Energy Efficiency and Emissions of a Biomass Pellet Fired Hydronic Heater
Using Multiple Fuels
Final Report
Prepared for the
New York State Energy Research and Development Authority
Albany, NY
nyserda.ny.gov
Ellen Burkhard, Ph.D.
Senior Project Manager
Prepared by:
U. S. Environmental Protection Agency, Office of Research and Development
Research Triangle Park, NC
John Kinsey, Michael Hays, Ingrid George, Amara Holder, Dennis Tabor, Edgar Thompson, Tiffany Yelverton,
Peter Kariher, and William Linak
AND
Jacobs Technology, Inc.,
Research Triangle Park, NC
Carl Singer
Oak Ridge Institute for Science Education
Assigned to the U. S. Environmental Protection Agency, Research Triangle Park, NC
Angelina Brashear
EPA Report No.: EPA/600/R-20/179
This document has been reviewed by the U.S. Environmental Protection Agency, Office of Research and
Development, and approved for publication.
l
-------
NOTICE
This report was prepared in the course of performing work sponsored by the New York State
Energy Research and Development Authority and the U.S. Environmental Protection Agency's
Office of Research and Development. The opinions expressed in this report do not necessarily
reflect those of NYSERDA or the State of New York, and reference to any specific product,
service, process, or method does not constitute an implied or expressed recommendation or
endorsement of it. Further, NYSERDA and the State of New York make no warranties or
representations, expressed or implied, as to the fitness for particular purpose or merchantability
of any product, apparatus, or service, or the usefulness, completeness, or accuracy of any
processes, methods, or other information contained, described, disclosed, or referred to in
this report. NYSERDA and the State of New York make no representation that the use of any
product, apparatus, process, method, or other information will not infringe privately owned rights
and will assume no liability for any loss, injury, or damage resulting from, or occurring in
connection with, the use of information contained, described, disclosed, or referred to in this
report.
This document has been reviewed by the U.S. Environmental Protection Agency, Office of
Research and Development, and approved for publication.
11
-------
ABSTRACT
This report characterizes the thermal performance and air pollutant emissions from a pellet-fired
hydronic heater (PBHH) burning both premium hardwood and switchgrass fuels. The PBHH was
operated under three load conditions: 25% (minimum); 100%; and during a simulated load
profile indicative of a typical 232 m2 house during the first two weeks in January in Syracuse,
NY (i.e., Syracuse cycle). Duplicate test runs were conducted at each load condition with
measurements made throughout each 6-hour run period. Measurements were made to determine
thermal efficiency and quantify a variety of air pollutants including criteria and related gases,
volatile organic compounds and carbonyls, gas-phase polycyclic aromatic hydrocarbons,
polycyclic dibenzodioxins and furans, gaseous hydrochloric acid, total particulate matter, particle
number and size, elemental/organic carbon, optical black carbon, particle morphology, particle
elemental composition, and particle-phase semivolatile organics. Samples of both the fuel and
bottom ash were also collected and analyzed for a variety of important parameters. The data from
the study showed that hardwood provided the highest thermal efficiency of the two fuels tested.
In addition, the emissions of most air pollutants were highest for hardwood combustion at 25%
load whereas switchgrass combustion produced higher emissions for the Syracuse cycle and at
100%) load. Comparison of these data to historical information for other appliances burning
similar fuels indicated at least generally comparable results.
Key Words
Pellet-fired hydronic heaters, hardwood pellets, switchgrass pellets, thermal efficiency, air
pollutant emissions, particulate matter composition, polycyclic aromatic hydrocarbons, volatile
organic compounds
in
-------
ACKNOWLEDGEMENTS
This research was funded by the New York State Energy Research and Development Authority
(NYSERDA) through Cooperative Agreement No. 32984 with additional support provided by
the U.S. Environmental Protection Agency (EPA), Office of Research and Development, through
Cooperative Research and Development Agreement 795-14. Jacobs Technology, Inc. was funded
by EPA through Contract No. EP-C-15-008. Ms. Brashear was supported by a grant from EPA
through the Oak Ridge Institute for Science Education.
The authors acknowledge the testing and analytical assistance of Bill Preston, John Nash, and
Daniel Janek of Jacobs Technology. We also appreciated the expert advice from Dr. Jerry
Cherney of Cornell University who also provided the PBHH unit and switchgrass pellets used
during testing. Also providing advice and guidance to the program were Ms. Amanda Aldridge
and Mr. Mike Toney of EPA's Office of Air Quality Planning and Standards.
iv
-------
TABLE OF CONTENTS
TABLES ix
FIGURES x
ACRONYMS AND ABBREVIATIONS xiv
EXECUTIVE SUMMARY S-l
1 Project Background and Objectives 1
2 Experimental Approach 2
2.1 Facility Description 2
2.1.1 Pellet Burning Hydronic Heater (PBHH) 3
2.1.2 Heat Load Profiles 4
2.1.3 Heat Load Demand Control 5
2.2 Test Protocol 6
3 Sampling Procedures 9
3.1 Overall Approach 9
3.2 Sampling Techniques 9
3.2.1 Flue Gas Volumetric Flow Rate (EPA Methods 1A and 2C) 9
3.2.2 Gaseous Continuous Emissions Monitoring 9
3.2.3 Hydrochloric Acid (HCl)/Semivolatile and Nonvolatile Sampling Train.... 10
3.2.4 VOC and Carbonyl Sampling Train 10
3.2.5 Total PM Mass Measurements 11
3.2.6 PM Number and Size Measurements 12
3.2.7 OCEC and PM Elemental Composition 12
3.2.8 Black Carbon Measurements (Aethalometer and PAX) 13
3.2.9 PM Semivolatile Organic Compounds (SVOCs) 14
3.2.10 Bottom Ash Evaluation 14
3.2.11 Fuel Sampling 14
3.3 Sample Recovery and Preservation 14
3.4 Sample Collection and Frequency 14
4 Measurement Methods and Procedures 16
4.1 Heat Load Demand Measurements and Efficiency Determination 16
4.2 Equipment Calibration 17
4.2.1 CEM Calibration Procedures 17
4.2.2 Sampling Equipment Calibration 17
v
-------
4.2.3 On-Line PM Instrumentation 18
4.3 Continuous Monitoring of Gaseous Pollutants 18
4.3.1 CEM Bench 18
4.3.2 Closed-Cell FTIR 19
4.4 Volatile Organic Compound and Carbonyl Analyses 20
4.5 Gaseous PAH Analyses 21
4.6 Gaseous PCDD/PCDF Analyses 21
4.7 Gaseous Halide (HC1) Analyses 22
4.8 Total PM Mass Analyses 22
4.9 Particle Number and Size Determination 22
4.10 OCEC and Black Carbon Analyses 23
4.10.1 Laboratory OCEC Analyses 23
4.10.2 Semi-Continuous OCEC Analyzer 23
4.10.3 Optical Black Carbon 23
4.11 PM Elemental Analyses 24
4.12 Particle Morphology 24
4.13 PM Semivolatile Organics Analyses 24
4.14 Fuel Analyses 25
4.15 Bottom Ash Analyses 25
5 Data Analysis 26
5.1 Thermal Efficiency 26
5.2 Emission Calculations 27
5.3 Dilution Factor 28
6 Quality Assurance and Quality Control 30
6.1 Overall Objectives 30
6.2 Data Quality Objectives (DQOs) 30
6.3 Data Quality Indicator Goals (DQIs) 31
6.3.1 Stack Testing and Thermal Measurement Parameters 31
6.3.2 Continuous Emission Monitoring 32
6.3.3 Total PM and Filter Mass 34
6.3.4 Total Halide Emissions 34
6.3.5 In-Stack PAHs and PCDD/Fs 34
6.3.6 VOCs and Carbonyls 35
6.3.7 Particle Number and Size 36
vi
-------
6.3.8 Optical Black Carbon 36
6.3.9 Laboratory and Semi-Continuous OCEC 37
6.3.10 PMSVOCs 37
6.3.11 Closed Cell FTIR 38
7 Experimental Results 39
7.1 Thermal Efficiency 40
7.2 Gas Phase Pollutants 41
7.2.1 Criteria and Related Gaseous Emissions 41
7.2.2 Volatile Organic and Carbonyl Compounds 48
7.2.3 Gaseous PAH Emissions 49
7.2.4 Dioxin and Furan Emissions 55
7.2.5 Total Halide Emissions 60
7.3 Particle Phase Pollutants 62
7.3.1 Total Particulate Matter Emissions 62
7.3.2 Particle Number Emissions 65
7.3.3 Particle Size Distributions 72
7.3.4 Elemental and Organic Carbon (ECOC) 72
7.3.5 Optical Black Carbon (OBC) 78
7.3.6 PM Elemental Composition 83
7.3.7 Particle Morphology 85
7.3.8 PM Semi-Volatile Organic Compounds 86
7.3.9 Fuel and Ash 93
8 Summary and Conclusions 97
8.1 Effect of Fuel Type 97
8.2 Comparison to Historical Emissions Data 98
8.3 Conclusions 101
9 References and Supporting Documentation 102
9.1 References 102
9.2 EPA Test Methods 103
9.3 ASTY1. NIOSH, and Other Methods 105
APPENDIX A A-l
Survey: Display and Menu A-3
Factory Adjustments A-4
APPENDIX B B-l
vii
-------
APPENDIX C C-l
APPENDIX D D-l
viii
-------
TABLES
Table E-l. Measurement Methods and Sampling Locations.
Table E-2. Test Average Thermal Efficiency and Emission Factors.
Table 2.1. Fuel Analysis Results.
Table 3-2. Measurement Methods and Sampling Locations.
Table 6.1. Daily Stack CO2 Quality Assurance Checks.
Table 6.2. Daily Dilution Duct CO2 Quality Assurance Checks.
Table 7-1. Thermal Efficiency Summary.
Table 7-2. Test Average Gaseous Emissions (Engineering Units).
Table 7-3. Test Average Gaseous Emissions (SI Units).
Table 7-4. Average Total VOC and Carbonyl Emissions.
Table 7-5. Test Average Total Gaseous PAHs.
Table 7-6. Test Average Total PCDD/PCDF Emission Factors.
Table 7-7. Test Average HC1 Emission Factors.
Table 7-8. Test Average Total Particulate Matter Emission Factors.
Table 7-9. Test Average Particle Number Emission Factors.
Table 7-10. Test Average ECOC Emission Factors21.
Table 7-11. Test Average Optical Black Carbon (BC) Emission Factors.
Table 7-12. Emission Factors for Air Toxic and Other Selected Elements'1.
Table 7-13. Average Total SVOC Emission Factors.
Table 7-14. Average OC Emission Factor Components.
Table 7-15. Data Population and Concentration Range by Compound Class.
Table 7-16. Results of Fuel Analyses21.
Table 7-17. Results of Bottom Ash Analyses21.
Table 8-1. Percent Difference Between Switchgrass and Hardwood.
ix
-------
FIGURES
Figure E-l. Hydronic heater testing facility.
Figure 2-1. Hydronic heater testing facility.
Figure 2-2. Reka HKRST/V-FSK 20 PBHH tested.
Figure 2-3. Simulated Syracuse, NY, area heat load profile for January.
Figure 3-1. Diagram of multi-filter sampler.
Figure 7-1. Average thermal efficiency by fuel and load condition.
Figure 7-2. Test average gaseous emission factors (engineering units) for nitrogen and sulfur
compounds (a) and (c) as well as organic gases (b) and (d).
Figure 7-3. Test average gaseous emission factors (SI units) for nitrogen and sulfur compounds
(a), (c), and (e) as well as organic gases (b), (d), and (f).
Figure 7-4. Total speciated VOC emission factors in terms of: (a) mass per mass of fuel burned;
(b) mass per heat input; and (c) mass per heat output. Bars represent range of values.
Figure 7-5. Speciated VOC emission factors in mass per heat input (mg/MJ) for the 16 most
abundant VOCs averaged over each test condition.
Figure 7-6. Test average total emission factors (engineering units) for gaseous PAH compounds
in terms of: (a) mass per heat input; and (b) mass per heat output.
Figure 7-7. Test average total emission factors (SI units) for gaseous PAH compounds in terms
of: (a) mass per mass of fuel burned; (b) mass per heat input; and (c) mass per heat output.
Figure 7-8. Test average emission factors for the 16 target PAH compounds determined using
EPA Method 23 for each fuel and load condition.
Figure 7-9. Composite total emission factors (engineering units) for PCDD/PCDF compounds in
terms of: (a) mass per heat input; and (b) mass per heat output.
Figure 7-10. Composite total emission factors (SI units) for PCDD/PCDF compounds in terms
of: (a) mass per mass of fuel burned; (b) mass per heat input; and (c) mass per heat output.
Figure 7-11. Composite emission factors for the 17 PCDD/PCDF compounds measured using
EPA Method 23_ for each fuel and load condition.
Figure 7-12. Test average total emission factors (SI units) for HC1 in terms of: (a) mass per mass
of fuel burned; (b) mass per heat input; and (c) mass per heat output.
Figure 7-13. Total PM mass emission factors (engineering units) in terms of: (a) mass per heat
input; and (b) mass per heat output.
x
-------
Figure 7-14. Total PM emission factors (SI units) in terms of: (a) mass per mass of fuel burned;
(b) mass per heat input; and (c) mass per heat output.
Figure 7-15. Total particle number concentration time histories for hardwood combustion at 25%
load (a and b), during the Syracuse cycle (c and d), and 100% load (e and f).
Figure 7-16. Total particle number concentration time histories for hardwood combustion at 25%
load (a and b), during the Syracuse cycle (c and d), and 100% load (e and f).
Figure 7-17. Background corrected test average particle number concentrations for both fuels.
Figure 7-18. Total particle number emission factors (engineering units) in terms of: (a) particles
per heat input; and (b) particles per heat output.
Figure 7-19. Total particle number emission factors (SI units) in terms of: (a) particles per mass
of fuel burned; (b) particles per heat input; and (c) particles per heat output. Also shown in (a) is
a similar factor for the European unit tested previously with hardwood during the Syracuse cycle.
Figure 7-20. Differential number particle size distributions and summary statistics for tests
burning hardwood pellets.
Figure 7-21. Differential number particle size distributions and summary statistics for tests
burning switchgrass pellets.
Figure 7-22. Elemental and organic carbon emission factors for all fuel and load conditions in
terms of: (a) mass per heat input; and (b) mass per heat output (engineering units).
Figure 7-23. Elemental and organic carbon emission factors for all fuel and load conditions in
terms of: (a) mass per mass fuel; (b) mass per heat input; and (c) mass per heat output (SI units).
Figure 7-24. Comparison of manual filter results to those from the semi-continuous carbon
analyzer for: (a) elemental carbon; and (b) organic carbon (all tests).
Figure 7-25. Optical black carbon emission factors for each fuel and load condition on the basis
of: (a) thermal input; and (b) thermal output (engineering units).
Figure 7-26. Optical black carbon emission factors for all fuel and load conditions in terms of:
(a) mass per mass fuel; (b) mass per heat input; and (c) mass per heat output (SI units).
Figure 7-27. Comparison of optical black carbon to elemental C for all tests conducted.
Figure 7-28. Average emissions factors for: (a) toxic metals; and (b) sulfur and chlorine as
derived from XRF analysis of the Teflon filter samples.
Figure 7-29. Soot aggregate with K inclusions emitted from wood pellet combustion.
Figure 7-30. (a) C, Fe, O particle; and (b) ash particle emitted from grass pellet combustion.
XI
-------
Figure 7-31. SVOC and unresolved OC emission factors for all fuel and load conditions in terms
of: (a) mass per mass fuel; (b) mass per heat input; and (c) mass per heat output (SI units).
Figure 7-32. Filter-based OC-EC ratios in PM for individual tests sorted by load conditions. SwG
= switchgrass.
Figure 7-33. Individual mean SVOC concentrations in PM emitted from boiler testing.
Concentrations are given in units of g/g OC. Error bars indicate the concentration range. The y-
axis is log scale. Symbols and colors are coded by compound class.
Figure 7-34. Quantile box plots of individual SVOC concentrations pooled by compound class.
Levoglucosan is the anhydrosugar.
Figure 7-35. Concentration sums (|ig/gOC) sorted by individual test, compound class, test load
conditions, and fuel type (H- hardwood pellet; SwG - switch grass pellet).
Figure 7-34. Bottom ash analyses in terms of: (a) loss-on-ignition; and (b) metal oxides. Data in
weight percent.
Figure 8-1. Comparison of REKA to Appliance D from the Chandrasekaran et al. study (2013a)
for all fuel and load conditions for: (a) total PM; (b) OC; and (c) CO.
Xll
-------
ACRONYMS AND ABBREVIATIONS
A A atomic absorption
APCS air pollution control system
ASTM American Society for Testing and Materials (now ASTM International)
BC black carbon
BTU British thermal units
CEM continuous emission monitor
CH4 methane
CO carbon monoxide
CO2 carbon dioxide
CRADA cooperative research and development agreement
DAS data acquisition system
DNPH dinitrophenylhydrazine
DQI data quality indicator
DQO data quality objective
dscm dry standard cubic meter(s)
EC elemental carbon
EF emission factor
ELPI electrical low-pressure impactor
EPA U.S. Environmental Protection Agency
FID flame ionization detector
GC gas chromatography
HC1 hydrochloric acid
HEPA high-efficiency particulate air
HHV high heating value
HPLC high performance liquid chromatography
HRGC high resolution gas chromatography
HRMS high resolution mass spectrometry
LOI loss on ignition
LRMS low resolution mass spectrometry
MJ megajoule
MOP miscellaneous operating procedure
MS mass spectrometry
NDIR non-dispersive infrared
NIOSH National Institute for Occupational Safety and Health
NIST National Institute of Standards and Technology
NO nitrogen monoxide
NO2 nitrogen dioxide
NOx nitrogen oxides
NRMRL National Risk Management Research Laboratory
xiv
-------
NYSERDA New York State Energy Research and Development Authority
oc
organic carbon
PAH
polycyclic aromatic hydrocarbon
PAX
photoacoustic extinctiometer
PBHH
pellet-burning hydronic heater
PCDD/F
polychlorinated dibenzo-/>dioxins and polychlorinated dibenzofurans
PFI
Pellet Fuels Institute
PM
particulate matter
PM2.5
particulate matter <2.5 microns
PTFE
polytetrafluoroethylene
QA
quality assurance
QAPP
quality assurance project plan
QC
quality control
RPD
relative percent difference
RTP
Research Triangle Park
SIM
selected ion monitoring
S02
sulfur dioxide
SOP
standard operating procedure
SVOC
semivolatile organic compound
TEM
transmission electron microscope
THC
total hydrocarbon
TOT
thermal-optical transmittance
TPS
thermophoretic sampler
VOC
volatile organic compound
XRF
X-ray fluorescence
xv
-------
xvi
-------
EXECUTIVE SUMMARY
Various types of biomass are readily available for New York State residential heating. Wood-pellet
hydronic heaters were found to be cleaner burning (less air pollutant emissions) and more efficient
than conventional outdoor wood-fired hydronic heaters designed to cycle from full load to idle or
gasification units (primarily of European design) (Kinsey et al., 2012). However, the unknown
performance of this type of technology using non-woody biomass pellets and the uncertainty over
emissions of particles and other products of incomplete combustion leave numerous questions
regarding the environmental, health, and energy implications for this type of alternative home
heating.
The goal of this research was to fully characterize the thermal efficiency and emissions of
particulate matter (PM) and other pollutants from a hydronic heater using non-woody pellets to
provide NYSERDA with a preliminary environmental and energy assessment of the environmental
impact of hydronic heaters using non-woody fuel pellets. Testing was conducted on a REKA
HKRST/V-FSK 20 pellet-burning hydronic heater (PBHH) while burning premium hardwood and
switchgrass pellets. The switchgrass pellets were supplied by Dr. Jerry Cherney of Cornell
University who also provided technical support.
The PBHH characterization determined the thermal performance and emissions for a broad array
of pollutants under 3 load conditions burning hardwood pellets and switchgrass pellets. Load
conditions were: 25% of nameplate rating; 100% of nameplate rating; and Syracuse load
conditions. Technical specifications for the PBHH identify a minimum load of 21 MJ/h (5.8 kW),
nominally 30% of nameplate rating. While 15% of nameplate rating was desired, 25% was the
lowest load we could reliably control without overheating of the PBHH. Controlled hot water
temperatures were lowered significantly to 66 °C to achieve this heat load. Hardwood pellets were
Pellet Fuel Institute (PFI) certified Premium Grade Hardwood Pellets (Fiber Energy Products,
LLC; Mountain View, AR) packed in 40-pound plastic bags. Switchgrass pellets were Biomass
Pellet Fuel (SwitchGreen, Kingston, Ontario CANADA) packed in 40-pound plastic bags. Fuel
samples collected during operations were analyzed for proximate, ultimate, and ash minerals by
Standard Laboratories (Cresson, PA).
The basic test protocol was based on ASTM Method E2515-11 (Figure E-l). However, several
modifications were necessary to accommodate the test program objectives and to accommodate
facility and funding limitations. One of the most significant modifications was the change in all
test periods to 6 hours to accommodate a reasonable simulation of the Syracuse load cycle. Further,
the heat capacity of the PBHH and recirculating water required over an hour to reach control
temperature from a cold start and longer to achieve conditions resembling a steady state
combustion bed. The PBHH was lit at the beginning of each operational week and allowed to run
overnight prior to testing in an attempt to achieve a representative fuel bed. Overnight operations
were performed using an auxiliary water/air heat exchanger to provide load since facility
requirements precluded over-night operation with cooling water. Cooling water was set to achieve
initial heat load for each test at least an hour before starting the test.
S-l
-------
To Air
Pollution
Control
System
Outside
Internal Sampling Platform
Canopy
PM
Mass
Metals
and
NIOSH
OCEC
ASTM
2515
Dilution Duct
(25.4 cm OD)
Modified
Method 23 &
n2 Diluent
CO + C02
C02 jo-
THC 11A
NO
S°2 TO-15
Semi-Continuous OCEC
Black Carbon, CO2, and TEM
Secondary
Diluters
Thermocouple
I
Building
Chilled
Water
so ation
©> —[>
-------
efficiency and emissions calculations. The hopper, screws, and firebox were emptied completely
of fuel and ash prior to changing fuel.
Emission sampling was conducted on a suite of air pollutants consisting of both continuous
measurements and time-integrated sampling to determine key appliance operating parameters as
well as gas- and particle-phase emissions. The continuous measurements included temperature and
flow throughout the system, as well as additional temperature measurements conducted in
locations such as the combustion zone, stack, and dilution tunnel. The measurement methods and
sampling locations used in the program are provided in Table E-l.
Table E-l. Measurement Methods and Sampling Locations
Pollutant
Measurement Technique
Test Method or
Instrument"
Time Scale
Sampling
Location
Total PM emissions
Filter gravimetric
Modified ASTM Method E2515-11
Time-integrated
Dilution duct
Particle size distribution
(including PM2.5)
On-line cascade impactor
Dekati Electrical Low Pressure
Impactor (ELPI)
Continuous
Dilution duct
Particle morphology15
Transmission electron microscopy
(TEM)
Thermophoretic sampler (TPS) +
analysis of TEM grids
Time-integrated
Secondary
dilution manifold
Organic C/Elemental C
(OC/EC)
Thermal-optical transmission
(TOT)
NIOSH Method 5040; pre-fired
quartz filters in multi-sampler
Time-integrated
Dilution duct
Sunset Model 4 OC/EC analyzer
(optional)
Semi-continuous
Secondary
dilution manifold
Optical black carbon
Optical absorption
Magee AE-22 Aethalometer and
PAX Extinctiometer (optional)
Continuous
Secondary
dilution manifold
Total gas- and particle-
phase polycyclic
aromatic hydrocarbons
Impingers by gas
chromatography/low resolution
mass spectroscopy (GC/LRMS)
and XAD resin + filter by
GC/LRMS
Modified EPA Method 5, 0010, 23,
and 26A (filter + XAD resin +
impingers)
Time-integrated
(1 sample/test)
Stack
Gaseous
polychlorinated
dibenzodioxins and
furans
XAD resin/filter by high-resolution
gas chromatography/mass
spectrometry (GC/HRMS)
Hydrochloric acid
Impingers by ion chromatography
Volatile organic
compounds (VOCs) and
carbonyls
SUMMA canisters by GC/MS
(VOCs) and DNPH sorbent
cartridges by high pressure liquid
chromatography (carbonyls)
EPA Methods TO-15 and TO-11A
Time-integrated (2
samples/ test)
Dilution duct
Filter-based
semivolatile organic
compounds
GC/MS
Thermal extraction of quartz
OC/EC filters
Time-integrated
Dilution duct
Particle elemental
composition
X-ray fluorescence (XRF)
Analysis of Teflon filters in multi-
sampler by EPA Method IO-3.3
Time-integrated
Dilution duct
S-3
-------
Pollutant
Measurement Technique
Test Method or
Instrument"
Time Scale
Sampling
Location
Bottom ash
Loss on ignition (LOI)
Muffle furnace or thermal
gravimetric analysis
1 sample/test
Grab samples
Elemental composition
Atomic absorption (AA)
S02
Fourier transform infrared
spectroscopy (FTIR)
Closed-cell IMACC spectrometer
Continuous
Dilution duct
CO
Non-dispersive infrared (NDIR)
analysis
EPA Method 10B
Continuous
Stack
C02
NDIR
EPA Method 3A
Continuous
Stack and
dilution duct
o2
Paramagnetic analysis
EPA Method 3A
Continuous
Dilution duct
NOx
Chemiluminescence analysis
EPA Method 7E
Continuous
Dilution duct
Total hydrocarbons
(THC)
Heated flame ionization detector
(FID)
EPA Method 3C
Continuous
Dilution duct
ch4
FTIR
Closed-cell IMACC spectrometer
Continuous
Dilution duct
a See main text for acronyms of analytical methods; additional compounds measured with FTIR are not given
in this Table, but discussed later in the report
Due to a leak discovered in the continuous emission monitoring system, either the closed-cell
Fourier Transform Infrared (FTIR) instrument (CO, CH4, NH3, NOx, and SO2) or analysis of the
SUMMA canisters (N2O) were used for data analysis. The NDIR, paramagnetic, FID, and
chemiluminescence instruments did not produce useful measurements, and no valid data were
collected for CO2 or THC.
For each test run, emission factors for the target pollutants were calculated in terms of mass of fuel
burned, energy input, and energy output according to the calculation scheme shown in Section 5.
For both gases and particles, the test average results for the two runs conducted at each fuel and
load condition are provided in Section 7 which, except for stack PAHs, PCDD/PCDFs, and HC1,
are background-corrected unless otherwise indicated. Since only two tests were conducted at each
condition, the standard error (deviation) could not be calculated as is usually done. Instead, the
summary data tables show the relative percent difference (RPD) for the two tests. RPD is defined
as the difference between emission factor values from duplicate tests divided by the average of the
duplicates multiplied by 100 and is an indicator of the variability observed between the two test
runs. The data are also provided in graphical form generally in both engineering and SI units. Note
that in the graphs, the bars indicate the range of values for each parameter and not the RPD or
standard error. The bars are provided to generally indicate the amount of variability observed
between the two test average values which was oftentimes considerable. A summary of the test
results is shown in Table E-2.
S-4
-------
Table E-2. Test Average Thermal Efficiency and Emission Factors
25% Load
Syracuse Cycle
100% Load
Pollutant
Units
Wood
Grass
Wood
Grass
Wood
Grass
Thermal efficiency
percent
79
63
94
72
89
81
CO
g/kg fuel
36.9
32.8
18.5
19.1
11.4
2.25
ch4
g/kg fuel
1.96
0.9
0.305
0.839
0.0751
0.138
nh3
mg/kg fuel
4.9
226
10.5
59.4
25.6
8.94
n2o
mg/kg fuel
76.9
209
27.9
168
20.4
42.8
NOx
mg/kg fuel
9.90
186
205
102
ND
10.2
so2
mg/kg fuel
117
ND
25.3
ND
390
438
Total VOCs and
mg/kg fuel
2490
1780
205
1320
21.4
88.7
carbonyls
Gaseous PAHs
mg/kg fuel
114
11.7
3.43
71.7
2.66
22.9
Total PCDD/Fs
ng TEQ/kg
fuel
0.158
0.223
0.0929
0.455
0.320
0.105
Total halides (HC1)
mg/kg fuel
1.53
2.70
ND
2.53
4.43
13.5
Total PM
g/kg fuel
2.91
1.30
0.269
0.761
0.401
0.662
Particle number
particles/kg
fuel
2.07E+14
2.30E+14
7.65E+13
1.48E+14
6.78E+13
7.18E+13
Elemental carbon
mg/kg fuel
20.0
11.1
10.2
83.8
90.0
292
Organic carbon
mg/kg fuel
1075
572
33.1
392
1.78
62.8
Optical black carbon
mg/kg fuel
145
62.3
29.1
162
180
420
Elemental Cr1,
g/kg fuel
1.85E-03
6.05E-04
3.94E-05
4.08E-04
1.53E-05
2.00E-05
Elemental Mnb
g/kg fuel
2.87E-03
1.45E-04
2.19E-04
5.34E-05
9.51E-05
2.89E-05
Elemental Pbb
g/kg fuel
6.25E-03
1.38E-04
2.55E-04
6.26E-04
2.03E-04
3.68E-04
Total SVOCs0
mg/kg fuel
439
124
3.70
79.4
0.640
9.16
Three significant figures. Numbers shown in red face type are a single test value. All others are an average of
two test runs.
b Note that the emissions of Cr, Mn, and Pb are components of the total PM emission factor shown above and
not additional emissions.
c Note that the SVOCs are components of the total organic carbon emission factor shown above and not
additional emissions.
Based on the study results, the following conclusions were reached:
1. The combustion of hardwood exhibited the highest thermal efficiency at all load
conditions. Of the three loads tested with hardwood, the highest efficiency was for
operation during the Syracuse cycle which was the most indicative of "real world"
conditions.
2. With respect to reduced and oxidized carbonaceous gases, the same general trend was
observed as was the case for the particle-phase constituents namely, hardwood produced
higher emissions at 25% load and switchgrass had the highest emissions during the
Syracuse cycle and at 100% load.
S-5
-------
3. For reduced and oxidized nitrogen compounds, grass combustion generally had the highest
gaseous emissions as compared to hardwood.
4. For the speciated gas-phase VOCs/carbonyls and PAHs, the combustion of hardwood
generally produced the highest emissions at 25% load whereas during the Syracuse cycle
and 100% load, switchgrass exhibited the higher emissions.
5. For particle-phase air pollutants, the combustion of hardwood generally produced the
highest emissions at 25% load whereas during the Syracuse cycle and 100% load
switchgrass exhibited the higher emissions.
6. Comparing the current results to historical data for similar hydronic heaters, the data
reported here were comparable to the unit tested for NYSERDA in 2010 using hardwood
pellets during the Syracuse cycle and generally similar to Appliance D tested by
Chandrasekaran et al. (2013a) burning both fuel types depending on load.
7. The lack of U. S. technical support severely hampered operation and testing of the REKA
appliance evaluated in this study.
S-6
-------
1
Project Background and Objectives
Various types of biomass are readily available for New York State residential heating. Wood-pellet
hydronic heaters were found to be cleaner burning (less air pollutant emissions) and more efficient
than conventional outdoor wood-fired hydronic heaters designed to cycle from full load to idle or
gasification units (primarily of European design) (Kinsey et al., 2012). However, the unknown
performance of this type of technology using non-woody biomass pellets and the uncertainty over
emissions of particles and other products of incomplete combustion leave numerous questions
regarding the environmental, health, and energy implications for this type of alternative home
heating.
The goal of this project is to fully characterize the thermal efficiency and emissions of particulate
matter (PM) and other pollutants from a hydronic heater using non-woody pellets. This research
effort will provide NYSERDA with a preliminary environmental and energy assessment of the
environmental impact of hydronic heaters using non-woody fuel pellets.
This report provides the results of testing conducted on a REKA HKRST/V-FSK 20 pellet-burning
hydronic heater (PBHH) while burning premium hardwood and switchgrass pellets. The
switchgrass pellets were supplied by Dr. Jerry Cherney of Cornell University who also provided
technical support. The project was designed to supplement previous research characterizing the
emissions and energy efficiency of four outdoor wood-burning hydronic heaters as described by
Kinsey et al. (2012). The project was predominantly funded by the New York State Energy
Research and Development Authority (NYSERDA; Agreement 32984) through a Cooperative
Research and Development Agreement (CRADA) 79514 with the U. S. Environmental Protection
Agency, National Risk Management Research Laboratory, Air and Energy Management Division
(EPA) located at Research Triangle Park, NC.
This document covers in detail the background, objectives, technical approach, quality
assurance/quality control (QA/QC), experimental results, and references associated with the
research. EPA was assisted in this effort by its in-house contractor, Jacobs Technology, through
EPA Contract EP-C-15-008. Mr. Carl Singer acted as the contractor lead for the study.
1
-------
2 Experimental Approach
2.1 Facility Description
Figure 2-1 depicts the components of the hydronic heater test facility located at EPA's research
campus in Research Triangle Park (RTP), NC. The unit was located outside the EPA High Bay
facility, allowing testing under ambient conditions. The diluted flue gas was ducted into the facility
for ease of sampling and connection to the air pollution control system (APCS). The facility duct
work configuration and flows were generally based on ASTM International (ASTM) Method
E2515-11, Standard Test Method for Determination of Particulate Matter Emissions Collected by
a Dilution Tunnel (see the References section for a listing of the ASTM and other methods
associated with this work). The stack from the PBHH is 0.2 m in diameter and is 1.35 m in length
insulated with 3.8 cm high temperature fiberglass insulation. Because the stack was shorter than
the 6-m installation requirements, an induced draft fan (Tjernlund Model AD-1; White Bear Lake,
MN) was installed per manufacturer's recommendation in the stack downstream of the sampling
point. A conical hood cone was placed above the outlet of the stack to entrain the unit exhaust and
ambient dilution air. The cone was connected to a 0.25 m (outside diameter) stainless steel duct
surrounded by an outdoor sampling platform outfitted with vertically oriented 7.6 cm ports to
support particulate sampling, continuous emission monitors, and velocity measurements.
Internal Sampling Platform
PM
Mass
Metals
and
NIOSH
OCEC
Outside
Canopy
Di ution Duct
(25.4 cm OD)
83.8 cm
Modified
Method 23 &
N2 Diluent
CO + co2
2 jo
THC 11A
NO
S°2 TO-15
Secondary
Diluters
Isolation
Va ves
Recirculation
Pump
Water-to-Air
Heat Exchanger
To Air
Pollution
Control
System
ASTM
2515
Semi-Continuous OCEC,
Black Carbon, CO2, and TEM
Thermocouples
Water-to- T (f^> [J]
(1) Rotameter Needle
Biilding
¦ Chilled
Water
Building
Return
20 kW REKA Pellet Heater
Figure 2-1. Hydronic heater testing facility
The outside duct was connected to a horizontal indoor sampling platform set up to accommodate
sensitive measurement equipment and other stack measurement methods. The diluted and cooled
2
-------
exhaust gases were transferred into the building through a 0.25-m (outside) diameter and
approximately 12 m long stainless steel duct with multiple sampling ports. The temperatures
within the indoor dilution sampling duct were slightly below ambient. This sampling section was
connected to the APCS manifold for treatment prior to release to the atmosphere. The air duct
system moves 19.8-20.4 dry standard cubic meters (dscm) per minute of air, which correlates to
an approximate dilution ratio of 15 to 1 from the PBHH stack at full firing rate. Flows and pressures
were controlled by adjusting the facility APCS induction draft fan. All instruments/samplers were
connected to their respective sampling port using 0.95- or 0.63-cm (outside diameter) stainless
steel probes and sample lines.
A secondary dilution manifold was installed for continuous measurements of black carbon and
particle morphology at the location indicated in Figure 2-1. A model DI-1000 stainless steel
eductor operated with nitrogen (Dekati, Kangasala, Finland) was used with a stainless steel porous
tube diluter (Lyyranen et al. 2004) to sample and dilute exhaust from the dilution duct. The dilution
ratio was controlled by changing the flow of nitrogen to the eductor and the probe to optimize
particle concentrations for the black carbon instrumentation. The dilution ratio was determined by
measuring the CO2 concentrations in the secondary dilution system with a CO2 analyzer (Model
820, Li-Cor Biosciences, Lincoln, NE) and comparing it to the CO2 concentrations in the dilution
duct. This is discussed in detail in Section 5.
2.1.1 Pellet Burning Hydronic Heater (PBHH)
The PBHH tested during this test program was a Reka HKRST/V-FSK 20, shown in Figure 2-2.
The complete unit consists of a flat-bottom fuel bin, a hydronic heater, and an ash bin. Fuel was
conveyed from the fuel bin with a feed screw and dropped onto a reciprocating push-grate in a
stair-step fashion where the fuel was burned by adding under-fire and overfire air. Ash and cinders
are conveyed to the ash bin by a feed screw at the discharge of the moving push-grate. The heat is
transferred within the hydronic heater by a two-pass steel-plate heat exchanger with 50 mm fire
tubes. The fire tubes were kept clear by an automated compressed air soot blower which back-
flushed the tubes creating high transient PM emissions during each event. Recirculating water
from the hydronic heater is passed to a water loop passing through a water-to-water heat exchanger
simulating the space (residential house) to be heated. Fuel feed was controlled by a computer
operating system using temperature and oxygen sensors in the stack which turns the screw feeder
on and off to meet the heat load demand. The unit was not equipped with a secondary combustion
zone for emissions control.
No U. S. technical support for the REKA PBHH was available for installation and tuning of the
hydronic heater which resulted in severe delays in execution of the project. System operators made
their best attempt to operate the system in as appropriate manner as possible for both fuels tested.
During commissioning, the PBHH was found to overheat when running in "idle" mode, i.e. with
no heat load, causing a water blowdown and shutdown of the PBHH. This PBHH controlled water
temperature by adjusting forced air for combustion. Under idle conditions at this test facility, even
without operating the induced draft fan, combustion continued under natural draft conditions. Fuel
feed was controlled by an independent system based on stack oxygen concentration and continued
3
-------
to provide fuel during idle conditions until overheating resulted in shutdown. An auxiliary water-
to-air heat exchanger (Modine Model MSB; West Kingston, RI) was added to the hot water
distribution system before testing to remove heat during periods when the unit was functioning in
an "idle" condition. The PBHH control scheme would seem to require some mitigation for in-
home operation under low load conditions however no guidance was provided in manufacturer's
installation instructions.
Moving gate
Figure 2-2. Reka HKRST/V-FSK 20 PBHH tested
2.1.2 Heat Load Profiles
Three heat load profiles were evaluated in the study. Two of these were steady state operation at
100% load, 72 Mj/h (68,300 Btu/h) and minimum attainable load which was 25% load or 18 MJ/h
heat output and the third was the Syracuse heat load cycle. The heat load profile (Figure 2-3)
referred to as the Syracuse heat load cycle, is derived from a simulation program for heat demand
for a 232 m2 home in Syracuse, NY, developed by Brookhaven National Laboratory. The program
uses an averaged hour-per-hour heat load for the first two weeks of January averaged over 25 years
(Kinsey et al, 2012). The average daily heat load for the first two weeks in January is
approximately 827 megajoules (MJ) (784,000 British Thermal Units [Btu]) with a maximum heat
load of approximately 36 MJ/h (34,000 Btu/h) (1 Btu = 0.00105587 MJ). The Syracuse cycle has
been approximated by a polynomial over a 24-hour period:
4
-------
O (BTU/h) =
(3-1)
-0.0143 x6 + 0.9261 x5 - 21.129 x4 + 204.05 x" - 852.18 x2 + 1674.7 x + 30947
Due to facility and financial limitations, all tests were limited to 6 hours of active testing. For this
testing, the Syracuse cycle was compressed from a 24-hour cycle to a 6-hour cycle. A target heat
load for any elapsed time in the cycle can be calculated. For this testing, heat load was adjusted
every 10 minutes to simulate the Syracuse cycle. Heat load targets were calculated for each 10-
minute interval through the test period.
-------
load removed from the hot water is calculated from the temperature difference in the chill water
across the heat exchanger multiplied by the mass flow of chill water using the following formula:
Q=FCPAT (3-2)
where F is the mass flow rate of cooling water through the heat exchanger, Cp is the heat capacity
of the cooling water, and AT is the difference between the inlet and outlet temperature of the
cooling water.
