PB84-155373

Performance Test Results and Comparative
Data for Designated Equilvalent
Methods for Sulfur Dioxide

Research Triangle Inst,
Research Triangle Park, NC

Prepared for

Environmental Monitoring Systems Lab.
Research Triangle Park, NC

Jan 84

NatM Tcdaiaf IdlmmVm %«r*M-


-------
P884-155373

EPA-600/4-84-015
January 1984

Performance Test Results and
Comparative Data for Designated Equivalent
Methods for Sulfur Dioxide

by

Raymond M. Micfiie, Jr.

Frederick W. Sexton
Center for Environmental

Measurements
Research Triangle Institute
Research Triangle Park, NIC 27709

Frank F. McElroy
Vinson L. Thompson
Methods Standardization Branch
Quality Assurance Division
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency
Research Trianglo Park, NC 27711

Contract Nos. 68-02-2714, 68-02-3222

EPA Project Officers: Frank F. McElroy, Vinson L. Thompson
Quality Assurance Division
Environmental Monitoring Systems Laboratory
U.S. Environmental Protection Agency

ENVIRONMENTAL MONITORING SYSTEMS LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
RESEARCH TRIANGLE PARK, NORTH CAROLINA 27711


-------
TECHNICAL REPORT DATA

(Please read Instructions on the reverse before completing}

1, REPORT NO. 2.

EPA-6Q0/4-84-Q15

3. RECIPIENT'S ACCESSION NO.

PR8 4 1 5 537, J/

4, TITLE AND SUBTITLE

Performance Test Results and Comparative Data for
Designated Equivalent Methods for Sulfur Dioxide

6. REPORT DATE . ' . '

January 1984

6 PERFORMING ORGANIZATION CODE

*

7**ul"W!iSllichiet Jr., F.W. Sexton, RT1, RTP/NC
F.F. McElroy, V.L. Thompson, EPA

B. PERFORMING ORGANIZATION REPORT NO.

9. PERFORMING ORGANIZATION NAME AND ADDRESS

Center for Environmental Measurements
Research Triangle Institue
Research Triangle Perk, NC 27709

10, PROGRAM ELEMENT NO.

11. CONTRACT/GRANT NO.

68-02-3222 and
68-02-2714

12. SPONSORING AGENCY NAM? AND.ADDRESS , .

Environmental Monitoring Systems Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711

13, TYPE CJF REPORT AND PERIOD COVERED

14. SPONSORING AGENCY CODE

EPA/600/08

18, SUPPLEMENTARY NOTES

16. ABSTRACT

The U.S. Environmental Protection Agency designates specific ambient
monitoring methods and analyzers as reference or equivalent methods acceptable
for use in certain required monitoring. Such designation is based on test
of the performance of the analyzer model by the manufacturer (or other
applicant) prior to designation. After designation, EPA carries out further
tests of the analyzer, including performance tests as well as comparative
tests to evaluate the performance, reliability, and operational peculiarities
of the analyzer with respect to other analyzers monitoring the same pollutant.

This report summarizes both the manufacturer's predesignation test
results and the available EPA postdesignation test results for currently
designated sulfur dioxide analyzers. Manufacturer's predesignation test
results are presented for thirteen analyzers, and all test results met or
exceeded the specifications. EPA postdesignation performance test results
are presented for ten analyzers, and similarly, all test results met or
exceeded the specifications. EPA postdesignation comparative test results
are presented for eight analyzers, with the results indicating that the
analyzers tested were generally stable and reliable and demonstrated a
high degree of comparability.		

17.	KEY WORDS AND DOCUMENT ANALYSIS

», DESCRIPTORS

b.IDENTIFIERS/OPEN ENDED TERMS

c. cosati Field/Group







B. DISTRIBUTION STATEMENT

RELEASE TO PUBLIC

19. SECURITY CLASS (This Report)

UNCLASSIFIED

21. NO. OF PAGES

139

20. SECURITY CLASS I This page)

UNCLASSIFIED

22. PRICE

EPA Form 2220-1 4-7?} previous edition i.« obsolete


-------
NOTICE

This document has been reviewed in accordance with
U.S. Environmental Protection Agency policy and
approved for publication. Mention of trade names
or commercial products does not constitute endorse-
ment or recommendation for use.

ii


-------
PREFACE

Measurement and monitoring research efforts are designed to anticipate
potential environmental problems, to support regulatory actions by develop-
ing an in-depth understanding of the nature and processes that impact
health and the ecology, to provide innovative means of monitoring compliance
with regulations, and to evaluate the effectiveness of health and environ-
mental protection efforts through the monitoring of long-term trends. The
Environmental Monitoring Systems Laboratory, Research Triangle Park, NC,
has responsibility for: assessing environmental monitoring technology and
systems; implementing Agency-wide quality assurance programs for air pollu-
tion measurement systems; and supplying technical support to other groups
in the Agency, including the Office of Air, Noise, and Radiation, the
Office of Toxic Substances, and the Office of Enforcement.

Assessment of environmental monitoring technology is performed in part
by the Quality Assurance Division. Through the Equivalency Testing Program,
automated analyzers are evaluated to ensure that explicit performance
specifications are met. This document is one in a series of four that
summarizes performance test results and comparative data obtained during
analyzer evaluations. The series will document evaluations of analyzers
that monitor ambient air for ozone, oxides of nitrogen, sulfur dioxide,
and carbon monoxide.

iii


-------
ABSTRACT

Under Part 53 of Title 40 of the Code of Federal Regulations (40 CF1
Part 53), the U.S. Environmental Protection. Agency designates specific
ambient monitoring methods and analyzera as reference or equivalent methods
acceptable for use in National Air Monitoring Stations (NAMS), State and
Local Air Monitoring Stations (SLAMS), and Prevention of Significant Deteri-
oration (PSD) monitoring. Such designation requires that extensive perform-
ance testing of the analyzer bs carried out by the manufacturer (or other
applicant) prior to designation. After designation, EPA carries out further
tests of the analyzer. These EPA tests include performance tests as well
as comparative tests to evaluate the performance, reliability, and opera-
tional peculiarities of the analyzer with respect to other analyzers moni-
toring the same pollutant.

This report summarizes both the manufacturer's predesignation test
results and the available EPA postdesignation test results for currently
designated sulfur dioxide analyzers. Manufacturer's predesignation test
results are presented for 13 analyzers. The manufacturer's tests were
conducted according to procedures presented in 40 CFR Part 53, and all test
results met or exceeded the specifications contained therein. EPA Phase I
postdesignation test results are presented for 10 analyzers. These tests
were conducted according to an abridged version of the 40 CFR Part 53
procedures. Again, all test results met or exceeded the CFR specifications.
EPA Phase II postdesignation test results are presented for eight analyzers.
Although these tests were primarily comparative in nature with no formal
performance specifications applicable to the data, the results indicated
that the analyzers tested were stable, reliable, and demonstrated a high
degree of comparability among themselves.

iv


-------
CONTENTS

Section.	Page

Foreword		ill

Abstract						iv

Figures ......... 	 .	vi

Tables				xii

Abbreviations and Symbols			xiii

Acknowledgment ....... 	 ................	xiv

1	Introduction . 				1

2	Postdesignation Testing ................ 	 .	4

3	Test Results and Data for Designated Sulfur Dioxide

Analyzers 					 .	11

ASARCO 500 				14

Beckraan 953 					17

Bendix 8303 . 				33

Lear Siegler SM1000 					41

Lear Siegler AM2020 				51

Meloy SA185-2A			53

Meloy SA285E . . 	 .............	63

Meloy SA700 . . 	 ................	72

Monitor Labs 8450 			77

Monitor Labs 8850 	.			87

Philips PW9700 	 ..... 		92

Philips PW9755 	 ...............	96

Thermo Electron. 43			105

Appendix A. Measurement Principle for SO2 Analyzers .... 	 .	117

v


-------
FIGURES

Number	Page

1	ASARCO Model 500 S02 analyzer (EQSA-0877-024). ASAJRCO

Incorporated, 3422 South 700 West, Salt Lake City,

Utah 04119	 14

2	Typical calibration curve for the ASARCO Model 500 ..... 16

3	Beckman Model 953 fluorescent ambient S02 analyzer
(F.QSA-0678-029) . Beckman Instruments, Inc., Process
Instruments Division, 2500 Harbor Drive, Fullerton,

CA 92634 		 1?

4	Typical calibration curve for the Beckman 953 	 20

5	Zero and span drift in the Beckman 953 SO2 readings

during the Phase II ambient monitoring test 	 21

6	Relationship between the Beckman 953 and the average
of the other analyzers during the Phase II ambient
monitoring test (before HC reactor replacement—see

General Comments)					 22

7	Frequency distribution of differences in hourly S02
ambient air readings between the Beckman 953 and the
average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test
(before HC reactor replacement--see General Comments).

All readings corrected for zero and span drift ....... 23

8	Frequency distribution of differences in hourly S02
ambient air readings between the Beckman 953 and the
average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test
(before HC reactor replacement--see General Comments).

Beckman 953 readings are not corrected for zero and

span drift		 24

9 Relationship between the Beckman 953 and the average
of the other analyzers during the Phase II ambient
monitoring test (after HC reactor replacement--see
General Comments) 	 ......... 25

vi


-------
FIGURES (continued)

Number

10

11

1 ?

13

14

15

16

17

18

19

Page

Frequency distribution of differences in hourly SO2

ambient air readings between the Beckraan 953 and

the average of simultaneous readings from the other

analyzers during the Phase II ambient monitoring test

(after HC reactor replacement--see General Comments),

All readings corrected for zero and span drift ....... 26

Frequency distribution of differences in hourly S02

ambient air readings between the Beckman 953 and the

average of simultaneous readings from the other

analyzers during the Phase II ambient monitoring test

(after HC reactor replacement--see General Comments).

Beckman 953 readings are not corrected for aero and

span drift	 27

Bendix Model 8303 sulfur dioxide analyzer (EQSA-1078-030).

The Bendix Corporation Environmental and Process Instru-
ments Division, P.O. Box 831, Louisburg, WV 24901 	 33

Typical calibration curve for the Bendix 8303 ....... 35

Zero and span drift in the Bendix 8303 SOg readings

during the Phase II ambient monitoring test ........ 36

Relationship between the Bendix S303 and the average

of the other analyzers during the Phase II ambient

monitoring test					 37

Frequency distribution of differences in hourly SO2

ambient air readings between the Bendix 8303 and the

average of simultaneous' readings from the other

analyzers during the Phase II ambient monitoring test.

All readings corrected for zero and span drift	 38

Frequency distribution of differences in hourly S02

ambient air readings between the Bendix 8303 and the

average of simultaneous readings from the other

analyzers during the Phase II ambient monitoring test.

Bendix 8303 readings are not corrected for zero and

span drift 	 ............. 	 39

Lear Siegler SM1000 SOg ambient monitor (EQSA-1275-005).

Lear Siegler, Inc., Environmental Technology Division,

74 Inverness Drive last, Englewood, CO 80112 	 41

Typical calibration curve for the Lear Siegler SM1000 ... 43

vii


-------
FIGURES (continued)

Number	Page

20	Zero and span drift in the Lear Siegler SN1000 S02

readings during the Phase II ambient monitoring test .... 44

21	Relationship between the tear Siegler SMI000 and the
average of the other analyzers during the Phase II

ambient monitoring test 	 45

22	Frequency distribution of differences in hourly S02
ambient air readings between the Lear Siegler SM1000
and the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test.

All readings corrected for zero and span drift . . 	 46

23 Frequency distribution of differences in hourly S02
ambient air readings between the Lear Siegler SM10Q0
and the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test,
tear Siegler SMI000 readings are not corrected for
zero and span drift . . , ,	 47

24	Lear Siegler Model AK2020 ambient SO2 monitor (EQSA-
1280-049). Lear Siegler, Inc., Environmental
Technology Division, 74 Inverness Drive East,

Englewood. CO 80112		 51

25	Meloy Model SA185-2A sulfur dioxide analyzer (EQSA-
1275-006). Columbia Scientific Industries, 11950

Jollyville Road, Austin, TX 78759 				 . 53

26	Typical calibration curve for the Meloy SA185-2A 	 56

27	Zero and span drift in the Meloy SA1S5-2A S02 readings

during the Phase II ambient monitoring test		 . 57

28	Relationship between the Meloy SA185-2A and the average
of the other analysers during the Phase II ambient
monitoring test					 58

29	Frequency distribution of differences in hourly S02
ambient air readings between the Meloy SA185-2A and
the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test.

All readings corrected for aero and span drift ....... 59

30 Frequency distribution of differences in hourly S02
ambient air readings between the Meloy SA185-2A and
the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring
test. Meloy SA185-2A readings are not corrected
for zero and span drift 				60

viii


-------
FIGURES (continued)

Number	Page

31	Meloy Mode- SA285E suliur dioxide analyzer (EQSA-
1078-032). Columbia Scientific Industries, 11950

Jollyville Road, Austin, TX 78759 	 63

32	Typical calibration curve for the Meloy SA285E 	 66

33	Zero and span drift in the Meloy SA285E S02 readings

during the Phase II ambient monitoring test 	 67

34	Relationship between the Meloy SA285E and the average
of the other analyzers during the Phase II ambient

monitoring test	 68

35	Frequency distribution of differences in hourly S02
ambient air readings between the Meloy SA285E and the
average of simultaneous readings iroro the other
analyzers during the Phase II ambient monitoring test.

All readings corrected for zero and span drift	 69

36	Frequency distribution of differences in hourly SOg

ambient air readings between the Meloy SA285E and the

average of simultaneous readings from the other

analyzers during the Phase II ambient monitoring test.

Meloy SA285E readings are not corrected for zero

and span drift	 70

37	Meloy Model SA700 fluorescence sulfur dioxide analyzer
(ESQA-0580-046). Columbia Scientific Industries,

11950 Jollyville Road, Austin, TX 78759 . 		 72

38	Typical calibration curve for the Meloy SA700 	 75

39	Monitor Labs Model 8450 sulfur monitor (EQSA-0876-013).

Monitor Labs, Incorporated, 10180 Scripps Ranch

Boulevard, San Diego, CA 92131 	 77

40	Typical calibration curve for the Monitor Labs 8450 .... 80

41	Zero and span drift in the Monitor Labs 8450 S02 readings
during the Phase II ambient monitoring test 	 81

42	Relationship between the Monitor Labs 8450 and the
average of the other analyzers during the Phase II

ambient monitoring test 	 82

43	Frequency distribution of differences in hourly S02
ambient air readings between the Monitor Labs 8450 and
the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test.

All readings corrected for zero and span drift	 83

ix


-------
FIGURES (continued)

Number	Page

44 Frequency distribution of differences in hourly S02
ambient air readings between the Monitor Labs 8450
and the average of simultaneous readings from the
other analyzers during the Phase II ambient monitoring
test. Monitor Labs 8450 readings are not corrected
for zero and span drift			 84

45	Monitor Labs Model 8850 fluorescent S02 analyzer (EQSA-
0779-039), Monitor Labs, Incorporated, 10180 Scripps

Ranch Boulevard, San Diego, CA 92131 ............ 87

46	Typical calibration curve for the Monitor Labs 8850 .... 89

47	Philips PW9700 SO2 analyzer (EQSA-0876-011). Philips
Electronic Instruments, Inc., 85 McKee Drive, Hahwah,

NJ 07430 		 			 -	. 92

48	Typical calibration, curve for the Philips PW9700 . 	 94

49	Philips PW9755 sulfur dioxide analyzer (EQSA-0676-01G).

Philips Electronic Instruments, Inc., 85 McKee Drive,

Mahwah, NJ 07430 	 ...... 96

50	Typical calibration curve for the Philips FW9755 ...... 98

51	Zero and span drift in the Philips PW9755 S02 readings

during the Phase II ambient monitoring test 	 99

52	Relationship between the Philips PW9755 and the average
of the other analyzers during the Phase II ambient

monitoring test ...... 	 ..... 	 100

53	Frequency "distribution of differences in hourly S02
ambient air readings between the Philips PW9755 and
the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test.

All readings corrected for zero and span drift ....... 101

54	Frequency distribution of differences in hourly S02

ambient air readings between the Philips PW9755 aad
the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test.

Philips PW9755 readings are not corrected for zero
and span drift 						102

55	Thermo Electron Model 43 pulsed fluorescent S02 analyzer,,

(EQSA-0276-009), Thermo Electron Corporation, Environ-
mental Instruments Division, 108 South Street, Hopkinton,

MA 01748 				 . 105

x


-------
FIGURES (continued)

Number	Page

56	Typical calibration curve for the Thermo Electron 43 ... . 107

57	Zero and span drift in the Thermo Electron A3 S02 readings
during thePhase II ambient monitoring test .... 	 108

58	Relationship between the Thermo Electron 43 and the
average of the other ana(yzars during the Phase II
ambient monitoring test (before manufacturer's

refurbishraent--see General Comments) .... 	 109

59	Frequency distribution of differences in hourly SO2
ambient air readings between the Thermo Electron 43
and the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test
(before manufacturer's refurbishment—see General Com-
ments). All readings corrected for zero and span drift . . 110

60	Frequency distribution of differences in hourly SO2
ambient air readings between the Thermo Electron 43
and the average of simultaneous readings from the other
analyzers during the Phasa II ambient monitoring test
(before manufacturer's refurbishment—see General
Comments). Thermo Electron 43 readings are not

corrected for zero and „i-an drift	Ill

61	Relationship between the Thermo Electron 43 and the
average of the other analyzers during the Phase II
ambient monitoring test (after manufacturer's

refurbishment—see General Comments) 	 112

62	Frequency distribution of differences in hourly S02
ambient air readings between the Thermo Electron 43
and the average of simultaneous readings from the other
analyzers during the Phase II ambient monitoring test
(after manufacturer's refurbi shment--see General Comments).

All readings corrected for zero and span drift ....... 113

63	Frequency distribution of differences in hourly S02
ambient air readings between the Thermo Electron 43
and the average of simultaneous readings from the
other analyzers dut*ing the Phase II ambient monitoring
test (after manufacturer's refurbishment--see General
Comments). Thermo Electron 43 readings are not

corrected for zero and span drift 		114

xi


-------
1

2

3

4

5

6

7

8

9

10

11

12

13

14

15

Pag

11

13

15

19

34

42

52

55

65

74

79

88

93

97

106

TABLES

Analyzers Included in the Postdesignation Testing of
Sulfur Dioxide Analyzers 	

Estimates of Total Operating Precision and Minimum
Detection Limits 	 .........

ASARCO 500 Laboratory Performance Test Results ....

Beckuian 953 Laboratory Performance Test Results . . .

Bendix 8303 Laboratory Performance Test Results . . .

Lear Siegler SHI000 Laboratory Performance Test Results

Lear Siegler AM2020 Laboratory Performance Test Results

Meloy SA185-2A Laboratory Performance Test Results . .

Meloy SA285E Laboratory Performance Test Results . . .

Meloy SA700 Laboratory Performance Test Results . . .

Monitor Labs 8450 Laboratory Performance Test Results

Monitor Labs 8850 Laboratory Performance Test Results

Philips PW9700 Laboratory Performance Test Results . .

Philips PW9755 Laboratory Performance Test Results . .

Thermo Electron 43 Laboratory Performance Test Results

xii


-------
ABBREVIATIONS AND SYMBOLS

ABBREVIATIONS

cm3/mxn

CFR

DAMDF

EMSL

EPA

FPD

h

Hz

kPa

L

1DL

HDL

mm

MSB

NAMS

NBS

PCS

PFD

PMT

ppb

ppm

PSD

psig

QAD

SLAMS

SRM

torr

TFE

(jm

URL

UV

Vac

SYMBOLS

CO

C02

H20

H2G(v)

NO

NO

HO?

03

S02

--cubic centimeter per minute

—Code of Federal Regulations

--Durham Air Monitoring Demonstration Facility

—Environmental Monitoring Systems Laboratory

—U.S. Environmental Protection Agency

—flame photometric detection

--hour

—hertz (frequency in cycles per second)

—kilo pascal
—liter

--lower detectable limit
—minimum detectable limit
--millimeter

—Methods Standardization Branch
—National Air Monitoring Station
—National Bureau of Standards
--printed circuit board
--pulsed fluorescent detection
—photomultiptier tube
—parts per billion
—parts per million

—prevention of significant deterioration

—pounds per square inch, gauge

—Quality Assurance Division

—State and Local Air Monitoring Station

—standard reference material

—unit of pressure (1 ram Hg @ 0° C or 133 N/ra2)

—tetrafluoroethylene (Teflon)

—micrometer

--upper range limit

—ultraviolet

--volts, alternating current

—approximately
—carbon monoxide
—carbon dioxide
—water
—water vapor
—nitric oxide
—oxides of nitrogen
—nitrogen dioxide
—ozone

—sulfur dioxide

xiii


-------
ACKNOWLEDGMENT

The authors gratefully acknowledge the invaluable statistical assistance
of Mr. David Holland of the Data Management and Analysis Division, Environ-
mental Monitoring Systems Laboratory, U.S. Environmental Protection Agency.

xiv


-------
SECTION 1

INTRODUCTION

Under Part 53 of Title 40 of the Code of Federal Regulations (40 CFR
Part 53), the U.S. Environmental Protection Agency (EPA) designates specific
ambient monitoring methods and analyzers as acceptable for use in National
Air Monitoring Stations (NAMS), State arid Local Air Monitoring Stations
(SLAMS), and Prevention of Significant Deterioration (PSD) monitoring,
among other monitoring applications. These methods are designated as
reference methods or equivalent methods, depending on whether or not the
method employs a particular measurement principle and calibration procedure
specified for each pollutant in 40 CFR Part 50, A list of all currently
designated methods is available from the Quality Assurance Division (QAD)
or from the Quality Control Coordinator at any EPA Regional Office.

