United States Atmospheric Research and Exposure
Environmental Protection Assessment Laboratory
Agency Research Triangle Park, NC 27711
June 1993
Research and Development
EPA600/R-93/130
&EPA
Workshop Report:
Identification of Performance
Parameters for Portable X-Ray
Fluorescence Measurement of
Lead in Paint
-------
June 1993
WORKSHOP REPORT: IDENTIFICATION OF
PERFORMANCE PARAMETERS FOR PORTABLE X-RAY
FLUORESCENCE MEASUREMENT OF LEAD IN PAINT
Research Triangle Park, NC
January 11-12,1993
Prepared by
E. D. Estes
W. F. Gutknecht
Center for Environmental Measurements and Quality Assurance
Research Triangle Institute
Research Triangle Park, North Carolina 27709-2194
EPA Contract No. 68D10009
RTI Project No. 5960-167
Mr. Warren Loseke, Work Assignment Manager
Atmospheric Research and Exposure Assessment Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Research Triangle Park, NC 27711
-------
DISCLAIMER
The Workshop described in these proceedings was sponsored by the United States
Environmental Protection Agency. The opinions, conclusions and recommendations are
solely those of the various participants and therefore, the contents of this document do
not necessarily reflect the views of the Agency. No official endorsement should be
inferred.
-------
ACKNOWLEDGEMENTS
This document was prepared under the direction of Mr. Warren Loseke,
Atmospheric Research and Exposure Assessment Laboratory (AREAL), U.S.
Environmental Protection Agency, Research Triangle Park, N.C.
Special acknowledgement is given to Ms. Sharon Harper and Mr. Michael E. Beard
(AREAL/EPA), for their careful review.
ii
-------
EXECUTIVE SUMMARY
The potential impact on health from environmental lead has resulted in increased
interest in lead exposure by Federal, State and local government agencies. As a result,
programs committed to sampling and analysis of lead are increasing nationwide. Public
housing authorities are required, by 1994, to randomly inspect ail their housing projects
for lead-based paint [Lead-Based Paint Poisoning Prevention Act, 42, U.S.C. 4:22
(d)(2)(A), 1971]. Currently, the most common approach to screening housing for the
presence of lead in paint is the use of the portable X-ray fluorescence (XRF) detector,
which gives rapid results and is non-destructive. More than six different portable XRF
instruments are commercially available. The EPA has sponsored studies of the
performance of each of several of these instruments.
To review the status of portable XRF technology, the U.S. Environmental
Protection Agency sponsored a 1-1/2 day workshop limited, to Federal and State
government personnel and government contractors. The goal of the workshop was to
identify problems and limitations that could result in measurement error and to develop
a set of performance parameters to generate or verify figures of merit that are
comparable across the technology.
The workshop was attended by 32 persons with experience in portable XRF
measurements. The morning session of the workshop began with a brief overview of the
purpose and goals of the workshop and was followed by presentations on the following
topics:
fundamentals
specific studies
future needs
a proposed test design
manufacturer's input
In the afternoon, the attendees discussed spectral and physical problems and
limitations of portable XRF and began to identify specific parameters or areas that need
Hi
-------
to be evaluated. On the second day, the attendees continued to identify parameters that
are experimentally verifiable and to discuss general approaches to the experimental
measurement of these parameters that would allow direct comparisons of portable XRF
instruments. Details of the experiments were not determined and will require consultation
with a statistician. The parameters judged to be the most important by consensus of the
workshop participants are the manufacturer's calibration, precision, accuracy, detection
limit, substrate effects and other interferences, and ruggedness. Other parameters that
should be considered in evaluating a portable XRF but are difficult to measure
experimentally include training, cost, safety features, physical configuration, portability,
data retention/storage capabilities, and manufacturer's support.
IV
-------
Table of Contents
DISCLAIMER i
ACKNOWLEDGEMENTS ii
EXECUTIVE SUMMARY iii
TABLE OF CONTENTS v
SECTION 1 INTRODUCTION 1
1.1 Goals of the Workshop 1
1.2 Structure of the Workshop 2
SECTION 2 SUMMARY OF PRESENTATIONS : 6
SECTION 3 GENERAL DISCUSSION 19
SECTION 4 PERFORMANCE PARAMETERS 23
4.1 Introduction 23
4.2 Manufacturer's Calibration 23
4.3 Precision 23
4.4 Accuracy 24
4.5 Detection Limit 25
4.6 Substrate Effects and Other Interferences 25
4.7 Ruggedness 26
4.8 Other Parameters 26
APPENDIX A Letter of Invitation and Note of Thanks
APPENDIX B Questionnaire Sent to XRF Manufacturers
APPENDIX C Lists of Attendees and Manufacturers
APPENDIX D Papers/Information Distributed by Speakers
and Manufacturers
-------
SECTION 1
INTRODUCTION
1.1 Goals of the Workshop
Numerous programs involving research on the toxicity and bioavailability of lead,
the environmental monitoring of lead, and the abatement of lead in and clearance testing
of housing are currently active. Examples of past projects include the Housing and Urban
Development (HUD) National Survey, the Three City Study and method evaluation studies
sponsored by the U.S. Environmental Protection Agency (EPA), and abatement programs
carried out by Maryland, Massachusetts, and many other active groups.
Public housing authorities are required, by 1994, to randomly inspect all their
housing projects for lead-based paint [Lead-Based Paint Poisoning Prevention Act, 42,
U.S.C. 4:22 (d)(2)(A), 1971]. That statute specifies the use of an on-site X-ray
fluorescence (XRF) analyzer for measurement of lead and denominates a reading of 1.0
mg/cm2 as a positive finding of lead-based paint. The HUD Interim Guidelines
(September 1990) use the XRF analyzer and the 1.0 mg/cm2 standard, but also specify
that atomic absorption analysis (AAS) or another comparable testing technique will be
used as a backup or confirmatory test when the XRF reading is inconclusive.
Alternatively, AAS may be used without first using on-site XRF. The HUD Guidelines set
the level of hazard for lead as 1.0 mg/cm2 (area basis) or 5000 |ig/g (weight basis).
Currently, the most common approach to screening housing for the presence of
lead in paint is the use of the portable X-ray fluorescence (XRF) detector, which gives
rapid results and is non-destructive. More than six different portable XRF instruments are
commercially available. The EPA has sponsored studies of the performance of each of
several of these instruments.
The U.S. Environmental Protection Agency sponsored a 1 -112 day workshop limited
to Federal and State government personnel and government contractors to review the
status of portable XRF technology. The goal of the workshop was to identify problems
and limitations that could result in measurement error and to develop a set of
performance parameters to generate or verify figures of merit that are comparable across
1
-------
the technology. Parameters to be included are detection limits, precision, bias,
interferences, productivity, safety, and use or generation of hazardous materials.
1.2 Structure of the Workshop
Following planning efforts by EPA and EPA-contractor staff, letters of invitation
were mailed to representatives from Federal and State governmental agencies and a
number of EPA contractors with experience in the use of portable XRF methods for
measurement of lead in paint. The intent was to invite recognized experts in the field,
while keeping the size of the workshop manageable and productive. Manufacturers of
portable XRF instruments were not invited to the workshop but were asked to complete
a questionnaire and to supply any additional information that they judged to be of value
to the workshop.
Thirty-two representatives accepted the invitation to attend the workshop (see
Appendix C for list). Upon arrival, attendees received a notebook containing the following
information:
Workshop Program
Agenda
List of Attendees
XRF Reports
"First Preliminary Draft Recommendation on Portable and
Transportable X-Ray Fluorescence Spectrometers for Field
Measurements of Hazardous Elemental Pollutants (International
Organization of Legal Metrology, OIML Reporting Secretariat
PS17/RS5: USA, October, 1992)
"Evaluation of Portable X-Ray Fluorescence Spectrometer for
Measurement of Lead in Paint, Soil and Dust," Hardison, D. L, C. O.
Whitaker, J. D. Neefus, E. D. Estes and W. F. Gutknecht,
Proceedings of 1992 EPA and AWMA International Symposium on
Measurement of Toxic and Other Related Air Pollutants. Durham,
NC, May 1992.
-------
The workshop began with a morning session of presentations on the current status
of portable XRF technology including fundamentals, specific studies, future needs, a
proposed test design, and manufacturers' input (see workshop agenda presented on page
4). In the afternoon, the attendees discussed spectral and physical problems and
limitations of portable XRF and began to identify specific parameters or areas that need
to be evaluated. On the second day, the attendees continued to identify parameters that
are experimentally verifiable and to suggest experiments that would allow evaluation and
direct comparisons of portable XRF instruments.
-------
GENDA
FOR
PORTABLE XRF SPECTROMETER WORKSHOP
JANUARY 11 - 12, 1993
EPA Environmental Research Center, RTP, NC
Time
Topic
Speaker
Day 1 - January 11
9:00 a.m.
9:15 a.m.
9:30 a.m.
9:45 a.m.
10:00 a.m.
10:15 a.m.
10:30 a.m.
10:50 a.m.
11:05 a.m.
11:20 a.m.
11:35 a.m.
11:50 a.m.
1:00 p.m.
1:15 p.m.
Welcome and Overview
Presentation by Attendees
Break
Presentation by Attendees
Lunch
Manufacturers' Input
Discussion of spectral and
physical problems/limitations
M. Beard/S. Harper
B. Gutknecht
B. Clickner
S. Weitz/M. McKnigJit
J. Pesce
J. Zilka
R. Petre
C. Papanicolopoulos
T. Spittler
Open
B. Gutknecht
Attendees
-------
Time
Topic
Speaker
Day 1 - January 11
2:40 p.m.
3:00 p.m.
4:30 p.m.
Day 2 - January 12
9:00 a.m.
10:30 a.m.
10:50 a.m.
Break
Development of prototype
performance tests (spectral,
physical and statistical aspects)
Adjourn
Discussion/revision of
performance tests
Break
Continuation of discussion/
revision of performance tests
Attendees
Attendees
Attendees
12:00 p.m.
Adjourn
-------
Section 2
SUMMARY OF PRESENTATIONS
Presentations were made to provide a background for the development of a set
of performance parameters that would be used to generate or verify figures of merit that
are comparable across portable XRF technology. The presentations covered the
fundamentals of the XRF method, specific studies and comparisons, future needs, and
manufacturers' input. Summaries are given below.
William Gutknecht/RTI - "A Comparison of the Scitec, Warrington and Princeton
Gamma-Tech Portable XRFs"
The preliminary (draft) results of a limited evaluation of three portable XRF
instruments (Scitec MAP-3, Warrington Microlead 1, and Princeton Gamma-Tech XK-3)
were presented. The evaluation was carried out using standard films prepared from oil-
based paint spiked with white lead and verified by acid digestion/atomic absorption
analysis. The lead concentration of the standard films ranged from 0 to 6 mg/cm2. The
films were placed on eight different substrates and analyzed by portable XRF, using the
manufacturer's calibration and following the manufacturer's instructions. The substrates
tested were 3/4" plasterboard, 3/4" plywood, 5" concrete, 4" cinderblock, 1/8" aluminum,
1/8" steel, brick and solid cinderblock. According to the manufacturer, no substrate
correction is necessary for the Scitec. Warrington and Princeton Gamma-Tech (PGT)
data were substrate corrected.
The Scitec was evaluated using two different "universal" calibrations installed by
the manufacturer. The first calibration covered the range 0.6 to 2.64 mg/cm2. The
standard paint films yielded correlation coefficients of 0.98 to 0.99 and standard errors
of estimate of x from y of -0.5 mg/cm2 for the five substrates tested (plasterboard,
plywood, 4" cinderblock, aluminum and brick). Biases ranged from + 0.13 mg/cm2 for 3/4"
plasterboard to -0.29 mg/cm2 for 3/4" plywood. The second calibration covered the range
from 0 to 6 mg/cm2. The standard films yielded correlation coefficients ranging from 0.93
-------
for solid cinderblock to 0.995 for steel. The standard error of the estimate ranged from
0.30 mg/cm2 for steel to 1.1 mg/cm2 for solid cinderblock. The bias ranged from 0.1
mg/cm2 for plasterboard to 4.4 mg/cm2 for solid cinderblock. The bias was found to be
concentration dependent for plasterboard, aluminum, and concrete (i.e., there is a change
in the value of the bias with a change in the value of the concentration) and constant for
plywood, steel, brick and 4" and solid cinderblock.
The Warrington was evaluated with all eight substrates. Correlation coefficients
were consistent, ranging from 0.990 for concrete to 0.998 for plasterboard, plywood and
steel. The standard error of estimate ranged from 0.2 mg/cm2 for plasterboard, plywood
and steel to 0.4 mg/cm2 for concrete. The bias ranged from -0.04 mg/cm2 (at 1.6 mg/cm2)
for plasterboard to 1.2 mg/cm2 (at 0.3 - 0.8 mg/cm2) for concrete and was found to be
concentration dependent for all substrates.
The PGT also was evaluated with all eight substrates. The correlation coefficients
were consistent, ranging from 0.977 for brick to 0.988 for 4" cinderblock. The standard
error of estimate ranged from 0.39 mg/cm2 for 4" cinderblock to 0.63 mg/cm2 for brick.
The bias ranged from 0.1 mg/cm2 (at 0.3 mg/cm2) for plasterboard and cinderblock to
-1.0 to -1.1 mg/cm2 (at 1.6 mg/cm2) for concrete, aluminum and steel. Bias was found
to be concentration dependent for all substrates, with the greatest dependency found for
concrete and 4" cinderblock.
To determine whether Scitec results could be improved by applying a substrate
correction, the Scitec was used to measure the substrate only for wood, sheetrock, 4"
cinderblock, brick, aluminum siding and steel. All substrates gave readings very close to
0 mg/cm2 (0.00 to 0.03 mg/cm2). When real-world painted boards were measured before
and after scraping, however, a different result was obtained. A board that measured 13
to 15 mg Pb/cm2 before scraping appeared to contain 1 to 3 mg Pb/cm2 after scraping;
a board that measured 6 to 7 mg Pb/cm2 before scraping yielded results of 0.1 to 3 mg
Pb/cm2 after scraping; and a board that measured 1.7 to 2.2 mg Pb/cm2 before scraping
measured 0.0 to 0.1 mg Pb/cm2 after scraping. One explanation for these observations
is that lead migrates or is driven into the substrate by the removal process. Variables
that must be considered in evaluating a portable XRF are the calibration, substrate
-------
correction, lead in the substrate, and paint variability.
Robert Clicknef/WESTAT - "On the Performance of the Scitec MAP/XRF in the
National Survey of Lead-Based Paint in Housing"
Dr. Clickner discussed the performance of the Scitec MAP-3 XRF used in the
National Survey of Lead-Based Paint in Housing from January to March, 1990. Scitec
delivered eight nominally identical MAPs, each containing a fresh 40 mCi Co57 radioactive
source and the "universal automatic substrate correction" software. Each MAP was
calibrated by the manufacturer over the range 0.0 to 2.64 mg/cm2. During the survey,
381 housing units across the United States were tested and thousands of 60-second XRF
measurements were made on dozens of different substrates.
Shims (standard paint films) from NIST with lead paint levels of 0.6 ± 0.02 and
2.99 ± 0.3 mg/cm2 and four substrates (wood, drywall, steel and concrete) were used to
continually check the MAPs' performance. Daily validation measurements were made for
each MAP/shim/substrate combination. Regression analyses on the validation data were
performed to estimate the precision and accuracy of the readings and to relate readings
to substrate and lead loading level. A different equation was obtained for each
machine/substrate.
For a wood substrate, readings generally were lower than the actual lead levels
(NIST certified levels) when the levels were <2.0 mg/cm2 and higher than the actual levels
when the levels were >2.0 mg/cm2. Eighty-eight percent of the validation readings of the
0.6 mg/cm2 shim on wood were <0.6 mg/cm2 and 52 percent were equal to 0.0 mg/cm2.
