Obstacles in the way

Delays in achieving economic viability of second-generation ethanol keeps investors at arms’ length, but some companies are staying in the race

Raízen’s second-generation ethanol plant in Piracicaba


The race to commercially produce second-generation ethanol from cellulose is taking much longer than expected. Several competitors have given up midway, and now the moment is decisive as the remainder struggle for leadership. Three industrial plants seem to be close to reaching technological efficiency and economic viability. Two are in Brazil: one belongs to Raízen, in Piracicaba, São Paulo, and another to GranBio, in São Miguel dos Campos, Alagoas. Both use sugarcane waste for raw materials, such as the leaves and bagasse left over from pressing in traditional bioethanol production. The third plant, owned by the consortium Poet-DSM, is located in Emmetsburg, Iowa, USA, and converts leftovers from corn production into fuel.

The persistence of these three companies is extraordinary in an environment that has been accumulating dashed expectations. A decade ago, evidence that mature technology already existed to produce biofuels on a large scale from bountiful raw materials (such as wood or agricultural waste) attracted a considerable volume of risk capital, which was invested in constructing what were to be the first commercial plants. High oil prices (peaking above US$100 per barrel between 2008 and 2011) and strong public incentives in the United States to produce second-generation ethanol made the scenario even more attractive. But the decision to build the plants proved much riskier than expected. Roadblocks in the production processes halted continuous operation of equipment and combined with high costs of some of the materials used, showing that research and investment were still needed to construct a more competitive path.

Raízen Converting sugarcane bagasse into fuel is nearing commercial viabilityRaízen

This was followed by a period of depression. Data from the consultancy Bloomberg New Energy Finance show that global investments in new-generation biofuels, which reached nearly US$3 billion in 2011, fell to less than US$1 billion in 2013 and little more than half a billion in 2016. Companies have revised their plans. The Spanish group Abengoa closed its plant in Hugoton, Kansas, in the United States in 2015. Last November, the multinational DowDuPont put its cellulosic ethanol plant in the American city of Nevada, Iowa, up for sale for US$225 million, and has not yet found a buyer. The company announced that it will continue to offer specialized inputs in the biofuels market, such as genetically modified yeasts capable of improving production yields. The unit has the capacity to produce 110 million liters of ethanol per year, but has never operated commercially.

Startups created in the United States to support the industry adjusted their business plans; this was the case at Solazyme in San Francisco, which has turned to food production, and Amyris, in Emeryville, California, which now manufactures cosmetics, fragrances, and anti-malaria medications. “The path toward developing this technology has been longer and substantially more expensive than anticipated by the specialists,” says engineer Viler Janeiro, director for cellulosic ethanol business at the Sugarcane Technology Center (CTC), a biotechnology company that has built a demonstration plant in the municipality of São Manuel in São Paulo State.

GranBio Bioflex, GranBio’s mill in São Miguel dos Campos, Alagoas: technology imported from Italy did not work as promisedGranBio

In first-generation ethanol production, only one-third of the biomass is used in the process of fermenting the sucrose in the cane juice. The challenge in second-generation production is to also utilize the bagasse and leaves, which are sources of cellulose, hemicellulose, and lignin and account for the remaining two-thirds of the plant’s energy and are not metabolized in the conventional system. In general, the technologies used in second-generation mills pretreat the biomass to break the structure of the lignocellulosic material; this is followed by hydrolysis, in which enzymes are used to convert cellulose and hemicellulose polymers into sugars, and then fermentation, using genetically modified yeasts that transform the sugars from the biomass into ethanol. Yeast development has made uneven progress in utilizing cellulose and hemicellulose. Other microorganisms more efficiently break apart hexoses (the six-carbon sugars from cellulose) than pentoses (five-carbon sugars resulting from hydrolyzing hemicellulose). Another important hurdle to achieving economic feasibility involves the cost of enzymes needed to generate the sugars, which is still considered very high.

For chemical engineer Carlos Eduardo Vaz Rossell, it is crucial to improve the efficiency of enzymes and reduce their price. “Brazil has an advantage, which is the availability of a large volume of biomass in the plant itself in the form of sugarcane bagasse. This helps in the search for economic viability and justifies new investments in research,” says Rossell, who from 2010 to 2016 coordinated the pilot plant of the Brazilian Laboratory of Bioethanol Science and Technology (CTBE), which is linked to the Brazilian Center for Research in Energy and Materials (CNPEM) in Campinas.

