&EPA
United States
Environmental Protection
Agency
REPORT
Cerex Environmental Services
UV Hound Point Sample
Air Monitor
Office of Research and Development
National Homeland Security
Research Center
-------
EPA600/R-06/142
June 2006
Technology Evaluation Report
Cerex Environmental Services
UV Hound Point Sample Air
Monitor
By
Tricia Derringer, Thomas Kelly, Dale Folsom,
and Zachary Willenberg
Battelle
505 King Avenue
Columbus, OH 43201
Eric Koglin
Task Order Project Officer
National Homeland Security Research Center
Office of Research and Development
U.S. Environmental Protection Agency
944 East Harmon Ave.
Las Vegas, NV 89119
-------
Notice
The U.S. Environmental Protection Agency (EPA), through its Office of Research and Development's
National Homeland Security Research Center, funded and managed this technology evaluation through a
Blanket Purchase Agreement under General Services Administration contract number GS23F0011L-3
with Battelle. This report has been peer and administratively reviewed and has been approved for
publication as an EPA document. Mention of trade names or commercial products does not constitute
endorsement or recommendation for use of a specific product.
11
-------
Preface
The U.S. Environmental Protection Agency (EPA) is charged by Congress with protecting the nation's
air, water, and land resources. Under a mandate of national environmental laws, the Agency strives to
formulate and implement actions leading to a compatible balance between human activities and the
ability of natural systems to support and nurture life. To meet this mandate, the EPA's Office of
Research and Development (ORD) provides data and science support that can be used to solve
environmental problems and to build the scientific knowledge base needed to manage our ecological
resources wisely, to understand how pollutants affect our health, and to prevent or reduce environmental
risks.
In September 2002, EPA announced the formation of the National Homeland Security Research Center
(NHSRC). The NHSRC is part of the ORD; it manages, coordinates, and supports a variety of research
and technical assistance efforts. These efforts are designed to provide appropriate, affordable, effective,
and validated technologies and methods for addressing risks posed by chemical, biological, and
radiological terrorist attacks. Research focuses on enhancing our ability to detect, contain, and clean up
in the event of such attacks.
NHSRC's team of world-renowned scientists and engineers is dedicated to understanding the terrorist
threat, communicating the risks, and mitigating the results of attacks. Guided by the roadmap set forth in
EPA's Strategic Plan for Homeland Security, NHSRC ensures rapid production and distribution of
security-related products.
The NHSRC has created the Technology Testing and Evaluation Program (TTEP) in an effort to provide
reliable information regarding the performance of homeland security related technologies. TTEP
provides independent, quality-assured performance information that is useful to decision makers in
purchasing or applying the tested technologies. It provides potential users with unbiased, third-party
information that can supplement vendor-provided information. Stakeholder involvement ensures that
user needs and perspectives are incorporated into the test design so that useful performance information
is produced for each of the tested technologies. The technology categories of interest include detection
and monitoring, water treatment, air purification, decontamination, and computer modeling tools for use
by those responsible for protecting buildings, drinking water supplies, and infrastructure and for
decontaminating structures and the outdoor environment.
The evaluation reported herein was conducted by Battelle as part of TTEP. Information on NHSRC and
TTEP can be found at http://www.epa.gov/ordnhsrc/index.htm.
in
-------
Acknowledgments
The authors wish to acknowledge the support of all those who helped plan and conduct the evaluation,
analyze the data, and prepare this report. We also would like to thank Donald Stedman of the University
of Denver and Stephen Billets and John Zimmerman of the U. S. Environmental Protection Agency's
Office of Research and Development, for their reviews of this report.
IV
-------
Contents
Page
Notice ii
Preface iii
Acknowledgments iv
Abbreviations/Acronyms vii
Executive Summary viii
1.0 Introduction 1
2.0 Technology Description 5
3.0 Quality Assurance/Quality Control 8
3.1 Equipment Calibration 8
3.1.1 Reference Methods 8
3.1.2 Instrument Checks 8
3.2 Audits 9
3.2.1 Performance Evaluation Audit 9
3.2.2 Data Quality Audit 9
3.3 QA/QC Reporting 9
4.0 Test Results 10
4.1 Response Time 11
4.2 Recovery Time 11
4.3 Accuracy 12
4.4 Response Threshold 13
4.5 Temperature and Humidity Effects 13
4.6 Interference Effects 13
4.6.1 False Positive 14
4.6.2 False Negative 14
4.7 Cold-/Hot-Start Behavior 15
4.8 Operational Characteristics 16
5.0 Performance Summary 18
6.0 References 20
Appendix A: Detection and Quantification Limits Calculated by the Vendor of the UV Hound
Based on Spectral Data Recorded in Response Threshold Tests A-l
v
-------
Figures
Figure 2-1. CerexUV Hound 5
Figure 2-2. Generation of Absorbance Spectra 6
Tables
Table 1-1. Target TIC and CW Agent Challenge Concentrations 3
Table 3-1. Performance Evaluation Audit Results 9
Table 4-1. Results from UV Hound Evaluation with Cb at 10 ppm Concentration 10
Table 4-2. Interference Effects 14
Table 4-3. Start State Effects 16
Table A-l. Detection Limits Estimated by Cerex for the UV Hound A-l
VI
-------
Abbreviations/Acronyms
AC
AEGL
CG
CK
C12
cw
DC
DEAE
EPA
FPD
GB
GC
HD
IDLH
ug/mL
ug/m3
ms
NHSRC
PE
ppb
ppm
ppm-m
QA
QC
QMP
RH
SA
THC
TIC
ISA
TTEP
USB
UV
hydrogen cyanide
acute exposure guideline level
phosgene
cyanogen chloride
chlorine
chemical warfare
direct current
N,N-diethylaminoethanol
U.S. Environmental Protection Agency
flame photometric detection
sarin
gas chromatography
sulfur mustard
immediately dangerous to life and health
microgram per milliliter
microgram per cubic meter
millisecond
National Homeland Security Research Center
performance evaluation
part per billion (by volume in air)
part per million (by volume in air)
ppm-meter
quality assurance
quality control
quality management plan
relative humidity
arsine
total hydrocarbon
toxic industrial chemical
technical systems audit
Technology Testing and Evaluation Program
universal serial bus
ultraviolet
vn
-------
Executive Summary
The U.S. Environmental Protection Agency's National Homeland Security Research Center Technology
Testing and Evaluation Program (TTEP) is helping to protect human health and the environment from
adverse impacts as a result of acts of terror by carrying out performance tests on homeland security
technologies. Under TTEP, Battelle evaluated the performance of the Cerex UV Hound point sample air
monitor in detecting toxic industrial chemicals (TICs) and chemical warfare (CW) agents in air.
The UV Hound operates on the principle that, when exposed to light, gases absorb various characteristic
wavelengths of the light to an extent proportional to the amount of gas in the light beam. Each gas has a
unique spectral fingerprint that can be used to identify and quantify gaseous components. In the
UV Hound, a xenon or deuterium lamp produces an ultraviolet light beam. Specially designed optics
focus the beam and project it through the air. A receiver then collects the light and focuses it into a
spectrometer. The spectrometer analyzes the wavelengths and magnitudes of the received light. The
resulting single-beam spectrum (the plot of signal strength versus wavelength of light) contains all of the
spectral information needed to identify and quantify the gases present in the air traversed by the light
beam.
