United States
Environmental Protection
Agency
Air and Radiation
(EN-341W)
EPA340/1-91-008
September 1991
vvEPA
Manual for Coordination of VOC
Emissions testing Using EPA
Methods 18> 21, 25, and 25A
V v.
o
kTMC
i'EPA
• 340
' i
91
008
c.2
U.S. EPA Headquarters Library
Mail code 3604 y^O^f
1200 Pennsylvania Avenue NW
Washington DC 20460
7±> Printed on Recycled Paper
-------
• *, *"
nsg
Y^t
i:.^>n
n't "
•rr
-------
EPA 340/1-91-008
MANUAL FOR COORDINATION OF VOC
EMISSIONS TESTING
USING EPA METHODS 18, 21, 25, AND 25A
U.S. Environmental Protection Agency
Office of Air and Radiation
Stationary Source Compliance Division
Washington, D.C. 20460
September 1991
SESI™«»
WASHINGTON, D.C.204&0
-------
4
-------
TABLE OF CONTENTS
Chapter page
'* -
1. INTRODUCTION ;.. M
:4 2. ESTABLISHING TEST OBJECTIVES 2-1
2.1 Review of Applicable Regulations and Background Information 2-1
2.2 Determination of Compliance Limits 2-2
2.3 Determination of Data Necessary to Show Compliance 2-3
2.4 Basis and Comparison of Results for EPA
Methods 18, 21, 25, and 25A : 2-3
3. PREPARATION FOR AND OBSERVATION OF COMPLIANCE TEST .. 3-1
3.1 Notification of Compliance Test .. 3-2
3.2 Conduct Pre-test Survey 3-3
3.3 Finalizing Compliance Test Protocol 3-4
3.4 Procedures Common to Most Types of VOC Testing 3-5
3.4.1 Measurement Errors 3-5
3.4.2 Determination of Flue Gas Flow Rate 3-6
3.4.3 Moisture Determination 3-7
3.4.4 Organic Compound Identification and Quantification
by Gas Chromatography 3-8
3.5 On-Site Observation Procedures 3-9
3.5.1 First Sampling Run '. 3-10
3.5.2 Second Sampling Run 3-10
3.5.3 Third Sampling Run 3-10
3.5.4 Sample Recovery and Transport 3-11
3.5.5 Analysis 3-11
3.5.6 Observation Report 3-12
4. MEASUREMENT OF GASEOUS ORGANIC COMPOUND EMISSIONS
BY GAS CHROMATOGRAPHY - METHOD 18 4-1
4.1 Applicability 4-1
4.2 Method Description 4-1
4.3 Precision, Accuracy, and Limit of Detection of the Method 4-2
4.4 Measurement and Location of Sampling Points 4-2
-------
TABLE OF CONTENTS (Continued)
Chapter
4.5
Page
Observation Procedures for Method 18 Testing 4-3
4.5.1 Selection of Proper Sampling and Analytical Technique 4-3
4.5.2 Preliminary Measurements and Setup 4-3
4.5.3 Observation of On-site Testing 4-4
4.6 Observation Procedures for Method 18 Analysis 4-10
4.6.1 Preparation of Calibration Standards 4-10
4.6.2 Sample Analysis 4-11
4.7 Use of Audit Materials and Interpretation of Data 4-12
4.7.1 Performance Audits 4-12
4.7.2 Systems Audit 4-15
5. DETERMINATION OF VOLATILE ORGANIC COMPOUND LEAKS
- METHOD 21 5-1
5.1 Applicability 5-1
5.2 Method Description 5-1
5.2.1 Regulations and Leak Definition 5-1
5.2.2 Portable Instrument Operating Principles ,. 5-2
53 Calibration Precision 5-6
5.3.1 Calibration of VOC Analyzers 5-6
5.3.2 Laboratory Calibrations 5-7
5.33 Field Span Check Procedure 5-10
5.3.4 Thermocouple 5-11
5.4 Location of Sampling Points 5-11
5.5 Observation Procedures and Checklists for VOC Testing 5-13
5.5.1 Performance Criteria and Evaluation Procedures for
Portable VOC Detectors 5-13
11
-------
TABLE OF CONTENTS (Continued)
Chapter Page
5.5.2 Laboratory and Shop Support Facilities 5-17
5.53 Routine Field-Oriented Evaluations of Instrument
Conditions and Performance 5-17
5.6 Typical Sampling Problems and Solutions 5-19
6. TOTAL GASEOUS NON-METHANE ORGANICS AS CARBON -
METHOD 25 6-1
6.1 Applicability 6-1
6.2 Method Description 6-2
6.2.1 Sampling Procedures 6-2
6.2.2 Sampling Equipment 6-2
6.2.3 Analytical Procedures 6-3
6.2.4 Analytical Equipment 6-4
63 Precision and Accuracy 6-4
6.4 Location of Sampling Points 6-4
6.5 Observation Procedures for Method 25 Testing 6-5
6.5.1 Equipment Specifications 6-5
6.5.2 Pre-test Leak Checks X 6-5
6.5.3 Pre-test Sampling Train Purge 6-7
6.5.4 Sampling Procedures 6-7
6.5.5 Post Sampling Procedures 6-8
6.6 Sampling Problems, Errors, Solutions, and Action Required 6-9
6.6.1 High Gas Sample Moisture Content and Freezing of Trap .... 6-9
6.6.2 Use of Electrical Service Not Permitted for Probe and Filter .. 6-10
6.6.3 Probe Exit or Filter Temperatures Not Within Specification .. 6-10
6.6.4 Non-constant Sample Flow Rate 6-10
6.6.5 Use of Method 25 for Measuring Low Levels of Organics 6-11
6.6.6 Sampling and Analysis by Different Companies 6-12
6.6.7 Measurement in Ducts Containing Organic Droplets 6-12
111
-------
TABLE OF CONTENTS (Continued)
Chapter Page
6.7 Analysis 6-12
6.7.1 Analytical System Performance Checks 6-12
6.72 Calculations 6-16
6.8 Audit Procedures 6-17
6.8.1 Performance Audits 6-17
6.8.2 Systems Audit 6-21
7. TOTAL GASEOUS ORGANIC CONCENTRATION USING A FLAME
IONIZATION ANALYZER - METHOD 25A 7-1
7.1 Applicability 7-1
7.2 Method Description ; 7-1
7.3 Precision and Accuracy 7-3
7.4 Sampling Point Location 7-4
7.5 Observation Procedures for Method 25A Sampling 7-4
7.5.1 Leak Check 7-4
7.5.2 Calibration 7-5
7.5.3 Response Time Test 7-6
7.5.4 Sampling Procedures 7-7
7.5.5 Establishing Response Factors 7-8
7.6 Sampling Problems and Solutions 7-10
7.6.1 Cold Spots in Sampling System 7-10
7.6.2 Sampling System Leaks 7-10
7.6.3 High Moisture 7-10
7.6.4 Adjustments to Gain or Zero Offset 7-11
7.6.5 High THC Concentrations 7-11
7.7 Audits 7-12
IV
-------
TABLE OF CONTENTS (Continued)
tei Page.
8. REVIEW PROCEDURES FOR VOC TEST REPORTS 8-1
8.1 VOC Compliance Testing Report Format 8-1
8.2 Report Review 8-1
8.2.1 Cover, Certification, and Introduction 8-2
82.2 Emission Results and Performance Audit Results 8-2
8.2.3 Facility Operations 84
8.2.4 Sampling and Analytical Procedures 8-5
$.2.5 Compliance Report Appendices 8-6
Appendix • Page
A. ORGANIC COMPOUND IDENTIFICATION AND QUANTIFICATION
.XI Organic Compound Identification by Retention Time A-l
A.2 Adequate Peak Resolution A-2
A.3 Proper Response Factors A-2
A.3.1 Different Response Factors for Different Detectors A-3
A.3.2 Different Response Factors for Different Compounds A-3
A.3.3 Different Response Factors for the Same Compound A-5
B. VOC OBSERVATION PROCEDURES
B.I Agency Use of Screening Measurement Methods During
Compliance Test B-l
B.2 On-site Observation Procedures Coupled with the Use of
Agency Screening Methods B-2
B.2.1 First Sampling Run \ B-2
B.2.2 Second Sampling Run B-3
B.2.3 Third Sampling Run B-4
-------
TABLE OF CONTENTS (Continued)
Appendix
C METHOD 18 OBSERVATION PROCEDURES
Page
C.1 Selection of Proper Sampling and Analytical Technique
for Method 18 C-l
C.2 Observation of On-site Testing C-14
G2.1 Evacuation Container Sampling (Heated and Unheated) C-15
C22 Direct Interface Sampling C-27
C.2.3 Dilution Interface Sampling C-30
C.2.4 Adsorption Tube Sampling C-32
C.3 VOC Sample Analysis C-36
C.3.1 Preparation of Calibration Standards C-36
C3.2 Analysis of Direct Interface Samples C-42
C.33 Analysis of Dilution Interface Samples C-43
C.3.4 Analysis of Adsorption Tube Samples C-45
C.4 Auditing Procedures C-48
C.5 References G49
D. METHOD 25 OBSERVATION PROCEDURES
D.I Specifications for Method 25 Sampling Equipment D-l
D.2 Specifications for Method 25 Analytical Equipment D-4
D.3 Method 25 Nomenclature and Equations D-4
VI
-------
LIST OF FIGURES
Figure Page
3.1 Preliminary Survey Data Forms 3-13
4.1 Integrated Bag Sampling System 4-16
42 Field Sampling Data Form for Container Sampling 4-17
4.3 Data Form for Analysis of Method 18 Field Samples 4-18
4.4 Field Audit Report Form 4-19
6.1 Method 25 Sampling Train 6-22
6.2 Method 25 Filter Housing 6-24
6.3 Method 25 Condensate Trap 6-25
6.4 NMO Analytical Cycle 6-26
6.5 NMO Sample Delivery Schematic 6-27
6.6 Condensate Recovery System 6-29
. t,
6.7 Liquid Sample Injection Unit 6-30
6.8 Field Audit Report Form 6-32
6.9 Schematic of Method 25 Audit System 6-33
7.1 Method 25A Sampling Train 7-14
7.2 Method 25A Sampling Checklist 7-15
73 Calibration/Sample Valve Assembly with Ambient Dump 7-16
8.1 VOC Compliance Test Report Format 8-8
8.2 Compliance Test Report Review Form - Report Contents 8-10
A.1 Adequate Peak Resolution A-4
vu
-------
LIST OF FIGURES (Continued)
Figure Page
C.1 General Scheme for Selection of Appropriate Sampling Techniques C-13
C2 On-site Measurement Checklist C-17
C3 Direct Pump Sampling System C-23
C.4 Explosion Risk Area Sampling System Option Using an Evacuated
Steel Container C-24
C5 Direct Interface Sampling System C-28
C.6 Direct Interface Sampling Form C-29
C.7 Dilution Interface Sampling System C-31
C.8 Adsorption Tube Sampling System C-34 •
C.9 Field Sampling Data Form for Adsorption Tube Sampling C-35
CIO Post Sampling Operations Checklist C-37
D.I Recommended Standard Format for Reporting Method 25
Data and Results D-8
vui
-------
LIST OF TABLES
Table Page
2-1 Partial Listing of Data for Compliance Test Protocol 2-4
5-1 Source Categories that Emit Fugitive VOCs 5-2
5-2 Portable Instruments Range, Sensitivity, and Response Time 5-4
5-3 Recommended Calibration Gases for Routine Instrument Service 5-8
5-4 Calibration Time Requirements 5-9
5-5 Performance Criteria for Portable VOC Detectors 5-14
6-1 Method 25 Sampling Equipment Component and Calibration
Specifications 6-23
6-2 Analytical Component and Calibration Schedule 6-28
6-3 Activity Matrix for Method 25 Auditing Procedures 6-31
C-l Status of Selected Organic Compounds for Method 18 Sampling and
Analysis Techniques C-2
' Kg
C-2 Method 18 Sampling Techniques for Selected Volatile Organic Compounds . C-4
C-3 GC Detectors for Selected Volatile Organic Compounds by Method 18 C-6
G4 Recommended Calibration Techniques for Selected Volatile Organic
' Compounds by Method 18 C-8
D-l Method 25 Equipment Checklist D-2
D-2 Method 25 Sampling Checklist D-3
IX
-------
-------
CHAPTER 1
INTRODUCTION
Under current environmental regulations, a plant or facility that emits volatile
organic compounds (VOCs) into the atmosphere must maintain emissions at or below
certain levels, as set forth in the applicable Federal, State, and local standards.
Compliance testing, in which emissions are sampled while the plant operates under those
normal maximum operational levels which are expected to produce maximum emissions,
is the "means by which emissions are documented and permits to operate are obtained.
Agency personnel observing the execution of compliance testing, and reviewing the test
protocol and compliance test report are Test Coordinators."*
/ .
The purpose of this report is to provide the test coordinator with procedures to
(1) identify the data necessary to determine compliance, (2) oversee the compliance test,
and (3) review the compliance test report written by the testing firm. This manual is
intended to provide guidance to agency representatives with various levels of experience
in coordinating VOC compliance tests. A detailed overview of the methods has been
provided for the more experienced test coordinator.
While the facility is responsible for the proper conduction of the compliance test,
it is the agency representative's responsibility to ascertain that the test is conducted in
accordance with established conditions. It is the agency representative's responsibility to
ascertain that the test is conducted in accordance with established conditions. The test
coordinator determines whether the test protocol, probably prepared by a testing firm
and submitted by the facility, will provide the data necessary to determine compliance
with a reasonable degree of certainty; conducts a pretest meeting, or other pretest
communication, to finalize the test conditions; coordinates the compliance test; and
reviews the data submitted in the compliance test report.
A performance test consists of the following steps:
1. The source owner/operator notifies the agency (responsible for
determining compliance) of the proposed test.
2. The proposed test plan is submitted to the agency.
3. A pretest meeting, or in some circumstances a pretest teleconference, is
held at the affected facility to finalize the test conditions.
4. The conditions established at the pretest meeting are formally documented
in writing by the agency, by letter, or a form suitable for this purpose, or
other acceptable means.
"The use of this terminology (i.e., Test Coordinators") instead of the more commonly
used terminology "observers," reflects the expanded role, hi addition to observing the
actual test, of EPA representatives in having a test performed in an acceptable manner.
1-1
-------
5. The performance tests are conducted as coordinated by the agency
representative.
6. Preparation of a report by the agency coordinator to document the events
concerning the tests.
7. Receipt and review of the test report and documentation of the official test
results.
This manual deals with the coordination of compliance testing for volatile organic
compounds. A volatile organic compound (VOC) is defined in 40 CFR Subpart A,
General Provisions, 60.2, as any compound of carbon, excluding carbon monoxide,
carbon dioxide, carbonic acid, metallic carbides or carbonates, and ammonium
carbonate, which participates in atmospheric photochemical reactions or which is
measured by a reference method, an equivalent method, or an alternative method; or
which is ^determined by procedures specified under any subpart. Negligibly
photochemically reactive solvents used hi inks to decrease drying time or other purposes
do not contribute to the total VOC emissions tally. These materials should not count
toward VOC emission levels if they are "exempt" from the applicable regulation. These
negligibly-reactive compounds shall not be considered VOC if the amount of such
compounds can be and is accurately quantified. The EPA considers the following
organic solvents to have negligible photochemical reactivity, and therefore does not
consider them to be VOCs.
Methane
Ethane
1,1,1-Trichloroethane (methyl chloroform)
Methylene chloride (dichloromethane)
Trichlorofluoromethane (CFC-11)
Dichlorodifluoromethane (CFC-12)
Chlorodifluoromethane (CFC-22)
Trifluoromethane (CFC-23)
Trichlorotrifluoroethane (CFC-113)
Dichlorotetrafluoroethane (CFC-114)
Chloropentafluoroethane (CFC-115)
1,1,1-Trifluoro 2,2-dichloroethane (HCFC-123)
1,1,1,2-Tetrafluoroethane (HFC-134a)
1,1-Dichloro 1-fluoroethane (HCFC-141b)
1-Chloro 1,1-difluoroethane (HCFC-142b)
2-Chloro-l, 1,1,2-tetrafluoroethane (HCFC-124)
Pentafluoroethane (HFC-125)
1,1,2,2-Tetrafluoroethane (HFC-134)
1,1,1-Trifluoroethane (HFC-143a)
1,1-Difluoroethane (HFC-152a)
1-2
4
-------
• Perfluorocarbon compounds which fall into these classes: (1) cyclic,
branched, or linear, completely fluorinated alkanes, (2) cyclic, branched, or
linear, completely fluorinated ethers with no unsaturations, (3) cyclic,
branched, or linear, completely fluorinated tertiary amines with no
unsaturations, and (4) sulfur containing perfluorocarbons with no
unsaturations and with sulfur bonds only to carbon and fluorine
Many States also do not consider some or all these materials to be VOCs.
Chapter 2 of this report provides the coordinator with procedures and references
for establishing the test objectives. Chapter 3 discusses the pretest survey and the
procedures for coordinating the compliance test. Chapters 4, 5, 6, and 7 present
sampling and analysis observation procedures for Methods 18, 21, 25, and 25A,
respectively. Chapter 8 presents review procedures for the compliance test report
submitted by the facility. The test coordinator must be familiar with the regulations and
the test methodology.
1-3
-------
-------
CHAPTER 2
ESTABLISHING TEST OBJECTIVES
Under Federal regulations, the facility is required to provide at least thirty days
prior notice of the scheduled compliance test The requirement for submission of a
proposal test protocol is usually established by the agency. To develop the test protocol
and conduct emissions measurements, the facility usually retains the services of a testing
firm. The test protocol is used as a starting point in establishing test conditions. The
test coordinator reviews the test protocol, requests changes if necessary, and approves
the final test protocol. This chapter provides the procedures for establishing the test
objectives and for determining the acceptability of the written test protocol.
2.1 REVIEW OF APPLICABLE REGULATIONS AND BACKGROUND
INFORMATION
The test coordinator must be familiar with the regulations which are applicable to
the facility to be tested. A more thorough understanding of all applicable regulations
and guidelines typically can be obtained through discussions with each of the applicable
agency groups. Typical agency groups (agency organizations vary) and types of
information needed from these groups are discussed below. The test coordinator
generally has copies of the regulations and is usually familiar with them.
Compliance Group - Copies of all applicable regulations should be obtained.
Previous compliance history and current compliance problems should be determined.
Permit Group - The test coordinator should obtain a copy of the existing permit
to operate, if one exists, and a copy of the permit to construct. These permits are
necessary because, depending on the agency, they list process and control equipment
operating requirements. All conditions and requirements for an existing permit should
be understood. If no previous operating permit exists, the construction permit is used.
Inspection or Enforcement Group - The test coordinator should determine what
measurements and other parameters are necessary to establish representative facility
operations.
Obtaining Background Information • If needed, the test coordinator may obtain
additional background information on testing methodology and process and control
equipment operations through State agency personnel, EPA technical manuals, and EPA
contact personnel. EPA has a computer bulletin board which agency test coordinators
can use to obtain background information. The Emissions Measurement Technical
Information Center (EMTIC) computer bulletin board service (BBS) can be reached by
calling, on a computer modem, (919) 541-5742. A listing of EPA manuals and contact
personnel related to emissions measurement can be found on the EMTIC BBS. The
agency emissions test coordinator must be a registered user of the BBS.
2-1
-------
The agency contacts and the EMT1C BBS may provide the following types of
information:
1. Background information on how the regulation was established and the
intent of the regulation.
2. Legal determinations and policy memorandums for facility operations,
interpretation of the regulation, and testing methodology.
3. Problems generally associated with facility operation and testing
methodology.
4. Process operational procedures and control equipment operation
procedures that provide short term emission reductions. These procedures
are occasionally used by source personnel to reduce emissions during
compliance testing.
5. Compliance test reports from similar sources.
6. Source history information. This will include all permits, compliance test
reports, and inspections. If the facility has been cited for noncompliance, it
is extremely important that the test coordinator be aware of these actions
and their status and that no discussion with facility personnel regarding
these problems occur without prior knowledge of the agency.
22 DETERMINATION OF COMPLIANCE LIMITS
The compliance limit will be specified in the applicable regulations or permit. It
is critical that the test coordinator understand the measurement units of the applicable
emission standard or limit and how they are determined. If process rate or weight is
used in expressing the limit, then the definitions for process rate or weight must be
understood. Many regulations exempt certain compounds from the limits (e.g., methane
or ethane may be exempted as VOC). These exemptions should be determined and
understood.
For limits expressed as a concentration, the units of the standards are generally
parts per million by volume (ppmv). If the emissions are expressed as parts per million
by weight (ppmw), significantly different values will be obtained. Also, the standards
typically require (1) correction of measurements expressed as parts per million by
volume to a dry basis at a standard temperature and pressure and (2) no dilution air.
The EPA Methods for VOC determination produce emission results on several
different units of measurement bases. Therefore, results from Methods 18, 21, 24, 25,
and 25A may not be directly comparable, unless additional procedures and calibrations
are performed. Emission measurements, uncontrolled emissions, and methods for
measuring collection efficiency, must yield data on the same unit basis. The basis of
results for EPA Methods for VOC determination and a comparison of VOC data
obtained by the use of different methods are discussed later. ,
2-2
-------
2J DETERMINATION OF DATA NECESSARY TO SHOW COMPLIANCE
The test coordinator should determine whether the test protocol submitted to the
agency provides the framework to collect data necessary to demonstrate compliance with
a reasonable degree of certainty. The test coordinator should: (1) understand the
requirements of the compliance test and (2) match requirements of the compliance test
to the specifications in the written test protocol.
The test coordinator may also find it helpful to use the pretest survey form,
presented and discussed in Chapter 3, to outline the requirements of the compliance test
A partial listing of necessary data requirements is presented in Table 2-1, page 2-4.
2.4 BASIS AND COMPARISON OF RESULTS FOR EPA METHODS 18,21, 25,
AND 25A
When the results from two VOC measurement methods are to be compared, the
test coordinator must understand the unit basis of method results (type of calibration
standards and emissions correlation to calibration standards) for Methods 18, 21, 25, and
25A. Method 18 identifies only those compounds for which sampling and analysis is
specifically conducted; results are expressed in terms of concentration of those specific
organic compounds. Method 18 does not identify or quantify unknown compounds.
Methods 21, 25, and 25A do not provide results on an organic compound specific
basis (i.e., the exact organic compounds measured cannot be determined from the
emission results); measurement results from these methods are expressed in terms of the
calibration standard (e.g., ppm as propane or ppm as carbon) or as in the .case of
Method 24, on a total organic basis (percent volatile organics). -.-
When the facility proposes the use of two different methods for collection
efficiency determination, the test coordinator must determine the acceptability of the
testing protocol.
The sampling and analytical methods used at each sampling location must provide
emission results on the same unit measurement basis.
^
Ensuring that measurements are obtained on the same unit measurement basis
can be complicated. When the test coordinator is uncertain of procedures, the EMTIC
representative who is listed as the VOC contact can be consulted.
2-3
-------
TABLE 2-1. PARTIAL LISTING OF DATA FOR COMPLIANCE TEST
PROTOCOL
Safety Considerations
• Required by EPA
• Required by plant
• Recommended by OSHA
Process
Facility identification and description
Parameters to be monitored and recorded (including recording times
time intervals)
Acceptable range for each parameter monitored
Raw materials to be used
Fuel
Process samples to be taken and analyzed
Process rate
Mode of operation
Manual or automatic operation
Cleaning and auxiliary systems
Normal period for process cycle
Materials processed - coverage and/or shape
Diversion or circumvention of pollutants from air pollution
control equipment
Operational personnel (Must be the same as scheduled for
normal day-to-day operations of facility)
Instruments to be added and/or calibrated
3., Control Equipment
Description of control equipment
Parameters to be monitored and recorded (including recording
tunes)
Acceptable values for each parameter
Control equipment effluent samples to be taken and analyzed
2-4
-------
TABLE 2-1. PARTIAL LISTING OF DATA FOR COMPLIANCE TEST
PROTOCOL (Concluded)
* Mode of operation
* Manual or automatic operation
* Collected pollutant removal cycle
* Cleaning cycle
* Auxiliary or gas conditioning systems
* Instruments to be used in monitoring operation
• Instruments used in monitoring operation to be calibrated
4. Measurement Methodology
• Basis of results
• Sample run time
• Portion of the process cycle to be'tested
• Portion of the control equipment cycle to be tested
• Sampling locations with drawing showing bends, obstructions,
control equipment, etc.
• QA audit samples to be provided by agency
• Fugitive emissions (not discussed in this manual)
• Transfer efficiency (not discussed in this manual)
• Sampling procedures to be used (emission, process and control
equipment samples) , '-«.
* Hood capture efficiency (not discussed in this manual)
• Material balance (not discussed in this manual)
• Sampling procedures not required -by method but required by
control agency
• Analytical procedures to be used (emission, process and control
equipment samples)
• Analytical procedures not required by method but required by
control agency
• QA/QC procedures required (e.g., performance audit samples)
• Report format and required data
• Time restrictions for submittal of report
2-5
-------
-------
CHAPTERS
PREPARATION FOR AND OBSERVATION OF COMPLIANCE TEST
This Chapter addresses preparation and procedures for the on-site observation of
the compliance test. Preparation for the test observations includes scheduling personnel,
scheduling equipment and ordering performance audit gases. A pretest meeting should
be conducted at the facility at least 15 days prior to the commencement of testing. This
gives the test coordinator the opportunity to inspect the facility, including any air
pollution control equipment and test location, determine which operational parameters
should be monitored, and obtain information to finalize the test conditions. Preparation
also includes conducting a pretest survey when necessary. The sampling procedures
common to all EPA VOC Methods are discussed in Section 4 of this chapter. The
specific sampling and analytical procedures for Methods 18,21,25, and 25A are
presented in Chapters 4, 5, 6, and 7, respectively.
Observation of VOC compliance tests is typically more difficult than for tests for
other criteria pollutants. Because of the complex nature of compliance testing for VOC,
Method 18 recommends that the tester conduct site specific and, if applicable, compound
specific pretest preparations. The test coordinator may also find it useful to conduct a
pretest survey.
The function of the test coordinator is to ascertain that the test are conducted in
accordance with the conditions established at the pretest meeting. The test coordinator
has the authority to require changes in the testers' procedures if they are not in
accordance with the regulatory requirements or not in agreement with the established
conditions.
The test coordinator's principle function is to evaluate the representativeness of
compliance testing. In other words, the compliance test results should represent
emissions typical of maximum operating conditions which produce maximum emissions.
Representativeness is typically evaluated in terms of five criteria; if any of the criteria
are not met, the compliance test is considered nonrepresentative:
1. Process and control equipment must be operated in such a manner as to
produce representative samples of controlled and uncontrolled emissions. By
measuring the emissions before and after control devices, the removal or
control efficiency can be determined.
2. Locations of the sampling ports and points must provide samples which are
representative of the total uncontrolled (if applicable) and controlled process
emissions.
3. Samples collected in a sampling train must be representative of the
concentration(s) at the sampling point(s).
4. Samples recovered and analyzed must be representative of samples collected
in the sampling train.
3-1
-------
5. Reported results must be representative of facility operations and the samples
recovered and analyzed.
3.1 NOTIFICATION OF COMPLIANCE TEST
For Federal New Source Performance Standards (NSPS), the general
requirements for performance (compliance) testing are presented in Title 40 Part 60.8 of
the Code of Federal Regulations. The requirements are as follows:
1. Performance (compliance) testing within 60 days after achieving the maximum
production rate at which the facility will operate, but not later than 180 days
after initial startup.
2. Performance testing and data reduction in accordance with the test methods
and procedures contained in each applicable subpart unless the Administrator
allow one of the four options listed.
3. Performance testing conducted under such conditions as the Administrator
shall specify to the plant operator based on representative performance of the
affected facility. r
4. At a least 30 days prior notice of any performance testing to the
Administrator.
5. Provision of (1) sampling ports, (2) safe sampling platforms, (3) safe access to
sampling platforms, and (4) utilities for sampling by the facility.
6. Three separate runs per compliance test. In the event of the items listed in
60.8(f) the Administrator may accept two sample runs as a test.
The Federal requirements ensure that the agency is notified prior to the test and
that the sampling site is acceptable. Many States have developed similar guidelines to
ensure proper notification of compliance tests. Testing arrangements and scheduling of
staff and equipment for the observation are also facilitated by an agency guideline. A
typical State agency guideline requires or includes the following:
1. A testing protocol submitted by the facility.
2. 30 Day notification of testing.
3. Testing to be conducted during normal agency business hours.
4. Operational conditions during testing.
5. Quality assurance requirements for the compliance testing (e.g., performance
audit samples).
6. Testing procedures required by the State which are deviations from the EPA
Methods.
7. Compliance test report format (see Chapter 8 for example format).
8. Safety and sampling access requirements.
Notification 30 days prior to the test and submission of a written test protocol
allows the test coordinator sufficient time to review the test protocol, conduct a pretest
3-2
-------
meeting to establish that the requirements for compliance testing will be met, and order
performance audit materials.
Performance Audit Materials - The test coordinator must obtain the proper audit
materials and/or devices. The testing firm analyzes the audit materials on-site as part of
the compliance test. To select the proper audit gas cylinder(s), the test coordinator will
consult the applicable method, and written test protocol supplied by the tester to
determine the type (specific organic compound) and proper concentration range of the
audit gas to order. Table C-l, page C-4, lists audit gases available from EPA for
common target organic compounds. Availability and ranges of audit gases can be
determined by contacting:
U.S. Environmental Protection Agency
Environmental Monitoring Systems Laboratory
Quality Assurance and Technical Support Division
Research Triangle Park, North Carolina 27711
Attention: Audit Cylinder Gas Coordinator
For audit gases obtained from a commercial gas manufacturer, ensure that the
manufacturer has (1) certified the gas in a manner similar to the procedure described in
40 CFR Part 61, Appendix A, Method 106, Section 5.2.3.1 and (2) obtained an indepen-
dent analysis of the audit cylinder to verify that the audit gas concentration is within 5
percent of the manufacturer's stated concentration.
To accurately assess the emission measurements, the performance audit sample
concentrations should fall within the range of approximately 50 to 200 percent of the
expected emissions concentration. Interpretation of audit results is discussed in
Chapter 8. Performance audit gases with concentrations 5 times greater or 5 times less
than the expected emission value or an organic compound different than that being
measured should be not used.
3.2 CONDUCT PRE-TEST SURVEY
Prior to compliance testing, the affected facility is often visited by a representative
of the testing firm and the test coordinator. This information gathering visit is referred
to as the pretest survey. Agencies should be strongly urged to conduct pretest meetings
at the affected facilities well in advance of the test date. During the pretest survey of
the process and control equipment, the test coordinator may find it useful to conduct
screening measurements to (1) establish some operating baseline values for process and
air pollution control equipment and (2) to determine problems with the methods to be
applied later in the compliance test. The test coordinator may prepare a form of
information to gather during the pretest survey. An example of a general pretest survey
checklist is shown in Figure 3.1, page 3-13. Because of the complex nature of most
organic processes, the test coordinator may provide the testing firm with the example
3-3
-------
checklist to be completed to enhance the probability that the testing will be performed
as required.
Since most organic gases are invisible, conducting independent testing to estimate
the emission levels and other key parameters using a portable organic analyzer (EPA
Method 21 instrumentation) can be extremely useful. Appendix B provides an example
of how these instruments can be used.
i
One of the primary concerns for any organic sampling program must be safety.
The test coordinator should always question the facility representative concerning general
plant safety requirements and safety in regard to sampling. Every test protocol should
address the safety considerations involved in performing the test. There are numerous
safety considerations involved in organic sampling, particularly regarding health effects
and explosion hazards, however, it is beyond the scope of this manual to discuss each
one in detail. It cannot be over-emphasized that the test coordinator must always be
aware of the safety hazards.
3.3 FINALIZING COMPLIANCE TEST PROTOCOL
Before the compliance testing, it is recommended that the agency test coordinator
meet with a representative of the testing firm, and a facility representative. At this
meeting the compliance test protocol can be finalized, the baseline (representative)
facility operating conditions can be established, and the testing schedule can be
coordinated. The pretest meeting agreements, or conditions established, should be
documented in a final letter from the agency test coordinator to the facility
owner/operator, or on a form/checklist completed by the test coordinator.
The testing firm representative must know the exact sampling procedures to be
used, the minimum data and reporting requirements, and the conditions that constitute
an invalid test. If the test coordinator will use a checklist to monitor sampling and
analytical procedures, it is beneficial to provide the checklist to the testing firm to ensure
that all required steps will be completed. Likewise, the facility representative should
explain what process and control equipment data will be recorded, the intervals of data
collection, the raw materials to be used, tested, and the conditions that could constitute
an invalid test. Since it is the facility representative's responsibility to obtain the
baseline operational parameters, it is the agency's responsibility to designate which
parameters will be monitored and recorded. Execution of the compliance test in
accordance with the established test protocol should constitute a valid test.
The lines of communication for the compliance test should be defined. It is
recommended that all official communications regarding facility operations, testing
methodology, and agency policy be limited to the test coordinator, facility representative,
and testing firm representative. It can be useful to know the names of the supervisors of
these individuals in the event of poor cooperation or when requests for information are
questioned.
3-4
-------
As a minimum for relatively simple processes, a teleconference should be
conducted to establish the test conditions, followed by formal documentation. In all
cases, whether the process and sampling is simple or complex, it is the test coordinator's
responsibility to be certain that all details of the test procedure are understood and
accepted before the test begins.
/ •
Procedures common to most types of VOC testing and relevant to on-site
observation are presented below. The sampling and analytical procedures specific for
.Methods 18, 21, 25, and 25A are addressed in Chapters 4, 5, 6, and 7, respectively.
3.4 PROCEDURES COMMON TO MOST TYPES OF VOC TESTING
Determination of measurement errors, measurement of flue gas flow rate,
moisture determination, organic compound identification by retention time, proper gas
chromatography (GC) peak resolution, and application of proper response factors are
required for most VOC testing. Each is discussed below.
3.4.1 Measurement Errors
The procedure for determining pollutant emission rates by stack sampling involves
measurement of a number of parameters. Errors of measurement associated with each
parameter combine to produce an error in the calculated emission rate. Measurement
errors are of three types: bias, blunders, and random errors. Bias errors, usually a result
of poor sampling and analytical technique, cause the measured value to differ from the
true value in one direction. Many bias errors can be eliminated through proper
calibration of the equipment. Most blunders occur during sampling, sample transport,
and sample preparation for analysis. Elimination of blunders should be a main concern.
Random errors (precision), which result from a variety of factors, cause measured values
to be either higher or lower than the true value. Such errors result from the inability of
sampling personnel to read scales precisely, poor equipment performance, and lack of
sensitivity in measurement devices. The usual assumption is that random errors are
normally distributed about the mean or true value and can be represented statistically in
terms of probabilities. All methods have some inherent random error (precision).
To make on-site decisions based on the significance of error, the test coordinator
must know three things to determine the importance of the error.
1. Does the facility have to prove compliance with the standards (Federal
regulations) or does the agency have to prove a violation of the standards
(State regulations)?
2. What are the direction and magnitude of any biases?
3. What is the acceptable bias that will be allowed before rejecting the results?
3-5
-------
If a facility is attempting to prove itself in compliance with most Federal
regulations, any magnitude of high bias in an outlet emission measurement (measured
emissions higher than actual emissions) conclusively showing compliance would be
allowed. However, these high biased results might not be valid for use in emissions
. trading or banking. A low bias in an outlet emission measurement would be acceptable
if it did not cause the reported results to show compliance rather than violation. Since
the final results are not generally known during the on-site testing, it is preferable for the
test coordinator to have a fixed value to apply in allowing or disallowing a test run while
on-site. EPA method development testing reports indicate that most VOC test methods
have a precision of about 10 percent. A listing of EPA method development reports and
the method's precision are on the EMTIC BBS. Therefore, a good rule of thumb for
allowing biases determined on-site is up to 10 percent high bias and 5 percent low bias.
When the State or local agency has the burden of proving that the source is in
violation with the applicable regulation, any low bias in a measurement (measured
results less than the true value) that still proves the source in violation would be
acceptable. The low bias will not be challenged by the facility unless the testing does not
comply with legal requirements as stated by the applicable test method. When the
agency bears the proof of violation, a good rule of thumb for acceptable bias is up to
5 percent high bias and 10 percent low bias. Again, it should be stressed that for proof
of violation, meeting the requirements of the applicable test method is mandatory.
Errors from the measurement of most sampling parameters have very little effect
on the final data results. The test coordinator or tester may be able to calculate the
direction and magnitude of a measurement error.
3.4.2 Determination of Flue Gas Flow Rate
For ducts equal to or greater than 12 inches in diameter, the number of sampling
points necessary to determine the flow rate is specified by EPA Method 1, Figure 1-2,
"Minimum number of traverse points for velocity (nonparticulate) traverses." For ducts
less than 12 inches in diameter, EPA Method 1A should be used to determine the point
location for velocity measurements. Sampling port locations upstream of air pollution
control equipment do not typically meet Method 1 requirements. If a sampling location
does not meet minimum requirements and the system is closed with no air entering or
leaving, then the flow rate at the outlet location (after control equipment) can be
measured and the standardized flow rate used for the inlet location.
Flow rate is determined by EPA Method 2 for large ducts equal to or greater
than 12 inches in diameter. Method 2 uses a "S" type pilot tube to determine the
average velocity pressure. The velocity pressure and the stack gas molecular weight
(from Method 3) and stack gas moisture content (from Method 4) are used to determine
the flue gas flow rate. For ducts less than 12 inches in diameter within the temperature
range of 0 to 50°C, Method 2A can be used to measure the gas volume flow rate directly
3-6
-------
with a gas meter. Method 2B is used to measure gas volume flow rate from gasoline
vapor incinerators. This method determines the flue gas flow rate prior to combustion
and then calculates flue gas flow rate after combustion based on a carbon balance.
Method 2C applies to ducts less than 12 inches in diameter and measures flue gas flow
rate with a standard pitot tube instead of a type S pitot tube. Method 2D also applies to
ducts less than 12 inches in diameter and uses the same approach as 2A by measuring
flow rate directly with a rotameter or orifice plate.
* Flue gas flow rates measured using Methods 1 and 2 are typically within
10 percent of the true flow rate values for all the methods shown above (Methods 2A,
2B, 2C, and 2D). Sections 3.0 and 3.1 of EPA's Quality Assurance Handbook, Volume
HI, EPA-600/4-77-027b, discusses errors associated with velocity measurements.
3.4.3 Moisture Determination
Stack gas moisture content must be determined when (1) flue gas flow rate is
determined or (2) stack gas concentration is measured (container sampling or direct
interface sampling). The moisture content is used to correct the emission concentration
or mass emission rate to a dry basis. EPA Method 4, Section 3.3 of EPA's Quality
Assurance Handbook, Volume in, EPA-600/4-77-027b, is used to measure stack gas
moisture content Section 3.3 of the Quality Assurance Handbook, Volume HI, provides
detailed information on the application of Method 4.
