GTL technologies focus on lowering costs

GTL OUTLOOK-1

Michael J. Corke
Purvin & Gertz Inc.
London

Difficulties in the development of major natural-gas production projects and the limitations imposed by saturated markets for LNG or pipeline gas have focused attention on alternative gas utilization approaches. At the same time, technology improvements have transformed the Fischer-Tropsch (F-T) conversion of natural gas-to-liquid (GTL) hydrocarbons from a technically interesting but uneconomic option into an option worthy of serious consideration.

This two-part series reviews GTL technology developments which have led to today's situation (Part 1) and examines the economics of GTL conversion (Part 2). The economic viability of GTL projects mainly depends on feed-gas pricing, investment costs, and the potential to produce liquids with natural-gas production.

Technology overview

The process of converting natural gas to marketable liquid hydrocarbons comprises three main elements ( Fig. 1 [99,518 bytes]): synthesis gas (syngas) production, GTL synthesis, and product work-up.

The feed gas will typically be treated prior to syngas production to remove sulfur compounds. Sulfur is a poison to the catalysts in syngas production (if catalysts are necessary) and in the F-T reaction. Sulfur compounds may also cause corrosion, safety, and environmental problems.

Sulfur removal may be accomplished by either adsorption onto zinc oxide or by molecular sieve absorption if the sulfur content of the gas is not excessive. Other solutions such as Merox washing are required for feed gas with higher sulfur contents.

It may also be necessary to remove trace amounts of heavy metals present in the feed gas. Mercury, for example, can act as a catalyst poison and lead to metallurgical problems.

Syngas production

Syngas is a mixture of carbon monoxide and hydrogen. It can be produced from a variety of hydrocarbon feedstocks, including natural gas. It is the building block for the production of hydrocarbons in the F-T conversion process. Syngas is also the building block for the production of methanol and ammonia.

For the production of hydrocarbons using the F-T conversion process, the ideal molar ratio of hydrogen to carbon monoxide is just below two.

Syngas of a composition ratio appropriate to the F-T conversion can be made by one or more of several processes working in parallel, in a combined mode, or with the addition or extraction of hydrogen.

The established syngas production technologies, summarized below, are all high temperature and high pressure processes. The exit gases must be reduced to temperatures suitable for the next process, F-T synthesis. This can be done by using the hot gases to generate steam or by direct water quenching.

• Steam Reforming. Steam reforming is the longest established technology. It is offered by licensors such as Foster Wheeler Corp., Haldor Tops e AS, Kinetics Technology International BV, Lurgi AG, and Uhde GmbH. Steam and natural gas are reacted at high temperatures (800-900° C.) and moderate pressures (20 bar) in the presence of a nickel catalyst. The reaction produces a mixture of hydrogen, carbon monoxide, and carbon dioxide (Fig. 2 [112,765 bytes]).

Steam reforming produces a high hydrogen to carbon monoxide ratio. To some extent, this can be controlled by adjusting the proportions of natural gas and steam in the reactor feed. Steam reforming in isolation, however, does not yield the ratio ideally required for the F-T conversion of syngas to hydrocarbons.

• Partial Oxidation. Partial oxidation is the second major technology used for the production of syngas. It can also be used for the production of syngas from coal or from heavy oil products. Licensed technology is offered by Texaco Inc. and the Royal Dutch/Shell Group.

In this process, natural gas is burned at very high temperatures (1,200-1,500° C.) and very high pressures (140+ bar). A catalyst is not used. For large scale operations, pure oxygen is normally supplied rather than air. An oxygen production (air separation) plant is therefore required. The reaction equation shows that this process produces a near-ideal hydrogen to carbon monoxide ratio with natural gas as the feedstock (Fig. 2).

• Autothermal reforming. Autothermal reforming is a hybrid process which concurrently reacts natural gas with steam and oxygen in the optional presence of carbon dioxide. Process licensors include Lurgi and Haldor Topsøe.

Carbon dioxide, when required, is typically provided through a recycle system. Partial oxidation occurs first, followed by steam reforming in the catalyst bed. Autothermal reforming takes place in a single reactor vessel.

Investment costs are therefore somewhat reduced. The composition of syngas may be controlled by varying the proportions of the feed streams.