The following performance characteristics of the UV Hound were evaluated in tests with the TIC
chlorine (Cb) as the target gas:
# Response time
# Recovery time
# Accuracy of hazard identification
# Response threshold
# Temperature and humidity effects
# Interference effects
# Cold-/hot-start behavior
# Operational characteristics.
The evaluation included sampling potential indoor interferents, both with and without Cb. The
interferents used were latex paint fumes, air freshener vapors, ammonia cleaner vapors, a mixture of
hydrocarbons representing motor vehicle exhaust, and N,N-diethylaminoethanol (DEAE), a boiler water
additive that can enter indoor air via steam humidification. A range of temperatures (5 to 35°C) and
relative humidities (<20 to 80%) was used to assess the effects of these conditions in detecting Cb.
In addition, response threshold tests also were conducted for the following TICs: hydrogen cyanide
(HCN; North Atlantic Treaty Organization military designation AC), arsine (SA), cyanogen chloride
(CK), and phosgene (CG), and for the CW agents sarin (GB) and sulfur mustard (HD). The UV Hound
had not been tested for detection of these six chemicals before this evaluation. Detection was
investigated at two concentrations, and Cerex personnel estimated quantitative detection limits based on
spectral data recorded during the challenges with these chemicals.
Summary results from testing of the Cerex UV Hound are presented below for each performance
parameter evaluated. Discussion of the observed performance can be found in Chapter 4 of this report.
viii
-------
Response Time: The UV Hound operated by compiling numerous spectra over 30-second intervals;
thus, response time and recovery time were both quantized in 30-second periods. When the UV Hound
responded to 10 ppm Cb challenges (i.e., approximately the immediately dangerous to life and health
[IDLH] level), its response times ranged from 30 to 270 seconds. Response times were higher (150 to
270 seconds) at the high-temperature/high-humidity condition. The response times for room temperature
at low humidity were 30 to 150 seconds, and for room temperature/high humidity, were 30 to
90 seconds.
Recovery Time: When the UV Hound responded to Cb challenges, recovery times ranged from 30
seconds to the maximum of 300 seconds allowed under the test procedures. For room temperature at low
humidity, the range was 60 to 210 seconds; and, at room temperature/high humidity, it was 120 to 240
seconds. Recovery time was highest (300 seconds) at high temperature/high humidity. At high
temperature/medium humidity, the recovery time was 150 seconds.
Accuracy: The UV Hound responded accurately (i.e., produced a stable green light indicating detection
of Cb) in 21 of 46 total challenges with Cb. The UV Hound was 100% accurate in identifying Cb in 16
total challenges delivered at 5 °C and 50% relative humidity (RH), or at 22 °C and either < 20% or 50%
RH. However, in 30 total challenges at 35 °C/50% RH, or at 80% RH at either 22 °C or 35 °C, accuracy
was 10 to 20%. Failure to respond to Cb challenges was the primary form of inaccuracy. Also, in one
trial in the accuracy test, the UV Hound initially indicated detection of Cb, but stopped indicating
detection of Cb while the Cb challenge was going on. Inspection of individual spectral results suggests
that clouding of the mirrors in the optical cell by water and Cb may have adversely affected detection of
Cb in the enclosed optical cell at these conditions.
Response Threshold: The UV Hound was able to detect CK, SA, CG, GB, and HD at challenge
concentrations typically a few times the respective IDLH or AEGL-2 levels. (A quantitative estimation
of detection limits for these five chemicals was conducted by the vendor; results are shown in Appendix
A.) However, the UV Hound did not detect AC at 100 parts per million (ppm) (twice the IDLH level for
that compound).
For Cb, a response threshold of 7.5 ppm was observed, as that was the lowest concentration at which the
majority of trials produced a positive response. At that concentration, three of five successive challenges
produced a positive indication of Cb from the UV Hound.
Temperature and Humidity Effects: The effects of temperature and humidity on the UV Hound are
summarized in the previous paragraphs.
Interference Effects: No interferent when tested alone produced a false positive result. Paint fumes and
ammonia floor cleaner interfered with the detection of Cbby producing false negative results, but air
freshener vapors, engine exhaust hydrocarbons, and DEAE did not.
Other erroneous positive responses were observed when testing accuracy, and in interference testing
with floor cleaner vapors, when the UV Hound indicated detection of Cb while sampling clean air after
the completion of a Cb challenge. Other erroneous negative responses were observed in testing
accuracy, cold-/hot-start behavior, and interferences, when the UV Hound stopped indicating detection
of Cb while the Cb challenge was going on.
ix
-------
Cold-/Hot-Start Behavior: The UV Hound responded to Cb with a stable green light in all five control
(i.e., fully warmed up) challenge cycles and in all five hot temperature cold-start challenge cycles. For
the cold temperature cold-start test, the UV Hound did not respond in any of the five Cb challenges,
almost certainly due to condensation of moisture on the mirrors in the cold optical cell. For the room
temperature cold-start test, the UV Hound responded in all five of the challenge cycles within 30 to 90
seconds, but in four of the challenges, that response ceased before the Cb challenge ended.
Operational Characteristics: The prototype UV Hound and associated software were easy to set up
and use after training provided by Cerex. The unit is designed to be mobile (via vehicle) rather than
portable, and is also suited for fixed-site monitoring. The purchase price of the UV Hound is $49,000 to
$69,000, depending on the options chosen.
Conclusion: The UV Hound detected CK, SA, CG, GB, and HD at concentrations typically a few times
their respective AEGL-2 or IDLH level. However, AC could not be detected with the UV Hound at
twice its IDLH level. A response threshold of about 7.5 ppm was found for Cb. At that level, the
majority of challenges produced a stable green light response, indicating a strong fit to reference spectral
data for Cb.
The UV Hound accurately detected Cb under moderate temperature and humidity conditions (i.e.,
< 22 °C and < 50% RH). However, higher temperature or humidity reduced accuracy to 10 to 20%, and
affected response time and recovery time. The UV Hound spectral data suggest that water/Cb collection
on the mirrors in the optical cell reduced light transmission and detection capabilities under these
conditions. Paint fumes and ammonia floor cleaner interfered with the detection of Cb. Erroneous
positive responses were also observed when sampling clean air, and erroneous negative responses were
seen when the UV Hound stopped indicating the detection of Cb while the Cb challenge was still going
on. The UV Hound responded rapidly and accurately to Cb upon startup, except after cold storage, when
moisture condensed in the cold optical cell of the unit.
x
-------
1.0 Introduction
The U.S. Environmental Protection Agency's (EPA's) National Homeland Security Research Center
(NHSRC) is helping to protect human health and the environment from adverse impacts as a result of
intentional acts of terror. With an emphasis on decontamination and consequence management, water
infrastructure protection, and threat and consequence assessment, NHRSC is working to develop tools
and information that will help detect the intentional introduction of chemical or biological contaminants
in buildings or water systems, the containment of these contaminants, the decontamination of buildings
and/or water systems, and the disposal of material resulting from clean-ups.