For flue gas streams at or below 60°C (140dF), flue gas moisture content can be
determined using wet bulb/dry bulb thermometers and the partial pressure equation
shown below. This is an approximation method, as described in Method 4. Obtain the
wet bulb/dry bulb temperatures as follows:
1. Moisten the wet bulb thermometer wick with deionized distilled water.
2. Insert thermometers into flue gas stream and monitor wet bulb temperature.
3. When wet bulb temperature has stabilized, record both wet bulb and dry
bulb thermometer temperatures.
4. Calculate flue gas moisture content (%H2O) using the equations listed below.
Equation 3-1
P. +
-------
x (I+(J*w-32yi571)) x 100 Equation 3-2
where:
w2 • Calculated constant, saturation % H2O at Tw
Tw = Wet bulb temperature, °F
Td = Dry bulb temperature, °F
Pb - Barometric pressure, in. Hg
Ps = Static pressure of duct, in. H2O
3.4.4 Organic Compound Identification and Quantification by Gas Chromatography
When using Method 18, the organic compounds to be measured must be known
prior to the test. To identify and quantify the major components of the organic
compounds known to exist in the sample, the retention time of each component is
matched with the retention times of the known compounds (the standard reference
material or calibration standard) under identical conditions. Separation of organic
compounds is performed with gas chromatographic columns, referred to as GC analysis.
The retention time is the time between sample injection into the GC and when the
organic compound reaches the detector. If GC conditions remain constant, the retention
time for each compound will be constant as well, and will serve as the identifying
parameter for each peak. Care must be taken to assure that two compounds do not
share the same retention time. The retention time shall be within 0.5 seconds or
1 percent of the retention time of the known compound's (calibration standard)
retention time (whichever is greater) to be considered acceptable. The retention time
will vary with (1) type of column or column material, (2) length of column, (3)
temperature of column, (4) organic compound, and (5) several other factors. The exact
seconds or minutes of the retention time do not matter, except the longer the retention
time, the longer the analysis time.
To obtain proper quantitative values, sample peaks (the result of organic
compounds as they reach the detector) must be properly separated to enable the
detector to analyze only the compound of interest.
Understanding the use of the response factor is important because (1) different
detectors can have a different response factor for the same compound, (2) each detector
can have a different response factor for different compounds, and (3) the same detector
can give a different response factor for the same compound at different conditions. The
response factor for each compound on any detector can be determined by dividing the
area units from the integrator printout of the standards by the concentration of the
standard (area units/ppm of standard). This is done for all concentrations of standards
used to calibrate the detector.
3-8
-------
Method 25 was developed to eliminate the reduced response factor problem when
the organic compounds are unknown. When the organic compounds in the sample are
unknown (e.g., after an incineration process), then proper calibration gases cannot be
selected. To minimize this problem, Method 25 removes all the elements that give
reduced response factors and analyzes the compound as methane in terms of carbon.
The results are then reported as parts per million as carbon. Unfortunately, the true
molecular weight of the compound is lost and a concentration or mass emission rate
cannot be calculated.
A detailed discussion of organic compound identification and quantification is.
presented in Appendix A.
3.5 ON-SITE OBSERVATION PROCEDURES
The attitude and behavior of the agency test coordinator during the pretest
meeting and compliance test, are of the utmost importance. The test coordinator should
conduct his/her duties thoughtfully and thoroughly, not disrupting the testers as they
perform the tests. The test coordinator must also avoid appearing meek or reluctant to
authoritatively represent the agency's interest. If problems with facility operation or
sampling are noted by the test coordinator during the compliance test, it is recommended
that the test coordinator deal solely with the designated testing firm representative and
facility representative; he/she should have a clear understanding with them if it is
necessary to communicate with other testing firm personnel or facility operations
personnel. Conversely, it may be advisable to refrain from answering inquiries from the
testing firm personnel and facility operations personnel concerning agency enforcement
policy. T
During the test, the test coordinator should check to ensure representative facility
operations and adherence to specified sampling procedures. To eliminate the possibility
of overlooking necessary checks and to provide the agency with documentation to use in
later enforcement actions, the test coordinator can use checklists covering details of
process operations, control equipment operations, and sampling procedures. The
pretest survey forms previously discussed in Chapter 3.2 can be used to develop the
process and control equipment checklist. Sampling checklists for Methods 18, 25, and
25A are presented in the applicable discussions later in this manual.
Since additional measurements are typically not made by the agency, the
recommended procedures for conducting on-site observations do not include use of
agency conducted screening measurements. Independent screening measurements
conducted by an agency during the compliance test are discussed in Appendix B.
The remainder of this Chapter presents a recommended scheme for the test
coordinator to use in conducting the test observation.
3-9
-------
3.5.1 First Sampling Run
Applicable methods include Method 18, 25, and 25A. If analyses are to be
conducted on-site, acceptable results must be obtained for audit sample(s) prior to any
field sample analyses. An inspection of the sample recovery area and observation of the
sampling train(s) assembly by the test coordinator may be useful in detecting and
eliminating errors before they occur. If only one agency test coordinator is present, the
schedule below will make the most effective use of observation time. These procedures
are provided to assist less experienced test coordinators in establishing a routine for on-
site observations. More experienced test coordinators will follow their established
routine.
For the first test run, after determining that the facility operations are as specified
in the test protocol, the test coordinator should go to the sampling location to observe
the test team recording the initial data. The initial sampling system leak check should
be observed. When the test coordinator is satisfied with the sampling train preparation
and the facility operation, he should allow the testing to begin. He should observe the
sampling procedures for the first 15 minutes of sampling and then conduct a check on
the facility operations. If the process and control equipment are operating satisfactorily
and the data are being recorded as specified, the test coordinator can return to the
sampling location to observe the completion of sampling, giving close attention to the
final readings and the final leak check. If conducted on-site, the sample analysis should
be observed closely during analysis of the audit gas cylinder and the first field sample.
The analyst should be required to conduct all necessary calculations to determine the
field results in terms of the units of the allowable emissions standard (e.g;, ppmv on a
dry, standard condition basis for the specified organic compound). All procedures and
calculations should be validated by the test coordinator. The distance between the
control room, or area, and the test location may create restrictions, which may be
compounded by difficult procedures. The agency coordinator again must use sound
judgement in determining his/her activities during the tests.
3.52 Second Sampling Ron
Applicable methods include Method 18, 25, and 25A. If the test coordinator is
satisfied that the sampling procedures applied during the first test run are proper, he
should spend most of the second run observing process operations, with intermittent
checks on the sampling procedures. At the end of the run, the test coordinator should
return to the sampling location and observe recording of the final data, the final leak
check, transport of the sampling train to the cleanup area, and sample recovery, as
applicable.
3-10
-------
3.53 Third Sampling Run
Applicable methods include Method 18, 25, and 25A. The focus of observation
for the third sampling run is based on observations made during the first and second
runs. The test coordinator's attempts to determine which element(s), facility operations,
sampling procedures, sample recovery procedures, and/or analytical procedures may
introduce the greatest degree of error in the emission measurements. The test
coordinator should then place the most emphasis on these elements. However, at least a
brief check of each should be included in observations made during the third run.
3.5.4 Sample Recovery and Transport
It is important that sample recovery, sample transport, and analysis are conducted
according to applicable method procedures and are well-documented. To reduce the
possibility of invalidating the test results, all of the samples should be placed in sealed,
nonreactive, properly identified containers. The samples must then be delivered to the
laboratory and analyzed within the sample stability time specified by the method or
determined by the preliminary evaluations. Each container must be uniquely identified,
at the time of sample recovery, to preclude the possibility of interchange. The number
of each container must be recorded on the sample recovery form (which documents the
chain of custody) and the analytical data sheet.
i
3.5.5 Analysis
When the analysis is conducted at the testing firm's laboratory or an outside
laboratory rather than in the field, the use of performance audit samples is the best
method for determining if the sampling and analytical procedures were followed.
Consult the applicable Method to determine the allowable error (acceptable results) for
the audit sample analysis. Acceptable audit sample results cannot assure acceptable
results on the field samples. Acceptable audit sample results only indicate that the
method was conducted properly and that the calibration standard values were correct.
Analytical errors are generally difficult to detect by reviewing the compliance test
report without the aid of performance audit information. Instrument integrators which
record and calculate laboratory data should reduce analytical laboratory error.
Computers used to calculate emissions data can greatly reduce calculated analytical
errors. The laboratory data, calibrations, and calculations must be well-documented and
presented in the compliance test report in such a manner that the test coordinator can
evaluate the validity of the data using procedures presented in the chapters on the
specific methods (e.g., Chapter 4 for Method 18).
3-11
-------
3.5.6 Observation Report
Chapter 8 discusses review of the compliance test report Information gathered
during the pretest survey, if performed by the test coordinator, can be useful in reviewing
the compliance test report The test coordinator may also find it beneficial to organize
the on-site observations upon returning from the field. The test coordinator should
prepare a formal report concerning the performance tests. This is a narrative report
with field notes and forms/checklists attached.
3-12
4
-------
General Information
Name of company Date_
Address
Contact Title.
Phone
Process to be tested
Source ID No. Permit No. Other ID No.
Duct or vent to be sampled
Current permit requirements (attach copy of operating or other permit)
n. General Plant Requirements
Plant safety requirements_
Vehicle traffic rules
Plant entry requirements_
Security agreements_
Potential problems_
Safety equipment requirements (glasses, hard hats, shoes, etc.)
Can photographs be taken?_
ffl. Process and Product Information
Process description
Raw material and fuels that produce the potentially highest
emissions
Raw materials and fuels for compliance test
Figure 3.1. Preliminary survey data forms.
Note: These forms apply to Methods 18, 25, and 25A. Some changes or modifications
are necessary to them when applying to Method 21.
3-13
-------
Will raw materials be sampled and analyzed?
If yes, describe procedures
Estimated precision and accuracy of procedures
Methods that the agency will use in analysis of raw materials
Do plant records demonstrate the raw materials used?
Are the normal recording intervals satisfactory for test?
Should these records be kept on file for future inspections?
Remarks:
Products that potentially produce the highest emissions
Products selected for compliance test
Products sampled and analyzed by source?
If yes, describe procedures
Estimated precision and accuracy of procedures
Methods that the agency will use in analysis of the products
Do plant records demonstrate the products produced?
Are normal record intervals satisfactory for test?
Should these records be kept on file for future inspections?_
Operating cycle
Check: Batch Continuous Cyclic
How is the cycle determined?
Timing of batch or cycle (hours and/or minutes)
Portion of cycle to be tested
Portion of cycle represented by each run
Maximum process rate or capacity
Method to determine process weight or rate
Estimated precision and accuracy of method
Dd~any instruments need calibration?_
changed for test ?
Methods that agency will demonstrate to check rate or weight_
Do plant records reflect process weight/rate?
Are normal record intervals satisfactory for test?
Should these records be kept on file for future inspections?
Other process parameters to be recorded (e.g., temperatures, air flow rate)
Figure 3.1. (Continued)
3-14
-------
Parameter How determined? Recorded
Estimated precision accuracy calibrated _
Acceptable value for parameter units
Method that agency will conduct to check parameter
Are normal record intervals satisfactory for test?
Parameter How determined? Recorded
Estimated precision accuracy calibrated
Acceptable value for parameter units
Method that agency will conduct to check parameter
Are normal record intervals satisfactory for test?
Parameter How determined? • Recorded
Estimated precision accuracy calibrated
Acceptable value for parameter units
Method that agency will conduct to check parameter
Are normal record intervals satisfactory for test?
Parameter How determined? Recorded
Estimated precision accuracy calibrated
Acceptable value for parameter units
Method that agency will conduct to check parameter
Are normal record intervals satisfactory for test?
Should these records be kept on file for future inspections?
Physical arrangement of process - doors open, hoods on, covers on
Normal mode of process operation: manual automatic
Mode of operation for test: manual automatic _
What constitutes a process malfunction? •
How are malfunctions handled with regard to process operation and
notification of agency?
Description of future, planned changes in operations
Normal maintenance schedule
List of parameter records which should be retained by facility
Figure 3.1. (Continued)
3-15
-------
IV. Air Pollution Control Equipment
Description of control equipment
length of cycle
cyclic
Control equipment operations: continuous
How is the cycle determined?
Do instruments determine cycle? Are control equipment data recorded?
Are these records kept on file for future inspections?
Portion of cycle to be tested
Normal mode of operation: manual
Mode of operation for test: manual
automatic
automatic
Control equipment parameters to be recorded (temperatures, air flow rate)
Parameter How determined? Recorded
accuracy calibrated
Estimated precision
Acceptable value for parameter
units
Method that agency will conduct to check parameter
Are normal record intervals satisfactory for test?
Should these records be kept on file for future inspections?
Parameter How determined? Recorded
Estimated precision accuracy calibrated
Acceptable value for parameter units
Method that agency will conduct to check parameter
Are normal record intervals satisfactory for test?
Should these records be kept on file for future inspections?
Parameter How determined? Recorded
accuracy
Estimated precision
Acceptable value for parameter units
Method that agency will conduct to check parameter
Are normal record intervals satisfactory for test?
calibrated
Should these records be kept on file for future inspections?
Removal procedure and sequence for collected materials
Collected materials sampled and analyzed: procedures for
Collected material type
Collected material volume or weight
Methods used by agency on collected materials: type and/or volume
Physical arrangement of control equipment: auxiliary systems or ducts
What constitutes a control equipment malfunction?
Figure 3.1. (Continued)
3-16
-------
How are malfunctions normally handled with regard to keeping the process
and control equipment on line?
Describe future, planned changes in operations
Normal maintenance schedule
V. Sampling Site and Procedures
A. Sampling Site
Site description
Duct shape and size_
Material(s) of construction
Wall thickness inches
Upstream distance inches No. of diameters
Downstream distance inches No. of diameters
Size of test port(s)
Size of access area for test personnel and equipment
Hazards Ambient temp °F
B. Properties of Gas Stream
Temperature °C °F, Data source
Velocity , Data source
Static pressure inches H2O, Data source
Moisture content %, Data source
Paniculate content , Data source
Gaseous components
N2 _ % Hydrocarbons (ppm) Toxics/ Acids (ppm)
CO _ % _ HQ
% HF
SO, _ % _ Other
C. Sampling Considerations
Sampling and analytical procedures to be used - attach procedures
Specific procedures - use check lists presented with Methods
(i.e., Method 18, Chapter 4; Method 21, Chapter 5; Method 25, Chapter 6
and Method 25A, Chapter 7)
Location to set up GC (when applicable)
Figure 3.1. (Continued)
3-17
-------
Special hazards to be considered^
Power available at duct
Power available for GC (when applicable)
Potential problems
Specific safety equipment (glasses, hard hats, shoes, etc.)
D. Other Sampling Considerations
Fugitive emissions: how determined
How controlled
Attach copy of method to determine fugitive emissions if present
Estimated precision and accuracy of method
Screening methods to be conducted by agency
Hood capture efficiency: how determined
Attach copy of method to determine hood capture efficiency
Estimated precision and accuracy of method
Screening methods to be conducted by agency
Parameters to show continuing compliance with capture efficiency
Interval of recording parameter for test
Are test parameters to be recorded in future and kept on file?
Transfer efficiency: how determined
Attach a copy of method to determine transfer efficiency.
Estimated precision and accuracy of method
Screening methods to be .conducted by agency
Parameters to show continuing compliance with transfer efficiency
Interval of recording parameter for test
Are test parameters to be recorded in future and kept on file?
E. Site Diagrams (attach additional sheets if required)
F. Quality Assurance Performance Audit Samples
Quality assurance audit samples taken?
Audit samples: proper compound
proper range(s)
Figure 3.1. (Continued)
4
348
-------
G. Emission Measurement Screening Techniques
Detector tubes or other screening techniques used_
Screening technique conducted: prior to testing
during testing between testing
Remarks
Figure 3.1. (Concluded)
Note: Some of the information in Figure 3.1 may be presented in tabular format, if
desired.
349
-------
-------
CHAPTER 4
MEASUREMENT OF GASEOUS ORGANIC COMPOUND EMISSIONS
BY GAS CHROMATOGRAPHY ~ METHOD 18
4.1 APPLICABILITY
Method 18 as promulgated on October 8, 1983 and revised on February 19,1987
is a generic method which is structured to analyze approximately 90 percent of the total
gaseous organics emitted from an industrial source. This method is used to measure
known organics that are in excess of 1 part per million by volume. It does not include
techniques to identify and measure trace amounts (less than 1 ppm) of organic
compounds, such as those found in building air and fugitive emission sources. Also, this
method will not determine compounds that (1) are polymeric (high molecular weight
compounds such as dioxins and furans), (2) polymerize before analysis (such as glues or
resins), or (3) have very low vapor pressures at stack or instrument conditions (such as
anilines).
4.2 METHOD DESCRIPTION
i
Method 18 is based on extracting a gas sample from the stack at a rate
proportional to the stack gas velocity using one of four techniques: (1) withdrawing the
gases directly from the stack into the analyzer (direct interface sampling), (2) collecting
gases in a container (integrated bag sampling), (3) dilution interface sampling, or (4)
collecting gases on a sorbent tube (adsorption tube sampling). The major gaseous
organic components of a gas mixture are then separated by gas chromatography, and
measured with a suitable detector.
For the first three techniques, the sample or diluted sample is introduced directly
into the sample loop of the gas chromatograph (GC). The measured sample is then
carried into the GC column with a carrier gas where the organic compounds are
separated. The organic compounds are each quantified by a GC detector such as a
flame ionization, photoionization, or electron capture detector. The qualitative analysis
is made by comparing the retention times (from injection to detection) of known
calibration standards to the retention times of the sample components. The quantitative
analysis is made by comparing the detector response for the sample compound to a
known quantity of that compound in a corresponding standard. Gas samples collected
on adsorption tubes are desorbed from the adsorption media using a solvent. A
measured volume of the desorption solution is injected into a heated injection port
where the mixture is vaporized and carried into the GC column with a carrier gas. The
sample is separated into the individual components, then qualitatively and quantatively
analyzed in the same manner as a gas sample.
Gas samples are analyzed immediately as taken from the stack or within a set
period of time after being collected in a container or on an adsorption tube. To select
4-1
-------
the correct GC column and establish proper GC analytical conditions, the analyst must
identify approximate concentrations of organic compounds to be measured beforehand.
With this information, the analyst can then prepare or purchase commercially available
standard mixtures to calibrate the GC under physical conditions identical to those that
are used for the samples. The analyst must also have prior information concerning
interferences arising from other compounds present in the emissions, the need for
sample dilution to avoid detector saturation, gas stream filtration to eliminate particulate
matter, and prevention of sample loss due to moisture condensation in the sampling
apparatus.
43 PRECISION, ACCURACY, AND LIMIT OF DETECTION OF THE METHOD
Precision of analytical procedures is quantified based on duplicate sample
analysis.* All EPA's Method 18 evaluation studies have demonstrated a relative standard
deviation of less than 5 percent for the analytical precision which is required by
Method 18.
Accuracy of sampling and analysis is quantified based on the required analysis of
two audit gas cylinders obtained through the EPA's audit sample repository. Field
validations of Method 18 conducted by EPA and numerous compliance test audits have
demonstrated, that when proper procedures and checks are conducted, Method 18 is
accurate within 10 percent of the true value (as required by the method).
The limit of detection, of Method 18 is typically about 1 part per million by
volume (ppmv) for most organic compounds. The actual limit of detection will vary for
each organic compound and type of detector and is defined as the minimum detectable
concentration of that compound, or the concentration that produces a signal-to-noise
ratio of 3:1.
4.4 MEASUREMENT AND LOCATION OF SAMPLING POINTS
The agency representative must determine if the final results need to be presented
on a concentration basis or a mass emission basis. For data presentation on a
concentration basis, only the concentrations of the specified organics and the stack gas
moisture content must be measured. If the mass emission rate of any compound is to be
presented, the flow rate of the stack gas must also be determined using the velocity
traverse. The number and locations velocity traverse points are selected according to
Method 1; the traverse is conducted according to Methods 2, 2A, 2C, or 2D, as
applicable. Although Method 18 requires sampling at a single point, it may be necessary
to perform a velocity traverse to obtain the emission data in the units of the standard.
Method 18 requires that samples are collected proportionally, meaning that the
sampling rate must be kept proportional to the stack gas velocity at the sampling point
during the sampling period. If the process has a steady state flow (constant), then the
4-2
-------
flow rate does not have to be varied during sampling. The majority of sources of organic
emissions are of this type because they use constant rate fans. If the testing firm can
confirm that the emission source of interest has a steady state flow (e.g., it uses a
constant rate fan), then sampling can be conducted at a constant rate and no concurrent
velocity measurements need to be made. If it cannot be determined whether the process
is steady state, .then velocity measurements (based on the velocity head, AP) must be
made at the point to be sampled. The average velocity head (phot reading, AP) and
range of fluctuation is determined and then utilized to establish the proper flow rate
settings during sampling.
4.5 OBSERVATION PROCEDURES FOR METHOD 18 TESTING
As previously mentioned, one of the primary concerns for any organic sampling
program must be safety. It is beyond the scope of this manual to discuss safety aspects
of organic sampling and analysis. However, it cannot be over-emphasized that the test
coordinator must always be aware of the safety hazards. The major two hazards are
explosion and health effects. For all source performance tests, there may be hazards in
gaining access to the test location and in moving about the facility to complete the
required activities in observing the tests.
4.5.1 Selection of Proper Sampling and Analytical Technique
• -t
Because of the number of different combinations of sampling procedures, sample
preparation procedures, calibration procedures, GC column packing materials, GC
operating conditions, and GC detectors covered under this method, a set of tables has
been developed to assist the tester in selecting (and the test coordinator in evaluating)
acceptable sampling and analytical techniques. The organic compounds included in
these tables were selected based on their current status as either presently regulated or
being evaluated for future regulation by the EPA and State and local agencies. Table
C-l, page C-2, provides the user with the following information for the selected
compounds: (1) the International Union of Pure and Applied Chemistry (IUPAC) name,
any synonyms, the chemical formula, the Chemical Abstracts Service (CAS) number; (2)
method, classification and corresponding references for more information; and (3)
information on whether EPA currently has an audit cylinder for this compound. Also,
detailed discussions on how to select the proper sampling and analytical techniques are
presented in Appendix C.
4.52 Preliminary Measurements and Setup
Method 18 recommends that a pretest survey and/or laboratory evaluation be
conducted by the testing firm prior to sampling and analysis. The pretest survey
measurements are needed to properly design the emission test protocol. The primary
objective of the preliminary survey is to collect a pretest survey sample for
(1) determining which sampling procedure is most appropriate and (2) optimizing the
4-3
-------
analytical procedures. Using the pretest survey sample, estimates of the emission
concentration(s) are made and the organic compounds in the gas stream are identified.
Also, any compounds that may interfere with the quantification of the target analyte(s)
are identified and appropriate modifications can be made to the analytical procedures.
Table C-l, page C-2, provides data on availability and ranges of EPA audit gases
for the target organic compounds. Further information can be obtained by contacting
EPA.
For audit gases obtained from a commercial gas manufacturer, check that the
manufacturer has (1) certified the gas in a manner similar to the procedure described in
40 CFR Part 61, Appendix B, Method 106, Section 5.2.3.1 and (2) obtained an
independent analysis of the audit cylinder that verifies that the audit gas concentration is
within 5; percent of the manufacturer's stated concentration.
4.53 Observation of On-site Testing
The on-site observation techniques and reference tables provided in Appendix C
should assist the test coordinator in determining if the testing firm has selected an
acceptable sampling and analysis technique. Data quality should be enhanced if the
testing firm conducts the recommended quality assurance/control checks and procedures
provided in this chapter and Appendix C. At some facilities, the testing firm may need
to use two or more sampling trains (different sampling techniques) to accurately measure
all the organic compounds of interest.
Because of the large number of approaches to the three different sampling
techniques (container sampling, adsorption tube sampling, and direct interface sampling),
only the most commonly used (evaluated container or Tedlar bag sampling) is discussed
in this chapter. The other sampling approaches are addressed in Appendix C in the
locations shown below. The test coordinator can use listing below to locate and review
the sampling approach of interest.
Chapter,,. Sampling Approaches
4.5.3
A
B
C
D
E
F
G
Evacuated Container Sampling
Sampling System Preparation
Proportional Sampling
Indirect Pumping Bag Sampling
Sample Recovery and Transport to Laboratory
Common Problems
Stability Check
Retention Check
Page
4-5
4-6
4-6
4-7
4-9
4-9
4-9
4-9
4-4
-------
Appendix Sampling Procedures Page
C2.1 Evacuated Container Sampling (heated and unheated) C-15
H Direct Pumping Bag Sampling C-22
I Explosion Risk Area Bag Sampling C-23
J Prefilled Bag Sampling C-23
K Heated Syringe Sampling C-26
C22 Direct Interface Sampling C-27
C23 Dilution Interface Sampling C-30
C2A Adsorption Tube Sampling C-32
Evacuated Container Sampling (Heated and Unheated) - In this sampling
technique, sample bags are filled by evacuating rigid air-tight containers that hold them.
The suitability of the bags for sampling is confirmed by permeation and retention checks
using the specific organic compounds of interest during the pretest survey
operations. The permeation and retention checks (discussed later) must be performed
on the field samples to ensure that the container sampling technique is acceptable.
On-site sampling includes the following steps:
1. Conducting preliminary measurements and setup. ~
2. Preparing and setup of sampling system.
3. Preparing of the probe.
4. Connecting electrical service and conducting leak check of sampling system.
5. Inserting probe into duct and sealing port.
6. Purging sampling system. '- -
7. Conducting proportional sampling.
8. Recording data.
9. Recovering sample and transporting to laboratory.
The "On-site Checklist" (Figure C.2, page C-17) includes checks for each of the
steps above and can be used by the test coordinator as an instructional guide. To assist
the test coordinator, the most critical items are printed in bold lettering.
It is the responsibility of the testing firm to ensure that the sampling and
analytical procedures are performed correctly. The following detailed information is
given only as training guide for the less experienced test coordinator and does not imply
mandatory actions by the test coordinator except when the discussions state that the test
coordinator "shall" conduct a given procedure.
Method 18 requires that samples be collected proportionally, meaning that the
sampling rate must be kept proportional to the stack gas velocity at the sampling point
during the sampling period. If the process has a steady state flow (constant), then the
flow rate does not have to be varied during sampling. The average velocity head (pitot
4-5
-------
reading) and range of fluctuation is determined and then utilized to establish the proper
flow rate settings during sampling. If it is found that the process is not steady state, then
the velocity head must be monitored during sampling to maintain a constant proportion
between the sample flow rate and the flow rate in the duct.
A total sampling time greater than or equal to the minimum total sampling time
specified in the applicable emission standard must be selected. The number of minutes
between readings while sampling should be recorded as an integer. It is desirable for the
time between readings to be such that the flow rate does not change more than 20
percent during this period.
If the sampling system must be heated during sample collection and analysis, the
system temperature should not decrease below the specified temperature. The average
stack temperature is used as the reference temperature for initial heating of the system
and should be determined. Then, the stack temperature at the sampling point is
measured and recorded during sampling to adjust the heating system just above the stack
temperature or above the dew point. The use of a heated sampling system requires on-
site analysis.
A. Sampling System Preparation - The test coordinator should observe the
preparation of the probe and sampling train in the laboratory area (&ee~Figure 4.1, page
4-16). The sampling apparatus must meet the following criteria (on-site measurements
checklist, Figure C.2, page C-17):
1. The probe must be constructed of stainless steel, glass or Teflon.
2. All connections must be either stainless steel or Teflon.
3. The probe, if required, must be capable of keeping the stack gases at or
just above the stack temperature.
4. The sample line must be Teflon.
5. The sample bags must be Tedlar or Teflon, leak checked and blank
checked.
6; A permeation check and retention check should have been conducted prior
to testing. If these checks have not been made, then they should be
conducted on the field samples.
7. The flowmeter must be calibrated, be in the proper range, and heated, if
applicable.
8. If located between the probe and bag, the pump must be of the Teflon
coated diaphragm type and heated, if applicable. If the pump is after the
rigid container, it may be any type of pump that provides the proper flow
rate.
9. If a dilution system is used and the probe must be heated, the dilution
system must be in a heated box.
4-6
-------
B. Proportional Sampling - Sampling must be conducted at a rate in constant
proportion to the stack gas flow at the sampling point Thus, for a steady state operation,
the sampling flow rate is not varied during the run. For a nonsteady state process, the
sampling flow rate is varied in proportion to the changing velocity. The velocity is
monitored by measuring the velocity head (AP) which is linearly related to the square of
the velocity. A recommended method for determining proportional sampling rates is as
follows:
1. Conduct a single point velocity check as previously specified, and deter-
mine the average velocity head (AP^g) to be sampled.
2. The average sampling flow rate for the test is .determined prior to the start
of the run. Typically, the average sampling flow rate is about 0.5 1/min
yielding approximately 30 liters of sample. The flow rate chosen in the
laboratory should fill the bag to about three-fourths of its capacity during
the sample run. The average flow rate chosen is then assigned to the
average velocity head measured.
3. The flow rate to be used during sampling when the velocity head varies
from the average is calculated using the following equation.
AP* Equation 4-1
where:
Qm = Average sampling rate, 1/min (r^/min),
Qs = Calculated sampling rate, 1/min (rrYmin),
AP = Actual velocity head, mm (in.) H2O, and
= Average velocity head, mm (in.) H2O.
4. Determine the rotameter setting for the sampling rate (Q5) from the rota-
meter calibration curve, and adjust the rotameter accordingly.
C. Tedlar Bag Sampling Procedures using an Indirect Pumping Technique - Use
of proportional sampling will provide for the correct sampling rate and the proper filling
of the sample bag. The tester should follow the procedure below to obtain an integrated
sample when the pump is located after a rigid container (Figure 4.1, page 4-16).
1. If a-heating system is required, turn on the heating system and set
container temperature at the average stack temperature determined from
the pretest measurements. If probe heating is required, then bag heating
would also likely be required.
2. Leak check the sampling train just prior to sampling by connecting a U-
tube, inclined manometer, or equivalent at the probe inlet and pulling a
4-7
-------
vacuum of a 10 in. H2O. Close the needle valve and then turn the pump
off. The vacuum should remain stable for at least 30 seconds. If a leak is
found, repair before proceeding; if not, slowly release the vacuum gauge.
This leak check is optional. The most critical place for a leak is between
the probe and the bag. Air inleakage into the sample bag will produce low
results. If the rigid sample container develops a small leak which only
effects the pumping rate, this should not cause a significant bias in the
results.
3. If the system is being heated, wait for it to come to the proper tempe-
rature. Place the probe in the stack at the sampling point: centroid of the
stack or no closer to the walls than 1 meter. Seal the sampling port to
prevent dilution of the stack gas by inleakage of ambient air. It can be
*• important that the port is sealed to prevent air inleakage. Also, if the
'•'' system must be heated, a significant loss of organics can result from poor
heating. It is better to heat the system above than below the specified
temperature.
4. Disconnect the flexible bag. Purge the system by turning on the pump and
drawing at least 5 times the sampling system volume through the train, or
purge for 10 minutes, whichever is greater. The system is purged to
equilibrate and remove ambient air. If the system is not purged, then a
negative error may be introduced.
5. Adjust the flow rate to the proper setting based on the velocity pressure
(measured during the purging, for nonsteady state processes). Proportional
sampling should be conducted for nonsteady state systems, but constant
rate sampling generally will not cause a significant bias unless the
concentration and flow rate are changing significantly.
6. Connect the flexible bag to the sampling train (the connections should
ensure a leakfree system), and begin sampling. The sampling rate must
remain proportional to the stack gas velocity for the total sampling time
specified by the applicable standard. Although Method 18 recommends a
rate of about 1 1pm be used, slower sampling rates and smaller sample
bags have been shown to be as accurate and precise.
Record all data required (at 5 minute intervals, minimum) on the field
H sampling data form similar to Figure 4.2, page 4-17. The flow rate and
sampling train heating system should be adjusted after every pitot and
temperature reading to the correct level. A shorter sampling time for each
point is typically used when the flow rate is changing significantly. For
emission sources with small changes in the flow rate, the sampling time
per point may be longer.
8. Disconnect and seal the flexible bag upon completion of sampling. Take
care not to dilute the contents with ambient air. Hie bag should be sealed
well to prevent leakage.
-
7.v
4-8
-------
9. Label eacb bag clearly and uniquely to identify it with its corresponding
data form and/or run. If the sampling system was heated, the sample bag
must be maintained at the stack temperature through sample analysis.
D. Sample Recovery and Transport to Laboratory - Sample recovery should be
performed so as to prevent contamination of the bag sample and maintain sample
integrity. The bag should remain leakfree, be protected from direct sunlight, be main-
tained at a temperature that will prevent condensation of any of the gases, and be stored
in a safe place to prevent damage or tampering prior to analysis. It is recommended
that bag samples be analyzed within two hours of sample collection; however, many of
the organic compounds are stable enough to allow a few days prior to analysis. Upon
completion of the testing and sample recovery, all the data forms should be checked for
completeness and the sample bags re-examined for proper identification. It is important
to check the bags for problems with permeation and retention of the sample.
E. Common Problems - The most common problems encountered with bag
sampling techniques are (1) adsorption of the gases on the bag, (2) permeation of the
gases through the bag, (3) reaction of gases in the bag, (4) condensation of the gases or
water vapor in the bag, and (5) leaks developing in the bag during testing, transport,
and/or analysis. The bags must be checked for stability and retention of the target
compound.
F. Stability Check - To assess the stability of the gas sample in Tedlar bagSj
perform a second analysis after a time period equalling the period between sample
collection and the first analysis. If the concentration of the sample collected in a Tedlar
bag decreases by more than 10 percent between the first and second analyses, then an
accepted sampling method other than Tedlar bags should be considered.
G. Retention Check - Perform a retention check on the bag sample by
successively evacuating the bag and refilling it with hydrocarbon-free air or nitrogen one
or more times. Analyze the bag contents for the target compound(s), then allow the gas
to sit in the bag overnight and reanalyze bag contents for the target compound(s). If any
target compound is detected in the bag at a concentration greater than 5 percent of the
original concentration, then an accepted sampling method other than Tedlar bags
should be considered.
One technique that can be used to reduce both retention and/or condensation in
the bag is addition of a heating system. Heating is generally applied during sample
collection and maintained through analysis. However, heating may increase the
permeation rate. Another option is the use of heat lamps applied to the sample bags
after sample collection and during sample analysis. Two other techniques that have been
used to prevent condensation are (1) addition of a knockout trap to remove water vapor
and heavy organics from the sample stream, and (2) use of sorbents such as Tenax to
remove the high boiling point organics. In these cases, the testing firm must demonstrate
4-9
-------
that the organic compound(s) of interest are not removed. Alternatively, sample and/or
water vapor condensation may be reduced by the use of the prefilled bag technique. The
prefilling of the bag lowers the concentrations of the organic and/or water vapor, thereby
eliminating condensation.
If gases are reacting in the bag, then the bag material can be changed, the time
between sample collection and analysis reduced, or a different sampling technique used
such as direct interface sampling. Methods to reduce bag leak problems are proper
construction of the sample bags, conducting additional runs, using a backup sample
collection technique such as an another bag sampling system or an adsorption tube
sampling system, and care in handling the sampling bags. Also, steel canisters can be
used in place of bags. If the organic compounds are stable with time, the use of steel
canistersjmay improve preservation of the samples especially if they must be air freighted
to the laboratory for analysis. ;
4.6 OBSERVATION PROCEDURES FOR METHOD 18 ANALYSIS
Unless it is conducted on-site, the test coordinator will not typically have an
opportunity to observe the analysis. The test coordinator should therefore check the
data and other documentation in the compliance test report for:
1. Audit sample results within 10 percent of true value (see Chapter 4.7 for
more details).
2. Proper preparation of calibration standards.
. 3. Proper resolution of compounds.
4. Additional unidentified peaks.
5. Proper analytical precision.
6. Acceptable collection efficiency for adsorption tube sampling.
7. Acceptable desorption efficiency for adsorption tube sampling.
8. Proper calculation of analytical data results.
A:test coordinator's postsampling operations checklist (Figure CIO, page C-37) is
provided; to assist in the review of the procedures for on-site analysis and the compliance
test report for off-site analysis.
4.6.1 Preparation of Calibration Standards
Calibration standards are prepared prior to sample analysis following the
procedures described in the following chapters. Refer to Table C-4, page C-8, for
recommendations on the procedures suitable for selected compounds. Note that there
are two basic types of standards, gaseous or liquid; the type prepared depends on the
type of sample collected. Gaseous calibration standards are needed for analysis of
pretest survey samples collected in glass flasks or bags, and final samples collected in
bags, by direct, or by dilution interface sampling. There are three techniques for
4-10
-------
preparing gaseous standards, depending on availability and the chemical characteristics
of the standard compound(s); gas cylinder standards may also be used directly, if the
proper concentration ranges are available. Liquid calibration standards are required for
the analysis of adsorption tube samples from the pretest survey and/or the final
sampling, as well as to determine the desorption efficiency; there are two techniques for
preparing liquid calibration standards. The concentrations of the calibration standards
should bracket the expected concentrations of the target compound(s) in the emissions
being sampled. Specific procedures for preparing and analyzing each type of standard
are described in Appendix C.3.
4.6.2 Sample Analysis
After the GC has been calibrated and the analysis of the audit sample(s) has been
conducted successfully, the samples are analyzed. The following are the key procedures
used to analyze emission samples collected in Tedlar bags using a GC calibrated with
gaseous calibration standards. The analytical procedures for adsorption tube sampling
and direct interface sampling are presented in Appendix C.3.
1. Note the time of injection on the strip chart recorder and/or actuate the
electronic integrator. Record the sample identity, detector attenuation
factor, chart speed, sample loop temperature, column temperature and
identity, and the carrier gas type and flow rate on a data form. It is also
recommended that the same information be recorded directly on the
chromatogram and on an analytical data sheet similar to Figure 4.3, page
4-18. Record the operating parameters for the particular detector being
used.
2. Examine the chromatogram to ensure that adequate resolution is being
achieved for the major components of the sample. If adequate resolution
is not being achieved, vary the GC conditions until resolution is achieved,
and reanalyze the standards to recalibrate the GC at the new conditions.
3. After conducting the analysis with acceptable peak resolution, determine
the retention time of the sample components and compare them to the
retention times for the standard compounds. To qualitatively identify an
individual sample component as a target compound, the retention time for
the component must match within 0.5 seconds or 1 percent, whichever is
greater, of the retention time of the target compound determined with the
calibration standards.
4. Repeat injection of the first sample until the area counts for each
identified target compound from two consecutive injections are within
5 percent of their average.
5. Multiply the average area count of the consecutive injections by the attenu-
ation factor to get the area value for that sample, and record the area
value on the data form.
4-11
-------
6. Immediately following the analysis of the last sample, reanalyze the cali-
bration- standards, and compare the area values for each standard to the
corresponding area values from the first calibration analysis. If the
individual area values are within 5 percent of their mean value, use the
mean values to generate a final calibration curve for determining the
sample concentrations. If the individual values are not within 5 percent of
their mean values, generate a calibration curve using the results of the
second analysis of the calibration standards, and report the sample results
compared to both standard curves.
