Peer Review of the U.S. Environmental Protection Agency's "Final
Report on the World Trade Center (WTC) Dust Screening Study"
Final Report
October 2005
Prepared for:
Office of Research and Development
U.S. Environmental Protection Agency
401 M Street, SW
Washington, DC 20460
EPA Contract No. 68-C-02-060
Task Order 107
Prepared by:
Dan Crane, Occupational Safety and Health Administration
Frank Ehrenfeld, International Asbestos Testing Laboratories
Mickey Gunter, University of Idaho
Ernest McConnell, ToxPath, Inc.
Shu-Chun Su, Hercules, Inc.
James Webber, New York State Department of Health
John Wilhelmi, Eastern Research Group, Inc.
-------
Peer Review Summary Report WTC Dust Screening Study
Note
This report was prepared by six peer reviewers and Eastern Research Group, Inc (ERG), under
contract to the U S. Environmental Protection Agency (EPA) (Contract No 68-C-02-060, Task
Order 107) This report summarizes discussions from a conference call during which the six peer
reviewers responded to nine charge questions regarding EPA's "Final Report on the World Trade
Center (WTC) Dust Screening Study " This report summarizes key points raised during the
conference call This report does not contain a verbatim transcript of all issues discussed during
the conference call, nor does it embellish, interpret or enlarge upon matters that were incomplete
or unclear. Except as specifically noted, no statements in this report represent analyses by or
positions of EPA or ERG
-------
Peer Review Summary Report WTC Dust Screening Study
Table of Contents
Note i
List of Abbreviations iv
Executive Summary 1
1.0 Introduction 3
1.1 Background 3
1.2 Scope of the Peer Review 4
1.3 Report Organization 6
2.0 Selection of a Signature 7
2.1 Charge Question 1: Selection Criteria 7
2.2 Charge Question 2: Slag Wool as a Signature 8
3.0 The Proposed Analytical Method 10
3.1 Charge Question 3: Detail of Proposed Analytical Method 10
3.2 Charge Question 4: Suggested Improvements to the Proposed Analytical Method 11
4.0 Data Analysis and Interpretation 12
4.1 Charge Question 5: Study Design 12
4.2 Charge Question 6: Statistical Analysis of Data 13
4.3 Charge Question 7: Data Interpretation 15
5.0 General Considerations and Alternate Approaches 16
5.1 Charge Question 8: Suggested Improvements 16
5.2 Charge Question 9: Other Comments 18
6.0 References 21
Appendix A. List of Peer Reviewers
Appendix B. Charge to Reviewers and Preliminary Written Comments
Appendix C. Responses to Questions of Clarification
Mi
-------
Peer Review Summary Report WTC Dust Screening Study
List of Abbreviations
COPC contaminant of potential concern
EDS energy dispersive spectrometry
EPA U S Environmental Protection Agency
ERG Eastern Research Group, Inc
NAIMA North American Insulation Manufacturers Association
PLM polarized light microscopy
PM participate matter
SEM scanning electron microscopy
TEM transmission electron microscopy
TIMA Thermal Insulation Management Association
USGS U S Geological Survey
WTC World Trade Center
iv
-------
Peer Review Summary Report WTC Dust Screening Study
Executive Summary
This report summarizes an independent peer review of the U S Environmental Protection
Agency's (EPA's) "Final Report on the World Trade Center (WTC) Dust Screening Study"
(EPA 2005a) In this study, EPA developed an analytical method to screen bulk dust samples for
mineral slag wool, particles consistent with concrete compositions, and gypsum (EPA 2005b)
The study included a method validation component, in which spiked background dust samples
with varying levels of WTC dust were prepared and then eight laboratories measured
concentrations of candidate WTC signatures in spiked and non-spiked samples Based on the
data collected during this study, EPA proposed slag wool as a signature constituent of WTC dust.
Six expert peer reviewers with various affiliations and from relevant scientific disciplines were
asked to provide an independent peer review of the WTC Dust Screening Study The peer
reviewers prepared preliminary written comments on the study and further discussed these
comments during a 4-hour conference call on October 4, 2005 The peer review of the WTC
Dust Screening Study focused on nine charge questions, which asked the peer reviewers to
comment on EPA's proposed analytical method, data analysis and interpretation, and selection of
slag wool as a signature for WTC dusts
Following are the peer reviewers' main findings, organized by topic The remainder of this report
documents the discussions among the reviewers that led up to these findings.
Proposed Analytical Method. The peer reviewers had concerns about laboratories'
abilities to implement the proposed analytical method consistently or correctly, given that
three out of the eight laboratories selected to participate in the method validation study
failed to produce data of acceptable quality. Even after representatives from the eight
laboratories "attended a 2-day session during which the method was further developed
and discussed" and "all laboratory participants held weekly conference calls as the
analytical program was proceeding to discuss general issues with the protocol" (EPA
2005a), data had to be massaged to differentiate WTC dust from background dust The
peer reviewers recommended several improvements to the proposed analytical method
and recommended that EPA establish strict criteria for identifying and enabling
"qualifying laboratories," should EPA decide to move forward with slag wool as a WTC
signature Refer to the summary statements in Sections 3 and 5 of this report for the peer
reviewers' specific recommendations for improving and implementing the proposed
analytical method
Data Analysis and Interpretation The peer reviewers were skeptical that EPA's
evaluation and interpretation of the study data were performed fairly Peer reviewers
pointed to several non-standard steps taken to enhance the study's ability to distinguish
WTC dust from background dust and noted that these steps could be interpreted as
attempts to prove the method's success rather than to objectively evaluate its real-world
potential for fingerprinting WTC dust The peer reviewers' two most notable concerns
were
o The reported difference in slag wool levels for impacted and non-impacted
locations was based on statistical analyses that excluded certain results from
-------
Peer Review Summary Report WTC Dust Screening Study
background samples but did not exclude results from spiked samples The
background results and spiked sample results were statistically indistinguishable
when the entire data set was considered
o The study authors disqualified three out of the eight original selected laboratories
during the screening study There was no follow-up investigation to determine
why these selected laboratories, presumably following the proposed method,
failed to differentiate background dusts from WTC dusts [After the peer review
conference call, one peer reviewer noted that a possibility exists that the three
laboratories were wrongly disqualified due to errors in an equation used to
calculate the numbers of slag wool fibers per gram of dust. Section 4 2 describes
this further.]
Refer to Section 4 of this report for more detailed information on the peer reviewers'
comments regarding data analysis and interpretation.
Selection of a WTC Signature. The peer reviewers supported EPA's conclusion that
gypsum and elements consistent with concrete do not meet the WTC signature selection
catena Regarding slag wool, the peer reviewers agreed that, from the data provided,
EPA has not made the case that its proposed analytical method can reliably discriminate
background dust from dust contaminated with WTC residue Thus, the proposed method
has not demonstrated the utility of slag wool as a successful signature constituent. This
finding was based on critical reservations, stated above, regarding the proposed analytical
method and the data analysis and interpretation Section 2 of this report further
summarizes the peer reviewers' discussions on this topic.
General Considerations and Alternate Approaches At the end of their conference call,
the peer reviewers offered several recommendations to EPA for identifying a WTC
signature The reviewers classified some of these recommendations as modifications to
the proposed analytical method, such as reconsidering the utility of polarized light
microscopy (PLM), using fiber dimensions (in addition to fiber counts) to "fingerprint"
dusts, and implementing a tiered sampling approach that would use slag wool as a
screening marker for the potential presence of WTC dust followed by transmission
electron microscopy (TEM) analysis of chrysotile asbestos as a confirmatory marker The
reviewers acknowledged that further evaluation would be necessary to assess the utility
of these and other suggested modifications The peer reviewers classified their remaining
recommendations as entirely new approaches that could only be investigated through new
research projects Section 5 of this report lists all of the general considerations and
alternate approaches the peer reviewers discussed during the conference call
The remainder of this report presents additional information on the independent peer review of
the WTC Dust Screening Study Section 1 presents a detailed account of the peer review process,
Sections 2 to 5 summarize how peer reviewers responded to the nine charge questions during the
peer review conference call, and Appendix B includes copies of the peer reviewers' preliminary
written comments
-------
Peer Review Summary Report WTC Dust Screening Study
1.0 Introduction
This report summarizes six experts' independent peer review of the U S Environmental
Protection Agency's (EPA's) "Final Report on the World Trade Center (WTC) Dust Screening
Study" (EPA 2005a) This report refers to the review document as the WTC Dust Screening
Study
Eastern Research Group, Inc (ERG), under contract to EPA, organized and implemented this
independent peer review according to procedures outlined in EPA's "Peer Review Handbook"
(EPA 2000) Between August 19, 2005 and September 23, 2005, the six peer reviewers prepared
preliminary written comments on the WTC Dust Screening Study On October 4, 2005, the peer
reviewers participated in a 4-hour conference call to discuss their preliminary written comments
and to develop summary statements on several key issues.
The six peer reviewers and a technical writer from ERG wrote all of the information that appears
in this summary report, except for Section 1 1, which comes almost entirely from EPA's charge
to the peer reviewers This introductory section provides background information on the WTC
Dust Screening Study (Section 1 1), describes the scope of the peer review (Section 1 2), and
outlines the organization of this report (Section 1 3).
1.1 Background
In the days following the terrorist attack on New York City's WTC towers, EPA, other federal
agencies, and New York City and New York State public health and environmental authorities
initiated numerous air monitoring activities to better understand the ongoing impact of emissions
to the outside environment from that disaster In 2002, EPA Region 2 turned its attention to the
indoor environment, providing a volunteer "clean and test" or "test only" program for residents
and homeowners who wished to have their apartments cleaned or tested. Asbestos was selected
as a surrogate for the presence of WTC-related contamination for this program
In March 2004, EPA convened the WTC Expert Technical Review Panel to interact with EPA
and the public on plans to monitor for the presence of any remaining WTC dust in indoor
environments For more information on the WTC Expert Technical Review Panel, refer to the
website, www epa gov/wtc/panei Approximately 750 units in an area extending north to
Houston Street in lower Manhattan and across the East River into a portion of Brooklyn were
proposed to be sampled for contaminants of potential concern (COPC), as well as for specific
constituents that can be used as markers to identify residual contamination by dust from the
WTC collapse. These "WTC dust signature" constituents are critical to the sampling program, as
they will provide the basis for estimating the geographic extent of the remaining residue in dust
of the WTC collapse. Using the results of dust sampling for the WTC collapse, EPA will decide
whether indoor cleanup or other activities are warranted
Based on previous work of the United States Geological Survey (USGS) and others, EPA
identified three components of the dust generated by the WTC collapse that could be used to
screen sampled dust for the presence of WTC dust These markers, or signature components,
were mineral slag wool, gypsum, and elements of concrete EPA developed the following
working hypothesis for the signature "A dust sample that contains WTC dust will have slag
-------
Peer Review Summary Report WTC Dust Screening Study
wool and elements of concrete and gypsum present in 'significant quantities' when compared to
typical indoor urban dust." Experts from the USGS, EPA's Office of Research and Development,
EPA's National Enforcement Investigations Center, and the commercial testing laboratory
community worked together to develop an analytical method to quantify the concentration of
these three markers in indoor dust
Between September 2004 and April 2005, numerous samples were taken in impacted buildings
near Ground Zero and at background locations These samples have been analyzed for these
three markers to determine whether they validly constitute a WTC signature, as suggested by
earlier USGS efforts. Also, the samples were used in a method validation study whose primary
purposes were twofold- (1) to evaluate the analytical method with regard to method variability
(both inter- and intra-laboratory variability) and cost and expediency and (2) to assist in the
determination of the lower limits of concentrations of the markers that could be reliably
measured and that could be reliably distinguished from background concentrations This method
validation study involved five contract laboratories and three federal government laboratories
This study entailed spiking known background dust with varying concentrations of known WTC
collapse dust, and then having the laboratories perform a blinded analysis on both spiked
samples and background samples
EPA's study is documented more fully in the WTC Dust Screening Study, which was the subject
of this peer review EPA contracted with ERG to implement an independent peer review of the
study. The remainder of this report documents this peer review More information on EPA's
response to environmental issues associated with the WTC collapse can be found on the
following website www epa gov/wtc
1.2 Scope of the Peer Review
ERG managed every aspect of this peer review, including selecting reviewers and coordinating
activities before, during, and after the peer review conference call The following sections
describe what each of these tasks entailed
1.2.1 Selecting the Peer Reviewers
ERG selected peer reviewers that met the selection catena EPA specified in its task order for this
project Those criteria noted that "The expertise that is desired include, but is not limited to,
these fields of study: polarized light microscopy (PLM), scanning electron microscopy (SEM),
transmission electron microscopy (TEM), paniculate matter (PM) sampling methods, quality
assurance procedures, analytical method development, and validation procedures "
Based on these selection criteria, ERG developed a list of candidates for this peer review Many
candidates were identified by an ERG search for subject matter experts Additional candidates
were identified through a peer reviewer nomination process In July, 2005, EPA announced the
procedure by which any interested party could nominate candidates for this peer review project
ERG contacted every nominee that was received Through these two search efforts, ERG
identified more than 50 highly qualified candidate peer reviewers ERG asked all candidates to
submit information on their expertise, availability, and potential conflicts of interest For this
peer review, a conflict of interest was defined as a situation in which an individual's activities,
-------
Peer Review Summary Report WTC Dust Screening Study
interests, or relationships create a situation where the candidate may benefit from the outcome of
the review
After carefully reviewing the candidates' expertise and credentials, ERG selected six peer
reviewers who, as a group, met the selection criteria and were able to provide a fair, independent,
and scientifically rigorous peer review of the WTC Dust Screening Study Appendix A lists the
peer reviewers' names and affiliations Recognizing that few individuals specialize in every
scientific discipline listed in the selection criteria, ERG ensured that the collective expertise of
the selected peer reviewers adequately covers the criteria (e g, at least one reviewer has
expertise in PLM, at least one reviewer has expertise in method validation studies, and so on).
ERG selected peer reviewers of varying affiliations (e g , academia, consulting companies,
analytical laboratories, government agencies) in hope that the peer reviewers would offer a
balanced perspective on the WTC Dust Screening Study
1.2.2 Activities Prior to the Peer Review Conference Call
Major activities that ERG and the peer reviewers conducted prior to the peer review conference
call follow
Distribute peer review documents and background information On August 19, 2005,
ERG sent the six peer reviewers packages that contained three primary review documents
and four background documents The primary review documents were the WTC Dust
Screening Study (EPA 2005a), the quality assurance project plan for the study (EPA
2005c), and written guidelines for the peer review These guidelines commonly called
a "charge" were nine questions that addressed various aspects of the WTC Dust
Screening Study, including an open-ended question that invited the peer reviewers to
comment on any topics that the other questions did not explicitly address A copy of the
charge is included in this report as part of Appendix B.
The packages ERG sent to the peer reviewers also included four documents with
background information These documents were distributed to give peer reviewers
additional insights on the previous characterization of WTC dusts These background
documents included three publications by USGS scientists (Meeker et al 2001; Lowers et
al 2005, Clark et al 2005) and an external review draft of EPA's proposed sampling
program to determine the extent of WTC impacts to indoor environments (EPA 2004)
The peer reviewers were also directed to the website for the WTC Expert Technical
Review Panel for online access to many additional documents that contain relevant
background information
Facilitate questions of clarification The peer reviewers worked independently for 5
weeks to review the WTC Dust Screening Study and to prepare preliminary written
responses to the nine charge questions During this time, ERG asked the peer reviewers to
refrain from discussing the scientific ment of the WTC Dust Screening Study with EPA
or any other party. However, ERG recognized that peer reviewers might have questions
of clarification about the study during the review process. ERG asked the peer reviewers
to forward all such questions in writing to ERG. ERG then sent these questions to EPA
for written responses By this approach, all of the reviewers' questions of clarification
-------
Peer Review Summary Report WTC Dust Screening Study
and all of EPA's written responses have been documented. Copies of the questions and
answers are included in this report as Appendix C
Obtain and compile the peer reviewers' preliminary written comments The peer
reviewers' preliminary written comments were due to ERG on September 23, 2005 After
receiving the complete set of comments, ERG then compiled, bound, and distributed
these comments to the six peer reviewers The peer reviewers' preliminary written
comments are included in this report, without modification, as Appendix B It should be
noted that'these written comments should be considered preliminary and some reviewers'
technical findings might have changed based on discussions during the conference call
The preliminary written comments do not necessarily reflect the reviewers' final
opinions
1.2.3 A ctivities During the Peer Review Conference Call
After the peer reviewers had ample time to read through the complete set of preliminary written
comments, the peer reviewers participated in a 4-hour conference call to discuss the comments
and to prepare summary statements of their findings The conference call occurred on October 4,
2005, and was attended by the six peer reviewers and a technical writer from ERG. The technical
discussions were exclusively among the six peer reviewers One of the peer reviewers (Dr James
Webber) served as the technical chair of the call During the conference call, the peer reviewers
discussed their answers to the nine charge questions
1.2.4 Activities Following the Peer Review Conference Call
After the peer review conference call, the ERG technical writer who attended the conference call
prepared a draft of this peer review summary report, which draws largely from summary
statements that the peer reviewers agreed upon during the conference call ERG distributed this
draft to the six peer reviewers and asked them to verify that it accurately reflects the tone and
content of the discussions during the conference call After every peer reviewer confirmed that
the draft summary report was a faithful account of the reviewers' discussions, ERG submitted
the final peer review summary report (i.e, this report) to EPA
1.3 Report Organization
The structure of this report reflects the main topic areas covered in the charge to the reviewers
selection of a signature (Section 2, charge questions 1 and 2), the proposed analytical method
(Section 3, charge questions 3 and 4), data analysis and interpretation (Section 4, charge
questions 5, 6, and 7), and general considerations and alternate approaches (Section 5, charge
questions 8 and 9) All references cited in the text are presented in Section 6
-------
Peer Review Summary Report WTC Dust Screening Study
2.0 Selection of a Signature
The first two charge questions addressed the selection of signature constituents of WTC dusts.
This section presents the peer reviewers' general summary statements in response to these
questions, along with a record of discussion of associated issues raised during the peer review
conference call
2.1 Charge Question 1: Selection Criteria
The first charge question to the peer reviewers was' "The following criteria were established to
assist EPA in the selection of appropriate constituents in dust to be characterized as WTC
signature constituents
a) They are present at levels unique to WTC dust (distinct from urban dust)
b) They are persistent for many months (not volatile)
c) They are sufficiently homogeneous in WTC dust
d) Available analytical methods are able to detect these screening materials with a small
sample size, low minimum detection limit, and low interference from other dust
components
Based on information in the supplied documents, and any other knowledge you may bring to the
table, are these criteria adequate for establishing a WTC signature marker'' If the answer is no,
please elaborate?"
Peer Reviewers' Summary Statement:
The peer reviewers found the four criteria that EPA used to select signature constituents to be
reasonable The peer reviewers noted that the fourth criterion is vague (e g, what precisely is
meant by a "small" sample and a "low" detection limit'') and listed two additional criteria that
EPA should have considered, as documented below
Record of Discussion:
Following is a summary of specific issues that the peer reviewers discussed during the peer
review conference call when responding to charge question 1
" Comments on criterion that signature constituents be "sufficiently homogeneous in WTC
dust " Two peer reviewers noted that existing data show that WTC dust samples are not
homogeneous, especially when comparing dust samples collected indoors to those
collected outdoors, dust samples collected at street level to those collected at elevation,
and dust samples collected near Ground Zero to those collected further downwind. These
reviewers acknowledged that this non-homogeneity complicates efforts to select
signature constituents A peer reviewer added that many candidate constituents (e g., slag
wool) might not be homogeneous in terms of their chemical and physical properties an
observation that led to the following two recommendations
-------
Peer Review Summary Report WTC Dust Screening Study
Recommended additional criterion that constituents have clearly defined chemical and
physical properties. A peer reviewer questioned whether the proposed analytical method
could reliably distinguish slag wool from rock wool based solely on the iron content
quantified using SEM (Section 5 of this report discusses this issue further) Based on this
concern, the reviewer recommended that an additional criterion for selecting WTC
signatures should have been considered Variations in signature constituents' chemical
and physical properties must fall within boundaries that can be clearly defined and
reliably differentiated from those of potential interferences at a pre-determmed
confidence level
Recommended additional criterion that standards be available for gauging measurement
accuracy Noting that the WTC Dust Screening Study had no provision for assessing
whether laboratories are reliably applying analytical methods to measure target
constituents at a pre-determmed confidence level, a peer reviewer recommended that an
additional criterion for selecting WTC signatures should have been Standards of various
concentrations can be created using the pure signature constituents for assessing if
analytical results can meet a pre-determmed measurement quality objective (e g , +/- 30%
accuracy for SEM analyses)
When discussing the use of standards, one peer reviewer noted that standard samples for
slag wool are likely available from the North American Insulation Manufacturers
Association (NAIMA) Other peer reviewers supported acquiring standard samples from
NAIMA, but provided that the standards have chemical and physical properties that are
representative of the slag wool found in WTC dust One peer reviewer noted a potential
challenge associated with acquiring a single representative slag wool standard1 because
the WTC towers were constructed over a relatively long time frame, slag wool found in
the insulation, ceiling tiles, and other materials likely have varying chemical and physical
properties Nonetheless, the peer reviewers unanimously agreed that standards are needed
to assess the accuracy of laboratories' measurements
2.2 Charge Question 2: Slag Wool as a Signature
The second charge question asked the peer reviewers "Based on information in the supplied
documents, and any other knowledge you may bring to the table, do you agree or disagree with
the conclusion that slag wool alone meets the criteria as a WTC marker? Please explain your
answer "
Peer Reviewers' Summary Statement:
The peer reviewers supported EPA's conclusion that gypsum and elements consistent with
concrete do not meet the WTC signature selection criteria Regarding slag wool, the peer
reviewers agreed that, from the data provided, EPA has not made the case that its proposed
analytical method can reliably discriminate background dust from dust contaminated with WTC
residue Thus, the proposed method has not demonstrated the utility of slag wool as a successful
signature constituent Critical reservations about selecting slag wool as a signature include the
following
8
-------
Peer Review Summary Report WTC Dust Screening Study
Based on the distributions of background sampling results and spiked sampling results,
several peer reviewers concluded that "false positives" will likely occur, especially when
samples are collected near known or unknown sources of slag wool, a common building
material
Three of the eight original laboratories invited to participate in this study were eliminated
from this study This removal of almost 40% of the selected laboratories brings into
question the robustness and reproducibihty of the method.
Some reviewers were concerned about possible degradation (or even disappearance) of
slag wool's characteristics in damp environments
Record of Discussion:
When answering this charge question during the conference call, the peer reviewers addressed
several relevant issues Some of these issues are described below and others were discussed in
the peer reviewers' responses to other charge questions
Challenges associated with selecting a signature When responding to this and other
charge questions, the peer reviewers identified several significant challenges associated
with selecting a WTC signature. For instance, because some materials used to construct
the WTC apparently were not considerably different from those used to construct many
other buildings in Lower Manhattan, many candidate constituents are ubiquitous in
background dust and therefore not unique to the WTC. Further, a peer reviewer noted
efforts to find constituents that are "sufficiently homogeneous" are confounded by the
fact that composition of indoor dusts associated with the WTC collapse is expected to
vary with numerous factors, including distance from Ground Zero, elevation, and
building configuration Finally, as years continue to pass since the terrorist attack, WTC
dust that previously entered indoor environments will continue to be diluted, thus
becoming more difficult to identify
Concerns regarding false positives Because slag wool is a common building material
(i e, not unique to the WTC), multiple peer reviewers' preliminary written comments
note that slag wool in dusts will likely be found in buildings with ceiling tiles or exposed
insulation and at locations near significant building construction, renovation, or
demolition activities Accordingly, this peer reviewer questioned how well slag wool
meets the first WTC signature selection criterion for uniqueness
Questions regarding the persistence of slag wool. Most peer reviewers noted in their
preliminary written comments that slag wool meets the second signature selection
criterion of being "persistent for many months (not volatile)." Citing studies that have
shown slag wool to be somewhat soluble in saline and biological environments (e g,
Bernstein et al 1996, IARC 2002), one reviewer questioned whether slag wool met this
second criterion. This reviewer noted that EPA could conduct solubility tests to verify
whether slag wool is indeed sufficiently persistent in damp environments.
