Dimethyl ether fuel proposed as an alternative to LNG

Yoshitsugi Kikkawa, Ichizo Aoki
Chiyoda Corp.
Yokohama, Japan

To cope with the emerging energy demand in Asia, alternative fuels to LNG must be considered. Alternative measures, which convert the natural gas to liquid fuel, include the Fischer-Tropsch conversion, methanol synthesis, and dimethyl ether (DME) synthesis. Comparisons are evaluated based on both transportation cost and feed-gas cost.

The analysis will show that DME, one alternative to LNG as transportation fuel, will be more economical for longer distances between the natural-gas source and the consumer. LNG requires a costly tanker and receiving terminal. The break-even distance will be around 5,000-7,000 km and vary depending on the transported volume. There will be risk, however, since there has never been a DME plant the size of an LNG-equivalent plant [6 million metric tons/year (mty)].

Asian energy demand

According to a forecast of energy demand by the Japanese Ministry of International Trade and Industry (MITI), the Asian APEC (Asia-Pacific Economic Cooperation) countries will need 3.0 billion mtoe (metric ton oil equivalent) of primary energy in 2010. ¹

The following five energy forms are expected to be the source of primary energy in 2010: oil, natural gas, coal, nuclear, and hydroelectric.

According to BP Statistics, 24% of the primary energy consumed by the world in 1996 was natural gas. Table 1 [36,973 bytes] shows a breakdown of fuel consumption by the world and by Asia-Pacific.²

The Asian region will need about 300-600 million mty of natural gas in 2010, which cannot be solely supplied from Asia. Asia will need to import gas from natural gas resources in the Middle East or Russia. Table 2 [68,142 bytes] lists the giant gas reserves that may be available for Asia.

India's demand for natural gas in 2005 is estimated at 7 bscfd.³ As a result of its recent economic growth, India will need 10-12 gw/year power capacity in the next decade.⁴

Asia, a large importer of LNG, predominantly uses LNG as fuel for power generation. The percentages of natural gas used for power generation in Japan, Korea, and Taiwan are 70%, 55%, and 40%, respectively.^{5 6}

LNG and pipeline

Natural gas is a favorable energy source to minimize SOx, NO _x, and CO2 emissions.

Historically, there are two transportation measures for natural gas transportation: LNG or pipeline. The pipeline infrastructure has been successfully built in North America and Europe, while LNG has been successfully applied between Algeria and Europe, Alaska and Japan, Southeast Asia and the Far East, the Middle East and Japan, and Australia and Japan.

LNG is used for transportation over long distances or from the deep sea, which may be uneconomical or a hurdle for pipeline construction.

Recently, several ideas have been introduced for the conversion of natural gas via liquids before transportation. The suggested liquids include syncrude or middle distillates (via the Fischer-Tropsch process), methanol, or dimethyl ether (DME).

Gas-to-liquids technology

The Fischer-Tropsch synthesis technology was developed during World War II in Germany. Sasol Ltd. and Royal Dutch Shell have successfully produced liquid fuel from synthesis gas, which is produced from coal or natural gas. Exxon Corp. developed AGC-21 (Advanced Gas Conversion Technology 21st Century) and is ready to build a commercial plant.

Fischer-Tropsch synthesis is economical when crude oil prices are over $20-$25/bbl in future.

Sasol Ltd.

Using a circulating fluidized-bed reactor, Sasol has successfully produced liquid fuel from synthesis gas derived from coal since 1955. Sasol developed a slurry-bed reactor which can scale up to 14,000 b/d. In addition, Sasol has been producing 2,500 b/d of liquids at its Sasolburg complex with this unit since May 1993.

South Africa currently consumes about 450,000 b/d of liquid fuels; net imports amount to 255,000 b/d. The remaining 195,000 b/d is manufactured both from coal by Sasol and from natural gas by state-owned Mossgas. Sasol provided technology for Mossgas' natural gas-to-synthetic fuel plant in South Africa, which began operations in 1993.⁷

Sasol claims, "The products will fetch at least conventional fuel prices, and potentially also some premium due to their environmental advantages. At crude oil prices in the $16-18/bbl range, the product prices are expected to lie in the $22-25/bbl range. At these prices, the pretax return on investment is between 12 and 15%. For multiple modules the return on investment will increase due to the advantage of economy of scale."

