City Research Online

City, University of London Institutional Repository

Citation: Arcoumanis, C., Bae, C., Crookes, R. and Kinoshita, E. (2008). The potential of
di-methyl ether (DME) as an alternative fuel for compression-ignition engines: A review. Fuel,
87(7), pp. 1014-1030. doi: 10.1016/j.fuel.2007.06.007

This is the accepted version of the paper.

This version of the publication may differ from the final published
version.

Permanent repository link: http://openaccess.city.ac.uk/15549/

Link to published version: http://dx.doi.org/10.1016/j.fuel.2007.06.007

Copyright and reuse: City Research Online aims to make research

outputs of City, University of London available to a wider audience.
Copyright and Moral Rights remain with the author(s) and/or copyright
holders. URLs from City Research Online may be freely distributed and
linked to.

City Research Online: http://openaccess.city.ac.uk/ [email protected]

 The potential of di-methyl ether (DME) as an alternative fuel
for compression-ignition engines: A review
a,* b,1 c,2 d,3
Constantine Arcoumanis , Choongsik Bae , Roy Crookes , Eiji Kinoshita
a
School of Engineering and Mathematical Sciences, The City University London, Northampton Square, London EC1V 0HB, United Kingdom
b
Department of Mechanical Engineering, Korea Advanced Institute of Science and Technology, 373-1 Kusong-Dong, Yusong-Gu,
Taejon 305-701, Republic of Korea
c
Department of Engineering, Queen Mary, University of London, Mile End Road, London E1 4NS, United Kingdom
d
Department of Mechanical Engineering, University of Kagoshima, 1-21-40 Korimoto, Kagoshima 890-0065, Japan

Abstract

This paper reviews the properties and application of di-methyl ether (DME) as a candidate fuel for compression-ignition engines.
DME is produced by the conversion of various feedstock such as natural gas, coal, oil residues and bio-mass. To determine the technical
feasibility of DME, the review compares its key properties with those of diesel fuel that are relevant to this application. DME’s diesel
engine-compatible properties are its high cetane number and low auto-ignition temperature. In addition, its simple chemical structure
and high oxygen content result in soot-free combustion in engines. Fuel injection of DME can be achieved through both conventional
mechanical and current common-rail systems but requires slight modification of the standard system to prevent corrosion and overcome
low lubricity. The spray characteristics of DME enable its application to compression-ignition engines despite some differences in its
properties such as easier evaporation and lower density. Overall, the low particulate matter production of DME provides adequate jus-
tification for its consideration as a candidate fuel in compression-ignition engines. Recent research and development shows comparable
output performance to a diesel fuel led engine but with lower particulate emissions. NOx emissions from DME-fuelled engines can meet
future regulations with high exhaust gas recirculation in combination with a lean NOx trap. Although more development work has
focused on medium or heavy-duty engines, this paper provides a comprehensive review of the technical feasibility of DME as a candidate
fuel for environmentally-friendly compression-ignition engines independent of size or application.

Abbreviations: BSFC, brake specific fuel consumption; BTX, benzene, toluene and xylene; C–C, carbon-to-carbon; CFC’s, chloro-fluoro-carbons; C:H,
carbon-to-hydrogen ratio; C2H2, acetylene; C2H4, ethylene; C3H3, proparagyl; CH4, methane; CH2O, formaldehyde; CI, compression-ignition; CNG,
compressed natural gas; CO2, carbon dioxide; CO, carbon monoxide; CR, compression ratio; DI, direct-injection; DMC, di-methyl carbonate; DME, di-
methyl ether; EGR, exhaust gas recirculation; FIE, fuel-injection equipment; H2, hydrogen; HC, hydrocarbon; HCCI, homogeneous charge compression-
ignition; HDV, heavy-duty vehicle; IMEP, indicated mean effective pressure; LDV, light-duty vehicle; LNT, lean NOx trap; LPG, liquefied petroleum gas;
NMHC, non-methane hydrocarbon; NOx, nitrogen oxide; Pa, ambient pressure; Pin, fuel- injection pressure in common-rail; PAH, polycyclic aromatic
hydrocarbon; PM, particulate matter; PTFE, polytetrafluoroethylene; SI, spark-ignition; SOF, soluble organic fraction; SO2, sulfur dioxides; SPI,
sustainable process index; syngas, synthetic gas; ULEV, ultra low emission vehicle.
*
Corresponding author. Tel.: +44 207 040 0113; fax: +44 207 040 8101.
E-mail addresses: [email protected] (C. Arcoumanis), [email protected] (C. Bae), [email protected] (R. Crookes), kinoshit@mech.
kagoshima-u.ac.jp (E. Kinoshita).
1
Tel.: +82 42 869 3044; fax: +82 42 869 5044.
2
Tel.: +44 207 882 5270; fax: +44 208 983 1007.
3
Tel.: +81 99 285 8268; fax: +81 99 250 3181.
Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1015
2. DME properties. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1016
2.1. Advantages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1016
2.2. Disadvantages. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1017
3. Fuel specification and production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1017
3.1. Fuel specification of DME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1017
3.2. DME production . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1017
4. Fuel-injection system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1018
4.1. Features of the fuel-injection system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1018
4.2. Flow phenomena in the injector nozzle . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1018
4.3. DME fuel-injection system options . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1019
4.3.1. Common-rail system . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1019
4.4. A liquid DME fuel tank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1020
5. Spray characteristics and combustion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1020
5.1. Spray characteristics . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1020
5.1.1. Pa < saturated vapour pressure of DME or low ambient pressure (0.1 MPa) [19,23,27] . . . . . . . . . . . . . . 1020
5.1.2. Saturated vapour pressure of DME < Pa < critical pressure of DME [10,11,27] . . . . . . . . . . . . . . . . . . . 1021
5.1.3. Pa > critical pressure of DME [10,11] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021
5.2. Spray combustion of DME [26–29] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021
5.3. Numerical simulation of DME spray combustion. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1021
6. Engine performance and emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1022
6.1. Exhaust emissions and fuel consumption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1022
6.1.1. Particulate matter (PM) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1022
6.1.2. NOx . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1023
6.1.3. HC and CO . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1024
6.1.4. Combustion noise [3,15] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025
6.1.5. Non-regulated exhaust emissions. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025
6.2. Effect of fuel-injection equipment (FIE) and combustion system parameters. . . . . . . . . . . . . . . . . . . . . . . . . . . . 1025
7. Energy efficiency and well-to-wheels emissions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1026
7.1. Well-to-wheels analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1026
7.2. Assessment of emissions by the sustainable process index. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1026
8. New DME concept engine . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1027
9. Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1028
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1029

1. Introduction DME-fuelled vehicles is projected to be 22.5% and 19%,

respectively [4]. This is comparable to LPG and com-
Di-methyl ether (DME) is a liquified gas with handling pressed natural gas (CNG) fuelled vehicles but less than
characteristics similar to those of liquified petroleum gas the highest energy efficiency of 26% in direct-injection
(LPG) [1]. It can be produced from a variety of feed-stock (DI) diesel fuelled vehicles [4]. On the other hand, the
such as natural gas, crude oil, residual oil, coal, waste prod- well-to-wheels carbon dioxide (CO2) emissions of a
ucts and bio-mass. Many investigations have been carried DME-fuelled vehicle is comparable to those using DI diesel
out on DME to determine its suitability for use as a fuel or CNG fuelled engines [4]. However, an oxidation catalyst
in diesel-cycle engines [1,2]. would be necessary to meet ULEV carbon monoxide (CO)
DME has the appearance of an excellent, efficient and hydrocarbon (HC) emission limits [5].
alternative fuel for use in a diesel engine, with almost DME was also found to be an excellent gas turbine fuel
smoke-free combustion. This is not only because of its with emission properties comparable to natural gas [6].
low auto-ignition temperature and its almost instantaneous DME-fuelled turbine also allows to achieve a significant
vapourization when injected into the cylinder, but also performance improvement through thermochemical recu-
because of its high oxygen content (around 35% by mass) peration with 44% higher power output and an 8% decrease
and the absence of C–C bonds in the molecular structure of the specific CO2 emissions compared to the present plant
[1,2]. With a properly designed DME injection and com- [7]. However, DME is not a suitable fuel for spark-ignition
bustion system, nitrogen oxides (NOx) emissions can also (SI) engines due to its high cetane number, though the
meet ultra low emission vehicle (ULEV) limits [3]. The burning velocity is similar to hydrocarbon fuels [8]; the eas-
well-to-wheels energy efficiency of heavy- and light-duty ily-induced knock would limit the operation of SI engines.
1016

Overall, the key to the development of a dedicated low-

emissions, DME-fuelled engine is the performance and
durability of its fuel-injection system [1,2]. In this review,
the features considered to be most important in developing
the potential for widespread utilization of DME to recipro-
cating engines are: the production and properties of DME,
the fuel-injection system and the spray characteristics
which contribute to the engine’s performance and exhaust
emissions.

2. DME properties
Fig. 1. Structure of di-methyl ether (DME).
The key properties of DME and diesel fuel are shown in
Table 1 [9,10]. In addition to the advantages above, it has a DME is non-toxic and environmentally benign. At low
low carbon-to-hydrogen ratio (C:H) with a chemical for- mole fractions (a few percent by volume) the gas has hardly
mula of CH3–O–CH3 as shown in Fig. 1. DME in a gas- any odour and causes no negative health effects. Even at
eous state is invisible under standard atmospheric high mole fractions (>10% by volume) there is no effect
conditions (0.1 MPa at 298 K). When it is pressurized on human health except narcotic effects after long exposure
above 0.5 MPa, it condenses to the liquid phase. Gaseous and it may be recognized by the odour. DME displays a
DME is denser than air while liquid DME has a density visible blue flame when burning over a wide range of air–
two thirds that of water. The vapour pressure is similar fuel ratios, similar to natural gas, which is an important
to that of LPG and requires the same handling and storage safety characteristic. Nevertheless, the operation of DME
precautions. It dissolves in water up to 6% by mass. How- combustion system needs the adoption of rigorous proce-
ever, it is not compatible with most elastomers due to its dures for safe operation due to the wide flammability
corrosiveness, so that careful selection of materials is neces- limits.
sary to prevent deterioration of seals after prolonged expo- The importance of the properties of DME can be iden-
sure to DME. The sealing of DME filled vessels, for tified by examining its advantages and disadvantages as a
example, can be achieved with inert material, such as poly- candidate fuel for replacing diesel fuel in compression-igni-
tetrafluoroethylene (PTFE). tion (CI) engines.

