet chemical values. The
chromatograph could be set up for this range but some difficulty was noted
with saturation.
7_2 WALDEN RESEARCH CORPORATION
-------
At the outlet of the CO boiler, all the NDIR CO instruments gave dif-
ferent readings which varied from 10 to 200 ppm. The wet chemical (Pd-o-phen)
method could not detect any CO in the effluent. This indicated that the con-
centration was less than about 50 ppm. Only one instrument which had an
interference filter to remove COp interference gave a value that low.
Confirmatory laboratory tests indicated that COp (at 12-16%) was a problem
with all the devices with gas filters, but not with the instrument which
had an interference-type filter. Thus, an NDIR device with such a filter
for C02 is necessary for monitoring at the outlet of a CO boiler or at
only very efficient combustion sources.
Hydrocarbon concentrations at the inlet of the CO boiler ranged from
150-300 ppm and was essentially all methane and was reflected in the
measured data. Both the total hydrocarbon and methane channels on the
chromatograph were in good agreement with the flame ionization analyzer.
No continuous HC data were obtained at the outlet of the CO boiler, but
manual samples varied from 0.5 to 5 ppm.
7-3 WALDEN RESEARCH CORPORATION
-------
8. RECOMMENDATIONS
Although the overall results of the study, as indicated above, show that
the existing state-of-the-art in hydrocarbon and carbon monoxide monitoring
is generally sufficient to satisfy stationary source requirements, there
are areas where further study and development are recommended as follows:
(a) Improvements have been made in several of the instrument types ex-
amined since the program was underway, particularly with regard to the proc-
ess gas chromatographic monitors. Additional study of this class of instru-
ments appears valuable, since this is the only monitor type capable of
determining non-methane hydrocarbons directly.
(b) The selective combustion technique should be examined further to
determine more fully its potential application in obtaining non-methane
hydrocarbon measurements in conjunction with the flame ionization detector
and non-dispersive infrared detector analyzers.
(c) Additional study should be made of the characteristics of sample
acquisitions sytems; particularly examining the potential for losses of
higher molecular weight components.
(d) Additional study of the dynamic response of gas chromatographic
analyzers should be made to obtain more complete data, and to determine
how to utilize these devices on sources with variable hydrocarbon levels,
such as foundries.
(e) Further study of the measurements by NDIR of low levels of carbon
monoxide in the presence of large amounts of carbon dioxide should be
made, since the study showed that, under these circumstances, the results
depend on the type of filtering concept used in the analyzer.
(f) Although the field and laboratory results provided adequate data
to characterize the applicability and accuracy of the several types of in-
struments, additional data would be very useful and further work is
recommended.
g_1 WALDEN RESEARCH CORPORATION
-------
9. REFERENCES
1. Allen, T. H. and W. S. Root, J. Biol. Chem. 216, 309 (1955).
2. Altshuller, A. P., et al.. Anal. Chem. Acta. 25_, 101 (1961).
3. Foundry Air Pollution Control Manual, American Foundrymens Society,
Des Plaines, 111. (1967).
4. Anon., Hydrocarbon Processing 56, No. 11, 119 (1971).
5. Ayers, C. W., Anal. Chem. Acta 15, 77 (1956).
6. Bellar, T. A. and J. E. Sigsby, ES&T 1. 242 (1967).
7. Burianec, Z. and J. Berianova, Czech. Chem. Comm. 28_ (11), 2895 (1963).
8. Burianec, Z. and J. Berianova, Czech. Chem. Comm. 28^ (11), 2895 (1963).
9. Cuffe, S. T., "Catalyst Regeneration," in Air Pollution Engineering
Manual. J. A. Danielson, Ed., AP642, U.S.H.E.W. (1967).
10. Dalton, C. and C. Kowalski, Clin. Chem. TJ., 744 (1967).
11. Dietz, W. A., J. Gas Chromat., 68 (February 1967).
12. Driscoll, J. N. and A. W. Berger, "Improved Manual Methods for Sampling
and Analysis of Flue Gas," Vol. 3 - CO, Contract No. CPA 22-69-95,
PB209 269, APTD 1109.
13. Duncombe, W. G., Biochem. J. 83_, 6 (1962).
14. Engels, G. and E. Weber, Cupola Emission Control Technology, Trans!. by
P. S. Cowans, Iron Founders Society, Cleveland, Ohio (1967).
15. Fegraus, C. E., et al., "Instrumentation for New Jersey's Vehicle Emis-
sion Testing," Instr. Soc. Am. Trans., 7(4):317-326 (1968).
16. Hinkly, D. and P. L. Kelley, Science. Vol. 171, p. 635 (1971).
17. Iwayama, Y., J. Pharm. Soc. 79^ Japan, 552 (1959).
18. Lorch, E. and K. F. Gey, Anal. Biochem. Ij5, 244 (1966).
19. LAAPD, "Emissions to Atmosphere from Petroleum Refinery in Los Angeles
County," Report No. 9, APCD, Los Angeles, Calif. (1958).
20. McNevon, W. M. and P. F. Clone, Polarographical Separation of Cool Flame
Oxidation Products, Reinhold, N.Y. (1955).
21. NAPCA, "Air Quality Criteria for Hydrocarbons," NAPCA, Pub. AP64 (March
1970).
g_] WALDEN RESEARCH CORPORATION
-------
22. NAPCA, "Air Quality Criteria for CO," NAPCA Pub. AP62 (March 1970).
23. Niessen, W. R., "System Study of Air Pollution Sources from Municipal
Incinerators," Final Report, Contract No. CPA 22-69-23 (March 1970).
24. Palanos, R., Proceedings from Joint Conference on Sensing of Environ-
mental Pollutants, Palo Alto, Calif. (November 1971).
25. Robbins, R. C., K. M. Borg, E. Robinson, APCA J., Vol. 18, p. 106 (1968).
26. Rose, A. H., et al., JAPCA 8 (4), 297 (1959).
27. Stevens, R. K., et al.. ES&T. Vol. 3, p. 652 (1969).
28. Taylor, H. A., Foundry Engineering, J. Wiley & Son, N.Y. (1959).
29. Van Sanat, W. A., et al.. AIHA Journal 16. 222 (1955).
30. Younghans, R., etal., C.A. 36142 V (1967).
g_2 WALDEN RESEARCH CORPORATION
-------
APPENDIX I
TRANSLATION OF CARBON MONOXIDE PAPER
A New Method for Carbon Monoxide Determination
Z. Burianec and J. Burianova
ABSTRACT
A new method to determine CO has been worked out which uses its re-
action with bis-(o-phenanthroline)-panadium chloride to yield a colored
_2
product. The method can determine, on the average, from .5 x 10 to
5 x 10~ % CO (by volume) spectrophotometrically or from 1.5 x 10 to
5 x 10 % CO (by volume) photometrically.
In the past, the determination of small amounts of CO has been car-
ried out in many ways; most have centered about the use of the reduction
reactions of this gas. One of the most widely used methods takes ad-
vantage of the reaction of CO with iodate. A number of the reactions
use the reduction of the salts of several metals, e.g., palladium (II),
silver, and mercury (II) oxides. The analyses usually are run in the
presence of another reducing gas such as dry carburetted hydrogen, or
hydrogen sulfide; the methods are only worked out for known concentra-
tions of other gases and, therefore, interfering gases must frequently
be separated out before the analysis can begin. From the viewpoint of
selectivity, there is only one exception to the last statement, the
method based on the reaction of CO with hemoglobin. In this method, the
hemoglobin reacts with CO to form a carbonyl compound which is identified
by a characteristic absorption band in the visible and near infrared re-
gion.
