energies

Article Alcohol Fuels for Spark-Ignition Engines: Performance, Efficiency, and Emission Effects at Mid to High Blend Rates for Ternary Mixtures

James W. G. Turner 1,* , Andrew G. J. Lewis 1, Sam Akehurst 1 , Chris J. Brace 1, Sebastian Verhelst 2 , Jeroen Vancoillie 2, Louis Sileghem 2, Felix C. P. Leach 3 and Peter P. Edwards 3 1 Institute for Advanced Automotive Propulsion Systems, University of Bath, Bath, Somerset BA2 7AY, UK; [email protected] (A.G.J.L.); [email protected] (S.A.); [email protected] (C.J.B.) 2 Department of Electromechanical, Systems and Metal Engineering, Campus UFO, Ghent University, T4, Sint-Pietersnieuwstraat 41, B-9000 Gent, Belgium; [email protected] (S.V.); [email protected] (J.V.); [email protected] (L.S.) 3 Departments of Chemistry and Engineering Science, University of Oxford, Oxford OX1 3PJ, UK; [email protected] (F.C.P.L.); [email protected] (P.P.E.) * Correspondence: [email protected]; Tel.: +44-1225-383466



Received: 21 October 2020; Accepted: 25 November 2020; Published: 3 December 2020


Abstract: This paper follows on from an earlier publication on high-blend-rate binary gasoline-alcohol mixtures and reports results for some equivalent ternary fuels from several investigation streams. In the present work, new findings are presented for high-load operation in a dedicated boosted multi-cylinder engine test facility, for operation in modified production engines, for knock performance in a single-cylinder test engine, and for exhaust particulate emissions at part load using both the prototype multi-cylinder engine and a separate single-cylinder engine. The wide variety of test engines employed have several differences, including their fuel delivery strategies. This range of engine specifications is considered beneficial with regard to the “drop-in fuel” conjecture, since the results presented here bear out the contention, already established in the literature, that when specified according to the known ternary blending rules, such fuels fundamentally perform identically to their binary equivalents in terms of engine performance, and outperform standard gasolines in terms of efficiency. However, in the present work, some differences in particulate emissions performance in direct-injection engines have been found at light load for the tested fuels, with a slight increase in particulate number observed with higher methanol contents than lower. A hypothesis is developed to explain this result but in general it was found that these fuels do not significantly affect PN emissions from such engines. As a result, this investigation supplies further evidence that renewable fuels can be introduced simply into the existing vehicle fleet, with the inherent backwards compatibility that this brings too.

Keywords: alcohols; gasoline-alcohol blends; ternary blends; renewable fuels; e-fuels

1. Introduction The affordability of transport for the end-user is the key factor in the success of the motor vehicle, which has accelerated mankind’s economic development since its inception; cheap transportation is a driver of economies worldwide and this was not possible before the internal combustion engine (ICE). This state of affairs has arisen solely because of its cost-effectiveness and the resulting affordability and utility of the vehicles it is fitted to, coupled to the fact that liquid fuels make the storage and distribution of the energy that they require similarly cost-effective. This overall transport system cost-to-utility

Energies 2020, 13, 6390; doi:10.3390/en13236390 www.mdpi.com/journal/energies Energies 2020, 13, 6390 2 of 31 Energies 2020, 13, x FOR PEER REVIEW 2 of 32 millionratio has vehicles given risewere to manufactured, a situation where, representing for the firsta year-round time in 2016, rate of almost three 95vehicles million per vehicles second, were and onemanufactured, billion passenger representing and commercial a year-round vehicles rate of had three been vehicles built perby 2010 second, [1]. andFurthermore, one billion in passenger a recent study,and commercial Gutzmer of vehicles INA-Scha hadeffler been presented built by data 2010 showing [1]. Furthermore, that car production in a recent rates study, are accelerating Gutzmer of andINA-Schae analysisffler of presented his data data yields showing an estimate that car productionthat the second rates arebillion accelerating vehicles and will analysis have ofbeen his manufactureddata yields an estimatein the 2021–2022 that the second time frame billion [2]. vehicles He estimates will have the been ratio manufactured of ICE-only, in plug-in the 2021–2022 hybrid electrictime frame vehicles [2]. He (PHEVs) estimates and the battery ratio ofelectric ICE-only, vehicl plug-ines (BEVs), hybrid and electric analysis vehicles of his (PHEVs) figures andshows battery that ifelectric there vehiclesis no paradigm (BEVs), shift and analysisin customer of his behavior figures showsthen three that ifbillion there vehicles is no paradigm will have shift been in customer built by 2030behavior [2]. Figure then three 1 presents billion the vehicles data of willGutzmer have in been these built terms; by 2030 Figure [2]. 2 shows Figure 1that presents even with the datasuch ofan optimisticGutzmer in view these of terms;the take Figure up of2 showspure BEVs, that eventhe IC withE will such continue an optimistic to domina viewte oflight-duty the take propulsion up of pure forBEVs, years the to ICE come, will continueand that toeven dominate if whatlight-duty might be propulsiontermed “peak for engine” years to come,occurs and in 2020, that eventhen ifstill what in 2030might nearly be termed as many “peak combustion engine” occurs engines in 2020, will then be still made in 2030 as nearlywere in as 2014, many when combustion they engineswill be manufacturedwill be made as at were a rate in of 2014, 2.66 when per second. they will While be manufactured there are many at a estimates rate of 2.66 in perthe second.literature, While this theredata hasare manybeen chosen estimates because in the INA-Schaeffler literature, this datais one has of been the Tier-1 chosen suppliers because INA-Schaewith a greatffl erinterest is one in of the accelerationTier-1 suppliers of electromobility, with a great interest and inso thethey acceleration can be assumed of electromobility, to present aand scenario so they with can a be rapid assumed take upto presentof electric a scenario vehicles with and a rapidcorresponding take up of substitu electric vehiclestion of ICEs. and corresponding Plainly, despite substitution various political of ICEs. pronouncementsPlainly, despite various concerning political the demi pronouncementsse of the ICE, concerningthe automotive the demise industry of and the ICE,its observers the automotive believe thatindustry they andwill itscontinue observers the believe foreseeable that theyfuture; will ot continueher estimates the foreseeable by industry future; observers other put estimates a global by penetrationindustry observers of PHEVs put and a global BEVs penetration combined ofat PHEVsslightly andless BEVsthan 20% combined by 2030 at [3]. slightly One lessmight than therefore 20% by suggest2030 [3]. that One reports might of therefore the impending suggest death that reportsof the ICE of theappear impending to be greatly death exaggerated of the ICE appear [4]. to be greatlySynergistic exaggerated with [4 ].the vehicle manufacturing industry developing over time, the fuel supply industrySynergistic has kept with pace. the vehicleBecause manufacturing there was noindustry fossil hydrocarbon developing overfuel time,supply the industry fuel supply then, industry early engineshas kept used pace. bio-sourced Because there fuels was and no fossil the hydrocarbonswitch to fossil fuel fuels supply was industry made then,early earlyon, despite engines early used concernsbio-sourced that fuels demand and the would switch toou fossiltstrip fuelssupply was [5]. made Ch earlyeapness on, despiteof the earlyenergy concerns store thatboth demand in the infrastructurewould outstrip and supply aboard [5]. the Cheapness vehicle is ofone the factor energy in the store affordability both in the of infrastructure the system, and and the aboard favorable the economicsvehicle is one of liquid factor infuel the supply affordability has given of the rise system, to a situation and the where favorable the economicsauthors’ analysis of liquid of fuel the supplyglobal usehas givenof oil risesuggests to a situation that it is where used continuously the authors’ analysis at a volume of the flow global rate use equivalent of oil suggests to 7.4% that of it isthat used of Niagaracontinuously Falls. at a volume flow rate equivalent to 7.4% of that of Niagara Falls.

100 5000

80 4000

60 3000

40 2000

20 1000 Cars / [Millions] Manufactured Year per Cumulative Cumulative Cars Manufactured / [Millions]

0 0 2005 2010 2015 2020 2025 2030 2035 Year / [-]

BEVs HEVs and PHEVs ICE-only Cumulative Cars Manufactured

Figure 1. Projected productionproduction volumes volumes of of cars cars split split into in ditoff differenterent propulsion propulsion types types (data (data adapted adapted from fromthat giventhat given in [2]). in BEV: [2]). batteryBEV: battery electric electric vehicle; vehicle; HEV: hybridHEV: hybrid electric electric vehicle; vehicle; PHEV: PHEV: plug-in plug-in hybrid hybridelectric electric vehicle. vehicle. ICE: internal ICE: internal combustion combustion engine. engine.

Energies 2020, 13, 6390 3 of 31 Energies 2020, 13, x FOR PEER REVIEW 3 of 32

100 5000

80 4000

60 3000

40 2000

20 1000 Cars / [Millions] Manufactured Year per Cumulative Cumulative Cars Manufactured / [Millions]

0 0 2005 2010 2015 2020 2025 2030 2035 Year / [-]

ICE (all) Cumulative Cars Manufactured

Figure 2. Data from Figure1 reworked to show vehicles with an internal combustion engine fitted Figure 2. Data from Figure 1 reworked to show vehicles with an internal combustion engine fitted (i.e., pure ICE and hybrid vehicles). The values for cumulative cars built is the same as on Figure1. (i.e., pure ICE and hybrid vehicles). The values for cumulative cars built is the same as on Figure 1. Note that in 2030, nearly as many engines are predicted to be built as in 2014, and the value of 84 Note that in 2030, nearly as many engines are predicted to be built as in 2014, and the value of 84 million engines per year still represents 2.66 engines per second. ICE: internal combustion engine. million engines per year still represents 2.66 engines per second. ICE: internal combustion engine. It should be remembered that the customer is de facto the only financial input to the whole system It should be remembered that the customer is de facto the only financial input to the whole system of vehicle manufacture and transport energy supply; all of the other stakeholders, being the vehicle of vehicle manufacture and transport energy supply; all of the other stakeholders, being the vehicle manufacturers (OEMs), fuel companies and governments, take this money out of the system either as manufacturers (OEMs), fuel companies and governments, take this money out of the system either sales or taxation. If the consumer can no longer afford transportation, then a very worrying state of as sales or taxation. If the consumer can no longer afford transportation, then a very worrying state affairs will ensue. of affairs will ensue. While it is increasingly true that in many urban and extra-urban driving situations, modern While it is increasingly true that in many urban and extra-urban driving situations, modern production IC engines and exhaust after treatment systems actually clean the air (insofar as their production IC engines and exhaust after treatment systems actually clean the air (insofar as their tailpipe emissions of hydrocarbons (HCs), carbon monoxide (CO) and oxides of nitrogen (NOx) are tailpipe emissions of hydrocarbons (HCs), carbon monoxide (CO) and oxides of nitrogen (NOx) are significantly lower than the background concentration) [6,7], a larger global problem is that of the significantly lower than the background concentration) [6,7], a larger global problem is that of the emission of fossil carbon dioxide (CO2). At present, the global atmospheric CO2 concentration is emission of fossil carbon dioxide (CO2). At present, the global atmospheric CO2 concentration is increasing by at least 3 ppm per year, despite global action being taken to limit it. While all possible increasing by at least 3 ppm per year, despite global action being taken to limit it. While all possible steps need to be taken to move from fossil fuels to other energy carriers, as mentioned above it seems steps need to be taken to move from fossil fuels to other energy carriers, as mentioned above it seems to be accepted as fact that an ICE has to operate on such fuels. This is plainly not the case. Referring to be accepted as fact that an ICE has to operate on such fuels. This is plainly not the case. Referring to the above comments on the dawn of motoring, when vehicles used biofuels because there was no to the above comments on the dawn of motoring, when vehicles used biofuels because there was no alternative and consequently there was no meaningful fossil carbon footprint in agriculture, it is an alternative and consequently there was no meaningful fossil carbon footprint in agriculture, it is an uncomfortable fact that early ICE vehicles were more carbon neutral than a BEV is today. Unfortunately, uncomfortable fact that early ICE vehicles were more carbon neutral than a BEV is today. due to economics alone, society made the switch to fossil fuel to ensure the energy supply kept pace Unfortunately, due to economics alone, society made the switch to fossil fuel to ensure the energy with the burgeoning demand. It is not fanciful to state that before this point, when it operated on supply kept pace with the burgeoning demand. It is not fanciful to state that before this point, when biofuels, the early ICE vehicle was probably the most environmentally-friendly mechanized transport it operated on biofuels, the early ICE vehicle was probably the most environmentally-friendly system we have had or will ever have. What we now desperately need to find an answer to is the real mechanized transport system we have had or will ever have. What we now desperately need to find problem: that we presently fuel the majority of vehicles on fossil oil. an answer to is the real problem: that we presently fuel the majority of vehicles on fossil oil. Alcohols are some of the most widely used alternative fuels in the automotive market. The fact Alcohols are some of the most widely used alternative fuels in the automotive market. The fact that they can be made fully renewably and blended with gasoline is potentially a useful tool in the that they can be made fully renewably and blended with gasoline is potentially a useful tool in the drive to reduce the release of fossil CO2 associated with transport, which is one of the most difficult drive to reduce the release of fossil CO2 associated with transport, which is one of the most difficult energy consumption modes to decarbonize. This is especially true when one considers the cost of energy consumption modes to decarbonize. This is especially true when one considers the cost of BEVs for all of the stakeholders in transport, and particularly for the consumer who (if the present BEVs for all of the stakeholders in transport, and particularly for the consumer who (if the present economic model stays in place for the foreseeable future) appears to be expected to pay significantly economic model stays in place for the foreseeable future) appears to be expected to pay significantly more for vehicles than he or she presently does. While electricity may be cheap on a cost-per-MJ basis

Energies 2020, 13, 6390 4 of 31 more for vehicles than he or she presently does. While electricity may be cheap on a cost-per-MJ basis at the moment, and BEVs are approximately three times more efficient in their energy utilization than combustion engines, the price of such vehicles remains cost-prohibitive for many people, and is likely to remain so (especially considering total cost of ownership) even if future stretch targets for automotive traction battery costs are met [8]. Although, for the consumer, electrical energy is cheap in relation to transportation (liquid) fuel costs, there is of course no guarantee that this will continue to be the case. The application of life-cycle analysis (LCA) may also see increasing numbers of questions raised over the ecological impact of electrification, especially since many of the minerals and elements used in the manufacture of traction batteries and motors are extremely rare (for example cobalt, neodymium, dysprosium, and to a lesser extent lithium). There is also the danger of energy security concerns being swapped for concerns over commodity security—although it is true that, unlike the use of fossil fuels in combustion-engine vehicles, commodity security affects only the construction of EVs and not necessarily their use. There are moral concerns regarding how cobalt in particular is obtained, too. Therefore, due to the fact that alcohols can either be used as pure fuels in their own right or blended in many ways with existing fossil hydrocarbon fuels [9–24], there is, in the authors’ opinion, considerable merit in continuing to investigate them as a potential future automotive energy vector. As stated, this paper follows on from a previous work in which the performance of several different mid- and high-blend alcohols were tested and compared to the performance of regular and premium gasolines in a variety of test engines [9]. A key tenet in the previous paper was that the binary blends tested, being either gasoline-ethanol or gasoline-methanol mixtures, were configured with the same stoichiometry. The paper showed that such fuels perform essentially identically in any given combustion system, at least in terms of engine torque, power, and brake specific fuel consumption (BSFC) at the same spark advance, and they could also reduce emissions of oxides of nitrogen dramatically; however, particulate matter reduction (tested in a direct-injection engine) was found to be more blend dependent, with that finding being presented for the first time there. This new work conducts similar tests for so-called ternary blends, configured from gasoline, ethanol, and methanol such that they have the same stoichiometry as the binary blends discussed in the previous work. Blends made from these components are termed ‘GEM’ blends here. The present work utilizes fundamentally the same test engines and equipment as the previous investigation. While the ternary blend results are presented for the first time here, some reference will be made to the previous tests where pertinent, since many of the binary gasoline-alcohol blends reported there have the same stoichiometry as the GEM fuels investigated here; in fact in the space of possible blends at any given stoichiometry, the binary ones actually represent limit cases. This will be returned to later. This paper is constructed in the following manner, similar to how the authors dealt with the subject of binary blending of alcohols in gasoline [9]: Section2 discusses the rationale for and a method of introducing methanol into the bulk transport fuel pool, namely iso-stoichiometric blend formulation. This helps to set the context for work presented in the rest of this paper. Section3 discusses the testing of iso-stoichiometric GEM blends in the “Ultraboost” extreme downsizing 4-cylinder demonstrator engine at the University of Bath. The very high brake mean effective pressure (BMEP) that this engine is capable of operating at means that the data shown here represents testing of the hypothesis of identical performance from any blend configured according to the rules to a level not published before. Furthermore, the interaction with exhaust gas recirculation (EGR), already known to have excellent synergies with alcohol fuels, is shown. Some particulate emissions results are reported here, pertinent to the high-load operating conditions reported. Section4 discusses tests conducted at Ghent University using three di fferent engines: two port-fuel-injection naturally-aspirated engines—one a production 4-cylinder and the other a single-cylinder research engine configured with a high compression ratio (CR)—and one multi-cylinder Energies 2020, 13, 6390 5 of 31 direct-injection engine. The blends tested correspond to the stoichiometry of E85 and in terms of knock all of the blends perform identically and fundamentally the same as E100. Section5 presents data gathered at the University of Oxford in which the particulate emissions performance of GEM blends is investigated in a single-cylinder test engine at light load. These findings have not been presented before, and how they reinforce the findings from the binary blend tests conducted in the first part of this investigation is discussed. Finally, this paper ends with a Conclusions section.