The South Africa government, however, provides a subsidy for Sasol to compete in the international oil market.

In July 1997, Sasol, Qatar General Petroleum Corp. (QGPC), and Phillips Petroleum Co. signed a memorandum of understanding for a feasibility study for a joint venture gas-to-liquid project in Qatar. The project would be located in Ras Laffan Industrial City, Qatar. The plant, if approved, is scheduled for start-up in 2002. It would produce about 20,000 b/d of high-quality distillates and naphtha. The project would draw its natural gas feedstock from Qatar's extensive reserves.^7-10

Royal Dutch Shell

The Shell Middle Distillate Synthesis (SMDS) process was applied to the Bintulu plant at a capital cost of $850 million. The plant converts 100 MMscfd of natural gas to 12,000 b/d of product ranging from naphtha to paraffinic waxes. The plant came on stream in May 1993.

Because the Bintulu SMDS plant is small and uneconomic in comparison with conventional middle-distillates production, Shell targets the specialized, high-return wax markets. Middle distillates from Bintulu are used by Shell and its customers, particularly in California, as an additive to their diesel fuels to bring them into compliance with stringent emissions regulations.^{7 11}

Exxon Corp.

Exxon Corp.'s AGC-21 process runs a 200 b/d gas-to-liquids (GTL) pilot plant at Baton Rouge. The company unveiled its process by announcing negotiations with QGPC to build a GTL plant to use some of the gas in Qatar's North field. Exxon wants to build a plant to convert 500 MMscfd-1 bscfd of natural gas to 50,000-100,000 b/d of middle distillates, naphtha, and catalytic cracker feedstock.

According to Exxon, the AGC-21 process has been developed to the point where scaling up to commercial plant size is within reach. A project in Qatar would cost $1.2-$2.4 billion, which equates to about $24,000/b/d of output. This is well below the $30,000/b/d cost reckoned to be the break-even point for GTL at 1997 oil prices.^{7 12}

Syntroleum Corp.

Syntroleum Corp., Tulsa, which developed a gas-to-liquids process, took out its first patents in 1989. The following year, Syntroleum built a 2 b/d pilot plant that is used to test catalyst performance, reactor designs, and operating procedures. The company is licensing the technology to petroleum majors.

Methanol from natural gas

Two decades ago, methanol was studied as an alternative to LNG. Table 3 [23,737 bytes] shows the results of one of those studies.

Methanol is an unlikely alternative for LNG as a future feed gas. Although its transportation and receiving terminal costs are much less than those of LNG and DME, its thermal efficiency is lower than both LNG and DME.

P. Soedjant said in a paper presented at the 4th International Conference on Liquefied Natural Gas that methanol is more economical for transportation than LNG if the distance is more than 6,000 nautical miles (11,000 km), given that the natural-gas cost is 50¢/million BTU.¹³ If the study were reviewed based on a today's natural gas cost, which is higher, the break-even distance would be higher.

Although there has never been such application, General Electric Co. has tested methanol as a fuel of gas turbine, with favorable results.

Since methanol is a toxic material and soluble in water, the danger of a methanol spill from a tanker accident will need to be assessed. The amount of methanol required for power generation is ten times that of current international trade of chemicals on tankers; thus, the chances of a tanker spill will increase.

Dimethyl ether

Recently, DME has been considered as an alternative to diesel fuel or LNG. The stringent 1998 Californian Ultra Low Emission Vehicles (ULEV) regulations can be met by using DME as fuel.

Today, DME is predominantly used as an aerosol propellant as a result of its environmentally benign characteristics. It is a colorless gas with a faint ethereal fragrance. DME is not harmful to the ozone layer as the previously used chloro-fluorocarbon (CFC) gases were. It is also virtually nontoxic and is easily degraded in the troposphere.

The vapor pressure at ambient temperature is about 5 bar, which makes DME similar to LPG with respect to its physical properties. DME is relatively inert and noncorrosive. Unlike diethyl ether, it is noncarcinogenic. Also, DME does not form peroxides by prolonged exposure to air.¹⁴ Basic chemical and physical properties are listed in Table 4 [42,784 bytes].

