United States
Environmental Protection
Agency
Particulate Matter From Carpet
Due to Human Activity
FINAL REPORT
-------
-------
EPA/600/R-07/131 November 2007 www.epa.gov/ord
Resuspension and Tracking of Particulate
Matter From Carpet Due to Human Activity
Final Report
Prepared by:
RTI International
Center for Aerosol Technology
Post Office Box 12194
Research Triangle Park, NC 27709
Order Number: 4C-R179-NALX
Submitted by:
Jonathan Thornburg, PhD
Project Manager
Prepared for:
Jacky Rosati, PhD
Project Officer
U.S. Environmental Protection Agency
National Homeland Security Research Center
Research Triangle Park, NC 27711
Office of Research and Development
National Homeland Security Research Center, Decontamination and Consequence Management Division
-------
Notice
This report is submitted in fulfillment of Order Number 4C-R179-NALXby RTI International under the sponsorship of the
United States Environmental Protection Agency. This report covers a period from May 25, 2004 to September 30, 2005, and
work was completed as of September 30, 2005.
-------
Abstract
Each day adults and children are exposed to paniculate matter (PM) from the flooring and other horizontal surfaces in their
homes and offices. Potential health risks from inhalation and dermal exposure to this paniculate matter exist. This paniculate
matter can include metals, pesticides, or terrorist-based materials such as anthrax, ricin or radiologies. The PM found on
flooring stems primarily from dusts tracked in on shoes and ambient particle penetration from outside that deposits on the
flooring and other horizontal surfaces. Once indoors, PM translocates throughout the residence. The major mechanisms of
translocation are hypothesized to be resuspension and tracking, but the relative importance of these mechanisms is unknown.
The goal of this research was to begin developing a fundamental understanding of generic PM movement within a residence.
This project provided resuspension and tracking data to define the approach and scope of future research to determine primary
routes of PM translocation within buildings.
The resuspension research measured the magnitude of resuspended PM vertical and lateral concentration gradients within a
room, confirmed that PM emission factors calculated from medium pile carpet agreed with laboratory-generated values, and
determined whether residential vacuum cleaners were an effective remediation technique. These experiments were conducted
inside seven occupied homes in the Research Triangle area of North Carolina. Resuspended PM concentrations exhibited
vertical and lateral gradients. On average, the resuspended concentration at 36 inches above the floor was 1.8 times lower
than the concentration measured at 18 inches. Spatial-temporal analysis of the data suggested a time lag of 2 to 5 minutes
between resuspension at the source and transport to instruments 8 feet away. Depending on the room configuration and
size, the concentration 8 feet away was 10% to 50% of the concentration at the source. Calculated emission factors, mass
resuspended per step per unit mass available in carpet, varied from 0.001 to 0.06 mg/step-mg, and followed patterns similar
to those determined during laboratory experiments. Emission factors between houses were statistically different. Previous
research showed variations in particle adhesion and loading characteristics affected emission factors. These data confirm that
carpet age is a significant variable, and fiber length becomes more important over time after the carpet has been in place for an
unknown number of years. Vacuuming was an effective remediation technique. A residential vacuum reduced the resuspended
PM mass by an average of 44%. However, vacuuming increased the measured emission factors by a factor of 4 because of
the reduction in mass available for resuspension. Therefore, carpet history and maintenance must be known when applying
emission factors to exposure models
Experimental methods were developed to collect dust samples from carpet to determine PM translocation rates via tracking.
Field and laboratory experiments used these methods to quantify PM movement rates. Mass movement via tracking from field
tests varied between 2.7 and 24.1 (jug per in2-week. The rate was highly dependent on weather conditions and the estimated
number of traverses across the carpet, determined from the number of occupants. Laboratory tests showed between 40% and
80% of the mass on a shoe is transferred to carpet on the first step after loading, with subsequent steps transferring about
2%. These tests also showed that approximately 1% of the PM mass in the carpet was transferred to the shoe with each step.
-------
-------
Table of Contents
NOTICE iv
ABSTRACT v
TABLE OF CONTENTS vii
LIST OF FIGURES viii
LIST OF TABLES ix
ACRONYMS x
SYMBOLS xi
1.0 INTRODUCTION 1
1.1 Scope 1
1.2 Research Objectives 2
1.3 General Approach 2
1.4 Underlying Variables 2
2.0 EXPERIMENTAL METHODS 5
2.1 Resuspension Instrumentation and Procedures 5
2.2 Tracking Instrumentation and Procedures 12
3.0 Results and Discussion 19
3.1 Resuspension Experiments 19
3.2 Household Tracking Experiments 29
3.3 Laboratory Tracking Experiments 31
4.0 Quality Assurance 37
4.1 Gravimetric Mass 37
4.2 Climet Optical Particle Counters 38
4.3 Aerodynamic Particle Sizer 39
4.4 Temperature and Relative Humidity 40
4.5MetOnes 40
4.6 Fluorometry 41
5.0 Conclusions 43
6.0 References 45
APPENDIX A Characteristics of Houses Participating in Resuspension Tests 47
APPENDIX B Characteristics of Houses Participating in Walk-off Mat Tracking Tests 61
APPENDIX C Climet Translocation Graphs 65
APPENDIX D Household Tracking Data 73
-------
List of Figures
Figure 2-1 Collection of microvacuum sample to characterize the PM mass available for resuspension 8
Figure 2-2. Arrangement of sample collection equipment around 9 ft2 resuspension test area 9
Figure 2-3. Sample collection during resuspension experiment 10
Figure 2-4. Illustration of walk-off mat and microvac sample collection template 14
Figure 2-5. Example of uniform loading of shoe with Uranine from stepping firmly onto
cookie sheet with uniform deposits of Uranine dust 15
Figure 2-6. Walking path and example steps during laboratory tracking experiments 16
Figure 2-7. Shoe wash collection after a step 17
Figure 2-8. Carpet sample collection 17
Figure 2-9. Fluorometer well-plate template identifying the type and quantity of samples analyzed
per test 18
Figure 3-1. Climet data collected at House 1, Test 1, showing temporal and spatial dispersion
of the resuspended PM larger than 5 jjum 28
Figure 3-2. Transfer of Uranine from shoe to carpet per step during low loading tests 33
Figure 3-3. Transfer of Uranine from shoe to carpet per step during high loading tests 34
-------
List of Tables
Table 2-1. PM measurements during each resuspension test 7
Table 2-2. Procedure for collecting PM resuspension data from private residences 11
Table 3-1. Experimental conditions recorded during each test 20
Table 3-2. Summary of data collected at each house 21
Table 3-3. P-values for resuspension data collected at seven residences 24
Table 3-4. P-values for calculated emission factors 25
Table 3-5. Average emission factors per test at each house 25
Table 3-6. Average ratio of resuspended mass at 18 inches versus 36 inches separated into
pre-vacuumed and post-vacuumed conditions 26
Table 3-7. P-values for spatial-temporal differences in Climet concentrations 27
Table 3-8. Results from the spatial-temporal autoregressive analysis of the full and reduced datasets 29
Table 3-9. Results from the Tukey-Kramer tests identifying differences in mass loading per week 30
Table 3-10. PM tracking rate per week at each house sampled 30
Table 3-11. Uranine concentrations available for collection on the shoe, loaded onto the shoe,
and transferred to the carpet during each step 31
Table 3-12. Tinopal concentrations available for collection on the shoe, loaded onto the shoe,
and transferred to the carpet during each step 31
Table 3-13. Results from the spatial-temporal autoregressive analysis of the Uranine data 32
Table 3-14. Total mass and mass fraction of Tinopal loaded onto carpet that was transferred
to the shoe during Step 1 35
Table 4-1. QA/QC criteria for measurements collected 37
Table 4-2. Quality control measures to implement during testing 37
Table 4-3. Metric sample completeness statistics 38
Table 4-4. Gravimetric analysis precision and accuracy statistics 38
Table 4-5. Filter blank statistics 38
Table 4-6. Climet sample completeness statistics 38
Table 4-7. Climet particle number concentration and precision statistics 39
Table 4-8. Climet particle number concentration accuracy statistics 39
Table 4-9. Aerodynamic Particle Sizer sample completeness statistics 39
Table 4-10. QA/QC results for the HOBO H8 temperature and relative humidity data 40
Table 4-11. MetOne concentrations at 18 and 36 inches 40
Table 4-12. MetOne concentration correction factors 41
Table 4-13. Fluorometry sample completeness statistics 41
Table 4-14. Fluorometer calibration curve statistics for each pair of tests 41
Table 4-15. Fluorescent tracer mass found in quality control samples 41
-------
Acronyms
°C degrees Celsius
°F degrees Fahrenheit
APS Aerodynamic Particle Sizer
Avg average
cm centimeter
CMD count median diameter
DI deionized
EmFa emission factor
ft foot
GSD geometric standard deviation
HEPA high efficiency paniculate air
HVAC heating, ventilation, and air conditioning
IDL instrument detection limit
in inch
Ib pound
Lpm liter per minute
m meter
MDL minimum detection limit
min minutes
mL milliliter
mm millimeter
MQL minimum quantitation limit
mW milliwatt
N/A not applicable
nm nanometer
OPC optical particle counter
Pb lead
PVC polyvinyl chloride
QA/QC quality assurance/quality control
QAPP Quality Assurance Project Plan
PM paniculate matter
PJ3 relative humidity
RSD relative standard deviation
s second
Std Dev standard deviation
T temperature
a confidence level in statistical analyses
|juL microliter
(jum micrometer
-------
Symbols
A sample collection area
Ayac area sampled with microvacuum
Ac test carpet area
CAPS counts resuspended as measured by APS
CBk background count concentration as measured by instrument
Copc counts resuspended as measured by OPC
CR count concentration resuspended as measured by instrument
D dilution volume
V fluid volume
I fluorescence intensity
Ibkg fluorescence background intensity
MAPS mass resuspended as measured by APS
MAvafl mass available on carpet for resuspension
MR mass concentration resuspended as measured by instrument
MBk background mass concentration as measured by instrument
Mde mass deposited on surface
Mfflter mass on Teflo filter
MR mass resuspended as measured by instrument
Mrem mass remaining on shoe
Mshoe mass loaded on shoe
MURQ mass resuspended measured by URG sampler
MVas mass available as measured by microvacuum
QAPS sample flow of APS
t sample time
X instrument counting efficiency
£, OPC unit specific correction factor
T| transport efficiency in sample line
-------
-------
1.
Introduction
1.1 Scope
Each day adults and children are exposed to paniculate matter (PM) from the flooring and other horizontal surfaces in their
homes and offices. Potential health risks from inhalation and dermal exposure to this paniculate matter exist. This paniculate
matter can include metals, pesticides, or terrorist-based materials such as anthrax, ricin or radiologies.
The PM found on flooring stems primarily from dust tracked in by shoes and ambient particles that settle on the flooring and
other horizontal surfaces. Once indoors, PM translocates throughout the residence. The major mechanisms of translocation are
hypothesized to be resuspension and tracking.
Resuspension emission factors express the ratio of the mass resuspended to the mass available on the carpet. Emission factors
can also be expressed as particle count. The mass available does not necessarily equal the total mass in the carpet. Carpet
provides depth that may allow migration of the particles, especially those >10 jjum, downward toward the backing so that they
are not available for resuspension. Also, particles smaller than 1 jjum strongly adhered to the fibers are not easily dislodged and
may not be "available" for resuspension. Therefore, only particles on the upper portion of the carpet fibers may be available for
resuspension.
Previous research conducted by RTI for U.S. EPA developed and refined experimental methods to measure emission factors for
generic paniculate matter, fibers, and metals in the laboratory and field (Rodes and Thornburg, 2004; Thornburg and Rodes,
2004a, 2004b). Resuspended particle mass and size distribution data can be measured with either an integrated gravimetric-
electron microscope combination or with real-time data collected with aerodynamic particle sizing instrumentation. The
denominator of the emission factor, mass available, can also be measured by multiple methods. The key is identification of the
method that best represents the quantity available. Sample collection mechanisms also affect the mass available estimate. Rodes
and Thornburg, 2004, and Thornburg and Rodes, 2004a, using seeded, new carpet, showed that microvacuum sampling and
scanning electron microscopy analysis of carpet fibers yielded mass data statistically correlated with the quantity loaded onto
the carpet. Hence, both methods provided a good estimate of the mass available.
Previous research showed that walking on medium carpet resuspended almost 2% of the PM between 1 and 10 jjum available
on the carpet fibers, but up to 40% could be resuspended under the proper conditions. Carpet age, applied force, and relative
humidity determined the emission factors. Emission factors of less than 1% were obtained for new carpet at low humidity and
low applied force (e.g., walking). Emission factors between 20 and 40% were obtained for older carpet at high force (e.g.,
stomping) and low humidity. These tests also confirmed that the reservoir of particles available for resuspension is finite.
As a result, depletion will occur in lightly loaded carpets and emission factors will drop to zero after five minutes or less of
resuspension activity. Because these emission factors were calculated in laboratory experiments, the representativeness of these
emission factors for "real" carpet inside homes is unknown.
