THE IMPACT OF ETHANOL BLENDED FUEL ON VEHICLE EMISSIONS OF VOLATILE ORGANIC COMPOUNDS

Transcription

1 THE IMPACT OF ETHANOL BLENDED FUEL ON VEHICLE EMISSIONS OF VOLATILE ORGANIC COMPOUNDS Anne Tibbett 1, Nicholas Coplin 2, Merched Azzi 1, John Carras 1 1 CSIRO Energy Technology, New Illawarra Rd, Lucas Heights, NSW, 2234, Australia, anne.tibbett@csiro.au 2 Orbital Corporation Limited, 4 Whipple St, Balcatta, WA, 6914, Australia, ncoplin@orbitalcorp.com.au Abstract A comprehensive study was undertaken by CSIRO and Orbital Australia, commissioned by the Australian Department of Environment, Water, Heritage and the Arts (DEWHA), to determine the effect of ethanol blended gasoline fuel on vehicle exhaust and evaporative emissions, and associated impacts on air quality and health. The program was undertaken using in-service, light duty petrol vehicles, representative of the current Australian fleet. The vehicles were tested for exhaust emissions over the recently developed Australian real world urban drive cycle, and for evaporative emissions under current ADR emissions certification protocols. This paper delivers the results of the emissions characterisation component of this study; specifically the speciation of over 1 volatile organic compounds, as air toxics and those contributing to the formation of ozone in the atmosphere, and the impact of ethanol blended fuel on vehicle emission rates. Keywords: VOCs, air toxics, ethanol fuel, E1, petrol, exhaust, evaporatives 1. Introduction The issues surrounding ethanol blended gasoline as a motor vehicle fuel has attracted attention in Australia over recent years, with a number of studies undertaken on various aspects of its use (Brown et.al, 1998, Orbital Australia, 23). In 25 the Federal Government established a Biofuels Taskforce whose report highlighted the need to assess the potential health impacts of the use of ethanol blended fuel in the Australian context (Biofuels Taskforce, 25). To this effect a comprehensive study was commissioned by the Australian Department of Environment, Water, Heritage and the Arts (DEWHA) to determine the effect of ethanol blended gasoline on vehicle exhaust and evaporative emissions and associated impacts on air quality and human health. The study was managed and conducted by Orbital Australia, a leading motor vehicle testing laboratory, and by CSIRO s air pollution specialist group. The work program consisted of a number of components, each measuring key data required for the health assessment of the use of 5% and 1% v/v ethanol blended fuel (E5 and E1). In overview, this involved the measurement of regulated and unregulated components in vehicle exhaust and evaporative emissions using a test group of vehicles selected to represent the Australian passenger vehicle fleet, the examination of ambient secondary organic aerosol formation associated with the use of these fuels, the modelling of photochemical smog formation and, based on these data, assessment of the health impact as a cost benefit analysis. The study was completed in 28 and the full report can be found on the DEWHA website (CSIRO and Orbital, 28). This paper reports the component of the study which addresses the characterisation of organic compounds in vehicle exhaust and fuel evaporatives, as air toxics and those contributing to the formation of ozone in the atmosphere, and the impact of ethanol blended fuel on vehicle emission rates for these compounds. 2. Speciation of Organics Vehicle emissions can be characterised as coming from two main sources; tailpipe exhaust emissions from the combustion of fuel and evaporative emissions from the release of fuel vapours. Volatile organic compounds (VOCs) in these emissions are found as C 1 to C 12 isomers of aliphatic and aromatic hydrocarbons, derived from combustion and from the raw fuel, and C 1 to C 8 aliphatic and aromatic carbonyl compounds, derived from

