-------
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
. REPORT NO.
EPA/600/4-85/012
4. TITLE AND SUBTITLE
Field Experience with Four Portable VOC Monitors
5. REPORT DATE
February 1985
6. PERFORMING ORGANIZATION CODE
'. AUTHOR(S)
Robert A. Ressl & Thomas C. Ponder, Jr.
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
PEI Associates, Inc.
1006 N. Eowen Road
Arlington, TX 76012
10. PROGRAM ELEMENT NO.
11. CONTRACT/GRANT NO.
68-02-3767
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Monitoring Systems Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
13. TYPE OF REPORT AND PERIOD COVERED
Final 2/84 - 7/84
14. SPONSORING AGENCY CODE
EPA 600/08
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This report discusses the field operation problems associated with use of
four portable volatile organic compound (VOC) detection instruments in con-
ducting Reference Method 21 VOC screenings. The report presents the results
of the field trials and summarizes the ease of use of each instrument. Informa-
tion on operational problems and recommendations are provided. Also included
are discussions of the features that would make all portable instruments more
reliable, durable, or convenient to use. Based on the data collected for this
study, three of the instruments report similar leak rates in the facility where
they were used.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
COSATi Field/Group
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (Tins Report/
UNCI.ASSIFTFn
21. NO. OF PAGES
66
20. SECURITY CLASS
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (R«». 4-77) PREVIOUS EDITION is OBSOLETE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
-------
ACKNOWLEDGMENT
Special thanks is given to Mr. Roosevelt Rollins, the U.S. Environmental
Protection Agency Project Officer, for his support and input in the prepara-
tion of this report. Thanks is also given to the EPA for providing the AID,
TLV, and HNu instruments for use in the field trials and to the four plants
that allowed PEI to enter their facilities to use the instruments.
-------
ABSTRACT
This report discusses the field operation problems associated with use of
four portable volatile organic compound (VOC) detection instruments, in con-
ducting Reference Method 21 VOC screenings. The report presents the results
of the field trials and summarizes the ease of use of each instrument. Infor-
mation on operational problems and recommendations are provided. Also in-
cluded are discussions of the features that would make all portable instru-
ments more reliable, durable, or convenient to use. Based on the data col-
lected for this study, three of the instruments report similar leak rates in
the facility where they were used.
This report was submitted in fulfillment of Contract No. 68-02-3767 by
PEI Associates, Inc. under the sponsorship of the U.S. Environmental Protec-
tion Agency. This report covers a period from February 1, 1984, to July 31,
1984; work was completed as of July 31, 1984.
IV
-------
CONTENTS
Disclaimer ii
Acknowledgment iii
Abstract iv
Figures vi
Tables vii
1. Introduction 1
2. Field Trial Methods 6
3. Summary of Evaluations 12
4. Desirable Instrument Features and Recommendations for
Future Studies 19
Appendices
A. Description of Instruments 22
Appendix A References 42
B. Summary of Field Screening Sheets at Four Sites 43
C. Summary of Comments from Field Evaluation Sheets 49
-------
FIGURES
Number Page
1 Portable Organic Vapor Analyzer OVA-108 23
2 Analytical Instrument Development, Inc., Model 712 28
3 Controls located on the top of the AID side pack 29
4 Controls located on the bottom of the AID side pack 30
5 TLV 34
6 Front of the TLV 35
7 HNu Systems, Inc., PI-101 38
8 HNu controls and readout 40
VI
-------
TABLES
Number Page
1 Reference Method 21 performance criteria for four instruments
based on manufacturers' data 3
2 Summary of operating problems 13
3 Approximate response factors for the OVA of various
hydrocarbons relative to methane 26
4 Approximate response factors for the OVA of various
organic compounds relative to methane 27
5 Approximate response factors for the AID of various
VOCs relative to methane 32
6 Multiplying factors for converting ppm meter readings of
hexane-calibrated instruments to ppm concentrations of
other gases (approximations) 37
7 Summary of sources screening with one or more readings
over 10,000 PPMV, Site 1 43
8 Summary of sources screening with one or more readings
over 10,000 PPMV evaluated Kay 10, 1984, Site 2 45
9 Summary of sources screening with one or more readings
over 10,000 PPMV evaluated May 30, 1984, Site 3 46
10 Summary of sources screening with one or more readings
over 10,000 PPMV evaluated June 21, 1984, Site 4 48
11 Summary of comments from daily instrument critiques 50
12 Summery of site critiques 55
-------
SECTION 1
INTRODUCTION
The U.S. Environmental Protection Agency (EPA) has issued performance
standards and guidelines to limit emissions of volatile organic compounds
(VOC) from several stationary source categories. These industries such as
petroleum refineries, synthetic organic chemical plants, and natural gas
processing plants emit significant quantities of VOCs from sources other than
classical point sources into the workplace and surrounding atmosphere. These
fugitive VOC emissions occur from valves, pumps, drains, pressure relief
devices, etc. In order to reduce these fugitive emissions, EPA has promul-
gated regulations that require periodic monitoring of the potential leaking
sources in order to identify leaking sources so they can be repaired. Spe-
cialized instruments are required since the leaking VOCs cannot normally be
detected without them.
As described in 40 CFR 60, Appendix A, Reference Method 21 (RM 21), De-
termination of Volatile Organic Compound Leaks, there are technically feasible
devices suitable for monitoring fugitive VOC leaks. These devices can be
placed near possible points of emissions and will respond to releases of the
organic compounds. Specific instruments suitable for this purpose include,
but are not limited to, catalytic oxidation, flame ionization, infrared
absorption, and photoionization detectors.
EPA has prepared a Technical Assistance Document (TAD) titled "The Use of
Portable Volatile Organic Compound Analyzers for Leak Detection." The Decem-
ber 1983 report was prepared by Ralph M. Riggin of Battelle Columbus Labora-
tories for Roosevelt Rollins of the EPA under Contract No. 68-02-3487 (WA-18).
The TAD describes in detail the four types of detectors and typical instru-
ments. Additionally, it contains information on manufacturers of these in-
struments. Another useful document is titled "Summary of Available Portable
VOC Detection Instruments," EPA-340/1-80-010, March 1980. This document also
contains lists of manufacturers of portable VOC detection instruments.
-------
Subsequent field use of portable VOC detectors has disclosed some instru-
ment-specific problems such as undetected flame-outs, plugged orifices from
dirt, high background readings due to chemical absorption on probe and tubing
surfaces, high humidity effects, varying or lack of response, long response
times, and calibration drift. The EPA, through its contractor, Research Tri-
angle Institute, proposed that four instruments undergo field use to identify
and document instrument-specific problems of the instruments used to conduct
screenings following RM 21. The four instruments chosen do not represent all
the available instrument types. Rather, they were selected for evaluation
because they were in common use. Other types and brands of instruments are
available, and the conclusions and recommendations in this report may not be
applicable to all of them.
This report is to document the field operational problems encountered
while using the four available instruments. The instruments were selected for
use on the basis of their availability and because two of them have been fre-
quently encountered in field use by PEL A comparison of the manufacturers'
specifications was made with the RM 21 test requirements and is shown in
Table 1. Although some of the instruments did not appear to meet the RM 21
requirements, all the instruments were used and their field experience re-
ported. No attempt was made to make a rigorous evaluation of each instrument
or of RM 21. However, where possible, the requirements of RM 21 were met.
Under contract to the U.S. Environmental Protection Agency (EPA) and
private clients, PEI has routinely conducted VOC screenings using the Foxbcro*
Century Systems Portable Organic Vapor Analyzer Model OVA-108 (OVA), following
RM 21. The screenings have been required as part of various state and federal
regulations. (Regulations encountered include 40 CFR 60 Subparts VV and GGG,
Santa Barbara County Air Pollution Control District Rule 331, Bay Area Air
Quality District Regulations Rules 18 and 25, and Louisiana Petroleum Refinery
Fugitive Emissions Control Regulations.) Although various VOC screening in-
struments were used to develop the background information for the EPA stand-
ards and RM 21, the operational problems of using those instruments
*Registered trademark.
-------
TABLE 1. REFERENCE METHOD 21 PERFORMANCE
CRITERIA FOR FOUR INSTRUMENTS BASED ON MANUFACTURERS' DATA
Reference
Method 21
section
3.1. la)
3.1.1b)
3.1.1c)
3.1. Id)
3.1.1e)
3.1.2a)
3.1.25)
3.1.2c)
Criteria
Responds to compounds being processed
Capable of measuring the leak defini-
tion concentration specified in the
regulation
Capable of scale reading to ±5% of the
specified leak definition
Equipped with a pump for continuous
flow of 1/2 to 3 1pm
Intrinsically safe
A response factor for the individual
compounds to be measured must be less
than 10T
The instrument response time must be
less than 30 seconds
The calibration precision must be _<10%
of the calibration gas value
Acceptability
by instrument
HNu
Yesa
Nob
Yesd
Yes
Yes
No9
Yes
(5
sec)
Yes,
(
-------
TABLE 1. (continued)
dThe HNu maximum scale reading is 2,000 ppm. However, scale divisions are
100 ppm; therefore, it is readable to 5% of a 10,000 ppm standard.
eThe OVA scale is logarithmic. At the upper end of the scale, the divisions
are greater than 500 ppm. However, at lower readings, the scale divisions
are less than 500 ppm.
fThe response factor varied for each plant where the evaluations occurred,
and a yes response indicates that in most of the plants the instrument
response factors were less than 10. However, in one of the plants the
response factors were greater than 10.
