Coskata Due Diligence

I hadn’t intended to spend the entire week talking about Coskata, but that’s the way things are shaping up. After Monday’s essay in which I expressed strong skepticism that Coskata could deliver, I received a number of e-mails and phone calls, and the essay received quite a few visitors. The story was picked up and reported via several outlets. People are certainly interested in the Coskata story, and they want to know whether the emperor has clothes. I had stated that based on my analysis of some of the published numbers, the emperor has no clothes, and this raised a number of eyebrows.

Some of the raised eyebrows were within Coskata. Following that essay I was contacted by Wes Bolsen, the Chief Marketing Officer & Vice President of Coskata. I give credit to Wes for stepping up and offering to answer questions in order to blunt my skepticism. After all, if my skepticism is unwarranted, then I may have brought undue suspicion on the company. I am always interested in getting the straight story, so I spent an hour and a half on the phone with Wes – and we dug deeply into the details. My previous post listed the questions I had in hand as we began our discussion. I won’t go line by line, but instead summarize the conversation from my perspective – and indicate a couple of areas in which I need to follow up.

Let me state for the record that I want Coskata, and for that matter Choren, LS9, Range Fuels, Virent, Nanosolar, and many more like them to be successful. The world needs solutions to our energy problems, and I applaud all of these companies for their efforts. But I never let what I want cloud my judgment when I am trying to determine what is true. In other words, even though I want Coskata to be successful, I still believe in scrutinizing their claims very closely, and stating for the record whether I believe their claims to be credible. And if they look to be the company best-suited for turning waste biomass into fuel, I will advise my contacts – some of whom are sitting on mountains of biomass – that this looks like an outlet they should explore.

Summary

Coskata’s intellectual property is focused on several areas. First, they have licensed five sets of patented organisms (bacteria). These are not Coskata patents, but are exclusively licensed to Coskata. Wes didn’t want to reveal the particular patents that are licensed, but they were developed at the University of Oklahoma and at Oklahoma State University. The nature of the organisms is that they consume syngas (carbon monoxide and hydrogen) which is the product of gasification of biomass (or coal, oil, or natural gas) and produce ethanol.

On the purification step, instead of a traditional distillation, they intend to use membrane separation. The same deal applies here as with the microbes; there is an exclusive license to Coskata, but Coskata does not own the patents.

While no patents have been granted to date to Coskata (understandable, since the company is only a couple of years old), they have multiple patents pending (as many as 18) around the bioreactors. They also have some pending around the microbes. While I have not yet done so, it should be easy to dig up their patent applications, which will provide additional details on the specific microbes in question.

I asked a lot of questions around the gasifier. I asked why they went with a relatively new technology like plasma gasification instead of one of the gasification technologies that has a long track record. Wes said that 1). Plasma gasification is much cheaper; and 2). The low operating pressures they use in their reactors (1 to 2 bar) are better suited for a low-pressure gasification technology like plasma gasification. I know that plasma gasification is used for waste disposal, but I asked if anyone is using it for the primary purpose of producing a fuel. The answer was no, they intend to be the first. He added that you wouldn’t use plasma gasification for a GTL or CTL plant because compression costs to go from low pressure to the pressure that FT runs at would be prohibitive. He also volunteered that an air separation unit (that’s a big ticket item) was required to provide high-purity oxygen to the gasifier. I asked if it took a lot of energy to maintain the plasma, and he said only 5% of the energy inputs into the gasification went into maintaining the plasma.

On the reactor, the gas makes a single pass through. There is no gas recycle, and the tail gas is used to provide energy for the separation. They use 1 gallon of water to make 1 gallon of ethanol, which would be a significant improvement over grain ethanol. Wes stated that the bacteria only produce ethanol; no byproducts. This would be a major advantage in the purification process. A very important point that Wes told me here is that the crude ethanol product is 3 to 4% ethanol in water. This got my attention, and I will explain why later.

I asked if their intent was to build plants or license the technology, and he said both. They would like to build 1 plant for every 8 licenses they grant.

I asked for a breakdown of the claim of $1/gallon production costs. Wes said everything except capital recovery is included in that cost. He said that almost half is feedstock costs. They have assumed $50 per dry ton of woody biomass, and a yield of (more than) 100 gallons per dry ton. Of the other 50 cents, 25 cents is maintenance and operating costs (including the gasifier), 25 cents is utilities and other. But if you charge the biomass costs to the gasification, the syngas cost is well more than half of that $1.00.

So, that suggests a sanity check that I have not yet had time to do. If you convert a ton of woody biomass into syngas under these assumptions, how do the syngas costs compare to syngas that you could purchase commercially? If it is much cheaper, then there are other implications, since syngas is in great demand for lots of commodity chemicals. If it is much more expensive, then that suggests that it may not be long before it becomes attractive to switch to natural gas as the gasification feed. But the gasification piece is not their focus. For all practical purposes, they could simply have a pipeline into their plant in which they were bringing syngas. The reactor and downstream technology is their focus.

However, I told Wes that I think the $1/gal claim is misleading, as most people I have talked to were under the impression that this is the total cost. It is not. Capital must paid for, and investors expect a return on capital. The total production cost is going to be very sensitive to overall capital costs, and that is a potential issue from my perspective (more on that below).

I asked about the scale of their lab facilities, as this is a very key issue. Coskata has a small pilot plant (what I would call lab-scale) outside of Chicago. They have an autothermal reformer (ATR) on the front end so they are producing their own syngas at the piloting facility. They are not dealing with cleanup processes of the syngas at this scale. After the ATR, the syngas feeds the fermenters which then go to a conventional distillation train. The operation runs 24/7, and the scale is gallons per day. I asked what they were doing with the ethanol product from that facility, but Wes wasn’t sure.

I have been involved in several projects that went from the lab to pilot scale and up to commercial scale. One thing I can say from my experience: You can’t design a plant based on lab scale experiments. Why? Because all kinds of issues that were minor at the lab scale can become problematic at pilot scale and major headaches at commercial scale. This is why you scale up from the lab to a pilot plant to a commercial plant: To get a better handle on these issues, and to get a better estimate of costs. Many projects die at the piloting scale when capital costs start to pile up. And this is one issue that I would flag with Coskata: They have only run small lab scale tests – which have neither the plasma gasification front end nor the membrane separation back end – and that is the basis for an estimate of $400 million to build a 100 million gallon a year plant. I think as they learn things at the pilot scale, that estimate is at risk to grow. At least that’s always been my experience at the pilot scale. (In fact, here is an article from four months ago in which the quoted cost was $300 million).

Now for the stickiest issue of all for me: The energy balance. I asked about this, and I felt like Wes repeated something he has said a thousand times before: “The Coskata process yields up to a 7.7 net energy balance.” I asked “Did you get that from Michael Wang at Argonne?” He told me that yes he did, and I went on a 5-minute digression about my issues with the way his work has been (mis)used.

But I have enough information to determine whether that can possibly be true. Intuitively, I don’t think it can. Here’s why. If you start with a ton of woody biomass, and end up with 100+ gallons of ethanol – and have to purify from a 3.5% solution of ethanol in water – you can make a pretty good estimate of the energy balance. I haven’t had time to do this yet, but I will. However, a 3% alcohol solution is close to the region at which you would classify the stream as a waste stream and send it to wastewater treatment if you were in a chemical plant (and in fact, when I worked in a butanol plant we did send 3% butanol to the wastewater ponds). Why? Because the energy you could gain by removing all of that water is negated by all the energy you have to put into removing the water. Since Wes said that the membrane technology uses half the energy of distillation, I can make a pretty good stab at seeing whether the energy balance adds up.

I told Wes that I felt like there was a disconnect between a 3.5% alcohol solution and an energy balance of 7.7. This energy balance is in the region of sugarcane ethanol, and they are producing an ethanol solution in the 8-9% range (which takes much less energy to purify than a 3.5% solution). My guess is that this is the ‘fossil fuel’ balance and not a true energy balance. To put into perspective the difference, if you had 100 BTUs of biomass, and only ended up with 1 BTU of usable fuel, you would probably view that as a great waste of BTUs. But if you had to input 0.1 BTUs of fossil fuel into the process, you could claim that your (fossil) energy return was (1/0.1) or 10/1. In summary, it can be a misleading metric.

In response, Wes said that they were using waste heat to drive the separation. I said that I understood this, but that waste heat still came from the original BTUs in the wood, and I needed to close the energy balance to see if everything added up. To be completely fair, I have not yet had time to do this, so right now I only have a question mark over the energy balance. A ton of woody biomass will have an energy value of maybe 13 million BTUs, and 100 gallons of ethanol contain 7.6 million BTUs. Is the other 5.4 million BTUs enough to separate out the water (2857 gallons of water for a 3.5% ethanol solution)? I have to calculate that.

Conclusions

So, do I still think Coskata is a ‘dead man walking’? Let me say that I still see question marks. There are some good and novel things about their process, and I learned quite a lot from talking with Wes. He does know his material, and was very good at answering almost every question I threw at him (and I threw a lot at him). He said they have an estimate to build a 100 million gallon a year plant for $400 million. That would put them at a disadvantage to the capital costs for a corn ethanol plant, an oil refinery, or even GTL, but in much better shape than most capital estimates I have seen for biomass gasification.

I have question marks around several areas. One is the capital estimate to build the 100 million gallon plant. I don’t think you can make a very accurate estimate from only lab scale experiments, and they won’t have a well-defined capital estimate until they have some experience running their larger pilot plant. This 40,000 gallon per year plant will be built in Madison, Pennsylvania and will be operational in 2009. It may be early 2010 before they have a capital estimate that you can be really confident has covered all the potential snags.

The second major question mark I have is around the energy balance. The information I have is that one ton of dry woody biomass produces 100+ gallons of ethanol. The ethanol must be purified from a solution that is 96.5% water. The claimed energy balance is 7.7. That doesn’t sound right to me, but until I do the energy balance (and I will try to do it tomorrow if I have time) I will just flag it as something to look into.

The other things to note are that they are using several pieces of novel technology. Syngas is a very common raw material in the chemical industry, but it isn’t often produced via plasma gasification. Likewise, distillation is a core process in the refining and chemical industries. Yet next to nobody is using membrane separation for their purification needs. If Coskata is correct that membrane separation for ethanol requires half the energy of distillation, I have to wonder why ethanol producers everywhere aren’t flocking to exchange their distillation trains for membrane units.

Finally, I want to thank Wes and Coskata for spending time to educate me on the process. I suspect it won’t be the last time I speak with Wes, and I do hope Coskata is successful (despite my questions). I will follow up once I have had time to work through the energy balance.

15 thoughts on “Coskata Due Diligence”

  1. Wood for biomass puts a big question mark in my mind. Specifically, what kinds of wood are they going to use? The reason I ask is that I heat in the winter with a wood stove, and use fuelwood from various tree species. I know from direct experience that the heating value per unit wood varies widely depending on the species. Are they taking this into consideration? Are they using only one tree species? Or does calculating BTU per ton make for an accurate measure of energy content even if different species are used?

  2. This seems like an ideal solution for the waste industry. I’ve got a friend that pays $40 a ton to dispose of trees he cuts down. It seems like it’d be a win-win for these companies if the cost of producing the ethanol was less or even equal to the cost of gasoline. They could do away with a lot of landfill space and fuel those big garbage trucks to boot. I think they’ll do very well in the licensing area,even if the costs are 4X higher than estimated.

  3. Their membrane-separation technology seems unproven in terms of process feasability and final cost. This is a cellulosic mandate/producer credit play.

    Wes will be out shopping his resume, looking for a new job in 2010.

  4. Not to interrupt the discussion, but I did want to take a moment and thank you for the Coskata reporting, Robert. I follow a number of energy-related sites and blogs, and this kind of detailed and considered analysis is fleetingly rare. I’m still digesting everything, but this is easily the most interesting thing you’ve done in ages. Kudos.

  5. Great post RR, and my congrats to Coskata for talking to you mano a mano.
    I hope Coskata works — but I also sense that in a broader sense, it doesn’t matter. Somebody is going to “crack the code” on biomass conversion sooner or later.
    As a society we are new to the commercialization of biomass fuels. But thanks to efforts of outfits like Coskata, we are getting closer and closer. Even failures contain important lessons.
    All hail the scientists, engineers and financiers of Coskata. It is people such as them — not doomer-snivelers on the sidelines — who will make life better for all of us.

  6. Not the last word in alcohol distillation, but really a pretty jam packed article:

    http://www.ces.purdue.edu/extmedia/AE/AE-117.html

    Comparing the energy balance involved with separating by distillation 3.5% wt/wt alcohol-water and a more optimal 9% concentration appears to be adding between 1.5 and 2 HETP’s to the column, and employing aggressive feedstock preheating with the efflux water stream through many stage heat exchangers.

    Membrane may be able to concentrate the primary stock from 3.5% to 10% more cost effectively in a single step: if so, this would mitigate much of the heat exchange costs and necessity.

    A 100,000,000 gallon/year plant will require 1,000,000 tons of feedstock, and will have to process 24/7 about 3,000 T per day of fermentation material. While this is a lot (about 40 railroad dry-bulk carriers), it isn’t unprecedented, and certainly not more than a mid-sized coal-fired power plant, so that reality check is OK.

    On the other hand, plasma reforming of 3,000 T of organics to syngas … is quite the technological challenge. I would sure like to know more about its yields and byproducts. Working with cellosic and hemicellosic starting materials is helpful: the oxygen is already incorporated, so the conversion of HCOH units to CO + H2 has a lot lower endothermic energy balance than working with either natural gas or coal and steam.

    No matter what, the plasma (or conventional) plant will have a lot of waste heat – ideal for powering the distillation side, if even but in a ‘preheater’ sense.

    Finally, a couple of aspects of this that continue to sound somewhat disingenuous: NO biological process produces a single byproduct of metabolism. To anyone familiar with making beer, a teaspoon of yeast (“start”) becomes several cups of sludge (“end”). Brewers at larger scales deal with the same, and are the source of animal (and human) feed supplements by that route.

    I remain interested but share Robert’s skepticism regarding the scalability of this technology. There are a lot of conventional engineering problems to be resolved, each of which consumes energy. Will the accumulated load of ameliorative process energy inputs exceed the reasonable product energy potential? Or, to put it differently, will it do better than 2:1, generally considered the ‘well let’s do it’ standard for alternative energy recovery?

  7. OT but…Aug. 21 (Bloomberg) — Crude oil speculators account for about 81 percent of all contracts traded on the New York Mercantile Exchange, the Washington Post said, citing Commodity Futures Trading Commission data.

    The share of contracts held by financial firms speculating for their clients or for themselves is “far greater” than the U.S. regulator had previously stated, the newspaper said. The share may rise in coming weeks, according to the paper.

    At this point, speculators ARE the market. It is senseless to debate whether they “influence” prices. The market price is the speculative price.

  8. Robert,

    Did you get a chance to ask how much parasitic load the ASU’s have on the process? I thought those were typically very energy intensive. The use of an ASU with a relatively low BTU feedstock seems to me counter-intuitive.

    If plasma torches require only 5% of energy input, why wouldn’t they be used in more syngas applications? 5% seems a very low figure – do they account for the generation and transmission losses for fossil fuel to electricity?

  9. Why not just burn the biomass for electricity and use the saved natural gas in vehicles? Both boilers and vehicles can be converted with existing tech.

  10. “Why not just burn the biomass for electricity and use the saved natural gas in vehicles?”

    If this technology works half as well as they think it will,it’ll make a heck of a lot more money than electricity would. They’re assuming $1 a gallon cost,with feedstock costs running half of that,right? But,this is supposed to work with things like used tires,and people pay to get rid of those. If they could make $3 to dispose of the tire,and another $3 for the ethanol…for a cost of 50 cents,these guy will be rolling in dough in no time.

  11. Maury, assume we get cap and trade for CO2 reduction. Then an idled coal plant is a sunk cost, as is the trained operating staff and the grid connection. The investment to retrofit the boiler has returns in months, not years, and requires no speculative technology.

  12. maybe membrane is here feasible due to the much lower biomass yield… compared to normal fermentation processes.

    mtf

  13. According to the USDA Forest Products Lab green wood, 50% moisture, contains 5.74 btu/ton and around here sell for about $40/ton. That drives the raw material cost up significantly.

  14. Great job, RR! I always think the ability of an engineer (or any technical person) to explain a technical topic to the non-technically inclined is a sign of intelligence. This essay (and some of your previous work) shows great intelligence, together with an unusual ability to distill the essentials. Keep it up!

    On to the topic at hand:
    Statement #1: The nature of the organisms is that they consume syngas (carbon monoxide and hydrogen) which is the product of gasification of biomass (or coal, oil, or natural gas) and produce ethanol.

    Statement #2: Wes stated that the bacteria only produce ethanol; no byproducts.
    As mentioned by Bob Lynch, statement #2 on itself is very suspicious. Mother nature just doesn't do things that way.

    Let's translate statement #1 to chemistry. The claim is:
    CO + H2 -> C2H5O (A)

    As anybody who passed high school chemistry can atest, equation A is not a balanced reaction: There is only one carbon atom on the lefthand side, but two on the righthand side. The hydrogen is not balanced either.

    This is where statement #2 comes under suspision: the syngas has a C:O ratio of 1:1. The ethanol has a C:O ratio of 2:1. Where is the excess oxygen going to go? Most likely to a waste product.

    Let's try this:
    4CO + 5H2 -> 2C2H5O + O2 (B)
    That would be great, as it would produce (some of) the oxygen required for the gasification step. Unfortunately, reaction B, has a water-flowing-uphill element to it: thermodynamically you can't produce oxygen that way.

    It is far more likely that the excess oxygen would leave the system as water vapor (H2O) [or worse CO2, but that is less likely], diverting more of the syngas to the waste product which statement #2 denies existing:
    4CO + 7H2 -> 2C2H5O + 2H2O (C)
    Or perhaps, but unlikely:
    6CO + 5H2 -> 2C2H5O + 2CO2 (D)

    Either way, I believe that leaves statement #2, sufficiency trashed. RR, does any of this change your conclusions?

  15. As mentioned by Bob Lynch, statement #2 on itself is very suspicious. Mother nature just doesn’t do things that way.

    It’s a given that they have gaseous by-products. But without knowing the particulars of the microbes involved, it is hard to say what specifically they are.

    That’s interesting from an efficiency point of view. I was interested in liquid by-products because that would tell me how complex the distillation system would need to be. That’s the context of his comment on “no by-products.”

    RR, does any of this change your conclusions?

    It is an important point, but not for the reason I asked the question. It is important because it defines the theoretical efficiency.

    RR

Comments are closed.