Exhaust emissions from light- and heavy-duty vehicles: chemical composition, impact of exhaust after treatment, and fuel parameters.

This paper presents results from the characterization of vehicle exhaust that were obtained primarily within the Swedish Urban Air Project, "Tätortsprojektet." Exhaust emissions from both gasoline- and diesel-fueled vehicles have been investigated with respect to regulated pollutants (carbon monoxide [CO], hydrocarbon [HC], nitrogen oxides [NOx], and particulate), unregulated pollutants, and in bioassay tests (Ames test, TCDD receptor affinity tests). Unregulated pollutants present in both the particle- and the semi-volatile phases were characterized. Special interest was focused on the impact of fuel composition on heavy-duty diesel vehicle emissions. It was confirmed that there exists a quantifiable relationship between diesel-fuel variables of the fuel blends, the chemical composition of the emissions, and their biological effects. According to the results from the multivariate analysis, the most important fuel parameters are: polycyclic aromatic hydrocarbons (PAH) content, 90% distillation point, final boiling point, specific heat, aromatic content, density, and sulfur content.


Introduction
The impact of internal combustion engines, Otto, and diesel engines on the environment and our lifestyles has been considerable. In Europe, research and development work on engines within the last two to three decades has been strictly focused on engine performance in terms of power output, fuel economy, reliability, etc., but not on engine emissions. During recent decades, pronounced interest has focused on exhaust emissions and their impact on health and the environment (1)(2)(3)(4). Due to the rapid increase in the number of vehicles in use, especially in urban areas, engine emissions have become suspected culprits for some of the health effects observed in urban populations (5).
Most of the interest in emissions has been focused on passenger cars and other light-duty vehicles, because these categories of vehicles exist in much greater numbers than the heavyduty vehicles. Vehicle emissions are usually divided into categories of regulated and unregulated pollutants. Regulated pollutants consist of carbon monoxide (CO), nitrogen oxides (NO , mainly nitrogen monoxide and nitrogen dioxide), unburned fuel, or partly oxidized hydrocarbons (HC), and particulates. These pollutants are specified by law in most of the industrially advanced countries. Unregulated pollutants are defined as compounds that are not specified by law. However, these unregulated pollutants may well belong to the group of unburned hydrocarbons, but not as individual compounds.
Several of the compounds present in diesel and gasoline engine exhausts are known to be carcinogenic and/or mutagenic (5). A group of compounds most often associated with this carcinogenic/ mutagenic property are the polycyclic aromatic compounds (PAC) (6). This is of interest because unregulated pollutants are generally measured under the same driving conditions as those developed for regulated pollutant evaluations. Consequently, the unregulated pollutant measurements are made under the same engine operation conditions as for regulated pollutants, although the exhaust emission measurements obtained have no legal bearing. This publication focuses primarily on vehicular pollutants that have been investigated and characterized within the Swedish Urban Air Project (SUAP), Tatortsprojektet (1979Tatortsprojektet ( to 1991, founded by the Swedish Environmental Protection Agency (SwEPA); these data were obtained largely from the time period 1985 to 1991. The effects of pollutants emitted from mobile sources are dependent on several factors that will be discussed herein.

