svEPA
United States
Environmental Protection
Agency
     Air Quality Criteria for
     Participate Matter

     Volume I

     Fourth External Review Draft
                                xxx

-------

-------
                                                            EPA/600/P-99/002aD
                                                                    June 2003
                                                      Fourth External Review Draft
Air Quality  Criteria for Participate Matter
                        Volume I
       National Center for Environmental Assessment-RTF Office
               Office of Research and Development
               U.S. Environmental Protection Agency
                   Research Triangle Park, NC
                                                  /"TV  Recycled/Recyclable
                                                        Printed wilh vegetable-based ink on
                                                        paper that contains a minimum of
                                                        50% post-consumer fiber content
                                                        processed chlorine free.

-------
 I                                         DISCLAIMER
 2
 3           This document is an external review draft for review purposes only and does not constitute
 4      U.S. Environmental Protection Agency policy. Mention of trade names or commercial products
 5      does not constitute endorsement or recommendation for use.
 6
 7
 8                                            PREFACE
 9
10          National Ambient Air Quality Standards (NAAQS) are promulgated by the United States
11      Environmental Protection Agency (EPA) to meet requirements set forth in Sections 108 and 109
12     of the U.S. Clean Air Act (CAA).  Sections 108 and 109 require the EPA Administrator (1) to
13     list widespread air pollutants that reasonably may be expected to endanger public health or
14     welfare; (2) to issue air quality criteria for them that assess the latest available scientific
15     information on nature and effects of ambient exposure to them; (3) to set "primary" NAAQS to
16     protect human health with adequate margin of safety and to  set "secondary" NAAQS to protect
17     against welfare effects (e.g., effects on vegetation, ecosystems, visibility, climate, manmade
18     materials, etc.); and (5) to periodically (every 5 years) review and revise, as appropriate, the
19     criteria and NAAQS for a given listed pollutant or class of pollutants.
20           The original NAAQS for particulate matter (PM), issued in 1971 as "total suspended
21     particulate" (TSP) standards, were revised in 1987 to focus  on protecting against human health
22      effects associated with exposure to ambient PM less than 10 microns (< 10 urn) that are capable
23      of being deposited in thoracic (tracheobronchial and alveolar) portions of the lower respiratory
 24      tract.  Later periodic reevaluation of newly available scientific information, as presented in the
 25      last previous version  of this "Air Quality Criteria for Particulate Matter" document published in
 26      1996, provided key scientific bases for PM NAAQS decisions published in July 1997.  More
 27      specifically, the PM10 NAAQS set in 1987 (150 ug/m3,24-h; 50 ug/m3, annual average) were
 28      retained in modified form and new standards (65 ug/m3,24-h; 15 ug/m3, annual average) for
 29      particles ^ 2.5 urn (PM2.5) were promulgated in July 1997.
 30           This Fourth External Review Draft of revised Air Quality Criteria for Particulate Matter
 31      assesses new scientific information that has become available mainly between early 1996

        June 2003                                 I-ii         DRAFT-DO NOT QUOTE OR CITE

-------
   1      through April 2002. The present draft is being released for public comment and review by the
  2      Clean Air Scientific Advisory Committee (CASAC) to obtain comments on the organization and
  3      structure of the document, the issues addressed, the approaches employed in assessing and
  4      interpreting the newly available information on PM exposures and effects, and the key findings
  5      and conclusions arrived at as a consequence of this assessment.  Public comments and CASAC
  6      review recommendations will be taken into account in making any appropriate further revisions
  7      to this document for incorporation into a final draft. Evaluations contained in the present
  8      document will be drawn on to provide inputs to associated PM Staff Paper analyses prepared by
  9      EPA's Office of Air Quality Planning and Standards (OAQPS) to pose alternatives for
 10      consideration by the EPA Administrator with regard to proposal and, ultimately, promulgation of
 11      decisions on potential retention or revision of the current PM NAAQS.
 12           Preparation of this document was coordinated by staff of EPA's National Center for
 13      Environmental Assessment in Research Triangle Park (NCEA-RTP). NCEA-RTP scientific
 14      staff,  together with experts from other EPA/ORD laboratories and academia, contributed to
 15      writing of document chapters; and earlier drafts of this document were reviewed by experts from
 16      federal and state government agencies, academia, industry, and non-governmental organizations
 17      (NGOs) for use by EPA in support of decision making on potential public health and
 18      environmental risks of ambient PM. The document describes the nature, sources, distribution,
 19     measurement, and concentrations of PM  in outdoor (ambient) and indoor environments.  It also
 20     evaluates the latest data on human exposures to ambient PM and consequent health effects in
 21      exposed human populations (to support decision making regarding primary, health-related PM
 22     NAAQS).  The document also evaluates ambient PM environmental effects on vegetation and
 23      ecosystems, visibility, and man-made materials, as well as atmospheric PM effects on climate
 24      change processes associated with alterations in atmospheric transmission of solar radiation or its
 25      reflectance from the Earth's surface or atmosphere (to support decision making on secondary
 26      PM NAAQS).
 27          The NCEA of EPA acknowledges the contributions provided by authors, contributors, and
28      reviewers and the diligence of its staff and contractors in the preparation of this document.
29
       June 2003                                 I_H{        DRAFT-DO NOT QUOTE OR CITE

-------
             Air Quality Criteria for Participate Matter
                            VOLUME I
EXECUTIVE SUMMARY

1.   INTRODUCTION
2.   PHYSICS, CHEMISTRY, AND MEASUREMENT OF PARTICULATE
    MATTER [[[ 2~l
    APPENDIX 2A:  Techniques for Measurement of Semivolatile Organic
                  Compounds ....................................  2A.-1
    APPENDIX 2B:  Analytical Techniques ............................. 2B-1

3.   CONCENTRATIONS, SOURCES, AND EMISSIONS OF ATMOSPHERIC
    PARTICULATE MATTER  ......................................... 3^
    APPENDIX 3A:  Spatial and Temporal Variability of the Nationwide
                  AIRS PM2.5 and PM10_2.S Data Sets  ....... ...........  3A-1
    APPENDIX 3B:  Aerosol Composition Data from the Speciation
                  Network ........................................ 3B~1
    APPENDIX 3C:  Organic Composition of Particulate Matter ............ 3C-1
    APPENDIX 3D:  Composition of Particulate Matter Source Emissions ----  3D-1
    APPENDIX 3E:  Variability Observed in PM25 and PM10_2.5 Concentrations
                  at IMPROVE Sites ............................... 3E-1

4.   ENVIRONMENTAL EFFECTS OF AIRBORNE PARTICULATE
    MATTER [[[ 4-1
    APPENDIX 4A:  Colloquial and Latin Names .......................  4A-1

3.   CONCENTRATIONS, SOURCES, AND EMISSIONS OF
    ATMOSPHERIC PARTICULATE MATTER ........................... 3-1

4.   ENVIRONMENTAL EFFECTS OF AIRBORNE PARTICULATE
    MATTER [[[ 4-1


-------
               Air Quality Criteria for Particulate Matter
                                 (cont'd)


                               VOLUME II


 6.   DOSIMETRY OF PARTICULATE MATTER	6-1

 7.   TOXICOLOGY OF PARTICULATE MATTER IN HUMANS AND
      LABORATORY ANIMALS 	7_1

 8.   EPIDEMIOLOGY OF HUMAN HEALTH EFFECTS ASSOCIATED
      WITH AMBIENT PARTICULATE MATTER	8-1
      APPENDIX 8A:  Short-Term PM Exposure-Mortality Studies:
                    Summary Tables	 8A-1
      APPENDIX 8B: Particulate Matter-Morbidity Studies: Summary Tables ... 8B-1

 9.   INTEGRATIVE SYNTHESIS	9_1
      APPENDIX 9A: Key Quantitative Estimates of Relative Risk for
                   Particulate Matter-Related Health Effects Based on
                   Epidemiologic Studies of U.S. and Canadian Cities
                   Assessed in the 1996 Particulate Matter Air Quality
                   Criteria Document	 9A-1
June 2003                          I-v    DRAFT-DO NOT QUOTE OR CITE

-------
                              Table of Contents
List of Tables	  l'xm
List of Figures 	'
Authors, Contributors, and Reviewers	1-xxxi
U.S. Environmental Protection Agency Project Team for Development of Air Quality
        Criteria for Particulate Matter	I-xxxix
Abbreviations and Acronyms  	 i"xlu

EXECUTIVE SUMMARY	E"1

1.    INTRODUCTION	}"}
      1.1    LEGISLATIVE REQUIREMENTS	M
      1.2    HISTORY OF PREVIOUS PM CRITERIA AND NAAQS REVIEWS	1-2
            1.2.1   The 1997PMNAAQSRevision 	I-3
            1.2.2   Coordinated Particulate Matter Research Program 	1-6
      1.3    CURRENT PM CRITERIA AND NAAQS REVIEW	1-9
            1.3.1   Key Milestones	*~9
            1.3.2   Methods and Procedures for Document Preparation	1-13
            1.3.3   Approach  	}~}5
      1.4    DOCUMENT ORGANIZATION AND CONTENT	1-15
      REFERENCES  	M7

2.    PHYSICS, CHEMISTRY, AND MEASUREMENT OF PARTICULATE
      MATTER	^"i
      2.1     PHYSICS AND CHEMISTRY OF PARTICULATE MATTER	2-2
             2.1.1    Basic Concepts	2"2
             2.1.2    Physical Properties and Processes 	2~3
                     2.1.2.1   Definitions of Particle Diameter	2-3
                     2.1.2.2   Aerosol Size Distributions 	2-5
                     2.1.2.3   Ultrafme Particles	2-20
             2.1.3    Chemistry of Atmospheric Particulate Matter  	2-28
                     2.1.3.1   Chemical Composition and Its Dependence on
                             Particle Size	2'29
                     2.1.3.2   Primary and Secondary Particulate Matter	2-29
                     2.1.3.3   Particle-Vapor Partitioning	2-31
                     2.1.3.4   Atmospheric Lifetimes and Removal Processes	2-34
             2 1 4    Comparison of Fine and Coarse Particles	2-36
      2.2    MEASUREMENT OF PARTICULATE MATTER	2-38
             2.2.1    Particle Measurements of Interest	2~38
             2.2.2    Issues in Measurement of Particulate Matter	2-40
                     2.2.2.1    Artifacts Due to Chemical Reactions	2-41
                     2.2.2.2    Treatment of Semivolatile Components of
                              Particulate Matter	2'42
                     2.2.2.3    UpperCutPoint 	2~43

 June 2003                              I-vi       DRAFT-DO NOT QUOTE OR CITE

-------
                                 Table of Contents
                                       (cont'd)
                      2.2.2.4    Cut Point for Separation of Fine and Coarse
                                Particulate Matter	2-44
                      2.2.2.5    Treatment of Pressure, Temperature, and Relative
                                Humidity	2-47
                      2.2.2.6    Time Resolution	2-48
                      2.2.2.7    Accuracy and Precision  	2-49
              2.2.3    Measurement of Semivolatile Particulate Matter	2-51
                      2.2.3.1    Particle-Bound Water	2-52
                      2.2.3.2    Nitrate and Organic Species	2-57
                      2.2.3.3    Continuous Measurement of Semivolatile and
                                Nonvolatile Mass 	2-64
              2.2.4    U. S. Environmental Protection Agency Monitoring Methods 	2-68
                      2.2.4.1    The Federal Reference Methods for Measurement
                                of Equilibrated Mass for PM10, PM2 5, and PMi0.2.5	2-68
              2.2.5    Speciation Monitoring  	2-82
              2.2.6    Inorganic Elemental Analyses 	2-85
              2.2.7    Elemental and Organic Carbon in Particulate Matter	2-86
              2.2.8    Ionic Species	2-88
              2.2.9    Continuous Monitoring	2-88
              2.2.10   Measurements of Individual Particles	2-90
              2.2.11   Low Flow Filter Samples for Multiday Collection of
                      Particulate Matter	2-92
      2.3     SUMMARY AND KEY POINTS	2-93
              2.3.1    Atmospheric Physics and Chemistry of Particles	2-93
              2.3.2    Measurement of Atmospheric Particles  	2-96
              2.3.3    Key Points	2-100
      REFERENCES 	2-102

APPENDIX 2A.  Techniques for Measurement of Semivolatile Organic Compounds	 2A-1
APPENDIX 2B.  Analytical Techniques 	2B-1
      2B.1  INORGANIC ELEMENTS  	2B-1
            2B.1.1    Energy Dispersive X-Ray Fluorescence	2B-1
            2B.1.2    Synchrotron Induced X-Ray Fluorescence	2B-1
            2B.1.3    Proton (or Particle) Induced X-Ray Emission	2B-2
            2B.1.4    Proton Elastic Scattering Analysis	2B-3
            2B.1.5    Total Reflection X-Ray Fluorescence 	2B-3
            2B.1.6    Instrumental Neutron Activation Analysis	2B-4
            2B.1.7    Atomic Absorption Spectrophotometry	2B-5
            2B.1.8    Inductively Coupled Plasma with Atomic Emission
                     Spectroscopy  	2B-6
            2B.1.9    Inductively Coupled Plasma with Mass Spectroscopy 	2B-6
            2B.1.10   Scanning Electron Microscopy 	2B-7

June 2003                               I-vii        DRAFT-DO NOT QUOTE OR CITE

-------
                              Table of Contents
                                   (cont'd)
     2B.2 ORGANIC AND INORGANIC CARBON ............................ 2B-8
     2B.3 CONTINUOUS METHODS  ...................................... 2B-17
          2B.3.1    Continuous Measurement of Mass .......................... 2B-17
          2B.3.2    Continuous Measurement of Organic and/or Elemental Carbon . . . 2B-21
          2B.3.3    Continuous Measurements of Nitrate and Sulfate .............. 2B-24
     2B.4 OTHER CONTINUOUS MEASUREMENTS ......................... 2B-26

3.    CONCENTRATIONS, SOURCES, AND EMISSIONS OF ATMOSPHERIC
     PARTICULATE MATTER .............................................. 3"1
     3 1   INTRODUCTION   ............................................... 3
     3~2  PATTERNS AND TRENDS IN AMBIENT PM CONCENTRATIONS ...... 3-2
          3.2.1     Seasonal Variability in PM Concentrations .................... 3-13
          3.2.2     Diurnal (Circadian) Variability in PM Concentrations ........... 3-30
          3.2.3     Relations Among Particulate Matter in Different Size Fractions  . . . 3-32
          3.2.4     Relations Between Mass and Chemical Component
                   Concentrations .......................................... 3-34
          325     Spatial Variability in Particulate Matter and its Components ...... 3-41
     3.3  SOURCES OF PRIMARY AND SECONDARY PARTICULATE
          MATTER [[[ 3'59
          3.3.1     Chemistry of Secondary PM Formation ....................... 3-62
          3.3.2     Source Contributions to Ambient PM Determined by Receptor
                   Models ................................................ 3'72
          3.3.3     Background Concentrations of PM in the United States .......... 3-82
          3.3.4     Emissions Estimates for Primary Particulate Matter, and
                   Precursors to Secondary Particulate Matter (SO2, NOX, VOCs,
                   andNH3) in the United States  .............................. 3-91
          3.3.5     Uncertainties of Emissions Inventories ....................... 3-97
     3.4   SUMMARY AND KEY CONCLUSIONS ............................ 3-99
     REFERENCES [[[ 3~105

APPENDIX 3A: Spatial and Temporal Variability of the Nationwide AIRS PM25
               and PM10_25 Data Sets
      REFERENCES  [[[ 3A"3
 APPENDIX 3B: Aerosol Composition Data from the Speciation Network ............. 3B-1
      REFERENCES  [[[ 3B'3
 APPENDIX 3C: Organic Composition of Particulate Matter ........................ 3C-1
      REFERENCES  [[[ 3C'14
 APPENDIX 3D: Composition of Particulate Matter Source Emissions  ............... 3D-1

-------
                                 Table of Contents
                                       (cont'd)
 4.    ENVIRONMENTAL EFFECTS OF AIRBORNE PARTICULAR MATTER        4-1
       4.1     INTRODUCTION	      4_!
       4.2     EFFECTS OF AMBIENT AIRBORNE PM ON VEGETATION AND	
              NATURAL ECOSYSTEMS	4_1
              4.2.1    Ecological Attributes 	4_2
              4.2.2    Ecosystem Exposures - Particle Deposition	4-4
                      4.2.2.1    Fine and Coarse Particulate Matter	4-5
                      4.2.2.2    Diversity of Deposition Modes	4-8
                      4.2.2.3    Magnitude of Deposition  	4.44
              4.2.3    Assessment of Atmospheric PM Deposition Effects  	4-55
                      4.2.3.1    Effects on Vegetation and Ecosystems	4-60
                      4.2.3.2    Ecosystem Response to Stresses	4-79
              4.2.4    Urban Ecosystems	    4-149
       4.3     AIRBORNE PARTICLE EFFECTS ON VISIBILITY	'........'.'.  4-153
              4.3.1    Introduction	4-153
              4.3.2    Factors Affecting Atmospheric Visibility	4-155
                      4.3.2.1    Optical Properties of the Atmosphere and
                               Atmospheric Particles	4-155
                      4.3.2.2    Relative Humidity Effects on Particle Size and
                               Light-Scattering Properties	4-163
              4.3.3    Relationships Between Particles and Visibility 	4-166
              4.3.4    Photographic Modeling of Visibility Impairment  	4-174
              4.3.5    Visibility Monitoring Methods and Networks  	4-175
              4.3.6    Visibility Impairment: Trends and Current Conditions	4-178
                      4.3.6.1    Trends in Visibility Impairment	4-178
                      4.4.6.2    Current Conditions	4-187
              4.3.7    Economics of Particulate Matter Visibility Effects  	     4-187
      4.4     PARTICULATE MATTER EFFECTS ON MATERIALS	4-191
              4.4.1     Corrosive Effects of Particles and Sulfur Dioxide on
                     Man-Made Surfaces  	4-192
                     4.4.1.1    Metals	4.192
                     4.4.1.2    Painted Finishes  	4-194
                     4.4.1.3    Stone and Concrete	4-197
             4.4.2    Soiling and Discoloration of Man-Made Surfaces	4-202
                     4.4.2.1    Stones and Concrete	4-203
                     4.4.2.2    Household and Industrial Paints	4-204
      4.5     ATMOSPHERIC PARTICULATE MATTER, CLIMATE CHANGE,
             AND EFFECTS ON SOLAR UVB RADIATION TRANSMISSION	4-205
             4.5.1    Atmospheric Particle Interactions with Solar and Terrestrial
                     Radiation Related to Climate Change	4-206
                     4.5.1.1     The Projected Impacts of Global Climate Change	4-206
June 2003                                I_ix       DRAFT-DO NOT QUOTE OR CITE

-------
                               Table of Contents
                                     (cont'd)
                    4.5.1.2    Airborne Particle Relationships to Global
                              Warming and Climate Change ................... 4-209
            4.5 .2   Atmospheric Paniculate Matter Effects on the Transmission
                    of Solar Ultraviolet Radiation Transmission: Impacts on
                    Human Health and the Environment ........................ 4-220
                    4.5.2.1    Potential Effects of Increased Ultraviolet Radiation
                              Transmission ................................. 4-220
                    4.5.2.2    Airborne Particle Effects on Atmospheric
                              Transmission of Solar Ultraviolet Radiation  ........ 4-223
     4.6    SUMMARY AND KEY CONCLUSIONS .......................... 4-227
            4.6.1   Particulate Matter Effects on Vegetation and Ecosystems ....... 4-227
            4.6.2   Paniculate Matter-Related Effects on Visibility ............... 4-235
            4.6.3   Particulate Matter-Related Effects on Materials ............... 4-236
            4.6.4   Effects of Atmospheric Particulate Matter on Global Warming
                    Processes and Transmission of Solar Ultraviolet Radiation ...... 4-237
     REFERENCES ......................... ............................. 4"240

APPENDIX 4A:  Colloquial and Latin Names ...................................  4A-1
5.    HUMAN EXPOSURE TO PARTICULATE MATTER AND ITS
      CONSTITUENTS
      5.1     INTRODUCTION
             5.1.1    Purpose ................................................ 5-1
             5.1.2    Particulate Matter Mass and Constituents  ..................... 5-2
             5.1.3    Relationship to Past Documents ............................. 5'3
             5.1.4    Chapter Structure ........................................ 5-4
      5.2    BASIC CONCEPTS OF EXPOSURE ................................ 5~5
             5.2.1    The Concept of Exposure .................................. 5'5
             5.2.2    Components of Exposure .................................. 5-6
             5.2.3    Quantification of Exposure ................................. 5~9
             5.2.4    Methods To Estimate Personal Exposure ..................... 5-10
                     5.2.4.1    Direct Measurement Methods  .................... 5-1 1
                     5.2.4.2    Indirect or Modeling Methods  ................ ---- 5-12
                     5.2.4.3    Methods for Estimating Personal Exposure to Ambient
                               Particulate Matter .............................. 5'20
      5.3    SUMMARY OF PARTICULATE MATTER MASS DATA . . ........... 5-26
             5.3.1    Types of Particulate Matter Exposure Measurement Studies ...... 5-26
             5.3.2    Available Data  ......................................... 5'27
                     5.3.2.1    Personal Exposure Data  ......................... 5-27
                     5.3.2.2    Microenvironmental Data ........................ 5-32
                     5.3.2.3    Traffic-Related Microenvironments ................ 5-40
 June 2003                                I-x         DRAFT-DO NOT QUOTE OR CITE

-------
                                Table of Contents
                                       (cont'd)
                     5.3.2.4    Reanalyses of Previously-Reported Particulate Matter
                               Exposure Data	5-40
             5.3.3    Factors Influencing and Key Findings on Particulate Matter
                     Exposures 	5-43
                     5.3.3.1    Relationship of Personal/Microenvironmental
                               Particulate Matter with Ambient Particulate Matter	5-43
                     5.3.3.2    Factors That Affect Relationships Between Personal
                               Exposure and Ambient PM	5-63
                     5.3.3.3    Effect of Ambient Sources on Exposures to Particulate
                               Matter	5-84
                     5.3.3.4    Correlations of Particulate Matter with Other
                               Pollutants  	5-87
      5.4    SUMMARY OF PARTICULATE MATTER CONSTITUENT DATA .... 5-91
             5.4.1    Introduction	5-91
             5.4.2    Monitoring Studies That Address Particulate Matter
                     Constituents  	5-91
             5.4.3    Key Findings	5-91
                     5.4.3.1    Correlations of Personal and Indoor Concentrations
                               with Ambient Concentrations of Particulate Matter
                               Constituents  	5-91
             5.4.4    Factors Affecting Correlations Between Ambient Measurements
                     and Personal or Microenvironmental Measurements of
                     Particulate Matter Constituents	5-102
             5.4.5    Limitations of Available Data	5-103
      5.5    IMPLICATIONS OF USING AMBIENT PARTICULATE MATTER
             CONCENTRATIONS IN TOXICOLOGICAL AND
             EPIDEMIOLOGICAL STUDIES OF PARTICULATE MATTER
             HEALTHEFFECTS  	5-104
             5.5.1    Toxicology  		:	5-104
             5.5.2    Potential Sources of Error Resulting from Using
                     Ambient Particulate Matter Concentrations in Epidemiological
                     Analyses 	'	5-104
                     5.5.2.1    Associations Between Personal Exposures and
                               Ambient Particulate Matter Concentrations	5-107
                     5.5.2.2    Role of Compositional Differences in Exposure
                               Characterization for Epidemiology	5-111
                     5.5.2.3    Role of Spatial Variability in Exposure
                               Characterization for Epidemiology	5-112
             5.5.3    Analysis of Exposure Measurement Error Issues in Particulate
                     Matter Epidemiology  	5-113
                     5.5.3.1    Time-Series Analyses	5-113
                     5.5.3.2    Studies of Chronic Effects 	5-114

June 2003                               I-xi        DRAFT-DO NOT QUOTE OR CITE

-------
                             Table of Contents
                                  (cont'd)
           5.5.4    Conclusions from Analysis of Exposure Measurement Errors
                   on Particulate Matter Epidemiology	5-116
     5.6    SUMMARY OF OBSERVATIONS AND LIMITATIONS	5-118
     REFERENCES 	5-128
June 2003                            I-xii       DRAFT-DO NOT QUOTE OR CITE

-------
                                     List of Tables

  Number                                                                         Page

  1-1       Key Scientific Uncertainties Related to the Source-To-Response Framework ____  1-8

  1-2       Research Topics and Questions Recommended by National Research Council
           to be Addressed by Expanded Particulate Matter Research Program .......... 1-10

  1-3       Schedule for Development of the Current Revised Particulate Matter Air
           Quality Criteria Document                                                _
 2-1       Comparison of Ambient Particles, Fine Particles (Ultrafine Plus
           Accumulation-Mode), and Coarse Particles ............................. 2-37

 2-2       Particulate Matter Components/Parameters of Interest for Health, Ecological,
           or Radiative Effects; for Source Category Apportionment Studies; or for
           Air Quality Modeling Evaluation Studies ............................... 2-39

 2-3       Summary of Sensitivity Studies of WINS Impactor Performance ............ 2-73

 2-4       PM2 5 Samplers Currently Designated as FRMs for PM2 5 Mass
           Concentrations ................................ ................     2-74

 2-5       Measurement Methods for Inorganic Elements ........................... 2-86

 2-6       Methods for Continuous Measurement of PM Mass, Components, etc ......... 2-89

 3-1       Distribution of Ratios of PM25 to PM10 and Correlations Between PM2 5
           and PM10, PM2 5 and PM10.2J, and PM]0.2 5 and PM]0 Found at Collocated
           Monitoring Sites in Seven Aerosol Characteristic Regions in 1999 ........... 3-33

 3-2       Concentrations of PM2 5, PM10_25, and Selected Elements in the PM2 5 and
           pHo-2.5 Size Ranges with Standard Deviations and Correlations Between
           Elements and PM2 5 Mass in Philadelphia, PA  ........................... 3.35

 3-3        Concentrations of PM25, PM10.25> and Selected Elements in the PM25 and
           PMio-2.5 Size Range with Standard Deviations and Correlations Between
           Elements and PM25 and PMI0.2 5 Mass in Phoenix, AZ ..................... 3-36

 3-4a-d     Measures of the Spatial Variability of PM25 Concentrations Within Selected
          Metropolitan Statistical Areas (MSAs) ................................. 3_42

 3-5       Measures of the Spatial Variability of PM]0_2 5 Concentrations Within Selected
          Metropolitan Statistical Areas (MSAs)  ................................. 3.49
June 2003                               I_xiii       DRAFT-DO NOT QUOTE OR CITE

-------
                                  List of Tables
                                      (cont'd)
Number
3-6       Correlation Coefficients for Spatial Variation of PM2.5 Mass and Different
          Sources for Pairs of Sampling Sites in the South Coast Air Basin (1986)  ......  3-57

3-7       Correlation Coefficients for Spatial Variation of PM25 Mass and Different
          Components for Pairs of Sampling Sites in Philadelphia (1994) .............  3-57

3-8       Constituents of Atmospheric Particles and their Major Sources ..............  3-60

3-9       Receptor Model Source Contributions to PM2 5 ............. - .............  3'76

3-10      Receptor Model Source Contributions to PMio ...........................  3'77

3-1 1      Emissions of Primary PM25 by Various Sources in 1999 ...................  3-93

3-12      Emissions of Precursors to Secondary PM25 Formation by Various Sources
          in 1999  [[[  3"94

3-13      Nationwide Changes in Ambient Concentrations and Emissions of PM]0 and
          Gaseous Precursors to Secondary Particulate Matter from 1992 to 2001  ....... 3-96

3A-1     Performance Metrics for PM25 from Collocated Samplers ................ 3A-31

3B-1     PM2 5 Speciation Samplers by Location: Sites Selected for PM CD
          Summary [[[
 3B-2a     Burlington, VT Summary Data

 3B-2b     Philadelphia, PA Summary Data
                                                                                QT> 7
 3B-2c     Atlanta, GA Summary Data  ......................................... JD~'

                                                                                •3D O
 3B-2d     Detroit, MI Summary Data .......................................... DD~°

                                                                                •3D Q
 3B-2e     Chicago, IL Summary Data  .........................................


 3B-2f     ST. Louis, MI Summary Data ....................................... 3B-1°

                                                                               •3D 1 1
 3B-2g     Houston, XX Summary Data ....... . ................................ JD'1 i

                                                                               •5D 17
 3B-2h     Minneapolis, MN Summary Data  .................................... :>E>'1Z-

                                                                               •5D 1 "2

-------
                                   List of Tables
                                       (cont'd)

 Number                                                                        page

 3B-2J     Phoenix, AZ Summary Data 	3B-14

 3B-2k     Seattle, WA Summary Data  	3B-15

 3B-21     Sacramento, CA Summary Data 	3B-16

 3B-2m    Riverside-Rubidoux, CA Summary Data  	3B-17

 3B-3      Blank Corrections for Elemental Carbon, Organic Carbon, and Total Carbon
           in the Speciation Network	-	3B-18

 3C-1      Particulate Organic and Elemental Carbon Concentrations Based on Studies
           Published after 1995  	3C-2

 3C-2      Particulate Organic Compound Concentrations Based on Studies Published
           after 1990 at Selected Sites  	3C-5

 3D-1      Average Abundances of Major Elements in Soil and Crustal Rock	 3D-2

 3D-2      Summary of Particle-Phase Organic Compounds Present in Fine Particle
           Road Dust Sample	 3D_5

 3D-3      Composition of Fine Particles Released by Various Stationary Sources in
          the Philadelphia Area	 3D-6

 3D-4a    Organic and Elemental Carbon Fractions of Diesel and Gasoline Engine
          Particulate Matter Exhaust	  3D-12

 3D-4b    Contribution of Organic Carbon to Particulate Matter Carbon Emissions
          in Motor Vehicle Exhaust Collected from Vehicles Operated on Chassis
          Dynamometers	  3D-13

 3D-5     Emission Rates for Constituents of Particulate Matter from Gasoline and
          Diesel Vehicles 	  3D-14

 3D-6     Summary of Particle-Phase Organic Compounds Emitted from Motor
          Vehicles	  3D-16

 3D-7     Mass Emissions, Organic Carbon, and Elemental Carbon Emissions from
          Residential Combustion of Wood 	  3D-19
June 2003                               I-xv        DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Tables
                                       (cont'd)
Number
3D-8      Summary of Particle-Phase Organic Compounds Emitted from the
          Combustion of Wood in Fireplaces  ..................................  3D-20

3D-9      Mean Aerosol Composition at Tropical Site (Sriwijaya University, Sumatra,
          Indonesia) Affected Heavily by Biomass Burning Emissions ......... - ----  3D-22

3E-1      The Range of Annual Mean PM25, PM10.2.5, and PM10 Concentrations at
          IMPROVE Particulate Matter Monitoring Sites .......................... 3E-3

3E-2      The Range of 24-Hour PM25, PM10_25, and PM10 Average Concentrations at
          the 90th Percentile Level at IMPROVE Particulate Matter Monitoring Sites ---- 3E-4

4-1       Essential Ecological Attributes and Reporting Categories  ...................  4-3

4-2       Types and Determinants of Particulate Deposition and Impact to Vegetation ... 4-10

4-3       Relative Importance of Wet, Dry, Particulate, and Total Deposition to Three
          Forest Sites [[[ 4"12

4-4       Reported Mean Deposition Velocities for Sulfate, Chlorine, Nitrate, and
          Ammonium and Ion-Containing Particles ............................... 4-37

4-5       Representative Empirical Measurements of Deposition Velocity for
          Particulate Deposition  .............................................. 4"38
 4-6      Reported Mean Deposition Velocities for Potassium, Sodium, Calcium, and
          Magnesium Base-Cation-Containing Particles  ........................... 4-38

 4-7      Mean Annual Nitrogen Deposition from Fine and Coarse Particles
          Compared to Total Nitrogen Deposition from all Sources to a Variety
          of Forest Ecosystems ............................................... 4~46

 4-8      Mean Annual Sulfate Deposition from Fine and Coarse Particles
          Compared to Total Sulfur Deposition from all Sources to a Variety of
          Forest Ecosystems ................................................. 4-4 '

 4-9      Mean Annual Base Cation Deposition from Fine and Coarse Particles
          Compared to Total Base Cation Deposition from all Sources to a Variety
          of Forest Ecosystems  ............................................... 4-48

 4-10     Mean Particle Size, Deposition Rates, and Derived Deposition Velocities for
          Heavy Metal Deposition to the Upper Canopy of an Upland Oak Forest ........ 4-52


-------
                                     List of Tables
                                         (cont'd)

  Number                                                                          Page

  4-11       Total Heavy Metal Deposition to Temperate Latitude Forests	4-52

  4-12       Primary Goods and Services Provided by Ecosystems	4-58

  4-13       Ecosystem Functions Impacted by Air Pollution Effects on Temperate Forest
            Ecosystems	        4,34

  4-14       Nitrogen-Saturated Forests in North America, Including Estimated N Inputs
            and Outputs 	            4-102

  4-15       Visibility Measurement Techniques 	4-172

  4-16       Residential Visibility Contingent Valuation Study Results	4-190

  4-17       Residential Visibility Valuation Study Results for Los Angeles and
            San Francisco	         4-191

 4-18       Corrosive Effects of Paniculate Matter and Sulfur Dioxide on Metals	4-195

 4-19      Corrosive Effects of Paniculate Matter and Sulfur Dioxide on Stone  	4-198

 4-20      Examples of Impacts Resulting from Projected Changes in Extreme
           Climate Events	          4-208

 4-21      Effects of Reactive Nitrogen	    4-231

 5-1        Types of Paniculate Matter Used in Exposure and Concentration Variables	5-7

 5-2        Activity Pattern Studies Included in the Consolidated Human Activity
           Database	               5_jc

 5-3        Personal Exposure Models for Paniculate Matter	5_16

 5-4        Summary of Recent PM Personal Exposure Studies	5-28

 5-5        Summary of Recent Microenvironmental PM Measurement Studies	5-33

 5-6        Papers Reporting Reanalyses of Paniculate Matter Exposure Studies	'5-41

 5-7       Personal Monitoring Studies for Paniculate Matter: Measured Concentrations
          and Correlation Coefficients 	5.45



June 2003                               I_xvii       DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Tables
                                       (cont'd)
Number
5-8       Mean Concentration for PM Mass Reported for the Baltimore
          (Williams et al., 2000a,b,c) and the Fresno (Evans et al., 2000) Studies 	5-55

5-9       Daily-average Correlation Coefficients Reported for the Baltimore
          (Williams et al., 2000a,b,c) and the Fresno (Evans et al., 2000) Studies 	5-55

5-10      Regression Analysis Reported for Indoor/Outdoor Relationships for PM25
          in the Baltimore (Williams et al., 2000a,b,c) and Fresno (Evans et al., 2000)
          Studies	5"56

5-11      Volume Mean Diameter and Maximum PM25 Concentrations of Indoor
          Particle Sources	5"81

5-12      Correlations Between Personal PM2 5 and Ambient Pollutant Concentrations  ... 5-88

5-13      Correlations Between Hourly Personal PM2.5 and Gaseous Pollutants	5-90

5-14      Studies That have Measured Particulate Matter Constituents in Personal
          Exposure Samples 	**"

5-15      Studies That Have Measured Particulate Matter Constituents in
          Microenvironmental Samples  	5"94

 5-16      Summary Statistics for Personal, Indoor, and Outdoor Concentrations of
          Selected Aerosol Components in Two Pennsylvania Communities	5-96

 5-17      Regression Analysis of Indoor Versus Outdoor Concentrations	5-98

 5-18      Mixed Model Analysis of Personal Versus Outdoor Concentrations	5-100

 5-19      Regression Analysis of Indoor Versus Outdoor Concentrations	5-100

 5-20     Regression Analysis of Indoor Versus Outdoor Concentrations	5-101
  June 2003                               I-xviii       DRAFT-DO NOT QUOTE OR CITE

-------
                                     List of Figures

 Number                                                                           page

  1-1       A general framework for integrating particulate-matter research	1-7

 2-1       Distribution of coarse, accumulation, and nuclei mode particles by three
           characteristics: (a) number, (b) surface area, and (c) volume for the grand
           average continental size distribution  	2-8

 2-2       Particle size distributions by number, (a) number concentrations are shown
           on a logarithmic scale to display the wide range by site and size;  (b) number
           concentrations for the average urban distribution are shown on a  linear scale	2-9

 2-3       Size distribution by volume (a) for the averaged rural and urban-influenced
           rural number distributions shown in Figure 2-2a and a distribution from south
           central New Mexico, and (b) for the averaged urban and freeway-influenced
           urban number distributions shown in Figure 2-2a	2-10

 2-4       Volume size distribution, measured in traffic, showing fine and coarse
           particles and the nuclei and accumulation modes of fine particles	2-11

 2-5       Submicron number size distribution observed in a boreal forest in Finland
           showing the tri-modal structure of fine particles	2-12

 2-6       Specified particle penetration (size-cut curves) through an ideal
           (no-particle-loss) inlet for five different size-selective sampling criteria	2-15

 2-7       An idealized distribution of ambient particulate matter showing fine-mode
           particles and coarse-mode particles and the fractions collected by
           size-selective samplers	2-16

 2-8       Comparison of penetration curves for two PM10 beta gauge samplers using
           cyclone inlets	    2-18

 2-9       Typical engine exhaust size distribution	2-23

 2-10       Number size distributions showing measurement of a nucleation burst made
           in a boreal forest in Finland	2-25

 2-11       Examples of the measured one hour average particle number size distributions
           and the log normal fits to the modes of the data  	2-26

 2-12(a-f)   Fitted multi-model particle size distribution at different sampling distances from
           Freeway 405 (a) 30 m downwind, (b) 60 m downwind, (c) 90 m downwind,
           (d) 150 m downwind, (e) 300 m downwind, (f) 300 m upwind  	2-27
June 2003                                I_xix        DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Figures
                                       (cont'd)
Number
2-12(g)   Ultrafme particle size distribution at different sampling locations near the
          405 Freeway in Los Angeles, CA	2"28

2-13      Particle growth curves showing fully reversible hygroscopic growth of sulfuric
          acid particles, deliquescent growth of ammonium sulfate particles at the
          deliquescent point, reversible hygroscopic growth of ammonium sulfate
          solution droplets at relative humidity relative humidity greater than 80%,
          and hysteresis until the crystallization point is reached  	2-34

2-14      Theoretical predictions and experimental measurements of growth of
          NH4HSO4 particles at relative humidity between 95 and 100% 	2-35

2-15      Schematic showing major nonvolatile and semivolatile components of PM25 ... 2-42

2-16      Particulate matter concentrations in Spokane, WA, during the August 30,
          1996 dust storm	2'46

2-17      Aerosol water content expressed as a mass percentage, as a function relative
          humidity	2~55

2-18      Amount of ammonium nitrate volatilized from Teflon filters, expressed as a
          percentage of the measured PM2 5 mass, for the SCAQS and CalTech studies,
          for spring and fall sampling periods 	2-58

2-19      Average concentration of nonvolatile and semivolatile PM components in
          three cities 	2~65

2-20      Comparison of mass measurements with collocated RAMS, PC-BOSS,
          FRM PM2 5 sample, and a conventional TEOM monitor	2-67

2-21      Schematic  diagram of the sample collection portion of the PM25 FRM
          sampler	2-70

 2-22     Schematic  view of the final design of the WINS	2~71

 2-23     Evaluation of the final version of the WINS	2~72

 2-24      Schematic diagram showing the principle of virtual impaction  	2-80

 2-25      Size distribution of particles divided by chemical classification into organic,
           marine, and crustal	2"91
 June 2003                                I-xx        DRAFT-DO NOT QUOTE OR CITE

-------
                                     List of Figures
                                         (cont'd)
  Number
  2B-1      This thermogram, for a sample containing rock dust and diesel exhaust,
            shows three traces that correspond to temperature, filter transmittance,'
            and FID detector response	;	      2B-10

  2B-2      Examples of thermograms obtained by (a) the IMPROVE protocol and by
            (b) the NIOSH protocol	2B-12

  3-la      1999-2001 county-wide average annual mean PMIO concentrations for
            counties with PM10 monitors	   3.7
 3-lb
1999-2001 highest county-wide 98th percentile 24-h average PM,0
           concentrations for counties with PM10 monitors		3.7

 3-2       Nationwide trend in ambient PMIO concentration from 1992 through 2001 	3-8

 3-3       Trend in PM10 annual mean concentrations by EPA region, 1992
           through 2001 	    3_9

 3-4a      1999-2001 county-wide average annual mean PM2 5 concentrations for
           counties with PM25 monitors	3_10

 3-4b      1999-2001 highest county-wide 98thpercentile 24-h average PM25
           concentrations for counties with PM25 monitors	3_10

 3-5       Collection of annual distribution of 24-h average PM25 concentrations
           observed in U.S. and Canadian health studies conducted during the 1980s
           and early 1990s 	                3_12
 3-6a      1999-2000 estimated county-wide average annual mean PM10_25
           concentrations for counties with collocated PM25 and PM10 monitors	3-14
 3-6b      1999-2000 estimated county-wide highest 98th percentile 24-h PMIO_2 5
           concentrations for counties with collocated PM2 5 and PM30 monitors	3-15

 3-7a,b     Quarterly distribution of 24-h average PM2 5 concentrations for selected
           monitors in the (a) Philadelphia, PA and (b) Cleveland, OH  	3-16

 3-7c,d     Quarterly distribution of 24-h average PM2 5 concentrations for selected
           monitors in the (c) Dallas, TX and (d) Los Angeles, CA MS As	3-17

3-8        Seasonal concentrations of PM25 and PMIO measured in the four
          MAACS cities .;	                  3_20


June 2003                                I.xxi        DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Figures
                                       (cont'd)
Number
3-9a b    Quarterly distribution of 24-h average PM10_2.5 concentrations for selected
          sites in the (a) Cleveland, OH; (b) Dallas, TX MSAs	3-22

3-9c      Quarterly distribution of 24-h average PM10_2.5 concentrations for selected
          sites in the Los Angeles, CA MSA 	3-23

3-10      Frequency distribution of 24-h average PM25 concentrations measured at
          the Presbyterian home monitoring site in southwestern Philadelphia from
          1992 to 1995	3"26

3-11      Concentrations of 24-h average PM2 5 measured at the EPA site in
          Phoenix, AZ from 1995 to 1997 	3-27

3-12     Frequency distribution of 24-h average PM2 5 concentrations measured at the
          EPA site in Phoenix, AZ from 1995 to 1997	3-27

3-13     Frequency distribution of 24-h average PM2.5 measurements obtained
          from all California Air Resources Board dichotomous sampler sites from
           1989to 1998  	3~28

 3-14     Frequency distribution of 24-h average PMI0.2.5 concentrations obtained
          from all California Air Resource Board Dichotomous sampler sites from
           1989 to 1998  	3'28

 3-15      Concentrations of 24-h average PM25 measured at the Riverside-Rubidoux
           site from 1989 to 1998	3'29

 3-16      Frequency distribution of 24-h average PM25 concentrations measured at the
           Riverside-Rubidoux site from 1989 to 1994	3-29

 3-17      Intraday variability of hourly average PM25 concentrations across the
           United States 	3~31

 3-18      Occurrence of differences between pairs of sites in three metropolitan
           statistical areas	

 3-19      Intersite correlation coefficients for PM25, PM10, and PM10_25 	3-52

 3-20      PM2 5 chemical components in downtown Los Angeles and Burbank (1986)
           have similar characteristics  	:*"->->
  June 2003                                I-xxii       DRAFT-DO NOT QUOTE OR CITE

-------
                                  List of Figures
                                      (cont'd)
 Number
 3-21      Concentrations of PM25 chemical components in Rubidoux and downtown
           Los Angeles (1986) 	3.55

 3-22      Monthly average Saharan dust components of the aerosol sampled in
           Miami, PL, from 1974 to 1996 	3_84

 3-23      PM2 5 and PMi0 concentrations measured at ChilHwack Airport, located in
           northwestern Washington State, just before and during the Asian desert
           dust episode of April and May 1998	3_85

 3-24      Time series of 24-h average PM10 concentrations observed in the Rio Grande
           Valley during May 1998  	3_gy

 3-25      PM10 concentrations observed in St. Louis, MO, during May 1998  	3-88

 3A-1      Philadelphia, PA-NJ MSA	  3A-4

 3A-2      Washington, DC MSA	  3A-5

 3A-3     Norfolk, VAMSA	  3A_6

 3A-4     Columbia, SC MSA	  3A-7

 3A-5     Atlanta, GAMSA  	  3A_8

 3A-6     Birmingham, AL MSA 	  3A_9

 3A-7     Tampa, FL MSA 	 3A-10

 3A-8     Cleveland, OHMSA 	 3A-11

 3A-9     Pittsburgh, PA MSA 	 3A-12

 3A-10    Steubenville, OH-Weirton, WVMSA	 3A-13

 3A-11    Detroit, MI MSA	 3A-14

 3A-12    Grand Rapids, MI MSA	 3A-15

 3A-13    Milwaukee, WI MSA	 3A-16

 3A-14    Chicago, IL MSA	 3A-17


June 2003                             I-xxiii      DRAFT-DO NOT QUOTE OR CITE

-------
                               List of Figures
                                   (cont'd)

Number

3A-15    Gary, IN MSA	  3A'18

3A-16    Louisville, KY MSA 	  3A'19

3A-17    St. Louis, MO MSA	  3A'20

3A-18    Baton Rouge, LA MSA	  3A~21

3A-19    Kansas City, KS-MO MSA	  3A-22

3A-20    Dallas, TX MSA 	  3A'23

3A-21    Boise, ID MSA ....'.	  3A'24

3A-22    Salt Lake City, UT MSA	  3A'25

3A-23    Seattle, WAMSA  	  3A-26

3A-24    Portland,ORMSA	  3A'27

3A-25    Los Angeles-Long Beach, CA MSA	  3A~28

3A-26    Riverside-San Bernadino, CA MSA 	  3A-29

3A-27    San Diego, CA MSA	  3A'30

3A-28    Columbia, SC MSA	  3A~32

3A-29    Tampa, FL MSA 	  3A'33

3A-30    Cleveland, OH MSA 	  3A'34

3A-31    Steubenville, OH MSA  	  3A"35

3A-32    Detroit, MI MSA	  3A'36

3A-33    Milwaukee, WI MSA	  3A-37

3A-34    Chicago, IL MSA	 3A'38

 3A-35    Gary, IN MSA	 3A'39


 June 2003                            I-xxiv       DRAFT-DO NOT QUOTE OR CITE

-------
                                    List of Figures
                                        (cont'd)

 Number                                                                        Page

 3A-36    Louisville, KYMSA	  3A_40

 3A-37    St. Louis, MI MSA 	  3A_41

 3A-38    Baton Rouge, LA MSA	  3A_42

 3A-39    Dallas, TX MSA 	  3A_43

 3A-40    Salt Lake City, UT MSA  	  3A_44

 3A-41    Portland, ORMSA	  3A_45

 3A-42    Los Angeles, CA MSA  	  3A_46

 3A-43    Riverside, CAMSA	  3A_47

 3A-44    San Diego, CAMSA	  3A_4g

 3D-1      Size distribution of particles generated in a laboratory resuspension
           chamber	         3D-2

 3D-2      Size distribution of California source emissions, 1986	 3D-4

 3E-1      Locations of the rural PM IMPROVE sites selected for the study  	3E-2

 3E-2a,b    Interannual variability in  24-h average PM2 5 concentrations observed at
           selected IMPROVE sites: (a) Bridger Wilderness, WY and (b) Yellowstone
           National Park, WY	3E_6

 3E-2c,d    Interannual variability in  24-h average PM2 5 concentrations observed at
           selected IMPROVE sites: (c) Dolly Sods/Otter Creek Wilderness, WV and
           (d) Brigantine National Wildlife Refuge, NJ	3E-7

 3E-3a,b    Interannual variability in 24-h average PMI0.2 5 concentrations observed at
           selected IMPROVE sites: (a) Bridger Wilderness, WY and (b) Yellowstone
          National Park, WY;	3E_8

 3E-3c,d   Interannual variability in 24-h average PM10_2 5 concentrations observed at
          selected IMPROVE sites:  (c) Dolly Sods/Otter Creek Wilderness, WV
          and (d) Brigantine National Wildlife Refuge, NJ  	3E-9
June 2003                               I_XXv       DRAFT-DO NOT QUOTE OR CITE

-------
                                    List of Figures
                                         (cont'd)
Number
3E-4a,b   Seasonal variability in 24-h average PM2 5 concentrations observed at
          selected IMPROVE sites: (a) Bridger Wilderness, WY and (b) Yellowstone
          National Park, WY ................................................ 3E'10

3E-4c,d   Seasonal variability in 24-h average PM2 5 concentrations observed at selected
          IMPROVE sites: (c) Dolly Sods/Otter Creek Wilderness, WV and
          (d) Brigantine National Wildlife Refuge, NJ ............................ 3E-! 1
3E-5a,b    Seasonal variability in 24-h average PM10.2 5 concentrations observed at
           selected IMPROVE sites: (a) Bridger Wilderness, WY and (b) Yellowstone
           National Park, WY ................................................ 3E'12

3E-5c,d    Seasonal variability in 24-h average PM10-2.5 concentrations observed
           at selected IMPROVE sites: (c) Doily Sods/Otter Creek Wilderness, WV
           and (d) Brigantine National Wildlife Refuge, NJ ........................ 3E-13

4-1        The diversity of fine PM from sites in the western and eastern U.S ............ 4-7

4-2        Relative importance of three modes of deposition of nitrate (A) and sulfate
           (B) at high elevation sites  ........................................... 4-1 1
 4-3        A simplified resistance catena representing the factors controlling deposition
           of particles to the surface ............................................ 4-16

 4-4        The relationship between deposition velocity of selected participate
           materials and the distribution of the material between the coarse- and
           fine-aerosol fractions  ............................................... 4~*7
 4-5       The relationship between particle diameter and deposition velocity for
           particles [[[ 4"18

 4-6       Vertical stratification of diverse, chemically speciated particles in a mixed
           oak forest [[[ 4~33

 4-7       The relationship between particle size and concentration below a spruce
           canopy with wind velocity at a height of 16.8 m equaling 5ms'1   ............ 4-34

 4-8       Range of percent of total deposition delivered in the dry paniculate
           fraction across the sites of the Integrated Forest Study ..................... 4-40

 4-9       Contribution of particulate deposition to total deposition of nitrogen, sulfur,

-------
                                     List of Figures
                                         (cont'd)
 Number
 4-10      Mean (±SE) percent of total nitrogen, sulfur, or base cation deposition
           contributed by fine plus coarse particles	4-49

 4-11      Annual total deposition of heavy metals to Hubbard Brook Experimental
           Forest, NH	4.54

 4-12      Sample stressors and the essential ecological attributes they affect	4-56

 4-13      Linkages among various ecosystem goods and services (food, water,
           biodiversity, forest products) and other driving forces (climate change)	4-61

 4-14      Effects of environmental stress on forest trees presented on a hierarchial
           scale for the leaf, branch, tree, and stand levels of organization	4-80

 4-15      Illustration of the nitrogen cascade showing the movement of the human-
           produced reactive nitrogen as it cycles through the various environmental
           reservoirs in the atmosphere, terrestrial ecosystems, and aquatic ecosystems  ... 4-97

 4-16      Nitrogen cycle 	4-100

 4-17      Diagrammatic overview of excess nitrogen in North America	4-104

 4-18      Schematic of sources and sinks of hydrogen ions in a forest  	4-118

 4-19      Key elements of proposed framework for determining critical loads for
           nitrogen and sulfur in the United States	4-125

 4-20      Calcium deposition in > 2-u-m particles, < 2-u.m particles, and wet forms
           and leaching in the Integrated Forest Study sites  	4-129

 4-21       Magnesium deposition in > 2-u-m particles, < 2-u.m particles, and wet forms
           and leaching in the Integrated Forest Study sites  	4-130

 4-22       Potassium deposition in > 2-u-m particles, < 2-um particles, and wet forms
           and leaching in the Integrated Forest Study sites  	4-131

 4-23       Base cation deposition  in > 2-u-m particles, < 2-u.m particles, and wet forms
           and leaching in the Integrated Forest Study sites  	4-132

 4-24       Total cation leaching balanced by sulfate and nitrate estimated from
           particuiate deposition and by other sources of sulfate and nitrate and
           by other anions in the Integrated Forest Study sites  	4-133

June 2003                                I-xxvii       DRAFT-DO NOT QUOTE OR CITE

-------
                                    List of Figures
                                        (cont'd)
Number
4-25      Soil exchangeable Ca+2 pools and net annual export of Ca+2 in the
          Integrated Forest Study sites	4-134

4-26      Soil exchangeable Mg+2 pools and net annual export of Mg+2 in the
          Integrated Forest Study sites	4-134

4-27      Soil exchangeable K+2 pools and net annual export of K+2 in the Integrated
          Forest Study sites	4~135

4-28a     Simulated soil solution mineral acid anions and base cations in the red
          spruce site with no change, 50% N and S deposition, and 50% base
          cation deposition  	4-138

4-28b     Simulated soil solution mineral acid anions and base cations in the red
          spruce site with no change, 50% N and S deposition, and 50% base
          cation deposition  	4-139

4-29a     Simulated soil solution Al and soil base saturation in the red spruce site
          with no change, 50% N and S deposition, and 50% base cation deposition	4-140

4-29b     Simulated soil solution Al and soil base saturation in the red spruce site
          with no change, 50% N and S deposition, and 50% base cation deposition	4-141

4-30a     Simulated soil solution mineral acid anions and base cations in the
          Coweeta site with no change, 50% N and S deposition, and 50% base
          cation deposition  	-	4-143

4-30b     Simulated soil solution mineral acid anions and base cations in the
          Coweeta site with no change, 50% N and S deposition, and 50% base
          cation deposition  	4-144

4-31      Relationship of plant nutrients and trace metals with vegetation  	4-148

4-32      Light reflected from a target toward an observer   	4-157

4-33      Light-scattering efficiency factor for a homogeneous sphere with an index
          of refraction of 1.50 as a function of the size parameter	4-161

4-34      Volume-specific light-scattering efficiency as a function of particle
          diameter	4-162
 June 2003                               I-xxviii      DRAFT-DO NOT QUOTE OR CITE

-------
                                     List of Figures
                                          (cont'd)
  Number
  4-35      Particle growth curve as a function of relative humidity showing deliquescent
            growth of ammonium sulfate particles at the deliquescent point, reversible
            hygroscopic growth of ammonium sulfate solution droplets at relative
            humidity greater than 80%, and hysteresis until the crystallization point
            is reached	4-164

  4-36      Comparison of extinction and visual range   	4-168

  4-37      Proportionality of observed daytime haziness to fine particle mass
            concentration in Los Angeles  	4-169

  4-38      Relative  humidity adjustment factor for ammonium sulfate as a function of
            relative humidity 	4-171

  4-39a      Aggregate visibility trends for 10 eastern Class  1 areas	4-181

  4-39b      Aggregate visibility trends for 26 western Class 1 areas	4-181

 4-40       (a) Eastern class  I area aggregate trends in aerosol light extinction on the
            20% haziest days, including trends by major aerosol component; (b) Western
            class I area aggregate trends in aerosol light extinction on the 20% haziest
           days,  including trends by major aerosol component	4-183

 4-41 a     Light extinction trends in Tucson, Arizona from 1993 to 2002  	4-185

 4-4 Ib     Light extinction trends in Phoenix, Arizona from 1994 to 2001	4-186

 4-42      Estimated global  mean radiative forcing exerted by gas and various particle
           phase species for the year 2000, relative to 1750  	4-213

 5-1        Hypothetical exposure time profile: Pollutant exposure as a function of time
           showing how the averaged exposure, integrated exposure, and peak exposure
           relate to the instantaneous exposure 	5.9

 5-2        Distribution of F^ (a) and a (b) for daytime as estimated from PTEAM
           study data	       5_23

 5-3        Comparison of correlation coefficients for longitudinal analyses of personal
           exposure versus ambient concentrations for individual subjects  for PM2 5
           and sulfate  	_        5_5g
June 2003                                I-xx{x        DRAFT-DO NOT QUOTE OR CITE

-------
                                   List of Figures
                                       (cont'd)
                                                                                 Page
Number
5-4       Personal exposure versus ambient concentrations for PM2.5 and sulfate	5-61

5-5       Regression analyses of aspects of daytime personal exposure to PM10
          estimated using data from the PTEAM study  	5'62

5-6       Air exchange rates measured in homes throughout the United States	5-69

5-7       Box plots of hourly air exchange rates stratified by season in Boston, MA,
          during 1998 	5"70

5-8       Geometric mean infiltration factor (indoor/outdoor ratio) for hourly nighttime,
          nonsource data for two seasons	-	

5-9       Regression of air exchange rate on absolute indoor-outdoor temperature
          difference	

5-10     Comparison of deposition rates from Long et al., 2001a with literature
          values	

 5-11     Penetration efficiencies and deposition rates from models of nightly
                  j .                                                	5-76
          average data	

 5-12     Mean hourly indoor/outdoor particle concentration ratio from an unoccupied
          residence in Fresno, CA, during spring 1999 	5'82

 5-13     Personal versus outdoor SO4'2 in State College, PA 	5-97

 5-14     Plots of nonambient exposure to PMIO, (a) daily individual daytime values
           from PTEAM data and (b) daily-average values from THEES data	5-110
  June 2003                                I-xxx       DRAFT-DO NOT QUOTE OR CITE

-------
                      Authors, Contributors, and Reviewers
                               EXECUTIVE SUMMARY
 Principal Authors
 Dr. William E. Wilson—National Center for Environmental Assessment (B243-01),
 U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Lester D. Grant—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Ms. Beverly Comfort—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. J.H.B. Garner—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Joseph P. Pinto—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
                           CHAPTER!. INTRODUCTION
 Principal Authors

 Dr. Lester D. Grant—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

 Dr. Dennis J. Kotchmar—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC  27711
June 2003                               I-xxxi       DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
            CHAPTER 2. PHYSICS, CHEMISTRY, AND MEASUREMENT
                            OF PARJJCULATE MATTER
Principal Authors

Dr. William Wilson—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr Candis S. Claiborn—Washington State University, Laboratory for Atmospheric Research,
Department of Civil and Environmental Engineering, P.O. Box 642910, Pullman, WA 99164

Dr. Brooke L. Hemming—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Contributing Authors

Dr. Tom Cahill, University of California, Davis, One Shields Ave., Davis, CA 95616

Dr. Judith C. Chow, Desert Research Institute, 2215 Raggio Parkway, Reno, NV 89512

Max R. Peterson, Research Triangle Institute, P.O. Box 12194, Research Triangle Park, NC
27709

Dr. James J. Schauer, University of Wisconsin, College of Engineering, Department of Civil and
Environmental Engineering, Madison, WI 53706

Dr Barbara J. Turpin, The State University of New Jersey, Rutgers, Department of
Environmental Sciences and Rutgers Cooperative Extension, New Brunswick, NJ 08901-8551

Dr. John G. Watson, Desert Research Institute, 2215 Raggio Parkway, Reno, NV  89512

 Contributors and Reviewers

 Dr. Edward O. Edney—National Exposure Research Laboratory (E205-02)
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. R.R. Eldred, Crocker Nuclear Laboratory, University  of California, Davis, Davis, CA 95616

 Mr. William Ewald—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
 June 2003                              I-xxxii      DRAFT-DO NOT QUOTE OR CITE

-------
                       Authors, Contributors, and Reviewers
                                         (cont'd)


  Contributors and Reviewers
  (cont'd)

  James B. Flanagan—Research Triangle Institute, P.O. Box 12194, Research Triangle Park, NC


  Dr. Judith Graham—American Chemistry Council, 1300 Wilson Boulevard, Arlington VA
  22207                                                                  6   '

  Mr. Jim Homolya—Office of Air Quality Planning and Standards (C339-02),
  U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 '

  R.K.M. Jayanty—Research Triangle Institute, P.O. Box 12194, Research Triangle Park NC
  27709

  Mr. Scott Mathias—Office of Air Quality Planning and Standards (C539-01),
  U.S. Environmental Protection Agency, Research Triangle Park, NC 27711 '

  Mr. Frank McElroy—National Exposure Research Laboratory (MDr46),
  U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

  Dr. Joseph Pinto—National Center for Environmental Assessment (B243-OI),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Joann Rice—Office of Air Quality Planning and Standards (C339-02),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27709

 Ed E. Rickman—Research Triangle Institute, P.O. Box 12194, Research Triangle Park, NC


 Dr. Bret Schichtel—Colorado State University, Cooperative Institute for Research in the
 Atmosphere, Foothills Campus, Fort Collins, CO  80523-1375

 Dr. Russell Wiener—National Exposure Research Laboratory (D205-03)
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
June 2003                              I-xxxiii       DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
                 CHAPTERS.  CONCENTRATIONS,SOURCES,AND
              EMISSIONS OF ATMOSPHERIC PARTICULATE MATTER
Principal Author

Dr. Allen S. Lefohn—A.S.L. & Associates, 111 North Last Chance Gulch, Helena, MT 59601

Dr. Joseph P. Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Contributing Authors

Dr. Barbara J. Turpin—The State University of New Jersey, Rutgers, Department of
Environmental Sciences and Rutgers Cooperative Extension, New Brunswick, NJ 08901-8551

Dr. James J. Schauer—University of Wisconsin, College of Engineering, Department of Civil
and Environmental Engineering, Madison, WI 53706

Contributors and Reviewers

Dr. JoEllen Bandemeyer—Research Triangle Institute, PO Box 12194, Research Triangle Park,
NC 27709

Dr. Lyle Chinkin—Sonoma Technology, 1360 Redwood Way, Suite C, Petaluma, CA 94549

Dr Candis S. Claiborn—Washington State University, Laboratory for Atmospheric Research,
Department of Civil and Environmental Engineering, P.O. Box 642910, Pullman, WA 99164

Dr. Steven Colome—Integrated Environmental Services, 5319 University Drive, #430, Irvine,
CA 92612

Mr. Tom Coulter—National Exposure Research Laboratory (D234-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Dr. Edward O. Edney—National Exposure Research Laboratory (E205-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

 Mr. Terence Fitz-Simons—Office of Air Quality Planning and Standards (C304-01),
 U. S. Environmental Protection Agency, Research Triangle Park, NC  27711
 June 2003                             I-xxxiv       DRAFT-DO NOT QUOTE OR CITE

-------
                      Authors, Contributors, and Reviewers
                                       (cont'd)
 Contributors and Reviewers
 (cont'd)

 Dr. Sheldon Friedlander—University of California at Los Angeles, Department of Chemical
 Engineering, 5531  Boelter Hall, Los Angeles, CA 90095

 Dr. Lynn Hildemann—Civil and Environmental Engineering Department, Stanford University,
 Stanford, CA 94305

 Mr. Phil Lorang—Office of Air Quality Planning and Standards (D205-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Ms. Karen Magliano—California Air Resources Board, 2020 L Street, Sacramento, CA 95814

 Mr. Scott Mathias—Office of Air Quality Planning and Standards (C539-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Ms. Nehzat Mottallebi—California Air Resources Board, 2020 L Street, Sacramento, CA 95814

 Mr. Tom Pace—Office of Air Quality Planning and Standards (D205-01), U.S. Environmental
 Protection Agency, Research Triangle Park, NC 27711

 Dr. Richard Poirot—VT Air Program, Building 3  South, 103 South Main Street, Waterbury, VT
 05671

 Mr. Win Setiawan—California Air Resources Board, 2020 L Street, Sacramento, CA 95814

 Mr. Dane Westerdahl—California Air Resources Board, 2020 L Street, Sacramento, CA  95814
       CHAPTER 4. ENVIRONMENTAL EFFECTS OFPARTICULATEMATTER
Principal Authors

Ms. Beverly Comfort—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711

Mr. William Ewald—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC  27711
June 2003                               I-xxxv       DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
Principal Authors
(cont'd)

Dr. J.H.B. Garner—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Lester D. Grant—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. David A. Grantz—University of California/Riverside, Kearney Agricultural Center,
9240 South Riverbend Avenue, Parlier, CA 93648

Dr. Brooke L. Hemming—National Center for Environmental Assessment (B243-OI),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Dale W. Johnson—Environmental and Resource Science, 1000 Valley Road, University of
Nevada, Reno, NV 89512

Dr. William Malm—National Park Service, Air Resources Division, CIRA, Colorado State
University, Fort Collins, CO

Dr. Joseph P. Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Contributing Authors

 Dr. Paul J. Hanson—Environmental Sciences Division, Oak Ridge National Laboratory,
 P.O. Box 2008, Bethel Valley Road, Building 1059, Oak Ridge, TO 37831-6422

 Dr. William H. Smith—Professor Emeritus, Yale University School of Forestry and
 Environmental Studies, 370 Prospect Street, New Haven, CT 06511

 Contributors and Reviewers

 Dr. Larry T. Cupitt—National Exposure Research Laboratory (D305-01),
 U. S. Environmental Protection Agency, Research Triangle Park, NC  27711

 Mr. Rich Damberg—Office of Air Quality Planning and Standards (C539-02),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Russell R. Dickerson—University of Maryland, Department of Meteorology,
 Stadium Drive, College Park, MD 20742
  June 2003                              I-xxxvi       DRAFT-DO NOT QUOTE OR CITE

-------
                      Authors, Contributors, and Reviewers
                                       (cont'd)


  Contributors and Reviewers
  (cont'd)

  Dr. Anne Grambsch—National Center for Environmental Assessment (860ID),
  U. S. Environmental Protection Agency, Washington, DC 20036

  Dr. Sagar V. Krupa—University of Minnesota, Department of Plant Pathology,
  495 Borlaug Hall, 1991 Upper Buford Circle, St. Paul, MN 55108

  Dr. Alan J. Krupnick—Quality of the Environment Division, Resources for the Future
  1616 P Street, NW, Washington, DC 20036

  Mr. Paul T. Roberts—Sonoma Technology, Inc., 1360 Redwood Way - Suite C
  Petaluma,CA 94954

  Mr. John Spence-Cary, NC 27511

 Ms. Debra Meyer Wefering—Duckterather Weg 61, Bergisch Gladbach, Germany  54169

 Ms. Kay Whitfield—Office of Air Quality Planning and Standards (C243-02),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Richard Zepp—National Exposure Research Laboratory (IOD),
 U. S. Environmental Protection Agency, Athens, GA
           CHAPTERS. HUMAN EXPOSURE TO PARTICULATE MATTER
                             AND ITS CONSTITUENTS
 Principal Authors

 Dr. Linda S. Sheldon—National Exposure Research Laboratory (E205-02),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. Haluk Ozkaynak—National Exposure Research Laboratory (E205-02),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Dr. William E. Wilson—National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711
June 2003                              I-xxxvii      DRAFT-DO NOT QUOTE OR CITE

-------
                     Authors, Contributors, and Reviewers
                                      (cont'd)
Contributing Authors

Dr. Janet Burke—National Exposure Research Laboratory (E205-02),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Roy Fortmann—National Exposure Research Laboratory (E205-02),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. David T. Mage—Institute for Survey Research, Temple University,
Philadelphia, PA 19122-6099

Mr. Thomas McCurdy—National Exposure Research Laboratory (E205-02),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711


Dr. Gary Norris—National Exposure Research Laboratory (E205-03),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Anne Rea—National Exposure Research Laboratory (E205-04),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Alan Vette—National  Exposure Research Laboratory (E205-04),
U. S. Environmental Protection Agency, Research Triangle Park, NC 27711

Contributors and Reviewers

Mr. John Langstaff—Office of Air Quality Planning and Standards (C539-01)
U.S. Environmental Protection Agency, Research Triangle park, NC 27711

Dr. Joseph Pinto—National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Mr. Harvey Richmond—Office of Air Quality Planning and Standards (C539-01),
U. S. Environmental Protection Agency, Research Triangle Park, NC  27711
 June 2003                             I-xxxviii     DRAFT-DO NOT QUOTE OR CITE

-------

             U.S. ENVIRONMENTAL PROTECTION AGENCY
 PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
                       FOR PARTICULATE MATTER
Executive Direction

Dr. Lester D. Grant—Director, National Center for Environmental Assessment (B243-01),
U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Scientific Staff

Dr. Robert W. Elias—PM Team Leader, Health Scientist, National Center for Environmental
Assessment (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711

Dr. William E. Wilson—Air Quality Coordinator, Physical Scientist, National Center for
Environmental Assessment (B243-01), U.S. Environmental Protection Agency, Research
Triangle Park, NC 27711

Ms. Beverly Comfort—Health Scientist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. J.H.B. Garner—Ecological Scientist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Dennis J. Kotchmar—Medical Officer, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Dr. Joseph P. Pinto—Physical Scientist, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

Technical Support Staff

Ms. Nancy Broom—Information Technology Manager, National Center for Environmental
Assessment (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711

Mr. Douglas B. Fennell—Technical Information Specialist, National Center for Environmental
Assessment (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC
27711

Ms. Emily R. Lee—Management Analyst, National Center for Environmental Assessment
(B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC  27711
 june 2003                             I-xxxix      DRAFT-DO NOT QUOTE OR CITE

-------
              U.S. ENVIRONMENTAL PROTECTION AGENCY
   PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
                        FOR PARTICULATE MATTER
                                       (cont'd)
 Technical Support Staff
 (cont'd)

 Ms. Diane H. Ray—Program Specialist, National Center for Environmental Assessment
 (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Ms. Donna Wicker—Administrative Officer, National Center for Environmental Assessment
 (B243-01), U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Mr. Richard Wilson—Clerk, National Center for Environmental Assessment (B243-01),
 U.S. Environmental Protection Agency, Research Triangle Park, NC 27711

 Document Production Staff

 Dr. Carol A. Seagle—Technical Editor, Computer Sciences Corporation, 2803 Slater Road, Suite
 220, Morrisville, NC 27560

 Ms. Jessica Long—Graphic Artist, Computer Sciences Corporation, 2803 Slater Road, Suite
 220, MorrisviJle, NC 27560

 Ms. Carolyn T. Perry—Word Processor, Computer Sciences Corporation, 2803 Slater Road,
 Suite 220, Morrisville, NC 27560

 Ms. Kelly Quinones—Word Processor, InfoPro, Inc., 8200 Greensboro Drive, Suite 1450,
 McLean, VA 22102

 Mr. Charles E. Gaul—Graphic Artist, Computer Sciences Corporation, 2803 Slater Road, Suite
 220, Morrisville, NC 27560

 Mr. John M. Havel—Graphic Artist, Computer Sciences Corporation, 2803 Slater Road, Suite
 220, Morrisville, NC 27560

 Ms. Laura A. Rhodes—Graphic Artist, Computer Sciences Corporation, 2803  Slater Road,  Suite
 220, Morrisville, NC 27560

 Mr. Keith A. Tarpley—Graphic Artist, Computer Sciences  Corporation, 2803 Slater Road,  Suite
 220, Morrisville, NC 27560
June 2003
I-xl
DRAFT-DO NOT QUOTE OR CITE

-------
               U.S. ENVIRONMENTAL PROTECTION AGENCY
   PROJECT TEAM FOR DEVELOPMENT OF AIR QUALITY CRITERIA
                        FOR PARTICULATE MATTER
                                      (cont'd)


  Technical Reference Staff

  Mr. John A. Bennett—Technical Information Specialist, SANAD Support Technologies, Inc.
  11820 Parklawn Drive, Suite 400, Rockville, MD 20852

  Ms. Lisa Greenbaum—Records Management Technician, Reference Retrieval and Database
  Entry Clerk, InfoPro, Inc., 8200 Greensboro Drive, Suite 1450, McLean, VA 22102

  Ms. Sandra L. Hughey—Technical Information Specialist, SANAD Support Technologies, Inc.,
  11820 Parklawn Drive, Suite 400, Rockville, MD 20852
June 2003                              I_xli       DRAFT-DO NOT QUOTE OR CITE

-------
                   Abbreviations and Acronyms
AAS
AC
ACGIH
ACS
ADS
AES
AIRS
ANC
AQCD
AQI
ARIES
ASOS
ATOFMS
AWOS
BaP
BASE
BC
BNF
BOSS
 BYU
 CAA
 CAAM
 CAMNET
 CARB
 CASAC
 CASTNet
atomic absorption spectrophotometry
air conditioning
American Conference of Governmental Industrial Hygienists
American Cancer Society
annular denuder system
atomic emission spectroscopy
Aerometric Information Retrieval System
acid neutralizing capacity
Air Quality Criteria Document
Air Quality Index
Aerosol Research and Inhalation Epidemiology Study
Automated Surface Observing System
time-of-flight mass spectrometer
Automated Weather Observing System
benzo(a)pyrene
Building Assessment and Survey Evaluation
black carbon
bacterial nitrogen fertilization
Brigham Young University Organic Sampling System
Bringham Young University
 Clean Air Act
 continuous ambient mass monitor
 Coordinated Air Monitoring Network
 California Air Resources Board
 Clean Air Scientific Advisory Committee
 Clean Air Status and Trends Network
June 2003
                                      I-xlii
                               DRAFT-DO NOT QUOTE OR CITE

-------
  cc
  CCPM
  CCSEM
  CEN
  CFCs
  CFR
  CHAD
  GIF
  CUT
  CMAQ
  CMB
  CMSA
  CO CD
  COD
  COPD
  CPC
  CRP
  CSIRO
  CSMCS
  CSS
  CTM
  CV
  CVM
  Da
  DAQ
  DCFH
  DMPS
 base cations
 carbonate carbon
 continuous coarse particle monitor
 computer-controlled scanning electron microscopy
 European Standardization Committee
 chlorofluorocarbons
 Code of Federal Regulations
 Consolidated Human Activity Database
 charcoal-impregnated cellulose fiber
 Chemical Industry Institute of Technology
 Community Multi-Scale Air Quality
 chemical mass balance
 Consolidated Metropolitan Statistical Area
 Air Quality Criteria Document for Carbon Monoxide
 coefficient of divergence
 chronic obstructive pulmonary disease
 condensation particle counter
 Coordinated Research Program
 Commonwealth Scientific and Industrial Research Organisation
 Carbonaceous Species Methods Comparison Study
 coastal sage scrub
 chemistry-transport model
 coefficient of variation
 contigent valuation method
 aerodynamic diameter
 Department of Air Quality
dichlorofluorescin
differential mobility particle sizer
June 2003
                   I-xliii
DRAFT-DO NOT QUOTE OR CITE

-------
DMS
DP
dv
BAD
EC
ECAO
EDXRF
EEA
ENSO
EPA
ESP
ETS
EXPOLIS

FID
FRM
GAM
GC
GCMs
GCVTC
GC/MSD
GHG
GSD
HBEF
HDS
HEADS
HEI
 HI
 hivol
dimethyl sulfide
particle diameter
deciview index
electrical aerosol detector
elemental carbon
Environmental Criteria and Assessment Office
energy dispersive X-ray fluorescence
Essential Ecological Attributes
El Nino-Southern Oscillation
Environmental Protection Agency
electrostatic precipitator
environmental tobacco smoke
Air Pollution Exposure Distribution within Adult Urban Populations in
Europe
flame ionization detection
Federal Reference Method
general additive models
gas chromatography
General Circulation Models
Grand Canyon Visibility Transport Commission
gas chromatography/mass-selective detection
greenhouse gases
geometric standard deviation
Hubbard Brook Experiment Forest
honeycomb denuder/filter pack sampler
Harvard-EPA Annular Denuder Sampler
Health Effects Institute
Harvard Impactors
High volume sampler
June 2003
                    I-xliv
DRAFT-DO NOT QUOTE OR CITE

-------
   HTGC-MS
   HVAC
   1C
   ICP
   IPS
   IMPROVE
   INAA
   IOVPS
   IPCC
   IPM
   IPN
   ISO
   K
   LAI
   LOD
  LWC
  LWCA
  MAA
  MAACS
  MADPro
  MAQSIP
  mCa
  MDL
  MOUDI
  MS
  MSA
  MSA
  mv
  high temperature gas chromotography-mass spectrometry
  heating, ventilation, or air conditioning
  ion chromatography
  inductively coupled plasma
  Integrated Forest Study
  Interagency Monitoring of Protected Visual Environments
  instrumental neutron activation analysis
  integrated organic vapor/particle sampler
  Intergovernmental Panel on Climate Change
  inhalable particulate matter
  Inhalable Particulate Network
 International Standards Organization
 Koschmieder constant
 leaf area indices
 level of detection
 liquid water content
 liquid water content analyzer
 mineral acid anion
 Metropolitan Acid Aerosol Characterization Study
 Mountain Acid Deposition Program
 Multiscale Air Quality Simulation Platform
 membrane-associated calcium
 minimum detection level
 micro-orifice uniform deposit impactor
 mass spectroscopy
 methane sulfonic acid
metropolitan statistical area
motor vehicle
June 2003
                    I-xIv
                                                     DRAFT-DO NOT QUOTE OR CITE

-------
NAAQS
NAMS
NAPAP
NARSTO
NAST
NCEA
NDDN
NERL
NESCAUM
NFRAQS
NGOs
NHAPS
NIOSH
NIR
NIST
NOPL
NOX
NPP
Nr
 NRC
 NuCM
 OAQPS
 OAR
 OC
 ORD
 PAH
 PAN
 PAR
National Ambient Air Quality Standards
National Ambient Monitoring Stations
National Acid Precipitation Assessment Program
North American Research Strategy for Tropospheric Ozone
National Assessment Synthesis Team
National Center for Environmental Assessment
National Dry Deposition Network
National Exposure Research Laboratory
Northeast States for Coordinated Air Use Management
North Frontal Range Air Quality Study
non-governmental organizations
National Human Activity Pattern Survey
National Institute for Occupational Safety and Health
near infared radiation
National Institute of Standards and Technology
naso-oro-pharyngo-laryngeal
nitrogen oxides
net primary production
 reactive nitrogen
National Research Council
 nutrient cycling model
 Office of Air Quality Planning and Standards
 Office of Air and Radiation
 organic carbon
 Office of Research and Development
 polymiclear aromatic hydrocarbon
 peroxyacetyl nitrate
 photosynthetically active radiation
 June 2003
                                        I-xlvi
                                DRAFT-DO NOT QUOTE OR CITE

-------
   PBL
   PBP
   PBY
   PC
   PC
   PC-BOSS

   PCA
   PCBs
   PCDD
   PCDF
   PCM
   PEM
   PESA
   PIXE
  PM
  PM10.25
  PM25
  PMF
  POP
  PRB
  PTEAMS
  PTEP
  PUF
  PWC
  RAMS
  RAMS
  RAS
  RCS
 planetary boundary layer
 primary biological particles
 Presbyterian Home
 particle concentrator
 pyrolitic carbon
 Participate Concentrator-Brigham Young University Organic Sampling
 System
 principal component analysis
 polychloronated biphenyls
 polychlorinated dibenzo-p-dioxins
 polychlorinated dibenzofurans
 particle composition monitor
 personal exposure monitor
 proton elastic scattering analysis
 proton induced X-ray emission
 paniculate matter
 coarse particulate matter
 fine particulate matter
 positive matrix factorization
 persistent organic pollutant
 policy-relevant background
 Particle Total Exposure Assessment Methodology
 PM10 Technical Enhancement Program
 polyurethane foam
 precipitation-weighted concentrations
 Regional Air Monitoring Study
 Real-Time Air Monitoring System
roll-around system
Random Component Superposition
June 2003
                   I-xlvii
                                                   DRAFT-DO NOT QUOTE OR CITE

-------
RFO
RH
RIVM
RPM
RPM
RRMS
RTF
RUE
SA
SAB
SCAQS
SCOS
sd
sec
SEM
SES
SEV
SIP
 SLAMS
 SMPS
 SMSAs
 soc
 SoCAB
 SOPM
 SP
 SRM
 STN
 SUVB
residual fuels oils
relative humidity
Directorate-General for Environmental Protection
Regional Particulate Model
respirable paniculate matter
relatively remote monitoring sites
Research Triangle Park
radiation use efficiency
Sierra Anderson
Science Advisory Board
Southern California Air Quality Study
Southern California Ozone Study
standard deviation
secondary
scanning electron microscopy
sample equilibration system
Sensor Equivalent Visibility
State Implementation Plans
State and Local Air Monitoring Stations
 scanning mobility particle sizer
 Standard Metropolitan Statistical Areas
 semivolatile organic compounds
 South Coast Air Basin
 secondary organic particulate matter
 Staff Paper
 standard reference method
 Speciation Trends Network
 solar ultraviolet B radiation
June 2003
                    I-xlviii
                                                     DRAFT-DO NOT QUOTE OR CITE

-------
   svoc
   SVM
   TAR
   TDMA
   TEO
   TEOM
   THEES
   TNF
   TOFMS
   TOR
   TOT
   TPM
   TRXRF
   TSP
   TVOC
   UAM-V
   UCM
  UNEP
  URG
  USGCRP
  VAPS
  VMD
  VOC
  WMO
  VR
  WINS
  WMO
  WRAC
  semivolatile organic compounds
  semivolatile material
  Third Assessment Report
  Tandem Differential Mobility Analyzer
  trace element oxides
  tapered element oscillating microbalance
  Total Human Environmental Exposure Study
  tumor necrosis factor
  aerosol time-of-flight mass spectroscopy
  thermal/optical reflectance
 thermal/optical transmission
 thoracic paniculate matter
 total reflection X-ray fluorescence
 total suspended particulates
 total volatile organic compounds
 Urban Airshed Model Version V
 unresolved complex mixture
 United Nations Environment Programme
 University Research Glassware
 U.S. Global Change Research Program
 Versatile Air Pollution Samplers
 volume mean diameter
 volatile organic compounds
 World Meteorological Organization
visual range
Well Impactor Ninety-Six
World Meteorological Organization
Wide Range Aerosol Classifier
June 2003
                   I-xlix
                                                   DRAFT-DO NOT QUOTE OR CITE

-------
WTA
WTP
XAD
XRP
willingness to accept
willingness to pay
polystyrene-divinyl benzene
X-ray fluorescence
 June 2003
                                         1-1
                                DRAFT-DO NOT QUOTE OR CITE

-------
   i                              EXECUTIVE SUMMARY
   2
   3
   4      E.I   INTRODUCTION
   5           The purpose of this document, Air Quality Criteria for Particulate Matter (PM AQCD),
   6      is to present air quality criteria for paniculate matter (PM) in accordance with Clean Air Act
   7      (CAA) Sections 108 and 109, which govern establishment, review, and revision of U.S. National
   8      Ambient Air Quality Standards (NAAQS) as follows:

   9      •  Section 108 directs the U.S. Environmental Protection Agency (EPA) Administrator to list
            pollutants that may reasonably be anticipated to endanger public health or welfare and to
            issue air quality criteria for them. The air quality criteria are to reflect the latest scientific
            information useful in indicating the kind and extent of all identifiable effects on public
            health and welfare expected from the presence of the pollutant in ambient air.

