Dimethyl Ether Fuel Literature Review

Executive Summary

Transport Canada (TC), through its ecoTECHNOLOGY for Vehicles (eTV) program, has retained the services of the National Research Council Canada (NRC) to undertake a comprehensive literature search to study the production, distribution, handling and use of Dimethyl Ether (DME) in road vehicles as a substitute for conventional pump diesel.  Although DME can theoretically be used to power vehicles in any sector (e.g. passenger vehicles, off-road equipment) the principal thrust of this study involves diesel engines that have been modified, or designed, to accept DME and used in the heavy duty on-road sector. The main purpose will be to present the attributes of DME as they pertain to heavy duty vehicles but also to present any limitations the fuel may have for use in Canada.

NRC-AST used the services of NRC's Knowledge Management (KM) group to retrieve as many DME related journals, papers, presentations, test results and dissertations from academics as well as marketing and specification documents from the commercial trucking sector. Although much of the information was retrieved from academic papers from all over the world, care was taken to not focus entirely on the chemical theories of DME. Rather, the theory was blended with the industry-specific requirements that were pertinent to the heavy haul sector in Canada to form a study that can be used to fully understand how DME can be used in Canada. More importantly, any of the limitations that must be understood by operators, drivers, refuellers, government regulators and the motoring public have also been explored. In all, over 70 papers, presentations and industry documents formed the basis of the references for this research study.

Dimethyl Ether (DME) is an organic compound with the chemical formula CH3OCH3. For decades it has been used in a variety of products and applications such as propellants in aerosol cans, cooking fuels, solvents and medical treatments due to its lack of odour and toxicity and its ability to be absorbed into the troposphere.  However, it can also be made into a viable alternative for diesel fuel, most notably for use in heavy haul transport vehicles.

Unlike conventional diesel which is produced from non-renewable crude oil, DME can be produced anywhere using renewable products like natural gas, crude oil, propane, residual oil, pulp and paper waste, agricultural by-products, municipal waste, fuel crops such as switchgrass, coal, and biomass such as forest products and animal waste. This provides a great deal of flexibility for production since facilities do not need to be located near sources of crude oil but can be setup any place where bio based feedstocks or natural gas can be found, or produced.

The current estimated yearly production of DME is approximately 5 million to 9 million tons, depending on the cited source. The literature review revealed many methods by which DME may be produced. In general though, DME is currently produced via the dehydrogenation reaction of methanol. Based on current projections, it is likely that the abundance of North American natural gas and a high level of animal waste could provide ample sources for future DME production without reliance on offshore resources. It takes approximately 1.4 tons of Methanol to produce approximately 1.0 ton of DME. Other forms of production such as the use of pulp and paper process waste such as black liquor were also identified. A joint venture between Volvo and Chemrec in the North of Sweden produced 4.3 tons of black liquor based DME per day to power a dedicated fleet of four tractor-trailers.

Oberon Fuels manufactures skid mounted small scale DME production units meant to replace large scale infrastructure projects that could cost hundreds of thousands of dollars. The units can produce between 3,000 and 10,000 gallons of DME a day (11,340 litres to 37,800 litres. This process is ideally suited to regional haulers with large fleets who see each of their vehicles daily and want to reduce costs by not only self fueling, but self-producing the fuel.

DME exists as an invisible gaseous ether compound under atmospheric conditions (0.1 MPa and 298 K) but must be condensed to the liquid phase by pressurization above its vapour pressure at about 0.5 MPa (5.1 bar/73 psi) at 25oC to be used as a diesel fuel alternative. One of the more significant features of DME is the lack of a direct carbon-to-carbon bond that is found in traditional diesel fuels. Conventional diesel contains no oxygen whereas DME is an oxygenated fuel and contains about 34.8% oxygen by mass with no carbon-to-carbon bonds. The increase in oxygen content can reduce the precursors to soot formation like C2H2, C2H4 and C3H3. The presence of oxygen can also reduce auto ignition since the C-O bond energy is lower than the C-H bond energy found in conventional diesel. DME has approximately 66% of the energy content, by mass, and about 50%, by volume, of diesel fuel.

