You are on page 1of 29

Fuel 285 (2021) 119140

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Review article

2,5-Dimethylfuran (DMF) as a promising biofuel for the spark ignition


engine application: A comparative analysis and review
Anh Tuan Hoang a, *, Sandro Nižetić b, Aykut I. Ölçer c
a
Institute of Engineering, Ho Chi Minh City University of Technology (HUTECH), Ho Chi Minh City, Viet Nam
b
LTEF – Head of Laboratory for Thermodynamics and Energy Efficiency, Faculty of Electrical Engineering, Mechanical Engineering and Naval Architecture, University of
Split, Split, Croatia
c
Head of Maritime Energy Management Specialization, World Maritime University, Malmö, Sweden

A R T I C L E I N F O A B S T R A C T

Keywords: Being one of the furan-based compounds, 2,5-dimethylfuran (DMF) is known as a critical platform substance and
2,5-Dimethylfuran an ideal green solution on the pathway of finding alternative fuels to assuage the ever-increasing shortage of
Spark ignition engine fossil energy as well as to reduce the negative and dangerous impacts on global climate and environment.
Engine performance
Therefore, DMF could be considered as a potential promising biofuel for in-future engines since it was generated
Emission characteristics
Engine knocking
from renewable lignocellulosic biomass. In this review paper, the DMF synthesis process from current biomass
Wear through the catalyst-based reactions was thoroughly analyzed. Furthermore, the spray and flame characteristics
of DMF in comparison to commercial gasoline and ethanol were completely evaluated. More interestingly, the
performance, combustion, and emission characteristics of spark ignition engine running on gasoline, ethanol, and
DMF-based fuels were discussed in detail. Finally, the comparison of the effects of DMF properties on the SI
engine knocking as well as the characteristics of lubrication and wear were also performed. In general, DMF
could become a new-fashioned alternative fuel for spark ignition engines although the optimization strategies on
the DMF production process should be conducted before commercializing and realizing in the near future.

relating to the areas of arable land for crops for the ethanol production,
the lower heating value and Research Octane Number (RON) of ethanol
1. Introduction
compared to gasoline were considered as the barriers for ethanol
becoming an independent “worker” to replace commercial gasoline in
A harsh and undeniable truth is the diminishing worldwide reserve
the existing engine applications. Therefore, next-generation advanced
of fossil energy sources and the rapid increase in global environmental
biofuels without ethanol could be considered as a potential candidate for
pollution [1], resulting in unpredicted and severe consequences for the
the pressing need in near future. A promising approach pointed out is to
future of our next generation [2,3]. Facing these serious problems, the
produce furan-based chemicals from cellulosic biomass like 2,5-dime­
search for new-fashioned energy sources with low cost, environmentally
thylfuran (DMF) [9–11]. Being one of the above-mentioned products
friendly and high safety properties has become the topical issue of in­
derived from furan, DMF is paid significant attention to due to its
terests on the way towards targeting sustainable development and
outstanding advantages, in which DMF could be produced from the
happier life [4,5]. Looking at many decades of research on the alterna­
carbohydrates such as glucose, fructose, starch-containing sources, and
tive energy sources, biofuels have been emerging as a potential candi­
even cellulose [12]. Compared to ethanol, DMF possesses higher energy
date [6]. However, a question was given out that which feedstock would
density, higher RON, greater boiling temperature, and lower aqueous
be suitable for biofuel production without any competitions? Gradually,
solubility [13,14]. More interestingly, blending DMF with commercial
the answer has been revealed as biomass has shown its outstanding
gasoline is also easier than that of bioethanol as well as DMF is immis­
characteristics compared to other feedstock for bio-fuel production.
cible with water [15]. Most studies have demonstrated that no need for
Therefore, the conversion of renewable biomass into liquid biofuels has
much modification for both spark and diesel engines while these engines
been considered as a “hotspot” research field in current years [7].
ran on DMF. These physicochemical properties of DMF have inspired
Indeed, ethanol is currently known as the only commercial biofuel
several researchers to explore highly efficient sustainable and advanced
satisfying renewable fuel requirements [8]. Nonetheless, some issues

* Corresponding author.
E-mail address: hatuan@hutech.edu.vn (A.T. Hoang).

https://doi.org/10.1016/j.fuel.2020.119140
Received 7 June 2020; Received in revised form 26 August 2020; Accepted 28 August 2020
Available online 28 September 2020
0016-2361/© 2020 Elsevier Ltd. All rights reserved.
A.T. Hoang et al. Fuel 285 (2021) 119140

Nomenclature IT Injection timing


ITE Indicated thermal efficiency
ATDC After top dead center Lb Markstein length
BTDC Before top dead center LHV Lower heating value
CA Crank angle LBV Laminar burning velocities
CMF 5-Chloromethylfurfural MAPi Intake manifold absolute pressures
CO Carbon monoxide MBT Maximum Brake Torque
COV Coefficient of variation MFB Mass fraction burned
DI Direct injection NOx Nitrogen oxides
DMF 2,5-Dimethylfuran RON Research octane number
EGT Exhaust gas temperature OR Octane rating
ER Evaporation rate SFC Specific fuel consumption
EOI End of injection PAH Polycyclic aromatic hydrocarbons
FC Friction coefficient PFI Port fuel injection
GDI Gasoline direct injection PM Particulate matter
HC Unburned hydrocarbons Pex_max Maximum in-cylinder pressure
HDO Hydrodeoxygenation PN Particle number
HMF 5-Hydroxymethylfurfural SI Spark ignition
HRR Heat release rate SMD Sauter Mean Diameter
HV Heat of vaporization SOC Start of combustion
ICE Internal combustion engine SOI Start of injection
ID Ignition delay Tex_max Maximum in-cylinder temperature
IMEP Indicated mean effective pressure THF Tetrahydrofuran
ISFCE Indicated specific fuel consumption equivalent UFPS Unstretched flame propagation speed

production technologies for DMF with the primary purpose of utilizing it thoroughly studied the synthesis processes, mechanical characteristics
as a safe liquid-fuel for the transportation sector. and features, suitable solvent and catalyst type for DMF processing
The leading efforts in producing DMF were suggested by many re­ [25–27], DMF’s physicochemical mechanism and financial performance
searchers [16–18], they introduced an approach for synthesizing DMF [28]. Furthermore, recent researches showed DMF as a promising fuel
based on catalytic from fructose that could be directly gained from used in internal combustion engines (ICE) without modifying engine
cellulosic biomass or via the glucose-to-fructose isomerization process, structure. Also, several studies on ICE indicated better knock resistance
which is a widely-distributed monosaccharide in nature and presented for the DMF combustion [29], as well as lower unburned hydrocarbons
in Fig. 1. Moreover, DMF is found as a source of sustainable oxygenated (HC), nitrogen oxides (NOx), and particulate matter (PM) compared to
liquid-fuel with insoluble characteristic in water [19,20]. Therefore, it is gasoline and ethanol [30,31]. Obviously, the recent studies have clari­
becoming an inevitable task to produce monosaccharides as well as fied the potential in using DMF as a novel alternative fuel in SI engine
polysaccharides into DMF with high performance to boost environ­ when some attributes of DMF were reported to be better than ethanol as
mental issues and maximize the energy-resource structure [21,22]. well as DMF could be produced from renewable biomass feedstock.
It could be seen from Fig. 1 that the development of DMF and However, the comparison of the applicability of DMF with ethanol is
traditional derivatives through the biomass conversion pathway have necessary since ethanol has become the commercial fuel in many
the mass production ability, it was recorded a process to transform the countries for many years. The main aim of the comparison is to highlight
plentiful raw material lignocellulosic biomass to DMF [24], the pro­ the strengths and weaknesses of DMF compared to ethanol, aiming to
duction technique of DMF in recent progress aiming to make up DMF come up the suggestion and proposal relating to the application of DMF
from biomass has achieved high efficiency. Several researchers have in the near future. In this review paper, the conversion process of

Fig. 1. DMF production pathway from biomass [23].

2
A.T. Hoang et al. Fuel 285 (2021) 119140

carbohydrates-based materials into DMF was reviewed aiming to build a HMF HDO process because Lewis acid sites are easy to be formed,
development premise of advanced catalysts for the strategies of increasing the activated degree of unsaturated C-heteroatom bonds
simplifying technology and production process. More importantly, the through the formation of adducts or via σ-coordination [35]. Indeed, the
spray and flame characteristics of DMF were thoroughly analyzed and chemical reactions based on the mechanism of the carbon-heteroatom
compared to those of gasoline and ethanol. Finally, the performance, bond, or transition metal-catalyzed carbon–carbon bond, or transition
combustion and emission features of spark ignition engines running on metal-mediated bond are considered as the most powerful tools in
DMF, gasoline, and ethanol were fully evaluated and compared. synthesizing the organic compounds [36]. As a result, the C-O bond is
broken, increasing the dehydration of C-OH groups in fructose or
2. DMF production and properties glucose to form HMF, resulting in higher DMF selective conversion ef­
ficiency. In 2007, Román-Leshkov and co-workers have presented their
2.1. DMF production effort in the DMF production [37], in which they selected biomass as the
input material to produce fructose that was considered as the interme­
To improve the conversion efficiency of DMF, and to minimize the diate substance for DMF synthesis by using the catalyst. As depicted in
byproducts formation, the catalyst is usually put into the reaction pro­ Fig. 2, HMF was synthesized from fructose based on the selective
cess with a reasonable concentration, which is optimized to reach the dehydration process under the support of a biphasic reactor with the
best efficiency. Therefore, the catalyst systems are believed to play a presence of HCI as a catalyzer and a solution of NaCl 35%. After that, the
central part in the DMF production from various input material sources evaporation process of HCl and water has occurred from the solvent
such as cellulosic/lignocellulosic biomass, glucose, fructose, 5-hydroxy­ containing HMF. Extracted HMF was converted into DMF by the bime­
methylfurfural (HMF), and 5-chloromethylfurfural (CMF). Nonetheless, tallic catalyst of Cu/Ru supported by carbon. As a result, the conversion
the DMF synthesis from cellulosic/lignocellulosic biomass, glucose, and efficiency of DMF was 62% to 70% without byproducts. More impor­
fructose must undergo the intermediate steps, which are to form HMF or tantly, the authors have detected that it could use glucose for large-scale
CMF. Finally, HMF or CMF is considered as a precursor to synthesize production of DMF while fructose has provided higher DMF selectivity.
DMF through the catalyst pathway [32]. Due to this result, the DMF By inheriting the previous study, recent studies have been focused on
synthesis from HMF or CMF is usually more preferable and has higher the catalyst development to improve the selective conversion efficiency
selective conversion efficiency than from cellulosic/lignocellulosic of DMF. In the study of Thananatthanachon et al. [19], they have uti­
biomass, glucose, and fructose. Indeed, there have been many catalytic lized formic acid (HCOOH) as the H2 donor to synthesize DMF from
types developed for the conversion process of HMF into DMF through fructose through the one-pot process. Achieved results have confirmed
the selective hydrodeoxygenation (HDO) [33]. Normally, the catalytic that HCOOH could play a catalyst role in the dehydration of fructose to
systems based on single transition metals (like Pt, Ru, Mo, Cu) or HMF. More interestingly, an excellent yield (>95%) of DMF was
bimetallic catalysts (like Ni-Co, Cu-Ru, Pt-Au) were found to be potential attained in case of heating HMF in refluxing tetrahydrofuran (THF) with
candidates to achieve high DMF yield although catalyst systems origi­ the support of HCOOH, H2SO4 and Pd/C-based catalyst system. With the
nated from noble metals and bimetal were proven to provide better same input source, Li et al. [38] have performed the one-step DMF
catalytic characteristics and stability [34]. Moreover, bifunctional production from fructose under the biphasic solvent of [BMIM]Cl/THF
catalyst systems could also be applied to the DMF production since these and carbon-supported Ru-based catalyst. However, only 66% of DMF
catalyst systems contain both the deoxygenating component and the yield could be obtained. In the case of using glucose for DMF production,
hydrogenation metal. For metal/acid-based catalyst systems, the syn­ Chidambaram et al. [39] have explored a novel catalyst based on
ergistic effect of acid-metal interaction is thought to be the key for the phosphor–molybdic acid with high conversion efficiency by a two-step

Fig. 2. The pathway of DMF production from various feedstock under carbon-supported Cu/Ru-bimetallic catalyst system [37].

3
A.T. Hoang et al. Fuel 285 (2021) 119140

approach in the ionic liquid. In the first step, around 99% of HMF was bimetallic catalyst based on Cu-Co. Unbelievably, around 99% of DMF
formed from glucose. After that, the HMF HDO process with the pres­ yield could be reached under 180 ◦ C of temperature. More interestingly,
ence of Pd/C catalyst and the acetonitrile-based additive has created the use of Cu-Co catalyst (Cu:Co = 1:3) supported by C has provided
DMF with 44–47% selective efficiency. In a recent study of Insyani et al. nearly 100% of DMF yield [54]. In study of Nilges et al. [55], electro­
[40], they have used a multifunctional Pd/UiO-66@SGO catalyst aiming chemical method was first introduced to convert HMF to DMF by using
to activate the dehydration in HMF formation from glucose for the one- Cu electrodes. This process was found to undergo a series of reduction
pot production of DMF. The achieved result has shown that a high yield steps of electron/proton for the conversion of HMF to DMF. As a result,
of 45.3 mol% for DMF production from glucose, this result was reported the highest DMF selective conversion efficiency was 35.6% as catalyst
to be much lower than 70.5% DMF yield from fructose under the same reaction was carried out in H2SO4 solution with the presence of water
catalyst condition. The combination of Lewis acid (CrCl3) catalyst and and ethanol. Taken as a whole, the production process of DMF from
Brønsted acid (HCl) catalyst for the HMF production from glucose has various precursors such as could be presented in Table 1.
been confirmed to bring higher efficiency [41]. It was detected that
Lewis CrCl3 acid catalyst has activated the isomerization reaction of 2.2. DMF properties
glucose to fructose, while HCl-supported Brønsted acid catalyst has the
obvious effects in dehydration/rehydration reactions, producing around As analyzed above, the hydrogenolysis of biomass-derived HMF or
59% HMF yield. Furthermore, HMF could be produced by lignocellu­ direct conversion of carbohydrates to DMF is considered as a core re­
losic biomass by combinational treatments, in which acid/base was used action in the up-gradation of platform compounds originated from car­
for the pretreatment process and catalyst was used for HMF synthesis bohydrates for extracting potential biofuels, in which DMF is known as
stage. Nguyen et al. [42] have utilized wood chips for HMF production one of the promising liquid fuels used for the transportation means. In
through the pretreatment with a diluted solution of 3% NaOH and the addition, DMF is demonstrated that only one-third of the energy is
catalyst reaction with CrCl3 in ([BMIM]Cl) solvent at a temperature of consumed in the evaporation stage. Blending DMF with commercial
120 ◦ C. After 1 h of reaction time, high conversion efficiency up to 79% gasoline is also easier than that of bioethanol as well as DMF is immis­
of HMF yield was observed. Similarly, Binder et al. [43] have used raw cible with water. In an experimental study on the use of DMF for a direct-
corn straw to synthesize DMF through the HMF intermediate substance injection (DI) spark-ignition (SI) engine, it has been indicated that the
by using C-supported Cu/Ru in the solvent of butan-1-ol. Not as ex­ performance of the engine fueled with DMF was comparable to gasoline.
pected, the HMF yield from corn straw was quite low, only 48% Such excellent properties make DMF become a more ideal alternative
compared to 92% from fructose and 81% from glucose. In general, the fuel in the near future. In brief, DMF properties in comparison with
HMF production from raw lignocellulosic biomass was found to provide ethanol and gasoline were in Table 2.
a quite low yield, even lower than 5% in some works [44,45]. Indeed, all It is clearly seen from Table 2 that lower volatility and comparable
production processes of DMF from carbohydrates undergo intermediate molecular mass compared to gasoline make DMF become a new green
substances such as HMF or CMF. Not surprisingly, it is affirmed that the fuel with high value. Higher RON of DMF (RON = 119) compared to
conversion efficiency of DMF from HMF or CMF is higher than other commercial gasoline (RON = 90–100) and ethanol (RON = 110) will
paths. Similar to other catalyst reactions, the catalyst systems in the allow DMF to use in spark-ignition engines with higher compression
DMF synthesis process play a key role in the conversion efficiency as ratios aiming to improve the fuel economy. More attractively, DMF’s
well as DMF yield. Wang et al. [46] have applied hollow carbon spheres- energy density is higher than that of ethanol, leading to more heat en­
supported PtCo@HCS catalyst to the hydrogenolysis reaction of HMF ergy released in the combustion. The boiling point of DMF (93 ◦ C) is
under the optimized condition such as 180 ◦ C of temperature and 10 bar higher than that of ethanol (77 ◦ C), resulting in DMF having proper
of H2 pressure. Surprisingly, the DMF yield was very high, up to 98%. gasification efficiency, which is beneficial to impairing engine inlet air
Also, another experiment using Pt-Co/MWCNTs as bimetallic catalyst resistance and achieving engine start-up performance at low tempera­
has shown a high DMF yield, corresponding to 92.3% under 160 ◦ C of tures. Moreover, latent heat vaporization of DMF is insignificantly
temperature and 10 bar of H2 pressure [47]. In addition, some catalyst higher in comparison with gasoline but it is much lower than ethanol
types based on Ru such as commercial Ru/C, Ru/C, and Ru/CNTs fuels, this is useful for DMF storage and transport. Another positive
(carbon nanotubes-supported Ru) were detected to enhance the selec­ aspect, the kinematic viscosity of DMF at the same test temperature is
tive hydrogenolysis of HMF. As a result, the DMF yield could be attained lower than that of ethanol, but insignificantly higher than that of gas­
as high as 94.7% for commercial Ru/C catalyst [48], 81% for Ru/C oline, leading to the helpfulness in ensuring the injection pressure as
catalyst [49], and 83.5% for Ru/CNTs catalyst [50]. With another noble well as protecting the fuel supply system. The above-mentioned proof is
metal, Nishimura et al. [51] have utilized the NaBH4 reduction method sufficient to demonstrate that DMF has enough physicochemical prop­
to prepare the catalyst based on Pd-Au/C for the conversion of HMF into erties to make it be a credible alternative fuel for combustion engines.
DMF in the solvent of HCl. An excellent DMF yield was reported, up to Moreover, it is also worth believing that these qualities bode well for the
96% after 12 h of reaction time. Although non-noble metals were nor­ safe transport and storage of DMF when considering as a new-fashioned
mally reported to catalyze the HMF conversion to DMF with a lower liquid fuel in the transportation field.
yield than with noble metals, they also showed remarkable results in
promoting the HMF HDO. In some case, they even provided high yields. 3. Spray and flame characteristics
For instance, Huang et al. [52] have explored a catalyst of 7Ni-30W2C
supported by active carbon for the DMF synthesis from HMF. A quite 3.1. Spray characteristics
high DMF yield of 96% could be observed as conducted at 180 ◦ C of
temperature and 4 MPa of H2 pressure. Alternatively, Ni was found to In addition to the excellent properties of DMF compared to ethanol or
have a key role in the hydrogenation of HMF, while W2C has offered a commercial gasoline given in Table 2, DMF must possess other specific
good ability in the deoxygenation process. The synergistic effect of Ni physical properties if DMF is employed as an alternative fuel for com­
and W2C has brought such a high DMF yield. Obviously, the bimetallic mercial engines. Therefore, another issue should be mentioned as the
catalyst systems have shown a much higher yield of DMF synthesis than effects of DMF properties on the spray characteristics because the fuel
with a single catalyst system. This was attributed to the synergistic effect spray characteristics are known to play a vital part in the characteristics
of metals joining in the reaction. In addition to Ru that was considered as of combustion and emissions, and engine performance, they are espe­
a promoter for Cu-supported catalyst system, Co was found to be a po­ cially important for direct injection (DI) engines. In general, some pa­
tential candidate for enhancing the DMF yield. Indeed, Guo et al. [53] rameters characterizing the fuel spray characteristics could include
have prepared a modified N-graphene-supported nanoparticle spray penetration, cone injection angle, droplet size, and velocity