2.2 Test Protocol
The PBHH characterization determined the thermal performance and emissions for a broad array
of pollutants under 3 load conditions burning hardwood pellets or switchgrass pellets. Load
conditions were 25% of nameplate rating, 100% of nameplate rating, and Syracuse Load
conditions. Technical Specifications for the PBHH identify minimum load of 21 MJ/h (5.8 kW),
nominally 30% of nameplate rating. While 15% of nameplate rating was desired, 25% was the
lowest load we could reliably control without overheating of the PBHH; controlled hot water
temperatures were lowered significantly to 66 °C to achieve this heat load. Hardwood pellets were
Pellet Fuel Institute (PFI) certified Premium Grade Hardwood Pellets (Fiber Energy Products,
LLC; Mountain View, AR) packed in 40-pound plastic bags. Switchgrass pellets were Biomass
Pellet Fuel (SwitchGreen, Kingston Ontario) packed in 40-pound plastic bags. Fuel samples
collected during operations were analyzed for Proximate, Ultimate, and Ash Mineral by Standard
Laboratories, Cresson, PA. The fuel analyses are summarized in Table 2.1
The test protocol was based on ASTM Method E2515-11 however several modifications were
necessary to accommodate the test program objectives and to accommodate facility and funding
limitations. One of the most significant changes was modification of all test periods to 6 hours to
accommodate a reasonable simulation of the Syracuse load cycle. Further, the heat capacity of the
PBHH and recirculating water required over an hour to reach control temperature from a cold start
and longer to achieve conditions resembling a steady state combustion bed. The PBHH was lit at
the beginning of each operational week and allowed to run overnight prior to testing in an attempt
to achieve a representative fuel bed. Overnight operations were performed using an auxiliary
water/air heat exchanger to provide load since facility requirements precluded over-night operation
with cooling water. Cooling water was set to achieve initial heat load for each test at least an hour
before starting the test.
6
-------
Table 2.1. Fuel Analysis Results
Parameter Measured
Hardwood Pellets
Switchgrass Pellets
Proximate Analysis (as received)
Moisture
4.66%
10.41%
Volatile Matter
80.34%
72.62%
Fixed Carbon
14.00%
12.71%
Ash
1.00%
4.26%
Ultimate Analysis (as received)
Sulfur
0.09%
0.12%
Carbon
47.91%
42.79%
Hydrogen
5.79%
5.44%
Nitrogen
0.31%
0.82%
Oxygen (by difference)
40.24%
36.16%
GCV (Btu/lb)
7736
7526
Ash Mineral Analysis
Silicon Dioxide
26.57%
62.99%
Aluminum Oxide
2.40%
7.55%
Ferric Oxide
4.89%
3.63%
Titanium Dioxide
0.27%
0.06%
Phosphorus Pentoxide
0.80%
1.00%
Calcium Oxide
36.63%
10.69%
Magnesium Oxide
2.58%
3.65%
Sodium Oxide
0.44%
1.41%
Potassium Oxide
11.76%
6.42%
Sulfur Trioxide
13.91%
2.32%
The PBHH was operated at constant conditions during each test, with chilled water flow the only
operating parameter adjusted to maintain desired load. Program conditions for each test are
summarized in Appendix A. The fuel feed, forced combustion air, reciprocating grate, etc.
operated during the testing according to program conditions. Induced draft fan at the stack and
induced draft fan on the dilution duct were operated at constant rates over the entire test. To achieve
low heat loads without overheating, the PBHH was set to 66 °C water temperature during 25%
load and Syracuse load tests while set to 75 °C during 100% load tests. During testing, the heat
transferred to the chilled water was adjusted by adjusting the flowrate of the chilled water. Flowrate
was adjusted every 10 minutes throughout the 6-hour test and maintained constant through the 10-
minute period despite changes in chilled water differential temperature or changes in inlet hot
water temperature. The temperature differential of the chilled water across the heat exchanger was
used to estimate the flow required to meet desired heat load. The heat load achieved during each
test is shown in Appendix B.
7
-------
A fill level was marked with tape around the interior of the fuel hopper. The PBHH was operational
overnight with test fuel in the hopper. Pellet fuel was added to the fill level immediately prior to
beginning each test. Fuel was added during each test in order to maintain fuel over the screw
conveyer and prevent air flow to or from the firebox. The lid was maintained in a closed and
latched position when not adding fuel. At the end of the test, fuel feed for the PBHH was paused
while fuel was added to the fill line. After re-initiating fuel feed, the firebox was opened and ash
was sampled from the end of the moving grate. Weight of fuel added was monitored throughout
the run for efficiency and emissions calculations. The hopper, screws, and firebox were emptied
completely of fuel and ash prior to changing fuel.
8
-------
3 Sampling Procedures
3.1 Overall Approach
Emission sampling was conducted on a suite of pollutants consisting of both continuous
measurements and time-integrated sampling to determine key appliance operating parameters as
well as gas- and particle-phase emissions. The continuous measurements included temperature and
flow throughout the system, as well as additional temperature measurements conducted in
locations such as the combustion zone, stack, and dilution tunnel.
3.2 Sampling Techniques
The conventional extractive sampling techniques are based on established EPA or ASTM methods,
or their modified versions, adapted to this PBHH source. In general, the data obtained from
conventional extractive methods were considered "reference" data and used to evaluate the data
obtained from the continuous measurement techniques.
3.2.1 Flue Gas Volumetric Flow Rate (EPA Methods 1A and 2C)
Dilution duct velocities were determined using a Shortridge airfoil due to the low velocity head at
this site and the small profile of the airfoil. The round 10-inch duct was traversed before and after
each test at sample points identified in EPA Method 1A, Sample and Velocity Traverses for
Stationary Sources with Small Stacks or Ducts: 6.7%, 25%, 75% and 93.3% of duct diameter
excluding nipple on orthogonal axis. The Shortridge output velocity at each point with an
arithmetic averaging of the data points. Velocity was measured at the centroid of the duct during
each test to assure steady state flow for particulate measurements. Average velocity was converted
to volumetric flow rate using the cross-sectional area of the duct.
3.2.2 Gaseous Continuous Emissions Monitoring
Continuous emissions monitoring (CEM) of gaseous pollutant concentrations were also performed
during each test. CO2 was monitored in the stack with CO2. SO2, nitrogen oxide (NOx), CO and
total hydrocarbons (THC) determined in the dilution duct using established EPA methods 3 A, 6C,
7E, 10, and 25 A. Due to resource constraints, calibration was performed prior to the test program
with all subsequent QA checks being performed through the bias line. However, during data
analysis a discrepancy was observed between CO2 concentrations measured in the dilution duct
and the mass of fuel burned. It was concluded that a bias, presumably a leak, had been introduced
into the dilution duct during repair and replacement operations that occurred immediately before
and during the first test. All dilution duct CEM data were considered unreliable for emissions
estimates. Therefore, either the closed-cell Fourier Transform Infrared (FTIR) instrument (CO,
CH4, NH3, NOx, and SO2) or analysis of the SUMMA canisters (N2O) were used. The NDIR,
paramagnetic, FID, and chemiluminescence instruments did not produce useful measurements,
and no valid data were available for stack CO2 or THC. The dilution duct CO2 remained the most
9
-------
reliable indicator of trends arising from PBHH cycles required to estimate dilution factors for
measurements made with secondary nitrogen dilution.
A closed-cell FTIR spectrometer was also used as outlined in Section 4.3.2. The use of the closed-
cell FTIR was a last-minute substitution due to the loss of other continuous analyzers for CH4 and
SO2 originally planned for use in the program. Due to the leak in the CEM bench, however, the
FTIR data were also used to determine CO, NH3, and NOx.
3.2.3 Hydrochloric Acid (HCI)/Semivolatile and Nonvolatile Sampling Train
The standard EPA Method 26A train, the EPA Method 23 train, and the EPA Method 0010
sampling train are versions of EPA Method 5 that were used for sampling. The Method 23 train
was modified to include impingers styled like the Method 26A train so that total halides (mainly
HC1) can also be collected. The modification consisted of using impingers that contain a 0.1 N
sodium hydroxide solution to absorb total halides. The gas was then measured with a calibrated
dry gas meter.
The train consists mainly of a heated probe, heated box containing a cyclone and filter, water-
cooled condenser, water-cooled XAD-2 cartridge, impinger train for water determination (which
has been modified for the HC1 collection), leak-free vacuum line, vacuum pump, and a dry gas and
orifice meter with flow control valves and vacuum gauge. Temperatures were measured and
recorded in the hot box (set at 125 °C), at the impinger train outlet, at the XAD-2 cartridge outlet
(maintained to be below ambient temperature), and at the inlet and outlet of the dry gas meter.
Leak checks were conducted at the beginning and end of each sampling run. Prior to sampling all
glassware, probes, glass wool, and aluminum foil were cleaned or purchased clean.
Hydrochloric acid from sodium hydroxide impingers were analyzed by high-performance liquid
chromatography (HPLC) ion-chromatograph (IC) using a conductivity detector (CD). The IC was
calibrated using commercial prepared standards.
3.2.4 VOC and Carbonyl Sampling Train
Volatile organics were sampled via Method TO-15, Determination of Volatile Organic
Compounds (VOCs) in Air Collected in Specially-Prepared Canisters and Analyzed by Gas
Chromatography/Mass Spectrometry (GC/MS). Sampling for VOCs were conducted using
laboratory-supplied 6-L SUMMA canisters connected to an Entech 1800 (Simi Valley, CA)
canister sampler and an in-line metal filter (frit). The canister sampler was equipped with a mass
flow controller assembly to allow for fill sampling times of 180 minutes. The canisters were
cleaned and evacuated before sampling using the Entech 3100 canister cleaner following the
standard operating procedure (SOP) associated with this work. Two 180-minute samples were
sequentially drawn for each 360-minute test.
Carbonyls were sampled via EPA Method TO-11A, Determination of Formaldehyde and Other
Aldehydes in Indoor Air Using a Solid Adsorbent Cartridge. This method utilizes commercially
10
-------
available cartridges (Supelco LpDNPH H30, PN 505323, Sigma-Aldrich, St. Louis, MO) followed
by high performance liquid chromatography (HPLC) analysis (Agilent 1100 LC, Agilent
Technologies, Santa Clara, CA). Samples were collected on 2,4-dinitrophenylhydrazine (DNPH)
cartridges by drawing air from the sampling line at a sampling rates in the range of 250-500
mL/min using a calibrated mass flow controller and pump. Carbonyl sampling was timed to
coincide with SUMMA canister sampling with a 180-minute duration per sample. Breakthrough
of carbonyls during sampling was determined during at least one test per condition by sampling
with two DNPH cartridges in series. After collection, samples were placed in the original pouch
and stored in a refrigerator before analysis. DNPH cartridges were analyzed within two weeks of
collection. The sampling was conducted according to the applicable miscellaneous operating
procedure (MOP).
3.2.5 Total PM Mass Measurements
For the determination of total PM mass emissions on a time-integrated basis, the PM emission
measurements followed the general procedures outlined in ASTM Method E2515-11. The
sampling system for this test method consists of duplicate dual-filter dry sampling trains sampling
from the dilution duct. Both particulate sampling trains were operated simultaneously at a sample
flow rate not exceeding 0.007 m3/min. The total particulate results obtained from the two sampling
trains were averaged to determine the particulate emissions and compared as a QC check on the
data. Each sampling train had two filters in series.
To achieve project objectives, several modifications to the sampling procedures in ASTM E2515-
11 were found necessary. The first modification consisted of replacing the glass filters with 47-
mm Emfab Pallflex TX40HI20WW filters (Pall Life Sciences, Ann Arbor, MI) which consist of
pure borosilicate glass microfibers reinforced with woven glass cloth and bonded with
polytetrafluoroethylene (PTFE); filters were changed during testing as required to maintain sample
flow. A second modification involved replacing the Teflon filter holders with two 47-mm stainless
steel filter holders (Pall Life Sciences, Ann Arbor, MI) which are grounded to prevent particles
losses. Finally, instead of using a straight probe oriented perpendicular to the flow stream, a
custom-made stainless-steel PM sampling probe oriented directly into the flow was used.
Considering these modifications, the mass of PM was determined using filter weights recovered
per EPA Method 5 instead of weighing the entire assembly. A front-half acetone rinse was
recovered and incorporated in total PM mass.
Samples for particle morphology were taken using a thermophoretic particle sampler (TPS; R.J.
Lee Group, Monroeville, PA). The TPS is designed to sample from ambient conditions, rather than
the duct. To obtain a representative sample from the dilution tunnel the TPS was contained within
a stainless-steel chamber (61 cm x 25 cm x 31 cm) connected to the secondary dilution system
(described in Section 2.1). A pump was used to draw a sample from the secondary dilution system
into the chamber. To prevent overloading of the sample the TPS was operated for only 10 minutes
of the test. Two samples were obtained, one for each fuel while the heater was operating on the
Syracuse cycle.
11
-------
3.2.6 PM Number and Size Measurements
A Dekati ELPI (ELPI software, Version 4.0) was used to provide real-time (10-sec) particle size
distributions (PSDs). The ELPI generates a PSD by first charging the particles with a unipolar
diode charger, which charges the particles based on geometrical diameter before they enter a
cascade impactor. The charged particles impact the stages on the impactor based on their inertia
(i.e., their aerodynamic diameter [equivalent unit density spheres]). A multi-channel electrometer
measures the charge of the particles as they land on each of the stages, giving current values for
each stage in f-amps. These current values are then converted to number of particles on each stage
and, if the density of the particle is known (or assumed), the mass of the particles on each of the
stages can also be found.
The differential number distribution, dN/dlog(Dp), is determined from the current distribution by
dividing these currents for each channel by conversion values for each channel. This conversion
vector was calculated by the manufacturer from the charger efficiency values for the stage's
midpoint diameter. The stage's midpoint diameter is the average of the cut point of the stage of
interest and the stage above. These midpoint diameters are determined for both the Stokes and
aerodynamic diameters. These particle numbers are then normalized by dividing by the logarithmic
width of the stage (either in terms of aerodynamic or Stokes diameter). Mass distributions,
dMZDlog(Dp), are then easily found by taking the number distribution and multiplying it by the
mass of each spherical particle assuming unit density.
3.2.1 OCEC and PM Elemental Composition
The National Institute for Occupational Safety and Health (NIOSH) Method 5040, Diesel
Particulate Matter (as Elemental Carbon), was used for OC/EC analyses. The first step in OC/EC
sampling involves filter preparation. The 47-mm quartz filters were pre-fired by placing them in
an oven at 900 °C overnight to remove any residual carbon present. Samples were collected
directly from the dilution duct onto the pre-fired quartz filters using an unheated polished stainless-
steel PM probe oriented into the flow, stainless-steel filter holder, calibrated mass flow controller,
and sampling pump. Three samples were drawn during a 6-hour test cycle. The exposed filters
were then sampled with a 1.5 cm2 punch in a radial fashion and analyzed by thermal-optical
transmittance (TOT).
Sampling was conducted using two multi-filter samplers (Figure 3-1). One multi-filter sampler
was equipped with four Pall 47-mm stainless-steel filter holders (Pall Life Sciences, Ann Arbor,
MI), each containing one pre-fired Pall Tissuquartz™ quartz filter acting as the primary filter for
OC/EC. The other multi-filter sampler was equipped with four dual-filter trains consisting of two
Pall 25-mm stainless-steel filter holders connected in series. The first filter holder contained a pre-
weighed Teflon filter (Pall Teflo™) with the backup filter holder containing a pre-fired Pall
Tissuquartz™ filter. The backup quartz filters were analyzed for OC/EC using NIOSH 5040. The
OC concentration on the backup quartz filters was subtracted from that found on the primary 47-
mm filters to compensate for gas-phase organic artifacts. The 25-mm Teflon filters were analyzed
gravimetrically to determine total PM mass and subsequently analyzed by an outside laboratory
(Chester LabNet, Tigard, OR) for elemental composition using EPA Method 10-3.3 (X-ray
12
-------
fluorescence [XRF]). Filter samples were collected at approximately the beginning, middle, and
end of each test period to represent the entire operating cycle using the same type particle probes
as ASTM 2515-11 described above. For the 47-mm quartz filters, three samples of one hour each
were collected. In the case of the three 25-mm filter sets, the sampling time was 20 minutes each.
Sample
from
Probes
Bypass Line
4-Way Sample
Splitter
Auto
Two-Way
Valves
D83-ED- csHEb
Vent Vent
25-mm Filter
Holders
Way Valve
Vacuum
Pump
Figure 3-1. Diagram of multi-filter sampler
3.2.8 Black Carbon Measurements (Aethalometer and PAX)
The AE-22 Aethalometer™ (Magee Scientific, Berkeley, CA) is an instrument that provided a
near real-time readout of the concentration of BC aerosol particles. The Aethalometer™ uses a
continuous filtration and optical absorption measurement method to give a continuous readout of
optical black carbon (OBC) real-time data. The Aethalometer used the SW:AF985d4 software
package.
A PAX photoacoustic extinctiometer (DMT, Longmont, CO) using the PAX.exe software package
was also used for the on-line measurement of optical black carbon. The primary quantity reported
by the instrument is the absorption coefficient Babs. The instrument also measures aerosol
scattering with an inverse nephelometer and a photomultiplier tube, which reports the scattering
coefficient Bscat. Both light absorption and scattering coefficients were measured with a 1-s
sampling rate.
The AE-22 and the PAX sampled emissions from the secondary dilution manifold. The diluted
emissions were split between the instruments using a custom-made stainless steel aerosol splitter.
13
-------
3.2.9 PM Semivolatile Organic Compounds (SVOCs)
After analysis for OC/EC, the 47-mm quartz filters were subsequently solvent extracted and
analyzed for SVOCs using the methodology described in Section 4 below. Like OC/EC, the
samples analyzed represented the three one-hour sampling periods conducted during each test run.
3.2.10 Bottom Ash Evaluation
Grab samples of the bottom ash were taken at the end of each test from the ash drop-out inside the
unit, cooled, and stored in sealed glass containers. Sample aliquots were sent to Standard
Laboratories, Inc. to be analyzed for elemental composition by atomic absorption (AA)
spectroscopy and mass loss upon heating in air (loss-on-ignition; LOI).
3.2.11 Fuel Sampling
A composite fuel sample was collected for each fuel type, hardwood pellets and switchgrass
pellets, for subsequent fuel analysis. Each fuel type was received as a single lot and considered
homogeneous. Each composite was comprised of 3 grab samples recovered from separate bags of
fuel. Samples were taken from newly opened bags prior to adding the remainder to achieve bin fill
level prior to testing. Grab samples were acquired on separate test days and stored in sealed freezer
bags prior to shipping for analysis.
3.3 Sample Recovery and Preservation
Following completion of a test run, each time-integrated sampling train was recovered in a clean
area, and the cleanup procedure started as soon as the probe was removed from the source location.
During transport between the test facility and the designated recovery (an adjacent laboratory),
both ends of the heated probe and openings of the impinger assembly were covered with aluminum
foil or sealed with ground glass caps. The organic rinses of the train were performed as specified
in EPA Method 23.
Samples were recovered on-site in the inorganics preparation laboratory located adjacent to the
test facility. Sample recovery procedures were followed as detailed in the appropriate methods.
Accordingly, samples shipped to outside laboratories were preserved as prescribed in the
method(s).
3.4 Sample Collection and Frequency
The sampling methods and sampling frequency for all the target pollutants are listed in Table 3-2.
Details of each method is described in detail in Section 4 below.
14
-------
Table 3-2. Measurement Methods and Sampling Locations
Pollutant
Measurement Technique
Test Method or Instrument
Time Scale
Sampling
Location
Total PM emissions
Filter gravimetric
Modified ASTM Method E2515-11
Time-integrated
Dilution duct
Particle size distribution
(including PM2.5)
On-line cascade impactor
Dekati ELPI
Continuous
Dilution duct
Particle morphology15
SEM/TEM
Thermophoretic sampler (TPS) +
analysis of TEM grids
Time-integrated
Secondary
dilution manifold
OC/EC
TOT
NIOSH Method 5040; pre-fired
quartz filters in multi-sampler
Time-integrated
Dilution duct
Sunset model 4 OC/EC analyzer
(optional)
Semi-continuous
Secondary
dilution manifold
BC
Optical absorption
Magee AE-22 Aethalometer and
PAX extinctiometer (optional)
Continuous
Secondary
dilution manifold
Total gas- and particle-
phase PAHs
Impingers by GC/LRMS and XAD
resin + filter by GC/LRMS
Modified EPA Method 5, 0010, 23,
and 26A (filter + XAD + impingers)
Time-integrated
(1 sample/test)
Stack
PCDD/Fs
XAD/filter by high-resolution
GC/MS
HC1
Impingers by ion chromatography
VOCs and carbonyls
SUMMA canisters by GC/MS and
DNPH cartridges by HPLC
EPA Methods TO-15 and TO-11A
Time-integrated (2
samples/ test)
Dilution duct
Filter-based SVOCs
GC/MS
Thermal extraction of quartz
OC/EC filters
Time-integrated
Dilution duct
Particle elemental
composition
XRF
Analysis of Teflon filters in multi-
sampler by EPA Method IO-3.3
Time-integrated
Dilution duct
Bottom ash
Loss on ignition (LOI)
Muffle furnace or thermal
gravimetric analysis
1 sample/test
Grab samples
Elemental composition
AA
so2
FTIR
IMACC
Continuous
Dilution duct
CO
Non-dispersive infrared (NDIR)
analysis
EPA Method 10B
Continuous
Stack
C02
NDIR
EPA Method 3A
Continuous
Stack and
dilution duct
o2
Paramagnetic analysis
EPA Method 3A
Continuous
Dilution duct
NOx
Chemiluminescence analysis
EPA Method 7E
Continuous
Dilution duct
THC
Heated flame ionization detector
(FID)
EPA Method 3C
Continuous
Dilution duct
15
-------
Pollutant
Measurement Technique
Test Method or Instrument
Time Scale
Sampling
Location
ch4
FTIR
IMACC
Continuous
Dilution duct
4 Measurement Methods and Procedures
Established methods were used to measure the thermal parameters and various pollutants of
interest described below. For standardized EPA and ASTM methods, no other written procedure
is generally required. However, for non-standard methods, a specialized protocol is available and
was used in the program.
4.1 Heat Load Demand Measurements and Efficiency Determination
Heat load represents the heat delivered through the heat exchanger to cooling water on this test
facility and is regulated by the flow of cooling water provided to the heat exchanger. Heat load
was determined as the product of the mass flow of cooling water, the heat capacity of the cooling
water, and the difference in temperature of the cooling water exiting and entering the heat
exchanger as presented above. The temperature of the water circulating to the PBHH does not
factor into the heat load calculation but is important in determining the amount of cooling water
required to produce a specified heat load.
Mass flow of cooling water was controlled by a manually controlled rotameter at the inlet to the
heat exchanger. Volumetric flow was manually recorded at 10-minute intervals during testing and
converted to mass flow using a density of 8.34 pounds per gallon. Mass flow was
corrected/calibrated using timed catches of water flowing through the heat exchanger at 11 L/min
(3.0 gallons per minute). Volumetric flow was adjusted based on real time estimates of heat load
with some consideration for the expected changes in PBHH water inlet temperature. The
temperature of the inlet water changes during the firing cycle resulting in changes in heat transfer
through the heat exchanger.
Temperatures across the heat exchanger were measured by Type K thermocouples and logged to
a Personal Daq/55 and PDQ2 expansion module data acquisition system (Measurement Computing
Corporation, Norton, MA). The analog outputs of the thermocouples were connected to a DAS for
monitoring and recording with a sampling frequency of 1 second.
Efficiency is calculated based on the total energy transferred to the cooling water during the tests
and the total energy input during the test: the gross calorific value of fuel fed to the PBHH. The
amount of fuel fed to the PBHH is determined by the mass of fuel added to the fuel hopper during
each test to maintain a constant volume in the hopper.
16
-------
The gross calorific value (aka higher heating value) of fuel fed was determined by fuel analysis on
an as received (wet) basis. Fuel samples were analyzed for Proximate (ASTM D2961, D3302,
D3173, D3175M, D3172, D3174), Ultimate (ASTM D4239 METHOD B, D3178, D5373, D3176),
and gross calorific value (ASTM D5865) by Standard Laboratories, Cresson PA. Carbon content,
a component of Ultimate analysis, was used for subsequent dilution calculations. In addition, the
ash mineral content of the fuel was analyzed by ASTM D2795 and ASTM D3682.
4.2 Equipment Calibration
System preventive maintenance was performed prior to the start of the test program. All major
components were checked to ensure operability and repaired or replaced if required. The EPA
Metrology Laboratory calibrated instruments such as meter boxes for sample volume prior to the
start of the sampling program. Laboratory equipment maintenance is conducted as recommended
by the manufacturer on an as-needed basis. Any leaks developed were repaired, parts lubricated as
recommended by the manufacturer, and manometers filled and checked for leaks. Replacement
parts, including fuses, pumps, spare tubing, compression fittings, etc., were maintained in the
laboratory to minimize downtime. Specific procedures are outlined below.
4.2.1 CEM Calibration Procedures
CEM calibration is performed using various standard gases (Airgas Specialty Gases, Durham, NC)
similar to the calibration procedures outlined in EPA Method 7E. A three-point calibration (zero,
mid, and span) was made prior to beginning the test program except for the dilution duct CO2
CEM; no mid-range calibration gas was available upon determining expected duct concentration.
Range considerations necessitated a relocation of dilution duct CO immediately prior to beginning
the test program resulting in loss of the three-point calibration for this instrument. A two-point
bias check was made daily before each test and a two-point drift check after each test, using the
same zero and span gases. All gas cylinders used for calibration are certified by the suppliers that
they are traceable to NIST standards within manufacturer-specified limits.
4.2.2 Sampling Equipment Calibration
EPA certified methods require that a laboratory record be maintained of all calibrations. The
requirements are based on the standard reference test method from which each respective method
was derived. The method specifies minimal calibration activities - standard pitot need not be
calibrated but should be inspected and cleaned, if necessary, prior to each certification test.
The volume metering systems were calibrated prior to the testing by the EPA Metrology
Laboratory using a wet-test meter, as permitted in the method. All thermocouples were calibrated
before and after the project. Thermometric fixed points (i.e., ice bath and boiling water) are
adequate standards for this task. These calibrations were also performed and documented by the
EPA Metrology Laboratory. The portion of the volume metering system from the pump to the
orifice meter must be leak checked following each test, using the procedure described in EPA
Method 5, section 5.6. Barometers must be calibrated semiannually by reference to a mercury
17
-------
barometer or a local National Weather Service station. Corrections should be made at a rate of -
0.1 in mercury (Hg) per 100 ft of elevation above sea level.
All instrument mass flow controllers and meter boxes were calibrated annually by the EPA
Metrology Laboratory and documented accordingly.
4.2.3 On-Line PM Instrumentation
Instruments were calibrated prior to the start of experiments according to the manufacturer's
instructions, unless noted. Specific information on the use and calibration of each instrument is
covered in instrument-specific manuals or Miscellaneous Operating Procedures.
4.3 Continuous Monitoring of Gaseous Pollutants
4.3.1 CEM Bench
Several primary gaseous flue-gas constituents were analyzed continuously using a CEM system
that includes monitors for CO, CO2, O2, NOx, and THC. The analog outputs of the analyzers were
connected to a Personal Daq/55 and PDQ2 expansion module data acquisition system
(Measurement Computing Corporation, Norton, MA). The analog outputs of the analyzers are
connected to a DAS for monitoring and recording with a sampling frequency of 1 second.
Sample gases are extracted for CEM analysis through a fixed stainless-steel probe at each location.
Sample from the stack was passed through a sample cooler to remove water and transported to the
CEM in Teflon tubing. Sample from the dilution duct was extracted with a heated head pump and
transported to a manifold through heated Teflon tubing. The dilution duct sample was then split
with moist sample transported to the total hydrocarbon analyzer through heated Teflon tubing
while the remaining sample was passed through a sample cooler to remove water and transported
to the CEMs in Teflon tubing.
Note: Analysis subsequent to data collection, indicated that these CEM measurements were
problematic, inconsistent with species mass balances, and in poor agreement with parallel FTIR
measurements. Further investigation revealed the presence of a systemic leak in the CEM
plumbing. As a result, the CEM data are not included. However, except for measurements of CO2
and CH4, parallel FTIR measurements of CO, O2, and NOx are substituted. Description of the CEM
measurements is included here for completeness.
The CO2 in the stack were determined using a California Analytical Instruments (Orange, CA)
Model ZRH NDIR analyzer. The CO2 in the dilution duct were measured using a LI-COR
(Lincoln, NE) Model 820. These analyzers operate by directing identical infrared beams through
an optical sample cell and a sealed optical reference cell. A detector located at the opposite end of
the cells continuously measures the difference in the amount of infrared energy absorbed within
each cell. This difference is a measure of the concentration of the component of interest in the
18
-------
sample. The infrared gas analyzer measures gas concentration based on the principle that each type
of gas component shows a unique absorption line spectrum in the infrared region.
CO was measured in the dilution duct using a California Analytical Instruments (Orange, CA)
Model 300 NDIR analyzer operating on the same principal as CO2 analyzers. This unit was
relocated from the stack immediately prior to beginning the test program to provide adequate range
for the location.
The NOx analyzer (Model 400-HCLD, California Analytical Instruments, Orange, CA) to be used
in the study operates via chemiluminescence. Sample is directed to a converter where the NO2
component is dissociated to form NO. A small portion of the sample flow is metered into a vacuum
(reaction chamber) where it is allowed contact with an excess of ozone from an integral ozonator.
NO and ozone react to form NO2, a portion (-10% at room temperature) of which is elevated to
an excited state. The excited molecules return to ground state and give off light of a characteristic
frequency. This light is detected by a photomultiplier tube, and the output is amplified and scaled
to read directly in parts per million by volume.
Total hydrocarbon concentrations were measured using a California Analytical Instruments Model
300 heated FID.
4.3.2 Closed-Cell FTIR
An extractive cell based FTIR spectrometer was configured for sampling from the dilution duct.
The FTIR system used for this study was an Industrial Monitor and Control Corporation (IMACC;
Round Rock, TX), spectrometer equipped with a Micheleson inferometer, a zinc selenide beam
splitter, a mercury cadmium telluride detector, and a 12-L, 1-m multi-pass gas cell with gold-
coated mirrors and a stainless steel coated body. The interferometer performs an optical inverse
Fourier transform on the entering IR radiation. This modulated IR beam passes through the gas
sample where it is absorbed to various extents at different wavelengths by the various molecules
present. Buried in the IR spectrum is the absorption "fingerprint" of all gases in the air sample
through which the IR beam passes. This is caused by IR radiation interacting with the molecules
and the interaction resulting in molecules absorbing specific wavelengths or "colors" of the
radiation. The absorption adds energy to the molecule and causes it to vibrate and rotate faster.
The vibrations and rotations of molecules are dictated by their structure. This means the patterns
of "colors" that are absorbed are also unique to each molecule. The presence of a specific pattern
is unequivocal evidence of the presence of a specific compound and the intensity of the absorption
is proportional to the concentration of the compound in the path.
The IR source used in the FTIR is a SiC ceramic at a temperature of 1550 K. The IR radiation goes
through an interferometer that modulates the infrared radiation. Spectra were produced from 300
co-added interferograms (5-minute scan time) that were apodized with Happ-Genzel function, and
then transformed to yield a single-beam spectrum with a nominal 0.5 cm"1 resolution. Reference
spectra were generated using E-trans from the Hitran database, and the use of the Pacific Northwest
19
-------
Laboratories (PNL) spectral library. A graphical software package was used to create and test
custom analysis routines for the FTIR. All routines created can have full linearity correction, lines-
shift correction, and cross interference correction as well as dynamic reference selection to provide
real-time matching of the method.
The IMACC spectrometer was interfaced to a laptop computer via computer cables. Spectral data
were collected at 5-minute intervals at a nominal 0.5 cm"1 resolution using IMACC FTIR Software
Suite. For this experimental design, a pump system was installed to pull the sample through the
cell at a rate of approximately 45-55 liters per minute. The sample was drawn from the constant
volume sampler. The sampling lines were heated to 100 °C.
For each analyte of interest there is a specific wavelength which is measured and the concentration
values are determined. Interfering species are identified as well. Single beam spectra collected
during the testing phase were converted to absorbance spectra. This was done by either selecting
a background spectrum prior to running a burn or by generating a synthetic background spectrum.
The IMACC software suite and IMACC Quantify was used to generate the concentration values.
It allows the user to shift references as needed, save residuals and do linearity and bias plots. The
software is based on linear regression.
4.4 Volatile Organic Compound and Carbonyl Analyses
VOCs were analyzed by EPA Method TO-15 in SIM mode using an Agilent Model 6890/5973N
GC/MS (Agilent Technologies, Santa Clara, CA) using Agilent ChemStation E.02.01 software.
Sample aliquots were taken from canister samplers by an Entech 7500A Autosampler (Entech
Instruments Inc., Simi Valley CA) and preconcentrated using an Entech 7150 Preconcentrator
followed by GC/MS analysis. NIST traceable VOC gas standards (Linde Electronics and Specialty
Gases, Stewartsville, NJ) were used to prepare calibration standard samples in canisters with an
Entech 4600A Dynamic Diluter, which were used to calibrate the GC/MS instrument response.
Samples were analyzed within a week of sampling. A lab blank sample was analyzed with each
GC/MS analytical sequence and was used to blank correct sample concentrations. The QC
procedures and data validation criteria specified in the method SOP were followed.
N2O concentrations were measured in the SUMMA canisters by analyzing canister samples on a
SRI Model 8610C GC with an electron capture detector (SRI Instruments, Torrance, CA). The GC
uses a backflush system controlled by a 10-port valve and a 183-cm (6-ft) Hayesep D precolumn
with a 366-cm (12-ft) Poropak Q analytical column. The GC detector is a 5 mCi 63 Ni 140 BN
electron capture detector. The makeup gas was 10% methane in argon (Airgas National Welders;
Raleigh, NC). The instrument was calibrated using specially prepared calibration standard samples
in canisters. Calibration standards were prepared from a certified N2O cylinder (Airgas-National
Welders; Raleigh, NC) following the same methods as preparing the VOC calibration standards in
canisters using the Entech 4600A Dynamic Diluter.
Carbonyls were analyzed via EPA Method TO-11A by HPLC. DNPH cartridge samples were
extracted with 6 mL of carbonyl-free acetonitrile (Burdick and Jackson). The exact volume of each
extract was determined gravimetrically and the density of acetonitrile. The extracts were analyzed
20
-------
by HPLC using an Agilent Model 1100 HPLC with a diode array detector and Agilent
ChemStation A. 10.02 software. Samples were extracted and analyzed within 2 weeks of sampling.
A lab blank DNPH sample was analyzed with each HPLC analytical sequence and was used to
blank correct sample concentrations. Carbonyl-DNPH standards (Sigma-Aldrich, St. Louis, MO)
were used to calibrate the HPLC instrument response for each target analyte. The extraction and
HPLC analytical procedures for the carbonyl analysis followed MOP 2700, Standard Operating
Procedure for Determination of Carbonyls in Ambient Air Collected on DNPH-Coated Silica
Cartridges Using the Agilent 1100 HPLC.
4.5 Gaseous PAH Analyses
The target PAH compounds from the Method 23 train filter and sorbent were analyzed using
modified EPA Method 8270D on a Thermo GC Trace 1310/ISQ (Thermo Scientific, Inc., Milan,
IT/Austin, TX USA) using Xcaliber 2.2 software. Labeled standards for PAHs were added to the
XAD-2 trap before the sample is collected. The surrogate recoveries were measured relative to the
internal standards and are a measure of the sampling train collection efficiency. Internal standards
were added before extraction. Before analysis, a third set of labeled standards were added to
quantitate the recovery through the extraction and concentration process. The semivolatile XAD
and filter samples were prepared for analysis by solvent extraction using toluene and then a
concentration by three-ball Snyder column; then the sample was split for PCDD/F and PAH
analysis. The portion for the PAH analysis was 10% of the total. The PAH portions were filtered
through silica gel and concentrated to final volume in a TurboVap II using nitrogen blowdown.
The extract was prescreened to determine the level of dilution needed for PAH analysis. Samples
were analyzed using selected ion monitoring (SIM) mode. All surrogate standard recoveries fell
within the standard method criteria (25% to 130%) except Naphthalene which was below on
several samples. This is an early adoption of the three-spike style of standards for PAH analysis,
modeled after the three-spike style of Method 23 for PCDD/F.
After being split for HC1 analysis, select impinger solutions were extracted by liquid-liquid
extraction and screened for PAHs. The amount of PAH was <5% of the amount found on the filter
and XAD resin, therefore the impingers were not extracted and analyzed for PAHs.
4.6 Gaseous PCDD/PCDF Analyses
The Method 23 XAD and TX40 filter samples were extracted and cleaned up according to EPA
Method 23 and analyzed for PCDD/F using HRGC/HRMS consisting of a Hewlett-Packard gas
chromatograph 6890 Series (Agilent Technologies Inc., Wilmington, DE) equipped with a CTC
Analytics Combi PAL autosampler (CTC Analytics, Switzerland) and coupled to a Micromass
Premiere (Waters Inc., UK) double-focusing high resolution mass spectrometer using Masslynx
4.1 software. The chromatographic column used was an RTX-Dioxin 2 (Restek, Bellefonte, PA,
USA).
21
-------
The standards used for chlorinated dioxin/furan identification and quantification were a mixture
of standards containing tetra- to octa-PCDD/F native and C-labeled congeners designed for
modified EPA Method 23 (ED-2521, EDF-4137A, EDF-4136A, EF-4134, ED-4135, Cambridge
Isotope Laboratories Inc., Andover, MA). The PCDD/F calibration solutions were prepared in-
house and contained native PCDD/F congeners at concentrations from 0.5 (ICAL1) to 20 (ICAL6)
ng/mL.
Initial concentration steps were performed using a three-ball Snyder column, and then 10% of the
extracts were separated for PAH analysis. The remainder of the extract were combined,
concentrated, and solvent-exchanged into hexane. The extract was cleaned by a PowerPrep (Fluid
Management Systems Inc., Watertown, MA) for PCDD/F analysis. The PowerPrep is an
automated device that performs the cleanup specified in Method 23. A keeper (decane) is used
after extract cleanup with the PowerPrep to prevent samples from going to dryness.
4.7 Gaseous Halide (HCI) Analyses
The sodium hydroxide solutions from the Method 23 samples were analyzed for HCI according to
Method 26A analytical procedure. Samples were analyzed for chloride using a Dionex (Sunnyvale,
CA) DX500 chromatography system. This HPLC system used a GP40 gradient HPLC pump, a
CD 20 conductivity detector, and an AS40 automated sampler. The IC system used a Dionex
AS 12a 4 mm x 200 mm analytical column, and an AS12A 4 mm x 50 mm guard column. The
analysis used a 2.7 mMNa2C03 and 0.3 mMNaHC03 eluent pumping at 1.5 mL per minute. The
chloride calibration standard was prepared by Dionex (Lot# 23-116VY) and is traceable to NIST
Standard 3182. Standards and sample 1 to 10 dilutions were prepared volumetrically using an
Eppendorf (Hamburg Germany) Repeater M4 pipette.
4.8 Total PM Mass Analyses
For the determination of total PM mass emissions on a time-integrated basis, the PM emission
measurements adhered to the procedures outlined in ASTM Method E2515-11, Standard Test
Method for Determination of Particulate Matter Emissions Collected by a Dilution Tunnel with
the exceptions noted in Section 3.2.1. The filter pre- and post-weighing was performed in the EPA
temperature and humidity-controlled weigh room using an ATI Cahn Model C-44 (Thermo Fischer
Scientific, Waltham, MA) analytical balance. The weighing room is kept at 22 ± 2 ° C and 35 ± 5
% relative humidity where the filters are equilibrated for at least 24 hours before tare and final
weighing.
4.9 Particle Number and Size Determination
As discussed above, particle number concentration and size distribution were determined using the
Dekati ELPI instrument. The instrument was set up to perform one complete scan every 10 sec
during the entire 6-hr run. The total number concentration was calculated by averaging the total
particle count for all the impactor stages during each scan conducted over the 6-hr test period. A
composite particle size distribution (PSD) was also produced by calculating the average
22
-------
differential number (dN/dlogDp) in each size bin for each test and combining these into a complete
PSD. The composite PSD was then converted to a mono-modal lognormal distribution and
summary statistics calculated from the data.
Daily checks on the ELPI were performed by zeroing the electrometers using a flush of high-
efficiency particulate air (HEPA) and by performing a leak check, flow rate check, and instrument
zero check using a HEPA filter. Multiple pre-test zeroing of the electrometers was performed to
assure reliable measurements.
4.10 OCEC and Black Carbon Analyses
4.10.1 Laboratory OCEC Analyses
Particulate OC/EC analysis of 47-mm and 25-mm quartz filter samples was performed in-house
according to a modified version of the NIOSH Method 5040 using a Sunset Model A TOT carbon
analyzer (Sunset Laboratories, Portland, OR) running the Sunset NIOSH870.par software
program. The laboratory analysis consists of heating up a filter section in steps from ambient
temperature to 870 °C. Carbon species are volatilized off the filter, oxidized to CO2, reduced to
CH4, and quantified with a flame ionization detector (FID). Laser transmittance is used to correct
for pyrolyzed OC. The split between organic (low temperature) and elemental (high temperature)
carbon is operationally defined. The detection limit is 0.2 |ig carbon/cm2.
4.10.2 Semi-Continuous OCEC Analyzer
Operation of the Sunset Laboratories Model 4 semi-continuous OC/EC analyzer followed the
procedures outlined in the manufacturer's operating manual. Like the laboratory system, the
instrument also ran the Sunset NIOSH870.par operating and analysis software. The principle
behind this instrument is very similar to that of the OC/EC model instrument described in Section
4.10.1 except CO2 is measured directly with a non-dispersive infrared spectrometer. As with the
integrated sample described above, quartz filters used with the semi-continuous instrument were
pre-conditioned in an internal oven following the instrument procedures prior to sampling. Each
sample has an area of 1.2 square centimeters and was collected semi-continuously, from the
secondary dilution system (discussed in detail in Section 2.1) for ten minutes followed by onboard
analysis of the sample for eighteen minutes. These measurements were considered non-critical as
they were sampled in addition to the collected time-integrated filters discussed in Section 4.10.1.