NHSRC's Technology Testing and Evaluation Program (TTEP) works in partnership with recognized
testing organizations; with stakeholder groups consisting of buyers, vendor organizations, and
permitters; and with the full participation of individual technology developers in carrying out
performance tests on homeland security technologies. The program evaluates the performance of
innovative homeland security technologies by developing test plans that are responsive to the needs of
stakeholders, conducting tests, collecting and analyzing data, and preparing peer-reviewed reports. All
evaluations are conducted in accordance with rigorous quality assurance (QA) protocols to ensure that
data of known and high quality are generated and that the results are defensible. TTEP provides high-
quality information that is useful to decision makers in purchasing or applying the tested technologies. It
provides potential users with unbiased, third-party information that can supplement vendor-provided
information. Stakeholder involvement ensures that user needs and perspectives are incorporated into the
test design so that useful performance information is produced for each of the tested technologies.
Under TTEP, Battelle recently evaluated the performance of the Cerex UV Hound point sample air
monitor in detecting toxic industrial chemicals (TICs) and chemical warfare (CW) agents in air. This
evaluation was conducted by adapting a peer-reviewed test/QA plan(1) that was developed in accordance
with the requirements of the quality management plan (QMP) for TTEP.(2) Amendments to that test/QA
plan(1) specific to evaluation of ultraviolet (UV) absorption instruments were established prior to this
evaluation, consistent with the requirements of the TTEP QMP.(2)
This evaluation included the first assessment of the UV Hound's ability to detect certain TICs and CW
agents, specifically the TICs hydrogen cyanide (HCN; North Atlantic Treaty Organization military
designation AC), arsine (SA), cyanogen chloride (CK), and phosgene (CG), and the CW agents sarin
(GB) and sulfur mustard (HD). For these six chemicals, the evaluation consisted solely of determining
whether the UV Hound could detect the chemicals at relevant concentrations in air. This determination
was based on recording spectra when challenging the UV Hound with clean air and with two challenge
concentrations of each of the six chemicals. A software program was used that directly recorded
individual spectra, so that results from clean air and challenge concentrations could be inspected to
assess detection of these compounds. In addition, the vendor of the UV Hound analyzed the spectral data
-------
to estimate detection limits for the detected chemicals. The results of the latter analysis are presented in
Appendix A of this report.
A much broader evaluation of UV Hound capabilities was conducted using the TIC chlorine (Cb) as the
challenge chemical. Chlorine was the only target TIC for which spectral information is stored in the
software library of the UV Hound, allowing it to quantify Cb concentrations in air. The software
program automatically compared recorded spectra to reference spectral data and displayed the measured
concentration of the target chemical, along with goodness-of-fit indicators relative to the reference
spectra. A green light indication on the UV Hound's laptop computer display meant a confirmed
detection of Cb, and was triggered by a coefficient of determination (r2) of 0.2 or greater for the spectral
fit. With Cb as the challenge gas, the full range of test procedures called for in the test/QA plan(1) were
conducted, and the following performance parameters of the UV Hound were evaluated:
# Response time
# Recovery time
# Accuracy of hazard identification
# Response threshold
# Temperature and humidity effects
# Interference effects
# Cold-/hot-start behavior
# Operational characteristics.
This evaluation addressed detection of chemicals in the vapor phase, because that application is most
relevant to use in a building contamination scenario. This evaluation took place between October 20,
2005 and January 3, 2006 in two phases: testing with CW agents (conducted in a certified surety
laboratory at Battelle's Hazardous Materials Research Center) and testing with TICs (conducted in a
non-surety laboratory at Battelle). In all evaluations, the UV Hound was operated using an internal
enclosed multi-pass optical cell with a volume of about 8.8 liters and a 14.8-meter optical path length.
Challenge gas flows entered and exited this cell through H-inch diameter ports located near the ends of
the cell. The optical cell that was used in CW agent testing was removed for disposal, and an identical
optical cell was installed in the UV Hound to perform the TIC testing. Evaluations with Cb were
conducted with challenge concentrations that were at the immediately dangerous to life and health
(IDLH) level, as specified in the test/QA plan.(1) Evaluations with AC, SA, CK, CG, GB, and HD were
conducted at concentrations well above the corresponding IDLH or similar levels, to increase the
likelihood of detecting absorption of UV light by these chemicals. Table 1-1 summarizes the challenge
concentrations used.
In response threshold tests for AC, CK, SA, and CG, the challenge concentrations were established
based on the known concentrations of primary source gas mixtures and dilution using mass flow control.
In the evaluations with Cb and in response threshold evaluations with CW agents, challenge
concentrations were confirmed by means of reference analysis of the challenge air stream. The reference
method for Cb was a commercial electrochemical Cb sensor (Drager MiniWarn). The reference method
for GB and HD was gas chromatography with flame photometric detection (GC/FPD), using bags for air
sample collection.(1)
-------
Table 1-1. Target TIC and CW Agent Challenge Concentrations
Chemical
AC
SA
C12
CK
CG
GB
HD
Target
Concentration
100 parts per million (ppm)
and 50 ppm
10 ppm and 5 ppm
10 ppm
100 ppm and 50 ppm
10 ppm and 5 ppm
0.15 ppm and 0.12 ppm
1.06 ppm and 0.79 ppm
IDLH or AEGL-200
Concentration
50 ppm
(50 mg/m3)
3 ppm
(10 mg/m3)
10 ppm
(29 mg/m3)
20 ppm(b)
(50 mg/m3)
2 ppm
(8 mg/m3)
0.037 ppm
(0.2 mg/m3)
0.09 ppm
(0.6 mg/m3)
(a) All values in this column are IDLH levels, except that for HD an acute exposure guideline level (AEGL) is given. The
AEGL-2 value of 0.09 ppm (0.6 mg/m3) for HD is that expected to produce a serious hindrance to efforts to escape in the
general population, based on a 10-minute exposure.
(b) Estimate based on IDLH for hydrogen cyanide (AC).
As described in the test/QA plan,(1) response time, recovery time, and accuracy were evaluated by
alternately challenging the UV Hound with clean air and known vapor concentrations of Cb- It must be
noted that the UV Hound reported a response every 30 seconds, based on numerous individual spectra
integrated over as little as 50 milliseconds (ms) each during that interval. Thus, response and recovery
times are necessarily "quantized" in 30-second intervals. Furthermore, response and recovery times were
limited by the gas changeover rate in the optical cell, which was approximately one air exchange every
1.7 minutes. Each clean air and Cb challenge was supplied for at least 5 minutes, and thus at least three
exchanges of the optical cell volume were achieved in each challenge. Response threshold was
evaluated by repeatedly stepping up from a 1 ppm Cb concentration until a concentration was reached at
which the majority of trials produced a stable green light response. Similar evaluations conducted over
the range of 5 to 35 °C and 20 to 80% relative humidity (RH) were used to establish the effects of
temperature and humidity on detection capabilities for Cb. The effects of potential indoor interferences
were assessed by sampling selected interferences both with and without Cb. The interferences used were
latex paint fumes, ammonia floor cleaner vapors, air freshener vapors, a mixture of gasoline exhaust
hydrocarbons, and N,N-diethylaminoethanol (DEAE), a boiler water additive potentially released to
indoor air by humidification systems. The concentrations of the interferents were checked during the
evaluation by means of a total hydrocarbon (THC) analyzer, calibrated with known concentrations of
propane. The UV Hound unit was also evaluated with Cb after a cold start (i.e., without the usual warm-
up period) from room temperature, from cold storage conditions (5 °C), and from hot storage conditions
(40 °C) to evaluate the delay time before readings could be obtained and the response speed and
accuracy once readings were obtained. Operational factors such as ease of use, data output, and cost
were assessed through observations made by evaluation personnel and through inquiries to the vendor.