4
A.T. Hoang et al. Fuel 285 (2021) 119140

Table 1
Introduced catalyst systems for high-yield DMF synthesis from various feedstock.
Feedstock Catalyst Solvent Condition Time, h Yield, % Ref

HMF Ru/Co3O4 THF T = 403K 24 93.4 [56]


p = 7 bar
RuCo/CoOx 1,4-dioxane T = 473K 2 96.5 [57]
p = 5 bar
Ru/CoFe-LDO THF T = 453K 6 98.2 [58]
p = 10 bar
Ru/Cu DMSO T = 423 K 0.3 91.0 [59]
Ni/Al2O3 1,4-dioxane T = 453 K 3 91.5 [60]
Ni-Co/C THF T = 483 K 24 90.0 [61]
NiZnAl 1,4-dioxane T = 453 K 15 93.6 [62]
Pt3Ni/C 1-propanol T = 473K – 98.0 [63]
p = 33 bar
Pd/C H2O T = 353K 2 100 [64]
p = 10 bar
Ni-OMD3 H2O T = 473K 6 98.7 [65]
p = 30 bar
Ni/perovskite ethanol T = 503 6 98.3 [66]
p = 50 bar
Ni2-Fe/CNTs n-butanol T = 473K 3 91.3 [67]
p = 30 bar
NiCu3/C 1-propanol T = 473K 3 98.7 [68]
p = 33 bar
NC-Cu/MgAlO cyclohexanol T = 493 K 0.5 96.1 [69]
CnZn-2 1,4-dioxane T = 493 K 5 91.8 [70]
p = 15 bar
Glucose 4.8Pd/UiO-66 @SGO THF T = 433 K 3 45.3 [40]
Fructose Cu-Ru/C HCl/KCl/1-butanol T = 343 K 10 73.0 [37]
Pd/C THF T = 343 K 15 51.0 [19]
Ru/C [BMIM]Cl/THF T = 403 K 0.5 50.0 [38]
2.4Pd/UiO-66@SGO THF T = 433 K 3 68.6 [40]
HT-Cu/ZnO/Al2O3 GBL/H2O T = 513 K – 40.6 [62]
Amberlyst15/Ni@WC FA-ethanol T = 383 K – 38.5 [71]

could be considered as the result of the secondary breakup after the


Table 2
downstream of the spray has reached around 24 mm. In the gasoline
Physicochemical properties of DMF in comparison to those of gasoline and
case, it could be seen a slight rise in SMD along the axis, while SMD of
ethanol [23,72–75].
D50 (50% DMF + 50% gasoline) and DMF were greater than for gaso­
Physicochemical properties at Unit DMF Ethanol Gasoline
line. Nonetheless, no remarkable difference was seen in terms of SMD
20 ◦ C
between DMF, D50, and gasoline with a distance of around 40 mm away
Molecular formula – C6H8O C2H6 C4–C12 from the injector. More importantly, they found that ethanol was much
Oxygen content % 16.67 34.78 0
different from DMF, D50, and gasoline. The smallest SMD for ethanol
Stoichiometric air/fuel ratio – 10.72 8.95 14.70
Latent heat of vaporization kJ/kg 389 920 351 near the nozzle could be self-evident and it became the highest as 40 mm
Energy density MJ/L 30 18.4–21.2 32–34.8 away from the injector. Indeed, the SMD variation in the range of 16.9
Boiling point (BP) ◦
C 93 77 27–225 μm to 20.2 μm along the axis in the ethanol case was found in com­
Auto-ignition temperature (AIT) C 286 434 420
parison with 16–22.5 μm for gasoline, 14.9–23.4 μm for DMF, and

Research octane number (RON) – 119 110 90–100


Liquid density (ρ) kg/ 890 791 745
14.8–22.3 μm for D50. This difference in SMD has resulted from the
m3 difference in triad physical properties of fuels such as viscosity, density,
Surface tension (σ) N/m 0.0259 0.0223 0.0200 and surface tension, which could be clearly seen in Table 2 and it was
Kinematic viscosity (µ) cSt 0.57 1.50 0.37–0.44 also demonstrated in some published works on the effects of triad
physical properties of fuels on spray characteristics [78,79]. Moreover,
distribution [76]. Spray penetration and cone injection angle depended they detected that ethanol possessed the shortest penetration length in
much on the injector design are the main parameters affecting signifi­ the primary breakup process, resulting in more break-ups occurring for
cantly the shape of the fuel spray. Meanwhile, fuel droplet size and fuel the ethanol case than for gasoline, DMF, and D50. Besides, they pointed
velocity distribution play the leading role in the breakup of liquid fuel, out that the increase in injection pressure has reduced the SMDs of DMF
the atomization, mixture formation, which have direct influences on the from 8.8 μm at the pressure of 50 bar to 7 μm at the pressure of 150 bar,
evaporation and combustion characteristics. As to assess DMF’s spray while the increase in injection pressure was more beneficial for DMF’s
behavior, there were several investigations using various techniques. spray behavior than that of ethanol.
For instance, Tian et al. [77] have used Phase Doppler Particle Analyzer In the study of Li et al. [80], they have employed the Cascade At­
to measure the main parameters of DMF’s spray characteristics under omization Breakup model to predict the droplet sizes of DMF as well as
various pressures in comparison to those of ethanol and gasoline. As a to investigate the preparation of DMF-air mixture in the combustion
result, the mean velocity of 100% DMF and blends of gasoline with DMF chamber of gasoline direct injection (GDI) engines. They reported the
was found to be slightly lower than that of 100% gasoline, although the relatively poor distribution of DMF-air because the mixing time was
difference of mean velocity is only<10 m/s for most cases (Fig. 3a). The insufficient along with spray-wall interaction was significant. Moreover,
Sauter Mean Diameter (SMD) of DMF and its blend in comparison with when DMF was injected at the end of the compression stroke in GDI
gasoline and ethanol along the axis were plotted in Fig. 3b. For four fuel engines, DMF has contributed to forming an extra-rich mixture with a
samples, the fuel-droplet size was rapidly reduced just after they were ratio of around 9.3%, and very-lean mixture with a ratio of around 4.8%
injected out. Fuel-droplet sizes were observed to become stable, which in comparison with the nearly-homogeneous mixture formed from

5
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 3. Spray characteristics of DMF and its bends compared to other fuels; (a) – Velocity of DMF droplets compared to gasoline, ethanol, and DMF-gasoline blends
from the nozzle [77]; (b) – SMD distribution [77]; (c) – Total volume for DMF/air and gasoline/air mixture under various equivalence ratios [80]; (d) – Rate of
evaporation for DMF compared to gasoline [80]; (e) – DMF spray images at 100 bar of injection pressure [77]; (f) – The spray shape of various fuel blends [81].

6
A.T. Hoang et al. Fuel 285 (2021) 119140

gasoline (extra-rich mixture with a ratio of around 0.5% and very-lean was corresponding to the ratio of the very-lean mixture. In the event of
mixture with a ratio of around 2.3%). It was also found around the an extra-rich mixture (case 3), its percentage was increased from SOI to
spark plug that the equivalence ratio was leaner than 0.9. More the end of injection (EOI); ER was then steadily reduced in case of
impressively, they have evaluated the effects of the spray characteristics gasoline but ER had an insignificant change for DMF case. More
of DMF on the quality of DMF-air mixture compared to gasoline under remarkably, a slow increase in the portion of the case 2 was found, while
various equivalence ratios (Φ) (case 1 with Φ < 0.5; case 2 with 0.5 < Φ a slight rise in the portion of extra-rich mixture (case 3) was also re­
< 1.5; case 3 with Φ > 1.5) through the temporal evolution (measured ported for DMF after 210◦ after top dead center (ATDC). Furthermore,
by crankcase angle) of the local mixture (Fig. 3c). The measurement the comparison between DMF and gasoline based on global fuel evap­
based on instantaneous cylinder volume was used to normalize the oration rate could be observed in Fig. 3d, in which the total amount of
evaporation rate (ER) at each position of the crankcase angle. It could be injected fuel was used to normalize the vapor fuel. During the injection
seen from Fig. 3c that case 2, corresponding to the ratio of ignitable period, the vaporization rate of gasoline was seen to be faster compared
mixture, was increased after the start of injection (SOI), while the to the DMF case, this result was because of smaller SMD of gasoline and
opposite trend of the ignitable mixture could be observed in case 1 that the difference in vapor pressure of DMF and gasoline [20,82]. More

Fig. 4. Combustion parameters for DMF compared to gasoline and ethanol; (a) – Flame images at engine loads of 3 bar and 4 bar IMEP versus crank case angle,
collected from Xu et al. [86]; (b) – Effects of equivalent ratio on laminar burning velocity [89]; (c) – Effects of initial temperature on LBV of various fuels [90–93]; (d)
– Markstein length at 70 ◦ C of initial temperature [89].

7
A.T. Hoang et al. Fuel 285 (2021) 119140

interestingly, they found that around 90% of fuel evaporated at also reported in study of Wu et al. [90]. Especially, LBV of gasoline was
165◦ ATDC for the gasoline case and at 275◦ ATDC for the DMF case, significantly higher that of DMF in the case of the equivalence ratio >1.2
indicating that the DMF spray has delayed around 90◦ CA to achieve this [89]. The effects of initial temperature on LBV of DMF at various
vaporization percentage. Furthermore, around 96% of DMF and 98.5% equivalent ratios compared to other fuels (such as gasoline, ethanol, and
of gasoline were injected by 340◦ ATDC, implying that it would need methyl furan) could be illustrated in Fig. 4c. In addition to the effects of
more time for DMF to turn into complete vapor before the start of equivalent ratio and initial temperature on LBV, diluent concentration,
combustion (SOC) [80]. For penetration length illustrated in Fig. 3e, it and initial pressure were found to be the parameters affecting LBV;
could be seen the evolution of DMF’ penetration length in the correla­ however, it was found that the effect of initial temperature on LBV was
tion with the time at 100 bar of injection pressure [77]. Compared to more considerably [95]. The changing trends of Markstein length (Lb)
results reported in the literature, the difference in penetration between were also illustrated in Fig. 4d, where Lb > 0 presented the decrease of
DMF, gasoline, and ethanol was around 10%, this was agreed with the flame speed with increasing the stretch rate [96]. In contrast, Lb <
finding in the study of Atsushi et al. [83] and Paolo et al. [84]. Moreover, 0 depicted the rise of flame speed with increasing the stretch rate.
an insignificant difference in penetration length between DMF/iso-oc­ Bradley et al. [97] have disclosed that the initial flame could be stable
tane blends and DMF could also be observed in Fig. 3f, even with a multi- until the flame radius reached the critical value if Lb > 1.5. Moreover, it
hole injector of DI spark engine [81]. Absolutely, these results were could be seen from Fig. 4d that the reduction of Lb for DMF, ethanol, and
suitable for those given in Table 2, where molecular weight, liquid gasoline has occurred along with the increase in equivalence ratio. This
density, and surface tension of DMF are considerably higher than that of was due to rapid decay, leading to lower Lb at higher equivalence ratios.
gasoline and ethanol In general, the stability of the ethanol flame was the highest among
DMF, ethanol and gasoline because of the highest of Lb value. None­
3.2. Flame characteristics theless, in the DMF case, the flame was observed to be always below the
ethanol’ and gasoline’s flames, suggesting that the flame of DMF was not
Being a novel chemical, the study on DMF combustion characteristics as stable as that of ethanol and gasoline.
is very important. Among global parameters characterizing for com­
bustion characteristics, laminar burning velocities (LBV) and 3.3. Combustion duration and ignition delay
unstretched flame propagation speed (UFPS) were thought to represent
combustion characteristics and to validate the kinetic chemical mech­ Normally, the combustion duration was considered as the interval
anism. Clearly, initial parameters such as temperature, pressure, between the spark timing and the 5% mass fraction burned (MFB) and
equivalence ratio, and mixture composition showed significant effects was measured by crank angle degrees (CA), in which MFB could be
on LBV due to the multi-lateral combined influences like reaction calculated by the analysis of heat release and using the standard method
mechanism, or physicochemical properties. Being a key indicator [98]. The variation in initial combustion duration could help to clarify
affecting directly the turbulent simulations, LBV was known as the the changing tendency of the maximum pressure in the cylinder. Indeed,
burning rate when applied to ICE. Most of the investigations have lower combustion duration would result in higher combustion pressures
concentrated on the use of combustion bombs under isochoric condi­ and the in-cylinder pressure rise could occur earlier in the combustion
tions or the heat flux-based method. For flame images and burning ve­ stroke. In the study of Daniel et al. [88], they reported that the initial
locities of DMF compared to ethanol and gasoline under various combustion duration for ethanol was 0.9◦ CA lower compared to that of
conditions, it could be observed in Fig. 4a and b, respectively. Fig. 4a DMF, this was due to higher LBV and faster burning rate of ethanol than
showed the flame images of DMF compared to those of ethanol and for DMF and gasoline [99,100]. It could be seen from Fig. 5a that the
gasoline various engine loads of 3 bar and 4 bar IMEP [85]. It could be combustion durations of ethanol, DMF, and gasoline in the case of CA50
seen that the flame images of ethanol (ETH) were increased to twice in (10–50% MFB) were higher than the case of (50–90% MFB). Obviously,
comparison to the original intensity, this result was because the flame the combustion duration for DMF was consistently shorted than for
luminance signal of DMF and gasoline were seen to be much stronger ethanol, this was attributed to the higher volatility and oxygen content,
than that of ethanol. Indeed, a considerably brighter flame of DMF in and higher cetane number of ethanol compared to DMF [101]. In cases
comparison to gasoline could be observed based on the image sequence of 10–50% MFB and 50–90% MFB, the combustion durations of DMF
in Fig. 4a. This was absolutely matched with previous research, which were 0.12◦ CA and 0.83◦ CA lower than those of ethanol. Moreover, the
declared a shorter combustion duration of DMF compared to that of decrease of combustion duration in the DMF case after the CA50 location
gasoline and ethanol [86–88]. Moreover, the flame images of ethanol was found to be 7% and 3% higher compared to ethanol and gasoline,
and gasoline have revealed that ethanol had the largest flame areas respectively, indicating higher burning speed for DMF. Achieved results
caused by the highest laminar burning velocity shown in Fig. 4b. have highlighted the advantage of the burning speed for DMF in com­
It is evident that ethanol exhibited the highest LBV among ethanol, parison with ethanol, although the LBV of the DMF combustion process
gasoline, and DMF at all initial temperatures. At 323 K, the peak of LBV was lower than ethanol and gasoline [89].
was 56 cm/s for ethanol compared to 43 cm/s for gasoline. In the case of Ignition delay (ID) is attributed to be a very important factor rep­
DMF and ethanol, the increase in maximum LBV from 40 cm/s to 49 cm/ resenting the combustion process in the engines. It is a preparatory
s for DMF, and from 56 cm/s to 64 cm/s for ethanol were observed as the duration since the fuels have been injected into the engine cylinder but
rise of initial temperature was up to 50 ◦ C [89]. This rise of LBV could the ignition process of fuels has not been initiated yet. In fact, the engine
exceed this result through the suggestion of Wu et al. [94], who declared design and performance are exerted a very great impact by ID, which
that the LBV of DMF reached 55 cm/s at an initial temperature of 120 ◦ C was found to be various with different fuels used. Functionally, ID could
and equivalence ratio of 1.2. Achieved results have demonstrated a great be classified into physical ID and chemical ID, in which physical ID is
potential of ethanol in fast burning combustion relating to LBV as well as considered as the duration between the injection start and the time point
very high burning magnitude, reporting that the difference of LBV be­ of attaining chemical reactions. During physical ID, it occurs the prep­
tween ethanol and DMF in the engine application with the range of aration process for fuels such as the atomization, mixing with in-cylinder
equivalence ratio from 0.9 to 1.2 was approximately 30%. At 393 K, the air, raising to self-ignition temperature, which depends much on the fuel
similarity of LBV of DMF compared to LBV of gasoline were observed at properties like viscosity, density. In addition to physical ID, chemical ID
lower equivalence ratios, although a distinct difference around 10% is found to be larger compared to the physical ID. Indeed, chemical ID
between LBV of DMF and LBV of gasoline could be recognized as the has a great dependence on the characteristics of the surroundings such
equivalence ratio was changed from 0.9 to 1.1, which was known as the as temperature, pressure, fuel components. Typically, ignition analysis
typical operating conditions of gasoline engines. Similar results were based on the sensitivity was performed to explore the critical reactions

8
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 5. (a) – Combustion duration of CA50 at 8.5 bar IMEP and KL-MBT timing [88]; (b) – Rate-of-product analysis for the DMF consumption (Left) and sensitivity
analysis of DMF ignition delay (Right) and (c) – Comparison of ignition delay between DMF, ethanol, and gasoline [102].

in the DMF ignition. In the study of Tanaka et al. [102], the OH addition C6H7O with HO2 radical and the decomposition reactions of H2O2 to
reaction to the double-bond carbon in DMF (C6H8O) molecular aiming produce OH radicals were also detected to be critical for DMF ignition
to form C6H9O2 (dmf252OH3j) was found to be the main pathway in (Fig. 5b (right)). In contrast, the reactions consuming C6H7O and the
DMF consumption, this could be seen in Fig. 5b (left). Moreover, the formation reactions of H2O2 were found to have a negative sensitivity
formation of C6H7O (dmf252j) radicals through the H-abstraction re­ for DMF ignition, indicating the negative effect on the ignition of DMF.
actions from DMF, and the reaction of resonance-stabilized radical In general, the ignition delay time of DMF was between ethanol and

