EPA/600/R-95/078
July 1995
Measurement and Analysis of
Vapor Sensors Used at
Underground Storage Tank Sites
by
Carnegie Mellon Research Institute
4400 Fifth Avenue
Pittsburgh, PA 15213-2683
A Division of Carnegie Mellon University
Project Report
#IV-0900-NAEX
Project Officer
Katrina E. Varner
U.S. Environmental Protection Agency
National Exposure Research Laboratory
Las Vegas, NV 89193
prepared by
Marc A. Portnoff, Richard Grace, Jeff Hibner
National Exposure Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
Printed on Recycled Paper
-------
NOTICE
The U.S. Environmental Protection Agency (EPA), through its Office of Research and
Development (ORD), wholly funded and managed the extramural research described here. It has been
peer reviewed by the Agency and approved as an EPA publication.
The information in this document has been wholly funded by the U.S. Environmental Protection
Agency under PO #IV-0900-NAEX to Carnegie Mellon Research Institute. Mention of trade names or com-
mercial products does not constitute endorsement or recommendation for use.
-------
ABSTRACT
This program is a continuation of an investigation to quantify the operating characteristics of vapor
sensor technologies used at underground storage tank (UST) sites. In the previous study [EPA/600/R-
92/219] the sensitivity, selectivity, and response time to simulated UST environments were measured for
metal oxide semiconductor gas sensors and polymer adsorption gas sensors.
During this study, a third vapor sensor technology, the catalytic bead sensor, was selected and
tested. The testing was performed using the Carnegie Mellon Research Institute (CMRI) automated gas
testing facilities. The automated laboratory system (located at the testing facility) monitored the sensors'
responses while dynamically exposing them to various mixtures of methane, butane and xylene. The sen-
sors were also tested to determine the effects of humidity on their responses. Sensor responses were
characterized by sensitivity, selectivity, and speed of response and recovery to selected test concentra-
tions of methane, butane, and xylene. The test results are presented as a list of sensor specifications to
allow for direct comparison of these three different types of sensors.
Also, during this study, the metal oxide semiconductor gas sensors and polymer adsorption gas
sensors were re-tested after one year of continuous operation. The test results are presented as a list of
sensor specifications to quantify the stability of sensor performance.
-------
CONTENTS
Page
Abstract jjj
Figures vi
Tables vii
1.0 FOREWORD 1
2.0 INTRODUCTION 2
2.1 Background 2
2.2 Project Overview 3
2.3 Catalytic Bead Sensors 4
2.4 Metal Oxide Semiconductor Sensor and Polymer Adsorption Sensor 4
3.0 EXPERIMENTAL PROCEDURE 6
4.0 SENSOR CONSTRUCTION AND MODEL EQUATIONS 9
4.1 Polymer Adsorption Sensor 9
4.2 Metal Oxide Semiconductor Sensor 9
4.3 Catalytic Bead Sensor 11
5.0 TEST DESCRIPTIONS 14
5.1 Individual Gas - Sensitivity and Selectivity Test 14
5.2 Multiple Gas - Sensitivity and Selectivity Test 14
5.3 Multiple Gas - Water Vapor Selectivity Test 14
5.4 Response and Recovery Time Test 14
6.0 RESULTS and DISCUSSIONS 17
6.1 Catalytic Bead Sensors 17
6.1.1 Reproducibility 17
6.1.2 Sensitivity 20
6.1.3 Humidity Response 20
6.1.4 Selectivity 21
6.1.5 Speed of Response and Recovery 22
6.2 Stability Data for the Metal Oxide Semiconductor Sensor and Polymer Adsorption Sensor .22
6.2.1 Polymer Adsorption Sensor 26
6.2.2 Metal Oxide Semiconductor Sensor 26
7.0 CONCLUSIONS 32
8.0 RECOMMENDATIONS 33
REFERENCES 34
iv
-------
CONTENTS (Continued)
Page
APPENDIX
A. Catalytic Bead Sensors 35
B. Catalytic Bead Sensor Test Data 39
C. Polymer Adsorption Sensor and Metal Oxide Semiconductor Sensor Test Data 45
-------
FIGURES
Number Page
3-1 CMRI Gas Sensor Characterization Facility 6
3-2 DET Catalytic Bead Sensor Circuit Diagram (Wheatstone Bridge Configuration) 8
3-3 GAT Catalytic Bead Sensor Circuit Diagram (Wheatstone Bridge Configuration) 8
4-1 Polymer Adsorption Sensor 10
4-2 Polymer Adsorption Sensor Response to Xylene 10
4-3 Metal Oxide Semiconductor Sensor 12
4-4 MOS 823 Sensor Response to Xylene 12
4-5 DET Catalytic Sensor's Response to Butane 13
5-1 Individual Gas - Sensitivity and Selectivity Test 15
5-2 Multiple Gas - Sensitivity and Selectivity Test 15
5-3 Multiple Gas - Water Vapor Selectivity Test 16
5-4 Response and Recovery Time Test 16
6-1 Catalytic Bead Sensor Response to Changes in Humidity, in a Mixture of Methane,
Butane, and Xylene 21
6-2 GAT Catalytic Bead Sensor - Standard Mode - Response to Methane, Butane,
and Xylene Concentration Ramps 23
6-3 GAT Catalytic Bead Sensor - Methane Insensitive Mode - Response to Methane,
Butane, and Xylene Concentration Ramps 23
6-4 Catalytic Bead Sensor Response to Mixtures of Methane, Butane, and Xylene 24
6-5 Catalytic Bead Sensor Butane Response and Recovery Plot 24
6-6 Expanded View - Catalytic Bead Sensor Butane Response and Recovery Plot 25
VI
-------
TABLES
Number
Page
2.1 Major Components of API PS-6 Gasoline 2
2.2 Vapor Sensor Signal Output 3
6.1 Catalytic Bead Sensor Specifications 18
6.2 Catalytic Bead Sensor Response to Dynamic Mixing of Methane, Butane, Xylene,
and Water Vapor 19
6.3 Polymer Adsorption Sensor Stability Data 27
6.4 MOS 823 Sensor Stability Data 28
6.5 MOS 822 Sensor Stability Data .29
6.6 MOS 812 Sensor Stability Data 30
6.7 MOS 813 Sensor Stability Data '' ' '31
B1 DET Catalytic Bead Sensor Measurements 39
B1 DET Catalytic Bead Sensor Measurements (cont.) 40
B2 GAT Catalytic Bead Sensor - Standard Mode Measurements 41
B2 GAT Catalytic Bead Sensor - Standard Mode Measurements (cont.) 42
B3 GAT Catalytic Bead Sensor - Ml Mode Measurements 43
B3 GAT Catalytic Bead Sensor - Ml Mode Measurements (cont.) 44
C1 Polymer Adsorption Sensor Stability Data -
1992 Response - Uncalibrated 45
C2 Polymer Adsorption Sensor Stability Data -
1992 Response -1 Point Calibration 46
C3 Polymer Adsorption Sensor Stability Data -
1992 Response - 2 Point Calibration 47
C4 MOS 823 Sensor Stability Data -1992 Response - Uncalibrated 48
C5 MOS 823 Sensor Stability Data -1992 Response -1 Point Calibration ^49
C6 MOS 823 Sensor Stability Data -1992 Response - 2 Point Calibration 50
VII
-------
-------
SECTION 1
FOREWORD
This report summarizes the work performed
under contract #IV-0900-NAEX. This is the sec-
ond of two projects performed for the United States
Environmental Protection Agency's Environmental
Monitoring Systems Laboratory. It was written to
complement the first project report1 [EPA/600/R-
92/219]. The reader is advised to read the first pro-
ject report before reading this document.
These projects were initiated to improve the
understanding of the performance of lower cost
vapor sensor technologies used at underground
storage tank (UST) sites in both leak detection
equipment and site surveying instruments.
A solicitation was placed in the Commerce
Business Daily requesting vendors of gas sensors
to respond to the Environmental Protection
Agency's Environmental Monitoring Systems
Laboratory. The vendors who responded provided
the sensors for which performance measurements
are described in this report.
Quantifying gas sensor performance is key to
interpreting the sensor's response and determining
how the user may apply the technology. Sensor
properties that define a vapor sensor's performance
include:
1) sensitivity to a target gas or class of gases;
2) selectivity so that the sensor readings are
not misinterpreted;
3) response and recovery to the target gas in a
reasonable time period;
4) stability during the time of data collection;
5) precision and accuracy of the sensors; and
6) the dynamic range and detection limits
which the sensors operate.
By understanding these properties, the appropriate
sensor can be selected for a given environment
realizing that the same vapor sensor can work well
in one application and poorly in another.
There are a variety of vapor sensor technolo-
gies commercially available. However, the infor-
mation describing sensor performance is limited
and difficult to interpret. This is the case for the
application of vapor sensors used at (UST) sites.
These projects were funded to demonstrate an
alternative method of collecting and processing
sensor data that compare distinct vapor sensor
technologies. These projects were not designed to
fully characterize the vapor sensors under study.
The first project focused on measuring the per-
formance of the two most commonly used UST
leak detection vapor sensors: 1) the metal oxide
semiconductor sensor and 2) the polymer adsorp-
tion sensor.
The second project was designed to measure
the performance of catalytic bead sensors and fiber
optic chemical sensors. The project's scope was
changed when the fiber optic chemical sensors did
not become available. In place of measuring the
fiber optic chemical sensors, the properties of the
metal oxide semiconductor sensor and the polymer
adsorption sensors were re-measured to determine
their long-term stability characteristics.
-------
SECTION 2
INTRODUCTION
2.1 BACKGROUND
Over two million underground storage tanks
(UST) are currently being regulated by the EPA.
By the end of 1993 the vast majority of these tanks
were to be equipped with one or more of the
approved leak detection methods. Vapor monitor-
ing equipment, housed in monitoring wells sur-
rounding the UST, is one of the options for pro-
tecting the environment from leaking gasoline
tanks or fuel spills.
The concept behind vapor monitoring is that a
small leak of volatile liquid will generate a large
increase in product vapor concentration. By prop-
er placement of the monitor wells, taking into
account adequate soil porosity, the product vapor
will readily migrate to the monitoring wells.
There, the vapor sensor will detect the increased
vapor concentrations and sound an alarm.
In practice, however, this concept has been dif-
ficult to implement because of some misunder-
standings about the properties of commonly used
vapor sensors. These include how sensitive a sen-
sor is to leaking gasoline; how insensitive a sensor
is to the background gases found at UST sites; and
how to property calibrate a sensor.
The selection of test gases was based upon a
study performed by Geoscience Consultants, Ltd.,
in 19882. This study detailed the hydrocarbon
vapor concentration at 27 gasoline service stations
from three diverse geographic regions in the
United States.
Their findings indicated that:
• all the surveyed locations had some evi-
dence of underground methane and
gasoline vapor products.
• methane existed in high concentrations at
many locations.
• tracking butane concentrations would be
useful in detecting recent gasoline leaks or
spills.
• m-xylene was a large component of gaso-
line product (Table 2.1).
In addition to the above characterization of the
UST environment, the humidity level at under-
ground UST sites is considered to be near saturation3.
For this environment, several important sensor
properties need to be measured to determine the
appropriate selection of a vapor sensor. Clearly a
sensor needs to be sensitive to gasoline. To simpli-
fy testing and because the composition of gasoline
varies, iso-butane and m-xylene were chosen as
target gases to identify a possible gasoline leak.
Iso-butane was selected to represent the C4 - C6
class of hydrocarbons and m-xylene was chosen to
represent the benzene-toluene-xylene class of com-
pounds. Both classes of compounds are major con-
stituents in gasoline.
Table 2.1. Major Components of API PS-6 Gasoline2
Compound
2-Methybutane
M-Xylene
2,2,4-Trimethylpentane
Toluene
2-Methylpentane
N-Butane
1 ,2,4-Trimethybenzene
N-Pentane
2,3,4-Trimethylpentane
2,3,3-Trimethylpentane
3-Methylpenlane
0-Xytene
Ethylbenzene
Benzene
P-Xylene
2,3-Dimethylbutane
N-Hexane
1 -Methyl, 3- Ethylbenzene
1 -Methyl, 4-Ethylbenzene
3-Methylhexane
Percent Weight
8.72
5.66
5.22
4.73
3.93
3.83
3.26
3.11
2.99
2.85
2.36
2.27
2.00
1.94
1.72
1.66
1.58
1.54
1.54
1.30
Also, based on the Geoscience Consultants study,
methane and water vapor content were chosen as
sources of potential interferences. That is to say
that sensor response could be misinterpreted, caus-
ing false alarms or masking of a leak because of
changes in methane concentration or humidity lev-
els. It should be noted that there are several other
potential sources of interferences such as hydrogen
sulfide, and changes in oxygen and carbon dioxide
background levels that were not studied because
they were outside the scope of this project.
-------
2.2 PROJECT OVERVIEW
The goal of this study is to provide information
that can aid in understanding the capabilities and
limitations of commercial vapor sensors used at
underground storage tank (UST) sites. There are a
variety of vapor sensors used at UST sites for sur-
veying or leak detection. Each vapor sensor tech-
nology has distinct advantages and disadvantages.
However, product literature often does not provide
sufficient information for direct comparison of the
competing technologies. Also, there is a lack of
information pertaining to how specific vapor sen-
sors are affected by the UST site environment.
This study focused on the metal oxide semicon-
ductor (MOS) sensors4, the polymer adsorption
(PA) sensor5, and two commercially available cat-
alytic bead sensors.6-7 Mention of trade names or
commercial products does not constitute endorse-
ment or recommendation for use.
All of the above sensor types are non-specific.
That is to say they respond to a variety of com-
bustible or toxic gases. For example the metal
oxide semiconductor (MOS) sensor responds, at
various concentrations, to methane, propane, hexa-
ne, xylene, alcohol, carbon monoxide, and ethyl-
ene oxide. It is sensitive to the ppm range of any
one gas, but is not specific. Therefore it is difficult
to interpret its response in an undefined environment
The responses of these sensor technologies
have been studied by both their manufacturers and
university laboratories. However, the sensor
responses are generally reported as sensor signal
(i.e. resistance and mV) versus gas concentration.
For some sensor types, resistance increases with
gas concentration; for others it decreases. Some
sensors respond exponentially to increases in gas
concentration, and others linearly. With all these
variations, how to compare and correlate the sen-
sors' response to gas concentration is not always
straightforward.
Table 2.2 summarizes the behavior of the sen-
sors being studied. In order to compare the
responses of these sensors, their individual
responses have been modeled, using simple model
equations. The sensor output is then converted to
units of parts per million by volume (ppm). In this
form, the results are presented as what one would
see from an instrument's digital display. For exam-
ple when the sensor is calibrated with 1,000 ppm of
gas A, the digital display will read 1,000.
Table 22. Vapor Sensor Signal Output
Sensor Type Signal Output Response to Increasing
Gas Concentration
MOS Resistance (Q) Decreases exponentially
Polymer Adsorption Resistance (ft) Increases exponentially
Catalytic Bead volts (mV) Increase linearly
Accordingly, other sensor properties such as
sensor selectivity to another gas can be judged by
what the digital display reads. For example, when
the sensor is exposed to a second gas, 1000 ppm
gas B, the digital display can read zero or some
number. If the display reads zero, then the sensor
does not respond to gas B and is considered selec-
tive with regard to gas B. If the sensor was not per-
fectly selective, the number displayed would indi-
cate the degree of cross sensitivity to that particu-
lar gas. If the display reads 1,000, then the sensor
has an equal response to Gas B as it did to Gas A
and is not selective.
By testing these sensors in a laboratory to
measure their sensitivity and selectivity (cross sen-
sitivities) to methane, butane, xylene, and humidi-
ty a better understanding of how to use these sen-
sors can be obtained. This can minimize problems
that can arise at UST sites. Some examples of this
include:
1) The UST owner is unaware that the sensor
responds to methane. The sensor is placed
in a monitoring well with a background
environment containing methane.
When the sensor responds to the methane,
instead of gasoline vapors, it would cause
the monitoring system to alarm. This
would cost the UST owner time and
money to determine whether the tanks were
leaking or it was a false alarm.
2) The UST owner is unaware of the sensor's
response to changes in humidity. The sen-
sor is calibrated with dry gas, and placed in
a monitoring well.
The humidity level underground at UST
sites is considered to be near saturation3.
This damp environment could lead to false
alarms if the sensor's response to gasoline
vapors increases with increased humidity.
Worse yet, if the sensor's response to gaso-
line vapors is suppressed with increased
humidity, then a real leak could occur with-
out detection.
-------
Other sensor properties like speed of response,
stability, and reproducibility also need to be evalu-
ated with respect to the UST environment. Leaks
at USTs generally occur slowly, and site monitor-
ing is done on a time scale of days and not minutes.
Therefore, a fast sensor response, on the order of
seconds or minutes, is not essential. For recovery
time, the important requirement is that the sensor
recovers in a reasonable time period (days) after
situations such as an accidental spill occur. In this
case, if a sensor takes too long to recover from the
spill, the detection of a true leak could be masked.
Stability plays an important role in determin-
ing how a sensor is employed in UST monitoring.
