SCERP Project Number: A-6
Principal Investigator: JoAnn S. Lighty, PI; David W. Pershing,
Co-PI
University of Utah
1. Objectives of the project
The overall objective of this project was to develop low- cost technology to reduce toxic hydrocarbon emissions from domestic heating systems and other incineration systems fired with various fuels. The focus of years 1 and 2 was on pollutant emissions from domestic heating systems which currently burn a variety of fuels. Information was collected regarding the existing systems and wastes used in the border region and a systematic study was conducted to evaluate the NO, CO, and total unburned hydrocarbons (THC) from these systems. This project was a combined effort between faculty and students at the University of Utah (U of U) and the Technical Institute of Juarez (ITCJ). The study was limited to the El Paso/Juarez region.
2. Experimental design and procedure
Three different types of wood were obtained from the researchers at ITCJ. A study has been done to determine the amount of each wood burned in the region and we are awaiting these results from ITCJ. Analyses of the woods are presented in Table I. The three woods are labeled US Pallets, Mexican Pallets, and Particle Board; the particle board is actually covered with a plastic, simulated-wood laminate and is typically used in the construction of indoor furniture. As seen in Table 1, the Mexican and US Pallets are quite similar in moisture and ash contents. The US Pallets contain more volatiles (less fixed C) and more oxygen. Chlorine and nitrogen levels were low in the pallets. In contrast, the particle board pieces were quite different. The moisture, volatile, and oxygen contents were lower, while the ash content was higher. Of special note is the high nitrogen content, over 2% (dry basis), and the higher chlorine concentration. This nitrogen content is higher than that of coal.
The experiment was started by burning 1.0 kg of pine in the furnace, which has been previously described in numerous reports and the attached article. Once the furnace cooled down to a skin temperature of approximately 300deg.F, 1.0 kg the test of wood was placed in the hot furnace and allowed to combust; a second loading was added following a decrease in skin temperature. On-line analyses of THC, CO, O2, CO2, NO, and hydrocarbon speciation (via fast response GC/MS) were recorded; in addition, particles were collected using EPA Method 5H.
Mass balances were calculated using a carbon balance to determine the accuracy of the experimental procedure and measurements. The mass balance closures were normally between 1-15%.
Table 1. Analysis of Wood Pieces
US Pallets Mexican Particle
Pallets Board
Proximate
Analysis (%)
Moisture 5.43 5.41 4.80
Ash 0.29 0.23 0.66
Volatile 84.12 81.03 76.68
Fixed Carbon 10.10 13.33 17.86
Ultimate
Analysis (%,
dry basis)
Carbon 49.25 51.16 50.17
Hydrogen 6.09 6.23 6.13
Nitrogen 0.11 0.06 2.65
Sulfur 0.02 0.03 0.04
Ash 0.31 0.24 0.69
Oxygen 44.22 42.28 40.32
Chlorine <0.01 <0.01 0.23
3. Experimental Results
General Results
As previously reported, the pallets burned more slowly than the particle board; reloading was approximately every 25 minutes for the pallet pieces versus 20 for the particle board. This result is probably due to the reduced amount of moisture in the particle board. For the pallets, the CO concentration increased at the end of the pine burning, remained low during the first load of pallet burning, and then increased when the second load was added. The THC followed an opposite trend, increasing during the initial pallet burn and then decreasing. For the pallets, the emissions were relatively similar and NO was quite low. For the particle board, the CO peaked three times, corresponding to the three different loadings. The concentration of NO during the peak burning was approximately 4 times higher for the particle board versus the pallets. This result is consistent with the increase in the amount of fuel nitrogen in this wood as compared to the others. It is difficult to evaluate individual peaks for these compounds; therefore, the NO, CO, and THC traces were integrated with respect to time to give a total amount emitted per kg of wood. These results are shown in Table 2.
The results of Table 2 indicate that the emissions of THC were relatively similar for the three types of wood, given the experimental variation, which is also shown in the table. The particle board had a larger quantity of CO versus the other two woods. Other experiments (Summit, et al. 1993) have indicated that the particle board clearly has larger emissions of CO while THC is relatively consistent between the three woods. Therefore, it appears as though the particle board is a higher contributor to CO emissions than the other wood fuels. The particle board is clearly a dominant source of NO versus the other two woods. As previously mentioned, we are anticipating data from our Mexican colleagues detailing the amount of each type of wood used in the border region so that we can put these numbers in the context of usage. Finally, it is difficult to translate these numbers into pure emission factors. Variations are numerous and to determine one experimental and residential emission factor would be quite difficult. The experiments were conducted to compare the three woods, not develop these factors. For comparison, the emission factors from AP-42 (USEPA, 1985) are shown. The CO numbers and NO numbers are relatively consistent between the pallets and the emission factors for residential wood burning. The THC number is rather high from AP-42 but the source states that there is a lot of uncertainty in this number.
