American Petroleum Institute Air Research Programs

1
American Petroleum Institute Air Research
Programs

• Howard Feldman, Director
• Regulatory Analysis and Scientific Affairs

Palo Alto, CA July 27, 2005
2
March 2005 PERF Meeting

• Two Research Programs
• Smart LDAR (Leak Detection and Repair)
• Mobile Source Air Toxics

3
Characterization of PM2.5 from Stationary
Sources
4
Background

• Chemically speciated PM2.5 emissions data are
  needed
• for SIP development
• environmental assessments
• emission inventories
• source species fingerprints
• source apportionment

5
Background

• Traditional air emissions test methods for
  stationary sources (hot filter/iced impinger
  methods e.g. EPA 5 202, ARB 5, etc.)
• Not accurate or precise enough for low
  concentrations of many current sources (e.g.,
  gas-fired power plants)
• Bias due to background and artifacts can be very
  significant
• Contributes significantly to larger-than-expected
  range of results for similar sources e.g.
  gas-fired power generation
• Limited capability for chemical/physical
  characterization of PM2.5

6
Project Overview

• Goals
• Develop and standardize improved dilution
  sampling technology/methods for PM
• Develop preliminary emission factors and
  speciation profiles for PM2.5 and precursors
• 4-year project (2000-2004)
• Build on prior work sponsored by API, GRI and DOE
• Laboratory (pilot-scale) tests method
  development
• Field tests emissions and species
• 3 Combined cycle and cogeneration power plants
  (gas)
• 1 Diesel engine for backup generator (no
  controls DPF)
• 1 boiler (gas No. 6 oil)
• 2 process heaters (gas)

7
Field Tests
Dilution Sampler
Traditional Stack Sampling

• PRIMARY PM2.5
• PM2.5 mass (gravimetric)
• 40 Elements Al-Zn (XRF)
• OC/EC (TOR)
• SVOC (PUF/XAD, GC/MS)
• Ions SO4, NO3-, Cl- (IC) NH4 (colorimetry)
• Chemically-speciated ultrafine particles (MOUDI)
• Ultrafine size distribution (SMPS)
• PM2.5 PRECURSORS
• SO2, NH3 (impregnated filters, IC, colorimetry)
• VOCC8 (Tenax, GC/MS)
• OTHER
• Carbonyls (sorbent tube, HPLC)
• VOCC2 (canisters, GC/MS)

stack
PM10, PM2.5, filterable and Condensable
Particulate (cyclones, heated filter, Impinger
train)
NO, NOx, SO2, CO, CO2, O2 (continuous gas
analyzers)
Solid/Condensable Particle size dist. (dual
cascade impactors)
Ammonia (in-stack filter, impingers)
SO3 (controlled condensation)
Ambient Air and Stack
Combustion Sources
8
PM2.5 Mass Gas Combustion

• In-stack method results dominated by sulfate in
  impingers (SO2 artifact)

Dilution results for gas combustion consistently
1/10 or lower compared to hot filter/iced
impinger method results
9
Speciated PM2.5 Natural Gas
OC accounts for gt90 of PM2.5 mass (measurement
background and artifact?)
10
Speciated PM2.5 Natural Gas
Elements
Most elements except S not significant
11
Summary

• Developed a compact dilution sampler technology
  for PM2.5 stationary source stack sampling
• Improved portability, accuracy sensitivity for
  stationary source PM2.5 measurements
• ASTM standard development in progress
  (D22.03/WI752)
• Developed new speciated PM2.5 and precursor
  emissions data for power plants and other sources
  (7 field tests)
• PM2.5 mass from gas-fired sources is much lower
  using dilution sampling than traditional hot
  filter/iced impinger methods
• Chemical/physical sampling artifacts positively
  bias iced impinger results
• PM2.5 mass speciation vary with source type
  fuel
• Carbon sulfate are important components

12
Recommendation

• Further validation refinement, especially for
  very clean sources
• Background levels in dilution air probe
  recovery solvent
• Organic carbon artifacts
• Relative variability of results

13
Emissions of Polyaromatic Hydrocarbons from
Refinery Heavy Liquid Streams
14
Background

• Actual PAH Emission Rates not Known
• Existing Estimation Techniques Likely
  Overestimate, Particularly Heavy Molecular Weight
  PAHs
• Growing Concerns about PAH Emissions,
  particularly Naphthalene
• Potentially Large Expense by Industry to Quantify
  PAH Emissions

15
Current Work

• 2004 Study Evaluated Five Leaks and Two Heavy
  Liquid Types
• Study Evaluated Volatile Emissions, Liquid
  Deposition, and Stream Composition
• Study used Conventional Bagging Methodologies to
  Capture Emissions
• Samples were Collected over an Approximate
  24-Hour Period

16
Preliminary Results

• Only the Lighter Molecular Weight PAHs Enter the
  Air Pathway
• Partitioning of Individual PAHs between Vapor and
  Liquid Phase Dependent on Molecular Weight
• None of the Carcinogenic PAHs Detected in Vapor
  Phase Emissions, All in Liquid Deposited on Valve