-------
QA oversight of this evaluation was provided by Battelle and EPA. As a result of scheduling conflicts,
no technical systems audit (TSA) was performed during this evaluation. However, all test procedures
and equipment had been subjected to TSAs in other recent evaluations conducted under the same
test/QA plan.(1) A data quality audit was conducted on all data from this evaluation.
-------
2.0 Technology Description
This report provides results for the evaluation of the UV Hound point sample air monitor. Following is a
description of the UV Hound, based on information provided by the Cerex vendor. (Contact: Peggy S.
Rackstraw, Vice President Sales, Cerex Environmental Services, Office: 865-777-0462, Fax: 865-777-
0463, Mobile: 703-623-1524 peggy@cerexenv.com.) The information provided below was not verified
in this evaluation.
Cerex air monitoring equipment is designed to continuously examine air to detect, identify, and quantify
chemical airborne threats. The UV Hound (Figure 2-1) uses UV differential optical absorption
broadband spectroscopy to determine the presence and concentration of target and interfering gases.
Target gases include combustion gases, light chain hydrocarbons, TICs, and others. This evaluation
assessed the ability of the UV Hound to also detect CW agents and several TICs for which spectral
information was not previously available in the UV Hound's software library.
The UV Hound operates on the
principle that, when exposed to light,
gases absorb various characteristic
wavelengths of light to an extent
proportional to the amount of gas in
the light beam. Each gas has a
unique spectral fingerprint that can
be used to identify gaseous
components. In the UV Hound, a
xenon or deuterium lamp produces a
UV light beam. Specially designed
optics focus the beam and project it
through the air. A receiver then
collects the light and focuses it into a
spectrometer. The spectrometer
analyzes the wavelengths and
magnitudes of the received light.
The resulting single-beam spectrum
(the plot of signal strength versus
wavelength of light) contains all of
the spectral information needed to identify and quantify the gases present in the air traversed by the light
beam. Quantitative analysis of the spectrum requires transformation of the single-beam spectrum to
reveal the actual absorbance features—the characteristic "fingerprints" that are unique to each chemical
compound.
Figure 2-1. Cerex UV Hound
-------
Figure 2-2 shows spectral information illustrating the principles of detection in the UV Hound. The
upper left portion of Figure 2-2 shows the typical source spectrum of a Cerex UV system that uses a
deuterium light source, along with the spectral locations of characteristic absorption features of several
target gases. The right-hand portion of Figure 2-2 illustrates data spectra in the atmosphere when a target
gas (phosgene) is not present as well as when the gas is present, and the lower left-hand portion of
Figure 2-2 shows the absorbance spectrum generated when the logarithm of the data spectrum is
subtracted from the logarithm of the clean air (background) spectrum. Once absorption spectra are
created, they can be compared to reference spectra (absorption spectra of known concentrations of gas)
to determine the actual concentration of gas in the atmosphere. UV Hound software, running in a laptop
computer connected to the spectrometer, performs this comparison and displays the calculated
concentration of each target species detected, along with indication of the fit to the reference data.
Location of Peak AbsorbanceFeaturesfor Agent and TICS
Figure 2-2. Generation of Absorbance Spectra
-------
A positive detection of a target chemical is indicated by a green light on the laptop display, and is
triggered by a coefficient of determination (r2) exceeding 0.2 for the spectral fit. The UV Hound
evaluated in this effort was equipped with two spectrometers to broaden the detected wavelength range,
and thus required two laptop computers. That extended wavelength range was needed to assess the
detection capabilities of the UV Hound for those six chemicals (AC, CK, SA, CG, GB, and HD) for
which absorbance spectra had not been previously obtained.
The UV Hound is a mobile analytical platform designed to be mounted on a vehicle, carried into a
chemical hotspot, or used as a fixed point monitoring system. The system can be DC-powered by the
cigarette lighter in a vehicle and can be configured with wireless communication so that multiple
systems can be networked together. The UV Hound is a portable, point-sample unit and can have an
internal path-length of up to 24 meters. With a potential dynamic operating range from 0.01 to several
ppm, the UV Hound is designed for quick response applications. The UV Hound may also be fitted with
an optional snorkel attachment, permitting it to remain safely inside a vehicle or shelter while an air
intake hose draws in outside air. The UV Hound ordinarily operates by using its own instrument case as
an open flow-through optical cell, in which a fan pulls ambient air in one end of the chassis and expels it
out the other end. In that mode, the flow rate at the inlet is approximately 10 cubic feet per minute. The
system can also be fitted with an internal optical cell that isolates the sample gas from the surrounding
area; the instrument was used in this manner for the evaluation described in this report.
The UV Hound is 7 by 8 by 35 inches in size and weighs approximately 20 pounds. The price ranges
from $49,000 to $69,000, depending upon options chosen.
-------
3.0 Quality Assurance/Quality Control
QA/quality control (QC) procedures for this evaluation were based on the program QMP(2) and the
test/QA plan.(1)
3.1 Equipment Calibration
3.1.1 Reference Methods
Calibration standards for the CW agents GB and HD were prepared by diluting stock agent to
micrograms per milliliter (^g/mL) concentrations and then injecting a 1-microliter volume of each
standard into the GC/FPD. Calibration was based on a regression of peak area versus amount of agent
injected. New calibration plots were prepared immediately prior to challenging the UV Hound. A GB
calibration plot was prepared on October 26, 2005. An HD calibration plot was prepared on October 21,
2005. The concentrations of the standards used ranged from 0.25 [ig/mL to 2.5 ^g/mL for GB and 0.50
[ig/mL to 7.5 [ig/mL for HD. All calibration plots for both agents were linear, with r2 values of greater
than 0.98.
The reference measurements for Cb relied upon the manufacturer's calibration of the commercial
electrochemical monitor used (Drager MiniWarn).
The THC analyzer used to document the interferent levels provided in testing was calibrated by filling a
25-liter Tedlar bag with a commercial certified 33-ppm propane standard. Since propane is a three-
carbon molecule, this standard constitutes a THC concentration of 99 ppm of carbon. This standard was
used for single-point calibration of the THC analyzer on each test day. Clean air from the analytical
laboratory was used for zeroing the analyzer.
3.1.2 Instrument Checks
The UV Hound was operated and maintained according to the vendor's instructions throughout the
evaluation. The assessment of any maintenance needs was based on predefined diagnostics in the UV
Hound software; no maintenance was needed during this evaluation.
