Coskata Refutes Energy Analyst’s Critique, Says It’s On Track to Make Ethanol for Under $1 Per Gallon

The idea that a strain of bacteria discovered at the bottom of a lagoon on the campus of Oklahoma State University could hold an answer to U.S. dependence on foreign oil sounds improbable to many. It certainly does to Robert Rapier. The chemical engineer at Arnhem, Netherlands-based Accsys Technologies published a blog post last weekend arguing that Coskata—the Warrenville, IL, startup that hopes to turn the bacteria into the engines of an industrial process that converts waste gases into millions of gallons of ethanol per year—has wildly exaggerated its claims that the process will eventually yield biofuels at a cost under $1 per gallon.

“Not only is this not the ‘slam dunk’ that is being projected, you probably have a better chance of hitting a blindfolded shot from mid-court than Coskata has of producing cost-competitive ethanol,” Rapier wrote. His post attracted the attention of other blogs that cover the energy sector, including Venture Beat.

But Coskata, which is funded in part by Waltham, MA-based Advanced Technology Ventures, is taking issue with Rapier’s analysis, saying that it’s based on faulty assumptions about the cost of building and operating demonstration ethanol plants versus commercial-scale plants. In a comment responding to Rapier’s post and in a separate conversation with Xconomy yesterday, Coskata vice president and chief marketing officer Wes Bolsen says the company sees no barriers to using its technology, which couples gasification of municipal waste and other hydrocarbon-rich materials with bacterial fermentation of the resulting gas, to produce ethanol for $1 per gallon or less, not counting the cost of the manufacturing facilities themselves. “Yes, gasification plants are expensive to put on the ground,” says Bolsen. “But the production cost is half that of gasoline, and you also have more feedstock flexibility—you can gasify trash, tires, agricultural residues, wood waste, any kind of biomass.”

Rapier’s critique centered around the cost of the demonstration plant that Coskata is building in Madison, PA, on the site of an existing gasification reactor owned by Westingthouse Plasma, a subsidiary of Canadian firm Alter Nrg Corp. The pilot plant, which is expected to begin production next spring, will cost $25 million and is expected to produce 40,000 gallons of ethanol per year. Rapier did some basic math, calculating that the plant will produce 2.6 barrels of ethanol per day, which puts the plant’s capital costs at $9.6 million per daily barrel. On the surface, that compares unfavorably to the capital costs of other fuel production technologies such as oil refineries ($10,000 to $20,000 per daily barrel) or even corn-to-ethanol plants ($20,000 to $30,000 per daily barrel). “The capital costs would have to go down by a factor of 100 before they could even start to get competitive,” Rapier commented.

Bolsen concedes that Coskata’s pilot plant will lack economies of scale. “It is very expensive to do commercial demonstrations; any facility producing under 1 to 2 million gallons per year is going to be unbelievably inefficient,” Bolsen told me by phone today. But it’s misleading to evaluate the economics of Coskata’s technology based solely on the cost of the demonstration plant, he argues. “You can never do your economics off of a demonstration,” he says.

Part of the Pennsylvania plant’s high cost, Bolsen says, is due to the fact that it’s an R&D facility, with two separate ethanol reactors—a membrane separation assembly and a traditional distillation assembly—being built side-by-side. He says Coskata is planning a true commercial-scale facility, with construction due to begin next year, that will cost $400 million and will produce 100 million gallons of ethanol per year. That translates into about $61,000 per daily barrel, or only about twice the capital cost of a corn-to-ethanol plant.

But 100 million gallons per year is a lot of ethanol—and the real question from an outsider’s point of view may be whether Coskata’s anaerobic bacterial fermentation process will work as well on an industrial scale as executives say it’s working in the laboratory.

The company’s technology is different from most other ethanol production techniques, in that it uses synthesis gas or “syngas” as its input rather than corn or high-cellulose corn alternatives such as switchgrass. Most so-called “cellulosic ethanol” technologies, such as the one being developed by Cambridge, MA-based Mascoma (which, like Coskata, has collected a big investment from General Motors), use various enzymes to break down the cellulose molecules in feedstock into sugars that can then be fermented into ethanol. In Coskata’s system, feedstocks such as municipal waste are simply vaporized inside … Next Page »

Single PageCurrently on Page: 1 2

Wade Roush is the producer and host of the podcast Soonish and a contributing editor at Xconomy. Follow @soonishpodcast

Trending on Xconomy

By posting a comment, you agree to our terms and conditions.

  • Curt Fischer

    The quality of Robert Rapier’s blog post is stunningly bad. Of course, this doesn’t mean his conclusions are wrong, just that they’re unsupported.

    First: A well-known approximation to chemical engineers is that costs scale as the 0.6 power of the capacity.

    In equation terms, where C is cost and P is capacity: (C_big / C_pilot) = (P_big / P_pilot)^0.6.

    If we accept the 0.6 as gospel for a second, that would give a extrapolated cost for a 150 MMgal / yr plant of $356 per daily barrel, not $9.6 million. That’s still ~2x higher than EIA Annual Energy Outlook 2006’s estimate for biomass-to-liquids capital costs. I’m confused why R. Rapier would quote both the $9.6 million number and the EIA number in the SAME POST, without noting the huge inconsistency.

    Second, the correlation is known to be inaccurate when making giant extrapolations. Our extrapolation here is indeed gigantic, by at least a factor of 3000 or so. A lot will depend on what the right value for the exponent turns out to be. For example, changing it from 0.6 to 0.55 reduces the capital cost by a third. Raising it to 0.65 makes the costs surge. But whatever the exponent might be, we know for sure that it won’t even be close to 1.0, which is what Rapier assumed.

    Third: Rapier says: One final thing I would point out is that the selectivity of these processes generally favors methanol over ethanol. They may be able to push the selectivity toward ethanol, but they will almost certainly end up producing mixed alcohols that will have to undergo purification to ethanol.

    This is ridiculous, as the selectivity for ethanol relative to methanol of the bacteria Coskata plans to use is virtually infinite. Bacterial metabolism is not the same as Fischer-Tropsch (no relation to me).

  • I always find it extremely ironic when someone is critical of a piece of work, and then completely botches the math when trying to demonstrate it. Check your numbers, Curt. $356 per daily barrel? And then you note that you are confused?

    Further, I have to presume the Curt didn’t actually read the post, since he is confused about a quote and states that I didn’t note a huge inconsistency. Yet that post states “Caveats are that the graphic is for full-sized plants, and the numbers are a couple of years old.” So the very issue that Curt is complaining about was addressed in the post. Curt is taking exception to something that he has misinterpreted in one case and overlooked in another.

    The issue in focus in the original post was how far out of the stratosphere the numbers were for their pilot plant – and I gave a specific example of another pilot plant (GTL). When comparing pilot to pilot, the numbers were $9.6 million to $188,000 – and the $188,000 plant didn’t go to commercialization because it was too expensive. It is known that the $9.6 million will fall as the process scales (Curt seems to believe this would be a revelation to me, despite the fact that I have actually been closely involved in the scale up of facilities), but the point was that they wouldn’t fall enough (based on the $25 million which Coskata now says includes things other than the plant) to be competitive.

    So, I am left to conclude that Curt thought the quality was bad because of his misinterpretations of what I ‘really meant’ instead of what I actually wrote.

    Cheers, RR

  • Curt Fischer


    Thanks for the response. First, thanks for catching my typo: instead of $356 per daily barrel, I meant $356k per daily barrel. The discrepancy between this number and the chart on your blog post, like I mentioned, was about twofold.

    Second, I fail to see how your caveat that the “graphic is for full-sized plants” explains away my objection on your use of the $9.6 million number.

    Third, your comparison of “pilot to pilot” is still not a fair comparison. The GTL pilot plant you mentioned was 400 bbl/d of capacity. The Coskata pilot is only 2.6 bbl/d. See the equation in my earlier comment? If we use that to scale down the GTL pilot to 2.6 bbl/d in size, the cost per daily barrel of the hypothetical super-small GTL plant would be $1.4 million. That’s still much lower than Coskata’s pilot, but not nearly as low as your comparison would (mis)lead one to believe.

    You seem to reject the idea that plant costs do not scale directly with capacity. Although I recognize that the 0.6 value for the scaling exponent is not universal, it is surely closer to the truth than the 1.0 you implicitly assume.

    Fourth, you have not responded to my criticism of your claims about selectivity vs. methanol.

  • Curt, that’s just the thing. I stated several times in my post “of course this is just pilot scale.” What do you think I mean by that? So when you suggest that I seem to reject the scale issue, or that I am “implicitly assuming” something – and that was the gist of your initial post – you have skipped over multiple instances of me indicating that I did not expect costs to scale directly with size. The $9.6 million number was not significant because I expected a full-sized plant to come in at that number. It was significant because it was so far away – even with scaling (and your own numbers above show this) from competing with corn ethanol, oil refining, GTL, or even CTL that it is a nonstarter. Thus, the ‘Dead Man Walking.’ Now, Coskata has since indicated that there is much more in that estimate than just a pilot plant. So maybe that’s a valid answer. It’s really hard to say when someone has only done lab-scale experiments.

    If you have kept up with the conversation – which has continued following the initial post – you will see exactly what my issues are. First, you can’t make any sort of reasonable plant estimate based on lab scale experiments. You are probably aware of this. You can’t reasonably extrapolate mass and energy balances from a few hundred gallons a year in a lab (their current pilot scale) to a hundred million gallons a year.

    Regarding the selectivity issues, most (alcohol) gasification processes do result in mixed alcohols (or at least have significant by-products). Coskata indicated to me – and I reported it in my followup – that their microbes produce no byproduct alcohols. This was your only legitimate criticism, but I had already reported that in my follow-up. This will simplify the purification step, but they also indicated that the alcohol coming off the fermenter is at 3 or 4% ethanol. If you have access to HYSYS, run a separation at that level of ethanol and see how much net energy is lost due to the BTUs required for removing water.

    To conclude, I actually hope they deliver on their claims. For that matter, I would help them in any way that I could. I have long been on the record that I think gasification is the future. I just see some areas that warrant a good investigation.


  • Pingback: Cutting Through the Coskata Cellulosic Ethanol Hype | Commodities Options | Commodities Futures | Commodities Prices()

  • Pingback: Cutting Through the Coskata Cellulosic Ethanol Hype | The Oil Report()

  • Sven Arunogen

    Wish it was true. Just another scam. Next.