2. A Strategy for Introducing Methanol into Transport Fuel by Configuring Ternary Blends as Drop-In Fuels The two “light alcohols”, i.e., those with no isomers, ethanol and methanol, have for years been blended with gasoline to create fuels for SI engines, with E85 and M85 being examples of binary blends which have been commercialized for several years. For these fuels, the number represents the volume percentage of the alcohol (abbreviated to E and M) in admixture with gasoline. In fact all of the alcohols are miscible with gasoline to a greater or lesser extent and this gives rise to the potential to create multi-component blends to a target stoichiometry, provided this value is within an upper and lower target range. The ternary—or three component—blends used to in this paper are all mixtures of gasoline and the light alcohols ethanol and methanol. Collectively, they are referred to as ‘GEM’ blends; and specifically “GXX EYY MZZ”, where XX, YY and ZZ are the volume percentages of the three individual components, respectively. The process by which such blends can be formulated and how they can offer a route to increased penetration of renewable energy in the fuel pool has been discussed before [17–20,22,24]. Both engine and vehicle tests of such blends have been reported, the latter showing that when ternary blends are configured with the same stoichiometric air-fuel ratio (AFR) as E85 (this nominally being a blend of 85% ethanol and 15% gasoline by volume)—approximately 9.7:1—then they are undetectable by a flex-fuel vehicle engine management system when such a vehicle has already been calibrated to be capable of taking any proportion of gasoline and E85 put into its fuel tank [17,18]. This is essentially due to near-identical volumetric energy content (the volumetric lower heating value (LHV)) in the blends as well their producing a near-identical response from a flex-fuel vehicle’s ethanol concentration sensor. Furthermore, in [18], it was shown that increasing the methanol content actually improved the cold startability of the vehicle, providing a further mechanism to keep increased renewable energy content in the fuel pool all year round. All of these previous reports show that if followed, these ternary blending rules permit a method of simply creating fuel blends equivalent to any original binary gasoline-ethanol mixture. Recently, Pearson and co-workers [24] have defined more completely the blending rules for such an approach, and have shown the effect of water on AFR and volumetric LHV if this is left in ethanol (which could be very beneficial from a production energetics point of view). (Note that unlike all the other alcohols, methanol does not form an azeotrope with water, and consequently simple distillation is sufficient to dry it fully. By contrast, ethanol cannot be dried by simple distillation beyond 4.4% v/v water. In the case of the distillation of the water formed in a production process utilizing CO2 to synthesize methanol, the process can utilize the waste heat generated within the exothermic synthesis process itself, supporting the contention that once the feed stocks have been obtained using sustainable processes, all downstream processes and products can be maintained as fully sustainable too.) Pearson and co-workers show that blending for constant oxygen content is just as viable as blending for constant stoichiometry or volumetric energy content, and furthermore that the magnitude of the differences between the approaches are effectively negligible [24]. Importantly, when such ternary blends are configured in this manner, all such possible blends have near-identical volumetric LHV, practically the same octane numbers, and extremely close enthalpies of vaporization (to +/ 2%, the variation arising − from treating the components on a mass basis) [17,18], all properties which need to be the same or similar for any fuel to capable of being ‘dropped in’ to an existing vehicle fuel and combustion system. Energies 2020, 13, 6390 6 of 31

It has also been shown that for any given amount of ethanol that can be introduced into the Energiestotal fuel 2020 pool,, 13, x theFOR ease PEER of REVIEW synthesis and hence supply of methanol means that introducing it to make6 of 32 GEM blends suitable for the vehicles already in the market effectively displaces gasoline [17]. In [17], authorsthe authors also also show show that that within within certain certain wide wide price price limits, limits, the the resulting resulting blends blends can can be be cheaper cheaper than gasoline on a per-unit-energy basis; furthermore in [25], [25], it was shown that a taxation system based on taxing units of energy and applying a multiplier based on the fossil CO2 impact of supplying the on taxing units of energy and applying a multiplier based on the fossil CO2 impact of supplying energythe energy in that in that form form could could provide provide a government a government with with the themeans means to encourage to encourage renewable renewable energy energy at theat the expense expense of offossil fossil fuels fuels while while keeping keeping the the over overallall taxation taxation revenue revenue gathered gathered constant. constant. Thus, Thus, a frameworka framework could could be becreated created to encourage to encourage the theconsumer consumer to move to move away away from from fossil fossil fuels fuelsbased based on their on beingtheir being more moreexpensive expensive to use to than use thanrenewable renewable fuels, fuels, while while maintaining maintaining economic economic viability viability for all for of all the of stakeholders.the stakeholders. For thethe case case of stoichiometryof stoichiometry equivalent equivalent to E85, to the E85, volume the relationshipvolume relationship between the between components the componentsis shown in Figureis shown3. This in Figure was determined 3. This was in previously-publisheddetermined in previously-published work using a proprietary work using fuel a proprietarycharacteristics fuel tool characteristics [18,25], and tool here [18,25], one can and see here that one as the can proportion see that as ofthe ethanol proportion is reduced, of ethanol so the is reduced,rate of increase so the rate in the of proportionincrease in ofthe methanol proportion is fasterof methanol than that is faster of the than gasoline. that of This the isgasoline. because This one isvolume because unit one of volume ethanol unit is equivalent of ethanol to is a equivalent binary gasoline-methanol to a binary gasoline-methanol volume unit with volume the proportions unit with theof 32.7:67.3 proportions (this of then 32.7:67.3 sharing (this the then 9:1 stoichiometric sharing the 9:1 AFR stoichiometric of ethanol). AFR of ethanol).

100

90

80

70

60

50

40 Blend / [% v/v] / [% Blend

30

20

10 Proportion of Gasoline or Methanol in Ternary MethanolorinGasoline Ternary Proportionof

0 0 102030405060708090100 Proportion of Ethanol in Ternary Blend / [% v/v]

Gasoline Methanol Figure 3. RepresentationRepresentation of of how how the the volume volume proportion proportionss of of three blend constituents change for a targeted stoichiometric AFR AFR equivalent equivalent to to that that of of E8 E855 (i.e., 9.7). Here Here these these constituents constituents are are gasoline, gasoline, ethanol, and methanol, and together in this form they are termed iso-stoichiometric ‘GEM’ ternary blends. When When reading reading this this figure, figure, note note that that E85 E85 is is effectively effectively on on the right; where the lines end is the limitlimit of being able to meet the stoichiometricstoichiometric AFR limit of 9.7.9.7. The The equivalent equivalent where where no no ethanol is present (i.e., M56) is on the extreme left.

Four such GEM blends were tested by Lotus En Engineeringgineering in a Saab 9-3 BioPower production flex-fuelflex-fuel vehicle employingemploying aa virtual virtual sensor sensor to to infer infer the the alcohol alcohol concentrations concentrations that that it wasit was operating operating on. on.These These initial initial tests tests showed showed that that all all of of the the blends blends were were invisible invisible to to the the vehicle’svehicle’s on-boardon-board diagnostics system which was unmodified unmodified [17], [17], provided provided that that a a certain certain minimum minimum level level of of cosolvent cosolvent was was present; present; at very low temperatures M56 (the binary gasoline-methanolgasoline-methanol mixture with the same stoichiometric AFR as E85) waswas foundfound toto illuminateilluminate the the malfunction malfunction indicator indicator lamp. lamp. This This was was presumed presumed to to be be due due to tophase phase separation separation because because of theof the absence absence of anyof any form form of cosolventof cosolvent in thisin this particular particular fuel fuel blend. blend. In GEM blends, the function of a cosolvent in mixtures of gasoline and methanol is undertaken by the ethanol provided some minimum volume is used. This minimum volume was investigated in later tests in a vehicle certified to a higher emissions standard utilizing a physical sensor for alcohol concentration. However, in this later series of tests, despite repeated cold soaks and related cold start tests at −20 °C, [18] a minimum requirement for ethanol was not found. Even M56 (i.e., the binary

Energies 2020, 13, 6390 7 of 31

In GEM blends, the function of a cosolvent in mixtures of gasoline and methanol is undertaken by the ethanol provided some minimum volume is used. This minimum volume was investigated in later tests in a vehicle certified to a higher emissions standard utilizing a physical sensor for alcohol concentration. However, in this later series of tests, despite repeated cold soaks and related cold start tests at 20 C, [18] a minimum requirement for ethanol was not found. Even M56 (i.e., the binary − ◦ gasoline-methanol mixture) started satisfactorily, suggesting that such GEM blends could help to displace more gasoline from the fuel pool year round (for a fixed volume of ethanol that is available). From the vehicle testing program conducted by Lotus, it was deduced that, since all of the vehicles in the parc do not all use the same alcohol concentration-sensing technology, a minimum ethanol concentration would in fact be needed to ensure the satisfactory operation of the entire fleet, and that this minimum level would be in the region of 10% by volume [18]. In turn, this implies that, without the addition of any other type of cosolvent, the E85-equivalent ternary blend with the maximum methanol concentration that could be used is G40 E10 M50. A further important finding from this earlier work was that regardless of the technology used to sense the alcohol concentration in the fuel system near to the engine, energetic efficiency improvements on the order of 5% were possible. This was true for any of the ternary blends tested at the E85-equivalent stoichiometric AFR of 9.7:1 [17,18]. This is important when considering the amount of renewable energy that would have to be supplied to synthesize such fuels by the electrofuel route as per the approach discussed by various authors [26–33]. Later sections in this paper will provide further validation of the efficiency improvements possible. Issues of material compatibility were investigated in another publication [20], together with results showing that exhaust emissions do not change significantly as the blends are altered, with all E85-equivalent ternary blends essentially out-performing gasoline in this respect. However, it should be noted that particulate and aldehyde emissions were not measured in these tests. Further, Pitcher et al. [21] investigated the spray morphology of GEM blends in an optical direct-injection engine, this being important because of its bearing on emissions and fuel consumption in engines using such fuel introduction technologies. They showed that there was no significant difference in morphology across the full spectrum of E85-equivalent blends. In the present work, Section5 builds on this work and presents new data from GEM blends with regards to their soot-formation tendency, which is especially important with respect to Euro 6c (and higher) emissions legislation. These tests were conducted in a different optical engine, and represent the first time such particulate matter data from such blends has been presented by this group. Considerable additional work on the GEM concept has been reported by Sileghem and co-workers in both port-fuel-injection (PFI) and direct-injection (DI) engines [34,35]. All of this work has shown that in terms of engine and combustion system performance any GEM blend configured according to the rules established by Pearson and Turner will perform in the same manner using the same engine control settings. Note that it is possible to produce ternary blends of other alcohols with gasoline, or even quaternary (or higher number) blends. Similarly, water can be included in such blends; the physicochemical characteristics of such these blends was investigated [24], although, practically speaking, phase separation would become an issue with the presence of water, the light alcohols being especially prone to this when in admixture with gasoline. Other authors have discussed physicochemical, performance, and blending aspects of binary gasoline-alcohol blend fuels in detail [36–39], but while it is accepted that challenges exist in blending alcohols with gasoline (such as high vapor pressure and evaporative emissions effects aboard the vehicle, or phase separation and blend stability in the distribution system, with materials compatibility issues in both), it is the authors’ belief that these are understood and by no means insurmountable, as evidenced by the amount of high-blend-rate binary mixtures of either ethanol or methanol that is or has been used around the world and the number of fully-homologated vehicles that have been sold that are capable of taking them. From all of this initial work, the inherent compatibility of existing production flex-fuel technology with ternary blends with stoichiometric AFRs matched to fuels already in the market place has been Energies 2020, 13, 6390 8 of 31 established. It is therefore believed that a progressive rollout of such fuels, by methanol percentage and region, could allow staged introduction of methanol as a transport fuel. This has the potential to meaningfully reduce greenhouse gas emissions, improve energy security, and provide better air quality. Once the market has accepted the new fuel, it enables widespread use of an energy carrier which could eventually be completely decarbonized using the electrofuel approach [26–33], with methanol then becoming a chemically-liquefied form of hydrogen as an extension of the hydrogen economy [40]. At this point in time, a practical liquid fuel that has been fully decarbonized will have been realized.

3. Ternary Gasoline-Alcohol Blend Tests in the “Ultraboost” Extreme Downsizing Demonstrator Engine The “Ultraboost” collaborative project, more properly known as “Ultra Boost for Economy”, was part-funded by the UK Technology Strategy Board (now Innovate UK) and started in 2010. It sought to demonstrate the potential to provide a vehicle-level CO2 reduction of 35% in an SUV by employing a heavily downsized (and ‘down-cylindered’) engine. One major target was that this new engine should be capable of generating the torque curve of the Jaguar Land Rover AJ133 V8 engine. That engine is of 5.0 L capacity and is naturally aspirated (NA). This performance had to be realized while operating on 95 research octane number (RON) standard pump gasoline [41] (in fact the project did indeed show that this CO2 reduction target was achievable and in fact greater than this was demonstrated by vehicle simulation using the dynamometer data gathered). At the start of Ultraboost, there was no pre-ordained engine swept volume, but the final version of the engine was of in-line 4-cylinder configuration, with a compound-charging system and 2.0 L swept volume [41]. However, the first iteration of the design (termed ‘UB100’) was externally boosted by a charge air handling unit (CAHU) system at the University of Bath facility, where all of the engine test work, including taking the results reported here, was carried out. Specifications for this engine are given in Table1. For detailed information on the other project partners and their responsibilities, the individual stages of the project, and the test facility, the reader is referred to [41].

Table 1. Specifications of the Ultraboost extreme downsizing demonstrator engine in the externally-boosted “UB100” configuration, as used for the fuel blends testing reported here [41].

General architecture 4-cylinder in-line with 4 valves per cylinder and double overhead camshafts All aluminum Jaguar Land Rover AJ133 cylinder block converted to single-bank operation on the Construction right-hand side Siamesed liner pack to facilitate reduced bore diameter Dedicated cylinder head Firing order 1–3–4–2 Bore (mm) 83 Stroke (mm) 92 Swept volume (cc) 1991 Pent-roof combustion chamber with asymmetric central direct injection and spark plug High-tumble intake ports Combustion system Auxiliary port-fuel injection Possible second spark plug position in an under-intake-port location Compression ratio (:1) 9.0 Chain-driven double overhead camshaft (DOHC) with fast-acting dual Valve gear continuously-variable camshaft phasers Cam profile switching tappets on inlet and exhaust Engine management system Lotus ‘T6’ engine control unit

As well as for its prime aim as being a platform for the development of a self-contained technology demonstrator package, the engine was used for an extensive series of fuel tests within the program. For these fuel investigations accurate control of the engine’s plenum temperature and pressure was Energies 2020, 13, 6390 9 of 31 required and so the CAHU system was used throughout. Cooled EGR was also utilized in some of these tests, this being provided by a dedicated rig especially developed for the process. This rig took exhaust gas from the engine and pressurized it for delivery to the intake of the engine, where it was mixed with the air supply provided separately by the CAHU. This EGR supply system is discussed further in [41,42]. The temperature of the resulting combined charge was also controlled by the test facility, and it is this test equipment configuration that is used in the results presented below. The exhaust back pressure was controlled by the test cell also. The results of most of the other conventional hydrocarbon fuels tests have been presented in other publications [9,43–47]. From its original work on the ternary blends concept (see Section2) and as a result of being a partner in the project, Lotus Engineering requested that E85 and an iso-stoichiometric ternary alcohol blend be tested within the Ultraboost project, and these fuels were then formulated by Shell; the specifications of these are shown in Table2. In this section the ternary blend is henceforth termed “GEM”. In the present work, the results from these iso-stoichiometric blends will be compared to two commercial gasolines (95 and 98 RON), and with the project’s control fuel. This control or ‘base fuel’, was a standard pump gasoline sold as 95 RON fuel, but with some octane giveaway (from 97.0 RON, see Table2) as a result of it having to achieve a motor octane number (MON) of 85, which is necessary for the European market. All of these fuels are also shown in Table2. The 98 RON gasoline contained no ethanol, but both of the 95 RON fuels had 5% ethanol and all were production fuels of EN228 standard; as a consequence none of these are considered to be a gasoline-alcohol blend fuel in the discussion that follows.