-------
3.2 Audits
3. 2. 1 Performance Evaluation Audit
A performance evaluation (PE) audit was conducted to assess the quality of reference measurements
made in the evaluation. For Cb, the PE audit was performed once during the evaluation by diluting and
analyzing a standard that was independent of the standard used during the evaluation. In each case, the
primary and audit standards were diluted in exactly the same way, and analytical results were then
compared, with allowance for differences in the nominal concentrations of the standards. The target
tolerance for this PE audit was ±20%. Table 3-1 shows that the result of the PE audit was well within the
target tolerance for
Independent PE audit samples do not exist for GB and HD. For the CW agents, check standards of GB
and HD were prepared by individuals other than the staff conducting the reference analyses. The check
standards were prepared in the same way as the reference calibration standards, i.e., by dilution of
military grade agent. The results obtained for these two sets of standards were then compared. For GB,
standards were prepared at concentrations of 0.25, 0.50, 1.0, and 2.5 ^g/mL. All results were within 4%
for the separate standards made by two individuals. For HD, standards were prepared at concentrations
of 0.5, 1.0, 2.5, 5.0, and 7.5 ^g/mL. All results were within 8% for the separate standards made by two
individuals.
Table 3-1. Performance Evaluation Audit Results
TIC
C12
Sample
Standard (Cylinder LL23078)
PE Audit Std (Cylinder QF 8866)
Date of
Audit
12/14/05
Concentration
6,015 ppm
5,811 ppm
Result
12.1 ppm
11.5 ppm
Agreement
(%)
1.6
3.2.2 Data Quality Audit
100 percent of the data acquired during the evaluation were audited. The Battelle Quality Assurance
Manager traced the data from the initial acquisition, through reduction, to final reporting, to ensure the
integrity of the reported results. All calculations performed on the data undergoing the audit were
checked.
3.3 QA/QC Reporting
Each audit was documented in accordance with the test/QA plan(1) and the QMP.(2) Once the audit
report was prepared by the Battelle Quality Manager, it was routed to the Test Coordinator and Battelle
TTEP Program Manager for review and approval. The Battelle Quality Manager then distributed the
final assessment report to the EPA Quality Manager and Battelle staff.
-------
4.0 Test Results
With the TIC Cb, the UV Hound was evaluated for the performance parameters listed in Chapter 1 using
the test procedures in the test/QA plan.(1) Test procedures were based on sets of five or more challenges
with Cb, alternating with intervals of sampling clean air. Only response threshold was evaluated for the
TICs AC, SA, CK, and CG and the CW agents HD and GB. One UV Hound unit was used during both
TIC and CW agent evaluations; the optical cell used in CW agent testing was replaced with an identical
cell before the TIC testing. The following sections summarize the findings of this evaluation; results for
both TIC and CW agents are included.
For Cb, the software program provided with the UV Hound gave an intensity reading for the signal, an
error reading in percent, an r2 value, and a quantitative indication of the concentration every 30 seconds.
The r2 value was a measure of the "fit" of the recorded spectral data for a compound. If the r2 value
exceeded 0.2 for a challenge, a green light on the software display indicated the confirmed presence of
the chemical of interest, in this case Cb. The occurrence of a green light was taken as the indication of
accurate identification of Cb in this evaluation. Results of this evaluation with Cb, including
temperature and humidity effects, are summarized below and presented in Table 4-1. As Table 4-1
shows, accurate responses (i.e., confirmed detection of the target chemical) were observed in 21 of 46
total challenges with Cb. All the inaccurate responses (i.e., failure to indicate detection of Cb) occurred
in the three tests involving high temperature (35 °C) and/or high RH (80% RH). The tests at these three
sets of conditions were repeated at the end of the evaluation (resulting in 10 total trials each) and
confirmed the results seen in the original tests.
Table 4-1. Results from UV Hound Evaluation with C12 at 10 ppm Concentration
Tj,, Environmental
Conditions
C12 Control (22°C - 50% RH)
22°C - <20% RH
22°C - 80% RH
35°C - 50% RH
35°C - 80% RH
5°C - 50% RH
UV Hound Alarms Response Time
Response (Indicated Chemical) Range (Seconds)
Green light (5/5)
Green light (6/6)
Green light (2/10)
NR(8/10)(a)
Green light (1/10)
NR(9/10)
Green light (2/1 0)00
NR(8/10)
Green light (5/5)
C12
C12
C12
C12
C12
C12
30
30-150
30-90
60
150-270
30-90
Recovery Time
Range
(Seconds)
120-180
60-210
120-240
150
300
30-180
NR = No response during the C12 challenge.
^ In three tests, the UV Hound displayed a green light confirming detection of C12 while sampling clean air after completion of the C12
challenge.
^ In one test, the UV Hound stopped displaying a green light while the C12 challenge continued.
10
-------
4.1 Response Time
For all challenges conducted during this testing, the UV Hound was set to collect individual spectra over
50 ms integration intervals and average the collected spectra to provide a response every 30 seconds.
This 30-second interval for response must be noted in considering the response time results in Table 4-1.
For room temperature (22 °C) at medium humidity (50% RH), the UV Hound provided a green light at
the end of the first 30-second cycle for all five challenge cycles (Cb challenge/clean air) conducted. The
low temperature (5 °C) test with medium humidity provided five green light responses with varying
response times, with the green light appearing at 30 seconds for the second challenge, 90 seconds for the
final challenge, and 60 seconds for the other three challenges. The high temperature (35 °C) test with
medium humidity was conducted twice for a total of 10 cycles. Nine out of 10 challenges resulted in no
green light for Cb. The one challenge that did respond with a green light did so at 60 seconds.
For room temperature at low humidity (<20% RH), a green light appeared in all challenges within
30 seconds to 150 seconds, respectively. For this test, six challenge cycles were conducted. Three
challenge cycles had response times of 30 seconds, 90 seconds, and 150 seconds, respectively. The other
three challenge cycles each had a response time of 120 seconds. For the room temperature test at high
humidity (80% RH), 10 challenge cycles were conducted. The UV Hound responded with a green light
for two of the 10 challenges and no green light in five others. In three other challenge cycles, the green
light appeared after the Cb challenge was ended and the UV Hound was sampling clean air. For the two
green light responses that occurred while the UV Hound was being exposed to Cb, the green light
response times were 30 seconds and 90 seconds. For the high temperature at high humidity test, the
UV Hound was exposed to 10 challenge cycles. The UV Hound responded with a green light during two
cycles. In one of those cycles, the UV Hound responded with a green light at 270 seconds. In the other
cycle, the UV Hound provided a response at 150 seconds; however, the response cleared (the green light
went off) while the UV Hound was still being exposed to Cb.
Inspection of the individual UV Hound spectra recorded in the challenges conducted at 35 °C and/or
80% RH showed that they were characterized by low signal intensities and by spectral integration times
far exceeding the intended 50-ms interval. In fact, many of the spectral records reached the maximum
allowable integration time of 1,000 ms. These observations are symptomatic of ineffective transmission
of light from the source to the spectrometer in the UV Hound and suggest clouding or condensation on
the mirrors in the optical cell at these temperature/RH conditions. The mechanism of this clouding is
unknown, but presumably involves an interaction of Cb, water vapor, and the optical cell surfaces under
these conditions. An alternative explanation—misalignment of the optical elements as a result of
temperature changes—seems unlikely because frequent failures to detect Cb also occurred at room
temperature/80% RH. Similarly, misalignment as a result of overpressurization of the optical cells is
ruled out because the gas flow rate through the optical cell was the same in all tests. Note that water
vapor itself does not absorb in the UV range, so spectral interference from water is not a factor.
