Abstract
In our work analyzing the technologies and market forces that impact biobased fuels and chemicals, we are often asked to assess the impact of a less-expensive feedstock, a “breakthrough” processing technology, or a shift in regulation or subsidy. We also see many studies claiming that biobased fuels or chemicals are (or are not) carbon-neutral, do (or do not) compete with food, will (or will not) deplete the world's water supply, and will (or will not) replace a petroleum-based equivalent. Which is right?
Everyone from government ministers to protesters in the streets wants a simple answer to these kinds of questions. But it is in seemingly simple questions like these that the devilishness of details is most evident. That is because, in biomass-to-chemicals processes, all the factors are highly dependent on each other. Only on the level of a specific plant in a specific place can accurate assessments be made, and scientists and engineers model these because models are necessary to provide useful answers. However, these accurate models do not scale up to valid assessments of the same question for the entire world, or a country, or even a small region. Many or all of the assumptions made for one installation are different everywhere else; hence, most simple answers about the big picture are wrong.
For example, a recent study, which claimed that 8 of 10 scenarios for rapeseed biodiesel production in the European Union (EU) failed to meet a targeted 35% greenhouse gas savings, received a lot of attention. This study—or rather, the public's and press reactions—illustrate the danger of top-line conclusions and simple averages: 2 of 10 scenarios are passing the 35% target, and the others fall along a spectrum from there on down. In other words, there is no one number, and the average is meaningless since we do not know whether or not the 10 scenarios are a representative sample of all growers in the EU. We do not even know how to pick a representative sample, because there are so many variables that go into farming (i.e., size of the farm, condition of the soil, rainfall, soil composition, age of the implements) and every other factor in the production chain. Rather than just setting a goal of 35%, biobased fuel and chemical developers (and regulators) should look at what the two successful scenarios have in common and promote practices and policies along those lines.
Technology—Panacea or Placebo?
While the costs of operating biorefineries using various configurations of gasification and enzymatic hydrolysis are difficult to calculate on a global average basis, a well-crafted model that enables exploration can identify local opportunities for improvement in cost and performance. All other factors being equal, these technologies can chip away at the cost, material, and environmental uncertainties that plague assessments of biobased fuels and chemicals, and hence might be applicable in a broader context. Many opportunities for improvement come from swapping out existing technologies for newer ones. And looking to adjacent industries—wastewater treatment in particular—provides insight and reduces technology risk. Some specific examples can help illustrate how technologies can address the claims of common truisms.
PROFITABILITY
Profitable biofuels and chemicals require greater volumes of cheap, local biomass
Biobased chemicals are mostly derived from the easily fermentable starchy parts of cane and corn, but these crops cannot supply nearly enough material to replace even a minute fraction of petroleum-based products. Developers have been trying to coax vastly more biomass from sources such as algae and cellulosic plants, whether they be purpose-grown crops like switchgrass and miscanthus or forest and agricultural waste. The high hope (and critical need) for biomass technologies has brought new funding to startups like NexSteppe (Malibu, CA) and Renmatix (King of Prussia, PA) from corporations like DuPont (Wilmington, DE) and BASF (Ludwigshafen, Germany). Other companies focused on growing more cellulosic biomass cheaply include Mendel Biotechnology (Hayward, CA), Viaspace (Irvine, CA), and Repreve Renewables (Soperton, GA), which claims it will be able to deliver dry miscanthus to biorefineries as far as 50 miles away for $50–60/ton. Around the world, companies are looking for species and strains that thrive in their local climate; a major Indonesian industrial conglomerate is experimenting with indigenous crops like kemiri sunan and nyamplung, while also exploring jathropha, miscanthus, and algae. This feedstock flexibility will always be the best hedge against high grower payments; if one commodity rises in price, developers such as TMO Renewables (Guildford, UK)—which claims to be able to use 28 different cellulosic feedstocks—will always have lower-cost alternatives.
