Profiling the Organic Emissions from a Light-Duty Direct Injection Diesel Engine over a range of Speeds and Loads

Abstract

Diesel engines account for a large percentage of the particulates in urban city environments. Polycyclic aromatic compounds (PAC), some proven carcinogens, have been found on diesel particulates. The trace level nitro-PAC emissions, such as 1-nitropyrene and dinitropyrenes, contribute a large proportion of the mutagenicity in the particulates; in the case of 1-nitropyrene between 10 to 40% of the total mutagenicity of the particulate has been claimed. The potential health hazards of PAC require the levels and sources of such emissions to be evaluated over a range of speeds and loads. PAC emissions are dependant on the engine specification, such as normally aspirated compared with turbocharged, and the operating conditions (speed and load). The effect of such variables can be determined using emission profiling, in which profiles of the exhaust are compared at varying engine powers. In this way the effect of speed and load on the combustion efficiency can be established. Identification of PAC sources may be further complicated when engine sampling systems, such as the conventional dilution tunnel/filter system, are prone to artefact formation. This is especially relevant to secondary nitro-PAC emissions, which are prone to forming as artefacts of the filter installed in the dilution tunnel. In this study, organic emission maps were constructed using the unique total exhaust solvent-stripping apparatus (TESSA) developed at the University of Plymouth. TESSA allowed rapid sampling with a minimum potential for artefact formation. The close proximity of TESSA to the engine allowed the role of the combustion chamber in the formation of emissions to be evaluated. Primary organic emissions, such as pyrene, are derived from survival of compounds in the fuel/oil and by combustion generation. Establishment of emission maps for the primary emissions are vital to resolving the formation of secondary emissions, such as 1-nitropyrene. Profiling of primary emissions sampled 26 different speeds and loads using TESSA (sampling times as low as IS seconds). Following simple work-up, quantification of the o-alkanes was by gas chromatography with flame ionisation detection and PAH by gas chromatography/mass spectrometry operated in electron impact mode. Emissions were expressed as a recovery of the compound emitted as a percentage of the same compound entering the chamber in the fuel. The o-alkane and PAH emission maps correlated with the gaseous unburnt hydrocarbon emissions, indicating that fuel survival was an important source of emissions, whereas lubricating oil contributions were minimal. Fuel survival contributions decreased with load; at 1000 rpm the average PAH survival of 0.95% at idling decreased to 0.2% at full load. High survivals under idling was a consequence of the low chamber temperatures and air:fuel ratios mixed beyond the lean flammability limits, whereas at full load, the high temperatures resulted in the greatest combustion efficiency. The o-alkane emission trends replicated those of PAH; at 1000 rpm the average o-alkane survival was 0.48% at idle compared with 0.084% at full load. Correlations between the distribution of the emissions and fuel at high load, suggested fuel survival unchanged was responsible. At low loads the exhaust/fuel PAH ratios were more varied, with the range of percentage recoveries at low loads increasing with speed (difference between percentage recovery of fluorene and phenanthrene at idling for 1000 rpm and 3000 rpm was 0.05 and 0.18 respectively). At high load, the combustion environment can be envisaged as producing areas under which complete combustion and survival unchanged occur. In between the complete combustion and unburnt fuel zones, a narrow range of temperatures and time for combustion reduce the opportunity for combustion generation reactions and/or preferential survival. Low load at 3000 rpm may increase the intermediate zone, allowing preferential survival and/or combustion generation reactions to evolve. Possible pyrolytic cracking of o-alkanes and demethylation of PAH at low loads and 3000 rpm was evident. The optimum time, swirl, and temperature for efficient combustion at low loads was generated at 2000 rpm to 2500 rpm, whereas at the higher temperatures corresponding to high loads, the effect of speed was much smaller. The primary emissions map show engineering improvements, particularly at low loads, could be implemented to lower the PAH emissions. The correlation between the emissions and fuel input suggest modifications to the PAH content of fuels may lower emissions. The formation of secondary nitro- and oxy-PAC emissions is by transformation of primary emissions. In the case of nitro-PAC, nitration has been proposed to occur by free radical processes between PAH and oxides of nitrogen, NO,, within the chamber and by electrophilic substitution of PAH surviving combustion by nitrogen dioxide and nitric acid, via the nitronium ion. The combustion contribution to nitro-PAC emissions was investigated using an upgraded TESSA system, and 3 speeds at low, mid, and high NO, for each speed were sampled. Following initial extraction, concentration and clean-up of the samples, the nitro-PAC fractions of interest were isolated by normal phase high performance liquid chromatography. The nitro-PAC were identified and quantified by gas chromatography with electron capture detection, and gas chromatography/mass spectrometry operated in the negative ion chemical ionisation mode (the detection limits for both analytical systems was of the order of 40-50 pg of nitro-PAC standards injected). The profiling indicated that a proportion of the fuel underwent nitration within the combustion chamber across a range of speeds and loads. The extract concentrations (average of 5.3 ppm) found in this study were much lower than those previous found (ranging from 55 to 2280 ppm). The majority of the previous studies relied on sampling using dilution tunnel/filter systems, for which post combustion contributions are simulated; suggesting that a major source of nitro-PAC is derived from post-combustion nitration of PAH surviving combustion, some of which may be artefacts of the filter. Different speeds produced different trends for nitro-PAC emissions with respect to engine load. It was not until the high temperature speed of 3500 rpm was reached that both NO, and nitro-PAC increased with load (R-sq= 0.989 & p=0.067 for 1-nitronaphthalene & NO,). Nitro-PAC emissions at 3500 rpm were primarily the result of combustion chamber nitration of PAH at high NO,. In the case of 1-nitropyrene, there was strong evidence to support pyrosynthetic contributions to the pyrene mass, which in turn became nitrated. The nitro-PAC emissions at low loads were the result of post-combustion nitration of PAH surviving combustion with the nitronium ion. The correlation of the PAH precursors to nitro-PAC in the fuel and nitro-PAC emissions suggest fuel modifications may to some extent lower the nitro-PAC emissions. The combustion generation of nitro-PAC at high engine powers may require post-combustion after treatment.