4.10.3 Optical Black Carbon
The AE-22 Aethalometer provided a continuous measurement (10 s time resolution) of the
attenuation at 880 nm of PM deposited on a filter spot. The attenuation is converted to OBC
concentration using the manufacturer's calibration. Filter based measurement of black carbon
attenuation is subject to a loading artifact leading to an underestimate of the OBC concentrations.
As strongly absorbing PM loads on the filter, the optical path length through the filter is reduced
23
-------
as the light scatter is reduced by the aerosol. The black carbon concentration is corrected by method
derived by Virkkula et al (2007):
BCcorrected. (1 "l~ /c X ATN) X BCmeasure(i
Where BCcorrected is the black carbon concentration corrected for filter loading, k is the correction
factor which is dependent upon the particle properties, ATN is the filter attenuation at 880 nm and
BCmeasured is the concentration reported by the AE-22. The k factor cannot be estimated when black
carbon concentrations are changing through filter advances. To overcome this limitation a
comparison between BCcorrected and Babs measured by the PAX was made to determine the k factor
that provided the highest correlation between the two measurements.
4.11 PM Elemental Analyses
Elemental analyses of Teflon filter samples were performed by Chester LabNet using XRF as
described in EPA Method 10-3.3. This method is applicable to the quantitative analysis of aerosols
deposited on a variety of filter types for the elements sodium (Na) through uranium (U). The QC
checks set by the laboratory include a QA standard, which is a multi-element thin-film vapor-
deposited National Institute of Standards and Technology (NIST) certified standard on Mylar
manufactured by Micromatter, Inc. (Vancouver, BC, Canada). Elements analyzed were aluminum
(Al), silicon (Si), potassium (K), sulfur (S), calcium (Ca), titanium (Ti), vanadium (V), manganese
(Mn), iron (Fe), copper (Cu), zinc (Zn), and lead (Pb).
4.12 Particle Morphology
Samples for particle visualization were deposited on TEM grids and analyzed by R. J. Lee Group
using a Transmission Electron Microscope. Samples were analyzed for particle count to determine
aerosol concentration, particle morphology and elemental composition.
4.13 PM Semivolatile Organics Analyses
Sixty-one pre-heated (550 °C, 12 h) quartz filters were collected as part of the study. These filters
were collected with the intent of performing organic matter speciation. Prior to speciation, all
quartz filter samples were stored at -65°C for less than one year.
The organic and elemental carbon (OC and EC) composition on each PM filter (1.5 cm2) was
measured using thermal-optical analysis and a modified NIOSH 5040 method (Cassinelli and
O'Connor, 1998). Total extractable OC was estimated using these OC-EC values. To ensure
adequate OC mass for a successful gas chromatography mass spectrometry (GC-MS) analysis, a
sample compositing strategy was required. The OC sample loads and resulting filter composite
strategy is available upon request. Past studies have demonstrated that at least 100 ug of filter OC
is required to achieve reasonable GC-MS results (or about 0.3 ug OC for a 300 |il final extract
volume). However, lower OC concentrations were used for this data set due to the high sensitivity
of the GC-MS. All filter-based, organic compound emission factors are normalized to OC in this
report.
24
-------
The extraction and GC-MS conditions used for this investigation were described earlier (Hays et
al., 2013; Hays et al., 2011; Hays et al., 2002). Briefly, prior to undergoing a solvent extraction,
each quartz filter composite was placed in a 50-ml glass jar and spiked with an internal standard
mixture containing d-8 naphthalene and C-13 levoglucosan compounds. Internal standard spike
volumes changed on the basis of anticipated final volume of extract. Use of the internal standard
method allowed us to compensate for extraction losses and changes in MS response over 24 hr.
Filters (typically n=3) were extracted twice (50 min and 5 min) ultrasonically with roughly 10 ml
of a 2/2/1 vol/vol hexane, benzene and isopropanol solution (HIB). Each extract was filtered with
a 0.2 |im PTFE filter (Supelco, Iso-Disc™) and then concentrated to between 0.3 ml and 1 ml
depending upon the OC concentration extracted. Sample extracts underwent derivation to convert
the organic acids and levoglucosan to their methyl ester and silyl-ester analogs. Methylation was
performed by reacting 50 |il of sample extract with 50 |il of in-house prepared diazomethane
reagent and 15 |il of methanol and allowing the reaction to proceed for at least 1 hour. The hydroxyl
groups on levoglucosan were silylated by reacting 10 |il of sample extract with 50 |il of BSTFA
reagent (Sigma Aldrich, St. Louis, MS). The reaction was allowed to proceed for 30 minutes at
70°C and then allowed to sit at room temperature overnight to ensure completion. The neutral and
derivatized extracts were analyzed as described below.
Sample extracts were analyzed by GC-MS for a total of 115 organic compounds representing
eleven compound classes. The compound classes included normal-alkanes, branched-alkanes,
polycyclic aromatic hydrocarbons (PAH), anhydrosugars, aromatic, resin, alkanoic, and fatty
acids, aliphatic diacids, phytosterols, and methoxyphenols. The methoxyphenols were analyzed
using thermal extraction (TE)-GC-MS (TDS3, Gerstel Inc, Baltimore MD, and Agilent
Technologies 6890/5973 MS [q]). For TE-GC-MS, a 1 ju.1 volume of each sample extract was
injected manually onto a baked Carbotrap F/Carbotrap C adsorbent tube. The solvent from each
sample spike was evaporated by flowing nitrogen across each adsorbent tube for 60 seconds at a
rate of 50 ml/min. All other organic compounds were analyzed using a GC-MS (Agilent
7673A/7000 series triple quadrupole [qqq] system interfaced to a liquid sample auto-injector).
4.14 Fuel Analyses
Ultimate, proximate, and ash mineral content by Standard Laboratories, Inc. ASTM Methods
D2961, D3302, D3173, D3175M, D3172, D3174, D4239 METHOD B, D3178, D5373, D3179,
D3176, D5865, D4208, D2795, and D3682 were used during the analyses conducted.
4.15 Bottom Ash Analyses
Collected bottom ash samples were analyzed for mass loss upon heating in air or oxygen
atmosphere by Standard Laboratories, Inc. Bottom ash samples were also analyzed by Standard
Laboratories, Inc. for elemental composition by AA spectroscopy using ASTM D3682.
25
-------
5 Data Analysis
The overall objective of this project is to assess the energy efficiency of the PBHH and the
emissions of target compounds as a function of fuel input, energy input, and energy output. Energy
efficiency calculations have been limited to thermal efficiency considering the heat delivered to a
simulated load and the gross calorific value of the fuel fed to the PBHH. Emissions are calculated
from measured concentrations and dilution duct flow. Measurements from the stack and from
secondary dilution are corrected to dilution duct concentrations in the process of calculating
emissions.
5.1 Thermal Efficiency
Thermal efficiency is used to determine the overall ability of the system to generate useful heat by
transferring it to meet the load demand (thermal efficiency). Thermal efficiency in this project was
defined as the heat delivered to the water/water heat exchanger (heat output) divided by the
calculated energy input of the fuel (gross calorific value [HHV]) of the heater defined as:
The useful heat delivered was calculated using the inlet and outlet temperatures of the cooling
water for the heat exchanger used to simulate the heat load demand, the water flow rate at each
temperature reading, and the heat capacity of water. In practice, the heat transfer rate was averaged
over the entire test and multiplied by the run time:
(5-1)
where:
Qt
the thermal efficiency,
the useful heat delivered to cooling water,
the energy input to the heater during the test.
_ ZiFCp AT
(5-2)
where:
To
Ti
AT
t
F
CP
¦p
mass flow rate of water,
heat capacity of water (4.18 kJ/kg -°C),
temperature difference between cooling
water outlet and cooling water inlet,
To - T;
outlet temperature,
inlet temperature, and
run time
26
-------
The energy input was calculated from the mass of fuel fed during the test and the gross calorific:
Qt=Mf.HHVf (5_3)
where:
Mf = mass of fuel burned and
HHVf = higher heating value of the fuel
5.2 Emission Calculations
For each test run, emission factors for the target pollutants were calculated in terms of mass of fuel
burned, energy input, and energy output. The mass emission (Mx) for a set burn time l for each
target compound is calculated in the dilution tunnel as follows:
Mx = Zt(CX)t - CXia) ¦ V ¦ t (5-4)
where:
Cx,t = the concentration (mass/volume) of
the target compound x in the dilution
duct,
Cx,a = the ambient concentration
(mass/volume) of the target
compound x, and
V = the volumetric flow rate
(volume/time) in the dilution tunnel
at time t.
Concentrations from the Method 23 train (HC1, PAH, PCDD/DF) at the stack are an exception to
this procedure as no ambient concentration was available. Most concentrations were determined
on a test average basis; volatiles and semi-volatiles (i.e. TOl 1 and T015 samples) were split into
two independent 3-hour samples. The ambient concentration for total PM utilized an ambient air
sample extracted beside the PBHH during 100% load operation firing switchgrass pellets.
Remaining concentrations were corrected with concentrations determined from dilution duct
sampling with no firing (i.e. cold) of the PBHH.
Volumetric flow rate was determined by multiplying the average of the dilution duct velocity
measured before and after each test by the cross-sectional area of the dilution duct at the point
measured. The 10-inch duct had a cross-sectional area of 0.0506 m2 (0.545 ft2). Emission factors
were calculated and reported in three bases: per mass of fuel burned, per unit of energy input, and
per unit of energy output.
27
-------
The emission factor per mass of fuel burned (EFm,x) is calculated as:
EF =¥*-
"* Mf (5-5)
The emission factor per energy input EFi„put,x is defined as the mass of pollutant (x) per fuel energy
generated by the PBHH unit:
EFInputiX = ^ (5-6)
vi
The emission factor per energy output is defined as the mass of pollutant (x) discharged per useful
room heat produced by the heating unit:
EF0utPutlX = 7T (5-7)
vo
5.3 Dilution Factor
Measurements made at the stack or after secondary dilution were corrected to dilution duct
concentrations using a dilution factor. Due to failure in the CEM measurements, estimates were
made for CO2 concentrations in the dilution duct for use in dilution factor calculations. The average
CO2 in the dilution duct was estimated based on the mass of fuel burned in each test (Mf), the
carbon concentration in the fuel from the ultimate analysis, and the volumetric flow rate
determined for the test. The calculation estimates the volume of CO2 emitted divided by the volume
of flow in the dilution duct corrected for ambient CO2 concentration:
_Mf-%C 1 SV
Lc°lt ~ 100-t MWCarbon V + Cc02'a V-*>
where:
Cco2,t = the CO2 concentration in the duct,
Cco2,a = the ambient concentration of CO2,
%oC = the carbon concentration in the fuel
(weight percent),
MWcarbon = the molecular weight of carbon, and
SV = the specific volume of an ideal gas at
20 °C and 1 atmosphere.
Concentrations measured at the stack or on secondary dilution were corrected to dilution duct
concentrations for emission calculations.
28
-------
The dilution factor at the stack was based on the estimated dilution duct CO2 concentration and
the average stack CO2 concentration:
r\ r* CCQ2,t CCQ2,a /c q\
~~ Stack C02 ^ '
The dilution used for secondary dilution was performed with nitrogen containing minimal CO2:
DF = £co2^ ^5_10^
SDCO 2,t
where:
SDco2,t = secondary dilution CO2
concentration during sample
interval
Because the nature of operations and the short-term nature of the measurements taken on secondary
dilution, each measurement was corrected using time specific dilution factors. Due to the quality
of duct CO2 measurements, time specific dilution duct CO2 concentrations were estimated using
the average dilution duct CO2 measurements for the specific sampling time, average dilution duct
CO2 concentrations for total test time, and the average fuel based estimate of dilution duct CO2
concentrations calculated as above:
„ Fuel Based Average C02 ,, n n N
Crn? t = CEM sample averaqe C02 (5-11)
CEM Test average C02 f u \ >
29
-------
6 Quality Assurance and Quality Control
6.1 Overall Objectives
The overall objectives of the program were to:
1. Develop PM and gaseous emission factors and chemical source profiles for a PBHH
operating at different load demand rates while burning multiple non-woody fuel types.
2. Determine the energy efficiency of the pellet-burning appliance using different fuel types
and load demands.
3. Determine, if possible, the effect of non-woody fuel properties on the PM and gas-phase
emissions as compared to premium wood pellets.
4. Assess, to the degree possible, any adverse effects of burning non-woody biomass fuels on
the appliance tested.
All of the above objectives were met as discussed in Section 7 below except for the testing of
multiple non-woody fuel types. In this study, due to resource constraints, only switchgrass pellets
were evaluated.
6.2 Data Quality Objectives (DQOs)
The DQOs for the project are as follows:
1. Determine the total PM mass emission factor (g/kg fuel) and thermal efficiency within
± 25% (relative percent difference [RPD]) for duplicate measurements conducted on each
fuel type at the same load demand.
2. Where possible, achieve an agreement within ± 25% (RPD) between the time-integrated
and continuous measurement of the same pollutant or chemical characteristic.
3. Attain a data recovery and analysis of at least 90% of the samples and/or 90% of the
continuous monitoring time scheduled for all sampling runs conducted.
4. If possible, ensure that the samples collected are representative of the normal operation of
the appliance as determined by the comparison to any similar data published in the
literature.
Regarding DQO 1 for total PM, the data shown in Table 7-5 below indicates that this goal was met
for all fuel and load conditions apart from switchgrass combustion at 25% and 100% load. At 25%
load, the RPD for the duplicate tests was ~ 35% and at 100% load ~ 40%. Even this level of
agreement is remarkable due to the high variability in the emissions observed during each test and
between test runs. For thermal efficiency, DQO 1 was met for switchgrass combustion at all load
conditions but not for hardwood combustion. In the case of hardwood, the variation between tests
30
-------
ranged from 30% for the Syracuse cycle to a factor of 2.5 for 25% load. This is not surprising due
to the high variability seen from test to test.
For DQO 2, this goal was not met since there were no concurrent time-integrated and continuous
measurements conducted for any of the parameters measured. In the case of DQO 3, the 90% goal
was met for all samples collected and continuous monitoring conducted in the program. Finally,
DQO 4 was met by comparing the data collected in this study to both prior work for NYSERDA
conducted by NRMRL and data published in the literature as described in Section 8 below.
6.3 Data Quality Indicator Goals (DQIs)
DQIs were established for the measurement program as outlined in the approved Quality
Assurance Project Plan (QAPP). The following sections provide the quality assurance and quality
control activities for each set of parameters measured along with compliance with the DQI goals
outlined in the QAPP and the implications on data quality for those not complying with a particular
DQI.
6.3.1 Stack Testing and Thermal Measurement Parameters
Stack concentrations were calculated as emissions using the duct flow measured with the
Shortridge airfoil. While this is consistent with the overall ASTM E2515 approach, it represents a
deviation from the planned emission factor-based approach. The Shortridge airfoil calibration was
checked before initiating the test program and was found to be within instrument specification of
± 7 feet per minute plus 3% of reading. Readings during the test program ranged from -900 to
1300 feet per minute. As a result, stack flow measurements are expected to be accurate within ±
3.5%.
Thermal performance of the PBHH was determined by the heat transferred from hot water
circulating through the boiler to cold water supplied by the facility. Cold water temperature
entering and exiting the heat exchanger was measured with Type K thermocouples calibrated over
the range of 10 °C to 100 °C. The expanded uncertainty for any single measurement was ± 0.93
°C at the heat exchanger inlet and ± 0.60 °C at the heat exchanger outlet. With one second polling
over a 6-hour test, the random error around a run average temperature would be minimal. The
combined standard uncertainty for each thermocouple (after calibration) was ± 0.06 °C
The chilled water flow through the heat exchanger was controlled and monitored through a
rotameter. Flow through the rotameter was checked gravimetrically during hardwood testing
(10/24/2016) and after completion of switchgrass testing (11/17/2016) yielding correction factors
of 1.04 and 1.02 respectively. Accuracy is well within the DQI goal of ± 10 %. A 1.03 correction
factor was applied to all flow readings, the average correction from both checks.
31
-------
6.3.2 Continuous Emission Monitoring
CEM's sampled from both the PBHH stack and from the dilution duct. On conclusion of the
sampling program, a substantial discrepancy was observed between the dilution duct CO2
measurements and the CO2 expected from fuel combustion. This anomaly had not been noticed
during the test campaign due to the decision to forego routine direct calibration checks relying
instead on bias and drift checks. Despite acceptable pre-campaign direct calibration, it was
concluded that a leak must have developed in the sampling train. Low dilution duct CO2
measurements were confirmed with analysis of TO 15 samples for CO2. The measurements for
most pollutants at the dilution duct were abandoned in favor of what was considered more reliable
FTIR measurements. The dilution duct CO2 measurements were adjusted based on fuel use to
allow time resolved estimates of dilution factors for measurements using secondary dilution off
the duct. Discussion of other dilution duct CEM measurements has been omitted for brevity as
they were not used in this report.
Prior to initiating the sampling program, the stack CO2 CEM was calibrated directly to the analyzer
with certified (± 5%) calibration gas. Calibration checks, directly to the analyzer, were performed
on 10/14/2016 with nitrogen, 9.0% CO2 in nitrogen, and 18.1% CO2 in nitrogen. Calibration error
was found to be 2.49% of span for zero gas, 0.31% of span for mid-range gas, and -2.62% of span
for high range gas. Daily bias checks were performed prior to testing except for the 10/25/2016
test; data acquisition difficulties resulted in testing delays on the 25th necessitating omission of this
QA check. In addition, daily drift checks were performed after each test. Furthermore, on
11/10/2016 the upscale drift check failed to achieve lineout due to unexpected expansion of
response time. Results for these checks are presented in Table 6.1. All data was corrected based
on daily bias and drift checks.
Prior to initiating the sampling program, the dilution duct CO2 CEM was calibrated directly to the
analyzer with certified (±5%) calibration gas. Calibration checks, directly to the analyzer, were
performed on 10/12/2016 with nitrogen and 4509 ppmv CO2 in nitrogen. Calibration error was
found to be 1.5% of span for zero gas and 0.2% of span for high range gas; no midrange gas was
tested for this instrument. Daily bias checks were performed prior to testing except for the
10/25/2016 test; data acquisition difficulties resulted in testing delays on the 25th necessitating
omission of this QA check. Furthermore, the bias check on 11/10/2016 failed to achieve lineout
due to unexpected expansion of response time. In addition, daily drift checks were performed after
testing.
The drift check failed to achieve lineout for 10/09/2016 test due to unexpected expansion of
response time. Results for these checks are presented in Table 6.2. All data was corrected based
on daily bias and drift checks.
32
-------
Table 6.1. Daily Stack CO2 Quality Assurance Checks
Date
10202016
10212016
10252016
10262016
10272016
11032016
11042016
11082016
11092016
11102016
11152016
Zero Bias
1.17%
-0.76%
NA
-1.19%
-1.46%
4.42%
-2.23%
-0.55%
0.06%
-0.29%
-0.55%
Upscale
Bias
7.13%
3.81%
NA
4.81%
4.10%
2.55%
7.07%
2.93%
1.16%
3.68%
0.74%
Zero Drift
-1.93%
8.38%
NA
-0.27%
-0.84%
-6.65%
2.04%
0.61%
-0.34%
Failed
0.22%
Upscale
Drift
-3.33%
-5.30%
NA
-0.71%
0.06%
4.52%
-1.12%
-1.77%
2.52%
Failed
2.60%
Table 6.2. Daily Dilution Duct CO2 Quality Assurance Checks
Date
10202016
10212016
10252016
10262016
10272016
11032016
11042016
11082016
11092016
11102016
11152016
Zero Bias
-4.95%
-6.47%
NA
0.95%
1.15%
0.82%
0.85%
0.87%
0.97%
Failed
0.78%
Upscale
Bias
1.20%
2.47%
NA
-8.31%
-11.76%
-10.05%
-9.59%
-10.84%
-8.26%
Failed
-7.99%
Zero Drift
-1.53%
1.29%
NA
0.20%
-0.42%
0.03%
-0.26%
0.10%
Failed
NA
0.14%
Upscale
Drift
1.27%
-2.88%
NA
-3.45%
-2.98%
0.46%
-1.09%
2.58%
Failed
NA
-1.84%
33
-------
6.3.3 Total PM and Filter Mass
Dry gas meters were calibrated approximately 4 months prior to beginning hydronic heater tests.
Two dry gas meters were used to meter sample volumes for the paired PM trains and a third dry
gas meter used for ambient air PM blank train. All calibration readings were within 2% of the
average gamma. No further pre-test volume check was performed; however, the accuracy of the
volume measurements was confirmed by successful calibration drift checks after the test program.
Drift was well within ± 5% drift tolerance: -2.2% and 0.7% for the paired PM train meters and
0.7% for the ambient air PM train meter.
All glass fiber filters for PM were weighed to constant weight with a stability requirement of ±
0.5mg as associated with Method 5. Except for two filters for one hardwood 25% Load run
(10/21/2017), all filters were stable at < ± 0.07 mg. Total catch per train ranged from 2.059 mg to
25.927 mg. Filters were weights were checked against a standard 200 mg weight with a tolerance
of 0.003 mg.
All PM concentrations were corrected for ambient air PM contributions. Ambient air was sampled
near the PBHH while running at full load with switchgrass on 11/16/2016. For the majority of
measurements, the ambient background was less than 5% of the average test concentration. For
the hardwood 100% load test on 10/19/2016, ambient background was 5.6% of the measured PM
concentration. For the hardwood Syracuse load tests on 10/26 2016 and 10/27/2016, ambient
background was 15% of measured PM concentration for both tests.
6.3.4 Total Halide Emissions
Samples collected for halide determination were collected from the Method 23 train used for PAHs
and PCDD/Fs sampling from the PBHH stack prior to dilution. Two meter boxes were used during
the testing for the Method 23 and HC1 tests; both were calibrated ~ 4 months prior to beginning
hydronic heater tests. All calibration readings were within 2% of the average gamma. No further
pre-test volume check was performed; however, the accuracy of the volume measurements was
confirmed by successful calibration drift checks after the test program. Drift was well within ±5%
drift tolerance: 0.5% and 0.7%.
The chloride concentrations and emissions are flagged because the chloride measured in the train
liquid are not large compared to the liquid in the field blank. The DQO goal blank sample
concentration having less than 5% of a test sample concentration. The field blank liquid chloride
concentration ranged from 16 to 106% of test liquid concentration. Following Method 26A
procedures, the test concentrations were background corrected for the chloride in the field blank.
6.3.5 In-Stack PAHs and PCDD/Fs
Samples collected for PAHs and PCDD/Fs s utilized a Method 23 train sampling from the PBHH
stack prior to dilution. Two meter boxes were used during the testing for the Method 23 trains;
both were calibrated ~ 4 months prior to beginning hydronic heater tests. All calibration readings
34
-------
were within 2% of the average gamma. No further pre-test volume check was performed; however,
the accuracy of the volume measurements was confirmed by successful calibration drift checks
after the test program. Drift was well within ± 5% drift tolerance: 0.5% and 0.7%.
A field blank for the Method 23 sample train was created during operations on 11/15/2016. The
field blank contained < 9% of any PAHs recovered compared to a DQI goal of <50%; sufficient
analyte was recovered during tests to attribute PAH's to the sample. The majority of analytes in
the field blank sample were below level of detection. Using the TEQ level of detection for
compounds not detected, the field blank remained <50% of each of the components quantified for
all samples. The majority of analytes achieved much better performance with these criteria with
notable exceptions for the low toxicity and low concentration compounds 1,2,3,4,6,7,8,9 - OCDD
and 1,2,3,4,6,7,8,9 - OCDF. Over the total TEQ, the blanks at the level of detection for compounds
not detected are less than 3.4% of test samples.
The Sampling to Extraction hold time DQI of 60 days was met for the first batch of samples but
was exceeded by around 40 days for the second batch of samples. The samples were stored in the
cold and dark, and the hold times are believed to have been established for methods which include
much more labile compounds not just PAHs and PCDD/Fs It is not expected for the extra 40 days
before extraction to have any significant effect on the samples.
The recovery criteria DQI for PAHs was set for this project similar to the criteria for Method 23.
Because this is a process that is designed to provide a more rigorous QA than Method 8270 the
criteria have not been fully developed. For the Pre-extraction spikes and the pre-sampling spikes,
the criteria were set at 25-130%). The pre-sampling spike was between 47 and 105% with the
exception of two very high level samples and interference is suspected. Because the actual
quantitation of the targets is by isotope dilution this is not expected to cause significant error in
the reported values. The Pre-extraction spike had 4 compounds that were consistently above the
criteria between 130-200%) and the Naphthalene recovery was below 25%> for half of the samples.
Again, because the actual target quantitation was by Isotope dilution this is not expected to cause
significant error to the values.
The Recovery Criteria for the PCDD/Fs were met except for the first two samples which had high
TeCDD Presampling recovery that because the actual quantitation is isotope dilution it is not
expected to cause significant error in the reported values.
6.3.6 VOCs and Carbonyls
The data quality indicators and associated corrective action for the analysis of VOCs and carbonyl
samples are outlined in Table 6-1 of the QAPP. These DQIs follow guidelines set out by EPA
Methods TO-15 and TO-11 A. The balance used to measure carbonyl extract weights was checked
every day and was within 0.2 mg of calibration weight meeting DQI goals. Pressure checks of
canisters used for TO-15 analysis were within 4.2 kPa of expected value within DQI acceptance
criteria. For carbonyl sampling, flow rate checks were within 10%> of DQI criteria, except one day
35
-------
where it was within 16%. Carbonyls measured in field blanks were all below 80 ng/cartridge levels
except one outlier that was 900 ng/cartridge. This value was still <30% of average measured values
for that day, meeting the DQI criteria.
For VOCs, a daily calibration runs were analyzed to determine system performance each day, and
acceptable recoveries within 30% of expected values were achieved with exception of 6% of
values. In the case of carbonyls, a daily calibration run was analyzed for every 5 sample runs that
were mostly 15% of expected values except three data points, where these were within 22%. One
sample was analyzed in replicate during each analytical sequence with acceptable precision that
was mostly within 10% relative percent difference. When this criterion is not met, the data were
flagged for further evaluation. Detailed information on the results from the QA samples is available
upon request.
6.3.1 Particle Number and Size
The Dekati Electrical Low Pressure Impactor was used to determine particle number and
aerodynamic size during the study. Three DQI goals were established in the approved QAPP for
the following parameters:
• Leak check: <10mbar/min
• Flow check: 10±lLpm
• Zero check (HEPA filter): < 50 particles/cm3
These checks were performed daily before each test conducted. Leak check and zero check results
were read directly off the instrument. The flow check was determined using a TSI Model 4140
portable flow meter checked against a Gilibrator® bubble flow meter.
All daily checks were easily within the DQI limits shown above. The following are the ranges of
values obtained for each parameter:
• Leak check: 0-4 mbar/min
• Flow check: 9.65-10.2 Lpm
• Zero check: 2-7 particles/cm3
Thus, the ELPI data are of high quality and acceptable for use during data analyses.
6.3.8 Optical Black Carbon
The Aethalometer (Model AE22) was used to measure optical black carbon and UV absorbing
particulate matter during the study. The DQI goal for the Aethalometer was a daily zero check
using a HEPA filter. Throughout the duration of the study the black carbon concentrations during
the daily zero check was less than the background value of 660 ng/m3. BC concentrations tended
36
-------
to vary during the zero check due to instrument noise; however, zero-check values were always
less than the 30 Mm-1 absolute DQI objective and at least three orders of magnitude lower than
the lowest average concentration measured during testing.
6.3.9 Laboratory and Semi-Continuous OCEC
Quality control procedures for the laboratory OCEC analysis were applied according to the
approved Quality Assurance Project Plan. A brief description of these practices is provided in
Section 6.3.10 below.
6.3.10 PM SVOCs
Either an average response factor or linear regression was used for calibration and to quantify
organic compound concentrations in the samples. The calibration range varied by target compound
class. It was 0.1 ng/|il - 1 ng/jal for most polycyclic aromatic hydrocarbons (PAH) and 0.625
ng/|il - 6.25 ng/|il for most alkanes. A 5-level levoglucosan standard range of 12.75 ng/jal to 130
ng/|il was used, while a the three-level organic acid calibration range was 2 ng/jal - 16 ng/|il. A
mid-level continuing calibration of 10 ng/|il was used for methoxyphenol analytes. A mid-level
check standard was run daily and used to assess target recovery. If the daily mid-level check
standard failed to pass the laboratory's minimum acceptance criterion (80% of compounds must
agree to within 25% of actual fixed concentration value of standard), it was used as a daily
continuing calibration that updated all target responses. All the methoxyphenol targets were
quantified using a continuing calibration. Detection limits were determined for all target organic
compounds as described in EPA document SW-846 (EPA, 2014) with n=7; t statistic= 3.14.
Typical detection limits for the instrument used in this study were provided elsewhere (Hays et al.,
2011). Values that fall below the detection limit threshold were reported as not detected (ND).
Matrix spikes that considered all standard compounds were performed to determine extraction
recovery. Matrix spike recoveries were used as an additional data quality check, and typical values
are also reported in Hays et al. (2011). Several of the methoxy phenols matrix spike targets were
acceptable while others were lower than expected.
Automated integration results for individual peaks were reviewed and corrected if applicable.
Retention times are critical for the predictability of target analyte components. Because the GC
was equipped with electronically programmable control (EPC), retention times shifted less than
0.1 min throughout the analysis period. Target analyte validity was also determined using fragment
isotopic ratios that exceeded the minimum signal/noise ratio of 3 to 1 and had good proximity to
mid-level check standard retention times. Additional quality control was performed by monitoring
the internal standard response of all samples. Precision was demonstrated by triplicate injection
checks of composite samples. Background correction was performed using dilution tunnel blank
tests for all samples except for those burning hardwood pellets at full load and one test at low load.
Those emission factor values are given 'as is'. In certain cases, background subtractions produced
negative values. Negative values and non-detects were treated as 'missing' during generation of
descriptive statistics. Elution of individual phytosterol compounds was putatively observed for
37
-------
experiments conducted for both hardwood and switchgrass pellets. However, the vast majority of
tests didn't show these compounds, which are not reported here due to the lack of phytosterol
standards. Compliance with the DQIs in the QAPP for PM analysis were achieved for this study
unless explicitly noted above.
6.3.11 Closed Cell FTIR
As mentioned above, a closed cell FTIR was a last-minute addition to the program and used to
measure CH4, SO2, CO, NH3, and NOx. In order to verify that the FTIR spectrometer is operating
correctly a series of Quality Control checks were performed daily. MOP-6807 was followed with
the certain alterations to the following procedures. 2.1.2 Stray Light, 2.1.3.2 Random Baseline
Noise, 2.1.3.3 Signal Strength, 2.1.3.4 Signal -beam Spectrum, and 2.1.3.5 Wavenumber Shifts
and Changes in Resolution were conducted and recorded daily. Since the cell based system is
closed and the path length fixed the zero path length descriptions cannot be followed in some
procedures.
38
-------
7 Experimental Results
The experimental results for the REKA hydronic heater tested are provided in this section. Thermal
efficiency is discussed first followed by gas- and particle-phase pollutants. For both gases and
particles, the test average results for the two runs conducted at each fuel and load condition are
provided which, except for stack PAHs, PCDD/PCDFs, and HC1, are background-corrected unless
otherwise indicated. Since only two tests were conducted at each condition, the standard error
(deviation) could not be calculated as is usually done. Instead, the summary data tables show the
relative percent difference (RPD) for the two tests. RPD is defined as the difference between
emission factor values from duplicate tests divided by the average of the duplicates multiplied by
100 and is an indicator of the variability observed between the two test runs. The data are also
provided in graphical form generally in both engineering and SI units. Note that in the graphs, the
bars indicate the range of values for each parameter and not the RPD or standard error. The bars
are provided to generally indicate the amount of variability observed between the two test average
values which was oftentimes considerable.
For the gaseous pollutants determined by on-line monitoring, the test averages are a simple mean
of the continuous measurements made over the 6-hour test period. In the case of the VOCs, two
sample sets (one set equals 1 SUMMA canister for VOCs and one DNPH cartridge for carbonyls)
were collected during each test each having a duration of 3 hours. The test average was determined
from the two sample sets. For the PAH data from the modified Method 23 sample train, data are
available for the 16 compounds determined over the entire period of each test. Test average
emissions are determined from the two sample sets collected for each fuel/load condition. Finally,
in the case of the PCDDs and PCDFs, the samples were composited prior to analysis since it is
normally difficult to obtain enough sample mass for this type of analysis. In this study, however,
the levels for most of the target compounds were in range but some were above the calibration
range. The blank train had very low levels as expected therefore the compounds are believed to be
from the combustor and not a method artifact.
For determination of average total PM mass emissions, the data from the modified ASTM Method
2515 sample train was used as operated over the entire 6-hour period. Note that each individual
test average can represent multiple filter sets (two filters sampling at the same time) collected
during the run depending on loading. Total PM mass was also determined from the 25-mm Teflon
filters with three samples of 20 minutes each collected during each run. Although these samples
only represent 1/6 of the total test time, these data correlated well with the 2515 results and thus
appear representative of the entire test run. For PM number, the average emission factor was
determined from the continuous EPLI data output for the 6-hour test period. The ELPI data were
also used to develop a composite particle size distribution (PSD) for each test by combining data
from all the scans conducted at 10-sec intervals throughout the 6-hour test. In the case of EC/OC,
the test averages were derived from the three 1-hour 47-mm quartz filters samples. These same
filters were also used for the determination of particle-phase SVOCs after EC/OC analysis. Optical
black carbon averages were determined from the continuous Aethalometer data output after
correction for secondary dilution of the sample extracted from the dilution duct. Elemental
analyses by XRF were conducted on the 25-mm Teflon filters after gravimetric analysis.
Finally, in the discussion of the test results, the data expressed in terms of mass of pollutant per
mass of fuel burned were generally used for the observations made. It should be noted that the
39
-------
same general trends were also present in the results expressed in terms of mass of pollutant per
heat input or heat output and thus the generalizations should be similar regardless of which
reporting convention is used. The experimental results are provided in the following subsections.
7.1 Thermal Efficiency
The thermal efficiency of the unit was determined as described above with the average results
provided in Table 7-1 and shown graphically in Figure 7-1. Note that the bars in the figure
represent the range of values obtained not the measurement uncertainty.
Table 7-1. Thermal Efficiency Summary
Fuel Type
Load
Condition
Value3
Average
Thermal
Efficiency (%)b
Hardwood
25%
Average
79
RPD
85
Syracuse Cycle
Average
94
RPD
27
100%
Average
89
RPD
Switchgrass
25%
Average
63
RPD
17
Syracuse Cycle
Average
72
RPD
9.7
100%
Average
81
RPD
12
a RPD = relative percent difference in efficiency for the two test runs conducted at each fuel/load condition.
b Two significant figures. "Average" for hardwood at 100% load is from a single test
40
-------
120
100
80
60
40
So 20
¦ Wood ¦ Grass
Bars = range of values
25% Load
Syracuse Cycle
Heater Load
100% Load
Figure 7-1. Average thermal efficiency by fuel and load condition
As shown in Figure 7-1, the thermal efficiency for both fuels generally increased with load and a
higher efficiency was always observed for hardwood combustion as compared to switchgrass. In
the case of wood, the Syracuse cycle had the highest efficiency at 94% followed by 89% for 100%
load and 79% load at 25% load, respectively. For grass pellet combustion, the efficiency varied
from 63 to 81% with the highest efficiency at 100% load. In addition, for hardwood operating at
25% load and during the Syracuse cycle the efficiency values obtained from the two tests were
highly variable as evidenced by the range of values shown in Figure 7-1.
7.2 Gas Phase Pollutants
7.2.1 Criteria and Related Gaseous Emissions
The test average emission factors (EFs) for the gaseous pollutants monitored during the study are
provided in Tables 7-2 and 7-3 expressed in terms of engineering and SI units, respectively. These
data are also shown graphically in Figures 7-2 and 7-3 with test-specific results provided in
Appendix C. Due to the leak in the CEM bench discovered during the data analysis, these results
were derived from either the closed-cell FTIR instrument (CO, CH4, NH3, NOx, and SO2) or
analysis of the SUMMA canisters (N2O) after appropriate background subtraction. As described
previously no valid data were available for CO2 or THC.
For the EFs of gaseous nitrogen compounds in Figure 7-3a several trends were observed. In the
case of NH3, grass produced 6-46x higher emissions except at 100% load where the EF for wood
was a factor of 3 higher. For hardwood pellets, the NH3 EF increased by a factor of ~ 5 from 4.9
mg/kg fuel at 25% load to ~ 25.6 mg/kg fuel at 100% load which is not what would be expected.
One would expect that reduced nitrogen should be most prevalent at low load and then decrease
with increasing load. In the case of switchgrass, the trend is the reverse with the emission factor at
41
-------
25% load (226 mg NFb/kg fuel) being a factor of 25 higher than at 100% load (8.94 mg NFb/kg
fuel). The trend for switchgrass is more understandable from combustion theory.
In the case of N2O (partially oxidized nitrogen), the EFs in Figure 7-3a show a factor of 2-6 higher
emissions for grass as compared to wood. The EFs for both fuels also generally decrease from 76.9
and 209 mg/kg fuel for wood and grass at 25% load, respectively, to 20.4 and 42.8 mg/kg fuel at
100%) load. This is counterintuitive since the emission factors of oxidized nitrogen species such as
N2O should increase with increasing combustion temperature indicative of higher load. Finally,
the NOx emissions from both fuels appear to be most prevalent for the Syracuse cycle rather than
at 100%) load. It would be expected that NOx should be highest at the highest load (highest
temperatures) tested. The EFs for grass were ~ 18x higher than wood for 25% load as compared
to the Syracuse cycle which showed the opposite trend of the NOx emissions being about half that
observed for wood. Emissions of nitrogen containing species is complicated by relative
contributions of thermal NO and fuel nitrogen, complex interactions between temperature and
excess oxygen, and large relative differences in the fuel nitrogen contents of the hardwood (0.31%>)
and switchgrass (0.82%>) fuels.
42
-------
Table 7-2. Test Average Gaseous Emissions (Engineering Units)
Test Average Pollutant Emissions3
Pollutantb
25% Load
Syracuse Cycle
100% Load
Reporting Units
Wood
Grass
Wood
Grass
Wood
Grass
lb/MMBTU Input
CO
4.77
4.36
2.39
2.54
1.47
0.299
RPD
64.2
1.25
36.9
13.2
28.3
65.0
ch4
0.253
0.120
0.0394
0.111
0.00971
0.0183
RPD
31.4
10.5
25.2
30.9
57.8
86.6
nh3
0.000633
0.0301
0.00135
0.00789
0.00331
0.00119
RPD
179
47.6
NA
80.4
38.3
25.4
N20
0.00994
0.0278
0.00361
0.0223
0.00263
0.00569
RPD
NA
28.3
40.8
1.78
61.5
3.80
NOx
0.00128
0.0247
0.0265
0.0135
0.00136
RPD
NA
56.6
NA
105
NA
NA
S02
0.0151
0.00327
0.0504
0.0582
RPD
NA
NA
NA
NA
32.4
4.13
lb/MMBTU
CO
6.37
7.01
2.51
3.51
1.43
0.362
Output
RPD
23.6
16.4
11.5
4.10
2.90
53.9
ch4
2.83
1.44
0.333
1.15
0.0718
0.165
RPD
57.0
7.13
50.1
21.9
33.6
76.5
nh3
0.00471
0.357
0.00981
0.0807
0.0247
0.0111
RPD
152
225
NA
72.6
13.2
37.3
N20
0.0686
0.333
0.0293
0.232
0.0194
0.0528
RPD
NA
10.8
15.6
10.9
37.5
8.40
NOx
0.00883
0.307
0.193
0.138
0.0134
RPD
NA
72.4
NA
98.1
NA
NA
S02
0.104
0.0237
0.378
0.541
RPD
NA
NA
NA
NA
7.09
16.3
a Red face type = single test value only; = non-detect; NA = not applicable. Three significant figures
b RPD = relative percent difference in emission factors for the two test runs conducted at each fuel/load
condition
43
-------
Table 7-3. Test Average Gaseous Emissions (SI Units)
Test Average Pollutant Emissions"
Reporting Units
Pollutantb
25% Load
Syracuse Load
100% Load
Wood
Grass
Wood
Grass
Wood
Grass
g/kg fuel
CO
36.9
32.8
18.5
19.1
11.4
2.25
RPD
64.2
1.25
36.9
13.2
28.3
65.0
ch4
1.96
0.900
0.305
0.839
0.0751
0.138
RPD
31.4
10.5
25.2
30.9
57.8
86.6
mg/kg fuel
nh3
4.90
226
10.5
59.4
25.6
8.94
RPD
179
47.6
NA
80.4
38.3
25.4
N20
76.9
209
27.9
168
20.4
42.8
RPD
NA
28.3
40.8
1.78
61.5
3.80
NOx
9.90
186
205
102
10.2
RPD
NA
56.6
NA
105
NA
NA
S02
117
ND
25.3
390
438
RPD
NA
NA
NA
NA
32.4
4.13
g/MJ Input
CO
2.05
1.88
1.03
1.10
0.632
0.129
RPD
64.2
1.25
36.9
13.2
28.3
65.0
ch4
0.109
0.0515
0.0170
0.0480
0.00418
0.00787
RPD
31.4
10.5
25.2
30.9
57.8
86.6
mg/MJ Input
nh3
0.273
12.9
0.582
3.40
1.42
0.512
RPD
179
47.6
NA
80.4
38.3
25.4
N20
4.28
12.0
1.55
9.59
1.13
2.45
RPD
NA
28.3
40.8
1.78
61.5
3.80
NOx
0.551
10.6
11.4
5.83
0.585
RPD
NA
56.6
NA
105
NA
NA
S02
6.51
ND
1.41
21.7
25.0
RPD
NA
NA
NA
NA
32.4
4.13
g/MJ Output
CO
2.74
3.02
1.08
1.51
0.615
0.156
RPD
23.6
16.4
11.5
4.10
2.90
53.9
ch4
1.22
0.621
0.144
0.496
0.0309
0.0712
RPD
56.96
7.13
50.07
21.91
33.6
76.5
mg/MJ Output
nh3
2.03
154
4.22
34.7
10.6
4.79
RPD
152
225
NA
72.6
13.2
37.3
N20
29.5
143
12.6
99.9
8.35
22.7
RPD
NA
10.8
15.6
10.9
37.5
8.40
NOx
3.80
132
82.9
59.2
5.77
RPD
NA
72.4
NA
98.1
NA
NA
44
-------
Reporting Units
Pollutantb
Test Average Pollutant Emissions"
25% Load
Syracuse Load
100% Load
Wood
Grass
Wood
Grass
Wood
Grass
mg/MJ Output
S02
44.9
ND
10.2
163
233
RPD
NA
NA
NA
NA
7.09
16.3
a Red face type = single test value only; = non-detect; NA = not applicable; Three significant figures
b RPD = relative percent difference in emission factors for the two test runs conducted at each fuel/load condition
45
-------
Average Emissions (Ib/MMBTU In)
^ © O © © O Q ©
— sssssas
- NH3 N20 .. NC
Ban = rang* of values
I
>x ::so2
..