HNu is not capable of measuring a 10,000 ppm concentration without dilut-
ing the sample, and there is no known dilution system for field use with the
HNu. Available information indicates the response factor for the HNu for
methane, ethane, and propane is greater than 10.
Assumes the same response factors as for the OVA since both instruments use
the same type of detector. However, response factors for the AID were not
determined.
Based on use of the 48 ppm benzene calibration gas. No other calibration
gases were available.
-------
were not published. Therefore, this task was established to perform field
screenings using four different available VOC screening instruments and
document and report on their ease of use and the operational problems that
occur during use.
The task was designed to conduct screenings under field conditions.
Therefore, PEI obtained permission from facilities where we were working with
our primary instrument, the OVA, to use the instruments listed below for VOC
screening:
Foxboro Century Systems Portable Organic Vapor Analyzer Model
OVA-108;
United Technology's Bacharach Instruments Model TLV Sniffer;*
Analytical Instruments Development, Inc., AID Model 712;* and
HNu Systems, Inc., Model PI-101.*
Brief descriptions of these instruments are contained in Appendix A.
The purpose of this report is to provide basic information on each
instrument, how they were used, what operational problems were encountered
during and between screenings, and the ease of use of each instrument in rela-
tionship to the other three. As part of the field use of each instrument, PEI
collected the screening data on each instrument's response to various leaks in
the facilities and compared the results. These data are summarized in Appen-
dix B. The data were collected using the four instruments and following RM 21
as closely as possible. However, because of equipment problems (AID and OVA
failures at some sites), the data are of limited usefulness in showing that
the screening results are the same regardless of the instrument used.
In the course of the study, special attention was given to documenting
operational problems and ease of use. Appendix C contains summaries of the
daily and site instrument critique sheets used to record the instrument prob-
lems.
*Registered trademark.
-------
SECTION 2
FIELD TRIAL METHODS
The instruments used represent three types of detectors. The OVA and
Analytical Instruments Development, Inc., Model 712 (AID) are flame ionization
detectors (FID). FIDs draw a gas sample into a combustion chamber where it is
burned in a hydrogen flame. The combustion products produce charged ions.
The number of charged ions and the type of chemical in the sample stream are
empirically proportional to the concentration of the chemical in the sample
stream. The charged ions cause a current flow in the detector where the
number of flowing ions is proportional to the VOC concentration in the sample
stream for the particular VOC under investigation. United Technology's
Bacharach Instruments, Inc., Model TLV Sniffer (TLV) uses catalytic oxidation
(such as a platinum-coated wire) and a Wheatstone bridge to detect resistance
changes in the catalytic oxidizer electrode. The catalytic oxidizer causes
the VOC (or other combustible) to burn, and the resulting released heat raises
the temperature of the electrode. As the temperature is raised, the
resistance in that portion of the Wheatstone bridge increases. As a result of
the resistance change, the bridge circuit produces a signal proportional to
the concentration of the combustible in the sample stream. The HNu Systems,
Inc., Model PI-101 (HNu) uses a photoionization detector (PID). In this sys-
tem, ultraviolet light is generated and strikes the organic molecules in the
sample stream. When the ultraviolet light is absorbed, an ion is emitted from
the molecule and collected on a detector. The number of ions collected is
proportional to the VOC concentration. Each of these instrument types has its
own unique characteristics. Appendix A presents more details taken from the
manufacturers' operating manuals on the OVA, AID, TLV, and HNu. Other infor-
mation on portable VOC instruments is contained in the report titled "The Use
of Portable Volatile Organic Compound Analyzers for Leak Detection," December
1983, and prepared by R. M. Riggin of Battelle Columbus Laboratories.
-------
There are other models and brands of VOC screening instruments available
besides the four used in the field trials. More information on available VOC
screening instruments may be found in two EPA documents: "Summary of Avail-
able Portable VOC Detection Instruments," EPA-340/1-80-010, March 1980; and
"Evaluation of Potential VOC Screening Instruments," EPA-600/7-82-062, Novem-
ber 1982.
METHODS OF EVALUATION
The first step in the evaluation was to use the information reported by
the manufacturer and compare each instrument's performance with the perform-
ance requirements in 40 CFR 60, Appendix A, Reference Method 21, Section 3.
The comparison was used not to disqualify any instrument from use but to pro-
vide information on how the instrument could be used. The intent of the field
trials was to develop information on the ease of use of these four instruments
and to report on any operating problems encountered during their use. The
evaluation is summarized in Table 1 and is based on information contained in
the manufacturers' literature and other published sources.
While preparing Table 1, the following three points were raised:
1. The response factors for the TLV and OVA are published in two EPA
documents. Those response factors differ from those published by
the manufacturers. Since the OVA and AID operate on the same
principles it is assumed that they have the same response factors.
However, even though the response factors published by the manu-
facturers of the OVA and AID are similar, they are not the same. No
attempt was made to verify the manufacturers' information or to
measure response factors for any of the instruments. We did not
find any published response factors for the HNu; however, the
manufacturer did publish relative sensitivity values which are
related to response factors but cannot be directly converted to
them. Most of the published response factors are for pure chemicals
which may not be correct for use with various chemical mixtures
encountered during screening. Additionally, the mixtures encounter-
ed in screening are poorly defined and change as the process stream
moves through the facility, making determination of an appropriate
response factor difficult.
2. The manufacture of the HNu suggests a method of detecting and mea-
suring high concentration samples using dilution. Although a dilu-
tion probe is available for use with the OVA, no attempt was made to
dilute samples or use any dilution probes. Previous experience with
-------
the dilution probe on the OVA found it unsatisfactory since the cal-
ibration was unreliable, and satisfactory screening results could be
obtained without using it on the OVA. PEI is unaware of any commer-
cially available dilution system for the HNu that can be used while
conducting RM 21 screenings.
3. The response factor is not normally used in determining whether or
not a source is leaking. It is simply a measure of an instrument's
responsiveness to the compounds in question and must be less than 10
to meet RM 21 requirements. At one plant the response factors for
the compounds in the screening area were greater than 10. However,
the instruments (OVA and TLV) would response to the compounds
(published response factors for the OVA and TLV were approximately
20 and 30, respectively, for one of the compounds). Therefore,
another instrument reading was chosen as the action level, and an
instrument response of that value or greater was considered as a
leak. The chosen value was selected to equate to a 10,000 ppm
concentration of the compound being screened. For a compound with a
response factor of 20, this equates to a 500 ppm reading on the
instrument. If a compound had a response factor of 9.0, it would
meet the requirement of RM 21, Section 3.1.2a. However, its actual
concentration in order to achieve a 10,000 ppm reading on an instru-
ment calibrated with methane as required in the Petroleum or SOCMI
regulations would have to be 90,000 ppmv. At this concentration,
many VOCs would flame out an OVA before giving a 10,000 ppm response
and would not be considered as a leaking source. Therefore, the
selection of an actual concentration of 10,000 ppmv is considered
conservative.
Although the above points have possible serious implications about the
applicability of the instruments' usefulness in conducting RM 21 screenings
for compliance with appropriate regulations, they do not interfere with
providing enough information to satisfy the project's intent. Therefore,
these points were basically ignored, and the instruments were used in order to
determine the information necessary to complete the field trials and report on
the problems and ease of use of the instruments in the field. The above three
points are considered part of the problems with the instruments' field use.
Only one of the instruments, the HNu, had a negative response to any of
the performance criteria. However, the OVA is equipped with a logarithmic
scale and does not have scale divisions of 500 ppm or less at the upper end of
the scale (it does have scale division of less than 500 ppm at the lower end
of the scale), so it technically does not meet the requirements of RM 21.
The next step was to formulate a method of evaluating the instruments.
Because the chief goals of the study were to document and evaluate ease of use
8
-------
and operational problems of these instruments, it was obviously necessary to
use each instrument as much as possible. However, since the plants were being
routinely screened with the OVA, PEI continued to use the OVA to screen the
entire plant. Because of the time required to screen the whole facility only
a part of the facility was used for the field trials. Operating time on the
other instruments was, therefore, not equal to that on the OVA.
The OVA, TLV, and AID instruments were calibrated using a 9,970 ppmv
methane-in-air certified calibration gas. Ambient air was used as zero air.
The calibration gas was taken from a compressed gas cylinder into an inert
bag. The bag was triple purged with the calibration gas before each use. The
calibration gas in the bag was then used to calibrate the instruments. The
bag sample was repeatedly attached and detached to the sample probe inlet
until the instrument reading was stable at the calibration gas value without
having to readjust the calibration dial. The meter reading was recorded on a
calibration log. The calibration procedure was performed before and after
each day's screening. Where appropriate, the calibration was checked at
midday. No changes were made in the instrument settings and the instruments
were not shut off until final calibration values were recorded.
With the HNu none of the lamps can be calibrated to a 10,000 ppm methane
or hexane standard as supplied; the instrument is not sensitive to methane and
a dilution system must be used to calibrate to hexane. Since no dilution
system was available, none was used, and the instrument instead was calibrated
with a 48 ppm benzene standard to verify the system was responding before it
was used in the field trials. All three lamps were used.