Methods for the determination of carbon monoxide are ordinarily not
selective and exact analysis is necessarily either a costly undertaking
Collection Czech. Chem. Commun., 28(11), 2895-2901 (German) (1963).
I_] WALDEN RESEARCH CORPORATION
-------
or involves complicated apparatus and a boring analytic procedure. Simple
and quick methods are generally not sensitive and exact.
The new method is based upon the reaction of CO with an aqueous so-
lution of bis-(o-phenanthroline)-palladium (II) chloride to form a red-
violet palladium carbonyl complex, probably Pd2(H20)4(CO)2(phen)2. Since
the method revolves around the characteristic color of the reaction prod-
uct, a carbonyl, this method is selective and, under certain conditions,
independent of the presence of other reducing gases.
MIXTURES OF CO WITH AIR
In this method, CO was produced by the decomposition of formic acid
(aided by the addition of concentrated H^SO.) and purified by sucking
through U-tubes containing potassium hydroxide, calcium chloride and
phosphorus pentoxide. The makeup of the gas was determined volumetric-
ally using a Hempel buret with copper (I) chloride absorbing solution.
As shown in Figure 1, low concentrations of the analyzed gas were driven
into flask S with an air flow, which was controlled with a manometer.
An exact amount of CO was let in this evacuated 5-liter flask from con-
tainer Z. The flask was brought to atmospheric pressure with air. The
pressure was read before and after CO was let into the flask and also
after flask S was filled with air to atmospheric pressure. The height
of the Hg column was measured with the aid of a cathetometer to 0.1 mm
Hg. The gas mixture was allowed to mix for 10-12 hours in order to be-
come uniform. The control for this experiment with low concentrations
of CO was the subsequent analysis of the mixture by IR at the Institute
for Hygiene and Job Safety, Prague.
REACTION SOLUTION
The following chemicals are needed to produce the reaction mixture
for CO: ^ 10% Pd (II) chloride solution of ^ pH2, analytical grade 29%
hydrazine hydrate solution and analytical grade o-phenanthroline.
1-2
WALDEN RESEARCH CORPORATION
-------
PREPARATION
Dissolve 0.3g o-phenanthroline in 850 ml of hot O; to this solu-
tion add 7.2 ml 1.08% Pd (II) chloride solution dropwise so that the im-
mediate precipitate dissolves. The concentration of the PdCl2 solution
was determined several times with the use of diacetyldioxime. The yellow
solution was cooled to room temperature, transferred to a 1-liter volu-
metric flask and diluted to volume with H^O. Its pH is ^ 3.
APPARATUS
The reaction flask R is a 250 ml round bottom flask that has been
connected to a tube with two stopcocks. With the help of one of the
stopcocks, the flask is evacuated with an oil pump to a pressure of
1 mm Hg and the gas flask (S) is filled. With the help of the second
stopcock, the absorbing solution is poured through a funnel into the
flask, which then is filled to atmospheric pressure with the gas to be
analyzed. The content of the flask with the reaction mixture and gas
was shaken during the length of the reaction by means of a Chirana
shaker. The solutions absorbances measured once with a Unicam 500 at
500 nm and another time with a Pulfrich photometer with filter S53. In
both cases, 1-cm glass cuvettes were used.
OPTIMAL CONDITIONS FOR CO REACTION
Since the rate of the reaction of CO with aqueous Pd(phen)?Cl? com-
plex solution is dependent upon the presence of the palladium and o-
phenanthroline as well as the concentration of these components, the
presence of reducing and complex forming substances, and, in addition,
the nature and length of contact of the gas with the absorbing solution,
the optimum conditions for the reaction were investigated.
The optimum palladium concentration is determined by the upper so-
lution limits of Pd(phen)Cl2, which precipitates as solid phase in the
formation of the bis-(o-phenanthroline)-palladium (II) chloride. It
has been our experience that PdCU concentrations from 4-6.0 x 10 M,
in which the molar ratio of o-phenanthroline to palladium is 4:1, is
1-3
WALDEN RESEARCH CORPORATION
-------
the best. A higher ratio of o-phen to Pd (8:1 to 20:1) causes a very
slight increase in the solubility of Pd(o-phen)C12 and does slow the
rate of reaction of the species with CO.
The presence of divalent Pd supports the reduction process and has-
tens the rate of reaction which is very small with low concentrations of
Pd (II). The reaction takes several hours. The hydrazine hydrate is
considered the most important substance. If the amount of hydrazine hy-
drate is exactly right for the reaction to occur, it has been our ex-
perience to see a 10- to 50-fold increase in the reaction rate. The op-
timum concentration of hydrazine hydrate in the reaction solution is
6-9.0 x 10"2M.
Due to the effect of hydrazine on the excess Pd, the solution must
be stabilized, preferably with ammonia. By examining the time dependence
of the solution absorbance with differing amounts of ammonia, the opti-
mum concentration of 0.2-0.3M was arrived at. With this concentration,
the absorbance of the final solution was constant for 120 minutes. The
stability and the absorption spectrum of the colored palladium carbonyl
are not influenced by hydrazine insofar as it is not present in a large
excess (over the optimum concentration) and the CO concentration is such
that all the Pd reacts to form the color complex. Since there were no
changes in the absorption spectrum of the carbonyl, it is suggested that
the hydrazine is not bound to the complex molecule. The reaction does
not appear to go completely when there is an interference with the color
complex formation either through previous reduction of Pd to Pd° or by
the introduction of a substance which reacts with CO faster than the o-
phenanthroline complex. When the CO concentration is so low that only a
part of the Pd in the solution reacts with it, a reaction occurs between
the hydrazine and the unreacted bis-(o-phenanthroline)-PdClp. The solu-
tion stops being stable after the reaction with CO and begins to turn
brown; after a few hours it polymerizes. In this way, the measurement
of the solution absorption normally protected by the stability of the
CO reaction with the Pd(o-phenanthroline) complex, is robbed of its
reproducibility and exactness. Reaction conditions need to be worked
out to circumvent this difficulty.
I_4 WALDEN RESEARCH CORPORATION
-------
Also checked were the reaction of ammonia, pyridine, ethylenediamine
and several other bases on the solution of the complex carbonyl, the unre-
acted bis-(o-phenanthroline) -PdCU and the hydrazine. As previous work
has shown, there is no substantial effect on the solution due to either
ammonia or any of the other basic substances. With PdCl?, ammonia gives
the complex [Pd(NH,).]Cl9 which, in comparison with the analogous hydra-
J ^T £
zine complex [Pd(N2H.).]Cl2 is very stable. Through its formation, the
destructuve influence of the hydrazine can be avoided and the bases can
assist in stabilizing a solution which has low concentrations of CO.
When ammonia or pyridine is added to the bis-(o-phenanthroline)PdCl?
solution, the solution does not turn brown or polymerize even after a
long time. A solution which has, after 10 minutes contact with CO, be-
gun to turn orange-red can be returned to red-violet (absorbance maximum
at 550 nm) by the addition of ammonia. This is probably due to the for-
mation of another complex from the decomposition process.
If the gas and the absorbing solution are shaken vigorously for 10
minutes, the reaction, even for low levels of CO, is quantitatively com-
plete (120-150 shakes/minute).
PROCEDURE
Ten ml of reaction solution and 0.1 ml 6.3M hydrazine hydrate solu-
tion are pipetted into the reaction vessel (R); the solution enters the
evacuated flask, whereupon a mixture of CO and air is allowed to fill
the flask. The flask is taken from the apparatus, attached to the shak-
ing apparatus, and after ten minutes of shaking, the remainder of the
gas is removed by evacuation; the flask is filled with air and 0.2 ml
concentration. NH.OH is added to the red absorbing solution. After ten
minutes, the absorbance of the solution can be determined spectrophoto-
metrically or photometrically.