Commercial analyzer models are designated as reference or equivalent
methods based on an application submitted by the manufacturer (or other
applicant). This application must show that a test analyzer representative
of the analyzer model meets the following criteria;

1.	It has been tested according to specific test procedures
prescribed in 40 CFR Part 53;

2.	It meets prescribed minimum performance specifications; and

3.	It meets certain other requirements such as having a suitable
calibration procedure and an adequate operation and instruc-
tion manual.

Any designated analyzer model should meet the requirements mentioned
above. However, these analyzer models may differ considerably in design,
features, reliability, measurement principle, ease of operation, and level
of performance above the required minimum. These factors allow for intelli-
gent selection of the best or most appropriate analyzer for a particular
application.

The Methods Standardization Branch (MSB) of the Environmental Monitoring
Systems Laboratory (EMSL) at Research Triangle Park, NC, is responsible for
EPA's reference and equivalent method designation program for ambient air
monitoring methods. In this capacity, MSB acquires performance data and
other information on these methods, particularly on the various commercial
analyzer models that have been designated as reference or equivalent methods.

1


-------
To summarize this information and make it available to those who may
find it beneficial in selecting analyzers, MSB, with the assistance of the
Research Triangle Institute (RTI), has prepared this report on designated
methods for sulfur dioxide (S02). Other reports discuss designated methods
for ozone, nitrogen dioxide, and carbon monoxide.

The information presented in this report is derived from the following
sources;

1.	Predesignation test data submitted by the applicant,

2.	Postdesignation tests conducted by MSB,

3.	Reports or comments from analyzer users, and

4.	Other pertinent information as needed,

The predesignation test data were extracted from the application for
designation submitted by the applicant (usually the analyzer manufacturer)
and were thus obtained from tests conducted at the applicant's own laboratory
or test site. These tests must be conducted in accordance with the explicit
test procedures specified in 40 CFR Part 53. Results must then be calculated
and interpreted according to the prescribed specifications.

The postdesignation tests conducted by MSB are part of an ongoing
postdesignation test program and may include one or more of the following
types;

1.	Phase I tests are laboratory performance tests similar to
the predesignation tests required by 40 CFR Part 53;

2.	Phase II tests simulate actual use conditions and serve to
intercompare designated methods; and

3.	Phase III tests are special purpose tests for specific
analyzers and are used when unique needs or circumstances
arise.

Postdesignation test results are the major source of information in this
report and are described in detail in Section 2. The tests are conducted
either by Quality Assurance Division personnel or by a contractor, under
QAD direction, who has no contractual or financial association with any
commercial manufacturers of designated analyzers. Because of resource
limitations, there may be substantial delays in completing the postdesigna-
tion tests of an analyzer model following its designation as a reference or
equivalent method.

The information in this report is based on files maintained by MSB for
each designated method. MSB encourages analyzer users to subn.it comments
that might be of interest to others using or contemplating purchase of a
designated analyzer. Such comments, particularly negative comments, are

2


-------
more valuable if they are accompanied by supporting data, records, or other
documentation. Pertinent information from other sources, such as other EPA
groups, testing laboratories, and manufacturers will also be included when
available. Publication of an updated report will be considered whenever
significant new information is available.

Additional information on any aspect of the report may be obtained

from MSB. The report is not intended to recommend or endorse any particular
analyzer over others. MSB welcomes any comments on the usefulness of these
reports or suggestions for improvement. Remarks may be mailed to the
address on the front page or made by phone to any of the following numbers:

Commercial: (919) 541-2622
(919) 541-3791

FTS:	629-2622

629-3791

3


-------
SECTION 2
POSTDESIGNATION TESTING

Because the primary predesign ition testing of an analyzer is carried
out by the manufacturer or other applicant, it is EPA's general policy to
perform followup performance tests on the analyzer after designation.

These tests are referred to as postdesignation tests and are of three
types: Phase I, Phase II, and Phase III. Phase I and II tests are described
below. Phase III tests include any special tests or special circumstances
not covered by Phase I and II and are conducted only when deemed necessary
or appropriate.

The postdesignation tests are usually carried out after designation on
an analyzer purchased by EPA through normal procurement procedures. Occa-
sionally, an analyzer purchased prior to designation is returned to the
manufacturer for upgrading to designated status. Commercially unavailable
analyzers are borrowed for the tests. The manufacturer is always notified
of the testing to be done and of the fact that the test analyzer obtained
is assumed to be representative of the model. Because of the complex
nature of most analyzers, variability in performance from one instrument to
another can be expected. Even though all analyzers must be within the
required specifications, this potential variability must be kept in mind
when the test results in this report are considered and compared with those
of other analyzers.

PHASE I TESTS

General Description

Phase I tests are laboratory performance tests that are conducted in
accordance with the same test procedures and specifications required for
the applicant's predesignation tests. The one exception is that fewer
trials (usually four) are performed for each test parameter to reduce
costs. All calibrations, test apparatus, pollutant standards, test proce-
dures, test atmospheres, and test documentation are as specified in 40 CFR
Part 53. An individual report containing more detailed information on the
Phase I test for each analyzer tested is available from MSB,

Analyzers to be Phase I tested are set up and allowed to operate for
several weeks prior to actual testing. During this startup period, prelim-
inary calibrations and linearity checks are performed to verify proper
operation. In the event that unusual behavior such as nonlinearity or
inoperability due to a major malfunction is identified, the manufacturer is

4


-------
notified. The manufacturer is encouraged to become involved throughout the
startup period so that potential problems may be identified prior to testing
If a failure occurs during testing, the manufacturer is notified and given
the opportunity to repair the analyzer. At the conclusion of the test, all
failures and manufacturer involvement are documented and reported.

Data interpretation

The Phase I test results for each analyzer are shown in Tables 3
through 15 along with the EPA performance specifications and the applicants'
own predesignation test data (as submitted in the equivalency application).
The results reported are averages of several trials for each parameter
(usually four trials for the EPA Phase I data and seven for the applicant's
predesignation data). The test results must generally be less than (or
within) the EPA performance specifications shown for an analyzer to achieve
designation, and, except for the lower detectable limit, these results are
comparable in a relative sense from one instrument to another. The manufac-
turer's advertised or claimed performance specifications, however, may not
be directly comparable to the test data or to other analyzers since manu-
facturers may choose to define their performance specifications, or how
they are measured, differently. The various EPA performance parameters are
described briefly below along with pertinent comments. See 40 CFR Part 53
for the formal definitions of the performance parameters and the prescribed
test procedure for each.

Noise--

Noise refers to short-term deviations in the analyzer output that are
not caused by changes in the input concentration. Noise is measured at
both 0 percent and 80 percent of the upper range limit (URL) and is the
standard deviation of 25 readings taken at 2-minute intervals. Normally,
this value will be slightly less than the average absolute deviation from
the mean value and is usually much less than the peak-to-peak noise.

tower Detectable Limit*--

The lower detectable limit (LDL) is the smallest pollutant concentra-
tion that produces an output change of at least twice the noise level. The
test does not determine the actual LDL; it only verifies that the output
signal at an input concentration of 0.01 ppm is at least two times the O-ppm
noise level. Therefore, the value reported is usually close to 0.01 ppm and
must be greater than twice the noise level. An output signal significantly
lower than 0.01 ppm may indicate a nonapparent negative zero offset or other
low-level response problems.

Interference Equivalent--

The interference equivalent is the relative degree of unwanted response
caused by compounds other than sulfur dioxide present in the ambient air
sample. The reported values should not be interpreted as the actual error
to be expected in ambient measurements because the test concentrations of
most of the interferents are higher than those normally encountered in
ambient air. However, the values for water and carbon dioxide (C02) do

5


-------
represent the approximate error that may result when the calibration gases
are dry and/or free of C02.

Zero Drift (12 h)--

The 12-h zero drift is the maximum change in continuous output response
to zero input concentration over a 12-h period of unadjusted operation.

The 12-h zero drift includes any peak-to-peak noise component, so it may be
higher than the 24-h zero drift.

Zero Drift (24 h)--

The 24-h zero drift is the net change in response to zero input concen-
tration between two points in time ~24 h apart during unadjusted analyzer

operation. It includes zero drift caused by changes in ambient temperature,
line voltage, or both. The 24-h zero drift tends to be sometimes positive,
sometimes negative; hence the average would be misleadingly low. Therefore,
the average of the absolute values of the analyzer's daily drift is reported
in Tables 3 through 15 as a more meaningful value.

Span Drift (20% URL) —

The span drift at 20 percent URL is the net change in response to an
input concentration of ~20 percent of the URL between two points in time
~24 h apart during unadjusted analyzer operation and including the effects
of changes in ambient temperature, line voltage, or both. Because the test
procedure specifies no correction for zero drift, the reported value represents
a combination of both zero drift and span drift over the period. The
average of the absolute values is reported.

Span Drift (80% URj,) —

The span drift; at 80 percent URL is the net change in response to an
input concentration of ~80 percent of URL between two points in time ~24 h
apart during unadjusted analyser operation and including the effects of
changes in ambient temperature, line voltage, or both. The 80-percent URL
span drift is also calculated without a correction for zero drift; therefore
it may be higher or lower than the actual change in the slope of the calibra-
tion curve. Again, the average of the absolute values of the individual
daily results is reported.

Lag Time—

Lag time is the time required to observe the first change in response
resulting from a sudden change in input concentration.

Rise Time—

Rise time is the time (not including the lag time) required for the
response to reach 95 percent of its final value after a sudden increase in
the input concentration.

Fall Time—

Fall time is the time (not including lag time) required for the response
to decrease to 5 percent of its previous value when the input concentration
is suddenly changed to 0 ppm.

6


-------
Long lag, rise, or fall times generally do not adversely affect meas-
urements expressed as hourly averages and may even be beneficial where
concentrations fluctuate rapidly. However, long response times may be an
annoyance or source of error during calibration and zero and span drift
adjustment.

Precision-

Precision is the variation in the response value to six repeated
measurements of the same upscale pollutant concentration, approached alter-
nately from lower and higher concentrations. The reported precision is a
measure of the repeatability of the analyzer and is determined at "both 20
and 80 percent of the URL.

PHASE II TESTS

General Description

Phase II tests are intended to test analyzers in a more or less typical
ambient monitoring configuration where their stability, reliability, general
performance, and operational peculiarities can be observed, recorded, and
compared. The tests are conducted simultaneously on a group of analyzers
during a period of several months. All analyzers measure the same pollutant
in the ambient air, which is sampled from a common manifold. The ambient
pollutant concentrations are sometimes augmented in the manifold with arti-
ficially generated pollutant to increase the concentration readings.

Unless otherwise specified, the Phase ,11 tests are carried out at
EPA's Durham Air Monitoring Demonstration Facility (DAMDF) in an urban/
industrial/commercial area of Durham, NC. Ambient air samples for the
analyzers are drawn from an intake on the building roof and distributed by
an all-glass, power-ventilated manifold system. Pollutant levels are
augmented, when appropriate. During such augmentation, however, all sample
air contains at least 90 percent unaltered ambignt air. The analyzers'
inlets are connected to the manifold via Teflon particulate filters and
either 3-mra (1/8-in) or 6-mm (1/4-in) Teflon tubing, depending on the
analyzer flow rate. The sample lines are kept as short as practical.

All test analyzers are installed, calibrated, operated, and maintained
in accordance with the manufacturer's operation/instruction manual and good
monitoring practice. Analyzers receive a multipoint calibration initially
and once per month during the test period. Two-point zero and span checks
are made two or three times per week, with zero adjustments only if the
zero response is not within ±3 percent of full-scale response from nominal,
and span adjustments only if the span (slope of calibration curve) changes
by more than ±7 percent from nominal.

Zero air is obtained from filtered ambient air by appropriate drying
and chemical scrubbing. Calibration and span gases are obtained by tech-
niques such as dilution of high concentration cylinder gases, permeation
devices, and gas phase titration. Ozone (O3) concentrations are generated
by ultraviolet (UV) radiation and are traceable to a UV standard; all other

7


-------
pollutant concentrations are traceable to National Bureau of Standards
Standard Reference Materials (NBS-SRMs).

Analyzer readings are recorded at 1-uiin intervals using a Monitor Labs
9300 data acquisition system. One-hour averages (50-rain minimum) are auto-
matically calculated and transferred to magnetic tape. Zero and span
responses are also recorded on the tape. The tapes are read into a Hewlett-
Packard 9845A Desk Top computer for verification, editing, corrections as
necessary, and calculation of the various statistical comparisons.

Data Interpretation

I II	II. .umiainim ,, ,		II,.,

Phase II testing is primarily intended to be comparative in nature.
As such, the test does not require conformance to any formal performance
specifications , and the analyzers neither "officially pass" nor "officially
fail" the comparisons. The Phase II test results and other data are pre-
sented in Section 3 as a series of figures for each analyzer. The figures
in each analyzer's series (Figures 4 through 8, for example) are discussed
in their order of appearance. The first (Figure 4, for example) is a
typical multipoint calibration curve for the analyzer, showing five or more
points spaced over the full-scale range, calculated by least squares
regression from the actual calibration response data. The span is normally
adjusted to a slope of ~200 (scale percent/ppm) for a 0~to-0,5~ppm range.
The correlation coefficient provides an empirical measure of the linear
association between the analyzer response and the calibration concentrations.

The chronological records of the zero and span drift for each analyser
are shown next in the series of figures (Figure 5, for example). Zero
drift is plotted in absolute units (ppra) and the span drift is plotted as
the percent difference of the span slope from the nominal slope (not
percent-of-scale change in an arbitrary span concentration). Both zero and
span drifts are total accumulated drifts except where manual adjustments
are explicitly indicated by diamonds situated on the x-axis. The drifts
are referenced to the nominal intercept (0 percent of scale) and nominal
slope (200) represented by the horizontal, dashed lines. Also shown are
the nominal adjustment limits for zero (±3 percent of full scale or ±15 ppb)
and span (±7 percent). All adjustments, whether made because nominal drift
limits were exceeded or during routine calibrations, are indicated by the
diamonds. It is often not possible to adjust the zero and spaa to the
exact nominal values desired because of the nature of thn controls. The
drift plots are intended mainly for visual comparison among the various
analyzers.

The statistics shown above each chart give (1) the total net drift
over the entire test period, ignoring adjustments and resets; (2) the
number of individual drift periods—i.e., periods between zero and span
checks; (3) the average length of the drift periods in days; (4) the average
absolute value of the drift for each period (i.e., ignoring the signs of
the drifts); and (5) the standard deviation of the individual drifts over
all drift periods. The marker at the right side of each chart illustrates
a drift range of ±3 standard deviations.

8


-------
The next figure (Figure 6, for example) is a scatter diagram comparing
each subject analyzer measurement (corrected for zero and span drift) with
the average of the corresponding measurements from the other analyzers
during simultaneous, continuous monitoring of ambient air over approximately
150 days. The linear regression relationship estimated from the data
(solid line) is shown along with a dashed line representing the ideal
linear relationship (slope = 1, intercept = 0). The linear regression
slope and intercept have been tested statistically to determine if they are
significantly different from 1 and 0, respectively. In general, the regres-
sion slopes and intercepts are significant even though they approximate the
ideal values because of the large number of points (N) used in the regression.
Also shown is the sample correlation coefficient (r).

The additional dashed lines on the scatter diagrau show statistical

bounds that should include 90 percent of future analyzer measurements (¥)
for any concentration (X) as measured by the average of the reference
analyzers. This is a 95-percent confidence interval, based on the estimated
linear regression and centered on the estimated regression line, applicable
to 90 percent of future observations.*

During the S02 Phase II test, ambient levels of SO2 were augmented
("spiked") with additional, extraneous S02 from a permeation system to
allow evaluation of the instruments at concentration levels greater than
the rather low ambient levels available at the test site. Concentration
levels for the augmentation were approximately 75 ppb (15 percent of full
scale), 125 ppb (25 percent of scale), and 250 ppb (50 percent of scale),

Each augmentation period lasted from 15 to 19 days. These augmentation
periods account for the grouping or clustering of the data points observed
on the scatter diagrams,

The next figure (Figure 7, for example) presents another comparison of
the differences in hourly ambient air readings between subject analyzer and
the average of simultaneous readings from other analyzers. Corrections for
zero and span drifts are made before the data are compared. The comparison
is shown in the form of a frequency distribution indicating numerically and
graphically the frequency (number of hours out of the total test period) of
the occurrence of each discrepancy versus its magnitude and sign. Also
given are the correlation coefficient, the mean difference (bias), the
standard deviation of the differences (spread), and the number of absolute
differences >20 ppb; all of these statistics are calculated with respect to
the differences between the hourly readings of the subject analyzer and the
average of the corresponding hourly readings from the other analyzers in
the test group.

*A more detailed description of
be found in Lieberman, G. J., and R.

intervals in regression, Biometrlka,

this type of confidence interval can
G. Miller, Simultaneous tolerance
50:155-168, 1963,

9


-------
Last is a figure (Figure 8, for example) showing the same type of
frequency distribution discussed above but without periodic zero and span
corrections; all hourly averages for the subject analyzers are based on the
original calibration curve. The reference data from the analysers on which
the comparison is based are still corrected for zero and span drift.
Comparison of the data in corresponding sets of frequency distributions,
particularly the mean difference and the standard deviation of the differ-
ences , illustrates the effect of making periodic zero and span corrections
to ambient monitoring data. Usually, the agreement of the hourly readings
(bias, spread, or both) is improved by the periodic zero and span adjust-
ments, and the practice is generally recommended where possible.

10


-------
SECTION 3

TEST RESULTS AND DATA FOR DESIGNATED SULFUR DIOXIDE ANALYZERS

This section contains summary information on Phase I and Phase II
postdesignation tests performed on 12 models of sulfur dioxide analyzers
(see Table 1) and predesignation test data for 13 models as reported in the
corresponding applications for designation as equivalent methods.

TEST RESULTS

Tables 3 through 15 summarize the available Phase I laboratory perform-
ance test results for each analyzer and compare these data with EPA specifi-
cations and with the manufacturer's predesignation test results. Phase II
test results for each tested analyzer are shown in a series of figures
contained in each analyzer subsection. See Section 2 for an explanation of
the tests and figures.

TABLE 1. ANALYZERS INCLUDED IN THE POSTDESIGNATION TESTING OF

SULFUR DIOXIDE ANALYZERS

Manufacturer



Model

Serial No.

Phase I tests

Phase II tests

ASARCO



500

None

5/78-1/79

Not anticipated

Beckman #1
Beckman #2



953
953

1000039
1000865

5/82-4/83

11/78-6/79

Bendix



8303

11252

Planned

2/79-6/79

Lear Siegler



SM1000

7564

9/76-5/77

11/78-6/79

Lear Siegler



AM2020

0910020

Planned

Not planned

Me loy



SA185-2A

4H024

10/76-11/76

11/78-6/79

Me loy



SA285E

7E140

Planned

11/78-6/79

Me loy



SA700

OF106

5/82-3/83

Not planned

Monitor Labs



8450

147

3/77-8/77

11/78-6/79

Monitor Labs
Monitor Labs

# 1
#2

8850
8850

90
717

8/82-6/83

Not planned

Philips



PW9700

D1415C

5/78-1/79

Not planned

Philips



PW9755 _

612

5/78-1/79

11/78-6/79

Thermo Electron

43

ASM-3294-64

10/76-1/78

11/78-6/79

11


-------
Determination of Minimum Detection Limit (MDL)

Of particular interest lor any air monitoring analyzer is its minimum
measurement capability. MDL can be defined as the lowest concentration
that an analyzer can reliably distinguish from zero concentration. Note
that this MDL is very similar to the lower detectable limit defined in
40 CFR Part 53 and discussed in Section 2. However, the test for LDL does
not actually measure the LDL, it only determines that the LDL is greater
than twice the noise level. In this study, an estimate of the MDL for each
analyzer was determined by the following equation;

MDL = 3 x precision,

where the value 3 is the minimum recommended by the "Guidelines for Data
Acquisition arid Data Quality Evaluation in Environmental Chemistry"* and
implies a risk of about 7 percent for concluding that a detectable concen-
tration of S02 is present when the concentration is actually zero.

Rather than use the precision defined in 40 CFR Part 53 and measured
in laboratory tests (Section 2) to estimate MDL, a more appropriate "total
operating precision" of an analyzer was obtained from the Phase II test
data. The total operating precision consists of the variance of two unpre-
dictable, random sources of measurement error—the method precision and
random bias. Random bias is caused by interfering substances that influence
the ability of an analyzer to measure SO2. The total operating precision
is the square root of the sum of these two variance components.

The estimation of these components is relatively straightforward when
there is a defined reference method measuring the "true" concentration or
when replicate measurements are obtained. If neither reference method nor
replicate measurements are available, total operating precision can be
estimated by using principal component estimates as described by Lawton
et al.f Using this procedure, estimates of total operating precision and
corresponding estimates of MDL for five of the analyzers tested were made
and are shown in Table 2.