2 incomplete combustion and partial oxidation of the fuel. In order to determine the air quality and health impacts of these emissions, speciation of individual compounds is undertaken as each compound has a different reactivity in the formation of ozone and photochemical smog, and a different toxicity in its potential health effect. The compounds contributing to photochemical reactivity (known as the ozone precursor group of VOCs) have been identified for consideration under the Tier II Australian Air Toxics NEPM, and a listing of priority components has been specified under US EPA and California Air Resources Board (CARB) protocols for photochemical assessment of ambient air (known as the PAMS component listing). These compounds have incremental reactivity (MIR) factors associated with them, as developed by Carter (22) for use in measurement of ozone formation potential, and are required for photochemical modelling. The PAMS listing comprises around 6 compounds as isomers of C 2 to C 12 aliphatic and aromatic hydrocarbons and C 1 to C 3 carbonyls. Certain VOCs are ascribed priority in their potential to be hazardous to human health (known as air toxic components). US EPA and CARB agencies identify many compounds in this category. In this study, the target VOCs were those prioritised for assessment under the Air Toxics NEPM (i.e. benzene, toluene, xylenes and formaldehyde) and selected NEPM Tier II priority organics (1,3-butadiene, styrene and acetaldehyde). Ethanol and minor alcohols were included in the emissions assessment. Although they are not currently considered a NEPM priority air toxic, and have only modest photochemical reactivity, they are measured in order to complete the measurement of VOC and hydrocarbon mass emissions and for the purposes of the MIR calculation of emissions reactivity in the formation of ozone. In this study, the VOC speciation aimed to quantify the majority of the organic mass found in the exhaust and evaporative emissions, thereby capturing all compounds of photochemical significance, and the air toxics and alcohols. 3. Emissions Measurement Vehicle testing and the measurement of regulated components in exhaust and evaporative emissions were carried out at the Orbital vehicle test facility. The determination of organic components in the emissions was conducted at the analytical laboratories of the CSIRO Energy Technology Division in Sydney, with the exception of 1,3-butadiene which was analysed at Orbital due to the requirement for determination within a short period of its collection. 3.1 Exhaust Emissions Measurement Exhaust emissions were measured using a test fleet consisting of 21 vehicles representing a sampling of the most popular ethanol suitable makes and models from 1999 to the date of the evaluation; 27. Consideration of various factors effecting the penetration of light duty petrol vehicles in the Australian marketplace were made in the process of vehicle selection. The test fleet comprised seven ADR37/1 compliant models with two samples of each targeting ~1999 and ~23 model years (14 vehicles). In some cases the 23 group had an ADR79/ standing, a revision which holds similar emissions standards. Seven vehicles with current ADR79/1 compliance, were selected from the model years 26+. This category reflects the requirement for onboard diagnostics and more stringent emissions standards, and includes the 24-hour diurnal evaporative test, as discussed in the next section. Vehicle details can be found in Appendix A. The base fuel used in the test program was unleaded 91 octane summer grade petrol (ULP) with a dry vapour pressure of 62.5 kpa and complying with the current fuel standard. This fuel was used to prepare the E5 and E1 ethanol fuels which were splash blended at 5% and 1% by volume, respectively. Their consequent vapour pressures were 7 and 68 kpa, respectively. A transient drive cycle, the Petrol Composite Urban Emissions Drive Cycle (CUEDC), was used to characterise the emissions performance of the ULP, and E5 and E1 fuels. The Petrol CUEDC was recently developed as a drive cycle representative of