Table 2. Specifications of the gasoline, E85, and ternary gasoline-alcohol “GEM” fuels tested in the Ultraboost extreme downsizing demonstrator engine. All data was supplied by Shell and measured using standard industry test procedures and equipment.

Alcohol Proportion Volumetric Stoichiometric Fuel Name RON MON S and Type LHV (MJ/L) Air-Fuel Ratio (:1) Base 97.0 85.3 11.7 5% Ethanol 31.29 14.18 95 RON Pump 95.1 85 10.1 5% Ethanol 31.51 14.19 98 RON Pump 98.7 86.5 12.2 No alcohol 31.57 14.45 E85 107.4 89.5 17.9 85% Ethanol 22.65 9.68 23% Ethanol GEM 106.0 88.1 17.9 22.71 9.78 43% Methanol

Table2 shows an interesting slight deviation from previous GEM fuel research insofar as the E85 and GEM did not have RON and MON values with closer alignment. Further, they did not have the same stoichiometric AFR. Here, they were made from different gasoline blend stocks, which was also a deviation from previous practice. Thus, before testing began, from their previous experience a slight difference in performance between these two fuels was expected by the authors; in the previously-reported work by Lotus Engineering the GEM blends were splash blended using the same gasoline each time and the achieved stoichiometric AFR range was tighter due to the use of Lotus Engineering’s proprietary fuels properties calculator (here the stoichiometric AFR range was just over +/ 0.5% for the two fuels; in the previous work by Lotus Engineering, +0.2/ 0.4% was achieved across − − six blends [17,19,20]). To illustrate this, Figures4 and5 show test data from ignition timing swings for the base fuel, E85, and GEM at the same operating conditions. These operating conditions and the rationale for adopting them have been discussed by Remmert et al. [43], together with other information on the testing conditions. Note that throughout this test campaign, daily checks on performance were maintained and no issues were raised on drift or repeatability by the researchers operating the engine. In Figure4, the engine operating condition is 2000 rpm speed, with intake plenum conditions of 40 ◦C and 130 kPa gauge temperature and pressure, respectively. This intake condition is referred to as test point A later in this paper. For the results in Figure5, the equivalent conditions are 3000 rpm, 60 ◦C, and 140 kPa gauge, Energies 2020, 13, 6390 10 of 31 respectively, and it included 10% cooled EGR; it is termed test point D. It is clear from these figures that both of these operating points are high-load conditions, and that both of the high-alcohol-content fuels yield significantly higher BMEP than the base gasoline for the same conditions; furthermore it should be noted that for all operating points discussed here it was these intake and exhaust boundary Energies 2020, 13, x FOR PEER REVIEW 10 of 32 conditionsEnergies 2020 that, 13, werex FOR maintained,PEER REVIEW fixing the air flow rate through the engine at each point, and10 notof 32 the fuelengine energy was flow operated rate through at a relative the engineAFR (or (despite λ) of 1 throughout). the fact that Test the points engine Bwas and operatedC will be discussed at a relative engine was operated at a relative AFR (or λ) of 1 throughout). Test points B and C will be discussed AFRlater. (or λ) of 1 throughout). Test points B and C will be discussed later. later. 32.0 32.0 31.5 31.5 31.0 31.0 30.5 30.5 30.0 30.0 29.5 29.5 29.0 29.0 28.5

BMEP / [bar] 28.5 BMEP / [bar]28.0 28.0 27.5 27.5 27.0 27.0 26.5 26.5 26.0 26.0 -8-7-6-5-4-3-2-1012345 -8-7-6-5-4-3-2-1012345 Spark Advance / [deg BTDC] Spark Advance / [deg BTDC] Base Fuel E85 GEM Base Fuel E85 GEM

Figure 4. Ignition timing swings at test point A for E85 and the GEM blend compared with the base FigureFigure 4. 4.Ignition Ignition timing timing swings swings atat testtest pointpoint A for E85 and and the the GEM GEM blend blend compared compared with with the the base base 95 RON gasoline fuel. These results are all at a matched engine speed of 2000 rpm and with 40 °C 9595 RON RON gasoline gasoline fuel. fuel. TheseThese resultsresults are all at a ma matchedtched engine engine speed speed of of 2000 2000 rpm rpm and and with with 40 40°C ◦C temperature and 130 kPa gauge pressure in the plenum but without any cooled EGR. Note that the temperaturetemperature and and 130 130 kPa kPa gauge gauge pressurepressure inin thethe plenumplenum but but without without any any cooled cooled EGR. EGR. Note Note that that the the engine load (i.e., BMEP) varied with spark advance. See text for discussion on accuracy of the results. engineengine load load (i.e., (i.e., BMEP) BMEP) varied varied with with sparkspark advance.advance. See See text text for for discussion discussion on on accuracy accuracy of ofthe the results. results.

26.5 26.5

26.0 26.0

25.5 25.5

25.0 25.0

24.5 24.5

24.0 24.0 BMEP / [bar]

BMEP / [bar]23.5 23.5

23.0 23.0

22.5 22.5

22.0 22.0 4 5 6 7 8 9 1011121314151617181920 4 5 6 7 8 9 1011121314151617181920 Spark Advance / [deg BTDC] Spark Advance / [deg BTDC] Base Fuel E85 GEM Base Fuel E85 GEM

FigureFigure 5. 5.Ignition Ignition timing timing swings swings atat testtest pointpoint D for for E85 E85 and and the the GEM GEM blend blend compared compared with with the the base base Figure 5. Ignition timing swings at test point D for E85 and the GEM blend compared with the base 9595 RON RON gasoline gasoline fuel. fuel. TheseThese resultsresults are all at a ma matchedtched engine engine speed speed of of 3000 3000 rpm rpm and and with with 60 60°C ◦C 95 RON gasoline fuel. These results are all at a matched engine speed of 3000 rpm and with 60 °C temperaturetemperature and and 140 140 kPa kPa gauge gauge pressurepressure inin the plenum with with 10% 10% cooled cooled EGR. EGR. Note Note that that the the engine engine temperature and 140 kPa gauge pressure in the plenum with 10% cooled EGR. Note that the engine loadload (i.e., (i.e., BMEP) BMEP) varied varied with with spark spark advance.advance. See text text for for discussion discussion on on accuracy accuracy of of the the results. results. load (i.e., BMEP) varied with spark advance. See text for discussion on accuracy of the results. ItIt is is apparent apparent that that here here thethe twotwo high-proportionhigh-proportion alcohol alcohol blend blend fuels fuels did did not not perform perform quite quite It is apparent that here the two high-proportion alcohol blend fuels did not perform quite identically.identically. They They bothboth reached reached nearly nearly the the same same maximum maximum BMEP, BMEP, with with E85 being E85 being slightly slightly higher, higher, but identically. They both reached nearly the same maximum BMEP, with E85 being slightly higher, but interestingly the GEM blend generates approximately 0.6 bar more BMEP for any given amount of interestingly the GEM blend generates approximately 0.6 bar more BMEP for any given amount of spark advance up to its knock-limited spark advance (KLSA) limit (this being the point at which the spark advance up to its knock-limited spark advance (KLSA) limit (this being the point at which the curves end, i.e., at these high BMEP conditions minimum park advance for best torque (MBT) could curves end, i.e., at these high BMEP conditions minimum park advance for best torque (MBT) could not be reached before this point, in spite of the high knock resistance of the fuels). It is presumed that not be reached before this point, in spite of the high knock resistance of the fuels). It is presumed that the reason that E85 can generate slightly more BMEP is the fact that it can travel sufficiently further the reason that E85 can generate slightly more BMEP is the fact that it can travel sufficiently further

Energies 2020, 13, 6390 11 of 31 but interestingly the GEM blend generates approximately 0.6 bar more BMEP for any given amount of spark advance up to its knock-limited spark advance (KLSA) limit (this being the point at which the curves end, i.e., at these high BMEP conditions minimum park advance for best torque (MBT) could not be reached before this point, in spite of the high knock resistance of the fuels). It is presumed that Energies 2020, 13, x FOR PEER REVIEW 11 of 32 theEnergies reason 2020 that, 13, E85 x FOR can PEER generate REVIEW slightly more BMEP is the fact that it can travel sufficiently further11 of 32 up itsup spark its spark curve curve before before it reaches it reaches its its KLSA; KLSA; this this is is likely likely a a result result ofof its higher higher RON RON and and MON MON values values up its spark curve before it reaches its KLSA; this is likely a result of its higher RON and MON values asas shown shown in in Table Table2. 2. The The reason reason that that the the GEM GEM blendblend generatesgenerates more more BMEP BMEP at at lower lower values values of ofspark spark as shown in Table 2. The reason that the GEM blend generates more BMEP at lower values of spark advanceadvance is is possibly possibly due due to to a a combination combination ofof factors, including including the the fact fact that that the the exhaust exhaust back back pressure pressure advance is possibly due to a combination of factors, including the fact that the exhaust back pressure (EBP)(EBP) was was slightly slightly lower lower forfor thethe GEM blend blend than than the the E85. E85. This This is shown is shown in Figure in Figure 6 for6 testfor point test point A, (EBP) was slightly lower for the GEM blend than the E85. This is shown in Figure 6 for test point A, A,where where it it can can also also be be seen seen that that the the gasoline gasoline case case actually actually has has the thelowest lowest EBP EBP of all. of For all. test For point test point D where it can also be seen that the gasoline case actually has the lowest EBP of all. For test point D Figure 7 shows the same thing, and the shape of the EBP curve for the GEM fuel here can be seen D Figure 77 showsshows thethe same thing, and and the the shape shape of of the the EBP EBP curve curve for for the the GEM GEM fuel fuel here here can can be be seen seen reflected in the BMEP curve in Figure 5, lending credence to this hypothesis; it should be noted that reflectedreflected in in the the BMEP BMEP curve curve in in Figure Figure5 ,5, lending lending credencecredence toto this hypothesis; it it should should be be noted noted that that the absolute difference between the EBP values for test point D is much smaller than it was for test thethe absolute absolute di differencefference between betweenthe the EBPEBP valuesvalues for te testst point point D D is is much much smaller smaller than than it itwas was for for test test point A. The difference between the alcohol fuel cases was unfortunately not discovered until after pointpoint A. A. The The di ffdifferenceerence between between the the alcohol alcohol fuel fuel cases cases was was unfortunately unfortunately not not discovered discovered until until after after the the tests had been completed. teststhe had tests been had completed.been completed.

45.5 45.5 45.0 45.0 44.5 44.5 44.0 44.0 43.5 43.5 43.0 43.0 42.5 42.5 42.0 42.0

Exhaust Back Pressure / [kPa Gauge] [kPa / Pressure Back Exhaust 41.5 Exhaust Back Pressure / [kPa Gauge] [kPa / Pressure Back Exhaust 41.5 41.0 41.0 -8 -7 -6 -5 -4 -3 -2 -1 0 1 2 3 4 5 -8 -7 -6 -5 -4Spark -3 Advance -2 / -1[deg BTDC] 0 1 2 3 4 5 Spark Advance / [deg BTDC] Base Fuel E85 GEM Base Fuel E85 GEM

Figure 6. Exhaust back pressure for the data presented in Figure 4, taken at test point A. See text for FigureFigure 6. 6.Exhaust Exhaust back back pressure pressure forfor thethe datadata presented in in Figure Figure 4,4, taken taken at at test test point point A. A. See See text text for for discussion on accuracy of the results. discussiondiscussion on on accuracy accuracy of of the the results. results.

119.5 119.5

119.0 119.0

118.5 118.5

118.0 118.0

117.5 117.5

117.0 117.0

116.5 116.5 Exhaust Back Pressure / [kPa Gauge] Exhaust Back Pressure / [kPa Gauge] 116.0 116.0 4567891011121314151617181920 4567891011121314151617181920Spark Advance / [deg BTDC] Spark Advance / [deg BTDC] Base Fuel E85 GEM Base Fuel E85 GEM

FigureFigure 7. 7.Exhaust Exhaust back back pressure pressure forfor thethe datadata presented in in Figure Figure 6,6, taken taken at at test test point point D. D. See See text text for for Figure 7. Exhaust back pressure for the data presented in Figure 6, taken at test point D. See text for discussiondiscussion on on accuracy accuracy of of the the results. results. discussion on accuracy of the results. The difference in plenum pressure and temperature across all of the test results shown in Figures The difference in plenum pressure and temperature across all of the test results shown in Figures 4–7 was less than 0.3% and 1 °C, respectively, indicating the excellent accuracy with which the CAHU 4–7 was less than 0.3% and 1 °C, respectively, indicating the excellent accuracy with which the CAHU could hold the intake conditions constant. Note that the valve timing was held constant for each test could hold the intake conditions constant. Note that the valve timing was held constant for each test condition and as a consequence the difference in back pressure is seen as the main reason for the condition and as a consequence the difference in back pressure is seen as the main reason for the

Energies 2020, 13, 6390 12 of 31

The difference in plenum pressure and temperature across all of the test results shown in Figures4–7 was less than 0.3% and 1 ◦C, respectively, indicating the excellent accuracy with which the CAHU could hold the intake conditions constant. Note that the valve timing was held constant for each test condition and as a consequence the difference in back pressure is seen as the main reason for the difference in performance between E85 and the GEM blend, because it would lead to a difference in volumetric efficiency. It is not believed that the GEM blend would in itself give better performance, although another difference between the two high-blend alcohol fuels is that different gasoline blend stock was used for each. This is thought to be one of the reasons why the octane numbers of these two fuels vary by 1.4 units, too (as shown in Table2), as well as the di fference in stoichiometric AFR. Whether it is calculated on a mass or volume basis, the E85 and ternary blend will have essentially the same enthalpy of vaporization, as was discussed separately by Chupka et al. for ethanol blends [48] and by Turner et al. for GEM blends [17]. Regardless of the above discussion, it is apparent that the two high-blend alcohol fuels allow the engine to generate significantly more load for the same plenum conditions than when it operates on the baseline fuel. A total of four different test points (shown in Table3 and including A and D mentioned above) were used to compare fuels during this part of the Ultraboost program, and the rationale for selecting these has been discussed in previous publications [24,43–45]. Salient results for these conditions are also presented in Figures8–10.

Table 3. Test operating points for Ultraboost engine testing.