It should be noted that these estimates of precision and MDL do not
include errors associated with other operators and laboratories. Thus,
these estimates are probably lower than would be determined if other vari-
ables were included.

^Analytical Chemistry, 52:2242, December 1980.

tLawton, W. H., E. A. Sylvestre, and B. J. Young-Ferraro, Statistical
comparison of multiple analytic procedures: Application of clinical chemis-
try, Technometrics, 2j_: 397-409, 1979.

12


-------
TABLE 2. ESTIMATES OF TOTAL OPERATING PRECISION AND MINIMUM

DETECTION LIMITS

Analyzer

Total operating
precision (ppm)

Minimum detection
limit (ppm)

Beckman 953

.0075

.0225

Meloy SA185-2A

.0032

.0096

Meloy SA285E

.0016

.0048

Monitor Labs 8450

.0040

.0120

Thermo Electron 43

.0059

.0177

13


-------
			____ _ _ 		ASARCO 500 Sulfur Dioxide

ASARCO 500

Figure 1. ASARCO Model 500 S02 Analyzer (EQSA-0877-024).
ASARCO Incorporated, 3422 South 700 West,

Salt Lake City, Utah 84119.

General Description

The ASARCO Model 500 S02 analyzer operates on the principle of con-
ductometric detection as described in Appendix A. It is a refined version
of tke original S02 monitor developed by Dr. Moyer D, Thomas in 1928, It
is equipped with an unheated chemical scrubber that removes all ambient
compounds other than SO2 that may cause conductivity changes in the absorb-
ing solution. Such compounds include carbon dioxide, nitrogen dioxide,
ammonia, and hydrogen chloride. Removal of these compounds makes the
instrument specific for SQg.

14


-------
ASARCO 500 Sulfur Dioxide

This analyzer, used only by ASARCO for monitoring at its own facil-
ities, is not in commercial production. For that reason, Phase II testing
is not planned for this analyzer.

TABLE 3. ASARCO 500 LABORATORY PERFORMANCE TEST RESULTS

Performance paramaters

Units

EPA

specs,8

i

Manufacturer'*
test results1*

EPA
last results6

Noise - 0% URL

ppm

0,005

<0,001

Not tested

Noise — 80% URL

PPm

0.005

0,002

Not tested

Lower detectable limit

ppm

0.01

0.006

0.008

Interferon ts

ppm







C02



±0,02

0.002

0.011

no2



+ 0.02

0.002

0.011

nh3



+ 0.02

0.001

0.000

HCI



±0.02

0.001

0.008

Total



<0,06

0.006d

0.030d

Zero drift - 12 h

ppm

±0.02

0,00td

0.001d

Zero drift — 24 h

ppm

±0.02

0.001d

0.001d

Span drift - 20% URL

%

±20.0

0.73d

2.6d

Span drift — 80% URL

%

±5.0

1,42d

2.8d

Lag time

min

20

0.5

Not tested

Rise time

miri

15

0.5

Not tested

Fall time

min

15

0.5

Not tested

Precision - 20% URL

PPm

0.010

0.001

0.005

Precision — 80% URL

ppm

0.015

0.006

0.012

aFrom EPA equivalency regulations, 40 CFR Part 53,

^Average, from manufacturer's application for equivalency determination.

cAverage, from EPA Phase I postdesignation tests.

^Average absolute values.

15


-------
ASARCO 500 Sulfur Dioxide

100

+*
C
(D
0

L

0)

a

2:
H
Cl

F

ui

QC

UJ
-J

cr
u
cn

0.0

. 1	.2	.3	.4

CONCENTRATION, ppm

ANALYZER!
HODEL:
SERIRL No.
POLLUTANT:
RANGEj
LOCATIONi
BATEs

DATE CODE;
NAME:

ASARCO
500
NONE
S02

.5 ppm
RTI

05^22/78

?0.142
PHASE I

Y - MX + R

M - 188.B4?
ssx c.i.- + 4.see

R - -.523
95* C.I.- + 1.265

r - .99963

DRTAs

X

Y

1

0.000

.750

2

.083

15.000

3

.039

18.200

4

. 150

27.000

5

. 225

41.600

6

.291

53.200

7

.375

70.200

8

.40!

75.300

S

.464

88.300

Figure 2. Typical calibration curve for the ASARCO Model 500.

16


-------
		Beckinan 953 Sulfur Dioxide

Beckman 953

ii k- 4- K-»vl fSrf

Figure 3. Beckman Model 953 Fluorescent Ambient SO2 Analyzer

(EQSA-0678-029). Beckman Instruments, Inc., Process
Instruments Division, 2500 Harbor Drive, Fullerton,

CA 92634.

General Description

The Beckman Model 953 sulfur dioxide analyzer operates on the prin-
ciple of fluorescence as described in Appendix A. It uses mechanically
chopped ultraviolet radiation to provide signal modulation for amplifica-
tion, The instrument is equipped with a heated "reactor" (scrubber) to
remove aromatic hydrocarbons that would otherwise produce a positive inter-
ference in the fluorescence technique. The heated reactor, however, will
convert ambient hydrogen sulfide (which does not interfere with the fluor-
escence technique) into Biilfur dioxide, causing an interference. To prevent
this interference, an unheated chemical scrubber for hydrogen sulfide is
placed upstream of the hydrocarbon reactor. Removal of moisture from the
sample is not required in this analyzer because this potential interference
is minimized by judicious selection of the wavelength of the excitation
radiation.

17


-------
Beckman 953 Sulfur Dioxide

To be used as an equivalent analyzer, the Model 953 must be operated
on the 0-to-0.5-ppni or O-to-l.O-ppm range, with a time constant setting of

2.0, 2,5, or 3.0 minutes and with a 5-to-10-pm membrane filter element
installed in the rear panel filter assembly. The analyzer can be operated
with or without any of the following options:

Remote operation kit (Catalog No. 641984)

Digital panel meter (Catalog No. 641710)

Rack mount kit (Catalog No. 641709)

Panel mount kit (Catalog No. 641708).

Ambient air surrounding the analyzer must be within the range of 20° to
30° C, inclusive, and line voltage to the analyzer must be within the range
of 105 to 125 Vac, inclusive.

18


-------
Beckman 953 Sulfur Dioxide

TABLE 4. BECKMAN 9S3 LABORATORY PERFORMANCE TEST RESULTS

Performance parameters

Units

EPA

specs.a

Manufacturer's
test results5

EPA
test results0

Noise - 0% URL

ppm

0.005

0,001

0.001

Noise - 80% URL

ppm

0.005

0.002

0.002

Lower detectable limits

ppm

0,01

0.010

0.009

Interferents

ppm







H2S



±0.02

0.001

0.001

Meta-xylene



±0.02

0.002

0.007

Naphthalene



+ 0,02

0.001

0.005

h2o



±0.02

0,006

-0.010

NO



±0.02

0.005

<0.001

no2



±0.02

0.004

-0.002

o3



±0.02

0.004

-0.014

Total



<0.06

0.023d

0.040d

Zero drift-12 h

ppm

±0.02

0.003d

0.003d

Zero drift—24 h

ppm

±0.02

0.00td

0.001d

Span drift-20% URL

%

±20.0

1,05d

3.23d

Span drift-80% URL

%

±5.0

0.58d

1.62d

Lag time

min

20

0.3

0.2

Rise time

min

15

6,2

2.3

Fall time

min

15

5.1

2,3

Precision — 20% URL

ppm

0.010

0.001

<0.001

Precision —80% URL

ppm

0.015

O.OQ2

0,001

"From EPA equivalency regulations, 40 CFR Part 53,
b Average, from manufacturer's application for equivalency determination.
cAverage, from EPA Phase I postdesignation tests.

^Average of absolute values.

19


-------
Beckman 953 Sulfur Dioxide

100

90

60 -

40 -

20 -

0
0. 0

.2	.3

CONCENTRRTION, ppm

ANALYZER:
MODEL:
SERIAL No.
POLLUTANT:
RANGE:
LOCATION:
DATE:

DATE CODE:
NAME:

BECKMAN
953

1000039
S02

. 5 ppm
DAMDF
03/30/?9

79.089
PHASE II

Y = MX + R
M - 194.000

SSH C.I.- + 2.459

fl «= 4.787
95* C.I.™ + .635

r *= . 39996

DATA:

X

V

1

0.000

5.200

2

.078

19.700

3

. 157

35.000

4

.260

55.000

5

.363

75.200

6

.412

85.000

Figure A. Typical calibration curve for the Beckman 953.

20


-------
BECKMRN 953

to

JU
Q.

a

g

S

u

N

C
®
o

*.
o
a

a

a

TOT. NET ZERO DRIFT, ppb -7.9
NUMBER OF DRIFT PERIODS! S3
RVE DRIFT PERIOD, days: 2.6
RVE |DRIFT|/PERIOD, ppb! 3.3
STD DEV. ZERO DRIFT, ppbi 4.S

-30

J	L	1	L

jOl



-L

U	L

_L



-I	L

J	L

J	1

30

45

75	90	105 128

DRY OF YEAR

135

150

165

180

TOTRL NET SPAN DRIFT, %s-55.3
NUMBER OF DRIFT PERIODS; 53
RVE DRIFT PERIOD, days: 2.6
AVE | DRIFT J /'PERIOD, X: 1.9
STD DEV. SPBN DRIFT, Zi 2.2

-15



j—i—1_

Xl

j	i—1_

JHL

_t_~I	1	L

J	L

J	0 I——i—i	I	i

30	45	60	75	90 105 120

DAY OF YEAR

135

150

165

180

Figure 5. Zero and span drift in the Bectaian 953 S02 readings during the Phase II

ambient monitoring test.


-------
Beekman 953 Sulfur Dioxide

.5

£
a
a.

ro
in
a?

U

LI

m

.4

3 -

. 1

0

0	,1	.2	.3	.4

RVERRGE OF REFERENCE RNRLY7ERS, ppm

.5

Figure 6. Relationship between the Beekman 953 and the average of the
other analyzers during the Phase II ambient monitoring test
(before HC reactor replacement--see General Comments),

22


-------
Beckman 953 Sulfur Dioxide

24
B
IS
91
90
20?
U) 307

y 2S,

~

ft 210

at

3
O

u
o

u.
o

a:
UJ

a

£
3

Z

130
72
43

30

29

31
27
21
26

7
1
0

+ 18
+17
+1S
+13
+11
+9
+7
+5
+3

y «¦

£

J2

a

a.

J

z

U

0£

u
u.
u.

t-i

CI

-1
-3

-s

-7

-11
-13
-15
-17
—IS

:i

TOTRL No. OF DRTR PRIRS:	1590

CORRELRTION COEFFICIENT:	.9S8131

MERN DIFFERENCE, ppbt	+5.16

STD. DEVIATION OF DIFFERENCES,	ppb: 6.5720

MRX. ABSOLUTE DIFFERENCE, ppb:	25

NO. ABSOLUTE DIFFERENCES >20:	21

Figure 7. Frequency distribution of differences in hourly S02 ambient
air readings between the Beckman 953 and the average of
simultaneous readings from the other analyzers during the
Phase II ambient monitoring test (before HC reactor rep]ace-
Eient--see General Comments). All readings corrected for
zero and span drift.

23


-------
Beckman 953 Sulfur Dioxide

(A
Ul

8
fc

a

TOTAL No. OF DRTfi PRIRS:

CORRELATION COEFFICIENT;

MEAN DIFFERENCE, ppb:

STD. DEVIRTION OF DIFFERENCES, ppb;

MRX. ABSOLUTE DIFFERENCE, ppb;

NO. RESOLUTE DIFFERENCES >20:

1590

.930363

+12.39
7.3404
33
170

Figure 8. Frequency distribution of differences in hourly S02 ambient
air readings between the Beckman 953 and the average of
simultaneous readings from the other analyzers during the
Phase II ambient monitoring test (before HC reactor replace-
ment—see General Comments). Beckman 953 readings are
not corrected for zero and span drift.

24


-------
Beckman yS3 Sulfur Dioxide

.5

.4

,2

. 1

0

0	.1	.2	.3	.4

RVERHGE: OF REFERENCE RNRLYZERS, ppm

Figure 9. Relationship between the Beckman 953 and the average of the
other analyzers during the Phase II ambient monitoring test
(after HC reactor replacement--see General Comments).

25


-------
Beckraan 953 Sulfur Dioxide

(0

Ul

U

o
o

TOTRL No. OF DHTB PAIRS;	1448

CORRELATION COEFFICIENT:	.396473

MERN DIFFERENCE, ppb:	+19.50

STD. DEVIRTION OF DIFFERENCES,	ppbs 5.5777

HfiX. RBSOLUTE DIFFERENCE, ppb:	30

NO. RBSOLUTE DIFFERENCES >20:	668

Figure 10, Frequency distribution of differences in hourly S02 ambient
air readings between the Beckman 953 and the average of
simultaneous readings from the other analyzers during the
Phase II ambient monitoring test (after HC reactor replace-
¦ent—see General Comments). All readings corrected for
zero and span drift.

26


-------
Beckraaii 953 Sulfur Dioxide

W
Ui

u

5

a:
an

u
o

u.

o

QL
Ui
CQ
T.

1181
01
37
52
33
27
11
4
0

0

1
8
i
e
e
e
a
0
0
0
0

JB

Q.
O.

8 ~!»

+19
+1?

+11
+13
+11
+8

+7
+5
+3

+ 1

-I
-3

-a

—7

-a

-u

-13

—15
-17
-18

TOTAL No. OF DRTft PHIRS:
CORRELATION COEFFICIENT;

MEAN DIFFERENCE, ppb:
STD. DEVIRTION OF DIFFERENCES,
HRX. RBSOLUTE DIFFERENCE, ppb:
NO. ABSOLUTE DIFFERENCES >20:

1448

.993255
+24. 13
ppb• S.5539

33
109?

Figure 11. Frequency distribution of differences in hourly S02 ambient
air readings between the Beckman 953 and the average cf
simultaneous readings front the other analyzers during the
Phase II ambient monitoring test (after HC reactor replace-
ment—see General Comments). Beckman 953 readings are
not corrected for zero and span drift.

27


-------
Beckraan 953 Sulfur Dioxide

Manufacturer's Recommended Maintenance

Replace sample filter.

Check span gas solenoid valve.

Refill or replace interferent reactor.

Replace selective scrubber cartridge.

Replace activated carbon scrubber.

Check and/or service pump.

Check or replace source lamp,

Operation

Frequency
Every 30 days
Every 30 days
Every 90 days
Every 24 mo
Every 24 mo
Every 24 no
Every 10,000 h
(~417 days of con-
tinuous operation)

Malfunctions

5-14-82

8-5-82
8-20-82

10-21-82

12-21-82

3-29-83*

4-14-83b

During Phase I postdesignation setup and calibration (see
General Comments for analyzer history), the analyzer's fuse
blew due to failure of the reaction chamber heater.
Replacement of the heater corrected the problem.

Phase I ozone interferent test failure (see General
Comments).

After being shut down for 2 weeks, the analyzer would not
respond to sulfur dioxide. The malfunction was caused by

failure of the demodulator board.

Phase I span drift test failure on day 3 of the test pro-
cedure (see General Comments),

Phase I hydrogen sulfide interferent test failure. The
malfunction was corrected by replacing the H2S scrubber
with a new scrubber (see General Comments) and repeating
the H2S interferent test procedure.

Phase I span drift test failure on day 1 of the test
procedure. Cause for malfunction not determined (see
General Comments).

Phase I span drift test failure on day 3 of the test
procedure. The malfunction was caused by failure of the
interferent reactor temperature control circuit (see
General Comments).

Modified 953, SN 1000039.
bNew 953, SN 1000865.

28


-------
Beckman 953 Sulfur Dioxide

General Comments

The Beckraan 953 SO2 analyzer was purchased in October 1978 and immedi-
ately included in the postdesignation Phase II testing program conducted
from November 27, 1978, to June 29, 1979.

The preventive maintenance section of the instrument manual for the
Beckman 953 states that the hydrocarbon reactor should be replaced routine-
ly every 3 months. In April 1979, during Phase II testing, the reactor was
replaced during routine, preventive maintenance, The sample inlet filter
was also replaced during the maintenance operation.

In addition to routine preventive maintenance, corrective maintenance
was performed, which included replacement and alignment of the lamp and
adjustment of the photoraultiplier tube (PMT) high voltage. This corrective
maintenance was suggested by Beckman personnel to sain additional range for
the span adjustment. The span setting had gradually increased from a pot
setting of 861 to 999 (fully clockwise) during the course of Phase II
testing. This fully clockwise setting was not sufficient to bring the span
response within ±7 percent of the nominal span value. To correct this
situation, the lamp was replaced and the PMT high voltage was adjusted from
-1080 volts to -1119 volts. After adjustment, the span pot setting necessary
to accomplish a nominal span response was 625.

A significant change in the instrument's response to ambient air
samples was noted after this maintenance. Before maintenance, the mean
difference and the standard deviation of the differences between the 953
responses and the average of the other analyzer responses had been +4.4 ppb
and 7,5 ppb, respectively (see Figures 6 through 8). After maintenance,
the mean difference jumped to +20.1 ppb, while the standard deviation
remained essentially unchanged at 7.1 ppb (see Figures 9 through 11).

However, responses to zero and span gases remained normal before and after
maintenance. These observations indicated that change in performance on
ambient air samples could not be attributed to the corrective maintenance
operations because these operations affect zero/span analyses and ambient
analyses equally. The performance change must therefore have resulted from
a change in the interaction between some compound in the ambient air sample
and the new reactor material. This change in interaction apparently produced
a positive response.

Because the change in performance was apparently caused by a positive
interference rather than an increased sensitivity for SOj, Phase-I-type in-
terference tests were conducted for two likely interferents: hydrogen sul-
fide and naphthalene. The resulting interference equivalents were +0,003
ppm for H2S and +0.001 ppm for naphthalene. Evidently, neither of these
compounds could have caused the problem.

At present, neither the cause of the suspected interference nor the
actual reason for the performance difference between the "old" and "new"
reactor is known. Additional testing to characterize this phenomenon is
planned during Phase I postdesignation testing of the instrument.

29


-------
Beckman 953 Sulfur Dioxide

The Phase II test program was completed on 6-29-79. At this time the
analyzer was removed from service.

In April 1980, prior to the initiation of Phase I testing, the analyzer
was sent to Beckman at Fullerton, California, for modification to current
production configuration. It was returned to IPA in May 1980.

On June 16, 1980, upon initial startup, the analyzer did not respond
to SO2. When the problem could not be resolved by the installation of new
high-voltage supply and amplifier boards (supplied by Beckman), a Beckman
service representative replaced the demodulator board (the board was loose
and the chopper blade was hitting the photosensor), adjusted the chopper
blade for proper clearance, checked the system for leaks, replaced the
amplifier board, and aligned the UV lamp. The analyzer remained operational
until September 25, 1980, when it again did not respond to SO?. Trouble-
shooting indicated a bad amplifier board. The analyzer was removed from
the test program.

After consultations with Beckman personnel concerning the poor opera-
tional history of this analyzer, the unit was returned to the factory in
the spring of 1982 for refurbishment under warranty. Corrective actions
performed during the refurbishment included replacing the pump and pump
shock mounts, the PNA converter, the H2S and zero air scrubbers, the PMT,
the PMT housing 0-ring, a socket assembly, strip marker-14, and the source
lamp. Replacements to update the analyzer to current manufacturing status
were also made during the refurbishment. They included addition of a
holder; replacement of the H2 lamp supply, high-voltage, filament control,
and interconnect boards; and replacement of the reaction chamber.

The analyzer was returned to EPA in April 1982. Beckman indicated
that, "Subsequent to the installation of these components, the analyzer was
tested in accordance with the PID test procedure 641968 for the Model 953.
The analyzer's performance was equal to or better than the specifications
required in the test procedure." Although originally purchased in October
1978, the analyzer, with the above stated refurbishment, could be considered
a new analyzer as of April 1982.

On August 5, 1982, the analyzer failed the ozone interferent test.
Additional o2one interferent tests conducted by Beckman on a current produc-
tion analyzer did not result in test failures but indicated somewhat larger
(but still within EPA specifications) ozone interference values than those
of the original test analyzer used for the original equivalency application
tests. The discrepancy was attributed to different PNA converter operational
temperatures. The temperature variance appears to be attributable to a
Beckman-instituted change in the supplier of the PNA converter heating
blanket used in the analyzer.

30


-------
Beckman 953 Sulfur Dioxide

On October 21, 1982, the analyzer failed a span drift test. Again
Beckman was contacted. Additional span drift tests conducted by Beckman on
a current production analyzer confirmed the span drift test failure. This
performance deficiency was attributed to inadequate reaction chamber and
air bath temperature control at high temperatures (>30° C) and was corrected
by modifying the analyzer back cover panel to allow the entrance of additional
ambient air to the analyzer.

On December 21, 1982, the analyzer failed the hydrogen sulfide inter-
ferent test. This performance deficiency was corrected by installing a new
selective scrubber. The cause of the selective scrubber failure has not
been determined.