 10      •  Section 109 directs the EPA Administrator to set and periodically revise, as appropriate,
            (a) primary NAAQS, which in the judgement of the Administrator, are requisite to protect
            public health, with an adequate margin of safety, and (b) secondary NAAQS which, in the
            judgement of the Administrator, are requisite to protect the public welfare from any known
            or anticipated adverse effects (e.g., impacts on vegetation, crops, ecosystems, visibility,
            climate, man-made materials, etc.).

 11      •   Section 109 of the CAA also requires periodic review and, if appropriate, revision of
            existing criteria and standards. Also, an independent committee of non-EPA experts, the
            Clean Air Scientific Advisory Committee (CASAC), is to provide the EPA Administrator
            advice and recommendations regarding the scientific soundness and appropriateness of
            criteria and NAAQS.

12          To meet these CAA mandates, this document assesses the latest scientific information
13      useful in deriving criteria  as scientific bases for decisions on possible revision of current
14      PM NAAQS. A separate EPA PM Staff Paper will draw upon assessments  in this document,
15      together with technical analyses and other information, to identify alternatives for consideration
16      by the EPA Administrator with regard to possible retention or revision of the PM NAAQS.
        June 2003                                 E-l         DRAFT-DO NOT QUOTE OR CITE

-------
1          The present document is organized into a set of 10 chapters, as follows:
2       • This Executive Summary summarizes key points from the ensuing chapters.

3       • Chapter 1 provides a general introduction, including information on legislative requirements
         and history of the PM NAAQS and an overview of approaches used to prepare this document.

4       • Chapters 2,3 and 5 provide background information on PM atmospheric science, air quality,
         and human exposure aspects to help place the succeeding discussions of PM health and
         environmental effects into perspective.

5       • Chapter 4 deals with environmental effects of PM on vegetation and ecosystems, visibility,
         man-made materials, and climate.

6       • Human health issues related to PM are addressed in Chapter 6 (Dosimetry); Chapter 7
         (Toxicology); and Chapter 8 (Community Epidemiology).

7       * Chapter 9 provides an integrative synthesis of key points from the preceding chapters.

 8
 9      E.2  AIR QUALITY AND EXPOSURE ASPECTS
10           The document's discussion of air quality and exposure aspects considers chemistry and
11      physics of atmospheric PM; analytical techniques for measuring PM mass, size, and chemical
12     composition; sources of ambient PM in the United States; temporal/spatial variability and trends
13      in ambient U.S. PM levels; and ambient concentration-human exposure relationships.  Overall,
14     the atmospheric science and air quality information provides further evidence substantiating the
15      1996 PM AQCD conclusion that distinctions between fine and coarse mode particles (in terms of
16      emission sources, formation mechanisms, atmospheric transformation, transport distances, air
17      quality patterns, and exposure relationships) warrant fine and coarse PM being viewed as
18      separate subclasses of ambient PM. Key findings are summarized in the next several sections.
19
20
        June 20o3                                E-2         DRAFT-DO NOT QUOTE OR CITE

-------
 1
E.2.1   Chemistry and Physics of Atmospheric Particles
 2       •  Airborne PM is not a single pollutant, but rather is a mixture of many subclasses of
           pollutants with each subclass containing many different chemical species. Particles
           suspended in the atmosphere originate from a variety of sources and possess a range of
           morphological, chemical, physical, and thermodynamic properties. A complete description
           of atmospheric particles would include an accounting of the chemical composition,
           morphology, and size of each particle, and the relative abundance of each particle type as a
           function of particle size. Recent developments in single particle analysis techniques are
           bringing such a description closer to reality.

 3      •  Particle size distributions show that atmospheric particles exist in two classes: fine particles
           and coarse particles.  Fine and coarse particles are defined primarily in terms of their
           formation mechanisms and  size; they also differ in sources, chemical composition, and
           removal processes (see Table 2-1). Subsequent chapters show that fine and coarse particles
           also differ in aspects of concentration, exposure, dosimetry, toxicology, and epidemiology.
           These differences support the setting of separate standards for fine and coarse particles.

4      •  Fine and coarse particles overlap in the size range between 1 and  3 urn aerodynamic
           diameter where ambient PM concentrations are at a minimum.  Coarse particles are generally
           larger than this minimum and are generally formed by mechanical processes.  Energy
          considerations limit the break-up of large mineral particles and small particle aggregates
          generally to a minimum size of about 1 urn in diameter.  Coarse particles and coarse-mode
          particles are equivalent terms.

5       •  Fine PM is derived primarily from combustion material that has volatilized and then
6         condensed to  form primary PM or from precursor gases reacting in the atmosphere to form
          secondary PM. New  fine particles are formed by the nucleation of gas phase species; they
          grow by coagulation (existing particles combining) or condensation (gases condensing on
          existing particles).  Fine particles are subdivided into accumulation, Aitkin, and nucleation
          modes. In earlier texts,  nuclei mode referred to the  size range now split into the Aitkin and
          nucleation modes (see Figures 2-4 and 2-5). Particles in the  size range below 0.1 um
          diameter are called ultrafine  or nanoparticles and include the Aitkin and nucleation modes.

      June 2003                                 E-3         DRAFT-DO NOT QUOTE OR CITE

-------
1       •  PM2 5 is an indicator for fine particles. PM10.2.5 is an indicator for thoracic coarse particles
          (coarse particles capable of reaching the thoracic portion of the respiratory system - trachea,
          bronchi, and alveolar regions). PM25 has been selected as the indicator for fine particles to
          include all accumulation-mode particles during high relative humidity, while recognition that
          it also includes some coarse-mode particles between 1 and 2.5 urn. PM2.S specifies a sample
          collected through a size-selective inlet with a specified penetration curve at 50% cut point at
          2.5 jim aerodynamic diameter.

2       •  Aerosol scientists use three different approaches or conventions in the classification of
          particles by size: (1) modes, based on the observed size distributions and formation
          mechanisms; (2) cut point, usually based on the 50% cut point of the specific sampling
          device, including legally specified, regulatory sizes for air quality standards; and
          (3) dosimetry or occupational health sizes, based on the entrance into various compartments
          of the respiratory system.

 3
 4      E.2.2  Sources of Airborne  Particles in the United States
 5       • Because of the complexity of the composition of ambient PM25 and PM10_25, sources are best
          discussed in terms of individual constituents of both primary and secondary PM25 and
          PM10_25. Each of these constituents can have anthropogenic and natural sources as shown in
          Table 3-8. Major components of fine particles are sulfates, strong acid, ammonium nitrate,
          organic compounds, trace elements (including metals), elemental carbon, and water.

 6       • Primary particles are emitted directly from sources. Secondary particles are formed from
          atmospheric reactions of sulfur dioxide (SO2), nitrogen oxides (NOJ, and certain organic
          compounds. NO reacts with ozone (O3) to form NO2. SO2 and NO2 react with hydroxy
          radical (OH) during the daytime to form sulfuric and nitric acid.  During the nighttime, N02
           reacts with ozone and forms nitric acid through a sequence of reactions involving the nitrate
           radical (NO3).  These acids may react further with ammonia to form ammonium sulfates and
           nitrates. Some types of higher molecular weight organic compounds react with OH radicals,
           and olefinic compounds also react  with ozone to form oxygenated organic compounds,

        June 2003                                 E-4        DRAFT-DO NOT QUOTE OR CITE

-------
10
            which nucleate or can condense onto existing particles.  SO2 also dissolves in cloud and fog
            droplets, where it may react with dissolved O3, H2O2, or, if catalyzed by certain metals, with
            O2, yielding sulfuric acid or sulfates, that lead to PM when the droplet evaporates.

            Organic compounds constitute from 10 to 70% of dry PM25 mass. Whereas the chemistry of
            participate nitrate and sulfate formation has been relatively well studied, the chemistry of
            secondary organic particulate matter formation is especially complex.  Although additional
            sources of secondary organic PM might still be identified, there appears to be a general
            consensus that biogenic compounds (monoterpenes, sesquiterpenes) and aromatic
            compounds (e.g., toluene and ethylbenzene) are the most significant precursors.
            Atmospheric transformations of the compounds, which are formed in the particle phase
            during the aging of particles, are still not adequately understood.

            The results of receptor modeling studies throughout the United States indicate that the
            combustion of fossil and biomass fuels is the major source of measured ambient PM25.
            Fugitive dust, found mainly in the PM10_25 range size, represents the largest source of
            measured ambient PMi0 in many locations in the western United States. The application of
            any of the source apportionment techniques is still limited by the availability of source
            profile data. Whereas the Chemical Mass Balance approach relies directly on source profile
            data, solutions from the Positive Matrix Factorization technique yield profiles for the factors
            that contribute to PM.

            The use of organic compounds in source apportionment studies could potentially result in the
            attribution of PM to many more source categories than is possible using only trace elements.
            However, in the relatively few studies of the composition of the organic fraction of ambient
            particles that have been performed, typically only about 10 to 20 % of organic compounds
            have been quantified. The  separation of contributions from diesel- and gasoline-fueled
            vehicles using organic marker compounds is still somewhat problematic. Additional efforts
            to develop protocols for extraction and analysis of organic markers are  needed.
       June 2003                                  E-5        DRAFT-DO NOT QUOTE OR CITE

-------
1     E.2.3  Atmospheric Transport and Fate of Airborne Particles
2       •  Primary and secondary fine particles have long lifetimes in the atmosphere (days to weeks)
          and travel long distances (hundreds to thousands of kilometers).  They tend to be uniformly
          distributed over urban areas and larger regions, especially in the eastern United States.  As a
          result, they are not easily traced back to their individual sources.

3       •  Coarse particles normally have shorter lifetimes (minutes to hours) and generally only travel
          short distances (<10's of km). Therefore, coarse particles tend to be unevenly distributed
          across urban areas and tend to have more localized effects than fine particles. However, dust
          storms occasionally cause long range transport of small coarse-mode particles.

4
5      E.2.4  Airborne Particle Measurement Methods
6       •  Measurements of ambient PM mass and chemical composition are needed to determine
          attainment of standards; to guide progress towards attainment of a standard (including
          determination of source categories and validation of air quality models); and to determine
          health, ecological, and radiative effects.  A comprehensive approach requires a combination
          of analytical techniques to assess (1) mass, (2) crustal and trace elements, (3) water-soluble
          ionic species including strong acidity, (4) elemental carbon, and (5) organic compounds.

 7       • There are no calibration standards for suspended particle mass; therefore, the accuracy of
 8         particle mass measurements cannot be definitively determined. The precision of particle
          mass measurements can be determined by comparing results from collocated samplers.
          Intercomparisons, using different techniques and samplers of different designs, coupled with
          mass balance studies (relating the sum of components to the measured mass), provide a
          method for gaining confidence in the reliability of PM measurements.

 9      • Mass concentration measurements with a precision of 10% or better have been obtained with
          collocated samplers of identical design. Field studies of EPA PM10 and PM25 reference
           methods and reviews of field data from  collocated PM10 and PM25 samplers show high
          precision (better than ± 10%). The use of more careful techniques, including double

        June 2003                                 E-6        DRAFT-DO NOT QUOTE OR CITE

-------
            weighing of filters, can provide higher precision and may be needed for precise
            determination of PM10_25 by difference.

 10      •  Analytical techniques exist for measurement of the mass and chemical composition of PM
            retained on a filter (nonvolatile mass) in terms of elements (except carbon) and certain key
            ions (sulfate, nitrate, hydrogen, and ammonium). Acceptable measurements can be made of
            the total carbon retained on a filter. However, the split into organic carbon and elemental
            carbon depends on the operational details of the analytical methods and varies somewhat
            among methods. Determination of the various organic compounds in the organic carbon
            fraction remains a challenge.

 11       •   Important components of atmospheric PM (particle-bound water, ammonium nitrate, and
            many organic compounds) are termed semivolatile because significant amounts of both the
            gaseous and condensed phases may exist in the atmosphere in equilibrium.  Particle-bound
            water is not considered a pollutant. Most of the particle-bound water is removed by heating
            the particles or by equilibration of the collected particles at a low relative humidity (40%)
            for 24 hours.  However, these processes also cause the loss of other semivolatile
           components.  Semivolatile components also evaporate from the filter during sampling due  to
           the pressure drop across the filter or due to a reduction in the atmospheric concentration
           during the sampling time.

12      •  Continuous methods must dry the PM to remove particle-bound water. If heating is used to
           dry the particles, more  of the semivolatile components may be removed than are lost in filter
           sampling.  Collection and retention of ammonium nitrate and semivolatile organic
           compounds represents a major challenge in the effort to move to continuous measurement of
           PM mass.  The use of diffusion dryers, which dehumidify the air stream without heating,
           represents a promising approach. Uncertainty in the efficiency of retention of ammonium
           nitrate and organic compounds on filters also affects source category attribution and
           epidemiologic studies.
      June 2003                                 E-7        DRAFT-DO NOT QUOTE OR CITE

-------
1      •  Semivolatile organic compounds and semivolatile ammonium compounds (such as NH4NO3)
          may be lost by volatilization during sampling. Such losses may be very important in
          woodsmoke impacted areas for organic compounds or in agricultural and other areas where
          low sulfate and high ammonia lead to high NH4NO3 concentrations. New techniques are
          now in use for measurement of nitrates and new research techniques are being tested for
          measurement of mass of semivolatile organic compounds in PM and of the total
          (semivolatile plus non-volatile) PM mass.  The Federal Reference Methods (FRM) for PM10
          and PM25 give precise (± 10%) measurements of "equilibrated mass." However, the loss of
          semivolatile PM (ammonium nitrate and organic compounds) and the possible retention of
          some particle-bound water in current PM mass measurements contribute to uncertainty in the
          measurement of the mass of PM as it exists suspended in the atmosphere.

2       •  Techniques are available to separate fine particles from coarse particles and collect the fine
          particles on a filter. No such technique exists for coarse particles. As yet, no consensus
          exists on the best technique for collecting a coarse particle sample for determination of mass
          and composition. Candidates include multistage impaction, virtual impaction,  and
          difference (subtracting PM25 mass or composition from PM10 mass or composition).
          Advances in the theory and practice of virtual impaction suggest that it should be possible to
          design virtual impactors with much less than the 10% of fine PM collected in the coarse PM
          sample, as is now the case for the dichotomous samplers used in air quality studies and with
          penetration curves as sharp as those used in the current FRM for PM2 5.

 3
 4      E.2.5  Ambient PM Concentrations in the U.S.
 5       • The recently deployed PM2 5 FRM network has provided data for a large number of sites
          across the United States. The data are stored in the Aerometric Information Retrieval
          System (AIRS).  Data have also been  collected at remote sites as part of the fMPROVE and
          NESCAUM networks. Annual mean U.S. PM25 concentrations from 1999 to 2000 range
          from about 5 ug/m3 to about 30 ^ig/m3. In the eastern United States, the data from 1999 to
          2001 indicate that highest quarterly mean concentrations and maximum concentrations most
          often occur during the summer. In the western United States, highest quarterly mean values

       June 2003                                E-8        DRAFT-DO NOT QUOTE OR CITE

-------
     and maximum values occur mainly during the winter at a number of sites although there
     were exceptions to these general patterns.  PM2.5 and PMIO concentrations many urban areas
     have generally declined over the past few decades. However, they appear to have leveled off
     in the past few years.

  •  The median PM25 concentration across the United States during 1999,2000, and 2001, the
     first three years of operation of the PM25 FRM network, was 13 ug/m3, with a 95th percentile
     value of 18 ug/m3. The corresponding median PM^s concentration was 10 ug/m3, with a
     95th percentile value  of 21 ng/m3.

  •  Although PM25 concentrations within a given Metropolitan Statistical Area (MSA) can be
     highly correlated between sites, there can still be significant differences in their
     concentrations. The  degree of spatial uniformity in PM25 concentrations and the strength of
     site to site correlations in urban areas varies across the country. These factors should be
     considered in using data obtained by  the PM2 5 FRM network to approximate community-
     scale human exposures, and caution should be exercised in extrapolating conclusions as to
     spatial uniformity or correlations obtained in one urban area to another. Limited information
     also suggests that the spatial variability in urban source contributions is likely to be larger
     than for regional source contributions to PM25 and for PM25, itself.

  •   Data for PMi0_25 concentrations are not as abundant as they are for PM25.  The difference
     method used in their  derivation is subject to the effects of uncertainties in measuring both
     PM10 and PM25.  As a result, estimates of PM10_25 concentrations, at times, come out as
     negative values, based on currently available data (e.g., in the EPA AIRS Database). In
     most cities where significant data is available, PMIO_25 is spatially less uniform than PM25.

  •  Ambient PM contains both primary and secondary  components. The results of ambient
    monitoring studies and receptor modeling studies indicate that PM2 5 is dominated by
    secondary components in the eastern United States. General statements about the origin of
    OC in ambient PM25  samples cannot yet be made and so the contribution of secondary
    components throughout the rest of the United States is still  highly uncertain. Primary
    constituents represent a smaller but still important component of PM25.  Crustal materials,
June 2003                                 E-9        DRAFT-DO NOT QUOTE OR CITE

-------
           which are primary constituents, constitute the largest measured fraction of PM10_25
           throughout the United States. Data for the concentration of bioaerosols in both the PM25 and
           PM10_2 5 size ranges are sparse.

10       •  Recent but limited information about policy-relevant background concentrations have not
           provided sufficient evidence to warrant any changes in estimates of the annual average
           background concentrations given in the 1996 PM AQCD. These are: 1 to 4 ^ig/m3 in the
           West and 2 to 5 ^g/m3 in the East for PM2 5; and approximately 3  ng/m3 in both the East and
           the West for PM10_2.5, with a range of 0 to 9 ng/m3 in the West and 0 to 7 ug/m3 in the East.
           Such concentrations are likely to be highly variable both spatially and temporally. Further
           information regarding the frequency distribution of 24-hour concentrations based on
           analyses of observations at relatively remote monitoring sites and on source apportionment
           analyses has become available and can be used for selected sites.

 11
 12      E.2.6  Human Exposure to PM
 13       •  Personal exposure to PM mass or its constituents results when individuals come in contact
           with particulate pollutant concentrations in locations or microenvironments that they
           frequent during a specific period of time. Various PM exposure metrics can be defined
           according to its source (i.e., ambient, nonambient) and the microenvironment where
           exposure occurs.

 14       •  Most people spend most of their time indoors where they are exposed to indoor-generated
           PM and ambient PM that has infiltrated indoors.

 15       •  Indoor-generated and ambient PM differ in sources, sizes, chemical composition, and
           toxic ity.

 16       -The ambient PM concentration and the indoor PM concentration can be measured by outdoor
           and indoor monitors. The total personal exposure can be measured by a personal exposure
           monitor carried by the person. However, the concentrations of indoor-generated PM and
        June 2003                                E-10       DRAFT-DO NOT QUOTE OR CITE

-------
  17
  18
 19
 20
21
   ambient PM that has infiltrated indoors and the related values of ambient and nonambient
   PM exposures must be estimated.

 •  The intercept of a regression of individual, daily values of total personal exposure on daily
   PM concentrations, gives the average nonambient PM exposure; and the slope gives the
   average attenuation factor (the ratio of ambient PM exposure to ambient PM concentration).

 -  Similarly, the intercept of a regression of individual, daily values of indoor PM concentration
   on daily ambient concentrations, gives the average concentration of indoor-generated PM;
   and the slope gives the average infiltration factor (concentrations of ambient PM that has
   infiltrated indoors/ambient PM concentration).

•  The attenuation factor and the infiltration factor depend on the penetration coefficient, the
   fraction of ambient PM that penetrates through the walls, doors, windows, etc.; the
   deposition or removal rate, a measure of how rapidly PM within the indoor
   microenvironment is removed by deposition to surfaces or by filtration in a heating/cooling
   system; and the air exchange rate, a measure of how rapidly indoor air is replaced by outdoor
  air. The attenuation factor also depends on the fraction of time spent outdoors.
i
  The air exchange rate is an important variable for determining the concentration of ambient
  PM found indoors. It can be measured by release and measurement of an inert tracer gas
  indoors. The air exchange rate increases with opening of windows or doors or operation of
 window or attic fans. It also increases as the indoor/outdoor temperature difference
 increases. For closed homes, i.e., no open windows or doors, the air exchange rate does not
 appear to be a function of wind speed or direction.

 The penetration coefficient and the deposition rate can be estimated from measurements of
 outdoor and indoor concentrations under conditions when there are no indoor sources (night
 time or unoccupied home). These parameters are functions of particle size. The penetration
 coefficient is high and the deposition rate is low for accumulation mode particles (0.1 to
 1.0 (im).  The penetration coefficient is lower and the deposition rate is  higher for ultrafine
 particles (< 0.1 urn) and coarse mode particles (> 1.0 urn).  The attenuation factor and the
       June 2003
                                                E-ll
                                                   DRAFT-DO NOT QUOTE OR CITE

-------
           infiltration factor are higher for particles in the accumulation mode than for ultrafme or
           coarse particles.

22       • The attenuation factor and the infiltration factor will vary as the air exchange rate does, and,
           therefore, will vary with season and housing characteristics. These factors will increase with
           increased opening of windows and doors. For closed homes, these factors will increase with
           an increase of the indoor/outdoor temperature difference; but they do not appear to be
           affected by wind speed or direction.

23       • The regression technique is useful for finding average values of the attenuation factor and the
           nonambient exposure and possibly for estimating the distribution of individual, daily values
           of the nonambient PM exposure.

24       • Individual, daily values of the ambient PM exposure, the nonambient PM exposure, and the
           attenuation factor may be determined from individual, daily values of the total PM personal
           exposure and daily ambient PM concentrations by several techniques:
25         - Mass balance technique. Direct measurement of the air exchange rate, measurement of
              the fraction of time spent outdoors by a diary of the subject's activity pattern, and use of
              the equilibrium mass balance equation for the attenuation factor with estimated values of
              the penetration coefficient and the deposition rate.

 26         -  Sulfate ratio technique. Individual, daily values of the attenuation factor (for PM2 5) will
               be given by individual, daily values of personal exposure to sulfate  /the daily ambient
               sulfate concentration provided there are no indoor sources of sulfate and sulfate and PM2.5
               have similar particle size distributions.

 27         _  Recursive technique.  Indoor-generated emissions, which tend to be episodic, can be
               removed from a continuous record of indoor PM concentration, allowing separation of
               indoor-generated PM and ambient PM that has infiltrated indoors.

 28       • In pooled studies (different subjects measured on different days), individual, daily values of
            total PM exposure are usually not well correlated with daily ambient PM concentrations.
         June 2003                                 E-12        DRAFT-DO NOT QUOTE OR CITE

-------
            In longitudinal studies (each subject measured for multiple days), individual, daily values of
            total PM personal exposure and daily ambient PM concentrations are highly correlated for
            some, but not all subjects.

 29       * Only one study has reported estimated individual, daily values of ambient and nonambient
            PM exposure. Individual, daily values of total PM personal exposure and daily ambient PM
            concentrations were poorly correlated. However, individual, daily values of ambient PM
            exposure and daily ambient PM concentrations were highly correlated. Individual daily
            values of ambient and nonambient PM exposure were not well correlated. Individual daily
            values of nonambient PM exposure and daily ambient PM concentrations were also not well
            correlated.

 30       -As long as the nonambient PM exposure is not correlated with the ambient PM exposure, it
            will not bias the estimated health effect of PM. However, the effect per ug/ambient PM
            concentration will be biased low compared to the health effect per ug/ambient PM exposure
            by the attenuation factor.  This effect probably explains some of the heterogeneity in PM10
            effects observed in multicity epidemiology studies, as indicated by a correlation of PM
            effects in different cities with air conditioning use in those cities (i.e., the higher the air
            conditioning use, the  lower the health effect estimate per ug/m3 of ambient PM).

 31       •  Exposure relationships also provide some insight into the issue of confounding.  While the
            data base is small, concentrations of gaseous co-pollutants, NO2, O3, and SO2 (and probably
            CO) are likely poorly correlated, and sometimes not significantly correlated, with personal
            exposure to the respective co-pollutant. However, they are frequently significantly
            correlated with both the ambient PM concentration and the ambient PM exposure. Thus, in a
            regression, where associations are found between gaseous co-pollutants and a health effect, it
            may be because they are a surrogate for PM rather than a confounder. That is, the health
            effect due to PM is transferred to the gaseous pollutant because of the positive correlation
            between the ambient concentration of the gas and the ambient PM exposure.

32
33
        June 2003                                E-13        DRAFT-DO NOT QUOTE OR CITE

-------
1     E.3   DOSIMETRY OF PARTICULATE MATTER
2      • Dosimetry establishes the relationship between PM exposure and the dose of inhaled PM
         delivered to and retained at the target site. Deposition, clearance, translocation, and retention
         comprise the essential elements of dosimetry.
3      • Dosimetric information is critical to extrapolating effects found in controlled exposure
         studies of laboratory animals to those observed in human exposure studies and for relating
         effects in normal healthy persons to those in potentially susceptible persons.

4      • Dosimetry separates the respiratory tract into three regions, extrathoracic (ET),
         tracheobronchial (TB), and alveolar (A), based on anatomical features and particle deposition
         and clearance phenomena that occur within each region.
5       • Particles in the accumulation mode size range (0.1 to 1.0 urn Dp) have the lowest deposition
         fraction in all three regions.
6       •  Coarse and ultrafine particles have higher fractional deposition. For coarse particles,
          fractional deposition peaks between 5 and 10  urn Dp for the TB region and 2.5 and 5 urn Dp
          for the A region.
 7       •  For ultrafine particles, fractional deposition peaks between 0.0025 and 0.005 urn Dp for the
          TB region and between 0.01 and 0.05 for the A region.
 8       •  A significant fraction of ultrafine and coarse particles, but not particles in the accumulation-
          mode size range, are deposited in the ET region.
 9      • Such transport could provide a mechanism whereby particles could affect cardiovascular
          function as reported in the epidemiologic studies
10      • Fractional deposition, as a function of particle size, depends on lung size, tidal volume, and
          breathing rate.  Exercising subjects receive higher doses of particles per cm2 of lung surface
          than subjects at rest.
        June 2003                                E-14        DRAFT-DO NOT QUOTE OR CITE

-------
  1       • Airway structure and physiological function vary with age. Such variations may alter the
           deposition patterns for inhaled particles. Airflow distribution is very uneven in diseased
           lungs, and deposition can be enhanced locally in areas of active ventilation. Total lung
           deposition is generally increased by obstructed airways so that particle deposition can be
           enhanced in people with chronic lung disease. Unfortunately, deposition studies in another
           susceptible population, the elderly, are still lacking.

 2       •  Particles depositing on airway surfaces may be completely cleared from the respiratory tract
           or translocated to other sites within this system by regionally specific clearance mechanisms.
           Clearance is either absorptive (dissolution) or nonabsorptive (transport of intact particles).
           Deposited particles may be dissolved in body fluids, taken up by phagocytic cells, or
           transported by the mucociliary system. Retained particles tend to be small (< 2.5 urn) and
           poorly soluble (e.g., silica, metals). Ultrafme particles can be rapidly cleared from the lungs
           into the systemic circulation where they can be transported to extrapulmonary regions.

 3      •  Tracheobronchial clearance has both a fast and a slow component. In the fast phase particles
           deposited in the TB region clear out rapidly during the first several hours and continue to
           clear out for 24 hours.  A small remaining portion may clear out over several days (slow
          phase).  Translocation of poorly soluble PM to the lymph nodes takes a few days and is more
          rapid for smaller (< 2 jim) particles; elimination rates of these retained particles are on the
          order of years. People with COPD have increased particle retention partly because of
          increased initial deposition and impaired mucociliary clearance and use cough to augment
          mucociliary clearance.

4       •  Alveolar clearance takes months to years. Particles may be taken up by alveolar
          macrophages within 24 hours, but some phagocytosed macrophages translocate into the
          interstitium or lymphatics whereas some remain on the alveolar surface. Penetration of
          uningested particles into the interstitium increases with increasing particle load and results in
          increased translocation to lymph nodes.

5       • Acute effects of PM are probably best related to deposited dose; whereas chronic effects may
         be related to  cumulative or retained dose.  Retention of particles is a function of depositi<
tion
      June 2003                                  E-15        DRAFT-DO NOT QUOTE OR CITE

-------
          site, clearance of particles by macrophages or the mucociliary system, and particle
          characteristics, especially solubility. Chronic effects may also arise from recurring cycles of
          pulmonary injury and repair.

6       •  Computational models allow calculation of fractional deposition and dose per cm2 of lung
          surface as a function of particle size and respiratory parameters for humans and some animals
          (including the laboratory rat). Such calculations can be used to predict the exposures needed
          to produce comparable doses for animal to human extrapolation. Computational models have
          been improved in recent years but experimental validation of model predictions is still
          required.

 7
 8      E.4  TOXICOLOGY OF PARTICULATE MATTER IN HUMANS AND
 9            LABORATORY ANIMALS
10      E.4.1  Health Effects of Specific PM Components
11       •  There is relatively little new information on the effects of acid aerosols, and the conclusions
           of the 1996 PM AQCD, i.e., Air Quality Criteria for Particulate Matter (U.S. Environmental
           Protection Agency, 1996a), are unchanged. It was previously concluded that acid aerosols
           cause little or no change in pulmonary function in healthy subjects, but asthmatics may
           develop small changes in pulmonary function. Although pulmonary effects of acid aerosols
           have been the subject of extensive research in past decades, the cardiovascular effects of acid
           aerosols have received little attention and should not be ruled out as possible mediators of
           PM health effects.
12       • Health effects of particle-associated soluble metals have been demonstrated by in vivo and
           in vitro studies using residual oil fly ash (ROFA) or soluble transition metals. Although
           there are some uncertainties about differential effects of one transition metal versus another,
           water soluble metals leached from ROFA or ambient filter extracts have been shown
           consistently (albeit at high concentrations) to cause cell injury and inflammatory changes
           in vitro and in vivo.
        June 2003                                E-16        DRAFT-DO NOT QUOTE OR CITE

-------
 1       •  Even though it is clear that combustion particles that have a high content of soluble metals
           can cause lung injury and even death in compromised animals and correlate well with
           epidemiologic findings in some cases (e.g., Utah Valley Studies), it has not been fully
           established that the small quantities of metals associated with ambient PM are sufficient to
           cause health effects. Moreover, it cannot be assumed that metals are the only or the primary
           toxic component of ambient PM,  rather than there possibly being many different toxic agents
           contributing to ambient PM health effects.

 2       •  There is growing toxicological evidence that diesel exhaust particles exacerbate the allergic
           response to inhaled antigens. The organic fraction of diesel exhaust has been linked to
           eosinophil degranulation and induction of cytokine production, suggesting that the organic
           constituents of diesel PM are the responsible part for the immune effects.  It is not known
           whether the adjuvant-like activity of diesel PM is unique or whether other combustion
           particles have similar effects. It is important to compare the immune effects of other
           source-specific emissions, as well as concentrated ambient PM, to diesel PM to determine
           the extent to which exposure to diesel exhaust may contribute to the incidence and  severity
           of allergic rhinitis and asthma.

 3       •  Published research on the acute effects of particle-associated organic carbon constituents is
           conspicuous by its relative absence, except for diesel exhaust particles.

 4       •  Studies with ultrafine particles have demonstrated a significantly greater inflammatory
          response than that seen with fine particles of the same chemical composition at similar mass
          doses.  In other more limited studies,  ultrafines also have generated greater oxidative stress
          in experimental animals.  However, when the particle surface area is used as a dose metric,
          the inflammatory response to both fine and ultrafine particles may be basically the same.
          Thus, it may be the higher surface area of ultrafine particles is the important factors
          contributing to health effects.

5       •  Concentrated ambient particle (CAPS) studies should be among the most relevant in helping
          to understand the characteristics of PM producing toxicity, susceptibility of individuals to
          PM, and the underlying mechanisms.  Studies have used collected urban PM for intratracheal
          administration to healthy and compromised animals.  Despite the difficulties in extrapolating

      June 2003                                 E-17        DRAFT-DO NOT QUOTE  OR CITE

-------
          from the bolus delivery used in such studies, they have provided strong evidence that the
          chemical composition of ambient particles can have a major influence on toxicity.

6       •  Ambient particle concentrators, that concentrate particles in situ, provide particles for
          inhalation studies.  The concentration of particles without collection avoids the changes in
          size and composition that occur during collection during collection and resuspension.
          Studies with inhaled CAPs have observed cardiopulmonary changes in rodents and dogs at
          high concentrations of PM between 0.15 and 2.5 um diameter. A new generation of ambient
          particle concentrators that allow separation and concentration of coarse-mode, accumulation-
          mode, and ultrafme particle separately will permit the direct toxicological comparison of
          these various ambient particle sizes.
7       • Recent studies support the conclusion of the 1996 PM AQCD, which stated that bioaerosols,
          at concentrations present in the ambient environment, cannot account for the reported health
          effects of ambient PM. However, it is possible that bioaerosols could contribute to the health
          effects of PM.

 8
 9     E.4.2  Mechanisms  of Action
10     E.4.2.1   Cardiovascular Effects
11      •  Changes in heart rate, heart rate variability, and conductance associated with ambient PM
           exposure have been  reported in animal studies, in several human panel studies, and in a
           reanalysis of data  from the MONICA study.  Some of these studies included endpoints
           related to respiratory effects but few significant adverse respiratory changes were detected.
           This raises the possibility that ambient PM may have effects on the heart that are
           independent of adverse changes in the lung.
12       •  Inhaled paniculate matter affects the heart by uptake of particles into the circulation
           or release of a soluble substances into the circulation.

13       -  Inhaled paniculate matter affects autonomic control of the heart and cardiovascular system.

14
15
        June 2003                                E-18       DRAFT-DO NOT QUOTE OR CITE

-------
  E.4.2.2  Respiratory Effects
   •  Particularly compelling evidence pointing towards ambient PM causing lung injury and
     inflammation derives from the study of ambient PM materials on filter extracts collected
     from community air monitors before, during the temporary closing of a steel mill in Utah
     Valley, and after its reopening. Intratracheal instillation of filter extract materials in human
     volunteers provoked greater lung inflammatory responses for materials obtained before and
     after the temporary closing versus that collected during the plant closing. Further, the
     instillation  in rats of extract materials from before and after the plant closing resulted in a
     50% increase in air way hyperresponsiveness to acetylcholine compared to 17 or 25%
     increases with saline or extract materials for the period when the plant was closed,
     respectively. Analysis of the extract materials revealed notably greater quantities of metals
     for when the plant was opened suggesting that such metals (e.g., Cu, Zn, Fe, Pb, As, Mn, Ni)
     may be important contributors to the pulmonary toxicity observed in the controlled exposure
     studies, as well as to health effects shown epidemiologically to vary with PM exposures of
     Utah Valley residents before, during, and after the steel mill closing.

  •  Still other toxicological studies point towards lung injury and inflammation being associated
     with exposure of lung tissue to complex combustion-related PM materials, with metals again
     being likely contributors.  Rats with SOrinduced bronchitis and monocrotaline-treated rats
     have been reported to have a greater inflammatory response to concentrated ambient PM
     than normal rats.  These studies suggest that exacerbation of respiratory disease by ambient
     PM may be  caused in part by lung injury and inflammation.

  •   Paniculate air pollution causes increased susceptibility to respiratory infections. Exposure
     of rats to ROFA and L monocytogenes, a bacterial pathogen, led rats treated with saline or
    L monocytogenes in the absence of ROFA. Preexposure of rats to ROFA significantly
    enhanced injury and delayed the pulmonary clearance of a subsequent challenge of L
    monocytogenes, when compared to saline-treated control rats. Acute exposure to ROFA
    appeared to slow pulmonary clearance of L. monocytogenes and to alter AM function.  Such
    changes could lead to increased susceptibility to lung infection in exposed populations.
June 2003
E-19
                                                      DRAFT-DO NOT QUOTE OR CITE

-------
1       •  Participate air pollution increases airway reactivity and exacerbates asthma.  Diesel
          paniculate matter (DPM) has been shown to increase production of antigen-specific IgE in
          mice and humans. In vitro studies have suggested that the organic fraction of DPM is
          involved in the increased IgE production. ROFA leachate also has been shown to enhance
          antigen-specific airway reactivity in mice, indicating that soluble metals can also enhance an
          allergic response.
2
3      E.4.2.3  Systemic Effects Secondary to Lung Injury
4       • Lung injury from  inhaled PM causes impairment of oxygenation and increased work of
          breathing that adversely affects the heart.
5       • Lung inflammation and cytokine production cause adverse systemic hemodynamic effects.

6       • Lung inflammation from  inhaled PM causes increased blood coagulability that increases the
          risk of heart attacks and strokes.

7       • Interaction of PM with the lung affects hematopoiesis.

 8
 9      E.4.3  Susceptibility
10       • Older animals or animals with certain types  of compromised health, either genetic or
          induced, are more susceptible to instilled or inhaled particles,  although the increased animal-
          to-animal variability in these models has created greater uncertainty for the interpretation of
          the findings. Moreover, because PM seems to affect broad categories of disease states,
          ranging from cardiac arrhythmias to pulmonary infection, it can be difficult to know what
          disease models to use in evaluating the biological plausibility of adverse health effects of
          PM. Nevertheless, particularly interesting new findings point toward ambient PM
          exacerbation of allergic airway hyperresponsiveness and/or antigen-induced immune
          responses. Both  metals and diesel particles have been implicated, with an expanding array of
           new studies showing DPM in particular as being effective in exacerbating allergic asthma
           responses.
        June 20o3                                 E-20        DRAFT-DO NOT QUOTE OR CITE

-------
  1     E.5  EPIDEMIOLOGY OF HUMAN HEALTH EFFECTS ASSOCIATED
  2           WITH AMBIENT PARTICIPATE MATTER
  3           Chapter 8 of this document assesses the extensive PM epidemiologic data base that has
  4     become newly available since the 1996 PM AQCD.  The most important types of additions to
  5     the epidemiologic database beyond that assessed in the 1996 PM AQCD include the following:
  6       •  New multi-city studies on a variety of endpoints which provide more precise estimates of the
            average PM effect sizes than most smaller-scale  individual city studies;

  7       •  More studies of various health endpoints using ambient PM10 and/or closely related mass
            concentration indices (e.g., PM13 and PM7), which substantially lessen the need to rely on
            non-gravimetric indices (e.g., BS or CoH);

  8       •  New studies evaluating relationships  of a variety of health endpoints to the ambient PM
            coarse fraction (PM10_2 5), the ambient fine-particle fraction (PM2 5), and even ambient
            ultrafme particles measures (PM0_, and smaller),  using direct mass measurements and/or
            estimated from site-specific calibrations;

  9       •  A few new studies in which the relationship of some health endpoints to ambient particle
            number concentrations were evaluated;

10       •  Many new studies which evaluated the sensitivity of estimated PM effects to the inclusion of
            gaseous co-pollutants in the model;

11       •  Preliminary attempts to evaluate the effects combinations or mixtures of air pollutants
            including PM components, based on factor analysis or source profiles;

12       • Numerous new studies of cardiovascular endpoints, with particular emphasis on assessment
           of cardiovascular risk factors as well as symptoms;

13       • Additional new studies on asthma and other respiratory conditions potentially exacerbated
           by PM exposure;

14       • New analyses of lung cancer associations with long-term exposures to ambient PM;
       June 2003                                E-21        DRAFT-DO NOT QUOTE OR CITE

-------
1       • New studies of infants and children as a potentially susceptible population.

2           In addition, Chapter 8 discusses statistical issues posed by use of General Additive Mode
3      (GAM) analyses involving default convergence criteria in widely used commercially-available
4      software employed in time-series analyses of ambient PM-health effects relationship reported in
5      a few published studies assessed in the 1996 PM AQCD and in numerous more recent studies
6      assessed in the present PM AQCD. This includes discussion of the following key points:
7        - Use of GAM analyses with default convergence criteria (GAM default) has variable impacts
          on PM effect size estimates from study to study, depending on many factors (numbers of
          observations, numbers of potential effect modifiers or potential confounders included,
          specific "smoothing" approaches used to control for their effects, numbers of degrees of
          freedom used, etc.). The effect of GAM (default) use tends most often to be some
          (unusually only slight) overestimation of the PM effect size compared to results obtained
          with use of GAM with stringent convergence criteria or other appropriate modeling
          approaches, e.g., general linear models (GLM).

 8        • The results of EPA-encouraged reanalyses of a number of (> 35) of important PM time-
          series analyses comparing PM effect sizes and standard error (confidence interval) estimates
          from GAM (default) analyses versus GAM (stringent) or other appropriate statistical
          approaches (GLM).  These reanalyses appear in a Health Effects Institute Special Report
          (HEI, 2003) that includes not only short communications on the GAM reanalyses by the
           original investigators, but also commentaries on the reanalyses and their implications for
           interpreting the PM time-series analyses results by a special peer-review panel convened by
           HEI at EPA's request.

 9
10      E.5.2  Key Epidemiologic Findings
11            The epidemiologic studies discussed in Chapter 8 demonstrate biologically-plausible
12      responses in humans exposed at ambient concentrations. These observational study findings are
13      further enhanced by supportive findings of causal studies from other scientific disciplines
14      (dosimetry, toxicology, etc.), in which other factors could be experimentally controlled, as

        June 2003                                E-22        DRAFT-DO NOT QUOTE OR CITE

-------
 1      discussed in Chapters 6 and 7.  The most salient findings derived from the PM epidemiologic
 2      studies include the following:
 3        •  A large and reasonably convincing body of epidemiology evidence confirms earlier
           associations between short- and long-term ambient PM10 exposures (inferred from stationary
           community air monitor measures) and mortality/morbidity effects and suggest that PM10
           (or one or more PMi0 components) is a probable contributing cause of adverse human health
           effects.

4        •  There appears to be some spatial heterogeneity in city-specific excess risk estimates for the
           relationships between short-term ambient PM10 concentrations and acute health effects. The
           reasons for such variation in effects estimates are not well understood at this time but do not
           negate ambient PM's likely causative contribution to observed PM-mortality and/or
           morbidity associations in many locations. Possible factors contributing to the apparent
           heterogeneity include geographic differences in air pollution mixtures, composition of PM
           components, and personal and sociodemographic factors affecting PM exposure (such as use
           of air conditioners, education, and so on).

5       - A growing body of epidemiologic studies confirm associations between short- and long-term
          ambient PM2 5 exposures (inferred from stationary air monitor measures) and adverse health
          effects and suggest that PM2 5 (or one or more PM2 5 components) is a probable contributing
          cause of observed PM-associated health effects. Some new epidemiologic findings also
          suggest that health effects are associated with mass or number concentrations of ultrafme
          (nuclei-mode) particles, but not necessarily more so than for other ambient fine PM
          components.

>       • A smaller body of evidence appears to support an association between short-term ambient
         thoracic coarse fraction (PM10_2.5) exposures (inferred from stationary air monitor measures)
         and short-term health effects in epidemiologic studies. This suggests that PM10_25, or some
         constituent component(s) of PM10_25, may be a contributory cause of adverse health effects
         in some locations.  Reasons for differences among findings on coarse-particle health effects
         reported for different cities are still poorly understood, but several of the locations where
     June 2003                                 E-23        DRAFT-DO NOT QUOTE OR CITE

-------
  significant PM10.2 s effects have been observed (e.g., Phoenix, Mexico City, Santiago) tend to
  be in drier climates and may have contributions to observed effects due to higher levels of
  organic particles from biogenic processes (endotoxins, molds, etc.) during warm months.
  Other studies suggest that particles of crustal origin are generally unlikely to exert notable
  health effects under most ambient exposure conditions. Also, in some western U.S. cities
  where PM10_25 is a large part of PM10, the relationship between hospital admissions and PM10
  may be an indicator of response to coarse thoracic particles from wood burning.

 • Long-term PM exposure durations, on the order of months to years, as well as on the order
  of a few days, are statistically associated with serious human health effects (indexed by
  mortality, hospital admissions/medical visits, etc.). More chronic PM exposures, on the
  order of years or decades, appear to be associated with life shortening well beyond that
  accounted for by the simple accumulation of the more acute effects of short-term PM
  exposures (on the order of a few days).  Substantial uncertainties remain regarding the
  magnitude of and mechanisms underlying chronic health effects of long-term PM exposures
   and the relationship between chronic exposure and acute responses to short-term exposure.

 • Recent investigations of the public health implications of such chronic PM exposure-
   mortality effect estimates were also reviewed. Life table calculations by Brunekreef (1997)
   found that relatively small differences in long-term exposure to airborne PM of ambient
   origin can have substantial effects on life expectancy. For example, a calculation for the
   1969-71 life table for U.S. white males  indicated that a chronic exposure increase of
   10 ^ig/m3 PM was associated with a reduction of 1.3 years for the entire population's life
   expectancy at age 25. Also, new evidence of associations of PM exposure with infant
   mortality and/or intrauterine growth retardation and consequent increase risk for many
   serious health conditions associated with low birth weight, if further substantiated, would
   imply that life shortening in the entire population from long-term PM exposure could well
   be significantly larger than that estimated by Brunekreef (1997).

 •  Considerable coherence exists among effect size estimates for ambient PM health effects.
   For example, results derived from several multi-city studies, based on pooled analyses of

June 2003                                 E-24        DRAFT-DO NOT QUOTE OR CITE

-------
            data combined across multiple cities (thought to yield the most precise estimates of mean
            effect size), show the percent excess total (non-accidental) deaths estimated per 50 ug/m3
            increase in 24-h PM10 to be -1.4% in the 90 largest U.S. cities with the estimate for the
            Northeast being the largest (approximately twice the nationwide estimate); -3.4% in
            10 large U.S. cities; -3.6% in the 8 largest Canadian cities; and -3.0% in western European
            cities.  These combined estimates are consistent with the range of PM10 estimates previously
            reported in the 1996 PM AQCD. These and excess risk estimates from many other
            individual-city studies, generally falling in the range of ca. 1.5 to 8.0% per 50 jig/m3 24-h
            PM10 increment, also comport well with numerous new studies confirming increased cause-
            specific cardiovascular- and respiratory-related mortality. They are also coherent with larger
            effect sizes reported for cardiovascular and respiratory hospital admissions and visits,  as
            would be expected for these morbidity endpoints versus those for PMIO-related mortality.

10        •  Several independent panel studies (but not all) that evaluated temporal associations between
            PM exposures and measures of heart beat rhythm in elderly subjects provide generally
            consistent indications of decreased heart rate variability (HRV) being associated with
            ambient PM exposure (decreased HRV being an indicator of increased risk for serious
            cardiovascular outcomes, e.g., heart attacks). Other studies point toward changes in blood
            characteristics (e.g., C-reactive protein levels) related to increased risk of ischemic heart
            disease also being associated with ambient PM exposures.  However, these heart rhythm and
            blood characteristics findings should currently be viewed as providing only limited or
            preliminary support for PM-related cardiovascular effects.

11        •  Notable new evidence now exists which substantiates positive associations between ambient
            PM concentrations and increased respiratory-related hospital admissions, emergency
            department, and other medical visits, particularly in relation to PM10 levels. Of much
            interest are new findings tending to implicate not only fine particle components but also
            coarse thoracic (e.g., PM10_25) particles as likely contributing to exacerbation of asthma
            conditions.  Also of much interest are emerging new findings indicative of likely increased
            occurrence of chronic bronchitis in association with (especially chronic) PM exposure. Also
           of particular interest are reanalyses or extensions of earlier prospective cohort studies of
       June 2003                                 E-25        DRAFT-DO NOT QUOTE OR CITE

-------
           long-term ambient PM exposure effects which demonstrate substantial evidence for
           association of increased lung cancer risk with such PM exposures, especially exposure to
           fine PM or its subcomponents.

12       • One major methodological issue affecting epidemiologic studies of both short-term and
           long-term PM exposure effects is that ambient PM of varying size ranges is typically found
           in association with other air pollutants, including gaseous criteria pollutants (e.g, O3, NO2,
           SO2, CO), air toxics, and/or bioaerosols.  Available statistical methods for assessing
           potential confounding arising from these  associations may not yet be fully adequate.  The
           inclusion of multiple pollutants often produces statistically unstable estimates.  Omission of
           other pollutants may incorrectly attribute their independent effects to PM.  Second-stage
           regression methods may have certain pitfalls that have not yet been fully evaluated. Much
           progress in sorting out relative contributions of ambient PM components versus other
           co-pollutants is nevertheless being made  and, overall, tends to substantiate that observed PM
           effects are at least partly due to ambient PM acting alone or in the presence of other
           covarying gaseous pollutants.  However,  the statistical association of health effects with PM
           acting alone or with other pollutants should not be taken as an indicator of a lack of effect of
           the other pollutants.

13        • It is possible that differences in observed health effects will be found to depend on site-
           specific differences hi chemical and physical composition characteristics of ambient
           particles and on factors affecting exposure (such as air conditioning) as well as on
           differences in PM mass concentration. For example, epidemiologic Utah Valley studies
            showed that PM10 particles, known to be richer in metals during exposure periods while the
            steel mill was operating, were more highly associated with adverse health effects than was
            PM10 during the PM exposure reduction while the steel mill was closed. In contrast, PM10 or
            PM25 was relatively higher in crustal particles during windblown dust episodes in Spokane
            and in three central Utah sites than at other times, but was not associated with higher total
            mortality. These differences require more research that may become more feasible as the
            PM25 sampling network produces air quality data related to speciated samples.
        June 2003                                 E-26        DRAFT-DO NOT QUOTE OR CITE

-------
  1       •  The above reasons suggest it is inadvisable to pool PM epidemiologic studies for different
            locations or time periods, with different population sub-groups, or different health endpoints
            (as is often done in "meta-analyses"), without careful assessment of potential causes and
            consequences of these differences and appropriate caveating of results. Published multi-city
            analyses using common data bases, measurement methods, analytical strategies, and
            extensive independent external review (as carried out in the APHEA and NMMAPS studies)
            are useful.  Pooled analyses of more diverse collections of independent studies of different
            cities, using varying methodology and/or data quality or representativeness, are likely less
           credible and should not, in general, be used without careful assessment of their underlying
           scientific comparability.

 2       •  It may be possible that different PM size components or particles with different composition
           or sources produce effects by different mechanisms manifested at different lags or that
           different preexisting conditions may lead to different delays between exposure and effect
           Thus, although maximum effect sizes for PM effects have often been reported for 0-1 day
           lags, evidence is also beginning to suggest that more consideration should be given to lags of
           several days.  Also, if it is considered that all health effects occurring at different lag days
           are all real effects, so that the risks for each lag day should be additive, then overall risks
           may exist that are higher than implied by maximum estimates for any particular single or
           two-day lags. In that case, multi-day averages or distributed lag models may provide more
           accurate estimate of the total impact of PM on the population.

3       -  Certain classes of ambient particles may be distinctly less toxic than others and may not
          exert human health effects at typical ambient exposure concentrations or only under special
          circumstances. Coarse thoracic particles of crustal origin, for example, may be relatively
          non-toxic under most circumstances compared to those of combustion origin such as wood
          burning. However, crustal particles may be sufficiently toxic to cause human health effects
          under some conditions; resuspended crustal particles, for example, may carry toxic trace
          elements and other components from previously deposited fine PM, e.g., metals from
          smelters (Phoenix) or steel mills (Steubenville, Utah Valley), PAH's from automobile
         exhaust, or pesticides from administration to agricultural lands.  Likewise, fine particles


      June 2003                                E_27        DRAFT-DO NOT QUOTE OR CITE

-------
  from different sources have different effect sizes. More research is needed to identify
  conditions under which one or another class of particles may cause little or no adverse health
  effects, as well as conditions under which different classes of particles may cause relatively
  more notable effects.

  Certain epidemiologic evidence from "intervention studies" suggests that reducing ambient
  PMIO concentrations may reduce a variety of health effects on a time scale from a few days
  to a few months, as found in epidemiologic studies of "natural experiments" such as in the
  Utah Valley, and supported by toxicology studies using particles extracts from ambient
  community sampling filters from the Utah Valley. Recent studies in Germany and in the
  Czech Republic also tend to support a hypothesis that reductions in particle-related air
  pollution are associated with reductions in the incidence of adverse health effects.

   Studies that combine the features of cross-sectional and cohort studies provide the best
   evidence for chronic effects of PM exposure.  Gauderman et al. (2000; 2002) have found
   significant decreases in lung function growth related to PM10 levels using these techniques.

   Adverse health effects in children reported in a limited number of more recent studies are
   emerging as an important area of more concern than was the case in the 1996 PM AQCD.
   Unfortunately, relatively little is known about the relationship of PM to the most serious
   health endpoints, neonatal and infant mortality,  emergency hospital admissions and
   mortality in older children, as well as low birth weight and preterm birth.

  • Little is yet known about involvement of PM exposure in the progression from less serious
   childhood conditions, such as asthma and respiratory symptoms, to more serious disease
   endpoints later in life. This is an important health issue because childhood illness or death
   may cost a very large number of productive life-years.

  . Lastly, new epidemiologic studies of ambient PM associations with increased non-hospital
   medical visits (physician visits) and asthma effects suggest likely much larger health effects
   and costs to society due to ambient PM than just those indexed by mortality and/or hospital
   admissions/visits.
June 2003                                 E-28        DRAFT-DO NOT QUOTE OR CITE

-------
 1      E.6  ENVIRONMENTAL EFFECTS OF PM
 2      E.6.1  Vegetation and Ecosystem Effects
 3       *  Deposition of particulate matter from the atmosphere has the potential to alter ecosystem
          structure and function. Human existence on this planet depends on the essential life-support
          services that ecosystem structure and functions provide. Concern has risen in recent years
          regarding the consequences of changing the biological diversity of ecosystems because
          human activities are creating disturbances that are altering the structure (complexity and
          stability) and functioning (producing changes in the processes of energy and water flow and
          nutrient cycling) of ecosystems.

 4       •  Human-induced changes in biotic diversity and alterations in the structure and ecosystems
          processes are the two most dramatic ecological trends in the past century.  Biodiversity
          encompasses all levels of biological organization, including individuals, populations, species,
          and ecosystems. For this reason, there is a need to understand the effects of PM deposition
          on vegetation and ecosystems and biodiversity.

 5       •  Ecosystem functions maintain clean water, pure air, biodiversity, and impart the following
          benefits: fixation of solar energy, absorption and breakdown of pollutants, cycling of
          nutrients, binding of soil, degradation of organic wastes, maintenance of a balance of
          atmospheric gases, regulation of radiation balance, and climate.

6      •  The deposition of PM onto vegetation and soil, depending on its chemical composition
          (acid^ase, trace metal, or nutrients, e.g., nitrates or sulfates), can produce direct or indirect
          responses within an ecosystem.

7      •  Ecosystem response to pollutant deposition is a direct function of the level of sensitivity of
          the ecosystem and its ability to ameliorate resulting changes.  Changes in ecosystem
          structural patterns and the functioning of ecological processes must be scaled in  both time
          and space and propagated to the more complex levels of community interaction to produce
          observable ecosystem changes.
      June 2003                                 E-29        DRAFT-DO NOT QUOTE OR CITE

-------
  The major effect of atmospheric PM on ecosystems is indirect and occurs through the
  deposition of nitrates and sulfates and the acidifying effects of the associated IT ion in wet
  and dry deposition in the soil environment.  In the soil, they have the potential to inhibit
  nutrient uptake, alter the ecological processes of energy flow and nutrient cycling, change
  ecosystem structure, and affect ecosystem biodiversity.  Most nitrate is not deposited or
  measured as PM but is a combination of wet and dry deposition.

  The soil environment, one of the most dynamic sites of biological interaction, is inhabited by
  communities of bacteria, fungi, and actinomycetes. Bacteria, as essential components of the
  nitrogen and sulfur cycles, and fungi in the uptake of mineral nutrients play an important role
  in plant nutrition. Their actions make these elements available for plant uptake.  Fungi form
  mycorrhizae, a mutualistic symbiotic relationship with plant roots that is integral to the
  uptake of mineral nutrients.  Changes in the soil environment resulting from deposition of
  nitrates and sulfates that alter the role of the bacteria in the nutrient cycles and mycorrhizal
  fungi in making minerals available for plant utilization, determine plant and, ultimately,
  ecosystem response.

  Although nitrogen as molecular nitrogen (N2) is the most abundant element in the
  atmosphere, it only becomes available to support plant growth after its conversion into
  reactive forms.  In nature, nitrogen may be divided into two groups:  reactive (Nr) and
  nonreactive (N^. Reactive nitrogen or Nr includes the inorganic reduced forms of nitrogen
   (e.g., ammonia  [NH3] and ammonium [NH4+]), inorganic oxidized forms (e.g., nitrogen
   oxides [NOJ, nitric acid [HNO3], nitrous oxide [N2O], and nitrate [NO/]), and organic
   compounds (e.g., urea, amine, proteins, and nucleic acids).

   Reactive nitrogen can be widely dispersed and accumulate in the environment when the rates
   of its formation exceed the rates of removal via denitrification. As a result of human food
   production, it is now accumulating in the environment on all spatial scales - local, regional
   and global; and its creation and accumulation is projected to increase as per capita use of
   resources by human populations increases.
June 2003                                 E-30        DRAFT-DO NOT QUOTE OR CITE

-------
   •  The major changes in the nitrogen cycle can be both beneficial and detrimental to the health
     and welfare of humans and ecosystems. The cascade of environmental effects resulting from
     reactive nitrogen accumulation include the following: (1) production of tropospheric ozone
     and aerosols that induce human health and ecosystem problems; (2) increases in the
     productivity in forests and grasslands followed by decreases wherever Nr deposition
     increases significantly and exceeds critical thresholds; (3) reactive nitrogen additions
     probably decrease biodiversity in many natural habitats; (4) in association with sulfur is
     responsible for acidification and loss of biodiversity in lakes and streams in many regions of
     the world; (5) eutrophication, hypoxia, loss of biodiversity, and habitat degradation in coastal
     ecosystems. [Eutrophication is now considered the biggest pollution problem in coastal
     waters.], and (6) contributes to global climate change and stratospheric ozone depletion,
     which can in turn affect ecosystems and human health.

     "Nitrogen saturation" results when reactive nitrogen concentrations exceed the capacity of a
     system to utilize it. Saturation implies that some resource other than nitrogen is limiting
     biotic function. Water and phosphorus for plants and carbon for microorganisms are most
     likely to be the secondary limiting factors. The appearance of nitrogen in soil solution is an
                     s
     early symptom of excess nitrogen. In the final stage, disruption of ecosystem structure
     becomes visible.

    Possible ecosystem responses to nitrogen saturation include (1) a permanent increase in foliar
    nitrogen and reduced foliar phosphorus and lignin caused by the lower availability of carbon,
    phosphorus, and water; (2) reduced productivity in conifer stands because of disruptions of
    physiological function; (3) decreased root biomass and increased nitrification  and nitrate
    leaching; (4) reduced soil fertility, resulting from increased cation leaching and increased
    nitrate and aluminum concentrations in streams, (5) decreased water quality, and (6) changes
    in soil microbial communities from predominantly fungal (mycorrhizal) communities to
    those dominated by bacteria during saturation.
June 2003
E-31
                                                       DRAFT-DO NOT QUOTE OR CITE

-------
>  Intensive research over nearly a decade indicates that, although the soils of most
  North American forests are nitrogen limited, severe symptoms of nitrogen saturation have
  been observed in the high-elevation, non-aggrading spruce-fir ecosystems in the Appalachian
  Mountains; the eastern hardwood watersheds at the Fernow Experimental Forest near
  Parsons, WV; forests in southern California; the southwestern Sierra Nevada in Central
  California; and the Front Range in northern Colorado. The mixed conifer forest and
  chaparral watershed with high smog exposure in the Los Angeles Air Basin exhibit the
  highest stream water NO3" concentrations in wilderness areas of North America.

• Increases in soil nitrogen play a selective role in ecosystems. Plant succession patterns and
  biodiversity are affected significantly by chronic nitrogen additions in some North American
  ecosystems.  Long-term nitrogen fertilization studies in both New England and Europe
  suggest that some forests receiving chronic inputs of nitrogen may decline in productivity
  and experience greater mortality. Declining coniferous forest stands with slow nitrogen
  cycling may be replaced by deciduous fast-growing forests that rapidly cycle nitrogen.

• In experimental studies nitrogen deposition over a 12-year period Minnesota grasslands
  dominated by native warm-season grasses shifted to low-diversity mixtures dominated by
  cool-season  grasses.

 •  Excess nitrogen inputs to  unmanaged heathlands in the Netherlands has  resulted in
   nitrophilous grass species replacing slower growing heath species. Over the past several
   decades the  composition of plants in the forest herb layers has been shifting toward species
   commonly found in nitrogen-rich areas. It also was observed that the fruiting bodies of
   mycorrhizal fungi had decreased in number, indicating that formation of mycorrhizae were
   affected.

 •  Soil nitrogen enrichment  of the soil significantly alters the composition and richness of the
   arbuscular mycorrhizal fungal community and markedly decreases overall diversity of the
   mycorrhizal community.  Decline in the coastal sage scrub in southern California was
   directly linked to the decline of arbuscular mycorrhizal fungal community.
 June 2003
                                          E-32
DRAFT-DO NOT QUOTE OR CITE

-------
  •  The effects of excessive deposition of nitrogen, particularly NH3 and NH4+, have led to
     changes in Dutch heathlands via:  (1) acidification of the soil and the loss of cations of K+,
     Ca2+ and Mg2+; and (2) nitrogen enrichment, which results in increased plant growth rates
     and altered competitive relationships. Alteration of any of a number of parameters
     (e.g., increased nitrogen) can alter ecosystem structure and function.

  *  Notable effects of excess nitrogen deposition have been observed with regard to aquatic
     systems. Atmospheric nitrogen deposition onto soils in watershed areas feeding into
     estuarine sound complexes (e.g., the Pamlico Sound of North Carolina) appear to contribute
     to excess nitrogen flows in runoff (especially during and after heavy rainfall events such as
     hurricanes) from agricultural practices or other uses (e.g., fertilization of lawns or gardens).
     Massive influxes of such nitrogen into watersheds and sounds can lead to algae blooms,
     which, upon degradation, deplete oxygen and cause extensive fish kills and damage to
     commercial fish and sea food harvesting.

  •  Acidic precipitation, a critical environmental stress  that affects forest landscapes and aquatic
    ecosystems in North America, Europe, and Asia, is  linked to the effects associated with the
    deposition of Nr and sulfates and the associated hydrogen ion. The resulting soil acidity can
    lead to nutrient deficiencies and to high aluminum-to-nutrient ratios that limit uptake of
    calcium and magnesium and create a nutrient deficiency adversely affect tree growth.
    Calcium is essential in the formation of wood and the maintenance of the primary plant
    tissues  necessary for tree growth.

 •  The critical loads concept is useful-for estimating the amounts of pollutants (e.g., reactive
    nitrogen and acidic precipitation) that sensitive ecosystems can absorb on a sustained basis
    without experiencing measurable degradation.  The estimation of ecosystem critical loads
    requires an understanding of how an ecosystem will respond to different loading rates in the
    long term and can be of special value for ecosystems receiving chronic deposition of Nr and
    sulfur independently and as acidic  deposition when in combination.
June 2003                                 E-33        DRAFT-DO NOT QUOTE OR CITE

-------
• The biological effect of PM is determined by the amount deposited during the wet and dry
  deposition onto the plant surfaces or soil in the vicinity of the roots. The three major routes
  involved are (1) precipitation scavenging in which particles are deposited in rain and snow;
  (2) occult (fog, cloud water, and mist interception); and (3) dry deposition, a much slower,
  yet more continuous removal to place surfaces.

• Few heavy metals (e.g., copper, nickel, zinc) have been documented to have direct
  phytoxicity under field conditions.  Ecosystems immediately downwind of major emissions
  sources such as power generating, industrial, or urban complexes can receive locally heavy
  inputs.  Heavy metal accumulation in the richly organic forest floor where biological activity
  is greatest by affecting litter decomposition present the greatest potential for influencing
  nutrient cycling.  The presence of cadmium, copper and nickel have been observed to affect
  the symbiotic activity of fungi, bacteria, and actinomycetes.

 • Phytochelatins produced by plants as a response to sublethal concentrations of heavy metals,
  are indicators of metal stress and can be used to indicate that heavy metals are involved in
  forest decline.  Increasing concentrations of phytochelatins with increasing altitude,  and their
   increases across regions that show increased levels of forest injury, have implicated  them in
   forest decline.

 •  Secondary organics formed in the atmosphere have been variously subsumed under  the
   following terms: toxic substances, pesticides, hazardous air pollutants (HAPS),  air toxics,
   semivolatile organic compounds (SOCs), and persistent organic pollutants (POPS).  Such
   substances are controlled under CAA SecLl 12, Hazardous Air Pollutants (not as criteria
   pollutants controlled by NAAQS under CAA Sections 108 and 109).  Their possible effects
   on humans and ecosystems are discussed in many other government documents and
   publications.  They are noted in this document because, in the atmosphere, many of the
   chemical compounds are partitioned between gas and particle phases and are deposited as
   particulate matter.  As particles, they become airborne and can be distributed over a wide
   area and affect remote ecosystems. Some are also of concern to humans because they may
   reach toxic levels in food chains of both animals and humans; whereas others may tend to
   maintain or decrease some toxicity as they move through the food chain.

 June 2003                                 E-34        DRAFT-DO NOT QUOTE OR CITE

-------
  1      E.6.2  Visibility
  2           Chapter 4 of this document includes information supplementary to several other significant
  3      reviews of the science of visibility, including the 1991 report of the National Acid Precipitation
  4      Assessment Program, the National Research Council's Protecting Visibility in National Parks
  5      and Wilderness Areas (1993), and EPA's 1995 Interim Findings on the Status of Visibility
  6      Research. The following points are made in Chapter 4 and/or in the above referenced
  7      documents.
  8      • The relationships between air quality and visibility are well understood.  Ambient fine
           particles are the major cause of visibility impairment. Significant scientific evidence exists
           showing that reducing fine particle concentrations will improve visibility.

  9      • The National Research Council defines visibility qualitatively as "the degree to which the
           atmosphere is transparent to visible light." This definition may be expressed quantitatively in
           terms of contrast transmittance.  The EPA has defined visibility impairment as a reduction in
           visual range (the farthest distance at which a large black object can be distinquished against
           the horizontal sky is the visual range) and/or atmospheric discoloration.

10      • Light, as it passes through the atmosphere from a scene to an observer, is both scattered and
           absorbed.  The rate of loss of transmitted light intensity with distance is measured by the
           light-extinction coefficient, which may be expressed as the sum of the coefficients for:
           (a) light scattering due to gases; (b) light scattering due to particles; (c) light absorption by
           gases, and; (d)  light absorption by particles. Light scattering by particles is the major
           component of light extinction. Light absorption by gases is almost entirely due to NO2, and is
           typically significant only near NO2 sources.  Light absorption by particles is primarily caused
           by elemental carbon.

