RADICALLY NEW GTL PROCESS DEVELOPED AT TEXAS A&M UNIVERSITY

At the American Institute of Chemical Engineers’ Spring Meeting held in Houston, Texas, in April, K. Hall et al. presented a radically new process for converting natural gas into hydrocarbon liquids.

Conventional wisdom has gravitated to some variation of Fischer-Tropsch technology to produce hydrocarbon liquids from natural gas. Texas A&M has developed an entirely different approach for this conversion and the Texas A&M University System has licensed the technology to Synfuels International of Dallas, Texas, for commercialization.

The new process has several advantages: It is simple; it should be economical for flows ranging from 3 million standard cubic feet per day (MMSCFD) through 500 MMSCFD; it can be skid-mounted (at lower flow rates) for easier transportation; it appears that the cost of fluids produced would be about $12 to $15 per barrel; it is energy self-sufficient, water can be a byproduct, and the product is a light naphtha with about a C6 molecular weight that can be converted into a heavier fraction with extra processing. The paper presents some details of the process and discusses a 100-thousand standard cubic feet per day pilot unit.

Fischer-Tropsch technology provides versatile processes that convert natural gas into syngas (a mixture of carbon monoxide and hydrogen) and then converts the syngas into liquids, gases and solids. A large amount of research has advanced the technology through catalyst development and process optimization. However, the process is complicated and requires many unit operations to achieve its purpose. As a result, the plants are rather expensive and require a relatively high price for oil to be competitive.

Another technology that can achieve the desired results is plasma processing. By subjecting the natural gas to a plasma torch or by using the natural gas as the plasma gas, it is possible to convert natural gas into acetylene and then to process the acetylene into liquids. While this technology is less complicated than Fischer-Tropsch, it is not much less expensive. The plasma is expensive to produce (requiring a large amount of electricity) and difficult to operate in a stable manner.

At Texas A&M, researchers began developing a plasma process that was thought to have some inherent advantages for converting natural gas into liquids. However, along the way they discovered a better way to perform the chemistry without resorting to a plasma process. The process discovered is less expensive than a plasma process and much more stable and easier to operate.

The paper by Hall et al. presents the new process and discusses bench-scale tests as well as a pilot unit. The pilot unit has a capacity of 100,000 standard cubic feet per day of natural gas and produces about 10 to 12 barrels per day of light naphtha as product.

The Synfuels process is essentially two reaction steps and two separation steps. Of course, the process contains compressors, heat exchangers and other unit operations for efficient operation, but the reaction and separation steps are the crux of the technology. It is assumed that the gas has been treated to remove most of the heavy components because it would be more efficient to do so using conventional technology. It is also assumed the gas has been sweetened because hydrogen sulfide would pass through to the product making it sour and less useful. The process would convert carbon dioxide in the feed into carbon monoxide and ultimately use it as fuel, but it is preferred to keep CO2 out of the process.

To begin the process, the inlet gas mixes with a hydrogen recycle stream to enter the first reactor through a preheat step that raises the temperature to approximately 500ºC or 1,000ºF. The hydrogen serves to reduce coke formation in the reactor. The reactor is an electrically heated furnace that cracks the natural gas hydrocarbons into olefins, primarily ethylene and/or acetylene. Residence time in this reactor is crucial and must be less than 70 milliseconds to keep coke production to a minimum. For relatively lean gas the residence time would be relatively longer (approaching 70 milliseconds), the temperature would be relatively higher (on the order of 1,300ºC or 2,400ºF), and acetylene would be the dominant, reactive C2 product. For relatively rich gas, say on the order of 10 mol percent C2+, the residence time would be relatively shorter (approaching 20 milliseconds), the temperature would be relatively lower (approximately 900ºC or 1,700ºF) and ethylene would be the dominant C2 product. In either case, the product stream contains significant unreacted methane and hydrogen (both from recycle and reaction). Nitrogen passes through unchanged (although hydrogen cyanide could result thermodynamically, the kinetics are too slow to produce measurable amounts before the reaction is quenched). The reaction is quenched with water at the exit of the reactor to approximately 500ºC or 1,000ºF.

After cooling and washing steps, this stream mixes with a light gas recycle stream from the liquid separator. The recycle stream contains C1 and small amounts of C2-C4 hydrocarbons, any carbon monoxide and nitrogen. The combined stream passes to a hydrogen separator in which nearly all of the hydrogen passes to recycle or to turbine fuel. This separator can be a membrane unit, a pressure swing adsorption unit or a fractionator. The hydrogen that does not recycle and a purge stream from the light gas recycle constitute turbine fuel to produce the electricity required by the process. The combined stream contains sufficient energy to produce enough electricity to run the entire process.

The hydrocarbon stream exiting the separator passes to a catalytic reactor. In this reactor, the reactive C2 molecules combine with the methane to make heavier hydrocarbons. Bench tests have produced a liquid product with about C5 average molecular weight. This light naphtha consists of mostly isoalkanes.

The outlet from the catalytic reactor passes to a standard separation unit that produces the liquid product and the light gas recycle.

A reasonably good estimate of cost for a 10-MMSCFD process is $20 million ($17,000 to $20,000 per barrel per day capacity). In addition, product cost is estimated at $12 to $15 per barrel (capital cost and operating expenses) assuming remote gas at $0.50 per MSCFD, 10-year depreciation on a $25-million capital investment and $1 to $3 per barrel for operating costs.