RM 21 allows calibration using either methane or hexane. Methane in air
was used rather than hexane in air because of the difficulty in obtaining a
10,000 ppm hexane in air standard and the reduced hazards (10,000 ppm methane
in air is not explosive while 10,000 ppm hexane in air is near the lower
explosion limit for hexane). A compressed gas cylinder containing 10,000 ppm
hexane in air is near saturation. Therefore, obtaining reproducible concen-
trations of a calibration standard from a compressed gas cylinder is difficult
because some hexane condenses out inside the cylinder and changes the volume-
tric concentration of the hexane in the air.
-------
We conducted two types of operational tests with the equipment, a relia-
bility check and field operations. In the reliability check, the instruments
were charged and allowed to operate for several 8-hour periods. This test
provided some assurance that the instruments could operate over an 8-hour pe-
riod. In the field operations, all four instruments were used to screen a
series of 200 to 300 sources. Following the screening, the instrument opera-
tor completed an evaluation sheet, reported any problems with the instrument,
and provided general subjective comments on the instrument as a leak detection
tool.
The log sheets collected during the evaluations are summarized in Appen-
dix C.
EVALUATION SITES
The first of the four plants screened, Site 1, is a small, integrated
refinery operation with approximately 3,000 sources subject to screening. For
the instrument field trials, approximately 200 sources were selected, and each
of the four instruments was used to screen those sources. Because of the dif-
ficulty of repeatedly locating the various sources, there were instances
during the screening process when some sources were not screened with all of
the instruments. This occurrence is reflected in the summaries contained in
Appendix B. We anticipated that the leaking materials would be mostly meth-
anes, ethanes, and propanes and suggest that this is the reason we obtained
low responses on the HNu. The HNu did respond in a few cases to some leaks;
however, the responses obtained on the HNu were all approximately 100 ppm,
well below the leak level definition.
The second facility screened, Site 2, manufactures fluorocarbons used for
refrigerants. The screening was limited to one of the three processing areas
making fluorinated hydrocarbons F-ll and F-12. The screening was conducted
predominantly in the reactor area; however, a number of sources from the
purification process, including F-ll blend and weigh tanks, were screened.
The only two VOCs that were.present in the test area were carbon tetrachloride
(a feedstock) and F-ll. Approximately 150 sources were chosen for the evalua-
tion.
10
-------
The AID was not available for use at Site 2. Prior to the screening, the
AID was operating properly. However, the instrument failed to ignite at the
plant. The testing staff made numerous attempts to adjust the instrument in
the field, including recharging the instrument with hydrogen and removing the
detector and cleaning it, but all attempts failed. The instrument was eventu-
ally returned to the manufacturer, who found and repaired a hydrogen leak and
readjusted the hydrogen flow rate.
The third facility screened, Site 3, is an integrated refinery with
approximately 14,000 sources. The screening operation took place in one
portion of the refinery where light hydrocarbons are combined to form heavier
hydrocarbons. Most of the leaking materials in this area are octane and
lighter aliphatics. Approximately 300 sources were chosen for evlauation in
this facility. Only the AID, HNu, and TLV were used on the day of the evalua-
tion because the OVA failed to operate properly. The OVA was subsequently
returned to a manufacturer's representative who replaced the preamplifier.
The HNu did not respond to any of the leaks within the test area.
Site 4 is a small natural gas processing plant. The plant receives sour
natural gas from an offshore field. The gas passes through a slug separator
to remove liquids then through an amine sweetening process. Some liquids are
removed but only natural gas was present in the portion of the plant where the
instrument field trials were conducted. Approximately 200 sources were chosen
for evaluation. Only three of the instruments were evaluated since the HNu
does not respond to methane, the only leaking substance in the evaluation
area. Although methane is an exempt hydrocarbon, Santa Barbara requires a RM
21 leak testing program. The proposed NSPS regulations for new natural gas
processing facilities do not address the fact that most leaks are methane or
ethane.
11
-------
SECTION 3
SUMMARY OF EVALUATIONS
Of the four instruments used in the field trials, the AID, HNu, and TLV
were new or nearly-new instruments. The OVA is approximately 4 years old and
has been used extensively for VOC screenings in many plants. It has been re-
conditioned once by Foxboro and repaired several times by various Foxboro man-
ufacturers' representatives; in addition, field maintenance has occasionally
been performed on the instrument. Prior to the field trials, the instrument
had been in routine screening use in a refinery in Louisiana.
The field trials were designed to provide subjective comments on the in-
struments' performance. Table 2 summarizes the comments developed from the
screenings.
The operator logs reveal some interesting information about the useabil-
ity of the instruments. (These logs are summarized in Appendix C.) For exam-
ple, the AID is equipped with a liquid crystal digital readout. This device
has a fairly narrow viewing angle, and it was frequently impossible to observe
the readout while moving the probe tip all around the source as required for
proper screening. To compensate for this difficulty, the operator used a
short, flexible extension made of plastic tubing on the end of the probe. The
operator could then manipulate the flexible probe tip around the sample inter-
face with one hand and still maintain a satisfactory viewing angle of the dis-
play. This modification, however, made the AID more difficult to use because
both hands were required to control the instrument probe.
All the instruments except the HNu were equipped with some kind of probe
filter, and these filters were always installed. However, we added to all the
instruments a 2-in. length of plastic tubing loosely stuffed with glass wool
to protect the probe from contamination by the greases and oils present at
many sources. The instruments' responses were the same with and without the
plastic tubing in place. We developed this modification during other screen-
ings and incorporated it for this study as a matter of course. The tubing
12
-------
TABLE 2. SUMMARY OF OPERATING PROBLEMS
Item
OVA
TLV
HNu
AID
Carrying strap
Battery
Battery charger
Instrument readout
Calibration knob
or zero/span
adjustment
The best arrangement of
the instruments evalu-
ated.
Acceptable during
period of evaluation.
Acceptable during
period of evaluation.
The analog readout
with logarithmic scale
was conveniently lo-
cated in the probe and
very easy to use.
The knob could not be
secured. However,
since it was located on
the control module,
which had a cover, it
did not require secur-
ing.
(continued)
No strap; instrument
was carried by a handle
that was sometimes
inconvenient.
Acceptable during
period of evaluation.
Acceptable during
period of evaluation.
The readout in the con-
trol unit was less con-
venient to read than on
the probe but was
acceptable. It
required frequent scale
changes that were some-
what inconvenient.
The zero adjust knob
(only adjustment) was
located on the control
module and could not be
secured. It was easily
and frequently bumped,
requiring re-zeroing of
the instrument, until
it was secured with
tape.
The strap was very nar-
row and after an hour
of carrying was quite
uncomfortable.
Acceptable during
period of evaluation.
Acceptable during
period of evaluation.
The readout in the con-
trol unit was less con-
venient to read than on
the probe but was
acceptable. It
required frequent scale
changes that were some-
what inconvenient.
The zero knob was some-
what protected and was
quite stiff to turn.
It is located on the
control module and did
not require securing.
The strap was unpadded
and, although reasonably
wide, the edge of the
strap became very uncom-
fortable after an hour
of carrying.
Acceptable during
period of evaluation.
Acceptable during
period of evaluation.
The digital readout was
difficult to read from
an angle, and the fre-
quency with which it had
to be updated made selec-
tion of a reading value
difficult.
The calibration (zero
and span) require a
screwdriver to adjust.
The response and level
knobs had locks to
secure them. All were
acceptable.
-------
TABLE 2. (continued)
Item
OVA
TLV
HNu
AID
On/off and other
controls
Sample line and
instrument
umbi1ical
Probe contamina-
tion
(continued)
The controls are on the
control module. The
instrument and pump
switches are easily
moved (newer models
have locking toggles).
The handles on the
hydrogen supply are
too short (newer
models have longer
ones). The gas select
knob was not used
since span gases were
used for calibration.
The sample line tends
to kink after long use
when the protective
sleeve slips. The line
could be longer. The
electrical connector at
the control module has
been weakened and has
shorted.
Since all the plants
had some sources where
the probe could get
dirty, all units were
affixed with a 2-in.
long piece of Tygon
tubing with a glass
wool plug as a primary
filter.
The on/off/standby,
battery, operate, and
range switch caused
no problems.
The sample hose could
be longer. During the
evaluation period, the
hose developed a kink
and would frequently
pinch off, causing the
pump to stall and the
instrument to operate
improperly.
See OVA comments.
The controls were
acceptable.
The umbilical was too
short.
See OVA comments.
The alarm, on/off switch,
and the battery/AC/charge
switch were frequently
confused, which resulted
in turning off the
instrument instead of the
audible alarm on several
occasions.
The umbilical was too
short.
See OVA comments.
-------
TABLE 2. (continued)
Item
OVA
TLV
HNu
AID
Probe contamina-
tion (continued)
Probe assembly
Audible alarms
Screening time
This flexible tip was
also helpful when
screening because it made
it easier to get the
probe tip close to the
source interface.
The assembly was con-
veniently sized and not
uncomfortably heavy.
The alarm adjust knob
on the back was broken
off when the assembly
was dropped.
The alarm cannot be
heard in most plant
environments. The ear
plug was very uncom-
fortable and the oper-
ators did not wear it.