Whether the resuspended PM rises into the breathing zone to be an inhalation exposure risk and be to carried throughout the
residence via general air currents is unknown. Preliminary testing at three residences indicated a majonty of the mass traveled
less than 2 m laterally or vertically before settling to the floor (Rodes and Thornburg, 2004). Additional testing is required to
confirm these initial findings. We hypothesize that the large particle size distribution of the resuspended dust (MMAD ~ 5.5
mm, GSD ~ 2.0, Rodes and Thornburg, 2004) promoted rapid gravitational settling to the carpet and prevented dispersion of
the particles to sufficient heights to be convectively transported by the air currents within the homes. In addition, any mass that
did migrate away from the location of resuspension was insufficient to raise the PM concentration above background levels
because of the low emission factors and the large mixing volume. Additional research is needed to determine the amount of
translocation of resuspended PM to different areas of the residences.
An unknown fraction of the PM on carpet fibers also will adhere to shoe soles and be carried throughout the home as the
residents move from room to room. Whether or not the PM is dislodged from the shoe soles onto other surfaces needs to be
investigated. PM dislodged from shoe soles onto carpet in other rooms may then become available for resuspension by walking
or other activities. Remediation methods such as vacuuming may have a significant influence on translocation and exposure,
depending on the type and frequency of the activity. A literature review indicates that tracking as a translocation mechanism has
not been studied previously. The exploratory research into tracking will develop sampling methods, a better understanding of
parameters affecting tracking, and a range of tracking translocation rates (mass per unit of time).
-------
1.2 Research Objectives
The goal of this pilot effort was to begin developing a fundamental understanding of generic PM movement within a residence.
The relative importance of resuspension and tracking for PM translocation is unknown. This project provided resuspension
and tracking data to define the approach and scope of future research to determine primary routes of PM translocation within
buildings. Specific objectives of this project were:
1) To measure PM resuspension emission factors from walking on medium-pile carpet (of various ages and fiber lengths)
in private homes at ambient relative humidity (40%-60%) to confirm that laboratory-generated emission factors are
representative of real environments
2) To measure resuspended PM concentrations at two heights to obtain vertical gradient data
3) To monitor PM concentrations in areas/rooms surrounding the test area to quantify the amount of translocation from
resuspension
4) To initially characterize the magnitude of tracking as a method for PM translocation within a residence
5) To evaluate whether vacuuming can significantly reduce PM resuspension and translocation within a residence
This research provided improved input data for the inhalation and dermal exposure models used in risk assessments for metals,
allergens, biologies, and pesticides associated with particles.
1.3 General Approach
The key elements of the research approach to address the objectives given in Section 1.2 were:
1) Develop a project Quality Assurance Project Plan (QAPP) before beginning experimental work. (See "Resuspension of
Paniculate Matter on Flooring - Quality Assurance Project Plan for Basic Research Projects," EPA Order No. 4C-R179-
NALX, EPA/NHSRC under Contract No. QT-OH-04-00315 RTI Project No. 09188.)
2) Identify seven homes with medium-pile carpet of various ages for collecting resuspension data for calculation of
emission factors.
3) Obtain particle counters suitable for measuring vertical and horizontal concentration gradients of resuspended PM.
4) Select a vacuum cleaner representative of a homeowner quality unit rather than a HEPA unit that would be used in
remediation efforts.
5) Conduct controlled walking experiments pre- and post-vacuuming in each residence to measure emission factors and
particle translocation vertically and laterally.
6) Develop procedures and methods for tracking tests in homes and the laboratory.
a) Identify four homes for deployment of a 12-ft long piece of new, medium-pile carpet. Collect microvacuum
samples every week at 3-ft intervals to monitor for PM mass loading increases over time.
b) For laboratory tests, identify fluorescent tracers and develop shoe/carpet loading procedures to quantify the mass
dislodged per step.
1.4 Underlying Variables
Carpet pile height and particle adhesion are two variables that probably influence resuspension and tracking of particles from
carpets. However, these variables were not studied during this project because of the limited scope of work. Our hypotheses and
suppositions regarding these variables based on our carpet research experience are presented below because these concepts aid
in the interpretation of the data collected during these experiments.
1.4.1 Flooring Surfaces
RTI data from previous projects showed that as the pile height decreased, the level of resuspension and tracking from normal
walking decreased substantially (Rodes, 1998). Tracking decreased because of the reduction in contact area between the shoe
and the surface. No measurable resuspension was observed from dust on bare flooring. Resuspending dust from bare floors may
require substantial turbulence from either stomping or very fast walking to provide the energy to both release particles from
the surface and elevate them into the air sufficiently to add to the air concentration. Low-pile indoor-outdoor carpeting also
provided essentially immeasurable resuspension. Thus, the current work focused only on medium-pile carpeting (-70% of all
new carpeting sold), which has been shown by RTI and others to contribute significantly to resuspension (Rodes, 2001; Ferro et
al., 2004).
1.4.2 Particle Adhesion
A potentially important factor in understanding resuspension of particles from carpet fibers is adhesion. Adhesion of particles
to carpet fibers has been reported to be influenced by relatively humidity, controlled primarily by electrical charging of both the
-------
fibers and the particles. Very low humidity is routinely found to increase the charging of certain formulations of carpet fibers.
An additional adhesive force considered here is surface tension — particle-to-particle and particle-to-fiber. This latter type of
force bonding particles together or to the fibers is potentially important when the relative humidity exceeds -45%. Adhesion
literature states that at 45% RH sufficient water is present to increase the surface adhesion force by increasing the contact angle
between the particle and a second surface. Both of these types of adhesive forces work together to bond particles to surfaces.
Undoubtedly, resuspension occurs when sufficient energy is imparted to exceed the cumulative adhesive force levels.
RTI also has observed that as carpet ages or becomes significantly soiled (including coating of the fibers over time by
aerosolized grease from cooking), its ability to generate static charging appears to decrease substantially. The coating of grease
may increase the surface adhesion for larger particles. This substantial change in potential adhesion characteristics for particles
to fibers between new and old carpeting was addressed in the current research by considering either new, unsoiled carpeting, or
soiled carpeting that was at least 1 year (or substantially more) old as a binomial variable.
No efforts were made to measure either type of adhesion in these experiments, but temperature and relative humidity were
recorded during all tests to determine whether the influences from these surface forces could be estimated categorically.
Successful modeling of particle resuspension will require more detailed investigation of the relationships among these factors.
-------
-------
2.0
Experimental Methods
2.1 Resuspension Instrumentation and Procedures
The instrumentation and procedures selected for resuspension tests determined the particle aerodynamic diameter and
concentration (either mass or number) of the particles resuspended from the carpet and available for resuspension.
2.1.1 Instrumentation and Materials
The instrumentation used to collect the emission factor and translocation data from the residences is described below. A brief
summary of the resuspension method is provided also.
2.1.1.1 Carpet Pile Height - A dial micrometer was used to estimate the carpet pile height at multiple locations within the
residences.
2.1.1.2 Carpet Microvacuum Samples - The quantity of paniculate matter available for resuspension was estimated by
collecting two vacuum samples onto 47-mm Teflo filters. The vacuumed area was a 3 -in by 3 -in square. The vacuum used
was custom designed by RTI. A modified ASTM method (D5755-95: Standard Test Method for Microvacuum Sampling and
Indirect Analysis of Dust by Transmission Electron Microscopy for Asbestos Structure Number Concentrations) was followed
for the sample collection. The RTI modifications included using a brush on the vacuum nozzle to trap carpet fibers, reducing
the sampling velocity to 45 cm/s to minimize the aggressiveness of the procedure, manually removing obvious carpet fibers
collected on the filter, and using gravimetric analysis of the filter instead of liquid extraction for microscopic analysis.
2.1.1.3 MetOne Optical Particle Counters -The MetOne GT-521 is an optical particle counter that uses light scattering
principles; it operates a 30-mW laser diode that emits 785 nm of light with a nominal flow rate of 2.83 1 Lpm. However, in its
application here, the flow rate was reduced to ~1.9 Lpm to improve the counting efficiency of the unit. Two MetOne GT-521
units were deployed at the edge of the resuspension area. One unit sampled at 18 inches above the floor, and the other sampled
at 36 inches above the floor. The units collected data continuously in two channels: >1 (Jim and >5 jjum. One data point was
collected every 6 seconds.
2.1.1.4 Climet 4302 Optical Particle Counters - Six Climet 4302 optical particle counters were located at various
locations within the test room. All monitors were within 8 feet of the resuspension source. The instruments collected particle
concentration data every 60 seconds. Two channels measured the cumulative number >0.5 jjum and >5 jjum, respectively.
Instruments were operated according to manufacturer's instructions. All data were corrected for background aerosol levels.
Data were used for translocation hypothesis testing.
2.1.1.5 Aerodynamic Particle Sizer - The TSI Aerodynamic Particle Sizer (Model 3321) was used to obtain resuspended PM
mass and count size distribution data from 0.5 to 20 jjum. The APS was operated according to manufacturer's instructions .
Counting efficiency errors were corrected using the data from Peters and Leith (2003).
2.1.1.6 URG Mass Sampler - URG mass samplers (University Research Glassware, Chapel Hill, NC) collected PM10 samples
isokinetically at 20 Lpm. Two samplers were used per test: one at 18 in and the other at 36 in above the floor. Samples were
collected on 47-mm Teflo filters. Filter mass was determined gravimetrically.
2.1.1.7 Temperature and Relative Humidity - Temperature and relative humidity within the homes were measured with a
HOBO H8 Data Logger (Onset Computer Corporation, Bourne, MA). Temperature and relative humidity were recorded at 1-
minute intervals.
2.1.1.8 Gravimetric Analysis - Aerosol mass collected on filters was weighed in RTFs temperature and humidity controlled
(23 °C, 35% RH) weighing chamber on a Mettler Toledo MT2 balance with 0. 1 (jug resolution.
2.1.1.9 Statistical Analysis - All statistical analyses were performed using SAS version 9. 1 (SAS Inc., Gary, NC). The General
Linear Mixed model (Proc Mixed) was used to determine the significance of experimental parameters on the resuspended dust.
This model structure allowed statistical analysis of the random and fixed effects for their influence on the resuspended dust.
The statistical models evaluated for the dependent variables are listed below. The statistical models included the first order
independent variables and all second order interactions (listed below collectively as INTERACTIONS).
I = Age+ Vacuums Lengths Heights INTERACTIONS (2-1)
I Step )
log(MArall )=Age + Vacuum + Length + Height + INTERA CTIONS (2-2)
-------
MMD = Age + Vacuum + Length + INTERACTIONS (2-3)
CMD = Age + Vacuum + Length + INTERACTIONS (2-4)
log
'MURG/
/Step
= Age + Vacuum + Length + Height + INTERA CTIONS (2-5)
where:
• MVRQ = mass collected by URG filter
• Step = numbers of steps made during the five-minute test
• Age = carpet age
• Vacuum = carpet was vacuumed (e.g., remediated) before test ("Yes" = 1; "No" = -1)
• Length = carpet fiber length (pile height)
• Height = sampling height
* MAvail = mass available on carpet for resuspension
• MMD = mass median diameter
• CMD = count median diameter
• MTO /MA , - Step = emission factor
^JRG Avail *
One source of experimental variability that could be controlled during the statistical analysis was the number of steps per test.
Therefore, the resuspended mass was normalized by dividing the measured value by the number of steps.
The collected data for the normalized "resuspended mass" and "mass available" variables were not normally distributed
because of the small dataset (n<55) and the large variation in masses between houses. The "log" transformation was required to
normalize the data.
2.1.2 Tests Conducted
PM resuspension and translocation experiments were conducted in seven private residences with medium-pile carpet. RTI
recruited volunteers for the residential testing. Houses with pets (dogs, cats) were included in the study. Volunteer residences
had at least 36 ft2 of open space in a frequently used room to increase the probability of obtaining detectable resuspended PM
concentrations. The area and frequency of use specifications attempted to equalize the dust reservoir and carpet characteristics
within a house. Four different carpet sections, one for each experiment, were tested to avoid depletion of the dust reservoir. The
HVAC system in each home was deactivated and all windows were closed during the tests to remove these variables from the
statistical analysis. Natural and HVAC-induced ventilation could increase the between-house variability in the resuspended and
translocated concentrations. Experience gained during RTI Project 0886 showed that the HVAC system can supply sufficient
quantities of "clean" air that dilutes the resuspended PM and introduces experimental error. Open windows could introduce
either "clean" or "dirty" air that could confound the experimental data.
The home and test room characteristics for each residence were recorded (see Appendix A). Each home was assigned a unique
identification number linked to the address. Information recorded covered home occupants, carpet characteristics, cleaning
history, and interior and exterior surveys. This information provided a qualitative understanding of the carpet condition and PM
loading to relate measured emission factors to the corresponding laboratory-generated values.
Four resuspension tests per home were conducted. Two resuspension tests were conducted on the carpet "as is." The carpet
was cleaned with a standard residential vacuum prior to the remaining two resuspension tests to test the efficacy of simple
remediation efforts. The vacuum was a 13-amp Mach 2.1 Hoover (Model # U5330-900) upright vacuum with beater bar. The
entire 9 ft2 area was vacuumed two times from each direction (left to right, right to left, top to bottom, bottom to top). The
entire vacuum was cleaned thoroughly prior to the experiments. A new Hoover Type Y Allergen® vacuum bag (99.98% filtration
efficiency) was installed for each experiment.