After discussing the issues listed above, as well as those listed in the other charge questions, the
peer reviewers debated how to phrase their summary statement on the utility of slag wool as a
-------
Peer Review Summary Report WTC Dust Screening Study
WTC signature. The peer reviewers eventually agreed that the information they were provided in
the WTC Dust Screening Study did not convincingly establish that slag wool met EPA's four
criteria for reliable signatures, however, they also concluded that the information provided was
insufficient to reject slag wool as a signature After this discussion, the peer reviewers agreed
that the method validation study and the proposed analytical method, in its current form, have
not yet demonstrated the utility of slag wool as a successful marker Refer to Sections 3 and 5 for
the peer reviewers' specific recommendations for improving the proposed method
3.0 The Proposed Analytical Method
Charge questions 3 and 4 asked the peer reviewers to comment on the level of detail in the
proposed analytical method and to identify necessary improvements. This section presents the
peer reviewers' general summary statements in response to these questions, along with a record
of discussion of associated issues raised during the peer review conference call
3.1 Charge Question 3: Detail of Proposed Analytical Method
The third charge question asked the peer reviewers "Are the analytical methods written in
sufficient detail such that a qualifying laboratory could follow the methodology and obtain valid
results without supplemental assistance from EPA or other sources'?"
Peer Reviewers' Summary Statement:
The peer reviewers had concerns about laboratories' abilities to implement the proposed
analytical method consistently or correctly, given that three out of the eight laboratories selected
to participate in the method validation study failed to produce data of acceptable quality Even
after representatives from the eight laboratories "attended a 2-day session during which the
method was further developed and discussed" and "all laboratory participants held weekly
conference calls as the analytical program was proceeding to discuss general issues with the
protocol" (EPA 2005a), data had to be massaged to differentiate WTC spiked samples from
background dust. Accordingly, the peer reviewers agreed that a critical element to the success of
future sampling is the criteria EPA eventually establishes for identifying and enabling
"qualifying laboratories " The peer reviewers recommended that these criteria might include:
Intensive hands-on training of analysts
Close monitoring of performance by extensive use of duplicates, spikes, and standards
Creation of "standards" of slag wool from indoor WTC dust and other sources
Record of Discussion:
The peer reviewers briefly discussed charge question 3 during the conference call They pointed
to the fact that three out of the eight laboratones failed to generate data of acceptable quality as
ample evidence that the analytical method is not sufficiently detailed or robust to enable
qualifying laboratories "to obtain valid results without supplemental assistance from EPA or
10
-------
Peer Review Summary Report WTC Dust Screening Study
other sources " One reviewer added that failure of these laboratories to generate quality data calls
into question whether EPA met one of its initial study objectives " that the analytical method
works well enough and is able to be carried out by enough analytical laboratories to . distinguish
WTC dust from background dust" (see page 9 of the WTC Dust Screening Study)
3.2 Charge Question 4: Suggested Improvements to the Proposed Analytical Method
The fourth charge question asked the peer reviewers: "If the answer to Question 3 is no, what
items or information could be added so that a qualified laboratory could follow this protocol and
obtain valid results without the assistance of EPA or other sources''"
Peer Reviewers' Summary Statement:
The peer reviewers concluded that the method, as written, could not be used to differentiate
WTC dusts from other dusts and recommended several modifications to the proposed method
that might make it more robust. The most important improvements identified (and agreed to by
all six peer reviewers) were
Eliminate the drop mount
The iron content catena used to differentiate the signature constituent slag wool from the
primary interfering constituent rock wool should be based on detailed and comprehensive
analysis of respective components in WTC dust and background dust samples, and not on
TIMA's general product information
Eliminate sieving and settling from sample preparation because these steps cause the loss
of large portions of particles
Measure fiber dimensions (see Section 5 for further discussion)
Provide more exacting detail on aliquot withdrawal protocols
Prepare all samples at a central laboratory for distribution to other laboratories
Generate spikes and standards using indoor, rather than outdoor, WTC dust
Require appropriate use of significant digits
Describe how gravimetric-reduction data will be interpreted
Record of Discussion:
The peer reviewers referred to their preliminary written comments on charge question 4 for the
reasons why they recommended the aforementioned improvements Additional items discussed
during the conference call were the need for explicit sample rejection criteria. For instance, a
reviewer noted that the analytical method could specify minimum sample sizes, maximum
moisture contents, or other criteria that, if met, would require laboratories to reject samples
11
-------
Peer Review Summary Report WTC Dust Screening Study
Another issue raised during this discussion was whether electron microscopic analyses were even
necessary, as several peer reviewers noted that PLM alone might be sufficient in this screening
study without the added expense of SEM analyses The peer reviewers revisited this issue when
discussing alternate approaches to the WTC Dust Screening Study (see Section 5).
4.0 Data Analysis and Interpretation
The charge to the reviewers included three questions that ask about the statistical analyses and
interpretations of data collected during the method validation study This section presents the
peer reviewers' responses to these charge questions, organized into five topics
4.1 Charge Question 5: Study Design
The fifth charge question asked the peer reviewers "The method validation study design entailed
spiking characterized background dust with characterized WTC dust at various levels, and then
sending 32 blind samples (16 originals and 16 duplicates) of background and spiked dust to each
of eight laboratories (five commercial laboratories and three government laboratories) These
laboratories characterized the dust using the protocol described above and then sent the results
back to EPA. Was this an appropriate design to achieve the goals of this method validation
study?"
Peer Reviewers' Summary Statement:
The peer reviewers agreed that the general study design was generally appropriate for method
validation Two shortcomings identified were that laboratories did not analyze standard samples
to gauge method accuracy and that the method's equations include unverified assumptions
regarding particle losses due to settling and sieving
Record of Discussion:
During the peer review conference call, additional issues were raised regarding EPA's study
design First, one peer reviewer noted that inconsistent use of sample preparation techniques
might have accounted for the considerable inter- and mtra-laboratory variability observed during
the study Accordingly, this peer reviewer noted that an improved study design would have
involved having a single central laboratory prepare all samples (i e, generate SEM stubs) and
distribute these prepared samples, rather than having the eight laboratories prepare samples on
their own Second, this peer reviewer questioned why a mixture of indoor and outdoor dusts
collected following the WTC disaster was used to spike samples, given that results from the
WTC Dust Screening Study are to be used to evaluate dusts in indoor settings
Another concern about the method study design was the assumption that iron content, as
measured by SEM, can be used to differentiate slag wool from rock wool One reviewer noted
that this assumption was apparently based entirely on data provided the Thermal Insulation
12
-------
Peer Review Summary Report WTC Dust Screening Study
Management Association (TIMA)' However, this peer reviewer presented data from other
sources (see page 10 of Shu-Chun Su's preliminary written comments in Appendix B)
suggesting that differences in iron content might be insufficient for distinguishing these
materials This peer reviewer noted that an improved study design would have thoroughly
characterized the compositional variation of slag wool and rock wool in WTC dust during the
initial stages of method development, rather than relying on published values that might not be
adequately representative of the material found in WTC dust
When discussing the unverified assumptions in the method's equations, a peer reviewer
explained that, when the participating laboratories sieved the original 32 samples through a
100-mesh screen, the partitioning ratio between coarse and fine slag wool was not shown to be
constant For instance, the sieving step could have removed 10% of the slag wool in some
samples, while this same step could have removed 20% of the slag wool in other samples
Similarly, this peer reviewer questioned the impact of the "one-minute waiting period" before
aliquots were extracted from the diluted suspension, given that inconsistent application of this
procedure could have resulted in variable fractions of coarser particles precipitating to the
bottom of the beaker The peer reviewers' recommendations (see Section 3 2) to eliminate
sieving and settling from sample preparation would address these concerns
4.2 Charge Question 6: Statistical Analysis of Data
The sixth charge question asked the peer reviewers "Did EPA and Versar adequately evaluate
and interpret the results from the eight laboratories, as documented by the supplied documents on
NYC dust analyses'?"
Peer Reviewers' Summary Statement:
The peer reviewers were skeptical that EPA's evaluation and interpretation of the study data
were performed fairly Peer reviewers pointed to several non-standard steps taken to enhance the
study's ability to distinguish WTC dust from background dust These steps could be interpreted
as attempts to prove the method's success rather than to objectively evaluate its real-world
potential for fingerprinting WTC dust The peer reviewers' two most notable concerns were1
The reported difference in slag wool for impacted and non-impacted locations was based
on statistical analyses that excluded certain sampling results from the background
locations, but not from the spiked samples The background results and spiked sample
results were statistically indistinguishable when the entire data sat was considered
The study authors disqualified three out of the eight originally selected laboratories that
participated in the WTC Dust Screening Study There was no follow-up investigation to
determine why these three selected laboratories, presumably following the proposed
method, failed to differentiate background dusts from WTC dusts
1 When reviewing a draft of this report, another peer reviewer noted that he thought USGS researchers noted
differences in iron content between slag wool and rock wool based on analyses of WTC dust samples, not based on
the TIMA data
13
-------
Peer Review Summary Report WTC Dust Screening Study
After the peer review conference call, one peer reviewer submitted this additional
summary statement Because the equation used to calculate the final analytical results
(the number of slag wool fibers per gram of dust) does not account for non-uniform
depletion of slag wool fibers due to sieving and due to coarse particle precipitation during
sample preparation, the equation underestimates the number of slag wool fibers per gram
of dust for all spiking and background samples analyzed by all eight participating
laboratories. Had the underestimations been corrected for in the analytical results, the
spiking sample results of the three disqualified laboratories could have been closer to the
expected values than the other five laboratories Therefore, their results might have been
wrongly disqualified. [Note This comment was submitted shortly before the final report
was drafted and reflects the opinion of one peer reviewer The other peer reviewers did
not comment on this matter ]
Record of Discussion:
The peer reviewers discussed many issues raised in their preliminary written comments when
responding to this charge question. Issues raised included, but were not limited to, the following-
Exclusion of data from three laboratories Nearly every peer reviewer was troubled by
the fact that EPA's conclusion regarding the use of slag wool as a signature constituent
was only reached after data from three laboratories were rejected Most peer reviewers
found the decision to exclude these laboratories' data to be unjustified and genuine cause
for concern Given that the stated purpose of the WTC Dust Screening Study was to
validate an analytical method, peer reviewers noted that all analytical data should have
been considered unless EPA had evidence of laboratories gross failure to perform
Significance of law fiber counts and large data extrapolations When applying the
proposed analytical method, laboratories first counted the number of slag wool fibers on
an SEM stub These raw fiber counts were relatively low (i e, generally less than 20
fibers per SEM stub). From these raw data, laboratories computed the numbers of fibers
per gram of dust using large extrapolation factors Thus, even though data throughout the
main body of the report for the WTC Dust Screening Study are presented in thousands of
fibers per gram of dust (using many significant figures), the actual observations made in
the analytical laboratory are based on relatively small fiber counts This was of particular
concern to reviewers because measurement error is known to increase as fiber count
decreases. To illustrate this concern, one peer reviewer stepped through back-of-the-
envelope calculations that, given the range of fibers counted in individual samples, the
measurement quality objective for accuracy (+/-30%) stated in the quality assurance
project plan likely could not be met (see Dan Crane's preliminary written comments in
Appendix B)
Comparison of data across laboratories After examining the raw fiber counts generated
by the eight analytical laboratones (see pages 58 and 59 in the WTC Dust Screening
Study), one peer reviewer noted that the summary report failed to examine or even
acknowledge one of the more notable trends among the data Laboratory D consistently
reported considerably higher fiber counts than did any other laboratory This observation
14
-------
Peer Review Summary Report WTC Dust Screening Study
raised further questions about laboratories' ability to implement the proposed method
consistently
Interpretation of background samples In addition to their concerns regarding excluding
data from selected laboratories, the peer reviewers were troubled by the fact that EPA
also excluded several background samples from its statistical analyses One peer reviewer
noted that, when all background samples are considered, levels of slag wool m
background dust are statistically indistinguishable from levels in the WTC dust from
4 Albany Street Peer reviewers were further concerned that no detailed explanations
were provided to adequately justify the decision to exclude background samples collected
in three locations (New Jersey, Long Island, and Research Triangle Park)
Validity of underlying data One peer reviewer noted that an inherent assumption in the
statistical analyses is that the underlying data are valid. Given concerns raised elsewhere
about the analytical method (e g, the potential inability to distinguish slag wool from
rock wool, the unverified assumptions regarding particle losses during sieving and
settling), this peer reviewer was not convinced that the original data were of sufficient
quality to warrant such detailed statistical analyses
4.3 Charge Question 7: Data Interpretation
The seventh charge question asked the peer reviewers "EPA has observed differences in slag
wool concentration which discriminate between background dust and dust contaminated with
WTC residue Do you agree that the data and analysis presented in the documents support this
observation? Please explain your answer"
Peer Reviewers' Summary Statement:
As indicated in the previous summary statements, the peer reviewers generally do not believe
that EPA made the case from the study data that the proposed method reliably discriminate
background dust from dust contaminated with WTC residue
Record of Discussion:
Issues discussed during the peer review conference call related to charge question 7 included the
following
Concerns about inadequate statistical power One of the main conclusions of the WTC
Dust Screening Study is that". . slag wool measurements appear to be sensitive enough
to distinguish WTC dust (defined as 4 Albany) spiked at the 10% level from background
dust" (EPA 2005a) However, when examining the analytical data (see page 58 and 59 of
the WTC Dust Screening Study), peer reviewers noted that this finding was reached by
comparing two samples spiked at the 10% level using dusts from 4 Albany Street and the
background samples that EPA included in its evaluation With this limited number of
samples and the broad error ranges associated with low fiber counts, a peer reviewer
questioned whether the number of samples considered in the method validation study was
sufficient to reliably discriminate the background dusts from WTC dusts
15
-------
Peer Review Summary Report WTC Dust Screening Study
Comments on required sensitivity When discussing whether the proposed method allows
for discrimination between background dusts and WTC dusts, one peer reviewer referred
to the following quote in the WTC Dust Screening Study " at the 10% spike level, the
slag wool concentration typically exceeds one standard deviation, but never exceeds two
standard deviations, above the background level" (EPA 2005a) This peer reviewer
disagreed with EPA's conclusion that this difference between 10% WTC dust and
background dust was "sensitive enough" to qualify slag wool as a signature, though it
was noted that the signature selection criteria did not establish a minimum concentration
difference (or sensitivity) between background dust and WTC dust that must be achieved
to designate a signature constituent
Interpretation of data in Figure 2 of the review document When responding to this
charge question, one peer reviewer noted that the USGS data plotted in Figure 2 of the
WTC Dust Screening Study appear to show great promise for slag wool as a signature
constituent, and this peer reviewer questioned why data generated dunng the method
validation study were not similarly convincing Other peer reviewers provided several
explanations First, these reviewers advised against focusing too much on trends in Figure
2, because these trends are based largely on outdoor dust samples collected shortly after
the WTC collapse, they noted that the data shown in Figure 3 are more relevant, given
that they represent indoor dusts collected much more recently (i e., September 2004)
Focusing on the data in Figure 3, these reviewers said the concentration differences
between the three spiking levels are not nearly as pronounced Further, noting that the
concentrations in Figure 3 are based on only three samples and on large extrapolations
from the raw fiber count data, a peer reviewer cautioned against drawing strong
conclusions from the trend line
5.0 General Considerations and Alternate Approaches
The final two charge questions were open-ended questions asking peer reviewers to discuss
general considerations and alternate approaches for identifying signature constituents in WTC
dusts This section summarizes how the peer reviewers responded to these questions during the
peer review conference call.
5.1 Charge Question 8: Suggested Improvements
The eighth charge question asked the peer reviewers. "Is there any other way in which the
proposed signature marker(s) can be used to determine the extent to which WTC dust may have
penetrated and remains in indoor environments'"
Peer Reviewers' Summary Statement:
Fiber dimensions and chemical composition of slag wool should be further investigated Fibers
from the WTC collapse would likely include a greater proportion of smaller fibers given the
impact energy of the building collapse, and indoor fibers from the WTC collapse are expected to
be smaller due to the removal of large fibers by passage from outdoor to indoor environments
16
-------
Peer Review Summary Report WTC Dust Screening Study
The chemical composition of slag wool in WTC dusts has apparently not been adequately
char act en zed
Record of Discussion:
Specific issues discussed during the peer review conference call related to charge question 8
include the following
Relevance of fiber dimension. Peer reviewers acknowledged throughout the conference
call that dust samples from non-impacted locations would undoubtedly contain slag wool,
given the widespread use of this material in building construction Accordingly, relying
entirely on fiber counts for identifying a WTC signature would ultimately result in "false
positives," especially for indoor environments with ceiling tiles and locations near recent
building construction, renovation, or demolition activities
However, some peer reviewers suspected that dimensions of slag wool in WTC dust
might be unique when compared to dimensions of slag wool in background dust Two
observations were noted in support of this statement (1) the tremendous impact energy
associated with the WTC collapse presumably led to smaller fiber dimensions in the
WTC dust cloud as compared to fiber dimensions observed in the original building
materials, and (2) larger fibers (as compared to smaller fibers) were more likely to be
removed from the WTC dust cloud due to gravitational settling and "filtering" as these
dusts passed from outdoor to indoor environments. Given these observations, the peer
reviewers recommended that EPA further investigate whether considering fiber
dimensions of slag wool might enhance the proposed signature They noted further that
EPA could begin to address this recommendation by comparing fiber dimensions
observed m two sets of existing samples the USGS WTC slag wool (which is largely
composed of outdoor dusts) and dusts from 4 Albany Street (which is entirely composed
of indoor dusts)
Chemical composition of slag wool in WTC dust A peer reviewer reiterated an earlier
comment that EPA's method development work relied entirely upon composition data
(most notably, iron content) reported by TIM A to distinguish slag wool from rock wool,
even though these published data might not be representative of the chemical
composition of slag wool found in WTC dusts This peer reviewer advocated a more
thorough characterization of the chemical, physical, and optical properties of all
candidate signature constituents prior to selecting a final signature
Further evaluation of gypsum and anhydrite. While the peer reviewers agreed that
gypsum and elements consistent with concrete alone did not meet EPA's signature
selection criteria (see Section 2.2), one peer reviewer noted that anhydrite might hold
promise as a potential WTC signature This suggestion was based on the fact that the heat
of combustion when jet fuel burned in the WTC towers and during subsequent fires in the
debns pile might have created anomalous amounts of anhydrite Accordingly, this peer
reviewer wondered if anhydrite, or the relative amounts of gypsum and anhydrite, could
be used as a WTC signature. The other peer reviewers did not support or reject this
suggestion, but one reviewer clarified that the overwhelming majority of dusts in indoor
17
-------
Peer Review Summary Report WTC Dust Screening Study
environments apparently originated from the dust cloud released dunng the physical
collapse of the WTC towers (and not from the combustion by-products from jet fuel or
the debris pile) Another issue discussed, but not resolved, was whether gypsum would be
sufficiently persistent for use as a signature, particularly in damp environments
The peer reviewers raised additional issues when responding to this charge question, such as
recommendations that EPA evaluate the utility of multi-component signatures and that EPA
reconsider the utility of PLM in its analytical method However, the reviewers discussed these
topics in greater detail when responding to the final charge question (see Section 5 2)
5.2 Charge Question 9: Other Comments
The final charge question asked the peer reviewers "Are there any additional comments or
concerns about this study that have not been addressed by the other questions?"
Peer Reviewers' Summary Statement:
Most of the peer reviewers' comments on this charge question fell into two general categories
modifying the proposed analytical method or investigating the utility of new approaches The
specific comments are documented in the record of discussion, below.
Record of Discussion:
Dunng the conference call, the peer reviewers discussed their responses to the final charge
question A summary of those responses follows References to individual peer reviewers'
preliminary written comments are included for further information on the specific
recommendations
(1) Recommendations for modifying or enhancing the proposed approach
Given the shortcomings of slag wool alone as a signature (e g , the likely false positives),
some reviewers recommended that EPA consider implementing a tiered sampling
approach that would use slag wool as a screening marker for the potential presence of
WTC dust followed by TEM analysis of chrysotile asbestos as a confirmatory marker.
Most reports indicate chrysotile's ubiquity in WTC dust, that the WTC collapse and
combustion produced unique chemical and morphological properties in individual fibers,
that standardized and robust collection and TEM analytical methods exist, and that there
are dozens of certified TEM laboratories with extensive chrysotile analytical capabilities.
This suggestion generated considerable discussion On the one hand, several peer
reviewers supported this recommendation after reviewing the supporting arguments
presented in the preliminary written comments (see Jim Webber's response to charge
question 9 and Frank Ehrenfeld's response to charge question 8, both in Appendix B). On
the other hand, some peer reviewers noted that the utility of asbestos as a signature (based
on indoor air samples) has already been investigated, but found to be of limited utility
especially when compared to the utility of various manmade vitreous fibers. In response,
one peer reviewer noted that a 2002 study (Chatfield and Kommsky 2002) documented
relatively high levels of chrysotile asbestos in indoor environments near Ground Zero
18
-------
Peer Review Summary Report WTC Dust Screening Study
using various bulk and surface dust sampling methods Based on these and other data,
this reviewer said that the potential utility of chrysotile asbestos as a confirmatory
signature warrants further consideration
Throughout the conference call, multiple peer reviewers recommended that EPA
reconsider the utility of PLM as a supplementary analysis to, or possibly in replacement
of, SEM The rationale behind this recommendation is stated earlier in this report and
also in reviewers' preliminary written comments in Appendix B (see Dan Crane's
response to charge question 8 and Mickey Gunter's and Shu-Chun Su's responses to
charge question 9).
If EPA proceeds with slag wool as its signature, a peer reviewer recommended that
positive detections for slag wool should trigger physical site surveys for evidence of other
slag wool sources These other sources include presence of ceiling tiles or exposed
insulation and proximity to large construction, renovation, or demolition sites
As Section 5 1 of this report notes, several peer reviewers recommended that EPA
investigate whether fiber dimension along with fiber count would allow a more definitive
differentiation between WTC dusts and background dusts
Building upon comments raised earlier in the conference call, some peer reviewers
recommended that EPA improve methods to identify critical features of slag wool
specific to WTC dusts Noting that iron content and its resolution by SEM/EDS is
currently used to differentiate slag wool from rock wool, the reviewers recommended
EPA use field studies of WTC dusts to thoroughly characterize the chemical and physical
properties of slag wool and other manmade vitreous fibers found in this matrix, rather
than relying on data published in the TIMA atlas Another suggestion was for EPA to
investigate the utility of the Emmons Double Variation Method when characterizing slag
wool in the dust samples (in Appendix B, see the first issue raised by Frank Ehrenfeld in
response to charge question 9)
One peer reviewer noted that, while the WTC Dust Screening Study acknowledges that
use of slag wool as a signature will likely result m some "false positives," the report fails
to evaluate the likelihood of observing "false negatives " Because the sampling protocol
was not developed to evaluate minimum detection limits, the reviewer concluded that the
study did not establish a lower bound of contamination that the method can reliably find
Should EPA adopt the proposed method, further discussion and evaluation of the
detection limit was recommended.