ACCURACY OF THE METHOD
With the help of the previously described procedure, the following
curves were constructed. Three tests were run in which freshly prepared
1-5 WALDEN RESEARCH CORPORATION
-------
gas mixtures were used each time. The gas was analyzed each time by the
same procedure; they all contained the same concentration of unsaturated
carburetted hydrogen and hydrogen. From the data in Table I, it can be
-2 -4
seen that CO concentrations of 2.1 x 10 to 5 x 10 % volume can be de-
_2
termined spectrophotometrically with 5-10% relative error and 1.5 x 10
to 5 x 10"4 with 6.6-10% relative error.
INFLUENCE OF OTHER GASES
The influence of hydrogen, H^S, ethylene propylene, butane and acety-
lene were studied. It was ascertained that the CO determination was not
affected by significant concentrations of ethylene, propane or butane.
Hydrogen in concentration over 15% volume, FLS and acetylene in concen-
trations over 1% volume, affected the measurement since they caused
either colloidal palladium or sulfide to precipitate out which, in turn,
raised the absorbance of the solution.
CONCLUSION
-4
With use of this method, CO concentration above 5 x 10 % by volume
can be determined; this is at least as good as most previous methods and
better than many others. The accuracy (10% relative) is certainly better
than the rest of the analytic procedures. The advantage of this method
is its rapidity (20-25 minutes for successful analysis) and the small
amount of gas needed to run a test (250 ml). This method would be par-
ticularly well suited for the analysis of gases from industrial sources
since the formation of the carbonyl is selective and, therefore, is
recommended as a general analytic method for CO determination.
1-6 WALDEN RESEARCH CORPORATIOt
-------
APPENDIX II
COMPUTER PROGRAM FOR MONITOR CALCULATIONS
AND STATISTICAL ANALYSIS
II-l
WALDEN RESEARCH CORPORATION
-------
t 1 (
1
M
S i !
K :
V. i '
* ' 1
0
,
1
1
!
i
1
t
i
IM »
! ** »
: *< x z
h- O >- X
o o ; o i- *
ft. M <
»---»*»; x
3 MX
O O * ' ^H (VI **
UJ o ; x x p) x
^ j ** ,j . * ^*
1- U, I- _J
» n t_ i_i uj h-
n o CD »- ' ' ox
O. ~ -^ ' X
1- _» -H r* O tO I
> * * * >*4 X, *-<
0> -~ Z n 'XV PI*
UJ o >-t euj n IP
a »\j >- x »-xox
< «i o « o ^*
>- Z O :> O IV *
»- O X < V CVJ -*
00 2 » C> < O I
a ~- ~i c _j »
-' UI'CP)
»-i\j«» >-_jxx
^ J ^- C*J*»
Ujt-^** X*^vtv
o.»
I-CUJO H- M >- »
* (V > O Xh-*<«^
i xacc ov«ruiv
CL ^>*^»^- * »(/
«r z^~» vac.oc<
(M OOZ(M ^ i-i o >-i »
(v o ~ » *: o _j cc
_l Z iCUJ *(/)»«<
U. i-ts»C X I «J
> < »»O Pllfl'-'XO
uj 9200 vvoir«
_J Cr->» *(!.
a
c, o
* t 1
» »
c c
*
» r-<
V
u.
<
<
1-
O 0
*. *>o o,
o oo o
o o * >
>-< >-> o o
* * V.
o o it) o o
O O O !
0 > 0
o o o
o o o
o o ~* o
- 0 O f-l
o o o
0 0 ~ 0
» * c
C C C C 0 0
O *-H
* C C fi
c c c c » o
«-i i C
^ c c J °
c o c o o e
o o
o fu -» *« o o x
c » » o ^* uj
* o o o > »»
.00^ *-< o »
»Olfl »-
c c .c c c c < «c
p- .- e .^i-~
» »^H » (/) Q: >
oooo»oi-aro
»c «VU.O
C(VCO C'O.^-U.
^locccoi: a r-
c- c r z c i i
*o.--7c_;» M ii
e>zZ"tvrr)
Clf-C « C Z Q
e ozu.u.aaa
t-r-ist-aaa
c x u.1 a1 u1
' ' i
', i '
t
i !
;'',.;
i : i ' !
' ' i ' i i * '
' '
1 t
i ; i i :*-'
1' 1 ,
1 i ,
( J
1 - 1
i
i i
i : '
1
ill 1 1 i I .
! i ; :
i ! ! i
1 '
i i
;
i i
i
i
i
i
! ;
1
1 ;
| (
' '
' 1 1 | '
1 »
1 ! ;
i i
t .
1 1 ;
1 i ! ' : ' ; '
i
1
i j '
i
I . ,
-
c
*t
' ** ' «
a: ... w
~ a i- , - Ifi
a r»- fv «-« <«
~~on<-« -< o »
u.cj>-*»_j rv< x
ox .- a ^-_j a
C UJ ^1 IIIIH HCC
l_)lh--» ~~> ^ i M c o e n en
n H » tn rvj ~3 coo ~o T
aao._j(/>tsc unite »- < z. z r c « »- ~ ~- z a
u.'iaju,i<(/!u.|i/iocz>yzcc«>a>-'>-'«-'a.a.MC «_>_j>-c
O" C c c
O1 a c c
4. « fV
;,
»-«
H-
«t
0
»
a
c
«
0
UJ
X
o
c
1
z
"
a
v. !
' C '
1 c
C:
C
> 0 '
! 1
p- pi Pi
0 0 O
000
O 0 C
coo
1000
, 1
1 «"
'o
0
c
o
e
in '
1 C
0
o
o
1°
1
-------
1
! ,
1 1 i
i
1 1 '
i
1
< i
i
i i '
(
rv ! i !
I^f i
t- '
> 1 i
r- .
N
X 1
* 1 '
0 ! i
i * '
i i
i
i i
i _ i
i ;
i t
1 !
! ; i
i ~
i ^ - "i
1 1 0
i 300
Q. rvj »
t- ~ .-.
3 ID 0 CVl
o «^ *-*
« ~ ~
* *-» -4
o o »
-»>«* m
UJ o
&. CVJ (M (/)
< _J _)
1 IL. >
4 «t H- »-
n o IT >-
UJ -< »
a. »-«>»
1 < -> t-lCVI
, >- _J -<-«
~
1 O> ~ 2 n
! u) 0 M
1 a cvi »- x
j; «i o
1 »-Z 03
K O
i eo r ~
Ul i
. a - --
i < c -~
t- IV! -!
« e
1 t- _J » ~-
U; C * **
Q. >C UJ
I »- O,..
»- C LU O
(V) >- Q
t- ~ » »
1 3 O
i a. -j o o
a »- cv, -
a r »- »
CVj O C Z M
(V O
_l Z *. U.'
U »- N » C.
> < ~ C
Ul J Z O O
_» c ~ »
x _ .. ~
QL *i v o o
IT e F rvi -
O C U.
i c : u cv
i a - i- x.
Q C * (A **
r- - cv r
CV1
2:
C '
If .
a i '
LL f
> c
c
Z c
2 ' e
a o
1 i
!
i '
1
! ,
i
i
!
!
; i
I
|
i
i
i
1
i
i
i
i
i \
I !