The air/fuel ratio of DME fuel at stoichiometric conditions is approximately 9 versus 14.6 for diesel meaning that complete combustion of 1 kg DME requires less air than that of 1 kg diesel fuel. However, more than 1 kg of DME is required to provide the same amount of energy as 1 kg of diesel. DME has a much higher, and wider, flammability range (i.e. the volume of fuel, expressed as a percentage in an air mixture at standard conditions, where ignition may occur) in air than the three hydrocarbon fuels (gasoline, diesel and propane but very similar to natural gas. DME is sulfur free whereas even ultra-low sulfur diesel (ULSD) contains some sulfur.

Most #1 and #2 pump diesel fuels have cetane numbers between 40 and 45 and many bio-diesels have CN greater than 50. DME has a cetane number between 55 and 60, which makes it very suitable for a diesel cycle engine.  This reduces engine knocking and engine noise when compared to engines powered with conventional diesel and also helps to provide a more complete combustion process with less wasted fuel, particularly at engine start up or when in-cylinder temperatures cool off.  Fuels such as propane and natural gas have high octane numbers but cetane numbers less than 10, making them impractical for dedicated use in a diesel cycle engine unless they are combined with at least some diesel as an ignition source.

DME in the liquid state has low viscosity and low lubricity, two properties which strongly affect the maximum achievable injection pressure in a fuel injection system: viscosity allowing it to readily pass through narrow passages and the lack of lubricity can accelerate the wear of surfaces moving relative to each other such as the feed pump, the high pressure injection pump, and injector nozzles. Due to the low viscosity and lubrication characteristics, fuel additives are mandatory to improve the fuel viscosity to make DME a viable fuel for on road engines.

In addition to its low lubricity and viscosity, DME adversely affects many types of plastics and rubbers and also dissolves nearly all known elastomers found in the fuel system. Retrofitting a vehicle to burn DME that is equipped with elastomers and certain plastics could result in very short service life of those components and possible fuel leaks or a reduction in working pressure. Laboratory tests have demonstrated that DME is compatible with Teflon® and Buna-N rubber. Tests have demonstrated that the bulk modulus of DME is approximately 1/3 that of conventional diesel. Research has demonstrated that due to DME's low elastic modulus, the compressibility of DME is higher than that of diesel fuel, which means that the compression energy in the DME fuel pump is greater than that in the diesel fuel pump. The differences between diesel and DME with regards to lubricity, viscosity, bulk modulus and energy density means that many components in the fuel system must be changed when converting from diesel to DME. The fuel injection timing and duration must also be altered.

Some laboratory tests using pure DME instead of diesel have caused pump failure in less than 30 minutes. Adding a small amount of lubrication significantly increased lubricity but still not to the point where it could be considered acceptable for the typical expected life of a highway tractor. They concluded that raising the lubricity of DME to acceptable levels may not be possible but changing the designs of the pumps to accept pure DME could be a much more viable option. Some researchers have concluded that one of the more significant challenges in using DME as a diesel-fuel substitute is the modification, tuning and management of the engine fuel delivery system.

Particulate Matter (PM), or soot formation, in a DME-fueled engine is almost zero because DME has an oxygen content of 35% and no carbon-to-carbon bonds. Many tests and research programs have demonstrated that DME powered vehicles have PM levels that are orders of magnitude less than diesel PM levels and can pass all current worldwide emissions regulations without the use of any type of diesel particulate filter or trap. Studies have shown that as much as 99% of the PM released from a DME engine is in the nano particle size, which can cause more damage to human health than the larger particle sizes. However, it is not clear from the research if the absolute volume and count of PM particles could pose a risk to human health as a result of tailpipe emissions from Canadian vehicles given that this could be 99% of what is already a miniscule value.