Engine Boost Plenum Test Load EGR Exhaust Back Speed Pressure Temperature Notes Point Condition (%) Pressure (Bar) (rpm) (Bar Gauge) (◦C) A 2000 High 0 1.3 40 0.45 Boost pressure increased to B 2000 High 10 1.5 40 0.45 compensate for presence of EGR Mimics supercharged C 3000 Mid-High 10 1.4 60 0.7 operation Mimics turbocharged D 3000 Mid-High 10 1.4 60 1.2 operation

Test point A was intended to represent a highly downsized engine operating near to the torque curve “knee point” but with no cooled EGR being used for knock mitigation. Test point B was effectively point A with the intake boosted to account for 10% external EGR being used by keeping the amount of oxygen flowing through the engine constant; the exhaust back pressure was, however, kept equal between these points. Points C and D are higher speed conditions but with markedly different back pressure. A supercharged engine condition was mimicked by point C, while point D intended to be representative of a turbocharged one with higher exhaust port back pressure; both of these use 10% EGR. Since the plenum conditions were the same for points C and D comparison between these was intended to see what the effect of increased hot residuals retention had on the knock limit. It is accepted, however, that with these two test points there is a conflating effect of reduced volumetric efficiency, even though they shared the same, unvarying, valve timing. Within the discussion above regarding the slight difference in back pressure for the GEM fuel test, the results shown in Figures8–10 essentially show the expected response. The high-alcohol blends generate significantly greater torque at all test points than the two gasolines. At the same time, the BSFC for the alcohol blends is significantly higher than the two gasolines. However, the brake thermal efficiency for the alcohol blends is also significantly better: generally 1–2% better in absolute terms, and approximately 6% better at test point B. Part of the reason for this will be due to the higher molar expansion ratios of the alcohols versus gasoline, and thus their blends. The importance of this characteristic has been discussed by Szybist et al. [49] and Nguyen et al. [50]. Energies 2020,, 13,, x 6390 FOR PEER REVIEW 1313 of of 32 31 Energies 2020, 13, x FOR PEER REVIEW 13 of 32 540 540 520 520 500 500 480 480 460 460 440 440 420 Torque / [Nm] Torque 420 Torque / [Nm] Torque 400 400 380 380 360 360 340 340 ABCD ABCDBase Fuel 95 RON E5 Pump E85 GEM Base Fuel 95 RON E5 Pump E85 GEM

Figure 8. Maximum torque for the base and 95 RON pump gasolines, E85 and the GEM blend, at four Figure 8. Maximum torque for the base and 95 RON pump gasolines, E85 and the GEM blend, at Figuredifferent 8. Maximumfixed-intake torque conditions, for the baseoperated and 95 at RON KLSA pump ignition gasolines, timing. E85 Forand descriptionthe GEM blend, of the at fourtest four different fixed-intake conditions, operated at KLSA ignition timing. For description of the test differentoperating fixed-intake points, labelled conditions, A to D here,operated see Tableat KL 3SA and ignition text. Note timing. that theFor twodescription high-alcohol of the blends test operating points, labelled A to D here, see Table3 and text. Note that the two high-alcohol blends have operatinghave nearly-matched points, labelled stoichiometry A to D here, and see enthalTable py3 and of text.vaporization, Note that RONthe two and high-alcohol MON, and blends as a nearly-matched stoichiometry and enthalpy of vaporization, RON and MON, and as a consequence have nearly-matched stoichiometry and enthalpy of vaporization, RON and MON, and as a consequenceperform in an perform essentially in an identical essentially manner. identical manner. consequence perform in an essentially identical manner. 400 400 380 380 360 360 340 340 320 320 300 300 280

BSFC / [g/kWhr] 280

BSFC / [g/kWhr] 260 260 240 240 220 220 200 200 ABCD ABCDBase Fuel 95 RON E5 Pump E85 GEM Base Fuel 95 RON E5 Pump E85 GEM

FigureFigure 9. 9. BSFCBSFC at at maximum maximum torque torque for for the the base base and and 95 95 RON RON pump pump gasolines, gasolines, E85 E85 and and the the GEM GEM blend, blend, Figureatat four four 9.different di BSFCfferent at fixed-intake fixed-intakemaximum torque conditions, for the operated base and at 95 KLSA RON ignition pump gasolines, timing. For For E85 description description and the GEM of of the theblend, test test atoperatingoperating four different points, fixed-intake labelledlabelled AA to conditions,to D D here, here, see seeoperated Table Table3 and at3 andKLSA text. text. Note ignition Note that timing.that the twothe For high-alcoholtwo description high-alcohol blends of the blends have test nearly-matched stoichiometry and enthalpy of vaporization, RON and MON, and as a consequence operatinghave nearly-matched points, labelled stoichiometry A to D here, and see enthalTable py3 and of text.vaporization, Note that RONthe two and high-alcohol MON, and blends as a perform in an essentially identical manner. haveconsequence nearly-matched perform instoichiometry an essentially and identical enthal manner.py of vaporization, RON and MON, and as a consequence perform in an essentially identical manner.

Energies 2020, 13, 6390 14 of 31 Energies 2020, 13, x FOR PEER REVIEW 14 of 32

42

40

38

36

34

32 Brake Thermal Efficiency / [%] / Efficiency Thermal Brake

30

28 ABCD Base Fuel 95 RON E5 Pump E85 GEM

Figure 10. Figure 10. BrakeBrake thermal thermal efficiency efficiency at at maximum maximum torque torque for the the base base and and 95 95 RON RON pump pump gasolines, gasolines, E85 E85 and the GEM blend, at four different fixed-intake conditions, operated at KLSA ignition timing. For and the GEM blend, at four different fixed-intake conditions, operated at KLSA ignition timing. For description of the test operating points, labelled A to D here, see Table3 and text. Note that the two description of the test operating points, labelled A to D here, see Table 3 and text. Note that the two high-alcohol blends have nearly-matched stoichiometry and enthalpy of vaporization, RON and MON, high-alcohol blends have nearly-matched stoichiometry and enthalpy of vaporization, RON and and as a consequence perform in an essentially identical manner. MON, and as a consequence perform in an essentially identical manner. The results for this section reinforce that the blending rules for ternary blends discussed The results for this section reinforce that the blending rules for ternary blends discussed in [17– in [17–20,22] are entirely borne out. As a counterpoint, it is interesting to note that the work of 20,22] are entirely borne out. As a counterpoint, it is interesting to note that the work of Waqas et al. Waqas et al. [23] suggested that the octane of the gasoline used for the blend can have an effect in [23] suggested that the octane of the gasoline used for the blend can have an effect in that a high RON that a high RON base fuel allows linearity in octane numbers, but that a low RON base fuel yields base fuel allows linearity in octane numbers, but that a low RON base fuel yields a non-linear a non-linear behaviors. However, in this section, the octane numbers of the blends were directly behaviors. However, in this section, the octane numbers of the blends were directly measured. measured. Nevertheless, Waqas et al.’s observations have obvious implications when splash blending Nevertheless, Waqas et al.’s observations have obvious implications when splash blending such such ternary blends. This will be mentioned in later sections. ternary blends. This will be mentioned in later sections. Some particulate emissions results are now reported here since they were taken as part of the Some particulate emissions results are now reported here since they were taken as part of the Ultraboost project fuels investigation; Section5 reports further results from a single-cylinder engine. Ultraboost project fuels investigation; Section 5 reports further results from a single-cylinder engine. Figure 11 shows the particulate emissions from the base fuel, E85 and GEM fuels at these test conditions. Figure 11 shows the particulate emissions from the base fuel, E85 and GEM fuels at these test These were measured using a Cambustion DMS500 sampling at 1 Hz, at steady state, over 30 s and conditions. These were measured using a Cambustion DMS500 sampling at 1 Hz, at steady state, over the mean and standard deviation (σ) of the data over that time were taken, with the methodology 30 s and the mean and standard deviation (σ) of the data over that time were taken, with the being equivalent to that discussed in Section5 and fully described in [ 47]. The results show that at methodology being equivalent to that discussed in Section 5 and fully described in [47]. The results the four different operating points, the E85 fuel has the lowest PN emissions and the GEM fuel has show that at the four different operating points, the E85 fuel has the lowest PN emissions and the higher particulate emissions compared to the base fuel. The high PN emissions from the GEM fuel GEM fuel has higher particulate emissions compared to the base fuel. The high PN emissions from are thought to be caused by the rapid evaporation of that fuel upon injection due to the presence of the GEM fuel are thought to be caused by the rapid evaporation of that fuel upon injection due to the the high-vapor-pressure methanol. This rapid evaporation then causes locally rich regions to form, presence of the high-vapor-pressure methanol. This rapid evaporation then causes locally rich which leads to higher particulate emissions. A fuller analysis of these particulate emissions is included regions to form, which leads to higher particulate emissions. A fuller analysis of these particulate in [46,47,51]. emissions is included in [46,47,51]. As a consequence of these tests, one can state that, as expected from previous work, the alcohols As a consequence of these tests, one can state that, as expected from previous work, the alcohols in general perform better than normal pump gasoline in terms of KLSA, maximum brake output, in general perform better than normal pump gasoline in terms of KLSA, maximum brake output, and and thermal efficiency, and with the expected disadvantage in terms BSFC. The results gathered here, thermal efficiency, and with the expected disadvantage in terms BSFC. The results gathered here, allowing for the difference in operating conditions (specifically, the higher back pressure evident when allowing for the difference in operating conditions (specifically, the higher back pressure evident the GEM blend was tested), show the potential for GEM blends to perform at least as well as the binary when the GEM blend was tested), show the potential for GEM blends to perform at least as well as gasoline-ethanol blends whose stoichiometric AFR they have been matched to, as has been shown in the binary gasoline-ethanol blends whose stoichiometric AFR they have been matched to, as has been the authors’ previous work. These results are wholly in line with the findings of the first report on shown in the authors’ previous work. These results are wholly in line with the findings of the first this alcohol fuel research using the Ultraboost research engine [9]. However, it must be noted that the report on this alcohol fuel research using the Ultraboost research engine [9]. However, it must be

Energies 2020, 13, 6390 15 of 31 Energies 2020, 13, x FOR PEER REVIEW 15 of 32 presencenoted of that methanol the presence in the GEM of methanol fuel leads in to the inferior GEM performancefuel leads to withinferior respect performance to particulate with number, respect to andparticulate so more work number, needs and to be so carried more work out in needs this area. to be carried out in this area.

Figure 11. PN emissions at maximum torque for standard gasoline and two high-alcohol blends at fourFigure different 11. fixed-intake PN emissions conditions. at maximum The GEMtorque blend for standard has the highest gasoline PN and emissions. two high-alcohol The error blends bars at correspondfour different to σ. fixed-intake conditions. The GEM blend has the highest PN emissions. The error bars ± correspond to ±σ. Nevertheless, it is interesting to reflect that through this approach methanol clearly has a better blendingNevertheless, effect with gasoline it is interesting than ethanol. to reflect The that justification through forthis this approach assertion methanol is as follows. clearly E85has anda better M56blending are matched effect for with the samegasoline stoichiometric than ethanol. AFR The and justification are known fromfor this earlier assertion work tois as have follows. essentially E85 and identicalM56 properties,are matched including for the latentsame heat,stoichiometric volumetric AFR LHV, and RON, are MON,known and from sensitivity earlier work [17–20 to,22 ].have Previousessentially researchers identical have properties, discussed including that rather latent than heat, use volumetric liquid volume LHV, proportionsRON, MON, toand compare sensitivity performance[17–20,22]. of Previous such fuels, researchers using the molar have ratiodiscussed would that provide rather a betterthan use basis, liquid since volume combustion proportions occurs to in thecompare gas phase performance [11,24]. The of molar such ratiosfuels, in using these the fuels molar have beenratio calculatedwould provide using aa fuelsbetter properties basis, since calculatorcombustion proprietary occurs to in the the Universitygas phase [11,24]. of Bath The [52] mola and usingr ratios the in logicthese propounded fuels have been in [ 24calculated], with the using samea fuels properties properties for gasoline calculator used propri throughoutetary to the University calculations, of i.e.,Bath which [52] and is not using the the case logic in the propounded actual fuelsin blended [24], with here. the Accepting same properties this limitation, for gasoline the resultsused throughout show that thethe ethanolcalculations, mole fractioni.e., which in E85is not is the 0.94,case whereas in thethe actual methanol fuels blended molar concentration here. Accepting in M56this limitation, is 0.83. Ceteris the results paribus, show since that it the can ethanol produce mole the samefraction effect in E85 as a is higher 0.94, whereas molar proportion the methanol of ethanol molar can, concentration the lower molar in M56 proportion is 0.83. Ceteris for methanol paribus, issince proofit can of its produce better blending the same e effectffect in as gasoline. a higher molar Using pr theoportion same calculator, of ethanol in can, the the GEM lower blend molar tested proportion here the ethanolfor methanol and methanol is proof of mole its better fractions blending are 0.24 effect and in 0.63, gasoline. respectively, Using the giving same a calculator, total alcohol in molethe GEM fractionblend of tested 0.87, i.e., here between the ethanol the two and binary methanol blends, mole which fractions is logically are 0.24 what and one 0.63, would respectively, expect. giving a totalThus, alcohol if one wantsmole fraction primarily of to 0.87, improve i.e., between performance the two of a binary gasoline-alcohol blends, which fuel, methanolis logically appears what one to bewould the better expect. blend component of the two alcohols. It is believed that this is borne out by the initial results ofThus, the Co-Optima if one wants project primarily in the United to improve States. performance This project, whichof a gasoline-alcohol is looking at biofuel fuel, options methanol to improveappears energy to be the security better forblend the component US and which of the is looking two alcohols. at medium It is believed blend rates that (below this is 30%),borne has out by foundthe that initial of theresults various of the di ffCo-Optimaerent chemical project families in the tried, United the States. alcohols This are project, the most which desirable is looking [53]. at Indeedbiofuel in [ 54options], in terms to improve of RON, energy for a given-percentagesecurity for the US blend-rate and which methanol is looking was at shownmedium to blend have arates better(below response 30%), than has found ethanol that when of the expressed various different up to ~20% chemical by volume. families From tried, thethe resultsalcohols discussed are the most heredesirable at higher [53]. blend Indeed rates, duein [54], to the in marked terms of reduction RON, for in thea given-percentage volume of methanol blend-rate required methanol to obtain was the sameshown engine to have e ffia betterciency response as ethanol than when ethanol in admixture when expressed with gasolineup to ~20% discussed by volume. above, From it the seems results self-evidentdiscussed to here the authorsat higher that blend in termsrates, ofdue performance to the marked methanol reduction is a in much the volume more favorable of methanol additive required whento comparedobtain thewith same ethanol engine for efficien transportcy as fuel.ethanol As mentionedwhen in admixture earlier, it with can also gasoline extend discussed the impact above, of it a givenseems fixed self-evident volume of to ethanol the authors (this being that limitedin terms by of biomass performance availability, methanol for example) is a much when more blended favorable additive when compared with ethanol for transport fuel. As mentioned earlier, it can also extend the impact of a given fixed volume of ethanol (this being limited by biomass availability, for example) Energies 2020, 13, 6390 16 of 31 together with gasoline, as has been discussed by Turner et al. [17]. Ternary blending of the three components can also make the distillation curve more acceptable than binary blending, overcoming one of the drawbacks listed by the Co-Optima project. As the Co-Optima project moves forward, methanol has been eliminated as a candidate blend agent, primarily on the grounds of blend stability at lower concentrations than discussed here. To the authors of this work methanol would seem to be a preferred choice when considering ternary blends and with engine performance as the sole criterion; unfortunately there has not been such consideration of ternary blends within Co-Optima. The effect of ternary blending of gasoline with the light alcohols on engine performance will be further investigated in the following sections of the present paper.

4. Performance and Emissions of Iso-Stoichiometric GEM Blends in Production Engines and Blend Knock-Limited Spark Advance Tests Ghent University has operated several engines using pure methanol in a research stream investigating the desirable characteristics of methanol combustion in optimized engines. The alcohols in general have the potential to increase power and efficiency of SI engines [55,56], and their interesting properties allow different load control strategies to be used as well [57]. These studies have shown that any potential improvement is dependent upon the method of engine operation, but with this biased towards the alcohol such benefits can be provided while still maintaining the functional advantage of flex-fuel capability (i.e. so that gasoline could still be used, albeit with lower performance or efficiency, when the alcohol fuel is unavailable). To investigate this concept further, three naturally-aspirated engines were operated on gasoline, methanol and several GEM blends. These were a 4-cylinder PFI production-type Volvo engine, a single-cylinder research engine based on an Audi PFI engine, and a 4-cylinder Hyundai production engine utilizing DI. Details of these engines are shown in Table4.

Table 4. Test engine specifications.