Determine the bag sample moisture content by measuring the temperature and
the barometric pressure near the bag. Use water saturation vapor pressure chart, assum-
ing the .relative humidity of the bag to be 100 percent unless a lower value is known, to
determine the water vapor content as a decimal figure (percent divided by 100). If the
bag has been heated during sampling and analysis, the flue gas or duct moisture content
should be determined using Method 4.
4.7 USE OF AUDIT MATERIALS AND INTERPRETATION OF DATA
An audit is an independent assessment of data quality. Based on the
requirements of Method 18 and the results of collaborative testing of other EPA
Methods, two specific performance audits are recommended:
1. An audit of the sampling and analysis procedures of Method 18 is required
under NSPS and recommended for other purposes.
2. An audit of the data processing is recommended.
A systems audit may be conducted by the test coordinator in addition to these
performance audits. Performance audits are described in detail in Chapter 4.7.1 and the
systems audit is explained in Chapter 4.7.2.
4.7.1 Performance Audits
r , •
-* -
Performance audits are conducted to evaluate quantitatively the quality of data
. produced by the total measurement system (sample collection, sample analysis, and data
processing). It is required that cylinder gas performance audits be performed once
during every NSPS compliance test utilizing Method 18, and it is recommended that a
cylinder gas audit be performed once during any compliance test utilizing Method 18
conducted under regulations other than NSPS.
Performance Audit of the Field Test - As stated in Section 6.5 of 40 CFR 60,
Appendix A, Method 18, immediately after the preparation of the calibration curves and
prior to the sample analysis, the analysis audit described in 40 CFR 61, Appendix C,
Procedure 2: "Procedure for Field Auditing GC Analysis," should be performed. The
4-12
-------
information required to document the analysis of the audit sample(s) has been included
on the example data sheet shown in Figure 4.4, page 4-19. The audit analyses shall
agree within 10 percent (or other specified value, as explained below) of the true values.
The test coordinator may obtain audit cylinders by contacting: U.S. Environmental
Protection Agency, Atmospheric Research and Exposure Assessment Laboratory, Quality
Assurance and Technical Support Division, Research Triangle Park, North Carolina
27711. Audit cylinders obtained from a commercial gas manufacturer may be used
provided that (1) the gas manufacturer certifies the audit cylinder in a manner similar to
the procedure described in 40 CFR 61, Appendix B, Method 106, Section 5.2.3.1, and
(2) the gas manufacturer obtains an independent analysis. Independent analysis is
defined as an analysis performed by an individual other than the individual who performs
the gas manufacturer's analysis, using calibration standards and analysis equipment dif-
ferent from those used for the gas manufacturer's analysis. Verification is completed and
acceptable when the independent analysis concentration is within 5 percent of die gas
manufacturer's concentration.
Responsibilities of the Test Coordinator - The primary responsibilities of the test
coordinator are to ensure that the proper audit gas cylinder(s) are ordered and safe-
guarded, and to interpret the results obtained by the analyst.
When auditing sampling systems that do not dilute the stack gas during sampling,
the audit gases ordered must consist of the same organic compound(s) that are being
measured; for emission standards on a concentration basis, the audit gas concentration^)
must be in the range of 25 percent to 250 percent of the applicable standard. The audit
should include analysis of two concentration levels. If two cylinders are not available,
then one cylinder can be used. It is strongly recommended that audit cylinder values
below 5 ppm not be used. For emission standards which specify a control efficiency, the
concentration of the audit gases should be in the range of 25 percent to 250 percent of
the expected stack gas concentration. The audit should include analysis of two
concentration levels. If two cylinders are not available, the audit can be conducted using
one cylinder.
The test coordinator must ensure that the audit gas cylinder(s) are shipped to the
correct address, and to prevent vandalism, verify that they are stored in a safe location
both before and after the audit. Audit cylinders should not be analyzed when their
pressure drops below 200 psi because the cylinder gas value may be unreliable.
The audit results must agree within 10 percent of the stated audit cylinder value
or true value. Agreement within 15 percent is allowed for cylinders between 5 and
20 ppm. When the measured value agrees within these limits, the test coordinator
directs the analyst to begin analyzing the field samples. For on-site analysis, when the
measured concentration does not agree, the analyst should first recheck the analytical
system and calculations, and then repeat the audit. When the results of the repeat audit
are within the limits, the analyst may conduct the field sample analysis. If the analyst
4-13
-------
li
fails the second audit, the agency may reject the compliance test results. Method 18
states "Audit supervisor judgement and/or supervisor policy determine action when
agreement in not within ± 10 percent." "When a consistent bias in excess of 10 percent is
found, it may be possible to proceed with the sample analysis, with a corrective factor to
be applied to the results at a later time." The test coordinator should therefore know
the policy of the agency related to audit failure.
During the audit, the test coordinator should record the appropriate cylinder
number(s), cylinder pressure(s) (at the end of the audit), and the calculated concen-
trations on the "Field audit report form,11 Figure 4.4, page 4-19. The individual being
audited must not, under any circumstances, be told the actual audit concentrations) until
the calculated concentration(s) have been submitted to the test coordinator and are
considered acceptable.
When auditing sampling systems that dilute the emissions during collection, the
audit gas concentration value used in the calculations can either be based on (1) the
undiluted concentration using the criteria discussed above or (2) the expected
concentration of the gases following dilution during collection using the same dilution
factor as used for the emission samples.
The audit procedures that follow are used for the evacuated container sampling
approach with either on-site or off-site analysis. Auditing procedures for the other
sampling techniques are presented in Appendix C according to the sampling approach
used to collect the organic emissions and whether the samples are analyzed on-site or
off-site.
Container (Bag, Syringe, and Canister) Sampling with On-site Analysis - The
cylinder gas performance audit for bag, syringe, or canister sampling with on-site analysis
is conducted on-site just prior to the analysis of the field samples. The recommended
procedures for conducting the audit are:
1: - The audit samples should be collected in the type of container used during
*» sample collection. However, to conserve audit gas, it is usually not
V necessary to involve the rest of the sampling system in audit sample
collection for unheated container sampling. Problems related to the
reaction or retention of the organic compounds will occur in the container.
Interferents in the stack gas such as water vapor and other organics are not
present in the audit cylinders and thus, related problems are not assessed
by the audit. For heated container systems, it may be necessary to use the
whole sampling system to collect the audit gas. However, if a gas must be
heated to prevent its condensation in the sampling system, it is likely that
audit gas cylinders are not available for this compound or level of
compound.
4-14
-------
2. Prior to analysis, the audit samples should remain in the appropriate
container approximately the same length of time as the field samples.
After the preparation of the calibration curve, a minimum of two
consecutive analyses of each audit cylinder gas should be conducted. The
analyses must agree within 5 percent of the average. The audit results
should be calculated by the analyst (or his representative) and given to the
test coordinator. The test coordinator will record all the information and
data on the "Field Audit Report Form" and then inform the analyst of the
status of the audit. The equations for calculation of error are included on
the form.
Container (Bag and Canister) Sampling with Off-site Analysis - For cylinder gas
performance audits associated with container samples analyzed off-site, it is
recommended that the audit be conducted off-site just prior to the compliance test (if
the agency desires) and then repeated during the off-site sample analysis as a quality
control measure. The use of the pretest audit will help ensure that the analytical system
will be acceptable prior to testing. Alternatively, the audit gases can be collected in the
appropriate containers on-site or off-site, and then analyzed just prior to the field sam-
ples analysis. It is recommended that the tester fill at least two containers with the audit
gas to guard against a container leak causing a failed audit. Since the use of the
performance audit is to both assess and improve the data quality, the use of .the pretest
audit will provide the tester/analyst with a better chance of obtaining acceptable data.
The recommended procedure for conducting the audit is the same as described above for
the on-site audit with the exception that the test coordinator will likely not be present
and the data will have to be reported by telephone.
Performance Audit of Data Processing - Calculation errors are prevalent in
processing data. Data processing errors can be determined by auditing the recorded data
on the field and laboratory forms. The original and audit (check) calculations should
agree within round-off error; if not, all of the remaining data should be checked. The
data processing may also be audited by providing the testing laboratory with specific data
sets (exactly as would appear in the field), and by requesting that the data calculation be
completed and that the results be returned to the agency. This audit is useful in
checking both computer programs and manual methods of data processing.
4.72 Systems Audit
A systems audit involves checking to ensure that the proper equipment and
procedures are used. The observation of the sampling and analytical procedures by the
test coordinator described in Chapters 4.5 and 4.6 constitutes a systems audit for Method
18. The systems audit results may be recorded on the sampling and analytical checklists
referenced in Chapters 4.5 and 4.6.
4-15
-------
"I
Q.
01
g>
"E.
2
so
•«*-
±
.§?
4-16
-------
>> c
u
4->
ro
0) E m U r—
«rt O) X O)
IF- O tflt— U C
"c'
e o
o-.-
c « =r
c o •«-»—
3 +J r—
Q Lw:
CJ4J
c o
.—. en t- c
"- 0>
F- O.
U 4-1
s_ u
(JJ Q)
«
*->
o c
^ o
E 0)
= -§
O ~ •
> e i
u s- •
m QJ
re re
c c u
o o >
^^ (^ '
iS
O> rB
.Q —
E-0
* 3
QJ
a. . _
o Q o:
0)
u
re
o
rg
-»
e
•^
•*-*
_
J
p
£
a
\>
»»
p
•
J
«
?
t
J
>
r
t
»
o
•r*
^
ro
U
O
c
o
c.
o
)
*Q.
tO
5
1
•*>
»
•»
M.
^^
u.
o
^M>
tJ
O
C
4«J
OJ
I/)
(JJ
*2C
^-,-*
C
*t«-
"N^gp*
E
*- t
QJ = l/j
«M Q/
03 *"*
£ '
^ «P •&
jQ W
^ *.>
" (/>
}*U ^
**^ frt'
If
O)
-a «_H
E ^
5 O
C'—'
^3 M^
^*. ^B
•B W
0)
1— 4->
c. o
E *J
«B ••-
oo e.
M
W
U
Q.
.U
C
J. _1
•*•'
01
£
O
(0
ca
o
cvj
•~*
.5
W
£!
^
•pv
*J
19
(/>
fl) <~»
ell
fete
15 •—
^T f. ^
Oo
U
X
o
^
S-LL.
V> 0)0
C V
•5 Se
I- U.
V 0)
fc. C
4-1 F^LW
r9 e
QJ F—
CL ^ g_*
E Ee
Ol (Q
I- v>
~~.
%0~
O *~*
O.U
e
J£ LL.
OO
ra •— •
4^
O
o
«IP.^
C
•*^
E
k *^
OJ cnro
ft! -F- li_
5 *J *—'
O OJ C
F~ irt ••-
LU E
73 •
ra O
O) CVJ
— C fc"
LC.^-
Si
^^ *
0 01 J=
0 E
C_> 4-1 CM
C
F^ oT e
0. E —
E— E
ra 4->
vo
r
,
o>
^
W
>
0*
^
*c
fMBk
<
O)
3^
<
D)
CD
^^
••v
ro
0
19
+*
a
o
0)
(VI
03
4-17
-------
Date:
Location:
Analyst:
Type of Calibration Standard:
Number of Standards: Date Prepared:
Plant:
Sample Type:
1
Target Compound:
Prepared By:
Column Used:
GC Used:
Carrier Gas Used: ___
Column Temperatures, Initial: ___^__
Sample Loop Volume: Loop Temperature:
Detector Temp.: Auxiliary Gases:
Carrier Gas Flow Rate:
Program Rate:
Final:
Inject. Port Temp.
Calibration Data Standard 1
First analysis/second analysis
Standard concentration (C^,)
Flow rate through loop(ml/min) /
Liquid injection volume (tubes) /
Injection time (24-hr clock)
Chart- speed (cm/min) /
Detector attenuation /
Peak retention time (min) I
Peak 'retention time range (min)
Peak area /
Peak area x attenuation factor /
Average peak area value (Y)
Percent deviation from average
Calculated concentration (CM)
% deviation from actual (%DM)
Standard 2
standard 3
Linear regression equation; slope (m):
y-intercept
Sample Analysis Data
First analysis/second analysis
Sample identification
Interface dilution factor
Flow rate through loop (ml/min)
Liquid injection volume (tubes)
Injection time (24-hr clock)
Chart speed (cm/min)
Detector attenuation
Peak retention time
Peak retention time range
Peak area
Peak.area x atten. factor
Samle 1
Sample 2
Sample 3
% deviation from average
Calculated concentration
c«d
or C, -
(Y - b)
m
(%D,
(C.)
-J
At - Y
Y
Xi no*: *n — ... ,. V i nni
Figure 43. Data form for analysis of Method 18 field samples.
4-18
-------
Part A. - To be filled out using information from organization
supplying audit cylinders.
1. Organization supplying audit sample(s) and shipping address
2. Audit supervisor, organization, and phone number
3. Shipping instructions: Name, Address, Attention
4. Guaranteed arrival date for cylinders -
5. Planned shipping date for cylinders -
6. Details on audit cylinders from last analysis
a. Date of last analysis
d. Audit gas (es) /balance gas*
e. Audit gas(es) > ppm
f. Cylinder construction
Low cone.
High cone.
Part B. - To be filled out for audit analysis.
1. Process sampled
2. Audit location
3
4,
5. Audit Results:
Name of individual audit
Audit date
d. Measured concentration, ppm
e. Actual audit concentration, ppm "
f . Audit accuracy : *
Percent accuracy* =
Measured Cone. - Actual Cone, x 100
Actual Cone.
Low
cylinder
High
cylinder
^Results of two consecutive injections that meet the criteria.
Figure 4.4. Field audit report form.
4-19
-------
-------
CHAPTERS
DETERMINATION OF VOLATILE ORGANIC
COMPOUND LEAKS - METHOD 21*
5.1 APPLICABILITY
Method 21 applies to the determination of volatile organic compound (VOC)
leaks from process equipment that is in VOC service. In VOC service includes any
fugitive emission source that contains or contacts a fluid composed of equal to or greater
than 10 percent VOC by weight. For the benzene fugitive emission regulation, in
benzene service includes any source that contains or contacts a fluid equal to or greater
than 10 percent benzene by weight.
VOC service can further be divided into light liquid or heavy liquid service. Light
liquid VOC service is defined as one or more of the stream components having a vapor
pressure greater than 0.3 Kpa (0.04 psia) at 20° (68°F). All VOC sources with a stream
component vapor pressure equal to or less than 0.3 Kpa at 20°C are in heavy liquid
service. The NSPS (New Source Performance Standard) for "Refinery Leaks" defines
heavy liquid as kerosene or heavier liquid.
Leaks are classified as fugitive emissions. Sources of fugitive emissions include,
but are not limited to, valves, flanges and other connections, pumps and compressors,
pressure relief devices, process drains, open-ended valves, pump and compressor sealing
systems, degassing vents, accumulator vessel vents, agitator seals, and access door seals.
5.2 METHOD DESCRIPTION
A portable instrument is used to detect VOC leaks from individual sources. The
instrument detector type is not specified, but it must meet certain specifications and
performance criteria contained in EPA Method 21, Section 3. This procedure is
intended to locate and classify leaks only, and is not to be used as a direct measure of
mass emission rates from individual sources.
5.2.1 Regulations and Leak Definition
Industries that emit fugitive VOCs and are affected by federal regulations are
shown in Table 5-1. The sources of fugitive emissions, methods by which emissions are
detected and repaired, and control procedures are very similar for each of these
industries.
"The majority of this Chapter is taken directly from EPA-340/1-86-015, "Portable
Instruments User's Manual for Monitoring VOC Sources." To reduce the references to
the cited manual, reference markings have been left out.
5-1
-------
TABLE 5-1. SOURCE CATEGORIES THAT EMIT FUGITIVE VOCs
Source Category
Type of Control Guidance*
Petroleum refineries
Synthetic organic chemicals
manufacturing industry
Polymers and resin manufacturing
Natural gas and natural gasoline
processing plants
Benzene in coke ovens/by-products
plants
Vinyl-_chloride sources
Benzene fugitive sources
CTG, NSPS
CTG, NSPS
CTG, NSPS
CTG, NSPS
NESHAPs
NESHAPs
NESHAPs
CTG: Control Technique Guideline
NSPS: New Source Performance Standard
NESHAPs: National Emission Standard for Hazardous Air Pollutants
A leak definition can be based on either a concentration value or a "no detectable
emissions." The "no detectable emission" standard is applied to sources designed to
operate in a leakless manner, such as pressure-relief devices with rupture or sealless
pumps. The concentration-based value most often used to define a leak is a
concentration equal to or greater than 10,000 ppmv. The "no detectable emission"
standard is not an absolute zero reading. A violation of the "no detectable emission"
limit is defined in Method 21 as a concentration greater than five percent of the
concentration-based leak definition. For example, based on the 10,000 ppmv definition
of a leak, a concentration greater the 500 ppmv would be in violation of the "no
detectable emission" standard.
522 Portable Instrument Operating Principles
Various types of instruments are available for detecting organic vapors. These
operate on different principles. Each detector has its own advantages, disadvantages,
and sensitivity.
In addition to the portable VOC detectors, other portable equipment used during
Method 21 testing includes temperature sensors, flow monitors, and pressure gauges.
This equipment is much smaller, less expensive, and easier to use than the portable VOC
detectors.
5-2
-------
Several types of portable VOC detectors can be used either as screening tools or
to meet the requirements of EPA Method 21. These include:
* Flame ionization detector (FID)
• Photoionization (ultraviolet) detector (PDD)
• Catalytic combustion or hot wire detector
• Nondispersive infrared detector (NDIR)
The specifications of these instruments vary significantly with regard to sensitivity,
range, and responsiveness. Table 5-2 lists the most common instruments currently in use
and the associated detection principle, range, sensitivity, and response time.
Flame Ionization Detector - In an FID, the sample is introduced into a hydrogen
flame. A concentration of as little as 0.1 ppm of hydrocarbon produces measurable
ionization, which is a function of the number of carbon ions present. A positively
charged collector surrounds the flame, and the ion current between the flame and the
collector is measured electronically. Pure hydrogen burning in air produces very little
ionization, so background effects are essentially masked by the hydrogen flame. The
calibration output current is read on a panel meter or chart recorder.
Organic compounds containing nitrogen, oxygen, or halogen atoms give a reduced
response in a FID when compared to compounds without these atoms. The FID
hydrocarbon analyzers are usually calibrated in terms of a gas such as methane or
hexane, and the output is read in parts per million of carbon measured as methane or
hexane.
Although nitrogen (N2), carbon monoxide (CO), carbon dioxide (CO2), and water
vapor (H2O) do not produce significant interferences, condensed water vapor can block
the sample entry tube and cause erratic readings. Also, when the oxygen (O2)
concentration exceeds 4 percent, it can significantly reduce the detector output. The
relative response of the FID to various organic compounds, including those with attached
oxygen, chlorine, and nitrogen atoms, varies from compound to compound.
Photoionization Detector - In the photoionization detector, ultraviolet light ionizes
a molecule as follows: R + hv'> R+ + e", where R* is the ionized species and hv
represents a photon with energy less than or equal to the ionization potential of the
molecule: Generally, all species with an ionization potential less than the ionization
energy of the lamp are detected. Because the ionization potential of all major
components of air (03, N^ CO, CO& and H2O) is greater than the ionization energy of
the lamps in general use, they are not detected.
The detector consists of an argon-filled, ultraviolet (UV) light source that emits
photons. A chamber adjacent to the sensor contains a pair of electrodes. When a
positive potential is applied to one electrode, the field that is created drives any ions
formed by the absorption of UV light to the collector electrode, where the current
(proportional to the concentration) is measured.
5-3
-------
TABLE 5-2.
Instrument
Trade Names
550, 551, 555
(AID, Inc.)
OVA 108, 128
Centuiy'Systems
(Foxboro)
PI - 101
(Hnu Systems,
Inc.)
TLV Sniffer
(Bacharach)
Ecolyzer 400
(Energetics
Science)
Miran 1A
(Foxboro)
PORTABLE
Detection
Principle
FID
FID
pro
Catalytic
combustion
Catayltic
combustion
NDIR
INSTRUMEK
RESPONSE
Range, ppm
0-200
0-2,000
0-10,000
0-10
0-100
0-1,000
0-20
0-200
0-2,000
0-500
0-5,000
0-50,000
0-100%
LEL
ppm to %
ITS RANGE, SENSITIVE
TIME
Sensitivity
0.1 ppm at
0-200 ppm
0.2 ppm (Model 128)
0.5 ppm (Model 108)
1 ppm
2 ppm
1% LEL
1 ppm
Y, AND
Response
Times
5
2
2
5
15
1, 4, 10
and 40*
* Response times for different ranges.
5-4
-------
Nondispersive Infrared Detector - Nondispersive infrared (NDIR) spectrometry Is
a technique based on the broadband absorption characteristics of certain gases. Infrared
radiation is typically directed through two separate absorption cells: a reference cell and
a sample cell. The sealed reference cell is filled with nonabsorbing gas, such as nitrogen
or argon. The sample cell is physically identical to the reference cell and receives a
continuous stream of the gas being analyzed. When a particular hydrocarbon is present,
the IR absorption is proportional to the molecular concentration of that gas. The
detector consists of a double chamber separated by an impermeable diaphragm. Radiant
energy passing through the two absorption cells heats the two portions of the detector
chamber differentially. The pressure difference causes the diaphragm between the cells
in a capacitor to distend and vary. This variation in capacitance, which is proportional to
the concentration of the component of gas present, is measured electronically.
Interferences in NDIR measurements are usually a result of other gases in the
sample absorbing at the same wavelength as the gas of interest. Efforts to eliminate
these interferences by use of reference cells or optical filters are only partially successful.
For hydrocarbon (HC) monitoring, the detector is filled with one or several different
hydrocarbons, which may be different from the HC contained hi the sample; this causes
a disproportionate response. Other sources of errors include gas leaks in the detector
and reference cells, inaccurate zero and span gases, nonlinear response, and electronic
drift.
Catalytic Combustion or Hot Wire Detector - The heat of combustion of a gas is
sometimes used for quantitative detection of that gas. Suffering the same limitations as
thermal conductivity, this method is nonspecific, and satisfactory results depend on
sampling and measurement conditions.
One type of thermal combustion cell uses a resistance bridge containing arms that
are heated filaments. The combustible gas is ignited in a gas cell upon contact with a
heated filament; the resulting heat release changes the filament resistance, which is
measured and related to the gas concentration.
Another combustion method uses catalytic heated filaments or oxidation catalysts.
Filament temperature change or resistance is measured and related to gas
concentrations.
Thermocouple - The temperature sensor most commonly used is the direct-
readout hand-held thermocouple. The thermocouple is composed of two wires of
dissimilar metals that are joined at one end. When the joined end is heated, a voltage
flow can be observed. A voltmeter is attached to the thermocouple, and the observed
voltage is proportional to the measured temperature. A portable thermocouple assembly
consists of a shielded probe, a connecting wire, and a voltmeter. The voltmeter may be
a temperature conversion unit on a multimeter or a dedicated direct readout
temperature unit. The voltmeter is battery-operated, small, and easily portable.
5-5
-------
I
Static Pressure Gauges - Among the several different available static pressure
gauges, the most commonly used for this type of field work are the inclined manometer
and the diaphragm gauge. A pressure tap is necessary for use of a portable static
pressure gauge. Hie primary feature of the pressure tap is a small opening in the wall of
a duct, which can be fitted with a connection and a hose to make pressure
measurements. The tap should be far enough away from such disturbances as elbows
and internal obstructions to make the effects of such disturbances negligible.
> ... '.
The appropriate side, positive or negative, of the manometer or pressure gauge is
connected by a rubber hose at the tap, and a pressure reading can be taken. It is often
advantageous to disconnect a permanent pressure gauge and take a pressure reading at
that point to compare it with the facility's instrumentation.
S3 CALIBRATION PRECISION
Calibration precision is the degree of agreement between measurements of the
. same known value. To ensure that the readings obtained are repeatable, a calibration
precision test must be completed before placing the analyzer in service, and at 3 month
intervals, or at the next use, whichever is later. The calibration precision must be equal
to or less than 10 percent of the calibration gas value.
To perform the calibration precision test, .a total of three test runs are required.
Measurements are made by first introducing zero gas and adjusting the analyzer to zero.
The specified calibration gas (reference) is then introduced and the meter reading is
recorded. The average algebraic difference between" the meter reading and the known
value of the calibration gas is then computed. This average difference is then divided by
the known calibration value and multiplied by 100 to express the resulting calibration
precision in percentages.
53.1 Calibration of VOC Analyzers
Calibration requirements for VOC instrumentation are specified in Method 21
and in the specific NSPS applicable to sources of fugitive VOC emissions. The
requirements pertaining to calibration are briefly summarized here.
The instrument should be calibrated daily.
The gas concentrations used for calibration should be close to the leak
definition concentration.
The calibrant gas should be either methane or hexane.
A calibration precision test should be conducted every 3 months.
If gas blending is used to prepare gas standards, it should provide a known
concentration with an accuracy of ± 2 percent.
5-6
-------
O
O
O
The daily calibration requirement specified in Method 21 and in the various NSPS
gives individual instrument operators some flexibility. The calibration could consist of a
multipoint calibration in the lab, or it could be a single-point "span-check."
Neither Method 21 nor the applicable NSPS specifies where the calibration must
•take place. It is simpler to conduct the calibration in the laboratory than at the facility
being tested; however, there is the possibility that the calibration may shift sufficiently to
affect the accuracy of the leak detection measurements. The degree of shift has not yet
been documented for the various commercially available instruments. Because of the
potential for calibration shift, one should consider conducting, at a Tninirnpni, a single-
point span check after the instrument arrives on-site. It is also suggested that a span test
be run at the midpoint of the day and at the conclusion of the field work.
Although the span checks discussed above would in most cases qualify as the daily
calibrations required by the NSPS, a separate calibration test for organic vapor analyzers
should be conducted whenever possible. Because uniform day-to-day calibration gas
temperatures and calibration gas flow rates-can be maintained in the laboratory,
calibrations performed in the regulatory agency laboratory are conducted under more
controlled conditions than those conducted in the field. Furthermore, the initial
calibration test provides an excellent opportunity to confirm that the entire instrument
system is working properly before it is taken in the field. The laboratory calibration data
should be carefully recorded in the instrument calibration/maintenance notebook. This
calibration should be considered the official calibration required by the regulations.
533 Laboratory Calibrations
As specified in the EPA-promulgated NSPS, the instruments used in accordance
with Method 21 must be calibrated by using either methane or hexane at concentrations
that are close to the leak-detection limits. In most cases, the leak-detection limit is
10,000 ppmv, however, for certain sources, it is 500 ppmv above the background levels.
Methane-in-air is generally the preferred calibrant gas for the high concentration
range. A hexane-in-air concentration of 10,000 ppmv should not be prepared because it
is too close to the lower explosive limit. Also, hexane may condense on the calibration
bag surface at this high concentration. If hexane-in-air calibrations are necessary, the
chosen concentration should reflect a compromise between the need for adequate
calibration of leak detection levels and the practical safety and reproducibility problems
inherent in the use of hexane. The EPA's position is that the choice of calibrant gas
does not affect the ability of the instruments to detect fugitive leaks.
5-7
-------
When charcoal beds are used to provide the VOC-free air, a routine check should
be made to assess breakthrough of organic compounds. This is done by passing a low-
hydrocarbon-concentration gas stream (approximately 10 to 50 ppmv) through the bed
for a period of 5 to 10 minutes. If the bed has not become saturated, the outlet
hydrocarbon concentration should be low. Methane should not be used as the
hydrocarbon because charcoal is ineffective in adsorbing methane.
533 Field Span Check Procedure
The following are some of the various ways to calibrate the portable instrument in
'the field: .
• Use large pressurized gas cylinders transported to inspection site.
'"-•' Use certified gas cylinders provided by the facility being inspected.
; •'•'• Use disposable gas cylinders with the appropriate gas composition and
"""•"•• concentration.
• Use a gas sampling cylinder with a gas blending system.
Transporting large pressurized gas cylinders is generally impractical because most
agencies do not have the necessary vehicles. It is not safe to transport unsecured,
'pressurized gas cylinders in personal or State-owned cars. Furthermore, there are
specific Department of Transportation (DOT) regulations governing the shipping of
compressed gases.
Using the facility's gas cylinders is certainly the least expensive approach for a
regulatory agency; however, the appropriate gas cylinders are not always available. Also,
use of the facility's cylinders prevents the agency from making a completely independent
assessment of the VOC fugitive leaks and from evaluating the adequacy of the facility's
leak-detection program.
Using disposable cylinders of certified calibration gas mixtures is relatively simple
because no on-site blending is necessary and the cylinders are easily transported. The
calibration gas mixture may be fed to the instrument directly by using a preset regulator
that provides constant gas flow and pressure; or the gas can be fed into a Tedlar or
Teflon bag, from which it is drawn into the portable instrument
A fourth approach involves the use of a stainless steel gas sample cylinder with a
small Tedlar sample bag. A small quantity of calibration gas is drawn from a large
cylinder of certified gas mixture (at the agency's main laboratory) into the small
transportable gas sample cylinder. The calibration gas is kept at a relatively low pressure
to minimize safety problems during transport of the material to the job site. The
compressed gas is transferred to the Tedlar bag through a regulator and needle valve.
At a pressure of 325 psig, a 1 liter sample cylinder should provide enough span check gas
for two field checks. Zero air can be supplied by drawing ambient air through a small
5-10
-------
charcoal filter. This approach is very inexpensive because the agency is using small
quantities of the certified calibration gas mixture from the main cylinder at the
laboratory and they are not purchasing any disposable cylinders. Some additional
development work on this simple approach is necessary to ensure that a regulator is
available to transfer the gas from the main cylinder to the sample cylinder at pressures
reaching several hundred psig. Most regulators have a delivery pressure limit of
100 psig. It is also necessary to confirm that the compressed gas can be transferred
safely. It should be noted, however, that this is the same approach used to fill the
hydrogen fuel cylinders on the portable flame ionization analyzers.
The field span check should be performed at the greater distance possible from
potential sources of fugitive VOC. It should also be performed in an area where there is
nojarge AC motors or other equipment that generate strong electrical fields, as such
equipment can have an adverse effect on certain types of instruments (e.g.,
photoionization analyzers). The charcoal filter used in the "clean air" supply should be
routinely regenerated to avoid the possibility of saturation. It should be checked
occasionally for saturation by supplying a moderate, known concentration of VOC and
then checking the measured exit concentration after several minutes.
Data concerning the span checks should be recorded in the field notes. If gauges
are provided with the instrument, the tester also should occasionally note the instrument
sample gas flow rate.
53.4 Thermocouple
Thermocouples may be calibrated in several ways. The simplest method is
immersion in an ice bath and boiling distilled water. Electronic "ice point" reference
circuits are also commercially available to check thermocouple operation. An isothermal
zone box may be used to test the thermocouple in a different range.
There are several suggestions for thermocouple operations. These include:
-*M
1. Use the largest wire possible that will not shunt heat away from the
measurement area.
2. Avoid mechanical stress and vibration that could strain the wires.
3. Avoid steep temperature gradients.
4. Use'the thermocouple wire well within its temperature rating.
5. Use the proper sheathing materials in hostile environments.
5.4 LOCATION OF SAMPLING POINTS
There are many potential sources of fugitive VOC emissions hi a given facility.
The sources that will be considered here include: pump seals, compressor seals, process
valves, pressure relief devices, and agitator seals.
5-11.
-------
Pumps are used extensively by the target industries to move organic liquid. The
most widely used pump is the centrifugal pump. Other pumps used are the positive
displacement, reciprocating and rotary action, canned-motor, and diaphragm pumps.
Most pumps have a moving shaft which is exposed to the atmosphere. The fluid being
moved inside the pump must be isolated from the atmosphere. This requires a seal.
Leaks can occur at the point of contact between the moving shaft and the stationary
casing. The canned-motor and the diaphragm pumps do not have seals; therefore, they
more effectively prevent leaks.
Compressors are, basically, pumps that are used in gas service. Gas compressors
used in process units can be driven by rotary or reciprocating shafts and therefore need
shaft seals to isolate the process gas from the atmosphere. Rotary shafts may use either
packed or mechanical seals, while reciprocating shafts must use packed seals. As with
the seals in pumps, the seals in compressors are the most likely source of fugitive
emissions from these units.
One of the most common pieces of equipment in an industrial plant is the valve.
Individually, process valves have a low emission rate. However, because of the large
number of valves present in most plants, as a group they usually constitute the largest
percentage of fugitive VOC emissions. For example, in a 100,000 gallon per day
petroleum refinery, there are usually 25,000 process valves as compared to about 250
pump seals. In some instances, valves may make up 90 percent of the process
components that must be checked for leaks.
Many different types of valves exist, such as globe, gate, plug, ball and check
valves. However, they can be grouped into three functional categories:
• Block: used for on/off control. Generally, these valves are used only
occasionally, such as when there is a process change (i.e., unit shutdown).
• Control: used for flow rate control.
* Check: used for directional control. Since check valves are enclosed within
. process piping, they have no stem or packing gland and are not considered
to be a potential source of fugitive emissions.
.The most common valves in use are the gate valve and the globe valve. These valves
can be found either in-line or at the end of a process line.
Engineering codes require that pressure-relieving devices or systems be used in
applications where the process pressure may exceed the maximum allowable working
pressure of the vessel. The most common pressure-relieving device used in process units
is the pressure relief valve. Typically, a relief valve is spring loaded. It is designed to
open when the process exceeds a set pressure. This allows the release of vapors or
liquids until the system pressure is reduced to a normal operating level. When the
normal pressure is retained, the valve reseats and a seal is again formed. There are two
5-12
-------
potential causes of leakage from relief valves. Simmering occurs when the operating
pressure is similar to the set pressure of the valve, while popping occurs when the
operating pressure exceeds the set pressure, generally for an extremely short period of
time. The other cause of leakage is improper valve reseating after a relieving operation.
Agitators are commonly used to stir or blend chemicals. Like pumps and
compressors, agitators may leak organic chemicals at the point where the shaft
penetrates the casing. Consequently, seals are required to minimi^ fugitive emissions
from agitators.
Flanges are bolted, gasket-sealed junctions between sections of pipe and pieces of
equipment. They are used whenever pipe or equipment components (vessels, pumps,
valves, heat exchangers, etc.) may require isolation or removal. The possibility of a leak
through the gasket seal makes flanges a potential source of fugitive emissions. However,
the results of EPA's refinery sampling programs have shown that flanges have a very low
emission factor. Even though there are many of them in any refinery or chemical plant,
their overall contribution to fugitive emissions is small.
5.5 OBSERVATION PROCEDURES AND CHECKLISTS FOR VOC TESTING
5.5.1 Performance Criteria and Evaluation Procedures for Portable VOC Detectors
As previously stated, any portable VOC detector may be used as long as it meets
the performance criteria specified in Method 21. The performance criteria and detector
evaluation procedure is summarized in Table 5-5.
In addition to the performance criteria, Method 21 also requires that the analyzer
meet the following specifications:
• The VOC detector shall respond to those organic compounds processed at
the facility (determined by the response factor).
• The analyzer shall be capable of measuring the leak definition specified in
the regulation (i.e., 10,000 ppmv or "no detectable limit").
• The scale of the analyzer shall be readable to ± 5 percent of the specified
leak definition concentratioa
• The analyzer shall be equipped with a pump so that a continuous sample is
provided at a nominal flow rate of between 0.5 and 3 liters per minute.
• The analyzer shall be intrinsically safe for operation in explosive
atmospheres as defined by the applicable standards.
5-13
-------
TABLE 5-5. PERFORMANCE CRITERIA FOR PORTABLE VOC DETECTORS
Criteria
Response factor
Response time
Calibration
precision
Requirement
Must be <10
Must be .<30
seconds
Must be .<10%
of calibration gas
value
Time Interval
One time, before detector
is put in service
One time, before detector
is put in service; if
modification to sample
pumping or flow
configuration is made, a
new test is required
Before detector is put in
service and at 3 month
intervals or next, use,
whichever is later
Also, criteria for the calibration gases to be used are specified in Method 21.
Two calibration gases are required for both monitoring and analyzer performance
evaluation. One is a zero gas which is air with less than 10 ppmv VOC. The other
calibration gas uses a reference compound/air mixture. This calibration gas is also
referred to as the reference gas. The concentration of the reference gas is approximately
equal to the leak definition. The leak definition and the reference compound are both
specified in the applicable regulations. Calibration may be performed using a compound
other than the reference compound if a conversion factor is determined for the alternate
compound. The resulting meter readings during source surveys can be converted to
reference compound results. Often instrument manufacturers list conversion factors for
other gases in their operator's manuals. Because of the nonlinear responses, however,
care must be taken to use the conversion factor at the action level.
Selection of the Necessary Types of Instruments - Selection of the types of
instruments needed for source evaluation is based primarily on a review of the types of
industrial facilities within the agency's jurisdiction and an evaluation of the measurement
'requirements inherent in the promulgated VOC regulations. Agencies should also
determine if it is possible to select instruments that can be used for future air toxic
control requirements as well as the already existing VOC regulations.
Organic Vapor Analyzers - One important criterion in the selection of organic
vapor detectors is the response of the instrument to the specific chemical or chemicals
present in the gas stream. The abilities of the major classes of organic vapor analyzers
to detect different organic chemicals differ substantially. The response factor provides a
5-14
-------
convenient index of this property. The response factor (RF) is defined by:
Response factor » Actual concentration of cpropount;} Equation 5-1
Observed concentration from detector
A response factor must be determined for each compound that is to be measured,
either by testing of from reference sources. The analyzer response factor for individual
compounds to be measured must be less than 10.0. The response factor tests are
required before placing the analyzer in service, but do not have to be repeated at
subsequent intervals.
Response factors can be determined by the following method. First the analyzer
is calibrated using the reference gas. Then, for each organic species that is to be
measured, a known standard in air is obtained or prepared The standard should be at a
concentration of approximately 80 percent of the leak definition unless limited by
volatility or explosivity. In these cases, a standard at either 90 percent of the saturation
concentration or 70 percent of the lower explosive limit (LEL) is prepared. This mixture
is then injected into the analyzer and the observed meter reading is recorded. The
analyzer is then zeroed by injecting zero air until a stable reading is obtained. The
procedure is repeated by alternating between the mixture and zero air until a total of
three measurements have been obtained. A response factor is calculated for each
repetition and then averaged over three runs.
Alternately, if response factors have been published for the compounds of interest
for the type of detector, the response factor determination is not required, and existing
results may be referenced. When published response factors of the organic compound
being monitored are greater than 1 (approaching 10) or much smaller than 1
(approaching 0.1), it is prudent to measure the response factor for these specific
compounds. When screening for leaks from a source containing cumene, and FID can
be used (RF=1.87), while the catalytic oxidation detector cannot (no RF value). The
same data shows that neither of these devices would be capable of detecting leaks from a
source containing carbon tetrachloride.
The concept of using response factors as a general guide to analyzer applicability
is especially important when dealing with chemical mixtures. Since many process streams
in industrial plants are composed of a mixture of compounds, having a simple method to
determine the response factor for a given detector type is important. One EPA study
has concluded that analyzer response factors for a mixture fall between the responses
expected for the pure components. Therefore, if desired, an interpolated or weighted
average can be used to predict the response for mixtures based on known responses for
individual compounds. For further information see EPA 600/2-81-110, "Response of
Portable VOC Analyzers to Chemical Mixtures."