Results and Discussion
Gasoline-fueled Vehides Organic Halides. Particle-associated organic halides have been identified in exhausts from gasoline and diesel-fueled vehicles (7). The most abundant bromofluorenone isomer determined as present in gasoline exhaust corresponds to an emission of approximately 2 pg/ km. Haglund et al. (8) have investigated the presence of halogenated polycyclic aromatic hydrocarbons (PAH) in urban air, snow, and automobile exhaust, determined qualitatively in all three samples. A possible source of bromine and chlorine in leaded gasoline is the addition of scavengers, such as ethylene dichloride (EDC) and ethylene dibromide (EDB) (7,8). However, the presence of bromine in diesel fuel has also been determined (7), indicating bromine to be a natural constituent in crude oil. Another explanation for the presence of bromine in diesel fuel may be the storage conditions of the crude oil. Since storage is often in underground cavities on a bed of sea water, bromine present in the sea water may leach into the oil store. Further, polyhalogenated dioxins and furans have been detected in the exhaust emissions of leaded gasoline-fueled Table 1. Mean emission factors (N= 2 to 4) from constant cruising speeds, 70 and 90 km/hr. gated: carburetor without exhaust after treatment (V1); three-way catalyst system, X-sond, turb (V2); three-way catalyst system and X-sond (V3) (14). CO cars (9,10). Haglund and co-workers (10) determined bromfurans to be more prominent than bromdioxins in leaded gasoline exhausts (10). Emissions of polychlorinated dibenzodioxins (PCDD) and polychlorinated dibenzofurans (PCDF) from gasoline-fueled vehicles are not considered to be a major environmental pollutant, as group profiles differ significantly from those obtained in ambient air samples and in the biota (11). This indicates that when emitted from mobile sources, organic bromine derivatives might have a greater impact on the environment than organic chlorine derivatives. Engine Operating Conditions. The amount of exhaust emitted from light-duty vehicles depends on driving conditions such as load and speed. This correspondence has been investigated with respect to regulated exhaust emissions (12). It was demonstrated that a correlation between increased speed of the vehicles and CO and NOX emissions was evident for speeds over 50 km/ hr for vehicles without a catalyst. However, studies of exhaust emissions from gasoline-fueled passenger vehicles, with and without a three-way catalyst, have been carried out in combination with bioassay tests to obtain emission factors for both regulated and unregulated pollutants (13,14). Both Toxic  27 tions (FTP-73). Both particleand semi-Toxic 7.9 volatile phase-associated PAH were sampled and quantified. A decrease in fuel-PAH content resulted in a decrease of emitted PAH. Using a simple mathematical method to distinguish PAH present in the fuel and given in Table PAH formed in the exhaust emission, the from catalyst-intercept value can be interpreted as PAH d a dramatic formed in the combustion process (15,17). from the vehi-Using this method, we found that a major ugh the emis-portion (>50%) of emitted PAH is formed Table 2. Emission of PAH (pg/ km), particulate phase associated, N= 4, mean value +/-standard deviation (SD) (%), semi-volatile associated, N= 2, mean value +/-SD (%). Lead-free gasoline, US FTP-73 test, k-sond without catalyst ( 15 Table 3. Emission of PAC (pg/ km), particulate phase-associated, N= 6, mean value +/standard deviation (SD) 1%), semi-volatile-associated, N= 2, mean value +/-SD 1%) (27 in the combustion process. It was found that when a full range of aromatics (mono, di, and larger aromatics) was present in the fuel, more mutagens were produced in the combustion process. Comparison of PAH distribution between the particleand the semi-volatile phases, where samples were collected using the dilution tunnel technique or from ambient air revealed that PAH profiles were similar in both of the collection methods: sampling of vehicle gasoline exhaust using the dilution technique generated particleand semi-volatile phaseassociated PAH profiles similar to those measured in urban environments (18).
The evaluation of biological activity of the semi-volatile phase leads initially to controversial results. Stump et al. showed that an adsorbent trap enriched very low amounts of mutagenic material compared to the level in the filter extracts (19). Schuetzle has also reported similar results (20). However, using the cryogenic condenser technique (21) the mutagenic activity of the semi-volatile phase was determined to be approximately 30 to 50% of the total mutagenicity (15,22,23). These results demonstrate the need to investigate both 48  particle-and semi-volatile phases when characterizing exhaust emissions from gasoline-fueled vehicles. This statement is also valid for characterization of diesel exhaust, further discussed in the section "Dieselfueled Vehicles." In Table 2, particulateassociated and semi-volatile-phase-associated PAC originating from the dilution tunnel technique are presented (16). Diesel-fueled Vehides Characterization. A literature survey of identified and tentatively identified constituents in diluted diesel exhaust emissions was performed within the SUAP (16), which indicated that approximately 450 individual compounds have been detected in diesel exhaust. Quite a few new compounds have been identified or tentatively identified in diesel exhaust samples since 1987 (24). However, several compounds present in diesel exhaust have not been identified and investigated with respect to possible environmental and health effects. Our evaluation of diesel exhaust emissions under the SUAP has employed a methodology designed for heavy-duty diesel exhaust emission characterizations, which was developed in close cooperation with the former SwEPA Emission Research Section in Studsvik (25). As for gasoline-fueled vehicles earlier presented, the evaluation, in conjunction with bioassay tests, characterizes both regulated and unregulated pollutants. Samples obtained consist of particulate material, semivolatile material, and gaseous components. Vehicles are operated during transient driving conditions on a chassis dynamometer, and prior to sampling, the exhaust is diluted in accordance with the Federal Register (26). The transient driving schedule used is the "bus cycle," which simulates public transportation conditions within a city. The bus cycle was originally developed at the Technical University of Braunschweig (Germany). The development of sampling methodology, chemical analysis of the particulate and semivolatile associated unregulated pollutants and bioassay tests are described in more detail elsewhere (25,27). A particulate filter was used downstream from the diluted diesel exhaust for collection of the particulate phase, and polyurethane foam plugs were used for sampling of the semi-volatile phase.