Engine Type Volvo 1.8 L Audi/NSU Hyundai Cylinders 4 in-line 1 4 in-line Valves 16 2 16 Double overhead Overhead camshaft Double overhead Valvetrain camshaft (DOHC) (OHC) camshaft (DOHC) Bore (mm) 83 77.5 88 Stroke (mm) 82.4 86.4 97 Displacement (cc) 1783 407.3 2360 Compression ratio 10.3 13.1 11.3 (:1) Fuel system PFI PFI DI Induction Atmospheric Atmospheric Atmospheric Engine management MoTeC M800 MoTeC M4 Hyundai system

All the measurements presented here were taken at steady-state operating conditions at various speeds and loads. Stoichiometric operation was adopted in order to enable the use of a three-way catalyst (TWC). For the Hyundai engine, fitted with DI, tests were conducted with early injection. This resulted in a homogeneous mixture using a closed-loop feedback to control the air/fuel ratio to stoichiometric conditions. MBT spark timing was applied, except for knock-limited operating conditions, where KLSA-2◦ crank angle was used. The notion that blends of methanol and ethanol with gasoline with the same stoichiometry as E85 can be used in flex-fuel engines has been investigated on the Volvo 1.8 L engine and Hyundai 2.4 L. The performance, efficiency and emissions of different GEM blends were measured and compared to Energies 2020, 13, 6390 17 of 31 those on the pure constituent components to evaluate their potential. Additionally, the Audi engine was used for a preliminary study of the knock behavior of these blends because it had a higher CR.

EnergiesThe blends 2020, 13, testedx FOR PEER in theREVIEW 1.8 L Volvo engine are shown in Table5. All were splash17 blendedof 32 using a conventional 95 RON pump gasoline. (Note that the blends tested on the Hyundai engine wereengine very similar was used in for terms a preliminary of volume study percentages of the knock but behavior used a of Haltermann these blends EEE because certification it had a higher gasoline.) In TableCR. 5, there are two binary blends: Blend A represents an idealized ‘normal’ E85 and Blend D is the iso-stoichiometricThe blends tested gasoline-methanol in the 1.8 L Volvo engine equivalent. are shown This in Table has 57% 5. Allv /werev methanol, splash blended 1% v/v usingmethanol a conventional 95 RON pump gasoline. (Note that the blends tested on the Hyundai engine were very more than in the earlier study of Turner et al. [22]. This difference is because of the stoichiometric AFR similar in terms of volume percentages but used a Haltermann EEE certification gasoline.) In Table of the gasoline in this blend. The two remaining blends in Table5 are true ternary blends. Blend B has 5, there are two binary blends: Blend A represents an idealized ‘normal’ E85 and Blend D is the iso- half thestoichiometric ethanol of gasoline-methanol Blend A, and Blend equivalent. C has halfThis thathas 57% ofBlend v/v methanol, B, and 1% in eachv/v methanol case, the more gasoline than and methanolin the componentearlier study volumes of Turner has et al. been [22]. adjusted This diff toerence maintain is because the targetof the stoichiometry.stoichiometric AFR of the gasoline in this blend. The two remaining blends in Table 5 are true ternary blends. Blend B has half the ethanolTable 5. ofProperties Blend A, ofand the Blend GEM C blends has half tested that on of the Blend 4-cylinder B, and 1.8 in L each PFI productioncase, the gasoline engine. and methanol component volumes has been adjusted to maintain the target stoichiometry. Fuel Blend Blend A Blend B Blend C Blend D

Table 5. Properties of the GEM blends testedG15 on the 4-cylinderG29.5 1.8 L PFI productionG37 engine.G43 Component ratios (% v/v) E85 E42.5 E21 E0 Fuel Blend M0 Blend A BlendM28 B Blend CM42 Blend D M57 Oxygen content (% m/m) 23.34G15 22.74G29.5 G37 22.38G43 22.54 Component ratios (% v/v) E85 E42.5 E21 E0 Gravimetric LHV (MJ/kg) 29.22 29.48 29.66 29.53 M0 M28 M42 M57 VolumetricOxygen LHV (MJ content/L) (% m/m) 22.82 23.34 22.7922.74 22.38 22.8122.54 22.59 AFRstoichGravimetric (kg/kg) LHV (MJ/kg) 9.72 29.22 29.489.76 29.66 9.8 29.53 9.73

∆hvapVolumetric(kJ/kg) LHV (MJ/L) 762.4 22.82 762.7 22.79 22.81 761.5 22.59 770.8 AFRstoich (kg/kg) 9.72 9.76 9.8 9.73 α (H/C in fuel) 2.83 2.86 2.87 2.9 ∆hvap (kJ/kg) 762.4 762.7 761.5 770.8 Specific CO2 emissionsα (H/C (gCO in2/ MJ)fuel) 71.69 2.83 71.56 2.86 2.87 71.49 2.9 71.41 Specific CO2 emissions (gCO2/MJ) 71.69 71.56 71.49 71.41

Apparent in this data is the fact that as the gasoline proportion increases so the specific CO2 Apparent in this data is the fact that as the gasoline proportion increases so the specific CO2 emissions (reported in gCO2/MJ) decrease; this is because methanol has the lowest carbon intensity in emissions (reported in gCO2/MJ) decrease; this is because methanol has the lowest carbon intensity terms of gCO2/MJ of the three blend components, and more methanol has to be added as the gasoline in terms of gCO2/MJ of the three blend components, and more methanol has to be added as the contentgasoline rises. content rises.

4.1. Performance4.1. Performance First,First, the resultsthe results from from the the Volvo Volvo 1.8 1.8 L L engineengine will be be discussed. discussed. In Inthis this engine, engine, four fourdifferent different GEM GEM blendsblends were were tested tested for for which which the the brake brake thermal thermal efficiency efficiency (BTE) (BTE) is isshown shown in Figure in Figure 12. In 12 this. In figure, this figure, the BTEthe BTE is shown is shown for for all all blends blends at at a a constant constant loadload of of 40 40 Nm Nm (2.8 (2.822 bar bar BMEP) BMEP) over over a speed a speed range range from from 15001500 to 3500 to 3500 rpm. rpm. Within Within the the range range of of experimental experimental uncertainty, uncertainty, all all of ofthe the iso-stoichiometric iso-stoichiometric blends blends havehave similar similar BTEs. BTEs. Furthermore, Furthermore, this this is is valid valid for all of of the the loads loads tested. tested.

Figure 12. BTE of various iso-stoichiometric GEM blends as a function of engine speed for a fixed Figure 12. BTE of various iso-stoichiometric GEM blends as a function of engine speed for a fixed brake torque of 40 Nm (2.82 bar BMEP). brake torque of 40 Nm (2.82 bar BMEP).

Energies 2020, 13, x FOR PEER REVIEW 18 of 32

Energies 2020, 13, x FOR PEER REVIEW 18 of 32 EnergiesFigure2020, 1313, 6390displays a comparison of the BTEs for an average for the GEM blends with18 their of 31 gasoline,Figure methanol, 13 displays ethanol a comparison components of at the fixed BTEs loads for of an 40 average and 80 Nm.for the This GEM was performedblends with for their the GEM blends because their results shown in Figure 12 are within experimental uncertainty. When gasoline,Figure methanol, 13 displays ethanol a comparison components of the at BTEsfixed forloads an averageof 40 and for 80 the Nm. GEM This blends was withperformed their gasoline, for the compared to gasoline, it is clear that the GEM blends show significant efficiency gains, being close to GEMmethanol, blends ethanol because components their results at shown fixed loads in Figure of 40 12 and are 80 within Nm. Thisexperimental was performed uncertainty. for the When GEM that of pure ethanol. The best for BTE is methanol, which might be expected from the results already comparedblends because to gasoline, their results it is clear shown that in the Figure GEM 12 blends are within show experimental significant efficiency uncertainty. gains, When being compared close to discussed. thatto gasoline, of pure itethanol. is clear The that best the GEMfor BTE blends is methanol, show significant which might efficiency be expected gains, being from closethe results to that already of pure discussed.ethanol. The best for BTE is methanol, which might be expected from the results already discussed.

Figure 13. Brake thermal efficiency when operating on various fuels as a function of engine speed for Figure 13. Brake thermal efficiency when operating on various fuels as a function of engine speed for different fixed brake torques of 40 Nm (2.82 bar BMEP) (dashed lines) and 80 Nm (5.64 bar BMEP) Figuredifferent 13. fixedBrake brake thermal torques efficiency of 40 when Nm (2.82 operating bar BMEP) on vari (dashedous fuels lines) as a function and 80 Nm of engine (5.64 bar speed BMEP) for (solid lines). different(solid lines). fixed brake torques of 40 Nm (2.82 bar BMEP) (dashed lines) and 80 Nm (5.64 bar BMEP) (solid lines). BecauseBecause of of the the near-identical near-identical LHV LHV of of the the four four is iso-stoichiometrico-stoichiometric GEM GEM blends blends (see (see Table Table 55)) the BSFCBSFCBecause results results ofdo do the not not near-identical depend on mixture LHV of composition.the four iso-stoichiometric This This is is shown shown GEM in in Figure Figure blends 14.14 (see. Figure Figure Table 15 155) istheis a a comparisonBSFCcomparison results of ofdo the the not BSFCs BSFCs depend of of the theon componentsmixture components composition. and and the the mean meanThis ofis of theshown the GEM GEM in blends blendsFigure for14. for 40 40Figure and and 8015 80 Nm.is Nm. a comparisonDespiteDespite better better of efficiencythe effi ciencyBSFCs with withof the GEM GEM components blends, blends, this this and en engine thegine mean consumes consumes of the ~32% ~32%GEM moreblends more fuel fuel for by by40 volume volumeand 80 thanNm. than Despitewhenwhen operating operating better efficiency on on GEM GEM mixtures mixtureswith GEM than than blends, when when this it it is is usen usinginggine gasoline. gasoline. consumes Because Because ~32% of ofmore th thee propertiesfuel properties by volume of of alcohols alcohols than whenthisthis difference di operatingfference would would on GEM be be expectedmixtures expected thanto to be be whensmalle smaller itr is for for us moderning modern gasoline. highly-downsized highly-downsized Because of the and properties and pressure-charged pressure-charged of alcohols thisdirect-injectiondirect-injection difference would (DI) (DI) engines.engines. be expected With With to aa befull full smalle optimization optimizationr for modern of of such such highly-downsized an an engine, engine, the the situationand situation pressure-charged would would in in all all direct-injectionlikelihoodlikelihood be be further further (DI) engines. improved. improved. With a full optimization of such an engine, the situation would in all likelihood be further improved.

Figure 14. BSFC of GEM blends as a function of engine speed for a fixed brake torque of 40 Nm (2.82 Figurebar BMEP). 14. BSFC of GEM blends as a function of engine speed for a fixed brake torque of 40 Nm (2.82 Figurebar BMEP). 14. BSFC of GEM blends as a function of engine speed for a fixed brake torque of 40 Nm (2.82 bar BMEP).

Energies 2020, 13, x FOR PEER REVIEW 19 of 32

Energies 2020, 13, 6390 19 of 31 Energies 2020, 13, x FOR PEER REVIEW 19 of 32

Figure 15. BSFC as a function of engine speed for different fixed brake torques of 40 Nm (dashed

lines)Figure and 15. 80BSFC Nm as(solid a function lines). of engine speed for different fixed brake torques of 40 Nm (dashed lines) Figure 15. BSFC as a function of engine speed for different fixed brake torques of 40 Nm (dashed and 80 Nm (solid lines). lines)In the and following 80 Nm (solid section, lines). the focus will be on the difference between the two binary GEM blends as measuredIn the following on the section,direct-injected the focus Hyundai will be onengine the di: ffE85erence and between M56. Note thetwo that binary because GEM a blendsdifferent as gasolinemeasuredIn the was onfollowing theused direct-injected to section, formulate the Hyundai focus the GEMwill engine: be blends on the E85 fodifference andr these M56. tests, between Note the that thecomposition because two binary a di ffoferent GEM thegasoline blendsbinary asgasoline-methanolwas measured used to formulate on the blend direct-injected the has GEM changed blends Hyundai slightly for these enginecompar tests,: the edE85 compositionto and the blendsM56. ofNote listed the binarythat in Table because gasoline-methanol 5, reducing a different by gasoline1%.blend From has was the changed resultsused slightlyto shown formulate compared in Figures the to GEM12–15 the blends blendsit is expected listedfor these in that Table tests, the5, results reducing the composition of byevery 1%. other From of GEMthe the resultsbinary blend gasoline-methanolshownshown there in Figures would 12 blend be–15 aligned it has is expected changed with the that slightly results the results comparof the of binaryed every to theblends other blends GEM tested listed blend in thisin shown Table study. 5, there reducing would by be 1%.aligned FromFirst, with themeasurements theresults results shown of were the in binaryFigures performed blends 12–15 using testedit is expectedE85 in thisat fixed study.that loadsthe results of 50, of 75 every and other150 Nm GEM across blend a shownspeedFirst, range there measurements ofwould 1500 beto aligned3000 were rpm, performedwith then the these results using were of E85 therepeated at binary fixed using loadsblends M56. of tested 50, In 75 order andin this 150 to study. Nmproduce across the a speedsame torquerangeFirst, ofoutput, 1500 measurements to very 3000 small rpm, wereadjustments then performed these were of the using repeated throttle E85 using atvalve fixed M56. were loads In permitted order of 50, to 75 produce with and the 150 the two Nm same fuels, across torque but a speedtheoutput, start range very of of smallinjection 1500 adjustments to 3000timing rpm, and of then the injectionthrottle these were valvepressure repeated were and permitted using ignition M56. with timingIn the order two were to fuels, produce kept but thethe the start same of torquethroughout.injection output, timing In veryFigure and injectionsmall 16, BTEadjustments pressure for the andtwo of ignitionthefuels throttle is timingshown valve wereat thewere kept different permitted the same engine throughout.with loads, the two and In fuels, Figurethe only but 16 , thesignificantBTE start for theof difference twoinjection fuels (i.e., timing is greater shown and than at injection the 1%) di ffcanerent pressure be seen engine at and 150 loads, Nmignition andand the2500timing only rpm, were significant where kept the errorthe diff erencesame bars throughout.do(i.e., not greater overlap. thanIn Figure 1%) can16, BTE be seen for atthe 150 two Nm fuels and is 2500 shown rpm, at wherethe different the error engine bars doloads, not overlap.and the only significant difference (i.e., greater than 1%) can be seen at 150 Nm and 2500 rpm, where the error bars do not overlap.

Figure 16. Cont.

EnergiesEnergies 20202020,, 1313,, x 6390 FOR PEER REVIEW 2020 of of 32 31 Energies 2020, 13, x FOR PEER REVIEW 20 of 32

Figure 16. BTE as a function of engine speed for different fixed brake torques of (A) 50 Nm, (B) 75 Figure 16. BTE as a function of engine speed for different fixed brake torques of (A) 50 Nm, (B) 75 Nm, Nm, and (C) 150 Nm. Figureand (C )16. 150 BTE Nm. as a function of engine speed for different fixed brake torques of (A) 50 Nm, (B) 75 Nm, and (C) 150 Nm. AsAs stated, stated, most most of of the the injection injection parameters parameters were were fixed, fixed, the the injection injection duration duration being being the the only only variable.variable.As stated, Figure most1717 showsshows of the the the injection di differencefference parameters in in this this variable vawereriable between fixed, between the E85 injection andE85 M56.and duration M56. The diTheff erencebeing difference the is bigger only is biggervariable.than 1% than forFigure 1% 50 Nmfor 17 50 andshows Nm 75 Nmandthe for75difference Nm one operatingfor inone this oper point,vaatingriable whereas point, between whereas for 150E85 Nm, andfor 150 theM56. diNm,ff Theerences the difference differences are larger is arebiggerbut larger still than only but 1% 2.7%still for only at50 maximum. Nm 2.7% and at maximum.75 Nm for one operating point, whereas for 150 Nm, the differences are larger but still only 2.7% at maximum.