11       •  Light scattering efficiency depends on particle size, falling off rapidly for particles below
           0.3 or above 1.0 urn in diameter.  Therefore, particles in the accumulation mode (of the fine
          particle mode) are most effective in scattering light and are more important in visibility
          degradation than either nuclei-mode or coarse-mode particles.  Light absorption is not a
          strong function of particle size. Under exceptional circumstances, such as dust storms, coarse
          particles can dominate scattering.

       June 2003                                  E-35        DRAFT-DO NOT QUOTE OR CITE

-------
- In addition to reducing the intensity of light carrying information about a scene (transmitted
  radiance), particles also scatter light into the observer's view. This extraneous light, called
  air light or path radiance, carries no information about the scene. The competition between
  these two sources of light, expressed as the ratio of transmitted radiance from the scene to
  path radiance, determines the contrast transmittance and the visual quality of the view.

• Visibility at any location is affected by air quality and non-air quality related effects. The
  visibility effects of atmospheric constituents are dependant upon not just the mass of
  pollutants, but on the size distribution and refractive index of particles, which are strongly
  influenced by relative humidity. Non-air quality effects include the angle between the sun
  and the observer's sight path, location of clouds, and reflectivity of the ground. These effects
  are independent of effects due to changes in atmospheric constituents.  Lighting and scene
  effects can be accounted for by defining a range of these effects when estimating visibility
  changes due to air quality influences.

 • The relationship between air pollution and the appearance of a scenic view is well
  understood. Models exist that, given an adequate description of the air quality and non-air
  quality variables, can produce  a simulated photograph that accurately depicts a cloud-free
  scene as it would appear to a human observer.

 • There are several potential quantitative indicators of visibility,  including: (a) fine particle
  mass and composition (fine particle mass alone provides less of both types of information);
  (b) scattering by dried ambient particles; (c) scattering by particles under ambient conditions;
  (d) extinction (calculated from measurements of scattering plus absorption); (e) light
   extinction measured directly; and (f) contrast transmittance.

 -  A new index, the deciview (dv), is now being used as a quantitative measure of haziness. It is
   related to the light extinction coefficient, b^ by Haziness (dv) = 10 ln(bJ10 Mn). The
   deciview is more nearly linearly related to perceived changes in haze level than either visual
   range or light extinction. A change of 1 or 2 dv in uniform haze under many viewing
   conditions will be seen as a small but noticeable change in the appearance of a scene
   regardless of the initial haze condition.

 June 2003                                 E-36        DRAFT-DO NOT QUOTE OR CITE

-------
 5
 6
 7
 8
 9
10
11
12
13
           Visibility in the United States is best in the western, intermountain region. Visibility
           impairment or haziness is greatest in the eastern United States and southern California.
           Haziness in the eastern United States is caused primarily by atmospheric sulfate. Haziness in
           southern California is primarily caused by nitrate and organic PM. Nitrates contribute about
           40% to the total light extinction in southern California. Nitrates account for 10 to 20% of the
           total extinction in other areas of the United States.

           Organics are the second largest contributors to light extinction in most areas in the United
           States. Organic carbon is the greatest cause of light extinction in the Pacific Northwest,
           Oregon, Idaho, and Montana, accounting for 40 to 45% of the total extinction.  Organic
           carbon contributes between 15 to 20% to the total extinction in most of the western United
           States and 20 to 30% in the remaining areas of the United States.

           Light absorption by carbon is relatively insignificant but is highest in the Pacific Northwest
           (up to 15%) and in the eastern United States (3%).

           High dust concentrations transported from southern California and the subtropics have
           contributed to regional haze in the Grand Canyon and other Class I areas in the southwestern
           United States.
E.6.3  Participate Matter-Related Effects on Materials
     Atmospheric PM and SO2 exert effects on materials that are related both to aesthetic appeal
and physical damage. Studies have demonstrated particles, primarily consisting of carbonaceous
compounds, cause soiling of commonly used building materials and culturally important items
such as statues and works of art. Physical damage from the dry deposition of SO2, particles, and
the absorption or adsorption of corrosive agents on deposited particles can also result in the
acceleration of the weathering of manmade building and naturally occurring cultural materials.
 •  The natural process of metal corrosion from exposure to environmental elements (wind,
   moisture, sun, temperature fluctuations, etc.) is enhanced by exposure to anthropogenic
   pollutants, in particular SO2.
       June 2003
                                          E-37
DRAFT-DO NOT QUOTE OR CITE

-------
1      • Dry deposition of SO2 enhances the effects of environmental elements on calcereous stones
         (limestone, marble, and carbonated cemented) by converting the calcium carbonate (calcite)
         in the stone to calcium sulphate dihydrate (gypsum). The rate of deterioration is determined
         by the SO2 concentration, the stone's permeability and moisture content, and the deposition
         rate.

2      • Sulfur dioxide limits the life expectancy of paints by causing discoloration, loss of gloss, and
         loss of thickness of the paint film layer.

3      • A significant detrimental effect of paniculate pollution is the soiling of painted surfaces and
         other building materials.  Soiling is a degradation process requiring remediation by cleaning
         or washing, and depending on the soiled surface, repainting. Soiling decreases the
         reflectance of a material and reduces the transmission of light through transparent materials.
          Soiling may reduce the life usefulness of the material soiled.
5      E.6.4  Effects of Atmospheric Participate Matter on Global Warming
6              Processes and Transmission of Solar Ultraviolet Radiation
7       • The physical processes (i.e., scattering and absorption) responsible for airborne particle
          effects on transmission of solar visible and ultraviolet radiation are the same as those
          responsible for visibility degradation. Scattering of solar radiation back to space and
          absorption of solar radiation determine the effects of an aerosol layer on solar radiation.

 8       • Atmospheric particles greatly complicate projections of future trends in global warming
          processes because of emissions of greenhouse gases; consequent increases in global mean
          temperature; resulting changes in regional and local weather patterns; and mainly deleterious
          (but some beneficial) location-specific human health and environmental effects.  The body of
          available evidence, ranging from satellite to in situ measurements of aerosol effects on
          radiation receipts and cloud properties, is strongly indicative of an important role in climate
          for aerosols. This role, however, is poorly quantified. No significant advances have been
          made in reducing the uncertainties assigned to forcing estimates provided by the IPCC for
          aerosol-related forcing, especially for black carbon-containing aerosol.  The IPCC

        June 2003                                E-38        DRAFT-DO NOT QUOTE OR CITE

-------
 10
11
12
  characterizes the scientific understanding of greenhouse gas-related forcing as "high" in
  contrast to that for aerosol, which it describes as "low" to "very low."

•  Quantification of the effect of anthropogenic aerosol on hydrological cycles requires more
  information than is presently available regarding ecosystems responses to reduced solar
  radiation and other changes occurring in the climate system. However, several global scale
  studies indicate that aerosol cooling alone can slow down the hydrological cycle, while
  cooling plus the nucleation of additional cloud droplets can dramatically reduce precipitation
  rates.

  In addition to direct climate effects through the scattering and absorption of solar radiation,
  particles also exert indirect effects on climate by serving as cloud condensation nuclei, thus
  affecting the abundance and vertical distribution of clouds. The direct and indirect effects of
  particles appear to have significantly offset global warming effects caused by the buildup of
  greenhouse gases on a globally-averaged basis.  However, because the lifetime of particles is
  much shorter than that required for complete mixing within the Northern Hemisphere, the
  climate effects of particles generally are  felt much less homogeneously than are the effects of
  long-lived greenhouse gases.

 Any effort to model the impacts of local  alterations in particle concentrations on projected
 global climate change or consequent local and regional weather patterns would be subject to
 considerable uncertainty.

 Atmospheric particles also complicate estimation of potential future impacts on human health
 and the environment projected as possible to occur because of increased transmission of solar
 ultraviolet radiation (UV-B) through the Earth's atmosphere, secondary to stratospheric
 ozone depletion due to anthropogenic emissions of chlorofluorcarbons (CFCs), halons, and
 certain other gases. The transmission of solar UV-B radiation is affected strongly by
 atmospheric particles. Measured attenuations of UV-B  under hazy conditions range up to
 37% of the incoming solar radiation.  Measurements relating variations in PM mass directly
 to UV-B transmission are lacking.
       June 2003
                                       E-39
                                                             DRAFT-DO NOT QUOTE OR CITE

-------
1      • Particles also can affect the rates of photochemical reactions occurring in the atmosphere,
         e.g., those involved in catalyzing tropospheric ozone formation. Depending on the amount of
         absorbing substances in the particles, photolysis rates either can be increased or decreased.
         Thus, atmospheric particle effects on UV-B radiation, which vary depending on size and
         composition of particles, can differ substantially over different geographic areas and from
         season to season over the same area. Any projection of effects of location-specific airborne
         PM alterations on increased atmospheric transmission of solar UV radiation (and associated
         potential human health or environmental effects) due to stratospheric ozone-depletion would,
         therefore, also be subject to considerable uncertainty.

2
3      E.7   KEY CONCLUSIONS
4       • Epidemiologic studies show consistent positive associations of exposure to ambient PM with
         health effects, including mortality and morbidity. The observed associations of ambient PM
         exposure with health effects must take into account the effects of other environmental or
         demographic  factors, depending on whether the effects are acute or chronic.  Effect-size
          estimates for PM-attributable excess relative risk for a given health outcome vary somewhat
          depending on specific analytic models used, but nevertheless have shown reasonable
          quantitative consistency across different studies.

 5        • Issues surrounding potential bias in PM risk estimates  from time-series studies using
          generalized additive models analyses and default convergence criteria were recently raised.
          Reanalyses of some important studies were included in an HEI Special Report (2003) that
          generally confirmed that PM effect estimates generally decline somewhat when using more
          stringent convergence criteria, as compared to the default GAM analyses, with the new
          estimates being well within confidence intervals of the original estimates. Overall, the
          absolute effect was relatively small, and the basic direction of effect and conclusions
          regarding the significance of the PM effect on mortality and hospital admissions remained
          unchanged in these analyses when the GAM convergence requirement was made more
          stringent. A  modeling issue of particular importance highlighted by HEI (2003) is the
        june 2003                                E-40        DRAFT-DO NOT QUOTE OR CITE

-------
            sensitivity of all models to the degrees of freedom allotted to potentially confounding weather
            variables and time. The HEI panel recommended further exploration of the sensitivity of
            these studies to a wider range of alternative degrees of smoothing and to alternative
            specifications of weather variables in time-series models.

  6       •  Individuals with cardiovascular or pulmonary disease, especially if they are elderly, are more
            likely to suffer severe health effects (mortality or hospitalization) related to PM exposure
            than are healthy young adults.  Children and asthmatics are also susceptible to certain PM
            effects, e.g., increased respiratory symptoms and decreased lung function. Smokers also
            constitute a population group at increased risk for ambient PM exposure effects.

  7       •  Daily ambient fine particle concentrations measured at a community air-monitoring site are
            useful surrogates for daily personal exposures to ambient fine particles. This is consistent,
            for example, with observed high correlations of personal sulfate exposures with ambient
            sulfate concentrations. The relationship between personal exposure to thoracic coarse
           particles and the ambient concentration of thoracic coarse fraction particles is not as strong,
           making effects due to coarse fraction particles harder to detect in epidemiologic studies.

  8       • Development of a comprehensive biologically based exposure-dose-response model to aid
           health risk assessment requires further dosimetry data characterizing differences among
           species in percent deposition and regional deposition patterns including differences in
           inhalability, airway geometry, and clearance rates. More information is also needed on
           mechanism(s) of clearance, pathological processes affecting deposition and clearance of
           particles, and factors influencing the response(s) of respiratory tract tissues to particle burden.

 9       • The percent deposition and regional patterns of deposition depend strongly on particle size.
           Percent deposition is higher in smaller lungs (children; women), during exercise, and in the
           functioning parts of the lungs in people with compromised lungs (e.g., those with chronic
           obstructive lung disease).

10       • Estimation of public health impacts of ambient airborne particle exposures in the United
           States would most credibly combine information on exposure-response relationships derived

       June 2003                                 £-41        DRAFT-DO NOT QUOTE OR CITE

-------
          for particular U.S. urban areas, in combination with estimates of exposures to ambient
          particle concentrations for the general population and/or specific susceptible subgroups (e.g.,
          the elderly) those particular areas. At this time, risk assessment is necessarily limited to use
          of available information from concentration-response relationships relating ambient
          concentrations to health effects in populations. In view of geographic differences in ambient
          PM mixtures and demographics, broad generalization of some single "best estimate" of
          relative risk for a given increment in concentration of a given particle indicator (e.g., PM10,
          PM25, etc.) would be subject to much uncertainty.

11       • Toxico logical studies of PM using controlled inhalation exposure of humans and laboratory
          animals, intratracheal instillation in humans and animals, and exposure of human and animal
          cells grown in culture find numerous biological effects which may be related to adverse
          health effects. Newer studies are finding different biological effects for a variety of different
          particle components.  Newer studies also are beginning to identify biological mechanisms
          whereby PM deposited in the lung can produce adverse effects on the cardiovascular and
          respiratory systems.

12       • Epidemiologic studies indicate increased health risks associated with exposure to PM, alone
          or in combination with other air  pollutants. PM-related increases in individual health risks
          are small, but likely significant from an overall public health perspective because of the large
          numbers of individuals in susceptible risk groups that are exposed to ambient PM.

13       • Numerous new studies, including multicity studies, continue to show a consistent association
          of PM10 exposure with mortality and various morbidity endpoints, thus substantiating the
          relationship of PM exposure with various health effects. However, some new studies using
          PM25 as an indicator find higher statistical significance and higher excess risk for PM25
          compared to PM10. Several studies have also observed statistically associations of PM10_25
          with health effects.

14       • Epidemiologic studies, in which factors identified with source  categories or individual
          chemical components of PM have been used as indicators, also show significant associations
        June 2003                                 E-42       DRAFT-DO NOT QUOTE OR CITE

-------
           with nitrates, sulfates, various indicators of elemental carbon, the organic component of PM,
           and some elements. PM-related source category factors, such as regional sulfate, vehicular
           traffic, fossil fuel combustion, vegetative combustion, and oil burning have also been found
           to be significantly associated with mortality.

15       • Data from multicity studies, comparisons of effects found in single and multiple regressions,
           exposure relationships, and toxicity suggest that the gaseous co-pollutants (CO, N03, SO2,O3)
           are not responsible for the relationships observed with PM indicators in community, time-
           series epidemiology. This does not, however, necessarily imply lack of an independent
           association of gaseous co-pollutants with health effects.

16       • Fine and thoracic coarse PM, indexed respectively by PM25 and PM,0_25, should be
           considered as separate subclasses of PM. Considerations of emissions sources, atmospheric
           chemistry, physical behavior, exposure relationships, respiratory deposition, toxicologic
           findings, and epidemiologic observations argue for monitoring fine and thoracic coarse
           particles separately.

17       •  Assessment of health risk in epidemiologic studies of ambient air pollutants, including PM,
           has relied largely on studies that focus on changes in health risks that occur in relation to
           normal changes in ambient air pollutant concentrations. Further evidence of the effects of air
           pollution on health may be deduced from intervention studies, i.e, studies of changes in
           health effects that occur when air pollution concentrations have been temporarily or
          permanently reduced through regulatory action, industrial shutdown, or other intervening
          factor(s).  Only a few epidemiologic intervention studies are available; however, taken
          together, these studies lend confidence that further reduction of ambient air pollution
          exposures in the U.S. would reduce both respiratory and cardiovascular health effects.
          Available studies also give reason to expect that further reductions in both particulate and
          gaseous air pollutants would benefit health.  Furthermore, experimental studies of Utah
          Valley filter extracts points to PM-associated metals as a likely cause or promoter of at least
          some of the health disorders associated with ambient PM.
      June 2003                                 £-43        DRAFT-DO NOT QUOTE OR CITE

-------
  The overall weight of evidence, based on information concerning PM exposure, dosimetry,
  toxicology, and epidemiology, supports the conclusion that PM, especially fine PM, is the
  primary contributor to a variety of adverse health effects associated with air pollution.
  However, difficult technical issues still remain in further separating the effects of fine and
  coarse particles and in delineating respective contributions of PM acting along or in
  combination with gaseous co-pollutants in increasing risks of health effects anticipated to
  occur in response to exposures to contemporary particle-containing ambient air mixes in the
  United States.
June 2003
E-44
DRAFT-DO NOT QUOTE OR CITE

-------
  1      REFERENCES
  2
  3      Brunekreef, B. (1997) Air pollution and life expectancy: is there a relation? Occup. Environ. Med. 54: 781-784.
  4      Health Effects Institute. (2003) Revised analyses of time-series studies of air pollution and health. Boston, MA:
  5            Health Effects Institute; special report. Available: http://www.healmeffects.org/Pubs/TimeSeries.pdf
  6            [27 June 2003].
  7      Intergovernmental Panel on Climate Change (IPCC). (200la) Climate change 2001: the scientific basis.
  8            Contribution of working group I to the third assessment report of the Intergovernmental Panel on Climate
  9            Change. Cambridge, United Kingdom: Cambridge University Press.
 10      Intergovernmental Panel on Climate Change (IPCC). (2001b) Climate change 2001: impacts, adaptation, and
 11            vulnerability. Contribution of working group II to the third assessment report of the Intergovernmental Panel
 12            on Climate Change. Cambridge, United Kingdom: Cambridge University Press.
 13      National Acid Precipitation Assessment Program. (1991) National Acid Precipitation Assessment Program 1990
 14            integrated assessment report. Washington, DC: National Acid Precipitation Assessment Program.
 15      National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
 16            National Academy Press. 3v.
 17      United Nations Environment Programme (UNEP). (1998) Environmental effects of ozone depletion:  1998
 18            assessment. J. Photochem. Photobiol. B 46: 1-4.
 19      United Nations Environment Programme (UNEP). (2000) Environmental effects of ozone depletion: interim
20            summary. Available at: http://www.gcrio.org/ozone/unep2000summary.htmi [9 April 2002].
21      U.S. Code. (1991) Clean Air Act,  §108, air quality criteria and control techniques, §109, national ambient air quality
22            standards. U. S. C. 42: §§7408-7409.
23      U.S. Environmental Protection Agency. (1995) Interim findings on the status of visibility research. Research
24            Triangle Park, NC: Office of Research and Development; report no. EPA/600/R-95/021.
25      U.S. Environmental Protection Agency. (1996) Air quality criteria for paniculate matter. Research Triangle Park,
26            NC: National Center for Environmental Assessment-RTF Office; report nos.  EPA/600/P-95/001aF-cF. 3v.
27      World Meteorological Organization (WMO). (1999) Scientific assessment of ozone depletion: 199S. Geneva,
28            Switzerland: World Meteorological Organization, Global Ozone and Monitoring Project; report no. 44.
        June 2003                                      E-45         DRAFT-DO NOT QUOTE OR CITE

-------

-------
   1
   2
   3
   4 .
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
30
31
                              1.  INTRODUCTION
       This document is an update of "Air Quality Criteria for Paniculate Matter" published by
  the U.S. Environmental Protection Agency (EPA) in 1996, and it will serve as the basis for
  Congressionally-mandated periodic review of the National Ambient Air Quality Standards
  (NAAQS) for paniculate matter (PM). The present document critically assesses the latest
  scientific information relative to determining the health and welfare effects associated with the
  presence of various concentrations of PM in ambient air. The. document builds upon the
  assessment in the previous 1996 EPA Air Quality Criteria Document for Particulate Matter (PM
  AQCD) by focusing on assessment and integration of information most relevant to PM NAAQS
  criteria development, based on pertinent peer-reviewed literature published or accepted for
  publication mainly through April 2002.  This introductory chapter presents a brief summary of
  legislative requirements and history of the PM NAAQS, provides an overview of issues
  addressed and procedures utilized in the preparation of the present document, and provides
  orientation to the general organizational structure of the document.
 1.1   LEGISLATIVE REQUIREMENTS
      As indicated in U.S. Code (1991), the U.S. Clean Air Act (CAA), Sections 108 and 109
 (42 U.S.C. Sections 7408 and 7409) govern the establishment, review, and revision of National
 Ambient Air Quality Standards (NAAQS). Section 108(a) directs the EPA Administrator to list
 pollutants, which, in the Administrator's judgement, cause or contribute to air pollution which
 may reasonably be anticipated to endanger either public health or welfare and to issue air quality
 criteria for them. The air quality criteria are to reflect the latest scientific information useful in
 indicating the kind and extent of all identifiable effects on public health and welfare that may be
 expected from the presence of the pollutant in ambient air.
      Section 109 directs the Administrator of EPA to propose and promulgate "primary" and
 "secondary" NAAQS for pollutants identified under Section 108. Section 109(b)(l) defines a
 primary standard as a level of air quality, the attainment and maintenance of which, in the
judgement of the Administrator, based on the criteria and allowing for an adequate margin of
       June 2003
                                                 1-1
                                                     DRAFT-DO NOT QUOTE OR CITE

-------
 1      safety, is requisite to protect the public health. Section 109(b)(2) defines a secondary standard as
 2      one which, in the judgement of the Administrator, based on the criteria, is requisite to protect
 3      public welfare from any known or anticipated adverse effects associated with the presence of
 4      such pollutants. Welfare effects, defined in Section 302(h), include, but are not limited to,
 5      effects on soils, water, crops, vegetation, man-made materials, animals, wildlife, weather,
 6      visibility and climate, damage to and deterioration of property, and hazards to transportation, as
 7      well as effects on economic values and personal comfort and well-being.  Section 109(d)(l)
 8      requires periodic review and, as appropriate, revision of existing criteria and standards.
 9      Section 109(d)(2) requires an independent committee of non-EPA experts, the Clean Air
10      Scientific Advisory Committee (CASAC), to provide advice and recommendations to the EPA
11      Administrator regarding the scientific soundness and appropriateness of criteria and NAAQS for
12     PM and other "criteria air pollutants" (i.e., ozone, nitrogen dioxide, sulfur oxides, carbon
13      monoxide, lead) regulated under CAA Sections  108-109.
14
15
16     1.2  HISTORY OF PREVIOUS PM CRITERIA AND NAAQS REVIEWS
17           "Particulate matter" is the generic term for a broad class of physically and chemically
18     diverse substances that exist in ambient air as discrete particles (liquid droplets or solids) over a
19     wide range of sizes. These airborne particles originate from a variety of stationary and mobile
20     sources. Primary particles are emitted directly into ambient air; whereas secondary particles are
21     formed in the atmosphere by transformation of gaseous emissions such as sulfur oxides (SOJ,
22     nitrogen oxides (NO,), and volatile organic compounds (VOCs). The physical and chemical
23     properties of PM vary greatly with time, region, meteorology,  and source category, thus
24      complicating assessment of ambient PM health and welfare effects. Particles in ambient air are
25      usually distributed bimodally in two somewhat overlapping size categories: (1) fine (diameter
26      generally less than 2.5 mm) and (2) coarse (diameter generally greater than 1.0 urn). Particles in
27      these two size fractions tend to differ in terms of formation mechanisms, sources of origin,
 28      composition, and behavior in the atmosphere and human respiratory tract.
 29           EPA first promulgated primary and secondary NAAQS for PM on April 30,1971 (Federal
 30      Register, 1971).  These standards measured PM as "total suspended paniculate" (TSP), which
 31      refers to ambient PM up to a nominal size of 25 to 45 micrometers (um). The primary standards

        June 2003                                  1-2         DRAFT-DO NOT QUOTE OR CITE

-------
  1     for PM (measured as TSP) were 260 fig/m3 (24-h average), not to be exceeded more than once
  2     per year, and 75 |ig/m3 (annual geometric mean). The secondary standard (measured as TSP)
  3     was 150 ug/m3 (24-h average), not to be exceeded more than once per year.
  4          EPA completed the next review of PM air quality criteria and standards in July 1987,
  5     revising the 1971 standards to protect against adverse health effects of inhalable airborne
  6     particles which can be deposited in the lower (thoracic) regions of the human respiratory tract,
  7     with "PM10" as the indicator (i.e., those particles collected by a sampler with a specified
  8     penetration curve yielding an upper 50% cut-point of 10-urn aerodynamic diameter (Federal
  9     Register, 1987). EPA established identical primary and secondary PM10 standards for two
 10     averaging times: 150 ng/m3 (24-h average), with no more than one expected exceedance per
 11     year and 50 ^g/m3 (expected annual arithmetic mean), averaged over three years.
 12
 13     1.2.1  The 1997 PM NAAQS Revision
 14          The last previous review of the air quality criteria and standards for PM was initiated in
 15     April 1994 by EPA announcing its intention to develop revised Air Quality Criteria for
 16     Particulate Matter. Several workshops were held by EPA's Environmental Criteria and
 17     Assessment Office in Research Triangle Park, NC (ECAO-RTP) in November 1994 and January
 18     1995 to discuss important new health effects information useful in preparing initial PM AQCD
 19     draft materials. Also, plans for review of the PM criteria and standards under a highly
 20     accelerated, court-ordered schedule were presented by EPA at a public meeting of the CASAC in
 21      December 1994. A court order entered in American Lung Association v. Browner, CIV-93-643-
 22      TUC-ACM (U.S. District Court of Arizona, 1995), as subsequently modified,  required
 23      publication of EPA's final decision on the review of the PM NAAQS by July  19,  1997.
 24           Several external review drafts of the revised PM AQCD were prepared by the RTP
 25      Division of EPA's newly created National Center for Environmental Assessment (i.e., by
 26      NCEA-RTP, the successor office to ECAO-RTP), and each were made available for public
 27      comment followed by CASAC review (at public meetings held in August 1995, December 1995,
 28      and February 1996).  The CASAC came to closure on its review of the PM AQCD in early 1996,
29      advising the EPA Administrator in a March 15,1996 closure letter (Wolff, 1996) that "although
30      our understanding of the health effects of PM is far from complete, a revised Criteria Document
31      which incorporates the Panel's latest comments will provide an adequate review of the available

        June 2003                                 1 -3         DRAFT-DO NOT  QUOTE OR CITE

-------
 1      scientific data and relevant studies of PM." Revisions made in response to public and CASAC
 2      comments were then incorporated by NCEA-RTP, as appropriate, into the final 1996 PM AQCD
 3      (U.S. Environmental Protection Agency, 1996a). A PM Staff Paper (SP), prepared by the Office
 4      of Air Quality Planning and Standards (OAQPS) within EPA's Office of Air and Radiation
 5      (OAR), drew upon the 1996 PM AQCD and other exposure and risk assessments to pose options
 6      for PM NAAQS decisions. Drafts of the PM SP also underwent public comment and CASAC
 7      review, with consequent revision into the final 1996 PM SP (U.S. Environmental Protection
 8      Agency, 1996b).
 9           The 1996 PM AQCD and the 1997 PM SP (U.S. Environmental Protection Agency,
10      1997a,b) provide detailed information on atmospheric formation, ambient concentrations, and
11      health effects of ambient air PM, as well as quantitative estimates of human health risks
12      associated with exposure to ambient PM. The principal focus of these documents was on recent
13      epidemiologic evidence reporting associations between ambient concentrations of PM and a
14     range of serious health effects. Special attention was given to several size-specific classes of
15      particles, including PM10 and the principal fractions of PM]0, referred to as the fine (PM2 5) and
16     coarse (PM10_2 5) fractions. PM2 5 refers to those particles collected by samplers having
17     penetration curves yielding an upper 50% cut point of 2.5 urn aerodynamic diameter. PM10_2 5
18     refers to those particles in an aggregate sample having an upper 50% cut point of 10 u.m and a
19     lower 50% cut point of 2.5 urn aerodynamic diameter. In other words, the coarse fraction
20     (PM10_2 5) refers to the inhalable particles that remain if fine (PM2.5) particles are removed from a
21      sample of PM10 particles. As discussed in the 1996 PM AQCD, fine  and coarse fraction particles
22     can be differentiated by their sources and formation processes and by their chemical and
23     physical properties, including behavior in the atmosphere.
24          Taking into account information and assessments presented in the 1996 PM AQCD and
25     PM SP, advice and recommendations of CASAC, and public comments received on proposed
26     revisions to the PM NAAQS published in December 1996 (Federal Register, 1996), the EPA
27     Administrator promulgated significant revisions to the PM NAAQS in July 1997 (Federal
28     Register,  1997).  In that decision, the PM NAAQS were revised in several respects. While it was
29     determined that the PM NAAQS should continue to focus on particles less than or equal to
30      10 nm in  diameter, it was also determined that the fine and coarse fractions of PM10 should be
31     considered separately. New standards were added, using PM2.5 as the indicator for fine particles,

       June 2003                                1-4         DRAFT-DO NOT QUOTE OR CITE

-------
   1      and PMi0 standards were retained for the purpose of regulating coarse-fraction particles. Two
   2      new PM2.5 standards were set:  an annual standard of 15 ^ig/m3, based on the 3-year average of
   3      annual arithmetic mean PM2 5 concentrations from single or multiple community-oriented
   4      monitors; and a 24-hour standard of 65 ng/m3, based on the 3-year average of the 98th percentile
   5      of 24-hour PM25 concentrations at each population-oriented monitor within an area. To continue
   6      to address coarse-fraction particles, the annual PM,0 standard was retained, and the form, but not
   7      the level, of the 24-hour PM10 standard was revised to be based on the 99th percentile of 24-hour
   8      PM10 concentrations at each monitor in an area. The secondary standards were revised by
   9      making them identical in all respects to the primary standards.
 10           Following promulgation of the revised PM NAAQS, legal challenges were filed by a large
 11      number of parties, addressing a broad range of issues.  In May 1998, the U.S. Court of Appeals
 12      for the District of Columbia Circuit issued an initial opinion that upheld EPA's decision to
 13      establish fine particle standards, holding that such standards were amply justified by the growing
 14      body of empirical evidence demonstrating a relationship between fine particle pollution and
 15      adverse health effects. American Trucking Associations v. Browner. 175 F. 3d 1027,  1055-56
 16      (D.C. Cir. 1999) (rehearing granted in part and denied in part, 195 F. 3d 4 (D.C. Cir. 1999b),
 17      affirmed in part and reversed in part, Whitman v. American Trucking Associations. 531 U.S. 457
 18      (2001). Further, the court found "ample support" for EPA's decision to regulate coarse fraction
 19      particles, although it vacated the revisions to the 1987 PMi0 standards on the basis of PM10 being
 20      a "poorly matched indicator for coarse particulate pollution" because PM!0 includes fine
 21     particles. Id. at 1053-55.  As a  result of this aspect of the court's ruling, which EPA did not
 22     appeal, the 1987 PM10 standards remain in effect.
 23          In addition, the court broadly held that EPA's approach to establishing the level of the
 24     standards in its 1997 decisions on both the PM and ozone NAAQS (which were promulgated on
 25     the same day and considered together by the court  in this aspect of its opinion) effected "an
 26     unconstitutional delegation of legislative authority." Id- at 1034-40. EPA appealed this aspect of
 27     the court's ruling to the United States Supreme Court. In February 2001, the U.S. Supreme
 28     Court unanimously reversed the Court of Appeals'  ruling on the constitutional issue, and sent the
29     case back to the Court of Appeals for resolution of any remaining issues that had not been
30     addressed in that court's earlier rulings. Whitman v. American Trucking Associations. 531 U.S.
31      457,475-76 (2001).  In March 2002, the Court of Appeals rejected all remaining challenges to

        June 2003                                 1-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1     the standards, finding that the 1997 PM25 standards were reasonably supported by the record and
 2     were not "arbitrary or capricious."  American Trucking Associations v. EPA. 283 F. 3d 355,
 3     369-72 (D.C. Cir. 2002). Thus, the 1997 PM25 standards are in effect.
 4
 5     1.2.2   Coordinated Particulate Matter Research Program
 6          Shortly after promulgation of the 1997 PM NAAQS decisions, NCEA-RTP published a
 7     PM Health Risk Research Needs Document (U.S. Environmental Protection Agency, 1998a) that
 8     identified research needed to improve scientific information supporting future reviews of the PM
 9     NAAQS.  The document provided a foundation for PM research coordination among Federal
10     agencies and other research organizations and also provided input to later National Research
11     Council (NRC) deliberations on PM research. The Office of Research and Development (ORD)
12     of EPA also moved quickly to broaden its ongoing PM research activities by developing, in
13     partnership with other Federal agencies, a coordinated interagency PM research program. This
14     interagency program has and continues to focus mainly on expanding scientific knowledge of
15     ambient PM exposure and health effects, as well as including development of improved
16     monitoring methods and cost-effective mitigation strategies. The interagency effort also
17     promotes substantially expanded coordination with other research organizations, including the
18     Health Effects Institute (HEI) and other state-, university-, and industry-sponsored research
19     groups. Beginning in the fall of 1997, public participation was and continues to be encouraged
20     through workshops and review  of EPA's PM Research Program documentation.
21          In response to Congressional requirements in EPA's Fiscal Year 1998 Appropriation, the
22     NRC established its Committee on Research Priorities for Airborne Particulate Matter in January
23      1998.  This NRC PM Research Committee's charge is to identify the most important research
24     priorities relevant to setting PM NAAQS, to develop a conceptual plan for PM research, and to
25     monitor EPA's research progress toward improved understanding of the relationship between
26     PM and public health. The Committee issued its first report in early 1998 (National Research
27     Council, 1998), a second one in 1999 (National Research Council, 1999), and a third one in 2001
28      (National Research Council, 2001). In the above-noted series of reports, the NRC PM Research
29      Committee recommended that expanded PM research efforts be planned and carried out in
30      relation to a general conceptual framework as shown in Figure 1-1. That framework essentially
31      calls for research aimed at: (a) identifying sources of airborne particles or gaseous precursor

        June 2003                                1-6        DRAFT-DO NOT QUOTE OR CITE

-------
          Sources of Airborne
          Parti culate Matter or
          Gaseous Precursor
             Emissions
                          Indicator in Ambient
                            (Outdoor) Air
                             (e.g. Mass
                            Concentration)
                 Mechanisms determining emissions,
                 chemical transformation (including
                 formation of secondary particles from
                     gaseous precursors), and
                        transport in air
                                     Human time-activity
                                     patterns, indoor (or
                                     micro environmental)
                                     sources and sinks of
                                      parti culate matter
Personal
Exposure

k.
A
W
L
Dose to
Target
Tissues

   Deposition,
clearance, retention
 and disposition of
 parti culate matter
  presented to
  an individual
 Mechanisms of
damage and repair
        Figure 1-1.  A general framework for integrating particulate-matter research. Note that
                     this figure is not intended to represent a framework for research management.
                     Such a framework would include multiple pathways for the flow of
                     information.
        Source: National Research Council (2001).
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
15
16
17
18
 emissions and characterization of processes involved in atmospheric transformation, transport,
 and fate of ambient PM; (b) delineation of temporal and spatial patterns of air quality indicators
 (e.g., PM25, PMio.2.5, PM10 mass concentrations) of ambient PM and apportionment of observed
 variations in such ambient PM indicators to various emission sources; (c) characterization of
 human exposures to ambient PM as one important component of total personal exposure to
 particles, as modified by time-activity patterns and varying microenvironmental exposure to
 particles of indoor or ambient origin; (d) characterization of resulting respiratory tract
 deposition, clearance, retention, and disposition of inhaled particles, as determinants of dose to
 target tissues (e.g., locally in the lungs or via systemic translocation to the heart or other organs);
 (e) delineation of mechanisms of damage and repair plausibly leading to (f) human health
 responses, as extrapolated from or quantified by experimental animal or human exposure
 (toxicology) studies and/or observational (epidemiology) studies.
      Research conducted under a PM Research Program structured in relation to the conceptual
 framework shown in Figure 1-1 would be expected (a) to reduce key scientific uncertainties
 regarding interrelationships between PM sources, ambient concentrations, exposures, dose to
target tissues, and resulting health effects and (b) thereby improve the scientific underpinnings
for both current and future periodic PM NAAQS reviews. Table 1 -1 highlights some types of
key uncertainties identified by the NRC PM Research Committee in relation to elements of the
        June 2003
                                            1-7
DRAFT-DO NOT QUOTE OR CITE

-------
        TABLE 1-1. KEY SCIENTIFIC UNCERTAINTIES RELATED TO THE
                       SOURCE-TO-RESPONSE FRAMEWORK

Source                >           Concentration (or other indicator)

    • Contribution of various emission sources to ambient and indoor particulate matter concentrations

    • Relative contribution of various sources to the most toxic components of particulate matter	

Concentration (indicator)                  >        Exposure

    • Relationship between ambient (indoor) particulate matter and the composition of particles to
     which people are exposed

    • Contribution of ambient particulate matter to total personal exposure for:

      - Susceptible subpopulations
      - General population

    • Variation in relationship of ambient particulate matter concentrations to human exposure by place

    • Variation in contribution of ambient particulate matter concentrations to total human exposure
     over time

    • Covariance of particulate matter exposures with exposures to other pollutants

    • Relationships between outdoor ambient and personal exposures for particulate matter and co-
     pollutants

Exposure             >        Dose

    • Relationship between inhaled concentration and dose of particulate matter and constituents at the
     tissue level in susceptible subjects
      -Asthma
      - Chronic obstructive pulmonary disease (COPD)
      - Heart disease
      - Age: infants  and elderly
      -Others	

Dose                  >        Response	

     • Mechanisms linking morbidity and mortality to particulate matter dose to or via the lungs

      - Inflammation
      - Host defenses
      - Neural mechanisms	,^-—^=^^==^=^=1

 Source: National Research Council (2001).
June 2003                                   1 -8         DRAFT-DO NOT QUOTE OR CITE

-------
  1     source-to-response conceptual framework illustrated in Figure 1-1. The NRC Committee went
  2     on to delineate a series of 10 research topics that they recommended be addressed in an
  3     expanded PM research program aimed at answering a set of broadly stated questions, as shown
  4     in Table 1-2.
  5          The EPA PM Research Program, structured to address topics shown in Table 1-2 includes,
  6     for example, studies to improve understanding of the formation and composition of fine PM,
  7     improved measurements and estimation of population exposures to ambient PM, the
  8     characteristics or components of PM that are responsible for its health effects, factors increasing
  9     susceptibility to PM effects in some subpopulations, and mechanisms by which these effects are
 10     produced. The results from these efforts, and related efforts by other Federal agencies and the
 11     general scientific community during the past several years, have substantially enhanced the
 12     scientific and technical bases for future decisions on the PM NAAQS.
 13
 14
 15     1.3   CURRENT PM CRITERIA AND NAAQS REVIEW
 16     1.3.1  Key Milestones
 17          As with other NAAQS reviews, rigorous assessment of relevant scientific information is
 18     presented in this updated, revised PM AQCD.  As shown in Table  1-3, development of the
 19     document has involved substantial external peer review through (a) public workshops involving
 20     the general aerosol scientific community, (b) iterative reviews of successive drafts by CASAC,
 21      and (c) comments from the public. The final document will reflect input received through these
 22     reviews and will serve to evaluate and integrate the latest available scientific information to
 23      ensure that the review of the PM standards is based on rigorous evaluation of the available
 24      science.
 25           An earlier (October 1999) First External Review Draft of this updated document was
 26      released in the fall of 1999 for public comment and CASAC review. A Second External  Review
 27      Draft (March 2001) took into account earlier public comments and  the December 1999 CASAC
 28      review and was reviewed by CASAC in July 2001. A Third External Review Draft similarly
29      took into account prior public comments and CASAC recommendations and was released in
30      early May 2002 for a 60-day public comment period; CASAC reviewed it at a public meeting in
31      July 2002.

        June 2003                                 1-9        DRAFT-DO NOT QUOTE OR CITE

-------
     TABLE 1-2. RESEARCH TOPICS AND QUESTIONS RECOMMENDED BY
   NATIONAL RESEARCH COUNCIL (NRC) TO BE ADDRESSED BY EXPANDED
            	PM RESEARCH PROGRAM

RESEARCH TOPIC 1.    OUTDOOR MEASURES VERSUS ACTUAL HUMAN EXPOSURES

    • What are the quantitative relationships between concentrations ofparticulate matter and gaseous co-
     pollutants measured at stationary outdoor air monitoring sites and the contributions of these concentrations
     to actual personal exposures, especially for subpopulations and individuals?	

RESEARCH TOPIC 2.    EXPOSURES OF SUSCEPTIBLE SUBPOPULATIONS TO TOXIC
                        PARTICULATE MATTER COMPONENTS

    • What are the exposures to biologically important constituents and specific characteristics ofparticulate
     matter that cause responses in potentially susceptible subpopulations and the general population?	

RESEARCH TOPIC 3.    CHARACTERIZATION OF EMISSION SOURCES

    • What are the size distribution, chemical composition, and mass-emission rates ofparticulate matter emitted
     from the collection of primary-particle sources in the United States, and what are the emissions of reactive
     gases that lead to secondary particle formation through atmospheric chemical reactions?	

RESEARCH TOPIC 4.    AIR-QUALITY MODEL DEVELOPMENT AND TESTING

    • What are the linkages between emission sources and ambient concentrations of the biologically important
     components ofparticulate matter?	,	

RESEARCH TOPIC 5.    ASSESSMENT OF HAZARDOUS PARTICULATE MATTER
                        COMPONENTS

    - What is the role ofphysicochemical characteristics ofparticulate matter in eliciting adverse health effects?

 RESEARCH TOPIC 6.    DOSIMETRY: DEPOSITION AND FATE OF PARTICLES IN THE
                        RESPIRATORY TRACT

    • What are the deposition patterns and fate of particles in the respiratory tract of individuals belonging to
     presumed susceptible subpopulations?	

 RESEARCH TOPIC 7.    COMBINED EFFECTS OF PARTICULATE MATTER AND GASEOUS
                        POLLUTANTS

    • How can the effects ofparticulate matter be disentangledfrom the effects of other pollutants? How can the
     effects of long-term exposure to particulate matter and other pollutants be better understood?	

 RESEARCH TOPIC 8.   SUSCEPTIBLE SUBPOPULATIONS

    • What subpopulations are at increased risk of adverse health outcomes from particulate matter?	

 RESEARCH TOPIC 9.   MECHANISMS OF INJURY

    • What are the underlying mechanisms (local pulmonary and systemic) that can explain the epidemiologic
     findings of mortality/morbidity associated with exposure to ambient particulate matter?

 RESEARCH TOPIC 10.  ANALYSIS AND MEASUREMENT	

    • To what extent does the choice of statistical methods in the analysis of data from epidemiologic studies
     influence estimates of health risks from exposures to particulate matter? Can existing methods be
     improved? What is the effect of measurement error and misclassification on estimates of the association
	between air pollution and health?

 Source: National Research Council (2001).


June 2003                                  1-10         DRAFT-DO NOT QUOTE OR CITE

-------
     TABLE 1-3.  SCHEDULE FOR DEVELOPMENT OF THE CURRENT REVISED
    PARTICIPATE MATTER AIR QUALITY CRITERIA DOCUMENT (PM AQCD)
  Major Milestones
            Target Dates
  PM NAAQS Review Plan to CASAC
  Prepare AQCD Development Plan
  Begin Literature Search
  Federal Register Call for Information/Sources Sought
  CASAC Meeting on AQCD Development Plan
  Prepare Workshop Drafts of Chapters
  Peer Review Workshop
  Prepare External Review Draft AQCD
  First External Review Draft
  Public Comment Period on Draft AQCD
  CASAC Meeting on First Draft AQCD
  Second External Review Draft
  Public Comment Period on Second Draft
  CASAC Meeting on Second Draft
  Third External Review Draft
  Public Comment Period on Third Draft
  CASAC Meeting on Third Draft
  Final PM AQCD
            October 1997
            November 1997 to January 1998
            February 1998
            April 1998
            May 1998
            May to December 1998
            April 1999
            March to September 1999.
            October 1999
            October 1999 to January 2000
            December 1999
            March 2001
            April to July 2001
            July 2001
            April 2002
            May to July 2002
            July 2002
            Oct/Nov 2002
     Shortly after EPA released the Third External Review Draft in May 2002, the Health
Effects Institute (HEI) announced that researchers at Johns Hopkins University had discovered
problems with applications of statistical software used in a number of important studies on links
between ambient air particulate matter (PM) and death and disease.  In response to the surfacing
of such statistical issues, which affected numerous PM time-series studies that used General
Additive Models (GAM) and were published post-1995, EPA took steps in consultation with
CASAC to encourage researchers to reanalyze affected studies and to submit them expeditiously
June 2003
1-11
DRAFT-DO NOT QUOTE OR CITE

-------
 1      for peer review by a special expert panel convened by HEI. The results of reanalyses for more
 2      than 30 studies of likely importance in informing U.S. standard setting decisions for ambient PM
 3      have now been peer-reviewed and are included, along with commentary by the HEI expert peer-
 4      review panel, in a recently released HEI Special Report (Health Effects Institute, May 2003).
 5      The subject statistical issues and reanalyses results are discussed in more detail in Chapter 8 of
 6      this document. After a 60-day public comment period, starting in June 2003, and CASAC
 7      review of this Fourth Draft, the final version of this PM AQCD is targeted for issuance by
 8      December 31,2003.
 9           After CASAC review of the First External Review Draft of this revised PM AQCD in
10      December 1999, EPA's OAQPS started to prepare the associated PM StaffPaper (SP).
11      A preliminary draft SP was made available to the public and was discussed at a July 2001
12      CASAC meeting; and additional consultations  have been held via teleconference with CASAC
13      to obtain feedback on proposed approaches to be used in conducting analyses to be included in
14     the First External Review Draft of the PM SP due to be released in the summer of 2003. The
15      draft PM SP will draw on the updated findings and conclusions from this draft PM AQCD and
16     will also undergo public comment and CASAC review (now being scheduled for November,
17     2003). Ultimately drawing on information in the final version of this newly revised PM AQCD,
18     the PM SP will evaluate policy implications of the key studies and scientific findings contained
19     in the AQCD, present related staff analyses of  air quality and  human health risk, and identify
20     critical elements that EPA staff believes should be considered in reviewing the PM standards.
21      The PM SP is intended to bridge the gap between the scientific review in the AQCD and the
22     public health and welfare policy judgements required of the Administrator in reviewing the PM
23     NAAQS.  In doing so, the PM SP will include  staff conclusions and recommendations of options
24     for the Administrator's consideration.
25          Based on the final versions of the PM AQCD and the PM SP, and on the advice of
26     CASAC, the Administrator will propose decisions as to whether to retain or revise the current
27     PM NAAQS and provide opportunities for public and CASAC comments on the proposed
28     decisions. Taking into account comments on the proposed decisions, the Administrator will then
29     make final decisions on the PM NAAQS, which are now expected to be published around the
30     end of 2005.
31

       June 2003                                1-12        DRAFT-DO NOT QUOTE OR CITE

-------
  1      1.3.2   Methods and Procedures for Document Preparation
  2          The procedures for developing this revised PM AQCD build on the experience and
  3      methods derived from the most recent previous PM, Ozone, and CO AQCD preparation efforts.
  4      Briefly, the respective responsibilities for production of the present PM AQCD are as follows.
  5      An NCEA-RTP PM Team was formed and is responsible for developing and implementing the
  6      project plan for preparation of the PM AQCD, taking into account inputs from individuals in
  7      other EPA program and policy offices identified as part of the EPA PM Work Group. The
  8      resulting project plan (i.e., the PM Document Development Plan) was discussed with CASAC in
  9      May 1998 and was appropriately revised. An ongoing literature search has continued to be
 10      conducted to identify, to the extent possible, all pertinent PM literature published since early
 11      1996. Additionally, EPA published in the Federal Register (1) a request for information asking
 12      for recently available research information on PM that may not yet be published and
 13      (2) a request for individuals with the appropriate type and level of expertise to contribute to the
 14     writing of PM AQCD materials to identify themselves (U.S. Environmental Protection Agency,
 15      1998b). The specific authors of chapters or sections of the proposed document were selected on
 16     the basis of their expertise on the subject areas and their familiarity with the relevant literature;
 17     these include both EPA and non-EPA scientific experts.  The project team defined critical issues
 18     and topics to be addressed by the authors and provided direction in order to emphasize
 19     evaluation of those studies most clearly identified as important for standard setting.  It should be
 20     noted that materials contributed by non-EPA authors are incorporated and, at times, modified by
 21      EPA PM team staff to reflect internal and/or external  review comments, e.g., by the public or
 22     CASAC, and that EPA is responsible for the ultimate content of the PM AQCD.
 23           The main focus of this revised  criteria document is the evaluation and interpretation of
 24      pertinent atmospheric science information, air quality data, human exposure information, and
 25      health and welfare effects information newly published since that assessed in the 1996 PM
 26      AQCD and likely to be useful in deriving criteria for U.S. EPA PM NAAQS decisions. Initial
 27      draft versions of AQCD chapters were evaluated via expert peer-review workshop discussions
 28      and/or written peer reviews that focused on the selection of pertinent studies included in the
29      chapters, the potential need for additional information to be added to the chapters, and the
30      quality of the summarization and interpretation of the literature. The authors of the draft
31      chapters then revised them on the basis of workshop and/or written expert review

        June 2003                                1-13        DRAFT-DO NOT QUOTE OR CITE

-------
 I      recommendations. These and other integrative summary materials were incorporated into the
 2      First External Review Draft of the PM AQCD (October 1999), which underwent public
 3      comment and was the subject of consultation with CASAC at a December 1999 public meeting.
 4           In order to foster timely presentation and publication of newly emerging PM research
 5      findings, EPA co-sponsored an Air and Waste Management Association International Speciality
 6      Conference, entitled "PM 2000: Particulate Matter and Health," held in January 2000 in
 7      Charleston, SC. The conference was co-sponsored in cooperation with several other government
 8      agencies and/or private organizations that also fund PM research. Topics covered included new
 9      research results concerning the latest advances in PM atmospheric sciences (e.g., PM formation,
10      transport, transformation), PM exposure, PM dosimetry and extrapolation modeling, PM
11      toxicology (e.g., mechanisms, laboratory animal models, human clinical responses), and PM
12      epidemiology. The main purpose of the conference was to facilitate having the latest scientific
13      information available in time for incorporation as quickly as possible into the Second External
14     Review Draft of this revised PM AQCD.  Hence, arrangements were made for scientists to
15      submit written manuscripts on papers or posters presented at the PM 2000 Conference for
16     expedited peer-review by several major journals, so that decisions on acceptance for publication
17     could be made by mid-2000.  The evaluations and findings set forth in the Second External
18     Review Draft (March 2001) of the revised PM AQCD included consideration of such published
19     PM 2000 papers and extensive additional information published elsewhere since the previous
20     First External Review Draft; and the Second Draft was reviewed by CASAC in July 2001.
21           Further revisions were then incorporated into the Third External Review Draft (April 2002)
22     to reflect both public comment and CASAC review of the Second Draft, as well as assessment of
23     additional extensive new information published since that addressed in the Second Draft. This
24     Fourth Draft AQCD incorporates changes made in response to earlier public comments and
25     CASAC reviews; and it includes pertinent peer-reviewed literature published or accepted for
26     publication mainly through April 2002. The final PM AQCD will include revisions made in
27     response to public comment and CASAC review of this draft document.
28
29
30
        June 2003                                1-14       DRAFT-DO NOT QUOTE OR CITE

-------
  1      1.3.3   Approach
  2          An initial step in development of this revised PM AQCD was to focus on selection of
  3      pertinent issues to include in the document as the basis for the derivation of PM NAAQS criteria.
  4      Preliminary issues were identified by the NCEA PM Team via inputs from other EPA program
  5      and policy offices, as derived from the 1996 PM AQCD and SP, their CASAC and public
  6      reviews, from the 1997 standard promulgation process, and from the 1998 PM Research Needs
  7      Document (alluded to in Section 1.2.2).  Still further identification and refinement of issues
  8      resulted from NRC review and reports on PM research priorities (also discussed in Section 1.2.2
  9      above). The CASAC review of the PM AQCD Development Plan and public comments on
 10      earlier draft AQCD materials  at various stages of their development also provided important
 11      inputs regarding issue identification.
 12          In developing draft materials for inclusion in the revised PM AQCD, detailed review of
 13      key new research was undertaken as a first step. However, instead of presenting a
 14      comprehensive review of all the literature, emphasis in this revised AQCD is placed on (1) first,
 15     the concise summary of key findings derived from previous PM criteria reviews (especially the
 16      1996 PM AQCD) and, then, (2) evaluation of the most pertinent new key information, with
 17     greater emphasis on more interpretive assessment — an approach reflecting CASAC
 18     recommendations. To aid  in the development of a more concise document than the 1996 PM
 19     AQCD, compilation of summary tables of relevant new literature published since completion of
 20     that previous document and selective text discussion of such new literature has been undertaken,
 21      with increased emphasis being placed in text discussions on interpretive evaluation and
 22     integration of key points derived from the newly summarized research results.
 23
 24
 25      1.4  DOCUMENT ORGANIZATION AND CONTENT
 26           The present draft document attempts to critically review and assess relevant scientific
 27      literature on PM published, since early 1996, including materials accepted for publication mainly
 28      through April 2002 (and thus appearing mostly during 2002).  Limited coverage of some more
29      recent studies is also included as deemed appropriate in light of its special importance. For
30      example, information derived from the recently released HEI Special Report (Health Effects
31      Institute, May 2003), discussed above in Section 1.3.1, is being integrated into this assessment.

        June 2003                                1-15        DRAFT-DO NOT QUOTE OR CITE

-------
 1     Literature discussed in detail in the 1996 EPA PM AQCD (U.S. Environmental Protection
 2     Agency, 1996a) generally is not discussed in depth in this document; rather key findings from
 3     that 1996 review are summarized as appropriate. Limited treatment is included, however, for
 4     some earlier studies judged to be of particular importance in the review of the PM NAAQS.
 5     Key literature is mainly summarized in tables; and the text mainly attempts to integrate and
 6     discuss overall interpretive points. An attempt is made to present alternative points of view
 7     where  scientific controversy exists.
 8          Emphasis is placed in the document on studies conducted at or near PM pollutant
 9     concentrations found in ambient air.  However, although emphasis has been placed on studies in
10     the range of current ambient levels, studies examining effects of higher concentrations have been
11     included if they contain unique data or documentation of a previously unreported effect or
12     mechanism.
13          The present document, basically organized to assess information related to topics along the
14     same flow of issues presented in the NRC  conceptual framework shown in Figure  1-1, includes
15     an Executive Summary and nine chapters presented in two volumes. Volume I contains the
16     Executive Summary, this general introduction (Chapter 1), and Chapters 2 through 5. Chapters 2
17     and 3 provide background information on  physical and chemical properties of PM and related
18     compounds; sources and emissions; atmospheric transport, transformation, and fate of PM;
19     methods for the collection and measurement of PM; and U.S. ambient air PM concentrations.
20     Chapter 4 assesses PM environmental effects on vegetation and ecosystems, visibility, man-
21     made materials, and climate-related effects (including effects on atmospheric transmission of
22     solar radiation), and includes limited information on economic impacts of some welfare effects.
23     Chapter 5 discusses factors affecting exposure of the general population to ambient PM.
24          The second volume contains Chapters 6 through 9.  Chapter 6 evaluates information
25     concerning dosimetry of inhaled particles  in the respiratory tract. Chapter 7 assesses the
26     toxicology of specific types of PM constituents and potential mechanisms of action, based on
27     both laboratory animal studies and controlled human exposure studies. Chapter 8  discusses
28     observational, i.e., epidemiologic, studies. Lastly, Chapter 9 integrates key information on
29     exposure, dosimetry, and critical  health risk issues derived from studies reviewed  in the prior
30     chapters, as well as highlighting key points regarding important welfare effects associated with
31     ambient PM.

       June 2003                                1-16        DRAFT-DO NOT QUOTE OR CITE

-------
   1       REFERENCES

   2       Federal Register. (1971) National primary and secondary ambient air quality standards. F. R. (April 30)
   3             36:8186-8201.
   4       Federal Register. (1987) Revisions to the national ambient air quality standards for particulate matter. F. R. (July 1)
   5             52:24,634-24,669.
   6       Federal Register. (1996) National ambient air quality standards for particulate matter, proposed rule. F. R.
   7             (December 13) 61:65,638-65,713.
   8       Federal Register. (1997) National ambient air quality standards for particulate matter, final rule. F. R. (July 18)
   9             62:38,652-38,752.
 10       Health Effects Institute. (2003) Revised analyses of timerseries studies of air pollution and health. Boston,  MA:
 11             Health Effects Institute; special report.
 12       Lioy, P. J. (1990) Assessing total human exposure to contaminants: a multidisciplinary approach. Environ.  Sci.
 13             Technol. 24: 938-945.
 14       National Research Council. (1983) Risk assessment in the federal government: managing the process. Washington,
 15             DC: National Academy Press. Available: http:/ftooks.nap.edu/books/PODI15/html/index.htnil (3 June 2003).
 16       National Research Council. (1994) Science and judgment in risk assessment. Washington, DC: National Academy
 17             Press. Available: http://www.nap.edu/books/030904894X/htmI/ (4 June 2003).
 18       National Research Council. (1998) Research priorities for airborne particulate matter. I. Immediate priorities and a
 19             long-range research portfolio. Washington, DC: National Academy Press. Available:
 20             http://www.nap.edU/cataIog/6131 .html (4 June 2003).
 21       National Research Council. (1999) Research priorities for airborne particulate matter. H. Evaluating research
 22             progress and updating the portfolio. Washington, DC: National Academy Press. Available:
 23             http://www.nap.edu/books/0309066387/html/ (4 June 2003).
 24       National Research Council. (2001) Research priorities for airborne particulate matter. III. Early research progress.
 25             Washington, DC: National Academy Press. Available: http://www.nap.edu/books/0309073375/html/ (4 June
 26             2003).
 27       Sexton, K.; Selevan, S. G.; Wagener, D. K.; Lybarger, J. A. (1992) Estimating human exposures to environmental
 28             pollutants: availability and utility of existing databases. Arch. Environ. Health 47: 398-407.
 29       U.S. Code. (1991) Clean Air Act, §108, air quality criteria and control techniques, §109, national ambient air quality
 30             standards. U. S. C. 42: §§7408-7409.
 31       U.S. Court of Appeals for the District of Columbia.  (I999a) American Trucking Associations, Inc. vs. U.S.
 32             Environmental Protection Agency. 175 F3d 1027 (D.C. Cir. 1999).
 33       U.S. Court of Appeals for the District of Columbia.  (1999b) American Trucking Associations, Inc. vs. U.S.
 34             Environmental Protection Agency. 195 F.3d 4 (D.C. Cir. 1999), affirmed in part, reversed in part,  and
 35             remanded..
 36       U.S. Court of Appeals for the District of Columbia.  (2002) American Trucking Associations, Inc. vs. U.S.
 37             Environmental Protection Agency. 283 F.3d 355, 378-79 (D.C. Cir. 2002).
 38       U.S. District Court of Arizona. (1995) American Lung Association v. Browner. West's Federal Supplement 884
 39             F.Supp.  345 (No. CIV 93-643 TUC ACM).
 40       U.S. Environmental Protection Agency. (1996a) Air quality criteria for particulate matter. Research Triangle Park,
 41            NC: National Center for Environmental Assessment-RTP Office; report nos. EPA/600/P-95/001aF-cF. 3v.
 42       U.S. Environmental Protection Agency. (1996b) Review of the national ambient air quality standards for particulate
 43            matter: policy assessment of scientific and technical information. OAQPS staff paper. Research Triangle
 44            Park, NC: Office of Air Quality Planning and Standards; report no. EPA/452/R-96-013. Available from:
 45            NTIS, Springfield, VA; PB97-115406REB.
 46      U.S. Environmental Protection Agency. (1998a) Particulate matter research needs for human health risk assessment
 47            to support future reviews of the national ambient air quality standards for particulate matter. Research
 48            Triangle Park, NC: National Center for Environmental Assessment; report no. EPA/600/R-97/132F.
 49      U.S. Environmental Protection Agency. (1998b) Review of national ambient air quality standards for particulate
 50            matter. Commer. Bus. Daily: February 19. Available: http://cbdnetaccess.gpo.gov/index.html [1999, November 24].
 51      U.S. Supreme Court. (2001) Whitman v. American Trucking Association. 531 U.S. 457 (nos. 99-1257 and 99-1426).
 52      SWolff, G. T. (1996) Closure by the Clean Air Scientific Advisory Committee (CASAC) on the staff paper for
53            particulate matter [letter to Carol M. Browner, Administrator, U.S. EPA]. Washington, DC: U.S.
54            Environmental Protection Agency, Clean Air Scientific Advisory Committee; EPA-SAB-CASAC-LTR-96-
55            008; June 13.
         June 2003                                      1-17         DRAFT-DO NOT QUOTE OR CITE

-------

-------
  i          2.  PHYSICS, CHEMISTRY, AND MEASUREMENT
  2                          OF PARTICULATE MATTER
  3
  4
  5           Chapter 3 of the 1996 EPA document Air Quality Criteria for Particulate Matter (1996 PM
  6      AQCD; U.S. Environmental Protection Agency, 1996a) contained an extensive review of the
  7      physics and chemistry of airborne particulate matter (PM).  Chapter 2 of this revised version of
  8      the PM AQCD also provides background information on the physics and chemistry of
  9      atmospheric particles, information useful in aiding the understanding of subsequent chapters.
 10      Those subsequent chapters are basically organized to follow the sequence of key elements
 11      comprising the risk assessment framework described in Chapter 1 (Section 1.2.2), beginning
 12      with sources and continuing to effects as shown in Figure 1-1. Thus, this chapter provides
 13      new information useful in evaluating PM effects of PM on human health and welfare and in
 14      preparing related risk assessments used to support PM standard-setting decisions. Information
 15      important for implementation of PM standards, but not essential to the standard setting process,
 16      is not the focus in this chapter. The reader is referred to the NARSTO Fine Particle Assessment
 17      (NARSTO, 2003) for information relevant to air quality management for PM.
 18          Unlike other criteria pollutants (O3, CO, SO2, NO2, and Pb), PM is not a specific chemical
 19      entity but is a mixture of particles from different sources and of different sizes, compositions,
 20      and  properties. Emphasis is placed here on discussion of differences between fine and coarse
 21      particles and differences between ultrafine particles and accumulation-mode particles within fine
 22      particles.
 23          PM is defined quantitatively by the measurement techniques employed.  Therefore, it will
 24      be useful to discuss our understanding of the relationship between PM suspended in the
 25      atmosphere, PM inhaled by people, and PM measured by various sampling and analytical
 26      techniques.  Chapter 4 of the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a)
27      contained a review of the state-of-the-art of PM measurement technology. Since that time,
28      considerable progress has been made in understanding problems in the measurement of PM
29      mass, chemical composition, and physical parameters. There also has been progress in
30      developing new and improved measurement techniques, especially for continuous
31      measurements. Therefore, a more extensive survey on measurement problems and on newly

        June 2003                                2-1         DRAFT-DO NOT QUOTE OR CITE

-------
 1     developed measurement techniques is included in Section 2.2. For more detail and older
 2     references, the reader is referred to Chapters 3 and 4 of the 1996 PM AQCD (U.S.
 3     Environmental Protection Agency, 1996a).
 4
 5
 6     2.1  PHYSICS AND CHEMISTRY OF PARTICIPATE MATTER
 7     2.1.1    Basic Concepts
 8          Atmospheric particles originate from a variety of sources and possess a range of
 9     morphological, chemical, physical, and thermodynamic properties. Examples of atmospheric
10     particles include combustion-generated particles, such as diesel soot or fly ash; photochemically
11     produced particles, such as those found in urban haze; salt particles formed from sea spray; and
12     soil-like particles from resuspended dust.  Some particles are liquid; some are solid. Others may
13     contain a solid core surrounded by liquid. Atmospheric particles contain inorganic ions, metallic
14     compounds, elemental carbon, organic compounds, and crustal compounds. Some atmospheric
15     particles are hygroscopic and contain particle-bound water. The organic fraction is especially
16     complex, containing hundreds  (probably thousands) of organic compounds. (See Appendix 3C
17     for information on the composition of the organic fraction and the concentration of specific
18     organic compounds.) Primary particles are emitted directly from sources; whereas secondary
19     particles are formed from gases through chemical reactions in the atmosphere involving
20     atmospheric oxygen (Oj) and water vapor (H2O); reactive species such as ozone (O3); radicals
21     such as the hydroxyl (-OH) and nitrate (*NO3) radicals; and pollutants such as sulfur dioxide
22     (SO2), nitrogen oxides (NOjJ, and organic gases from natural and anthropogenic sources.  The
23     particle formation process includes nucleation of particles from low-vapor pressure gases
24     emitted from sources or formed in the atmosphere by chemical reactions, condensation of
25     low-vapor pressure gases on existing particles, and coagulation of particles. Thus, any given
26     particle may contain PM from  many sources. Because a particle from a given source is likely to
27     be composed of a mixture of chemical components and particles from different sources may
28     coagulate to form a new particle, atmospheric particles may be considered a mixture of mixtures.
29          The composition and behavior of particles are fundamentally linked with those of the
30     surrounding gas. An aerosol may be defined as a suspension of solid or liquid particles in air.
31     The term aerosol includes both the particles and all vapor or gas phase components of air.

       June 2003                                 2-2         DRAFT-DO NOT QUOTE OR CITE

-------
   1      However, the term aerosol is sometimes used to refer to the suspended particles only.  In this
   2      document, "paniculate" is used only as an adjective, as in particulate matter.
   3           A complete description of the atmospheric aerosol would include an accounting of the
   4      chemical composition, morphology, and size of each particle and the relative abundance of each
   5      particle type as a function of particle size (Friedlander, 1970). However, the physical and
   6      chemical characteristics of particles are usually measured separately. Size distributions by
   7      particle number, used to calculate surface area and volume distributions, often are determined by
   8      physical means, such as electrical mobility, aerodynamic behavior, or light scattering.  Chemical
   9      composition usually is determined by analysis of collected samples although some species can
 10      be measured in situ.  The mass and average chemical composition of particles, segregated
 11      according to aerodynamic diameter by cyclones or impactors, can also be determined.  However,
 12      recent developments in single particle analysis techniques by electron microscopy with X-ray
 13      analysis of single particles (but not agglomerates) collected on a substrate or by mass
 14      spectroscopy of individual suspended particles provide elemental composition of individual
 15      particles by particle size and, thus, are bringing the description envisioned by Friedlander closer
 16      to reality.
 17
 18      2.1.2    Physical Properties  and Processes
 19      2.1.2.1   Definitions of Particle Diameter
 20          The diameter of a spherical particle may be determined by optical or electron microscopy,
 21      by light scattering and Mie theory, by its electrical mobility, or by its aerodynamic behavior.
 22      However, atmospheric particles often are not spherical.  Therefore, their diameters are described
 23      by an "equivalent" diameter (i.e., the diameter of a sphere that would have the same physical
 24     behavior). An optical diameter is the diameter of a spherical particle, with the same refractive
 25     index as the particle used to calibrate the optical particle sizer, that scatters the same amount of
 26     light into the solid angle measured. Diffusion and gravitational settling are important physical
 27     behaviors for particle transport, collection, and removal processes, including deposition in the
 28     respiratory tract. Different equivalent diameters are used depending on which process is more
29     important. For smaller particles diffusion is more important and the Stokes diameter is often
30     used. For larger particles gravitational setting is more important and the aerodynamic diameter
31      is often used.

        June 2003                                  2-3         DRAFT-DO NOT QUOTE OR CITE

-------
 1           The Stokes diameter, Dp, describes particle size based on the aerodynamic drag force
 2      imparted on a particle when its velocity differs from that of the surrounding fluid.  For a smooth,
 3      spherically shaped particle, Dp exactly equals the physical diameter of the particle. For
 4      irregularly shaped particles, Dp is the diameter of an equivalent sphere that would have the same
 5      aerodynamic resistance. Electrical mobility analyzers classify particles according to their
 6      electrical mobility.  Particles of equal Stokes diameters that carry the same electric charge will
 7      have the same electrical mobility. Hence, for  spherical particles, the electrical mobility diameter
 8      would equal the Stokes diameter. The mobility diameter can be considered the diameter of a
 9      spherical particle that would have the same electrical mobility. The particle mobility  can be
10      related to the particle diffusion coefficient and Brownian diffusion velocity through the Stokes-
11      Einstein equation. Thus, the Stokes diameter  is the appropriate parameter for particle behavior
12     governed by diffusion. The Stokes diameter, Dp, is used in size distributions based on light
13     scattering and mobility analysis. The Stokes diameter is independent of density.
14           The aerodynamic diameter, Da, however, depends on particle density. It is defined as the
15     diameter of a spherical particle with an equal  gravitational settling velocity but a material density
16     of 1 g/cm3. Cascade impactors separate particles based on their aerodynamic diameter and
17     aerodynamic particle sizers measure the aerodynamic diameter.  Respirable, thoracic, and
18     inhalable sampling and PM2.5 and PM10 sampling are based on particle aerodynamic diameter.
19     For particles greater than about 0.5 urn, the aerodynamic diameter is generally the quantity of
20      interest. For smaller particles, the Stokes diameter may be more useful. Particles with the same
21     physical size and shape but different densities will have the same Stokes diameter but different
22      aerodynamic diameters.
23           Aerodynamic diameter, D^ is related to the Stokes diameter, Dp, by:
24
                                               f ~s-< \ 1/2
                                                                                            (2-1)
25
26     where p is the particle density, and Cp and Ca are the Cunningham slip factors evaluated for the
27     particle diameters Dp and Da respectively. The slip factor is a function of the ratio between
28     particle diameter and mean free path of the suspending gas (0.066 urn for air at one atmosphere

        June 2003                                 2-4         DRAFT-DO NOT QUOTE OR CITE

-------
  1      pressure and 20 °C). C is an empirical factor that accounts for the reduction in the drag force on
  2      particles due to the "slip" of the gas molecules at the particle surface.  It is important for
  3      particles less than 1 um in diameter, for which the surrounding air cannot be modeled by a
  4      continuous fluid. For large particles (Dp > 5 urn) C = 1; while for smaller particles C> 1.
  5           For particles with diameters greater than the mean free path, A., the aerodynamic diameter
  6      given by equation (2- 1 ) is approximated by:
  7
                                                                 >                        (2-2)
  8
  9      This expression, which shows that aerodynamic diameter is directly proportional to the square
 10      root of the particle density, is often used for particles as small as 0.5 um.  For particles with
 1 1      diameters much smaller than the mean free path, the slip factor must be taken into account.
 12      In this case the aerodynamic diameter is directly proportional to the particle density,
 13
D=                                                  (2.3)
                                       a
 14
 15     Detailed definitions of the various sizes and their relationships are given in standard aerosol
 16     textbooks (e.g., Friedlander [2000], Reist [1984, 1993], Seinfeld and Pandis [1998], Hinds
 17     [1999], Vincent [1989, 1995], Willeke and Baron [1993], Baron and Willeke [2002], and Fuchs
 18     [1964, 1989]).
 19
 20     2.1.2.2  Aerosol Size Distributions
 21          Particle size as indexed by one of the "equivalent" diameters is an important parameter in
 22     determining the properties, effects, and fate of atmospheric particles. The atmospheric
 23     deposition rates of particles and therefore their residence times in the atmosphere are a strong
 24     function of their Stokes and aerodynamic diameters. The diameter also influences deposition
 25     patterns of particles within the lung.  Because light scattering is strongly dependent on the
26     optical particle size, the amount of light scattering per unit PM mass will be dependent on the
                                                                                            r
27     size distribution of atmospheric particles. Therefore, the effects of atmospheric particles on


        June 2003                                  2-5         DRAFT-DO NOT QUOTE OR CITE

-------
 1     visibility, radiative balance, and climate will be influenced by the size distribution of the
 2     particles. Studies using cascade impactors or cyclones measure the particle-size distribution
 3     directly in aerodynamic diameter. The diameters of atmospheric particles range from 1 nm to
 4     100 urn, spanning 5 orders of magnitude.  A variety of different instruments, measuring a variety
 5     of equivalent diameters, are required to cover this range.
 6           Older particle counting studies used optical particle counters to cover the range of 0.3 to
 7     30 um diameter. Diameters of particles below 0.5 urn were measured as mobility diameters.
 8     The particle diameters used in size distribution graphs from these studies usually are given as
 9     physical or Stokes diameters rather than aerodynamic diameters.  In recent years, aerodynamic
10     particle sizers have been developed that give a direct measurement of the aerodynamic diameter
11     in the range of approximately 0.7 to 10 um diameter. These instruments have been used with
12     electrical mobility analyzers that measure the mobility diameter of particles from 3 nm to
13     approximately 0.5 um (McMurry, 2000).  Unfortunately, there is no agreed-upon technique for
14     combining the various equivalent diameters.  Some workers use various assumptions to combine
15     the various measurements into one presentation; others report each instrument separately.
16     Therefore, the user of size distribution data should be careful to determine exactly which
17     equivalent diameter is reported.
18
19     Particle Size Distribution Functions
20           The distribution of particles with respect to size is an important physical parameter
21     governing their behavior. Because atmospheric particles cover several orders of magnitude in
22     particle size, size distributions often are expressed in terms of the logarithm of the particle
23     diameter on the X-axis and the measured differential concentration on the Y-axis: AN/A(logDp)
24     = the number of particles per cm3 of air having diameters in the size range from log Dp to log(Dp
25     + ADA Because logarithms do not have dimensions, it is necessary to think of the distribution
26     as a function of logCDp/Dpo), where the reference diameter Dpo = 1 um is not explicitly stated. If
27     AN/A(logDp) is plotted on a linear scale, the number of particles between Dp and Dp + ADp is
28     proportional to the area under the curve of AN/A(logDp) versus logDp. Similar considerations
29     apply to distributions of surface, volume, and mass. When approximated by a function, the
30     distributions are usually given as dN/d(log Dp) rather than AN/A(log Dp).
31

       June 2003                                 2-6         DRAFT-DO NOT QUOTE OR CITE

-------
   1      Atmospheric Aerosol Size Distributions
   2           In 1978, Whitby (1978) published an analysis of over 1000 particle size distributions
   3      measured at various locations in the U.S. Figure 2-1 shows the number, surface area, and
   4      volume distributions for the grand average  continental size distribution. Volume, surface area,
   5      and number distributions are plotted on an arithmetic scale such that the volume, surface area, or
   6      number of particles in any specified size range is proportional to the corresponding area under
   7      the curve. These distributions show that most of the particles are quite small, below 0.1 urn;
   8      whereas most of the particle volume (and therefore most of the mass) is found in particles
   9      > 0.1 um. Other averaged atmospheric size distributions are shown in Figure 2-2 and 2-3
 10      (Whitby, 1978; Whitby and Sverdrup, 1980). Figure 2-2a and b describe the number of particles
 11      as a function of particle diameter for rural, urban-influenced rural, urban, and freeway-
 12      influenced urban aerosols. For some of the same data, the particle volume distributions are
 13      shown in Figure 2-3a and b. Whitby (1978) observed that the size distributions typically had
 14      three peaks which he called "modes."  The entire size distribution could be characterized well by
 15      a trimodal model consisting of three additive log-normal distributions.  The mode with a peak
 16      between 5 and 30 um diameter, formed by mechanical processes, was named the coarse particle
 17      mode; the mode with a peak between 0.15 and 0.5 um, formed by condensation and coagulation,
 18      was called the accumulation mode; and the mode with a peak between 0.015 and 0.04 um,
 19      whose size was influenced by nucleation as  well as by condensation and coagulation, was called
 20      the transient nuclei or Aiken nuclei range, subsequently  shortened to the nuclei mode. The
 21      nuclei mode could be seen in the number and surface distribution but only in special situations
 22     was it noticeable in the mass or volume distributions. The accumulation and nuclei modes taken
 23     together were called fine particles.  An idealized size distribution showing modes and formation
 24     mechanisms is shown in Figure 2-4.
 25          Whitby (1978) concluded
 26
 27                 The distinction between "fine particles" and "coarse particles" is a fundamental one.
 28           There is now an overwhelming amount of evidence that not only are two modes in the mass or
 29           volume distribution usually observed, but that these fine and coarse modes are usually
3®           chemically quite different.  The physical separation of the fine and coarse modes originates
31            because condensation produces fine particles while mechanical processes produce mostly
32            coarse  particles ... the dynamics of fine particle growth ordinarily operate to prevent the fine

        June 2003                                  2-7         DRAFT-DO NOT QUOTE OR CITE

-------
                                  Nn = 7.7X10
                                DGNn = 0,013
                                                                      100
Figure 2-1. Distribution of coarse (c), accumulation (a), and nuclei (n) mode particles by
           three characteristics, a) number (N), b) surface area (S) and c) volume (V) for
           the grand average continental size distribution. DGV = geometric mean
           diameter by volume; DGS = geometric mean diameter by surface area;
           DGN = geometric mean diameter by number; Dp = particle diameter.