9
A.T. Hoang et al. Fuel 285 (2021) 119140

gasoline (Fig. 5c), this result was also reported in references [103,104]. IMEP for DMF lesser than that of gasoline, while the spark timing
sensitivity of ethanol was the least. Due to this reason, DMF was thus
4. Engine performance and emissions concluded to have a tuning window for spark timing sensitivity larger
than gasoline. As a result, the factors characterizing the combustion
There is no doubt that DMF is suitable for existing commercial en­ sensitivity such as the combustion phase and combustion stability was
gines running on gasoline. To admit this identification, the experimental ordered as gasoline > DMF > ethanol [108]. Indeed, the feasibility of
and simulation studies must be further carried out on the real engines, in employing DMF as an alternative fuel in SI engines was highlighted
which the engine performance is known as the key parameter in all tests through the direct comparison of the engine performance running on
related to the use of fuels in engines. In this section, the engine perfor­ DMF, gasoline, and ethanol [109]. All experiments were tested under
mance such as brake thermal efficiency, specific fuel consumption, internal features (fixed speed at1500 rpm and various engine loads via
exhaust gas temperature, in-cylinder pressure, and output power was changing IMEP) with a stoichiometric ratio of air/fuel and fixed injec­
thoroughly analyzed for DMF and its blends in the application to gaso­ tion timing (IT) = 280◦ BTDC. However, the throttle was set at different
line. Moreover, the performance and emissions of engines fueled with positions, while spark timing was fixed at 34◦ BTDC for all throttle po­
DMF and its blends were also compared to that of other fuels to have an sitions and test fuels. As a result, the specific fuel consumption (SFC) for
overall insight into the using strategies of DMF in the future. DMF was found to be the most noticeable influence, SFC was considered
as a major factor for the vehicle users because it was directly linked to
4.1. Engine performance the mileage range of those vehicles. When having an insight on the
initial combustion phasing, it was found the similar trend for DMF,
In the SI engine application, spark timing is called as the most vital gasoline, and ethanol as increasing IMEP. At all IMEPs, initial combus­
parameter to tune and control the combustion phase, which has a tion duration of DMF was lower than that of gasoline and ethanol, this
noticeable effect on the engine-out power and emissions, as well as fuel parameter was magnified at higher IMEP (around 7.1 bar). However, it
consumption. The adjustment for spark timing is typical to maximize the could not confirm that the ignitability of DMF was better than that of
brake thermal efficiency and minimize fuel consumption. Normally for gasoline and ethanol if considering the differences in the pressures and
during the cold start process, the strategies for retarding spark timing are temperatures in the combustion chamber. Also, the fluctuation trends of
applied to fast catalyst light-off, which was identified as the temperature combustion durations for DMF, gasoline, and ethanol were also
to achieve the efficiency of 50% due to shifting the combustion phase revealed, in which combustion duration was considered the time cor­
towards the expansion stroke leading to the increase in the exhaust responding to 10% to 90% of mass fuel burned. Mostly, ethanol has the
temperature [105]. Besides, the fast catalyst light-off was also known as longest combustion duration among the test fuels due to the lowest
a paramount need in the cold-start process because around 80–90% of pressure and temperature in the cylinder. It was worth noting that
total emissions in the cold-start could be contributed from the HC relatively more ethanol than DMF and gasoline for a stoichiometric
emission source [106]. Therefore, the focus on the spark timing study for mixture was the main cause of increasing the combustion duration.
the new-fashioned fuels such as DMF has been considered as an Moreover, the energy density of DMF and gasoline were much higher
extremely important work before assessing the engine performance and than that of ethanol resulting in the rapid heat release, the combustion
other indicators. Indeed, Wang et al. [107] and Daniel et al. [88] have duration was thus more rapidly decrease in case of gasoline and DMF as
performed their experiments on the same SI engine, at a speed of 1500 rising the engine loads. These results were also suitable for the fast flame
rpm, the range of IMEP (Indicated Mean Effective Pressure) from 3.5 bar propagation rate of DMF and gasoline reported in the above section. In
to 8.5 bar under the stoichiometric ratio of fuel/air and various engine general, the combustion process under optimized working conditions for
loads in case of utilizing gasoline, DMF, and ethanol as fuels. They found gasoline was also more suitable for DMF than ethanol. More impor­
that the timing for achieving maximum brake torque was retarded by tantly, the SFC of DMF was reported to be around 30% lower than that of
2◦ CA when the engine knock occurred and the combustion process was ethanol but it was similar to gasoline. However, a limited difference for
instable with the covariance of IMEP higher 5%. They suggested that the DMF associated with the spark timing was pointed out in case of the
knock-limited spark advance was represented as the optimum ignition same spark timing for each individual test fuel. At 7.1 bar of IMEP, the
timing for such cases. At low IMEP, the difference of knock-limited spark combustion duration of DMF was 3◦ CA and 4◦ CA shorter than that of
advance and maximum brake torque was negligible for gasoline, gasoline and ethanol, respectively, leading to a 5 bar higher maximum
ethanol, and DMF. Nonetheless, the changing trends of both knock- pressure in the cylinder compared to ethanol.
limited spark advance and maximum brake torque were observed as In SI engines, the application of injection strategies based on the
increasing the load from 3.5 bar to 4.5 bar IMEP. Indeed, maximum simultaneous combination between DI event and port fuel injection
brake torque and knock-limited spark advance of DMF was more (PFI) was confirmed to blend two fuel types at any blending ratio. This
advanced than that of gasoline, while the spark timing of ethanol was was found to increase the flexibility of fuel use and the potential for
the most advanced. This trend was seen more clearly at the highest IMEP optimizing the combustion process of engines. By detecting the advan­
of 8.5 bar, corresponding to less advanced 6◦ CA for gasoline than DMF, tages of combining injection events between DI and PFI, Daniel et al.
while this maximum difference was 5◦ CA between ethanol and DMF. [31] have compared the use of DMF-gasoline for a single-cylinder 4-
After the spark timings for all test fuels were optimized, DMF had shorter stroke SI engine operated under DI option and DI-PFI strategy. The
combustion duration in comparison with gasoline, around 1◦ CA at 3.5 used SI engine was operated at a fixed engine speed of 1500 rpm and
bar IMEP. Especially at high loads, combustion duration of DMF was various engine loads (IMEP = 3.5 bar to 8.5 bar) at the stoichiometry of
much shorter than that of gasoline, around 4◦ CA at 8.5 bar IMEP. As a air–fuel mixture. As a result, the combustion duration of SI engine fueled
result, such fast burning rate at higher load of 8.5 bar IMEP for DMF with G-DMF25DI was lower than DMF25DI, resulting in lower COV of
caused maximum in-cylinder temperature that was 393 K higher than IMEP and higher combustion stabilities of G-DMF25DI (Fig. 6a). In
that of gasoline, while maximum in-cylinder pressure of DMF was 14 bar addition, the lowest value of COV of IMEP, corresponding to 1.5%, was
higher than that of gasoline. Moreover, ethanol has higher latent heat of found at an engine load of 6.5 bar IMEP. This COV of IMEP was also
vaporization (920 kJ/kg) than DMF (389 kJ/kg), resulting in the in­ lower than the case of applying individual injection event, PFI, or dual-
crease in the cooling effect of ethanol. Consequently, the maximum in- DI. Finally, the dual injection was reported as a promising strategy to
cylinder temperature of DMF was about 100 K higher than that of optimize the combustion process for a blend of 75% gasoline and 25%
ethanol at all engine loads. In another scene, the spark timing sensitivity DMF. Similarly, Wu et al. have performed their experiments on an SI
of DMF has already compared to ethanol and gasoline in the SI engine engine combined two injection strategies as DI and PFI under various
application. The retardation of the spark timing was found to reduce intake manifold absolute pressures (MAPi) such as 0.65 bar, 0.8 bar, and

10
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 6. Comparison of the effects of used fuels (DMF, gasoline, ethanol) and injection strategies on the combustion characteristics and the engine efficiency; (a) – The
dependence of combustion stability on IMEP under various injection event [31]; (b) – The change of indicated efficiency depending on the start of injection in split-
injection event [13]; (c) – Specific fuel consumption and indicated thermal efficiency at various loads [31,107].

0.95 bar, in which only gasoline was used for PFI event while gasoline, suppression indirect injection event and the octane numbers of fuels
ethanol or DMF was directly injected into the cylinder for DI event. They [109]. Theoretically, higher efficiency was thought to be the result of a
realized that the total input-energies at each MAPi for gasoline, DMF, shorter combustion duration. Nonetheless, DMF did not meet the afore-
and ethanol were slightly increased with the reduction of the gasoline mentioned trend, even at the lowest MAPi. Consequently, the lowest
fraction in the PFI event although these increases were different due to indicated efficiency for DMF was observed at lower MAPi. At the highest
the differences of the stoichiometric ratio of air–fuel, the mass of MAPi of 0.95 bar, the indicated thermal efficiency (ITE) of DMF was
injected fuel, and heating value. At 0.65 bar of MAPi, the DMF provided higher than that of gasoline due to the positive effects of shorter com­
slightly higher total input-energies compared to ethanol as rising the bustion duration for DMF [87]. To evaluate the conversion efficiency of
DMF fraction in the DI event. Nonetheless, the opposed trend for ethanol each fuel, the measurement relating to ITE or the indicated specific fuel
and DMF occurred at higher MAPi, this was attributed to the cooling consumption equivalent to gasoline (ISFCE) is performed. Normally,
effect of ethanol greater than DMF. The increase in the total input- indicated efficiency is inversely proportional to ISFCE. In addition to
energies was believed to offer a greater opportunity to achieve a indicated efficiency, ISFCE is also a parameter to compare directly the
higher output power. In addition, the combustion duration for gasoline energy conversion efficiency between as-used fuels. In general, the en­
was observed to increase when reducing the gasoline fraction in the PFI ergy conversion efficiency of DMF compared to gasoline and ethanol
event although this increase in combustion duration was decreased with could be illustrated in Fig. 6 (b&c).
increased MAPi. After optimizing spark timing, the combustion duration Saying about the relationship between ISFCE and indicated effi­
could be further decreased because of the better quality for the knock ciency with IMEP, the study of Daniel et al. [88] have revealed that the

11
A.T. Hoang et al. Fuel 285 (2021) 119140

peak indicated efficiency for DMF, gasoline, and ethanol was achieved necessary to get the equivalent energy released from 1 L of gasoline
with the range of IMEP from 7 bar to 7.5 bar. They reported that such a because the stoichiometric ratios for each fuel are different as the
decrease of indicated efficiency for DMF might be due to higher com­ throttle was posited corresponding to λ = 1.0 at the fixed engine loads.
bustion temperature resulting in rising heat loss and lowering useful Therefore, the fuel consumption for DMF was lower than ethanol and
work [88]. Moreover, the higher loss of heat exchange was thought to slightly higher than gasoline to achieve the same output power. In the
compensate for this effect, it was reflected from the result of ISFCE. As a study of Zhong et al. [109], they reported that the fuel rate of gasoline
result, lower efficiency of DMF compared to ethanol although both was around 33% less than that of ethanol, while this fuel rate for DMF
ethanol and DMF could be competitive to commercial gasoline, espe­ was only slightly higher than that of gasoline. Moreover, in the study on
cially at higher loads. For the split-injection strategy, the changing trend the comparison of specific fuel consumption (SFC) between blends of
of indicated efficiency for DMF, ethanol, and gasoline was also depicted DMF5G (5% DMF and 95% gasoline), E10G (10% ethanol and 90%
in Fig. 6b. Since the application of a split-injection event was believed to gasoline), and gasoline has also revealed that the SFCs for DMF5G and
improve full-load torque, it was utilized in the study of Daniel et al. [13] E10G were similar to that of gasoline at all engine loads. Nonetheless,
to explore the effects of gasoline, ethanol, DMF on the engine perfor­ unlike E10G, DMF5G has exhibited a slightly lower SFC compared to
mance and emissions. They have performed their experiments on a that of gasoline at higher engine loads [110]. Besides, DMF was reported
single-cylinder 4-stroke DISI engine operated at a fixed speed of 1500 to provide around 30% lower SFC compared to ethanol [107]. More
rpm under the stoichiometric ratio of air/fuel at two ratios of split- importantly, Daniel et al. [31] have detected around 3.2% reduction of
injection technique as SOI1:SOI2 = 1:1 and SOI1:SOI2 = 2:1 (SOI was SFC at 7.5 bar of IMEP for D25DI blends. This result was thought to be
defined as the start of injection). In their study, SOI1 was fixed at extremely promising and counter-intuitive due to the efficiency
270◦ BTDC while SOI2 was swept in the range of 90oBTDC-240oBTDC improvement offsetting the lower heating value of DMF. Due to this
with the step of 30◦ CA. The similar results to the previous works relating reason, the synergy between the dual-injection event and the use of
to the achieved maximum IMEP with single-pulse injection event were blend DMF-gasoline was further highlighted although SFC of G-D25DI in
reported and explained due to the increased octane index and decreased case of employing PFI was lower than that of DI when the engine was
stoichiometric ratio of air/fuel for DMF and ethanol, in which this operated at IMEPs of below 6.7 bar. Even with the use strategies of
decrease of stoichiometric ratio of air/fuel was attributed to the oxygen combined fuel such as dual-fuel or PFI, or the application of split-
content in molecular of for DMF and ethanol. In terms of split-injection injection strategy, or the change of injection timings, all studies have
event, the higher IMEP compared to single-pulse injection in the case of indicated that SFC of DMF was comparable to gasoline, while SFC of
DMF and ethanol was found to be more sensitive than with gasoline. The ethanol was much higher than that of gasoline and DMF [13,88]. In a
indicated efficiency of gasoline was found to increase significantly as the recent study of Liu et al. [111] on comparing the volumetric fuel con­
engine was operated under a split-injection strategy. However, the sumption (VFC) of a 4-cylinder 4-stroke DISI engine, which ran on
benefits of ethanol and DMF were seen less clearly. It is worth ethanol, n-butanol, DMF in comparison to their blends with various
mentioning for ethanol and DMF that the utilization of the split-injection ratios of 20%, 50%, 75%, and was operated at different speeds of
event was only beneficial at SOI1:SOI2 = 1:1 and SOI2 = 210◦ BTDC, at 15–120 km/h with intervals of 25 km/h. An important finding relating
which the indicated efficiency was increased up to 0.5% for ethanol and to the use of DMF is that there was an insignificant influence on VFC
1.2% for DMF compared to the case of single-pulse injection. Further­ relative to gasoline as using blends of DMF/gasoline at various DMF
more, there was an improvement up to 2.3% for engine-out power in ratios. The major reason pointed out was that the lower heating value of
case of using DMF with the SOI2 = 200–240◦ BTDC, this power DMF was comparable with gasoline but higher than that of ethanol and
enhancement was mainly attributed to better atomization and the n-butanol. The VFCs of ethanol/gasoline blends were highest, similar to
reduction of piston-wetting phenomenon. In addition, the similarity of the previous works. At blending ratios of 20% and 50% for blends of
ITE and combustion characteristics of DMF compared to gasoline was DMF/gasoline and n-butanol/gasoline, their achieved VFCs were similar
thought to result in the comparable ITE of the engine [107]. However, to the case of using gasoline. Nonetheless, the blends of DMF/gasoline
the application of different injection strategies was detected to bring the and n-butanol/gasoline with>75% of DMF and n-butanol have disclosed
various brake thermal efficiencies (Fig. 6c). For instance, the use of G- higher VFCs than that of gasoline. In general, VFC as adding DMF was
D25DI (duel-injection with gasoline for PFI and blend of 25% DMF and reported to be better than that of adding n-butanol or ethanol. There­
75% gasoline for DI) has produced around 0.9–1.4% higher ITE than the fore, the use of DMF for SI engines was believed to bring more beneficial
case of D25DI (blend of 25% DMF and 75% gasoline for DI), although as considering the fuel economy.
the efficiency of the engine running on G-D25DI condition at higher The exhaust gas temperature (EGT) is a critical parameter reflecting
loads was influenced by PFI [31]. In general, the use of gasoline under the combustion process and fuel conversion efficiency. The EGT was
PFI has reinforced the improvement of combustion chemistry because found to have remarkable effects on the efficiency of the catalyst con­
gasoline has higher volatility, leading to easily pre-mixing in the cylin­ verter and the performance of the turbine-compressor turbocharger. The
der and promoting the DMF combustion. More impressively, due to the EGT was significantly dominated by the maximum in-cylinder temper­
inability of mixing with gasoline and the synergetic characteristic of the ature (Tex_max), the focus on determining the maximum in-cylinder
separate injecting mechanism, DMF has been demonstrated to bring temperature was considered as a common and useful method to assess
higher effectiveness in the application of dual-injection event. the EGT. It could be seen from Fig. 7a that the lowest in-cylinder tem­
Up to now, most of the studies have been in common through the perature has resulted from the combustion of ethanol. This result could
evaluation of the combustion characteristics of DMF for the SI engine; be due to, on the one hand, the latent heat of evaporation of ethanol is
they have affirmed DMF as a promising green fuel as compared to higher than that of DMF and gasoline, leading to more heat absorbed in
ethanol and gasoline. In addition to the above-mentioned observation on the compression stroke, while DMF and gasoline were producing more
indicated efficiency, ISFCE, combustion duration, and COV of IMEP, the net heat because of the less required energy for the phase change. On the
changing progress of in-cylinder pressure, in-cylinder temperature, fuel other hand, the maximum brake torque timing in the case of ethanol was
consumption, brake thermal efficiency, and exhaust gas temperature observed to be much more advanced compared to both gasoline and
were also measured and determined, although the number of study on DMF. As a result, more energy was taken off from the released heat in the
these parameters on the SI engines were limited. Relating to fuel con­ combustion to produce useful work, which was compatible with the
sumption, there is an undeniable fact that the calorific value of DMF was explanation in Fig. 6 (b&c). Experimental study of Daniel et al. [88] has
comparable to gasoline, only lower 7%, while the calorific value of concluded that the Tex_max in case of using DMF was around 30℃ higher
ethanol was 34% lower than that of gasoline. This means that a than the case of gasoline at the lowest IMEP based on the fuel-specific
requirement of 1.512 L and 1.073 L of ethanol and DMF, respectively, is MBT timing but the reducing extent for ethanol was seen to be around

12
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 7. Changing tendency of engine performance and in-cylinder parameters for SI engine running on gasoline, ethanol, and DMF; (a) – Maximum in-cylinder
temperature at various loads [107,109]; (b)- Maximum in-cylinder pressure at various loads [107,109]; (c) – Evolution of in-cylinder pressure and HRR at the
MAPi = 0.95 bar with various fractions of PFI [87].