Figure 17. Difference in injection duration of M56 compared to E85 as a function of engine speed for Figure 17. Difference in injection duration of M56 compared to E85 as a function of engine speed for different fixed brake torques of 50, 75, and 150 Nm. differentFigure 17. fixed Difference brake torques in injection of 50, duration 75, and 150of M56 Nm. compar ed to E85 as a function of engine speed for 4.2. NOxdifferent Emissions fixed brake torques of 50, 75, and 150 Nm. 4.2. NOx Emissions Figure 18 compares the four GEM blends tested on the 1.8 L PFI Volvo engine in terms of their 4.2. NOx Emissions engine-outFigure 18 NOx compares emissions. the four Gasoline GEM fuelingblends tested generates on the the 1.8 highest L PFI NOxVolvo emissions engine in andterms methanol of their engine-outthe lowest,Figure NOx 18 while compares emissions. those fromthe Gasoline four the GEM fueling blends blends gene tested lierates somewhere on the the highest 1.8 between L PFINOx Volvo emissions thetwo. engine Since and in termsmethanol most of NOx their the is lowest,engine-outproduced while by NOx the those emissions. thermal from mechanism the Gasoline GEM fueling (itselfblends highly genelie sorates dependentmewhere the highest between on the NOx maximum the emissions two. temperatureSince and mostmethanol reached)NOx the is producedlowest,alcohol fuelswhile by the generate those thermal from lower mechanism the levels GEM of (itself thisblends pollutant highly lie so dependent becausemewhere of onbetween their the lower maximum the adiabatic two. temperature Since flame most temperature. reached) NOx is alcoholproducedAt lower fuels engine by the generate speedsthermal and lowermechanism loads levels lower (itself of NOx this highly emissions pollutant dependent are because believed on the of maximum to betheir partly lower temperature the resultadiabatic of reached) elevated flame temperature.alcohollevels of fuels internal Atgenerate lower EGR; engine thislower is because speedslevels and withof thisloads the usepollutant lower of conventional NOx because emissions of throttling theirare believed lower to control adiabaticto be the partly engine’s flame the resulttemperature.output of to elevated a fixed At lower load, levels theengine of vacuum internal speeds in EGR;the and intake this loads is is lobecause higherwer NOx forwith theemissions the alcohol use ofare fuels conventional believed due to theto be factthrottling partly that theythe to controlresulthave higher of the elevated engine’s BTE.Note levels output that of no internalto knocka fixed EGR; occurred load, this the duringis vacuum because the measurementsin with the theintake use is of using higher conventional all for the the GEM alcoholthrottling blends fuels and to duecontrolMBT to timing thethe factengine’s could that be theyoutput set forhave to every highera fixed load BTE.load, and engineNotethe vacuum that speed. no inknock In the Figure intakeoccurred 18, is generally higher during for speaking, the the measurements alcohol the higher fuels usingduethe gasolineto all the the fact GEM content that blends they the higherandhave MBT higher the timing NOx BTE. output, could Note be regardlessthat set nofor knockevery of the loadoccurred proportions and engineduring of speed. the typemeasurements In Figure of alcohol 18, generallyusingused toall create the speaking, GEM the blend.blends the higher Furtherand MBT the research timinggasoline incould thiscontent be areas se thet wouldfor higher every be load warranted.the NOxand engine output, speed. regardless In Figure of the 18, proportionsgenerally speaking, of the type the ofhigher alcohol the used gasoline to create content the blend.the higher Further the researchNOx output, in this regardless areas would of thebe warranted.proportions of the type of alcohol used to create the blend. Further research in this areas would be warranted.

Energies 2020, 13, x FOR PEER REVIEW 21 of 32 Energies 2020, 13, 6390 21 of 31 Energies 2020, 13, x FOR PEER REVIEW 21 of 32

Figure 18. NOx emissions as a function of engine speed for a fixed brake torque of 40 Nm (2.82 bar Figure 18. NOx emissions as a function of engine speed for a fixed brake torque of 40 Nm (2.82 bar BMEP). FigureBMEP). 18. NOx emissions as a function of engine speed for a fixed brake torque of 40 Nm (2.82 bar BMEP). TheThe engine-out engine-out NOx NOx emissions emissions of of E85 E85 and and M56 M56 as as a afunction function of of engine engine speed, speed, measured measured on on the the DIDI Hyundai HyundaiThe engine-out engine, engine, NOxare are shown shownemissions in in Figure Figure of E85 19. 19 and .Here Here M56 the the as load a load function is is higher, higher, of engineat at 75 75 Nm, Nm,speed, and and measured again again MBT MBT on was the was DIachievableachievable Hyundai throughout. throughout. engine, are ForshownFor NOx,NOx, in E85 FigureE85 performs performs 19. Here slightly slightly the load better better is thanhigher, than M56 at M56 although 75 Nm, although and the erroragain the barserror MBT overlap barswas achievableoverlapfor most for points. throughout.most points. For NOx, E85 performs slightly better than M56 although the error bars overlap for most points.

Figure 19. NOx emissions as a function of engine speed for a fixed brake torque of 75 Nm Figure(5.29 bar 19. BMEP).NOx emissions as a function of engine speed for a fixed brake torque of 75 Nm (5.29 bar BMEP).Figure 19. NOx emissions as a function of engine speed for a fixed brake torque of 75 Nm (5.29 bar 4.3.BMEP). Knock Behavior 4.3. KnockFrom Behavior all of the previous work, it is expected that the different GEM fuels would have very similar 4.3.knock KnockFrom behavior Behaviorall of the [ 17 ,19previous,20]. To work, investigate it is expected this, all thethat blends the different and the GEM individual fuels would components have very were similartestedFrom onknock theall behaviorof Audi-based the previous [17,19,20]. single-cylinder work, To itinvestigate is expected engine, th butis, that all configured thethe blendsdifferent with and GEM a th higher-than-standarde individualfuels would components have CRvery of weresimilar13.1:1 tested asknock shown on behavior the in TableAudi-based [17,19,20].4 above. single-cylinder The To engineinvestigate was en operatedthgine,is, all but the as configured a blends fixed BMEP and with th ofe 7.71individuala higher-than-standard bar (i.e., components 25 Nm) and wereCRthe of speed tested13.1:1 was onas shown fixedthe Audi-based at in 2000 Table rpm. 4 single-cylinder above. For each The fuel, engine theengine, ignitionwas butoperated timingconfigured as was a fixed advanced with BMEP a higher-than-standard until of 7.71 an intermediatebar (i.e., 25 CRNm)knocking of and 13.1:1 the condition as speed shown was wasin fixedTable obtained. at 4 above.2000 Inrpm. The addition For engine each to was thefuel, operated RON the ignition 95 gasolineas a fixedtiming used BMEP was as advancedof the 7.71 component bar until (i.e., an25 to Nm)intermediateformulate and the the speedknocking blends, was acondition fixed RON at 98 2000 gasolinewas rpm. obtained. wasFor each also In investigated.fuel,addition the ignitionto the The RONtiming results 95 was ofgasoline thisadvanced experiment used until as the an are intermediatecomponentshown in Figure to knocking formulate 20. condition the blends, was a obtained.RON 98 gasoIn additionline was toalso the investigated. RON 95 gasoline The results used ofas thisthe componentexperimentIn Figure areto formulate 20shown, all the in GEMFigurethe blends, blends 20. a have RON the 98 same gaso ignitionline was timing also investigated. at incipient knock The (determinedresults of this by experimenta combination are shown of audible in Figure knock 20. and the third derivative of the pressure signal passing a threshold value). It is therefore reasonable to suppose that they have constant octane numbers. Furthermore, pure ethanol produces the same result. The GEM blends were configured to be equivalent to E85 which

Energies 2020, 13, 6390 22 of 31 has 93% molar concentration of ethanol as discussed earlier and so this is understandable. Methanol Energiesstill displays 2020, 13,superior x FOR PEER knock REVIEW resistance to all of the other fuels. 22 of 32

FigureFigure 20. 20. Knock-limitedKnock-limited spark spark advance advance of all of GEM all GEM blends blends and individual and individual blend components. blend components. Audi- basedAudi-based single-cylinder single-cylinder research research engine configured engine configured with a higher-than-standard with a higher-than-standard CR of 13.1:1. CR Operating of 13.1:1. Operating point of 25 Nm (BMEP = 7.71 bar) and 2000 rpm. point of 25 Nm (BMEP = 7.71 bar) and 2000 rpm. Since the fuels used in this section were splash blended using relatively high-octane gasoline, this In Figure 20, all the GEM blends have the same ignition timing at incipient knock (determined work is considered to support the observations of Waqas et al. that using such a gasoline as a blend by a combination of audible knock and the third derivative of the pressure signal passing a threshold component ensures linearity in the knock resistance of the resulting fuel [23]. value). It is therefore reasonable to suppose that they have constant octane numbers. Furthermore, pure5. Particulate ethanol produces Emissions the from same GEM result. Fuels The in GEM a Direct-Injection blends were Spark-Ignitionconfigured to be Engine equivalent to E85 which has 93% molar concentration of ethanol as discussed earlier and so this is understandable. MethanolIn Section still displays3, high-load superior PN emissionsknock resistance from the to Ultraboostall of the other 4-cylinder fuels. engine operating on GEM blendsSince were the reported. fuels used Here, in this particulate section were emissions splash areblended presented using from relatively four GEMhigh-octane fuels and gasoline, a base thisgasoline, work measuredis considered from to asupport single-cylinder the observations DI engine of at Waqas lighter et engine al. that loads. using Thissuch wasa gasoline performed as a blendbecause component while gaseous ensures emissions linearity havein the been knoc presentedk resistance several of the times resulting for GEM fuel [23]. blends, PN emission have not been recorded as much. 5. Particulate Emissions from GEM Fuels in a Direct-Injection Spark-Ignition Engine 5.1. Fuels Tested In Section 3, high-load PN emissions from the Ultraboost 4-cylinder engine operating on GEM For this experiment, four GEM fuels were splash blended from the three components. Their blends were reported. Here, particulate emissions are presented from four GEM fuels and a base composition was the same as those tested in the first paper published by Turner et al. [17] and is gasoline, measured from a single-cylinder DI engine at lighter engine loads. This was performed reproduced in Table6, together with their predicted dry vapor pressure equivalent (DVPE). because while gaseous emissions have been presented several times for GEM blends, PN emission have not been recorded as much. Table 6. Composition of the GEM fuels tested in the single-cylinder DI engine.

5.1. Fuels Tested Gasoline Ethanol Methanol Predicted DVPE * Fuel (% v/v) (% v/v) (% v/v) (kPa) For this experiment, four GEM fuels were splash blended from the three components. Their Gasoline 100.0 0.0 0.0 62 composition was the same as those tested in the first paper published by Turner et al. [17] and is reproduced GEMin Table A 6, together 15.0 with their pr 85.0edicted dry vapor 0.0 pressure equivalent 39 (DVPE). GEM B 29.5 42.5 28.0 67 GEMTable C 6. Composition 37.0 of the GEM 21.0 fuels tested in the 42.0 single-cylinder DI engine. 78 GEM DGasoline 44.0 Ethanol 0.0 Methanol 56.0 Predicted DVPE 87* Fuel (%* Predictionv/v) derived(% v/v) from a UNIFAC(% v/v) model [58]. (kPa) Gasoline 100.0 0.0 0.0 62 The base gasolineGEM A for these 15.0 blends was 85.0 a CEC RF-02-03-specification 0.0 39 gasoline supplied by BP. This gasoline hasGEM no B oxygenated 29.5 components 42.5 present in28.0 it, and its properties 67 are shown in Table7. GEM C 37.0 21.0 42.0 78 GEM D 44.0 0.0 56.0 87 * Prediction derived from a UNIFAC model [58].

Energies 2020, 13, 6390 23 of 31

Note that in this table the abbreviation DVPE refers to dry vapor pressure equivalent, and IBP to the initial boiling point in the distillation curve, T90 to the point at which 90% of the fuel mass has boiled off, and FBP to the final boiling point at which the last components boil off.

Table 7. Composition of the base CEC RF-02-03 gasoline test fuel.

Parameter Unit Result RON 97.0 MON 87.7 Density @ 15 ◦C kg/L 0.7439 DVPE kPa 60.0 Aromatics % v/v 29.0 Olefins % v/v 1.8 Saturates % v/v 69.2 Carbon % m/m 86.74 Hydrogen % m/m 13.26 Gross Calorific Value MJ/kg 46.19 Net Calorific Value MJ/kg 43.38 IBP ◦C 28.8 T90 ◦C 166.6 FBP ◦C 200.8

5.2. Expected Fuel Effects on PN Emissions It has previously been noted [59] that AFRs rich of stoichiometric increase particulate emissions. This is true of ethanol, but the increase in particulate matter (PM) emissions rich of stoichiometric for ethanol is very small, and anyway compared to gasoline it is effectively a dramatic reduction; as an example, E10 has been shown to produce a greater than one order of magnitude reduction in PM emissions compared with straight gasoline [60]. This is thought to be due to the presence of oxygen in the fuel molecule which reduces the concentration of intermediate species that are important to the formation of precursors to soot [61]. As methanol is added to a fuel, its vapor pressure increases, as can be seen by the change in predicted DVPE in Table6. High vapor pressures can result in the flash boiling of the fuel on injection, leading to very poor mixture preparation and concomitantly high levels of PM emissions. Flash boiling occurs when a liquid jet is injected into a volume with a pressure lower than the saturation vapor pressure of the liquid. Here the fuel evaporates so quickly that it essentially does so as it leaves the injector, occurring in an “onion skin” manner, with the outer layers evaporating first. This is regardless of the composition of the layers and is known to occur in DI engines [62]. The liquid is then rapidly depressurized, becoming superheated and consequently thermodynamically unstable, only regaining its stability by flash boiling.

5.3. Test Engine The single-cylinder engine used for these tests was based on that of the Jaguar AJ133 5.0 L V8 engine [63]. Its specification are shown in Table8. It is described more fully in [ 58] and here it was not used in its optical configuration. Energies 2020, 13, 6390 24 of 31

Table 8. Single-cylinder DI engine specifications. IVO = intake valve opens; IVC = intake valve closes; EVO = exhaust valve opens; EVC = exhaust valve closes; ATDC = after top dead center; ABDC = after bottom dead center; BBDC = before bottom dead center.

Bore (mm) 89 Stroke (mm) 90.3 Cylinder capacity (cc) 562 Compression ratio (:1) 11.1 Fuel injection pressure (bar) 150 IVO (◦ATDC) 33.5 IVC (◦ABDC) 61.5 EVO (◦BBDC) 65 EVC (◦ATDC) 5

5.4. Particulate Measurements A Cambustion DMS500 [64] was used to measure PN emissions. While this was not a legislative test, in order to make these results representative of PN emissions measured during such a test, a Wiebe filter was used to mimic the effect of the PMP legislatively-compliant protocol [65]. Measurements at 1 Hz were taken for 60 s at steady state, and the mean and standard deviation (σ) of the data over that time were taken.

5.5. Results For these experiments, the engine was operated at two fixed indicated mean effective pressure (IMEP) points (corresponding to low and medium loads) at 1500 rpm as shown as in Table9. For each load, open-loop AFR control was used at λ = 0.9 and 1.01. λ = 1.01 was chosen to avoid any small rich mixture excursions which are known to have a large effect on the PM emissions from non-oxygenated fuels [59]. Because of this, these rich excursions tend to dominate drive cycle PM results, hence testing was also conducted at λ = 0.9.

Table 9. Engine operating conditions for GEM fuels and gasoline. BTDC = before top dead center.

IMEP (bar) 2.6 and 4.9 Inlet air temperature (◦C) 40 Coolant temperature (◦C) 60 Relative AFR (λ) 0.9 and 1.01 Engine speed (rpm) 1500 Start of ignition (◦BTDC) 35 Start of injection (◦BTDC) 280

Figure 21 shows the PN emissions from the GEM fuels at light load (2.6 bar IMEP). In these results, there is a clear trend of increasing PN emissions for the GEM fuels from blends A to C as the methanol quantity increases. Confidence in these results is provided by the fact that the experimental sequence was randomized and a single repeat of GEM D was undertaken with highly repeatable results. In Figure 21, it also is shown clearly that once a certain level of methanol is passed in the fuel (say GEM Blend C), there is no further increase in PN emissions. Furthermore the base gasoline values fall half way in the sequence, i.e., gasoline gives higher levels of PN emissions than GEM B, but lower than GEM C. At this load point, this in turn suggests that the presence of ethanol reduces PN while methanol does the opposite, despite a higher proportion of oxygenated components; it is flash boiling that also explains this. Since the methanol component flashes first, this results in poor mixture preparation and hence increases PN emissions. As the amount of methanol in the fuel is reduced, its deleterious effect in this regard is reduced, and instead other factors that cause oxygenate fuels to reduce particulate emissions become important, causing PN emissions to then be reduced from the gasoline levels. Energies 2020, 13, x FOR PEER REVIEW 25 of 32

EnergiesEnergies2020 2020, 13, 13, 6390, x FOR PEER REVIEW 2525 of of 31 32

Figure 21. PN emissions from GEM fuels at 2.6 bar IMEP. Error bars correspond to ±σ.