Other hybrid technologies available include the sequential use of steam reforming with autothermal reforming.

F-T conversion

The second step in a typical GTL process converts syngas into paraffinic and olefinic hydrocarbons of varying chain lengths. Of course, the syngas could be used for the production of other liquid products, including methanol, ammonia, or dimethyl ether (DME). In turn, methanol could be subsequently converted into other liquid products, including gasoline blending components, and ammonia could be converted into fertilizers. However, the production of longer chain hydrocarbons is the focus of this study.

The F-T process was first developed by Hans Fischer and Franz Tropsch working in Germany in the 1920s. The first large scale application of the technology was also in Germany, where it was used for the production of synthetic liquid fuels from coal during World War II. Further development occurred in the U.S. in the late 1940s and early 1950s. Next, South Africa, prompted by its political isolation from world markets and its concerns about energy security, also developed an F-T process.

Essentially all of this early work on syngas was based on the gasification of coal. GTL projects based on natural gas are a recent development. Today, there are only two natural-gas based commercial plants. One plant, commissioned in 1991, is owned by Mossgas located at Mossel Bay in South Africa. The second plant is Shell's Bintulu, Malaysia, plant, which was commissioned in 1993.

The F-T conversion process typically uses iron or cobalt based catalysts. The process takes place at moderate temperatures (200-300° C.) and moderate pressures (10-40 bar). The primary reaction in converting natural gas to hydrocarbons is:

nCO + (2_n+1)H₂ = C_nH_2n+2 + nH₂O

Other reactions also take place in the process, resulting in the production of olefins and alcohols, and there are a variety of side reactions.

Much attention has focused on developing catalysts with appropriate selectivity and physical properties. Catalyst selectivity, syngas composition, and process conditions (principally temperature) govern the relative importance of the various reactions and the extent to which paraffinic chain lengths are limited. Higher process temperatures result in the production of higher yields of the lighter and more olefinic hydrocarbons and reduced yields of the heavier paraffins.

The F-T process typically involves the recycle of unconverted gases to the reactor, carbon dioxide removal from the recycle loop, and dehydration of the recycle gas. Sometimes, the process separates hydrogen for use in final product-upgrading processes. Some recycle gas may be used as fuel gas.

Product work-up

The hydrocarbon products from the F-T synthesis have various chain lengths. They are predominantly paraffins and n-alpha olefins if the product is directed towards middle distillate production. Some oxygenates will also be present.

The mixture may be cooled, routed to tankage, and shipped as a synthetic crude (syncrude) for processing as a feedstock in a refinery at another location. Transportation by conventional tankers may not be an option because of the high pour point of syncrude, particularly for a plant that maximizes distillate yields.

Alternatively, the syncrude may be separated and further treated at the production location to produce finished fuels, fuel blending components, or specialty products (like waxes), for local use or for export.

Various processes may be used for finishing. A mild hydrocracking/hydroisomerization process breaks down longer chain paraffins into shorter chain normal and isoparaffins with improved cold properties and boiling ranges. High quality jet-fuel and diesel-fuel blending components can be prepared in this way.

Mild hydrotreating might similarly be used to eliminate olefins and alcohols. This finishing is not necessarily an integral part of a GTL project. The decision to include a product finishing section to a GTL project is an economic and marketing question.

Current licensed options

Many companies and organizations are engaged in GTL research and development. Exxon Corp., Sasol Ltd., Shell, Syntroleum Corp., and others have developed licensed technologies that are either in use or proposed for use with regard to both the production of syngas and F-T conversion.

The technology options and developments reviewed below are based on published information and do not represent a comprehensive list.

Exxon

Exxon is operating a small scale (200 b/d) pilot plant. It is interested in developing a full scale GTL project based on its proprietary technology. Exxon's processed is termed AGC-21 (Advanced Gas Conversion for the 21st Century).

Syngas is produced by simultaneous partial oxidation and steam reforming in a single fluidized-bed reactor. Economies of scale are achieved by the use of a single, large reactor vessel.

Prior to the F-T step, heat is removed from the syngas in a steam generator. The F-T synthesis is based on a multiphase-slurry reactor which uses a proprietary cobalt-based catalyst.