If a sensor changes with time, independent of the
actual conditions, it could lead to false alarms
and/or not being able to detect a leak. The stabili-
ty measurement indicates how sensor properties
change with time, thus determining the need and
frequency of calibration.
The reproducibility of sensor properties from
one sensor to another is also essential in maintain-
ing instrument quality control. When a sensor fails
and is replaced, if the replacement sensor unknow-
ingly behaves differently, errors in monitoring a
site are very likely. By knowing the limitations of
reproducibility of various sensor types, steps can
be taken to properly calibrate replacement sensors
to assure the monitoring equipment's performance.
In the previous project1, MOS and PA gas sen-
sors were tested to measure their responses to com-
ponents of gasoline (iso-butane and m-xylene) and
how they were affected by interfering compounds
such as methane vapors and changes in humidity.
The sensor responses were characterized by sensi-
tivity, selectivity, and speed of response and recov-
ery and the data presented as a list of sensor speci-
fications. The conclusions of that program showed
that both sensor types, if used properly, had suit-
able properties for UST leak detection.
The work described in this report covers two
areas of investigation. The first area is to expand
the list of sensor technologies tested to include cat-
alytic bead sensors. The second area is to measure
properties of MOS sensors and PA sensors after
one year of continuous operation.
2.3 CATALYTIC BEAD SENSORS
Catalytic bead sensors are widely used for the
detection of combustible gases. Several companies
manufacture gas detection instruments using catalytic
bead sensors. A solicitation was placed in the
Commerce Business Daily requesting companies
that manufacture and/or sell gas sensors, either cat-
alytic bead and/or fiber optic gas sensors to
respond to the advertisement. Additionally, the
sensor had to be a stand alone product. Those com-
panies that met that criteria were asked to partici-
pate. Of the companies surveyed, GasTech (GAT)
and Detronics (DET) agreed to participate in the
project. Appendix A contains their respective cat-
alytic bead information. It is not the intent of this
project to offer commercial advantage of one or
more of these vendors. The objective is to provide
information on how this type of vapor sensor works
in a UST environment.
One of the primary uses of catalytic bead sen-
sors is for the detection of explosive levels of com-
bustible gases. The DET sensor is used to repre-
sent this type of catalytic bead sensor. The GAT
catalytic bead sensor is used to represent catalytic
bead sensors designed for use at UST sites where
the detection of combustible gases (volatile organ-
ic compounds) at lower concentrations is the objective.
The GAT sensor is designed to operate in two
modes. In the standard mode, it operates in a man-
ner similar to the DET sensor. In the methane
insensitive mode, the sensor operating voltage is
lowered, so the sensor no longer responds to
methane vapors but still responds to other com-
bustible vapors. This is considered a positive
development since methane at UST sites is viewed
as an interference gas2.
During this program phase, catalytic bead sen-
sors were tested using the same test procedures
established in the previous project. The sensors'
responses were characterized by sensitivity, selec-
tivity, and speed of response and recovery to select-
ed test concentrations of methane, butane, and
xylene. The test results are presented to allow
direct comparisons with the MOS and PA sensors.
2.4 METAL OXIDE SEMICONDUCTOR
SENSOR AND POLYMER ADSORPTION
SENSOR
Also during this work, the stability of the
Figaro MOS and Adsistor PA sensors was mea-
sured to determine the effects of continuous opera-
tion on sensor performance.
The Figaro MOS sensors, model numbers 812,
813, 822, 823, and the Adsistor PA sensors, were
re-tested using the same test procedures as in the
-------
previous project1. Since the end of the previous test results are presented so that the stability of sen-
project, over one year ago, the sensors have been sor performance can be evaluated.
powered in a clean humidified environment. The
-------
SECTION 3
EXPERIMENTAL PROCEDURE
The same procedures and tests were employed
as in the previous project1. The data presented
were collected using the CMRI laboratory auto-
mated gas sensor characterization facility. The
facility was designed to study the behavior of gas
sensors and to characterize their response in terms
of sensitivity, selectivity, speed of response and
recovery, and stability. A computer-controlled gas
delivery and data acquisition system (CDS), Figure
3-1, creates the test atmosphere in the sensor test
chamber and records the corresponding sensor
responses. The CDS controls and sets proper lev-
els of oxygen, nitrogen, and water vapor to create a
clean baseline environment through a network of
mass flow dilution modules. This clean air can
then be contaminated with up to five different
vapor compounds. For this study, the facility was
modified to independently set concentrations for
methane (CH4), butane (C4Hg), and m-Xylene
(C8H10). The CDS was set to maintain a constant
flow rate of one liter/minute.
A second gas system, delivering clean humid-
ified air, was used to maintain the sensor atmos-
phere when the sensor chambers were not connect-
ed to the CDS.
An on-line gas chromatograph was used to
verify the delivery of gases to the test chamber both
during and in between tests.
In the previous project, three test chambers
were built to house the sensors. One chamber was
built to test nine PA sensors and two chambers to
house 12 MOS sensors, 6 of each type. All the
materials used in the construction of the chambers
were chosen to minimize undesirable out-gassing
that might contaminate the test atmosphere. The
chambers also have the capabilities to power the
sensors and monitor their responses in accordance
with manufacturers' literature. The volume of each
test chamber was 1.2 liters.
The PA sensor chamber consisted of an alu-
minum plate and a glass-epoxy based printed cir-
cuit board mounted on standoffs. Standard clamp
pins were inserted into the circuit board for con-
DATA
ACQUISITION
SYSTEM
LABORATORY
COMPUTER
GAS CYLINDERS
f METHANE }
( ETHANE J
C i -BUTANE J
(CARBON "\
MONOXIDE J
SOLVENT VAPOR
BUBBLER
m-Xylene
'
,,
•^J
i
-^>
\
-^
SENSOR ELECTRICAL
RESPONSE MEASUREMENT
Network of mass flow
controllers and
valves for mixing
gases from cylinders
and bubblers .
GAS CYLINDERS
( OXYGEN )
C NITROGEN J
WATER VAPOR
BUBBLER
GAS
CHROMATOGRAPH
r*l
\
\
1 I Sensor 1
•^ V Chamber /
f PA 1
' — ^i Sensor 1
V Chamber /
^^^y
BACKGROUND GAS SYSTEM
Maintains chamber
environment to set levels
(i.e. O2 / N2/ humidity/
contaminant level)
Figure 3-1. CMRI Gas Sensor Characterization Facility.
-------
necting to the data acquisition unit, and mounting
the PA sensor. The PA sensors were soldered onto
clamp pins and their resistance measured with the
CDS multi-meter. Precautions were taken to
insure that the solder flux did not interfere with the
PA sensors, as per manufacturer's recommenda-
tions. A glass lid was used to complete the cham-
ber construction. Gas flowed into the chamber
through a feed-through in the bottom of the alu-
minum plate.
The two MOS sensor chambers contained six
822 and six 823 sensors and six 812 and six 813
sensors, respectively. Both consisted of an alu-
minum plate mounted with 12 MOS sensor sock-
ets. These 12 sockets were mounted to form a 5.75
inch diameter circle. Thermocouples were also
installed to monitor chamber temperature. As with
the PA sensor chamber, a feed-through was tapped
into the center of the aluminum plate for the gases
to enter into the chamber, and a glass lid was used
to complete the chamber housing.
The sensors were powered and measured
according to manufacturers' instructions. Sensor
heaters were all powered using a 5-volt power sup-
ply. The sensor bias voltage was maintained at 10
volts. Precision load resistors (RL=3920 ohm ±
1%) were installed in series with the sensor leads.
Sensor signals were measured by reading the volt-
age across the load resistor. All wiring was done
on the outside of the chamber to prevent interfer-
ence with sensor responses.
Test chamber temperatures were monitored
during testing. The PA sensor test chamber tem-
perature operated at room temperature, 22°C ±
1°C. The MOS sensor test chambers ran hotter, at
33°C ± 1°C, due to the local heating induced by the
MOS sensors' operating power requirements.
For over one year, the three test chambers,
housing the MOS sensors and the PA sensors, were
attached to the auxiliary gas system which deliv-
ered clean humidified air. The MOS sensors were
powered continuously during this time with the
exception of a half dozen temporary power out-
ages.
An additional chamber was designed and built
to house three DET sensors (model #61-8100E-01)
and three GAT catalytic bead sensors (model
#226530-01). The chamber was built to allow the
sensors to be powered and monitored in accor-
dance with manufacturers' literature. The chamber
consisted of a Hoffman gasketed enclosure with a
volume of 960 cm3. Holes were tapped into the lid
so that six sensors could be mounted, and so that
gases could enter and exit the chamber. A thermo-
couple was also installed to monitor chamber tem-
perature.
Both the DET and GAT catalytic bead sensors
were monitored using a wheatstone bridge config-
uration, Figures 3-2 and 3-3. The resistors used to
complete the wheatstone bridge were specified by
their respective manufacturers and mounted on the
external circuit boards shown in Figures 3-2 and 3-
3. These figures also show where the voltmeter
was connected to measure the sensor response. All
power and voltmeter connections were on the out-
side of the chamber to prevent interference with
sensor responses.
The catalytic bead sensors were powered inter-
mittently during the test program. This was done
to simulate the use of these sensors in portable
instruments. The sensors were turned on several
hours before being tested and powered off after
testing. The DET sensors were powered at 3.3
volts. The GAT sensors were powered at 6.0 volts
for standard operation and 3.8 volts for methane
insensitive operation.
Test chamber temperatures were monitored
during testing. Like the MOS sensor chambers, the
catalytic bead test chamber ran hotter, 36 ± 1°C,
due to the local heating induced by the sensors'
operating power requirements.
The relative humidity that a given set of gas
sensors is exposed to is dependent on the tempera-
ture of their test chamber. The CDS controls the
absolute humidity (ppm H2O) of the test gas. If the
absolute humidity is set at 15,000 ppm I^O, the
relative humidity of a test chamber at 22°C is 58%,
and at 36°C is 26%.
-------
Exterior
Electronics
Circuit Board
Power Supply
3.30 Volts @ 0.5 amps
DET —
Sensor Housing
Active ~ Reference
Catalytic Catalytic
Bead Bead
Element Element
Figure 3-2. DET Catalytic Bead Sensor Circuit Diagram (Wheatstone Bridge Configuration).
Exterior
Electronics
Circuit Board
Power Supply
Standard Mode - 6.0 Volts @ 0.5 amps
Methane Insensitive Mode - 3.8 Volts @ 0.5 amps
GAT
Sensor Housing
Common
(Red) 1 Reference
(Black)
Active ' Reference
Catalytic Catalytic
Bead Bead
Element Element
Figure 3-3. GAT Catalytic Bead Sensor Circuit Diagram (Wheatstone Bridge Configuration).
-------
SECTION 4
SENSOR CONSTRUCTION AND MODEL EQUATIONS
As in the previous project1, mathematical mod-
els were used to convert both non-linear and linear
sensor signals into gas concentration values (ppm).
It was not within the scope of this work to use best-
fit models, which would have more closely charac-
terized sensor response. The models chosen for the
PA sensors and catalytic bead sensors are the ones
suggested by their respective manufacturers5-6-7.
The model selected for the MOS sensors is the one
commonly cited in the literature8.
Tests were performed to determine which tar-
get gas (methane, butane, xylene) each sensor type
was most sensitive to. Then using the most appro-
priate model equation, the sensor's signal would be
converted to units of ppm target gas. Thus, if a sen-
sor was most sensitive to butane, its response
throughout all of the testing would be reported in
units of ppm butane.
The MOS sensor and PA sensor were most
sensitive to xylene. The catalytic bead sensors
were most sensitive to butane. For all cases, the
sensors were calibrated in a background of humid
air, with the absolute humidity set at 15,000 ppm
A summary of how the MOS, PA, and catalyt-
ic bead sensors are constructed and the model
equations used are presented below. The sections
on the MOS and PA sensors are reprinted for con-
venience from the first project report.
4.1 POLYMER ADSORPTION SENSOR
The Adsistor polymer adsorption sensor looks
like a small resistor, Figure 4-1. It is specially
coated to make it sensitive to gas vapors. The PA
sensor requires no power to operate and is moni-
tored by measuring its resistance like a common
resistor.
The base of the coating is a non-conductive,
resilient polymer which holds in place conductive
particles. The phenomenon of adsorption is the
basis for the sensor's sensitivity. In an ambient air
environment, the conductive particles, each inde-
pendently anchored to the polymer surface, are in
contact with each other forming an electrical path.
When a contaminant vapor comes in contact with
the particle surface, a mono-layer is adsorbed onto
the particle surfaces. Van der Waal's adsorption
forces (adhesion of gas molecules to the surface of
a solid) cause separation between each of the parti-
cles, increasing the electrical path's resistance. The
electrical resistance measured across a PA sensor is
determined by the amount and type of gas mole-
cule adsorbed to its surface9.
PA sensor data was collected by measuring the
sensor's electrical resistance. The resistance is
related to concentration for most gas vapor con-
centrations by equation 4.11.
Eqn.4.11
R =
where R = Measured resistance
Rb = Resistance in clean air
k = Gas constant at ambient temperature
c = Gas concentration (ppm)
The PA sensor resistance versus concentration
is reported to be a straight line when plotted on a
semi-log graph5.
The model was tested for xylene by exposing
the sensors to a xylene ramp of 100 ppm to 1000
ppm in 100 ppm steps. The resistance versus
xylene concentration curve is plotted in Figure 4.2.
This curve is not a straight line. This may indicate
that the sensor is not sufficiently sensitive to the
lower xylene concentration range.
Because the sensors did not respond to the
lower test concentration, a two-point calibration, at
100 and 1000 ppm xylene, was used to determine
Rb and k in equation 4. 1 1 . Solving equation 4. 1 1
fore yields equation 4.12 which is used to translate
the measured PA sensor resistance into a measured
gas concentration.
Eqn.4.12
The two-point calibration line is also plotted in
Figure 4.2.
4.2 METAL OXIDE SEMICONDUCTOR
SENSOR
The construction of a Figaro MOS sensor is shown
in Figure 4-3a. The sensor is primarily composed
-------
Conductive Polymer
for Electrical Contact
Metal Leads
Conductive Particles
embedded in Polymer Coating
for Vapor Sensing
Figure 4-1. Polymer Adsorption Sensor.
1000 -
8
CO
o
rr
100
Absolute humidity = 15,000 ppm H20
n PA sensor #8
— Two Point Model Fit
—i—
200
400
—r—
600
800
1
1000
Xylene Concentration (ppm)
Figure 4-2. Polymer Adsorption Sensor Response to Xylene.
10
-------
of tin oxide sintered on a small ceramic tube.
Noble metal wires are used to provide electrical
contact between the sintered tin oxide and the elec-
tronics used to measure its resistance. The noble
metal wires also provide mechanical support.
Through the center of the ceramic tube, a coiled
wire is positioned to serve as the sensor heater.
The MOS sensors require between 300 mW
and 600 mW of power to operate the sensor ele-
ment at elevated temperatures between 200°C to
450°C. By varying the composition of the sensor
element and/or the operating temperature, the sen-
sor's response to various combustible gases is
altered.
For this project, the sensors were powered and
measured according to the manufacturers' instruc-
tions. Sensor heaters were all powered using a 5-
volt power supply. The sensor bias voltage was
maintained at 10 volts. Precision Load resistors
(RL=3920 ohm ±1%) were installed in series with
the sensor leads (Figure 4-3b). Sensor signals were
measured by reading the voltage across the load
resistor.
The MOS sensors respond to changes in the
partial pressure of oxygen. At a set oxygen level,
oxygen is adsorbed on the surface of the gas sens-
ing MOS sensor. This adsorption of oxygen on the
semiconductor is strong enough to promote elec-
tron transport from the semiconductor to the
adsorbed oxygen. In the presence of a fixed oxy-
gen environment such as ambient air, an equilibri-
um state is achieved and the sensor electrical resis-
tance (baseline) is established. If the environment
is then contaminated with a combustible type gas,
a surface catalyzed combustion reaction occurs.
This reaction causes the surface adsorbed and neg-
atively charged oxygen to be reduced, returning the
shared electron to the semiconductor, and decreas-
ing the semiconductor's electrical resistance. The
relationship between the amount of change in
resistance to the concentration of a combustible gas
is non-linear and can be expressed by a power law
equation.
MOS sensor data was collected and converted
to sensor resistance using equation 4.21.
Eqn.4.21 R = RL (VB - VR)/VR
where R =
RL =
v —
VB ~~
Resistance (ohms)
Load resistor (3920 ohms)
Voltage bias (10 volts)
Voltage across RL
The resistance concentration curve was observed to
be approximately linear on a log - log plot.
Therefore, a power law model was adopted for
these sensors as seen in equation 4.22.
Eqn. 4.22 (a) Log(R) - log(Ro) = Blog(c)
(b) R/Ro = cB
where R = sensor resistance
c = gas concentration (ppm)
6 = power law slope
Ro = sensor resistance when c=l
The two parameters Ro and 6 are determined
by considering measurements taken at c =100, and
c =1000 ppm for the gas in question. Once the
parameters are determined, the sensor resistance is
translated into concentration by inverting equation
4.22 and as shown in equation 4.23.