1 ,
i
| i
1. .
l ,
!z :
1 K- X
, t- ; "<
< «
; X
tn x
» N-«
o tn »
lone
s o nx
x x~
1 CO 1*1
» » CJ^
V . -« CVIX
i rvi^-t
; cvi
x x a
( ro to x
j * ~*
1 -J »
i ijj ^
! o i
o ^*
J r
X X«C
o ift X
^H X *-<
xv n
c ui PI in
~> »- I O I
r<. O
O CVI
X < V CVI ^
«J < O X
V »- >
C _J
u) c n
1- _l X X
«t < < i
c v> »
I * V CV
x iDtr x
O V. UJ < ->
*- NJ »-
X. >-<-'
x a < Q x
f 3 X <~<
O X <* IT. V
* (/
v o, c oc «;
^ >-< O l-> »
* * o _i a
^ (/! » « «3
X I - U
n in *~ x c
>. V C IT «3
-* a »
-» X. J -«
C (/> £ O < V
>- 1- K h- K 1
< «t < I «i X
C" C C IP C "*
'
coo c
COO c-
o c. c c
C O C 0
o o o o
*
»
o o
M * «
» *
O G
^4 *
» »>
o o
i-H *
»
0 C
-I . 1
» »
0 0
*-« ^H
» >
O 0
1-^ *H
» *
o o
*
*» ^^
* »
o o
^* "*^
»
o o
^* ^^
«. »
c c
»
* » r-*
V
U.
t
4
<
K
4
c*
f^t
C
O
O
0
o
»
0 0
«
00
«< ^H
» «
0 0
rvio
0
»
o c
o c
r+ ^^
0 0
CVI 0
0 r-c
» >
O O
0 0
1 1 *-l
» >
0 0
c c
«-t ^~
» »
c c
c c
»- ^
c o
orvi
^- c
«
.0 0
c c
f
» »
c c-
c CV
*-> o
i\ r
1
i
1
1
1
1
!
i
'
,
[
1
'
j
I
,
!
|
» .
o o
o o
o o
^* f«4
» e.
O O
V
Ifl O O
o o o
o -« o
* > «-4 »H
o o
o o
O O 4
c o o o
^H r^ f^ f*
O O O O
Ul C -* O
o o «
O rt O
» o
o o o
» ^H O
O O w->
o o o
_-l . O
« « o
C C 0 0
O r-.
O O -H «
0 O O
. ^- c
» » o
o c -
o o c c
o o
rt i-t O O
> » o *-^
0 0 0 »
. c
O Ift »
iC C C, C-
- c
»»»«»
o c o
c
c c o
c c c o
^- ^ o c
» » 0
o o «
* c
C IT C
c c
^- ^~
j | i
' ' 1 1 1
1 ' '
, *
t ' ' I
1
1 : : i : !
1 i ! j
1 ' t ' , 1 1
1 ' ' 1
1 . 1
i . !
i 1 1 i 1 ' 1
1 i i 1 1 ' i 1
i 1 1 1 1 1 1 1 1
( ' ' 1
1 1 1 i , 1
! '
I 1 ; !
' i i ! i I
1 lilt
1 i -
! i ' i
t 1 '
1 1 1
1 1 1 '
! , ! f
1 1
: J ; i
: ! ; i ' ! i : i
' 1 ' 1 t
1 1 1 1
1
, 1 t 1
'ill
i i ! ! , 1
i ' ! ' 1 i
i , , I , ' i
1 1 , '
; , ! I '
i ' !
' ' . !
i . l
' p i ' I '
x
ui a
>-r o: K-
> Ul <
»- O f« K
« >( C. U
t- UJ a «-
i/i a > a r^ cvi
»-aco o m
XU.U U. U K _!
a»-u. us ^.^-cr
z a ^-.^c^uj* M ii n
Z'Z ClllJIt--* i- ~)
"5 C -i H H H II II II tfi CVI ~3
z &i z ** cv* f> «^ m *c i/^ z o* O1 **^
c z o ~ - ecc<^~
su.u.aaocott_j>-M »^-^icc
ii»-Qocaoa_jv xje (JHII
O>-XU.iUlU'U-'UJUJ«IOC352
a
a
^
(-
prj «r AT fT| ^ if1 ^ fh. n fv^ ir tf ffc c ^ CV f^ CV'
ocoococccococccccc
ooooooccooccoccoco
CCCCCOCTOCOCCOOCOCO
oooooooococooooooo
1 I 1 1
1 ! 1 1 ! i i
i - 1 ! t i 1 i
1
! i ; : : ! i
1 1 1
1 !
1 I i 1 ' 1 1 1 i
i . 1 i ' ! 1 |
i l :
1 . j 1
! 1 ' ! '
' ' .
i , i I '
' i 1 : i | '
1 1 i '
i I 1 ' i : i
1
i i j ! ! j
i ; , i i j
: ' i ' 1 1 1
' , j
1 : ' ' i ' 1 '
! : ! 1
! , | , i ! i
i i : i ; ' ,
' i i ' < i 1 i
t '
i ! i I i i
1 ; l ' i
1 ; ' i
ii
i
i i i i
i
II i i 1 ;
; '
! j I
\ < '
'1
! ! ' !
1 iii
1 i !
' ' i
; i *i ; | i
i \ ' . i
' ' . i
i < i
i i . i
i
i ' ' ' ; l
i ' i 1 l \ '
i ' f ' i :-
i
i i ' ' !
' i '
i
' ' ' i c
" r
: ' < »i
t~
CVl 0
-iiccc H cm c
COO >-iO «^--i-i yc
C K«=i.ZZC II »- «- ~ Z a «
cc< a»->-«-a.OLMO u-iv-ou.1
?CrCvJ>J«J«l_JC^C~NJ^C}ZU.Qr
c c c e
a c c if .
^ ^ IV ^
w i ^- ix ^
1 *-
1
1* ^ if C. ^* CV C^ <4 if ^ V^- C ^* CV ^ f^ p*1
f> f^i I* »««««4'««U'U IfU If IT
ccce.cc ccccccccccc-
c ccooococoooooeroc.
COCOC/OC CC/CC.CCOOCC
OOCOOOCOC.OCOOOCOO
II-3
-------
' I
i > !
1
«M
^
|w
(V)
X
*
o
z
-1
z
i
i
a
CT1
(VJ
liJ
Ul
1
a
i
i
(\J
c
i-
VI
a
u
>
z
^
ae
-M
o
(VJ
o
i-
a c
o> ts
9* *-
» «--4
a- "
o> «
o
m UJ
C U,
Ul X
U U
ct
0
- »
r- r-
c -
c v
00
1
i
'
M
(VJ C
C? ^
(VJ 1
0 0
1- 1-
c c
t2 v
*» M-k
O (\J
C("l
v_'
C C1
uj uj
*
%$
u u.
<-. c
r- o
' t
C. 0
0 0
o
in
o
« !
o
»-
c
<_5
«.
^
^
C
Ul
£
it
u a.
(V W
0 C
c o
C 0
0 0
1
c
IT
r-> ,-.
II
O S
O *
C
in
Ul
IT £
c c
0 0
0 O
0 0
,
1
, '
1
: i
1
1 '
' ^H
2
«.
^
*
' TT
o
3>
o
t~*
r^
-«
*
J
JC
w
X
O i
3
o
o -»
lO *v Q.
.-* ^ 2
~ UJ
C 2 *-
^- » » i
O H- *^ - j^i O «f
K » H « w
& ~* 2. < *- «
* t/i -.itr iO uj H"
!!< I 0. C X
U 2. L.. tt
a. c n u ' u.' o
*-* c o >- o u.
^ a
l-Ti ft*
U ' Ob
.1J
. »- (V (V D
O C C C C O
o o o o c. o
o o o a o 0
^
..«
-»
Ml
n
K4
» ,
«_»
»-4
»
Z
5C
t/>
_l
1-1 .-«
1- 0
^ ^>
» «ff
(X -4
O *
-. X
in
^ i*>
~ <
»
4Q Q
U.' O
Or u
,f\
U'
J
(», ir
C 0
O 0
O 0
«.*
^
(*1
w
I
1
-I
*
z
1-1
o
Ul
(£
1-
as
c
vCO
a
>
^*
UJUI
1- _l
i- 1-
OL ***
^ ^.
