-------
Preliminary Analysis of Collected Samples for Slag Wool
Most of the collected samples were analyzed for slag wool content by the EPA's National
Exposure Research Laboratory (NERL) Scanning Electron Microscopy (SEM) Laboratory. This
analysis was performed as part of the EPA's development of a protocol for sample preparation
and analysis and for preliminary sample characterization. These samples were not analyzed for
elements of concrete or gypsum as an analytical method for these components had not yet been
developed The data acquired during this method/protocol development effort are presented in
Appendix B Caution should be used with these data as it was obtained while the method was
being developed Post-study data acquired by NERL are also presented in Appendix C.
In evaluating the method development data acquired by NERL (Appendix B), there appears to be
a distinction between samples taken in impacted areas versus background samples Eighteen of
the 21 samples from impacted areas had slag wool at concentrations of greater than 100,000 slag
wool fibers per gram of dust, with a range of 69,000 to 13,400,000, while all of the samples from
background areas had concentrations less than 100,000 fibers/gram, ranging from no slag wool
detected (in 12 of 47 samples) to 92,800 fibers/gram of dust
Based on this preliminary work, the USGS, the EPA's Office of Research and Development
(ORD), the EPA's National Enforcement Investigations Center (NEIC), and experts five
commercial testing laboratories (denoted labs A-H in Appendix E), worked together to develop
an analytical method to identify the presence and concentration of the screening constituents (i e.
slag wool, gypsum and elements of concrete) in indoor dust This method was reviewed by the
WTC Expert Technical Panel's signature subcommittee and is presented in Appendix D The
composition of this technical panel can be found at http //www epa.gov/wtc/panel
III. METHOD VALIDATION STUDY
Study Design
The basis for the WTC dust screening method discussed above is as follows if a unit has been
impacted, those materials that are found in WTC dust will be found in the dust collected from the
unit The materials under consideration are: 1) slag wool, 2) elements consistent with concrete
and 3) gypsum. The study described herein was intended to validate the WTC dust screening
method by demonstrating the following things
1) that the above described materials are reasonable markers for WTC dust (by showing that
these markers distinguish WTC-laden dust from background dust),
2) that WTC dust at a diluted concentration can be distinguished from background, and
3) that the analytical method works well enough and is able to be carried out by enough
analytical laboratories to 1) evaluate the above materials as markers and 2) distinguish
WTC dust from background dust
The first of these three objectives was partially addressed in method development work, which
focused on slag wool As described in the previous section, slag wool was found to be elevated
in locations deemed "impacted", while slag wool was not detected or detected at low
concentrations in "background" areas
-------
Five independent laboratones and three government laboratories participated in this method
validation phase One government laboratory analyzed only a small portion of the samples, but
this lab was critical in the method development Each laboratory attended a two day session
during which the method was further developed and discussed, and the protocol was adapted to
suit each laboratory's equipment.
Following this session, the laboratories received dust samples consisting of both confirmed
background samples (10 samples plus duplicates for a total of 20) and confirmed non-impacted
dust spiked with varying amounts of confirmed WTC dust (6 spiked samples plus duplicates for
a total of 12). Specifically, a sample that was characterized and confirmed as non-impacted
(designated in Appendix B as NE Queens maid service) was split, and the splits were spiked at
levels of 1, 5, and 10% total mass with two different characterized and confirmed WTC dusts
These spiked samples were then homogenized as documented in the QAPP for this study
(Appendix A). The two spiking dusts were 1) a composite sample of predominantly outdoor dust
collected in September of 2001 by USGS, and 2) dust collected by the U S EPA from the
Deutsche Bank building at 4 Albany Street in September of 2004 The 4 Albany Street building
borders the south side of the WTC complex. Six spiked samples were prepared for each
laboratory, these were split so that each laboratory received 12 spiked samples. Each laboratory
also received 10 non-impacted background samples that were also split, resulting in a total of 20
background samples. Thirty-two samples in all were sent for analysis to the eight labs
In addition to the 32 samples, one of the five commercial laboratones also received 28
background samples to increase the available data characterizing background locations.
The labs were provided the 32 samples "blind", they did not know which samples were pure
background dust, and which were the spiked dust To ensure sufficient results for spiked
samples, the government laboratory that was only able to analyze a small portion of the samples
was asked to analyze only the 12 spiked samples Again, they were not told the identity of these
samples (Lab C) The labs had five weeks to analyze all samples The final data from all
laboratories, including the data for the additional 28 background samples, were reviewed,
evaluated and analyzed by the EPA and the EPA's prime contractor This prime contractor's
from this analysis is presented m Appendix E
Composition of Spiked Samples
The USGS performed an analysis of the spiked, homogenized samples pnor to the samples being
sent to the labs The measured levels were in the approximate range for the spiking percent (1, 5,
and 10%) based on the undiluted concentration level of each WTC dust and, in all but one case,
each percent level was fully distinguishable from the others (Figures 2 and 3) The variability in
the measured levels was expected due to the difficulty in homogenizing dusts that have large
particle size distributions, and the fact that components of WTC dust will vary within a sample
because of the nature of the source Given these difficulties and the measurement results, these
dusts were determined to be reasonably homogeneous
As seen in Figures 2 and 3, the level of slag wool differs between the two WTC dusts, with the
pure dust that was collected from 4 Albany Street in 2004 more than an order of magnitude
lower than the dust collected by the USGS in September of 2001. The pure dust from 4 Albany
10
-------
Street had slag wool levels at 500,000 fibers/gram of dust versus approximately 11,000,000
fibers/gram of dust for the USGS collected sample. There are likely explanations for this large
difference in slag wool levels. The USGS sample was a composite of multiple outdoor samples
and one indoor sample taken during September of 2001. The 4 Albany was an indoor sample
was taken three years post 9/11 in September of 2004. As this 4 Albany sample was taken
exclusively inside of a building, it was not only diluted by three years accumulation of urban
background dust, but was also characteristic of dust that had penetrated the shell of a building as
opposed to that deposited on the ground outside.
USGS Spiking Material
1,400,000-
•+*
0
2
o>
u.
200,000
0
^
t
j
j
0
Q
f5>
^
s
J1234S678910
% Spiking
Figure 2: USGS Spiking Material Results. Analysis was conducted by USGS prior to being
sent to labs for study. Pure dust averaged approx. 11,000,000 fibers/gram.
(Figure provided by USGS)
11
-------
3
Q
"5
2
o
0)
.a
90,000
80,000
70,000
60,000
50,000
40,000
20,000
10,000
4 Albany Street Spiking Material
X
23456
% Spiking
10 11
Figure 3: 4 Albany Street Spiking Material Results. Analysis was conducted by USGS
prior to being sent to labs. Pure dust averaged approx. 500,000 fibers/gram.
(Figure provided by USGS)
IV. RESULTS AND DISCUSSION
Development of Study Results
The final report from the prime contractor with all raw analytical and calibration data can be
found in Appendix E. A summary of the study results that includes the data from the 28
additional background samples analyzed by a single commercial laboratory is provided in Table
I, as well as Figures 4-7. A map of the origin of the samples analyzed during this study is shown
in Figure 8.
All background sample data used in Table I and Figures 4-7 are from the Greater NY City area.
Background samples taken in Research Triangle Park, North Carolina are not included as they
are not representative of NY City background dust. Data for all background sample results may
be found in Tables 3 and 4 of the Versar report in Appendix E. It should be noted that the
Research Triangle Park samples show higher slag wool levels than NY City area background
samples. This is due to the presence of slag wool containing ceiling tiles in the building
sampled. Note also that Table I indicates two average values for background slag wool. These
values reflect the inclusion and exclusion of two samples collected in New Jersey (NJ) and Long
Island (LI) that were extremely high in slag wool fibers, likely due to their insulation,
fireproofmg or ceiling tiles. Based on these results it is likely that some false positive results will
occur in buildings with slag wool-based ceiling tiles, fireproofmg or insulation. .
12
-------
Three of the commercial laboratones, 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 The statistical analysis performed to make this
determination is presented in Appendix F In addition, Lab H was not considered when
determining concrete and gypsum levels as their data were at least two times higher than the
sample average without these data (Table I and Figures 6 and 7)
In discussions with the commercial laboratories, it was determined that some labs did not have
the personnel or the equipment to perform the required analysis in the given timeframe, thus,
data quality became an issue Additionally, labs that had less experience with slag wool analysis
felt that a clearer definition, in addition to that provided in the catalog developed by USGS in
Lowers et al., 2005b, of slag wool was needed to distinguish it from other mineral wools
Finally, labs that were unable to automate the gypsum and concrete analysis expressed their
belief that the method was too long and complicated for accurate quantitative dust analysis All
laboratory comments will be taken into consideration in when finalizing the protocol
13
-------
Slag Wool
Average
(fibers/g
dust)
Elements of
Concrete
(% Area)
Gypsum
(% Area)
Background
(Greater NY Area)
AVG + SD
35,950 + 74,300
17,740+15,835*
Range of Samples
ND* - 369,230
ND* - 60,000**
AVG + SD
156 + 57
Range of Samples
6 - 30.5
AVG±SD
95 + 34
Range of Samples
4-165
USGS Spiked (Collected
9/01)
1%
94,000 + 25,740
5%
452,510+100,640
10%
870,280 + 310,420
1%
20 + 6
5%
19 + 7
70%
16 + 2
1%
9±6
5%
7 + 3
70%
6 + 05
4 Albany Spiked
(Collected 9/04)
1%
17,270 ±7,880
5%
52,510 + 26,140
10%
88,540+18,300
1%
15 + 1
5%
18±4
10%
16 + 3
1%
9 + 4
5%
5 + 2
70%
7 + 2
• **ND=Non Detect (Zero slag wool fibers)
• *Two extremely high values from NJ and LI removed
Table 1: Avg, Standard Dev., and Range of Results for Background and Spiked Samples
(Data Summarized from Tables 1,2,3 and 4 of Appendix E).
14
-------
1000000 -i
100000 —
in
a
•o
2
5
2
o
g
D)
a
CO
10000
1000
100
High Slag Wool
Values in NJ and LI
Background
Spiked Samples
Figure 4: Average Slag Wool (Fibers/Gram of Dust) in background and spiked samples.
(Data from Tables 3 and 4 Appendix E)
15
-------
10000000 3
1000000 :
(A
3
•o
ro
5
"5
1
o
o
(A
100000
10000
1000
100
• Background
A 4 Albany 1%
$4 Albany 5%
84 Albany 10%
«» USGS 5%
&USGS 10%
'$• Impacted
High Slag Wool
Values in NJ and LI
0
Background
Impacted
Spiked Samples
Figure 5: Average Slag Wool (Fibers/Gram of Dust) in background, spiked and impacted
samples. Impacted samples are locations that are shown in satellite pictures to have been
affected by WTC Collapse Dust. Slag wool results for impacted samples were derived
during method development and were not part of this method validation; they are provided
for comparative purposes. These impacted samples range from 0.1 to 1.6 miles from the
WTC site (see Figure 8 for sample origin location). Data from Appendix B (Impacted) and
Tables 3 and 4 of Appendix E (Background and Spiked).
16
-------
30
25-
20
ra
4)
15
10
0
$
T
Background
Spiked Samples
Figure 6: Average of Elements of Concrete (% Area) in background and spiked samples.
(Data from Tables 1 of Appendix E)
17
-------
18 r
16 •
14 •
12
10
I
•
19
O
0
o
0
T
o-i-
Background
Spiked Samples
Figure 7: Average of Gypsum (% Area) in background and spiked samples.
(Data from Tables 2 of Appendix E)
18
-------
DESIGNATION
LOCATION
3,4
5,6
7
B"
10
il',HS3<1),HS(2)
Stony Brook, LI
West End Ave between 72nd and 73rd Streets
30th Ave between 21st and 23rd, Queens
70th Street between 20th and 21st Ave, Brooklyn
79thi St between York and East End Ave, Manhattan
:92nd Street between Columbus and CFW, Manhattan
iColumbia'lviedicai'Center.w; 'i 68th St', Manhattan
''teaneck,NJ
14,15
16
89th Street between Amsterdam and Columbia, Manhat
:8fJth Street between' Riverside and East End Ave , Ma
West End Ave between 1 0Stn and 1 6eih sireets
:Research Tiiangje Park, NC
24,25,26
CMC(1),CMC(2)
vvoa;i),wGsi;2)"
USCCi),USC(2)''
Chttenden Avenue, Manhattan
' Colombia Medical'Center , w! 'i 68th SI , Manhattan
.Nassau County, LI
:Federai Courthouse, White Plains. NY
12,13
28
27
FP(i'JiFP(2J"
Northern Manhattan, Above 70th Street
Nassau County
NE Queens
'long Beachi is land, NJ
Federal Courthouse, Central islip, NY
Figure S: Map of the origin of the samples analyzed during this study
(Reference Appendix D for sampling data).
19
-------
Discussion
Slag wool appears to be an indicator for WTC dust and can be distinguished from background
dust at all three spiking levels for the USGS dust and at the 10% level of the 4 Albany Street
dust The 4 Albany Street dust is considered to be WTC impacted dust but as noted earlier, the 4
Albany dust likely had lower levels of slag wool due to the fact that it was an indoor dust that
was not sampled until three years after the WTC collapse
Levels of gypsum and elements of concrete have no discemable relationship to the level of WTC
dust There does not appear to be a distinguishable difference between levels of concrete and
gypsum in background dust and the samples spiked with WTC dust, despite USGS analysis of
WTC dust from 2001 (Meeker, 2005) showing elevated levels of these components. This is
likely due to the fact that while these components may seem high in WTC dust, they are also
high m general background dust as they are common building materials
While method development (Appendix B and summarized in Section II above) work showed that
dusts from known impacted locations generally had slag wool levels above 100,000 fibers/gram,
several samples taken within this impacted zone and analyzed during method development
showed lower levels of slag wool Two likely explanations can be offered for these results
First, as the data in Appendix B was acquired during method development, it must be viewed as
such, and second, multiple cleanings of the inhabited areas since September 11, 2001 may have
removed residual WTC collapse contamination. The majority of these samples were taken in
fully inhabited buildings, from locations within the buildings that can be characterized as either
'accessible5 or 'infrequently accessed' areas. These terms are described in the final draft EPA
sampling program, and they denote areas that are accessed by people over the course of time,
such as counter tops or rugs (accessible) or underneath furniture (infrequently accessed) For
this reason alone, it is encouraging that a substantial amount of the dust sampled in late 2004 and
beyond had high levels of slag wool
While there was ample evidence of higher levels of slag wool associated with the WTC dust and
lower levels associated with background, there is high variability in slag wool measurements
within and between labs Estimates of within lab relative standard deviations based on analysis
of duplicate samples of the 4 Albany Street data are 55%, 24% and 14% for the 1%, 5% and 10%
dilution levels, respectively Estimates of between lab relative standard deviations based on the
4 Albany Street data are 64%, 70% and 29% for the 1%, 5% and 10% dilution levels,
respectively (looking at results from analysis of the same spike level samples by multiple labs)
Causes of the high levels of variability may include1
• Procedures to homogenize the spiked samples did not result in complete
mixing and distribution of fibers; they instead resulted in a 'reasonably'
homogeneous sample given the large size variation of the dust components
• Components of both non-impacted/background and WTC dusts will vary
within a sample because of the inherent nature of the dust samples Thus, the
samples received by the labs may vary in content
• Operator experience with the target components appeared to be an issue -
post-study discussion indicated that labs representatives with less familiarity
with slag wool expressed a belief that further guidance as to its definition was
needed
20
-------
• The variability in the mass of dust used for the analysis, as the protocol allows
for a range, not a specific mass, to be used This range is essential due to the
extreme differences in slag wool levels possible between background and
spiked samples
Finally, it is noted that Table I indicates two average values for background slag wool. These
values reflect the inclusion and exclusion of two samples (and their duplicates) collected in New
Jersey (NJ) and Long Island (LI) that were extremely high in slag wool fibers, likely due to their
insulation, fireproofing or ceiling tiles. Similarly, it was earlier noted that samples taken from a
North Carolina building due also to slag wool used in ceiling tiles were not included in the
interpretative analyses Based on these results, it is likely that some false positive results will
occur in buildings with slag wool-based ceiling tiles, fireproofing or insulation
V. CONCLUSIONS
The interlaboratory results indicate that the better performing labs are capable of distinguishing
the difference between 1,5 and 10% 4 Albany Street dust Also, despite the high levels of within
sample and within lab variability, the method using slag wool appears to be sensitive enough to
distinguish 10% 4 Albany Street dust from background dust Additional evaluation of the data
will be performed to further understand the variability Measures will be taken (i e standards
will be sent regularly to each lab) during EPA's planned sampling program to evaluate the
accuracy and precision of the laboratones
In summary, the data developed in this study support the following findings
1) Five of the eight laboratones were able to reasonably measure the slag wool
concentrations in background dust spiked with confirmed WTC dust
2) High levels of variability in slag wool measurements, both within labs and between
labs, were observed in the data Despite this variability, the slag wool method appears to
be sensitive enough to distinguish WTC dust from background dust at the 10% level
(defined as, 4 Albany Street)
3) The levels of gypsum and elements of concrete m the spiked samples were
indistinguishable from the levels in the background samples This observation suggests
that, while these components may have been elevated in dust samples collected near
September 2001, as found by USGS in their studies on WTC dust, they are also
commonly found in the indoor environment and would not be useful as WTC signature
components
4) Analysis of samples during method development generally showed slag wool levels m
samples from impacted locations to be greater than slag wool levels in samples from
background locations
21
-------
VI. REFERENCES:
Lowers, H A , G.P Meeker, and I K Brownfield (2005a) 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
http //pubs usgs gov/of/2005/1073/
Lowers, H A , Meeker, G P , I K Brownfield. (2005b) World Trade Center Dust Particle Atlas
U S Geological Survey Open-File Report 2005-1165 http //pubs uses gov/of/2005/1165/
Meeker, G P , A M Bern, H A Lowers, and I K. Brownfield. (2005) Determination of a
Diagnostic Signature for World Trade Center Dust using Scanning Electron Microscopy Point
Counting Techniques URL http //pubs usgs gov/of/2005/1031/
U S. Geological Survey Open File Report 2005-1031
NYCDOHMH/ATSDR (2002) New York Department of Health and Mental Hygiene and
Agency for Toxic Substances and Disease Registry Final Technical Report of the Public Health
Investigation To Assess Potential Exposures to Airborne and Settled Surface Dust in Residential
Areas of Lower Manhattan URL: http://www epa gov/wtc/panel/ATSDRFinal-report-
lowermanhattan-02.pdf Agency for Toxic Substances and Disease Registry, US Department of
Health and Human Services, Atlanta, GA
R J Lee (2004) Signature Assessment 130 Liberty Street Property Expert Report WTC Dust
Signature Prepared for Deutsche Bank May, 2004. R.J Lee Group, Inc 350 Hochberg Road,
Monroeville, PA 15146
US EPA (2004) Mapping the Spatial Extent of Ground Dust and Debns from the Collapse of the
World Trade Center Buildings, DRAFT in peer review, EPA/600/X-03/018, URL
http./Avww epa aov/wtc/panel/backdocs html
United States Environmental Protection Agency, Office of Research and Development,
Washington, D.C , July 2004, 35pp.
US EPA (2002) Exposure and Human Health Evaluation of Airborne Pollution from the World
Trade Center Disaster, Draft in peer review, EPA/ URL'
bUp..//cfpub..ej3a.gw^ United States Environmental
Protection Agency, Office of Research and Development, Washington, D C , October 2002
VII. CONTRIBUTORS TO THIS STUDY:
Principal Investigators:
Jacky Ann Rosati U S EPA, ORD
David Friedman U S EPA, ORD
Contributors (in alphabetical order):
Nancy Adams U S EPA, ORD
Gma L Andrews U S. EPA, ORD
Lara Autry U S. EPA, ORD
22
-------
Amy Bern US EPA,NEIC
Isabella Brownfield USGS
Ten Conner U S EPA, ORD
EvanEnglund US EPA, ORD
Pat Evangehsta U S EPA, Region 2
Henry Kahn US EPA, ORD
Matthew Lorber U.S EPA, ORD
Heather Lowers USGS
Mark Maddalom U S EPA, Region 2
Lisa Matthews US EPA, ORD
Greg Meeker USGS
TimOppelt US EPA, ORD
Joachim Pleil U S EPA, ORD
Dennis Santella US EPA, Region 2
Raj Smghvi U S EPA, Region 2, ERT
Stanley Stephanson U S EPA, Region 2
Shirley Wasson U S. EPA, ORD
Steve Wilson USGS
Supporting Contractors:
Prime Contractors
Alion Scientific
Lockheed-Martin
Versar
Subcontractors
RJ Lee Group, Inc
EMSL Analytical Inc.
MVA Scientific Consultants
Reservoir Environmental
MAS, Inc
Other Contributors:
John Holland SEE US EPA, ORD
VIII. ACKNOWLEDGEMENTS:
The U S EPA would like to acknowledge the U.S General Services Administration (GSA),
NY/NJ Port Authority, the U S National Park Service, Deutsche Bank, and Columbia University
(K Crowley) for allowing samples to be collected at their facilities
In addition, the EPA would like to acknowledge the residents of NY and NJ who allowed us to
sample in their homes, and the maid services that collected vacuum cleaner bags for use in this
study. Finally, the EPA would like to thank the Sierra Club for helping to recruit sampling
locations
23
-------
IX. APPENDICES
APPENDIX A: QUALITY ASSURANCE PROJECT PLAN FOR THE WORLD TRADE
CENTER (WTC) SCREENING METHOD STUDY
(Due to formatting - this document will be provided under separate cover)
24
-------
APPENDIX B: DATA ACQUIRED BY EPA NERL DURING METHOD
DEVELOPMENT
Samples Collected at Background Locations
Residential
West End Ave between 72nd and 73rd Streets, Manhattan
30th Avenue between 21st and 23rd St, Queens
E 79th Street between York and East End Ave, Manhattan
Chittenden Avenue, Manhattan
92nd Street between Columbus and CPW, Manhattan
80th Street between Riverside and West End Ave, Manhattan
Edison, NJ
Stony Brook, LI
70th Street between 20th and 21st Ave. Brooklyn
Teaneck NJ
Long Beach Island, NJ
West End Avenue between 105th and 106th Streets, Manhattan
Edison, NJ
88th Street between Amsterdam and Columbia, Manhattan
North East Queens (Maid Service)
Nassau County, Long Island (Maid Service)
Bus/ness
Port Authority Bldg, Port of Newark, NJ
slag wool fibers/
gram of dust
2 53E+04
5 47E+04
2 80E+04
2 26E+04
4 93E+04
1 53E+04
2 87E+04
2 42E+03
1 46E+04
0 OOE+00
1 79E+04
2 90E+04
4 09E+04
4 77E+04
0 OOE+00
0 OOE+00
1 77E+04
412E+03
8 35E+03
0 OOE+00
5 74E+03
0 OOE+00
0 OOE+00
0 OOE+00
5 37E+03
1 02E+04
1 27E+04
0 OOE+00
1 63E+04
6 43E+03
0 OOE+00
1 65E+04
0 OOE+00
0 OOE+00
1 95E+04
3 86E+04
3 45E+04
7 32E+04
5 09E+04
1 85E+04
6 60E+04
Average of Duplicates
(slag wool fibers/gram
dust)
2 20E+04
4 43E+04
25
-------
Columbia Medical Center, W 168" St, Manhattan
Edison, NJ
Federal Courthouse, Quarropas St, White Plains
Federal Courthouse, (slip, Long Island
8 58E+04
0 OOE+00
1 33E+04
9 09E+04
9 28E+04
9 OOE+04
Samples Collected at Known Impacted Locations
Business
290 Broadway, Manhattan
Broadway between Maiden Lane and John Street, Manhattan
Deutsche Bank Bldg, 130 Liberty Street, Manhattan
Deutsche Bank Bldg, 4 Albany Street, Manhattan
USGS Composite Sample Collected Sept 2001
Samples Collected at Locations with Unknown Impact
Residential
John Street between Gold and Pearl, Manhattan
South End Avenue between Albany and Liberty, Manhattan
River Terrace, Manhattan
40th Street between Tunnel Exit St and 2nd Ave. Manhattan
Orange Street between Henry and Hicks, Brooklyn
24th Street between 8th and 9th Ave, Manhattan
Montague between Montague Terrace and Hicks Street, Manhattan
Houston and Mulberry Streets, Manhattan
Bus/ness
Port Authority Bldg, Columbia St, Brooklyn
6.92E+04
881E+04
1 64E+05
1 95E+05
8.35E+04
1 33E+05
2 79E+05
471E+06
5.77E+06
6 60E+06
1 18E+07
1 22E+07
1 13E+05
2 06E+05
214E+05
2 25E+05
2 28E+05
2 78E+05
6 36E+05
1 67E+06
1 34E+07
1 26E+04
917E+03
0 OOE+00
291E+03
1 11E+04
3 32E+03
5 03E+03
6 30E+03
2 06E+05
9 89E+04
1 30E+05
1 94E+05
1 12E+04
3 06E+05
1 20E+05
619E+06
2 30E+05
26
-------
Governor's Island 5 07E+04
5 75E+05
8 79E+04
Vanck Street, Manhattan 9 57E+04
Samples Collected Outside of NY City
Business
Research Triangle Park, NC 5 OOE+04
8 96E+04
27
-------
APPENDIX C: DATA ACQUIRED BY EPA NERL POST-STUDY
Samples Collected at Background Locations
Residential
Composite -North East Queens (Maid Service) 1 06E+04
1 49E+04 1 28E+04
Business
Port Authority - Port of Newark, N J 9 77E+03
Samples Collected at Impacted Locations
Business
Governor's Island 1 93E+04
6 39E+05
1 21E+06
Port Authority Bldg, Columbia St, Brooklyn 1 22E+05
28
-------
APPENDIX D: PROTOCOL USED FOR THE SCREENING METHOD STUDY
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
June 27, 2005
Prepared by:
U.S. Environmental Protection Agency
National Enforcement Investigations Center/ National Exposure Research
Laboratory/National Homeland Security Research Center
Denver, CO and Research Triangle Park, NC
The use of trade names does not imply endorsement and are used for illustrative purposes only
29
-------
Contents
1.0 Purpose 31
2.0 Scope/Application 31
2.1 Limitations of the Method and Future Considerations 31
3.0 Definitions 31
4.0 Summary of Method 31
5.0 Interferences 32
6.0 Safety 32
7.0 Apparatus and Materials 32
8.0 Reagents 33
9.0 Sample Storage 33
10.0 Quality Control 33
10.1 Calibration 34
11.0 Procedure 34
11.1 Weighing and Splitting 34
11.2 Ashing 35
11.3 Sieving 35
11.4 Preparation of Sample for Polarized Light Microscopy 35
11.4 Mounting Sample on SEM Sample Stubs 35
12.0 Analysis 37
12.1 Analysis by Polarized Light Microscopy 37
12.2 Analysis by SEM/EDS 35
12.2.1 Screening for Slag Wool 35
12.2.2 EDX Screening for Gypsum/Anhydrite 3 5
12.2.3 X-Ray Mapping for Gypsum 36
12.2.4 X-Ray Mapping for Ca-rich Particles 37
12.2.5 Particle Analysis for Gypsum and Concrete 37
13.0 Data Analysis and Calculations 38
14.0 References .39
15.0 Appendix 40.
30
-------
1.0 Purpose
This document describes sample preparation and analytical screening procedures for bulk samples of dust
collected from residential and commercial office environments These methods are collectively referred to
as the protocol
2.0 Scope/Application
The protocol describes polarized light microscopy (PLM) and scanning electron
microscopy (SEM) with energy dispersive spectrometry (EDS) to screen bulk dust
samples for mineral slag wool, particles consistent with concrete compositions, and
gypsum The analysis methods include operating parameters and particle identification
criteria.