Source: Whitby (1978).
June 2003
2-8
DRAFT-DO NOT QUOTE OR CITE

-------
   1,000,000 -

     10,000 -

pC    loo-

X—*•      Jt
 a.      1
D
 o>
%     001 -I
z
     0.0001 -

   0.000001 -
                           • Clean Rural
                            Urban influenced Rural
                          - - Average Urban
                          — Urban •*- Freeway
                    0.01    0,1     1      10    100
                       Particle Diameter, Dp (pm)
                                                      T3
                                                          200,000
                                                          150,000
                                                      ,0
                                                      Q-  100,000
                                                      O)
                                                      _p
                                                      T3
                                                           50,000 -
                                                      0,01        0.1        1         10
                                                          Particle Diameter, Dp (um)
        Figure 2-2. Particle size distributions by number, (a) number concentrations are shown on
                    a logarithmic scale to display the wide range by site and size; (b) number
                    concentrations for the average urban distribution are shown on a linear scale.
                    For the linear scale, the area under any part of the curve is proportional to
                    particle number in that size range.

        Source: Whitby (1978); Whitby and Sverdrup (1980).
 1            particles from growing larger than about 1 um. Thus, the fine and coarse modes originate
 2            separately, are transformed separately, are removed separately, and are usually chemically
 3            different... practically all of the sulfur found in atmospheric aerosol is found in the fine
 4            particle fraction. Thus, the distinction between fine and coarse fractions is of fundamental
 5            importance to any discussion of aerosol physics, chemistry, measurement, or aerosol air quality
 6            standards.
 7
 8           Whitby's  (1978) conclusions were based on extensive studies of size distributions in a

 9      number of western and midwestern  locations during the 1970s (Whitby et ah, 1974; Willeke

10      and Whitby, 1975; Whitby, 1978; Wilson et al, 1977; Whitby and  Sverdrup, 1980).
11      No size-distribution studies of similar scope have been published since then. Newer results
        June 2003
                                          2-9
DRAFT-DO NOT QUOTE OR CITE

-------
   70
   65-
   60-
   55-
t50-
-y 45-
co
i. 40~
*CL 35-
Q
O) 30-
O
^ 25-
•O 20-
   15-
   10-
    5-
                                      Clean Rural
                                      Urban Influenced
                                      Rural
                                      South-Central
                                      New Mexico
     0.01      0.1        1        10
           Particle Diameter, Dp (Mm)
   70
   65-
   60-
   55-
*£ 50-
-^ 45-
rt
 E 40-
^ 35-
Q
 O) 30-
5 25-
"O 20-
   15-
   10-
    5-
                                                           ET
                                                 100
                                Average Urban
                                Urban + Freeway
        0.01      0.1       1       10      100
             Particle Diameter, Dp (pm)
       Figure 2-3. Size distribution by volume (a) for the averaged rural and urban-influenced
                   rural number distributions shown in Figure 2-2a and a distribution from
                   south central New Mexico, and (b) for the averaged urban and freeway-
                   influenced urban number distributions shown in Figure 2-2a.
       Source: Whitby and Sverdrup (1980); Kim et al. (1993) south central New Mexico.
 I     from particle counting and impactor techniques, including data from Europe (U.S.
 2     Environmental Protection Agency, 1996a) and Australia (Keywood et al., 1999,2000), show
 3     similar results for the accumulation and coarse modes. Extensive measurements of particle size
 4     distributions, as part of EPA's Supersites program, are providing much new data for analysis.
 5          Whitby's (1978) conclusions have held up remarkably well. However,  ideas about the
 6     sub-0.1  um diameter range have changed somewhat as newer instruments provided
 7     measurements extending to smaller sizes and with greater resolution in size and time (McMurry
 8     et al., 2000). Depending on the source, temperature, saturated vapor pressure of the components,
 9     and the  age of the aerosol, size distributions have been observed with peaks (including multiple
10     peaks) throughout the sub-0.1 jxm diameter size range. Sub-0.1 ^m diameter peaks have been
       June 2003
2-10
                                                   DRAFT-DO NOT QUOTE OR CITE

-------
                                                                Mechanically
                                                                 Generated
                     0.002
                              0.01
                           Nuclei Mode
                                     0.1            1
                                   Particle Diameter, Dp(pm)
                                   Accumulation Mode
                                                                                  100
                                    Fine Particles
       Coarse Mode
     Coarse Particles
       Figure 2-4.  Volume size distribution, measured in traffic, showing fine and coarse
                   particles and the nuclei and accumulation modes of fine particles.  DGV
                   (geometric mean diameter by volume, equivalent to volume median diameter)
                   and ag (geometric standard deviation) are shown for each mode. Also shown
                   are transformation and growth mechanisms (e.g., nucleation, condensation,
                   and coagulation).
       Source: Adapted from Wilson and Suh (1997).
 1
 2
 3
 4
 5
 6
 7
 8
9
observed in rural areas (O'Dowd, 2002) as well as for brief periods (nucleation bursts) in urban
areas (Woo et al., 2001a). Based on these and other observations, discussed in detail in
Section 2.1.2.3, aerosol scientists now classify particles in the sub-0.1 urn size range as ultrafme
particles and divide this size range into a nucleation region (< 10 nm) and an Aitkin (nuclei)
region (10-100 nm), as shown in Figure 2-5. Other studies, discussed in detail in the 1996 PM
AQCD (U.S. Environmental Protection Agency, 1996a), have shown that in fog or clouds or at
very high relative humidities the accumulation mode may split into a larger size (more
hygroscopic or droplet) submode and a smaller size (less hygroscopic or condensation) submode.
      June 2003
                                         2-11
DRAFT-DO NOT QUOTE OR CITE

-------
2.000-
"E 1.500-
g_ 1.000-
dN/dlog Dp
in
§


Nudeation
Mode
r
-1
l,r>T


i-





— |



Aitken Mode

~


-, Accumulation
Mode
ih" rhp
                                        ID     20       50    100
                                              Particle Diameter (nm)
                                                                 200
                                                                         500
                                                                              1.000
       Figure 2-5.  Submicron number size distribution observed in a boreal forest in Finland
                   showing the tri-modal structure of fine particles.  The total particle number
                   concentration was 1011 particles/cm3 (10 minute average).
       Source: Makelaetal. (1997).
 1      Definitions of Particle Size Fractions
 2           In the preceding discussion several subdivisions of the aerosol size distribution were
 3      identified. Aerosol scientists use three different approaches or conventions in the classification
 4      of particles by size: (1) modes, based on the observed size distributions and formation
 5      mechanisms; (2) cut point, usually based on the 50% cut point of the specific sampling device,
 6      including legally specified, regulatory sizes for air quality standards; and (3) dosimetry or
 7      occupational health sizes, based on the entrance into various compartments of the respiratory
 8      system.
 9
10          Modal. The modal classification as first proposed by Whitby (1978) is shown in
11      Figures 2-1 and 2-4.  New modes introduced since 1978 are shown in Figure 2-5. The
12     nucleation and Aitkin modes are best observed in the number distribution.  However, the Aitkin
13     mode can be seen in the volume distribution in traffic or near traffic or other sources of ultrafme
14     particles (Figures 2-3b and 2-4). The observed modal structure is frequently approximated by
        June 2003
2-12
DRAFT-DO NOT QUOTE OR CITE

-------
   1      several log-normal distributions.  Definitions of terms used to describe size distributions in
   2      modal terms are given below.
   3
   4           Nucleation Mode: Freshly formed particles with diameters below 10 nm, observed during
   5           active nucleation events. The lower limit, where particles and large molecules overlap, is
   6           uncertain.  Current techniques limit measurements to particles 3 nm or greater.
   7
   8           Aitkin Mode: Larger particles with diameters between 10 and 100 nm.  The Aitken mode
   9           may result from growth of smaller particles or nucleation from higher concentrations of
 10           precursors. Nucleation and Aitldn nuclei modes are normally observed in the number
 11           distribution.
 12
 13          Accumulation Mode: Particles with diameters from about 0.1 nm to just above the
 14          minimum in the mass or volume distributions which usually occurs between 1 and 3 urn.
 15          Accumulation-mode particles normally do not grow into the coarse mode. Nucleation-
 16          mode and Aitkin-mode particles grow by coagulation (two particles combining to form
 17          one) or by condensation (low-equilibrium vapor pressure gas molecules condensing on a
 18          particle) and "accumulate" in this size range.
 19
 20          Coarse Mode or Coarse Particles: Particles with diameters mostly greater than the
 21          minimum in the particle mass or volume distributions, which generally occurs between
 22          1 and 3 urn. These particles are usually formed by mechanical breakup of larger particles
 23          or bulk material.
 24
 25          Fine Particles: Fine particles include the nucleation, Aitkin, and accumulation modes, i.e.,
 26          particles from the lowest measurable size, currently about 3 nm, to just above the minimum
 27          in the mass or volume distribution which generally occurs between 1 and 3 urn. These
 28           particles are generated during combustion or formed from gases.
 29
30           Ultrafme Particles: That portion of fine particles with diameters below about 0.1  urn
31           (100 nm), i.e., the Aitkin and nucleation modes.

       June 2003                                2-13        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Modes are defined primarily in terms of their formation mechanisms but also differ in
 2      terms of sources, composition, age, and size. Nucleation mode applies to newly formed particles
 3      which have had little chance to grow by condensation or coagulation.  Aitkin mode particles are
 4      also recently formed particles that are still actively undergoing coagulation. However, because
 5      of higher concentrations of precursors or more time for condensation and coagulation, the
 6      particles have grown to larger sizes. Accumulation mode applies to the final stage as particles,
 7      originally formed as nuclei, grow to a point where growth slows down. These three modes,
 8      which together are called fine particles, are formed primarily by combustion or chemical
 9     reactions of gases yielding products with low saturated vapor pressures. Fine particles include
10     metals and elemental and organic carbon (primary PM) and sulfate, nitrate, ammonium ions, and
11      organic compounds (secondary PM).
12           The coarse mode refers to particles formed by mechanical breakdown of minerals, crustal
13     material, and organic debris.  The composition includes primary minerals and organic material.
14     The accumulation mode and the coarse mode overlap in the region between 1 and 3 um (and
15     occasionally over an even larger range).  In this region, chemical composition of individual
16     particles can usually, but not always, allow identification of a source or formation mechanism
17     and so permit identification of a particle as belonging to the accumulation or coarse mode.
18
19           Sampler Cut Point. Another set of definitions of particle size fractions arises from
20     considerations of size-selective sampling. Size-selective sampling refers to the collection of
21     particles below or within a specified aerodynamic size range. Size fractions are usually specified
22     by the 50% cut point size; e.g., PM2.5 refers to particles collected by a sampling device that
23      collects 50% of 2.5 um particles and rejects 50% of 2.5 urn particles.  However, size fractions
24      are defined, not merely by the 50% cut point, but by the entire penetration curve. Examples of
25     penetration curves are given in Figure 2-6. Thus, as shown by Figure 2-6, a PM2 5 sampler,  as
26      defined by the Federal Reference Method, rejects 94% of 3 um particles, 50% of 2.5 um
27      particles, and 16% of 2 um. Samplers with the same 50% cut point but differently shaped
28      penetration curves would collect different fractions of PM. Size-selective sampling has arisen in
29      an effort to measure particle size fractions with some special significance (e.g., health, visibility,
30      source apportionment, etc.), to measure mass size distributions, or to collect size-segregated
 31      particles for chemical analysis.  Dichotomous samplers split the particles into smaller and larger

        June 2003                                 2-14        DRAFT-DO NOT QUOTE OR CITE

-------
                          100
                                     2.5   4         10   20        50
                                          Aerodynamic Diameter (um)
       Figure 2-6. Specified particle penetration (size-cut curves) through an ideal (no-particle-
                   loss) inlet for five different size-selective sampling criteria. Regulatory size
                   cuts are defined in the Code of Federal Regulations; PM2.5 (2001c), PMIO
                   (2001a). PM2.5 is also defined in the Federal Register (1997). Size-cut curves
                   for inhalable participate matter (IPM), thoracic particulate matter (TPM) and
                   respirablc particulate matter (RPM) size cuts are computed from definitions
                   given by American Conference of Governmental and Industrial Hygienists
                   (1994).
 1     fractions that may be collected on separate filters. However, some fine particles (=10%) are
 2     collected with the coarse particle fraction. Cascade impactors use multiple size cuts to obtain a
 3     distribution of size cuts for mass or chemical composition measurements. One-filter samplers
 4     with a variety of upper size cuts are also used, e.g., PM2,5, PMI0.
 5          Regulatory size cuts are a specific example of size-selective sampling.  In 1987, the
 6     NAAQS for PM were revised to use PM10, rather than total suspended particulate matter (TSP),
 7     as the indicator for the NAAQS for PM (Federal Register, 1987). The use of PMj0 as an
 8     indicator is an example of size-selective sampling based on a regulatory size cut (Federal
 9     Register, 1987). The selection of PM10 as an indicator was based on health considerations and
10     was intended to focus regulatory concern on those particles small enough to enter the thoracic
11     region of the human respiratory tract. The PM25 standard set in 1997 is also an example of
       June 2003
2-15
DRAFT-DO NOT QUOTE OR CITE

-------
 1      size-selective sampling based on a regulatory size cut (Federal Register, 1997). The PM25
 2      standard was based primarily on epidemiologic studies using concentrations measured with
 3      PM2 5 samplers as an exposure index. However, the PM2 5 sampler was not designed to collect all
 4      respirable particles; it was designed to collect fine particles because of their different sources and
 5      properties (Whitby et al., 1974). Thus, the need to attain a PM25 standard will tend to focus
 6      regulatory concern on control of sources of fine particles.
 7          Prior to 1987, the indicator for the NAAQS for PM was TSP.  TSP is defined by the design
 8      of the High Volume Sampler (hivol) that collects all of the fine particles but only part of the
 9      coarse particles (Figure 2-7). The upper cut-off size of the hivol depends on the wind speed and
10      direction and may vary from 25 to 40 urn.  The Wide Range Aerosol Classifier (WRAC) was
11      designed specifically to collect the entire coarse mode using an impaction system designed by
12      Lundgren to collect particles up to 100 urn in diameter (Lundgren and Burton, 1995).
13
                         70
                     &
60 -

50 -

40 -

30 -

20 -

10 -
                          0.1
                               Accumulation Mode
                                    Coarse Mode
                                0.2
               i^    I     l^      1     I
               0.5     1     2       5    10
                   Particle Diameter, Dp{jjm)
               Total Suspended Particles (TSP) —
              	  PM,
                                              "10
                                       PM
                                         '2.5
                               PM
                                                                   20
                                                                          TSP
                                                                          HiVbl
             WRAC
                                                                          SO
                                                                               10C
                                                         1045
       Figure 2-7.  An idealized distribution of ambient participate matter showing fine-mode
                   particles and coarse-mode particles and the fractions collected by size-selective
                   samplers. (WRAC is the Wide Range Aerosol Classifier which collects the
                   entire coarse mode [Lundgren and Burton, 1995].)
       Source: Adapted from Wilson and Suh (1997) and Whitby (1978).
       June 2003
                        2-16
DRAFT-DO NOT QUOTE OR CITE

-------
   1           An idealized distribution with the normally observed division of ambient aerosols into
   2      fine-mode particles and coarse-mode particles and the size fractions collected by the WRAC,
   3      TSP, PM,0, PM2 5 and PM]0_2 5 samplers is shown in Figure 2-7. PM10 samplers, as defined in
   4      Appendix J to Title 40 Code of Federal Regulations (40 CFR) Part 50 (Code of Federal
   5      Regulations, 2001a; Federal Register, 1987), collect all of the fine-mode particles and part of the
   6      coarse-mode particles. The upper cut point is defined as having a 50% collection efficiency at
   7      10 ± 0.5 um aerodynamic diameter. The slope of the collection efficiency curve is defined in
   8      amendments to 40 CFR, Part 53, (Code of Federal Regulations, 2001b). An example of a PM10
   9      size-cut curve is shown in Figure 2-6.
 10           An example of a PM25 size-cut curve is also shown in Figure 2-6.  The PM2 5 size-cut
 11      curve, however, is defined by the design of the Federal Reference Method (FRM) Sampler.  The
 12      basic design of the FRM sampler is given in the Federal Register (1997, 1998) and as 40 CFR
 13      Part 50, Appendix L (Code of Federal Regulations, 200ic). Additional performance
 14      specifications are given in 40 CFR Parts 53 and 58 (Code of Federal Regulations, 2001b,d).
 15      In order to be used for measurement of PM2 5 to determine compliance with the PM2 5 NAAQS,
 16      each specific sampler design and its associated manual of operational procedures must be
 17      designated as a reference method under 40 CFR Part 53 in Section 1.2 of Appendix L (Code of
 18      Federal Regulations, 200 Ic).  Thus PM25 FRM samplers may have somewhat different designs
 19      (see Table 2-4 in Section 2.2.4.1.2).
 20          Papers discussing PM10 or PM2 5 frequently insert an explanation such as "PMX (particles
 21     less than x um diameter)" or "PMX (nominally, particles with aerodynamic diameter < x urn)."
 22     While these explanations may seem easier than "PMX, (particles collected with an upper 50% cut
 23     point of x um aerodynamic diameter and a specified penetration curve)," they are not entirely
 24     correct and may be misleading since they suggest an upper 100% cut point of x urn. Some
 25     countries use PM10 to refer not to  samplers with a 50% cut at 10 um Da but to samplers with
 26     100% rejection of all particles greater than 10 um Da.  Such samplers miss a fraction of coarse
 27     thoracic PM.  An example is shown in Figure 2-8.
 28         PMi0, as defined by EPA, refers to particles collected by a sampler with an upper 50% cut
 29     point of 10 um Da and a specific, fairly sharp, penetration curve. PM25 is analogously defined.
30      Although there is not yet an FRM, PM^ refers either to particles collected by a sampler with
31      an upper 50% cut point of 10 um Da and a lower 50% cut point of 2.5 um Da or to the difference

        June 2003                                2-17         DRAFT-DO NOT QUOTE OR CITE

-------

                         100
                          90-
                          80-
                          70-
                          60
                          50
                          40
                          30
                          20
                          10
                                 Wedding Cvdonic Inlet
                                 |  O u=2 km/h
                                 i  D U=8 km/h
                                 I  A U=24 km/h
                                 I Wedding and Wsigand
                               ft !      <1«S>
  Kimoto cydonic inlet
      • Manufacturer
       Tsai and Cheng (1996)
1
                                                4     6    8  10
                                           Aerodynamic Diameter (pm)
                                           20
                                                                          e—t
30
       Figure 2-8.  Comparison of penetration curves for two PM10 beta gauge samplers using
                   cyclone inlets. The Wedding PMIO sampler uses the U.S. EPA definition of
                   PM, as x = 50% cut point. The Kimoto PM10 defines PMX as x = the 100% cut
                   point (or zero penetration).
       Source: Tsai and Cheng (1996).
 1      between the particle concentration measured by a PMi0 monitor and a PM2.5 monitor. In all
 2      cases, the fraction of PM collected depends on the entire penetration curve (or curves); i.e., for
 3      PM2 5 some particles > 2.5 um Da are collected and not all particles < 2.5 urn Da are collected.
 4           In an analysis reported in 1979, EPA scientists endorsed the need to measure fine and
 5      coarse particles separately (Miller et al., 1979). Based on the availability of a dichotomous
 6      sampler with a separation size of 2.5 um D^ they recommended 2.5 um Da as the cut point
 7      between fine and coarse particles. Because of the wide use of this cut point, the PM2 5 fraction is
 8      frequently referred to as "fine" particles. However, although the PM2 5 sample will usually
 9     contain all of the fine particles, it may collect a small fraction of the coarse particles, especially
10     in dry areas or during dry conditions.  A PM10_2.5 size fraction may be obtained from a
11      dichotomous sampler or by subtracting the mass collected by a PM2 5 sampler from the mass
12     collected by a PM10 sampler. The resulting PM10.2 5 mass, or PM10i5, is sometimes called
13     "coarse" particles. However, it would be more correct to call PM2 5 an indicator of fine particles
        June 2003
                                                  2-18
                                 DRAFT-DO NOT QUOTE OR CITE

-------
   1     (because it contains some coarse particles) and PM10_2.5 an indicator of the thoracic component of
   2     coarse particles (because it excludes some coarse particles below 2.5 urn Da and above 10 urn
   3     DJ. It would be appropriate to call PMi0 an indicator of thoracic particles. PM10 and thoracic
   4     PM, as shown in Figure 2-6, have the same  50% cut point.  However, the thoracic cut is not as
   5     sharp as the PM,0 cut; therefore, thoracic PM contains some particles between 10 and 30 urn
   6     diameter that are excluded from PM,0.
   7          Over the years, the terms fine and coarse, as applied to particles, have lost the precise
   8     meaning given in Whitby's (1978) definition.  In any given article, therefore, the meaning of fine
   9     and coarse, unless defined, must be inferred from the author's usage. In this document, fine
  10     particles means all particles in the nucleation, Aitken, and accumulation modes; and coarse
  11      particles means all particles in the coarse mode. PM25 and fine particles are not equivalent
  12      terms.
  13
  14           Occupational Health or Dosimetric Size Cuts. The occupational health community has
  15      defined size fractions in terms of their entrance into various compartments of the respiratory
  16      system. This convention classifies particles  into inhalable, thoracic, and respirable particles
  17      according to their upper size cuts.  Inhalable particles enter the respiratory tract, including the
  18      head airways.  Thoracic particles travel past the larynx and reach the lung airways and the
 19      gas-exchange regions of the lung.  Respirable particles are a subset of thoracic particles that are
 20      more likely to reach the gas-exchange region of the lung.  In the past, exact definitions of these
 21      terms have varied among organizations.  As of 1993, a unified set of definitions was adopted by
 22      the American Conference of Governmental Industrial Hygienists (ACGIH, 1994), the
 23      International Standards Organization (ISO), and the European Standardization Committee
 24      (CEN). The curves which define inhalable (IPM), thoracic (TPM), and respirable (RPM)
 25     paniculate matter are shown in Figure 2-6. These curves should not be taken to indicate that
 26     particles > 4 urn Da do not reach the gas exchange regions or that particles < 4 urn Da do not
 27     deposit in the bronchi.  See Figure 6-13 for a graphical characterization of particle deposition in
28      regions of the respiratory system as a function of particle size.
29
30
       June 2003                                 2-19        DRAFT-DO NOT QUOTE OR CITE

-------
 1      2.1.2.3   Ultrafine Particles
 2           As discussed in Chapter 7 (Toxicology of Paniculate Matter in Humans and Laboratory
 3      Animals) and in Chapter 8 (Epidemiology of Human Health Effects Associated with Ambient
 4      Paniculate Matter), some scientists argue that ultrafme particles may pose potential health
 5      problems and that some health effects may be associated with particle number or particle surface
 6      area as well as or more closely than with particle mass.  Some additional attention will be given
 7      to ultrafme particles because they contribute the major portion of particle number and a
 8      significant portion of particle surface area.
 9
10      Formation and Growth of Fine Particles
11           Several processes influence the formation and growth of particles. New particles may be
12     formed by nucleation from gas phase material.  Particles may grow by condensation as gas phase
13      material condenses on existing particles; and particles also may grow by coagulation as two
14     particles combine to form one.  Gas phase material condenses preferentially on smaller particles,
15     and the rate constant for coagulation of two particles decreases as the particle size increases.
16     Therefore, ultrafme particles grow into the accumulation mode; but accumulation-mode particles
17     do not normally grow into the coarse mode (see Figure 2-4). More information and references
18     on formation and growth of fine particles can be found in the 1996 AQCD PM (U.S.
19     Environmental Protection Agency, 1996a).
20
21     Equilibrium Vapor Pressures
22           An important parameter in particle nucleation and in particle growth by condensation is  the
23      saturation ratio, S, defined as the ratio of the partial pressure of a species, p, to its equilibrium
24      vapor pressure above a flat surface at a specified temperature, p0: S = p/p0. For either
25      condensation or nucleation to occur, the species vapor pressure must exceed its equilibrium
26      vapor pressure.  For particles, the equilibrium vapor pressure is not the same as p0.  Two effects
27      are important:  (1) the Kelvin effect, which is an increase in the equilibrium vapor pressure
28      above the surface due to its curvature (very small  particles have higher vapor pressures and will
29      not be stable to evaporation until they attain a critical size) and (2) the solute effect, which is a
 30      decrease in the equilibrium vapor pressure of the liquid due to the presence of other compounds
 31      in solution. Organic compounds may also be adsorbed on ultrafme  carbonaceous particles.

        June 2003                                2-20        DRAFT-DO NOT QUOTE OR CITE

-------
   1           For an aqueous solution of a nonvolatile salt, the presence of the salt decreases the
   2      equilibrium vapor pressure of the water over the droplet. This effect is.'in the opposite direction
   3      of the Kelvin effect, which increases the equilibrium vapor pressure above a droplet because of
   4      its curvature. The existence of an aqueous solution will also influence the vapor pressure of
   5      water-soluble species. The vapor pressure behavior of mixtures of several liquids or of liquids
   6      containing several solutes is complex.
   7
   8     New Particle Formation
   9           When the vapor concentration of a species exceeds its equilibrium concentration
  10     (expressed as its equilibrium vapor pressure), it is considered condensable. Condensable species
  11      can either condense on the surface of existing particles or can nucleate to form new particles.
  12      The relative  importance of nucleation versus condensation depends on the rate of formation of
  13      the condensable species and on the surface or cross-sectional area of existing particles (McMurry
  14      and Friedlander, 1979). In ambient urban environments, the available particle surface area is
  15      usually sufficient to rapidly scavenge the newly formed condensable species. Formation of new,
  16      ultrafme particles is usually not observable in mass or volume distributions except near sources
  17      of condensable species. Wilson et al. (1977) report observations of the Aitkin nuclei mode in
  18      traffic. However, bursts of new particle formation can be observed in urban areas in the number
  19      distribution (Woo et al., 2001a; McMurray et al., 2000).  New particle formation also can be
 20      observed in cleaner,  remote regions.  Bursts of new particle formation in the atmosphere under
 21      clean conditions usually occur when aerosol surface area concentrations are low (Covert et al.,
 22      1992). High  concentrations of nuclei mode particles have been  observed in regions with low
 23     particle mass concentrations indicating that new particle formation is inversely related to the
 24     available aerosol surface area (Clarke, 1992).
 25
 26     Sources of Ultrafine Particles
 27          Ultrafme particles are the result of nucleation of gas phase species to form condensed
 28      phase species with very low equilibrium vapor pressure. In the atmosphere there are four major
 29      classes of substances that yield particulate matter with equilibrium vapor pressures low enough
30      to form nuclei mode particles:
31
       June 2003                                 2-21         DRAFT-DO NOT QUOTE OR CITE

-------
1      (1)  Particles containing heavy metals. Nuclei mode particles of metal oxides or other metal
            compounds are generated when metallic impurities in coal or oil are vaporized during
            combustion and the vapor undergoes nucleation. Metallic ultrafine particles also may be
            formed from metals in lubricating oil or fuel additives that are vaporized during
            combustion of gasoline or diesel fuels.  Ultrafine metallic particles were discussed in
            Section 6.9 of the 1996 PM AQCD (U.S. Environmental Protection Agency, 1996a).
2
3      (2)  Elemental carbon or soot (EC).  EC particles are formed primarily by condensation of C2
            molecules generated during the combustion process.  Because EC has a very low
            equilibrium vapor pressure, ultrafine EC particles can nucleate even at high temperatures
            (Kittelson, 1998; Morawska et al., 1998).
4
5      (3)  Organic carbon (OC). Recent smog chamber studies and indoor experiments show that
            atmospheric oxidation of certain organic compounds found in the atmosphere can
            produce highly oxidized organic compounds with an equilibrium vapor pressure
            sufficiently low to result in nucleation (Kamens et al., 1999; Weschler and Shields,
            1999).
6
7      (4)  Sulfates.  Sulfuric acid (H2SO4) molecules are generated in the atmosphere by conversion
            of sulfur dioxide (SOj) to H2SO4. As H2SO4 is formed, it can either nucleate to form new
            ultrafine particles, or it can condense on existing ultrafine or accumulation mode particles
            (Clark and Whitby, 1975; Whitby, 1978). Nucleation theory allows calculation of
            nucleation rates for both binary nucleation, involving water and sulfuric acid (Easter and
            Peters, 1994) or ternary nucleation, which requires sulfuric acid, ammonia (NH3), and
            water (Korhonen et al.,  1999). Kulmala et al. (2000) compared nucleation rates of binary
            and ternary nucleation and concluded that the ternary rate is 1,000 times the binary rate.
            Results from an aerosol dynamics model with a ternary nucleation scheme indicate that
            nucleation in the troposphere should be ubiquitous and yield a reservoir of
            thermodynamically stable sulfate clusters 1-3 nm in diameter. The growth of these
            clusters to a detectable size (> 3 nm diameter) is limited by the availability of
            condensable vapor.  Observations of atmospheric particle formation and growth from a
            continental and a coastal site suggest that a growth process including ternary nucleation is

       June 2003                                 2-22       DRAFT-DO NOT QUOTE OR CITE

-------
 8
 9
10
11
12
13
14
15
16
      responsible for the formation of cloud condensation nuclei. Nucleation processes in the
      atmosphere may also involve organic compounds as well as sulruric acid, ammonia, and
      water.  However, current formulations of nucleation theory only include the three
      inorganic components.  (The possible formation of ultrafme NH4NO3 by reaction of NH3
      and nitric acid (HNO3) vapor has not been investigated.)

      Vehicle engine exhaust may include all these substances. Ultrafme particles are observed
in the emissions from spark, diesel, and jet engines (Kittelson, 1998). In these cases it seems
likely that elemental carbon, organic compounds, ammonia and sulfuric acid from sulfur in the
fuel, as well as metal additives in the fuel or fuel oil, may contribute to the formation of ultrafme
particles (Tobias et al., 2001).  An example of particles from a diesel engine showing number
and volume distributions is given in Figure 2-9.
                                               Ultraflne Particles
                                               Dp < 100 nm
                                                     Accumulation v
                                                       Mode     ^
                                                                 Coarse
                                                                  Mode
                        0.001           0.01
                         Numb«r Weighting	
                                            0.1
                                         Diameter (pm)
10
                                                               Ma» Weighting
             Figure 2-9. Typical engine exhaust size distribution.
             Source: Kittelson (1998).
      June 2003
                                         2-23
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1      Recent Measurements of Ultrafme Particles
 2           Instruments, developed during the past decade, permit measurement of size distributions
 3      down to 3 nm diameter particles. Use of these techniques have led to new information on the
 4      formation of new particles by nucleation. Such measurements have been carried out during
 5      intensive field measurement campaigns, during continuous measurements in urban areas in
 6      several European cities, and in the U.S. as a part of the Supersite program (McMurry et al.,
 7      2000; Woo et al., 200la). Nucleation has been observed in the free troposphere (Weber et al.,
 8      1999; Clarke, 1992; Schroder and Strom, 1997; Raes et al., 1997); in outflows of evaporating
 9      convective clouds (Clarke etal, 1998; Heggetal., 1990,1991 ;Radke and Hobbs, 1991;  Peny
10      and Hobbs, 1994); in the marine boundary layer (Covert et al., 1992; Hoppel et al.,  1994; Van
11      Dingenenetal., 1995; Weber etal., 1998; Clarke etal., 1998); downwind of coastal regions
12     during low tide (McGovern et al., 1996; McGovern, 1999); on mountains (Weber et al., 1995,
13      1997; Raes et al., 1997; Wiedensohler et al., 1997); over forests (MaKela et al.,  1997; Kulmala
14     et al., 1998; O'Dowd et al., 2002); downwind of certain biogenic emissions (Weber et al., 1998);
15     in urban areas (Birmili and Wiedensohler, 1998; McMurry et al., 2000; Woo et al.,  2001a); near
16     freeways (Zhu et al., 2002a,b); in engine exhaust (Kittelson, 1998; Tobias et al., 2001); and in
17     homes (Wallace and Howard-Reed, 2002). Nucleation events in outdoor air almost always occur
18     during daylight, indicating that photochemistry plays a role in producing the gas phase
19     precursors of new particles.
20          The number size distributions observed over a boreal forest in Finland before and during
21     the initial stages of a nucleation event are shown in Figure 2-10. The Aitken and accumulation
22     modes can be seen clearly before the nucleation event. The nucleation mode, with a peak
23     between 3 and 7 nm, appears during the event. Figure 2-11 shows the variety of size
24     distributions  that may be observed as nuclei are formed and grow, based on size distributions
25     measured in the Arctic marine boundary layer (Covert et al.,  1996). These distributions all show
26      a trimodal distribution within the fine particle size range. The changes in size distribution due to
 27      coagulation (and dilution) immediately downwind of a freeway (Zhu et al., 2002b) are shown in
 28      Figure 2-12(a)-(g). At 30 m downwind the nucleation mode  is larger than the Aitken mode, but
 29      by 60 m downwind coagulation has removed particles from the nucleation mode and added
 30      particles to the Aitken mode so that the Aitken mode is larger than the nucleation mode.
 31

        June 2003                                2-24       DRAFT-DO NOT QUOTE OR CITE

-------
              tn
               E
               o
               O
                 102.
                                                                            Before nucleatian
                                                                            During nudeation
                                          10                     100
                                            Particle Diameter (nm)
                                    1,000
        Figure 2-10.  Number size distributions showing measurement of a nucleation burst made
                     in a boreal forest in Finland.
        Source: O'Dowd et al. (2002).
 1           There is strong evidence that sulfuric acid vapor often participates in nucleation.  However,
 2      condensation of sulfuric acid and its associated water and ammonium ions typically can account
 3      for only 10% to 20% of the observed growth rates for freshly nucleated particles. Therefore,
 4      organic compounds may account for much of the formation and growth of freshly nucleated
 5      particles. Evidence of nucleation of organic particles comes from smog chamber studies
 6      (Kamens et al., 1999) and from field studies over forests (MaKelS et al., 1997; Kulmala et al.,
 7      1998; O'Dowd et al., 2002). Nucleation of organic particles may also occur indoors due to the
 8      reaction of infiltrated ozone with indoor terpenes from air fresheners or cleaning solutions
 9      (Weschler and Shields, 1999). The observation of bursts of nuclei-mode particles in Atlanta
10      (Woo et al., 2001a), perhaps due to unusually high rates of production of condensable species,
11      suggests that high concentrations of ultrafine particles may be a normal occurrence in polluted
12      urban areas.
13
14      Concentration of Ultrafine Particles: A Balance Between Formation and Removal
15           Nuclei-mode particles may be removed by dry deposition or by growth into the
16      accumulation mode. This growth takes place as other low vapor pressure material condenses on
17      the particles or as nuclei-mode particles coagulate with themselves or with accumulation mode
       June 2003
2-25
DRAFT-DO NOT QUOTE OR CITE

-------
         1600
         1400-
         1200-
         1000-
          300-
          600-
          400-
          200-
           0
          600
August 30, 1991, 1000 UTC
                        10
                                    100
                                                                                          1000
                                           August 31, 1991, O200 LTTC  K
                                                         September 22,1991,1200 UTC
                             J  MH
                             Q  40-
                             CP
                             _O
                             |  20H
                             ^   0
                                    September 24,1991, 0900 UTC
                                                                               100
                                                                                          1000
                                              10
                                                          100
                                                                    1000
      Figure 2-11.  Examples of the measured one hour average particle number size
                    distributions and the log normal fits to the modes of the data. Squares are
                    measured data, solid lines are the fitted log normal modes determined by
                    DistFit™. These modes, nucleation between 3 and 20 nm, Aitken between
                    20 and 100 nm, and accumulation above 100 nm can be observed in most
                    examples.
      Source: Covert etal. (1996).
1     particles.  Because the rate of coagulation would vary with the concentration of accumulation-
2     mode particles, it might be expected that the concentration of nuclei-mode particles would
3     increase with a decrease in accumulation-mode mass. On the other hand, the concentration of
4     particles would be expected to decrease with a decrease in the rate of generation of particles by
5     reduction in emissions of metal and carbon particles or a decrease in the rate of generation of
6     H2SO4 or condensable organic vapor.  The rate of generation of H2SO4 depends on the
       June 2003
                    2-26
DRAFT-DO NOT QUOTE OR CITE

-------
       2.0e+S
     Z
     f 5.0e+4
        0.0
  1.0e+4


  S.Oe+3

*£
^ 6.0e+3

Q
O)
£ 4.06+3
•o
•^.
•o
  2.06+3
        0.0
                                                 4.0*+4
                                                 2.0e+4
                                                  0.0
                                                                                    TE
                                                                            »   pg-40.1nm
                                                                           **  X(T-»1.52
                                                                       ;*^
-------
                                    10                        100
                                          Particle Diameter (nm)
      Figure 2-12(g). (Combination of a-e with dN/d logDp scale.) Ultrafme particle size
                     distribution at different sampling locations near the 405 freeway in
                     Los Angeles, CA.
      Source: Zhu et al. (2002b).
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
concentration of SO2 and hydroxyl radical (-OH), which is generated primarily by reactions
involving ozone (O3). Thus, reductions in SO2 and O3 would lead to a decrease in the rate of
generation of H2SO4 and condensable organic vapor and to a decrease in the concentration of
nuclei-mode particles. The balance between formation and removal is uncertain. However,
these processes can be modeled using a general dynamic equation for particle size distribution
(Friedlander, 2000) or by aerosol dynamics modules in newer air quality models (Binkowski and
Shanker, 1995; Binkowski and Ching, 1995).

2.1.3   Chemistry of Atmospheric Particulate Matter
     The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
ions; particle-bound water; elemental carbon; a great variety of organic compounds; and crustal
material. Atmospheric PM also contains a large number of elements in various compounds and
concentrations.  More information and references on the composition of PM measured in a large
number of studies in the United  States, may be found in 1996 PM AQCD (U.S. Environmental
       June 2003
                                        2-28
DRAFT-DO NOT QUOTE OR CITE

-------
  1     Protection Agency, 1996a). In this document, the composition and concentrations of PM are
  2     discussed in Chapter 3, Section 3.1, Patterns and Trends in Ambient PM25 Concentrations and
  3     ambient data for concentrations and composition of PM25 are given in Appendices 3A, 3B,
  4     and 3C.
  5
  6     2.1.3.1   Chemical Composition and Its Dependence on Particle Size
  7          Studies conducted in most parts of the United States indicate that sulfate, ammonium, and
  8     hydrogen ions; elemental carbon, secondary organic compounds and primary organic species
  9     from cooking and combustion; and certain transition metals are found predominantly in the fine
 10     particle mode. Crustal materials such as calcium, aluminum, silicon, magnesium, and iron are
 11      found predominately in the coarse particles. Some primary organic materials such as pollen,
 12     spores, and plant and animal debris are also found predominantly in the coarse mode. Some
 13      components such as potassium and nitrate may be found in both the fine and coarse particle
 14      modes but from different sources or mechanisms.  Potassium in coarse particles comes from soil.
 15      Potassium also is found in fine particles in emissions from burning wood or cooking meat.
 16      Nitrate in fine particles comes primarily from the reaction of gas-phase nitric acid with gas-
 17      phase ammonia to form particulate ammonium nitrate. Nitrate in coarse particles comes
 18      primarily from the reaction of gas-phase nitric acid with preexisting coarse particles.
 19
 20      2.1.3.2  Primary and Secondary Particulate Matter
 21           Particulate material can be primary or secondary. PM is called "primary" if it is in the
 22      same chemical form in which it was emitted into the atmosphere.  PM is called "secondary" if it
 23      is formed by chemical reactions in the atmosphere. Primary coarse particles are usually formed
 24      by mechanical processes. This includes material emitted in particulate form such as wind-blown
 25      dust, sea salt, road dust, and combustion-generated particles such as fly ash and soot. Primary
 26      fine particles are emitted from sources either directly as particles or as vapors that rapidly
 27      condense to form ultrafine or nuclei-mode particles. This includes soot from diesel engines,
 28      a great variety of organic compounds condensed from incomplete combustion or cooking, and
29      compounds of As, Se, Zn, etc., that condense from vapor formed during combustion or smelting.
30      The concentration of primary particles depends on their emission rate, transport and dispersion,
31      and removal rate from the atmosphere.

        June 2003                                2-29       DRAFT-DO NOT QUOTE OR CITE

-------
 1          Secondary PM is formed by chemical reactions of free, adsorbed, or dissolved gases. Most
 2     secondary fine PM is formed from condensable vapors generated by chemical reactions of
 3     gas-phase precursors.  Secondary formation processes can result in either the formation of new
 4     particles or the addition of particulate material to pre-existing particles.  Most of the sulfate and
 5     nitrate and a portion of the organic compounds in atmospheric particles are formed by chemical
 6     reactions hi the atmosphere. Secondary aerosol formation depends on numerous factors
 7     including the concentrations of precursors; the concentrations of other gaseous reactive species
 8     such as ozone, hydroxyl radical, peroxy radicals, or hydrogen peroxide; atmospheric conditions
 9     including solar radiation and relative humidity (RH); and the interactions of precursors and
10     pre-existing particles within cloud or fog droplets or in the liquid film on solid particles. As a
11     result, it is considerably more difficult to relate ambient concentrations of secondary species to
12     sources of precursor emissions than it is to identify the sources of primary particles.
13     A significant effort is currently being directed toward the identification and modeling of organic
14     products of photochemical smog including the conversion of gases to particulate matter. More
15     information of the transformation of precursor gases into secondary PM is given in Chapter 3,
16     Section 3.3.1, Chemistry of Secondary PM Formation.
17           Particle strong acidity is due almost entirely to H2SO4 orNH4HSO4. Thus, the acidity of
18     atmospheric particles depends on both the amount of SO2 that is oxidized to SO3 and
19     subsequently forms H2SO4 and the amount of ammonia available to react with the sulruric acid.
20     Nitric acid is more volatile than sulruric acid. Thus, if gas phase SO3 or sulruric acid or particles
21     containing H2SO4 or NH4HSO4 contact particles containing NH4NO3, nitric acid gas will be
22     released with the remaining ammonia contributing to further neutralization of the acid. Little
23     NH4NO3 is found in atmospheres containing significant particle strong acidity. However, as SO2
24      emissions are reduced to the point that there is more than enough ammonia to neutralize the
25      sulruric acid, NH4NO3 particles will begin to form. Thus, ammonia emissions and
26      concentrations relative to those of SO2 and H2SO4 are important in determining the strong acidity
27      in the atmosphere and the concentration of particulate NH4N03. Therefore, once SO2 emissions
28      have been reduced to the point that ammonia and sulfate are in balance to form (NH4)2SO4,
29      further reductions in SO2 will not result in an equivalent reduction in airborne PM because one
30      (NH4)2SO4 unit will be replaced by two NH4NO3 units.
31

        June 2003                                 2-30        DRAFT-DO NOT QUOTE OR CITE

-------
   1      2.1.3.3   Particle-Vapor Partitioning
   2           Several atmospheric aerosol species, such as ammonium nitrate and certain organic
   3      compounds, are semivolatile and are found in both gas and particle phases. A variety of
   4      thermodynamic models have been developed to predict the temperature and relative humidity
   5      dependence of the ammonium nitrate equilibria with gaseous nitric acid and ammonia.
   6      However, under some atmospheric conditions, such as cool, cold, or very clean air, the relative
   7      concentrations of the gas and solid phases are not accurately predicted by equilibrium
   8      considerations alone, and transport kinetics can be important.  The gas-particle distribution of
   9      semivolatile organic compounds depends on the equilibrium vapor pressure of the compound,
 10      total particle surface area, particle composition, atmospheric temperature, and relative humidity.
 11      Although it generally is assumed that the gas-particle partitioning of semivolatile organics is in
 12      equilibrium in the atmosphere, neither the equilibria nor the kinetics of redistribution are well
 13      understood. Diurnal temperature fluctuations cause gas-particle partitioning to be dynamic on  a
 14      time scale of a few hours and can cause semivolatile compounds to evaporate during the
 15      sampling process. The pressure drop across the filter can also contribute to the loss of
 16      semivolatile compounds. The dynamic changes in gas-particle partitioning caused by changes  in
 17      temperature, pressure, and gas-phase concentration, both in the atmosphere and after collection,
 18      cause serious sampling problems that are discussed in Section 2.2.3, Measurement of
 19      Semivolatile Particulate Matter.
 20
 21     Equilibria with Water Vapor
 22          As a result of the equilibrium of water vapor with liquid water in hygroscopic particles,
 23     many ambient particles contain liquid water (particle-bound water). Unless removed, this
 24     particle-bound water will be measured as a component of the particle mass. Particle-bound
 25     water is important in that it influences the size of the particles, and in turn, their light scattering
 26     properties and their aerodynamic properties, which are  important for deposition to surfaces, to
 27     airways following inhalation, and in sampling instrumentation. The aqueous solution provides a
 28     medium for reactions of dissolved gases including reactions that do not take place in the gas
29     phase. The aqueous solutions also may act as a carrier to convey soluble toxic species to the
30      gas-exchange regions of the respiratory system, including species that would be removed by
31      deposition in the upper airways if the particles had remained in the gas phase (Friedlander and

        June 2003                                 2-31        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Yeh, 1998; Kao and Friedlander, 1995; Wilson, 1995).  An extensive review of equilibrium with
 2      water vapor as it pertains to ambient aerosols was given in Chapter 3 of the 1996 PM AQCD
 3      (U.S. Environmental Protection, Agency, 1996a).
 4           The interaction of particles with water vapor may be described briefly as follows.
 5      As relative humidity increases, particles of crystalline soluble salts, such as (NH4)2SO4,
 6      NH4HSO4, or NH4NO3, undergo a phase transition to become aqueous solution particles.
 7      According to the phase rule, for particles consisting of a single component, this phase transition
 8      is abrupt, taking place at a relative humidity that corresponds to the vapor pressure of water
 9      above the saturated solution (the deliquescence point).  With a further increase in relative
10     humidity, the solution particle adds water (and the concentration of the solute decreases) so that
11      the vapor pressure of the solution is maintained equal to that of the surrounding relative
12     humidity; thus, the solution particle tends to follow the equilibrium growth curve. As relative
13     humidity decreases, the solution particle follows the equilibrium curve to the deliquescence
14     point. However, rather than crystallizing at the deliquescence relative humidity, the solute
15     remains dissolved in a supersaturated solution to considerably lower relative humidities.
16     Ultimately the solution particle abruptly loses its water vapor (efflorescence) and typically
17     returns to the initial crystalline form.
18           For particles consisting of more than one component, the solid to liquid transition will take
19     place over a range of relative humidities with an abrupt onset at the lowest deliquescence point
20     of the several components and with subsequent growth as crystalline material  in the particle
21     dissolves according to the phase diagram for the particular multicomponent system. Under such
22     circumstances, a single particle may undergo several more or less abrupt phase transitions until
23     the soluble material is fully dissolved. At decreasing relative humidity, such particles tend to
24     remain in solution to relative humidities well below the several deliquescence points.  In the case
25     of the sulfuric acid-ammonium sulfate-water system, the phase diagram is fairly well
26      understood. For particles of composition intermediate between NH4HSO4 and (NH4)2SO4, this
27      transition occurs in the range from 40% to below 10%, indicating that for certain compositions
28      the solution cannot be dried in the atmosphere. At low relative humidities, particles of this
29      composition would likely be present in the atmosphere as supersaturated solution droplets (liquid
30      particles) rather than as solid particles. Thus, they would exhibit hygroscopic rather than
31      deliquescent behavior during relative humidity cycles.

        June 2003                                2-32        DRAFT-DO NOT QUOTE  OR CITE

-------
  1          Other pure compounds, such as sulfuric acid, are hygroscopic (i.e., they form aqueous
  2     solutions at any relative humidity and maintain a solution vapor pressure over the entire range of
  3     relative humidity). Soluble organic compounds may also contribute to the hygroscopicity of the
  4     atmospheric aerosol (Saxena et al., 1995; Saxena and Hildeman, 1996), but the equilibria
  5     involving organic compounds and water vapor, and, especially for mixtures of salts, organic
  6     compounds, and water, are not so well understood. These equilibrium processes may cause an
  7     ambient particle to significantly increase its diameter at relative humidities above about 40%
  8     (Figure 2-13).  A particle can grow to five times its dry diameter as the RH approaches 100%
  9     (Figure 2-14).  The Federal Reference Methods, for filter measurements of PM25 and PM10 mass,
 10     require equilibration at a specified, low relative humidity after collection (for PM25, between
 11     30 and 40% RH with control of ±5%  RH [Code of Federal Regulations, 2001aJ).  This
 12     equilibration removes much of the particle-bound water and provides a relatively stable PM
 13     mass for gravimetric measurements (see Section 2.2 for details and references). Otherwise,
 14     particle mass would be a function of relative humidity, and the particle mass would be largely
 15     particle-bound water at higher relative humidities. However, some particle-bound water may be
 16     retained even after equilibration. Recent studies have  shown that significant amounts of particle-
 17     bound water are retained in particles collected on impaction surfaces even after equilibration and
 18     that the amount of retained particle-bound water increases with relative humidity during
 19     collection (Hitzenberger et al., 1997).
 20          The retention of particle-bound  water is a greater problem for continuous monitors that
 21      measure changes in mass collected on a filter over long sampling times. If particle-bound water
 22     is not removed, changes in relative humidity would cause changes in the mass of PM collected
 23      over previous hours or  days. These changes could be much greater than amount of PM mass
 24     added in one hour. Therefore, continuous monitoring techniques generally attempt to remove
 25      particle-bound water before measurement either by heating or dehumidification. However, other
 26      semivolatile materials (e.g., ammonium nitrate and organic compounds) that may be partially
27      lost during sampling or equilibration of an unheated filter are certainly lost when the collected
28      sample  is heated above ambient temperature. These changes in particle size with relative
29      humidity also mean that particle measurements such as surface area or volume, or composition
30      as a function of size, should be made at the same RH in order for the results are to be
31

        June 2003                                2-33        DRAFT-DO NOT QUOTE OR CITE

-------
                             2.0-
                          o
                          Q
                           O
                          *-«
                           (0
                          o:

                          I
                           o
                          O
                           k_
                           0)
                          -s
                           CO
                          Q
1.5-
                             1.0
       Hygroscopic Growth
        Curve for (H2SO«)
       ••-^
       (
                                     30
 Hysteresis Loop
 for (NHJ, SO,
	M	
                50       70
                 RH (%)
                                                                	 •?
                                   - 6
      Figure 2-13.  Particle growth curves showing fully reversible hygroscopic growth of
                   sulfuric acid (H2SO4) particles, deliquescent growth of ammonium sulfate
                   [(NH4)2 SO4] particles at the deliquescent point (A, about 80% relative
                   humidity [RH]), reversible hygroscopic growth of ammonium sulfate
                   solution droplets at RH greater than 80%, and hysteresis (the droplet
                   remains supersaturated as the RH decreases below 80%) until the
                   crystallization point (B, about 38% RH) is reached (adapted from National
                   Research Council, 1993 and Tang, 1980).

      Source: National Research Council (1993) adapted from Tang (1980).
1     comparable. These problems are addressed in more detail in Section 2.2, Measurement of

2     Particulate Matter.

3
4     2.1.3.4   Atmospheric Lifetimes and Removal Processes
5          The lifetimes of particles vary with size. Nuclei-mode particles rapidly grow into the
6     accumulation mode. However, the accumulation mode does not grow into the coarse mode.
7     Accumulation-mode fine particles are kept suspended by normal air motions and have very low
8     deposition rates to surfaces. They can be transported thousands of km and remain in the
9     atmosphere for a number of days. Coarse particles can settle rapidly from the atmosphere within
       June 2003
                  2-34
                DRAFT-DO NOT QUOTE OR CITE

-------
                                                                           216
                                 ooooo Experimental Measurements
                                        50
                                     NH4HSO4  Dry Particle Diameter (nm)
        Figure 2-14.  Theoretical predictions and experimental measurements of growth of
                      NH4HSO4 particles at relative humidity between 95 and 100%.
        Source: Li et al. (1992).
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
 hours and normally travel only short distances.  However, when mixed high into the atmosphere,
 as in dust storms, the smaller-sized coarse-mode particles have longer lives and travel greater
 distances. Dry deposition rates are expressed in terms of a deposition velocity that varies with
 particle size, reaching a minimum between 0.1 and 1.0 \im aerodynamic diameter (e.g., Lin
 et al., 1994). Accumulation-mode particles are removed from the atmosphere primarily by cloud
 processes. Fine particles, especially particles with a hygroscopic component, grow as the
 relative humidity increases, serve as cloud condensation nuclei, and grow into cloud droplets.
 If the cloud droplets grow large enough to form rain, the particles are removed in the rain.
 Falling rain drops impact coarse particles and remove them. Ultrafme or nuclei-mode particles
 are small enough to diffuse to the falling drop, be captured, and be removed in rain. Falling rain
 drops, however, are not nearly as effective in removing accumulation-mode particles as the
cloud processes mentioned above.  A more detailed discussion of particle deposition, including
acid deposition, especially as it applies to deposition to vegetation, soil, and water surfaces is
given in Chapter  4 (Environmental Effects of Airborne Particulate Matter). Acid deposition and
       June 2003
                                         2-35
DRAFT-DO NOT QUOTE OR CITE

-------
 1      PM are intimately related, first, because particles contribute to the acidification of rain and,
 2      secondly, because the gas-phase species that lead to dry deposition of acidity are also precursors
 3      of particles. Therefore, reductions in SO2 and NOX emissions will decrease both acidic
 4      deposition and PM concentrations.
 5           Sulfate, nitrate, and some partially oxidized organic compounds are hygroscopic and act as
 6      nuclei  for the formation of cloud droplets. These droplets serve as chemical reactors in which
 7      (even slightly) soluble gases can dissolve and react. Thus, SO2 can dissolve in cloud droplets
 8      and be oxidized to sulfuric acid by dissolved  ozone or hydrogen peroxide.  These reactions take
 9      place only in aqueous solution, not in the gas phase. Sulfur dioxide also may be oxidized by
10      dissolved oxygen.  This process will be faster if metal catalysts such as iron or manganese are
11      present in solution. If the droplets evaporate, larger particles are  left behind. If the droplets
12     grow large enough, they will fall as rain; and the particles will be removed from the atmosphere
13     with potential effects on the materials, plants, or soil on which the rain falls. (Similar
14     considerations apply to dew.) Atmospheric particles that nucleate cloud droplets also may
15     contain other soluble or nonsoluble materials such as metal salts  and organic compounds that
16     may add to the toxicity of the rain.  Sulfuric acid, ammonium nitrate, ammonium sulfates, and
17     organic particles also are deposited on surfaces by dry deposition.  The utilization of ammonium
18     by plants leads to the production of acidity. Therefore, dry deposition of particles can also
19     contribute to the ecological impacts of acid deposition. These effects are discussed in Chapter 4
20     (Environmental Effects of Airborne Particulate Matter).
21
22     2.1.4   Comparison of Fine and Coarse Particles
23           The physical and chemical properties of fine particles (including ultrafine particles and
24      accumulation-mode particles) and coarse particles are summarized for comparison purposes in
25      Table 2-1. These include important differences in sources, formation mechanisms, composition,,
26      atmospheric residence time, removal processes, and travel distances. Ensuing chapters in this
27      document will also show that fine and coarse particles differ in aspects of concentrations,
28      exposure, dosimetry, toxicology, and epidemiology.  Collectively, these differences continue to
29      warrant consideration of fine particles as a separate air pollutant class from coarse particles and
 30      the setting of separate standards for fine and coarse particles.
 31

        June 2003                                 2-36        DRAFT-DO NOT QUOTE OR CITE

-------
                   TABLE 2-1. COMPARISON OF AMBIENT PARTICLES,
    FINE PARTICLES (Ultrafme Plus Accumulation-Mode) AND COARSE PARTICLES
                                      Fine
                      Ultrafine
                               Accumulation
                                                                                Coarse
  Formation
  Processes:
             Combustion, high-temperature
          processes, and atmospheric reactions
  Formed by:   Nucleation
             - Condensation
               Coagulation
  Composed
  of:
 Solubility:
 Sources:
 Atmospheric
 half-life:

 Removal
 Processes:
 Travel
 distance:
 Sulfate
 Elemental Carbon
 Metal compounds
 Organic compounds
 with very low
 saturation vapor
 pressure at ambient
 temperature
 Probably less soluble
 than accumulation
 mode  .

 Combustion
 Atmospheric
 transformation of SO2
 and some organic
 compounds
 High temperature
 processes
Minutes to hours


Grows into
accumulation mode
Diffuses to raindrops

<1 to 10s of km
 Condensation
 Coagulation
 Reactions of gases in or on
 particles
 Reactions of gases in or on
 particles
 Evaporation of fog and cloud
 droplets in which gases have
 dissolved and reacted

 Sulfate, Nitrate, Ammonium,
 and Hydrogen ions
 Elemental carbon
 Large variety of organic
 compounds
 Metals: compounds of Pb, Cd,
 V, Ni, Cu, Zn, Mn, Fe, etc.
 Particle-bound water
 Largely soluble, hygroscopic,
 and deliquescent


 Combustion of coal, oil,
 gasoline, diesel fuel, wood
 Atmospheric transformation
 products of NOxl SO2, and
 organic compounds, including
 biogenic organic species
 (e.g., terpenes)
 High-temperature processes,
 smelters, steel mills, etc.

 Days to weeks
Forms cloud droplets and rains
out
Dry deposition

100s to 1000s of km
 Break-up of large solids/droplets


 Mechanical disruption (crushing,
 grinding, abrasion of surfaces)
 Evaporation of sprays
 Suspension of dusts
 Reactions of gases in or on particles
 Suspended soil or street dust
 Fly ash from uncontrolled combustion
 of coal, oil, and wood
 Nitrates/chlorides from HNCyHCl
 Oxides of crustal elements
 (Si, Al, Ti, Fe)
 CaCOj, NaCI, sea salt
 Pollen, mold, fungal spores
 Plant and animal fragments
 Tire, brake pad, and road wear debris

 Largely insoluble and nonhygroscopic
Resuspension of industrial dust and
soil tracked onto roads and streets
Suspension from disturbed soil (e.g.,
farming, mining, unpaved roads)
Construction and demolition
Uncontrolled coal and oil combustion
Ocean spray
Biological sources


Minutes to hours
Dry deposition by fallout
Scavenging by falling rain drops


<1 to 10s of km
(100s to 1000s in dust storms)
 Source: Adapted from Wilson and Suh (1997).
June 2003
                                 2-37
                        DRAFT-DO NOT QUOTE OR CITE

-------
 1      2.2  MEASUREMENT OF PARTICULATE MATTER
 2           The 1996 PM AQCD (U..S. Environmental Protection Agency, 1996a) summarized
 3      sampling and analytical techniques for PM and acid deposition that had appeared in the literature
 4      since the earlier 1982 PM AQCD (U.S. Environmental Protection Agency, 1982). Excellent
 5      reviews have also been published by Chow (1995) and McMurry (2000). This section discusses
 6      problems in measuring PM; new techniques that attempt to alleviate these problems or measure
 7      problem species; Federal Reference Methods, speciation monitors, analytical methods for
 8      inorganic elements, organic and elemental carbon, and ionic species; and continuous and
 9      multiday monitors.
10
11      2.2.1    Particle Measurements of Interest
12          There are many PM components and parameters that are of interest across the various
13      types of uses to which PM measurement data are applied.  These uses include analyses of
14     compliance with air quality standards and trends; source category apportionment studies related
15     to the develop of pollution reduction strategies and the validation of air quality models; studies
16     related to health, ecological, and radiative effects; and characterization of current air quality for
17     presentation to the public in the context of EPA's Air Quality Index. PM measurement
18     components and parameters of specific interest for these various purposes are noted below and
19     summarized in Table 2-2.
20          Particle measurements are needed to determine if a location is in compliance with air
21     quality  standards, to determine long-term.trends in air quality patterns, and for epidemiologic
22     studies. For these purposes, precision of the measurements by a variety of measurement
23     instruments in use is a critical consideration. Therefore, intercomparisons of various samplers
24     under a variety of atmospheric and air quality conditions are important.
25          In order to reduce pollution to attain a standard, pollution control agencies and national
26     research organizations need measurements to identify source categories and to develop and
27     validate air quality models. For these purposes, PM parameters other than mass, such as
28      chemical composition and size distribution, must also be measured. Moreover, measurements
29      are needed with shorter time resolution in order to match changes in pollution with diurnal
30      changes in the boundary layer.
31

        June 2003                                2-38        DRAFT-DO NOT QUOTE OR CITE

-------
             TABLE 2-2. PARTICULATE MATTER COMPONENTS/PARAMETERS OF
               INTEREST FOR HEALTH, ECOLOGICAL, OR RADIATIVE EFFECTS;
                     FOR SOURCE CATEGORY APPORTIONMENT STUDIES;
                   OR FOR AIR QUALITY MODELING EVALUATION STUDIES
        • Particle number
        • Particle surface area
        • Particle size distribution
        • PM mass (fine PM mass [PM25] and coarse thoracic PM mass [PM10_25]) including both
          nonvolatile mass as measured by the current Federal Reference method and total mass
          (including semivolatile components such as ammonium nitrate and semivolatile organic
          compounds, but not particle-bound water)
        • Ions (sulfate, nitrate, and ammonium)
        • Strong acidity (H+)
        • Elemental carbon
        • Organic carbon (total, nonvolatile, and semivolatile; functional groups and individual
          species)
        • Transition metals (water soluble, bioavailable, oxidant generation)
        • Specific toxic elements and organic compounds
        *  Crustal elements
        •  Bioaerosols
        •  Particle refractive index (real and imaginary)
        •  Particle density
        *  Particle size change with changes in relative humidity
 1           A number of PM measurements are needed for use in epidemiologic and exposure studies
2      and to determine components of PM to guide planning and interpretation of toxicologic studies.
3      For these purposes, size and chemical composition measurements are important, as is
4      measurement across different time intervals.  For epidemiologic studies of acute (i.e., short-term)
5      PM exposures, 1-h or continuous measurements can provide important information beyond that
6      provided by 24-h measurements. However, for epidemiologic studies of chronic PM exposures,
7      measurements that permit integration over longer intervals (e.g., a week to a month) are more
8      relevant. For dosimetric studies and modeling, information will be needed on the particle size
       June 2003                               2-39       DRAFT-DO NOT QUOTE OR CITE

-------
 1      distribution and on the behavior of particles as the relative humidity and temperature changes
 2      found in the atmosphere are increased to those found in the respiratory system.
 3           For studies of ecological effects and materials damage, measurements of particles and of
 4      the chemical components of particulate matter in rain, fog, and dew are needed to understand the
 5      contributions of PM to soiling of surfaces and damage to materials and to understand the wet and
 6      dry deposition of acidity and toxic substances to surface water, soil, and plants.  Some
 7      differentiation into particle size is needed to determine dry deposition.
 8           For studies of visibility impairment and radiative effects, information is needed that relates
 9      to how particles scatter and absorb light, including refractive index, ratio of scattering to
10      absorption, size distribution, and change in particle size with change in relative humidity.
11
12      2.2.2    Issues in Measurement of Particulate Matter
13           The EPA decision to revise the PM standards by adding  daily and yearly standards for
14     PM2 5 has led to a renewed interest in the measurement of atmospheric particles and also to a
15     better understanding of the problems in developing precise and accurate measurements of
16     particles. It is very difficult to measure and characterize particles suspended in the atmosphere;
17     however, numerous improvements in PM monitoring are in use and others are in development.
18     EPA's PM standards are based, in part, on epidemiologic relationships between health effects
19     and PM concentrations as measured with existing monitoring  methods. As understanding of
20     suspended PM has advanced and new monitoring information has become available, EPA has
21     changed the indicator for the PM NAAQS from TSP to PMIO, and added PM2 5. During the
22     current PM NAAQS review, consideration will be given to  a standard for coarse thoracic PM.
23           The U.S. Federal Reference Methods (FRM) for PM25 and PM10 provide relatively precise
24     (±10 %) methods for determining the mass of material remaining on a Teflon filter after
25     equilibration.  However, numerous uncertainties remain as to  the relationship between the mass
26     and composition of material remaining on the filter as determined by  the FRM measurement
27     procedure and the mass and composition  of material that existed in the atmosphere as suspended
28     PM. As a result, EPA defines accuracy for PM measurements in terms of agreement of a
29      candidate sampler with a reference sampler. Therefore, intercomparison of samplers is very
30      important in determining how well various samplers agree and how various design choices
31      influence what is actually measured.