Table 1
2.1. Advantages
Properties of DME and diesel fuel
Property (unit/condition) Unit DME Diesel fuel
(i) High oxygen content: Together with the absence of
any C–C bonds it is responsible for its smokeless
Chemical structure CH3–O– –
CH3
combustion; low formation and high oxidation rates
Molar mass g/mol 46 170 of particulates would therefore be expected under
Carbon content mass% 52.2 86 compression-ignition (CI) engine operation.
Hydrogen content mass% 13 14 (ii) Low boiling point: Leads to quick evaporation when a
Oxygen content mass% 34.8 0 liquid-phase DME spray is injected into the engine
Carbon-to-hydrogen ratio 0.337 0.516
Critical temperature K 400 708
cylinder.
Critical pressure MPa 5.37 3.00a (iii) High cetane number: Results from the low auto-igni-
Critical density kg/m3 259 – tion temperature and almost instantaneous vapouri-
Liquid density kg/m3 667 831 zation. Teng et al. [11] examined the high cetane
Relative gas density (air = 1) 1.59 – number (>55) of DME from the viewpoint of its ther-
Cetane number >55 40–50
Auto-ignition temperature K 508 523
mochemical characteristics. Its critical temperature,
Stoichiometric air/fuel mass 9.0 14.6 400 K (127 °C) is lower than the compressed air tem-
ratio perature at the later stages of the compression stroke,
Boiling point at 1 atm K 248.1 450–643 which allows the DME injected into the cylinder to
Enthalpy of vapourization kJ/kg 467.13 300 evaporate immediately. When the temperature of
Lower heating value MJ/kg 27.6 42.5
Gaseous specific heat capacity kJ/kg K 2.99 1.7
DME is higher than 400 K, it becomes superheated
Ignition limits vol% in 3.4/18.6 0.6/6.5 vapour and no evaporation is associated with the
air mixing. The chain combustion reaction, however,
Modulus of elasticity N/m2 6.37E+08 14.86E+08 may be initiated by either (or both) of two competing
Kinematic viscosity of liquid cSt <.1 3 path ways [12,13]:
Surface tension (at 298 K) N/m 0.012 0.027
Vapour pressure (at 298 K) kPa 530 10
(a) C–O bond fission (pyrolysis mechanism):
a
Estimated on the basis of the equivalent chemical formula. CH3 OCH3 ¼ CH3 O þ CH3 ð1Þ
 1017

(b) Hydrogen abstraction (oxidation mechanism): like natural gas and renewable sources like wood. Its pro-
duction method most closely resembles that of methanol.
4CH3 OCH3 þ O2 ¼ 4CH3 OCH2 þ 2H2 O ð2Þ
DME production methods are of the following type
CH3 OCH2 ¼ CH2 O þ CH3 ð3Þ [1,2,15]

The C–O bond energy is smaller than that of the C–H (a) De-hydrogenation of methanol
bond (414 kJ/mol [14]) and the distortion of the C–O (b) Direct conversion from synthesis gas (syngas)
bonds in the DME molecule weakens the bonding strength;
thus, the C–O bond breaks easier than the C–H bond. which have essentially similar features.
Therefore, the pyrolysis mechanism may be more able to Currently, most DME is produced by de-hydrogenation
start the chain reaction at relatively low temperatures, of methanol from syngas which is a mixture of hydrogen
which would result in the low auto-ignition temperature. (H2), carbon monoxide (CO) and carbon dioxide (CO2),
The physical delay of DME is thus much shorter than that but may also contain impurities like methane (CH4).
for many conventional fuels, resulting in a shorter total The direct conversion method is the simultaneous pro-
ignition delay and a higher cetane number. duction of DME and methanol from syngas using appro-
priate catalysts. The first step of direct DME production
2.2. Disadvantages is the conversion of the feedstock to syngas, most common
by steam reforming of natural gas and for coal, oil residues
(i) Low combustion enthalpy: Lower than that of diesel and bio-mass by partial oxidation through gasification with
fuel due to the oxygen content of the molecules, pure oxygen. The second step for the direct route is via
which necessitates a larger injected volume and methanol synthesis using a copper-based catalyst while
longer injection period for DME in order to deliver the third step is the de-hydrogenation of methanol to
the same amount of energy to that provided by diesel. DME using alumina- or zeolite-based catalysts. With the
(ii) Low viscosity: Lower than that of diesel fuel, causing direct conversion of syngas to DME, the second and third
leakage from the fuel supply system which relies on steps can occur simultaneously in one reactor using appro-
small clearances for sealing. Its lower lubricity char- priate catalysts (different to those described above). The
acteristics can cause intensified surface wear of mov- final step is the purification of the raw product, which
ing parts within the fuel-injection system which may also contain some methanol and water.
requires attention. Although it is possible to produce DME from both fos-
(iii) Low modulus of elasticity: Lower than that of diesel sil and renewable sources, bio-DME is very attractive;
fuel, with the difference being greatest at lower pres- however, producing DME from fossil sources is the only
sures. Thus, the compressibility of DME is usually viable route at present from a cost perspective since bio-
four to six times higher than that of diesel in a closed DME is very expensive relative to diesel fuel. Because nat-
system and the compression work of the fuel pump ural gas at remote locations will, more likely, be the lowest-
for DME is greater than it is for diesel fuel by up cost feedstock [4], the most cost-effective route for DME
to 10% in an open system [10]. production is by direct synthesis of syngas (possibly with
methanol as an intermediate step or co-product), from nat-
3. Fuel specification and production ural gas. From both the cost and product-yield points of
view, the synthesis of DME directly from syngas may be
3.1. Fuel specification of DME the preferred route for large scale production. It has been
claimed that DME produced from natural gas wells in
Recommendations have been made concerning permissi- Southeast Asia or the Middle East is economical enough
ble levels of impurities present in DME [2]. Most impurities [16] to be supplied at a lower price than diesel by up to
are recommended to be negligible such that the propor- 40% [17].
tions in mass are less than 0.05% for methanol, higher alco- Increasing the purity requirements of the final product
hols, higher ethers and ketones. Water is recommended to will increase production costs of DME. Nevertheless, it is
be less than 0.01% while methyl ethyl ether should be less still considered that DME, with its very low percentages
than 0.2%. An odorant such as ethylmercaptane can be of methanol and water, should not exhibit problems when
included up to 20 ppm. used in internal combustion engines.
Flexible plants can be set-up to produce methanol and
3.2. DME production DME in different ratios, depending on the relative demand,
so that a large production capacity can be realized in a lim-
DME has been produced world wide in quantities of ited amount of time. Converting a methanol plant to pro-
100,000–150,000 tonnes per annum from 1996 or earlier duce DME, besides building new DME facilities, is an
[1,2], and is used as a propellant for spray cans for cosmet- attractive option for DME production. DME can also be
ics instead of chloro-fluoro-carbons (CFC’s). It can be pro- produced in a coal-fired power plant and in plants close
duced from carbonaceous feedstock, both from fossil fuels to urban or industrial areas, where waste-heat from the
1018

