Apr 25, 2012 - Dalton Research Institute, Manchester Metropolitan University, Manchester, ... Research, Missouri University of Science and Technology, Rolla,.
Article pubs.acs.org/est
Polycyclic Aromatic Hydrocarbon Emissions from the Combustion of Alternative Fuels in a Gas Turbine Engine Simon Christie,*,† David Raper,† David S. Lee,† Paul I. Williams,‡ Lucas Rye,§ Simon Blakey,§ Chris W. Wilson,§ Prem Lobo,∥ Donald Hagen,∥ and Philip D. Whitefield∥ †
Dalton Research Institute, Manchester Metropolitan University, Manchester, M1 5GD, United Kingdom School of Earth, Atmospheric and Environmental Science, University of Manchester, Manchester, M13 9PL, United Kingdom § Department of Mechanical Engineering, University of Sheffield, Sheffield, S11 8JG, United Kingdom ∥ Centre of Excellence for Aerospace Particulate Emissions Reduction Research, Missouri University of Science and Technology, Rolla, Missouri 65409, United States ‡
ABSTRACT: We report on the particulate-bound polycyclic aromatic hydrocarbons (PAH) in the exhaust of a test-bed gas turbine engine when powered by Jet A-1 aviation fuel and a number of alternative fuels: Sasol fully synthetic jet fuel (FSJF), Shell gas-to-liquid (GTL) kerosene, and Jet A-1/GTL 50:50 blended kerosene. The concentration of PAH compounds in the exhaust emissions vary greatly between fuels. Combustion of FSJF produces the greatest total concentration of PAH compounds while combustion of GTL produces the least. However, when PAHs in the exhaust sample are measured in terms of the regulatory marker compound benzo[a]pyrene, then all of the alternative fuels emit a lower concentration of PAH in comparison to Jet A-1. Emissions from the combustion of Jet A-1/GTL blended kerosene were found to have a disproportionately low concentration of PAHs and appear to inherit a greater proportion of the GTL emission characteristics than would be expected from volume fraction alone. The data imply the presence of a nonlinear relation between fuel blend composition and the emission of PAH compounds. For each of the fuels, the speciation of PAH compounds present in the exhaust emissions were found to be remarkably similar (R2 = 0.94−0.62), and the results do provide evidence to support the premise that PAH speciation is to some extent indicative of the emission source. In contrast, no correlation was found between the PAH species present in the fuel with those subsequently emitted in the exhaust. The results strongly suggests that local air quality measured in terms of the particulate-bound PAH burden could be significantly improved by the use of GTL kerosene either blended with or in place of Jet A-1 kerosene.
■
INTRODUCTION Security of supply and growing environmental concerns are placing increasing pressure on the transport sector to diversify away from petroleum-derived fuels. In the European Union, directives are in place to encourage fuel source diversification and the production of fuels refined from renewable feed stocks. The most recent directive1 specifies that at least 10% of the energy used in each member state’s transport sector must come from renewable resources by 2020. This directive was introduced to amend concerns that targets specified in the previous directive2 were damaging the environment and causing social issues.3 © 2012 American Chemical Society
Within the aviation sector, the development and certification of alternative drop-in fuels are progressing at a rapid pace: a standard specification for aviation fuel containing synthesized hydrocarbons was approved by ASTM in 2009,4 Hydrogenated esters and fatty acids (HEFA), also often referred to as hydrotreated renewable jet (HRJ), qualified as a 50/50 blend with petroleum Jet A-1 in 2011,4 and the Commercial Aviation Alternative Fuels Initiative (CAAFI) anticipate fully synthetic Received: Revised: Accepted: Published: 6393
January 23, 2012 April 23, 2012 April 25, 2012 April 25, 2012 dx.doi.org/10.1021/es300301k | Environ. Sci. Technol. 2012, 46, 6393−6400
Environmental Science & Technology
Article
formed from the incomplete combustion of organic matter.15 As products of thermal decomposition, potential sources of PAH are wide-ranging and include forest fires,16 incineration,17 heating,18,19 transport,20,21 and even cooking.22 PAH compounds are also present in crude oil, coal deposits, and in the majority of processed liquid fuels.23 In general, the concentration of PAH found in a particular fuel is widely variable and dependent upon the original source of the crude oil. Nevertheless, the absolute concentrations of PAH in aviation fuel are typically much lower than those found in other fuels.24 Within the United Kingdom, emissions from road transport represent the largest single source category on the national PAH inventory with an aggregate contribution of 58% for the combined U.S. Environmental Protection Agency (EPA) group of 16 priority PAH compounds.25 In contrast, over the same period the main source categories for the PAH marker compound benzo[a]pyrene are residential and institutional combustion with a contribution of 48%, while the contribution from road transport is only 11%.25,26 Hence, the method of assessment used to establish PAH concentrations must be defined when comparing source categories. The toxicity of an individual PAH compound is highly structurally dependent, with different isomers varying from toxic to extremely toxic, carcinogenic, mutagenic, and teratogenic.27 