The relationship between NOx formation, the use of selective catalytic reduction (SCR) and exhaust gas recirculation (EGR) is less clear than the formation of particulate matter. The near zero levels of PM means that PM burn off in the engine is not required thus in cylinder temperatures can be lowered, which reduces the levels of NOx. Some tests have demonstrated that DME powered vehicles can be operated without the use of SCR for NOx reduction with the addition of light EGR to reduce NOx levels and diesel oxidation catalysts to reduce carbon moNOxide and hydrocarbon levels to below emissions regulation levels. However, most of these tests were conducted when emissions regulations were slightly less stringent than today. More testing will be required to determine if a DME powered vehicle can pass current emissions regulations for PM, NOx, CO and HC without the use of SCR and what level of EGR and catalysing would be required to compensate for this lack of SCR. Additionally, it will be important to maintain a level of EGR that does not increase fuel consumption or contribute to engine wear. The primary benefits of removing SCR from vehicles are as follows: reducing cost, reducing weight, reducing the need to replenish a consumable fluid and removing a piece of equipment that can cause an engine derate when a fault occurs, be it actual or nuisance.

Research has shown that HC and CO emissions from a DME-fueled engine are usually lower than or equal to that of a diesel engine. However, if SCR is removed and EGR is used to cool the in cylinder temperature it will result in higher levels of CO and HC which can then require the use of a diesel oxidation catalyst to reduce CO and HC levels. Additionally, if high levels of engine injection advance are requested (more than 20 degrees) it can result in levels of CO and HC that are significantly higher than what is currently allowed in North America for on-road and off-road vehicles. Therefore, careful consideration must be given to the inter relationship between engine timing and EGR levels otherwise the reductions of PM may be offset by increase in other pollutants.

DME produces less exhaust, by mass, than diesel and less CO2 than diesel. DME produces more water than diesel but not orders of magnitude more (as is the case with fuel cells) therefore it is assumed that the incremental effect to cold road surfaces would be minimal, when compared to diesel. For every 43 MJ of fuel energy, DME produces approximately 0.70 kg less total exhaust, 0.30 kg less CO2, and 0.55 kg more H2O than diesel fuel. This may not be a large reduction in CO2 but it does mean that an engine that is slightly above the maximum allowable levels for CO2 under Canada's new GHG reduction strategy could potentially pass the regulations if fueled with DME, assuming horsepower and fuel consumption remained the same.

The Carbonyl compounds such as formaldehyde (HCHO), acetaldehyde (CH3CHO), and formic acid (HCOOH) have all been found to be higher for DME than for diesel. This is due to the high Oxygen content and additives in DME whereas diesel is Oxygen-free. Formaldehyde can be reduced to negligible levels by the use of a diesel oxidation catalyst whereas catalysts have no effect on the reduction of acetaldehyde. Careful attention will need to be paid to the lubricity additives that are mixed with DME as they can adversely affect the generation of these Oxygen based compounds and the reduction in PM could be offset by an increase in other toxins.

It is clear that any DME powered vehicle can pass any emission standard in the world (at the time of report preparation) for PM without the use of a particulate filter or trap.

DME's energy density (1.88 DGE) on a volumetric basis is lower than that of LNG (1.56 DGE) but higher than that of CNG (3.98 DGE). Other alternative fuels such as CNG, LNG and LPG are all used in spark ignition engines and therefore require special maintenance such as spark plug replacement. DME powered engines can be maintained similarly to diesel engines with more attention being paid to fuel lubricity to prevent premature failure of pumping and fuel injection components. 

DME is stored at pressures that are higher than diesel, but similar to LPG and much lower than CNG.

CNG and LNG fuel tanks are significantly heavier and more expensive than DME fuel tanks. Some papers indicate that LNG and DME tanks can be more than 350 kg heavier than similarly sized DME tanks. Exact figures were difficult to obtain but the size and weight of DME fuel tanks should only be marginally higher than diesel fuel tanks, however, range will be reduced due to the lower energy content of DME.