Sampling results indicated that the semivolatile phase PAH content was approximately 280% higher than the particle-phase PAH. Most of the gas phase PAH mass consisted of three-ring compounds (27). The contribution of semi-volatile-phaseassociated mutagenicity to the total mutagenicity was approximately 20% in strain TAIOO +/-S9; in strain TA98-S9, the contribution was 10%; and in strain TA98+S9, 37%. Both particle-and semi-volatile phase crude extracts were fractioned into five fractions by a method originally developed for crude gasoline extracts (28,29). The greatest mutagenic activity was found in both phases in the fractions containing monoand dinitro-PAH (27). Chemical analyses of PAC measured in particulate and semivolatile phases as determined from diluted diesel exhaust that was sampled from the bus cycle are displayed in Table 3.
Fuel Effects. Eight diesel fuels and two heavy-duty vehicles (bus denoted vehicle 1, truck denoted vehicle 2, respectively) were selected for investigation of fuel composition for regulated pollutants, unregulated pollutants, and biological effects (via mutagenicity and TCDD-receptor affinity tests) in the exhaust emissions. Exhaust emission tests were repeated in triplicate for each fuel blend and vehicle. The investigation also focused on specification of fuel parameters in order to obtain less polluting diesel fuels for vehicles operating in cities. One common, commercial, standard diesel fuel was used as a reference for commercial Volume 102, Supplement 4, October 1994 The PLS regression coefficient K= 0.33 (correlation coefficient R = 0.79) was obtained. The PLS regression is displayed in Figure 2, in which the x-axis is the PLS component X1 representing fuel parameters, and the y-axis is the PLS component Y1 representing unregulated pollutants emitted (aldehydes, olefins, light aromatics, PAH, 1-nitropyrene). The PLS regression coefficient K= 0.64 (correlation coefficient R = 0.88) was obtained. In a similar man-4 5 6 ner, the PLS regression model for biological effects is displayed in Figure 3, where the y-D9) for vehicle 2 We conduded from the study that there exists a quantifiable relationship between the variables of the diesel-fuel blends and the variables of the chemical emission and their biological effects. Figure 4 illustrates the z-axis, the PLS-component X1 that represents the emission of both regulated and unregulated exhaust components, and the y-axis, the PLScomponent Y1 that represents measured biological effects in the exhaust emissions. From the figure, we observe how the PLS regression coefficient K= 0.62 (correlation coefficient R= 0.85) is obtained. An important emission parameter with respect to biological activity is the sum of PAH in both the particle and semi-volatile phases. Diesel fuel-related emission factors emanating from a standard diesel fuel (D6, Swedish summer diesel-fuel quality), a diesel fuel for city buses (D8), and a test fuel (Dl) are presented in Table 4. The PLS regression coefficients obtained in fuels available on the Swedish market; three are commercially available and commonly used for city buses, and four are test fuels.
The entire investigation is described in detail in an earlier publication (25).
This presentation is a very condensed version of the results and findings from our investigation of the bus cycle, using a chassis dynamometer and the dilution tunnel technique. Chemometrics or multivariate analysis methodologies were used to help analyze the large quantities of data produced by this study (25). In order to build quantitative relations between different blocks of data, such as fuel parameters, regulated pollutants, unregulated pollutants, and bioassay emission data, partial least squares regression (PLS) was used (30). The relationships demonstrated were relatively vehicle-independent.