Very good; ~30 seconds
per source.
The probe is very
lightweight and easy
to manipulate.
The assembly was quite
heavy and very diffi-
cult to manipulate.
See OVA comments.
Not applicable.
Somewhat slow; "45
seconds per source.
Unknown; no response
to sources.
The assembly had a com-
fortable feel. How-
ever, the plastic bezel
damaged during the
second screening falls
off frequently.
See OVA comments.
Very good; ~30 seconds
per source.
-------
is easily replaced and reduces time spent in the field cleaning the metal sam-
ple probes. The tubing was normally replaced on an as required basis which
was fairly frequently. The replacement criteria was based on appearance of
the tubing if there was any visible contamination on it it was replaced.
The TLV comes without a shoulder strap making it very awkward to use for
screening. We used straps from the other instruments when screening with the
TLV. Also, since the zero knob is easily moved, we used a piece of tape to
secure it. This prevented us from having to constantly check and reset the
instrument zero.
The OVA comes with one strap. However, after several screenings we real-
ized that if it could be worn as a backpack, it would be easier to use, make
the operator more mobile, and speed up the screenings. Therefore, we pur-
chased a second strap from the manufacturer and wore the OVA as a backpack,
using the two carrying straps as a shoulder harness. Since the OVA has a
readout on the hand-held probe, this did not create any problem. However,
when the instrument flamed out, relighting it was a little awkward since the
operator had to either get someone else to press the igniter button on the
case or let the case slip forward on the shoulders and stretch around to reach
it.
The OVA has been used extensively, and a common problem encountered has
been battery failure. After consultation with several technicians, we learned
that the requirements for long battery life were frequent use with care not to
overcharge or allow the battery to go into deep discharge. Since the instru-
ment was used by many people on an irregular basis, no satisfactory solution
to the battery problem was found. Therefore, we frequently replace the
battery on the OVA. However, with regular use and maintenance, the battery
problem should be minimal.
Our experience with the battery in the OVA highlighted the fact that the
battery problem was really one associated with the battery type (rechargeable
nickle cadmium batteries) and not the Century Systems battery. Basically, the
nickle cadmium rechargeable batteries all require the same kind of treatment
for long life; i.e., frequent use (regular discharging and recharging), pro-
tection from overcharging, and protection from deep discharge. Therefore, any
device with rechargeable nickle cadmium batteries should be used with the same
precautions to ensure long battery life.
16
-------
The HNu did not detect any leaks. The HNu response was surprising, es-
pecially at Site 2, because this plant processes chlorinated hydrocarbons.
Previous EPA reports indicated that the HNu would respond to chlorinated hy-
drocarbons. The chlorinated hydrocarbons, carbon tetrachloride and F-ll, were
present in pure form in the screening area, but the HNu did not respond to
them even though it was properly calibrated on a 48 ppm benzene standard. We
do not know why the HNU did not respond.
As a test of the HNu's response small samples of carbon tetrachloride and
F-ll were taken from process sample drains after purging them. All three HNu
detectors were used to see if any would respond to the two samples. Each
probe was attached and the appropriate calibration settings dialed into the
instrument. The probe tip was then lowered into a container of one of the
compounds until the tip was about 1/2 inch from the liquid surface. This was
repeated at least three times. A comment on the instrument's responses was
then made.
In all cases, the HNu responded with a zero or slightly below zero read-
ing when the probe tip was placed near the free surface of the two samples.
Sometimes as the probe was brought near the liquid, an upscale response would
occur but when the instrument reading stabilized with the probe tip near the
liquid surface, it always indicated zero or below. The upscale responses were
erratic and unpredictable with each probe.
During the screenings, we collected data on the leak rate over the same
group of sources in four different plants. These data are presented in Appen-
dix B. Although in some cases the data were collected using procedures that
do not exactly follow RM 21 procedures and for sources that are not covered by
federal regulations, the data still demonstrate how using different instru-
ments produce similar results.
The screening at Site 1 resulted in leak rates of 8.60, 7.46, and 8.05
percent from the AID, TLV, and OVA, respectively (results for the HNu are not
included). These are higher than the overall plant leak rate since the
screening area was chosen to include as many leaking sources as practical.
At Site 2 three different people screened the area with two different
instruments. The leak rates were 5.16, 5.16, and 5.16 percent with the OVA
and 3.23, 3.23, and 2.58 with the TLV. At this plant the response factors for
17
-------
the instruments were quite high (20 and 30, respectively). This is probably
why the results do not agree well and points out that high response factors
are undesirable.
At Site 3 the OVA was not operational during the evaluation; the pream-
plifier did not function on the day of the evaluation (the OVA had operated
properly only two weeks earlier). However, data were available from a screen-
ing in the area approximately three months earlier. The plant personnel re-
ported that there had been no attempted repairs of leaking sources in the area
in that three-month interval and that the process was operating at the same
level as when the OVA was used as when the TLV, AID, and HNu were used. The
screening results were 4.93, 2.46, and 8.45 percent for the AID, TLV, and OVA,
respectively. (The HNU did not respond to any leaks in the area.) The higher
rate for the OVA is unexplained. The difference between the AID and TLV ap-
pears to be because of a lower response factor for the TLV than the AID. How-
ever, the area supposedly only contained methane, ethane, and propane which
have very low response factors for both instruments.
At Site 4 the leak rates were 1.63, 1.09, and 1.63 percent for the AID,
TLV, and OVA, respectively. The only leaking material at this plant was
methane, and all three instruments were calibrated with a 10,000 ppm methane
calibration gas which definitely seems to enhance the reproducibility of re-
sults among the instruments.
18
-------
SECTION 4
DESIRABLE INSTRUMENT FEATURES AND
RECOMMENDATIONS FOR FUTURE STUDIES
Rather than recommending improvements in each instrument, we have devel-
oped a description of the desirable features that should be included in a
field screening instrument:
The strap should allow for carrying the instrument on the back, out
of the way, leaving the hands free for climbing, handling log
sheets, and manipulating the probe and readout (assuming the read-
out is attached to the probe).
° The calibration controls should be located in the backpack and pro-
tected by the instrument case cover. All controls should have locks
to prevent unintentional movement.
0 The readout should be analog and use a logarithmic scale that ranges
from 10 to 100,000 ppm. RM 21 must be modified to accept the re-
sulting scale divisions.
0 The readout should have a lock-and-hold reading capabilities and/or
hold-highest-reading function switch.
0 There should be provisions to use the instrument as a go/no-go de-
tector with indicator lights to show whether the reading is above or
below the calibration point.
0 An igniter button should be located on the probe/readout if the
instrument uses an FID.
0 A series of status indication lights should appear on the probe/
readout assembly to show if the instrument has sufficient battery
charge and if it is responding.
0 The sample line connecting the backpack and probe/readout should be
at least 4 ft. long and very flexible.
0 A holster should be provided for the probe/readout so that both
hands can be freed for climbing and handling data sheets.
0 The probe assembly should have provisions for frequent cleaning
because the probe becomes contaminated with grease and other mate-
rials during the screening process.
19
-------
0 The system should be protected from the elements and be able to
operate in light rain, high humidity, and high ambient temperatures.
0 The system should protect the rechargeable battery from overcharging
and deep discharge. Provisions should be made for easy battery
removal and replacement. Ideally, the charging system should be
capable of being left on charge at all times. There should also be
an indication of when the unit is within one hour of being too weak
to operate effectively. This would provide time to check the
instrument calibration before the unit required recharging.
0 The calibration system should allow for easy calibration to multiple
calibration gases.
As screening instruments, all the instruments used could be improved.
The above list of desirable features would enhance the usefulness of any of
the instruments.
The object of a screening program is to identify leaking sources so that
they can be repaired and thus reduce fugitive emissions. RM 21 and the
associated regulations define a leak based on a specified instrument reading
with the instrument calibrated with a specified calibration gas. The most
commonly used definition is a 10,000 ppm response when calibrated with a
10,000 ppm methane calibration gas. This is really a go-no-go measurement and
the exact instrument reading is not important as long as we know it is either
above or below a calibration mark established by a certified gas. There is no
requirement to measure the actual concentration.
The above is not a problem with the use of portable VOC instruments but
must be recognized by those conducting screenings and using the results. The
screening results are not meant to be a quantification of the leaking materi-
als but a count of the leaking sources where the leaking source definition is
based on a screening method. Keeping this in mind, to measure a leak we only
need to be concerned that we have a certified calibration gas, an instrument
that will consistently give the same reading when exposed to that calibration
gas and an instrument that responds in a known way to the leaking gases. An
ideal instrument for screening would be one that had only two indicators on
it; i.e., a red light indicating a value above the calibration point and a
green one for a value below the calibration point. This would require cali-
bration with two gases: one slightly above the leak rate definition point and
one slightly below.
20
-------
It would be beneficial to field test other instruments. A March 1980
report, "Summary of Available Portable VOC Detection Instruments," EPA 340/1-
80-010, on available VOC detection instruments should be updated to include
current prices and to indicate whether or not the instruments meet the re-
quirements of RM 21. Available and additional data on leak rates, repeatabil-
ity of screening values, and response factors for VOC instruments should be
assembled in a screening handbook to present the kinds of information neces-
sary for a proper screening program. Information should be included on how to
screen various types of sources, how to prepare log and repair sheets, re-
sponse factors of various instruments, lists of typical compounds encountered
in various types of facilities, etc. Screening procedures should be developed
for organic materials with response factors greater than 10.