A peer reviewer recommended that EPA consider revising its method to quantify the
mass of slag wool in samples (which would account for fiber dimension), instead of
quantifying the counts of slag wool in these samples (see Emest McConnell's response to
charge question 9 in Appendix B)
A peer reviewer agreed with the WTC Dust Screening Study's finding that vacuuming
dust from a square-meter area renders the space unusable for replicate samples However,
this reviewer added that the study did not quantify sampling efficiency is expected to
19
-------
Peer Review Summary Report WTC Dust Screening Study
vary with different surface types. Questioning the assumption made during method
development that canister vacuums are 100% efficient in collecting dusts, this reviewer
recommended further evaluation to judge and verify the proposed field collection
techniques
(2) Recommendations for investigating new approaches
The peer reviewers' responses to charge question 9 also included recommendations that
essentially involved considenng new approaches to identifying WTC signatures, as opposed to
mere modifications to EPA's proposed approach The following recommendations are discussed
in greater detail primarily in one peer reviewer's preliminary written comments (see Mickey
Gunter's responses to charge question 9 in Appendix B):
Add academic scientists to the study design team to take advantage of their m-depth
expertise, flexibility to pursue research, and immunity from political fallout
Explore use of multiple analytical methods (especially "bulk" methods) with the potential
for creating a multi-component signature Specific analytical methods mentioned
included PLM, SEM, inductively coupled plasma (ICP) spectroscopy, X-ray fluorescence
(XRF) spectroscopy, and X-ray diffraction (XRD) techniques One peer reviewer
wondered if very simple markers, such as pH, might hold promise when searching for
multi -component signatures
Compare the chemical composition of particulate filters (both PMio and PM2s) collected
at ambient air monitoring stations nearest Ground Zero before the WTC collapse to the
chemical composition of corresponding filters collected after the WTC collapse
Recognizing that improved characterization of WTC building materials would likely help
EPA identify the most appropriate signatures, one peer reviewer recommended that EPA
consider researching WTC building construction documentation or analyzing samples of
debris that has been disposed of at Fresh Kills Landfill Other reviewers, however, were
not convinced of the utility of testing materials in the landfill, given the level of effort
required to obtain a representative sample and the fact that even a statistically-based
sample of debris in the landfill might not be representative of the material in the WTC
dust cloud that infiltrated into indoor environments
Throughout their discussions, peer reviewers wondered if all available sampling results
were considered in the selection of WTC signatures Noting that a large number of
samples have been collected since 2001, both by public and private parties, one peer
reviewer suggested that EPA have a contractor attempt to acquire as much existing data
as possible and then conduct further statistical analyses of the combined data sets These
analyses, he noted, would have to account for numerous factors, such as the sampling
method used, the analytical method used, whether samples were collected before or after
cleaning, and so on. This peer reviewer noted that multivanate statistical analyses of a
broader set of data might reveal viable multi-component signatures that are not readily
discemable from the smaller set of samples that EPA considered in the WTC Dust
Screening Study
20
-------
Peer Review Summary Report WTC Dust Screening Study
6.0 References
Bernstein et al 1996 Evaluation of soluble fibers using the inhalation biopersistence model, a
nine-fiber comparison Inhalation Toxicology 8 345-385
EJ Chatfield, JR Kommsky 2002 Characterization of Paniculate Found in Apartments after the
Destruction of the WTC Chatfield Technical Consulting Limited, Ontario, Canada 2002
RN Clark, RO Green, GA Svvayze, G Meeker, S Sutley, TM Hoefen, KE Livo, G Plumlee,
B Pavri, C Sarture, S Wilson, P Hageman, P Lamothe, JS Vance, J Boardman, I Brownfield,
C Gent, L Morath, J Taggart, PM Theodorakos, M Adams 2001. Environmental Studies of the
World Trade Center Area after the September 11, 2001 Attack. U S Geological Survey. Open
File Report 01-0429. November 27, 2001
EPA 2000 Peer Review Handbook, Second Edition U S Environmental Protection Agency,
Office of Science Policy EPA 100-B-OO-OOl December 2000
EPA 2004 Draft Proposed Sampling Program to Determine Extent of World Trade Center
Impacts to the Indoor Environment External Review Draft (EPA/600/R-04/169A). U S
Environmental Protection Agency October 15, 2004
EPA 2005a Final Report on the World Trade Center (WTC) Dust Screening Method Study. U S
Environmental Protection Agency, Office of Research and Development and Region 2 August
17, 2005
EPA 2005b. Protocol for Preparation and Analysis of Residential and Office Space Dust by
Polarized Light Microscopy and Scanning Electron Microscopy with Energy Dispersive X-Ray
Spectroscopy U S Environmental Protection Agency, National Enforcement Investigations
Center/National Exposure Research Laboratory/National Homeland Security Research Center.
June 27, 2005 (Note This document also appears as Appendix D in the previous reference )
EPA 2005c Al Quality Assurance Project Plan for World Trade Center (WTC) Screening
Method Study Project #WTC-1 Revision #3 U S Environmental Protection Agency, Office of
Research and Development August 8, 2005.
IARC. 2002. Man-made Vitreous Fibers IARC Monographs on the Evaluation of Carcinogenic
Risks to Humans (Volume 81) International Agency for Research on Cancer (World Health
Organization), Lyon, France
HA Lowers, GP Meeker, IK Brownfield Analysis of Background Residential Dust for World
Trade Center Signature Components Using Scanning Electron Microscopy and X-Ray
Microanalysis. U S. Geological Survey Open File Report 2005-1073 2005.
GP Meeker, AM Bern, HA Lowers, IK Brownfield. Determination of a Diagnostic Signature for
World Trade Center Dust using Scanning Electron Microscopy Point Counting Techniques. U S
Geological Survey Open File Report 2205-1031 2005
21
-------
Appendix A. List of Peer Reviewers
-------
vxEPA
United States
Environmental Protection Agency
Independent Peer Reviewers of EPA's World
Trade Center (WTC) Dust Screening Study
Daniel Crane
Lead Physical Scientist
Occupational Safety and Health Administration
U.S. Department of Labor
8660 South Sandy Parkway
Sandy, UT 84070
801-233-4900
Fax: 801-233-5000
Email: crane.dan@dol qov
Frank E. Ehrenfeld III
Laboratory Director
International Asbestos Testing Laboratories
16000 Horizon Way
Mt Laurel, NJ 08054
856-231-9449
Fax: 856-231-9818
Email: frankehrenfeld@iatl com
Mickey Gunter
Professor
Department of Geological Sciences
University of Idaho
Moscow, ID 83844-3022
208-885-6015
Fax: 208-885-5724
Email: mgunter@.uidaho edu
Ernest McConnell
President
ToxPath, Inc.
3028 Ethan Lane
Raleigh, NC 27613
919-848-1576
Fax: 919-848-1576
toxpathmcc@bellsouth.net
Shu-Chun Su
Senior Research Scientist
Hercules Incorporated
500 Hercules Road
Wilmington, DE 19808
302-995-3498
Fax-302-995-4135
Email: ssu@herc com
James Webber
Research Scientist IV
New York State Department of Health
Wadsworth Center
Empire State Plaza
Albany, NY 12237
518-474-0009
Fax: 518-473-2895
Email: webber@wadsworth org
-------
Appendix B. Charge to Reviewers and Preliminary Written Comments
Note: Each peer reviewer prepared the preliminary written comments in this appendix prior to
the peer review conference call. These written comments are preliminary and were used
to help identify key issues for discussion. After reviewing the entire set of preliminary
written comments, some reviewers' opinions on specific matters might have changed
Accordingly, these comments should be viewed as preliminary, while the content of the
main body of this report reflects the peer reviewers' final findings
-------
CHARGE TO REVIEWERS
-------
CHARGE QUESTIONS (August 17, 2005)
World Trade Center Signature Study Peer Review
BACKGROUND
In the days following the terrorist attack on New York City's World Trade Center (WTC)
towers, EPA, other federal agencies, and New York City and New York State public health and
environmental authorities initiated numerous air monitoring activities to better understand the
ongoing impact of emissions to the outside environment from that disaster In 2002, EPA's
Region 2 turned its attention to the indoor environment, providing a volunteer "clean and test" or
"test only" program for residents and homeowners who wished to have their apartments cleaned
and/or tested Asbestos was selected as a surrogate for the presence of WTC-related
contamination for this program
In March 2004, EPA convened the WTC Expert Technical Review Panel to interact with
the Agency and the public on plans to monitor for the presence of any remaining WTC dust in
indoor environments near Ground Zero For more information on the WTC Expert Panel, see
http //www.epa gov/wtc/panel Approximately 750 units in an area extending north to Houston
Street in lower Manhattan and across the East River into a portion of Brooklyn were proposed to
be sampled for contaminants of potential concern (COPC), as well as for specific constituents that
can be used as markers to identify residual contamination by dust from the collapse of the WTC
These "WTC dust signature" constituents are the cornerstone of the sampling program as they
will provide the basis for estimating the geographic extent of the remaining residue in dust of the
WTC towers collapse. Using the results of dust sampling for the WTC building collapse
signature, EPA will decide whether indoor cleanup or other activities are warranted at this time
Based on the previous work of the United States Geological Survey (USGS) and others,
EPA identified three components of the dust generated by the WTC towers collapse that could be
used to screen sampled dust for the presence of WTC dust These markers, or signature
components, were mineral slag wool, gypsum and elements of concrete EPA developed the
following working hypothesis for the signature "A dust sample that contains WTC dust will
have slag wool and elements of concrete and gypsum present in 'significant quantities' when
compared to typical indoor urban dust" Experts from the USGS, EPA's Office of Research and
Development, EPA's National Enforcement Investigations Center, and the commercial testing
laboratory community worked together to develop an analytical method to quantify the
concentration of these three markers in indoor dust
Between September 2004 and Apnl 2005, numerous samples were taken in impacted
buildings near Ground Zero and at background locations. These samples have been analyzed for
these three markers to determine whether or not they validly constitute a WTC signature, as
suggested by earlier USGS efforts Also, they were used in a method validation study whose
primary purposes were twofold. 1) to evaluate the analytical method with regard to method
variability (as measured by both inter- and mtra-laboratory variability), and cost and expediency
issues, and 2) to assist in the determination of the lower limits of concentrations of the markers
that could be reliably measured, and that could be reliably distinguished from background
concentrations This method validation study involved five contract laboratories and three
-------
federal government laboratories This study entailed spiking known background dust with
varying concentrations of known WTC collapse dust, and then having the laboratories perform a
blind analysis on both the spiked samples and background samples
The charge questions focus on the following the basis for identifying signature
components, the validity of the hypothesis, whether the existence of a signature has been
adequately demonstrated, and the work done to develop, validate and apply an analytical method
for the three signature components
The following documents are the primary ones being reviewed. 1) the analytical protocol
used by the eight laboratories and 2) EPA and Versar (EPA contractor) reports documenting the
validation study including the background and hypothesis, analyses of samples taken to verify
the hypothesis, and the results and interpretation of the method validation study Several
background documents will also be available to the reviewers. These include the draft final
sampling plan to be used to evaluate the presence and levels of contaminants of potential
concern of any remaining WTC dust in indoor environments near Ground Zero (this plan
includes an overview of the Signature Study), two USGS reports on the signature development,
and other pertinent reports that provide additional background on the characterization of WTC
dust The reviewers are not being asked to provide comments on these background documents
If the reviewers believe they need any additional documents or clarification of information
provided in the documentation provided, they will be supplied through the peer review contractor
CHARGE QUESTIONS
Basis for Development of a Signature.
Ql) The following criteria were established to assist EPA in the selection of appropriate
constituents in dust to be characterized as WTC signature constituents
a) They are present at levels unique to WTC dust (distinct from urban dust),
b) They are persistent for many months (not volatile),
c) They are sufficiently homogeneous in WTC dust, and
d) Available analytical methods are able to detect these screening materials with a
small sample size, low minimum detection limit, and low interference from other
dust components.
Based on information in the supplied documents, and any other knowledge you may bring to the
table, are these criteria adequate for establishing a WTC signature marker? If the answer is no,
please elaborate
Documentation of the Existence of a Signature
Q2) Based on information in the supplied documents, and any other knowledge you may
bring to the table, do you agree or disagree with the conclusion that slag wool alone meets the
criteria as a WTC marker? Please explain your answer
-------
Analytical Method Development
Q3) Are the analytical methods written in sufficient detail such that a qualifying laboratory could
follow the methodology and obtain valid results without supplemental assistance from EPA or
other sources'?
Q4) If the answer to Question 3 is no, what items or information could be added so that a
qualified laboratory could follow this protocol and obtain valid results without the assistance of
the EPA or other sources'?
Method Validation Study
Q5) The method validation study design entailed spiking characterized background dust with
characterized WTC dust at various levels, and then sending 32 blind samples (16 originals and
16 duplicates) of background and spiked dust to each of eight laboratories (five commercial
laboratories and three government laboratories) These laboratories characterized the dust using
the protocol described above and then sent the results back to EPA Was this an appropriate
design to achieve the goals of this method validation study*?
Q6) Did EPA and Versar adequately evaluate and interpret the results from the eight
laboratories, as documented by the supplied documents on NYC dust analyses9
Dust collected from currently occupied buildings is expected to have lower levels of the key
WTC constituents as compared to dust sampled near September 11,2001 in time or sampled
more recently but in uninhabited heavily impacted buildings EPA will use the results of this
method validation study to determine the final distinguishing concentrations for the WTC
marker(s) If currently sampled dust has this marker(s) at or above such a distinguishing
concentration, EPA would consider the sampled dust to "contain residues of WTC dust" for
purposes of estimating the geographic extent of WTC impacts and making cleanup decisions
The key requirement for a distinguishing concentration is that it be adequate to distinguish
between dusts that do not contain WTC residues from those that do with a reasonably low false
positive error rate
Q7) EPA has observed differences in slag wool concentration which discriminate
between background dust and dust contaminated with WTC residue Do you agree that
the data and analysis presented in the documents support this observation*? Please
explain your answer
Q8) Is there any other way in which the proposed signature marker(s) can be used to determine
the extent to which WTC dust may have penetrated and remains in indoor environments'?
Q9) Are there any additional comments or concerns about this study that have not been
addressed by these questions'?
-------
PRELIMINARY COMMENTS
FROM
DANIEL T. CRANE, M.E.
-------
Daniel T Crane
Draft Response to Charge Questions
CHARGE QUESTIONS (August 17, 2005)
World Trade Center Signature Study Peer Review
Preface to Response to Charge Questions'
In the days following the terrorist attack on New York City's World Trade Center (WTC)
towers, EPA, other federal agencies, and New York City and New York State public
health and environmental authorities initiated numerous air monitoring activities to better
understand the ongoing impact of emissions to the outside environment from that
disaster In 2002, EPA's Region 2 turned its attention to the indoor environment,
providing a volunteer "clean and test" or "test only" program for residents and
homeowners who wished to have their apartments cleaned and/or tested. Asbestos was
selected as a surrogate for the presence of WTC-related contamination for this program
In March 2004, EPA convened the WTC Expert Technical Review Panel to interact with
the Agency and the public on plans to monitor for the presence of any remaining WTC
dust in indoor environments near Ground Zero For more information on the WTC Expert
Panel, see http //www epa.gov/wtc/panel Approximately 750 units in an area extending
north to Houston Street m lower Manhattan and across the East River into a portion of
Brooklyn were proposed to be sampled for contaminants of potential concern (COPC), as
well as for specific constituents that can be used as markers to identify residual
contamination by dust from the collapse of the WTC. These "WTC dust signature"
constituents are the cornerstone of the sampling program as they will provide the basis
for estimating the geographic extent of the remaining residue in dust of the WTC towers
collapse. Using the results of dust sampling for the WTC building collapse signature,
EPA will decide whether indoor cleanup or other activities are warranted at this time.
Based on the previous work of the United States Geological Survey (USGS) and others,
EPA identified three components of the dust generated by the WTC towers collapse that
could be used to screen sampled dust for the presence of WTC dust These markers, or
signature components, were mineral slag wool, gypsum and elements of concrete. EPA
developed the following working hypothesis for the signature "A dust sample that
contains WTC dust will have slag wool and elements of concrete and gypsum present in
'significant quantities' when compared to typical indoor urban dust." Experts from the
USGS, EPA's Office of Research and Development, EPA's National Enforcement
Investigations Center, and the commercial testing laboratory community worked together
to develop an analytical method to quantify the concentration of these three markers in
indoor dust
Between September 2004 and April 2005, numerous samples were taken in impacted
buildings near Ground Zero and at background locations These samples have been
analyzed for these three markers to determine whether or not they validly constitute a
WTC signature, as suggested by earlier USGS efforts. Also, they were used in a method
validation study whose primary purposes were twofold 1) to evaluate the analytical
method with regard to method variability (as measured by both inter- and intra-laboratory
variability), and cost and expediency issues, and 2) to assist in the determination of the
-------
Daniel T Crane
Draft Response to Charge Questions
lower limits of concentrations of the markers that could be reliably measured, and that
could be reliably distinguished from background concentrations This method validation
study involved five contract laboratories and three federal government laboratories. This
study entailed spiking known background dust with varying concentrations of known
WTC collapse dust, and then having the laboratones perform a blind analysis on both the
spiked samples and background samples
The charge questions focus on the following the basis for identifying signature
components; the validity of the hypothesis, whether the existence of a signature has been
adequately demonstrated; and the work done to develop, validate and apply an analytical
method for the three signature components
The following documents are the primary ones being reviewed 1) the analytical protocol
used by the eight laboratories and 2) EPA and Versar (EPA contractor) reports
documenting the validation study including the background and hypothesis, analyses of
samples taken to verify the hypothesis, and the results and interpretation of the method
validation study
The answers to the charge questions below reflect my experience with industrial hygiene
laboratories and the way in which they analyze particulates by microscopy They further
reflect my knowledge and expenence with the issues surrounding the measurement of the
COPCs.
The EPA requested that the Light Microscopy portion of the study not be reviewed as it
was applied in this protocol
CHARGE QUESTIONS
Basis for Development of a Signature:
Ql) The following cntena were established to assist EPA in the selection of appropnate
constituents in dust to be characterized as WTC signature constituents
a) They are present at levels unique to WTC dust (distinct from urban dust),
Answer to Ol a)
With respect to constituents of concrete dust, and gypsum/anhydrite as markers, 1
agree with the study that while comprising some of the most common constituents
of the dust, these components are not unique in any way in their physical
properties as components of the WTC debns In addition, these components are
ubiquitous in urban settings While there must be an elevated level of these
components in the shadow of the dust plume, their inspecificity of composition
necessarily excludes them from the marker suite because it is expected that they
will lead to an unacceptably high number of false positive results
Considering slag wool as a marker, the evaluation study provides an answer,
albeit not in a straightforward manner. The collected background samples
-------
Daniel T Crane
Draft Response to Charge Questions
showed that slag wool was not broadly prevalent away from ground zero, or the
drift zone of the debris cloud The exceptions to this were the samples excluded
on the basis of the presence of ceiling tiles or other building material containing
slag wool m the sampled location The data from this study indicates that slag
wool could be used as a marker, in the impacted area There is no special
attribute of the slag wool itself to tie it to the World Trade Centers. It was and is a
common component in the insulation and other building materials used in the
WTC as well as many other buildings in the area. However, the number of fibers
in the collected dust reflect that in the considered region, with the exception of
buildings with an internal source, the slag wool fibers seen reflect mostly fibers
generated in the collapse of the World Trade Center buildings
Because of the known positive interference, the use of this marker demands a site
survey for slag wool containing materials m any sampled building where a
positive survey result is found Any positive results in the presence of an internal
source must render as suspect the decision that WTC signature dust is present.
b) They are persistent for many months (not volatile),
Answer to Ol b")
This report demonstrated that the slag wool component has been persistent in the
sampled locations since the collapse of the World Trade Centers I believe that
the slag wool component, due to its physical nature is sufficiently persistent and
pervasive in the dust to serve as a marker
With respect to the persistence of concrete and gypsum/anhydrite, it appears that
these components were indeed found. This demonstrates that they are persistent
In the case of concrete dust, it is to be expected that it is persistent over the
relatively short time since the event In the case of gypsum/anhydrite, it is known
that these materials react with water and this may alter the persistence of the
component dependent upon the exposure history of the dust
Note however, that any problems referent to concrete and gypsum/anhydrite are
moot if they are discounted as potential marker components.
c) They are sufficiently homogeneous in WTC dust
Answer to Ol c)
Virtually all samples m the dispersion zone contained the three marker
components. However, presence does not imply homogeneity at any particular
sampling size At the individual analytical sample size, the heterogeneity of the
collected material virtually guarantees that the samples are inhomogeneous This
does not disqualify the material as a marker material It does limit the precision
with which it can be quantified, and subsequently re-sampled after cleanup to
certify the cleanup
-------
Daniel T Crane
Draft Response to Charge Questions
d) Available analytical methods are able to detect these screening materials with a small
sample size, low minimum detection limit, and low interference from other dust
components
Answer to Ql d)
The analytical technique has a putative detection limit of 1 target particle counted
in the SEM analyzed area This has not been reduced to a protocol detection limit
in terms of fibers or particles per gram of collected material. The recovery
efficiency of fibers and other particles from the sample surfaces is suspected to be
variable and unknown, but assumed by the framers of the protocol to be
sufficiently high to make a clean-up decision This is based on their statement in
several places that a second sample would not be expected to provide an
equivalent sample if taken from the same area This seems to be a reasonable
assumption However, no data is given in the study that documents the rate of
collection in resampling the same areas
The particle size detection limit is set to 0 5 urn. At 500 X magnification this is
approximately 0 25 mm on the SEM screen That approximates the viewing pitch
of a color screen, or near single pixel resolution This seems too small to be of
practical use, although it will certainly pick up the very smallest particles visible
larger than that
With reference to interferences, using the USGS catalog of WTC particles, the
analyst should be able to sort out most, if not all of the particles of interest Other
MMVF present in the samples are sufficiently different in chemistry such as to be
distinguishable by competent microscopists
Based on information in the supplied documents, and any other knowledge you may
bnng to the table, are these catena adequate for establishing a WTC signature marker'' If
the answer is no, please elaborate
General answer to Ql'
Having investigated a variety of materials for the purpose of establishing the
original COPCs and knowing that the three suggested materials are the most
commonly found materials m the WTC debris, it is apparent that they provide the
highest potential to be found in locations impacted by WTC dust. They may
provide a surrogate for measurement for the COPCs which may be present at
levels too low to be reliably measured
Concrete dust and gypsum/anhydrite dust are too widespread in areas unaffected
by WTC and are at levels too unpredictable within the impacted area to be of any
real value as a tracer or surrogate for dispersion of WTC dust
In the study, dispersion of slag wool seemed to be confined to the impacted area.
Where it can be shown that the slag wool did not originate from internal sources
-------
Daniel T Crane
Draft Response to Charge Questions
in a sampled building, It would appear that slag wool represents the best candidate
as a surrogate for the COPCs
Slag wool alone of the three has the potential to be used as a marker However, it
is apparent that any protocol which hopes to adequately determine WTC impact
on a particular building must be accompanied by a building survey as an adjunct
for any positive result to determine whether or not there is an internal source
which can account for the positive sample In addition, if a history of open-air
demolition is available, it should be added to the record as a potential source for
the target materials This is especially true if the demolition was by implosion or
other very dusty mode
It is also apparent that the actual detection limit for the protocol is unknown If it
is assumed that the vacuum sampling technique has an adequate sampling
efficiency such that a decision can be made, then slag wool represents the best
surrogate for WTC dust sampling However, the high uncertainty associated with
the analysis makes resampling problematic in that a second sample after clean-up
may not adequately represent the level of remaining contaminant
Documentation of the Existence of a Signature.
Q2) Based on information in the supplied documents, and any other knowledge you may
bring to the table, do you agree or disagree with the conclusion that slag wool alone
meets the catena as a WTC marker? Please explain your answer.
Answer to O2
Slag wool is not specific to the WTC collapse It is essential to remember that
while it is a significant source in the impacted area, there may be other significant
sources which may confound the proposed surveys A history of open-air
demolitions, if any should be sought Also, it is well to remember that slag wool is
being used as a surrogate for the detection of other COPCs which may be at levels
below current technology to measure This is because of the ubiquity of slag
wool in WTC dust. That same ubiquity must be respected as a confounder m the
interpretation of any results
Slag wool alone of the three has the potential to be used as a marker However, it
is apparent that any protocol which hopes to adequately determine WTC impact
on a particular building must be accompanied by a building survey as an adjunct
for any positive result to determine whether or not there is an internal source
which can account for the positive sample
It is also apparent that the actual detection limit for the protocol is unknown If it
is assumed that the vacuum sampling technique has an adequate sampling
efficiency such that a decision can be made, then slag wool represents the best
surrogate for WTC dust sampling However, the high uncertainty associated with
-------
Daniel T. Crane
Draft Response to Charge Questions
the analysis makes resampling problematic in that a second sample after clean-up
may not adequately represent the level of remaining contaminant
Analytical Method Development
Q3) Are the analytical methods written in sufficient detail such that a qualifying
laboratory could follow the methodology and obtain valid results without supplemental
assistance from EPA or other sources9
Answer to O3 No
The study was begun with a joint meeting to correlate the analyses of the
laboratories Even in the face of this training session, three of the laboratories
were adjudged unable to adequately perform the method as written It is unclear
whether the attendees to the sessions were the actual analysts or other laboratory
representatives With the level of technique required to properly apply the
protocol, it is essential that more intensive procedural correlation take place
Q4) If the answer to Question 3 is no, what items or information could be added so that a
qualified laboratory could follow this protocol and obtain valid results without the
assistance of the EPA or other sources9
Answer to O4
This analytical protocol is heavily reliant on technique It is essential that
correlation of effort be accomplished with the actual analysts who will be
performing the tests This collaboration should consist of hands-on mentored
training with sufficient trials such that the analysts are demonstrably proficient.