DME has a heating value (41.0 MJ/kg) lower than that of diesel (about 42.5 MJ/kg) and LNG (about 50 MJ/kg) but higher than that of methanol (19.7 MJ/kg). Its vapor pressure is between that of propane and butane, so it can be handled as LPG. Due care, however, is necessary in the selection of materials for gaskets.

Therefore, DME is suitable for the modern gas-turbine power generator. The thermal efficiency of the process is higher than that of methanol by 5-10%. The combined cycle will perform with high thermal efficiency, over 55%, as well as LNG using GE's modern FA technology.

Synthesis of DME

To become an alternative to LNG, DME must be produced in large quantities to meet the demand. DME is currently produced by fixed-bed catalytic dehydration of methanol.

In large-scale manufacture, however, there are significant advantages in integrating the methanol and DME synthesis steps into one single process. This process will convert synthesis gas directly into DME.

The advantages obtained by combined synthesis of DME and methanol can be appreciated by thermodynamic considerations: DME synthesis from synthesis gas involves three reactions (heat of reaction given in kJ/mole in parentheses):

CO₂+ 3 H₂ CH₃OH + H₂O (50.1) (1)
H₂O + CO H₂ + CO₂ (40.9) (2)
2 CH₃OH CH₃OCH₃ + H₂O (23.4) (3)

A flow diagram of DME synthesis is shown in Fig. 1 [66,460 bytes].

All three reactions must take place simultaneously. The simplest way to achieve this is to apply a physical mixture of classical methanol and DME catalysts. Such a solution, however, could create problems with regard to selectivity, particularly at high temperatures. The problems could lead to excessive formation of byproducts, mainly higher alcohols and hydrocarbons. Because the overall reaction is highly exothermic, the ideal catalyst must exhibit high stability and preserve high selectivity at high temperatures.

Haldor Topsøe has already successfully developed catalysts with such properties, designed specifically for application in acetyls manufacture in which high selectivity is important.¹⁴

Costs for LNG vs. DME

The transportation costs for an LNG case and a DME case were reviewed based on several assumptions for transportation distances, capacity, economic factors, and performance.

The LNG project costs, culminating in the LNG costs to the consumer, are shown in Table 5 [159,353 bytes]. The DME project costs and DME costs to the consumer (i.e., the outlet of DME receiving terminal) are shown in Table 6 [174,484 bytes].

The gas costs for the consumer are summarized in Fig. 2 [100,926 bytes]. For long distances, more than 5,000-7,000 km, the DME case has a lower cost per million BTU than LNG.

Cost bases

The maximum transport distance to transport natural gas was assumed to be 12,000 km. The minimum transport distance was assumed to be 2,000 km. The Middle East will be a major supplier of natural gas to East Asia. Table 7 [33,041 bytes] shows how these distances were estimated. Eventually, the distances were taken as 2,000 km, 6,000 km, and 12,000 km, respectively.

For the capacity of an LNG plant, nominal capacities of 6 million mty, 12 million mty, and 18 million mty were assumed. These capacities require feed-gas quantities of 1,100 MMscfd, 2,200 MMscfd, and 3,300 MMscfd, respectively.

The low-end nominal capacity was based on a recent 3.3 million mty LNG plant expansion in Oman;¹⁵ the plant capacity was taken as 6.6 million mty since most LNG plants need at least two trains for reliability reasons. The maximum capacity was assumed to be about 20 million mty which is similar to the LNG plant in Badak, Indonesia, which will have eight trains of around 2.5 million mty each.¹⁶ Capacities over 20 million mty will be unlikely, from a safety and security viewpoint.

Capacities of the DME plant were determined by looking at LNG capacities with the same product heating value. The train capacity was assumed to be 5,200 tons/day which is the LNG equivalent 1.0 million mty, which will cope with a single large oxygen plant. For the equivalent LNG 6 million mty, 12 million mty, and 18 million mty plants, the DME plants will need feed gas of 1,250, 2,500, and 3,750 MMscfd, respectively.