Environmental Health Perspectives ponents that have an adverse effect on urban air quality. A possible technique to reduce particulate emissions is the use of an exhaust after treatment system, such as a particulate trap, catalyst, or a combination of both. In Table 5, emission results from two different particle traps are presented (31). Table 6 shows the emission results of exhaust that is after treated with a catalyst in combination with a particulate trap (32). From these results it can be concluded that diesel exhaust after treatment reduces emissions. For the wire mesh particulate trap (31), although general NOX emissions were unaffected, nitrogen dioxide appeared at increased levels in certain modes with increased load and speed. This was investigated in a 13-mode diesel engine driving cycle, which does not include transient driving conditions. A general issue regarding particulate traps is the need to investigate in more detail the regeneration process with respect to chemical composition and biological effects of the exhaust gases.
Alternative Fuel Vehicles. Besides gasoline and diesel, other fuels can be used as the energy source in automotive combustion engines. Egeback and Westerholm made a literature survey of Swedish research programs to evaluate possibilities and present aspects of using alcohols as automotive fuels (33). Research programs for alternative fuels depend strongly on political considerations, as well as the price and access to crude oil. The use of alcohol fuels depends not only upon possible replacement of some share of imported motor fuels and crude oils with domestic fuel constituents, but also on possible environmental improvements, such as reduction of exhaust emissions. Table 7 lists emissions data from light-duty vehicles, originating from the U.S. FTP-72 driving cycle (34) fueled with 15 and 95% methanol in gasoline. Despite the measures taken for light-duty vehicles, there is still a need to reduce exhaust emissions from heavyduty vehicles, especially vehides intended for operation in urban areas. The technologies used in Sweden for compression ignition engines are a dual fuel system (35,36), and engines run on alcohol with an ignition improver (36,37). Of these two engine concepts, the latter seems to have the greater potential to reduce emissions when used in vehides operated in cities. Presented in Table  8   emissions from cars fueled with either gasoline or a 15% methanol-gasoline blend. As can be seen in Table 9 there is an increase of emissions at lower ambient temperatures. This shows that published vehicular emission factors that are normally obtained at approximately 20°C are underestimated for those areas with mean ambient temperatures below 20°C. This fact must be con-sidered when performing risk assessment of air pollutants.  air toxin; it is also harmful to used to clean the exhaust er vehicles. Also, there are con diesel fuel that should be i reduced. The requirement to motive emissions has led to a for the development of internal engines and vehicles. For a lo main parameters for improvem nal combustion engines were ft tion and driveability, and thes are still of high priority for developer. The challenge is i grate the emission requirement design, and to minimize the flict between good emission and high fuel efficiency. Wh( missions obtained in the bus cycle driving conditions (N= 3). Two vehicles no fast process for developing standards enger bus (Vi) and a heavy-duty diesel truck (V2). Only three of eight fuels and regulations in order to reduce the andard diesel fuel) (D6), commercial diesel fuel for city buses (08), and one tailpipe emissions. Sweden  cles are commonly equipped with this engine type. Swedish manufacturers have successfully developed fuel-efficient, turbocharged engines with intercooler. Today the catalysts emission technology, the vehicle itself also the great majority of new engines for heavy mitted from must be integrated into the discussion. The duty vehicles on the Swedish market are nponents in design of the vehicle body, rolling friction turbo-charged. Since the early 1980s, removed or coefficient, and the weight and load are repeated investigations of emissions from reduce auto-greatly important since they impact emis-heavy-duty diesel-fueled vehicles in Sweden new strategy sions, as well as fuel consumption.
have yielded some important results, which [combustion Gasoline-fueled Vehicles. In Europe, air have been presented here. Since Sweden )ng time, the pollution from motor vehicles had become has not yet introduced mandatory emission ient of inter-a problem by the early 1960s. The num-standards for heavy-duty vehicles (except iel consump-ber of vehicles, especially passenger cars, for smoke), there is great variation in emis-,e parameters increased rapidly, and the emissions levels sion behavior from one type of vehicle to the engine of carbon monoxide, hydrocarbons, and another. The manufacturer of the vehicle now to inte-lead from the individual vehicles became has not had any other obligation for tech-been agreed upon; the customers, which are primarily community-owned bus companies, require that the new buses meet certain emission standards. In the 1993 model, Sweden introduced mandatory emission requirements for all heavy-duty vehicles meaning that they will be required to meet the following standards: 1.2 g/kWh, 4.9 g/kWh, 9.0 g/kWh, and 0.4 g/kWh for HC, CO, NOX, and particulates, respectively, when measured according to the 13-mode test method defined by the Stockholm group. The technologies that are being developed to reduce emissions from heavy-duty dieselfueled vehicles include engine modification, improved diesel-fuel, and exhaust-gas after treatment (38). The question of which technology should be used also depends on the air quality objectives. If the objective is to reduce those substances in the exhaust suspected of increasing cancer incidence, improvement of fuel quality, in combination with a catalyst or a particulate filter most probably will have to be used. On the other hand, to meet the 1993 Swedish emission standards, the only measure the vehicle manufacturer will have to take is to modify the engine, a modification that is not clearly defined. There is a full range of technical improvements available, such as turbo-charging, intercooling, in-cylinder air motion or turbulence (swirl or noswirl) matched with the type of fuel injection system, and combustion chamber configuration (48). No-or low-swirl motions require high injection pressures and multihole injectors in order to atomize the fuel, and to achieve a complete air-fuel mixture.
