THE ENVIRONMENTAL TECHNOLOGY VERIFICATION
PROGRAM
4»EPA Baitelle
U.S. Environmental Protection Agency ^e Business of Innovation
ETV Joint Verification Statement
TECHNOLOGY TYPE: Trace Metals Analysis System
APPLICATION: ANALYSIS OF ARSENIC IN WATER
TECHNOLOGY
NAME: SafeGuard Trace Metals Analyzer
COMPANY: TraceDetect
ADDRESS: 180 North Canal Street PHONE: (206)523-2009
Seattle, Washington 98103 FAX: (206) 523-2042
WEB SITE: www.tracedetect.com
E-MAIL: richardb(S>tracedetect.com
The U.S. Environmental Protection Agency (EPA) has established the Environmental Technology
Verification (ETV) Program to facilitate the deployment of innovative or improved environmental
technologies through performance verification and dissemination of information. The goal of the ETV
Program is to further environmental protection by accelerating the acceptance and use of improved and cost-
effective technologies. ETV seeks to achieve this goal by providing high-quality, peer-reviewed data on
technology performance to those involved in the design, distribution, financing, permitting, purchase, and use
of environmental technologies. Information and ETV documents are available at www.epa.gov/etv.
ETV works in partnership with recognized standards and testing organizations, with stakeholder groups
(consisting of buyers, vendor organizations, and permitters), and with individual technology developers. The
program evaluates the performance of innovative technologies by developing test plans that are responsive to
the needs of stakeholders, conducting field or laboratory tests (as appropriate), 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 adequate quality are generated and that the results
are defensible.
The Advanced Monitoring Systems (AMS) Center, one of six technology areas under ETV, is operated by
Battelle in cooperation with EPA's National Exposure Research Laboratory. The AMS Center evaluated the
performance of the TraceDetect SafeGuard Trace Metals Analyzer in measuring total arsenic in water. This
verification statement provides a summary of the test results.
-------
VERIFICATION TEST DESCRIPTION
The SafeGuard was verified by comparing its arsenic measurements to those from a laboratory-based
reference method—inductively coupled plasma mass spectrometry (ICP-MS) performed according to EPA
Method 200.8. The SafeGuard performance was verified by analyzing laboratory-prepared performance test
(PT) samples, quality control (QC) samples, and environmental samples. All samples were tested using both
the SafeGuard and the reference method. The SafeGuard was verified by evaluating accuracy, precision,
linearity, method detection limit (MDL), matrix interference effects, operator bias, inter-unit reproducibility,
and rate of false positives/false negatives.
Samples were prepared and analyzed according to the vendor's recommended procedures and the test/QA
plan. All samples were analyzed without pretreatment except the drinking water samples collected from
plumbing. These samples were filtered to remove potential copper contaminants. The results from the
SafeGuard were compared to those from the reference method to assess accuracy and linearity. Four aliquots
of PT samples and environmental samples were analyzed to assess precision. Seven aliquots of a 5 ppb PT
sample were analyzed to assess the detection limit of the SafeGuard. Potential matrix interference effects
were assessed by challenging the SafeGuard with PT samples of 10 ppb arsenic concentration that contained
both low levels and high levels of potentially interfering substances. All samples were analyzed using two
identical SafeGuard units (designated Unit #1 and Unit #2). Results of analyses from the two units were
statistically compared to evaluate inter-unit reproducibility. Operator bias was assessed by statistically
comparing data from two operators (technical and non-technical) analyzing identical sets of samples on both
units. The rates of false positive and false negative results were evaluated relative to the 10-parts-per-billion
(ppb) maximum contaminant level for arsenic in drinking water. Other factors that were qualitatively assessed
during the test included ease of use, time required for sample analysis, and reliability.
QA oversight of verification testing was provided by Battelle and EPA. Battelle QA staff conducted a technical
systems audit, a performance evaluation audit, and a data quality audit of 10% of the test data.
This verification statement, the full report on which it is based, and the test/QA plan for this verification test are
all available atwww.epa.gov/etv/centers/centerl.html.