production process can be used for heating or generating To prevent this, most sealing material made of con-
hot water. ventional elastomers should be replaced with anti-
corrosive material such as Teflon-coated o-rings or
4. Fuel-injection system PTFE based high-tension sealing components, if
there is an expectation for DME use in fuel-injection
The features of a fuel-injection system for DME and the equipment to reach mass production [19].
characteristics of the flow in the injector nozzle are different (v) Low lubricity: The lower lubricity of DME than that
from those for diesel fuel due to the different physical and of diesel fuel leads to wear problems. The lubricity
chemical properties. In what follows these features and a can be enhanced with up to 2000 ppm of a lubricity
suggested appropriate fuel-injection system for DME are additive; examples include Lubrizol (1000 ppm) [12],
summarized. Hitec 560 (100 ppm) [9] and Infineum R655
(500 ppm) [15]. Even diesel fuel, di-methyl carbonate
4.1. Features of the fuel-injection system (DMC) and fatty acids have been suggested as viscos-
ity enhancing additives [20,21]. However, it might be
(i) Closed pressurized fuel system: Because of the low possible to adapt the materials of the bearing surfaces
boiling point of DME (248 K = 25 °C), it is a gas and plunger/barrel interfaces to the low lubricity of
under standard atmospheric conditions and therefore DME [1]. It is expected that the experience gained
must be pressurized in a fuel system, including a stor- in the past with the use of methanol in fuel-injection
age tank, and handled like a liquefied gas. Thus, the systems will prove useful in developing a leakage-free
low boiling point of DME necessitates a closed pres- and durable injection system for DME operating at
surized fuel system. The vapour pressure of DME, injection pressures not exceeding 30 MPa [1], while
roughly the same as LPG, demands the same kind current common-rail systems for diesel fuel could be
of handling and storage considerations as for LPG. adapted for DME with the use of lubricity additives
(ii) Cavitation in the fuel-injection system: The high at injection pressures up to 55 MPa [19].
vapour pressure of DME means that cavitation, (vi) Long injection period: The low liquid density and low
which prevents stable fuel-injection operation, is calorific value require a higher volume of DME to be
unavoidable in the fuel-injection equipment [2]. injected into the cylinder, compared with that for die-
Although DME is kept in the liquid phase under sel fuel. In particular, 1.8 times the volume of diesel
atmospheric temperature conditions, when pressur- fuel is needed (to supply the same amount of energy)
ized to above 0.5 MPa the fuel-supply pressure (the which necessitates a longer injection period and
feed pressure from the storage tank to the fuel pump) advanced injection timing.
must be between 1.2 and 3 MPa to avoid cavitation in (vii) Large compression work: The compression work of
the fuel line. This is the case not only because of the the fuel pump for DME will always be larger than
higher temperatures during engine operation, when that for diesel fuel, since DME has lower density
the saturated vapour pressure of DME is higher, and higher compressibility than diesel [11].
but also due to dynamic flow effects leading to the (viii) Sealing material: DME is not compatible with most
formation of vapour zones in the fuel line [12]. elastomers and can chemically attack some com-
(iii) Low injection pressure: DME gasifies immediately monly used sealing materials and other plastic com-
during injection, due to its low boiling point, even ponents, raising questions about the durability of
though it is injected as a liquid. Therefore, the high injection systems handling DME. A careful selection
fuel injection pressures, such as 50–150 MPa, used of sealing materials is necessary to prevent deteriora-
in modern diesel injection systems are not required tion after prolonged exposure to DME. Sealing of
for DME and there is a consensus that 20–30 MPa DME-filled storage vessels and supply lines can, for
is sufficient as long as the required amount of DME example, be achieved with PTFE (as mentioned in
can be supplied during the corresponding injection Section 2).
period even at high engine speeds [18].
(iv) Leakage: Because of its low viscosity, currently avail- 4.2. Flow phenomena in the injector nozzle
able fuel-injection systems are not suitable for DME
due to leakage problems. Even at atmospheric condi- The physical properties of DME give rise to different
tions the leakage of DME can be significant in characteristics of the needle lift of the injector, of the pres-
regions of small clearance such as between the plun- sure oscillation in the fuel line, and of the flow phenomena
ger and barrel of a rotary-type fuel-injection system. in the injector nozzle compared to those for diesel fuel. The
Leakage rates of DME along the plungers, of up to flow in the nozzle is of prime importance with respect to
40–50% of the fuel, have been reported in the litera- the fuel-spray development and mixture preparation in
ture [12]. In heavy-duty engines, leakage along the the combustion chamber [9]. In particular flow cavitation
plungers is more serious than in light-duty engines in the nozzle orifice was found to play a crucial role in
where the drive mechanism is engine-oil lubricated. the formation of the emerging from the nozzle spray.
 1019

(i) Needle lift behaviour: In view of DME’s lower modu- by lowering the back pressure. In this case, complex
lus of elasticity compared to diesel, the needle lift pro- gas-dynamic processes lead to an increase in the exit
file of the conventional injector is expected to be angle of the discharging spray.
different. With the resulting higher compressibility
(relative to diesel) the nozzle opening stage is slower 4.3. DME fuel-injection system options
in the case of DME [22].
Kajitani et al. [23] suggested that the feed pressure is The two methods of developing a fuel-injection system
an important factor affecting the nozzle-opening time for DME involve (1) modification of conventional pump–
and that the nozzle opens earlier when the feed pres- pipe–nozzle fuel-injection systems or (2) new system with
sure is higher. With DME, the feed pressure in the a variable displacement pump or common-rail. In light-
fuel line is required to supply the fuel from the stor- duty vehicles, it is common to use either dual-fuel or
age tank to the fuel pump. Thus due to the higher retrofit engines by converting existing pump–pipe–nozzle
feed pressure of DME in the fuel line, the start of systems. However, in view of the high reliability require-
nozzle opening should be earlier for DME than for ments, this method is difficult to use and it is uncertain
diesel fuel. Considerable residual needle bouncing whether it can be applied in practice to light-duty vehicles
was also observed with the low nozzle-opening pres- [1,2]. State of the art common-rail diesel systems are also
sure (6.86 MPa), because of the reduced spring pre- promising for implementation in passenger car engines
load on the nozzle needle. This is in contrast to the operating with DME [18,22,23]; this is because these sys-
relatively well-defined injection behaviour at higher tems may require no more than 8 MPa injection pressure
opening pressures (8.82 MPa) and, as a result, it has to operate satisfactorily. In heavy-duty vehicles, however,
become necessary to optimize the injector specifica- new dedicated DME fuel-injection systems are under devel-
tions in order to prevent needle bouncing. opment based on the common-rail concept. For demon-
(ii) Pressure drop: In the case of DME, there is a consid- stration purposes it is probably more cost effective to
erable pressure drop across the valve seat due to the convert conventional fuel injection equipment [1]. On the
high fuel compressibility [22]. A loss of pressure other hand, Tsuchiya and Sato [18] found that an in-line
across the valve seat reduces the pressure drop across fuel-injection pump (Jerk type) with conventional spring-
the nozzle holes which, in turn, reduces the velocity of loaded needle injector was more suitable than a common-
the injected spray. rail system for heavy-duty DME engines due to the maxi-
(iii) Pressure oscillations and residual pressure in the injec- mum injection pressure characteristics.
tion line: After injection, DME exhibits larger pres- In order to meet future emissions legislation for vehicles,
sure oscillations and a higher residual pressure in it is necessary to develop a fuel-injection system that is able
the injection line between the fuel pump and the injec- to control injection timing, injection duration, quantity of
tor, compared to diesel fuel [10]. The reason for the injection and rate-shaping in combination with EGR sys-
larger pressure oscillations with DME is believed to tems. Injection rate-shaping which reduces the rate of injec-
be its higher compressibility compared to diesel. A tion during the ignition delay period, can decrease the rate
larger amount of energy is stored in the compressed of heat release just after ignition (‘premixed’ combustion)
DME than in the diesel fuel during the injection pro- as well as the peak firing pressure and temperature in the
cess which gives rise to larger pressure oscillations in cylinder, thus leading to NOx and noise reduction [5]. High
the injection line; these slowly decay between injec- EGR rates can also be employed due to the very low soot-
tions giving rise to the high residual pressure. producing capability of DME. It is considered that by opti-
(iv) Flow phenomena and flow rate in the nozzle orifice: mization of the injection system at high EGR rates, NOx
Kapus and Ofner [9], using model orifice experiments and PM can both be reduced simultaneously. For NOx
and numerical methods, described the flow in the reduction using a catalyst, post-injection after the main
model orifice for DME to be in the gas–liquid two- injection event is crucial for controlling the exhaust gas
phase regime. In contrast, the same analysis showed temperature and composition. This is important for new
diesel fuel to exhibit both liquid single-phase and generation diesel engines equipped with storage type De-
gas–liquid two-phase flow characteristics. NOx catalysts which necessitates development of flexible
For a high pressure drop between the inlet and the multi-injection systems capable of variable post-injection
outlet edge of the orifice, gas bubbles emerging from patterns.
cavitating locations with high fluid velocities and low
static pressures (mainly at the inlet edges of the con- 4.3.1. Common-rail system
traction) are mixed with the liquid to form an appar- A purpose-built common-rail fuel-injection system is
ently homogeneous zone of gas and liquid. Once this probably the best solution for a dedicated DME engine
zone covers the whole cross-sectional area of the ori- [1]. The independence of injection pressure on engine speed
fice, the flow characteristics become independent of gives a significant boost to the DI engine irrespective of the
the back pressure. This means that, for a given inlet fuel used and allows optimization of injection timing, injec-
pressure, the flow rate can not be increased further tion duration, rate of injection, and rate-shaping to be
1020

achieved. AVL has developed a DME common-rail fuel- separating the two fluids. The propane pressurized the
injection system with solenoid valves (which are three- DME into a sub-cooled-liquid state and functioned as a
way-valves) specifically for large truck and bus engines driving fluid that pumped the liquid DME from the tank
[24]. This system, which has leakage control with double to the injection pump by means of its vapour pressure.
seals and a purge system, has shown promising results for
engine performance, exhaust emissions and safety in 5. Spray characteristics and combustion
operation.
In order to accurately measure the rated injection of 5.1. Spray characteristics
DME and to investigate the possibility of rate-shaping of
the injection period, Ikeda et al. [25] connected an injection When it is injected into the engine cylinder, DME
rate meter (based on the Zeuch method) to a common-rail vapourizes immediately due to its low boiling point at
fuel-injection system. It was shown that, if the dimensions atmospheric pressure; as a result of its low critical temper-
of the measuring system are modified appropriately, the ature and critical pressure, the behaviour of liquid DME is
injection rate meter can be applied to the common-rail fuel sensitive to the in-cylinder thermo-dynamic conditions [11].
injection system to characterize the DME injection pattern. Investigations into the spray characteristics of DME have
The maximum injected quantity seemed to depend on the been carried out by various groups [9,10,19,23,27,28].
cavitation factor (number) in the case of DME. The cavita- Wakai et al. [27] performed experiments using a Schlieren
tion factor was defined as the pressure drop of the nozzle optical system on a DME spray injected by means of a die-
normalized by the inlet pressure minus the vapour pressure sel injector into a vessel in the low- to middle-pressure
[9,25]. They also suggested that multiple fuel-injection, range of 0.1–1.5 MPa. Sorenson et al. [10] studied high-
including pre- and post-injection, was only possible by pressure liquid DME injection into a nitrogen environment
the use of a common-rail fuel-injection system. Ikeda at temperatures <413 K (40 °C) and pressures of 1.5, 2.5,
et al. [25] calibrated and assessed the accuracy of the injec- 4.0 and 5.5 MPa, using a lubricity additive to the fuel. Teng
tion rate meter with multiple injection rates and quantities, et al. [11] analysed the hydrodynamic behaviour of a DME
including five-stage injection, and showed that the com- spray in a DI diesel engine cylinder by referring to the
mon-rail injection system was able to realize the required investigations carried out by Glensvig et al. [28] and Soren-
programmed injection accurately. Yu and Bae [19] also uti- son et al. [10]. Kajitani et al. [23] observed sprays of DME
lized commercial diesel common-rail and solenoid-driven at atmospheric pressure while Yu and Bae [19] character-
diesel injectors for DME injection tests with the help of ized the injection rate and spray development with a com-
pneumatic pumps to compress DME up to 55 MPa. They mon-rail system operating at pressures up to 55 MPa. The
measured DME injection rates from the solenoid injector experimental conditions of these investigations are shown
using the so-called Bosch tube method based on the in Table 2 and the spray behaviour of DME can be classi-
hydraulic pulse theory. fied according to the ambient pressure Pa as explained and
summarized below.
4.4. A liquid DME fuel tank
5.1.1. Pa < saturated vapour pressure of DME or low
Although the DME tank could be designed similarly to ambient pressure (0.1 MPa) [19,23,27]
those used for LPG, a low-pressure pump and cooling sys- When DME is injected into a lower ambient pressure
tem would be needed between the DME fuel tank and the than its saturated vapour pressure, the pressure of the
injection pump, further complicating the fuel design. To DME spray falls below the saturated vapour pressure
simplify the McCandless et al. [26] developed a liquid and flash boiling occurs. Several rugged clusters have been
DME fuel tank by considering the thermodynamic behav- observed at the peripheral region of the spray, when viewed
iour of DME. The fuel tank, of cast aluminum with a on a relatively large scale, considered to be the vapour
capacity of 40 l, was essentially a thermodynamic pump cloud. The tip penetration of the DME spray is less than
containing two fluids: liquid DME and a vapour–liquid or almost the same as that of a diesel fuel spray but a wider
mixture of driving fluid (mainly propane) with a diaphragm spray angle is observed with DME compared to diesel. The