Other engine modifications include injection duration and time, as well as combustion duration. An advanced technology is the use of electronically governed fuel injection of the engine. All measures taken to reduce the peak temperature in the combustion chamber during the combustion mode will decrease NO. emission. Several possibilities for reducing the peak temperature include turbo-charging in combination with intercooling (if not used to increase the power output of the engine), retarded ignition timing, and exhaust gas recirculation (EGR). On the other hand, there is the risk that retarded ignition timing and EGR will increase particle emission. The increase of the smoke level because of retarded ignition timing can be overcome by an increase of the injection rate and in-cylinder air motion. The example shows that those techniques which do not increase other emission components must be used. It also shows that optimization of engine and combustion parameters is possible. Hydrocarbon emis-   Table 7. Mean emission factors levels obtained in the U.S. FTP-72 driving cycle using different light-duty passenger vehicle fuel/ engine concepts: lead free, 15% methanol (V1); lead free, 15% and three-way catalyst X-sond (V2); and 95% methanol and catalyst without feedback control (V3) (34  sion levels from a well functioning diesel engine are usually low if compared to the standards. Specific components in the exhaust, such as PAC, light aromatics, aldehydes, etc., can be efficiently reduced with a catalytic converter (39). Catalysts coated with precious metals (platinum, palladium, rhodium, and silver) are effective in oxidizing CO and gaseous HC, particulate soluble organic fraction (SOF), and other components, such as aldehydes, ketones, etc., but are not efficient in oxidizing the soot emissions (40). There are differing opinions among the heavy-duty vehicle manufacturers about particulate emissions. Some manufacturers are convinced that the U.S. 0.1 g/ Bhph (gram/ brake horse power hour) particulate emission standard can be met by an efficient engine modification in combination with an oxidation catalyst. Other manufacturers believe that a particle filter may solve the problem by trapping the particles with a specially designed filter element. This filter poses the difficulty of getting rid of the trapped particles to regenerate the filter. Different ideas for filter regeneration have been tested or are under investigation (41). Many solutions are under consideration, such as heating the filter with a burner or electricity, or by using a special additive. One solution may be to coat the filter element with a precious metal.

Fuels
In this section we discuss gasoline, diesel fuel, methanol, ethanol, and natural gas as automotive fuels. There are alternative fuels to methanol, ethanol, and natural gas, such as rape oil and hydrogen, but there is very little tested experience with these fuels. Important gasoline fuel parameters are octane numbers (RON and MON), density, lead and benzene content, volatility, vapor pressure and content of aromatics, olefins, and paraffins (42). To meet car manufacturers' requirements for a highoctane fuel, lead was added to the fuel. There has also been a strong demand for a highly volatile fuel, especially during winter.
It became evident 20 to 30 years ago that measures had to been taken to improve fuel quality for environmental reasons. Lead in fuel has been proven to cause brain damage in children living in areas with high levels of pollution from gasoline-fueled vehicles. Another reason for keeping lead out of gasoline is that lead acts as a poison to the catalyst. In the section of this article titled Organic Halides, we showed that the scavengers used in connection with lead contribute to the emission of dioxins as well. It is well known that benzene is a harmful substance to man, and that a high level of aromatics in the fuel increases the exhaust emission of benzene and PAH. It has also been shown that high fuel volatility increases the evaporative emission, and that a certain level of olefin may cause a buildup of deposits in the engine's intake system (carburetor and intake valves).