TECHNOLOGY DESCRIPTION
The following description of the SafeGuard is based on information provided by the vendor. This technology
description was not verified in this test.
TraceDetect's SafeGuard is designed to automatically measure total arsenic concentrations in drinking water
samples (including raw water and treated water) over a range from 1 ppb to over 100 ppb. Once the operator
has introduced the sample vial and selected "measure" on the control computer, all calibrations, dilutions,
reductions, standard additions, and measurements are performed by the SafeGuard with the results displayed
and logged in a data file. The SafeGuard consists of three main components: the expert system, the fluidics
system, and TraceDetect's patented NanoBand™ sensor and potentiostat. Each of these components has a
part in the measurement process—from controlling the pumps, to adding chemicals, to making measurements
and interpreting the results. The SafeGuard uses anodic stripping voltammetry (ASV) and the method of
standard addition to make metals measurements. ASV is an electro-analytical method that detects ions in a
solution by the potential at which they oxidize and strip away from the surface of an electrode. The SafeGuard
is able to measure As (III) and reduce As (V) to As (III) to measure total arsenic. It can be configured to
analyze copper, lead, zinc, cadmium, and mercury in water.
The SafeGuard stores data for every measurement and operation. The base of the SafeGuard is 15 inches by
28 inches (381 mm by 711 mm). It is 22 inches (559 mm) high and requires a computer, mouse, monitor, and
keyboard. The TraceDetect SafeGuard as configured for measuring arsenic during this verification test was
priced at $35,000, excluding options that the customer may require for unique sample preparation (e.g.,
copper removal from samples, filters for high turbidity samples).
-------
VERIFICATION RESULTS
Accuracy was assessed by comparing the results to Method 200.8(2) results from ICP-MS analysis. The
quantitative assessment of accuracy indicated that the relative bias for the SafeGuard ranged from -28% to 7%
for the technical operator and -28% to 11% for the non-technical operator (excluding residential well water
samples at approximately 600%).
Precision was assessed by analyzing four replicates of each sample. For the technical operator, precision
expressed as relative standard deviation (RSD) ranged from 3% to 44%, and for the non-technical operator
2% to 38%. The average RSD for PT samples only was 10% for the technical operator and 9% for the non-
technical operator. These results exclude samples for which one or more of the replicate results were not
detected by the SafeGuard.
The linearity of response was evaluated by plotting the SafeGuard results against the reference analysis results
for the PT samples. The table below summarizes the equations for the linear regressions and presents the 95%
confidence interval for the slopes as +/- error. All linear regressions against the reference method results had
coefficients of determination (r2) greater than 0.99. The 95% confidence intervals for the slopes indicate that
only the technical operator data for Unit # 1 were consistent with a slope of 1 and were not significantly
different from the reference analysis results. The 95% confidence intervals for the y-axis intercept included
zero for both operators on both units indicating no significant difference from the reference analysis results.
Summary of Linear Regression Equations for SafeGuard and Reference Results
Slope Intercept Coefficient of
Description (+/- Error) (+/- Error) Determination
Unit #1, technical operator
Unit #2, technical operator
Unit #1, non-technical operator
Unit #2, non-technical operator
1.005 (0.044)
0.808 (0.034)
0.874 (0.027)
0.796 (0.019)
-1.618(2.32)
0.060(1.70)
0.155(1.27)
0.960 (0.96)
0.9942
0.9936
0.9961
0.9979
The MDL was assessed by analyzing seven replicates of a sample spiked at a level approximately five times
the manufacturer's estimated detection limit for the SafeGuard (i.e., 1 ppb x 5 = 5 ppb). The MDLs calculated
using the precision data from these replicates ranged from 2.0 ppb to 3.8 ppb.
Results for samples containing low and high levels of interfering compounds indicated that neither level of
interference appeared to affect the detection of arsenic, with bias ranging from -19% to 7%, consistent with
the bias observed in the absence of interferences. The SafeGuard performance was affected by one of the
environmental samples, the residential well water. The native (unspiked) replicates of this sample from both
operators and both SafeGuard units reported an arsenic concentration from 2.50 ppb to 9.70 ppb, whereas the
reference method reported this sample at 0.89 ppb to 1.12 ppb.