real world Australian driving conditions for light duty petrol fuelled vehicles. It consists of four phases representing driving behaviour in residential (cold start), arterial, freeway and congested traffic conditions. Figure 1 details the speed/time profile for the four phases of the CUEDC cycle. Speed (km/h) Composite Urban Drive Cycle for light duty gasoline vehicles as developed for NISE2 Residential Arterial Freeway Congested Time (secs) Figure 1. Drive Trace for the Petrol CUEDC The exhaust emissions sampling system consisted of a full flow dilution tunnel meeting ADR (Australian Design Rule) and CFR (US Code of Federal Regulations) specifications. The constant

3 volume sampling system (CVS) incorporated a 254 mm primary dilution tunnel and critical flow venturi. Engine exhaust was directed from the tailpipe to the tunnel for dilution using HEPA filtered ambient air. Samples for regulated and unregulated emissions measurement were taken from ports located on the dilution tunnel under sampling parameters applicable to the target species. Separate sampling lines were used to draw exhaust from the dilution tunnel to systems for collection of hydrocarbon VOCs, carbonyl VOCs and ethanol. The emissions measurement system is illustrated in Figure 2. Figure 2. Schematic of Dilution Tunnel and Emissions Measurement System. Three replicate tests were performed for the measurement of organic components in exhaust emissions. This included one test where sampling of each individual phase of the drive cycle was undertaken to achieve a phase-by-phase result and calculated per cycle determination, and two tests where sampling over the total period of the four phases was undertaken to achieve an integrated per cycle result. For the purposes of the study it was assumed that each of the four phases has equal contribution to the total cycle. Iterations of the CUEDC can include estimations for average weightings to be applied to the various phases for particular driving patterns for modelling of emissions to urban centres and the national average. In this paper the emissions results based on the total cycle will be discussed. Vehicle exhaust emissions results are presented in units of mass/kilometre, after having been converted from the units of their sample collection of either concentration or mass. Exhaust samples for hydrocarbon VOCs were collected into passivated stainless steel canisters from test bags used for collection of total hydrocarbons, as per standard procedures for regulated testing. The canister sample was pressurised with dry nitrogen in preparation for transport and analysis. Background samples of the dilution air were also taken over the period corresponding to each test. Analysis of hydrocarbon VOCs was performed using instrumental techniques which incorporate systems for sample cold-trapping and pre-concentration, thermal desorption and gas chromatography. Analysis of hydrocarbon compounds was performed by both flame ionisation detection (GC-FID) and mass spectrometry (GC-MS). Standard methods for preparative and instrumental analysis were followed with primary reference to California Air Resources Board (CARB) methods for determination from vehicular sources (CARB, 22). These methods were optimised for full speciation and quantification of individual compounds and to the specific requirements engendered by the testing. A 57 component certified TO-15/PAMS hydrocarbon gas mixture (Scott Speciality Gases) was used for compound identification and instrument calibration. Additional compounds were identified by mass spectral elucidation and structural library search along with information obtained from hydrocarbon speciation of the gasoline fuel used in the study. This operation resulted in the speciation of 98 hydrocarbon components and quantification of at least 9% of the hydrocarbon mass present as C 2 to C 12 VOCs in the emissions samples. Samples for carbonyl VOCs were collected using DNPH (2,4-dinitrophenylhydrazine) cartridges designed