4.2 Recovery Time
Results of the recovery time analysis are summarized below and presented in Table 4-1.
11
-------
At room temperature with medium humidity, the green light went off at 120 seconds after cessation of
the Cb challenge for the final challenge, 180 seconds for the second challenge, and 150 seconds for the
other three challenges. The low temperature at medium humidity test also had varying recovery times
with 30 seconds for the third and fourth challenges, 150 seconds for the first challenge, and 180 seconds
for the second and final challenges. The high temperature at medium humidity test had only one green
light response, which cleared at 150 seconds.
At room temperature with low humidity, five of the six cycles had a recovery time of 60 seconds. The
final challenge had a recovery time of 210 seconds. At room temperature with high humidity, the two
recovery times were 120 and 240 seconds. At high temperature with high humidity in the one test where
the UV Hound still had a green light at the end of the Cb challenge, the UV Hound recovery time was
recorded as 300 seconds, which was the maximum time allowed in the test procedure for the alarm
response to stop after the end of a Cb challenge. In the case where the UV Hound cleared while still
being challenged with Cb, that clearance occurred 60 seconds after the UV Hound began alarming to the
presence of Cb.
4.3 Accuracy
Results of the accuracy analysis are summarized below and are based on the data presented in Table 4-1.
The UV Hound was considered accurate if it confirmed the presence of Cb by displaying a green light,
indicating a close fit to the reference spectral data, and continued to display that indication until the Cb
challenge stopped.
Accuracy was defined as the proportion of trials in which the unit registered an accurate response to the
challenge. The UV Hound was 100% accurate at room temperature and medium humidity, room
temperature and low humidity, and low temperature and medium humidity (total of 16 challenges). The
UV Hound was 20% accurate at the room temperature/high humidity condition, and 10% accurate at the
high temperature/medium humidity and high temperature/high humidity conditions (total of 30
challenges). At the latter condition, one response was judged inaccurate because the confirmed
indication of Cb stopped during the Cb challenge.
High/Low—The high/low test evaluated the response of the UV Hound to alternating higher or lower
concentrations of the target chemical (Cb). In this high/low test, when a higher challenge (10 ppm) was
delivered first, the UV Hound intensity response did not decrease when the challenge was then switched
to a lower concentration (5 ppm). However, when the lower challenge was delivered first, the
UV Hound intensity response did increase when the challenge was then switched to the higher
concentration. In four out of the six challenge cycles (two high/low/clean air and two low/high/clean
air), the UV Hound responded with a green light within 60 to 120 seconds. The final high/low and
low/high cycles did not provide green light responses. Recovery times for the four challenge cycles that
produced responses ranged from 60 to 240 seconds.
12
-------
4.4 Response Threshold
For the TICs AC, CK, SA, and CG and the CW agents GB and HD, the detection capabilities of the UV
Hound were assessed by challenges with the concentrations shown in Table 1-1. The aim of this
evaluation was to assess whether these chemicals could be detected by the UV Hound, as the UV Hound
had not been challenged with these chemicals before. The result of this evaluation was that CK, SA, CG,
GB, and HD all produced detectable spectral response at both of the respective challenge concentrations
shown in Table 1-1. Thus, the UV Hound, as tested, is capable of detecting these five chemicals at the
challenge concentrations, which were typically a few times the respective IDLH or AEGL-2 level.
However, no response was observed in either challenge with hydrogen cyanide (AC).
The vendor of the UV Hound also performed a quantitative evaluation of the spectral data from the
response threshold tests for CK, SA, CG, GB, and HD. The results of that evaluation are presented in
Appendix A.
A response threshold test was also conducted for Cb by starting at a low concentration (1 ppm) and
stepping the concentration up until the UV Hound responded with a green light for the majority of the
challenge cycles at a concentration. At 1 ppm and 2 ppm, the UV Hound showed no green light
response. At 5 ppm, one of four cycles produced a green light. At 7.5 ppm, the UV Hound responded
with a green light in three of the five cycles within 120 to 180 seconds. This result indicates a response
threshold for Cb of approximately 7.5 ppm.
4.5 Temperature and Humidity Effects
The effects of temperature and humidity on the UV Hound are summarized in Sections 4.1 through 4.3.
4.6 Interference Effects
Five interferents (latex paint fumes, ammonia floor cleaner vapors, air freshener vapors, gasoline engine
exhaust hydrocarbons, and DEAE) were used in the evaluation. The concentrations of paint fumes, floor
cleaner vapors, air freshener vapors, and gasoline exhaust hydrocarbons were monitored by a total
hydrocarbon monitor and maintained at 10, 10, 1, and 2.5 parts per million carbon (ppm C),
respectively.(1) DEAE was delivered from a compressed gas standard and diluted to a concentration of
10 parts per billion (ppb) carbon (ppb C).(1) The effect of these interferences on the UV Hound response
to Cb is summarized below and in Table 4-2.
Because of the reactivity of Cb, the potential existed for reduction of the delivered Cb challenge level
due to reaction with the interferent vapors. To avoid this, the Cb level was monitored with the
electrochemical sensor at the outflow port of the UV Hound's optical cell, as well as at the inflow, and
the delivered Cb standard flow was increased as necessary to maintain approximately 10 ppm Cb in the
cell. Such an adjustment was needed only when using the ammonia floor cleaner as interferent. There
was no indication that the electrochemical sensor responded to any reaction product of ammonia with
the floor cleaner vapors.
13
-------
Table 4-2. Interference Effects
TICorCW ¥ , ,
. Interferent
Agent
C12 Control
Paint Fumes
Floor Cleaner
Air Freshener
Gasoline Engine Exhaust
DEAE
UV Hound
Response
Green light (5/5)
Green light (l/5)(a)
NR (4/5)
NR (10/10) ^
Green light (5/5) (a)
Green light (5/5)
Green light (5/5) (a)
Alarms
(Indicated
Chemical)
C12
C12
-
C12
C12
C12
Response Time
Range (Seconds)
30
120
-
30-90
30-90
30-120
Recovery Time
Range (Seconds)
120-180
-
-
30
0-120
30-150
NR = No response.
" In one or more trials, the UV Hound stopped displaying a green light while the C12 challenge continued.
^ In two trials, the UV Hound displayed a green light confirming detection of C12 while sampling clean air after completion of C12
challenge.
4.6.1 False Positive
A false positive was defined as a response from the UV Hound when challenged with an interferent in
air in the absence of a TIC or CW agent. None of the five interferents produced a false positive in the
five trials conducted for each interferent.
A different type of erroneous positive response was observed in a few challenges with Cb, when the
UV Hound showed a confirmed identification of Cb while the optical cell was being purged with clean
air after the end of a Cb challenge. Such occurrences are noted in Section 4.1 through 4.3. Similar
erroneous positive responses were also noted after challenges with Cb and interferences together, as
noted in Table 4-3 and described below in Section 4.6.2.