On March 29, 1983, the modified analyzer (slotted back cover panel)
failed a span drift test. A Beckman service representative arrived the
week of April 4th to try to determine the cause of the malfunction and, if
possible, to repair the analyzer. After 3 days of unsuccessful attempts to
determine the cause of the malfunction, a new 953 analyzer (SN 1000865) was
sent from Beckman.

On April 14, 1983, the new 953 analyzer, with a modified back cover
panel, failed a span drift test. On April 20, 1983, a Beckman service
representative attributed the malfunction to a failure of the interferent
reactor temperature control circuit. Subsequent to replacement of the
interferent reactor temperature control board, the new 953 analyzer success-
fully passed a repeated span drift test.

Phase II Testing Operator's Comments

This analyzer required 20 to 40 minutes to achieve equilibrium on span
gases, a significant inconvenience during calibration.

During 150 days of Phase II testing, data were lost for 12 days due to

malfunctions.

The compartmentalized design of this analyzer was a convenient feature
during maintenance procedures.

Omission of Data from the Beckman 953 in Computation of Averages Used for
Phase II Comparisons

In the frequency distributions developed from Phase II test data, each
analyzer is compared to the average response from the other analyzers.

Ideally, this average is assumed to represent the "real" or theoretical
concentration of the sample. Nonrandom deviations (e.g., malfunctions or
consistent bias) in any analyzer included in this averaging process would
degrade the quality of the average and, thus, the entire comparison.

Although it is not possible to eliminate nonrandom deviations completely,
it is possible to eliminate or at least minimize obvious ones.

31


-------
Beckraan 953 Sulfur Dioxide

A significant change in this analyzer's performance during the Phase II
test was noted following routine, preventive maintenance procedures during
which the hydrocarbon reactor column was replaced. As stated earlier,
before-maintenance mean difference and standard deviation, of the difference
were +4,4 ppb and 7.5 ppb, respectively. Following replacement of the
reactor, the mean increased to +20,1 ppb while the standard deviation
remained essentially unchanged at 7.1 ppb. Responses to zero and span
gases were nominal both before and after the replacement. Because of this
change in performance, data from this analyzer were considered suspect and
nonrepresentative of the average performance observed during the test.

They have therefore been omitted from the computation of the "average"
data.

32


-------
Bendix 8303

Bendix 8303 Sulfur Dioxide

Figure 12. Bendix Model 8303 Sulfur Dioxide Analyzer (EQSA-1078-030) .

The Bendix Corporation Environmental and Process Instru-
ments Division, P.O. Box 831, Louisburg, WV 24901.

General Description

The Bendix 8303 sulfur dioxide analyzer operates on the principle of
flame photometric detection as described in Appendix A. It is equipped
with a heated silver scrubber to remove ambient hydrogen sulfide, rendering
the analyzer essentially specific for S02- Interference from other reduced
sulfur compounds (e.g., methyl mercaptan) is possible but of minor signifi-
cance at most monitoring sites. This scrubber is supplied in an independent
module separate from the analyzer itself.

Because a silver scrubber has a tendency to convert H2S to S02 in the
presence of ozone, the analyzer incorporates an "ozone decomposer" stage
prior to the silver scrubber. In this decomposer, ozone is removed from
the sample stream by reaction with ethylene in a heated chamber. The
ethylene is introduced into the sample stream by permeation from a small
pressurized tank within the scrubber assembly.

The 8303 also includes a background sulfur addition device. A con-
stant concentration of about 50 ppb of methyl mercaptan is added to the
flame by permeation via the hydrogen supply. The purpose of this sulfur
addition is to improve low-level sensitivity by improving the signal-to-
noise ratio at very low S02 concentrations. It also avoids the potential

33


-------
Bendix 8303 Sulfur Dioxide

problem of a hidden negative zero offset resulting from the inability of
the electronics to process negative signals caused by negative zero drift.

To be used as an equivalent analyzer, the Model 8303 must be operated
on a range of either 0 to 0.5 or 0 to 1.0 ppm, with a Teflon filter installed
on the sample inlet of the H2S scrubber assembly. Ambient air surrounding
the analyzer must be within the range of 20° to 30° C, inclusive,
and line voltage to the analyzer must be within the ran^e of 105 to 125 Vac
inclusive.

TABLE 5. BENDIX 8303 LABORATORY PERFORMANCE TEST RESULTS





EPA

Manufacturer's

EPA

Performance parameters

Units

specs.3

test results'1

test results0

Noise — 0% URL

ppm

0.005

0.001



Noise - 80% URL

ppm

0.005

<0.001



Lower detectable limit

ppm

0.01

0.009



Interferents

ppm







C02



±0.02

-0.002



CO



+ 0.02

- 0.001

TESTING

o3



±0.02

0.002

NOT YET

h2s



±0.02

0.002

PERFORMED

h2o



±0.02

- 0.003



Total



< 0,06

0.010d



Zero drift — 12 h

ppm

±0.02

0.001d



Zero drift — 24 h

ppm

±0.02

0.001d



Span drift -20% URL

%

+ 20.0

1.73d



Span drift - 80% URL

%

+ 5.0

0.71d



Lag time

min

20

0,3



Rise time

min

15

0.3



Fall time

min

15

0.3



Precision - 20% URL

ppm

0.010

<0.001



Precision — 80% URL

ppm

0.015

0.001



aFrom EPA equivalency regulations, 40 CFR Part 53.

^Average, from manufacturer's application for equivalency determination.

cAverage, from EPA Phase I postdesignation tests.

^Average absolute values.

34


-------
Bendix 8303 Sulfur Dioxide

100

+>
C
CD
O

u.

m
Q.

(J
2

M

P

a;

UJ

oi
u

tr

o

U)

0.0

1	.2	.3	.4

CONCENTRATION, ppm

ANALYZER:

BENDIX

y «

MX + R

DATA:

__X	

	Y	

MODEL:

8303











SERIAL No.

: 11252

M -

199.721

1

0.000

5.300

POLLUTANT:

S02

2

.078

20.300

RANGE:

.5 ppm

95 %

C.I.« + 2.311

3

. 157

36.100

LOCATION:

DAHDF

R «=

4.926

4

.260

56.600

DATE:

03/30/79

5

.363

77.400

DATE CODE:

79.089

to
tn

C.I.- + ,596

8

.412

87.500

NAME:

PHASE II

r =

.99997



. . •



Figure 13. Typical calibration curve for the Bendix 8303.

35


-------
u>

JQ

a
a

H
U.

M

ft:
Q

o
a:
u

N

TOT. NET ZERO DRIFT, ppb+22.
NUMBER OF DRIFT PERIODS: 32
RVE DRIFT PERIOD, days:
RVE |DRIFT|/PERIOD, ppbs
STD DEV. ZERO DRIFT, ppbs

90 105 120
DRY OF YERR

180

c
e
c

c
o
a.

t—

u.

M

X

a

z

-------
Bendix 8303 Sulfur Dioxide

.5

.4

£

CL

CL

- .3

-------
Bendix 8303 Sulfur Dioxide

w
u
u
z
u
a.
a.

o
u
o

Lu
O

0

a
b
0
0

7
33
54
93
214
452
222

S 215

m
r

Z)

279
254
92
23
23
4
3
1

XI
Q.
Q.

+ 19
+ 17
+ 15
+ 13 +
+ 1 1
+9
+7
+5
+3
+ 1
-1
-3
-5
-7
-9
-1 )
-13
-15
-17

¦'i

TOTAL No. OF DRTR PRIRS:

CORRELATION COEFFICIENT:

MERN DIFFERENCE, ppb:

STD. DEVIATION OF DIFFERENCES, ppb

MAX. ABSOLUTE DIFFERENCE, ppb:

NO. ABSOLUTE DIFFERENCES >20:

1969

. 997375
-2.G1

4 .6767
2 1
1

Figure 16. Frequency distribution of differences in hourly S02 ambient
air readings between the Bendix 8303 and the average of
simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. All readings corrected
for zero and span drift.

38


-------
Bendix 8303 Sulfur Dioxide

U)
Ui

m
x.

z

0

23
12

24
89
78

352

ui

^ 267
U

a 163
a:

3 2?3

o

O 310
& 234
K 1 14

U

23
4
0
0
2
0
0
0

JO

a.
a.

+19
+17
+15
+ 13
+ 11
+9
+7
+5
+3
+ 1
-1
-3
-5

-9
-11 --
-13 - -
-15 - -
-17 --
-19 --

:

?

TOTRL No. OF DRTR PRIRS:
CORRELATION COEFFICIENT:

MERN DIFFERENCE, ppb:
STD. DEVIATION OF DIFFERENCES, ppb:
MRX. ABSOLUTE DIFFERENCE, ppb:
NO. ABSOLUTE DIFFERENCES >i?0:

1969

.997076
+3.70

4.9744
18
0

Figure 17. Frequency distribution of differences in hourly S0£ ambient
air readings between the Bendix 8303 and the average of
simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. Bendix 8303 readings are
not corrected for zero and span drift.

39


-------
Bendix 8303 Sulfur Dioxide

Manufacturer's Recommended Maintenance

Frequency

Replace sample filter.

Inspect dilution air filter.

Service vacuum pump.

Replace Hz filter.

Clean instrument.

Leak check.

Flow check.

Check sulfur addition device.

Check scrubber permeation tank pressure.

Monthly
Monthly

When vacuum is <23 in. Hg

As required

As required

None specified

None specified

None specified

None specified

Malfunctions

Factory checkout data could not be duplicated during the initial setup
tor Phase II.

General Coisr,ents

During initial startup of this analyzer on 2-14-79, the factory check-
out data (flow rates, pressures, linearity, span, and zero) could not be
duplicated. Bendix service personnel replaced the dilution air capillary
and repeated the entire factory checkout procedure onsite. After the
checkout, the analyzer was in operating condition.

During 130 days of Phase II testing, data were lost for 0 days.

Currently, Phase I testing has not been performed.

This instrument achieved equilibrium on span gases in 5 to 7 min.

The instrument manual maintenance instructions seemed incomplete in
that many of the maintenance operations were described in detail, but no
mention was made of when or under what conditions the operation was to be
performed.

40


-------
	Lear Siegler SM100Q Sulfur Dioxide

Lear Siegler SM1000

Figure 18, Lear Siegler SMI000 SO2 Ambient Monitor (EQSA-1275-005).

Lear Siegler, Inc., Environmental Technology Division,
74 Inverness Drive East, Englewood, CO 80112.

General Description

The Lear Siegler Model SM1000 sulfur dioxide analyzer is no longer in
production, having been replaced by the Model AM2020. Some Model SM1000
analyzers are still in use, however, and are covered by this equivalent
method designation. The Model SM1000 operates on the principle of second
derivative spectroscopy as described in Appendix A. Use of the second
derivative technique, along with judicious selection of the analytical
wavelength, makes the analyzer specific for SO2 without the need for chemical
scrubbers or other conditioning. The analyzer is packaged in two separate
modules, one containing the electronics and the other containing the optics
and pneumatics. An optional outdoor enclosure is available for the analyzer,
permitting it to be used over an extended temperature range of -34° to
38° C (-30° to 100° F).

To be used as an equivalent analyzer, the Model SM1000 must be oper-
ated on the 0-to-0.5-ppm range, at a wavelength of 299.5 nm, with the
"slow" (300-second) response time, and with or without any of the following
options:

41


-------
Lear Siegler SM10Q0 Sulfur Dioxide

SM-1	Internal zero/span

SM-2	Span timer card

SM-3	0- to 0.1-volt output

SM-4	0- to 5-volt output

SM-5	Alternate sample pump

SM-6	Outdoor enclosure.

Line voltage to	the analyzer must be within the range of 105 to 125 Vac,
inclusive,

TABLE 6. LEAR SIEGLER SM1000 LABORATORY PERFORMANCE

TEST RESULTS

EPA	Manufacturer's	EPA

Performance parameters Units	specs,8	test results'*	test results®

Noise - 0% URL

ppm

0.005

0.003

0.002

Noise - 80% URL

ppm

0.005

0.003

0.002

Lower detectable limits

ppm

0.01

0.013

0.010

Interferents

ppm







Meta-xylene



±0.02

-0.002

-0,001

no2



±0.02

-0.001

-0.004

NO



±0.02

0.001

0.001

°3



±0,02

0.002

-0.003

Total



<0.06

0.006d

0.009d

Zero drift —12 h

ppm

±0.02

0.018d

0.014d

Zero drift —24 h

ppm

±0.02

0.004d

0.010d

Span drift -20% URL

%

±20,0

3.98d

5.50d

Span drift — 80% URL

%

±5.0

0.86d

2.18d

Lag time

min

20

0.9

0.3

Rise time

min

15

8.4

7.5

Fall time

min

15

8.0

7.0

Precision —20% URL

ppm

0.010

0.002

0,002

Precision — 80% URL

ppm

0.015

0.002

0.002

aFrom EPA equivalency regulations, 40 CFR Part 53.
b Average, from manufacturer's application for equivalency determination,
cAverage, from EPA Phase I postdesignation tests.

^Average of absolute values.

42


-------
Lear Siegler SM1000 Sulfur Dioxide

CONCENTRRTION, ppm

RNRLYZER:
MODEL:
SERIRL No.
POLLUTRNT:
RRNGE:
LOCRTION:
DRTE:

DRTE CODE:
NRME:

LERR SIEGLER
SM 1000
7564
SOS

. 5 ppm
DRMDF
03/30/79

79.089
PHRSE II

Y - MX + R
M - 200.722

95% C.I.- + 3.597

R = 5.931

95*4 C.I.- + .92S

r - .99992

DRTRi

1

0.000

5.500

2

.070

22.100

3

. 157

37.200

4

.260

58.500

5

.363

79.000

6

.412

66.200

Figure 19. Typical calibration curve for the Lear Siegler SM1000.

43


-------
¦c-
¦c-

+»
c

4)

O

t-

o
a

h-
Lu

M

QC

a
z

CE

a.
cn

TL NET ZERO DRIFT, ppb:+132.5
NUMBERiOF DRIFT PERIODS: 29

S3

a
a

h-
Lu

M

a:

p

o
a:
u

N

-30

30

+ 15 r

+ 10 -

+5 -

+0 -

-5 -

-10 -

-15
30

j	i	i	i	i	i	i	i	i	i	i	i0 i	i	i Oi	i	v	£ i	i	&.

45	60	75	90	105 120 135 150

DRY OF YEAR

J	l	(Xti	1

165	180

TOTAL NET SPAN DRIFT, X:-27.5
NUMBER OF DRIFT PERIODS: 29
AVE DRIFT PERIOD, days: 2.7
RVE |DRIFT|/PERIOD, Si: 4.S
STD/OEV. SPRN DRIFT, J4: 5.3

J	L

J	I	I	I	l	l	I	L

J	1	L

j—i—i—iw i—iQ iQ i	i		iO i Q	i

45

60

75

90	105	120

DAY OF YEAR

135

150

165

180

Figure 20. Zero and span drift in the Lear Siegler SM1000 S02 readings during the

Phase II ambient monitoring test.


-------
Lear Siegler SM1000 Sulfur Dioxide

.5

E

a.

o.

<3
<3
<3

9 ' I

I

2;

U)

QL
U
_J
(1J

u

w

cn

QL
CE
Id

.4 -

.3 -

.2 -

. 1 -

0

0

.1	.2	.3	.4

RVERRGE OF REFERENCE RNRLYZERS, ppm

Figure 21. Relationship between the Lear Siegler SM1000 and the average
of the other analyzers during the Phase II ambient monitoring test.

45


-------
Lear Siegler SM1000 Sulfur Dioxide

U

u
z
u
a
a

3
U

o
o

o

IX.
u
m

TC

3
Z

304



+19

70



+17

(16



+15

91



+13

105



+11

106



+9

104



+7

30





JU

n

+5

a?

a.

+3

104

U
o

+1

93

z





u

-1

70

OL
LI



U.

-3

64

Li_



65

H

a

-5
-7

51



-e

26



-li

31



-13

30



-15

33



-17

31



-19

124





TOTRL No. OF DATA PAIRS:	1797

CORRELATION COEFFICIENT:	.336291

MEAN DIFFERENCE, ppb:	+3.74

STD. DEVIATION OF DIFFERENCES,	ppb: 20.7309

MAX. ABSOLUTE DIFFERENCE, ppb:	100

NO. ABSOLUTE DIFFERENCES >20:	390

Figure 22. Frequency distribution of differences in hourly S02 ambient
air readings between the Lear Siegler SM1000 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. All readings corrected
for zero and span drift.

46


-------
Lear Siegler SM1Q00 Sulfur Dioxide

CO

UI

u

•mm

y

Q£

~
U
U
o

u.
o

u
n
z:

3

289
61
66
127
11?
95
90

74
84
84
04
94

75
57
56
SB
50
46

34

35
121

+ 19
+ 17
+15
+ 13
+ 11
+9
+7
+5
+3

UJ .,
(j +1

A

a.

a.

z
UJ
K
UI

u.
u.

-1
-3
-5
-7
-9
-11
-13
-IS
-17
-19

TOTRL No. OF DATA PAIRS:
CORRELATION COEFFICIENT;

MEAN DIFFERENCE, ppb:
STD. DEVIATION OF DIFFERENCES,
MAX. ABSOLUTE DIFFERENCE, ppb:
NO. ABSOLUTE DIFFERENCES >20:

1797

.924336
+3.25
ppb: 21.0543
100
377

Figure 23. Frequency distribution of differences in hourly S02 ambient
air readings between the Lear Siegler SM1000 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. Lear Siegler SM1000
readings are not corrected for zero and span drift.

47


-------
Lear Siegler SM1Q0Q Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation	Frequency

Replace ultraviolet (UV) lamp.	Every 60 days

Clean sample cell mirrors.	Every 60 days

Replace zero air filters.	Every 60 days

Malfunctions

1-2-76	to 4-5-76	Excessive signal noise and drift.

4-5-76	Defective wobbler coil (returned to factory).

6-15-76 to 8-3-76	Signal noise and drift less severe but not acceptable.

9-21-76	Mirror alignment incorrect.

9-23-76 to 10-14-76	Excessive signal noise.

2-10-77	to 3-18-77	Excessive signal noise.

3-25-77	Analyzer lamp replaced.

11-1-78	Excessive signal noise--UV-tuned mirrors Installed.

11-8-78	Analyzer lamp replaced.

11-20-78	Insufficient span pot.

11-27-78	Electronics corrected—span pot acceptable.

12-12-78	Excessive signal noise--power supply card replaced.

1-10-79	"Test" button light failed.

2-2-79	Negative response noted—analyzer taken off-line.

Factory service requested,

2-7-79	Factory service personnel determined that the

power supply was defective. New card ordered
from factory.

3-27-79	New power supply card arrived and was installed—

analyzer back on-line.

4-1-79	to 4-29-79	Signal noise steadily increased.

48


-------
Lear Siegler SM1000 Sulfur Dioxide

4-30-79	Analyzer optics cleaned and aligned. New lamp

installed—signal noise substantially reduced.

5-23-79	Fro*, t panel fuse blew for no apparent reason.

General Comments

The instrument was set up initially in January 1976. Excessive drift
and signal noise were unacceptable and persisted despite efforts to correct
the problem, including mirror realignments, voltage adjustments, and routine
zero and span checks. The instrument was returned to the factory for
maintenance on 4-6-76. On 6-15-76, the instrument was received from the
manufacturer and placed on-line. Drift continued to be a problem. The in-
strument was powered down on 8-3-76.

Phase I testing was to have begun on 9-21-76. However, it was delayed
for 5 mo until excessive signal noise problems were corrected.

Between 5-3-77 and 11-3-78, the instrument was returned to the manu-
facturer for maintenance, repair, and application of the designation sticker.

On 11-3-78, in preparation for Phase II testing, new UV-tuned mirrors
and a new lamp were installed.

Phase II testing was delayed for 18 days because of excessive signal
noise.

On 4-30-79, corrective maintenance was performed in an attempt to
reduce noise in the analyzer's response. The maintenance included:

Cleaning the sample cell mirrors
Cleaning the lamp lens
Changing the UV lamp
Aligning the sample cell mirrors
Adjusting the analytical wavelength
Peaking the optics.

Although maintenance required 6 h, the noise was reduced substantially.

During 194 days of Phase II testing, data were lost for 64 days be-
cause of malfunctions.

Phase II Testing Operator's Comment

This instrument repeatedly experienced malfunctions and required
repair or adjustments. The instrument signal was characteristically noisy
throughout the testing. At one point during the test, the anlyzer was
off-line for 48 days awaiting arrival of a power supply board from the

49


-------
Lear Siegler SM1000 Sulfur Dioxide

factory. The instrument manual was outdated. Several times, factory
representatives either performed or suggested procedures other than those
listed in the manual. The alignment of the mirrors was a difficult and
time-consuming process. The recommended alignment procedure had to be
performed in a dark room even though the instrument manual did not specify
this condition. Any movement or vibration of the analyzer had the potential
for upsetting the alignment of the optics.

The analyzer has a sample demand flow rate of 4.3 stdL/min. This
demand rate is greater than the capability of many dynamic calibration
systems at span concentrations of approximately 0.45 ppm.

Omission of Data from the Lear Siegler SM1000 in Computation of Averages
Used for Phase II Comparisons

In the frequency distributions, each analyzer is compared to the
average response from the other analyzers. Ideally, this average is assumed
to represent the "real" or theoretical concentration of the sample. Non-
random deviations (e.g., malfunctions or consistent bias) in any analyzer
included in this averaging process would degrade the quality of the average
and thus the entire comparison. Although it is not possible to eliminate
nonrandon deviations completely, it is possible to eliminate or at least
minimize obvious ones.