specifically for collection of carbonyl species in the gas phase. The exhaust was drawn from the dilution tunnel through heated lines to the cartridge using a pump. Actual flow rates were proportioned from the CVS system flow and were optimised dependent on the concentrations expected from the test to 1-2 litres/minute. Ambient backgrounds were taken for each test and field cartridge blank samples were taken routinely. The cartridges were transported under ice. Determination of carbonyl compounds in the samples was performed using high performance liquid chromatography (HPLC), with diode array detection operated in the UV range, using standard methodology (CARB, 22). The system was optimised for separation and quantification of C 1 to C 8 carbonyl compounds based on a certified 13-component carbonyl-dnph standard mixture (Supelco CARB 14). The analysis targeted the compounds contributing significantly in the carbonyl emissions profile and as such eight carbonyl compounds were reported. The profile was dominated by formaldehyde and acetaldehyde and these will be the subject of discussion in this paper. Ethanol samples were collected using a wet-scrubbing technique into impingers, at flow rates of.25 1 litre/minute, under a sampling system of controlled flow similar to the carbonyls procedure. A second impinger was positioned in series to provide a backup for analyte carryover and the two impingers were placed in an ice-bath

4 during sampling. Test background and impinger blank samples were also taken. The impinger solutions were transferred to vials prior to transport. The determination of alcohols was performed using gas chromatography with flame ionisation detection (GC/FID) based on standard methodology (CARB, 22). The chromatographic sequence was optimised for separation of ethanol from minor alcohols and other interfering compounds. Alcohols were quantified against pure liquid standards. Methanol was the only other compound observed in the chromatographic profile and was found as a minor component in certain vehicle emissions only. This discussion will address ethanol emissions. 3.2 Evaporative Emissions Measurement Evaporative emissions were measured on a subset of 6 vehicles from the 21 vehicle test fleet. Three vehicles were selected from each of the ~1999 (ADR37/1) and 26+ (ADR79/1) categories. The ADR79/1 evaporative emissions testing protocol was used, which is the current Australian Design Rule emissions certification standard for light duty petrol vehicles. The evaporative test incorporates a 24-hour diurnal breathing loss test and a 1-hour hot soak test. The diurnal test represents the effect of daily ambient temperature change and the potential for vapour to form and escape from the vehicle s fuel system when the vehicle is parked. The 24-hour diurnal was considered to be more representative of real world behaviour than the previous 1-hour test used in earlier ADR revisions. The temperature is profiled from 2 C to 35 C and back to 2 C over a 24-hour period. The hot soak test represents the condition immediately after vehicle operation as a result of heating the fuel tank vapour space. The measurement of organic compounds was undertaken on three replicates of both the diurnal and hot soak test. For both tests, a sample is taken at the start and finish of the test period in order to generate the evaporative emission result. In this paper the results of the total test will be discussed. This is the sum of the diurnal and hot soak tests and, under ADR protocol, assumes that each has equal contribution. The vehicle evaporative emission results are presented in units of mass/test after being converted from their original sample collection units of either concentration or mass. Evaporatives emissions sampling is undertaken using probes fitted through the wall of the evaporative test sheds. The overall system of transfer to the collection vessels for determination of organics is similar to that used for exhaust. The measurement of 1,3-butadiene and carbonyl compounds is not required as these are combustion generated and hence not a component of fuel evaporative emissions. Analytical determinations were performed in a similar manner as the exhaust emissions, with optimisation of the methodologies in response to differences in composition and concentration for evaporative emissions. 