Catalyst recovery is important because cobalt is expensive and unacceptable in food-grade or pharmaceutical-grade petroleum products. Exxon uses a proprietary technology to recover any catalyst entrained with the reactor products. The F-T reaction is exothermic, and heat is removed by generating steam from water tubes in the reactor.

Sasol

Sasol has been converting syngas based on coal gasification into petroleum products in South Africa since 1955.

Sasol's first commercial scale plants used the fixed and fluidized-bed processes. Three circulating fluidized-bed reactors, each of 1,500 b/d capacity, were commissioned at Sasolburg in 1955. The Arge process uses a fixed-bed reactor to manufacture wax.

Subsequent technology changes resulted in increasing reactor sizes. Sasol's Synthol process was used at Secunda where circulating fluidized-bed reactors with a capacity of 6,500 b/d were commissioned in 1980. Subsequently, Advanced Synthol gas-solid fluidized-bed reactors with capacities of 11,000 b/d were commissioned in 1995.

Currently, about 16 reactors with a total capacity of about 150,000 b/d are operated at Secunda. A project to replace these with eight advanced Synthol reactors, four of 11,000 b/d capacity and four of 20,000 b/d capacity, is presently in progress. Sasol's fluidized bed plants are directed towards the production of light olefins and gasoline blendstocks.

All of Sasol's South African plants were based on syngas produced by coal gasification. In 1991, however, South Africa's Mossgas commissioned a plant at Mossel Bay using Sasol's Synthol process. It produced gasoline and diesel fuel from a natural-gas feedstock. This plant has three reactors, each with 7,500 b/d of capacity.

More recently, Sasol has developed a slurry-phase reactor aimed at producing high-grade distillates. A demonstration-scale plant (2,500 b/d) was commissioned in 1993 and full commercial-scale technology is now being offered. A slurry reactor plant with two reactors, each of 10,000 b/d of capacity, is currently proposed for Qatar, where natural gas would be the feedstock.

Sasol is not tied to any particular syngas-production technology. For the proposed Qatar project, autothermal reforming licensed by Haldor Tops e is envisaged.

The slurry-reactor synthesis step uses a proprietary cobalt-based catalyst. Other Sasol plants utilize iron-based catalysts. Heat produced during the reaction is removed by generating steam. Operating conditions are understood to about 25 bar and 240° C.

Broadly, the Exxon and Sasol F-T reactor technologies appear similar although details differ.

Shell

Shell has operated a commercial scale GTL plant at Bintulu since 1993. Liquid-production capacity is about 12,500 b/d and the plant manufactures distillate fuels, specialty chemicals, and waxes. The plant was shut down on Dec. 25, 1997, as a result of an explosion, understood to be associated with the air-separation plant, and remains out of service at the time of writing. Recently, Shell has announced plans to rebuild the plant.

Shell is one of the major licensors of gasification technology. The Bintulu plant uses Shell's proprietary version of partial oxidation, the Shell Gasification Process (SGP). The syngas derived from the partial oxidation has a somewhat lower hydrogen content than required. This is corrected by adding hydrogen produced by a separate hydrogen plant.

In contrast to the proposed Exxon and Sasol technologies, the Bintulu plant uses a multi-tube fixed-bed reactor. The reactor operates at 40-60 bar and 1,200-1,300° C. Heat released by the reaction is removed by generating steam. The catalyst sits in the reactor tubes, and water is in the jacket. A promoted cobalt-based catalyst is used.

The syngas is treated for ammonia removal before the F-T process. Although Bintulu uses a fixed-bed reactor, Shell is understood to be working on the development of a slurry-type reactor, which may be preferred for future commercial-scale plants.

The economics of the plant have been supported by the high prices obtainable for its specialty products. Limited markets for these products suggest that any expansion or additional plant in the region would need to be based on the sale of mainstream fuel products.

Syntroleum

Syntroleum, Tulsa, has been working to develop a GTL approach that is economic for smaller scale units. The technology has been licensed to Marathon Oil Co. and Texaco. Syntroleum is operating a laboratory-scale pilot plant.

Syntroleum has announced that its technology could be economic for plant sizes as low as 2,000 b/d. On the contrary, Exxon, Sasol, and Shell are proposing plants with up to 100,000 b/d of capacity to achieve economies of scale.