5-15
-------
Range and Accuracy - The ability of an instrument to measure 10,000 ppmv
should be carefully considered if the instrument will be used to determine compliance
with EPA Method 21 regulations. As indicated in Table 5-1, only a few of the currently
available units can operate at 10,000 ppmv or above. Other units can operate at this
concentration only by using dilution probes. Although dilution probes can be used
accurately, they can also be a large source of error. Both changes in flow rate through
the dilution probe and saturation of the charcoal tubes used to remove organic vapors
from the dilution air can lead to large errors in the indicated organic vapor
concentration. Dilution probes also complicate calibration and field span checks. For
these reasons, they should be avoided whenever possible.
. Generally, an instrument should have the desired accuracy at the concentration of
interest. It should be noted that an accuracy of ± 5 percent is required for Method 21
work. "'*
Response Time - The response time of an analyzer is defined as the time interval
from a step change in VOC concentration at the input of a sampling system to the time
at which 90 percent of the corresponding final value is reached as displayed on the
analyzer readout meter. The response time must be equal to or less than 30 seconds.
The response time must be determined for the analyzer configuration that will be used
during testing. The response time test is required before placing and analyzer in service.
If a modification to the sample pumping system or flow configuration is made that would
change the response time, a new test is required before further use.
The response time of an analyzer is determined by first introducing zero gas into
the sample probe. When the meter has stabilized, the system is quickly switched to the
specified calibration gas. The time, from the switching to when 90 percent of the final
stable reading is reached, is noted and recorded. This test sequence must be performed
three times. The reported response time is the average of the three tests.
Safety - All instruments used during field inspections of VOC emission sources
and air toxics emission sources must be intrinsically safe if they are to be used in
potentially explosive atmospheres. Localized pockets of gas (and even particulates)
within the explosive range can result from fugitive leaks and malfunctioning control
devices. Intrinsically safe means that the instrument will not provide a source of ignition
for the explosive materials when used properly. Instrument designs are certified as
intrinsically safe for certain types of atmospheres by organizations such as the Factory
Mutual Research Corporation.
The large majority of the organic vapor analyzers are designed to be intrinsically
safe in Class I areas. Factory Mutual, however, has certified only a few of the currently
available commercial instruments to be intrinsically safe for Class n areas.
5-16
-------
It should also be noted that battery-powered thermocouple are not designed as
intrinsically safe for either Qass I or Qass n atmospheres. Therefore, these instruments
cannot be taken into or through areas where there is a possibility of encountering
explosive mixtures of organic vapors and/or dust Conventional flashlights are also not
intrinsically safe, and they should be replaced by explosion-proof flashlights.
•>''
5.5.2 Laboratory And Shop Support Facilities
Because of their level of sophistication, organic vapor analyzers require laboratory
and instrument shop support facilities. Regulatory agency inspectors should not attempt
to store and calibrate the instruments in their offices, as this practice can lead to
significant safety problems and complicate the routine maintenance of the instruments.
Gas Flow Evaluation - Many of the organic vapor analyzers, especially the flame
ionization detectors, are sensitive to the sample flow rate. Routine confirmation of
proper flow rate is important, especially for those instruments that do not include a flow
sensor. Flow rates are normally measured by use of a rotameter designed for flow rates
between 0.5 and 5.0 liters per minute. The rotameter should be calibrated against a
soap bubble flow meter.
Electrical Diagnostic Equipment - The extent to which malfunctioning organic
vapor analyzers can be serviced by agency personnel is limited because the ^intrinsic
safety of the instrument can be voided inadvertently. Nevertheless, qualified agency
instrument technicians should be equipped to check such operating parameters as the
lamp voltage of photoionization units and the battery output voltage of all portable
instruments.
Thermocouple Calibration Equipment - The thermocouple readout device and
thermocouple probes should be calibrated at least twice a year. For convenience, the
calibrations should be performed in-house with a conventional tube furnace. The field
instrument and probes are compared against National Institute of Standards and
Technology (NIST) traceable thermocouple probes.
Static Pressure Calibration Equipment - All diaphragm-type static.pressure
gauges must be calibrated on at least a weekly basis. A relatively large U-tube
manometer can be permanently mounted in the agency laboratory for calibration of
0 to 10 inch W.C. and 0 to 60 inch W.C. gauges. An inclined manometer is needed for
calibration of the 0 to 2 inch W.C. gauges.
5.53 Routine Field-Oriented Evaluations of Instrument Conditions and Performance
Several instrument performance checks should be made before the inspector
leaves for the job site and during the routine screening of possible fugitive VOC sources.
The field-check procedures are in addition to, not a replacement for, the calibration
5-17
-------
procedures discussed earlier. The daily calibration, the field span check, and the routine
field performance checks are necessary to confirm that the instrument is operating
properly. Preferably, the initial instrument checks should be made by the agency's
instrument specialist assigned responsibility for the analyzers. Brief notes concerning
each day's initial instrument checks should be included in the main instrument
evaluation/maintenance notebook kept in the instrument laboratory. The inspectors
make the field checks by using the instruments at the job site and documentation of
these field checks should be a part of the inspectors' field notes.
Initial Instrument Checks - It is very important that a few simple instrument
checks be made before the inspector leaves for the job site. The appropriate field
checks for each instrument can be found in the instruction manual supplied by the
instrument manufacturer. The following common factors, however, should be checked
regardless of the type of instrument:
•. Leak checks including integrity of sample line and adequacy of pump
operation
* Probe condition
• Battery pack status
• Detector conditions
• Spare parts and supplies
These checks can be made in a period of 5 to 15 minutes. Repairs to the detectors,
batteries, and probes usually can be accomplished quickly if a set of spare parts is kept
on hand. Some of the checks that should be made before field work is begun are
discussed in the following discussions.
Leak Checks - To leak check the probes on units with flow meters, the probe
outlet should be plugged for 1 to 2 seconds while the sample pump is running. If the
sample flow rate drops to zero, there are no significant leaks in the entire sampling line.
If any detectable sample flow rate is noted, further leak checks will be necessary to
prevent dilution of the VOC sample gas during sampling. The leak checks involve a
step-by-step disassembly of the probe/sample line starting at the probe inlet and working
back toward the instrument. At each step, the probe/sample line is briefly plugged to
determine if inleakage is still occurring at an upstream location. Once the site of
leakage has been determined, the probe/sample line is repaired and reassembled. To
confirm that the probe/sample line is now free of air infiltration, the probe is again
briefly plugged at the inlet to demonstrate that the sample flow rate drops to zero.
When leaks are detected, there is sometimes a tendency to over-tighten the
fittings, especially those between the instrument body and the end of the sample line.
With some types of fittings (e.g., Swagelok fittings) over-tightening can damage the fitting
and even lead to persistent leaks.
5-18
-------
Units that do not have flow monitors should be leak-tested by installing a
rotameter on the sample line as close as possible to the inlet to the instrument body.
The leak-testing procedure described above can then be followed. Also, the sound of the
pump should be noted, as this provides one qualitative means of identifying clogs. It
should be noted, however, that pump noise is useless for identification of probe leakage
^because the pump continues to receive air due to the infiltration.
Some catalytic combustion units should not be leak tested by plugging the probe.
Short-term loss of sample flow would reportedly lead to high detector temperatures.
When more than one probe can be attached to the same instrument body, each
probe should be tested. Only those that can be sealed properly should be packed for
field use.
Probe Condition - The probes for some instruments can contain a number of
independent components, especially those that dilute the sample before analysis. The
physical condition of the probe should be checked visually before use. These checks
include, but are not limited to, an examination for:
• Presence of any organic deposits on the inside of the probe
• Presence of clean a particulate filter in the probe :
• Condition of orifice
5.6 TYPICAL SAMPLING PROBLEMS AND SOLUTIONS
One of the main problems in monitoring organic vapors is locating or pinpointing
the leaking source. Organic vapors are dispersed by the wind, sometimes making it
difficult to determine their source. It is important that the probe be moved slowly; the
slower the instrument response time, the slower the probe must be moved. Placing a
notebook or something similar (to block the wind) on the upward side of the suspected
leaking source may help locate the leak, but is not required by Method 21.
..'•
In some cases, it may be difficult to determine whether .a meter response is
caused by high ambient air hydrocarbons or by a source leak, particularly when the
* ambient reading is highly variable. This problem is commonly experienced in enclosed
areas. One method to determine if a source is leaking is to place the probe at the leak
source and then remove it from the leak source. This operation is repeated at regular
intervals. If the movement of the needle corresponds to the placement and removal of
the probe (keeping in mind the analyzer response time), the source is probably leaking.
•" The screening value is then determined by subtracting the ambient reading from the
measured screening results. A variety of such situations may be encountered and
judgement on the part of the operator may be required to obtain a representative
reading.
5-19
-------
Occasionally, a source may be encountered which has a highly variable leak rate.
In general, the maximum sustained reading or the maximum repeatable reading should
be recorded. Again, judgement on the part of the operator may be required to obtain a
representative reading.
Further difficulty may arise when emission sources contain heavier hydrocarbon
streams, particularly hot sources. When these sources are sampled, some of the organic
vapor tends to condense on the internal surfaces of the probe or sample hoses. The
response of the meter is considerably slower for the heavier hydrocarbons than for the
lighter ones. And, the meter may require more time to return to zero. When sampling
heavier hydrocarbons, the meter should be allowed to stabilize before reading the results.
Before sampling the next source, sufficient time should be allowed for the meter to
stabilize or return to zero. Often the meter will not return completely to zero and a
considerable adjustment may be required.
',:.'- Under no circumstances should the end of the probe be placed in contact with
liquid. If liquid is drawn into the system through the sample hose, it may damage the
analyzer. A liquid trap, connected between the analyzer and the sample probe, can be
used. In addition, the equipment being sampling may be covered with a film of grease
or dirt If the probe touches these components, the grease may plug the probe. The
inspector can carry a package of pipe cleaners to clean out the probe. Alternatively, a
Teflon probe extension can be used and the end cut off if it becomes clogged.
When using a portable VOC detector, the following safety practices are suggested:
1. Do not place a rigid probe in contact with a moving part such as a rotating
pump shaft. A short, flexible probe extension may be used.
2. Do not place the umbilical cord from the detector on a heated surface such
as a pipe, valve, heat exchanger, or furnace.
5-20
-------
CHAPTER 6
TOTAL GASEOUS NON-METHANE ORGAN1CS AS CARBON - METHOD 25
6.1 APPLICABILITY
,- Method 25 is designed to measure Total Gaseous Non-Methane Organics
(TGNMO's). Organic' compounds which exist as a gas or which have significant vapor
pressure at or below 250°F are subject to measurement by this method. Methane is
excepted from regulation and is not included in the reported organic emissions. During
analysis of the samples, all organics are catalyzed to methane. With proper analytical
procedures methane in the sample does not bias the TGNMO result The catalyzed
methane generated during sample analysis corresponds one-to-one with the carbon
.content of the sample, therefore the TGNMO results are reported in "ppm as carbon.11
Method 25 exhibits a 1:1 response for all carbon present in the sample and
therefore shows no bias due to differing response factors for different compounds.
Method 25 should be used when multiple organics are present or when the make-up of
the gas stream is not known. Products of incomplete combustion can be present at any
combustion source, therefore the make-up of the exhaust gas cannot be known with
certainty. For this reason Method 25 should be used for most combustion sources.
During Method 25 analysis, all carbon present is catalyzed to methane. Because
of this, the relative contribution to the total carbon content from different compounds
cannot be determined. If specific organic compounds are to be identified and quantified,
EPA Method 18 or some other method must employed, as Method 25 is not a compound
specific method.
Method 25 is applicable to sources with VOC concentrations of 100 ppm to
several percent by volume as carbon. The general application of the method allows
detection of concentrations as low as 100 ppmv, but with modifications cited in
Section 3.17 of EPA's "Quality Assurance Handbook, Volume HI" (EPA-600/4-77-027b)
and prior approval of these modification by the Administrator, a lower detectable limit
of 50 ppmv can be achieved.
Organic compound concentrations are expressed as carbon by adjusting the ppm
values for the number of carbon atoms per molecule. For example, an audit cylinder
containing 50 ppmv of toluene would have a concentration of 350 ppm as carbon
because toluene has seven carbons per molecule. A mixture of 50 ppm CO, 50 ppm
CH4, 2 percent CO2, and 20 ppm propane would have an organic concentration of 60
ppm as carbon. The 50 ppm CH4 is not counted because Method 25 excludes methane,
20 ppm propane is counted three times because there are three carbons per molecule in
propane. CO and CO2 are not counted because they are not organic constituents.
6-1
-------
62 METHOD DESCRIPTION
62.1 Sampling Procedures
EPA Method 25 uses an evacuated cylinder or tank to draw gas from the emission
source at a constant rate from a single point in the gas stream. The sample is integrated
evenly over the period of the test run. The sample is withdrawn from the source through
a heated probe, passed through a heated particulate filter and cold condensate trap, and
drawn into the evacuated cylinder. Heavy molecular weight organics are condensed out
of the sample into the chilled trap, and lighter organics are trapped in the evacuated
cylinder. The contents of both the condensate trap and the evacuated cylinder are
analyzed for organics following the test run.
A gaseous organic is defined as any organic which is in the gaseous state at
standard pressure and 121°C (250°F). Therefore, the probe is kept at a temperature
above 121°C at 129°C (265°F) and the filter housing is kept at 121±3°C (250±5°F). The
filter ensures that only gaseous organics and no organic particulate matter or mist passes
through to the sample. Particulate matter or mist could significantly bias the results
high. A thermocouple well should be placed at the probe exit and the filter housing.
The temperatures at these locations are monitored every 5 minutes during testing.
After passing through the filter, "heavy" organics condense in the chilled
condensate trap. The trap is kept on dry ice to maintain the coldest possible
temperature. The dry ice should be kept in an insulated container to ensure that it does
not sublime during the run. The dry ice level should be checked periodically during the
run. After the condensate trap, the remaining sample flows through a rotameter and
fine metering valve used to control the sampling rate, then into the evacuated tank or
cylinder. Method 25 states that the sample flow rate shall be between 60 cc/min and
100 cc/min. For a one hour ran at a sampling rate of 60 cc/min, a sample volume of
3600 cc or 3.6 liters will be collected. At a flow rate of 100 cc/min, the sample volume
will be 6 liters. The evacuated tank should have a volume of at least 4.5 liters to allow
for a minimum of 3.6 liters sample with room for error. The tank should not exceed 12
liters unless the sampling time is planned to be greater than one hour. If the volume of
the tank^is too large, the sample becomes diluted when the tank is pressurized, and the
sensitivity of the analysis is decreased. A diagram of the Method 25 sampling train is
shown in Figure 6.1 at the end of the chapter, page 6-22. All other figures and tables
are also presented at the end of the chapter.
622 Sampling Equipment
The sampling system consists of a heated probe, heated filter, condensate trap,
flow control system, sample purge pump, and evacuated sample tank. Complete systems
are commercially available, however a system can be fabricated from easily obtained
materials. Any system should meet the specifications listed below. Table 6-1, page 6-23,
6-2
-------
lists the calibration specifications and frequencies for the sampling system. Specifications
for the sampling equipment are presented in Table D-l, page D-2. Figure 6.2, page
6-24, is an example of an acceptable filter housing. Figure 6.3, page 6-25, is an example
of an acceptable condensate trap design.
6.2 J Analytical Procedures
Both the condensate trap and the sample tank are analyzed for nonmethane
organics (NMO's). The condensible organics are recovered from the trap by volatilizing
the organics and catalytically oxidizing them to carbon dioxide and collecting the CO3 in
an intermediate collection vessel (ICV). The carbon dioxide concentration in the ICV is
then measured. The non-condensible NMO's are recovered from the sample tank by
pressurizing the tank and analyzing the contents by using a flame ionization detector
(FID).
Both the CO2 evolved from the condensibles recovery and the non-condensible
NMO's are analyzed by using a gas chfomatograph with an FID. This analysis differs
from other VOC methods because the sample is first conditioned such that all organics
are reduced to methane before being introduced to the FID. A separation column is
used tc separate methane, carbon monoxide, carbon dioxide, and NMO's. As each
compound elutes from the separation column, it is catalytically oxidized to CO2. The
CO2 is then passed over a reduction catalyst in the presence of hydrogen. The CO2 is
reduced to methane. In this process, only methane is passed to the detector. The FID
response is assigned to CO, CO2, methane and NMO's by the elution time in the cycle.
Figure 6.4, page 6-26, is a schematic of the analysis cycle, and Figure 6.5, page 6-27, is a
schematic of the sample delivery valve and flow path for the analysis.
The condensible organics are recovered by "burning " the trap, i.e., heating the
trap in an oven to 200°C. After burning the trap, the CO2 generated by catalytic
oxidation is measured. Any CO2 present in the trap prior to burning will bias the results
high. To eliminate this bias, a "cold purge " is done. The cold purge consists of purging
the trap with pure air while the trap is immersed in dry ice to drive; off any CO2 present.
The purge effluent is monitored with a non-dispersive infrared (NDER) analyzer to
assure that all the CO2 is gone before ending the purge. The NDIR analysis is non-
destructive, and the purge effluent is collected at the exit of the NDIR in a second ICV.
As the NDIR response approaches zero, a 10 ml syringe is used to extract a sample from
an injection port located before the NDIR. The 10 ml sample is analyzed using the
NMO analyzer for CO2 concentration. The purge is considered complete when the CO2
concentration is below 10 ppm. This ICV is analyzed for NMO's as some organics may
volatilize during the purge.
After the cold purge, the trap bum is done. The trap is placed in the oven at
room temperature. Then it is heated to 200°C. During the burn, the trap is purged with
pure oxygen. As the trap is heated, organics are carried by the purge to the catalyst bed
6-3
-------
where they are oxidized to CO2. The purge is again monitored for end-point using the
NDIR, and 10 ml samples are withdrawn from the injection port and checked until the
CO2 concentration is less than 1 ppm. The purge gas is collected in the ICV and
analyzed for CO2 and NMO's (in case some organics passed through the catalyst). Care
must be taken that the organics are not volatilized too fast or the catalyst may become
saturated. If this happens, some organics will pass through to the ICV. They might not
be accounted for by NMO analysis. However, If the organics have high boiling points,
they may condense on the side of the ICV and bias the results low. This can be avoided
by heating the trap slowly or by adding auxiliary oxygen to the system through a tap
between the oven and the oxidation catalyst. Each analytical system should be equipped
with such a tap for handling samples with high organic catches. High NMO
concentrations in the ICV vessel may indicate breakthrough of unoxidized hydrocarbons.
NMO concentrations from the sample tank, the ICV from the cold purge,
and the ICV from the trap burn, as well as the CO2 concentration from the trap burn are
weighted according to the vessel volume associated with each concentration to determine
the carbon content of the original sample.
6.2.4 Analytical Equipment
Since most Method 25 analyses are done post-test in a laboratory, it will be
difficult to observe the equipment and procedures. A list of the major equipment
components needed for analysis is presented in Appendix D.2, page D-4. Table 6-2,
page 6-28, contains the major components of the analytical system and the calibration
schedule for each.
6.3 PRECISION AND ACCURACY
Very little data is available concerning the precision and accuracy of Method 25.
A limited number of laboratories actively perform the NMO analysis and interlaboratory
studies are not complete as of this writing. The generally accepted limits for precision
and accuracy are both 20 percent of the mean value. Preliminary results of
interlaboratory studies indicate that at lower concentrations, less than 200 ppm, the
accuracy limit becomes much higher than 20 percent. At high concentrations, greater
than 1000 ppm, the accuracy limit should be less than 20 percent. High concentrations
of CO2 and water will further increase the limits of the accuracy and precision, especially
at low organic concentrations.
6.4 LOCATION OF SAMPLING POINTS
Method 25 sampling is performed at a constant rate from a single point in the
duct. No traversing or isokinetic sampling is required. The nozzle on the sample probe
6-4
-------
can be of any size, but should be constructed such that it can be turned away from the
direction of flow in the stack to avoid collecting paniculate matter. The nozzle should
be placed at the point in the stack or duct with average velocity.
A velocity traverse should be done prior to sampling, and the average impact
pressure differential of the gas stream should be determined. The nozzle should be
placed at the point where the delta P (AP) is closest to the average AP. If the velocity
profile across the duct is level, place the nozzle at the center of the duct.
63 OBSERVATION PROCEDURES FOR METHOD 25 TESTING
It is the responsibility of the testing firm to ensure that the sampling and
analytical procedures are performed correctly. The following detailed information is
given only as training guide for the less experienced test coordinator and does not imply
mandatory actions by the test coordinator except when the discussions state that the test
coordinator "shall" conduct a given procedure.
6.5.1 Equipment Specifications
All equipment used during sampling should be checked to ensure that it conforms
to the requirements of Method 25. Table D-l, page D-2, is a checklist for sampling
equipment specifications and calibration which may be completed or used as a guide by
the test coordinator or testing firm. Thermocouples and rotameters should be calibrated
and the calibration values and last date of calibration should be noted. Rotameters
should have a gamma between 0.9 and 1.1, and thermocouples must agree with the
calibration standard within 3°C (5°F). The sample tanks should have been calibrated for
standard volume and tagged with a unique identification code. The volume and tag
number for each tank should be recorded by the testing firm.
6.5.2 Pre-test Leak Checks
A leak check must be conducted on the entire sampling system prior to sampling.
Table D-2, page D-4, is for sampling operations checklist which may be completed or
used as a guide by the test coordinator. The allowable leak rate is 1 percent of the
sample flow rate. The system is assembled and the probe and filter housing are heated
to their respective set-points. A system which is leak tight when cold may not be leak
tight when hot, so leak checks should be conducted with the system heated. The
condensate trap is placed in the container of dry ice. The top of the trap should be 2.5
to 5 cm above the top of the dry ice. If the 1/8 inch tubing connecting the trap to the
filter is covered with dry ice, water and CO2 will freeze in the tubing and block the flow
during sampling. As the trap temperature reaches equilibrium with the dry ice, frost will
form on the trap. The frost line should not extend onto the connecting tubing. The inlet
tubing to the condensate trap should extend into the filter housing so that organics don't
condense in the tubing, but condense in the packing of the trap.
6-5
-------
The leak rate is determined by attaching a mercury manometer to the inlet of the
probe and evacuating the system to within 10 mm Hg of absolute. The metering valve
used to control sample flow rate should be wide open to allow the system to evacuate as
quickly as possible. Each sampling system will allow diversion of sample flow to a purge
pump. This is usually done with a 3-way valve. Be sure that the 3-way valve is turned
toward the sampling position.
As the system is evacuated, the air moving from the sampling train to the
evacuated tank can be monitored with the rotameter. The flow through the rotameter
should drop to zero, indicating that all the air is out of the system, before the leak check
is started. If no flow is initially seen on the rotameter, the flow control valve may be
closed or the sample/purge valve might be at the neutral or purge position. The sample
valve at the sample tank is switched off to isolate the system from the tank. The system
should maintain the same vacuum, as read on the Hg manometer, for 5 or 10 minutes.
The allowable leak rate is based on a check time of 10 minutes, so if 5 minutes is used,
divide the allowable leak rate by 2. A tank other than the one intended for sample
should be used for the leak check because the air initially present in the sampling system
is drawn into the leak check tank. The allowable leak rate is calculated by the following
equation:
A P = 0.01 x F x Pb x t/ Vt
Equation 6-1
where:
A P = allowable pressure change, mm Hg
F = sample flow rate, cc/min
Pb = barometric pressure, mm Hg
t = leak check time, minutes
Vt = volume of the sampling train, cc
For example, if the sampling train volume were 30 cc's, the intended flow rate
60 cc/min, and 760 mm Hg and 10 minutes are used for barometric pressure and leak
check time. Then the maximum allowable pressure change over the 10 minute period is:
AP = 0.01 x 60 cc/min x 760 mm Hg x 10 mins/30 cc
= 152mmHg = 5.98 in. Hg
Equation 6-2
After the leak check, the sample/purge valve is turned to the neutral position to
prevent in-Ieakage of ambient air. The flow control valve is closed all the way so that
the sample flow rate can be accurately controlled when the run is started. The leak
check tank is replaced with a fresh sample tank. When the probe exit temperature and
filter housing temperature are at their set points ±3°C (±5°F), remove the manometer
fitting from the tip of the probe and place the probe in the stack at the previously
determined sampling point.
6-6
-------
Each sample tank should also be leak checked. This is most easily accomplished
by evacuating the sample tanks one day prior to the test date. The absolute pressure in
each tank as well as the barometric pressure and tank temperature are recorded. On the
day of testing, the tank pressure and temperature are measured again. After allowing for
differences in barometric pressure and tank temperature, the tank absolute pressure
should be identical to the previous day's reading. If the sample tanks cannot be
evacuated on the day preceding testing, they should be evacuated and left for at least
one hour. The absolute pressure in the tank should not change in that hour. If the tank
leak checks are started the day prior to testing, the absolute pressure change should be
no more than 5 mm Hg. For a one hour leak check, the pressure change should be no
more than 1 mm Hg.
6.5.3 Pre-test Sampling Train Purge
A minimum of 10 minutes before the start of the run, the sample/purge valve is
turned to purge,, and the purge pump is turned on. The purge rate should be set at 60 to
100 cc/min. The purge pump draws stack gas through the probe and filter, but not the
condensate trap or sample tank. Just before sampling starts, the purge pump is turned
off and the sample/purge valve is returned to the neutral position.
6.5.4 Sampling Procedures
To start the run, the sampler needs to perform the next four steps nearly
simultaneously:
1. Start the run timer.
2. Open the sample/purge valve to the sample position
3. Open the valve to the sample tank. If a sealing quick connect is used, push
the quick connect to the "locked-in" position.
4. Open the flow control valve to the desired flow rate.
The pressure differential between the duct and the sample tank is the driving
force in the sampling train. As the sample tank vacuum decreases, the flow rate will
decrease if left unattended. The sample flow rate, probe exit temperature, and filter
housing temperature should be monitored and recorded every 5 minutes during the run.
The flow control valve should be used to adjust the sample flow rate such that it stays
constant to ± 10 percent of the intended flow rate.
If the vacuum in the sample tank decreases to the point that the flow rate can no
longer be maintained, the following procedure should be followed: Turn off the tank
sample valve. Disconnect the tank from the system without disconnecting any other part
of the system. Take another leak check and evacuate the sample tank. Record the tank
vacuum and temperature. Attach it to the sampling train and resume sampling until the
required run time has been met or exceeded.
6-7
-------
To end the sampling run, the sample/purge valve is moved to the neutral position
and the sample tank valve is turned off. Before the train is taken apart, the condensate
trap and sample tank should be tagged with the date, ran number, facility and sampling
location designations, and the reference number assigned to the project by the testing
company. This will assure that no mix-up or confusion arises over sample identity during
analysis. The condensate trap and sample tank identification numbers must be recorded
on the data sheet for that run.
A post-test leak check is not required by Method 25. Under no circumstances
should a leak check be done using a fresh sample tank or a leak check tank, or the purge
pump. Since sampling is completed at a vacuum much lower than that of a completely
evacuated sample tank, organics will volatilize and be carried from the cold trap to the
leak check tank or purge pump. An alternative that is acceptable to verify the integrity
of the sampling system at the end of the run is as follows:
Turn the sample/purge valve to the neutral position. Reconnect the sample tank
or open the sample tank valve. Wait for the rotameter to drop to zero. Record the
system vacuum on the gauge installed in the system (this may be close to zero, but is
usually 10 mm Hg). Turn off the sample tank valve and remove the sample tank from
the system. Wait 5 minutes and check the system vacuum again. The vacuum should
not decrease by more than 2 mm Hg.
This procedure is not required and is at the option of the test coordinator. This
procedure will not cause sample loss because it is conducted at the lowest vacuum in the
system during the run. If any lighter organics are drawn from the trap to the sample
tank, they will be recorded in the non-condensible NMO analysis.
6.5.5 Post Sampling Procedures
Proper procedures must be followed on-site to insure the samples are recovered
for analysis. The sample tank is removed from the train and the final tank absolute
pressure is recorded to the nearest mm Hg. The condensate trap is removed from the
sampling system promptly and both the inlet and outlet are plugged to prevent leakage
into or out of the trap. The trap must be kept cold until condensate recovery. To
accomplish this, sufficient dry ice must be available to keep the traps cold until they are
transported to the analytical lab. The dry ice should be kept in specially designed
coolers which will maintain dry ice for several days. The sample tanks may be
pressurized on-site. If they are, the final positive pressure must be recorded as well as
the tank temperature at that time.
Pressurizing the sample tanks is done by using a "Y" connector to attach the
sample tank to both a cylinder of carrier grade air, and the mercury manometer. An
on/off valve is located in the leg of the "Y" between the sample tank and the air
cylinder. The on/off valve must be closed at this point. The tank pressure is read on
6-8
-------
the manometer and the tank temperature is measured and recorded. The on/off valve is
opened and air is pushed into the sample tank until approximately 300 mm Hg pressure
is indicated by the manometer. The on/off valve is closed again to isolate the sample
tank from the air cylinder and the new tank temperature and pressure are measured and
recorded. The sample tank is removed from the T" and prepared for shipment to the
laboratory.
Inspection of the test data sheets before leaving the site may disclose omissions or
problems which can easily be remedied now, but which could cause the data to be
unacceptable if undetected until later. A copy of an acceptable data sheet is included in
Figure D.I, page D-8. Items which must be recorded are:
Date • Run Number
Company Name • Source Designation
Test Start Time • Test Finish Time
Operator • Sample Train I.D.
Sample Tank I.D. • Sample Tank Nominal Volume
Trap I.D. • System Leak Check Rate
Train Volume • Tank Temperatures & Pressures -
pre-test and post-test
Some missing data may be filled in at the end of the test run with the correct
values. Failure to record other values at the appropriate time may be cause for
repeating a run. One such item is the pre-test or post-test tank pressure and
temperature. If the tank was pressurized before this data was recorded, the run must be
repeated, because the sample volume cannot be calculated without it. A missing tank
I.D. or trap I.D. may be recovered from the tags on the tanks and traps, but if the traps
and tanks were not tagged and no other way exists to assign the proper tanks or traps to
the proper runs, the runs must be repeated.
6.6 SAMPLING PROBLEMS, ERRORS, SOLUTIONS, AND ACTION REQUIRED
Because of the large number and variety of organic processes, it is not possible to
discuss all of the sampling problems related to Method 25 sampling. Only the most
common problems will be addressed.
6.6.1 High Gas Sample Moisture Content and Freezing of Trap
Due to the condensate trap temperature maintained by the dry ice, any moisture
and some CO2 present in the sample will freeze in the trap. If the tubing leading to the
trap becomes too cold, water will freeze in the tubing causing a plug. This condition is
indicated by a sudden loss of flow during sampling or by difficulty in maintaining
proper flow with the flow control valve entirely open. If a plug develops, the timer
should be stopped. The purge/sample valve is turned to the neutral position and the
6-9
-------
sample tank valve is closed. Any simultaneous trains must also be stopped. The trap is
then raised out of the dry ice until the connecting tubing is 2.5 to 5 cm above the dry ice.
When the frost film on the connecting tubing melts, sampling can be resumed.
If the sample gas has a high moisture content, the freezing problem at the inlet to
the trap may become chronic. If raising the trap out of the dry ice bath does not
alleviate the problem, the inlet line can be insulated. Also, a second trap may be placed
in front of the condensate trap. The second trap should be kept in an ice water bath.
The water will condense out in this trap without freezing and the condensate trap will
not collect as much moisture. This second trap must be analyzed in the same manner as
the condensate trap. Both traps are kept on dry ice during shipping and storage.
6.62 Use of Electrical Service Not Permitted for Probe and Filter
If for safety reasons, the plant cannot allow the use of electrical service at the
sampling site, sampling should be conducted using an in-stack filter. The filter should
consist of a stainless steel tube packed with quartz wool. The condensate trap is then
connected directly to the in-stack filter.
6.63 Probe Exit or Filter Temperatures Not Within Specification
The temperature at the probe exit and the filter housing are measured every
5 minutes during the test run. Method 25 requires that these temperatures be
maintained above 129°C and at 121 ±3°C, respectively. If specifications, the test
coordinator may allow or disallow the run. Since an NMO is defined as a non-methane
organic existing at or below 121°C, a probe temperature or filter temperature
significantly below 121°C may allow some organics to condense in the sampling train.
Therefore, if the probe exit or filter temperature is maintained at or falls below 121°C
for an appreciable length of time, the run should be considered invalid. If, however, the
probe exit or filter temperature exceeds the limits, organics which have higher boiling
points may pass through the filter. If the temperature specifications are exceeded, the
bias willvbe toward higher organic concentrations and the agency may choose to accept
the runs.:
x.'
6.6.4 Non-constant Sample Flow Rate
Method 25 sampling requires a constant flow rate. The flow rate is monitored
every 5 minutes throughout the run. However, because of the variable driving force
from the sample tank, the flow may not have been maintained at the proper setting
± 10 percent This can be detected at the end of the run by examining the vacuum gauge
readings for each five minute point. The difference between readings should be constant
over the course of the run. For instance, if a 4.5 liter tank is used for a 1 hour run, the
vacuum change over each 5 minute period should be between 3 and 5 cm Hg. If the
flow rate falls below the set point the vacuum change would be less than 3 cm Hg. If
6-10
-------
the flow is higher than the set point the vacuum change will be greater than 5 cm Hg.
Vacuum gauges normally have increments of 1 cm Hg. So, a vacuum reading is only
'accurate to within about 1 cm Hg, and this should be considered when reviewing the
data A difference of greater than 2 cm Hg from the normal vacuum change for any
5 minute period would indicate a problem with that reading.
The appropriate response when inconsistent flow is detected is dependent on the
emission source characteristics. If the emission source is a batch operation where one
full batch comprises one test run, the flow problems would weight the results toward the
period of highest flow and away from the period of lowest flow. This could be significant
depending on process conditions. If the source operates in a steady state condition, then
each 5 minute sampling period would be fairly equivalent in terms of the emission rate,
and the inconsistent flow would have little effect on the results.
One situation which is serious regardless of process conditions is very high flow.
The condensate'trap is designed for a flow rate of 60 to 100 cc/min. If the flow is much
higher, then breakthrough of heavy organics may occur. The heavy organics would then
condense on the sides of the sample tank and not be included in the NMO analysis. If
the vacuum change for any period indicates that the flow rate during that period was
greater than 200 cc/min, die run should be repeated. High flows are commonly seen at
the start of sampling. If the flow control valve is not closed after being wide open during
sampling a large amount of gas will be drawn through the system before the flow can be
set properly.
6.6.5 Use of Method 25 for Measuring Low Levels of Organics
The lower detectable limit of Method 25 is 100 ppm as carbon. Due to the large
number of factors contributing to imprecision in both sampling and analysis, the accuracy
of the results is questionable at concentrations near the lower limit. A simple way to
increase the accuracy at lower concentrations is to extend the sampling time from
1 to 2 hours. This will require that the vacuum in the sample tank be sufficient to draw
sample for twice as long. This can be accomplished by using a larger sample tank
( >9 liters) or by changing sample tanks halfway through the run. When the vacuum in
the first sample tank becomes too low to maintain the proper flow rate, stop sampling
and disconnect the sample tank. The condensate trap should not be disconnected. The
same condensate trap is used for the entire test run. The tester should then install a
new sample tank and record its volume and I.D. number on the data sheet Do not
perform a leak check with the new sample tank in place. Organics will be drawn from
the condensate trap to the sample tank by the higher vacuum. Proceed with the
extended test run.
If the source emissions have a high moisture content, the extended test period
may cause the condensate trap to fill up with water or ice. This can be avoided by using
an auxiliary trap placed in an ice water bath prior to the condensate trap (see
6-11
-------
Section 6.6.1). Both traps must be analyzed. The NMO contribution from both tanks
and traps is summed.
Method 25 was not intended to measure organics at levels below 100 ppm as
carbon. However, if the tester has no other options, Method 25 can be used under the
following conditions: (1) extreme caution must be used in preparing the traps and tanks
and (2) two traps and two tanks should be set aside as field blanks with the analytical
results subtracted from the field sample values. This approach will improve
measurements at low level sources, but the accuracy and precision will be poor.
6.6.6 Sampling and Analysis by Different Companies
Because of the small number of laboratories that conduct Method 25 analysis, a
large-portion of the Method 25 sampling and analysis is conducted by two different
companies. This creates problems in assigning responsibility when audit sample results
are not acceptable. If the sampling company wants to check the consistency of the
analytical results, they should obtain extra traps and cylinders from the laboratory.
These clean traps and cylinders should not be opened, marked as if they were a sample,
and submitted for analysis.
X
6.6.7 Measurement in Ducts Containing Organic Droplets
If the gas stream to be sampled contains organic droplets, Method 25 results can
be significantly biased-high. The testing firm should first try to find another sampling
location. If this is not possible, an in-stack filter may be added to the sampling system
with both the in-stack and out-of-stack filters being replaced after each run. The addition
of an in-stack filter should help collect organic droplets and will reduce the loading on
the out-of-stack filter.
6.7 ANALYSIS
6.7.1 Analytical System Performance Checks
Method 25 analysis is rarely conducted on-site. Therefore, direct observation of
the analytical procedures is seldom possible. However, documentation of the quality
control checks required by Method 25 should be included in the compliance test report.
Method 25 requires the following checks to be done at system start-up or after any
period where the analytical system has been unused for 6 months or longer:
Oxidation Catalyst Efficiency Test
Reduction Catalyst Efficiency Test
NMO Response Linearity Test
CO2 Response Linearity Test
NMO Analyzer Performance Check
6-12
-------
• Condensible Organic Recovery System Check
Although these tests and checks are only required at system start-up,
documentation showing each test to be within the specifications of the method should be
included with each report. The checks are described below.
The last portion of this chapter addresses performance checks that should be
conducted each day that the system is used for NMO analysis.
Oxidation Catalyst Efficiency Test - With both the oxidation and reduction
catalysts unbeated, analyze the high level methane standard (nominal 1 percent CH4 in
air) in triplicate. With only the oxidation catalyst heated to its operating temperature,
reanalyze the high level methane standard in triplicate. Record data and calculate the
oxidation catalyst efficiency using the following equation:
Oxidation Catalyst Efficiency = (Rl - R2)/R1 x 100 Equation 6-2
where:
Rl = response with both catalysts unheated
R2 = response with only oxidation catalyst heated
If the oxidation catalyst is working properly, the methane is all oxidized to CO2
when the catalyst is heated. The system response would then be zero during condition
R2. The average response with the oxidation catalyst heated should be less than
1 percent of the average response obtained with both catalysts unheated.