Eqn. 4.23
A plot showing how the model corresponds to
the sensor response of an MOS 823 sensor is
shown in Figure 4-4.
4.3 CATALYTIC BEAD SENSOR
The catalytic bead sensors have been used for
decades in the detection of combustible gases, par-
ticularly methane in air in concentrations above
0.1% (1000 ppm). They use a small amount of
power (50 - 300 mW) to operate and are widely
used in portable gas detection instruments.
The catalytic bead sensor is composed of an
active and passive element. Both elements are
made by embedding a coiled platinum wire in a
porous ceramic such as alumina. The active ele-
ment is coated with a catalyst such as platinum.
The passive element is coated with an inert glass.
Both elements are powered to a specific elevated
operating temperature ranging from 300°C to
800°C
When a combustible gas such as methane
comes in contact with the elements, the element
coated with platinum bums the vapor causing local
heating. The glass coated element does not react
with the combustible gas. The platinum coil's
resistance changes as a function of temperature.
Thus, this local heating raises the platinum coil's
11
-------
•f 10 VOltS
Noble metal wire
19.5mm—H
100-mesnSUS316
Stainless steel
guaze (double)
Heater coil
Sensor element
Ni plated
brass ring
Ceramic base
Kovar pin
MOS (&
Sensor VT.
Load Resistor
(RL» 3.92 KQ)
Vort
Meter
a) Sensor Construction
reprinted from Figaro Literature
b) Measurement Circuit
Figure 4-3. Metal Oxide Semiconductor Sensor.
100000
CO
o,
§
CO
10000-
1000-
Absolute Huimidity » 15,000 ppm H2O
o 823 #4
- Two Point Model Rt
10
100
Xylene Concentration (ppm)
1000
Figure 4-4. MOS 823 Sensor Response to Xylene.
12
-------
resistance. The passive element serves as a refer-
ence element to compensate for environmental
conditions.
The two elements are connected into a wheat-
stone bridge circuit, so that small resistance
changes, due to heating by the gas combustion, can
be measured as a change in voltage, Figures 3-2
and 3-3.
Both Gastech (GAT) and Detronics (DET) cat-
alytic bead sensors have a linear voltage output
response with respect to changes in gas concentra-
tion. Both manufacturers suggest calibrating the
sensors using at least two points, a zero or baseline
value, and a set concentration value.
The DET sensors were powered at 3.3 volts.
The GAT sensors were powered at 6.0 volts for
standard operation and 3.8 volts for methane insen-
sitive operation. The lower power voltage for the
methane insensitive mode of operation of the GAT
sensors corresponds to a lower catalytic bead oper-
ating temperature. At this lower temperature, the
catalytic beads do not combust the more stable
methane molecules, but can still combust the other
volatile organic hydrocarbons.
Catalytic bead sensor data was collected by
measuring the voltage across the wheatstone
bridge. This voltage is related to gas vapor con-
centrations by a linear relationship, equation 4.31.
Eqn. 4.31 C = (Vo-Z)xG
where Vo = output voltage, mV
Z = output voltage in baseline clean air,
mV
G = sensor gain, ppm/mV
C = gas concentration, ppm
Sensor linearity was tested for butane by exposing
the sensors to a butane ramp of 50 ppm to 5000
ppm. Figure 4-5 is a graph showing the sensor
voltage output versus butane concentration. Also
plotted in Figure 4-5 is the calibration line. This
line was based on drawing a straight line through
the sensor output values in clean air and in air con-
taining 5000 ppm butane. For both conditions the
absolute humidity level was held at 15,000 ppm.
30
Absolute Humidity = 15,000 ppm H2O
>
E
D)
C
OS
CD
or
k.
o
CO
C
CD
CO
10
1000
2000
3000
4000
Two Point Calibration Line
Sensor Response
DET#6
5000
Butane Concentration (ppm)
Figure 4-5. DET Catalytic Sensor's Response to Butane.
13
-------
SECTION 5
TEST DESCRIPTIONS
Four types of tests were performed to charac-
terize sensor response. Each of the following tests
was designed to measure one or more specific sen-
sor properties:
5.1 INDIVIDUAL GAS-SENSITIVITY AND
SELECTIVITY TEST
This test measures a sensor's sensitivity and
selectivity (cross sensitivity) to individual test
gases, Figure 5-1. This test exposes the sensors to
individual test gases at five different concentra-
tions. The concentration ranges were 50,150,500,
1500, and 5000 ppm for methane and butane and
10, 30, 100, 300, and 1000 ppm for xylene. Each
concentration was held for thirty minutes before
proceeding to the next level. The sensors were
exposed to clean air for two hours between each
ramp.
Tests were performed at two humidity levels.
The first set was conducted at 15,000 ppm Hf).
This level was chosen to represent the humidity
present at underground storage sites (97% relative
humidity at 55°F). The second set was done in dry
air, 0 ppm Hf) (less than 50 ppm water vapor), to
simulate sensor response when exposed to dry cal-
ibration gases.
5.2 MULTIPLE GAS - SENSITIVITY AND
SELECTIVITY TEST
This test determines how the presence of mul-
tiple test gases affects a sensor's sensitivity and
selectivity (cross sensitivity), Figure 5-2. This test
creates a background test atmosphere composed of
500 ppm methane, 500 ppm butane, and 100 ppm
xylene in air containing 15,000 ppm H2O. During
the test, each gas is then individually increased to
10 times its background level for thirty minutes.
The methane and butane levels are raised from 500
to 5000 ppm respectively, and the xylene level is
raised from 100 to 1000 ppm.
5.3 MULTIPLE GAS - WATER VAPOR
SELECTIVITY TEST
This test measures sensor response to the
changes in humidity in the presence of multiple test
gases, Figure 5-3. This test creates a background
test atmosphere composed of 500 ppm methane,
500 ppm butane, and 100 ppm xylene in air con-
taining 15,000 ppm Hf>. The water vapor concen-
tration is then changed in sixty minute steps from
15,000 ppm, to 5000 ppm, to 1667 ppm, to 0 ppm
Hf> (less than 50 ppm water vapor), and then set
back to 15,000 ppm
This test was also performed at the increased
test gas concentrations 5000 ppm methane, 5000
ppm butane, and 1000 ppm xylene.
5.4 RESPONSE AND RECOVERY TIME
TEST
The Response and Recovery Time Test deter-
mines how fast a sensor responds to changes in gas
concentration, Figure 5-4. The tests were per-
formed in air humidified to 15,000 ppm H^O. The
sensors were measured at one minute intervals dur-
ing the test. The butane concentration changed in
thirty minute steps from 0 ppm, to 5000 ppm, to
500 ppm, to 5000 ppm and back to 0 ppm.
The response time is defined as the time
required for the sensor to reach 95% of its final
reading. For this series of tests, the final reading is
measured at 30 minutes after the new gas concen-
tration is introduced. The recovery time is defined
as the time needed for the sensor to reach 95% of
the total change in the sensor reading. For exam-
ple when changing from 5000 to 500 ppm butane,
the 95% recovery time would be the time interval
until the sensor reads below 725 ppm butane.
14
-------
10000:
Concentration (ppm)
-* o
-j. O O
-•• 0 0 0
fl* 1 J 1 1
Absolute Humidity set at 15,000 ppm H2O or 0 ppm H20
«••* .-- 1
•
•
«J — •
• i
• !
r-* .—'
• j
J J
• i
• *
,,-n . , , , ,.,,,/ , ?
**"i
Actual Gas
Concentrations
Butane (ppm)
Xylene (ppm)
0 0 60 1 20 1 80 240 300 360 420 480 540 600 660 720
Time (min.)
Figure 5-1. Individual Gas - Sensitivity and Selectivity Test.
10000:
E*
Q. 1000 -
c !
o
3
+— •
d
-------
10000 :
?
£ 1000:
c '
0
V-
OJ
•+-•
c
0>
O1 00 -
c
o :
O ;
m -
500 ppm Methane, 500 ppm Butane, and
1 00 ppm Xylene are maintained constant
throughout this test.
15,000 i 5,000 i 1,667 i 0 | 15,000
ppm H2O [ ppm H2O [ ppm H2O| Ppm H2O [ ppm H2O
i i i i
0 30 60 90 120 150 180 210 240 270 300
Time (min.)
Actual Gas
Concentrations
- - - - Methane (ppm)
Butane (ppm)
Xylene (ppm)
Figure 5-3. Multiple Gas - Water Vapor Selectivity Test.
10000
E
Q.
a.
c
.0
03
•*-•
c
o
O
1000-
100
Absolute Humidity = 15,000 ppm H2O
Actual Butane
Cone, (ppm)
-10 0 10 20 30 40 50 60 70 80 90 100 110
Time (min.)
Figure 5-4. Response and Recovery Time Test.
16
-------
The scope of this project was limited to data
collection and presentation in a laboratory environ-
ment. It does not include a tutorial as to how to
maximize the use of this data, or how to apply this
data to all situations. As such, specific value judg-
ments on sensor performance being good or bad,
which are application specific, are not made.
6.1 CATALYTIC BEAD SENSORS
The set of tests, described in Section 5, covers
a range of conditions that vapor sensors may be
expected to encounter at UST sites. Tables 6-1 and
6-2 summarize the performance of the DET and
GAT catalytic bead sensors as a list of sensor spec-
ifications. Appendix B contains the test responses
for the individual catalytic bead sensors.
The test results are presented as if each sensor
were operating in an instrument with a digital dis-
play. For the catalytic bead sensors, the sensor dis-
play reads ppm butane. The butane readings were
determined by the sensor response and calibration
curves described in section 4.3.
Each sensor was individually calibrated for
butane. This was accomplished by exposing the
sensor first to clean humid air and then to humid air
containing 5000 ppm butane. For both calibration
points, the absolute humidity was set at 15,000
ppm H2O.
The data are presented in the same manner as
the previous study1 and are the average of three
DET sensors and three GAT sensors. The GAT
sensors were tested in two distinct operating
modes: 1) Standard mode and 2) Methane insensi-
tive (MI) mode. In the standard mode, the sensors
are powered at 6 volts (sensor I.D. #1 - #3); in the
MI mode, they are powered at 3.8 volts (sensor I.D.
#4 - #6). [The lower operating power of 3.8 volts,
reduces the sensor's operating temperature to pre-
vent the combustion of methane]. The data are
reported as the average measured sensor response
along with the standard deviation and percent stan-
dard deviation.
Table 6.1 reports the results of individual test
gases with regard to sensor:
• value and variation of model parameters
• sensitivity to butane
• humidity effects on butane sensitivity
• cross sensitivity to xylene and methane
• response time
• recovery time
Table 6.2 lists the results of multiple test gases
with respect to the effects of humidity and cross
sensitivities on sensor response to xylene and
methane.
6.1.1 Reproducibility
Reproducibility is a measure of how much sen-
sor performance varies from one sensor to the next.
It is an important variable in drawing meaningful
conclusions from the data. For instance, a large
variation in butane sensitivity (gain) would make it
difficult to determine the expected sensor response
to butane during field operations. For example, it
might be assumed that two instruments using the
same type of sensors would produce identical read-
ings. They are calibrated in the field using only
one-point. If one of the sensors had a higher sen-
sitivity, the calibration would not correct for this
difference, and one sensor would yield a higher
reading than the other.
For this project, each sensor was individually
calibrated. The reproducibility of each set of three
sensors can be inferred by the size of the standard
deviation or percent standard deviation. The larger
the percent deviation the less reproducible the sen-
sors in that set are.
The percent deviation of each sensor set's gain
and zero values is shown in Table 6.1. The gain of
the DET sensors varied ±4%, while the GAT sen-
sor's gain varied ±6% in the standard mode and
±10% in the MI mode. The sensor output voltages
in clean air (zero value) were close to zero for all
the catalytic bead sensors. These small changes in
signal correspond to large percent deviations rang-
ing from 37% for the DET sensors to 165% for the
GAT-MI sensor.
17
-------
(0
c
o
T5
o
'o
o
Q.
CO
C
o
CO
CO
o
IK
CO
T5
O
CD
d>
i>!
ZE CO co
to «
"DO?
CO <0 co
o ^8 c?
= 11
CO %&
S 5
„!
§ ™ "6
co lg g>
k 11
0 § a:
n
OJ
55
^
2
CO
CD
CD
55
.
Q
2
CO
CD
9
CD
i
55
Q
2
CO
CD
1
<
Except where noted, the absolute
humidity was held at 1 5,000 ppm H2O
Butane Model Parameters
b
to
8
CO
CO
00
*
CM
m
CO
1
CM
O)
§
CM
co
CM
CO
I
00
o
CM
55
>•
O
2
CO
CD
I
1
55
.
Q
2
CO
CD
1
I
55
O
2
CO
CD
C?
1
<
Gain (ppm butane/mv)
Zero (mV)
Sensitivity
Butane Readings (ppm)
Calibrated at 0 and 5000 ppm Butane
o
o
CO
3
CO
oo
0
S
CM
CM
CM
1
to
CM
O
•
o
0
o
8
18
00
CO
o
*~
at
CM
Jo
Sj
£
£
CM
in
IE
3
o>
CO
CO
«
o
o
o
s
i
CM
O)
JI
00
00
in
CM
CO
i
§
«
8
CM
CO
CO
8
O
o
o
s
o o o o o
in in o o o
•r- in m o
••- in
I
1
CD
ta
m
£
>
6
Q
2
CO
CD
1
I
i
2
CO
CD
1
I
2
CO
CD
CD
<
Effect of Humidity
Butane Readings (ppm) in Dry Air
(Absolute Humidity < 50 ppm H2O)
i
CO
CO
!
3
s
0
s
CM
CM
CM
CM
CM
Si
O)
o>
1
o
o
0
s
s
oo
CM
O
CO
§
o
CO
s
1
5?
«
T"
^^
^^
55
to
-
CM
1
«
to
CO
CM
i
1
o
«
CM
CO
9
i
CM
O>
£
*
«
in
o
CO
5j
00
CO
So o o o
in o o o
I
1
£
CD
OJ
3
m
£
>
6
Q
2
CO
CD
1
I
55
I
2
CO
CD
1
CD
55
cS
O
2
CO
CD
?
5
<
Cross Sensitivity
Butane reading (ppm) when:
«
o
o
f
CM
CM
CM
CO
§
oo
CO
0)
CM
5j
5
CM
CO
O
3?
N.
CO
CO
CM
*
O
CM
CO
2
58
.
2
CO
CD
1
%Dev.
d
2
CO
CD
CD
|
cS
O
2
CO
CD
?
5
<
5000 ppm Methane is delivered
1 000 ppm Xylene is delivered
Response & Recovery
95% Response Time (mln.)
•
1
^»
s
3
o
CM'
•
.
^
•
.
^
•
•
^
•
•
V
0 to 5000 ppm Butane
500 to 5000 ppm Butane
95% Recovery Time (min.)
•
•
^
•
•
^
•
.
^
•
.
^
s
o
oo
CO
*
O
Tf
CO
co
5000 to 500 ppm Butane
5000 to 0 ppm Butane
18
-------
O
a.
CD
CO
C
CO
CD"
c
X
CD
C
CO
•«— >
«
' 00 ^
CO CO °
Q *- ?
c * =§
ail
5 is
O CL <
0>
T3 _
O (A
2 §
B §5
CO > »
£ 2 °
Q *- oi
W oS .E
m "O «J
w eg
i-2
in
"«
o
co
>«
CO 2
f/J co °
8*f
111
til*
Q 0. <
•s-s
E 2
a-
~ E
Q ex
8 °-
IO g
O Iff
§T
9 &
a S
3 E co
m § CD
S X ffl
fi « «
•03 V->
1 1 w
II ,2
o® ^
.1
CD
Q
1
55
S
Q
2
to
CD
5^
0 Q.
Is
I
55
§s
o
2
co
CD
§>?
CD °-
Is
I
ss
35
2
CO
o
8
o
g
o
8
1
o
s
in
8
m
55
8
J5
"t
00
00
8
CO
C\J
0
0)
0
o
o
8
§
1
8
8S
8
o
in
r^
oo
in
«
CO
CO
CM
O)
5?
8
CO
0
O)
o
o
o
8
8
1
8
s?
o
in
CO
o
in
<£
CNJ
CO
CO
CM
O
as
8
CO
o
O)
o
8
o
8
8
1
8
8?
o
In
CO
8
«
E
3
3
g
s
00
r»*
3
8
o
8
1
1
§
3?
o
in
CO
o
m
5?
CM
CO
CO
%
8?
10
R
<*
O)
o
o
1
1
i
$
i
m
CM
T—
U)
CN
U)
8S
•^~
in
O)
1
Jf
O)
O)
1
o
o
1
§
i
i
1
s
r»
00
m
8?
CM
CO
CO
CM
Ol
8?
m
S
?
8
o
8
o
8
i
o
CM
|
CM
8
CO
§
*
£
0)
(O
f*
£
cS
U)
s
8
o
§
§
i
o
s
3S
%
o
m
r*.
oo
in
55
CM
CO
CO
s
3?
m
S
•-
r--
00
S
8
§
§
i
0
CO
85
a>
to
oo
00
CM
00
5
85
8
in
8
00
s
g
CM
in
S
^1
.