100 ◦ C in comparison to DMF at lowest IMEP. At maximum brake torque of the Tex_max in case of using DMF compared to gasoline, especially at
timing, a negligible difference of gasoline and DMF was reported. higher loads, was explored. At the highest load, corresponding to 8.5 bar
Nonetheless, the Tex_max for DMF was found to be noticeably higher of IMEP, the Tex_max in case of DMF was around 100 ◦ C lower than that of
relative to gasoline in the case of applying the fuel-specific maximum gasoline, while three remaining fuel samples (ethanol, methanol,
brake torque/knock-limited spark advance timing. Namely, the Tex_max butanol) have provided much lower in-cylinder temperature than with
as utilizing DMF was approximately 120 ◦ C higher than that of gasoline gasoline and DMF [108]. In this situation, the reduction of Tex_max for
at IMEP = 8.5 bar. Indeed, the Tex_max at these values were in accordance DMF compared to gasoline may be due to lower heating value of DMF.
with the efficient working condition of three-way catalysts. More spe­ More interestingly, DMF was also demonstrated as suitable alternative
cifically, the higher in-cylinder temperature was also more useful for fuel for the cold-start conditions rather than other alcoholic fuels. In
turbocharger-based engines with a downsized and compact dimension, addition maximum in-cylinder temperature, the data relating to the
leading to more potential to improve the output power. The similar pressure in the cylinder, in which the in-cylinder pressure was repre­
changing trends of Tex_max for DMF, gasoline, and ethanol were seen in sented through the maximum in-cylinder pressure (Pex_max) or and heat
Fig. 7a, when they have further investigated that the increase in engine release rate (HRR), was believed to be one of key parameters to explain
loads was the main cause of raising the Tex_max. Moreover, peak in- more clearly the changing trend of performance and efficiency. In the
cylinder temperature was found to depend much on ST and combus­ study of Daniel et al, they indicated that the difference between Pex_max
tion duration, this finding was received a good agreement from Wang of three fuel samples (DMF, gasoline, and ethanol) was negligible at low
et al. [107] and Zhong et al. [109]. In contrast, another study performed engine loads. Nonetheless, a significant increase in the Pex_max of ethanol
at SR10 (10◦ CA of spark retarded) condition of an SI engine fueled with and DMF compared to gasoline was found. Moreover, the peak in-
DMF compared to gasoline, ethanol, methanol, butanol to evaluate the cylinder pressure at engine loads >6.5 bar IMEP for the ethanol case
effects of as-used fuels on the cold-start state of the engine. A reduction has exceeded the DMF case [31]. Especially, the Pex_max for the ethanol

13
A.T. Hoang et al. Fuel 285 (2021) 119140

case was around 7 bar and 20 bar greater than that of DMF and gasoline, 4.2. Emission characteristics
respectively, when the engine was operated at a load of 8.5 bar IMEP.
Additionally, the Pex_max for DMF and ethanol was also measured Either theoretically or practically, emission characteristics of SI en­
through maximum brake torque timing for gasoline, where the marginal gines depend primarily on operating conditions, fuel properties, and
difference between DMF, gasoline, and ethanol was reported and this engine design. As mentioned above, the physicochemical properties of
difference was due to the difference in the burning speeds. A similar DMF had significant effects on engine performance as the SI engines are
trend on the change of Pex_max was explored by another study of this fueled with DMF under various injection strategies. Normally, the
group. They have examined the absolute behavior at fuel-specific different changing propensity on engine performance is the main cause
maximum brake torque timings and retarded gasoline knock-limited of varying engine-out emissions. Better engine efficiency is considered
maximum brake torque timing timings and evaluated the detrimental as having a proportional relation to better combustion and more
impacts at SR10 [108]. Generally, a sharp reduction in the Pex_max was completely burnt fuel, resulting in lower pollutant emissions. Emission
caused by the impact of the retardation of ignition timing with load. The characteristics from engines could be evaluated through NOx, PM and
similarity of the decay in Pex_max at low engine loads between as-used soot, HC, CO, and other compounds.
fuels could be illustrated in Fig. 7b, in which the difference range of Being one of the gas emissions characterizing the combustion tem­
Pex_max at 3.5 bar IMEP was lower than 2%. This difference was observed perature, NOx is considered as the most concerning emission. Typically,
more self-evidently as increasing the loads, whereby the greatest NOx emissions are formed under the combustion temperature exceeding
changes were come from gasoline at SR10, resulting in a 48% decrease 1800 K in the case of an oxygen-locally rich mixture [112,113]. More­
in Pex_max. A similar reduction in Pex_max in the DMF case to gasoline was over, sensitive are NOx emissions with the engine loads, this means that
also declared, this was attributed to the similarity of knock suppression NOx emissions are increased with increasing the engine loads [114]. In
abilities for DMF and gasoline. Absolutely, the findings over maximum recent studies on the emission characteristics of SI engines running on
in-cylinder pressure and temperature matched with the in-cylinder gasoline, DMF, and ethanol, it was found that NOx emission produced
pressure plotted in Fig. 7c, where the in-cylinder pressure and HRR from ethanol was lower than that of gasoline, however, NOx emissions
for the DMF case were higher than those of gasoline and ethanol, even at from DMF were observed to be higher than gasoline, especially at higher
low loads [107]. At higher loads, an evident difference in the peak in- loads. These results could be due to higher combustion temperatures for
cylinder pressure between DMF and gasoline was recognized. As the DMF case [13]. Even with the ignition timing optimized for specific
mentioned above, the constant ignition timing might induce the engine fuel in the study of Daniel et al. [88], NOx emissions were observed to
knock for the case of using DMF and gasoline although the knock trend increase compared to both gasoline and ethanol, however, at loads
for DMF seemed to be lower than that of gasoline, especially at higher higher than 7.5 bar IMEP, NOx emission in the optimum case was
loads. Due to this reason, the following studies on the engine knock as slightly higher than with gasoline. For the combined injection strategy,
employing DMF must play a vital role in the application study of DMF in PFI and DI in the study of Wu et al. [87], NOx emissions at all MAPi in
SI engines. the same working condition and PFI fueling were the highest. Obviously,
Fig. 7c has presented the in-cylinder pressure trace and corre­ the peak in-cylinder pressure for ethanol case was similar to the DMF
sponding HRR for DMF, gasoline, and ethanol in the dependence on the case but the peak in-cylinder temperature of DMF was somewhat higher
crank angle at the highest MAPi = 0.95 bar with fixed ignition timing = because of higher heating value and lower cooling effects of DMF than
7oBTDC [87]. It could be seen the comparison of the in-cylinder pressure with ethanol. Furthermore, the H/C ratio was also thought to have an
trace and HRR in the case of PFI strategy with 100%, 55%, and 0% of inversely proportional relation to the relative NOx emissions [115]. The
fuel mass fractions. The slightly earlier and greater peak pressures were fact is that the H/C ratio (corresponding to 3) of ethanol is the highest,
reached as decreasing the PFI mass fractions of fuels, in which DMF and followed by gasoline with H/C ratio > 2 and the H/C ratio of DMF is
ethanol were emphasized by this behavior. As reported, DMF possessed only = 1.33. In addition to NOx emissions, CO2 emission was also
a faster burning velocity than gasoline, it was possibly as a consequence mentioned in a few studies on the use of DMF for SI engines. It is evident
of higher temperatures in the combustion chamber [88]. The combus­ that CO2 emission was considered as an indication of evaluating the
tion duration for DMF was thus considerably decreased compared to the combustion quality. In general, the authors have reported that the in­
ethanol case when DMF was used in 55% PFI case, in which it was crease in the engine loads and the use of optimum advanced ST appeared
partially introduced by the DI event. As a result, Pex_max was shifted to reduce CO2 emissions. On the one hand, increasing the engine loads
towards the left (meaning earlier), this could be self-evident with the 0% was the main cause of more complete combustion. On the other hand,
PFI case. This achieved result has disclosed and provided further proof to the combustion efficiency of gasoline, DMF, and ethanol was higher,
confirm the negative effects of PFI in combustion efficiency. Moreover, it peaking at the range of 6–7 bar IMEP, resulting in the minimal CO2
could also be seen that the dual-injection event has caused the negligible emission [88]. For the split-injection event, CO2 emission for the DMF
effects for all test fuels on HRR in the initial stage. Nonetheless, the peak case was also much higher than with ethanol and gasoline [13]. In
HRR reached the new one in the later stages when reducing the mass contrast, CO2 emission for gasoline was found to higher than with DMF
fraction in PFI strategy but HRR for the gasoline case showed a marginal and gasoline as considering the effects of SR10 (spark retard by 10◦ CA).
influence by the decrease of PFI fueling. Indeed, HRR in the case of As a result, CO2 emissions for gasoline, ethanol, and DMF were increased
100% PFI fueling has been proven to peak lower and slightly later than to 19.4%, 10.7%, and 5.6%, respectively [108]. For the real test cycles
that of the 100% DI fueling case. The changing tendency of HRR was presented in the study of Liu et al. [111], NOx and CO2 emissions for D20
similar to the DMF case, however, the peak of HRR in the DMF case was (20% DMF and 80% gasoline of volume) cases were concluded to be
found to be slightly earlier and higher than that of gasoline because, for slightly higher than those of gasoline. The range of NOx increase was
DMF case, the in-cylinder temperature was higher and the LBV was around 2.7% to 5.54% and was around 0.64% to 1.65% for CO2 emission
faster than those of gasoline [89]. In contrast, the peak of HRR for DMF as applying the UDC, EUDC, and NEDC test cycles. However, the use of
was reported to be marginally higher compared to the ethanol case DMF has been found to increase the engine-out CO2 emission compared
although the LBV of DMF was slower than with ethanol. Perhaps, this to gasoline, DMF has the added benefits for CO2 consuming from the raw
was due to higher heating value and higher DMF mass injected into the production because DMF was extracted from carbohydrates.
combustion chamber compared to ethanol. One more, the DI strategy For formaldehyde (CH2O) and acetaldehyde (CH3CHO) emissions,
was demonstrated to have more benefits for DMF cases compared to the they were found for the DMF case to be around 50% lower than those of
PFI event when HRR was enhanced and the combustion duration was gasoline and ethanol [14] because CH2O and CH3CHO were generated
shortened for DMF. during the oxidation of fuels under low temperatures [116]. Nonethe­
less, combustion temperatures for DMF appeared to be higher than with

14
A.T. Hoang et al. Fuel 285 (2021) 119140

gasoline and ethanol or, at least, the DMF combustion process has taken and gasoline, it could be seen the difference in mole fraction of the major
place at greater temperatures in case of optimized ST [88], helping to species such as CO, C4H4, C2H2, CH4, C2H4, C2H6, HCHO, and CH3CHO.
inhibit the formation of CH2O and CH3CHO. Moreover, the DMF struc­ Obviously, CO and C5H8 isomers were produced by the migration of CH3
ture was also believed to produce a lower concentration of CH2O and from C2 to C3 in DMF, while CO, C3H4, and C2H4 were known as the
CH3CHO emissions. As presented in the decomposition and oxidation products of additional ring-opening reaction of DMF [86]. Due to quite
mechanism, DMF was first consumed by oxidizing CH3 group, leading to slow H abstraction from the ring, ketone emissions of DMF were seen to
forming ketenyl (HCCO) and propyne (C3H4) [117]. Simultaneously, be much higher than that of ethanol, even higher than that of gasoline
toluene (C7H8) was generated from the reaction between benzene [120]; this could be observed in Fig. 8e. In contrast, the HCHO and
(C6H6) and (CH3) radical during the DMF combustion process. After CH3CHO species produced from DMF were observed to be lower than
that, benzaldehyde was formed by further oxidation of toluene [118]. In those of gasoline and ethanol. Therefore, emissions of CH2O and
addition to furan and its derivatives, the presence of many intermediates CH3CHO in the DMF case were reduced, this trend could be clearly seen
produced from the combustion such as benzene, oxygenated hydrocar­ in Fig. 8f.
bons, hydrocarbons, and free radicals was also reported in the laminar Fig. 8e and f showed the emissions of some carbonyl species in
DMF flame, in which benzene was detected to readily observe in the addition to three common aldehydes such as CH2O, CH3CHO, C6H5CHO
DMF-rich flame, when benzene was believed to form based on the produced from the DMF combustion. Overall, the total concentration of
combination of two C3H3 radicals. Moreover, C3H3 was found to be one carbonyls shown in Fig. 8e was ordered as followed: ethanol (15.4 ppm)
of the derivatives of the degradation product of DMF, this could be seen < DMF (89.6 ppm) < gasoline (115.5 ppm). However, aldehyde emis­
in the consumption pathways of DMF shown in Fig. 8a [119]. Based on sions for DMF were much lower than gasoline and other alcoholic fuels
the mole fraction analysis shown in Fig. 8 (b-d) between DMF, ethanol, (Fig. 8f), this result was also matched with the study of Wang et al.

Fig. 8. Mechanism of consumption reaction for DMF and emissions; (a) – Possible consumption pathways of DMF for the formation of soot precursor and emissions
[119]; (b) – Mole fractions of DMF [121]; (c) – Mole fraction of ethanol [122]; (d) – Mole fraction of gasoline [123]; (e) – Ketone emissions; (f) – Aldehyde emissions
[14]; (g) – Emissions of HC, CO, and NO from the cold start combustion [110].

15
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 8. (continued).

[107]. In summary, achieved results relating to HC and carbonyl emis­ In order to evaluate the combustion efficiency, HC and CO emissions
sions were also matched with the achieved results from the combustion are very important indicators. Normally, these emission components are
evolution afore-mentioned. inversely proportional to NOx and CO2 emissions, and they are also

16
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 8. (continued).

sensitive to the fuel properties, engine operating conditions, and engine much on the oxygen content in fuels. In all tests as considering the whole
design. For the engine operating conditions, the HC and CO emissions engine loads, ethanol with around 34.78% of oxygen content produced
were typically decreased as increasing the engine loads. Alternatively, the lowest CO emissions, followed by DMF with around 16.67% and
the oxygen content in fuels was detected to aid better oxidation. For the gasoline with non-oxygen. However, the changing trend of CO emission
level of HC emissions, all studies have confirmed that HC emission was different from HC emission. For instance, Zhong et al. [109] re­
produced from the DMF combustion lay between gasoline and ethanol, ported that the peak value of the parabola curve for CO emission in the
even with the combined injection strategies or optimized ST ethanol case was matched with the minimum value of the parabola
[13,87,109]. However, in the case of spark timing retard, HC emissions curve for DMF and gasoline. The support of oxygen content at lower
for DMF tended to decrease in comparison with gasoline or ethanol at loads has been revealed through the maximum difference of CO emis­
lower loads < 5.5 bar IMEP [108]. Generally, HC emissions for DMF was sions between DMF and gasoline at these loads, corresponding 7 g/kWh.
around 30% lower than that of gasoline in the range of engine loads. Moreover, in the study of Daniel et al. [108], they have indicated that
More detailed, the method of gas chromatography-mass spectrometry CO emissions for DMF at loads <5 bar were lower than with gasoline,
was used to analyze the emissions of HCs and carbonyl (aldehydes and ethanol, methanol, and butanol. At higher loads, the fluctuation trends
ketones) for a DISI engine running on gasoline, DMF, and ethanol. of CO emissions were observed in the comparison between the
Shukla et al. [110] have studied the performance and emission charac­ oxygenated fuels although CO emissions for all oxygenated fuels were
teristics of a small-carbureted SI electrical generator fueled with DMF5G consistently lower than with gasoline. In contrast, the application of
(5% DMF + 95% gasoline) in comparison to E10 (10% ethanol + 90% split-injection for SI engine running on DMF appeared to bring low ef­
gasoline) and gasoline (Fig. 8g). Surprisingly, DMF5G was found to be ficiency in the decreasing strategy of CO emissions, especially with early
better suited to the cold-start combustion than with E10, this was injection event [13]. This result was due to the lower in-cylinder tem­
because of the similar distillation behavior of DMF to gasoline, leading peratures when DMF was directly injected into the combustion chamber,
to the decrease of cold-start emissions. Similarly to HC emissions, CO resulting in poorer combustion. In a recent study, a similar tendency was
emissions were reported from all experiments to reduce with the found when an SI engine has experienced the test cycles such as UDC,
increased engine loads. This tendency appeared to be the inverse of EUDC, and NEDC with various speeds aiming to evaluate the real
indicated efficiency, in which the highest efficiency has provided the emissions [111]. As a result, both HC and CO emissions for D20 for all
lowest CO emissions. The CO emissions were demonstrated to depend test cycles were higher than with gasoline, although this increase was