4. Results and Discussion 4.1 Exhaust Emissions The results for emissions of priority air toxics (as benzene, toluene, ethylbenzene, total xylenes and 1,3-butadiene) in exhaust generated from ULP and E1 fuels, over the CUEDC drive cycle, for each of the 21 test vehicles are presented in Figure 3. The E5 result has been excluded from this chart for reasons of simplicity, but will be discussed briefly. The marked improvement in ULP emissions is clearly seen for the 26+ vehicles with the change in regulatory limits to ADR79/1. This observation follows for the reduction in total nonmethane volatile organic compounds (TNMVOC) obtained from the VOC speciation, and the total hydrocarbon (THC) from regulated testing. Based on the average for the 26+ model year of 11 mg/km, around 8% lower air toxics emissions are found compared with the ~1999 grouping (51 mg/km), i.e. 4-5 times higher emissions for the older vehicles. This follows the THC emissions of 48 mg/km and 227 mg/km (average) for the two age groups, respectively. The total of BTEX + 1,3-butadiene air toxics contribute around 2% to the THC emissions for all vehicle age groups. Emissions Rate, mg/km Commodore 2 Corolla 3 Falcon 4 Mazda 3 5 Yaris Exhaust Toxics Emissions: ULP vs E1 Benzene Toluene Ethylbenzene Total Xylenes 1,3-Butadiene 1st Bar: ULP 2nd Bar: E1 6 Hilux 7 Camry 8 Commodore 9 Falcon 1 Camry 11 Corolla 12 Lancer 13 Astra 14 Magna 15 Commodore 16 Falcon 17 Camry 18 Corolla 19 Lancer 2 Astra 21 Magna 26+ ADR 79/1 ~ 23 ADR 37/1, 79/ ~ 1999 ADR 37/1 Figure 3. Priority air toxics in CUEDC exhaust emissions for the vehicle test fleet. Across all age groups and for most vehicles, the E1 fuel had a positive impact in reduction of individual air toxics, although the extent of change varied. The older ADR37/1 certified vehicles (~1999 and ~23) showed greatest variation within each grouping, an effect compounded by the age of their catalysts systems. Overall, E1 fuel produced a fleet average reduction in exhaust emissions of 11% for air toxics, a result also obtained for THC. No real trend was seen in the extent of reduction between the three model years, as differences in individual vehicles blurred any possible correlation. The E5 fuel result generally

5 exhibited changes consistent with the trend E1>E5>ULP. The exhaust emissions of speciated VOCs found similar trends, as seen in Figure 4 using Vehicle 18 as an example (Toyota Corolla ~1999, ADR37/1 compliance). Here the VOC speciation is ranked in order of concentration of the ULP emissions. The emissions profile for all fuels is dominated by the aromatics, which are primarily fuel derived, and the combustion derived components; ethene, ethane, propene and acetylene, along with other light hydrocarbons from fuel and combustion origins. Emissions Rate, mg/km Exhaust VOC Emissions: ULP, E5, E1 Toyota Corolla (~1999, ADR37/1) ULP E5 E1 Toluene Ethene m- + p-xylene Isopentane Benzene Propene Ethane o-xylene 2-Methylpentane Acetylene 1,2,4-Trimethylbenzene Ethylbenzene 1-Butene + Isobutene 3-Methylpentane 2,2-Dimethylbutane 3-Methylhexane Methylcyclopentane 3-Ethyltoluene n-pentane 2-Methylhexane Cyclohexane n-hexane n-heptane n-butane 2-Methyl- + 4-Methylheptane 3-Methylheptane 2,3-Dimethylbutane 2-Ethyltoluene 1,3,5-Trimethylbenzene 1,2,3-Trimethylbenzene 4-Ethyltoluene Methylcyclohexane 2,3-Dimethylpentane n-octane C7 alkanes 1,4-Diethylbenzene C1 aromatics C9 alkane trans-2-butene n-propylbenzene Propyne trans-2-pentene 2-Methyl-2-Butene C8 alkane Isooctane C8 aliphatics 2,4-Dimethylpentane C1 aromatics cis-2-butene C7/C8 alkane C9 alkane Cyclopentane 1-Pentene 2,3-Dihydroindene 1-Buten-3-yne n-nonane C7 