The limit value for PAH in the environment is based upon the measurement of benzo[a]pyrene (BaP), which is used as a marker compound and indicator of the carcinogenic risk for ambient mixtures of PAH. The European Union fourth air quality daughter directive28 has set a limit value for PAH measured as an annual average of BaP collected on airborne particulate matter (PM10) of 1 ng/m3. The U.K. expert panel on air quality standards29 issued a report that recommends a maximum annual average concentration for BaP 4 times lower at 0.25 ng/m3. It is considered that exposure to this level of PAH would make the risk to human health insignificant. It is often reported that PAH speciation profiles are sourceindicative: that is, different types of combustion produce different quantities and speciated distributions of PAH compounds.18−22,30,31 Hence PAH compounds are potentially useful as chemical markers to identify and apportion the emissions from individual combustion sources by use of receptor models.32−34 In fact, the existence of source-specific PAH speciation profiles is the fundamental assumption on which the use of receptor models for the apportionment of PAH to their respective emission sources have been based. Nevertheless, many authors disagree with this hypothesis and argue that this approach is fraught with uncertainty because of the many transport and removal mechanisms (such as deposition, photolysis, radical reactions, and vaporization) that become increasingly important as the distance between the source and the point of sampling increases. 35,36 Furthermore, there is evidence that, in some cases, fuel composition may also directly affect the PAH emissions.37,38 Despite this lack of consensus, standard PAH speciation profiles and emission factors for many different combustion sources do exist in the literature.39 PAH source apportionment studies often invoke these speciation profiles and they are also the foundation on which many PAH emission inventories are built. Airports and aircraft are considered to be relatively minor emitters of PAH, and within the United Kingdom, emissions of PAH from aircraft are omitted from the national PAH
Fischer−Tropsch (FT) fuel to qualify in 2012. In support of the certification process, several demonstration flights of commercial aircraft burning various blends of Jet A-1 and either biomass or FT fuels have also been conducted.5 First-generation Fischer−Tropsch fuels such as gas-to-liquid (GTL) and coal-to-liquid (CTL) can offer security of supply. However, in a full “well to wake” environmental analysis, the greenhouse gas (GHG) emissions from the combustion of these fuels do exceed the equivalent emissions from the use of conventional jet fuel. 6 Nevertheless, the economics of production for GTL and CTL are favorable at current and predicted future oil prices.7 The perceived opportunity for these fuels is as an effective mechanism to dampen price volatility in the jet fuel supply market,8 although other authors argue that these fuels will have a limited effect on price volatility over the near term.9 Second-generation alternative jet fuels such as HEFAs produced from renewable sources have reduced life-cycle GHG emissions relative to petroleum-based jet fuel.6 In the future, HEFA fuels could play a central role within a basket of measures to help mitigate aviation’s contribution to climate change.10,11 There are, however, many significant challenges that must be overcome before second-generation alternative fuels are economically viable and widely available.12,13 Overall, and considering the current fuel technology readiness level,9 it would seem probable that GTL and CTL will be the first commonly available drop-in alternative fuels (with due recognition of the past and current market presence of Sasols South African Natref Refinery). These fuels may form a small but notable portion of aviation fuel requirements in the near future,14 while second-generation fuels are on a more midterm horizon. The combustion of fuel in a gas turbine engine is a highly efficient process. Nevertheless, there is no reason to assume that the emissions from the combustion of alternative drop-in fuels in gas turbine engines will be identical to those from the combustion of Jet A-1. Accordingly, the introduction and use of GTL and CTL as alternative fuels in aviation may bring with them a shift in emphasis for the priority pollutant compounds in the emissions. Our atmosphere is a highly sensitive environment with complex and far-reaching interactions. As such, due diligence must require that changes in the emission profile for the combustion of alternative fuels in terms of non-CO2 effects should come under scientific scrutiny ahead of the large-scale introduction of new fuels on climate or economic grounds. The objective of this work is to specifically investigate the relative emission of polycicylic aromatic hydrocarbon (PAH) compounds. We report here on the speciated PAH profiles from a test-bed gas turbine engine when powered by Jet A-1, Sasol fully synthetic jet fuel, Shell gas-to-liquid kerosene, and Jet A-1/ GTL (50:50 by volume) blended kerosene. Speciated PAH profiles for each of the fuels are also given. The scope of the test program corresponds to intermediatescale combustion trials that fit into the range of possible analysis between laboratory bench-scale testing of fuels (requiring ∼1 L of fuel) and full engine tests (requiring ∼1000 L of fuel).