17
A.T. Hoang et al. Fuel 285 (2021) 119140

insignificant. Namely, the maximum increase for HC and CO emissions was around 30% higher than that of ethanol, even higher than that of
in the D20 case was indicated for UDC test cycle with 5.03% (mgHC/km) gasoline. They believed that PM emissions consisted of accumulation
and 6.17% (mgCO/km) compared to gasoline. Although various particles with diameter >100 nm and nucleation particles <100 nm in
changing tendencies for emissions could be observed depending on diameter (Fig. 9a). As increasing the engine loads (at 5.5 bar IMEP), the
many factors, better emission trend of HC and CO emissions for DMF total number of particles in the nucleation mode for DMF was still higher
compared to gasoline was an undeniable fact. than that of gasoline and ethanol, although the total number of particles
It is evident that the emission regulations for vehicles equipped with in the accumulation mode for DMF was much lower than with gasoline.
DISI engines are more and more stringent to improve air quality. This However, they are in common that total PM emission for DMF was lower
could be seen from Euro 5b regulations, in which the PM emissions were than with gasoline in the low-to-moderate loads. This difference in the
regulated with lower level compared to the previous regulations, and formation mechanism of PM under the accumulation and nucleation
particle number (PN) emissions were considered as an adding indicator modes might be attributed to the measuring condition and method, and
to control the emissions since Euro 6 regulations were issued. As using the surrounding environment. Wang et al. [126], they have concluded
DMF as fuel, it was expected to produce lower PM emissions than gas­ that the major components of PM emissions came from volatile com­
oline because DMF is a partially-oxygenated fuel [12]. It was reported ponents including HCs, lubricants, etc. The achieved results from their
that, under DI condition, PM emissions, on which composed of solid soot study were matched with the previous studies that soot component
with condensed hydrocarbon or adsorbed hydrocarbon, were signifi­ produced from DMF, which was considered as the main component in
cantly decreased in comparison to the gasoline case, even at lower loads. PM emissions, were lower than with gasoline but higher than that of
Indeed, lower PM emissions were demonstrated through the correlation ethanol [107]. Besides, the engine operating conditions were proven to
to HC emissions [124]. This could be explained that lower PM emissions affect significantly the PM emissions of DISI engines [88,109,126]. Facts
have caused less soot surface for the condensation and adsorption of HC, show that the engine operating conditions were proven to affect
resulting in shifting the HC ratio to nucleated particles, and PN emis­ significantly the PM emissions. At higher loads (8.5 bar IMEP), there was
sions were thus increased in comparison to gasoline. There are typically a separation between the accumulation and nucleation modes at the
two modes such as the accumulation and nucleation mechanism for the point of about 50 nm in diameter of particles. Indeed, the distribution of
distribution of PM size [125]. The separation between the two afore­ particle size was dominated by the nucleation mode when the particles
mentioned mechanisms could be observed in previous works [88,109] with around 25–30 nm in diameter have peaked [88]. At this load
although the antagonistic results were also reported as all experiments extent, the ratios of accumulation particles of the total concentration of
were carried out on the same engine. In the study of Daniel et al. [88], particles were only 2.1% for ethanol, 1.7% for DMF, and 18.3% for
for the accumulation mechanism, PM emission was thought to consist of gasoline.
solid carbonaceous particles with diameter >50 nm, while the diameter More specifically, <4000 particles/cm3 was produced from ethanol
of particles in the nucleation mechanism was lower than 50 nm. They and DMF compared to 21,000 particles/cm3 for gasoline. Likely, the
found that the size distribution of particles, at lower loads, in the droplet velocity and mean droplet diameter of gasoline were higher than
nucleation mode was slightly less than that of accumulation one. those of DMF and ethanol, this confirmation could be self-evident in
Namely, the total particles under the accumulation mode were 67.1%, Fig. 3a&b, causing an increase in the piston impingement. Other reasons
64.4%, and 62.1% for ethanol, DMF, and gasoline, respectively. Mean­ to explain the changing tendency of PM emissions were indicated
while, the total particle concentration of DMF was around 21,805 par­ thermal efficiency (Fig. 6c) and in-cylinder temperatures (Fig. 7a), in
ticles/cm3 lower than with ethanol. This positive result was due to lower which higher in-cylinder temperature and higher thermal efficiency
surface tension and viscosity of DMF compared to those of ethanol, have promoted the pyrolysis of fuel as well as reduced the solid carbo­
resulting in smaller fuel droplets injected into the combustion chamber naceous emissions. Generally, it is thought that the oxygenated fuels
[77]. Furthermore, in-cylinder temperatures for the ethanol case were have significant effects on reducing the formation of the soot precursor,
much lower compared to the DMF case, causing a greater cooling effect consequently reducing PM emissions. For DMF20, the combustion
of ethanol. As a result, the benefits resulted from the ethanol with higher temperature is somewhat higher than that of gasoline and E20, pro­
oxygen content were counteracted by a higher cooling effect. On the moting further the oxidation of soot. Thus, a remarkable decline of PM
contrary, in the study of Zhong et al. [109], the size distribution of emissions can be achieved when using DMF20 for the SI engine. In the
particles in the nucleation mode for DMF, at lower loads at 3.5 bar IMEP, case of E20, the latent heat of vaporization of ethanol is the higher than

Fig. 9. PM emission for different fuels at various operating conditions; (a) – PM size distribution at fixed spark timing of 10oBTDC and load of 8.5 IMEP [88]; (b) –
PM emission for various fuel depending on test cycle [111].

18
A.T. Hoang et al. Fuel 285 (2021) 119140

that of DMF but octane number and heating value of ethanol are lower had lower concentration than gasoline. In addition to traditional influ­
than DMF, resulting in the lower combustion temperature and the encing factors on PM emissions as mentioned above, PM and PN emis­
decrease of oxidized soot rate compared to DMF. As a result, PM emis­ sions were found to depend much on the test cycle, which was self-
sions from DMF20 were observed to be lower than those of E20 and evident in Fig. 9b. According to the UDC test cycle, the maximum
gasoline, although DMF has still higher C/O ratio than ethanol [111]. reduction of PM for DMF20 was 48.06% compared to gasoline although
Nonetheless, the opposite tendency has occurred as considering the this reduction extent was smaller based on other test cycles such as
nucleation particles at the same loads of 8.5 bar IMEP, the worst EUDC or NEDC. More specifically, all test cycles also have provided a
offender with around 122,000 particles/cm3 was found as DMF. This spectacular result on lower PM emission for D20 than with E20, Bu20,
sharp increase in nucleation particle number was because the ring- and gasoline [111]. Not only that, PN for DMF20 was reported to
structured DMF was difficult to burn completely, products produced decrease for all test cycles, in which the largest reduction extent of PN
from the incomplete combustion of DMF were known as soot precursors for DMF20 was 69.43% for UDC test cycle in comparison with the gas­
[23,127]. Generally, PM emissions in the DMF case at low loads were oline case. Furthermore, PN for DMF20 was comparable to that of E20
found to be comparable to the gasoline case. However, at high loads, although it was lower than with Bu20.
nucleation particles for DMF were much higher than with gasoline, Somewhat, the flame images showed its velocity of propagation and
while accumulation particles for DMF had similar sizes to ethanol and its luminance. In general, the oxygenated fuels such as DMF, MF, or

Fig. 10. Soot formation extent depending on various factors; (a) – Impact of fuel on the soot profile based on derivative weight (upper) and activation energy (lower)
in the cases of gasoline, DMF, E25, and ethanol at engine parameters of 8.5 bar IMEP, speed = 1500 rpm, equivalent ratio = 0.9, SOI = 100◦ BTDC [126]; (b) – Effects
of temperature and DMF concentration on sooting tendency [131]; (c) – Effects of engine loads on soot formation [126]; (d) – HRTEM images for gasoline, DMF20,
and E20 soot samples at various resolutions [136].

19
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 10. (continued).

ethanol have brighter flames than with non-oxygen containing fuels required activation energies (Fig. 10a), it was also suitable for the
such as iso-octane or gasoline. More importantly, since the flame lumi­ highest in-cylinder temperatures for DMF. It could be seen from Fig. 10a
nance was proportional to the in-cylinder combustion temperatures, the that the profiles based on the analysis of derivative weight for the soot
higher the flame luminance was, the greater the soot concentrations production from DMF in comparison to gasoline and ethanol. From the
were [128]. It could be reported in reference [129] that the luminance of derivative weight profiles, the activation energy (Ea, KJ/mol) was
the DMF flame was the highest, it was thus believed that the soot con­ shown in the lower figure, in which the integrated area on the diagram
centration was considered as leading dominating factor for the DMF indicated the relationship between temperature and derivative weight
flame luminance and the brighter flame of DMF was the main cause of was proportional to the soot weight [126]. Obviously, DMF combustion
higher soot production. Nonetheless, an ever-aforementioned inter­ has produced less soot than gasoline and E25, although it was higher
esting contrast was lower PM emissions for DMF compared to gasoline, than with neat ethanol. Moreover, the activation energy was also
which was because of the oxygen-containing molecule of DMF revealed to be highest for soot formation from gasoline, corresponding
[87,107,109]. Therefore, DMF, on the one hand, could generate more to 153 kJ/mol, followed by E25 (124 kJ/mol), DMF (109 kJ/mol), and
soot than the gasoline case because of the circular structures of molec­ ethanol (83 kJ/mol). As a result, the soot reactivity generated from DMF
ular; on the other hand, the oxygen content in DMF was also thought to was much higher than with gasoline. More impressively, four major
enhance the oxidization process, resulting in the reduction of the soot factors including (i) – carbon structure, (ii) – primary particle, (iii) –
concentration. Finally, this behavior of DMF and gasoline appeared to agglomerated particle size, and (iv) – soot oxidization mode were
generate a low soot emission for DMF [130]. From experimental studies, explored to explain why DMF was easier to oxidize than with gasoline
it was found that the oxidation of soot generated from DMF was easier [126]. More interestingly, an adding idea was found to perfect the soot
than gasoline, even than gasoline to some extent. This was indicated formation mechanism when the activation energy and the combustion
through the lowest maximum mass loss rate temperature and lowest temperature were proportional to the engine loads, this was indicated a

20
A.T. Hoang et al. Fuel 285 (2021) 119140

difficulty in the soot oxidation if soot was formed at higher combustion parallel stacks might be not formed because of a high amount of non-
temperatures, which were illustrated in Fig. 10b&c. It could be the aromatic hydrocarbons, resulting in the difference in the nano­
sooting tendency of DMF depending on the temperature and concen­ structures that have a significant effect on the reactivity of soot particles
tration from Fig. 10b, in which it was self-evident that soot yield (dot) towards oxygen [141]. The elemental analysis from FTIR spectra from
and soot amount (bar) in the pyrolysis mechanism of DMF were the combustion process of gasoline, DMF20, and E20 was added to
increased proportionally to the increase in temperature and DMF con­ clarify the mechanism of soot formation. On the basis of the analysis
centration [131]. Alternatively, Fig. 10c has pointed out the correlation from FTIR spectra reported in reference [136], the highest contents of
between engine load and the soot oxidization. In general, the profile of oxygen and hydrogen came from E20, in which soot samples produced
soot oxidization in the gasoline case has shifted to higher temperatures from E20 and DMF20 were seen to provide aliphatic and oxygenated
along with the increase in engine loads. At lower loads (5.5 bar IMEP), functional groups with higher concentrations compared to gasoline, this
around 80% of soot originated from gasoline was oxidized at a tem­ finding was also agreed by Conturso et al. [142,143]. However, C–C
perature of 700 K, while there was around 85% of soot joining in the functional groups containing soot for the DMF20 and E20 cases were
oxidation process for the DMF case at this temperature. However, the lower than with gasoline. The FTIR spectra have shown a broad peak in
temperature was required to be higher to yield 80% of oxidized soot the range of 3300–3700 cm− 1 for the soot, indicating a stretched vi­
when the engine was operated on gasoline at a higher load (8.5 bar bration from the hydroxyl (OH) group [136]. Facts show that ROH,
IMEP) [126]. In the DMF case, the profiles of soot oxidization were C6H5OH, and R-COOH groups, which were arisen from the aforemen­
negligible shifted as compared to gasoline. The increase in engine loads tioned OH groups, were considered as intermediate substances during
has increased the net soot production (including formation-growth- the oxidation process of soot, although the OH peak could not find the
oxidation for soot) for both DMF and gasoline although a lower extent contribution of R-COOH groups because these R-COOH groups were fast
for DMF [132]. These results were matched with the calculated activa­ decomposed in flame structures [140].
tion energy and maximum mass loss rate temperature for DMF and In general, the presence of functional groups such as aliphatic and
gasoline as changing the engine loads. However, it was thought to be oxygenated groups in the soot was revealed through the FTIR analysis,
difficult to entirely explain the cause of soot reduction for the blends of in which the amount of aromatic C–C bonds in gasoline soot was highest
gasoline with DMF or ethanol because of the differences in fuel-bound but aliphatic and oxygenated group-rich soot were also found for the
oxygen mechanism [133,134]. Moreover, the oxygen atom in DMF or DMF20 and E20 cases. However, increasing surface oxygen groups could
other oxygenated fuels was believed to promote the formation of CO/ have an evident effect [144], or no effect on the soot oxidation [145],
CO2 by decomposing the intermediate species based on oxygenated while the enhancement of aliphatic character and the reduction of PAH
hydrocarbons although a different extent was observed because of the sizes were pointed out to influence the soot oxidation more. In order to
differences in the structure of fuel molecular. Typically, the increase in have the answer relating to the demonstration of the quantification
the number of oxygenated radicals was the major reason for oxidizing ability on the soot formation between DMF and gasoline, Tran et al.
soot, resulting in the soot reduction [135]. Furthermore, it was also [146] have simulated the flame of DMF in comparison to that of a sur­
detected that C2H5O radicals were produced when adding ethanol to rogate mixture (including n-heptane, iso-octane, and toluene with the
favor the CO/CO2 formation instead of C2H4 or C2H2 aiming to break the ratio of 13.7%:42.9%:43.4% by volume). They found that the mole
pathways forming soot. However, a lower extent was found with the fraction of a major product such as C5H6 from the DMF combustion was
presence of DMF because the O/C ratio of DMF (O/C = 0.17) was lower around 20 times lower than that of the surrogate mixture. Meanwhile,
than with ethanol (O/C = 0.5), resulting in the reduction of the fuel- the mole fractions of C6H5-C2H5, C8H8, and C6H6 in the surrogate
bound oxygen effect [136]. Moreover, the DMF combustion process mixture case were around 30, 6, and 2 times, respectively, greater
has produced higher amounts of species, (for instance C6H6, C6H7O, compared to the DMF case. These results have shown the significantly
CH2 = CH-CH = CH2, C5H6O, and C5H6) with larger molecular mass lower mole fractions of PAH components in the DMF flame compared to
dominating the soot oxidation reactivity compared to the ethanol case the surrogate mixture, affirming DMF as an interesting and promising
[137]. In addition, the amounts of C–C bonds in DMF were also much ON improver. In general, the obtained study results from the use of DMF
higher than that of ethanol, favoring the formation of soot precursors. for SI engines were comparable to the existing commercial fuels, even
Furthermore, smaller agglomerated particles from the DMF combustion having some better characteristics. However, the utilization of DMF for
mentioned above were also attributed to easy oxidation of DMF because SI engine application should be thoroughly analyzed and evaluated on
of a higher ratio of surface to volume, leading to the greater acceleration other aspects. In summary, the effects of DMF properties on the SI engine
of the oxidation for DMF than with gasoline, in which the soot genera­ performance, combustion and emission characteristics were given in
tion was believed to experience this oxidization. It was proposed that the Table 3.
utilization of oxygenated fuels could make the after-treatment system
less its function because the concentration of PM emissions was lower as 5. Engine knocking and noise
well as the energy demand for the soot oxidation was lower than the
requirement [138]. In the case of using DMF20 and E20, it was indicated By analyzing the reaction kinetics based on the thermal, chemical,
that slower PAH condensation and soot-growth rates might be the main and elemental mechanism of DMF along with associated compounds,
cause of the reduction in the size of primary particles for DMF20 and E20 Friese et al. [148] have already calculated the enthalpies and dissocia­
compared to gasoline [136]. From Fig. 10d, it could be seen smaller size tion energies for the bonds in DMF molecular and related radical species
of primary particles in soot from E20 and DMF20 flame compared to that by employing quantum chemical tools. The highest bond energy was
of gasoline through HRTEM images. This finding showed a good found to lie on the ring-carbon-H bonds with a value of 120 kcal/mol
agreement with previous works, in which less carbon atoms reduced the compared to 115 kcal/mol for ring-carbon-CH3 bonds and 85 kcal/mol
formation of soot precursor, and then, soot [135,139]. Moreover, some for the C–H bond of CH3 groups in DMF. This result relating to DMF was
ordered stacks of parallel PAH layers and longer fringe lengths from explained due to the strong chemical bond as well as the geometrical
gasoline soot compared to DMF20 and E20 soot samples could be stiffness of a five-carbon membered ring-based DMF, resulting in having
observed in Fig. 10d. In addition, amorphous character of E20 and little tendency to react chemically at high temperatures but exhibiting
DMF20 soot samples were seen to be higher than that of gasoline. These an excellent knock-inhibitor. Alternatively, a higher RON of DMF
results might be due to poor arrangement of PAHs in E20 and DMF20 compared to commercial gasoline was considered as an important
soot samples, in which the residence time of E20 and DMF20 soot parameter to inhibit the engine knocking [149]. Because of this reason,
samples in flames was detected to be shorter than the gasoline case, DMF, as a new-fashioned green fuel, was paid much attention to the
reducing the degree of graphitization [140]. Furthermore, graphite-like application in engines without any modifications. Indeed, audible

21
A.T. Hoang et al. Fuel 285 (2021) 119140

Table 3
Emission and performance characteristics of SI engines running on DMF-based fuels.
Engine parameter Operating condition Fuels Engine performance Emission characteristics Ref