Table 2
Experimental conditions of DME spray investigations
Investigators Wakai et al. [27] Sorenson et al. [10], Glensvig et al. Yu and Bae [19]
[28], Teng et al. [9]
Injection system In-line type Bosch injection pump, In-line pump, single-hole-nozzle and Common-rail, solenoid-driven injector
single-hole-nozzle pintle-nozzle five-hole mini-sac nozzle
Nozzle diameter (mm) 0.2 0.513 0.168
Valve opening – or common-rail 9.8 13 at Pa = 1.5 and 2.5, 20 at 25, 40 and 55 (common-rail pressure)
pressure (MPa) Pa = 4.0 and 5.0
Ambient pressure (MPa) 0.1 and 1.5 1.5, 2.5, 4.0 and 5.5 0.1 and 3.0
Ambient temperature (K) (°C) 293 K (20 °C) 411 K (38 °C) <413 K (40 °C)
 1021

increased spread of the DME spray (wider spray angle and DME has a higher gas-to-liquid density ratio in the spray,
larger spray volume) is thought to be mainly due to the which also tends to increase the spray angle. Because of the
enhancement of fuel atomization by flash boiling and faster large spray angle, the fuel-rich boundaries (due to the rapid
evaporation of DME. fuel evaporation) of the DME sprays from the different
nozzle holes tend to overlap under long injection duration
5.1.2. Saturated vapour pressure of DME < Pa < critical and create strong swirl which is unfavourable for complete
pressure of DME [10,11,27] combustion. The characteristics of the DME spray suggest
In this case the behaviour of the DME spray is similar to that the DME may need a larger combustion chamber for
that of diesel fuel in terms of general shape. The tip pene- better mixing and combustion than diesel fuel. Fig. 2 shows
tration is less than or almost the same as that of the diesel typical DME spray images compared to diesel spray.
fuel spray and the spray angle is greater than that of diesel
fuel. The evaporation of the DME spray is more rapid than 5.2. Spray combustion of DME [26–29]
that of the diesel fuel spray and enhanced evaporation rates
have been observed at higher pressures. Unlike a diesel fuel Wakai et al. [27] also performed experiments on the
spray, the boundary of the DME spray seems to be irregu- spray combustion characteristics of DME, using a constant
lar and this irregularity increases with increasing cylinder volume vessel at 3.1 MPa and 823 K, and found that the
pressure. At cylinder pressures near the DME critical pres- DME spray was ignited at 2.6 ms after the start of injec-
sure, breakup of the DME spray occurs unlike the behav- tion. The first ignition occured around the middle portion
iour of the diesel fuel spray. of the spray, and the flame spread mainly in the down-
stream direction. The flame was non-luminous and light
5.1.3. Pa > critical pressure of DME [10,11] emission from the flame was very weak. Under lower ambi-
When the pressure of DME droplets enters the super- ent pressures (2.1 and 1.1 MPa), the ignition delay of the
critical region, the spray turns into a miscible turbulent DME spray was shorter than that of a diesel fuel spray,
jet (a single-phase phenomenon). In this case, the liquid– whereas it was almost the same as the diesel fuel at higher
vapour interface disappears and the droplets spread to ambient pressures (4.1 and 3.1 MPa). At an ambient pres-
the neighborhood in an explosive way. The spray angle sure of 1.1 MPa, the DME spray could be ignited at tem-
increases as the ambient pressure decreases, whereas a die- peratures >770 K, contrary to the diesel fuel which could
sel fuel spray demonstrates the opposite tendency. The not be ignited.
transition from a spray to a miscible turbulent jet may be
encountered during the spray development. If the size of 5.3. Numerical simulation of DME spray combustion
the droplets is small enough, the droplets may completely
vaporize before reaching the spray tip due to the tempera- Numerical simulations of spray combustion for DME
ture distribution in the spray. In this case, during the spray were carried out by Golovitchev et al. [30,31]. Their numer-
development, the dispersing droplets may dominate the ical model was based on the KIVA-3 code, with modified
spray hydrodynamics in the early stages, while air entrain- sub-models accounting for spray atomization, detailed oxi-
ment may considerably affect the later stages of the spray dation chemistry, soot formation and the effects of chemis-
evolution. try-turbulence interaction. The leading role of the methyl
Wakai et al. [27] suggested that, in the low-pressure formate reaction path was recognized in the ignition pro-
regions in the nozzle sac chamber and in the holes, cavita- cess and the detailed chemical mechanism (43 species and
tion in DME occurs more easily than in the diesel fuel. It 222 reactions), including soot formation, was used. The
was considered that such cavitation formation in the form predictive capability of the model was evaluated through
of vapour bubbles can enhance the atomization of the comparison of numerical and experimental data on liquid
DME spray and influence the spray characteristics. Kapus and vapour penetration and on ignition in a constant-vol-
and Ofner [9] also suggested that complex gas–liquid two- ume chamber under diesel-like conditions for both n-hep-
phase dynamic flow with cavitation, may occur for DME tane and DME. The predictions of the ignition point and
which increases the exit angle of the discharging spray. delay time were found to be in agreement with the experi-
The effect of cavitation on the enhancement of atomization mental data. The calculations showed ignition to start after
and on the spray global characteristics for DME is still evaporation of a major portion of the fuel which allowed
unclear. an auto-ignition wave to propagate in a partially premixed
Teng and McCandless [29] conducted a comparative mixture. The flame propagation process exhibited oscilla-
study of the characteristics of diesel fuel and DME sprays tory characteristics where the auto-ignition kernels moved
under realistic engine conditions on the basis of momentum forwards and backwards, due to pressure waves and the
conservation. Their analysis revealed that the DME spray cooling effect of the evaporating fuel. A high-temperature
in the diesel combustion system may not develop as well zone became visible in the central part of the spray. Simu-
as that of the diesel fuel at high engine loads and speeds lation of an engine fuelled with neat DME was also carried
due to the longer injection duration and increased spray out using the same CFD model. The DME fuel spray was
angle. Being less dense and more volatile than diesel, rapidly atomised and vapourised. In the initial stage of
1022

Fig. 2. DME sprays at various conditions compared to diesel sprays [18]; atmospheric (10.1 MPa) and high pressure (3 MPa) ambient conditions with
fuel-injection pressures at 25 and 55 MPa.

ignition the heat release zone appeared to be associated fuel, while DME fuel consumption can be kept close to the
with a diffusion flame but later on the flame propagated DI diesel value on an energy basis. Fig. 3 illustrates some of
in a partially premixed mixture. Under these conditions, the most recent data on brake specific fuel consumption
the ignition mechanism showed the well-known two-step (BSFC), NOx and CO2 specific emissions for DME com-
characteristics, with the maximum heat release taking place pared with diesel fuel from a six-cylinder 7 l turbo-
after TDC. Better performance and peak combustion pres- charged/intercooled heavy-duty diesel engine operating in
sures were demonstrated at lower injection pressures and the Japanese JE-05 test mode [18]. The data demonstrates
reduced injection velocities (<170 m/s). The validation data the reduction in NOx and CO2 emissions that can be
for the CFD model was given by Zhao et. al [32]. Follow- achieved with DME at a similar fuel economy. Smoke
ing this it was argue that the application of neat DME (PM) and THC together with NOx are shown in Fig. 4
would require the development of low-pressure fuel-injec- from experiments obtained in a small single-cylinder engine
tion equipment, properly matched to the combustion of 638 cm3 [33]. The data in Figs. 3 and 4 show the reduc-
system. tion in NOx and particulate emissions that can be achieved
with DME at equal power and torque and equal fuel econ-
6. Engine performance and emissions omy. It can be argued that any particulate emissions come
from the lubricant oil and not from DME. Because ignition
6.1. Exhaust emissions and fuel consumption delay is shorter with DME, NOx emissions should, in prin-
ciple, be lower than those for diesel fuel.
In previous investigations [3,5,9], it has been demon-
strated that emissions of particulate matter (PM), NOx, 6.1.1. Particulate matter (PM)
and combustion noise from compression-ignition engines It is well known that soot is formed in fuel-rich regions
are significantly lower when DME is used instead of diesel under high temperature conditions. The precursors of soot
 1023