alkane 2,3-Dimethylhexane cis-2-pentene Isobutane 2-Methyl-1-Butene 2-Methyl-1-Pentene + 1- Propane n-decane trans-2-hexene 3,3-Dimethylpentane Isopropylbenzene 2-Methyl-2-Pentene 1,3-Diethylbenzene Cyclopentene 3-Methyl-1-Butene 2,3,4-Trimethylpentane trans/cis-3-hexene 3,3-Dimethyl-1-Pentene n-undecane cis-3-methyl-2-pentene Styrene cis-2-hexene C7 alkene 4-Methyl- + 3-Methyl-1- n-dodecane n-tridecane Isoprene C8 alkene Figure 4. Speciated VOC profile for ULP, E5 and E1 exhaust from a ~1999 Toyota Corolla The emissions rate of the carbonyl VOCs, as formaldehyde and acetaldehyde, are presented in Figure 5. These compounds dominate the carbonyl emissions profile, contributing around 7% to the total of major C 1 C 8 components measured. Once again a significant decrease in carbonyl emissions is seen for the later model vehicles with around 85% lower emissions for 26+ (average 1.3 mg/km for ULP) compared to the ~1999 grouping (7.6 mg/km). The carbonyls make up around 1-3% of the THC. Emissions Rate, mg/km Commodore 2 Corolla 3 Falcon 4 Mazda 3 Exhaust Carbonyls Emissions: ULP vs E1 Formaldehyde 1st Bar: ULP Acetaldehyde 2nd Bar: E1 5 Yaris 6 Hilux 7 Camry 8 Commodore 9 Falcon 1 Camry 11 Corolla 12 Lancer 13 Astra 14 Magna 15 Commodore 16 Falcon 17 Camry 18 Corolla 19 Lancer 2 Astra 21 Magna 26+ ADR 79/1 ~ 23 ADR 37/1, 79/ ~ 1999 ADR 37/1 Figure 5. Formaldehyde and acetaldehyde in CUEDC exhaust emissions for the test fleet. Unlike the hydrocarbon VOCs, the E5 and E1 fuels generally had a negative impact on carbonyl exhaust emission rates. The major contributor to this effect was the acetaldehyde component of the carbonyl emissions, where substantial increases were seen for all vehicles. The presence of ethanol promotes the production of acetaldehyde, and to a lesser extent formaldehyde, via direct oxidation and via oxidative processes in the conversion of fuel hydrocarbons. Formaldehyde emissions tended to be variable between vehicles, a result not unexpected due to its generation dependence largely on combustion characteristics of the engine. In the case of E1, most vehicles showed similar emissions or modest increases (up to around 35% on average). Exceptions were major increases for two vehicles; vehicle 5 (26+) showed a 126% increase (2.3 times higher emissions but from a low absolute emissions base of.15 mg/km) and vehicle 15 (~1999) showed an 88% increase (1.9-times higher and from a high emissions base of 4.1 mg/km). Decreases of 22% and 3% for vehicles 6 and 7 (26+) were also seen. Acetaldehyde emissions for the E1 fuel were up to 5 times higher, and on average 3-times higher than those for the ULP (increases ranged from 9% to 41%). In general, the E5 emissions sat between the ULP and E1. Ethanol was found in the E5 and E1 exhaust at relatively low emission rates, with a general trend towards higher emissions with increasing fuel content, particularly for the older vehicles. For E1, the 26+ age group averaged an emission rate of.8 mg/km with nil detected for a number of vehicles in this, and other, groups. On average, the 26+ vehicle group produced around 85% lower emissions for the E1 fuel compared to the older ~1999 vehicles. 4.2 Evaporative Emissions The results for emissions of priority air toxics (as benzene, toluene, ethylbenzene and total xylenes) present in evaporative emissions generated from ULP, E5 and E1 fuels, over the ADR79/1 test, for each of the six test vehicles are presented in Figure 6. 1,3-butadiene and carbonyl compounds are not included in the air toxics suite for evaporative emissions as these are combustion derived component, present in exhaust only. Emissions Mass, mg/test Evaporative Toxics Emissions: ULP, E5, E1 Benzene Toluene Ethylbenzene Total Xylenes ULP E5 E1 ULP E5 E1 ULP E5 E1 ULP E5 E1 ULP E5 E1 ULP E5 E1 Vehicle 1 Vehicle 2 Vehicle 5 Vehicle 15 Vehicle 17 Vehicle 18 Commodore Corolla Yaris Commodore Camry Corolla 26+ ~ 1999 Figure 6. Priority air toxics in evaporative emissions for the vehicle test fleet.