        June 2003                                 2-40        DRAFT-DO NOT QUOTE OR CITE

-------
  1           There are eight general areas where choices are made in the design and use of an aerosol
  2      sampler. These include (1) consideration of positive artifacts due to chemical reaction or
  3      adsorption; (2) treatment of semivolatile components; (3) selection of particle size cut
  4      characteristics for the upper cut point; (4) separation of fine and coarse PM; (5) treatment of
  5      pressure, temperature, and relative humidity; (6) time resolution; (7) assessment of the reliability
  6      of the measurement technique; and (8) operation and maintenance procedures needed to
  7      maintain consistent measurements over time.  In many cases, choices have been made without
  8      adequate recognition of the consequences. As a result, measurement methods developed by
  9      different organizations may give different results when sampling the same atmosphere even
 10      though the techniques appear to be similar.
 11
 12      2.2.2.1   Artifacts Due to Chemical Reactions
 13           When TSP was collected on glass fiber filters, the reaction of SO2 (and other acid gases)
 14      with basic sites on the glass fiber or with basic coarse particles on the filter led to the formation
 15      of sulfate (or other nonvolatile sales, e.g., nitrate, chloride). These positive artifacts led to
 16      overestimation of mass, sulfate (and probably also of nitrate). The metal impurities in the glass
 17      fiber caused a high background that led to low precision in the measurement of trace metals.
 18      These problems were largely overcome by changing to quartz fiber or Teflon filters and by
 19      separate collection of PM25.  However, the possible reaction of acidic gases with basic coarse
20      particles remains a possibility, especially with measurements of PM10 and PM10_25.  The reaction
21      of NH3 with acidic particles, either during sampling or during transportation, storage, and
22      equilibration remains a problem in areas such as the eastern U.S. where PM is frequently acidic.
23      Techniques have been developed that overcome this problem by use of a denuder to remove NH3
24      during sampling and to protect the collected PM from NH3 (Suh et al, 1992, 1994; Brauer et al.,
25      1991; Koutrakis et al., 1988a,b). However, this  technique has been applied primarily for
26      measurement of particle strong acidity, not for the measurement of artifact-free ammonium or
27      mass concentrations.  In the measurement of particle strong acidity, basic coarse particles must
28      be separated from acidic fine particles (Koutrakis et al., 1992).
29
30
31

        June 2003                                2-41        DRAFT-DO NOT QUOTE OR  CITE

-------
2.2.2.2  Treatment of Semivolatile Components of Participate Matter
     Current filtration-based mass measurements can experience significant evaporative losses,
during and possibly after collection, of a variety of semivolatile components (i.e., species that
exist in the atmosphere in dynamic equilibrium between the condensed phase and gas phase).
Important examples include ammonium nitrate, semivolatile organic compounds, and particle-
bound water. This problem is illustrated in Figure 2-15.
                    a.
                   Q
                    O)
                    o
                    at
                    
-------
  1        1.   Collect/measure all components present in the atmosphere in the condensed phase except
               particle-bound water. (Examples: Brigham Young absorptive sampler and Harvard
               pressure drop monitor. Both require preconcentration of the accumulation mode and
               reduction of ambient humidity.)
  2
  3        2.   Stabilize PM at a specified temperature high enough to remove all, or almost all, particle-
               bound water.  This results in loss of much of the semivolatilePM.  (Examples: tapered
               element oscillating microbalance (TEOM) operated at 50 °C, beta gauge with heated
               inlet.)
  4
  5        3.   Equilibrate collected material at fixed, near-room temperature and moderate relative
               humidity to reduce particle-bound water. Accept the loss of an unknown but possibly
               significant fraction of semivolatile PM.  (Example:  U.S. Federal Reference Method and
               most filter-weighing techniques.) Equilibration originally was designed to remove
               adsorbed water vapor from glass fiber filters in order to maintain a stable filter weight.
               The designated RH (35±5%) was a compromise.  If the RH is too low, electrostatic
               charging becomes a problem. The equilibration process does help provide a stable and
               reproducible mass. It also reduces the particle-bound water. However, it may not
               remove all particle-bound water.
  6
  7      2.2.2.3  Upper Cut Point
  8           The upper cut point of the high volume sampler varied with wind speed and direction.
  9      Newer PM samplers are usually designed to have an upper cut point and its standard deviation
10      that are independent of wind direction and relatively independent of wind speed. Current PM
11      samplers have upper cut points that are  stable under normal operating conditions. However,
12      problems may occur under unusual or adverse conditions. Ono et al. (2000) reported the results
13      of a study in which several PMIO samplers were collocated and operated at various sites at
14      Owens Lake, CA, a location with high concentrations of coarse PM. Samplers included the
15      Partisol sampler, the TEOM, a dichotomous sampler, the Wedding high-volume sampler, and the
16      Graseby high-volume sampler.  They found that the TEOM and Partisol samplers agreed to
17      within 6% on average. The dichotomous sampler and the Graseby and Wedding high-volume

       June 2003                                2-43       DRAFT-DO NOT QUOTE OR CITE

-------
 1      samplers, however, measured significantly lower PM10 concentrations than the TEOM (on
 2      average 10, 25, and 35% lower, respectively). These lower concentrations were attributed to a
 3      decrease in the cut point at higher wind speeds and to a dirty inlet.
 4            The choice of the cut point characteristics depends upon the application for the sampling
 5      device. A  separation that simulates the removal of particles by the human upper respiratory
 6      system might appear to be a good choice for both health risk and regulatory monitoring (i.e.,
 7      measure what gets into the lungs). The ACGIH-ISO-CEN penetration curve for thoracic
 8      particles (particles able to pass the larynx and penetrate into the bronchial and alveolar regions of
 9     the lung) has a 50% cut point at 10 urn aerodynamic diameter (DJ. The U.S. PM10 separation
10     curve is sharper than the thoracic penetration curve but has the advantage of reducing the
11      problem of maintaining the finite collection efficiency specified by the thoracic penetration
12     curve for particles larger than 10 urn Da. (See Section 2.1.2.2 and Figure 2-6).
13
14     2.2.2.4  Cut Point for Separation of Fine and Coarse Particulate Matter
15           As shown in Table 2-1, fine and coarse particles differ not only in size but also in
16     formation  mechanisms; sources; and chemical, physical, and biological properties.  They also
17     differ in concentration-exposure relationships, dosimetry (deposition and retention in the
18     respiratory system), toxicity, and health effects as observed by epidemiologic studies. Thus, it is
19     desirable to measure fine and coarse PM separately in order to properly allocate health effects to
20      either fine or coarse PM and to correctly determine sources by receptor modeling approaches.
21      For example, sulfates in fine particles are associated with hydrogen or ammonium ions while
22      sulfates in coarse particles are associated with basic metal ions.  Transition metals in coarse
23      particles are likely to be associated with soil and tend to be less soluble (and presumably less
24      bioavailable) than transition metals  in fresh combustion particles found in fine particles.
25           The  2.5 urn Dacut point was chosen in the early 1970s as the cut point for a new
26      dichotomous sampler (Loo et al., 1976; Jaklevic et al., 1977) for use in the Regional Air
27      Pollution Study in St. Louis, MO. At that time aerosol scientists were beginning to realize that
28      there was  a minimum between 1 and 3 urn in the distribution of particle size by volume (Whitby
29      et al., 1974). The 2.5 urn cut point was subsequently used as an indicator of fine-mode PM in a
 30      number of studies including the Harvard Six-City Studies of the relationships between mortality
 31      and PM concentrations (Dockery et al., 1993; Schwartz et al., 1996).  A2.5 urn cut point was

        June 2003                                2-44        DRAFT-DO NOT QUOTE OR CITE

-------
   1      also used in the Inhalable Particle Network (Suggs and Burton, 1983) which provided data for
   2      another major epidemiologic study of PM - mortality relationships using an American Cancer
   3      Society cohort (Pope et al., 1995). Therefore, at the time of the last review of the NAAQS for
   4      PM (U.S. Environmental Protection Agency, 1996a,b), there were a number of epidemiologic
   5      studies demonstrating a statistical relationship between PM2 5 concentrations and mortality.
   6           It is now understood that the size range between 1.0 and 2.5 urn, sometimes called the
   7      intermodal region may contain either accumulation-mode or coarse-mode material or both, i.e.,
   8      the two modes may overlap in this region (Kegler et al., 2001). The experimental information on
   9      the composition and source of the intermodal mass was discussed extensively in the 1996 PM
 10      AQCD (U.S. Environmental Protection Agency, 1996a).  Depending on conditions, a significant
 11      amount of either accumulation- or coarse-mode material may be found in the intermodal region
 12      between 1.0 and 2.5 urn. The analysis demonstrated the important role of relative humidity in
 13      influencing the size of particles in both the accumulation and coarse modes.
 14          As the relative humidity increases, hygroscopic accumulation-mode particles will increase
 15      in size due to accumulation of particle-bound water. At high relative humidities, some originally
 16      submicrometer accumulation-mode PM may be found with a Da above 1  urn. At a relative
 17      humidity of 100%, such as found in fog and clouds, accumulation-mode PM may extend above
 18      2.5 ^m Da. What is not well understood is whether such particles will shrink to diameters below
 19      1 ^irn as the RH decreases or whether reactions occurring in the wet particles will result in an
 20      increase in non-aqueous mass so that even at low RH the diameters would exceed 1 urn. On the
 21     other hand, at very low relative humidity, coarse particles may be fragmented into smaller sizes,
 22     and small amounts of coarse PM may be found with an Da below 2.5 urn (Lundgren et al., 1984;
 23     Lundgren and Burton, 1995).  Thus, a PM25 sample will contain all of the accumulation-mode
 24     PM except during periods of RH near 100 %.  However, under conditions of low RH, it may also
 25     contain a small fraction of the coarse PM.  The selection of a cut point of 2.5 urn as a basis for
 26     EPA's  1997 NAAQS for fine particles (Federal Register, 1997) and its continued use in many
 27     health effects studies reflect the importance placed on more complete inclusion of accumulation-
 28     mode particles while recognizing that intrusion of coarse particles can occur under some
29     conditions with this cut point.
30          In addition to the influence of relative humidity, in areas where winds  cause high
31      concentrations of windblown soil there is evidence that a significant amount of coarse-mode PM

        June 2003                                2-45        DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
may be found below 2.5 urn. An example, taken from data collected during the August 1996
dust storm in Spokane, WA, is shown in Figure 2-16. Note that the PM10 scale is 10 times that
of the other size fractions.  PMb although high in the morning, goes down as the wind increases
and PM10, PM25, and PM2.5_, go up. During the peak of the dust storm, around 9 pm, PM^.! was
88% of PM2^. For the 24-h period, PM25_, was 54% of PM25.  However, PM, was not affected
by the intrusion of coarse particles. Similar considerations probably apply to short-term
intrusions of dust transported from distant sources such as the Sahara and Gobi deserts (Husar
etal.,2001).
                                                                                    600
                                                                                    500
       51  10
                                                                                    400
                                                                                    300
                                                                                    200
                                                                                    100
                                                                                  ,5
                                                                                   o
                                                                                  E~
                                                                                  Q_
              12am 2am  4am  6am  Sam 10am 12pm 2pm  4pm  6pm  8pm  10pm
                                   Local Time, August 30,1996

       Figure 2-16.  Particulate matter concentrations in Spokane, WA, during the August 30,
                    1996 dust storm.
       Source: Claiborn et al. (2000).
  1          A cut point of 1.0 ^im could reduce the misclassification of coarse-mode material as fine,
  2      especially in areas with high levels of wind blown soil, but under high RH conditions could
  3      result in some accumulation-mode material being misclassified as coarse. A reduction in RH,
        June 2003
                                         2-46
DRAFT-DO NOT QUOTE OR CITE

-------
   1     either intentionally or inadvertently, will reduce the mass mean diameter of the accumulation-
   2     mode particles.  A sufficient reduction in RH should yield a dry accumulation mode with very
   3     little material above 1.0 urn. Studies of the changes in particle size with changes in relative
   4     humidity suggest that only a small fraction of accumulation-mode particles will be above 1.0 urn
   5     in diameter at RH below 60%, but a substantial fraction will grow above 1.0  um for RH above
   6     80% (Hitzenberger et al., 1997; McMurry and Stolzenburg, 1989; U.S. Environmental Protection
   7     Agency, 1996a).
   8          Under high relative humidity circumstances, a monitor using a 1.0 um Da cut point can
   9     achieve better modal separation if the air stream is dehumidified to some fixed humidity that
 10     would remove all or most particle-bound water without evaporating semivolatile components.
 11      New techniques which require reduction of RH prior to collection have been  developed for
 12     measurement of fine paniculate matter minus particle-bound water but including semivolatile
 13      nitrate and organic compounds. With such techniques, PM, measurements in conjunction with
 14      concurrent PM2 5 measurements would be useful  for exposure, epidemiologic, and source
 15      apportionment studies especially in areas where intrusion of coarse-mode particles into the
 16      intermodal range is likely.
 17
 18     2.2.2.5   Treatment of Pressure, Temperature, and Relative Humidity
 19           There are a variety of techniques for defining (or ignoring) the pressure, temperature, and
 20     relative humidity during and after sampling.  For example, the sample volume may be based on
 21     the mass or volumetric flow corrected to standard temperature and pressure (273 °K and 1 atm.)
 22     (current FRM for PM10), or it may be based on the volumetric flow at ambient conditions of
 23     temperature and pressure (current FRM for PM2 5).
 24          There are also a variety of options for the control of temperature during collection.  The
 25     particles may be heated enough to remove much of the particle-bound water (i.e., TEOM at
 26     50 °C); the particles may be heated several degrees, just enough to prevent condensation of
 27     water in the sampling system; the particles and the sampler may be maintained near ambient
 28      temperature (±5 °C of ambient temperature is required for EPA FRM samplers); or the particles
29      and sampler may be maintained at constant temperature inside a heated or air conditioned
30      shelter. There are also options for control of temperature after collection: (a) no control (room
       June 2003                                2-47        DRAFT-DO NOT QUOTE OR CITE

-------
 1      temperature) or (b) ship and store at cool temperature (4 °C is the current EPA FRM
 2      requirement).
 3           Consideration must also be given to relative humidity. Changes in relative humidity cause
 4      changes in particle size of hygroscopic or deliquescent particles.  Changing relative humidity by
 5      adding or removing water vapor affects measurements of particle number, particle surface area,
 6      and particle size distribution and the amount of overlap of accumulation-mode and coarse-mode
 7      particles.  Changing relative humidity by intentional or inadvertent changes in temperature also
 8      affects the amount of loss of ammonium nitrate and semivolatile organic compounds.
 9      Monitoring personnel should be aware of the various options for treatment of pressure,
10      temperature, and relative humidity; make appropriate selections; and document which options
11      are used.
12          Studies of relationships between personal/indoor/outdoor measurements present special
13      problems.  Indoor environments are typically dryer than outdoors and may be warmer or, if
14     air-conditioned, cooler.  These differences may change particle size and the amount of
15     volatilization of semivolatile components. Such changes between indoors and outdoors will
16     complicate the comparison of indoor to outdoor PM concentrations; the modeling of personal
17     exposure to all particles; and apportionment of personal exposure into particles of ambient
18     origin, particles of indoor origin, and particles originating from personal activity.
19
20     2.2.2.6  Time Resolution
21           The classic 24-hour filter collection technique is being supplemented by a variety of
22      continuous monitors for various PM constituents. This process is being accelerated by the lower
23      operational cost of continuous monitors and the availability of new  continuous monitors for
24      mass, number, and certain chemical components, as well as refinements of older methods based
25      on beta attenuation or light scattering. Most epidemiologic studies have used 24-hour
26      concentrations as exposure indicators. However, one epidemiologic study of chronic effects
27      uses a filter sampler with a two-week collection period (Gauderman et al., 2000).  Another
 28      recent study used 1-2 h concentrations (see Peters et al., 2000). Continuous methods are
 29      discussed in Section 2.2.5.
 30
 31

        June 2003                                2-48        DRAFT-DO NOT QUOTE OR CITE

-------
  1     2.2.2.7  Accuracy and Precision
  2           Precision is typically determined by comparison of collocated samplers or through
  3     replicate analyses; whereas accuracy is determined through the use of traceable calibration
  4     standards. Unfortunately, no standard reference calibration material or procedure has been
  5     developed for suspended, atmospheric PM. It is possible to determine the accuracy of certain
  6     components of the PM measurement system (e.g., flow control, inlet aspiration, PM2 5 cut,
  7     weighing, etc.).  The absolute accuracy for collecting a test aerosol can also be determined by
  8     isokinetic sampling in a wind tunnel. However, it is not currently feasible to provide a simulated
  9     atmospheric aerosol with naturally occurring semivolatile components. It is particularly
 10     challenging to develop an atmospheric aerosol calibration standard suitable for testing samplers
 11     in the field.  Therefore, it is not possible at the present time to establish the absolute accuracy of
 12     a PM monitoring technique. Intercomparison studies may be used to establish the precision of
 1-3     identical monitors and the extent of agreement between different types of monitors. Such studies
 14     are important for establishing the reliability of PM measurements. Intercomparison studies have
 15     contributed greatly to our understanding of the problems in PM measurement.  Such studies will
 16     be discussed as they apply to specific measurement problems, monitoring instruments, or
 17     analytical techniques.
 18          Some measurement errors of concern in PM]0 sampling, including those that arise due to
 19     uncertainty tolerances in cutpoint, particle bounce and reentrainment, impactor surface
 20     overloading, and losses to sampler internal surfaces, were discussed in detail in the 1996 PM
 21      AQCD (U.S. Environmental Protection Agency, 1996a). Other measurement errors of concern
 22     in PM2 5 sampling arise because of our inability to assess accuracy in an absolute sense due to a
 23     lack of an atmospheric aerosol calibration standard, the inclusion in PM25 of a  small amount of
 24     coarse particles as discussed in Section 2.2.1.3, problems associated with the definition of PM25
 25      as what remains on a filter after collection and equilibration rather than as the mass of particles
 26     as they exist in the air.  Still, it is possible to measure PM indicators with high precision.
 27      Detailed information on precision and quality assurance may be found on EPA's Technology
 28      Transfer Network website (U.S. Environmental Protection Agency, 2002a). See discussion in
29      Section 2.2.4.
30           Because of the difficulties associated with determining the accuracy of PM measurements,
31      EPA has sought to make FRM measurements equivalent by specifying operating conditions and,

        June 2003                                 2-49       DRAFT-DO NOT QUOTE OR CITE

-------
 1      in the case of PM25 samplers, by specifying details of the sampler design. Thus, both the PM10
 2      as well as the PM2 5 standards are defined with consistency of measurement technique rather than
 3      with the accuracy of the true mass concentration measurement in mind (McMurry, 2000). It is
 4      acknowledged in the Federal Register (1997) that, "because the size and volatility of the particles
 5      making up ambient particulate matter vary over a wide range and the mass concentration of
 6      particles varies with particle size, it is difficult to define the accuracy of PM2 5 measurements in
 7      an absolute sense...." Thus, accuracy is defined as the degree of agreement between a field
 8      PM2 5 sampler and a collocated PM2 5 reference method audit sampler (McMurry, 2000). The
 9      Federal Reference Method for PM2.5 is discussed in Section 2.2.3.3. As mentioned earlier,
10      volatilization of organic compounds and ammonium nitrate during sampling or post-sampling
11      handling can lead to significant underestimation of the fine particulate mass concentration in
12      some locations. Sources of error in the measurement of mass of PM25 suspended in the
13      atmosphere also arise because of adsorption or desorption of semivolatile vapors onto or from
14     collected PM, filter media, or other sampler surfaces; neutralization of acid or basic vapors on
15      either filter media or collected PM; and artifacts associated with particle-bound water.
16          During the past 25 years, there have been advancements in the generation and
17     classification of monodisperse aerosols as well as in the development of electron microscopy and
18     imaging analysis that have contributed to the advancement in aerosol calibration (Chen, 1993).
19     Still, one of the limitations in PM sampling  and analysis remains the lack of primary calibration
20     standards for evaluating analytical methods  and for intercomparing laboratories. Klouda et al.
21     (1996) examined the possibility of resuspending the National Institute of Science and
22     Technology (NIST) Standard Reference Material 1649 (Urban Dust) in air for collection on up to
23     320 filters simultaneously using Standard Research International's dust generation and collection
24     system. However, the fine component is not resuspended and the semivolatile  component has
25     evaporated. Consequently, this material is not a suitable standard for suspended PM.  NIST  is
26     continuing work in this area with EPA support.
27          Methods validation was discussed in the 1996 PM AQCD (U.S. Environmental Protection
28     Agency, 1996a), and the usefulness of intercomparisons and "internal redundancy" was
29     emphasized. For example, a number of internal consistency checks are applied to the IMPROVE
30     network (Malm et al., 1994). These include mass balances, sulfur measurements by both
31     proton-induced X-ray emission (PIXE) and ion chromatography (1C), and comparison of organic

       June 2003                                2-50        DRAFT-DO NOT  QUOTE OR CITE

-------
   1      matter by combustion and by proton elastic scattering (PESA) of hydrogen. Mass balances
  2      compare the gravimetrically determined mass with the mass calculated from the sum of the
  3      major chemical components (i.e., crustal elements plus associated oxygen, organic carbon,
  4      elemental carbon, sulfate, nitrate, ammonium, and hydrogen ions). Mass balances are useful
  5      validation techniques; however, they do not check for, or account for, artifacts associated with
  6      the absorption of gases during sampling or the loss of semivolatile material during sampling.
  7      The mass balance check may appear reasonable even if such artifacts are present because only
  8      the material collected on the filter is included in the balance.  Mass balance checks may also
  9      suffer from errors due to some particle-bound water remaining in the PM even after equilibration
 10      and from the use of an arbitrary factor to account for the amount of oxygen and hydrogen atoms
 11      per carbon atom in the organic carbon fractions (Andrews et al., 2000).
 12
 13      2.2.3   Measurement of Semivolatile Participate Matter
 14           PM suspended in the atmosphere is composed of many chemical species having a wide
 15      range of vapor pressures.  Substances with vapor pressures below 10'9 Torr (mmHg) will
 16      typically be nonvolatile; substances with vapor pressures above 10'1 Torr will be primarily in the
 17      gas phase; substances with vapor pressures between 10'1  and 10'9 Torr may exist in an
 18      equilibrium state with some material in both the gas phase and the condensed (particulate) phase
 19     and are known semivolatile material (SVM) (Pankow, 1994a). SVM, originally in the
 20     atmosphere in the particulate phase and collected on a filter, may subsequently be lost from the
 21      filter.  SVM may evaporate during sampling due to a reduction in its concentration in the
 22     atmosphere being sampled or due to the pressure drop across the filter.  SVM may evaporate
 23      after sampling; during intentional equilibration at a low relative humidity; or during transport,
 24      handling, and storage if exposed to an atmosphere in which the vapor pressure of one or more
 25      semivolatile components is lower than in the atmosphere sampled.  Since water is not a
 26      pollutant, it is necessary to remove most of the particle-bound water before  weighing (Chow,
 27      1995). However, collection and measurement of ammonium nitrate and semivolatile organic
 28      compounds in suspended atmospheric PM represents a major analytical challenge (McMurry,
29      2000).
30
31

       June 2003                                2-51        DRAFT-DO NOT QUOTE OR CITE

-------
 1      2.2.3.1  Particle-Bound Water
 2           It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
 3      compounds. However, for many measurements of suspended particle mass, it is desirable to
 4      remove the particle-bound water before determining the mass. The mass of particle-bound water
 5      is strongly dependent on the relative humidity. However, the dependence is not linear since
 6      there is significant hysteresis in the water adsorptiori-desorption pathways (Seinfeld and Pandis,
 7      1998). Water vapor is not a pollutant and cannot be controlled. Particle-bound water is not
 8      included in the mass of PM subject to regulation and control. Because the mass of particle-
 9      bound water could be equal or greater than that of the other components, a measurement of PM
10      mass including particle-bound water would depend more on relative humidity that pollution.
11      For all these reasons, it is usually desirable to remove most, if not all, particle-bound water
12      before weighing collected PM. However, in some situations it may be important to know how
13      much of the suspended particle's mass or volume results from particle-bound water.
14     Figures 2-13 and 2-14 show the change in diameter of sulfate particles as a function of relative
15     humidity. Figure 2-13 also shows hysteresis resulting from the difference between
16     deliquescence and crystallization points.
17           Pilinis et al. (1989) calculated the water content of atmospheric particulate matter above
18     and below the deliquescent point. They predicted that aerosol water content is strongly
19     dependent on composition and concluded from their calculations that liquid water could
20     represent a significant mass fraction of aerosol concentration at relative humidities above 60%.
21      Since then, a few researchers have attempted to measure the water content of atmospheric
22     aerosol. Most techniques have focused on tracking the particle mass as the relative humidity is
23      changed and are still in the development phase. There have been only a few demonstrations
24      using actual ambient aerosol to date. Of interest, in particular, is the development of the Tandem
25      Differential Mobility Analyzer (TDMA) and its applications in investigations of the effects of
26      relative humidity on particle growth.
27           Lee et al. (1997) examined the influence of relative humidity on the  size of atmospheric
28      aerosol using a TDMA coupled with a scanning mobility particle sizer (SMPS). They reported
29      that the use of the TDMA/SMPS system allowed for the abrupt size changes of aerosols at the
 30      deliquescence point to be observed precisely. They also reported that at relative humidities
        June 2003                                 2-52        DRAFT-DO NOT QUOTE OR CITE

-------
  1      between 81 and 89% the water content of ammonium sulfate aerosols (by mass) ranged from
  2      47 to 66%.
  3           Andrews and Larson (1993) investigated the interactions of single aerosol particles coated
  4      with an organic film within a humid environment. Using an electrodynamic balance, they
  5      conducted laboratory experiments in which sodium chloride and carbon black particles were
  6      coated with individual organic surfactants (intended to simulate the surface-active, organic films
  7      that many atmospheric aerosol particles may exhibit) and their water sorption curves were
  8      examined. Their results showed that when ordinarily hydrophobic carbon black particles were
  9      coated with an organic surfactant, they sorbed significant amounts of water (20 to 40% of the dry
 10      mass of the particle).
 11           Liang and Chan (1997) developed a fast technique using the electrodynamic balance to
 12      measure the water activity of atmospheric aerosols.  In their technique, the mass of a levitated
 13      particle is determined as the particle either evaporates or grows in response to a step change in
 14      the relative humidity. Their technique was demonstrated using laboratory experiments with
 15      NaCl, (NH4)2SO4, NaN03, and (NH4)2SO4/NH4NO3 solutions. They concluded  that one of the
 16      advantages of their fast method is the ability to measure the water activity of aerosols containing
 17      volatile solutes such as ammonium chloride and some organics.
 18           Mclnnes et al. (1996) measured aerosol  mass concentration, ionic composition, and
 19      associated water mass of marine aerosol over  the remote Pacific Ocean. The mass of
20      particle-bound water was determined by taking the difference between the mass obtained at 48%
21      RH and at 19% RH, assuming the aerosol particles were dry at 19% RH. Based on a comparison
22      of the remote Pacific aerosol to aerosol collected at a site at the marine/continental interface of
23      the Washington coast, the amount of water associated with the aerosol was observed to be a
24      function of the ammonium to sulfate ratio. They found that the amount of water associated with
25      the submicrometer aerosol comprised 29% of the total aerosol mass collected at 47% RH and
26      9% of the total mass at 35% RH.
27           Ohta et al. (1998) characterized the chemical composition of atmospheric  fine particles
28      (50% cut point of 2 um) in Sapporo, Japan, and as part of their measurements, determined the
29      water content using the Karl Fischer method (Meyer and Boyd, 1959).  After exposing a Teflon
30      filter, a portion of the filter was equilibrated at 30% RH for 24 h. Then the filter piece was
31      placed in a water evaporator heated at 150 DC, vaporizing the particle-bound water. The vapor

        June 2003                                 2-53        DRAFT-DO NOT QUOTE OR CITE

-------
 1      evolved was analyzed for water in an aqua-counter where it was titrated coulometrically in Karl
 2      Fischer reagent solution (containing iodine, sulfur, and methanol). The accuracy of the aqua-
 3      counter is ±1 mg. Using this technique, they determined that the water content of the particles
 4      ranged from 0.4 to 3.2% of the total paniculate mass (at RH < 30%).  This represents a smaller
 5      portion of water compared to their previous reported values (Ohta and Okita, 1990) that were
 6      determined by calculation at RH of 50%.
 7           Speer et al. (1997) developed an aerosol liquid water content analyzer (LWCA) in which
 8      aerosol samples are collected on PTFE filters and then placed in a closed chamber in which the
 9      relative humidity is closely controlled. The aerosol mass is monitored using a beta-gauge, first
10      as the relative humidity is increased from low RH to high RH, and then as the RH is decreased
11      again.  They demonstrated the LWCA on laboratory-generated aerosol and on an ambient PM2 5
12      sample collected in Research Triangle Park, NC. The ambient aerosol sample was also analyzed
13      for chemical constituents.  It is interesting to note that, although their laboratory-generated
14     (NH4)2SO4 aerosol demonstrated a sharp deliquescent point, their atmospheric aerosol, which
15      was essentially (NH4)2SO4, did not show a sharp deliquescent point.
16          Hygroscopic properties of aerosols have been studied from the viewpoint of their ability to
17     act as condensation nuclei. The hygroscopic properties of fresh and aged carbon and diesel soot
18     particles were examined by Weingartner et al. (1997) who found that fresh, submicron-size
19     particles tended to shrink with increasing relative humidity because of a structural change.
20     Lammel and Novakov (1995)  found, in laboratory studies, that the hygroscopicity of soot
21      particles could be increased by chemical modification and that the cloud condensation nucleation
22     characteristics of diesel soot were similar to those of wood smoke aerosol.
23          The results of several of the above studies in which aerosol water content as a function of
24     relative humidity was determined are summarized in Figure 2-17. In this figure, the results of
25     Lee et al. (1997), Mclnnes et al. (1996), and Ohta et al. (1998) are included.  Relative humidity
26     ranged from 9%, at which the aerosol water content was assumed to be zero (Mclnnes et al.,
27      1996), to 89%, at which the aerosol water content was determined to be 66% by mass (Lee et al.,
28      1997). Koutrakis et al. (1989) and Koutrakis and Kelly (1993) also have reported field
29     measurements of the equilibrium size of atmospheric sulfate particles as a function of relative
30     humidity and acidity.
31

        June 2003                                 2-54        DRAFT-DO NOT QUOTE OR CITE

-------
                        100
                         90
                         80
                         70
                         60-
                         50
                      O  40
                      to
                      
-------
 1      Canyon, AZ, during the 1990 Navajo Generating Station Visibility Study were reported by
 2      Zhang et al. (1993). They used a tandem differential mobility analyzer (TDMA; McMurry and
 3      Stolzenburg, 1989) to measure the hygroscopic properties for particles in the 0.05- to 0.5-um
 4      range. In their experimental technique, monodisperse particles of a known size are selected from
 5      the atmospheric aerosol with the first DMA. Then, the relative humidity of the monodisperse
 6      aerosol is adjusted, and the new particle size distribution is measured with the second DMA.
 7      At both sites, they observed that monodisperse particles could be classified according to "more"
 8      hygroscopic and "less" hygroscopic. Aerosol behavior observed at the two sites differed
 9      markedly. Within the experimental uncertainty (±2%) the "less" hygroscopic particles sampled
10      in Los Angeles did not grow when the RH was increased to 90%; whereas at the Grand Canyon,
11      the growth of the "less" hygroscopic particles varied from day to day, but ranged from near 0 to
12     40% when the RH was increased to 90%. The growth of the "more" hygroscopic particles in
13      Los Angeles was dependent on particle size (15% at 0.05 urn to 60% at 0.5 urn); whereas at the
14     Grand Canyon, the "more" hygroscopic particles grew by about 50% with the growth not
15     varying significantly with particle size. By comparison of the TDMA data to impactor data,
16     Zhang et al. (1993) surmised that the more hygroscopic particles contained more sulfates and
17     nitrates while the less hygroscopic particles contained more carbon and crustal components.
18          Although most of the work to date on the hygroscopic properties of atmospheric aerosols
 19     has focused on the inorganic fraction, the determination of the contribution of particle-bound
20     water to atmospheric particulate mass is greatly complicated by the presence of organics. The
21     effect of RH on adsorption of semivolatile organic compounds is discussed elsewhere in this
22     chapter. Saxena et al. (1995) observed that particulate organic compounds also can affect the
23     hygroscopic behavior of atmospheric particles.  They idealized the organic component of aerosol
 24      as containing a hydrophobic fraction (high-molecular weight alkanes, alkanoic acids, alkenoic
 25      acids, aldehydes, and ketones) and a hydrophilic fraction (e.g., lower molecular weight
 26      carboxylic acids, dicarboxylic acids, alcohols, aldehydes, etc.) that would be likely to absorb
 27      water.  They then analyzed data from a tandem differential mobility analyzer in conjunction with
 28      particle composition observations from an urban site (Claremont, CA) and from a nonurban site
 29      (Grand Canyon) to test the hypothesis that, by adding particulate organics to an inorganic
 30      aerosol, the amount of water absorbed would be affected, and the effect could be positive or
 31      negative, depending on the nature of the organics added. They further presumed that the

        June 2003                                 2-56        DRAFT-DO NOT QUOTE OR CITE

-------
  1     particulate organic matter in nonurban areas would be predominantly secondary and thus
  2     hydrophilic, compared to the urban aerosol that was presumed to be derived from primary
  3     emissions and thus hydrophobic in nature. Their observations were consistent with their
  4     hypothesis, in that at the Grand Canyon, the presence of organics tended to increase the water
  5     uptake by aerosols; whereas at the Los Angeles site, the presence of organics tended to decrease
  6     water uptake.
  7          Peng and Chan (2001) also recently studied the hygroscopic properties of nine water
  8     soluble organic salts of atmospheric interest using an electrodynamic balance operated at 25°C.
  9     Salts studied included sodium formate, sodium acetate, sodium succinate, sodium pyruvate,
 10     sodium methanesulfonate, sodium oxalate, ammonium oxalate, sodium malonate, and sodium
 11     maleate.  They observed that hygroscopic organic salts have a growth factor of 1.76-2.18 from
 12     RH=10-90%, comparable to that of typical hygroscopic inorganic salts such as NaCl and
 13     (NH4)2S04.
 14          Nonequilibrium issues may be important for the TDMA, as well as for other methods of
 15     measuring water content. Although approach to equilibrium when the RH is increased is
 16     expected to be rapid for pure  salts, it may be much slower for aerosols containing a complex mix
 17     of components (Saxena et al., 1995). For example, if an aerosol contains an organic film or
 18     coating, that film may impede the transport of water across the particle surface, thus increasing
 19     the time required for equilibrium (Saxena et al., 1995).  Insufficient time to achieve equilibrium
 20     in the TDMA could result in underestimation of the water content.
 21
 22     2.2.3.2   Nitrate and Organic Species
 23      Particulate Nitrates
 24          It is now well known that volatilization losses of particulate nitrates occur during sampling
 25      on Teflon filters (e.g., Zhang and McMurry [1992]; see also Hering and Cass [1999] and Babich
 26      et al., [2000]). The effect on the accuracy of atmospheric particulate measurements from these
 27      volatilization losses is more significant for PM25 than for PM10.  The FRM for PM25 will likely
28      suffer loss of nitrates similar to that experienced with other simple filter collection systems.
29      Sampling artifacts resulting from the loss of particulate nitrates represents a significant problem
30      in areas such as southern California that experience high amounts of nitrates. Hering and Cass
31      (1999) reported on errors in PM25 mass measurements due to volatilization of particulate nitrate

        June 2003                                2-57        DRAFT-DO NOT QUOTE OR CITE

-------
 1      (Figure 2-18). They examined data from two field measurement campaigns conducted in
 2      southern California: (1) the Southern California Air Quality Study (SCAQS) (Lawson, 1990)
 3      and (2) the 1986 CalTech study (Solomon et al., 1992). In both these studies, side-by-side
 4      sampling of PM2 5 was conducted.  One sampler collected particles directly onto a Teflon filter.
 5      The second sampler consisted of a denuder to remove gaseous nitric acid followed by a nylon
 6      filter that absorbs the HNO3 which evaporates from ammonium nitrate. In both studies, the
 7      denuder consisted of MgO-coated glass tubes (Appel et al., 1981). Fine particulate nitrate
 8      collected on the Teflon filter was compared to fine particulate nitrate collected on the denuded
 9      nylon filter. In both studies, the PM25 mass lost because of volatilization of ammonium nitrate
10      represented a significant fraction of the total PM2.5 mass.  The fraction of mass lost was higher
11      during summer than during fall (17% versus 9% during the SCAQS study and 21% versus 13%
12     during the CalTech study; Figure 2-18). In regard to percentage loss of nitrate, as opposed to
13      percentage loss of mass discussed above, Hering and Cass (1999) found that nitrate remaining on
14     the Teflon filter samples was on average 28% lower than that on the denuded nylon filters.
15
            80%
           60%-
         £  40%-
            20%'
            0%
                                 SCAQS Data Set
                                o Summer Measurements
                                • Fal Measurements
 50     100    150    200    250
PM,s Gravimetric Mass (M9/rn3)
                                                         80%
                                                      £ 60%-
                                                         40%-
                                                       a 20%-
                                                          Caltech Data Set
                                                           o April - September
                                                           • October-March
                                                                         100
                                                                               150
                                                                                      200
                                                                                            250
                                                                    PM,, Gravimetric Mass (|ag/m )
        Figure 2-18. Amount of ammonium nitrate volatilized from Teflon filters, expressed as a
                     percentage of the measured PM2>5 mass, for the SCAQS and CalTech studies,
                     for spring and fall sampling periods.
        Source: Hering and Cass (1999).
        June 2003
                             2-58
DRAFT-DO NOT QUOTE OR CITE

-------
   1           Hering and Cass (1999) also analyzed these data by extending the evaporative model
   2      developed by Zhang and McMurry (1987). The extended model utilized by Hering and Cass
   3      (1999) takes into account dissociation of collected paniculate ammonium nitrate on Teflon
   4      filters into nitric acid and ammonia via three mechanisms:  (1) scrubbing of nitric acid and
   5      ammonia in the sampler inlet (John et al., 1988 showed that clean PMj0 inlet surfaces serve as an
   6      effective denuder for nitric acid), (2) heating of the filter substrate above ambient temperature by
   7      sampling, and (3) pressure drop across the Teflon filter. For the sampling systems modeled, the
   8      flow-induced pressure drop was measured to be less than 0.02 atm, and the corresponding
   9      change in vapor pressure was 2%, so losses driven by pressure drop were not considered to be
 10      significant in this work.  Losses from Teflon filters  were found to be higher during the summer
 11      compared to the winter, higher during the day compared to  night, and reasonably consistent with
 12      modeled predictions.
 13           Finally, during the SCAQS study, particulate samples also were collected using a Berner
 14      impactor and greased Tedlar substrates in size ranges from 0.05 to 10 urn in aerodynamic
 15      diameter.  The Berner impactor PM2 5 nitrate values were much closer to those from the denuded
 16      nylon filter than those from the Teflon filter with the impactor nitrate being approximately
 17      2% lower than the nylon filter nitrate for the fall measurements and approximately 7% lower
 18      during the summer measurements. When the impactor collection was compared to the Teflon
 19      filter collection for a nonvolatile species (sulfate), the results were in agreement.
 20          ft should be noted that filters or collection surfaces were removed immediately after
 21      sampling and placed into vials containing a basic extraction solution during these
 22      intercomparison studies.  Therefore, losses that might occur during handling, storage, and
 23     equilibration of filters or impaction surfaces were avoided. The loss of nitrate observed from
 24     Teflon filters and impaction surfaces in this study, therefore, is a lower limit compared to losses
 25     that might occur during the normal processes involved in equilibration and weighing of filters
 26     and impaction surfaces. Brook and Dann (1999) observed much higher nitrate losses during a
 27     study in which they measured particulate nitrate in Windsor  and Hamilton, Ontario, Canada, by
 28     three techniques:  (1) a single Teflon filter in a dichotomous sampler, (2) the Teflon filter in an
 29     annular denuder system (ADS), and (3) total nitrate including both the Teflon filter and the
30     nylon back-up filter from the ADS. The Teflon filter from the dichotomous sampler averaged
31      only 13% of the total nitrate.  The Teflon filter from  the ADS averaged 46% of the total nitrate.

        June, 2003                                2-59        DRAFT-DO NOT QUOTE OR CITE

-------
 1      The authors concluded that considerable nitrate was lost from the dichotomous sampler filters
 2      during handling, which included weighing and x-ray fluorescence (XRF) measurement in a
 3      vacuum.
 4           Kim et al. (1999) also examined nitrate sampling artifacts by comparing denuded and
 5      undenuded quartz and nylon filters, during the PM10 Technical Enhancement Program (PTEP) in
 6      the South Coast Air Basin of California. They observed negative nitrate artifacts (losses) for
 7      most measurements; however, for a significant number of measurements they observed positive
 8      nitrate artifacts. Kim et al. (1999) pointed out that random measurement errors make it difficult
 9      to measure true amounts of nitrate loss.
10           Diffusion denuder samplers, developed primarily to measure particle strong acidity
11      (Koutrakis et al., 1988a,b, 1992), also can be used to study nitrate volatilization. Such
12     techniques were used to measure loss of particulate nitrate from Teflon filters in seven U.S.
13     cities (Babich et al., 2000). Measurements were made with two versions of the Harvard-EPA
14     Annular Denuder System (HEADS).  Nitric acid vapor was removed by a Na2CO3-coated
15     denuder. Particulate nitrate was the sum of nonvolatile nitrate collected on a Teflon filter and
16     volatized nitrate collected on a Na2CO3-coated denuder downstream of the Teflon filter (full
17     HEADS) or on a Nylon filter downstream of the Teflon filter (Nylon HEADS). It was found that
 18     lull HEADS (using a Na2CO3 denuder) consistently underestimated the total particulate nitrate
 19     by approximately 20% compared to the nylon HEADS. Nonvolatilized nitrate was also
20     measured on a Teflon filter from collocated Harvard Impactors (HI). The PM2.5 HI sampler, like
21     the PM2.5 FRM, use impactors with a 50% cut at 2.5 urn.  The HI uses a 37 mm filter and a flow
 22     rate of 10 L/min (0.465 L/min/cm2) which the FRM uses a 47 mm filter and a flow rate of 16.7
 23     L/min (0.481 L/min/cm2). Therefore, the flow rate and pressure drop across the filters are
 24      similar and the loss of nitrate should be similar in both types of samples.  Babich et al. (2000)
 25      found significant nitrate losses in Riverside, CA;  Philadelphia, PA; and Boston, MA but not in
 26      Bakersfield, CA; Chicago, IL; Dallas, TX; or Phoenix, AZ where measurements were made only
 27      during winter.
 28           Eatough et al. (1999a) developed a high-volume diffusion denuder system in which
 29      diffusion denuder and particle concentrator techniques were combined (see Section 2.2.3.2). The
 30     particle concentrator reduces the flow through the denuder so that the denuder can be operated
 31      for weeks without a loss of collection efficiency, making the sampler suitable for routine field
         June 2003
                                                 2-60
DRAFT-DO NOT QUOTE OR CITE

-------
  1      sampling. The system was evaluated for the collection of fine participate sulfate and nitrate in
  2      Riverside, CA (Eatough et al., 1999b).  Concentrations of PM25 nitrate obtained from the PC-
  3      BOSS agreed with those obtained using the Harvard-EPA Annular Denuder Sampler, HEADS
  4      (Koutrakis et al., 1988b).
  5
  6      Paniculate Organic Compounds
  1           Many semivolatile organic compounds (SVOC) are of interest because of their possible
  8      health effects.  SVOC include products of incomplete combustion such as polycyclic aromatic
  9      hydrocarbons (PAHs) and polycyclic organic matter, which has been identified as a hazardous
 10      air pollutant. PAHs also have been suggested as alternative paniculate tracers for automobile
 11       emissions because the phase-out of organo-lead additives to gasoline means that lead is no
 12      longer a good tracer for automobiles (Venkataraman et al., 1994). PAHs also are emitted during
 13       biomass burning, including burning of cereal crop residues and wood fuels (Jenkins et al., 1996;
 14       Roberts and Corkill, 1998).
 15           A number of studies have used absorbing material following quartz filters to determine
 16      phase equilibria of specific organic compounds (Liang et al., 1997; Gundel et al., 1995; Kamens
 17      et al., 1995). Much work has also gone into the development of a theory to help understand the
 18      phase relationships (Yamasaki et al., 1982; Rounds and Pankow, 1990; Pankow, 1987,1994a,b;
 19      Pankow etal., 1993; Rounds etal., 1993; Odum etal., 1994). The theory describing phase
 20      equilibria of semivolative organic compounds (SVOC) and the gas/particle partitioning of SVOC
 21      on inorganic, organic, and ambient smog aerosols continues to be developed (Liang et al., 1997;
 22      Jang et al., 1997,1999; Strommen and Kamens, 1997; Jang and Kamens, 1998,1999,2001;
 23      Leach et al., 1999; Kamens et al., 1999, 2001).
 24           The mass of organic and elemental carbon is usually determined by analysis of PM
 25      collected on a quartz filter. However, quartz fiber filters have a large specific surface area on
 26      which adsorption of gases can occur. Possible artifacts associated with adsorption of organic
 27      gases onto quartz filters have been examined in experiments in which two quartz fiber filters
 28      were deployed in series. The second quartz filter may indicate gaseous volatile organic
29      compounds (VOC) adsorbed on both filters (positive artifact), SVOC evaporated from particles
30      on the first filter and subsequently adsorbed on the second filter (negative artifact), or a
31      combination of both effects. Unless the individual compounds are identified, the investigator

        June 2003                                2-61        DRAFT-DO NOT QUOTE OR CITE

-------
 1      does not know what to do with the loading value on the second filter (i.e., to add or subtract from
 2      the first filter loading value). Moreover, even if the individual compounds were identified on the
 3      back-up filter, the decision concerning adding or subtracting the back-up filter loading would not
 4      be straightforward.
 5           The positive quartz filter artifact has been discussed by Gundel et al. (1995) and Turpin
 6      et al. (2000). It is also possible that some SVOC may desorb from the filter resulting in a
 7      negative artifact (Eatough et al., 1993; Tang et al., 1994; Eatough et al., 1995; Gundel et al.,
 8      1995; Cui et al., 1998; Pang et al., 2001; Finn et al, 2001).  Semivolatile organic compounds can
 9      similarly be lost from Teflon filters because of volatilization, causing the PM2 5 mass to be
10      significantly underestimated (negative artifact). Like paniculate nitrates, the FRM for PM2 5 will
11      suffer loss of SVOC, similar to the losses experienced with other simple filter collection
12      systems. Most studies that have focused on the positive and negative sampling artifacts
13      associated with SVOC compounds have utilized either diffusion denuder technology (Eatough
14     et al., 1995; Mader et al.,  2001) or placed an adsorbent media, such as a back-up quartz filter or a
15     polyurethane foam adsorbent behind the main filter (Wallace and Kites, 1995). Further
16     information on denuder techniques are given in Appendix 2A.
17          Using their multichannel  diffusion denuder sampling  system (BOSS), Eatough et al. (1995)
18     reported that, for  samples collected at the South Coast Air Quality Management District
19     sampling site at Azusa, CA, changes in the phase distribution of SVOC could result in a loss on
20     average of 35% of the particulate organic material. Cui et al. (1998) found that losses of SVOC
21     from particles in the Los Angeles Basin during the summer were greater during the night
22     (average = 62%) than during the day (average = 42%).
23           The percent SVOC  lost from the front filter in a filter-denuder system may be greater than
24     that lost in a filter-only system such as the FRM. In a filter-denuder system, the gas-phase
25     component of the SVOC  is removed. The absence of the gas-phase causes the gas-particle
26     equilibrium to shift so the SVOC collected on the filter may evaporate more rapidly in a filter-
27     denuder system than in a filter-only collection system. To  determine the fraction of SVOC lost
28      from a Teflon filter in a filter-only system, it is necessary to compare the amount measured by a
29      nondenuder system with  that measured by a denuder system.  At present, little information is
30      available on the volatilization losses of SVOC. However, in one study (Pang et al., 2001), the
31      total mass on denuded and undenuded filters were compared and found to be identical within

        June 2003                                 2-62        DRAFT-DO NOT QUOTE OR CITE

-------
   1      error limits (R2 = 0.816, slope = 0.961 ± 0.027 for total mass compared to R2 - 0.940, slope =
   2      0.986 ± 0.020 for sulfate). Pang et al. interpreted this result as suggesting that the major cause of
   3      loss of SVOC is the pressure drop across the filter.
   4           Positive artifacts may occur during sample collection because of the adsorption of gases
   5      onto the filter materials (e.g., Gundel et al., 1995). Using a quartz filter behind a Teflon filter,
   6      Kim et al. (200 la) estimated that on  an annual average basis 30% of the PM25 organic carbon
   7      concentration resulted from positive  artifacts. There is a larger positive artifact because of
   8      greater adsorption of organic vapor onto quartz fiber filters than onto Teflon filters (Turpin et al.,
   9      1994; Chow et al., 1994a,b, 1996; Eatough et al., 1996; Finn et al., 2001; Kirchstetter et al.,
 10      2001).
 11           Kirchstetter et al. (2001) report that adsorptive properties of quartz fiber filters vary with
 12      lot number; therefore, front and back-up filters should be taken from the same lot. The literature
 13      suggests that a Teflon filter followed by a quartz back-up filter appears to provide a better
 14      estimate of the adsorption of gases on a quartz fiber front filter than does a quartz filter followed
 15      by a quartz backup and that the difference between these two adsorption estimates can be
 16      substantial for short durations (Novakov et al., 1997; Kirchstetter et al., 2001; Turpin et al.,
 17      2000). The typically lower organic carbon loadings on concurrently collected quartz followed
 18      by quartz filters relative to Teflon followed by quartz filters are believed to occur because
 19      adsorption on the quartz front filter acts to reduce the gas-phase concentration downstream until
 20      adsorption equilibrium has been achieved in the vicinity of the front quartz filter surface.
 21      Because Teflon filters have little affinity for organic vapors, this equilibrium occurs almost
 22      instantaneously for Teflon filters, and the Teflon-quartz back-up filter is exposed to the ambient
 23      concentration of organic vapors from  the beginning of the sampling period.  It might be expected
 24     that the quantity of organic vapor adsorbed on quartz filters would depend on the organic
 25     composition and would vary by season and location. However, it is also possible that the quartz
 26     possesses a limited number of adsorption sites which are rapidly occupied so that the quantity of
 27     OC on the back up filter would be relatively constant and depend on the pretreatment of the
28     quartz.
29
30
31


        June 2003                                 2-63        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Combined Measurement of Semivolatile Nitrate and OC and Nonvolatile Species (OC, EC,
 2      Nitrate, andSulfate)
 3           Fine particles in urban atmospheres contain substantial quantities of semivolatile material
 4      (e.g., NH4N03 and semivolatile organic compounds [SVOC]) that are lost from particles during
 5      collection on a filter.  Several diffusion denuder samplers have been developed for the
 6      determination of both NO3' and organic semivolatile fine particulate components as well as
 7      nonvolatile nitrate and organic compounds and nonvolatile sulfate (Pang et al., 2001; Eatough
 8      et al., 1993). The combination of technology used in the BOSS diffusion denuder sampler and
 9      the Harvard particle concentrator has resulted in the Particle Concentrator-Brigham Young
10      University Organic Sampling System (PC-BOSS) for the 24-hr integrated collection of PM2.5,
11      including NH4NO3 and semivolatile organic material. Modifications of the BOSS sampler allow
12      for the determination of these same species on a time scale from a few hours to weekly (Ding et
13      al., 2002; Eatough et al., 1999a,b; 2001). Episodic studies have been conducted in Riverside,
14     CA and Bakersfield, CA (Obeidi et al., 2002) and Provo, UT (Obeidi and Eatough, 2002). The
15     average concentration of semivolatile and nonvolatile components in the three cities, during the
16     study periods, are shown in Figure 2-19.
17
18     2.2.3.3   Continuous Measurement of Semivolatile and Nonvolatile Mass
19           Techniques for the continuous measurement of PM mass is needed both to provide
20     real-time information on pollution levels (Long et al., 2002) and to reduce the costs involved in
21     visiting sites to change filters and in the equilibration and weighing of filters. Two techniques
22     are currently  in use for the continuous measurement of PM mass. The TEOM is normally
23     operated at 50° C in order to remove particle-bound water. However, at 50° C most semivolatile
24     material is also evaporated. Therefore, the TEOM, operated at 50° C, may be considered to
25     measure the mass of nonvolatile PM.  The beta gauge mass monitor changes the filter more
26     frequently than the TEOM and is less sensitive to changes in mass caused by changes in relative
27     humidity.  It  does not control the temperature at the filter. However, most beta gauge monitors
28      heat the inlet. This heating causes evaporation of a substantial fraction of the particle-bound
 29      water and an unknown fraction of the semivolatile PM. Thus, the beta gauge may be considered
 30      to measure the nonvolatile PM plus a small fraction of the particle-bound water and an unknown
 31      fraction of the semivolatile PM. Three new techniques have been developed to address the issue
 32      of lost semivolatile PM mass.

        June 2003                                2-64        DRAFT-DO NOT QUOTE OR CITE

-------
                                              Riverside, CA; September 1997
                                                                -;= ~ J^Non-volatll*
                                                               - EL= a :: |a,C«mpon«nti
                                       Lo
                                    Ammonium
                                        Ntrate
                                           23
                                            Ammonium
                                             Nitrate   H oriental Ammonium
                                              2-6     Carbon   Sutfats
                                                      3.1       1.2
                                               BakersHefd, CA; March 1998
                                       S«mlvolatlj«
                                     Component! ,
                                                                    Non-volatll*
                                                          EH  ~ ^  ^=\ Components
                                    Lost
                                 Ammonium
                                     Nitrate
                                       as
 •:•::• Ammonium
       Sulfate
          1.7
                                              Prove, LIT; September 1998
                                                                 ^ Non-volatl!*
                                                                     Components
                                                                   Ammonium
Figure 2-19.   Average concentration of nonvolatile and semivolatile PM components in
                 three cities (u.g/m3).
Source: Obeidi and Eatough (2002); Obeidi et al. (2002).


June 2003                                        2-65
DRAFT-DO NOT QUOTE OR CITE

-------
 1      Real-Time Total Ambient Mass Sampler (RAMS)
 2           The RAMS, a monitor based on diffusion denuder and TEOM monitor technology, has
 3      been developed, validated, and field tested for the real-time determination of total fine PM mass,
 4      including semivolatile PM (Eatough et al., 1999a; Obeidi and Eatough, 2002; Obedi et al, 2002;
 5      Pang et al., 2001). The RAMS measures the total mass of collected particles including
 6      semivolatile species with a TEOM monitor using a "sandwich filter." The sandwich contains a
 7      Teflon coated particle collection filter followed by a charcoal-impregnated filter (GIF) to collect
 8      any semivolatile species lost from the particles during sampling.  Because the instrument
 9      measures total mass collected by the sandwich filter, all gas phase compounds that can be
10      adsorbed by a GIF must be removed from the sampling stream prior to the TEOM monitor.
11      Laboratory and field validation data indicate that the precision of fine PM mass determination is
12     better than 10%. The RAMS uses a Nafion dryer to remove particle-bound water from the
13      suspended particles and a particle concentrator to reduce the amount of gas phase organics that
14     must be removed by the denuder. An  example of data from the RAMS, the TEOM, and the
15     PC-BOSS is shown in Figure 2-20.  This figure also shows the PM25 mass from the TEOM as
16     being negative for the hours of 16 to 19. Thislikely results from the loss of volatile materials
17     from the heated filter.
18
19     Differential TEOM
20           Patashnick et al. (2001) developed a differential TEOM system that is based on a pair of
21     TEOM sensors, each of which is preceded by its own electrostatic precipitator (ESP) and
22     downstream from a common size selective inlet. By alternately switching the ESPs on and off
23     and out of phase with each other, the two sensors measure "effective mass" that includes both
24     the nonvolatile component and the volatile component sampled by the TEOM, less the volatile
25     component that vaporized during the sampling interval. On the sensor side with the ESP turned
26     on, there is no particle collection on that filter so that only volatilization of previously collected
27     particles continues. This allows a correction for the effective mass as measured by the first
28      sensor by subtracting out the volatilization artifact and leaving the nonvolatile and volatile
29      components of the paniculate matter. This system  has yet to be well characterized for other
30      biases or interferences such as reactions on the filters, particle collection efficiency of the ESPs,
31      and particle and semivolatile material losses.

        June 2003                                 2-66       DRAFT-DO NOT QUOTE OR CITE

-------
                       PC-BOSS (Nonvolatile Material)
                                                         PC-BOSS (Lost From Particles)
                     100
                              TEOM
                              at35C
                                     ™B—  RAMS
                                           at35C
	FRMPM«
     24 h average
                                           Riverside, CA
                                                     I 1 I  I I I  M I  I I II I 11 I I  1 I
                        13
                       15
                                   17
                                  19
                                             21
                                               Time of Day
       Figure 2-20.  Comparison of mass measurements with collocated RAMS (real-time data),
                    PC-BOSS (1-h data), FRM PM2S sampler (average of 24-h data), and a
                    conventional TEOM monitor (real-time data). The semivolatile fine
                    particulate matter is sampled with the RAMS and PC-BOSS, but not with
                    the TEOM monitor or the FRM PMW single filter sampler.  The PC-BOSS
                    provides information on both the nonvolatile component (NV) and the
                    semivolatile organic component (SVOC).
       Source: Eatoughetal. (1999a).
 1
 2
 3
 4
 5
 6
7
8
9
 Continuous Ambient Mass Monitor (CAMM)
     Koutrakis and colleagues (Koutrakis et al., 1996; Wang, 1997) have developed CAMM, a
 technique for the continuous measurement of ambient particulate matter mass concentration
 based on the measurement of pressure drop increase with particle loading across a membrane
 filter. Recently, Sioutas et al. (1999) examined the increase in pressure drop with increasing
particle loading on Nuclepore filters. They tested filters with two pore diameters (2 and 5 um)
and filter face velocities ranging from 4 to 52 cm s': and examined the effects of relative
humidity in the range of 10 to 50%. They found that, for hygroscopic ammonium sulfate
particles, the change in pressure drop per unit time and concentration was a strong function of
      June 2003
                                        2-67
                                                         DRAFT-DO NOT QUOTE OR CITE

-------
 1      relative humidity, decreasing with increasing relative humidity. These results suggest that
 2      paniculate concentration measurements, similar to the method of Koutrakis et al. (1996) that
 3      uses the pressure drop method, may be subject to additional uncertainties if used in an
 4      environment where the ambient relative humidity is quite variable and the relative humidity
 5      where the particles are measured cannot be controlled accurately. The current version of the
 6      CAMM (Wang, 1997) uses a particle concentrator, a Nafion dryer, and frequently moves the
 7      filter tape to avoid artifacts due to evaporation of semivolatile components from the active
 8      portion of the filter tape which would occur if the atmospheric concentration of the semivolatile
 9      components decreased.
10           The CAMM was recently operated alongside a gravimetric PM method (the Harvard
11      Impactor, or HI) in seven U.S. cities selected for their distinctly different ambient paniculate
12      compositions and densities. The correlation between the two methods was high, with an overall
13      r2 of 0.90 and average CAMM/HI ratio of 1.07 (Babich et al., 2000).
14
15      2.2.4    U. S. Environmental Protection Agency Monitoring Methods
16     2.2.4.1   The Federal Reference Methods for Measurement of Equilibrated Mass for PM10,
17              PM2.5, and PM10_2.5
18          jn 1997, EPA promulgated new standards for PM25 to address fine-fraction thoracic
19     particles and retained with minor revisions the 1987 PMIO standards to continue to address
20     coarse-fraction thoracic particles (Federal Register, 1997). In partial response to numerous
21     challenges to these standards, the U.S. Court of Appeals for the District of Columbia Circuit in
22     American Trucking Association v. EPA, 175 F. 3d  1027 (U.S. Court of Appeals, D.C. Cir. 1999)
23     found "ample support" for regulating coarse-fraction particles but revoked the revised PM]0
24     standards (leaving in effect the 1987 PM10 standards) on the basis of PM10 being a "poorly
25     matched indicator for coarse particulate pollution" because PMi0 includes fine particles.
26     Consistent with this specific aspect of the Court's ruling, which EPA did not appeal, EPA is now
27     considering use of PM10_2.5 as the indicator for coarse-fraction thoracic particles, in conjunction
28     with PM2 5 standards that address fine-fraction thoracic particles. Thus, EPA is now developing
29     a Federal Reference Method for the measurement of PM10_2 5.
30
 31

        June 2003                                2-68        DRAFT-DO NOT QUOTE OR CITE

-------
  1      2.2.4.1.1  PMIO
  2           The FRM specified for measuring PM]0 (Code of Federal Regulations, 2001a,b) has been
  3      discussed in previous PM AQCD's and will only be mentioned briefly. The PM!0 FRM defines
  4      performance specifications for samplers in which particles are inertially separated with a
  5      penetration efficiency of 50% at an aerodynamic diameter (DJ of 10 ± 0.5 jim. The collection
  6      efficiency increases to ~ 100% for smaller particles and drops to = 0% for larger particles.
  7      Particles are collected on filters and mass concentrations are determined gravimetrically.
  8      Instrument manufacturers are required to demonstrate through field tests a measurement
  9      precision for 24-h samples of ± 5 ug/m3 for PM10 concentrations below 80 u.g/m3 and 7% above
 10      this value. A number of samplers have been designated as PM10 reference samplers. The TEOM
 11      and several beta gauge samplers with 1-h time resolution have been designated as automated
 12      equivalent methods (U.S. Environmental Protection Agency, 2001).
 13
 14      2.2.4.1.2 PM2S
 15           In contrast to the performance-based FRM standard for PM10, the FRM for PM2.5 (Code of
 16      Federal Regulations, 2001a) specifies certain details of the sampler design, as well as of sample
 17      handling and analysis, whereas other aspects have performance specifications (Noble et al.,
 18      2001).  The PM2.5 FRM sampler consists of a PM10 inlet/impactor, a PM2 5 impactor with an oil-
 19      soaked impaction substrate to remove particles larger than 2.5 jjm Da, and a 47-mih PTFE filter
 20      with a particle collection efficiency greater than 99.7%. The sample duration is 24 h, during
 21      which time the sample temperature is not to exceed ambient temperatures by more than 5 "C.
 22      A schematic diagram of the PM25 FRM sample collection system is shown in Figure 2-21. After
 23      collection, samples are equilibrated for 24 h at temperatures in the range of 20 to 23 °C (± 2 °C)
 24      and at relative humidities in the range of 30 to 40% (± 5%).  The equilibration tends to reduce
 25      particle-bound water and stabilizes the filter plus  sample weight. Filters are weighed before and
 26      after sampling under the same temperature and relative humidity conditions. For sampling
 27      conducted at ambient relative humidity less than 30%, mass measurements at relative humidities
 28      down to 20% are permissible (Code of Federal Regulations, 200 la).
29          The PMi0 inlet specified for the PM2 5 FRM is modified from a previous low flow-rate
30      PM10 inlet that was acceptable in both EPA-designated reference and equivalent PM]0 methods.
31      The modification corrects a flaw that was reported for the previous sampler, in that under some

        June 2003                                2-69       DRAFT-DO NOT QUOTE OR CITE

-------
                                          fractionator
                                            Oowntube
                                             WINS
                                             PM2.5
                                           •fractionator
       Figure 2-21.  Schematic diagram of the sample collection portion of the PM2^ FRM
                    sampler.
       Source: Noble etal. (2001).
 1     meteorological conditions, the inlet may allow precipitation to penetrate the inlet. The
 2     modification includes a larger drain hole, a one-piece top plate, and louvers. Tolocka et al.
 3     (200 la) evaluated the performance of this modified inlet in a series of wind tunnel experiments.
 4     The modified inlet was found to provide a size cut comparable to the original inlet, for both
 5     PM2 5 and PM10 sampling. Since the modification did not change the characteristics of the size
 6     cut, the modified inlet may be substituted for the original inlet as part of a reference or
 7     equivalent method for PM10 and PM2 5 (Tolocka et al., 2001a).
 8
 9           WINS Impactor. Design and calibration of the EPA PM2 5 Well Impactor Ninety-Six
10     (WINS) is given by Peters et al. (2001a). The WINS impactor was designed to be deployed
11     downstream of the Graseby-Andersen 246B PM]0 inlet as part of a sampler operating at a flow
12     rateof!6.7L/m.  The WINS is pictured in Figure 2-22. The PM25 inlet consists of a single jet,
13     directed toward a round hole, with a jet exit impaction surface comprised of a 37 mm diameter
       June 2003
2-70
DRAFT-DO NOT QUOTE OR CITE

-------
                                                 PM-10 Aerosol
                                                   ffomtntet
                                                            Nozzle
                                                             ColtectJoa cap
                                                            with ar&spill ring
                                                            ImpacSon surface:
                                                          tHter immersed in 1 mL
                                                            Dow Corning 704
                                                            diffusion pumpofl
                                                  PfiS-&6 aerosol to
                                                  sample collection
                                                      fflter
               Figure 2-22.  Schematic view of the final design of the WINS.
               Source: Peters et al. (2001a).
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
glass fiber filter immersed in 1 ml of low volatility diffusion pump oil (i.e., the well).  Particles
not having enough inertia to be removed by the impactor are captured downstream on the sample
collection filter. This design was selected to minimize impactor overloading that would
otherwise result in particle bounce.  The oil wicks through the particulate deposit on the
impactor to provide a continuously wetted surface for impaction. The penetration curve
indicated a 50% cutpoint of 2.48 urn Da with a geometric standard deviation of 1.18
(Figure 2-23).
     •The WINS separator was evaluated for its loading characteristics (Vanderpool et al., 2001)
by monitoring the performance after repeated operation in an artificially generated, high
concentration, coarse-mode aerosol composed of Arizona Test Dust, as well as in the field in
Rubidoux, Phoenix, Philadelphia, Research Triangle Park, and Atlanta. In the wind tunnel
experiments, the WINS performance was found to be a monotonic function of loading. A minus
5% bias in the PM25 measurement resulted from a coarse particulate loading of approximately
16 mg. This negative bias was due to a slight reduction in the separator cutpoint. It was also
       June 2003
                                          2-71
DRAFT-DO NOT QUOTE OR CITE

-------
                                   100
                                   80
                                   60
                                                       WINS Final Version  ,
                                                      	 WINS Best Fit Line
                                                       •£v   Aeioiizer Detector
                                                       X   Fluorometer Detector
                                    20
                                       1.0        2.0        3.0       4.0
                                            Aerodynamic Diameter (pm)

              Figure 2-23. Evaluation of the final version of the WINS.

              Source:  Peters et al. (2001a).
 1      found that the predictable results from the controlled laboratory experiments could not be
 2      extrapolated to field settings and that the WINS performance was more sensitive to the impactor
 3      loading in the field than it was in experiments with the single component aerosol.  Significant
 4      particle bounce was not observed in either the laboratory or the field experiments. Vanderpool
 5      et al. (2001) concluded that their study supports the recommendation that the FRM WINS wells
 6      should be replaced after every 5 days of 24-h operation (U.S. Environmental Protection Agency,
 7      1998).
 8           A detailed sensitivity study of the WINS impactor was conducted (Vanderpool et al., 2001)
 9      in which the effects on the impactor performance of a number of parameters were examined.
10      The results of this study are summarized in Table 2-3.
11           The regulations also allow for Class I, II, and III equivalent methods for PM2 5 (Code of
12     Federal Regulations, 2001c). Class I equivalent methods use samplers with relatively small
13      deviations from the sampler described in the FRM. Class II equivalent methods include "all
14     other PM2 5 methods that  are based upon 24-h integrated filter samplers that are subjected to
        June 2003
2-72
DRAFT-DO NOT QUOTE OR CITE

-------
              TABLE 2-3. SUMMARY OF SENSITIVITY STUDIES OF WINS IMPACTOR
                                           PERFORMANCE
Parameter
Manufacturing tolerances
on WINS components
Flow control biases
T and P measurement
Diffusion oil volume
Impactor loading
Ambient P variations
Air Properties
Impactor oil crystallization
Impactor oil viscosity
Amount of variance
Specified tolerances
4%
Allowable ambient
0.75 ml to 3 ml
After 5 24 h events

OC

-20 C
-35 C
Outpoint variation
0.05 micrometers
0.05 micrometers
+ 0.02 micrometers
No effect
-0.07 micrometers
Negligible
2.40 micrometers
No effect
No effect
PMZ5 mass
concentration bias
<1%
Outpoint shift partially offset
volume bias
+ 0.4%

<1.5%
Negligible
NA
No effect
No effect
Need to change WINS more frequently than every
5 days
         Source: Vanderpool etal. (2001).
  1
  2
  3
  4
  5
  6
  7
  8
  9
10
11
12
13
14
subsequent moisture equilibration and gravimetric mass analysis." Class III equivalent methods
include non-filter-based methods such as beta attenuation, harmonic oscillating elements, or
nephelometry (McMurry, 2000). As of July 2001,11 PM25 samplers (listed in Table 2-4) had
been tested and led to the conclusion that the PM10 sampling systems can be designed such that
concentration measurements are precise to ±  10%. For PM25, cutpoint tolerances are not
expected to affect the mass concentration as much as for PMIO, because the 2.5 urn Da cutpoint
generally occurs near a minimum in the mass distribution (e.g., Figure 2-5).
     The PM2 5 mass concentration will be affected, on the other hand, by other sampling issues
mentioned but not discussed extensively  in the previous 1996 PM AQCD (U.S. Environmental
Protection Agency, 1996a). These issues have been discussed earlier in this chapter and include
gas/particle, particle/particle, and particle/substrate interactions for sulfates and nitrates (e.g.,
Appel et al., 1984), volatilization losses of nitrates (Zhang and McMurry, 1992), semivolatile
organic compound artifacts (e.g., Eatough et al., 1993), and relative humidity effects (e.g.,
Keeleretal., 1988).
       June 2003
                                         2-73
DRAFT-DO NOT QUOTE OR CITE

-------
            TABLE 2-4. PM, < SAMPLERS CURRENTLY DESIGNATED AS FRMs FOR
                                PM,. MASS CONCENTRATIONS
Sampler
RAAS2.5-100
RAAS2.5-300
RAAS2.5-200
Partisol 2000
Partisol-Plus 2025
Partisol 2000 audit
PQ200
PQ 200A
605 CAPS
MASS 100
MASS 300
Manufacturer
Andersen Instruments
Andersen Instruments
Andersen Instruments
Rupprecht & Patashnick
Rupprecht & Patashnick
Rupprecht & Patashnick
BGI, Inc.
BGI, Inc.
ThermoEnvironmental Instruments
URC
URC
Description
FRM single
FRM sequential
FRM audit
FRM single
FRM sequential
FRM audit
FRM single
FRM audit
FRM single
FRM single
FRM sequential
Federal Register Reference
Vol. 63, p. 3 199 1,6/1 1/98
Vol.63,p.3l991,6/ll/98
Vol. 64, p. 12167,3/11/99
Vol. 63, p. 18911,4/16/98
Vol. 63, p. 18911,4/16/98
Vol. 64, p. 19153,4/19/99
Vol. 63, p. 18911,4/16/98
Vol. 63, p. 18911,4/16/98
Vol. 63, p. 58036, 10/29/98
Vol. 65, p. 26603, 05/08/00
Vol. 65, p. 26603, 05/08/00
        Source: Peters et al. (2001b); U.S. Environmental Protection Agency (2001).
 1          Several studies now have been reported in which the FRM was collocated with other PM2 5
 2     samplers in intercomparison studies. During the Aerosol Research and Inhalation Epidemiology
 3     Study (ARIES), several PM25 samplers were collocated at a mixed industrial-residential site near
 4     Atlanta, GA (Van Loy et al., 2000).  These samplers included a standard PM2 5 FRM, a TEOM
 5     with Nation drier, a participate composition monitor (PCM; Atmospheric Research and
 6     Analysis, Cary, NC), a medium-volume (113 L/min flow rate) fine particle (PM25) and
 7     semivolatile organic compound (i.e., a filter followed by a solid adsorbent) sampler, operated by
 8     the Desert Research Institute, a HEADS sampler, and a dichotomous sampler for coarse PM.
 9     The PCM sampler has three channels, all of which have PM10 cyclone inlets. The first two
10     channels both have two denuders preceding a 2.5-um WINS impactor and filter packs. The first
11     denuder is coated with sodium carbonate to remove acid gases, and the second is coated with
12     citric acid to remove ammonia. The third channel has a carbon coated parallel-plate denuder
13     preceding the WINS impactor. Measurements of 24-h mass from the FRM, PCM, and TEOM
14     samplers, as well as reconstructed PM25 mass (RPM), were compared for a 12-mo period. The
15     slopes for the TEOM-FRM, PCM-FRM, and RPM-FRM correlations were 1.01, 0.94, and 0.91,
       June 2003
2-74
DRAFT-DO NOT QUOTE OR CITE

-------
  1     respectively; whereas the y-intercepts for each were 0.68, 0.04, and 0.98. Participate sulfate
  2     measurements on the FRM Teflon filter, the PCM Teflon filter, and PCM Nylon filter were
  3     nearly identical. Nitrate results from the three filters were much less consistent, with the FRM
  4     collecting substantially less nitrate than that collected on either the denuded nylon filter or a
  5     denuder followed by a Teflon-nylon filter sandwich. Particulate ammonia measurements were
  6     also compared and showed more scatter than the sulfate measurements but less than the nitrate
  7     measurements.
  8          An intercomparison of both PM10 and PM25 mass measurements was conducted during the
  9     1998 Baltimore PM Study (Williams et al., 2000). PM monitors were collocated at a residential
 10     indoor, residential outdoor, and ambient monitoring site within Baltimore County, MD. PM
 11     samplers included TEOMs, PM2 5 FRMs, cyclone-based inlets manufactured by University
 12     Research Glassware (URG), and Versatile Air Pollution Samplers (VAPS). The VAPS sampler
 13     is a dichotomous sampler operating, at 33 L/miri (one coarse particle channel at 3 L/min, and two
 14     fine particle channels at 15 L/min, each). In the configuration employed during this study, one
 15     fine particle channel was operated with a Teflon filter backed by a nylon filter and preceded by a
 16 .    sodium carbonate coated annular denuder; the second fine particle channel had a quartz filter
 17     preceded by a citric acid-coated annular denuder; and the coarse particle channel had a
 18     polycarbonate filter followed by a Zefluor filter for flow distribution. Differences in PM2 5 mass
 19     concentrations between the samplers, although not large, were attributed to potential particle
 20     nitrate  losses, denuder losses, and losses of SVOC for some samplers. Differences between
 21      coarse paniculate mass concentrations, on the other hand, varied widely between the
 22      instruments.
 23           In another intercomparison study, Tolocka et al. (2001b) examined the magnitude of
 24      potential sampling artifacts associated with the use of the FRM by collocating FRMs alongside
 25      other chemical speciation samplers at four U.S. cities. The locations included a high nitrate and
 26      carbon, low sulfate site (Rubidoux, CA); high crustal, moderate carbon and nitrate site
 27      (Phoenix); high sulfate, moderate carbon, and low nitrate (Philadelphia); and low PM25 mass
 28      (Research Triangle Park, NC). The use of Teflon and heat-treated quartz filters was also
29      examined in this study.  The Teflon filters collected less nitrate than the heat-treated quartz
30      filters.  Filters in samplers using denuders to remove organic gases collected less organic PM
31      than filters in samplers without denuders.