2.1 Limitations of the Method and Future Considerations
This protocol provides a means of analyzing for particles consistent with those found in dust
present after the collapse of the World Trade Center (WTC) in New York City Components of
WTC Dust have been documented and catalogued by the U S Geological Survey Denver
Microbeam Facility and the images and charactenstics shall be used in identification of particles
0)
The x-ray mapping procedure in sections 1223 and 1224 and the calculations presented in
section 130 only determine the maximum percentage of non-gypsum, calcium-nch particles,
which may include non-concrete matenals The particle analysis procedure presented in section
12 2 5 is the preferred procedure for determining the percentages of gypsum and concrete particles
in the sample
The x-ray mapping and image analysis procedure relies heavily on the thresholds for backscattered
electron images Binary (particles white and background black) backscattered electron images
(BEI) should be used to reduce errors in setting thresholds in Photoshop
3.0 Definitions
1 PLM - Polarized Light Microscopy
2 SEM - Scanning Electron Microscope
3 EDS - Energy Dispersive Spectrometry
4 SEI - Secondary Electron Image
5 BEI - Backscattered Electron Image
6 Mineral Wool - lightweight vitreous fibrous material composed of rock wool and slag wool and used
especially for heat and sound insulation
7 Rock Wool - a man-made vitreous fiber (MMVF) component of mineral wool containing magnesium,
aluminum, silicon, and calcium Sodium and potassium may also be present Iron oxide is typically 3-
12% by weight
8 Slag Wool - a man-made vitreous fiber (MMVF) component of mineral wool containing magnesium,
aluminum, silicon, and calcium Sodium and potassium may also be present Iron oxide is typically
less than 2% by weight
9 HEPA - High-Efficiency-Particulate-Air Filter
4.0 Summary of Method
1. Weigh sample to nearest 0.0005 g
31
-------
2 Split the sample, archive half and keep half for analysis
3 Ash half of the sample for analysis
4 Sieve the ashed sample to ISO um
5. Split the <150 um ashed portion Archive three quarters of the sample Keep one
quarter for PLM and SEM/EDS analysis
6 Weigh the quarter and place it in enough isopropanol to get a 10-20 mg per mL
dilution Apply an aliquot to a glass slide, let dry, and add 1 55 (or 1 605) refractive
index oil. Analyze by PLM for mineral wool.
7 Prepare a sample for SEM/EDS analysis using the same dilution prepared for PLM.
8 Apply an aliquot of the sample to an aluminum sample stub with a carbon adhesive
tab covered by a piece of polycarbonate filter (13-mm diameter or punched out of a
larger filter to fit the size of the stub)
9 Identify fibers by EDS and record the occurrence of fibers > 25 um in length at 100 x
magnification to get a statistical representation of fiber compositions
10. Prepare 10-fold dilution of the suspension from step 7 and apply an aliquot to a
polycarbonate/adhesive tab substrate affixed to an aluminum sample stub
Alternatively, a lighter loading can be prepared by filtering the diluted suspension
through a 25-mm diameter, 0.4-um pore size, polycarbonate filter and affix this to a
carbon adhesive tab affixed to an aluminum sample stub
11 Collect x-ray maps of 10 fields at 500 x magnification for major elements, especially
Ca, S, and Fe and use Adobe Photoshop or similar software to determine the area
percent of gypsum and Ca-nch particles Fe-nch particles may also be identified in
this step
12 Perform particle analysis via computer-controlled SEM/EDX analysis
5.0 Interferences
Interferences include possible contamination of samples by airborne dust or through improperly cleaned
glassware and sieves Interferences are minimized by performing all procedures involving dry dust in a
clean room, cleaning countertops and glassware thoroughly before proceeding and placing particle-free
wipes on all working surfaces To avoid cross-contamination, properly clean all glassware, sieves, and
tools between samples
6.0 Safety
Respirable particles which may present a health hazard may exist in the sample Bulk samples may release
respirable particles during handling All procedures involving dry dust samples will be performed under a
negative flow High-Efficiency-Particulate-Air Filter (HEPA) hood Samples handled outside of the HEPA
hood will be covered with aluminum foil or placed in sealed glass jars
7.0 Apparatus and Materials
1 HEPA negative flow hood
2 Forceps
3 Kimwipes
4 Stainless steel spatula
S Weighing paper
6 Programmable furnace [not required for validation study]
7 Ceramic crucibles with lids [not required for validation study]
32
-------
8 Analytical balance (accuracy to 0 OOOS g)
9 Retsch ultrasonic sieve shaker (AS200 Basic), or similar [not required for validation study]
10 Sample sieves, 3-inch diameter (recommended), 150-nm (100-mesh) opening, with lid and bottom pan
similar [not required for validation study]
11 SEM aluminum sample stubs
12 Conductive carbon adhesive tabs
13 Eppendorf pipette, 10-uL capacity
14 Disposable pipette tips
15 1 - 10 mL pipette
16 Glass vials for sonicating dust in isopropanol suspension (holds 10-mL volume)
17 Razor blade
18 Ultrasonic bath
19 50 mL glass beaker
20 Polycarbonate filters (25-mm diameter, 0 4-nm pore size)
21 Polycarbonate filters (13-mm diameter, 0 4-um pore size), or borer to cut larger filters to SEM stub
size
22 11 -mm diameter cork borer
23 Milhpore filter apparatus for use with 25 mm filters
24 125 mLNalgene bottles
25 Hand-held vacuum pump
26 High-vacuum carbon evaporator with rotating stage
27 Glass etri dishes with lids
28 Adobe Photoshop Software, or similar
29 Glass petrographic slides
30 Glass cover slips
31 Polarized light microscope for mineral identifications
32 Scanning Electron Microscope with the following attributes
a Resolution 5 run (at 25 k V, WD=10 mm - system dependent) or better
b Accelerating Voltage 10to20kV
c Minimum magnification range 50x to 200,000x
d SEI (secondary electron image)
e 6EI (backscattered electron image)
f Energy dispersive x-ray detector and analyzer for EDS analysis
g Ability to collect x-ray maps or particle analysis software (preferably both)
8.0 Reagents
1 Isopropanol, reagent grade [CAS No 67-63-0]
2 1 55 or 1 605 Refractive Index Oil
9.0 Sample Storage
Dust samples will be stored in an air-tight container, such as a sealed glass jar Samples placed in reagents
will be labeled appropriately and stored according to laboratory safety standards Samples prepared for
analyses will be stored in a protective container, such as a plastic case or covered etn dish, to prevent
contamination
10.0 Quality Control
Quality control is implemented by thoroughly cleaning glassware and spatulas, keeping working surfaces
clean, and preventing cross contamination. During ashing, particles may be suspended if slow heating is
not achieved Following the ashing program as outlined will minimize flashing, which can cause particles
to become airborne Covered crucibles will be used to prevent contamination caused by flashing Used
Eppendorf pipette tips and weighing papers will be discarded and new tips and papers will be used for each
33
-------
sample
Duplicate samples shall be prepared to determine the precision of the analysis In addition, sample blanks
shall be prepared These blanks are checks for cross contamination during handling of the samples Blanks
shall be prepared at the same time and in the same manner as samples
10.1 Calibration
Calibration of the EDS system must be completed at least once at the beginning and again at the
end of each analytical session Backscattered electron image (BEI) calibration should be
performed at the beginning of the session and anytime the backscattered image brightness and/or
contrast is adjusted
EDS calibration for both qualitative and quantitative (not required by this method but could be
useful for identification of particle type) analysis is accomplished by the analysis of a polished
carbon-coated reference standard The recommended material is USGS BIR1-G basalt glass
mounted in epoxy in a brass tube, polished, and carbon coated using a carbon evaporator (2, 3)
The calibration reference matenal should be analyzed at the same operating conditions to be used
for the analysis including beam current, accelerating voltage, working distance, detector dead
tune, and sample tilt (= 0°) For BIR1 -G the analysis should be performed with a beam size of 10-
20 um or equivalent area raster All calibration spectra will be saved with the corresponding data
set The calibration data will be used for inter- as well as intra-laboratory comparisons This
calibration is in addition to, and not a substitute for the normal EDS calibration recommended by
the EDS manufacturer which will be performed at regular intervals as specified by the EDS
manufacturer
Backscattered electron detector calibration can be performed on the same BIR1-G material by
adjusting the detector brightness and contrast to achieve the following conditions The epoxy on
the BIR1-G reference material will be at 0 in a 256 grayscale image and the brass mounting tube
will be at 256 The BIR1-G basalt glass should fall at approximately 130-140 gray scale units
11.0 Procedure
11.1 Weighing and Splitting
Weighing and splitting should be performed under a negative flow HEPA hood
If the fan speed is set too high, loss of particles may occur. The fan speed may
need to be adjusted to prevent the loss of fine particles
Obtain an analytical balance with an accuracy of 0 0005 g and preweigh a clean
piece of weighing paper Transfer the dust from the sample vial to the weighing
paper and determine the weight of the dust Split the sample with a clean razor
blade using the cone-and-quarter method If there are large clumps of organic
fibers, such as hair or lint, temporarily remove the hair with a pair of forceps and
tap the forceps lightly with another tool over a piece of weighing paper to remove
fine particles. Center the fine fraction on the paper and split the sample into four
equal parts using a razor blade Collect opposite corners (!/2 of the sample) for
analysis and archive the other half Quarter the larger organic fiber bundles the
same way, keeping half to proceed to the ashing step and half for archival
purposes
34
-------
Place the two quarters for ashing into a preweighed crucible Weigh the split and
record the results
11.2 Ashing
Place the ceramic crucibles containing the samples into a furnace
The furnace program should proceed as follows
1 Increase temperature by 1 "C/minute until sample reaches 250 °C
2 Hold temperature at 250 °C for 4 hours
3 Increase temperature by 1 °C/minute until sample reaches 480 °C
4 Hold temperature at 480 °C (sufficient for decomposing orgamcs) for 8 hours Do not exceed
500 °C
5 Shut off furnace
6 Allow sample to cool before removing from furnace
7 Weigh the ashed sample to the nearest 0 0005 g and record the result
11.3 Sieving
Sieve the sample through a 150-um sieve using a Retsch ultrasonic sieve shaker,
or similar Three-inch diameter sieves are recommended to minimize sample loss
from particles being trapped in the sieve The ultrasonic shaker will be operated
at 20-minute intervals at the following settings. 20, 40, 60, 70, 80, then back
down to 50 and 20. This will provide amplitudes ranging from 0 to 1 5 mm
Transfer the large and small fractions to clean pieces of weighing paper and
weigh to the nearest 0 0005 g Archive the fraction greater than 150-um
11.4 Preparation of Sample for Polarized Light Microscopy
Split the less than 150-um sample fraction using the cone and quarter method Collect one comer
for analysis and archive the other three quarters Weigh the quarter split to the nearest 0 0005 g
and place it into a glass vial Make a suspension of 10-20 mg dust per mL of isopropanol The
amount of isopropanol needed will vary depending on the amount of dust, the target dilution is 10-
20 mg per mL
Cut an Eppendorf pipette tip with a razor blade to increase the opening to
approximately 1 mm
Place the suspension in an ultrasonic bath for one minute, then remove the suspension from the
ultrasonic bath and shake it gently to suspend all particles Collect a 10-uL aliquot of the mixture
using an Eppendorf pipette with the modified tip and transfer to a glass slide Prepare 4 such
slides Allow them to dry, then add a drop of 1 55 (or 1 605) refractive index oil
11.5 Preparation of Sample for SEM Analysis
Prepare the SEM substrate on aluminum stubs using 04-um pore size
polycarbonate filters, carbon adhesive tabs. Using an 11 mm filter punch and
placing the filter between two filter separators, punch a circle the size of the
35
-------
carbon tab into the filter Place carbon adhesive tab affixed to an aluminum stub
on the dull side of the 11-mm polycarbonate filter such that the shiny side of the
filter exposed If available, a 13-mm diameter polycarbonate filter may be used in
place of the punched out 11-mm filter
Collect a 10-uL aliquot of the mixture from the PLM sample preparation using
the Eppendorf pipette with the modified tip and transfer to a prepared
polycarbonate/adhesive tab substrate This will yield a loading on a 12-mm SEM
stub of about 100-200 ug, which is a moderately heavy loading. Adjust the
number of aliquots as needed to obtain the target loading
Prepare a 10-fold dilution of the above suspension to get a suspension of 1-2 mg
dust per mL of isopropanol Sonicate the suspension in an ultrasonic bath for one
minutes. Remove the suspension and gently shake it to suspend all particles.
Wait one minute to allow the coarse particles to settle Collect a 10-^iL aliquot of
the suspended mixture using an Eppendorf pipette with the modified tip and
transfer to a prepared polycarbonate/adhesive tab substrate This will yield a
loading on a 12-mm SEM stub of about 10-20 ug, which is a light loading
Adjust the number of aliquots as needed to obtain the target loading.
Alternatively, prepare a lightly loaded sample using the filtration method as
follows Use a Millipore filter apparatus for use with 25-mm filters for filtration.
Place a few drops of isopropanol on the fritted glass surface and place the 25-mm
polycarbonate filter (0 4-um pore size) on the isopropanol. Attach the top of the
apparatus and add a few milliliters of isopropanol to the filter so that no part of it
is exposed to air Sonicate the suspension (diluted as described in previous
paragraph) in an ultrasonic bath for one minute. Remove the suspension and
gently shake it to suspend all particles Wait one minute to allow the coarse
particles to settle. Collect 1 mL of the suspended mixture using a pipette and
filter it through the polycarbonate filter Actual amounts for filtration will vary
based on sample loading. The goal is to have a loading on a 12-mm SEM stub of
about 10-20 ug, or about 5-10 percent area coverage, which is a light loading
Adjust the volume of the aliquot to filter as needed to obtain the target loading
Place the filter on a carbon adhesive tab on a standard SEM aluminum mount
The filter needs to be completely flat on the SEM stub. This can be achieved by
forming the wet filter into a gentle U-shape using forceps and the side of the
forefinger, then placing the bottom curve of the filter onto the center of the carbon
adhesive tab and slowly releasing the sides so they lay flat. Tnm the edges of the
filter using a razor blade
After drying, coat the samples on the polycarbonate or polycarbonate/adhesive tab substrates with
carbon using a carbon evaporator with a rotating stage Transfer the stubs to the SEM in a clean,
covered container
12.0 Analysis
36
-------
12.1 Analysis by Polarized Light Microscopy
Polarized light microscopy will be conducted using the general techniques outlined in EPA
600/R93/116 (4) For this procedure, four slides (prepared as descnbed in section 11 4) will be
analyzed 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 The fraction of fibers with refractive index less than 1 55 (or 1 605) will contain
primarily soda-lime glass fibers For the validation study, numbers of fibers greater than and less
than 1 55 (1 605) refractive index will be counted Dispersion staining and becke line techniques
may be used Fiber point counting will be performed at 100 x magnification
If more than 20 mineral wool fibers are found, continue counting and recording all of the fibers
above and below the index oil refractive index Report both raw fiber counts per refractive index
category and number of fibers from each category per gram of sample Continue on to step 1221
to determine the ratio of slag wool to other fibers with refractive index greater than 1 55 (or 1 605)
using EDS as described below
If less than 20 mineral wool fibers are found on each slide, count the number of slag wool fibers
using SEM/EDS and report as number of fibers per gram of sample
12.2 Analysis by SEM/EDS
12.2.1 Screening for Slag Wool
Operating conditions for the JEOL 6460-LV SEM are 15 kV, 0 5-5-nA beam current, 10-
mm working distance (system dependent), and zero degree tilt
Place the more concentrated sample deposited directly on the polycarbonate/adhesive tab
substrate into the SEM Use the backscattered electron mode at lOOx magnification to
quickly distinguish carbon fibers from inorganic fibers (carbon fibers may be visible, but
not as bright in a BEI) Identify all inorganic fibers over 25 um in length (smaller fibers
cannot be reliably detected at the 1 OOx operating magnification) When an inorganic fiber
is found, identify the composition of the particle by EDS Slag wool is the primary fiber
of interest Record all inorganic fiber results as number of fibers for each fiber type
For the samples with high fiber loading, as determined by PLM as descnbed in section
121, count fibers per type until a statistical representation of the ratios of fiber
compositions in the sample is achieved Report the ratio (by fiber number) of slag wool
fibers to total MMVF fibers corresponding to the high RI Use this ratio to correct the
total number for high RI fibers counted by PLM to number of slag wool fibers present
For the samples with low fiber loading, as determined by PLM as descnbed in section
121, scan the entire stub to determine the number of fibers per type Report the slag
wool fiber results as the number of slag wool fibers/gram of sample
12.2.2 EDS Screening for Gypsum/Anhydrite
Place the more concentrated sample deposited directly on the polycarbonate/adhesive tab
substrate in the SEM Choose a random field at lOOx magnification and perform an EDS
analysis on the entire field Look for the presence of sulfur in this field If sulfur is
present, continue to Section 12 2 3 or 12 2 5 for analysis of gypsum and concrete by
mapping or particle analysis If it is not present, repeat the analysis on another random
field If sulfur is still not present, mark the sample as non-detect (ND) for sulfur
37
-------
12.2.3 X-Ray Mapping for Gypsum
Place a more dilute sample, deposited directly on the polycarbonate/adhesive tab
substrate or prepared by filtration, in the SEM Collect binary backscattered electron
images (particles white and background black, shadow off) and secondary electron
images for 10 non-overlapping, random fields at 500 x magnification Collect x-ray
maps for Na, Mg, Al, Si, S, Ca, and Fe at each of these fields Fields containing MMVF
will not be used for this analysis Operating parameters for the SEM are the same as
those for analyzing slag wool Acquisition parameters for x-ray mapping using the
NORAN System Six Software are time constant 14 (mapping mode, 11333 cps), 10-20 %
deadtime, 256 x 256 image resolution, 20 second frame time, and 100 frames collected
(about 40 minutes total acquisition time) Secondary electron images will be used for
reference only Save all of the maps and electron images in TIFF format
Open the backscattered electron image and the Ca and S x-ray maps in Adobe Photoshop
Make sure that all of the element maps are the same size and resolution by choosing
Image Size from the Image Menu and changing the pixel size or the resolution as needed
The presence of gypsum can be determined by overlapping the Ca and S maps
Perform the following functions in Adobe PhotoShop (A macro is in development to
perform the following functions to decrease user time and human errors in adjusting the
threshold )
1 Convert each of the three images to grayscale (Image —> Mode —» Grayscale)
2 Perform an auto contrast and brightness on each image and map to increase the scale
of colors (Image —> Adjustments —» Auto Levels)
3 Threshold each element map, Ca and S (do not analyze the backscattered electron
image at this time), by going to the Image Menu and choosing Adjustments —>
Threshold Adjust the threshold to 128 The background will be black and the
particles white
4 Invert the image (Image—* Adjustments —>Invert) to make the background white and
the particles black
5 Copy the S map and paste it over the Ca map in a separate layer in the file and
change the opacity (located in the Layers window) to 50 % for the S map layer The
black areas are gypsum/anhydnte
6 Display a histogram of the image in expanded mode by selecting the Histogram tab
on the Navigator Window (or under the Image Menu in some versions of
Photoshop) Place the cursor over the line for the black area and record the
percentile for the black area This is the percentage of particles containing Ca and S
in the entire field
NOTE If a binary backscattered electron image is obtained during data collection, then
steps 7-11 may be deleted The Invert function will, however, need to be applied to make
the particles black and the background white before continuing to step 12
7 Begin analysis of the backscattered electron image Select the particles by going to
the Select Menu and choosing Color Range Go to the selection pulldown menu and
choose Highlights
8 Fill the selection with black by going to the Edit Menu —» Fill and choosing black
from the color pulldown menu
9 Select the inverse areas by going to the Select Menu and selecting Inverse
10 Fill the selection with white by going to the Edit Menu —> Fill and choosing white
from the color pull down menu
11 Deselect the area by clicking on the image
12 Perform the Threshold and Histogram functions for the backscattered electron image
38
-------
as outlined in 3 and 6 Record the histogram result for the backscattered electron
image
Determine the area percent of gypsum by performing the calculations in Section 13 0
12.2.4 X-Ray Mapping for Ca-Rich Particles
Analysis of components of concrete will be performed on the same fields as the
gypsum/anhydnte analysis At this time, only a method for the determination of the area
percent of Ca-nch particles is presented See Section 2 1 for discussion
Perform the following steps on the Ca x-ray map Tiff file in Adobe Photoshop
1 Convert the Ca x-ray map to grayscale (Image -» Mode —> Grayscale)
2 Perform an auto contrast and brightness on the map to increase the scale of colors
(Image —» Adjustments —» Auto Levels)
3 Threshold the Ca map by going to the Image Menu and choosing Adjustments —>
Threshold Adjust the threshold to 128 The background will be black and the
particles white
4. Invert the image (Image—» Adjustments —>Invert) to make the background white and
the particles black
5 Display a histogram of the image Place the cursor over the line for the black area
and record the percent lie for the black area This is the area percent coverage of
particles containing Ca in the entire field
Determine the maximum area percent coverage of non-gypsum, Ca-nch particles by
performing the calculation in Section 13 0
12.2.5 Particle Analysis for Identification of Gypsum and Concrete.
Place the more dilute sample, deposited directly on the polycarbonate/adhesive tab
substrate or prepared by filtration, in the SEM Particle analysis will be used to identify
gypsum and concrete particles
Perform particle analysis at 500 x magnification All other operating parameters for the
SEM are the same as those used to analyze for slag wool (Section 1221) A binary
backscattered electron image should be used in particle analysis mode Particle analysis
parameters should be set to analyze all particles in the field greater than 0 5 um and to
separate touching particles For particles greater than 5 um, scan the entire particle, spot
analysis is adequate for smaller particles The x-ray spectrum and counts for all particles,
and an image of particles > 20 um long, will be recorded and saved Other particle
parameters to be reported will include the maximum, minimum, and average diameters,
the aspect ratio, and area of each particle
It will be necessary to review data collected by automated software to ensure data
integrity An Excel spreadsheet, in conjunction with images and x-ray data, may be used
for this purpose Particles should be sorted into one of three categories Ca-S (gypsum),
Ca-nch, and Other Aid in identification of particles may by facilitated by referencing
the U S Geological Survey's WTC Dust Particle Atlas (1) A particle classification
protocol will be developed based on the data from the validation study
The number of particles analyzed will be determined using the results of the validation
study For the study, the area percent of each component should be within 10% relative
error or better Typically, data for 1000 - 1200 particles should be acquired
39
-------
Results for particle analysis will be recorded as area percent gypsum and area percent concrete
particles for each field and average area percent for the each component in the sample
13.0 Data Analysis and Calculations
Table 60 To determine the concentration of slag wool in fibers/gram, perform the
following calculations
Determine the number of fibers with RI > 1 55 (or 1 605)
# fibers identified - mg of sample on slide * 1000 = fibers/gram on slide
Determine the percentage of fibers with the composition of slag wool with RI > 1 55 (or 1 605)
Fibers/gram on slide * # fibers identified as slag wool = fibers slag wool/gram on slide
Total number of fibers identified by EDS with RI > 1 55 (or 1 605)
Back calculate to the number of fibers per gram of the original sample
Fibers slag wool/g on slide x g after sieving * g sample after ashing = Total f/g of sample
g before sieving x g sample before ashing
Table 61 To determine the area percent of gypsum/anhydrite from the x-ray mapping
procedure, perform the following calculations
Determine the area percent of gypsum/anhydrite in each field of view
% of black area in Ca-S map overlay x IQO = area % gypsum
% of black area in BSE image
Calculate the average percentage of gypsum/anhydnte for the sample
(area % gypsuml^ + (area % gypsum^ + = Avg area % gypsum
number of fields
Table 62 To determine the maximum area percentage of Ca-nch particles, which includes
concrete particles, from the x-ray mapping procedure, perform the following calculations
Determine the area percent of non-gypsum Ca-nch particles in each field of view
(% black area Ca map") - (% black area Ca-S mapl = % non-gypsum Ca-nch particles
% black area on BSE image
Calculate the average percentage of non-gypsum Ca-nch particles for the sample
(area % Ca-nch particlesV + (area % Ca-nch particles')?; + = Avg area % Ca-nch particles
number of fields
Table 63 Calculate the area percent for gypsum and concrete by summing the areas of
each particle in for each particle type and dividing by the total area analyzed
area gypsum 1 + area gypsum 2 + x 100 = area percent gypsum (do likewise for concrete)
total area analyzed
Rules for concrete and gypsum classification are currently being developed
40
-------
14.0 References
Lowers, Heather A , Meeker, Gregory P, Brownfield, Isabelle K, 2005 World Trade Center
Dust Particle Atlas U S Geological Survey Open-File Report 2005-1165 On the web at
litto //pub" uses gov/of/2QOS/l 165/
Meeker, G P, Taggart, J E, and Wilson, S A , 1998 A Basalt Glass Standard for Multiple
Microanalytical Techniques Proceedings Microscopy and Microanalysis 1998 Microscopy
Society of America
A polished and carbon coated calibration reference sample of BIR1-G may be obtained by
contacting Stephen Wilson, U S Geological Survery, MS 973, Denver Federal Center, Denver,
CO, 80225, svvilson^usgs gov
Perkins, R L and Harvey, B W, 1993, TEST METHOD Method for the Determination of
Asbestos in Bulk Building Materials, EPA/600/R-93/116
41
-------
15.0 Appendix:
DATA SHEETS
Determination of Slag Wool Fibers in Dust- PLM with Dispersion Staining
Sample ID Project
Analyst
Circle One Original Duplicate Triplicate Date
General Sample Appearance.