FEEDSTOCK RANGE
Converting a wider range of feedstocks such as waste and cellulosic material will help
The technology for fermenting easily available sugars in starchy plant parts like wheat, potatoes, and grapes is thousands of years old. However, methods of transforming other biomass to useful intermediates is newer and less efficient, and dozens of contenders in gasification, pyrolysis, and acid and enzymatic hydrolysis are competing to be the conversion technology of choice for producing intermediates like sugars, syngas, and pyrolysis oil. As in cultivation, the picture is frustratingly fuzzy. Successes in the transformation of cellulosic ethanol to biofuel, in the form of new loan guarantees for Abengoa (Seville, Spain), Fiberight (Catonsville, MD), and ZeaChem (Lakewood, CO); POET's (Sioux Falls, SD) and DSM's (Herleen, the Netherlands) joint venture; and Novozymes' (Bagsværd, Denmark) and M&G's (Milan) big new Italian facility, bode well for capacity increases. But stubbornly high production costs, difficulty dealing with inhibitors, feedstock variability, and other factors are killing off firms like Qteros (Marlborough, MA) and letting technologies such as gasification—which is feedstock-agnostic—catch up.
HIGH-VALUE INTERMEDIATES
Processing intermediates into higher-value chemicals (not ethanol) improves the profit picture
Given the market challenges facing ethanol, many developers are looking to produce higher-value chemicals from sugars and syngas. For example, Dow Chemical (Midland, MI) and Mitsui (Tokyo, Japan) have a joint venture to produce polyethylene (PE) plastic from biobased ethanol, and Braskem (São Paulo, Brazil) is already marketing its biobased PE at a price premium. Cellulosic-processing startups that initially focused on ethanol have found new genes and microbes that produce higher-value materials directly; for example, ZeaChem is developing a new organism that excretes three-carbon (C3) propionic acid, which it converts to industrial chemicals like propylene. In fact, there are known catalytic and fermentation routes from syngas and sugars to nearly every chemical derived from petroleum; one leading chemical company believes it can make any such molecule from biomass, if the incentive is high enough.
Translating Technology Promises into Techno/Econometric Analysis
The road to economic viability requires improvements in dozens of steps in cultivation, conversion, and processing to higher-value output, and increasingly detailed techno/econometric analysis shows areas where costs might be cut or a higher-value product might be produced. Based on detailed models of the biobased fuels and chemicals value chain, linking process steps to their input and output materials, it is possible to quantify the effect of a given technology in one representative situation. For one recent study, we drew on complex biorefinery design models such as the National Renewable Energy Laboratory's (NREL) Aspen model and Oklahoma State University's (OSU) Cellulosic Ethanol Feasibility Template and designed a new model that focuses on operational costs technology can help address. Wary of unproven claims, we populated the model with plausible values from the middle of the observed range, rather than calculated industry averages or observed local examples. Companies such as Mascoma (Cambridge, MA) and TMO Renewables, for example, claim yields of 60 to more than 100 gallons of ethanol per ton of feedstock; our model assumed 70 gal/ton. That baseline allows for an objective assessment of the impact of a new technology on a typical case, but the flexibility to explore tradeoffs and relative advantages of different improvements.
In one example, better baling machines might improve the economics of cellulosic biomass more than cheaper enzymes. Enzymatic hydrolysis is still an emerging science, so while it has been at a cost disadvantage that kept it largely confined to laboratories, today it is being commercialized in new facilities like GraalBio's (São Paulo, Brazil) 82-million-liter plant in Brazil. The site will use Chemtex's (Wilmington, NC) Proesa technology as well as the latest enzymes from Novozymes and DSM. Enzymes are expensive, but being improved; Novozymes says it takes 50 kg of its Cellic CTec3 enzymes to make one ton of ethanol, compared to 250 kg of competitors' products. Engineering and manufacturing these enzymes is a complex process, requiring special skills and equipment that companies like Novozymes guard fiercely and keep in house, despite the logistical and cost burden of trucking them in from a central facility. It is also a touchy topic. To quote Novozymes' website: “Many believe that there are several potential scenarios for enzyme production for cellulosic ethanol production. Yet onsite production is not viable today, nor will it be in the future.”