-------
A single volunteer walked randomly within the 3-ft by 3-ft carpet area for five minutes during an experiment. The volunteer
was a 71 inch, 165-lb male with size 12 shoes (120 in2). A pedometer worn by the volunteer determined the number of steps
taken during the experiment. The subject walked so that a constant foot pressure (energy) was imparted to the carpet since
previous work showed emission factors varied with energy level. There is no way to ensure each step applied the same foot
pressure. Any variability in foot pressure contributing to experimental error was minimized by the large number of steps
possible in a five-minute period (~ 250 steps). Also, the highest possible foot pressure imparted by walking still should be much
less than the foot pressure generated by "stomping."
Microvacuum samples characterized the quantity of PM available for resuspension (Table 2-1, Figure 2-1). Resuspended PM
concentrations and size distributions were measured with the time-of-flight instrument (APS), optical particle counters (Climet
OPCs and MetOne OPCs), and gravimetric samplers (URG units) listed in Table 2-1 and shown in Figures 2-2 and 2-3. APS
and Climet sampling artifacts were accounted for when calculating PM emission factors. Large particles suffered deposition
losses within the sampling lines leading to both instruments. These losses were characterized for both instruments during RTI
Project 0886. The same sampling lines were used in this research. In addition, the APS does not count every particle that enters
the instrument. The correction factor developed by Peters and Leith (2003) was used.
Table 2-1. PM measurements during each resuspension test.
Background PM
concentration and size
distribution
Resuspended PM
concentration and size
distribution
APSxl
Climet x 6
MetOne x 2
APSx2
Climet x 6
MetOne x 2
URGx2
10 files over 10
minutes
10 files over 10
minutes
10 files over 10
minutes
10 files over 10
minutes
10 files over 10
minutes
10 files over 10
minutes
1 filter over 10
minutes
Calculation of total and size dependent
mass and number concentration using
Excel® correcting for instrumentation
artifacts
Gravimetric analysis for mass and
SEM analysis for size distribution
PM reservoir strength:
concentration and size
distribution
Microvac
2 filters per test area
Gravimetric analysis for mass
-------
Figure 2-1. Collection of microvacuum samples to characterize the PM mass available for resuspension. Template area
is 9 in2. Fourteen passes were made across the template. Samples were collected prior to walking in test area.
-------
Figure 2-2. Arrangement of sample collection equipment around 9 ft2 resuspension test area. Note that instruments
were located at 18 inches and 36 inches above the floor. URG filter samplers were not installed. Also, note that
Climet was located at entry into room.
-------
Figure 2-3. Sample collection during resuspension experiment.
A preliminary experimental protocol is presented in Table 2-2. The 9 ft2 test area was outlined in masking tape. All sampling
instruments were positioned around the edge of the template in the direction of most likely air movement. The sample inlet
height for the four Climets set up away from the test area boundary was 36 in. Presence of furniture, walls, or other obstructions
to airflow sometimes required Climet inlets to be as low as 32 in or as high as 40 in, depending on professional judgment. These
changes were noted on the test datasheets.
Once all equipment was ready, movement within the house ceased to allow the PM concentration to return to quiescent levels
prior to collection of the background PM samples with the APS, Climets, and MetOnes. All instrumentation within the house
was programmed to start after 15 minutes of quiescent conditions. After collection of 10 minutes of background data at each
sampling height, the PM mass samplers were started and the test volunteer began the scripted activity within the template. The
volunteer wore a particle-free clean room suit to prevent the "personal cloud" from biasing the data. The volunteer walked for 5
minutes, then waited for another 10 minutes to allow the PM concentration to return to background levels. After the 10 minutes
allocated for the PM resuspension expired, all instrumentation was stopped, mass samples collected, and data files saved.
Ancillary indoor temperature and relative humidity data were collected during each test. One experiment required about 60
minutes to prepare for the test and about 60 minutes to collect the data.
-------
Table 2-2. Procedure for collecting PM resuspension data from private residences.
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Deploy HOBO temperature and relative humidity monitor.
Outline 3 ft by 3 ft resuspension area.
Vacuum test area if remediation evaluation test is being conducted.
Collect two microvac samples in two separate 9 in2 sections within test area.
Set up APS, Climet, MetOne, and URG units at the boundary of the area. Sampler inlets,
one of each type, should be at 18 inches and 36 inches above the floor. APS only at 24
inches.
Set up remaining four Climets no more than 8 feet from area boundary and with sample
inlets at about 36 inches above the floor or as determined by professional judgment
accounting for obstacles and expected air movement.
Allow PM concentrations to return to background levels.
Start background sample collection with all APS and Climets. Collect data for 5 minutes.
Reset all instrumentation for resuspension test.
Allow PM concentrations to return to background levels.
Start all APS, Climet, and URG samplers.
Begin walking within test area to resuspend PM. Walk for 5 minutes.
Stop movement and allow PM to settle. Wait for 10 minutes.
Stop all sample and data collection instrumentation.
Retrieve URG filters and save APS and Climet data files.
Retrieve HOBO temperature and relative humidity monitor.
TOTAL
5 min
5 min
5 min
10 min
10 min
10 min
10 min
5 min
10 min
5-10 min
1 min
5 min
10 min
1 min
15 min
5 min
120 min
2.1.3 Data Reduction
Emission factors were calculated in multiple ways. Resuspended mass concentrations were measured with the URG and APS
instruments. Corresponding resuspended count data were provided by the APS, Climet, and MetOne units. The mass available
for resuspension was calculated from microvacuum samples.
2.1.3.1 PM Resuspended by APS - The mass resuspended as measured by APS was calculated using Eq. 2-6. The mass
concentration in each size bin (CR) corrected for the background resuspended particle concentration (CB/cg) and transport
efficiency into the APS (T|) were summed to yield a total concentration. This concentration was multiplied by the flow through
the chamber (QAPS =5.0 Lpm), the total time particles were resuspended (/), and a dimensionless correction factor for APS
counting efficiency (x=2). Total time particles were resuspended was not constant across all tests. Depletion occurred within
three minutes during some experiments because of low mass load in the carpet. Other experiments had airborne concentrations
greater than background that lingered for up to five minutes past the end of the resuspension activity because of high mass load
in the carpet and low air exchange rate. These factors were accounted for in the data analysis.
MAPS = IZ &
Q
APS
X
(2-6)
APS also provided resuspended PM number concentration data. The total counts resuspended (CAPS) were calculated using an
equation similar to Eq. 2-6.
CAPS =
X
(2-7)
-------
2.1.3.2 PM Re-suspended by Gravimetric Data - Mass resuspended as measured by the URG samplers was equal to the
gravimetric mass collected on the filters, Mfilter. Mass resuspended was not corrected for background because concentrations
either were not statistically different from zero or contributed less than 5% of the total mass collected (as measured by APS).
MVRG = Mmter (2-8)
2.1.3.3 PM Resuspended by Optical Particle Counters - Resuspended PM counts (Copc) were calculated from the number
concentrations measured by the Climets and MetOnes (CR) and their respective sample flow (Qopc). Measured concentrations
were corrected for background PM concentration (CBk), sample line transport efficiency (T|), and a unit-specific correction
factor (|). Collocated sampling (as part of the quality control procedures outlined in the QAPP) for the Climets and MetOnes
indicated concentrations measured by individual units differed by up to a factor of 10. Reference instruments for calculation of
the correction factors were Climet #958304 and MetOne #01.
n y t y ? (2-9~)
^-OPC ^3 ^ ^
2.1.3.4 PM Available by Microvac - Mass available by microvac was calculated from the gravimetric mass collected on the
filters, Mfilter, divided by the total area vacuumed
(Avac = 9 in2) and multiplied by the area of the test carpet piece (Ac = 1296 in2).
(2-10)
2.2 Tracking Instrumentation and Procedures
Two experimental approaches determined whether PM moves through a residence via tracking. The first approach used real
homes to bracket the range of expected translocation rates via tracking. These tests were followed by laboratory experiments
that identified the salient parameters influencing particle translocation via tracking.
2.2.1 Instrumentation and Materials
The residential and laboratory tracking tests required different experimental methods. Whether a method was used for a
residential or laboratory experiment is clearly delineated in the subsection heading.
2.2.1.1 Carpet Microvacuum Samples (Residential) - The quantity of paniculate matter available for tracking on the test
carpet placed within each residence was determined by collecting microvacuum samples. PM available for tracking was
collected on 47-mm Teflo filters. The vacuumed area was a 3-in by 3-in square. The locations sampled on the carpet each week
are described in Section 2.2.2. The vacuum used was custom designed by RTI. A modified ASTM method (D5755-95: Standard
Test Method for Microvacuum Sampling and Indirect Analysis of Dust by Transmission Electron Microscopy for Asbestos
Structure Number Concentrations) was followed for the sample collection. The RTI modifications included using a brush on the
vacuum nozzle to trap carpet fibers, reducing the sampling velocity to 45 cm/s to minimize the aggressiveness of the procedure,
manually removing obvious carpet fibers collected on the filter, and using gravimetric analysis of the filter instead of liquid
extraction for microscopic analysis.
2.2.1.2 Gravimetric Analysis (Residential) - Aerosol mass collected on filters was weighed in RTFs temperature and humidity
controlled (23 °C, 35% RH) weighing chamber on a Mettler Toledo MT2 balance with 0.1 jjug resolution.
2.2.1.3 Fluorescent Particles (Laboratory) - The test dust used was amorphous silica (Syloid, W.R. Grace Co., Baltimore,
MD) tagged with Uranine (Fisher Scientific, Fair Lawn, NJ) or Tinopal (Ciba-Geigy, Greensboro, NC). Uranine and Tinopal are
tracers that fluoresce at different wavelengths. Uranine particles determined the quantity of dust dislodged from the shoe onto
the carpet. Tinopal particles determined the amount of paniculate matter transferred from the carpet to the shoe sole. This dust
had a median diameter of ~2.5 jjum and a standard deviation of -1.1 |jjn when aerosolized or applied to the surface. This size
distribution had 10% of the mass contained in 10 jjum particles. An established RTI procedure was used to generate the dust.
2.2.1.4 Fluorometer (Laboratory) - A GENios TECAN fluorometer using XFLUOR v4.51 software measured the intensity of
the fluorescence emitted from the samples. This system analyzed 96 100 jjuL samples for both Uranine and Tinopal fluorescence
in one batch. The instrument software was automatically programmed to change the excitation and emission wavelengths
to the desired wavelength. The Uranine excitation wavelength was 485 nm and the emission wavelength was 535 nm. The
Tinopal excitation wavelength was 360 nm and the emission wavelength was 465 nm. Dilution of samples was sometimes
-------
necessary to stay below the upper detection limit of the fluorometer. Progressive dilutions of 1/10, 1/20, and 1/30 were used
until a valid fluorometer reading was obtained. Calibration curves relating the fluorescence intensity to the fluorescent particle
mass concentration were generated for each batch. All measurements were corrected for residual fluorescence from the sample
collection fluid (0.01 N NaOH solution) and the well-plate. The Uranine and Tinopal detection limits were
0.2 |jig/mL and 0.7 |Jig/mL, respectively.
2.2.1.5 Statistical Analysis (Residential and Laboratory) - The nonlinear, spatial-temporal autoregressive analysis necessary
to analyze the tracking data required use of the Proc Mixed procedure in SAS version 9.1 (SAS Inc., Gary, NC). The spatial-
temporal analysis accounted for the variability in conditions influencing the loading on the carpet sections or shoe over time.
The autoregressive structure controlled for the variability in experimental conditions found between houses or laboratory test
conditions so that comparisons between these independent variables could be made. The Tukey-Kramer least squares analysis
procedure identified whether differences between houses existed.
2.2.2 Residence Experiments
Farfel et al. (2001) showed significant increases in lead (Pb) mass within 3 ft of the home main entry after 3 weeks of deploying
a walk-off mat to collect lead particles. This work extended the size of the walk-off mats to cover a larger area of the residence
and extended the sampling period to provide estimates of tracking translocation rates. Walk-off mats were installed inside
four volunteers' residences. One residence (HI) participated in the resuspension testing, and the other three participated only
in the tracking study. Walk-off mats were new, medium-pile carpet about 30 in wide and at least 144 in long. The walk-off
mats were located immediately inside the main entry of the home. Home selection criteria for installation of the walk-off mats
depended on: 1) resident participation, 2) space availability, 3) number of people living in the home, 4) absence of dogs inside
the home, and 5) main entry being not through the garage. Volunteers were instructed not to vacuum or clean the walk-off mat.
Characteristics of these houses are presented in Appendix B.
Walk-off mats were deployed inside the home for four weeks. Figure 2-4 summarizes the details of the walk-off mat data
collection. Microcvacuum samples were collected in each of four segments of the carpet immediately after deployment and at
the end of each week. All microvacuum samples were collected in situ. The four segments of the walk-off mat corresponded
to specified distances, 3-ft increments, from the main entry into the home. Microvacuum samples were collected using a 30-in
by 36-in metal template consisting often 9-in2 sample points. The corners of the template had "feet" to raise the grid above
the carpet surface to avoid cross-contamination between segments. Ten sample points allowed collection of duplicate samples
per week, one sample per row. Each sample grid per walk-off mat segment was used only once per residence. The grids were
spaced to match the normal adult step width of ~ 18 in. Two background samples were collected from each carpet segment.