Reduction Catalyst Efficiency Test - With the oxidation catalyst unheated and the
reduction catalyst heated to its operating temperature, analyze the high level methane
standard in triplicate. Repeat the analysis in triplicate with both catalysts heated to their
operating temperatures. Record the data and calculate the reduction catalyst efficiency
using the equation below:
Reduction Catalyst Efficiency = R4/R3 x 100 • Equation 6-3
where:
R3 = response with reduction catalyst only heated
R4 = response with both catalysts heated
When the oxidation catalyst is heated (condition R4), the methane is oxidized to
CO2 and the reduction catalyst must then reduce the CO2 back to methane. If the
reduction catalyst is not 100 percent efficient, then the CO2 will pass through to the FED
and the response will be lower than the response at condition R3. The responses
observed under these two conditions should agree within 5 percent.
6-13
-------
NMO Response Linearity Test and Initial Calibration - With both catalysts at
their operating temperatures, perform triplicate injections of each of the following .
propane standards: 20 ppm, 200 ppm, and 3,000 ppm in air nominal. Convert certified
concentrations in ppm to ppm C by multiplying the ppm concentrations by 3. Record
these concentrations on a data sheet, along with the area responses observed for each
injection. Calculate the mean response factor as ppm C/mean area for each standard
and the overall mean response factor for all three standards. The NMO response
linearity is acceptable if the average response factor of each calibration gas standard is
within 2.5 percent of the overall mean response factor and if the relative standard
deviation for each set of triplicate injections is less than 2 percent. The overall mean
response factor is used as the initial NMO calibration response factor
CO2 Response Linearity Test and Initial Calibration - Perform the linearity test
as described above, except use CO2 calibration standards of 50 ppm, 500 ppm, and
1 percent in air. The overall mean response factor is used as the initial CO2 calibration
response factor (RF^). The CO2 calibration response factor (RFC02) should be within
10 percent of the NMO calibration response factor
NMO Analyzer Performance Test - After calibration of the NMO response as
described above, analyze each of the following four gas standard mixtures in triplicate.
Standard 1 is nominally 50 ppm CO, 50 ppm CH4, 2 percent CO2, and 20 ppm propane
in air; Standard 2 is nominally 50 ppm hexane in air; Standard 3 is nominally 20 ppm
toluene in air; and Standard 4 is nominally 100 ppm methanol in air. Record the NMO
area responses for each standard on the data sheet. Convert the certified organic
compound concentrations of the standards to ppm C by multiplying by the carbon
number of the compound (3 for propane, 6 for hexane, and 7 for toluene). Record these
concentrations on the data sheet as the expected concentrations. Calculate the mean
NMO concentration of the test gas. The analyzer performance is acceptable if the
average measured NMO concentration for each mixture or standard is within 5 percent
of the expected value.
Condensible Organic Recovery System Check - This check is conducted in three
stages. First, the carrier gas is checked for its blank concentration. The carrier gas
blank value should be less than 5 ppm. Second, the oxidation catalyst is checked, then
the system is checked by spiking with a known organic concentration. A schematic of the
condensate recovery system is shown in Figure 6.6, page 6-29.
The oxidation catalyst efficiency is tested by the following sequence.
1. The system is set-up as normal using a clean condensate trap, and clean
immediate collection vessel (ICV) which has been evacuated.
2. The recovery valve is set to the vent position, and the carrier gas is
replaced with the 1 percent methane in air cylinder.
3. Set the flow from the 1 percent methane standard equal to the normal
6-14
-------
carrier gas flow rate. Allow the NDIR response to stabilize, then turn the
recovery valve to the ICV. Using the flow control valve on the ICV,
maintain the system pressure near atmospheric. Continue flow until the
ICV has been pressurized to 300 mm Hg.
4. Analyze the CO2 concentration in the ICV using the NMO analyzer. The
CO2 concentration should agree with the certified concentration of the 1
percent methane standard within 2 percent
Hie condensible organic recovery efficiency is checked by setting the recovery
system up as normal, except the condensate trap is replaced with a liquid sample
injection unit similar to that shown in Figure 6.7, page 6-30. The recovery valve is
turned to the collect position, then 50 ftl of hexane is injected into the liquid sample
injection unit The liquid organic is collected by the normal procedures of the method.
Hexane has six carbons and the percent recovery is calculated using the following
equation:
Percent Recovery=1.604 x - J^gt Equation 6-4
where:
M = molecular weight of compound injected, g/g-mole
Vv = volume of ICV tank, m3
Pf = final pressure of ICV tank, mm Hg absolute
COB = measured concentration (NMO analyzer) for the condensate
trap ICV, ppm CO2
L = volume of liquid injected, ftl
p = density of liquid injected, g/cc
Tf = final temperature of ICV, °K
N = carbon number of liquid compound injected (N = 12 for
decane, N = 6 for hexane).
The recovery efficiency should be checked in triplicate with 50 fd hexane, then in
triplicate with 10 /A hexane, 50 fA decane, and 10 pi decane. The percent recovery is
acceptable if the average percent recovery is 100 ± 10 percent with a relative standard
deviation of less than 5 percent for each set of triplicate injections.
The following set of performance checks should be performed each day that the
system is used for NMO analysis.
Leak Test of Condensibles Recovery System - A clean trap should be installed in
the system, then the vacuum pump is used to evacuate the system to 10 mm Hg absolute.
6-15
-------
The system is closed off and the system pressure is monitored for 10 minutes. The
system pressure should change by no more than 2 mm Hg over this time period.
System Background Check for Condensibles Recovery System - The carrier air
and auxiliary air are set to their normal operating parameters. The recovery valve is set
to the vent position and a 10 ml syringe is used to extract a sample from the injection
port upstream of the NDIR. The 10 ml sample is injected into the NMO analyzer and
the CO2 concentration is recorded. The CO2 concentration should be less than 10 ppm.
Oxidation Catalyst Efficiency - This test is performed in the same manner as
described previously for system start-up.
CO2 Analyzer Response - The highest level CO2 calibration gas is analyzed in
triplicate. The average peak area is used to calculate a daily response factor for CO2.
The daily CO2 response factor should agree with the system CO2 response factor
determined during system start-up.
NMO Response Check - The gas mixture containing nominally 50 ppm CO,
50 ppm CH4, 2 percent CO^ and 20 ppm propane in air is analyzed in triplicate. The
average area count value is used to calculate a daily response factor as carbon. The
daily response factor should be within 5 percent of the initial NMO response factor.
If audit cylinder samples are included with the field samples to be analyzed, the
test results may be accepted or rejected on the basis of the audit results, and the
performance check documentation may be optional. However, if no audit is performed,
the test results cannot be evaluated without proper documentation of system
performance and calibration.
6.72 Calculations
The calculations required to compute the TGNMO concentration of the original
sample from the sampling and analytical data recorded are reproduced in Appendix D.3,
page D^t. The number of variables is large and some of the equations are complex. It
is recommended that a computer program or spreadsheet software be used to handle all
calculations. A copy of the program used for calculations should be included with the
test results. Also, example calculations using data from one of the test runs for each test
series should be included. The run number of the example calculation must be stated.
Check to be sure that the example calculation agrees with the reported concentration
within reasonable round-off error. Choose a different test run and perform the
calculations yourself. The answer should again agree with the reported result to within
reasonable round-off error.
Calculations should be carried out to at least one extra decimal place beyond that
of the acquired data and should be rounded off after final calculation to two significant
6-16
-------
digits for each run or sample. All rounding of numbers should conform to ASTM 380-76
procedures.
6.8 AUDIT PROCEDURES
An audit is an independent assessment of data quality. Independence is achieved
if the individual(s) administering the audit and their standards and equipment are
different from the regular field team and their standards and equipment. Routine
quality assurance checks by a field team are necessary to generate good quality data, but
they are not part of the auditing procedure. Table 63, page 6-31, summarizes the quality
assurance functions for auditing.
Based on the requirements of Method 25 and the results of collaborative testing
of other EPA Test Methods, one performance audit is required when testing for
compliance for Standards of New Source Performance and is recommended when testing
for other purposes; and a second performance audit is recommended. The 2
performance audits are:
1. An audit of the sampling and analysis of Method 25 is required for NSPS
and recommended for other purposes.
2. An audit of the data processing is recommended.
It is suggested that a systems audit be conducted as specified by the test coordinator in
addition to these performance audits. The two performance audits and the systems audit
are described in detail in Chapters 6.8.1 and 6.8.2, respectively.
6.8.1 Performance Audits
Performance audits are conducted to evaluate quantitatively the quality of data
produced by the total measurement system (sample collection, sample analysis, and data
processing). It is required that a cylinder gas performance audit be performed once
during every NSPS compliance test utilizing Method 25 and it is recommended that a
cylinder gas audit be performed once during any test utilizing Method 25 conducted
under regulations other than NSPS.
Performance Audit of the Field Test - As stated in Section 4.5 of Method 25
(40 CFR 60, Appendix A) and the "Instructions for the Sampling and Analysis of Total
Gaseous Nonmethane Organics from Quality Assurance Audit Cylinders using EPA
Method 25 Procedures" (supplied with the EPA audit gas cylinders), a set of two audit
samples are to be collected in the field (not laboratory) from two different concentration
gas cylinders at the same time the compliance test samples are being collected. The two
audit samples are then analyzed concurrently and in exactly the same manner as the
compliance samples to evaluate the tester's and analyst's technique and the instrument
calibration. The information required to document the collection and analysis of the
6-17
-------
audit samples has been included on the example data sheet shown in Figure 6.8,
page 6-32. The audit analyses shall agree within 20 percent of the actual cylinder
concentrations. The testing firm may obtain audit cylinders by contacting the agency
responsible for observing and/or evaluating the compliance test and informing the
agency the time and location of the compliance test. The test coordinator will then
contact: U.S. Environmental Protection Agency, Atmospheric Research and Exposure
Laboratory, Quality Assurance and Technical Support Division, Research Triangle Park,
North Carolina 27711 and have the cylinders shipped to the specified site.
Responsibilities of the Test Coordinator - The primary responsibilities of the
observer are to ensure that the proper audit gas cylinders are ordered and safe-guarded,
and to interpret the results obtained by the analyst
When notified by the testing firm that a compliance test is to be conducted, the
test coorinator orders the proper cylinders from the EPA's Quality Assurance Division.
Generally the audit cylinders will be shipped (at EPA's expense) directly to the specified
test site. However, if the test coordinator will be on-site during the compliance test, the
audit cylinders may be shipped to the testing firm for transport to the sampling site.
Since the audit cylinders are sealed by EPA, the testing firm will not be allowed to
collect any audit gas without breaking the seal.
The audit gas concentration(s) should be in the range of 50 percent below to 100
percent above the applicable standard. If two cylinders are not available, then one
cylinder can be used. If the applicable regulation is based on removal efficiency rather
than emission limits, an audit cylinder should be provided near the expected
concentration of both the inlet and outlet of the control system. The testing firm should
provide the test coordinator with the best available information to approximate the
concentration at the control system inlet. The expected concentration at the control
system outlet can be calculated using the regulated removal efficiency.
The test coordinator must ensure that the audit gas cylinder(s) are shipped to the
correct-address, and to prevent vandalism, verify that they are stored in a safe location
both before and after the audit. Also, audit cylinders should not be analyzed when the
pressurevdrops below 200 psi. The test coordinator ensures that the audits are conducted
as described below. At the conclusion of the collection of the audit samples, if the
testing firm will transport the audit cylinders to the home laboratory for shipment back
to the EPA/QAD contractor, the test coordinator seals both cylinders to ensure that
additional audit sample gas cannot be collected without breaking the seal.
The test coordinator must interpret the audit results. Indication of acceptable
results may be obtained by the testing firm immediately following analysis by telephoning
the responsible agency with the audit and compliance test results in ppm C. The testing
firm must include the results of both audit samples, their identification numbers, and the
analyst's name along with the results of the compliance test samples in the appropriate
6-18
-------
reports to the EPA regional office or other appropriate agency during the 30 day period
following the test.
When the measured audit concentration agrees within 20 percent of the true
value, the audit results are considered acceptable. Failure to meet the 20 percent
specification may require reanalysis of the audit samples and compliance test samples,
reauditing, or retests until the audit problems are resolved. However, if the audit results
do not affect the compliance or noncompliance status of the affected facility, the agency
may waive the reanalysis, further audits, or retest requirements and accept the results of
the compliance test For example, if the audit results average 38.6 percent low, the
compliance results would be divided by (1 - 0386) to determine the correlated effect. If
the audit results average 58.3 percent high, the compliance sample results would be
divided by (1 + 0.583) to determine the effect. When the compliance status of the
facility is the same with and without the correlated value,: then the responsible agency
may accept the results of the compliance test. While steps are being taken to resolve
audit analysis problems, the agency may also choose to use the test data to determine
the compliance or noncompliance of the affected facility.
The same analyst, analytical reagents, and analytical system shall be used for
analysis of the compliance test samples and the EPA audit samples; if this condition is
met, and the same testing firm is collecting other sets of compliance test samples,
auditing of subsequent compliance analyses for the same agency within 30 days is not
required. An audit sample set may not be used to validate different sets of compliance
test samples under the jurisdiction of different agencies, unless prior arrangements are
made with both agencies.
During the audit, the test coordinator should record the coded cylinder number(s)
and cylinder pressure(s) on the "Field Audit Report Form," Figure 6.8, page 6-32. The
individual being audited must not be told the actual audit concentrations or the
calculated audit percent accuracy.
On-site Collection of Audit Sample(s) - The cylinder gas performance audit
sample collection must be conducted in the field (not laboratory) at the same time the
compliance test samples are being taken. A maximum of 5 liters of audit gas is to be
used for each test run unless multiple sample tanks are required for sampling. The
testing firm is required to supply a 2 stage regulator (CGA - 350), a glass manifold or
Teflon tee connection, and other suitable Swagelok fittings (they are not supplied) for
use with the audit gas cylinder. The recommended procedures for conducting the on-site
audit sample collection are as follows:
1. The test coordinator should verify that the seal affixed to the audit cylinder
by the shipping laboratory is still intact. After the seal has been checked
by the test coordinator, the testing firm may break the seal. However, if
the test coordinator is not present at the time of the audit, the testing firm
may break the seal and proceed with the audit.
6-19
-------
2. The tester should set up the Method 25 sampling train and perform the
leak check.
3. The audit gas from the cylinder has to be sampled at atmospheric pressure
either from a glass manifold or through a Teflon tee connection. This can
be done by attaching both the cylinder and the probe of the Method 25
sampling train to two of the manifold or tee connections while excess gas
flows out through the remaining connection as shown in Figure 6.9,
page 6-33. This can be accomplished by starting the cylinder gas flow into
the manifold or tee with the sampling train flow turned off. Then, turn on
the sampling train flow while adjusting the flow from the audit cylinder to
ensure excess audit gas flows from the manifold or tee. After the proper
''' sampling flow rate has been obtained in the sampling train, adjust the audit
cylinder so only a few cubic centimeters of excess gas is discharged from
'"'~" "~ the manifold or tee. The testing firm must ensure that the audit gas is
conserved.
'"""'"' 4. Use the same sampling flow rate and sample volume as used for the field
test samples. When a constant flow rate can no longer be maintained by
the sampling train, it should be turned off and then the audit cylinder shut
off. Ensure that the audit cylinder is closed tight to prevent leakage. If
the compliance test requires more than one sample tank to complete a run,
each audit sample should use the same number of tanks required by the
average run.
5. The same procedures are repeated for the second audit cylinder using a
separate sampling train.
6. The sampling trains containing the audit samples should be recovered and
shipped in the same manner as and along with the field test samples.
7. In all cases, it is recommended that the test coordinator reseal the audit
cylinders to ensure against tampering. However, if the testing firm is to
return the cylinder to the EPA/QAD contractor, it is mandatory that the
audit cylinders are resealed by the test coordinator.
8. The audit cylinders are to be shipped back immediately after the test to
the EPA/QAD contractor at the cost of the responsible agency or the
testing firm either by ground transportation or air cargo. They are not to
be shipped collect.
Analysis of Audit Sample(s) - The collected audit sample fractions (condensate
trap and evacuated tank) are analyzed at the same time as the Method 25 compliance
test samples. Follow the procedures described in Method 25 for sample analysis,
calibration, and calculations. The same analysts, analytical reagents, and analytical
system shall be used for both the compliance test samples and the EPA audit samples.
Reporting of Audit Sample(s) Results • The audit sample results are to be
reported to the responsible agency by the testing firm in terms of condensibles
(condensate trap fraction), noncondensibles (tank fraction), and total (sum of both
6-20
-------
fractions) as ppm C. The agency will in turn report the results to the EPA/QAD
contractor for continuing evaluation of the Method 25 audit program. The testing firm
must also supply the agency with the results of both audit samples as described above,
their identification numbers, and the analyst's name along with the results of the
compliance test samples in written reports to the EPA regional office or the appropriate
agency during the 30 day period.
Performance Audit of Data Processing - Calculation errors are prevalent in
processing data. Data processing errors can be identified by auditing the recorded data
on the field and laboratory forms. The original and audit (check) calculations should
agree within round-off error; if not, all of the remaining data should be checked. The
data processing may also be audited by providing the testing firm with specific data sets
(exactly as would appear in the field), and by requesting that the data calculation be
completed and that the results be returned to the agency. This audit is useful in
checking both computer programs and manual methods of data processing.
6.8.2 Systems Audit
A systems audit is an on-site, qualitative inspection and review of the total
measurement system (sample collection, sample analysis, etc.). Initially, a systems audit
is recommended for each compliance test, defined here as a series of three runs at one
facility. After the testing firm gains experience with the method, the frequency of
auditing may be reduced - for example, to once every four tests.
While on site, the auditor observes the testing firm's overall performance,
including the following specific operations:
1. Setting up and leak testing the sampling train.
2. Collecting the sample at a constant rate at the specified flow rate.
3. Conducting the final leak check and recovery of the samples.
4. Sample documentation procedures, sample recovery, and preparation of
samples for shipment.
6-21
-------
REGULATING
VALVE
TEMPERATURE
CONTROLLER
DUAL RANGE
MANOMETER
VACUUM PUMP
PURGE VALVE
THERMOCOUPLE
SAMPLE
VALVE
Jl
STAINLESS STEEL
FILTER HOLDER
HEATED BOX
FLOW
CONTROL
VALVE
ROTAMETER
STAINLESS
STEEL PROBE
•CONDENSATE
TRAP
SAMPLE
TANK
Figure 6.1. Method 25 sampling train.
6-22
-------
TABLE 6-1. METHOD 25 SAMPLING EQUIPMENT COMPONENT
AND CALIBRATION SPECIFICATIONS
Apparatus
Limits
Frequency and Method
Sample Tank
Volume
Sampling Train
Volume
±5 cc or 5 g
No Limit
(suggest ±2 cc)
Rotameters
0.9 to 1.1
Thermometers
Within 3°C
(5.4'F)
Barometer
Within 2.5 mm Hg
(0.1 in Hg)
Initially
Initially and after
replacing any
component; should
have condensate trap
installed during
calibration.
Initially and whenever
calculated sample volume
does not match volume
expected from sample
flow rate (deviation
of > 10%)
Calibrate against
mercury-in-glass
thermometer and in
boiling water;#•
check at ambient temp
before each test;
recalibrate if ambient
check is outside limit
Check against mercury-
in-glass barometer or
National Weather Service
value. Check before and
after each test.
6-23
-------
VACUUM PUMP
CONNECTOR
SAMPLE
SHUT-OFF
VALVE
25.4
1.0
RBERFAX
INSULATION
DIMENSIONS:
PROBE
CONNECTOR
m
m
3.1^5
0.125
CONDENSATE
TRAP PROBE
BULKHEAD
CONNECTOR
D
PROBE LINE
THERMOCOUPLE
TO TEMPERATURE
CONTROLLER
a a
FILTER HEAT CONDENSATE
TEMPERATURE TRAP PROBE
CONTROLLER " CONNECTOR
THERMOCOUPLE THERMOCOUPLE
Figure 6.2. Method 25 filter housing.
6-24
-------
: 0.375
316 SS NUT
DIMENSIONS:
inn
in
-WA11 :
WALL
COARSE QUARTZ
WOOL PACKING
225
Figure 6.3. Method 25 condensate trap.
6-25
-------
CARRIER GAS
CALIBRATION STANDARDS-
SAMPLE TANK-
SAMPLE
INJECTION
LOOP
INTERMEDIATE COLLECTION
• VESSEL (CONDmONED
TRAP SAMPLE)
SEPARATION
COLUMN
BACKFLUSH
CO.CH4lC02
NONMETHANE
ORGAN1CS
OXIDATION
CATALYST
REDUCTION
CATALYST
i
FLAME
IONEATJON
DETECTOR
1
•HYDROGEN
COMBUSTION
AIR
DATA RECORDER
Figure 6.4. NMO analytical cycle.
6-26
-------
COLUMN OVEN
Figure 6.5. NMO sample delivery schematic.
6-27
-------
H
1
g
U
Z
o
H
£•<
£
n
H
j
8
H
1
CJ
I
\o
o
0
0*
0
fa
1^
o
0
u
c
o
•H
.p
id
M
43
•H
3
i Q
i}
i
•P431-I
O (d
C «J C
H 0 «J
U
C
0
•H
O
IH
id
u
» I
s s
I J Jl Jf
05 0) jS
-P (0
§ILJ
T3 «4
•v (C
^j (^
10 H C
>i O O
09 B
^uj
>•
O
0
-H
O
•H
M
O >i
^- H
e S
0 i<
•P
M 0
^H SC
CO Z
« «
id 4->
4^ H 3
CO 0 42
+J (Q
§042
10 H C
>iO 0
(0 8
>"1
4>
•H
(4
•
id id
A
Jg
O ^J
id o
4J
(0
0
4J
0
to
C
a
(0
«
D*
(0
4J
iO O
CO E
ft
iO 0
[A £
ft
<3> 3 vo
O
C
0
•H
U
•rl
IH
«H
0
c
o
•H
jj
I
S
dP
If)
at
At
4J
10
l-l
id
4J
id
y
c
o
•H
O
%
0
K
id
JJ
M
0
to
a
a*
0
o
c
id
M
O
O
N« ilf
e
0
4J
CO
^"f
to
1
M 3
(043
H C
O O
B
ft
3 «
O
0
6-28
-------
FLOW METERS
HEAT TRACE (100"C)
SAMPLE
RECOVERY
VALVE
FLOW
CONTROL
VALVE
SYRINGE PORT
VACUUM PUMP
Figure 6.6. Condensate recovery system.
6-29
-------
CONNECTING T
INJECTION
SEPTUM
FROM ,
CARRIER^1
DIMENSIONS: "
CONNECTING ELBOW
SS TUBING
TO
CATALYST
Figure 6.7. Liquid sample injection unit
6-30
-------
n
VD
Ed
CO
tr>
c u
n
C 0
o a SL,T\ o
0 0 0 0 0 U
OS 4J M k J-i <0
D> to
C JJ 0
-H -H 3
U ' «C 0 H
3 3 b <0
c a
00 0 E 0
U E fc 0 3
C 0 3 O J-I
o o * to jj
j-i jj nj -a
» 0 M 0 C 0
>ji U O M -I
D J-i J-i £ a 3
0 0 3 JJ E M
J-i > O 0 (9 0
fc* 0 tn E to j^
to
0
^
0 c a
> m E
•H .C (0
JJ 4-> JJ 10
(0 -H
•H *O tO £
0 3 <0 JJ
S-l i
<«-i -H fH
J-i "O RJ
0 3 C
a- < <
jj 0
U £ ffl .
01 JJ jj
U g AT)
O 0 -O
o i-i o -a *o
i O O 3
D J-i J-i £ O A
0 0 3 JJ H JJ
J-i > O 0 « -H
0
^ 0
0 Jw ^H
_f* yri u_i
O fl O
I
13 10 T3
•c c c
10 O 3
•H 0
•H JJ i*
(0 A
C .H C
•H 3 -H U
tyi O £ O
•H ^H JJ Jj
k (0 -H J^
O 0 S 0
cr>
c
•H
(0
to
0
U
0
& tO
^
(0 O
JJ J-I
(0 J-l
o u
1
0
JJ-H 1 1 0
tn > o jz jj
0 0 0 u o co
JJ T3 Jj 0 C •
4J IO
0 J-l E -O
JJ-H OT3-C 0
0 J-l 0 (0 }J •
c x! m.J£ .
e 0
O O * 9*4-1 0 Q 0
Jj JJ tO JC
.. 0 M 0 >i J3 ffl V to
>4iUb. O C 3 S-l
D h J-l .0 C 4J.JS JS.-4J-3
fli ^ 3 C^ Q) CO ^^ fj ^^
^j ^ O X ^2 ' fl) GJ flj ^^ *^H
^y fl} tJl Q) ^J ^^ ^P ^j ?5 C^
*O
1 0
£J2
O-H J*
Qj ^ flj
jj 0 4J •
to a,
H 0 10
0 -O JS -- •
C 0
o to
•H 1C M
JJ -H
0 0 J2
U 3 JJ
0) D1
Cb'H C
O C-H
1
1
jj
•H 0
•0 3
3 0 O1
< O -H
C C
tD (0 £
E > 0
fl) J^ 0
JJ 0) JJ
(0 10
MOO
.0 E
a w
o cu
.0)
(0
01 0)
83 E
-H 0
-Jj O
o «w 3
o c o
a, c
6)
fl)
.J5
(0
0
.
o o a
.:«J
JJ (Q
C 0 0
a> to j=
e o jj
a) a
O ^ >i
M 3 J3
O A
~«w •'O
.e )-> a)
. aro c
IH--H
-h 6
.O 10 tj
.jw^J 01
10 JJ
-JJ 0 0
-ffl
.O O
-.3 0
0
O
U
-- IM
a 0
O 0
• - jj j:
flj 3 jj
14 O
Ot O4J
JC ffl
0
"0 C»O
0 O 0
C -P
«w tf'O
0 3
*H3 ffl iff
030
6-31
-------
Part A. - To be filled out using information from organization
supplying audit cylinders.
1. Company supplying audit sample(s) and shipping address
2. Test coordinator, organization, and phone number
3. Shipping instructions: Name, Address, Attention
4. Guaranteed arrival date for cylinders -
5. Planned shipping date for cylinders -
6. Details on audit cylinders from last analysis
d. Audit gas (es) /balance .gas. .
Low cone.
N2
Aluminum
High cone
N2
Aluminum
Part B. - To be filled out by test coordinator.
1. Process sampled
2. Audit location
Name of individual audit
Audit date
3.
4.
5. Audit cylinders sealed
6. Audit results:
b. Cylinder pressure before audit, psi. .......
c. Cylinder pressure after audit, psi. ........
d. Measured concentration, ppm C
U— tube fraction ...........................
TanJc fraction .............................
Total concentration .......................
e. Actual audit concentration, ppm C
f. Audit accuracy:1
Low cone . cvlinder .........................
Hi ah cone, cvlinder ........................
Percent accuracy =
Measured Cone. — Actual Conc^ x 100
Actual Cone.
9 Problems detected ( if anvl .................
Low
cone.
cylinder
High
cone.
cylinder
, —
aThe audit accuracy is calculated on the total concentration.
Figure 6.8. Field audit report form.
I
6-32
-------
EXCESS
FLOW
ROTAMETER
ROTAMETErl
TEFLON TEE*
OR MANIFOLD
FLOW
CONTROL
VALVE
MANOMETER
i> r—i
«»m ^ " ^ i
SAMPLE
TANK
VALVE
CONDENSATE
TRAP
AUDfT
CYLINDER
Figure 6.9. Schematic of Method 25 audit svstem.
SAMPLE
TANK.
6-33
-------
-------
CHAPTER?
TOTAL GASEOUS ORGANIC CONCENTRATION USING
A FLAME IONIZATION ANALYZER - METHOD 25A
7.1 APPLICABILITY
Method 25A is designed to measure Total Hydrocarbon (THC) concentrations in
flue gas streams. Total hydrocarbons are measured by a flame ionization detector (FID)
on a continuous, real-time basis. No compound separation takes place, so Method 25A
is noncompound specific, like Method 25. Not all hydrocarbons are suitable for analysis
by FID. Highly substituted or halogenated hydrocarbons, in particular, are not amenable
to FID analysis. In general alkanes, alkenes, and aromatics are the most appropriate
compound groups for Method 25A sampling and analysis. Method 25A results are
measured on a wet basis and the concentrations must be adjusted for the percent
moisture in the sample gas stream for the purpose of emissions calculations.
Method 25A has some advantages over other methods used for measuring organic
compounds. Specifically, the data is generated continuously in real time. Method 25A
can only be used in situations where an appropriate response factor for the stack gas can
be determined. In gas streams that cannot be characterized or which have changing
composition, the response factor for the stream cannot be determined; Method 25A is
not applicable for such a gas stream.
Method 25A can be used in other situations which require a response factor to be
established or estimated, for instance, a surface coating operation which uses which four
solvents. The amount of each solvent used as a percentage of the total solvent usage is
known. A standard is prepared using the weighted percent of each solvent used in the
coating operation. The concentration in ppm as carbon is calculated for the standard.
The standard is introduced to the sampling system and the response factor is calculated.
The analyzed concentrations from the emission source are then adjusted by dividing by
the response factor.
7.2 METHOD DESCRIPTION
Method 25A sampling is performed continuously using a THC monitor. Sample
gas is extracted from the emission source from a single point in the duct and pumped to
the monitor at a constant rate. The sample is analyzed by an FID detector and the
resulting electrical signal is proportional to the carbon content of the sample stream
passing through the detector.
The sample gas is pumped through a Teflon line 1/4 inches to 3/8 inches in
diameter from the sampling point to the analyzer. The line and the sample pump are
heated to prevent condensation of water vapor or hydrocarbons. The analyzer and the
detector are also heated. All parts of the sampling system are kept above 121°C (250°F).
7-1
-------
The sample is pumped to the detector at a constant pressure, which is monitored by a
gauge on the analyzer. The analyzer is calibrated and operated at the same sample
delivery pressure to insure equivalent instrument response for the same concentration.
The FID consists of a flame, fed by hydrogen and air, a collection ring, and an
amplifier for the electrical signal. A small portion of the sample gas extracted from the
duct is passed through the flame using a capillary tube to restrict the flow. If the sample
delivery pressure is not kept constant, the flow rate through the capillary is not constant,
and the instrument response will vary accordingly. As the sample passes through the
flame, the organic molecules are burned. The flame and the collector ring are
maintained with a very high electrical potential between them. As the molecules are
burned, the gas between the collector ring and the jet becomes conductive due to the
presence of ions given off during combustion. The ions carry a small current from the
jet to the collector ring. The current is proportional to the amount of organics being
combusted in the flame. The current is amplified and output as a 0 to 1 volt signal.
An FID is capable of measuring hydrocarbons in ranges varying from 0 to 10 ppm
to 100,000 ppm. The analyzer is calibrated in the range applicable to the emission limit,
given flow rate estimates. Calibration gases are prepared and the concentrations
certified by National Institute of Standards and Technology (NIST) methods or EPA
Protocol 1. The gases should be prepared such that three gas concentrations, high, mid-,
and low range, are used for each instrument range. The concentrations should
correspond to 80 to 90 percent of range limit, 45 to 55 percent of range limit, and 25 to
35 percent of range limit. The range selected (e.g., 0 to 100 ppm ) should correspond to
100 percent of full scale response at one of the available signal amplifier settings. The
highest available response on each range is the range span. The high range calibration
standard is introduced to the analyzer and the analyzer output is adjusted so that the
output matches the standard concentration. If the analyzer response is linear, the mid-
and low range gases should give responses equal to their certified concentrations.
The analyzer can be calibrated by introducing the calibration gasses directly to the
instrument (direct cal), or by introducing the gases to the inlet of the sampling system
(system cal). If a direct cal is done, a performance check must be conducted to validate
the sampling system. After calibration, the mid-level gas is introduced at the sampling
system inlet and the response is recorded at the analyzer. The instrument response
should match the calibration gas concentration. To allow calibration gases to be
introduced into the sampling system as close as possible to the sample gas inlet, a three-
way valve is installed at the back of the sample probe. The three-way valve is used to
allow either stack gas or calibration gas to be drawn through the sampling system. A
diagram of the Method 25A sampling system is shown in Figure 7.1, page 7-14. The
major components of the sampling system are listed below:
Probe • 1/4 inch or 3/8 inch in diameter. Constructed of Teflon or stainless steel.
7-2
-------
Col/Sample Valve - A three-way valve which allows the analyzer to draw from
either the sampling probe or the calibration line. This should be located at the rear of
the probe and must be heated to 250°F.
Paniculate Filter - A 25-mm glass mat filter with a cut size of-2 microns or less. It
must be heated to 250°F.
Sample Line - 1/4-inch or 3/8-inch Teflon line encased in an insulated shell with
heat tracing. It must be capable of maintaining 250°F along the entire length of.the line.
Sample Pump - Teflon-sealed diaphragm piimp capable tif pulling at Least
1 liter/minute and attaining 10 inch Hg vacuum. The pi'inp-heads .must Jse heated to at
least 250°F.
THC Analyzer - Hydrocarbon analyzer equipped a flame iomzation detector. The
analyzer should keep the sample stream above 250°F until delivered to the'FID. The
FID must be equipped with an output amplifier compatible wim^ strip chart recorder or
data acquisition system.
Recorder • A strip chart recorder or data acquisition system capable of recording
the analyzer output continuously or recording the analyzer output average response at
intervals of no greater than 1 minute. The instrument output voltage should represent
full scale deflection of the strip chart.
73 PRECISION AND ACCURACY ~
The precision and accuracy of Method 25A are defined by the calibration
parameters imposed on the tester before and after the run. Calibration error is defined
as the deviation of the analyzer response from the true value of a calibration gas.
Method accuracy is commonly determined in this manner. Therefore, the calibration
error can be used to assess the accuracy of Method 25A The specification for
calibration error is ±5 percent of the true value of the calibration gas.
Calibration drift is defined as the deviation between the value determined for a
calibration gas measured before the test run and the value determined for the same gas
after the test run. This is a measure of the reproducibility of the. analysis and represents
the precision of the method. The limit for calibration drift is ±3 percent of the span
used for analysis.
.The precision of the method should remain the same lor all volatile organic
compounds introduced to the system. However, the accuracy determination is based on
use of a single compound calibration gas. In situations that require a response .factor to
be used to adjust the analyzed concentrations, the accuracy is only as good as the
response factor. If the response factor is generated for a specific compound,-the
7-3
-------
accuracy of the response factor is determined by the care taken in preparing the
compound standard, and the accuracy of the analyzer when used to generate the
response factor. A standard of a different concentration can be used to check the
response factor by comparing the actual reading to the predicted response using the
response factor. The cumulative accuracy limits can be estimated as the square root of
the sum of the squares of contributing limits. If the analyzer accuracy is ±5 percent and
the response factor accuracy is ± 10 percent, the cumulative accuracy limit is (0.052 +
0.102)*4 = 11 percent. For emission sources where the response factor is estimated, the
accuracy is no better than the estimate of the response factor.
7.4 SAMPLING POINT LOCATION
Method 25A sampling is conducted nonisokinetically at a constant rate. A
sampling point is selected in one of three ways: (1) the probe can be placed at a single
point in the center of the duct, (2) the probe can be placed at a single point at a position
in the duct with average gas velocity, (3) or a rake-type probe may be used. If no
stratification is expected in the duct, a single point sample can be extracted. If the test
location is immediately downstream of the inlet of an auxiliary gas stream or ambient
intake, the concentration may not be constant across the duct. If this might be the case,
a rake-type probe should be used. The probe is sized to be the same length as the stack
diameter. Several holes of equal size are drilled in the side of the probe such that the
spacing between the holes is equal. No holes should be placed within an inch of the
duct walls, and it is most important that no holes extend into the port nipple or out of
the duct. The end of the probe should be blocked, or the majority of sample will enter
the probe through the end. If paniculate matter is expected to be present, the holes
should be turned away from the direction of flow to minimize plugging.
7.5 OBSERVATION PROCEDURES FOR METHOD 2SA SAMPLING
7.5.1 Leak Check
A*leak check of the sampling system is recommended prior to testing. The .
method does not require a leak check, but it is a good idea to verify the integrity of the
sampling'system before starting a test run. If the tester can show that the system
calibration has a calibration error of less than ±5 percent of the certified value of the
cal gas, he may decline to perform a leak check. If a leak check is performed, it can be
done in one of two manners. The first method is to place a vacuum gauge at the probe
tip and draw a vacuum on the system of 10 inch Hg. A valve is used to isolate the
system from the sample pump. The system vacuum should remain constant for a period
of 5 minutes. The second method is to plug the probe tip and place a rotameter or
water bubbler at the pump exhaust. The pump should evacuate the system until there is
no flow through the pump. One disadvantage to this approach is that the system vacuum
will be as high as the pulling capacity of the sample pump, while sampling should never
exceed 10 inch Hg.
7-4
-------
The leak check and other sampling procedures are summarized in a sampling
checklist (Figure 7.2, page 7-15) to be completed or used a guide by the test coordinator.
7.5.2 Calibration
Prior to the start of testing, the hydrocarbon analyzer must be calibrated. The
calibration is done by introducing the high, mid-, and low range calibration gases to the
analyzer to show that the analyzer response reflects the true concentration of the gases.
The analyzer range must be chosen so that the source THC limit is 10 to 100 percent of
the range. Preliminary traverses may be done to determine the emission source flow
rate so that the allowable concentration limit can be calculated from the mass emission
limits. The allowable concentration limit should be higher than 10 percent of the
analyzer range. If not, a lower range should be selected until the limit is above
10 percent of the range or the lowest range of the analyzer is reached. The allowable
concentration limit should not be above 100 percent of the analyzer range unless the
source concentration is expected to be less than 10 percent of the allowable
concentration limit.
The calibration gases are usually propane in air or propane in nitrogen. Some
regulatory agencies may require that the calibration gases be methane in air or nitrogen.
If the calibration gases are propane, the propane concentrations should be multiplied by
three to represent the concentration as carbon in the cylinder. If the calibration cylinder
concentrations are not adjusted to carbon concentration, the test results must later be
multiplied by a K factor of 3.0 to adjust for the number of carbons in the calibration gas.
Any gas standard may be used as a calibration standard if it is National Institute of
Standards and Technology traceable and the K factor is known. The K factor for
methane is 1.0.
The calibration gas may be introduced directly into the analyzer or through the
sampling system. A cylinder of carrier grade purity zero air containing no hydrocarbons
is introduced to the analyzer and the back pressure to the FID is set The zero offset
adjustment of the analyzer is set such that the analyzer output is 0 ppm ±3 percent of
the span. The high range calibration gas is then introduced to the analyzer. The
amplifier gain adjustment is set such that the analyzer output matches the calibration gas
certified value within 5 percent of that value. The actual analyzer response is recorded
and a calibration factor is calculated. The calibration factor is equal to the certified
cylinder concentration divided by the analyzer response in divisions:
CF = Certified Cylinder Value Equation 7-1
# Divisions Analyzer Response
If the high range cylinder value is 85.7 ppm and the instrument response is 84.6 divisions,
the calibration factor is:
7-5
-------
CF
84.6 Divisions
1/013 ppm
Division
Equation 7-2
Since the amplifier gain is set such that the recorded value is ±5 percent of the
true value, the calibration factor should always fall in the range of 0.95 to 1.05. If higher
instrument ranges are used, the factor will be some factor of ten, e.g., 9.5 to 10.5 (for a
factor of ten) or 950 to 1050 (for a factor of one-thousand).