Fig. 3. Fuel consumption (BSFC), NOx and CO2 emissions for DME compared to diesel from a six-cylinder 7 l turbocharged/intercooled heavy-duty
diesel engine operating in Japanese D13 mode [18].

increased oxygen content in the fuel [34] and with

decreased number of C–C bonds [35].
The effect of oxygen content in the fuel on the PM from
a diesel engine was investigated experimentally by Miyam-
oto et al. [36]. The PM emissions reduced from the value
for conventional diesel fuel with no oxygen content to less
than 2% of that value when the oxygen content of the fuel
was 25–30% by mass. Curran et al. [37] showed the influ-
ence of the oxygen content of fuels on the production of
soot precursors using a detailed chemical kinetic model;
the calculated values of the soot precursors decreased with
increased oxygen content in the fuel. Therefore, it can be
concluded that the soot formed in DME combustion
should be almost zero at an oxygen content of 35% and
no C–C bonds. This implies that DME-fuelled engines
would not need a particulate filter in the aftertreatment sys-
tem. Sidhu et al. found in [38] that DME yields much lower
particulates (0.026%) than other fuels under compression-
ignition conditions such as diesel (0.51%), biodiesel
(0.52%) and CNG (0.30%). A DME-fuelled engine was also
found to have in the exhaust gas higher (71%) soluble
organic fraction (SOF) than in a diesel (20%) of total
carbon.

6.1.2. NOx
Comparative values of NOx emissions from DME CI
engines and those from diesel fuel seem to vary depending
on the engine conditions and the fuel supply system,
according to various sources in the literature. NOx was
found to be lower with DME than with diesel fuel [3,9],
which was attributed to the shorter ignition delay for
Fig. 4. Smoke, THC and NOx emission for DME and diesel from a single DME than for diesel, the smaller amount of fuel injected
cylinder of 638 cm3 DI compression-ignition engine [33]. during the ignition delay period and the smaller amount
of fuel burned during the pre-mixed burning phase. The
initial enthalpy change was thus lower, resulting in a reduc-
are unsaturated hydrocarbons such as acetylene (C2H2), tion of the peak combustion temperature. However, it is
ethylene (C2H4), and propargyl (C3H3), found in diesel possible that higher NOx can be produced from DME than
combustion products [34]. The proportion of fuel carbon from diesel fuel for an early start of injection since the
forming soot precursors has been found to decrease with duration of the peak combustion temperature would be
1024

longer in the initial combustion period due to the shorter 6.1.3. HC and CO
ignition delay of DME [39]. When injection retardation is Table 3 shows HC and CO emissions in the literature
optimized for each fuel, NOx from DME is lower than from CI engines fuelled with DME, compared to those
from diesel fuel [39,40]. Teng et al. [11] suggested that, with diesel fuel. It can be seen that HC emissions from
although DME is an oxygenated fuel, the release of free DME are usually lower than or equal to those from the
oxygen is very limited as revealed by chemical kinetics cal- combustion of diesel fuel; HC emissions consist of partially
culations [12] which makes the effect of the oxygen content or completely unburned fuel, produced in locations where
in DME on NOx formation insignificant. Egnell [22] stud- combustion takes place under fuel-rich conditions, due to
ied NOx formation for DME using a reaction enthalpy incomplete air-fuel mixing. Teng et al. [11] explained that,
analysis, by maintaining similar injection conditions for because DME has a short ignition delay period, the over-
DME and diesel fuel. It was shown that the rate of reaction rich and over-lean mixture regions formed during the igni-
is not fundamentally different for DME and diesel fuel tion delay period might be smaller, resulting in significantly
under these conditions. One explanation for the lower reduced HC emissions. It should be added that DME is an
NOx emissions with DME, when compared to diesel fuel oxygenated fuel containing 35% by mass of oxygen, has
injected at considerably higher pressures, was that the local good mixing characteristics and is a superheated vapour
equivalence ratio was higher for DME, giving a lower local after entering the engine cylinder. The volume of fuel-rich
oxygen concentration and thus NOx levels. It was sug- regions existing during the combustion period could thus
gested that because the gaseous specific heat capacity of be less, resulting in reduced HC emissions.
DME is higher than that of diesel fuel, the adiabatic flame The data of CO emissions show some contradictions
temperature for DME is slightly lower than for diesel fuel depending on the engine system and operating conditions.
(represented in this analysis by dodecane). This resulted in The higher CO emissions sometimes associated with DME
the proposition that lower NOx emissions for DME should could result from the longer injection duration, coupled
be expected if all other conditions are identical; this could with lower injection pressures and larger spray holes. Since
again be explained by the lower rate of energy release [41]. there is production of CHO and CH2O involved in the
A zero-dimensional multizone model has also been imple- combustion of DME, depending on the reaction process,
mented to qualitatively simulate model the combustion a larger amount of CO may be produced compared with
characteristics of DME within a compression-ignition diesel fuel [24]. Impingement of the DME fuel spray on
engine with particular reference to NOx formation; the pre- the cooled combustion chamber wall could raise HC and
dictions indicated that DME forms less NOx than diesel CO emissions by quenching the DME reaction process
fuel [42,43]. [48]. CO may also be produced in locations of over-lean
Exhaust gas recirculation (EGR) reduces both burned conditions. Due to the faster evaporation and better mixing
gas temperature and NOx formation inside the burned with air, despite the shorter ignition delay for DME, more
gas region. DME engines allow a higher EGR rate for over-leaning can occur and, if the local equivalence ratio
NOx reduction without being restricted by the NOx-soot becomes too low to support combustion, an increase of
tradeoff, because DME has a high cetane number and fuel CO emissions may result. Higher emissions of CO and
droplets in a DME spray within the engine cylinder evapo- HC have been reported along with lower NOx in many
rate – about three times faster than diesel fuel [44]; further investigations [49–53].
NOx reduction can be achieved with higher EGR rate [45]. On the other hand, DME has good mixing characteris-
Teng and McCandless [46] found that ‘‘engine-out’’ NOx tics, so that the locations of the fuel-rich regions in the
emissions were less than the 2007 regulated level during combustion period could be reduced, resulting in lower
tests in a heavy-duty truck engine. The longer injection CO emissions [11]. Egnell [22] argued that the lower emis-
duration for DME with current fuel-system technology sions of HC and CO are linked to the structure of the DME
was, however, unfavourable for NOx reduction; to prevent molecule, similarly to soot formation. The low C/H ratio,
this a multiple-injection strategy was proposed [47]. the lack of C–C bonds and the high oxygen content of

Table 3
Comparison of HC and CO emissions for DME with those for diesel fuel
Year Test condition HC CO
1995 McCarthy et al. [54] ECE R49 13-mode cycle – estimation Almost equal Higher
1995 Fleisch and Meurer [3] ECE R49 13-mode cycle – estimation Almost equal
1995 Kapus and Cartellieri [5] US-FTP 75 cycle – estimation Almost equal Higher
Lower (with oxid. cat.) Higher (with oxid. cat.)
1997 Kajitani et al. [39] Steady-state test Lower
1998 Kajitani et al. [23] Steady-state test Lower Lower
1998 Ofner et al. [24] ECE R49 13-mode cycle test Lower Higher
2001 Egnell [22] Steady-state test Lower Lower
2005 Goto et al. [21] Japanese D13-mode Lower Lower
 1025

the fuel should give faster and more effective oxidation of

intermediate species. The interaction between fuel-injection
system behaviour and the combustion chamber design
should be optimized to minimize HC and CO emissions.
HC and CO could nevertheless be treated relatively easily
with an oxidation catalyst in the aftertreatment system.

6.1.4. Combustion noise [3,15]

Since ignition delay is shorter for DME than for diesel,
the pressure rise for DME during the premixed burning
stage is slower than with diesel fuel which results in reduced
combustion noise.

6.1.5. Non-regulated exhaust emissions

It is likely that some formaldehyde (CH2O) is formed
during combustion of DME in a CI engine, but significant
emissions of sulfur dioxide (SO2), polycyclic aromatic
hydrocarbons (PAH) and benzene, toluene and xylene
(BTX) are expected. The formaldehyde emissions could
be 1–10% of the non-methane hydrocarbon emissions
[1,35]. Aldehydes may also be high because associated
CO emissions are high, although formaldehydes can be eas-
ily reduced by oxidation catalysts to a negligible level [55].