The PM mass collected each week, in each segment, required spatial-temporal statistical analysis. The statistical model
accounted for the number of people inside the homes, segment number, and week number. Due to the small sample of four
houses and data collection over five weeks, the paniculate matter masses collected were not normally distributed. Several
transformations of the data were considered. The "log" transformation provided the appropriate correction and parametric tests
were performed under normal assumptions.
2.2.3 Laboratory Experiments
2.2.3.1 Sample Collection and Analysis - Additional tracking experiments were conducted in the laboratory. These tests
identified salient parameters affecting translocation and quantified the range of particle translocation rates for comparison with
the field tests. Because little is known about the mechanism of particle adhesion/removal on surfaces (e.g., shoes, carpet fibers),
this exploratory effort required the development of sampling methods and judicial selection of test conditions.
One challenge was determining how to uniformly load a known quantity of the fluorescent tracer to a surface. The best
approach was to suspend a known mass of tracer in 25 mL of DI water, then apply 2 mL aliquots to the surface by syringe.
The aliquots were dispersed evenly in a 1.5-in diameter circle outlined by a piece of PVC pipe. As the DI water evaporated,
the fluorescent particles were deposited uniformly across the 1.5-in circle. Eleven aliquots were applied to each surface. Eight
aliquots were applied into an area matching the shoe sole area. The other three aliquots were applied outside the shoe sole area
to quantify the amount deposited.
Uranine and Tinopal particles were applied as follows. Uranine tracer was applied indirectly to the shoe sole. The above
procedure was used to load Uranine onto a 9-in by 13-in cookie sheet (Figure 2-5). Once the Uranine solution dried, the shoe
was evenly pressed into the cookies sheet to evenly coat the sole surface (Figure 2-5). The shoe surface was marked with eight
1.5-in diameter circles to correspond to the sample locations. Following this step of the procedure, the three wash samples
outside the footprint were collected to quantify the amount deposited. [Awash sample was collected with 10 mL of 0.01 N
NaOH solution (twice), a disposable pipette, and a piece of 1.5-in diameter PVC pipe. The PVC pipe was held firmly against
the flat shoe sole by the technician. Using the other hand, the same technician poured 10 mL of the NaOH solution into the PVC
pipe, mixed the solution with the pipette, and then pipetted the solution back into the vial. The area on the shoe sole was washed
twice with 10-mL aliquots of the NaOH solution.] Also, the eight wash samples within the footprint were collected to calculate
-------
the amount transferred to the shoe via a mass balance. Tinopal particle solution was applied to the carpet (Step 1 location)
using the above procedure and allowed to dry. Tinopal particles were transferred from the carpet to the shoe during a normal
step on the first location only. Again, the three samples outside the footprint quantified the amount deposited and the eight
samples within the footprints were used to calculate the amount transferred to the shoe via a mass balance. As discussed below,
carpet samples were removed to extract the Tinopal for quantification. Subsequent steps transferred Tinopal from the shoe sole
to the carpet.
The sample collection procedure for the tracking tests is described below. The test volunteer was a 6-ft, 170-lb male wearing
size 10 shoes (160 in2) with flat, smooth soles. The volunteer stepped onto the Uranine-coated cookie sheet to evenly load the
shoe sole of his right foot (Figure 2-5). Then, he stepped onto the Step 1 template (Figure 2-6). After the step, the volunteer
stopped movement, lifted his foot and a wash sample was collected from the specified location on the shoe sole (Figure 2-7).
Steps 2-6 followed this same procedure. After the last step, the shoe was removed and the remaining two shoe wash samples
were collected.
Figure 2-4. Illustration of walk-off mat and microvac sample collection template. Sample collection template is
shown in Segment #3. Numbers within boxes correspond to order of sample collection:
b = background, # = week number.
30 in
Row
>144 in
Segment 4
Segment 2
Segment 1
9-in2 grid for
"" microvac
Following the steps, carpet squares were removed to extract the fluorescent particles for measurement of the amount remaining
in the carpet (Tinopal: Step 1 only) or transferred to the carpet (Uranine: all steps; Tinopal: Steps 2-6). Squares approximately
1 inch by 1 inch were cut from the carpet (Figure 2-8). The squares corresponded to the marked sample collection point on the
carpet. The squares were placed in a disposable beaker filled with 40 mL of 0.01 N NaOH solution and sonicated for
20 minutes. Then, the carpet squares were removed from the beaker. The sonication extracted the fluorescent particles from the
carpet fibers and suspended them in the fluid.
Once samples for two tests were collected (96 samples: 48 per test), the samples were transferred to the well-plate for
fluorometry. Each sample was mixed for five seconds to resuspend the fluorescent particles in solution immediately before
100 (juL was pipetted into the well-plate (Figure 2-9). If necessary, 1/10, 1/20, or 1/30 dilutions of a sample were prepared to
obtain a measurement below the maximum detection limit of the fluorometer.
-------
2.2.3.2 Data Reduction - The fluorescent particle mass loadings (|jig/in2) on the shoe and carpet following each step were
calculated directly from the fluorescence intensity data from the fluorometer (I) corrected for the background fluorescence (Ibk),
any dilutions (D), fluid volume (V), and sample collection area (A).
Mass = •
(l-Ib!tg)xDxV
(2-11)
Mass loaded onto the shoe (|jig/in2) was calculated via a mass balance between the average quantities deposited minus the
average amount remaining following the step.
-3
M
dep, i
(2-12)
Figure 2-5. Example of uniform loading of shoe with Uranine from stepping firmly onto cookie sheet with uniform
deposits of Uranine dust. Left picture shows clean shoe and cookie sheet prior to step. Right picture shows loading of
shoe following step.
-------
StepS
Step 6 (not shown)
Figure 2-6. Walking path and example steps during laboratory tracking experiments. 1) Target walking path. Shoe
template shows where each step should occur. Circles indicate where Tinopal loaded onto carpet (Step 1 only) and
where carpet and shoe wash samples were collected after each step. 2) First step onto carpet. Notice alignment of
shoe and the template to identify where carpet samples should be collected and to ensure Tinopal is loaded onto
shoe. 3) Normal step off carpet.
-------
Figure 2-7. Shoe wash collection after a step. Each circle on shoe corresponds to a sample to be collected after
the appropriate step. Only one location sampled per step. 10 ml of 0.01 N NaOH was poured into PVC pipe firmly
held against shoe sole to prevent leaks. Solution mixed and pipetted back into sample vial. Procedure repeated with
another 10 ml of fluid. 20 ml of fluid extracts > 99% of fluorescent particles off surface. All fluid combined for
fluorometric analysis.
The initial loading of Tinopal
and depositing of Uraine
occur on the first step.
A 1-in cut of carpet was used for
particle extraction and fluorometry.
Figure 2-8. Carpet sample collection. 1) Uranine deposited on carpet following Step 1. 2) Removal of 1-in2 carpet
sample for sonic extraction then fluorometry. 3) Step 1 carpet after removal of all samples. 4) Number of wash,
carpet, and QA/QC samples collected per experiment.
-------
a)
a.
-a
a>
N
03
C
03
Q.
E
03
03
^
CT
O)
Q.
00
a)
T3
a)
03
Q.
E
a>
a>
03
"o.
a>
"CD
E
o
^_
o
CNI
CN
CT)
oo
CD
LO
co
CN
-* I
E O
mz
.* i
^2
mz
o
oo
o
CN
CO
o
H
o
CD
O
^
CO
0
o
o
CD
CO
o
o
CN
o
oo
"co
1-
n
11
_ o
f v £Q
D°
o
^"
o
CD
"co
o
=>
o
CN
O
OO
"co
0
o
o
o
"co
0
22
^
CD
o
.c
CO
~§
CD
o
.c
CO
O)
00
CO
o
O)
00
p
CO
o
O)
00
^
J_
CO
o
2
00
CD
o
.c
CO
1
o W
o W
p
"O CD
8 w
CO
^
1
D
CO
o
fxg
1^
1
•a
CO
o
i
1
•a
CO
o
oo
CD
CO
CO
a.
CD
CO
LO
CD
CO
CO
Q.
CO
CN
CD
CO
CO
a.
CD
CO
CN
>
CO
a.
CD
CO
oo
CD
CO
CN
a.
CD
CO
CD
CO
CN
a.
CD
CO
$^
CN „_
Q.CO
CD <
CO
•
^~ CD
CO
LO
CD
CO
CO
CD
CO
CD
CO
i
CO
oo
CD
CO
CD
a.
CD
CO
LO
CD
CO
CD
CO
^
CD
CO
CD
Q.
CD
CO
LO
>
*R
CD
CO
oo
CD
CO
LO
a.
CD
CO
LO
CD
CO
LO
a.
CD
CO
p
LO
Q. C:
CD <
CO
4
^" CD
<55W
LO
CD
CO
J
CO
^
CD
CO
a.
CD
CO
1
I
CO
1
ID
"D Q_
CO M
O
— 1
i
i LO
"a Q.
CO M
o
1
^
"D CL
CO |j_
O
,
? CO
"a CL
CO 1 1
o "-
^CM
"S °-
CO u_
|
ID
"D CL
CO M
O
it
O
i LO
"O CL
CO LL
O i
-1 D
ib
"O n
CO [i
o
1
^ nn
CO |j_
O
i
^CN
"D CL
CO M
O
_l
^
^S:
o
P2
^
CD
O
.C
CO
^
CD
O
.C
CO
O)
00
CO
o
O)
00
p
CO
o
O)
00
^
J_
CO
O
2
00
CD
O
.C
CO
1 co
§ CO
•a CD
% co
CO
^
1
ID
•a
CO
o
fxg
i^
i
•a
CO
o
i
•a
CO
o
oo
CD
CO
CO
a.
CD
CO
LO
CD
CO
CO
Q.
CO
CN
CD
CO
CO
a.
CD
CO
CN
CO
Q.
CD
CO
oo
CD
CO
CN
a.
CD
CO
LO
CD
CO
CN
Q.
CD
CO
1
Q_CO
1*
CN „
Q.CO
CD <
CO
•
^" CD
CO
LO
CD
CO
1
CO
CD
CO
1
CD
CO
i
CO
oo
CD
CO
CD
a.
CD
CO
LO
CD
CO
CO
CD
CO
^
CD
CO
CD
a.
CD
CO
LO
CD
CO
oo
CD
CO
LO
a.
CD
CO
LO
CD
CO
LO
a.
CD
CO
p
LO .
Q. fc
CD <
CO
4
^~ CD
CO
LO
CD
CO
1L
£
CO
^
CD
CO
1
a.
CD
CO
I
CO
* T:
•a .2 g .2
§co > co
^ UJ ^
CD CD
a. a.
E E
CO CO
CO CO
TD
CD
.c
C/)
CD
o
o
o
o
O)
,~
Uranine load
E
o
O
T5
e-
8
o
O)
^
CO
_o
"co
Q.
O
P
E
o
O
T5
e-
8
o
O)
^
CO
_o
"co
Q.
O
P
CD
O
CO
CD
O
.c
CO
E
o
4=
CD
"o.
E
CO
CO
.c
CO
1
Q.
CD
To
O)
1
TJ
Q.
"o
O)
CO
o
Cookie sheel
Q.
£
To
M—
o
CO
CO
|
^~
CD
Q.
E
CO
CO
o
o
CD
O
.c
CO
CD
Q.
CN
c:
_o
To
E
CO
i
CD
O
.C
CO
Background
•a
(^
backgroui
.
CD
CD
.C
CO
CD
O
O
O
CD
a.
E
CO
CO
CD
O
O
.Q
O
To
I
CO
"c
O
O
"a;
Q.
CO
O
Background
CO
Q.
To
CO
O)
1
CO
CD
Q.
CO
CO
.c
CO
CO
CD
O
.C
CO
i
T3
CO
O
CD
CO
p
co
o
CO
CD
CO
p
co
o
i
CD
CO
CL
LJ_
i
T3
CO
O
tep#-Se#
CO
O)
m
CD
O
.C
CO
00
CO
o
tt
CO
O)
00
CO
o
CD
o
.c
CO
O
CO
co
O
00
O
LU
I
O
CO
CO
a.
o
c
p
co
o
JO
00
O
o
o
T~
CO
o
D
o
CN
O
OO
CO
o
D
o
o
CD
CO
O
D
o
CD
O
^~
CO
O
D
o
oo
o
CM
CO
O
D
o
o
0
CO
p1
o
CN
0
oo
CO
p1
o
0
CD
CO
o
CD
0
CO
p1
o
oo
0
CN
CO
p1
T
o
CO
•7
CO
00
-------
3.
Results and Discussion
3.1 Resuspension Experiments
3.1.1 General Findings
All experiments successfully provided data to achieve the research objectives, although several experiments lacked one or
more types of data. General experimental conditions for all tests are presented in Table 3-1. Table 3-2 summarizes the data
collected at each house. As described in Section 4.0, data quality objectives were achieved for all metrics except the MetOnes.
Reduction of the MetOne data showed the expected vertical concentration gradients were not present. Therefore, MetOne data
are not included in Table 3-2 and were not included in any statistical analyses. More information on the MetOne data quality is
presented in Section 4.5.