(The MAP will not give a negative reading.) For the 2.99 mg/cm2 shim, 60 percent of the
readings were >3.0 mg/cm2. The standard deviation of repeated readings was 0.25
mg/cm2.
The performance of the MAP on drywall or plaster was similar to the performance
on wood, with readings lower than the actual levels when the levels were <2.5 mg/cm2
and higher than the actuallevels when the levels were >2.5 mg/cm2. Eighty-two percent
of the validation readings of the 0.6 mg/cm2 shim on drywall or plaster were <0.6 mg/cm2
8
-------
and 52 percent were equal to 0.0 mg/cm2. At 2.99 mg/cm2, 44 percent of the readings
were <3.0 mg/cm2 and 45 percent were >3.0 mg/cm2. The standard deviation of repeated
readings was 0.25 mg/cm2.
For a steel substrate, the readings were higher than the actual levels for all
observed levels of lead, and there were substantial differences among the eight
instruments. Eighty-eight percent of the validation readings of the 0.6 mg/cm2 shim on
steel were >0.6 mg/cm2. At 2.99 mg/cm2, 60 percent of the readings were >3.0 mg/cm2
although one machine had no readings above 3.0 mg/cm2. The standard deviation of
repeated readings was 0.21 mg/cm2.
For concrete, brick, and other related substrates, the MAP had difficulty detecting
low to moderate levels of lead, and there was much variation both among and within
instruments. Ninety-five percent of the validation readings of the 0.6 mg/cm2 shim were
equal to 0.0 mg/cm2, and only one machine had more than one non-zero reading. At
2.99 mg/cm2, 95 percent of the readings were <3.0 mg/cm2 and 11 percent of the
readings were equal to 0.0 mg/cm2. The standard deviation of repeated readings was
0.49 mg/cm2.
In summary, the MAP readings were found to be systematically different from the
amounts of lead in the paint film standards with the direction and magnitude of the
differences being related to the substrate material and the lead levels in the paint. The
precision of the readings also was found to be dependent upon the substrate. On the
basis of this survey, WEST AT concluded that a substrate correction is necessary to
accurately determine the presence and amount of lead-based paint on surfaces when
using XRFs similar to the Scitec MAP-3 used in the national survey of lead-based paint
in housing.
Steve Weitz/HUD - "HUD Needs"
Mr. Weitz briefly presented the needs of the Department of Housing and Urban
Development (HUD) with respect to portable XRF technology. The HUD Guidelines are
to be revised totally by October 1993. HUD expends considerable funding on testing and
is dedicated to producing data of the highest quality. Because lead-in-paint data are the
-------
basis for important decisions, improved instruments are needed to obtain better data. To
develop improved instruments, a method for evaluating the current instruments is needed.
Mary McKnight/NIST - "Evaluation of Methods for Field Measurement of Lead in
Paint Films"
Dr. McKnight presented a plan for evaluating field methods for measurement of
lead in paint. The first step is to characterize the problem by establishing performance
criteria, observing field practices and identifying key factors that could affect field results.
A limited simulated field study and a limited field study will then be designed. The
simulated field study will be carried out in a test chamber using real walls and painted
boards that closely represent field samples. The field study will be carried out in actual
housing units. If the results of the simulated field study are consistent with the results of
the actual field study, an expanded simulated field study will be conducted over a range
of field possibilities (lead concentrations, environmental conditions, substrate types, etc.)
and a knowledge-based system will be developed for assessment of equipment or field
measurement procedures. The data obtained in the simulated field study will be used to
develop a standard method or protocol for assessing equipment or other measurement
procedures for lead in paint. Other benefits of a study such as this include gaining
precision and bias data for the measurements and improved understanding of factors
affecting measurement quality.
John Pesce/Star Environmental - "An Examination of Substrate Effect on Portable
X-Ray Fluorescence Instrumentation"
Mr. Pesce presented the results of an investigation of substrate effect on portable
X-ray fluorescence (XRF) instrumentation. The data presented were derived from both
actual field use and round-robin information.
Princeton Gamma-Tech (PGT) XK-3 XRF analyzers were used to test 421 units
in an apartment complex for lead in paint. More than 20,000 readings were taken. The
testing protocol included measurement of substrate effect levels (SELs) on all surfaces.
10
-------
The SEL was subtracted from the Apparent Lead Concentration (ALC) to obtain a
Corrected Lead Concentration (CLC). Sodium sulfide solution was used for confirmation
when the ALCs were 1.2 mg/cm2, and samples were taken for atomic absorption analysis
when CLCs were 0.8 - 1.5 mg/cm2. The SEL locations also were tested with the
Warrington and Scitec XRFs, using a "shoe" or template to fix the orientation of each
instrument.
A variety of unpainted substrates was tested, including masonry, cinder block,
plaster, brick, sheetrock, concrete, wood and metal. The PGT and Warrington
instruments showed significant variability forthese building materials. PGT values ranged
from 0.0 mg/cm2 for pressed board to 1.9 mg/cm2 for a metal window sill. Warrington
values ranged from -1.1 mg/cm2 for a metal window sill and mailbox to 1.7 mg/cm2 for a
plaster column and a concrete column. The Scitec measured 0.0 mg/cm2 for most
substrates. The Scitec response was 0.3 - 0.4 mg/cm2 for metal, 0.3 mg/cm2 for a
concrete column and oak, and 0.2 mg/cm2 for sheetrock. It should be noted that the PGT
and Warrington instruments can give a negative reading, but the Scitec cannot.
For metals, the PGT and Warrington respond to substrate effects differently, with
the elevated (+) readings of the PGT similar in absolute value to the depressed (-)
readings of the Warrington. All XRFs appear to work well on less dense materials such
as wood.
Substrate effects also were examined as a part of the round-robin study
established by a group of private lead inspectors in Massachusetts. Experiments were
performed using steel channel, red brick, poplar, belly casing, and sheetrock substrates.
The SEL was measured for a piece of each substrate containing no lead paint. The
apparent lead concentration (ALC) was measured for the same substrate painted with
lead paint. The SEL was subtracted to give a corrected lead concentration (CLC) for
each substrate. For steel channel, the CLCs of the PGT and Scitec showed some lead;
the CLC of the Warrington showed no lead. For red brick, the CLCs of the PGT and
Warrington showed some lead; the Scitec detected no lead on the painted or unpainted
brick, indicating that a false negative reading can be observed. All three XRFs identified
lead on poplar wood surfaces although the quantitative results were not in good
11
-------
agreement, ranging from 0.5 mg/cm2 for the Scitec to 1.2 mg/cm2 for the PGT. A convex
wood molding called belly casing was used to test the effects of an air gap between the
XRF probe and the sample. Results were consistent with the results for the poplar wood
sample. Sheet rock samples produced consistent results with the three XRFs.
To test for read-through effects, measurements were made on lead flashing under
1-1/2" of wood. The PGT appeared less susceptible to detect this type of lead, giving a
reading of 1.7 mg/cm2. The Warrington and Scitec measured 10.8 and 4.6 mg/cm2,
respectively.
Based upon this study, it was concluded that portable XRFs can be used with a
reasonable amount of confidence for a wide variety of substrates, assuming that the
operator has the experience to know what the substrate is and to make the appropriate
corrections. The operator must know when to doubt the XRF and use secondary and
tertiary screening methods. Neither automatic or manual correction can compensate for
all possible substrates, and the correction values vary from instrument to instrument from
the same manufacturer as well as from different manufacturers. Confirmatory testing
should be performed prior to making abatement decisions.
John Zilka/Applied Systems - "XRF Problems and Applications"
Mr. Zilka presented some problems and needs with respect to use of field portable
XRFs. Applied Systems uses instruments manufactured by PGT, Scitec and Warrington
and has observed variances and reliability problems in the field, particularly with the
Scitec. The manufacturers have provided little support in the way of standard operating
procedures (SOPs) and validation.
Two major areas of concern that Mr. Zilka would target for study are the reliability
of the instrument and the reliability of the inspector. Currently, Applied Systems runs
validation checks at the beginning of the day, hourly, and at the end of the day, but
detailed validation procedures are needed to determine that the instrument is operating
properly. Training field personnel also is a big problem. To do the job correctly, a field
tester needs the proper tools including a reliable instrument, site specific standard
12
-------
operating procedures (SOPs) and a basic knowledge of construction materials. Currently,
there are considerable differences in what is required for an inspection and in what is
being done to meet these requirements. For example, it is unclear how many inspectors
are actually doing substrate corrections.
Finally, there is a need for auditing an inspector's work to give the client some
degree of confidence in the results.
Roy Petre/State of Massachusetts Department of Public Health - "General
Concerns"
Mr. Petre discussed some policy issues with respect to current standards and
future needs. In the past a health-based standard for lead expressed in mass/area was
used because children eating paint chips was the focus of concern. Currently, dust is
considered to be the major source of exposure. The Childhood Lead Prevention standard
of 0.5 percent by weight for paint is high relative to dust. Mr. Petre suggests that the
current standards be examined carefully with respect to exposure models, animal studies,
and total dust exposure from future disturbances.
Mr. Petre also expressed a need for a cost-effective method for testing on-site.
Laboratory testing is expensive and requires considerable delay and detailed chain-of-
custody procedures. In some cases, it may be cheaper to replace the architectural
elements than to test them. Taking representative field samples for laboratory analysis
is also a problem, and scraping creates a potential hazard by putting the lead in its most
ingestible form. One on-site method is portable XRF. For a portable XRF to be used
with confidence, however, it should have an accuracy and precision of 0.1 to 0.2 mg/cm2.
A "wand" for testing ceilings would be a useful design feature. A self-correction for
substrate effects is also a consideration since scraping creates dust. It would be
desirable to design a machine that can read by layer to eliminate the read-through effect.
For example, an XRF will sometimes detect leaded paint on a plaster surface that has
been covered with wood panelling. On the other hand, thin paint films containing 2 to 4
percent lead may be cleared when an XRF is used for testing because of the thin layer
13
-------
effect. Since the portable XRF reads in concentration units of mass/area, painted
surfaces containing high percentages of lead by weight may show lead concentrations
below the regulatory limit if the paint is spread thinly enough. Theoretically, a correction
can be made for this effect. The State of Massachusetts currently uses sodium sulfide
for confirmatory testing, with good results, when concentrations below the abatement
threshold are obtained with the portable XRF.
Mr. Petre stressed that operator effect must be addressed. Massachusetts has
been training and licensing lead paint inspectors since 1990. An inspector must complete
an apprenticeship under a master inspector. In addition, Massachusetts is considering
a program to certify individual XRFs.
Chris Papanicolopoulos/Georgia Tech Research Institute - "XRF Instrumentation"
Dr. Papanicolopoulos presented the results of an investigation of portable XRF
instruments. From 1987 through 1989, three Princeton Gamma Tech XK-3s were tested,
and in 1990 some comparisons of the XK-3s, the Scitec MAP, and the Warrington
Microlead instruments were made. Many combinations of sample substrates and paints
were analyzed in the field. A goal of the study was to identify problems and to determine
whether they are instrument specific or generic.
The Microlead, which has a scintillation counter detector, uses a grease between
the scintillation tube and photomultiplier tube that breaks down after repeated use and
also in the summer (i.e., at high ambient temperature), causing inconsistent results. Also,
a design problem was discovered in the electronics of the Microlead. When it was
corrected, measurements were more reliable.
The PGT XK-3's detector is a proportional counter that is temperature sensitive.
When the sample substrate is a dense material, backscatter can saturate the counter.
Dr. Papanicolopoulos stated that proportional counters have "reached their limit" and
cannot be improved.
The Scitec detector is a semiconductor crystal that must be "quiet" to operate
properly. The detector can get very hot when subjected to large numbers of X-rays in a
14
-------
short period of time and should be cooled in some way such as using the Peltier method.
Large numbers of X-rays can arise from a large amount of lead present in the sample
and/or a very dense substrate causing backscatter of a large percentage of the incident
or excitation X-rays.
One of the major problems with the portable XRF is substrate interference. The
PGT XK-3 does not adjust the peak height for backscatter. If the backscatter goes up
or down, it will be reflected in the measured lead concentration. The Scitec has a
proprietary formula for correcting for backscatter, but it is not appropriate for the soil
method. A better method for handling the backscatter problem is needed.
Dr. Papanicolopoulos suggested that instruments be certified two times a year with
the resulting data forwarded to the EPA. He also suggested that there be an independent
quality assurance person to see that the instrument is being used to obtain valid results.
Thomas Spittler/EPA Region I - "Laboratory XRF"
Dr. Spittler presented the results of analysis of paint chips by laboratory XRF.
Peeling paint chips are usually multilayered, with the inner layers containing more lead.
The outer layers frequently are newer paints containing titanium, zinc, and iron that will
absorb lead X-rays from the inner layers. Spectra from a laboratory XRF were presented
to demonstrate that significantly different lead concentrations could result when front and
back sides of paint chips were measured independently.
Dr. Spittler stated that paint on walls is not a threat to children unless it is peeling
or on chewable surfaces. Attempting to remove and analyze paint that is intact could
generate another "asbestos situation." However, he is not convinced that portable XRFs
will give reliable results. Data from an alternative method using a laboratory XRF were
presented. Three standards were prepared using different weight ratios of National
Institute of Standards and Technology (NIST) Standard Reference Material (SRM) 1579
and somar mix (an inert material) to cover a range of from 2.00 to 9.38 percent lead. An
actual paint sample then was ground and blended with somar mix in a dental
amalgamator. The standards and samples were analyzed by laboratory XRF with peak
15
-------
height used for quantitation. A measured value of 23.1 percent lead was obtained
compared to the "true" value of 24 percent obtained by a more rigorous XRF method.
The somar dilution procedure (three standards and one sample) was performed in
approximately five minutes. Dr. Spittler stated that the cost of a good laboratory XRF is
-40,000 (down from -$100,000) and that it can be operated by a trained high school
student.
Dr. Spittler also emphasized that to determine what is poisoning a child, house
dust and soil samples should be taken in addition to paint samples that might be posing
a threat.
William Gut knee ht/RTI - "Portable XRF Manufacturers' Input"
Prior to the workshop, questionnaires were mailed to six manufacturers of portable
XRF instruments that were identified by RTI through contacts with a variety of people
working on lead studies (see Appendix C). The questionnaire was designed to obtain
information on detection limit, sensitivity, accuracy, calibration procedure, operating
procedure, analysis time, substrate correction procedures, safety considerations,
problems/limitations, and any envisioned improvements. Responses were received from
the manufacturers of the Scitec MAP-3, the Spectrace 9000 and the Princeton Gamma
Tech XK-3. The responses and other information can be found in Appendix D.
The manufacturers approached the various performance parameters (detection
limit, accuracy, precision) in different ways. The PGT XK-3's detection limit of 0.5 mg
Pb/cm2 is taken as the single-reading standard deviation of a series of readings. The
Spectrace 9000's detection limit of 0.01 to 0.1 mg Pb/cm2 (depending upon substrate) for
a 200 sec assay is determined as three times the standard deviation of repeated blank
analyses. The Scitec manufacturer pointed out that for XRF analysis of lead-based paint,
the detection limit is a function of measurement condition, measurement time, confidence
level, and source strength. For a 60 sec measurement, K-shell detection limits of 0.10 and
0.43 mg Pb/cm2 were found for paint on wood and concrete, respectively, based on data
from 46 different instruments with 25 repeated measurements each.
16
-------
Sensitivity is generally defined as the change in response that results from a
specified change in concentration. Only the Spectrace manufacturer reported sensitivity
in this form as ~50(PbJ, ~4(PbK) cts/sec/mg Pb/cm2/mCi. Also, only the Spectrace
manufacturer reported an accuracy; the Spectrace literature states the accuracy of their
instrument to be +. 0.12 mg Pb for < 1 mg Pb/cm2 on wood using a 60 sec assay.