6.2. Effect of fuel-injection equipment (FIE) and

combustion system parameters
Fig. 5. Effect of fuel-injection pressure in the low speed and high load
The effect of the fuel-injection equipment and combus- operating conditions for DME with a single cylinder of 638 ml DI
compression-ignition engine [18].
tion system parameters on fuel consumption, NOx emis-
sions and combustion noise has been investigated for
various DME-fuelled engines [5,9,18,36,56]. The parame- ever, another investigation with a common-rail sys-
ters examined included injection pressure, start of injection, tem in a smaller engine (0.5 l/cylinder) showed an
rate of injection, compression ratio, piston-bowl shape and opposite trend [56] i.e. that higher injection pressure
EGR level; each of these parameters is examined later on. in the range 25–55 MPa, improved fuel economy with
Fig. 5 shows an example of exhaust emissions affected by comparable or even reduced CO and THC emissions.
injection pressure [18]. Kook et al. [56] showed that the implications of better
Optimization of these parameters for a DME-fuelled mixing between DME and air through higher injec-
engine (single cylinder, 2 l) equipped with a pintle nozzle tion pressure are higher power output but higher
injector and EGR [9] achieved combustion noise of about NOx; the higher the injection pressure of DME
76 dB(A) at both part load (IMEP = 0.2 MPa) and full (between 20 and 55 MPa of common-rail pressure),
load (0.75 MPa). On the other hand, for diesel fuel without the higher was the IMEP, due to the improvement
EGR the noise levels were 80 dB(A) at part load and of the atomization process and mixing over a wide
86 dB(A) at full load. NOx emissions were below 2 g/ range of injection timings.
kWh, i.e. over 50% lower than for diesel, when the fuel con- On the basis of the above it is hard to ascertain the
sumption for DME was almost equal to that of diesel fuel. effect of injection pressure and further investigation
For a DME-fuelled engine (four cylinder, 2 l) equipped is needed for different fuel-injection equipment of
with a five-hole injector nozzle and no EGR [5], although parameters such as the injector geometry and the
the fuel consumption was slightly higher at higher load interaction between the emerging sprays, the in-cylin-
than with diesel, the NOx emissions were over 50% lower. der flow and the combustion chamber geometry.
At full load, the combustion noise reached levels of about (ii) Start of injection: Retardation of the start of injection
80–82 dB(A), about 10% lower than diesel fuel. led to a reduction of NOx emissions and combustion
noise, but deterioration of the fuel consumption and
(i) Injection pressure: It was found that higher injection increased CO emissions [9].
pressures, between 19 and 25 MPa, degraded fuel (iii) Nozzle type: The change of nozzle configuration,
economy and increased CO and THC emissions in from needle-hole nozzle to pintle nozzle, not only
a heavy-duty engine (1.15 l/cylinder) equipped with changed the spray formation but also the injection
a jerk pump type FIE, as shown in Fig. 5 [18]. How- process or the rate-shaping of injection. It was shown
1026

that the nozzle type should be matched to the engine out a central cone) was better than a re-entrant bowl
design for better fuel distribution and mixing, which for the case of a pintle nozzle injector. Kapus and
would in turn increase power output and reduce Cartellieri [5] also found that, with respect to fuel
exhaust emissions. consumption, a wide shallow bowl offered more
(iv) Injection process: Rate shaping of injection [5,9] was advantages than a narrow deep bowl when a multi-
carried out using multiple injections. With pilot-injec- hole nozzle was used.
tion both NOx emissions and combustion noise
decreased while fuel consumption increased. As the 7. Energy efficiency and well-to-wheels emissions
amount of fuel injected during the initial part of the
injection period decreased, combustion noise also 7.1. Well-to-wheels analysis
decreased. The increase of injection pressure later
on further decreased combustion noise over the The well-to-wheels energy efficiency is estimated as the
whole load region. product of the energy efficiency of each individual utiliza-
(v) Number of nozzle holes for a multi-hole nozzle: Three- tion step. The well-to-filling station step includes the recov-
and seven-hole nozzles were used as alternative ery and transport of the feedstock, fuel production and fuel
configurations to the standard five-hole nozzle in distribution efficiency. For the production of DME and
multi-hole nozzles [5]. With the three-hole nozzle, methanol, it is assumed for financial reasons that remotely
NOx emissions were slightly lower than with both located gas sites are used. Since waste heating from the fuel
the five- and seven-hole nozzles but fuel consumption production cannot be recovered for district heating, the
deteriorated considerably at higher loads. With the efficiencies of the fuel production are relatively low. Conse-
seven-hole nozzle, fuel consumption was improved quently, the well-to-filling station energy efficiency for
at higher loads due to the better fuel distribution in DME and methanol are among the lowest. Energy effi-
the combustion chamber but combustion noise ciency for the vehicle includes engine efficiency, transmis-
increased considerably. It was concluded that the sion efficiency and a weight assessed correction factor.
five-hole nozzle was the optimum configuration for Furthermore, the well-to-wheels CO2 emissions are deter-
DME. mined by the CO2 emissions during the production and dis-
(vi) Intake swirl level: Kapus and Ofner [9] found that tribution of the fuel, the CO2 produced during the
swirl flow in the combustion chamber was not neces- combustion of the fuel, and the overall vehicle efficiency.
sary with DME because of its fast rate of evaporation The CO2 emissions during fuel production are primarily
and good mixing with air. When the swirl level determined by the energy supplied for the production of
decreased from 1.8 to 0.1 (AVL swirl number) the the fuel but CO2 can also be emitted due to molecular
wall heat-transfer rate was also reduced due to the changes in the process of making fuel from the relevant
weaker air motion in the combustion chamber, result- feedstock. The emissions of CO2 during combustion are
ing in lower fuel consumption. Combustion noise was determined from the carbon content of the fuel and its
also reduced with the reduced charge motion. lower heating-value. Detailed analyses can be found in
(vii) Compression ratio (CR): With higher CR, ignition [1,4].
delay was reduced and the amount of fuel burnt dur- Table 4 shows the well-to-wheels energy efficiency and
ing the initial energy release decreased, causing a less- CO2 emissions of both light-duty vehicles (LDV) and
rapid pressure rise in the cylinder. Both NOx and heavy-duty vehicles (HDV). The well-to-wheels energy effi-
combustion noise were correspondingly reduced. ciency of DME-fuelled vehicles is generally comparable to
(viii) EGR: NOx emissions from DME combustion, simi- that of the LPG and CNG fuelled vehicles, while the energy
larly to other engine fuels, decreased with increased efficiency is the highest for DI vehicles operating with diesel
EGR rate, although CO and HC emissions became fuel. The well-to-wheels CO2 emissions of DME-fuelled
higher and fuel consumption increased with higher vehicles are comparable to those with DI diesel engines
EGR. EGR is an ideal mechanism, especially at low and CNG fuelled engines. Amongst the fossil fuel deriva-
loads, for decreasing NOx emissions in DME engines, tives, diesel and DME from natural gas have the lowest
due to the smoke-free combustion. At high loads, CO well-to-wheels CO2 emissions. As expected the fuels pro-
emissions limit the use of EGR for normally aspi- duced from renewable feedstock have very low well-to-
rated engines as the air-excess ratio, which is already wheel CO2 emission levels.
at a low level, is further reduced by EGR.
(ix) Piston-bowl shape: It is generally desirable in engines 7.2. Assessment of emissions by the sustainable process index
with a piston-bowl configuration that spray-wall wet-
ting (or spray-wall impingement) should be mini- The overall ecological impact of a fuel can be judged by
mized in order to allow low exhaust emissions, considering other emissions such as NOx. Ofner et al. [24]
combustion noise and fuel consumption. Kapus and carried out an ecological assessment of DME as a fuel,
Ofner [9], found that, with respect to combustion using the sustainable process index (SPI) proposed by
noise and CO emissions, a normal piston-bowl (with- Krotscheck and Narodoslawsky which takes into account
 1027

Table 4
Well-to-wheels energy efficiency and CO2 emissions of light-duty vehicles (LDV) and heavy-duty vehicles (HDV) [1,4]
Fuel Engine type Well-to-wheel efficiency (%) Relative well-to-wheel CO2 emissions (Diesel = 1)
LDV HDV LDV HDV
Diesel CI, DI 25.7 30.3 1.00 1.00
DME CI, DI 19.0 22.5 1.01 1.02
DME renewable CI, DI – – 0.17 0.17
Gasoline SI, k = 1 16.9 20.1 1.47 1.46
LPG SI, k = 1 18.2 21.6 1.28 1.28
CNG SI, k = 1 16.8 20.7 1.14 1.09

both CO2 and NOx emissions. The concept of the SPI is properties. As a result the overall environmental impact
based on a comprehensive knowledge of the interactions shifts between the values of the fossil and bio-genic primary
of human-induced mass and energy flows with those of energy source (e.g. wood) in proportion to the relative mix-
the environment although this comprehensive knowledge ture ratio. From Fig. 6, it can be deduced that 17% of
does not really exist. With respect to fuels, the SPI is calcu- DME must be produced from wood in order to lower the
lated by comparing the natural carbon flow of the life-cycle SPI to the CNG level [24].
of the fuel to the human-induced carbon fuel flow. The nat-
ural carbon flow for fossil fuels is taken as the rate of car- 8. New DME concept engine
bon sedimentation in the oceans. The carbon flow in the
natural cycle can be expressed in terms of the area required With the objective of improving both engine thermal
for circulation of a kg of carbon per year; for bio-fuels, this efficiency and exhaust emissions, several new engine-design
area amounts to approximately 7 m2. For fossil fuels, sed- concepts such as low compression ratio DI diesel engines
imentation to accumulate 1 kg of carbon requires approx- and homogeneous charge compression ignition (HCCI)
imately 500 m2. Thus, the natural carbon flow for fossil engines have been investigated using DME as an alterna-
fuels is approximately 70 times lower than for bio-fuels. tive fuel.
This method transforms the carbon flow estimate into Kajitani et al. [33] investigated the concept of using
‘area-year’ in terms of square meters annual (m2a). If an DME as fuel in a direct-injection compression-ignition
amount of carbon, say in wood, is used for providing engine with a low compression ratio, in an effort to identify
energy of 1 kWh at the end use, the carbon flow during a combustion regime with the highest possible thermal effi-
production, combustion and utilization of the fuel and ciency. It was shown that:
the corresponding assimilation by trees, can be expressed
in terms of ‘area-year’ per kWh. (i) The lowest compression ratio for easy start and stable
Fig. 6 shows the SPI in terms of m2a/kWh-wheel for operation of a small passenger-car engine was around
different fuels [24]. It is evident that the CO2 emissions of 12:1.
fossil fuels contribute significantly to the SPI. DME pro- (ii) The DME-fuelled engine was superior to the conven-
duced from natural gas has a 10% lower overall impact tional diesel engine in terms of thermal efficiency,
compared to diesel EUROIII because of the reduced exhaust emissions and engine noise.
NOx emissions. (iii) The ignition delay became longer with a reduction in
The main advantage of DME is that it can be produced compression ratio from 18 to 12.
from both fossil and renewable resources. DME produced (iv) The brake thermal efficiency, THC and CO emissions
from renewable feedstock can be added to the DME pro- as well as engine noise for DME remained almost
duced from natural gas without any change of the fuel constant over this compression ratio range.