In addition, the protocol provides for two different mounting techniques for SEM
analysis liquid drop and filtration These two mounting methods should not be
expected to provide equivalent mounts For a random field-choice counting
method, the particulate must be randomly distributed on the filter In my
experience, the drop method tends to deposit the particulate in a non-random
fashion due to surface tension effects, or deposition techniques, often leaving
areas with visibly higher concentrations of material For low concentration
samples, the non-random deposition may not be readily apparent The drop
method is perfect for quick work requiring identification of particles. The liquid
drop technique has the advantage that it is relatively quick, while the filtration
technique may take a very long time, due to pore blockage (polycarbonate filters
do not have a very high porosity.) Also, if the hydrophihc treatment is removed
from the polycarbonate filters by allowing them to dry after wetting with the
isopropanol and before the filtration supernatant is added to the filtration
chimney, the filters will then by hydrophobic and it will be very difficult to pull
the charge through the filter.
-------
Daniel T Crane
Draft Response to Charge Questions
However, unless the test laboratories agree to and use the same technique
(preferably the filtration technique,) there is a significant window for
disagreement and precision broadening If the method used is the liquid drop,
sample preparation technique correlation between laboratories is critical and may
be uncontrollable even with a higher degree on inter-laboratory coordination
Method Validation Study
QS) The method validation study design entailed spiking characterized background dust
with characterized WTC dust at various levels, and then sending 32 blind samples (16
originals and 16 duplicates) of background and spiked dust to each of eight laboratories
(five commercial laboratories and three government laboratories) These laboratories
characterized the dust using the protocol described above and then sent the results back to
EPA Was this an appropriate design to achieve the goals of this method validation
study?
Answer to O5
The goals of this study can be simply reduced to validation of the sampling and
analytical protocol for the identification of three common components of WTC
dust. The study aimed to determine whether or not the method could discriminate
between various spiked concentration levels In addition, it was established that
areas inside the impacted area show significantly increased levels of the target
substances
Given that this was designed as a screening method, it still appears to be time
intensive, especially with respect to the post-processing required for the concrete
and gypsum/anhydrite analysis The study did not comprise a large number of
samples As a result, the precision of the method cannot be tightly specified
This, however, does not preclude the study from meeting its basic goals The
study had sufficient power to demonstrate that two of the target substances could
not be used as viable markers for WTC dust It further showed that the test
laboratories could detect and measure the concentration of slag wool fibers in
spiked and background samples.
It is less clear that the method, as tned, has the power to provide quantitative
results. This will be addressed more in-depth m the answer to question 6 It is
appropnate to remember that it was conceived as a screening method, and as
such, should not be expected to provide highly quantitative results
As a screening method, the protocol is adequate to establish that slag wool fibers
are present in the samples to some undetermined level of detection This leaves
the question unanswered as to how little contamination this method can detect
The upshot is that when the method detects fibers, it is certain that slag wool is
-------
Daniel T Crane
Draft Response to Charge Questions
present (and similarly assumed that the COPCs, for which it is surrogate, are
present) When the result is negative, it is unclear whether or not slag wool is
present This is of some concern as certain surfaces with greasy or otherwise
"sticky" properties may not be adequately sampled.
Q6) Did EPA and Versar adequately evaluate and interpret the results from the eight
laboratories, as documented by the supplied documents on NYC dust analyses?
Dust collected from currently occupied buildings is expected to have lower levels of the
key WTC constituents as compared to dust sampled near September 11, 2001 in time or
sampled more recently but in uninhabited heavily impacted buildings EPA will use the
results of this method validation study to determine the final distinguishing
concentrations for the WTC marker(s) If currently sampled dust has this marker(s) at or
above such a distinguishing concentration, EPA would consider the sampled dust to
"contain residues of WTC dust" for purposes of estimating the geographic extent of WTC
impacts and making cleanup decisions The key requirement for a distinguishing
concentration is that it be adequate to distinguish between dusts that do not contain WTC
residues from those that do with a reasonably low false positive error rate.
Answer to 06
The method was adjudged adequate by Versar to determine whether or not a
property is contaminated with WTC dust To the extent that the protocol is
considered a screening technique, it shows that areas expected to contain the
target contaminants do indeed show them, and the areas expected to be free of the
contaminants, apparently do not This begs the question as to why some of the
urban samples outside the impacted area (background) do not show any fiber slag
wool Slag wool was not unique to the WTC, and it is expected that some
samples of dust from these background areas should show some slag wool fibers
The stark stratification of results serves to highlight the lack of power for the
study to investigate the detection level of the procedure
Simply, in the absence of a site source of fiber, a positive sample is likely to
represent a property contaminated with WTC dust The opposite cannot be
inferred from the data in this study
The exclusion of the results from three of the laboratories in establishing
acceptable confidence limits, for fiber counting, is unjustified Given the known
expertise of the particular laboratories recruited for the study, exclusion of any of
the three is genuine cause for concern In Appendix F, upper and lower 95%
confidence limits are established using data from laboratories designated A, B, C,
D, and H Laboratories E, F, and G were designated as commercial laboratories
which had unacceptable performance. This left only two of the original five
commercial laboratories The analysis in Appendix F highlights some important
differences between the laboratories, or the techniques they used Reasons for the
differences were not discussed at length in the report However, it was not the
stated aim of this project to identify or capacitate any lab or group of labs It was
to validate the protocol Except in the case of gross failure of a laboratory to
-------
Daniel T. Crane
Draft Response to Charge Questions
perform, the data from all laboratories should be included in any evaluation of the
data Further, because of the known breadth of error m particle counting, it is
questionable whether the number of samples in the study provided sufficient
power to the analysis to legitimately make an exclusion decision. Finally,
contract commercial laboratories selected by the bidding process are unlikely to
perform differently than the aggregate of the five selected laboratories Indeed,
one or more of the excluded laboratories may win the bid The entire suite of data
should be used to characterize the performance of the protocol
When examining the data from the perspective of fibers per gram as in the Versar
preliminary report Table 4 (page 56), the data is represented as relatively large
numbers with significant precision (many significant figures ) However, when the
data is examined from the perspective of the raw number of fibers examined,
Table 5, (page 58,) a different picture is seen With the exception of laboratory D,
the actual number of fibers counted is seen to be quite low Indeed, laboratory D
systematically shows significantly higher counts than any of the other
laboratories No mention is made in the report of this quite sinking result
The counting of particles is only approximated by normal statistics when there are
large numbers of particles When the numbers or counted particles are low, the
statistics may be log-normal, as suggested m the text, or may sometimes be
described by Poisson statistics It may be that the data need be analyzed using
distributionless analysis if the data does not suggest a particular distribution.
Also, if the counts are sufficiently low, a bias correction may be necessary to
correctly model the data. Such bias corrections are empirically determined In
addition, it is usually the case that there is a subjective component to the variance
in addition to that expected directly from the distribution
It is sufficient to note that a Data Quality Objective of ±30 or ±35% is
inappropriate for the small numbers of fibers seen An unsophisticated look at the
data assumes a Poisson approximation is for ease of calculation The variance
goes as 1/n, and the standard deviation goes as 1/Vn If the background results
from table 3, p 54 of the Versar report are used to calculate a pooled average for
1/Vn, then a relative standard deviation can be seen to be 41% (A value of 1 was
used m place of 0 counts to represent a detection limit) Using a normal
approximation for 95% confidence limits, they are seen to be ±81%. A
distributionless estimation would be higher This provides broader acceptance
limits than those outlined in the review documents. Again, this does not include
any estimate of subjective uncertainty, always present in particle counts
The impact of this is that clean-up decisions will be made on the basis of very few
fiber counts, even though the reported number in terms of fibers per gram will be
very high Any measurements made subsequent to the cleanup will be expected
to be lower than the initial survey The decision band may be sufficiently broad
as to make such comparisons impossible.
-------
Daniel T Crane
Draft Response to Charge Questions
Q7) EPA has observed differences in slag wool concentration which discriminate
between background dust and dust contaminated with WTC residue Do you agree that
the data and analysis presented in the documents support this observation? Please explain
your answer
Answer to O7
Limited to the data contained in this report, it appears that slag wool alone can be
used to discriminate against background concentrations where there are no
discoverable slag wool sources which could account for a positive result These
may include, but are not limited to slag wool containing ceiling tile, insulation, or
known demolition sites which could have contaminated the site It has been noted
elsewhere in this evaluation that the converse may not necessarily be true This
study did not investigate the detection limit in sufficient detail to determine the
minimum concentrations that could be found in potentially impacted buildings
Q8) Is there any other way in which the proposed signature marker(s) can be used to
determine the extent to which WTC dust may have penetrated and remains in indoor
environments9
Answer to O8-
Given the length of time since the collapse, surface sampling in the way followed
in this protocol appears to have the best chance to collect consistent samples (for
the purpose of the study) Other surface sampling techniques, such as tape
sampling, might provide a more efficient sampling of target surfaces, and thereby
provide a measure of lower contamination amounts (speculative ) But, tape
sampling would have the disadvantage of being from a much smaller surface and
consequently being less precise, providing less opportunity for replicate analysis
and would not integrate well with the data from the current study Aggressive air
sampling, such as previously used for the original measurement of COPCs, has
the known deficiency of being very close to the detection limits for the
methodology involved and is not practical in re-occupied spaces
With respect to Polarized Light Microscopy (PLM), it should be noted that so
long as the study was centered on all three of the marker substances, the PLM
was, perhaps, extraneous. However, when the study excluded concrete and
Gypsum/anhydrite, PLM should have re-emerged as a viable alternative. It can be
used quickly as a go/no-go test to determine the presence of mineral wool As-~
such, it would not necessarily be used quantitatively, although many industrial
hygiene analysts are already trained in semi-quantitatively determining
percentages using this technique. An analytical infrastructure already exists
nationwide for accreditation of the techniques involved, along with the training
availability. For the semi-quantitative purposes of the study, 1 believe that PLM
still offers a less-expensive and quicker alternative
Q9) Are there any additional comments or concerns about this study that have not been
addressed by these questions9
-------
Daniel T Crane
Draft Response to Charge Questions
Answer to O9
The study properly chose to step away from the concrete dust and
gypsum/anhydrite markers due to their widespread presence in the target
environment. Although slag wool was not specific to the World Trade Center
construction, this limited study seems to show that it is possible that it can be used
as a marker for the intrusion of WTC dust to spaces in the impacted area. The
bulk of the comments in this section refer to the protocol as it is used in the
establishment of slag wool as a marker
As a small study, it necessarily has limitations, and some of these should be
noted The report correctly recognizes that vacuuming up sample from a 1 square
meter area (or more) renders that space unusable for a replicate sample.
However, it does not quantify the sampling efficiency on any surface, even a few
representative carpet, fabric or hard surfaces This could have been done on soft
surfaces by replicate vacuuming, on hard surfaces by the same or by pre/post tape
samples Canister vacuums are not any where near 100% efficient This is
demonstrated daily by door-to-door salesmen using collection techniques very
similar to those contained in the sampling protocol
Related to this is the fact that this overall protocol is likely to have false positives
No false negatives were noted There were indeed negatives The sampling
protocol was not constructed in such a way as to evaluate the detection limit for
contamination It is not known how small a contamination this method can
reliably find Presumably, the analytical protocol has a relatively low detection
limit, but this was not addressed in the report The spiked samples at 1% were all
positive definite, indicating that a lower detection limit is probable
The analytical method itself is workable as it relates to slag wool The
gypsum/anhydrite and concrete analysis would be quite cumbersome and time
consuming for any laboratory unequipped with automated chemical and spatial
analysis. That must be part of the bid if it is included in the final protocol
In applying the analytical procedures, there are some physical laboratory
correlation issues related to lab technique that must be addressed in a group
setting with all of the actual analysts involved in the work. Laboratory technique
is learned by doing rather than reading It is an art It is essential that the
technicians actually performing the work be jointly trained, or uniformly trained
in some way; either by coming together as in the study or by a key person training
all of the technicians Representative training is questionable in quality and
ability to provide the necessary precision and technical expertise
Along with this, within the laboratory protocol there were two distinct methods of
sample preparation that, in my expenence, will lead to different results Only the
filtration technique should be used Liquid drop techniques do not generally
provide a sufficiently random particle distribution on the sample
-------
Daniel T Crane
Draft Response to Charge Questions
Either or both of the issues above may have resulted in the perceived deficiencies
of the excluded laboratories, and in the notably higher results overall for
laboratory D.
In the SEM analysis, the laboratories used the USGS BIR1-G for calibration of
the chemical analysis. Although not critical to the study at hand, the
magnification calibration should be documented somewhere for reference,
especially because some of the labs did document fiber size As noted above, the
lower size selection criterion was 0.5um Because the analytical magnification of
choice was 500x, this is 0 25mm on the viewing screen. (Unless the
magnification is calibrated to the hard-copy device, and the screen is
uncalibrated ) If this is the case, the viewing screen should be calibrated in order
that the analyst knows how big the analytical targets are
With respect to the aim of the project, the protocol design must be sufficiently
robust as to be used to establish that the space is clean after any cleanup or other
remediation Such validation should be accomplished before any chance of re-
entramment can take place either from missed sources in the space, from common
ventilation plenums, or from sources generally exterior to the target space
-------
PRELIMINARY COMMENTS
FROM
FRANK E. EHRENFELD III
-------
EPA/ERG Project 2005 Ehrenfeld / 1ATL
World Trade Center Signature Study Peer Review
General Outline for Response
Acknowledgements
o H Sonny Robb, Craig Liska, and Thorn Snyder for conversations and analytical work on
WTC dust samples Specifically, work on MMVF characterization by PLM, SEM, and
TEM using archived samples from various clients and locations in and around the Lower
Manhattan site. Thanks also to John Napolitan for literature and extended database
searches Special thanks to Lou Solebello for his vast knowledge of fiber
characterization techniques
Charge Questions
Basic Development of a Signature
Ql - Are the listed criteria adequate for establishing a WTC signature marker'
Documentation of the Existence of a Signature
Q2 - Does the slag wool marker alone meet the cntena as a WTC marker'
Analytical Method Development
Q3 - Are the analytical methods written in sufficient detail such that a qualifying laboratory
could follow the methodology and obtain valid results without supplemental assistance from EPA
or other sources?
Q4 - What information would I add so that a qualified laboratory could follow this protocol
and obtain valid results without the assistance of EPA or other sources'
Method Validation Study
Q5 - Was this method validation study appropnate to achieve the goals outlined?
Q6 - Did EPA/Versar adequately evaluate and interpret the results correctly'
Q7 - Does this study's data support the observations made by Lowers in OFR 2005-1073?
Q8- What underlying issues relating to these marker(s) and regarding the penetration and
persistence of the WTC dust on indoor environments can be further measured and explored?
Q9 - What additional comments or concerns about this study have not been addressed by these
charge questions'
-------
EPA/ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Basic Development of a Signature1
Ql - Are the listed cntena adequate for establishing a WTC signature marker?
The listed criteria to establish an appropriate marker include
a) It must be present at levels unique to WTC dust.
b) It must be persistent
c) It must be sufficiently homogeneous
d) Available methods are able to detect these screening materials with a small sample size,
low minimum detected limit, and low interference from other dust components
Based upon the review of previous work outlined by EPA, ves it is mv opinion that these are
adequate for establishing a WTC signature marker It is obvious that any marker would have to
be present at levels that are unique Can the marker be successfully differentiated from
background or ambient levels? Yes, this should be included as part of the cntena Should the
marker be persistent enough to be sampled and measured years after the catastrophic event?
Absolutely. Should the marker be sufficiently homogeneous? Of course
Should available (and established) analytical methods be able to detect the marker even when
there is little sample size, even when there would be a terrible detection limit, and even when
other matenals might interfere with these measurements? Ideally yes Is this charge question
asking about the specifics of slag wool as a marker? No, it only concerns a general list of criteria
that should be met for WTC dust As such, I agree that any marker should meet these criteria
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Documentation of the Existence of a Signature.
Q2 - Does the slag wool marker alone meet the criteria as a WTC marker'
If this question is 'does slag wool' meet the criteria for a WTC marker? Then ves there has been
evidence presented to draw that conclusion Is the evidence overwhelming or compelling? No
Does this translate into 'the only marker that meets the criteria is slag wool', or 'slag wool is the
only marker that meets the critena exclusive of other components established in the dust'' Then
no. there is strong evidence that other components meet the critena
Regardless of its catastrophic source, WTC dust may be the most studied mixture of building dust
and debris in the history of forensic and environmental science Though only a limited series of
samples, certainly the evidence presented by Lowers1 and Meeker2 is enough to conclude that
slag wool meets the above catena Their findings are clear, beyond reproach, and influential in
this determination That said, there is some concern over the homogeneity of slag wool at
geographically extreme locations and at elevated sites in the impacted area
It would be helpful to learn more about the background samples collected at RTI since the slag
wool component of these dust samples far exceeded those studied by Lowers Were these
collected using the same field sampling protocols3? In fact the question still remains that though
they "are not included as they are not representative of NY City background dust"(page 12,
paragraph 2, section IV)4, they might still represent background dust in many building throughout
NYC that have slag wool containing building materials See also Q9.
But what about other well defined and studied markers9 Do these also meet the criteria9 Some
do and some do not. For the purposes of this being a brief response, here are a few questions to
consider
1 Will 'qualified' labs be able to differentiate these markers9 See Q3
2 What are the expectations for sampling and analysis costs9 See Q9
3 Will sample analysis be cost effective9 SeeQ8andQ9
4 Though asbestos is present at known5 levels in NYC background dust, why was asbestos
abandoned as a marker? Clearly the asbestos in the published ambient levels could be
differentiated from the asbestos in the WTC dust Newman's6 opinion is clear on the
subject, though he does not list adequate* reasons for his conclusions There are not
-------
EPA / ERG Project 2005 Ehrenfeld /1ATL
World Trade Center Signature Study Peer Review
persuasive reasons for this oversight' [Recently I have been made aware that Newman
was referring to the asbestos as a surrogate issue and not necessarily asbestos as a
marker] SeeQ8andQ9
5 Please also consider the distribution of slag wool (and heavier dust components) in a
diminished gradient from ground zero to limited geographic locations (and higher
elevations). This is a real weakness in the study that may need to be further explored
See Q8 and Q9
6 Could a primary and secondary marker system be employed to verify WTC dust origins?
This was postulated by Meeker2 in his conclusion's third paragraph See also Q8 and Q9
7 Is there significant separation between background and WTC slag wool levels to
conclude that it alone meets the marker criteria? Please note (page 48 of the Final
Reportl that Versar points out that for the DB source samples "at the 10% spike level, the
slag wool concentration typically exceeds one standard deviation, but never exceeds two
standard deviations above the average background sample concentration " Again even
for the USGS source samples Versar states that "the 5 and 10% levels are essentially all
more than two standard deviations above the average background " EPA savs (page 21
Conclusions) that this is "sensitive enough" I'm not sure that I can agree
Regardless of the case for additional markers, a very compelling argument is repeatedly made in
the final report4 (p 6 section I) and in ancillary documents7 (Appendix Al, p 10, section A5 1.)
that the basis for an "appropriate screening test" would result in some false positives but very
few, if any, false negative Slag wool meets the criteria
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Analytical Method Development
Q3 - Are the analytical methods written in sufficient detail such that a qualifying laboratory
could follow the methodology and obtain valid results without supplemental assistance from EPA
or other sources'
Overall, ves the protocols and step bv step analytical procedures should be sufficient for a
'qualified' lab to employ successfully without assistance. There are comments/suggestions in O4
that may clarify certain steps found in the proposed method
There is some question about what defines a 'qualified' laboratory Considering the
abandonment of data from three of the eight laboratones retained by Versar for this study, it is
questionable if enough independent laboratories could be qualified Indeed at the bottom of page
19 of Appendix A mention is made that the method evaluation will include evidence "that the
analytical method works well enough, and is able to be carried out by enough analytical
laboratories " What are the expectations of the panel concerning 'works well enough' and
'enough analytical laboratones nr> Will the panel develop criteria concerning laboratory
qualifications'' Would they, should they consider allowing a government laboratory to handle the
analytical work with a contracted QC lab? See also Q9
The DQOs and MQOs that are outlined in A7 seem to allow for a wide enough range of
variability. Please consider the impact of a sentence of Appendix A again (page 13 section
A.6 3.) that briefly addresses the subject of method validity by saying that "criteria such as time
for analysis, and mtra- and inter- laboratory variability will be considered" Now consider that
almost 40% of the data did not meet muster. Does this mean that the method's procedures were
unable to be followed or that the laboratory staff was not competent enough to follow through' I
think the latter
What are the criteria for determining if a lab is qualified' Appendix Al (p 14 section A8) lists
only training by PLM and SEM as qualifiers On page 19 of the Appendix (section C) it notes
that all contract labs are required to employ standard QA practices and be under in-house QA
personnel.
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Versar's preliminary report Addendum dated August 5, 2005 (page 60 of the Final Report)
indicates that each of the five laboratories analyzed a glass standard reference material to
determine statistical acceptance limits The accompanying table seems to display data for three
laboratories, not five See also Q4
Versar's statistical analysis of laboratory performance notes on page 63 of the Final Report that
"Because Labs E, F, and G fail in more than one category in both spiked data sets, they are not
included in the remaining study " That's it1?! The one paragraph coverage on page 13 of the
Final Report touches upon this issue and does conclude that "all laboratory comments will be
taken into consideration when finalizing the protocol." This deserves an investigation that can, m
a one or two page memo, list possible reasons why these labs failed to meet qualifying status
This memo would go beyond the reasons listed on page 20 of the Final Report for high
variability These include operator/analyst expenence This is understandable in many analytical
investigations, but surely technicians must be engaged who have basic knowledge of the target
analytes This begs the question how were the labs in this study selected?
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Q4 - What information would I add so that a qualified laboratory could follow this protocol
and obtain valid results without the assistance of EPA or other sources7
This answer assumes that a 'qualified' laboratory meets some sort of criteria for staff experience
and background as well as minimum facilities and instrumentation requirements
1. The initial mass of samples and sub-samples seems to be an issue noted at the top of page
21 of the Final Report Please amend the method to include a tighter range.
2. Please also list sample rejection criteria Please consider- Too little sample (ex If less
then 200mg7), too many interferences (ex Loaded with ash or combustion product), too
much moisture, cross contamination with other submitted samples, etc.
3 Standardize or list the acceptable particle analysis software so that there is no variation
(Appendix p 39 Section 7).
4 Please standardize or provide acceptable parameters for somcation energy and calibration
information. (Appendix page 42 Section 11 5)
5 Define "the more concentrated sample" with an empirical range (Appendix A page 43
Section 1221)
6 Please expand on step 12 2 1 on page 43 of Appendix A referring to "counting fibers per
type until a statistical representation of the ratios of fiber compositions in the sample is
achieved "
7 Please specify the duration and sensitivity of the scan used to determine
Gypsum/Anhydrite in Section 12 2 2 on page 44
8 Section 12.2 3 notes 40 minute total acquisition time per map for each of 10 fields Is
this overkill? Won't 400 plus minutes be terribly expensive for each sample? Is this
really a 'screening' method7 Could this be tnmmed down without affecting sensitivity?