        June 2003                                 2-75        DRAFT-DO NOT QUOTE OR CITE

-------
 1           Peters et al. (2001 b) compiled the results of several field studies in which a number of
 2      FRM and other PM2 5 samplers were intercompared.  In addition to the FRM samplers listed in
 3      Table 2-3, other PM2.5 samplers included the Sierra Instruments dichotomous sampler, the
 4      Harvard impactor, the IMPROVE sampler, and the Air Metrics saturation monitor.  Results were
 5      compiled from PM2 5 field studies conducted in Birmingham, Denver, Bakersfield, Phoenix,
 6      Research Triangle Park, Atlanta,  and Rubidoux. Limited studies on precision for the non-FRM
 7      samplers suggest that the Harvard Impactor and dichotomous samplers had the lowest coefficient
 8      of variations (CV), with both under 10%. The CV forthis study was calculated by dividing the
 9      sample standard deviation by the average concentration. The IMPROVE samplers had CVs
10      between 10 and 12%, and the Air Metrics samplers had the highest observed CVs, over 15%.
11      In intercomparisons with FRM samplers, the non-FRM samplers showed strong linear
12     relationships in comparison to the FRM sampler; however, none of the comparisons passed the
13      current EPA Subpart C equivalent method criteria, which EPA is in the process of revising.
14          Detailed information on precision of PM samplers used in monitoring networks may be
15     found in EPA's Technology Transfer Network website (U.S. Environmental Protection Agency,
16     2002a).
17
18     2.2.4.1.3    PMlft.2.5
19          Measurement techniques for PM10_2 5 are somewhat more complex than those for PM2 5 or
20     PM10 because, for PM10_2 5, it is necessary to isolate a size fraction between an upper 50% cut
21     point of 10 um Da and a lower 50% cut point of 2.5 urn Da. EPA is currently developing an
22     FRM for PM10_25. Several candidate techniques are discussed below.
23
24          The Difference Method. One approach to measurement of PMI0.25 is to make separate
25     measurements of PMi0 and PM2.5 and take the difference of the resulting equilibrated masses.
26     One problem is that, if either the PM2.5 or the PM10 sampler fails, no PM10_25 measurement can be
27     obtained. In addition, errors in cut-point, flow rate,  and filter weights (both before use and after
28     collection and equilibration of particles) and uncertainties due to loss of semivolatile
29     components of PM may occur for each size cut. Careful control of flow rate and equivalent
 30     treatment of PM10 and PM2 5 filters in terms of pressure drop across the filter and temperature of
 31     the filter during and after collection can improve precision and accuracy.  Allen et al. (1999a)

        June 2003                                2-76        DRAFT-DO NOT QUOTE OR CITE

-------
  1     summarized several sampling issues to consider in measuring coarse particulate mass by
  2     difference, including the use of identical instrumentation (except outpoints), filter media, filter
  3     face velocity, and ambient-filter temperature differences; common flow measurement devices;
  4     use of higher sampler flow rates (10 L/min minimum for 24-h sample is recommended); and
  5     avoiding excessive filter loading. The concern, expressed by Allen et al. (1999a), that the "pie-
  6     plate" inlet required by the final version of the PM2 5 FRM might have a different cut point than
  7     the "flat-top" inlet of the PM10 FRM, has been addressed by a wind tunnel study which found
  8     both to have an appropriate PM10 cut point (Tolocka et al., 2001 a).
  9           Since the difference method requires weighing two filters, the key to obtaining high
 10     precision in the coarse mass measurement is precise measurements of filter weights. Allen et al.
 11     (1999a) discuss techniques for increasing the precision of the difference method by reducing
 12     errors in filter weights. These include proper temperature and humidity controls,  use of a high
 13     quality microbalance, 100% replicate weighings, control of static charge, aging of new filters,
 14     weighing of a sufficient number of laboratory blank filters, and accounting for buoyancy  errors
 15     caused by variability in barometric pressure. Allen et al. (1999a) emphasize the necessity of
 16     replicate weighing of filters and a third weighing if the difference between the first two weights
 17     exceeds a specified minimum. Lawless et al. (2001) investigated the magnitude of uncertainties
 18     attributed to fluctuations in some of these parameters (humidity, temperature, drafts, vibration,
 19     and electrostatic charges) and recommended methods for improving their control. Koistinen
 20     et al. (1999) and Hanninen et al. (2002) give a excellent discussion of the procedures developed
 21      to overcome problems associated with gravimetric measurements of PM25 mass in the EXPOLIS
 22     (The Air Pollution Exposure Distributions  Within Adult Urban Populations in Europe) Study.
 23      They discuss factors such as corrections for buoyancy, elimination of static charge, and increases
 24      in the mass of blank filters with time. The establishment of a temperature and humidity
 25      controlled room required for the equilibration and weighing of filters for the FRM is expensive.
 26      Allen et al. (2001) describe a more cost-effective technique that uses a chamber with relative
 27      humidity controlled at 34% relative humidity by a saturated aqueous solution of MgCl2.
28           Allen et al. (1999a) recommend that, in reporting precision from collocated samplers both
29      the (CV) and the square of the correlation coefficient (r2) be reported. For a study in Boston with
30      27 pairs  of mass data from collocated PM10 and PM25 using standard weighing methods, they
31      obtained a CV of 4.7% and an r2 of 0.991 for PM25, a CV of 4.4% and an r2 of 0.994 for PM]0,

        June 2003                                 2-77        DRAFT-DO NOT QUOTE OR CITE

-------
 1     and a CV of 15% and an r2 of 0.88 for PM10_2 5. By using duplicate weighings and other
 2     techniques suggested for improving precision, they obtained a CV of 1.3% and an r2 of 0.998 for
 3     PM2 5 in a study in Chicago with 38 collocated measurements. On the basis of the improvement
 4     in the CV for PM25, they estimate that use of the recommended techniques for PM10_25 by
 5     difference would have yielded a CV of 3.8% and an r2 of 0.98 if they had been applied in the
 6     Chicago  study.
 7
 8          Multistage Impaction. A second technique involves the use of impaction to isolate the
 9     size fraction between 2.5 and 10 um Da. In the impaction process, the air stream is first
10     accelerated through a small hole (nozzle) or slit. The air stream is directed so that it "impacts"
11     on a surface. Depending on the velocity and pressure of the air stream, particles smaller than a
12     certain size will follow the air stream around the impactor surface.  Larger particles will impact
13     on the surface. In practice, impactors have 50% cut points similar to those for the rejection of
14     larger particles in PM2 5 and PMi0 samples (see Figure 2-6).
15           Multistage impactors are used to separate particles into several size fractions for the
16     determination of mass and chemical composition as a function of size (Wang and John, 1988;
17     Marple et al., 1991). The major problem with the use of impactors to separate the 10-2.5 urn Da
18     fraction of coarse particles (thoracic coarse PM) is bounce.  Coarse particles tend to be dry, solid
19     particles. When they hit a hard surface, they can bounce and be carried away with the air stream
20     (e.g., Dzubay et al., 1976; Wesolowski et al., 1977; Rao and Whitby, 1978; Cheng and Yen,
21     1979; Wang and John, 1987; John and Sethi, 1993).  Various techniques have been used to
22     reduce bounce. One technique is to use a porous substance such as a glass or quartz fiber filter
23     (Chang et al.,  1999) material or a polyurethane foam (Breum, 2000; Kavouras and Koutrakis,
24     2001). These techniques may result in less precise separation and yield a sample that must be
25     extracted before chemical analyses can be performed. Another technique is to coat the impactor
26     with a soft wax or grease (Rao and Whitby, 1977; Turner and Hering, 1987; Pak et al., 1992).
27     This can cause problems with weighing and chemical analyses. In addition, as the deposit of
28     particles builds up,  incoming particles may not hit the soft surface, but  instead hit a previously
29     collected hard particle and bounce off of it. The WINS impactor discussed earlier uses a filter in
30     a well of low volatility oil to ensure a wetted surface at all times. However, such a technique,
        June 2003                                2-78        DRAFT-DO NOT QUOTE OR CITE

-------
   1     while appropriate for removing unwanted particles, would not yield a particle sample suitable for
   2     weighing or for chemical analyses.
   3
   4          Virtual Impaction. The problems of bounce and blow off of particles from impactors,
   5     especially for the collection of large quantities of particles, was addressed by aerosol scientists in
   6     the mid-1960s by the development of what is now known as 'Virtual" impaction (Hounam and
   7     Sherwood, 1965; Conner, 1966).
   8          In a virtual impactor, a hole is placed in the impaction plate just below the accelerating jet.
   9     Two controlled flows allow a fraction, e.g., 10% (or another predetermined fraction, typically 5 -
  10     20%), of the air containing the coarse particles to go through the hole and through a filter (minor
  11      flow). A 10% minor flow gives a coarse channel enrichment factor of 10.  The remaining
  12      fraction (e.g., 90% of the airflow) containing the fine particles follows a different path and  goes
  13      through a second filter (major flow). The upper cutpoint is usually set by the inlet (e.g.,
  14      10 urn DJ. The flow rates, pressures, and distance from the nozzle to the virtual impactor
  15      surface can be varied to direct particles with an Da greater than the lower cutpoint (i.e.,
  16      > 2.5 um) to go through the hole and be collected on the first filter and to direct smaller particles
  17      (i.e., < 2.5 ^im) to flow around the impactor be collected on the second filter.  Large particles
 18      "impact" into the hole with a small amount of the air flow.  The smaller particles follow the
 19      major air flow around the impactor plate.  This technique overcomes the problem of bounce.
 20      An example of the separation into fine and coarse particles is shown in Figure 2-24.
 21          The usefulness of this technique for collecting samples of fine and coarse particles for
 22     chemical analysis was recognized by EPA in the mid-1970s. A development program was
 23     undertaken leading to the development of the now well known  "dichotomous sampler*' (a virtual
 24     impactor that separates particles into two size fractions) and an associated XRF analyzer
 25     (Dzubay and Stevens, 1975;LooetaI.,  1976; Jaklevic etal., 1977;Dzubay etal., 1977). The
 26     dichotomous sampler was originally developed for use in the Regional Air Monitoring Study
 27     (RAMS), part of the Regional Air Pollution Study (RAPS), conducted in St. Louis, Missouri in
 28      the mid-1970s. Dichotomous samplers were operated at 10 RAMS sites from March 1975 to
29      March 1977; and 33,695 filters were collected and analyzed by  XRF with an overall sampling
30      efficiency of 97.25% (Strothmann and Schiermeier, 1979; Loo et al,  1976; Loo et al., 1978;
31      Dzubay, 1980; Lewis and Macias, 1980).  Dichotomous samplers were a novel concept at that

        June 2003                                2-79       DRAFT-DO NOT QUOTE OR CITE

-------
       Figure 2-24.  Schematic diagram showing the principle of virtual impaction. The initial
                    flow, Qo, is split into a minor flow, Q19 which carries the larger particles,
                    which impact into the hole, to the coarse particle filter and a major flow, Q2,
                    which carries the smaller particles, which can follow the airflow, to the fine
                    particle filter.
       Source: Loo et al. (1976).
 1      time. Concern over particle losses and other problems at cut point sizes smaller than 2.5 urn Da
 2      influenced the decision to choose 2.5 instead of 1.0 as the cut point diameter.
 3           Subsequent to the use of the dichotomous sampler in RAPS, considerable progress has
 4      been made in the theory and practice of designing virtual impactors, especially in how to reduce
 5      losses and provide a sharp cut (Masuda et al., 1979; Marple and Chien, 1980; Chen at al., 1985,
 6      1986; Loo and Cork, 1988). Now virtual impactors, with rectangular slits or round holes, are
 7      used (a) to provide cut point sizes as low as 0.15 urn Da and (b) to concentrate coarse,
 8      accumulation, and ultrafme mode particles for use in health studies (Solomon et al., 1983;
 9      Marple et al., 1990; Sioutas et al., 1994a,b,c). Dichotomous samplers were also used in a
10      national network to measure PM25 and PMIO_2.5 in the Harvard Six City Study (Spengler and
11      Thurston, 1983; Dockery et al., 1993) and the Inhalable Particulate Network (Suggs and Burton,
12      1983). A trichotomous high volume sampler has also been developed that provides samples of
13      particles less than 1.0 urn, 1.0 urn - 2.5 um, and 2.5 urn (Marple and Olsen, 1995). This
       June 2003
2-80
DRAFT-DO NOT QUOTE OR CITE

-------
   1
   2
   3
   4
   5
   6
   7
   8
   9
  10
  11
  12
  13
  14
  15
  16
  17
  18
  19
 20
 21
 22
 23
 24
 25
 26
 27
 28
29
30
31
  sampler was intended for study of the composition of the intermodal mass between 1.0 and
  2.5 urn Da.
       More recently, two dichotomous sequential PM air samplers were collocated with a manual
  Federal Reference Method (FRM) air sampler and operated for over one year at a waterfront site
  on Tampa Bay (Poor et ah, 2002). The FRM sampler was alternately configured as a PM2 5, then
  as a PM10 sampler. For the dichotomous sampler measurements, daily 24-h integrated PM2 5 and
  PMIO_25 ambient air samples were collected at a total flow rate of 16.71  min"1. As was the case in
  earlier versions of the dichotomous sampler, a virtual impactor split the air into flow rates of
  1.67 and 15.01 min'1 and collected PM10_25 and PM25 on 47-mm diameter PTFE® filters.
  Between the two dichotomous air samplers, the average concentration, relative bias and relative
  precision for PM25 were 13.3 pg m'3, 0.02% and 5.2% concentrations (n = 282), and for PMIO_2.5
  were 12.3 fig m'3, 3.9% and 7.7% (n = 282). FRM measurements were alternate day 24-h
  integrated PM2 5 or PM10 ambient air samples collected onto 47-mm diameter PTFE® filters at a
  flow rate of 16.71 min1. Between a dichotomous and a PM25 FRM air sampler, the average
  concentration, relative bias and relative precision were 12.4 ^g m"3, -5.6% and 8.2% (n = 43);
 and between a dichotomous and a PMIO FRM air sampler, the average concentration, relative
 bias and relative precision were 25.7 ug nv3, -4.0% and 5.8% (n = 102). The PM2.5 concentration
 measurement standard errors for two dichotomous and one FRM samplers were 0.95, 0.79 and
 1.02 ug m-3; and for PM10, the standard errors were 1.06,1.59, and 1.70  ug m'3.  The authors
 (Poor et al., 2002) concluded that their results  indicate that "the dichotomous samplers have
 superior technical merit" and demonstrate 'the potential for the dichotomous sequential air
 sampler to replace the combination of the PM2.5 and PM10 FRM air samplers, offering the
 capability of making simultaneous, self-consistent determinations of these participate matter
 fractions in a routine ambient monitoring mode."
     The dichotomous sampler provides two separate samples. However, a fraction of the
 smaller particles, equal to the minor flow, will go through the virtual impaction opening with the
 air stream and be collected on the coarse particle filter. In the dichotomous sampler used in the
 RAPS program, 10% of the fine particles were collected with the coarse particles. Thus, in order
to determine the mass or composition of the coarse particles, it is necessary to determine the
mass and composition of the fine particles and subtract the appropriate fraction from the mass or
composition of the particles collected on the coarse particle filter.  Allen et al. (1999b) discuss
       June 2003
                                         2-81
                                                            DRAFT-DO NOT QUOTE OR CITE

-------
 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
potential errors in the dichotomous sampler caused by uncertainties in the coarse mass channel
enrichment factor.  Virtual impactors have also been designed with a clean air jet in the center of
the round nozzle. This makes possible lower contamination of coarse particles by fine particles
but maintains low losses and sharp cuts (Chen and Yeh, 1986; Chien and Lundgren, 1993).  The
fine particle intrusion into the coarse particle sample can also be reduced by operating two
virtual impactors in series (Dzubay and Stevens, 1975).
     Aerosol physicists have also conducted theoretical and experimental investigations of
virtual impaction using slits instead of round holes (Forney et al, 1978,1982; RavenhaU et al.,
 1978; Masuda and Nakasita, 1988; Sioutas et al., 1994b, c, d; Ding and Koutrakis, 2000). The
 slit virtual impactor permits much higher flow rate than round hole virtual impactors and
 resolves problems  that occur with multihole virtual impactors (Marple et al., 1990; Fang et al.,
 1991).  The slit technique has been used to develop virtual impaction systems for concentrating
 particles in the size range 0.1 to 2.5 nm Da for exposure studies using animals and people
 (Sioutas et al., 1995a, b). The slit impactor can also be used to concentrate coarse particles for
 measurement (Misra et al., 2001) or exposure studies (Chang et al, 2002). It has also been
 possible to concentrate ultrafine particles (> 0.1 urn) by first separating ultrafme particles from
 larger particles, adding water vapor to saturate the air containing the ultrafine particles, cooling
 the air to cause supersaturation and growth of the ultrafine particles into the 1.0 - 4.0 urn size
 range, concentrating these particles with a slit virtual impactor, and heating the air to return the
 particles to their original size (Sioutas and Koutrakis, 1996; Sioutas et al, 1999; Sioutas et al,
 2000; Kim et al, 2001b,c; Geller et al, 2002).

 2.2.5   Speciation  Monitoring
 Speciation Network and Monitoring
       In addition to FRM sampling to determine compliance with PM standards, EPA requires
 states to conduct chemical speciation sampling primarily to determine source categories and
 trends (Code of Federal Regulations, 2001b). Source category apportionment calculations are
 discussed in Chapter 3. A PM,5 chemical Speciation Trends Network (STO) has been deployed
 that consists of 54 core National Ambient Monitoring Stations (NAMS) and approximately
  250 State and Local Air Monitoring Stations (SLAMS). In addition, over 100 IMPROVE
  (Interagency Monitoring of Protected Visual Environments) samplers located at regional
         June 2003
                                                  2-82
                                                        DRAFT-DO NOT QUOTE OR CITE

-------
   1      background and transport sites can be used to fulfill SLAMS requirements. The overall goal of
   2      the speciation program is "to provide ambient data that support the Nation's air quality program
   3      objectives" (U.S. Environmental Protection Agency, 1999). Information and reports on EPA's
   4      speciation monitoring program may be found on EPA's Technology Transfer Network at
   5      http://www.epa.gov/ttn/amtic/pmspec.html. The NAMS speciation sites will provide routine
   6      chemical speciation data that will be used to develop annual and seasonal aerosol
   7      characterization, air quality trends analysis, and emission control strategies. The SLAMS
   8      speciation sites will further support the NAMS network and provide information for
   9      development of State Implementation Plans (SIPs).
 10           At both NAMs and SLAMs sites, aerosol samples will be collected  for analysis of trace
 11      elements, ions (sulfate, nitrate, ammonium, sodium, and potassium), and total carbon. The
 12      NAMS speciation sites will operate on a 1  in 3 day schedule, with  10 of these sites augmented
 13      with continuous speciation analyses for everyday operation. The SLAMS speciation sites will
 14      generally operate on a 1 in  6 day basis; however, many sites may be operated on a 1 in 3 day
 15      basis in locations where increased data collection is  needed. There are several samplers that are
 16      suitable for use  in the NAMS/SLAMS network. These samplers include an inlet cutpoint with
 17      size cut characteristics comparable to the WINS FRM; proven denuder technology for nitrate;
 18      and sampler face velocity and sample volume similar to that of the  FRM.  The current samplers
 19      include three filters: (1) Teflon for equilibrated mass and elemental analysis by energy
 20      dispersive x-ray fluorescence (EDXRF), (2) a nitric acid denuded Nylon filter for ion analysis
 21     (ion chromatography), (3) a quartz fiber filter for elemental and organic carbon.  EC and OC are
 22     determined by thermal optical analysis via a modification of the NIOSH (National Institute for
 23     Occupational Safety and Health) method 5040 (Thermal Optical Transmission) [TOT]).
 24     However, no corrections are made for positive artifacts caused by adsorption on organic gases or
 25     the quartz filters or negative artifacts due to evaporation of semivolatile organic compounds
 26     from the collected particles.
 27          Since 1987, the IMPROVE network has provided measurements of ambient PM and
 28     associated light extinction in order to quantify PM chemical components that affect visibility at
29     Federal Class 1 areas that include designated national parks, national monuments, and
30     wilderness areas. Management of this network is a cooperative effort between U.S. EPA, federal
31      land management agencies,  and state governments. The IMPROVE program has established

        June 2003                                2-83        DRAFT-DO NOT QUOTE OR CITE

-------
 1      protocols for analysis of aerosol measurements that provide ambient concentrations for PM]0,
 2      PM2 5, sulfates, nitrates, organic and elemental carbon, crustal material, and a number of other
 3      elements.  Information on the IMPROVE program may be found at
 4      http://vista.cira.colostate.edu/improve.
 5           IMPROVE aerosol monitoring consists of a combination of particle sampling and sample
 6      analysis. The IMPROVE sampler, which collects two 24-hour duration samples per week,
 7      simultaneously collects one sample of PM10 on a Teflon filter, and three samples of PM25 on
 8      Teflon, nylon, and quartz filters. PM10 mass concentrations are determined gravimetrically from
 9      the PMjo filter sample, while PM2 5 mass concentrations are determined gravimetrically from the
10      PM2 5 Teflon filter sample. The PM2 5 Teflon filter sample is also used to determine
11      concentrations of selected elements using particle-induced x-ray emission (PIXE), x-ray
12      fluorescence (XRF), and Proton Elastic Scattering Analysis (PESA). The PM25 nylon filter
13      sample, which is preceded by a denuder to remove acidic gases, is analyzed to determine nitrate
14     and sulfate aerosol concentrations using Ion Chromatography (1C). Finally, the PM25 quartz
15      filter sample is analyzed for organic  and elemental carbon using the Thermal Optical Reflectance
16     (TOR) method. Corrections are made for positive artifacts but not for negative artifacts.
17          The STN and the IMPROVE networks represent a major advance in the measurement of
18     nitrate since the combination of a denuder to remove metric acid vapor and a Nylon filter to
19     adsorb nitric acid vapor that volatilizes from the collected ammonium nitrate particles
20     overcomes the loss of nitrate from Teflon filters. However, the different techniques used for the
21     measurement of OC and EC lead to significant differences between OC and EC measurements
22     when the two techniques are intercompared (Chow et al., 2001). IMPROVE yields higher EC
23     and lower OC although there is good agreement for TC. Another difference arises from the
24     treatment of the positive artifact due to the absorption of organic gases by the quartz filters used
25     in IMPROVE and STN samplers. More information on these differences is given in Section
26     2.2.7 and details are discussed in Appendix 2B.
27           Several of the PM25 size selectors developed for use in the EPA National PM2 5 STN were
28     recently evaluated by comparing their penetration curves under clean room experiments with
29     that of the WINS impactor (Peters et al., 2001c). The corresponding speciation monitors were
 30     then compared to the FRM in four cities. The PM25 inlets tested were the SCC 2.141 cyclone
 31      (6.7 L/min) that is in the Met One Instruments SASS sampler; the SCC 1.829 cyclone

        June 2003                               2-84        DRAFT-DO NOT QUOTE OR CITE

-------
  1     (5.0 L/min) that is proposed for use in the Rupprecht and Patashnik real-time sulfate/nitrate
  2     monitor; the AN 3.68 cyclone (24.0 L/min) that is in the Andersen RAAS; and the spiral
  3     separator (7.0 L/min) that was previously in the Met One SASS.  The outpoints of the SCC
  4     cyclones compared reasonably well with the WINS (2.52 and 2.44 micrometers for the SCC
  5     2.141 and SCC 1.829, respectively, at their design flowrates), but both demonstrated a tail
  6     extending into the coarse particle mode. The AN inlet had the sharpest cutpoint curve, but the
  7     50% cutpoint diameter was 2.7 urn Da at its design flowrate. The spiral inlet had the shallowest
  8     cutpoint curve, and the 50% cut point was 2.69 and 2.67 u.m Da for an ungreased and greased
  9     inlet, respectively. The speciation samplers were also compared to the FRM sampler with WINS
 10     inlet under ambient conditions in four cities. The Andersen RAAS equipped with the AN 3.68
 11     cyclone compared well to the FRM in all four cities, when compared on the basis of PM25 mass,
 12     sulfate, and crustal concentrations.  Greasing the spiral inlet in the Met One sampler improved
 13     the performance of that sampler, which tended to give much higher PM25 concentrations than the
 14     FRM in cities with high crustal particulate matter.
 15
 16     2.2.6   Inorganic Elemental Analyses
 17          In addition to the lighter elements, hydrogen, carbon, oxygen and nitrogen,  the following
 18     40 heavier elements are commonly found in ambient air samples: sodium, magnesium,
 19     aluminum, silicon, phosphorus, sulfur, chlorine, potassium, calcium, titanium, vanadium,
 20     chromium, manganese, iron, cobalt, nickel, copper, zinc, gallium, arsenic, selenium, bromine,
 21      rubidium, strontium, yttrium, zirconium, molybdenum, palladium, silver, cadmium, indium, tin,
 22     antimony, barium, lanthanum, gold, mercury, thallium, lead,  and uranium. These often indicate
 23      air pollution sources and several of them are considered to be toxic (transition metals,
 24      water-soluble metals, and metals in certain valence states [e.g., Fe(II), Fe(III), Cr(III),  Cr(VI),
 25      As(III), As(V)]). Measurement methods for inorganic elements are listed in Table 2-5. These
 26      methods differ with respect to detection limits, sample preparation, and cost (Chow, 1995).
 27      EDXRF and PIXE are the most commonly applied methods because they quantify more than
28      40 detectable elements, they are non-destructive, and they are relatively inexpensive. Both were
29      discussed in the previous 1996 PM AQCD.  TRXRF and S-XRF are newer techniques with
30      lower detection limits. AAS, ICP-AES, and ICP-MS are also appropriate  for ionic
31      measurements of elements that can be dissolved. PESA provides a means for measuring

        June 2003                                2-85        DRAFT-DO NOT QUOTE OR CITE

-------
             TABLE 2-5.  MEASUREMENT METHODS FOR INORGANIC ELEMENTS
#
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
Acronym
EDXRF
S-XRF
PIXE
PESA
TRXRF
INAA
AAS
ICP-AES
ICP-MS
SEM
Full Name
energy dispersive x-ray fluorescence
synchrotron induced X-ray emission
proton induced x-ray emission
proton (or particle) elastic scattering
analysis
total reflection X-ray fluorescence
instrumental neutron activation
analysis
atomic absorption spectrophotometry
inductively coupled plasma with
atomic emission spectroscopy
inductively coupled plasma with
mass spectroscopy
scanning electron microscopy
Comments
heavier elements
heavier elements; lower detection limits
than EDXRF
heavier elements; lower detection limits
than EDXRF
lighter elements
heavier elements; lower detection limits
than EDXRF
many elements; sensitivity different than
EDXRF
many elements that can be dissolved
many elements that can be dissolved
many elements that can be dissolved
heavier elements
 1     elements with lower atomic numbers from hydrogen to carbon. More detailed information on
 2     each technique is given in Appendix 2B.1.
 3
 4     2.2.7   Elemental and Organic Carbon in Particulate Matter
 5          Ambient particles from combustion sources contain carbon in several chemically and
 6     optically distinct forms. Health- and visibility-related studies of these particles require
 7     information about the relative contributions to total particle mass by these different forms of
 8     carbon. With the exception of carbonate-based carbon, however, a clear classification scheme
 9     has not yet been established to distinguish organic carbon, light-absorbing carbon, black carbon,
10     soot and elemental carbon. The absence of clear, physically-based definitions results in
11     confusion  in the interpretation of speciation data. For example, depending on the radiation
12     wavelength specified, "light absorbing" carbon can include compounds that volatilize without
13     oxidation. "Black" carbon includes various mixtures containing "elemental" (graphitic) carbon;
       June 2003
2-86
DRAFT-DO NOT QUOTE OR CITE

-------
  1     partially degraded, oxidized graphitic fragments; and partially-oxidized amorphous aromatic
  2     carbon. For studying visibility reduction, a measurement of light-absorbing carbon may be more
  3     useful than one of elemental carbon.  For source apportionment by receptor models, several
  4     consistent but distinct fractions of carbon in both source and receptor samples are desired,
  5     regardless of their light-absorbing or chemical properties.  Differences in ratios of the carbon
  6     concentrations in these fractions form part of the source profiles that distinguish the contribution
  7     of one source from the contributions of other sources (Watson et al., 1994a,b).
  8           Three method-dependent operational classes of carbon are commonly measured in ambient
  9     aerosol samples collected on quartz-fiber filters: (1) semi-volatile organic or non-visible light-
 10     absorbing carbon, termed "organic carbon (OC)"; (2) elemental carbon, soot, black carbon, or
 11     fight-absorbing carbon, termed "elemental carbon (EC)"; and (3) carbon present as K2CO3,
 12     Na2CO3, MgCO3, CaCO3, termed "carbonate carbon (CC)." The sum of OC, EC, and CC in PM
 13     gives the total carbon (TC).
 14           The thermal/optical reflectance (TOR), thermal/optical transmission (TOT), and thermal
 15     manganese oxidation (TMO) methods are most commonly used for the analysis of organic and
 16     elemental carbon in atmospheric  PM. In thermal separation methods, OC is vaporized and the
 17     EC remaining on the filter is then oxidized to CO2 and quantified by nondispersive infrared
 18     detection, by electrochemical techniques or by reducing the CO2 to CH4 and quantifying CH4 via
 19     flame ionization detection (FID).  OC that does not vaporize below 550 C can pyrolyze at higher
 20     temperatures to form additional black carbon. Thermal/optical methods must correct for this
 21      effect in order to correctly distinguish OC from EC. The various methods give  similar results for
 22      TC, but not for EC or OC, due  to differing assumptions regarding the thermal behavior of
 23      ambient aerosol carbon. These methods are discussed in detail in Appendix 2B.2.
 24           Carbonate carbon (i.e., K2CO3, Na2CO3, MgCO3, CaCO3) can be determined thermally, or
 25      on a separate filter section by measurement of the carbon dioxide (CO^ evolved upon
 26      acidification (Johnson et al.,  1980). It is usually on the order of 5% or less of TC  for ambient
 27      particulate samples collected in urban areas (Appel, 1993).
28           The forms of carbon present in natural materials that may be burnt to generate atmospheric
29      aerosol tend to be poorly defined.  Thus, the pyrolysis products of these materials  during
30      thermal/optical analysis cannot be predicted. The Geochemical  Society convened an
31      international steering committee in 1999 to define a set of representative black carbon and black

        June 2003                                2-87        DRAFT-DO NOT QUOTE OR CITE

-------
 1     carbon-containing benchmark materials to be used to support ambient aerosol sample analysis.
 2     These standard materials may be used to provide thermal/optical "fingerprints" for deducing
 3     primary aerosol sources, and to establish characteristic analytic interferences or artifacts
 4     associated with such sources. The materials recommended to date include n-hexane soot,
 5     lignocellulosic chars, soils, marine sediments, and the NIST urban dust standard reference
 6     material (SRM 1649a). The committee has also recommended a set of standard materials that
 7     may potentially interfere with black carbon analyses: shale, natural organic matter, melanoidin
 8     (an amino acid-based material) and coals.  These recommendations are discussed on the steering
 9     committee website: http://www.du.edu/~-dwismith/bcsteer.html.
10
11     2.2.8    Ionic Species
12          Ion chromatography (1C) is widely used for analyzing  ionic species in the water-soluble
13     portion of suspended PM. 1C is the method of choice for the measurement of sulfate, nitrate,
14     ammonium, sodium, and potassium ions for the NAMS program. Aerosol strong acidity, H+, is
15     determined by titration of a water solution of PM collected following a series of annular
16     denuders to remove acid and basic gases with back-up filters to collect NH3 and HNO3 that
17     volatilize from the PM during collection. The 1996 PM AQCD (U.S. Environmental Protection
18     Agency, 1996a) discussed measurement of ions by 1C (Section 4.3.3.1) and of strong acidity
19     (Sections 3.3.1.1 and 4.3.3.1); so, no further details will be discussed here.
20
21     2.2.9   Continuous Monitoring
22           The EPA expects that many local environmental agencies will operate continuous PM
23     monitors. All currently available continuous measurements of suspended particle mass share the
24     problem of dealing with semivolatile PM components. So as not to include particle-bound water
25     as part of the mass, the particle-bound water must be removed by heating or dehumidification.
26     However, heating also causes loss of ammonium nitrate and semivolatile organic components.
27     A variety of potential candidates for continuous measurement of particle mass and related
28     properties are listed in Table 2-6. These techniques are discussed in more detail in Appendix
29     2B.3.
30
31

        June 2003                                2-88        DRAFT-DO NOT QUOTE OR CITE

-------
      TABLE 2-6. METHODS FOR CONTINUOUS MEASUREMENT OF PM MASS,
                                       COMPONENTS, ETC.	

   #    Acronym            Name                                  Comments
    1.    TEOM
 Tapered Element
 Oscillating Microbalance
 Measures only nonvolatile components of PM.
   2.    TEOM    TEOM with Sample
        with SES   Equilibration System
  10.
                          By using dehumidification instead of heat for drying and a lower
                          temperature for stabilization, includes some of the semivolatile
                          components of PM.
3. — Differential TEOM
4. FDMS Filter Dynamics
Measurement System
. 5. RAMS Real-Time Total Ambient
Mass Sampler
6. CAMM Continuous Ambient
Mass Monitor
7. — Beta gauge
8. — Piezoelectric
microbalance
9. CCPM Continuous Coarse
Particle Monitor
Research instrument designed to measure both the nonvolatile
and semivolatile components of PM.
Commercial version of the differential TEOM.
Research instrument designed to measure total PM using
denuders and an adsorptive filter.
Measures total PM by pressure drop across a frequently
changed filter.
Measures PM mass by beta attenuation. Unless dried by heat or
dehumidification will also measure particle-bound water.
Measures mass by change in resonance frequency when particles
are deported on a crystal.
Virtual impaction is used to concentrate PM1MJ which is then
measured by a TEOM.
 Semi-continuous EC/OC
                                            Several commercially available instruments automate the thermal
                                            technique for EC/OC and provide hourly measurements.
  11.      —
 Semi-continuous nitrate
                                           Collection of PM followed by flash vaporization and
                                           determination of NOX provides 10 minute measurements of
                                           particulate nitrate.
  12.      —
Semi-continuous sulfate
                                           Several techniques are available in which particulate sulfate is
                                           measured using flame ionization detection.
  13.      —
Continuous ion
chromatography of
water-soluble ions
Particles are grown by mixing with water vapor, collected in
water, and injected into an automatic ion chromatography.
  14.
Mass spectroscopy of
individual particles
                                           Single particles are evaporated, ionized and the components
                                           analyzed by mass spectroscopy. Several different systems are in
                                           use in research studies.
  15.     EAD     Electrical Aerosol
                  Detector
                         This instrument measures charge collected by particles and gives
                         a continuous signal that is proportional to the integral of the
                         particle diameter.
 16.
Integrating nephelometer
                                           Light scattering by suspended particles, collected over a large
                                           solid angle, provides an indicator for particle mass including
                                           particle-bound water unless the air sample is dried by heating or
                                           dehumidification.
June 2003
                              2-89
                  DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.2.10  Measurements of Individual Particles
 2          Recently, several researchers have developed instruments for real-time in situ analysis of
 3     single particles (e.g., Noble and Prather, 1996; Card et al, 1997; Johnson and Wexler, 1995;
 4     Silva and Prather, 1997; Thomson and Murphy, 1994). Although the technique varies from one
 5     laboratory to another, the underlying principle is to fragment each particle into ions, using either
 6     a high-power laser or a heated surface and, then, a time-of-flight mass spectrometer (TOFMS) to
 7     measure the ion fragments in a vacuum. Each particle is analyzed in a suspended state in the air
 8     stream (i.e., without collection), avoiding sampling artifacts associated with impactors and
 9     filters. The technique is called aerosol time-of-flight mass spectrometry (ATOFMS).
10     By measuring both positive and negative ions from the same particle, information can be
11     obtained about the composition, of individual particles of known aerodynamic diameter. This
12     information is  especially useful in determining sources of particles. Noble and Prather (1996)
13     used ATOFMS to provide compositionally resolved particle-size distributions. Their instrument
14     is capable of analyzing size and chemical composition of 50 to 100 particles/min at typical
15     ambient concentrations and up to 600/min at high particle concentrations.  Four systems for
16     measurement of single particles using mass spectroscopy are reviewed by  Middlebrook (2002).
17     An example of the type of information that can be determined is shown in Figure 2-25.
18          Because  particles are analyzed individually, biases in particle sampling (the efficiency of
19     particle transmission into the sensor chamber as a function of size; particle size measurement,
20     and detection of particles prior to fragmentation)  represent a major challenge for these
21     instruments. Moreover, the mass spectrometer has a relatively large variability in ion yields
22     (i.e., identical samples would yield relatively large differences in mass spectrometer signals
23     [Thomson and Murphy, 1994]); therefore, quantitation is inherently difficult (Murphy and
24     Thomson, 1997).  Quantitation will be even more challenging for complex organic mixtures
25     because of the following two reasons: (1) a large number of fragments are generated from each
26     molecule, and  (2) ion peaks for organics can be influenced or obscured by inorganic ions
27     (Middlebrook  et al., 1998). Nonetheless, scientists have been successful in using these
28     techniques to identify the presence of organics in atmospheric particles and laboratory-generated
29     particles (i.e., as contaminants in laboratory-generated sulfuric acid droplets) as well as the
30     identification of specific compound classes such  as PAHs in combustion emissions (Castaldi and
31     Senkan, 1998; Hinz et al., 1994; Middlebrook et  al.,  1998; Murphy and Thomson, 1997;

       June 2003                               2-90        DRAFT-DO NOT QUOTE OR CITE

-------
                        0.2    0.3 0.40.5 0.60.70.80.91,0
                                                    r
                                                   2.0
                                                                              Soil
                                                                     3.0   4.0
                                        Aerodynamic Diameter (urn)
        Figure 2-25.  Size distribution of particles divided by chemical classification into organic,
                      marine, and crustal.
        Source: Noble and Prather (1998).
  1
  2
  3
  4
  5
  6
  7
  8
 9
10
11
12
13
Neubauer et al., 1998; Noble and Prather, 1998; Reilly et al., 1998; Silva and Prather, 1997).
A new multivariate technique for calibration of ATOFMS using microorifice impactors shows
promise for simplifying the calibration process (Fergenson et al., 2001). This calibration
technique has been applied to gasoline and diesel particles to demonstrate the feasibility of using
this technique for the source apportionment of gasoline and diesel particles in an atmospheric
mixture (Song et al., 2001).
     Until recently, ATOFMS systems have only been able to characterize particles that are
larger than approximately 0.2 to 0.3 \im in diameter.  Wexler and colleagues (Carson et al.,
1997; Ge et al., 1998) have developed an ATOFMS instrument that is able to size, count, and
provide chemical composition on individual particles ranging in size from 10 run to 2 urn.
       June 2003
                                          2-91
DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.2.11   Low Flow Filter Samples for Multiday Collection of
 2              Particulate Matter
 3          For some purposes, such as demonstrating attainment of an annual standard or as an
 4     exposure indicator for epidemiologic studies of chronic health effects, 24-h measurements are
 5     not essential. Annual or seasonal averages may be adequate. Multiday sampling techniques can
 6     result in lower costs for weighing, chemical analysis, and travel time to change filters. The
 7     multiday sampler serves a second purpose. Most commercially available samplers are optimized
 8     for collecting 24-h samples of the PM concentrations found in the U.S., Europe, or Japan. Many
 9     cities in other parts of the world have significantly higher PM concentrations.  Under these
10     conditions, the 16.7 L/min flow through 37 or 47 mm diameter filters may overload the filter and
11     prevent the sampler from maintaining the prescribed flow rate for 24 h. A low flow sampler
12     with a 0.4 L/min flow rate and a 47 mm diameter filter has been designed by Aerosol Dynamics,
13     Inc. With this sampler, the sample collection time can be chosen to suit the ambient
14     concentration level. This sampler, with a one-week collection period, has been used to
15     characterize  PM2 5 in Beijing, PRC (He et al., 2001). With a two-week collection period, it is
16     being used in a chronic epidemiologic  study in southern California, USA (Gauderman, et al.,
17     2000).
18          The sampler, as described by He et al. (2001), has three PM2.5 channels.  One channel
19     collects PM  on a Teflon filter for gravimetric mass measurement and elemental analysis by XRF.
20     A second channel collects PM on a quartz filter for organic and elemental carbon analysis.
21     A denuder to remove organic gases and a backup filter to collect semivolatile organic
22     compounds may be added. The third channel uses a carbonate denuder to remove acid gases
23     (HNO3 and SO2), a Teflon filter to collect PM for analysis of ions by ion chromatography, and a
24     nylon fitter to collect volatilized nitrate.  The Teflon filter can also be weighed prior to
25     extraction. Thus, the multiday sampler can provide the information needed for source
26     apportionment by Chemical Element Balance techniques (Watson et al., 1990a,b; U.S.
27     Environmental Protection Agency, 2002b).
28          For monitoring  sites with high day-to-day variability  in PM concentrations, a sample
29     integrated over a week may provide a more accurate measurement of the annual average than
30     can be obtained by l-in-3 or l-in-6 day sampling schedules. Daily PM data from Spokane, WA
31     were resampled to simulate common sampling schedules and the error due to less-than-everyday
32     sampling was computed (Rumburg et al., 2001). The error in the annual mean concentration for
       June 2003                               2-92        DRAFT-DO NOT QUOTE OR CITE

-------
 1      PM2 5, expressed as a percentage difference from the everyday sampling mean, was 1.7, 3.4, and
 2      7.7% for l-in-2 day, l-in-3 day, and l-in-6 day sampling, respectively.
 3
 4
 5      2.3  SUMMARY AND KEY POINTS
 6      2.3.1   Atmospheric Physics and Chemistry of Particles
 7           Atmospheric particles originate from a variety of sources and possess a range of
 8      morphological, chemical, physical, and thermodynamic properties. The composition and
 9      behavior of aerosols are linked with those of the surrounding gas.  Aerosol may be defined as a
10      suspension of solid or liquid particles in air and includes both the particles and all vapor or gas
11      phase components of air. However, the term aerosol often is used to refer to the suspended
12      particles only.
13           A complete description of the atmospheric aerosol would include an accounting of the
14      chemical composition, morphology, and size of each particle, and the relative abundance of each
15      particle type as a function of particle size. Recent developments in single particle analysis
16      techniques are bringing such a description closer to reality.
17           The diameter of a spherical particle may be determined geometrically, from optical or
18      electron microscopy, by light scattering.and Mie theory, or by a particle's behavior (e.g.,
19      electrical mobility or its aerodynamic behavior). However, the various types of diameters may
20      be different, and atmospheric particles often are not spherical. Therefore, particle diameters are
21      described by an "equivalent" diameter. Aerodynamic diameter, Da (the diameter of a unit
22      density sphere that would have the same terminal settling velocity as the real particle), and the
23      Stokes diameter, Dp (the diameter of a sphere of the same density as the particle that would have
24      the same aerodynamic resistance or drag), are the most widely used equivalent diameters.
25           Atmospheric size distributions show that most atmospheric particles are  quite small, below
26      0.1 jim; whereas most of the particle volume (and therefore most of the mass) is found in
27      particles greater than 0.1 Jim.  An important feature of the mass or volume size distributions of
28      atmospheric particles is their multimodal nature. Volume distributions, measured in ambient air
29      in the United States, are almost always found to have a minimum between 1.0 and 3.0 urn. That
30      portion of the size distribution that contains particles that are mostly larger than the minimum is
31      called "coarse" particles or the "coarse" mode.  That portion of the size distribution that contains

        June 2003                                2-93        DRAFT-DO NOT QUOTE OR CITE

-------
 1     particles that are mostly smaller than the minimum is called "fine" particles and includes several
 2     modes. "Accumulation-mode" refers to that portion of fine particles with diameters above about
 3     0.1 urn. That portion of fine particles with diameters below 0.1 |im are usually called "ultrafine"
 4     by toxicologists and epidemiologists and "nanoparticles" by aerosol physicists and material
 5     scientists. In the number distribution of ultrafine particles, particles in the size range below 0.01
 6     are called the nucleation mode and particles between 0.01 and 0.1 are  called the Aitken mode.
 7     The Aitken mode can be observed as a separate mode in mass or volume distributions only in
 8     clean or remote areas or near sources of new particle formation by nucleation.
 9           The aerosol community uses three different approaches or conventions in the classification
10     of particles by size: (1) modes, based on the observed size distributions and formation
11     mechanisms; (2) cut point, usually based on the 50% cut point of the specific sampling device,
12     including legally specified, regulatory sizes for air quality standards; and (3) dosimetry or
13     occupational sizes, based on the entrance into various compartments of the respiratory system.
14     Over the years, the terms fine and coarse as applied to particle sizes have lost the original precise
15     meaning of fine and coarse.  In any given article, therefore, the meaning of fine and coarse,
16     unless defined, must be inferred from the author's usage. In particular, PM25 and fine particles
17     are not equivalent. PM25 refers to the aggregate sample of PM that is collected following a size-
18     selective inlet with a specified penetration as a function of size and a 50% cutpoint of 2.5 urn Da.
19     It may also be used to refer to number (or other measure of particles suspended in the
20     atmosphere that would be collected by such a sampler).  PM10 is defined similarly. PMia.2 5
21     refers to the sample that would be collected if the PM25 component could be removed from a
22     PM10 sample.
23           Several processes influence the formation and growth of particles. New particles may be
24     formed by nucleation from gas phase material. Particles may grow by condensation as gas phase
25     material condenses onto existing particles. Particles may also grow by coagulation as two
26     particles combine to  form one.  Gas phase material condenses preferentially on smaller particles,
27     and the rate constant for coagulation of two particles decreases as the particle size increases.
28     Therefore, nuclei mode particles grow into the accumulation mode, but growth of accumulation
29     mode particles into the coarse mode is rare.
30           The major constituents of atmospheric PM are sulfate, nitrate, ammonium, and hydrogen
31      ions; particle-bound  water; elemental carbon; a great variety of organic  compounds; and crustal

        June 2003                                 2-94        DRAFT-DO NOT QUOTE OR CITE

-------
   1      material. Atmospheric PM contains a large number of elements in various compounds and
  2      concentrations and hundreds of specific organic compounds. Paniculate material can be primary
  3      or secondary. PM is called primary if it is in the same chemical form in which it was emitted
  4      into the atmosphere. PM is called secondary if it is formed by chemical reactions in the
  5      atmosphere.  Primary coarse particles are usually formed by mechanical processes; whereas
  6      primary fine particles are emitted from sources either directly as particles or as vapors that
  7      rapidly condense to form particles.
  8           Most of the sulfate and nitrate and a portion of the organic compounds in atmospheric
  9      particles are secondary.  Secondary aerosol formation depends on numerous factors including the
 10      concentrations of precursors; the concentrations of other gaseous reactive species such as ozone,
 11      hydroxyl radical, peroxy radicals, and hydrogen peroxide; atmospheric conditions, including
 12      solar radiation and relative humidity; and the interactions of precursors and preexisting particles
 13      within cloud or fog droplets or on or in the liquid film on solid particles. As a result, it is
 14      considerably more difficult to relate ambient concentrations of secondary species to sources of
 15      precursor emissions than it is to identify the sources of primary particles.
 16          The lifetimes of particles vary with particle size.  Coarse particles can settle rapidly from
 17      the atmosphere within minutes or hours and normally travel only short distances. However,
 18      when mixed high into the atmosphere, as in dust storms, the smaller-sized, coarse-mode particles
 19      may have longer lives and travel greater distances.  Accumulation-mode particles are kept
 20      suspended by normal air motions and have a lower deposition velocity than coarse-mode
 21      particles. They can be transported thousands of kilometers and remain in the atmosphere  for a
 22     number of days.  Accumulation-mode particles are removed from the atmosphere primarily by
 23      cloud processes.  Dry deposition rates are expressed in terms of a deposition velocity that varies
 24      with the particle size, reaching a minimum between 0.1 and 1.0 urn aerodynamic diameter.
 25           PM is a factor in acid deposition.  Particles serve as cloud condensation nuclei and
 26      contribute directly to the acidification of rain. In addition, the gas-phase species that  lead to dry
 27      deposition of acidity are also precursors of particles. Therefore, reductions in S02 and NOX
 28      emissions will decrease both acid deposition and PM concentrations.  Sulfuric acid, ammonium
29      nitrate, and organic particles also are deposited on surfaces by dry deposition and can contribute
30      to ecological effects.
31
       June 2003                                 2-95        DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.3.2    Measurement of Atmospheric Particles
 2          The decision by the EPA to revise the PM standards by adding daily and yearly standards
 3     for PM2 5 has led to a renewed interest in the measurement of atmospheric particles and to a
 4     better understanding of the problems in developing precise and accurate measurements of
 5     particles. Unfortunately, it is very difficult to measure and characterize particles suspended in
 6     the atmosphere.
 7          PM monitoring is used to develop information to guide implementation of standards (i.e.,
 8     to identify sources of particles; to determine whether or not a standard has been attained; and to
 9     determine health, ecological, and radiative effects).  Federal Reference Methods (FRM) specify
10     techniques for measuring PMi0 and PM25. Particles are collected on filters and mass
11     concentrations are determined gravimetrically. The PM10 FRM sampler consists of a PM10
12     inlet/impactor and a 47-mm Teflon filter with a particle collection efficiency greater than 99.7%.
13     The PM2 5 FRM is similar except that it includes a PM2 5 impactor with an oil-covered impaction
14     substrate to remove particles larger than 2.5  urn. Both techniques provide relatively precise
15     (±10%) methods for determining the mass of material remaining on a Teflon filter after
16     equilibration. Despite considerable progress in measuring the atmospheric PM mass
17     concentration, numerous uncertainties continue to exist as to the relationship between the mass
18     and composition of material remaining on the filter as measured by the FRMs and the mass and
19     composition of material that exists in the atmosphere as suspended PM. There is no reference
20     standard for particles suspended in the atmosphere, nor is there an accepted way to remove
21     particle-bound water without losing some of the semivolatile components of PM such as
22     ammonium nitrate and semivolatile organic  compounds. It also is difficult to cleanly separate
23     fine and  coarse PM. As a result, EPA defines accuracy for PM measurements in terms of
24     agreement of a candidate sampler with a reference sampler. Therefore, intercomparisons of
25     samplers become very important in determining how well various samplers agree and how
26     various design choices influence what is actually measured.
27           Current filtration-based mass measurements lead to significant evaporative losses of a
28     variety of semivolatile components (i.e., species that exist in the atmosphere in dynamic
29     equilibrium between the condensed phase and gas phase) during and possibly after collection.
30     Important examples include ammonium nitrate, semivolatile organic compounds, and particle-
31     bound water. Loss of these components may significantly affect the quality of the measurement

       June 2003                                2-96        DRAFT-DO NOT QUOTE OR CITE

-------
  1     and can lead to both positive and negative sampling artifacts. Negative artifacts resulting from
  2     loss of ammonium nitrate and semivolatile organic compounds may occur during sampling
  3     because of changes in temperature, relative humidity, or composition of the aerosol or because
  4     of the pressure drop across the filter. Negative artifacts also may occur during handling and
  5     storage because of evaporation.  Positive artifacts occur when gas-phase compounds (H2O,
  6     HNO3, SO2, and organic compounds) absorb onto or react with filter media or collected PM or
  7     when some particle-bound water is not removed.
  8          Sampling systems for semivolatile PM components make use of denuders to remove the
  9     gas-phase fraction and absorptive filters to remove the condensed phase and retain any material
 10     that subsequently evaporates from the collected PM. The loss of particulate nitrate may be
 11     determined by comparing nitrate collected on a Teflon filter to that collected on a nylon filter
 12     (which absorbs nitric acid which evaporates from ammonium nitrate particles) preceded by a
 13     denuder to remove gas-phase nitric acid. In two studies in southern California, the PM2 5 mass
 14     lost because of volatilization of ammonium nitrate was found to represent 10 to 20% of the total
 15     PM25 mass and almost a third of the nitrate. Denuder/absorptive filter sampling systems also
 16     have been developed for measuring particulate phase organic compounds. This technique is an
 17     improvement over the filter/adsorbent collection method. However, the denuder systems
 18     currently discussed in the literature are not straightforward in their use, and the user must have a
 19     thorough understanding of the technology. The FRM for PM2 5 will likely suffer loss of
 20     particulate nitrates and semivolatile organic compounds, similar to the losses experienced with
 21      other single filter collection systems.
 22          It is generally desirable to collect and measure ammonium nitrate and semivolatile organic
 23      compounds as part of particulate matter mass. However, it is usually desirable to remove the
 24     particle-bound water before determining the mass.  In some situations, it may be important to
 25      know how much of the suspended particle's mass or volume results from particle-bound water.
 26      Calculations and measurements indicate that aerosol water content is strongly dependent on
 27      relative humidity and composition. Particle-bound water can represent a significant mass
 28      fraction of the PM concentration  at relative humidities above 60%.  A substantial fraction of
29      accumulation-mode PM is hygroscopic or deliquescent. The more hygroscopic particles tend to
30      contain more sulfates, nitrates, and secondary organic compounds, while the less hygroscopic
31      particles tend to contain more elemental carbon, primary organic compounds, and crustal

        June 2003                                2-97        DRAFT-DO NOT QUOTE OR CITE

-------
 1     components.  Fresh, submicron-size soot particles may tend to shrink with increasing relative
 2     humidity because of a structural change. The effects of relative humidity on the sorption of
 3     semivolatile organic compounds on particles are not well understood. The amount of water
 4     sorbed to an atmospheric aerosol may be affected by the presence of an organic film on the
 5     particle, which may impede the transport of water across the surface.
 6          Fine and coarse particles differ not only in formation mechanisms and size, but also in
 7     sources; composition; and chemical, physical, and biological properties.  Fine and coarse
 8     particles overlap in the intermodal size range (1-2.5 urn DJ. As relative  humidity increases, fine
 9     particles grow into this size range; as relative humidity decreases, more coarse particles may be
10     suspended in this size range.  It is desirable to measure fine PM and coarse PM separately in
11     order to properly allocate health effects to either fine PM or coarse PM and to correctly
12     determine sources by factor analysis or chemical mass balance.  The selection of a cut point of
13     2.5 urn as a basis for EPA's 1997 NAAQS for fine particles (Federal Register, 1997) and its
14     continued use in many health effects studies reflects the importance placed on more complete
15     inclusion of accumulation-mode particles, while recognizing that intrusion of coarse-mode
16     particles can occur under some conditions with this cut point.
17          In addition to FRM sampling of equilibrated mass to determine compliance with PM
18     standards, EPA requires states to conduct speciation sampling primarily to determine source
19     categories and trends. The current speciation samplers collect PM2.5 on three filters:
20     (1) a Teflon filter for gravimetric determination of mass and for analysis of heavy elements by
21     X-ray fluorescence; (2) a Nylon filter preceded by a nitric acid denuder for artifact-free
22     determination of nitrate and measurement of other ionic species by ion chromatography; and
23     (3) a quartz filter for measurement of elemental carbon (EC) and organic carbon (OC). In
24     addition, IMPROVE (Interagency Monitoring of Protected Visual Environments) samplers
25     provide information on regional PM background and transport.  IMPROVE samplers, in addition
26     to the three types of filters collected by the speciation samplers, also collect a PM10 sample. The
27     IMPROVE and speciation networks use slightly different methods for determination of EC and
28     OC. The two methods agree on total carbon but differ in the split of total carbon into EC and
29     OC. The two methods also differ in their correction for positive artifacts due to absorption of
30     volatile organic  compounds  on the quartz filters. Neither EC/OC method provides for any
        June 2003                                2-98        DRAFT-DO NOT QUOTE OR CITE

-------
  1     correction for negative artifacts due to evaporation of semivolatile organic compounds from the
  2     collected particles.
  3          The EPA expects that monitoring agencies will operate continuous PM monitors; and EPA
  4     is in the process of providing guidance regarding appropriate continuous monitoring techniques.
  5     All currently available techniques for continuous measurements of suspended particle mass, e.g.,
  6     the integrating nephelometer, the beta-absorption monitor, and the Tapered Element Oscillating
  7     Microbalance (TEOM), share the problem of dealing with semivolatile PM components: that is,
  8     in order not to include particle-bound water as part of the mass, the particle-bound water must be
  9     removed by heating or dehumidification; however, heating also causes ammonium nitrate and
 10     semivolatile organic compounds to evaporate. The TEOM monitor operates at a constant, but
 11     higher than ambient, temperature to remove particle-bound water; whereas, the FRM is required
 12     to operate at no more than 5 °C above the ambient temperature.  Subsequently, much of the
 13     particle-bound water is removed during equilibration at 40% relative humidity. This difference
 14     in techniques for removal of particle-bound water causes differences in the measured mass
 15     concentration between the TEOM and FRMs.
 16          Several new techniques for continuous PM mass measurements are currently being field
 17     tested.  The Real-Time Total Ambient Mass Sampler (RAMS) measures the total mass of
 18     collected particles including semivolatile species with a TEOM monitor using a "sandwich
 19     filter."  The sandwich contains a Teflon-coated particle-collection filter followed by a charcoal-
 20     impregnated filter to collect any semivolatile species lost from the particles during sampling.
 21      The RAMS uses a Nafion dryer to remove particle-bound water from the suspended particles and
 22      a particle concentrator to reduce the quantity of gas phase organic compounds that must be
 23      removed by the denuder. The Continuous Ambient Mass Monitor (CAMM) estimates ambient
 24      particulate matter mass by measurement of the increase in the pressure drop across a membrane
 25      filter caused by particle loading.  It also uses a Nafion dryer to remove particle-bound water.
 26      A new differential TEOM offers the possibility of measuring both nonvolatile and semivolatile
 27      PM mass. In addition to continuous mass measurement, a number of techniques for continuous
28      measurement of sulfate, nitrate, or elements are being tested.
29
30
31


        June 2003                                2-99         DRAFT-DO NOT QUOTE OR CITE

-------
 1     2.3.3    Key Points
 2     Fine and Coarse Particles
 3          Particle size distributions show that atmospheric particles exist in two classes, fine
 4     particles and coarse particles.  Fine and coarse particles are defined primarily in terms of their
 5     formation mechanisms and size; an they also differ in sources, chemical composition, and
 6     removal processes (see Table 2-1). Subsequent chapters will show that fine and coarse particles
 7     also differ in aspects of concentration, exposure, dosimetry, toxicology,  and epidemiology.
 8          These differences support the setting of separate standards for fine and coarse particles.
 9     Fine and coarse particles overlap in the size range between 1 and 3 \im aerodynamic diameter
10     where particulate matter (PM) concentrations are at a minimum.  Coarse particles are generally
11     larger than this minimum and are generally formed by mechanical processes. Coarse particles
12     and coarse-mode particles are  equivalent terms. Fine particles are generally smaller than the
13     minimum and are generally formed by coagulation and condensation of gases. Fine particles
14     are subdivided into accumulation, Aitkin, and nucleation modes. In earlier texts, nuclei mode
15     referred to the size range now  split into the Aitkin and nucleation modes (see Figures 2-4
16     and 2-5).
17
18     Measurement of Mass and Composition
19     Nonvolatile PM. Analytical techniques exist for measurement of the mass and chemical
20     composition of PM retained on a filter (nonvolatile mass) in terms of elements (except carbon)
21     and certain key ions (sulfate, nitrate, hydrogen, and ammonium). Acceptable measurements can
22     be made of the total carbon retained on a filter. However, the split into organic carbon and
23     elemental carbon depends on the operational details of the analytical methods and varies
24     somewhat among methods.  Determination of the various organic compounds in the organic
25     carbon fraction remains a challenge.
26
27     Semivolatile PM. Important components of atmospheric PM (particle-bound water, ammonium
28     nitrate,  and many organic compounds) are termed semivolatile because significant amounts of
29     both the gaseous and condensed phases may exist in the atmosphere in equilibrium. Particle-
30     bound water is not considered a pollutant. Most of the particle-bound water is removed by
31     heating the particles or by equilibration of the collected particles at a low relative humidity

       June 2003                               2-100       DRAFT-DO NOT QUOTE OR CITE

-------
 1      (40%) for 24 hours. However, these processes also cause the loss of other semivolatile
 2      components. Semivolatile components also evaporate from the filter during sampling due to the
 3      pressure drop across the filter or due to a reduction in the atmospheric concentration during the
 4      sampling time. Filter collection and equilibration techniques for PM, such as prescribed by the
 5      Federal Reference Methods, lose a fraction of the semivolatile PM.  Continuous methods must
 6      dry the PM to remove particle-bound water. If heating is used to dry the particles, more of the
 7      semivolatile components may be removed than are lost in filter sampling.  Collection and
 8      retention of ammonium nitrate and semivolatile organic compounds represents a major challenge
 9      in the effort to move to continuous measurement of PM mass. The use of diffusion dryers,
10      which dehumidify the air stream without heating, represents a promising approach. Uncertainty
11      in the efficiency of retention of ammonium nitrate and organic compounds on filters also
12      impacts source category attribution and epidemiologic studies.
13
14      Separation of Fine and  Coarse PM
15           Techniques are available to separate fine particles from coarse particles and collect the fine
16      particles on a filter. No such technique exists for coarse particles. As yet, no consensus exists
17      on the best technique for collecting a coarse particle sample for determination of mass and
18      composition. Candidates include multistage impaction, virtual impaction, and difference
19      (subtracting PM2 5 mass or composition from PM10 mass or composition). Advances in the
20      theory and practice of virtual impaction suggest that it would be possible to design virtual
21      impactors with much less than the 10% of fine PM collected in the coarse PM sample as is now
22      the case for the dichotomous samplers  used in air quality studies and with penetration curves as
23      sharp as those used in the current Federal Reference Method for PM2.5.
       June 2003                                2-101        DRAFT-DO NOT QUOTE OR CITE

-------
        REFERENCES

 1      Aiginger, H. (1991) Historical development and principles of total reflection x-ray fluorescence analysis (TXRF).
 2            Spectrochim. ActaPartB 46: 1313-1321.
 3      Aiginger, H.; Streli, C. (1997) Total reflection x-ray fluorescence spectrometry (TXRF) for trace element analysis.
 4            Spectrosc. Eur. 9: 16-21.
 5      Allen, G.; Sioutas, C.; Koutrakis, P.; Reiss, R.; Lurmann, F. W.; Roberts, P. T. (1997) Evaluation of the TEOM®
 6            method for measurement of ambient particulate mass in urban areas. J. Air Waste Manage. Assoc.
 7            47:682-689.
 8      Allen, G. A.; Oh, J. A.; Koutrakis, P.; Sioutas, C. (1999a) Techniques for high-quality ambient coarse particulate
 9            mass measurements. J. Air Waste Manage. Assoc. 49(special issue): PM133-PM141.
10      Allen, J. O.; Paschkewitz, J. S.; Plummer, E. F.; Lafleur, A. L.; Sarofim, A. F.; Smith, K. A. (1999b) Absorption of
11            semi-volatile compounds in oiled impaction substrates: measurement of pyrene absorption. Aerosol Sci.
12            Technol. 30:16-29.
13      Allen, G. A.; Lawrence, J.; Koutrakis, P. (1999c) Field validation of a semi-continuous method for aerosol black
14            carbon (aethalometer) and temporal patterns of summertime hourly black carbon measurements in
15            southwestern PA. Atmos. Environ. 33: 817-823.
16      Allen, R.; Box, M.; Larson, T. V.; Liu, L. J. (2001) A cost-effective weighing chamber for particulate matter filters.
17            J. Air Waste Manage. Assoc. 51:1650-1653.
18      American Conference of Governmental Industrial Hygienists. (1994) 1994-1995 threshold limit values for chemical
19            substances and physical agents and biological exposure indices. Cincinnati, OH: American Conference of
20            Governmental Industrial Hygienists.
21      Andrews, E.; Larson, S. M. (1993) Effect of surfactant layers on the size changes of aerosol particles as a function of
22            relative humidity. Environ. Sci. Technol. 27: 857-865.
23      Andrews, E.; Saxena, P.; Musarra, S.; Hildemann, L .M.; Koutrakis, P.; McMurry, P. H.; Olmez, L; White, W. H.
24            (2000) Concentration and composition of atmospheric aerosols from the 1995 SEAVS experiment and a
25            review of the closure between chemical and gravimetric measurements. J. Air Waste Manage. Assoc.
26            50:648-664.
27      Appel, B. R. (1993) Atmospheric sample analysis and sampling artifacts. In: Willeke, K.; Baron, P. A., eds. Aerosol
28            measurement:  principles, techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 233-259.
29      Appel, B. R.; Tokiwa, Y.; Haik, M. (1981) Sampling of nitrates in ambient air. Atmos. Environ. 15: 283-289.
30      Appel, B. R.; Tokiwa, Y.; Haik, M.; Kothny, E. L. (1984) Artifact particulate sulfate and nitrate formation on filter
31            media. Atmos. Environ. 18:409-416.
32      Amott, W. P.; Moosmuller, H.; Rogers, C. F.; Jin, T.; Bruch, R. (1999) Photoacoustic spectrometer for measuring
33            light absorption by aerosol; instrument description. Atmos. Environ. 33:2845-2852.
34      Arnott, W. P.; Moosmuller, H.; Walker, J. W. (2000) Nitrogen dioxide and kerosene-flame soot calibration of
35            photoacoustic  instruments for measurement of light absorption by aerosols. Rev. Sci. Instruments
36            71:4545-4552.
37      Ayers, G. P.; Keywood, M. D.; Gras, J. L. (1999) TEOM vs. manual gravimetric methods for determination of PMZ5
38             aerosol mass concentrations. Atmos. Environ. 33: 3717-3721.
39       Babich, P.; Davey, M.; Allen, G.; Koutrakis, P. (2000) Method comparisons for particulate nitrate, elemental carbon,
40             and PMZ5 mass in seven U.S. cities. J. Air Waste Manage. Assoc. 50:1095-1105.
41       Bacon, J. R.; Chenery, S. R. N.; Ellis, A. T.; Fisher, A.; McMahon, A. W.; Potts, P. J.; Williams, J. G. (1995)
42             Atomic spectrometry update-atomic mass spectrometry and x-ray fluorescence spectrometry. J. Anal. At.
43             Spectrom. 10:253R-310R.
44       Baldwin, D. P.; Zamzow, D. S.; D'Silva, A. P. (1994) Aerosol mass measurement and solution standard additions for
45             quantitation in laser ablation-inductively coupled plasma atomic emission spectrometry. Anal. Chem.
46             66:1911-1917.
47       Ballach, J.; Hitzenberger, R.; Schultz, E.; Jaeschke, W. (2001) Development of an improved optical transmission
48             technique for black carbon (BC) analysis. Atmos. Environ. 35:2089-2100.
49       Baron,  P. A.; Willeke, K., eds. (2002) Aerosol measurement: principles, techniques, and applications. 2nd ed.
50             New York, NY: John Wiley & Sons.
51       Benner, R. L.; Stedman, D. H. (1989) Universal sulfur detection by chemiluminescence. Anal. Chem.
52             61:1268-1271.
53       Benner, R. L.; Stedman, D. H. (1990) Field evaluation of the sulfur chemiluminescence detector. Environ. Sci.
54            Technol. 24: 1592-1596.


         June 2003                                    2-102         DRAFT-DO NOT QUOTE OR CITE

-------
   1      Bergm, M. H.; Ogren, J. A.; Schwartz, S. E.; Mclnnes, L. M. (1997) Evaporation of ammonium nitrate aerosol in a
   2            heated nephelometer: implications for field measurements. Environ. Sci. Technol. 31: 2878-2883.
   3      Bertoni, G.; Febo, A.; Perrino, C.; Possanzini, M. (1984) Annular active diffuse sampler: a new device for the
   4            collection of organic vapours. Ann. Chim. (Rome) 74: 97-104.
   5      Billiet, J.; Dams, R.; Hoste, J. (1980) Multielement thin film standards for XRF analysis. X-Ray Spectrom
   6            9:206-211.
   7      Binkowski, F. S.; Ching, J. K. S. (1995) Modeling fine particulate mass and visibility using the EPA regional
   8            particulate model. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of Research
  .9            and Development, National Exposure Research Laboratory; report no. EPA/600/A-95/143. Available from-
  10            NTIS, Springfield, VA; PB96-139100.
  11      Binkowski, F. S.; Shankar, U. (1995) The regional particulate matter model. I. Model description and preliminary
  12            results. J. Geophys. Res. [Atmos.] 100: 26,191-26,209.
  13      Birch, M. E. (1998) Analysis of carbonaceous aerosols: interlaboratory comparison. Analyst (Cambridge  UK)
  14            123:851-857.                                                                              '
  15      Birch, M. E.; Gary, R. A. (1996) Elemental carbon-based method for occupational monitoring of particulate diesel
  16            exhaust: methodology and exposure issues. Analyst 121: 1183-1190.
  17      Birmili, W.; Wiedensohler, A. (1998) The  influence of meteorological parameters on ultrafine particle production at
  18            a continental site. J. Aerosol Sci. 29(suppl. 1): S1015-S1016.
  19      Bowers, W. D.; Chuan, R. L. (1989) Surface acoustic-wave piezoelectric crystal aerosol mass monitor. Rev Sci
 20            Instrum. 60:  1297-1302.
 21       Brauer, M.; Koutrakis, P.; Keeler, G. J.; Spengler, J. D. (1991) Indoor and outdoor concentrations of inorganic acidic
 22            aerosols and gases. J. Air Waste Manage. Assoc. 41: 171-181.
 23      Breum, N. O. (2000) The dust holding capacity of porous plastic foam used in particle size-selective sampling
 24            J. Aerosol Sci. 31:379-385.
 25       Brook, J.  R.; Dann, T. F. (1999) Contribution of nitrate and carbonaceous species to PM25 observed in Canadian
 26            cities. J. Air Waste Manage. Assoc.  49: 193-199.
 27      Brorstrom, E.; Grennfelt, P.; Lindskog, A. (1983) The effect of nitrogen dioxide and ozone on the decomposition of
 £o             particle-associated polycyclic aromatic hydrocarbons during sampling from the atmosphere. Atmos Environ
 29             17:601-605.
 30       Bukowiecki, N.; Kitteison, D. B.; Watts, W. F.; Burtscher, H.; Weingartner, E.; Baltensperger, U. (2002) Real-time
 31             characterization of ultrafine and accumulation mode particles in ambient combustion aerosols. J. Aerosol Sci
 32             33: 1139-1154.
 33       Bunz, H.;  Koyro, M.; Mohler, O.; Saathoff, H. (1996) Temperature dependence of the redistribution mechanisms for
 -*4            volatile components of atmospheric aerosol particles. J. Aerosol Sci. 27(suppl. 1): S653-S654.
 35       Cachier, H.; Bremond, M.-P.; Buat-Menard, P. (1989) Determination of atmospheric soot carbon with a simple
 36            thermal method. Tellus 41B: 379-390.
 37       Cadle, S. H.; Groblicki, P. J. (1982) An evaluation of methods for the determination of organic and elemental carbon
 3°            in particulate samples. In: Wolff, G.  T.; KHmisch, R. L., eds. Particulate carbon: atmospheric life cycle New
 39            York, NY: Plenum Press; pp. 89-109.
 40       Cadle, S. H.; Groblicki, P. J.; Mulawa, P. A. (1983) Problems in the sampling and analysis of carbon particles.
 41            Atmos. Environ. 17: 593-600.
 42       Cadle, S. H.; Mulawa, P.A. (1990) Atmospheric carbonaceous species measurement methods comparison study:
 43            General Motors results. Aerosol Sci. Technol. 12: 128-131.
 44       Cahill, T. A. (1981) Innovative aerosol sampling devices based upon PIXE capabilities. In: Johansson, S. A. E., ed.
 45            Particle induced x-ray emission and its analytical applications: proceedings of the 2nd international
 |5            conference on PIXE  and its analytical applications; June, 1980; Lund, Sweden. Amsterdam, The Netherlands:
 47            North-Holland Publishing Co. Nucl. Instr. Meth. Phys. Res. 181:473-480.
 48       Cahill, T. A. (1985) Proton-induced X-ray emission (PIXE). In: Res, J. P., ed. Metals Handbook: Materials
 49            Characterization. Cleveland, OH: American Society for Metals.
 50      Cahill, T. A. (1990) Particle-induced x-ray emission. In: Mills, K., ed. Metals Handbook; v. 10, p. 102.9th ed.
 51             Cleveland, OH: American Society for Metals.
 52      Cahill, T. A.; Eldred, R. A.; Motallebi, N.; Malm, W. C. (1989) Indirect measurement of hydrocarbon aerosols
 53             across the United States by nonsulfate hydrogen-remaining gravimetric mass correlations. Aerosol Sci
54             Technol. 10:421-429.
55       Cahill, T. A.; Wilkinson, K.; Schnell, R. (1992) Composition analyses of size-resolved aerosol samples taken from
56             aircraft downwind of Kuwait, spring  1991. J. Geophys. Res. [Atmos.] 97: 14,513-14,520.



         June 2003                                      2-103        DRAFT-DO NOT QUOTE OR CITE

-------
 1       Cahill, T. A.; Wakabayashi, P.; James, T. A. (1996) Chemical states of sulfate at Shenandoah National Park during
 2            'summer, 1991. Nucl. Instrum. Methods Phys. Res. B 109/110: 542-547.
 3       Camp D. C.; Stevens, R. K.; Cobourn, W. G.; Husar, R. B.; Collins, J. F.; Huntzicker, J. J.; Husar, J. D.; Jaklevic,
 4            ' J. M.; McKenzie, R. L.; Tanner, R. L.; Tesch, J. W. (1982) Intercomparison of concentration results from fine
 5            particle sulfur monitors. Atmos. Environ. 16: 911-916.
 6       Campbell, A. J.; Humayun, M. (1999) Trace element microanalysis in iron meteorites by laser ablation ICPMS.
 7            Anal. Chem. 71:  939-946.
 8       Carson, P. G.; Johnston, M. V.; Wexler, A. S. (1997) Laser desorption/ionization of ultra fine aerosol particles.
 9            Rapid Commun.  Mass Spectrom. 11:993-996.                                               f
10       Castaldi, M. J.; Senkan, S. M. (1998) Real-time, ultrasensitive monitoring of air toxics by laser photoiomzation
11            time-of-flight mass spectroscopy. J. Air Waste Manage. Assoc. 48: 77-81.
12       Chandramouli, B.; Kamens, R. (2001) The photochemical formation and gas-particle partitioning of oxidation
13            products of decamethyl cyclopentasiloxane and  decamethyl tetrasiloxane in the atmosphere. Atmos. Environ.
14             35:87-95.                                                           .   ,  .,
15       Chandramouli, B.; Jang, M.; Kamens, R. M. (2003) Gas-particle partitioning of semi-volatile orgamcs on organic
16             aerosols using a predictive activity coefficient model: analysis of the effects of parameter choices on model
17             performance. Atmos. Environ. 37: 853-864.
18       Chang, M.; Kim, S.; Sioutas, C. (1999) Experimental studies on particle impaction and bounce: effects of substrate
19             design and material. Atmos. Environ. 33:2313-2322.
20       Chang, M. C.; Sioutas, C.; Kim, S.; Gong, H., Jr.; Linn, W. S. (2000) Reduction of nitrate losses from filter and
21             impactor samplers by means of concentration enrichment. Atmos. Environ.  34: 85-98.
22       Chang, M.-C; Geller, M. D.; Sioutas, C.; Fokkens, P. H. B.; Cassee, F. R. (2002) Development and evaluation of a
23             compact, highly  efficient coarse particle concentrator for lexicological  studies. Aerosol Sci. Technol.
24             36:492-501.                                                                      ,   ..
25       Chein, H.; Lundgren, D. A. (1993) A virtual impactor with clean air core for the generation of aerosols with narrow
26             size distributions. Aerosol Sci. Technol. 18: 376-388.
27       Chen, B. T. (1993) Instrument calibration. In: Willeke, K.; Baron, P. A., eds. Aerosol measurement: principles,
28             techniques, and applications. New York, NY: Van Nostrand Reinhold; pp. 493-520.
29       Chen, B. T.; Yeh, H. C. (1987) An improved virtual impactor: design and performance. J. Aerosol Sci. 18: 203-214.
30       Chen^ B. T.; Yeh^ H. C.; Cheng, Y. S. (1985) A novel  virtual impacton calibration and use. J. Aerosol Sci. 16:
31            ' 343-354. '                                                                      ie. • T  i.   i
32       Chen, B. T.; Yeh, H. C.; Cheng, Y. S. (1986) Performance of a modified virtual impactor. Aerosol Sci. Technol.
OQ             ff 'l^iQ ^7/^
34       Cheng, Y.-S.; Yeh, H.-C. (1979) Particle bounce in cascade impactors. Environ. Sci. Technol. 13:1392-1396.
35       Chow,'j. C. (1995) Measurement methods to determine compliance with ambient air quality standards for suspended
36            particles. J. Air Waste Manage. Assoc. 45:320-382.
37      Chow, J. C.; Watson, J. G. (1998) Guideline on speciated particulate monitoring (Draft 3). Research Triangle Park,
38           'wC: U.S. Environmental Protection Agency; Office of Air Quality Planning and Standards. Available:
39            http://www.epa.gov/ttn/amtic/files/ambient/pm25/spec/drispec.pdf (20 March 2002).
40      Chow  J C • Watson, J. G. (1999) Ion chromatography in elemental analysis of airborne particles. In: Landsberger,
41            ' S.; Creatchman, M., eds. Elemental Analysis of Airborne Particles. Newark, NJ: Gordon and Breach Science
42            Publishers; pp. 97-137.
43      Chow, J. C.; Watson, J. G.; Pritchett, L. C.; Pierson, W. R.; Frazier, C. A.; Purcell, R. G. (1993) The DRI
44            thermal/optical reflectance carbon analysis system: description, evaluation  and applications in U.S. air quality
45             studies. In: Lodge, J. P., Jr.  Fourth international conference on carbonaceous particles in the atmosphere;
46             April 1991; Vienna, Austria. Atmos. Environ. Part A 27:1185-1201.
47      Chow  J C • Fujita, E.  M.; Watson, J. G.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994a) Evaluation of filter-based
48            ' aerosol measurements during the 1987 Southern California Air Quality Study. Environ. Monit. Assess.
49             30:49-80.                                                                        ,   _,
 50      Chow, J. C.; Watson, J. G.; Fujita, E. M.; Lu, Z.; Lawson, D. R.; Ashbaugh, L. L. (1994b) Temporal and spatial
 51             variations of PM15 and PM10 aerosol in the Southern California Air Quality Study. Atmos. Environ.
 52             28:2061-2080.
 53      Chow, J. C.; Watson, J. G.; Lu, Z.; Lowenthal, D. R; Frazier, C.  A.; Solomon, P. A.; Thuilher, R. H.; Magliano, K.
 54            ' (1996) Descriptive analysis of PM2.5 and PM10 at regionally representative locations during
 55             SJVAQS/AUSPEX. In: Parrish, D.; Trainer, M.; Rao, S. T.; Solomon, P. A., eds. A&WMA international
 56             specialty conference on regional photochemical measurements and modeling, part 2; November 1993;
 57             San Diego, CA. Atmos. Environ. 30: 2079-2112.