««
m
c'
o
Ul*
z -*
u o
l-U.
a
fi*
v^
ir >c
C 0
C 0
0 0
I
1
o
in
o .-<
?£
a o
O 0
1
«- (VJ
1 i
ij
i x
o
: CK
I 0
UJ
o
0
Q£
o
i
o
a. ui
ace
O -~ Ul
(V U
rv c o
0 .C X
-H »
-t> X
^ ^-* 4
Ul Z
Ul 1-
1- i- Ui
-* OL C.
axe
7 *» O
-~ .-
C jQ.
.-. o *
»- c x
i O Ul
Z u u C
ir i- - u.
.- (V
5 S
i
r ir ir uri
«t <.: f c
coco
coca
o o o o
i
.M
IT
« 2
* 1-
O
,»
VI (VI
ui a.
c c
c
UJ
2 I
O
Z C1
C (V
I .
= a
1^-
IT
C
c-
o
o
.
'
at
o ;
(VJ
c
f
^
»o
«^
Ul
>-
IM
a
_^
f*n Q,
rn e
C M
*-.
UJ Z"
2 iC
»
I Uw
ts
1
IV
^ CVi
C 0
0 C-
o p
(VJ
a.
^
^
»
(V IX
u. n
si
cv sr
r r.
0 0
C 0
o p
.
*»
*^
%-*
Z
^ri
r^ ft
* *t
* C
^* M
N
I--I I-H
* Z
- Ul
c
c o
*
*
ir K
r. c,
(V (V
c c
e o
0 0
,
i
i
i
o
r- o
^ i*
*Q
O <-
1-
C
C 1-
^ z
o ^
O C II
c
U" o C
" UJ Ul -
t- o J-: --
O U- U- O
o
* -C C .-
4 ir ir
^ ^ ^ fv.'
C C 0 0
c o o o
o o o a
CVJ
..-* 1
^
,,-H
c
I" ,
0
o
*.* i
^H
^* i
» 1
»-"i
IV
UJ
u
C Ul
Ul O
z.
C 1-
U 0
*
»_.
«.>
bJ
C 0
«^ ^c
"3 ->
^
« O
«*
«:
* r~
fV. (V1
O C3
c o
0 0
, '
t
1
[ (
i
'
;
i
.
o
, Z '
o
i 1 ** i
.*_«
H-t
«
,^
l_
^_.
(VJ
1
_K
»_4
«^
r..
-* J
> ~
1 . ^
^-> » 1 I O
Z ii !--«
i < U
1- Ul C
« c
C C (N. X
ii o a ~
-^ .- £ C 5
c c
O O O 0 O 0
i
!
i
1
,
i
,
"-
1
1
'
j
;
1
J
I
1
1
1
1 (
1 '
' 1
o
in
in
«p-
c
0 1-
in
-- 0
o
o *
D- _l
I
C
e e
UJ
_i un
a !!.-_)-
> i ~* «_>
C 1 Z
lA » i .k--.- H
u> >^ i^ n
c ~ «v c ^
v H ** *....
1
j
4 »*« I/"
c o c o o
coc.cc
0 O 0 O O
-------
,
1
, (
!
J
I I
1 '
1 )
1
'
£ 1
X, '
*.
I*1 , j
X. ,
* ' !
o
1
1
i
1
1
, ;
i
i
i
;
z
^
^
X
|
,
i t
1 i
i i
<
in
« '
! O .
»--'
n *
0 C 0
(VI O «M
t o _t o
t i- ec *~
a
o- c c c
(X, O U) O
_i « .. * a
iu r _) ^ u.
> < n _i i- M
UJ . _ .
"^ 3 £ -! J _O £
a: zcipMt-iiz*
ir »-air*<«-i-a
i z - u r ~
1 CU.C_JU.(VCU-
r
c
->->- o
»-
C C C K
t£ O t2 Q- C
(V
c o c *
^ . lij t« . ^ *^
« -- 1- a
a Q a u. < c
- «- - - a r-
u. u. u. a ca
~ l-i »- T U. C
ir
«v
j i
t
,
i
1
i
1
i
1
'
'
1 |
!
« !
i
! !
I 1
! 1
,
1
i
I
i
( ]
1
i i
** (
ft
c
2
c *- o *-. c -<
n ^« t/i
C UJ C _' C «-
< C i
C C C 0 C (/>
tr < c a c «-
e c
r- c e
-» u -*
t
1 i
I
(
J
| |
;
'
t
t-
I
1
i i
i i
!
'
i
i
t
!
1
'
'
1
4~>
^+
^- ^^
Z ~
I-^iT "
a. - n
O *~* '^
n w ir
y i\ r. sr c
(A * C II 11^ i
i ' 1 - .
' 1 '
J | 1 I 1 .
I [ ! ' i
>
; i 'iti,! . '<
! ' i ! ' l ' '
. 1 1 !
! i
1 : i i
t i ,1 i '
i , i i
i ; t i
i (
1 . ' ': ' ' '
i 1 . 1 t , 1
* t t , | '
1 ^ (
' 1 ' : i
i :
! i <
|
1 , , , ( '
1 i ' '
I . t | t
1 ' '
i ) 'i
i ; . i ;
; ! j , ' '
i l ' \
t '
1 i '.
. : ! ; . | ;
; i i ! > ( i i
i :
i ' . i ' 1
1 r '
i II;
' ; x
1 . : *
I i X
: ; M
1 ' , 1 , i
1 : , «C
: ' : ' u.
1 : ! °
; ! ; ~
. *
i
, i
! l ' ; CD
1 i ; i ! -
: J i 2
. ,o
i ! . : .*
i i i i
X
1 . ' 1 *-<
; , ' X
! i _» , , . i\)
; _J . ' ,
. _» : o
', " i ' , i u-
* ~ , 0
0 _l ' U- «M '
: iw d . t £,'
o z
t- I- Z
z ' »- «n
o - z c
cm ' tt
' O UJ
0 t- t-t _J M
fVJ J <\J *
(M ^- f«- _l
, >- _l O O OX
>-4 »s_jo erM "
c _i - M
~(\.W _ CC >.
-_;* _)C C l- f- ZOlf
«-'*l/) »z o ZO<
fV ^» *~^ 3C ^ ^* CT1 C3 ^3 ^ * *
_| »> » IT C ^-fVjlTO C (V -> *~ »<
H C «C ft» »- »- » i-i .-
^^^4«>-MO C*H**-(O O «-^U.-'_iii>-i_ju~ 'd*
c/ijt X~«-
X ITU-CC _IU._)U.OQ.CU.OCC_IU.C)_IU-UCa CiTf-CtC
»--)«-CC _J»-_I«-OC tf»-C_lC_J»-J_J>--l»->CCZCC ?-U.
r-r* (C^cv^- r-r- ^»-
-C 0*CVfV, CC «*
<»v rv
-------
(VI
f-
X
X
^
o
1
[
z
<
X
1
1
cc
o-
(V
_l .
UJ
>
UJ .
_l I
a
*/-,
v
a
1
f
f\j
0 '
- 1
o !
u. i
> ;
«!
a i
1
i
^»
*<
c
U,
O
*
Iff
i. 1
_l
^*
«
*<
*+
X
*o
^4
o
f\J
«
*
V
z
*-
^.