The Syntroleum approach to syngas production differs from those of the licensors reviewed above in two ways: It employs autothermal reforming of natural gas for syngas production, and it uses air rather than pure oxygen in the partial oxidation step.

As outlined earlier in this section, autothermal reforming involves partial oxidation and steam reforming in a combined process. Using air instead of pure oxygen in the partial oxidation step eliminates the need for an air-separation plant. Elimination of the air-separation plant saves much cost. It results, however, in the need to handle large volumes of nitrogen within the process.

Using air rather than oxygen is not new; the production of syngas for ammonia production uses air.

It has been reported that Syntroleum is using a fluidized-bed reactor and a cobalt-based catalyst. Reactor conditions are about 20-35 bar and 190-230° C. Syntroleum is also reportedly working on other reactor configurations.

Technology developments

Essentially, the basic GTL technology has been proven, and the main barrier to the development of GTL projects is marginal or unfavorable economics. Much research and development attention is focused on the syngas production step, which for conventional processes, accounts for more than 50% of total investment costs.

Syngas production

Developments in the area of syngas production may be divided into two categories; those that represent developments of existing technology and those that represent novel technology.

The developmental approach is perhaps typified by the technology offered by Exxon. Syngas is produced by simultaneous partial oxidation and steam reforming carried out in a single fluidized-bed reactor, which allows for economies of scale. Other licensors, including BG Technology, British Petroleum Co. plc (BP), Haldor Tops e, and ICI Katalco, also appear to be focused on this type of improvement.

BG's catalytic partial oxidation approach involves partial oxidation and reforming within a fixed-nickel catalyst bed. Air, rather than oxygen, is used, avoiding the need for an expensive air-separation plant. BG claims that the technology, which remains under development, permits much higher space velocities to be achieved in the reactor, allowing much smaller and less costly reactor vessels to be used.

ICI is drawing on its latest ammonia and methanol production experience to develop a compact reformer which could be integrated into a GTL process.

In contrast to the above developmental approaches, one novel approach to the production of syngas is ceramic membrane utilization. This technology involves separating streams of natural gas and air by an appropriate ceramic membrane which allows the oxygen from the air to migrate into the natural gas, where it produces syngas via partial oxidation.

The U.S. Department of Energy (DOE) has sponsored a research program led by the Air Products and Chemical Inc., Babcock & Wilcox Co., Atlantic Richfield Co. (ARCO), and a number of other technology and research organizations. It appears, however, that commercial application would be at least 10 years away if the research is successful.

One technical or safety issue which will no doubt be considered is the physical integrity of the membrane. A failure involving uncontrolled mixing of air and natural gas is to be avoided. Ceramic membrane technology is also under study by a technology alliance between Praxair Inc., Amoco Corp., BP, Sasol, and Statoil.

The DOE study is reportedly funded to $84 million. The DOE views a commercially viable GTL technology as an opportunity to economically produce and market Alaskan North Slope natural gas.

F-T synthesis

Early F-T reactors employed fixed beds and generally used iron-based catalysts. Fluidized-catalyst beds, as in Sasol's Synthol process, subsequently developed. Recent activity, however, has focused on the development of slurry-type reactors ( Fig. 3 [139,753 bytes]).

Sasol has been operating a 2,500 b/d demonstration scale plant since 1993. Exxon's and Energy International's technologies use a slurry reactor. Shell may use a slurry reactor for new projects.

A number of advantages have been demonstrated with the slurry reactor. Improved heat transfer improves temperature control and allows the use of higher activity catalysts. The F-T reactions are exothermic, and catalyst activity is limited by the ability to remove heat in a fixed bed reactor.

In turn, higher activity catalysts allow for increased reactor capacities for any given reactor size - a cost reduction. Catalyst can also be removed and added onstream to maintain the required activity level.

Although fluidized-bed reactors also allow for improved heat removal, they are more complex and can function only where the reaction products are essentially in the vapor phase. They are therefore mainly suitable for processes that produce lighter gasoline-type hydrocarbons.

In parallel, research continues in catalyst design. Activity, selectivity, and resistance to abrasion and deactivation are key areas. Most recent work focuses on cobalt-based catalysts, which have a higher activity than iron-based catalysts.