         June 2003                                     2-104         DRAFT-DO NOT QUOTE OR CITE

-------
  1      Chow, J. C.; Zielinska, B.; Watson, J. G.; Fujita, E. M.; Richards, H. W.; Neff, W. D.; Dietrich, D.; Hering, S. V.
  2             (1998) Northern Front Range Air Quality Study. Volume A: Ambient measurements. Fort Collins, CO:
  3             Colorado State University, Cooperative Institute for Research in the Atmosphere. Available:
  4             http://www.nfraqs.colostate.edu/dri.html [4 February 2002].
  5      Chow, J. C.; Watson, J. G.; Crow, D.; Lowenthal, D. H.; Merrifield, T. (2000) Comparison of IMPROVE and
  6             NIOSH carbon measurements. Aerosol Sci. Technol. 34: 23-34.
  7      Chow, J. C.; Watson, J. G.; Crow, D.; Lowenthal, D. H.; Merrifield, T. (2001) Comparison of IMPROVE and
  8             NIOSH carbon measurements. Aerosol Sci. Technol. 34: 23-34.
  9      Claiborn, C. S.; Finn, D.; Larson, T. V.; Koenig, J. Q. (2000) Windblown dust contributes to high PM2 s
 10             concentrations. J. Air Waste Manage. Assoc. 50: 1440-1445.
 11      Clark, W. E.; Whitby, K. T. (1975) Measurements of aerosols produced by the photochemical oxidation of SO2 in
 12             air. J. Colloid Interface Sci. 51: 477-490.
 13      Clarke, A. D. (1992) Atmospheric nuclei in the remote free-troposphere. J. Atmos. Chem. 14:479-488.
 14      Clarke, A. D.; Varner, J. L.; Eisele, F.; Mauldin, R. L.; Tanner, D.; Litchy, M. (1998) Particle production in the
 15             remote marine atmosphere: cloud outflow and subsidence during ACE 1. J. Geophys. Res. A 103:
 16             16,397-16,409.
 17      Cobourn, W. G.; Husar, R. B.; Husar, J. D. (1978) Continuous in situ monitoring of ambient particulate sulfur using
 18             flame photometry and thermal analysis. In: Husar, R. B.; Lodge, J. P., Jr.; Moore, D. J., eds. Sulfur in the
 19             atmosphere: proceedings of the international symposium; September 1977; Dubrovnik, Yugoslavia. Atmos.
 20             Environ. 12: 89-98.
 21      Cobourn, W. G.; Husar, R. B. (1982) Diurnal and seasonal patterns of particulate sulfur and sulfuric acid in St.
 22             Louis, July 1977-June 1978. Atmos. Environ.  16: 1441-1450.
 23      Code of Federal Regulations. (2001a) Appendix J to Part 50-reference method for the determination of particulate
 24             matter as PM10 in the atmosphere. C. F. R. 40: pt. 50, app. J.
 25      Code of Federal Regulations. (2001b) Ambient air monitoring reference and equivalent methods. C. F. R. 40: §53.
 26      Code of Federal Regulations. (200 lc) Appendix L to Part 50~reference method for the determination of fine
 27             particulate matter as PM2 5 in the atmosphere. C. F. R. 40: pt. 50, app. L.
 28      Code of Federal Regulations. (2001d) Ambient air quality surveillance. C. F. R. 40: §58.
 29      Conner, W. D. (1966) An inertial-type particle separator for collecting large samples. J. Air Pollut. Control Assoc.
 30             16:35-38.
 31      Conner, T. L.; Norris, G. A.; Landis, M. S.; Williams, R. W. (2001) Individual particle analysis of indoor, outdoor,
 32             and community samples from the 1998 Baltimore particulate matter study. Atmos. Environ. 35:3935-3946.
 33      Gotham, W. E.; Bidleman, T. F. (1992) Laboratory investigations of the partitioning of organochlorine compounds
 34            between the gas phase and atmospheric aerosols on glass fiber filters. Environ. Sci. Technol. 26:469-478.
 35      Countess, R. J. (1990) Interlaboratory analyses of carbonaceous aerosol samples. Aerosol Sci. Technol. 12: 114-121.
 36      Courtney, W. J.; Shaw, R. W.; Dzubay, T. G. (1982) Precision and accuracy of a p gauge for aerosol mass
 37            determinations. Environ. Sci. Technol. 16: 236-239.
 38      Coutant, R. W.; Callahan, P. J.; Chuang, J. C.; Lewis, R, G. (1992) Efficiency of silicone-grease-coated denuders for
 39            collection of polynuclear aromatic hydrocarbons. Atmos. Environ. Part A 26:2831-2834.
 40      Covert, D. S.; Kapustin, V. N.; Quinn, P. K.; Bates, T. S. (1992) New particle formation in the marine boundary
 41            layer. J. Geophys. Res. [Atmos.] 97: 20,581-20,589.
 42      Covert, D. S.; Wiedensohler, A.; Aalto, P.; Heintzenberg,  J.; McMurry, P. H.; Leek, C. (1996) Aerosol number size
 43            distributions from 3 to 500 nm diameter in the arctic marine boundary layer during summer and autumn.
 44            Tellus48B: 197-212.
 45      Cui, W.; Machir, J.; Lewis, L.; Eatough, D. J.; Eatough, N. L. (1997) Fine particulate organic material at Meadview
 46            during the project MOHAVE summer intensive study. J. Air Waste Manage. Assoc. 47: 357-369.
 47      Cui, W.; Eatough, D. J.; Eatough, N. L. (1998) Fine particulate organic material in the Los Angeles Basin -1:
 48            assessment of the high-volume Brigham Young University organic sampling system, BIG BOSS. J. Air
49            Waste Manage. Assoc. 48: 1024-1037.
 50      Dagnall, R. M.; Thompson, K. C.; West, T. S. (1967) Molecular-emission in cool flames. Part I. the behaviour of
 51             sulphur species in a hydrogen-nitrogen diffusion flame and in a shielded air-hydrogen flame. Analyst
 52            92:506-512.
53       Dams, R.; Robbins, J. A.; Rahn, K. A.; Winchester, J. W. (1970) Nondestructive neutron activation analysis of air
54            pollution particulates. Anal. Chem. 42: 861-867.
55       Dane, A. D.; Timmermans, P. A. M.; Van Sprang, H. A.; Buydens, L. M. C. (1996) A genetic algorithm for
56            model-free x-ray fluorescence analysis of thin film. Anal. Chem. 68:2419-2425.
         June 2003                                     2-105         DRAFT-DO NOT QUOTE OR CITE

-------
 1       Davis, D. W.; Reynolds, R. L.; Tsou, G. C.; Zafonte, L. (1977) Filter attenuation corrections forx-ray fluorescence
 2             analysis of atmospheric aerosols. Anal. Chem. 49: 1990-1993.
 3       Davis, C. S.; Fellin, P.; Otson, R. (1987) A review of sampling methods for polyaromatic hydrocarbons in air.
 4             JAPCA 37: 1397-1408.
 5       DeBoer, D. K. G.; Leenaers, A. J. G.; Van Den Hoogenhof, W. W. (1995) Glancing-incidence x-ray analysis of
 6             thin-layered materials: a review. X-Ray Spectrom. 24: 91-102.
 7       Ding, Y.; Lee, M. L.; Eatough, D. J. (1998a) The determination of total nitrite and JV-nitroso compounds in
 8             atmospheric samples. Int. J. Environ. Anal. Chem. 69: 243-255.
 9       Dbg, Y.; Cui, W.; Lee, M. L.; Eatough, D. J. (1998b) Fine paniculate JV-nitroso and nitrite organic compounds in
10             the atmosphere. Appl. Occup. Environ. Hyg. 13: 432-438.
11       Ding, Y.;. Koutrakis, P. (2000) Development of a dichotomous slit nozzle virtual impactor. J. Aerosol Sci.
12             31:1421-1431.
13       Ding, Y.; Pang, Y.; Eatough, D. J.; Eatough, N. L.; Tanner, R. L. (2002) High volume diffusion denuder sampler for
14             the routine monitoring of fine paniculate marten II. field evaluation of the PC BOSS. Aerosol Sci. Technol.:
15             36:383-396.
16       Dockery, D. W.; Pope, C. A., IH; Xu, X.; Spengler, J. D.; Ware, J. H.; Fay, M. E.; Ferris, B. G., Jr.; Speizer, F. E.
17             (1993) An association between air pollution and mortality in six U.S. cities. N. Engl. J. Med. 329:  1753-1759.
18       Durham, J. L.; Wilson, W. E.; Bailey, E. B. (1978) Application of an SO2-denuder for continuous measurement of
19             sulfur in submicrometric aerosols. Atmos. Environ. 12: 883-886.
20       Dzubay, T. G. (1980) Chemical element balance method applied to dichotomous sampler data. Ann. N. Y. Acad.
21             Sci. 338:126-144.
22       Dzubay, T. G.; Stevens, R. K. (1975) Ambient air analysis with dichotomous sampler and X-ray fluorescence
23             spectrometer. Environ. Sci. Technol. 9: 663-668.
24       Dzubay, T. G.; Hines, L. E.; Stevens, R. K. (1976) Particle bounce errors in cascade impactors. Atmos. Environ.
25             10:229-234.
26       Dzubay, T. G.; Stevens, R. K.; Peterson, C. M. (1977) Application of the dichotomous sampler to the
27             characterization of ambient aerosols. In: Dzubay, T. G., ed. X-ray fluorescence analysis of environmental
28             samples. Ann Arbor, MI: Ann Arbor Science Publishers, Inc.
29       Easter, R. C.; Peters, L. K. (1994) Binary homogeneous nucleation: temperature and relative humidity fluctuations,
30             nonlinearity, and aspects of new particle production in the atmosphere. J. Appl. Meteorol. 33: 775-784.
31       Eatough, D. J. (1999) BOSS, the Brigham Young University Organic Sampling System: determination  of paniculate
32             carbonaceous material using diffision denuder sampling technology. In: Lane, D., ed. Gas and particle phase
33             measurements of atmospheric organic compounds. New York, NY: Gordon and Breach Science Publishers.
34             (Advances in environmental, industrial and process control technologies: v. 2).
35       Eatough, D. J.; Wadsworth, A.; Eatough, D. A.; Crawford, J. W.; Hansen, L. D.; Lewis, E. A. (1993) A
36             multiple-system, multi-channel diffusion denuder sampler for the determination of fine-particulate organic
37             material in the atmosphere. Atmos. Environ. Part A 27:1213-1219.
38       Eatough, D. J.; Tang, H.; Cui, W.; Machir, J. (1995) Determination of the size distribution and chemical composition
39             of fine particulate semi-volatile organic material hi urban environments using diffusion denuder technology.
40             In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on  particulate air pollution and human
41             mortality and morbidity, part II; January 1994; Irvine, CA. Inhalation Toxicol. 7: 691-710.
42       Eatough, D. J.; Eatough, D. A.; Lewis, L.; Lewis, E. A. (1996) Fine particulate chemical composition and light
43             extinction at Canyonlands National Park using organic particulate material concentrations obtained with a
44             multisystem, multichannel diffusion denuder sampler. J. Geophys. Res. [Atmos.] 101: 19,515-19,531.
45       Eatough, D. J.; Obeidi, F.; Pang, Y.; Ding, Y.; Eatough, N. L.; Wilson, W. E. (I999a) Integrated and real-time
46             diffusion denuder sampler for PMrs. Atmos. Environ. 33:2835-2844.
47       Eatough, D. J.; Pang, Y.; Eatough, N. L. (1999b) Determination of PMr5 sulfate and nitrate with a PC-BOSS
48             designed for routine sampling for semi-volatile particulate matter. J. Air Waste Manage. Assoc. 49(special
49             issue): PM69-75.
50       Eatough, D. J.; Eatough, N. L.; Obeidi, F.; Pang, Y.; Modey, W.; Long, R. (2001) Continuous determination of
51             PMi5 mass, including semi-volatile species. Aerosol Sci. Technol. 34: 1-8.
52       Fairchild, C. L; Wheat, L. D. (1984) Calibration and evaluation of a real-time cascade impactor. Am. Ind. Hyg.
53             Assoc. J. 45: 205-211.
54       Falconer, R. L.; Bidleman, T. F.; Gotham, W. E. (1995) Preferential sorption of non- and
55             mono-ortho-polychlorinated biphenyls to urban aerosols. Environ. Sci. Technol. 29: 1666-1673.
         June 2003                                     2-106        DRAFT-DO NOT QUOTE OR CITE

-------
  1      Fang, C. P.; McMurry, P. H.; Marple, V. A.; Rubow, K. L. (1991) Effect of flow-induced relative humidity changes
  2            on size cuts for sulfuric acid droplets in the microorifice uniform deposit impactor (MOUDI). Aerosol Sci.
  3            Technol. 14: 266-277.
  4      Farwell, S. O.; Rasmussen, R. A. (1976) Limitations of the FPD and BCD in atmospheric analysis: a review. J.
  5            Chromatogr. Sci. 14:224-234.
  6      Fassel, V. A.; Kniseley, R. N. (1974) Inductively coupled plasma-optical emission spectroscopy. Anal. Chem.
  7            46:1110A-1120A.
  8      Federal Register. (1987) Revisions to the national ambient air quality standards for particulate matter. F. R. (July 1)
  9            52:24,634-24,669.
 10      Federal Register. (1997) National ambient air quality standards for particulate matter; final rule. F. R. (July 18)
 11            62:38,652-38,752.
 12      Federal Register. (1998) National ambient air quality standards for particulate matter and revised requirements for
 13            designation of reference and equivalent methods for PM2 5 and ambient air quality surveillance for particulate
 14            matter; correction. F. R. (February 17) 63: 7710-7718.
 15      Fergenson, D. P.;  Song, X.-H.; Ramadan, Z.; Allen, J. O.; Hughes, L. S.; Cass, G. R.; Hopke, P. K.; Prather, K. A.
 16            (2001) Quantification of ATOFMS data by multivariate methods. Anal. Chem. 73:3535-3541.
 17      Fernandez, F. J. (1989) Atomic absorption spectroscopy. In: Lodge, J. P., Jr., ed. Methods of Air Sampling and
 18            Analysis. 3rd ed. Chelsea, MI: Lewis Publishers, Inc.; pp. 83-89.
 19      Finn, D.; Rumburg, B.; Claibora, C.; Bamesberger, L.; Siems, W. F.; Koenig, J.; Larson, T.; Norris, G. (2001)
 20            Sampling artifacts from the use of denuder tubes with glycerol based coatings in the measurement of
 21            atmospheric particulate matter. Environ. Sci. Technol. 35:40-44.
 22      Fitz, D. R. (1990) Reduction of the positive organic artifact on quartz filters. Aerosol. Sci. Technol. 12: 142-148.
 23      Flocchini, R. G.; Cahill, T. A.; Shadoan, D. J.; Lange, S. J.; Eldred, R. A.; Feeney, P. J.; Wolfe, G. W.; Simmeroth,
 24            D. C.; Suder, J. K. (1976) Monitoring California's aerosols by size and elemental composition. Environ. Sci.
 25            Technol. 10: 76-82.
 26      Forney, L. J.; Ravenhall, D. G.; Winn, D. S. (1978) Aerosol impactors: a study of a fluid jet  impinging upon a void.
 27            J. Appl. Phys. 49: 2339-2345.
 28      Forney, L. J.; Ravenhall, D. G.; Lee, S. S. (1982) Experimental and theoretical study of a two-dimensional virtual
 29            impactor. Environ. Sci. Technol. 16: 492-497.
 30      Friedlander, S. K. (1970) The characterization of aerosols distributed with respect to size and chemical composition.
 31            J. Aerosol Sci. 1: 295-307.
 32      Friedlander, S. K. (1977) Smoke, dust and haze: fundamentals  of aerosol behavior. New York, NY: John Wiley &
 33            Sons.
 34      Friedlander, S. K.; Yeh, E. K. (1998) The submicron atmospheric aerosol as a carrier of reactive chemical species:
 35            case of peroxides. Appl. Occup. Environ. Hyg. 13:416-420.
 36      Friedlander, S. K. (2000). Smoke, dust and haze: fundamentals of aerosol dynamics. 2nd ed.  New York, NY:
 37            Oxford University Press.
 38      Fuchs, N. A. (1964) The mechanics of aerosols. New York, NY: Pergamon Press.
 39      Fuchs, N. A. (1989) The mechanics of aerosols. New York, NY: Dover Publications.
 40      Fung, K. (1990) Particulate carbon speciation by MnO2 oxidation. Aerosol Sci. Technol. 12: 122-127.
 41      Fung, K..; Chow, J. C.; Watson, J. G. (2002) Evaluation of OC/EC speciation by thermal manganese dioxide
 42            oxidation and the IMPROVE method. J. Air Waste Manage. Assoc. 52: 1333-1341.
 43      Gallorini, M. (1995) Trace element monitoring in atmospheric pollution processes by neutron activation analysis.
 44            Microchem. J. 51: 127-137.
 45      Card, E.; Mayer, J. E.; Metrical, B. D.; Dienes, T.; Fergenson, D. P.; Prather, K. A. (1997) Real-time analysis of
 46            individual atmospheric aerosol particles: design and performance of a portable ATOFMS. Anal. Chem.
47            69:4083-4091.
48      Gauderman, W. J.; Mcconnell, R.; Gilliland, F.; London, S.; Thomas, D.; Avoi,  E.; Vora, H.; Berhane, K.;
49            Rappaport, E. B.; Lurmann, F.; Margolis, H. G.; Peters, J. (2000) Association between air pollution and lung
 50            function growth in southern California children. Am. J. Respir. Crit. Care Med. 162: 1383-1390.
51      Ge, Z.; Wexler, A. S.; Johnston, M. V. (1998) Laser desorption/ionization of single ultrafine  multicomponent
52            aerosols. Environ. Sci. Technol. 32: 3218-3223.
53      Geller, M. D.; Kim, S.; Misra, C.; Sioutas, C.; Olson, B. A.; Marple, V. A. (2002) Methodology for measuring
54            size-dependent chemical composition of ultrafine particles. Aerosol Sci. Technol. 36:  748-763.
55      Gong, H., Jr.; Linn, W. S.; Sioutas, C.; Terrell, S. L.; Clark, K. W.; Anderson, K. R.; Terrell, L. (2003) Controlled
56            exposures of healthy and asthmatic volunteers to concentrated ambient fine  particles in Los Angeles.
57            Inhalation Toxicol. 15: 305-325.


         June 2003                                     2-107        DRAFT-DO NOT QUOTE OR CITE

-------
  I       Goss, K.-U.; Eisenreich, S. J. (1997) Sorption of volatile organic compounds to particles from a combustion source
 2             at different temperatures and relative humidities. Atmos. Environ. 31: 2827-2834.
 3       Gray, A. L.; Wiliiams, J. G. (1987a) Oxide and doubly charged ion response of a commercial inductively coupled
 4             plasma mass spectrometry instrument J. Anal. At. Spectrom. 2: 81-82.
 5       Gray, A. L.; Williams, J. G. (1987b) System optimisation and the effect on polyatomic, oxide and doubly charged
 6             ion response of a commercial inductively coupled plasma mass spectrometry instrument. J. Anal. At.
 7             Spectrom. 2: 599-606.
 8       Grosjean, D. (1983) Polycyclic aromatic hydrocarbons in Los Angeles air from samples collected on Teflon, glass
 9             and quartz filters. Atmos. Environ. 17: 2565-2573.
10       Grosjean, D.; Parmar, S. S. (1990) Interferences from aldehydes and peroxyacetyl nitrate when sampling urban air
11             organic acids on alkaline traps. Environ. Sci. Technol. 24:1021-1026.
12       Grosjean, D.; Williams, E. L., II; Grosjean, E.; Novakov, T. (1994) Evolved gas analysis of secondary organic
13             aerosols. Aerosol Sci. Technol. 21: 306-324.
14       Gundel, L. A.; Lane, D. A. (1999) Sorbent-coated diffusion denuders for direct measurement of gas/particle
15             partitioning by semi-volatile organic compounds. In: Lane, D. A., ed. Gas and particle phase measurements of
16             atmospheric organic compounds. Amsterdam, The Netherlands: Gordon and Breach Science Publishers.
17             (Advances in environmental, industrial and process control technologies: v. 2).
18       Gundel, L. A.; Lee, V. C.; Mahanama, K. R. R.; Stevens, R. K.; Daisey, J. M. (1995) Direct determination of the
19             phase distributions of semi-volatile polycyclic aromatic hydrocarbons using annular denuders. Atmos.
20             Environ. 29:1719-1733.
21       Hanninen, O. O.; Koistinen, K. J.; Kousa, A.; Keski-Karhu, J.; Jantunen, M. J. (2002) Quantitative analysis of
22             environmental factors in differential weighing of blank teflon filters. J. Air Waste Manage. Assoc.
23             52: 134-139.
24       Hansen, A. D. A.; McMuny, P. H. (1990) An intercomparison of measurements of aerosol elemental carbon during
25             the 1986 carbonaceous species method comparison study.  J. Air Waste Manage. Assoc. 40: 894-895.
26       Hansen, A. D. A.; Rosen, H.; Novakov, T. (1984) The aethalometer - an instrument for the real-time measurement of
27             optical  absorption by aerosol particles. Sci. Total Environ. 36:  191-196.
28       Harman, J. N.  (1989) Inductively coupled plasma emission spectroscopy. In: Lodge, J. P.,  Jr., ed. Methods of air
29             sampling and analysis. 3rd ed. Chelsea, MI: Lewis Publishers,  Inc.; pp.  89-92.
30       Haupt, O.; Klane, B.; Schaefer, C.; Dannecker, W. (1995) Preparation of quartz fibre filter standards for x-ray
31             fluorescence analysis of aerosol samples. X-Ray Spectrom. 24: 267-275.
32       He, K.; Yang, F.; Ma, Y.; Zhang, Q.; Yao, X.; Chan, C. K.; Cadle, S.; Chan, T.; Mulawa, P. (2001) The
33             characteristics of PMzi in Beijing, China. Atmos. Environ. 35:4959-4970.
34       Hegg, D.  A.; Radke, L. F.; Hobbs, P. V. (1990) Particle production associated with marine clouds. J. Geophys. Res.
35             [Atmos.] 95: 13,917-13,926.
36       Hegg, D.  A.; Radke, L. F.; Hobbs, P. V. (1991) Measurements of Aitken nuclei and cloud condensation nuclei in the
37             marine  atmosphere and their relation to the DMS-cloud-climate hypothesis. J. Geophys. Res. [Atmos.]
38             96: 18,727-18,733.
39       Hering, S.; Cass, G. (1999) The magnitude of bias in the measurement of PM2S arising from volatilization of
40             particulate nitrate from Teflon filters. J. Air Waste Manage. Assoc. 49: 725-733.
41       Hering, S. V.;  Appel, B. R.; Cheng, W.; Salaymeh, F.; Cadle, S. H.; Mulawa, P. A.; Cahill, T. A.; Eldred, R. A.;
42             Surovik, M.; Fitz, D.; Howes, J. E.; Knapp, K. T.; Stockburger, L.; Turpin, B. J.; Huntzicker, J. J.; Zhang,
43             X.-Q.; McMurry, P. H. (1990) Comparison of sampling methods for carbonaceous aerosols in ambient air.
44             Aerosol Sci. Technol. 12: 200-213.
45       Hinds, W. C. (1999) Aerosol technology: properties, behavior, and measurement of airborne particles. 2nd ed.
46             New York, NY:  John Wiley & Sons.
47       Hinz, K.-P.; Kaufrnann, R.; Spengler, B. (1994) Laser-induced mass analysis of single particles in the airborne state.
48             Anal. Chem. 66: 2071-2076.
49       Hitzenberger,  R.; Dusek, U.; Beraer, A. (1996) Black carbon measurements using an integrating sphere. J. Geophys.
50             Res. [Atmos.]  101: 19,601-19,606.
51       Hitzenberger,  R.; Berner, A.; Dusek, U.; Alabashi, R. (1997) Humidity-dependent growth of size-segregated aerosol
52             samples. Aerosol Sci. Technol. 27: 116-130.
53       Hitzenberger,  R.; Jennings, S. G.; Larson, S. M.; Dillner, A.; Cachier, H.; Galambos, Z.; Roue, A.;  Spain, T. G.
54             (1999)  Intercomparison of measurement methods for black carbon aerosols. Atmos. Environ. 33: 2823-2833.
55       Holynska, B.;  Ptasinski, J.; Maenhaut, W.; Annegarn, H. J. (1997) Energy-dispersive x-ray fluorescence
56             spectrometer with capillary optics for the chemical analysis of atmospheric aerosols with high time resolution.
57             J. Aerosol Sci. 28: 1455-1463.


         June 2003                                      2-108        DRAFT-DO NOT QUOTE OR CITE

-------
   1       Hoppel, W. A.; Frick, G. M.; Fitzgerald, J. W.; Larson, R. E. (1994) Marine boundary layer measurements of new
   2             particle formation and the effects nonprectpitating clouds have on aerosol size distribution. J. Geophys. Res.
   3             [Atmos.] 99: 14,443-14,459.
   4       Houck, J. E.; Chow, J. C; Watson, J. G.; Simons, C. A.; Pritchett, L. C; Goulet, J. M.; Frazier, C. A. (1989)
   5             Determination of particle size distribution and chemical composition of particulate matter from selected
   6             sources in California: volume I and executive summary (final report). Sacramento, CA: California Air
   7             Resources Board.
   8       Hounam, R. F.; Sherwood, R. J. (1965) The cascade centripeter: a device for determining the concentration and size
   9             distribution of aerosols. Am. Ind. Hyg. Assoc. J. 26:122-131.
 10       Huntzicker, J. J.; Hoffman, R. S.; Ling, C.-S. (1978) Continuous measurement and speciation of sulfur-containing
 11             aerosols by flame photometry. Atmos. Environ. 12: 83-88.
 12       Huntzicker, J. J.; Johnson, R. L.; Shah, J. J.; Gary, R. A. (1982) Analysis of organic and elemental carbon in ambient
 13             aerosols by a thermal-optical method. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate carbon - atmospheric
 14             life cycle: proceedings of an international symposium; October 1980; Warren, MI. New York, NY: Plenum
 15             Press; pp. 79-88.
 16       Husar, R. B.; Tratt, D. M.; Schichtel, B. A.; Falke, S. R.; Li, F.; Jaffe, D.; Gasso, S.; Gill, T.; Laulainen, N. S.; Lu,
 17             F.; Reheis, M. C.; Chun, Y.; Westphal, D.; Holben, B. N.; Gueymard, C.; McKendry, L; Kuring, N.;
 18             Feldman, G. C.; McClain, C.; Frouin, R. J.; Merrill, J.; DuBois, D.; Vignola, F.; Murayama, T.; Nickovic, S.;
 19             Wilson, W. E.; Sassen, K.; Sugimoto, N.; Malm, W. C. (2001) The Asian dust events of April 1998. J.
 20             Geophys. Res. 106: 18,317-18,330.
 21       Jackson, K. W.; Mahmood, T. M. (1994) Atomic absorption, atomic emission, and flame emission spectrometry.
 22             Anal. Chem. 66: 252R-279R.
 23       Jaklevic, J. M.; Loo, B. W.; Goulding, F. S. (1977) Photon-induced X-ray fluorescence analysis using
 24             energy-dispersive detector and dichotomous sampler. In: Dzubay, T.  G., ed. X-ray fluorescence analysis of
 25             environmental samples. Ann Arbor, MI:  Ann Arbor Science Publishers, Inc.; pp. 3-18.
 26       Jaklevic, J. M.; Gatti, R. C.; Goulding, F. S.; Loo, B. W. (1981a) A  p-gauge method applied to aerosol samples.
 27             Environ. Sci. Technol. 15:680-686.
 28       Jaklevic, J. M.; Loo, B. W.; Fujita, T. Y. (1981b) Automatic particulate sulfur measurements with a dichotomous
 29             sampler and on-line x-ray fluorescence analysis. Environ. Sci. Technol. 15: 687-690.
 30       Jang, M.; Kamens, R. M. (1998) A thermodynamic approach for modeling partitioning of semivolatile organic
 31             compounds on atmospheric particulate matter: humidity effects. Environ. Sci. Technol. 32: 1237-1243.
 32       Jang, M.; Kamens, R. M. (1999) Exploratory model approach for adsorption partitioning of SOCs on inorganic
 33             particles. Environ. Sci. Technol.
 34       Jang, M.; Kamens, R. M.; Leach, K. B.; Strommen, M. R. (1997) A  thermodynamic approach using group
 35             contribution methods to model the partitioning of semivolatile organic compounds on atmospheric particulate
 36             matter. Environ. Sci. Technol. 31:2805-2811.
 37       Jenkins, B. M.; Jones, A. D.; Turn, S. Q.; Williams, R. B. (1996) Particle concentrations, gas-particle partitioning,
 38             and species intercorrelations for polycyclic aromatic hydrocarbons (PAH) emitted during biomass burning.
 39             Atmos. Environ. 30: 3825-3835.
 40       Johansson, T. B.; Askelsson, R.; Johansson, S. A. E. (1970) X-ray analysis: elemental trace analysis at the 10'12g
 41             level. Nucl. Instrum. Methods 84: 141-143.
 42       John, W.; Sethi, V. (1993) Threshold for resuspension by particle impaction. Aerosol Sci. Technol. 19:69-79.
 43       John, W.; Wall, S. M.; Ondo, J. L. (1988) A new method for nitric acid and nitrate aerosol measurement using the
 44             dichotomous sampler. Atmos. Environ. 22: 1627-1635.
 45       Johnson, M V.; Wexler, A. S. (1995) MS of individual aerosol particles. Anal.  Chem. 67: 721A-726A.
 46       Johnson, R. L.; Shah, J. J.; Huntzicker, J. J. (1980) Analysis of organic, elemental, and carbonate carbon in ambient
 47             aerosols. In: Verner, S. S. Sampling and analysis of toxic organics in the atmosphere: proceedings of a
 48             symposium; August, 1979; Boulder, CO. Philadelphia, PA: American Society for Testing and Materials,
 49             pp.  111-119. [ASTM Special Technical Publication 721].
 50      Kamens, R. M.; Coe, D. L. (1997) A large gas-phase stripping device to investigate rates of PAH evaporation from
 51             airborne diesel soot particles. Environ. Sci. Technol. 31: 1830-1833.
 52      Kamens, R. M.; Jaoui, M. (2001) Modeling aerosol formation from a-pinene + NOX in the presence of natural
 53            sunlight using gas phase kinetics and gas-particle partitioning theory. Environ. Sci. Technol. 35: 1394-1405.
54      Kamens, R.; Odum, J.; Fan, Z.-H. (1995) Some observations on times to equilibrium for semivolatile polycyclic
55            aromatic hydrocarbons. Environ. Sci. Technol. 29:43-50.
56      Kamens, R.; Jang, M.; Chien, C.-J.; Leach, K. (1999) Aerosol formation from the reaction of a-pinene and ozone
57            using a gas-phase kinetics-aerosol partitioning model. Environ. Sci. Technol. 33: 1430-1438.


         June 2003                                      2-109         DRAFT-DO NOT QUOTE OR CITE

-------
 1       Kao, A. S.; Friedlander, S. K. (1995) Frequency distributions of PM10 chemical components and their sources.
 2             Environ. Sci. TechnoL 29: 19-28.
 3       Kavouras, I. G.; Koutrakis, P. (2001) Use of polyurethane foam as the impaction substrate/collection medium in
 4             conventional inertia! impactors. Aerosol Sci. Technol. 34:46-56.
 5       Keeler, G. J.; Brachaczek, W. W.; Gorse, R. A., Jr.; Japar, S. M.; Pierson, W. R. (1988) Effect of ambient humidity
 6             on dichotomous sampler coarse/fine ratios. Atmos. Environ. 22:  1715-1720.
 7       Keeler, G. J.; Pirrone, N. (1996) Atmospheric transport and deposition of trace elements to Lake Erie from urban
 8             areas. Water Sci. Technol. 33: 259-265.
 9       Kegler, S. R.; Wilson, W. E.; Marcus, A. H. (2001) PM1, intermodal (PM15-,) mass, and the soil component of
10             PM25 in Phoenix, AZ, 1995-96. Aerosol Sci. Technol. 35: 914-920.
11       Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (1999) Relationships between size
12             segregated mass concentration data and ultrafine particle number concentrations in urban areas. Atmos.
13             Environ. 33: 2907-2913.
14       Keywood, M. D.; Ayers, G. P.; Gras, J. L.; Gillett, R. W.; Cohen, D. D. (2000) Size distribution and sources of
15             aerosol in Launceston, Australia, during winter 1997. J. Air Waste Manage. Assoc. 50: 418-427.
16       Khlystov, A.; Wyers, G. P.; Slanina, J. (1995) The steam-jet aerosol collector. Atmos. Environ. 29: 2229-2234.
17       Kim, Y. J.; Boatman, J. F.; Gunter, R. L.; Wellman, D. L.; Wilkison, S. W. (1993) Vertical distribution of
18             atmospheric aerosol size distribution over south-central New Mexico. Atmos. Environ. Part A 27: 1351-1362.
19       Kim, B. M.; Lester, J.; Tisopulos, L.; Zeldin, M. D. (1999) Nitrate artifacts during PMi5 sampling in the South Coast
20             Air Basin of California. J. Air Waste Manage. Assoc. 49(special issue): PM142-153.
21       Kim, B. M.; Cassmassi, J.; Hogo, H.; Zeldin, M. D. (2001a) Positive organic carbon artifacts on filter medium
22:             during PM2 5 sampling in the South Coast Air Basin. Aerosol Sci. Technol. 34: 35-41.
23       Kim, S.; Jaques, P. A.; Chang, M. C; Froines, J. R.; Sioutas, C. (2001b) Versatile aerosol concentration enrichment
24             system (VACES) for simultaneous in vivo and in vitro evaluation of toxic effects of ultrafine, fine and coarse
25             ambient particles Part I: development and laboratory characterization. J. Aerosol Sci. 32:1281-1297.
26       Kim, S.; Jaques, P. A.; Chang, M. C.; Barone, T.; Xiong, C.; Friedlander, S. K.; Sioutas, C. (2001c) Versatile
27             aerosol concentration enrichment system (VACES) for simultaneous in vivo and in vitro evaluation of toxic
28             effects of ultrafine, fine and coarse ambient particles Part II: field evaluation. J. Aerosol Sci. 32:1299-1314.
29       Kirchstetter, T. W.; Corrigan, C. E.; Novakov, T. (2001) Laboratory and field investigation of the adsorption of
30             gaseous organic compounds onto quartz filters. Atmos. Environ. 35:  1663-1671.
31       Kittelson, D. B. (1998) Engines and nanoparticles: a review. J. Aerosol Sci. 29: 575-588.
32       Kittelson, D. B.; McKenzie, R.; Vermeersch, M.; Dorman, F.; Pui, D.; Linne, M.; Liu, B.; Whitby, K. (1978) Total
33             sulfur aerosol concentration with an electrostatically pulsed flame photometric detector system. Atmos.
34             Environ. 12:105-111.
35       Klouda, G. A.; Klinedinst, D. B.; Steel, E. B.; Benner, B. A., Jr.; Parish, H. J. (1996) Exploring a method to produce
36             an urban dust particle filter standard. J. Aerosol Sci. 27(suppl. 1): S351-S352.
37       Koistinen, K. J.; Kousa, A.; Tenhola, V.; Hanninen, O.; Jantunen, M. J.; Oglesby, L.; Kuenzli, N.; Georgoulis, L.
38             (1999) Fine particle (PM2,) measurement methodology, quality assurance procedures and pilot results of the
39             EXPOLIS study. J. Air Waste Manage. Assoc. 49:1212-1220.
40       Korhonen, P.; Kulmala, M.; Laaksonen, A.; Viisanen, Y.; McGraw, R.; Seinfeld, J. H. (1999) Ternary nucleation of
41             H2SO4, NH3, and H2O in the atmosphere. J. Geophys. Res. 104:26,349-26,353.
42       Koutrakis, P.; Kelly, B. P. (1993) Equilibrium size of atmospheric aerosol sulfates as a function of particle acidity
43             and ambient relative humidity. J. Geophys. Res. [Atmos.] 98:7141-7147.
44       Koutrakis, P.; Wolfson, J. M.; Slater, J. L.; Brauer, M.; Spengler, J. D.; Stevens, R. K.; Stone, C. L. (1988a)
45             Evaluation of an annular denuder/filter pack system to collect acidic aerosols and gases. Environ. Sci.
46             Technol. 22: 1463-1468.
47       Koutrakis, P.; Wolfson, J. M.; Spengler, J. D. (1988b) An improved method for measuring aerosol strong acidity:
48             results from a nine-month study in St. Louis, Missouri and Kingston, Tennessee. Atmos. Environ.
49             22:157-162.
50       Koutrakis, P.; Wolfson, J. M.; Spengler, J. D.; Stern, B.; Franklin, C. A. (1989) Equilibrium size of atmospheric
51             aerosol sulfates as a function of the relative humidity. J. Geophys. Res. [Atmos.] 94: 6442-6448.
52       Koutrakis, P.; Thompson, K. M.; Wolfson, J. M.; Spengler, J. D.; Keeler, G. J.; Slater, J. L. (1992) Determination of
53             aerosol strong acidity losses due to interactions of collected particles: results from laboratory and field
54            studies. Atmos. Environ. Part A 26:987-995.
55       Koutrakis, P.; Sioutas, C.; Ferguson, S. T.; Wolfson, J. M.; Mulik, J. D.; Burton, R. M. (1993) Development and
56            evaluation of a glass honeycomb denuder/filter pack system to collect atmospheric gases and particles.
57       .     Environ. Sci. Technol. 27:2497-2501.


         June 2003                                      2-110        DRAFT-DO NOT QUOTE OR  CITE

-------
  1       Koutrakis, P.; Wang, P.-Y.; Wolfson, J. M.; Sioutas, C., inventors. (1996) Method and apparatus to measure
  2             particulate matter in gas. U.S. patent 5,571,945. November 5. Available: www.patents.ibm.com [1999,
  3             November 24].
  4       Krieger, M. S.; Hites, R. A. (1992) Diffusion denuder for the collection of semivolatile organic compounds. Environ.
  5             Sci. Technol. 26: 1551-1555.
  6       Kulmala, M.; Toivonen, A.; Makela", J. M.; Laaksonen, A. (1998) Analysis of the growth of nucieation mode
  7             particles observed in Boreal forest. Tellus SOB: 449-462.
  8       Kulmala, M.; Pirjola, L.; Makela, J. M. (2000) Stable sulphate clusters as a source of new atmospheric particles.
  9             Nature (London, U, K.) 404: 66-69.
 10       Kusko, B. H.; Cahill, T. A.; Eldred, R. A.; Matsuda, Y.; Miyake, H. (1989) Nondestructive analysis of total
 11             nonvolatile carbon by Forward Alpha Scattering Technique (FAST). Aerosol Sci. Technol. 10: 390-396.
 12       Lammel, G.; Novakov, T. (1995) Water nucieation properties of carbon black and diesel soot particles. Atmos.
 13             Environ. 29: 813-823.
 14       Landsberger, S. (1988) Improved methodology for the determination of the seven elemental tracer long-distance
 15             pollution signatures using thermal and epithermal neutron activation analysis. Anal. Chem. 60: 1842-1845.
 16       Landsberger, S..; Wu, D. (1993) Improvement of analytical sensitivities for the determination of antimony, arsenic,
 17             cadmium, indium, iodine, molybdenum, silicon and uranium in airborne particulate matter by epithermal
 18             neutron activation analysis. J. Radioanal. Nucl. Chem. 167:219-225.
 19       Landsberger, S.; Wu, D.; Vermette, S. J.; Cizek, W. (1997) Intercomparison of IAEA airborne particulate matter
 20             reference material. J. Radioanal. Nucl. Chem. 215: 117-127.
 21       Lane, D. A.; Johnson, N. D.; Barton, S. C.; Thomas, G. H. S.; Schroeder, W. H. (1988) Development and evaluation
 22             of a novel gas and particle sampler for semivolatile chlorinated organic compounds in ambient air. Environ.
 23             Sci. Technol. 22: 941-947.
 24       Lavanchy, V. M. H.; Gaggeler, H. W.; Nyeki, S.; Baltensperger, U. (1999) Elemental carbon (EC) and black carbon
 25             (BC) measurements with a thermal method and an aethalometer at the high-alpine research station
 26             Jungfraujoch. Atmos. Environ. 33: 2759-2769.
 27       Lawless, P. A.; Rodes, C. E.; Evans, G.; Sheldon, L.; Creason, J. (2001) Aerosol concentrations during the 1999
 28             Fresno Exposure Studies as functions of size, season and meteorology. Aerosol  Sci. Technol. 34:66-74.
 29       Lawrence, J. E.; Koutrakis, P. (1994) Measurement of atmospheric formic and acetic acids: methods evaluation and
 30             results from field studies. Environ. Sci. Technol.  28: 957-964.
 31       Lawrence, J.; Koutrakis, P. (1996a) Measurement and speciation of gas and particulate phase organic acidity in an
 32             urban environment: 1. analytical. J. Geophys. Res. [Atmos.] 101: 9159-9169.
 33       Lawrence, J.; Koutrakis, P. (1996b) Measurement and speciation of gas and particulate phase organic acidity in an
 34             urban environment: 2. speciation. J. Geophys. Res. [Atmos.] 101:9171-9184.
 35       Lawson, D. R. (1990) The Southern California Air Quality Study. J. Air Waste Manage. Assoc. 40: 156-165.
 36       Leach, K. B.; Kamens, R. M.; Strommen, M. R.; Jang, M. (1999) Partitioning of semivolatile organic compounds in
 37             the presence of a secondary organic aerosol in a controlled atmosphere. J. Atmos. Chem. 33:241-264.
 38       Lee, W.-M. G.; Shih, P.-M.; Wang, C.-S. (1997) The influence of relative humidity on the size of atmospheric
 39             aerosol. J. Environ. Sci. Health Part A: Environ. Sci. Eng. Toxic Hazard. Subst.  Control 32: 1085-1097.
 40       Lewis, C. W.; Macias, E. S. (1980) Composition of size-fractionated aerosol in Charleston, West Virginia. Atmos.
 41             Environ. 14: 185-194.
 42       Lewtas, J.; Pang, Y.; Booth, D.; Reimer, S.; Eatough, D. J.; Gundel, L. A. (2001) Comparison of sampling methods
 43            for semi-volatile organic carbon associated  with PM^5. Aerosol Sci. Technol. 34:9-22.
 44       Li, W.; Montassier, N.; Hopke, P. K. (1992) A system to measure the hygroscopicity of aerosol particles. Aerosol
 45             Sci. Technol.  17:25-35.
 46      Li, N.; Kim, S.; Wang, M.; Froines, J. R.; Sioutas, C.; Nel, A. (2002) Use of a stratified oxidative stress model to
 47            study the biological effects of ambient concentrated and diesel exhaust particulate matter. Inhalation Toxicol.
 48              14:459-486.
 49       Li, N.; Sioutas, C.; Froines, J. R.; Cho, A.; Misra, C.; Nel, A. (2003) Ultrafme particulate pollutants induce oxidative
 50             stress and mitochondrial damage. Environ. Health Perspect. Ill: 455-460.
 51       Liang, Z.; Chan, C. K. (1997) A fast technique for measuring water activity of atmospheric aerosols. Aerosol Sci.
 52             Technol. 26: 255-268.
53       Liang, C.; Pankow, J. F.; Odum, J. R.; Seinfeld, J.  H. (1997) Gas/particle partitioning of semivolatile organic
54             compounds to model inorganic, organic, and ambient smog aerosols. Environ. Sci. Technol. 31: 3086-3092.
55       Lin, J. J.; Noll, K. E.; Holsen, T. M. (1994) Dry deposition velocities as a function of particle size in the ambient
56             atmosphere. Aerosol Sci. Technol. 20: 239-252.
         June 2003                                      2-111         DRAFT-DO NOT QUOTE OR CITE

-------
 I       Lindskog, A.; Brorstrom-Lunden, E.; Alfheim, I; Hagen, I. (1987) Chemical transformation of PAH on airborne
 2             particles by exposure to NO2 during sampling: a comparison between two filter media. Sci. Total Environ.
 3             61:51-57.
 4       Liousse, C.; Cachier, H.; Jennings, S. G. (1993) Optical and thermal measurements of black carbon aerosol content
 5             in different environments: variation of the specific attenuation cross-section, sigma a. Atmos. Environ. Part A
 6             27:1203-1211.
 7       Long, R. W.; Smith, R.; Smith, S.; Eatough, N. L.; Mangelson, N. F.; Eatough, D. J.; Pope, C. A., Ill; Wilson, W. E.
 8             (2002) Sources of fine paniculate material along the Wasatch Front. Ener. Fuels 16: 282-293.
 9       Loo, B. W.; Cork, C. P. (1988) Development of high efficiency virtual impactors. Aerosol Sci. Technol. 9: 167-176.
10       Loo! B. W.; Jaklevic, J. M; Goulding, F. S. (1976) Dichotomous virtual impactors for large scale monitoring of
11          '  airborne paniculate matter. In: Liu, B. Y. H., ed. Fine particles:  aerosol generation, measurement, sampling,
12             and analysis. New York, NY: Academic Press; pp. 311-350.
13       Loo, B. W.; French, W. R.; Gatti, R. C.; Goulding, F. S.; Jaklevic, J. M.; Llacer, J.; Thompson, A. C. (1978)
14             Large-scale measurement of airborne paniculate sulfur. Atmos.  Environ. 12: 759-771.
15       Lundgren, D. A.;  Burton, R. M. (1995) Effect of particle size distribution on the cut point between fine and coarse
16             ambient mass fractions. In: Phalen, R. F.; Bates, D. V., eds. Proceedings of the colloquium on paniculate air
17             pollution and human mortality and morbidity; January 1994; Irvine, CA. Inhalation Toxicol. 7: 131-148.
18       Lundgren, D. A.;  Hausknecht, B. J.; Burton, R. M.  (1984) Large particle size distribution in five U.S. cities and the
19             effect on a new ambient paniculate matter standard (PM10). Aerosol Sci. Technol. 7:467-473.
20       Lynch, A. J.; McQuaker, N. R.; Brown, D. F. (1980) ICP/AES analysis and the composition of airborne and soil
21             materials in the vicinity of a lead/zinc smelter complex. J. Air Pollut. Control Assoc. 30: 257-260.
22       Mader, B. T.; Flagan, R. C.; Seinfeld, J. H. (2001) Sampling atmospheric carbonaceous aerosols using a particle trap
23             impactor/denuder sampler. Environ. Sci. Technol. 35: 4857-4867.
24       Mahanama, K. R. R.; Gundel, L. A.; Daisey, J. M. (1994) Selective fluorescence detection of polycyclic aromatic
25             hydrocarbons in environmental tobacco smoke and other airborne particles. Int. J. Environ. Anal. Chem.
26             56:289-309.
27       Makela, J. M.; Aalto, P.; Jokinen, V.; Pohja, T.; Nissinen, A.; Palmroth, S.; Markkanen, T.; Seitsonen, K.;
28             Lihavainen, H.; Kulmala, M. (1997) Observations of ultrafine aerosol particle formation and growth in boreal
29             forest. Geophys. Res. Lett. 24:1219-1222.
30       Malm, W. C.; Sister, J. F.; Huffman, D.; Eldred, R. A.; Cahill, T. A. (1994) Spatial and seasonal trends in particle
31             concentration and optical extinction in the United States. J. Geophys. Res. [Atmos.] 99:1347-1370.
32       Mamane, Y.; Willis, R.; Conner, T. (2001) Evaluation of computer-controlled scanning electron microscopy applied
33             to an ambient urban aerosol sample. Aerosol Sci. Techno!. 34: 97-107.
34       Marple, V. A.; Chien, C. M. (1980) Virtual impactors: a theoretical study. Environ. Sci. Technol. 14: 976-985.
35       Marple, V. A.; Olson, B. A. (1995) A high volume PM10/PM2.s/PM1.0 trichotomous sampler. In: Paniculate matter.
36             health and regulatory issues: proceedings of an international specialty conference; April; Pittsburgh, PA.
37             Pittsburgh, PA: Air & Waste Management Association; pp. 237-261. (A&WMA publication VIP-49).
38       Marple, V. A.; Liu, B. Y. H.; Burton, R. M. (1990) High-volume impactor for sampling fine and coarse particles. J.
39             Air Waste Manage. Assoc. 40: 762-767.
40       Marple, V. A.; Rubow, K. L.; Behm, S. M. (1991)  A microorifice uniform deposit impactor (MOUDI): description,
41             calibration, and use. Aerosol Sci. Technol. 14:434-446.
42       Masuda, H.; Nakasita, S. (1988) Classification performance of a rectangular jet virtual impactor-effects of nozzle
43             width ratio of collection nozzle to acceleration nozzle. J. Aerosol Sci.  19: 243-252.
44       Masuda, H.; Hochrainer, D.; StSber, W. (1979) An improved virtual impactor for particle classification and
45             generation of test aerosols with narrow size  distributions. J. Aerosol Sci. 10: 275-287.
46       McGovern, F. M. (1999) An analysis of condensation nuclei levels at Mace Head, Ireland. Atmos. Environ.
47             33: 1711-1723.
48       McGovern, F. M.; Jennings, S. G.; O'Connor, T. C.; Simmonds, P. G.  (1996) Aerosol and trace gas measurements
49             during the Mace Head experiment. Atmos. Environ. 30: 3891-3902.
50       Mclnnes, L. M.; Quinn, P. K.; Covert, D. S.; Anderson, T. L. (1996) Gravimetric analysis, ionic composition, and
51             associated water mass of the marine aerosol. Atmos. Environ. 30: 869-884.
52       McMurry, P. H. (2000) A review of atmospheric aerosol measurements. Atmos. Environ. 34: 1959-1999.
53       McMurry, P. H.; Friedlander, S. K. (1979) New particle formation in the presence of an aerosol. Atmos. Environ.
54             13: 1635-1651.
55       McMurry, P. H.; Stolzenburg, M. R. (1989) On the sensitivity of particle size to relative humidity for Los Angeles
56             aerosols. Atmos. Environ. 23; 497-507.



         June 2003                                     2-112         DRAFT-DO NOT QUOTE OR CITE

-------
   1       McMurry, P. H.; Woo, K. S.; Weber, R.; Chen, D.-R.; Pui, D. Y. H. (2000) Size distributions of 3 to 10 nm
   2             atmospheric particles: implications for nucleation mechanisms. Philos. Trans. R, Soc. London Ser. A
   3             358:2625-2642.
   4       McQuaker, N. R.; Kluckner, P. D.; Chang, G. N. (1979) Calibration of an inductively coupled plasma-atomic
   5             emission spectrometer for the analysis of environmental materials. Anal. Chem. 51: 888-895.
   6       Meyer, A. S., Jr.; Boyd, C. M. (1959) Determination of water by titration with coulometrically generated Karl
   7             Fischer reagent. Anal. Chem. 31: 215-219.
   8       Meyer, M. B.; Rupprecht, E. G. (1996) Particulate matter sampling methods: the importance of standardization.
   9             J. Aerosol. Sci. 27(suppl. 1): S349-S350.
 10       Meyer, M. B.; Rupprecht, E.; Patashnick, H. (1995) Considerations for the sampling and measurement of ambient
 11             paniculate matter.  In: Particulate matter: health and regulatory issues: proceedings of an international
 12             specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste Management Association;
 13             pp. 204-210. (A&WMA publication VIP-49).
 14       Meyer, M. B.; Patashnick, H.; Ambs, J. L.; Rupprecht, E. (2000) Development of a sample equilibration system for
 15             the TEOM continuous PM monitor. J. Air Waste Manage. Assoc. 50: 1345-1349.
 16       Middlebrook, A. M.; Murphy, D. M.; Thomson,  D. S. (1998) Observations of organic material in individual marine
 17             particles at Cape Grim during the first aerosol characterization experiment (ACE 1). J. Geophys. Res.
 18             103: 16475-16483.
 19       Middlebrook, A. M.; et al. (2002) A comparison of particle mass spectrometers during the 1999 Atlanta Supersites
 20             project. J. Geophys. Res.: in press.
 21       Mignacca, D.; Stubbs, K.  (1999) Effects of equilibration temperature on PMi0 concentrations from the TEOM
 22             method in the Lower Eraser Valley. J. Air Waste Manage. Assoc. 49: 1250-1254.
 23       Miller, F. J.; Gardner, D. E.; Graham, J. A.; Lee, R. E., Jr.; Wilson, W. E.; Bachmann, J. D. (1979) Size
 24             considerations for establishing a standard  for inhalable particles. J. Air Pollut. Control Assoc. 29:610-615.
 25       Misra, C.; Geller, M. D.; Shah, P.; Sioutas, C.; Solomon, P. A. (2001) Development and evaluation of a continuous
 26             coarse (PM10-PM25) particle monitor. J. Air Waste Manage. Assoc. 51: 1309-1317.
 27       Modey, W. K.; Pang, Y.; Eatough, N. L.; Eatough, D. J. (2001) Fine particulate (PMX5)  composition in Atlanta,
 28             USA: assessment of the particle concentrator-Brigham Young University organic sampling system,
 29             PC-BOSS, during the EPA supersite study. Atmos. Environ. 35: 6493-6502.
 30       Moosmuller, H.; Arnott, W. P.; Rogers, C. F. (1998) Photoacoustic and filter measurements related to aerosol light
 31             absorption during the Northern Front Range Air Quality Study (Colorado 1996.1997). J. Geophys. Res.
 32             103:28,149-28,157.
 33       Morawska, L.; Bofmger, N. D.; Kocis, L.; Nwankwoala, A. (1998) Submicrometer and supermicrometer particles
 34             from diesel vehicle emissions. Environ. Sci. Technol. 32:2033-2042.
 35       Mueller, P. K.; Collins, J.  F. (1980) Development of a particulate sulfate analyzer. Palo Alto, CA: Electric Power
 36             Research Institute; report no. P-1382F.
 37       Mueller, P. K.; Fung, K. K.; Heisler, S. L.; Grosjean, D.; Hidy, G. M. (1982) Atmospheric particulate carbon
 38             observations in urban and rural areas of the United States. In: Wolff, G. T.; Klimisch, R. L., eds. Particulate
 39             carbon: atmospheric life cycle: proceedings of an international symposium; October 1980; Warren, MI.
 40            New York, NY: Plenum Press; pp. 343-370.
 41       Mueller, P. K.; Hansen, D. A.; Watson, J. G., Jr. (1986) The subregional cooperative electric utility, Department of
 42            Defense, National Park Service, and EPA study (SCENES) on visibility: an overview. Palo Alto, CA: Electric
 43            Power Research Institute; report no. EA-4664-SR.
 44       Mukerjee, S.; Ellenson, W. D.; Lewis, R. G.;  Stevens, R. K.; Somerville, M. C.; Shadwick, D. S.; Willis, R. D.
 45            (1997) An environmental scoping study in the lower Rio Grande Valley of Texas-Ill, residential
 46            microenvironmental monitoring for air, house dust, and soil. Environ. Int. 23:657-673.
 47      Munksgaard, N. C.; Parry, D. L. (1998) Lead isotope ratios determined by ICP-MS: monitoring of mining-derived
 48            metal particulates in atmospheric fallout, Northern Territory, Australia. Sci. Total Environ. 217: 113-126.
 49      Murphy, D. M.; Thomson, D. S. (1997) Chemical composition of single aerosol particles at Idaho Hill: negative ion
 50            measurements. J. Geophys. Res. 102:6353-6368.
 51      Nadkami, R. A. (1975) Multielement analysis of coal and coal fly ash standards by instrumental neutron activation
 52            analysis. Radiochem. Radioanal. Lett. 21:  161-176.
53      Nam, S.-H.; Masamba, W. R. L.; Montaser, A. (1993) Investigation of helium inductively coupled plasma-mass
54            spectrometry for the detection of metals and nonmetals in  aqueous solutions. Anal. Chem. 65: 2784-2790.
55      NARSTO. (2003) Fine particle assessment (review draft). Oak Ridge, TN: U.S. Department of Energy, Oak Ridge
56            National Laboratory, North American Research Strategy for Tropospheric Ozone, Quality Systems Science
57            Center. Available: http://www.cgenv.com/Narsto/ (22 April 2003).


         June 2003                                     2-113         DRAFT-DO NOT QUOTE OR CITE

-------
 1       National Research Council. (1993) Protecting visibility in national parks and wilderness areas. Washington, DC:
 2             National Academy Press. 3v.
 3       Nelson, M. P.; Zugates, C. T; Treado, P. J.; Casucio, G. S.; Exline, D. L.; Schlaegle, S. F. (2000) Combining Raman
 4             chemical imaging scanning electron mocroscopy (SEM) to characterize ambient fine particulate. Aerosol Sci.
 5             Technol.34:I08-117.
 6       Neubauer, K. R.; Johnston, M. V.; Wexler, A. S. (1998) Humidity effects on the mass spectra of single aerosol
 7             particles. Atmos. Environ. 32: 2521-2529.
 8       Nham, T. T. (1996) Cool plasma technique forthe ultra-trace level determination of 56Fe, 40Ca, 39K, TLi, 24Mg
 9             and 23Na for semi-conductor applications. Mulgrave, Victoria, Australia: Varian Australia Pty Ltd; ICP-MS
10             instruments at work no. ICP-MS-7. Available:
11             http^/www.varianinc.com/cgi-bin/nav?varinc/docs/osyicpms/atwork/icpms7&cid=680052 [18 January,

12             2002]-                                                                          ^   *• *• -A  ,
13       Noble, C. A.; Prather, K. A. (1996) Real-time measurement of correlated size and composition profiles of individual
14             atmospheric aerosol particles. Environ. Sci. Technol. 30:2667-2680.
15       Noble, C. A.; Prather, K. A. (1998) Single particle characterization of albuterol metered dose inhaler aerosol in near
16            'real-time. Aerosol Sci. Technol. 29: 294-306.
17       Noble, C. A.; Vanderpool, R. W.; Peters, T. M.; McElroy, F. F.; Gemmill, D. B.; Wiener, R. W. (2001) Federal
1 g             reference and equivalent methods for measuring fine particulate matter. Aerosol Sci. Technol. 34: 457-464.
19       Noel, M. A.; Topart, P. A. (1994) High-frequency impedance analysis of quartz crystal microbalances. 1. General
20             considerations. Anal. Chem. 66: 484-491.
21       Novakov, T.; Corrigan, C. E.; Penner, J. E.; Chuang, C. C.; Rosario, O.; Mayol Bracero, O. L. (1997) Organic
22             aerosols in the Caribbean trade winds: a natural source? J. Geophys. Res. 102: 21307-21313.
23       O'Dowd, C. D.; Aalto, P.; Hameri, K.; Kulmala, M.; Hoffinann, T. (2002) Atmospheric particles from organic
24             vapours. Nature (London, U. K.) 416: 497-498.
25       Obeidi, F. D.; Eatough, D. J. (2002) Continuous measurement of semi-volatile fine particulate mass in Provo, UT.
26             Aerosol Sci. Technol. 36:191-203.
27       Obeidi, F.; Eatough, N. L.; Eatough, D. J. (2002) Use of the RAMS to measure semivolatile fine particulate matter at
28             Riverside and Bakersfield, California.  Aerosol Sci. Technol. 36: 204-216.
29       Odum, J. R.; Yu, J.; Kamens, R. M. (1994) Modeling the mass transfer of semivolatile organics in combustion
30             'aerosols. Environ. Sci. Technol. 28: 2278-2285.
31       Ohta, S.; Okita, T. (1990) A chemical characterization of atmospheric aerosol in Sapporo. Atmos. Environ. Part A
32             24A: 815-822.
33       Ohta, S.; Hori, M.; Yamagata, S.; Murao, N. (1998) Chemical characterization of atmospheric fine particles in
34             Sapporo with determination of water content. Atmos. Environ. 32: 1021-1025.
35       Olin, J. G.; Sem, G. J. (1971) Piezoelectric microbalance for monitoring the mass concentration of suspended
36             particles. Atmos. Environ. 5: 653-668.
37       Olmez, I. (1989) Trace element signatures in groundwater pollution. In: Watson, J. G., ed. Receptor models in air
38             resources management: an international specialty conference; February 1988; San Francisco, CA. Pittsburgh,
39             PA: Air & Waste Management Association; pp. 3-11. (APCA transactions series: no. 14).
40      Ondov, J. M. (1996) Particulate tracers for source attribution: potential for application to California's San Joaquin
41             Valley. J. Aerosol Sci. 27(suppl. 1): S687-S688.
42      Ondov, J. M.; Kelly, W. R.; Holland, J. Z.; Lin, Z. C.; Wight, S. A. (1992) Tracing fly ash emitted from a coal-fired
43             power plant with enriched rare-earth isotopes: an urban scale test. Atmos. Environ. Part B  26B: 453-462.
44      Ondov, J. M.; Divita, F., Jr. (1993) Size spectra for trace elements in urban aerosol particles by instrumental neutron
45            activation analysis. J. Radioanal. Nucl. Chem. 167: 247-258.
46      Ono, D. M.; Hardebeck, E.; Parker, J.; Cox, B. G. (2000) Systematic biases in measured PMIO values with U.S.
47            Environmental Protection Agency-approved samplers at Owens Lake, California. J. Air Waste Manage.
48            Assoc. 50:1144-1156.
49      Pak, S. S.; Liu, B. Y. H.; Rubow, K. L. (1992) Effect of coating thickness on particle bounce in mertial impactors.
50          '  Aerosol Sci. Technol. 16:141-150.
51       Pang, Y.; Ren, Y.; Obeidi, F.; Hastings, R.; Eatough, D. J.; Wilson, W. E. (2001) Semi-volatile species in PM25:
52            comparison of integrated and continuous samplers for PMZ5 research or monitoring. J. Air Waste Manage.
53            Assoc. 51: 25-36.
54      Pang, Y.; Eatough, N. L.; Wilson, J.; Eatough, D. J. (2002a) Effect of semivolatile material on PMi5 measurement
55            by the PM1S federal reference method sampler at Bakersfield, California. Aerosol Sci. Technol. 36:289-299.
56      Pang, Y.; Eatough, N. L.; Eatough, D. J. (2002b) PM2 5 semivolatile organic material at Riverside, CA: implications
57            forthe PM25 federal reference method sampler. Aerosol Sci. Technol. 36: 277-288.


         June 2003                                     2-114         DRAFT-DO NOT QUOTE OR CITE

-------
   1      Pankow, J. F. (1987) Review and comparative analysis of the theories on partitioning between the gas and aerosol
   2            paniculate phases in the atmosphere. Atmos. Environ. 21: 2275-2283.
   3      Pankow, J. F. (1994a) An absorption model of the gas/aerosol partitioning involved in the formation of secondary
   4            organic aerosol. Atmos. Environ. 28: 189-193.
   5      Pankow, J. F. (1994b) An absorption model of gas/particle partitioning of organic compounds in the atmosphere
   0            Atmos. Environ. 28:185-188.
   7      Pankow, J. F.; Bidleman, T. F. (1991) Effects of temperature, TSP and per cent non-exchangeable material in
   8            determining the gas-particle partitioning of organic compounds. Atmos. Environ. Part A 25: 2241-2249.
   9      Pankow, J. F.; Storey, J. M. E.; Yamasaki, H. (1993) Effects of relative humidity on gas/particle partitioning of
  } Y            semivolatile organic compounds to urban paniculate matter. Environ. Sci. Technol. 27:2220-2226.
  11      Parmar, S. S.; Grosjean, D. (1990) Laboratory tests of KI and alkaline annular denuders. Atmos. Environ Part A
  12            24:2695-2698.
  13      Patashnick, H.; Rupprecht, G.; Arabs, J. L.; Meyer, M. B. (2001) Development of a reference standard for
  *4            paniculate matter mass in ambient air. Aerosol Sci. Technol. 34:42-45.
  15      Pellizzari, E. D.; Krost, K. J. (1984) Chemical transformations during ambient air sampling for organic vapors Anal
  16            Chem. 56: 1813-1819.
  17      Peng, C.; Chan, C. K. (2001) The water cycles of water-soluble organic salts of atmospheric importance
  18            Atmos. Environ. 35:  1183-1192.
  19      Penner, J. E.; Novakov, T. (1996) Carbonaceous particles in the atmosphere: a historical perspective to the fifth
 20            international conference on carbonaceous particles in the atmosphere. J. Geophys. Res FAtmos 1
 21             101: 19,373-19,378.
 22      Perry, K. D.; Hobbs, P. V. (1994) Further evidence for particle nucleation in clear air adjacent to marine cumulus
 23             clouds. J. Geophys. Res. [Atmos.] 99:22,803-22,818.
 24      Perry, K. p.; Cahill, T. A.; Schnell, R. C.; Harris, J. M. (1999) Long-range transport of anthropogenic aerosols to the
 25             National Oceanic and Atmospheric Administration baseline station at Mauna Loa Observatory Hawaii
 26             J. Geophys. Res. Atmos. 104: 18,521-18,533.
 27       Peters, A.; Liu, E.; Verrier, R. L.; Schwartz, J.; Gold, D. R.; Mittleman, M.; Baliff, J.; Oh, J. A.; Allen, G.;
 28             Monahan, K.;  Dockery, D. W. (2000) Air pollution and incidence of cardiac arrhythmia. Epidemiology
 29             11:11-17.                                                                               **
 30       Peters, T. M.; Vanderpool, R. W.; Wiener, R. W. (2001a) Design and calibration of the EPA PM25 well impactor
 31             ninety-six (WINS). Aerosol Sci. Technol. 34:389-397.
 32       Peters, T. M.; Norris, G. A.; Vanderpool, R. W.; Gemmill, D. B.; Wiener, R. W.; Murdoch, R. W.; McEIroy, F. F.;
 33             Pitchford, M. (2001b) Field performance of PM^ federal reference method samplers. Aerosol Sci. Technol'
 34           -  34:433-443.
 35       Peters, T. M.; Gussman, R. A.; Kenny, L. C.; Vanderpool, R. W. (200Ic) Evaluation of PM25 size selectors used in
 JO             speciation samplers. Aerosol Sci. Technol. 34: 422-429.
 37       Petzold, A.; Niessner, R. (1996) Photoacoustic soot sensor for in-situ black carbon monitoring. Appl Phys B-
 38             Lasers Opt. 63: 191-197.
 39       Petzold, A.; Kopp, C.; Niessner, R. (1997) The dependence of the specific attenuation cross-section on black carbon
 40             mass fraction and particle size. Atmos. Environ. 31: 661-672.
 41       Pilinis, C.; Seinfeld, J. H.; Grosjean, D. (1989) Water content of atmospheric aerosols. Atmos. Environ
 42             23:1601-1606.
 43       Piorek, S. (1994) Modem, PC based, high resolution portable EDXRF analyzer offers laboratory performance for
 44             field, in-situ analysis of environmental contaminants. Nucl. Instr. Meth. Phys. Res. A 353: 528-533.
 45       Pirrone, N.; Keeler, G. J. (1996) The Rouge River watershed pollution by trace elements: atmospheric depositions
 4o            and emission sources. Water Sci. Technol. 33:267-275.
 47       Plantz, M. (1996) Common molecular ion interferences in ICP-MS. Mulgrave, Victoria, Australia: Varian Australia
 4o            Pry Ltd; ICP-MS instruments at work no. ICP-MS-6. Available:
 49            n^://www.varianinc.com/cgi-bin/nav?varinc/docs/osi/icpms/arwork/index&cid=680052 [18 January, 2002]
 50      Poor, N.;  Clark, T.; Nye, L.; Tamanini,  T.; Tate, K.; Stevens, R.; Atkeson, T. (2002) Field performance of
 51            dichotomous sequential PM air samplers. Atmos. Environ. 36: 3289-3298.
 52      Pope, C. A., Ill; Thun, M. J.; Namboodiri, M. M.; Dockery, D. W.; Evans, J. S.; Speizer, F. E.; Heath, C. W., Jr.
 53            (1995) Paniculate air pollution as a predictor of mortality in a prospective study of U.S. adults. Am. J^Respir
54            Crit. Care Med. 151: 669-674.
55      Potts, P. J.; Webb, P. C. (1992) X-ray fluorescence spectrometry. J. Geochem. Explor. 44:251-296.
56      Radke, L. F.; Hobbs, P. V. (1991) Humidity and particle fields around some small cumulus clouds. J. Atmos Sci
57            48:1190-1193.


         June 2003                                     2-115        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Raes  F • Van Dingenen, R.; Cuevas, E.; Van Velthoven, P. F. J.; Prospero, L M. (1997) Observations of aerosols in
 2          ' the free troposphere and marine boundary layer of the subtropical Northeast Atlantic: d.scussion of processes
 3            determining their size distribution. J. Geophys. Res. [Atmos.] 102:21315-21328.
 4      Ranweiler, L. E.; Moyers, J. L. (1974) Atomic absorption procedure for analysis of metals in atmospheric paniculate
 5            matter. Environ. Sci. Technol. 8: 152-156.                   ^
 6      Rao, A. K.; Whitby, K. T. (1977) Nonideal collection characteristics of single stage and cascade impactors. Am. ina.
 7          '  Hyg. Assoc. J. 38:174-177.                                                    .          .
 8      Rao, A. K.; Whitby, K. T. (1978) Non-ideal  collection characteristics of inertial impactors-I. single-stage impactors
 9          '  and solid particles. J. Aerosol Sci. 9: 77-86.
10      Ravenhali, D. G.; Forney, L. J.; Jazayeri, M. (1978) Aerosol sizing with a slotted virtual impactor. J. Colloid
11            Interface Sci. 65: 108-117.                                                       .
12      Reid  J  S • Cahill, T. A.; Wakabayashi, P. H.; Dunlap, M. R. (1994) Geometric/aerodynamic equivalent diameter
13          ' ratios of ash aggregate aerosols collected in burning Kuwaiti well fields. Atmos. Environ. 28: 2227-2234.
14      Reilly, P. T. A.; Gieray, R. A.; Whitten, W.  B.; Ramsey, J. M. (1998) Real-time characterization of ^organic
15             composition and size of individual diesel engine smoke particles. Environ. Sci. Technol. 32:2672-26 /».
16      Reist, P. C. (1984) Introduction to aerosol science. New York, NY: Macmillan Publishing Company.
17      Reist P C. (1993) Introduction to aerosol science. New York, NY: McGraw-Hill.               -,.«..
18      Risse, U.; Flammenkamp, E.; Kettrup, A. (1996) Determination of aromatic hydrocarbons in air using diffusion
19            ' denuders. Fresenius1 J. Anal. Chem. 356: 390-395.                                      .    u^uhtn
20      Roberts, R. A.; Corkill, J. (1998) Grass seed field smoke and its impact on respiratory health. J. Environ. Health ou

22      Rounds, s! A.; Pankow, J. F. (1990) Application of a radial diffusion model to describe gas/particle sorption
23             kinetics. Environ. Sci. Technol. 24:1378-1386.                          	     .
24       Rounds, S. A.; Tiffeny, B. A.; Pankow, J. F. (1993) Description of gas/particle sorption kinetics with an mtraparticle
25             diffusion model: desorption experiments. Environ. Sci. Technol. 27: 366-377.
26       Rumburg, B.; Alldredge, R.; Claiborn, C. (2001) Statistical distributions of paniculate matter and the error
27             associated with sampling frequency.  Atmos. Environ. 35: 2907-2920.
28       Rupprecht, G.; Patashnick, H.; Beeson, D. E, Green, R. N.; Meyer, M. B. (1995) A new automated monitor for the
29             measurement of paniculate carbon in the atmosphere. In: Paniculate matter: health and regulatory issues.
30             proceedings of an international specialty conference; April; Pittsburgh, PA. Pittsburgh, PA: Air & Waste
31             Management Association; pp. 309-319. (A&WMA publication VIP-49).
32       Sakata, K.; Kawabata, K. (1994) Reduction of fundamental polyatomic ions in inductively coupled plasma mass
33             spectrometry. Spectrochim. Acta 49B: 1027-1038.                                     ,
34      Saxena, P.; Hildemann, L. M. (1996) Water-soluble organics in atmospheric particles: a critical review of the
35            literature and applications of thermodynamics to identify candidate compounds.  J. Atmos. Chem. 24: 5/-iiw.
36      Saxena, P.; Hildemann, L. M.; McMurry, P. H.; Seinfeld, J. R (1995) Organics alter hygroscopic behavior of
37            atmospheric particles.!. Geophys. Res. [Atmos.] 100: 18,755-18,770.      ,„,„«.-   „ r^v
38      Schmid, H.; Laskus, L.; Abraham, J. H.; Baltensperger, U.; Lavanchy, V.; Bizjak, M.; Burba, P.; Cachier, H, Crow,
 ^9            D • Chow J • Gnauk, T.; Even, A.; Ten Brink, H. M.; Giesen, K.-P.; Hitzenberger, R.; Hueglm, C.;
40            Maenhau( W.; Pio, C.; Carvalho, A.; Putaud, J.-P.; Toom-Sauntry, D.; Puxbaum, H. (20011 Results of the
 41            "carbon conference" international aerosol carbon round robin test stage I. Atmos. Environ. 35:2111-2121.
 42      Schroder, F.; Strom, J. (1997) Aircraft measurements of submicrometer aerosol particles (> 7 nm) m the midaltitude
 43            free troposphere and tropopause region. Atmos. Res. 44:333-356                           ^wo T  A;r
 44      Schwartz, J.; Dockery, D. W.; Neas, L. M.  (1996) Is daily mortality associated specifically with fine particles? J. Air
 45            Waste Manage. Assoc. 46: 927-939.
 46      Seinfeld, J. H.; Pandis, S. N. (1998) Atmospheric chemistry and physics: from air pollution to climate change.
 47            New York, NY: John Wiley & Sons, Inc.                                _
 48       Sem, G. J.; Tsurubayashi, K.; Homma, K. (1977) Performance of the piezoelectric microbalance respirable aerosol
 49             sensor. Am. Ind. Hyg. Assoc. J. 38:580-588.                                          .
 50       Shah, J.  J.; Rau, J. A. (1990) Carbonaceous methods comparison study: interlaboratory round robin interpretation
 51           ' results. Sacramento, CA: California Air Resources Board. Available:
 52             http://www.arb.ca.gov/research/apr/past/pm.htm (21 March 2002).
 53       Silva, P. J.; Prather, K. A. (1997) On-line characterization of individual particles from  automobile emissions.
 54             Environ. ScL Technol. 31:3074-3080.                                          .
 55       Simon, P. K.; Dasgupta, P. K. (1995) Continuous automated measurement of the soluble fraction of atmospheric
 56             paniculate matter. Anal. Chem. 67:  71-78.