■
BACKGROUND Polycyclic aromatic hydrocarbons (PAH) are some of the most ubiquitous and widespread organic pollutants in the environment. Most PAH compounds found in the environment are 6394
dx.doi.org/10.1021/es300301k | Environ. Sci. Technol. 2012, 46, 6393−6400
Environmental Science & Technology
Article
constructed with a minimum bend radius of 10 internal diameters was not heated. The sampling head and filter element were positioned a further 3 m downstream of the probe tip. Exhaust gases were extracted at a constant 300 L/min set by critical aperture, and the temperature of the exhaust at the filter medium was typically 45 °C (approx 40 ± 2 °C at the start of the stabilized sampling period and 50 ± 2 °C by the end of sampling). The PM samples were collected on 47 mm Pall A/E glass fiber filters (Pall Gelman, Ann Arbor, MI). These filters have a high retention efficiency and are recommended by the EPA for high-volume sampling to collect atmospheric particulates and aerosols.46 Nominal pore size is 1 μm and nominal thickness is 330 μm. These glass fiber filters have a dioctyl phthalate test collection efficiency of 99.98% at 0.3 μm47 and a maximum operating temperature in air of 550 °C. All filters were accurately weighed at constant humidity to a precision of 1 μg prior to sampling. Similarly, all filters were accurately weighed at constant humidity following collection of a sample. Sampling with glass fiber filters will result in a reduced efficiency of capture for the low molecular weight PAH species (such as naphthalenes) due to the relatively high vapor pressure of these species, compounded by the elevated temperature at the sampling head. In this sense, the results reported here cannot be considered absolute. Nevertheless, this sampling medium was selected as it is directly comparable with current environmental PAH sampling practice.48 Exhaust PM sampling at the selected test conditions was conducted over a 6-min experimental window that commenced once the APU had stabilized. Determination of PAH in Parent Fuels. A sample of each parent fuel was analyzed for the presence of PAH compounds. For each fuel, a 1% dilution in dichloromethane was prepared and spiked with quantification and syringe standards: The quantification standard consisting of 100 ng each of naphthalene-d8, acenaphthylene-d8, acenaphthene-d10, fluorene-d10, phenanthrened10, anthracene-d10, fluoranthene-d10, pyrene-d10, benz[a]anthracene-d12, chrysene-d12, benzo[b]fluoranthene-d12, benzo[k]fluoranthene-d12, indeno[123-cd]pyrene-d12, dibenzo[ah]anthracene-d14, benzo[ghi]perylene-d12, and 2100 ng of benzo[a]pyrene-d12. The syringe standard consisting of 100 ng of 9methylanthracene-d12 and 2000 ng of perylene-d12. The spiked fuel samples were analyzed on an Agilent 6890/5973 gas chromatograph−mass spectrometer running in selected ion monitoring mode. Determination of PAH in Exhaust Emission Sample. Exhaust PM samples were collected for each of the different fuels at three different engine conditions. However, because of the limited sampling time at each engine condition (due to limited fuel) it was necessary to bulk the samples from the three different engine conditions together to form a composite sample for each fuel type to obtain sufficient sample mass for PAH analysis. For this reason, the concentration and speciation of PAH compounds in the APU exhaust emissions are categorized by fuel type only. Each of the composite filter samples was spiked with a quantification standard consisting of 1000 ng each of n8hthalened8, acenaphthylene-d12, acenaphthene-d10, fluorene-d10, phenanthrene-d10, anthracene-d10, fluoranthene-d10, pyrene-d10, benz[a]anthracene-d12, chrysene-d12, benzo[b]fluoranthene-d12, benzo[k]fluoranthene-d12, indeno[123-cd]pyrene-d12, dibenzo[ah]anthracene-d14, benzo[ghi]perylene-d12, and 21 000 ng of benzo[a]pyrene-d12. Each sample was Soxhlet-extracted with