Measurement times for a test are 60 seconds for the Scitec MAP-3 and Spectrace 9000,
and 12 seconds for the PGT.
The Scitec and PGT instruments are pre-calibrated by the manufacturer and there
are no plans to allow for direct user calibration or recalibration. The PGT XK-3 is
calibrated using 10 readings each of the 0 and 1.53 mg/cm2 HUD standards. The Scitec
is calibrated using NIST 2759 standard paint films on different substrates. The Spectrace
9000 is pre-calibrated by the manufacturer using several pure element standards, various
substrates, and at least one NIST 2759 reference material. However, the calibration may
be modified by user accessible routines via either the keyboard or the RS232 port. The
port may be protected to ensure only approved access. Keyboard changes are indicated
on the printed (or stored) record and default values can be restored.
The portable XRFs also use different approaches for substrate correction. The
PGT XK-3 manufacturer recommends manual substrates correction in accordance with
the HUD Guidelines. The Scitec MAP comes with a general or "universal" compensation
for K and L shell X-rays that is built into the system's software. Scitec also provides a
"dedicated" substrate correction for high density substrates such as steel, brick, concrete
and thin metals. If lead is suspected to be present in the substrate, Scitec recommends
use of NIST SRMs over the painted surface to characterize the substrate. For the
Spectrace 9000, problems with overlapping X-ray peaks or signals are identified and
characterized when the instrument is calibrated and correction factors are applied
automatically to the spectral data to obtain a substrate corrected readout.
All three manufacturers reported some type of instrument check for satisfactory
operation. A calibration check sample is provided with the PGT XK-3. The Scitec MAP
system conducts a "self test" during a 15 sec warm-up immediately after the instrument
is turned on to determine if all system components are operating within specification.
17
-------
Scitec software used to download measurement values verifies both measurement
integrity and instrument performance and evaluates the calibration check readings. The
Spectrace 9000 sounds an alarm if 1) peaks are not in the correct position; 2) the battery
is low; 3) there is unusual electronic noise; or 4) there is loss of signal. All alarms are
accompanied by a descriptive message on the screen.
Neither PGT nor Scitec released information on planned improvements. Spectrace
envisions 1) increasing the detector area; 2) using a battery with a longer life; 3) reducing
the size of the probe for easier access to narrow spaces; and 4) more sophisticated
substrate corrections.
18
-------
SECTION 3
GENERAL DISCUSSION
Following the presentations, there was a general discussion of current needs for
portable XRF and the spectral and physical problems/limitations that need to be
addressed.
Public housing authorities are required, by 1994, to randomly inspect all their
housing projects for lead-based paint. There is, therefore, an urgent need for a reliable
method for taking representative samples and for producing defensible data with no false
negatives. The workshop attendees generally agreed that an on-site, real-time,
nondestructive lead measurement method such as the portable XRF is of value for a
number of reasons. Use of a reliable portable XRF would result in a considerable savings
of time and would not require "touching up" sampling sites after destructive sample
collection.
Jim Hayes of the State of N.C. Environmental Health Services is conducting an
epidemiological evaluation of lead poisoned children. Two of the goals of the program
are to teach parents the reason for concern and to identify potential sources of the
contamination. Mr. Hayes pointed out that it is of real value to have an on-site
measurement method that can give the parents a lead concentration at the time of the
visit without causing damage to painted surfaces. Susan Guyaux of the Maryland
Department of the Environment stated that day-care center personnel may react
negatively to taking paint scrapings, and that a hazard is actually created by taking
samples .in this way. She stressed the need for an on-site, non-destructive lead analysis
method with appropriate laboratory confirmation. Both the field and laboratory methods
need to be the best that they can be.
Dr. William Gutknecht of the Research Triangle Institute stated that portable XRFs
"are a reality" and that their sales are increasing. It is anticipated that they will be used
more in the future for soil and dust analysis. With this in mind, the workshop attendees
turned their attention to answering questions such as:
How "good" are the portable XRFs that are currently on the market?
19
-------
Where and/or under what conditions can portable XRFs be used reliably?
When is laboratory confirmation necessary?
What are the problems and limitations of portable XRF technology?
Is the technology improvable?
What are the critical parameters needed to evaluate and compare portable
XRFs?
How can these parameters be tested?
It is sufficient to certify manufacturers or is certification of individual
instruments needed?
Several groups represented at the workshop had conducted studies to compare
lead measurement results obtained by portable XRF and by atomic absorption
spectrometry (AAS). Bill Gutknecht reported that RTI found that the XRF and AA
generally agree although occasional significant differences were found. David Cox of
David Cox and Associates reported that in a data base of approximately 4000 AAS vs.
XRF measurements examined by his company, only 5 to 10 percent were serious outliers
and there were ~4 percent false positive results using the portable XRF. The problem
is that there is no way to identify the outliers when only one measurement method is used
and there is no way to predict when they will occur. John Zilka of Applied Systems
emphasized the need for defensible data. From his experience, hourly checks indicate
that the portable XRF generally gives consistent readings, and he is usually able to tell
if the equipment is not operating properly. However, Chris Papanicolopoulos of Georgia
Tech Research Institute cautioned that a portable XRF can be consistently inaccurate.
Based upon their experience and expertise, workshop attendees expressed
opinions concerning where and under what conditions portable XRFs could be used.
Kevin Ashley of NIOSH stated that the portable XRF is a useful screening tool but that
the data obtained are not defensible in court. Merrill Brophy of the Maryland Department
of the Environment supported Dr. Ashley's comment and added that "for a vast majority
20
-------
of housing," portable XRF is a first step. However, more confidence in the XRF data is
needed. Bob Clickner of Westat reported that the portable XRF works "fairly well" for
measurement of paint on wood and drywall, but generally gives a high reading for paint
on steel. Sharon Harper of the Office of Research and Development (ORD) EPA agreed
that results were good for paint on wood and drywall. David Cox warned against focusing
too much attention on a single XRF measurement. He suggested that a better approach
is to take a set of measurements on an aggregate to answer questions such as "Do the
doors contain lead?" and he has found that portable XRF technology works well for this.
Much of the discussion centered upon the "gray area" where laboratory
confirmation of an XRF measurement is needed. John Scalera of Office of Pollution
Prevention and Toxics (OPPT), EPA posed two questions to the group: 1) "Is there a
definable concentration limit for XRF at which there is 95 percent confidence in the XRF
data?" and 2) "Is the XRF useful at this concentration?" For example, if there is 95
percent confidence in the XRF at 6 or 7 mg/cm2 and this lead concentration is commonly
found, the XRF will be a useful screening tool. Bob Clickner reported that few readings
above 6 mg/cm2 were found in the national Survey of Lead-Based Paint in Housing. If
the confirmation limit was set at 6 mg/cm2 but most levels were <6 mg/cm2, almost every
measurement would require laboratory confirmation. John Zilka, however, stated that he
finds levels far in excess of 5 or 6 mg/cm2, especially in building exteriors.
Another factor to be considered in determining the utility of portable XRFs is the
concentration level of lead that is considered dangerous. John Pesce of Star
Environmental pointed out that there is a difference in philosophy about how much lead
is dangerous and that more data are needed to determine the capabilities of the XRF.
John Zilka stated that we need to answer the question "Is there a safe level of lead, and
is it tied in with the accuracy, precision, etc. of the XRF?" Chris Papanicolopoulos
cautioned that the regulatory levels will probably go lower and lower. Since XRF
technology is already struggling, inspectors may be forced to use laboratory analyses for
all measurements.
Jim DeVoe of NIST stated that there is a problem with the design of the portable
XRF and that more sensitivity is needed. John Pesce reported that he sees differences
21
-------
in day-to-day field use of the XRF, that the XRF still has trouble with the simplest (non-
real-world) samples and that substrate effects should be more closely examined. Jim
DeVoe expressed a need for the development and availability of "reference walls" with
different substrate layering orders to simulate wall types that are encountered in the field.
Chris Papanicolopoulos stated that the instrument should have a feedback mechanism
to alert the user to improper operation or an automatic shutoff. Kenn White of T. C.
Analytics said that, rather than asking manufacturers their capabilities, users should tell
manufacturers their needs and ask if they can be met.
22
-------
SECTION 4
PERFORMANCE PARAMETERS
4.1 Introduction
After the individual presentations and the general group discussion, the workshop
attendees developed a list of performance parameters that could be used to make a
direct comparison of portable XRFs. General approaches to the experimental
measurement of these parameters were also discussed. However, details of the
experiments were not determined and will require consultation with a statistician. The
parameters judged to be the most important by a consensus of the workshop participants
are discussed in the following sections.
4.2 Manufacturer's Calibration
Proper calibration is essential for obtaining reliable results using any instrumental
method for analysis. Most portable XRFs currently on the market are calibrated by the
manufacturer and have no provision for user interaction or recalibration. Much of the
workshop discussion focused on procedures for checking the manufacturer's calibration.
The general agreement among workshop attendees is that calibration checks should be
performed at the beginning and end of each testing day and at hourly intervals during
use. Suggested standards are the NIST films at 0.0,1.0 and 3.5 mg Pb/cm2 on both low
and high density materials representing site specific substrates. The periodic calibration
checks can be used to evaluate the stability of the calibration and to test for drift. The
accuracy of the calibration is tested as described in Section 4.4.
4.3 Precision
One performance parameter that can be experimentally measured is precision,
which is generally defined as the scatter of a group of measurements, made under the
same specific conditions, about their average value. Precision may be expressed as the
standard deviation (a) or the coefficient of variation (100a/x where x is the average
value). Before precision can be used to directly compare portable XRFs, an experiment
23
-------
must be carefully designed in terms of the number of tests to be performed and the
equation used for the calculation of precision. The workshop attendees agreed that a
"test" should be performed as described by the manufacturer. For the Scitec, a test is
defined as a single 60 second reading. For the PGT XK-3, a test is defined as the
average of three cycles collected manually. For the Warrington, a test is defined as the
average of three cycles collected automatically. For both the PGT XK-3 and the
Warrington, an initial single reading is taken at the beginning of each test but is not used.
Test were not defined for other brands of portable XRFs.
It was decided that NIST films at concentrations of 0.0, 1.0, 2.0, and 3.5 mg/cm2
should be used for the precision tests. Suggested substrates for precision testing are a
low density material such as pine or plastic foam, a high density material such as
concrete, and a layered substrate containing both high and low density materials such as
wood over concrete. The latter has been found to be a major problem for some portable
XRFs.
The attendees agreed that the precision tests should be performed both on new
(off the shelf) instruments and on instruments that have been used for field
measurements for several months. In terms of source depletion, a new instrument would
be defined as one whose source is less than 10 percent depleted and an old instrument
as one whose source depleted beyond 1-1/2 half lives of the excitation source. In all
cases, a "blind" process should be used for selection of instruments to be tested.
4.4 Accuracy
Another performance parameter that can be measured experimentally is the
accuracy or bias. Bias (B) is defined as the signed difference between the average (x)
of a set of measurements of a reference material and the "true" value (T) of the reference
material and is calculated in the units of measurement using the equation
B = X-T
or as a percentage of the true value using the equation
24
-------
%B = 100 (X-T)/T
Bias can be determined from the precision test data since NIST SRMs will be
used. The data will be examined to determine whether the bias is a function of lead
concentration.
Of concern is the need to describe clearly the chances of making an improper
decision when an X-ray measurement result for lead in paint is compared to a decision
level. This involves characterizing the random and systematic error components of the
method at or near the decision level. Determination of an acceptable accuracy figure for
each instrument will be possible after a definitive field study has been conducted.
4.5 Detection Limit
Because the regulatory limit for lead may be lowered as environmental lead
programs continue, it is important to determine the limits of detection and quantitation.
Again, this involves characterizing the random and systematic error components at or
near the limit of detection. Estimation of each instrument's detection limit over various
substrates is expected to be possible using field study data.
4.6 Substrate Effects and Other Interferences
The buildings to be tested under the HUD Guidelines contain a wide variety of
building materials that can be painted with lead-based paint. Typical materials that are
encountered include wood, plaster, gypsum, concrete, brick, metal, sheetrock, and cinder
block, and often these materials are in tandem. Wood over concrete and plaster over
metal mesh pose major problems for some portable XRFs. Some instruments will give
readings higher than the regulatory limit for certain substrates when no lead is present.
Other substrates may yield negative readings that will result in a lead concentration
reading that is lower than the true value.
The different manufacturers of portable XRFs treat substrate effects (background)
in different ways. To determine whether a portable XRF is suitable for a particular
purpose, the effects of various substrates should be evaluated. The workshop attendees
25
-------
suggested that experiments be performed using NIST films and a variety of substrates
that cover a range of densities and thickness. The effects of "read-through" should also
be tested by placing an NIST film between two substrates and measuring the apparent
lead concentration using the portable XRF.
4.7 Ruggedness
Because of the enormous number of lead measurements to be made and the wide
variety of environmental conditions that are likely to be encountered, the portable XRF
must be field-rugged. Workshop participants discussed a number of parameters that
should be tested to evaluate the ruggedness of portable XRFs.
Experiments should be performed to determine the temperature range over which
an XRF operates satisfactorily and to determine the necessary temperature equilibration
time. A range of -20 °F to 120 °F was suggested. The International Organization of
Legal Metrology (VOIML) recommends that the instrument be placed in an environmental
chamber and warmed or cooled until it is in thermal equilibrium at the desired
temperature. Tests should then be carried out to determine the relative standard
deviation of the output signal.
The XRF should also be able to withstand a reasonable mechanical shock to the
control box such as a bump against a door facing. A "bump test" was discussed, but the
specific procedure is yet to be determined. The OIML describes a mechanical shock test
that involves tilting the instrument about one bottom edge to a height of at least 50 mm
and allowing it to fall once to the surface.
A third parameter for ruggedness is susceptibility to interference from
electromagnetic fields. The OIML recommends exposing the instrument to
electromagnetic fields at a field strength of 10 V/m and determining the relative standard
deviation of the output signal during this exposure.
4.8 Other Parameters
Other parameters that should be considered in evaluating a portable XRF but are
26
-------
difficult to measure experimentally are as follows:
Ease of operation
Training (level required; usefulness of the manufacturer's manual)
Cost
Safety features/radiation leakage/operator exposure
Physical configuration (handle; shutter; probe size and weight; cord)
Data retention/storage capabilities
Manufacturer's support (licensing; availability of a user "hot line;"
training)
Portability/weight of detector/probe assembly
27
-------
APPENDIX A
Letter of Invitation and Note of Thanks
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
RCH AND EXPOSURE ASSE
ESEARCH TRIANGLE PAR
NORTH CAROLINA 277 11
~* ATMOSPHERIC RESEARCH AND EXPOSURE ASSESSMENT LABORATORY
l"*^ RESEARCH TRIANGLE PARK
November 24, 1992
Dear :
Numerous programs involving research on the toxitity and bioavailability of lead,
the environmental monitoring of lead, and the abatement and clearance of lead in
housing are currently active. Examples of past projects include the HUD National
Survey, the EPA Three City Study, method evaluation studies sponsored by EPA and
abatement programs carried out by Maryland, Massachusetts and many other groups.
Public housing authorities are required, by 1994, to randomly inspect all their
housing projects for lead-based paint. Currently, the most common approach to
screening housing for the presence of lead in paint is the use of the portable X-ray
fluorescence (XRF) detector, which gives rapid results and is non-destructive. Currently,
more than six different portable XRF instruments are commercially available. The U. S.
Environmental Protection Agency has sponsored studies of the performance of several
of these instruments.
The U.S. Environmental Protection Agency is sponsoring a 1-1/2 day workshop
limited to Federal and State government personnel and government contractors to
review the status of portable XRF technology, to identify problems and limitations that
could result in measurement error, and to develop a matrix of performance parameters
to generate or verify figures of merit that are comparable across the technology.