SB
85
CM
CO
CO
CM
O)
85
in
S
M-
o>
8
§
I
o
o
s
0
•*•
CM
£
CM
0>
in
o
CD
(O
CO
X
CM
S
CO
00
I
00
8
a>
r^.
CO
r^
8
1
s
o
o
8
o
£
a>
in
S
$
X
s
s
s
1
in
S
S
s
8
8
1
I
0
CO
*
O)
s
1
i!
§
S
3
g
s
3
eo
o>
8
1
1
i
T-
O
CM
£
3
g
§
s
1
o
o
s
o
*•
CM
19
-------
The variations in sensor response have an
added component due to the truncation of data col-
lected. The Gas Delivery System truncated voltage
readings to one mV. This measurement limitation
dictated the smallest butane reading (ppm) incre-
ment that could be made for each sensor type. For
the DET sensor, a one mV change in sensor
response corresponded to a reading change of 326
ppm butane. For the GAT sensors in the standard
and MI mode, a one mV change in sensor response
corresponded to a change in butane readings of 50
and 32 ppm, respectively. Thus, a difference of one
mV in the measurement between sensors, affects
the DET sensors a factor of 6 to 10 times more than
the GAT sensors. For example, in Table 6.1 the
average response to 500 ppm butane at 15,000 ppm
HjO, for the DET sensors was 549 ± 204 ppm
butane. The standard deviation of 204 is slightly
more than 0.5 mV. Thus, a large component of this
variation could be caused by data truncation. Due
to this effect, it is difficult to determine repro-
ducibility at the lower gas concentration levels.
The DET and GAT sensors show similar vari-
ations in sensor response at gas concentration lev-
els above 1500 ppm butane; the GAT - MI sensors
had a slightly wider range of responses. For exam-
ple at 1500 ppm butane, the percent standard devi-
ations are listed below:
GAT - Standard 3.2%
GAT-MI mode 20.1%
DET 3.7%
6.1.2 Sensitivity
Sensitivity refers to the amount of gas the sen-
sor responds to. Does the sensor respond to 10
ppm, 100 ppm, 1000 ppm of a specific gas? The
individual gas sensitivity and selectivity tests were
used to determine this parameter for the selected
test gases.
The DET and GAT Standard sensors respond-
ed to methane, butane, and xylene. The GAT - MI
sensors responded to butane and xylene. All three
sensor types were most sensitive to butane and
were modeled and calibrated for butane responses.
The butane sensitivity data that is presented in
Table 6.1 indicates the DET and GAT Standard
sensors are sensitive to butane concentrations as
low as 50 ppm while the GAT - MI sensors respond
to butane concentrations of 150 ppm. It should be
noted however, that at these concentrations, the
large standard deviations indicate lower sensor
reproducibility. For all three sets of sensors, sen-
sor-to-sensor reproducibility increases consider-
ably at butane concentrations of 1500 ppm and
above.
6.1.3 Humidity Response
Tests were performed to determine how the
sensors' butane response was affected by changes
in the humidity level. This is presented in Table 6.1
under the heading "Effect of Humidity". The sen-
sors were exposed to the same butane concentra-
tions, but the absolute humidity was changed from
15,000 ppm HjO to 0 ppm t^O (less than 50 ppm
water vapor). If the sensors were not affected by
this change in humidity, their readings should be
similar to those in the previous section.
The butane responses of the DET and GAT
Standard sensors are both affected by changes in
the level of humidity. As the absolute humidity is
changed, from 15,000 ppm to 0 ppm HjO, the sen-
sor signals are lowered uniformly throughout the
concentration range. These effects, tabulated in
Table 6.1, are most noticeable at butane concentra-
tions below 1500 ppm. Here the negative values
indicate that the sensors' responses have dropped
below what had been the zero value calibration
point. The sensor still responds, but the changes in
humidity have lowered the raw sensor output
below the calibration setting. If the sensor had
been calibrated with dry gas and placed in a damp
environment, the reverse would be seen. In this
case, the sensor's output, and thus its sensitivity,
would be higher than expected.
The butane responses of the GAT-MI sensors
were less affected by these same changes in humid-
ity, Table 6.1. Their average zero reading was low-
ered, but the remaining readings showed little
change.
The results of the multiple gas water selectivi-
ty test, quantified in Table 6.2, show the DET and
GAT Standard sensors' reading, decreasing from
40% to 50% as the absolute humidity is reduced
from 15,000 ppm to 0 ppm H2O. Also to be noted
are variations of 20 - 37% in sensor-to-sensor
response. The GAT-MI sensors showed a change
in sensor reading of 20% with the sensor-to-sensor
variations as high as 69%. This indicates that, at
these concentration levels, the catalytic bead sen-
sor responses are affected by changes in humidity.
The sensor-to-sensor variations can be seen in
Figure 6-1. This is a plot of the response of three
20
-------
10000
Q.
C
.g
T3
is
I
o
O
1000-*
100
500 ppm Methane, 500 ppm Butane, and
1 00 ppm Xylene are maintained constant
throughout this test.
Sensor Response (ppm)
DET #4
DET #5
DET #6
Concentration (ppm)
15,000 j 5,000 i 1,667 | 0 j 15,000
ppm H2O ] ppm H2O [ ppm H2O [ ppm H2O [ ppm H2O
30 60 90 120 150 180 210 240 270 300
Time (min.)
Figure 6-1. Catalytic Bead Sensor Response to Changes In Humidity, in a Mixture of Methane,
Butane, and Xylene.
DET sensors, computed as ppm butane, during the
multiple gas water selectivity test. The plot shows
how each sensor behaves differently throughout the
test. Sensor #4 readings change when the absolute
humidity changes from 15,000 ppm to 5000 ppm
and 1677 ppm to 0 ppm, but not when it changes
from 5000 ppm to 1677 ppm. Sensor #5 readings
change with each change in humidity, and Sensor
#6 only changes at 15,000 ppm to 5000 ppm
humidity step change.
This random effect seen in the changes in
absolute humidity affecting the butane readings, is
also representative of the GAT Sensor behaviors in
both the standard and MI mode.
These results led to a multiple gas water selec-
tivity test being performed at increased test gas
concentrations. Instead of testing the sensors at
500 ppm methane, 500 ppm butane, and 100 ppm
xylene, while varying the humidity, the sensors
were tested at 5000 ppm methane, 5000 ppm
butane, and 1000 ppm xylene. The results from
these tests are tabulated in Table 6.2. At these con-
centration levels, the effect in changes in humidity
is smaller, 23% for the DET sensor, and around 1%
each for the GAT sensors.
6.1.4 Selectivity
While sensitivity is the measure of how a sen-
sor responds to the target gas of interest, selectivi-
ty or cross sensitivity is the measure of how the
sensor responds to all other gases. Although the
testing performed does not systematically define
this parameter for all gases, the data does indicate
how the sensors will respond to several important
gases present at a UST environment.
If the sensors were perfectly selective for
butane, when xylene or methane were present, the
sensor display would read zero. If the sensor was
not perfectly selective, the number displayed
would indicate the degree of cross sensitivity to
that particular gas. By understanding this sensor
property, one can better interpret field data.
21
-------
The average cross sensitivity response to
methane for the DET and GAT standard sensors
was basically the same. In the presence of 5000
ppm methane, the DET and GAT standard sensors
read 2826 ppm butane and 2739 ppm butane,
respectively, Table 6.1. For xylene, the DET sen-
sors have a larger cross sensitivity than the GAT
standard sensors. In the presence of 1000 ppm
xylene, the DET sensor read 1736 ppm butane, and
GAT standard sensors read 1062 ppm butane,
Table 6.1.
The GAT-MI sensors showed no cross sensi-
tivity to methane, reading zero in the presence of
5000 ppm methane. The GAT-MI sensors did
respond in the presence of 1000 ppm xylene, read-
ing 1845 ppm butane. A comparison of a GAT
sensor's response to a gas concentration ramp test,
powered in both the standard and MI mode is plot-
ted in Figures 6-2 and 6-3, respectively.
The DET and GAT sensors' response to multi-
ple gases is shown in Table 6-2. The DET and
GAT standard sensors show cross sensitivity to
both methane and xylene, while the GAT-MI sen-
sors again showed zero response methane, and a
cross sensitivity to xylene.
In the presence of 500 ppm methane, 500 ppm
butane, and 100 ppm xylene, the DET and GAT
standard sensors read 1090 ppm and 923 ppm,
respectively. The GAT-MI sensors read 503 ppm.
When the methane level was increased to 5000
ppm, the DET and GAT standard sensors readings
increased to 3478 ppm and 3264 ppm respectively
while the GAT-MI reading remained unchanged at
503 ppm. When the butane concentration was
raised to 5000 ppm, all the sensors responded with
readings on the order of 5300 ppm. When the
xylene concentration was increased to 1000 ppm,
again all the sensor readings increased. The DET,
GAT standard, and GAT-MI sensors read 2715
ppm, 1769 ppm, and 2013, respectively. Figure
6.4 displays these results showing the response of
the catalytic bead sensors, computed as ppm
butane, during a multiple gas sensitivity and selec-
tivity test.
6.1.5 Speed of Response and Recovery
The DET and GAT sensors' responses to
changes in butane concentration are plotted in
Figures 6-5 and 6-6. The measured response times
can mainly be attributed to the one-minute flush
time of the test chamber.
The test chamber response time is calculated
by dividing the volume of the test chamber (0.96
liter) by the flow rate (one liter/minute). Assuming
uniform mixing, the time to reach 95% and
99.99% changes in gas concentration would be
2.88 and 4.80 minutes, respectively.
With the given test conditions, sensor response
times could not be measured directly for the faster
sensor responses, but they can be estimated to be
under one minute based on equation 6.1.
Eqn.6.1 Rg = (Rm2 - R,2)"2
where
Rs = Sensor Response Time
Rm = Measured Response Time
Rc = Chamber Response Time
In general, the catalytic bead sensors' response
and recovery times were under one minute. The
exceptions for the DET sensors' recovery time of
3.84 minutes, and the GAT-MI sensors' of 2.54
minutes could be attributed to system measure-
ment truncation.
6.2 STABILITY DATA FOR THE METAL
OXIDE SEMICONDUCTOR SENSOR AND
POLYMER ADSORPTION SENSOR
The MOS and PA sensor stability data were
analyzed in a manner that highlights the perfor-
mance of an instrument that uses these sensors in
the field. The MOS and PA sensors were re-tested
after one year of continuous operation in clean
humidified air. The data is presented in Tables 6.3
- 6.7 to allow for sensor performance in 1991 to be
directly compared to sensor performance in 1992.
Individual sensor stability data can be found in
Appendix C. If these sensors were perfectly sta-
ble, their response during this phase of testing
(1992) would be identical to their responses in the
previous project (1991).
Sensor drift can be described by 1) a change in
the sensor's baseline output or zero drift, 2) a
change in the sensor's sensitivity response or gain,
or 3) a combination of both zero and gain drift.
Zero drift is simply a shifting, up or down, of the
sensor's sensitivity curve. Gain drift indicates to a
change in slope of the sensor's sensitivity curve.
Either a zero or gain drift can be corrected with a
calibration at one gas concentration (one-point cal-
ibration). In the case where both the zero and gain
drift occur, a calibration using at least two gas con-
22
-------
10000 ;
!
1
Q. 1000 :
Q i
c '.
O
+3
CO
£ 100-;
CD :
o
c
o
O
10 1
Absolute Humidity = 15,000 ppm H2O
***^ ^8&k
' • t
¥fy '•
r* f*
• •
!** i
• i jw
• • fr
too I
i <> *& $8SSSB>
• J
^_p _j
«KS
j
i
f^
|
P
I
i
P
j
" Actual Gas
Concentrations
& J ,. AL^ /•fcMfc-MfcX
n^ Adh^b,*. / \
^ butane (ppm)
V*. .1 / \
Xylene (ppm)
Sensor Response
* GAT-Std #2
-60 0 60 120180240300360420480540600660720
Time (min.)
Figure 6-2. GAT Catalytic Bead Sensor - Standard Mode - Response to Methane, Butane, and
Xylene Concentration Ramps.
10000-
"c"
GL
>— -
c !
o
CO
Sz 100 :
c :
CD :
0
c
o
o
4 A
10 i
•i .
Absolute Humidity = 15,000 ppm H2O
•-» ffy
i
f-1"
!
r^
i
•
•
•
j
•
fS&fr
IT"
—i
•
^
J
r
•
i
>
^i
«
^W^
«d>
!
Actual Gas
Concentrations
ft J 4*4L« «%M«^« fin. an-irmA
^ r» * / \
butane (ppm)
Xylene (ppm)
Sensor Respone
* GAT-MI #5
-60 0 60 120180240300360420480540600660720
Figure 6-3. GAT Catalytic Bead Sensor - Methane Insensitive Mode - Response to Methane,
Butane, and Xylene Concentration Ramps.
23
-------
10000
Q.
a.
I
o
O
1000
NHHUEBI
100
10
Absolute Humidity = 15,000 ppm H2O
7
Note: GAT Ml #5 does
not respond to methane
IHIBUI Mil I
Actual Gas
Concentrations
•--• Methane (ppm)
Butane (ppm)
—— Xylene (ppm)
Sensor Response (ppm)
D GATStd#1
GAT Ml #5
o DET#5
60 90 120 150 180 210 240 270 300 330
Time (mln.)
Figure 6-4. Catalytic Bead Sensor Response to Mixtures of Methane, Butane, and Xylene.
10000
c
o
I
8
8
1000-
100
Absolute Humidity = 15,000 ppm H2O
Actual Butane
Cone, (ppm)
Sensor Response
a DET #5
* GAT STD #2
O GAT Ml #5
-10 0 10 20 30 40 50 60 70 80 90 100 110
Time (min.)
Figure 6-5. Catalytic Bead Sensor Butane Response and Recovery Plot.
24
-------
10000-
Q.
2 1000 •
CD
o
o
o
100
Absolute Humidity - 15,000 ppm H2O
DD
aaaaaaaaaaaaaaaag a
Actual Butane
Cone, (ppm)
Sensor Response
a DET#5
* GATSTD#2
O GAT Ml #5
6
20 25 30 35 40 45 50 55 60 65 70 75 80
Time (min.)
Figure 6-6. Expanded View - Catalytic Bead Sensor Butane Response and Recovery Plot.
centrations (two-point calibration) would be
required to re-characterize a sensor.
Zero drift is seen in the MOS and PA sensors
by a change in the Ro and Rb values, respectively.
A change in gain is seen by a change in the B and
K values, respectively. These parameters are pre-
viously defined in sections 4.1 and 4.2.
It should be noted that a change in sensor sen-
sitivity to a target gas can also affect the sensor's
cross sensitivity to other target gases and interfer-
ence gases. For example, if a sensor becomes less
sensitive to a target gas, but its sensitivity to an
interference gas remains stable, then its cross sen-
sitivity to the interference gas has increased.
The MOS and PA sensor stability data is pre-
sented in tabular form, Tables 6.3 - 6.7. Each table
contains four sets of readings which are arranged to
indicate if and how the sensors' responses have
changed with time.
The first column lists the specifications
obtained from the previous project (1991
Response). Each sensor was individually calibrat-
ed at 100 ppm and 1000 ppm xylene (two-point
calibration).
The next three columns list the sensor's
responses collected during this project (1992
response), but analyzed in three distinct ways.
* Uncalibrated
* One-Point Calibration
* Two-Point Calibration
Each method takes the measurements per-
formed during this project and uses a different set
of calibration coefficients to generate a set of read-
ings to indicate the extent of sensor stability or
drift. Each of these methods is described below.
Uncalibrated
This second set of readings was generated
using the calibration constants obtained from the
1991 data. This set of readings shows what an
instrument would display after one year of opera-
tion, without calibration. The sensor's stability can
be judged by comparing these readings to those
from the sensor's 1991 Response.
One-Point Calibration
This third set of readings is obtained by adjust-
ing one of the two model parameters based on the
1992 measurement at 100 ppm xylene. This is
equivalent to a one-point calibration in the field. If
25
-------
the sensor readings are comparable to the sensor's
1991 Response, then only a zero drift has occurred.
If the sensor readings are not comparable, then a
change in sensor gain has also occurred.
This set of readings also shows that the sensors'
responses are basically equivalent to the 1991
Responses. This confirms that, primarily, a zero
drift has occurred.
Two-Point Calibration
This fourth set of readings is obtained by
adjusting both model parameters based on the 1992
measurement at 100 ppm and 1000 ppm xylene.
This calibration, performed in the same manner as
the previous project, completely characterizes the
sensor response. Thus, comparisons between this
set of readings and the sensor's 1991 Response
determine the extent of changes in the sensor's
cross sensitivities to methane, butane, and humidity.
6.2.1 POLYMER ADSORPTION SENSOR
Table 6.3 shows the stability data for nine PA
sensors. Comparing the PA sensors' 1991
Response to the Uncalibrated 1992 response shows
the sensors have become less sensitive to xylene.
Instead of reading 1000 ppm xylene in the presence
of 1000 ppm xylene, the average sensor reading
was 624 ± 94 ppm xylene - a loss in xylene sensi-
tivity of about 37%. This decreased sensitivity is
seen throughout the set of readings as negative
numbers. These negative numbers would have
translated into a zero response for a leak detection
system. This means the lower detection limit has
increased from 300 ppm xylene to 1000 ppm
xylene over the course of one year of continuous
operation.