All equipment was not available for testing at each house. Experiments at two houses were conducted before the Climets were
received as Government Furnished Equipment. Study of PM translocation via resuspension was not hindered by the lack of the
Climets for experiments at these two houses. Climet data from five houses were sufficient for achieving this study objective.
The MetOnes were not available for use at three houses. The MetOnes' availability for this project was a matter of convenience.
The instruments were purchased for another project and were available for the resuspension studies only when not needed for
their primary project. However, the poor quality data obtained limited their value for achieving the study objectives.
Table 3-3 contains the statistical analysis of the data examined for influence of different carpet ages, fiber lengths, vacuuming,
and sampling height on the resuspended PM concentration and size distribution, and the quantity of PM available for
resuspension. The statistical models used were listed in Eqs. 2-1 thru 2-4. The quantity of mass resuspended followed the
expected trends. The differences in carpet age, quality, and maintenance between houses definitely had an impact. Older homes
with poor maintenance released more PM during the resuspension tests than homes with new, clean carpet (Table 3-3, Appendix
A). Vacuuming the carpet also decreased the amount of mass resuspended by an average of 44%, independent of sampling
height and mass available for resuspension. Only if the carpet was new and already extremely clean did vacuuming not have
an influence. There was a significant difference in the PM mass available for resuspension (as measured by the microvacuum)
between houses, with the houses with the newest, most frequently cleaned carpet having the lowest values. The PM mass
resuspended also followed the expected gradient as a function of height from the floor. More details on the vertical gradient
in resuspended PM concentrations are discussed in Section 3.1.3. Statistical analysis of the resuspended PM size distribution
gave conflicting results. The mass median diameter of the resuspended PM was not statistically associated with the carpet age,
carpet pile height, sample height, or whether or not the test area had been vacuumed (remediated). However, the count median
diameter was statistically influenced by the carpet age.
-------
Table 3-1. Experimental conditions recorded during each test. Pre-vacuum indicates test occurred prior to simple
remediation (cleaning). Post-vacuum indicates the test area was remediated via vacuuming with a regular household
unit.
j* S ' ' ' ' '
JffiSitfl;
1
^)
0
A
c
f.
H
••#&':
*tf?J
i
2
o
J
4
1
2
3
4
1
2
o
J
4
1
2
3
4
1
2
3
4
1
2
o
J
4
1
2
3
4
™"&ittvv™
Walk, pre-vacuum
Walk, post-vacuum
Walk, pre-vacuum
Walk, post-vacuum
Walk, pre-vacuum
Walk, pre-vacuum
Walk, post-vacuum
Walk, post-vacuum
Walk, pre-vacuum
Walk, post-vacuum
Walk, pre-vacuum
Walk, post-vacuum
Walk, pre-vacuum
Walk, pre-vacuum
Walk, post-vacuum
Walk, post-vacuum
Walk, pre-vacuum
Walk, pre-vacuum
Walk, post-vacuum
Walk, post-vacuum
Walk, pre-vacuum
Walk, pre-vacuum
Walk, post-vacuum
Walk, post-vacuum
Walk, pre-vacuum
Walk, pre-vacuum
Walk, post-vacuum
Walk, post-vacuum
,\'"{. rff
-;$&,WJ&
54
52
51
52
56
57
55
55
59
58
57
57
60
59
60
59
45
45
46
45
41
40
41
40
46
47
49
49
' •$$*&;
,'-f*f)-«*
73.2
75.2
77.7
78.0
72.1
72.3
72.4
72.6
74.5
75.2
76.8
77.3
77.3
78.0
79.5
80.8
68.1
68.1
68.3
68.5
73.3
73.8
74.1
74.1
74.4
75.9
78.7
79.4
'^t'itelJS;
350
400
325
und
250
275
240
280
300
280
260
340
300
260
260
260
275
300
290
310
320
340
350
330
280
295
300
240
&<•' :^:f;^l!!!!!!!!!!i;r'
f '-^^, ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;,.
Pedometer not reset before test.
No Climets or MetOnes.
Climets not received as GFE.
MetOnes being used on primary project
18-in URG mass data invalid. Pump shut off early.
No Climets or MetOnes.
Climets not received as GFE.
MetOnes being used on primary project
No MetOnes.
MetOnes being used on primary project.
-------
rn
O
-------
1
0
tn
till! ilKII KllilKllilKllil
*g'
^M
W
Wt!.!,.!,!.!,!.!,!.!,!.!,!.!,!.!/S,
lil Kl i! Kl ! ! Kl !i! Kl !» Kl »
'/f "' V<^
ill
^fil
r-g!|
iSi
; <• / i<-
.:,..:,..:,..J;;g
v-'- "si
tt" " '
jk
; |*|
ttS • '
,' v ' ; -S--1
"VVVA,
-i'lll
"' ''' ''' ''' ''' ''' ''' ''' ''' ''«£
. . . . . . . . . S>
; ; ; , T , T , T «.
p,,««|
- ' ' '..v'.-US.
I/
T
o
i/
c
r
^
r
V
t-
r
-1-
O
0
90
— H
V
90
O
— H
O
5
£
>,
i
3
^
i
*•'
H
0
i
5
H
H
5
-4
S
H
-j
OS
o
0
vo
-'
V
vo
oo
o
o
3
J
u
if
3
>
H
oo
0
T,
0
vq
OS
CN
•*'
•*
OS
oo >n
o o
0 0
90 VO
O O
V V
90 VO
O O
o o
o o
S
3
c5
s,
^
S
fS
c
0
I/
c
r
T
o
w
^
c
tN
o
0
90
-'
V
90
0
O
o
i
\
',
**
f>
3
n
5
3
H
t
1
t
H
~j
^t"
O
0
vo
oo
o
V
vo
t-
o
o
3
3
->
u
1
>
1
•*'
0
CN
0
vq
t-
Tt
CN
^H
ts'
•—i OO
o o
0 0
90 VO
o o
V V
90 VO
oo oo
o o
o o
o o
S
3
e5
1
IS1
a
^
S
TT
0
CN
r-
vq
OS
OS
CN
-------
oo
o
o
o
C3
o
OS
f-
r-
O
C3
O
^H
wi
I
OS
OO
00
r-'
T,
OO
p
t-'
00
f-
o
OO
OO
p
OO
OS
OS
•*'
OO
-------
&
.»
*>.
** •
#..
0
O
OO
oo
ts
0
o
•n
•n
ts
ts
90
vo
Walk, pre-vacuu
0
o
0
o
p
CN
O
OO
ts
CN
90
vo
vacu
alk, po
o
o
0
o
oo
in'
ts
CN
90
vo
alk, pre-vacuum
0
o
OS
0
o
w ro g
-
CO E >=
t^ ^ ro
03 ,_ ^
£ O> LO
« «q
^ E o
*" .- ii
- ^ cc
!= Q_ 03
(DO)
CO T3
8
Q. t
— 03 O
O - -^
O _V .<£
03
03
03
>
Q-
(Y)
OO
-------
3.1.2 Resuspension Emission Factors
Statistical analysis of the calculated emission factors using Eq. 2-5 identified carpet age, vacuuming, and the carpet age-fiber
length interaction as significant variables (Table 3-4). Previous field and laboratory tests showed emission factors typically were
between 0.01 and 0.1 but could be as high as 0.4 or as low as 0.001 under the proper conditions (Rodes and Thornburg, 2004;
Thornburg and Rodes, 2004a; Thornburg and Rodes, 2004b). These experiments showed that emission factors from a wide
variety of houses varied between 0.001 and 0.064 (Table 3-5), and were within the previously identified typical range.
Table 3-4. P-values for calculated emission factors. Statistically significant
parameters (a = 0.05) are shown in bold-italics.
.•^M'i^Mt^^^'^^^^
X;:™:™:™:™X;^^^
Age
Vacuum
Height
Length
Age * Vacuum
Age * Height
Age * Length
Vacuum * Height
Length * Vacuum
Length * Height
y||I||^
!:i''.'"M'''f''*''""''4'''^ •
0.0059
< 0.0001
0.4972
0.9770
0.3388
0.6959
0.0047
0.8413
0.7934
0.9999
Table 3-5. Average emission factors per test at each house.
jiM^&^,
1
2
3
4
5
6
7
Average
Std Dev
,:;$;!;;^
'.mi^fflMfr
0.0048
0.0117
0.0011
0.0009
0.0014
0.0030
0.0021
0.0016
0.0082
0.0033
0.0025
0.0045
0.0198
0.0237
0.0063
0.0072
-¥*M*$eyC,vj;l
0.0012
0.0027
0.0005
0.0010
0.0008
0.0006
0.0014
0.0007
0.0059
0.0005
0.0020
0.0008
0.0103
0.0105
/:l::l::l::l::l|^t-feu{Wlf!d,.':::*::::::'";"^
* ..* ..* ..* ..* ..* ..* ..* ..* ..* ..* ..* .; ..* .. „ 3?. „ !.H J!..H™.. „ Jr Jr .. .... in'ir.' .'. .'. .'. .'. .'. .'. .'. .'. .'. .'. ...
,''ff'jf'fS^f^t,v",',t,';
0.0124
0.0210
0.0081
0.0121
0.0369
0.0170
0.0189
0.0097
0.0222
0.0198
0.0177
0.0110
0.0640
0.0471
0.0227
0.0160
r '-stit&e^'::.-'
0.0064
0.0035
0.0055
0.0061
0.0092
0.0022
0.0141
0.0055
0.0155
0.0133
0.0090
0.0068
0.0344
0.0205
Emission factors were expected to vary between houses, as indicated by the statistical significance of the carpet age and the
carpet age-pile height interaction term. We hypothesize that older carpet had higher emission factors because of the weaker
adhesion forces between the dust and carpet fibers. Adhesion forces present on new carpet should not vary as function of fiber
length. Hence, fiber length was not significant individually. The interaction between age and length was required to influence
the emission factor. We hypothesize that after carpet loses its electrostatic adhesive force (after an unknown period of time),
-------
emission factors will increase with increasing fiber length, as shown by Rodes (1998). Future research should be conducted to
confirm these two hypotheses.
Vacuuming also influenced the emission factors. Emission factors post-vacuuming were about 4x greater than those pre-
vacuuming (Table 3-5). Emission factors pre-vacuuming were lower even though more mass is resuspended during the pre-
vacuum experiments, regardless of carpet age or fiber length. The change in emission factors was driven by the mass available
on carpet, which was included in the denominator of the emission factor equation. Across all houses, vacuuming decreased the
mass available for resuspension by an order of magnitude, but the total mass resuspended decreased by a factor of 2.5, thereby
resulting in a net increase in the emission factors.
3.1.3 Resuspended PM Vertical Gradients
Confirmation of the vertical gradient in resuspended PM concentration was referred to in Section 3.1.1. Table 3 -6 provides
additional evidence. For all experiments except one, the resuspended PM mass at 18 in was always greater than the mass
measured at 36 in. The probability that the average ratio across all experiments was greater than unity was statistically
significant at a = 0.05.
Table 3-6. Average ratio of resuspended mass at 18 in versus 36 in separated into pre-vacuumed and
post-vacuumed conditions.
mmmmmmmmmm
$'" "HlWusfck'"" '"s:|
i
^
7
/I
c
T
Mean
Std Dev
Probability (p-value)
that the Ratio is > 1
&f':, ;;;;;; r^M^rSit^w^
W" ''^i^-^mii^^1'1 l"^I
1.16
a
1.71
1.33
2.76
1.33
2.42
1.24
1.29
2.09
1.28
1.37
3.09
1.95
1.77
0.64
< 0.0001
r3©iti'iafiK;, ;;;;;;; ,'::;i;r
1.84
1.33
b
2.00
1.55
1.25
2.93
1.39
2.67
1.80
3.48
1.19
0.87
1.93
1.86
0.76
< 0.0001
"No ratio because 18-in mass invalid. Pump shut off during sample collection.
bNo ratio because 36-in mass was below gravimetric analysis minimum detection
limit.
-------
3.1.4 PM Translocation
PM translocation within a house was measured with the Climets. Climet concentration data were not normally distributed.
A "log" transformation of the Climet data satisfied the normality parametric tests. The results from the spatial-temporal
autoregressive analysis of the transformed database are shown in Table 3-7. Location stratified the Climet locations into two
categories: within 1 ft of the resuspension area boundary or greater than 1 ft from the boundary. The spatial-temporal analysis
of the dependent variables "house," "vacuum," and "height" independently confirms their statistical significance on the
resuspended PM concentration measured by the Climets. The analysis also shows that resuspended PM concentrations varied
spatially and temporally.
Table 3-7. P-values for spatial-temporal differences in Climet concentrations. Statistically significant parameters
(a = 0.05) are shown in bold-italics.
Be^n
-------
Figure 3 1. Climet data collected at House 1, Test 1, showing the temporal and spatial dispersion of the
resuspended PM larger than 5 jjum. Red vertical line demarcates when resuspension activity stopped.
House 1 - Test 1
No Vacuum,dp > 5 pm
8000
7000
-18 inch
36 inch
Corner by Closet
Entrance to Room
-*~Mid Wall, Front of Bed
"•"Corner Opposite Door
Time
-------
3.2 Household Tracking Experiments
Paniculate matter data collected at each house, each week, and each segment are presented in Appendix D.