Wood Gr
2596 Load
ass Wood Grass Woo
Syracuse Cycle
i Grass
100% Load
Wood
Wood
Wood
100% Load
SNH3 v N20 sNOx «S02
Bart = range of values
<£ 0.50
Wood
Wood
25% Load
3
?.
m
2
2 5
«•»».
xt
I4
I 3
I
I"
s.
(d)
* CO «CH4
Bars = range of valu«i
III
Wood Gras
Syracuse Cycle
Wood Crass
100% Load
Figure 7-2. Test average gaseous emission factors (engineering units) for nitrogen and sulfur
compounds (a) and (c) as well as organic gases (b) and (d)
46
-------
500
45©
J *m
.1.300
| 250
I 2W
gj 150
I 100
<
50
00
j MH3 N2Q ,7 NOx .7 S02
Bars = range of values
!
i. r t
Wood Grass
Wood Grass
Wood Grass
25%lesd
Syracuse Cyde
100% Load
V
Y
»
Si
5
00
a NH3 # N20 m NO* m S02
Bars = range of values
Wood Grass
25%load
Wood Grass
Syracuse Cycle
Wood Grass
100* load
g 250
"as
J, 200
•§s ISO
f 100
£
| 50
«
(e)
= NH3 H20 NOx :: S02
Ssrs = range of vslucs
Wood Grass
25%load
Wood Grass
Syracuse Cyde
§>
!'
(b)
SCO ¦ CK4
Bars = range of values
Wood Grass
25% Load
Wood
Syrati
Grass
e Cyde
Wood Grass
100% Load
1-
*
I 0.5
<
(d)
- CO : CH4
Bars = range of values
l" _ - .
Wood Grass Wood Grass Wood Grass
25% Load Syracuse Cyde lQOYi Load
25% Load
> CO CH4
Bars = range of values
Wood Grass
Syracuse Cyde
Wood Grass
108% Load
Figure 7-3. Test average gaseous emission factors (SI units) for nitrogen and sulfur compounds
(a), (c), and (e) as well as organic gases (b), (d), and (f)
For sulfur and organic pollutants in Figure 7-3a and 3b the emission trends are generally more of
what would be expected. Here, the SO2 generally increases with load (e.g., a factor of more than
3x for hardwood at 100% load as compared to 25% load) with CO and CH4 generally decreasing
with increasing combustion temperature for both fuels typical at higher load. There does not,
47
-------
however, seem to be any significant SO2 measured for switchgrass pellets burned at 25% load and
during the Syracuse cycle.
7.2.2 Volatile Organic and Carbonyl Compounds
Speciated volatile organic compounds (VOCs) were characterized in the hydronic heater emissions
during operation using hardwood and switchgrass fuels and under the three heat load conditions.
VOC samples were taken during the first and second half of each emissions test and were analyzed
following EPA Methods TO-15 and TO-11A for a total of 132 target VOCs. All tests were
performed in duplicate for each fuel/load condition. Emission factors were calculated for speciated
VOCs on a mass to mass fuel burned and mass to heat input and output bases and EFs were
averaged over each test condition. Average speciated total VOC emission factor values are given
in Table 7-4 for each test condition in mass per fuel burned units and per heat input and output.
RPDs listed in the table was calculated for the duplicate runs. Detailed results are provided in
Appendix C.
Table 7-4. Average Total VOC and Carbonyl Emissions.
Fuel Type
Load
Condition
Average Total VOC and Carbonyl Emissions3
Value
mg/kg
fuel
mg/MJ
Input
lb/MMBTU
Input
mg/MJ
Output
lb/MMBTU
Output
Hardwood
25%
Average
2940
163
0.379
241
0.560
RPDb
22.0
22.0
66.0
Syracuse
Cycle
Average
205
11.4
0.0265
12.6
0.0293
RPD
45.0
45.0
69.0
100%
Average
21.4
1.19
0.00276
1.13
0.00262
RPD
61.0
61.0
37.0
Switchgrass
25%
Average
1780
102
0.237
162
0.376
RPD
24.0
24.0
6.00
Syracuse
Cycle
Average
1320
75.2
0.175
104
0.242
RPD
24.0
24.0
15.0
100%
Average
88.7
5.07
0.0118
6.16
0.0143
RPD
24.0
54.0
42.0
a Three significant figures
b RPD = relative percent difference in emission factors for the two test runs conducted at each fuel/load
condition
48
-------
Figure 7-4 summarizes the total VOC emission factor values for each test condition on a mass per
mass fuel burned basis (Figure 7-4a), mass per heat input basis (Figure 7-4b) and mass per heat
output basis (Figure 7-4c) all in SI units. Note that the bars shown in these figures represent the
range in values for the two tests conducted. Similar trends were observed for the total speciated
VOC emission factors by mass per fuel burned and per heat input/output as shown in Figure 7-4.
Generally, and similar to the CO and CH4 data, total speciated VOC emissions were highest for
the 25% heat load conditions using both hardwood and switchgrass fuels and for Syracuse cycle
using switchgrass. As expected, the 100% heat load had the lowest total VOC emissions for both
fuels. Large differences in emission factors for the two fuels were observed under the same heat
load conditions. The trends in VOC emissions between the two fuels were consistent regardless of
whether results were normalized as mass/fuel or mass/heat input or output. The hydronic heater
operating on switchgrass produced between 4-8 times higher total VOC emissions compared to
hardwood pellets for both 100% load and Syracuse cycle. However, for 25% load conditions, total
VOC emissions for hardwood tests were approximately 50-60% higher than the switchgrass tests.
Figure 7-5 shows the individual VOC emission factors for the 16 most abundant VOCs measured
in the hydronic heater emissions for each test condition in mass per heat input. The VOCs with the
highest emission factors measured from the hydronic heater included carbonyls (e.g.
formaldehyde, acetaldehyde, acetone, acrolein), aromatics, and unsaturated hydrocarbons
compared to 4-12% for other test conditions). Many of the major VOCs measured in the hydronic
heater emissions are considered as partial combustion products typically found in biomass burning
and other combustion related emissions. Some of these VOCs are also classified as air toxics and
hazardous air pollutants of concern.
7.2.3 Gaseous PAH Emissions
The background-corrected test average EFs for total speciated gas phase PAHs are shown in Table
7-5 along with the relative percent difference (RPD) between the two tests conducted at each
fuel/load condition. These data are also shown graphically in Figures 7-6 and 7-7 in engineering
and SI units, respectively. Detailed data for each test is shown in Appendix C.
49
-------
3.5
¦ Grass
1 .1
I 2'°
i/i
{A
£ 1-5
LU
U
g 1.0
0.5
0.0
25% Load Syracuse Cycle 100% Load
Heater Load Condition
*•>
3
Wood
J
¦ Grass
£
vi
c
1
in
«/»
i
LU
§ 20
[b>
25% Load
Syracuse Cycle 100% Load
Heater Load Condition
¦ Wood
Q.
*¦>
3
¦ Grass
I
E,
c 150
O
v\
inn
¦
LU
A
q !>U
I
I
_¦ _
25% Load
Syracuse Cycle 100% Load
w
Heater Load Condition
Figure 7-4. Total speciated VOC emission factors in terms of: (a) mass per mass of fuel burned;
(b) mass per heat input; and (c) mass per heat output. Bars represent range of values
50
-------
1.4E+02
-n 1.2E+02
ao
-§• 1.0E+02
8.0E+01
£ 6.0E+01
5 4.0E+01
Q
U1
2.0E+01
<-> 0.0E+00
25% Syracuse
Cycle
Hardwood
100%
25%
=
Syracuse
Cycle
Switchgrass
100%
Formaldehyde
Acetaldehyde
Benzene
Propylene
i Acetone
Acrolein
i Propane
i Vinyl Acetate
I Propanal
I 2-Butanone
I Toluene
11-Butene
Acetonitrile
Naphthalene
Cyclopentane
1,3-Butadiene
Figure 7-5. Speciated VOC emission factors in mass per heat input (mg/MJ) for the 16 most
abundant VOCs averaged over each test condition
51
-------
Table 7-5. Test Average Total Gaseous PAHs
Fuel Type
Load Condition
Average Total Gaseous PAH Emissions"
Valueb
mg/kg
fuel
mg/MJ
Input
lb/MMBTU
Input
mg/MJ
Output
lb/MMBTU
Output
Hardwood
25%
Average
114
6.32
0.0147
6.19
0.0144
RPD
173
173
140
Syracuse Cycle
Average
3.43
0.191
0.000443
0.204
0.000473
RPD
13.2
13.2
12.5
100%
Average
2.66
0.148
0.000343
0.132
0.000307
RPD
153
153
141
Switchgrass
25%
Average
11.7
0.670
0.00156
1.09
0.00253
RPD
17.4
17.4
34.8
Syracuse Cycle
Average
71.7
4.10
0.00951
5.69
0.0132
RPD
18.8
18.8
27.9
100%
Average
22.9
1.31
0.00304
1.56
0.00363
RPD
101
101
91.7
a Three significant figures
b RPD = relative percent difference in emission factors for the two test runs conducted at each fuel/load
condition
As indicated in Figure 7-7a, the PAH EF for hardwood drops by a factor of - 42 with increasing
load from 25% to 100% whereas the opposite is the case for switchgrass. For switchgrass, the PAH
emissions rise with increasing load most notably for the Syracuse cycle which exhibited a factor
of - 6x higher emissions than at 25% load. This trend is counterintuitive. Heavy organics should
be more easily consumed at the higher combustion temperature occurring during the Syracuse
cycle and at 100% load. Obviously, there are processes occurring in the REKA heater at these two
load conditions for grass pellets which are not consistent with the wood fuel combustion. Also,
except at 25% load, the PAH emissions for switchgrass are a factor of - 9-2lx higher than
hardwood indicating a significant fuel effect. It is interesting to note that the independent VOC
and PAH measurements both indicate decreased emissions with increasing load for the hardwood
fuel, but increasing emissions with increasing load for the switchgrass fuel, even though both fuels
exhibit similar volatile and fixed carbon values. While both fuels contain large fractions of volatile
carbon (80 and 73%), it is possible that this volatile matter may be more easily liberated at lower
temperatures for the switchgrass fuel leading to increased VOC and PAH emissions if this carbon
volatilization is enhanced by high (high load) combustion temperatures while the fuel is still in the
process of being fed by auger into the firebox. This combination could result in volatile species
bypassing or being partially oxidized within the flame. Determination of loss-on-ignition via
thermogravimetric analysis may shed light upon any differences in volatile matter evolution.
52
-------
0.030
=" 0.025
CO
<
CL
(a) t
0.020
3, 0.015
to
C
O
2 0.010
0.005
0.000
25% Load
¦ Wood ¦ Grass
Bars = range of values
1J
Syracuse Cycle
Heater Load Condition
100% Load
0.030
Wood ¦ Grass
Bars = range of values
0.025
CO 0.020
0.015
O 0.010
w 0.005
0.000
25% Load Syracuse Cycle 100% Load
Heater Load Condition
Figure 7-6. Test average total emission factors (engineering units) for gaseous PAH compounds
in terms of: (a) mass per heat input; and (b) mass per heat output
53
-------
(a)
2 200
Wood ¦ Grass
Bars = range of values
11. ¦ ¦
25% Load Syracuse Cycle 100% Load
Heater Load Condition
25% Load Syracuse Cycle 100% Load
Heater Load Condition
(c)
¦ Wood ~ Grass
Bars = range of values
3
CL
+¦>
3
i
bO
tfi
C
•S 40 -
i
-------
shown in this figure for hardwood at 25% load, significant quantities of acenapthylene, flourene,
phenanthrene, anthracene, fluoranthene, and pyrene are present. At higher loads, PAHs are evident
only in trace quantities. In the case of switchgrass, napthalene is the most predominate species
present in the emissions for all load conditions, especially during the Syracuse cycle. For grass
during the Syracuse cycle, substantial quantities of acenapthylene, phenanthrene, and chrysene are
also present. The same species are also present in similar proportions at 100% load except at much
lower levels. These compounds are considered as hazardous air pollutants under the Clean Air Act.
Tables detailing the specific emission factor for each compound determined in each test is provided
in Appendix C.
7.0
9 1.0
25% Syracuse Cycle
Hardwood
100%
"
25%
Syracuse Cycle
Switchgrass
100%
Naphthalene
Acenapthylene
Acenaphthene
Flourene
i Phenanthrene
Anthracene
i Fluoranthene
i Pyrene
i Benzo(a)Anthracene
I Chrysene
I Benzo(b)Flouranthene
I Benzo(a)pyrene
lndeno(l,2,3-cd)pyrene
Dibenz(a,h)anthracene
Benzo(g,h,i)perylene
Benzo(k)Flouranthene
Figure 7-8. Test average emission factors for the 16 target PAH compounds determined using
EPA Method 23 for each fuel and load condition.
7.2.4 Dioxin and Furan Emissions
The EFs for total speciated PCDDs and PCDFs in terms of toxic equivalent mass for each fuel and
load condition are shown in Table 7-6 and graphically in Figures 7-9 and 7-10 in terms of
engineering and SI units, respectively. Recall that these results reflect composite emission factors
derived from the samples collected for each of the two tests at each fuel/load condition which were
combined prior to GC/MS analysis. Thus, RPD could not be calculated. Compound-specific results
are shown in Appendix C.
55
-------
Table 7-6. Test Average Total PCDD/PCDF Emission Factors
Fuel Type
Load
Condition
Composite Total PCDD/PCDF Emissions"
ng TEQ/kg
fuel
ng TEQ/MJ
Input
lb TEQ/MMBTU
Input
ng TEQ/MJ
Output
lb TEQ/MMBTU
Output
Hardwood
25%
0.158
0.00880
2.04E-11
0.0136
3.16E-11
Syracuse
Cycle
0.0929
0.00517
1.20E-11
0.00556
1.29E-11
100%
0.320
0.0178
4.14E-11
0.0199
4.63E-11
Switchgrass
25%
0.223
0.0128
2.97E-11
0.0206
4.78E-11
Syracuse
Cycle
0.455
0.0260
6.03E-11
0.0359
8.35E-11
100%
0.105
0.00601
1.40E-11
0.00742
1.72E-11
a Three significant figures
As shown in the Figure 7-10a, except for operation at 100% load, the combustion of switchgrass
pellets produced ~ 1.4 - 5x more PCDD/PCDFs that was the case for hardwood. For wood at 100%
load, the PCCD/PCDFs produced were a factor of - 3x higher than was the case for grass. This
result is counterintuitive since the chlorine content of the switchgrass pellets was about twice that
of hardwood. It might be expected, therefore, that the fuel with the highest chlorine content should
always create the highest PCDD/PCDF emissions which was found not to be the case at 100%
load.
The relative abundances of the PCDD/PCDF compounds observed is also informative. Figure 7-
11 shows the composite emission factor for each compound analyzed in terms of heat input. As
indicated by the figure, the highest emission factors observed for all fuels and loads were 2,3,7,8-
TCDD, 1,2,3,7,8-PeCDD, 2,3,7,8- TCDF, and 2,3,4,7,8-PeCDF the exact amount of which
depended on fuel and load condition. Only relatively minor amounts of the other compounds were
present in the samples collected. Unlike certain VOCs and PAHs, all of the compounds identified
in Figure 7-8 are considered by EPA to be air toxics.
56
-------
25% Load Syracuse Cycle 100% Load
Heater Load Condition
^ 9.0E-11
| 8.0E-11
^ 7.0E-11
a
£ 6.0E-11
=• -p 5.0E-11
J2 3
O
tCO O
C Q.
.2 ^ 4.0E-11
3.0E-11
il]
£ 2.0E-11
U
^ 1.0E-11
G
R 0.0E+00
(b)
¦ Wood
¦ Grass
¦
¦
_
_
_
25% Load Syracuse Cycle 100% Load
Heater Load Condition
Figure 7-9. Composite total emission factors (engineering units) for PCDD/PCDF compounds in
terms of: (a) mass per heat input; and (b) mass per heat output
57
-------
(a) 1 Wood ¦ Grass
2
°-w
b0
co
"J7!
LU
G
¦
¦
Q U.UD
u
¦
¦
25% Load
Syracuse Cycle
Heater Load Condition
100% Load
~ 0.030
(b) ¦ Wood ¦ Grass
^ 0.025
P
g> 0.020
i/i
.2 0.015
IS)
«/»
£
^ 0.010
U_
G
u
•5; 0.005
o
O
U
0.000
L_
25% Load Syracuse Cycle 100% Load
Heater Load Condition
O
(c) _
¦ Wood ¦ Grass
g U.u5j
&D
to
'175
M
LU
¦1
L
U
0-005 ¦
G
II
¦1
25% Load
Syracuse Cycle
Heater Load Condition
100% Load
Figure 7-10. Composite total emission factors (SI units) for PCDD/PCDF compounds in terms of:
(a) mass per mass of fuel burned; (b) mass per heat input; and (c) mass per heat output.
58
-------
0.030
0.025
0.020
= 0.015
2. o.oio
•| 0.005
0.000
25% Load Syracuse 100% Load
Cycle
Hardwood Pellets
25% Load Syracuse 100% Load
Cycle
Switchgrass Pellets
2,3,7,8 - TCDD
l 1,2,3,7,8 - PeCDD
1,2,3,4,7,8 - HxCDD
1.2.3.6.7.8 - HxCDD
l 1,2,3,7,8,9 - HxCDD
I 1,2,3,4,6,7,8 - HpCDD
I 1,2,3,4,6,7,8,9 - OCDD
l 2,3,7,8 - TCDF
I 1,2,3,7,8 - PeCDF
l 2,3,4,7,8 - PeCDF
I 1,2,3,4,7,8 - HxCDF
I 1,2,3,6,7,8 - HxCDF
1.2.3.7.8.9 - HxCDF
2,3,4,6,7,8 - HxCDF
1.2.3.4.6.7.8 - HpCDF
1.2.3.4.7.8.9 - HpCDF
1,2,3,4,6,7,8,9 - OCDF
Figure 7-11. Composite emission factors for the 17 PCDD/PCDF compounds measured using
EPA Method 23 for each fuel and load condition.
59
-------
7.2.5 Total Halide Emissions
HC1 impinger sample emission factor data is presented in Table 7-7. The plots for emission factors
in terms of g of HC1 per kg of fuel, g of HC1 per MJ input, and in terms of g of HC1 per MJ output
are presented in Figure 7-12. As shown in Figure 7-12, the HC1 emissions for wood were about
the same regardless of load condition. Also, one set of samples for the hardwood/Syracuse cycle
tests had concentrations below the field blank value that resulted in chlorides below detection
limits after the blank correction. It can also be observed in all load cases that switchgrass
contributed higher emission rates than the wood pellets with the emissions increasing with load.
Table 7-7. Test Average HC1 Emission Factors
Fuel Type
Load Condition
Value
Average HC1 Emissions
mg/kg
fuel
mg/MJ
Input
lb/MMBTU
Input
mg/MJ
Output
lb/MMBTU
Output
Hardwood
25%
Average
1.53
85.1
0.000198
138
0.000320
RPD
200
23.2
103
Syracuse Cycle
Average
NDC
ND
ND
ND
ND
RPD
ND
ND
ND
100%
Average
4.43
246
0.000572
226
0.000526
RPD
115
115
96.7
Switchgrass
25%
Average
2.70
154
0.000358
245
0.000570
RPD
25.4
25.4
7.9
Syracuse Cycle
Average
2.53
145
0.000335
201
0.000468
RPD
33.6
33.6
42.4
100%
Average
13.5
769
0.00179
913
0.00212
RPD
126
126
119
a Three significant figures
b RPD = relative percent difference in emission factors for the two test runs conducted at each fuel/load
condition
CND = not detected after blank correction
60
-------
25.000
(a)
¦ Wood ¦ Grass
Bars= range of values
-§• 15.000
C
0
'
-------
7.3 Particle Phase Pollutants
7.3.1 Total Particulate Matter Emissions
The average total PM emission factors derived from the modified ASTM 2515 sample trains are
shown in Table 7-8 along with the associated RPD for the two tests conducted at each fuel/load
condition. These data are also shown graphically in Figure 7-13 and 7-14 in terms of both
engineering and SI units, respectively. Also shown in these figures are data from the pellet-fired
hydronic heater tested in the previous study (labeled "Former PBHH") conducted by EPA for
NYSERDA (Kinsey et al., 2012). More detailed results for the individual tests are provided in
Appendix D.
Table 7-8. Test Average Total Particulate Matter Emission Factors
Fuel Type
Load
Condition
Valueb
Average Total Particulate Emissions"
g/kg
fuel
g/MJ
Input
lb/MMBTU
Input
g/MJ
Output
lb/MMBTU
Output
Hardwood
25%
Average
2.91
0.162
0.376
0.249
0.578
RPD
2.02
2.02
82.9
Syracuse
Cycle
Average
0.269
0.0150
0.0348
0.0159
0.0368
RPD
25.1
25.1
0.540
100%
Average
0.401
0.0223
0.0518
0.0219
0.0509
RPD
11.0
11.0
14.6
Switchgrass
25%
Average
1.30
0.0744
0.173
0.118
0.274
RPD
34.9
34.9
17.6
Syracuse
Cycle
Average
0.761
0.0435
0.101
0.0599
0.139
RPD
23.5
23.5
14.4
100%
Average
0.662
0.0378
0.0879
0.0462
0.107
RPD
39.7
39.7
27.8
a Three significant figures
b RPD = relative percent difference in emission factors for the two test runs conducted at each fuel/load
condition
As indicated by the emission factors provided in Figure 7-14a, the total PM emission factor
generally drops by a factor of- 8 for hardwood and a factor of- 1.8 for switchgrass with increasing
load. The only exception is for wood during the Syracuse cycle which produces slightly lower PM
than at 100%. In addition, except at 25% load, the PM emission factors for switchgrass combustion
are a factor of - 2-3x higher than for hardwood. It is interesting that wood combustion at 25% load
produces total PM mass/mass fuel which is more than twice that for grass. Finally, it appears that
the total PM emission factors for the burning of hardwood pellets over the Syracuse cycle were
generally similar to total PM emissions from the European 2-stage unit tested previously for
NYSERDA.
62
-------
-=¦ °-40
¦ Wood ¦ Grass
Bars = range of values
« 0.05 A
< 0.00
0.30 -
0.25 -
o 0.20 -
£ 0.15 -
0.10 -
hi ll
Former PBHH
25% Load Syracuse Cycle 100% Load
Heater Load Condition
5 0.90
O
^ 0.80 -
OQ
2 0.70
^ 0.60 H
£ 0.50
O
0.40
jj 0.30 -
— 0.20
01
00
2 0.10
(b) .
0.00
¦ Wood ¦ Grass
Bars = range of values
Former PBHH
25% Load
Syracuse Cycle
Heater Load Condition
100% Load
Figure 7-13. Total PM mass emission factors (engineering units) in terms of: (a) mass per heat
input; and (b) mass per heat output; Dashed line represents "Former PBHH" which refers to pellet-
fired hydronic heater tested previously (Kinsey et al., 2012)
63
-------
3.00
= 2.50
~ 2.00
c
o
.2 1.50
£ l.oo
Q)
bQ
re
0) 0.50
>
<
0.00
(a)
Former PBHH
*
¦ Wood ¦ Grass
Bars = range of values
25% Load Syracuse Cycle
Heater Load Condition
100% Load
0.18
a. °-16 1
Q.
C
-> 0.14 -
5
3° 0.12 -
{A
i 0.10 -
*17!
i/i _ __
0.08 ¦
LU
^ 0.06 H
u 0.02
<
0.00
(b)
Former PBHH
¦ Wood ¦ Grass
Bars = range of values
25% Load Syracuse Cycle
Heater Load Condition
100% Load
—. °"40
+¦»
3
S- 0.35 -
3
o
0.30 H
3 0.25 -
(/)
C
.2 0.20
I/)
£ 0.15 A
Uj
i o.io -
0.05 ¦
0.00
(C)
¦ Wood ¦ Grass
Bars = range of values
Former PBHH
25% Load Syracuse Cycle
Heater Load Condition
100% Load
Figure 7-14. Total PM emission factors (SI units) in terms of: (a) mass per mass of fuel burned;
(b) mass per heat input; and (c) mass per heat output; Dashed line represents "Former PBHH"
which refers to pellet-fired hydronic heater tested previously (Kinsey et al., 2012)
64
-------
Emission factor data from the Method 2515 filters were compared to similar results for the 25-mm
Teflon filters by simple linear regression analysis. The regression constant of 1 from this analysis
showed that the Teflon filters produced total PM emission factors similar to those determined by
Method 2515 with a correlation constant (r2) of 0.8. Recall that 3 Teflon filter samples of 20
minutes each were collected during each test and as such are more "snapshots" of the emissions
from the hydronic heater rather than overall test averages. It was surprising that the Teflon results
agree so well with those from the Method 2515 trains considering the high variability in the PM
emissions observed by the ELPI instrument as discussed below. Therefore, total PM emission
factors for the Teflon filter sampling will not be reported here but are included in Appendix D.
7.3.2 Particle Number Emissions
Time histories of the total particle number concentrations for each test from the ELPI are shown
in Figure 7-15 for hardwood and 7-16 for switchgrass. These plots represent particles ranging in
size from 30 nm to < 10 |im in aerodynamic diameter (equivalent unit density spheres). It should
also be noted that a soot blow event generally preceded the transition from low fire to high fire
producing large quantities of black carbon soot contributing substantially to the high variability in
emissions observed over the 6-hour test runs.
As illustrated by Figures 7-15 and 7-16, the operational pattern for the duplicate tests were
generally similar except for switchgrass combustion during the Syracuse cycle (Figure 7-16c and
7-16d). Here, the particle concentration time histories are substantially different. Also of note is
the concentration histories of grass combustion at 100% load (Figures 7-16e and 7-16f) which
were quite different from those of wood combustion (Figures 7-15e and 7-15f). In Figures 7-16e
and 7-16f a sawtooth pattern was seen for switchgrass as compared to Figures 7-15e and 7-15f for
hardwood where a more typical high/low fire cycle was evident. It should be noted, however, that
the time scales plotted on the abscissas are not identical for the six plots presented in Figures 7-15
and 7-16.
The average total particle number concentration and standard deviation determined during each
test run from the on-line ELPI analyzer is shown in Figure 7-17 for both fuels. As shown in Figure
7-17, switchgrass generally produced higher particle number concentrations for all load conditions
as reflected in the emission factors provided below.
The average total particle number emission factors calculated from the ELPI data are provided in
Table 7-9 along with the associated RPD for the two tests conducted at each fuel and load
condition. The emissions factor data are also shown graphically in Figure 7-18 and 7-19 in terms
of both engineering and SI units, respectively. Also provided in Figure 7-19a is the emission factor
for the European hydronic heater tested previously for NYSERDA (Kinsey et al., 2012). Test
specific PM number emission factor results are provided in Appendix D.
65
-------
Particle Number Concentration vs. Time-Test4 (10/21/16)
I
iiiiiiiiasssaiiiisisisi
Time of Day
Particle Number Concentration vs. Time-Test 6 (10/25/16)
(b) 25% Load
Wood Fuel
L
P
li
II
rf
Ni 1
F
r
\,i,
f
I
Time of Day
Particle Number Concentration vs. Time-Test 7 (10/26/16)
3.0E+05
2.0E+05
(c)
Syracuse Cycle
Wood Fuel
1
'
r
i
l
1
1
8:51:02
8:59:42
9:08:22
9:17:02
9:25:42
9:34:22
9:43:02
9:51:42
II11I
oooooo
10:52:22
11:01:02
11:09:42
11:18:22
11:27:02
11:35:42
11:44:22
11:53:02
12:01:42
12:10:22
12:19:02
12:27:42
12:36:22
12:45:02
12:53:42
13:02:22
13:11:02
13:28:22
13:37:02
13:45:42
13:54:22
14:03:02
14:11:42
14:20:22
14:29:02
14:37:42
14:46:22
Particle Number Concentration vs. Time-Test 8 (10/27/16)
liliillllilillllllillllll
mn
e of Day
Particle Number Concentration vs. Time-Test 2 (10/19/16)
illllllililiiiililiiiiiil
Particle Number Concentration vs. Time-Test 3 (10/20/16)
!|!!!!!IIll!!lSli!!ll!IIl|l|I!l!IS!
Figure 7-15. Total particle number concentration time histories for hardwood combustion at 25%
load (a and b), during the Syracuse cycle (c and d), and 100% load (e and f)
66
-------
Particle Number Concentration vs. Time-Test 11 (11/03/16)
(a)
1
Tl
iisiiiiifiiiiiiiiiisisiiiiii
o 5 .
Particle Number Concentration vs. Time-Test 12 (11/04/16)
(b)
Ih
In
N
IIIIilSSISISliSIIISIIIISIlIIIIHIIII
Particle Number Concentration vs. Time-Test 13 (11/08/16)
Syracuse Cycle
Switchgrass Fuel
3 0E+")5
7 0E+j5
4.0E+05
o 0EKI5
2 0E+")5
1 0E+j5
of Day
Particle Number Concentration vs. Time—Test 14 (11/09/16)
sssaa5ssaaaaaaaaaaa|ssssaaaisssa5sssssS55
Time of Day
Particle Number Concentration vs. Time-Test 15 (11/10/16)
Q.0E+05
8.0E+05
7.0E+05
; 6.0E+05
5.0E+05
2.0E+05
1.0E+05
O.OE+OO
(e)
100% Load
Switchgrass Fuel
IMi
!ISll3!II!!!!!lill!!lf!!!!!l
Time of Day
Particle Number Concentration vs. Time—Test 17 (11/15/16)
9.0E+05 ¦
8.0E+05 ¦
7.0E+05 ¦
: 6.0E+05 ¦
5.0E+05 ¦
4.0E+05 ¦
3.0E+05 ¦
2.0E+05 •
1.0E+05 •
0.0E+00 •
(f)
100% Load
Switchgrass Fuel
I
Time of Day
Figure 7-16. Total particle number concentration time histories for hardwood combustion at 25%
load (a and b), during the Syracuse cycle (c and d), and 100% load (e and f)
67
-------
Particle Concentration by Operating Condition and Fuel
Syracuse
Syracuse
Wood
Syracuse
Background Corrected
1.00E+00 2.00E+05 4.00E+05 6.00E+05 8.00E+05
Test Average Particle Concentration (particles/cm3)
Figure 7-17. Background corrected test average particle number concentrations for both fuels.
As shown in Figure 7-19a, hardwood combustion generally produced fewer particles per fuel mass
than switchgrass for all test conditions. This decrease varied from a factor of 1.1 at 100% load to
a factor of greater than 1.9 for operation during the Syracuse cycle. Also, for both fuels, the total
PM number emission factor dropped by about a factor of 3 with increasing load, similar to total
PM mass. For hardwood, the number emission factor dropped from 2.07 x 1014 particles/kg fuel at
25% load to 6.78 x 1013 particles/kg fuel at 100%. In the case of switchgrass, the number emission
factor was 2.3 x 1014 particles/kg fuel at 25% load and 7.18 x 1013 particles/kg fuel at 100% load.
Finally, it appears that the REKA unit tested under the Syracuse cycle using hardwood pellets has
a substantially higher average particle number emission factor as compared to the appliance tested
previously for NYSERDA in 2010 also under the Syracuse cycle. (Note that the European unit
was operated using hardwood pellets only.) The European pellet heater had an average emission
factor of 8.5 x 1010 particles/kg fuel whereas the REKA unit produced an emission factor of 7.65
x 1013 particles/kg fuel which is almost a 3 order-of-magnitude increase in particle number.
68
-------
Table 7-9. Test Average Particle Number Emission Factors
Fuel Type
Load
Condition
Valueb
Average Particle Number Emissions3
particles/kg
fuel
particles/MJ
input
particles/MMBTU
input
particles/MJ
output
particles/MMBTU
output
Hardwood
25%
Average
2.07E+14
1.15E+13
1.21E+16
1.47E+13
1.55E+16
RPD
81.3
81.3
4.00
Syracuse Cycle
Average
7.65E+13
4.25E+12
4.48E+15
4.52E+12
4.77E+15
RPD
19.5
19.5
6.19
100%
Average
6.78E+13
3.77E+12
3.97E+15
3.70E+12
3.90E+15
RPD
13.8
13.8
11.7
Switchgrass
25%
Average
2.30E+14
1.31E+13
1.39E+16
2.10E+13
2.22E+16
RPD
13.3
13.3
4.31
Syracuse Cycle
Average
1.48E+14
8.45E+12
8.91E+15
1.18E+13
1.25E+16
RPD
49.3
49.3
57.8
100%
Average
7.18E+13
3.77E+12
4.33E+15
5.03E+12
5.30E+15
RPD
25.3
25.3
13.2
a Three significant figures
b RPD = relative percent difference in emission factors for the two test runs conducted at each fuel/load condition; Three significant figures
69
-------
CO
1.8E+16
Wood ¦ Grass
1.6E+16 -
Bars = range of values
1.4E+16 -
1.2E+16 -
8.0E+15 -
« 1.0E+16 H
v
» 6.0E+15 -
O
E
$ 4.0E+15 H
2.0E+15 H
S
a. 0.0E+00
25% Load Syracuse Cycle 100% Load
Heater Load Condition
2.5E+16
Wood ¦ Grass
Bars = range of values
m 2.0E+16
2
£ 1.5E+16
— 1.0E+16
= 5.0E+15
0.0E+00
25% Load Syracuse Cycle 100% Load
Heater Load Condition
Figure 7-18. Total particle number emission factors (engineering units) in terms of: (a) particles
per heat input; and (b) particles per heat output
70
-------
3.5E+14
3.0E+14 -
2.5E+14 -
2.0E+14 -
1.5E+14
l.OE+14 -
5.0E+13 -
O.OE+OO
(a)
¦ Wood ¦ Grass
Bars = range of values
Former PBH
H = 8.5 x 1030
25% Load Syracuse Cycle 100% Load
Heater Load Condition
l.SE+13
3
1.6E+13
Q.
C
—
1.4E+13
1.2E+13
V
U
'¦£
1.0E+13
re
fi.
8.0E+12
E
4.0E+12
P
2.0E+12
a.
0.0E+00
(b)
1
¦ Wood I Grass
Bars = range of values
ll
25% Load Syracuse Cycle
Heater Load Condition
100% Load
2.5E+13
0.0E+00
2.0E+13 ¦
1.5E+13 -
3
o
-I
s
_aj
u
'E
fU
¦S 1.0E+13 -
IA
C
O
5.0E+12 -
(C)
¦ Wood ¦ Grass
Bars = range of values
25% Load Syracuse Cycle 100% Load
Heater Load Condition
Figure 7-19. Total particle number emission factors (SI units) in terms of: (a) particles per mass
of fuel burned; (b) particles per heat input; and (c) particles per heat output. Also shown in (a) is a
similar factor for the European unit tested previously with hardwood during the Syracuse cycle.
71
-------
7.3.3 Particle Size Distributions
The average differential number particle size distribution for each test conducted is shown in
Figure 7-20 and Figure 7-21 for hardwood and switchgrass, respectively, as determined from the
ELPI data. These data were then converted to a single equivalent lognormal PSD and summary
statistics calculated in the form of the geometric mean particle diameter (GMD) and geometric
standard deviation (GSD) of the distribution. These statistics are likewise shown in Figures 7-20
and 7-21 for each test run.
Looking at Figure 7-20 for hardwood, all the PSDs exhibited at least two modes, a major mode
centered slightly greater or less than ~ 100 nm and a minor mode in the range of ~ 300 nm. For
the 25% load condition, however, a third large particle mode centered around ~ 500 nm was also
observed. The smaller 300 nm mode is likely the result of flame generated soot, and the larger 100
nm mode the result of condensed organic carbon formed post flame. The existence of two modes
can be explained by high supersaturation vapor pressures caused by steep temperature profiles
forcing homogeneous nucleation of new organic carbon particles rather than heterogeneous
condensation of the organic carbon on existing soot particles. In addition, the shape of the PSDs
was generally similar for the two tests conducted at each appliance operating condition except for
25% load. At 25% load, the particles were generally smaller and the distribution narrower during
Test 6 (Figure 7-20b) compared to Test 4 (Figure 7-20a) which is reflected by differences in the
GMD and GSD for the two tests. It should also be noted that a significant number of large particles
> 1 |im were measured by the ELPI for all load conditions contributing to the relatively high mass
emissions observed.
For the switchgrass PSDs shown in Figure 7-21, similar trends were observed as was the case for
hardwood except that the number concentrations were generally higher, especially for the 25%
load condition. Substantially differences in the PSD was also shown for the two tests conducted
using the Syracuse cycle. For Test 14 (Figure 7-2Id), the major mode was centered at ~ 300 nm
rather than slightly less than 100 nm as was observed during Test 13 (Figure 7-21c). This is
reflected by the shape of the PSD as well as the higher GMD for Test 14.
Finally, the PSDs generated for hardwood during the Syracuse cycle (Figures 7-20c and 7-20d)
were compared to those determined for the European pellet burner tested previously (Kinsey et al.,
2012). Figures 7-20c and 7-20d show a much broader bi-modal PSD as compared to the European
unit which was narrower and mono-modal. This would, of course, indicate that combustion
conditions inside the REKA hydronic heater were substantially different as also reflected by the
higher particle number emission factors for the REKA discussed above.
7.3.4 Elemental and Organic Carbon (ECOC)
The averages emission factors for EC and OC determined from the time-integrated quartz filter
sampling are provided in Table 7-10 along with the associated RPD for the two tests at each fuel
and load condition. These data are also shown graphically in Figures 7-22 and 7-23 in terms of
both engineering and SI units, respectively. Test-specific EC and OC emission factor data are
provided in Appendix D.
72
-------
Overall Particle Size Distribution-Test 4 (10/21/16)
1.0E+06
9.0E+05
8.0E+05
F
7.0E+05
o
CD
6.0E+05
O
r
5.0E+05
Q.
4.0E+05
O)
-1
3.0E+05
~o
¦C
2.0E+05
1.0E+05
0.0E+00
(a)
—Average Data
GMD = 144 nm
GSD = 1.92
25% Load
Wood Fuel
0.1 1.0
Dp (|jm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 6 (10/25/16)
1.0E+06
9.0E+05
8.0E+05
F
7.0E+05
o
¦
o
6.0E+05
o
¦c
5.0E+05
a
a.