The one-point calibration set of readings
shows an improved sensor xylene response, read-
ing to 1056 ± 44 ppm xylene in the presence of
1000 ppm xylene. Also, the one-point calibration
data shows mat the sensor's response with respect
to humidity, methane, and butane cross sensitivities
is comparable to that of the 1991 Response. This
indicates that a zero drift has occurred and one-
point calibration is sufficient to re-characterize the
PA sensors.
The two-point calibration set of readings
shows that the model parameters have only
changed slightly over the year, and that the varia-
tion in sensor responses is approximately 11%.
6.2.2 METAL OXIDE SEMICONDUCTOR
SENSOR
The MOS 823 Sensor was chosen for illustrat-
ing the behavior of the MOS sensors tested for sev-
eral reasons: First, the test results document that
the MOS 812, 822, and 823 sensors all have com-
parable responses, considering the statistical
spread in their respective responses. Second, the
manufacture of the MOS 812 sensor has been dis-
continued, being replaced by the 822 model.
Third, the 822 and 823 sensors are described by the
manufacturer as being the same sensor, but pack-
aged differently. Finally, the MOS 813 sensors are
very sensitive to methane and are of limited use for
monitoring UST product leaks.
Table 6.4 shows the stability data for six MOS
823 sensors. Comparing the 1991 Response to the
Uncalibrated 1992 Response, shows that the per-
cent variation in the sensor's xylene response has
greatly increased. In the presence of 1000 ppm
xylene, the sensor's reading has changed from 1000
± 0 ppm xylene to 886 ± 464 ppm xylene. The sen-
sors appear to be unstable. Some sensors became
less sensitive, and others more.
One-point calibration yielded a sensor reading
of 1399 ± 560 ppm xylene in the presence of 1000
ppm xylene. This indicates that the characteristic
model changed over time as well as a baseline drift.
The two-point calibration results show that the
sensor's model parameters have changed. The
slope parameter, B, has increased in value indicat-
ing an increased sensitivity to xylene. The data
also indicates that the sensors have become slight-
ly more sensitive to changes in humidity, the
humidity interference increasing from 56% to
63%, and less sensitive to butane, the xylene read-
ings decreasing from 793 ppm to 404 ppm.
These data suggest that a two-point calibration
is the minimum needed to characterize this sensor
type after a year of continuous operation.
26
-------
o
0)
c
0)
to
0)
c
CO
to
o
lo
CO
^--
CO
k_
o
0)
o
CO
c
o
o
*
0>
.>»
o
Q_
CO
CD
*
jjc _^^ ^x
10 •Z3 ^^
^L ^Q o
a "5 §
CM ^ I*
o> it* o
r- D_ O
NS
• T-
*2 ^ <§>
*""
Q> 13
g §
O (Q
8* .S
o> =
CM O
O) c
o> 3
CD
C
Q) CD
(0 c >«
c o x
III
0) co g
CC 0 o
O) o
O> CM o
T- *~
iQj
£
s?
1
2
CO
CD
f
1
<
s?
Q
2
CO
CD
CD
£5
^
O
2
CO
CD
D)
5
CD
1
^
£
2
CO
CD
9
CD
^J
<
£
CM
oo
CO
CM
£
8
CM
O
00
CO
oo
o
CM
£
8
co
CM
«£
£
CO
8
00
oo
0)
CM
S?
O
3
co
g
§
CO
CO
en
CM
8?
o
3
CO
I
S?
1
2
CO
CD
2
1
*
55
O
2
CO
CD
CD
t§
^
1
2
CO
CD
CD
I
a-
1
2
CO
CD
CD
^
**•
*
-
1C
g
-
CO
g
o
o
o
g
CO
CM
CM
g
O
o
o
o
£
CM
JS
IS
-
^^
fs^
g
o
0
0
ft
CO
CO
CM
g
5
O
s?
CO
CO
m
CO
s?
^
in
in
CO
**
CO
8
CM
CO
CO
j£
CO
s
o
CM
*s
jo
s
CM
CD
g
CO
CO
£
CM
co
CO
g
O
0
o
ft
*
co
co
CM
ft
o
o
o
o
&
1
2
CO
CD
1
>
8?
Q
2
CO
CD
I1
CD
&
#
1
2
CO
CD
O)
2
CD
K
o
^
1
2
CO
CD
1"
CD
^
<
co
CO
^
g
00
^
g
CO
CO
^
o
K
T-
o
in
CO
o
#
CO
in
(O
00
O
m
T—
g
o>
K
g
CO
CM
co
o
#
8
%
CO
C\l
S?
8
%
CM
CO
CM
#
CM
!o
T—
CO
CM
*s
5
CO
s
s?
8
CO
CO
*
'
CO
o>
g
*
CO
CO
CM
&
CM
O)
CO
#
^
in
CM
g
o
00
O)
o>
1
#
1
2
CO
CD
1"
1
i
s?
s
2
CO
CD
CD
&
#
I
2
CO
CD
CD
i
#
I
2
CO
CD
I1
CD
>
<
£
CM
CO
OO
*
-
S
£
CO
CO
$
CM
S
#
£
U)
in
#
co
®
o
CO
CO
&
•»
s
g
CO
CM
CO
o
CM
Except where noted, the absolute
humidity was held at 15,000 ppm H
Xylene Model Parameters
Gas Constant K
Resistance (a) in Clean Air Rb
Sensitivity
Xylene Readings (ppm)
o o o o 52
*- CO O O O
T- CO O
o.
Q.
1
.i
8
CD
C
CD
>s
X
11
Q. V
2 [? 1
3 tO ffl
Iff
LU
o o o o o
•«- co o o o
T- CO O
I
1
CD
_CD
Cross Sensitivity
Xylene Reading (ppm) when:
5000 ppm Methane is delivered
5000 ppm Butane is delivered
21
-------
o
(0
c
o
CO
x
«
>s
^—'
!5
I
C
CD
CO
CO
CM
03
CO
O
CD
CO
CD
C
C CD
o> O >»
2X ^^ ^^
vl 1^5 ^^
§ g E
or "CO °
O ^
°* ** o
V» ^1 ^>
*~ CM *~
^ ^ fftl
^
g § |
2 2 £
§•£ E
« ~m &
pr CO O.
*M O O
IE®
T-
S) "2
Is
*!
«\l O
£ =>
CD
C
CO o
DC O o
O> ^ o
O> CM o
T- "*"
>
&
55
1
2
CO
CD
CO
CD
>
<
I
55
1
2
CO
CD
O)
2
CD
i
55
1
2
CO
CD
O)
2
CD
1
^$
0
>
O
2
co
CD
CD
#
§
Q
CM
CD
O
J?S
*
Tt
UJ
CD
in*
in
o
-»-
LU
*-
CM
CM
o"
CO
in
CD
!
LU
co
in
o
LU
CO
CM
CM
o
o
0
LU
CO
CO
LU
CM
CM
£j
CD
in
o
55
c|
LU
CO
CO
LU
>
tS
55
i
2
CO
CD
CO
CD
>
<
% Dev.
1
2
CO
CD
O)
2
CD
I
55
1
2
CD
i
CD
1
^8
>
O
2
CO
CD
O)
2
CD
#
Q
"^*
CO
•
Q
^
I
2
CO
CD
(0
u-
^
<
I
i
2
CO
CD
CD
1
55
1
2
o>
CD
1
>5
0
>
o
I
CD
CD
ss
Q
^
o
o
85
CO
***
-
CM
#
CO
^*
"
O)
*5
CO
CM
CM
*5
CO
s
CO
^*»
CO
CO
o
o
1
CM
CM
g
CO
in
co
1
m
o>
CO
g
£
CO
CM
"
O
o
1
-
T~
56
3
CO
in
1
O)
CO
CM
55
3
CM
55
^~
o
0
g
T—
*-
T~
55
K
•t
CD
1
CM
CM
CO
55
CO
CO
00
>
Q
S5
1
5
CO
CD
O)
to
CD
^
<
1
55
1
2
CO
CD
CD
1
55
1
2
CO
CD
CD
I
>5
0
>
o
2
CO
CD
I1
CD
^
JO
^*
o
CM
CM
55
in
8
^
o
"*
^
^
00
i
CM
S
CO
LO
1
m
*~
g
T"
5
00
CO
g
O)
CO
CM
g
^M.
CO
r^»
CO
O)
IN.
O
CM
|JE
.2 Q.
O O_ (ft
to 0 2
-So ®
ni = *-
«s i
1s i
|1 1
•*= ^>
£ to -o
CD (0 O
•c * 5
5 >, 2
*- .t= 0)
8-2 c
|E £
LU J 5*
^
Q
CD >- W
Q. < c?
— c "6
CO ro (0
> CD ®
| 0 ^
CD ^ CD
§ a ^
°- 8 £
s ^
-n 1
CD C
DC 9
(/)
O O O O O
i- CO O O O
r- CO O
ex
a.
£
CD
CD
^^
X
* I
.E Q.
Q. V
.-s* ^ 1
2 o> E
2 I ^
3 (0 jo
I r? *
"S CD 58
c 5
!t~
UJ
o o o o o
t- CO O O O
t- CO O
Q.
a.
1
&
CD
CD
^H.
X
0
f
|
o>
c
>» ^
±£ CO
£ CD
w o
(/) >«
^^
9
O
£ £
CD CD
« S
.S2 .J2
CD CD
C C
(0 CO
£ "3
CD CD
E 1
&tt
0
^5 J5?
s
28
-------
CD
CO
X
C£
CO
OB
o
JD
cd
CO
o
0)
c
0)
CO
CM
CVJ
00
CO
O
CD
CD
C
^ o
O> CM 0
<§>
1
35
£
o
B
CO
CD
£
CD
>
<
35
1
B
CO
CD
»—
>
<
1
35
O
B
CO
CD
*JT
5
<
i
85
O
B
CD
?
CD
<
85
8
CM
O
O
O
*
*~
in
UJ
CO
in
o
LLJ
^.
m
35
CO
a
^
_^
>9
CO
1
UJ
CM
0
III
00
CO
CM
O
£
f^
*
CM
m
UJ
•^
CD
in
o
LU
m
£
CO
o
s
o
35
m
CM
in
9
UJ
CD
m
o
UJ
in
8
85
£
B
CO
1
CD
>
<
>'
35
O
B
CO
CD
k_
CD
<
1
35
I
B
CO
CD
9
5
<
i
35
1
§
CD
<
t
^
in
1
00
o
g
0
o
o
^
s
o
CO
CM
g
o
o
o
o
>?
m
o>
CO
>?
X
-
^-
vo
*
O
o
o
1—
^
CO
CM
S
o
CO
CM
>°
a
CM
CO
O)
I
(O
o
«
CO
00
in
CM
85
o
(O
o>
in
85
o>
in
CM
m
CO
T"~
35
cti
CO
s
LO
g
o>
in
T—
35
CM
CM
0
*
g
o
o
o
1
3
O
CO
CM
g
O
o
o
o
>
85
1
B
CO
CD
O)
ro
CD
>
<
>
&
35
1
B
CO
CD
9
il
CD
^
1
35
O
B
CO
CD
i_
CD
<
1
35
1
B
CO
CD
CD
<
CO
'*~
CM
<-
o
*~
^*"
^
35
§
o
CM
in
Si
5
35
§
r^
00
CO
„
CM
CM
CM
;s°
S
in
in
sP
CM
0
s
cli
in
0
CM
3>
3
CO
CM
-
*-
o
CO
CO
£
r*.
CO
35
CO
CM
85
CM
CM
36
m
CO
CM
*-
85
m
m
in
o>
CO
CO
1
S
s
*
CO
3
CM
S
i
35
1
B
(O
CD
4_
CD
<
85
O
B
CO
CD
2
CD
•^
1
35
i
ig
CO
CD
O)
&
CD
<
1
86
i
S
CD
I1
CD
<
35
In
2
CO
CO
85
CM
00
00
£
00
CO
sp
^
CO
CO
0
8
O
CD
g
O)
CM
CM
I
S
in
CM
85
to
9
I
i
s
00
o
CM
.5 Q.
9 Q- M
Except where noted, the abs
humidity was held at 15,000
Xylene Model Parameter
„ 0 f
•2:
Power Law Slope
Resistance (ft) in Clean Air
Sensitivity
Xylene Readings
O O O O O
^- CO O O O
i- CO O
T~
Xylene Delivered (ppm)
k_ — >
S E
Effect of Humidity
Xylene Readings (ppm) in
(Absolute Humidity < 50 pp
o o o o o
f- CO O O O
T- CO O
Xylene Delivered (ppm)
C
CD
$
Cross Sensitivity
Xylene Reading (ppm)
T3 TJ
£ 2
.i >
.J2 .«2
(0 (0
£ 3
CD m
| E
o °
o m
29
-------
c
(D
CO
X
CO,
CO
to
Q
!5
co
o
(0
0)
CO
CM
00
CO
O
CD
CO
_0)
*
CD
c I
q> O >»
w '£2 *>
C tS* r-
5 CO E
X f>
* i= o"
rv* (0 S
*M O ^
1 ZL |
* ^ ffti
w
« 0 «
& .2 c
lit
co ±= E
OJ ^^5 ^^
fu JR Q_
^^J ^5
o> • °
T-
$ is
o to
O. ft.
W ft
o ±:
CM O
O> c
CD
C
o> .2
C .Q *
O 7n t
g- JD a
fl> "co 8
OC O o
ir" ST *°
O) CM 0
@
35
>
a
S
CO
CD
CO
2
CD
>
<
i
85
>
O
2
CO
CD
CD
<
>
85
a5
o
2
CO
CD
<
1
3?
Q
2
CO
CD
CD
<
35
CO
co
d
o
o
1
IO
UJ
CO
co
o
UJ
o
"*
85
10
rr
Q
O)
o
1
10
UJ
CO
CM
m
UJ
CO
CO
5)
-»
o
en
d
35
1
UJ
CO
in
UJ
CD
CM
3)
^
O)
O
35
S
IO
5
m
CO
in
UJ
CD
CM
I
85
>
Q
2
CO
CD
CO
2
CD
>
<
>'
85
>
o
2
CO
CD
CD
<
>
35
£
O
2
CO
CD
I1
CD
<
>
35
O
2
CO
CD
CD
<
OJ
CO
^
2
j£
r*.
o
g
o
o
o
CO
•*
-
in
o
CO
g
o
o
o
o
g
^
CM
g
(O
r-
co
g
o
o
o
T~
£
CO
CM
h-.
CO
CO
g
s
5
CM
1
CO
o
*
IO
CM
CO
CM
CM
CM
CO
85
*r
CM
CO
T—
CO
I
O)
•t
CM
h*
O
1
**
CO
35
IO
CO
5!
g
O
o
o
35
CO
o
UO
CO
g
o
o
o
o
1
35
>
Q
2
CO
CD
9
CD
^
<
i
85
s
o
2
CO
CD
c?
il
CD
<
1
85
55
O
2
CO
CD
CD
<
1
35
Q
2
CO
CD
CD
<
35
T-
co
1
CO
0)
35
CM
CM
CD
CM
35
CO
CD
CM
O
35
ft
T—
CM
5
85
-
CM
CO
CM
N.
35
CM
IO
a
85
CO
1^
8
1~
CO
CO
T—
CD
"*
85
-
CM
35
CM
CO
35
IO
co
o
CM
35
CM
CO
ft
O
o
g
cn
CO
1
'*"
CM
CO
CO
cn
1
IO
CO
CM
1
8
r-.
CM
35
-
o
35
>
Q
2
CO
CD
CD
>
<
I
35
1
Q
2
CO
CD
e?
il
CD
<
1
35
65
o
2
CO
CD
CO
2
CD
<
1
85
Q
2
CO
CD
9
CD
<
*
•^
*~
O
CO
^
Q
^
O
g
rr
h-
CM
g
£
00
T~
35
O
CM
£
CO
10
2
I
IO
o
CO
1
s
o
CM
O
CM
1 i.
00- (0
3 g §
« 8 |
r- in c
±3 T- C
- .^ CO
"° W CO
•C 0)
(D CO ft
-I 5 §
£* «
S--1 g
8 E .£
UJ .C x
015 I
~ »_ w
CD ._ rj)
0- < c
•2 c =6
CO tO CO
1 5 *
•d .s 8
CD ^^ ^
2 g" £X
C ^
'to 2
CD C
c/>
o o o o o
i- CO O 0 O
i- CO 0
ne Delivered (ppm)
5
X
|i
c a
-gg
oL v
^^ £-
.« co 1
V co E
E i ^
3 2 £
1 C 1
0 g |
UJ
O O O O O
*- CO O O O
i- CO O
ie Delivered (ppm)
*
x
C
CD
"1
O.
CO
« ^
CO
i
O
^n ^n
1 1
^O ^O
.CO .CO
CD Q
c c
CO OJ
£ 3
CD CQ
1 I
QL o
8
30
-------
o
to
c
o
CO
X
CO
o
Q
CO
k_
o
O)
CO
CO
GO
CO
o
CD
_
JD
CO
CO
4) O "^
W TJ3 ^N
§ CO E
g-.g &
«^ CD ^^
So 2
la: 8
w
c
8) £3 ®
Cm "^
0 2 X
(0 «^ L.
it* CO Q.