9. Section 1225 implies that 1000-1200 particles will have to be individually characterized
to achieve the required sensitivity of 10% relative error Is this accurate7 Again, is this
really a 'screening' method7 Please amend.
10 This same section (Appendix A page 46 section 1225) indicates that particles over 20
\un long will have to scanned and images recorded and saved If you digest the
information displayed in Meeker2 Figure 9 is this a burden of documentation that is un-
necessary7 Please amend.
-------
EPA / ERG Project 2005 Ehrenfeld /1ATL
World Trade Center Signature Study Peer Review
11 It would be helpful to see a mock data report or Certificate of Analysis listing all of the
expected reporting limits, variables, concentrations, densities, LOQs etc This might be
supplied along with the examples of Data Sheets supplied in the Final Report's Appendix
section 15
12 Page 60 of the report copies a Versar Memo concerning a glass standard reference
material The interpretation of this data should be limited to the inter-laboratory variation
of instrumental detection Is this a commentary about the method's ability to successfully
measure these elements in real-world particles' It only partially achieves this. Should a
section of the method include expanded detail on expected calibrations of software, EDS
sensitivity, etc ?
13 What could be implemented to further qualify a laboratory' Any on-site inspections or
demonstration of proficiency using prepared standards and spiked samples'
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Method Validation Study
Q5 - Was this method validation study's design appropriate to achieve the goals outlined1?
Was the design appropriate? Yes The MQOs, DQOs, quality assurance, and other controls were
selected and placed in the study with sensible foresight. The design seems to be sound There is
some question as to the abandonment of PLM and exclusion of TEM as a screening test. See Q8
There is also some concern about the evolution and final course that the method followed See
Q9
Was this method validation study able to achieve the goals outlined on page 9 Section III of the
Final Report9 Or put another wav. did the study validate the method? Not quite These are the
three goals as outlined in the method
1) To demonstrate that slag wool is a 'reasonable' marker for WTC dust This has been
achieved and is further discussed in Q2 and Q7
2) That WTC dust at a diluted concentration can be distinguished from background. The
method has limited utility and too many loose ends to pronounce it valid
3) That the analytical method works well enough and is able to be earned out by enough
analytical laboratories to a) evaluate the markers and b) to distinguish WTC dust from
background dust There is persuasive evidence that this study has demonstrated that the
lab community has not achieved a level of competency to meet this last objective
Reasoning for the last item concerns the ever increasing complex and convoluted analytical
procedures and the production of data beyond what I would consider reasonable for a 'screening'
method Instead, this proposed 'screening' method seems to lean to a full inorganic particle
investigation, complete with quantitative and qualitative analytical procedures and the subtle
nuances that define both characterization and forensic science See Q9
Please see my comments concerning 'qualifying' laboratones in Q3 above
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Q6 - Did EPA/Versar adequately evaluate and interpret the results correctly7
Versar's analysis starts on page 45 of the Final Report. Their evaluation of the analytical data
summarized in three bullet points on that page seems to be adequate Indeed their conclusion on
page 72 clearly states that "the conclusions are based solely on the analytical data provided from
the laboratory test and a few analyses of the 100% WTC spike samples." If their task was to
provide analysis of the data alone, then they have indeed fulfilled their mission EPA's
responsibility, on the other hand, is probably broader. Certainly the WTC panel will expect a
thorough review of all the attributes of the study in addition to the analysis of the data Taking
this into consideration, then no. EPA did not adequately evaluate and interpret the results from
the eight labs etc
It was not eight labs it was five The EPA's Executive Summary (page 4 of the Final Report and
page 21 of the Conclusion) made no mention of the significance of the failure of almost half of
the contract labs to produce acceptable data This is an omission that though may not be senous,
certainly raises speculation that other factors may not have been properly addressed (ex RTI, NJ,
and LI samples) Page 13 of the Final Report briefly mentions this issue but without its proper
weight Instead of (page 4 of the Executive Summary and page 21 of the Conclusions) the weak
pronouncement that 60% of the labs 'reasonably' were able to measure the marker, it would have
been refreshing to have read that either "the method has clearly validated the hypothesis that slag
wool meets all of the qualifications of a marker and that the method works well enough that
enough labs can successfully employ the method" or that "the method, in its current form has not
yet demonstrated its utility"
Is there significant separation between background and WTC slag wool levels to conclude that it
alone meets the marker catena? Please note (page 48 of the Final Report) that Versar points out
that for the DB source samples "at the 10% spike level, the slag wool concentration typically
exceeds one standard deviation, but never exceeds two standard deviations above the average
background sample concentration " Again even for the USGS source samples Versar states that
"the 5 and 10% levels are essentially all more than two standard deviations above the average
background " EPA says (page 21 Conclusions') that this is "sensitive enough". Considering the
strengths and weaknesses of most commercial testing laboratories with these capabilities. I'm not
sure that I can agree
10
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
This shortcoming in interpretation seems even more relevant when accounting for the interjection
in this charge question of additional caveats and disclaimers about the employment of this method
on currently occupies spaces and their expected lower levels of COPC. The charge question
states that "the key requirement for a distinguishing concentration is that it be adequate to
distinguish between dusts that do not contain WTC residues from those that do with a reasonably
false positive error rate" There is a question of confidence regarding this study's limited data set
and the 10% spiked level that it claims is 'reasonably' measured Perhaps the study might be
expanded to increase confidence in its efficacy. See Q9.
Versar's data reduction is well reasoned and appropriate I agree with their logic and their
statement (page 63 of the Final Report) that C and D labs should be added to the preferred
performer group of labs Lab H's inclusion is probably warranted for the reasons listed Versar's
model for precision is satisfactory I have trouble interpreting Table 5 on page 71!
Both Versar's data analysis and EPA's conclusions regarding gypsum and concrete components
seems to be validated
11
-------
EPA / ERG Project 2005 Ehrenfelcl / IATL
World Trade Center Signature Study Peer Review
Q7 - Does this study's data support the observations made by Lowers in OFR 2005-1073 and
Meeker in OFF 2005-1031? Does the study's data support the observations by NERL and its
preliminary analysis7
Evaluating the three studies recently presented bv EPA and comparing them to this method's
conclusions yields mixed results
In her discussion Lowers states that "the maximum amount of COPC, such as Chrysotile asbestos
derived from WTC dust could be calculated by using the relative proportions of phases found in
WTC dust by Meeker". Maybe, but there has been other studies9' 10 offering data on various
COPC (often expressed in different units of measurement - see also Q9) that also could be used
to estimate dust components (even those that might meet the definition of marker) Further work
on slag wool containing background samples (ex. RTI samples discussed earlier) would have to
be done to connect-the-dots towards this conclusion Lowers did not analyze (or at least report)
any Chrysotile detected in the 1073 study Furthermore, the technology employed (SEM/EDS)
may not have been sufficient to accurately detect asbestos as a COPC Her data supports the
hypothesis that this study promulgates that slag wool is a sufficient marker for WTC dust
The most significant statement in Lowers is the last paragraph "This study has examined only
six background samples In order to arrive at a statistically significant representation of New
York City residential background dust compositions more analyses are needed In addition,
sampling background dust from office buildings and other interior environments is needed to
determine background ranges of signature components in these types of interior spaces"
Therefore the 'strength' of the Lowers report is its data suggesting slag wool at insignificant to
none detected levels in NYC background dust The limited sample set (six) underscores its
weakness Indeed, EPA should be cautioned about proceeding with implementing any method
based on so few samples A more robust set of data would certainly have to be established
Meeker's conclusions are worthy of a second look I agree with the conclusion that "if slag wool
fibers are not found in settled dust samples above a predetermined critical level, it is unlikely that
COPC derived from the WTC could be present at significant levels in the samples " His last
paragraph is critical Here he states that the ability of slag wool to be a marker depends on its
levels in other background samples This continues with the statement that there is a real
12
-------
EPA/ERG Project 2005 Ehrenfeld /IATL
World Trade Center Signature Study Peer Review
possibility of slag wool being present in background samples (ex ceiling tiles etc ) Finally,
Meeker adds that the size distribution might prove useful in distinguishing WTC and non-WTC
dust samples. I agree with his conclusions in that slag wool has the potential to be a marker and
that the particle size distributions may be useful, especially taking into account the physical
forces inflicted upon much of the dust (pulverized much finer than virgin components found m-
situ) Like Lowers, the 'strength' of the Meeker report is its data suggesting slag wool at
significant levels in known WTC impacted samples The strength of Meeker's 1031 observations
that support this study's premise continues in his bnef mention of the correlation between the
proposed marker and other COPCs Conversely, the limited sample set again underscores its
weakness Once more, EPA should be cautioned about proceeding with implementing any
method based on so few samples
The NERL preliminary analysis data in Appendix B has more samples. The duplicate QC results
are admirable Was there an order of magnitude difference in slag wool background versus
impacted sites7 Was this evidence of 'distinguishing' concentrations between the two types of
samples?
This screening method study did have a more vigorous set of samples (32 total background and
spiked samples) then the previous preliminary EPA studies1'2. Did the screening method study
show conclusive evidence that slag wool levels between background dust and impacted dust be
can routinely discriminated17 No Did the screening method study conclude that slag wool levels
between background dust and 'spiked dust' can be discriminated at low levels (10%) by a select
few labs' Yes
Please consider that only the NERL samples used the field sampling protocol24 that the study's
method outlines Was there a model for the efficacy of the vacuum collection protocol using
settled dust (preferably dust that had been settled for a few years) spiked with the various WTC
reference materials7 Preparing the spiked standards in a laboratory setting and, from the limited
data, extrapolating potentially far-reaching conclusions, may not be wise See Q9
Overall, the background and spiked samples examined in this study bv the five qualified
laboratories suooort the observations previously presented bv Lowers. Meeker, and NERL
13
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Q8- Is there any other way in which the proposed signature marker(s) can be used to
determine the extent to which WTC dust may have penetrated and remains in indoor
environments'7
For slag wool as the proposed signature marker limited For other potential markers perhaps
First, by process of elimination, let's consider all of the COPC and eliminate those that will not
meet the proposed cnteria as markers
There is compelling evidence that both gypsum and concrete in the WTC dust should not be
considered as markers (page 21 of the Final Report) Indeed, four years after the catastrophic
event, these components of the dust may be significantly removed in occupied spaces and greatly
reduced in any space where a cursory clean-up was attempted The data suggests that any
exposure to moisture (outside samples, ram exposed, carpet cleaning, etc ) would alter (by partial
dissolution and disbursement) these constituents Finally, buildings renovated in the last four
years (and any pre-existing construction debris) may preclude any significant discrimination
between these markers and WTC dust
The heavy metals and organic compounds listed as COPC may have some utility as secondary
confirmation markers though their persistence and ability to significantly discriminate from
typical NYC background levels may pose a problem20'21> 22
Please consider the distribution of slag wool (and heavier dust components) in a diminished
gradient from ground zero to both limited geographic locations and higher elevations What
COPC has been well studied, will easily penetrate higher elevations, is persistent, will remain
unaltered (ex gypsum and concrete are water soluble), and remain in measurable quantities even
after cursory clean-ups' What COPC has established analytical methods' What COPC has
guidelines in place to protect workers and establish re-occupancy thresholds' What COPC has
experienced field engineering and project design in place' What COPC meets the criteria for a
marker set forth by the WTC panel and listed in the first charge question' What COPC also
meets the study design criteria on page 9 of the Final Report' What COPC has an experienced,
well qualified, and economically competitive laboratory structure in place' The answer is
obvious and the issue has been debated previously Chrysotile asbestos in WTC dust'
14
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
The previous debate11 may have been off-center as to the utility of Chrysotile as a marker That
is, the previous review concerned asbestos in air as a surrogate (travels with and is linked to other
components) and not as a marker It concluded that Chrysotile in air would serve as a suitable -
but not stand-alone - surrogate. Why not Chrysotile asbestos in the surface dust as a marker? At
least why not as a secondary marker1' That peer review panel also concluded that a multi-tiered
approach is feasible "asbestos sampling results would be strengthened with the addition of lead
wipe samples "
It has already been proposed that Chrysotile and lead on surfaces be considered as a paired
marker-like team22 Meeker also proposed a multi-tiered approach in his conclusions2 "An
analysis strategy for routine samples could evolve using rapid scans of settled dust by SEM to
look for the presence of the MMVF. If found, these fibers could then be analyzed using EDS to
determine fiber compositions. If the majority of fibers (>85%) detected were of slag wool
composition, or if slag wool was found at a predetermined critical concentration, the sample
would then be searched for gypsum and concrete particles along with other MMVF components
Further conformation of the presence of WTC dust could then be reached by looking for
secondary components in the approximate abundances found in this study Alternatively, if slag
wool, gypsum, and concrete were present, the sample could then be analyzed for COPC such as
asbestos, lead, and organic compounds "
Though asbestos is present5 in NYC background dust, why was asbestos abandoned as a marker?
Clearly the asbestos in the published ambient levels could be differentiated from the asbestos in
the WTC dust Newman's6 opinion is clear on the subject, though he does not list adequate9
reasons for his conclusions There are not persuasive reasons for this oversight? [Recently I have
been made aware that Newman was referring to the asbestos as a surrogate issue and not
necessarily asbestos as a marker],
The public health (not to mention the risk assessment and public relations) implications of
sampling for and utilizing a dual asbestos and lead approach is huge25.
Therefore, slag wool (and all MMVF components') and Chrysotile examined in the surface dust
(not the air) in potentially impacted areas along with lead analysis might be re-visited There are
several advantages and limitations that would have to be considered using this approach:
15
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
1 What are the ambient levels5 of Chrysotile asbestos in NYC? How much information
exists in both public and private databases7 Certainly this will vary greatly especially in
older buildings that have poor housekeeping issues
2 What are the levels of Chrysotile in WTC dust? This has been studied9 Perhaps there
exists enough statistically sound data in the private sector (building owner and insurance
industry?) to establish these levels with confidence The concentrations might be studied
not in percent of the WTC dust (well studied at 0 02-2 0%) but in structures per cm2 of
the surface area tested?
3. Porous surface sampling and the 'carpet as a reservoir' issues for these COPCs must be
addressed6 It may be considered an analytical bonus that these micron-sized COPCs
remain in carpet etc even after substantial vacuuming and cleaning. Indeed, methods
have been used to fully investigate some of these concerns12
4 The same physical forces that were mentioned by Meeker concerning MMVF and its
ability to be entrained in carpet (electrostatic interactions) and to not be easily
homogenized on a sample are also a concern of asbestos (strong surface charges)
5 Re-entrainment issues are well studied '3l '"1S, and could be used as a final re-occupancy
test along with lead on surfaces
6. There are m-place engineering and design companies that could implement off-the-shelf
protocols for clean-up and testing This would be more cost effective than having to go
through the typical learning curve with new guidelines and protocols.
7 There are m-place laboratory methods for sampling and analysis of asbestos in surface
dust17> 18> 19 There are m-place models for asbestos in surface dust, that though not risk
based, may provide utility for this project23
8 The asbestos and lead laboratory community is extremely competitive and reasonably
cost effective
9 Commercial asbestos laboratories have trained staff, adequate facilities for the
preparation and analysis of this target analyte, and are accredited by local, state, and
national organizations that require strict QAQC programs, proficiency testing, and on-site
assessments
10 Though no asbestos in dust proficiencies exist, EPA might promote this to establish inter-
laboratory precision There have been samples generated and studies conducted among
several well qualified laboratones by ASTM16 towards its objective of establishing
accuracy and precision data for their methods.
16
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
11 Slag wool and other MMVF are also identifiable by TEM/EDS In fact, at the micro
level of analysis of TEM, perhaps this is an easier analysis then the study's proposed
method
17
-------
EPA / ERG Project 2005 Ehrenfeld /1ATL
World Trade Center Signature Study Peer Review
Q9 - What additional comments or concerns about this study have not been addressed by these
charge questions7
There are several issues that have been addressed within the context of the charge question
responses Issues dealing with standards preparation, size of the data set for the various
preliminary investigations, concerns over why selected laboratories failed to produce data within
statistically allowable bounds, field sampling efficacy, and the choice of slag wool as a primary
marker have been discussed
Here are additional comments and concerns in no real order of importance
1. How would the vast amount of borosilicate present interfere with proper discrimination
of the other MMVFs7 The Na sensitivity must be at a maximum to differentiate by
SEM/EDS Has there been any thought to employing the Emmons Double Variation
Method26*7 This is a relatively accurate and inexpensive screening tool to compare known
versus unknown glass fragments
2 The failure of those few labs to produce qualified data is still disturbing Could this peer
review panel see information as to each lab's reasons why there were shortcomings'7
This is especially a concern since each lab knew the high profile importance of their task.
3 The lack of uniformity27 between studies done on the WTC dust is probably very
frustrating for the WTC Panel and those working on making sense of the data pool
Could EPA task a contractor to pool all of the public and pnvate data from the last few
years worth of efforts to clean impacted areas7 Specifically, to add even more depth as to
the efficacy of sampling procedures, cleaning procedures, engineering controls,
measurable markers and/or surrogates, reasoning for re-cleaning, and baselme/MCL
threshold guidelines
4 IATL and others28 have noted that a significant population of Chrysotile asbestos from
WTC dust has been 'damaged' That is, at high magnification, it is clear that there is a
surface pitting on the Chrysotile surface. Most times the Chrysotile appears as single
fibrils In extreme cases these structures have a measurable sulfur peak Might this be
used as a confirmation tool to differentiate WTC Chrysotile from ambient Chrysotile in
buildings that contain ACM7
5 Most times if you go to a surgeon for a medical opinion they will recommend surgery
Perhaps a poor example, but if the panel tasked an SEM-based laboratory to develop a
18
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
screening method, then the expected was realized SEM has limited utility29 for
analyzing Chrysotile asbestos fibrils SEM also has to slog through time-consuming EDS
analysis of hundreds (even thousands30) of particles This is by no means a slight against
the excellent USGS lab and those who have employed their skills and expertise to assist
the panel In fact, my bias tends towards materials investigations using PLM, SEM, and
TEM techniques This is evidenced in my response in Q8.
6 What are the panel's expectation for finding 'enough qualified laboratories'9
7 What are the panel's expectations regarding costs per sample' Including collection,
planning, and analysis' Appendix A section 6 3, p 13 indicates that time/cost will be a
factor in assessing the method.
8 Could a multi-tier approach be considered?
a Perhaps a simple pH test might be a good starting point. All WTC dust samples
measured in the first month or two after the collapse had high (>9) pH
b This might be followed by PLM screening for the characteristic minerals and
their fingerprint percentages. By screening a lOOmg portion would be
examined under the stereoscope, a few 5-10mg sub-samples would be ftirther
analyzed to qualify the minerals (isotropic versus anisotropic) and one sub-
sample more closely characterized using multiple RI oils and comparing to
known WTC dusts
c. Similar sub-samples might be sent for lead analysis.
d. Perhaps 100cm2 carpet swatches (or microvac samples) might then be analyzed
for Chrysotile asbestos and MMVF phases
Finally, what constitutes a 'screening' method7 This is perhaps the 'elephant in the room' that
many panelists may ignore Is this WTC Signature Screening Method really a screening method?
The concept that may have started as a simple screening protocol has become a complex and
convoluted analytical procedures with the required production of data beyond what I would
consider reasonable for a 'screening' method Instead, this proposed 'screening' method seems to
lean to a full inorganic particle investigation, complete with quantitative and qualitative analytical
procedures and the subtle nuances that define both characterization and forensic science. Some
examples
Meeker2 sought to utilize "routine analytical techniques". The proposed method may not
have evolved in that direction
19
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
Qualitative and quantitative approach for slag wool (40minutes x 10 fields/scans each =
400 minutes per sample) as proposed is beyond the concept of screening.
Qualitative and quantitative approach for gypsum and associated minerals (1000-1200
particles analyzed) as proposed is beyond the concept of screening.
Many real samples will have limited matrix material to measure This low level of
residual dust may produce horrendous limits of detection by SEM The conclusion that
SEM can 'quickly and easily find' slag wool may be debatable.
20
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
References.
1 Lowers et al, Analysis of Background Residential Dust for World Trade Center Signature
Components Using Scanning Electron Microscopy and X-Ray Microanalysis Open File
Report 2005-1073
2- Meeker et al Determination of a Diagnostic Signature for World Trade Center Dust using
Scanning Electron Microscopy Point Counting Techniques Open File Report 2005-
1031.
3- ASTM D-5438-00 Standard Practice for Collection of Dust from Carpeted Floor for
Chemical Analysis
4- US EPA, Final Report on the World Trade Center (WTC) Dust Screening Method Study,
August 17, 2005
5- Ewmg , W M., Further Observations of Settled Dust in Buildings ASTM STP 1342, Ed
Beard and Rook, 2000
6- Newman, D M Comments to the EPA World Trade Center Expert Technical Review
Panel March 1, 2005
7- Appendix Al Quality Assurance Project Plan for World Trade Center (WTC) Screening
Method Study Revision 3 August 8, 2005.
8- Newman, David E, On The Issue of Microvac Sampling EPA World Trade Center
Expert Technical Review Panel, comments May 3, 2004
9- Chatfield, E J and Kommsky, J R, 2002, Characterization of Paniculate Found in
Apartments after the Destruction of the WTC: Chatfield Technical Consulting Limited,
Ontario, Canada 2002.
10- Loiy et al, Characterization of the Dust/Smoke Aerosol that Settled east of the World
Trade Center un Lower Manhattan after the Collapse of the WTC, EHP 110 (7) 703-714,
2002
11- US EPA 68C02060, Summary Report for the Peer Review on the Use of Asbestos as a
Surrogate Contaminant for Determining the Risk from Other Contaminants April 29,
2004
12- Millette et al, Methods for the Analysis of Carpet Samples for Asbestos, in Settled
Asbestos Dust - Sampling and Analysis Lewis Publishers CRC Press 1994.
13- Millette et al, Applications of the ASTM Asbestos in Dust Method D5755 in Beard and
Rook, Advances in Environmental Measurement Methods for Asbestos, ASTM STP
1342, 2000
14- Crankshaw et al, An Overview of Settled Dust Analytical Methods and Their Relative
Effectiveness" in Beard and Rook, Advances in Environmental Measurement Methods
for Asbestos, ASTM STP 1342, 2000
21
-------
EPA / ERG Project 2005 Ehrenfeld / IATL
World Trade Center Signature Study Peer Review
15- US EPA, Interim Final WTC Residential Confirmation Cleaning Study, Vol 1, May
2003
16- Millette, James Conversations and report from ASTM Committee D 22 07 in 2002 and
2003
17-ASTM D5755 Asbestos in Surface Dust Structure Concentrations by Microvacuum
Technique using TEM 2002.
18- ASTM D5756 Asbestos in Surface Dust Mass Concentrations by Microvacuum
Technique using TEM 2002
19- ASTM D6480 Asbestos in Surface Dust by Wipe Technique using TEM 2000
20- R J. Lee Group Mercury Assessment, 130 Liberty Street Property, Expert Report
Mercury May 2004
21- Meeker and Lowers. Particle Atlas of the World Trade Center Dust Open File Report
2005-1165 p4 2005
22- US EPA June 30, 2005. Draft Final Plan Proposed Sampling Program to Determine
Extent of WTC Impacts to the Indoor Environment p9-12
23- US EPA June 30, 2005. Draft Final Plan Proposed Sampling Program to Determine
Extent of WTC Impacts to the Indoor Environment pi0-11
24- US EPA June 30, 2005 Draft Final Plan Proposed Sampling Program to Determine
Extent of WTC Impacts to the Indoor Environment P24
25- US EPA 68C02060, Summary Report for the Peer Review on the Use of Asbestos as a
Surrogate Contaminant for Determining the Risk from Other Contaminants April 29,
2004 p9.
26- Emmons, R C American Mineralogist, 13, 504-514 (1929? And 14, 415-426 (1930?)