Homogeneous?
Structure #
Rl Fluid
_155_
Dispersion Staining
Becke Line
Fiber
non-MW
chrvsotlle
Comments
42
-------
SEM Sheet
Reference ASTM - D5755-03
Report Number1 _
Sample Number:.
File Name.
Sample Description:
Preparation Date:
Analysis Date:
Computer Entry Date:
Sample weiaht:
Dilution Volume:
Volume Aliquot:
Magnification:
Bv:
Bv:
Bv:
qrams
ml
uL
X
Structure #
Field #
Fiber Type Length (Microns) Width (Microns)
Image
EDS
43
-------
TJGbTET
Ota
-------
APPENDIX E: REPORT FROM THE U.S. EPA CONTRACTOR ON THE SCREENING
METHOD STUDY
Versar
6850 Versar Center
Springfield, VA 22151
Ms Jacky Rosati
US Environmental Protection Agency
E-305-03 109 TW Alexander Drive
Research Triangle Park, NC 27711 July 21,2005
Dear Ms Rosati
Attached is a preliminary report based on analytical data thus far received, for dust
samples collected primarily m the New York City area Most of the samples were taken in areas
that, it is believed, were not affected by paniculate matter generated during the World Trade
Center (WTC) collapse (i e , background samples) Some of the samples were spiked with one
or the other of two dusts that are believed to have originated from the WTC collapse. The
analytical protocol was developed by the government, specifically for this project, and was
modified as the project developed. The purpose of the testing was to determine if the spiked
background dusts could be distinguished from those samples that were not spiked
Three parameters were measured to make this determination- (1) slag wool fiber content,
(2) calcium-rich particle content, and (3) gypsum particle content
The analytical data indicate that
• With respect to calcium-nch particles and gypsum particles, spiked samples cannot readily be distinguished
from background samples
• With respect to slag wool content in the samples spiked with the first of the two WTC dusts, spikes at the
10% level may be statistically identifiable as WTC-contammation, although spikes at or below the 5% level
are probably not identifiable
• With respect to slag wool content, samples spiked with 5% and 10% of the second of the two WTC dusts
are easily identifiable as WTC-contammated Even at the 1% spike level, samples may be statistically
identifiable
The attached preliminary report will explain the above conclusions in more detail However, it must be
noted that all of the analytical data from the eight laboratories that performed the analysis has not yet been
received Nevertheless, it is believed that the above conclusions will not likely change once those additional
data are incorporated
Sincerely,
Stephen M Schwartz, P E, Q E P
Project Manager
45
-------
Versar
Preliminary Report
of
Analysis of New York City Area Dust Samples
Purpose:
The objective of this study is to determine if New York City area dusts that are contaminated with
varying levels of dusts known to originate from the collapse of the World Trade Center (WTC) can be
distinguished from background dusts that are believed not to be contaminated with WTC dusts
Project Summary:
In the initial portion of the testing, 10 dust samples from New York City areas that are believed
not to be contaminated with dusts originating from the collapse of the WTC were used These are
referred to as the first set of background samples An additional background dust sample was spiked at
1, 5, and 10 percent levels (by weight) with dust believed to have originated from the WTC collapse
An additional background sample was spiked at 1, 5, and 10 percent levels with a second dust sample
that is believed to have originated from the WTC collapse Therefore, a set of 16 samples was
generated
• 10 different background dusts
• 3 samples, each consisting of one background dust sample spiked with one source of WTC
dust at 1, 5, and 10% levels
• 3 samples, each consisting of one background dust sample spiked with a second source of
WTC dust at 1, 5, and 10% levels
Initially, 32 samples were sent to each of eight analytical laboratories (three U S government, and
five private) The 32 samples consisted of two identical sets (i e, duplicates) of the 16 samples
discussed above The pnvate laboratones did not know that there were duplicate samples Further,
they did not know which, if any, of the samples contained WTC spikes
Subsequently, a second set of 28 different background samples was analyzed to obtain a better
understanding of the variability of background dusts These 28 samples were sent to only one of the
five private laboratones
It was ultimately agreed that each of the laboratones would perform the following
three Scanning Electron Microscopy-based (SEM) analyses on each of the
samples they received (see Methodology and Data Analysis section)
• Slag wool fiber content (in number of fibers per gram of dust) Slag wool was a significant
component of the WTC insulation material
• Calcium-rich particle content (in area percent concentration in the SEM field) Such particles
are assumed to be indicative of cement/concrete-like particles
• Gypsum particle content (in area percent concentration in the SEM field) Such particles are
assumed to be indicative of "dry wall" (i e , gypsum-containing wall board)
Conclusions:
-------
A number of conclusions can be drawn from the analytical results thus far
obtained. It is not expected that data that are subsequently received will
substantially change these conclusions. It must be noted that there are several
caveats that affect the quality of the data Those are discussed later in this report
2 With respect to calcium-rich particles and gypsum particles, spiked samples
cannot readily be distinguished from background samples
Tables 1 and 2 present the analytical data thus far available for calcium-rich and
gypsum content respectively Analysis was performed using SEM and x-ray
mapping (XRM) techniques The shaded areas represent the samples spiked
with 1, 5, and 10% WTC dust. The others areas are background samples
Sample designations followed by "(1)" and "(2)" are duplicate samples
(Samples received by the laboratories had random identification numbers, so
that the laboratories did not know if any samples were duplicates, nor did they
know if any samples contained WTC dust) In addition, Table 3 is the
analysis of a subsequent 28 background samples, analyzed by only laboratory
"B" Analysis of calcium-rich and gypsum particles for this sample set is
shown on Table 3
The average of all background samples (including the second set of 28 samples) for
calcium-rich particles is 22 3 area percent, with a high value of 66 5% and a
low value of 4 2% The average for the spiked samples is 20 7%, with the
highest value being 25 9% The 1, 5, and 10% spiked samples do not show
any trend with respect to calcium-rich particle content (i e., they do not show
any increase as the spike level increases).
The average of all background samples (including the second set of 28 samples) for
gypsum particles is 11 7 area percent, with a high value of 56 5% and a low
value of 0 1 % The average for the spiked samples is 9 3%, with the highest
value being 32 8% The 1,5, and 10% spiked samples do not show any trend
with respect to gypsum particle content
3. With respect to slag wool content in the samples spiked with the first of the
two WTC dusts, spikes at the 10% level may be statistically identifiable as
WTC-contammation, although spikes at or below the 5% level are probably
not identifiable
Table 4 presents all the analytical data thus far available for SEM slag wool fiber
analysis (as the number of slag wool fibers per gram of dust). The shaded
areas represent samples that are spiked at the 1, 5, and 10% levels with WTC
dust Table 3 also presents additional slag wool fiber background-only sample
data (next to last column) It can be seen from Figure 1 that for those spiked
samples designated as "DB" that at the 5% spike level, the slag wool
concentrations probably do not exceed one standard deviation above the
average slag wool background concentration (including the Table 3
47
-------
background data) However, at the 10% spike level, the slag wool
concentration typically exceeds one standard deviation (see Figure 2), but
never exceeds two standard deviations above the average background sample
concentration The average background concentration is about 27,400 fibers
per gram. The standard deviation is about 40,100 fibers per gram '
It should be noted that there is a trend showing a clear increase in slag wool fiber
concentration from the 1% to the 10% spike level (see "DB" sample shaded
area on Table 4) However, the numerical values of those concentrations, as
noted above, are still less than two standard deviations above the average
concentration.
4. With respect to slag wool content, samples spiked with 5% and 10% of the
second of the two WTC dusts are easily identifiable as WTC-contaminated
Even at the 1% spike level, samples may be statistically identifiable.
The slag wool content data for the samples spiked with the WTC dust shown in Table
4 as "USGS" are easily identifiable As can be seen in Figures 4 and 5,
samples spiked with the USGS WTC dust at the 5 and 10% levels are
essentially all more than two standard deviations above the average
background sample concentration. (Average plus two standard deviations
would be about 108,000 fibers per gram.2) At the 1% spike level though,
WTC dust is more difficult to identify because the slag wool concentrations
are mostly between one and two standard deviations above the average
background sample (see Figure 3)
5 With respect to slag wool content, clearly, there is a large difference between
the two WTC dust spikes used In the "DB"-spiked samples, as noted above,
it is expected to be more difficult to determine a significant slag wool fiber
concentration difference from background The "USGS"-spiked samples
clearly had significantly more slag wool fiber content than the "DB" samples
6 Examining Tables 1, 2, and 4 and the Figures, it can be seen that the analyses
for the duplicate samples rarely replicate one another However, the variation
between duplicate sample values (i e, intralab) is about half of the variation
between individual laboratory values (mterlab).3
1 Background concentration data for this analysis excluded several samples that were known to have high
slag wool content, specifically the Cl-RTP samples (see Table 4), and samples C2,3,4,5,6 (see Table 3)
2 Ibid
3 For slag wool fiber analysis, the average difference between the analyses of duplicates (i e, intralab
differences) is about 50% of one standard deviation of the between-laboratones analyses (i e, inUrlab
differences) For both calcium-rich and gypsum particle analysis the average intralab difference is 20% of
the interlab difference
48
-------
Methodology and Data Analysis:
The analytical protocol was developed specifically for this project by one of the
government laboratories, and modified by all laboratory participants at a meeting held
for that purpose All laboratory participants held weekly conference calls as the
analytical program was proceeding to discuss general issues with the protocol
Additional modifications were made to the protocol based on those conference calls
The original protocol included analysis by Polarized Light Microscopy (PLM), so
data are also available for PLM analysis The PLM analyses were curtailed because it
became obvious that PLM could not adequately differentiate between fiber types
Further, total fiber concentrations were also determined, both by PLM and SEM
methods, but those data are not presented in this report
Caveats:
There are a few factors that may contribute to data uncertainty Nevertheless, it is
unlikely that these factors will alter the above major conclusions Some of these factors
are as follows:
1 As noted earlier, not all of the analytical data have been received
2 Dust samples were collected by several methods Evaluation of the
sampling methodology was not part of the study
3 To determine fiber concentration, fibers were counted using an SEM
Different laboratories diluted samples to different levels before counting,
introducing some variability of results
4 Laboratory equipment capabilities and personnel skills varied
49
-------
TABLE 1: SEM X-Ray Mapping - Calcium-Rich Area Percent
Sample
Designations
AP5(1)
AP5(2)
CMC(1)
CMC(2)
HS3(1)
HS3(2)
WGS(1)
WGS(2)
MW(1)
MW(2)
DB1%(1)
DB1%(2)
DB5%{1)
DB5%(2>
DB10%(1)
DB10%(2)
C1-RTP(1)
C1-RTP(2)
USGS1%{1)
USGS1%(2)
USGS5%{1)
USGS5%(2)
USGS10%(1)
USGS10%(2)
USC(1)
USC(2)
FP(1)
Laboratory Letter Codes
A
B
234
224
277
341
103
178
228
199
122
120
18.2
131
13,4
20,5
14.4
126
132
162
16.5
210
148
146
17.0
179
123
94
130
C
156
230
18.9
26.6
26.4
D
204
221
21 8
21 5
141
133
132
163
142
109
14.0
141
161
129
15,2
149
11 3
11 9
174
111
142
162
15.8
121
11 1
95
11 6
E
144
168
67
204
64
148
139
57
126
83
183
15.4
108
4,1
8.6
10.8
75
56
124
57
11 9
10 .7
109
98
195
66
59
F
11 6
98
79
101
153
65
77
74
76
57
8,9
9.4
9.0
75
8.0
81
61
42
7.6
74
6.7
83
8.9
83
54
71
62
G
307
396
551
389
291
440
584
493
55,9
400
50.7
571
435
424
H
458
489
604
552
631
497
534
523
463
498
502
522
490
39,6
406
48.8
665
61 0
432
419
536
51.8
400
45.1
461
403
709
Location Key:
Chittenden Avenue, Manhattan
Columbia Medical Center, W68*1 Street, Manhattan
Teaneck, NJ
Nassau County, LI
West End Ave Between 105* and 106* Streets, Manhattan
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
Research Triangle Park, NC
Research Triangle Park, NC
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
Federal Courthouse, White Plains, NY
Federal Courthouse, Central I slip, LI
-------
FP(2)
MUNYCK1)
MUNYCK2)
MUNYC2(1)
MUNYC2(2)
105
254
197
204
176
103
171
14.0
199
144
103
152
31 9
277
136
82
63
80
89
73
556
454
61 5
396
366
568
578
Northern Manhattan, Above 70th Street
Northern Manhattan, Above 70th Street
Samples spiked with WTC dust, at 1, 5. and 10% levels are shaded. All others are
background samples
51
-------
TABLE 2: SEM X-Ray Mapping - Gypsum Area Percent
Sample
Designations
AP5(1)
AP5(2)
CMC(1)
CMC(2)
HS3(1)
HS3(2)
WGS(1)
WGS{2)
MW(1)
MW(2)
DB1%(1)
DB1%(2)
DB5%(1)
DB5%(2)
DB10%{1)
DB10%[2)
C1-RTP(1)
C1-RTP(2)
Laboratory Letter Codes
A
USGS1%<1) 1
USGS1%(2)
USGS5%{1 )
USGS5%(2)
USGS10%(1)
USGS10%(2) !
USC(1)
USC(2)
B
80
203
43
69
59
149
29
61
38
54
72
71
7.3
61
65
50
85
87
63
54
77
25
63
4.8
48
62
C
13.8
3.4
8.7
152
9.8
D
144
11 3
48
30
92
11 0
54
47
70
53
5.7
5.2
55
5.5
7.8
4.8
97
82
5.8
4.1
5.7
4.1
7.1
4.8
52
42
E
09
1 8
02
1 0
03
23
02
02
02
01
3.0
1.1
0.7
0.1
0.5
0.€
02
03
1.0
0.2
09
0.5
1.2
0.7
1 2
02
F
25
1 6
1 1
1 0
57
1 5
04
03
07
1 1
0.6
1.3
1.2
1.6
1.0
1 9
1 3
08
0.9
09
1.1
24
1 1
1.4
07
24
G
341
31 3
261
308
440
290
190
232
22.0
291
25.7
245
249
H
261
337
224
176
429
405
422
391
376
41 6
28.0
30.0
24.3
28.9
27.0
28.5
534
504
29.4
29.2
29.3
21.7
30.9
328
271
324
Location Key:
Chittenden Avenue, Manhattan
Columbia Medical Center, W68*1 Street, Manhattan
Teaneck, NJ
Nassau County, LI
West End Ave Between 105* and 106* Streets, Manhattan
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
Research Triangle Park, NC
Research Triangle Park, NC
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
Federal Courthouse, White Plains, NY
52
-------
FP(1)
FP(2)
MUNYC1(1)
MUNYCK2)
MUNYC2(1)
MUNYC2(2)
11 6
44
105
30
55
42
54
61
92
55
61
60
03
06
1 2
1 4
92
07
1 2
1 5
09
1 0
25
1 8
245
268
31 0
565
400
241
263
308
295
Federal Courthouse, Central I slip, LI
Northern Manhattan, Above 70th Street
Northern Manhattan, Above 70th Street
Samples spiked with WTC dust, at 1, 5. and 10% levels are shaded. All others are
background samples
53
-------
Table 3: New York City Background Dust Samples
Sam
pie
No.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
EPA Sample ID
HS1 -06-01"
HS1 -06-02
AP2-07-01
AP2-07-02
AP3-08-01
AP3-08-02
HS2-09-01
AP4-10-01
AP7-14-01
CMC-17-01
HS3-18-01
WGS6557
WGS5826-1
PT1 52W88
PT1 52W88-2ndFI
CY321W80
MW924WEAve
C2**
C3
C4
C4 (no date)
C5
C6
N-01S
SEM XRM
Calcium-
Rich (area
%)
58
163
124
99
105
173
256
131
175
305
143
102
181
175
158
147
21 1
79
168
137
175
167
100
123
Gypsum
(area %)
58
40
85
72
47
76
96
11 5
60
104
90
42
62
84
56
122
101
61
34
38
88
72
109
78
SEM (Heavv
Slag Wool
(fibers/g)*
35,565
230,769
32,432
7,692
12,500
<3.636
7,605
42,857
3,333
4,651
11,858
34,826
15,564
17,021
19,305
30,888
8,097
46,703
170,309
160,772
488,372
74,236
280,762
369,231
Loading)
Total
Fibers
(fibers/g)
104,603
523,077
113.514
130.769
212.500
21.818
22.814
485,714
23,333
23,256
71,146
44,776
54,475
46,809
42,471
34,749
28,340
102,890
321 ,696
227,760
790,698
148,472
415,039
523,077
Particle Count
Slag
Wool
9
15
6
2
2
0
2
3
1
1
3
7
4
4
5
8
2
11
18
24
21
17
23
24
Total
Fibers
25
34
21
34
34
6
6
34
7
5
18
9
14
11
11
9
7
24
34
34
34
34
34
34
Stony Brook, LI
West End Ave between 72nd and 73rd Streets
30th Ave between 21st and 23rd, Queens
70th Street between 20th and 21st Ave, Brooklyn
79th St between York and East End Ave, Manhattan
92nd Street between Columbus and CPW, Manhattan
Columbia Medical Center, W 168th St, Manhattan
Teaneck, NJ
Nassau County, LI
Nassau County LI
88th Street between Amsterdam and Columbia, Manhattan
88th Street between Amsterdam and Columbia, Manhattan
80th Street between Riverside and East End Ave, Manhattan
West End Ave between 105th and 106th Streets
Research Triangle Park, NC
Research Triangle Park, NC
Research Triangle Park, NC
Research Triangle Park, NC
Research Triangle Park, NC
Research Triangle Park, NC
Edison, NJ
54
-------
25
26
27
28
Aver
age
Stan
dard
Devi
ation
Coeff
. Of
Varia
nee
Nevms Ct
E Curtis Ave**
LBI
Mixture
160
79
73
191
14.9
5.5
0.4
96
90
163
11 8
8.1
3.0
0.4
<4,367
5,173
<3,636
7,194
84,709
128,759
1.5
91,703
24,138
61,818
35,971
168,837
200,808
1.2
0
2
0
2
21
7
17
10
Edison, NJ
Edison, NJ
Long Beach Island, NJ
NE Queens
* A fiber count of one fiber was used to calculate the analytical sensitivity for non-
detects
" Internal laboratory duplicates were run on these samples The result shown is the
average of the two duplicates ("<" samples were assumed to be 0}
55
-------
TABLE 4: SEM - Slag Wool Fiber Count/Gram of Sample
Sample
Designations
AP5(1)
AP5(2)
CMC(1)
CMC(2)
HS3(1)
HS3(2)
WGS(1)
WGS(2)
MW(1)
MW(2)
DB1%(1>
DB1%(2)
DB5%(1)
DB5%(2)
DB10%(1)
DB10%{2)
C1-RTP{1)
C1-RTP(2)
USGS1%(1)
USGS1%(2)
USGS5%(1)
USGS5%(2)
USGS10%(1)
USGS10%(2)
USC(1)
USC(2)
FP(1)
FP(2)
Laboratory Letter Codes
A
non-det
non-det
16.393
5,900
12,232
5,747
34,826
72,562
67,797
104.575
84,746
246.914
98.039
600,000
1,218.855
73,394
18,519
B
3.663
<3636
3,448
<3875
7,299
7,692
34,221
10.753
18,939
3,717
10.909
17.422
29.197
25,271
66.421
77.778
159,011
173,585
109,091
83,032
404,332
343,284
840,231
1.366,470
56,025
41.199
18,051
16,470
C
5,451
9.133
32.385
33,646
74.837
57.644
50.293
50.160
364,813
531,277
521,212
D
non-det
6,980
11,800
9,620
19,000
18,600
26,400
18,100
18,700
31,800
29.900
27,300
50.800
35,800
113,000
95.100
269,000
165,000
119,000
104,000
681,000
146,000
1,620.000
238,000
91,800
40.700
16.300
31,800
E
<249
<667
<282
309
<286
<667
<256
6,990
1,320
893
<2,000
3,770
31,000
6,900
108,000
20.400
168,000
21,900
366.000
18,700
227,900
191,000
1.410.000
271,000
33,700
7,890
1,100
3,920
F
<500
500
<4,500
667
2,750
5,060
1,630
<30,500
1,000
<45,500
1.920
12,500
1.700
14,700
7.000
34.100
38,000
160,000
79.800
79,500
433,000
197.000
629,000
372,000
15,600
48,400
12,400
30,500
G
2,470
13,910
5,780
6.100
<6.320
7,370
9,480
3,520
13,630
18,080
7.650
1.320
6,230
13.040
12,900
25.210
84,650
39,930
9.200
25.370
66,450
73.330
144,120
33,040
<3,230
3,540
11,920
<1,181
H
<7.386
<7,698
<7,241
<6.289
<7,576
34,813
16,077
18,399
17,301
<9,497
15,924
16.038
107.143
70,472
114,638
96.696
188,088
318.143
90,992
137,363
672,926
347,904
734.767
413,153
29,268
74,212
28,249
25,489
Location Key:
Chittenden Avenue, Manhattan
Columbia Medical Center, W68* Street, Manhattan
Teaneck, NJ
Nassau County, LI
West End Ave Between 105*1 and 106th Streets, Manhattan
»
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
4 Albany Street Spiked into NE Queens background dust
Research Triangle Park, NC
Research Triangle Park, NC
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
Federal Courthouse, White Plains, NY
Federal Courthouse. Central Islip, LI
56
-------
MUNYCK1)
MUNYCK2)
MUNYC2(1)
MUNYC2(2)
10.840
41,298
7.220
3.745
28.777
48.507
14,400
20.200
66,500
45,500
14,900
1.960
1.390
24,200
13,100
<22,300
17,800
30,500
<2,545
<1,228
<12,453
2,330
6,803
41,118
123,106
59,473
Samples spiked with WTC dus^ at 1, 5, and 10% levels are shaded. All others are background samples.
For data analysis purposes
• Non-det = Non-detect - zero slag wool fibers were noted in the sample
• <# indicates that the value was less than the detection limit of the respective laboratory
by the square root of 2.