To improve the quality of gasoline, different measures have already been taken and more will be taken in the future. The most progressive step taken has been an agree-ment between the U.S. Environmental Protection Agency and the oil industry to establish rules and implementation guidelines for introduction of reformulated gasoline (43). The proposal states: Reformulated gasoline produced before March 1, 1997, will be certified by EPA if it results in no increase in oxides of nitrogen (NOr); contains no more than 1.0 volume percent of benzene; contains at least 2.0 weight percent of oxygen; contains no heavy metals unless valved; and meets or is below the following Reid Vapor Pressure (RVP) specification during the ozone season: In Class B areas, 7.2 psi RVP, and in Class C areas, 8.1 psi RVP.
Similarly, in Sweden, a new standard that specifies both summer and winter gasolines has been approved (44). The summergrade fuel will have an RVP of 45 to 75 kPa (6.53 to 10.88 psi) and the winter grade fuel 65 to 95 kPa (9.43 to 13.78 psi). The specification for benzene remains at 5.0% by volume.
Although diesel fuels can be of different blends, there is a well-pronounced demand from engine and vehicle manufacturers that the diesel fuel fulfill a certain standard. Important parameters from the fuel and engine manufacturers' points of view are cetane number/ cetane index, density, distillation Coald Filtering Pluging Point (CFPP), viscosity, specific energy, and sulfur content. From the environmental point of view, there are certain fuel parameters of special importance: sulfur content, aromatic content (including the PAH content), initial boiling point (IBP), 95% distillation point, density, and cetane number.
In the discussion between environmental authorities and engine manufacturers, much emphasis has focused on sulfur and aromatic content in fuels. The impact of sulfur and aromatics in the fuel on the exhaust emissions has been the subject of investigations (45,46). The impact of sulfur has been stressed particularly, since sulfur has been shown to contribute to the particulate emission because of sulfate formation in the exhaust. A secondary sulfur reaction contributes to the deterioration of catalysts and particle filters. Reactions between sulfur and components such as phosphorus and calcium in the oil or fuel may also dog the exhaust treatment systems. In the earlier section, Fuel Effects, we discussed the impact of aromatics and PAH in diesel fuel. Results from an extensive investigation of eight fuels used in two different Swedish manufactured vehicles have clearly shown that aromatics and PAH have a measurable and certain impact on the emissions (25). Since the vehicles in the cited Environmental Health Perspectives 20 investigation were not equipped with exhaust after treatment devices, the question of whether the impact of aromatics will diminish with use of such exhaust after treatment devices remains to be answered.
Methanol and ethanol are two alternative fuels which have been tested in both spark ignition and compression ignition engines. A widely accepted idea is that the use of alcohols as automotive fuels will decrease exhaust emissions. Many investigations have supported this idea, especially those concerning the emission of PAH and NO. from alcohol-fueled vehicles. The investigations have also shown that emission components such as aldehydes, alcohols, and alkylnitrites will increase to a certain extent. Therefore, all vehicles fueled with methanol or ethanol should also be equipped with an efficient catalyst in order to perform as clean vehicles. Physically, methanol and ethanol are more suitable fuels for ignition combustion engines than for compression ignition engines, since both methanol and ethanol have a high octane number and a low cetane number. To be used in a compression engine, either some type of ignition improver must be used, or the engine must be equipped with a spark plug or glow plug, or be modified in some other way. In addition, it is common that the compression ratio must be increased to take advantage of alcohols as automotive fuels.
Natural gas is used both as a raw material in the methanol production process and as an automotive fuel. Attempts to use natural gas as a fuel for compression ignition engines have not been successful, especially when the fuel has been used as a means to reduce emissions. Natural gas has a low cetane number, so ignition must be supported in some way. In practice, this can be done by using a dual fuel system, in which diesel fuel can support the ignition of the gas. Since natural gas is a clean fuel, there is a great potential in using it in ignition compression engines to reduce harmful emissions. The problem with natural gas is that more emission of nitrogen oxides may occur than when using gasoline. The vehicle should therefore be equipped with a closed-loop fuel system in combination with a three-way catalyst.
Liquefied petroleum gas (LPG) is also used as an automotive fuel. Most of what has been said about natural gas can be applied to LPG. Tests have confirmed that the level of the NOX emission can be three times higher when using LPG than when using diesel fuel in a heavy-duty engine. On the other hand, the level of the emission of particles, PAH, and some other  components was very low, as was the mutagenicity (37).