£ ° 8
la! ®
T-
g ^>
0 OB
Q. i-
(0 ja
0) »
CM O
^^
CD
C
(I) CD
(0 c >>
c o x
§-1 1
(0 ~
0> co 8
OC 0 o
O> ^- o
O) C\J o
S
1
38
1
2
CO
CD
O)
2
®
<
1
38
3
Q
2
CO
CD
O)
2
>
38
1
•
2
CD
1
<
&
38
cs
o
2
to
CD
2
CD
38
d
38
IO
III
00
CM
m
o
LU
f"
JS
s
0
IN.
O
g
IO
UJ
^
o
LU
CM
*
3
o
0
38
CO
1
LU
O
's*
CO
00
s
K
h.
CO
ur>
1
S
0>
CO
1
O)
IO
CM
1
IO
CM
CO
in
3j
1
^~
CM
CO
CM
t
CO
CM
CM
38
o
CM
co
o
CM
g
O
0
o
in
^
5
0
CO
CO
g
o
o
o
o
in
1
38
Q
8
CD
O)
2
CD
<
I
38
cB
Q
2
to
CD
CD
1
38
1
W
CD
9
CD
<
I
38
1
2
to
CD
I1
1
CO
00
OJ
1
S
•<*•
CO
38
co
P
CM
38
CM
3
m
CO
#
fc
IO
E
1
IO
O)
1
00
o
CD
38
S
ft
O
CO
CM
g
CO
CM
00
00
1
s
CD
3
38
CM
oo
fN,
O)
38
S
IN,
CM
38
00
oS
s
I
CM
CO
5>
CO
|
CM
V"
0)
g
CO
r>.
38
0
*
g
$
o
o
CM
38
IO
CM
g
&
CD
fO
IN.
m
CD
CO
i
38
i
2
CO
CD
?
CD
<
I
38
i
o
2
CD
I
38
i
2
to
CD
CD
•^
I
38
1
8
CD
I1
1
T—
S
fN,
#
8
o>
s
38
^
O
00
IN.
a
ft
CM
in
CD
00
1
g
CM
1
38
CO
81
%
#
19
"
CO
CM
£
S
co
o
CM
S E
20. e
OQ- m
o O **
!§ 1
i « |
is *
§i «
£ <« o
li 1
*- .-ti C.
Q. TD (0
81 £
•2 J |
„ 0 |
o.
Power Law Slope
Resistance (ft) in Clean Air
Sensitivity
Methane Readings
O O O O O
10 - S ? |
Methane Delivered (ppm)
|l
r-\ P
Effect of Humidity
Methane Readings (ppm) in I
(Absolute Humidity < 50 ppi
o o o o o
in in o Q o
i- in in o
Methane Delivered (ppm)
CD
*
Cross Sensitivity
Methane Reading (ppm)
1 1
i .1
« 3
a> .50
0 0
« CD
3 >.
m X
1 1
8 8
S 5
31
-------
SECTION 7
CONCLUSIONS
This series of laboratory investigations has ful-
filled its primary objective which was to allow for
direct comparison of the properties of distinct
vapor sensor technologies. Also, this program has
demonstrated that simulating the UST environment
in a controlled laboratory setting can provide cost
effective and meaningful data about vapor sensors
used at UST sites. This information can be useful
in making appropriate decisions in sensor selection
and use before conducting costly field tests.
A set of performance specifications has been
measured and presented for the DET and GAT
(Standard and MI modes) catalytic bead sensors.
The results show that the DET and GAT-Standard
mode sensors had comparable responses. The sen-
sors responded to methane, butane, and xylene, but
were most sensitive to butane. Variations in sen-
sor-to-sensor response to butane at concentration
levels below 1500 ppm were significant, as were
the effects of humidity. At butane levels above
1500 ppm these effects are minimized.
Throughout the testing, these sensors were shown
to respond at normal UST vapor concentrations.
The butane sensitivity and speed of response
of the GAT sensors were shown to be comparable
whether operated in the MI or Standard mode.
However, in the MI mode the GAT sensors were
shown to be methane insensitive with an increased
response to xylene. Also, they were relatively
insensitive to changes in humidty, except for read-
ings in clean air.
The goal of this series of laboratory investiga-
tions was to evaluate the DET and GAT catalytic
bead sensors. This has not been a comprehensive
study of all aspects affecting the DET and GAT
catalytic bead sensors. Therefore, the results
should not be misinterpreted as being directly
transferable to field conditions.
The stability data collected for the Metal
Oxide Semiconductor sensors and Polymer
Adsorption sensors indicate that they did drift. The
Polymer Adsorption sensors were more repro-
ducible and a one-point calibration seemed suffi-
cient to re-characterize the sensors for UST leak
detection. The variations in drift observed in the
Metal Oxide Semiconductor sensors suggest that at
least a two-point calibration would be required to
re-characterize these sensors.
The extent of drift and the time between cali-
bration required for these sensors at an actual UST
site cannot be concluded from the data of this
study. This is because the sensors were maintained
in clean humidified air for this test program. At a
UST site, the sensors would be continually
exposed to hydrocarbons and other combustible
gases such as hydrogen sulfide. The long-term
effects of this environment on these sensors is
unknown.
32
-------
SECTION 8
RECOMMENDATIONS
This program was a demonstration pro-
ject. Due to the limited scope, tests measuring
the range of environmental effects or vapors one
would likely encounter at a UST site were not
performed. Also, the method of data presentation
can be improved so that a wider audience can
take advantage of the data collected.
It is recommended to:
1. Continue sensor technology studies by test-
ing additional vapor sensor technologies, such as
those that are fiber optic based.
2. Expand the program to examine reliable
methods to field calibrate vapor sensors used with
UST leak detection systems.
3. Design the presentation of the data so that it
can reach a wider audience.
4. Expand the program to examine the sensors
under exposed hydrocarbon with other combustible
gases such as hydrogen sulfide under field condi-
tions.
5. Expand the program to include long-term
effects on the sensors. The program should include
statistics on precision, accuracy, detection limits
and dynamic range as a function of time.
33
-------
REFERENCES
1) Portnoff, M. A., Grace, R., Guzman, A. M.,
Hibner, J., Measurement and Analysis of
Adsistor and Figaro Gas Sensors Used for
Underground Storage Tank Leak Detection,
Final Project Report, U. S. Environmental
Protection Agency, Environmental
Monitoring Systems Laboratory, Office of
Underground Storage Tanks, Las Vegas,
Nevada, PO #OV-1255-NAEX, August 26,
1991. [EPA/600/R-92/219, December, 1992].
2) Schlez, C, "Background Hydrocarbon Vapor
Concentration Study for Underground Fuel
Storage Tanks," Draft Final Report for U.S.
EPA, Contract No. 68-03-3409, February 29,
1988.
3) Personal communication with Philip B.
Durgin, Ph.D., U. S. Environmental
Protection Agency, Environmental
Monitoring Systems Laboratory, Las Vegas,
Nevada, November 1990.
4) Figaro Taguchi sensors are a product of
Figaro Engineering of Japan represented by
Figaro USA, Inc., P. O. Box 357, Wilmette,
IL 60091.
5) Adsistor Vapor Sensors are products of
Adsistor Technology, P.O. Box 51160,
Seattle, Washington 98115.
6) Detronics sensors are a product of Detector
Electronics Corporation, 6901 West 110th
Street, Minneapolis, Minnesota 55438.
7) GasTech sensors are a product of Gas Tech
Inc., 8445 Central Avenue, Newark, CA
94560.
8) Grace, R., Guzman, A., M., PortnofT, M. A.,
Runco, P., Yannopoulos, L. "Computational
Enhancement of MOS Gas Sensor
Selectivity," P-33, Proceedings of the Third
International Meeting on Chemical Sensors,
Cleveland, OH, September, 1990.
9) Dolan, J., Jordan, W., "Detection Device",
U.S. Patent* 3,045,198, July 17, 1962.
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development
(ORD), funded the research described here. Mention of trade names or commercial products does not con-
stitute endorsement or recommendation for use.
34
-------
APPENDIX A
CATALYTIC BEAD SENSORS
THE FOLLOWING PRODUCT INFOR-
MATION WAS ABBREVIATED FOR
INCLUSION IN THIS REPORT.
Reprinted from:
Detector Electronics Specification Data Sheets
90-1041-01 (3/89) and 90-1043-01 (1189).
DESCRIPTION
The sensor is a very critical component in a
combustible gas detection system. The reliability
of the system requires consistent and dependable
performance of the sensor under extremes of oper-
ating and environmental conditions.
The sensor is essentially two matched resis-
tors, a catalytic element to sense combustible gases
and a reference element to compensate for envi-
ronmental conditions. These elements are elec-
tronic components in the same sense as the other
electronic parts in the detection systems circuitry.
However, unlike the other components, the element
pair must operate at a much higher temperature
while exposed to the ambient atmosphere. For this
reason, Det-Tronics (previously referred to as
Detronics) has developed METROSEAL™ a spe-
cial high density alumina sheath to protect the
resistance wire of each element. This results in
greater mechanical strength, less drift, greater sig-
nal stability, more uniform characteristics, and
increased sensor life. The unique tubular shape of
the Det-Tronics sensor element provides other sig-
nificant benefits. Because its mass is reduced, the
Det-Tronics sensor is less susceptible to mechani-
cal shock and long term vibration effects. In addi-
tion, the cylindrical shape provides more uniform
heating characteristics and fast still-air response
time.
Electronic parts produced by component man-
ufacturers have a known reliability as a result of
their extensive testing and characterization of the
devices. At Det-Tronics, sensor elements are man-
ufactured with the same built-in reliability as the
other more common electronic components of the
system. Det-Tronics sensor elements are burned in
and extensively tested at a number of voltages and
under a variety of test gas conditions. The data of
these tests are stored in a computer. Devices fail-
ing to meet Det-Tronics stringent requirements are
rejected. The acceptable active and reference ele-
ments are then computer matched as closely as
possible in electrical and thermal characteristics.
The purpose of this test is to match the active and
reference elements to criteria that minimize drift
and provide better long-term stability over the wide
range of temperatures encountered by operational
sensors.
Det-Tronics sensors are universal in operation
and configuration with gas sensor technology.
They detect all combustible gases and are supplied
in a single configuration that is used with all Det-
Tronics gas detection systems. This considerably
simplifies installation and spares provisioning
since only one type of sensor needs to be stocked.
FEATURES
• Computer matched active and reference
elements minimize zero and span drift.
• Exclusive Metroseal ceramic coating and
tubular shape provide lower vibration sen-
sitivity, more uniform heating, greater sta-
bilization, and fast still-air response time.
• Thermal isolation within the sensor hous-
ing reduces thermal interaction and
assures equal exposure to ambient condi-
tions for both sensor elements.
Gold plated corrosion inhibiting contacts
and keyed connectors simplify mainte-
nance and installation.
Optional clip-on dust covers are easily
installed for protection in dusty or windy
environments.
Full two year warranty on sensors and one
year electronics in all Det-Tronics gas
detection systems by the Det-Tronics
company.
35
-------
• Single sensor for all combustible gas sim-
plifying installation and replacement.
• Sintered stainless steel flame arrester.
SPECIFICATIONS
Storage Temperature -67° to +257°F
(-55°C to +125°Q
Optional High temperature model -67°F to +400°F
(-55°C to +205°C)
Ambient Operating Temperature -40° to +185°F
(-40°C to +85°C)
Optional high temperature model -40°F to +400°F
(-40°Cto+205°C)
Response Time* Less than 10 seconds to reach
50% of full scale reading and less than 30 seconds
to reach 90% of full scale reading with methane (in
still air)
Recovery Time* Less than 30 seconds after expo-
sure to pure methane
Sensor Life 3 to 5 years expected
Calibration Cycle typical 90 days
Linearity* ±5% linearity for all combustible gases
(± 3% Full scale up to 50% LFL, ± 5% Full scale
50% - 100% LFL)
*For additional information contact the factory for
technical notes and tests on specific gases.
36
-------
THESE INSTRUCTIONS WERE ABBREVIATED FOR INCLUSION IN THIS REPORT.
DET TRONICS
Communication with Ken Freiteg
Please use the circuit below for evaluating the Det Tronics Catalytic Bead Sensors.
"I Black
Red
=_ V1
/"\
White
^ To
^ D.A.S.
Sensor
A = Active Bead
R = Reference Bead
Rl = R2 = 1KQ, 1%, 50 ppm metal film resistors
VI = 3.30 VDC, 0.5 amp or better
Note: In this configuration, the Red lead will swing positive with applied gas.
37
-------
THESE INSTRUCTIONS WERE ABBRE-
VIATED FOR INCLUSION IN THIS
REPORT.
Reprinted from:
A report by Marc Portnofffrom Carnegie Mellon
Research Institute based on communication with
Gas Tech Engineering with
permission of Gas Tech.
Gas Tech, Inc.
10/4/91
The sketch below shows how to electrically
connect the GasTech Catalytic Bead Sensor into a
wheatstone bridge circuit for measuring purposes.
Note that the bridge resistors don't really have to
be exactly 442 £1, but they do need to be carefully
selected to be accurately identical.
For a perfectly balanced sensor and bridge
resistor set, there will be exactly 1/2 of the VDC
measured at the common points of sensor and
resistors, and therefore no current flow in an
ammeter connected between them. (This is rare;
there will always be a little off balance).
A gas response creates burning gas at the
active element, which increases its temperature and
therefore its resistance. This change in resistance
creates a change in balance of the bridge, which
creates a current flow change in the ammeter. For
PPM measurements, the change is very slight
(about 1 mV offset change or less for 100 ppm
hexane).
Note: many of the instruments for UST work
have the "No Methane Response" setup, or a
switch enabling user to switch between 3.8 VDC
and 6.0 VDC on the sensor, for use in an area
known to have natural methane in the ground. The
"No Methane Response" mode eliminates = 99.5%
of the methane response, but also reduces or elim-
inates some other gases.
Gas Tech Catalytic Bead Sensor Sketch
RED = Active (A)
WHITE = Common (C)
GREEN = Reference (R)
FDC for full response = 6.0 VDC
FDC for "No Methane Response" = 3.8 FDC
38
-------
APPENDIX B. CATALYTIC BEAD SENSOR TEST DATA
Table B1. DET Catalytic Bead Sensor Measurements
(#) Time period for changes in concentrations is 30 minutes.
*
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1/
18
19
20
21
22
23
24
25
26
21
28
29
30
31
32
33
34
35
36
3/
38
39
40
41
42
43
44
45
H20
ppm
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Operating Power
Gain (ppm butane/mv)
CH4
ppm
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Zero (V)
C4H10
ppm
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
C8H10
ppm
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
30
100
300
1000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
DET #4
3.3V
333
0.027
333
333
333
667
1000
2999
333
333
333
0
333
333
667
1666
4999
333
333
0
0
333
0
0
667
1666
0
-333
-333
-333
0
667
2666
-333
-333
-333
-333
-333
-333
333
1333
4999
-333
-333
-333
-333
-333
DET #5
3.3V
312
0.012
0
0
312
312
937
2812
0
0
0
0
0
312
312
1562
4999
312
312
0
0
0
0
312
625
1875
0
-625
-625
-625
-312
312
2500
-625
-625
-625
-625
-625
-312
0
1250
4687
-625
-625
-625
-625
-625
DET #6
3.3V
333
0.021
0
0
0
333
667
2666
0
0
0
0
0
0
667
1666
4999
333
0
0
0
0
0
0
667
1666
0
-333
-667
-333
0
667
2666
-333
-333
-333
-333
-333
-333
0
1333
4666
-333
-333
-667
-667
-333
Average
326
0.020
111
111
215
437
868
2826
111
111
111
0
111
215
549
1632
4999
326
215
0
0
111
0
104
653
1736
0
-430
-542
-430
-104
549
2611
-430
-430
-430
-430
-430
-326
111
1305
4784
-430
-430
-542
-542
-430
DET Sensor
Std. Dev.
12
0.008
192
192
187
199
177
167
192
192
192
0
192
187
204
60
0
12
187
0
0
192
0
180
24
120
0
168
182
168
180
204
96
168
168
168
168
168
12
192
48
187
168
168
182
182
168
3
% Dev.
4%
38%
173%
173%
87%
45%
20%
6%
173%
173%
173%
-
173%
87%
37%
4%
0%
4%
87%
~
-
173%
-
173%
4%
7%
-
-39%
-34%
-39%
-173%
37%
4%
-39%
-39%
-39%
-39%
-39%
-4%
173%
4%
4%
-39%
-39%
-34%
-34%
-39%
39
-------
Table B1. DET Catalytic Bead Sensor Measurements (cont.)
(#) Time period for changes in concentrations is 30 minutes.