Paniculate matter mass tracked into the four houses was not normally distributed due to the small sample of test houses. The
"log" transformation provided the appropriate correction, and parametric tests satisfied normal distribution assumptions. Two
sets of data were considered. All data were used in one analysis for a total of 160 data points. The second set removed the data
from House 1, Week 1, for a total of 150 data points. There was heavy rainfall the day before Week 1 data collection and thus
heavy soiling of the carpet.
The statistical output from the analysis of both datasets is shown in Table 3-8. The full dataset and the reduced dataset indicated
only "week" and "house" variables were statistically significant. This finding means there was a temporal component to the
mass loading and the mass loading was dependent on the number of people living in the residence.
Table 3-8. Results from the spatial-temporal autoregressive analysis of the full and reduced datasets.
Statistically significant variables (95% confidence) listed in italics.
Segment
Week
House
Segment x Week
12
0.5730
< 0. 0001
< 0.0001
0.9999
12
0.5513
< 0. 0001
< 0.0001
0.9981
Surprisingly, the expected spatial gradient between carpet segments was not observed due to movement patterns within the
home and the sample collection frequency. Residents walked across the carpet in more than one direction. In three homes,
the residents had to cross the carpet horizontally to get between rooms. This movement evenly distributed the mass laterally
within a segment so that there was no statistical difference in the mass loading collected in different boxes, which is a desired
result. Residents also were instructed to enter and exit the home by walking across the carpet. The longitudinal movement in
two directions evenly distributed the paniculate matter vertically. As a result of these factors, weekly sample collection was too
infrequent to discern the spatial gradient expected. Future tests should limit longitudinal movement to one direction, preferably
entry into the residence. In addition, more frequent sample collection is required.
The spatial-temporal statistical analysis was conducted again considering only the significant variable to identify statistically
significant contrasts in mass loading on the carpet between weeks. The Tukey-Kramer least square means test identified
differences between weeks. The full dataset produced contrasts that were significant for the differences between Week 0 and the
remaining four weeks (Table 3-9). The only other significant contrasts were between Weeks 1-3 and Weeks 1-4. The reduced
dataset produced contrasts that were significant for the differences between Week 0 and the other weeks. In addition, there was
at least a 95% confidence of difference between Week 1-3, Week 1-4, Week 2-3, and Week 2-4. The reduced datasets better
show the temporal differences without the bias of House 1, which was affected by the rainstorm.
These results help determine the sample collection frequency that should be implemented in future tests. The rapid soiling of
the carpet during Week 1 indicates two to three sample sets should be collected during the first week.
_
-------
Table 3-9. Results from the Tukey-Kramer tests identifying differences in mass loading per week. Statistically
significant differences (95% confidence) listed in italics.
Residential tracking test comparison between weeks. Rain event (HI, Week 2) included.
WeekO
Weekl
Week 2
Week3
Week 4
WeekO
Weekl
< 0.0001
Week 2
< 0. 0001
0.4415
WeekS
< 0.0001
0.0036
0.9957
Week 4
< 0. 0001
0.0004
0.9921
1.000
Residential tracking test comparison between weeks. Rain event (HI, Week 2) removed.
WeekO
Weekl
Week 2
Week3
Week 4
WeekO
Weekl
< 0.0001
Week 2
< 0. 0001
0.0809
WeekS
< 0.0001
0.0451
0.0003
Week 4
< 0. 0001
0.0360
< 0.0001
0.4769
Although a PM tracking rate per carpet segment could not be calculated, rates for the entire carpet were calculated for each
house (Table 3-10). The average PM gain in the carpet per week varied between 2.7 and 24.1 (jug per in2-week. The mass gain
(or loss) per week probably is dependent on the number of traverses across the carpet by the occupants. Although not recorded
directly, the number of occupants (Appendix B) is a possible surrogate variable that is directly related to the loading via the
"house" variable (Table 3-8). Recording the number and direction of traverses would increase the sensitivity of the tracking
results and provide more insight into the tracking mechanism.
Table 3-10. PM tracking rate per week at each house sampled. Weeks 1-2 were combined into one value per house.
House
Tl
T2
T3
T4
Mass Change (|jig/in2- week)
Cumulative Mass (|jig/in2)
Mass Change (|jig/in2- week)
Cumulative Mass (|jig/in2)
Mass Change (|jig/in2- week)
Cumulative Mass (|jig/in2)
Mass Change (|jig/in2- week)
Cumulative Mass (|jig/in2)
•-flfc&'^'fti ;
8.04
8.04
25.03
25.03
12.98
12.98
4.65
4.65
"^fltfi&ifif
8.20
16.24
39.95
64.98
2.19
15.17
1.39
6.04
''''''' '•'"'''WfeteK' 4iL_ li
-6.11
10.13
7.24
72.21
-1.50
13.67
1.93
7.97
''^l^'.i'Sli^Detf"'1
3.38 ±8.22
24.07 ±16.38
4.56 ±7.53
2.66 ±1.75
-------
3.3 Laboratory Tracking Experiments
Laboratory tracking experiments examined the amount of paniculate matter transferred from the shoe to carpet and vice versa.
The fluorescent particles Tinopal and Uranine were selected for these experiments. General results from these experiments are
presented in Tables 3-11 and 3-12.
Table 3-11. Uranine concentrations available for collection on the shoe, loaded onto the shoe, and transferred to the
carpet during each step.
i ::;***"
i
2
3
4
5
6
7
8
Available •
: |Jig$jSa)kv!
2,973
± 629
853
± 110
462
± 134
3,465
±233
15,983
± 868
7,962
± 4,197
14,269
± 2,944
14,731
± 1,588
•.;!%*:!!
^wPift^'''
863
726
308
457
32,533
7,491
5,873
7,829
Step^l""
116
± 56
557
± 173
240
±178
158
± 110
21,395
± 3,157
7813
± 8,118
4,274
± 1,313
6,233
± 737
-••..Nbttlb*
:|stei»:^;;'
34
± 15
16
± 9
2
±2
39
± 17
385
± 84
205 ±
149
593
± 267
582
± 290
sferredFrom
stepjs,-!
47
± 16
12
± 13
0.04
±0
47
±6
129
±23
37
±27
483
± 376
458
± 355
Shi*to&p
!!:Lsi^*i-;j'
36
± 16
8
± 1
0.04
±0
13
± 10
41
± 10
25
± 10
220
± 83
164
± 66
\ Step^'l
23
± 10
3
± 1
0.04
±0
13
±0.6
50
± 33
9
± 1
152
± 45
106
± 64
:£&*&'•
38
± 15
3
± 3
0.7
± 1
14
±4
93
± 58
8
± 5
127
± 56
106
± 37
Table 3-12. Tinopal concentrations available for collection on the shoe, loaded onto the shoe, and transferred to the
carpet during each step. Mass transferred from carpet to shoe during Step 1.
'inijkMpwW'' w>
i
2
3
4
5
6
7
8
. Available
"""-•toh1^:!::
7,039
±790
11,234
± 2,947
5,227
± 1,302
13,333
±4,145
110,951
± 2,362
49,581
± 23,268
64,126
±3,192
59,197
± 21,604
: ifrarisferrtitein" :
*<:$ar^'to-8fi*»: ;
"'"•-•
ESi^
0.0
38
±5
0.0
10
±18
92
±78
78
± 11
39
±7
0.0
~inr
-------
3.3.1 Uranine Results
The variability in the mass loaded onto the shoes required normalization to allow comparison of the data between tests.
The mass transferred from the shoe to the carpet (Steps 1-6) was normalized by the mass loading on the shoe (Step 0) to
yield the mass fraction deposited per step. This conversion also allowed easy calculation of the fraction remaining on the
shoe and cumulative fraction deposited on the carpet. The fraction deposited per step was not normally distributed. A "log"
transformation of the data satisfied the normality parametric tests. The results from the spatial-temporal autoregressive analysis
of the transformed database are shown in Table 3-13. The Tukey-Kramer least square means tests identified that the Step 1
mass fraction transferred was statistically significant different from the remaining steps at the 95% confidence level. The mass
fractions transferred during Steps 2-6 were not statistically different.
Table 3-13. Results from the spatial-temporal autoregressive analysis of the Uranine data.
Statistically significant variables (95% confidence) listed in italics.
p^rife^rS
Carpet Style
Carpet Age
Load
Step#
Load * Step
^WfrtiM*11-'^
1
1
1
5
5
"ftiiffiiMHfoiiii'f!™
0.1598
0.1311
0.0827
< 0.0001
0.9685
Mass load onto the shoe was almost statistically significant at the 95% confidence level. The similarity in the fraction deposited
during Steps 2-6 independent of mass loading was the primary cause. Tighter control over the mass loaded onto the shoe also
may increase the probability of the parameter becoming statistically significant.
The relationship between step number and mass fraction remaining on the shoe load and the cumulative fraction transferred
for the low mass loading tests is shown in Figure 3-2. A similar graph for high mass loading tests is shown in Figure 3-3.
The figures confirm between 40% (low loading) and 80% (high loading) of the mass is transferred to the carpet on Step 1.
Regardless of mass loading, only 10% of the mass on the shoe is transferred during subsequent steps. The mean (± std dev)
mass fraction transferred per step for Steps 2-5 was 0.03 (± 0.02).
-------
Figure 3-2. Transfer of Uranine from shoe to carpet per step during low loading tests. Mass fractions are averages of
four tests. Fraction remaining on shoe after each step and cumulative fraction transferred per step are shown.
•Fraction Remaining on Shoe
•Cumulative Fraction Transferred
-------
Figure 3-3. Transfer of Uranine from shoe to carpet per step during high loading tests. Mass fractions are averages of
four tests. Fraction remaining on shoe after each step and cumulative fraction transferred per step are shown.
"Fraction Remaining on Shoe
"Cumulative Fraction Transferred
-------
3.3.2 Tinopal Results
Tinopal loaded onto carpet determined the amount transferred from the carpet to the shoe during the first step. Subsequent steps
determined the amount transferred from the shoe back into the carpet. Tinopal shoe-to-carpet transfer data, when combined with
the Uranine data, would double the amount of data and increase the confidence in the statistical analysis.
The quantity of Tinopal loaded onto the carpet, the mass transferred to the shoe during the first step, and the fraction transferred
data are presented in Table 3-14. The quantity of Tinopal from the carpet to the shoe during the first step was 1.1%. For
unknown reasons, the quantity of Tinopal transferred to the shoe was a very small percentage of the total loaded onto the carpet.
Because of the small quantity on the shoe, the shoe-to-carpet transfer data on subsequent steps for Tests 1, 2, 3, 4, and 8 did not
follow the trend expected from the Uranine results. Thus, the results were inconclusive and additional statistical analysis was
not possible.
Table 3-14. Total mass and mass fraction of Tinopal loaded onto carpet that was transferred to the shoe
during Step 1.
p*i#**<«
i
2
o
J
4
5
6
7
8
P^w&onl^^ISi^^
7,039 ± 790
11,234 ±2,947
5,227 ± 1,302
13,333 ±4,145
110,951 ±2,362
49,581 ±23,268
64,126 ±3, 192
59,197 ±21,604
!!(<(!(('".'.'.' !"•'•'• '/!i/tii/t;-i"t!& ,',"/?'('" "A'.
^Ma»0rv;ShM*I^W^
42.1
58.6
154.9
166.4
1,226.1
699.1
819.6
0.0
Mean Transferred
f»'«1tert!»reP'l;
0.6
0.5
2.8
1.2
1.1
1.4
1.3
0.0
1.1 ± 0.8
-------
-------
4.0
Quality Assurance
Quality assurance and quality control measures for the project were outlined in the QAPP entitled "Resuspension of Paniculate
Matter on Flooring - Quality Assurance Project Plan for Basic Research Projects," EPA Order No. 4C-R179-NALX, EPA/
NHSRC under Contract No. QT-OH-04-00315 RTI Project No. 09188). Tables 4-1 and 4-2 summarize the QA/QC measures for
the project. The QA/QC results for each metric are presented below.
Table 4-1. QA/QC criteria for measurements collected.
• :,:,:i,Metrlc;,ia,.'
Mass
APS
Climet
Fluorometry
T
RH
:;,:$x^Mm, ),;',';
± 20%
± 5%
± 10%
± 5%
± 5%
± 5%
\,,. iMmm&t':-^
±5%
a± 15%
a± 20%
± 10
±5%
± 10%
'-, CbjWpietentesife',,
± 95%
± 95%
± 95%
± 95%
± 95%
± 95%
\v;..^Ifc;.ȣ:<
2 |Jig/m3
0.5 (jum
0.5 (jum
0.3 jjug
1°C
1%
v. /'MDL:VV, ,
1 |Jig/m3
NA
NA
o.i ^g
NA
NA
'•jil'.iiWfc .///;..'
3 |Jig/m3
NA
NA
0.3 ^g
NA
NA
NA = not applicable
^Determined from manufacturer's calibration certificate.
Table 4-2. Quality control measures to implement during testing.
