1. Introduction
Modern society is critically dependent on the transport of goods and people. In 2015, the world had around 1.1 billion light-duty vehicles (LDVs) and 380 million trucks [1], and these numbers are expected to grow, mostly in non-OECD (short for Organization for Economic Co-operation and Development) countries such as India and China, to 1.7–1.9 billion by 2040 [1], [2], [3], [4]. Transport accounts for about 20% of all energy used and around 23% of global carbon dioxide (CO2) emissions [5]. However, transport contributes only around 14% of global greenhouse gas (GHG) emissions—an amount that is comparable to the share of livestock farming for meat and dairy products [6]—if gases such as methane are included. At present, internal combustion engines (ICEs) power transport almost entirely (> 99.9%), with reciprocating engines powering land and marine transport, and jet engines powering air transport. Spark ignition (SI) engines power around 80% of all passenger cars across the world [4], while diesel engines dominate the commercial sector (road and marine use). LDVs use around 44% of the global transport energy [7] although they are much greater in number compared to commercial vehicles.
Petroleum-derived liquid fuels currently provide around 95% of transport energy, and roughly 60% of crude oil produced is used to make transport fuels [2], [3], [4], [7]. The demand for transport fuels across the world is very large, at around 4.9 billion liters each of gasoline and diesel and 1.3 billion liters of jet fuel each day [8], with an expected yearly growth of around 1% [2], [7]. Changes in the transport sector are occurring because of increasing demand driven by increasing population and prosperity; the need to ensure energy security, control GHG emissions, and improve local air quality; and in response to consumer preferences and demands. The importance of each of these drivers for change will differ in different countries and at different times. The battery electric vehicle (BEV) and the fuel cell could replace ICEs, and fuels such as natural gas and biofuels are possible alternatives to conventional liquid fuels made from crude oil. However, these alternatives to the existing system all start from a very small base, face critical barriers to unrestrained and quick growth [8], and—even by 2040—are not projected to account for more than around 10% of global transport energy [2], [3], [7]. It is important that these alternatives be assessed on a life-cycle basis to ensure that the environmental and other benefits are real, and that the burdens are not simply shifted from the engine tailpipe to somewhere else.
Thus, transport will be powered largely by ICEs using mostly petroleum-based fuels for decades to come [2], [3], [7], [8]. Furthermore, over this time scale, the shortage of oil will not constrain growth in transport; known reserves of oil have been increasing faster than consumption over the last several decades, and the current reserves should last for at least the next 50 years at current consumption rates [8], [9]. Therefore, it is imperative to improve the efficiency, environmental impact, and affordability of ICEs, which will mostly continue to power transport in the foreseeable future.
In general, there is much more scope for reducing the fuel consumption of LDVs than of commercial vehicles, with the result that the demand for transport energy in the commercial sector is expected to rise much faster than demand in the passenger car sector in future [2], [3]. More oil will have to be processed to produce the increasing amounts of diesel and jet fuel needed for the commercial sector, and the production of low-octane gasoline components, which are collectively known as naphtha, will increase proportionately. Naphtha is usually processed further to make gasoline, although it is also used to make petrochemicals. Since gasoline demand is not expected to increase at the same rate as the demand for diesel and jet fuel, it is highly likely that the availability of naphtha will increase in future. Refineries will be required to make very large investments to meet this changing fuel-demand structure, and will also need to find an economic use for naphtha in transport fuels in order to maintain their commercial viability [10], [11].
A great deal of potential exists for improving the efficiency of ICEs via improvements in combustion, control, and after-treatment systems. Fuel consumption in SI engines [2] can be further reduced in future by partial electrification in the form of hybridization, which is likely to be widely deployed [2]. Hybridization allows an SI engine to run more efficiently and enables the recovery of energy lost in braking. Heavy-duty vehicles are less likely to have the driving modes with frequent starts and stops that are more common to passenger cars, and usually run on diesel engines, which are already efficient; thus, electric hybridization will be less beneficial in such vehicles. However, if future engines are not obliged to run on current market fuels, there is greater scope in developing engines and fuels together for increased benefits in efficiency and emissions at affordable cost. Such approaches could also provide a pathway to use fuel components such as low-octane gasoline components, which are likely to be in surplus and could therefore be available at a lower price to the consumer.