17
Next Steps

• Additional Study Scheduled for Summer 2005
• Eight More Sources to be Evaluated
• Additional Heavy Liquid Types
• Increase Concentration (Method 21) Range of
  Leaking Components to get Better Method 21 vs
  Emission Rate Correlations

18
Development of SO3/SOX Emission Factors for
Gas-Fired Sources FCCUs
19
Background

• TRI requires H2SO4 emissions reporting for
  refineries, power plants and other sources
• Future PM2.5 issues
• Key Refinery Sources
• Gas combustion (process and natural gas)
• Fluid Catalytic Cracking Units (FCCUs)
• Emission factors may be biased
• No data or validated methods for gas
• Potential positive bias in Method 8 (SO2 to SO4)
• Old data/less sensitive methods
• Need improved SO3/SOX factors for more reliable
  estimates

20
FCCU SO2 to SO3 Conversion

• FCCUs burn coke to to regenerate catalyst
• Sulfur and trace elements tend to concentrate in
  the coke
• Some elements catalyze SO2 to SO3 conversion
• Literature review shows high SO3 conversions
  (gtgt2-3) for some sources esp. for low
  concentrations

SO2 gt 200 ppm
(Nie et al., Oil Gas Journal, Feb 2004)
21
Objectives

• Lab Study
• Evaluate potential SO2 and NH3 bias in EPA Method
  8 and controlled condensation method
• SO2 conversion to SO4 accelerated by NH3?
• Field Study
• Develop SO3/SOX factors for gas combustion and
  FCCU
• Compare Method 8 and controlled condensation

22
Field Tests

• Refinery sources
• 2 gas-fired units
• FCCU
• Simultaneous Method 8 and Controlled Condensation
• Paired sampling trains
• 1-hour test runs

23
Next Steps

• Complete lab tests (2Q 2005)
• Conduct field tests (3Q 2005)
• Data Analysis (4Q 2005)
• Report (1Q 2006)

24
MERCURY in U.S. CRUDE OIL
25
PARTICIPANTS

• U.S. EPA Office of Research and Development
• David Kirchgessner, Program Manager
• Mercury Technology Services
• S. Mark Wilhelm, Project Manager
• American Petroleum Institute
• National Petrochemical and Refiners Association

26
OBJECTIVES

• Determine the mean concentration and range of
  concentrations of mercury in crude oil processed
  in the U.S.
• Data must be statistically significant
• Sampling and analysis methods must reflect the
  best science currently available

27
LABORATORIES

• CEBAM ANALYTICAL (Lian Liang)
• FRONTIER GEOSCIENCES (Carl Hensman)

28
TECHNICAL APPROACH

•  Phase 1 Analytical Methods
• Phase 2 Sampling Methods and Oil Variability
• Phase 3 Statistical Sampling and Analysis

29
STATUS

• Approximately 100 market-named oils have been
  sampled and analyzed.
• Oils come from both domestic and foreign sources.
• The goal for statistical certainty is
  approximately 200 - 300 oils.
• Each oil is sampled and analyzed a minimum of 3
  times
• The presently estimated mean is less than 10 ppb

30
Ozone Health Impacts
31
O3 Airway Inflammation (AI)

• Goal - Optimize noninvasive AI assay
• Study Assay AI breath markers, lung
  function-symptoms in responders non-responders
  to 0.35 ppm-hr O3 exposures
• Anticipated Findings Responders have elevated
  AI (lagged 4 hrs) function-symptom responses
  but non-responders do not

32
AI Biomarker Response
33
Lung Function Response
34
6.6-Hr O3 Chamber Study

• Goal Alternative O3 standard format
• Study Assay airway inflammation, lung function,
  symptoms in 30 subjects at variable hourly O3
  levels ventilation for 0.08 ppm average 6-hour
  exposures
• Anticipated Findings Weighted form of standard
  provides better metric of acute (1-hr) and
  prolonged (8-hr) responses

35
Ambient O3 Monitor Bias

• Goal Assess design value day bias
• Study Compare collocated smog chamber
  reference network monitors
• Findings Monitors susceptible to 20-40 ppb
  positive bias from Hg organics (naphthalene,
  phenols, nitro-aromatics) during hot, stagnant,
  polluted conditions

36
Smog Chamber Response
37
O3 "Delta" (UV-CL) by Collocated FRM/FEM
250
200
150
ppb
3
O
100
50
0
1
13
25
37
49
61
73
85
97
109
121
133
145
157
Hour
38
O3 Rollback Model

• Goal Improve risk-benefit assessment
• Study Develop 3-parameter algorithm projecting
  annual hourly O3 time-series at attainment of
  alternative standards
• Anticipated Finding Currently used 2-parameter
  algorithms overestimate the risks benefits at
  standard compliance

39
PM Health Impacts
40
Fine Particles from Showers

• Goal Assess aqueous FP emissions
• Study Assay shower FP (0.3-10 um) mass as
  function of spray flow, splash, temperature, and
  total dissolved solids
• Findings Mg/m3 levels of PM10 (300 ug/m3 PM2.5)
  accumulate in ventilated (6/hr) bathrooms during
  showering

41
Shower PM Production
42
API Contacts

• Stationary Source Emissions
• Karin Ritter (ritterk_at_api.org)
• NAAQS Health Effects
• Will Ollison (ollisonw_at_api.org)