In Figure 21, it also is shown clearly that once a certain level of methanol is passed in the fuel (say GEM Blend C), there is no further increase in PN emissions. Furthermore the base gasoline values fall half way in the sequence, i.e., gasoline gives higher levels of PN emissions than GEM B, but lower than GEM C. At this load point, this in turn suggests that the presence of ethanol reduces PN while methanol does the opposite, despite a higher proportion of oxygenated components; it is flash boiling that also explains this. Since the methanol component flashes first, this results in poor mixture preparation and hence increases PN emissions. As the amount of methanol in the fuel is reduced, its deleterious effect in this regard is reduced, and instead other factors that cause oxygenate fuels to reduceFigure particulate 21. PN emissions emissions from become GEM fuels important, at 2.6 bar IMEP.causing Error PN bars emissions correspond to tothenσ .be reduced from the gasolineFigure levels. 21. PN emissions from GEM fuels at 2.6 bar IMEP. Error bars correspond to± ±σ. Figure 2222 shows shows the the PN PN emissions emissions from from Blend Blend A, Blend A, Blend D and D gasoline and gasoline at the higher at the load higher operating load In Figure 21, it also is shown clearly that once a certain level of methanol is passed in the fuel pointoperating (4.9 barpoint IMEP). (4.9 bar While IMEP). the lightWhile load the trendslight load are repeated,trends are the repeated, PN emissions the PN ofemissions Blend D of are Blend now (say GEM Blend C), there is no further increase in PN emissions. Furthermore the base gasoline belowD are now those below of the those gasoline of the baseline. gasoline It baseline. is hypothesized It is hypothesized that at higher that pressuresat higher thepressures flash boiling the flash of values fall half way in the sequence, i.e., gasoline gives higher levels of PN emissions than GEM B, methanolboiling of ismethanol suppressed, is suppressed, allowing those allowing factors those which factors generally which cause generally oxygenate cause fuels oxygenate to reduce fuels PN to but lower than GEM C. At this load point, this in turn suggests that the presence of ethanol reduces dominatereduce PN in to the dominate same way in the that same they way do for that ethanol. they do for ethanol. PN while methanol does the opposite, despite a higher proportion of oxygenated components; it is flash boiling that also explains this. Since the methanol component flashes first, this results in poor mixture preparation and hence increases PN emissions. As the amount of methanol in the fuel is reduced, its deleterious effect in this regard is reduced, and instead other factors that cause oxygenate fuels to reduce particulate emissions become important, causing PN emissions to then be reduced from the gasoline levels. Figure 22 shows the PN emissions from Blend A, Blend D and gasoline at the higher load operating point (4.9 bar IMEP). While the light load trends are repeated, the PN emissions of Blend D are now below those of the gasoline baseline. It is hypothesized that at higher pressures the flash boiling of methanol is suppressed, allowing those factors which generally cause oxygenate fuels to reduce PN to dominate in the same way that they do for ethanol.

Figure 22. PN emissions from gasoline and two GEM fuels at 4.9 bar IMEP. Error bars correspond Figure 22. PN emissions from gasoline and two GEM fuels at 4.9 bar IMEP. Error bars correspond to to σ. ±σ.± Overall, the results in this section show that using GEM fuels in DI engines will not significantly Overall, the results in this section show that using GEM fuels in DI engines will not significantly impact PN emissions versus gasoline, other than possibly to decrease them significantly. The caveat is impact PN emissions versus gasoline, other than possibly to decrease them significantly. The caveat that at light load, and when methanol is present in a high proportion, the results suggest that a slight is that at light load, and when methanol is present in a high proportion, the results suggest that a increase in PN emissions over those from gasoline may be seen, but as the load increases, then all slight increase in PN emissions over those from gasoline may be seen, but as the load increases, then such GEM fuels will then give lower PN emissions. This is likely due to methanol flash boiling at low all such GEM fuels will then give lower PN emissions. This is likely due to methanol flash boiling at pressures, and this effect reducing as the BMEP increases. The equivalent conditions are not reached for the ethanol in the fuel for either test point, so it tends to suppress PN emission formation. Further research is warranted into this phenomenon and whether, where other alcohols are being considered as blendFigure components 22. PN emissions in future from fuels gasoline (for example, and two GEM butanol), fuels at they 4.9 showbar IMEP. any Error similar bars trends. correspond to ±σ. 6. Conclusions ThroughOverall, investigationsthe results in this in several section di showfferent that research using GEM engines, fuels this in paperDI engines has shown will not that, significantly at mid-to high-blendimpact PN rates: emissions versus gasoline, other than possibly to decrease them significantly. The caveat is that at light load, and when methanol is present in a high proportion, the results suggest that a slight increase in PN emissions over those from gasoline may be seen, but as the load increases, then all such GEM fuels will then give lower PN emissions. This is likely due to methanol flash boiling at

Energies 2020, 13, 6390 26 of 31

1. In support of earlier work by the authors and other research groups, ternary (three component) mixtures of gasoline, ethanol, and methanol (or GEM) can be formulated using an iso-stoichiometric approach to provide a drop-in alternative fuel for any given binary gasoline-ethanol blend. 2. When so configured, such GEM fuels have matched gravimetric lower heating values, heats of vaporization, and molar concentrations of reactants in the charge. The latter two characteristics gives rise to similar molar expansion ratios and extremely close RON and MON values. With these come very close sensitivities. 3. The improvement in full-load performance of such high-blend-rate fuels in terms of KLSA, and the associated better BMEP and thermal efficiency, is entirely in line with what is expected for high-blend alcohols over conventional gasoline fuels. There is also the expected disadvantage in terms of BSFC due to the reduction in gravimetric energy content in relation to pure hydrocarbon fuels. These results are consistent with the literature and regardless of EGR being used or the amount of exhaust back pressure applied to the combustion system (to simulate turbocharged or supercharged operation of an engine). 4. NOx emissions are shown to be lower with the alcohol blends (and lowest with pure methanol). Furthermore they are consistent for the iso-stoichiometric formulations, with a suggestion that within such GEM blends increased methanol content tends to lower NOx emissions further. 5. When tested in a single-cylinder engine, the knock performance of the tested blends is essentially identical, and at the high blend rate (equivalent to the stoichiometry of E85) the same as pure ethanol and very close to that of methanol. This result is in line with the previous observations on the RON, MON, and sensitivity performance of such blends. 6. In terms of particulate emissions, the results presented here show that using GEM fuels is unlikely to impact the PN emissions from DI engines significantly, and that many of these fuels can decrease the magnitude observed. 7. GEM fuels with a higher proportion of methanol do tend to generate more PN emissions than gasoline at low load, which is attributed to higher degrees of flash boiling at lower in-cylinder pressures. However, this result reverses with increased load and may be due to the change in the dominance of flash boiling effects, with the presence of oxygen in the fuel then reducing the number of particulate precursors formed in combustion. 8. At lower loads the presence of a high proportion of ethanol in a blend significantly reduces PN emissions compared to gasoline. This is again believed to be due to the molecule containing oxygen; the flash boiling effect is also lower for ethanol than it is for methanol, which is believed to contribute to this result. Because methanol can be manufactured using low- and zero-net-carbon pathways (using biological routes which remove CO2 from the atmosphere via plant growth or by direct air capture and utilization in a fully synthetic manufacturing route, respectively), there is thus the potential to introduce it into the fuel pool used by the more than 9 million E85-gasoline flex-fuel vehicles in existence via GEM ternary blends. The lower price of methanol could also beneficially impact the fuel cost on a per unit energy basis, which is generally not the case for E85. Furthermore, although only demonstrated here at a blend rate equivalent to E85, the approach is also viable for any gasoline-ethanol binary blend.

Author Contributions: Conceptualization, J.W.G.T., S.V., F.C.P.L., and P.P.E.; methodology, J.W.G.T., A.G.J.L., L.S., and F.C.P.L.; validation, A.G.J.L., J.V., L.S., and F.C.P.L.; formal analysis, J.W.G.T., S.V., L.S., and F.C.P.L.; investigation, A.G.J.L., L.S., and F.C.P.L.; resources, S.A., S.V., and F.C.P.L.; data curation, J.W.G.T., A.G.J.L., L.S., and F.C.P.L.; writing—original draft preparation, J.W.G.T, L.S., S.V., and F.C.P.L.; writing—review and editing, A.G.J.L., S.A., C.J.B., J.V., L.S., F.C.P.L., and P.P.E.; funding acquisition, S.A., C.J.B., S.V., and F.C.P.L. All authors have read and agreed to the published version of the manuscript. Funding: Elements of the reported research took place as part of the “Ultra Boost for Economy” project part funded by the Technology Strategy Board, now Innovate UK, the United Kingdom’s innovation agency; their support is similarly gratefully acknowledged. J. Vancoillie and L. Sileghem gratefully acknowledge fellowships from the Research Foundation—Flanders (FWO09/ASP/030 and FWO11/ASP/056). Energies 2020, 13, 6390 27 of 31

Acknowledgments: Although too numerous to detail here, the historical support and encouragement of many individuals and organizations involved in the methanol industry and flex-fuel vehicle production is very gratefully acknowledged. Their faith in what was initially a purely academic finding concerning the properties of iso-stoichiometric GEM fuels has been extremely valuable and encouraging throughout the conduct of all aspects of this research. The partners in the Ultraboost project were Jaguar Land Rover, Lotus Engineering, GE Precision Engineering, Shell Global Solutions, CD-adapco, the University of Bath, Imperial College London, and the University of Leeds, and all of their help and contributions to the reported work are gratefully acknowledged. Conflicts of Interest: The authors declare no conflict of interest.

Abbreviations

ABDC After top dead center AFR Air-fuel ratio ATDC After top dead center BBDC Before bottom dead center BEV Battery electric vehicle BMEP Brake mean effective pressure BSFC Brake specific fuel consumption BTDC Before top dead center BTE Brake thermal efficiency CAHU Charge air handling unit CO Carbon monoxide CO2 Carbon dioxide CR Compression ratio DI Direct injection DOHC Double overhead camshaft DVPE Dry vapor pressure equivalent EGR Exhaust gas recirculation EVC Exhaust valve closes EVO Exhaust valve opens FBP Final boiling point (at which the last components boil off in the distillation curve) GEM Gasoline-ethanol–methanol HC Hydrocarbons IBP Initial boiling point (at which the first components boil off in the distillation curve) ICE Internal combustion engine IMEP Indicated mean effective pressure ITE Indicated thermal efficiency IVC Intake valve closes IVO Intake valve opens KLSA Knock-limited spark advance LHV Lower heating value MBT Minimum spark advance for best torque MON Motor octane number NOx Oxides of nitrogen OHC Overhead camshaft PFI Port fuel injection PM Particulate matter PN Particle number ppm Parts per million RON Research octane number SI Spark ignition T90 Point at which 90% of the fuel mass has boiled off (in the distillation curve) TWC Three-way catalyst ∆hvap Heat of vaporization σ Standard deviation Energies 2020, 13, 6390 28 of 31

References

1. Gross, M. A planet with two billion cars. Curr. Biol. 2016, 26, R307–R310. [CrossRef] 2. Gutzmer, P. Mobility for Tomorrow—Between the Poles of Electrification. In Proceedings of the 26th Aachen Colloquium, Aachen, Germany, 9–11 October 2017. 3. Woodward, M. New Market. New Entrants. New Challenges. Battery Electric Vehicles. Deloitte. 2019. Available online: https://www2.deloitte.com/content/dam/Deloitte/uk/Documents/manufacturing/deloitte- uk-battery-electric-vehicles.pdf (accessed on 19 June 2020). 4. Leach, F.C.P.; Akehurst, S.; Brace, C.; Busch, H.; Cairns, A.; Davy, M.; Gavaises, M.; Heikal, M.H.; Linne, M.; Shayler, P.; et al. Road bumps for electric cars (letter). Economist 2017, 413, 16. 5. White, T.L. Alcohol as a Fuel for the Automobile Motor; SAE Technical Paper Series; SAE International: Warrendale, PA, USA, 1907. [CrossRef] 6. Niizato, T. Future direction of Engine and PHEV technology at Honda. In Proceedings of the 14th International Conference on Engines for Vehicles, Hokkaido, Japan, 27–31 August 2017. 7. Fraidl, G.; Kapus, P.; Mitterecker, H.; Weißbäck, M. Internal Combustion Engine 4.0. MTZ Worldw. 2018, 79, 26–33. [CrossRef] 8. Powell, N.; Hill, N.; Bates, J.; Bottrell, N.; Biedka, M.; White, B.; Pine, T.; Carter, S.; Patterson, J.; Yucel, S. Impact Analysis of Mass EV Adoption and Low Carbon Intensity Fuels Scenarios. CONCAWE Report RD18-001538-4; 24 August 2018. Available online: https://www.fuelseurope.eu/wp-content/uploads/Mass-EV-Adoption- and-Low-Carbon-Fuels-Scenarios.pdf (accessed on 30 November 2020). 9. Turner, J.W.G.; Lewis, A.G.J.; Akehurst, S.; Brace, C.J.; Verhelst, S.; Vancoillie, J.; Sileghem, L.; Leach, F.C.P.; Edwards, P.P. Alcohol fuels for spark-ignition engines: Performance, efficiency and emission effects at mid to high blend rates for binary mixtures and pure components. Proc. Inst. Mech. Eng. Part D J. Automob. Eng. 2018, 232, 36–56. [CrossRef] 10. Foong, T.M.; Morganti, K.J.; Brear, M.J.; da Silva, G.; Yang, Y.; Dryer, F.L. The octane numbers of ethanol blended with gasoline and its surrogates. Fuel 2014, 115, 727–739. [CrossRef] 11. Anderson, J.E.; Leone, T.G.; Shelby, M.H.; Wallington, T.J.; Bizub, J.J.; Foster, M.; Lynskey, M.G.; Polovina, D. Octane Numbers of Ethanol-Gasoline Blends: Measurements and Novel Estimation Method from Molar Composition. SAE Technical Paper 2012-01-1274. In Proceedings of the SAE 2012 World Congress, Detroit, MI, USA, 24–26 April 2012. [CrossRef] 12. Bergström, K.; Melin, S.-A.; Jones, C.C. The New ECOTEC Turbo BioPower Engine from GM Powertrain—Utilizing the Power of Nature’s resources. In Proceedings of the 28th International Vienna Motor Symposium, Vienna, Austria, 26–27 April 2007. 13. Bergström, K.; Nordin, H.; Königstein, A.; Marriott, C.D.; Wiles, M.A. ABC—Alcohol Based Combustion Engines—Challenges and Opportunities. In Proceedings of the 16th Aachen Colloquium, Aachen, Germany, 7 October 2007; pp. 1031–1071. 14. Turner, J.W.G.; Pearson, R.J.; Holland, B.; Peck, B. Alcohol-Based Fuels in High Performance Engines. SAE Technical Paper 2007-01-0056. In Proceedings of the SAE Fuels and Emissions Conference, Cape Town, South Africa, 23–25 January 2007. [CrossRef] 15. Wallner, T.; Miers, S.A. Combustion Behavior of Gasoline and Gasoline/Ethanol Blends in a Modern Direct-Injection 4-Cylinder Engine. SAE Technical Paper 2008-01-0077. In Proceedings of the SAE 2008 World Congress, Detroit, MI, USA, 14–17 April 2008. [CrossRef] 16. West, B.H.; López, A.J.; Theiss, T.J.; Graves, R.L.; Storey, J.M.; Lewis, S.A. Fuel Economy and Emissions of the Ethanol-Optimized Saab 9-5-Biopower. SAE Technical Paper 2007-01-3994. In Proceedings of the SAE Powertrain & Fluid Systems Conference and Exhibition, Rosemont, IL, USA, 29 October–1 November 2007. [CrossRef] 17. Turner, J.W.G.; Pearson, R.J.; Purvis, R.; Dekker, E.; Johansson, K.; Bergström, K. GEM Ternary Blends: Removing the Biomass Limit by using Iso-Stoichiometric mixtures of Gasoline, Ethanol and Methanol. SAE Paper Number 2011-24-0113. In Proceedings of the 10th International Conference on Engines and Vehicles, Capri, Naples, Italy, 11–16 September 2011. [CrossRef] Energies 2020, 13, 6390 29 of 31