Northern Manhattan, Above 70 Street
Northern Manhattan, Above 70* Street
When this result was reached, the value was divided
57
-------
Table 5: SEM - Slag Wool Fiber Count
Sample Designations
AP5(1)
AP5(2)
CMC(1)
CMC(2)
HS3(1)
HS3(2)
WGS(1)
WGS(2)
MW(1)
MW(2)
DB1%m
DB1%(2)
DB5%(1)
DB5%(2)
DB10%(1)
DB10%(2)
C1-RTP(1)
C1-RTP(2)
USGS1%(1)
USGS1%(2)
USGS5%(1)
USGS5%(2)
Laboratory Letter Codes
A
0
0
3
1
2
1
7
8
12
16
15
20
15
99
B
1
0
1
0
2
2
9
3
5
6
3
5
8
7
18
21
45
46
30
23
112
92
C
1
2
7
7
12
12
9
11
62
D
0
3
5
4
8
8
11
8
8
14
13
12
22
16
48
42
116
72
54
47
194
65
E
0
0
0
1
0
0
1
7
6
2
0
4
7
8
•V3
9
16
30
27
23
25
21
F
0
1
0
0
1
3
1
0
1
0
1
2
1
2
2
3
4
4
11
4
>20
19
G
2
6
4
5
0
4
6
3
6
22
7
1
6
11
10
25
22
17
11
H
0
0
0
0
0
4
2
2
2
0
2
2
6
10
13
12
24
37
10
22 I 15
64
43
27 ; 39
Location Key:
Chittenden Avenue, Manhattan
Teaneck, NJ
Nassau County, LI
4 Albany Street Spiked into NE Queens background
dust
4 Albany Street Spiked into NE Queens background
dust
4 Albany Street Spiked into NE Queens background
dust
4 Albany Street Spiked into NE Queens background
dust
4 Albany Street Spiked into NE Queens background
dust
4 Albany Street Spiked into NE Queens background
dust
Research Triangle Park, NC
Research Triangle Park, NC
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
58
-------
USGS10%(1)
USGS10%(2)
USC(1)
USC(2)
FP(1)
FP(2)
MUNYCK1)
MUNYC1(2)
MUNYC2(1)
MUNYC2(2)
181
6
3
1
7
45
45
13
11
5
5
2
1
8
13
124
129
450
105
39
18
7
14
6
9
28
20
38
19
6
6
3
4
4
2
3
24
19
16
13
4
2
1
3
0
3
3
18
9
0
2
2
0
0
0
0
1
41
49
3
8
3
3
1
5
13
7
USGS Dust Spiked into NE Queens background dust
USGS Dust Spiked into NE Queens background dust
Samples spiked with WTC dust, at 1,5. and 10% levels are highlighted in yellow
59
-------
ADDENDUM TO VERSAR REPORT: SEM CALIBRATION DATA
MEMORANDUM
TO. Jacky Rosati
CC. David Friedman
FROM Stephen Schwartz
DATE August 5, 2005
SUBJECT BIR-1G Sample Analyses
Identical mounted and polished reference samples, each designated BIR-1G, were
sent to each of the five private laboratories participating in the analyses of dust samples
from New York City and elsewhere The samples were analyzed by Scanning Electron
Microscopy/Energy Dispersive X-ray Spectrometry (SEM/EDX) to determine their
elemental content The purpose of the study was to determine the variation within and
between each of the laboratories, and to assess their ability to identify elements using this
technology
Each of the five laboratories analyzed their BIR-1 G sample between 4 and 11
times, as convenient (there was no requirement for a specific number of analyses). The
average elemental concentration data for each laboratory is presented in the attached
table For calcium (Ca), magnesium (Mg), silicon (Si), and oxygen (O)4, which
constitute over 80% by weight of the elemental composition, the standard deviation
within each laboratory, for each element, was typically much less than 10% (i.e, the
coefficient of variation) Likewise, the coefficient of variation between laboratories for
Ca, Mg, Si, and 0, as shown on the attached table, was also much less than a 10% (The
graphic presentations of the EDX spectra within and between laboratories also appear to
be extremely similar.
Therefore, it can be concluded that each of the laboratories was easily able to
achieve excellent precision, by SEM/EDX, in quantifying the elements that were present
in larger concentrations
4 Some of the laboratories reported results as the weight percent of the elemental oxides, specifically Na2O,
MgO, A12O3, SiO2, CaO, TiOj, FeO, K2O, and MnO2 Oxide values were converted to individual elemental
values (e g, A12O3 is about 53% Aluminum, and 47% oxygen by weight)
60
-------
AVERAGE ELEMENTAL CONCENTRATION REPORTED FOR BIR-1G SAMPLES (Weight Percent of Sample)
Lab
A
B
C
D
E
F
G
H
AVERAGES
Standard Dev.
(%of
Average)
Sodium
044
1 14
1 84
1 14
61 40
Magnesium
502
582
510
531
833
Aluminum
811
875
788
825
548
Silicon
21 60
2476
2220
2285
735
Calcium
966
768
1030
921
1483
Titanium
068
052
064
061
1358
Iron
968
556
837
787
2673
Potassium
NR
000
003
002
141 42
Manganese
NR
000
021
011
141 42
Oxygen
4484
4578
4343
4468
265
NR -NotReported
61
-------
APPENDIX F: STATISTICAL ANALYSIS AND INTERPRETATION OF TEST
RESULTS- LABORATORY QUALIFICATION
Slag wool fiber content as a discriminator for residual WTC contamination in
indoor dust sample: Interpretation of multi-laboratory test results
Introduction
Eight laboratories were each challenged with a number of blinded dust sample aliquots to
determine number of slag wool fibers per gram These samples included background
dusts from various locations in NYC and a series of samples of common household dust
spiked with different levels of WTC collapse dust collected in 2001 or in 2004 The
purpose of this endeavor was to assess whether or not the method developed can be used
by qualified laboratories to discriminate between WTC and non-WTC impacted dust
samples.
In the following discussion, individual laboratories are evaluated and ranked for validity
and precision, and then the top performers are further evaluated as groups to determine
the expected confidence level for the slag wool content of any individual and randomly
assigned)sample.
Laboratory Qualification
Validity: Assessment of validity was conducted by analysis of a series of spiked samples
where the expected response ratios are known. The challenge samples consisted of a large
volume of non-impacted background dust collected in 2004 from locations in Northeast
Queens over ten miles from the WTC site This dust was subsequently spiked with 1,5,
and 10% WTC dusts by weight using either bulk collapse dusts collected in September
2001 immediately following the disaster (designated as USGS dust), or nominally
undisturbed dusts collected in 2004 in the abandoned Deutsche Bank (DB) complex that
borders the south side of the WTC complex (designated as 4 Albany dust).
Using units of (# slagwool fibers)/(gram of dust), preliminary analyses showed a mean
value of 12,200,000 (s d 1,697,056) for USGS dust, 579,667 (s d 173,782) for 4
ALBANY dust, and a nominal background level of 7,190 Based on these data, the
expected values of slope expressed as [(# slagwool fibers)/(gram dust)]/[% spike level]
are 121,928 and 5,725, respectively for USGS and 4 Albany spikes Specifically, each lab
was furnished two samples each of 1, 5,10% spikes from both 4 Albany and USGS series
for a total of 12 spiked samples plus a series of 20 additional background samples
collected from random locations all over the greater NYC area
Scatterplots and linear least squares regressions were constructed for each lab and for
each of the two spike series Preliminary inspection showed no apparent violations of
underlying assumptions required for regression analysis (primarily homogeneity of
variance), as such no lognormal transformation was performed. Using a forward selection
strategy, it was found that a higher order polynomial model does not statistically improve
the linear fit, this is expected as the sample set is designed as a linear progression Also, a
simpler but more general "runs" test for each linear regression confirmed these results
Data handling and manipulation was performed with Microsoft Excel SP-2, statistical
analyses were performed with SAS 9.1.3 XP-Pro (proc rsreg/lackfit, proc reg, proc
62
-------
mixed, and proc unwanate); graphing, ANOVA, and various other statistical results were
performed or verified with GraphPad Prism 3 03 The linear regression results are given
in Table 1, a summary of SAS proc rsreg/lackfit results are given in Table 2
Table r Summary of linear least squares regression results
Lab
A
B
G
D
E
F
G
H
DB spskes (4
slope
8126
6541
6554
8538
6936
1523
1630
9695
Albany)
95% C!
<*/•)
1760
971
691
1515
3632
1288
610
2645
sig siope
p-valus
0,0099
0.0025
0.0007
00049
0 *t28S
G.3027
0.0559
0,0215
f2
0,8421
G91SO
09573
08881
04769
G2SS8
05404
0.7705
USGS spikes
(2001 dusts)
slope 95% Cl ssg
124500
1 1 3300
52380
91340
74240
45370
7750
49510
{+•'-* P-
488
23460
5914
58230
49310
S40SO
4607
21790
slope
valtfs
0.0025
0.0085
00030
0,19t8
0.2104
00298
0,1873
0.0855
f2
1 0000
08537
O.S632
0,3608
0,3571
07325
0,4 144
05634
note: n =3
Table 2 Summary of "lack of fit" tests
Lab
A
B
C
D
E
F
G
H
A B.C.and D
A,B,C,D and H
r2
linear
00421
0.9190
Q.S573
0.888 1
04769
OZ588
0.6404
0.7705
07550
07027
r2
quad
00590
00564
OOOQ1
0.0706
0.0197
00445
00069
0 1455
0.0032
0.0023
p-v«lue
lin«ar
00149
0.0058
0 003$
0,0040
0.5904
03386
0,10:S
0.0 » 35
00001
0.000 1
pvaiue
quad
G2732
&079S
0944J
0 1037
07547
G6913
0.8240
01C70
06038
06473
Based on these summaries the linear model is appropriate Laboratories A and B
demonstrate excellent performance across the board each has r2 values > 0 80, significant
positive slopes with p < 0 05, and slopes with the expected magnitude We caution that
Lab A only has three points for the USGS spike results (yellow highlights, Table 1)
Fields highlighted in blue indicate potential problem areas If only the 4 Albany spike
series are considered, then Labs C and D can be added to the preferred performer group
This is reasonable because the range covered here is more likely to reflect the range of
concern for unknown samples Although the Lab H results demonstrate a lower r2 value
and a larger 95% CI for slope, this is caused by a single outlying point. As such, there is
no reason to exclude Lab H from the analysis 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
analysis
63
-------
From the above discussion, we can construct two groups of laboratories based on
the estimated validity of their results: "Best", consisting of Labs A, B, C, and D and
"very good" consisting of the best group plus Lab H. In the following series of figures,
the individual and composite linear regression results for the groups are demonstrated
graphically.
120000 •
100000 -
80000 •
60000
40000
20000
Spiked Samples - 4 Albany
"Best" Group
& LabB
D Lab A
4 LabC
* LabD
123456789
% spike level
10 11
64
-------
120000 -
100000 -
80000
60000
40000
20000
Spiked Samples - 4 Albany
"Best" Group Combined
(Labs A, B, C, and D)
Slope: 7440+-903
r2: 0.7550
01234567
% spike level
8 9 10 11
120000
100000
80000
60000
40000
20000
O LabB
D Lab A
.4 Lab C
4 LabD
* LabH
0 1
Spiked samples - 4 Albany
"Very Good" Group
(Labs A, B, C, D and H)
456
% spike level
8 9 10 11
65
-------
120000
100000 -
80000 •
60000
40000
20000
Spiked samples - 4 Albany
"Very Good" Group Combined
(Labs A, B, C, D and H)
Slope: 7891 +- 970
r2: 0.7027
0123456789 10 11
% spike level
120000
100000
80000
60000
40000
20000
O LabF
C Lab G
O LabE
0 1
Spiked samples - 4 Albany
"Outlying" Group
(Labs E, F and G)
456
% spike level
O
8 9 10 11
66
-------
Spiked samples - 4 Albany
"Outlying" Group Combined
(Labs E, F and G)
120000 -
100000 -
80000
60000
40000
20000
0 1
456
% spike level
1U 11
Precision: Up until this point, validity has been assessed only with those samples for
which there is some prior knowledge of content. For assessing precision, however, one
can use all of the samples (including unknowns) because each laboratory received
aliquots of the same set of 32 samples. Furthermore, the sample structure is such that
these 32 samples are comprised of 16 paired samples allowing within laboratory
precision estimates as well. Although there are a number of statistical options for
proceeding, an analysis of variance (ANOVA) and intra-class correlation coefficients
(ICC) are pragmatic for these circumstances as samples and laboratories are used in
groups. Preliminary analyses of "within" and "among" laboratory results indicate that the
underlying distributions (considering all 32 sample results) are not normal based on the
Shapiro-Wilk (S-W) test, and that natural log transformation of the data should used to
perform analysis of variance. The only exception is the USGS data set where only three
pairs of samples are reported and thus the natural space numbers did not require
transformation. Table 3 shows the results for the ICC analyses within laboratories, and
also for the groups (Labs A, B, C, D) and (Labs A, B, C, D, H) aggregated. The variance
components and p-values for the S-W normality test are also given. The lower part of
Table 3 gives the aggregated results for the background samples only; Laboratories A and
C did not contribute to these statistics but it is expected that they would perform
similarly.
67
-------
Table 3 Summary statistics for mtra-class correlation coefficients
Ml available Pairs
IntraClass Cotretaiion Calculations: from SAS proc mixed
Log Space data
Lab
A
AA
B
C
0
E
H
A.B.C.snd D
A,e,C,D and H
A natural space
Ali NYC Backgrou
nobs
6
5
32
10
31
32
32
32
32
80
112
Int
7.6570E-G1
1.331QE+09
2.7493E+GO
1 9870£+00
1,396e£*03
5.5988£+00
2.8559E+00
1 0962E+C3
2.064 7£* 00
2.0491E+00
2.0820E*00
Res
5.4910E-Q1
2 1025E+08
1.7470E-01
7 3Q50£'Q1
2.7860E-01
1 7150F+QG
1 3172E+GG
67860E-01
3.2810E-01
3.02SDE-01
2.8740E-01
!CC
0.5824
08536
09403
0.7312
08337
0.7655
06844
0,6176
08629
0.8714
0.8787
S-W p-valus
02786
0.767S
O.i 059
QG342
0.8847
06012
OSOS7
03571
0.1571
0.3296
0 '-540
nd Pairs
intraCiass Correlation Calculations: from SAS proc mixed
Log Space data
Ubx
A.S.C.and D
A(8,C,D and H
nobs
36
54
int
64570E-01
6-SSOOE'QI
fics
3.5070E-01
3.4290E-Q1
!CC
06430
06706
S W p value
02657
0.2571
*Laboratones A and C did not report paired New York City background data
From this exercise, we see that all of the individual laboratories demonstrate reasonable
ICCs (generally above 0 6) Furthermore, the laboratory groups chosen to demonstrate
good validity show ICCs greater than 0.87 when all data are considered. When only the
New York City background samples are analyzed, the ICCs are somewhat lower. These
results can be interpreted to mean that about 35% of the variance is attributable to
variability in the pooled laboratory analyses, and the remainder to true differences among
the background samples
As a further assessment of inter-laboratory precision, the between laboratory ANOVA
shows no reason to reject the null hypothesis (Ho = no difference, in natural log space)
among Laboratories A, B, D, and H Laboratory C was left out of this analysis because
they reported no background data at all
68
-------
Evaluation of Unknown Samples
In the previous section we qualified a group of laboratories for measurement of
unknowns based on spiked samples (validity), and comparative precision measures based
on ICC and ANOVA. We now assume that these laboratories are statistically similar and
combine their spike results into a single response graph. Based on these results, we
calculate 95% confidence intervals and 95% prediction bands as illustrated in the figures
below.
140000
120000-
100000
01
I
80000-
u- 60000-
o
I
01 40000-
V)
20000-
Splked samples - 4 Albany
Labs A, B, C, D, and H combined
0123456789 10 11 12
140000-
120000-
1
•o
E 100000-
2
Dl
^
8. 80000-
12
£
E 60000-1
01 40000
20000-
0
Spiked samples - 4 Albany
Lab B only
upper 95% prediction band
lower 95% prediction band
4567
percent 4 Albany spike
8 9
10 11
69
-------
The major effort here is to estimate the performance of the aggregate laboratory group
(A, B, C, D, and H) with respect to the group of samples from the greater New York City
area designated as "background" or "non-WTC impacted". The composite behavior of
these samples is illustrated below with respect to the analytical laboratories. The graph
indicates no consistent (high or low) percent bias from the cross laboratory means. This
confirms the conjecture made earlier that these laboratories are statistically similar. We
caution that Laboratory C did not provide any background data at all and could not be
directly included here, however, it is assumed that it would behave like the others.
Sample # vs percent difference by labs
(Slag Wool Fibers per gram dust)
NYC Background only
1 234567
8 9 10 11 12 13 14 15 16 17 18
sample!
The next step is to assess how an individual (presumably unknown) dust sample assay
relates to the amount of spiked 4 Albany dust percentage. Given the graph and underlying
statistics of the above figure entitled "Spiked Samples - 4 Albany, Labs A. B, C, D and H
Combined), one can calculate the x-value in % spiked 4 Albany equivalent and the 95%
confidence interval for the prediction for any unknown sample measurement from any
laboratory. This is essentially the use of the prediction band graph above in reverse. As
such the prediction of "x" and the CI take the following form:
Xpredicted = (Ybar - a)/(b)
Cl = Xprcd,cted ± [t(RSE)/b] * {1/m + 1/n + [(Ybar - ybar)2/(b2(n-l)sx2)}'2
70
-------
where Ybar is the mean laboratory measurement, a and b are the intercept and slope of the
regression, t is the critical t-value for n-2 degrees of freedom, RSE is the residual
standard error, m is the # of replicate measurements, n is the number of calibration points,
ybaris the mean of the regression y data, and sx is the standard deviation of the x values of
the regression data.
The reported slag wool results for the background samples can now be interpreted Table
5 shows the results for each background sample measurement across all participating
laboratories as a prediction of the percent equivalent 4 Albany spike level and half of the
95% confidence interval associated with the measurement There are a total of 63
measurement results in the table
Table 5' Results for each background sample across all participating laboratories as a
prediction of the percent equivalent 4 Albany spike level and ± 95% confidence interval.
I Lab* A B D H
I Sample Percent CI +. percent CI +- Percent CI +- Percent CI +-
: *
: 1 -1 35 5 43
3 -135 543
4
5 0 72 5 29
6
7 -061 5.38
8
9 0 20 5 33
10
25 7 95 5 21
26
27 0 99 5 28
28 .....
29 662 534
30
31 388 518
• 32
* Laboratory C was left out of this analysis because they reported no background data
We note that negative entries above are only statistical constructs Of the 63 background
measurements in this table, 7 (or about 11%) exceed the 4 Albany 5% spike level, 2 of
the 63 measurements exceed the 10% 4 Albany spike level If the upper confidence
limits are considered, 42 out of 63 (67%) exceed the 5% spike level and 7 of 63 (1 1%)
71
-089
-1 03
-092
-101
-0.43
-038
298
001
1 05
-088
575
387
0.93
073
-044
-0.88
2.29
4.79
540
541
5.40
541
537
536
520
534
528
540
5 17
518
528
5.29
5.37
540
522
5 17
-1.35
-0.47
0.14
-014
1 05
1.00
199
094
102
268
10.28
3.80
071
' 2.68
047
1 21
7.07
441
543
537
533
535
528
528
523
528
528
521
531
518
530
52l"
531
" 5"27
518
5 17
-069
-066
-071
-079
-068
306
068
098
0.84
-050
235
805
223
1.88
-0.49
386
1425
618
538
538
539
539
538
520
530
528
529
537
522
5.21
522
5.24
5 37
"518
5.63
517
-------
exceed the 10% spike level For instance for sample 25 at lab A the percent equivalent of
the fiber measurement is 7 95% and the upper confidence limit is 7 95% + 521% =
13 16%
As a further exercise, we calculated the same statistics for data within only one laboratory
(choosing Laboratory B as the example), these results do not include scatter in the
regression from the other qualified laboratories Here we find some improvement we see
3 of 18 values (16 7%) exceed the 5% 4 Albany dust level and 0 of 18 values exceed the
10% 4 Albany dust level For the upper confidence levels, 6 of 18 values exceed the 5%
4 Albany dust level and 1 of 18 values exceed the 10% 4 Albany dust level.
Conclusions
The conclusions are based solely on the analytical data provided from the laboratory test
and a few analyses of the 100% WTC spike samples From validity estimates based on
expected slopes and data scatter of WTC spiked samples, five of eight laboratories (A, B,
C, D, and H) were used for further analysis Intra-class correlation coefficients (with
natural log transformation) for individual labs and for the group of five demonstrate
similar and reasonable values (>0 7) when all available data are considered One-way
ANOVA analysis of Laboratories A, B, D, and H results provides no evidence to reject
the null hypothesis (that the results are from the same distribution) Laboratory C was
not included here because of insufficient reported data but, based on spike sample
statistics, it is likely that that they too would fall into this category
Under the practical constraints that the five laboratories are used at random with one
analysis per unknown sample, we cannot expect statistical discrimination at the 1% or 5%
4 Albany spike equivalent level because the upper 95% prediction bounds exceeds the
5% spike equivalent level across the board Reasonable discrimination is possible at the
10% 4 Albany spike equivalent level because the lower bound onlO% equivalent
measurements is approximately equal to the mean at 5% Albany spike equivalent level
72
-------
Al. Quality Assurance Project Plan for World Trade
Center (WTC) Screening Method Study
Project # WTC-1
Revision # 3
August 8, 2005
Prepared by:
U.S. Environmental Protection Agency
Office of Research and Development
Research Triangle Park, NC 27711
Approvals given below indicate that technical and administrative reviews have been conducted,
and reviewer comments for the document preceding the signature date have been resolved
Dr Jacky A Rosati, EPA Principal Investigator, ORD, NHSRC Date
David Friedman, EPA Principal Investigator & Project Officer, ORD, IOAA Date
Shirley Wasson, EPA QA Representative, ORD, NRMRL Date
-------
ACRONYMS
COPCs Contaminants of Potential Concern
DQI Data Quality Indicator
DQO Data Quality Objective
EPA U S Environmental Protection Agency
ERT Emergency Response Team (EPA)
HEPA High Efficiency Particulate Air
HVAC Heating, Ventilation and Air Conditioning
IOAA Immediate Office of the Assistant Administrator
MQI Measurement Quality Indicator
NERL National Exposure Research Laboratory (EPA)
MQO Measurement Quality Objective
NHSRC National Homeland Secunty Research Center (EPA)
NRMRL National Risk Management Research Laboratory (EPA)
ORD Office of Research and Development (EPA)
PLM Polarized Light Microscopy
QA Quality Assurance
QC Quality Control
QAPP Quality Assurance Project Plan
SEM Scanning Electron Microscopy
SOP Standard Operating Procedure
TEM Transmission Electron Microscopy
USGS US Geological Survey
WTC World Trade Center
-------
A2. TABLE OF CONTENTS
A PROJECT MANAGEMENT 5
A3 Distnbutton List 5
A4 Project Organization 5
A4 1 Responsibilities and Roles 5
A4 1 1 Analytical Laboratory Responsibilities 6
A4.1 2 Sampling Responsibilities 6
A4.2 Reporting Relationships 7
A5 PROBLEM DEFINITION AND BACKGROUND 9
A5.1 Introduction 9
A5.2 Method Development 10
A.5 3 Screening Method Study 11
A6. Project/Task Description 12
A 6 1 Sampling 12
A 6 2 Sample Preparation 12
A 6 3 Analytical Study 12
A 6 4 Tasks and Timeframes 13
A 7 Quality Objectives and Criteria 13
A7.1 Data Quality Objectives (DQOs) 13
A 7 2 Measurement Quality Objectives (MQOs) 13
A8 Special Training/Certifications 14
A9 Documents and Records 14
B DATA GENERATION AND ACQUISITION 15
Bl Sampling Process Design 15
B2 Sampling Methods 15
B3 Sample Handling and Custody 15
B3 1 Homogemzation -15
B4 Analytical Methods 16
B5 Quality Control 16
B6 Instrument/Equipment Testing, Inspection and Maintenance 18
B7 Instrument/Equipment Calibration and Frequency 18
-------
B8 Inspection/Acceptance of Supplies and Consumables 18
B9 Non-Direct Measurements 18
B10 Data Management 18
C ASSESSMENT AND OVERSIGHT 19
D DATA VALIDAT1ON AND USABILITY 19
APPENDICES 20
REFERENCES 20
FIGURES
Figure 1 Project Organizational Chart 7
TABLES
Table 1 Intralaboratory Measurement Quality Objectives (MQOs) 13
Table 2 Interlaboratory Measurement Quality Objectives (MQOs) 14
Table 3 Mass Concentrations of Spiked Samples 16
-------
A. PROJECT MANAGEMENT
Presented herein is the site Quality Assurance Project Plan (QAPP) for the World Trade Center
Screening Method Validation Study The QAPP has been developed in accordance with the
United States Environmental Protection Agency (EPA) guidance for Quality Assurance Project
Plans (EPA QA/G-5, EPA/240/R-02/009) December 2002 This plan is based on information
currently available and may be modified on site in light of field observations/results and other
acquired information Any modifications or deviations from this QAPP shall be approved by EPA
and documented
A3. Distribution List
The Distribution List documents who shall receive copies of the approved QAPP and any
subsequent revisions or amendments to the QAPP The US EPA shall distribute the QAPP to all
project team members and shall ensure that the project team members are familiar with any and all
QA issues A complete copy of the QAPP and any subsequent revisions shall be maintained on
file at the U S EPA RTF office and shall be available upon request. The following personnel will
receive copies of the approve QAPP for the WTC Screening Method Study (contact information
for these people can be found in A4)-
EPA
1. Jacky Rosati, EPA, ORD, NHSRC, Pnncipal Investigator
2 David Friedman, EPA, ORD, OAA, Principal Investigator
3 Rajeshmal Smghvi, EPA/ERT, WAM - Sampling Contract
4 Ten Conner, EPA, ORD, NERL ,WAM - SEM contract
5. Shirley Wasson, EPA, ORD, NRMRL, APPCD - Quality Assurance
Other Government Agencies
6 Greg Meeker, USGS - IAG Manager
Contractors
7 Cindy Klemman, Lockheed Martin -Sampling Contractor
8 Bob Willis, Alion Sciences - SEM Contractor
9 Steve Schwartz, Versar - Analysis Prime Contractor
10 Keith Rickabaugh, RJ Lee Group, Inc - Analysis Subcontractor
11 Rich Brown, MVA Scientific Consultants - Analysis Subcontractor
12. Garth Freeman, MAS, Inc - Analysis Subcontractor
13 John Newton, EMSL Analytical Inc - Analysis Subcontractor
14 Jeannie Orr, Reservoir Environmental, Inc - Analysis Subcontractor
A4. Project Organization
A4.1 Responsibilities and Roles
-------
Dr Jacky Rosati of the EPA, ORD, NHSRC and David Fnedman, ORD, OAA shall have the
oversight authority for all work conducted for this project and shall act as backup and work
assignment manager on the Versar analytical contract, respectively Dr Rosati has prepared the
Validation Study QAPP, and Shirley Wasson, EPA, ORD, NRMRL, APPCD will perform the QA
review of this QAPP Dr Rosati and Mr Fnedman shall provide technical assistance to ensure
that sample collection and analysis work is completed efficiently, and in compliance with the
applicable Scope of Work and all applicable rules.