Economies of scale

Studies have looked at plant sizes up to 100,000 b/d to achieve economies of scale. The capacity of Shell's Bintulu plant is nominally 12,500 b/d. The project proposed for Qatar by Sasol, Qatar General Petroleum Corp., and Phillips Petroleum Co. involves the construction of a two-reactor plant with a total capacity of 20,000 b/d.

A generalized mathematical relationship between investment cost and capacity can be expressed as follows:

Cost = constant * capacity^y

In this equation, the exponent y reflects the degree to which a particular process facility or equipment item benefits from economies of scale. For refining and petrochemical plants, its value is typically 0.5-0.8. A value of 1.0 would indicate a facility for which no economies of scale were obtained. Facilities for which additional capacity involves the construction of parallel trains rather than larger trains have exponent values approaching 1.0.

For GTL facilities, published cost estimates suggest an exponent of 0.66. Indicative economies of scale are illustrated by Fig. 4 [131,696 bytes]. based on investment costs for a 20,000 b/d plant in the range of $25,000-$35,000 per b/d.

Despite economies of scale, a plant with 100,000 b/d of capacity costs about $2 billion. Because of the technology risk, smaller scale projects have to be developed and proven before lenders will commit to such a large investment.

Other technology developments

Sun Refining & Marketing Co., Marcus Hook, Pa., has been working on a project supported by the U.S. DOE and Gas Research Institute. Sun is studying the direct production of liquids from natural gas using catalytic processes which mirror the action of biological enzymes. This is a leading-edge approach which will not likely show commercial results in the near term.

Technology to produce gasoline from natural gas via the intermediate step of methanol has been in existence since the 1980s although the New Zealand plant, originally built by Mobil Corp. and now owned by Methanex, is currently producing methanol only. Research is reportedly being done to modify the methanol-to-gasoline conversion process in such a way that it will be capable of producing distillate fuels.

An area of considerable interest is the potential production of dimethyl ether (DME) from natural gas via syngas. DME is a clean fuel with physical properties similar to those of LPG and appears to have advantageous environmental properties. It is suitable as a diesel fuel and for general fuel purposes.

DME technology is being developed by Amoco and Haldor Tops e. Product transportation costs would be higher than those of syncrude, which can be transported in heated conventional tankers. Similar to LPG, DME would be transported in pressurized, semi-refrigerated or fully refrigerated vessels. Although this technology is available today, its development depends mainly on economics.

Large scale, land-based projects have been the focus of most GTL work. Moreover, existing plants are all land-based. Several groups, however, have studied the development of floating GTL plants. These groups have embraced the production of both methanol and distillate hydrocarbons. A group led by Statoil and Sasol is currently active in this area.

The economics of floating systems will be driven by the value of their feed-gas stream. Gas which would otherwise be flared essentially has no value. If flaring is not permitted, then a disposal route for associated gas must be found in order to develop an oil or condensate field. Floating GTL plants have the advantage of moving on to another location once the initial source of feed gas has been exhausted. However, this is a secondary issue until technology improvements reduce costs for small to medium scale plants to reasonable levels.

Proposed projects

A number of laboratory-scale plants are being operated in connection with technology development. Rentech, a U.S.-based company, is reportedly installing a small scale plant (some hundreds of b/d) in India. This plant was previously operated in the U.S. before its feed-gas supply was exhausted.

Currently, the highest profile project proposed is the project for Qatar based on Sasol slurry-reactor technology. This would involve two reactors, each of 10,000 b/d capacity and syngas production via auto-thermal reforming of natural gas produced from Qatar's North field. Haldor Tops e licensed technology for auto-thermal reforming would be used.

The plant would be owned by Qatar General Petroleum Corp. (51%), Sasol (34%), and Phillips (15%). Phillips plans to contribute equity but not licensed technology. Foster Wheeler carried out a detailed feasibility study. The decision on whether to go forward with detailed engineering work should be made by late 1998.

Exxon's name was also mentioned in connection with a GTL project in Qatar. It has been reported that this would be a very large scale plant of up to 100,000 b/d capacity using Exxon's proprietary technology. The status of this project is unclear.

Hints about a variety of other projects, including projects using Syntroleum GTL technology and a project using Amoco/Haldor Tops e technology for DME production, have also been noted. In addition, a possible Sasol-Chevron project in Nigeria has been reported. Details of these projects, however, remain confidential or preliminary.

The Author