Mptrir > '•"•'• ""
.fflSMjSu/,/,/,/,/,/,/,/,/,/,/,/,,';,,
APS
Climet
Mass
Fluorometer
•Qililj^'Confrol-lvaloattoi '•' — ""— ^^
Precision: Collocated instruments once/week
Zero: HEPA filter installed on inlet daily
Background: Collected prior to each test
Precision: Collocated instruments once/week
Zero: HEPA filter installed on inlet daily
Background: Collected prior to each test
Precision: Collocated instruments once/week
Field Blanks: 5% of filters collected
Laboratory Blanks: 5% of filters collected
Background: Collected prior to each test
Precision: < 5% error compared against known standard
Zero: Measured daily with clean, deionized water
Background: At least 1 sample per experiment per surface
4.1 Gravimetric Mass
Gravimetric mass measurements determined the mass available for resuspension via the microvacuum and the mass
resuspended via the URG samplers. The microvacuum and the URG samplers collected the mass on a filter that was analyzed
gravimetrically.
The number of microvacuum and URG samples attempted and successfully collected is outlined in Table 4-3. The cumulative
completeness percentage (99.2%) exceeded the data quality objective.
Precision and accuracy of the gravimetric method were other quality assurance criteria (Table 4-4). Precision in the gravimetric
analysis was assessed by collecting collocated URG samples once per week. Precision was calculated as the % relative standard
deviation (%RSD). Accuracy was assessed every gravimetric analysis session (both pre-weighing and post-weighing) by
weighing a 100 (jug standard weight. A gravimetric analysis session was not started until the measured weight was within 5% of
the stated value.
-------
Table 4-3. Metric sample completeness statistics.
NMftil
Valid Samples
Planned Samples
% Completed
t»te«artiFtla>
55
56
98.2
Microvacuum
205
206
99.5
Cumulative
260
262
99.2
Table 4-4. Gravimetric analysis precision and accuracy statistics.
Prfecisran
Number Sample Pairs
Mean %RSD
8
10.9%
Xchutxy
Number Measurements
Mean
% Difference
31
99.6 jjug
0.4%
A series of filter blanks assessed quality control in the field and laboratory (Table 4-5). The mean masses collected on the filter
blanks for each metric were used as correction factors for the experimental samples. Field blanks typically had much higher
variation in the mass gain/loss due to processing of the samples in the field where many potential sources of contamination
existed. The percentage of blanks collected (10.8%) exceeded the planned percentage (10%).
Table 4-5. Filter blank statistics.
U RG Field Blanks
Number
4
Mass (|j,g)
-0.1 ±7.1
Microvacuum Field Blanks
Number
15
Mass (|j,g)
1.3 ±7.0
iy»raV^'BIanks
Number
9
Mass (|j,g)
0.1 ± 0.6
4.2 Climet Optical Particle Counters
Climet units measured resuspended particle number concentration throughout the test area. The QA/QC results for this metric
are summarized below.
The number of Climet samples attempted and successfully collected is outlined in Table 4-6. The cumulative completeness
percentage (95%) met the data quality objective. The primary reason for invalid samples was forgetting to start the instrument
data collection.
Table 4-6. Climet sample completeness statistics.
ftilric
Valid Samples
Planned Samples
% Completed
Background
112
120
93.3
Test
116
120
96.7
Cumulative
228
240
95.0
-------
Precision (Table 4-7) and accuracy of the Climet measurements were other quality assurance criteria. Precision was assessed by
collecting collocated Climet samples once per week, corresponding to the weeks a series of resuspension tests were conducted
in a house. The correction factor was applied to all data because of known differences between units. Precision was calculated
as the %RSD against the reference instrument (Climet #958304). Precision data quality objectives were met for all Climets.
Accuracy was supposed to be determined via the manufacturer's calibration certificate for each instrument. However, the
calibration certificates for these instruments had expired. The manufacturer also no longer supports these instruments and
calibration would have exceeded the available resources for this project (~$ 1,000 per instrument for calibration). Instead, the
Climets were collocated with the APS (reference), and Pearson correlation coefficients were calculated (Table 4-8). The APS
was selected as a reference because it has a current calibration certificate. A correlation coefficient >0.8 satisfied the data quality
objective.
The quality control assessment in the field was to make sure the Climets measured a concentration of 0 particles per cm3 in the
field. A HEPA filter was installed on the inlet to each unit once per house, for a total of 8 samples. All Climets always measured
0 particles per cm3 when the HEPA filter was installed.
Table 4-7. Climet particle number concentration precision statistics. Climet #958304 was the reference instrument
for application of correction factors.
rif facf&ff;
'
923954
1.3
1.4
7.3%
8.5%
958301
1.9
1.9
10.2%
10.8%
958302
1.4
2.3
7.7%
10.1%
958303
0.8
2.0
5.1%
9.7%
958304
Reference
Reference
958305
1.2
1.3
.3%
9.5%
Table 4-8. Climet particle number concentration accuracy statistics. Accuracy determined by Pearson correlation
coefficients between the APS and the Climet unit selected.
923954
958301
958302
958303
958304
958305
120
120
120
120
120
120
0.88
0.94
0.85
0.91
0.96
0.90
0.86
0.92
0.82
0.87
0.92
0.86
4.3 Aerodynamic Particle Sizer
The APS measured resuspended particle number and mass concentration at the resuspension area boundary. The QA/QC results
for this metric are summarized below.
The number of APS samples attempted and successfully collected is outlined in Table 4-9. The cumulative completeness
percentage (100%) exceeded the data quality objective.
Table 4-9. Aerodynamic Particle Sizer sample completeness statistics.
" :,;i» •' •• ', ' '-'/M-'/M', '
1w6wlC
Valid Samples
Planned Samples
% Completed
'"' " " " ' ft "- • ii '"-: V ' ''~jf .'".'",'",'",'"' '
•;;;;;; ;?5iR?l^ftR?IR¥i ,',',',',',%•
28
28
100%
.,",'• - - - T •'»,,,,,,,,,,;•
KSf ' .'.'.'.'•'..
28
28
100%
;,/••'••' j» '"""i'f'*'" :"'''':':':':-:'
•i.!'!V!Vf!r!WWiwJ.w^:;:;:;:;:;:;:;:
56
56
100%
-------
Precision and accuracy of the APS measurements were other quality assurance criteria. Precision could not be assessed because
only one APS was available. Accuracy was determined to be within specifications because the APS manufacturer's calibration
certificate was still valid. The APS was calibrated on December 15, 2004. The APS also agreed with the particle number
concentrations measured by the Climet optical particle counters.
The quality control assessment in the field was to make sure the APS measured a concentration of 0 particles per cm3 in the
field. A HEPA filter was installed on the inlet to the APS once per house, for a total of 8 samples. The APS always measured 0
particles per cm3 when the HEPA filter was installed.
4.4 Temperature and Relative Humidity
The HOBO H8 measured the temperature and relative humidity within each house during resuspension and tracking data
collection. Only one HOBO was used, so precision was not assessed. Accuracy was measured by placing the unit in a
temperature and humidity controlled chamber. The QA/QC results for this metric are summarized in Table 4-10. All data quality
objectives were achieved.
Table 4-10. QA/QC results for the HOBO H8 temperature and relative humidity data.
Valid Samples
Planned Samples
% Completed
28
28
100%
Temperature
RH
97.1
4.5 MetOnes
All MetOne data were invalid and not used in any statistical analysis. MetOne optical particle counters did not exhibit the
vertical concentration gradient demonstrated by the gravimetric and Climet optical particle counter data (Table 4-11). The
corrections for differences in counting efficiency were included in the MetOne data analysis, although the correction factors
were small (Table 4-12). The inability of the MetOnes to reach an equilibrium concentration during the six-second sampling
interval combined with the rapidly changing resuspended PM concentration increased the signal-to-noise ratio in the data.
Visual observation of the display screen during operation showed the measured concentration fluctuated greatly during sample
collection. As a result, larger concentration differences between the two heights were required for the MetOnes to detect a
height-dependent concentration gradient.
Table 4-11. MetOne concentrations at 18 and 36 inches. All data corrected for background
and differences between units.
House 1
House 3
House 4
House 7
Testl
Test 2
Test3
Test 4
Testl
Test 2
Tests
Test 4
Testl
Test 2
Tests
Test 4
Testl
Test 2
Tests
Test 4
1.77
1.34
1.99
no data
0.36
0.18
0.25
0.16
0.54
0.27
0.42
0.21
0.34
0.32
0.03
-0.16
1.48
1.37
1.50
0.15
0.30
0.19
0.22
0.21
0.49
0.27
0.64
0.19
0.72
1.18
0.18
0.27
-------
Table 4-12. MetOne concentration correction factors.
House 1
House 3
House 4
House 7
Means
1.03
1.01
1.18
0.83
1.01 ±0.14
0.77
0.80
1.23
1.20
1.00 ±0.25
4.6 Fluorometry
Fluorometer measurements quantified the mass of fluorescent material in each sample collected. A total of 768 measurements
were collected, 384 of each type (Table 4-13). More than 99% of the samples collected were valid, although 8% of the Tinopal
fluorescence readings were below the instrument detection limit when some fluorescence was expected. Invalid samples were
caused by accidental spilling of the wash fluid.
Table 4-13. Fluorometry sample completeness statistics.
Valid Samples
Planned Samples
% Completed
381
384
99.2
380
384
99.0
761
768
99.1
The precision in the fluorometry measurements was assessed by collecting duplicate readings for each sample. The
nondestructive nature of the analysis allowed repeat fluorescence measurements if initial precision criteria (> 95%) were
not achieved for a sample. If a sample was reanalyzed, a new aliquot of the sample was pipetted into a well-plate. The mean
precision across all samples was 97.8%.
The accuracy of the fluorometry measurements corresponded to the calibration curve generated for each pair of tests (Table 4-
14). From the calibration curve regression statistics, the accuracy in the measured mass concentrations was > 99% for all tests.
Table 4-14. Fluorometer calibration curve statistics for each pair of tests.
5&6
7&8
Slope
0.0014
0.0014
0.0013
0.0015
0.998
0.997
0.997
0.998
Slope
0.0020
0.0019
0.0018
0.0019
R2
0.993
0.996
0.998
0.997
Quality control during the experiments was assessed by collecting blank samples during each test (Table 4-15). Sample
fluorometric masses collected from each source were corrected for the background mass. On average, the fluorometric mass
found in background samples was less than 1% of the sample fluorometric mass.
Table 4-15. Fluorescent tracer mass found in quality control samples.
Blank 0.0 INNaOH
Shoe
Carpet-Step #1
Carpet - Steps #2-6
Cookie Sheet
0.0 ±0.3
0.4 ±0.8
0.5 ±1.5
1.0 ±1.8
0.0 ±0.0
0.0 ±0.7
0.0 ±0.0
1.3 ±3.2
0.0 ±0.0
0.0 ±0.0
-------
-------
5.0
Conclusions
Resuspension Experiments
1) Older carpet with poor maintenance released more PM mass during resuspension experiments than new, well-maintained
carpet. Similarly, the mass available for resuspension varied between houses, with the old and poorly maintained carpet
having greater quantities available. These findings agree with previously reported laboratory findings.
2) Vacuuming decreased the amount of PM mass resuspended by approximately 44% (independent of the mass available for
resuspension) for most carpets. Only new, frequently vacuumed carpet did not show this decrease. A normal, residential
vacuum could be an effective remediation measure. However, the quantity of mass removed must be balanced against the
amount deposited onto the carpet by tracking or other mechanisms.
3) Resuspended PM concentrations did decrease as height above the floor decreased. PM concentrations 36 in above
the floor were about 1.8 times lower than those 18 in above the floor. Vacuuming the carpet did not affect this ratio,
indicating that vacuuming removed all particle sizes with uniform efficiency.
4) Emission factors varied between 0.006 (new, clean carpet) and 0.023 (old, recently vacuumed carpet). The emission
factors fall within the range found during laboratory experiments.
5) As expected, emission factors varied with carpet age and carpet age-fiber length interaction. Rodes and Thornburg
(2004) and Thornburg and Rodes (2004a, 2004b) showed carpet age affected emission factors because of variations in
particle adhesion and loading characteristics. The hypothesized change in adhesion forces with age would cause the fiber
length to become important after a period of time. Rodes (1998) showed emission factors vary with carpet fiber length.
6) Emission factors measured after vacuuming the carpet were 4 times greater than those prior to vacuuming. Although
vacuuming reduced the total mass resuspended by a factor of 2.5, vacuuming also reduced the PM mass available for
resuspension by a factor of 10. The net result is an increase in the measured emission factors. The unintuitive nature of
this finding suggests carpet age, cleaning frequency, and other characteristics must be known when applying emission
factors to exposure models.
7) Spatial-temporal analysis of the resuspension data suggests significant PM mass translocation occurs at distances of
approximately 8 ft. Development of statistical or physical models to predict the amount of translocation and the salient
characteristics (e.g., room size) was beyond the scope of this research. Simple ratios of the data indicate 10% to 50% of
the resuspended PM migrates at least 8 feet (at a height of 36 in) from the source. However, the data are available for
more thorough analysis.
Residential Tracking Experiments
1) A method and sample collection equipment were developed to collect microvacuum samples within homes to determine
the PM movement rate due to tracking.
2) Rain or other events that change the moisture and adhesion properties of the PM can greatly influence the tracking rate.
3) The cumulative mass tracked into buildings varied between homes and between weeks. As expected, the cumulative mass
accumulated in the carpet increased steadily during the four weeks and more mass was collected in homes with more
occupants.
4) Tracking rates varied between 2.7 and 24.1 (jug per in2-week. The rate probably is highly dependent on the number of
traverses across the carpet, currently identified by the number of residence occupants.