This paper discusses the issues described above, but draws heavily on previous reviews by this author [8], [12], [13], [14]. The paper first discusses current fuels and engines and current projections on fuel supply and demand in brief. It then discusses future developments in ICEs and fuel/engine systems.
2. ICE fuels and combustion systems
Details about engine fuels and combustion processes have already been discussed in several books [15], [16], [17], [18], so this section contains only a brief summary.
2.1. ICE fuels
Transport has evolved to be almost entirely powered by liquid fuels because of their high energy density and ease of transport and storage. For example, at normal temperature and pressure, a liter of gasoline contains over 700 times more energy than a liter of natural gas and over 3100 times more energy than a liter of hydrogen gas. In order to carry enough mass on a vehicle to get a reasonable range, a great deal of energy must be used to liquefy or compress a gaseous fuel. Over the past century, an extensive global network worth trillions of dollars, which will be difficult and expensive to replace or replicate, has developed around the use of liquid fuels for transport.
Transport fuels are mostly made by refining crude oil (petroleum). The first step is the distillation of the crude oil. When oil is heated above ambient temperature, gases dissolved in the crude oil are released; these gases make up liquid petroleum gas (LPG). LPG can be up to 2% of crude oil, and consists mostly of propane and some butane. The fraction of crude oil that lies within the gasoline boiling range, with boiling points between ∼20 and ∼200 °C, from the initial distillation is known as straight run gasoline (SRG). Diesel fuels are made up of heavier components with boiling points ranging from ∼160 to ∼380 °C. Heavy components, with boiling points higher than 380 °C, can constitute 40%–60% of the weight of crude oil, depending on the source of the oil. In the refinery, these heavy components are first “cracked” into smaller molecules, which are further processed to produce useful products, for example by reducing sulfur or changing their octane/cetane number. The products in the gasoline boiling range from different parts of the refinery are collectively known under the generic term “naphtha.” Naphtha is usually processed further to increase its octane number; it is also used in the petrochemicals industry. Other nonpetroleum components such as biofuels and high-octane components such as methyl tertiary butyl ether (MTBE) and ethanol are blended with refinery components, along with some fuel additives, to meet the required fuel specifications [15], [16].
Knock, an abnormal combustion phenomenon, limits efficiency in SI engines. Hence, gasolines need to have high anti-knock quality, as specified by the research octane number (RON) and motor octane number (MON) [15], [19], [20]. Most market gasolines have RON > 90. For diesel fuels, autoignition quality is measured by the cetane number (CN). Diesel fuels generally require a high CN because they need to autoignite easily; practical diesel fuels have CN > 40. The higher the RON of a fuel, the lower is its CN, and vice versa [15], [19]. The CN of jet fuel is lower than that of conventional diesel fuel, and jet fuel is blended using more volatile components in the diesel boiling range. Marine transport fuels are blended from the heaviest components in the fuel pool, and have a high sulfur content. In the future, marine engines could be forced to run on conventional diesel fuel because of current moves to reduce the sulfur content of marine fuels; such a shift would further contribute to increasing the demand for diesel fuel in the future.
Gasoline-like fuels are defined in this paper as fuels with CN < 30 or RON > 60—that is, as fuels within the gasoline autoignition range, as specified in Ref. [19].