A4.1.1 Analytical Responsibilities
Steve Schwartz, Versar, shall arrange and oversee the analysis work by analytical laboratories
Mr Schwartz and his subcontractors shall adapt, adopt and follow the Quality Assurance Project
Plan prepared by EPA for all sample analysis activities performed for this project All appropriate
data, original field forms/data sheets, shall be collected and completed in accordance with the
instructions contained in the contract and provided to EPA
The analytical laboratories retained by Versar include RJ Lee Group, MVA Scientific, MAS,
EMSL Analytical, and Reservoir Environmental Dr Rosati will arrange and oversee the analysis
work by the U S government analytical laboratories (Greg Meeker, USGS, Ten Conner, EPA,
NERL) All analytical and government laboratones shall conduct the required analysis within the
requested turnaround time, input the relevant analytical data into the appropnate spreadsheets, and
other similar duties as descnbed in its contract Scope of Work, IAG or as requested by Dr. Rosati
All data for this project are considered confidential and only the EPA is authorized to allow for
their release
The pnmary contractor for the environmental sampling shall follow the QAPP that they have
prepared entitled "Generic Quality Assurance Project Plan for WTC Residue Sampling New York
City, NY, March 2005 (Appendix A). All appropriate data, original field forms/data sheets, and
cham-of-custody forms shall be collected and completed in accordance with the instructions
contained in the contract and provided to EPA All samples shall be handled as instructed by
EPA, to include sample sieving, ashing, splitting, archiving and distnbuting
A4.1.2 Sampling Responsibilities
Raj Smghvi, EPA, ERT Work Assignment Manager shall oversee the sample collection performed
by Lockheed Martin (Cindy Kleinman), as directed by Dr Jacky Rosati. All sampling
appointments shall be arranged by Raj Smghvi, EPA/ERT for Lockheed Martin (Cindy
Kleinman) All samples are to be archived and stored in a safe location as described in Section
31
Lockheed Martin, the pnmary contractor for the environmental sampling shall follow the QAPP
that they have prepared entitled "Generic Quality Assurance Project Plan for WTC Residue
Sampling New York City, NY, March 2005 (Appendix A) All appropriate data, original field
forms/data sheets, and chain-of-custody forms shall be collected and completed in accordance
with the instructions contained in the contract and provided to EPA. All samples shall be handled
as instructed by Dr Rosati, to include sample sieving, ashing, splitting, archiving and distributing
All data shall be considered confidential Only the EPA is authonzed to release any data collected
for this project
-------
A4.2 Reporting Relationships
The Project Organizational Chart (Figure 1) shows the reporting relationships between all of
organizations involved in this project, including the lead organization (i.e., EPA) and all
contractors and subcontractors.
Dt -Us. (ft
J
l-s U,kl
Kasfe huul .s ni-h' i
»\ • \ s.i
U-eUwJ Mirtai i
:B5JF »-•> |
.
JIG&CT
s>
'5 i> i
VVoik .
MX" (.
5jJ!(T.U 'v't
-.Hi
Figure 1: Project Organizational Chart
Contact Information for the above listed personnel (alphabetical):
-------
Rich Brown
MVA Scientific Consultants
5500 Oakbrook Parkway - Suite 200
Norcross, GA 30093
770-662-8509
Ten Conner
U S EPA National Exposure Research Laboratory
109 TW Alexander Drive, D205-03
Research Triangle Park, NC 27711
919-541-3157
Garth Freeman
MAS, Inc
3945 Lakefield Court
Suwanee, GA 30024
678-687-5990
David Friedman
U S EPA Headquarters
Anel Rios Building
1200 Pennsylvania Avenue, N W
8101R
Washington, DC 20460
Cindy Kleinman
Lockheed Martin
USEPA Facilities
Raritan Depot
2890 Woodbndge Avenue
800MS800
Edison, NJ 08837-3679
732-321-4252
Greg Meeker
USGS
Bldg53Sl, Mail Stop
Denver Federal Center
Denver, CO 80225
303-303-236-3188
John Newton
EMSL Analytical Inc
108 Haddon Avenue
Westmont, NJ 08108
856-858-4800
Jeannie Orr
-------
Reservoir Environmental, Inc
2059 Bryan Street
Denver, CO 80211
303-964-1986
Keith Rickabaugh
R J Lee Group, Inc
350 Hochberg Road
Monroeville, PA 15146
724-325-1776
Jacky Rosati
U S EPA National Homeland Security Research Center
109 TW Alexander Drive, E305-03
Research Triangle Park, NC 27711
919-541-9429
Steve Schwartz
Versar, Inc
6850 Versar Center
Springfield, VA 22151
703-642-6787
Rajeshmal Singhvi
US EPA ERT, Region 2
Raritan Depot
2890 Woodbndge Avenue
101MS101
Edison, NJ 08837-3679
732-321-6761
Bob Willis
Ahon Scientific
109 TW Alexander Drive, E205-06
Research Triangle Park, NC 27711
919-541-2809
AS. PROBLEM DEFINITION AND BACKGROUND
A5.1 Introduction
The objective of this effort is to develop and evaluate a means of determining whether dust
sampled as part of EPA's future sampling program contains residual contamination attributable to
the collapse of the WTC towers The tested screening method is a critical component of the
sampling program as it will be used, along with the results from contaminants of potential concern
(COPC) testing, to determine the need for cleanup
The USGS has published two reports which provide the basis for the WTC dust signature adopted
in this sampling program The first report discusses the analysis and interpretation of indoor and
9
-------
outdoor WTC dust samples collected near Ground Zero, days and weeks after September 11, 2001
(Meeker, et al, 2005) From this work, we see that the WTC dust samples are dominated by
gypsum, concrete, and man-made vitreous fibers (MMVF), mainly slag wool It is on the basis of
these key results that gypsum, elements of concrete, and slag wool were identified as candidates
for a WTC signature The second report discusses the analysis of EPA supplied samples taken
from several indoor locations well outside of the WTC impacted area (background) These
samples were taken between September of 2004 and April of 2005 Slag wool is absent from
many of these background samples, but Lowers et al (2005a) state that the samples do have
gypsum present, which they speculate might be due to the presence of wall board in the sampled
apartments Because of the lack of slag wool in these samples, it was concluded that these
samples did not contain WTC dust It was also concluded that perhaps slag wool is the single
most critical of the three WTC dust constituents when distinguishing WTC dust from other
common dusts.
Other studies also identified MMVF and gypsum as predominant components of WTC dust In a
study of air and settled dust quality in apartments in Lower Manhattan, the Agency for Toxic
Substances and Disease Registry (ATSDR) and the New York City Department of Health and
Mental Hygiene (NYCDOMH) found significantly more MMVF and gypsum in Lower Manhattan
apartments as compared to comparison areas above 59th St (NYCDOMH/ATSDR, 2002).
Meanwhile, no MMVF was found in comparison locations. They also concluded that gypsum was
seen at a higher percentage in dust m Lower Manhattan samples as compared to the comparison
area samples In a comprehensive study of the composition of settled dust in the Deutsche Bank
building at 130 Liberty St, R J Lee identified numerous hazardous contaminants that were present
in the dust at levels much higher than in background office buildings, and among those substances
identified in their "WTC signature" were mineral wool and gypsum (R J Lee, 2004).
If the WTC building collapse signature components of slag wool, gypsum, and elements of
concrete are not present, then one could conclude that WTC building collapse dust is not present
However, since these components might be present in typical New York City dust, as slag wool is
a component of insulating materials in currently constructed buildings, it is possible that a test
might show them to be present even though WTC dust never impacted the sampled area. A
'screening test' will, by its design, result in some fraction of such false positives (a location
without residual WTC dust that tests positive for the above components) However, an
appropnate 'screening test' would result in very few, if any, false negatives (a location with
residual WTC dust that tests negative for the above components).
A5.2. Method Development
EPA acquired 117 dust samples dunng the time period of September 2004 to April 2005. Twenty-
one 'impacted' samples were taken by the EPA at two buildings that were part of the Deutsche
Bank complex located at 130 Liberty Street and 4 Albany Street Both buildings were uninhabited
and slated for demolition Fifty samples were taken from locations well beyond the impacted
zone, these samples are considered to be 'background' dust Forty-six samples were taken from
locations that were possibly impacted but were a bit farther from the WTC site than the known
'impacted' samples None of these forty-six samples are used in the study, but several were
evaluated during the development of the analytical method. In addition, one impacted sample was
obtained from the USGS. This sample was a composite sample of outdoor and indoor WTC dust
collected in September of 2001
10
-------
While a standard method using a HEPA vacuum collector was used by EPA to collect most bulk
dust samples (Appendix C), some bulk dust samples were collected from residential and
commercial vacuum cleaner bags. Whether collected by HEPA vac or by acquiring vacuum
cleaner bag, all samples are handled (sieved, split and stored) as described in the sampling QAPP
in Appendix C Many of the above samples were analyzed for slag wool content by the EPA's
National Exposure Research Laboratory (NERL) Scanning Electron Microscopy (SEM)
Laboratory This analysis was performed as part of the EPA's development of a standard protocol
for sample preparation and analysis (Appendix B), to determine the sample status (background or
impacted) and content.
There appears to be a clear distinction between samples taken in impacted areas versus
background samples All of the impacted samples had slag wool at concentrations of greater than
100,000 fibers per gram of dust, with a range of 113,000 to 13,400,000, while all of the
background samples had concentrations less than 100,000 ppm, ranging from no slag wool
identified inlO samples to 92,800 fibers of slag wool per gram of dust Based on this preliminary
work, the USGS, the EPA's Office of Research and Development, the EPA's National
Enforcement Investigations Center (NEIC), and a number of experts from the commercial testing
laboratory community, worked together to develop an analytical method to identify the presence
and concentration of the screening constituents (slag wool, gypsum and elements of concrete) in
indoor dust This method was reviewed by the WTC Expert Technical Panel's signature
subcommittee and is presented in Appendix B The composition of this technical panel can be
found at http //www.epa gov/wtc/panel
A.5.3 Screening Method Study
The hypothesis that is the foundation for the WTC dust screening method is as follows If a unit
has been impacted, those materials that are found in WTC dust (markers) will be found in the dust
collected from the unit. The materials under consideration are. 1) slag wool, 2) elements
consistent with concrete, and 3) gypsum Since slag wool is a major component of WTC collapse
dust, if a sample does not contain 'significant' levels of this marker, the unit would not be
considered to contain WTC residuals The other markers will be used to distinguish samples
containing non-WTC slag wool from those containing WTC slag wool. It is expected that data
from this study will define the term 'significant level'
Five independent laboratories and three government laboratories will participate in this final
method validation phase One government laboratory will analyze only a small portion of the
samples, but this lab was critical in the method development. Each laboratory attended a two day
session during which the method was further developed and discussed, and procedures to adapt the
method to suit each laboratory's equipment were determined Following this session, the
laboratories received dust samples consisting of both confirmed background samples (10 samples
plus duplicates) and a confirmed non-impacted dust spiked with varying amounts of confirmed
WTC dust (6 spiked samples plus duplicates) The spiked dust contains known quantities
(concentrations) of the screening materials The labs were provided the samples "blind", thus, they
did not know which samples were pure background dust, and which were the spiked dust
While the goal was to validate a method of differentiating between samples of dust that contain
residues from the WTC collapse from those that do not, since the three primary materials (slag
11
-------
wool, and elements of concrete and gypsum) identified above are all normally found in dusts
present in the New York area, it is possible that the proposed screen may yield some percentage of
false positive identifications of WTC dust
A6. Project/Task Description
A.6.1 Sampling
Dispersion models, photos, interviews, and satellite data were reviewed to discern areas that were
likely impacted by WTC emissions and those that were not. Samples to be used to study the
above discussed protocol were collected from within both of these areas (background and
impacted) Samples were analyzed by EPA's NERL and USGS for content verification, and
confirmation of background or impacted status by evaluating levels of slag wool in the dust
collected by SEM.
A.6.2 Sample Preparation
WTC dust was spiked into confirmed non-impacted dust at three levels (1, 5, and 10% of total
mass) and homogenized The dusts were all characterized by the USGS and U S EPA NERL
prior to spiking The two spiking dusts were 1) a composite sample from USGS of predominantly
outdoor dust collected in September of 2001, and 2) dust collected by the U S EPA from the
Deutsche Bank building at 4 Albany Street in September of 2004 The 4 Albany Street building
borders the south side of the WTC complex The USGS performed an analysis of the spiked
samples prior to the samples being sent to labs. The spiked samples showed varied levels of slag
wool, this was expected due to the difficulty in homogenizing dust containing large fibers, and the
fact that components of WTC dust will vary within a sample because of the nature of the source
Despite this variability, the measured levels were in the approximate range
expected for the spiking percent (1, 5, and 10%) and, in all but one case, each percent level was
fully distinguishable from the other in all but one case
Analysis by USGS, NEIC and NERL determined that the levels of slag wool differs between the
two WTC dusts, with the pure dust from 4 Albany Street more than an order of magnitude lower
in slag wool than that provided by USGS (approximately 500,000 fibers/gram of dust vs
approximately 11,000,000 fibers/gram of dust, respectively) There are likely two explanations for
this significant difference in slag wool levels. The USGS sample was a composite of multiple
outdoor samples and one indoor sample taken during September of 2001 The 4 Albany sample
was taken three years post 9/11 in September of 2004 This sample was taken exclusively inside
of a building, thus, the dust was not only diluted by three years of urban background dust, but was
also characteristic of dust that had penetrated the shell of an unopened building as opposed to that
dropping on the ground outside
A.6.3 Analytical Study
Five independent laboratories were recruited for a final test of the screening method Each
laboratory attended a two day session during which the method was further developed and
discussed, and procedures to adapt the method to suit each laboratory's equipment was
determined These laboratories have received 32 of the collected dust samples consisting of both
confirmed background samples and confirmed background samples spiked with varying amounts
of confirmed WTC dust (Sample Distribution Table shown in Section B 5) The spiked dust
contained known quantities (concentrations) of the screening materials and reasonable
homogeneity was confirmed by USGS The labs were provided the samples "blind", thus, they did
12
-------
not know which samples were pure background dust, and which were background dust samples
spiked with WTC dust The labs have several weeks to analyze all dust samples. They have been
asked to provide data as to the quantity of screening materials present in the dust in a standardized
format (Appendix B). The final data from all laboratories will be evaluated to determine if they
were able to distinguish background samples from WTC spiked samples In addition, criteria such
as time for analysis, and mtra- and interlaboratory variability will be considered when determining
validity of the method
A.6.4 Tasks and Timeframes
Sample Collection
Method/Protocol Development
Screening Study
Completion of Reports
Peer Review
September 2004-May 2005
February 2005-June 2005
June 2005-August 2005
August 2005
August 2005-September 2005
A7. Quality Objectives and Criteria
A7.1 Data Quality Objectives (DQOs)
The data quality objectives for this project are based on data acquired in the methods development
stage of this study. An inter-lab data quality objective was determined using the variability within
each dust sample for slag wool fibers, one of the markers for WTC dust It was determined that
data could be acquired with a relative certainty of + 35%. An mtra-lab data quality objective was
determined using the vanability within each dust sample for slag wool as well It was determined
that data could be acquired with a relative certainty of ± 30%
A7.2 Measurement Quality Objectives (MQOs)
The accompanying tables (Tables 1 and 2) list Measurement Quality Objectives (MQOs) for this
mtralaboratory (within lab) and interlaboratory (within sample) variability. Accuracies and
precision were taken from preliminary data and manufacturer's specifications.
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%
Table 1 Measurement Quality Objectives (MQOs) for mtralaboratory Vanability (within lab)
13
-------
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%
Table 2 Measurement Quality Objectives (MQOs) for Intel-laboratory Variability (within sample)
Intralaboratory MQOs will be calculated based on results within each lab for the 32 samples
Interlaboratory MQOs will be calculated based on the composite result for each lab and compared
with other labs for the 32 samples Both sets of MQOs will be compared with the target MQOs
listed above Accuracy will be based on how close the labs are to the calculated overall laboratory
mean for each sample or set of samples (i e each spiking percentage or background set of
samples) and precision will be based on the duplicate results within each lab (relative %
difference)
AS Special Training/Certifications
All laboratories and analysts chosen for this work will have training in both Polarized Light
Microscopy (PLM) and Scanning Electron Microscopy (SEM)
A9 Documents and Records
Documents generated (or to be generated) during this project and responsible party
1) Sampling Access Agreement - EPA
2) Sampling Information Sheet - EPA
3) Sampling and Sample Handling QAPP - Lockheed Martin
4) Analytical Method/Protocol to be used in study - EPA
5) Screening Study QAPP - EPA
6) Prime Contractor Report on Screening Study - Versar
7) EPA Report (separate from Prime Contractor Report) on Screening Study - EPA
The Screening Study QAPP will be distributed as indicated in Section A3 of this document. Dr
Rosati will distribute the QAPP to Versar and the government labs, USGS and EPA NERL as well
as EPA, ERT Versar will be responsible for distributing the QAPP to all analytical
subcontractors (RJ Lee, MVA, MAS, EMSL and Reservoir) and EPA, ERT will be responsible for
distributing the QAPP to its sampling contractor, Lockheed Martin.
Sampling and analytical data will be reported to Dr Rosati by EPA/ERT and Versar, respectively,
on a weekly basis These data will be presented in a spreadsheet format The final data shall be
presented to Dr Rosati in a report format, both electronically (including a final data spreadsheet)
and hard copy
14
-------
B. DATA GENERATION AND ACQUISITION
Bl. Sampling Process Design
Dispersion models, photos, interviews, and satellite data were reviewed to discern areas that were
likely impacted by WTC dust and those that were not Impacted samples were collected very
close to the WTC site, and background samples were collected from areas distinctly outside of
those that were 'likely' impacted Samples were collected from federal buildings, office buildings
and private residences on a volunteer basis A pre-sampling survey of building and sampling
areas, including photos of sampling areas (if permitted by building owners) and notes on building
usage, to identify conditions that might compromise samples (e g , smoking or cooking areas) was
developed Additionally, an access agreement was signed by the unit occupant/owner and an
information sheet regarding the sampling was provided by the sampling contractor to the
occupant/owner (Appendix D and E, respectively)
B2. Sampling Methods
Sampling followed the approved "Generic Quality Assurance Project Plan for WTC Residue
Sampling New York City, NY, March 2005 (Appendix A) In each building identified for
sampling, dust samples were collected from at least three areas 1) one sample from a track-in
area near a building entrance, preferably in a carpeted area, 2) two samples from relatively
undisturbed areas (e g, on top of bookcases, under furniture), and 3) other areas showing visible
accumulation of settled dust, including HVAC ducts A standard method (REAC SOP 2040 -
Collection of Indoor Dust Samples from Carpeted Surfaces for Chemical Analysis using a Nilfisk
GS-80 Vacuum Cleaner) using a HEP A vacuum was used by EPA to collect bulk dust samples
(Appendix C)
B3. Sample Handling and Custody
Once samples were collected, they were sieved to 150 microns Sieving was performed by the
method in REAC SOP 2040 - Collection of Indoor Dust Samples from Carpeted Surfaces for
Chemical Analysis using a Nilfisk GS-80 Vacuum Cleaner (Appendix C).
Once samples were sieved, they were ashed as described in 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, May 18, 2005 (Appendix B) 0 25 mg
of each ashed sample was archived Ashed samples were disseminated as instructed by EPA (Dr.
Rosati) As not all collected samples were used in the study, remaining samples will be sealed and
stored in a limited access area All samples are accompanied by cham-of-custody forms To
ensure that these important samples are properly collected, tracked, stored, and distributed, quality
assurance (QA) procedures were in place prior to any sample collection (Appendix C)
B3.1 Homogenization
Homogemzation of spiked samples was performed by the USGS Standards Laboratory after
ashing occurred First, the background material was transferred to a 16 ounce glass container and
an expanded metal mixing card inserted The container was sealed, and placed on a roller mixer
where the contents were mixed for a total of eight hours Each spiking material was then
15
-------
transferred to individual four ounce glass containers. A laminated customized paper mixing card
was inserted and the container was reseated and placed on a horizontal roller mixer Once pre-
mixing of the two dust types was complete, three concentrations of spiking material were prepared
according to Table 3 as follows
Mass Target
Cone Wt. %
1
5
10
background
material, g
29.7
285
270
spiking
material, g
0.3
1 5
30
Table 3: Mass concentrations of spiked samples
The appropnate amount of background and spiking material were weighed into plastic weighing
boats and then transferred to pre-labeled four ounce glass bottles. A laminated customized paper
mixing card was inserted and the container sealed and placed on a horizontal roller mixer The
samples were blended for a total often hours After blending a total of six ahquots, ~0 5g was
removed from each container using a spatula and transferred to individual one ounce vials The
six samples from each concentration underwent SEM analysis at the USGS to assure reasonable
sample homogeneity has been accomplished. The samples were then shipped to EPA, ERT for
archiving and distribution
B4. Analytical Methods
The USGS, the EPA's National Exposure Research Laboratory (NERL) and EPA's National
Enforcement Investigations Center (NEIC) developed a method to screen for the three key
materials slag wool, elements of concrete, and gypsum This method involves the use of
polarized light microscopy (PLM) or scanning electron microscopy (SEM) to determine the
quantity of each of the materials present (Appendix B)
Data will be reported on the standardized sheets in the appendix of this protocol Data to be
reported includes
• Slag wool (fibers/gram of dust) and length/width
• Elements of concrete (area %)
• Gypsum (area %)
B5. Quality Control
Quality control is addressed in Section 10 0 of the standard protocol in Appendix B entitled
"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"
In addition to the items referred to in this section, several measures have been taken to ensure the
quality of this study
16
-------
• Standard protocol - a standardized protocol will be used for all sample preparation and
analysis. This was determined to be a necessity during the scoping part of this work. This
standardized protocol should help to minimize interlaboratory variability
• Duplicates- duplicates of all samples were provided for analysis Analysis of these
duplicates will allow us to determine intralaboratory variability when using a standardized
protocol
• Blind samples - all labs received the same 32 samples, and all samples were coded so that
this was a 'blind' study, no lab knows what they are analyzing nor will they be able to
compare their results with other laboratories
Sample Distribution Table (note, each of the eight laboratories has been given a letter A-H)
Sample
Designations
AP5(1)
AP5(2)
CMC(1)
CMC(2)
HS3(1)
HS3(2)
WGSd)
WGS(2)
MW(1)
MW(2)
DB1%(1)
DB1%(2)
DB5%(1)
DB5%(2)
DB10%(1)
DB10%(2)
C1-RTP(1)
C1-RTP(2)
USGS1%(1)
USGS1%(2)
USGS5%(1)
USGS5%(2)
USGS10%(1)
USGS10%(2)
USC(1)
USC(2)
FP(1)
FP(2)
MUNYdd)
MUNYCK2)
Laboratory Letter Codes
A
B
C
D
E
F
G
H
17
-------
MUNYC2(1)
MUNYC2(2)
Samples spiked with WTC dust, at 1, 5, and 10% levels are shaded. All others are
background samples. Total of 20 background samples (10 samples + 10 duplicates) and
twelve spiked samples (6 samples + 6 duplicates)
B6. Instrument/Equipment Testing, Inspection and Maintenance
Vacuum cleaners used for sampling were maintained as described in Appendix A Generic
Quality Assurance Project Plan for WTC Residue Sampling New York City, NY, March 2005.