Laboratory Tracking Experiments
1) Experimental procedures were developed to evenly load a surface with a known quantity of fluorescent particles, collect
samples from a variety of surfaces, extract and analyze the samples to determine the quantity of fluorescent mass, and
reduce the data to determine the amount of mass per unit surface area. These procedures will be useful for conducting
future tracking experiments to expand the preliminary findings reported here.
2) Data quality objectives were achieved for these experiments.
3) The amount of PM transferred was associated with the step number. More than 40% of the mass was transferred on the
first step. The remaining 10% was transferred during the subsequent five steps in approximate 2% increments.
-------
4) Uranine particle mass load on the shoe surface, carpet age, and carpet style did not influence the PM transfer from the
shoe to the carpet. Uranine mass load was statistically significant at the 90% confidence level. It is possible additional
testing, to increase the degrees of freedom, will prove PM load is a significant variable.
5) On average, 1.1% of the mass is transferred from the carpet to the shoe during a step. The fraction transferred was
independent of the experimental conditions. Additional research is needed to understand the carpet-to-shoe transfer
process. The small quantity transferred prohibited confirmation of the Uranine findings because there were not sufficient
data for statistical analysis.
-------
6.0
References
Farfel, M.R., Orlova, A.O., Lees, P.S.J., Bowen, C., Elias, R., Ashley, RJ., and Chisolm, JJ. "Comparison of Two Floor Mat
Lead Dust Collection Methods and Their Application in Pre-1950 and New Urban Houses." Environ. Sci. & Techno!., 35:2078-
2083 (2001).
Ferro, A.R., Kopperud, R.J., and Hildemann, L.M. "Source Strengths for Indoor Human Activities That Resuspend Paniculate
Matter" Environ. Sci. & Technol., 38:1759-1764 (2004).
Peters, T.M., and Leith, D., "Concentration Measurement and Counting Efficiency of the Aerodynamic Particle Sizer 3321."
J. Aerosol Sci., 34:627-634 (2003).
Rodes, C.E. 1998. "Experimental Methodologies and Preliminary Mass Transfer Factor Data for Estimation of Dermal
Exposures to Particles From Surfaces." Final Report, RTI Project 6980, U.S. EPA Contract 68-D5-0040 WA019 and WA 023.
Rodes, C.E., and Thornburg, J. 2004. "Study of Resuspension of Paniculate Matter on Flooring Surfaces Due to Human
Activity." Final Report, RTI Project 08886. U.S. EPA Contract 3C-R185-NALX.
Thornburg, J., and Rodes, C.E. 2004a. "Resuspension of Fibers From Indoor Surfaces Due to Human Activity." Final Report,
RTI Project 08924. U.S. EPA Contract 3C-R321-NANX.
Thornburg, J., and Rodes, C.E. 2004b. "Indoor PM Resuspension Testing for Metals." Final Report, RTI Project 08931. U.S.
EPA Contract 3C- R340-NALX.
-------
-------
APPENDIX A
Characteristics of Houses Participating
in Resuspension Tests
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
HI
Bedroom, 10.5 ft x 14 ft
2
0
3 cats Outside? yes
medium pile
15 mm
3. 5-3. 75 years
no
-every 2 weeks
on occasion
never
no
bonus room (other side of house)
no
no, minimal furniture
front door and garage door
good
concrete; 6 ft x 20 ft, slightly dirty
yes; large deck
gravel drive
natural front, grass lawn left and backyard
-------
House #1
FUTON
BED
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
H2
Family Room
2
1
1 dog, 2 cats Outside? no
medium pile, cut loop
15 mm
3 years
somewhat; heavy foot traffic areas.
twice per week
yes
no
no
no
no
kids 'toys
garage and front door
good
yes; 6 ft x 6 ft, concrete
wood; 12 ft x 12 ft
yes; drive and walkway
grass in front and backyards
-------
House #2
To Upstairs
D
0
0
R
Q_
i
Table
A Z_
Fireplace
SOFA
Tl
T4
T2
T3
D
0
0
R
Q_
i
Open Archway to Kitchen
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
H3
Family Room
2
2
1 dog, 1 cat Outside? yes, both
medium pile, cut loop
10 mm
6 years
somewhat; heavy foot traffic areas
once per week
no
November, 2004
no
no
no, some dirt on floor/tables
kids' toys
garage
good
yes; 8 ft x 6 ft, concrete.
wood; 16 ft x 12 ft, access thru sun porch
yes; drive and walkway
grass in front and backyards.
-------
House #3
o
o
3-
(V
Window
CHAIR
SOFA
Open Archway to Entryway
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
H4
Upstairs Bedroom
2
0
2 cats Outside? no
medium pile
10.5mm
1.5 years
no
once per week
yes, with pet hair release
no, but uses spot remover occ.
no
no
no
no
through garage
vinyl siding
no; stoop
yes; 8 ft x 8 ft
yes; into garage
yes, doesn't walk through
-------
House #4
Q.
O
TV
D
8
SOFA
DO
SO
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
H5
Master Bedroom
2
0
1 dog, 2 cats Outside? yes
medium pile, cut loop
8 mm
> 10 years
yes
weekly
yes, monthly
~ 4 years ago
no
no
some dust
lots of furniture
kitchen door
good
yes; wood, 8 ft x 20 ft
no
concrete drive and walk to porch
grass front and backyard
-------
House #5
Q_
i
D
r
e
s
s
e
r
Window
BED
To Bath
Door
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
H6
Living Room
1
0
0 Outside? no
medium pile, cut loop
5 mm
> 10 years
yes
every 2 months
no
> 5 years
no
no
yes, not cleaned > 3 months
no, minimal furniture
front door
good
no
yes; 10 ft x 10 ft
concrete drive and walk to porch
dirt front yard, grass backyard
-------
House #6
BED
Window
Dresser
Q.
i
Climet
To
Bath
D
r
e
s
s
e
r
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
H7
Living Room
2
0
1 dog, 2 cats Outside? no
medium pile, cut loop
5 mm
8 years
somewhat
every month
yes
> 2.5 years
no
no
yes, not cleaned in 2 months
yes
front door
good
yes; wood, 5 ft x 18. 5 ft
yes; 10 ft x 13 ft
concrete drive and walk to porch
grass front and backyard
-------
House #7
o
o
8
I-*
o
o
T3
Tl
A Z
Fireplace
T4
T2
SOFA
Open Archway to Kitchen
FUTON DESK
O
m
-i
m
Climet
Door Open
§
3
Q.
-------
APPENDIX B
Characteristics of Houses Participating
in Walk-off Mat Tracking Tests
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
Tl (Same as HI)
Front Hall (at Door) 3.5 ft x 12 ft
2
0
3 cats Outside? yes
Mohawk Horizon
11.5 mm
new
no
DO NOT VACUUM
N/A
N/A
no
no
no
no
front door
good
yes; small, concrete 6 ft x 20 ft
yes; large, off dining rm
no; gravel, leads to front steps
grass lawn on left and backyard,
woods in remainder of yard
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
T2
Entry Hall, 3 ft x 12 ft
2
1
2 cats, 1 dog Outsic
e? dog
Mohawk Horizon
11.5 mm
new
no
DO NOT VACUUM
N/A
N/A
no
yes, outside
no
no
front door
flaking on columns and porch ceiling
yes; 12 ft x 40 ft
N/A
no; gravel drive, concrete walkway
both; walk through lawn to porch
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
T3
Front Hall
2
0
2 cats Outside? no
Mohawk Horizon
11.5 mm
new
no
DO NOT VACUUM
N/A
N/A
no
no
no
no
back door
good
10 ft x 30 ft
small porch/stoop
gravel drive
grass front, mulch backyard
-------
House ID
Room Description
# Occupants
Adults
Kids
Pets
Carpet Characteristics
Type
Pile height
Age
Matted?
Cleaning History
Vacuum frequency?
Deodorizer?
Steam cleaned/shampoo?
Interior Survey
Water damage?
Flaking paint?
Dusty surfaces?
Clutter?
Exterior Survey
Primary entrance?
Exterior paint condition?
Front porch? Size?
Back deck? Size?
Concrete drive? Entrance?
Dirt/grass yard?
T4
Hallway
1
1
0 Outside?
Mohawk Horizon
11.5 mm
new
no
DO NOT VACUUM
N/A
N/A
no
no
no
no
front door
vinyl siding - excellent condition
no
8 ft x 6 ft
sidewalk and breezeway
in back of unit
-------
APPENDIX C
Climet Translocation Graphs
House 1 - Test 1
No Vacuum, dp > 5 jim
8000
6000
5000
4000
3000
2000
Entrance to Room
Mid Wall, Front of Bed
Corner Opposite Door
Time
House 1 - Test 2
Vacuumed, dp > 5 urn
5000
4500
18 inch
36 inch
Corner by Closet
Entrance to Room
Mid Wall, FrontofBed
CornerOpposite Door
Time
-------
House 1 -Tests
No Vacuum, dp > 5 urn
Time
House 1 - Test 4
Vacuumed, dp > 5 urn
"18 inch
"36 inch
Corner by Closet
Entrance to Room
"Mid Wall, FrontofBed
"CornerOpposite Door
Time
-------
House 3 - Test 1
No Vacuum, dp > 5 urn
4000
3500
3000
2500
2000
1500
500
Time
House 3 - Test 2
Vacuumed, dp > 5 urn
1600
1200
800
600
200
Time
-------
House 3 -Tests
No Vacuum, dp > 5 urn
4000
3500
18 in ch
36 inch
Fireplace
Sofa End Table
TV/Window
Love Seat
Time
House 4 - Test 1
No Vacuum, dp > 5 urn
7000
6000
5000
4000
3000
2000
1000
Time
-------
House 4 - Test 1
No Vacuum, dp > 5 urn
7000
Time
House 4 -Test 2
No vacuum, dp > 5 urn
4500
4000
3500
3000
2500
* 2000
1500
500
Time
-------
House 4 -Tests
Vacuum, dp > 5 urn
4000
3500
3000
2500
. 2000
1500
1000
500
18 inch
36 inch
Corner Cabine
Corner Bathroom
Door
Corner by Sofa
Time
House 4 - Test 4
Vacuum, dp > 5 urn
1400
1200
1000
Time
-------
House 7 - Test 1
No Vacuum, dp > 5 urn
30000
25000
Time
House 7 -Test 2
No vacuum, dp > 5 urn
35000
30000
25000
20000
15000
10000
5000
Time
-------
House 7 -Tests
Vacuumed, dp > 5 urn
Time
House 7- Test 4
Vacuumed, dp > 5 urn
9000
Time
-------
APPENDIX D
Household Tracking Data
Note: Position indicates row number and location number specified in Figure 2-4. The first digit is the row number and the
second digit is the location number. Week 0 corresponds to the background samples collected immediately after deployment of
the carpet inside the house.
-------
House Tl
House T2
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
o
J
o
J
o
J
3
3
3
3
o
J
4
4
4
4
4
4
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
11
25
11
25
11
25
11
25
22
14
22
14
22
14
22
14
21
13
21
13
21
13
21
13
24
12
24
12
24
12
24
12
23
15
23
15
23
15
23
15
20.90
12.60
14.90
14.50
15.50
21.20
8.70
29.70
100.30
120.20
73.70
80.10
100.80
115.90
73.80
73.20
564.90
653.30
389.90
604.60
785.00
628.80
329.30
616.30
221.50
178.60
157.60
176.00
130.50
183.20
138.10
142.40
176.90
112.70
153.40
90.50
118.00
55.90
163.20
82.70
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
o
J
o
J
o
J
3
3
3
3
o
J
4
4
4
4
4
4
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
11
25
11
25
11
25
11
25
22
14
22
14
22
14
22
14
21
13
21
13
21
13
21
13
24
12
24
12
24
12
24
12
23
15
23
15
23
15
23
15
14.50
26.10
26.90
45.00
51.70
41.40
24.50
44.60
260.10
444.70
356.90
447.00
580.60
440.80
327.40
467.50
82.30
167.20
115.20
150.80
71.40
156.70
80.80
167.60
526.00
627.60
440.20
431.00
699.30
775.80
748.90
853.40
772.00
666.60
836.50
352.80
742.30
636.70
727.40
955.80
-------
House T3
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
1
1
2
2
3
3
4
4
11
25
11
25
11
25
11
25
22
14
22
14
22
14
22
14
21
13
21
13
21
13
21
13
24
12
24
12
24
12
24
12
23
15
23
15
23
15
23
15
8.60
18.10
19.50
6.30
10.10
11.70
15.10
7.40
186.60
71.50
84.20
124.90
88.70
76.80
89.20
82.50
279.60
262.50
142.10
159.90
158.40
120.20
128.50
129.30
275.70
161.10
121.10
145.60
154.40
134.60
111.20
138.50
159.80
100.80
159.90
136.10
151.60
148.40
110.10
132.50
-------
&EPA
United States
Environmental Protection
Agency
PRESORTED STANDARD
POSTAGES FEES PAID
EPA
PERMIT NO. G-35
Office of Research and Development
National Homeland Security Research Center
Cincinnati, OH 45268
Official Business
Penalty for Private Use
$300
Recycled/Recyclable
Printed with vegetable-based ink on
paper that contains a minimum of
50% post-consumer fiber content
processed chlorine free
------- |