Parameters to be included in the matrix are detection limits, precision, bias, interferences,
productivity, safety, and use or generation of hazardous materials.
-------
As one of the recognized experts in this field, you are invited to attend this
workshop, which is scheduled for January 11-12, at the EPA Environmental Research
Center in Research Triangle Park, NC. Unfortunately, funds are not available to pay for
attendance at the workshop. However, this workshop offers an opportunity to make a
contribution to solving what has been called the nation's number one preventable
childhood health problem, and it also offers an opportunity to identify future research
needs.
If you are able to attend, please inform Dr. Eva Estes at (919) 541-5926 at your
earliest convenience. Thank you, and we hope to see you at the portable XRF
Workshop.
Sincerely,
Warren Loseke
-------
' UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
^tt/ ATMOSPHERIC RESEARCH AND EXPOSURE ASSESSMENT LABORATORY
RESEARCH TRIANGLE PARK
NORTH CAROLINA 2771 1
January 26, 1993
Dear :
Thank you for attending the Portable XRF Workshop held at the EPA
Environmental Research Center on January 11 -12, 1993 and for presenting the results
of your work in this area. As a result of the workshop we have begun to develop a matrix
of performance parameters to generate figures of merit that are comparable across the
portable XRF technology. Your hands-on experience, knowledge of methodologies, and
real-world data were instrumental in its success. We appreciate not only your
contributions, but also the preparation and effort required for attending.
In approximately one month you will receive a draft report of the workshop for your
comments. We look forward to continuing to work with you on this important issue.
Again, thank you for your contributions and for your interest in lead measurement
programs.
Warren Loseke
:acp
-------
APPENDIX B
Questionnaire Sent to XRF Manufacturers
-------
UNITED STATES ENVIRONMENTAL PROTECTION AGENCY
ATMOSPHERIC RESEARCH AND EXPOSURE ASSESSMENT LABORATORY
RESEARCH TRIANGLE PARK
NORTH CAROLINA 27711
December 8, 1992
Dear :
Numerous programs involving research on the toxicity and bioavailability of lead,
the environmental monitoring of lead, and the abatement and clearance of lead in
housing are currently active. Examples of past projects include the HUD National
Survey, the EPA Three City Study, method evaluation studies sponsored by EPA and
abatement programs carried out by Maryland, Massachusetts and many other groups.
Public housing authorities are required, by 1994, to randomly inspect all their
housing projects for lead-based paint. Currently, the most common approach to
screening housing for the presence of lead in paint is the use of the portable X-ray
fluorescence (XRF) detector, which gives rapid results and is non-destructive. Currently,
more than six different portable XRF instruments are commercially available. The U. S.
Environmental Protection Agency has sponsored studies of the performance of several
of these instruments.
The U.S. Environmental Protection Agency is sponsoring a 1-1/2 day workshop
\
limited to Federal and State government personnel and government contractors to
review the status of portable XRF technology, to identify problems and limitations that
could result in measurement error, and to develop a matrix of performance parameters
to generate or verify figures of merit that are comparable across the technology.
Parameters to be included in the matrix are detection limits, precision, bias, interferences,
productivity, safety, and use or generation of hazardous materials.
-------
Such discussions would be of limited productivity without input from the people
who actually develop and provide portable XRF spectrometers. As one of the
recognized suppliers of these instruments, your input is crucial to the success of this
workshop. Consequently, a questionnaire has been enclosed in an attempt not only to
address the issues described above but also to provide a means for introduction of
additional topics you may wish to see discussed. Your timely response is needed in
order to incorporate your input into the workshop agenda. I ask that responses be
returned by December 31.
A self-addressed envelope has been included for your convenience. If you have
any questions please call Eva Estes at (919) 541-5926. Thank you in advance for your
cooperation.
Sincerely,
Warren Loseke
-------
Portable XRF Spectrometer
Name or Model # of spectrometer
'Please reproduce this form if other models are available.
Applications
(Paint, Soil, Dust, etc.)
Detection Limit How is the detection limit
determined?
Sensitivity How is the sensitivity
determined?
Accuracy How is the accuracy
determined?
Briefly describe the calibration procedure
Are there any plans to allow for direct user calibration or recalibration?
-------
Are you aware of the new MIST paint film SRMs?
Have they been of any value to you in calibrating or testing your instruments?
Briefly describe the procedure for use of your instrument
Approximate analysis time
Interferences
What substrate correction procedures are used?
Are their any new developments or guidelines for using substrate correction procedures?
Describe.
Are there any built-in instrument checks to test for satisfactory operation? Describe. __
-------
Briefly describe safety procedures/considerations
Are there any problems/limitations with the instrument at present?
Are improvements to the current instruments envisioned?
When might an improved instrument be available?
What improvements will be implemented?
What do you view as the future needs of the public and/or lead testing organizations
regarding portable XRF spectrometers?
What topics should be added to the XRF workshop agenda?
-------
Please provide any standard operating procedures, Instrumentations or other literature
you believe would be informative. :
-------
APPENDIX C
Lists of Attendees and Manufacturers
-------
Portable XRF Workshop Attendees
January 11 - 12, 1993
1. Mr. Michael E. Beard
USEPA/AREAL
MD-78
RTF, NC 27711
Phone: (919) 541-2623
FAX: (919)541-0239
2. Dr. David A. Binstock
Research Triangle Institute
P.O. Box 12194, Bldg. 6
RTF, NC 27707
Phone: (919) 541-6896
FAX: (919)541-8778
3. Ms. Merrill Brophy
Maryland Department of Environment
Lead Program
2500 Broening Highway
Baltimore, MD 21224
Phone: (410) 631-3820
FAX: (410)631-4112
4. Mr. Bob Clickner
Westat
1650 Research Blvd.
Rockville, MD 20850-3129
Phone: (301) 294-2815
FAX: (301)294-2829
5. Mr. David Cox
David Cox and Associates
1511 K Street, NW, Suite 738
Washington, DC 20005
Phone: (202) 347-3090
FAX: (202)347-3106
6. Dr. Jim DeVoe
National Institute of Standards and Technology
Inorganic Analytical Research Division
U.S. Department of Commerce
Gaithersburg, MD 20899
Phone: (301) 975-4144
FAX: (301) 926-6182
-------
7. Dr. Robert Elias
USEPA
ECAO
MD-72
RTP, NC 27711
Phone: (919)541-4167
FAX: (919)541-0245
8. Dr. Eva Estes
Research Triangle Institute
P.O. Box 12194, Bldg. 6
RTP, NC 27707
Phone: (919)541-5926
FAX: (919)541-8778
9. Ms. Nancy Friederich
Midwest Research Institute
425 Volker Blvd.
Kansas City, MO 64110
Phone: (816)753-7600
FAX: (816)753-8420
10. Dr. William F. Gutknecht
Research Triangle Institute
P.O. Box 12194, Bldg. 6
RTP, NC 27707
Phone: (919) 541-6883
FAX: (919)541-8778
11. Ms. Susan Guyaux
Maryland Department of Environment
Lead Program
2500 Broening Highway
Baltimore, MD 21224
Phone: (410)631-3824
FAX: (410)631-4112
12. Ms. Sharon L. Harper
USEPA/AREAL
MD-78
RTP, NC 27711
Phone: (919) 541-2443
FAX: (919)541-3527
-------
13. Mr. Jim Hayes
N.C. State Government
Environmental Health Services Section
P. O. Box 27687
Raleigh, NC 27611-7687
Phone: (919) 733-2884
FAX: (919)733-0488
14. Mr. Lynn Hill
356 Amhurst Blvd.
Dayton, OH 45440
Phone: (513)427-1499
15. Mr. Jon Lathers
Wayne County Environmental Health Dept.
5454 Venoy Street
Wayne, Ml 48184
Phone: (313) 326-4900
FAX: (313)326-7221
16. Mr. Warren A. Loseke
USEPA/AREAL
MD-78
RTP, NC 27711
Phone: (919)541-2173
FAX: (919)541-3527
17. Dr. Mary McKnight
National Institute of Standards of Technology
Building 226, Room B348
Gaithersburg, MD 20899
Phone: (301) 975-6714
FAX: (301)975-4032
18. Dr. John D. Neefus
Research Triangle Institute
P.O. Box 12194, Bldg. 7
RTP, NC 27707
Phone: (919)541-6578
FAX: (919)541-8778
19. Dr. Chris Papanicolopoulos
Georgia Tech Research Institute
Environmental Sciences and Technology Laboratory
O'Keefe Building, Room 107
Atlanta, GA 30332-0800
Phone: (404) 894-3617
FAX: (404) 894-3906
-------
20. Mr. John Pesce
Star Environmental Services, Inc.
P. O. Box 1027
Melrose, MA 02176
Phone: (617)662-2220
FAX: (617)979-0060
21. Mr. Roy Petre
Childhood Lead Poisoning Prevention Program
Massachusetts Dept. of Public Health
305 South Street
Jamaica Plain, MA 02130
Phone: (617) 522-3700
FAX: (617)522-8735
22. Ms. Angel C. Price
Research Triangle Institute
P.O. Box 12194, Bldg. 6
RTP, NC 27707
Phone: (919)541-6555
FAX: (919)541-8778
23. Mr. John Scalera
USEPA
Mail Code TS-798
401 M. Street, SW
Washington, DC 20460
Phone: (202) 260-6709
FAX: (202)260-0001
24. Mr. Brad Schultz
USEPA
Mail Code TS-798
401 M. Street, SW
Washington, DC 20460
Phone: (202) 260-3896
FAX: (202)260-0001
-------
25. Mr. John Schwemberger
USEPA
Mail Code TS-798
401 M. Street, SW
Washington, DC 20460
Phone: (202) 260-7195
FAX: (202) 260-0001
26. Dr. Jim Simpson
Center for Disease Control
4770 Buford Highway, NE
Mail Stop F-42
Atlanta, GA 30341-3724
Phone: (404) 488-7330
FAX: (404) 488-7335
27. Dr. Tom Spittler
U.S. Environmental Protection Agency
60 Westview Street
Lexington, MA 02173
Phone: (617) 860-4334
FAX: (617) 860-4397
28. Ms. Cynthia (Cindy) R. Stroup
USEPA
Mail Code TS-798
401 M. Street, SW
Washington, DC 20460
Phone: (202) 260-3886
FAX: (202) 260-0001
29. Mr. Steve Weitz
Department of Housing and Urban Development
Room 8136
451 Seventh Street, SW
Washington, DC 20410
Phone: (202) 755-1805
FAX: (202) 1000
-------
30. Mr. Kenn White
T.C. Analytics
1200 Boissevain Avenue
Norfolk, VA 23507
Phone: (804) 627-0400
FAX: (804)627-1118
31. Ms. Emily E. Williams
Research Triangle Institute
P.O. Box 12194, Bldg. 7
RTP, NC 27707
Phone: (919) 541-6217
FAX: (919)541-8778
32. Mr. John Zilka
Applied Systems
2003 Sheffield Road, Suite B
Aliquippa, PA 15001
Phone: (412) 378-3066
FAX: (412) 378-8324
-------
Suppliers of Portable/Transportable XRF Instruments
May 1992
1. MAP
Scitec Corporation
2000 Logston Blvd.
Richland, WA 99352
TEL (509) 375-5000
FAX (509) 375-4931
Larry Lynott
2. Spectrace 9000
Spectrace Instruments
(Corporate Office)
345 East Middlefield Road
Mountain View, CA 94043
TEL (415) 967-0350
FAX (415) 967-6316
Todd Rhea
(Regional Office)
2401 Research Blvd.
Suite 206
Fort Collins, CO 80526
TEL (303) 493-2219
FAX (303) 493-2520
3. X-MET 880
Outokumpu Electronicxs
1900 N.E. Division St., #204
Bend, OR 97701
TEL (800) 229-9209
FAX (503) 385-6750
Stan Piorek
-------
4. PGT XK-3
Princeton Gamma-Tech, Inc.
1200 State Road
Princeton, NJ 08540
TEL (609) 924-7310
FAX (609) 924-1729
5. Microlead I
Warrington, Inc.
2113 Wells Branch Parkway
Suite 6700
Austin, TX 78728
TEL (512) 251-7771
FAX (512) 251-7744
Gary Stafford
6. Model SEFA-P
HNU Systems, Inc.
160 Charlemont St.
Newton, MA 02161-9987
TEL (617) 964-6690
FAX (617) 965-5812
Robert Petitti
-------
APPENDIX D
Papers/Information Distributed by Speakers and Manufacturers
-------
Data Presented by
Dr. Robert Clickner
-------
January 6. 1993
On the Performance of
the Scitec MAP/XRF
in the
National Survey of
Lead-Based Paint in Housing
by
Robert Clickner and John Rogers
WESTAT
January 11,1993
This work was sponsored by
Department of Housing and Urban Development
and the Environmental Protection Agency.
-1-
-------
January 8, 1993
National Survey of Lead Paint in Housing
o 381 housing units across the United States
o January to March, 1990
o Scitec MAP-3 spectrum analyzer X-ray fluorescence devices
o Used 8 nominally identical MAPs
o Each MAP delivered from Scitec with
o Fresh 40 mci Co57 radioactive source
o Calibrated to lead in paint standards of 0.0, 0.22,
1.20, and 2.64 mg/cm2
o "Automatic substrate correction" software
o Scitec claim: automatic substrate correction eliminates
the need for substrate corrections
o Thousands of 60-second XRF measurements on
dozens of different substrates
-2-
-------
Januarys, 1993
Validation Measurements in National Survey
o Continual check of MAPs' performance
o Shims (from NIST) with lead paint loadings:
0.6 + 0.02 mg/cm2
2.99 + 0.3 mg/cm2.
o Four substrates:
Wood, drywall, steel, and concrete
o Daily validation measurements for each MAP/shim/substrate
o Regression analyses on the validation data
o Estimate the precision and accuracy of the readings
o Relate readings to
o Substrate
o Lead loading level
o Different equation for each machine/substrate
o Equations inverted to calibrate readings for analyses.
-3-
-------
Januarys, 1993
Scitec MAP Performance Characteristics
o Readings are systematically different from the amount
of lead in the paint.
o The direction and magnitude of the differences are related to
o Substrate material
o Lead loading in the paint
o Two nominally identical MAPs may exhibit
significantly different performance characteristics
D Precision of the readings depends on the substrate
Therefore, in contrast to the Interim Guidelines and
Scitec's claims, substrate correction is in fact a
necessary step in the accurate determination of the
presence and amount of lead-based paint on surfaces.
o Findings apply only to XRF devices like the Scitec MAP-3s
described above and to closely related XRF devices.
-------
Januarys, 1993
Wood
o Readings < actual loadings when loadings < 2.0 mg/cm2
Readings > actual loadings when loadings > 2.0 mg/cm2
o At 0.6 mg/cm2
a 88% of the readings were < 0.6 mg/cm2
o 52% of the readings were equal to 0.0 mg/cm2.
o (MAP never produces a negative reading.)
o At 2.99 mg/cm2
a 60% of the readings were > 3.0 mg/cm2.
o Standard deviation of repeated readings is 0.25 mg/cm2.
o 7 of the 8 MAPs performed similarly
-5-
-------
January 8, 1993
Drywall or Plaster
a Performance is similar to the performance on wood
o Readings < actual loadings when loadings < 2.5 mg/cm2
Readings > actual loadings when loadings > 2.5 mg/cm2
o At 0.6 mg/cm2
o 82% of the readings were < 0.6 mg/cm2
o 52% were equal to 0.0 mg/cm2
o At 2.99 mg/cm2
o 44% of the readings < 3.0 mg/cm2
o 45% were > 3.0 mg/cm2.
o Standard deviation of repeated readings is 0.25 mg/cm2
o 6 of 8 MAPs exhibited similar performance characteristics
-6-
-------
January 8, 1993
Steel
o Readings > actual loadings, for all observed levels of lead
o At 0.6 mg/cm2
o 88% of the readings were > 0.6 mg/cm2
o At 2.99 mg/cm2
o 60% of the readings were > 3.0 mg/cm2
o 1 machine had no readings above 3.0 mg/cm2
a Substantial differences among the 8 machines
a Standard deviation of repeated readings is 0.21 mg/cm2
-7-
-------
Januarys, 1993
Concrete, brick, and other related substrates
o MAP has difficulty detecting low/moderate levels of lead
o At 0.6 mg/cm2
o 95% of the readings were equal to 0.0 mg/cm2
o Only 1 machine had more than one non-zero reading
o At 2.99 mg/cm2
o 95% of the readings were < 3.0 mg/cm2
o 11 % of these were equal to 0.0 mg/cm2
o Standard deviation of repeated readings is 0.49 mg/cm2
a There is much variation, both among and within machines.