Table 2. Integrated CO, THC, and NO Emissions Data as a
Function of Waste Wood Fuel
Number of CO THC NO
Experiment (g/kg (g/kg (g/kg
s wood) wood) wood)
US Pallet 2 60. +/- 4.7 +/- 0.3 0.8 +/- 0.1
3
Mexican 2 37. +/- 3.5 +/- 1.3 0.7 +/- 0.1
Pallet 9
Particle 2 95. +/- 3.1 +/- 0.6 3.7 +/- 0.1
Board 3.5
Residential - 61.1 95.1 0.9
wood comb.
USEPA,
1985
Particle and Gas Speciation Data
Since the emissions of CO and THC did not show depreciable differences, particles were collected and analyzed and the flue gas was analyzed using GC/MS techniques to determine if certain species were present in the gas.
Soot (particle) samples were collected from the chimney using EPA Method 5H. The samples were then examined under transmission electron microscope (Jeol Electron Microscope JEL-200C). Magnifications from 33,000 to 300,000 were used. The particle area, perimeter, and equivalent diameter were determined. The mass concentrations of soot were determined by classical gravimetric method. The parameters obtained for the three different woods are presented in Table 3.
The run for the US Pallets had a low CO2 concentration during the first loading of the wood, when the particle data were taken. Analysis of the second loading, which exhibited normal CO2 concentrations showed a slightly higher mass concentration, 2.1 mg/ft3 and particle number density 13.8e5 1/m3. The results for the Mexican Pallets are surprising. Clearly one would expect these results to more closely follow those of the US Pallets given the data previously discussed similarities and the fact that the ash contents are similar. Prior to completing more experiments, we would like to obtain the data from Mexico showing the relative importance of each of the wood in terms of usage. If the numbers warrant further investigation, then we will complete more experiments on this wood to verify our findings.
In general, however, the US pallets appear to generate less particulate
than the other woods. In contrast, the other woods generated much larger
quantities of particles with a smaller size range. The US pallets did have
the lowest amount of fixed C, indicating that in fact less soot might be
formed. The aggregate size for the US pallets and Mexican pallets/particle
board differ considerably. This result indicates that the soot aggregate
growth rate was significantly higher during the US pallet combustion. The
building blocks of an aggregate (individual spherical particles) were also
larger. The aggregates were simple, rather short chains. Soot material
formed from Mexican pallets and particle board was more complicated with
many side chains built from small spherical particles. Smaller building
blocks were probably a result of more intensive combustion of component
particles of the aggregates. The smaller particles are more dangerous for
living organisms since they can be more easily transported through the
natural respiratory systems.
Table 3. Particle Data for the Fuels
Mass Particle Aggregate
Concentration Number Size Range
(mg/ft3) Density (nm)
(1/m3)
US Pallets 1.3 8.9e5 1000-1300
Mexican 3.3 28.9e5 800-1000
Pallets
Particle 3.6 36.1e5 700-950
Board
Further analysis was done on the particles to determine the hydrocarbons present on the particles. These compounds can form via a variety of pathways. As fuel-hydrocarbons undergo pyrolysis and combustion, major gases, such as CO2, H2O, and CO, are formed along with light hydrocarbons. The lighter hydrocarbons can participate in secondary reactions that lead to the formation of single aromatic species. The sequential growth of polycyclic aromatic hydrocarbons (PAHs) can occur through the reaction of single aromatic species with C1, C2, and C3 hydrocarbons, leading to higher molecular weight aromatic compounds. PAH can either convert to soot through the addition of polyacetylenes and grow to small soot particles, can form relatively inactive polynuclear aromatics, or can undergo inter-conversion reactions to form other PAH. Each pathway combines a complex network of individual reactions.
The vapor phase measurements of PAHs were performed using a Finigan Mat 700 Ion Trap Detector GC/MS system. Gas samples were aspirated from the chimney at the same cross-section were the particles were collected. The solid phase measurements were performed on the soot carbonaceous material. A fragment of the fiberglass filter was thermally desorbed at 590K and analyzed using the same GC/MS system. The typical hydrocarbons identified during the combustion of the three types of wood are shown in Table 4 for both the particles and the vapor phase.
As shown in Table 4, no large hydrocarbon molecules were detected in the gas phase. The flue gas sampling location was below 500K in all the experiments, so it is likely that the large hydrocarbons had adsorbed onto the particles at this location. For the US pallets, the four major compounds identified, furan, benzene, toluene, and furaldehyde, are the usual products of cellulose combustion. Similar products were found in the vapor phase for the combustion of the Mexican pallets and the particle board. In addition, aliphatic hydrocarbons and toluene were found for the Mexican pallets and particle board. It is interesting to note that the chromatograms obtained in this experiment are similar to the ones obtained by others for samples collected at the US/Mexico border (McClennen and Dworzanski, 1992).