• SI, 1 cylinder, 4 • DI with injection pressure = DMF; ULG; ETH 1. For single DI • NOx for ULG < NOx for ETH < [13]
S; 15 MPa; • Optimum IMEP as SOI = 260- NOx for DMF;
• Compression • Fixed speed at 1500 rpm and 280oBTDC; • HC for ETH < HC for ULG < NOx
ratio 11.5:1; various loads; • Volumetric efficiency (VE): peak VE for for DMF;
• Single DI with various SOI; ULG = 75.9% at SOI = 210oBTDC < • CO for ETH < CO for ULG < CO
• Split DI with SOI1:SOI2 = 1:1 peak VE for DMF = 77.9% at SOI = for DMF; At SOI2 = 90oBTDC, CO
and 1:2; 210oBTDC < peak VE for ETH = 77.9% for DMF < CO for ULG;
at SOI = 270oBTDC; • CO2 for ETH < CO2 for ULG < CO2
• CD for ETH < CD for DMF < CD for for DMF;
ULG;
• Peak ITE for ETH = 40.3% > Peak ITE
for DMF = 38.1% > Peak ITE for ULG =
36.2%;
• SFC for ULG < SFC for DMF < SFC for
ETH;
• 2. For split DI with fixed SOI1 =
270oBTDC
• Peak IMEP for SOI1:SOI2 = 2:1 > Peak
IMEP for SOI1:SOI2 = 1:1; IMEP for
ETH > IMEP for DMF > IMEP for ULG;
• Peak VE for ULG < peak VE for DMF <
peak VE for ETH;
• Compared to single DI: ↑ 0.5% ITE for
ETH for SOI1:SOI2 = 1:1; ↑ 1.2% ITE
for DMF for SOI1:SOI2 = 1:1; ↑ 0.8%
ITE for ULG for SOI1:SOI2 = 1:1;
• ↑↓ SFC compared to single DI;
• SI, 1 cylinder, 4 • DI with injection pressure = ULG, DMF, ETH – • ↓62% HCHO, ↓ 52% CH3CHO, and [14]
S; 15 MPa at SOI = 280oBTDC; ↓ 6% C6H5CHO for DMF
• Compression • Fixed speed at 1500 rpm and compared to ULG;
ratio 11.5:1; load of 5.5 bar IMEP; • Σ(HCHO and CH3CHO) for DMF <
• Injection timing = 21oBTDC; Σ (HCHO and CH3CHO) for ETH;
• SI, 1 cylinder, 4 • PFI with injection pressure = • Gasoline for PFI and • ITE for G-D25DI (gasoline for PFI and – [31]
S; 3 bar at SOI = 50oBTDC; DI; DMF for DI) > 0.9–1.4% of ITE for
• Compression • DI with injection pressure = • DMF for DI; D25DI (duel-injection blend with 25%
ratio 11.5:1; 150 bar at SOI = 280oBTDC; DMF of volume);
• Fixed speed at 1500 rpm and • SFC of DDI (DMF for DI) > SFC of
various loads; D25DI > SFC of GDI (gasoline for DI) >
SFC of G-D25DI;
• IT of DDI > IT of D25DI ≈ IT of GDI >
IT of G-D25DI;
• Maximum in-cylinder pressure (MiP)
for DDI > MiP for G-D25DI;
• Combustion duration (CD) for DDI <
CD of G-D25DI < CD of (GDI and
D25DI);
• SI, 1 cylinder, 4 • Dual-injection with injection DMF, ULG, ETH • ITE for DMF < ITE for ULG < ITE for • HC for ETH < HC for HC for DMF [87]
S; pressure = 150 bar at SOI = ETH; < HC for ULG;
• Compression 280oBTDC for DI; injection • CD for DMF < CD of ETH < CD of ULG; • CO2 for ETH ≈ CO2 for ULG < CO2
ratio 11.5:1; pressure = 3 bar at SOI = • MiP for ULG < MiP for ETH MiP for for DMF;
50oBTDC for DI; DMF; • NOx for ETH < NOx for ULG <
• Fixed speed at 1500 rpm and NOx for DMF;
various MAPi = 0. 65 bar, 0.8
bar, 0.9 bar
• SI, 1 cylinder, 4 • DI with injection pressure = MF; DMF; unlead • CD for MF < CD of DMF < CD of ETH < • NOx for ETH < NOx for ULG < [107]
S; 15 MPa at SOI = 280oBTDC; gasoline (ULG); ethanol CD of ULG; NOx for DMF < NOx for MF;
• Compression • Fixed speed at 1500 rpm and (ETH) • Maximum in-cylinder temperature • HC for ETH < HC for MF < HC for
ratio 11.5:1; various loads; (MiT) for ETH < MiT for ULG < MiT for DMF < HC for ULG;
• Optimized spark timings; DMF < MiT for MF; • CO for ULG < CO for ETH < NOx
• MiP for ULG < MiP for DMF < MiP for for (DMF and MF);
ETH < MiP for MF; • Σ(HCHO and CH3CHO) for ETH >
• ITE for DMF < ITE for ULG < ITE for Σ(HCHO and CH3CHO) for
MF < ITE for ETH at loads < 7.5 IMEP; gasoline > Σ(HCHO and CH3CHO)
ITE for DMF > ITE for ULG at loads > for DMF;
7.5 IMEP;
• SFC for ULG < SFC for DMF < SFC for
MF < SFC for ETH;
• SI, 1 cylinder, 4 • DI with injection pressure = DMF, ULG, ETH, • Peak ITE for MTH > Peak ITE for ETH • NOx for MTH < NOx for ETH < [108]
S; 15 MPa at SOI = 280oBTDC; methanol (MTH); butanol > Peak ITE for DMF > Peak ITE for BUT NOx for BUT < NOx for ULG <
• Compression • Fixed speed at 1500 rpm and (BUT) > Peak ITE for ULG; NOx for DMF;
ratio 11.5:1; various loads; • EGT for ULG > EGT forDMF > EGT for • HC for (DMF, ULG, ETH) < HC for
• Optimized spark timings; ETH and BUT > EGT for MTH; BUT < NOx for MTH;
• ↑↓ CO at loads < 5.5 bar;
(continued on next page)

22
A.T. Hoang et al. Fuel 285 (2021) 119140

Table 3 (continued )
Engine parameter Operating condition Fuels Engine performance Emission characteristics Ref

• CO2 for ETH < CO2 for ULG < CO2


for DMF;
• Maximum CO2 for ELG; minimum
CO2 for MTH;
• Maximum CO2 for ELG; minimum
CO2 for MTH;
• PM for DMF > PM for ETH > PM
for ULG; PM diameter for DMF <
PM diameter for ETH < PM
diameter for ULG at SR10 and
load = 8.5 bar IMEP;
• SI, 1 cylinder, 4 • Various power ULG, DMF5G (5% DMF • ↓ 4% maximum brake power for • ↓ >30% CO for DMF5G and E10G [110]
S; by volume), E10G (10% DMF5G and 16% maximum brake compared to ULG;
• Compression ethanol by volume) power for E10G compared to gasoline; • ↓ HC for DMF5G and E10G
ratio 5.1:1; • SFC for E10G > SFC for DMF5G > SFC compared to ULG;
for ULG; • ↓ NOx for DMF5G and E10G
compared to ULG at power < 0.6
kW; ↑ NOx for DMF5G and E10G
compared to ULG at power > 0.6
kW;
• SI, 4 cylinders, • DI with NECD test cycle; ULG and blends of ULG • ↓ ID (ID of D20 < ID of E20 < ID of B20 • ↑ 2.57% CO for D20; ↑ 32.58% CO [111]
4 S; • Various speed from 2 to 120 with 20% DMF (D20), < ID of ULG); for E20; ↓ 5.19% CO for B20;
• Turbocharged; km/h with intervals of 15 km/ 20% ethanol (E20), 20% • ↑ 0.72% SFC for D20, ↑ 11.87% SFC for • ↑ 3.5% HC for D20; ↑ 2.33% HC
• Compression h; butanol (B20) of volume E20; ↓ 1.94% for B20; for E20; ↓ 2.91% HC for B20;
ratio 10.5:1; • ↑ 3.09% NOx for D20; ↓ 12.15%
NOx for E20; ↓ 13.21% NOx for
B20;
• ↑ 1.18% CO2 for D20, ↑ 2.39%
CO2 for B20; ↓ 5.7% for E20;
• ↓ 46.46% PM for D20; ↓ 19.59%
PM for E20; ↓ 13.43% PM for B20;
• ↓ 65.88% PN for D20; ↓ 72.87%
PN for E20; ↓ 52.88% PN for B20;
• SI, 1 cylinder, 4 • DI with injection pressure = ULG, DMF, ETH – • Soot and PM for ETH < Soot and [126]
S; 15 MPa at SOI = 280oBTDC; PM for DMF < Soot and PM for
• Compression • Fixed speed at 1500 rpm and ULG
ratio 11.5:1; under fuel-rich condition λ =
0.9;
• SI, 1 cylinder, 4 • DI with injection pressure = ULG, DMF, ETH • ITE for ETH < ITE for ULG < ITE for • ↑↓ HC; ↑ HC for ETH at SOI < [147]
S; 15 MPa at various SOI; DMF; 200oBTDC or SOI > 300oBTDC;
• Compression • Fixed speed at 1500 rpm and λ • Peak ITE for ETH at 270oBTDC; • ↑↓ NOx; ↑ NOx for ETH at SOI >
ratio 11.5:1; = 1; • Peak ITE for ULG and DMF at 325oBTDC; lowest NOx for DMF at
210oBTDC; SOI = 1755oBTDC
• ↑↓ IMEP; • ↓ NOx with retarding SOI;

engine knocking was detected first in an SI engine running on DMF and considered as an octane improver in gasoline [152]. For the combining
gasoline at an engine load of 6 bar IMEP by Zhong et al. [109]. However, injection strategy of PFI and DI reported in the study of Wu et al. [87],
when the engine reached a specific load such as 6.7 bar IMEP for DMF the cooling effect was also found to reduce the charge temperature in the
and 7.2 bar IMEP for gasoline, the engine knocking was the most severe combustion chamber and to raise the volumetric efficiency, resulting in
but the engine knocking magnitude observed from the gasoline com­ suppressing knock. The efficiency of suppressing knock was also indi­
bustion was higher than with DMF. As aforementioned, the octane rating cated as the highest for ethanol, and lowest for gasoline. One more proof
(OR) of each fuel was known as the major factor used to measure the on the anti-knock for DMF compared to gasoline and ethanol was also
anti-knock tolerance in the SI engine via the premixed charge supply. In investigated by Daniel et al. [88], who used fixed MBT/knock-limited
addition to OR, the history of the combustion pressure and temperature spark advance timing for gasoline to apply to the cases of DMF and
was considered as other critical factors influencing the engine knocking ethanol.
[150]. For DI engine, the cooling effect due to the fuel evaporation as As evidently illustrated in Fig. 11a for DMF, gasoline, and ethanol,
injected into the combustion chamber affected noticeably the combus­ there was not any difference in the maximum brake torque location at
tion efficiency. The cooling effect of each fuel could be determined low engine load, corresponding to 3.5 bar of IMEP. It was seen as a
through the ratio between the heat of vaporization and the lower relatively flat curve based on the relationship between IMEP and spark
heating value of that fuel. Therefore, this ratio (heat of vaporization advance. Nonetheless, the most advanced ST could be observed for the
(HV)/lower heating value (LHV)) was around 0.008 for ethanol, 0.01 for ethanol case throughout the remaining engine loads, this result was
DMF, and 0.032 for gasoline. As a result, ethanol had the highest anti- believed to result from higher knock resistance and higher LBV of
knock ability, followed by DMF and gasoline based on the cooling ef­ ethanol. More clearly, the maximum brake torque ST as utilizing DMF
fect and OR. It was worth mentioning that an SI engine, which was and gasoline was 5◦ CA and 11◦ CA, respectively, less advanced than with
designed and calibrated to run on gasoline, would show a lower ethanol at the highest load, also leading to lower anti-knock of gasoline
knocking tendency as fueled with DMF or blends of DMF/gasoline. and DMF compared to ethanol, even with spark advance. In the study
Undeniably, Nisbet reported in the year of 1946 that the knock of an SI Sarathy et al. [153], they indicated that the fuels containing oxygen had
engine was decreased as using DMF [151]. More impressively, a com­ the capability of knock suppression, relating to the relative oxygen
mercial company working on the petroleum field such as British Petro­ content. Perhaps, there was not surprising when DMF possessed a poorer
leum Company has claimed their patent on the potential of DMF as anti-knock capacity in comparison with ethanol due to DMF’s lower

23
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 11. The anti-knocking of an SI engine running on various fuels; (a) – Maximum brake torque/knock-limited spark advance timing for DMF, ethanol, and gasoline
at different engine loads [88]; (b) – Cumulative distribution functions (CDF) at 25◦ CA for an 0.08 bar of intake manifold absolute pressures [29].

oxygen content. It could be seen the comparison on the engine knocking than 2% DMF have shown little effect on the reduction of FC, this result
combustion between EEE standard fuel and other fuels such as DMF5, could be observed in Fig. 12a. As compared with ethanol and gasoline
DMF10, DMF15 (5%, 10%, 15% of DMF in gasoline), E10DMF10 (10% based on the FC under a relative humidity of 35–45%, DMF was also
of ethanol, 10% of DMF in gasoline), E10 (10% of ethanol) from found to have lower FC (0.271) than with gasoline (0.466) and ethanol
Fig. 11b. In this study, the knock-limited spark advance timing at (0.355). Furthermore, the tendency of FC for DMF was observed to be
− 20◦ CA was used to evaluate the cumulative distribution functions relatively stable compared to the fluctuating tendency of gasoline and
(CDF) for all test fuels. With the largest number of cycles for EEE CDF, it ethanol [158]. Besides, the viscosity of DMF was also higher than for
has shown the significant knock, while all cases based on DMF appeared gasoline and ethanol, the afore-mentioned factors are why DMF was
to work similarly to each other with a lower knocking extent than the proven to have better lubricity compared to gasoline and ethanol. More
EEE case. However, the knocking tendency of DMF cases was slightly interestingly, the mean coverage percentage calculated on the basis of
greater compared to the non-knocking results in the ethanol-containing lubrication film was around 68% for DMF, while this value was only
cases. Furthermore, they also reported the statistical test by applying the 53% for gasoline and 46% for ethanol [158]. The fact shows that higher
KS test [29]. As a result, with a confidence of 99%, blends of DMF at film coverage for the DMF case could be the key factor leading to a
knock-limited spark advance timing have shown negligibly different smaller wear scar (Fig. 12b) compared to the cases of gasoline and
behavior. The relative knocking propensity was further investigated as ethanol [159]. It could be seen from Fig. 12b that the largest wear scar
ST was fixed at − 25◦ CA. The increase in the amount of knocking for all has come to gasoline, this was matched with the findings on the FC of
test fuels was observed at this advanced ST. More importantly, less se­ gasoline that was much higher than that of DMF and ethanol. The
vere knocking for the ethanol blend (E10, E10DMF10) in comparison spectra of FT–IR of the steel surface before and after being immerging in
with the DMF blends was recognized. Finally, at the same blending ratio, DMF could be illustrated in Fig. 12c, in which showed a flex vibration
the use of ethanol was concluded to reduce engine knock more effec­ peak of OH group. This peak was then shifted to the lower wavenumber
tively than with DMF. Nonetheless, DMF was affirmed to be a compet­ due to the association effect originated from hydrogen atoms. Moreover,
itive potential additive due to its high energy and low water solubility. It other flex vibration peaks for radical groups of CH3, CH2, COOH, C– –O
was also suggested to employ the blends of gasoline-ethanol-DMF for the could be observed, confirming that hydroxyl (OH) and carboxyl (COOH)
SI engine to bring many benefits. groups have promoted the antifriction capacity of DMF as well as DMF/
gasoline blends after adding a reasonable concentration of DMF. It was
6. Lubricity and wear also concluded that the presence of (OH) and (COOH) groups after
immersing the sample into DMF was considered as the chemical re­
In addition to the performance and emissions of engine fueled with actions induced by the friction phenomenon of DMF and rubbing pairs
DMF and its blend, a matter given out is the concern to lubricating [119]. In addition, thin and slim wear scars in the DMF case were seen in
properties and wear of DMF as used in both SI and CI engine. The fact is Fig. 12d, while deep and wide wear scars could be showed for the
said that the oxygen/nitrogen-containing polar compounds were gasoline-lubricated specimens. The iron (Fe) elements from the test steel
demonstrated to be key to improving the lubrication capacity for used could react with the active functional groups such as (OH) and (COOH),
fuels in engines [154]. This means the oxygen in DMF might contribute resulting in the formation of a robust anti-wear film [160]. Simulta­
to the increase in the lubricity. Indeed, Hu et al. [155] have reported that neously, higher viscosity fuels were believed to absorb well the pro­
DMF possessed the better lubricity in comparison to gasoline because duced components in the friction duration [161]. In general, the anti-
the friction coefficient (FC) of DMF was 0.038, while the gasoline FC was wear and the lubricity of DMF are much higher than gasoline and
0.089. Even with DMF/gasoline blends, they were also confirmed to ethanol, this aspect is said to be one of the advantages of using DMF as
reduce the FC compared to the case of pure gasoline. Furthermore, the an alternative fuel.
FC was figured out increasing along with time and the engine loads due
to the rise of the contact area [156]. 7. Conclusion
In a recent study, Dearn et al. [157] have explored that as low as
around 2% DMF by volume in DMF/gasoline blends was considered The production of new-fashioned biofuels from lignocellulosic
enough to reduce significantly the FC, enhancing the antifriction char­ biomass feedstock is playing a key role in replacing fossil fuels to pro­
acteristic of the blends. However, the DMF/gasoline blends with more duce energy. Many researches have been performed to analyze and

24
A.T. Hoang et al. Fuel 285 (2021) 119140

Fig. 12. Lubrication and wear characteristics of DMF; (a) – The tendency of changing FC depending on DMF blending ration [157]; (b)- The size of wear scar for
DMF, ethanol, and ethanol [158]; (c) – The spectra of FT-IR for DMF before (black) and after (red) friction [159]; (d) – Images of the steel surface after immerging in
gasoline (upper) and DMF (lower) [159] (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

evaluate the synthesis process of DMF from available and renewable bifunctional catalyst systems such as Ru-Co/SiO2, Pt3Ni/C, and
lignocellulosic biomass feedstock under either a one-step or two-step PdAu4/GC800 were found to bring the high selective yield in the
process with various catalyst systems and reaction conditions aiming conversion reaction of HMF to DMF.
to achieve a high yield of DMF. In addition, the comprehensive evalu­ • The DMF spray was comparable to gasoline, while penetration length
ation of the applicability of DMF for existing spark ignition engine with and SMD of DMF were slightly higher than those of ethanol. More­
regards to the characteristics of performance and emissions, knocking over, laminar burning velocity and Markstein length of DMF were
and noise, and wear was also thoroughly reviewed to identify the lower compared to ethanol and gasoline, indicating that the DMF
strengths and weaknesses of DMF compared with gasoline and ethanol. flame was more unstable than that of ethanol and gasoline case
Moreover, the comparison of the spray and flame characteristics of DMF although the burning rate of DMF was found to be higher than with
with gasoline and ethanol was also conducted. In general, some specific ethanol and gasoline. More interestingly, DMF had shorter combus­
conclusions could be withdrawn from this review paper as followed: tion duration comparison to gasoline and ethanol along with the
ignition delay lay between gasoline and ethanol, suggesting that the
• On the catalyst front, bifunctional catalyst systems containing hy­ improvement of combustion, performance, and emissions of SI en­
drogenation metal have provided higher efficiency in the selective gine running on DMF compared to gasoline and ethanol.
DMF production than with mono-metallic catalysts. Normally, under • For the SI engine, DMF has shown similar combustion characteristics
the same reaction condition, noble metal-based bifunctional catalyst to gasoline and ethanol. More specially with optimized spark timing,
systems could produce DMF with the yield > 90%. Some typical burning rate due to the presence of oxygen and high octane number