-8-
-------
WESTAT
MAP Validation Readings on 0.6
mg/sq cm Shim
Concrete
Steel
Plaster, Drywall
Wood Substrate
Reading, mg/sq cm
178/93
-------
WESTAT
MAP Validation Readings on
2.99 mg/sq cm Shim
Concrete
// Steel
Plaster, Drywall Substrate
Wood
Reading, mg/sq cm
MAP299.XLC
M
-------
WESTAT
Estimating Lead Concentrations on Wood: Calibration Equations
for 8 Spectrum Analyzer XRFs.
3.50
3.00 —
2.50
u
o-
•)
I 2.00
0.00
Bold Line: Reading = Concentration
0.00
0.50
1.00 1.50 2.00
XRF Reading, mg./sq. cm.
2.50
3.00
PBWOOD.XLC
1/8/93
-------
WESTAT
Estimating Lead Concentrations on Plaster or Dry wall: Calibration
Equations for 8 Spectrum Analyzer XRFs.
4.50
4.00
3.50 —
3.00 -+-
u
I 2-50
0.00
-0.50
Bold Line: Reading = Concentration
0.50 1.00 1.50 2.00 2.50 3.00
XRF Reading, mg./sq. cm.
PBPLASTR.XLC
1/8/93
-------
WESTAT
Estimating Lead Concentrations on Steel: Calibration Equations
for 8 Spectrum Analyzer XRFs.
5.00 -i—
4.50 —
4.00 4-
Bold line: Reading = Concentration
1.50 2.00 2.50 3.00 3.50 4.00
-1.00 -*-
XRF Reading, mg./sq. cm.
PBSTEEL.XLC
1/8/93
-------
WESTAT
Estimating Lead Concentrations on Concrete: Calibration
Equations for 8 Spectrum Analyzer XRFs.
5.00
4p f\ _.
.50
4.00
3.50 —
u
CT
"Si
o
'€
3.00
2.50 -1—
u
o
u
Q.
»*
•>
0.00
Bold Line: Reading = Concentration
2.00 H-
1.50 •+-)
1.00 --
0.50 —
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00
XRF Reading, mg./sq. cm.
PBCONCRT.XLC
1/8/93
-------
Figure D-18 Validation XRF Readings versus Shim Lead Concentration
for XRF Instrument #36 on Wood
Lead concentration (mg/sq cm) - 0.5779 + 0.6977*XRF reading + 0.00069*(Days since 2/2/90)
5 T
4 3-
3 *
o
•
XRF Reading Histogram
Outliers (not used)
Calibration Equation
XRF & 1 mg/sq cm
XRF - Lead Concentration
Shim Concentration
XRF @ 1 mg/sq cm Is approximately 0.59
Median used for 0.6 shim
Lead Concentration (mg/sq cm)
-------
Figure D-34 Validation XRF Readings versus Shim Lead Concentration
for XRF Instrument #36 on Drywall
Lead concentration (mg/sq cm) • 0.1694 + 0.8220'XRF reading 4 0.00081 "(Days since 2/2/90)
4.5 T
o XRF Reading Histogram
• Outliers (not used)
Calibration Equation
XRF @ t mg/sq cm
XRF - Lead Concentration
Shim Concentration
XRF @ 1 mg/sq cm Is approximately 0.99
-0.5 -1-
Lead Concentration (mg/sq cm)
-------
Figure D-35 Validation XRF Readings versus Shim Lead Concentration
for XRF Instrument #37 on Drywall
Lead concentration (mg/sq cm) - 0.2311 + 0.9186*XRF reading + 0.00091*(Days since 2/2/90)
4 -r
XRF Reading Histogram
Outliers (not used)
Calibration Equation
XRF @ 1 mg/sq cm
XRF - Lead Concentration
Shim Concentration
XRF @ 1 mg/sq cm Is approximately 0.81
Lead Concentration (mg/sq cm)
-------
Figure D-10 Validation XRF Readings versus Shim Lead Concentration
for XRF Instrument #36 on Steel
Lead concentration (mg/sq cm) • -1.0265 + 1.0580'XRF reading + 0.00105*(Days since 2/2/90)
5 T
4.5 --
0.5
1 1.5 2 2.5 3
Lead Concentration (mg/sq cm)
o
XRF Reading Histogram
Outliers (not used)
Calibration Equation
XRF 9 1 mg/sq cm
XRF - Lead Concentration
Shim Concentration
XRF @ 1 mg/sq cm Is approximately 1.90
-------
Figure D-13 Validation XRF Readings versus Shim Lead Concentration
(or XRF Instrument #39 on Steel
Lead concentration (mg/sq cm) - -0.9832 + 1.1274'XRF reading + 0.00112*(Days since 2/2/90)
4.5 T
r i i
0.5 1 1.5 2 2.5 3
Lead Concentration (mg/sq cm)
3.5
XRF Reading Histogram
Outliers (not used)
Calibration Equation
XRF <8> 1 mg/sq cm
XRF • Lead Concentration
Shim Concentration .
XRF @ 1 mg/sq cm Is approximately 1.73
-------
Figure D-26 Validation XRF Readings versus Shim Lead Concentration
for XRF Instrument 036 on Concrete
Lead concentration (mg/sq cm) - 0.5692 + 1.5546'XRF reading + 0.00154*(Days since 2/2/90)
3.5 T
XRF Reading Histogram
Outliers (not used)
Calibration Equation
XRF® 1 mg/sq cm
XRF - Lead Concentration
Shim Concentration
XRF @ 1 mg/sq cm Is approximately 0.26
Median used for 0.6 shim
Lead Concentration (mg/sq cm)
-------
Figure D-28 Validation XRF Readings versus Shim Lead Concentration
for XRF Instrument #38 on Concrete
Lead concentration (mg/sq cm) » 0.5753 + 1.1586'XRF reading + 0.00115*(Days since 2/2/90)
o
XRF Reading Histogram
Outliers (not used)
Calibration Equation
XRF @ 1 mg/sq cm
XRF - Lead Concentration
Shim Concentration
•I XRF @ 1 mg/sq cm Is approximately 0.34
4 Median used for 0.6 shim
Lead Concentration (mg/sq cm)
-------
Flow Diagram Presented by
Dr. Mary McKnight
-------
Evaluation of Methods for Field Measurement of Lead In Paint Films
ITatnKllsrift IVarfr
kmKlwm '
TOOldll '
"*V»a»-ar«f»rlrr«i-Hrm ObMTVB field |»m^h»a*
review knowledge
Idniuty key \
affect field re
jn^a*TTlTlg*My^| \jPfntff\ llll^ff^t^l Hj*|"flJpJ*T^^| iBMfl
•Mdgn study Qaboraioty study)
t
fVwwiiK't JtinrtpJ «ifniilfl*«fl
Snalflflltlfln erf A^kltfl mt**A*r 1
ntwntnnr_fMi^fw1 — i '
ai^jKcuutjr-iLNUICU i
^id *J^imil^T^o*\ *. .— ..
/XF
XresuBs
V. - -** »
Lrmn.«i_ruM. ^*« HA
1
agcot
*
actors that
sutts
4-
Deslg
field
/Vwu4
(jono.
study
leM \\
ii ii •• •••Ji» ai^^
^ _A* _ «« X
ttr
is--
y
n
sfa
uc
r
i
«l
4*
ai
H
uc
^
n
be wrtt
mttea
»7
1 in. iff A
o
i other
d4tt»M
nciu
«-
»
^
^-
<
1
if
/
f
1
\
4
1
1
I
I
?
i
»,
*
it
%
».
j.
1
i
J
^
>
equtpment or
measurement procedures
study; anatyze
Develop fcpowledge^
system
Benefits
Standard Method (protocol) for
equjpment or other
measurement procedures for lead
precision and bias of measumients
Ahfltty to cover range of field
(e-g.. lead cone., environmenta
conditions, substrate types)
uvul
ifng of faction
quattty
Dedslan-support
-------
Data Presented by
Mr. John Pesce
-------
AN EXAMINATION OF SUBSTRATE
EFFECT ON PORTABLE X-RAY
FLUORESCENCE INSTRUMENTATION
Pesce, J., Star Environmental Services, Inc. P.O. Box 1027,
Melrose, MA 02176; Martin, K.P., Massachusetts Institute of
Technology, 77 Mass. Ave., Cambridge, MA 02139; Straub, W.E.,
Straub Industrial Hygiene, 250 Grove St., Melrose, MA 02176;
Edwards, R., Massachusetts Institute of Technology, 77 Mass. Ave.,
Cambridge, MA 02139.
ABSTRACT
Mobile, hand held X-ray fluorescence analyzers (XRF) are
used to measure concentrations of lead in paint. The
three manufacturers of the instruments treat background
readings in different ways. The same surface can give
significantly different readings from using different
instruments. Some instruments will give readings in
excess of legal standards for substrate materials when
there is no lead present. Others will give a negative
reading for a substrate which will lower the true lead
concentration reading. Readings can be corrected by
taking readings from bare substrate surfaces, but this is
an expensive and time consuming process. Surfaces can
also be checked by other confirmatory techniques. This
paper examines problems in the use of X-ray analyzers
when measuring lead concentrations in painted surfaces.
It also demonstrates the need for confirmatory methods.
Test results are presented on the effects of over twenty
different substrate materials and conditions. Materials
were tested at several lead concentrations. A testing
protocol using XRF with confirmatory testing of surfaces
using sodium sulfide and atomic absorption spectroscopy
is presented with the experience of using it in six
buildings containing over four hundred apartment units.
X-ray analyzers alone are not adequate to determine lead
concentrations in painted surfaces. Substrate effects
and confirmatory testing should be considered.
Government regulators should not exclude techniques that
are adjunct to the use of the XRF.
-------
INTRODUCTION
The purpose of this presentation is to share our experiences of how
various substrates effect the hand held X-ray fluorescent analyzers
designed to measure lead in paint. The substrate effect is usually
called substrate effect level and sometime referred to as substrate
equivalent lead, and abbreviated SEL. Our data is derived from
both actual field use and recently developed Round Robin
information. The Round Robin encompassed over thirty different
instruments that are being used in the fields daily.
The mobile X-ray fluorescent (XRF) analyzers to be discuss are
those frequently used for in-situ for lead in paint.
The first to be discussed is that Princeton Gamma Tech,
specifically the model XK3, This device uses a cobalt 57 source of
approximately 10 millicuries. The instrument counts the X-ray
events using a krypton filled proportional tube. Background
effects are compensated by monitoring a single lower X-ray
frequency, which changes the high voltage in the tube.
The second type is the Warrington Microlead I, Revision 4. The
Microlead also uses a cobalt 57 source of approximately 10
millicuries. This instrument uses a high and low band pass
screening with a two filter system. The filters are alternated
during the reading cycle. The filtered X-rays are then
scintillated and counted by a photomultiplier tube. The counts at
the high and low frequencies are referenced by computer logic to
adjust the voltage of the dynodes in the photomultipier tube.
The third type of XRF to be talked about is the SCITEC Metal
Analysis Probe analyzer. This instrument almost always uses a 40
millicurie cobalt 57 source. The detector used is a crystal. The
analyzer uses an algorhithem based upon information received from
five reference channels at different frequencies. The algorhithem
is designed to correct for most substrate effects. Although the
SCITEC has the ability to be spectrum read, most of time it is
operating in the direct read mode.
FIELD DATA
The broad base of building materials that can be painted with lead
based paint, and consequently must be tested, is staggering.
Residential wood frame construction poses many wood and plaster
items inside, many times these materials are in tandem. Exteriors
of these types of buildings also often have various types of
coverings over wood, including wood over wood. Larger buildings
both residential and commercial generally favor more concrete,
brick, and denser plaster. These buildings also have a tendency to
contain more metal architectural features and less wood millwork.
-------
*
*
*
*
FIGURE 1
FIELD TESTING
421 units tested in an apartment complex
All interior and exterior surfaces were tested/
resulting in over 20/000 readings
PGT-XX3/ XRF instruments were used
32 substrate effect levels (SEL) were measured
A comprehensive lead inspection of 421 units in high-rise apartment
building were performed using all PGT's. The inspection was
conducted to examine surfaces for lead on 100% of all interior,
common and exterior surfaces.
The protocol used for testing included SEL's on all surfaces and
confirmatory use of both sodium sulfide solution (6-8%) and atomic
absorption. Confirmation techniques were engaged when certain
levels were found by XRF. Sodium sulfide solution was used when
Apparent Lead Concentrations were 1.2 mg/cm2 and samples, taken for
AA analysis when Corrected Lead Concentrations were between 0.8
mg/cm2 to 1.5 mg/cm2. Consequently SEL's had to be take on ceilings,
walls, windows, etc. Having this significant amount of data on the
SEL's for the building, we took the opportunity to test the SEL
locations with the PGT, Warrington and the SCITEC. The following
is the data collected:
FIGURE 2
ALL
Masonry
Cinder Block
Plaster type A
Plaster column
Red Brick
Cinder Block
Sheet Rock
Concrete Column
Wood
Pressed Board
Hollow Door
Oak
Metal
Window Sill
Radiator Cover
Vent Duct
Mail Box
TYPES OF
PGT
1.1
0.6
1.2
1.1
1.1
0.2
0.9
0.0
0.4
0.2
1.9
1.6
1.1
1.5
SUBSTRATES
WAR
1.2
-0.1
1.7
1.3
1.2
-0.3
1.7
-0.1
-0.3
-0.2
-1.1
-1.0
-0.7
-1.1
SCI
0.0
0.0
0.0
0.0
0.0
0.2
0.3
0.0
0.0
0.3
0.4
0.3
0.3
0.3
-------
Since it is expected that these instruments will be used over a
wide variety of materials, we selected representative samples.
The tremendous variability of the XRF's SEL on groups within the
same type of building materials such as plaster/masonry, cause us
to take a more microscopic view of the SEL's within similar
building components.
FIGURE 3
MASONRY
Cinder Block
Plaster type A
Plaster column
Plaster type B
Red Brick
Cinder Block
Sheet Rock
Concrete Column
TYPES OF
PGT
1.1
0.6
1.2
0.6
1.1
1.1
0.2
0.9
SUBSTRATE
WAR
1.2
-0.1
1.7
0.1
1.3
1.2
-0.3
1.7
SCI
0.0
0.0
0.0
0.0
0.0
0.0
0.2
0.3
Note the variability of the PGT and Warrington with respect to
these type of building materials. Also note how the SCITEC almost
seems to have a flat response to most of these substrates. It will
also be interesting to note that both the PGT and Warrington can
give negative readings and the SCITEC cannot. This list
demonstrates how important it is to know which type of wall or
c.eiling that is being tested before any final determination can be
stated about its lead content. That is, if the wall is a
supporting concrete wall as opposed to a dividing wall, significant
differences of substrate effect occur.