4.0E+05
U)
3.0E+05
TJ
Z
T3
2.0E+05
1.0E+05
0.0E+00
(b)
—Average Data
GMD = 117 nm
GSD = 1.76
Wood Fuel
0.1 1.0 10.0
Dp (pm Aerodynamic Particle Diameter)
1.0E+06
9.0E+05
8.0E+05
t
7.0E+05
0)
o
6.0E+05
C
ro
Q-
5.0E+05
cT
4.0E+05
o>
o
3.0E+05
z
TJ
2.0E+05
1.0E+05
0.0E+00
Overall Particle Size Distribution-Test 7 (10/26/16)
—Averaa
GMD = 135 nm
GSD = 1.65
Syracuse Cycle
Wood Fuel
0.1 1.0
Dp (pm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 8 (10/27/16)
1.0E+06
9.0E+05
8.0E+05
7.0E+05
6.0E+05
5.0E+05
' 4.0E+05
3.0E+05
2.0E+05
1.0E+05
0.0E+00
(d)
—Average Data
GMD = 145 nm
GSD = 1.65
\ Syracuse Cycle
\ Wood Fuel
0.1 1.0
Dp (jjm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 2 (10/19/16)
1.0E+06
9.0E+05
0E+05
7.0E+05
6.0E+05
5.0E+05
GMD = 172 nm
GSD = 1.67
4.0E+05
3.0E+05
2.0E+05
100% Load
Wood Fuel
1.0E+05
0E+00
0.1 1.0
Dp (pm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 3 (10/20/16)
1.0E+06
9.0E+05
8.0E+05
7.0E+05
6.0E+05
5.0E+05
4.0E+05
3.0E+05
2.0E+05
1.0E+05
0.0E+00
(f)
—Average Data
f \ GMD = 174 nm
/ \GSD = 1.71
\ 100% Load
Wood Fuel
0.1 1.0
Dp (|jm Aerodynamic Particle Diameter)
Figure 7-20. Differential number particle size distributions and summary statistics for tests
burning hardwood pellets
73
-------
Overall Particle Size Distribution-Test 11 (11/03/16)
1.0E+06
9.0E+05
8.0E+05
7.0E+05
6.0E+05
5.0E+05
4.0E+05
3.0E+05
2.0E+05
1.0E+05
0.0E+00
(a)
—Average Data
GMD = 111 nm
GSD = 1.72
25% Load
Switchgrass Fuel
0.1 1.0
D- (|jm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 12 (11/04/16)
1.0E+06
9.0E+05
8.0E+05
7.0E+05
6.0E+05
5.0E+05
4.0E+05
3.0E+05
2.0E+05
1.0E+05
0.0E+00
(b)
—Average Data
A GMD =109 nm
/ \ GSD = 1.67
25 ^ Load
Switchgrass Fuel
0.1 1.0
Dp (pm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 13 (11/08/16)
1.0E+06
9.0E+05
8.0E+05
7.0E+05
6.0E+05
5.0E+05
4.0E+05
3.0E+05
2.0E+05
1.0E+05
0.0E+00
(c)
—Average Data
GMD = 136 nm
W GSD = 1.80
/ \
/ \
/ \
I
Syracuse Cycle
Switchgrass Fuel
0.1 1.0
Dp (Mm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 14 (11/09/16)
1.0E+06
9.0E+05
8.0E+05
7.0E+05
6.0E+05
5.0E+05
4.0E+05
3.0E+05
2.0E+05
1.0E+05
0.0E+00
(d)
—Average Data
GMD =
GSD = 1
60 nm
.77
/ \
/ \
/ \
/ \ ~
Syracuse Cycle
Switchgrass Fuel
0.1 1.0 10.0
Dp (|jm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 15 (11/10/16)
1.0E+06
Average Data
9.0E+05
8.0E+05
7.0E+05
6.0E+05
162 nm
5.0E+05
4.0E+05
3.0E+05
2.0E+05
1.0E+05
100% Load
Switchgrass Fuel
O.OE+OO
0.1 1.0
Dp (pm Aerodynamic Particle Diameter)
Overall Particle Size Distribution-Test 17 (11/15/16)
1.0E+06
Average Data
9.0E+05
8-0E+05
7.0E+05
GMD = 160 nm
GSD = 1.84
6.0E+05
5.0E+05
4.0E+05
3.0E+05
2.0E+05
1.0E+05
100% Load
Switchgrass Fuel
0.0E+00
0.1 1.0
Dp (pm Aerodynamic Particle Diameter)
Figure 7-21. Differential number particle size distributions and summary statistics for tests
burning switchgrass pellets
74
-------
Table 7-10. Test Average ECOC Emission Factors3
Carbon
Type
Fuel Type
Load
Condition
Value
Average Carbon Emissions
mg/kg
fuel
mg/MJ
Input
lb/MMBTU
Input
mg/MJ
Output
lb/MMBTU
Output
Elemental
Hardwood
25%
Average
20.0
1.11
0.00258
1.56
0.00362
RPD
44.6
44.6
44.2
Syracuse
Cycle
Average
10.2
0.568
0.00132
0.648
0.00150
RPD
92.8
92.8
112
100%
Average
90.0
5.00
0.0116
4.66
0.0108
RPD
95.5
95.5
74.6
Switchgrass
25%
Average
11.1
0.636
0.00148
1.01
0.00236
RPD
23.5
23.5
5.91
Syracuse
Cycle
Average
83.8
4.79
0.0111
6.53
0.0152
RPD
60.5
60.5
52.0
100%
Average
292
16.7
0.0387
20.3
0.0472
RPD
44.7
44.7
33.0
Organic
Hardwood
25%
Average
1075
59.8
0.139
83.5
0.194
RPD
45.7
45.7
43.1
Syracuse
Cycle
Average
33.1
1.84
0.00427
1.83
0.00424
RPD
122
122
105
100%
Average
1.78
0.0992
0.000230
0.0858
0.000199
RPD
200
200
200
Switchgrass
25%
Average
572
32.7
0.0759
52.2
0.121
RPD
20.2
20.2
2.58
Syracuse
Cycle
Average
392
22.4
0.0520
30.7
0.0713
RPD
38.4
38.4
29.5
100%
Average
62.8
3.59
0.00833
4.27
0.00993
RPD
115
115
106
a Three significant figures. RPD = relative percent difference in emission factors for the two test runs at each
fuel/load condition.
75
-------
0.18
0.16
^ 0.14
CO
^ 0.12
£
-0.10
c
O
0.08
£
LU
(j 0.06
0.04
0.02
0.00
(a) t
(LI
00
(5
Wood Grass
25% Load
Wood Grass
Syracuse Cycle
¦ Elemental C Organic C
Bars= range of values
I. I.
Wood Grass
100% Load
0.25
3
o 0.20
CO
(b) t
¦ Elemental C ¦ Organic C
Bars = range of values
0.15
c
o
•= 0.10
u
0)
00
2 0.05
>
<
0.00
Wood Grass
25% Load
Wood Grass
Syracuse Cycle
. I
Wood | Grass
100% Load
Figure 7-22. Elemental and organic carbon emission factors for all fuel and load conditions in
terms of: (a) mass per heat input; and (b) mass per heat output (engineering units)
76
-------
o
0)
n? 400
(a)
¦ Elemental C ¦ Organic C
Bars= Range of values
Wood Grass
25% Load
Wood Grass
Syracuse Cycle
. I
Wood Grass
100% Load
£ 60
(b)
¦ Elemental C ~ Organic C
Bars = range of values
Wood Grass
25% Load
I. I
Wood Grass
Syracuse Cycle
Wood Grass
100% Load
(c)
¦ Elemental C ¦ Organic C
Bars= range of values
II
a
En
_i
Wood | Grass
Syracuse Cycle
. I
Wood | Grass
100% Load
Figure 7-23. Elemental and organic carbon emission factors for all fuel and load conditions in
terms of: (a) mass per mass fuel; (b) mass per heat input; and (c) mass per heat output (SI units)
77
-------
Figure 7-23a shows that, except for wood at 25% load, the elemental carbon emissions generally
increase with load for both fuel types whereas the organic carbon emissions decrease with
increasing load. For wood, there was a factor of 4.5 increase in emissions between 25% and 100%
load but for grass it was a 26x increase. The general trend is understandable since significant
amounts of elemental carbon should only be produced at higher combustion temperatures whereas
at lower temperatures, such as at 25% load, elemental carbon is generally minimized and greater
amounts of organic carbon in the form of particle-phase semi-volatile organic compounds should
be present which was indeed observed here. What cannot be explained is that the OC emission
factors for hardwood combustion at 25% load is about twice that for switchgrass whereas just the
opposite was the case at the other two appliance operating conditions where grass produced a
factor of - ll-35x higher emissions. The same trend was also evident for EC. The EC emissions
for grass were generally a factor of ~ 3 to 8x higher than wood except at 25% load where grass
had about half the emissions of wood. It is obvious from these trends that the two fuels burn quite
differently at minimal load as compared to higher, more stable load conditions. This is also
consistent with the VOC and PAH emission factors discussed earlier, and suggest fundamental
differences in the combustion of hardwood and switchgrass fuels perhaps related to fuel volatility.
In addition to the time-integrated filter data, ECOC was also measured using the Sunset Model 4
Semi-Continuous Carbon Analyzer. The test average emission factors from the semi-continuous
instrument are compared to those derived from the manual filter sampling for both EC and OC in
Figure 7-24. As shown by the linear regression results, the data from the two instruments appear
to be well correlated with each other. However, the automated EC analyses provide 23% higher
emission factors and a factor of 2x higher emission factors for OC as compared to the time-
integrated data. One reason for the factor of 2 difference for OC is potentially one very high value
for the semi-continuous instrument at 25% load using hardwood fuel drives the regression constant
significantly upward. If this value was eliminated from the linear regression, the new regression
constant would show the time-integrated analysis as being similar to that observed for EC.
It is not surprising, however, that the semi-continuous emission factors are always higher than the
time-integrated filter results since the duration of each sample collection period is much shorter
and thus more susceptible to the frequent excursions in emissions such as soot blowing which were
characteristic of the REKA heater. These excursions were illustrated previously in Figures 7-15
and 7-16 where emission spikes of a factor of 2x or more in particle number concentration were
oftentimes observed. It is recommended, therefore, that only the time-integrated filter OCEC
emission factor results be used for reporting purposes since the longer sampling periods make them
more representative of the entire test run.
1.3.5 Optical Black Carbon (OBC)
The test average OBC EFs determined by optical absorption using the Aethalometer are provided
in Table 7-11 along with the associated RPD calculated for the two tests at each fuel and load
condition. The EFs are also shown graphically in Figures 7-25 and 7-26 in terms of both
engineering and SI units, respectively. More detailed OBC emission factor data are provided in
Appendix D.
78
-------
0.50
0.45
a
0.40
~
0.35
0.30
-------
Table 7-11. Test Average Optical Black Carbon (BC) Emission Factors
Fuel Type
Load Condition
Valueb
Average Optical BC Emissions3
mg/kg
fuel
mg/MJ
Input
lb/MMBTU
Input
mg/MJ
Output
lb/MMBTU
Output
Hardwood
25%
Average
145
8.1
0.0188
7.2
0.0167
RPD
Syracuse Cycle
Average
29.1
1.62
0.00375
1.66
0.00385
RPD
72.2
72.2
48.8
100%
Average
180
10.0
0.0233
9.66
0.0224
RPD
42.1
42.1
17.1
Switchgrass
25%
Average
62.3
3.56
0.00827
5.71
0.0133
RPD
9.94
9.94
7.72
Syracuse Cycle
Average
162
9.28
0.0215
12.8
0.0297
RPD
21.2
21.2
12.1
100%
Average
420
24.0
0.0557
29.39
0.0683
RPD
25.9
25.9
13.8
a Three significant figures. Numbers shown in red face type are a single test value
b RPD = relative percent difference in emission factors for the two test runs conducted at each load condition
Looking at the data shown in Figure 7-26a, in a trend similar to EC, switchgrass produced a factor
of 2-6x higher OBC as compared to hardwood except at 25% load where wood combustion gave
a factor of ~ 3 higher emissions. Also, the OBC EF drops or stays about the same with increasing
load for hardwood but increases with load for switchgrass. The drop in OBC for hardwood was
between the 25% load condition and the Syracuse cycle where a factor of- 5 decrease in emissions
was observed. In the case of switchgrass, the OBC emissions increased by almost a factor of ~ 7
at 100%) load as compared to 25% load. This is a similar trend to that found for EC as discussed
above.
Another interesting comparison is between the emissions of OBC and EC, as these two parameters
are oftentimes used interchangeably. The OBC emission factors are compared to the manual EC
results for all fuel and load conditions in Figure 7-27 along with a linear regression performed on
the data. As shown by the regression constant, EC represented only about 63% of the OBC
measured. Such results would indicate that, at least for the combustion of hardwood and
switchgrass pellets in this study, the two parameters are not equivalent to each other and thus
should not be used interchangeably.
Few studies on residential wood combustion have investigated black carbon emissions, and only
one, to our knowledge, has investigated black carbon from a pellet stove. Bertrand et al. (2017)
measured black carbon emissions from a 6-kW pellet stove operated on spruce/pine pellets and
observed emission factors of 1.2 g/kg, which is nearly three times larger than the largest black
carbon emissions observed in this study (0.4 g/kg, high load with the grass pellets).
80
-------
0.07
— 0.06
F- 0.05
CO
2 0.04
£
TT 0.03
IA
.2 0.02
E
LLI
y 0.01
m
O
0.00
1. J
Fi
25% Load Syracuse Cycle 100% Load
Heater Load Condition
0.08
i Wood ¦ Grass
0.07 ¦i Bars = range of values
3
o
^ 0.06
CO
S 0.05
2
!q 0.04
to
= 0.03
•| 0.02
LU
y o.oi
CO
O
0.00
25% Load Syracuse Cycle 100% Load
Heater Load Condition
Figure 7-25. Optical black carbon emission factors for each fuel and load condition on the basis
of (a) thermal input and (b) thermal output (engineering units)
81
-------
jUU
"3
3
*+-
¦ Wood ¦ Grass T
Bars = range of values
(a)
CkO
£.
f
o
'S
l/l
J
Uj
u
CO
i. ai
'¦M
Q.
o
25% Load Syracuse Cycle 100% Load
Heater Load Condition
— 30
3
Q.
£ 25
—»
s
|P 20
O 15
u 10
(J
CO
"ro C
u
Q.
o
o
¦ Wood ¦ Grass
Bars = range of values
-------
0.50
0.45
_ 0.40
~aj
= 0.35
oa
zi 0.30
22
U 0.25
"ro
£ 0.20
-------
most prevalent element found for the Syracuse cycle and at 100% load where the EFs for grass
were a factor of 1.8 to 2.5x higher than for wood. No significant levels of Pb were found in the
two fuels.
Table 7-12. Emission Factors for Air Toxic and Other Selected Elements"
Element
Hardwood
Switchgrass
25% Load
Syracuse Cycle
100% Load
25% Load
Syracuse Cycle
100% Load
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
g/kg fuel
Cr
1.85E-03
3.94E-05
1.53E-05
6.05E-04
4.08E-04
2.00E-05
Mil
2.87E-03
2.19E-04
9.51E-05
1.45E-04
5.34E-05
2.89E-05
Pb
6.25E-03
2.55E-04
2.03E-04
1.38E-04
6.26E-04
3.68E-04
S
1.50E-01
1.13E-02
2.03E-02
6.40E-03
1.06E-02
5.80E-03
CI
1.32E-01
9.05E-03
1.49E-02
9.38E-03
3.00E-02
3.81E-02
a Three significant figures; Values represent the average of all valid Teflon filter samples collected for both
tests at each fuel and load condition; Not corrected for ambient background
0.012
0.010
0.008
0.006
0.004
0.002
0.000
I Cr ¦ Mn ¦ Pb
(a)
25% Load Syracuse 100% Load
Cycle
25% Load Syracuse 100% Load
Cycle
Switchgrass
25% Load
Syracuse
100% Load
25% Load
(b)
Syracuse
Cycle
Switchgrass
100% Load
Figure 7-28. Average emissions factors for: (a) toxic metals; and (b) sulfur and chlorine
as derived from XRF analysis of the Teflon filter sample
The same general trend was also seen for S and CI as illustrated in Figure 7-28b. However, the
magnitude of the EFs were much larger than for the 3 toxic metals discussed above. At 25% load
for wood combustion, significant quantities of both S and CI were found with much smaller
amounts determined for grass combustion at the same load condition. For S during the Syracuse
cycle and at 100% load, wood had EFs which were a factor of - 2x higher than grass. This is not
consistent with the fuel analysis which shows that grass had a higher S content than wood. In the
case of CI at these load conditions, grass exhibited a factor of 2-3x higher emissions than wood.
This is consistent with the fuel analysis which showed a similar difference in CI content between
the two fuels as discussed below.
84
-------
7.3.7 Particle Morphology
Particles emitted from wood pellet combustion were mostly soot, i.e. aggregated carbonaceous
primary particles. The primary particles were approximately 30 nm in diameter. These particles
also had some electron dense inclusions that exhibited an EDS spectrum with K, S, and O (Figure
7-29). This composition is consistent with the XRF analysis of the bulk PM from wood pellets
from the Syracuse cycle, which had 26% K and 6% S. These particles are consistent with the
general understanding of biomass combustion in which inorganic compounds in the fuel are
vaporized during combustion and rapidly condense forming alkali salts, which often serve as the
nuclei for other particles (Torvala et al. 2014). The inclusions are likely KSO4 particles that have
aggregated with carbonaceous soot.
S-5500 30.0kV 0.0mm x90.0k BF-STEM 500nm
Figure 7-29. Soot aggregate with K inclusions emitted from wood pellet combustion
The particles from grass pellet combustion also exhibited a high carbon content and the typical
soot structure seen in Figure 7-29. Ftowever, some particles had a more compact morphology and
varying elemental composition. The particle in Figure 7-30a was typical of many of the collected
particles from grass pellet combustion, exhibiting very small primary particles (<10 nm) that are
composed of C, Fe, and O. Figure 7-30b shows an example of an ash particle composed primarily
of P, Ca, O and Mg in lesser amounts. This particle is an aggregate of several large primary
particles (Dp 50 - 100 nm), surrounded by a coating composed mostly of carbon. This
carbonaceous coating unlike the carbonaceous aggregate in Figure 7-29, is likely very low
volatility organic carbon that can persist in the high vacuum in the microscope and the electron
bombardment.
85
-------
(a)
(b)
b
¦ £ A ¦
In ^ 1
M I
S 5500 30 OfcV 0 Omm xl 5C* SF STEM JOOnrfi
Figure 7-30. (a) C, Fe, O particle; and (b) ash particle emitted from grass pellet combustion
7.3.8 PM Semi-Volatile Organic Compounds
Total SVOC Emissions. Speciated SVOCs were determined by GC/MS analysis of extracts from
the 47-mm quartz filters after determination of the total OC content. As such, the total SVOCs
represent a component of the total OC with the remainder being an unresolved complex mixture.
The test average total SVOC EFs are provided in Table 7-13 along with the RPD calculated for
the two tests at each fuel and load condition. Table 7-14 shows the EFs for both the speciated and
unresolved portions of the total OC. The EFs are also shown graphically in Figure 7-31 in SI units.
More detailed results for total SVOCs are provided in Appendix D.
As shown in Figure 7-31, the highest EF determined for total SVOCs was for hardwood at 25%
load followed by switchgrass at 25% load. For the Syracuse cycle and 100% load, wood had little
total OC emissions and thus the speciated SVOCs were very low. In the case of grass, the SVOC
content of the OC emission decreased with increasing load by a factor of up to ~ 18 at 100% load.
86
-------
Table 7-13. Average Total SVOC Emission Factors
Fuel Type
Load
Condition
Valueb
Average Total SVOC Emissions"
mg/kg fuel
mg/MJ
Input
lb/MMBTU
Input
mg/MJ
Output
lb/MMBTU
Output
Hardwood
25%
Average
439
24.4
0.0567
34.4
0.0799
RPD
41.8
41.8
46.9
Syracuse
Cycle
Average
3.70
0.206
0.000478
0.210
0.000487
RPD
82.6
82.6
60.2
100%
Average
0.640
0.0356
0.0000826
0.0000715
0.000000166
RPD
Switchgrass
25%
Average
124
7.10
0.0165
11.4
0.0265
RPD
8.21
8.21
9.45
Syracuse
Cycle
Average
79.4
4.54
0.0105
6.21
0.0144
RPD
44.5
44.5
35.8
100%
Average
9.16
0.524
0.00122
0.624
0.00145
RPD
116.00
116
107
a Three significant figures
b RPD = relative percent difference in emission factors for the two test runs conducted at each load condition
Table 7-14. Average OC Emission Factor Components
Fuel Type
Load
Condition
OC Emission Factor Components3
mg/MJ Input
mg/MJ Output
mg/kg fuel
Speciated
SVOCs
Unresolved
OC
Speciated
SVOCs
Unresolved
OC
Speciated
SVOCs
Unresolved
OC
Hardwood
25%
24.4
35.4
34.4
49.1
439
636
Syracuse Cycle
0.206
1.63
0.210
1.62
3.70
29.4
100%
0.0356
0.0637
0.0000715
0.0858
0.640
1.14
Switchgrass
25%
7.10
25.6
11.4
40.8
124
447
Syracuse Cycle
4.54
17.8
6.21
24.5
79.4
312
100%
0.524
3.06
0.624
3.65
9.16
53.6
a Three significant figures
87
-------
J* 1000
£.
2 800
8 600
£
LLJ
u
o 400
re
£
So 200
: (a
~ Speciated SVOCs
~ Unresolved OC
25% Syracuse 100%
Cycle
Hardwood
25% Syracuse 100%
Cycle
Switch grass
3
Q.
C
(b)
¦ Speciated SVOCs
S 60
bp
£50
O
u
re
J
LU
u
0
1 1 .
01
re 10
«
< o
25%
Syracuse Cycle 100%
Hardwood
25% Syracuse Cycle 100%
Switch grass
o 80
2
*S5 70
E.
T 6°
re
50
.52 40
E
g 20
a>
re 10
:(c)
¦ Speciated SVOCs
¦ Unresolved OC
¦
_
25%
Syracuse Cycle 100%
Hardwood
25%
Syracuse Cycle 100%
Switch grass
Figure 7-31. SVOC and unresolved OC emission factors for all fuel and load conditions in terms
of: (a) mass per mass fuel; (b) mass per heat input; and (c) mass per heat output (SI units)
88
-------
Speciated SVOCs. Results of the laboratory-based thermal-optical measurements were used to
determine organic carbon loadings on the filters. Composite filter OC loads ranged from 27 |ig to
1455 |ig. On average, 569 |ig of filter OC was composited and extracted. For the laboratory-based
thermal-optical instrument, an overall test composite mean and standard deviation for OC and EC
was determined to be 669.5 ±309.0 |ig m"3 and 696.2 ±482.1 |ig m"3, respectively. All filter-based
OC and EC values being reported were corrected for artifact and background. Figure 7-32 shows
the OC-EC ratios for each fuel and load condition. Full load testing produced significantly more
EC in filter PM than either Syracuse or low loads. Although test load composites show no
significant influence on the OC-EC ratio due to pellet type, OC-EC ratios produced from burning
hardwood (HW) and switchgrass (SwG) pellets were significantly different at full- and low-load
conditions. Generally, the low load and Syracuse load cycles produced more OC than EC
irrespective of the fuel type used. It is interesting to note that the OC-EC ratio for switchgrass at
full load is consistently higher that for hardwood.
100
10
0
05
i_
U 1
LU
1
u
O
0.1
0.01
0.001
*
~
~
+F
~
~
~
-6
~
1—1
¦=b
~
cm
~
+
f
+
Pellet type
+ Hardwood
~ SwG
v°
P*
.O*
Figure 7-32. Filter-based OC-EC ratios in PM for individual tests sorted by load conditions. SwG
= switchgrass
89
-------
A total of 1325 individual compound concentrations survived detection limit and background
subtraction criteria; 138 of which are injection replicates. Table 7-15 provides the data population
(TV) and the concentration data range grouped by compound class. We assumed an OC to OM ratio
of 1.7 and determined on average that 12% w/w ±5.8% of the organic matter on the filters was
identified using the GC-MS technique. The anhydrosugar, levoglucosan—a cellulose pyrolysis
product sometimes used as an atmospheric tracer of biomass burning—exhibited the greatest
relative concentration at 7.1% w/w of the organic matter. Additionally, the alkanoic acid, methoxy -
phenol, and PAH compound classes accounted for 2.0% w/w, 1.6% w/w, and 0.89% w/w of the
organic matter in the boiler particle emissions; whereas, the saturated hydrocarbons accounted for
less than 0.4% w/w. In general, oxygenated compounds were detected and quantified at higher
concentrations compared with the hydrocarbons; median concentrations were 436 |ig/g OC and
108 |ig/g OC for these respective chemical categories.
Table 7-15. Data Population and Concentration Range by Compound Class
Min
Max
Compound class
N
ng/g OC
Aliphatic diacid
93
2.685
3638
Alkanoic acid
182
40.75
61161
Anhydrosugars
17
11265
320299
Aromatic acid
117
0.2910
3417
Braiiched-a\ k an e
32
2.336
461.2
Fatty acid
56
10.13
2019
Methoxy-phenol
94
51.55
38283
Normal-alkane
320
3.173
2962
PAH
370
0.4370
16590
Resin acid
44
13.30
4303
Figure 7-33 presents mean concentrations (|ig of compound/g OC) of the individual organic
compounds in the boiler fine PM emissions. Concentration ranges representing all test conditions
are indicated by error bars and varied by greater than 3 orders of magnitude for nearly half of the
compounds. The vast majority of compound concentration means was within 10-1000 |ig/g OC.
Figure 7-34 pools these individual compound concentrations by compound class and shows the
relative enrichment of the methoxy-phenols and organic acids.
90
-------
u
o
00
3
o
c
o
u
c
ro
(D
fl
A
"1 ;r ¦¦
A
P
•"tl
¦«i
Hj
.],
1.
ti
in
I
Compound class
* aliphatic diacid
+ alkanoic acid
X anhydrosugar
~ aromatic acid
A b-alkane
.Z' fatty acid
¦ methoxy-phenol
<3] n-alkane
A PAH
\ resin acid
s I i i
£ !£•.£•!.9
ill!-!
Is-1
IT!
Figure 7-33. Individual mean SVOC concentrations in PM emitted from boiler testing. Concentrations are given in units of g/g OC.
Error bars indicate the concentration range. The y-axis is log scale. Symbols and colors are coded by compound class.
91
-------
1000000
tooooo
_ 10000
o
no
--- 1000
M
u
c
o
100
10
0.1
1
If _
1
ii'j
If
, 1
1 ¦
I 1
i|TTl
'1
IT
;4>
x!>
y// // ^ .
/ r / /
Figure 7-34. Quantile box plots of individual SVOC concentrations pooled by compound class.
Levoglucosan is the anhydrosugar
Individual compounds remained pooled within their respective classes in order to further examine
the effects of pellet type and test load conditions on the organic matter emissions. Figure 7-35
shows the concentration sums differentiated by individual test, compound class, test load
conditions (full, low, and Syracuse), and fuel type (H-hardwood and SwG-switchgrass pellets).
Test pairs were also examined using the Tukey-Kramer honestly significant difference test (a =
0.05). Mean concentrations of resin acids, fatty acids, and methoxy-phenols, showed no significant
difference under the different test load conditions. However, for several cases (aliphatic diacids,
alkanoic acids, //-alkanes, and PAH) the full load conditions produced significantly higher mean
concentrations (|ig/g OC) than both Syracuse and low load conditions. Moreover, compared with
full load testing, low load tests produced significantly higher levoglucosan and lower aromatic
acid concentrations in the organic aerosol particles. The effect of pellet fuel on the emissions also
varied by compound class. SwG pellets produced significantly lower average aliphatic diacid,
alkanoic acid, and methoxy-phenol concentrations. Although, oddly enough, use of SwG showed
significantly higher PAH in the OC fraction of PM. Pellet type had no significant effect on
levoglucosan, aromatic, resin and fatty acids, and b- and //-alkanes concentrations in the particle
emissions. Finally, a one-way analysis using a data pool that considered all measured organic
92
-------
compounds irrespective of class showed no significant difference among pairs of means
representing test load conditions and pellet type.
o
^.1 „ 1
jfci-c e a;c
¦*wdrei_o» ¦x.Tir.c *ce
Compound : aij
mn^- -cfwu
RAJ-t
rwr kc
H j SwG
H i SwG
H SwG H iSwG
H
SwG H SwG
H SwG H
SwG
H SwG
H SwG
1
+
+¦ i
i +4
+
* L-+
+
+ _ i±
+ +
r
m
¦¦
"M-
++
+
-
=
+-t
¦
+4- 1
L—1„,
i
" "2
K 4
+
+
f
^
+
-
-
¦
Mill
111 i
1111 ill 11111111 r
IrrTrr
TTT1
Urn-
i ii i i li ¦ ¦ i I i I i ii
11 i ¦
L
Mill
Figure 7-35. Concentration sums (|ig/gOC) sorted by individual test, compound class, test load
conditions, and fuel type (H- hardwood pellet; SwG - switch grass pellet)
7.3.9 Fuel and Ash
The results of the fuel analyses are provided in Table 7-16. These results represent a composite
sample of fuel added to the heater during testing. Parameters of interest are shown in red face type.
As shown in Table 7-16, there was about 4 times as much ash and about twice the chlorine in the
grass pellets as compared to the wood pellets. The sulfur content of the grass pellets was also
slightly higher than for wood. Finally, trace levels of Hg were found in both fuels but were not in
the PM emissions.
Similar results for the bottom ash are presented in Table 7-17 and graphically in Figure 7-34 which
represent averages of the two tests conducted. With regards to loss-on-ignition (LOI) in Figure 7-
34a, the amount of unburned fuel generally increased with increasing load for both fuels. For wood,
the LOI about doubled from 25% to 100% load and for grass, the LOI increased by a factor of ~ 8
from low to high load. These results are not unexpected since the appearance of the ash samples
as they were collected showed a similar trend. There appeared to be a great deal of unburned
carbon left in the ash samples especially at high load. This is likely the result of the unit's auger
and automatic feeding system moving fuel more quickly through the firebox during high load
conditions. The LOI for grass was about double that for wood except at 25% load where the grass
LOI was slightly more than half that of wood. This is at least partially consistent with the lower
93
-------
thermal efficiency of the unit for switchgrass as compared to wood and may also be related to
higher emissions for switchgrass at high loads.
Table 7-16. Results of Fuel Analyses"
PFI-Certified
Wood
Switchgrass Pellets
Parameter
Pellets
As-Received
Dry
As-Received
Dry
Moisture (Weight %)
4.66
10.41
Volatiles (Weight %)
80.34
84.27
72.62
81.06
Fixed Carbon (Weight
14.00
14.68
12.71
14.18
%)
Ash (Weight %)
1.00
1.05
4.26
4.76
Sulfur (Weight %)
0.09
0.09
0.12
0.13
Carbon (Weight %)
47.91
50.25
42.79
47.76
Hydrogen (Weight %)
5.79
6.08
5.44
6.08
Nitrogen (Weight %)
0.31
0.32
0.82
0.92
Oxygen (Weight %)
40.24
42.21
36.16
40.35
Heat Content (BTU/lb)
7736
8115
7526
8401
Chlorine (ppmw)b
66
69
134
150
Si02 (%)
0.2657
0.2790
2.6834
2.9785
AlO (%)
0.0240
0.0252
0.3233
0.3589
FeO (%)
0.0489
0.0513
0.1546
0.1716
Ti02 (%)
0.0027
0.0028
0.0026
0.0028
P05 (%)
0.0080
0.0084
0.0426
0.0473
CaO (%)
0.3663
0.3846
0.4511
0.5008
MgO (%)
0.0258
0.0271
0.1555
0.1726
NaO (%)
0.0044
0.0046
0.0601
0.0667
KO (%)
0.1176
0.1235
0.2735
0.3036
S03 (%)
0.1391
0.1461
0.0988
0.1097
Hg (ppb)
5.5
12.38
Pb (ppm)
<0.10
<0.10
a Results from the analysis of composite sample of each fuel type. Particularly important parameters are shown
in bold face type
b ppmw = parts per million by weight
94
-------
Table 7-17. Results of Bottom Ash Analyses3
Parameter
Hardwood Pellets
Switchgrass Pellets
25% Load
(weight %)
Syracuse
Cycle
(weight %)
100%
Load
(weight %)
25% Load
(weight %)
Syracuse
Cycle
(weight %)
100%
Load
(weight %)
Si02
21.5
23.2
15.5
56.3
45.8
28.0
AI2O3
1.32
1.76
1.00
6.72
5.35
3.27
Fe2C>3
1.91
1.97
1.66
2.90
2.47
1.47
Ti02
0.199
0.206
0.092
0.494
0.287
0.304
P4O10
0.374
0.391
0.392
0.877
0.772
0.428
CaO
36.6
38.0
30.5
12.1
9.2
4.6
MgO
2.02
2.42
1.94
3.37
2.52
1.42
Na20
0.361
0.383
0.360
1.38
1.12
0.649
K20
18.1
18.0
13.4
6.71
4.80
2.61
S03
0.968
1.05
0.530
1.22
0.539
0.527
LOI
17.1
12.95
34.43
7.70
27.16
57.18
a Average of both test runs at each fuel and load condition. LOI = loss-on-ignition
95
-------
100%
90%
o
+3 80%
'E
¦SP 70%
c
0
1
IX)
i/)
o
c
0)
u
i_
0)
0.
-C
V)
<
60%
50%
40%
30%
20%
10%
0%
1 Wood ¦ Grass
(a)
Bars = range of values
_ 1
¦
II
1. -1
1 II
¦ ¦
25% Load
Syracuse Cycle
100% Load
100%
H 80%
t 70%
,£P 60%
•| 30%
| 20%
- 10%
25% Load Syracuse 100% Load
Cycle
25% Load Syracuse 100% Load
Cycle
Si02 {%)
AI203 {%)
Fe203(%)
Ti02 {%)
P4O10 {%)
CaO (%)
I MgO {%)
Na20 {%)
K20 (%)
l S03(%)
Hardwood Pellets
Switchgrass Pellets
Figure 7-36. Bottom ash analyses in terms of: (a) loss-on-ignition; and (b) metal oxides. Data in
weight percent
Finally, for the metal oxide data shown in Figure 7-34b, oxides of silicon, calcium, and potassium
appear to be most abundant in the samples analyzed. The proportion of all the metal oxides was
generally consistent over all load conditions for wood but silica dominated the ash composition
for grass. This is also expected due to characteristics of the biological feedstock used for
production of the two fuels, and consistent with the fuel analyses.
96
-------
8 Summary and Conclusions
8.1 Effect of Fuel Type
The main objective of the program was to fully characterize the thermal efficiency and emissions
of particulate matter (PM) and other pollutants from a hydronic heater using non-woody pellets as
compared to combustion of premium wood pellets. The results of the individual measurements are
extensive and presented in detail in Section 7 and will not be repeated here. Instead, the difference
between the combustion of switchgrass and hardwood pellets was determined for the various
measured parameters as shown in Table 8-1. In this table, the percent difference is shown assuming
that the non-woody fuel produces higher values for each parameter measured. A negative value in
Table 8-1 indicates that hardwood combustion exhibited the higher value instead of switchgrass.
Table 8-1. Percent Difference Between Switchgrass and Hardwood
Measured Parameter
Grass-Woot
Difference in Average Emissions
25% Load
Syracuse Cycle
100% Load
Thermal efficiency
-25.4%
-30.6%
-9.88%
Total PM
-123%
182%
65.2%
PM number
11.2%
93.3%
5.97%
Elemental carbon
-79.6%
720%
224%
Organic carbon
-88.1%
1080%
3417%
Optical BC
-133%
458%
133%
Total particle-phase SVOCs
-253%
2045%
1333%
CO
-12.5%
3.61%
-404%
ch4
-118%
175%
83.1%
nh3
4520%
468%
-186%
n2o
172%
500%
110%
NOx
1780%
-102%
NA
S02
NA
NA
12.3%
Total VOCs and carbonyls
-65.2%
544%
314%
Total gaseous PAHs
-869
1990
761
Total PCDD/PCDFs
41.2
389
-205
a Based on emission factors in mass/mass fuel burned
b Negative values (indicated in bold face type) show when the measured parameter for wood is greater
than grass
As Table 8-1 shows, hardwood combustion always results in higher thermal efficiency as
compared to switchgrass at all load conditions. It also indicates that at 25% load, hardwood
generally had the higher particle phase emission factors (i.e., 5 out of 6 pollutants) whereas for
97
-------
operation during the Syracuse cycle and at 100% load switchgrass always had the highest
emissions. More specifically, compared to wood at the two high load conditions, switchgrass had:
• 65-182% higher total PM emissions;
• 6-93% higher PM number emissions;
• 224-720% higher EC emissions;
• 341-1080% higher OC emissions;
• 133-458% higher OBC; and
• 1333-2045%) higher total SVOC emissions.
At 25% load, compared to switchgrass, hardwood had:
• 123%) higher total PM emissions;
• 80%) higher EC emissions;
• 88%o higher OC emissions;
• 133%o higher OBC emissions; and
• 253%o higher total PM SVOC emissions
In the case of criteria and related gas-phase contaminants, the results were more mixed. For CO,
wood had higher emissions (13-404%>) at 25% load and 100%> load as compared to the Syracuse
cycle where the emissions for wood and grass were about the same. In the case of CH4, switchgrass
had 83-175%o more emissions for the Syracuse cycle and at 100%> load whereas hardwood
exhibited 118%> higher emissions at 25% load. Switchgrass had 468-4520%) higher NH3 emissions
at 25%o load and for the Syracuse cycle whereas at 100%> load, hardwood had 186%> higher
emissions. The N2O emissions were always highest (110-500%>) for the combustion of switchgrass
at all load conditions whereas for NOx, grass had 1780%) higher emissions at 25% load but for the
Syracuse cycle wood exhibited 102% higher emissions.
Finally, for the speciated gases, the emission factor results were again mixed. In the case of total
VOCs and carbonyls, grass pellet emissions were 314-255%) higher than wood for the Syracuse
cycle and at 100% load but at 25% load wood had 65% higher emissions. For total PAHs, grass
combustion produced 761-1990%) higher emissions for the Syracuse cycle and at 100% load
whereas at 25% load wood had 869%> higher emissions. In the case of the PCDD/PCDF emission
factors, grass produced 41-389%) higher emissions for 25% load and the Syracuse cycle while
wood provided 205% higher emissions at 100% load.
8.2 Comparison to Historical Emissions Data
There were a few studies found in the literature pertaining to either the burning of switchgrass
pellets in hydronic heaters or the emissions from other REKA models (Winther, 2006;
Chandrasekaran et al., 2013a; Chandrasekaran et al., 2013a; Valente et al., 2014). Of these, the
98
-------
most applicable to the current work is the study conducted by Chandrasekaran et al. (2013a).
Although this study did not use a dilution tunnel and ASTM 2515 for collection and sampling of
the emissions and only reports PM-10 rather than total PM, it did use EPA Conditional Test
Method 039 (EPA, 2004) which employs another type of stack gas dilution system. Of the six units
tested in the Chandrasekaran et al. study, Appliance D appears to be the most similar to the
appliance described here and thus will be used for comparison purposes.
Figure 8-1 compares the emissions of total PM, OC, and CO from the REKA appliance at 25%
and 100% load to Appliance D at "low" and "high" load. Looking at Figure 8-la for total PM, at
low load the REKA appliance produced a factor of 2-12x higher emissions than Appliance D
whereas for high load the two units produced comparable emissions. In the case of OC in Figure
8-lb, the same general trend was observed but here the REKA emitted a factor of 22-260 higher
OC emissions at low load as compared to Appliance D. Finally, Figure 8-lc shows that the REKA
produced substantially higher CO while burning wood pellets at both low and high load as
compared to Appliance D. For grass combustion at low load, the CO from both units were
comparable whereas at high load the REKA produced a factor of ~ 4 lower emissions.
From the above comparison, it appears that for some fuel and load conditions the emissions from
the two appliances were similar whereas for other fuel/load combinations the emissions can be
substantially different. There are several possible explanations for these results including
differences in fuel composition and sampling strategy as well as the lack of definition for high and
low load in the Chandrasekaran et al. study as compared to the current work. It should also be
noted that a direct comparison of the current results to those from the previous NYSERDA study
(Kinsey et al., 2012) described in Section 7 showed good agreement for hardwood combustion
while operating under the Syracuse cycle, with the exception of particles per kilograms fuel which
differed by several orders of magnitude. Based on this analysis, it was concluded that the results
presented here are generally comparable to the emissions from units of a similar type tested by
others.
99
-------
REKA Appliance
D*
(a)
REKA Appliance
D
Low Load High Load
Wood Pellets
REKA Appliance
D
REKA Appliance
D
Low Load High Load
Switchgrass Pellets
70
60
50
40
30
20
10
0
(b)
REKA Appliance
D
Low Load
REKA Appliance
D
High Load
Wood Pellets
REKA Appliance
D
Low Load
REKA Appliance
D
High Load
Switchgrass Pellets
500
2000 -
S1500 ¦
(C)
REKA Appliance
D
Low Load
REKA
Appliance
D
High Load
Wood Pellets
REKA Appliance
D
Low Load
REKA Appliance
D
High Load
Switchgrass Pellets
Figure 8-1. Comparison of REKA to Appliance D from the Chandrasekaran et al. study (2013a)
for all fuel and load conditions for: (a) total PM; (b) OC; and (c) CO.