25
A.T. Hoang et al. Fuel 285 (2021) 119140

for DMF was found to be faster than with gasoline and ethanol, the work reported in this paper.
leading to higher in-cylinder temperature and pressure compared to
the cases of gasoline and ethanol. Nonetheless, the energy density of References
DMF was around 7–15% lower than that of gasoline, lowering
thermal efficiency and increasing specific fuel consumption [1] Nguyen HP, et al. The electric propulsion system as a green solution for
management strategy of CO2 emission in ocean shipping: a comprehensive
compared to gasoline, while the combustion efficiency was assessed review. Int Trans Electr Energy Syst 2020;e12580. https://doi.org/10.1002/
to be higher than with gasoline. Regarding engine-out emissions, the 2050-7038.12580.
increase in NOx for DMF in comparison to gasoline was detected, [2] Nižetić S, Djilali N, Papadopoulos A, Rodrigues JJPC. Smart technologies for
promotion of energy efficiency, utilization of sustainable resources and waste
while the fluctuating trends of CO and HC emissions due to the dif­ management. J Clean Prod 2019;231:565–91.
ferences in working and experimental conditions such as the cold- [3] Gosovic B. The quest for world environmental cooperation: the case of the UN
start condition or injection event were also reported. More interest­ Global Environment Monitoring System, vol. 8. Routledge; 2019.
[4] Cheng G, Zhang C, Cao Y, Jiang Z. Review of energy-consumption measuring
ingly, PM emission of the SI engine running on DMF was much lower techniques for the flotation process. Energy Sources Part A Recover Util Environ
than the gasoline case. Due to the energy density of DMF was Eff 2018;40(19):2367–77.
29–33% higher than that of ethanol, performance characteristics of [5] Hoang AT, Nguyen TH, Nguyen HP. Scrap tire pyrolysis as a potential strategy for
waste management pathway: a review. Energy Sources, Part A 2020:1–18.
an engine fueled with DMF such as thermal efficiency and specific
https://doi.org/10.1080/15567036.2020.1745336.
fuel consumption were thus improved, while engine-out emissions [6] Hoang AT, et al. Rice bran oil-based biodiesel as a promising renewable fuel
were also better than the case of ethanol in some criteria. In general, alternative to petrodiesel: a review. Renew Sustain Energy Rev 2020. https://doi.
the performance and emissions of the SI engine fueled with DMF lay org/10.1016/j.rser.2020.110204.
[7] Wang X, Liang X, Li J, Li Q. Catalytic hydrogenolysis of biomass-derived 5-
between gasoline and ethanol, these findings were demonstrated hydroxymethylfurfural to biofuel 2, 5-dimethylfuran. Appl. Catal. A Gen. 2019.
even with other strategies of DMF use such as port injection or split [8] Saha B, Bohn CM, Abu-Omar MM. Zinc-assisted hydrodeoxygenation of biomass-
injection. derived 5-hydroxymethylfurfural to 2, 5-dimethylfuran. ChemSusChem 2014;7
(11):3095–101.
• DMF was detected to have not any influences on the knock since used [9] Kumalaputri AJ, Bottari G, Erne PM, Heeres HJ, Barta K. Tunable and selective
for SI engine due to its higher RON and octane number than with conversion of 5-HMF to 2, 5-furandimethanol and 2, 5-dimethylfuran over
gasoline, this is believed to be an outstanding advantage of DMF in copper-doped porous metal oxides. ChemSusChem 2014;7(8):2266–75.
[10] Hu L, Lin L, Wu Z, Zhou S, Liu S. Recent advances in catalytic transformation of
the SI application. In addition, lubrication and wear characteristics biomass-derived 5-hydroxymethylfurfural into the innovative fuels and
of DMF were better than those of gasoline and ethanol. chemicals. Renew Sustain Energy Rev 2017;74:230–57.
[11] Saha B, Abu-Omar MM. Advances in 5-hydroxymethylfurfural production from
biomass in biphasic solvents. Green Chem 2014;16(1):24–38.
In closing, this review paper on a new-fashioned class of fuels like [12] Xu H, Wang C. A comprehensive review of 2, 5-dimethylfuran as a biofuel
DMF has highlighted the pathway from the production process to the candidate. Biofuels from Lignocellul Biomass Innov beyond Bioethanol 2016:
application on the engines. Compared to other alternative fuels, DMF 105–29.
[13] Daniel R, Wang C, Xu H, Tian G. Split-injection strategies under full-load using
has shown excellent effects on performance and emission characteristics
DMF, a new biofuel candidate, compared to ethanol in a GDI engine, SAE
of the SI engine, even when SI engines were operated under the Technical Paper; 2012. doi: https://doi.org/10.4271/2012-01-0403.
advanced combustion modes. Due to such wide range of applications, [14] Daniel R, Wei L, Xu H, Wang C, Wyszynski ML, Shuai S. Speciation of
DMF has been considered as one of the potential biofuel and the hydrocarbon and carbonyl emissions of 2, 5-dimethylfuran combustion in a DISI
engine. Energy Fuels 2012;26(11):6661–8.
promising value-added chemical, which could be synthesized from [15] Nguyen DC, Hoang AT, Tran QV, Hadiyanto H, Wattanavichien K, Pham VV.
lignocellulosic biomass or starch-rich feedstock although the environ­ A review on the performance, combustion and emission characteristics of SI
mental concerns, yield, and selectivity effectiveness were thought to be engine fueled with 2,5-Dimethylfuran (DMF) compared to ethanol and gasoline.
J Energy Resour Technol 2020. https://doi.org/10.1115/1.4048228.
the main drawbacks of DMF producing processes. Therefore, as the [16] Mhadmhan S, Franco A, Pineda A, Reubroycharoen P, Luque R. Continuous flow
global need for DMF is ever-increasing, producing processes and selective hydrogenation of 5-hydroxymethylfurfural to 2, 5-dimethylfuran using
environment-related concerns as well as the input materials should be highly active and stable Cu–Pd/reduced graphene oxide. ACS Sustain Chem Eng
2019;7(16):14210–6.
carefully and thoroughly studied to target the low-cost products and [17] Esteves LM, et al. Effect of support on selective 5-hydroxymethylfurfural
have the minimum impact on the environment. In order to deal with hydrogenation towards 2, 5-dimethylfuran over copper catalysts. Fuel 2020;270:
these issues, finding the input materials and feedstock with available 117524. https://doi.org/10.1016/j.fuel.2020.117524.
[18] Guo D, et al. Selective hydrogenolysis of 5-hydroxymethylfurfural to produce
and low-cost attributes are considered as one of the key factors, in which biofuel 2, 5-dimethylfuran over Ni/ZSM-5 catalysts. Fuel 2020;274:117853.
biomass waste from agro- and forestry-based industries and https://doi.org/10.1016/j.fuel.2020.117853.
carbohydrate-based waste must be the main selection. In addition, green [19] Thananatthanachon T, Rauchfuss TB. Efficient production of the liquid fuel 2, 5-
dimethylfuran from fructose using formic acid as a reagent. Angew Chemie 2010;
and inexpensive catalyst systems for the DMF synthesis process should
122(37):6766–8.
be developed, the use of bi-phasic catalyst systems was found to bring [20] Hoang AT, Nguyen DC. Properties of DMF-fossil gasoline RON95 blends in the
more promising results in terms of DMF yields and selectivity. More consideration as the alternative fuel. Int J Adv Sci Eng Inf Technol 2018;8(6):
noticeably, the recovery of solvents in the DMF synthesis reactions was 2555–60.
[21] Caes BR, Teixeira RE, Knapp KG, Raines RT. Biomass to furanics: renewable
not an easy task, resulting in the challenges in the purification process as routes to chemicals and fuels. ACS Sustain Chem Eng 2015;3(11):2591–605.
well as the quality of DMF. Although DMF has shown many advantages [22] van Putten R-J, Van Der Waal JC, De Jong ED, Rasrendra CB, Heeres HJ, de
as using in existing SI engines, it must admit that DMF could be only Vries JG. Hydroxymethylfurfural, a versatile platform chemical made from
renewable resources. Chem Rev 2013;113(3):1499–597.
used as an alternative fuel as ethanol once the selling price of DMF is [23] Qian Y, Zhu L, Wang Y, Lu X. Recent progress in the development of biofuel 2, 5-
comparable to gasoline or ethanol. Due to this reason, it is necessary to dimethylfuran. Renew Sustain Energy Rev 2015;41:633–46.
optimize the DMF production process and catalyst-based reaction sys­ [24] Wang J, Liu X, Hu B, Lu G, Wang Y. Efficient catalytic conversion of
lignocellulosic biomass into renewable liquid biofuels via furan derivatives. Rsc
tems to target the commercialization and realization strategy that Adv 2014;4(59):31101–7.
should be addressed. Last but not least important, the support policy for [25] Kong X, et al. Catalytic conversion of 5-hydroxymethylfurfural to some value-
DMF production should be deployed more synchronously and effectively added derivatives. Green Chem 2018;20(16):3657–82.
[26] Bhaumik P, Dhepe PL. Solid acid catalyzed synthesis of furans from
to avoid following the same path of failed policies such as ethanol carbohydrates. Catal Rev 2016;58(1):36–112.
production in some countries. [27] Iris KM, Tsang DCW. Conversion of biomass to hydroxymethylfurfural: a review
of catalytic systems and underlying mechanisms. Bioresour Technol 2017;238:
716–32.
Declaration of Competing Interest [28] Gandini A. Furans as offspring of sugars and polysaccharides and progenitors of a
family of remarkable polymers: a review of recent progress. Polym Chem 2010;1
The authors declare that they have no known competing financial (3):245–51.
interests or personal relationships that could have appeared to influence

26
A.T. Hoang et al. Fuel 285 (2021) 119140

[29] Rothamer DA, Jennings JH. Study of the knocking propensity of 2, 5- [58] Li Q, Man P, Yuan L, Zhang P, Li Y, Ai S. Ruthenium supported on CoFe layered
dimethylfuran–gasoline and ethanol–gasoline blends. Fuel 2012;98:203–12. double oxide for selective hydrogenation of 5-hydroxymethylfurfural. Mol Catal
[30] Alexandrino K, Millera A, Bilbao R, Alzueta MU. Interaction between 2, 5-dime­ 2017;431:32–8.
thylfuran and nitric oxide: Experimental and modeling study. Energy Fuels 2014; [59] Jeong G-Y, Singh AK, Sharma S, Gyak KW, Maurya RA, Kim D-P. One-flow
28(6):4193–8. syntheses of diverse heterocyclic furan chemicals directly from fructose via
[31] Daniel R, Xu H, Wang C, Richardson D, Shuai S. Combustion performance of 2, 5- tandem transformation platform. NPG asia Mater 2015;7(4):e173.
dimethylfuran blends using dual-injection compared to direct-injection in a SI [60] Kong X, Zheng R, Zhu Y, Ding G, Zhu Y, Li Y-W. Rational design of Ni-based
engine. Appl Energy 2012;98:59–68. catalysts derived from hydrotalcite for selective hydrogenation of 5-
[32] Wang H, et al. Recent advances in catalytic conversion of biomass to 5-hydrox­ hydroxymethylfurfural. Green Chem 2015;17(4):2504–14.
ymethylfurfural and 2, 5-dimethylfuran. Renew Sustain Energy Rev 2019;103: [61] Yang P, Xia Q, Liu X, Wang Y. Catalytic transfer hydrogenation/hydrogenolysis of
227–47. 5-hydroxymethylfurfural to 2, 5-dimethylfuran over Ni-Co/C catalyst. Fuel 2017;
[33] Hu L, Lin L, Liu S. Chemoselective hydrogenation of biomass-derived 5-hydrox­ 187:159–66.
ymethylfurfural into the liquid biofuel 2, 5-dimethylfuran. Ind Eng Chem Res [62] Kong X, Zhu Y, Zheng H, Zhu Y, Fang Z. Inclusion of Zn into metallic Ni enables
2014;53(24):9969–78. selective and effective synthesis of 2, 5-dimethylfuran from bioderived 5-
[34] Mani CM, Braun M, Molinari V, Antonietti M, Fechler N. A high-throughput hydroxymethylfurfural. ACS Sustain Chem Eng 2017;5(12):11280–9.
composite catalyst based on nickel carbon cubes for the hydrogenation of 5- [63] Luo J, et al. Base metal-Pt alloys: a general route to high selectivity and stability
hydroxymethylfurfural to 2, 5-dimethylfuran. ChemCatChem 2017;9(17): in the production of biofuels from HMF. Appl Catal B Environ 2016;199:439–46.
3388–94. [64] Chatterjee M, Ishizaka T, Kawanami H. Hydrogenation of 5-hydroxymethylfur­
[35] Dou Y, Zhou S, Oldani C, Fang W, Cao Q. 5-Hydroxymethylfurfural production fural in supercritical carbon dioxide–water: a tunable approach to dimethylfuran
from dehydration of fructose catalyzed by Aquivion@ silica solid acid. Fuel 2018; selectivity. Green Chem 2014;16(3):1543–51.
214:45–54. [65] Goyal R, et al. Studies of synergy between metal–support interfaces and selective
[36] Wang C, Xi Z. Co-operative effect of Lewis acids with transition metals for organic hydrogenation of HMF to DMF in water. J Catal 2016;340:248–60.
synthesis. Chem Soc Rev 2007;36(9):1395–406. [66] Chen M-Y, Chen C-B, Zada B, Fu Y. Perovskite type oxide-supported Ni catalysts
[37] Román-Leshkov Y, Barrett CJ, Liu ZY, Dumesic JA. Production of dimethylfuran for the production of 2, 5-dimethylfuran from biomass-derived 5-
for liquid fuels from biomass-derived carbohydrates. Nature 2007;447(7147): hydroxymethylfurfural. Green Chem 2016;18(13):3858–66.
982. [67] Yu L, et al. Robust and recyclable nonprecious bimetallic nanoparticles on carbon
[38] Li C, et al. Tailored one-pot production of furan-based fuels from fructose in an nanotubes for the hydrogenation and hydrogenolysis of 5-hydroxymethylfurfural.
ionic liquid biphasic solvent system. Chinese J Catal 2015;36(9):1638–46. ChemCatChem 2015;7(11):1701–7.
[39] Chidambaram M, Bell AT. A two-step approach for the catalytic conversion of [68] Luo J, et al. Unraveling the surface state and composition of highly selective
glucose to 2, 5-dimethylfuran in ionic liquids. Green Chem 2010;12(7):1253–62. nanocrystalline Ni–Cu alloy catalysts for hydrodeoxygenation of HMF. Catal Sci
[40] Insyani R, Verma D, Kim SM, Kim J. Direct one-pot conversion of Technol 2017;7(8):1735–43.
monosaccharides into high-yield 2, 5-dimethylfuran over a multifunctional Pd/ [69] Gao Z, Li C, Fan G, Yang L, Li F. Nitrogen-doped carbon-decorated copper catalyst
Zr-based metal–organic framework@ sulfonated graphene oxide catalyst. Green for highly efficient transfer hydrogenolysis of 5-hydroxymethylfurfural to
Chem 2017;19(11):2482–90. convertibly produce 2, 5-dimethylfuran or 2, 5-dimethyltetrahydrofuran. Appl
[41] Choudhary V, et al. Insights into the interplay of Lewis and Brønsted acid Catal B Environ 2018;226:523–33.
catalysts in glucose and fructose conversion to 5-(hydroxymethyl) furfural and [70] Zhu Y, Kong X, Zheng H, Ding G, Zhu Y, Li Y-W. Efficient synthesis of 2, 5-
levulinic acid in aqueous media. J Am Chem Soc 2013;135(10):3997–4006. dihydroxymethylfuran and 2, 5-dimethylfuran from 5-hydroxymethylfurfural
[42] Van Nguyen C, et al. Combined treatments for producing 5-hydroxymethylfur­ using mineral-derived Cu catalysts as versatile catalysts. Catal Sci Technol 2015;5
fural (HMF) from lignocellulosic biomass. Catal Today 2016;278:344–9. (8):4208–17.
[43] Binder JB, Raines RT. Simple chemical transformation of lignocellulosic biomass [71] Braun M, Antonietti M. A continuous flow process for the production of 2, 5-
into furans for fuels and chemicals. J Am Chem Soc 2009;131(5):1979–85. dimethylfuran from fructose using (non-noble metal based) heterogeneous
[44] Iryani DA, Kumagai S, Nonaka M, Sasaki K, Hirajima T. Production of 5- catalysis. Green Chem 2017;19(16):3813–9.
hydroxymethyl furfural from sugarcane bagasse under hot compressed water. [72] An H, Yang WM, Maghbouli A, Chou SK, Chua KJ. Detailed physical properties
Procedia Earth Planet Sci 2013;6:441–7. prediction of pure methyl esters for biodiesel combustion modeling. Appl Energy
[45] Ye X-N, et al. Fast pyrolysis of corn stalks at different growth stages to selectively 2013;102:647–56.
produce 4-vinyl phenol and 5-hydroxymethyl furfural. Waste Biomass [73] Hoang AT, Tran QV, Al-Tawaha ARMS, Pham VV, Nguyen XP. Comparative
Valorization 2019;10(12):3867–78. analysis on performance and emission characteristics of an in-Vietnam popular 4-
[46] Wang G-H, et al. Platinum–cobalt bimetallic nanoparticles in hollow carbon stroke motorcycle engine running on biogasoline and mineral gasoline. Renew
nanospheres for hydrogenolysis of 5-hydroxymethylfurfural. Nat Mater 2014;13 Energy Focus 2019;28:47–55.
(3):293–300. [74] Dutta S. Deoxygenation of biomass-derived feedstocks: hurdles and opportunities.
[47] Wang X, Liu Y, Liang X. Hydrogenolysis of 5-hydroxymethylfurfural to 2, 5- ChemSusChem 2012;5(11):2125–7.
dimethylfuran over supported Pt–Co bimetallic catalysts under mild conditions. [75] Hu L, Jiang Y, Xu J, He A, Wu Z, Xu J. Chemocatalytic pathways for high-
Green Chem 2018;20(12):2894–902. efficiency production of 2, 5-dimethylfuran from biomass-derived 5-
[48] Hu L, Tang X, Xu J, Wu Z, Lin L, Liu S. Selective transformation of 5-hydroxy­ hydroxymethylfurfural. Biomass Biochem Biofuels 2020:377–94.
methylfurfural into the liquid fuel 2, 5-dimethylfuran over carbon-supported [76] Hoang AT. Experimental study on spray and emission characteristics of a diesel
ruthenium. Ind Eng Chem Res 2014;53(8):3056–64. engine fueled with preheated bio-oils and diesel fuel. Energy 2019;171:795–808.
[49] Jae J, Zheng W, Lobo RF, Vlachos DG. Production of dimethylfuran from [77] Tian G, Li H, Xu H, Li Y, Raj S.M. Spray characteristics study of DMF using phase
hydroxymethylfurfural through catalytic transfer hydrogenation with ruthenium doppler particle analyzer, SAE Int. J. Passeng. Cars-Mechanical Syst., vol. 3, no.
supported on carbon. ChemSusChem 2013;6(7):1158–62. 2010-01–1505, pp. 948–958, 2010.
[50] Priecel P, Endot NA, Carà PD, Lopez-Sanchez JA. Fast catalytic hydrogenation of [78] Hoang AT, Le AT. Trilateral correlation of spray characteristics, combustion
2, 5-hydroxymethylfurfural to 2, 5-dimethylfuran with ruthenium on carbon parameters, and deposit formation in the injector hole of a diesel engine running
nanotubes. Ind Eng Chem Res 2018;57(6):1991–2002. on preheated Jatropha oil and fossil diesel fuel. Biofuel Res J 2019;6(1):909–19.
[51] Nishimura S, Ikeda N, Ebitani K. Selective hydrogenation of biomass-derived 5- [79] Hoang AT. Prediction of the density and viscosity of biodiesel and the influence of
hydroxymethylfurfural (HMF) to 2, 5-dimethylfuran (DMF) under atmospheric biodiesel properties on a diesel engine fuel supply system. J Mar Eng Technol
hydrogen pressure over carbon supported PdAu bimetallic catalyst. Catal Today 2018. https://doi.org/10.1080/20464177.2018.1532734.
2014;232:89–98. [80] Li H, Ma X, PoWen TU, Xu H, Shuai S-J, Ghafourian A, Numerical study of DMF
[52] Huang Y, Chen M, Yan L, Guo Q, Fu Y. Nickel–tungsten carbide catalysts for the and gasoline spray and mixture preparation in a GDI engine, SAE Technical
production of 2, 5-dimethylfuran from biomass-derived molecules. Paper; 2013.
ChemSusChem 2014;7(4):1068–72. [81] Tu P, et al., Investigation on the spray characteristics of DMF-isooctane blends
[53] Guo W, et al. Efficient hydrogenolysis of 5-hydroxymethylfurfural to 2, 5-dime­ using PDPA, SAE Technical Paper; 2014.
thylfuran over a cobalt and copper bimetallic catalyst on N-graphene-modified Al [82] Jężak S, Dzida M, Zorębski M. High pressure physicochemical properties of 2-
2 O 3. Green Chem 2016;18(23):6222–8. methylfuran and 2, 5-dimethylfuran–second generation biofuels. Fuel 2016;184:
[54] Chen B, Li F, Huang Z, Yuan G. Carbon-coated Cu-Co bimetallic nanoparticles as 334–43.
selective and recyclable catalysts for production of biofuel 2, 5-dimethylfuran. [83] Matsumoto A, et al. Spray characterization of ethanol gasoline blends and
Appl Catal B Environ 2017;200:192–9. comparison to a CFD model for a gasoline direct injector. SAE Int J Engines 2010;
[55] Nilges P, Schröder U. Electrochemistry for biofuel generation: production of 3(1):402–25.
furans by electrocatalytic hydrogenation of furfurals. Energy Environ Sci 2013;6 [84] Sementa P, Vaglieco BM, Catapano F. Thermodynamic and optical
(10):2925–31. characterizations of a high performance GDI engine operating in homogeneous
[56] Zu Y, et al. Efficient production of the liquid fuel 2, 5-dimethylfuran from 5- and stratified charge mixture conditions fueled with gasoline and bio-ethanol.
hydroxymethylfurfural over Ru/Co3O4 catalyst. Appl Catal B Environ 2014;146: Fuel 2012;96:204–19.
244–8. [85] Ma X, Jiang C, Xu H, Richardson S., In-cylinder optical study on combustion of
[57] Gao Z, Fan G, Liu M, Yang L, Li F. Dandelion-like cobalt oxide microsphere- DMF and DMF fuel blends, SAE Technical Paper; 2012. doi: https://doi.org/1
supported RuCo bimetallic catalyst for highly efficient hydrogenolysis of 5- 0.4271/2012-01-1235.
hydroxymethylfurfural. Appl Catal B Environ 2018;237:649–59. [86] Xu N, Gong J, Huang Z. Review on the production methods and fundamental
combustion characteristics of furan derivatives. Renew Sustain Energy Rev 2016;
54:1189–211.