6 Immediately noticeable in the results from evaporative testing is the significant increase in the emissions for the E5 and E1 fuels compared to the ULP, and the fact that E5 has the highest emission rate for all vehicles. This result is aligned with that for the THC emissions. The evaporative emissions follow the vapour properties of the fuel, and fuels with small quantities of ethanol, such as E5 and E1, have a higher vapour pressure than the base fuel to which the ethanol is added. In this study the E5 blend had a slightly higher vapour pressure than the E1. As such the E5 tests produced higher evaporative emissions than the E1, with ULP giving the lowest emissions. In relative terms, E5 generally lead to at least a doubling of the evaporative emissions of air toxics, and of THC, whilst E1 emissions were 5%, or more, higher than the ULP. The diurnal phase of the test dominated the total evaporative emissions, typically making up 8-9% of the total. The evaporative emissions trend between the fuels is also seen for speciated VOCs, as shown in Figure 7 using Vehicle 1, the Holden Commodore, as an example (26+, ADR79/1). Here the VOC speciation is ranked in order of concentration of the ULP. The emissions profile for all fuels is dominated by compounds of higher volatility and those presenting at higher concentration in the fuel itself. For most vehicles, toluene, m- + p-xylene, 2-methylpentane and the lighter C 4 and C 5 hydrocarbons contribute the majority of the emissions mass. The emissions profile can be affected by the performance of the vehicles fuel vapour recovery canister used for emissions mitigation, and the effect of ethanol on canister performance in its adsorption of hydrocarbons. Emission Mass, mg/test Evaporative VOC Emissions: ULP, E5, E1 Holden Commodore (26+, ADR79/1) ULP E5 E1 Isopentane n-butane Toluene n-pentane 2-Methylpentane Isobutane trans-2-butene 2,2-Dimethylbutane 3-Methylpentane 2-Methyl-2-Butene 1-Butene + Isobutene m- + p-xylene cis-2-butene n-hexane trans-2-pentene Methylcyclopentane 2-Methyl-1-Butene Benzene Cyclohexane 3-Methylhexane 2,3-Dimethylbutane n-heptane cis-2-pentene 2-Methylhexane o-xylene Ethylbenzene 1-Pentene Cyclopentane 2-Methyl-2-Pentene 3,3-Dimethyl-1-Pentene 2-Methyl-1-Pentene + 1-Hexene Methylcyclohexane 1,2,4-Trimethylbenzene trans-2-hexene 2-Methyl- + 4-Methylheptane Cyclopentene cis-3-methyl-2-pentene 2,3-Dimethylpentane 2,4-Dimethylpentane 3-Methylheptane 3-Methyl-1-Butene C7 alkanes 3-Ethyltoluene cis-2-hexene n-octane Isooctane C7/C8 alkane C8 aliphatics trans/cis-3-hexene 2,3,4-Trimethylpentane C8 alkane 2-Ethyltoluene C7 alkane 4-Ethyltoluene 1,3,5-Trimethylbenzene C9 alkane 1,2,3-Trimethylbenzene 2,3-Dihydroindene 2,3-Dimethylhexane n-propylbenzene 3,3-Dimethylpentane Isoprene Methyl-1-Pentene C9 alkane n-nonane Isopropylbenzene 1,4-Diethylbenzene Styrene 1,3-Diethylbenzene n-decane C1 aromatics n-undecane Figure 7. Speciated VOC profile for ULP, E5 and E1 evaporative emissions from a 26+ Holden Commodore. Ethanol is a major component of the vapour generated by ethanol blended fuels, at E5 and E1 levels, and hence also contributes significantly to evaporative emissions. This effect is due to the decrease in hydrogen bonding between ethanol molecules and hence the fuel mixture exhibits non-ideal behaviour and ethanol contributes to the vapour composition to a greater extent than its vapour pressure would theoretically dictate (Harley and Coulter-Burke, 2). Certainly ethanol contributed significantly to evaporative emissions in this study, however the results were quite variable between vehicles and a significant trend was difficult to deduce. The emission rate is also complicated by ethanol s effect on the efficiency of the vehicle s fuel vapour recovery canister. 5. Conclusion Organic compounds, as speciated hydrocarbon, carbonyl and ethanol VOCs, were evaluated in exhaust and evaporative emissions from vehicles fuelled with regular gasoline and E5 and E1 ethanol blended gasoline fuels. Emission rates for vehicles holding current ADR79/1 certification, had significantly improved exhaust emission rates and showed some improvement in evaporatives emissions. The impact of ethanol blended fuel was generally favourable for exhaust emissions of hydrocarbon VOCs and this effect aligned with that for total hydrocarbon. Ethanol had a negative impact on exhaust acetaldehyde emissions where major increases were seen, and on hydrocarbon components in evaporative emissions, an effect which was more pronounced for the higher volatility E5 fuel. Ethanol emissions were minor in exhaust but significant in evaporative emissions. Speciation of organic compounds has provided insight into the emissions impact of fuels and vehicle technology necessary for air quality and health evaluations. Acknowledgments The authors wish to acknowledge the enormous effort made by our colleagues in achieving the analytical outcomes of this project; Ian Campbell, Rosemary Wood, Owen Farrell, Steve Lavrencic, Katelyn Edge and David Jacyna, and advice on aspects of sampling from Brendan Halliburton. Also our collaboration in many aspects of the project with David Worth from Orbital, and his staff. References Biofuels Taskforce Report to the Prime Minister 25, Brown, S., Pengilley, M. and Patterson, D. 1998, Petrohol in-service vehicle emissions study. Report No. MV-A-35 to the NSW EPA. Carter, W and Malkina, I. 22, Development and application of improved methods for measurement of ozone formation potentials of volatile organic compounds. Final report to the California Air Resources Board. Contract CARB 22, Test methods for the determination of hydrocarbons from vehicular sources, California Environmental Protection Agency Air Resources Board, Sacramento, USA.