The calibration factor is used to predict the response for the mid- and low range
gases using the following equation:
Analyzer Response = Cylinder Concentration
Calibration Factor
Equation 7-3
The actual response must be within 5 percent of the predicted response. This
procedure is a linearity check on the analyzer. A linearity check must be done once for
each set of test runs. If any range is to be used other than the one in which the linearity
check is performed then a mid-level gas in that range is introduced to confirm the
multiplier for that range. The analyzer response for the mid-level gas in each range
must agree with the cylinder value within 5 percent.
When the linearity check is done through the sampling system using the cal
sample valve at the rear of the sample probe, the integrity of the sampling system is not
in question. However, if the calibration is done directly through the analyzer, a
performance check must be conducted using the high range gas. The high range gas is
purged through the calibration line to the cal/sample valve. A T" with a rotameter on
the tap leg is placed in the calibration line before the cal/sample valve. The rotameter
allows excess calibration gas to dump to the atmosphere. The flow rate through the
calibration valve is set such that, with the sample pump on, the dump rate at the
rotameter is less than 2 liters per minute. If no dump were provided in the system, the
calibration gas would pressurize the system and system leaks would not affect the
performance check results. When the cal/sample valve is switched to sample, the system
is under negative pressure and system leaks will be indicated by the test results. If the
dump rate falls to zero flow, ambient air will be drawn in through the rotameter and the
performance check results will be low. If the dump rate is too high, the ability of the
rotameter to relieve the delivery pressure of the calibration gas will be exceeded and the
sampling system will still be under positive pressure. A diagram of the
calibration/sample valve with an ambient dump is included in Figure 7.3, page 7-16.
7-6
-------
7.5.3 Response Time Test
Response time is defined as the time required for a step change in system
conditions to show a response at the analyzer equal to 95 percent of the step change.
The response time of the sampling system must be checked to assure that the sample is
delivered from the emission source to the analyzer in an acceptable period of time. No
limit for response time is set in Method 25A. A typical response time is less than one
minute, and should not be over two minutes. A long response time indicates that the
flow rate of sample being drawn by the sample pump is not high enough for the test
conditions. Any test which requires more than 100 feet of sample line will probably
require a pump that can pull at least 1 liter per minute to keep the response time down
to an acceptable limit.
., The response time test is accomplished by first purging the sampling system with
zero air with the calibration/sample valve turned to calibration. The zero air flow is
stopped and replaced as quickly as possible with the high level calibration gas. A
stopwatch is used to record the time taken for the instrument response to equal
95 percent of the high level cal gas value. This procedure is repeated until three step
changes have been measured. The three response times are averaged and recorded as
the response time for the test series.
. • ' .«
7.5.4 Sampling Procedures . ".
Sampling is initiated by turning the cal/sample valve to the sample position and
starting the sample pump and data recorder. A period of time longer than-the response
time of the system must be allowed to purge the system with sample gas. The start time
•for the test run is marked on the data recorder. If a strip chart is used for'recording
data, a separate data sheet should be used to record one-minute averages from the strip
.chart trace. The delivery back pressure to the FID must be maintained at the same
Tahie used for calibrations. The analyzer and recorder are allowed to operate with no
adjustments except those needed to maintain FID delivery pressure for the required test
: duration. Any process changes or interruptions should be recorded on the strip chart or
in a test log.
For test runs longer than one hour, it may become necessary to interrupt the test
run to check the calibration drift and zero drift of the analyzer. No adjustments to the
output multiplier may be made between the pretest calibration and the calibration drift
check. The drift check is also performed immediately following completion of the test
run. On longer test runs, the drift check may be performed as often as every hour. If a
drift check is done and the drift exceeds 3 percent of the span value or the zero drift
exceeds 3 percent of the span value, the data collected prior to the drift check are
invalid. The system must be adjusted and recalibrated and the testing repeated. For
example, if a four-hour test run is in progress with a drift check performed every hour,
.and the drift check after the third hour is outside.the specifications, the data recorded
7-7
-------
between the second-hour drift check and the third-hour drift check is considered invalid.
The analyzer is recalibrated and the testing is resumed. The third-hour of testing is
repeated and if the hourly drift check is acceptable, the fourth hour of testing is then
completed.
Alternately, the sampling system may be recalibrated without adjustments. The
calibration factor from the new calibration is recorded and the results are reported using
both the pretest calibration factor and the post-test calibration factor. The agency will
accept the data showing worst case results.
7.5.5 .Establishing Response Factors
, Response factors can be established in three ways: (1) using a gas standard of
known concentration, (2) using a liquid standard of known concentration, and (3) using a
liquid standard of unknown concentration. The proper approach will be determined by
the specifics of each test for which Method 25A is applied. The testing firm should
insure that the procedures for establishing response factors have been approved by the
appropriate regulatory agency prior to testing. The three methods are described below.
Gas Standard - If the molar fraction of each compound present in the emissions
is known before testing, a gas standard in the same ratio of mole fractions can be
prepared for the response factor test. The concentration as carbon is calculated for the
gas standard. The gas standard is introduced to the calibrated analyzer and the analyzer
response is recorded. The response factor is equal to the predicted response divided by
the actual response.
The response factor should be established on the day of testing. An FID uses a
flame to oxidize hydrocarbons. The response factor is a function of the efficiency of the
flame to perform this oxidation. Each time the flame is extinguished and relit, its
efficiency may change. Therefore, the same flame should be used to establish the
response factor as is used for testing. If multiple days of testing are planned, the
response factor should be reestablished each day. An alternative to establishing the
response factor on the day of testing is to determine the response factor in the lab
before testing. The flame must then be extinguished and the system allowed to cool.
The analyzer is then relit from a cold start. The response factor test is performed again,
and the results compared to the previous test. If the deviation between the two response
factors is <5 percent of the first response factor, the two are averaged and the average
response factor is used for the test.
Liquid Standard of Known Composition - A liquid standard of known
composition can be used to make a gaseous standard by volatilizing a small amount of
liquid into a gas cylinder or Tedlar bag. A heated injection port is recommended to
assure that all of the liquid is volatilized. The cylinder or Tedlar bag should be filled
with a metered amount of zero grade air or nitrogen. The composition of the liquid,
7-8
-------
volume of liquid injected, molecular weight of each component-density of the liquid, and
volume of diluent gas used must be known to calculate the concentration of the resulting
gas standard. The response factor determination is then done exactly as for a gas
standard. Polar compounds such as alcohols are not stable in Tedlar bags; it is
recommended that cylinders be used for alcohols or that the response factor be
determined immediately after the standard is prepared in a Tedlar bag.
Liquid Standard of Unknown Composition - If the test program entails measuring
the emissions from a process which uses a solvent which is a complex mixture that
remains constant hi composition, a known mass of the solvent can .be used to generate a
response factor. This is not truly a response factor since there is jio^predicted response.
Normally the response factor has no units. A response factor produced from a liquid of
unknown concentration has the units of divisions fesponse/mass/vohime. Since ihe
carbon content of the liquid is never determined, this response factor cannot be used
when the emission limit is expressed as Ibs carbon, but may be used for mass balance
determination such as a capture efficiency test.
The solvent is used to make a gas sample in the same manner as described for
the liquid of known concentration. The mass of solvent used must be measured or can
be calculated if the density of the solvent is known. The total volume of the air or
nitrogen introduced into to the cylinder or bag is also measured. The mass
concentration of the gas standard is calculated by dividing the mass of solvent injected by
the volume in the container. The concentration units are mass/volume (e.g^ nag/liter).
When the gas standard is introduced into the analyzer, the response is recorded and the
response factor is calculated by dividing the divisions of response by the standard
concentrations. Since the composition of the liquid is unknown, the boiling point of the
liquid is also unknown, therefore, a heated injection port must be used to make the gas
standard. The container should be checked for condensation and the injection port
temperature should be kept above 250°F.
An example of this technique follows: The exhaust over a couniercuriEnl solvent
rinsing station is sampled. As long as the operation remains steady state, the
concentration in the dip tank remains constant, but it is difficult to estimate the
composition in the tank. A sample is taken from the tank and 10 ml is weighed to
establish the density of the solvent. 50 fil of solvent at a specific gravity of 0.88 are used
to make a gas standard in a bag containing 10.5 liters of nitrogen. The resulting gas -
standard is introduced to the analyzer and the response is 54.5 divs.
Gas Concentration = (50 u\ x 0.88 mg/al
10 J liter Equation 7-4
4.19 mg/1
Response Factor = 54.5 divs
4.19 mg/1 Equation 7-5
13.01 divs/mg/1
7-9
-------
The first test run resulted in an average instrument response of 41 divisions and the flow
rate through the exhaust was calculated to be 1000 scfm which translates to
28320 standard liters per minute. The emission rate of solvent through the exhaust is:
Concentration of Exhaust
41 divs
13.01 divs/mg/1
3.15 mg/1
Equation 7-6
Mass Emissions
3.15 mg/1 x 28320 liters/min x 60 min/hr
5352480 mg/hr x 1 lb/454000 mg
11.79 Ib/hr Equation 7-7
7.6 SAMPLING PROBLEMS AND SOLUTIONS
7.6.1 Gold Spots in Sampling System
The sampling system is heated to 250°F for two reasons: first to prevent
condensation of hydrocarbons in the sampling system. Second, to prevent condensation
of moisture in the system. Some organics are soluble hi water, and water condensed in
the sampling system could act as a scrubber causing sample loss. Also, water droplets
carried into the analyzer can cause malfunction of the gauge reading the back pressure
to the FID. Although the FID will still be functional, the faulty back pressure reading
could cause the flow to the FID to change and result in incorrect readings.
No part of the sampling system may drop below 121°C (250°F). Any part of the
system found to be less than 121°C (250°F) must be heated or replaced. Since heated
lines are insulated, it is hard to tell how hot the sample line is at any one point. One
quick check is to disconnect the sampling system at die entrance to the analyzer. With
the lines heated and the sample pump operating at its normal flow rate, the exit
temperature from the pump is measured and should be a 121°C (250°F). Some other
things to-look for are: (1) temperature drop in unions between two lengths of sample
line, (2) unheated or uninsulated Teflon showing anywhere hi the system, (3) sudden
concentration spikes from what should be a steady state process, and (4) an inadequately
heated filter or calibration/sample assembly.
Two inches of unheated stainless steel or four inches of unheated Teflon is
enough to cause condensation in a sample line. Unions between lines should be too hot
to touch or they should be wrapped with a heat tape to keep them above 121°C (250°F).
The filter and calibration/sample assembly should also be wrapped in heat tape. Sudden
concentration spikes which cannot be explained by process changes may indicate
moisture condensation in the line which is passing through the back pressure regulator in
droplets.
7-10
-------
7.62 Sampling System Leaks
Anytime a performance check or calibration drift check is conducted and the
results are lower than expected, the cause could be a sampling system leak. If the cause
of the low value cannot be found, a leak check should be done on the system. If a leak
is found, the results from the preceding test run should be invalidated. A leak cannot be
considered constant, and the results of the preceding test should not be reported using
either the pretest or post-test calibration.
7.6.3 High Moisture
Because the THC analyzer operates under positive pressure and because the back
pressure gauge is mounted on the front of the analyzer and acts as a cold sink, a high
moisture content in the sample gas can cause problems using Method 25A. The oven
temperature of the analyzer can be turned up to help alleviate the problem. However,
the oven temperature should be set before the calibration. Moisture condensing in the
pressure gauge would be part of the bypass stream and would not be considered a loss of
sample. However, this moisture will affect the reading on the gauge. The gauge may be
replaced by running a line to a mercury manometer. The line to the manometer should
be heated. A small amount of condensation in the manometer will not greatly affect the
reading due to the difference in density between mercury and water. A column of
13.6 inches of water would have to collect in the manometer before affecting the
manometer reading by one inch of Hg.
Method 25A may not be applicable for testing emissions containing more than 40
percent by volume of moisture. These situations should be reviewed on a'case-by-case
basis to determine the most appropriate method of testing. •—
7.6.4 Adjustments to Gain or Zero Offset
When multiple test runs are being done, the calibration drift is checked after each
run. If the calibration has drifted, but is still within the limits, the run data is acceptable.
The tester may want to readjust the output gain to eliminate the drift. If the pot is not
reset, the drift may exceed the limit by the end of the following run. This may also be
true for the zero offset
The gain or offset may be adjusted by the following sequence:
1. The results of the calibration drift check and zero drift check performed
prior to adjustment are recorded as the post-test drift check for the
preceding run.
2. The gain and offset may then be adjusted.
3. The analyzer calibration and zero offset are checked again. The results are
recorded as the calibration and zero offset pretest check for the next run.
7-11
-------
7.6.5 High THC Concentrations
Some hydrocarbon analyzers can be used to measure concentrations up to
10 percent by volume as carbon. Many, however, are not linear above concentrations of
4 to 6 percent carbon by volume. When emission concentrations at a facility are higher
than this, two strategies may be used.
The first strategy is to dilute the sample before introducing it to the analyzer.
Any of a variety of dilution techniques may be used. The dilution ratio of the sampling
system must be calibrated after the analyzer calibration is completed. A high range
calibration gas of the same compound used for the analyzer calibration is introduced to
the system through the calibration position of the calibration/sample valve. The dilution
ratio is the ratio of the known calibration concentration and the analyzer response.
Propane standards are available at concentrations as high as 8 percent (24 percent as
carbon), or liquid propane tanks can be used to provide 100 percent propane gas. If
8 percent propane is used as the dilution standard and the instrument response were
5300 ppm as carbon, then the dilution ratio is:
Dilution Ratio
240.000 ppm C
5300 ppm C
45.3
Equation 7-8
The second strategy is to reduce the flow rate of the sample to the FDD. This will
extend the range in which the FID is linear. Conversely, the sensitivity of the FID will
be reduced. The flow rate can be changed most easily by installing a smaller capillary to
the FID. Using the same back pressure, the flow change through the capillary is
proportional to the ratio of the square of the capillary diameter. The analyzer must be
calibrated after changing the capillary. The calibration gases must bracket the expected
concentrations from the emission source. Since propane standard upper concentrations
are limited, methane or ethane would provide a more appropriate calibration standard.
7.7 AUDITS
A performance evaluation audit is not required by Method 25A because the
calibration gases must be NIST traceable and serve as an audit each time a system
calibration is performed. The testing firm should not, however, refuse an audit if the test
coordinator deems one necessary.
The audit material must be limited to the gas compound in the calibration
standard(s). Also, the audit sample concentration must be in the range used for the test
runs. In order to assure that the audit samples are appropriate, the test coordinator
must contact the testing firm prior to the test and inform them that an audit will be
performed. The compound used in the standards and the calibration ranges expected to
be used during testing must be identified.
7-12
-------
The audit gases must be prepared according to MIST guidelines or EPA
Protocol 1 guidelines. The cylinder concentrations should be certified to ±2 percent of
the tag value. Three to six weeks should be allowed for preparation and shipment of the
audit cylinders.
The audit gas is introduced into the sampling system in the same manner as the
calibration gas used for calibration drift checks.
No limit is specified for audit accuracy, however, the audit accuracy should fall
within the 5 percent limit imposed on calibration error. The tester may be allowed to "^
reset the output gain and zero offset before the audit if the calibration drift and zero "'
drift checks have been performed and recorded for the previous runs.
7-13
-------
7-14
-------
Test coordinator complete once per test series.
Check if acceptable; "X" if not acceptable.
Leak Xest
Was a Leak Check done
TIES What was the Leak Rate .intended mow Rate
NO _____ See Performance Check
Calibration
Is applicable limit >10% of analyzer range <100% analyzer .range
Are calibration gasses NBS traceable or EPA Protocol 1
What gae is used as Calibration Standard "__
What is the K factor Methane = 1 Propane » 3
Ethane = 2 Butane —-4
Was calibration direct or system cal
High Range Calibration Gas Tag Value Analyzer response .±5%
Mid-Range Gas Tag Predicted Response ..-Analyxer .Response ±5% __
Low Range Gas Tag Predicted Response Analyzer Response ±5%
Performance Check - Not Necessary if Calibration was done Through System
Tag Value of Calibration Gas
Analyzer Response
Analyzer Response ± 5% of Tag Value
Response Time Test
Three repetitions performed
Average Response Time < 2 mini
Post-test Calibration Checks
Calibration Drift Check - Pretest Response Post-test Response
Deviation s 3% of Span
Zero Drift Check - Pretest Response , Post-test Response
Deviation S 3% of Span
Response Factor
Was a response factor used
How was the Response Factor Generated Gas Standard
Known Liquid Standard
Unknown Liquid Standard _____
What is the Response Factor Units of Response Factor _____
Unitless if Gas standard or Liquid Standard used.
. Units of mass/volume/division if Unknown Liquid Standard used.
Figure 7.2. Method 25A sampling checklist
7-15
-------
SAMPLE
PROBE
STACK
HEATED
THREE-WAY
VALVE
iu
z
ti|
E
<
ZJ
CO
I
/J
^
AMBIENT
DUMP
H ROTAMETER
CALIBRATION
GAS
SAMPLE
TO ANALYZER
Figure 13. Calibration/sample valve assembly with ambient dump.
7-16
-------
CHAPTERS
REVIEW PROCEDURES FOR VOC TEST REPORTS
The compliance test report is written by the testing firm and submitted by the
facility. It is typically the test coordinator's responsibility to review the report. Some
states lack written guidelines that specify the compliance test report format and
data reporting requirements. This chapter provides a standardized report
format that can be used by the testing firm in the preparation of the compliance test
report and instructions on compliance test report review for the test coordinator.
8.1 VOC COMPLIANCE TESTING REPORT FORMAT
EPA's NSPS does not have or need a deadline for submission of the report after
completion of the compliance testing, since there is a deadline for submission of the
report without regard to timing of the compliance test. Unless the state and local
agencies have a deadline for VOC compliance test results to be submitted to the
responsible agency by the facility representative, it is recommended that a sixty (60) day
limit from completion of the field work is recommended. The report should include, but
is not limited to, the following:
1. Basis format and information shown in Figure 8.1, Page 8-8.
2. Certification by testing firm representative stating that sampling
procedures, analytical procedures, and data presented in the report are
authentic and accurate.
3. Certification by testing firm representative (preferably by a professional
engineer) that all testing details and conclusions are accurate and valid.
4. Certification by a facility representative that the facility process data are
correct to the best of his/her knowledge.
5. Calculations made using applicable equations shown in the applicable
method. An example calculation should be shown for one run,
6. Final results presented in English and metric units and containing two
significant digits for each run. Values may be rounded off to three
significant digits after calculation of each equation and two digits for the
final results or all digits may be carried in the computer and rounded to
two significant digits only for the final results. All rounding of numbers
should be performed in accordance with the ASTM 380-76 procedures.
8.2 REPORT REVIEW
The following discussion of report review procedures assumes that the test
coordinator has (1) reviewed a written test protocol, (2) stated the requirements of the
compliance test to the testing firm and facility, (3) conducted a pretest survey of the
facility, (4) observed facility operations during the test, (5) observed the sampling
procedures during the test, and (6) required the testing firm to conduct a performance audit.
8-1
-------
Figure 8.2, page 8-10, is an example compliance test report review form. The test
coordinator may complete the form or use it as a guide for the report review. This
report review form contains three sections: "Report Contents," "Report Comments," and
"Summary of Results." The first section of the form is presented in order of the report
format. The test coordinator should verify whether each item was included by checking
"yes," "no," or "OK." "OK" means the information is not complete, but it does not effect
the report review. When the "no" or "OK" column is marked, the test coordinator should
state the reason for that mark. Deficiencies should be noted and specific
recommendations and conclusions made in the "Report Comments" Section. The test
report should be reviewed in its entirety.
The following subsections provide some discussion of the content and review of
specific parts of the compliance test report.
8.2.1 Coyer, Certification, and Introduction
The cover and certifications are self explanatory. The following items should be
in the Introduction of the report.
Test Purpose - The purpose for testing will typically be to comply with one or
more of the state or federal regulations. A complete description of purpose(s) and/or
applicable regulation numbers should be given, for example: state regulation (list
regulation No.), federal regulation (list regulation No.), obtain a permit, certify a
monitor, establish an alternative emission limit, or establish a control or transfer
efficiency.
Test Location * The test location description should provide sufficient information
to ensure that there is no confusion as to which process and control equipment emissions
were tested.
Source Identification No. - Many agencies assign a source identification number
to each facility and each process or emission point in the facility. If an ED number exists,
the correct number should be used.
Test Dates, Pollutants Tested, Plant Representative Name, and Any Other
Background Information - This is self-explanatory.
8.2.2 Emission Results and Performance Audit Results
A quick review of the results, including appendices, to assure completeness of
the reported results and determine if the values seem reasonable (i.e., correct moisture,
temperature, and pressure corrections) may be helpful. Report revisions should be made
by die testing firm.
8-2
-------
Comparison of compliance test emission results with screening method emission
results, if performed, should reveal obvious errors.
The performance audit results should be compared to the EPA audit sample
values. For Method 18, audit results should fall within 10 percent of EPA audit sample
values. This method allows for discretionary acceptance for any audit value which does
not fall within the 10 percent range based on the effect on the facility's compliance
status.
The is another consideration in interpreting results form analysis of EPA audit
gases. Acceptable results on EPA performance audit gases does not ensure correct field
sample results. The atmosphere surrounding any compound of interest is referred to as
a "sample matrix." The sample matrix for audit gases is typically pure nitrogen. The
target compound in a stack gas will have a more complicated matrix. Occasionally there
is a component in the stack gas matrix which causes an analytical interference. If this
interferant has the same retention time as the target compound (the volatile organic
compound found in the audit gas cylinder), then the analyzed audit gas concentration
could be correct, while the additive effect of the interferant compound would cause the
analyzed concentration of the compound in the stack gas to be biased high. The target
compound and the interferant compound would create a single peak. The concentration
value supplied by the area of this peak, would not be representative of the true
concentration of either component.
EPA audit gas "true value" results are accurate to about ±5 percent. EPA audit
materials are obtained from and analyzed by an EPA contractor. Yearly audits of these
contractors and their audit materials have shown error and Variance as great.as
5 percent. EPA audit reports issued annually summarize all audits conducted, and when
necessary, assign new values to each audit cylinder. The test coordinator can use the
EPA audit summary report to assess the variation in the reported audit value for the
actual cylinder used. The test coordinator may make the acceptability criteria the 10
percent allowable error plus the deviation in EPA audit values for that cylinder.
Audit sample results should never be used to correct field data results. To
determine the effect the audit results have on the compliance status of the facility, the
test coordinator should mathematically adjust the field results to determine if the
compliance status would change. For example: the allowable emissions were 100 ppm,
the average emission results were 80 ppm, and the audit results were 25 percent low (75
percent of the true value). Corrected, the 80 ppm emission (results which are in
compliance) would become (80 divided by 0.75) 107 ppm, which shows noncompliance.
The test data, for this example, should be rejected because the audit results were outside
of the acceptable range and the results of the audit samples affect the compliance status
of the facility. Use of the same values with audit results with 25 percent high would
yield adjusted data showing the facility in compliance, but by a greater margin. The test
coordinator could accept the results, if desired.
8-3
-------
Process and Control Equipment Data - The test coordinator can review his field
observation notes to ensure that the reported process data are consistent with agency
observations. If data regarding raw materials, product, or collected materials are related
to determination of compliance, they should also be presented.
Allowable Emissions - The facility should restate the emissions standards. This
demonstrates an understanding of the applicable regulation(s). It is beneficial for the
test coordinator to make an independent assessment of the allowable emissions. Any
differences should be noted and the report corrected accordingly.
Discussion of Errors (real and apparent) - This discussion is important to the
future validity of the data. If the testing firm and the facility state that the data is valid,
then they-will not be able to easily discredit the data at a later time. Testing firms do
not like to admit errors. However, when they do state that the data is invalid, the
agency typically requires a retest. If no errors were noted by the testing firm or facility,
it should be recorded in the report that "no errors were noted." If errors are noted, the
test coordinator writes his observations regarding these errors on the review form.
8,2.3 Facility Operations
Facility operations may be critical in establishing compliance with the emission
standard(s). The following items should be evaluated.
Description of Process and Control Equipment - The process and control
equipment descriptions in the test report must be sufficient to (1) identify the process
and control equipment in comparison to the permit to operate and (2) allow the agency
enforcement inspector to determine if the facility does make major modification to
process or control equipment.
Process and Control Equipment Flow Diagrams - The report should include a
flow diagram which shows all ducts leading in and out of the process and control
equipment of interest and any auxiliary systems that can be enabled or disabled. The
diagram will allow the test coordinator to determine whether the system has been
modified since it was last tested.
Process Parameter, Materials and Product Results (with example calculations) -
The process operating parameter, raw materials, and product results should be presented
as required to determine compliance with applicable regulation(s). Some facilities
cannot directly determine process rates and use indirect measurements with assumed
factors to calculate process rates. For example, if a facility assumes that the final
product contains 5 percent of the organics input as residues, then they must state
whether they consider this to be a constant factor or a variable factor. The effect on the
compliance status may be small, but it is to the advantage of the agency that the facility
not be allowed to change their assumptions to fit their needs at a later date.
8-4
-------
Calculations should be checked to ensure that reported values include the proper
assumptions, are based on the correct equations, and are~mathematicaliy correct. The
test coordinator may wish to assign a confidence limit to each value.
Describe Process Operations Tested - If the process is cychc or a batch operation,
the report should describe the portion(s) of process cycle covered by the testing and the
rationale for selecting them. This is important because facility could stretch out a cycle
to reduce emissions during sampling. When cycles .of operation or batches are
compressed or reduced, emissions are typically higher. The portion of the process cycle
tested should be approved by the agency prior to sampling.
Representativeness of Process Parameters, Raw Material, and Products - It is
preferable if the facility representative or testing firm state that the process parameters,
raw materials, and products were representative. If this information Js.notistated.ia a
report showing noncompliance, facility representative may later claim that the facility
was tested under conditions of upset or malfunction.
Two items not typically presented in a compliance test report^tiiat should be
determined during the pretest survey are (1) the normal process maintenance schedule
and (2) how malfunctions are handled. These may be specified in the permit to operate
and should be checked by inspectors in the future.
Describe Control Equipment Operations Tested - Many VOC control systems
involve cyclic operations. For example, two carbon bed absorbers in parallel may be
used as a control device. One is cleaned while the other is collecting emissions.
Depending on the methods of bed cleaning and switching from one bed to^the other,
emission rates may be the highest at the beginriing or end of the cycle. Removal or
recycling of the collected material is another area of concern. The faculty should not be
allowed to change their cycle of operation or method of removing collected .material for
the compliance test.
Representativeness of Control Equipment Parameters and QUtected^Materials -
The same concerns apply here as for the process parameters.
8.2.4 Sampling and Analytical Procedures
Sampling Port Location(s) and Dimensions of Cross-section -Ensure that the
sampling location drawings are accurate and that the mass emission.rates are directly
related to the area of the stack.
Sampling Point Description (including labeling) - Check on the accuracy of the
description.
8-5
-------
Brief Description of Sampling Procedures (including equipment and diagram of
the sampling train) - Hie EPA Methods include many options regarding equipment and
procedures. For example, Method 18 sampling can be performed by at least six
approaches. A detailed and accurate diagram of sampling equipment and a description
of the sampling procedures used should be included in the compliance test report
Deviations of Sampling Procedures from Method -The compliance test report
should address any deviations (planned or not) from the standard method including an
explanation justifying the deviation(s). Planned deviations to the standard test method
should obtain pretest approval from the agency.
For accidental deviations, the report should provide a description of the deviation
and expected effect When appropriate, the test coordinator should comment on what
effect the accidental deviation had on the validity of and the degree of confidence in the
emission results. If the test coordinator does not feel qualified to make this
determination, the VOC emission measurement contact for the agency can be consulted.
Brief Description of Analytical Procedures (including calibration) - The analytical
equipment (e.g., the detector and column), analytical conditions (e.g., oven temperatures,
isothermal analysis, program temperature ramping analysis, gases, and fuels), calibration
procedures, and calibration materials should be described. The report cannot simply
state that Method 18 allows at least five approaches for preparing the calibration gases.
Analytical Deviations from Standard Method - See the discussion above on
sampling procedure deviations.
8.2.5 Compliance Report Appendices
As previously noted, the test coordinator may find it more productive to review
appendix material first The essential elements of the appendices are discussed below.
Complete Results with Example Calculations - Example calculations should be
provided in the report showing the individual equations and input data. These
calculations should be checked, with special attention given to all assumed values or
factors used to determine emissions presented in the test report (e.g., molecular weight,
moisture content, and correction factors). The testing firm should include raw data,
equations, and calculations for all measurement and calibration procedures in the report.
Raw Field Data - Check the field data for completeness. It is a good policy for
the test coordinator to sign all completed field data forms while in the field. This
practice discourages the alteration of raw data by the testing firm.
8-6
-------
Laboratory Reports, Signed Chain-of-Custody Sheets - Acceptable audit sample
results ensure that the proper detector was used and that the calibration gases were
prepared correctly. When acceptable audit sample results are obtained, errors in
compliance sample results are generally biased positive to positive interferences.
The test coordinator should refer to the appropriate chapter in this manual (e.g.,
Chapter 4 for Method 18) to determine exact procedures for validating the analytical
results for a specific method. Although it is the responsibility of the analyst to identify
and eliminate interferences, the test coordinator may wish to compare the retention
times, peak resolution, and peak shapes for the calibration standards, audit samples, and
field samples.
When performance audit samples are not used in the test program, the test
coordinator will need to conduct a evaluation of the calibration standards used. When
no audit samples are analyzed, all test results are based on the value assigned to the
calibration standards. Therefore, it is critical that the calibration standards are assigned
the correct values.
Calibration Procedures and Results - Calibration results are used to calculate
emission concentrations and mass emission results. Complete documentation of
sampling equipment and analytical instrumentation calibration, including complete
presentation of results and calculations should be included in the compliance test report.
All calculated values must be within round off error of the true value.
Raw Process and Control Equipment Data (signed by plant representative) - The
test coordinator and facility representative should sign process and control equipment
data sheets.
Test Logs, Project Participants and Titles, and Related Correspondence - Self-
explanatory.
8-7
-------
Cover
1. Plant name and location
2. Emission source sampled
3. Dates of testing
4. Testing company name and address
Certification
1. Certification by team leader
2. Certification by reviewer (Professional Engineer preferred)
Introduction
1. Test purpose
2. Test location, type of process and control equipment
3. Any source identification numbers, if applicable
4. Test dates
5. Pollutants tested
6. Name of plant representative
7. Other important background information
Summary of Results
1. Emission(s) results and performance audit results
2. Process and control equipment data, related to determination of
compliance
3. Allowable emissions
4. Discussion of errors, both real and apparent
Facility Operations
1. Description of the process and control equipment
2. Process and control equipment flow diagrams
3. Process parameter, material, and product results, with example calculations
4. Describe portions of process operation tested
5. Representativeness of process parameters, raw materials, and products
6. Describe portions of control equipment operation tested
7. Representativeness of control equipment parameters, and collected
materials
Sampling and Analytical Procedures
1. Sampling port location(s) and dimensions of cross-section
2. Sampling point description, including labeling system
3. Brief description of sampling procedures, including equipment and diagram
of sampling train
Figure 8.1. VOC compliance test report format
8-8
I
-------
4. Description of sampling procedures (planned and accidental) that deviated
from standard method
5. Brief description of analytical procedures, including calibration
6. Description of analytical procedures (planned and accidental) that deviated
from standard method
Appendices
1. Complete results with example calculations
2. Raw field data (original, not computer printouts)
3. Laboratory report, with signed chain-of-custody forms
4. Calibration procedures and results
5. Raw process and control equipment data, signed by plant representative
6. Test log
7. Project participants and titles
8. Related correspondence
Figure 8.1. (Concluded)
8-9
-------
4>
B>
«J
Q
ia
£33
O
•H
4*
(8
CO
w
0)
Of
•o
flj
§ S
i*H o^
C >i
-------
e
3
4*
M
O
&
s
o
•H
>
2
41
M
O
D
a
1
O
«H
O
0
*1*
O1
id
a
CD
4*
a
o
g
Q
U
M
0
o
%
to
s
•d
c
10
to
to
0
0
0 -P
ri c
a 0
B
0 a
43 -H
£ 3
tr»
t
4* •
•H H
rH
•H
2
Pd
C
IO
I-P
hd
r-j
3
U
rH
(0
O
0
rH
•d a
e e
id id
•H B
c - 0
Q) rH
e m £
Sf- -&A
& £*
• S*
H iH
O to - 3
h B VH W
41 * 0 0
C rl 41 fc
O D» 0
o id g to
-H (0 -P
•d-d rJ O
C A3
idga-d
01 H w M
10
CO
0
u
o
a
•d
0
•H 41
41 0
id 6
41 id
C rl
0 id
•d
0
41
r-t U)
O 0
r<4l
41
C C
O O
O'H
41
<4H Id
O rl
0
to a
C 0
o
•H4>
4J C
rl 0
0 E
aa
a&
to a-H 0
0
rl
a
S
•
m
^
0
(0
s
•
vo
Icn
H
[ C
•H <|)
•p e
« a
4>-H
0 &
M 0
0
a
£
«
r^
C
o
•H
•d4i
c o
id 0
to
-. i
en w in
0 ^ 01
M C O
9 O h
«-H O
O 41
o id«p.
0 O O
H O
Pi H 01
C
•-*41 0
« X4-H
O O (0
•H ac
41 0
>«CPfi
iH C-H
rt-H-d
as
•d id
0 CO
CP
-HH
H
a
§
a
C
•r»
10
«j
iJ
?
c
•H
rH
t
(d
to
(tH
0
E
id
r4
|0>
id
l*Hi
I'd
•d
id
41
c
0
8,
•rH
0141
0 id
(HA
341
g 3 -d
0
41
- Ul
C >i
o to
•H
41 D»
ac
•H-H
^1 rH
O 0
U A
0 id
•drH
4J tP
C C
•H -H
o-d
a3
rH
&1O
C C
•H -H
rH
a*
id
CO
•
r«
tjiCf
C 0
-rl
i-t CT>
ac
e--i
id *d
(0 3
f-H
MH O
0 C
•rl
C
O »
•H tO
41 0
au
•H 3
h *d
U 0
0) 0
0 O
•d rl
a
<4-l
0
•r|
£
•
n
0^
OrH
O Id
r!4l
ac
0
tj^-d
C-H
•H O
rH O
ant
g
id-d
w c
id
,0»
H C
BJ-H
'
r7
«
41
C
0
•d
-H
o
o
id
•g
a)
t-i
td-na
c 'd u 0
m 3 -H c
H
a »H o n 0
0-rl U 341
41
id
•H
i
•d
rl^J
U 0
to u
0 O
•d u
a
C*d
O 0
•H O
41 O
au
•H a
1
«
2
o
c
o
m
41
A
O
i
Q
iH
S
o
•H
4*
''^
« -
2
'S'
•4rir
M
w
3
- g
•H
it
4^
*
r4
ft
-H
fa
8-11
-------
o
0)
BJ
CA O^
.^J
fl
9
41
£^
^j
9
OS
1
g -
o
PM
*
9
•H
41
1
PS
a
9
E«
9
O
fi
0
ft
H
t-
0
n
41
fl
9
I
i
o
• •
K
O
i
a
H
X
<0
e
(0
0)
45
•H
•
CO
41
3 .
to
9 f
03 f
• i
rH
a
•8
M
9*4 *
W 1 *
ft
9
a
*
i
*
•i
to
5*
H
P-l
id
H 1
3
0 n
H i*
T3 1
a-
4
*
•
I
C
r
P
g
%
_*
™T «
10
c.
^
H .1
Li
"1
O-
—
10
P
10
3 -1
3
_i
0
-1 <
-1
*
§
4
^^
U> .
r
C
H-
i
M
0
p
3 :
i
54
j <
;
'
'
H
p
O
a>
H
a
c<
P
*j
P
3
i
ui
M
>>
L( -,
5 i
^
J J
0 1
M
3
•>
g i
U
M
?
PI
i
1
H
0
1
C
H n
H J
V H
3 i
J i
J
•§
r
(
C
^
to
P
H
w
9
O
C
r
»
H J
3
0
1)
3,
4
A
c
0 1
H
_l ,
^ '
0
^ !
H
^ :
o !
i f
f U
1
f
1
g
a>
B
H
a
&.
0)
H»
D
£
3
0 i
o
C
10-
n
to
5J
iJI
M i
%
t
0
K
•>
^
H
P
10
4->
V
to
9
ft
0)
P
C
H
P.
£
0
L>
5>
H
10
^
Si
•* 1
0 I
3 r
(
E
VI
|
1
1
«
•1
.
•i
*
4
;
i
<
^
?1
(
J-r
0 (
2^
9 r
to
9
H
P
H
P
pd
3
C
10
10
p
C '
s,
J
H
J
L{
0
%
J 1
J
0-1
•n (
3 r
-1 (
M C
- a
9
U
C
9
!
a
u
j
3
U
3
1)
J
0
H
8
0
0
•H
U
9
4*
1
M
•H
a
9
a
£
0
O
a
41
9
1
U
H
0
•H
•4
V
.
M
9 .
a
•H
9 *"
Pi ^
•g -s
0 41
s §
g u
5 ^
& «
^ eo
2
&
•H
&•
8-12
-------
«
4*
O
4*
O*
Oi
Q
K
I
•w
M
0
§•
M
•P
a
S
o
§
«
Comp
o
c
01
^
fl)
Jn
o
c
01
4)
c
o
•H
•p
(0
c
•H
R
4)
.U
•o
4)
o
C
10
•H
Qi
o
J_l
o
lu
o
01
4)
•o
0)
4J
o
0)
"O
^,
fl)
X)
jj
esults in acceptable ram
udit results acceptable :
recommended that the tes'
1 1
01
4J
•H
•o
JJ
10
Q)
0)
(0
^
fl)
jj
«
o
c
H
43
•H
OJ
^
^
o
C
H
O
nd recommended action
flS
c
o
•H
id
g
(0
H
a
c
01
Q)
^^"
TJ
y
O
O<
0
u
o
-H
(0
CT
O
•2
c
10
01
4J
-H
c
»
01
results on correct basi
c
0
•H
CO
01
-H
g
§
fll
Q]
•H
(0
iH
P*
^'
0)
„
o
c
H
^1
u
(0
3
u
0
0
c
01
fl)
01
(0
o
c
01
0)
>l
01
O
C
01 &-P
«) H
1
o
c
»
1
)
H -P
-H
•o
3
(0
x<
4i
4 -H
3 >
-(
£4 O
01 fl) 0)
0)
>1•)
a
s
o
-I tPO
•1 (ft
3
TJ
3 0
3f*
*M
3 -P
x:
+j
•H
^
41
calculations validated
results within normal ro
ening VOC measurement me
measurement method agree
operations representativ
n problems
e
fl)
0 01 fc
•H b 0
0)
01 4) M >
to :
-H
g
*
01
«. •»
U
c
a>
01
01
i 0)
U
O
LJ
0)01 >i &> ft
<
1) U
o> >> e
M 01
0) -H
Q
0)
4)
•H
(0
CL,
y^
0)
«.
O
c
H
01
c
0
H
u
Q)
B,
0)
c
•H
0)
^
3
•P
3
*w
T(
§
-P
•H
e
^
0)
Oi
o^
c
•H
4J
(0
^
0)
o<
o
^3
(0
•H
key parameters to establ
01
•1-1
H
^
0)
fl)
r"1
H
0
C
01
fl)
>^
Q)
^
•H
.p
(0
+J
c
0)
CO
at
^4
equipment operations rep
.n problems
rH
o
t-t
•P
c
o
u
fl)
1
01
c
fH
(0
a
x
Q)
«.
o
c
H
O
•*-4
0
0)
a
01
c
•H
0)
^4
3
•P
3
^4
•d
c
10
Jj
•H
@
(4
0)
a
o*
c
*rt
-p
(0
^
fl)
a
o
x:
0)
•H
key parameters to establ
4->
0)
•H
fH
^
0)
Q)
5H
£J
•P
. c
8
CO
0)
k
0
o»
•H
8-13
-------
0
t*
(0
o
R
W
0
^^1
:ceptabl
v
10
e sampling procedures
no, explain problems
H
0
-------
-------
-------
APPENDIX A
ORGANIC COMPOUND DDENTDFICAT1ON AND QUANTIFICATION
A.1 Organic Compound Identification by Retention Time
A2 Adequate Peak Resolution
A3 Proper Response Factors
-------
-------
A.1 ORGANIC COMPOUND IDENTIFICATION BY RETENTION TIME
For Method 18, the organic compounds to be measured must be known prior to
the test. To identify and quantify the major components of the organic compounds
known to exist in the sample, the retention time of each component is matched with the
retention times of known compounds (standard reference materials or calibration
standards) under identical conditions. Separation of organic compounds is performed
with gas chromatographic columns, referred to as GC analysis. The retention time is the
time between sample injection into the gas chromatograph (GC) until the organic
compound reaches the detector. If conditions are maintained constant, the retention
time for each compound will be constant as well, and will serve as the identifying
parameter for each peak. Care must be taken to assure that two compounds do not
share the same retention time. The retention time shall be within 0.5 seconds or 1
percent of the retention time of the known compound's (calibration standard) retention
time (whichever is greater) to be considered acceptable. The retention time will vary
with (1) type of column or column material, (2) length of column, (3) temperature of
column, (4) organic compound and several other factors (e.g., other organics present).
The exact seconds or minutes of the retention time do not matter, although the longer
the retention time is, the longer the analysis time will be.
The major problem with the use of retention time in identifying organic
compounds is that other organic compounds could share the same retention time.
Compound identification is often achieved by using two GC columns with greatly
different packing materials which have differing physical properties used for separation.
If the proper retention time for a given organic compound is obtained for both columns,
then the analyst would assume that the compound has been identified. A listing of
organic compounds and their retention times, with respect to certain analytical
conditions, is presented in Section 3.16 of the QA Handbook, Volume III titled, "Kovats
Retention Indices." The Kovats Retention Indices are generally used to help the analyst
select the proper column for separating the organic compounds known to be present in
the sample. When the analyst only uses a single column at a set condition, there will
always be some doubt as to whether other organic compounds have the same retention
time and are being reported as the compound of interest It is not uncommon to have as
many as three compounds share the same retention time.
It is the responsibility of the analyst to ensure that the proper column has been
selected. When additional compounds share the same retention time, the reported value
will be higher than the true value.
To select the correct GC column or establish appropriate GC conditions, the
analyst must identify approximate concentrations of organic emission components. With
this information, the analyst can prepare or purchase commercially available standard
mixtures to calibrate the GC under physical conditions identical to those that will be
used for the samples. The analyst must also have presurvey information concerning
A-l
-------
interferences arising from other compounds present in the sample matrix and indicating
whether there is need for sample dilution to avoid detector saturation, gas stream
filtration to eliminate paniculate matter, and/or prevention of sample loss by moisture
condensation in the sampling apparatus.
Most analysts today use an integrator to determine retention times and sample
peak areas. The integrator will print out the retention time for all peaks and the
integrated area of all peaks. The observer should confirm that the peaks match within
the specified time of 0.5 seconds or 1 percent of the retention time, whichever is greater.
The first peak on the printout immediately after injection of the sample is typically not
an organic compound from the source. It is usually the air peak or solvent peak and is
not counted.
A3 ADEQUATE PEAK RESOLUTION
To obtain proper quantitative values, sample peaks (organic compounds as they
reach the analyzer) must be properly separated. As previously mentioned, compounds
must be separated to enable the detector to analyze only the compound of interest. The
analyst will perform initial tests using the calibration standards to determine the
optimum GC conditions for minimizing analysis time while still maintaining sufficient
resolution. Sufficient resolution shall be determined following the procedure described
by Knoll or in EPA Method 625 where the baseline to valley height (V) between two
adjacent peaks must be less than 25 percent of the sum of the two peak heights
(P1+P2). The equation is shown in Figure A.1. Both methods for determining peak
resolution give about the same results. EPA Method 625 is easier to calculate and
understand.
Most integrators will show where they are starting and stopping the peak area
integration by placing tick marks on the integrator printout. It is important that the
integrator is set properly to determine the proper area The observer may wish to check
where the start and stop marks are placed.
A3 PROPER RESPONSE FACTORS
Understanding the use of the response factor is important because (1) different
detectors can have a different response factor for the same compound, (2) each detector
can have a different response factor for different compounds, and (3) the same detector
can give a different response factor for the same compound at different conditions. The
response factor for each compound on any detector can be determined by dividing the
area units from the integrator printout from the standard by the concentration of the
standard (area units/ppm of standard). This is done for all concentrations of the
standard used to calibrate the detector. A best fit line, which is not forced through zero,
is typically calculated for these different response factors. This line is then used to
calculate the ppm from each sample's peak area based on the number of area Figure
A-2
-------
A.1. Adequate -peak resolution, units (area units of sample divided by response factor
which is area units per ppm). Today's integrators often make this calculation for the
analyst.
A3.1 Different Response .Factors for Different Detectors
It is important to note that different detectors are more or less sensitive to
different types of organic compounds. For example, the most commonly used detector,
. the -flame ionization detector, has the highest response factor for organic compounds that
:are straight chain molecules and contain only carbon and hydrogen. The flame
ionization detector's response factor for compounds that contain chlorine or oxygen is
reduced. Therefore, if it is calibrated with a compound that contains only carbon and
hydrogen like methane, propane, or butane, its measured values for chlorinated or
.oxygenated .compounds would be low.
The electron -capture detector has the highest response for chlorinated
.compounds. Its response to methane, propane, and butane would be lower than if they
were chlorinated. .Selection of a detector with a high response for the compound(s)
being measured is desirable, but is not critical unless the organic compound will not be
adequately measured by the detector.
. A32 .Different Response Factors for Different Compounds
After a suitable detector has been selected, the observer should be aware, as
-mentioned above, that the same detector may have a different response factor for each
compound analyzed. Therefore, a corresponding calibration gas must be used for each
compound. If the test is conducted for four compounds, the method may require that
four different calibration gases be used.
Method 25A is designed to measure emissions consisting of organic compounds
with only carbon and hydrogen and therefore requires instrument calibration with
propane. Method 25 A can provide accurate measurement of organic compounds
.xontaiBmg elements other than carbon and hydrogen if the detector is calibrated with a
..-mixture of .organic gases that closely represents gases in the stack. Response factors are
published for measurement of different organic compounds on each detector. These may
:not be accurate, but they are a good indication of the variation in response. For the
.flame ionization detector, most compounds will give at least a SO percent response when
compared to methane; propane, and butane.
Method-25 was developed to eliminate the reduced response factor problem when
"the organic compounds are unknown. When the organic compounds in the sample are
unknown (like after an incineration process), proper calibration gases cannot be selected.
this problem, Method 25 converts all the elements that give reduced
Tesponse factors and analyzes the compound as methane in terms of carbon. The results
A-3
-------
CM
o
•a
I
I
A-4
-------
are then reported as parts per million as carbon. Unfortunately, the true molecular
weights of the compounds are lost and a true concentration or mass emission rate .cannot
be calculated.
Although it may not always be desirable, there are cases when the use of a
surrogate calibration compound has been allowed. For example, if the regulations
require that all compounds be analyzed, the use of a calibration standard (specific
organic compound can be selected by the tester) for unknown compounds that consists of
less than 20 percent of the total area of all peaks (total response) for all compounds in
the sample would produce only a small error in the total emissions measured.
A surrogate calibration gas may also be allowed when (1) appropriate calibration
gases are not available, or (2) calibrating with multiple gases in the field is anticipated to
be too inconvenient. The agency may allow the tester to use a single calibration gas
(as a surrogate standard) and correct the measured values based on the response factors
obtained in the laboratory.
A*33 Different Response Factors for the Same Compound
Manufacturers of GC detectors have published response factors of the organic
compounds for their detector. These values should serve as estimates. The exact
response factor will vary with sample composition and conditions of the detector. The
response factor should be similar from day-to-day on a laboratory instrument analyzing
the same compound and composition of gas under the same conditions. However, the
repeatability of response factors is not sufficient to eliminate the requirement for
calibration. It is important to conduct single point response factor checks every couple
of hours or when the environment or sample conditions change significantly.
A-5
-------
-------
APPENDIX B
VOC OBSERVATION PROCEDURES
B.I Agency Use of Screening Measurement Methods During Compliance Test
B.2 On-site Observation Procedures Coupled with the Use of Agency Screening
Methods ~
-------
-------
B.1 AGENCY USE OF SCREENING MEASUREMENT METHODS DURING
COMPLIANCE TEST
As previously stated, most organic gases are invisible. Therefore, it can be
beneficial for the agency to conduct independent testing to determine the emission levels
and other key parameters. The agency's portable organic analyzer (EPA Method 21
instrumentation) can be extremely useful. The discussion below provides a example of
how these instruments can be used.
A printing or coating bed operation with a carbon adsorber as the control
equipment is to be tested. The same principles apply to other types of processes and
control equipment The procedures are as follows:
1. Determine which organic compounds are used in the process.
2. Obtain the proper calibration gases or determine the correct response
factor (see Chapter 5) and use a common calibration gas (e.g., methane).
The analytical instrument must accurately measure low concentrations of
volatile organic compound emissions (i.e., 0 to 100 ppm or 0 to 500 ppm).
The observer may determine the organic materials process during the
pretest survey. An instrument recorder or other device which provides
continuous recording of the emission data is highly recommended. An .
external pump may also be required.
3. Prepare instruments and calibration gas(es) for the pretest (presite) survey.
Often pretest surveys are conducted before compliance testing. Typically
the testing firm representative performs the pretest survey. Agency
personnel should participate if the compliance test is expected to be
unusual or difficult.
4. During the pretest survey, use the portable VOC analyzer to determine the
concentration levels at the edge of any hoods, points of possible fugitive
leaks, at the inlet sampling location, and at the outlet sampling location. If
problems exist, such as poor hood capture efficiency or fugitive leaks, the
facility should be informed and the problems corrected prior to the
compliance test The facility should also be informed that inspectors will
use the same procedures and instrumentation during future inspections to
determine if the problems still exist
5. During the compliance test, the same problem areas should be checked for
-fugitive emissions, and concentrations around the edge of the hood should
be determined. Once the observer is satisfied that all problems have been
corrected, Ihe test program may start If only one VOC analyzer is
available to the observer, it should be used at the outlet sampling location.
The emissions to the atmosphere should be determined before the first
run, during the first run, between the first and second runs, during the
second run, between the second and third runs, and during the third run.
When two analyzers are available, periodic checks for fugitive emissions
B-l
-------
B.2
should be made before, during, and between runs. Determining the
emissions before and between runs will prevent the facility from conducting
testing during portions of the process and control equipment operations
that have reduced emissions. If emissions increase between the runs, then
the observer may have to double the testing tune and require testing to be
conducted on a continuous basis. The second run starts immediately after
the first run is completed. If temperature sensors or other screening
instruments are used, it is best to have a primary standard (i.e., ASTM
mercury-in-glass thermometer) available to check the measured results.
ON-SITE OBSERVATION PROCEDURES COUPLED WITH THE USE OF
AGENCY SCREENING METHODS
It can be beneficial for the agency to conduct independent screening
measurements during the compliance test. An approach to conducting independent
screening measurements during the compliance test is presented below. VOC screening
instruments (EPA Method 21) can be used for either spot checks or equipped with
f pumps and continuous recorders to provide continuous emissions measurements. The
discussions of on-site observation presented in Chapter 3 will still apply. The concerns
about safety are even more critical when the agency is conducting their own screening
measurements. The observer will be in closer contact with the VOC's thereby increasing
the health and explosion risk. All instruments and methods of using these instrument
must be intrinsically safe. For using these portable instruments, the EPA manual EPA-
340/1-86-015, "Portable Instruments User's Manual for Monitoring VOC Sources*
should be read.
B.2.1 First Sampling Run
The observer's emission results should be compared with the testing firm's
emission results. On-site analysis by the test team does not diminish the need for the
agency to conduct independent on-site measurements; it improves the chances of
eliminating errors and obtaining a valid test. If only one agency observer is present, the
schedule below will make the most effective use of observation time. These procedures
are provided for less experienced observers to help establish a routine for on-site
observations. More experienced observers'will follow their established routine.
For the first test run, after determining that the facility operations are as specified
in the protocol, the observer should go to the sampling location to observe the test team
as they record the initial data. If a post test leak check is required, the initial sampling
system leak check need not be observed. When the observer is satisfied with the
sampling train preparation, he should allow the testing to begin. The agency's analyzer
(VOC screening method) should also be started. The observer should observe the test
team's sampling procedures for the first 15 minutes of sampling. When satisfied with the
test team performance and the agency's emissions results, the observer will note the
B-2
-------
average emissions for the last 10-minute period and then conduct a check on the facility
operations.
If the process and control equipment are operating satisfactorily and the data are
being recorded as specified, the time of the process and control equipment observation is
noted and the observer returns to the sampling site. The observer should determine the
emissions during the time that the facility operations were being verified by die observer
(assuming that a pump and recorder are on the analyzer) or take 10 minutes of emission
readings if the analyzer does not have a pump and recorder and must be operated
manually. If the observer is satisfied with the emission results, the analyzer probe should
be removed from the stack and audit gas should be drawn through the probe and
analyzer. The value for the audit sample should be recorded and the appropriate
calculation made to determine the relative accuracy. If any inajor problems exist with
the analyzer, they should be corrected.
- The observer should observe the completion of the test team's sampling,
particularly the final readings and the final leak check. He should then observe
transport of the samples and sample recovery. If the test team's analysis is to be
conducted on-site, the analysis should be closely observed during the audit gas cylinder
analysis and the analysis of the first test run sample(s). The tester's analyst should be
required to conduct all necessary calculations to determine the field results in terms of
units of allowable emissions standard (e.g., ppmv on a dry, standard condition basis for
the specified organic compound). All procedures and calculations should be validated by
the observer.
If the agency's organic analyzer is equipped for continuous analysis, it should
continue to run and record the emissions between completion of the first run and start of
the second run. If the emissions increase significantly during this period, the process and
control equipment parameters should be checked to detennine the reason for the
increase. As previously stated, many VOC process and control equipment operations can
be altered to provide short term emission reductions. If the observer feels that.this has
happened, it may be necessary to require an additional sampling ran (total of four runs),
informing the source that the emissions will not be allowed to increase between sampling
runs. Another option is to double the sampling time and require that subsequent run to
be started immediately after the prior run.
BJL2 Second Sampling Run
If the observer is satisfied with all sampling procedures during the first test run,
he should spend most of the second run observing process operations with intermediate
checks on the emission levels. During the second run, the observer should record the
emissions during two 10-minute periods, conduct one check with the audit gas, and
monitor the facility operations between emission readings. Ideally, .the emissions will be
recorded continuously, so the observer can correlate the emission levels with facility
operations.
B-3
-------
The observer should observe the test team recording final data of the second test
run, final leak check, transport of the sampling train to the cleanup area, and recovery of
the second run samples. The observer should determine what his emphasis will be for
the third run by considering the facility operations, emission levels measured by the test
firm, agency measured emissions, and observation of the sampling procedures. If the
analysis is conducted on-site, these results will also be taken into account The
observer's analyzer should continue to monitor the emissions between the second and .
third run to ensure that they do not increase between runs. ..
B.2 J Third Sampling Run
The observer should use available information to determine which procedures
need the most attention for the third run by assessing the observations of the facility
operations, measured emissions (if applicable), and observation of sampling procedures.
If the analysis is being conducted on-site, these results must also be taken into account.
All these elements should be used to determine which procedures may have the greatest
degree of error. The observer should then focus on these procedures. A check of
facility operations, sampling procedures, sample recovery, and sample analysis (if
applicable) should be included in the observation procedures of the third run.
B-4
-------
APPENDIX C
METHOD 18 OBSERVATION PROCEDURES
C.1 Selection of Proper Sampling and Analytical Technique
C.2 Observation of On-site Testing
C3 Preparation of Calibration Standards
C.4 Auditing Procedures
C5 References
-------
-------
C.1 SELECTION OF PROPER SAMPLING AND ANALYTICAL TECHNIQUE FOR
METHOD 18
Because of the number of different combinations of sampling, sample preparation,
calibration procedures, GC columns packing material and operating procedures, and GC
detectors covered under this method, a set of tables has been developed to assist the
tester in selecting (and the observer in evaluating) an acceptable sampling and analytical
technique. The compounds listed in these tables were selected based on their current
status as either presently regulated or being evaluated for future regulations .by. the EPA
and state and local agencies. The selected organic compounds for Method 18 presented
in Table C-l provide the user with: (1) the International Union of Pure and Applied
Chemistry (IUPAC) name, any synonyms, the chemical formula, the Chemical Abstracts
Service (CAS) number; (2) method, classification and corresponding references for more
information; and (3) information on whether EPA currently has an audit cylinder for this
compound.
For a given compound, the sampling or analytical techniques described in Tables
C-2, C-3, and C4 are classified into one of five categories as follows:
1. Reference. This is a method promulgated by EPA as the compliance test
method for one or more EPA emission regulations.
2. Tentative. This is a method where EPA method development is completed
and documented, but the method has not been promulgated.
3. ^Development This is a method currently under development by EPA
4. Other. A method developed and documented by an organization other
than EPA
5. None. This is a sampling or analytical technique that has not been
developed or validated but based on experience with similar situations, this
may work.
Table C-2 shows.all the sampling techniques allowed for Method 18: (1) direct
interface, (2) .Tedlar bag, and (3) adsorption tube sampling. For each compound, each
.of the allowed sampling techniques is rated either: (1) recommended, (2) acceptable, (3)
not recommended or (4) unknown. A particular sampling technique is rated first based
on current EPA methodology. "Where EPA methodology does not exist, methodology
provided by organizations other than the EPA is used for rating. As an example on how
to use Table C-2, the rating for benzene is R-2 for Tedlar bags and A-15,16 with carbon
disulfide for adsorbent tubes. This means that for sampling, a Tedlar bag is
recommended as a sampling technique and Reference 2 (Appendix C.5) provides further
.description, while charcoal tubes using carbon disulfide as the desorption liquid are
acceptable and Reference 15,and 16 (Appendix C5) provide further description.
•Before a final sampling technique is selected, the observer and source tester will
;need to consider the general strengths and weaknesses of each technique in addition to
C-l
-------
TABLE CL STATUS OF SELECTED ORGANIC COMPOUNDS
FOR METHOD 18 SAMPLING AND ANALYSIS TECHNIQUES
Chemical Abstracts Mam
Synonyms
Formula
CAS Mo.
Method
Class
EPA Audit
Cylinder(ppn)*
Alcohols
Hethenol •
Ethanol . '
Isopropyl Alcohol
n-Propyl Alcohol
n*8utyl Alcohol
Methyl Alcohol
Ethyl Alcohol
2-Propanol
1-Prcpanol
1-Iutanol
CU.O
C.H.O
CAO
CAO
C.*J3
(67-56-1)
(64-17-5)
(67-63-0)
(71-23-8)
(71-36-3)
0-6
0-7 -
0-7
0-8
0-8
30-80
No
No
No
No
Alkanes
Cyclohexane
Hexane
•i
C.MU
C,HM
(110-S2-7)
(110-54-3)
0-9
0-9
80-200
20-90,1000-3000
Alkenes
Ethylene
Propylene '
Ethene
?ropeoe
c^J,
V.
(74-SS-1)
(115-07-1)
•N
N
5-20,300-700
5-20,300-700
Dienes
1 ,3-3utadiene
Hexaeh 1 orocyc I opentadi ene ,
Butadiene
P erch 1 orocyc I opentadi ene
C,H.
c,ct.
(106-99-0)
(T7-47-4)'
0-10
0-11
5-60
NO
Aromatic
Benzene
Hesftylene
Ethylbenzene
Cunene .
Xylene (m-,o-,p-)
Toluene
Styrene
2-Maphthylamine
Benzol
1 ,3,5-Trinethylbenzene
1 -Methylethylbenzene
Dimethyl benzene
Methylberuene
Ethenylbenzene
2-Naphthytenamine
C.H.
tyi»
c^,.
CJa
C.H,.
C,H,
CA
c,A«
(71-43-2)
(108-67-8)
(100-41-4)
(98-82-8)
(1330-20-7)
(108-38-3)
(100-42-5)
(91-59-8)
T-12
N
0-13
0-13
0-13
0-9,13
0-13
0-14
5-20,60-400
No
No
No
5-20,300-700
5-20,100-700
No
No
Ketones
Acetone
Methyl Ethyl Cetone
Methyl isobutyl Ketone
2-Propanone
2-Butanone
4-Kethy I -2-pentanone
C^(.0
w
C.HB0
(67-64-1)
(78-93-3)
(108-10-1)
0-15
0-16
0-15
No
• 30-80
5-20
Epoxides
Ethylene Oxfde
Propylene Oxide
Epoxy Ethane
1,2-Epoxy Propane
«y«.o
W.O
(75-21 -8)
(75-56-9)
0-17
0-18
5-20
5-20,75-200
Sutfides
bisC2-Chlopoethyl) Sulffde
Mustard Gas
C.H.CI.S (505-60-2) N
No
C-2
-------
TABLE C-l. (Concluded)
Chemical Abstracts Name
Synonyms
Fermi la
Halogenated
Ethyl idene Chloride
Ethylene Oibromtde
Ethylene Oichloride
Propylene Bichloride
1,1,1 -Trichloroethane
Bromodich lorcmethane
Ch lorodibromomethane
Chloroform
Carbon Tetrachloride
D i ch I orodi f I uoremethane
Methyl Bromide .
Methyl Chloride
Methylene Chloride
Tetraeh I oroethyl ene
. Brcmoform
Trichloroethylene
Trichlorotrifluoroethane
Vinylidene Chloride
Ethyl Chloride
Chlorobenzene
Vinyl Chloride
1 ,2-Oibromo-3-ehloropropane
.1,1-Oichloroethane
1,2*0ibranoethane
1 , 2-0 i ch I oroethane
,£-uicnioraprupene
Nethylchlwofani
Tr ieh loraaethane
Freon 12
Bromome thane
Chloronethane
Dichlorcmethane
Perch loroethyl ene
Tribromomathane
Trichloroethene .
Freon 113
Chloreethane
Monoch I or obenzene
Chi oroethyl ene
DBCP
CJJ.CI,
CXBrCl,
CH3r,Cl
OfCl,
cci.
CHfl '-
c,cV
CHBr,
C.HCI,
C,H,Cl*
C,H*Cl
as NO.
Method
Class
EPA Audit
Cylinder(ppm)*
*
(106-93-4)
(107-06-2)
(78-87-5)
(71-55-6)
(75-27-4)
(124-48-1)
(56-23-5)
(75-71-8)
<75-fl9-2)
(75-25-2)
(79-01-4)
(76-13-1)
<75-3S-<>
(75-08-3)
(75-01-4)
(96-12-8)
0-19
0-20
T-21
0-22
T-21
N
N
T-23
T-23
0-24
0-25
0-26
T-27
-T-21 '
0-19
T-21
0-28
0-29
tr-T9
ft-30
0-37
5-20,50-300
5-20,100-600
3-20,300-700
5-20
No
No
5-20,300-700
5-20
No
No
Mo
1-20
i>"-20 ,300-700
No
S-20, 100-600
5-20
5-20,100-600
No
5-20
5-30
No
Method Classification Code
R a Reference - EPA promulgated method.
T B Tentative - EPA UKthod development complete; EPA reference available.
0 " Development
0 * Other -
N * None -
EPA method currently utder development.
Method development completed by organizations other, than EPA; reference available.
No reference available; reeoonendation based on experience.
The codes in the method classification coluan describe the current status of a sampling and analysis
method for each selected coopound. For example, the method classification code for benzene is: T-12.
This means the current method for benzene is a tentative EPA method with development complete and the •
reference for the method is citation number 12 in Appendix (j.5
The availability af EPA audit cylinders is shorn fn this column where:
( ) 3 Audit cylinders for this particular compound are available fran EPA in the concentration ranges
indicated (Reference 4}.
No * Audit cylinders for this particular compound are not avallable frca EPA. The source tester must
obtain audit gas cylinders from comercial gas vendors certified fey-irmependcnt-anaiysi* to be
within 5 percent of the concentration claimed by the vendor.
C-3
-------
TABLE C-2. METHOD 18 SAMPLING TECHNIQUES FOR SELECTED
;,.;.-!; "A VOLATILE ORGANIC COMPOUNDS
Selected
Conpcu^ds
Direct
Interface
Tedlar
SaS*
•;- Adso,-bent Tubes and Desorption Liquid
Charcoal*
Other «•
' Desbrption Liquid"**.
Alcohols' *
Hethanol "•'
Eshanol * ,
Isopropyl Alcohol
n-?ropyl Alcohol
n-3utyl Alcohol
T
T
T
T
T .
N
N
M
N
M
»
T-7 ' •
T-7 '
T-8
T-8 ' '
A-6; Silica tel
.
-
-
•
'Distilled Water
IX 2-Butanol in CS2
1X 2>3utanol fn CS2
Carbon Disulfide
Carbon Disulfide
Alkanes
Cyclohexane < .
Hexane -
T
T
,*.
ethyl ene '
Propylene
T
T
1 ,3-3utadiene , ,
Hexachlorocyclopentadiene
T
T
U
U
T-9
T-9 ' '
- •
Carton Bisulfide
Carbon Bisulfide
Atkenes
N
U
U
U
U
U
" ' ' U
U
Oieoes . ...
A-10
U
' A-41
N " ' " "
U
A- 11; Porapak"*
'•' ' Carbon Disulfide
""" Hexane
Aromatic • . '
Benzene
Nesitylene
Erhy I benzene
Cunene
Xylen* (m-,o-,p->
Toluene
Styrene < • •
e-Kapthylamine
T
T '
T
T
T
T
T
T
, • .
Acetone
Methyl Ethyl ICetone
Methyl Isobutyl ICetone
T
T
T
.-. •
Ethylene Oxide
Propylene Oxide
T
T
bisCZ-Chloroethyl) Sulfide
T
R-t2
U
U
U
U
U
U
U
T-9, 13
U- '
T-13
T-13
T-13
T-9, 13
T-13 '
T-U
-
Carbon Bisulfide
0
Carbon Disulfide
. , Carbon Disulfide
Carbon Oisulfide
Carbon Oisulfide
.Carbon Disulfide
Carbon Disulfide
Ketones-
K
N
M
T-1S
M
T-15
A- 16; Ambersoro
Carbon Disulfide
Carbon Oisulfide
Carbon Disulfide
Epoxides
A
U
T-17
T-18
.
99:1 Benzene:CS2
Carbon Disulfide
Sulfides
U
U
U
U
C-4
-------
Table C-2. (Concluded)
Selected
Compounds
Street
Interface
Tedlar .
Bag*
Adsorbent Tubes and Desorptioo
Charcoal*-
Other **
Desorptioo
Liquid
liquid***
Halogenated
Ethyl idem Chloride
Ethylene Dibronide
Ethylene Diehloride
Propylene Dichleride
1,1, 1 -Tr ichloroethano
Bromodichloroaethane
Oi I orodi brojBcmethane
Chloroform
Carbon Tetraehloride
D i ch I orodi f I uoronethane
-. Methyl Bromide
Methyl Chloride
Methylene Chloride
Tetrachloroethylene ' '
Sromoform
Trichloroethylene
Tr i ch lorotri f luoroethene
Vinyl idene Chloride
Ethyl Chloride
Chlorobenzene
Vinyl Chloride
1,2-Dibromo-3-chlor9propar>e
T
T
T
T
T
T
T
T
T
T
T
T
T
• T
T
T
T
T
T
T
T
T
U , .1- T-19
K-31
R-21-
U
R-21 '
U
U
R-23
R-23
U
U ,
U
R-27
R-21 .
U
R-21
R-21
U
U
U
R-30
U
T-20
T-19
T-22
T-19
U
U
T-19
T-19
. T-24
T-E5
T-26
T-32
T-33.
T-19
T-34
T-35
T-28
- T-29
T-19
T-36
T-37 .
.
-
*
-
.
-
•
-
-
•
-
-
.
*
.
*
•
.
-
-
»
•
Carbcti Disulfi'd*
99:1 8trtnr*;HsOH
Caracn Bisulfide
15X Acetone in CycJ-ohexane
Carocrt Risulficte
U
U
Cartoon Oisulfide
Carbon oisytftde
Hethyl«f« Chloride
Carixn Oisulfide
Watlranol
CarDon DiauLtioe
Cartefi 3 i su I f i d»
Carbon DisulHde
Carbon O'stjlHde
Carson Oisulfide
Career; Bisulfide
Cartoon Oisulfide
Carbon Disulfiae
Carbon O'sulfide
Carbon OisulfiAj
Rating Code
ft > Reconmnded.
A a Acceptable.
Based on actual source tests experience (sampling and analysis) this method is
valid and is the method of choice among Method 18 users.
Based on actual source tests or similar source test experience (sanpling and
analysis), this method is valid. The tester must evaluate for specific test.
Method has no documented experience, but in theory could be valid.
ided. Based on actual source tests or similar source test experience and/or theory, this
method is invalid.
U » unknown. Method has no documented experience and the theoretical aspects of saivling by this
method are inconclusive. The tester must demonstrate that this samp I ing- method is
valid. - • • -
The rating codes for stapling are based on the extent of method validation, for example, the rating
code for benzene is: T; 8-12; A-9,13. This means that direct interface is theoretically possible for
benzene, but no documented experience has been found; Tedlar bags are the recommended sanptinq mthod for
benzene by the tentative EPA method referenced in citation 12 in AppendixC.5; and sampling with
charcoal adsorption tubes is acceptable following the two methods referenced in citations 9 and 13 in
Appendix .5.
* - If eondensibles exist, use the procedure described in Section 3.16.4 of the EPA Quality Assurance Handbook
Volume HI.
** * Solid sorbents ether than charcoal recommended.
*** » The recommended desorption solution is given in this column. Analyst should consult the appro-
priate'reference for details.
C-5
-------
TABUE C-3. GC DETECTORS FOR SELECTED VOLATILE ORGANIC
COMPOUNDS BY METHOD 18
Selected Conpoirris
Cas Chromatograph Detector *
FID
ECO
PIB
1 EUCD
Alcohols
Hethanol
Ethsnol
Isopropyl Alcohol
n-Propyl Alcohol
n-autyl Alcohol
R-4,6
R-7
8-7
R-8
R*8 >
X
N
N
N
N
T-38
T-38
T-38
T-38
T-38
N
H
N
N
N
Alkane*
Cyclohexane
Hexane
R-4,9'
R-4,9
N
K
T-38
T-38
M
N
Alkenes
Ethytem
Propylene
A-4
A-4
K
M
T-38
T-38
. H
N
Dfenes
1,3-8utadiene
Hexachloroeyclopentadiene
R-4,10,ii
«-11
H
T
T-38
U ,-
H
T
Aromatic
Benzene
Mesitylene
Ethylbenzene
Cumene
Xylene
Toluene
Styrene
2-Napthylanine
R-4,12
T
R-13
R-13
R-4,13
«-4,9,13
R-13
R-14
N
N
M
M
N
N
N
N
T-38
T-38
T-38
T-38
T-38
T-38
T-38
U
M
N
N
H
M
II
N
K
Ketones
Epoxides
Sul fides
bis(2-Chloroethyl> Sulfide j U
Acetone
Methyl Ethyl Ketone
Methyl Isobutyl Ketone
R-1S
R-4.16
R-*,t5
N
a
K
T-38
T-38
T-38
N
H
K
Ethytene Oxide
Propylene Oxide
R-4,17
R-4,18
N
N
T-38
T-38
M
H
C-6
-------
TABLE C-3. (Concluded)
Selected Casiwunds
Gas Chromtograph Detector •
FID
.EO | P10
ELO
Halogenated
Ethyl idem Chloride
ithylene Dibremide
Ethylene Oiehloride
Propylene DicMoride
1 , 1 , l-Tricfilaroethane
Bronodidi l-orsnethene
Ch t or od i browmethane
CMorofoni
Carbon Tetrschloride
Otchlorodifluoremethane
Methyl Broaride
Methyl Chloride
Methylene Oloride
Tetrachloroethylene
Bromofora
Trichloroethylem
T r i eh 1 orotri f I uoroethane
Vinyl idem Chloride
Ethyl Chloride
Chlorocenzene
Vinyl Chloride
1 , 2-0 i broMo-3-ch I oropr opane
R-19
A-4
R-4,21
A-4
8-4,21
U
U
. R-4,23
R-4,23
R-24
a-25
R-26
R-4, 27,32
R-4.21
R-19
R-4,2t
R-4,21
R-4,28
R-29
R-4,19
»-4,30
U
• T
R-20
T
T
r
T
T
• • T ' '
T
T ' '
T
T
T •
T
- T
T
' T
••• T '
T
T
T *
R-37
U
U
T-38
T-38
U
U
U
T-38
T.-38
H-38
T-38
T-38
T-38
T-38
T-38
T-38
H-38
T-38
T-38
T-38
T-38
U -
T
T
T
R-22
T
T
T
A-23
A-23
T
T
T
T
T
T
T
T
T
T
T
T
T
Code
ft ec UMIIIGI HJed*
A * Acceptable.
Theoretical.
Based on actual source test experiences (sanpling and analysis)
this method is valid and is the method of choice aocng Method 18
users.
Based on actual source tests or similar source test experience
(sampling and analysis), this aethod is valid. The tester oust
evaluate for specific test.
Method has no documented experience, but in theory could be .
valid.
4 N * Mot Recomended. Based on actual source tests or similar source test experience
and/or theory, this wtftod is invalid.
U * unknown. Method has no doeunented experience and the theoretical aspects
are not conclusive. The tester oust demonstrate that this
detection nthod is valid.
The rating codes for GC detectors are based on the detector specified in the method
that is referenced. For example, the rating code for benzene is: 8-4,12; M; T-38; N.
JThis means that the F!0 is recanaended for detection of benzene by both references 4 and
12 cited in Appendix £.5; the SO and the ELO are not receamended for benzene; and
detection of benzene tnth a P1D is theoretically possible based on the ioniration
potential found in reference 33.
i.
* The following abbreviations are used for the gas diromatography detectors:
FID 3 Flaw !cnization Detector
ELO a Electraconductivity Detector
(Hall Detector)
ED * Electron Capture Detector
PID 3 Photoionization Detector
(with lamps up to 11.7 electron
volts)
C-7
-------
|V
Is--:
TABLE C-4. RECOMMENDED CALIBRATION TECHNIQUES FOR SELECTED
VOLATILE ORGANIC COMPOUNDS BY METHOD 18
Selected Compounds
Nethods for Direct Interface
and Tedlar Bag Sanples
Gas
Cylinders
Gas
Injection
into
Tedlar Bag
Liquid
Injection
into
Tedlsr Bag
Methods for Adsorption
Tube Sanples
Prepare
Standard in
Oesorption
Liquid
Prepare
Standard
on Tube
and Desarb
Alcohols
Aromatics
Cetones
Nethanol
ithanol
Isopropyl Alcohol
n-Propyl Alcohol
n-Butyl Alcohol
T-4
U
U
U
U
H
N
N
N
N
U
V
U
U
U
»~6
8-7
8-7
8-8
8-8
T
T
T
T
T
Alkanes
Cyctohexane
Hexane
Ethylene
Propylene
1,3-3utadiene
Nexach lorocyclopentadiene
T-4
T-4
H
N
Alkenes
T-4
T-4
U
U
U
U
N
N
8-9
R-9
U
U
T
T
U
U
bienes
A-10
U
R-10
N
N
U
8-41
8-11
U
T
Benzene
Mesitylene
Ethyl benzene
Cunene
Xyler* (m-,o-,p->
Toluene
Styrene
2-Kapthylanrine
B-12CSRK 1806)
U
U
U
T-4
T-4
U
U
X
H
N
N
K
N
M
U
A- 12
U
U
U
U
U
U
U
8-9,13
U
8-13
8-13
8-13
8-9,13
8-13
8-14
T
U
T
T
T
T
T
T
Acetone "
Kethyl Ethyl Ketone
Methyl Isobutyl Ketone
U
T-4
T-4
Ethylene Oxide
Propylene Oxide
bis(2-Chloro«thyl} Sulfide
T-4
T-4
U
N
N
N-
Epoxides
U
U
Sulfides
U
U
U
U
N
N
U
8-15
8-16
8-15
T
T
T
8-17
8-18
U
T
T
U
C-8
-------
TABLE CM. (Concluded)
Selected Compounds
Methods or Direct Interface
and T dlar Bag Samples
Gas
Cylinders
Gas
Injection
into
Tedlar Bag
Liquid
Injection
into
Tedlar Sag
Methods for Adsorption
Tube Saacles
Prepare
Standard in
Desorption
Liquid
Prepare
Standard
on Tube
and Desorb
Halogenated
Ethyl idene Chloride
Ethylene Oi bromide
Ethytene Oichloride
Prepylene Oicftloride
1,1,1-Trichloroethane
Bromodichlorooethane
CJUorosibrcHioptethene
Chloroform
Carbon Tetrachloride
D ich lorodi f Igorcmethane
Methyl Bromide
Methyl Chloride
Methytene Chloride
T et r ach I or oethy I ene
Bremoform
Trichloroethylene
Triehlorotriftuoroe thane
Vinyl idene Chloride
Ethyl Chloride
Cilorobenzene
Vinyl Chloride
1 , 2-Oibrofflo-3-chloropropane
U
T-4
R-21
T-4
R-21
U
U
R-23
R-23
U
U
U
R-21
R-2KSRM 1809}
U
R-21
R-21
T-*
U
T-4
R-30
U
N
H
• M
N
N
U
U
M
N
U
U
U
N
H
• H
N
M
N
N
N
.A-30
H
U
N-31
A-21
U
A-21
U
U
A-23
A-23
M
N
M
A-21
A-21
U
A-21
A-21
U
U
U
N
U
R-19
R-20
R-19
R-22
R-19
U
U
R-19
R-19
8-24
R-25
R-26
R-32
R-33
R-19
R-34
R-35
R-28
U-29
R-19
R-36
R-37
T
T
T
T
T
U
U
T
T
T
T
T
T
T
T
T
T
T
T
T
T
T
Rating Code
M m Reconiiiended.