18. Turner, J.W.G.; Pearson, R.J.; Dekker, E.; Iosefa, B.; Dolan, G.A.; Johansson, K.; Bergström, K. Evolution of alcohol fuel blends towards a sustainable transport energy economy. In Proceedings of the 2012 MIT Energy Initiative Symposium: Prospects for Flexible- and Bi-Fuel Light Duty Vehicles, Cambridge, MA, USA, 19 April 2012. 19. Turner, J.W.G.; Pearson, R.J.; McGregor, M.A.; Ramsay, J.M.; Dekker, E.; Iosefa, B.; Dolan, G.A.; Johansson, K.; Bergström, K. GEM Ternary Blends: Testing Iso-Stoichiometric mixtures of Gasoline, Ethanol and Methanol in a Production Flex-Fuel Vehicle Fitted with a Physical Alcohol Sensor. SAE Paper Number 2012-01-1279. In Proceedings of the SAE 2012 World Congress, Detroit, MI, USA, 24–26 April 2012. [CrossRef] 20. Turner, J.W.G.; Pearson, R.J.; de Goede, S.B.A.; Woolard, C. GEM Ternary Blends: Investigations into Exhaust Emissions, Blend Properties and Octane Numbers. SAE Paper Number 2012-01-1586. In Proceedings of the SAE Powertrains, Fuels and Lubricants Meeting, Malmo, Sweden, 18–20 September 2012. 21. Pitcher, G.; Turner, J.W.G.; Pearson, R.J. GEM Ternary Blends of Gasoline, Ethanol and Methanol: An Initial Investigation into Fuel Spray and Combustion Characteristics in a Direct-Injected Spark-Ignition Optical Engine using Mie Imaging. SAE Technical Paper 2012-01-1740. In Proceedings of the SAE Powertrains, Fuels and Lubricants Meeting, Malmo, Sweden, 18–20 September 2012. [CrossRef] 22. Turner, J.W.G.; Pearson, R.J.; Dekker, E.; Iosefa, B.; Dolan, G.; Johansson, K.; Bergström, K. Extending the role of alcohols as transport fuels using iso-stoichiometric ternary blends of gasoline, ethanol and methanol. Appl. Energy 2013, 102, 72–86. [CrossRef] 23. Waqas, M.; Naser, N.; Sarathy, M.; Feijs, J.; Morganti, K.; Nyrenstedt, G.; Johansson, B. Auto-Ignition of Iso-Stoichiometric Blends of Gasoline-Ethanol-Methanol (GEM) in SI, HCCI and CI Combustion Modes. SAE Technical Paper 2017-01-0726. In Proceedings of the SAE 2017 World Congress, Detroit, MI, USA, 4–6 April 2017. [CrossRef] 24. Pearson, R.J.; Turner, J.W.G.; Bell, A.; de Goede, S.; Woolard, C.; Davy, M.H. Iso-stoichiometric fuel blends: Characterisation of physicochemical properties for mixtures of gasoline, ethanol, methanol and water. Proc. Inst. Mech. Eng. Part D J. Automob. Eng. 2014, 229, 111–139. [CrossRef] 25. Turner, J.W.G.; Pearson, R.J.; Harrison, P.; Marmont, A.; Jennings, R.; Verhelst, S.; Vancoillie, J.; Sileghem, L.; Pecqueur, M.; Martens, K.; et al. Evolutionary decarbonization of transport: A contiguous roadmap to affordable mobility using Sustainable Organic Fuels for Transport. In Proceedings of the Institution of Mechanical Engineers Sustainable Vehicle Technologies Conference, Gaydon, UK, 15–16 November 2012. 26. Pearson RJTurner, J.W.G.; Eisaman, M.D.; Littau, K.A. Extending the Supply of Alcohol Fuels for Energy Security and Carbon Reduction. SAE Technical Paper 2009-01-2764. In Proceedings of the SAE Powertrain, Fuels and Lubricants Meeting, San Antonio, TX, USA, 2–4 November 2009. [CrossRef] 27. Pearson, R.J.; Eisaman, M.D.; Turner, J.W.G.; Edwards, P.P.; Jiang, Z.; Kuznetsov, V.L.; Littau, K.A.; di Marco, L.;

Taylor, S.R.G. Energy storage via carbon-neutral fuels made from CO2, water, and renewable energy. Proc. IEEE 2012, 100, 440–460. [CrossRef] 28. Verhelst, S.; Turner, J.W.G.; Sileghem, L.; Vancoillie, J. Methanol as a fuel for internal combustion engines. Prog. Energy Combust. Sci. 2019, 70, 43–88. [CrossRef] 29. Steinberg, M. Production of synthetic methanol from air and water using controlled thermonuclear reactor power—I. Technology and energy requirement. Energy Convers. 1977, 17, 97–112. [CrossRef] 30. Olah, G.A.; Goeppert, A.; Prakash, G.K.S. Beyond Oil and Gas: The Methanol Economy, 2nd ed.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, Germany, 2009; ISBN 98-3-527-32422-4.

31. Bandi, A.; Specht, M.; Weimer, T.; Schaber, K. CO2 recycling for hydrogen storage and transportation—Electrochemical CO2 removal and fixation. Energy Convers. Manag. 1995, 36, 899–902. [CrossRef]

32. Stucki, S.; Schuler, A.; Constantinescu, M. Coupled CO2 recovery from the atmosphere and water electrolysis: Feasibility of a new process for hydrogen storage. Int. J. Hydrog. Energy 1995, 20, 653–663. [CrossRef] 33. Weimer, T.; Schaber, K.; Specht, M.; Bandi, A. Methanol from atmospheric carbon dioxide: A liquid zero emission fuel for the future. Energy Convers. Manag. 1996, 37, 1351–1356. [CrossRef] 34. Sileghem, L.; Coppens, A.; Casier, B.; Vancoillie, J.; Verhelst, S. Performance and emissions of iso-stoichiometric ternary GEM blends on a production SI engine. Fuel 2014, 117, 286–293. [CrossRef] 35. Sileghem, L.; Ickes, A.; Wallner, T.; Verhelst, S. Experimental Investigation of a DISI Production Engine Fuelled with Methanol, Ethanol, Butanol and ISO-Stoichiometric Alcohol Blends. SAE Technical Paper 2015-01-0768. In Proceedings of the SAE 2015 World Congress, Detroit, MI, USA, 21–23 April 2015. [CrossRef] Energies 2020, 13, 6390 30 of 31

36. Shirazi, S.A.; Abdollahipoor, B.; Martinson, J.; Windom, B.; Foust, T.D.; Reardon, K.F. Effects of dual-alcohol gasoline blends on physiochemical properties and volatility behavior. Fuel 2019, 252, 542–552. [CrossRef] 37. Larsson, T.; Stenlaas, O.; Erlandsson, A. Future Fuels for DISI Engines: A Review on Oxygenated, Liquid Biofuels. SAE Technical Paper 2019-01-0036. In Proceedings of the SAE 2018 World Congress, Detroit, MI, USA, 10–12 April 2018. [CrossRef] 38. Shirazi, S.A.; Abdollahipoor, B.; Windom, B.; Reardon, K.F.; Foust, T.D. Effects of blending C3-C4 alcohols on motor gasoline properties and performance of spark ignition engines: A review. Fuel Process. Technol. 2020, 197, 106194. [CrossRef] 39. Bergthorson, J.M.; Thomson, M.J. A review of the combustion and emissions properties of advanced transportation biofuels and their impact on existing and future engines. Renew. Sustain. Energy Rev. 2015, 42, 1393–1417. [CrossRef] 40. Züttel, A.; Remhof, A.; Borgschulte, A.; Friedrichs, O. Hydrogen: The future energy carrier. Phil. Trans. R. Soc. Part A 2010, 368, 3329–3342. [CrossRef] 41. Turner, J.W.G.; Popplewell, A.; Patel, R.; Johnson, T.; Darnton, N.J.; Richardson, S.; Bredda, S.W.; Jackson, R.; Bithell, C.; Tudor, R.; et al. Ultra Boost for Economy: Extending the Limits of Extreme Engine Downsizing. SAE Technical Paper 2014-01-1185. In Proceedings of the SAE 2014 World Congress, Detroit, MI, USA, 8–10 April 2014; Volume 7. [CrossRef] 42. Lewis, A.G.J.; Akehurst, S.; Turner, J.W.G.; Patel, R.; Popplewell, A. Observations on the measurement and performance impact of catalyzed vs. non-catalyzed EGR on a heavily downsized DISI engine. SAE Technical Paper 2014-01-1196. In Proceedings of the SAE 2014 World Congress, Detroit, MI, USA, 8–10 April 2014; Volume 7. [CrossRef] 43. Remmert, S.M.; Cracknell, R.; Head, B.; Schuetze, A.; Lewis, A.G.J.; Akehurst, S.; Turner, J.W.G.; Popplewell, A. Octane Response in a Downsized, Highly Boosted Direct Injection Spark Ignition Engine. SAE Technical Paper 2014-01-1397. In Proceedings of the SAE 2014 World Congress, Detroit, MI, USA, 8–10 April 2014; Volume 7. [CrossRef] 44. Campbell, S.; Remmert, S.; Cracknell, R.; Schuetze, A.; Warnecke, W.; Lewis, A.; Akehurst, S.; Turner, J.W.G.; Patel, R.; Popplewell, A. Fuel Effects in a Downsized, Highly Boosted Direct Injection Spark Ignition Engine. In Proceedings of the 23rd Aachen Colloquium, Aachen, Germany, 6–8 October 2014. 45. Remmert, S.; Campbell, S.; Cracknell, R.; Schuetze, A.; Lewis, A.; Giles, K.; Akehurst, S.; Turner, J.W.G.; Popplewell, A.; Patel, R. Octane Appetite: The Relevance of a Lower Limit to the MON Specification in a Downsized, Highly Boosted DISI Engine. SAE Technical Paper 2014-01-2718. In Proceedings of the SAE 2014 Powertrain, Fuels and Lubricants Conference, Birmingham, UK, 20–23 October 2014; Volume 7. [CrossRef] 46. Leach, F.C.P.; Stone, C.R.; Richardson, D.; Lewis, A.G.J.; Akehurst, S.; Turner, J.W.G.; Remmert, S.; Campbell, S.; Cracknell, R.F. The effect of oxygenate fuels on PN emissions from a highly boosted GDI engine. Fuel 2018, 225, 277–286. [CrossRef] 47. Leach, F.C.P.; Stone, R.; Richardson, D.; Lewis, A.G.J.; Akehurst, S.; Turner, J.W.G.; Remmert, S.; Campbell, S.; Cracknell, R.F. Particulate emissions from a highly boosted gasoline direct injection engine. Int. J. Engine Res. 2018, 19, 347–359. [CrossRef] 48. Chupka, G.M.; Christensen, E.; Fouts, L.; Alleman, T.L.; Ratcliff, M.A.; McCormick, R.L. Heat of Vaporization Measurements for Ethanol Blends Up to 50 Volume Percent in Several Hydrocarbon Blendstocks and Implications for Knock in SI Engines. SAE Technical Paper 2015-01-0763. In Proceedings of the SAE 2015 World Congress, Detroit, MI, USA, 21–23 April 2015; Volume 8. [CrossRef] 49. Szybist, J.P.; Chakravathy, K.; Daw, C.S. Analysis of the Impact of Selected Fuel Thermochemical Properties on Internal Combustion Engine Efficiency. Energy Fuels 2012, 26, 2798–2810. [CrossRef] 50. Nguyen, K.-D.; Szybist, J.; Sileghem, L.; Verhelst, S. Effects of molar expansion ratio of fuels on engine efficiency. Fuel 2020, 263, 116743. [CrossRef] 51. Leach, F.C.P.; Stone, R.; Richardson, D.; Lewis, A.G.J.; Akehurst, S.; Turner, J.W.G.; Shankar, V.; Chahal, J.; Cracknell, R.F.; Aradi, A. The effect of fuel composition on particulate emissions from a highly boosted GDI engine—An evaluation of three particulate indices. Fuel 2019, 252, 598–611. [CrossRef] 52. Millwood, D.; Turner, J.W.G. Fuel Properties Calculator; Proprietary Software; University of Bath: Bath, UK, 2020. Energies 2020, 13, 6390 31 of 31

53. Sluder, C.S. Estimation of the Fuel Efficiency Potential of Six Gasoline Blendstocks Identified by the U.S. Department of Energy’s Co-Optimization of Fuels and Engines Program. SAE Technical Paper 2019-01-0017. In Proceedings of the SAE 2019 Powertrain, Fuels and Lubricants Conference, San Antonio, TX, USA, 22–24 January 2019. [CrossRef] 54. McCormick, R.L.; Fioroni, G.; Fouts, L.; Christensen, E.; Yanowitz, J.; Polikarpov, E.; Albrecht, K.; Gaspar, D.J.; Gladden, J.; George, A. Selection Criteria and Screening of Potential Biomass-Derived Streams as Fuel Blendstocks for Advanced Spark-Ignition Engines. SAE Int. J. Fuels Lubr. 2017, 10, 442–460. [CrossRef] 55. Vancoillie, J.; Verhelst, S.; Demuynck, J.; Sierens, R.; Rusu, E.; Pantile, V. Efficiency Comparison between Hydrogen, Methanol and Gasoline on a Production-Type Four-Cylinder Engine. In Proceedings of the 13th EAEC European Automotive Conference, Valencia, Spain, 14–16 June 2011. 56. Pearson, R.J.; Turner, J.W.G.; Peck, A.J. Gasoline-ethanol-methanol tri-fuel vehicle development and its role in expediting sustainable organic fuels for transport. In Proceedings of the IMechE Low Carbon Vehicles Conference, London, UK, 20–21 May 2009. 57. Naganuma, K.; Vancoillie, J.; Verhelst, S.; Sileghem, L.; Turner, J.W.G.; Pearson, R.J.; Martens, K. Drive cycle analysis of load control strategies for methanol fuelled ICE vehicle. SAE Technical Paper 2012-01-1606. In Proceedings of the SAE International Powertrains, Fuels and Lubricants Meeting, Malmo, Sweden, 18–20 September 2012. 58. Leach, F.C.P.; Richardson, D.; Stone, R. The Influence of Fuel Properties on Particulate Number Emissions from a Direct Injection Spark Ignition Engine. SAE Technical Paper 2013-01-1558. In Proceedings of the SAE 2013 World Congress, Detroit, MI, USA, 16–18 April 2013. [CrossRef] 59. Peckham, M.S.; Finch, A.; Campbell, B.; Price, P.; Davies, M.T. Study of Particle Number Emissions from a Turbocharged Gasoline Direct Injection (GDI) Engine Including Data from a Fast-Response Particle Size Spectrometer. SAE Technical Paper 2011-01-1224. In Proceedings of the SAE 2011 World Congress, Detroit, MI, USA, 12–14 April 2011. [CrossRef] 60. Chen, L.; Braisher, M.; Crossley, A.; Stone, R. The Influence of Ethanol Blends on Particulate Matter Emissions from Gasoline Direct Injection Engines. SAE Technical Paper 2010-01-0793. In Proceedings of the SAE 2010 World Congress, Detroit, MI, USA, 13–15 April 2010. [CrossRef] 61. Litzinger, T.; Stoner, M.; Hess, H.; Boehman, A. Effects of oxygenated blending compounds on emissions from a turbocharged direct injection diesel engine. Int. J. Engine Res. 2000, 1, 57–70. [CrossRef] 62. Davy, M.H.; Williams, P.A. The Effects of Flash Boiling on Mixture Formation in a Firing Direct-Injection Spark-Ignition (DISI) Engine. In Proceedings of the International Congress on Direkteinspritzung im Ottomotor II (Gasoline Direct Injection Engines), Munich, Germany, 16–17 November 1999; pp. 154–170. 63. Sandford, M.; Page, G.; Crawford, P. The All New AJV8. SAE Technical Paper 2009-01-1060. In Proceedings of the SAE 2009 World Congress, Detroit, MI, USA, 20–23 April 2009. [CrossRef] 64. Reavell, K.; Hands, T.; Collings, N. A Fast Response Particulate Spectrometer for Combustion Aerosols. SAE Technical Paper 2002-01-2714. In Proceedings of the SAE Powertrain & Fluid Systems Conference, San Diego, CA, USA, 21–24 October 2002. [CrossRef] 65. Andersson, J.; Giechaskiel, B.; Muñoz-Bueno, R.; Sandbach, E.; Dilara, P. Particle Measurement Programme (PMP) Light-Duty Inter-Laboratory Correlation Exercise (ILCE_LD) Final Report; European Commission Joint Research Centre Institute for Environment and Sustainability: Ispra, Italy, 2007.

Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).