B7. Instrument/Equipment Calibration and Frequency
All microbalances used in this study shall calibrated annually Scanning Electron Microscopes
(EDS system) shall be calibrated on a daily basis as discussed in Appendix B entitled "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, Section 10 1
B8. Inspection/Acceptance of Supplies and Consumables
Inspection and acceptance of all consumables used during sampling will be performed as
described in Appendix A Generic Quality Assurance Project Plan for WTC Residue Sampling
New York City, NY, March 2005
Inspection and acceptance of all consumables used during sample analysis will be performed by
the analytical and government laboratones Consumables are listed under apparatus and materials
in Appendix B entitled "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.
B9. Non-Direct Measurements
Not applicable
BIO. Data Management
Data from this study will be reported by the subcontractors and government labs to the prime
contractor on standardized data sheets found in Appendix B "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, May 18, 2005." Data will be
compiled and analyzed by the prime contractor in a electronic spreadsheet, and a report of this
data will be written
18
-------
C. ASSESSMENT AND OVERSIGHT
• All work will be overseen by the principal investigators of this study. Weekly conference
calls will be held with EPA, the prime contractor and all subcontractors to assess progress
and discuss any issues or problems that rtiay have arisen Data is to be submitted to the
prime contractor by the subcontractors on a weekly basis as this data is obtained Due to
the rapid nature of this study, no interim reports are required.
• All stakeholders (EPA, USGS and contractors) will be provided a copy of the QAPP and
will review it for correctness
• All sampling and sample preparation performed by Lockheed Martin will be under direct
oversight of quality assurance personnel and audits will be conducted as noted in Appendix
A (Sampling QAPP)
• All contracting laboratories are required to employ standard QA practices and all work
should be performed under the oversight of in-house quality assurance personnel
• All data will undergo evaluation and review by the prime contractor prior to being
assembled into a report to the EPA The EPA will perform its own assessment and
evaluation of the study data and issues, and will assemble this information into an overall
EPA report
• The study will undergo a formal EPA peer review once it has been completed Peer
reviewers will be provided all data and reports, and will be given 6 weeks to perform a full
evaluation of the study In addition, all data and reports will also be provided to the WTC
Technical Panel for their review and comments
D. DATA VALIDATION AND USABILITY
• All data shall be provided by the subcontractors to the prime contractor on the
spreadsheets provided in Appendix B "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, May 18, 2005.",
Section 15 0
• The prime contractor will review and verify the data before compiling it into a report
• EPA will evaluate the prime contractor report, along with the compiled data to
determine'
o whether the MQO's presented in Table I of this QAPP were met
o whether the study described herein demonstrated the following
• that slag wool, gypsum and elements of concrete are reasonable markers
for WTC dust (by showing that these markers distinguish WTC-laden
dust from background dust),
• that WTC dust at a diluted concentration can be distinguished from
background, and
• that the analytical method works well enough, and is able to be earned
out by enough analytical laboratories to 1) evaluate the above materials
as markers and 2) distinguish WTC dust from background dust
19
-------
• EPA will prepare a final report documenting all data, analysis and conclusions based
on the above evaluation
APPENDICES
A Genenc Quality Assurance Project Plan for WTC Residue Sampling New York City, NY,
March 2005
B 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, May 18, 2005
C REAC SOP 2040 - Collection of Indoor Dust Samples from Carpeted Surfaces for
Chemical Analysis using aNilfisk GS-80 Vacuum Cleaner
D Access Agreement
E Information Sheet
REFERENCES
Lowers, H A, G.P Meeker, and IK Brownfield (2005a) 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
Lowers, H A, Meeker, G P , I K Brownfield (2005b) World Trade Center Dust Particle Atlas.
U S Geological Survey Open-File Report 2005-1165. http //Eubs.usgs^Qv/of/2Q05/].165/
Meeker, G P , A.M Bern, H A Lowers, and I K. Brownfield (2005) Determination of a
Diagnostic Signature for World Trade Center Dust using Scanning Electron Microscopy Point
Counting Techniques U S Geological Survey Open File Report 2005-1031.
NYCDOHMH/ATSDR (2002) New York Department of Health and Mental Hygiene and
Agency for Toxic Substances and Disease Registry. Final Technical Report of the Public Health
Investigation To Assess Potential Exposures to Airborne and Settled Surface Dust in Residential
Areas of Lower Manhattan Agency for Toxic Substances and Disease Registry, US Department
of Health and Human Services, Atlanta, GA.
R J Lee (2004) Signature Assessment 130 Liberty Street Property Expert Report WTC Dust
Signature Prepared for Deutsche Bank May, 2004 R J Lee Group, Inc 350 Hochberg Road,
Monroeville, PA 15146
20
-------
APPENDIX A: SAMPLING QAPP
Al. GENERIC QUALITY ASSURANCE PROJECT PLAN FOR
WORLD TRADE CENTER (WTC) RESIDUE SAMPLING NEW
YORK CITY. NEW YORK
U.S. EPA Work Assignment No 0-089
Lockheed Martin Work Order No • EACOO089
U S EPA Contract No • EP-C-04-032
Prepared For
United States Environmental Protection Agency/Environmental Response Team
Edison, NJ
March 2005
21
-------
A2. TABLE OF CONTENTS
A PROJECT MANAGEMENT
Al TITLE PAGE . . . 1
A2 TABLE OF CONTENTS . . ... 2
A3 DISTRIBUTION LIST . . 3
A4 PROJECT ORGANIZATION .... .3
A5 PROBLEM DEFINITION . 3
A6 PROJECT DESCRIPTION AND SCHEDULE ... . .4
A7 DATA QUALITY OBJECTIVES AND CRITERIA FOR MEASUREMENT
OF DATA 4
A8 TRAINING AND CERTIFICATION .. 4
A9 DOCUMENTS AND RECORDS ... 5
B DATA GENERATION AND ACQUISITION 5
Bl SAMPLING PLAN DESIGN . . . .... 5
B2 SAMPLING METHODS . . 6
B3 SAMPLE HANDLING AND CUSTODY . 6
B4 ANALYTICAL METHODS .... ..6
B5 QUALITY CONTROL ... .7
B6 INSTRUMENT/EQUIPMENT TESTING, INSPECTION AND
MAINTENANCE ... 7
B7 INSTRUMENT/EQUIPMENT CALIBRATION AND FREQUENCY . ... 7
BS INSPECTION/ACCEPTANCE OF SUPPLIES AND CONSUMABLES . 7
B9, NON-DIRECT MEASUREMENTS 7
BIO DATA MANAGEMENT .. .. .7
C ASSESSMENT/OVERSIGHT .. . 8
Cl ASSESSMENT AND RESPONSE ACTIONS 8
C2 REPORTS TO MANAGEMENT. . . 8
D DATA VALIDATION AND US ABILITY . . . . .9
DI DATA REVIEW, VERIFICATION AND VALIDATION . . 9
D2 VERIFICATION AND VALIDATION METHODS . . 9
D3 RECONCILIATION WITH USER REQUIREMENTS .. . 9
REFERENCES 9
TABLE 1 - FIELD SAMPLING SUMMARY ... . . . . 10
0089-DQAPP-031805
22
-------
A PROJECT MANAGEMENT
This project generic Quality Assurance Project Plan (QAPP) was prepared in accordance with
EPA Requirements for Quality Assurance Project Plans (QAPPs), EPA QA/R-5 and the Response
Engineering and Analytical Contract (REAC) Program QAPP,
A3 DISTRIBUTION LIST
The following personnel will receive copies of the approved QAPP for the World Trade Center
(WTC) Residue Sampling, Work Assignment (WA) No 0-089
1. Rajeshmal Smghvi, Environmental Protection Agency/Environmental Response
Team (EPA/ERT)
Work Assignment Manager (WAM)
2. Jacky Rosati, EPA, Research Triangle Park North Carolina (NC)
3 Eletha Brady-Roberts, National Homeland Security Research Center (NHSRC)
4 Jeffrey Bradstreet, REAC Air Response Section Leader/Task Leader
(TL)/Quality Control (QC)
Coordinator
5 Deborah Killeen, REAC Quality Assurance Officer (QAO)
6 Dennis Miller, REAC Program Manager
A4 PROJECT ORGANIZATION
The following individuals will participate in the project
EPA/ERT
Rajeshmal Singhvi -WAM
Jacky Rosati - Project Coordinator
Jeff Catanzanta - Technical Auditor
Eletha Brady-Roberts - NHRSC Quality Assurance (QA)
REAC
Jeffrey Bradstreet -TL/QC Coordinator
Miguel Trespalacios - Senior Air Sampling Scientist
Michael Hoppe - Environmental Scientist/Sampler
TBD - Environmental Scientists/Samplers
Richard Magan - Technician/Sampler
Deborah Killeen - QAO
Laboratories that will receive residue samples for chemical marker/signature identification on
this project include1
Department of Environmental Sciences and Engineering, University of
North Carolina, United States Geological Survey (USGS) Denver
Microbeam Laboratory, and Other commercial laboratones to be
determined
The REAC TL/QC Coordinator for the project is the pnmary point of contact with the EPA/ERT
WAM. The XL is responsible for the completion of the Work Plan (WP) and QAPP, project
team organization, and supervision of all project tasks, including reporting and deliverables The
EPA NHSRC will provide oversight and guidance in the field through the WAM
A5, PROBLEM DEFINITION
23
-------
The World Trade Center (WTC) attacks on 11 September 2001 caused the airborne release of two
types of dusts those related to the building collapse and fine participate matter from the subsequent
fires There are concerns among residents of New York City (NYC)about the potential health
effects of WTC dusts that might remain in buildings in NYC. The goal of this study is to collect
dust samples from areas near the WTC and distant from the WTC (background NYC dusts)
These dust samples will be used to validate chemical markers or signatures for WTC dust - as
compared to background dust - for a larger sampling effort to identify indoor areas still
contaminated with WTC dust. The markers or signatures for WTC dust are being developed by
laboratories at EPA, USGS, and several universities By sampling a number of contaminated and
uncontammated sites and by utilizing recently collected samples, the WTC signatures can be
validated or improved for the larger sampling study to delineate contaminated areas
The residue from the collapse of the WTC Towers may contain heavy metals (primarily lead,
arsenic, and mercury), polynuclear aromatic hydrocarbons (PAHs), and other contaminants that
EPA designated laboratories are investigating to associate specific analytes with the WTC.
A6. PROJECT DESCRIPTION AND SCHEDULE
The purpose of this project is to collect dust samples from designated buildings that may be from
the collapse of the WTC towers or from background sites for comparison EPA will identify
contaminated and uncontammated buildings in NYC and obtain access for REAC personnel to
conduct sampling To the extent possible, contaminated buildings that are within both the dust
and the fire plumes will be selected, so that current dust samples for these two types of emissions
can be collected EPA may identify several groups of buildings over time, as permission for access is
obtained EPA will provide REAC personnel with the street address and the name and phone
number for a contact person in each building identified for surveying
This activity will involve conducting scoping surveys of buildings identified by the EPA in the
NYC Area, preparing sampling plans, collecting samples, splitting samples for multiple
laboratories, shipping samples, and archiving samples for up to two years for future analysis and
report preparation
The schedule of activities and reports is as follows
WP 6 October 2004
Draft Generic QAPP 6 October 2004
Final Generic QAPP 18 March 2005
Draft Building Survey Form 15 October 2004
• Collect Residue Samples As scheduled by EPA
• Prepare and Send Sample Aliquots 4 Days after sampling
• final Report 10 Days after sampling
A7 DATA QUAIJTY OBJECTIVES AND CRITERIA FOR MEASUREMENT OF
DATA
The focus of this project is to collect dust samples that may be contaminated with materials from
the destruction of the WTC towers or are potentially uncontaminated This QAPP covers the
collection, storage and shipment of the samples to EPA designated laboratories The specific
chemical markers or signatures associated with WTC dust samples are investigatory and will be
used to further define the project Once defined, the specific chemical markers or signatures will
be used to further define the project
A8 TRAINING AND CERTIFICATION
24
-------
The training of all field personnel involved with sampling activities is intrinsic to their position
and required responsibilities They will have the following documented training
n Occupational Safety and Health Administration (OSHA) 40-hour and 8-hour
refresher in Hazardous
Waste Operations (20 CFR1910 120)
25
-------
• Department of Transportation (DOT) hazardous materials shipping
First Aid and Cardiopulmonary Resuscitation (CPR) training
A9 DOCUMENTS AND RECORDS
The RE AC Program QAPP serves as the basis for this generic QAPP The most current approved
version is available to all REAC technical personnel as an uncontrolled copy on the REAC Local
Area Network (LAN) Documents and records that will be generated during this project include.
WP
Draft Generic QAPP
Final Generic QAPP
Field logbooks
Site maps
Photos of Sampling Locations
Chain of Custody forms
Final Reports
The Final Report will provide a description of the project, field procedures, sample preparation
procedures, difficulties encountered and will include validated final copies of chain of custody
forms. All documentation will be recorded in accordance with REAC standard operating
procedure (SOP) #2002, Sample Documentation and REAC SOP #4001, Logbook
Documentation The final report will be prepared using REAC SOP #4021, Preparation of Final
Reports
B, DATA GENERATION AND ACQUISITION
Bl SAMPLING PLAN DESIGN
Judgmental sampling will be used to select sample locations that are most likely to represent WTC
residue or background dust This will be based on historical information, visual inspection and best
professional judgment of the WAM and sampling team. This type of sampling is used to identify
contaminants present in areas potentially having the highest concentration of contaminants.
Additional samples may be collected when requested by the WAM and EPA NHSRC personnel
During the sampling of EPA specified buildings, dust samples will be collected from each of the
buildings up to 20 in accordance with the EPA approved generic QAPP. Sampling will likely be
performed at two types of areas in each building a high traffic area (to characterize tracked-in dust)
and a lowtraffic area (to represent settled indoor dust) Two low traffic areas will be specified for a
total of three areas that will be sampled. The desired high traffic area is to be an area near a main
entrance, preferably carpeted The desired low-traffic areas include areas infrequently cleaned,
such as the top of elevator housing, under refrigerators, behind file cabinets, above ceiling tiles, on
high shelves, or m other areas that show visible dust accumulation and are infrequently disturbed
Sampling will not be restricted to carpeted areas as the intent of the sampling is to obtain the
desired residue. If vacuum sampling is not possible or preferred, sweep sampling will be used to
collect the residue Sampling will not be conducted m areas that would likely contain chemicals in
dusts that would interfere with the analysis for the WTC markers The following areas will be
avoided m the sampling effort
• Areas with significant cigarette or cigar smoke, incense, or
burning candles
Areas near major outdoor combustion sources (e g, power plants)
Due to the inability to obtain triplicate samples (once an area is sampled, little residual remains),
three samples will be collected in the same general area, for a total of nine samples from a building
(i e, three sample areas times three samples in each general area). The proximity of the samples in
26
-------
each general area will be determined as a result of visual inspection in the field and discussions
with the WAM
27
-------
B2 SAMPLING METHODS
Vacuum sampling will be performed in accordance with modified REAC SOP #2040, Collection of
Indoor Dust Samples From Carpeted Surfaces for Chemical Analysis Using a Nilfisk GS-80
Vacuum Cleaner This method may afford collection of samples large enough for analysis of both
purported organic and inorganic signatures Although the method specifies the size and shape of the
areas to be sampled and the mass to be collected, the sample collection procedure will vary to
accommodate the site-specific conditions and ensure that an adequate sample is obtained If it is
not feasible to use the vacuum method of sampling, samples will be collected in bulk by sweeping
the residue into a pan or sample bag in accordance with modified ERT/REAC SOP #2011 Chip,
Wipe and Sweep Sampling The sample handling and data collection requirements specified in
modified REAC SOP #2040 will be followed
The area to be sampled is not measured before sampling, but after the sample is collected This is a
modification of both REAC SOP #2040 and ERT/REAC #2011. REAC SOP #2040 is further
modified in that samples will also be collected from non-carpeted surfaces, the amount of sample
collected will be visibly checked and dust weight calculations will not be performed Sweep
sampling utilizes a dedicated, hand held sweeper brush to acquire the sample from an area The
area sampled is measured after sampling
Sample Volume, Container, Preservation and Holding Time. The collected samples are
placed into appropriately sized glass jars or zip-lock plastic bags Storage of the samples
collected by sweep or vacuum are maintained in a refrigerated unit at 4 ± 2 degrees Celsius (°C)
after sieving
Sampling Equipment Decontamination. The nozzles, wands and hoses are decontaminated
after use with a bottle brush, to remove any accumulated dust in the hose and nozzle When the
nozzle is clean, it is removed and sprayed with reagent grade methanol and allowed to air dry on a
clean surface The wand and hose are then cleaned with the bottle brush To continue a new
polylmer and collection bag for the collection of another sample is installed
B3 SAMPLE HANDLING AND CUSTODY
In the field, sampling data are recorded on a Vacuum Sampling Work Sheet or in a dedicated
project logbook Cham of custody (COC) records will be used to document the collection of dust
samples by vacuum or bulk. All COC records will receive a peer review in the field prior to
shipment of the samples in accordance with REAC SOP #4005, Chain of Custody Procedures
All samples will be delivered to the REAC facility and sieved in accordance with modified REAC
SOP 2040, Collection of Indoor Dust Samples From Carpeted Surfaces for Chemical Analysis
Using a Nilfisk GS-80 Vacuum Cleaner and modified ERT/REAC SOP #2011 Chip, Wipe and
Sweep Sampling The samples will be sieved through a No 100 sieve (150 microns [firnl) After
sieving, the samples will either be transferred to jars, which will be placed into Ziplock* storage
bags, or directly into Ziplock™ storage bags, and then placed into a holding refrigerator with the
corresponding COC record.
Scnbe* spreadsheet formats will be used for sample management REAC is required by contract to
use Scnbe" to track and log the samples In addition a unique sample numbering system has been
established to each sample, which identifies the site identification, event number and the sample
number Additional information is provided with the sample number to identify whether it is a
sieved (S) or coarse (C) fraction and the weight of the fraction in grams, e g, SO. 1
The samples collected by REAC personnel will be shipped to the designated laboratory for
analysis in accordance with REAC SOP #2004, Sample Packaging and Shipment One of the four
aliquots of each sample will be retained and stored by REAC staff for up to two years in a secure
refrigerator,
28
-------
B4 ANALYTICAL METHODS
Once specific chemical markers or signatures have been defined, EPA personnel in consultation
with NHSRC will be able to determine which analyses will be appropriate The laboratories
specified under Section A4 are conducting the investigatory work
B5 QUALITY CONTROL
This QAPP covers the collection, storage and shipment of the samples to EPA designated
laboratories for analysis Quality control for the field and storage procedures are as follows-
• Field documentation on Field Sampling Worksheets or in logbooks
Documentation of temperature for the dedicated secure refrigerator
Duplicate samples will not be taken due to the nature of the sampling method Once an area is
vacuumed, little residual sample remains Quality control for the laboratory procedures will be
specified by the EPA/NHSRC
B6. INSTRUMENT/EQUIPMENT TESTING, INSPECTION AND MAINTENANCE
The Nilfisk vacuums used in the collection of the residue samples will be maintained in
accordance with established specifications On a quarterly basis, the parts of the Nilfisk vacuum
cleaners are inspected for cracks and breaks An inventory of available supplies is conducted
every three months.
B7. INSTRUMENT/EQUIPMENT CALIBRATION AND
FREQUENCY The instrument/equipment calibration frequency is
not applicable to this QAPP
B8 INSPECTION/ACCEPTANCE OF SUPPLIES AND
CONSUMABLES
REAC personnel are responsible for the procurement, inspection, and acceptance of supplies and
consumables for this WA. The vacuum cleaner filters purchased by REAC personnel must meet
the requirements specified by the manufacturer The REAC TL and Group Leaders are
responsible for ensuring that the correct filters and sampling bags are specified in the purchase
orders and verifying upon receipt that the correct parts have been shipped It is the responsibility
of the EP AVERT to provide adequate facilities, equipment and supplies for REAC to perform all
field related tasks for this WA
B9 NON-DIRECT
MEASUREMENTS This section is not
applicable to this QAPP BIO
DATA MANAGEMENT
The QAPP is identified by the footer located on the bottom left hand comer of the page The file
identification represents the structure and the filename The filename starts with the 3-digit WA
number preceded by a "zero", then the deliverable type (D or N) to identify the document as a
deliverable or non-deliverable followed by the document type For amended or revised documents,
the letters "A" and "R" for amended and revised, respectively, and the appropriate amendment or
29
-------
revision number (eg 1.2,3 ) are added after the document type After the document type and
revision/amendment code (if any), a six-digit code based on the month, day and year (mmddyy) is
added to indicate the date the document was delivered to the client.
Field sampling data will initially be recorded on field data sheets and in field notebooks Samples
will be identified by the field assigned sample number Paper versions of all dehverables (Work
Plan, Generic QAPP and Final Reports) will be provided to the ERT WAM and stored in the
REAC Central Files Electronic versions of ail dehverables will be saved on the REAC archive
drive in accordance with Administrative Procedures (AP) #34, Archiving Electronic Files All data
dehverables for this WA will be posted to the ERT-Information Management System (IMS) web
site as either a Scribe* electronic data deliverable (EDO) or in portable document format (pfd)
Submission of the deliverable to the appropriate ERT-IMS website will be considered delivery to
the WAM as of the date and time such dehverables are received on the website
Field log books will also be archived once the project is completed and the Work Assignment 0-089
is closed All SOPs referenced in this QAPP are available on the REAC LAN.
C. ASSESSMENT/OVERSIGHT
Cl. ASSESSMENT AND RESPONSE ACTIONS
The REAC TL, Air Response Section Leader, QAO and QC Coordinator are responsible for QC
assessments and corrective action for this WA These personnel have the authority to issue stop
work orders The tasks associated with this QAPP are assessed through the use of peer reviews,
technical reviews and/or technical system audits, and management system reviews Peer review
enables the reviewers to identify and correct reporting errors before reports are submitted
Technical reviews are conducted by those immediately responsible for overseeing or performing the
work (self-assessments) An independent assessment or technical audit will be performed by Jeff
Catanzanta Management system reviews establish compliance with prevailing management
structure, policies and procedures, and ensures that the required data are obtained
Peer reviews are conducted on project dehverables to ensure a technical review with respect to
content, completeness and the overall quality of the deliverable prior to submittal to the EPA/ERT
The responsibilities of the review team and the sequence in which the deliverable is reviewed, is
outlined in REAC AP #22, Peer Review of REAC Deliverable* The REAC QAO will audit data
dehverables on a biannual basis to determine compliance with the peer review procedures
The EPA/ERT WAM for this task will be present and will have the responsibility for verifying that
the proper SOPs and sampling procedures are followed If any technical issues or deficiencies are
identified, they will be reported to the REAC TL for immediate resolution or corrective action
Any changes in scope of work will be documented on a Field Change Form and approved by the
WAM
C2 REPORTS TO MANAGEMENT
Monthly technical reports will be prepared for this WA when hours have been charged on a monthly
basis. These reports will detail the accomplishments for the past month, any problems encountered,
solutions to rectify the problem, contacts and meetings, goals for the next month, and an estimate
of the of the total labor hours and costs for the next reporting period The monthly technical
reports are submitted to the EPA/ERT Project Officer and WAM
On a quarterly basis, the REAC QAO provides a report to the REAC Program manager and the
ERT QA Manager that summarizes the quality assurance (QA) activities on a quarterly penod
These reports include results of performance evaluation samples, system audits (internal and
external), summary of non-cpnformance and corrective actions, preparation of SOPs for
analytical and operational activities, training, contacts/meetings and other QA activities
30
-------
REAC Report
Monthly Progress
Quarterly Q A Reports
Recipients
EPA/ERT Project Officer and WAM
EPA/ERT Project Officer and WAM
31
-------
D DATA VALIDATION AND USABILITY
Dl DATA REVIEW, VERIFICATION AND VALIDATION
For field activities, it is necessary to determine whether the samples were collected using the
sampling design specified in element Bl, whether the samples were collected according to a
specific method or SOP as specified in element B2, and whether the collected samples have been
recorded and handled properly as in element B3 Field sampling worksheets and field notes will
be reviewed by the RE AC TL for completeness The COC records will be reviewed to ensure that
the field information has been accurately reflected on the COC records
D2 VERIFICATION AND VALIDATION METHODS
Verification occurs at eacii level in the field to ensure that appropnate outputs are being generated
routinely Records produced electronically or maintained as hard copies are subject to data
verification During field activities, records associated with sample collection such as field data
sheets, COC records, logbook documentation, or electronic devices to log samples are verified
Naming conventions for the initial samples and samples fractions produced during sieving are
verified by the RE AC TL Chain of custody records are verified along with refrigerator and
freezer logs to ensure the integrity of the samples
There is no analytical data being generated under this WA, therefore, procedures for verifying
and validating data, including the chain of custody for data throughout the life cycle is not
applicable
D3 RECONCILIATION WITH USER REQUIREMENTS
Responsibility lies with the EPA, thus, this element is not applicable to this QAPP
REFERENCES
Response Engineering and Analytical Contract 2003 Quality Assurance Project Plan for the
Response, Engineering, and Analytical Contract, Revision 0 0
U S Environmental Protection Agency 1990, Quality Assurance/Quality Control Guidance for
Removal Activities, EPA/540/G-9/004, Office of Emergency and Remedial Response
U S. Environmental Protection Agency 2001 EPA Requirements for Quality Assurance Project
Plans (QAPPs), EPA/240/B -01/003, Office of Environmental Information.