Only a few PAH compounds were identified in the soot material. It is suspected that resolution of the capillary column used was not effective in yielding all the large hydrocarbons present. In future experiments a new column will be used. No mutagenic PAH was found during the US Pallet combustion; however, benzo(a,e)pyrene and benzo(b)fluoranthene were present in the Mexican wood and particle board combustion. Significantly higher amounts of these hydrocarbons were found in the Mexican pallet samples. Ion count areas are shown in Table 5.
Table 4. Hydrocarbons in the Vapor Phase and on the Solids
Vapor-Phase Solid-Phase
Analysis Analysis
US Pallets Furan 1-
Phenylnaphthalene
Benzene Benzophenathrene
Toluene Rethene
Furaldehydes Pyrene
Mexican Pallets Furan 1-
Phenylnaphthalene
Benzene Benzo(g,h,i)peryle
ne
Furaldehydes Benzo(a,e)pyrene
Aliphatic HC (C8- Rethene
C24)
Benzo(b)fluoranthe
ne
Dibenzo(def,mno)ch
rysene
Pyrene
Particle Board Furan 1-Phenylnaphthalene
Benzene Dibenzo(def,mno)chrysene
Furaldehydes Phenanthrene
Aliphatic HC (C8- Benzo(a,e)pyrene
C24)
Benzo(b)fluoranthene
Rethene
Table 5. Estimated Ion Area Counts for Hydrocarbons on the
Particles
Mexican Pallets Benzo(a)pyrene 13,613
Benzo(e)pyrene 48,798
Benzo(b)fluoranthene64,680
Rethene 77,060
1-phenylnaphthalene12,492
Particle Board Benzo(a)pyrene 8,608
Benzo(e)pyrene 16,703
Benzo(b)fluoranthene4,577
Rethene 176,870
1-phenylnaphthalene183,701
Mutagenic benzo(g,h,i)perylene was detected only in the Mexican pallets (9,211 ion count area). The highest amounts of 1-phenylnaphthalene and rethene were found during particle board combustion. Pyrene was not found in the particle board solids; however, it was present in a high amount (298,012) in the Mexican pallet solids. A small amount of pyrene (2,079 ion count area) was found for the US pallet solids. In general, the US pallets yielded smaller amount of all the PAH compounds.
In these experiments we did not analyze for chlorinated hydrocarbons
in either the vapor phase or on the solids. Due to the larger amount of
chlorine in the particle board, if this fuel is found to be used extensively
in the border region, future experiments will include analyzing for chorinated
compounds.
4. Conclusions
The experiments have shown that the combustion of wood fuels in typical Mexican heater systems can lead to emissions of CO, hydrocarbons, and particles which contain some hydrocarbons. The particle board has slightly higher emissions of CO as compared to the other fuels. The loadings of CO for the other fuels, however, are on the order of those found with residential wood burning (USEPA, 1985).
It is apparent that the emission of NO from the particle board is extremely high as compared to the other fuels used. This is due to the high nitrogen content in this wood, as result of the urea-based resin that is used in the particle board and the plastic laminate. As previously mentioned, we are awaiting data from Mexico which will show the usage of this wood versus the others. If indeed the burning of this material is widespread, then it could possibly be a major contributor to the NO emissions in the border area.
Some limited experiments have been conducted with a state-of- the-art wood stove, but those results were inconclusive as to the NO and CO emissions; particles were not collected during these initial experiments. The data do suggest, however, that the NO concentrations from this stove are much less. The state-of-the-art wood stove is a two-stage combustion system which is operated lean in the chamber with secondary air injection prior to the exhaust. This staging has been shown to be effective in reducing NO for coal combustors and is presently a common control strategy.
A larger amount of particles was released during the combustion of the
particle board and the Mexican pallets. Future work will hopefully explain
the differences in the particle data for the Mexican versus US Pallets
when no other differences were found in gas-phase emissions. The particulate
contained numerous hydrocarbons, and in the case of the Mexican pallets
and the particle board, certain mutagenics were identified. In addition,
some of the hydrocarbons found have been previously identified by the Texas
Air Control Board during 1990 testing in the El Paso/Juarez region, namely,
phenanthrene and pyrene (Dattner, 1992).
5. References
Dattner, S., personal communication (1992).
McClennen. W. and J. Dworzanski, "PAH Identification at the US/Mexico Border," unpublished data (1992).
Summit, G. D., J. S. Lighty, D. W. Pershing, W. D. Owens, C. Diaz-Quiz, "Control of pollutant emissions from waste burning in residential heaters of the US/Mexico border region," presented at the 1993 Spring Meeting, Western States Section, The Combustion Institute, Salt Lake City, UT, March 1993.
USEPA, "Compilation of air pollutant emission factors,"Fourth Edition,
Office of Air Quality Planning and Standards, Research Triangle Park, NC,
AP-42.
The FY91 SCERP-supported phase of this project:: A-6
and H-7
The FY92 SCERP-supported phase of this project: A-6
The FY93 SCERP-supported phase of this project:: AQ93-4
The FY96 SCERP-supported phase of this project:: AQ96-5
Last updated 7/1/99