4.6.2 False Negative
A false negative response was defined as a reduction or elimination of response to Cb, when Cb and the
interferent were present together in the challenge. In the control test (i.e., Cb challenge with no
interferent), the UV Hound identified the Cb challenge in all tests with a stable green light response
within 30 seconds.
For the paint fumes with Cb, the UV Hound responded with a green light in only one out of five
challenge cycles at 120 seconds. That response then cleared before the end of the Cb/interferent
challenge.
For the ammonia floor cleaner vapors with Cb, 10 challenge cycles were conducted. In two cycles, the
green light came on after the Cb/interferent challenge was complete and the UV Hound was sampling
clean air. In the other eight challenges, no green light responses occurred. The UV Hound's intensity
readings for this test varied widely from large negative intensities to large positive intensities.
14
-------
When Cb was tested in the presence of air freshener vapors, the UV Hound responded with a green light
within 30 to 90 seconds for all five challenge cycles. However, in four of the five cycles, the UV Hound
green light then went out prior to completing the Cb/interferent challenge cycle. The time from first
response to clearing of the green light was 120 seconds in most cases.
For the gasoline engine exhaust interferent with Cb, the UV Hound green light came on in all five cases
within 90 seconds, and three of the five cycles responded within 30 seconds. The green light remained
on until the Cb/interferent challenge was replaced with clean air. The recovery times ranged from 0 to
120 seconds.
For Cb in the presence of DEAE, a green light was recorded for all five challenge cycles. The response
times ranged from 30 to 120 seconds. In one of the five challenge cycles, the green light went out prior
to completing the Cb/interferent challenge. For the other four challenge cycles, the green light indication
remained stable until the clean air challenge began, and recovery times ranged from 30 to 150 seconds.
These results show both actual false negative responses, i.e., failure of the UV Hound to identify the Cb
challenge in the presence of the interferent, and another type of erroneous negative response, in which
the UV Hound's accurate identification of Cb ceased while the Cb/interferent challenge continued. The
former were observed with floor cleaner vapors and latex paint fumes, and the latter with air freshener
vapors, latex paint fumes, and DEAE. Neither was observed with engine exhaust hydrocarbons. The
latter type of erroneous negative response was also observed during testing of cold-/hot-start behavior,
as described in Section 4.7.
4.7 Cold-/Hot-Start Behavior
Analysis of the effects of insufficient warm-up time, under start-up conditions ranging from cold (5 to
8 °C) to hot (40 °C), are summarized below. Table 4-3 illustrates the data obtained in testing for cold-
/hot-start effects, showing the start condition, sequential experiment number, response reading, response
and recovery times, and indicated chemical.
In the control test, the UV Hound responded to each Cb challenge with a confirmed (i.e., stable green
light) identification within 30 seconds after the start of the challenge, and cleared within 120 to 180
seconds after the Cb challenge ended.
For the room temperature cold-start test, the UV Hound was held at room temperature overnight. The
UV Hound was turned on at 10:02 a.m. and the first test was conducted at 10:03 a.m. The UV Hound
responded with a green light in all of five challenge cycles. However, in four of the five cycles, the
UV Hound's green light went out prior to completing the Cb challenge.
For the cold-temperature, cold-start test, the UV Hound was stored overnight at 5 to 8 °C. The
UV Hound was removed from the cold storage at 8:43 a.m., and the first challenge was conducted at
8:47 a.m. Testing personnel noted that almost immediately the scan times of the UV Hound increased
greatly. The green light did not come on during any of the five Cb challenges, and the intensity readings
of the UV Hound continuously decreased as the challenge cycles were conducted. It is highly likely that
15
-------
Table 4-3. Start State Effects
UV Houn
dUnit
Start Condition
Control
Room
Temperature
(Cold Start)
Cold
Temperature
(Cold Start)
Hot
Temperature
(Cold Start)
Experiment
Number
1
2
o
3
4
5
1
2
o
3
4
5
1
2
o
3
4
5
1
2
o
3
4
5
UV Hound
Response
Green light
Green light
Green light
Green light
Green light
Green light
Green light
Green light
Green light
Green light
NR(b)
NR(b)
NR(b)
NR(b)
NR(b)
Green light
Green light
Green light
Green light
Green light
Response
Time
(Seconds)
30
30
30
30
30
90
60
60
30
60
-
-
-
-
-
60
30
30
30
60
Recovery
Time
(Seconds)
150
180
150
150
120
.(a)
_(a)
30
.(a)
.(a)
-
-
-
-
-
150
180
210
150
120
Alarm
(Indicated
Chemical)
C12
C12
C12
C12
C12
C12
C12
C12
C12
C12
-
-
-
-
-
C12
C12
C12
C12
C12
^'Green light indication ceased before end of CI2 challenge.
(b) No response, likely due to condensation of moisture on the mirrors in the optical cell.
this behavior was the result of moisture condensing on the cold surfaces of the mirrors in the optical cell,
thereby obscuring the UV light beam, as the moisture content of the challenge air (at 22 °C and 50%
RH) exceeded the saturation vapor pressure at the temperature to which the UV Hound had been cooled
overnight. This suggestion is supported by the return of the UV Hound to normal operation after it had
warmed in the laboratory.
For the hot-temperature, cold-start test, the UV Hound was held overnight at approximately 40 °C. The
UV Hound was removed from storage at 8:28 a.m., and the first test was conducted at 8:30 a.m. The
UV Hound produced a stable green light in all five of the challenge cycles within 30 to 60 seconds. The
recovery time after completing the challenge cycle was 120 to 210 seconds.
4.8 Operational Characteristics
The UV Hound tested was a prototype, in that a second spectrometer had been installed to extend the
wavelength region detected. This was done to evaluate detection of the four TICs and two CW agents
that had not been detected previously by the instrument. Each spectrometer was connected to a laptop
computer with a universal serial bus (USB) cable. Two types of software were used, a basic program for
recording raw spectral data, and an advanced program that compared recorded spectra to a library of
spectral reference data to identify and quantify selected TICs. The former software was used for
evaluating response thresholds for AC, CK, SA, CG, GB, and HD; the latter software was used in all
16
-------
evaluations with Cb. The UV Hound and software were easy to set up and use after training provided by
Cerex.
The UV Hound operation was simple. The deuterium lamp and internal fan were turned on by
connecting the electrical power cord to the UV Hound itself. Each of the two spectrometers was
powered by its respective laptop computer. All control of the UV Hound was through the laptop
software, which always started correctly and performed reliably. The instrument manual was well
written and clear, including its instructions for use of the software.
The UV Hound is most suited for monitoring at a fixed site to detect or diagnose chemical
contamination. Its reliance on a laptop computer for control and data acquisition limits its portability,
though it is designed for mobile operation aboard a vehicle.
17
-------
5.0 Performance Summary
Summary results from the evaluation of the Cerex UV Hound are presented below. Discussion of the
observed performance can be found in Chapter 4 of this report. All results are from testing one
UV Hound unit.
Response Time: The UV Hound operated by compiling numerous spectra over 30-second intervals;
thus, response time and recovery time were both quantized in 30-second periods. When the UV Hound
responded to 10 ppm Cb challenges (i.e., approximately the IDLH), its response times ranged from 30 to
270 seconds. Response times were higher (150 to 270 seconds) at the high-temperature/high-humidity
condition. The response times for room temperature at low humidity were 30 to 150 seconds, and for
room temperature/high humidity, were 30 to 90 seconds.