21
-------
APPENDIX A
FOXBORO CENTURY SYSTEMS PORTABLE
ORGANIC VAPOR ANALYZER MODEL OVA-108
The OVA consists of two primary units connected with an umbilical cord,
as shown in Figure 1. It is designed to measure trace quantities of organic
materials in air employing a hydrogen flame ionization detector similar to
that utilized in laboratory gas chromatographs. The flame ionization detector
is an almost universal detector for organic compounds although it is more sen-
sitive to some organics than others. It has the sensitivity to analyze in the
parts per million (ppm) range in air, and in the presence of moisture, nitro-
gen oxides, carbon monoxide, and carbon dioxide.
The instrument has broad application and can readily be calibrated to
measure many organic vapors. The instrument has an adjustment so that the
meter readout can be expressed on any reference calibration compound. How-
ever, the instrument responds to all the compounds present in the sample
stream, and the meter reading is a measure of all the compounds in the sample
stream in terms of the calibrant gas. Specific knowledge of the instrument
response to the compound(s) must be known in order to determine the concentra-
tion of the compound(s). It has a single logarithmically scaled readout from
1 ppm to 10,000 ppm. It is designed for use as a portable survey instrument.
The OVA is certified as intrinsically safe by Factory Mutual Research Corpora-
tion (FM) for use in Class 1, Division 1, Groups A, B, C, and D hazardous en-
vironments.
The OVA is designed to detect and measure VOCs found in almost all indus-
tries, including those manufacturing petrochemicals and engaged in natural gas
transmission and distribution. It measures organic vapor concentration by
producing a response to an unknown sample that is related to a gas of known
composition to which the instrument has previously been calibrated. During
operation, a continuous sample is drawn into the probe and transmitted to the
detector chamber by an internal pumping system. The sample flow rate is
22
-------
r
Figure 1. Portable Organic Vapor Analyzer OVA-108.
-------
metered and passed through particle filters before reaching the detector cham-
ber. Inside the detector chamber, the sample is exposed to a hydrogen flame
that ionizes the organic vapors. When most organic vapors burn, they leave
positively charged carbon-containing ions that are located by a negative col-
lecting electrode in the chamber. The electric field that exists between the
conductors surrounding the flame and the collecting electrode drives the ions
to the collecting electrode. As the positive ions are collected, a current
corresponding to the collection rate is generated on the input electrode.
This current is measured with a logarithmic electrometer preamplifier that has
an output signal proportional to the log of the input or ionization current.
A signal conditioning amplifier is used to modify the signal from the preamp
and to condition it for subsequent meter or external monitor display. The
meter display is an integral part of the probe-readout assembly and displays a
range of 1 to 10,000 ppm (1 percent) on the OVA.
The basic instrument consists of two major assemblies, the probe-readout
assembly and the side pack assembly (see Figure 1). The output meter and
alarm level adjustments are incorporated in the probe-readout assembly, which
is operated with one hand. The side pack assembly contains the remaining op-
erating controls and indicators, the electronic circuitry, detector chamber,
hydrogen fuel supply, and electrical power supply. It is a quantitative in-
strument with sensitivity to 1 ppm methane and is capable of readouts from 1
to 10,000 ppm concentration by volume of methane in air.
Other major features include:
1. 250° logarithmically scaled readout;
2. internal electronic calibration;
3. less than 2-second response time; and
4. at the least, 8-hour service life for fuel supply and
battery pack.
Internal two-point electronic calibration is provided for checking the
instrument and supplying reference signals for setting up the gas select ad-
justment. A battery test feature shows charge condition to be read on the
meter. Hydrogen flameout is signified by an audible alarm as well as a visual
24
-------
indication on the meter. The instrument contains a frequencysM*l!®teii ofefcec
tion alarm that can be preset to sound at a desired concentrstnim lieviffiTL THe
frequency of the detection alarm varies as a function of detects®! "to-eH,, aim-
ing an audible indication of organic vapor concentration. Tte iirasrfcumiBKTt iis
designed for one-person, one-hand .operation, and the entire inriitt wehgjte aa
total of less than 12 pounds, including fuel supply and battery. /%« e
-------
TABLE 3. APPROXIMATE RELATIVE RESPONSE FACTORS FOR THE OVA
FOR VARIOUS HYDROCARBONS RELATIVE TO METHANE
Compound
Methane
Ethane
Propane
n-Butane
n-Pentane
Ethyl ene
Acetylene
Benzene
Toluene
Relative response
with methane as the
calibrant
1.00
.90
.64
.61
1.00
.85
2.00
1.50
1.20
gas
Nitrogen-containing compounds (i.e., amines, amides, and nitriles) re-
spond similarly to oxygenated materials. Halogenated compounds also show a
lower relative response as compared with hydrocarbons. Materials containing
no hydrogen, such as carbon tetrachloride, give the lowest response; the pre-
sence of hydrogen in the compounds results in higher relative responses.
Thus, CHCla gives a much higher response than does CCU. As in the other
cases, when the carbon to halogen ratio is 5:1 or greater, the response will
be similar to that observed for simple hydrocarbons.
The typical relative response of various organic compounds to methane is
shown in Table 4.
The OVA has negligible response to carbon monoxide and carbon dioxide.
Thus, other organic materials may be analyzed in the presence of CO and C02.
EPA has extensively measured response factors for the OVA. Three reports of
interest are: "Response Factors of VOC Analyzers Calibrated with Methane for
Selected Organic Chemicals," EPA-600/2-81-002, May 1981; "Response Factors of
VOC Analyzers at a Meter Reading of 10,000 ppmv for Selected Organic Com-
pounds," EPA-600/2-81-051; and "Response of Portable VOC Analyzers to Chemical
Mixtures," EPA-600/2-81-110, June 1981.l
26
-------
TABLE 4. APPROXIMATE RESPONSE FACTORS FOR THE OVA OF VARIOUS
ORGANIC COMPOUNDS RELATIVE TO METHANE
Compound
Methane
Ketones
Acetone
Methyl ethyl ketone
Methyl isobutyl ketone
Alcohols
Methyl
Ethyl
Isopropyl
Halogen compounds
Carbon tetrachloride
Chloroform
Trichloroethylene
Vinyl chloride
Relative response with methane
as the calibrant gas
1.00
.60
.80
1.00
.15
.25
.65
.10
.65
.70
.35
ANALYTICAL INSTRUMENT DEVELOPMENT, INC., MODEL 712
The AID consists of two primary pieces connected with an umbilical cord,
as shown in Figure 2. The instrument is designed for the continuous monitor-
ing of organic compounds in air. The measuring technique used is hydrogen
flame ionization. The system is self contained and completely portable-. The
AID is designed to meet Class 1, Division 1, Groups A, B, C, and D of the Na-
tional Electrical Code. (As of the writing of the report, the manufacturer
reports that the instrument has not been certified as intrinsically safe.)
Some of the controls, most of the electrical circuits, the battery pack, the
hydrogen cylinder, flow control system, and pump are located in the side pack,
which is normally carried over the shoulder by the attached shoulder straps.
Figure 3 shows the controls located on the top of the side pack, and Figure 4
shows the controls on the bottom of the side pack. The umbilical cord con-
tains the gaslines and the electrical connections between the side pack and
the gun. The gun contains the flame ionization detector, the electrometer
27
-------
c
• ro •
co "
Figure 2. Analytical Instrument Development, Inc., Model 712=
-------
m • • — —
'-V ^
_ I CO
i K
:.TOTAL HYDROCARBON ANALYZER
Figure 3. Controls located on the top of the AID side pack,
-------
"I • I
CO
O
ANALYTICAL ,
INSTRUMENT- '.r-/.-
OEVELOPMENT1NC.
AVONDALE PA 19311
Figure 4. Controls located on the bottom of the AID side pack.
-------
amplifier circuit, the digital readout and its circuitry, the sensitivity con-
trol and range switch, and the ignitor circuit and flameout indicator. The
gun is carried in one hand and directed toward the sample to be analyzed.
Sample air is drawn into the flame ionization detector through a short
length of stainless steel tubing. The air sample does not come in contact
with the sampling pump until after the sample passes through the flame ioniza-
tion detector and, thus, has been measured. The electrical signal from the
flame ionization detector is amplified and presented in the digital readout
which shows the concentration in ppm when the instrument is calibrated. There
are two ranges: on the "XI" range the concentration of the sampled gas is
read directly in ppm; on the "X10" range, the value shown on the readout must
be multiplied by 10 to obtain the concentration of the sampled gas. All the
readings displayed on the digital readout must be multiplied by 10. Thus, the
reading 0 to 200 ppm on the "XI" range is actually 0 to 2,000 ppm, and on the
110 range the reading 0 to 2,000 ppm is actually 0 to 20,000 ppm.
The analyzer incorporates two separate flameout indicators. If the flame
in the FID goes out, a small indicator light beside the digital readout is
illuminated. In addition, an audible alarm sounds from the side pack. Fol-
lowing a flameout, the values of the readings generally become very low or
even negative. Thus, the readings themselves provide a possible third indica-
tion of flameout.