Throughout the Phase II test, the Lear Siegler SM1000 exhibited a
higher standard deviation (s = 20.8 ppb) and a lower correlation coefficient
(r = 0.934) than any of the other analyzers. This situation probably
reflects the large amount of noise and variability constantly present in
the analyzer's response. In addition, the analyzer experienced several
malfunctions and required extensive repair and adjustment. Therefore, data
from this instrument were considered nonrepresentative of the average
concentrations observed during the test and have been omitted from computa-
tion of the "average"' data.

50


-------
Lear Siegler AM2020

Lear Siegler AM2020 Sulfur Dioxide

Figure 24. Lear Siegler Model AM2020 Ambient S02 Monitor
(EQSA-1280-049). Lear Siegler, Inc., Environ-
mental Technology Division, 74 Inverness Drive
East, Englewood, CO 80112.

General Description

The Lear Siegler Model AM2020 sulfur dioxide analyzer operates on the
principle of second derivative spectroscopy as described in Appendix A.
Use of the second derivative technique, along with judicious selection of
the analytical wavelength, makes the analyzer specific for S02 without the
need for chemical scrubbers or other conditioning. The AM2020 differs from
the SMI000 in that the modulation of the analytical wavelength required to
measure the second derivative signal directly is accomplished by a continu-
ously rotating "scanner" rather than a reciprocating mechanical slit.

51


-------
Lear Siegler AM2020 Sulfur Dioxide

Additionally, the AM2020 contains an on-board microprocessor that monitors
and compensates for a variety of instrument condi tions and performs diagnos-
tic and self-testing operations.

To be used as an equivalent analyzer, the Model AM2020 must be operated
on the O-to-O.5- or the 0-to-l.Q ppm range, at a wavelength of 299.5 nm,
with a 5-min integration period. Ambient air surrounding the analyzer must
be within any 10° C temperature range between 20° and 45° C, inclusive.

Line voltage to the analyzer must be within the range of 105 to 125 Vac,
inclusive. The equivalent analyzer may be operated with or without the
automatic zero and span correction feature.

TABLE 7. LEAR SIEGLER AM2020 LABORATORY PERFORMANCE

TEST RESULTS





EPA

Manufacturer's

EPA.

Performance parameters

Units

specs.®

test resultsh

test results0

Noise - 0% URL

ppm

0.005

0.001



Noise - 80% URL

ppm

0.005

0.002



Lower detectable limits

ppm

0.01

0.009



Interferents

ppm





!

Meta-xylene



±0.02

0.001



imo2



±0.02

-0.004

TESTING

NO



±0.02

< -0.001

NOT YET





±0.02

-0.006

PERFORMED

Total



<0.06

o
o



Zero drift—12 h

ppm

±0.02

0.007d



Zero drift — 24 h

ppm

±0.02

0.004d



Span drift—20% URL

%

±20,0

4.77d



Span drift —80% URL

%

±5.0

1.90d



Lag time

min

20

0.5



Rise time

min

15

5.0



Fall time

min

15

5.1



Precision — 20% URL

ppm

0.010

0.001



Precision — 80% URL

ppm

0.015

0.003



aFrom EPA equivalency regulations, 40 CFR Part 53.

Average, from manufacturer's application for equivalency determination.
0Average, from EPA Phase I postdesignation tests.

^Average of absolute values.

52


-------
Meloy SA185-2A

Meloy SA185-2A Sulfur Dioxide

Figure 25. Meloy Model SA135-2A Sulfur Dioxide Analyzer

(EQSA-1275-0G6). Columbia Scientific Industries,
11950 Jollyville Road, Austin, TX 78759.

General Description

The Meloy SA185-2A operates on the principle of flame photometric
detection as described in Appendix A. It is equipped with a chemical
scrubber to remove ambient hydrogen sulfide, making it practically specific
for SOg. Interference from other reduced sulfur compounds (e.g., methyl
mercaptan) is possible but is usually of only minor significance at most
ambient monitoring sites.

The standard output is exponential, following the characteristic of the
flame photometric detector, but the analyzer may be equipped with a linear
output option. This linear option, however, uses a logarithmic amplifier
that by nature will not process negative signals. Therefore, any negative
zero drift is truncated by the instrument output circuits, resulting in a
zero output and a possibly nonapparent negative zero offset. Careful
adjustment of the zero control is necessary to avoid such a negative zero
offset.

To be used as an equivalent analyzer, the Model SA185-2A must be
operated on the 0-to-0.5-ppm range, with or without any of the following
options:

53


-------
Meloy SA185-2A Sulfur Dioxide

S-l	Linearized output

S-2	Modified recorder output

S-5	Teflon-coated block

S-6A	Re ignite timer circuit

S-7	Press to read

S-11A	Manual zero and span

S-11B	Automatic zero arid span

S-13	Status lights

S-14	Output booster amplifier

S-148	Line transmitter board

S-18	Rack mount conversion

S-18A	Rack mount conversion

S-21	Front panel digital voltmeter

S-22	Remote zero/span control and status (timer)

S-22A	Remote zero/span control

S-23	Automatic zero adjust

S-23A	Automatic/manual zero adjust

S-24	Dual range linearized output

S-33	Remote range control and status (signals)

S-34	Remote control

S-35	Fro..c panel digital meter with BCD output

S-36	Dual range log-linear output

S-38	Sampling mode status.

The analyzer may also be operated on the 0-to-l.0-ppm range; however, it
must be equipped with option S-36 or options S-l and S-24, with or without
any of the other options. Ambient air surrounding the analyzer must be
within the range of 20° to 30° C, inclusive, and line voltage must be
within the range of 105 to 125 Vac, inclusive.

54


-------
Meloy SA185-2A Sulfur Dioxide

TABLE 8. MELOY SA185-2A LABORATORY PERFORMANCE TEST RESULTS

Manufacturer's	EPA

test results'1	test results0

Performance parameters

Units

specs.8

Log

Linear

Log

Linear

Noise-0% URL

ppm

0.005

<0.001

<0.001

<0.001

<0.001

Noise-80% URL

ppm

0.005

0.001

0.001

0.001

0.001

Lower detectable limit

ppm

0.01

0.007

0.007

0.005

0.005

Intsrferents

ppm











o3



±0.02

-0.001

- 0.002

not tested

not tested

CO



±0.02

-0.001

-0.001

- 0.003

- 0.003

h2o



+0.02

- 0.004

-0.003

-0.001

- 0.002

co2



±0.02

-0.010

-0.010

- 0.025d

- 0.026d

h2s



±0.02

0.003

0.004

0.001

< 0.001

Total



<0.06

0.019e

0,Q20e

0.030e

0.031e

Zero drift - 12 h

ppm

±0.02

0.001e

0.001e

< 0.001e

0.001e

Zero drift ~ 24 h

ppm

±0.02

0.001e

<0.001e

0.002e

0.003e

Span drift - 20% URL

%

±20.0

2.84e

3.23e

0.98e

0.49e

Span drift-80% URL

%

±5.0

2,16e

2,99e

0,65e

1.05e

Lag time

min

20

0.1

0.1

not tested

0.1

Rise time

min

15

1.1

1.1

not tested

0.7

Fall time

min

15

0.5

0.5

not tested

0.6

Precision - 20% URL

ppm

0.010

0.001

<0.001

< 0.001

0.001

Precision — 80% URL

ppm

0.015

0.002

0.002

<0.001

0.002

aFrom EPA equivalency regulations, 40 CFR Part 53.
b Average, from manufacturer's application for equivalency determination.
cAverage, from EPA Phase I postdesignation tests, Serial Number 4H024.
dSee general comments.
eAverage of absolute values.

55


-------
Meloy SA185-2A Sulfur Dioxide

100

•P
C
0)

o

i.

m

CL

U?
Z

M

CJ

cc

LI

&

UJ
_j

cr
u

0.0

.1	.2	.3	.4

CONCENTRRTION, ppm

RNRLYZER:

MODEL:

SERIRL No.

POLLUTRNT:

RANGE:

LOCRTION:

DRTE:

DRTE CODE:
NRME:

MELOY
SR 105
: 4H024
S02

. 5 ppm
DRMDF
03/30/79

79.089
PHRSE II

Y - MX + fl
M «= 199.814

95* C.I.- + 3.685

R = 5.456
95* C.I.- + .951

r - .99991

DRTR:

X

Y

1

0.000

6. 100

2

.078

20 ,,500

3

. 157

36„500

4

.260

57.300

5

.363

78.300

6

.412

87.800

Figure 26. Typical calibration curve for the Meloy SA185-2A.

56


-------
TOT. NET ZERO DRIFT, ppb+15.9

DRY OF YEFIR

DRY OF YERR

Figure 27. Zero and span drift in the Meloy SA185-2A S02 readings during the

Phase II ambient monitoring test.


-------
MeJLoy SA185-2A Sulfur Dioxide

.4

£
a

GL

CE 3
OJ *3

i

in

ao

wHI

-
o

-J
id

z

i -

0

0

¦ 1	• 2	. 3	i 4

RVERRGE OF REFERENCE RNflLYZERS, ppm

Figure 28. Relationship between the Meloy SA185-2A and the average
of the other analyzers during the Phase II ambient
monitoring test.

58


-------
Meloy SA185-2A Sulfur Dioxide

0
0
0

0

1
3

e

31

m

X.
ZJ

z



in
u
u

2
LJ

a. 107
(K

ZJ

u
u
o

161
404
706

ja
a
a.

;1

CK 652

U

631
290
151
69
45
17
9
1

+ 19
+ 17 +
+ 15
+ 13
+ 11
+9
+7
+5
+3
+ 1
-1
-3
-5
-7
-9
-11
-13
-15

'"J

-19 -1

TOTRL No. OF DRTfl PRIRS:

CORRELRTION COEFFICIENT:

MERN DIFFERENCE, ppb:

STD. DEVIATION OF DIFFERENCES, ppb:

MRX. RBSOLUTE DIFFERENCE, ppb:

NO. RBSOLUTE DIFFERENCES >20:

3285

.998796
-3.8 1

3.9524
20
0

Figure 29. Frequency distribution of differences in hourly S02 ambient
air readings between the Meloy SA185-2A and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. All readings corrected
for zero and span drift.

59


-------
Meloy SA185-2A Sulfur Dioxide



20











+ 19



26











+ 17



22











+ 15



28











+ 13



37











+11



71











+9

0)

104





y





+7

o

323





z
in

JO
Q

+5

t*

732

a.



oc





+3

d

520

iii

+ 1

o



<	)

o

428

2







U

-1



,475

ffi





u.

-3

K

229

u.



bJ

m
r

101

M

a

-S

Z3





-7



59











-9



49











-11



27











-13



12











-15



B











-17



3











-19



10





1

TOTAL No. OF DATA PRIRS:

CORRELATION COEFFICIENT:

MEAN DIFFERENCE, ppb:

STD. DEVIATION OF DIFFERENCESf ppb;
MAX. ABSOLUTE DIFFERENCE, ppb:
NO. ABSOLUTE DIFFERENCES >20:

3285

.937855
+ 1 .58

5.4001
28
23

Figure 30. Frequency distribution of differences in hourly S02 ambient
air readings between the Meloy SA185-2A and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. Meloy SA185-2A readings
are not corrected for zero and span drift.

60




-------
Meloy SA185-2A Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation

Frequency

Replace charcoal filter.

Every 6 mo

Replace dilution air orifice.

Every 6 mo

Replace Teflon lines.

Annually

Replace flow restrictor capillary.

Annually

Clean rotameters.

Annually

Clean burner chamber.

Annually

Replace optical window.

Annually

Clean fan filter.

Annually

Replace H2S scrubber cartridge.

Annually

Replace II2 regulator.

Every 24 mo

Replace solenoid valves.

Every 24 mo

Replace vacuum pump.

Every 24 mo

Malfunctions



11-12-76	Phase I C02 interferent

test failed (see general
comments).

4-4-79	Pump diaphragm was

replaced.

General Comments

On 11-12-76, during Phase I testing, the analyzer failed the C02
interferent test with an average interference equivalent of -0.026 ppm.
Hydrogen flow was adjusted per conversations with a Meloy Laboratories
representative. A second failure prompted Meloy to send a representative
who adjusted the hydrogen flow and the oxygen content in the test atmos-
phere. The analyzer was never formally retested; however, after the adjust-
ments, the analyzer did pass an informal test with an interference equivalent
of approximately -18 to -19 ppb (specification: ±20 ppb). At this point,
MSB personnel questioned whether or not this performance was representative
of all Model SA185-2A analyzers, and Meloy representatives agreed to test
10 randomly chosen production units. All 10 of the units passed the tests
with interference equivalents ranging from 10 to 18 ppb and averaging
14 ppb.

During Phase II testing, a special zero adjustment procedure was
adopted to avoid the nonapparent negative zero drift possible with this
analyzer. The zero setting was adjusted whenever the analyzer's front
panel meter indicated a reading of less than zero while the analyzer was
sampling zero gas. This adjustment procedure superseded the ±3 percent
criterion applied to zero adjustments of the other analyzers in the test.

61


-------
Meloy SA185-2A Sulfur Dioxide

During Phase II testing, a misalignment in the analyzer's electronics
caused the recorder to read 6.0 percent of chart when the analyzer front
panel meter read zero. This occurred even though the recorder was zeroed
at 5.0 percent of chart. The analyzer is equipped with circuits that allow
this condition to be corrected. However, the condition was dealt with
during the Phase II tests in the data reduction routines rather than by
analyzer adjustment. Nevertheless, this misalignment was responsible for
the consistent +5.0-ppb offset noted in the zero drift chart for this
analyzer (Figure 27). The analysis of zero drift assumes a nominal zero of
5.0 percent of chart. A chart reading of 6.0 percent corresponds to +5.0 ppb.

During 215 days of Phase II testing, data were lost for 0 days because
of malfunctions.

Phase II Testing Operator Comments

Signal noise levels at span concentrations were higher than the average
noise experienced from the other analyzers during Phase II testing.

Engaging the locking mechanisms on the zero and span controls caused
the zero/span settings to change, which was annoying at times.

Unlike most analyzers, the zero and span controls do not have graduated
dials, which prohibited recording the control settings.

62


-------
Meloy SA285E

Meloy SA285E Sulfur Dioxide

Figure 31. Meloy Model SA285E Sulfur Dioxide Analyzer
(F.QSA- 1078-032). Columbia Scientific
Industries, 11950 Jollyville Road, Austin,
TX 78759.

General Description

The Meloy SA285E, a newer, improved version of the Model SA185-2A,
operates on the flame photometric detection principle as described in
Appendix A. It is equipped with a chemical scrubber to remove ambient
hydrogen sulfide, making the analyzer essentially specific for S02. Inter-
ference from other reduced sulfur compounds such as methyl mercaptan is
possible but of minor significance at most ambient monitoring sites.

Linear output is standard on the SA285E, and it does not have the potential
for nonapparent negative zero offset mentioned for the SA185-2A.

To be used as an equivalent analyzer, the Model SA285E must be operated
on the following ranges and time constant switch positions:

63


-------
Meloy SA285E Sulfur Dioxide

Range, ppb

0-50*
0-100*
0-500
0-1,000

Time constant setting

1 or 10
1 or 10
Off, 1, or 10
Off, 1, or 10

The analyzer	may be operated with or without any of the following

options:

S-5	Teflon-coated block

S-14B	Line transmitter board

S-18	Rack mount conversion

S-18A	Rack mount conversion

S-21	Front panel digital meter

S-22	Remote zero/span control and status (timer)

S-22A	Remote zero/span control

S-22B	Remote zero/span control and status (pulse)

S-23	Auto zero adjust

S-23A	Auto/manual zero adjust

S-25	Press to read

S-26	Manual zero and span

S-27	Auto manual zero/span

S-28	Auto range and status

S-30	Auto reignite

S-32	Remote range control and status

S-35	Front panel digital meter with BCD output

S-37	Temperature status lights

S-38	Sample mode status.

Ambient air surrounding the analyzer must be within the range of 10° to

40° C, inclusive,	and line voltage to the analyzer must be within the range
of 105 to 130 Vac, inclusive.

''"Designation of ranges less than 500 ppb is based on meeting the same
absolute performance specifications required for the 0-to-500-ppb range;
proportionately more restrictive performance specifications applicable to
lower ranges have not been established.

64


-------
Meloy SA285E Sulfur Dioxide

TABLE 9. MELOY SA285E LABORATORY PERFORMANCE TEST RESULTS





EPA

Manufacturer's

EPA

Performance parameters

Units

specs.3

test results'3

test results0

Noise-0% URL

ppm

0.005

<0.001



Noise — 80% URL

ppm

0.005

0.001



Lower detectable limit

ppm

0.01

0.009



Interferents

ppm







C02



±0.02

-0.014



CO



+0.02

-0.001

TESTING

H20



±0.02

- 0.001

NOT YET

h2s



±0.02

0.004

PERFORMED

o3



±0.02

-0.002



Total



<0.06

0.022d



Zero drift - 12 h

ppm

+0.02

0.001d



Zero drift - 24 h

ppm

±0.02

0.002d



Span drift — 20% URL

%

+20.0

2.25d



Span drift-80% URL

%

±5.0

1.22d



Lag time

min

20

<0.1



Rise time

min

15

0.8



Fall time

min

15

0.4



Precision — 20% URL

ppm

0.010

< 0.001



Precision — 80% URL

ppm

0.015

0.002



aFrom EPA equivalency regulations, 40 CFR Part 53,

^Average, from manufacturer's application for equivalency determination.

c Average, from EPA Phase I postdesignation tests.

^Average of absolute values.

65


-------
Meloy SA285E Sulfur Dioxide

CONCENTRATION, ppm

ANALYZER:	MELOY

MODELS	SR 285 E
SERIRL No. : 7E140

POLLUTRNT:	SOS
RANGE: .5 ppm

LOCRTION:	DRMDF

DRTE:	03/30/79
DATE CODE; 7S.089

NRME;	PHRSE II

Y - MX + R

M - 202.713
95* C.I.- + 2.776

R - 4.G42
95* C.I.- + .717

r « .99995

DATR:	__Y

1

0.000

5. 100

2

.078

20.000

3

, 157

36.300

4

.260

57.300

5

.363

78.500

6

.412

88.100

Figure 32. Typical calibration curve for the Meloy SA285E.

66


-------


-£2
Q.
CL

t-
Li_
M

ql

a

o

£K

UJ
N

¦P
C
(D
O

£-
ID

a

H

Lu
M

CK

a

"Z.

cc

OL

cn

+30
+20
+ 10
+0
-10
-20
-30

TOT. NET ZERO DRIFT, ppb +.9
NUMBER OF DRIFT PERIODS: 55
AVE DRIFT PERIOH, days: 2.7
RVE [DRIFTj/PERIOD, ppb: 1.5
STD DEV. ZERO DRIFT, ppb; 2.4

J	L

A

A



_L_

_L

A

A

45

50	75	90	105	120

DRY OF YERR

135

150

165

180

TOTRL NET SPRN DRIFT, %: +15.4
NUMBER OF DRIFT PERIODS! 55
RVE DRIFT PERIOD, days: 2.7
RVE |DRIFT|/PERIOD,	1.8

STD DEV. SPRN DRIFT,	2.3

90	105	120

DAY OF YERR

180

Figure 33. Zero and span drift in the Meloy SA285E S02 readings during the

Phase II ambient monitoring test.


-------
Meloy SA285E Sulfur Dioxide

. 5

.4

.3

. 1

0

s/t



//*



0



Y-MX+fl
M= 1.018

(not s i g . )
R= ~~. 001

C s i gn i f . )
r — .99926
N= 3290

.1	.2	.3	.4

RVERRGE OF REFERENCE HNRLYZERS, ppm

.5

Figure 34. Relationship between the Meloy SA285E and the average
of the other analyzers during the Phase II ambient
monitoring test.

68


-------
Meloy SA285E Sulfur Dioxide

(n bts
u

z 169

Ul

K 435

ac

3 735
U

O 713
£ 657
K 298

ul

m
z

Z3
Z

TOTRL No. OF DRTfl PRIRS:	3290

CORRELATION COEFFICIENT:	.999261

HERN DIFFERENCE, ppb;	+.50

STD. DEVIRTION OF DIFFERENCES,	ppb; 3.4297

MRX. ABSOLUTE DIFFERENCE, ppb:	17

NO. ABSOLUTE DIFFERENCES >20;	0

Figure 35. Frequency distribution of differences in hourly S02 ambient
air readings between the Meloy SA285E and the average of
simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. All readings corrected
for zero and span drift.

69


-------
Meloy SA285E Sulfur Dioxide

I
12
54
SB
121
58

01 **

hi

SJ 50
z

Id

Of	62
OL

g	204

o	3?e

£	549

Q£ 42?