2.2. Engine combustion systems
Two major practical ICE combustion systems are used in land and marine transport. In SI engines, an electric spark initiates flame propagation to release energy from a mixture of fuel and air that is compressed after premixing. In modern SI engines, the fuel/air ratio must be maintained at stoichiometric levels to enable the three-way catalyst to be used effectively in order to reduce carbon monoxide (CO), unburned hydrocarbons (HC), and nitrogen oxides (NOx) in the exhaust to acceptably low levels. In terms of mass, soot emissions from SI engines are negligible; however, the number of nanoparticles (i.e., particles less than 100 nm in diameter) is of increasing concern. Gasoline particulate filtersare likely to be increasingly required in future to address particulate emissionsfrom SI engines. For any given fuel, the maximum temperature and pressure that can be reached ahead of the expanding flame front, in the “end gas,” are limited by knock, which is caused by autoignition in the end gas [15], [19], [20]. Knock depends on the time available for autoignition and on the anti-knock quality of the fuel. Load is reduced in SI engines by the use of a throttle to reduce airflow, since the amount of fuel energy cannot be reduced independently of airflow due to the fixed air/fuel ratio. In contrast, compression ignition (CI) engines do not use a throttle because load is controlled by controlling the amount of fuel that is injected. SI engines compress a mixture of fuel and air; in contrast, CI engines compress mostly air during the compression stroke [17], [18]. As a result, pumping losses are higher in SI engines than in CI engines. For all these reasons, SI engines convert only 20%–25% of the fuel energy to motive power over a typical driving cycle and are less efficient compared to CI engines. However, their impact on pollution in terms of HC, NOx, CO, and particulate mass is low because of the use of effective after-treatment. The engines used in commercial transport are larger and heavier than those used in passenger cars, so they have to run at lower speeds. Knock is more likely to occur in larger engines at lower speeds than in smaller engines running at higher speeds because of the greater amount of time available for autoignition. Hence, SI engines are not usually used in commercial transport.
In CI engines, fuel is injected into the high-pressure and high-temperature environment near the top of the compression stroke and heat release is initiated by autoignition as the fuel mixes with oxygen. At present, all practical CI engines use diesel fuel and are diesel engines. Soot (particulates) and NOxemissions are a significant problem for diesel engines, and technology such as complex after-treatment and a high-pressure injection system are needed to control them. Modern diesel engines are hence much more expensive than SI engines of similar size, although they are more efficient.
Recently, there has been much interest in homogeneous charge compression ignition (HCCI) combustion. In an HCCI engine, fuel and air are fully premixed, as in an SI engine, but heat release occurs by autoignition, as in knock in an SI engine. The thermal efficiency of HCCI engines is very high but they are limited to operating at low loads (i.e., lean mixture strengths) because of excessive pressure rise rates at richer equivalence ratios. Friction losses are proportionately higher at lower loads, and the brake efficiency of HCCI engines, at the loads they can operate, will be lower. Greater mixture or temperature stratification can increase the upper load limit of an HCCI engine but such engines should not be termed HCCI engines because the mixture is not homogeneous. In HCCI engines, at the low loads they are constrained to run, the soot and NOx levels can be exceptionally low. In diesel and SI engines, in-cycle control over the phasing of heat release is provided by the timing of the final fuel injection and of the spark, respectively. Such in-cycle control is not possible in HCCI combustion, which makes its implementation difficult in practical engines. If a certain level of inhomogeneity/stratification is obtained by the late injection of the final fuel pulse, in-cycle control of combustion can be restored and HCCI-like combustion—in which fuel and air are “premixed enough” to result in low soot and NOx—can be obtained. Gasoline compression ignition (GCI) and reactivity-controlled compression ignition (RCCI) are two such approaches, and are discussed below.
The evolution of new technology such as GCI requires collaboration among all stakeholders—including the auto and oil industries and governments—and is affected by strategic issues such as the supply and demand of transport energy.
3. Future development
Current efforts to improve ICEs are focused on using existing market fuels. This is clearly essential because it is very difficult to change both fuels and engines simultaneously in the marketplace.