The equations for the linear regressions that were performed to evaluate operator bias and inter-unit
reproducibility are summarized below. The 95% confidence interval includes a slope of 1 for Unit # 2, but the
95% confidence interval does not include a slope of 1 for Unit # 1, indicating a significant operator bias
(technical results > non-technical results) with that unit. Paired t-tests of the two sets of data indicate that the
SafeGuard results were not significantly different at a 0.05 level of significance depending on the operator.
Overall, these results indicate at most a small operator bias with one of the two SafeGuard units.
Inter-unit reproducibility was evaluated by comparing the data for the two SafeGuard units used by the technical
and non-technical operators. Linear regressions of the data for each unit show that Unit # 2 readings were lower
than Unit # 1 readings with both operators, but more strongly with the technical operator. Neither 95%
-------
confidence interval includes a slope of 1, indicating a significant inter-unit bias that is more pronounced with the
technical operator than with the non-technical operator. A paired t-test of the data indicated that the results from
the two units with the technical operator were significantly different at a 0.05 level of significance; however, the
results from the two units with the non-technical operator were not significantly different. Overall, these results
show an inter-unit bias with the technical operator, but minimal bias with the non-technical operator.
Summary of Linear Regression Equations for Assessing Operator Bias and Inter-unit Reproducibility
Slope Intercept Coefficient of
Description (+/- Error) (+/- Error) Determination
Unit #1, Operator bias
Unit #2, Operator bias
Technical operator, Inter-unit reproducibility
Non-technical operator, Inter-unit reproducibility
0.872 (0.033)
0.983 (0.033)
0.802 (0.029)
0.907 (0.023)
0.957(1.18)
0.566 (0.98)
1.208(1.03)
0.864 (0.74)
0.9886
0.9909
0.9897
0.9947
A false positive was defined as a SafeGuard result that was greater than 10 ppb and greater than 125% of the
reference concentration, when the reference concentration was less than or equal to 10 ppb. The rates of false
positives for the SafeGuard were 0% for both units for the technical operator and 2% and 0% for the non-
technical operator (Units # 1 and # 2, respectively). A false negative was defined as a SafeGuard result that
was less than or equal to 10 ppb and less than 75% of the reference concentration, when the reference
concentration was greater than 10 ppb. The rates of false negatives for the SafeGuard units were 4% and 22%
for the technical operator and 18% for both units for the non-technical operator. By averaging these rates, the
results indicate that the SafeGuard correctly identified water below the federal drinking water standard (<10
ppb) 99.5% of the time (0.995 = sensitivity) and identified water that did not meet the federal standard
(>10 ppb) 85.4% of the time (0.854 = specificity).
The SafeGuard system was easy to use, and the manual and software program were clear and easy to follow.
All reagent mixing and instrument flushing are automated. No solution or sample preparation was necessary.
The analysis time per sample at room temperature was 30 to 50 minutes. The listed price for SafeGuard at the
time of the verification test was $35,000, and the cost for a 45-sample reagent kit was $80. Replacement
reagents and supplies are available without purchasing entire kits.
Original signed by Gregory A. Mack 08/22/06 Original signed by Linda Shelton for 09/12/06
Gregory A. Mack Date Lawrence W. Reiter Date
Vice President Director
Energy, Transportation, and Environment Division National Exposure Research Laboratory
Battelle Office of Research and Development
U.S. Environmental Protection Agency
NOTICE: ETV verifications are based on an evaluation of technology performance under specific,
predetermined criteria and the appropriate quality assurance procedures. EPA and Battelle make no expressed or
implied warranties as to the performance of the technology and do not certify that a technology will always
operate as verified. The end user is solely responsible for complying with any and all applicable federal, state,
and local requirements. Mention of commercial product names does not imply endorsement.
------- |