<
*
»
X
**
x
^
*-«
o
(X '
*
o
IP
v> ~«
Z^
*_»
«tf l-
o o
9 (3
^
i
pi
if.
if
c;
o
o
'
i
o
o
H
rvj
ii ^
rs. »-i
* J
,4-
c
fv
< IT-
IT IT-
If IT
C O
O Cl
O O
, 1
I
1
i
; i
1 1
\
1 ,'
i
i
i ill
i I * '
i
i
i
i
j
i
I
1
t i
!
00 t
-*
i
1 1 ' r
i ; . ; '
i
1
t
i
i
1 i . "
1 t
fc
.
4
i
1
i
i
| 1 'OP
+
1
i j
I
1
t
I
1
i
i ,
i
i
i
i
i
!
I 1
1
i 1
,
1
;
[
; (
1 !
|
: ' i
,
'
-1
i
_i
f
w i 1 , ! »"
1 4
' K
. 00
.
t
: i ' 2
1 ' ; ^-*
r-l
; '
1 ;
***
i I < ^ > «6 I -»
"5 ' f> O ' (\J 1
' «0 « »- 0
(u (u < -* * o a. z
I -I , -I »O.^ |-I<
tf C IS ^ UJ*
a a . or K * CHC
o ;
C ' ' C '<£
: ' O )C V* H- »- - X O
o»-* o>^ >ot/i"^o o^4cc^^^
(Vtin* , , cvtin* . , <^«^I^AJ<^O . (vi«fv>»«
C »»C « « c jt 19 uJ (\J h-
*^ »-H UJ ! » *MUJ C***-)!/)*"^*^^^ *^Z **(V)fv
H»«(V '*' K Oa«lf>
i^i'O^^UJh-OOUJ^O 'OO»'*'"^**UjaCC3 *UJ«O 'O^HVI^UJ^-iUJ"^ UJUJ »-4 »-t* U.Z(VJ**"UJ
TOtV-»«D(Vi OZ ~53(ViMO3(Vj «(Vl»» Z ~>O_J3O «-> lua.(V.I3
1
;
! "-
|
'
'
J
,
1 ' 1
I
j
i
f
1 '
i ' t
t
1 ' )
I
f
i
)
&
^
(\J
a:
rv o
*\j
f*- »
^ ~
\jj *^ »-* t
_J O ~* '
(VIN-»»Z O-»Z '^O(Vllx- Z «»IIZ (*!» J£(VI»-*-»UJOtZZQ. lfl~'-t-Z ZXZ 1C
(V<--J>-D--.~»-K(Vi^-C>(V(-»- >-l->-OU.t-CI>-^-C-*i:^^-l-l- H U(V_lO-Ci KK»-
COU-OC*-tJOCrNJ»-^OOOU.GOO-JOCNJ»-*OUJOU.OU.O Hu.*-COQ.U. OU.UJl*-'C'*'U.CO
00*^ o ; -j o c o>-«oir ^ o o c1 v: _jdoru.»-*o^4Dif i** < cJc?c»-"0^'H-c.it cc^*c*o
1 '
f""-ct(vjin ^ccAjo*A.c ^f*^»crif c^r^tf^if c~(v(vt^<£of^h~^*^h*
ur ^o ^ f** r*- ^> ^ ^ ^ o c c< c c> o " *~ ^^ cv fv iv INJ cv (V ^i i*"' 1^1 p" ^t & ^ \f^ IT tT1 if ^ ^ *c ^ ^ ^~ ^- ^ c ^* "
if If1 IT If' IT If' if tfllT *fi *i£ *t' < ^^^«f^/^4>4 ^) < "i rf if 4" ^! ^ ^ <- ^ ^£. *£ ^f ^C sf ^ if «f << ^^.^»rw^>
OOOC'-CC.OOOOOOOOOOOOCOCOCOOCOO&OCOOC'OCi&COOOCC'CC.O
oacooooooooocoooc-Qooooocjooeooocooo&occcoccoooo
O 0
00
0000
O 0
O O 0 O
i
e> o
O O 0 O O O
O 0 O 0 O O
1 1 -6 ',
cooooooooc.
1 , 1
cooooooo
H * *r r^ u ' .
1- II * (VJ '
o u. u. c ,
< «J i-i C '
IT h-
^ ^
(V (V
' ! 1
i
r t f '
r- (V (\ (V (V
r* r- r- r- r- ,
o o o o o j
C 0 C 0 0 1
O O C O O '
i
-------
I
I
M
X
h-
N
^
!»
y
**
X
1
1
a
a
(V
_j
u.
>
u.
_)
a
ir
a
i
i
i'
(V
2
c
V
a
u
_
3
a
1
' i
1
'
1
, I
i ' 1
' ' I 1
f
I
1
i
.. |
1 I
1 1
1
I
i
! i ,
! '
i i , i
1
1
!
1
i '
i
!
i
i
i
j .
1
t
1
,
i ;
i
1
i
' : ' ! :
1 I ,
I ,
i i
i
i
i
t
O
in
1
ivi -< !
O
1
i
1
(
t
1
1
I
I
i
i
i
i
i
i
i -j
1 OD
1 ; ' i ' '
i
! i
1 i
i ' ' ' ' ;
' !
' '
i
! ' ' ' i
i ! !
1 i ; i
1 1
i ' ' ' !
,t; , iii
i ! , ' ; ' ' ' i
' ' ' 1 '
' I , '
! ' '
, i 1
, ) ' ' i . i ! i '
1 1
: ! ' ! :
1 ( < , 1 ,
« ' ; ' ''
, i , j
ii i t
i i '
i i 'i ,
: ' i ' , 1
' ' ' . i '
I , ' 1 I
» - ' ' ; i i i
; ; i ;
.
' i !
o ,
I
f»
i i : i
1 ! 1
X . .
o
4
«
X
i ; 1
i f J
: ! i i I
i ill i
S i
v 1 : i '<
^
* 1 , 1
! x"
i -
! o
i . *
^H
*^ ,
H- ~ 1 i
M 0
o *
e> -J) ~
. VJ j
i
«
i
:
: ~i
(
^
i ~
i
i *
M
~- K
e z - z c ~; ~
UJ *~ C? UJ ^ , t_ ^»
w >-i -a. t- ivj 'x »
» c H- o » o o <
»-i m u. «r M (\j o «s o a
~ ^,t- + a. t- ~ ~ _j
a. *i ~. - z. _j x
c » »-i u.' o cro iu
a »- ~ a *- c ^ c *« »-
c ~ar.~ » ~ a c -> * *>
C"-1 -j»-ccia «-> ~jor^-rcx rv
~i-<2 irrv«>-i^ OTUJ o*"* c*-|u-
< >- C» OK~'»-
V C» »t- C ». Z » ~
IX 2. (Vi UJ e. Q.< «+OZO<-x
m^» st ^- ~>(T1*^ c*^^ o^--^ c? Z5C or -*
xin » »»-»(V'x »-»(v»»-t-. » M »
x«^- ^ ^- -j«-r-oc »- rv » if »u.>- -! *-* F- iv
i*+ II N »«- _| %O UJ ~-icfvi «- rv x z c- n o c^u.*-1 ca*
~.IT« 2 ~ U «-" «tUJ»«^»-1lf^--~i; UJ*->- '-''-SU.
**^^aoin ^CHI_-*^^' II ^ 5C *" _J !!* fc'c *~UOUJ>- Z ~ - M ^H -C.H» ttUJ>-»-
Qx<2 >- i~it;ct-j3i:»-az z. M ~ H n - 3 C. - a
oma«-'C iiccjuuiu-'c iit-icxc'-ccu.vjocu.r2:~ujz~-c
u. _j_)C5ircc«-t-cu-_ict'3u._icc>-krar»-iZTv>-u.«u-
f
c c-c >r ac if iv
If" l*»(^" '^ (^
N Vc nif«r f- ivriiff^ iv< K%C
c cc-coccc.c.oocc.cccococceCiCcccc.oc
e oooooooooocccocococoococoocao
II-7
-------
1
i
(VI
' X.