Significant developments are underway to improve the efficiency of SI engines almost to the level of diesel engines [21], [22], [23], [24], [25]. For example, fuel consumption for the best-in-class passenger car in the US market is already around 16% lower than the US fleet average for cars of similar size and performance [21]. Approaches under development include lean burn, downsizing, and turbocharging, along with CI using market gasolines [24], [25]. Many supporting technologies are also being developed [21], [22], [23] to ensure that these efficient engines meet stringent exhaust emissions requirements. The gasoline direct-injection compression ignition (GDCI) engine using partially premixed compression ignition (PCI) has demonstrated diesel-like efficiency [24] using US market gasoline. The expectation is that combustion system developments alone will reduce fuel consumption by around 30%, in LDVs in comparison with the 2015 fleet average using SI engines [21], [22], [23]. With additional technologies such as light-weighting and hybridization, this improvement could reach around 50%. As fuel consumption decreases, the GHG impact of ICEs will decrease proportionately, and will reduce any advantages BEVs running on renewable electricity might have in terms of GHG emissions.
Similarly, after-treatment systems have been and continue to be developed to reduce exhaust pollutants such as particulates, NOx, CO, and HC [26], [27], [28]. For example, modern diesel particulate filters (DPFs) and gasoline particulate filters almost entirely eliminate particulates from ICE exhausts [27], [28]. A warmed-up catalyst in a modern car can reduce HC emissions in the exhaust to almost zero—certainly well below ambient air levels in many urban areas [26]. Even NOx levels in diesels can be reduced to levels 10 times lower than European limits set for 2020 with a modern exhaust catalyst and intelligent management of combustion temperatures and modes [28].
3.1. Fuel implications in the short term
Engine combustion system developments also have implications for fuels. For example, the design trend in SI engines has been to increase the pressure in the cylinder for a given unburned gas temperature in order to improve power density and efficiency. This makes autoignition in the end gas, leading to knock, more likely. High anti-knock fuel quality will help to avoid knock and enable higher efficiency SI engines. Pressure to increase the anti-knock quality of gasolines in order to enable high-efficiency SI engines will grow. For example, there are suggestions [29] that by 2040 all gasoline in the United States should have RON > 98, whereas currently US regular, the most commonly used gasoline in the United States, has a RON of around 92. Whether such a change will bring about benefits in terms of GHG emission reductions needs to be assessed on a life-cycle basis, and the answer may differ for different refinery configurations. Such increases in RON will require big changes and investments in refineries and will lead to a further increase in the availability of low-octane gasoline components such as naphtha [10], [11] because the opportunity to blend them in gasoline will decrease. The importance of high-octane components such as ethanol, MTBE, di-isobutylene, and methanol will also increase.
The question of how fuel anti-knock quality should be defined in such modern engines is an important one [14], [19], [20], as this definition has major implications for fuels manufacture, which is geared toward meeting gasoline anti-knock specifications. Gasoline anti-knock quality is currently defined by RON and MON. These are measured by comparing the gasoline with blends of iso-octane and n-heptane, known as primary reference fuels (PRFs), in the single-cylinder Cooperative Fuels Research (CFR) engine, according to test procedures set by the American Society for Testing and Materials (ASTM). For practical fuels, RON is higher than MON; the difference between them is known as the sensitivity, S. The MON test is run at a higher intake temperaturecompared to the RON test and the pressure for a given unburned mixture temperature is lower in the MON test than in the RON test. Practical gasolines contain aromatics, olefins, and oxygenates, which respond very differently in chemical kinetic terms to increasing pressure in comparison with the PRFs that are used to define the RON and MON scales. Practical fuels are much more prone to autoignition and knock under the MON test conditions than PRFs. However, SI engines have been moving away from the MON test conditions, as designers have increased the mass of air (pressure) in the engine without increasing the unburned gas temperature too much in order to increase efficiency and power density [19], [30]. In fact, a lower MON fuel, for a given RON—that is, a fuel with higher sensitivity for a fixed RON—has better anti-knock quality in modern engines [14], [19], [20]. However, in many areas including the United States and Europe, the MON is considered to contribute to anti-knock quality—that is, a high MON is considered desirable. As engine designers seek to further improve engine efficiency, this mismatch between specifications and engine requirements will widen and will have to be addressed. One approach might be to replace the octane scale with a different scale based on toluene/n-heptane mixtures (toluene reference fuels, TRFs) rather than PRFs. The fuel would be tested using the RON test and assigned a toluene number (TN), the volume percent of toluene in the TRF that matches the test fuel for knock [31]. At the very least, countries that specify gasoline anti-knock quality using RON alone, such as Japan, should not introduce a minimum MON specification.