          June 2003                                     2-116        DRAFT-DO NOT QUOTE OR CITE

-------
  1       Sioutas, C.; Koutrakis, P. (1996) Inertial separation of ultrafme particles using a condensational growth/virtual
  2             impaction system. Aerosol Sci. Technol. 25: 424-436.
  3       Sioutas, C.; Koutrakis, P.; Burton, R. M. (1994a) Development of a low outpoint slit virtual impactor for sampling
  4             ambient fine particles. J. Aerosol Sci. 25: 1321-1330.
  5       Sioutas, C.; Koutrakis, P.; Burton, R. M. (1994b) A high-volume small outpoint virtual impactor for separation of
  6             atmospheric particulate from gaseous pollutants. Part. Sci. Technol. 12:207-221.
  7       Sioutas, C.; Koutrakis, P.; Olson, B. A. (1994c) Development and evaluation of a low cutpoint virtual impactor.
  8             Aerosol Sci. Technol. 21: 223-235.
  9       Sioutas, C.; Koutrakis, P.; Wolfson, J. M. (1994d) Particle losses in glass honeycomb denuder samplers. Aerosol
 10             Sci. Technol. 21:  137-148.
 11       Sioutas, C.; Koutrakis, P.; Burton, R. M. (1995a) A technique to expose animals to concentrated fine ambient
 12             aerosols. Environ. Health Perspect. 103:172-177.
 13       Sioutas, D.; Koutrakis, P.; Ferguson, S. T.; Burton, R. M. (I995b) Development and evaluation of a prototype
 14             ambient particle concentrator for inhalation exposure studies. Inhalation Toxicol. 7: 633-644.
 15       Sioutas, C.; Koutrakis, P.; Wolfson, J. M.; Azaroff, L. S.; Mulik, J. D. (1996a) Field evaluation of a glass
 16             honeycomb denuder/filter pack system to collect atmospheric gases and particles. In: Gay, B. W.; Fuerst,
 17             R. G.; Jayanty, R. K. M., eds. Measurement of toxic and related air pollutants: proceedings of a conference;
 18             May, 1994; Durham, NC. Research Triangle Park, NC: U.S. Environmental Protection Agency, Office of
 19             Research and Development; report no. EPA/600/R-94/136; pp. 426-435.
 20       .Sioutas, C.; Wang, P. Y.; Ferguson, S. T.; Koutrakis, P.; Mulik, J. D. (I996b) Laboratory and field evaluation of an
 21             improved glass honeycomb denuder/filter pack sampler. Atmos. Environ. 30: .885-895.
 22       Sioutas, C.; Koutrakis, P.; Wang, P.-Y.; Babich, P.; Wolfson, J. M. (1999) Experimental investigation of pressure
 23             drop with particle loading in Nuclepore filters. Aerosol Sci. Technol. 30: 71-83.
 24       Sioutas, C.; Kim, S.; Chang, M.; Terrell, L. L.; Gong, H., Jr. (2000) Field evaluation of a modified DataRAM MIE
 25             scattering monitor for real-time PM2 5 mass concentration measurements. Atmos. Environ. 34:4829-4838.
 26       Smith, K. R.; Kim, S.; Recendez, J. J.; Sioutas, C.; Pinkerton, K. E. (2003) Airborne particles of the California
 27             Central Valley alter the lungs of health adult rats. Environ. Health Perspect.
 28       Solomon, P. A.; Moyers, J. L.; Fletcher, R. A. (1983) High-volume dichotomous virtual impactor for the
 29             fractionation and collection of particles according to aerodynamic size. Aerosol Sci. Technol. 2: 455-464.
 30       Solomon, P. A.; Salmon, L.  G.; Fall, T.; Cass, G. R. (1992) Spatial and temporal distribution of atmospheric nitric
 31             acid and particulate nitrate concentrations in the Los Angeles area. Environ. Sci. Technol. 26: 1596-1601.
 32       Song, X.-H.; Faber, N. M.; Hopke, P. K.; Suess, D. T.; Prather, K. A.; Schauer, J. J.; Cass, G. R. (2001) Source
 33             apportionment of gasoline and diesel by multivariate calibration based on single particle mass spectral data.
 34             Anal. Chim. Acta 446: 329-343.
 35       Speer, R. E.; Barnes, H. M.; Brown, R. (1997) An instrument for measuring the liquid water content of aerosols.
 36             Aerosol Sci. Technol. 27: 50-61.
 37       Spengler, J. D.; Thurston, G. D. (1983) Mass and elemental composition of fine and coarse particles in six U.S.
 38             cities. J. AirPoImt. Control Assoc. 33: 1162-1171.
 39       Stahlschmidt, T.; Schulz, M.; Dannecker, W. (1997) Application of total-reflection X-ray fluorescence for the
 40             determination of lead, calcium and zinc in size-fractionated marine aerosols. Spectrochim. Acta B
 41             52:995-1001.
 42       Stein, S. W.; Turpin, B. J.; Cai, X.; Huang, P.-F.; McMurry, P. H. (1994) Measurements of relative
 43            humidity-dependent bounce and density for atmospheric particles using the DMA-impactor technique. Atmos.
 44            Environ. 28: 1739-1746.
 45       Stevens, R. K.; O'Keeffe, A. E.; Ortman, G. C. (1969) Absolute calibration of a flame photometric detector to
 46            volatile sulfur compounds at sub-part-per-million levels. Environ. Sci. Technol. 3:652-655.
 47       Stevens, R. K.; Mulik, J. D.; O'Keefe, A. E.; Krost, K. J. (1971) Gas chromatography of reactive sulfur gases in air
48            at the parts-per-billion level. Anal. Chem. 43: 827-831.
49       Stolzenburg, M. R.; Hering,  S. V. (2000) Method for the automated measurement of fine particle nitrate in the
 50            atmosphere. Environ. Sci. Technol. 34: 907-914.
51        Streit, N.; Weingartner, E.; Zellweger, C.; Schwikowski, M.; Gaggeler, H. W.; Baltensperger, U. (2000)
52            Characterization of size-fractionated aerosol from the Jungfraujoch (3580 m asl) using total reflection x-ray
53            fluorescence (TXRF). Int. J. Environ. Anal. Chem. 76: 1-16.
54       Strobel, H. A.; Heineman, W. R. (1989) Chemical instrumentation: a systematic approach. 3rd ed. New York, NY:
55             Wiley and Sons, Inc.
         June 2003                                      2-117         DRAFT-DO NOT QUOTE OR CITE

-------
 1       Strommen, M. R.; Kamens, R. M. (1997) Development and application of a dual-impedance radial diffusion model
 2             to simulate the partitioning of semivolatile organic compounds in combustion aerosols. Environ. Sci. Technol.
 3             31:2983-2990.
 4       Strothmann, J. A.; Schiermeier, F. A. (1979) Documentation of the regional air pollution study (RAPS) and related
 5             investigations in the St. Louis air quality control region. Research Triangle Park, NC: U.S. Environmental
 6             Protection Agency, Environmental Science Research Laboratory; report no. EPA-600/4-79-076.
 7       Suarez, A. E.; Caffrey, P. F.; Ondov, J. M.; Thaung, K. C. (1996) Size distribution of Ir-tagged soot from diesel
 8             sanitation trucks. J. Aerosol Sci. 27(suppl. 1): S697-S698.
 9       Suggs, J. C.; Burton, R. M. (1983) Spatial characteristics of inhalable particles in the Philadelphia metropolitan area.
10            'j. Air Pollut. Control Assoc. 33: 688-691.
11       Suh, K H.; Spengler, J. D.; Koutrakis, P. (1992) Personal exposures to acid aerosols and ammonia. Environ. Sci.
12          '  Technol. 26: 2507-2517.
13       Suh, H. H.; Koutrakis, P.; Spengler, J. D. (1994) The relationship between airborne acidity and ammonia in indoor
14             environments. J. Exposure Anal. Environ. Epidemiol. 4:1-23.
15       Sutherland, J. L.; Bhardwaja, P. S. (1987) Composition of the aerosol in northern Arizona and southern Utah. In:
16             Bhardwaja, P. S., ed. Transactions, Visibility Protection: Research and Policy Aspects. Pittsburgh, PA: Air
17             Pollution Control Association; pp. 373-385.
18       Tan, S. R; Horlick, G. (1986) Background spectral features in inductively coupled plasma/mass spectrometry. Appl.
19          '  Spe'ctrosc. 40:445-460.
20       Tang, I. N. (1980) Deliquescence properties and particle size change of hygroscopic aerosols. In: Willeke, K., ed.
21             Generation of aerosols and facilities for exposure experiments. Ann Arbor, MI: Ann Arbor Science
22             Publishers, Inc.; pp. 153-167.
23       Tang, H.; Lewis, E. A.; Eatough, D. J.; Burton, R. M.; Farber, R. J. (1994) Determination of the particle size
24             distribution and chemical composition of semi-volatile organic compounds in atmospheric fine particles with
25             a diffusion denuder sampling system. Atmos. Environ. 28:939-947.
26       Tanner, R. L.; D'Ottavio, T.; Garber, R.; Newman, L. (1980) Determination of ambient aerosol sulfur using a
27             continuous flame photometric detection system. I. Sampling system for aerosol sulfate and sulfuric acid.
28             Atmos. Environ. 14:121-127.
29       Thomson, D. S.; Murphy, D. M. (1994) Analyzing single aerosol particles in real time. CHEMTECH 24: 30-35.
30       Tobias, H. J.; Beving, D. E.; Ziemann, P. J.; Sakurai, H.; Zuk, M.;  McMurry, P. H.;  Zarling, D.; Waymlonis, R.;
31             Kittelson, D. B. (2001) Chemical analysis of diesel engine nanoparticles using a nano-DMA/thermal
32             desorption particle beam mass spectrometer. Environ. Sci. Technol. 35: 2233-2243.
33       Tolocka, M. P.; Peters, T. M.; Vanderpool, R. W.; Chen, F. L.; Wiener, R. -W. (200 la) On the modification of the
34             low flow-rate PM!0 dichotomous sampler inlet. Aerosol Sci. Technol. 34: 407-415.
35       Tolocka, M. P.; Solomon, P. A.; Mitchell, W.; Norris, G. A.; Gemmill, D. B.; Wiener, R. W.; Vanderpool, R. W.;
36             Homdlya, J. B.; Rice, J. (2001b) East vs. west in the US: chemical characteristics of PM2 s during the winter
37             of 1999. Aerosol Sci. Technol. 34: 88-96.
38       Torok, S. B.; Labar, J.; Schmeling, M.; Van Grieken, R. E. (1998) X-ray spectrometry. Anal. Chem. 70:495R-517R.
39       Tsai, C.-J.; Cheng, Y.-R (1996) Comparison of two ambient beta gauge PM10 samplers. J. Air Waste Manage.
40          '  Assoc. 46: 142-147.
41       Tsai, C.-J.; Huang, H.-Y. (1995) Atmospheric aerosol sampling by an annular denuder system and a high-volume
42          '  PM10 sampler. Environ. Int.  21: 283-291.
43       Turner, D. B. (1994) Workbook of atmospheric dispersion estimates: an introduction to dispersion modeling. 2nd ed.
44            Boca Raton, FL: Lewis Publishers.
45      Turner, J. R.; Hering, S. V. (1987)  Greased and oiled substrates as bounce-free impaction surfaces. J. Aerosol Sci.
46            18:215-224.
47      Turpin, B. J.; Huntzicker, J. J. (1991) Secondary formation of organic aerosol in the Los Angeles basin: a descriptive
48            analysis of organic and elemental carbon concentrations. Atmos. Environ. Part A 25: 207-215.
49      Turpin, B. J.; Huntzicker, J. J. (1995) Identification of secondary organic aerosol episodes and quantitation of
50            primary and secondary organic aerosol concentrations during SCAQS. Atmos. Environ. 29: 3527-3544.
51       Turpin, B. J.; Gary, R. A.; Huntzicker, J. J. (1990) An in situ, time-resolved analyzer for aerosol organic and
52            elemental carbon. Aerosol Sci. Technol.  12: 161-171.
53      Turpin, B. J.; Liu, S.-P.; Podolske, K. S.; Gomes, M. S. P.; Eisenreich, S. J.; McMurry, P. H. (1993) Design and
54            evaluation of a novel diffusion separator for measuring gas/particle distributions of semivolatile organic
55            compounds. Environ.  Sci. Technol. 27: 2441-2449.
56      Turpin, B. J.; Huntzicker, J. J.;  Hering, S. V. (1994) Investigation  of organic aerosol sampling artifacts in the
57            Los Angeles basin. Atmos. Environ. 28:  3061-3071.


         June 2003                                     2-118         DRAFT-DO NOT QUOTE OR CITE

-------
   1       Turpin, B. J.; Saxena, P.; Andrews, E. (2000) Measuring and simulating paniculate organics in the atmosphere:
   2             problems and prospects. Atmos. Environ. 34: 2983-3013.
   3       Tyler, G. (1992) ICP vs. AA: how do you choose? Environ. Lab. (October/November): 31-34.
   4       U.S. Court of Appeals for the District of Columbia. (1999) American Trucking Associations, Inc. vs. U.S.
   5             Environmental Protection Agency. 175 F.3d 1027 (D.C. Cir. 1999).
   6       U.S. Environmental Protection Agency. (1982) Air quality criteria for paniculate matter and sulfur oxides. Research
   7             Triangle Park, NC: Office of Health and Environmental Assessment, Environmental Criteria and Assessment
   8             Office; EPA report no. EPA-600/8-82-029aF-cF. 3v. Available from: NTIS, Springfield, VA; PB84-156777.
   9       U.S. Environmental Protection Agency. (1996a) Air quality criteria for particulate matter. Research Triangle Park,
 JO             NC: National Center for Environmental Assessment-RTP Office; report nos. EPA/600/P-95/001aF-cF. 3v.
 11       U.S. Environmental Protection Agency. (1996b) Review of the national ambient air quality standards for particulate
 12             matter: policy assessment of scientific and technical information. OAQPS staff paper. Research Triangle
 13             Park, NC: Office of Air Quality Planning and Standards; report no. EPA/452/R-96-013. Available from-
 14             NTIS, Springfield, VA; PB97-I15406REB.
 15       U.S. Environmental Protection Agency. (1998) Monitoring PM25 in ambient air using designated reference of class I
 } °             equivalent methods. In: Quality Assurance Handbook for Air Pollution Measurement Systems, V. II, Part II,
 17             Section 2.12. Research Triangle Park, NC: Research and Development, Environmental Monitoring Systems
 18             Laboratory.
 19       U.S. Environmental Protection Agency.  (1999) Particulate matter (PM2 5) speciation guidance. Final draft. Edition 1.
 20             Research Triangle Park, NC: Office of Air Quality Planning and Standards. Available:
 21             www.epa.gov/ttn/amtic/files/ambient/pm25/spec/specfinl.pdf [2000, October 11].
 22       U.S. Environmental Protection Agency.  (2001) List of designated reference and equivalent methods. Available:
 23             http://www.epa.gov/ttn/amtic/files/ambienl/criteria/repmlist.pdft05 March 02].
 24       U.S. Environmental Protection Agency.  (2002a) Ambient monitoring technology information center (AMTIC).
 25             Washington, DC:  Office of Air Quality Planning and Standards; Monitoring and Quality Assurance Group.
 26             Available: http://www.epa.gov/ttn/amtic/ (20 March 2002).
 27       U.S. Environmental Protection Agency.  (2002b) Support center for regulatory air models. Washington, DC: Office
 28             of Air Quality Planning and Standards. Available: http://www.epa.gov/scram001 (21 March 2002).
 29       Van Dingenen, R.; Raes, F.; Jensen, N. R. (1995) Evidence for anthropogenic impact on number concentration and
 30             sulfate content of cloud-processed aerosol particles over the North Atlantic. J. Geophys Res
 31             100:21,057-21,067.
 32       Van Loy, M.; Bahadori, T.; Wyzga, R.; Hartsell, B.; Edgerton, E. (2000) The Aerosol Research and Inhalation
 33             Epidemiology Study (ARIES): PM2 5 mass and aerosol component concentrations and sampler
 34             intercomparisons.  J. Air Waste Manage. Assoc. 50: 1446-1458.
 35       Vanderpool, R. W.; Peters, T. M.; Natarajan, S.;  Gemmill, D. B.; Wiener, R. W. (2001) Evaluation of the loading
 36            characteristics of the EPA WINS PM2 3 separator. Aerosol Sci. Technol. 34: 444-456.
 37       Vasiliou, J. G.; Sorenson, D.; McMurry, P. H. (1999) Sampling at controlled relative humidity with a cascade
 38            impactor. Atmos. Environ. 30: 1049-1056.
 39       Venkataraman, C; Lyons, J. M.; Friedlander, S. K. (1994) Size distributions of polycyclic aromatic hydrocarbons
 40            and elemental carbon. 1. Sampling, measurement methods, and source characterization. Environ Sci
 41            Technol. 28: 555-562.
 42       Vincent, J. H. (1989) Aerosol sampling: science and practice. New York, NY: John Wiley & Sons.
 43      Vincent, J. H. (1995) Aerosol science for industrial hygienists. Oxford, United Kingdom: Pergamon.
 44      Wallace, J. C.; Kites, R. A. (1995) Computer-controlled low-volume air sampler for the measurement of semivolatile
 45            organic compounds. Environ. Sci. Technol. 29:2099-2106.
 46      Wallace, L.; Howard-Reed, C. (2002) Continuous monitoring of ultrafine, fine, and  coarse particles in a residence
 47            for 18 months in 1999-2000. J. Air Waste  Manage. Assoc. 52: 828-844.
 48      Wang, H.-C.; John, W. (1987) Comparative bounce properties of particle materials.  Aerosol Sci. Technol
49            7:285-299.
 50      Wang, H.-C.; John, W. (1988) Characteristics of the Berner impactor for sampling inorganic ions. Aerosol Sci
 51            Technol. 8:157-172.
52      Wang, P. Y. (1997) Continuous aerosol mass measurement by flow obstruction. Boston, MA: Harvard University
53            School of Public Health.
54      Ward, M. D.; Buttry, D. A. (1990) In situ interfacial mass detection with piezoelectric transducers. Science
55           (Washington, DC)  249: 1000-1007.
        June 2003                                     2-119        DRAFT-DO NOT QUOTE OR CITE

-------
 1      Watson, J. G.; Chow, J. C; Richards, L. W.; Neff, W. D.; Andersen, S. R.; Dietrich, D. L.; Houck, J. E.; Olmez, I.
 2            (1988a) The 1987-88 metro Denver brown cloud study volume II: measurements. Reno, NV: Desert Research
 3            Institute; final report no. 88I0.1F2 (2 of 3).
 4      Watson, J. G.; Chow, J. C.; Richards, L. W.; Neff, W. D.; Andersen, S. R.; Dietrich, D. L.; Houck, J. E.; Olmez, I.
 5            (1988b) The 1987-88 metro Denver brown cloud study volume HI: data interpretation. Reno, NV: Desert
 6            Research Institute; final report no. 8810.1F3 (3 of 3).
 7      Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B. M.; Pace, T. G.; Meyer, E. L.; Nguyen Q
 8            (1990a) The USEPA/DRI chemical mass balance receptor model, CMB 7.0. Environ. Software 5: 38-49.
 9      Watson, J. G.; Robinson, N. F.; Chow, J. C.; Henry, R. C.; Kim, B.; Nguyen, Q. T.; Meyer, E. L.; Pace, T. G
10            (1990b) Receptor model technical series, v. IH (1989 revision): CMB7 user's manual. Research Triangle Park,
11            NC: U.S. Environmental Protection Agency, Office of Air Quality Planning and Standards; report no.
12            EPA-450/4-90-004. Available: http://www.epa.gov/scrarn001/ [19 February 2002].
13      Watson, J. G.; Chow, J. C.; Lu, Z.; Gertler, A. W. (1993) Participate and gaseous organic receptor modeling for the
14            Southern California Air Quality Study. Sacramento, CA: California Air Resource Board; report no. 8733.1F.
15            Available: http://www.arb.ca.gov/research/apr/past/atmospheric.htmMir%20Quality%20Study (5  June

17      Watson, J. G.; Chow, J. C.; Lu, Z.; Fujita, E. M; Lowenthal, D. H.; Lawson, D. R.; Ashbaugh, L. L. (1994a)
18            Chemical mass balance source apportionment of PM10 during the Southern California Air Quality Study.
19            Aerosol Sci. Technol. 21:1-36.
20      Watson J G.; Chow, J. C.; Lowenthal, D. H.; Pritchett, L. C.; Frazier, C. A.; Neuroth, G. R.; Robbins, R. (1994b)
21            Differences in the carbon composition of source profiles for diesel- and gasoline-powered vehicles. Atmos.
22            Environ. 28:2493-2505.                                                        ,,™0,,T   L
23      Watson J G.; Fujita, E. M.; Chow, J. C.; Zielinska, B.; Richards, L. W.; Neff, W.; Dietrich, D. (1998) Northern
24            front range air quality study. Final report. Fort Collins, CO: Colorado State University, Cooperative Institute
25            for Research in the Atmosphere. Available: http://www.niraqs.colostate.edu/mdex2.html (16 Jan 2002).
26      Watson J G.; Chow, J. C.; Frazier, C. A. (1999) X-ray fluorescence analysis of ambient air samples. In:
27            Landsberger, S.; Creatchman, M., eds. Elemental Analysis of Airborne Particles. Newark, NJ: Gordon and
28            Breach Science Publishers; pp. 67-96.
29      Weber, R. J.; McMurry, P. H.; Eisele, F. L.; Tanner, D. J. (1995) Measurement of expected nucleation precursor
30             species and 3-500-nm diameter particles at Mauna Loa Observatory, Hawaii. J. Atmos. Sci. 52: 2242-2257.
31      Weber, R. J.; McMurry, P. H.; Mauldin, L.; Tanner, D. J.; Eisele, F. L.; Brechtel, F. J.; Kreidenweis, S. M ; Kok, G.
32            L.; Schillawski, R. D.; Baumgardner, D. (1998) A study of new particle formation and growth involving
33             biogenic and trace gas species measured during ACE I. J. Geophys. Res. A 103: 16,385-16,396.
34       Weber R. J.; Marti, J. J.; McMurry, P. H.; Mauldin, L.; Tanner, D.; Eisele, F.; Brechtel, F.; Kreidenweis, S.; Kok,
35             G.; Schillawski, R.; Baumgardner, D. (1997) Measurements of new particle formation and ultrafine particle
36             growth rates at a clean continental site. J. Geophys. Res. 102: 4375-4385.
37       Weber, R. J.; McMurry, P. H.; Mauldin, R. L., HI; Tanner, D. J.; Eisele, F. L.; Clarke, A. D.; Kapustm, V. N. (1999)
38             New particle formation in the remote troposphere: a comparison of observations at various sites. Geophys.
39             Res. Lett. 26:307-310.                                                	
40       Weber, R. J.; Orsini, D.; Daun, Y.; Lee, Y.-N.; kotz, P. J.; Brechtel, F. (2001) A particle-mto-hquid collector for
41             rapid measurement of aerosol bulk chemical composition. Aerosol Sci. Technol. 35:718-727.
42       Wedding, J. B.; Weigand, M. A. (1993) An automatic particle sampler with beta gauging. J. Air Waste Manage.
43             Assoc. 43: 475-479.
44      Weingartner, E.; Burtscher, H.; Baltensperger, U. (1997) Hygroscopic properties of carbon and diesel soot particles.
45             Atmos. Environ. 31:2311-2327.
46      Weschler, C. J.; Shields, H. C. (1999) Indoor ozone/terpene reactions as a source of indoor particles. Atmos.
47            Environ. 33:2301-2312.                                                  .    „„«, „ „   -
48      Wesolowski, J. J.; John, W.; Devor, W.; Cahill, T. A.; Feeney, P. J.; Wolfe, G.; Flocchmi, R. (1977) Collection
49            surfaces of cascade impactors. In: Dzubay, T. G., ed. X-ray fluorescence analysis of environmental samples.
 50            Ann Arbor, MI: Ann Arbor Science Publishers, Inc.; pp. 121-131.
 51       Whitby  K T. (1978) The physical characteristics of sulfur aerosols. Atmos. Environ. 12: 135-159.
 52      Whitby 1 K. T.; Charlson, R. E.; Wilson, W. E.;  Stevens, R. K. (1974) The size of suspended particle matter in air.
 53            Science (Washington, DC) 183:  1098-1099.
 54      Whitby  K T • Sverdrup, G. M. (1980) California aerosols: their physical and chemical characteristics. In: Hidy,
 55            G M.; Mueller, P. K.; Grosjean, D.; Appel, B. R.; Wesolowski, J. J., eds. The character and origins of smog
 56            aerosols: a digest of results from the California Aerosol Characterization Experiment (ACHEX). New York,
 57            NY; John Wiley & Sons, Inc.; pp. 477-517. (Advances in environmental science and technology: v. 9).


         June 2003                                     2-120         DRAFT-DO NOT QUOTE OR CITE

-------
  1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
46
47
48
49
50

51
 White, W. H. (1998) Statistical considerations in the interpretation of size-resolved particulate mass data. J. Air
       Waste Manage, Assoc. 48:454-458.
 Wiedensohler, A.; Hansson, H.-C.; Orsini, D.; Wendisch, M.; Wagner, F.; Bower, K. N.; Chourlarton, T. W.;
       Wells, M.; Parkin, M.; Acker, K.; Wieprecht, W.; Facchini, M. C; Lind, J. A.; Fuzzi, S.; Arends, B. G.;
       Kulmala, M. (1997) Night-time formation and occurrence of new particles associated with orographic clouds.
       Atmos. Environ. 31:-2545-2559.
 Willeke, K.; Baron, P. A., eds. (1993) Aerosol measurement: principles, techniques and applications. New York,
       NY: Van Nostrand Reinhold Publishers.
 Willeke, K.; Whitby, K. T. (1975) Atmosperic aerosols: size distribution interpretation. J. Air Pollut. Control Assoc.
       25: 529-534.
 Williams, E. L., n; Grosjean, D. (1990) Removal of atmospheric oxidants with annular denuders. Environ. Sci.
       Technol. 24: 811-814.
 Williams, R.; Suggs, J.; Zweidinger, R.; Evans, G.; Creason, J.; Kwok, R.; Rodes, C.; Lawless, P.; Sheldon, L.
       (2000) The 1998 Baltimore Particulate Matter Epidemiology-Exposure Study: Part 1. Comparison of ambient,
       residential outdoor, indoor and apartment particulate matter monitoring. J. Exposure Anal. Environ.
       Epidemiol. 10:518-532.
 Wilson, W. E. (1995) Aerosol exposure, physics, and chemistry. Inhalation Toxicol. 7: 769-772.
 Wilson, W. E.; Suh, H. H. (1997) Fine particles and coarse particles: concentration relationships relevant to
       epidemiologic studies. J. Air Waste Manage. Assoc. 47: 1238-1249.
 Wilson, W. E.; Spiller, L. L.; Ellestad, T. G.; Lamothe, P. J.; Dzubay, T. G.; Stevens, R. K.; Macias, E. S.; Fletcher,
       R.  A.; Husar, J. D.; Husar, R. B.; Whitby, K. T.; Kittelson, D. B.; Cantrell, B. K. (1977) General Motors
       sulfate dispersion experiment:  summary of EPA measurements. J. Air Pollut. Control Assoc. 27: 46-51.
 Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; McMurry, P. H. (200Ia) Measurement of Atlanta aerosol size distributions:
       observations of ultrafine particle events. Aerosol Sci. Technol. 34: 75-87.
 Woo, K.-S.; Chen, D.-R.; Pui, D. Y. H.; Wilson, W. E, (2001b) Use of continuous measurements of integral aerosol
       parameters to estimate particle surface area. Aerosol Sci. Technol. 34:  57-65.
 Wu, C. C.; Suarez, A. E.; Lin, Z.; Kidwell, C. B.; Borgoul, P. V.; Caffrey, P.  F.; Ondov, J. M.; Sattler, B. (1998) .
       Application of an Ir tracer to determine soot exposure to students commuting to school on Baltimore public
       buses. Atmos. Environ. 32:1911-1919.
 Yamasaki, H.; Kuwata, K.; Miyamoto, H. (1982) Effects of ambient temperature on aspects of airborne polycyclic
       aromatic hydrocarbons. Environ. Sci. Technol. 16:189-194.
 Yatin, M.; Tuncel, S. G.; Tuncel, G.; Aras, N. K. (1994).Trace element composition of atmospheric aerosols in
       Ankara, Turkey, determined by instrumental neutron activation analysis. J. Radioanal. Nucl. Chem.
       181:401-411.
 Zeng, X.;  Wu, X.; Yao, H.; Yang, F.; Cahill, T. A. (1993) PIXE-induced XRF with transmission geometry. Nucl.
       Instr. Meth. Phys. Res. B 75: 99-104.
 Zhang, X. Q.; McMurry, P. H. (1987) Theoretical analysis of evaporative losses from impactor and filter deposits.
       Atmos. Environ. 21: 1779-1789.
 Zhang, X.; McMurry, P. H. (1992) Evaporative losses of fine particulate nitrates during sampling. Atmos. Environ.
       Part A 26: 3305-3312.
 Zhang, X. Q.; McMurry, P. H.; Hering, S. V.; Casuccio, G. S. (1993) Mixing  characteristics and water content  of
       submicron aerosols measured in Los Angeles and at the Grand Canyon. Atmos. Environ. Part A
       27: 1593-1607.
Zhu, Y.; Hinds, W. C.; Kim, S.; Shen, S.; Sioutas, C. (2002a) Study of ultrafine particles near a major highway with
      heavy-duty diesel traffic. Atmos. Environ. 36: 4323-4335.
Zhu, Y.; Hinds, W. C.; Kim, S.; Sioutas, C. (2002b) Concentration and size distribution of ultrafine particles near a
      major highway. J.  Air Waste Manage. Assoc. 52:1032-1042.
Zoller, W. H.; Gordon, G. E. (1970) Instrumental neutron activation analysis of atmospheric pollutants utilizing
      Ge(Li) y-ray detectors. Anal. Chem. 42:257-265.
         June 2003
                                                2-121
DRAFT-DO NOT QUOTE OR CITE

-------

-------
   1               APPENDIX 2A.  TECHNIQUES FOR MEASUREMENT OF
   2                       SEMIVOLATILE ORGANIC COMPOUNDS
   3
   4
   5      Use ofDenuder Systems To Measure Semivolattte Organic Compounds
   6           Phase distribution of semivolatile organic species has been the subject of several studies
   7      that have employed denuder technology (see Gundel et al., 1995; Gundel and Lane,  1999) to
   8      directly determine the phase distributions while avoiding some of the positive and negative
   9      sampling artifacts associated with using back-up quartz filters. In an ideal system with a
 10      denuder that is 100% efficient, the gas phase would be collected in the denuder and the particle
 11      phase would be the sum of the material collected on the filter and the adsorbent downstream.
 12      Denuder collection efficiency depends on the denuder surface area (+), the diffusivity (+) and
 13      vapor pressure (-) of the compound, the temperature (-) and flow rate (-) of the air stream, and
 14      the presence of competing species (-), including water vapor (Cui et al.,  1998; Kamens and Coe,
 15      1997; Lane et al., 1988).  (The + and - symbols in parentheses indicate qualitatively the effect
 16      increasing each parameter would have on efficiency). In a system with a denuder collection
 17      efficiency less than 100%, the collection efficiency must be known to accurately attribute
 18      adsorbed organics from denuder breakthrough to the gas phase and adsorbed organics volatilized
 19      from collected particles to the particle phase. In calculating the overall phase distributions of
 20      SVOC PAH from a denuder system, the collection efficiency for each compound is needed.
 21          The efficiency of silicone-grease-coated denuders for the collection of polynuclear
 22     aromatic hydrocarbons was examined by Coutant et al. (1992), who examined the effects of
 23     uncertainties in the diffusion coefficients and in the collisional reaction efficiencies on the
 24     overall phase distributions of SVOC PAH calculated using denuder technology. In their study,
 25     they used a single stage, silicone-grease-coated aluminum annular denuder with a filter holder
 26     mounted ahead of the denuder and an XAD trap deployed downstream of the denuder. In a
 27     series of laboratory experiments, they  spiked the filter with a mixture of perdeuterated PAH,
 28     swept the system with ultra-high purity air for several hours, and then analyzed the filter and the
 29     XAD. They found that the effects of these uncertainties, introduced by using a single compound
 30     as a surrogate PAH (in their case, naphthalene), for validation of the denuder collection
31      efficiency, are less significant than normal variations because of sampling and analytical effects.
32     Results on field studies using their sampling system have not been published.
        June 2003                                2A-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1           For measuring paniculate phase organic compounds, the denuder-based sampling system
 2      represents an improvement over the filter/adsorbent collection method (Turpin et al., 1993).
 3      Some researchers, however, have reported that denuder coatings themselves can introduce
 4      contamination (Mukerjee et al., 1997) and that the adsorbed species may be difficult to remove
 5      from the coating (Eatough et al., 1993).
 6           In a study conducted in southern California (Eatough et al., 1995), the Brigham Young
 7      University Organic Sampling System (BOSS; Eatough et al., 1993) was used for determining
 g      POM composition, and a high-volume version (BIG BOSS; flow rate 200 L/min) was utilized
 9      for determining the paniculate size distribution and the chemical composition of SVOC in fine
10      particles. The BOSS, a multi-channel diffusion denuder sampling system, consists of two
11      separate samplers (each operating at 35 L/min). The first sampler consists of a multi-parallel
12     plate diffusion denuder with charcoal-impregnated filter papers as the collection surfaces
13      followed by a two-stage quartz filter pack and a two-stage charcoal-impregnated filter pack. The
14     second sampler operating in parallel with the first consists of a two-stage quartz filter pack,
15     followed by the parallel plate denuder, followed by the two-stage charcoal-impregnated filter
16     pack.  The filter samples collected by the BOSS sampler were analyzed by temperature-
17     programmed volatilization analysis. The second channel allows calculations of the efficiency of
18     the denuder in removing gas-phase specifics that would be absorbed by the charcoal impregnated
 19     filter.  Eatough et al. (1995) also operated a two-stage quartz filter pack alongside the BOSS
20     sampler. The BIG BOSS system (Tang et al., 1994) consists of four systems (each with a
21     flowrate of 200 L/min).  Particle size cuts of 2.5, 0.8, and 0.4 urn are achieved by virtual
22     impaction, and the sample subsequently flows through a denuder, then is split, with the major
23      flow (150 L/min) flowing through a quartz filter followed by an XAD-II bed.  The minor flow is
 24      sampled through a quartz filter backed by a charcoal-impregnated filter paper. The samples
 25      derived from the major flow (quartz filters and XAD-II traps) were extracted with organic
 26      solvents and analyzed by gas chromatography (GC) and GC-mass spectroscopy. The organic
 27      material lost from the particles was found to represent all classes of organic compounds.
 28            Eatough et al. (1996) operated the BOSS sampler for a year at the IMPROVE site at
 29      Canyonlands National Park, UT, alongside the IMPROVE monitor and alongside  a separate
 30      sampler consisting of a two-stage quartz filter pack.  They found that concentrations of
 31      paniculate carbon determined from the quartz filter pack sampling system were low on average

        June 2oo3                                2A-2        DRAFT-DO NOT QUOTE OR CITE

-------
  1      by 39%, and this was attributed to volatilization losses of SVOC from the quartz filters.
  2      In another study conducted with the BOSS in southern California, losses of 35% of the POM, on
  3      average, were found and attributed to losses of the SVOC during sampling (Eatough et al.,
  4      1995).
  5           The denuder used in the various BOSS samplers consists of charcoal-impregnated cellulose
  6      fiber filter material.  Denuder collection efficiencies of greater than 95% have been reported for
  7      organic gases that adsorb on quartz and charcoal-impregnated filters (Eatough et al., 1999a; Ding
  8      et al., 2002; Lewtas et al., 2001). However, because the mass concentration of gas phase species
  9      that adsorb on quartz and charcoal-impregnated filters is so much greater than the mass of
 10      semivolatile organic material in the particulate phase, it is necessary to measure and account for
 11       the inefficiency of the denuder in the BOSS samplers.  To address this problem, Brigham Young
 12       University (BYU) developed a particle-concentrator (PC)-BOSS system (Ding et al., 2002;
 13       Eatough et al., 1999b; Lewtas et al., 2001; Modey et al., 2001; Pang et al., 2001, 2002a,b). The
 14       PC-BOSS includes a virtual impactor upstream of the denuder to improve the denuder collection
 15       efficiency by removing a majority of the gases from the aerosol flow. With this system, denuder
 16       collection efficiencies of greater than 99% have been reported for organic gases, SO2(g),
 17       HNO3(g) and other species that adsorb on quartz and charcoal-impregnated filters (Pang et al.,
 18       2001). Since the concentrations of semivolatile organic and other gases in the presence of the
 19       concentrated particles is not altered by this process the virtual impaction concentration of
 20       accumulation mode particles (except possibly by slight changes in pressure), it is anticipated that
 21      the gas-particle distribution will not be significantly altered by the concentration process.  The
 22      virtual impactor has a 50% cut point at 0.1 urn aerodynamic diameter. As a result, some
 23      particles  in the 0.05 to 0.2 um diameter size range will be removed in the major flow along with
 24      the majority of the gases. Therefore, the mass collection efficiency of the virtual impactor
 25      concentrator will be a function of the particle size distribution in the 0.05 to 0.1 urn size range.
 26      This collection efficiency is measured by comparing the concentration of nonvolatile
 27      components measured in the concentrated sample with that measured in an unconcentrated
 28      sample.  The concentration efficiency varies from 50 to 75%. It is relatively constant over
 29      periods of weeks but varies by season and site, presumably as the particle size distribution
30      changes.  Previous studies at Harvard (Sioutas et al., 1995a,b) have shown that the composition
31      of the sampled aerosol is little changed by the concentration process. The BYU studies listed

        June 2003                                2A-3        DRAFT-DO NOT QUOTE OR CITE

-------
 1      above have shown that the concentration efficiencies for sulfate, organic carbon (OC) and
 2      elemental carbon (EC) are comparable for a given sampling location. Furthermore, the
 3      concentrations of these species and of fine particulate nitrate determined using the PC-BOSS
 4      have been shown to be comparable to those determined using more conventional samplers for
 5      sulfate or EC or using simpler denuder systems for OC and nitrate.
 6           Ding et al. (1998a) developed a method for the determination of total n-nitroso compounds
 7      in air samples and used the method to examine organic compounds formed fromNOx chemistry
 8      in Provo, UT (Ding et al.,  1998b).  In their method, n-nitroso compounds are selectively
 9      decomposed to yield nitric oxide, which is then detected using chemiluminescence. From the
10      samples from Provo, they found that the majority of the n-nitroso and nitrite organic compounds
11      that were present in fine particulate matter were semivolatile organic compounds that could be
12      evaporated from the particles during sampling. They found particulate n-nitroso compound
13      concentrations ranging between <1 and 3 nmoles/m3 and gas-phase n-nitroso compound
14     concentrations in the same range. Particulate organic nitrite concentrations were found in the
15      range of <1 to «5 nmoles/m3, and gas-phase concentrations as high as 10 nmoles/m3 were found.
16          Turpin et al. (1993) developed a sampling system that corrects for the loss of semivolatile
17     organic compounds during sampling by removal of most of the gas phase material from the
18     particles in a diffusion separator sampling system. Unlike the previously mentioned systems,
19     wherein the particulate phase is measured directly, in the system of Turpin et al. (1993) the
20     gas-phase is measured directly. In the laminar flow system, ambient, particle-laden air enters the
21      sampler as an annular flow.  Clean, particle-free  air is pushed through the core inlet of the
22     separator. The clean air and ambient aerosol join downstream of the core inlet section, and flow
23     parallel to each other through the diffusion zone. Because of the much higher diffusivities for
24     gases compared to particles, the SVOC in the ambient air diffuses to the clean, core flow.  The
25     aerosol exits the separator in the annular flow, and the core flow exiting the separator now
26     contains a known fraction of the ambient SVOC. Downstream of the diffusion separator, the
27     core exit flow goes into a polyurethane foam (PUF) plug, where the SVOC is collected. The
28     adsorbed gas phase on the PUF plug is extracted with supercritical fluid CO2 and analyzed by
29     gas chromatography/mass-selective detection (GC/MSD). The gas-phase SVOC is thus
30     determined. Ultimately, to determine particulate phase SVOC concentrations, the total
31     compound concentration will also  be measured and the particulate phase obtained by difference.

       June 2003                               2A-4       DRAFT-DO NOT QUOTE OR CITE

-------
   1     The system was tested for the collection of PAH. The diffiisional transport of gas-phase PAHs
   2     and particle concentrations agreed well with theory. Breakthrough was problematic for low
   3     molecular weight PAHs (MW < 160). Detection limits ranged from 20 to 50 pg of injected mass
   4     for all PAHs.
   5          Gundel et al. (1995) recently developed a technique for the direct determination of phase
   6     distributions of semivolatile polycyclic aromatic hydrocarbons using annular denuder
   7     technology.  The method, called the integrated organic vapor/particle sampler (IOVPS), uses a
   8     cyclone inlet with a 50% cutpoint of 2.5 urn at a sampling rate of 10 L/min. The airstream then
   9     goes through two or three sandblasted glass annular denuders that are coated with ground
 10     adsorbent resin material (XAD-4 was initially examined) that traps vapor-phase organics. The
 11      airstream subsequently passes through a filter, followed by a backup denuder.  The denuder
 12     collection efficiency is high and compares well with predictions based on the diffusivity of the
 13      compounds.  The denuder can also be extracted to obtain gas-phase concentrations directly
 14      (Gundel and Lane, 1999). Particle-phase PAHs are taken to be the sum of material on the filter
 15      and XAD adsorbent downstream after correction for denuder collection efficiency. The IOVPS
 16      was tested for sampling semivolatile PAH in laboratory indoor air and in environmental tobacco
 17      smoke (ETS). After exposure, the denuders, filters, and sorbent traps were extracted with
 18      cyclohexane (Gundel et al., 1995) and analyzed for PAHs from naphthalene  to chrysene using
 19      dual-fluorescence detection (Mahanama et al, 1994).  Recoveries from both denuders and filters
 20      were approximately 70% for 30 samples. Detection limits (defined as 3 times the standard
 21      deviation of the blanks) for gas-phase SVOC PAHs ranged from 0.06 ng for  anthracene to 19 ng
 22      for 2-methylnaphthalene.  The 95% confidence interval for reproduction of an internal standard
 23      concentration was 6.5% of the mean value. Relative precision, from a propagation of errors
 24     analysis or from the 95% confidence interval from replicate analyses of standard reference
 25     material SRM 1649 (urban dust/organics), was 12% on average (8% for naphthalene to 22% for
 26     fluorene). Sources of error included sampling flow rate, internal standard concentration, and
 27     co-eluting peaks. Gundel and Lane (1999) reported that roughly two-thirds of particulate PAH
28     fluoranthene, pyrene, benz[a]anthracene, and chrysene were found on the postfilter denuders, so
29     that it is likely that considerable desorption from the collected particles took place.
30          Solid adsorbent-based denuder systems have been investigated by other researchers as
31      well. Bertoni et al. (1984) described the development of a charcoal-based denuder system for

        June 2003                                2A-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1      the collection of organic vapors. Risse et al. (1996) developed a diffusion denuder system to
 2      sample aromatic hydrocarbons. In their system, denuder tubes with charcoal coating and
 3      charcoal paper precede a filter pack for paniculate collection and an adsorption tube to capture
 4      particle blow-off from the filter sample.  Breakthrough curves for benzene, toluene, ortho-
 5      xylene, and meta-xylene were developed for 60-, 90-, and 120-cm denuder tubes.  The effects of
 6      relative humidity on the adsorption capacities of the denuder system were examined, and it was
 7      found that the capacity of the charcoal was not affected significantly by increases in relative
 8      humidity. The feasibility of outdoor air sampling with the system was demonstrated.
 9           Krieger and Hites (1992) designed a diffusion denuder system that uses capillary gas
10      chromatographic columns as the tubes for SVOC collection. The denuder was followed by a
11      filter to collect particles, which in turn was followed by a PUF plug to collect organic material
12      volatilizing off the filter. Denuder samples were analyzed by liquid solvent extraction (CHjCy
13      followed by GC-mass spectrometric analysis. The PUF plugs and filters were extracted with
14     supercritical fluid extraction using supercritical N2O. Using this system, an indoor air sample
15     was found to contain primarily chlorinated biphenyls, ranging from trichlorobiphenyls (vapor
16     pressures 10'3 - 10^ Torr at 25 °C) to octachlorobiphenyls (W6 - 10'7 Torr). This demonstrated
17     that the sampler collects compounds with a wide range of volatility. They also found that on-
18     line desorption is successful in maintaining good chromatographic peak shape and resolution.
19     The entire method, from sample collection to the end of the chromatographic separation, took
20     2h.
21           Organic acids in both the vapor and paniculate phases may be important contributors to
22     ambient acidity, as well as representing an important fraction of organic paniculate matter.
23     Lawrence and Koutrakis (1996a,b) used a modified Harvard/EPA annular denuder system
24     (HEADS) to sample both gas and particulate phase organic acids in Philadelphia, PA, in the
25     summer of 1992. The HEADS sampler inlet had a 2.1-um outpoint impactor (at 10 L/min),
26     followed by two denuder tubes, and finally a filter pack with a Teflon filter. The first denuder
27     tube was coated with potassium hydroxide (KOH) to trap gas phase organic acids. The second
28      denuder tube was coated with citric acid to remove ammonia and thus to avoid neutralizing
29      particle phase acids collected on the filter.  The KOH-coated denuder tube was reported to
 30      collect gas phase formic and acetic acids at better than 98.5% efficiency and with precisions of
 31      5% or better (Lawrence and Koutrakis,  1994). It was noted that for future field measurements of

        June 2003                                2A-6       DRAFT-DO NOT QUOTE OR CITE

-------
  1     paniculate organic acids, a Na2CO3-coated filter should be deployed downstream of the Teflon
  2     filter to trap organic acids that may evaporate from the Teflon filter during sampling.
  3
  4     Rote of the Collection Media
  5          The role of the collection media was recently examined in a study conducted in Seattle
  6     (Lewtas et al., 2001). In that study, the influence of denuder sampling methods and filter
  7     collection media on  the measurement of SVOC associated with PM2 5 was evaluated. Activated
  8     carbon and XAD collection media were used in diffusion denuders and impregnated back-up
  9     filters in two different samplers, the Versatile Air Pollution Sampler (VAPS) and the-PC-BOSS.
 10     XAD-coated glass annular denuders and charcoal-impregnated cellulose fiber (GIF) filter
 11     denuders also were used. CIF  filters also were compared to XAD-coated quartz filters as backup
 12     filter collection media.  Lewtas et al. (2001) found that the two denuder types resulted in an
 13     equivalent measurement of particulate organic carbon and particle mass.  The carbon-coated
 14     denuders in the BOSS sampler were more efficient than the XAD-coated denuders for the
 15     collection of the more highly volatile organic compounds (MHVOC).  Lewtas et al. (2001)
 16     concluded that this MHVOC that is collected in the carbon-coated BOSS  denuder does not
 17     contribute substantially to the particle mass or to the SVOC measured as OC on quartz filters.
 18     However, this MHVOC would be captured in the carbon impregnated filters placed behind the
 19     quartz filters, so that, in the XAD denuder configuration, the captured MHVOC would  cause a
 20     higher OC concentration and an overestimation of the SVOC.
 21           Some of the recent research in denuder technology also has focused on reduction  in the
 22     size of the denuder, optimization of the residence time in the denuder, understanding the effect
 23      of diffusion denuders on the positive quartz filter artifact, identifying changes in chemical
 24      composition that occur during sampling, determining the effects of changes  in temperature and
 25      relative humidity, and identifying possible losses by absorption in coatings.
 26
 27      Reducing the Size of Denuders
28           The typical denuder configuration is an annular diffusion denuder tube of significant length
29      (e.g., 26.5 cm for 10  L/min [Koutrakis et al., 1988a,b]). A more compact design based on a
30      honeycomb configuration was shown to significantly increase the capacity (Koutrakis et al.,
31      1993). However, in intercomparisons with an annular denuder/filter pack  system (Koutrakis

        June 2003                                2A-7       DRAFT-DO NOT QUOTE  OR CITE

-------
 1      et al., 1988), significant losses of ammonia and nitric acid were observed for the honeycomb
 2      configuration and were attributed to the large inlet surface area and long sample residence time
 3      of the honeycomb design relative to the annular denuder system. Sioutas et al. (1996a)
 4      subsequently designed a modified glass honeycomb denuder/filter pack sampler (HDS) with an
 5      inlet that minimizes vapor losses on the inlet surfaces.  The modified HDS has reduced inlet
 6      surfaces and decreased residence time for sampled gases (NH3 and HNO3) compared to its
 7      predecessor (Sioutas et al., 1994d). Sioutas et al. (1996b) also tested various inlet materials
 8      (glass, PFA, and polytetrafluoroethylene [PTFE]) in laboratory tests and found that a PTFE
 9      Teflon coated inlet minimized loss of sampled  gases (1 to 8% loss of HNO3 observed, and -4 to
10      2% loss of NH3 observed). The highest inlet losses were observed for HNO3 lost to PFA
11      surfaces (14 to 25%). The modified HDS was  tested in laboratory and field tests and found to
12      agree within 10% with the annular denuder system.
13
14     Residence Time in the Denuder
15            The  efficiency of a diffusion denuder sampler for the removal of gas phase material can be
16     improved  by increasing the residence time of the sampled aerosol in the denuder.  However, the
17     residence  time can only be increased within certain limits.  Because the diffusion denuder
18     reduces the concentration of gas-phase semivolatile organic material, semivolatile organic matter
19     present in the particles passing through the denuder will be in a thermodynamically unstable
20     environment and will tend to outgas SVOC during passage through the denuder. The residence
21     time of the aerosol in the denuder, therefore, should be short enough to prevent significant loss
22     of particulate phase SVOC to the denuder.  Various studies have suggested that the residence
23     time in the denuder should be less than about 2 s (Gundel and Lane, 1999; Kamens and Coe,
24      1997; Kamens et al., 1995). The residence times in the various denuder designs described by
25     Gundel and Lane (1999) are from 1.5 to 0.2 s.  The equilibria and evaporation rates are not as
26     well understood for organic components as they are for NH4NO3 (Zhang and McMurry, 1987,
27      1992; Hering and Cass, 1999).
28
29     Effect of Diffusion Denuders on the Positive  Quartz Filter Artifact
30           The adsorption of organic compounds by a second quartz filter has been shown to be
31      reduced, but not eliminated, in samples collected in the Los Angeles Basin if a multi-channel

        June 2003                               2A-8       DRAFT-DO NOT QUOTE OR CITE

-------
   1      diffusion denuder with quartz filter material as the denuder collection surface preceded the
  2      quartz filters (Fitz, 1990). This artifact can be further reduced by the use of activated charcoal as
  3      the denuder surface and the use of a particle concentrator to reduce the amount of gas phase
  4      organic compounds relative to condensed phase organic compounds (Cui et al., 1997,1998;
  5      Eatough, 1999). Other experiments (Gotham and Bidleman, 1992; Cui et al.,  1998; Eatough
  6      et al., 1995, 1996) have shown that the quartz filter artifact can result both from the collection of
  7      gas phase organic compounds and from the collection of semivolatile organic compounds lost
  8      from particles during sampling. Thus, results available to date suggest that both a "positive" and
  9      a "negative" artifact can be present in the determination of particulate phase organic compounds
 10      using two tandem quartz filters.
 11           The importance of the adsorption of organic vapors on filters or PM relative to the
 12      volatilization of organic compounds from PM collected on a filter continues to be a topic of
 13      active debate. The relative importance of positive and negative artifacts will be different for
 14      denuded and undenuded filters; will depend on face velocity, sample loading, and the vapor
 15      pressures of the compounds of interest; and may vary with season and location because of
 16      variations in the composition of volatile and semivolatile organic material. Evidence exists for
 17      substantial positive and negative artifacts in the collection of organic PM.
 18
 19      Changes in Chemical Composition During Sampling
 20           The use of sampling systems designed to correctly identify the atmospheric gas and
 21      particulate phase distributions of collected organic material has been outlined above.
 22     An additional sampling artifact that has received little consideration in the collection of
 23      atmospheric samples is the potential alteration of organic compounds as a result of the sampling
 24      process. These alterations appear to result from the movement of ambient air containing
 25      oxidants and other reactive compounds past the collected particles.  The addition of NO2
 26      (]pyrene and benzo[a]anthracene from a few % to 38%,
 28      with the observed reduction increasing with increased concentration of the added gases
29      (Brorstrom et al., 1983).  Spiking a filter with an amine resulted in an increase in measured
30      concentrations of nitrosamines in both the filter and a following XAD sorbent bed for a
31      mid-volume sampler (Ding et al., 1998a,b). Similar results have been obtained for the exposure

        June 2003                                 2A-9       DRAFT-DO NOT QUOTE OR CITE

-------
 1      of a deuterated amine on a filter to NOX (Pellizzari and Krost, 1984). When Tenax columns
 2      spiked with deuterated styrene and cyclohexene were exposed to ppm concentrations of ozone or
 3      halogens, oxygenated and halogenated compounds were formed (Pellizzari and Krost, 1984).
 4      Similar oxidation of aldehydes and peroxyacetylnitrate (PAN) during sampling has been
 5      observed (Grosjean and Parmar, 1990). Collected PAH compounds can be oxygenated or
 6      nitrated on a filter (Davis et al., 1987; Lindskog and Brorstrom-Lunden, 1987), but 1-nitropyrene
 7      has been shown to be resistant to additional nitration (Grosjean, 1983). These various chemical
 8      transformations of collected organic compounds can be eliminated by removal of the gas phase
 9      oxidants, NOX, HNO3, etc., by reaction or adsorption prior to collection of the particles (Ding,
10      1998a,b; Grosjean and Parmar, 1990; Parmar and Grosjean, 1990; Pellizzari and Krost, 1984;
11      Williams and Grosjean, 1990).  The BOSS denuder should be effective in eliminating most of
12     the chemical transformation artifacts because reactive gases are removed by the charcoal
13      denuder that precedes the particle collection filter.
14
15     Temperature and Relative Humidity Effects
16           The problems of sampling artifacts associated with SVOC adsorption and evaporation are
17     compounded by temperature  and relative humidity effects (Pankow and Bidleman, 1991;
18     Pankow et al., 1993; Falconer et al., 1995; Goss and Eisenreich, 1997). Effects of temperature
19     on the partitioning of PAH were examined by Yamasaki et al. (1982), who found that the
20     partition coefficient AHvapor/PAHpart) was inversely related to temperature and could be
21     described using the Langmuir adsorption concept. The dissociation of ammonium nitrate aerosol
22     is also a function of temperature.  Bunz et al. (1996) examined the dissociation and subsequent
23     redistribution of NH4NO3 within a bimodal distribution using a nine-stage low-pressure Berner
24      impactor followed by analysis by ion chromatography and found a strong temperature
25      dependency on the redistribution. Bunz et al. (1996) found that at lower temperatures (below
26      10 °C) there was little change in the aerosol size distribution.  At temperatures between 25 and
27      45 °C, however, the lifetime  of NH4NO3 particles decreases by more than a factor of 10,  and size
28      redistribution, as measured by average ending particle diameter, increased more for higher
29      temperatures than for lower temperatures.
 30           The effects of relative humidity on the sorption of SVOC on particles are not well
 31      understood.  In a series of laboratory experiments, Goss and Eisenreich (1997) examined the

        June 2003                                2A-10       DRAFT-DO NOT QUOTE OR CITE

-------
  1     sorption of both nonpolar (hydrocarbons and chlorinated hydrocarbons) and polar (ethyl ether
  2     and acetone) volatile organic compounds onto combustion soot particles as a function of
  3     temperature and relative humidity. The soot particles used in their experiments were collected
  4     from oil furnaces and contained 60% (w/w) iron sulfate (water-soluble fraction) and 9% (w/w)
  5     elemental and organic carbon. They found that, for all  compounds, the sorption of VOC onto
  6     soot particles decreased with increasing relative humidity over the range of 10 to 95%.  They
  7     also observed hysteresis in the relative humidity dependency, with sorption coefficients at a
  8     given relative humidity higher when the RH is being increased than when the RH is being
  9     decreased. The sorption coefficients were fit with an exponential function to the RH so that the
 10     slope of the regression line would provide a measure of the influence of relative humidity.
 11     Based on the magnitude of the slope, they concluded that the RH-dependency of sorption was
 12     stronger for water-soluble organic compounds.
 13          In another study by Jang and Kamens (1998), humidity effects on gas-particle partitioning
 14     of S VOC were examined using outdoor environmental  chambers and the experimentally
 15     determined partitioning coefficients were compared to theoretical values. They examined the
 16     partitioning of SVOC onto wood soot, diesel soot, and secondary aerosols and concluded that
 17     "the humidity effect on partitioning was most significant for hydrophobic compounds adsorbing
 18     onto polar aerosols." Although these two studies seem  to be contradictory, on closer
 19     examination, it is difficult to compare the two studies for several reasons. The experiments
 20     conducted by Jang and Kamens (1998) were conducted in outdoor chambers at ambient
 21      temperatures and humidities.  Their model was for absorptive partitioning of SVOC on
 22     liquid-like atmospheric particulate matter. In contrast, the results of Goss and Eisenreich (1997)
 23      were obtained from a gas chromatographic system operated at 70 °C higher than ambient
 24      conditions.  The model of Goss and Eisenreich (1997) was for adsorptive partitioning of VOC on
 25      solid-like atmospheric particulate matter. In the study of Jang and Kamens (1998), calculated
 26      theoretical values for water activity coefficients for diesel soot were based on an inorganic salt
27      content of 1 to 2%; whereas, the combustion particles studied by Goss and Eisenreich (1997)
28      contained 60% water-soluble, inorganic salt content.  Jang and Kamens (1998) obtained their
29      diesel soot from their outdoor chamber, extracted it with organic solvent (mixtures of hexane and
30      methylene chloride), and measured the organic fraction. The resulting salt content of 2% of the
       June 2003                                2A-11       DRAFT-DO NOT QUOTE OR CITE

-------
 1     participate matter studied in Jang and Kamens (1998) is enough to affect water uptake but
 2     presumably not to affect the sorption partitioning of organics.
 3
 4     Impactor Coatings
 5          fmpactors are used as a means to achieve a size cutpoint and as particle collection surfaces.
 6     Particles collected on impactors are exposed to smaller pressure drops than filter-collected
 7     particles, making them less susceptible to volatile losses (Zhang and McMurry, 1987).
 8     However, size resolution can be affected by bounce when samples are collected at low
 9     humidities (Stein et al., 1994).  There are other sources of error inherent in some of the currently
10     acceptable practices that could potentially affect paniculate mass concentration measurements
11     and that will surely become even more important as more emphasis is placed on chemical
12     speciation. Allen et al. (1999a) reported that the practice of greasing impaction substrates may
13     introduce an artifact from the absorption of semivolatile species from the gas phase by the grease
14     because the grease could artificially increase the amount of PAHs and other organic compounds
15     attributed to the aerosol. Allen et al. (I999b) offer several criteria to ensure that this absorption
16     artifact is negligible, including selecting impaction oils in which analytes of interest are
17     negligibly soluble and ensuring that species do not have time to equilibrate between the vapor
18     and oil phases (criterion is met for nonvolatile species). They recommend using oiled impaction
19     substrates only if the absorption artifact is negligible as determined from these criteria.
20     Application of greases and impaction oils for preventing or reducing bounce when sampling with
21     impactors is not suitable for carbon analysis because the greases contain carbon (Vasilou et al.,
22     1999).
23           Kavouras and Koutrakis (2001) investigated the use of polyurethane foam (PUF) as a
24     substrate for conventional inertial impactors. The PUF impactor substrate is not rigid like the
25     traditional impactor substrate so particle bounce and reentrainment artifacts are reduced
26     significantly.  Kavouras and Koutrakis (2001) found that the PUF impaction substrate resulted in
27     a much smaller 50% cut point at the same flow rate and Reynolds number. Moreover, the lower
28     50% cut point was obtained at a lower pressure drop than with the conventional substrate, which
29     could lead to a reduction of artifact vaporization of semivolatile components.
30
        June 2003                                2A-12       DRAFT-DO NOT QUOTE OR CITE

-------
   1                     APPENDIX 2B.  ANALYTICAL TECHNIQUES
  2
  3
  4      2B.1  INORGANIC ELEMENTS
  5      2B.1.1  Energy Dispersive X-Ray Fluorescence (EDXRF)
  6          EDXRF has usually been the method of choice for analysis of trace elements on filters.
  7      EDXRF is preferred for aerosol analysis over wavelength dispersive XRF because it allows fast
  8      and simultaneous analysis over the total spectrum, allowing for the analysis of numerous
  9      elements simultaneously. EDXRF can accommodate small sample sizes and requires little
 10      sample preparation or operator time after the samples are placed into the analyzer.  It also leaves
 11      the sample intact after analysis; so, further analysis is possible. XRF irradiates a uniform
 12      particle deposit on the surface of a membrane filter with 1 to 50 kev x-rays that eject inner shell
 13      electrons from the atoms of each element in the sample (Dzubay and Stevens, 1975; Jaklevic et
 14      al., 1977; Billiet et al., 1980; Potts and Webb, 1992; Piorek, 1994; Bacon et al., 1995; deBoer et
 15      al., 1995; Holynska et al., 1997; Torok et al., 1998; Watson et al., 1999). When a higher energy
 16     electron drops into the vacant lower energy orbital, a fluorescent x-ray photon is released. The
 17     energy of this photon is unique to each element, and the number of photons is proportional to the
 18     concentration of the element. Concentrations are quantified by comparing photon counts for a
 19     sample with those obtained from thin-film standards of known concentration (Dane et al., 1996).
 20     The previous 1996 PM AQCD included a detailed discussion of EDXRF.
 21           Emitted x-rays with energies less than ~4 kev (affecting the elements sodium, magnesium,
 22     aluminum, silicon, phosphorus, sulfur, chlorine, and potassium) are absorbed in the filter, in a
 23      thick particle deposit, or even by large particles in which these elements are contained. Very
 24      thick filters also scatter much of the excitation radiation or protons, thereby lowering the
 25      signal-to-noise ratio for XRF and PIXE. For this reason, thin membrane filters with deposits in
 26      the range of 10 to 50 jig/cm2 provide the best accuracy and precision for XRF and PIXE analysis
 27      (Davis et al, 1977; Haupt et al., 1995).
 28
 29      2B.1.2  Synchrotron Induced X-ray  Fluorescence (S-XRF)
 3®           S-XRF is a form of EDXRF in which the exciting x-rays are derived from a synchrotron.
31      Bremmstrahlung x-rays are generated when energetic electrons (generally several GeV in

        June 2003                                2B-1        DRAFT-DO NOT QUOTE OR CITE

-------
 1      energy) are forced by a magnetic field to make a bend in their path.  The advantages of the
 2      technique are that an extremely high flux of x-rays may be obtained and that the x-rays are 100%
 3      polarized in the plane of the electron beam. The former allows for x-ray beams generally of
 4      50 to 500 urn in diameter. However, the beams can be focused into x-ray microprobes, with spot
 5      sizes on the order of one urn diameter. The x-ray polarization allows for removal of most of the
 6      background normally found under the characteristic x-ray peaks, greatly improving sensitivity
 7      compared to other XRF techniques.  The primary disadvantages are the limited number of
 8      synchrotrons, and that few synchrotrons have S-XRF capabilities. Thus, the technique has been
 9      relatively little used for PM, and then generally for special problems such as the smoke from the
10      Kuwaiti oil fires (Cahill et al., 1992, Reid et al., 1994).  However, with the increasing
11      availability of S-XRF facilities dedicated to PM analysis, the first of which was the Advanced
12     Light Source opened at Lawrence Berkeley National Laboratory last year, utilization of S-XRF
13      for PM analysis is increasing.
14
15     2B.1.3  Proton (or Particle) Induced X-ray Emission (PIXE)
16           PIXE differs from XRF analysis in the excitation source for producing fluorescence. The
17     filter deposit is bombarded with high-energy protons to remove inner shell electrons and the
18     resulting characteristic x-rays are analyzed as in XRF (Johansson, 1970, Cahill, 1981,1985;
19     Zeng et al., 1993). Small accelerators, generally Van de Graaffs, generate intense beams of low
20     energy protons, generally of a few MeV in energy. These have the  ability to remove electrons
21     from inner shells of atoms of any element. Thus, PIXE can see a very wide range of elements in
22     a single analysis. The cross section for producing x-rays using protons of a few MeV in energy
23     tends to favor lighter elements, Na through Ca, but sensitivities for equivalent PIXE and
24     multi-wavelength XRF analysis are roughly comparable.  The technique has been widely used in
25     the U.S. (Flocchini et al., 1976, Malm et al., 1994) and around the world, as many universities
26     have the small accelerators needed for the method.  Like S-XRF, the proton beams can be
27     focused into urn size beams, but these have been relatively little used for PM. However, the mm
28      size beams used in both S-XRF and PIXE are well suited to analyzing  the limited mass and small
29      deposits that result from detailed particle size measurements by impactors (Perry et al.,  1999).
 30
 31

        June 2003                                2B-2        DRAFT-DO NOT QUOTE  OR CITE

-------
  1     2B.1.4  Proton Elastic Scattering Analysis (PESA)
  2          With the routine availability of elemental analyses for all elements sodium and heavier,
  3     organic components remain the major unmeasured species for mass balance. For programs like
  4     IMPROVE (Malm et al., 1994), parallel filters are collected for separate organic and elemental
  5     carbon determinations. Aerosol programs that use PIXE can directly measure hydrogen
  6     simultaneously by scattering protons from Teflon filters that lack hydrogen (Cahill et al., 1989,
  7     1992). Generally, results from organic matter by carbon combustion from quartz filters and
  8     organic matter by hydrogen from Teflon filters are in agreement, assuming certain assumptions
  9     about the chemical states of sulfates and nitrates are met (Malm et al., 1994, Cahill et al., 1996).
 10
 11      2B.1.5  Total Reflection X-ray Fluorescence (TRXRF)
 12           One of the limitations of the EDXRF method is the minimum detection limit, which may
 13      be high due to high background values  (Streit et al., 2000). By implementation of x-ray optical
 14      geometries that use the total reflection of the primary radiation on flat surfaces, scattering on the
 15      substrate is reduced, so that detection limits can be reduced. This is the basis for the total
 16      reflection x-ray fluorescence (TRXRF) method (Aiginger and Streli, 1997). This modification to
 17      the EDXRF technique improves detection limits and avoids the need to correct for matrix
 18      effects. Despite its apparent advantages, TRXRF has not yet become widely in use for
 19      atmospheric aerosol analysis but has been used in the analysis of marine aerosol (Stahlschmidt et
 20      al., 1997) and at a high elevation site (Streit et al., 2000).  Streit et al. sampled ambient air at the
 21      High Alpine Research Station (3580 m  above sealevel) in the Bernese Alps, Switzerland, using a
 22      nine-stage, single-jet, low-pressure, cascade impactor equipped with quartz impactor plates
 23      coated with silicon oil diluted in 2-propanol. The typical sample volume for a weekly sample
 24      was 10 m3. The quartz plates were analyzed directly by TRXRF. Streit et al. reported that the
 25      minimum detection limits, defined by the 3o values of the blanks, ranged from 25 ng for S,
 26      decreased monotonically with increasing atomic number down to 5 pg for Rb, and decreased
 27      after that.  The use of TRXRF is expected to increase as EDXRF users become aware of the
 28      method. A relatively low-cost, add-on unit has been developed that would allow EDXRF users
29      to test the TRXRF technique (Aiginger, 1991).
30
31

        June 2003                                2B-3        DRAFT-DO NOT QUOTE OR CITE

-------
 1      2B.1.6  Instrumental Neutron Activation Analysis (INAA)
 2           INAA irradiates a sample in the core of a nuclear reactor for few minutes to several hours,
 3      depending on the elements being quantified (Dams et al., 1970; Zoller and Gordon, 1970;
 4      Nadkarni, 1975; Landsberger, 1988; Olmez, 1989; Ondov and Divita, 1993). The neutron
 5      bombardment chemically transforms many elements into radioactive isotopes. The energies of
 6      the gamma rays emitted by these isotopes identify them and, therefore, their parent elements.
 7      The intensity of these gamma rays is proportional to the amount of the parent element present in
 8      the sample. Different irradiation times and cooling periods are used before counting with a
 9      germanium detector. In source apportionment studies, it is possible to use a combination of XRF
10      and INAA to develop a relatively complete set of elemental measurements.  Between these two
11      analytical techniques, good sensitivity is possible for many elements, including most of the toxic
12      metals of interest. In general, XRF provides better sensitivity for some metals (e.g., Ni, Pb, Cu,
13      and Fe); whereas INAA provides better sensitivity for others (Sb, As, Cr, Co, Se, and  Cd). Both
14     methods provide similar detection limits for still other elements (V, Zn, and Mn).  INAA does
15     not quantify some of the abundant species in ambient particulate matter such as silicon, nickel,
16     tin, and lead. While INAA is technically nondestructive, sample preparation involves folding
17     the sample tightly and sealing it in plastic, and the irradiation process makes the filter membrane
18     brittle and radioactive.  These factors limit the use of the sample for subsequent analyses.
19           INAA has been used to examine the chemical composition of atmospheric aerosols in
20     several studies either as the only method of analysis or in addition to XRF (e.g., Yatin et al.,
21     1994; Gallorini, 1995). INAA has higher sensitivity for many trace species, and it is particularly
22     useful in analyzing for many trace metals. Landsberger and Wu (1993) analyzed air samples
23     collected near Lake Ontario for Sb, As, Cd, In, I, Mo, Si, and V using INAA. They
24     demonstrated that using INAA in conjunction with epithermal neutrons and Compton
25     suppression produces very precise values with relatively low detection limits.
26           Enriched rare-earth isotopes have been analyzed via INAA and used to trace sources of
27     particulate matter from a coal-fired power plant (Ondov et al., 1992), from various sources in the
28     San Joaquin Valley (Ondov, 1996), from intentially tagged (iridium) diesel emissions from
29     sanitation trucks (Suarez et al., 1996; Wu et al., 1998), and from iridium-tagged emissions from
30     school buses (Wu et al., 1998).
        June 2003                                2B-4        DRAFT-DO NOT QUOTE OR CITE

-------
  1          An intercomparison was conducted in which 18 pairs of filters were sent to participants in
  2     the Coordinated Research Program (CRP) on Applied Research on Waste Using Nuclear Related
  3     Analytical Techniques (Landsberger et al., 1997). As part of that study, participants used PIXE,
  4     INAA, XRF, or AAS to analyze the samples. Many of the results for XRF and PIXE in the
  5     coarse fraction were observed to be biased low compared to INAA. The authors speculated that
  6     there is a systematic error because of self-attenuation of the x-rays resulting from the particle
  7     size effect.
  8
  9     2B.1.7  Atomic Absorption Spectrophotometry (AAS)
 10          AAS is applied to the residue of a filter extracted in a strong solvent to dissolve the solid
 11      material; the filter or a portion of it is also dissolved during this process (Ranweiler and Moyers,
 12     1974; Fernandez, 1989; Jackson and Mahmood, 1994; Chow et al., 2000a).  A few milliliters of
 13      this extract are injected into a flame where the elements are vaporized. Elements absorb light at
 14      certain wavelengths in the visible spectrum, and a light beam with wavelengths specific to the
 15      elements being measured is directed through the flame to be detected by a monochrometer. The
 16      light absorbed by the flame containing the extract is compared with the absorption from known
 17      standards to quantify the elemental concentrations. AAS requires an individual analysis for each
 18      element, and a large filter or several filters are needed to obtain  concentrations for a large variety
 19      of elements.  AAS is a useful complement to other methods, such as XRF and PIXE, for species
 20      such as beryllium, sodium, and magnesium which are not well-quantified by these methods.
 21      Airborne particles are chemically complex and do not dissolve easily into complete solution,
 22      regardless of the strength of the solvent. There is always a possibility that insoluble residues are
 23      left behind and that soluble species may co-precipitate on them or on container walls.
 24           AAS was used to characterize the atmospheric deposition of trace elements Zn, Ni, Cr, Cd,
 25      Pb, and Hg to the Rouge River watershed by paniculate deposition (Pirrone and Keeler, 1996).
 26      The modeled deposition rates were compared to annual emissions of trace elements that were
 27      estimated from the emissions inventory for coal and oil combustion utilities, iron and steel
28      manufacturing, metal production, cement manufacturing, and solid waste and sewage sludge
29      incinerators.  They found generally good agreement between the trend observed in atmospheric
30      inputs to the river (dry + wet deposition) and annual emissions of trace elements, with R2s
31      varying from =0.84 to 0.98. Both atmospheric inputs and emissions were found to have

        June 2003                                2B-5        DRAFT-DO NOT QUOTE OR CITE

-------
 1      followed downward trends for Pb. For the period of 1987 to 1992, steady increases were
 2      observed for Cd (major sources are municipal solid waste incineration, coal combustion, sludge
 3      incineration, and iron and steel manufacturing), Cr and Ni (major sources are iron and steel
 4      production and coal combustion), and Hg (major sources are coal, the contribution from which
 5      had decreased from 53 to 45%, and municipal, solid, and medical waste incineration, the
 6      contribution from which has increased).
 7
 8      2B.1.8 Inductively Coupled Plasma with Atomic Emission Spectroscopy
 9              (ICP-AES)
10           ICP-AES introduces an extracted sample into an atmosphere of argon gas seeded with free
11      electrons induced by high voltage from a surrounding Tesla coil (Fassel and Kniseley, 1974;
12      McQuaker et al., 1979; Lynch et al., 1980; Harman,  1989; Tyler, 1992; Baldwin et ah, 1994).
13      The high temperatures in the induced plasma raise valence electrons above their normally stable
14     states.  When these electrons return to their stable states, a photon of light is emitted that is
15      unique to the element that was excited.  This light is detected at specified wavelengths to identify
16     the elements in the sample.  ICP-AES acquires a large number of elemental concentrations using
17     small sample volumes with acceptable detection limits for atmospheric samples. As with AAS,
18     this method requires complete extraction and destruction of the sample.
19
20     2B.1.9 Inductively Coupled Plasma with Mass Spectroscopy (ICP-MS)
21          ICP-MS has been applied in the analysis of personal exposure samples (Tan and Horlick,
22     1986; Gray and Williams, 1987a,b; Nam et ah, 1993; Munksgaard and Parry, 1998; Campbell
23     and Humayun, 1999). Ion species generated from ICP and from the sample matrix can produce a
24     significant background at certain masses resulting in formation of polyatomic ions that can limit
25     the ability of ICP-MS to determine some elements of interest. Cool plasma techniques have
26     demonstrated the potential to detect elements at the ultra-trace level (Nham et ah, 1996) and to
27     minimize common molecular ion interferences (Sakata and Kawabata, 1994; Turner, 1994;
28     Plantz, 1996). Detection limits of ICP-MS using a one-second scan are typically in the range of
29      10"3 ng/m3, which is an order of magnitude lower than other elemental analysis methods.  The
30      instrument can also be set up to analyze a wide dynamic range of aerosol concentrations.
        June 2003                                2B-6       DRAFT-DO NOT QUOTE OR CITE

-------
  1      Isotope analysis can also be performed with ICP-MS. Intercomparison studies are needed to
  2      establish the comparability of ICP-MS with other non-destructive filter analysis methods.
  3           Keeler and Pirrone (1996) used ICP-MS to determine trace elements Cd, Mn, V, As, Se,
  4      and Pb in atmospheric fine paniculate (PM2.5) and total suspended particulate samples collected
  5      in two Detroit sites. The results were used in a deposition model with additional measurements
  6      using AAS to estimate the dry deposition flux of trace elements to Lake Erie.
  7
  8      2B.1.10  Scanning Electron Microscopy (SEM)
  9           Mamane et al. (2001) investigated the use of computer-controlled scanning electron
 10      microscopy (CCSEM) as a way of supplementing XRP analysis and providing automated
 11      analysis of particle size, chemistry, and particle classification. An ambient coarse particulate
 12      sample from Baltimore was collected on a polycarbonate filter for this analysis. CCSEM
 13      analyses were conducted for 2819 particles in 78 randomly selected fields of view during an
 14      unattended 8-h run. Mamane et al. confirmed the stability of the CCSEM-instrument over
 15      several hours of operation. The physical properties of the sample such as particle diameter, mass
 16      loading per field, and particle number per field were well represented by analyzing
 17      approximately 360 particles with little additional information gained by analyzing more
 18      particles. Teflon filters are not well suited for SEM analyses.  Analysis of fine PM is expected
 19      to pose analytical challenges not addressed in the present study (Mamane et al., 2001).
20           Nelson et al. (2000) applied Raman chemical imaging and SEM (Raman/SEM) to study the
21      size, morphology, elemental and molecular composition, and molecular structure of fine
22      particulate matter.  In their study, filter compatibility was examined, and Raman/SEM chemical
23      imaging was conducted for several standard materials as well as for ambient PM2 5 samples.
24      Polycarbonate was determined to be a suitable substrate for both SEM and Raman chemical
25      imaging analysis.
26           Conner et al. (2001) used  CCSEM with individual X-ray analysis to study the chemical and
27      physical attributes of indoor and outdoor aerosols collected around a retirement home in
28      Baltimore. The CCSEM technique was demonstrated to be capable  of identifying spherical
29      particles typical of combustion or other high temperature (presumably industrial) processes as
30      well as pollens and spores. Indoor particles originating from cosmetics were also identified.
31

        June 2003                                2B-7        DRAFT-DO NOT QUOTE OR CITE