The AID is provided with a variable concentration level alarm. If the
concentration exceeds the level set on the alarm, a buzzer sounds. This
permits monitoring for a predetermined concentration level of VOCs without
continuous observation of the digital readout on the analyzer. The alarm ad-
justment potentiometer can be set anywhere between 0 and 10,000 ppm. When the
digital readout indicates a reading above this level, the alarm sounds. A
switch is available to disable the alarm when the instrument is being cali-
brated. This defeats the alarm circuit, both for flameout indication and for
readings above the set point on the alarm circuit. A flameout will still be
indicated, however, by the small red indicator light located to the side of
the digital readout. The switch affects only the audible alarm.
Immediately above the range switch in the gun panel is the response con-
trol, which changes the gain of the amplification system and can be used to
31
-------
reproduce various response factors for different materials in the AID. Nor-
mally, it is calibrated with methane when the response control is set at 100.
Higher settings on this dial provide greater sensitivity^for the instrument,
and these settings are then used for the materials whose responses are less
than the response obtained for methane. Lower numbers provide less sensiti-
vity and are used for materials with higher responses. For RM 21 screenings,
this dial would be set for methane or hexane and not changed.
The zero control compensates for the background current in the detector
when no organic material is present in the sample. The zero adjustment is
made to provide a zero readout when sampling zero air. The span control is
used to adjust the readout of the system to provide the proper ppm readout for
methane when presented with a known methane standard and the response control
on the gun is set at 100.
The response factors listed in Table 5 are the numbers (the actual con-
centration in ppmv divided by the reference concentration in ppmv as reported
by the manufacturer) by which the reading must be multiplied to give the cor-
rect concentration in ppm for that particular VOC, assuming the AID was cali-
brated correctly against methane as a standard. The lower the number of the
response factor, the more sensitive the AID is for the material.
TABLE 5. APPROXIMATE RESPONSE FACTORS FOR THE AID
OF VARIOUS VOCs RELATIVE TO METHANE
Compound
Methane
Propane
Butane
Hexane
Benzene
Toluene
Ethanol
Isopropanol
Methyl ethyl ketone
Methyl isobutyl ketone
Chloroform
Carbon tetrachloride
Trichloroethylene
Response factor
1.00
1.30
1.04
0.87
0.48
0.65
2.4
0.75
0.90
0.85
2
10
1.3
32
-------
As Table 5 shows, the hydrocarbons have approximately equal responses on
the AID versus methane standards. This means that the AID is giving about the
same response for each mole or volume ppm. The aromatic hydrocarbons show
slightly greater sensitivity than do the aliphatics. As can be seen, the in-
troduction of chlorine into the molecule significantly reduces the sensitivity
of the detector system for the molecules. The size of the molecules increases
the effects of chlorine substituted into it. These effects are most notable
in one- and two-carbon molecules. The substitution of oxygen into the mole-
cule also reduces sensitivity for the one- and two-carbon molecules. Again,
it should be pointed out that these response factors are approximate and will
vary somewhat among individual instruments. However, on a given instrument,
these will remain constant over time.2
BACHARACH INSTRUMENTS TLV SNIFFER
The TLV consists of a hand-held control module with a light sample line
and metal probe attached, as illustrated in Figure 5. The TLV is housed in a
brushed-aluminum and blue plastic case. The instrument weighs 5-1/2 Ib and
measures 8 x 6-1/2 x 3-1/4 in. A front panel contains a meter reading direct-
ly in ppm and a control knob for range selection and zero setting of the indi-
cating meter pointer. Figure 6 shows the front of the TLV. The side panels
of the instrument provide plugs and connectors for an air sample probe, ear-
phones, battery charger, and recorder. Removal of 10 screws holding the plas-
tic cover to the rigid aluminum case gives access to batteries and calibration
adjustment controls located within the interior of the case. The TLV has been
certified as intrinsically safe by the Factory Mutual Laboratories fcr use in
hazardous areas designated Class I, Divisions 1 or 2 by the National Electri-
cal Code.
The TLV air sampling system consists of a short intake connection leading
directly into an interior aluminum air chamber holding the detector, a 6-in.
length of tubing connecting the chamber and a miniature sample pump, and a 3-
in. length of tubing leading to an exhaust port on the left side of the in-
strument case.
The combustible gas detector consists of an active catalyst-coated re-
sistance element to oxidize combustible gas and an identical second resistance
33
-------
75T *»
to
-F*
Figure 5. TLV.
-------
CM IHHWD fQt\ CPU \ HEXAfJf |
TIV SNIFFER
Figure 6. Front of the TLV.
-------
element without the catalyst coating that provides a reference resistance.
Since both operate at approximately equal temperatures, only changes in gas
content of sampled air cause differences in resistance between the two to pro-
duce signals to the meter and gas alarm circuits. Both the catalyst-coated
(active) element and the reference element are incorporated in a Wheatstone
bridge circuit in such a way as to produce an electrical output proportional
to their differences in resistance. Because changes in air sample temperature
and humidity affect both active and reference elements equally, the electrical
signal output is proportional to the heat transferred to the sensor element,
which is proportional to the heat of combustion and concentration of the com-
pound in the sample of air (expressed in volumetric terms as ppm). However,
sudden changes in humidity may affect the zero reading on the X-l range. The
instrument should, therefore, be zeroed at the same relative humidity prevail-
ing during use by using clean ambient air instead of prepared air (dried com-
pressed air).
An audible alarm will sound at a preset gas concentration level. The
alarm response is controlled by amplification of the difference between the
gas concentration signal level and an internal reference voltage.
The TLV circuitry and meter provide readings from 0 to 10,000 ppm in
three range settings. The meter readout is scaled as 0-100 in 2 percent in-
crements. The amplifying circuits provide three decade settings, so the meter
readings can represent 0-100, 0-1,000, and 0-10,000 ranges. Each range
setting requires an easily made adjustment of the ZERO ADJUST control knob on
the instrument front panel to set the meter indicating pointer to zero. Span
adjustments for full-scale pointer deflection within each range are made
periodically as necessary by means of three gain potentiometers located within
the instrument.
The extent to which meter readings correspond to actual ppm of combusti-
ble gases in sampled air depends upon: (1) the internal electrical stability
of the instrument; (2) proper calibration of the instrument on gas mixtures of
known concentration; and (3) purity of the air sample used for meter pointer
zero setting.
Hexane gas is commonly used for factory calibration and subsequent in-
service recalibrations of the TLV. According to the manufacturer, the ppm
36
-------
TABLE 6. MULTIPLYING FACTORS FOR CONVERTING PPM METER READINGS
OF HEXANE-CALIBRATED INSTRUMENTS TO PPM CONCENTRATIONS OF OTHER GASES3
(Approximations)
Gas detected
Acetone
Acetylene
Acrylonitrile
Benzene
1,3-Butadiene
Butane
Butyl acetate
Carbon disulfide
Carbon monoxide
Cyclohexane
Ethane
Ethanol
Ethyl acetate
Ethyl ether
Ethyl ene
Ethylene oxide
Heptane
Hexane
Hydrogen
Factor
1.50
1.78
1.54
1.02
1.52
1.04
2.08
5.92
5.11
1.02
1.36
1.90
2.22
1.30
1.38
2.05
1.05
1.00
1.45
Gas detected
Hydrogen sulfide
Isopropanol
Methyl ethyl ketone
Methane
Methanol
Methyl acrylate
Methyl chloride
Methyl chloroform
Pentane
Perch! oroethyl ene
Propane
Propylene
Styrene
Tetrahydrofuran
Toluene
Trichloroethylene
Vinyl acetate
Vinyl chloride
o-Xylene
Factor
18.60
1.59
1.60
1.58
3.71
3.37
4.02
4.44
1.04
13.66
1.14
1.30
2.25
1.41
1.03
6.40
2.00
2.24
1.64
concentrations of gases other than hexane with instruments calibrated for
hexane are determined by multiplying the ppm meter reading by the factor for
the gas detected as listed in Table 6.3 The table values are equivalent to
the response factor referred to in RM 21.
HNu SYSTEMS, INC., MODEL PI-101
The HNu consists of two units, a side pack with all the controls and an
interchangeable hand-held unit with the fan and detector connected by an um-
bilical cord to the side pack. There are three flame and detector assemblies
available for the unit. Each one has a different lamp (sensor) for use with
different groups or types of chemicals. The three lamps are 9.5, 10.2, and
11.7 ev. Each lamp will have a different calibration setting and a different
group of compounds for wJiich it is most sensitive. The HNu has been certified
as intrinsically safe by Factory Mutual Laboratories for use in hazardous
areas designated Class 1, Division 2, Classes A, B, C, and D. Figure 7 shows
37
-------
CO
03
Figure 7. HNu Systems, Inc., PI-101. .
-------
the assembled instrument and Figure 8 shows the instrument controls and read-
out.
The HNu has been designed to measure the concentration of trace gases in
many industrial or plant atmospheres. The analyzer employs the principle of
photoionization detection (PID). This process is termed photoionization be-
cause the absorption of ultraviolet light (a photon) by a molecule leads to
ionization via:
RH + hu -»• RH+ + e"
where RH = trace gas
hu = a photon with an energy _> Ionization Potential of RH
The sensor consists of a sealed ultraviolet light source that emits pho-
tons energetic enough to ionize many trace species (particularly organics) but
which do not ionize the major components of air such as 02, N2, CO, C02, or
H20. A chamber adjacent to the ultraviolet source contains a pair of elec-
trodes. When a positive potential is applied to one electrode, the field
created drives any ions formed by absorption of UV light to the collector
electrode where the current (proportional to concentration) is measured.