^
ru
X
#
o
2
<
Z
1
1
a
^
(VI
J
U.'
UJ
J
cr
i
i
fr
(VI
2.
tfl
a
u.
>
2
a
!
I
(VI
c
in
o
t-
G
IP
«*
0
1T
^!
UJ «
^ fx
2 Z
C U.
»M
r-
i
c ft!
*- f^
(v rvi
o o
o o
i
o
M
f-
: j i
1
, .
,
1 ' i
; i
1
,
1
' j
i
t
1
\ \
! ' <
\ 1
1 ' '
;
;
'
! !
o -
X X
oo
< 1
a J
OB f. ad
i o .4
_JX £
«t OU
0 1
*
X
*
^
o
4
a
*
X
*
«
o
H
_J 4
ff
*
^
(9
a. u
i i
<
UJ O
1-
< »
C X
4 IT
cr
o x
*C X
X
o ~+
~ H-
u <
-a
1 C
f
c
IT
IT ^
«- «VI
(V M
c o
od
1
i
\
\
\
I
i ;
i
4" \fl ' '
0 rH
I X!
00 ! ' ,
,
d ^ '
0 -H
r x
J
i
i
i
i i
(
i i '
j
1 | |
I
1
'
0.-
X X
»-o c
hi
O O ~-
ax x
uio c
N
,
X
t
1
i |
1
i
f
0 J
o «
X 33
|
IT -<«># «
<-« IM ' o IM > a ' cc
C^- UJ 1 U, I » *
XX -~ >- 1 UJ JVJ^-
oo ' ' *4 in z « >- -o o
' ."-« oZ*>-"~ * 0. UJ
1 '~ If) ~* t\i -It-
O<-< ' ' < i 1-1 Z Q. O <
0 -« . «-i O X h- UJ C
T x < a cu. seu, »-
c <_ , c »-»- r- - K c. < »
a «!* (5CX
*' o 4Q****«'O *^. < in
» o «^ o '^ ^» o ^^ >^ *-^ or o ** ^
X« , - CV)«K- HH ~ (VI IT. OX
m ^^ M o B tj ** c *^ c ^* ** >* o o. i » i
< ~ i5ozin^'»5t«-'Ujxinx ~«tu.'«o<-'
xo-«o_) _l 2 a Z Z UJ»-2Z »-
(V.I«I H BU,BCCCO>-»-COCO-.BC B«-«U.<
X X2CH-2 2 II " B B«-ICK.Z.~'>-a
^" (V, (*1 '
evi 0
O. o '
* O (V ' ' :
» O . '
X -1 f-
O ' *" * » «
ifl »-« ^* X O
« »» o o z P> in , i
(VXaa:icc»-2
^ (v r- i
(V O1
^ ^ ^
- < o- ,
i
(V *
» <
J.
' "^ "^ ^ t/* If) If' tfl ^
' ^) n f^ f> fO ^*j r*i r^
OOOCfOOOO
i 1 II-Ol
ooooocoo
a o a
o o e o
o
-------
ii1'.;
1 1 < (1
i i
i
|
!
i ;
1 !
t
j
| ,
i
; i
,
, 1
1
i '
i i ' '
j
r-
' X
»- i
IVI
1 X
*
1 O
1
1
1
1
'
i
I
I
1
I
|
,
f ;
(
1
1
I 1
i
1
1
t .
1
j i
1
i
1
t
j
! i *
,
»- ;
^* -
X »-
j UJ (/)
1
1 »- O
o > < t- '
9 O t-
IVI C> uj s
uj r - o
_J - -) UI
>- X '
a D 2 x z
IT c c u.' a-
a. a 5 - -
1 Of i - t- C
1 D C U. U.' 7
r «/) c> ~ a u.'
IV' i
e
<-> 1 . . , ;
i
a
u
^
IV IV, 0 r- |
<-. 0 r-
i c c c o
Z o o c o
3 , . e o o o .
or i o o o o ,
i .
i 1 i i
'
1
i
i
i
i
(
'
t
I
,
,
I
i
!
T
i
,
f
,
t
J
'
I
t
,
i
}
1
|
'
i
i
i
J
1
t
i
!
1
1
1 !
i
i .
i
'
i : ;
t , '
1
i
i
', i
i
; ,
l ! I '
1 ' i
1
1 1
' i
1
1 ,
1 1
1
1
' ' 1 (
\
t i
;
1
1 '
1
I
i
:
' i '
'
t !
% i ' i
: i
I
t
'
r
( ,
' 1 1
1 i
I
I
II-9
i
i
ii'i'il
-------
j
1
VI
i 1-
X
^^
X
^^
1
1
a
a
IM
UJ
U.'
a.
V.
a
i
i
rv.
z
c
V.
a
Z
D
a
^.
a
c
i
«
u
o
»-
»-
3
a
c
^H
H
UJ
a.
<
t-
o
i^
$
*<.
I/) (VI
M
~s
(VI ~
(VI
M
X (VI
tft
*. «^
(V U.
"- u
a £
z c -
C (VJ p-
^ ** UJ Z
z rvi n c
i-i » Z O
c ~ p-
1
I
1 1
1
1
1
(
|
I
f
;
!
j
,
' i
1 i
/
1
i j
\
O :
Z
I -
o
*«t '
* ' i
i
i
4|
x a
o C
x r- c e o
c c «? o c o
o o c oo o
'
IT
O
n
o
r-
(Vt
« O
a. c
**
-< *
(C
a «*i
^ o
Ul
u.
C 0
u c
u u.
t- t-4
(V,
o
»-
o . «
tr .
» a a
00
XtF tf f- *"
If « V C *.?
n in iri
O O II O* O1
ui r>«- -~ M M
M (V' * JC
o c
c »- e c c c.
u. c r z ui UJ
*- tr « < a o
IT
o
^
fnffi mn *
c c c c c o
o c o c c o
c oo o o o
1
f 1
1
(
\ '
1
( (
1
1
i
1
(
I >
(
:
]
1
i !
1
t
\
1
,
,
1 1
1
'
j t
. n
: f 0
n
0
1-
o
-^ *
***« - ffi
M M ^^ O
d ^- UJ
PH lyi ^ p^ ^
f»- -« (VJ *-»
c c K e m «
UJ W
c x o o o u.
c < e -i c p-
^ ^-
c c
n n
ir in « < < r+
*t IL U
~ Tz "i
Z 2
< OC U.
C ~J «J r-
IV f
c o
r- d
o c o
o o c o
o o o o
i
i
i
' ' '
1 '
! i i
i i (
i t
j
i
.
1
i
i i
,
! j
r
i
i
i
, i
i
* t ; ; ,
i i
! : t
i
i i ! , (
i i
i ! , i
t
.
* , i
;
,
i i
i
' ' ;
1
1 i
I
1 f
i
1 i
i %
! '
f I
J 1 1 i 1
1 i i i
1 , (
! ' : I '
t . '
i i < i ' j
IT
O
in
o
~ t-
O (VI
« p-t o o
. ^ p It
a (v
(VI (VI c <
» O O C C O *CXp-l
Hoc NOCOOCO ir »o- m
p-«» -5 H n n a n in ^ »
OZ.Z XOXXXXUU.OUJOU.U.
cirirzc3r«ri/iO (VI
«» c c o
«vjir v P-
c c c o c C'« p- (v,(vjrv,(V;m<<*
ocoooooccoooccco
coeooococcoccooc
ooooooooooooocoo
,
,
;
1
i ,
!