FIGURE 4
METAL SUBSTRATES
Window Sill
Radiator Cover
Vent Duct
Mail Box
PGT
1.9
1.6
1.1
1.5
WAR
-1.1
-1.0
-0.7
-1.1
SCI
0.4
0.4
0.3
0.3
As shown in figure 4, for metals both the PGT and Warrington seem
to treat the substrate effects completely opposite. The elevated
readings for the PGT are similar in absolute values as the
depressed readings of the Warrington. Again the SCITEC seems only
modestly influenced by the substrate.
-------
FIGURE 5
WOOD SUBSTRATES
Pressed Board
Hollow Door
Oak
PGT
0.0
0.4
0.2
WAR
-0.1
-0.3
-0.2
SCI
0.0
0.0
0.3
As shown in figure 5 the XRFs apparently work on less dense
materials. The standard deviations reported by the manufactures of
.3 mg/cm2 can usually be depended upon for wood type substrates.
ROUND ROBIN INFORMATION
Over the past year, a group of private lead inspectors in
Massachusetts have established and participated in a voluntary
Round Robin testing procedure for portable X-ray fluorescence
instruments. As part of this Round Robin program, substrate
effects were examined. A detailed presentation of this Round Robin
program will be presented on Wednesday by another author of this
paper.
ALC
SEL
CLC
FIGURE 6
STEEL CHANNEL
PGT ( 9 )
1.8
0.9
0.9
WAR (5)
-0.8
-0.8
0.0
SCI(l)
1.2
0.4
0.8
In this slide the Substrate Effect Level, called the SEL, is a
piece of steel channel, with no lead paint. The Apparent Lead
Concentration, called the ALC, is the same steel channel with lead
paint. To make a substrate effect correction to the steel channel
with lead paint, the reading for the steel channel without lead
paint is subtracted. This gives a Corrected Lead Concentration
called the CLC. All readings are expressed in mg/cm2. For the
steel channel samples prepared for the Round Robin the CLC of PGT
and SCITEC showed some measure of lead while the Warrington's CLC
showed there to be no lead.
-------
ALC
SEL
CLC
FIGURE 7
RED BRICK
PGT ( 9 )
1.8
1.2
0.6
WAR (5)
1.4
0.3
1.2
SCI(l)
0.0
0.0
0.0
Red bricks were also prepared with and without lead paint on their
surfaces. What is shown in figure 7 is obvious that the SCITEC did
not pick up any amount of lead. While the zero readings of the
SCITEC on substrates is, at first, comforting, it can be seen from
this slide that like the Warrington on metals, a false negative
reading can result.
ALC
SEL
CLC
FIGURE 8
POPLAR
PGT (9)
1.2
0.0
1.2
WAR (5)
0.9
-0.1
1.0
SCI(l)
0.6
0.1
0.5
Poplar wood samples prepared for the Round Robin and shown in
figure 8 indicates that the XRF's were able to identify leaded
surfaces, although quantification are not in good agreement.
FIGURE 9
BELLY CAS I KG
ALC
SEL
CLC
PGT (9)
1.0
O.O
1.0
WAR(5)
1.2
0.3
0.9
SCI(l)
0.7
0.0
0.7
A decorative wood molding called belly casing was also used during
a Round Robin test. This was done to check the effects of an air
gap between the XRF probe and the sample. The results shown in
figure 9 are consistent with the flat wood sample shown previously.
This type of belly casing has a convexed fluted surface that would
allow only direct contact with a partial portion of the instrument
probe area.
-------
ALC
SEL
CLC
FIGURE 10
SHEET ROCK
PGT ( 9 )
1.4
0.2
1.2
WAR (5)
0.8
0.1
0.7
SCI(l)
1.2
0.4
0.8
The Round Robin sheetrock samples also produced a consistent result
for the three different XRF's.
ALC
SEL
CLC
FIGURE 11
READ-THROUGH EFFECTS
Lead Flashing under 1 1/2" of wood
PGT (13) WAR (11)
1.7 10.7
0.0 -0.1
1.7 10.8
SCI(l)
4.7
0.1
4.6
The penetration of the gamma ray and consequently the ability of
the K shell X-ray to also penetrate materials, can produce a
reading on the XRF's when lead is deep beneath the surface. It is
interesting to notice that in figure 11 the PGT is less susceptible
to detect this type of lead.
The instruments seem to do well over most of the less dense
materials, those which would most likely be found in wood frame
houses with less than four families. Higher density materials and
metals pose a problem for the XRF's and provide a need to correct
for the type of substrate being examined. This would indicate that
buildings with construction components of higher density would
require a more keen attention to substrate effect on the lead X-ray
frequency.
At present most of the level driven lead abatement is based upon
XRF reading of around 1.0 mg/cm2. This is well within a range for
which substrate effects alone can be contained. It is possible to
read above this level and not have any lead present.
In an effort to reduce the error resultant from XRF testing alone, we
used secondary and tertiary screening methods. The secondary
screening method used was AA. The tertiary screening was done with
the use of sodium sulfide solution (6-8%). (The fact that this
testing procedure is currently only accepted by the Commonwealth of
Massachusetts should not mitigate it usefulness as a confirmatory
method.)
-------
CONCLUSIONS
It is our conclusions that for a wide variety of substrates the
instruments can be used with a reasonable amount of confidence.
This can be aided by using substrate corrections but assumes the
operator has the experience to know what the substrate is and how to
make appropriate corrections. For example structural concrete walls
and those simply that divide rooms can look deceptively similar, yet
the XRF can have profoundly different readings. Lead/lead paint
behind even one inch thick wood can be mistaken as surface lead. The
operator of the XRF must have enough knowledge to know when to doubt
the XRF and refer back to secondary and tertiary screening methods.
The XRFs are also not able to compensate for all possible
substrates. Substrate correction values can be different from
instruments of the same manufacturer as well as differ from those
of each manufacturer.
Someone's mistaken perception about the level of lead on a surface
could indicate that no lead is present when dangerous levels of lead
are present, or could affect the cost of abatement significantly by
indicating abatement when none is needed. Although it is possible to
use SEL's to correct for these effects to some extent, the degree of
correction has limited accuracy. Confirmatory testing should be used
on certain substrates when lead levels that could influence the
abatement decision are marginal.
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — A
SAMPLE DESCRIPTION
This sample consisted of three layers:
-Top layer> two pieces of unpainted, 3/4" thick, wood.
-Middle layer> lead roofing flashing.
-Bottom layer> 1/4" plywood.
>»The flashing was the only lead in the sample.
PGT
ID Code
watch
dvblt
cegbm
gblam
mkcfx
mejsc
adlwz
walzj
gapcu
pwfal
bdfhj
kmprt
acegi
mean
aviation
Reading
2
.3
2.4
1
1
1
1,
1.
2.
1.
1.
1.
1.
1.
1.
0.
.4
.0
.5
.1
.5
1
7
9
8
4
6
7
42
Warrington
ID Code
tolpk
Iqrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
10.
11.
9.
11.
10.
10.
10.
11.
12.
10.
10.
9
5
8
3
5
7
5
6
3
1
3
10.7
0.74
SciTec
ID Code
Ingbv
Reading
4.68
i
4.68
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — B
SAMPLE DESCRIPTION
This sample consisted of four layers:
-Top layer> one sheet of heavy paper.
-Second layer> lead paint.
-Third layer> two inch thick concrete (from cinder block)
-Bottom layer> 1/4" plywood.
PGT
ID Code
watch
dvblt
cegbm
gblam
mkcfx
mejsc
adlwz
walzj
qapcu
pwfal
bdfhj
kmprt
acegi
mean
aviation
Reading
2
2
2
1
1
1
3
2
2
2.
2
2
1,
2.
0.
.6
.2
.3
.5
.3
.5
.0
.7
.2
.2
.8
.0
.9
2
54
Warrington
ID Code
tolpk
Igrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
1
1,
1.
1.
1.
2.
1.
2.
1.
2.
1.
1.
0.
.6
.4
9
7
4
2
4
3
4
0
7
7
33
SciTec
ID Code
Ingbv
Reading
0.13
0.13
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — C
SAMPLE DESCRIPTION
This sample consisted of two layers;
-Top layer> lead paint.
-Bottom layer> wood door, 1 3/4 " thick.
PGT
ID Code
watch
dvblt
cegbm
gblam
mJccfx
mejsc
adlwz
walzj
qapcu
pwfal
bdfhj
kmprt
acegi
mean
aviation
Reading
2.
2.
1,
1
2
1
2
2
2
2.
3.
2.
2.
2.
0.
.3
.6
.4
.9
.1
.6
.3
.7
7
7
1
0
5
3
.48
Warrington
ID Code
tolpk
Igrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
1.
1.
0.
1.
2.
2.
1.
2.
1.
1.
1.
1.
0.
2
6
7
9
0
5
5
0
6
7
7
,7
46
SciTec
ID Code
Inqbv
Reading
2.13
2.13
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — D
SAMPLE DESCRIPTION
This sample consisted of two layers:
-Top layer> non-lead paint.
-Bottom layer> wood door, 1 3/4 " thick.
PGT
ID Code
watch
dvblt
cegbm
qblam
mkcfx
mejsc
adlwz
walzj
qapcu
pwfal
bdfhj
kmprt
acegi
mean
aviation
Reading
0.3
0.1
-0.3
0.2
-0.1
0.0
0.0
'0.2
0.3
0.0
0.4
0.8
-0.2
0.1
0.29
Warrington
ID Code
tolpk
Iqrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
-0.1
-0.2
2.2
0.1
-0.4
0.9
0.6
0.4
0.0
0.0
0.8
SciTec
ID Code Reading
Inqbv 0 . 04
.
0.4
0.73
0.04
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — E
SAMPLE DESCRIPTION
This sample consisted of two layers:
-Top layer> lead paint.
-Bottom layer> wood door, 1 3/4 " thick.
PGT
ID Code
watch
dvblt
cegbm
qblam
mkcfx
mejsc
adlwz
walzj
gapcu
pwfal
bdfhj
Jonprt
acegi
mean
aviation
Reading
1.3
1.4
0.2
0.5
0.2
0.3
0.6
1.3
0.8
1.0
1.0
0.3
0.6
0.7
0.43
Warrington
ID Code
tolpk
Igrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
•
Reading
0.6
0.7
1.0
0.9
0.6
1.5
0.8
1.0
0.5
0.5
1.2
SciTec
ID Code Reading
Inqbv 0.63
_
0.9
0.31
0.63
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — F
SAMPLE DESCRIPTION
This sample consisted of three layers:
-Top layer> lead paint.
-Middle layer> two layers of cedar shingles.
-Bottom layer> 3/4 " wood.
PGT
ID Code
watch
dvblt
cegbm
qblam
mkcfx
mejsc
adlwz
walzj
gapcu
pwfal
bdfhj
kmprt
acegi
mean
aviation
Reading
0.6
0.9
0.4
0.4
0.6
0.5
0.6
0.9
0.9
0.6
1.2
0.4
0.6
0.7
0.25
Warrington
ID Code
tolpk
Igrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
1.4
1.8
1.5
1.8
1.4
1.8
1.3
2.0
1.7
1.4
1.5
SciTec
ID Code Reading
Ingbv 1.102
1.6
0.23
1.10
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — G
SAMPLE DESCRIPTION
This sample consisted of three layers:
-Top layer> one sheet of heavy paper.
-Middle layer> non-lead paint primer.
-Bottom layer> metal door jam.
PGT
ID Code
watch
dvblt
cegbm
gblam
mkcfx
mejsc
adlwz
walzj
qapcu
pwfal
bdfhj
kmprt
acegi
mean
Aviation
Reading
0.
0.
0.
0.
0.
0.
1,
1.
1.
0.
1,
0.
0.
0.
0.
.5
9
.2
.2
1
5
0
3
2
6
0
5
6
7
38
Warrington
ID Code
tolpk
Igrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
-0.
-1.
-0.
7
1
6
-0.8
-1.
-0.
-1.
-1.
-1.
-0.
-1.
-0
0
2
5
1
3
0
5
3
.9
.31
SciTec
ID Code
Inqbv
Reading
0.34
0.34
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE. — H
SAMPLE DESCRIPTION
This sample consists of four layers:
-Top layer> anodized aluminum metal house siding.
-Middle layer> lead paint (painted on bottom layer.
-Bottom layer> 3/4 " wood.
PGT
ID Code
watch
dvblt
cegbm
gblam
mkcfx
mejsc
adlwz
walzj
gapcu
pwfal
bdfhj
kmprt
acegi
mean
Aviation
Reading
1.3
1.5
1.1
1.0
0.6
0.4
1.3
1.9
1.5
1.5
1.&
1.5
1.3
1.3
0.42
Warrington
ID Code
tolpk
Igrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
1.2
1.1
0.9
1.5
1.0
1.1
0.7
1.4
0.3
1.0
0.7
SciTec
ID Code Reading
Ingbv 1.15
1.0
0.34
1.15
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — I
SAMPLE DESCRIPTION
This sample consisted of two layers;
-Top layer> lead paint.
-Bottom layer> 3/8 " sheetrock.
PGT
ID Code
watch
dvblt
cegbm
gblam
mkcfx
mejsc
adlwz
walzj
gapcu
pwfal
bdfhj
kmprt
acegi
mean
aviation
Reading
1,
2.
1.
1.
1.
0
1.
1.
1,
1.
.4
.0
.0
.2
.0
.9
.3
8
5
8
1.4
1.
1.
1.
0.
2
1
3
36
Warrington
ID Code
tolpk
Iqrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
0.
0.
0.
1.
0.
0.
0.
0.
0.
1.
0.
0.
0.
7
9
8
5
8
9
8
9
4
1
5
.9
.29
SciTec
ID Code
Ingbv
Reading
1.14
1.14
-------
ROUND ROBIN ROUND NUMBER ONE
DATE 10/03/91 AND 10/04/91
SAMPLE ID CODE — J
SAMPLE DESCRIPTION
This sample consists of two layers:
-Top layer> non-lead paint.
-Bottom layer> 3/8 " sheetrock.
PGT
ID Code
watch
dvblt
cegbm
gblam
mkcfx
mejsc
adlwz
walzj
qapcu
pwfal
bdfhj
kmprt
acegi
mean
iviation
Reading
0.2
0.5
-0.9
0.3
0.1
0.1
0.0
0.1
0.0
0.1
0.2
-0.1
0.0
0.1
0.32
Warrington
ID Code
tolpk
Igrom
azbyc
notrl
dfdce
hcdjr
npxtc
andkm
baino
dyzag
fbcps
Reading
-0.1
0.0
0.0
0.2
-0.2
0.0
-0.3
0.2
-0.4
-0.2
-0.3
SciTec
ID Code Reading
Inqbv 0.12
-0.1
0.2
0.12
-------
Data Presented by
Dr. Tom Spittler
-------
COUNTS
16260-X-4
1DOSEC
1000-
16360 VS 16260 (REVERSE SIDE) I
80°" DOT MODE - 0.7% PAINT CHIP
600T
4004
200-
S R
XES
KEY
-------
COUNTS
18282-X-4
100SEC
2000 >
1600-- DOT MODE - 9.6* PB PAINT CHIP
1200-
800'
400-
"XT
16262 VS 16262 (REVERSE SIDE) IM
KEV
XES
-------
COUNTS X 1000
14620-X-4
100SEC
14620 VS 14620CREVERSE SIDE
DOT MODE
4 8 % P B PAINT CHIP
0
KEY
XES
-------
COUNTS X 1000
14135-X-4
88SEC
8
1-4135 VS 1413
\ '•"*,
DOT MODE
2 4
REVERSE SID
B PAINT CHI
I N
< PB >
KEV
XES
-------
DATA FOR SOMAR DILUTION TEST
SAMP. #
MG 1579 + SOMAR MIX
7 8
2 1 0
3 6
2 1 5
1 6 9
2 1 6
M G 14135
6 5
209
-------
COUNTS
3-X-4
4000
3300 •
2900-
2000-
1500-
lOOOf
SOOr
BAVS 1.2 & 4 (IN DOT MODE)
KEY
u
XES
-------
-------
SAMP. # PEAK H WT. RATIO % PB
N B S 1
46 MM
0.371
4.45
N B 5 2
2 0 - M M
0.167
2.00
N B S 3
89 MM
0,782
9.38
14153
68 MM
0.311
23.1
-------
Data Presented by
Scitec Corporation
-------
April 27, 1993
Warren Loseke
U.S. EPA
Atmospheric Research and Exposure
Assessment Laboratory
Research Triangle Park, NC 27711
Dear Mr. Loseke:
During the 12 years which the MAP technology has been in existence, we've found that
the use of X-ray fluorescence for field measurements is often misunderstood, oversimplified, or
even disregarded. We appreciate your efforts to bring government and state personnel and their
contractors together to accurately identify and resolve the questions, concerns, and confusion
which exists.