Fig. 6. Sustainable process index (SPI) for various fuels [24].

1028

(v) NOx emissions became lower with a reduction in loads, combined with a high EGR rate, may allow NOx
compression ratio but remained at unacceptably high emissions from DME engines to meet the US 2010 stan-
levels. dard [46].

Lowering the compression ratio allowed reduction of 9. Summary

the peak firing pressure in the cylinder. Consequently,
engines could be designed with a lighter structure than cur- Numerous investigations of DME-fuelled engines have
rent diesel engines and lower cost. By combining low com- indicated that it offers excellent promise as an alternative
pression ratio with rate-shaped injection, NOx emissions, fuel for compression-ignition operation in the automotive
peak firing pressure and noise could all be further reduced. sector.
There is consensus that the HCCI engine has great DME is a liquefied gas that can be produced from a
potential to simultaneously provide high thermal efficiency, variety of feed stock, is non-toxic and environmen-
low combustion noise, lower NOx levels and soot-free com- tally benign. DME is as easy to handle as LPG since it is
bustion. On the other hand, it appears to be limited to condensed by pressurizing above 0.5 MPa. DME’s main
lighter load operation, exhibit difficulty of controlling igni- feature as an efficient alternative fuel for use in compres-
tion timing and produce more unburned-HC emissions. sion-ignition engines is its high cetane number. The fast
The HCCI engine is yet in the laboratory research phase evaporation of DME can lead to better mixing with air
owing to the difficulty of making it appropriate for mass in the engine cylinder and its high oxygen content can
production in the transportation sector. To address the achieve smokeless combustion through low formation
problems of the HCCI engine and to overcome future emis- and high oxidation rates of particulates. However, DME
sion legislation limits for vehicle operation, investigations requires a higher injected volume to supply the same
into fuel property and quality (including DME) are being amount of energy as the diesel fuel, due to its lower density
conducted. Since DME injected into the cylinder as a liquid and combustion enthalpy. DME-fuelled systems also need
fuel quickly evaporates, it is likely that a homogeneous air– lubricity-enhancing additives and anti-corrosive sealing
fuel mixture can be more easily formed. Rapidly evaporat- materials to secure leakage-free operation.
ing DME showed near zero PM and NOx emissions with- The DME spray characteristics and combustion process
out power loss, due to the advantage of no wall-wetting have been investigated for various fuel-injection equip-
[56]. DME was found to achieve HCCI operation with a ment, ranging from mechanical pump systems to advanced
diesel FIE system without engine modification. Combus- common-rail systems, in a variety of prototype engines
tion phasing could be controlled with either EGR or by with occasional slight modifications. These engine tests
the addition of high-octane/low-cetane fuels such as alco- with DME proved its potential as a clean alternative fuel
hols, methane and hydrogen, through suppression of the achieving smoke-free high power operation provided an
OH radical production in the premixed charge [57,58]. optimized fuel injection equipment and the engine configu-
Takatsuto et al. [59] experimentally investigated auto- ration are closely matched to reduce fuel consumption and
ignition and combustion characteristics in homogeneous exhaust emissions. Optimisation of the fuel injection equip-
lean mixtures. Auto-ignition was realized for DME over ment, overcoming the problem of low density, low lubricity
a wide equivalence ratio range compared to n-butane. and corrosiveness, may allow mass production of the
NOx emissions with DME were much lower than those DME-fuelled compression-ignition engines when adequate
with n-butane and THC concentrations were also low. DME supplies become available in the market.
For DME, auto-ignition and combustion were always It can be argued however that the well-to-wheels energy
found to include high and low temperature reactions, at efficiency and CO2 emissions for DME produced from nat-
all equivalence ratios, which were analysed through various ural gas are not better than for most other fuels, as a result
simulation and diagnostics tools [60–63]. Since DME has a of the high fuel production energy and relevant CO2 emis-
high cetane number and fuel drops in the DME spray sions, and that diesel fuel has the highest well-to-wheels
within an engine cylinder evaporate about three times fas- energy efficiency. Nevertheless, DME has the advantage
ter than diesel fuel, low-temperature combustion may be that it can be produced from both fossil and renewable
more appropriate for DME engines. This could be resources. DME produced from renewable feedstock may
achieved by using a multiple-injection strategy with timing be added to that produced from natural gas, without any
for the primary injection determined by the cylinder tem- change in fuel properties, for use as an alternative automo-
perature [46]. DME as an additive or an ignition enhancer tive fuel. However, extensive studies of the environmental
even enables the use of various other fuels for conventional impact of DME are needed to evaluate its full potential,
diesel combustion, dual-fuel operation and HCCI opera- using assessments based on different model assumptions
tion. Propane or LPG, hydrogen and natural gas have all and focusing on local conditions in urban areas or on bio-
been tested as alternatives to diesel with the assistance of logical processes.
DME [22,49,64–68]. Recent studies on DME-fuelled engines showed the
It is also expected that by employing the Miller cycle potential of exploiting new combustion concepts such as
at low loads and a multiple injection strategy at high low-compression DI diesel engines and homogeneous
 1029

charge compression-ignition engines operating in a low- [21] Goto S, Oguma M, Suzuki S. Research and development of a medium
temperature combustion regime. duty DME truck. SAE Paper 2005-01-2194;2005.
[22] Egnell R. Comparison of heat release and NOx formation in a DI
Overall, DME has been found to be a very promising diesel engine running on DME and diesel fuel. SAE Paper 2001-01-
alternative fuel for compression-ignition engines, capable 0651, SAE Trans J Fuel Lubr 2001;110(4):492–506.
of providing high thermal efficiency, low combustion noise, [23] Kajitani S, Chen CL, Oguma M, Alam M, Rhee KT. Direct injection
low NOx levels and soot-free combustion; it thus merits diesel engine operated with propane – DME blended fuel. SAE Paper
further research and development before a final decision 982536; 1998.
[24] Ofner H, Gill DW, Krotscheck C. Di-methyl ether as fuel for CI
is taken on its potential as a mass production fuel for the engines – a new technology and its environmental potential. SAE
automotive market. Paper 981158, SAE Trans J Fuel Lubr 1998;107(4):425–37.
[25] Ikeda T, Ohmori Y, Takamura A, Sato Y, Jun L, Kamimoto T.
Measurement of the rate of multiple fuel injection with diesel fuel and
References DME. SAE Paper 2001-01-0527, SAE Trans J Engine 2001;110(3):
372–80.
[1] Verbeek RP, Van Doom A, van Walwijk M. Global assessment of [26] McCandless JC, Teng H, Schneyer JB. Development of a liquid-DME
di-methyl ether as an automotive fuel. second ed., 96.OR.VM.029. fuel tank – a two-fluid thermodynamic pump. SAE Paper 2001-01-
1/RV, TNO Road-Vehicles Research Institute; 1996. 0652, SAE Trans J Fuel Lubr 2001;110(4):507–13.
[2] Arcoumanis C. The second European Auto-Oil programme (AOLII). [27] Wakai K, Nishida K, Yoshizaki T, Hiroyasu H. Spray and ignition
European Commission, vol. 2. Alternative Fuels for Transportation; characteristics of di-methyl ether injected by a DI diesel injector. In:
2000. Proceedings of the fourth international symposium COMODIA’98;
[3] Fleisch TH, Meurer C. DME, The diesel fuel for the 21st Century? In: 1998. p. 537–42.
AVL conference engine and environment; 1995. [28] Glensvig M. Sorenson SC, Abata DL. An investigation of the
[4] Verbeek R, Van der Weide J. Global assessment of di-methyl ether: injection characteristics of di-methyl ether. ICE-vol. 29-3, ASME;
comparison with other fuels. SAE Paper 971607; 1997. 1997.
[5] Kapus PE, Cartellieri WP. ULEV potential of a DI/TCI diesel [29] Teng H, McCandless JC. Comparative study of characteristics of
passenger car engine operated on di-methyl ether. SAE Paper 952754; diesel-fuel and dimethyl-ether sprays in the engine. SAE Paper 2005-
1995. 01-1723, SAE Trans J Engine 2005;114(3):1202–13.
[6] Basu A, Wainwright JM. DME as a power generation fuel: [30] Golovitchev VI, Nordin N, Chomiak J. Neat di-methyl ether: is it
performance in gas turbines. In: PETROTECH-2001 conference; really diesel fuel of promise? SAE Paper 982537;1998.
2001. [31] Golovitchev VI, Nordin N, Chomiak J, Nishida N, Wakai K.
[7] Cocco D, Tola V, Cau G. Performance evaluation of chemically Evaluation of ignition quality of DME at diesel engine conditions. In:
recuperated gas turbine (CRGT) power plants fuelled by di-methyl- Proceedings of the fourth international conference of internal
ether (DME). Energy 2005;31:1446–58. combustion engines 99 (ICE99): experiments and modeling; 1999. p.
[8] Djebaı¨Li N, Paillard C. Burning velocities of dimethyl ether and air. 299–306.
Combus Flame 2001;125:1329–40. [32] Zhao Z, Kazakov A, Dryer FL. Measurements of dimethyl ether/air
[9] Kapus P, Ofner H. Development of fuel injection equipment and mixture burning velocities by using particle image velocimetry.
combustion system for DI diesels operated on di-methyl ether. SAE Combus Flame 2004;139:52–60.
Paper 950062, SAE Trans J Fuel Lubr 1995;104(4):54–9. [33] Kajitani S, Chen Z, Oguma M, Konno M. A study of low-
[10] Sorenson SC, Glensvig M, Abata D. Di-metyl ether in diesel fuel compression-ratio di-methyl ether diesel engines. Int J Engine Res
injection systems. SAE Paper 981159, SAE Trans J Fuel Lubr 2002;2:1–11.
1998;107(4):438–49. [34] Westbrook CK. Chemical kinetic modeling of oxygenated diesel fuels
[11] Teng H, McCandless JC, Schneyer JB. Thermo-chemical character- in advanced petroleum-based and alternative fuels. DOE Report;
istics of di-methyl ether – an alternative fuel for compression-ignition 1999.
engines. SAE Paper 2001-01-0154, SAE Trans J Fuel Lubr [35] Ogawa H, Miyamoto N, Yagi M. Chemical-kinetic analysis on PAH
2001;110(4):96–106. formation mechanisms of oxygenated fuels. SAE Paper 2003-01-3190,
[12] Edgar B, Dibble RW, Naegeli DW. Autoignition of di-methyl ether SAE Trans J Fuel Lubr 2003;112(4):2413–21.
and di-methoxy methane sprays at high pressures. SAE Paper 971677, [36] Miyamoto H, Ogawa H, Arima T, Miyakawa K. Improvement of
SAE Trans J Fuel Lubr 1997;106(4):625–639. diesel combustion and emissions with various oxygenated fuel
[13] Francisco JS. On the competition between hydrogen abstraction additives. SAE Paper 962115;1996.
versus C–O bond fission in initiating di-methyl ether combustion. [37] Curran EM, et al. Detailed chemical kinetic modeling of diesel
Combus Flame 1999;177:312–6. combustion with oxygenated fuels. SAE Paper 2001-01-0653, SAE
[14] Masterton WL, Slowinski EJ. Chemical principles. W.B. Saunders Trans J Fuel Lubr 2001;110(4):514–21.
Company; 1997. [38] Sidhu Sukh, Graham J, Striebich. Semi-volatile and particulate
[15] Alex Mills. Status and future opportunities for conversion of emissions from the combustion of alternative diesel fuel. Chemo-
synthesis gas to liquid fuels. Fuel 1994;73:1243–79. sphere 2001;42:681–90.
[16] Semelsberger TA, Borup RL, Greene HL. Dimethyl ether (DME) as [39] Kajitani S, Chen CL, Konno M. Engine performance and exhaust
an alternative fuel. J Power Sources 2005;156(2):497–511. characteristics of direct-injection diesel engine operated with DME.
[17] Mii T, Uchida M. Fuel DME plant in East Asia. In: Proceedings of SAE Paper 972973, SAE Trans J Fuel Lubr 1997;106(4):1568–77.
15th Saudi–Japan joint symposium; 2005. [40] Longbao Z, Hewu W, Deming J, Zuohua H. Study of performance
[18] Tsuchiya T, Sato Y. Development of DME engine for heavy-duty and combustion characteristics of a DME-fueled light-duty direct-
truck. SAE Paper 2006-01-0052; 2006. injection diesel engine. SAE Paper 1999-01-3669;1999.
[19] Yu J, Bae C. Dimethyl ether (DME) spray characteristics compared [41] Cipolat D. Combustion aspects of a compression ignition engine
to diesel in a common-rail fuel injection system. J Autom Eng Proc fueled on DME. In: Proceedings of the seventh international
IMechE Pt D 2003;217:1135–44. conference on energy for a clean environment; 2003. p. 156.
[20] Zhao X, Ren Meifeng, Liu Z. Critical solubility of dimethyl ether [42] Konno M, Kajitani S, Oguma M, Iwasa T, Shima K. NO emission
(DME) + diesel fuel and dimethyl carbonate (DMC) + diesel fuel. characteristics of a CI engine fueled with neat dimethyl ether. SAE
Fuel 2005;84:2380–3. Paper 1999-01-1116, SAE Trans J Fuel Lubr 1999;108(4):460–7.
1030