Recovery Time: When the UV Hound responded to Cb challenges, recovery times ranged from
30 seconds to the maximum of 300 seconds allowed under the test procedures. For room temperature at
low humidity, the range was 60 to 210 seconds; and, at room temperature/high humidity, it was 120 to
240 seconds. Recovery time was highest (300 seconds) at high temperature/high humidity. At high
temperature/medium humidity, the recovery time was 150 seconds.
Accuracy: The UV Hound responded accurately (i.e., produced a stable green light indicating detection
of Cb) in 21 of 46 total challenges with Cb. The UV Hound was 100% accurate in identifying Cb in 16
total challenges delivered at 5 °C and 50% RH, or at 22 °C and either < 20% or 50% RH. However, in 30
total challenges at 35 °C/50% RH, or at 80% RH at either 22 °C or 35 °C, accuracy was 10 to 20%.
Failure to respond to Cb challenges was the primary form of inaccuracy. Also, in one trial in the
accuracy test, the UV Hound initially indicated detection of Cb, but stopped indicating detection of Cb
while the Cb challenge was going on. Inspection of individual spectral results suggests that clouding of
the mirrors in the optical cell by water and Cb may have adversely affected detection of Cb at these
conditions.
Response Threshold: The UV Hound was able to detect CK, SA, CG, GB, and HD at challenge
concentrations typically a few times the respective IDLH or AEGL-2 levels. (A quantitative estimation
of detection limits for these five chemicals was conducted by the vendor; results are shown in Appendix
A.) However, the UV Hound did not detect AC at 100 ppm (twice the IDLH level for that compound).
For Cb, a response threshold of 7.5 ppm was observed, as that was the lowest concentration at which the
majority of trials produced a positive response. At that concentration, three of five successive challenges
produced a positive indication of Cb from the UV Hound.
Temperature and Humidity Effects: The effects of temperature and humidity on the UV Hound are
summarized in the previous paragraphs.
18
-------
Interference Effects: No interferent when tested alone produced a false positive result. Paint fumes and
ammonia floor cleaner interfered with the detection of Cbby producing false negative results, but air
freshener vapors, engine exhaust hydrocarbons, and DEAE did not.
Other erroneous positive responses were observed when testing accuracy, and in interference testing
with floor cleaner vapors, when the UV Hound indicated detection of Cb while sampling clean air after
the completion of a Cb challenge. Other erroneous negative responses were observed in testing
accuracy, cold-/hot-start behavior, and interferences, when the UV Hound stopped indicating detection
of Cb while the Cb challenge was going on.
Cold-/Hot-Start Behavior: The UV Hound responded to Cb with a stable green light in all five control
(i.e., fully warmed up) challenge cycles and in all five hot temperature cold-start challenge cycles. For
the cold temperature cold-start test, the UV Hound did not respond in any of the five Cb challenges,
almost certainly due to condensation of moisture on the mirrors in the cold optical cell. For the room
temperature cold-start test, the UV Hound responded in all five of the challenge cycles within 30 to
90 seconds, but in four of the challenges, that response ceased before the Cb challenge ended.
Operational Characteristics: The prototype UV Hound and associated software were easy to set up
and use after training provided by Cerex. The unit is designed to be mobile (via vehicle) rather than
portable, and is also suited for fixed-site monitoring. The purchase price of the UV Hound is $49,000 to
$69,000, depending on the options chosen.
Conclusion: The UV Hound detected CK, SA, CG, GB, and HD at concentrations typically a few times
their respective AEGL-2 or IDLH level. However, AC could not be detected with the UV Hound at
twice its IDLH level. A response threshold of about 7.5 ppm was found for Cb. At that level, the
majority of challenges produced a stable green light response, indicating a strong fit to reference spectral
data for Cb.
The UV Hound accurately detected Cb under moderate temperature and humidity conditions (i.e.,
< 22 °C and < 50% RH). However, higher temperature or humidity reduced accuracy to 10 to 20%, and
affected response time and recovery time. The UV Hound spectral data suggest that water/Cb collection
on the mirrors in the optical cell reduced light transmission and detection capabilities under these
conditions. Paint fumes and ammonia floor cleaner interfered with the detection of Cb. Erroneous
positive responses were also observed when sampling clean air, and erroneous negative responses were
seen when the UV Hound stopped indicating the detection of Cb while the Cb challenge was still going
on. The UV Hound responded rapidly and accurately to Cb upon startup, except after cold storage, when
moisture condensed in the cold optical cell of the unit.
19
-------
6.0 References
1. Technology Testing and Evaluation Program Test/QA Plan for Evaluation of Portable Ion Mobility
Spectrometers for Detection of Chemicals and Chemical Agents, Version 1, Battelle, Columbus,
Ohio, February 2005, as amended for testing of UV absorption instruments.
2. Quality Management Plan (QMP) for the Technology Testing and Evaluation Program (TTEP),
Version 1, Battelle, Columbus, Ohio, January 2005.
20
-------
Appendix A:
Detection and Quantification Limits Calculated by the
Vendor of the UV Hound Based on
Spectral Data Recorded in Response Threshold Tests
As described in Sections 1.0 and 4.4 of the report, for AC, CK, SA, CG, GB, and HD, the detection
capabilities of the UV Hound were assessed by challenging the UV Hound with two concentrations of
each TIC or CW agent. Cerex personnel compared the recorded spectra with spectra from clean air
challenges, and calculated detection limits for each chemical detected. The results are shown in Table A-
1 for each chemical that could be detected. Note that both minimal detection limit and quantification
limit are shown in Table A-l. Detection limit is defined as the concentration that would produce a
response equal to three times the standard deviation of the baseline response. Quantification limit is the
concentration that would produce a response equal to nine times the standard deviation; i.e.,
quantification limit is three times detection limit. The units of ppm-meters (ppm-m) shown in Table A-l
reflect the importance of both chemical concentration and optical path length in contributing to the
detection of a target chemical.
Table A-l. Detection Limits Estimated by Cerex for the UV Hound
TIC/CW Agent
CK
SA
CG
GB
HD
Detection Limit
(ppm-m)
111
3.0
1.8
0.054
0.69
Quantification
Limit (ppm-m)
332
9.0
5.3
0.16
2.1
Minimum Sample
Path (meters)(a)
16
3.0
2
4.7
23
ND = No detection.
NA = Not applicable.
^ Minimum path length needed to achieve the limits shown.
The calculated detection and quantification limits shown in Table A-l can be converted to
corresponding concentrations in air, by dividing the values shown in ppm-m by the path length of the
cell (i.e., 14.8 meters). When this calculation is performed, the resulting detectable and quantifiable
concentrations for SA, CG, and GB are lower than the respective IDLH levels listed for these
compounds in Table 1-1. For CK and HD, the resulting detectable concentrations are lower, and the
resulting quantifiable concentrations slightly higher, than the respective IDLH level and AEGL-2 listed
A-l
-------
for these compounds in Table 1-1. Overall, the results in Table A-l suggest that the UV Hound is
capable of detecting these chemicals at concentrations near or below their immediately dangerous levels.
A-2
------- |