The purpose of the use of portable XRF is the identification of lead hazards quickly and
economically, and in sufficient detail for cost effective abatement strategies to be implemented.
Policy makers and government agencies seem to focus their attention on the performance of the
XRF instrument as though the device alone is responsible for the valid identification of lead
hazards.
However, a majority of the errors that occur in the identification of lead hazards are
attributable to the sampling design used, the operator, the customer, and a lack of effective
quality control practices—not the sampling device. Attention must be given to determining how
to control and minimize errors and bias from all sources; sampling designs; operators;
instruments; laboratories; record keeping; report generation; and customers.
As the single largest customer for lead paint testing services, the Department of Housing
and Urban Development has significant influence over how lead paint testing and reporting is
accomplished. However, the objectives, goals, roles, and responsibilities in HUD's LBP testing
program are inconsistent and not aligned properly with incentives and control measures. This
accounts for most of the errors and data quality problems which result.
WI993L.2
2000 Logston Blvd. Telephone: 1-800-4-NO-Lead
FAX: 509-375-4931
Richland, WA 99352
-------
The customer should know what to specify in requests for bids and proposals to do
testing work, they need to know how to evaluate bids, how to oversee testing, and how to
evaluate the test data. Unfortunately most PHA customers today do not have the knowledge or
skills to handle these responsibilities. Without fundamental knowledge and skills, they lack the
ability to effectively let and administrate contracts. This deficiency is a major reason for the
poor quality of test data that is received by HUD.
Few HUD lead paint testing contracts even deal with the issue of "quality," or even
specify that the contractor is to provide accurate and verifiable test data! Many PHA's assume
they are tasked only with turning in test data to HUD, they have no feeling of obligation to make
sure the test data is good. Most PHA's believe they've done their job when they've awarded
a contract to the low bidder...as though their only job is to get a test report at a low price.
The sampling design and protocol required by HUD will not generate highly axurate
data nor result in identification of lead hazards 95% of the time. HUD's Guidelines for testing
by XRF require only the minimum level of testing. The apparent objective of the Guidelines
is to satisfy HUD's legislative directive to test...not necessarily to locate and determine the
presence and extent of lead hazards.
The Guidelines fail to recognize the fact that lead paint is heterogeneous, instead they
assume that paint is homogeneous and don't even require that such a critical assumption be
checked or verified.
A trained inspector will test painted surfaces in a way that will reveal the presence of
lead. Current HUD Guidelines reduce inspection requirements to the lowest possible standards
and thereby restrict inspectors' ability to utilize their skills and knowledge.
A well trained operator will know how to use the XRF instrument to generate statistically
valid data. The instruments are capable of generating valid test data if they are operating
properly and used properly. The current HUD testing guidelines don't even require an inspector
to prove that he/she conducted any testing, let alone prove that testing was accomplished with
a properly working instrument!
The current testing protocols underutilize existing XRF technology and worker skills
thereby causing total testing costs to be substantially higher than necessary. The full benefits
of XRF instruments are ignored and expensive backup laboratory testing required much more
frequently than is necessary. Alarmingly, this expensive backup testing is not controlled,
checked, monitored, or regulated!
Effective testing protocols would standardize the way dwellings and structures are tested
and test data reported so contractors are required to test in a similar manner, check instrument
041993L.2
-------
performance regularly, and report test data in a consistent format. Most importantly, contractors
should be required to prove that they checked instrument performance, tested surfaces, and
followed protocols.
Housing authorities should be made responsible for the quality of test data, released from
their obligation to award contracts to the low bidder, and trained in how to write bid
specifications and administrate contracts.
Policy makers need to focus their attention on all of the issues related to lead paint
testing: The objectives of testing need to be clearly stated and protocols developed to meet the
objectives. Housing authorities need their role defined, responsibilities identified, and the
training to carry them out so they can be held accountable.
Most of all, if HUD wants high quality test data they need to specify and require
inspectors to provide and prove the quality of the data, AND, HUD must be willing to pay for
it.
041993L.2
-------
In response to your survey we submit the following:
PORTABLE XRF SPECTROMETER
Metals Analysis Probe ("MAP") Model 3B, used with the Ambient Scanner 310B for lead
in paint and soil analysis.
MAP 3 - FA2C PS
Detection Limit
Detection limit is defined as the lowest quantity of total lead detectable in optimum
conditions with a reasonable measurement time. Detection limits are quantified in terms of Pb
K-line X-ray detection.
With XRF used for LBP analysis, the detection limit is a function of:
1. Measurement conditions
2. Measurement time
3. Confidence level desired
4. Source strength.
In LBP analysis the detection limit varies with the density of the substrate, the length of
the measurement or number of repeated measurements, and the source strength.
Generally, as measurement conditions worsen (substrate density increases), the detection
limit deteriorates, and, as measurement time increases or lengthens, the detection limit improves
(gets lower in real terms).
Detection limits for soil vary as a function of measurement time, matrix structure, and
the elements contained in the sample.
For lead based paint measurement, the following detection limits are stated in terms of
K- and L-line precision levels at 1 standard deviation on various substrates or combinations of
building materials.
041993L.2
-------
DETECTION LIMITS - LEAD IN PAINT
The detection limits for lead in paint on low, medium, and high density substrates are
stated in terms of precision at 15, 60, and 240 seconds with a 40 mCi Co" radioisotope.
Precision is stated as one standard deviation of a series of consecutive measurements.
Substrate
(density)
Thin wood (low)
Wallboard (medium)
Concrete (high)
K-Shell Detection Limits in mg/cm2 at
Preprogrammed Measurement Times
15 sees
0.2
0.3
0.8
60 sees
0.1
0.1
0.4
240 sees
0.05
0.1
0.2
The above stated K-shell detection limits were derived from data recently and randomly
collected from 46 different instruments with 25 repeated measurements each on plaster wallboard
backed with 2x4 (nominal) wood, a hollow core door, and concrete block. The average scatter
for stud backed wallboard for all 46 MAP instruments was 0.1938, with one standard deviation
of 0.307 mg/cm2.
Substrate
(density)
Thin wood (low)
Wallboard (medium)
Concrete (high)
L-Shell Detection Limits in mg/cm2 at
Preprogrammed Measurement Times
15 sees
0.1
0.2
0.6
60 sees
0.05
0.1
0.3
240 sees
0.025
0.075
0.15
Q. Are there any plans to allow for direct user calibration or recalibration?
A. Scitec's policy does not allow for premature announcement of new products, features,
or services. However, in general, our product strategy has not included user calibration
or recalibration features for three important reasons:
1. Accurate field calibration requires highly skilled, trained, and experienced
technicians, which are short in supply, difficult to get to do fieldwork, and
relatively expensive. This type of person is usually not willing to do paint testing
on a regular basis.
0419WL.2
-------
2. It is difficult to design parameters that would allow quality control of field
calibrations. Allowing user calibration would allow users to set up instruments
to produce biased measurements, e.g., "control" data output if desired.
Q. Are you aware of the new MIST paint film SRM's?
A. Yes. Scitec has purchased several sets of these new standards.
Q. Have they been of any value to you in calibration or testing your instruments?
A. No, not yet. Only a small percentage of the XRF instruments in use are calibrated with
the new standards. Recalibration requires instrument downtime and expense. It will be
some time before all instruments are returned from the field and recalibrated. Many
contractors are not willing to purchase the MIST standards at the price NIST is asking
nor pay for recalibration of their instruments.
Q. Briefly describe the procedure for use of your instrument.
A. 1. Turn it on
2. Enter current date
3. Enter calibration check ID code
4. Make calibration checks
5. Plot cal check results on graph
6. Verify instrument performance
If OK,
7. Enter sample reference "ID" number
8. Make a "TEST" type measurement (@ 60 seconds)
9. When apparently homogeneous surfaces show patterns of either no lead or high
lead, a shorter "SCREEN" type measurement is sometimes used to reduce labor
time.
The results of the measurement(s), the length of the measurement(s), the sample ID
number, and the raw spectral data are all stored in the system's on board memory,
automatically. The operator cannot select or control which measurements are stored or
manipulate the data during transfer to a computer.
Q. Approximate analysis time.
A. User selected. Actual measurement time varies as a function of the source size and the
level of precision the operator chooses.
W1993L..2
-------
The instrument allows the operator to choose one of three levels of precision, which
equate to measurement time, which with a 40 mCi source, are: 15, 60, 240 seconds
(adjusted automatically by the instrument for source decay). The actual measurement
time is a combination of "SCREEN" and "TEST" type readings. Contractors report an
average of about 1 minute (includes source decay time) per sample point. This time
includes screen and test mode measurements and source decay.
Q. Interferences
A. With K-shell X-rays no known interferences exist. With L-shell X-rays interferences can
occur from zinc and copper.
Q. What substrate correction procedures are used?
A. The MAP LBP system comes with a general or "universal" substrate compensation
intrinsic in the system's software. Substrate correction is achieved through measurement
of the substrate effects during calibration. Proprietary algorithms installed in the
instrument during manufacture correct for substrate effects before the measurement result
is displayed by the instrument.
Q. Are there any new developments or guidelines for using substrate correction procedures?
A. Substrate characterization using NIST SRM's, dedicated calibrations, longer
measurements. We do not recommend using bare substrate measurements unless the
operator suspects there is lead in the substrate.
Q. Are there any built-in instrument checks to test for satisfactory operation?
A. Yes. During a 15 second warm up immediately after the instrument is turned on, the
system conducts a "self test" to determine if all system components are operating within
specification.
After any measurement the operator can view the actual spectrum of energies detected
and measured during the analysis. Operators are trained how to evaluate the spectrum
to determine if the instrument was functioning properly during the analysis.
Software used to download measurements verifies both measurement integrity and
instrument performance. The software analyzes the spectrum to determine measurement
validity (total counts, source strength, algorithm calculation, date, etc.). The software
also evaluates the calibration check readings, calculates their average and standard
7 W1993L.2
-------
deviation and compares the daily data with the same performance criteria for the
instrument when it was manufactured.
Measurements which do not meet valid criteria or which are taken with an instrument out
of calibration are not transferred from the console but rather are shown as "invalid."
Q. Briefly describe safety procedures/considerations.
A. 100% positive visual and manual safety shutter with locking ON-OFF key. A "passive-
OFF" or dead man OFF feature provides secondary shielding independent of the ON-
OFF key. The operator controlled key provides positive off. Since any/all safety
ultimately relies on the operator, emphasis is placed on operator training. Operators are
required to use dosimetry monitoring and keep observers ten feet from the XRF device
when in use.
Q. Are there any problems/limitations with the instrument at present?
A. Problems: Policy makers do not allow for proper or optimum use of the instrument.
Limitations'.
1. Calibrations are based on the use of "representative" substrate standards
which may not always represent every material encountered in the field.
2. Precision of measurements on curved, profiled, or contoured surfaces is
variable and unknown.
3. Instrument should not be used in extreme heat (> 110°F).
041993L.2
-------
Related Issues:
Some "influence centers" claim that a "limitation" of the instrument is that negative
readings are "truncated to zero." Therefore, results "must be biased high."
This claim is indicative of the lack of understanding of the practical implications of field
work, a lack of understanding of XRF technology, and a fundamental understanding of
statistics.
First of all, it is impossible to have less than zero mg/cm2 lead in a paint film, therefore,
negative readings are confusing to most inspectors. In order for an XRF operator to
fully understand a negative value they must be taught basic statistics. Few field workers
have the time, money, or inclination to learn statistics.
When the detection limit of one reading from an instrument is larger than the "target or
action level," negative numbers can occur, but are useful or meaningful only if an
averaging process is used, that is, if a series of repeated readings are taken on the
identical spot and averaged together (typical for direct readers but not for the Spectrum
Analyzer).
However, if the detection limit of one reading is less than the action level, and multiple
readings are not used, negative values are meaningless and add confusion to many. Since
a negative value is confusing to most people, and the action level is well above the
detection limit for the Spectrum Analyzer, we do not normally display a negative reading
because there is no need to do so.
If the "action level" was at or near zero (0 mg/cm2), or the precision of a group of
measurements at or near zero was being determined, indication of negative results would
be useful for the Spectrum Analyzer.
It is not a problem for us to display negative readings. We offer our customers the
option of having the instrument display negative values.
Q. Are improvements to the current instruments envisioned?
A. Yes. Scitec's policy is to not release new development information prematurely.
However, we can indicate that we spend a sizeable portion of our operating budget on
research and development, all directed at improving instrument performance and reducing
sources of bias-whether instrument or operator generated.
We recently introduced a new data download software program called "AcuTransfer0"
which checks and verifies measurement integrity and instrument performance during the
data download process.
M1W3L.2
-------
We also offer "Advanced Report Manager0" (ARM0) software which produces complete
HUD type reports and analysis.
These programs provide a complete field testing "system" that includes a number of
quality control features.
1. Minimization of operator bias or influence. All measurement data are
automatically recorded by the instrument. That is, the measurement
time/precision, test location, component, paint conditions, substrate, and
calibration checks are automatically recorded, analyzed, and reported by the
system. The operator cannot bias test data by selecting which measurements to
report.
2. Standardized testing and sample numbering protocols allow for fast and easy
(accurate) QC oversight during the data transfer and evaluation process.
3. Elimination of field notetaking errors.
4. Elimination of typing, transposing, and analytical errors.
Q. When might an improved instrument be available?
A. We regularly upgrade and enhance our product and software. The version we currently
offer is the second model and fifth overall version of the LBP system. All performance
related changes are systematically made to older versions during source replacement or
service. The current model is adequate for lead based paint testing, but current
guidelines do not allow for optimum use of its capabilities.
Q. What do you view as the future needs of the public and/or lead testing organizations
regarding portable XRF spectrometers?
A. 1. Training of housing authorities.
2. Clear, concise testing protocols/guidelines with logical objectives.
3. Uniform bidding information.
The major problem contractors have to deal with is the lack of knowledge the PHA has
concerning lead testing. Bid requests and requests for proposals are incomplete and
poorly written. This does not allow contractors to compete on an equal basis.
10
041993L.2
-------
Since the Guidelines do not require contractors to provide proof of testing and
compliance with protocols, contractors regularly cheat. Housing authorities are not
trained or equipped to deal with these problems.
Q. What topics should be added to the workshop agenda?
A. We suggest emphasis be placed on:
1. Ways of eliminating sources of errors and bias from:
a. Operator influence
b. Bid requests, contract awards, contract administration
c. Data handling, recording, reporting.
2. Optimizing the use of XRF for field testing. Current, and even the proposed
Guidelines do not allow for or provide incentives for use of the full capabilities
of XRF instruments. The Guidelines are written to accommodate the oldest form
of technology rather than the most advanced form of technology. Therefore HUD
receives poor quality data at very high total costs.
Attention should be given to:
a. Limited use of L-Shell XRF Analysis. This type of XRF analysis has
been overlooked, or not even allowed. However, we believe there is a
potential for the use of L-shell analysis in conjunction with K-shell
analysis.