Because DME has approximately only two thirds of the energy content of diesel by mass, and 50% by volume and only 80% of the physical density, the fuel consumption data must be multiplied by the diesel gallon equivalency (DGE). Typical DME raw test results are between 2.5 mpg and 3.0 mpg at cruising speed. However, when DGEs are considered, the values correspond to fuel consumption rates of over 5.00 mpg which is consistent with current heavy duty tractor fuel consumption values.       

DME can be stored for long periods of time in outdoor storage tanks that are exposed to direct solar radiation without any boil off or venting.

As a minimum, DME powered vehicles will require a new fuel tank, new fuel lines, new fuel pumps and injectors as well as new/improved seals and gaskets and modified engine management mapping and software to control the timing of the fuel injections.  The addition of a larger and heavier pressure vessel fuel tank (similar to LPG) and some form of EGR could possibly be offset by the removal of the diesel particulate filter and possibly the SCR componentry.  Whereas LNG and CNG vehicles must carry very heavy pressure vessel tanks that can reduce payload capacity, the net effect on weight to a heavy duty vehicle powered by DME should be negligible when compared to current diesel powered vehicles. If SCR can be removed there will likely be a small weight savings for DME. However, the range of the vehicles will be reduced to approximately 50% of current distances, which could necessitate a much larger tank which would then increase the weight to current diesel truck levels.

Many other minor vehicle modifications must be implemented such as pressure relief valves, non sparking metal components such as brass, shields and valve covers etc. However, none of these components will add enough weight to seriously affect the payload capacity of a load carrying vehicle.

Unlike most LNG/CNG/LPG engines that require spark plug replacement at specified intervals, DME engines can be serviced in a similar fashion as other compression ignition engines. Despite the fact that many components on DME engines must be changed from conventional diesel engines, they are still the same type of components so the maintenance philosophy can remain.

The low lubricity of DME could necessitate, at least in the near term, an increase in fuel component inspections and or replacement. More data will be required to determine the long term effects on the reliability of all fuel wetted components for long haul operations where fuel pump failure far away from the maintenance base would be considered a highly undesirable situation.

Approximately 7% less air, and presumably 7% less dirt and contaminants, will be drawn into a DME engine compared to diesel engines. It is not likely that air cleaner/filter intervals will need to be modified based on this relatively small change in air flow. Nonetheless, DME engines should be slightly cleaner than diesel engines under identical situations.

Diesel engines currently require oils that are specifically formulated to manage the soot that is generated inside the combustion chamber. More study will be required to determine if the lack of soot in DME powered engines could, or should, result in the use of a different grade of oil, perhaps more similar to those used in gasoline engines.

Vapour fuel leaks under the hood can cause engine over-speeding as raw fuel is ingested into the air intake system.

More study would be required to better understand how DME would affect the mean time between failure of any fuel wetted component, including the fuel system and the engine itself.

In general, DME behaves like propane/LPG and can be handled as such. As with LPG, DME is heavier than air and can pool on the floors of underground garages thus preventing an explosion hazard if a source of ignition were to be dropped into the pool.

Fire fighters will be required to receive training on the ways to extinguish a DME fed fire since improper technique can lead to explosions. DME is considered a dangerous good in the hazard class 2.1 and must be transported in a vehicle with a placard mounted that contains the necessary information.

The MSDS sheets for DME indicate that personnel who handle the product should wear gloves and eye protection as well foot protection. The level of protection is similar to personnel who are dispensing LPG and CNG and slightly less than those dispensing LNG.

From the driver's perspective there is very little to be concerned with outside of the need to fully understand how to fill the DME tank and understanding any range limitations that may differ from their diesel powered vehicles so that they are not stranded in inclement weather. The vehicle will look and feel the same and will likely be slightly quieter with less knocking at start up when compared to diesel.