32
-------
TABLE I Field Sampling Summary
World Trade Center (WTC) Residue Sampling
March 2005
Analybca
1
Parameter
Dust/Settled Partculate
Dust/Settled Paniculate
Sampling
Method
Nilfisk GS-80
Vacuum
Cleaner
Sweep
Preservation
Up to 2 years at 4
degrees C +1-2 degrees
C
Up to 2 years at 4
degrees C +/-7 degrees C
Total Samples
Up to 9 per Building
Up to 9 per Building
Maximum
Number
Samples
9
9
33
-------
APPENDIX B: PROTOCOL
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
June 27, 2005
Prepared by:
U.S. Environmental Protection Agency
National Enforcement Investigations Center/ National Exposure Research
Laboratory/National Homeland Security Research Center
Denver, CO and Research Triangle Park, NC
The use of trade names does not imply endorsement and are used for illustrative purposes only.
-------
Contents
1 0 Purpose 36
2.0 Scope/Application ... .... .... 36
21 Limitations of the Method and Future Considerations 36
3 0 Definitions ... . . • • 36
40 Summary of Method .. -37
5 0 Interferences 37
6.0 Safety 37
70 Apparatus and Materials 38
8.0 Reagents 39
9.0 Sample Storage 39
100 Quality Control 39
101 Calibration 39
110 Procedure ... . . 40
111 Weighing and Splitting . . . ... ... 40
11 2 Ashing . . . .. 40
113 Sieving . .. - ... 41
11 4 Preparation of Sample for Polarized Light Microscopy... .. 41
114 Preparation of Sample for SEM Analysis .. .. 41
120 Analysis ... - - - • 9
121 Analysis by Polarized Light Microscopy 9
122 Analysis by SEM/EDS -10
12.21 Screening for Slag Wool 10
12.22 EDX Screening for Gypsum/Anhydnte . 11
1223 X-Ray Mapping for Gypsum 11
1224 X-Ray Mapping for Ca-nch Particles 12
1225 Particle Analysis for Gypsum and Concrete . .13
130 Data Analysis and Calculations .... .. . .14
140 References 15
150 Appendix 16
35
-------
1.0 Purpose
This document describes sample preparation and analytical screening procedures for bulk
samples of dust collected from residential and commercial office environments. These
methods are collectively referred to as the protocol
2.0 Scope/Application
The protocol describes polarized light microscopy (PLM) and scanning electron
microscopy (SEM) with energy dispersive spectrometry (EDS) to screen bulk dust
samples for mineral slag wool, particles consistent with concrete compositions, and
gypsum. The analysis methods include operating parameters and particle identification
cntena
2.1 Limitations of the Method and Future Considerations
This protocol provides a means of analyzing for particles consistent with those
found in dust present after the collapse of the World Trade Center (WTC) in New
York City Components of WTC Dust have been documented and catalogued by
the U S Geological Survey Denver Microbeam Facility and the images and
characteristics shall be used in identification of particles (1)
The x-ray mapping procedure in sections 1223 and 1224 and the calculations
presented in section 130 only determine the maximum percentage of non-
gypsum, calcium-nch particles, which may include non-concrete materials The
particle analysis procedure presented in section 12 2 5 is the preferred procedure
for determining the percentages of gypsum and concrete particles in the sample
The x-ray mapping and image analysis procedure relies heavily on the thresholds
for backscattered electron images Binary (particles white and background black)
backscattered electron images (BE1) should be used to reduce errors in setting
thresholds in Photoshop
3.0 Definitions
1 PLM - Polarized Light Microscopy
2 SEM - Scanning Electron Microscope
3 EDS - Energy Dispersive Spectrometry
4 SEI - Secondary Electron Image
5 BEI - Backscattered Electron Image
6 Mineral Wool - lightweight vitreous fibrous material composed of rock wool and slag
wool and used especially for heat and sound insulation
7 Rock Wool - a man-made vitreous fiber (MMVF) component of mineral wool
containing magnesium, aluminum, silicon, and calcium Sodium and potassium may
also be present Iron oxide is typically 3-12% by weight
8 Slag Wool - a man-made vitreous fiber (MMVF) component of mineral wool
36
-------
containing magnesium, aluminum, silicon, and calcium. Sodium and potassium may
also be present. Iron oxide is typically less than 2% by weight
9. HEPA - High-Efficiency-Particulate-Air Filter
4.0 Summary of Method
1 Weigh sample to nearest 0.0005 g
2 Split the sample, archive half and keep half for analysis
3 Ash half of the sample for analysis
4. Sieve the ashed sample to 1 SO um.
5 Split the <150 um ashed portion Archive three quarters of the sample Keep one
quarter for PLM and SEM/EDS analysis
6 Weigh the quarter and place it in enough isopropanol to get a 10-20 mg per mL
dilution. Apply an aliquot to a glass slide, let dry, and add 1 55 (or 1 605) refractive
index oil Analyze by PLM for mineral wool.
7 Prepare a sample for SEM/EDS analysis using the same dilution prepared for PLM
8 Apply an aliquot of the sample to an aluminum sample stub with a carbon adhesive
tab covered by a piece of polycarbonate filter (13-mm diameter or punched out of a
larger filter to fit the size of the stub)
9 Identify fibers by EDS and record the occurrence of fibers > 25 um in length at 100 x
magnification to get a statistical representation of fiber compositions
10 Prepare 10-fold dilution of the suspension from step 7 and apply an aliquot to a
polycarbonate/adhesive tab substrate affixed to an aluminum sample stub
Alternatively, a lighter loading can be prepared by filtering the diluted suspension
through a 25-mm diameter, 0 4-um pore size, polycarbonate filter and affix this to a
carbon adhesive tab affixed to an aluminum sample stub
11. Collect x-ray maps of 10 fields at 500 x magnification for major elements, especially
Ca, S, and Fe and use Adobe Photoshop or similar software to determine the area
percent of gypsum and Ca-nch particles Fe-nch particles may also be identified in
this step
12 Perform particle analysis via computer-controlled SEM/EDX analysis
5.0 Interferences
Interferences include possible contamination of samples by airborne dust or through
improperly cleaned glassware and sieves. Interferences are minimized by performing all
procedures involving dry dust in a clean room, cleaning countertops and glassware
thoroughly before proceeding and placing particle-free wipes on all working surfaces To
avoid cross-contamination, properly clean all glassware, sieves, and tools between
samples
6.0 Safety
Respirable particles which may present a health hazard may exist in the sample Bulk
samples may release respirable particles during handling All procedures involving dry
dust samples will be performed under a negative flow High-Efficiency-Particulate-Air
37
-------
Filter (HEP A) hood Samples handled outside of the HEP A hood will be covered with
aluminum foil or placed in sealed glass jars
7.0 Apparatus and Materials
1 HEPA negative flow hood
2 Forceps
3 Kimwipes
4 Stainless steel spatula
5 Weighing paper
6 Programmable furnace [not required for validation study]
7 Ceramic crucibles with lids [not required for validation study]
8. Analytical balance (accuracy to 0 0005 g)
9 Retsch ultrasonic sieve shaker (AS200 Basic), or similar [not required for validation
study]
10 Sample sieves, 3-inch diameter (recommended), 150-um (100-mesh) opening, with
lid and bottom pan similar [not required for validation study]
11 SEM aluminum sample stubs
12 Conductive carbon adhesive tabs
13 Eppendorf pipette, 10-uL capacity
14 Disposable pipette tips
15. 1 - 10 mL pipette
16 Glass vials for sonicating dust in isopropanol suspension (holds 10-mL volume)
17 Razor blade
18 Ultrasonic bath
19 50 mL glass beaker
20 Polycarbonate filters (25-mm diameter, 0 4-um pore size)
21 Polycarbonate filters (13-mm diameter, 0 4-um pore size), or borer to cut larger filters
to SEM stub size
22 11-mm diameter cork borer
23 Milhpore filter apparatus for use with 25 mm filters
24 125 mLNalgene bottles
25 Hand-held vacuum pump
26 High-vacuum carbon evaporator with rotating stage
27 Glass petn dishes with lids
28 Adobe Photoshop Software, or similar
29 Glass petrographic slides
30 Glass cover slips
31 Polarized light microscope for mineral identifications
32 Scanning Electron Microscope with the following attributes
a. Resolution 5 nm (at 25 kV, WD=10 mm - system dependent) or better
b Accelerating Voltage. 10 to 20 kV
c Minimum magnification range 50x to 200,000x
d SEI (secondary electron image)
e BEI (backscattered electron image)
f Energy dispersive x-ray detector and analyzer for EDS analysis
38
-------
g. Ability to collect x-ray maps or particle analysis software (preferably both)
8.0 Reagents
1 Isopropanol, reagent grade [CAS No. 67-63-0]
2 1 55 or 1.605 Refractive Index Oil
9.0 Sample Storage
Dust samples will be stored in an air-tight container, such as a sealed glass jar Samples
placed in reagents will be labeled appropriately and stored according to laboratory safety
standards Samples prepared for analyses will be stored in a protective container, such as
a plastic case or covered petn dish, to prevent contamination
10.0 Quality Control
Quality control is implemented by thoroughly cleaning glassware and spatulas, keeping
working surfaces clean, and preventing cross contamination During ashing, particles
may be suspended if slow heating is not achieved Following the ashing program as
outlined will minimize flashing, which can cause particles to become airborne Covered
crucibles will be used to prevent contamination caused by flashing Used Eppendorf
pipette tips and weighing papers will be discarded and new tips and papers will be used
for each sample
Duplicate samples shall be prepared to determine the precision of the analysis In
addition, sample blanks shall be prepared These blanks are checks for cross
contamination during handling of the samples. Blanks shall be prepared at the same time
and in the same manner as samples
10.1 Calibration
Calibration of the EDS system must be completed at least once at the beginning
and again at the end of each analytical session Backscattered electron image
(BEI) calibration should be performed at the beginning of the session and anytime
the backscattered image brightness and/or contrast is adjusted
EDS calibration for both qualitative and quantitative (not required by this method
but could be useful for identification of particle type) analysis is accomplished by
the analysis of a polished carbon-coated reference standard The recommended
material is USGS BIR1-G basalt glass mounted in epoxy in a brass tube, polished,
and carbon coated using a carbon evaporator (2, 3)
The calibration reference material should be analyzed at the same operating
conditions to be used for the analysis including beam current, accelerating
voltage, working distance, detector dead time, and sample tilt (= 0°) For BIR1-G
39
-------
the analysis should be performed with a beam size of 10-20 urn or equivalent area
raster All calibration spectra will be saved with the corresponding data set. The
calibration data will be used for inter- as well as mtra-laboratory comparisons
This calibration is in addition to, and not a substitute for the normal EDS
calibration recommended by the EDS manufacturer which will be performed at
regular intervals as specified by the EDS manufacturer
Backscattered electron detector calibration can be performed on the same BIR1-G
material by adjusting the detector brightness and contrast to achieve the following
conditions The epoxy on the BIR1-G reference material will be at 0 in a 256
grayscale image and the brass mounting tube will be at 256 The BIR1-G basalt
glass should fall at approximately 130-140 gray scale units
11.0 Procedure
11.1 Weighing and Splitting
Weighing and splitting should be performed under a negative flow HEPA hood
If the fan speed is set too high, loss of particles may occur The fan speed may
need to be adjusted to prevent the loss of fine particles
Obtain an analytical balance with an accuracy of 0 0005 g and preweigh a clean
piece of weighing paper Transfer the dust from the sample vial to the weighing
paper and determine the weight of the dust Split the sample with a clean razor
blade using the cone-and-quarter method If there are large clumps of organic
fibers, such as hair or lint, temporarily remove the hair with a pair of forceps and
tap the forceps lightly with another tool over a piece of weighing paper to remove
fine particles. Center the fine fraction on the paper and split the sample into four
equal parts using a razor blade Collect opposite comers ('/z of the sample) for
analysis and archive the other half Quarter the larger organic fiber bundles the
same way, keeping half to proceed to the ashing step and half for archival
purposes
Place the two quarters for ashing into a pre-vveighed crucible Weigh the split and
record the results
11.2 Ashing
Place the ceramic crucibles containing the samples into a furnace.
The furnace program should proceed as follows
1 Increase temperature by 1 °C/rmnute until sample reaches 250 °C
2 Hold temperature at 250 °C for 4 hours
3 Increase temperature by 1 °C/rmnute until sample reaches 480 °C
4. Hold temperature at 480 °C (sufficient for decomposing organics) for 8 hours
Do not exceed 500 °C
40
-------
5 Shut off furnace
6 Allow sample to cool before removing from furnace.
7 Weigh the ashed sample to the nearest 0 0005 g and record the result
11.3 Sieving
Sieve the sample through a 150-um sieve using a Retsch ultrasonic sieve shaker,
or similar Three-inch diameter sieves are recommended to minimize sample loss
from particles being trapped in the sieve. The ultrasonic shaker will be operated
at 20-minute intervals at the following settings' 20, 40, 60, 70, 80, then back
down to 50 and 20. This will provide amplitudes ranging from 0 to 1 5 mm
Transfer the large and small fractions to clean pieces of weighing paper and
weigh to the nearest 0 0005 g Archive the fraction greater than 150-um
11.4 Preparation of Sample for Polarized Light Microscopy
Split the less than 150-um sample fraction using the cone and quarter method
Collect one comer for analysis and archive the other three quarters Weigh the
quarter split to the nearest 0 0005 g and place it into a glass vial Make a
suspension of 10-20 mg dust per mL of isopropanol. The amount of isopropanol
needed will vary depending on the amount of dust; the target dilution is 10-20 mg
per mL.
Cut an Eppendorf pipette tip with a razor blade to increase the opening to
approximately 1 mm
Place the suspension in an ultrasonic bath for one minute, then remove the
suspension from the ultrasonic bath and shake it gently to suspend all particles
Collect a 10-uL aliquot of the mixture using an Eppendorf pipette with the
modified tip and transfer to a glass slide. Prepare 4 such slides Allow them to
dry, then add a drop of 1 55 (or 1 605) refractive index oil
11.5 Preparation of Sample for SEM Analysis
Prepare the SEM substrate on aluminum stubs using 0 4-um pore size
polycarbonate filters, carbon adhesive tabs Using an 11 mm filter punch and
placing the filter between two filter separators, punch a circle the size of the
carbon tab into the filter Place carbon adhesive tab affixed to an aluminum stub
on the dull side of the 11-mm polycarbonate filter such that the shiny side of the
filter exposed If available, a 13-mm diameter polycarbonate filter may be used in
place of the punched out 11-mm filter
Collect a 10-u.L aliquot of the mixture from the PLM sample preparation using
the Eppendorf pipette with the modified tip and transfer to a prepared
polycarbonate/adhesive tab substrate This will yield a loading on a 12-mm SEM
41
-------
stub of about 100-200 ug, which is a moderately heavy loading Adjust the
number of aliquots as needed to obtain the target loading
Prepare a 10-fold dilution of the above suspension to get a suspension of 1-2 mg
dust per mL of isopropanol Sonicate the suspension in an ultrasonic bath for one
minutes Remove the suspension and gently shake it to suspend all particles
Wait one minute to allow the coarse particles to settle Collect a 10-u.L aliquot of
the suspended mixture using an Eppendorf pipette with the modified tip and
transfer to a prepared polycarbonate/adhesive tab substrate. This will yield a
loading on a 12-mm SEM stub of about 10-20 ug, which is a light loading
Adjust the number of aliquots as needed to obtain the target loading.
Alternatively, prepare a lightly loaded sample using the filtration method as
follows Use a Millipore filter apparatus for use with 25-mm filters for filtration
Place a few drops of isopropanol on the fritted glass surface and place the 25-mm
polycarbonate filter (0 4-um pore size) on the isopropanol Attach the top of the
apparatus and add a few milhhters of isopropanol to the filter so that no part of it
is exposed to air. Sonicate the suspension (diluted as described in previous
paragraph) in an ultrasonic bath for one minute. Remove the suspension and
gently shake it to suspend all particles Wait one minute to allow the coarse
particles to settle Collect 1 mL of the suspended mixture using a pipette and
filter it through the polycarbonate filter Actual amounts for filtration will vary
based on sample loading The goal is to have a loading on a 12-mm SEM stub of
about 10-20 ug, or about 5-10 percent area coverage, which is a light loading
Adjust the volume of the aliquot to filter as needed to obtain the target loading
Place the filter on a carbon adhesive tab on a standard SEM aluminum mount
The filter needs to be completely flat on the SEM stub This can be achieved by
forming the wet filter into a gentle U-shape using forceps and the side of the
forefinger, then placing the bottom curve of the filter onto the center of the carbon
adhesive tab and slowly releasing the sides so they lay flat Trim the edges of the
filter using a razor blade.
After drying, coat the samples on the polycarbonate or polycarbonate/adhesive tab
substrates with carbon using a carbon evaporator with a rotating stage Transfer
the stubs to the SEM in a clean, covered container
12.0 Analysis
12.1 Analysis by Polarized Light Microscopy
Pol an zed light microscopy will be conducted using the general techniques
outlined in EPA 600/R93/116 (4) For this procedure, four slides (prepared as
described in section 114) will be analyzed 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 The
42
-------
fraction of fibers with refractive index less than 1.55 (or 1.605) will contain
primarily soda-lime glass fibers. For the validation study, numbers of fibers
greater than and less than 1 55 (1.605) refractive index will be counted
Dispersion staining and becke line techniques may be used. Fiber point counting
will be performed at 100 x magnification
If more than 20 mineral wool fibers are found, continue counting and recording
all of the fibers above and below the index oil refractive index Report both raw
fiber counts per refractive index category and number of fibers from each
category per gram of sample. Continue on to step 122 1 to determine the ratio of
slag wool to other fibers with refractive index greater than 1.55 (or 1 605) using
EDS as described below
If less than 20 mineral wool fibers are found on each slide, count the number of
slag wool fibers using SEM/EDS and report as number of fibers per gram of
sample
12.2 Analysis by SEM/EDS
12.2.1 Screening for Slag Wool
Operating conditions for the JEOL 6460-LV SEM are 15 kV, 0 5-5-nA
beam current, 10-mm working distance (system dependent), and zero
degree tilt
Place the more concentrated sample deposited directly on the
polycarbonate/adhesive tab substrate into the SEM. Use the backscattered
electron mode at lOOx magnification to quickly distinguish carbon fibers
from inorganic fibers (carbon fibers may be visible, but not as bright in a
BEI). Identify all inorganic fibers over 25 urn in length (smaller fibers
cannot be reliably detected at the lOOx operating magnification) When an
inorganic fiber is found, identify the composition of the particle by EDS.
Slag wool is the primary fiber of interest. Record all inorganic fiber
results as number of fibers for each fiber type
For the samples with high fiber loading, as determined by PLM as
described in section 121, count fibers per type until a statistical
representation of the ratios of fiber compositions in the sample is
achieved Report the ratio (by fiber number) of slag wool fibers to total
MMVF fibers corresponding to the high RI Use this ratio to correct the
total number for high RI fibers counted by PLM to number of slag wool
fibers present
For the samples with low fiber loading, as determined by PLM as
described in section 12.1, scan the entire stub to determine the number of
43
-------
fibers per type Report the slag wool fiber results as the number of slag
wool fibers/gram of sample
12.2.2 EDS Screening for Gypsum/Anhydrite
Place the more concentrated sample deposited directly on the
polycarbonate/adhesive tab substrate in the SEM Choose a random field
at lOOx magnification and perform an EDS analysis on the entire field
Look for the presence of sulfur in this field If sulfur is present, continue
to Section 12 2 3 or 12 2 5 for analysis of gypsum and concrete by
mapping or particle analysis If it is not present, repeat the analysis on
another random field If sulfur is still not present, mark the sample as non-
detect (ND) for sulfur
12.2.3 X-Ray Mapping for Gypsum
Place a more dilute sample, deposited directly on the
polycarbonate/adhesive tab substrate or prepared by filtration, in the SEM
Collect binary backscattered electron images (particles white and
background black, shadow off) and secondary electron images for 10 non-
overlapping, random fields at 500 x magnification Collect x-ray maps for
Na, Mg, Al, Si, S, Ca, and Fe at each of these fields Fields containing
MMVF will not be used for this analysis Operating parameters for the
SEM are the same as those for analyzing slag wool. Acquisition
parameters for x-ray mapping using the NORAN System Six Software are
time constant 14 (mapping mode, 11333 cps), 10-20 % deadtime, 256 x
256 image resolution, 20 second frame time, and 100 frames collected
(about 40 minutes total acquisition time) Secondary electron images will
be used for reference only Save all of the maps and electron images in
TIFF format
Open the backscattered electron image and the Ca and S x-ray maps in
Adobe Photoshop Make sure that all of the element maps are the same
size and resolution by choosing Image Size from the Image Menu and
changing the pixel size or the resolution as needed The presence of
gypsum can be determined by overlapping the Ca and S maps
Perform the following functions in Adobe PhotoShop (A macro is in
development to perform the following functions to decrease user time and
human errors in adjusting the threshold )
1 Convert each of the three images to grayscale (Image —» Mode —>
Grayscale)
2 Perform an auto contrast and brightness on each image and map to
increase the scale of colors (Image —> Adjustments —» Auto Levels)
3 Threshold each element map, Ca and S (do not analyze the
44
-------
backscattered electron image at this time), by going to the Image Menu
and choosing Adjustments —* Threshold. Adjust the threshold to 128
The background will be black and the particles white
4 Invert the image (Image—* Adjustments —»Invert) to make the
background white and the particles black
5 Copy the S map and paste it over the Ca map in a separate layer in the
file and change the opacity (located in the Layers window) to 50 % for
the S map layer The black areas are gypsum/anhydrite
6 Display a histogram of the image m expanded mode by selecting the
Histogram tab on the Navigator Window (or under the Image Menu in
some versions of Photoshop) Place the cursor over the line for the
black area and record the percentile for the black area. This is the
percentage of particles containing Ca and S in the entire field.
NOTE. If a binary backscattered electron image is obtained during data
collection, then steps 7-11 may be deleted The Invert function will,
however, need to be applied to make the particles black and the
background white before continuing to step 12.
7 Begin analysis of the backscattered electron image Select the
particles by going to the Select Menu and choosing Color Range Go
to the selection pulldown menu and choose Highlights
8 Fill the selection with black by going to the Edit Menu —» Fill and
choosing black from the color pulldown menu.
9. Select the inverse areas by going to the Select Menu and selecting
Inverse
10 Fill the selection with white by going to the Edit Menu —» Fill and
choosing white from the color pulldown menu
11 Deselect the area by clicking on the image
12 Perform the Threshold and Histogram functions for the backscattered
electron image as outlined in 3 and 6. Record the histogram result for
the backscattered electron image
Determine the area percent of gypsum by performing the calculations in
Section 13.0.
12.2.4 X-Ray Mapping for Ca-Rich Particles
Analysis of components of concrete will be performed on the same fields
as the gypsum/anhydrite analysis At this time, only a method for the
determination of the area percent of Ca-rich particles is presented See
Section 2 1 for discussion.
Perform the following steps on the Ca x-ray map Tiff file in Adobe
Photoshop
45
-------
1 Convert the Ca x-ray map to grayscale (Image —> Mode —» Grayscale)
2 Perform an auto contrast and brightness on the map to increase the
scale of colors (Image —» Adjustments —» Auto Levels)
3 Threshold the Ca map by going to the Image Menu and choosing
Adjustments —» Threshold Adjust the threshold to 128. The
background will be black and the particles white
4 Invert the image (Image—> Adjustments —»Invert) to make the
background white and the particles black
5. Display a histogram of the image Place the cursor over the line for
the black area and record the percentile for the black area This is the
area percent coverage of particles containing Ca in the entire field
Determine the maximum area percent coverage of non-gypsum, Ca-nch
particles by performing the calculation in Section 130
12.2.5 Particle Analysis for Identification of Gypsum and Concrete.
Place the more dilute sample, deposited directly on the
polycarbonate/adhesive tab substrate or prepared by filtration, in the SEM
Particle analysis will be used to identify gypsum and concrete particles
Perform particle analysis at 500 x magnification. All other operating
parameters for the SEM are the same as those used to analyze for slag
wool (Section 1221) A binary backscattered electron image should be
used in particle analysis mode. Particle analysis parameters should be set
to analyze all particles m the field greater than 0 5 urn and to separate
touching particles For particles greater than 5 urn, scan the entire
particle, spot analysis is adequate for smaller particles The x-ray
spectrum and counts for all particles, and an image of particles > 20 um
long, will be recorded and saved Other particle parameters to be reported
will include the maximum, minimum, and average diameters, the aspect
ratio, and area of each particle
It will be necessary to review data collected by automated software to
ensure data integrity An Excel spreadsheet, in conjunction with images
and x-ray data, may be used for this purpose Particles should be sorted
into one of three categories. Ca-S (gypsum), Ca-nch, and Other. Aid in
identification of particles may by facilitated by referencing the U.S
Geological Survey's WTC Dust Particle Atlas (1) A particle
classification protocol will be developed based on the data from the
validation study
The number of particles analyzed will be determined using the results of
the validation study For the study, the area percent of each component
should be within 10% relative error or better. Typically, data for 1000 -
1200 particles should be acquired.