7 CSIRO and Orbital Australia Corporation 28, Evaluating the health impacts of ethanol blend petrol, Report to the Department of Environment, Water, Heritage and the Arts (DEWHA). uality/publications/ethanol-health-impacts.html Harley R.A. and Coulter-Burke, S.C. 2, Relating liquid fuel and headspace composition for California reformulated gasoline fuel samples containing ethanol, Environmental Science and Technology 34, Orbital Australia Corporation 23, A testing based assessment to determine the impacts of a 2% ethanol gasoline fuel blend on the Australian passenger vehicle fleet - 2 hour material compatibility testing.

8 Appendix A: Test Fleet Vehicle Details ID Group Make Model Model Variant Body Compliance Date ADR Certification Odometer km Engine Displ., L Cylinders Holden Commodore VE Omega Sedan Aug, 26 ADR79/1 9, Auto Toyota Corolla ZZE122R 5Y Ascent Hatch Jan, 27 ADR79/1 9, Auto Ford Falcon Fairmont Sedan Sept, 27 ADR79/1 3, Auto Mazda Mazda3 BK1F1 Neo Sedan July, 26 ADR79/1 8, Manual Toyota Yaris NCP9R YR Hatch Oct, 26 ADR79/1 7, Manual Toyota Hilux 15 SER Cabin July, 26 ADR79/1 27, Manual Toyota Camry Altise Sedan April, 27 ADR79/1 7, Auto 8 ~23 Holden Commodore VY Executive Sedan Nov, 23 ADR37/1 68, Auto 9 ~23 Ford Falcon XT Sedan Sept, 23 ADR79/ 78, Auto 1 ~23 Toyota Camry Aveta Sedan Feb, 23 ADR79/ 15, Auto 11 ~23 Toyota Corolla Ascent Sedan Nov, 23 ADR79/ 57, Manual 12 ~23 Mitsubishi Lancer Gli Coupe Oct, 23 ADR37/1 99, Manual 13 ~23 Holden Astra Sxi Hatch July, 23 ADR37/1 69, Auto 14 ~23 Mitsubishi Magna TJ Series 2 Sedan Mar, 23 ADR37/1 11, Auto 15 ~1999 Holden Commodore VT II Acclaim Sedan June, 1999 ADR37/1 17, Auto 16 ~1999 Ford Falcon AU Model Sedan Oct, 1999 ADR37/1 166, Auto 17 ~1999 Toyota Camry SXV2R CSX Sedan Aug, 1998 ADR37/1 163, Auto 18 ~1999 Toyota Corolla AE112R Conquest Hatch Sept, 1998 ADR37/1 159, Auto 19 ~1999 Mitsubishi Lancer MR Coupe Jan, 1999 ADR37/1 145, Auto 2 ~1999 Holden Astra CD Hatch Mar, 1999 ADR37/1 144, Manual 21 ~1999 Mitsubishi Magna TH Series Sedan July, 2 ADR37/1 152, Auto Vehicles 1, 2, 5, 15, 17, 18 subset for evaporatives testing Trans.