1
J
1
j
| i
j
1
i
t
,
1
i
| 1
1
'
j i
[ t
.
t
'
(
i
a.
z
UJ
*^
f*
a
UJ
K
X
0 M (VI
"7 CC
» (V IT 0
N » OC (VI
p- *- C U 0.
^" *^* 7 N
a < w «
a H ui >- p-
»- a c 2 - a
z t; a p-
c ui ui o " o
C P- C U (T O
p- <>
a o>
(V C
I**!*"
IT in ITi (VI
Ul Z 7
H U
~ p- rv
u. r r
p- IT ir
r- f- f-
coo
coo
000
11-10
-------
CVI
r-
x
(VI
X
I i
o
o
ct
*
CV,
*
*
-4 t-l «-t »-l O
I
I
oc
CVI
cr
1/1
a.
i
CM
z
o
i-i
If!
a
u
O <-> CC CD CO
oo
* *
C
e
CVI
*
CVI * -i
in i
I
Z Z Z
Z Z Z
I I
x f\j OJ
IT * *
* * *
in
x
-j cr >-
X
1 C Z
H UJ II
-> .
>- ~< CVI CVI
XX X U. U.
V) T i ->n
om »» «»
~ > *-« »« »-»
*» ^* ~*
<-»«->l\J(VI
ru
X
(/> o
X X X
if
x
-
O U. X
O i- Z
>- (V CC -CV,
X X X H U. U.
vi 1/1 v. o c o
zz
OOO
OU.XXOXXCX X
o o cr -» «11 ->
O - U,'
>- X X X
I M Z
Z
CVI
o m
inm in
in
> « ~> z
» z
- UJ - Z
^. CM -i
U. Z U. O
c z c o
z »
^- ~ CV. i
C U- C O
O * O Q
i
I
CM or
C i
o in
X
O O O "
e> c »c
O -. C (VI G C
B ~) O
». o
-> a ~> cvi > ~>
>-
l/> »-
i/> a
C
a in
o ~
X
o >
xx
e t/i IT
> CM
(/>
(/> u.
OJ x
CD V)
i i/i
r i
IM
-<
i
X
*
v\
a.
x
ui
IM
X
o -»
(VI »
^ x -<
CO
-J M tx
«l
CVI
I
X
O
X
Z
«* X
- > c
x cr u.'
T 10 O I
a o
.-IX OC UJ G
o m
or *- o (v PI
N o »*»uj
~> D -o
K z
x
o
in o
x
« -
z
< CVI
o «
^« »
(VI O
in ^
i > x
« w
i t
(Vi CV, CV.
u. u. u. c u.
O »H C O C O>
1 f" CM
G O
G G
:«- fvcv f^.^ c\.«v
OCJCC.CC.C-CCOCCCC. CCOCOCCCCOOOOCCOCOC.OOCCOCOCOCOOCOOOC.OOOOO
GGOOCCCCOOCOGOGOOGOeoOOOGOOeOGOOGOOOOQGOOGOCiOCCCiOOOGCOO
0000000000000000000000000000000000000000000000000000000
I
I
I
I
1
n-n
I
I
-------
{Jf
X
«M
V
0
z
<
z
1
1
«c
&
(VI
_l
>
UI
*J
tt
«/>
t
fu
o
*-
a
b
*
Z
D
tt
»* C
»-«->
^ U.
-111.
UIUI
a c
a u
0
u z
zx
c c
r* «-c
luu
^1
1- »-
tfVi IV
f*l 1>
in IT
0 0
o e
e o
to
*-
z.
UI
I X
o c
u. u
1- »-
« If
IT «fl
C O
C O
e o
"Z
o
\-
v\
u
Y-
**
> »-
X X
I/I C
a ~
i ^
u t-
K C
IT IT
C C
0 C
0 0
y
z
UI
a
UJ
u.
u.
»«
o v
X X
c- c
1
CV f
IT IT
c e-
e o
o o
!
f
f
1
,
!
|
i
I
1
i
i
i
1
i
1
!
z t~ in
o a >-i
»-t UJ
«/> u >
in a o
UI U. (t
a »-
IS Z
«. UI « 1
ru a:
U X X X
C C C 0
IT «e c "
II"- ir^ IT* in
C C' C C
c c o a
o o o o
t
a
<
>
a
u
o
M.
X o
C (E
1*1 «
IT UP
C C1
c- c
o a
i
j
i
i i
!
'
1
(
i
(
!
1
i
1
i
!
i
i
;
i
i
,
i
i i
.
UI
a.
o
^ j
VI
^H.
T r-
0 1C
ui m
*" ^ ^
a
u. c
»- -J C C
*
ac
c c.
-------
(VI
^
(V)
X
o
1
1
1
a
a
UJ
>
UJ
_J
or
v>
a
i
i
»
IV
z
o
VI
or
u
>
2
^
a
P
3
O
«.
lL
t-
r-i
a
Ul
z
5
c
a
ff
3
VI
<\J
1-4
t
CV
H- UJ
2 -J
C t-
**
z \-
o
I/I Ul
z c
u, u.
3 f-
i 2?
0 »-
f i m
c c
0 0
0 0
0 c:
0 0
**
o
T
O *
-> *-
*
» *>
"^ >-«
C -
UJ Z
I <
« O
~'»
(VI UJ
"* _
z -
< H
c
c
u.
» X.
* t
UJ
_l C
H- 0
** *
V-
z
c u.
X 1-
y -
o a
m r
C 0
o o
o o
c o
c a
0*
0
«M
"*
**
X
O
V
o
*
4;
^
X
0
(M
V
O
*-*
a
(VI
1-1
X
o
X
x
r+
4
(VJ
X
o
in a
X ~3
X
f N
» n
»- 0
« c
Z (VI
a
o c
u. c
c
c
f^
(V (V
(V (V
c e
c a
c c.
C 0
,
j
1
!
! .'
i
1
i
1
1
1
1
I
+* '
O
"^
» 1
i
^ !
«.
"^
»
»
t-
o
*
^.
~) 0
*~lf
z u.
« (VI
o ^^
^« »
^< 0
0 -«
(VJ«
»x
~»- z
U' < a
K 5 O
-a >- o
a o u' z
» u a uj
c- **
o c
IV (V
1
If . - ^- ^
rv ir IT un
c c o o
o o o o
c c c o
c c o o
! t
' ,
, 1 '
1 [
' J
1 '
1
1
: i
\
t
1
! 1
\
I
: i
1
'
i
1
t
1
i
; i
t
1
1
1 1
(
[
' ' 1
1
1 1
1 1 \
I 1 I . '
1 '
t
! i
t
1 1
I
1 *
1
1 1 i ,
1 1 1
I
1
1 ,
1 1
' 1 ' i
', i ' I
1 ' 1 1
1 ! i ! I
. . 1
i ' 1
! ! ' 1
1
1
,
i
1 ' 1
1
1 1
1
1
1
1
1
i
1
1
J t
1 1
'
1
i I !
: J ' (
!
1 i ;
:
i
!
1 '
II -13
-------
ENVIRONMENTAL
PROTECTION
AGENCY
DALLAS, TEXAS
LIBRARY
EPA-R2-72-106
Evaluation of M'
-------
ENVIRONMENTAL PROTECTION AGENCY
TECHNICAL PUBLICATIONS BRANCH
OFFICE OF ADMINISTRATION
RESEARCH TRIANGLE PARK. N.C. 27711
OFFICIAL BUSINESS
PENALTY FOR PRIVATE USE. $300
^
V
V
POSTAGE AND FEES PAID
ENVIRONMENTAL PR6TECTION AGENCY
E PA - 33*5
U.S.MAIL
\
------- |