To minimize adsorption of various sample gases, the ion chamber, made of
an inert fluorocarbon material, is located at the sampling port, and a rapid
flow of sample gas is maintained through the small ion chamber volume.
The analyzer will operate either from a rechargeable battery for more
than 10 hours or continuously from the AC battery charger. A solid-state am-
plifier board in the probe and a removable power supply board in the readout
module enable rapid servicing of the unit in the field.
The instrument was designed to measure trace gases over a concentration
range from less than 1 ppm to 2,000 ppm. The manufacturer states that higher
levels of various gases (to percentage range) can be measured by diluting the
sample with clean air to a concentration of less than 500 ppm. This is gen-
erally within the linear range of the instrument, and if the measured concen-
tration is multiplied by the dilution ratio the correct concentration in the
stream can be determined. No attempt was made to use this technique during
any of the field trials.
39
-------
photo'ionizer
systems inc
Figure 8. HNu controls and readout.
-------
If the probe is held close to AC power lines or power transformers, an
erroneous reading may be observed.
The instrument is equipped with an automatic solid state battery protec-
tion circuit. When the battery voltage drops below ^11 volts, the circuit
will automatically turn off the power to the instrument. This protection cir-
cuit prevents deep discharging of the battery and considerably extends the
battery life. If the instrument is unintentionally left on overnight, the
battery will be unharmed because of the battery protection circuit.k
41
-------
APPENDIX A REFERENCES
1. Foxboro Analytical Division of Foxboro Company. Instructions and Service
Manual, Century Systems Portable Organic Vapor Analyzer Model OVA 108,
MI2R900AD. Undated.
2. Analytical Instrument Development, Inc. Model 710 Portable Total Hydro-
carbon Analyzer. Undated.
3. United Technologies Bacharach. Bacharach Instruction Manual, TLV Snif-
fer, Instruction 23-9613, Revision 1, September 1982.
4. HNu Systems, Inc. Instruction Manual for Model PI-101 Photoionization
Analyzer. 1975.
42
-------
APPENDIX B
SUMMARY OF FIELD SCREENING SHEETS AT FOUR SITES
TABLE 7. SUMMARY OF SOURCES WITH ONE OR MORE READINGS OVER 10,000 PPMV
EVALUATED APRIL 26-27, 1984
SITE 1
Tag number
11701
11731
11732
12001
12022
12041
12047
12101
12112
12119
12121
12125
12603
12604
12606
12629
12641
12645
12654
12677
12700
12810
\-l--i mil dA I
Instrument readings
AID
osa
OS
15,000
OS
OS
12,000
OS
OS
660
315
110
170
0
1,400
OS
OS
0
1,500
1,500
0
OS
OS
TLV
OS
OS
OS
1,000
NS
8,800
2,200
OS
100
1,800
0
200
400
0
OS
NS
0
7,000
0
NS
OS
OS
OVA
50,000
FO
FO
NS
NS
FO
40,000
75,000
150
10,000
300
100
1,000
500
FO
100
100
75,000
900
FO
FO
FO
HNu 9.5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
100
0
0
0
0
0
0
0
HNu 10.2
0
0
0
0
0
0
0
100
0
100
0
0
0
0
100
100
0
100
0
0
0
0
HNu 11.7
0
0
0
0
0
60
60
100
100
110
100
100
100
80
100
0
100
80
90
0
0
0
43
-------
TABLE 7. (continued)
Tag number
12817
12865
12907
12924
12926
12958
12959
Total sources
Leak rate, %b
Instrument readings
AID
OS
OS
OS
OS
18,000
OS
OS
221
8.60
TLV
OS
OS
OS
OS
OS
OS
OS
228
7.46
OVA
FO
FO
FO
FO
FO
FO
FO
236
8.05
HNu 9.5
0
0
0
0
0
0
0
226
0
HNu 10.2
0
0
0
0
0
0
0
223
0
HNu 11.7
0
0
0
0
0
0
0
219
0
NS = Not screened.
OS = Off scale.
FO = Flameout.
DBased on sources with a 10,000 ppm response for the AID and OVA and a 6,000
ppm response for the TLV.
44
-------
TABLE 8. SUMMARY OF SOURCES WITH ONE OR MORE READINGS OVER 10,000 PPMV
EVALUATED MAY 10, 1984
SITE 2
Taq number
2019
2130
2139
2172
2791
2795
2805
2811
Total sources
Leak rate, %
Instrument readings
OVA
SJ
F0a
1,000
1,400
600
500
FO
1,200
2,000
155
5.16
JA
FO
FO
1,500
1,100
600
FO
1,000
1,500
155
5.16
JL
FO
L
1,200
700
500
FO
1,100
1,800
155
5.16
TLV
SJ
2,000
700
300
200
250
1,600
500
1,400
155
3.23
JA
2,000
300
800
400
200
1,600
700
500
155
3.23
JL
300
2,000
500
200
300
1,800
600
900
155
2.58
AID
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
N/A
HNu; all
3 lamps
0
0
0
0
0
0
0
0
0
FO = Flameout.
L = Leaker.
N/A = Not available.
SJ = Skip Jones.
JA = Jack Allen.
JL = Justin Lejeune.
The AID failed to operate and could not be used,
DBased on sources with a 500 ppm response.
45
-------
TABLE 9.
SUMMARY OF SOURCES WITH ONE OR MORE READINGS OVER 10,000 PPMV
EVALUATED MAY 30, 1984
SITE 3
(continued)
Tag number
9035
9048
9151
9162
9280
9281
9283
9286
9290
9292
9294
629
3677
9401
9027
9033
9034
9036
9059
9062
9065
9067
9146
9166
Instrument readings
AID
13,000
10,500
11,000
FO
FO
FO
18,000
10,000
FO
FO
FO
FO
FO
FO
0
490
3,000
4,500
5,600
3,000
800
6,500
8,000
1,500
TLV
4,000
3,000
3,500
6,800
6,000
OS
4,600
1,100
OS
OS
5,300
OS
0
OS
0
0
600
1,200
1,400
1,400
100
2,600
1,800
0
OVAa
NLb
NL
ML
NL
L
L
NL
L
L
L
L
NL
Unknown
L
L
L
L
L
L
L
L
L
L
L
HNu; all
3 lamps
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
46
-------
TABLE 9. (continued)
Tag number
9177
9178
9179
9282
9400
9416
9427
Total sources
Leak rate, %c
Instrument readings
AID
6,800
900
0
100
400
0
700
284
4.93
TLV
3,700
0
0
0
0
0
200
284
2.46
OVA
L
L
L
L
L
L
L
284
8.45
HNu; all
3 lamps
0
0
0
0
0
0
0
284
0
OVA based on screening on 2/29/84, except for source 629, which is based on
screening 5/22/84. The process operation is reportedly the same during the
time of all screenings, and there was reportedly no maintenance performed on
any of the equipment.
FO = Flameout.
NL = Nonleaker, <10,000 ppm instrument response.
L = Leaker, >_10,000 ppm instrument response.
cBased on sources with >_10,000 ppm response on the OVA and AID or a 6,000 ppm
response for the TLV.
47
-------
TABLE 10. SUMMARY OF SOURCES WITH ONE OR MORE READINGS OVER 10,000 PPMV
EVALUATED JUNE 21, 1984
SITE 4
Tag number
1530
1532
1533
1559
1565
1585
1589
1615
1633 Flange
1637
1638
1641
1643
Leak rate, %b
Total sources
(134 valves,
50 flanges)
Instrument readings
AID
12,000
osa
3,000
500
0
0
0
680
10,000
800
1,000
5,000
0
1.63
184
TLV
0
OS
0
0
0
0
400
0
6,000
0
200
2,000
0
1.09
184
OVA
0
75,000
0
0
75,000
9,000
3,000
0
10,000
0
3,000
0
3,000
1.63
184
OS = Off scale.
3Based on sources with a 10,000 ppm response for the OVA and AID and 6,000
ppm response for the TLV.
48
-------
APPENDIX C
SUMMARY OF COMMENTS FROM FIELD EVALUATION SHEETS
Tables 11 and 12 are composite summaries of the comments from the daily
and site evaluation sheets. These comments are summarized in the report in
Section 3.
49
-------
TABLE 11. SUMMARY OF COMMENTS FROM DAILY INSTRUMENT CRITIQUES
Number of evaluation sheets: 29
1. Date of operation: 4/25/84 through 6/19/84.
2. Persons making evaluation: Robert Ressl, Stan Lueck, Justin LeJeune,
Skip Jones, Jeff Hall, Jack Allen.
3. Instrument: HNu (with 9.8, 10.7, or 11.2 bulb) , OVA-108 , TLV ,
or AID . (ALL)
4. Period of operation (use 24-hour time): 2-9 hours.
5. Was instrument on continuously or intermittently during the period of
operation? Continuously except for accidental shutoffs.
6. Operational or screening location (plant name, process unit, etc.):
Hercules, CA; Goleta, CA; Calvert City, KY; and Lake Charles, LA.