inventory although emissions from aircraft ground-support vehicles are present.23 Nevertheless, some apportionment studies have identified aircraft as potentially sizable contributors to local PAH burdens,33 and airport air quality measurements have plainly identified a sizable source contribution from aircraft.40,41 While PAH speciation profiles and emission factors for aircraft gas turbine engines do exist in the literature,39 it is also recognized that the available data are sparse and appreciable differences in the mass of emitted PAHs occur for different engine types, sampling schemes, fuel types, and engine operating conditions.42 To our knowledge, the results presented here represent the first comparative assessment of PAH emissions from the combustion of alternative fuels in a gas turbine engine.
■
EXPERIMENTAL DETAILS Engine. A series of experiments were performed at the University of Sheffield’s Low Carbon Combustion Centre to study the relative emissions from gas turbine engine when powered by alternative fuels. The test-bed gas turbine engine was an Artouste Mk113 auxiliary power unit (APU), which has a two-stage turbine connected to a centrifugal compressor through a single shaft. The engine test-bed facility provided an ideal experimental platform to evaluate the performance of alternative fuels, as despite the apparent age of the engine (ca. 1963), the simplicity of the hardware allowed for easy instrumentation and logging of key engine and thermodynamic parameters. However, for this engine the emission of products from incomplete combustion (including unburned hydrocarbons, PAHs, and PM) are relatively high in comparison to current engines that incorporate improved combustor design. Further details of this experimental setup have been reported elsewhere.43 The test program involved an idle condition and a full power condition, before the engine was returned to a “hot idle” condition prior to shutdown. Limited fuel availability prevented sampling at other test points. At these engine conditions, the air fuel ratio (AFR) and fuel flow for each of the different test fuels were matched to within the experimental error. At engine idle, the AFR was 81 ± 2 and the fuel flow was 16.0 ± 0.25 g/s (124 ± 2 lb/h). At full power, the AFR was 77 ± 1 and the fuel flow was 30.2 ± 0.13 g/s (240 ± 1 lb/h). The ambient air temperature remained reasonably constant throughout at 17.0 ± 0.5 °C. Hence we may assert that exhaust emission differences may be directly attributed to differences in fuel composition. Fuels. The test-bed APU was run on standard aviation Jet A-1 (Merox and hydrotreated combination) kerosene from Shell, a coal-derived (FSJF) fully synthetic jet fuel from Sasol, a gas-to-liquid (GTL) kerosene from Shell (Sarasol 40), and a Jet A-1/GTL kerosene blended 50:50 in-house by volume. Each of these fuels either meets or exceeds the relevant fuel specifications: Def Stan 91-91,44 ASTM D1655,45 and ASTM D-7566,4 with the exception of the 100% GTL, which is currently not certified for use as an unblended fuel. Sasol FSJF has a “near zero” sulfur content and a minimum of 8% aromatic content. Shell GTL has a “near zero” content for both sulfur and aromatics. Particulate Matter Sampling. A high-volume collection probe (38 mm internal diameter stainless steel) aligned with the axis of the engine and positioned at a downstream distance of 3 m from the exit nozzle was used to collect a particulate matter (PM) sample from the exhaust. The sample line 6395
dx.doi.org/10.1021/es300301k | Environ. Sci. Technol. 2012, 46, 6393−6400
Environmental Science & Technology
Article
dichloromethane for a minimum of 14 h. The extract was reduced to 10 mL in volume by nitrogen blow-down and then chromatographically cleaned up. The final extract was concentrated to 1 mL in volume and spiked with a syringe standard (100 ng of 9-methylanthracene-d12 and 2000 ng of perylene-d12). The extracted samples were analyzed on an Agilent 6890/5973 gas chromatograph−mass spectrometer running in selected ion monitoring mode.