27
A.T. Hoang et al. Fuel 285 (2021) 119140

[87] Wu X, Daniel R, Tian G, Xu H, Huang Z, Richardson D. Dual-injection: The [118] Murakami Y, Oguchi T, Hashimoto K, Nosaka Y. Theoretical study of the benzyl+
flexible, bi-fuel concept for spark-ignition engines fuelled with various gasoline O2 reaction: kinetics, mechanism, and product branching ratios. J Phys Chem A
and biofuel blends. Appl Energy 2011;88(7):2305–14. 2007;111(50):13200–8.
[88] Daniel R, Tian G, Xu H, Wyszynski ML, Wu X, Huang Z. Effect of spark timing and [119] Wu X, Huang Z, Yuan T, Zhang K, Wei L. Identification of combustion
load on a DISI engine fuelled with 2, 5-dimethylfuran. Fuel 2011;90(2):449–58. intermediates in a low-pressure premixed laminar 2, 5-dimethylfuran/oxygen/
[89] Tian G, Daniel R, Li H, Xu H, Shuai S, Richards P. Laminar burning velocities of 2, argon flame with tunable synchrotron photoionization. Combust Flame 2009;156
5-dimethylfuran compared with ethanol and gasoline. Energy Fuels 2010;24(7): (7):1365–76.
3898–905. [120] Somers KP, et al. A comprehensive experimental and detailed chemical kinetic
[90] Wu X, Huang Z, Jin C, Wang X, Wei L. Laminar burning velocities and Markstein modelling study of 2, 5-dimethylfuran pyrolysis and oxidation. Combust Flame
lengths of 2, 5-dimethylfuran-air premixed flames at elevated temperatures. 2013;160(11):2291–318.
Combust Sci Technol 2010;183(3):220–37. [121] Xu N, et al. Experimental study of 2, 5-dimethylfuran and 2-methylfuran in a
[91] Bhattacharya A, Datta A. Effects of blending 2, 5-dimethylfuran on the laminar rapid compression machine: Comparison of the ignition delay times and reactivity
burning velocity and ignition delay time of isooctane/air mixture. Combust at low to intermediate temperature. Combust Flame 2016;168:216–27.
Theory Model 2019;23(1):105–26. [122] Leplat N, Dagaut P, Togbé C, Vandooren J. Numerical and experimental study of
[92] Ma X, Jiang C, Xu H, Ding H, Shuai S. Laminar burning characteristics of 2- ethanol combustion and oxidation in laminar premixed flames and in jet-stirred
methylfuran and isooctane blend fuels. Fuel 2014;116:281–91. reactor. Combust Flame 2011;158(4):705–25.
[93] Hu E, Xu Z, Gao Z, Xu J, Huang Z. Experimental and numerical study on laminar [123] Shao C, et al. Polycyclic aromatic hydrocarbons in pyrolysis of gasoline surrogates
burning velocity of gasoline and gasoline surrogates. Fuel 2019;256:115933. (n-heptane/iso-octane/toluene). Proc Combust Inst 2019;37(1):993–1001.
https://doi.org/10.1016/j.fuel.2019.115933. [124] Tian G, Xu H, Daniel R. DMF-a new biofuel candidate. InTech; 2011.
[94] Wu X, et al. Measurements of laminar burning velocities and Markstein lengths of [125] Rönkkö T, Virtanen A, Kannosto J, Keskinen J, Lappi M, Pirjola L. Nucleation
2, 5-dimethylfuran− air− diluent premixed flames. Energy Fuels 2009;23(9): mode particles with a nonvolatile core in the exhaust of a heavy duty diesel
4355–62. vehicle. Environ Sci Technol 2007;41(18):6384–9.
[95] Tran L-S, Wang Z, Carstensen H-H, Hemken C, Battin-Leclerc F, Kohse- [126] Wang C, Xu H, Herreros JM, Lattimore T, Shuai S. Fuel effect on particulate
Höinghaus K. Comparative experimental and modeling study of the low-to matter composition and soot oxidation in a direct-injection spark ignition (DISI)
moderate-temperature oxidation chemistry of 2, 5-dimethylfuran, 2-methylfuran, engine. Energy Fuels 2014;28(3):2003–12.
and furan. Combust Flame 2017;181:251–69. [127] Wang F, Zheng ZL, He ZW. A soot precursor formation embedded reaction
[96] Bechtold JK, Matalon M. The dependence of the Markstein length on mechanism of diesel surrogate fuel. Energy Sources Part A Recover Util Environ
stoichiometry. Combust Flame 2001;127(1–2):1906–13. Eff 2015;37(12):1323–31.
[97] Bradley D, Hicks RA, Lawes M, Sheppard CGW, Woolley R. The measurement of [128] X. He, X. Ma, F. Wu, J. Wang, and S. Shuai, “Investigation of soot formation in
laminar burning velocities and Markstein numbers for iso-octane–air and iso- laminar diesel diffusion flame by two-color laser induced incandescence,” SAE
octane–n-heptane–air mixtures at elevated temperatures and pressures in an Technical Paper; 2008.
explosion bomb. Combust Flame 1998;115(1–2):126–44. [129] Togbé C, et al. Combustion chemistry and flame structure of furan group biofuels
[98] Benson RS, Whitehouse ND. Internal combustion engines: a detailed introduction using molecular-beam mass spectrometry and gas chromatography–Part III: 2, 5-
to the thermodynamics of spark and compression ignition engines, their design Dimethylfuran. Combust Flame 2014;161(3):780–97.
and development, vol. 1. Elsevier; 2013. [130] Ma X, Xu H, Jiang C, Shuai S. Ultra-high speed imaging and OH-LIF study of DMF
[99] Aleiferis PG, Malcolm JS, Todd AR, Cairns A, Hoffmann H. An optical study of and MF combustion in a DISI optical engine. Appl Energy 2014;122:247–60.
spray development and combustion of ethanol, iso-octane and gasoline blends in [131] Alexandrino K, Salvo P, Millera Á, Bilbao R, Alzueta MU. Influence of the
a DISI engine, SAE Technical Paper; 2008. doi: https://doi.org/10.4271/2008-01 temperature and 2, 5-dimethylfuran concentration on its sooting tendency.
-0073. Combust Sci Technol 2016;188(4–5):651–66.
[100] Cooney C, Worm J, Naber J. The calculation of mass fraction burn of ethanol- [132] Donkerbroek AJ, Boot MD, Luijten CCM, Dam NJ, Ter Meulen JJ. Flame lift-off
gasoline blended fuels using single and two-zone models, SAE Technical Paper; length and soot production of oxygenated fuels in relation with ignition delay in a
2008. doi: https://doi.org/10.4271/2008-01-0320. DI heavy-duty diesel engine. Combust Flame 2011;158(3):525–38.
[101] Wei M, Li S, Xiao H, Guo G. A Comparison Study on the Combustion and [133] Pepiot-Desjardins P, Pitsch H, Malhotra R, Kirby SR, Boehman AL. Structural
Particulate emissions of 2, 5-dimethylfuran/diesel and ethanol/diesel in a diesel group analysis for soot reduction tendency of oxygenated fuels. Combust Flame
engine. Therm Sci 2018;22(3):1351–61. 2008;154(1–2):191–205.
[102] Tanaka K, et al. Ignition characteristics of 2, 5-dimethylfuran compared with [134] Khan O, Yadav AK, Khan ME, Parvez M. Characterization of bioethanol obtained
gasoline and ethanol. SAE Int J Engines 2016;9(1):39–46. from Eichhornia Crassipes plant; its emission and performance analysis on CI
[103] Eldeeb MA, Akih-Kumgeh B. Investigation of 2, 5-dimethyl furan and iso-octane engine, Energy Sources, Part A Recover. Util. Environ. Eff., pp. 1–11, 2019.
ignition. Combust Flame 2015;162(6):2454–65. [135] Jiang B, Wang P, Ying Y, Luo M, Liu D. Nanoscale characteristics and reactivity of
[104] Roy S, Zare S, Askari O. Understanding the effect of oxygenated additives on nascent soot from n-heptane/2, 5-dimethylfuran inverse diffusion flames with/
combustion characteristics of gasoline. J Energy Resour Technol 2019;141(2). without magnetic fields. Energies 2018;11(7):1698.
https://doi.org/10.1115/1.4041316. [136] Peña GDJG, Hammid YA, Raj A, Stephen S, Anjana T, Balasubramanian V. On the
[105] Van Basshuysen R, Schäfer F. Internal combustion engine handbook-basics, characteristics and reactivity of soot particles from ethanol-gasoline and 2, 5-
components, systems and perspectives, vol. 345; 2004. dimethylfuran-gasoline blends. Fuel 2018;222:42–55.
[106] Hu C, Song X, Liu N, Li W. Investigation on cold starting and warming up of [137] Liu D, et al. Combustion chemistry and flame structure of furan group biofuels
gasoline engines with EFI, SAE Technical Paper; 2007. using molecular-beam mass spectrometry and gas chromatography–Part I: Furan.
[107] Wang C, et al. Combustion characteristics and emissions of 2-methylfuran Combust Flame 2014;161(3):748–65.
compared to 2, 5-dimethylfuran, gasoline and ethanol in a DISI engine. Fuel 2013; [138] Wang L, Song C, Song J, Lv G, Pang H, Zhang W. Aliphatic C-H and oxygenated
103:200–11. surface functional groups of diesel in-cylinder soot: characterizations and impact
[108] Daniel R, Tian G, Xu H, Shuai S. Ignition timing sensitivities of oxygenated on soot oxidation behavior. Proc Combust Inst 2013;34(2):3099–106.
biofuels compared to gasoline in a direct-injection SI engine. Fuel 2012;99:72–82. [139] Gogoi B, et al. Effects of 2, 5-dimethylfuran addition to diesel on soot
[109] Zhong S, et al. Combustion and emissions of 2, 5-dimethylfuran in a direct- nanostructures and reactivity. Fuel 2015;159:766–75.
injection spark-ignition engine. Energy Fuels 2010;24(5):2891–9. [140] Peña GDJG, et al. Effects of methyl group on aromatic hydrocarbons on the
[110] Shukla MK, Singh E, Singh N, Singal SK. Prospects of 2, 5-dimethylfuran as a fuel: nanostructures and oxidative reactivity of combustion-generated soot. Combust
physico-chemical and engine performance characteristics evaluation. J Mater Flame 2016;172:1–12.
cycles waste Manag 2015;17(3):459–64. [141] Peña GDJG, et al. Physicochemical properties of soot generated from toluene
[111] Liu H, et al. Investigation on blending effects of gasoline fuel with N-butanol, diffusion flames: effects of fuel flow rate. Combust Flame 2017;178:286–96.
DMF, and ethanol on the fuel consumption and harmful emissions in a GDI [142] Russo C, D’Anna A, Ciajolo A, Sirignano M. Analysis of the chemical features of
Vehicle. Energies 2019;12(10):1845. https://doi.org/10.3390/en12101845. particles generated from ethylene and ethylene/2, 5 dimethyl furan flames.
[112] Hoang AT, Pham VV. A study of emission characteristic, deposits, and lubrication Combust Flame 2016;167:268–73.
oil degradation of a diesel engine running on preheated vegetable oil and diesel [143] Conturso M, Sirignano M, D’Anna A. Effect of 2, 5-dimethylfuran doping on
oil. Energy Sources Part A Recover Util Environ Eff 2019;41(5):611–25. particle size distributions measured in premixed ethylene/air flames. Proc
[113] Hoang AT, Le AT, Pham VV. A core correlation of spray characteristics, deposit Combust Inst 2017;36(1):985–92.
formation, and combustion of a high-speed diesel engine fueled with Jatropha oil [144] Song J, Alam M, Boehman AL. Impact of alternative fuels on soot properties and
and diesel fuel. Fuel 2019, 2019,;244(15):159–75. DPF regeneration. Combust Sci Technol 2007;179(9):1991–2037.
[114] Hoang AT. “Critical review on the characteristics of performance, combustion and [145] Yehliu K, Vander Wal RL, Armas O, Boehman AL. Impact of fuel formulation on
emissions of PCCI engine controlled by early injection strategy based on narrow- the nanostructure and reactivity of diesel soot. Combust Flame 2012;159(12):
angle direct injection (NADI)”, Energy Sources. Part A Recover Util Environ Eff 3597–606.
2020. https://doi.org/10.1080/15567036.2020.1805048. [146] Tran L-S, Sirjean B, Glaude P-A, Kohse-Höinghaus K, Battin-Leclerc F. Influence of
[115] Bhasker JP, Porpatham E. Effects of compression ratio and hydrogen addition on substituted furans on the formation of Polycyclic Aromatic Hydrocarbons in
lean combustion characteristics and emission formation in a Compressed Natural flames. Proc Combust Inst 2015;35(2):1735–43.
Gas fuelled spark ignition engine. Fuel 2017;208:260–70. [147] Daniel R, Wang C, Xu H, Tian G. Effects of combustion phasing, injection timing,
[116] Djokic M, Carstensen H-H, Van Geem KM, Marin GB. The thermal decomposition relative air-fuel ratio and variable valve timing on SI engine performance and
of 2, 5-dimethylfuran. Proc Combust Inst 2013;34(1):251–8. emissions using 2, 5-dimethylfuran. SAE Int J Fuels Lubr 2012;5(2):855–66.
[117] Tian Z, et al. An experimental and kinetic investigation of premixed furan/
oxygen/argon flames. Combust Flame 2011;158(4):756–73.

28
A.T. Hoang et al. Fuel 285 (2021) 119140

[148] Friese P, Simmie JM, Olzmann M. The reaction of 2, 5-dimethylfuran with [156] Qiongjie W, Yufu X, Xianguo H, Xifeng Z. Experimental study on friction and wear
hydrogen atoms–An experimental and theoretical study. Proc Combust Inst 2013; characteristics of bio-oil. Trans Chinese Soc Agric Eng 2008;9:2008.
34(1):233–9. [157] Dearn KD, et al. The tribology of fructose derived biofuels for DISI gasoline
[149] Wang C, Prakash A, Aradi A, Cracknell R, Xu H. Significance of RON and MON to engines. Fuel 2018;224:226–34.
a modern DISI engine. Fuel 2017;209:172–83. [158] Eslami F, Wyszyński ML, Tsolaskis A, Xu H, Norouzi S, Dearn K. Experimental
[150] Tiunov IA, Kotelev MS, Vinokurov VA, Gushchin PA, Bardin ME, Novikov AA. investigation on lubricity of 2, 5-dimethylfuran blends. Silniki Spalinowe 2012;
Antiknock properties of blends of 2-methylfuran and 2, 5-dimethylfuran with 51:3–10.
reference fuel. Chem Technol Fuels Oils 2017;53(2):147–53. [159] Hu E, Dearn K, Yang B, Song R, Xu Y, Hu X. Tribofilm formation and
[151] Nisbet HB. The blending octane numbers of 2, 5-dimethylfuran. J Inst Pet 1946; characterization of lubricating oils with biofuel soot and inorganic fluorides.
32:162–6. Tribol Int 2017;107:163–72.
[152] B. MT, S. DJH, S. DG, Fuel composition. European Patent 0082689 A2; 1983. [160] Hoang AT, Tabatabaei M, Aghbashlo M. A review of the effect of biodiesel on the
[153] Sarathy SM, Oßwald P, Hansen N, Kohse-Höinghaus K. Alcohol combustion corrosion behavior of metals/alloys in diesel engines. Energy Sources Part A
chemistry. Prog Energy Combust Sci 2014;44:40–102. Recover Util Environ Eff 2019;42(23):2923–43.
[154] Voice AK, Tzanetakis T, Traver M. Lubricity of Light-End Fuels with Commercial [161] Bekal S, Bhat NR. Bio-lubricant as an alternative to mineral oil for a CI engine—an
Diesel Lubricity Additives, SAE Technical Paper; 2017. experimental investigation with pongamia oil as a lubricant. Energy Sources Part
[155] Hu E, Hu X, Wang X, Xu Y, Dearn KD, Xu H. On the fundamental lubricity of 2, 5- A Recover Util Environ Eff 2012;34(11):1016–26.
dimethylfuran as a synthetic engine fuel. Tribol Int 2012;55:119–25.

29

You might also like