7. Names of all individuals using instrument during period of operation.
Same as 2.
8. Number of sources screened or sampled (indicate the location or number of
data log sheets or attach copies of the log sheets): 200-300.
9. Indicate location of calibration log sheet or attach copies for calibra-
tion before operation. N/A
10. Indicate location of calibration log sheet or attach copies for calibra-
tion after the period of operation. N/A
11. Instrument scales used: Low X Medium X High X
12. Report the calibration settings of the instrument used for the AID,
OVA-108, and HNu. Varies
13. PROBLEMS: (Indicate "yes" answers only. Lack of a response is con-
sidered a "no.")
Yes
a. Any problems with, initial or final calibrations? _6_
b. Any problems with calibration knobs? 9
c. Any flameouts of instruments? 13
(continued)
50
-------
TABLE 11. (continued)
13. (continued) Yes
d. Any problems with stability of instrument readings? 13_
e. Any problems with instrument background readings? 3
f. Any problems with instrument response time? 15
g. Any problems with weak battery or battery failure? 7
h. Any mechanical problems? 3
i. Any failures with the electrical components? 1
j. Any failures of the sample hose? 6
k. Any failures of the electrical connectors between the
probe and the instrument? l
14. For all "yes" answers to Item 13, explain in detail the particular
nature of the problem and how it was resolved.
a. During calibration of the AID on a 9,970 ppm methane standard, had
problems in adjusting to a correct value because readout updated so
frequently. Also had to set calibration to 200 versus 100 recom-
mended by the manufacturer (3 responses).
The HNu calibration was not as indicated on the sample gas supplied.
The lamp window was cleaned as suggested but the result was the
same.
The TLV could not be calibrated to a 10,000 ppm reading. The 10,000
ppm standard showed as 6,000 ppm.
b. Tape down the zero adjust knob on the TLV (9 responses).
The AID calibration setting was other than recommended by the
manufacturer.
c. AID and OVA flame out on high concentrations (6 responses).
AID tended to flame out frequently and was often hard to relight and
keep lit (3 responses).
The AID flameout light is impossible to see in bright sunlight.
(continued)
51
-------
TABLE 11. (continued)
14. (continued)
The OVA flameouts posed little problem since it was easily relit.
The AID would not stay lit and could not be evaluated.
d. AID is difficult to read at an angle and the display changes too
quickly (9 responses).
TLV does not seem to pick up all leakers.
The HNu needle responds strangely because of a weak battery.
The HNu was slow to respond if at all. When it did respond, it
would peak, then drop back to zero reading (2 responses).
The TLV readings are not in accordance with the OVA readings.
e. With the OVA, where leaks were detected close to the background, it
was difficult to distinguish them since instrument was showing a
high background (100 ppm) (3 responses).
f. Slow response of the TLV (13 responses).
The HNu response was slow, if it responded at all (2 responses).
The HNu response was slow and would peak, then drop to a zero
reading.
g. When the TLV was turned on, battery checked OK but after only a
couple of hours of screening, noted erratic response. Battery check
showed low battery. Therefore, screening results are doubtful (2
responses).
The operating manual says a battery charge for the TLV will last 6-8
hours. This may be too short for a long screening day.
The HNu battery checked OK but within an hour in the field was weak
and instrument responded erratically (2 responses).
When the TLV battery became weak, the audible alarm went off. The
instrument had been responding strangely for about 10 minutes before
the alarm was heard. Therefore, it had probably gone off and was
not heard because-of the noise in the plant.
A weak TLV battery seemed to cause even slower response.
(continued)
52
-------
TABLE 11. (continued)
14. (continued)
h. The AID needs a shade over the red flameout light so it can be read
in bright sunlight.
The HNu carrying strap is uncomfortable.
The AID probe cover came off in shipping.
i. The AID would not work 2 days after the screening and was returned
to the manufacturer for service. The manufacturer reported that the
detector was contaminated with a brown coating identified as a sul-
fite.
j. The TLV sample hose has a tendency to kink (6 responses).
The OVA sample line crimped near the metal "L." The outer protec-
tive sleeve was slipped over the crimp and that seemed to prevent it
from recurring.
k. The electrical connector between the probe and the side pack on the
HNu is hard to turn during detachment and hard to align when attach-
ing.
15. ANY OTHER COMMENTS:
The OVA was the easiest, most comfortable instrument to use, and the
double strap back pack configuration made it very easy tc use.
The OVA is quite heavy; however, the two straps make it much easier to
carry and use.
The OVA needs a wrist strap or holster to hold the hand-held unit to free
the operator's hands for climbing and handling log sheets.
The two shoulder straps and analog meter on the OVA made it very easy to
use, and its response was very fast in relation to the other instruments.
When OVA first lit off, it read off scale and took several minutes to get
back on scale.
There should be some way to lock and hold the highest reading on the AID.
The AID single strap was uncomfortable.
The AID was easy to use except for noted problems.
(continued)
53
-------
TABLE 11. (continued)
15. (continued)
The switches on both ends of the AID side pack were inconvenient and
resulted in turning the instrument off instead of the alarm.
The AID umbilical is too short to allow full reach.
The AID readings should be stabilized.
The light weight of the TLV is very desirable.
The TLV sample hose was awkward to use (3 responses).
It was inconvenient to have to look away from the probe to read the
instrument with the TLV.
The TLV needs a carrying strap (5 responses).
The TLV probe was easy to handle (2 responses).
The TLV needs some way to store the probe and hose, freeing the opera-
tor's hands.
The TLV is slow and would take longer to screen with.
The HNu can only read up to 2,000 ppm but even on the worst leaks it
would not read higher than 150 ppm.
The HNu probe is uncomfortable to use (3 responses).
The HNu did not respond to any of the compounds in the area (2
responses).
The HNu was quite awkward to carry for any length of time.
The HNu is very difficult to climb ladders with.
On the TLV, HNu, and AID, the low voltage connector on the battery
charger is of a type that has caused problems with other similar battery
chargers.
The audible alarm on the OVA, AID, and TLV cannot be heard in most
plants.
54
-------
TABLE 12. SUMMARY OF SITE CRITIQUES
1. Identify instrument used:
HNu 3
OVA 5
TLV 5
AID 5
2. Sketch the arrangement of the instrument as used and show any problem
areas:
Sketches not shown - modifications discussed in report Section 3.
3. Describe ease of use in relationship to the other instrument(s) used at
the evaluation site.
Ratings: 1st 2nd 3rd 4th
OVA 5
TLV 2 3
AID 3 1
HNu 2
The TLV was light and easy to carry but very slow.
4. Describe any problems in using the instrument to comply with Reference
Method 21 procedures:
OVA - None.
AID - The calibration was difficult because of the changing display (2
responses).
TLV - There are no calibration adjustments. Consequently, the instrument
reading on the 10,000 ppm methane standard is noted and a scale
conversion factor can be used to approximate actual readings as
methane (3 responses).
HNu - Can't calibrate to 10,000 ppm standard (2 responses).
The instrument didn't respond to the chemicals in the screening
area.
5. Describe any problems with handling the instrument:
OVA - None.
AID - The readout was difficult to read from an angle (2 responses).
(continued)
55
-------
TABLE 12. (continued)
5. (continued)
AID - (continued)
The instrument was very heavy with the single strap arrangement (3
responses).
TLV - No shoulder strap makes carrying the instrument difficult and the
sample line will crimp and shut off the sample flows (2 responses).
The sample line was cumbersome (3 responses).
HNu - The probe was awkward to handle (3 responses).
6. Describe any signs of damage that the instrument shows as a result of the
use at the evaluation site.
OVA - None.
TLV - Kink in sample hose (2 responses).
HNu - None.
7. Describe any recommendations for modifications in the instrument design
or function:
OVA - Use with two straps and wear as a back pack.
Develop a holster to hold the hand-held unit when not in use.
AID - The instrument needs a better strap (3 responses).
Make the umbilical longer.
Change the display.
Reduce how fast the digital display changes readings and shield
the flame and light so it can be seen in bright sunlight.
TLV - The instrument needs a shoulder strap (3 responses).
Need a more durable and powerful battery pack and faster response.
Add capabilities to conduct field calibration.
HNu - Make the instrument respond to more chemicals (2 responses).
(continued)
56
-------
TABLE 12. (continued)
escribe the instrument recharging procedure for the next day's use:
ALL - Charged overnight.
. Describe any problems with the hydrogen supplies or the power packs:
OVA - None.
AID - None.
TLV - None.
HMu - The instrument appeared charged when the screening started; how-
ever, after a short time of screening, the battery went dead.
10. Describe how the calibration gases were used:
OVA, AID, TLV -
A bag of a 10,000 ppm methane in air standard was prepared by fill-
ing and purging the bag three times. Then the bagged standard was
attached to the instrument and the calibration set. The instrument
was detached and reattached to the bagged standard gas several
times until the readings were stable.
HNu - Calibrated using 48 ppm standard only.
11. Any other comments:
OVA - The instrument failed to operate on 5/30/84. The instrument manu-
facturer's representative replaced the preamp in the detector.
AID - Concept and design seem practical. However, problems intrinsic to
the instrument reed to be solved.
TLV - The battery charge did not last long enough.
HNu - The instrument did not respond to the chemicals being measured.
57
------- |