27- US EPA June 30, 2005 Draft Final Plan. Proposed Sampling Program to Determine
Extent of WTC Impacts to the Indoor Environment. P22
28- Conversations with Dr James Millette in 2002 and 2003
29- NYSDOH ELAP Certification Manual 1992-1998 versions
30- US EPA Section 12 2 5 of Appendix Al Quality Assurance Project Plan for WTC
Screening Method Study Revision 3, August 8, 2005
22
-------
PRELIMINARY COMMENTS
FROM
MICKEY E. GUNTER, Ph.D.
-------
Mickey Gunter
Mickey Gunter's comments to "Charge Questions" for WTC signature
September 11,2005
Q1) The following criteria were established to assist EPA in the selection of appropriate
constituents in dust to be characterized as WTC signature constituents
a) They are present at levels unique to WTC dust (distinct from urban dust)
b) They are persistent for many months (not volatile)
c) They are sufficiently homogeneous in WTC dust, and
d) Available analytical methods are able to detect these screening materials with small sample
size, low minimum detection limit, and low interference from other dust components
Based on information in the supplied documents, and any other knowledge you may bring to the
table, are these criteria adequate for establishing a WTC signature marker'' If the answer is no,
please elaborate
Yes the above criteria are adequate However, the word "they" needs some discussion
and there are some rather loaded words in "d " To me "they" implies that what is sought
is some combination of constituents that could characterize WTC dust from other forms
of New York City dust, yet only slag wool has been selected As I will discuss more later
m my comments, I think that some unique combination of materials could have been
found
In regard to "loaded words" in "d," who might these "available analytical methods" be
available to? Commercial labs only, or more research-based labs7 Also, many of the
commercial labs, at least the ones used for this study, have their roots in asbestos work, it
seems that the WTC dust is being treated more like asbestos type material (i e , particle-
based analyses are being performed) than a bulk material "Small sample size" should
have been defined, does this mean 1 mg, 1 gram, or 100 grams7 Again, for typical
"asbestos" methods often only sub-mg sample sizes are available, these small sample
sizes in turn dictate what analytical methods can be used. But for WTC dust, multi-gram
samples are available and different analytical methods might be chosen based on these
larger sample sizes The same question should be applied to what is meant by "low
detection limit" The issue in "detection limit" as used herein really is one of the
sensitivity of a method to correctly determine if a dust sample contains a WTC signature
This might mean being able to measure one material at the ppm level, or use several
different materials measured at the percent level
Q2) Based on information in the supplied documents, and any other Knowledge you may bring to
the table, do you agree or disagree with the conclusion that slag wool alone meets the criteria as
a WTC marker' Please explain you answer
Based on what has been provided, I agree that slag wool can be used to detect 10% and
higher WTC components of indoor dust I am not sure how well slag wool meets the
criteria in Ql, however For instance, the data on background levels of slag wool (in
fibers/gram) are given in Table 1 (p 14) with an average and standard deviation of 35,950
and 74,300, respectively Thus, the standard deviation appears to be larger than the
-------
Mickey Gunter
mean, which most certainly does not assure usefulness of these background data After
removing a couple of high values (and this troubles me to remove observations so the
results "fit" better for the proposed use), these values drop to 17,740 and 15,835,
respectively Regardless of which numbers are used, the variability of slag wool is very
high in the background samples These high variations, in turn, decrease the sensitivity
of using slag wool to determine if WTC dust is a constituent of indoor dust to 10%
Thus, while slag wool is a WTC dust marker, its variation in the background samples
limits it use
Q3) Are the analytical methods written in sufficient details such that a qualifying laboratory could
follow the methodology and obtain valid results without supplemental assistance from EPA or
other sources'
Yes
Q4) If the answer to Question 3 is no, what items or information could be added so that a
qualified laboratory could follow this protocol and obtain valid results without the assistance of the
EPA or other sources?
NA
Q5) The method validation study design entailed spiking characterized background dust with
characterized WTC dust at various levels, and then sending 32 blind samples (16 originals and
16 duplicates) of background and spiked dust to each of eight laboratories (five commercial
laboratories and three government laboratories) These laboratories characterized the dust using
the protocol described above and then sent the results back to EPA Was this an appropriate
design to achieve the goals of this method validation study?
In theory, this method was the correct approach, but there were problems in
implementation
Spiked sample homogeneity (i.e., standard preparation): The main intent of this
portion of the study was for the labs to characterize the WTC dust for the WTC
signature(s) As seen in all the reported data, there was a large amount of inter- and intra-
laboratory variability On page 49, four caveats are given why this variability may have
occurred, but I think a fifth (and probably the most significant) caveat needs to be added
the labs prepared their own samples.
There is no doubt it is difficult to homogenize the spiked samples, and a major portion of
the variability in the lab measurements might have been a combination of differences in
the samples they received and/or their methods of preparing the samples A much better
test of this method would have been if each of the labs made measurements on the exact
same sample set (i e, one set of SEM mounts could have been made by USGS and sent to
each of the participating labs) If this would have been done, I believe a much better idea
of precision and accuracy would have been obtained on the slag wool contents
-------
Mickey Gunter
Choice of spiking samples (outdoor vs. indoor): One of the intents of this project was
to determine the WTC signature for indoor air, so it was a little strange why a mixed
outdoor/indoor WTC dust sample was used for spiking I think the size distributions (and
probably compositions) of indoor and outdoor dust would differ Thus, I do not think the
indoor/outdoor WTC spiking sample should have been used, and results from it are not
really helpful to address the issue of indoor air
Fibers vs. areas for different constituents It seemed strange that slag wool was
measured based on fiber counts and the other two components were based on area I
think the fiber counting is a holdover from measunng asbestos, and it makes no sense to
use it In reality, it would have made much more sense to determine the percent mass of
the components, which would be more directly related to the area than to the number of
fibers
Q6) Did EPA and Verser adequately evaluate and interpret the results from the eight
laboratories, as documented by the supplied documents on the NYC dust analyses?
I think the results are basically over-interpreted I really found little use for the statistical
analysis presented on pages 62-72 The same interpretations can be gained from just
looking at Figures 4-7 (i e., that only slag wool is a reliable maker, and then only for
greater than 10% WTC dust in an indoor sample)
In the fourth paragraph on page 62, values are given for the mean and standard deviation
for the two spiking materials followed by a value for "nominal background level of
7,190 which I interpret to be the background However, I have no idea where this
number came from Table 1 gives values of 35,950 or 17,740 for the background values
Following this is a discussion on theoretical slopes for regression lines for the two spiked
sets of samples, but yet no mention is made of theoretical intercept It appears the sample
the USGS spiked (a non-impacted sample from NE Queens Maid service, Appendix B)
contained no slag wool, so the theoretical intercept would be zero This theoretical
intercept is confirmed by the experimental data produced by the USGS m Figures 2 and
3 Applying this zero intercept most certainly would effect the results shown in the
statistical analysis section
I was also troubled by removing labs from the data analysis (i.e., just throwing anyway
this data) There is a reason the labs obtained these numbers. And, even though
individual labs are given by number, the labs participating in this study all have
experience in doing these sorts of analyses Also, the three labs that were thrown out
were all commercial labs, while lab H, another commercial lab, did not perform as well
as the three government labs and the one commercial lab deemed to be of higher quality
So if these labs have trouble, won't all commercial labs? Possibly, as I stated above, the
problem was in the sample preparation and not in the analytical work It also seemed
strange to me to exclude data in other parts of the report, such as the two high slag wool
values that are excluded from the data in Table 1 and Figure 4 Rationalizing throwing
away data so the results "fit" is troublesome at best
-------
Mickey Gunter
Q7) EPA has observed differences in slag wool concentration which discriminate between
background dust and dust contaminated with WTC residue Do you agree that the data and
analysis presented in the documents support this observation? Please explain your answer
Yes I think it is fairly clear from the data in Table 1 (p 14) that the proposed method
herein can be used However, as pointed out above, this method will only work to a
detection limit of 10%, or so, for WTC indoor dust in NYC indoor dust I am not sure if
that sensitivity is sufficiently low
Q8) Is there any other way in which the proposed signature marker(s) can be used to determine
the extent to which WTC dust may have penetrated and remains in indoor environments?
Yes, but I think it would require considerable more research I can only speculate on
some of the uses because I have not studied the WTC dust in detail, but I can offer some
ideas
As pointed out in the documents, gypsum will dehydrate to anhydrite with sufficient
heating Possibly anhydrite might occur in some of the combusted products from the
WTC and could be used as a marker for that portion of the debris In fact, both gypsum
and anhydrite show up in an X-ray diffraction pattern shown in the USGS (2001) open
file report (ofr-01-0429) on the World Trade Center (Of course a study would need to be
done to determine the background levels of anhydrite in New York residences, but I
assume there would be very little )
Even though outdoor dust is less of a concern than indoor dust, I think it would have been
useful to obtain PM10 and PM2 5 filters and analyze them to gam a better idea of the
composition of New York dust before and after September 11, 2001 I assume the state
of New York would have these filters; I have done similar work in Idaho and we obtained
the filters from the Idaho Department of Environmental Quality
It seems logical to me to better characterize the materials that comprised the World Trade
Center, especially the slag wool, concrete, and gypsum This could be accomplished in
two ways 1) obtain construction records and 2) obtain samples of the building which are
now in a landfill. There might have been some unique property of these constituents that
would earmark them to WTC dust For instance, a special quarry may have been opened
solely to provide aggregate for the concrete, or there may be some special chemical
characterizes (possibly the trace element composition) of the slag wool or a similar
unique composition to the gypsum or the window glass
Q9) Are there any additional comments or concerns about this study that have not been
addressed by these questions?
I have several comments that did not seem to fit in the above I appreciate having this
section for those comments Hopefully these comments will be helpful in this work and
possibly in future complicated analytical problems. My comments are intended to be
constructive criticisms, but some are rather blunt and should not be taken personally by
-------
Mickey Gunter
those who have done the work Many of these people have been thrown into these issues,
forced to work under political and time pressures, and may not have in-depth
backgrounds in the issues in which they are trying to find solutions I think working
under these conditions is problematic in itself and will not yield the type of results that
our society deserves
Polarized light microscopy (PLM): I was taken aback, and do not believe the statement
on p 49, second paragraph that states "The PLM analyses were curtailed because it
became obvious that PLM could not adequately differentiate between fiber types " The
only obvious aspect of this statement is that whoever concluded this has little to no
knowledge of how to use a PLM or how to interpret the optical properties of materials'
Refractive indices of materials are very sensitive to chemical composition and because
the chemistry of the fiber types differ, so too will their refractive indices Observation of
the spectra of the slag wool given in the USGS open file report (2005-1165) showed
similar compositions, so one would assume the refractive indices of the slag wool might
not vary that much Possibly someone has looked into these relationships, but without
being told this was considered it is hard to make that assumption In fact, the PLM might
be a much more useful method for differentiating among the different fibers and
obtaining slag wool content than SEM I think it would be worth pursuing this area of
research with a skilled microscopist who understands the relationships between the
optical properties of materials and their compositions
Likewise, the optical properties of gypsum and anhydrite could be used to differentiate
these two minerals Similar methods as those with the SEM and percent area detection
could be developed Gypsum and anhydrite can also be easily distinguished form each
other optically, while this would be almost impossible with chemical data from the SEM
And if anhydrite could be related to combustion products of the WTC dust, it could also
be used as a marker.
Use of multiple analytical methods and different markers At the outset of this
section, I want to state that Greg Meeker and the USGS staff have done a wonderful job
over the past 3-4 years in developing and employing micro-analytical methods to help
understand societal issues surrounding Libby, Montana amphiboles and now analysis of
the WTC dust However, I think other methods could also be used to help characterize
and fingerprint WTC dust As I stated above, I have done no research on WTC dust, but
I often work in similar areas. For instance, I just finished a project where I used PLM
(for particle identification and morphology), SEM-EDX (for particle morphology,
chemistry, and identification), ICP (for bulk chemical analysis), XRF (for bulk chemical
analysis), and powder X-ray diffraction (XRD, for bulk mineral identification and phase
quantification) all integrated into one project It seems like other analytical methods
could have been used in conjunction with SEM work. Clearly, PLM could have been
used, but it appears those attempting to use it were not knowledgeable on how it might
have been used
In geological research, we typically ask questions like "what is the source of a mineral or
rock?" To answer this we often look for the rare and uncommon constituents in the
-------
Mickey Gunter
materials. For instance, trace elements and isotopes are often used For WTC dust,
chemical analysis by XRF or ICP might reveal a trace-element signature. Possibly
someone performed these experiments and nothing could be found, if so, this should be
noted While isotope analyses are hard to perform and expensive, they too might provide
a definite answer with less subjectivity than counting fibers Since it appears that
millions of dollars are going to spent dealing with WTC dust, it might be worth the
expense to do this type of analytical work up front
As I have stated in this section, as a practicing mineralogist I use multiple analytical
methods, and these methods change depending on the questions at hand Another
advantage of multiple methods is they can tie a research project together In other words,
one can obtain, for example, phases percentages by two separate methods When this is
done, the two values had better be in agreement. When only one method is used, there is
no way to cross-check the results Although this may seem like a silly analogy to use in
such a serious matter, it was told to me years ago when I was a mineralogy student
"When the only tool you have is a hammer, every problem is a nail " It seems to me that
this could be the case in the particle analysis business, where the only way to analyze
something is to use an electron beam instrument because this is the only "tool" the
commercial labs have
Bulk vs. particle analysis The concluding statement from the last section serves as the
introduction to this section I attended a series of talks about WTC dust in 2002 Many
of the speakers talked about their sample filters being "overloaded" and hard to analyze
I had no idea what they meant and was intrigued to discover they had too much sample to
analyze by electron beam methods As a mineralogist, too much sample is a good thing,
it opens the door for other types of analytical methods such as powder X-ray diffraction
or "whole-rock" chemical analysis by XRF Both of these methods require gram-scale
amounts of sample However, I think the problem is that most of the commercial labs
working on these "environmental" issues have cut their teeth on nuneralogical analysis
by doing particle analysis with TEMs and SEMs These are wonderful instruments and
rmneralogical knowledge has greatly advanced since their development, but they should
be integrated into a project and not the sole instrument utilized
Another major advantage of these bulk methods is they are much less subjective For
instance, in particle analysis particles are counted in a small amount of material and these
values are extrapolated to larger volumes. In the case of slag wool fibers, only 0 01 to
0 05 grams of samples were analyzed, from samples sizes that ranged from 10 grams and
higher. The fiber counts ranged from 0 to 450 (Table 5, p 58). These values are then
extrapolated to fibers/gram and such values as non-detect to 1,6200,000 fibers/gram
appear in Table 4 (p 56). Thus, the whole is described in terms of only a small portion of
the sample We in the geological community do not really use these point-counting
methods any longer if we have sufficiently large samples to perform other types of
analysis There is no doubt in my mind that bulk methods would be very useful to
fingerprint the WTC dust. Once done, the bulk methods could complement and be
integrated into particle methods and make a more convincing argument It appears to me
-------
Mickey Gunter
that the only reason to use these particle analysis methods only is that they are what is
routinely used by commercial environmental labs
Approaching multi-faceted analytical issues Lastly, I think an issue like WTC dust is
a national concern and should be treated as such, and bring to bear the scientific resources
of our entire country. If another such incident occurs, and one might be occurring
currently m New Orleans, the federal government should attempt to tap the intellectual
capital mat exits in the USA, especially among university-based researchers This is not
to belittle the government labs, there are many fine people in these organizations (e g ,
Meeker and company at USGS) who have risen to the call to help with these disasters
My statement is also not intended to be self-serving, but to point out that academicians
are much more flexible to pursue research projects and do not fear any political fallout
from their results. Also, academicians are often experts in certain fields based on decades
of education and self-selecting research, while scientists at government labs are often
thrown into projects of which they have no in-depth background For instance, in issues
surrounding Libby, Montana, I worked with some individuals whose backgrounds were
in organic chemistry While they were competent scientists, they had no background in
mineralogy I could not imagine if I were thrown into some issue that would require me
to have an in-depth knowledge of the nomenclature and analytical methods used in
organic chemistry' I propose that EPA, USGS, and other government agencies work to
set up a program with the academic community to bring to bear their expertise to work in
concert with the government labs on these important issues I am sure many faculty, and
their students, would be very interested in working on such issues of societal importance
-------
PRELIMINARY COMMENTS
FROM
ERNEST E. McCONNELL, Ph.D.
-------
Ernest McConnell
Ernest E. McConnell, D.V.M., M.S. (Path), DACVP, DABT
President, ToxPath, Inc.
Office Telephone/FAX 3028 Ethan Lane
919-848-1576 Laurdane Est
Raleigh3NC27613
7 September 2005
To Eastern Research Group, Inc
Subj Determination of a Diagnostic Signature for World Trade Center Dust using Scanning
Electron Microscopy Point Counting Techniques
I have reviewed the subject document and offer the following comments per the charge
Basis for Development of a Signature
Ql) The following criteria were established to assist EPA m the selection of appropriate
constituents in dust to be characterized as WTC signature constituents
a) They are present at levels unique to WTC dust (distinct from urban dust,
b) They are persistent for many months (not volatile), wet for long periods,
c) They are sufficiently homogeneous in WRC dust, and
d) Available analytical methods are able to detect these screening materials with a
small sample size, low minimum detection limit, and low interference from other
dust samples
Based on information in the supplied documents, and any other knowledge you may bring to the
table, are these criteria adequate for establishing a WRC signature marker? If the answer is no,
please elaborate
Answer. The criteria appear reasonable and adequate to identify a signature WTC component
that will differentiate it from background dust
Documentation of the Existence of a Signature
Q2) Based on information in the supplied documents, and any other knowledge you may bring
to the table, do you agree or disagree with the conclusion that slag wool alone meets the en ten a
as a WRC marker? Please explain your answer
Answer From the material provided, the selection of slag wool should be suitable for most
cases. None of the other possibilities are as good and using them would be redundant in most
cases However, I would like to point out that slag wool is somewhat soluble in physiological
saline and biological environments While most of this solubility data has been derived to
-------
Ernest McConnell
answer biological questions, e.g how long slag wool fibers persist in the lung after inhalation,
the data probably have some relevance to the WTC analysis For example, in in vitro systems
for solubility, slag wool dissolves at a rate of!19± 41 ng/cm2/hr (IARC, 2002). In the lung slag
wool also disappears over a relatively short period of time, albeit this is also a Junction of
breakage and physiological removal in addition to dissolution For example, the reported half-
life of fibers >20 um L in the lung is only 8 1 days (Bernstein et al, 1996) Therefore, it is
imperative that the EPA (or others) establish how stable slag wool is in damp and wet
environments and factor this into its sampling protocol and evaluation of the findings While
this does not dissuade me from agreeing to the use of slag wool as a signature mineral, it needs
to be factored into the evaluation of a given dust sample, i e was it from a wet environment? At
a minimum, the EPA, etc should do a solubility study of pure slag wool and WTC dust
containing slag wool to see the impact of a water environment, i e does it dissolve and if so, how
rapidly? If such studies showed that slag wool dissolves, then one might want to use a different
endpomt, e g cement, as the signature material in cases where the dust was subjected to water
for some period of time The problem of relative solubility should be answered prior to
establishing a sample analysis protocol
Analytical Method Development:
Q3) Are the analytical methods written in sufficient detail such that a qualifying laboratory
could follow the methodology and obtain valid results without supplemental assistance from
EPA or other sources?
Answer The methods are written in enough detail and should easily be adopted by a testing
laboratory This is not a difficult analysis to conduct However, I question the need for
scanning electron microscopy (SEM) to identify the slag wool All of the man-made vitreous
fibers (MVF) of concern (slag wool, rock wool and fiberglass) are visible in the light microscopy
(LM) range, but just as importantly, there is a minimal amount of these fibers in the
submicroscopic range (<0 25 um) Finally, as noted in 12 1 (pg 37) of the Final Report, the
refractive index using polarized LM can differentiate mineral wool fibers (slag and rock wool)
from other MVFs. An option that might be worth pursuing is to first examine a given dust
sample with polarized LM using a refractive index cutoff of 1 55 If the number of these fibers
doesn 't meet the "cut-off" criterion for a WTC dust sample, one could stop there. If it exceeds
the level of concern, then it could be further analyzed with SEM to determine whether it is rock
wool or slag wool or some other MVF But, considering the objectives of the dust analysis, SEM
may still be "overkill" even in this case. The use of LM alone would certainly cut down on the
amount of time required and significantly reduce the cost
Q4) If the answer to Question 3 is no, what items or information could be added so that a
qualified laboratory could follow this protocol and obtain valid results without the assistance of
the EPA or other sources.
Answer: See above comment I would add, however, that evaluations of this import should only
be conducted by a highly qualified laboratory and should always be validated by the use of
positive and negative control samples (blind).
-------
Ernest McConnell
Method Validation Study
QS) The method validation study design entailed spiking characterized background dust with
characterized WRC dust at various levels, and then sending 32 blind samples (16 originals and
16 duplicates) of background and spiked dust to each of eight laboratories (5 commercial
laboratories and 3 government laboratories). These laboratories characterized the dust using the
protocol descnbed above and then sent the results back to EPA Was this an appropriate design
to achieve the goals of this method validation study ?
Answer This is a standard and acceptable approach However, I was surprised by how much
difference there was in the results among the various labs One explanation that was provided
was that the mixing may not have been adequate Since this was such a critical part of the
protocol, it is perplexing why this simple procedure would have been a problem
Q6) Did the EPA and Versar adequately evaluate and interpret the results from the eight
laboratories, as documented by the supplied documents on NYC dust analyses'?
Answer I think the results were properly interpreted However, before determining the "final
cut-off" for the number of fibers to be used as the level to identify a WTC dust sample, the
Agency may want to revisit this data and censor the data from the labs that are obvious outliers
as was done by Versar, rather than using an average of all of the labs
Dust collected from currently occupied buildings is expected to have lower levels of the key
WTC constituents as compared to dust sampled near September 11, 2001 in time or sampled
more recently but in uninhabited heavily impacted buildings EPA will use the results of this
method validation study to determine the final distinguishing concentrations for the WTC
markers (s) If currently sampled dust has this marker(s) at or above such a distinguishing
concentration, EPA would consider the sampled dust to "contain residues of WTC dust" for
purposes of estimating the geographic extent of WTC impacts and making cleanup decisions
The key requirement for a distinguishing concentration is that it be adequate to distinguish
between dusts that do not contain WTC residues from those that do with a reasonably low false
positive error rate
Q7) EPA has observed differences in slag wool concentration which discriminate between
background dust and dust contaminated with WTC residue Do you agree that the data and
analysis presented in the documents support this observation? Please explain your answer
Answer. It is apparent that there are differences in the WTC dust and background dust It also
seems reasonable to use slag wool as a signature component However, I think the "cut-off"
level needs to be better substantiated, especially in light of the findings of the inter laboratory
pilot study This can be addressed to some degree by using a conservative number In addition,
one might select one number to identify environments that require radical clean-up but a
different number for minor amelioration Finally, the issue of slag wool solubility as noted
above, and its effect on potential fiber numbers needs to be considered
-------
Ernest McConnell
Q8) Is there any other way in which the proposed signature marker(s) can be used to determine
the extent to which WTC dust may have penetrated and remains in indoor environments?
Answer It is imperative that the "sampling person" clearly identify any preexisting source of
slag wool This was clearly identified as a possible problem by analysis data from the EPA
Research Triangle Park samples which showed high levels of slag wool due to natural
contamination from their ceiling tiles
Has any effort been made to see if there is a simple chemistry method that would identify a
signature marker? For example, if cement is the major component of WTC dust, there would be
a significant level of calcium in such a sample. Slag wool is also calcium rich (38%) Would
calcium be in, or as high in background dust? If calcium levels qualify as a signature marker,
this methodology would certainly be a much more efficient way of answering the two primary
purposes of the survey, i.e 1) determine the geographic extent of the remaining WTC dust, and
2) determine the need for clean-up. A cut-off level of calcium could be established as was done
for slag wool.