100
-------
8.3 Conclusions
Based on the results presented in Section 7, the following conclusions were reached:
1. The combustion of hardwood exhibited the highest thermal efficiency at all load
conditions. Of the three loads tested with hardwood, the highest efficiency was for
operation during the Syracuse cycle which was the most indicative of "real world"
conditions.
2. With respect to reduced and oxidized carbonaceous gases, the same general trend was
observed as was the case for the particle-phase constituents namely, hardwood produced
higher emissions at 25% load and switchgrass had the highest emissions during the
Syracuse cycle and at 100% load.
3. For reduced and oxidized nitrogen compounds, grass combustion generally had the highest
gaseous emissions as compared to hardwood.
4. For the speciated gas-phase VOCs, carbonyls, PAHs, and total PCDD/PCDFs, the
combustion of hardwood generally produced the highest emissions at 25% load whereas
during the Syracuse cycle and 100% load, switchgrass exhibited the higher emissions.
5. For particle-phase air pollutants, the combustion of hardwood produced the highest
emissions at 25% load whereas during the Syracuse cycle and 100% load switchgrass
exhibited the higher emissions.
6. Comparing the current results to historical data for similar hydronic heaters, the data
reported here were at least generally comparable to the unit tested for NYSERDA in 2010
using hardwood pellets during the Syracuse cycle and to a similar appliance tested by
Chandrasekaran et al. (2013a) burning both fuel types depending on load.
7. The lack of U. S. technical support severely hampered operation and testing of the REKA
appliance evaluated in this study.
101
-------
9 References and Supporting Documentation
9.1 References
Bertrand, A.; Stefenelli, G.; Bruns, E. A.; Pieber, S. M.; Temine-Rousel, B.; Slowik, J. G.; Prevot,
A. S. H.; Wortham, H.; El Haddad, I.; Marchand, N. Primary Emissions and Secondary Aerosol
Production Potential from Woodstoves for Residential Heating: Influence of the Stove Technology
and Combustion Efficiency. Atmospheric Environment 2017, 169, 65-79.
Cassinelli, M. E.; O'Connor, P. F., NIOSH Method 5040. In NIOSH Manual of Analytical Methods
(NMAM), 4th, 2nd supplement ed.; Cassinelli, M. E.; O'Connor, P. F., Eds.; Vol. Supplement to
DHHS (NIOSH) Publication No. 94-113, 1998.
Chandrasekaran, S. R.; Hopke, P. K.; Newtown, M.; Hurlbut, A. Residential-Scale Biomass Boiler
Emissions and Efficiency Characterization for Several Fuels. Energy & Fuels 2013a, 27,
4840-4849.
Chandrasekaran, S. R.; Hopke, P. K.; Hurlbut, A; Newtown, M. Characterization of Emissions
from Grass Pellet Combustion. Energy & Fuels 2013b, 27, 5298-5306.
Fournel, S.; Palacios, J. H.; Morissette, R.; Villeneuve, J.; Godbout, S.; Heitz, M. Particulate
Concentrations during On-Farm Combustion of Energy Crops of Different Shapes and Harvest
Seasons. Atmospheric Environment 2015, 104, 50-58.
Hays, M. D.; Preston, W.; George, B. J.; Schmid, J.; Baldauf, R.; Snow, R; Robinson, J. R.; Long,
T.; Faircloth, J., Carbonaceous Aerosols Emitted from Light-Duty Vehicles Operating on Gasoline
and Ethanol Fuel Blends. Environmental Science & Technology 2013, 47, (24), 14502-14509.
Hays, M. D.; Gullett, B.; King, C.; Robinson, J.; Preston, W.; Touati, A., Characterization of
Carbonaceous Aerosols Emitted from Outdoor Wood Boilers. Energy & Fuels 2011, 25, (12),
5632-5638.
Hays, M. D.; Geron, C. D.; Linna, K. J.; Smith, N. D.; Schauer, J. J., Speciation of Gas-Phase and
Fine Particle Emissions from Burning of Foliar Fuels. Environmental Science and Technology
2002, 36, (11), 2281-2295.
Kinsey, J. S.; Touati, A.; Yelverton, T. L. B.; Aurell, J.; Cho, S.-H.; Linak, W. P.; Gullett, B. K.,
Emissions Characterization of Residential Wood-Fired Hydronic Heater Technologies.
Atmospheric Environment 2012, 63, 239-249.
Lyraanen, J.; Jokiniemi, J.; Kauppinen, E. I.; Backman, U.; Vesala, H. Comparison of Different
Dilution Methods for Measuring Diesel Particle Emissions. Aerosol Science and Technology 2004,
38, 12-23.
102
-------
Torvela, T.; Tissari, J.; Sippula, O.; Kaivosoja, T.; Leskinen, J.; Viren, A.; Lahde, A.; Jokiniemi,
J. Effect of Wood Combustion Conditions on the Morphology of Freshly Emitted Fine Particles,
Atmospheric Environment 2014, 87, 65-76.
Valente, M.; Trouve, G.; Schonnenbeck, C.; Brilhac, J-F. Emission Factors of Gas and Particulate
Matter During the Energy Recovery of Grape Marc in a Domestic Boiler. Proceedings of the First
International Conference on Atmospheric Dust, ProScience 1 2014, 219-224, (DOI:
10.14644/dust.2014.036).
Vicente, E. D.; Duarte, M. A.; Tarelho, L. A. C.; Nunes, T. F.; Amato, F.; Querol, X.; Colombi,
C.; Gianelle, V.; Alves, C. A. Particulate and Gaseous Emissions from the Combustion of Different
Biofuels in a Pellet Stove. Atmospheric Environment 2015, 120, 15-27.
Virkkula, A.; Makela, T.; Hillamo, R.; Yli-Tuomi, T.; Hirsikko, A.; Hameri, K.; Koponen, I. K.
A Simple Procedure for Correcting Loading Effects of Aethalometer Data. Journal of the Air &
Waste Management Association 2007, 57, 1214-1222.
Winther, K., EN 303 Test Report, REKA HKRST-FSK 60, Report No. 300-ELAB-l 132, Danish
Technological Institute, Arhus C, DEMARK, September 25, 2006.
9.2 EPA Test Methods
U.S. Environmental Protection Agency, EPA Test Method 0010, Modified Method 5 Sampling
Train, Test Methods for Evaluating Solid Wastes, Volume II; SW-846 (NTIS PB88-239223);
Washington, DC, September 1986.
U.S. Environmental Protection Agency, EPA Test Method 1 A, Sample and Velocity Traverses for
Stationary Sources with Small Stacks or Ducts. Code of Federal Regulations, Part 60, Title 40,
Appendix A, 1996. http://www.epa.gov/ttnemc01/promgate/m-01a.pdf. (accessed 3/30/2016).
U.S. Environmental Protection Agency, EPA Test Method 2C, Determination of Stack Gas
Velocity and Volumetric Flow Rate in Small Stacks and Ducts (Standard Pitot Tube). Code of
Federal Regulations, Part 60, Title 40, Appendix A, 1996.
http://www.epa.gov/ttnemc01/promgate/m-02c.pdf. (accessed 3/30/2016).
U.S. Environmental Protection Agency, EPA Test Method 3A, Determination of Oxygen and
Carbon Dioxide Concentration in Emissions from Stationary Sources (Instrument Analyzer
Procedure). Washington, DC, Code of Federal Regulations, Title 40, Part 60, Appendix A, 1989.
U.S. Environmental Protection Agency, EPA Test Method 3C, Determination of Carbon Dioxide,
Methane, Nitrogen, and Oxygen from Stationary Sources. Washington, DC, Code of Federal
Regulations, Title 40, Part 60, Appendix A, 1989.
103
-------
U.S. Environmental Protection Agency, EPA Test Method 5, Determination of Particulate
Emissions from Stationary Sources. Washington, DC, Code of Federal Regulations, Title 40, Part
60, Appendix A, 1996.
U.S. Environmental Protection Agency, EPA Test Method 6C, Determination of Sulfur Dioxide
Emissions from Stationary Sources (Instrument Analyzer Procedure), Washington, DC, Code of
Federal Regulations, Title 40, Part 60, Appendix A, 1995.
U.S. Environmental Protection Agency, EPA Test Method 7E, Determination of Nitrogen Oxides
Emissions from Stationary Sources (Instrument Analyzer Procedure). Washington, DC, Code of
Federal Regulations, Title 40, Part 60, Appendix A, 1990.
U.S. Environmental Protection Agency, EPA Test Method 10B, Determination of Carbon
Monoxide Emissions from Stationary Sources (Instrument Analyzer Procedure). Washington, DC,
Code of Federal Regulations, Title 40, Part 60, Appendix A, 1996.
U.S. Environmental Protection Agency, EPA Test Method 23, Determination of Poly chlorinated
Dibenzo-p-dioxins and Polychlorinated Dibenzofurans from Stationary Sources. Washington, DC,
Code of Federal Regulations, Title 40, Part 60, Appendix A, 1995.
http://www.epa.gov/ttn/emc/promgate/m-23.pdf. (accessed 3/30/2016).
U.S. Environmental Protection Agency, EPA Test Method 25A, Determination of Total Gaseous
Organic Concentration Using Flame Ionization Analyzer. Washington, DC, Code of Federal
Regulations, Title 40, Part 60, Appendix A, 1996.
U.S. Environmental Protection Agency, EPA Test Method 26A, Determination of Hydrogen
Halide and Halogen Emissions from Stationary Sources Non-Isokinetic Method, Code of Federal
Regulations, Part 60, Title 40, Appendix A, 2016.
U.S. Environmental Protection Agency, EPA Test Method 9056A Determination of Inorganic
Anions by Ion Chromatography. Test Methods for Evaluating Solid Wastes, Volume II, SW-846,
Washington, DC, 2007.
U.S. Environmental Protection Agency, EPA Test Method 8270D, Semivolatile Organic
Compounds by Gas Chromatography/Mass Spectrometry (GC/MS). Test Methods for Evaluating
Solid Wastes, Volume II, SW-846, Washington, DC, 2007.
U.S. Environmental Protection Agency, EPA Compendium Method 10-3.3: Determination of
Metals in Ambient Particulate Matter Using X-Ray Fluorescence (XRF) Spectroscopy.
Compendium of Methods for the Determination of Inorganic Compounds in Ambient Air.
Cincinnati, OH, EPA/625/R-96/010A, January 1999.
104
-------
U.S. Environmental Protection Agency, EPA Compendium Method TO-11A, Determination of
Formaldehyde in Ambient Air Using Adsorbent Cartridge Followed by High Performance Liquid
Chromatography (HPLC) [Active Sampling Methodology], Compendium of Methods for the
Determination of Toxic Organic Compounds in Ambient Air, EPA/625/R-96/010b, January 1999.
http://www.epa.gov/ttn/amtic/files/ambient/airtox/to-l lar.pdf. (accessed 3/30/2016).
U.S. Environmental Protection Agency, EPA Compendium Method TO-15, Determination of
Volatile Organic Compounds (VOCs) in Air Collected in Specially-Prepared Canisters and
Analyzed by Gas Chromatography/Mass Spectrometry (GC/MS). Compendium of Methods for
the Determination of Toxic Organic Compounds in Ambient Air, EPA/625/R-96/010b, January
1999.
U.S. Environmental Protection Agency, Test Methods for Evaluating Solid Waste (SW-846)
Physical/Chemical Methods; Office of Solid Waste; United States Environmental Protection
Agency: Washington, DC, 2014.
U.S. Environmental Protection Agency, Measurement of PM2.5 and PM10 Emissions y Dilution
Sampling (Constant Sampling Rate Procedures), Revised 7/07, Conditional Test Method (CTM)
039, Research Triangle Park, NC, July 2004.
9.3 ASTM, NIOSH, and Other Methods
ASTM International, Method D3172-13, Standard Practice for Proximate Analysis of Coal and
Coke. West Conshohocken, PA, 2013.
ASTM D3176-15 Standard Practice for Ultimate Analysis of Coal and Coke, ASTM International,
West Conshohocken, PA, 2015.
ASTM International, Method D3682-13, Standard Test Method for Major and Minor Elements in
Combustion Residues from Coal Utilization Processes. West Conshohocken, PA, 2013.
ASTM D3683-11 Standard Test Method for Trace Elements in Coal and Coke Ash by Atomic
Absorption, ASTM International, West Conshohocken, PA, 2011.
ASTM International, Method E2515-11, Standard Test Method for Determination of Particulate
Matter Emissions Collected by a Dilution Tunnel. West Conshohocken, PA, 2011.
http://www.astm.org/Standards/E2515.htm (accessed 3/30/2016.)
ASTM International, Method E2618-13. Standard Test Method for Measurement of Particulate
Emissions and Heating Efficiency of Solid Fuel-Fired Hydronic Heating Appliances. West
Conshohocken, PA, 2013.
105
-------
ASTM International, Method D7348-13A, Standard Test Methods for Loss on Ignition (LOI) of
Solid Combustion Residues. West Conshohocken, PA, 2013.
Canadian Standards Association, Method B415.1-10, Performance Testing of Solid-Fuel Burning
Heating Appliances. Mississauga, Ontario, Canada.
National Institute of Safety and Health, method 5040, Diesel Particulate Matter (as Elemental
Carbon). In NIOSH Manual of Analytical Methods (NMAM). Fourth Edition, pp 2-5, 2003.
http://www.cdc.gov/niosh/docs/2003-154/pdfs/5040f3.pdf. (accessed 3/30/2016).
106
-------
APPENDIX A
Hydronic Heater Operating Parameters
A-l
-------
Table A-l. REKA Unit Operating Parameters
Harthrood
Hardwood
(100% &
(Syracuse
Switchgrass
Switchgrass
Switchgrass
25% Load)*
Cycle)
(25% Load)
(Syracuse Cycle)
(100% Loadt)
Program 1
Program 1
Program 2
Program 2
Program 2
Internal Computer Operating Parameter
Set Points
Set Points
Set Points
Set Points
Set Points
Boiler recirculation water temperature (°C)
167
150
167
150
167
Pause (min)
5
5
20
20
20
Oxy
5
5
0
0
0
Manual/Auto (sec)
0
0
0
0
0
Moving grate on (sec)
1.00
1.00
20
20
20
Moving grate off (min)
1.00
1.00
1.00
1.00
1.00
Ash screw on (sec)
1.0
1.0
10
10
10
Ash screw off (min)
30
30
20
20
20
Fuel step 0 (sec)
0.16
0.16
0.10
0.50
0.50
Fuel step 1 (sec)
0.50
0.50
0.23
0.70
0.70
Fuel step 2 (sec)
1.00
1.00
0.50
1.4
1.4
Fuel step 3 (sec)
1.6
1.6
0.83
2.0
2.0
Fan step 0
8
8
8
8
8
Fan step 1
10
10
10
10
10
Fan step 2
20
20
20
20
20
Fan step 3
80
80
80
80
80
Oxy step 0 (%)
11
11
11
11
11
Oxy step 1 (%)
10
10
10
10
10
Oxy step 2 (%)
9
9
9
9
9
Oxy step 3 (%)
8
8
8
8
8
Start up time (min)
15
15
15
15
15
Start up oxy (%)
0
0
0
0
0
Start up fan
40
40
40
40
40
Fligh oxy level (%)
15
15
15
15
15
Pause fan
5
5
5
5
5
Pause fan time (sec)
10
10
10
10
10
Pause fuel (sec)
1.00
1.00
1.00
1.00
1.00
Pause airing (%)
8
8
8
8
8
Start fuel (sec)
4.00
4.00
4.00
4.00
4.00
After run (sec)
1
1
1
1
1
Boiler temperature hysterisis
3
3
3
3
3
Oxy blower down (%)
0
0
0
0
0
Motor 3 (%)
100
100
100
100
100
RSM
0
0
0
0
0
Step 3 always
0
0
0
0
0
Auto ignition (min)
0
0
0
0
0
VP2
0
0
0
0
0
Draft fan select
0
0
0
0
0
* Second test at 25% load used recirculating water temperature of 150 °C.
A-2
-------
REKA Manual
May 2009
Survey: Display and Menu
Nr
Text in Display
Menu. Description of Adjustments
(Minimum
- Maximum)
1
OPERATION TIME
Shows the number of operation hours
2
DANSK DK
Language options
3
FUEL TYPE
Selection of fuel
4
BOILER TEMP
Selection of boiler temperature (5 - 90)
5
PAUSE TIME
Selection of Interval time. Interval between operation in Interval mode
(3 - 99)
6
OXY %
Higher oxygen for fuels with high moisture content
(addition 0-3 %)
?
MAN/AUT
0 = automatic operation with Lambda probe
1-29 = manual operation without Lambda probe. The figures 1-29
determine the operation interval (f.inst. 20 = 20 seconds between
manual stoking. (Please note: No operation at 30 and 50)
(0 - 99)
8
MAN / OUTPUT
0 = All outlets stopped
1 = Transport screw/silo screw
2 = Overdraught blower (main blower) Blower up
(0-8)
3 = Under draught blower (spare blower) Blower down ( Small I. draft fan R2A 150)
4 = Moving grate
5 = Ash screw
6 = Draft fan
7 = Clearance (Stoker screw / cell sluice)
8 = Ash screw/I. draft fan. (Cleaning of boiler.)
9
MOVING GRATE ON
Selection of operating time of the moving grate
(0 - 52)
10
MOVING GRATE OFF
Selection of interval operating time of the moving grate
(0-255)
11
ASH SCREW ON
Selection of operating time of the ash screw
(0 - 52)
12
ASH SCREW OFF
Selection of interval operating time of the ash screw
(0-255)
13
FUEL STEP 0
Selection of fuel quantity at stage 0
(0.00-52)
14
FUEL STEP 1
Selection of fuel quantity at stage 1
(0.00-52)
15
FUEL STEP 2
Selection of fuel quantity at stage 2
(0.00-52)
16
FUEL STEP 3
Selection of fuel quantity at stage 3
(0.00-52)
17
Stop or press
"START-UP and ~"
18
FAN STEP 0
Selection of blower speed in stage 0 (0 - 80)
19
FAN STEP 1
Selection of blower speed in stage 1 (0 - 80)
20
FAN STEP 2
Selection of blower speed in stage 2 (0-80)
21
FAN STEP 3
Selection of blower speed in stage 3 (0 - 80)
A-3
-------
22
OX % STEP 0
Selection of set point for oxygen content in stage 0
(7 - 14)
23
OXY % STEP 1
Selection of set point for oxygen content in stage 1
(7 - 12)
24
OXY % STEP 2
Selection of set point for oxygen content in stage 2
(7-11)
25
OXY % STEP 3
Selection of set point for oxygen content in stage 3
(7-10)
26
START-UP TIME
Selection of operation time in start-up mode
©
1
27
START-UP OXY
Oxygen content below "high oxygen level" before shift to
operation (f. inst. 3 minus 15 = 12 %)
(0-5)
28
START-UP FAN
Blower speed during start-up
©
00
o
29
HIGH OXY LEVEL
Oxygen level during start-up
(10-21)
30
PAUSE FAN
Blower speed during interval operation
©
00
1
o
31
PAUSE FAN TIME
Time (sec.) the blower has to operate after interval firing
(0-255)
32
PAUSE FUEL
Fuel quantity during interval firing
©
1
to
33
PAUSE AIRING
Level for stop of interval ventilation
(6-21)
34
START FUEL
Fuel quantity by start after interval operation / start-up
(0.00-52)
35
AFTER RUN
Seconds. Empties the stoker screw at stop or interval
(0-255)
36
BOILER TEMP HYS.
Starting point for operation after interval (number of degrees
below the boiler temperature)
(0-16)
37
OXY BLOWER DOWN
Oxygen level for stop of under draught
(0-16)
38
MOTOR 3
This function is used together with after run function. If setting =150 then Stoker Screw and Cell
sluice is running 50% longer than Transport Screw and Silo Screw. If setting = 180 then Stoker
Screw and Celle sluice is running 80% longer then Transport Screw and Silo Screw. If setting = 205
then Stoker Screw and Celle sluice is running 2,5 sec. longer than Transport Screw and Silo Screw.
If setting = 210 then Stoker Screw and Cell sluice is running 5 sec. longer than Transport Screw and
Silo Screw. This will mean that every no. larger than 200 = 0,5 sec.(0 - 255)
39
RSM
Available
40
DISP.
Available
41
STEP 3 ALWAYS
On/off. Always high output or interval (no low output)
-1)
(0
42
AUT IGNITION
0 = no electric ignition /Electric ignition is active at numbers
between 1 and 6) (0 - 6)
2 = 2 minutes between fuel feeding (start portion)
43
VP2
Show time between the fuel steps at the display. Time in seconds when doses screw stands still.(0 -
1)
44
DRAFT FAN SELECT
0 = Normal operation
1 = Pulsates with draft fan
2 = Pulsates 3 seconds every 90 seconds (In interval operation)
3 = 1+2 mixed operation
(0-3)
Factory Adjustments
Nr
Text in Display
Type 1: REKA 10 kW
Type 2: Reka 20 - 30 - 60 kW
1
OPERATION TIME
00000
00000
2
DANSK DK
DK
DK
3
FUEL TYPE
Pellets
Programme 1
Chipwood
Programme 2
Various Fuels
Programme 3
Pellets
Programme 1
Chipwood
Programme 2
Various Fuels
Programme 3
4
BOILER TEMP
75°C
75°C
75°C
75°C
75°C
75°C
5
PAUSE TIME
30 m
20 m
15m
30 m
20 m
15m
6
OXY %
000
000
000
000
000
000
7
MAN/AUT
000
000
000
000
000
000
8
MAN/OUTPUT
000
000
000
000
000
000
9
MOVING GRATE ON
1.00s
1.00s
1.00s
1.00s
1.00s
1.00s
A-4
-------
10
MOVING GRATE OFF
001 m
001 m
001 m
001 m
001 m
001 m
11
ASH SCREW ON
0.50 s
0.50 s
0.50 s
1.00 s
1.00 s
1.00 s
12
ASH SCREW OFF
030 m
030 m
030 m
030 m
030 m
030 m
13
FUEL STEP 0
0.03s
0.10s
0.06s
0.16s- 0,50s
1.20s- 2,00s
0.60s- 1,00s
14
FUEL STEP 1
0.06s
0.30s
0.13s
0.50s- 1,00s
2.40s- 4,00s
1.20s- 2,00s
15
FUEL STEP 2
0.13s
0.60s
0.26s
1.00s- 2,00s
4.80s- 8s
2.40s- 4s
16
FUEL STEP 3
0.26s
1.00s
0.50s
1.60s- 3,00s
7 s- 12s
3,50 s- 6s
17
Stop or press "Start-up and ~ "
18
FAN STEP 0
025
025
025
008
008
008
19
FAN STEP 1
025
025
025
010
010
010
20
FAN STEP 2
040
040
040
020
020
020
21
FAN STEP 3
080
080
080
080
080
080
22
OX % STEP 0
011 %
011 %
011 %
011 %
011 %
011 %
23
OXY % STEP 1
010 %
010 %
010 %
010 %
010 %
010 %
2.4
OXY % STEP 2
009 %
009 %
009 %
009 %
009 %
009 %
25
OXY % STEP 3
008 %
008 %
008 %
008 %
008 %
008 %
26
START-UP TIME
015m
015m
015m
015m
015m
015m
27
START-UP OXY
000 %
000 %
000 %
000 %
000 %
000 %
28
START-UP FAN
050
050
050
040
040
040
29
HIGH OXY LEVEL
015%
015%
015%
015%
015%
015%
30
PAUSE FAN
020
020
020
005
005
005
31
PAUSE FAN TIME
010s
010s
010s
010s
010s
010s
32
PAUSE FUEL
1.00s
1.00s
1.00s
1.00s
1.00s
1.00s
33
PAUSE AIRING
008 %
008 %
008 %
008 %
008 %
008 %
34
START FUEL
2.50s
2.50s
2.50s
4.00s
4.00s
4.00s
35
AFTER RUN
QQ1
QQ1
QQ1
QQ1
QQ1
QQ1
36
BOILER TEMP HYS.
003°C
003°C
003°C
003°C
003°C
003°C
37
OXY BLOWER DOWN
QQQ
QQQ
QQQ
QQQ
QQQ
QQQ
38
MOTOR 3
1QQ
1QQ
1QQ
1QQ
1QQ
1QQ
39
RSM
000
000
000
000
000
000
40
DISP.
41
STEP 3 ALWAYS
000
000
000
000
000
000
A-5
-------
42
AUT IGNITION
000
000
000
000
000
000
43
VP2
000
000
000
000
000
000
44
DRAFT FAN SELECT
001
001
001
000
000
000
A-6
-------
APPENDIX B
Labview Temperatures
B-l
-------
PBHH Load - Hardwood/100% Load
-Test Load
¦Target Load
90000
80000
^ 70000
3 60000
CO
T3- 50000
o 40000
¦£ 30000
x 20000
10000
0
8:09
9:21
10:33
11:45
19-Oct-16
12:57
14:09
15:21
90000
80000
^ 70000
3 60000
CO
T3- 50000
o 40000
¦£ 30000
x 20000
10000
0
10:04
PBHH Load - Hardwood/ 100% Load
Test Load Target Load
11:16
12:28
13:40
20-Oct-16
14:52
16:04
17:16
B-2
-------
PBHH Load - Hardwood/25% Load
Test Load
20000
=5. 15000
3
4->
CO
ra 10000
o
i
+-<
5000
0
8:38 9:50 11:02 12:14 13:26 14:38 15:50
21-Oct-16
PBHH Load - Hardwood/25% Load
Test Load
25000
_c 20000
K 15000
¦a
ro
o
i
10000
5000
9:07
10:19
11:31
12:43
13:55
15:07
16:19
25-Oct-16
B-3
-------
PBHH Load - Hardwood/Syracuse Cycle
¦Test Load
¦Planned Load
45000
40000
^ 35000
3 30000
CO
T3- 25000
o 20000
¦£ 15000
x 10000
5000
8:38
9:50
11:02
12:14
26-Oct-16
13:26
14:38
15:50
40000
35000
30000
S 25000
ra 20000
0
1 15000
10000
5000
0
8:
PBHH Load - Hardwood/Syracuse Cycle
Test Load Target Load
A
ji
k
^1
jjT
*****
38 9:50 11:02 12:14 13:26 14:38 15:50
27-Oct-2016
B-4
-------
PBHH Load - Switchgrass/25% Load
Test Load
20000
=5. 15000
3
4->
CO
ra 10000
o
i
+-<
5000
0
8:38 9:50 11:02 12:14 13:26 14:38 15:50
3-Nov-16
PBHH Load - Switchgrass/25% Load
Test Load
20000
=5. 15000
3
4->
CO
ra 10000
o
i
+-<
5000
0
8:24 9:36 10:48 12:00 13:12 14:24 15:36
4-Nov-16
B-5
-------
45000
40000
^ 35000
3 30000
CO
T3- 25000
o 20000
% 15000
x 10000
5000
0
8:
PBHH Load - Switchgrass/Syracuse Cycle
Test Load Target Load
iJW
ftJL
4NP
jj
In
up
39 9:21 10:33 11:45 12:57 14:09 15:21
8-Nov-16
45000
40000
^ 35000
3 30000
CO
T3- 25000
o 20000
% 15000
x 10000
5000
0
8:
PBHH Load - Switchgrass/Syracuse Cycle
Test Load Target Load
^ i iii.ft
jlA
pNr**
A
M
iW_
Jf
39 9:21 10:33 11:45 12:57 14:09 15:21
9-Nov-16
B-6
-------
80000
70000
=5. 60000
S 50000
ra 40000
0
1 30000
20000
10000
0
8:
PBHH Load - Switchgrass/100% Load
Test Load Target Load
39 9:21 10:33 11:45 12:57 14:09 15:21
10-Nov-16
80000
70000
60000
S 50000
ra 40000
0
1 30000
20000
10000
0
8:
PBHH Load - Switchgrass/100% Load
Test Load Target Load
39 9:21 10:33 11:45 12:57 14:09 15:21
15-Nov-16
B-7
-------
APPENDIX C
Gas-Phase Emission Factors
C-l
-------
Table C-l. Test Average Emission Factors for Criteria and Related Gases
Average CO emissions
Average CH4 emissions
Average NH3 emissions
Fuel/Load Condition
Test Date
g/kg fuel
g/MJ Input
g/MJ Output
g/kg fuel
g/MJ Input
g/MJ Output
g/kg fuel
g/MJ Input
g/MJ Output
Hardwood - Full Load
10/19/2016
1.30E+01
7.21E-01
6.24E-01
9.69E-02
5.39E-03
3.61 E-02
3.05E-02
1.70E-03
1.14E-02
Hardwood - Full Load
10/20/2016
9.75E+00
5.43 E-01
6.06E-01
5.34E-02
2.97E-03
2.57E-02
2.07E-02
1.15 E-03
9.95 E-03
Hardwood - Low Load
10/21/2016
2.50E+01
1.39E+00
3.07E+00
1.65E+00
9.19E-02
1.56E+00
5.19 E-04
2.89E-05
4.91E-04
Hardwood - Low Load
10/25/2016
4.87E+01
2.71 E+00
2.42E+00
2.27E+00
1.26E-01
8.70E-01
9.28 E-03
5.16E-04
3.5 6 E-03
Hardwood - Svracusc
10/26/2016
2.19E+01
1.22E+00
1.14E+00
2.66E-01
1.48E-02
1.08E-01
1.05 E-02
5.82E-04
4.22E-03
Hardwood - Svracusc
10/27/2016
1.51 E+01
8.39E-01
1.02E+00
3.43 E-01
1.91 E-02
1.79 E-01
ND
ND
ND
Swilchgrass - Low Load
11/3/2016
3.30E+01
1.89E+00
2.77E+00
9.47E-01
5.42E-02
5.99E-01
2.80E-01
1.60E-02
1.77E-01
Swilchgrass - Low Load
11/4/2016
3.26E+01
1.86E+00
3.27E+00
8.52E-01
4.88E-02
6.43 E-01
1.72 E-01
9.87E-03
1.30E-01
Switchgrass - Syracuse
11/8/2016
2.04E+01
1.17E+00
1.54E+00
9.69E-01
5.54E-02
5.51E-01
8.33E-02
4.76E-03
4.74E-02
Swilchgrass - Syracuse
11/9/2016
1.79E+01
1.02E+00
1.48E+00
7.09E-01
4.06E-02
4.42E-01
3.55E-02
2.03 E-03
2.21 E-02
Switchgrass - Full Load
11/10/2016
1.52E+00
8.70E-02
1.14E-01
7.80E-02
4.46E-03
4.40E-02
1.01 E-02
5.76E-04
5.68E-03
Swilchgrass - Full Load
11/15/2016
2.98E+00
1.71E-01
1.98E-01
1.97E-01
1.13E-02
9.84E-02
7.81E-03
4.47E-04
3.90E-03
Average NOx emissions
(as N02)
Average S02 Emissions
Average N20 Emissions
Fuel/Load Condition
Test Date
g/kg fuel
g/MJ Input
g/MJ Output
g/kg fuel
g/MJ Input
g/MJ Output
g/kg fuel
g/MJ Input
g/MJ Output
Hardwood - Full Load
10/19/2016
ND
ND
ND
4.53 E-01
2.52E-02
1.69E-01
2.66E-02
1.48E-03
9.92E-03
Hardwood - Full Load
10/20/2016
ND
ND
ND
3.27E-01
1.82E-02
1.57E-01
1.41 E-02
7.85E-04
6.79E-03
Hardwood - Low Load
10/21/2016
ND
ND
ND
ND
ND
ND
ND
ND
ND
Hardwood - Low Load
10/25/2016
9.90E-03
5.51E-04
3.80E-03
1.17E-01
6.51E-03
4.49E-02
7.69E-02
4.28E-03
2.95 E-02
Hardwood - Svracusc
10/26/2016
2.05E-01
1.14E-02
8.29E-02
2.53E-02
1.41 E-03
1.02E-02
3.36E-02
1.87E-03
1.36E-02
Hardwood - Svracusc
10/27/2016
ND
ND
ND
ND
ND
ND
2.22E-02
1.24E-03
1.16E-02
Switchgrass - Low Load
11/3/2016
1.33E-01
7.62E-03
8.43 E-02
ND
ND
ND
2.39E-01
1.37E-02
1.51 E-01
Switchgrass - Low Load
11/4/2016
2.38E-01
1.36E-02
1.80E-01
ND
ND
ND
1.80E-01
1.03 E-02
1.36E-01
Switchgrass - Syracuse
11/8/2016
1.55E-01
8.89E-03
8.83E-02
ND
ND
ND
1.66 E-01
9.5 1 E-03
9.45 E-02
Switchgrass - Syracuse
11/9/2016
4.84E-02
2.77E-03
3.02E-02
ND
ND
ND
1.69E-01
9.68E-03
1.05E-01
Switchgrass - Full Load
11/10/2016
1.02E-02
5.85E-04
5.77E-03
4.47E-01
2.56E-02
2.52E-01
4.20E-02
2.40E-03
2.37E-02
Switchgrass - Full Load
11/15/2016
ND
ND
ND
4.29E-01
2.45 E-02
2.14E-01
4.36E-02
2.50E-03
2.18E-02
C-2
-------
Table C-2. Test Average Speciated VOC Emission Factors (Mass/Mass Fuel Burned)3
Compound Analysis
Hardwood Pellets
S wife harass Pelkts
25% Load
Syracuse
ffisXa,
:cn3c
mg
Load
ks
25'% L
mg4
I
:jC%Load
mzU
AVG |
RPD
AVG | RPD
AVG
RPD
AVG |
RPD
AVG | RPD
A\ b | R°D
Pr opvlene
TOO
202
51%
s:3c
ND
OS
2m
ns
23%
0.161
110%
Propane-
TOO
108
47%
-
*4%
ND
84.4
30%
33%
ND
Die W.C4 odfluca om ethane
TO 13
0.542
156%
1 te^2
4 4ac
0.00233
2M%
ND
ND
ND
Isoprop^iAlcohd
TOO
1.14
40%
0.806
;?:%
0.12S
2S%
l.O
8%
0.773
22%
0.053
I-Pentene
TOO
13.2
84%
0332
^%
ND
6.99
22%
4.81
36%
0.00843
Acrytenitrile
TOO
2.85
50%
0.379
5%
0.0828
24%
12.2
30%
9.55
19%
0.532
n-Pentane
TOO
4.2
mi
0.220
36%
ND
6.82
38%
3.97
42%
ND
Isoprene
TOO
9.49
50%
0.566
77%
0.201
6%
6.39
42%
4.26
27%
ND
trans-2~pettfenfi
TOO
6.96
45%
0.22 S
82%
0.0112
49%
6.26
34%
3.60
36%
0.0229
56%
cis-2-perttene
TOO*
3.54
4S%
0.138
62%
0.00466
200%
2.92
28%
1.84
36%
0.01 IS
200%
Tert-Butand
TOO
48.1
34%
1.21
80%
0.103
4^%
12.7
39%
7.99
35%
0.345
6%
JJ-Dictioroethene
TOO
ND
ND
ND
ND
0.0380
200%
ND
Me%teneCHoride
TOO
ND
0.0849
86%
0.0430
30%
0.00296
200%
ND
ND
3-CMoro-l -Propene
TOO
0.919
7m
0.0364
2®}%
ND
ND
ND
ND
llJ-Trk-hkff
TOO
ND
0.0574
"> A-Mt
ZWt'O
0.0573
200%
ND
ND
ND
CarboriDsdfrfe
TOO
ND
ND
0.00460
200%
ND
ND
0.0415
200%
22-Dimeth\ibutane
TOO
m
0.0205
50%
ND
ND
ND
ND
tram- IJ-DicHcfoethene
TOO
ND
ND
ND
KD
ND
ND
Oycfopentane
TOO
43.3
45%
2.24
67%
0.0930
89%
32. S
38%
22.1
16%
0-04 42-
44%
23-Dimeth\ibutane
TOO
3.33
0.206
17%
ND
3.41.
44%
1.59
71%
ND
II-DkHoroethanfi
TOO
ND
ND
ND
ND
ND
ND
TOO
0.22$
200%
ND
ND
0.203
200%
ND
ND
VinyLAcetate
TOO
111
46%
6.03
123%
0.4S6
77.4
10%
45.8
24%
0.0999
200%
2~Me thjipentane
TOO
2.69
38%
0.236
1%
0.0111
2.44
76%
1.22
88%
ND
2-Butanone
TOO
107
42%
3.57
70%
0.0683
81.7
32%
49.1
29%
0.0597
200%
3-Meth^pertane
TOO
0.132
l§3%
0.0699
114%
0.00329
1 JV'O
0.331
54%
0.185
11%
ND
2-CHoropreng
TOO
ND
ND
ND
ND
ND
ND
1-Eexene
TO 15
361
42%
0.161
63%
0.0194
$0%
5-22
31%
3.93
2914
0.0721
43%
eis-I.2-D3C.Woioethenfi
TOO
125
68%
0.0306
200%
ND
0 501
200'%
0.158
200%
ND
Dusaprcpylether
TO 15
ND
ND
ND
ND
ND'
ND
EtteiAcetate
TOO
ND
ND
ND
ND
\D
ND
n-Kexane
TOO
1.27
23%
0.236
121%
0.0354
2.24
43%
45%
ND
C-Ha'tfotra
TOO
0.52
28%
0.0210
70%
O.OOSS5
0.248
6-2%
195%
0.000906
200%
TetrahvAcfuran
TOO
5.90
21%
0.167
54%
ND
4.93
39%
57%
ND
E-thvil ert-B iMEte
TOO
2.0S
27%
0.0478
108%
ND
ND
200%
ND
Meitefcyctope mane
TOO
0.230
101%
0.071
167%
N D
0.142
230%
2§0%
ND
12-Dichloroethane
TOO
nd
0 0086
11%
0.0000062
200%
ND
ND
ND
2.4-Dun ethyiperffane
TOO
0.245
50%
0.0181
55%
ND
ND
ND
ND
1,1. t - Ti jchbr oethane
TOO
ND
ND
ND
ND
ND
ND
Benzene
TOO
124
48%
25.6
25%
6.S4
84,6
13%
102
43%
54.6
56%
CarbonT etracHoride
TOO
0.0251
200%
0.0152
64%
0.007S
0.0218
40%
0.0170
0%
0.00146
200%
Cycfohexang
TOO
0.673
59%
ND
ND
0J09
200%
0.282
47%
ND
2-Metfrvihexane
TOO
6.73
21%
0.2S6
200%
ND
0.0824
61%
0.0658
200%
ND
23-Dm effrvipertfane
TOO
ND
O.OOB66
200%
ND
ND
ND
ND
TertAmyiMelhylElher
TOO
ND
ND1
ND
ND
ND
ND
3-methylhexane
TOO
0.115
200%
ND
0.00308
200%
0.109
200%
0.0307
200%
ND
12-DicMorapropane
TOO
ND
ND
ND
ND
ND
ND
Brcmiodichlaromethane
TOO
ND
ND
ND
ND
ND
ND
14-Dioxane
TOO
ND
ND
ND
ND
ND
ND
C-3
-------
Table C-2 (continued)
Hardwood Pellets
Swtchaass Pellets
25% Load
Syracuse
!Q®%
Load
25%
Load
Syracuse
!:¦;%
Load
male
2,
m.g%2.
m
ms'ka
ms,k
ms;
ks
Ccmpomd
Analysis
AVG 1
RPD
AVG |
RPD
AVG
RPD
AVG
RPD
AVG 1
RPD
AVG
RPD
Trie Her oetfrene
TO! 5
ND
ND
ND
ND
ND
ND
IsQoctane
T015
ND
0.&D937
um%
ND
ND
ND
ND
MethvlNIeftia:r\tete
T015
ND
ND
ND
ND
ND
ND
Heptane
TO! 5
0.915
37%
0.00735
2m%
ND
2.34
31%
1.15
75%
ND
wis-t 3-DtcHQroptopene
T015
ND
ND
ND
ND
ND
ND
4-Methv-2-P entanone
TO 15
0.323
200%
ND
ND
0.672
17%
0.547
44%
ND
Methvk vc lote\ are
TO!5
0.0367
200%
ND
ND
0.128
7 /%
0.0968
m%
ND
tuns-: S-Djciiotopiopene
TO! 5
ND
ND
ND
ND
ND
ND
12 2-Tiis.Hotoetfiane
T015
ND
ND
ND
ND
ND
ND
23 4-Irunetfrvlpentane
TO! 5
ND
0.00675
73%
ND
ND
ND
ND
Toluene
TOl5
56.8
43%
4.36
35%
0.0709
23%
39.3
28%
45.8
47%
0.776
73%
2-\retteiheptanfi
TOI5
0.!52
200%
ND
ND
0202
200%
0.140
200%
ND
2-Ke\ai one
T015
ND
ND
ND
ND
ND
ND
Dibiom oclics orn ethane
T015
ND
ND
ND
ND
ND
ND
3-Meth<.ib?ptane
T015
0.159
20G%
ND
ND
ND
ND
ND
12-Dibtomoethane
TOI5
ND
ND
ND
ND
ND
ND
Octane
T015
0.2&I
200%
ND
ND
0.779
40%
0.380
«6%
ND
letiachkMQethene
TO!5
ND
o.oiso
49%
0.00850
200%
ND
ND
ND
1 I 12-TetiacblQtoethaiie
TO!5
ND
ND
ND
ND
ND
ND
Qiorobemene
T015
ND
ND
0.0415
20%
ND
ND
0.0393
200%
Ethvibetrene
T015
3.3?
38%
0.202
ND
6.05
33%
4.07
30%
ND
ji-Xylene
T013
4.91
21%
0254
ND
4.33
26%
3.71
62%
ND
p-Xvtene
T015
4.14
im
021!
ND
3.06
21%
2.57
37%
ND
Bromcfonn
T015
ND
ND
ND
ND
ND
ND
Sfjieoe
T015
3.42
35%
0..302
28%
ND
4.8
31%
4,39
31%
0.236
111%
l,t 22-T etractioroethane
T015
ND
ND
ND
ND
ND
ND
o-Xylene
TOi5
2.70
33%
0.160
S3%
ND
2M
30%
1.94
57%
ND
Nctnane
TO!5
ND
ND
ND
0.16S
200%
0.132
im%
ND
Cum ene
TO!5
ND
ND
ND
ND
ND
ND
CHarotduenes
TO! 5
ND
ND
ND
ND
ND
ND
n^Piopvlbeiizene
TO 13
0.112
200%
ND
ND
0.390
33%
• "V
58%
ND
m-Etfaitduene
T015
0.484
200%
ND
ND
0.647
48%
.
60%
ND
li 5-Tt anethvlbenzetie
T015
ND
ND
ND
ND
qa
200%
ND
12 4-Timiethvfwn2erifi
TO! 5
1.67
!%
0.0349
200%
ND
1.86
36%
116
45%
ND
Iert-But>iBen;ene
TO!5
ND
ND
ND
0.105
200%
0.0954
200%
ND
I -E thvM -MethvlB en: ene
TO! 5
ND
ND
ND
0.456
200%
ND
ND
o-Etetolufine
TO! 5
0238
200%
ND
ND
0.728
45%
0.470
63%
ND
1^-Dichlcrobenz ene
TO 15
ND
ND
ND
ND
ND
ND
M-DkHorobeitz ene
TO! 3
ND
ND
ND
ND
ND
ND
n-Decane
TO! 5
ND
ND
ND
0.195
200%
0.123
200%
ND
S ec-B u£y!B enzene
TO!5
ND
ND
ND
ND
ND
ND
12J-Trkn ethvlbenz ene
TO! 5
0.2.S
2»%
ND
ND
0.858
41%
0.411
60%
ND
12-DieHorobenz ene
TO!5
ND
ND
ND
ND
ND
ND
Q-Cyme-ne
TO!5
ND
ND
ND
ND
ND
ND
13-EHethvibenzene
T015
ND
ND
ND
0.225
33%
200%
ND
12-Diedi\t3enzene
TO! 5
§277
200%
ND
ND
0.685
58%
50%
ND
n-Bu^iBenzene.
TOI3
0.382
200%
ND
ND
0.5S5
58%
73%
ND
Undecame
TOI5
ND
ND
ND
0.525
33%
200%
ND
12.4-TncMorobenzene
TO! 5
ND
ND
ND
ND
ND
Naphthalene
TO!5
21.0
33%
4.55
44%
1.01
127%
9.25
16%
65%
8.18
31%
Dodecane
T015
0.162
2®%
ND
0.0233
200%
0.918
90%
61%
ND
Hexachlarobutadiene
T015
ND
ND
ND
ND
ND
ND
Formaldehyde
TO! LA
5W
17%
76.2
55%
7.14
31%
232
6%
216
5%
15.8
Acetaldehyde
TO!! A
35&
17%
20.5
§4%
0.735
27%
232
22%
159
2.6%
1.30
Propanal
T01!A
142
3 1%
ND
ND
72.3
27%
42.1
24%
0.288
Crotonaldehyde
TOIIA
23.7
19%
ND
ND
14.00
34%
10.1
51%
ND
Butyraldehyde
TO!! A
40.5
200%
ND
ND
0.139
200%
ND
ND
Benzaldehyde
TOO A
212
200%
3.13
200%
ND
20.3
66%
19.2
4%
0.763
200%
Isovaleraldehyde
TOIIA
4.78
200%
ND
ND
152
22%
10.2
1%
ND
Yaleraldehyde
TO!! A
0.612
200%
ND
ND
0.116
200%
ND
§.00287
200%
o-T dualde hyde
TO!! A
4.56
200%
2.18
200%
ND
ND
2.95
200%
ND
m-Tdualdehv'dfi
TOIIA
ND
ND
ND
ND
ND
ND
p-T dualde hyde
TOIIA
171
200%
ND
ND
ND
ND
ND
Hexaldehyde
TOIIA
34.3
200%
ND
ND
32.4
22%
ND
ND
-J-Dime tfrvibenzaldehyde
TO!! A
ND
ND
ND
ND'
6.68
m%
ND
TOTAL YOCs
2940
22%
203
45%
21.4
61%
1780
24%
1320
24%
SS "
54%
aND=not detected. RPD=relative percent difference.
C-4
-------
Table C-3. Test Average Speciated VOC Emission Factors (Mass/Heat Input)3
Compound Analysis
Hai dv ood v eflets
Switcher ass P eUets
23% Load
m.sMf input
Svta.
smor
use
input
-l?%Load
n:M trput
2:% Load
input
mgMF»pu£
.
maM.
Load
input
AVG |
RPD
A"G 1
RPD
G
PPD
WO
| RPD
AVG |
RPD
AVG
/TD
PrcpriAene
1015
111
U ^
ND
8.89
29%
6.74
23%
0.0092
xim
Propane
1013
5.99
0.274
t4%
ND
4. S3
30%
3,15
33%
ND
Dichioi adtfbot am ethane
TO! 5
0.O3O2
156%
[ J a*
4^
0.0OOS27
200%
0.0269
10%
23%
ND
CMorom ethane
TO! 5
0.490
lM-,8*
ST t
0.0746
28%
0.338
29%
_
22%
0.0506
13%
[sobutane
1013
1 13
31%
0.0679
40" L
0.00503
6m
1.07
3!%
21%
0.00700
37%
Dichks otettailucaroethane
TO! 5
ND
ND'
ND
ND
ND
\'imiCHcnde
TO! 5
ND
ND
ND
ND
200%
0.000655
200%
S-Butene
TO! 3
335
46%
ft. 144
%%
0.0147
.
3.19
31%
31%
0.0614
6%
13-Butadifine
TOI 5
1..5S
37%
0.108
000301
S%
1 0S
34%
r
27%
0.0183
93%
Bute
TO! 5
0.90!
37%
0.0622
0.000285
2 n\
1.03
39%
34%
ND
trans-2-butene
TO! 5
1.26
44%
0.0445
ND
1.05
30%
0 636
34%
0.00131
200%
Bromamelhane
TO! 5
ND
0.000693
0.00209
12%
0.0245
19%
25%
0.00445
0%
cis-2-butene
TO! 5
0.839
39%
0.0317
-
ND
0.7 SO
27%
34%
ND
Ghktfoe thane
TO! 5
ND
0.000331
0.000725
34%
MD
^ D
0.000808
200%
Ethand
TO! 5
0.679
35%
0.0476
0.0281
231/©
0.379
36%
3%
0.0178
27%
VinvJBronide
TO! 5
ND
ND
ND
ND
ND
ND
Acetonitrile
TO! 3
0.690
S!%
0.0752
0.0178
3~%
2.75
33%
1.83
33%
0.0779
15%
Acfdein
TOI5
9.37
39%
0.577
0.0346
51%
4,66
27%
3.82
!%
0.0559
32%
Acetone
TOI 5
12.2
49%
0.574
0.0445
24%
9.39
31%
* 44
18%
0.00523
200%
iso-Pentane
TO! 5
0.93.2
90%
0.066S
0.O032S-
45%
1.05
37%
22%
ND
Trkhkcaffoor omethane
TO! 5
ND
0.000745
* "U
0.000129
200%
ND
ND
ND
IsoproBtttohd
TO! 5
0.0636
40%
0.0448
0.00709
2S%
0.0658
8%
0 1442
22%
0.00303
1-Pentene
TOI 5
0.733
84%
0.OI85
99%
ND
0.399
22%
.
36%
0.000482
A^rvlonitrile
TO! 5
§.139
50%
0.0211
5%
0.00460
24%
0.697
30%
19%
0.0304
rvPenfcane
TO! 3
0.233
38%
00122
36%
ND
0.390
38%
42%
ND
IsojareM
TOI 5
0.528
30%
0.0315
77%
0.0112
6%
0.365
42%
27%
ND
trans-2-penrene
T0I5
§.3 87
43%
0.0127
S2%
0.000625
49%
0.358
34%
215
36%
0.00131
56%
ds-2-perSene
TO! 5
§197
48%
0.00767
62%
0 000259
200%
0.167
28%
36%
0.000675
200%
Tert-Butand
TOI 5
2.67
34%
0.0672
80%
0.00573
49%
0.727
39%
35%
0.0197
6%
1 J-Dkltaoetheiw
TO! 5
ND
\D
ND
ND
'
200%
ND
MeteieneCHcsride
TO! 3
ND
0.OO239
30%
0.000169
2.00%
ND
ND
3-t Itao-L-Piqpene
TOI 5
0.0511
79%
0.00202
ND
ND
ND
ND
] 1 2-Tnchlaro-l 2.2-QtfTuDroethane
TOI 3
ND
0.00319
0.00319
ND
ND
ND
C arbonDtaifide
TO! 3
ND
ND'
0.000256
ND
ND
0.00237
200%
2.2-DimethvtJutane
TO! 5
ND
0.00114
50%
ND
ND
ND
ND
trans-L2-Dicli or oettene
TOI 5
ND
ND
ND
ND
ND
ND
Cydopentane
TOI 3
2.4!
45%
0. !25
€7%
0.00517
§9%
i r
38%
1.27
1.6%
0.00252
44%
2J-Dimethvft>utane
TO! 5
mm
39%
0.0114
17%
ND
o is?
44%
0.090S
71%
ND
] i - Dickto oethane
TO! 3
ND
ND
ND
ND
ND
ND
\ letfoi-t-B utyi-E .thai
TO! 3
§.§127
200%
ND
ND
J 3115
200%
ND
ND
VimiAcetate
TO! 3
6.19
4®%
0.335
123%
0.0270
S5%
11}
10%
2,62
24%
0.0057!
200%
2-\iethv%jen£ane
TOl5
0.15
38%
0.0131
1%
0.000619
200%
J. 139
76%
0.0695
S8%
ND
2-Butanone
TO! 3
¦'1 u-i
42%
0.199
im
0.003S
62%
4.67
32%
2.80
29%
0.00341
2®)%
3-Klatt^^seitaog
TO1! 3
103%
0.00389
114%
0.000!S3
ro%
0.0139
34%
0.0106
11%
ND
2-Chkxroprerw
TO 13
% D
ND
ND
ND
ND
ND
1-Hexene
TO! 5
42%
0.CK)893
63%
0.00108
10%
0.298
31%
0.225
29%
0.00412
43%
cis- 12-Dichbtoefhene
TO! 5
m%
0.0017
2-00%
ND
0.0287
200%
O.00903
200%
ND
Dtisopiopviether
TO! 3
ND
ND
ND
ND
ND
ND
EtteiAcetate
TO! 5
^D
ND
ND
ND
ND
ND
n-He\ ane
TO! 5
23%
0.0131
121%
0.00197
122%
0.12S
43%
0.0702
45%
ND
CHwtfcra
TO! 5
1 CS
28%
0.0011?
0.000492
42%
0.0142
62%
0.0042!
195%
0.0000515
200%
Tetrahydrofuran
TO! 3
2!%
0.0093
ND
0.282
39%
0.167
37%
ND
EthMX ert-B utyEttier
TO! 3
0.116
27%
0.00266
ND
ND
0.0133
200%
ND
Methyk^lopentane
TO! 5
0.0125
101%
0.00395
ND
0.0OSO9
20O%
0.00223
2§0%
ND
12-Dichtaroethane
TO! 5
ND
0.000479
3.45E-07
200%
ND
ND
ND
2,4- Dwi e thylpentane
TO! 5
0.0136
30'%
0.0010!
ND
ND
ND
ND
lrlJ-TrichloroeUianfi
TO! 5
ND
ND
ND
ND
ND
ND
Benzene
TO'! 3
48%
1.43
25%
0.38
96%
4.84
13%
5.S!
43%
3.12
56%
CarbonTe trachloride
TO! 3
200%
0.000845
64%
0.000434
102%
0.00125
40%
0.000971
0%
0.0000832
200%
Cydohexane
TO! 5
39%
ND
ND
0.0177
200%
0.0161
47%
ND
2-Nlethylhexane
TO! 5
21%
0.0159
200%
ND
0.CKM7!
61%
0.00376
200%
ND
23-Dimethv^-erttanfi
TO! 3
D
0.000482
200%
ND
ND
ND
ND
TO! 5
ND
ND
ND
ND
ND
ND
3-methylhexane
TO! 3
0.0O638
200%
ND
0.00017!
200%
0.00620
200%
O.OOI75
200%
ND
12-Dictkroprcpanfi
TO! 5
ND
ND
ND
ND
ND
ND
Br amodchtar amethane
TOI 3
ND
ND
ND
ND
ND
ND
L4-Diaxane
TO! 5
ND
ND
ND
ND
ND
ND
C-5
-------
Table C-3 (continued)
Hai to ood Pellets
Switch® ass Pellets
25% L
oad
Syracuse
vW
c Load
Load
Syracuse
icr
Load
ins*Z input
rmput
mzVJ input
ma MJ *put
msMJ irpi
Compound
Analysis
AVG I
RPD
AVG |
RPD
AVG
I P.PD
A\G
| RPD
AVG \
RPD
AVG
RPD
Tricliaroethene
1015
ND
ND
ND
ND
ND
ND
Isooctane
TO! 5
ND
0.000521
146%
ND
ND
ND
ND
J..&tei!v!ethacr\iate
TO! 5
ND
ND
ND
ND
I
ND
Heptane
TO! 5
0.0509
37%
0.C5W4W
200%
ND
0.134
31%
0.0660
75%
ND
cis- [ 5-Di zldoiopropene
1013
ND
ND
ND
\D
ND
4-Methy-2-P entanone
T015
200%
nd
ND
0.03S4
17%
_
44%
ND
Metfrvbwlohexanfi
TO1! 5
-
200%
ND
ND
0.00730
77%
0.00553
68%
ND
ttans-1.3 -DicMoropropene
TO! 5
>D
ND
ND
ND
\D
ND
I! 2-1 icHoroethane
TO! 5
ND
ND
ND
ND
ND
ND
2 5,4-TimietteipeitoK.
TO! 5
ND
0.000375
73%
ND
ND
ND
ND
Toluene
TO! 5
3.16
43%
0.243
35%
0.00394
23%
3.40
28%
2.62
47%
0.0444
73%
2~\IettefTeptane
TO! 5
0.00847
im%
ND
ND
0.0116
200%
0.00798
2-CK)%
ND
2-Kexanone
TO! 5
ND
ND
ND
ND
ND
ND
Dibiomoctka omethane
TO! 3
ND
ND
ND
ND
ND
ND
3-Ivfettefaptane
TO! 5
0.0088?
200%
nd
ND
ND
ND
ND
! ,2-Ditrom oethane
TO! 3
ND
ND
ND
ND
ND
ND
Octane
TO! 5
0.0113
2W%
ND
ND
0.0445
4-0%
0.0217
66%
ND
TetracHaroethfine
TO! 5
ND
0.00100
49%
0.000473
200%
ND
ND
ND
IJ J2-TetracWciioethane
TO! 5
ND
ND
ND
ND
ND
ND
Cliorobenzene
TO! 5
ND
ND
0.00231
20%
ND
ND
0.00224
200%
Ethylbenzene
TO! 5
0.199
38'%
0.0112
70%
ND
0.346
33%
0.233
50%
ND
m-X\iene
TO! 5
§.273
21%
0.014!
84%
ND
0.247
26%
0.212
62%
ND
p-X>tene
TO! 3
0.230
28%
0.0118
64%
ND
0.175
21%
0.147
37%
ND
Bromofarm
TO! 5
ND
ND
ND
ND
ND
ND
Styrene
TO! 5
0.190
35%
0.0168
28%
ND
0.275
31%
0.262
51%
0.0135
!!!%
i:t.22-TetracMoroethane
TO! 5
ND
ND
ND
ND
ND
ND
o-Xviene
TO 15
0.15
33%
0.00892
85%
ND
0.14!
30%
0111
57%
ND
X crane
TO! 5
ND
ND
ND
0.00961
200%
0.00753
200%
ND
tumene
TO! 5
ND
ND
ND
ND
ND
ND
CHototolufine s
TO! 5
ND
ND
ND
ND
ND
ND
n-Propylbenzene
TO! 5
0.00621
2tHJ%
ND
ND
0.0223
35%
58%
ND
m-Ethyftoluene
TO! 5
0.0269
2«%
ND
ND
0.037
48%
60%
ND
13,5-Trimeth^fbenzene
TO! 5
ND
ND
ND
ND
200%
ND
S ,2,4-Tiiniethvfbenzene
TO! 5
0.0926
1%
0.00194
200%
ND
" "r
36%
45%
ND
Tert-B utyffi ensene
TO! 5
ND
ND
ND
200%
200%
ND
UEth^Mett^ilB enzene
TO! 5
ND
ND
ND
200%
^D
ND
o-Eteltcfcere
TO! 5
0.0132
200%
ND
ND
45%
63%
ND
13-DicMarobenzene
TO! 5
ND
ND
ND
ND
ND
ND
14-DicKlor obenzene
TO'13
ND
ND
ND
ND
ND
ND
n-Decane
TO! 5
ND
ND
ND
0.0112
200%
200%
ND
S ec-B utylB enz ene
TO! 5
ND
ND
ND
ND
ND
ND
ll.S-Innieth^lberizene
TO! 5
0.0156
200%
ND
ND
0.0491
41%
0.0235
60%
ND
1^-DicHnr obenzene
TO! 5
ND
ND
ND
ND
ND
ND
o-Cvmene
TO! 5
ND
ND
ND
ND
ND
ND
IJ-Diet^ibenzene
TGI 5
\D
ND
ND
0.0129
33%
200%
ND
li-Diethvibenzene
TO! 5
206%
ND
ND
0.0392
58%
0.0211
50%
ND
t>Bia?4Benzeri£
TO! 5
200%
ND
ND
0.0335
58%
73%
ND
Umkcme
TO! 5
\D
ND
ND
0.0300
53%
200?''®
ND
0,4-TricWorobenzene
TO! 5
ND
ND
ND
ND
AD
ND
NapMmtene
TO! 5
1.17
53%
0.253
44%
0.0563
0.530
16%
65%
0.46S
31%
D'odecans
TO! 5
0.00901
2$0%
ND
0.00130
0.0525
90%
0.019S
61%
ND
HexacMarobutadiene
TO! 5
ND
ND
ND
ND
ND
ND
Formal
TO!! A
32.6
17%
4.24
55%
0.397
51%
133
6%
12 3
5%
0.902
Ac etaktehyde
TO! 1A
19.9
17%
i 14
S4%
0.0409
27%
13.3
22%
9.11
26%
0.0741
•C e
Propanal
TO!! A
7.88
3!%
ND
ND
4.13
27%
2.40
24%
0.0164
:o:.
Crotanaldebvde
TOilA
1.32
19%
ND
ND
0.803
54%
0.576
51%
ND
BuftTaldehyde
TO! LA
2 25
200%
ND
ND
0.00792
200%
ND
ND
Benzaldeb/de
TO!! A
1.57
200%
0.174
200%
ND
1.16
«%
t.i
4%
0.0436
200%
Isovaleialdetyde
TO!! A
0.266
200%
ND
ND
0.87
22%
05S4
1%
ND
Valeralctehvde
T011A
0.034
200%
ND
ND
0.0066
200%
ND
0.000164
200%
o-Tctaldfitedg
T011A
0.254
200%
0.121
200%
ND
ND
0.169
200%
ND
m-TduaMefiyde
TO!! A
ND
ND
ND
ND
ND
ND
p-Tolualdehyde
T011A
9.92
200%
ND
ND
ND
ND
ND
Hcxaldebyde
TO!! A
3.02
200%
ND
ND
1.85
22%
ND
ND
2J-Dimethvf)enz aldehyde
TO!! A
ND
ND
ND
ND
0.382
m%
ND
Total VOC s
163
22%
11.4
45%
1.19
61%
102
24%
75.2
24%
5.07
54%
aND=not detected. RPD=relative percent difference.
C-6
-------
Table C-4. Test Average Speciated VOC Emission Factors (Mass/Heat Output)3
Hardwood Pellets
Swifchai asi^ell.ts
25% Load
Syracuse
i033
Load
21 %
Load
Svucuse
. "
Load
ma/MS output
mg'vif
output
me MJ output
msM
output
nieVJ
output
rnsMJ ousput
Compound
Analysis
AVG I
RPD
AVG
RPD
AVG
RPD
AVO
I RPD
V. It
?-:¦
A*- G
RPD
Propylene
TOl 5
15.5
5!%
3 685
ND
14.1
29%
Q ^
23%
0.0! 10
110%
Propane
1013
S..35
4^c
0 338
*4\
ND
7.67
30%
-
33%
ND
Dichbi odrfTuot am ethane
TOl 5
0.0312
!!6%
0 SQ362
4^
0
:: %
0.043!
10%
"
23%
ND
CMcarom ethane
TO! 5
mm
?*•%
0 0322
2Sde
0.538
29%
22%
0.0627
13%
[sobutane
TOl 3
1.33
51%
i "-"i
W%
1.69
31%
-
21%
0.Q0S74
37%
Dicbks otettallucaroethane
TOl 5
ND
ND
ND
ND
ND
ND
VimC Monde
TO! 5
ND
ND
ND
ND
0.00172
200%
O.OOOS58
200%
S-Butene
TO! 3
4.68
46%
0.162
73%
0.0157
132%
5.06
31%
1 Q7
31%
0.0759
6%
13-Butadifine
TO! 5
2,16
37%
0.120
000300
8%
! 71
34%
¦¦ >a
27%
0.022
93%
Butane
TO! 5
1.28
37%
0.0690
0.000318
200%
1 63
39%
34%
ND
trans-2-butene
TO! 5
1.76
44%
0.0503
^ 0
ND
1.66
30%
S"^
34%
0.00172
200%
Bromamefriane
TO! 5
ND
0.000842
2v'if%
0.00205
12%
0.0391
19%
25%
0.0055
0%
cis-2-butene
TOl 5
1-22
sm
0.03®
84%
ND
1.240
27%
34%
ND
CMofoethane
TOl 5
ND
0.000402
w.
0.000703
34%
ND
ND
0.00106
200%
Ethand
TO! 5
0.929
35%
O.0472
12 2%
0.0274
23%
0.60!
36%
3%
0.0218
27%
VirMBromide
TO! 5
ND
ND
ND
ND
ND
ND
Acetonicrite
TO! 3
0.SS3
81%
0,0814
m
0.0172
¦T'
4.37
33%
2.5!
33%
0.0957
15%
Acrdein
TOl 5
12.7
39%
0.638
42%
0.0332
7.41
27%
3
1%
0.0683
32%
Acetone
TOl 5
15.9
49%
0.641
5S%
0.0434
14 9
31%
18%
0.006S5
200%
iso-Pentane
TO! 5
1.17
90%
0.0739
4!%
0.00334
-¦* c
1.67
37%
22%
ND
TricWorcflucr omethane
TO! 3
ND
0.0OOS92
!"4%
0.000145
ND
\D
ND
IsopropjiAlcdid
TOl 5
0.0900
40%
mm
0.0069!
0.106
8%
0 1
22%
0.00375
10%
1-Pentene
TOl 5
0.935
84%
0.0212
ND
0.637
22%
36%
0.OOO631
200%
A^tMonitrie
TOl 5
30%
0.0227
5%
0.00450
24%
111
30%
19%
0.0369
52%
tvPeriaiifi
TOl 3
0 3 ->2
3S%
0.0135
36%
ND
0.618
38%
42%
ND
Isoprene
TOl 5
0
30%
} .133&
77%
0.0110
6%
0.577
42%
27%
ND
trans-2-pentene
T0I5
¦>542
43%
0 3 Hi
82%
0.000639
49%
0.567
34%
36%
0.00159
56%
cis-2-pentene
TOl 5
0 2
4S%
3 0SS3S
62%
0.00029
200%
0.265
28%
36%
0.000SS4
200%
Tert-Butand
TOl 5
3 Co
34%
0 J"fi 1
80%
0.00585
49%
1.15
39%
35%
0.0244
6%
1 !-Dkftiforoethene
TO! 5
ND
ND
ND
ND'
0.00287
200%
ND
MefeieneC blonde
TOl 3
ND
0.00537
8fi%
0.0024!
30%
0.000249
2.00%
^ D
ND
3-C Moro-L-Piopene
TO! 5
0.0657
7m
0.00246
205%
ND
ND
•iMii
ND
] 1 2-Tnchlaro-l 2.2-otfTuciroethane
TOl 3
ND
0.00388
:30%
0.00356
200%
ND
ND
ND
C arbonDtaifide
TOl 3
ND
ND
0.0002S6
200%
ND
ND
0.00311
200%
2 2- Dun etMbut ane
TO! 3
ND
0.00127
30%
ND
ND
ND
ND
trans-'«.2-Dicliaroethene
TOl 5
ND
ND
ND
ND
ND
ND
Cydopentane
TO! 5
3 3"
45%
0.140
67%
0.00484
§9%
2.97
38%
1.74
16%
0.00308
44%
2J-Dimeth\4bulane
TO! 5
$264
39%
0.0124
17%
ND
0.30S
44%
0.I24
71%
ND
] l-DicMoroethane
TOl 3
N'D
ND
ND
ND
ND
ND
\ letfoi-t-B i%i-E thei
TOl 3
13279
200%
NO'
ND
0.0171
200%
ND
ND
VmvLAcetate
TO! 3
8.63
4®%
0.389
123%
0.0252
9m
7.10
10%
3.60
24%
0.00748
200%
2-\iethv%jen£ane
TOl 5
0.213
38%
0.0141
i%
0.000535
200%
0.217
76%
C W42
S-S%
ND
2-Butanone
TO! 5
8.39
42%
0.223
70%
0.00362
62%
7.4!
32%
3 &
29%
0.00447
200%
3-Metb^entane
TOl 5
0.0138
103%
0.00388
114%
0.000162
170%
0.0298
34%
n out
11%
ND
2-CHoroprene
TO 13
ND
ND
ND
\D
ND
ND
1-Hex ene
TOl 5
0.2 S3
42%
0.0100
63%
0.00106
10%
*4~4
31%
0 3]Q
29%
0.00502
43%
cis- 1-2-DicHotoeChene
TOl 5
0.123
m%
000207
200%
ND
" hj421
200%
3 3I3S
200%
ND
Dtisopiopviether
TO! 3
ND
ND
ND
ND
ND
ND
EtteiAcetate
TOl 5
^D
ND
ND
ND
ND
ND
n-He\ ane
TOl 5
23%
0.0130
m%
0.00180
122%
0.202
43%
0.0961
45%
ND
CHorcform
TO! 3
28%
0.00131
70%
0.000501
42%
0.0222
62%
195%
0.0000678
200%
letrahydrofuran
TO! 3
•)in
2!%
0.0104
54%
ND
0.446
39%
57%
ND
EthyO* ert-B utyEttier
TOl 3
27%
0.00306
108%
ND
ND
200%
ND
Methyk^lopentane
TO! 3
101%
0.0038
167%
ND
0.0119
200%
200%
ND
12-Dichtaroefhane
TOl 5
\D
0.000511
11%
3.86E-07
200%
ND
ND
ND
2,4- Dwi e tftylpentane
TOl 3
30%
0.00112
33%
ND
ND
ND
ND
UJ-TricMaroeUiane
TO! 5
>D
ND
ND
ND
ND
ND
Benzene
TOl 5
48%
LSI
0.354
M%
7.73
13%
7.M
43%
3.79
56%
CarbonTe traction de
TO! 3
J i1 24
200%
0.000947
51 c
0.000458
102%
0.00197
40%
0.00134
0%
0.000109
200%
Cydohexane
TOl 3
59%
ND
ND
0.0259
200%
0.0221
47%
ND
2-Nlethyihexane
TOl 5
21%
0.0193
200%
ND
0.00779
61%
0.0O543
200%
ND
2J-Dimethylpentane
TOl 3
" D
0.000452
2
ND'
ND
ND
ND
TertAmviMetfoJLlfoe r
TO! 5
ND
ND
ND
ND
ND
ND
3-methvthexane
TO! 3
0.0140
200%
ND
0.000148
200%
0.00911
200%
0.00232
200%
ND
O-DicHaropropanfi
TOl 3
ND
ND
ND
ND
ND
ND
Br omodcHor amethane
TOl 3
ND
ND
ND
ND
ND
ND
1.4-Diaxanfi
TOl 5
ND
ND
ND
ND
ND
ND
C-7
-------
Table C-4 (continued)
Hai ood ? eflets
Switchsrass Pellets
:5s
Lead
Syracuse
Load
21%
Load
Syracuse
.
Load
n^\
U auipu:
m,s3.r
output
outpti
xneM
C0fipU.t
mgMJ output
msMJ outpttt
Compound
Analysis
AVO
I RPD
AVG
RPD
\\ G
PPD
AVG
I RPD
AVG |
RPD
AVG
RPD
Tricliaroethene
TO 15
ND
ND
ND
ND
ND
ND
Isooctane
TO! 5
ND
0.00050S
146%
ND
ND
ND
ND
J..&tei!v!ethacr\iate
TO! 5
ND
ND
ND
ND
ND
ND
Heptane
TO! 5
.] 3S4S
37%
0.000496
200%
ND
§.212
31%
o as^s
75%
ND
cis- [ 5-Di zldoiopropene
TOl 3
nd
ND
ND
ND
ND
ND
4-Methy-2-P entanone
TO! 5
¦} 53^:
200%
ND
ND
0.0614
17%
?
44%
ND
Metfrvbvdohexane
TO1! 5
3^4^
200%
ND
ND
0.01140
77%
3'?0"54
6S%
ND
ttans-1.3 -Diclrtoropropene
TO! 5
XD
m
ND
ND
XD
ND
I! 2-1 icliaroethane
TO! 5
ND
ND
ND
ND
ND
ND
2 5,4-TimietteipeitoK.
TO! 5
XD
0.000385
73%
ND
ND
ND
ND
Toluene
TO! 5
l ii
43%
0.267
35%
0.003 S5
23%
5.4!
28%
3.5S
47%
0.0536
73%
2~\Iettefieptane
TO! 5
.
200%
ND
ND
§.0!7
200%
0.0105
200%
ND
2-Kexanone
TO! 5
VD
ND
ND
ND
ND
ND
Dibiomoctka omethane
TO! 5
ND
ND
ND
ND
ND
ND
3-Ivfettefaptane
TO! 5
irs:
200%
nd
ND
ND
ND
ND
! ,2-Ditrom oethane
TO! 3
ND
ND
ND
ND
ND
ND
Octane
TO! 5
0.025
200%
ND
ND
0.0704
4-0%
0.0296
66%
ND
TetracHaroethfine
TO! 5
ND
0.00104
49%
o.ooos2&
200%
ND
ND
ND
IJ J2-TetracWciioethane
TO! 5
ND
ND
ND
ND
ND
ND
Cliorobenzene
TO! 5
ND
ND
0.00226
20%
ND
ND
0.00294
2.00%
Ethylbenzene
TO! 5
38%
0.0126
70%
ND
0.549
33%
0315
50%
ND
m-X\iene
TO! 5
a 4
21%
0.0160
84%
ND
0.394
26%
0.289
62%
ND
p-Xvtene
TO! 3
0 334
28%
0.0132
64%
ND
0.2S0
21%
0.201
57%
ND
Bromofarm
TO! 5
ND
ND
ND
ND
ND
ND
Styrene
TO! 5
n:^
35%
0.01S4
28%
ND
0.436
31%
0359
51%
0.0161
111%
i:t.22-TetracMoroethane
TO! 5
ND
ND
ND
ND
ND
ND
o-Xviene
TO! 5
0.216
33%
0.010!
85%
ND
§.224
30%
0.15!
57%
ND
X crane
TO! 5
ND
ND
ND
0.0141
200%
0.00995
200%
ND
Cumene
TO! 5
ND
ND
ND
ND
ND
ND
CHototolufine s
TO! 5
ND
ND
ND
:o
\TD
ND
n-Propylbenzene
TO! 5
0.0137
200%
ND
ND
0 'j3!4
35%
58%
ND
m-Ethyftoluene
TO! 5
0.0592
2®%
ND
ND
1 3583
48%
ao%
ND
13,5-Trimeth^fbenzene
TO! 5
ND
ND
ND
XD
200%
ND
S ,2,4-Tiiniethvfbenzene
TO! 5
0.143
1%
0.00235
200%
ND
36%
45%
ND
Tert-B utyffi ensene
TO! 5
ND
ND
ND
J
200%
0.00721
200%
ND
UEth^Mett^ilB enzene
TO! 5
ND
ND
ND
200%
ND
o-Eteltcfcere
TO! 5
0.029!
200%
ND
ND
45%
63%
ND
13-Dichtarobenzene
TO! 5
ND
ND
ND
ND
ND
1.4-DicKlor obenzene
TO! 3
ND
ND
ND
ND
ND
n-Decane
TOl 5
ND
ND
ND
0.0164
200%
200%
ND
S ec-B utylB em ene
TO! 5
ND
ND
ND
XD
ND
ND
ll.S-Innieth^lberizene
TO! 5
0.0342
200%
ND
ND
41%
0.0321
60%
ND
1^-DicHnr obenzene
TOl 5
ND
ND
ND
.O
ND
ND
o-Cvmene
TOl 5
ND
ND
ND
xd
ND
ND
IJ-Diet^ibenzene
TOl 5
XD
ND
ND
-
33%
0.00506
200%
ND
li-Diethvibenzene
TO! 5
206%
ND
ND
2 JMi
58%
. <2^
50%
ND
t>Bia?4Benzeri£
TOl 5
I 14^
200%
ND
ND
58%
"> ,Q
73%
ND
Umkcme
TO! 5
ND
ND
ND
¦>
53%
200%
ND
0,4-TricWorobenzene
TOl 5
ND
ND
ND
O
AD
ND
NapMmtene
TO! 5
2.01
53%
0.265
44%
0.0513
'2"1!
< £~n
16%
nQ1?
63%
0.572
31%
D'odecans
TOl 5
0.0198
2$0%
ND
0.00112
2 ' .
1 3T2
90%
0.027
61%
ND
HexacMarobutadiene
TO! 5
ND
ND
ND
ND
ND
ND
Formal
T011A
48.5
17%
4.72
55%
0.381
51%
2! J
6%
17 1
5%
1.09
Ac etaktehyde
TOM A
29.7
17%
1.29
S4%
0.0398
27%
212
22%
12.5
26%
0.0903
Propanal
TOl! A
114
3!%
ND
ND
€M
27%
331
24%
0.019!
Crotanaldebvde
TO!! A
im
ND
ND
1.26
54%
0.80?
51%
ND
BuftTaldehyde
TOl LA
200%
ND
ND
0.0116
200%
ND
ND
Benzaldeb/de
TOllA
200%
0.212
200%
ND
1.93
m%
1.52
4%
0.0506
200%
Isovaleialdetyde
TOllA
200%
ND
ND
1.39
22%
O.SOS
1%
ND
Valeralctehvde
TOllA
-
200%
ND
ND
0.0097
200%
ND
0.000215
200%
o-Tctaldfitedg
TOllA
d s
200%
0.113
200%
ND
ND
0.223
200%
ND
m-TduaMefiyde
TOllA
\D
ND
ND
ND
ND
ND
p-Tolualdehyde
TOllA
21.8
200%
ND
ND
ND
ND
ND
Hcxaldebyde
TOllA
2.70
200%
ND
ND
2.96
22%
ND
ND
2J-Dimetlrrv1berizaMehv^ie
TOllA
ND
ND
ND
ND
§.522
em
ND
Total VOC s
24!
«%
12.6
$9%
1.13
37%
162
6%
104
15%
6.16
42%
aND=not detected. RPD=relative percent difference.
C-8
-------
Table C-5. Test Average Total PAH Emission Factors
Test Average Total PAH Emission Factor
Tcsl Condition
Dale
nig/kg fuel
mg/MJ Input lb/MMBTU Tnpul
mg/MJ Output
lb/MMBTU Out
Hardwood - Full Load
10/19/2016
4.69E+00
2.61 E-01
6.06E-04
2.26E-01
5.24E-04
Hardwood - Full Load
10/20/2016
6.24E-01
3.47E-02
8.06E-05
3.88E-02
9.01 E-05
Hardwood - Low Load
10/21/2016
1.52E+01
8.43 E-01
1.96E-03
1.85E+00
4.31E-03
Hardwood - Low Load
10/25/2016
2.12E+02
1.18E+01
2.74E-02
1.05E+01
2.44E-02
Hardwood - Svracusc
10/26/2016
3.66E+00
2.03E-01
4.73E-04
1.91 E-01
4.43 E-04
Hardwood - Svracusc
10/27/2016
3.20E+00
1.78E-01
4.14E-04
2.16E-01
5.03E-04
Swilchgrass - Low Load
11/3/2016
1.07E+01
6.12E-01
1.42E-03
8.99E-01
2.09E-03
Swilchgrass - Low Load
11/4/2016
1.27E+01
7.29E-01
1.69E-03
1.28E+00
2.97E-03
Swilchgrass - Syracuse
11/8/2016
6.49E+01
3.71E+00
8.62E-03
4.90E+00
1.14E-02
Swilchgrass - Syracuse
11/9/2016
7.84E+01
4.48E+00
1.04E-02
6.49E+00
1.51E-02
Swilchgrass - Full Load
11/10/2016
1.13E+01
6.47E-01
1.50E-03
8.47E-01
1.97E-03
Swilchgrass - Full Load
11/15/2016
3.44E+01
1.97E+00
4.57E-03
2.28E+00
5.30E-03
C-9
-------
Table C-6. Test Average Speciated PAH Emission Factors (Mass/Mass Fuel Burned)
Hardwood Pellets
Switchgrass Pellets
25% Load
Syracuse Cycle
100%
Load
25%
Load
Syracuse Cycle
O
o
¦ o
ox
Load
10/21/2016
10/25/2016
10/26/2016
10/27/2016
10/19/2016
10/20/2016
11/3/2016
11/4/2016
11/8/2016
11/9/2016
11/10/2016
11 15 2016
PAH Compound
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
mg/kg fuel
Naphthalene
1.05E+01
4.53E+00
2.26E+00
1.97E+00
2.83E+00
3.90E-01
7.35E+00
9.10E+00
3.94E+01
4.97E+01
5.15E+00
1.69E+01
Acenapthylene
1.29E+00
4.68E+01
2.10E-01
9.90E-02
1.98E-01
1.88E-02
5.27E-01
6.36E-01
7.36E+00
5.72E+00
1.93E+00
5.85E+00
Acenaphthene
6.44&02
4.12E+00
2.06E-02
7.34&03
4.28E-02
1.07&03
8.17E-02
9.75E-02
2.64E-01
4.30E-01
3.48E-02
9.48E-02
Flourene
2.53E-01
1.67E+01
5.81E-02
4.49E-02
4.34E-02
3.80E-03
5.73E-01
6.24E-01
2.58E+00
2.36E+00
1.78E-01
6.16E-01
Phenanthrene
1.53E+00
7.03E+01
4.67E-01
4.63E-01
5.78E-01
7.10&02
1.09E+00
1.23E+00
6.91E+00
9.21E+00
1.24E+00
3.70E+00
Anthracene
3.17E-01
1.66E+01
4.14E-02
3.10E-02
2.27E-02
2.87E-03
1.86E-01
1.98E-01
9.38E-01
1.12E+00
1.04E-01
3.41E-01
Fluoranthene
3.50&01
1.98E+01
2.15E-01
2.40&01
3.46E-01
5.25&02
2.97E-01
2.84E-01
2.54E+00
3.64E+00
9.52E-01
2.43E+00
Pyrene
3.50E-01
1.82E+01
2.34E-01
2.53E-01
3.91E-01
5.42E-02
3.14E-01
2.67E-01
2.35E+00
3.17E+00
1.09E+00
2.66E+00
Benzo(a)Anthracene
6.83E-02
2.72E+00
1.44E-02
1.23E-02
2.07E-02
4.42E-03
5.98E-02
6.20E-02
4.19E-01
4.50E-01
5.76E-02
1.34E-01
Chrysene
1.08E-01
4.74E+00
2.60E-02
3.03E-02
3.88E-02
8.35E-03
1.15E-01
1.09E-01
5.25E-01
7.35E-01
8.91E-02
2.29E-01
Benzo(b)Flouranthene
6.17E-02
1.62E+00
1.80E-02
1.37E-02
2.88E-02
4.27E-03
3.14E-02
3.62E-02
3.49&01
4.17E-01
8.18E-02
2.51E-01
Benzo(k)flouranthene
6.66E-02
1.58E+00
2.15E-02
1.40E-02
2.47E-02
3.86E-03
3.34E-02
3.80E-02
3.79E-01
4.85E-01
8.49E-02
2.72E-01
Benzo(a)pyrene
7.49E-02
1.73E+00
1.75E-02
4.26E-03
1.98&02
1.17E-03
2.36&02
3.04E-02
2.90E-01
3.16E-01
5.70E-02
2.30E-01
Indeno(l,2,3-cd)pyrene
6.96E-02
1.27E+00
1.95E-02
8.54E-03
3.46E-02
2.74E-03
1.36E-02
1.69E-02
2.64E-01
2.88E-01
8.43E-02
2.99E-01
Dibenz(a,h)anthracene
6.29E-03
1.88E-01
1.21E-03
8.10E-04
9.56E-04
1.16E-04
2.70E-03
3.82E-03
3.30&02
3.81E-02
4.15&03
1.53E-02
Benzo(g,h,i)perylene
7.95E-02
1.15E+00
3.66E-02
1.48E-02
6.44E-02
4.80E-03
1.36E-02
1.65E-02
3.24E-01
3.46E-01
1.63E-01
4.38E-01
Totals
1.52E+01
2.12E+02
3.66E+00
3.20E+00
4.69E+00
6.24E-01
1.07E+01
1.27E+01
6.49E+01
7.84E+01
1.13E+01
3.44E+01
C-10
-------
Table C-7. Test Average Speciated PAH Emission Factors (Mass/Heat Input)
Hardwood Pellets
Switchgrass Pellets
25% Load
Syracuse Cycle
100%
Load
25%
Load
Syracuse Cycle
100%
Load
10/21/2016
10/25/2016
10/26/2016
10/27/2016
10/19/2016
10/20/2016
11/3/2016
11/4/2016
11/8/2016
11/9/2016
11/10/2016
11/15/2016
PAH Compound
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
Naphthalene
5.83E-01
2.52E-01
1.26E-01
1.09E-01
1.58E-01
2.17E-02
4.20E-01
5.20&01
2.25E+00
2.84E+00
2.95E-01
9.65E-01
Acenapthylene
7.17E-02
2.60E+00
1.17E-02
5.51E-03
1.10E-02
1.04E-03
3.01E-02
3.63E-02
4.20E-01
3.27E-01
1.10E-01
3.35E-01
Acenaphthene
3.58&03
2.29E-01
1.15E-03
4.08&04
2.38E-03
5.98&05
4.67E-03
5.58E-03
1.51E-02
2.46E-02
1.99E-03
5.42E-03
Flourene
1.41E-02
9.29E-01
3.23 E-03
2.49E-03
2.41E-03
2.11E-04
3.27E-02
3.57E-02
1.47E-01
1.35E-01
1.02E-02
3.52E-02
Phenanthrene
8.51E-02
3.91E+00
2.60E-02
2.57E-02
3.21E-02
3.95E-03
6.23E-02
7.03E-02
3.95E-01
5.26E-01
7.10E-02
2.11E-01
Anthracene
1.76E-02
9.22E-01
2.30E-03
1.72E-03
1.26E-03
1.59E-04
1.06E-02
1.13E-02
5.36E-02
6.40E-02
5.95E-03
1.95E-02
Fluoranthene
1.94E-02
1.10E+00
1.20E-02
1.33E-02
1.93E-02
2.92E-03
1.70E-02
1.62E-02
1.45E-01
2.08E-01
5.44E-02
1.39E-01
Pyrene
1.95E-02
1.01E+00
1.30E-02
1.41E-02
2.17E-02
3.02E-03
1.79E-02
1.53E-02
1.35E-01
1.81E-01
6.25E-02
1.52E-01
Benzo(a)Anthracene
3.80E-03
1.51E-01
8.02E-04
6.85E-04
1.15&03
2.46E-04
3.42&03
3.55E-03
2.40E-02
2.57E-02
3.29E-03
7.67E-03
Chrysene
6.01E-03
2.63E-01
1.44E-03
1.69E-03
2.16E-03
4.65E-04
6.59E-03
6.24E-03
3.00E-02
4.20E-02
5.09E-03
1.31E-02
Benzo(b)Flouranthene
3.43E-03
9.01E-02
1.00E-03
7.60E-04
1.60E-03
2.37E-04
1.79E-03
2.07E-03
1.99E-02
2.38E-02
4.68&03
1.43E-02
Benzo(k)flouranthene
3.70E-03
8.81E-02
1.20E-03
7.77E-04
1.37E-03
2.15E-04
1.91E-03
2.17E-03
2.16E-02
2.77E-02
4.85E-03
1.56E-02
Benzo(a)pyrene
4.16E-03
9.60E-02
9.74E-04
2.37E-04
1.10&03
6.49E-05
1.35E-03
1.74E-03
1.66E-02
1.81E-02
3.26E-03
1.31E-02
Indeno(l,2,3-cd)pyrene
3.87E-03
7.04E-02
1.08E-03
4.75E-04
1.92E-03
1.53E-04
7.76E-04
9.66E-04
1.51E-02
1.65E-02
4.82E-03
1.71E-02
Dibenz(a,h)anthracene
3.50E-04
1.05E-02
6.75E-05
4.51E-05
5.32E-05
6.43E-06
1.54E-04
2.18E-04
1.88E-03
2.18E-03
2.37E-04
8.73&04
Benzo(g,h,i)perylene
4.42E-03
6.38E-02
2.04E-03
8.22E-04
3.58E-03
2.67E-04
7.80E-04
9.43E-04
1.85E-02
1.98E-02
9.32E-03
2.50&02
Totals
8.43E-01
1.18E+01
2.03E-01
1.78E-01
2.61E-01
3.47E-02
6.12E-01
7.29E-01
3.71E+00
4.48E+00
6.47E-01
1.97E+00
C-ll
-------
Table C-8. Test Average Speciated PAH Emission Factors (Mass/Heat Output)
Hardwood Pellets
Switchgrass Pellets
25% Load
Syracuse Cycle
100%
Load
25%
Load
Syracuse Cycle
100%
Load
10/21/2016
10/25/2016
10/26/2016
10/27/2016
10/19/2016
10/20/2016
11/3/2016
11/4/2016
11/8/2016
11/9/2016
11/10/2016
11/15/2016
Isomer
mg/MJ Output mg/MJ Outpu
mg/MJ Output mg/MJ Output
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
mg/MJ Input
Naphthalene
1.28E+00
2.25E-01
1.18E-01
1.33E-01
1.36E-01
2.43E-02
6.17E-01
9.12&01
2.97E+00
4.11E+00
3.86E-01
1.12E+00
Acenapthylene
1.58E-01
2.32E+00
1.10E-02
6.69E-03
9.53E-03
1.17E-03
4.43E-02
6.37E-02
5.55E-01
4.73E-01
1.45E-01
3.88E-01
Acenaphthene
7.87&03
2.04E-01
1.07E-03
4.96&04
2.06E-03
6.68&05
6.86E-03
9.77E-03
1.99E-02
3.56E-02
2.61E-03
6.28E-03
Flourene
3.10E-02
8.29E-01
3.03E-03
3.03E-03
2.09E-03
2.36E-04
4.81E-02
6.25E-02
1.95E-01
1.96 E-01
1.33E-02
4.08E-02
Phenanthrene
1.87E-01
3.49E+00
2.44E-02
3.12E-02
2.78E-02
4.41&03
9.15E-02
1.231>01
5.21E-01
7.61E-01
9.30E-02
2.45E-01
Anthracene
3.88E-02
8.22E-01
2.16E-03
2.09E-03
1.09E-03
1.78E-04
1.56E-02
1.98E-02
7.08E-02
9.26E-02
7.79E-03
2.26E-02
Fluoranthene
4.28&02
9.82E-01
1.12E-02
1.62&02
1.67E-02
3.26E-03
2.49E-02
2.84E-02
1.92E-01
3.01E-01
7.13E-02
1.61E-01
Pyrene
4.28E-02
9.05E-01
1.22E-02
1.71E-02
1.88E-02
3.37E-03
2.63E-02
2.67E-02
1.78E-01
2.62E-01
8.18E-02
1.76E-01
Benzo(a)Anthracene
8.36E-03
1.35E-01
7.52E-04
8.32E-04
9.98&04
2.75E-04
5.02&03
6.21E-03
3.17E-02
3.73E-02
4.31E-03
8.90E-03
Chrysene
1.32E-02
2.35E-01
1.35E-03
2.05E-03
1.87E-03
5.19E-04
9.68E-03
1.09E-02
3.96E-02
6.08E-02
6.67E-03
1.52E-02
Benzo(b)Flouranthene
7.55E-03
8.04E-02
9.41E-04
9.23 E-04
1.39E-03
2.65E-04
2.63E-03
3.63E-03
2.63&02
3.45E-02
6.13E-03
1.66E-02
Benzo(k)flouranthene
8.15E-03
7.86E-02
1.12E-03
9.43E-04
1.19E-03
2.40E-04
2.80E-03
3.80E-03
2.86E-02
4.01E-02
6.36E-03
1.80E-02
Benzo(a)pyrene
9.16E-03
8.56&02
9.14E-04
2.88E-04
9.52&04
7.25E-05
1.98E-03
3.05E-03
2.19E-02
2.62E-02
4.27E-03
1.52E-02
Indeno( 1,2,3- cd)pyrene
8.51E-03
6.28E-02
1.02E-03
5.77E-04
1.66E-03
1.71E-04
1.14E-03
1.69E-03
1.99E-02
2.39E-02
6.31E-03
1.98E-02
Dibenz(a,h)anthracene
7.69E-04
9.33E-03
6.33E-05
5.47E-05
4.60E-05
7.18E-06
2.27E-04
3.83E-04
2.49E-03
3.15E-03
3.11E-04
1.01&03
Benzo(g,h,i)perylene
9.72E-03
5.69E-02
1.91E-03
9.98E-04
3.10E-03
2.98E-04
1.15E-03
1.65E-03
2.45E-02
2.86E-02
1.22E-02
2.90&02
Totals
1.85E+00
1.05E+01
1.91E-01
2.16E-01
2.26E-01
3.88E-02
8.99E-01
1.28E+00
4.90E+00
6.49E+00
8.47E-01
2.28E+00
C-12
-------
Table C-9. Composite Speciated PCDD/PCDF Emission Factors (Mass/Mass Fuel Burned)
Hardwood Pellets
Switchgrass Pellets
25% Load
Syracuse Cycle
100% Load
25% Load
Syracuse Cycle
100% Load
Isomer
ng TEQ/kg fuel
ng TEQ/kg fuel
ng TEQ/kg fuel
ng TEQ/kg fuel
ng TEQ/kg fuel
ng TEQ/kg fuel
2,3,7,8 - TCDD
0.027
0.020
0.062
0.036
0.084
0.022
1,2,3,7,8 - PeCDD
0.042
0.019
0.063
0.089
0.168
0.024
1,2,3,4,7,8 - HxCDD
0.001
0.000
0.001
0.002
0.004
0.000
1,2,3,6,7,8 - HxCDD
0.003
0.001
0.001
0.011
0.013
0.001
1,2,3,7,8,9- HxCDD
0.002
0.000
0.001
0.006
0.008
0.000
1,2,3,4,6,7,8 - HpCDD
0.002
0.000
0.000
0.005
0.006
0.000
1,2,3,4,6,7,8,9- OCDD
0.000
0.000
0.000
0.001
0.001
0.000
2,3,7,8 - TCDF
0.039
0.026
0.088
0.029
0.079
0.031
1,2,3,7,8 - PeCDF
0.003
0.002
0.008
0.003
0.006
0.002
2,3,4,7,8 - PeCDF
0.030
0.020
0.079
0.026
0.064
0.019
1,2,3,4,7,8 - HxCDF
0.003
0.001
0.005
0.003
0.005
0.001
1,2,3,6,7,8 - HxCDF
0.004
0.002
0.006
0.003
0.007
0.002
1,2,3,7,8,9- HxCDF
0.000
0.000
0.001
0.001
0.001
0.000
2,3,4,6,7,8 - HxCDF
0.002
0.001
0.005
0.003
0.006
0.001
1,2,3,4,6,7,8 - HpCDF
0.000
0.000
0.001
0.003
0.002
0.000
1,2,3,4,7,8,9 - HpCDF
0.000
0.000
0.000
0.000
0.000
0.000
1,2,3,4,6,7,8,9 - OCDF
0.000
0.000
0.000
0.001
0.001
0.000
Totals
0.158
0.093
0.320
0.223
0.455
0.105
C-13
-------
Table C-10. Composite Speciated PCDD/PCDF Emission Factors (Mass/Heat Input)
Isomer
Hardwood Pellets
Switchgrass Pellets
25% Load
Syracuse
Cycle
100%
Load
25% Load
Syracuse
Cycle
100%
Load
ny TEQ/MJ
Ib/MMBTU
ny TEQ/MJ
Ib/MMBTU
ng TEQ/MJ
Ib/MMBTU
ng TEQ/MJ
Ib/MMBTU
ng TEQ/MJ
Ib/MMBTU
ng TEQ/MJ
Ib/MMBTU
2.3.7.8 - TCDD
1.50E-03
3.48E-12
1.09E-03
2.53E-12
3.44E-03
7.99E-12
2.05E-03
4.77E-12
4.78E-03
1.11 E-11
1.23E-03
2.8636E-12
1.2.3.7.8 - PeCDD
2.35E-03
5.47E-12
1.08E-03
2.51E-12
3.51E-03
8.15E-12
5.11E-03
1.19E-11
9.62E-03
2.23E-11
1.40E-03
3.2474E-12
1,2,3,4.7.8 - HxCDD
5.35E-05
1.24E-13
1.39E-05
3.22E-14
4.12E-05
9.57E-14
1.25E-04
2.90E-13
2.07E-04
4.80E-13
2.03E-05
4.7234E-14
1.2.3.6.7.8 - HxCDD
1.57E-04
3.66E-13
2.85E-05
6.61E-14
8.15E-05
1.89E-13
6.04E-04
1.40E-12
7.56E-04
1.75E-12
4.32E-05
1.0037E-13
1,2,3.7.8.9 - HxCDD
9.40E-05
2.18E-13
1.67E-05
3.87E-14
4.53E-05
1.05E-13
3.58E-04
8.32E-13
4.70E-04
1.09E-12
2.16E-05
5.0186E-14
1.2.3.4.6.7.8 - HpCDD
1.10E-04
2.54E-13
8.19E-06
1.90E-14
1.87E-05
4.34E-14
3.01E-04
6.99E-13
3.41E-04
7.91E-13
1.28E-05
2.9817E-14
1.2.3,4,6,7,8,9 - OCDD
9.77E-06
2.27E-14
8.50E-07
1.97E-15
1.37E-06
3.19E-15
5.32E-05
1.24E-13
4.48E-05
1.04E-13
2.07E-06
4.809E-15
2.3.7.8 - TCDF
2.15E-03
5.00E-12
1.45E-03
3.36E-12
4.87E-03
1.13E-11
1.68E-03
3.91E-12
4.49E-03
1.04E-11
1.79E-03
4.1684E-12
1.2,3,7,8 - PeCDF
1.93E-04
4.48E-13
1.12E-04
2.61E-13
4.30E-04
9.99E-13
1.47E-04
3.42E-13
3.40E-04
7.90E-13
1.05E-04
2.4444E-13
2.3.4.7.8 - PeCDF
1.66E-03
3.86E-12
1.11E-03
2.57E-12
4.39E-03
1.02E-11
1.51E-03
3.51E-12
3.68E-03
8.54E-12
1.11E-03
2.5684E-12
1.2.3.4.7.8 - HxCDF
1.51E-04
3.50E-13
6.87E-05
1.60E-13
2.70E-04
6.28E-13
1.53E-04
3.55E-13
3.07E-04
7.12E-13
7.50E-05
1.7418E-13
1.2.3.6.7.8 - HxCDF
2.02E-04
4.69E-13
9.16E-05
2.13E-13
3.56E-04
8.27E-13
1.97E-04
4.58E-13
3.92E-04
9.10E-13
9.53E-05
2.2141E-13
1.2.3.7.8.9 - HxCDF
2.60E-05
6.04E-14
1.87E-05
4.35E-14
6.84E-05
1.59E-13
4.43E-05
1.03E-13
7.13E-05
1.66E-13
1.78E-05
4.133E-14
2.3.4.6.7.8 - HxCDF
1.10E-04
2.56E-13
7.08E-05
1.64E-13
2.61E-04
6.07E-13
1.81E-04
4.21E-13
3.28E-04
7.61E-13
6.99E-05
1.6237E-13
1.2.3.4.6.7.8 - HpCDF
2.24E-05
5.20E-14
1.05E-05
2.43E-14
3.02E-05
7.01E-14
1.96E-04
4.54E-13
1.01E-04
2.35E-13
1.04E-05
2.4208E-14
1,2,3,4,7,8.9 - HpCDF
2.45E-06
5.68E-15
1.46E-06
3.39E-15
3.22E-06
7.47E-15
9.66E-06
2.24E-14
1.07E-05
2.48E-14
1.53E-06
3.5426E-15
1.2.3.4.6.7.8.9 - OCDF
3.82E-06
8.87E-15
1.14E-06
2.64E-15
1.95E-06
4.54E-15
4.22E-05
9.80E-14
4.87E-05
1.13E-13
2.60E-06
6.0401E-15
Totals
8.80E-03
2.04E-11
5.17E-03
1.20E-11
1.78E-02
4.14E-11
1.28E-02
2.97E-11
2.60E-02
6.03E-11
6.01E-03
1.3958E-11
C-14
-------
Table C-ll. Composite Speciated PCDD/PCDF Emission Factors (Mass/Heat Output)
Hardwood Pellets
Svvitcligras
s Pellets
25%
,oad
Syracuse
Cvcle
100°o
Load
25%
Load
Syracuse
Cvcle
100%
Load
Isomer
ny TEQ/MJ
IbMMBTU
nu TEQ/MJ
Ib/MMBTU
nu TEQ/MJ
Ib/MMBTU
ny TEQ/MJ
Ib/MMBTU
ng TEQ/MJ
Ib'MMBTU
ng TEQ/MJ
lb MMBTU
2.3.7.8 - TCDD
2.32E-03
5.38E-12
1.17E-03
2.72E-12
3.84E-03
8.93E-12
3.31E-03
7.68E-12
6.61E-03
1.53E-11
1.52E-03
3.54E-12
1.2.3.7,8 - PeCDD
3.64E-03
8.46E-12
1.17E-03
2.71E-12
3.92E-03
9.11E-12
8.23E-03
1.91E-11
1.33E-02
3.09E-11
1.73E-03
4.01E-12
1.2.3.4.7.8 - HxCDD
8.28E-05
1.92E-13
1.49E-05
3.47E-14
4.61E-05
1.07E-13
2.01E-04
4.67E-13
2.86E-04
6.64E-13
2.51E-05
5.83E-14
1.2.3.6,7,8 - HxCDD
2.44E-04
5.66E-13
3.06E-05
7.12E-14
9.10E-05
2.11E-13
9.72E-04
2.26E-12
1.05E-03
2.43E-12
5.34E-05
1.24E-13
1.2.3.7.8.9 - HxCDD
1.46E-04
3.38E-13
1.79E-05
4.17E-14
5.06E-05
1.17E-13
5.77E-04
1.34E-12
6.51E-04
1.51E-12
2.67E-05
6.20E-14
1.2.3.4.6.7.8 - HpCDD
1.70E-04
3.94E-13
8.82E-06
2.05E-14
2.09E-05
4.85E-14
4.85E-04
1.13E-12
4.71E-04
1.09E-12
1.59E-05
3.68E-14
1,2.3.4.6.7.8.9 - OCDD
1.51E-05
3.51E-14
9.15E-07
2.12E-15
1.53E-06
3.56E-15
8.57E-05
1.99E-13
6.19E-05
1.44E-13
2.56E-06
5.94E-15
2,3,7,8 - TCDF
3.33E-03
7.73E-12
1.56E-03
3.61E-12
5.44E-03
1.26E-11
2.71E-03
6.29E-12
6.21E-03
1.44E-11
2.22E-03
5.15E-12
1,2,3,7,8 - PeCDF
2.98E-04
6.93E-13
1.21E-04
2.81E-13
4.81E-04
1.12E-12
2.37E-04
5.51E-13
4.70E-04
1.09E-12
1.30E-04
3.02E-13
2.3.4.7.8 - PeCDF
2.57E-03
5.98E-12
1.19E-03
2.76E-12
4.90E-03
1.14E-11
2.43E-03
5.65E-12
5.09E-03
1.18E-11
1.37E-03
3.17E-12
1,2,3,4,7,8 - HxCDF
2.33E-04
5.41E-13
7.40E-05
1.72E-13
3.02E-04
7.02E-13
2.46E-04
5.72E-13
4.24E-04
9.85E-13
9.26E-05
2.15E-13
1.2.3.6.7.8 - HxCDF
3.12E-04
7.25E-13
9.86E-05
2.29E-13
3.98E-04
9.24E-13
3.18E-04
7.38E-13
5.42E-04
1.26E-12
1.18E-04
2.73E-13
1,2,3,7,8,9 - HxCDF
4.02E-05
9.34E-14
2.02E-05
4.69E-14
7.64E-05
1.77E-13
7.13E-05
1.66E-13
9.86E-05
2.29E-13
2.20E-05
5.10E-14
2.3.4.6.7.8 - HxCDF
1.70E-04
3.96E-13
7.62E-05
1.77E-13
2.92E-04
6.79E-13
2.92E-04
6.77E-13
4.54E-04
1.05E-12
8.63E-05
2.00E-13
1.2.3.4.6.7.8 - HpCDF
3.47E-05
8.05E-14
1.13E-05
2.62E-14
3.37E-05
7.83E-14
3.15E-04
7.32E-13
1.40E-04
3.25E-13
1.29E-05
2.99E-14
1.2.3.4.7.8.9 - HpCDF
3.79E-06
8.79E-15
1.57E-06
3.64E-15
3.60E-06
8.35E-15
1.56E-05
3.61E-14
1.48E-05
3.43E-14
1.88E-06
4.37E-15
1,2,3,4,6,7,8,9 - OCDF
5.91E-06
1.37E-14
1.22E-06
2.84E-15
2.18E-06
5.07E-15
6.79E-05
1.58E-13
6.74E-05
1.56E-13
3.21E-06
7.46E-15
Totals
1.36E-02
3.16E-11
5.56E-03
1.29E-11
1.99E-02
4.62E-11
2.06E-02
4.78E-11
3.59E-02
8.35E-11
7.42E-03
1.72E-11
C-15
-------
APPENDIX D
Particle-Phase Emission Factors
D-l
-------
Table D-l. Test Average Total PM Mass Emissions and Thermal Efficiency Summary (ASTM 2515)
Average
Total PM EF
Total PM EF
Total PM EF
Total PM EF
Total PM EF
Test Condition
Date
Efficiency
lb/MMBTU Input
g/MJ Input
lb/MMBTU Output g/MJ Output
g/kg fuel
Hardwood - Full Load
10/19/20
16
89.5%
5.46E-02
2.35E-02
4.72E-02
2.03 E-02
4.23 E-01
Hardwood - Full Load
10/20/20
16
4.89E-02
2.11E-02
5.47E-02
2.35E-02
3.79E-01
Hardwood - Low Load
10/21/20
16
78.8%
3.72E-01
1.60E-01
8.18E-01
3.52E-01
2.88E+00
Hardwood - Low Load
10/25/20
16
3.80E-01
1.63 E-01
3.39E-01
1.46E-01
2.94E+00
Hardwood - Syracuse
10/26/20
16
94.5%
3.92E-02
1.69E-02
3.67E-02
1.58E-02
3.03 E-01
Hardwood - Syracuse
10/27/20
16
3.04E-02
1.31 E-02
3.69E-02
1.59E-02
2.36E-01
Switchgrass - Low Load
11/3/201
6
62.6%
2.03E-01
8.74E-02
2.98E-01
1.28E-01
1.53E+00
Switchgrass - Low Load
11/4/201
6
1.43 E-01
6.14E-02
2.50E-01
1.08E-01
1.07E+00
Switchgrass - Syracuse
11/8/201
6
72.4%
1.13 E- 01
4.86E-02
1.49E-01
6.42E-02
8.50E-01
Switchgrass - Syracuse
11/9/201
6
8.92E-02
3.84E-02
1.29E-01
5.56E-02
6.72E-01
Switchgrass - Full Load
11/10/20
16
8 1 ^o/n
7.04E-02
3.03E-02
9.23 E-02
3.97E-02
5.31 E-01
Switchgrass - Full Load
11/15/20
16
O IJ /o
1.05E-01
4.53E-02
1.22E-01
5.26E-02
7.93E-01
D-l
-------
Table D-2. Test Average Total PM Mass Emissions Summary (Teflon Filters)
Average TP Emission Factor
Test Condition
Date
g/MJ Input g/MJ Output
g/kg fuel
Hardwood - Full Load
10/19/20
16
Void
Void
Void
Hardwood - Full Load
10/20/20
16
9.90E-03
1.11E-02
1.78E-01
Hardwood - Low Load
10/21/20
16
1.15E-01
2.53E-01
2.07E+00
Hardwood - Low Load
10/25/20
16
2.18E-01
1.94E-01
3.92E+00
Hardwood - Syracuse
10/26/20
16
1.10E-02
1.03 E-02
1.98E-01
Hardwood - Syracuse
10/27/20
16
8.90E-03
1.08E-02
1.60E-01
Switchgrass - Low Load
11/3/20
16
5.50E-02
8.08E-02
9.63E-01
Switchgrass - Low Load
11/4/20
16
9.47E-02
1.66E-01
1.66E+00
Switchgrass - Syracuse
11/8/20
16
3.07E-02
4.05E-02
5.36E-01
Switchgrass - Syracuse
11/9/20
16
5.45E-02
7.88E-02
9.53E-01
Switchgrass - Full Load
11/10/20
16
2.42E-02
3.17E-02
4.24E-01
Switchgrass - Full Load
11/15/20
16
5.53E-02
6.41 E-02
9.67E-01
D-2
-------
Table D-3. Test Average PM Number Emissions Summary (ELPI)
Emission Factor
Standard Deviation |
Emission Factor
Standard Deviation
Emission Factor
Standard Deviation
Emission Factor
Standard Deviation
Emission Factor
Standard Deviation
Test Condition
Date
part/MMBTU Input
part/MMBTU Input
part/MJ Input
part/MJ Input
part/MMBTU Out
part/MMBTU Out
part/MJ Output
part/MJ Output
part/kg fuel
part/kg fuel
Hardwood - Full Load
10/19/2016
4.25E+15
1.76E+15
4.03E+12
1.67E+12
3.68E+15
1.52E+15
3.49E+12
1.44E+12
7.24E+13
1.61E+06
Hardwood - Full Load
10/20/2016
3.70E+15
1.10E+15
3.51E+12
1.05E+12
4.13E+15
1.23E+15
3.92E+12
1.17E+12
6.31E+13
4.23E+06
Hardwood - Low Load
10/21/2016
7.20E+15
4.80E+15
6.83E+12
4.55E+12
1.58E+16
1.06E+16
1.50E+13
1.00E+13
1.23E+14
2.33E+05
Hardwood - Low Load
10/25/2016
1.70E+16
1.25E+16
1.62E+13
1.19E+13
1.52E+16
1.12E+16
1.44E+13
1.06E+13
2.91E+14
1.45E+04
Hardwood - Syracuse
10/26/2016
4.92E+15
5.25E+15
4.67E+12
4.98E+12
4.62E+15
4.92E+15
4.38E+12
4.67E+12
8.39E+13
1.14E+05
Hardwood - Syracuse
10/27/2016
4.05E+15
4.02E+15
3.84E+12
3.81E+12
4.91E+15
4.88E+15
4.66E+12
4.62E+12
6.90E+13
2.51E+05
Switchgrass - Low Load
11/3/2016
1.48E+16
7.45E+15
1.40E+13
7.07E+12
2.17E+16
1.09E+16
2.06E+13
1.04E+13
2.45E+14
6.60E+04
Switchgrass - Low Load
11/4/2016
1.29E+16
6.23E+15
1.23E+13
5.91E+12
2.27E+16
1.09E+16
2.15E+13
1.03E+13
2.15E+14
1.11E+05
Switchgrass - Syracuse
11/8/2016
6.72E+15
4.95E+15
6.37E+12
4.69E+12
8.87E+15
6.54E+15
8.41E+12
6.20E+12
1.11E+14
2.54E+05
Switchgrass - Syracuse
11/9/2016
1.11E+16
1.12E+16
1.05E+13
1.06E+13
1.61E+16
1.62E+16
1.52E+13
1.54E+13
1.84E+14
1.30E+05
Switchgrass - Full Load
11/10/2016
3.78E+15
1.03E+15
3.59E+12
9.76E+11
4.95E+15
1.35E+15
4.70E+12
1.28E+12
6.27E+13
6.54E+06
Switchgrass - Full Load
11/15/2016
4.88E+15
1.01E+15
4.62E+12
9.58E+11
5.65E+15
1.17E+15
5.36E+12
1.11E+12
8.09E+13
5.18E+06
D-3
-------
Table D-4. Test Average Elemental Carbon Emissions Summary (Manual NIOSH 5040)
Test Condition
Date
Test Average Elemental Carbon
Average EC
lb/MMBTU Input
Standard Dev.
lb/MMBTU Input
Average EC
lb/MMBTU Out
Standard Dev.
lb/MMBTU Out
Average EC
mg/kg fuel
Standard Dev.
mg/kg fuel
Hardwood - 100% Load
10/19/2016
1.72E-02
1.00E-02
1.49E-02
8.68E-03
1.33E+02
7.77E+01
Hardwood - 100% Load
10/20/2016
6.07E-03
1.41E-03
6.78E-03
1.57E-03
4.70E+01
1.09E+01
Hardwood - 25% Load
10/21/2016
2.0 IE-03
6.76 E-04
4.41E-03
1.49E-03
1.55E+01
5.23E+00
Hardwood - 25% Load
10/25/2016
3.16E-03
1.08E-03
2.82E-03
9.67E-04
2.44E+01
8.39E+00
Hardwood - Syracuse
10/26/2016
7.07E-04
1.06E-04
6.63 E-04
9.97E-05
5.48E+00
8.23 E-01
Hardwood - Syracuse
10/27/2016
1.93E-03
2.06 E-03
2.35 E-03
2.50E-03
1.50E+01
1.60E+01
Swilchgrass - 25% Load
11/3/2016
1.65E-03
2.40E-04
2.43E-03
3.53E-04
1.24E+01
1.81E+00
Switchgrass - 25% Load
11/4/2016
1.30E-03
2.32E-04
2.29E-03
4.06E-04
9.83E+00
1.75E+00
Switchgrass - Syracuse
11/8/2016
1.45E-02
1.54E-02
1.91 E-02
2.04 E-02
1.09E+02
1.16E+02
Switchgrass - Syracuse
11/9/2016
7.76E-03
4.05 E-03
1.12E-02
5.86E-03
5.84E+01
3.05E+01
Switchgrass - 100% Load
11/10/2016
3.0 IE-02
8.14E-03
3.94E-02
1.07E-02
2.26E+02
6.13E+01
Switchgrass - 100% Load
11/15/2016
4.74E-02
3.34 E-03
5.49E-02
3.87E-03
3.57E+02
2.51E+01
D-4
-------
Table D-5. Test Average Organic Carbon Emissions Summary (Manual NIOSH 5040)
Test Average Organic Carbon
Average OC
Standard Dev.
Average OC
Standard Dev.
Average OC
Standard Dev.
Test Condition
Date
lb/MMBTU Input
lb/MMBTU Input
lb/MMBTU Out
lb/MMBTU Out
mg/kg fuel
mg/kg fuel
Hardwood - 100% Load
10/19/2016
4.6 IE-04
NA
3.99E-04
NA
3.57E+00
Hardwood - 100% Load
10/20/2016
ND
ND
ND
Hardwood - 25% Load
10/21/2016
1.07E-01
5.68E-02
2.36E-01
1.25E-01
8.29E+02
4.40E+02
Hardwood - 25% Load
10/25/2016
1.70E-01
8.29E-02
1.52E-01
7.40 E-02
1.32E+03
6.42 E+02
Hardwood - Syracuse
10/26/2016
6.89 E-03
1.35E-03
6.46E-03
1.27E-03
5.33E+01
1.05E+01
Hardwood - Syracuse
10/27/2016
1.66E-03
2.3 5 E-03
2.0 IE-03
2.85 E-03
1.28E+01
1.82E+01
Swilchgrass - 25% Load
11/3/2016
8.35E-02
2.39E-02
1.23E-01
3.51 E-02
6.29E+02
1.80E+02
Swilchgrass - 25% Load
11/4/2016
6.82E-02
2.47E-02
1.20E-01
4.34E-02
5.14E+02
1.86E+02
Swilchgrass - Syracuse
11/8/2016
6.20 E-02
7.60E-02
8.18E-02
1.00E-01
4.67E+02
5.73E+02
Swilchgrass - Syracuse
11/9/2016
4.20E-02
2.87E-02
6.08E-02
4.15 E-02
3.16E+02
2.16E+02
Swilchgrass - 100% Load
11/10/2016
3.55E-03
1.87E-03
4.66E-03
2.44 E-03
2.68E+01
1.41E+01
Swilchgrass - 100% Load
11/15/2016
1.31E-02
5.10E-03
1.52E-02
5.92 E-03
9.87E+01
3.84E+01
D-5
-------
Table D-6. Test Average Optical Black Carbon Summary (Aethalometer)
Test Average Emission Factors
Average BC
Standard Deviation
Average BC
Standard Deviation
Average BC
Standard Deviation
Test Condition
Date
g/MJ Input
g/MJ Input
g/MJ Output
g/MJ Output
g/kg fuel
g/kg fuel
Hardwood - Full Load
10/19/20
16
1.21E-02
1.54E-02
1.05E-02
1.34E-02
2.18 E- 01
2.78E-01
Hardwood - Full Load
10/20/2016
7.90E-03
7.60E-03
8.83E-03
8.50E-03
1.42E-01
1.37E-01
Hardwood - Low Load
10/21/2016
0.00E+00
0.00E+00
0.00E+00
0.00E+00
Void
Hardwood - Low Load
10/25/201
16
8.08E-03
4.74E-02
7.21E-03
4.23E-02
1.45E-01
8.53E-01
Hardwood - Syracuse
10/26/201
16
2.20E-03
1.20E-02
2.06E-03
1.13E-02
3.96E-02
2.16E-01
Hardwood - Syracuse
10/27/20]
16
1.03E-03
3.1 1E-03
1.25E-03
3.77E-03
1.86E-02
5.59E-02
Switchgrass - Low Load
11/3/20]
16
3.74E-03
1.23E-02
5.49E-03
1.80E-02
6.54E-02
2.14E-01
Switchgrass - Low Load
11/4/20]
16
3.38E-03
2.08E-02
5.93E-03
3.65E-02
5.92E-02
3.64E-01
Switchgrass - Syracuse
11/8/20]
16
1.03E-02
2.39E-02
1.35E-02
3.15E-02
1.79E-01
4.18 E- 01
Switchgrass - Syracuse
11/9/20]
16
8.30E-03
2.00E-02
1.20E-02
2.89E-02
1.45E-01
3.50E-01
Switchgrass - Full Load
11/10/20]
16
2.09E-02
2.96E-02
2.74E-02
3.87E-02
3.65E-01
5.17 E- 01
Switchgrass - Full Load
11/15/20]
16
2.71E-02
3.63E-02
3.14E-02
4.21E-02
4.74E-01
6.35E-01
D-6
-------
Table D-7. Test Average Emission Factors for Selected and Toxic Metals (XRF)
Load Fuel Date Sample No.
Element Mass/Mass Total PM (ug/ug)
S CI Cr Mn Pb
Average Element Mass/Mass Total PM (ug/ug)
Sulfur
Average Std. Dev.
Chlorine
Average Std. Dev.
Chromium
Average Std. Dev.
Manganese
Average Std. Dev.
Lead
Average Std. Dev.
100% Wood 10/19/2016 1
2
3
10/20/2016 1
2
3
Void Void Void Void Void
Void Void Void Void Void
V oid V oid V oid V oid V oid
9.38E-02 1.11E-02 2.22E-05 5.74E-05 1.70E-04
1.30E-01 8.45E-03 2.68E-06 5.86E-05 1.00E-04
1.19E-01 1.58E-02 1.42E-05 1.05E-04 2.14E-04
1.14E-01 1.86E-02
8.35E-02 6.69E-02
8.61E-05 6.72E-05
5.34E-04 4.63E-04
1.14E-03 9.07E-04
25% Wood 10/21/2016 1
2
10/25/2016 1
2
1.98E-02 1.53E-03 3.83E-06 4.29E-05 0.00E+00
1.94E-04 1.76E-03 1.03E-05 5.09E-05 0.00E+00
3.09E-02 2.90E-04 1.15E-06 7.16E-05 0.00E+00
1.70E-02 1.55E-02
1.01E-02 1.20E-02
5.17E-05 7.50E-05
2.96E-04 3.42E-04
0.00E+00
1.35E-01 4.90E-03 7.92E-05 9.76E-05 2.56E-04
8.78E-04 7.59E-04 2.68E-06 2.33E-05 3.37E-05
Void Void Void Void Void
6.77E-02 9.46E-02
6.19E-02 6.84E-02
9.16E-04 1.23E-03
1.31E-03 1.26E-03
3.19E-03 3.66E-03
Syracuse Wood 10/26/2016 1
2
3
10/27/2016 1
2
9.36E-02 7.78E-03 1.42E-05 1.52E-04 1.91E-04
5.46E-02 1.25E-03 2.68E-06 6.39E-05 6.24E-05
2.88E-02 9.46E-04 1.03E-05 7.69E-05 0.00E+00
5.90E-02 3.26E-02
6.39E-03 6.34E-03
5.08E-05 7.63E-05
4.54E-04 5.26E-04
1.08E-04 1.86E-04
9.34E-02 5.54E-03 1.15E-06 1.33E-04 1.13E-04
6.51E-02 1.05E-03 6.51E-05 7.54E-05 0.00E+00
4.73E-02 1.19E-02 3.90E-05 2.19E-04 3.80E-04
6.86E-02 2.33E-02
1.05E-01 1.27E-01
4.30E-04 4.03E-04
2.17E-03 2.12E-03
3.05E-03 4.28E-03
25% Grass 11/3/2016 1
2
11/7/2016 1
2
3
Void Void Void Void Void
4.04E-03 6.41E-04 4.82E-05 0.00E+00 0.00E+00
2.99E-03 3.58E-03 1.30E-05 1.04E-04 0.00E+00
3.51E-03 7.41E-04
5.22E-03 7.39E-03
1.90E-05 2.69E-05
1.52E-04 2.15E-04
0.00E+00 0.00E+00
8.22E-03 1.79E-02 5.59E-05 3.90E-05 3.50E-04
7.17E-03 1.30E-02 5.47E-05 8.84E-05 3.13E-04
1.68E-03 4.42E-03 3.25E-03 2.74E-04 5.59E-05
5.69E-03 3.51E-03
8.29E-03 5.81E-03
7.20E-04 1.18E-03
8.64E-05 7.83E-05
1.67E-04 1.24E-04
Syracuse Grass 11/8/2016 1
2
3
11/9/2016 1
2
5.34E-02 7.08E-03 2.40E-04 2.72E-05 1.83E-04
1.53E-02 3.17E-02 0.00E+00 1.84E-05 7.00E-04
2.06E-03 2.32E-03 0.00E+00 3.14E-05 0.00E+00
2.36E-02 2.66E-02
9.20E-02 1.55E-01
3.50E-05 6.06E-05
6.56E-05 7.94E-05
2.02E-03 3.43E-03
4.88E-03 3.37E-03 6.51E-06 0.00E+00 5.97E-05
2.09E-02 2.34E-02 1.34E-04 3.79E-05 4.22E-04
9.89E-04 2.14E-03 1.37E-03 1.07E-04 0.00E+00
8.94E-03 1.06E-02
1.11E-02 1.23E-02
8.37E-04 1.32E-03
7.51E-05 9.68E-05
1.78E-04 2.45E-04
100% Grass 11/10/2016 1
2
3
11/15/2016 1
2
3
9.01E-03 4.13E-02 1.31E-04 3.64E-05 4.41E-04
1.65E-02 4.28E-02 0.00E+00 4.29E-05 3.11E-04
1.07E-02 3.72E-02 1.19E-05 6.51E-06 3.15E-04
1.21E-02 3.94E-03
6.93E-02 4.95E-02
3.58E-05 3.46E-05
3.54E-05 2.65E-05
5.77E-04 4.03E-04
Void Void Void Void Void
Void Void Void Void Void
6.71E-03 4.41E-02 2.33E-05 4.02E-05 4.62E-04
6.71E-03
4.84E-02
2.57E-05
4.42E-05
5.08E-04
Specific data for all elements measured are available electronically upon request.
D-7
-------
Table D-8. Test Average Particle-Phase SVOC Summary (GC/MS)
Test Average SVOC Emission Factors
Pellet Type
Test Condition
Date
lb/MMBTU Input
mg/MJ Input
lb/MMBTU Output
mg/MJ Output
mg/l
------- |