W

m

X 438

3

Z ^gg
l?6
47
10
0
0
0

J3

a.
a.

u
o
z
u
a:
u
u,

Lu

+ 19
+1
+ 15
+ 13
+ 11
+3
+7
+5
+3
+ 1
-1
-3
-5
-7

-a

-li
13
15
17
19

"I

7

"!3 7

— IS —I

TOTRL No. OF DRTR PRIRS:
CORRELRTION COEFFICIENT:

MERN DIFFERENCE, ppb:
STD. DEVIATION OF DIFFERENCES,
MRX. RBSOLUTE DIFFERENCE, ppb;
NO. RBSOLUTE DIFFERENCES >20;

ppb:

3290

-1 ,
6,
21
1

998228
75

4814

Figure 36. Frequency distribution of differences in hourly S02 ambient
air readings between the Meloy SA285E and the average of
simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. Meloy SA285E readings are
not corrected for zero and span drift.

70


-------
Meloy SA285E Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation

Replace exhaust filter.

Replace zero air filter.

Replace Teflon sample line.

Clean rotameters.

Rebuild vacuum pump.

Replace H2S scrubber.

Clean burner chamber.

Replace optical window.

Clean exhaust orifice.

Clean dilution orifice.

Clean chassis.

Clean fan.

Clean fan filter.

Replace H2 filter (if necessary).

Replace flow capillary (if necessary).

Replace H2 regulator.

Repl ace solenoid valves.

Malfunctions

Frequency

Every 3 mo

Every 3 mo

Annually

Annually

Annually

Annually

Annually

Annually

Annually

Annually

Annually

Annually

Annually

Annually

Every 24 mo

Every 24 mo

Every 24 mo

12-11-78

H2 filter was clogged
(see General Comments)

General Comments

On 12-11-78, during a routine Phase II zero/span check, it was noted
that the analyzer H2 flow rate had decreased significantly while the air
flow rate had increased significantly. The dilution air orifice was cleaned
and, per the manufacturer's recommendation, the H2 filter assembly (a
sintered frit that was badly clogged) was removed. The removal of the H2
filter corrected both flow rate problems, and normal operation was resumed.
A new filter assembly was ordered; it arrived and was installed on 2-26-79.

During 215 days of Phase II testing, data were lost for 0 days because
of malfunctions.

Currently, Phase I testing has not been performed.

Phase II Testing Operator's Comments

This analyzer achieved equilibrium on span, gases in less than 15 min.

The voltage test modes featured in this analyzer serve as convenient
quality control checks on the day-to-day condition of the instrument's
electronics.

71


-------
Meloy SA700

Meloy SA700 Sulfur Dioxide

Figure 37. Meloy Model SA700 Fluorescence Sulfur Dioxide Analyzer
(ESQA-0580-046). Columbia Scientific Industries,
11950 Jollyvilie Road, Austin, TX 78759.

General Description

The Meloy Model SA700 fluorescence sulfur dioxide analyzer operates on
the principle of fluorescence as described in Appendix A. It differs from
many fluorescence-type analyzers in that the source UV radiation is continuous
rather than modulated by mechanical chopping or electronic puis ing. The
analyzer is equipped with a temperature-controlled scrubber to remove
aromatic hydrocarbons from the sample that would otherwise produce a positive
interference with the fluorescence technique. Removal of ambient moisture
from the sample is effected by a permeation-type dryer, although moisture
removal is not required for interference rejection due to judicious selection
of the excitation wavelength that minimizes the effect of this potential
interferent.

To be used as an equivalent analyzer, the Model SA700 must be operated
on the G-to-250,* 0-to-500, or the 0-to-l,000-ppb range with a time constant

"Designation of ranges less than 500 ppb is based on meeting the same
absolute performance specifications required for the 0-to-500-ppb range;
proportionately more restrictive performance specifications applicable to
lower ranges have not been established.

72


-------
Meloy SA700 Sulfur Dioxide

switch position of either 2 or 3, with or without any of the following
options:

FS-I	Current output

FS-2	Rack mount conversion

FS-2A	Rack mount conversion

FS-2B	Rack mount conversion

FS-3	Front panel mounted digital meter

FS-5	Auto/manual zero/span with status

FS-6	Remote/manual zero/span with status

FS-7	Auto zero adjust.

Ambient air surrounding the analyzer must be within the range of 20° to
30° C, inclusive, and line voltage to the analyzer must be within the range
of 105 to 130 Vac, inclusive.

73


-------
Meloy SA700 Sulfur Dioxide

TABLE 10. MELOY SA700 LABORATORY PERFORMANCE TEST RESULTS





EPA

Manufacturer's

EPA

Performance parameters

Units

specs.8

test results'*

test results0

Noise —0% URL

ppm

0.005

<0.001

0.001

Noise-80% URL

ppm

0.005

0.001

0.001

Lower detectable limits

ppm

0.01

0.010

0.009

Interferents

ppm







Meta-xylene



±0.02

Not tested

0.009

no2



+ 0.02

-0.006

0.000

NO



±0.02

0.004

0.009

°3



±0.02

-0.004

-0.003

Naphthalene



±0.02

0.000

< -0.001

H2S



±0.02

Not tested

0.001

h2o



±0.02

< -0.001

0.005

Total



<0.06

0.015d

0.028d

Zero drift — 12 h

ppm

±0.02

0.003d

0.012d

Zero drift —24 h

ppm

±0.02

0.002d

0.003d

Span drift—20% URL

%

±20.0

1.77d

2,21d

Span drift-80% URL

%

±5.0

0.52d

1.28d

Lag time

min

20

0.2

0.8

Rise time

min

15

2.3

1.8

Fall time

min

15

3.0

1.8

Precision-20% URL

ppm

0.010

<0.001

0,000

Precision-80% URL

ppm

0.015

<0.001

0.001

aFrom EPA equivalency regulations, 40 CFR Part 53.
b Average, from manufacturer's application for equivalency determination,
cAverage, from EPA Phase I postdesignation tests.

^Average of absolute values.

74


-------
Meloy SA7Q0 Sulfur Dioxide

100

C

-------
Meloy SA7QQ Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation

Clean or replace pneumatic lines
Clean rotameter
Clean flow control valve
Repack zero air filter
Replace solenoid valve
Rebuild or replace pump
Clean dryer

Replace hydrocarbon scrubber
Clean optics

Clean fan, air filter, chassis
Clean reactor

Frequency

Annually

Annually

Annually

Quarterly

Biannual 1y

Annually

Annually

Annually

Quarterly

Quarterly

Annually

Malfunctions

May 1982

During initial setup and calibration, the intensity
circuit was out of the control range and readjustment
was required. After consultation with Meloy and
performance of the alignment procedure, the problem
was corrected.

July 1-8, 1982

January 26, 1983

Analyzer displayed excessive zero drift, approximately
2 to 3 percent/day. Meloy indicated that this was
not normal performance and, at their suggestion, the
analyzer was returned for warranty service.

Analyzer output was erratic. A Meloy representative
replaced the UV lamp, installed a new-style lamp
mount, and performed alignment and intensity circuit
control adjustments to correct the problem.

March 18, 1983

Span drift test failure. The lamp intensity control
circuit was out of adjustment (see General Comments).

General Comments

On March 18, 1983, the SA700 analyzer failed a span drift test due to
a malfunction of the lamp intensity control circuit. Meloy representatives
arrived with a new SA700 analyzer on March 24, 1983. After performing the
intensity circuit control adjustment procedure and checking the analyzer
for 2 days, Meloy representatives could find no cause for the malfunction.
To avoid additional delay, EPA and Meloy agreed to repeat the span drift
test on the new SA700 analyzer. No test failures were observed on the new
analyzer. Performance of the malfunctioning analyzer's intensity control
circuit later appeared to be normal and no reason for the malfunction was
found.

76


-------
		Monitor Labs 8450 Sulfur Dioxide

Monitor Labs 8450

Figure 39. Monitor Labs Model 8450 Sulfur Monitor (EQSA-0876-013).
Monitor Labs, Incorporated, 10180 Scripps Ranch
Boulevard, San Diego, CA 92131.

General Description

The Monitor Labs 8450 sulfur dioxide analyzer operates on the prin-
ciple of flame photometric detection as described in Appendix A. It is
equipped with a heated silver gauze scrubber to remove ambient hydrogen
sulfide, making it unresponsive to that compound. This type of scrubber
has been known to produce both positive and negative interference to the
S02 analysis when ambient levels of ozone are present in the sample. (See
description of the measurement principle in Appendix A.) These interfer-
ences, however, are minimized by passing the sample through a heated chamber
upstream of the scrubber, thus destroying any ozone present before it
reaches the silver gauze.

The digital electronic techniques used to produce a linear output also
accept and process negative data. This capability prevents the possibility
of nonapparent negative zero offset in the linearization process.

The analyzer is packaged in two separate modules. One contains the
signal-processing electronics and the other contains the detector and the
sample-handling pneumatics. To be used as an equivalent analyzer, the
Model 8450 must be operated on a range of either 0 to 0.5 or 0 to 1.0 ppm,




-------
Monitor Labs 8450 Sulfur Dioxide

with a 5-second time constant, with a Model 8740 hydrogen sulfide scrubber
in the sample line, and with or without any of the following options:

BP	Bipolar signal processor

V	Zero/span valves

VT	Zero/span valves and timer

TF	TFE sample particulate filter

IZS	Internal zero/span module

CLO	Current loop output

DO	Status remote interface.

Ambient air surrounding the analyzer must be within the range of 20° to
30° C, inclusive, and line voltage for the analyzer must be within 105 to
125 Vac, inclusive.

78


-------
Monitor Labs 8450 Sulfur Dioxide

TABLE 11. MONITOR LABS 8450 LABORATORY PERFORMANCE

TEST RESULTS

EPA	Manufacturer's	EPA

Performance parameters Units	specs.8	test results'5	test results0

Noise - 0% URL

ppm

0.005

0.001

<0.001

Noise - 80% URL

ppm

0.005

0.001

0.001

Lower detectable limits

ppm

0.01

0.013

0.009

Interferents

ppm







CO



+ 0,02

< -0.001

-0.002

h2o



±0.02

-0.012

-0.003

C02



±0.02

-0.018

-0.019

h2s



+0.02

0.001

0.017

°3



+ 0,02

0.004

-0.007

Total



<0.06

0.035d

0.048d

Zero drift —12 h

ppm

±0.02

0.001d

0.001d

Zero drift —24 h

ppm

±0.02

0.001d

0.004d

Span drift — 20% URL

%

±20.0

1.06d

1.49d

Span drift-80% URL

%

+ 5.0

0.85d

1.00d

Lag time

min

20

<1

0.3

Rise time

min

15

<1

1.6

Fall time

min

15

< 1

2.1

Precision-20% URL

ppm

0.010

0.001

0.002

Precision — 80% URL

ppm

0.015

0.001

0.002

aFrom EPA equivalency regulations, 40 CFR Part 53.
k Average, from manufacturer's application for equivalency determination.
cAverage, from EPA Phase I postdesignation tests, Serial Number 147.
d Average of absolute values.

79


-------
Monitor Labs 8450 Sulfur Dioxide

100

0.0

J	.2	.3

CONCENTRRTION, ppm

ANALYZER:

MODEL;

SERIfiL No.

POLLUTANT:

RANGE:

LOCATION:

DATE:

DATE CODE:
NAME:

MONITOR 1
8450
14?

S02

. 5 ppm
DAMDF
03/30/?S

79.089
PHASE II

.ABS

Y

MX + R

M = 203.674

95V, C.I." + 2.764

R « 5.939

35'4 C.I.- + .713

r «= .99995

DATA:

_„X„„

	Y	

1

0.000

5.700

2

.078

21.700

3

. 15?

38.300

4

.260

59.100

5

.363

80.100

6

.412

89.400

Figure 40. Typical calibration curve for the Monitor Labs 8450,

80


-------
CO

St

o.
a.

i-t
£K
O

o

Q£

U
N

~»

c
a

o

Q

a

a:
o

z
a:

JL

w

+30
+20
+ 10
+0
-10
-20

TOT. NET ZERO DRIFT, ppb -4.8
NUMBER OF DRIFT PERIODS; 51
RVE DRIFT PERIOD, days: 2.?
RVE |DRIFT|/PERIOD, ppb: 1.3
STD DE^. ZERO DRIFT, ppb: 2.0

-30

J	1	1	L

JL

J	L

_l_



J	L

A

-I	L

30	45	60	75	90 105 120

DRY OF YEAR

135

150

165

+ 15
+ 10
+5
+0
-5
-10

TOTRL NET SPRN DRIFT, Xs
NUMBER OF DRIFT PERIODS:
RVE DRIFT PERIOD, days:
RVE |DRIFT|/PERIOD, X:
STD DEV. SPRN DRIFT, Xs

•15

A.

J	L

JL

J	L

-J	L

A

J	1	1	L

-L

6 .0

J	1	1	L

jflL

J	L

30	45	60	75	90 105 120

DRY OF YERR

135

150

165

j	1	1	1

180

J—i

180

Figure 41. Zero and span drift in the Monitor Labs 8450 S02 readings during the

Phase II ambient monitoring test.


-------
Monitor Labs 8450 Sulfur Dioxide

0

0

. 1

.3

.4

.5

RVERRGE OF REFERENCE RNRLYZERS, ppm

Figure 42, Relationship between the Monitor Labs 8450 and the average
of the other analyzers during the Phase II ambient
monitoring test.

82


-------
Monitor Labs 8450 Sulfur Dioxide

0

4
3

16

IB

40

U) 199
UJ

u 242
UJ

QC 361
(K

g 428
U

O 445
368
0, 603

UJ
m

S 29?
3

Z 104
39

5

0
0
0

ja
a.
a.

UJ

u
z

UJ

m
UJ
u_
u.

+ 19
+ 17
+ 15
+ 13
+ 11
+9
+7
+5
+3
+ 1
-1
-3
-5
-7
-3
-11
13
-15
-17
-19

"nT

-ia -

TOTRL No. OF DRTR PRIRS:
CORRELRTION COEFFICIENT:

MERN DIFFERENCE, ppb:
STD. DEVIRTION OF DIFFERENCES, ppb
MRX. RBSOLUTE DIFFERENCE, ppb;
NO. RBSOLUTE DIFFERENCES >20:

3170

.999087
+. 02
4.7655
19
0

Figure 43. Frequency distribution of differences in hourly S02 ambient
air readings between the Monitor Labs 8450 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. All readings corrected
for zero and span drift.

83


-------
Monitor Labs 8450 Sulfur Dioxide

0
0
0

0

1

3
9
65
160
223

U)

U
U

z
u
on
on
¦3
u
u

O 428

306

U.

O

P- 504

UJ
CQ

£ 385
D

z 571
314
87
29
3
2
0

jQ
Q.

a.

U
u
z
u
q:
u
u.
u,

M

Q

+ 19
+ 17
+ 15
+ 13
+ 11
+9
+7
+5
+3
+ 1
-1
-3
-5
-7
-9
-11
-13
-15
-17
-19

TOTRL No. OF DRTR PRIRS:

C0RRELRTI0N COEFFICIENT:

MERN DIFFERENCE, ppb:

STD. DEVIATION OF DIFFERENCES, ppb:

MRX. RB50LUTE DIFFERENCE, ppb:

NO. ABSOLUTE DIFFERENCES >20:

3170

.998799
-4.01

4 .5471
18
0

Figure 44. Frequency distribution of differences in hourly S02 ambient
air readings between the Monitor Labs 8450 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. Monitor Labs 8450
readings are not corrected for zero and span drift.

84


-------
Monitor Labj 8450 Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation

Frequency

Examine rotameter assembly.

Examine exhaust lines.

Check vacuum.

Examine, clean, or replace fan filter.

Check or replace chopper drive belt.

Replace sample filter.

Check H2 regulator.

Check exhaust orifice and filter.

Replace zero/span charcoal filter.

Replace H2S scrubber.

Examine or clean window.

Examine or clean purge air filter.

Examine or clean zero/span air filter.

Weekly
Weekly
Weekly

Every 15-30 days

Every 30 days
Every 30 days
Every 150 days
Every 150 days
Every 300 days
Every 300 days
Every 300 days
Every 300 days
Every 300 days

Various other maintenance operations are specified on an "as required"
basis.

Malfunctions

General Comments

During Phase I testing, the analyzer failed initial ozone and hydrogen
sulfide interferent tests. Additionally, performance during the C02 inter-
ferent test was marginal. According to the manufacturer's recommendation,
the H2S scrubber pads were replaced but with no improvement in test perform-
ance. The manufacturer then changed the formulation of the scrubber pads
from plated to pure silver pads With the new, pure silver pads installed,
the analyzer passed the ozone and H2S interferent tests. The manufacturer
supplied the reformulated scrubber pads to all analyzer owners who were
originally supplied with the old pads. The C02 interference, however,
remained marginal.

During 215 days of Phase II testing, data were lost for 17 days be-
cause of malfunctions.

February 1977 (initial startup)

Linearizer card replaced.
Power supply card replaced.

During Phase I testing

Failed initial ozone and
and hydrogen sulfide
interferent tests (see
General Comments).

1-17-79

Chopper motor was
replaced.

85


-------
Monitor Labs 8450 Sulfur Dioxide

Phase II Testing Operator's Comments

The analyzer required 30 to 45 min to come to equilibrium on span gas,
which lengthened the time required for routine zero/span operations.

The electrical and optical test modes featured in this analyzer serve
as convenient quality control checks on day-to-day condition of the in-
strument .

The LED display is more convenient to read than a conventional meter
display.

86


-------
	Monitor Labs 8850 Sulfur Dioxide

Monitor Labs 8850

Figure 45. Monitor Labs Model 8850 Fluorescent S02 Analyzer
(EQSA-0779-039). Monitor Labs, Incorporated,
10180 Scripps Ranch Boulevard, San Diego, CA 92131.

General Description

The Monitor Labs 8850 sulfur dioxide analyzer operates on the prin-
ciple of fluorescence as described in Appendix A. It uses mechanically
chopped ultraviolet excitation radiation to provide signal modulation for
amplification. The analyzer is equipped with an unheated scrubber to
remove aromatic hydrocarbons that would otherwise produce a positive inter-
ference with the fluorescence technique. Removal of ambient moisture is
not required because judicious selection of the excitation wavelength
minimizes the effect of this potential interferent.

To be used as an equivalent analyzer, the Model 8850 must be operated
on a range of either 0 to 0.5 or 0 to 1.0 ppm, with an internal time constant
setting of 55 seconds, with a TFE sample filter installed on the sample
inlet line, and with or without any of the following options:

03A

Rack

03B

Slides

05A

Valves zero/span

06A

IZS, internal zero/span source

06B,C,D

NBS-traceable permeation tubes

08A

Pump

09A

Rack mount for option 08A

010

Status output w/connector

013

Recorder output options

014

DAS output options

017

Low flow option.

87


-------
Monitor Labs 8850 Sulfur Dioxide

Ambient air surrounding the analyzer must be within the range of 20° to
30° C, inclusive, and line voltage to the analyzer must be within the range
of 105 to 125 Vac, inclusive.

TABLE 12. MONITOR LABS 8850 LABORATORY PERFORMANCE

TEST RESULTS

EPA	Manufacturer's	EPA

Performance parameters Units specs.8	test results'*	test results0

Noise - 0% URL

ppm

0.005

<0.001

<0.001

Noise - 80% URL

ppm

0.005

0.001

0.001

Lower detectable limits

ppm

0.01

0.019

0.010

Interferents

ppm







h2o



±0.02

0.003

0.007

NO



±0.02

0.003

0.005

NO,



±0.02

0.001

0.002

o3



±0.02

-0.004

-0.004

Naphthalene



±0.02

< -0.001

0.001

Meta-xylene



±0.02

Not tested

0.002

H2S



±0.02

Not tested

0.000

Total



<0.06

0.011d

0.021d

Zero drift—12 h

ppm

±0.02d

0.001d

0.002d

Zero drift —24 h

ppm

±0.02

0.001d

0.005d

Span drift-20% URL

%

±20.0

1.38d

3.46d

Span drift -80% URL

%

± 5,0

1.41d

0.56d

Lag time

min

20

0.3

0.5

Rise time

min

15

3.8

3.9

Fall time

min

15

3.8

3.9

Precision-20% URL

ppm

0.010

0.001

<0.001

Precision-80% URL

ppm

0.015

0.002

0.001

aFrom EPA equivalency regulations, 40 CFR Part 53.
kAverage, from manufacturer's application for equivalency determination.
cAverage, from EPA Phase I postdesignation tests.

^Average of absolute values.

88


-------
Monitor Labs 8850 Sulfur Dioxide

100

0.0

. 1	.2	.3	.4

CONCENTRATION, ppm

RNRLYZER:
MODEL:
SERIAL No,
POLLUTANTi
RANGE:
LOCRTION:
DRTE:

DATE CODE:
NAME;

MONITOR LRBS

8850

90

SOS

.5 ppm
135

07/14/82

82.195
C F SMITH

Y - MX + R DRTfls

M - 201.720
95* C.I.- + 2.093

R - 5.172
95* C.I.- ± .578

r - .99936

1

0.000

5.100

2

.094

24.100

3

.154

36.750

4

.228

50.803

5

.313

68.200

6

.388

83.250

7

.447

95.600

Figure 46. Typical calibration curve for the Monitor Labs 8850.

89


-------
Monitor Labs 8850 Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation

Recharge or replace zero/span module charcoal filter

Replace gas sample particulate filter

Examine rotameter assembly

Examine exhaust lines and vacuum

Examine exhaust orifice

Examine exhaust orifice filter

Examine fan filter

Examine zero/span air filter

Examine UV lamp

Replace UV lamp

Replace catalyst

Frequency

Every 90 days
Every 30 days
Weekly
Weekly

Every 180 days
Every 180 days
Every 15-30 days
Annually
Every 30 days
Annually
Annually

Various other maintenance operations are specified on an "as-required" basis,
Malfunctions

Span drift test failure (see General Comments).