Table 8 [14,177 bytes] shows the feed gas composition that was assumed in the calculations. The feed gas pressure was assumed to be 65 bar. The molecular weight used was 17.23. The high-heating value and the low-heating values used were 1,065 BTU/standard cu ft and 961 BTU/standard cu ft, respectively.

Economic factors used were: $0.50/million BTU for the feed gas cost at the wellhead, 15% fixed-charge factor, a 2¢/kw-hr rate for exported plant electric power, and an operating and maintenance factor (excluding fuel) of 4.5% for tankers and 4% for a plant.

The performance assumptions of the LNG and DME plants are shown in Table 9 [24,859 bytes]. That for the tankers is shown in Table 10 [22,485 bytes].

DME plant cost estimate

The DME process is similar to methanol synthesis. There are significant advantages in combining the methanol, water-gas shift, and DME synthesis into one single step. The reforming of the natural gas for large capacity uses an auto-thermal reforming reaction. Although an oxygen plant is required, this reaction has several merits over steam reforming:14

• Compact design
• Possible stoichiometric synthesis gas
• High reformer pressure eliminating or reducing make-up gas compression
• No NO_x emissions
• Low cost for large single train.

A plant consists of the process battery limits (BL), utility, and off site processes.

The process BL cost was estimated based on a methanol plant cost and a scaling factor listed in literature.14 The methanol plant cost was estimated based on Stanford Research Institute reports.¹⁷ The power generator in the offsite/utility facility will supply the plant required power including the process BL, the cooling water pumps, and the refrigeration power to cool the DME for the atmospheric storage tanks. Excess power will be exported for use outside of the plant.

The storage tank capacity will be about double the size of the DME tanker. The off site facilities include cooling-water treatment, steam generation, storage and loading, refrigeration, power generation and export, the flare system, and waste-water treatment. These costs were estimated by Chiyoda's in-house data.

DME project size

Because no DME plant of such large capacity has ever been built, the technology will be extrapolated from current methanol technology.

For cost comparison, the DME train capacity was assumed to be 5,200 tons/day which is equivalent in heating value to a methanol plant of 7,300 tons/day in plant size. This capacity is about three times that of current large-scale methanol plants; current methanol plant size is about 2,500 tons/day. If the train capacity is increased, the scale will result in a more competitive energy cost. It took 2 decades to increase the LNG plant's train capacity by three times.

The DME project will need a big investment similar to LNG project and will have technical risks since such a plant has never existed. Therefore, a detailed RAM (reliability, availability and maintainability) assessment will have an important role.

After successful operation of the whole DME chain, however, the DME will serve as fuel for power generation for the longer transportation distances. Moreover, as a result of environmental regulations, DME may be used in the future as an alternative to diesel fuel for road transportation.

References

1. Toichi, T., Asian Natural Gas VI, Singapore 1996.
2. BP Statistical Review of World Energy 1997, British Petroleum Co., p. 38.
3. World Gas Intelligence, Oct. 27 1995.
4. Sharama, R.P., 2nd Doha Conference on Natural Gas, 1997.
5. Kwan, Y.J., 2nd Doha Conference on Natural Gas, 1997.
6. Chang, H.C., 2nd Doha Conference on Natural Gas, 1997.
7. Knott, David, "Synthetic fuels aiding South Africa to advance its world energy role," OGJ, Mar. 17, 1997, p. 23.
8. Project News 1997(Japanese Magazine).
9. Jager, B., Paper No. FRT1-03 World Gas Conference 1997.
10. News, Phillips Petroleum Home Page.
11. Jacometti, J., Paper No. FRT1-02 World Gas Conference 1997.
12. Eisenberg, B., et. al., 73rd Annual GPA Convention 1994.
13. Soedjant, P., et. al. 4th International Conference on Liquefied Natural Gas, 1974.
14. Hansen, J.B., et al., SAE Paper 950063, 1995.
15. The LNG Observer, Winter 1995-1996.
16. The LNG Observer, Fall 1995.
17. SRI Project 3978-10, Global Outlook for Oxygenates 1990-2000, 1992.

The Authors