Q9) Are there any additional comments or concerns about this study that have not been
addressed by the questions''
Answer I have one additional comment The Final Report (Tbl 1, pg 14) suggests that fiber
number will be used to establish the level of concern In my experience the "total" fiber number
ofMVFs can be misleading unless one knows the length distribution For example, the number
of fibers in a given mass of dust would be much higher if the fibers are short than one where they
are short This may be important in a scenario such as the WTC because the large fibers would
tend to settle out sooner than the short ones, even though the total numbers may not be that
much different For this reason I think it would be better to us the mass of slag wool rather than
fiber numbers per gm of dust This could be expressed as a percent of the dust or ppm as has
been done in the USGS document (Tbl 1).
References:
Bernstein et al (1996) Evaluation of soluble fibers using the inhalation biopersistence model, a
nine-fiber comparison Inhal. Toxicol. 8.345-385
IARC (2002) Man-made Vitreous Fibres IARC Monographs on the Evaluation of Carcinogenic
Risks to Humans (Vol 81). International Agency for Research on Cancer (World Health
Organization), Lyon, France
-------
PRELIMINARY COMMENTS
FROM
SHU-CHUN SU, Ph.D.
-------
Shu-Chun Su
COMMENTS ON
FINAL REPORT ON THE WORLD TRADE CENTER (WTC) DUST SCREENING STUDY
(CONTRACT NO. 68-C-02-060 TASK ORDER NO. 107)
Shu-Chun Su, Ph D
Research Center
Hercules Incorporated
Wilmington, Delaware 19808
September 23, 2005
Summary
The vahdilj of the key analytical results (number of slag wool fiber per gram
of dust) is highly questionable because of
- the lack of any standard for evaluating the accuracy of results.
- the incorrect equation used foi Hie calculation of final results as a result
of the ignoring of slag fiber depletion through sieving and the required onc-
mmute coarse particle precipitation during sample preparation;
- the unprovcn reliability of differentiating the signature constituent slag
wool from the interfering constituent rock wool basing on the detection of low
concentration of Fe by energy dispeisive spectroscopy analysis with scanning
electron microscopy, and
- the huge extrapolation factors of 5,000 to 10,000 for calculating the final
results for 1 gram of samples from the slag wool fiber count obtained from the
measurements of 0 0001 to 0 0002 gram of samples
Therefoie, the conclusions in the Final Repoit need to be re-evahiated when
valid analytical data arc available
Ql) The following catena were established to assist EPA in the selection of appropriate constituents in
dust to be characterized as WTC signature constituents-
a) They are present at levels unique to WTC dust (distinct from urban dust),
b) They are persistent for many months (not volatile),
c) They are sufficiently homogeneous in WTC dust, and
d) Available analytical methods are able to detect these screening materials with a
small sample size, low minimum detection limit, and low interference from other
dust components
a) Yes
b) Yes, if not significantly diluted by urban dusts
c) Yes However, the homogeneity of signature constituents needs to be clarified and better
defined If it means only the concentration, u should be expanded to include also the chemistry and
-------
Shu-Chun Su
physical pioperltes used by the analytical protocol to identify and differentiate them from interference
constituents. Otherwise, if those diagnostic properties vary too much reliable identification will be rather
difficult, resulting in unacceptable rales of false negatives and/or false positives The reason that I
emphasued the inipoiiance of the homogeneity of chemical and physical pioperties was because I noticed
the limited effort to characterize the chemistry of SW (slag wool") and other interfering MMVF
constituents and the complete lack of effort to characterize their optical properties, such as refractive
indices, during the initial phase of the method development. Tt was nor clear that besides the 3 SW EDS
(energy dispersive X-ray spectroscopyj spectra presented in the ATLAS' how many more SW and RW
(rock wool) fibers have been analyzed to deleitnine their lespective variation ranges of Fe, which is the
key element used to distinguish SW from RW by the analytical protocol, using a more accurate and
reliable technique than EDS. I will discuss tins issue in more details later
d) This is not a well-defined question There is no wa> one can answei this question without
knowing
- how small a sample sue is consideied to be small" x micrograms. x milligrams. 01 x
grams0
- how low a minimum detection limn i& considered to be low1' 100 ppm, }%. or 10.000
fibeispcr gram?
- what does n mean by "low interference'"' The number of interference constituents' The
concentration of interference constituents,9 The degree of difficulty to diffeientiate
interference constituents from the signature components;?
1 believed that two more criteria for signatuie constituents should be added
ti) Tim varitmotis »/ their diemiciil and physical properties are within boundaries time >..<»!
be clearly defined and reliably differentiated from thiise t)J potential interference
conktituetils ur apntdtaerm'med confidence level (95%, Qy%. 99.9% etc.)
f) Standards i>j various- concattratlfmst L-an be created lining the pure signature constituents
far assessing if thf analytical mult* can meet a predetermined M{K) (Mea^uri'meni
Qualify Ubjet-tive), suck as ±30%.
1 consider that the flaws 1 found in the WTC dust screening method development and validation
studies, which will de discussed in greatei details, are parualh due to the failuie in the recognition of the
above two criteria
1 Lowers, H A , Meeker, G P , I K Brownfield (2005) World Trade Center Dust Particle Atlas U S Geological
Survey Open-File Report 2005-1165
-------
Shu-Chun Su
Q2) Based on information in the supplied documents, and any other knowledge you may bring to the
table, do you agree or disagree with the conclusion that slag wool alone meets the critena as a WTC
marker?
Tlie current study and results only show that S\V (slag wool) MIGHT BE a good candidate to
meet the first 3 criteria
Q3) Are the analytical methods written in sufficient detail such that a qualifying laboratory could follow
the methodology and obtain valid results without supplemental assistance from EPA or other sources7
The FR2 seems to have already provided a clear answer to this question Among the five
commercial laboialones contracted for ihis study. RJ Lee. MVA, EMSL, MAS, and Reservoir, EPA
determined that throe of them (Labs E, F. and G) had failed to produce data of acceptable quality
It can be assumed that there must ha\e been a rigorous process in selecting contract laboratoiies
for this important study All five laboiatoncs must have been considered to be highly qualifying ones
among many environmental and analytical laboratories in the U.S , \\hich have not only PLM (polarized
light microscopy) but also SEM (scanning election microscopy) and EDS capabilities.
It is an alarming fact that 60% of highly qualifying commercial laboratories selected by fcPA for
this important study have failed to produce valid lesults
It is equally alarming that Q3 seems to correlate the inability ro follow the analytical protocol and
produce valid results ONLY with the insufficient detail of the documented analytical methodology as if
had the analytical protocol been documented with mote details the high late of failure would not have
happened
There could be other reasons foi a laboratory to fail lo pioduce valid results besides the
insufficient detail of the documented analytical methodology.
- the lack of competence of the designated analyst(s) in performing the written analytical
methods.
- some inherited flaws in the analytical protocol itself or
- everything above
But 1 certainly could not agree that "some Infa did not have the personnel or the equipment to
perform the required analysis in the given timeframe" ip 13 of FR) was one of the reasons It is a well-
known fact that EPA would not have selected a laboratory without the requited equipment ("PLM. SEM.
and EDS) and the EXCEL data files provided by EPA for all 8 laboratories, including Labs E F. and G,
showed that every laboratory has the required equipment. The lack of personnel should not have been an
issue either. Any laboratory that had been awarded this important EPA contract would have to fulfill its
obligations by commuting adequate personnel to meet any deadline set by EPA.
2 USEPA et al (2005) Final Report on the World Trade Center (WTC) Dust Screening Method Study 72pp
-------
Shu-Chun Su
Since Q3 also emphasizes "withuut supplemental assistance Jrom EPA or other sources,'' lei's not
forget the fact that "Each laboratory hem attended a fu
-------
Shu-Chun Su
to ensure a repioducible particle loading whether by the same analyst or duTerent analysts Even a few
seconds difference, e g, 2 seconds vs 10 seconds after the one-minute waiting period, in collecting the
aliquot with a pipette may result in different particle loadings thai are significantly enough to affect
accurate calculation of the sample weight on the SEM stub Another inherited uncertainty of this
procedure is that the exact position of the pipette tip in the liquid column will also affect the amount of
particles in the 10-nL aliquot because there is a particle concentration gradient from the top to the bottom
What makes the execution of the analytical piocedures more complicated is dial tlieie ts another
step of aliquot extraction from the initial suspension to prepare the said 10-fold diluted suspension The
method developers assumed that the sample loading in each aliquot could be accurately calculated fiom
the ratio of the ahquor volume to the volume of the sampled suspension. In ideal world such assumption
is undeniably correct In real world, however, such assumption should not be taken for granted and must
be rigorously tested and verified so that the variability of sample weight on sample stub can be
determined and used to icfine the final result calculation
Q5) The method validation study design entailed spiking characterized background dust with
characterized WTC dust at various levels, and then sending 32 blind samples (16 originals and 16
duplicates) of background and spiked dust to each of eight laboratories (five commercial laboratories and
three government laboratories) These laboratones characterized the dust using the protocol described
above and then sent the results back to EPA Was this an appropriate design to achieve the goals of this
method validation study''
This would have been an appropnate approach if appropriated standards with known
concentrations of SW are also analy/.ed by each participating laboratory to determine whether the ±30%
accuracy objective has been achieved. One of the critical flaws of this study was the lack of any standard
Therefore, the MQO for accuracy (±30%) stated in A7 2 of QAPP have been completely ignored by the
study The Project Management should have caught tins nonconformancc with QAPP and taken
appropnate corrective action to address this issue of utmost importance
Q6) Did EPA and Versar adequately evaluate and interpret the results from the eight laboratories, as
documented by the supplied documents on NYC dust analyses'
Dust collected from currently occupied buildings is expected to have lower levels of the key WTC
constituents as compared to dust sampled near September 11, 2001 in time or sampled more recently but
in uninhabited heavily impacted buildings EPA will use the results of this method validation study to
determine the final distinguishing concentrations for the WTC marker(s). If currently sampled dust has
this marker(s) at or above such a distinguishing concentration, EPA would consider the sampled dust to
"contain residues of WTC dust" for purposes of estimating the geographic extent of WTC impacts and
making cleanup decisions The key requirement for a distinguishing concentration is that it be adequate to
distinguish between dusts that do not contain WTC residues from those that do with a reasonably low
false positive error rate.
I disagreed
-------
Shu-Chun Su
1) The evaluation and interpretation as well as conclusions ptesented in the FR were all based on
two assumptions.
- the SWG (SW Fiber couni per gram of duslj dala of the "Very Good" Gioup (Labs A, B. C D.
and PI) were valid and could be used to lehably validate the analytical protocol.
- the SWG data of the "Outlying" Group (Labs E. F. and G) were invalid and should be discarded
because they didn't fit with the results of the "Very Good" Group
A.re we sure that the SWG data of the "Very Good" Group arc valid''
Even though QAPP has detailed procedures for quality control (B5. p. 16), data validation (D,
p 19), etc , it seems thai there have been no serious effort to check whether the equation used to calculate
SWG was correct and all the assumptions adopted by the analytical protocol were valid
The following equation is from Section 13 0 (Data Analysis and Calculations) of Appendix D
Fibers slag wool/g on slide * e after sieving * g sample after ashing = Total f/g of sample
g before sieving * g sample before ashing
Because the FR has not presented an equation for calculating the fiber counts based on SEM/EDS
analysis, 1 guessed that Fibers slag wool/g on slide in the equation can be icplaced with Fibers slag
wool/g on SEM stub
This equation implies at least two assumptions
a) all SW fibers would be in the fraction < 150 ^n after sieving Since it is a grven fact that there
are SW and other MMVF fibers longer than J50 |jni and sieving has nevei been a ven effects e wa> to
separate fibrous material by their length, there must be SW and other MMVF fibers not only longer but
also shorter than 150 pin in the coarse fraction 'Therefore, this implied assumption and the above
equation can't be correct Because all calculations of the SW fiber count results were based on this
incorrect equation, the validity of all SW fiber residts presented in the FR is highly questionable even
though the SW fiber count ratios among the 1%, 5%, and 10% spiking samples showed a marginally
reasonable linear relationship for the five "Very Good" laboratories
Had this factor, i.e , //*- ;'«"<; ofWd in ~'/.5tf prr, froi'tim (') AW ' in --./JO /en/iMtiun, been a
constant for every single one of the 20 background samples and 12 spiked samples prepared by FPA.
there would be no need to make any correction to the existing SWG data Such a possibility is most
likely not true
b) all SW fibers would remain in the diluted suspension after letting coarse particle precipitate for
one minute during SEM sample preparation because there is no correction factor in the above equation to
address the depletion of SW fibers in the suspension after the one minute precipitation period It can be
safely concluded that the coarse fraction at the bottom of glass vial contains also SW fibers If the
-------
Shu-Chun Su
depletion of SW fibers in the suspension due lo the piecipitauon duruig the one-minute waiting period is
not factored in the calculation ot'SWG, the results could not be \alid
Had this factor, i e , the ratio vfSWG .>«/vwi/vi £#<
-------
Shu-Chun Su
very low concentration level. The selection of BIR-1G as the sianclard for EDS precision assessment is
therefore not adequate if SW is the only signature constituent to be used for screening WTC dust from
background dust because of BIR-lG's high concentration levels of FeO (8.34%) and Fe20? (2.06%)9. A
proper standard should have FeO Fe2O3 (EDS can not differentiate Fe:' from Fe") concendation no
more than 5%wt. An additional standard with FeO+ Fe20? around l%wt will also be helpful.
Slag wool
Rock Wool
Fig. 1. The comparison of EDS spectra of SW and RW (ATLAS: SLAGWOOL-02.TIF
and ROCKWOOL-02.TIF)
Smith, DB (1998) USGS Certificate of Analysis: Icelandic Basalt, BIR-1.
-------
Shu-Chun Su
Q7) EPA has observed differences in slag wool concentration that discriminate between background dust
and dust contaminated with WTC residue Do you agree that the data and analysis presented in the
documents support this observation? Please explain your answer.
No, because such an observation would be valid only if the SWG results were valid As shown
by my comments for Q6,1 am not convinced that the SWG results used to form this observation are valid
Q8) Is there any other way in which the proposed signature marker(s) can be used to determine the extent
to which WTC dust may have penetrated and remains in indoor environments?
1 don't know
Q9) Are there any additional comments or concerns about this study that have not been addressed by
these questions'7
1) The reliability of differentiating the SW from the interfering RW basing on the detection of
low concentration of Fe by SEM/EDS
EPA relied on the geneiali/ed data provided by TIMAlu for the chemical and physical pioperlies
of MMVF fibers in WTC dusts and believed that SEM/EDS could be used to reliably differentiate SW
whose FeO varies from 0 to 2%v\l. from RW \\hoss FeO varies between 3 to 12%wt Instead of
conducting a comprehensive study of the compositional variations of MMVFs. especially SW and RW. in
WTC and background dusts to determine their actual elemental compositions. EPA based its method
development work on the above FeO concentration data provided in a single TIMA document I believed
that this was a critical flaw of the method development
Unlike those strictly formulated products, neither SW fibers nor RW fibers ore known for the
stability of then chemical and physical properties because the diversity of the raw materials and the not-
so-stnngcnt specifications of the final products The data in the table" on the ne\t page indicates that
RW's Fe2Oi content could be as lov> as 0.30% If RW's Fe2C>3 content could be this low, Fe is no longer
a reliable diagnostic element for SW-RW differentiation Tins example shows that there might be a
possibility of overlapping FeO%wt ranges for SW and RW. resulting either false negatives or false
positives tn SW counts
There should be a very rigorous and comprehensive study of the compositional variation of
MMVFs in WTC and background dusts, especially SW and RW fibers, to establish the variation range of
Fe content at the very beginning of method development Many polished sections of dust samples should
be prepared and analyzed by WDS (wavelength dispersive spectroscopy) iising a microprobe so that the
10 Thermal Insulation Management Association, 1991, Nomenclature of Man-Made Vitreous Fibers TIMAInc 72p
11 Commission on Life Sciences (2000)Review of the U S Navy's Exposure Standard for Manufactured Vitreous
Fibers hup. //www n.?r;.edii/l)ix.Ks/(.'3(^p700375iln:J/15.hUnt
-------
Shu-Chun Su
ACTUAL chemical compositions of various MMVF fibers m WTC dusts can be accurately determined
and the variation ranges of Fc, the key element used to differentiate SW fibers from RW fibers, can be
accurately established These data are indispensable to determine \\helher Fe can be used to reliably
differentiate SW from RW at a piedeleimined confidence level
Name13 11
Class Glass
Components
TABLE 2-1 Typical Chemical Composition of Some Commercial MVFa
Composition (%)
A
Glass
C
Glass
21
Rock
F
Rock
G
Rock
22
Slag
RCF-1
RCF
X-607
RCF sub!
Insofrax
RCF sub «
SiO2
Fe203
TiO2
A1203
CaO
MgO
Na2O
K20
B2O2
P205
S03
Cr203
MnO
ZrOj
Total
6340
030
006
390
740
280
1540
1 30
450
000
030
000
001
000
9940
6500
010
002
1 90
740
260
1610
070
470
1 10
003
000
000
000
9960
6170
010
002
100
720
290
1610
060
920
1 10
020
000
001
000
10000
4630
1320
260
1350
1000
910
3 10
1 40
000
040
000
004
020
000
9980
5630
0.30
0 10
320
26 10
640
320
070
000
290
000
000
000
000
9910
6010
610
005
040
1880
830
550
020
000
008
005
000
000
000
9950
3840
000
050
1060
3800
990
040
050
000
000
1 80
000
070
000
10080
4770
100
210
4800
007
008
000
020
001
010
000
003
000
010
9940
5830
010
005
1 30
3870
040
030
010
000
040
000
000
000
000
9930
7620
030
008
1 40
020
21 50
007
0 10
000
003
000
000
001
000
9990
"Data denved from Bernstein et al (1996), Maxim etal (1999a), andMcConnell etal (1994, 1995, and 1999),
Material Safety Data sheets for Isofrax fibers from the Umfrax Corporation, Niagara Falls, NY
*Name 11, Certain Teed B glass wool fiber, A, new glass wool, C, new glass wool, 21, rock wool, F, rock wool,
G, rock wool, 22, slag wool, RCF-1 - kaolin-based refractory ceramic fiber, X-607, rock wool produced by
Umfrax, Isofrax, refractory ceramic fiber
'Substituted RCF
Another practical concern is the criteria used by SEM analyst to differentiate SW EDS spectrum
ftoni that of RW dui ing routine analysis As shown in Fig 1 which is entirely based on EPA's data, how
could an analyst make a reliable identification simply from the visual inspection of the Fe peak height1?
2) Problematic sampling procedures
- the lesult of a suigle 12 nun SEM stub with 10 to 20 jug of fine dust samples was used 10
determine the total SW fiber counts in one gram of dust samples The extrapolation factor was too high
(5,000 to 10,000) to ensure the accuracy of the final SW fiber counts The data in Table 1 clearly show
the effect of this excessive extrapolation the error propagation is most pronounced for low concentration
samples
10
-------
Shu-Chun Su
- the sample pteparation piocediire contains two steps of aliquot sampling the first one was fiom
the initial dust suspension and the second one was from the diluted dust suspension As discussed
previously, the accuracy of dust weight calculation based on the volumes of aliquots is highly
questionable
- the "wiii! <)ne mimile" problem as discussed in the comments for Q4
3) The PI. M procedure has been ruled our as a feasible method for SW fibei quantification. I
consider that this conclusion is premature and should be re-examined because if the fundamental flaws'"
in the method development of the PLM piocedure can be collected there might be a possibility thai PLM
could be a more accurate, efficient, and cost-effective technique than SEM/EOS. However, this could
not be accomplished if the method development team has no experts with stiong background in optical
crystallogt aphy, optical mineralogy, stercology. and automated image processing and analysis
12 The discussion of this topic could be another lengthy review
11
-------
PRELIMINARY COMMENTS
FROM
JAMES S. WEBBER, Ph.D.
-------
James S Webber
Peer Review of
Final Report on the World Trade Center (WTQ Dust
Screening Method (August 17, 2005)
Prepared by
James S. Webber, PhD
September 23,2005
Page 1 of 12
-------
James S Webber
Qt)
in tft/s* f o be vkam&ettzeti 3$ WTO signatory ponstttventiK
fy They&e p^sfere/tf for many #*&*&& && vc^fe};
cj TAey ar« stfffictBntty homogeneous in WTC dust; and
sampt&sfz&, tow minimum detection firnit, and low interference tram other dtist
components,
Basedan information m tn& suppled documents, and any otfi&r knowledge you tnay bring to the
table, are these criteria adequate for establishing a WTC signature marker? if the answer/* no.
While rather broad, I agree that these criteria should be adequate for establishing a WTC
signature marker
However, the use of low minimum detection limit invites controversy In a strict analytical
sense, minimum detection limit is predicated on a level of acceptable uncertainty, or precision To
many microscopists, however, detection limit refers to a microscope's capability to detect a single
particle of interest in a field of other particles The difference between the two interpretations is
usually several orders of magnitude because, despite the extreme sensitivity of the microscope,
the reproducibility is often poor because of non-uniform particle distribution and analyst
subjectivity, among other factors The poor precision of the proposed method is discussed later
Page 2 of 12
-------
James S Webber
Based on fafoansiiQn in fre suppferf documenfe, and any other kftovdedge you mayteincf to&e
tabte, do you agree or disagree wSh the condosiort fret stag woof atone meets the craters a$ a
WTC matk&r? Pfease exptaiityour answer
While I concur with the Final Report's conclusion that gypsum and concrete dust are too
ubiquitous to be useful fingerprints, I disagree with the conclusion that slag wool alone meets the
criteria as a WTC marker Specifically, Ql)a) requires a marker to be unique to WTC dust Slag
wool is a common building material which can be expected near any building demolition site,
near major renovations, or even where a workman's ladder has brushed ceiling tiles or insulation
The second paragraph on page 9 points to an overlap of slag wool (around the 100,000
fibers/gram concentration) in samples from both the impacted and background areas The
elimination of NJ and LI "outliers" from Table 1 (page 14) is another indicator of slag wool's
inadequacy as a solo marker Simply put, false positives are guaranteed because of slag wool's
broad application in past and present building products, as acknowledged twice ( likely that
false positive results ) in the report, on the last sentence on page 12 and again in the final
sentence before the Final Report's Conclusions on page 21
Page 3 of 12
-------
James S Webber
Q3) Are the analytical methods written in sufficient detail such that a quatifyfng laboratory could follow
the methodology and obtain valfd results without supplemental assistance from EPA or other
sources?
On the basis of the poor interlaboratory results (evidenced by the disqualification of three of
eight (almost half!) of the laboratories), I'm skeptical that the analytical methods applied as-
written by a typical laboratory would produce valid results The five surviving laboratories are
laboratories with stellar national and international reputations resulting from substantial
experience in non-routine research projects Yet the high inter-laboratory relative standard
deviations (RSD) from these elite laboratories on page 20 indicate how difficult it will be to
derive meaningful results if there is a need for newly initiated laboratones Despite carefully
controlled conditions during the study, poor reproducibihty was ascnbed (bottom of page 20) to
possible non-homogeneity of prepared samples and operator experience This poor precision
would indicate violation of Ql)d)'s requirement of low minimum detection limit The
acknowledgment that the protocol was adapted to suit each laboratory's equipment (top of page
10) and the need for weekly conference calls (top of page 49) by laboratory participants adds to
my doubt that this protocol will be uniformly adoptable.
Of course, the term qualifying laboratory in Q3 could be the caveat by which analyses are
limited to a select cadre of laboratones If this term is applied to laboratories that are winnowed
out in a process similar to that used to eliminate three of the eight original laboratones, results
from real-world samples might approach the quality of this inter-laboratory study
Page 4 of 12
-------
James S Webber
Q4) tffhe answer to Question 3 is no, what items or information cootd be added so that a qualified
laboratory could fofowtbis protocol and obtain vafid results without (he ass/stence oftfte gPA or
other sources?