Span drift test failure (see General Comments),

Span drift test failure (see General Comments).

October 13, 1982*
t

October 29, 1982
May 24, 1983+

General Comments

On October 13, 1982, the analyzer (SN 90) failed a span drift test.
The cause of the failure was not apparent. The analyzer was originally
purchased in March 1980 and was operated continuously from September 23,
1980, until the end of June 1982. A second (new) 8850 analyzer (SN 717)
was tested for span drift to determine whether the span drift failure was
caused by prolonged use (approximately 2 years) of the analyzer prior to
the tests or whether it was typical of the performance of the analyzer.

On October 29, 1982, the new 8850 analyzer (SN 717) also failed a span
drift test. Monitor Labs was immediately contacted concerning the test

^Monitor Labs 8850 SN 90.

^Monitor Labs 8850 SN 717.

^Monitor Labs 8850 SN 90, modified as
#885-0833 and #885-0834.

per Monitor Labs change drawing

90




-------
Monitor Labs 8850 Sulfur Dioxide

failures. After conducting some abbreviated span drift tests on 10 current
production analyzers, Monitor Labs observed the same performance deficiency
on approximately half of the tested analyzers. The performance deficiency
was attributed to the PMT thermostatic contro1 circuits. The performance
deficiency has been corrected with a simple modification of the mixer
circuit board assembly. To maintain the designation status of the Model
8850 analyzers sold prior to implementation of this modification, Monitor
Labs has supplied the ultimate purchasers of those previously sold analyzers
with an update kit to implement the modification.

On May 24, 1983, a modified 8850 analyzer (SN 90) again failed a span
drift test. This failure was attributed to a malfunction of the PMT thermo-
electric cooler. Subsequent to replacement of the thermoelectric cooler,
the modified analyzer successfully passed the repeated span drift tests.

91


-------
Philips PW9700

Philips PW970D Sulfur Dioxide

Figure 47. Philips PW97QQ SO2 Analyzer (EQSA-0876-011).
Philips Electronic Instruments, Inc.,
85 McKee Drive, Mahwah, NJ 07430.

General Description

The Philips Model PW9700 sulfur dioxide analyzer operates 011 the
principle of coulometric detection as described in Appendix A. It is
equipped with a heated chemical scrubber that removes species interfering
with the coulometric detection of S02 (e.g., oxidizing or reducing species),
thus making it essentially specific for SOg. The analyzer is supplied in
two separate modu!es--one containing the power supply and the vacuum pump
and the other containing the detector, sample handling pneumatics, and
electronics.

This model has been superseded by the Philips Model PW9755, but its
designation is still maintained. To be used as an equivalent analyzer, the
model PW9700 must be operated on the 0-to-0.5-ppm range and with a reference
voltage of 760 mV. Ambient air surrounding the analyzer must be within the
range of 20° to 30° C, inclusive, and line voltage to the analyzer must be
within the range of 105 to 125 Vac, inclusive.

92


-------
Philips PW9700 Sulfur Dioxide

TABLE 13. PHILIPS PW9700 LABORATORY PEftFOfifeJANCE TEST RESULTS





EPA

M&;.v facturer's

EPA

Performance parameter";

Units

specs.®

test results1'

test results0

Noise - 0% URL

ppm

0.005

<0.001

0.000

Noise — 80% URL

ppm

0.005

0.001

0.001

Lower detectable limits

ppni

0.01

0.009

0.007

Interferents

ppm







o3



±0.02

0.008

-0.008

NO



±0.02

0.000

0.001

no2



±0.02

0.004

-0.001

h2s



±0.02

0.001

<0.001

CoHn



±0.02

0.000

-0.001

h2o



±0.02

0.000

0.009

nh3



±0.02

0.000

0.000

HCI



±0.02

<0.001

-0.001

Total



=£0.06

0.013d

0.021d

Zero drift—12 h

ppm

±0.02

<0.001d

0.00ld

Zero drift —24 h

ppm

±0.02

0.002d

0.002d

Span drift-20% URL

%

±20.0

3.00^

5.0^

Span drift-80% URL

%

±5.0

1.01d

1.3d

Lag time

min

20

0,8

0.2

Rise time

min

15

3.5

2.2

Fall time

min

15

3.8

1.4

Precision — 20% URL

ppm

0.010

<0.001

0.001

Precision-80% URL

ppm

0.015

0.000

0.002

aFrom EPA equivalency regulations, 40 CFR Part 53.
k Average, from manufacturer's application for equivalency determination.
cAverage, from EPA Phase I postdesignation tests.

^Average of absolute values.

93


-------
Philips PW9700 Sulfur Dioxide

100

+>
C
Q)
0

L,
QJ
Q.

£
H

Cq
CC

UJ

u

.j

a:
o
tn

.2	.3

CONCENTRATION, ppm

ANALYZER:

MODEL:

SERIAL No.:

POLLUTANT:

RANGE:

LOCRTION:

DRTE:

DRTE CODE:
NRME:

PHILIPS
PW 9700
D1415C
SOH

. 5 ppm
RTI

05/16/78

78.136
PHASE I

Y «= MX + R
M- 200.136

95JJ C.I.- + 1.378

R •= 4.291
95% C.I.- + .382

r •= .9999?

DATA:

X

Y

1

0.000

4.500

2

.083

21.000

3

.099

24.200

4

. 150

33.800

5

.225

49.100

6

.291

62.800

7

.375

79.200

8

.401

84.700

9

.464

97.200

Figure 48. Typical calibration curve for the Philips PW9700.

94


-------
Philips PW970Q Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation

Frequency

Replace glass fiber membrane
in sampler unit

Every 3 months

Flush vacuum pump

Every 3 months

Replace silver gauge of the
selectivity filter

Every 3 months

Replace glass fiber membrane
of aerosol filter

Every 3 months

Replace measuring solution

Every 3 months

Calibrate S02 source

Every 3 months

Mai functions

September 22, 1978

Failure of reference electrode

General Comments

Only Phase I testing has been performed; Phase II testing is not
anticipated as this analyzer has been superseded by the Philips PW 9750.

During 153 days of Phase I testing, data were lost for 23 days due to
malfunctions.

95


-------
			Philips PW9755 Sulfur Dioxide

Philips PW9755

Figure 49. Philips PW9755 Sulfur Dioxide Analyzer (EQSA-0676-010).

Philips Electronic Instruments, Inc., 85 McKee Drive,
Mahwah, NJ 07430.

General Description

The Philips Model PW9755 sulfui lioxide analyzer operates on the
principle of coulometric detection as described in Appendix A. It is
equipped with a heated, chemical scrubber, which removes species interfering
with the coulometric detection of SO2 (e.g., oxidizing or reducing species),
thus making it essentially specific for SO2. The analyzer is packaged in
two separate modules—one containing the power supply and vacuum pump and
the other containing the detector, sample-handling pneumatics, and elec-
tronics. The power supply module can be equipped to power three Philips
instruments simultaneously.

To be used as an equivalent analyzer, the Model PW9755 must be operated
on the 0-to-0.5-ppm range and with a reference voltage setting of 760 mV,
with or without any of the following options:

PW9752/00	Air sampler manifold

PW9753/00	Mounting rack for accessories

PW9750/30	Frame for MTT

PW9750/41	Control clock 60 Hz

PW9754/00	Air distributor.

Ambient air surrounding the analyzer must be within the range of 20°
to 30° C, inclusive, and line voltage to the analyzer must be within the
range of 105 to 125 Vac> inclusive.

96


-------
Philips PW9755 Sulfur Dioxide

TABLE 14. PHILIPS PW9755 LABORATORY PERFORMANCE TEST DATA





EPA

Manufacturer's

EPA

Performance parameters

Units

specs.3

test results'1

test results®

Noise-0% URL

ppm

0.005

<0.001

0.000

Noise-80% URL

ppm

0.005

0.003

0.002

Lower detectable limit

ppm

0.01

0.009

0.009

Interferents

ppm







NO



±0.02

0.000

0.000

o3



±0.02

0.008

-0.004

no2



±0.02

0.004

0.000

h2s



±0.02

0.001

0.000

h2o



±0.02

0.000

0.002

HCI



±0.02

<0.001

0.000

c2h4



±0.02

0.000

< 0.001

nh3



±0.02

0.000

<0.001

Total



<0.06

0.013d

0.006d

Zero drift - 12 h

ppm

±0.02

0.001d

0.002d

Zero drift - 24 h

ppm

±0.02

0.001d

0.003d

Span drift-20% URL

%

±20.0

1.27d

1.35d

Span drift — 80% URL

%

±5.0

0.61d

0.3d

Lag time

min

20

0.6

0.6

Rise time

min

15

2.6

0.7

Fall time

min

15

2.7

0.8

Precision-20% URL

ppm

0.010

<0.001

<0.001

Precision - 80% URL

ppm

0.015

<0.001

0.001

aFrom EPA equivahncy regulations, 40 CFR Part 53.

''Average, from manufacturer's application for equivalency determination.

cAverage, from EPA Phase I postdesignation tests.

^Average absolute values.

97


-------
Philips PW9755 Sulfur Dioxide

CONCENTRRTION, ppm

RNRLYZER;

MODEL!

SERIAL No.

POLLUTRNTs

RRNGE:

LOCRTION:

DRTEs

DRTE CODE:
NRMEt

PHILIPS
PW 9755
612
S02

. 5 ppm
DRMDF
03/30/79

79.009
PHRSE II

Y - MX + R

M - 200.753
35% C.I.- + 2.462

fl - 5.074
95fc C.I.- + .641

r ¦= .99996

DRTR:

X

Y

1

0.000

5.000

2

.078

20.400

3

. 157

37.000

4

.260

57.500

5

.363

76.000

6

.412

87.500

Figure 50. Typical calibration curve for the Philips PW9755.

98


-------
JO
flL

a

fx.
a

o

QL

u

N

O

vO

4*
C
©
o

c
©
a

IK
Q

HE

Q-

0)

+30
+20
+ 10
+0
-10
-20

-30

TOT. NET ZERO DRIFT, ppb
NUMBER OF DRIFT PERIODS:
RVE DRIFT PERIOD, days:
FIVE J DRIFT |/PERIOD, ppb:
STD DEV. ZERO DRIFT, ppb:

J	L

-I	L

J	I	1	I	L

A

J	I	_l	L

A

J	I	J	L

A

j—1_—i—i

30

45

60

75

90 105 120
DRY OF YEAR

135

150

165

180

TOTAL NET SPRN DRIFT, Hi -2.2
NUMBER OF DRIFT PERIODS! 51
RVE DRIFT PERIOD, days:
RVE |DRIFT|/PERIOD, Hz
STD DEV. SPRN DRIFT, %z

+0
-5
¦10

15

j	1	1	1	1	L

J	L

A

J	!	1	1	L

A

4	L

J	L

A

J	L

J	1

30

45

60

75

90 105 120
DRY OF YERR

135

150

165

180

Figure 51. Zero and span drift in the Philips PW9755 S02 readings during the

Phase II ambient monitoring test.


-------
Philips PW9755 Sulfur Pi oxide

.5

?71

E

a
a.

in
in
r^-

CT)
2
Q.

U)
CL

X
Q.

.4 -

.3

.2

. 1

0

0



Y-MX+fl
M~ 1.005

(s1gn1f.)
fl~ .005

(s1gn1f.)
r~ .99839
N«= 3055

.1	.2	.3	.4

AVERAGE OF REFERENCE ANALYZERS, ppm

.5

Figure 52. Relationship between the Philips PW9755 and the average
of the other analyzers during the Phase II ambient
monitoring test.

100


-------
Philips PW9755 Sulfur Dioxide

TOTAL No. OF DflTfl PRIRS:
CORRELATION COEFFICIENT:

MEAN DIFFERENCE, ppbs
STD. DEVIATION OF DIFFERENCES, ppb
MAX. ABSOLUTE DIFFERENCE, ppb;
NO. ABSOLUTE DIFFERENCES >20;

3055

.998388
+5.68
4.6768

25
15

Figure 53. Frequency distribution of differences in hourly S02 ambient
air readings between the Philips PW9755 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. All readings corrected
for zero and span drift.

101


-------
Philips PW9755 Sulfur Dioxide

U1
U
Z
u

QL
OC.
Zt
U
O

u.
o

QL
U
ff»
X,

Z3
Z

45



+18

60



+17

56



+15

127



+13

158



+ 11

224



+9

253



+7

310





34?

M

a

a

+5
+3

446

u
o

+1

500

?





Ui

-1

376

IK

til



U.

-3

96

It.



21

M

P

-5
-7

13



-9

3



-11

3



-13

3



-15

0



-17

0



-19

0





1

TOTAL No. OF DRTfl PAIRS:	3055

CORRELATION COEFFICIENT:	.997944

MEAN DIFFERENCE, ppb:	+4.46

STD. DEVIATION OF DIFFERENCES,	ppb: 5.8336

MAX. ABSOLUTE DIFFERENCE, ppb:	2?

NO. ABSOLUTE DIFFERENCES >20:	33

Figure 54. Frequency distribution of differences in hourly S02 ambient
air readings between the Philips PW9755 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test. Philips PW9755 readings
are not corrected for zero and span drift.

102


-------
Philips PW9755 Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation

Frequency

Replace sample filter.

Replace scrubber cartridge.

Check electrolyte level.

Measure flow through detector.

Measure total flow.

Replace pump diaphragms.

Check pump vacuum.

Replace electrolyte.

Clean metering orifice.

Replace metering orifice filter materials.

Clean pump silencer.

Clean Peltier element.

Check pump bearings.

Replace pump crank.

Clean flushing orifice.

Replace flushing orifice filter materials.

Repack zero air filter.

Monthly

Every 3 mo
Every 3 mo
Every 3 mo
Every 3 mo
Every 3 mo
Every 3 mo
Every 6 mo
Every 6 mo
Every 6 mo
Every 6 mo
Annually

Annually
Annually

Every 24 mo
Every 24 mo
Every 24 mo

Malfunctions

None.

General Comments

During Phase II testing, the flow rate through the detector dropped
continuously from 145 cm3/min (-3.3 percent) on 2-1-79 to 135 cm3/min
(-10 percent) on 6-4-79. The flow specification for the metering orifice
is 150 cm3/min (±3 percent) and is defined at conditions of 1 atm and
20° C. Ultrasonic cleaning of the orifice improved the flow rate only
marginally. According to factory service personnel^ this flow degradation
results from corrosion of the orifice by the Br2/Br measuring solution.
Orifices can be purchased with bore inserts that are manufactured from a
material that is less susceptible to this corrosion (e.g., sapphire).

This instrument features no provision for adjusting the zero response
level. The zero voltage is always some small, upscale value (specification
<10 mV) that depends on the purity and age of the electrolyte. During
Phase II testing, it was noted that a span adjustment also produced a
proportional change in the zero response. This change results because the
span adjustment is, in reality, an amplifier or amplified gain adjustment.
Because the zero baseline is always some finite voltage greater than zero,
any change in the span adjustment (amplification factor) will cause a
corresponding change in the zero response.

103


-------
Philips PW9755 Sulfur Dioxide

During 172 days of Phase II testing, data were lost for 0 days because
of malfunctions.

Phase II Testing Operator's Comments

This instrument achieved equilibrium on span gases in less than 5
minutes.

Although the instrument manual was complicated—containing great
detail on all subsystems and options—the individual sections were well
written and easy to follow.

Throughout the Phase I and II testing periods this instrument required
only routine preventive maintenance.

104


-------
Thermo Electron 43

Thermo Electron 43 Sulfur Dioxide

Figure 55. Thermo Electron Model 43 Pulsed Fluorescent SO2 Analyzer

(EQSA-0276-009). Thermo Electron Corporation, Environmental
Instruments Division, 108 South Street, Hopkinton, MA 01748.

General Description

The Thermo Electron Model 43 sulfur dioxide analyzer operates on the
principle of fluorescence as described in Appendix A. It uses pulsed
ultraviolet radiation. Water vapor, which can be an interfeient in this
technique, is removed from the sample stream by a permeation dryer.

The analyzer is equipped with an unheated chemical scrubber called a
"cutter." The cutter removes aromatic hydrocarbons (e.g., naphthalene)
that would otherwise produce a positive interference in the fluorescence
technique.

To be used as an equivalent analyzer, the Model 43 must be operated on
a 0-to-0.5- or 0-to-l.0-ppm range, with or without any of the following
options:

001	Rack mounting for standard 19-inch relay rack

002	Automatic actuation of zero and span solenoid valves

003	Type S flash lamp power supply

004	Low flow.

Ambient air surrounding the analyzer must be within the range of 20°
to 30° C, inclusive, and line voltage must be within the range of 105 to
125 Vac, inclusive.

105


-------
Thermo Electron 43 Sulfur Dioxide

TABLE 15. THERMO ELECTRON 43 LABORATORY PERFORMANCE

TEST RESULTS





EPA

Manufacturer's

EPA

Performance parameters

Units

specs.®

test resuttsb

test results6

Noise - 0% URL

ppm

0.005

0.001

0.001

Noise - 80% URL

ppm

0.005

0.001

0.002

Lower detectable limits

ppm

0.01

0.003d

0.009

Interferents

ppm







Meta-xylene



±0.02

0.000

< -0.001

no2



±0.02

0.000

<0.001

NO



±0.02

0.003

0.003

03



±0.02

-0.001

-0.002

H2S



±0.02

0.000

Not tested

C02



±0.02

0.000

Not tested

CO



±0.02

0.001

Not tested

Naphthp'gne



±0.02

_e

_e

Total



<0.06

0.005r

0.005'

Zero drift —12 h

ppm

±0.02'

0.005'

0.004'

Zero drift —24 h

ppm

±0.02

0.0011

0.003'

Span drift — 20% URL

%

±20.0

0.52f

2.22'

Span drift-80% URL

%

± 5.0

0.46'

1.82f

Lag time

min

20

0.2

0.5

Rise time

min

15

3.9

4.9

Fall time

min

15

4.7

5.3

Precision-20% URL

ppm

0.010

0.001

0001

Precision-80% URL

ppm

0.015

0.001

0.001

aFrom EPA equivalency regulations, 40 CFR Part 53,
kAverage, from manufacturer's application for equivalency determination.
cAverage, from EPA Phase I postdesignation tests.

^Concentration significantly below the target value of 10 ppb allowed for this peiformance
parameter since decreased test concentrations here represent a more severe test condition.

Quantitative tests with naphthalene were not conducted; instead, the test analyzer was shown
to have no measurable response to a concentration of naphthalene which produced at least
a mid-scale response when the cutter was removed.

'Average of absolute values.

106


-------
Thermo Electron 43 Sulfur Dioxide

ANALYZER: THERMO ELECTRON Y

MX + fi

MODEL:

SERIRL No.

POLLUTANT:

RANGli:

LOCATION:

DATE:

DATE CODE:
NAME:

43

329464
S02

.5 ppm
DAMDF
03/30/79

79.089
PHRSE II

M «= 199. 142
95*4 C.I.- + 2.275

R = 4.698
95% C.I.- ± .587

r «= .9999?

DATA:

X

Y

1

0.000

5.000

2

.076

20.000

3

. 15?

36.000

4

.260

56.100

5

.363

77.000

6

.412

87.000

Figure 56. Typical calibration curve for the Thermo Electron 43.

107


-------
JO
Q-

a

h-

l f
»-i

a.
cj

o
oc

ill

N

TOT. NET ZERO DRIFT, ppb-10.5
NUMBER OF DRIFT PERIODSs 55
RVE DRIFT PERIOD, days: 2.?
RVE fDRIFT J/PERIOD, ppb: 3.3
STD DEV. ZERO DRIFT, ppb; 4.9

90	105 120

DRY OF YEAR

180

o
oo

~»
c
m

o

c.
©

QL

a:

f|

z
w

TOTAL NET SPRN DRIFT, K:-20.4
NUMBER OF DRIFT PERIODS:
RVE DRIFT PERIOD, dayss
RVE J DRIFT J/PERIOD, X:
STD DEV. SPRN DRIFT, Hi

90 105 120
DRY OF YETRR

135

150

165

180

Figure 57. Zero and span drift in the Thermo Electron 43 S02 readings during the

Phase II ambient monitoring test.


-------
Thermo Electron 43 Sulfur Dioxide

.5

.4

.3 -

.2

. 1

0

Y~MX+R

.949
Cs1gn1f . )
R« .003

(not s 1 g. )
rc .99651
N*= 2163

.1	.2	.3	.4

RVERRGE OF REFERENCE RNRLYZERS, ppm

.5

Figure 58. Relationship between the Thermo Electron 43 and the average

of the other analyzers during the Phase II ambient monitoring
test (before manufacturer's refurbishment—see General
Comments).