46
-------
Results for particle analysis will be recorded as area percent gypsum and area
percent concrete particles for each field and average area percent for the each
component m the sample
13.0 Data Analysis and Calculations
1. To determine the concentration of slag wool in fibers/gram, perform the following
calculations:
Determine the number of fibers with RI > 1 55 (or 1 605)
# fibers identified - mg of sample on slide * 1000 = fibers/gram on slide
Determine the percentage of fibers with the composition of slag wool with RI >
1 55 (or 1 605)
Fibers/gram on slide x # fibers identified as slae wool = fibers slag wool/gram on slide
Total number of fibers identified by EDS with RI > 1 55 (or 1 605)
Back calculate to the number of fibers per gram of the original sample
Fibers slag wool/g on slide * g after sieving x E sample after ashing = Total f/g of sample
g before sieving * g sample before ashing
2 To determine the area percent of gypsum/anhydnte from the x-ray mapping
procedure, perform the following calculations:
Determine the area percent of gypsum/anhydnte in each field of view
% of black area in Ca-S map overlay * 100= area % gypsum
% of black area in BSE image
Calculate the average percentage of gypsum/anhydnte for the sample
(area % gypsum)(i + (area % gypsumVi + = Avg area % gypsum
number of fields
3 To determine the maximum area percentage of Ca-rich particles, which includes
concrete particles, from the x-ray mapping procedure, perform the following calculations
Determine the area percent of non-gypsum Ca-nch particles in each field of view:
f/o black area Ca map") - f% black area Ca-S map") = % non-gypsum Ca-nch particles
% black area on BSE image
Calculate the average percentage of non-gypsum Ca-nch particles for the sample-
(area. % Ca-nch particlesVi + (area % Ca-nch oarticleslp + Avg area % Ca-nch particles
47
-------
number of fields
4. Calculate the area percent for gypsum and concrete by summing the areas of each
particle in for each particle type and dividing by the total area analyzed.
area gypsum 1 + area evpsum 2 + x 100 = area percent gypsum (do likewise for concrete)
total area analyzed
Rules for concrete and gypsum classification are currently being developed
14.0 References
1 Lowers, Heather A, Meeker, Gregory P, Brownfield, Isabelle K, 2005 World
Trade Center Dust Particle Atlas U S Geological Survey Open-File Report 2005-
1165 On the web athttp //pubs uses uoy/of/2005/J J65/
2 Meeker, G P , Taggart, J E , and Wilson, S A, 1998 A Basalt Glass Standard for
Multiple Microanalytical Techniques. Proceedings Microscopy and
Microanalysis 1998 Microscopy Society of America.
3 A polished and carbon coated calibration reference sample of BIR1-G may be
obtained by contacting Stephen Wilson, U.S Geological Survery, MS 973,
Denver Federal Center, Denver, CO, 80225, svyijson^iis^s goy
4 Perkins, R.L and Harvey, B.W., 1993, TEST METHOD: Method for the
Determination of Asbestos in Bulk Building Materials, EPA/600/R-93/116
48
-------
15.0 Appendix: DATA SHEETS
Determination of Slag Wool Fibers In Dust- PLM with Dispersion Staining
Sample ID
Circle One
General Sample
Homogeneous?
Structure #
Project
Analyst
Original Duplicate Triplicate Date
Appearance
Y
Rl Fluid
156
1 60S
Dispersion Staining
>RI
RI
-------
SEM Sheet
Reference ASTM - D5755-03
Report Number: _
Sample Number:.
File Name:
Sample Description.
Preparation Date:
Analysis Date:
Computer Entry Date:
Sample weiaht:
Dilution Volume:
Volume Aliquot:
Magnification:
Bv:
Bv:
Bv:
grams
mL
uL
X
Structure #
Field #
Fiber Type
Length (Microns) Width (Microns)
Image
EDS
50
-------
Ai
Ctte
*y^ Rvttdcs in ^B
-------
APPENDIX C: REAC SOP 2040
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE of 10
REV 00
DATE
05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
CONTENTS
1 0 SCOPE AND APPLICATION*
2 0 METHOD SUMMARY*
3 0 SAMPLE PRESERVATION, CONTAINERS, HANDLING, AND STORAGE*
4 0 INTERFERENCES AND POTENTIAL PROBLEMS
50 EQUIPMENT/APPARATUS*
5 1 Sampling Equipment*
5 2 Sieving Equipment*
6 0 REAGENTS*
7 0 PROCEDURES*
71 Preparation*
7 2 Calibration Procedures*
7 3 Field Operations*
7 4 Sieving Procedures*
7 5 Sampling Tram Decontamination*
8 0 CALCULATIONS*
90 QUALITY ASSURANCE/QUALITY CONTROL
10 0 DATA VALIDATION*
52
-------
11 0 HEALTH AND SAFETY
120 REFERENCES*
130 APPENDICES*
These sections affected by Revision 0 0
SUPERCEDES SOP #2040, Revision 0 0,11/18/98, USEPA Contract 68-C4-0022
53
-------
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 54 of 10
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
1 0 SCOPE AND APPLICATION
The purpose of this Standard Operating Procedure (SOP) is to define the procedures for the collection of
carpet-embedded dust samples that can be analyzed for lead, pesticides, or any other chemicals or elements
This procedure is applicable for the collection of samples on a variety of carpeted surfaces This SOP may
be modified to include the collection of dust adhering to floor surfaces but is not intended for the collection
of dust containing asbestos fibers
These are standard (i e, typically applicable) operating procedures which may be varied or changed as
required, dependent upon site conditions, equipment limitations or limitations imposed by the procedure
In all instances, the ultimate procedures employed should be documented and associated with the final
report
Mention of trade names or commercial products does not constitute United States Environmental Protection
Agency (U S EPA) endorsement or recommendation for use
2 OMETHOD SUMMARY
Sample collection is performed utilizing the Nilfisk GS-80 vacuum cleaner equipped with a high efficiency
particulate air (HEPA) filter A diagram of the Nilfisk GS-80 dust sampling apparatus is presented in
Figure 1, Appendix A Soil and other particulate matter with aerodynamic diameters of approximately 5
microns (^m) and larger that are embedded within the carpet are collected, sieved and submitted to the
laboratory for analysis
3 0 SAMPLE PRESERVATION, CONTAINERS, HANDLING AND STORAGE
Following collection of a sample into a dedicated collection bag, the bag is removed from the vacuum
cleaner and placed into a 32-ounce(oz) glass jar or a zip-lock plastic bag Storage of the samples at
ambient temperature is appropriate for samples that will be analyzed only for metals Samples for organic
analysis should be maintained at approximately 4 ± 2 degrees Celsius (°C)
4 OINTERFERENCES AND POTENTIAL PROBLEMS
There are no known interferences with this method
5 OEQUIPMENT/APPARATUS
5 1 Sampling Equipment
• Nilfisk Model GS-80 vacuum cleaner
• Two-meter folding ruler or similar device
• Masking tape
54
-------
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 55 of 10
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
Deiomzer or distilled water
Methanol, ACS grade
Kimwipes ™ or equivalent
Vacuum collection bags
Bottle brush
Scrub brush
Plotlmes
32-ounce glass jars or Ziploc plastic bags
Disposable gloves
5 2 Sieving Equipment
• 100-mesh sieve, 150-Dn mean diameter, as specified in ASTM D 422,
consisting of the cover, sieve and receiver pan
• Sieve shaker for mechanical sieving (CSC Scientific, Catalog Number 18480,
Thomas Scientific, Catalog Number 8324-A10) or equivalent
• Analytical balance, capable of weighing 0 1 milligrams (mg) and a range of 0 1
mgto 1000 grams (g)
• Disposable gloves
• Disposable dust mask
» Clean aluminum foil
• Kimwipes ™ or equivalent
• Camel hair brush (TFisher Scientific, Catalog Number 03-655) or equivalent
6 0 REAGENTS
Methanol and deiomzer/distilled water are required for sampling tram cleaning and decontamination
7 0 PROCEDURES
7 1 Preparation
The overall sampling strategy should be designed to address the goals of the study Users should
consider factors such as foot traffic volume, types of activities, and proximity to potential sources
The sampling strategy should be descnbed in the Work Plan (WP), Quality Assurance Project Plan
(QAPP), or Sampling and Analysis Plan (SAP) prepared prior to the sampling event The ideal
sampling locations are those areas that conform to the overall sampling strategy For example,
protocol may require the selection of a carpeted area for sampling where small children play or are
likely to play
55
-------
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 56 of 11
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
1 Determine the extent of the sampling effort, the sampling methods to be employed, the
amount of dust needed to reach the desired detection limit and the types and amounts of
equipment and supplies needed
2 Obtain and organize the necessary sampling and monitoring equipment
3 Decontaminate or pre-clean equipment, as specified in Section 7 5, and ensure that it is
in working order
4 Prepare schedule and coordinate with staff, client, regulatory agency, as appropnate
5 Perform a general site survey pnor to site entry in accordance with the site-specific
Health and Safety Plan
6 Measure the area to be sampled and outline it using masking tape or other appropriate
methods Draw a diagram of the room(s) where the sample(s) were taken, locating the
sampled area(s)
7 2Cahbration Procedures
The Nilfisk GS-80 vacuum cleaner has no flow devices that require calibration pnor to sampling
The sampling train shall be thoroughly inspected to ensure that it has been cleaned, properly
assembled, and complete
7 3 Field Operations
1 Prior to collecting a sample at a specific location, complete a Vacuum Sampling Work
Sheet (Figure 2, Appendix A) recording all required information and sketch the area to be sampled
2 Select a sampling area according to the data collection design outlined in the WP, QAPP
or SAP Typically, three rooms per floor are selected for sampling in each building Each sample is collected
with a dedicated sampling train that has been properly assembled, cleaned, and decontaminated to ensure
sample integrity The size/weight of each sample is dependent on the goals and objectives of the sampling
event, the analyses requested, and the desired method detection levels (MDLs) A 100-g sample is highly
desirable if multiple analyses (metals, pesticides, etc ) are requested A minimum 5- to 10-g sample is required
for metal analysis only
3 Using the 2-meter folding ruler or any other measuring device, outline and mark the
recommended 1-square meter (m2) portion of the carpet to be sampled
4 Begin collecting sample at one corner of the delineated sample area, moving the sampler
back and forth four times over a strip running in a straight line between the defined
56
-------
PsKpoittft Hwjfuasw-y onrt Atx&fwui Cortiiw.*
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 57 of II
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
sampling area edges The width of the strip is defined by the width of the sampling
nozzle After completing the first stnp, angle over to the second strip gradually on the
next pass, again completing four double passes
5 Continue sampling the delineated area until an adequate sample is collected Visual
observation is used to determine if enough sample has been collected from the
recommended 1-m2 area or if a larger area is required If sampling a larger area, measure
the area accurately and document accordingly
6 Wearing surgical gloves, be sure to tap with your hand on the nozzle inlet to dislodge any
dust remaining in the nozzle or the hose This procedure will ensure complete sample recovery Turn off the
vacuum cleaner and allow to sit undisturbed for at least 30 seconds Unsnap the two vacuum container clips to
access the inside of the container Remove the polylmer and the vacuum collection bag within it Seal off the
polyliner with the vacuum collection bag inside, and transfer to a properly labeled 32-oz glass jar or plastic bag
depending on the analysis(es) to be performed Document the sample information on the Vacuum Sampling
Work Sheet and pack properly for shipment to the laboratory
7 Remove the hose and the nozzle, and install a new polyliner and collection bag for the
collection of additional samples
8 Decontaminate the vacuum components using the steps outlined in Section 7 5
7 4 Sieving Procedures
Prior to submitting dust samples to the laboratory for analysis, the samples are sieved through a
100-mesh sieve using the following procedure
1 Select a clean working area in a facility equipped with a fume hood (a 4-foot by 4-foot
area is sufficient) Weigh the receiver pan on an analytical balance and record the weight
2 Wearing clean surgical gloves and a dust mask, retrieve the vacuum collection bags from
the 32-ounce glass jars used to transport the bags from the field to the laboratory
3 Empty the entire contents of the bag into the 100-mesh sieve with the receiver pan
attached Remove the plastic adaptor (blue ring) from the collection bag inlet and shake the bag as
necessary to ensure all the contents have been transferred into the sieve
4 Place the cover on the sieve and manually or mechanically shake the sieve for a
minimum of 5 minutes and a maximum of 10 minutes until all the fine dust particles are collected in the
bottom receiver pan If manual shaking is performed, follow the instructions given in American Society for
Testing and Materials (ASTM) D-422 "Conduct the sieving operation by means of a lateral and vertical
57
-------
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 58 of 11
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
motion of the sieve, accompanied by a jarring action in order to keep the sample moving continuously over
the surface of the sieve Continue sieving until not more than 1 mass percent of the residue on a sieve
passes that sieve dunng 1 minute of sieving"
If mechanical shaking is performed, set up the recommended sieve shaker on an even and
stable surface Proceed with the sieving operation following directions in the
manufacturer's manual
5 Re-weigh the receiver pan using an analytical balance The difference in weight is the
weight of the sieved sample If total weight of material is desired, the coarse material
remaining on top of the sieve must be collected on a pre-weighed sheet of aluminum foil,
re-weighed and the weight added to the weight of the sieved sample
6 Transfer the sieved sample from the receiver pan to an 8-oz wide-mouth glass jar Use a
camel hair brush to ensure complete transfer of the sample Cap the glass jar securely
7 Document each sample Each sample must be provided with the following information
identification number, date of sampling, location, analysis requested Each sample must
be recorded onto a Cham of Custody form before delivery to the analytical laboratory
8 Before processing the next sample, thoroughly wipe clean the cover, sieve and receiver
pan using a Kimwipe™ and deiomzed/distilled water Let dry prior to sieving additional
samples
7 5 Sampling Tram Decontamination
To decontaminate the sampling trains, move them to a well-ventilated area and perform the
following
3 Assemble one of the sampling trains to be used as the decontamination unit for
decontaminating the nozzles, hoses, and wands This unit must be equipped with a clean polylmer and dust bag
! With the vacuum cleaner turned on, decontaminate the nozzles, wands, and hoses using a
bottle brush to remove any accumulated dust in the hose and nozzle Be sure to tap the nozzle with your hand to
remove any visible dirt that has accumulated, and use the scrub brush to remove any hair or fibers entangled on
the nozzle's brush When the nozzle is considered to be clean, remove and spray with reagent grade methanol
and allow to air dry on a clean surface The wand and hose are then cleaned with the bottle brush Tap your
hand on the wand inlet while cleaning with the bottle brush to remove any visible dirt Repeat this procedure to
decontaminate any remaining nozzles, wands, and hoses
2 Remove the used dust bag from the decontamination unit and wipe clean the inside of the
58
-------
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 59 of 11
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
container with deiomzed/distilled water Spray the inside of the containers with methanol
and allow to air dry When decontaminating in between residential homes, cleaning the
inside of the containers with deiomzed/distilled water is sufficient
8 OCALCULATIONS
The dust weight calculations for the final sieved dust fraction is performed in accordance with ASTM
Method D-422 Dividing the final dust weight by the area sampled (expressed in m2) provides dust loading
in grams per squared meter ( g/m2 ) When the analysis results are received, the loading of analyte in
micrograms per square meter of carpet area (ug/m2) can be calculated in the same way The analysis
provides concentrations in milligrams/kilogram (mg/kg) or micrograms/kilogram (Cg/kg) If total (gross)
dust loading of the sampled area needs to be calculated, the total dust weight before sieving must be
obtained The total dust weight is divided by the area sampled to obtain total dust loading in g/m2
9 0 QUALITY ASSURANCE/QUALITY CONTROL
There are no specific quality assurance activities which apply to the implementation of these procedures
However, the following general QA procedures apply
1 All data must be documented on field data sheets or within site logbooks
2 All instruments must be operated in accordance with operating instructions as supplied by the
manufacturer, unless otherwise specified in the work plan Equipment checkout and calibration
activities must occur prior to sampling/operation and they must be documented
10 0 DATA VALIDATION
The information recorded dunng sampling will be used in conjunction with the analytical data during
validation
11 0 HEALTH AND SAFETY
When working with potential hazardous matenals, follow U S EPA, Occupational Safety and Health
(OSHA) and corporate health and safety procedures
120REFERENCES
American Society For Testing And Matenals 2000 Standard Practice for Collection of Dust from
Carpeted Floor for Chemical Analysis, Designation D 5438-00, Repnnted from the Annual Book of ASTM
Standards, Philadelphia, PA
59
-------
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 60 of 11
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
American Society For Testing And Matenals 1998 Standard Test Method for Particle Size Analysis of
Soils, Designation D 422-63, Repnnted from the Annual Book of ASTM Standards, Philadelphia, PA
Instructions for Use-Nilfisk Model GS 80, Nilfisk of America, Inc .Malvern, PA (1987)
130 APPENDICES
A - Figures
APPENDIX A
Figures
SOP #2040
May 2002
60
-------
STANDARD OPERATING PROCEDURES
SOP: 2040
PAGE: 61ofll
REV: 0.0
DATE: 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
NILFISK
^O Qf\
GooO
FIGURE 1. GS-80
Dust Sampling Apparatus
61
-------
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 62 of 11
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
V.S BPM
P»g. of
Rnpentt E8(lMcrfB| AMtytfetl CoMmt
Vienna Stnpftef Work hk«M
L««n< Mmn Ckm. Eton. Nl
OPA/SftTCWAM
KE AC Tttt Uriel.
U»l Otto-
OAu
FIGURE 2 Vacuum Sampling Work Sheet
62
-------
ftnptttaa HmjtnwsN.1 x<) oral Anrt-vV'oJ Cws>sw.»
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 63 of 11
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
SIEVE CLEANING
SCOPE AND APPLICATION
Sieves must be cleaned before each dust sample is separated into fractions Most of the "near-
mesh size" particles can usually be removed from the apertures by inverting the sieve and gently
tapping the frame of the sieve For sieves with apertures less than 1 millimeter (mm) (e g, 100-
mesh, 150 micron [Dn] sieve), the most effective method for cleaning the apertures is the use of
an ultrasonic bath
EQUIPMENT/APPARATUS
• Ultrasonic bath, capable of holding a standard sieve
• Magnifying glass
• Source of air, standard hair dryer or compressed air
• Spray bottle
REAGENTS
• Ultrasonic cleaner or laboratory-grade detergent that leaves no interfering residues
• Deionized (DI) water, Type II water or equivalent
Methanol, American Chemical Society (ACS) grade or equivalent
PROCEDURE
The following cleaning procedure will be used to clean sieves prior to use and after each sample
1 Place the sieve into an ultrasonic bath containing detergent and DI water and sonicate for approximately 10
minutes
2 Remove the sieve from the ultrasonic bath and nnse well with DI water
63
-------
STANDARD OPERATING PROCEDURES
SOP 2040
PAGE 64 of 11
REV 00
DATE 05/17/02
COLLECTION OF INDOOR DUST SAMPLES FROM CARPETED SURFACES FOR
CHEMICAL ANALYSIS USING A NILFISK GS-80 VACUUM CLEANER
3 Spray the sieve with methanol
4 Dry the sieve using a standard hair dryer or a compressed air source
5 Visually inspect the sieve to ensure that there are no remaining particles present in the apertures A
magnifying glass may be used to aid in this process
6 Repeat steps 1 through 5 prior to sieving subsequent samples
64
Rev 0, 3/11/05
-------
APPENDIX D: ACCESS AGREEMENT
REQUEST FORM
for the U.S. Environmental Protection Agency's Effort
to Develop and Validate a WTC Dust Signature
and to Characterize Background Dust in New York City
Name of Occupant:
Address:
Apartment Number (if applicable)
Contact Phone Numbers
REQUEST
I have read the fact sheet on the U S Environmental Protection Agency's Program to develop and
validate a WTC dust signature and to characterize background dust in New York City and considered
the information provided to me by the U S Environmental Protection Agency (EPA) Having
considered the information regarding the sampling program, I would like to participate
AGREEMENT
On behalf of myself and any other occupants, I agree to the following
I consent to employees, authorized representatives and contractors of the EPA having access to the above
referenced space for as long as necessary to conduct dust sampling activities
1 agree to obtain any required permission for sampling activities from my building management I also
agree to inform the EPA Project Monitor at least one business day prior to the scheduled work of any
building rules that are applicable to the program, including time restrictions, appropriate entrances to the
building, and elevator usage
I understand that the sampling will be performed by contractors retained by EPA I also understand that the
contractors performing sampling activities are required to maintain insurance coverage for commercial
general liability, workers compensation, dishonest acts of their employees and environmental impairment
liability related to this work The contractors are required to maintain such insurance at all times that they
are conducting sampling activities The contractors are responsible for damage or loss of property
I understand that the activities will require access to interior spaces and the use of electricity Sampling
activities will be performed throughout the entire space
I understand that the program will employ various methods of dust removal from surfaces, including, but
not limited to, vacuuming and wet wiping
65
-------
COMMUNICATION OF RESULTS
I understand that I will receive a copy of the sampling results for the residence once an analysis has
been completed and the data are quality assured Depending upon analysis and review time, these
results may not be available until up to six months after sampling.
I understand that results provided for locations sampled under the signature study will only indicate
whether WTC dust signature components are present, absent or inconclusive Results provided for
locations sampled under the background study will indicate the presence of WTC signature dust, as
well as the presence and levels of the contaminants of potential concern (COPC)
I understand that an explanation of the findings will be included in these results along with the name
and contact information for a U S. EPA toxicologists/nsk assessor This person will be able to answer
questions regarding data interpretation and health-related issues
I understand that monitoring data in EPA's database for this effort will be made available to the public,
but the identity of the participants and the specific location of the sampling will be kept confidential
AUTHORIZED SIGNATURE
I certify that I am authorized to grant this request on behalf of all the occupants of the above specified
space, and I grant this request and agree to its terms
Signature Date
Name and Title (PRINT)
Signature of U S EPA Representative Date
66
-------
APPENDIX E: INFORMATION SHEET
The U.S. Environmental Protection Agency's Program
to Develop and Validate a WTC Dust Signature
and to Characterize Background Dust in New York City
The September 11, 2001 attack on the WTC covered a large area with dust, debris, and
combustion by-products In order to determine if residual contamination exists, and to identify areas
that may be in need of clean up, the US EPA has undertaken studies both to identify a unique WTC
dust signature, and to characterize typical indoor dust from NY City. In order to complete these
studies, the EPA is seeking to acquire samples of urban dust from buildings both inside and outside of
the area of lower Manhattan that was impacted by the WTC collapse You are being asked to
participate in the study checked below
WTC DUST SIGNATURE STUDY
The U.S Environmental Protection Agency has initiated a study to define signatures for WTC dusts
The purpose of this study is to develop and validate one or more "signatures" in indoor dust that can be
used to determine whether dust sampled is from the collapse of the World Trade Center towers or not.
A "signature" is a chemical or physical characteristic of a material that can be used to identify that
specific material and discriminate between the material sought (WTC dust, in this case) and other
similar materials (NYC urban dusts). The signature materials are not necessarily related to health
concerns The signature could be something harmless but unique to the WTC source, measured only to
identify the origin of other chemicals of concern that occur in the same sample The WTC signatures,
if they can be developed, will support analysis to discriminate between normal indoor dusts and WTC-
generated dusts
Samples from approximately 20 buildings are needed for validation of the proposed signatures
Samples will be collected from approximately 10 buildings in the area that is suspected to be affected
by WTC emissions, and samples will be obtained from 10 buildings that are not suspected of being
affected
SAMPLING METHODS
Dispersion models, photos, interviews, and satellite data will be reviewed to discern areas that were
likely impacted by WTC emissions In each building identified for sampling, dust samples will be
collected from at least three areas 1) one sample from a track-in area near a building entrance,
preferably in a carpeted area, 2) two samples from relatively undisturbed areas (e g , on top of
bookcases, under furniture), and 3) other areas showing visible accumulation of settled dust, including
HVAC ducts A standard method using a HEPA vacuum collector will be used by EPA to collect bulk
dust samples Samples will be sealed and stored under refrigeration in a limited access area
To ensure that these important samples are properly collected, tracked, stored, and distnbuted,
comprehensive quality assurance (QA) procedures will be in place prior to any sample collection
There will be a pre-sampling survey of building and sampling areas, to include photos of sampling
areas (if permitted by building owners) and notes on building usage, to identify conditions that might
compromise samples (e g , smoking or cooking areas)
67
-------
Dust samples from background and affected locations will be made available to researchers involved in
developing and evaluating WTC signatures, as well as researchers characterizing typical NY City dust
When the results of this work are complete, EPA will develop and release reports on these studies
COMMUNICATING RESULTS
Publicly released results will not be provided by name or specific location, thus a resident's privacy
will always be preserved. The occupant will receive a copy of the sampling results for their residence
once an analysis has been completed and the data are quality assured. Depending upon analysis and
review time, this may take up to six months An explanation of the findings will be included in these
results along with the name and contact information for a U S EPA toxicologists/nsk assessor This
person will be able to answer questions regarding data interpretation and health-related issues Finally,
results provided for residences sampled under the signature study will only indicate whether WTC dust
signature components are present, absent or inconclusive
68
-------