*
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
H20
ppm
0
0
0
0
0
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
4999
1667
0
15002
15002
15002
15002
15002
15002
15002
15002
4999
1667
0
15002
Operating Power
Gain (ppm butane/mv)
CH4
ppm
0
0
0
0
0
0
500
0
0
0
0
0
500
500
500
500
4999
500
500
500
500
500
500
0
500
500
500
500
500
0
0
0
0
0
0
4999
4999
4999
4999
4999
Zero (V)
C4H10
ppm
0
0
0
0
0
0
0
0
500
0
0
0
500
500
500
500
500
500
4999
500
500
500
500
0
500
500
500
500
500
0
4999
500
4999
0
0
4999
4999
4999
4999
4999
C8H10
ppm
30
100
300
1000
0
0
0
0
0
0
100
0
100
100
100
100
100
100
100
100
1000
100
100
0
100
100
100
100
100
0
0
0
0
0
0
1000
1000
1000
1000
1000
DET #4
3.3V
333
0.027
-333
-333
333
1666
-333
333
667
333
667
333
667
333
1333
1333
1333
1333
3666
1333
5666
1333
2666
1333
1333
333
1333
1000
1000
667
1333
667
5666
1333
5666
667
333
9665
9331
9331
9665
9665
DET #5
3.3V
312
0.012
-625
-312
0
1562
-625
0
312
0
312
0
312
0
937
937
937
937
3437
1250
5311
1250
2812
1250
1250
0
1250
937
625
312
1250
312
5311
625
5311
312
0
4374
4062
9373
9373
9686
DET #6
3.3V
333
0.021
-333
-333
0
1666
-333
0
333
0
667
0
333
0
1000
1000
1000
1000
3333
1000
5332
1000
2666
1000
1000
L 0
1000
667
667
667
1000
667
5332
1000
5332
667
0
8998
8998
9331
9331
9331
Average
326
0.020
-430
-326
111
1632
-430
111
437
111
549
111
437
111
1090
1090
1090
1090
3478
1194
5436
1194
2715
1194
1194
111
1194
868
764
549
1194
549
5436
986
5436
549
111
7679
7464
9345
9456
9561
DET Sensor;
Std. Dev.
12
0.008
168
12
192
60
168
192
199
192
204
192
199
192
213
213
213
213
170
173
199
173
84
173
173
192
173
177
205
204
173
204
199
354
199
204
192
2881
2951
24
182
199
I
% Dev.
4%
38%
-39%
-4%
173%
4%
-39%
173%
45%
173%
37%
173%
45%
173%
20%
20%
20%
20%
5%
15%
4%
15%
3%
15%
15%
173%
15%
20%
27%
37%
15%
37%
4%
36%
4%
37%
173%
38%
40%
0%
2%
2%
40
-------
Table B2. GAT Catalytic Bead Sensor-Standard Mode Measurements
(#) Time period for changes in concentrations is 30 minutes.
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
H20
ppm
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Operating Power
Gain (ppm butane/mv)
CH4
ppm
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Zero (V)
C4H10
ppm
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
C8H10
ppm
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
30
100
300
1000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
GAT#1
6.0V
37
0.035
0
0
185
185
741
2777
0
0
0
0
185
185
555
1666
4999
0
0
0
0
0
185
185
370
1111
0
-370
-370
-185
-185
555
2407
-370
-370
-370
-370
-370
-185
185
1296
4999
-370
-370
-370
-370
-370
GAT #2
6.0V
33
-0.013
0
0
0
231
808
2854
0
0
0
0
0
115
462
1617
4999
0
0
0
0
0
0
115
346
1039
0
-561
-561
-445
-214
346
2359
-561
-561
-561
-561
-561
-445
0
1155
4669
-676
-676
-676
-676
-676
GAT #3
6.0V
34
0.035
0
172
172
172
862
2586
172
172
172
0
172
172
517
1724
4999
172
172
172
172
172
172
172
345
1034
172
-172
0
0
172
690
2586
-172
-172
-172
-172
0
172
345
1551
4827
-172
-172
-172
-172
-172
c
Average
35
0.019
0
57
119
196
804
2739
57
57
57
0
119
158
512
1669
4999
57
57
57
57
57
119
158
354
1062
57
-368
-310
-210
-76
530
2451
-368
-368
-368
-368
-310
-153
177
1334
4832
-406
-406
-406
-406
-406
GAT Senson
(tandard Moc
Std. Dev.
2
0.028
0
100
103
31
61
138
100
100
100
0
103
37
47
54
0
100
100
100
100
100
103
37
14
43
100
194
285
224
215
173
119
194
194
194
194
285
310
173
201
165
254
254
254
254
254
)
ie
% Dev.
6%
148%
~
173%
87%
16%
8%
5%
173%
173%
173%
-
87%
24%
9%
3%
0%
173%
173%
173%
173%
173%
87%
24%
4%
4%
173%
-53%
-92%
-106%
-284%
33%
5%
-53%
-53%
-53%
-53%
-92%
-203%
98%
15%
3%
-62%
-62%
-62%
-62%
-62%
41
-------
Table B2. GAT Catalytic Bead Sensor-Standard Mode Measurements (cont.)
(#) Time period for changes in concentrations is 30 minutes.
#
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
H20
ppm
0
0
0
0
0
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
4999
1667
0
15002
15002
15002
15002
15002
15002
15002
15002
4999
1667
0
15002
Operating Power
Gain (ppm butane/mv)
CH4
ppm
0
0
0
0
0
0
500
0
0
0
0
0
500
500
500
500
4999
500
500
500
500
500
500
0
500
500
500
500
500
0
0
0
0
0
0
4999
4999
4999
4999
4999
Zero (V)
C4H10
ppm
0
0
0
0
0
0
0
0
500
0
0
0
500
500
500
500
500
500
4999
500
500
500
500
0
500
500
500
500
500
0
4999
500
4999
0
0
4999
4999
4999
4999
4999
C8H10
ppm
30
100
300
1000
0
0
0
0
0
0
100
0
100
100
100
100
100
100
100
100
1000
100
100
0
100
100
100
100
100
0
0
0
0
0
0
1000
1000
1000
1000
1000
GAT#1
6.0V
37
0.035
-370
-185
0
741
-370
0
185
0
555
0
185
0
926
926
926
926
3333
926
5554
926
1851
926
926
185
926
741
741
741
926
185
4999
555
4999
185
185
8887
8887
8887
8887
8887
GAT #2
6.0V
33
-0.013
-561
-561
-214
577
-676
-115
115
-115
346
-115
0
-115
693
693
808
808
3184
808
5164
808
1732
808
808
-115
808
577
462
346
808
-115
4834
346
4834
-115
-115
8629
8629
8464
8464
8629
GAT #3
6.0V
34
0.035
0
0
172
862
-172
172
345
172
690
172
172
172
1034
1034
1034
1034
3275
1034
5344
1034
1724
1034
1034
172
1034
862
862
690
1034
172
4999
690
4999
172
172
8447
8447
8447
8447
8447
s
Average
35
0.019
-310
-249
-14
727
-406
19
215
19
530
19
119
19
884
884
923
923
3264
923
5354
923
1769
923
923
81
923
727
688
592
923
81
4944
530
4944
81
81
8654
8654
8599
8599
8654
GAT Sensors
tandard Moc
Std. Dev.
2
0.028
285
286
194
143
254
145
118
145
173
145
103
145
174
174
113
113
75
113
195
113
71
113
113
170
113
143
205
214
113
170
95
173
95
170
170
221
221
250
250
221
)
te
% Dev.
6%
148%
-92%
-115%
-1381%
20%
-62%
764%
55%
764%
33%
764%
87%
764%
20%
20%
12%
12%
2%
12%
4%
12%
4%
12%
12%
211%
12%
20%
30%
36%
12%
211%
2%
33%
2%
211%
211%
3%
3%
3%
3%
3%
42
-------
Table B3. GAT Catalytic Bead Sensor-Ml Mode Measurements
(#) Time period for changes in concentrations is 30 minutes.
#
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
36
36
37
38
39
40
41
42
43
44
4b
H20
ppm
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Operating Power
Gain (ppm butane/mv)
CH4
ppm
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Zero (V)
C4H10
ppm
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
50
150
500
1500
4999
0
0
0
0
0
C8H10
ppm
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
30
100
300
1000
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
10
GAT #4
3.8V
56
0.018
0
0
0
0
0
0
0
0
0
0
0
0
0
833
4999
0
0
0
0
0
0
0
555
1944
0
0
0
0
0
0
0
0
0
0
0
0
0
0
833
4999
0
0
0
0
0
GAT #5
3.8V
45
-0.007
0
0
0
0
0
0
0
0
0
0
0
0
227
1136
4999
0
0
0
0
0
0
227
454
1591
0
-227
-227
-227
-227
-227
0
-227
-227
-227
-227
-227
0
227
909
4999
-227
-227
-227
-227
-227
GAT #6
3.8V
50
0.013
0
0
0
0
0
0
0
0
0
0
0
250
500
1250
4999
0
0
0
0
0
250
250
750
2000
0
0
0
0
0
0
0
0
0
0
0
0
250
500
1250
4999
0
0
0
0
0
Metha
Average
50
0.008
0
0
0
0
0
0
0
0
0
0
0
83
242
1073
4999
0
0
0
0
0
83
159
587
1845
0
-76
-76
-76
-76
-76
0
-76
-76
-76
-76
-76
83
242
997
4999
-76
-76
-76
-76
-76
GAT Sensor.
ne Insensitiv
Std. Dev.
5
0.013
0
0
0
0
0
0
0
0
0
0
0
144
250
215
0
0
0
0
0
0
144
138
150
222
0
131
131
131
131
131
0
131
131
131
131
131
144
250
222
0
131
131
131
131
131
3
eMode
% Dev.
10%
165%
-
-
--
--
--
-
-
-
-
-
--
173%
103%
20%
0%
-
-
--
-
-
173%
87%
26%
12%
-
-173%
-173%
-173%
-173%
-173%
--
-173%
-173%
-173%
-173%
-173%
173%
103%
22%
0%
-173%
-173%
-173%
-173%
-173%
43
-------
Table B3. GAT Catalytic Bead Sensor - Ml Mode Measurements (cont.)
(#) Time period for changes in concentrations is 30 minutes.
#
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
H20
ppm
0
0
0
0
0
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
15002
4999
1667
0
15002
15002
15002
15002
15002
15002
15002
15002
4999
1667
0
15002
Operating Power
Gain (ppm butane/mv)
CH4
ppm
0
0
0
0
0
0
500
0
0
0
0
0
500
500
500
500
4999
500
500
500
500
500
500
0
500
500
500
500
500
0
0
0
0
0
0
4999
4999
4999
4999
4999
Zero (V)
C4H10
ppm
0
0
0
0
0
0
0
0
500
0
0
0
500
500
500
500
500
500
4999
500
500
500
500
0
500
500
500
500
500
0
4999
500
4999
0
0
4999
4999
4999
4999
4999
C8H10
ppm
30
100
300
1000
0
0
0
0
0
0
100
0
100
100
100
100
100
100
100
100
1000
100
100
0
100
100
100
100
100
0
0
0
0
0
0
1000
1000
1000
1000
1000
GAT #4
3.8V
56
0.018
0
0
555
1944
0
0
0
0
278
0
0
0
555
555
555
555
555
555
5277
555
2222
555
555
0
555
555
278
278
555
0
4999
278
4999
0
-555
6665
6665
6943
6943
6943
GAT #5
3.8V
45
-0.007
-227
0
227
1591
-227
0
0
0
227
0
227
0
454
454
454
454
454
454
5226
454
1818
454
454
0
454
454
227
227
454
0
4999
227
4999
0
0
6817
6817
6817
6817
6817
GAT #6
3.8V
50
0.013
0
250
750
2000
0
0
0
0
500
0
250
0
750
500
750
500
500
500
5249
750
2000
750
500
0
500
750
750
750
750
0
4999
500
4999
0
-250
6499
6499
6499
6499
6499
Methai
Average
50
0.008
-76
83
511
1845
-76
0
0
0
335
0
159
0
587
503
587
503
503
503
5251
587
2013
587
503
0
503
587
418
418
587
0
4999
335
4999
0
-268
6660
6660
6753
6753
6753
GAT Sensors
10 Insensitiv
Std. Dev.
5
0.013
131
144
264
222
131
0
0
0
145
0
138
0
150
51
150
51
51
51
25
150
202
150
51
0
51
150
288
288
150
0
0
145
0
0
278
159
159
229
229
229
1
} Mode
% Dev.
10%
165%
-173%
173%
52%
12%
-173%
-
--
--
43%
-
87%
-
26%
10%
26%
10%
10%
10%
0%
26%
10%
26%
10%
-
10%
26%
69%
69%
26%
--
0%
43%
0%
-
-104%
2%
2%
3%
3%
3%
44
-------
APPENDIX C.
POLYMER ADSORPTION SENSOR AND METAL OXIDE SEMICONDUCTOR SENSOR TEST DATA
T5
(D
S
r>
TS
O
c
Z>
•
*
^-»
JQ
(Q
35
k_
w
c
CD
CO
c
.g
*>?
Tabled. Polymer Adsorp
O)
c M
" CO
(0 O
'•&
c
0
il
fi
2 uf
"S «"
11
2 «
D JO
$ >»
S-'
81
.2 J
1
55
CD
•d
(5
CD
O)
2
1
O)
*
£
CO
%
2
r*.
2
CO
2
»
2
Tf
2
2
=tt:
2
eg
2
2
Xylene Model Parameters
|
P)
S
r^
!•>-
a
I*
i
•*•
i
,-
s
8
o
CO
8
•*
S
8
CO
1
^
8
£
Gas Constant K
s«
o
in
3
f«*
CO
3
co
£
CM
O
8
CO
8
0
X
,-
CO
«
CO
CM
CO
CO
t^
S
CO
Is.
co
•
S
^-
i
CO
^
•*
0)
CO
^
o
^
to
CM
CO
IO
C\j
T
T-
S
CO
CO
CVJ
~
CO
CM
CM
T
0>
5
O
g
CM
3
o>
1
0)
cv
a
CO
IO
CO
CO
£
CO
o
1
^
co
5
CNJ
R
CO
•«•
00
?
CM
CD
X?
IO
co
5
o
CO
i
IO
3
at
3
p
S
•*
co'
3
CM
IO
?
•*
a
IO
?
CO
CM
?
CM
CM
?
IO
1
,_
IO
-
o
o
Q.
Q.
1
1
8
CD
CD
>,
X
X
CO
r^
3
CO
?
IO
CJ
•*
•*t
S
r^
oi
r*.
3
N.
?
CM
CM
~
Tt
S
^t
f
co
o>
?
o
o
CO
S«
IO
o
3
CM
•*
3
00
CD
CO
3
o
00
$
r*.
§
CM
§
CO
8
CO
•*
CM
00
O
CM
3
IO
§
o
o
o
&
#
£
O
TJ
CO
CD
9
I
s
2
s
2
r*.
4fe
2
CO
%
2
in
*
2
•^r
:«:
2
CO
«
2
sy
*
2
,_
*
2
*- zz~
< 8
C^ z
0 H
.E S
^8
Q. V
x-a *
S &1
1=1 5
3 W CD
= 41
o » S
T; I <
|^
LU
1
•<*
$
^.
i^.
00
CM
q
r^.
c\i
%
o>
co
CO
^.
?'
IO
00
3
•«»
8
h-
s
,_
en
^
o>
3
o
1
CM
%
CM
CM
c3
CM
OC
^
CO
S
co
CM
CO
^_
IO
S
CM
?
O
•*
?
CO
00
oo
o>
co
3
r-.
1
o
CO
i
K
$
^_
^
0)
N.'
8
CM
r*.
CM
*?
.^
CO
CM
•*
o>
co
o
s
en
3
^
•*
CO
0
co
co
o
o
*—
ex
Q
TJ
£
.1
8
CD
CD
>>
X
#
5
00
8
oo
o>
^»
^*
$
o>
8
^
S
»•
CO
o
i
00
g
o
oS
O)
1
IO
£
o
0
CO
£
co
8
^
§
•*
co
^
r^
o>
@
o
co
^
co
IO
S
CM
IO
3
N.
^_
i
CO
o
o
o
1
#
1
o
i
CD
f
1
S
2
$
2
h*
*
2
CO
%
2
U)
*
2
^j-
%
2
2
CM
4*
£
«
2
Cross Sensitivity
Xylene Reading (ppm) when:
j6
^
^
CO
1
i
CM
1^
$
T-
1
1
f
h.
CM
^
fs.
O>
co
K
?
5000 ppm Methane is delivered
#
«
fN.
a
5
to
o>
V
c
N
CO
i
i
o
•*•
co
CO
ff
3
CD
S
co
(x.
S
CM
CO
co
?
CO
cJ
?
5000 ppm Butane is delivered
45
-------
c
o
1
.0
16
O
"c
"o
CL
CO
o
CL
co
tr
CM
o>
T—
I
CO
Q
C7D
>*.
O
CO
c
CD
CO
e-
o
CO
5
CD
_>»
o
CL
c\i
o
CO
O)
i *
m o
O (0
«
a*
2O>
„. ••_
§ o
<
0)
c VL
4) O)
*f
il
Q. k.
ii
% $>
^ 1
(0 TJ
.a >
— "O
sl
o
Except where noted, the absolute
humidity was held at 15,000 ppm HJ
*:
<
0.
co
%
£
1^.
%
<
Q.