109


-------
Thermo Electron 43 Sulfur Dioxide

W
Ld
U

z
y

QC
QC
D
U

o
o

u.

o

w
m

T.
D
Z

9









+ 19

17









+ 17

27









+ 15

49









+ 13

61









+ 11

67









+9

35









+7

171

JO

n



+5

210









+3

212

til

+ 1



u

266

z





u

-1

247

Hi

111



Lu

-3

124

U-



97

w
a

-5





-7

121









-9

127









-11

77









-13

64









-15

36









-17

19









-19

47







TOTAL No. OF DRTfl PRIRS:	2163

CORRELATION COEFFICIENT:	.996505

MERN DIFFERENCE, ppb:	-.74

STD. DEVIRTION OF DIFFERENCES,	ppb: 8.2988

MAX. RESOLUTE DIFFERENCE, ppb:	30

NO. ABSOLUTE DIFFERENCES >20:	49

?ure 59. Frequency distribution of differences in hourly S02 ambient
air readings between the Thermo Electron 43 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test (before manufacturer's
refurbishment—see General Comments). All readings corrected
for zero and span drift.

110


-------
Thermo Electron 43 Sulfur Dioxide

W
Id
O

z
u
a

QL
D
o

U
O

a
u
m
s:

3

z

10



+ 19

9



+17

12



+15

21



+13

44



+11

34



+9

62



+7

64





J3
CI.

+5

106

a

+3

92

«k

Ul

o

+1

156

2





Ui

-1

196

QC
lil



u.

-3

166

u.



146

M

p

-5
-7

172



-9

152



-11

173



-13

136



-15

116



-17

51



-19

217





TOTRL No. OF DATA PAIRS:	2163

CORRELRTION COEFFICIENT:	.994618

MEAN DIFFERENCE, ppb:	-6.99

STD. DEVIATION OF DIFFERENCES,	ppb: 11.4140

MAX. ABSOLUTE DIFFERENCE, ppb:	54

NO. ABSOLUTE DIFFERENCES >20:	211

Figure 60. Frequency distribution of differences in hourly S02 amt-'ent
air readings between the Thermo Electron 43 and the av = -age
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test (before manufacturer's
refurbishment—see General Comments). Thermo Electron 43
readings are not corrected for zero and span drift.

111


-------
Thermo Electron 43 Sulfur Dioxide

.5

£

a

a

cn
*r

z
o
on

H

-------
Thermo Electron 43 Sulfur Dioxide

TOTAL No. OF DATA PAIRS:	1046

CORRELATION COEFFICIENT:

MEAN DIFFERENCE, ppb:

STD. DEVIATION OF DIFFERENCES, ppb;
MAX. ABSOLUTE DIFFERENCE, ppb;
NO. ABSOLUTE DIFFERENCES >20:

.998475
+ .20
4.0650
14
0

Figure 62. Frequency distribution of differences in hourly S02 ambient
air readings between the Thermo Electron 43 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test (after manufacturer's
refurbishment—see General Comments). All readings corrected
for zero and span drift.

113


-------
Thermo Electron 43 Sulfur Dioxide

m
u
u
z
ui

QC.

m

Z3
U

u
o

u.
o

UJ
m
x

3
Z

0



+13

0



+17

0



+1S

e



+13

0



+11

0



+8

0



+7

0





jU
CL

+5

1

a

+3

3

UJ

u

+1

31

z



UJ

-1

199

IK
UJ



U.

-3

346

U.



49

H

a

-S
-7

114



-a

76



-li

109



-13

65



-15

43



-17

13



-19

I



TOTAL No. OF DRTfl PAIRS:	1046

CORRELATION COEFFICIENT:	.998607

MEAN DIFFERENCE, ppb:	-6.53

STD. DEVIATION OF DIFFERENCES,	ppb: 4.5567

MAX. ABSOLUTE DIFFERENCE, ppb:	20

NO. ABSOLUTE DIFFERENCES >20:	1

Figure 63. Frequency distribution of differences in hourly S02 ambient
air readings between the Thermo Electron 43 and the average
of simultaneous readings from the other analyzers during the
Phase II ambient monitoring test (after manufacturer's
refurbishment—see General Comments). Thermo Electron 43
readings are not corrected for zero and span drift.

114


-------
Thermo Electron 43 Sulfur Dioxide

Manufacturer's Recommended Maintenance

Operation

Frequency

Inspect sample filter.

Inspect, clean, or replace sample capillary.
Inspect and clean fan filter.

Replace hydrocarbon cutter.

Every 6 mo
Every 6 mo
Every 18 mo

Every 30 days

Malfunctions

Refurbishment of analyzer because of noise response on 5-9-79.
General Comments

Special hydrocarbon interferent tests were run during Phase I testing
using the following compounds as potential interferents:

2,3-dimethylbutane

2-methylpentane
• 3-methylpentane

2 ,3-dimethylpentane
2,2,4-trimethylpentane

2.3.4-trimethylpentane
1-pentene

3-methylhexane
heptane
benzene

1.3.5-trimethylbenzene
l-ethyl-3-methylbenzene.

Of these compounds, only the three aromatic hydrocarbons (benzene, 1,3,5-
trimethylbenzene, and l-ethyl-3-methylbenzene) produced a response on the
analyzer even without the hydrocarbon cutter. Of these three, benzene and
1,3,5-trimethylbenzene produced a response on the analyzer equipped with
the cutter at concentrations greater than 2 ppm.

Initial evaluation of Phase II test data in April 1979 showed that the
standard deviation of the differences between the Thermo Electron analyzer
and the average of the other analyzers was larger than expected (see Figures
58 through 60). This deviation seemed to result from random variability in
the instrument response. Because this instrument had been in service for
22 months before Phase II testing began, this apparent variability problem
was attributed to aging of the instrument subassemblies. Thermo Electron
service personnel were contacted, and they agreed that aging could be the
cause. On May 9, 1979, the instrument was refurbished by factory personnel
onsite at the EPA test facility. The refurbishment included replacement of
the following:

115


-------
Thermo Electron 43 Sulfur Dioxide

Fluorescence chamber
PMT

Flash lamp

Zero and span solenoids
Permeation dryer
Hydrocarbon cutter
Optics subassembly.

Before service, the mean difference and standard deviation of the ambient
data were -0.7 ppb and 8.3 ppb, respectively. After service, the mean dif-
ference and standard deviation became +0.2 ppb and 4.1 ppb, respectively
(see Figures 61 through 63).

During 215 days of Phase II testing, data were lost for 1 day because
of malfunctions.

Phase II Testing Operator's Comments

Throughout the duration of the Phase II testing, this instrument
showed a tendency to drift in the direction of a zero or span adjustment
for approximately 24 h after adjustment. This drifting took place even
though the instrument seemed to produce a stable strip chart trace about 10
min after the adjustment. Typically, the instrument would drift about
5 percent of scale beyond the adjustment point in 24 h.

An unusual characteristic of this analyzer is that the span pot must
be reduced to increase the span response.

This analyzer achieved equilibrium on span gases in less than 5 min.

The instruction manual for this analyzer is confusing in certain
places where it describes, in detail, how to perform a certain operation
(e.g., how to clean the fluorescence chamber) but makes no mention of when
or under what circumstances to perform the operation.

116


-------
APPENDIX A

MEASUREMENT PRINCIPLES EMPLOYED BY DESIGNATED

S02 ANALYZERS

The S02 instruments currently designated operate according to one of
the following measurement principles:

Flame photometric detection
Fluorescence detection

Second derivative spectroscopic detection
Coulometric detection
Conductometric detection.

Brief descriptions of each of these measurement principles are presented
below to enhance understanding of the similarities and differences among
various designated analyzers.

FLAME PHOTOMETRIC DETECTION

This principle is based on the photometric detection of the chemilumi-
nescence (light produced by a chemical reaction) from sulfur atoms in a hy-
drogen-rich flame. The sample gas (containing the sulfur compounds) is
mixed with hydrogen and burned, producing (among other products) a molecular
sulfur species, S2. This species is raised to an elevated electronic
energy state (S2'f) by interaction with hydrogen (H*) or hydroxyl (*0H)
radicals also produced in the burning process. As the activated species
decays to its energy ground state, it releases energy in the form of light.
This light covers a broad range of wavelengths with a maximum intensity at
394 nm. The light intensity is essentially proportional to the square of
the sulfur concentration in the flame and thus is related to the volumetric
concentration of sulfur-containing molecules in the sample gas. The light
energy is sensed by a photomultiplier tube (PMT), which produces an output
current proportional to the light energy. This current, converted to a vol-

117


-------
tage and conditioned as necessary by the electronic circuits, becomes the
analyzer's output reading or signal. Characteristically, flame photometric
detectors (FPD) produce an output signal that is exponentially related to
the sulfur concentration. However, all currently designated analyzers have
electronics capable of linearizing this output. Most analyzers include this
linearizer as standard equipment, but on at least one model it must be
procured as an option.

Several important points must be considered concerning the use of
flame photometry for ambient S02 analyses. First, flame photometric detec-
tion is sensitive to all sulfur-containing molecules. For this principle
to be specific for SO2, the sample stream must be scrubbed by appropriate
means upstream of the detector to remove ambient sulfur species other than
SO2. Scrubbing is usually accomplished by either an ambient-temperature
chemical scrubber or a heated silver scrubber. These scrubbers remove
ambient hydrogen sulfide but do not remove other reduced sulfur species
such as methyl mercaptan. However, this is a minor problem because the air
at most monitoring sites rarely contains significant quantities of reduced
sulfur compounds other than H2S.

One prevalent problem with heated silver scrubbers is that the scrubber
can generate both positive and negative interferences in the presence of
ambient ozone. The positive interferences may occur according to the fol-
lowing mechanism. Ambient hydrogen sulfide is removed from the sample
stream by reaction with the silver, forming silver sulfide according to the
equation:

H2S + 2Ag -* Ag2S + H2 .

In the presence of—or on subsequent exposure to—ozone, however, the sil-
ver sulfide probably continues to react where:

3Ag2S + 203 6Ag + 3S02
or

Ag2S + 03-* Ag«>0 + S02 .

Either way, the H2S entering the scrubber is eventually converted to S02

and detected as a positive interference.

118


-------
The negative interference in the presence of O3 probably follows a
more obscure mechanism. The result, nevertheless, is the removal of SO2
from the sample stream, perhaps by conversion (in the presence of the
silver) to the sulfate ion occurring in the form of sulfuric acid. Both
the positive and negative interferences in the presence of ozone can be
minimized by decomposing the ozone in the sample stream prior to its reach-
ing the silver. This process is most often accomplished thermally, but one
presently designated method uses ethylene to assist in decomposition.

A second point to consider when FPD is used is that the technique is
subject to quenching of the chemiluminescence by common ambient gases such
as oxygen and carbon dioxide. Therefore, it is important to maintain rea-
sonably constant concentrations of these gases between the calibration
matrix and the sample matrix. Difference in oxygen concentrations between
the two matrices can become significant if a low-level S02~in-nitrogen
cylinder gas is diluted to prepare a calibration standard. In such a case,
the nitrogen in the pollutant flow stream may "dilute" the oxygen in the
dilution air stream, significantly decreasing oxygen concentration. This
situation can be avoided by keeping the concentration of S02 in the cylinder
gas high enough that the nitrogen contributed by the pollutant flow stream
is insignificant with respect to the total flow volume.

Differences in C02 concentrations between calibration and sample
matrices become significant when sources of zero air that contain no C02
(ambient C02 concentration is at least 320 ppm) are used for calibration.
This often occurs because many systems used for preparing zero air remove
CO2. Also, most cylinders of compressed synthetic air sold as zero-grade
air contain no C02. This problem can be avoided by using zero air systems
that specifically pass C02 and by supplying their intakes with outdoor
ambient air. Zero-grade compressed air, which contains approximately
320 ppm CO2, may also be ordered from commercial gas suppliers.

One further point to remember when FPD is employed for SO2 analyses is
that FPD analyzers require a source of pure, dry hydrogen. This requirement
involves extra effort for the handling and maintenance of support gas
cylinders or hydrogen generators. The use and handling of hydrogen also
involves additional safety precautions because this gas is flammable or
explosive when mixed with air.

119


-------
FLUORESCENCE DETECTION

This principle is based on detection of the characteristic fluorescence
released by the sulfur dioxide molecule when it is irradiated by ultraviolet
light. This fluorescent light is also in the ultraviolet region of the
spectrum, but at a different wavelength than the incident radiation. The
fluorescent wavelengths usually monitored are between 190 and 230 nm. In
this region of the spectrum, there is relatively little quenching of the
fluorescence by other molecules occurring in ambient air. As in flame
photometry, the light is detected by a PMT that, through the use of elec-
tronics, produces a voltage proportional to the light intensity and S02
concentration.

The fluorescent light reaching the PMT is usually modulated to facili-
tate the high degree of amplification necessary. Some analyzers mechani-
cally "chop" the incident irradiation before it enters the reaction chamber.
This process is accomplished by a fan-blade-like chopper rotating at a
constant speed, which alternately blocks uid passes the 1ight to the chamber.
Other instruments electronically pulse the incident light source at a
constant rate.

Potential interferences to the fluorescence technique include any
species that either quenches or exhibits fluorescence. Both water vapor
and oxygen strongly quench the fluorescence of SO2 at some wavelengths.

Where water vapor presents a problem, it can be removed by a dryer within
the instrument. In most analyzers, the water interference is minimized by
careful selection of the incident radiation wavelength. The effect of
oxygen quenching can be minimized by maintaining identical oxygen concentra-
tions in the calibration and sample matrices, as discussed with respect to
flame photometry.

Aromatic hydrocarbons such as naphthalene exhibit strong fluorescence
in the same spectral regions as S02 and are major interferents. These aro-
matics must be removed from the sample gas stream by an appropriate scrub-
ber upstream of the reaction chamber. The scrubbers may operate at ambient
or elevated temperature. Certain elevated-temperature scrubbers, however,
have the potential for converting ambient hydrogen sulfide (which normally
does not interfere with the fluorescence technique) into SO2. In these
cases, the hydrocarbon scrubber must be preceded by a scrubber for ii2S.

120


-------
SECOND DERIVATIVE SPECTROSCOPIC DETECTION

Many gases absorb light at various wavelengths. In particular, gases
often absorb light in narrow, specific intervals of wavelength. This means
that a gas may absorb strongly at a certain wavelength but may exhibit less
absorption just a few wavenumbers to either side of that wavelength. These
narrow intervals, called absorption bands, are characteristic of a specific
gas, and the sharpness of a specific absorption peak is used as the basis
of second derivative spectroscopy.

In conventional spectroscopy, light of a fixed wavelength is passed
thiough a gas and detected by a PMT. The reduction in light intensity be-
cause of absorption at that fixed wavelength is exponentially related to
the concentration of the gas according to Beer's Law. In the case of an
ambient air sample, other gases may be present that also absorb at the
fixed wavelength and therefore interfere with the measurement. However, if
a wavelength can be found at which the gas of interest has a very sharp,
narrow absorption peak, while other gases that might be present exhibit a
broad, widely distributed absorption, second derivative spectroscopy can be
used to measure the gas of interest with minimal interference. This is
possible because second derivative spectroscopy measures the "sharpness"
(i.e., the second derivative) rat* .tan the magnitude of the peak.

The absorption peak sharpness is measured by selecting a wavelength
centered in a narrow absorption band of the gas of interest—S02—in the
near ultraviolet spectral region. The specific absorption band selected
must have a strong absorption peak and be free of overlapping sharp absorp-
tion peaks from other gases likely to be present in the sample. The second
derivative spectrometer has a way to vary the wavelength periodically
slightly above and slightly below the center wavelength so higher and lower
wavelengths fall on the sides of the absorption peak, well beiow the maximum.
The instrument compares the absorption at the center wavelength to the ab-
sorption at the two side wavelengths, which is a measure of the curvature
of the absorption peak. Because the other absorbing gases have broadband
absorption, their contribution to the absorption is almost identical at the
three wavelengths; hence, they have little effect on the curvature measure-
ment. The sharpness of the absorption peak is directly proportional to the

121


-------
height of the peak, which is, iri turn, related to the gas concentration of
the S02 according to Beer's Law. Hence, the concentration measurement is
highly specific for S02 without the use of chemical scrubbers or other sam-
ple conditioning components. Suitable electronic circuits in the analyzer
control the wavelength variation, make the absorption measurements, compute
the curvature, and provide an output signal representative of the concentra-
tion measurement.

A point to remember when using this technique is that the incident
light must pass through a "long" sample path to achieve sufficient sensi-
tivity. A typical pathlength of 12 meters is obtained with a "folded path"
cell, where the light is bounced back and forth several times through a
shorter cell. This necessitates the use of a precise, ultraviolet optical
system, which has significant optical losses and is often difficult to
align and keep aligned. In addition, optical systems exposed to ambient
air may need frequent cleaning. Finally, the optical system necessary to
obtain and vary the narrow, accurate wavelength band is often delicate and
difficult to adjust.

COULOMETRIC TITRATION DETECTION

The principle of continuous coulometric titration depends on maintain-
ing an equilibrium between ionic species in an electrolyte so the redox
(reduction-oxidation) potential between two electrodes in the electrolyte
remains constant. For S02 monitoring, the electrolyte is usually dilute
sulfuric acid (H2S04) and the species in equilibrium are usually free
bromine (Br2) and the bromide ion (Br ).

In an acid medium, free bromine and the bromide ion interact to estab-
lish an equilibrium according to the equation:

2Br~ I Br2 + 2e~ .

As lon6 as the temperature of this medium remains constant, the Nernst
equation predicts that the equilibrium potential produced by this interac-
tion will remain constant. When S02 is introduced into this equilibrium by
bubbling the sample gas stream through the H2S04 solution, free bromine is
reduced to the bromide ion according to the equation:

S02 + Br2 + 2H20 ~ H2S04 + 2Br~ + 2H+ .

122


-------
This reaction upsets the previously established equilibrium, resulting in
changed potential across the electrodes.

When this change in potential is sensed by the aralyzer's electronics,
a current is applied to a second set of electrodes in the medium. This
current causes the generation of free bromine from the bromide ion and
tends to reestablish the original equilibrium, which reestablishes the
original equilibrium potential. The current necessary to reestablish the
equilibrium is linearly proportional to the mass of bromine consumed, which
is, in turn, linearly proportional to the mass of SO2 entering the reaction.
This current is converted electronically into a suitable voltage that
becomes the analyzer's concentration measurement output signal.

Any species in the sample gas stream that has the ability to reduce
bromine chemically (or, conversely, be oxidized by bromine) represents an
interference to the measurement principle. In fact, any species that can
upset the bromine/bromide equilibrium, such as through consumption or aug-
mentation of the bromide ion, is a potential interferent. Careful, selec-
tive scrubbing of the sample gas stream is necessary to avoid such inter-
ference .

Analyzers using the coulometric technique require periodic replenish-
ment or replacement of the electrolyte. The entire measuring chamber, and
particularly the electrodes, must be kept scrupulously clean. Stable flow
control is essential for mass-sensitive analyzers. Flow control devices
(e.g., orifices and valves) may require substantial maintenance when used
in wet-chemical systems where they are subject to continuous contamination
and degradation by corrosive chemicals used in the system.

CONDUCTIMETRIC DETECTION

This principle is based on increased electrical conductivity of an
oxidizing, acid-measuring solution when it absorbs SO2• The only currently
designated instrument that uses this principle (the ASARCO 500) employs a
dilute solution of sulfuric acid containing a small amount of hydrogen
peroxide as the oxidizing agent.

As ambient air (containing S02) is bubbled through the measuring solu-
tion, the S02 is oxidized to the sulfate ion by the action of the hydrogen
peroxide according to the equation:

123


-------
S02 + H202 -»¦ S04 + 2H+ .

The formation of additional sulfate ions increases the conductivity of the
solution. Two platinum electrodes, placed in the solution and connected to
a balanced bridge circuit in a recorder, record this increased conductivity.
Because the conductivity of any solution is temperature dependent, provi-
sions are made for temperature compensation through the use of a tempera-
ture compensator immersed in the solution.

Any gas present in the atmosphere that will ionize in an aqueous medi-
um (e.g., dilute sulfuric acid) will increase the conductivity of the
measuring solution used in this principle. Such an increase represents a
positive interference in the measurement of S02. Nitrogen dioxide and
carbon dioxide are examples of these gases, which are routinely present in
ambient air. Ammonia and hydrogen chloride will also ionize in solution
but are not so prevalent in the atmosphere. Also, any species present in
the atmosphere that can decrease solution conductivity (perhaps by chemi-
cally complexing previously ionized species) represents a negative inter-
ference. To prevent such interferences, the sample gas must be chemically
scrubbed of all such species before it is introduced into the measuring
solution.

Although a conductometric analyzer could be designed with a dynamic,
continuously flowing measurement cell, the ASARCO uses a batch operation
with a fixed volume of static electrolyte. Because the conductivity of the
static electrolyte increases in proportion to the S02 concentration and re-
mains constant when the S02 concentration is zero, the result is that the
measured conductivity represents the accumulated (integrated) S02 concen-
tration. Periodically, the cell is (automatically) emptied and refilled
with fresh electrolyte to start a new cycle. The instantaneous S02 concen-
tration can be derived by differentiating (finding the slope with respect
to time) of the conductivity measurement. However, because hourly average
S02 concentrations are normally needed, the integrated S02 reading repre-
sented by the maximum conductivity reading for each cycle (obtained just
prior to refilling) is conveniently used to obtain the hourly average (two
cycles per hour). A special dynamic calibration procedure (described in
the analyzer's operation manual) is used to accommodate the cyclic, batch
operation.

124


-------