In addition to knock, conventional fuel-related concerns such as spark ignition, flame development, deposit formation and control [12], and pollutant formation will continue to be of importance as SI engines seek ever higher efficiency. There is persistent pressure on fuel manufacturers to reduce sulfur levels in both gasoline and diesel to enable effective after-treatment systems. Fuel additives [12], [32] are routinely used to control deposits in the fuel system.
There is greater scope for the development of affordable and highly efficient new fuel/engine systems to meet increasingly stringent requirements on GHG emissions and local air quality if engines are not confined to using current market fuels. Of course, such a shift will require cooperation between auto and oil industries and other stakeholders, and will probably happen in the mid to long term. Some of these possibilities are discussed below.
3.2. Gasoline compression ignition
This section borrows heavily from a recent review of GCI [33], which also carries a more comprehensive list of relevant papers.
Diesel engines have a lower GHG footprint because of their high efficiency; however, they bring increasing concerns regarding particulate and NOxemissions. Diesel fuel ignites very quickly after injection and burns in a quasi-steady jet diffusion flame [34] before it has a chance to mix sufficiently with oxygen in the cylinder, under most operating conditions. There is a much better chance of minimizing soot formation if the equivalence ratio, φ, of the mixture packets where combustion occurs is no greater than around 2 [34]. If the combustion temperatures are kept below around 2200 K, usually by using exhaust gas recirculation (EGR) in practical engines [34], [35], NOx formation can be minimized.
Most of the soot formed is oxidized inside the diesel engine. Both the oxygen content and the temperature in the cylinder decrease if the EGR level is increased in order to control NOx. Then soot oxidation decreases and, if any soot has been formed, the engine-out particulate matter (PM) or soot increases, leading to the well-known PM/NOx tradeoff in diesel engines. If fuel and oxygen are mixed sufficiently well before combustion starts, soot formation can be reduced or avoided. NOx emissions can then be controlled using EGR without increasing engine-out soot. However, such a combustion system increases the engine-out HC and CO, which must be controlled with appropriate after-treatment.
The final injection of fuel must be completed sufficiently before combustion starts in order to avoid soot formation. Partially premixed combustion (PPC) or PCI to avoid soot formation has been defined by Ref. [36] as occurring when the final fuel injection is completed sufficiently before combustion starts in order to ensure that the engine-out smoke has a filter smoke number (FSN) of below 0.05. A time constant relevant to PPC is the ignition dwell, IDW = SOC – EOI, where SOC and EOI are the crank angles at the start of combustion and the end of the final fuel injection, respectively. The IDW must be positive in PPC.
When diesel fuel is used, PPC can be promoted by reducing the injection pulse width and accelerating mixing by increasing the injection pressure at a given operating condition. Even with modern injection systems, PPC is only possible at low loads if diesel fuel is used, and NOx and particulates must be controlled by sophisticated after-treatment systems in conventional diesel engines. The global mixture strength is lean of stoichiometric and the engine exhaust always has oxygen in diesel engines; it is particularly difficult to reduce NOx in such an oxygen-rich environment. The difficulties presented by the use of diesel fuel can be addressed by sophisticated after-treatment technology, which makes the modern diesel engine expensive. Pollution-reduction strategies can also increase fuel consumption. For example, when the exhaust temperature is low at low loads, the DPF accumulates soot that must occasionally be burned off to regenerate the DPF using extra fuel.