[43] Saeed A, Saeed K, Ahmed A, Malik KA. Multizones modeling of the [56] Kook S, Kong J, Min K, Bae C. Homogeneous charge compression
combustion characteristics of oxygenated fuels in CI engines. SAE ignition engine with two-stage diesel fuel injection. In: Proceedings of
Paper 2006-01-0051;2006. the THIESEL 2004 (thermo- and fluid dynamic processes in diesel
[44] Teng H, McCandless JC, Schneyer JB. Compression ignition delay engines); 2004. p. 497–510.
(physical + chemical) of di-methyl ether – an alternative fuel for [57] Ogawa H, Miyamoto N, Kaneko N, Ando H. Combustion control
compression-ignition engines. SAE Paper 2003-01-0759, SAE Trans J and operating range expansion with direct injection of reaction
Fuel Lubr 2003;112(4):377–89. suppressors in a premixed DME HCCI engine. SAE Paper 2003-01-
[45] Kinoshita K, Oguma M, Goto S, Sugiyama K, Mori M, Watanabe T. 0746;2003.
Effects of fuel injection conditions on driving performance of a DME [58] Shudo T, Kitahara S, Ogawa H. Influence of carbon dioxide on
diesel vehicle. SAE Paper 2003-01-3193;2003. combustion in an HCCI engine with the ignition-control by hydro-
[46] Teng H, McCandless JC. Can heavy-duty diesel engines fueled with gen. SAE Paper 2006-01-3248;2006.
DME meet US 2007/2010 emissions standard with a simplified [59] Takatsuto R, Igarashi T, Iida N. Autoignition and combustion of
aftertreatment system? SAE Paper 2006-01-0053;2006. DME and n-butane–air mixtures in homogeneous charge compres-
[47] Teng H, Regner G. Fuel injection strategy for reducing NOx sion ignition engine. In: Proceedings of the fourth international
emissions from heavy-duty diesel engine fueled with DME. SAE symposium COMODIA’98; 1998. p. 185–90.
Paper 2006-01-3324;2006. [60] Sato S, Iida N. Analysis of DME homogeneous charge compression
[48] Oguma M, Goto S, Hatsuzawa H, Konno M, Chen Z, Watanabe T. ignition combustion. SAE Paper 2003-01-1825;2003.
Chemiluminescence analysis from in-cylinder combustion of a DME- [61] Kim H, Cho S, Min K. Reduced chemical kinetic model of DME for
fueled DI diesel engine. SAE Paper 2003-01-3192;2003. HCCI combustion. SAE Paper 2003-01-182522, SAE Trans J Fuel
[49] Crookes RJ, Bob-Manuel KDH. Di-methyl ether or rapeseed methyl Lubr 2003;112(4):1180–7.
ester: a preferred alternative fuel option for future diesel engine [62] Kawahara N, Tomita E, Kagajyo H. Homogeneous charge compres-
operation. In: Proceedings of the fifth international colloquium on sion ignition combustion with dimethyl ether – spectrum analysis of
fuels; 2005. p. 181–7. chemiluminescence. SAE Paper 2003-01-1828, SAE Trans J Fuel
[50] Mingfa Y, Zunquing Z, Sidu X, Maoling F. Experimental study on Lubr 2003;112(4):1214–21.
the combustion process of dimethyle ether (DME). SAE Paper 2003- [63] Yamada H, Goto Y. Analysis of reaction mechanisms controlling
01-3194, SAE Trans J Fuel Lubr 2003;112(4):2422–9. cool and thermal flame with DME fueled HCCI engines. SAE Paper
[51] Song J, Huang Z, Qiao X, Wang W. Performance of a controllable 2006-01-3299;2006.
premixed combustion engine fueled with dimethyl ether. J. Energy [64] Chen Z, Konno M, Oguma M, Yanai T. Experimental study of CI
Convers Manage 2004;45:2223–32. natural-gas/DME homogeneous charge engine. SAE Paper 2000-01-
[52] Sorenson SC, Mikkelsen S. Performance and emissions of a 0.273 l 0329, SAE Trans J Engine 2000;109(3):442–51.
direct injection diesel engine fuelled with neat dimethyl ether. SAE [65] Narioka Y, Yokoyama T, Lio S, Takagi Y. HCCI Combustion
Paper 950064, SAE Trans J Fuel Lubr 1995;104(4):80–90. characteristics of hydrogen and hydrogen-rich natural gas reformate
[53] Pana C, Negurescu N, Popa MG, Boboc G, Cernat Al. Performance supported by DME supplement. SAE Paper 2006-01-0628;2006.
of a DME fueled diesel engine. In: Proceedings of the fifth [66] Crookes RJ. Biofuels performance in internal combustion engines. In:
international colloquium on fuels; 2005. p. 227–32. Proceedings of the fourth international colloquium on fuels; 2003. p.
[54] Fleisch T, McCarthy C, Basu A, Udovich C, Charbonneau P, 201–7.
Slodowske W, Mikkelsen SE, McCandless JC. A new clean diesel [67] Yeom K, Jang J, Bae C. Homogeneous charge compression ignition
technology: demonstration of ULEV emissions on a navistar diesel of LPG and gasoline using variable valve timing in an engine. Fuel
engine using a new alternative fuel. SAE Paper 950061, SAE Trans J 2007;86(4):494–503.
Fuel Lubr 1995;104(4):42–53. [68] Yeom K, Bae C. Gasoline – di-methyl ether homogeneous charge
[55] Oguma M, Shiotani H, Goto S, Suzuki S. Measurement of trace levels compression ignition engine, energy and fuels, in press.
of harmful substances emitted from a DME DI diesel engine. SAE
Paper 2005-01-2202;2005.