Tuesday, January 15, 2013

Could Diesel Fuel Made from US Natural Gas Compete with CNG and LNG?

The announcement last month of a $21 billion project to capitalize on abundant, low-cost US natural gas should have caught the attention of everyone interested in this resource. As reported in the New York Times, Sasol, a South African energy company, intends to build a 96,000 barrel-per-day gas-to-liquids (GTL) plant in southwestern Louisiana, in conjunction with a new gas processing plant and ethylene cracker. The synthetic diesel fuel produced by this facility would provide a different pathway for shale gas to displace imported crude oil in the US transportation sector, in competition with compressed or liquefied natural gas (CNG or LNG.)

GTL involves a two-step conversion of the methane that makes up the bulk of natural gas into synthesis gas and hydrogen, which are recombined into liquid hydrocarbons by means of the decades-old Fischer-Tropsch (FT) process. GTL is also energy-intensive, with an overall efficiency around 60%. South African companies have vast experience with such synthetic fuels. Sasol are partners in the Oryx GTL plant in Qatar, and their coal-to-liquids plants in South Africa utilize a similar syngas step and the same FT process as GTL.

With the US suddenly perceived to be sitting atop a century's worth of natural gas, mainly in the form of unconventional gas from shale, tight gas formations and coal-bed methane, T. Boone Pickens isn't the only one to see an opportunity to displace imported oil with gas. Yet as attractive as that sounds for reasons of energy security and trade, it isn't obvious whether the public or even fleet operators are willing to switch on a larger scale to a lower-density gaseous fuel requiring both new distribution networks and new or modified powertrains. Only 0.1% of the natural gas consumed in the US now finds its way into vehicles, equivalent to less than 0.1% of US oil demand. Under the circumstances, it would be surprising if someone weren't looking seriously at GTL, one of the few practical ways to circumvent the mechanical and logistical barriers that have impeded the fueling of more US cars and trucks with natural gas.

When I read about Sasol's proposed project, I immediately thought of another, less well-known South African synfuels facility. Since 1992 the Mossel Bay GTL plant has been turning natural gas into gasoline, diesel and other fuels, drawing first on the Mossel Bay gas field and then on newer fields as the original one depleted. Although owned by another firm, the ongoing struggles to keep the "Mossgas" plant supplied are well-known in South African energy circles. I can't imagine Sasol embarking on a project like the one in Louisiana if they had any doubt about their ability to keep it supplied for decades.

Of course volume and price are two very different aspects of supply. A decade ago, conventional wisdom held that GTL required a gas cost of around $1 per million BTUs to be viable. Even with the shale bonanza today's US natural gas price is well above that level. What now makes it possible to conceive of GTL in the US is that the price of the crude oil used to make diesel and other fuels has risen so much higher than that of natural gas. That comparison is more obvious when one converts natural gas prices into their energy equivalent in crude oil. Today's US natural gas price is below the $23 per equivalent barrel that it was in 2001. Meanwhile crude oil has increased from about $26 to $95 per barrel. The drastically improved attraction of GTL becomes even clearer when comparing ten years of wholesale US Gulf Coast diesel prices to natural gas prices using the approximate GTL conversion rate of 10 million BTUs of gas per barrel of liquid product.

As the chart above reveals, this theoretical GTL margin has exploded since 2009. Yet it also shows that if gas prices returned to the levels we experienced just a few years earlier, the proposed project would encounter significant risks. Perhaps that helps explain Sasol's concept of a larger integrated gas complex with multiple sources of margin, capitalizing on the waste heat from the GTL process and the lighter hydrocarbons it yields as byproducts.

It remains to be seen whether GTL will prove an attractive means of leveraging the US shale gas revolution to back out imported oil. However, if Sasol and others proceed with US GTL projects, anyone eyeing our gas surplus for other purposes, whether in manufacturing, fertilizer production or power generation, would face serious competition linked to the global oil market. That includes potential LNG exporters, who passed an important hurdle with the publication of a favorable analysis by the Department of Energy.

A slightly different version of this posting was previously published on the website of Pacific Energy Development Corporation


jtf said...

Another aspect that I believe you've missed here, Mr. Styles, is that a big advantage of GTL for Sasol is that F-T diesel is ultra-low sulfur without requiring any in-house processing.

Pipeline grade gas has sulfur already scrubbed, making any diesel produced have a negligible sulfur content. On the other side of the equation, diesel fuel sulfur standards in the US have tightened noticeably in recent years, while the crude slate being processed in the US gets steadily more sour.

You can see how Sasol would love to exploit this premium. (In the interest of disclosure, I have to attribute this insight to a recent edition of Chemical Engineering Magazine).

Pearl GTL in Qatar is also pushing its position to arbitrage cheap natural gas with expensive liquid fuels.

Despite the potential commercial advantages, though, I can't help but find myself leery of GTL processes. Notwithstanding the rather asinine comparisons one tends to see (Forbes is particularly guilty), I can't stomach GTL any more than I can stomach corn ethanol because of that energy balance. I've taken to agreeing with my professors from way back in undergrad, who put it succinctly: "Fischer-Tropsch is a sign of desperation. If you're doing it at a large scale, things are going very wrong."

Geoffrey Styles said...

Thanks for your thoughtful and insightful comment. Not only is GTL diesel extremely low in sulfur, it also has a very high cetane--one source indicates 70-75, compared to 45 cetane index for typical US diesel fuel. However, it does present some challenges in cold climates, since its high paraffin content results in high cloud and pour points. I.e., it could turn to gel on a cold day in a northern climate. Taken together, these properties indicate its best value is probably as a blend stock to improve the quality of other diesel streams.

As for the energy balance, from an engineering perspective I understand your concern. However, I'm glad you mentioned ethanol, which after all is effectively another GTL mechanism, leveraged with a bit of sunlight. (Nitrogen fertilizer used to grow corn is derived from gas, and gas is the main source of process energy for many ethanol plants.) So one process effectively turns natural gas into a low-BTU liquid fuel that is only partially compatible with existing engines and incompatible with pipelines, which constitute the heart of our fuel distribution system. The other turns gas into a high-BTU liquid fuel fully compatible with both existing vehicles and the entire fuel distribution network. These aren't the only issues, but from a fuel perspective GTL beats ethanol hands down.

And while it might be more thermally efficient to turn gas into CNG and LNG for use in vehicles built or modified to use them-- compression and liquefaction aren't a free ride, either-- the vehicle conversion costs and convenience trade-offs still deter many potential users, despite the large cost advantage of gas per gasoline gallon equivalent . So the barriers to using the more energy-efficient pathway look much bigger than the barriers to GTL. If we were starting with a clean slate, I think CNG and LNG would have the edge, but we're not. Policy could certainly tip the balance, but I suspect policy makers are much more focused on other aspects of energy to give this much attention. The "NATGAS bill" has been floating around the Congress for several years, without much success: http://www.hartfordbusiness.com/apps/pbcs.dll/article?AID=/20120314/NEWS01/303149980.

Alice Finkel said...

GTL can be powered by heat from high temperature gas cooled nuclear reactors, which can shift the energy balance significantly in a favorable direction.

This is what the Chinese are looking to do, a decade or so down the road. Unless the Obama administration has a Damascus moment on new nuclear development, the Chinese will be the leaders in HTGR development, production, and applications.

The Chinese and Russians will still have a lot of competitively priced natural gas by then, and probably Canada. The US? Wait and see.

Anonymous said...

"That comparison is more obvious when one converts natural gas prices into their energy equivalent in crude oil." I am wondering how you did this conversion?


jtf said...

I was under the impression that it was possible to refine out most of the F-T waxes from diesel streams. If the cold flow properties of GTL diesel are as you say, then you're absolutely right about its value proposition.

One quibble with what you mentioned on corn ethanol: despite a large amount of NG-derived inputs into corn farming processes, my most recent look at USDA cost of production models suggested up to a third of the fossil energy inputs come from diesel fuel. But yes, we can certainly agree that corn ethanol doesn't have much going for it in the end-use department.

An interesting aside that you might find amusing is Celanese's GTL process that produces mixed alcohols, predominantly ethanol. You and I probably both know that Celanese is primarily a chemicals company and in all likelihood would focus on marketing highly pure chemicals-grade alcohols, but that didn't stop a few technically illiterate writers from advocating relaxing the RFS to allow GTL-based ethanol as well. Suddenly we go from a trade-off between a process that is slightly better in terms of energy balance but produces a worse fuel (corn ethanol) and a process that produces a superb fuel with a slightly worse energy balance (F-T), to a process that is the worst of both worlds. Hard to believe people will lap that up.

Geoffrey Styles said...

Nuclear energy offers synergies with various hydrocarbon processes, including oil sands extraction. I wasn't aware of the GTL angle; thanks.

Geoffrey Styles said...

Convert gas to barrels of oil equivalent (BOE) by dividing gas BTUs by 5.8 million BTU/BOE.

Geoffrey Styles said...

Sorry for any confusion re cold flow. I wasn't referring to waxes but to paraffin in the general sense of alkanes.

I'd be fascinated to see an energy balance for ethanol that attributes 1/3 to diesel, which as far as I know has 3 roles in the ethanol value chain: corn cultivation, hauling of harvested grain, and transportation of ethanol product (by rail or truck.) Compared to those, the energy loads associated with distillation/separation and fertilizer production should be much larger per unit of input or output.

As for the Celanese process, I discussed it back in 2011: http://energyoutlook.blogspot.com/2011/04/industrial-scale-ethanol.html At the time, I didn't have access to an energy balance for the process. From your comment, it sounds like you do. On a qualitative basis, avoiding the need to separate the ethanol from water looked like a big energy win for gas-to-ethanol. Isn't that the case?

jtf said...

I lost the model in question in an inbox clearing, but I pinged some of my colleagues about it, so hopefully there will be a link soon. There is of course always the possibility that my memory is faulty!

Truthfully I do not have a full energy balance on the Celanese process, but I put in some thought on it and a cursory order-of-magntiude analysis and I believe the energy return more closely resembles F-T GTL than it resembles corn ethanol*.

Celanese's process trivially falls below the -10% threshold because of the heat required for the partial oxidation to syngas. The reforming to methanol is exothermic but it doesn't offset the heat requirements of methanol carbonylation. There is also the need to provide hydrogen for the hydrodgenation of the resulting acetic acid to ethanol, which of course costs more natural gas, with the hydrogen required in a 1:1 stoichiometric ratio per molecule of ethanol produced. All of this leads me to believe the energy balance will more closely resemble F-T than corn ethanol.

Regarding your point about separation, I can see where it might be easy to overlook but there's still significant amounts of energy that need to be expended in separating from water. Celanese's methanol carbonylation technology is essentially a slightly improved version of the well-known Cativa and Monsanto acetic acid processes. All of these do not have single-pass efficiencies necessary to justify burning the remaining syngas for heating value, so all operate with water scrubbing systems and corresponding stripper columns. There's nothing I've found in the patents for Celanese's process to suggest that they will not first separate the acetic acid from their carbonylation reactor effluent using at least an absorber - after all, keeping the syngas around would both create useless side products from valuable syngas and consume valuable hydrogen. Depending on the economics, you could see an intermediate separation, but more likely it will just be a hydrogenation feed stream in the vapor phase with significant water.

Secondly, while the hydrogenation can be achieved in a single pass the resulting vapor-phase ethanol still needs to be separated from excess hydrogen, and all of Celanese's patents indicate that once again they will be doing this with a water absorption column followed by thermal stripping, much like the separations in the obsolete direct ethylene hydration process. If the absorber-stripper system following acetic acid hydrogenation is considered by itself, it will consume less energy than a corn ethanol separation train just by virtue of having a higher titer than raw beer can hope to achieve, but that energy expenditure is not going to be negligible. The patents are characteristically vague, but the range for separations they give is an absorber bottoms containing 24-76% ethanol, which of course must go through both a distillation and molecular sieve to output as anhydrous fuel blendstock.

* If memory serves, the energy return range for ethanol in the literature is about -10% to +30%.

jtf said...

Er, sorry, quick correction, I believe the hydrogen required for acetic acid hydrogenation will be required in 1.5:1 stoichiometric ratio, not 1:1.

1.5 H_2 + CH_3COOH --> H2O + CH_3CH2OH

Geoffrey Styles said...

You raise some excellent points. Ultimately, the US has two distinct but overlapping issues here: how best to use the growing natural gas surplus and how best to supplement liquid fuel supplies. These will be resolved by some combination of market forces and government policy. We can only hope that the latter is informed by the kind of discussion we've been having here about efficiencies and product compatability, rather than the political considerations that gave us our current, obsolete ethanol policy.

jtf said...

I managed to dig up the model in question we were discussing earlier, by the way. Mea culpa - the model was only for corn growing, and diesel fuel constituted 1/3 of the energy-related costs, not raw energy, accounting for the requisite NPK fertilizers and pesticides as fossil-derived.

Anonymous said...

thanks for share...