Table 1. Concentrations of Individual PAH Compounds in Parent Fuelsa PAH concn (μg/mL)
■
RESULTS PAH in Parent Fuels. The concentrations of PAH compounds in each of the parent fuels are given in Table 1. The results have been calculated for each PAH compound by use of the closest eluting deuterated PAH in the quantification standard. Limits of detection were calculated from the analysis of a blank sample. Where individual PAH compounds in the sample are below the limit of detection, data are prefixed with the < symbol and are omitted from the calculation of total PAH concentration. Uncertainty in the measurement of individual compounds is estimated to be ±20%. This estimate is consistent with a cross-checking comparison between the experimentally determined concentration of PAHs in the GTL/Jet A-1 blended fuel and the expected concentration calculated from the Jet A-1 and GTL fuel fractions. The results show that all of the fuels contain PAH compounds, although in different concentrations and with different speciation. The concentration of individual PAH compounds may typically change by more than 2 orders of magnitude between GTL and Jet A-1. In Table 1, the total PAH concentration is the sum of all PAH species measured. For each of the fuels, the total concentration of PAHs is dominated by the presence of naphthalene compounds that account for approximately 90% of that total. The summary data at the end of Table 1 highlight the appreciable differences in the concentrations of PAHs in the different fuels and their dependence upon the method of the assessment. Overall, Jet A-1 contains the highest total concentration of PAHs while GTL contains the least; FSJF contains the highest concentration of PAHs measured in terms of the EPA16 grouping while GTL contains the least; whereas GTL was the only fuel to contain a measurable level of benzo[a]pyrene. Hence the results indicate that BaP may not be an appropriate PAH marker compound in this application. An intercomparison of the individual PAH species present in each of the fuels was also considered. It was found that the speciated PAH compounds in the FSJF fuel do not correlate with those in the other fuels. In contrast, a degree of similarity between the speciated PAH compounds present in the GTL and Jet A-1 fuels does exist [R2 = 0.855 (n = 4)], although the number of data points is low and so may not be considered robust. Clearly a relationship between the speciated PAHs in the Jet A-1/GTL blended fuel and its constituent components is expected. A summary of all the R2 correlation data from a linear regression fit is given in Table 3. PAH in Exhaust PM Sample. The concentration of PAH compounds in the collected exhaust samples for each of the fuels are given in Table 2. The results for each PAH compound have been calculated by use of the closest eluting deuterated PAH in the quantification standard. Limits of detection were calculated from the analysis of a blank sample. Where individual PAH compounds in the sample are below the limit of detection, data are prefixed with the ’ symbol and are omitted from
PAH in fuel sample
Jet A-1
FSJF
GTL
GTL/Jet A-1 blend
naphthalene 2-methylnaphthalene 1-methylnaphthalene biphenyl acenaphthylene acenaphthene fluorene phenanthrene anthracene 2-methylphenanthrene 2-methylanthracene 1-methylanthracene 1-methylphenanthrene 9-methylanthracene 4,5methylenephenanthrene fluoranthene pyrene retene benzo[c]phenanthrene benzo[a]anthracene chrysene cyclopenta[c,d]pyrene benzo[b]naph[2,1-d] thiophene 5-methylchrysene benzo[b+j]fluoranthene benzo[k]fluoranthene cholanthrene benzo[e]pyrene benzo[a]pyrene perylene indeno[1,2,3-cd]pyrene dibenzo[ah/ac] anthracene benzo[ghi]perylene anthanthrene dibenzo[al]pyrene coronene dibenzo[ae]pyrene dibenzo[ai]pyrene dibenzo[ah]pyrene
900 1500 1200 340 21 14 31 4.5