A s Acceptable.
Theoretical.
Based on actual source test experience (sampling and analysis} this method
is valid and is the method of choice among Method 18 users.
Based on actual source tests or similar source test experience ling and
analysis}, this method is valid. The tester must evaluate for specific
test. • . •
Method has no documented sampling and analysis experience, but in theory
could be valid.
N * Mot Recommended. Based on actual source tests or similar source test experience and/or
theory, this method is invalid.
U « Unknown. Method has no docunented experience and the theoretical aspects are not
conclusive. The tester must demonstrate that this calibration Method is
valid.
The rating codes for calibration procedures are based on procedures specified in applicable sampling and/or
... analytical methods. For exaople, the rating code for benzene is: R-12CSXM 1806); N; A-12; R-9,13; T. This
neans that for beniene, the reeonaended calibration procedure for direct interface and Tedlar bag samples
involves the use of gas cylinders with the procedures described in citation 12 in AppendixC.5 and Standard
Reference Material 1806 (available from the National Bureau of Standards, Gaithersburg, MO); calibration
standards for benzene prepared by gas injection into Tedlar bags is not recommended; calibration standards
prepared by liquid injection into Tedlar begs Is acceptable following the procedures described in citation 12
in Appendix .5; preparation of calibration.standards in desorption liquid is the reccBBaended procedure for the
.„.adsorption tube methods described in citations 9 and 13 in Appendix .5; preparation of calibration standards
on adsorption tubes followed by desorption is theoretically valid for use with adsorption tube samples.
C-9
-------
the guidance provided in Table C-2. The strengths and weaknesses for the allowed
sampling techniques are as follows:
Direct Interface or Dilution Interface
Strengths:
1.
3.
Tedlar Bag
Strengths:
1.
2.
3.
4 —
Weaknesses:
1.
2.
3.
Samples are collected in a manner that retain the same compounds and
concentrations as the stack emissions.
No loss or alteration in compounds due to sampling since a sample
collection media (bag or adsorbent) is not used.
Method of choice for steady state sources when duct temperature is below
100°C and organic concentrations are suitable for the GC detector.
Weaknesses:
1.
2.
3.
4.
GC must be located at the sampling site.
GC cannot be operated at a sampling site if the presence of the H2
flame will be hazardous.
Cannot sample proportionally or obtain a time integrated sample.
Results represent only grab samples and should not be used for non
steady state processes.
Samples are collected in a manner that retains the same compounds and
concentrations as the stack emissions.
Samples may be returned to the laboratory for GC analysis.
Multiple analyses, if necessary, may be performed on each collected
sample.
Samples can be collected proportionally.
Unless protected, Tedlar bags are awkward and bulky for shipping back
to the laboratory. Caution must be taken to prevent bag leaks.
Stability of compound(s) of interest in Tedlar bags must be known and
sample storage time is generally less than 24 hours.
Polar compounds should not be collected due to bag absorption. Direct
interface or dilution interface is the method of choice for polar compounds.
C-10
-------
Adsorbent Tubes
Strengths:
1. Samples collected are compact and easy to return to the laboratory for
analysis.
2. Samples may be returned to the laboratory for GC analysis.
3. Sample storage time generally can be extended to a week by keeping
samples at O°C.
Weaknesses:
1. Quantitative recovery of organic compounds from the adsorbent material
must be known.
2. Breakthrough sample gas volume for organic compounds for the adsorbent
material must be known.
3. Any effect of moisture (in the stack gas) on the adsorbent material
collection capacity must be known. Moisture in the sample above
2 to 3 percent may severely reduce the adsorptive capacity.
4. Generally, samples are collected at a constant rate.
VJL
Table C-3 shows the GC detectors commonly used with Method 18. For each
compound, each GC detector is rated as: (1) recommended, (2) acceptable, (3) not
recommended, or (4) unknown. A particular GC detector is rated first based on current
EPA methodology. Where EPA methodology does not exist, methodology provided by
organizations other than the EPA is used for rating. As an example on how to use
Table C-3, the rating for benzene is R-2 for FID. This means FDD is recommended as
the GC detector and References 4 and 12 (Appendix C.5) provide further description.
Table C-4 shows the GC calibration preference for each compound based on the
technique used for sampling. Where appropriate, the source of calibration standards is
also shown. For each compound, the calibration technique shown is rated either:
(1) recommended, (2) acceptable, (3) not recommended, or (4) unknown. A particular
calibration technique is rated first based on current EPA methodology. Where EPA
methodology does not exist, methodology provided by organizations other than the EPA
is used for rating. As an example on how to use Table C-4, the rating for benzene is.
R-12 (SRM 1806) for gas cylinders. This means gas cylinders assayed and certified
against a National Institute of Standards (NIST) gaseous Standard Reference Material
(SRM) using EPA Traceability Protocol No. 1 (Reference 4) are recommended as the
calibration standard. Reference 12 provides further description on the source of the
calibration standard. NIST SRM 1806 would be used to assay and certify the calibration
standard.
C-ll
-------
For those compounds not listed in the tables, a general approach of classifying
compounds and then selecting the method typically used for that classification is
provided (Figure C.1, page C-13). The first classification in this general approach is
sample concentration. Method 18 can generally be used for samples having a
concentration of greater than approximately one part per million by volume. For
samples in the part per billion volatile organic compounds range, typically the volatile
organic sampling train is used. For samples in the pan per billion semivolatile
compounds, generally the modified method 5 sample train is used. The discussions on
the volatile organic sampling train (VOST) and modified method 5 (MM5) trains is
beyond the scope of this manual but can be found in the Solid Waste Sampling Handbook
SW-896. Method 18 can be used for most gaseous organic compounds with a
concentration greater than 1 ppm.
The second classification of compounds depends on stack temperature and
moisture content. In general, higher stack temperatures are obtained through direct
combustion which means higher moisture. The temperature/moisture are classified as
low temperature and moisture, medium temperature and moisture, and high moisture
and temperature.
t
The first temperature/moisture classification is low moisture and temperature.
Low moisture is defined as less than 1 percent. The second temperature/moisture
category is medium temperature and moisture (3 to 10 percent). The significance of
3 percent to 10 percent moisture is:
1. If the container sample is allowed to sit at room temperature moisture will
condense.
2. At between 3 to 10 percent moisture, techniques of heating the containers
can be used to keep the water in a vapor form.
The final temperature/moisture category is for moisture content greater than
10 percent. This moisture content renders most of the heating techniques impractical
because the moisture is more easily condensed and therefore eliminates most of the
standard sampling analytical techniques.
After the sampling conditions have been selected according to temperature and
moisture, the next classification deals with the polarity of the organic compound. Polar
compounds are generally those compounds that mix with water, such as alcohols, and act
as water since water is a polar compound. Silica is a good sorbent for both water and
alcohols. The nonpolar compounds, such as the chlorinated organics, typically do not
mix with water and can therefore be easily purged out of water.
For nOnpolar compounds at low levels or ambient temperatures and moisture
content all the collection techniques would work which include: Tedlar bags, adsorption
tubes, and direct and dilution interface. Charcoal as we know is the most widely used
sorbent for nonpolar compounds.
C-12
-------
LU
UJ LU O
i^£
LU
35
°l*z
5CH-.5
3 o o j=
s°K"5
— "^ ••• «
T
u_
QC
UJ
a.
O
c
O
Q.
-------
For collection of polar compounds at ambient temperatures and moistures (Figure
C.1, page C-13), all sampling techniques with the exception of Tedlar bags work. Bags
tend to adsorb polar compounds. Silica is the sorbent most commonly used for polar
compounds. Direct and dilution interface will work for all cases of polar and nonpolar
compounds in steady state emissions and where the compounds can be separated quickly.
Sampling under medium temperature and moisture conditions (Figure C.1, page
C-13) is the same as under low temperature and moisture condition with the exception
that containers must be heated to prevent moisture and organic condensation.
Therefore, nonpolar compound sampling would include heated container, adsorption
tube, direct and dilution interface methods. The polar compounds would best utilize
adsorption tube, direct and dilution interface techniques.
It should be noted that the adsorption tubes must be kept cool and that sorbent
collection, efficiency may be severely affected by water. One approach when moisture is
present is to sacrifice the first adsorption tube for collection of water and then have a
third collection tube as the backup tube.
The high temperature/high moisture conditions (Figure C.I, page C-13)
necessitate direct and dilution interface sampling or the recently EPA developed
adsorption tube sampling technique for nonpolar compounds such as chlorinated organics
that are not soluble in water. This technique will not work for compounds that are
soluble in water.
s
C2 OBSERVATION OF ON-SITE TESTING
The test coordinator should use the techniques and tables provided above in
Appendix C.1 to ensure that the sampling and analytical techniques selected by the tester
are acceptable. Because of the complexity involved in sampling organic compounds from
the variety of potential source types, only the more common problems are addressed for
each sampling method. The observer should have the tester conduct the recommended
quality assurance/control checks and procedures provided in this Chapter to assess the
suitability:of the sampling technique.
Specific sampling system descriptions and observers checklists are provided below.
The procedures are presented by sampling techniques as shown. The observer can
therefore read only the material of interest.
Chapter
4.53
A
B
C
Sampling Approaches
Evacuated Container Sampling
Sampling System Preparation
Proportional Sampling
Indirect Pumping Bag Sampling
Page
4-5
4-6
4-6
4-7
C-14
-------
Chapter Sampling Approaches Page
D Sample Recovery and Transport to Laboratory 4-9
£ Common Problems 4*9
F Stability Check 4-9
G Retention Check 4-9
Appendix Sampling Approaches Page
-
C.2.1. Evacuated Container Sampling (Heated and Unheated) C-15
H Direct Pumping Bag Sampling . C-22
I Explosion Risk Area Bag Sampling C-23
J Prefilled Bag Sampling C-23
C2.2 Direct Interface Sampling C-27
C.2.3 Dilution Interface Sampling C-30
C.2.4 Adsorption Tube Sampling C-32
G2.1 Evacuated Container Sampling (Heated and Unheated)
In this sampling technique, sample bags are filled by evacuating the rigid air-tight
containers that hold them. The suitability of the bags for sampling should have been
confirmed by permeation and retention checks using the specific organic compounds of
interest during the presurvey operations. The permeation and retention checks must be
performed on the field samples to ensure that the container sampling technique is
acceptable. ~-
The means of transporting the bags to the laboratory for analysis within the
specified time should also have been determined. Delays in shipping and/or analysis can
result in significant changes in concentration for many compounds. EPA has conducted
several field evaluations of Method 18 bag sampling techniques on a variety of organic
compounds. In the EPA Field validation reports the specified time between sample
: collection and analysis is shown. The permeation and retention checks are not required
by Method 18 but are highly recommended for compounds and source categories that
have not been validated by EPA.
On-site sampling includes the following steps:
1. Conducting preliminary measurements and setup.
2. Preparation and setup of sampling system.
3. Preparation of the probe.
4. Connection of electrical service and leak check of sampling system.
5. Insertion of probe into duct and sealing of port.
6. Purging of sampling system.
7. Proportional sampling.
C-15
-------
8. Recording data.
9. Recovering sample and transportation to laboratory.
The "On-site Checklist" (Figure C.2, page C-17) includes checks for each of the
following procedures and should be completed by the observer.
To assist the observer in noting the most critical items to observe, the key points are
printed in bold lettering.
Method 18 requires that samples be collected proportionally, meaning that the
sampling rate must be kept proportional to the stack gas velocity at the sampling point
during the sampling period. If the process has a steady state flow (constant), then the
flow rate "does not have to be varied during sampling. The average velocity head (phot
reading) and range of fluctuation is determined and then utilized to establish the proper
flow rate-settings during sampling. If it is found that the process is not steady state, then
the velocity head must be monitored during sampling to maintain a constant proportion
between the sample flow rate and the flow rate in the duct.
A total sampling time greater than or equal to the minimum total sampling time
specified in the applicable emission standard must be selected. The number of minutes
between readings while sampling should be an integer. It is desirable for the time
between readings to be such that the flow rate does not change more than 20 percent
during, this period.
s
If it was determined from the literature or the preliminary survey laboratory work
that the sampling system must be heated during sample collection and analysis, the
observer must ensure that the sample system does not go below the specified
temperature. The average stack temperature is used as the reference temperature for
the initial heating of the system and should be determined. Then, the stack temperature
at the sampling point is measured and recorded during sampling to adjust the heating
system just above the stack temperature or the dew point. In addition, the use of a
heated sampling system typically requires that the analysis be conducted on-site since it is
not practical to maintain the sample bag at elevated temperatures for long periods of
time.
A. Sampling System Preparation - See Chapter 4.5
B. Proportional Sampling - See Chapter 4.5
C. Indirect Pump Bag Sampling - See Chapter 4.5
D. Sample Recovery and Transport to Laboratory - See Chapter 4.5
£. Common Problems - See Chapter 4.5
C-16
-------
EQUIPMENT SETUP
Flue Gas Flowrate constant, variable
.Sampling rate constant, proportional
Sample tune required, actual
Time per point minutes, probe heat required _ yes no
DIRECT OR DILUTION BAG SAMPLING
Apparatus
Pitot tube: Type S Other , Properly attached
Pressure gauge: Manometer Other , Sensitivity
Probe liner: Borosilicate , Stainless steel Teflon
Clean , Probe heater (if applicable) on Glass wool filter
(if applicable) in place Stainless steel or Teflon unions used
to connect to sample line
Sample line: Teflon , Cleaned , Heated (if applicable)
Bag: Tedlar Other , Blank checked , Leak checked
Reactivity check , Retention check
Flowmeter: Proper range , Heated (if applicable) , Calibrated ^
Pump: Teflon coated diaphragm , Positive displacement pump > .
Evacuated canister , Personnel pump
Heated box with temperature control system: Maintained at proper temperature
Charcoal adsorption tube to adsorb organic vapors: Sufficient capacity : •_-
Dilution equipment: N2 gas , Hydrocarbon-free air , Cleaned and
dried ambient air , Dry gas meter
Barometer: Mercury , Aneroid , Other
Stack and ambient temperature: Thermometer , Thermocouple ,
Calibrated _ :
Procedures
Recent calibration (if applicable): Pitot tube , Flowmeter ,
Positive displacement pump* , Dry gas meter* , Thermometer ,
Thermocouple , Barometer
Sampling technique: Indirect bag , Direct bag , Explosion risk bag ,
Dilution bag , Heated syringe , Adsorption tube ,
Proportional rate , Constant rate , Direct interface ,
Dilution interface
Filter end of probe and pitot tube placed at centroid of duct (or no closer than 1 meter
to stack wall) and sample purged through the probe and sample lines*
Figure C.2. On-site measurements checklist.
C-17
-------
Vacuum line attached to sample bag and system evacuated until the flowmeter
indicates no flow (leakless)*
Heated box (if applicable) same temperature as duct*
Velocity pressure recorded and sample flow set
Proportional rate sampling maintained during run*
Stack temperature, barometric pressure, ambient temperature, velocity pressure
at regular intervals, sampling flow rate at regular intervals, and initial and
final sampling times recorded*
At conclusion of run, pump shut off, sample line and vacuum line disconnected
and valve on bag closed
Heated box (if applicable) maintained at same temperature as duct until analysis
conducted
No condensation visible in bag*
Sample bag and its container protected from the sunlight
Audit gases collected in bags using sampling system*
Explosive area bag sampling: (with following exceptions same as above)
Pump is replaced with an evacuated canister or sufficient additional line is added
between the sample bag container and the pump to remove the pump from the
explosive area . .
Audit gases collected in bags using sampling system*
Prefilled bag: Proportional rate Constant rate
Dilution factor determined to prevent condensation*
Proper amount of inert gas metered into bag through a properly calibrated dry gas
meter*
Filter end of probe (if applicable) and pilot tube placed at centroid of duct (or
no closer than 1 meter to stack wall) and sample purged through the heated probe,
heated sample line, and heated flowmeter or positive displacement pump*
Leak checked and partially filled bag attached to sample line
Stack temperature, barometric pressure, ambient temperature, velocity pressure at
regular intervals, sampling rate at regular intervals, and initial and final
sampling times recorded*
Probe, sample line, and properly calibrated flowmeter or positive displacement pump
maintained at the stack temperature*
Sampling iconducted at the predetermined rate, proportionally or constant for entire
run*
No condensation visible in probe, sample lines, or bag*
At conclusion of run, pump shut off, sample line disconnected and valve on bag
dosed
Sample bag and its container protected from sunlight
Audit gases collected in bags using dilution system*
Figure C.2. (Continued)
C-18
-------
Sample Recovery and Analysis
(As described in "Post sampling operations checklist," Figure CIO, page C-37)
DIRECT AND DILUTION INTERFACE
Apparatus
Probe: Stainless steel , Glass , Teflon , Heated system (if
applicable) , Checked
Heated sample line: Checked*
Thermocouple readout devise for stack and sample line: Checked*
Heated gas sample valve: Checked*
Leakless Teflon-coated diaphragm pump: Checked*
Flowmeter: Suitable range
Charcoal adsorber to adsorb organic vapors
Gas chromatograph and calibration standards (as shown in "Post sampling operations
checklist," Figure C.10)*
For dilution interface sampling only:
Dilution pump: Positive displacement pump or calibrated flowmeter with Teflon-
coated diaphragm pump checked*
Valves: Two three-way attached to dilution system
Flowmeters: Two to measure dilution gas, checked* -"
Heated box: Capable of maintaining 120°C and contains three pumps, three-way
valves, and connections, checked*
Diluent gas and regulators: N2 gas , Hydrocarbon-free air , Cleaned air _,
Checked
Procedures
All gas chromatograph procedures shown in "Post sampling operations checklist"
(Figure CIO)
Recent calibration: Thermocouples , Flowmeter , Dilution system
(for dilution system only)*
Filter end of heated probe placed at centroid of duct (or no closer than 1 meter to
stack wall), probe and sample line heat turned on and maintained at a temperature of
0°C to 3°C above the source temperature while purging stack gas
Gas chromatograph calibrated while sample line purged*
After calibration, performance audit conducted and acceptable*
Sample line attached to GC and sample analyzed after thorough flushing*
Figure C2. (Continued)
C-19
-------
With probe removed from stack for 5 min, ambient air or cleaned air analysis is
less than 5% of the emission results* '
Probe placed back in duct and duplicate analysis of next calibration conducted
until acceptable agreement obtained*
All samples, calibration mixtures, and audits are analyzed at the same pressure
through the sample loop*
Sample Analysis
(As shown in "Post sampling operations checklist," Figure C.10.)
If a dilution system is used, check the following:
With the sample probe, sample line, and dilution box heating systems on, probe and
source: thermocouple inserted into stack and all heating systems adjusted to a
temperature of 0°C to 3°C above the stack temperature
Hie dilution system's dilution factor is verified with a high concentration gas of
known concentration (within 10%)
The gas chromatograph operation verified by diverting a low concentration gas into
sample loop
The same dilution setting used throughout the run
The analysis criteria is the same shown as for the direct interface and in the
"Post sampling operations checklist," (Figure C.10.)
ADSORPTION TUBES
Apparatus
Probe: Stainless steel
Glass
Teflon
Heated
system and filter (if applicable)
Silica gel tube or extra adsorption tube used prior to adsorption tube when
moisture content is greater than 3 percent
Leakless sample pump calibrated with limiting (sonic) orifice or flowmeter.
Rotameter to detect changes in flow
Adsorption tube: Charcoal (800/200 mg), Silica gel (1040/260 mg)
Stopwatch to accurately measure sample time
Figure C.2. (Continued)
C-20
-------
Procedures
Recent calibration of pump and flowmeter with bubble meter
Extreme care is taken to ensure that no sample is lost in the probe or sample line
prior to the adsorption tube
- Pretest leak check is acceptable (no flow indicated on meter)
Total sample time, sample flow rate, barometric pressure, and ambient temperature
recorded
Total sample volume commensurate with expected concentration and recommended
sample loading factors •
Silica gel tube or extra adsorption tube used prior to adsorption tube when
moisture content is greater than 3 percent •
•' Post-test leak check and volume rate meter check is acceptable (no flow indicated on
meter, post-test calculated flow rate within 5 percent of pretest flow rate)
Sample Analysis
(As shown in the "Post sampling operations checklist," Figure CIO, page C-37.)
•Most significant items/parameters to be checked.
Figure C.2. (Concluded)
C-21
-------
F. Stability Check - See Chapter 4.5
G. Retention Check - See Chapter 4.5
H. Direct Pump Bag Sampling - Direct pump sampling is conducted in a manner
similar to evacuated container sampling, with the exception that the needle valve
and the pump are located between the probe and sample bag and the sample
exposed surfaces of both must be constructed of stainless steel, Teflon or other
material not affected by the stack gas (Figure CJ, page C-23), Due to the
additional likelihood that sample may be lost in the needle valve and pump, it is
recommended that the probe, sample line, needle valve, and pump be heated. If
it has or can be shown that this not a concern, then the heating may be
eliminated. All precautions, procedures, data forms and criteria can be applied.
Ensure that the system has been adequately purged before attaching the bag and
sampling.
I. Explosion Risk Area Bag Sampling - Explosion risk area bag sampling is also
similar to evacuated container sampling. The major difference is that no electri-
cal components can be used in the explosion risk area. The first option of the
tester is to locate the electrical equipment (e.g., the pump) outside the explosion
risk area and run a long flexible line to the container. If that option is not
possible, an evacuated steel container may be used as shown in Figure C.4, page
C-24. This option may involve a potential spark hazard and must be checked
through the plant safety officer. It is unlikely that electrical heating of the system
will likely be allowed. If an evacuated steel container is used, the leak check can
be conducted outside the explosion risk area and the probe can be purged with a
hand squeeze pump. The tester may wish to consider an alternative method of
sampling such as adsorption tubes and an intrinsically safe personnel sampling
pump or the syringe method. The primary concern must be safety in an explosion
risk area and all operations must be outlined in writing and cleared through the
Plant Safety Officer. The same criteria as described above for suitability of the
bag will apply and must be met.
J. Prefilled Bag Sampling - The prefilled bag sampling technique is similar to the
heated direct pump sampling method. The major difference is that the sample
bag is prefilled with a known volume of nitrogen, hydrocarbon-free air, or
cleaned, dried ambient air prior to sampling and the volume of gas sampled must
be accurately determined (Figure G3, page C-23). When using a flowmeter or
metering pump, the maximum dilution that should be attempted is 10 to 1.
Alternatively, a heated, gas tight syringe may be used to collect the source and
inject it into the sample bag. A heated, gas tight syringe can be used for dilutions
of 5 to 1 when the dilution is performed in the syringe and 50 to 1 when
performed in the bag. The use of a heated, gas tight syringe should follow the
procedures shown below. Both techniques should be verified in the laboratory
C-22
-------
o
VJ
SO
"E.
E
I
£.
I
o >
£ cr>
fi
o r* o
C-23
-------
C-24
4
o
vS
C
S3
SO
"35
gj
£0
e
C.
C
C
I
C.
u
-------
using higher concentrations of calibration gases and must be within 10 percent of
the calculated value. The technique is verified in the field by diluting the audit
gases in the same manner as the stack gases (see Chapter 4.7 for auditing
procedures).
Following are the recommended steps to conduct prefilled bag sampling:
1. Hie sampling should be conducted proportionally as described above in
Appendix G2.1. Calculation of the average sampling rate versus the average AP
will be the same with the exception that the volume of the prefilled inert gas must
be taken into account.
2. The suitability of the prefilled bag sampling technique should have been checked
in the laboratory. This would include calculating the dilution factor required to
obtain an acceptable sample concentration. The dilution factor must be properly
calculated since the concentration of the sample will be corrected for the inert gas
volume.
3. In the laboratory area, fill the sample bag (previously leak checked) with the
calculated volume of inert gas. Because of the potential for leaks, bags should be
filled the same day they are used. The inert gas volume must be determined with
a calibrated dry gas meter or mass flowmeter. The bag should be sealed and
taken to the sampling site.
4. At the sampling site, the sampling system is leak checked without the sampling
bag attached. Turn on the heating system and heat the system to the stack
temperature. Connect a U-tube H2O manometer or equivalent to the inlet of the
probe. After the system comes to the desired temperature, turn on the pump and
pull a vacuum of about 10 in. of H2O. Turn off the needle valve and shut off the
pump. If there is no noticeable leak within 30 seconds, then the system is leak
free. The heating and leak check are again important.
5. Place the probe in the stack at the sampling point (centroid or no less than 1
meter from the wall) and seal the port so there will be no in-leakage of ambient
air. Turn on the pump and purge the system for 10 minutes. While the system is
purging, determine and set the proper flow rate based on the AP.
6. Turn off the pump and attach the sample bag. Compare the heating system.
7. The sampling will be conducted proportionally. The stack temperature and
heating system temperature should be monitored and recorded. Record the data
on the sampling data form.
8. At the conclusion of the run, turn off the pump and remove the probe from the
duct. Remove the bag and seal it.
9. Conduct a final leak check. The system should pass the leak check; if it does not
pass, repeat the run.
C-25
-------
K. Heated Syringe Sampling - The heated syringe technique can be used with the
prior approval of the Administrator. This technique should only be used when
other techniques are impractical. The heated syringe technique requires on-site
analysis with three syringes collected and analyzed for each run. The require-
ments for the use of the syringes are the same as for the bag with regard to the
reaction of the gases with time and.the retention of the gases in the syringe.
Following are the procedures recommended for the syringe sampling technique:
1. If heating is required, then the syringe must be encased in material that has a
high density to maintain the proper temperature. Alternatively, an external
heating system can be used that keeps the syringe at the proper temperature until
just before use and to which the syringe can be immediately returned. The
syringe must be properly heated. A check can be made on the heating system by
filling the syringe with inert gas after the sample injection, reheating it and then
mject the inert gas. If the system give a concentration of 10 percent or more of
the original sample concentration, then the sampling system is unacceptable.
2. The access port should be extremely small to prevent in-leakage of ambient air.
The port may be covered with Teflon or other nonreactive material that will allow
the syringe to penetrate the material for sampling.
3. For the direct injection method (no dilution), place the syringe needle into the
stack and fill and discharge the full volume that will be sampled three times.
Then, draw the emission sample into the syringe, immediately seal the syringe and
return to the heating system, if applicable. The second and third syringes are
sampled at equal time intervals spanning the required sample (run) time. The
syringe samples must not be taken one immediately after another.
4. For the diluted syringe method, the inert gas is introduced into the syringe three
times and discharged. Following this, the proper volume of inert gas is pulled
into the syringe. The syringe is then placed into the duct and the proper volume
of stack gas is added. Immediately remove the syringe needle from the duct, seal
the syringe, and return to the heating system, if applicable. If a dilution approach
is :iised it should be checked as shown in Item 1 and the dilution factor should be
checked using calibration gases.
5. Foi>the bag diluted syringe method, the bag should be prefilled with the proper
volume of inert gas. The sampling is conducted as described above and the
sample injected into the bag through a septum.
6. Record the data on a field sampling data form.
7. Since the method requires that a proportional sample collected, the velocity head
(AP) should be recorded at about the same time that each sample is collected.
The concentrations can then be mathematically corrected to provide an integrated
value. If the process is a constant source operation (less than 10 percent change
in flow over the sampling period), it is not necessary to correct the measured
values.
C-26
-------
C22 Direct Interface Sampling - The direct interface procedure can be used provided
that the moisture content of the stack gas does not interfere with the analysis procedure,
the physical requirements of the equipment can be met at the site, and the source gas
concentration is low enough that detector saturation is not a problem. Adhere to all
safety requirements when using this method. Because of the amount of time the GC
takes to resolve the organic compounds prior to their analysis, the GC can only typically
make three analyses in a one-hour period. Therefore, the number of injections hi the
direct interface method is greatly limited by the resolution time. At least three injections
must be conducted per sample run.
Following are the procedures recommended for extracting a sample from the stack,
transporting the sample through a heated sample line, and introducing it to the heated
sample loop and the GC. The analysis of the sample is described in Appendix G3.
1. Assemble the system as shown Figure C.5, making all connections tight.
2. Turn on the sampling system heaters. Set the heaters to maintain the stack
temperature as indicated by the stack thermocouple. If this temperature is above
the safe operating temperature of the Teflon components, adjust the heating
system to maintain a temperature adequate to prevent condensation of water and
organic compounds.
3. Turn on the sampling pumps and set the flow rate at the proper setting. Typically
11/min is used. The sample rate may vary for the type of system used. The
system may use either an internal pump or external pump.
4. After the system reaches the same temperature as the stack, connect a U-tube
H2O manometer or equivalent to the inlet of the probe. Pull a vacuum of about
10 in. of H2O, and shut off the needle valve and then the pump. The vacuum
should remain stable for 30 seconds. If the system leaks, repair and:then recheck
the system.
5. Calibrate the system as described in Appendix C.3. Repeat'until duplicate
analyses are within 5 percent of their mean value (Appendix C.3). The calibration
of the system is critical.
6. Conduct the analyses of the two audit samples as described in Chapter 4.8. The
results must agree within 10 percent of the true value (or greater, if specified on
the cylinder). If the results do not agree, repair the system and repeat the
analyses until agreement is met or until approval is given by the representative of
the Administrator. The performance audit is critical.
7. After the audit has been successfully completed, place the inlet of the probe at
the centroid of the duct, or at a point no closer to the walls than 1 meter, and
draw stack gas into the probe, heated line, and sample loop. Purge the system for
a least 10 minutes.
8. Record the field sampling data on a form such as the form shown in Figure C.6.
9. 'Conduct the analysis of the sample as described in Appendix C3. Record the
data on the applicable data form. Ensure that the probe and sample lines are
maintained at 0°C to 3°C above the stack temperature (or a temperature which
C-27
-------
tz
111
i
111
C-28
.
(5
(fl
Q
0
I
u.
-------
.0 J= JB _S E
O O O
O OJ (9
> *J J-
. . .
o o s
O O ft)
u S || -,- u.
•— E *-» O*^ -^ +» •— ~ £
,CL3f-Oel..3fl»3 S
E— CS-—j_,_e3 4->
nj o -^ a<+- 10 •»- -^ Is >o
E _. _
re (O O M-
= u_ e
O
o
o>
C
I/)
*J _
o eu e c =
•»- u o a>.°
'
= >, S. O
ns 4-> OJ *J c
• O. fl 3
a- «_> o o o:
0)
•Q
M
M
a, a>
x e j_
»««-
s-^5
0) U 4->
0)
4-1
1.
o •
&«-»
LL.
e o
0 — •
*J U
uo
u
•••»
c
M 0)
01 c
c -^ — »
— ^ u.
•9 0
m at «— •
OJ i—
J- Q.O
EO •
V 19
U VI
3
03
S- .-*
a> u_
£.010
e JD— '
u o
I— i- (J
0.0
^*
Ji U.
U 0
fO *~*
4«£
v»0
J-> C
4/| gj ^M»
5 15
** •»- "E
*^T3
a>
«~»4-> C
v> e t~
«"• W E
W 3 -<*.
O) ^- i—
4-> — E
g-o
E
«— tt) C
LL. o i-
X- E
3 ^**«,
O i—
o E
Is.
*J <1>
u -a
0) E
k«u ^V
^ 7 ^
c e
«- 0
O — JZ
^*
0 U «•
E a>
-------
10.
prevents condensation). The sample lines must be properly heated and
equilibrated. A good check on the system is to pull the sample probe out of the
stack at the conclusion of a sample run. The system should return to less than
about 5 percent of the sample concentration within about 5 minutes.
Conduct the post-test calibration as described in Appendix C3. System
calibrations are critical.
C2.3 Dilution Interface Sampling - Source samples that contain a high concentration of
organic materials may require dilution prior to analysis to prevent saturating the GC
detector. The apparatus required for this direct interface procedure is basically the same
as described above, except a dilution system is added between the heated sample line
and the gas sampling valve. The apparatus is arranged so that either a 10:1 or 100:1
dilution of the source gas can be directed to the chromatograph.
Following are the procedures recommended for extracting a sample from the stack,
diluting the gas to the proper level, transporting the sample through a heated sample
line, and introducing it to the heated sample loop and the GC. The analysis of the
sample is described in Appendix C.3.
1.
3.
4.
C-30
Assemble the apparatus by connecting the heated box, as shown in Figure C.7,
between the heated sample line from the probe and the gas sampling valve on the
chromatograph. Vent the source gas from the gas sampling valve directly to the
charcoal filter, eliminating the pump and rotameter. ,--'
Measure the stack temperature, and adjust all heating units to a temperature 0°C
to 3°C above this temperature. If the temperature is above the safe operating
temperature of the Teflon components, adjust the heating to maintain a
temperature high enough to prevent condensation of water and organic com-
pounds. Heating is typically more critical for stacks that require dilution. The
check of removing the probe from the stack as recommended above should be
demonstrated.
After the heaters have come to the proper temperature, connect a U-tube H2O
manometer or equivalent to the inlet of the probe. Turn on the pump and pull a
vacuum of about 10 in. of H2O. Shut off the needle valve and then turn off the
pump. The vacuum reading should remain stable for 30 seconds. If a leak is
present, repair and then recheck the system.
Verify operation of the dilution system by introducing a calibration gas at the inlet
of the probe. The diluted calibration gas should be within 10 percent of the
calculated value. If the results for the diluted calibration gas are not within
10 percent of the expected values, determine whether the GC and/or the dilution
system is in error. If the analyses are not within acceptable limits because of the
dilution system, correct it to provide the proper dilution factors. Make this
correction by diluting a high concentration standard gas mixture to adjust the
dilution ratio as required. The dilution factor must be correct to obtain the true
value. A dilution system can give proper results, but it can also provide very
poor results when improperly conducted.
-------
I
IQ
6
a
I
u
I
an
§
U
« 5
.S.g
C-31
-------
5. Verify the GC operation using a low concentration standard by diverting the gas
into the sample loop and bypassing the dilution system. If these analyses are not
within acceptable limits, correct the GC by recalibration, etc. The test
coordinator should verify these calculations.
6. Conduct the analyses of the two audit samples as described in Appendix G3 using
either the dilution system or directly connect the gas sampling valve as required.
The results must agree within 10 percent of the true value or greater value if
specified on the cylinder. If the results do not agree, repair the system and repeat
the analyses until agreement is met or until approval is given by the
representative of the Administrator. The performance audit is critical and should
go through the dilution system when possible.
7. After the dilution system and GC operations are properly verified and the audit
successfully completed, place the probe at the centroid of the duct or at a point
no closer to the walls than 1 meter, and purge the sampling system for at least 10
minutes at the proper flow rate. Conduct the analysis of the sample as described
in-Appendix C.3. Record the field and analytical data on the applicable data
forms. Ensure that the probe, dilution system, and sample lines are maintained at
0°C to 3°C above the stack temperature (or a temperature which prevents
condensation).
8. Conduct the post-test calibration and verification of the dilution system as
described in Appendix C.3. Check the calibration calculations.
If the dilution system is used for bag sampling, the procedures for verifying operation
of the dilution system will be the same as shown above. The diluted calibration gas will
be collected in a bag and then verified. Also the audit samples will be collected in a bag
and analyzed. Acceptable results must be obtained for the audit samples prior to
analysis of the field samples.
G2.4 Adsorption Tube Sampling - Adsorption tube sampling can be used for those
organics specified in Table C-2, page C-4, and for other compounds as specified in the
National Institute of Occupational Safety and Health (NIOSH) methods. The selection
and use of adsorption tubes must be validated in the laboratory or through the use of the
literature.- This check will include selecting the proper adsorption material, and then
checking the capacity, breakthrough volume, adsorption efficiency, and desorption
efficiency. The adsorption efficiency can be greatly affected by the presence of water
vapor and other organics in, and temperature of, the stack gas. If sampling is
conducted for more than one organic compound, the adsorption and desorption effic-
iency checks must consider each. Because changes in process and control equipment
conditions can greatly affect all of the parameters stated above, it is recommended as a
standard operating procedure that more than one adsorption tube be used. The first
tube is analyzed as described in Appendix C3. If no problems are found, then the
second tube can be discarded. If problems with the first tube's adsorption efficiency are
discovered, then the primary section of the second tube can still be analyzed and the
results included with those of the primary portion of the first tube.
Following are the recommended procedures for adsorption tube sampling:
C-32
-------
.;-
1. The sampling system is assembled as shown in Figure C.8, page C-34. The
adsorption tube(s) must be maintained in a vertical direction for sampling. This
is done to prevent channeling of the gases along the side of a tube. It is
recommended that the sampling probe be eliminated when possible. If a sample
probe is used, it should be cleaned prior to its initial use with the extraction
solvent. Teflon tubing should be used for the probe and sample line.
2. Just prior to sampling, break off the ends of the adsorption tubes to provide an
opening at least one-half of the internal diameter. Audit samples must be
collected on the adsorption tubes during the test program as described in
Chapter 4.8. Since on-site analysis is typically not conducted when using
adsorption tubes, it is recommended that two samples be collected from each of
the two audit cylinders. This allows the tester a second chance to obtain the
proper value for each audit cylinder.
3. Prior to sampling and the collection of the audit samples, the sampling system
must be leak checked by connecting a U-tube H2O manometer or equivalent to
the inlet of the sample probe or adsorption tube. Turn the pump on and pull a
vacuum of about 10 in. of H2O. Shut off the needle valve and then turn off the
pump. The vacuum must remain stable for 30 seconds. If a leak is present,
repair and recheck the system. The leak check must be passed.
4. If the flow rate in the duct varies by more than 10 percent during the sampling
period, the sample should be collected proportionally. The proportional sampling
procedures will be the same as described for the bag sampling. The only
difference is that instead of using the volume of the bag as the limiting factor to
determine the average sampling rate, the breakthrough volume is the limiting
factor. If the source is a constant rate source (less than a 10 percent change in
flow rate for the sampling period), the samples can be collected at a constant rate.
5. Prepare the field blank just prior to sampling. The field blank will be handled in
be same manner as the field samples and should be from the same lot as the
other adsorption tubes. Blank correction can be allowed.
6. The flow rate meter must have been calibrated in the laboratory prior to the field
trip. The volume of sample collected must be accurately known for adsorption
tube sampling. The calibration data should be checked.
7. The sample run time must be equal to or greater than that specified by the
applicable regulation. During each sample run, the data should be recorded on
the sample data form as shown in Figure C.9, page C-35.
8. At the conclusion of each run, conduct another leak check as described above. If
the system does not pass the leak check, the run should be rejected, the leak
located and repaired, and another run conducted.
C-33
-------
CD fi
O E
w
2LJJO
2
c_
C-34
c.
O
o
!A
oo
u
8>
E
-------
"u"
g
at
^•k.
u
jm
•k
5
u
<«M
«r
e
"i
o
(V
.
e
s— *
CS r 1 S _
**** ^ *"" ^ KS
>» c o>+j e
ts •*- e «/> •»-
, •• gp ^^ ** m ^
pB yj jit gS ^A 1 1
«£•»
U
0)
^^K
0) X v» 0) 2 0>
4^ O * 4-» O JS O>
w> r— O «/t r— U C
|