£
£
if)
*
2
2
4*
<
CL
S
5
CM
*t
2
5
2
Xylene Model Parameters
5?
o
CO
00
8
r*-
l<
co
r>;
o>
Ot
CM
"t
g
8
"T
£
8
o
CO
3
rf
§
8
CO
8
3
1
Gas Constant K
8?
00
_
8
CO
Tt
CM
h-
cJ
h.
CJ
O
•*
in
CM
at
00
S
CO
£
CM
CO
in
CM
CM
O
Ot
CO
CM
m
CD
Si
,_
h.
CM
CM
CO
Oi
in
CM
Resistance (ft) in Clean Air Rb
CD
o
S*
CO
CM
|5
CM
R
o
£
Tf
o
P:
CM
R
CO
f:
r^
S
^-
F:
^
•t
fc
r^.
Si
CO
P:
CO
R
o
CO
g
o
o
0
0
o
o
o
o
o
o
o
o
o
o
o
8
o
8
o
8
o
o
o
o
8
o
8
o
o
Q.
Q.
•o
£
.i
8
s
X
X
00
r>
CO
CO
CM
CO
CO
CM
CM
o>
00
CM
CM
m
CO
CM
CM
h-.
5
00
t^
%
«<*•
a
00
00
•t
CM
T-
K
00
in
SI
o
o
CO
§
o
5
ti
o
O)
in
0
CVJ
t>:
8
CO
CO
CM
O
r^
in
o
h.
at
o
CM
O)
in
o
m
S
m
%
o
CO
co
o
o
o
o
T-
®
Q
S?
«:
2
3
4*
2
CO
4*
<
CL
OJ
4*
2
T—
4*
2
5 i?
^ i
0 §.
.E CL
-E 8
Q. V
Jt?f
2 i1 I
E =6 x
3 (0 Oj
= 4 1
o ® JS
S ® ^
|£
UJ
1
CM
00
r^
•*'
•t
o
00
CO
r^
d
CO
CO
00
CO
CO
r^
•t
0
CO
•^1
CM
CO
•
o
o>
t^
g
CM
CO
co
CM
in
m
in
CM
CO
CO
S
in
r*.
m
00
CO
CO
i<
N.
•*
CM
00
O
g'
Xylene Delivered (ppm) 1 00
g
o
CO
CM
S
CO
S
CM
in
CM
%
CM
f^
cS
t^
•*
CO
CM
'"t
CO
R
CM
Tt
8
CO
O)
00
CM
in
co
S
o
o
CO
*
CO
9
CO
S
o
in
r^
8
CM
§
CM
g
O
CO
1
1^
r>.
8
at
$
CO
8
o
•*
0
o
o
®
Q
S°
0>
•d
CO
(D
9
o
£
CO
pi
CO
Ri
CM
S
co
R
CM
^
00
R
m
i^.'
CO
o
R
co
R
5000 ppm Butane is delivered
46
-------
c
o
1
.9
"cc
O
•»—•
'o
Q_
CD
CO
C
o
Q.
CO
CD
rr
CM
05
O)
CO
03
Q
!5
co
_
o
CO
C
CD
CO
c
o
CO
5
CD
^
O
Q.
CO
O
s
0)
c ,«
'•B £
(0 O
0 CO
tt §
&*
(0 O>
M. ••-
> °
<
0)
c
V
>\
X
is.
M
§2
° QC
08 8
E i
CL m
O.C/5
§
m S
<0 >
"S ^
CD JE
75 -
.0
1
VJ
a
Except where noted, the absolute
humidity was held at 15,000 ppm H
4*
2
CO
4t
2
r-.
4t
3.
CO
4t
2
m
4t
2
2
%
2
2
2
CM
2
T-
2
Xylene Model Parameters
^
o>
o>
co
O)
h«
8
CO
8
in
CM
,-
1
CM
h>
s
a
in
in
§
CM
O)
00
CO
CO
8
CO
00
8
3
CO
s
8
CO
i
Gas Constant K
i
CO
8
O)
CO
Tt
CM
Tf
hi
CM
T-
co
in
CM
TT
00
Tf
CM
00
h-
r>-
CM
CO
•«fr
CM
CM
en
r~-
CO
CM
CM
CO
a
in
CO
CM
CM
o
00
in
CM
Resistance (ft) in Clean Air Rb
i
#
o
Q
•d
CO
0)
CO
0)
^
en
%
<
Q.
oo
4t
2
h*
%
<
Q_
co
=**
2
m
it
2
2
:«:
2
2
:**
<
Q.
CM
2
T-
=**:
2
Sensitivity
Xylene Readings (ppm)
2
CO
d
,_
15
CO
15
o
R
CO
R
^-
R
CO
s
T-
R
CO
s
CM
R
CO
R
o
2
00
d
CO
R
CO
R
m
R
CM
R
f«.
R
in
fc
oo
R
00
fc
CM
R
O
fc
o
CO
g
o
d
o
8
o
o
o
o
o
o
o
8
o
8
o
I
o
8
o
o
o
o
8
0
8
o
o
Q.
&
T3
£
.1
&
CD
CD
>,
X
#
en
c\i
in
CO
R!
CO
c^
00
CO
CM
CM
•*
O
CM
CM
00
CO
CM
CM
O>
O>
CM
CM
CO
CM
CM
CM
CO
m
CM
CM
O)
SI
m
CO
£
o
o
CO
g
o
d
o
8
o
o
8
o
o
8
o
o
1
o
i
o
1
o
8
o
o
i
o
1
o
1
o
8
CD
O
c£
CD
Q
•d
55
(D
C?
1
o>
4t
<
Q_
CO
4*
2
r^
4*
2
CD
%
2
m
%
2
3
4*
<
0_
2
fvi
4t
2
,_
%
2
Effect of Humidity
Xylene Readings (ppm) in Dry Air
(Absolute Humidity < 50 ppm H2O)
1
o>
r>:
,_
CM
•*
cn
m
CO
in
o>
CM
•*
h.
CO
m
•*
•*
•*
CO
•*
in
o
•t
o
|
o
m
in
CM
CM
m
o
T—
1
CM
CO
h.
CO
•^1
in
o
•*
h-
CO
CO
CO
•*
CM
00
•*
^
00
•*
co
m
^~
in
in
•t
^_
o
CO
CO
!>-
in
o
CO
3?
in
•t
00
CO
£
N.
O)
in
o
CO
in
O)
8
O)
00
CO
00
R
CO
i
in
$
o
s
co
£
8
I
i
.1
&
CD
CD
£
*
h-
O>
•*
hi
CM
T_
tn
%
o
s
o
CO
%
h.
co
h-
CM
m
R
CM
O>
m
«
CO
CM
h-
CM
00
S
^_
in
^
o
8
£
h»
O)
CM
8
o
CO
8
o
o
a
o
CM
8
o
h-
8
o
o>
CM
Tt
O
00
%
o
CO
h-
CM
0
O
8
O
O)
i
i
s?
cS
Q
•o
m
CD
?
1
0)
%
<
CL
CO
%
2
h*.
4fc
2
CO
4*
2
m
4*
2
Tj-
4*
2
CO
4*
2
CM
4fc
2
,_
4fc
2
Cross Sensitivity
Xylene Reading (ppm) when:
&
0
CM
0)
£
•*
%
h-
£
^-
S
^.
N.
oo
s
m
8
^-
S
o
8
in
fe
5000 ppm Methane is delivered
£
00
d
h.
R
,_
R
h.
R
00
h*
o
s
00
R
in
h*
in
R!
h*
CO
?!
5000 ppm Butane is delivered
-------
1
ID
o>
0)
o
CL
co
CD
DC
CM
O)
O)
CO
73
o
!5
co
05
c
CD
CO
CO
CM
GO
CO
O
o
I
O)
.E co
« O
O (0
s
a*
co o
|°
0
0)
£
X
II
§1
;= oc
08 §
E 2
o.(3i
8^
- §
si
P
£
£
<§
S
Except where noted, the absolute
humidity was held at 15,000 ppm H;
9
Q
<£
5
Q
T>
55
2
1
%
CO
CM
CO
10
%
CO
CM
00
s
CO
%
%
8
00
CM
4*
CO
CM
00
5
CO
CM
00
Xylene Model Parameters
£
CM
d
ft
o
In
0
£
0
%
o
•*
o
$
o
2
I
CO
4t
CO
CM
00
10
=fc
CO
CM
CO
$
CO
3
CO
4*
CO
CM
CO
3
CO
CM
CO
5
CO
CM
CO
Sensitivity
Xylene Readings (ppm)
#
5
co
o
r^
CO
r^
h-
*-
CM
CO
U>
•*
•*
(O
CO
03
O
I
•*
r^
(O
SI
•*
c\l
CM
«O
CM
CM
^
^>
O>
CO
in
00
CM
o
co
«
r^
S
N.
3
en
^
to
3
o
CM
IO
co
5
(O
CM
00
s
Xylene Delivered (ppm) 1 00
S
TJ-
S
0>
£
•*
1
CM
to
O>
to
O
CO
CO
IO
CO
en
g
O
o
CO
&
IO
CM
«
h*
IO
3
to
r*.
S
IO
8
CM
CO
£
00
i
CD
S
N.
00
CM
S
I
*°
2
CO
3
.
0
O
o
o
o
o
0
CM
O
O
1
00
o
en
o
CO
CM
CO
o
CM
o
CO
0
IO
o
CO
#
3
O)
CM
IO
•<»
•t
IO
CO
00
CM
CM
to
CO
CM
•*•
r^
I
ex
ex
1
I
&
9
0>
>%
X
5*
Si
00
00
IO
CO
CM
CM
8
IO
co
0)
IO
to
CM
IO
CM
CM
$
8
CO
s
co
3
s
O)
o
a
CM
i
IO
co
oo
o
r-
£
IO
•*
CM
•
2
I
se
CO
CM
CO
8
8
00
$
§
2
CO
CM
CO
2
CO
CM
CO
*t
CO
CM
00
Cross Sensitivity
Xylene Reading (ppm) when:
1
o
IO
*~
CM
o
T-
IO
CO
10
0
o
00
00
IO
!»«.
5000 ppm Methane is delivered
1
en
co
3
N.
o
§8
en
CO
co
CO
CM
£
0
s
to
§
0
£
CM
IO
en
IO
5000 ppm Butane is delivered
48
-------
c
g
I?
.9
"cB
O
"c
'o
Q_
CO
o
Q.
CO
CD
cc
CM
O)
O)
CO
^5
Q
>»
*-*
!5
co
C7D
O
co
c
CD
CO
CO
C\J
CO
CO
o
O
a>
&
CO
O)
c
— (0
"D «-
(0 O
0) CO
QC C
8>*
22
5
c
CD
CL .E
i!
T- C
* 9
E g
Q
OL c/5
8_
CO
T- 3
"S "1
<3 —
.0
0
o
2
CD
CO
4*
CO
CM
00
LO
4k
CO
CM
CO
3
CO
CM
CO
CO
4k
CO
CM
CO
CM
4k
CO
CM
CO
4k
CO
CM
00
Xylene Model Parameters
CM
CM
d
8
o
10
d
k
d
d
d
d
to
d
Power Law Slope B
1
6.73E+04
$
in
•^
cS
S
IO
o
+
CM
IO
CM
§
LU
O
cS
CM
10
cS
+
+
CO
Cri
Resistance (ft) in Clean Air Ro
I
c£
CO
Q
CD
i
5
CO
4k
CO
CM
CO
LO
4fc
CO
CM
CO
5
CO
CM
00
CO
4k
CO
CM
CO
CM
4k
CO
CM
00
5
CO
CM
00
Sensitivity
Xylene Readings (ppm)
IO
en
to
to
CO
CO
en
CM
to
o
*-
r>»
IO
en
*-
o
ir-
CM
s
00
o
8
f—
*
CO
CM
CM
3
CO
00
CM
^
9
O
CO
*
o
d
o
8
o
8
o
8
o
8
o
8
o
o
o
o
8
o
0
Q-
0.
T3
CD
1
CD
_CD
x"
1
co
CO
K
Tt
CO
CM
O
r-~
CM
o
CO
00
8
00
CM
o>
to
CO
CM
o
o
CO
1
IO
O)
CO
*~
0
o
00
00
8
IO
CO
IO
CO
cti
00
IO
CM
CM
o
o
o
CD
Q
o£
i
2
CO
CD
9
5
CO
4k
co
CM
OO
IO
4k
CO
CM
00
s
CO
CM
CO
CO
4k
CO
CM
CO
CM
CO
CM
00
4k
CO
f £
.£ ex
Q. V
>,3, |-
'•& p) E
I i 5
3 CO o>
X o> "3
O G) 55
UJ
CO
IO
d
d
•t
o
CO
o
T-
o
o
o
CO
o
0
i
10
to
0
CM
CO
"*
IO
0
IO
o
CM
o
o
CM
o
CO
t
IO
to
en
en
^
IO
CM
00
CO
T-
CM
o
o
o
0
Q.
ex
1
CD
_CD
Xs
I
co
to
8'
oo
!?
en
s
to
8
IO
8
CM
00
00
*
o
o
CO
fi
co
en
IO
N.
§
CM
§
IO
CM
O
CM
CM
CO
O
fv.
CO
9
O
O
O
1
3?
i
CD
9
*
CO
4k
CO
CM
CO
LO
4k
§
2
CO
co
4k
CO
CM
4k
CO
CM
CO
4k
CO
CM
CO
Cross Sensitivity
Xylene Reading (ppm) when:
2
5
0
00
to
00
CM
A
0
CO
s
fv.
CO
o
5000 ppm Methane is delivered
i
CM
CNJ
CO
IO
IO
h.
$
CO
r*.
IO
^
to
o
to
CM
5000 ppm Butane is delivered
49
-------
o
Is
.Q
"CO
O
"c
'o
Q.
C\l
CD
CO
C
o
CL
CO
CD
CC
CVJ
O)
CO
*—•
Q
.a
co
°
<
0
c
£
X
E "
a *
8*
0 °
c 8
6$
11
*- "O
CO *>
?!
?£
A
0
CVJ
.2 ex
o ex
II
"D CO
•S "O
81
Q) CO
!!
S-P
8 E
X =J
0)
o
55
m
-_j
55
CD
O)
2
CD
41:
CO
CVJ
00
in
4k
CO
CVI
co
*
CO
cvj
CO
CO
CO
CVJ
oo
CVI
CO
CVJ
CO
4fc
CO
CVI
co
del Parameters
i
c
Q
o
CO
1
a
0
CL
I
-
+
UJ
CM
CO
IO
+
HI
*
8
UJ
8
8
UJ
IO
8
UJ
-
8
UJ
1.13E+05
+
in
IN.
CO
O
cc
03
CD
O
§
8
n
CO
CO
CD
CC
O
»
**
]>
**
CO
c
&
|
(V.
IO
CM
•«*
O
0>
CM
CO
CO
CO
s
o>
CO
o
IO
V-
co
CO
$
°
o
5
O)
CO
$
oq
en
Eo
o
CO
<
o
o
o
o
o
o
8
0
o
o
o
o
o
o
8
0
8
o
o
o
o
o
Delivered loom)
CD
C
a
>
X
&
,_
CO
CO
00
CO
CM
CO
CO
CM
CM
CO
£
CM
CO
CO
CO
CM
CM
O
O
CO
«
0
o
0
8
o
o
8
o
0
§
o
8
*-
o
I
o
o
i
o
o
o
I
SN
>
0
Q
•o
CO
CD
0)
at
CD
CO
CVI
CO
in
4t
CO
CVI
oo
3
CO
CVJ
00
8
CO
CVI
CO
CVI
CO
CVI
00
£
CO
CVJ
CO
r^"1 ^
? -
.£ a.
~ ^ 1
2 g §
E 1 o
J cc I
Q — ^^
4) ^
UJ
i
<^
o
IO
o
o
CO
^
CM
0
CO
o
d
CO
o
o
I
CO
*~
^-
CM
CM
CM
CO
"*
O
^~
CO
CM
CO
d
00
o
CO
$
0
•t
CM
a>
CO
o
"fr
10
^
CO
^
a>
CO
IO
a>
o
o
Delivered (opm)
a>
c
">
X
8?
co
JI
•t
CM
O>
5§
O
s
CO
8
IO
CM
0
s
o
*
o
o
CO
%
1
CM
3
o>
CVI
CO
CO
§
CO
IN.
a>
CM
o>
e
CO
8
CO
9
o
0
o
£
a
>
5
o
T3
CO
CD
I
CD
4t
CO
in
CO
CVI
00
3
CO
CVI
CO
8
CO
CVI
CO
CVJ
4*
CO
4*
CO
CVJ
CO
c
CD
$
ex
Q.
o>
>» =0
2 c
«!
0
1
IN.
o>
IO
CM
IO
a>
CM
00
CM
co
CO
o
8
5
co
o>
I
(D
>m Methane is deliv
0
o
o
U)
10
CM
2
cu
CO
CO
00
CO
3
oo
IO
CO
CM
CO
CM
p
R
IN.
0)
c\
1
CD
ppm Butane is deliv
- o
0
8
50
i!rUS. GOVERNMENT PRINTING OFFICE: 1995 • 650-006/22065
------- |