Two SEM preparation methods are allowed in §11.5. 1) the drop method in the 3rd paragraph
of page 36 and 2) the filtration method in the 4th paragraph I would strongly recommend that the
drop method be eliminated because of the likelihood of uneven particle distribution caused by
surface tension and other physical anomalies of the drop. Filtration invariably produces a more
uniform particle distribution than produced by a drop mount. That being said, I would further
recommend that the filtration set-up in that section be revised to prevent bubble (void) formations
on the filter The filtration apparatus should be assembled dry and vacuum applied before
introducing the isopropanol suspension See precautions in Section 632 of EPA Method 100 1
(Chatfield, 1984)
The method requires that the sample be ashed, with gravimetric tracking, to remove organic
materials The use of this data is not explained WTC dust was primarily inorganic in nature,
which would imply that a high concentration of organic material in a field sample represents
substantial dilution by local dusts Unfortunately, this was not addressed in the study
Page5ofl2
-------
James S Webber
The method validation study design entailed spiking characterized background dust with
characterized WTC dust at various tevets, and then sending 32 btind samp/es (16 originate and
1& duplicates) of background and spiked dust to each of eight laboratories (five commercial
laboratories and three government laboratories). These laboratories characterized the dast using
the protocol described above and then sent the results back to EPA. Was this an appropriate
design to achieve the goats of this method validation study?
The method validation study design was appropriate for determining the suitability of a WTC
dust marker. The results, however, as detailed in my response to Q6 below, did not convincingly
demonstrate that suitability for slag wool
Page 6 of 12
-------
James S. Webber
QG) Did gPA and Versar adequately evaluate and interpret the results from the eight laboratories, as
documented by the supplied documents on /VVC dust analyses?
Dust collected from currently occupied buildings is expected to have lower levels of the key WTG
constituents as compared to dust sampled near September 1i, 2001 fn time or sampled more
recently but in uninhabited heavity impacted buildings. EPA 'will use the results of this method
vaffdation study to determine the final distinguishing concentrations foe the WTG markerfs}. W
currently sampled dust has this marker($) &t or above such & distinguishing concentration, EPA
would consider the sample dust to "contain residue* of WTC dust" tor purposes of estimating the
geographic extent of WTC impact* and making cleanup decisions. The. key requirement lor a
dtsttnguishino. concentration is that it 6e adequate to distinguish between dusts that do not
contain WTC residues from those that do with a reasonably low false positive error rate.
The fact that a linear regression fit the spiked samples' concentrations versus results reported
by the commercial laboratories was promising in that a linear relationship was expected between
these spike levels. Furthermore, using a Standard-Additions approach, extrapolation of the Spiked
Samples - Albany "Best" Group Combined sample regression (top figure on page 65) to the %
spike level baseline yields approximately (-)1.2%, which is equivalent to 7000 fibers per gram in
the background dust This is equivalent to the 7194 fibers per gram of the NE Queens background
material (from Table 3, page 55)
However, the need to eliminate various laboratories to achieve this fit is troubling
Eliminating three of eight (37%) commercial laboratories for poor performance raises questions
about the universal applicability of the method -1 would assume that the eight laboratories chosen
initially were neither bargain-basement laboratories nor chosen at random from the yellow pages.
The additional temporary elimination of another of the surviving laboratories (H in several
figures) increases concern about the rigor of the method
Only laboratories A & B performed well with the USGS spikes, raising further doubts about
the effectiveness of the method to distinguish a WTC signature Convincing demonstration of the
method's suitability for widespread use would have been inter-laboratory regressions that
matched the USGS results presented in Figure 3, where the best-fit regression line intercepted
near zero.
As to reasonably low false positive error rate, I reiterate my conclusions under Q2) they are
a given
Page 7 of 12
-------
James S Webber
Q7) EPA has observed differences in stag wool concentration which discriminate between
background dust and dust contaminated vtftb WTC residue, Do you agree that the data and
analysis presented in the documents support this observation? Please explain your answer,
The elimination of so many background samples from statistical consideration is vexing
Appendix E, Table 3 (pages 54-55) lists 28 background samples Six of these, or 21%, contain
greater than the landmark 100,000 fibers per gram With these samples included, RSD for slag
wool exceeds 150%, which effectively runs confidence limits from (less than) zero to well over
the average for 4 Albany. In other words, background samples would be statistically
indistinguishable from WTC samples The dismissal of the high-fiber background samples from
the study portends problems in the real world how will an analyst or decision maker know a
priori which high-fiber samples are simply from high background?
Page 8 of 12
-------
James S Webber
Q8) Is there another way the proposed signature martens} can be used t&determine the extent to
wfe'cfc WTC dust may have penetrated and remains m indoor environments?
Method development should have evaluated measurements of fiber lengths and widths -
WTC fibers are probably smaller than background fibers because the impact energy of the WTC
collapse shattered many Additionally, any indoor WTC dust will probably contain smaller slag
wool fibers than outdoor WTC dust because of the removal of larger fibers by settling during
dust's passage from outdoor to indoor environments, as hinted in the last sentence on page 11 It
would be worthwhile to compare fiber dimensions of USGS (outside) WTC slag wool to fiber
dimensions of 4 Albany (inside) slag wool
Page 9 of 12
-------
James S Webber
Q9) Are there any additional concerns or comments about tfi/s study that have not been addressed
by these questions?
I am obviously not persuaded that the proposed method's limitation of a WTC marker to slag
wool will be effective in determining WTC dust boundaries The convoluted disqualification of
laboratories and elimination of various background samples demonstrates the tenuous nature of
this simplistic approach. Below I make several cases for possible improvement of a WTC dust
signature
The Case for More Conditions on Slag Wool Analysis
A slag-wool-only marker would be useful in determining extent of WTC plume only with
several additional conditions
1) Investigate the particle-size distribution of slag wool in WTC dust If differences are
measured between these fibers and those in background fibers, better discrimination could
be made
2) When positive results are reported, investigate the surroundings of the sample's on gin for
potential contamination, e g., slag wool in building materials or demolition of nearby
structures
3) Participating laboratories would have to be screened with excruciating care Qualifying
critena would be similar to those that culled out almost half of the laboratories in the
original study
4) Design the survey with an overload (>25%) of QC Mix large numbers of samples with
spikes and known concentrations with the field samples to track the laboratory's (and
method's) performance in real time
The Case for Chrysotile as a WTC Dust Fingerprint
After encountenng the multiple difficulties with the slag-wool-only design, I am truly
puzzled by the absence of consideration of one of the most promising WTC markers. Chrysotile
is never mentioned in the report and asbestos occurs only once, in a reference This despite an
earlier WTC report by R J. Lee (2003) in which chrysotile was detailed as the most
important constituent of a multi-component WTC Dust Signature Chrysotile almost
certainly exceeds slag wool in meeting the marker cntena of Ql
Ql a) Chrysotile is more unique to WTC dust than is slag wool Because of the ban on most
uses of asbestos thirty years ago, chrysotile is no longer a ubiquitous component of
dust Certainly not at the levels found in WTC dust. According to the R J Lee report,
Page 10 of 12
-------
James S Webber
chrysotile is at least 4 orders of magnitude greater in WTC dust than in background
dust. Furthermore, the report continues, the chrysotile in WTC dust is often
distinguishable from other chrysolite-
1) Mg/Si ratio is significantly reduced in WTC fibers
2) WTC fiber size differs from chrysotile fibers typically found in ambient dust
a) WTC fibers are highly fibnlized (sheared apart) by the energy of the collapse,
with mean width <0.1 urn.
b) Fibers tend to be longer than in ambient dust
3) WTC fibers are often coated with residual metals (Al, Fe, Zn) released during
combustion of the WTC materials
Adding to their advantage over slag wool, the generally smaller size of chrysotile fibers
makes them more likely to infiltrate the indoor areas of concern
Ql b) Chrysotile is certainly durable - it's been around for millions of years
Ql c) Chrysotile has been detected at levels between 0 5% and 2% in ongmal analyses of
WTC dust, probably the most homogeneous distribution of chrysotile in any large-scale
environmental sample
Ql d) Chrysotile at a -1% level would be easily detected by SEM (and certainly by TEM, as
discussed below) Our TEM analysis of WTC materials in storm-sewer run-off during
the first post-collapse rainfall (September 14) revealed a substantial population of
fibers wide enough (>0 25 urn) for easy detection by SEM at a magnification of 4000X
(At that magnification, widths on the screen would be > 1 mm ) Concentrations of
fibers longer than 10 urn and wider than 0 25 um were 100 million fibers per gram
debris, most of which was the fine plaster and cement particles common in WTC dust.
This is about an order of magnitude greater than the slag wool concentrations
The Case for Transmission Electron Microscope Analysis of Chrysotile
I would further recommend that TEM be considered for analysis of chrysotile in dust TEM
holds many advantages over SEM for analysis of asbestos
1) Detection of the narrowest chrysotile fibers is possible by TEM's superior resolution To
the point, on the basis of our analysis of the run-off sample, if all chrysotile fiber sizes
were included, chrysotile concentrations would be almost 20 billion fibers per gram This
would allow use of greater dilution, which would substantially reduce the amount of
obscuring material in a field of view Using Chatfield's (1984) recommended upper limit
of loading, 20 ug/cm2, a 0.001 mm2 TEM grid opening would contain 35 chrysotile fibers
Page 11 of 12
-------
James S Webber
- an easy counting job with much less obscuring background than SEM analysis of the
same dust
2) Standardized methods exist for preparation and analysis by TEM that would require
minimum modification
3) TEM is a more standardized instrument for chrysotile analysis Differences between
TEMs are less than between SEMs, which have problems in standardization of contrast,
scan rates, etc There will be no need to have the protocol adapted to suit each
laboratory's equipment (top of page 10)
4) There exists an army of several dozen certified/accredited TEM laboratories with
extensive expenence analyzing asbestos under conditions similar to those of the proposed
protocol
5) Analysis of a fiber in a TEM preparation will probably have less interference from x-rays
from other particles because of the more dispersed sample
6) If there is any question about a fiber's identity, TEM can utilize electron diffraction to
unequivocally identify chrysotile SEM lacks this capability
Finally, the Case for a Stratified Approach
Because of slag wool's many shortcomings as a fingerprint, application of the suggested slag-
wool-only protocol should be limited to screening. If a dust sample is negative at the 10% level as
outlined in the proposed protocol, consider the dust to have originated outside the WTC dust
plume On the other hand, a positive result would trigger analysis for chrysotile using TEM
Quantitation of chrysotile concentrations along with qualitative characteristics of WTC chrysotile
would settle the question of WTC source
This stratified approach would be only moderately more expensive because methods already
exist and TEM laboratories are competitive in their pricing To paraphrase retired EPA guru,
Mike Beard, "It is folly to base a million-dollar abatement project on a five dollar analysis "
References
Chatfield, E J , and M J. Dillon. 1984 Analytical Method for Determination of Asbestos Fibers in
Water EPA-600/4-84-043
R. J Lee, 2003 130 Liberty Street Property, WTC Dust Signature Report, Asbestos.
December 2003 21 pages
Page 12 of 12
-------
Appendix C. Responses to Questions of Clarification
-------
Question 1:
Final Report, page 13, top "Three of the commercial laboratories, designated as
labs E, F and G, reported analytical data that are not consistent with other five
labs Generally, these labs were not able to distinguish differences between the
three spiking levels In addition, these labs did not meet the measurement quality
objectives (MQOs) for the spiked samples put forth in the QAPP for this study
(Appendix A Section A 7 1) Thus, the data from these three labs are not
considered in the results presented in Table I and Figures 4-7 " Please provide
details of MQO deficiencies for all eight of the original commercial laboratories.
EPA Response:
MQO comparisons were only made for the spiked sample data The following is
a synopsis of the analysis.
Intralaboratory
Once it was determined that Labs E, F and G were not able to distinguish
differences between the three levels for the spiked WTC samples, we compared
the data to the MQOs for intralaboratory precision For labs E, F and G, data for
five of the six sets of spiking duplicates fell outside of the + 30% intralaboratory
precision level All other labs did reasonably well with regard to precision Thus,
based on both of these results (not being able to distinguish between the spiking
levels and not meeting intralaboratory precision MQOs), the data from Labs E, F
and G were not used
Data for intralaboratory accuracy showed a great deal of variability This is, in
part, due to the fact that we do not have a 'standard' test dust against which we
can compare our data - instead, MQO qualification was calculated in two ways:
1) each labs average was compared to an assumed amount of 'pure dust' This
assumed amount was based on the percentage level and the original dust that was
used for spiking (i.e pure USGS dust contained 11 E+06 fibers per gram of dust,
thus, the 10% spiking level should contain 1.1 E+06 fibers per gram of dust, etc.)
and 2) each labs average was compared to the overall average for all labs for each
sample In both instances, the results showed high levels of variability, thus, the
MQO for accuracy (+ 30%) was not helpful in choosing lab data to consider in
our analysis
Two labs (A and C) failed to meet the 95% MQO for intralaboratory
completeness Note that it only took one missing value to fail meeting this MQO
Inter laboratory
Interlaboratory precision was determined by averaging the duplicate samples for
each lab, and then determining an average, standard deviation and relative
standard deviation for the set of eight labs When the three labs (E, F, and G) are
left out, the estimated overall precision of the study improves substantially
-------
Intel-laboratory accuracy was evaluated by averaging the duplicate samples for
each lab, and then determining an average for the set of eight labs. This showed a
great deal of variability. As stated previously, some of this variability is due to
the fact that we do not have a 'standard' test dust against which we can compare
our data MQO qualification was calculated by assuming the 'amount of pure
dust' per sample based on the percentage level and the original dust that was used
for spiking (i.e pure USGS dust contained 11 E+06 fibers per gram of dust, thus,
the 10% spiking level should contain 1 1E+06 fibers per gram of dust) When the
three labs (E, F, and G) are left out, the accuracy of the study improves for one set
of spiked samples (USGS, all %'s), and meets the MQO for those samples When
the three labs are left in the analysis, the overall accuracy of the study does not
meet the MQO for either set of spiked samples
The study meets the interlaboratory 95% MQO for completeness when only
considering spiked samples
Question 2:
Final Report, Appendix D §12.1 Analysis by Polarized Light Microscopy Is this
the full bench protocol that the laboratories used or was further detail provided
during interlaboratory discussions or by other means1'
EPA Response:
The full bench protocol was provided in the report It was realized part way
through the study that PLM was not a time-saving measure and thus, will not be
part of a final protocol for the proposed sampling
Question 3:
QAPP, §A5 2 Method Development (penultimate sentence) This method was
reviewed by the WTC Expert Panel's signature subcommittee and is presented in
Appendix B The Appendices list that appears on page 20 identifies Appendix B
as the PLM/SEM protocol Please provide a copy of the review referenced in
A52
EPA Response:
A formal written review was not provided by the subcommittee members Drs
Paul Lioy and Mort Lippmann provided comments by email A PDF of these
emailed comments is provided under separate cover. Mr Meeker was involved in
the development of the protocol and did not provide written comments
Question 4:
Draft Final Plan, page 1, paragraph 5- Using a peer review contract, EPA solicited
comment from non-panel experts on the use of asbestos as a surrogate for
determining risk from other contaminants and provided a report on those
comments back to the panel The non-panel experts generally supported the use of
asbestos as a surrogate, but encouraged the concurrent testing for lead Many
individual members of the panel, however, did not support the position that
-------
asbestos was an appropriate surrogate in determining nsk for other contaminants
What was the panel's rationale for elimination of chrysotile as a surrogate?
EPA Response:
First, by "panel, in the question just above, EPA assumes this refers to the WTC
Expert Technical Review Panel (WTC Panel), and not the asbestos-as-surrogate
peer review sub-panel In any case, it would be appropriate to simply comment
that the WTC Panel did not concur with the primary finding of the asbestos-as-
surrogate sub-panel While some on the sub-panel did not support the position
that asbestos was an appropnate surrogate, it is fair to conclude that the
"majority" opinion, although the sub-panel was unambiguously not a consensus
panel, was that asbestos would serve as an appropnate surrogate.
The report from that panel is posted on the WTC Panel web site under the
heading, "Background Documents " A good summary of the individual sub-panel
comments on asbestos as a surrogate can be found on pages 7-9 of this report
The WTC Panel concluded that a "WTC signature" would be a more powerful
tool for the purpose of determining the extent of remaining contamination from
the WTC collapse The contaminants of potential concern (COPC) EPA tested for
during its cleanup program (asbestos, PAHs, dioxms, lead and mercury) are all
common urban contaminants With the exception of lead, the COPC were found
at low levels and low frequency during EPA testing A distinct "WTC signature"
would allow attribution of COPC found in sampled dust to the WTC rather than
other sources
Question 5:
It's clear the five commercial labs were provided blind samples, but were the
three government labs'? I think they were, I just wanted to be sure
EPA Response:
All labs were provided with blinded samples
Question 6:
Nowhere (that I kind find) do the documents we have discuss sample size There
is mention of a 32 oz jar for the vacuum cleaner and a 1,000 gram maximum
reading for a scale. I get the idea the collected samples are on the order of 500
grams Is this correct?
EPA Response:
As much sample as possible was collected Sample sizes ranged from 10 grams or
so up to many hundreds of grams
-------
Question 7:
Would EPA provide a copy of the Occupational Safety and Health Administration
Field Operations Manual, "Instruction CPL 2-2.0A, March 1984, Chapter VII.
Sampling for Surface Contamination l.h"?
EPA Response:
The reference requested can be viewed online at
http //www osha gov/dts/osta/otm/otm_ii/otm_n_2 html
Question 8:
Has anyone measured the refractive indices of the MMVF in the WTC dust*? I'm
specifically interested in the slag wool and mineral wool discussed in our review
documents
EPA Response:
Message forwarded to ERG by EPA reads. "I don't think the specific refractive
index or the Rl range was measured for the slag wool present in the WTC dust I
do know that we did not measure it here at NEIC, but instead used the values
from TIMA, 1991 I also know that the slag/mmeral/rock wool fibers were easily
identified using both 1 550 and 1 605 RI oils "
Question 9:
Please provide copies of all of the data sheets generated by the laboratories for the
PLM and SEM analyses
EPA Response:
EPA provided ERG with 16 data files that were generated by 8 laboratories (some
laboratories' data are documented in more than one file) All of these files were
forwarded to the peer reviewers
Question 10:
Section 11 4 (Preparation of Sample for Polarized Light Microscopy) includes the
following statement- "Allow them to dry, then add a drop of 1 55 (or 1.605)
refractive index oil " Since no guidelines are provided to determine which oil,
1 550 or 1 605, is to be used, does this mean it was at the discretion of analysts to
decide which oil should be used?
EPA Response:
Yes
Question 11:
Section 11 4 (Preparation of Sample for Polarized Light Microscopy) includes the
following statement' "Allow them to dry, then add a drop of 1 55 (or 1 605)
refractive index oil." For the 10% spiked sample, which of the 8 laboratories
(A,B,C,D,E,F,G,H) used 1.550?
-------
EPA Response:
Laboratories F and H.
Question 12:
Section 121 (Analysis by Polarized Light Microscopy) includes the following
statement' "The fraction of fibers with refractive index greater than 1 55 (or
1 605) will contain mineral wool, which includes both slag wool and rock wool,
and possibly some E-type glass and ceramic fibers " Please explain in details in
which occasions all fibers > 1 550 are slag wool (SW), rock wool (RW), and E-
type glass (EG) and ceramic fibers (CF) and in which occasions only those >
1 605 are SW, RW, EG, and CF
EPA Response:
According to TIM A, 1991, slag wool and rock wool have a refractive index (RI)
of 1 6-1 8. E-type glass and ceramic fibers have an RI of 1 55-1 57 So,
depending on which RI oil was used, 1.55 or 1 605, EG and CF may or may not
be identified These four fiber types are the types that have been identified in
WTC Dust samples and background samples which have RI>1 55 Clarification
of fiber type, by chemistry, is to be performed m the SEM/EDS analysis This
secondary clarification is part of the reason that PLM was not a time-saving step
as expected
Question 13:
Section 1221 (Screening for Slag Wool) includes the following statement
"When an inorganic fiber is found, identify the composition of the particle by
EDS Slag wool is the primary fiber of interest" Please explain in details the
criteria used to differentiate slag wool fibers from other MMVF fibers by EDX.
EPA Response:
The criteria used to differentiate slag wool from other MMVF is from the USGS
Open File Report 2005-1165 "Particle Atlas of World Trade Center Dust" by
Heather Lowers and Gregory Meeker This atlas is available at
httD://Dubs.usQS.QOV/of/2005/1165/508OF05-l 165.html This was the
definition used for identification of MMVF for this purpose and was available for
all laboratories to use The pertinent text is as follows
MMVF and glass fragments
Man-made vitreous fibers (MMVF) are abundant in WTC
dust Glass fibers range in diameter from < 1 um to > 50 um
with lengths up to several hundred micrometers The best
compositional match for the majority (>85%) of WTC glass
fibers is slag wool, a by-product of pig iron production (TIMA,
1991) Pieces of yellow thermal insulation found in bulk
samples are composed of soda-lime glass fibers (Meeker and
others. 2005b) Very few fibers of this composition exist in the
-------
fine (<150 u.m) microscopic portion of the dust Rock wool is
also present as a trace constituent of the fine dust portion
Rock wool and slag wool can have similar EDS spectra
The two can be distinguished based on the presence of iron
Slag wool will generally have less than 2 weight percent FeO
whereas rock wool contains from 3 to 12 weight percent FeO
(TIMA, 1991) Soda-lime glass has a distinct EDS spectrum
from both slag wool and rock wool. The Na peak is higher and
Ca, Mg, and Al peaks are smaller in the soda-lime glass
spectrum than the slag wool and rock wool spectra.
Glass shards, fragments, and spheres are also present in the
dust samples The microscopic glass shards and fragments are
less abundant than the ubiquitous slag wool fibers in the fine
dust (<150 urn) Most of the glass fragments fall within the
compositional range for soda lime glass, a common type used
as window glass (TIMA, 1991) Other glass fragments are
present which contain mostly Si with trace amounts of Na, K,
and/or Al The majority (> 90%) of glass spheres, generally
less than 500 u,m in diameter, are of slag wool composition.
Sample EDS spectra from the atlas can be found following links from the text
Question 14:
Section A7 2 of the QAPP (Measurement Quality Objectives) includes the
following statement. "The accompanying tables (Tables 1 and 2) list
Measurement Quality Objectives (MQOs) for this mtralaboratory (within lab) and
mterlaboratory (within sample) variability Accuracies and precision were taken
from preliminary data and manufacturer's specifications " Please explain in detail
how accuracy was evaluated for the slag wool analysis by PLM
Measurement
Parameter
Individual dust
sample mass
Fibers/Concrete
Particles/Gypsum
Particles
Fibers
Analysis
Method
Microbalance
SEM
PLM
MQO for
Accuracy
+/- 5%
+/- 30%
+/- 30%
MQO for
Precision
+/- 5%
+/- 30%
+/- 30%
MQO for
Completeness
85%
85%
85%
EPA Response:
It wasn't. It was decided part way through the study not to use PLM as it was
determined not to be a time saving measure (as intended)
-------
Question 15:
Section A5 2 of the QAPP (Method Development) includes the following
statement. "This method was reviewed by the WTC Expert Panel's signature
subcommittee and is presented in Appendix B " The Appendices list that appears
on page 20 identifies Appendix B as the PLM/SEM protocol Please provide a
copy of the review referenced above
EPA Response:
A formal written review was not provided by the subcommittee members Drs
Paul Lioy and Mort Lippmann provided comments by email A PDF of these
emailed comments is provided under separate cover. Mr Meeker was involved in
the development of the protocol and did not provide written comments
EPA asked that the peer reviewers be aware of the web site for the World Trade
Center Expert Technical Review Panel (http //www epa.gov/wtc/panel/)
Presentations and summary reports from panel meetings are posted on the web
site, along with a number of background documents related to development of the
draft WTC sampling plan and the WTC dust screening study On the
"Background Documents" page, EPA has also posted comments from individual
members of the WTC panel
One panelist, David Newman of the New York Committee for Occupational
Safety and Health, has specifically asked the reviewers be made aware of some of
his comments that EPA has posted links to below
March 1 (page 13)
http //www epa gov/wtc/panel/pdfs/comments/02-23-05 PanelMembersComments pdf
May 27 (page 3)
http //www epa gov/wtc/panel/pdfs/comments/combo comments 052405 pdf
June 6 (page 10)
http //www epa govAvtc/panel/pdfs/comments/combo comments 052405 pdf
------- |