L-line X-rays can be detected only if there is lead on the immediate
surface coats of paint (i.e., the top 1 or 2 layers). Because of this many
people think that L-shell analysis is not useful since most lead paint is on
the bottom, or lowest layers, and therefore only K-line X-rays are useful.
However, if lead is on the surface, 1) it is a more immediate potential
hazard and 2) it can be measured more accurately.
In some cases use of an L-shell reading can reduce the need for
confirmatory sampling by AA.
L-shell lead X-ray measurement provides lower detection limits and better
precision than K-shell lead X-ray measurement. This is due to the higher
yield achievable and the reduced effects from substrate conditions.
11 041993L.2
-------
We do not propose that L-shell measurement be a substitute for K-shell
measurement, but an allowable alternative in some cases...if there is
greater than 1.10 mg/cm2 of lead on the surface coat of paint the condition
is positive!
In addition, L-shell analysis can be used to help determine whether
detected lead is behind an outer substrate (e.g., a pipe, painted wall,
flashing, etc.), or in the substrate.
b. Use of short and/or long XRF readings. The current and proposed
Guidelines limit the usefulness of XRF by limiting the number of, or
length of readings allowed and fixing a single inconclusive range where
chip sampling/lab analysis are required. This policy unnecessarily
increases the total testing costs incurred by HUD.
With the Spectrum Analyzer, a longer, or "CONFIRM" type reading will
provide a measurement with twice the precision of the "TEST"
measurement which is specified by HUD. Use of a "CONFIRM" type
measurement only when the apparent lead concentration is at or near the
action level would reduce the amount of lab sampling by 50%.
On the other hand, allowing inspectors to use shorter "SCREEN" type
measurements when lead concentrations are at or near zero, or very high
(above 1.5 mg/cm2) will reduce field testing costs by 60%.
3. Operator, EPA, HUD Training. The current XRF operator training provided
during EPA courses by the universities is not focused on the most important field
issues-how to generate statistically valid test results. The training is often
provided by scholastic types who have little or no field experience, or by selected
contractors whose technical experience is limited, and sometimes biased.
Training often includes special interest sub-agendas as the EPA has no policy
regarding trainers having conflicts of interest.
We've found that public housing authorities and contract administrators lack a
fundamental understanding of XRF, sampling, and/or statistics.
Public housing authorities, for the most part, do not fully understand what it is
they are tasked to do, so their requests for proposals and/or requests for bids are
poorly worded, structured, and contain inadequate information for contractors to
respond to in a meaningful way. Many contractors do not respond to bid
invitations simply because of the lack of information.
12
OU993L.2
-------
Most often bids are awarded to the "low bidder," without regard to the technical
skills of the bidding firm or the total testing costs. By awarding contracts in this
way, public housing authorities and government agencies are providing
contractors with the means to rip them off. Many bids allow, even require
conflict of interests to take place!
Most contract administrators do not know how to evaluate, inspect or interpret
lead testing reports once they get them. Contractors can turn in just about
anything, call it a test report, swear to its authenticity (with several disclaimers
thrown in throughout the report) and get paid!
Standardized bid requests, evaluation methods and testing contracts are needed.
Fundamental training in how to administrate and supervise testing work is needed,
and standardized test report and data handling methods will reduce the burden on
housing agencies and improve the quality of the test data.
Most of all, testing contractors should provide proof that statistically valid
random testing was conducted, and conducted with an instrument which was
working properly.
4. Laboratory confirmation. HUD's Guidelines require laboratory confirmation of
all "inconclusive" and "positive" XRF test results in spite of an absence of any
study or data about the accuracy and quality of the chip sampling/lab analysis
method of sampling.
Our informal surveys and evaluation of test data from contractors and labs
indicates that the chip sampling/lab analysis method tends to produce results with
a low bias, thus providing "false negatives" on inconclusive XRF tests.
This method is expensive and very time consuming, and does not necessarily
provide suitable back-up or "confirmatory" defensible data.
The main issue here is not the ability of laboratory instruments and technicians
to accurately measure lead in paint films (although there is concern about this
from one lab to another), but the ability of the contractor to remove a good,
representative sample.
We recommend:
a. A study of the accuracy and cost of the chip sampling/lab analysis method
be conducted.
b. Quality control methods need to be developed for the chip samplingAab
analysis procedure.
c. Alternative methods of dealing with preliminary "inconclusive" XRF tests
be investigated, and/or allowed.
13 041993L.2
-------
We hope our response is helpful to you and will further your efforts to develop effective
evaluations of XRF instruments and guidelines for their use. We urge you to involve instrument
manufacturers in all processes; we're a source of expertise and experience! Everyone has their
biases, or preferences and prejudices, so you should at least hear and take all input. Scitec is
willing to participate openly in any and all discussions. We're looking forward to working with
you.
Sincerely,
Larry T. Lynott
President
encl: HAD
cc: Jim Baugh
Darren Small
14
i^ 041993L.2
-------
Data Presented by
TN Technologies, Inc.
-------
Portable XRF Spectrometer
Name or Model # of spectrometer Spectrace 9000 (Pb Analyzer)
Applications Several applications are provided for; with the primary one
being an optimized configuration for the in-situ measurement of Pb in
paint film via its "K" and "L " x-ray excitation. This is a mass per unit area
analysis which reports the values of both the PbL and PbK results on the
same screen. Other unit area analyses provided include an application for
dust wipes and filtered (air/water) particulates. An application for the
determination of Pb by mass concentration (in ppm. %, or other units) in
bulk materials such as soil, dust, water, etc. is also contained. The
analyses are not limited to Pb; elements such as Hg, As, and many other
metal pollutants can be quantified in the same measurement. This
response relates only to the Pb-paint application.
Detection Limit For a single 60 sec measurement the approximate MDLs
are as follows; « 0.002mgPb/cm2 for the PbL analysis on any substrate,
and in the range 0.04 to 0.12 mgPb/cm2 for the PbK analysis for
substrates ranging from thin wood to thick concrete. It will be ~ 0.06 on
gypsum-over-wood. On thick steel the PbK MDL will be = 0.2mgPb/cm2.
How is the detection limit determined? It is equated to THREE TIMES the
standard deviation (3o) in the results of a measurement (expressed in
calibrated concentration units) performed on a blank, or low metal content
paint film in contact with a defined substrate. (J-values may be derived
from the results of repetitive measurement (negative values can be
displayed or printed) or obtained directly from the analysis report. The
latter reports the o-value of each result according to the conditions of the
measurement (i.e. time, signal strength, source decay etc) and taking into
account the accumulative statistical errors of all aspects of the data
processing including the spectrum strip and the substrate correction
mathematics.
Sensitivity In terms of the net (less substrate) counts/sec/mgPb/cm2, the
sensitivity for PbL analysis is ~ 300, and that for the PbK is ~ 5. In units
of "per millicurie of exciting source radiation" the respective numbers
would be 60 and 5. A Cd-109 source with the emission equivalent of
5mCi x-rays and 1mCi gamma is used in the instrument.
-------
How is the sensitivity determined? By measurement on a series of known
mgPb paint films and calculation of the initial slope of the source-decay-
corrected (SDC) countrate/mgPb response. The NIST 2759 Pb standards
present a suitable range. SDC countrate values can be viewed on the
instrument screen.
Accuracy For a single 60 sec the accuracy of a PbK determination will
be ~ ±0.125mgPb/cm2 at a loading of 0.5mgPb/cm2 and
« ±0.16mgPb/cm2 at a loading of 1.0mbPb/cm2. The influence of paint
thickness on the PbL measurement makes it difficult to assess the
accuracy of that analysis.
How is the accuracy determined? From the magnitude of the difference
between the x-ray results and the "mass/area " chemical analyses of a
representative range of samples on a variety of substrates comprising
wood, plaster and concrete. The expected accuracies (listed) are based
on the 2-standard deviation criteria and include no allowance for
knowledge of the substrate. In reality the accuracy is slightly better on a
wood substrate and that can be evaluated on an individual results basis
since the statistical uncertainties are always reported with each
measurement. For example an accuracy of ~ 0.1mgPb could be expected
on wood.
Briefly describe the calibration procedure. The primary calibration
performed at the factory entails measurement of several pure elements,
various bare substrates and at least one NIST 2759 standard reference
sample. A secondary calibration using an equivalent of the substrate
material and a Pb reference sample can be performed in the field. All
calibration constants are retained in a non-volatile on-board memory. They
are also importable from a floppy disk.
Are there any plans to allow for direct user calibration or recalibration?
As indicated there is provision for user recalibration.
Are you aware of the new NIST paint film SRMs? Yes.
Have they been of any value to you in calibrating or testing your
instruments? Yes. They are used in the primary calibration and will be an
essential part of the calibration maintenance.
-------
Briefly describe the procedure for use of your instrument. Operation is
menu driven so all steps are shown on the screen. Measurement can
begin ~5min after turn-on. It is push-button initiated and program
controlled so that the source shutter will automatically close at the end
of a short period of measurement. The probe (of ~ 4 Ibwt) must be held
steady but there is no need to maintain any pressure against the surface.
The result's display of both PbK and PbL, together with their error level
is self-explanatory. A menu choice will automatically store the results, as
well as the spectrum if desired. Prior to beginning a series of
measurements the operator will perform a check for ~ WOsec. If an error
condition is detected the screen will direct the operator to perform an
adjustment. Several automatic checks, such as the spectrum calibration,
are also performed during measurement and the operator is alerted to any
problem. The need for battery renewal will be indicated with enough
warning to protect stored results.
Approximate analysis time The time (in real clock seconds) is user
selectable from 1 sec up. Recommended time is ~ 15 sees for an initial
check. This should provide for a reliable (95% confidence) x-ray based
assessment at the 1mgPb level on all results except for those in an
"uncertain" zone of 0.75 to 1.25 mgPb. Results within that zone will
require additional time.
Interferences The main interference on the analysis is that of the
substrate scattered radiation in the spectral window of the Pb x-rays.
Other element fluorescent x-rays are excluded by the high resolution of
the x-ray detector.
\
What substrate correction procedures are used? The spectrum overlap
from the substrate is relatively small, due to the choice of source and
good detector resolution. It must stiff be corrected for however and this
is accomplished by a special stripping algorithm that is applied to each
spectrum as it is acquired in the measurement. The coefficients of the
stripping algorithm are maintainable by the operator using a set of
substrate-equivalent standards.
Are there any new developments or guidelines for using substrate
correction procedures? Describe. None are known at this time however
-------
there is a/ways room for improvement in this area. The ability to store the
entire, high resolution spectrum for further evaluation should prove useful
in this direction.
Are there any built-in instrument checks to test for satisfactory operation?
Describe. Yes as described earlier. The spectrum is continuously checked
for gain calibration and the operator has a CHECK procedure to conduct
periodically. Other continuous checks are made on the signal quality and
any malfunction will be indicated on the screen.
Briefly describe safety procedures/considerations. The main concern must
always be over the chance of accidental exposure to the source. On the
9000 the source gamma ray emission is only 1/35 of that used in other
spectrum analyzers and is also of lower gamma ray energy. Adequate
shielding is built-in for the source closed position.
Are there any problems/limitations with the instrument at present? The
probe face may be considered wide f~3") for some situations.
Are improvements to the current instruments envisioned? Yes. For
example more efficient larger detectors could become available to allow
even lower activity source usage or shorter measurement times. A smaller
footprint, lighter probe may also be possible in the near future.
When might an improved instrument be available? Some improvements
could be achieved within a 12 month time frame.
What improvements will be implemented? Possibly a lighter, smaller
footprint measurement probe
What do you view as the future needs of the public and/or lead testing
organizations regarding portable XRF spectrometers? Better procedures
for sampling in regard to XRF instrument verification of the mass/unit
analysis and a resolution of the % wt versus mgPb dilemma in the current
code.
What topics should be added to the XRF workshop agenda? Other
analytical techniques. Standardization progress reports of other
committees. Opinions on the "% " vs "mg" dilemma.
-------
Please provide any standard operation procedures, instrumentations or
other literature you believe would be informative. Copies of the standard
instrument manual were provided. The attached pictorial comparison of
current PbK analysis based instruments shows a wide range of values of
the "net signal to background" parameter as a result of the source choice
and the system spectral resolution. A high signal to background reduces
the effect of the background variation, i.e. the substrate problem, and
allows one to expect accuracies that are limited only by the statistical
factors imposed by the analysis time.
-------
SCINTILLATION
plus FILTERS
GAS
PROPORTIONAL
T I
~iIII " I
111
SILICON
90
ENERGY (keV)
110
Comparison of "NET-to-BACKGROUND" Ratio at
15mgPb/cm2 on Wood for Various Detectors
ENERGY (keV)
110
-------
Data Presented by
Princeton Gamma-Tech
-------
Portable XRF Spectrometer
Name or Model # of spectrometer PCT Model XK-3
'Ptaeae reproduce this form {f other rtwf&s are available.
APPficaik>n& Painted surfaces
(Paint, Soil, Dust, etc.)
Detection Limit 0.5 mg FWtac2 How is the detection Rmit
determined? It is taken as the single-reading standard deviation of a series
of readings.
Sensitivity 0.5 majpb/ca2 How is the sensitivity
determined? It la taken as the single-reading standard deviation of
a. aeries of readincfi.
Accuracy How is the accuracy
determined? Accuracy is determined by comparing instrument readings with
U.S. Pqpt. of HUD lead standards of CL60. \t'ft pftd 2.99 mp/r-m2 ^.traceable
to HBS standards. ROTE: Precision can be increased by averaging 9. number of
readings.
Briefly describe the Cafbratioa procedure A series of IP readings are taken on
0 and 1.53 Bg/ca^ HUD standards. Internal adjustments are Chen ma'de to brine
the unit into calibration and the procedure repeated until both readings
vlthln -f/- 0.1
Are there any plans to allow for direct user caffbration or recafibration? NO.
-------
Are you aware of the new NIST pair* film SRMs? Yes
Have they been of any value to you in cafibrating or testing your instalments?.
No. We were lucky enoagb to have old HDD standards.
Briefly describe the procedure for use of your instalment After checking calibration and
stabilisation (varmup), the face of the XK-3 is placed against the surface to be testet
and the handle pushed in_to. the stop. The hoodie ia held infirmly until a reading ->
appears on the display, signaled by an audible beep. The reading remains displayed
uctll a new ateagareaent is Initiated. We attach an Operator'a Maonal for your
further information.
Approximate analysis time About 12 seconds for a fresh source.
Interferences Only backscattered gamma rays. __
What SUbStraie Correction procedures are Used? Manual substrate corrections
arc -oa.de, in accordance vlth the recommendations of the BUD guidelines.
Are their any new developments or guidelines for using substrate correction procedures?
Describe.^ »o.
Are there any built-in instalment checks to test for satisfactory operation? Describe,
Only the calibration check block provided with the instrument.
-------
Briefly describe safety proceduresAjonskterafions The XK-3. is unique, its radiation
source shutter cannot be opened unless the face of the XK-3 is against a solid surface.
This prevents anyone from being exposad by the gamma-ray beam, which is possible with
other available inscrircenta. Specific safety practices are covered In detail la our
User School, attendance at which Is strongly recommended for anyone using the instrunei
Are there any problems/Hmftations with the instrument at present?
See cqyer letter.
Are improvements to the current instruments envisioned?.
Sae cover letter.
When might an improved instrument be available?.
See cover letter.
What improvements will be implemented?
See cover letter.
What do you view as the future needs of the pubic and/or lead testing organizations
regarding portable XRF spectrometers?
See cover letter.
What topics should be added to the XRF workshop agenda?
See cover letter. •
-------
Pfease provide any standard operating procedures, Instrumentations or other literature
you believe would be informative. __
attached
** TOTAL PAGE.006 **
------- |