In a single-cylinder heavy-duty engine, Kalghatgi et al. [37], [38] have demonstrated that extremely low soot and NOx could be obtained if gasoline-like fuels with greater resistance to autoignition were used. This is GCI, which is in fact a more practical way of achieving HCCI-like combustion. The important difference is that the fuel and air are not fully premixed in GCI, unlike in an HCCI engine. Combustion phasing can be controlled in-cycle as in a diesel engine by the timing of the last fuel injection (when multiple injections are used). This ensures that there is enough inhomogeneity within the cylinder to ensure that autoignition starts even under operating conditions when autoignition and hence HCCI combustion might not be possible at all (e.g., low load). Thus, in GCI, while fuel and air are “premixed enough” to ensure low or no soot formation, they are not fully premixed. Several other research groups have also demonstrated that if more mixing time is enabled by the use of gasoline-like fuels with greater resistance to autoignition, the control of particulates and NOx in CI engines is much easier [24], [39], [40], [41], [42], [43], [44], [45], [46], [47], [48], [49], [50], [51], [52], [53], [54], [55], [56].
3.2.1. Advantages of GCI engines
The injection system in a GCI engine should be cheaper than in a modern diesel engine because GCI engines do not require high injection pressures. In fact, in such engines, lower injection pressures improve the stability of combustion at low loads, presumably by enabling increased inhomogeneity [33], [48], [51]. In GCI engines, the focus of after-treatment shifts from the simultaneous control of soot and NOx to the oxidation of HC and CO. A GCI engine, like a diesel engine, will run lean overall and the exhaust will contain oxygen; thus, the oxidation of CO and HC should be easier to accomplish than the reduction of NOx. A DPF might be needed to cope with higher soot emissions at high loads in a GCI engine. However, soot is not formed in GCI engines at low loads, and the DPF does not accumulate soot and will need to be regenerated less often, if at all. Although soot might be formed at high loads, the DPF might be mostly self-regenerating, without requiring extra fuel, as the exhaust temperature may be high enough. Hence, the after-treatment system could be simpler and cheaper in a GCI engine than in a modern diesel engine.
GCI engines have been demonstrated to have very high efficiencies. Indicated fuel efficiencies of up to 56% have been measured in a heavy-duty engine [39]. Fuel consumption could be reduced by 25% or more in light-duty engines over an operating cycle [24], [45], [46] compared with running the engine on gasoline in SI mode. Further efficiency improvements might be possible in light-duty engines in comparison with conventional diesel operation. For example, in light-duty diesel engines using diesel fuel, noise is a problem at low loads because combustion is initiated by autoignition in a rich mixture. This is true even if the fuel injection is completed before combustion, because of the lower ignition delay (ID, ID = SOC – SOI, where SOI is the crank angle at the start of injection) of the diesel fuel [48], [51], [52]. In modern diesel engines, pilot injection is employed to increase temperatures in order to promote more diffusive combustion of the main fuel injection to alleviate noise; however, doing so results in lower efficiency and a greater amount of soot, which loads up the DPF. This negative result can be avoided by using fuels with a high ID because autoignition occurs in a much leaner mixture than diesel fuel; engine noise is low because the pressure rise rate is very low at low loads [33], [48], [51], and a pilot injection will not be needed. For a given speed, soot and NOxcan be controlled at higher loads in GCI compared with a diesel engine, and the engine could be downsized to meet the requirements over a given operating cycle. There might be further benefits in efficiency because lower injection pressures reduce parasitic losses, and because of less frequent regeneration of the DPF. There will be additional benefits in terms of overall energy consumption and GHG emission reduction from fuels manufacture if the fuels used are less processed than conventional diesel or gasoline.