Previously, I described a portion of my role in the early development of the MixAlco Process. Developed in the laboratories of Professor Mark Holtzapple at Texas A&M University, the process has undergone significant further developments, which I report on in this essay.
Details of the MixAlco Process
Here I will describe the process in a nutshell, but Wikipedia describes the process in significant detail. In fact, the details there are so thorough I suspect it was written at least in part by Professor Holtzapple’s graduate students.
The MixAlco process utilizes naturally occurring microbes to convert cellulose into chemical intermediates and fuels. The focus of the early work was to identify organisms that utilize cellulose as an energy source, and then to select microbes that are hardy and efficient for use in industrial fermentation. Unlike a conventional cellulosic ethanol process, the MixAlco Process does not add enzymes to break down cellulose. Cattle, for instance, host a mixture of microbes that produce their own enzymes that break down cellulose and convert it into primarily acetic acid. Think of a cow’s stomach as an industrial reactor, and you start to get the picture of what the process is attempting to accomplish.
From an economic point of view, the process has several advantages over conventional cellulosic ethanol. First, as noted, the process requires no enzyme addition. Second, because a mixed culture of hardy microbes is used, the process does not require sterile conditions. As you might imagine, the rumen digestive system is not a sterile environment. Finally, the process ultimately yields mixed alcohols or hydrocarbons, which have a higher energy content than does just ethanol.
The Present Status of the MixAlco Process
A company called Terrabon was formed in 1995 to commercialize the MixAlco process (and two other technologies). I graduated that same year, and plans were being formulated to construct a pilot plant based on the process. Construction of the pilot plant, designed to process up to 200 pounds of biomass per day, started in 2000. Because funding was scarce in those early days, it took several years to complete the pilot plant, but it has been in continuous operation since 2006.
Construction of a larger demonstration plant began in 2008, and was started up less than one year later. The demonstration plant is designed to process up to 10 tons of biomass per day, and has been in continuous operation since early 2009.
In 2009, the company attracted two large strategic investors. Valero invested in Terrabon in April 2009, and Waste Management followed in August. These investments bring together the technology in Terrabon, a refining and distribution capability with Valero, and lots of municipal solid waste to feed the reactors from Waste Management.
The Hard Questions
To evaluate this process objectively — admittedly more of a challenge than normal because of my history with the process and with Professor Holtzapple — I have to ask tough questions around the energy balance, conversion efficiency, capital costs, etc. If I look at the process completely objectively, it has characteristics that I do like, but also some that I don’t. One of the things I don’t like is the large quantities of water in fermentation processes that must be dealt with. That isn’t a show-stopper; after all Brazilian ethanol is a fermentation process that contains a lot of water. But they remove that water by using very low-cost biomass as fuel. Still, if the water wasn’t there in the first place, the energy efficiency could be a lot higher.
Professor Holtzapple sent me a number of presentations and papers (some linked below), which I have slowly worked through. These answered some questions — and triggered many more — so I have exchanged several e-mails with Professor Holtzapple in order to gain better insight into some areas. I present portions of the exchanges below.
Q&A with Professor Mark Holtzapple
RR: We were talking about a pilot plant in 1995, the year I graduated. What took so long to get it built?
MH: The pilot plant was started with seed funding of about $300,000. The funding was VERY welcome and resulted from initiatives taken by Jack Hopper at the Texas Hazardous Waste Research Center, which got its money from EPA. Unfortunately, this was not sufficient to build the entire pilot plant. We got some free equipment (e.g., steam-jacketed kettles) when Texaco shut down its research facility. Terrabon also provided major pieces of equipment over the years (e.g., vapor-compression dewatering system) and support structures. Recently, the pilot plant was upgraded for the DARPA project, the goal of which is to produce 100 L of jet fuel. We have added a centrifuge and driers, so the pilot plant is an evolving facility. It took a long time to build it because funding was hard to get, particularly in the early years when oil prices were low.
The demonstration plant was started in early 2008 and first became operational in early 2009…LESS than one year after construction began. To me, this is quite remarkable because it is quick for the first plant of any process. The detailed engineering was done by a company that designs waste water treatment plants and it was built by a general contractor that constructs buildings on campus and elsewhere. Much of the equipment was purchased from Grainger and McMaster-Carr. The reason I mention these points is that the design skills, construction skills, and equipment are commonly available, which makes scale-up easy. In contrast, biological processes that require sterility require specialized skills more commonly found in the pharmaceutical industry. They need sanitary valves, pumps, heat exchangers, and tanks, all of which are specialty items. Further, there are only a limited number of engineering and construction companies who have the skills…this makes scale up expensive and difficult.
The demonstration plant continues to evolve. Initially, it was only capable of doing pretreatment and fermentation. This summer, they are installing equipment that can do downstream processing. If you want more details, it is best to ask the Terrabon folks.
RR: Do you have a good idea of the energy returns of the process? I presume you will burn lignin (but probably aren’t in the pilot plant), but am curious about the fuel output/fossil fuel input ratio.
MH: The attached paper (Reference 2) shows the energy return can be as high as 18:1.
RR: OK a follow-up to that. I believe the 18:1 you mentioned is theoretical, but am really interested in what you have actually observed to date. My biggest concern is the amount of water that has to be removed from the process, and while I could envision scenarios that would yield a very high energy return, it would be at the expense of yield of final product (because I would have to burn much of the biomass for process heat).
MH: The main purpose of the demonstration plant is to prove that the fermentation works. Most engineers who have reviewed our process believe the downstream processing steps will work. Most of the uncertainty involves the fermentation, which is reasonable because no one has ever operated a mixed-acid fermentation for this purpose. As much as possible, Terrabon wants to use conventional, off-the-shelf technology for the downstream steps. The philosophy is to pursue one invention at a time. At the moment, the fermentation is the “invention” being pursued.
Currently, the demonstration plant does not employ our advanced vapor-compression desalination technology. Instead, Terrabon opted to use a less sophisticated, off-the-shelf technology that can be put in place rapidly.
Terrabon has built a large advanced vapor-compression distillation (AdVE) system for the City of Laredo, which be used to desalinate brackish water. It is still in the debugging stage, but will be shipped in late July.
We have put a LOT of effort into dewatering. This step can be very energy and capital intensive if not done properly. We have made some remarkable breakthroughs (in my opinion) that related to promoting dropwise condensation on the heat exchanger surface. In our laboratory apparatus, we have measured heat transfer coefficients as high as 42,500 Btu/(h ft2 F). For comparison purposes, a conventional heat exchanger at the same operating conditions would have a heat transfer coefficient of 3000 Btu/(h ft2 F). I am in the process of preparing the patent application, so the patent has not yet been filed. Unfortunately, I cannot give more details at this time.
RR: I generally view water as something I would rather exclude from my process, due to the energy required to remove it. In a conventional ethanol process (e.g., corn or sugarcane) a lot of the process energy is devoted to removal of water. There is also the issue that water resources are a problem in many areas. Can you comment?
MH: Regarding the water issue:
1. The fermentation water is recycled, so the process does not consume much water (other than for cooling towers).
2. If the feedstock is wet, water must be purged from the system. Because we distill water from the salt, we can purge distilled water from the system, which can be used for irrigation or industrial purposes.
3. All biological processes involve water in the fermentation. Critical issue are (1) concentration and (2) ease of separation. Brazilian ethanol tends to be pretty concentrated (~10%) whereas we tend to be more dilute (2 to 5%).
4. Although we must separate more water per unit of product, our separation process is easier. With ethanol, both water and ethanol are volatile, so multiple-stage distillation with reflux is required. Further, they have an azeotrope, which creates a complex separation step to reach purity. In contrast, we are separating a volatile component (water) from a nonvolatile component (salt). This allows the separation to occur in a single stage, rather than the multiple distillation stages needed with ethanol. In our process, because the temperature of the condensing steam and boiling salt water are very similar, it is amenable to heat pumping, which dramatically lowers the energy input. In contrast, with ethanol, the boiling and condensing occur at widely different temperatures, so heat pumps are not as effective.
RR: What have been some of the biggest challenges to overcome in the piloting?
MH: I think anyone who works with biomass will say that solids handling is a big challenge. Also, mixing is an issue. I believe these issues are tractable when scaling up.
We do NOT have any issues with maintaining sterility…all bugs are welcome to the table…may the best bugs win!! Our process is driven very close to the low-energy state (acetic acid) whereas sugar or ethanol are reactive intermediates that want to become acetic acid. The only other product of a lower energy state is methane + carbon dioxide, which are produced by methanogens. Fortunately, methanogens are ‘fragile” microorganisms… they are strict anaerobes so small amounts of oxygen wipe them out. Also, they do not grow well at low pH or at high ammonia concentrations. We put inhibitors into the system to kill the methanogens. We have had good luck with iodoform, but there are MANY other known inhibitors such as monensin, BES (bromoethanesulfonic acid), chloroform, etc.
RR: What is the yield of product from a ton of biomass? (e.g., gallons of alcohols, ketones, acids)
MH: The yield depends upon the route taken.
Using the acid route (see Reference 2), the reported yields are
Mixed alcohols = 141 gal/tonne = 127 gal/ton (from Table 1)
Hydrocarbons = 0.635 × mixed alcohol = 89 gal/tonne = 80 gal/ton from Page 551
Using the ketone route (see Paper submitted version)
Hydrocarbons = 81 gal/tonne = 73 gal/ton (from Table 1)
Mixed alcohols = 0.8 lb digested/lb fed × 0.65 lb acid/lb digested × 0.583 lb ketone/lb acid × 1.0225 lb alcohol/lb ketone × gal/6.6 lb × 2000 lb/ton = 94 gal/ton = 104 gal/tonne (from Table 1)
(Note 1: tonne = 1000 kg, ton = 2000 lb)
(Note 2: yields are quoted on an ash-free basis)
(Note 3: Yields are based upon laboratory studies. They must be confirmed at industrial scale.)
RR: Related to the above, have you used the leftover lignin in a biomass boiler? I have glanced through various information on the demo plant, but haven’t see whether this piece in integrated.
MH: The boilers in the demonstration plant are gas- or propane-fired. This reduces capital and hassle. In my opinion, biomass boilers only make sense at large scale.
RR: What are you doing with the product coming out of the pilot/demo plants?
MH: The product from the early runs has been stored for later processing. The stored product and the currently produced product will be converted into ketones using newly installed downstream equipment.
As always, readers are strongly encouraged to ask their own critical questions before forming strong opinions on a particular technology. As I have stressed many times, there are no silver bullets in the world of energy. Every energy production process comes with baggage of one sort or another. The key to successfully transitioning away from fossil fuels in the most sustainable manner will be to develop technologies in which the excess baggage is minimized.
The MixAlco technology discussed here does address many of the negative aspects related to certain other cellulosic technologies. I think the keys to their success in the long run will be how well they are able to integrate the combustion of leftover biomass for process heat into their process. If that aspect is successful, then the process should be able to produce liquid fuels with relatively low fossil fuel inputs.
There are numerous resources that describe the MixAlco process. Professor Holtzapple has made a couple of presentations available:
Terrabon’s site has a detailed process description:
The Wikipedia site is also very informative:
Finally, one can go to the Professor Holtzapple’s lab website to read about the latest developments:
1. Granda, C., Holtzapple, M., Luce, G., Searcy, K., and Mamrosh, D. (2009). Carboxylate Platform: The MixAlco Process Part 2: Process Economics Appl Biochem Biotechnol, 156:537–554.
2. *Granda, C., Zhu, L., and Holtzapple, M. (2007). Sustainable Liquid Biofuels and Their Environmental Impact. American Institute of Chemical Engineers. Environ Prog, 26: 233–250.
3. Holtzapple, M., and Granda, C. (2009). Carboxylate Platform: The MixAlco Process Part 1: Comparison of Three Biomass Conversion Platforms Appl Biochem Biotechnol, 156:525–536.