Ignition Compression Engine Development
The technical development of the ignition compression engine and catalyst systems has been very successful. The emission of all except a few components, such as nitrous oxide and hydrogen sulfide, has been reduced to very low levels, as has the associated mutagenicity. The need still remains to take advantage of the potential of the three-way catalyst system; the emission during startup and warmup of the vehicle must be improved, especially the cold-start behavior. The emission standards applied in Sweden, as well as in other European countries and the United States, should be strengthened in order to force car manufacturers to take advantage of the three-way catalyst technology. Today, the compression ignition engine is the most energy-efficient engine. The problem remains that the emission potential is not as good as for the ignition compression engine. Despite the fact that some important improvements have been reached as a result of engine modifications which so far have been made, much more can be done. The possibilities for environmentally improving the engine are found in further engine modifications, exhaust after treatment devices, and improved diesel fuels or alternative automotive fuels. There is still considerable potential to reduce the emission from heavy-duty vehicles. Methanol, ethanol, natural gas, or LPG are alternative automotive fuels with a certain deaner emission potential. Ethanol is meant to be a typical biofuel (i.e., it is not manufactured from petroleum products). To take advantage of the emission potential of the alternative fuels, there is a need for extensive development and adaptation of the ignition compression engine and of the compression ignition engine. There is also a need for extensive development of new catalytic converter systems. When developing the engines and exhaust gas after treatment devices, the difference between alcohols and gaseous fuels must be considered.

Conclusions
In this project, the importance of measuring the contribution of semi-volatile associated compounds in automotive exhaust has been confirmed for both gasolineand diesel-fueled engines. Furthermore, exhaust emissions must be thoroughly characterized, induding both regulated and unregulated exhaust emission constituents and bioassay tests. Gasoline-fueled vehicles were investigated at different cruising speeds, and an increased emission of PAH, 1-nitropyrene, particulates, and mutagenic activity was determined at higher cruising speeds. Catalyst-equipped Volume 102, Supplement 4, October 1994 vehicles showed a dramatic decrease of these emission factors when compared with those from the vehicles without catalyst.
Fuel-dependent gasoline exhaust emissions exhibited an increase in PAH (particulateand semi-volatile) emission in relation to increased fuel PAH content. The increase was linearly dependent, giving a term that can interpreted as the PAH formed in the combustion process. A large proportion of fuel PAH (>95%) is decomposed in the combustion process. Lowering the ambient temperatures will result in an increase of exhaust emissions (both regulated and unregulated pollutants) from engines (cold-start). It is important to consider this fact when performing risk assessment of air pollutants specifically for countries with mean ambient temperatures lower than 20°C. It has been confirmed from a diesel-fuel investigation that there exists a quantifiable relationship between fuel variables of the diesel-fuel blends, and the variables of the chemical emissions and their biological effects. Important diesel-fuel parameters are: density, 90% distillation point, final boiling point, specific energy, total aromatics, di-aromatics, tri-aromatics, and PAH contents.
From multivariate analysis performed, the PLS regression coefficient for regulated pollutants was found to be lower than that for unregulated pollutants; biological activity indicates that improving the quality of the fuel can have a greater impact on the unregulated pollutants in the exhaust emission and the biological effects of constituents in the exhaust. The PLS regression coefficients obtained are most likely different if the vehicles are equipped with exhaust after treatment. A general issue with the use of particulate traps is the need to investigate in more detail the regeneration process with respect to chemical composition and biological effects of the exhaust gases. It has been shown for a particulate trap that an increased nitrogen dioxide emission was measured in modes with increased load and speed compared to no exhaust after treatment.
Our study indicates that there is a need to update emission factors from research and today's engines/ vehicles that are alternatively fueled, with respect to both regulated and unregulated exhaust emissions, to be performed in conjunction with bioassay tests. The future engine for heavy-duty vehicles will probably be an Otto-cycle-equippedengine with a three-way catalyst, or some other type of engine, such as a lean burn, stratified charge or Sterling engine in a hybrid system, especially for city buses. Because of extended research and increased understanding of the health impact of exhaust emission constituents, and the increasing variety of fuels that may be used, the list of regulated pollutants is expected to be extended in the future. Further studies are needed on formed secondary pollutants that originate from mobile sources to estimate future potential environmental and health impacts.