Despite this leading firm's cautionary words (or perhaps because of them), many are trying. Mascoma and the now-defunct Qteros developed “consolidated bioprocessing” organisms that expressed cellulases to release sugars from biomass then fermented those sugars into chemicals. Mascoma, an IPO hopeful, even acquired SunOpta (Brampton, Canada) to bolster its pretreatment abilities. In yet another approach, Agrivida (Medford, MA) and Syngenta (Basel, Switzerland), with its Enogen corn, are integrating the enzymes in the feedstock organism itself. And rising competition from the likes of DSM and DuPont is sure to bring prices down. Assuming these kinds of advances in enzyme production lead to a new low-cost scenario, enzyme production still adds $20 per ton, or about $0.30 per gallon, to the total cost of cellulosic ethanol.
Tinkering with enzymes is difficult work, and improvements are hard-won. However, feedstock costs are almost always the biggest driver of the cost of biobased materials and chemicals, accounting for roughly 30% to 40% of the total—typically $50 to $75 per ton, or about $0.85 per gallon. Simpler technologies that can make a bigger dent in cultivation, harvesting, or densification might make a much bigger dent in the overall cost. For example, the Kansas Alliance for BioRefining and BioEnergy (KABB, Wichita, KS) believes technology and process improvements, such as new baling machines, might cut stover feedstock costs by $25 per ton, reducing the cost by some $0.35 per gallon of cellulosic ethanol.
Lessons for Modeling Cost and Consequence of Technologies in the Biobased Value Chain
Policymakers, markets, and Mother Nature do not make it easy for biobased fuels and chemicals firms, policymakers, or regulators. The road to environmental and economic viability requires improvements in dozens of steps in cultivation, conversion, and processing to higher-value output, and each modification and improvement has to be assessed and adjusted to fit the specific designs of plants being built in various locations around the world. Obviously, any specific plant design needs a detailed and specific engineering plan, but general techno/econometric analysis does show areas where costs usually creep in. Following are three key recommendations for decision-makers that want to understand where these cost bottlenecks can be alleviated with new technologies and how those improvements could play out on a regional or global level.
MODEL BUILDING
Start simply, build up to a credible and useful level of detail, and then stop
Models exist on a spectrum of precision, from a back-of-the-napkin flowchart to a real-time climate simulator driven by satellite data. A good model stained with your cocktail can be more valuable than another running on a government supercomputer! This is particularly important in areas of emerging technology, biology, economics, and environmental systems—the nexus of biobased materials and chemicals—since data are seldom comprehensive or accurate, equations have not been validated by experiment, and the actions of players in one place and time change the parameters for actors in other places and times. Even the most detailed model is a simplification of reality, and a high level of detail can lull the audience into a false sense of security. A simpler structure reminds the audience that much is still left to be discovered, decided, and done.
FLEXIBILITY
Determine performance and cost drivers (such as energy inputs or market prices) for each product and process in the model
Test ranges for the steps and cost factors that are well-characterized and quantifiable (and for which actual and theoretical ranges are known), leaving these variables open so they can be re-evaluated in a variety of scenarios. Overall industry averages or specific company claims are meaningless in light of the extreme variation in local conditions from site to site, or even at a single site over time—a warning to makers and users of any model—so interactivity is key.
ADDING DETAIL
Moving into still greater detail, model not only the flows of mass and money through the system, but the underlying physical principles behind them
For example, the amount of sunlight hitting a cultivation facility has both positive and negative effects: it accelerates growth, so more biomass feedstock is produced, but it also increases evaporation, and replacing lost water can become an enormous expense. As before, use values based on the range of high, mid, and low observations, but provide both financial and physical equations to link the components of the system together.