Very little data could be found related to the performance of DME in cold weather climates such as Canada.  The joint Volvo and Chemrec DME project was a two year project that operated continuously throughout all of Sweden's four seasons, including their cold winter months.  Some of the operations were conducted as far north as the 65th parallel which runs through all three Canadian territories. The test team did not report any operational issues related to ice, snow or cold temperatures. It is not known if Sweden uses salt on their roads to the same extent that is used in some Canadian provinces. Some specific recommendations for cold weather operations were found involving the use of corrosion resistant tanks and fasteners to minimize the risks of salt induced corrosion perforation. However, these recommendations are likely already 'best practices' for diesel vehicles. Rigorous yearly inspections may be required to ensure that fuel tank integrity has not been compromised due to corrosion. DME tanks are pressure vessels and are therefore more rugged than conventional diesel tanks thus the risk of tank failure is actually lower. However, the consequences of a tank failure are potentially more severe since it is a pressurized gas that could be expelled at high velocity whereas diesel fuel remains, at all times, in the liquid unpressurized phase.

At the time of report preparation there were no known DME fuel stations in Canada. This is in contrast to the approximately 100 public and private LNG and CNG stations found across Canada and the abundance of diesel fuel stations. This lack of publicly available DME means that the use of DME in heavy vehicles will likely have to be staged if is to be accepted by the industry. The first phase of DME use would most certainly be for use in fleets that return back to their base every evening for refuelling, be they public or private. These could include urban transit buses, waste collection vehicles, vocational vehicles such as concrete mixers etc. The current lack of infrastructure would make it nearly impossible for long haul tractor trailer operations or for intercity motor coach buses. The reduced range of DME vehicles combined with the fact that, once converted, DME vehicles can no longer operate on any other fuel will dictate that a significant DME network must be constructed along major corridors to support fleet operators who choose to commit to using DME.

The consumer cost of a litre of DME is extremely difficult to quantify at this time. The International DME Association states that the consumer cost of DME is roughly 75% to 90% that of LPG. They also noted that pure LPG prices fluctuate more since LPG is a petroleum based fuel and must follow global petroleum pricing. Oberon, the largest producer of DME in the United States simply states that DME is "competitive with diesel prices" but does not list a consumer cost on their website.

Until more accurate data are available, it is fair to assume that the cost per km of delivered DME, on a diesel gallon equivalency, is approximately the same as diesel fuel.  The cost per litre may be lower, but since more DME must be burned per distance driven, the cost per km of the fuel should be similar. However, it must be restated that not enough data were available to perform such a calculation with any certainty.

More study may be required to determine the quantity and effects of the nano PM particles that are being created by a DME powered engine, not only in terms of human health but in terms of engine wear;

Engine oil and engine wear analysis could be performed to determine the long term effects of using diesel engine oil, gasoline engine oil or combinations of these to understand the full effects of engine wear by DME;

The levels of carbonyl compounds were difficult to quantify therefore further work in this area may be warranted in order to fully quantify these emissions and their potential to affect human health and infrastructure;

A computational fluid dynamics study could be undertaken to determine how DME pools and accumulates in enclosed and confined spaces such as parking garages. This will be important information to have for property owners and vehicle maintenance shop owners who need to mitigate the risks of possible leaks on their properties;

A full well to wheel analysis of a few of the production methods of DME would be useful in determining how DME compares to, say, biodiesel in terms of emissions and production costs;

A full economic analysis could be performed using CNG, LNG, LPG, DME and conventional diesel to quantify the cost of fuel, consumables and maintenance to operate a long haul tractor for its entire life cycle. It is likely that this has already been performed for the non DME fuels therefore the addition of DME would provide a useful comparison as no such study for those who are considering using the fuel;

One of the greatest potential benefits to operators is the possible removal of the SCR system. A project could be undertaken to test a vehicle with variable EGR and no SCR to determine if a DME powered vehicle could pass the current set of EC emissions standards without SCR, without any increase in fuel consumption. A second set of tests with SCR would be required to compare the results.

The full report can be found at: 
http://nparc.cisti-icist.nrc-cnrc.gc.ca/eng/view/object/?id=b22a2fd1-f2fc-40d9-a6f4-cf109f3ea344

Date modified: