11.07.2014 Views

Study of Air Toxics Released from the Pre- Harvest Burning of ...

Study of Air Toxics Released from the Pre- Harvest Burning of ...

Study of Air Toxics Released from the Pre- Harvest Burning of ...

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

A&WMA International Specialty Conference<br />

Leapfrogging Opportunities for <strong>Air</strong> Quality Improvement<br />

<strong>Study</strong> <strong>of</strong> <strong>Air</strong> <strong>Toxics</strong> <strong>Released</strong> <strong>from</strong> <strong>the</strong> <strong>Pre</strong>-<br />

<strong>Harvest</strong> <strong>Burning</strong> <strong>of</strong> Sugarcane<br />

Danielle Hall, Jun Wang, Kuei-Min Yu, Krisha Capeto, Chang-Yu<br />

Wu, James Stormer, Guenter Engling, Yu-Mei Hsu<br />

May 11 th , 2010<br />

Department <strong>of</strong> Environmental Engineering Sciences<br />

University <strong>of</strong> Florida


Introduction: Sugarcane <strong>Burning</strong> Practice<br />

The pre-harvest burning <strong>of</strong> sugarcane<br />

is a common practice used to<br />

facilitate harvesting.<br />

– Removes unwanted biomass<br />

– Reduces snake and insect hazards<br />

– Concentrates sugar through water<br />

evaporation<br />

Palm Beach County’s 2008<br />

emissions inventory showed<br />

sugarcane pre-harvest burning<br />

contributed to:<br />

– 20% <strong>of</strong> VOC emissions<br />

– 48% <strong>of</strong> PM emissions<br />

– 22% <strong>of</strong> CO emissions<br />

– 11% <strong>of</strong> NO x emissions


Introduction: Sugarcane EFs<br />

Current EFs are based <strong>from</strong> one study <strong>of</strong> Hawaiian<br />

sugarcane (Darley, 1974) and are rated unreliable<br />

(category “D”) in AP-42.<br />

- Limited data set<br />

- Sugarcane <strong>from</strong> different areas may<br />

exhibit significant EF differences<br />

- Limited data available for specific HAPs


Objective<br />

Investigate <strong>the</strong> emission factors<br />

– Hazardous <strong>Air</strong> Pollutants<br />

Polycyclic Aromatic Hydrocarbons (PAHs)<br />

– 16 “priority PAH Pollutants” + 3 o<strong>the</strong>r PAH <strong>of</strong> concern.<br />

Carbonyls<br />

– Formaldehyde, acetaldehyde, propionaldehyde,<br />

crotonaldehyde, butyraldehyde, benzaldehyde, valeraldehyde,<br />

2,5-dimethylbenzaldyde<br />

Volatile Organic Compounds (VOCs)<br />

– Benzene, toluene, ethylbenzene, o,m,p-xylenes, styrene<br />

– PM 2.5<br />

Elemental Carbon (EC)<br />

Organic Carbon (OC)


Methodology: Chamber Design<br />

A combustion chamber used to simulate<br />

field burning.<br />

Stack sampling methods used.<br />

Stack velocity and chamber flowrate was<br />

determined following EPA Method 2.<br />

– <strong>Pre</strong>ssure drop and temperature were measured<br />

with a s-type pitot tube and <strong>the</strong>rmocouple across<br />

a horizontal traverse.<br />

CO and CO 2<br />

flue gases were continuously<br />

monitored to determine <strong>the</strong> combustion<br />

efficiency<br />

MCE =<br />

[ ]<br />

[ ]<br />

∆ CO 2<br />

∆[ CO]+ ∆ CO 2


Methodology: Sampling<br />

Two experimental conditions tested:<br />

•Dry sugarcane leaves<br />

-Feed rate ~ 100g / 40 sec<br />

•Whole sugarcane stalks<br />

(containing wet + dry leaves)<br />

-Fed to maintain near constant<br />

burning conditions.<br />

-Heterogeneous nature <strong>of</strong> biomass<br />

led to more variable combustion<br />

conditions<br />

Combustion chamber


Methodology: PAHs<br />

Sampling and analysis based<br />

<strong>from</strong> EPA Method TO-13A<br />

(adapted for stack sampling).<br />

PAHs isokinetically sampled and<br />

collected on quartz filters and<br />

PUF/XAD-2 resin cartridges.<br />

Filters and cartridges were sent<br />

to Columbia Analytical Services<br />

(CAS) where <strong>the</strong>y were Soxhlet<br />

extracted, concentrated, and<br />

analyzed by GC/MS.<br />

Filter<br />

PUF/XAD-2<br />

cartridge<br />

holder<br />

Sampling<br />

nozzle


Methodology: Carbonyls & VOCs<br />

CARBONYLS:<br />

Sampling and analysis based<br />

on EPA Method TO-11A.<br />

– DNPH sorbent cartridges with KI<br />

ozone scrubbers<br />

Samples extracted with<br />

acetonitrile, and analyzed by<br />

HPLC (performed by CAS).<br />

Ozone<br />

DNPH<br />

Scrubber<br />

cartridge<br />

VOCs:<br />

• Gas samples collected in<br />

Tedlar bags via negative<br />

pressure.<br />

• Samples analyzed for BTEX<br />

& styrene by GC/MS<br />

(performed by CAS)<br />

Sample<br />

Probe<br />

Vac-U-<br />

Chamber<br />

Tedlar bag<br />

Teflon sampling<br />

line<br />

Exhaust<br />

port<br />

Carbonyl Sampling System<br />

Vac-U-Chamber and Tedlar bag


Methodology: PM 2.5<br />

PM 2.5 sampling followed EPA O<strong>the</strong>r Test Methods 27 & 28 (modified)<br />

– Cyclone used to separate particles based on size.<br />

– Filterable PM 2.5 collected on a glass fiber filters and tissuquartz filters (for<br />

EC/OC analysis)<br />

– Condensable particulate matter (CPM) collected in a dry impinger train<br />

and on Teflon CPM filter.<br />

Glass fiber filters were pre- and post weighed.<br />

Impingers and <strong>the</strong> Teflon filter rinsed with water and solvent to collect CPM.<br />

The extracts were evaporated and <strong>the</strong> remaining residue (CPM) was<br />

weighed.<br />

EC/OC fraction <strong>of</strong> <strong>the</strong> PM 2.5 was determined using an OCEC Carbon<br />

Aerosol Analyzer (Sunset Laboratory) following NIOSH method 5040.<br />

• Analyzed at <strong>the</strong> Research Center for Environmental Changes, Academia<br />

Sinica, Taipei, Taiwan


Emission Factor Calculation<br />

EF (mg/ kg) = ∆C ×Q × t<br />

x<br />

m<br />

C x = compound concentration (in excess <strong>of</strong> background)<br />

Q= flowrate through chamber<br />

t= time <strong>of</strong> sampling<br />

m= mass <strong>of</strong> sugarcane burned


Results: Chamber CE<br />

Combustion efficiency ranges <strong>from</strong> 80-100%, with an<br />

average around 98.5%flaming combustion<br />

Table: Sugarcane CO and CO 2 EF<br />

comparison<br />

MCE (%)<br />

CO EF<br />

(g/kg)<br />

CO 2<br />

EF<br />

(g/kg)<br />

1255±28<br />

7<br />

<strong>Pre</strong>sent <strong>Study</strong> 98.5±0.2 9.2±3.3<br />

AP-42<br />

(Darley, 1974) NA 30-40 NA<br />

Yokelson et<br />

al., 2008 97.6 28.3 1838<br />

Figure: Real-time flue gas concentrations


Results: PAHs<br />

The average PAH EFs were 7.13 ± 0.94 mg/kg (n=4) and 8.18 ± 3.26<br />

mg/kg (n=3) for dry and whole stalk experiments, respectively.<br />

Emissions dominated by low molecular weight compounds.<br />

– 2-ring PAH compounds comprise 66%<br />

– 3-ring PAH compounds comprise 27%


Results: PAHs (cont’d)<br />

PAH EFs are comparable, but on <strong>the</strong> low end <strong>of</strong> o<strong>the</strong>r EFs<br />

reported for agricultural residue burning.<br />

Consistent dominance <strong>of</strong> phenanthrene and acenaphthylene<br />

compounds.


Results: PAHs (cont’d)<br />

PAH compound ratios were found that can<br />

possibly serve as source markers for source<br />

apportionment studies.<br />

Table: Characteristic PAH ratios


Results: Carbonyls<br />

• Total carbonyl EFs were 231.8±52.3 mg/kg (n=5) and<br />

909.6±527.7 mg/kg (n=4) for dry and whole stalk<br />

experiments, respectively.<br />

• EFs for whole stalk experiments exhibited more<br />

variability and were higher than dry leaf experiments.<br />

– Variable combustion conditions<br />

– Biomass composition<br />

Moisture content may inhibit<br />

complete combustion leading to<br />

higher pollutant emissions.<br />

Sugarcane sources also differed<br />

– may have different treatment<br />

practices (i.e., fertilizer and<br />

pesticide application)<br />

Table: Comparison <strong>of</strong> Combustion<br />

Conditions<br />

Experiment<br />

<strong>Burning</strong> Rate<br />

Average<br />

EF<br />

Average<br />

Temp<br />

(°F)<br />

Dry Leaves 232±52 311 1 kg/10 min<br />

Whole Stalks 482±16 600 1 kg/3 min<br />

(test 1)<br />

Whole Stalks 1401±166 145 0.24 kg/4 min<br />

(test 2)


Results: Carbonyls (cont’d)<br />

Emissions dominated by low molecular weight<br />

compounds


Results: VOCs<br />

Benzene and toluene dominate VOC emissions.<br />

– Benzene/toluene ratio was 0.32, which may be a unique marker<br />

pattern.<br />

Comparable to EFs for almond and walnut prunings.


Results: VOCs (cont’d)<br />

In general, VOC EFs are lower than o<strong>the</strong>r reported<br />

EFs, including those for sugarcane.<br />

Differences attributed to:<br />

– Measurement technique<br />

– Sugarcane source, condition, and burning characteristics (i.e., CE)<br />

Table: VOC EF (mg/kg) comparison


Results: PM 2.5<br />

CPM was not statistically higher (p=0.27) in <strong>the</strong><br />

sample than in <strong>the</strong> method blanks neglected<br />

The average PM 2.5 EF was 2.49±0.66 g/kg (n=4)<br />

for dry leaf experiments.<br />

Agrees very well with current EFs and o<strong>the</strong>r<br />

agricultural burning studies.<br />

Table: PM EF (mg/kg) comparison


Results: EC/OC<br />

EC emissions dominated OC emissions.<br />

Sugarcane EC emissions are high compared to<br />

o<strong>the</strong>r studies and OC emissions are low<br />

– Function <strong>of</strong> high CE and biomass composition<br />

Unique trend may be helpful for source<br />

apportionment studies.<br />

Table: EC and OC EF comparison


HAPs Emission Estimates<br />

HAPs emissions were estimated and compared<br />

with <strong>the</strong> 2005 National <strong>Air</strong> <strong>Toxics</strong> Assessment Data<br />

for Palm Beach County (2005) and <strong>the</strong> state <strong>of</strong> FL.<br />

*Disclaimer: <strong>the</strong>se estimates and statements do not<br />

represent <strong>the</strong> conclusions <strong>of</strong> <strong>the</strong> Palm Beach County<br />

Health Department.<br />

Inputs:<br />

– <strong>the</strong> upper limit EF <strong>of</strong> <strong>the</strong> 95% confidence interval<br />

– Assumed 335,650 acres <strong>of</strong> sugarcane burned (based on<br />

2008)<br />

– Fuel loading = 7 tons/acre


HAP Emission Inventory Estimates


HAP Emission Inventory Estimates (cont’d)


HAP Emission Inventory Estimates (cont’d)


Summary & Conclusions<br />

The data <strong>from</strong> this research fur<strong>the</strong>r validate and expand <strong>the</strong><br />

current AP-42 emission factors.<br />

– EFs are expected to highly variable during <strong>the</strong> fire event and throughout<br />

harvesting season—dependant on burning conditions and biomass<br />

conditions.<br />

Marker and tracer compounds and patterns identified can be<br />

used in future source apportionment studies to allocate ambient<br />

pollution to specific sources.<br />

With a more reliable and comprehensive understanding <strong>of</strong> <strong>the</strong><br />

emissions <strong>from</strong> sugarcane pre-harvest burning, regulators can<br />

make better decisions about <strong>the</strong> permitting and management <strong>of</strong><br />

this practice to better protect human health and <strong>the</strong> environment.


Thanks for your attention!<br />

ANY QUESTIONS?

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!