As for the use of the cane leaves, there are challenges with regard to productivity and logistics, since this resource must be gathered from the fields and brought to the industrial facility. Chemical engineer Antonio Bonomi, coordinator of the process intelligence division at CTBE, says that we do not yet know for sure how much leaf material should be removed from the fields for use as raw material. “Maintaining part of the leaves improves the productivity of the sugarcane, preserving moisture and nutrients.” In addition to these challenges, unforeseen engineering problems have arisen; some require expensive and complex solutions and made it impossible for most of the ethanol mills to meet their deadline for starting operations.

Enzymes needed to produce second-generation ethanol are still expensive

Artur Milanez, biofuels manager at the Brazilian Development Bank (BNDES), disagrees with the current perception that the delay in economic viability indicates a failure of the technology. “We are currently in the final stage of research and development (R&D). When technological routes are tested, it is common for there to be a process of funneling until the most efficient process is obtained,” he states. In 2011, BNDES and the Brazilian Funding Authority for Studies and Projects (FINEP) contributed R$3 billion to the Joint Plan to Support Technological Innovation in the Sugarcane Energy and Chemical Industrial Sectors (PAISS) to develop new technologies for processing sugarcane biomass. The portfolio of projects included Raízen and GranBio, as well as companies such as Abengoa, the CTC, and Petrobras.

It appears that Raízen is closest to reaching commercial scale operations. This company is a joint venture between the Brazilian Cosan and the multinational Shell, and in 2014 established a plant in Piracicaba which simultaneously produces both generations of biofuel. During the last harvest, the mill produced 12 million liters of cellulosic ethanol; the goal is to reach 25 million liters this year and 40 million liters (the total installed capacity) next year. Raízen benefited from Shell’s investment a decade ago in the Iogen energy company, which has a pilot plant extracting ethanol from wheat straw in Ottawa, Canada. “We analyzed the viability of the technology and saw that it was interesting,” says Antonio Stuchi, executive director of technologies and projects at Raízen.

In Canada, the notion of using wheat straw did not come to fruition: the process required quantities of water considered economically unfeasible. But the knowledge was useful for Raízen’s strategy. “We were unsure whether the enzymes would be effective and we were in the dark about the yeasts’ ability to work with the pentoses,” he recalls. Neither of these fears materialized. “But engineering problems caught us by surprise.” High levels of impurities in the biomass led to the creation of a pre-cleaning step. The process of separating the sugars and lignin needed to be redesigned, adding more filters and centrifuges. But the main problem was the corrosion of equipment in the pretreatment step, which led to intermittent operations. “Cane leaves, which are rich in silica, can cause much more damage than wheat straw,” says Stuchi.

Agronomist-engineer Gonçalo Pereira, a professor in UNICAMP’s Institute of Biology who was the cofounder and chief scientist at GranBio, notes that the companies made the mistake of transferring technologies from the pilot plants to the mills without testing them at an intermediate level. “The main hurdles were expected to be seen in the biotechnological steps, with the enzymes and yeasts, which were overcome. Nobody imagined that there would be problems in the mechanical part of the process, in the pretreatment equipment,” he says. He adds that cellulose was expected to behave as it does in the paper industry. “While the structure of lignocellulose in wood repels water and softens during the pretreatment, sugarcane bagasse acts like a sponge and forms a kind of porridge with fiber and silica. This material cannot be easily transported or dried out, which leads to erosion, clogged pipes and valves, and seized pipe threads,” he says.

The problems Raízen faced during the experimental production of second-generation fuel did not stop the plant from continuing to produce first-generation ethanol, while its competitors constructed units to exclusively produce cellulosic ethanol and were unable to get them to work. Stuchi explains that the strategy to produce both types of ethanol permitted several synergies. “For first-generation fuel, we obtain a surplus of biomass that is part of the operational cost. We use this raw material that is in the mill for the second-generation ethanol. There are also converging processes of fermentation and distillation. Another point is the demand for water. You start with solid biomass and this must be diluted. I can use the water that evaporated from the first-generation unit. This has helped contain the need for additional investments and demonstrate to shareholders that our processes could be more competitive.” Joint production of first- and second-generation fuel is proving useful to ensure the sustainability of the process, but this does not mean that the hybrid model will dominate. “The advantage of a mill that operates only with second-generation is that it would have more flexibility in production, bringing raw materials from other crops suited to the productive process,” says Artur Milanez of BNDES.

One technique to speed up construction of second-generation mills was to combine technologies licensed from other companies. Milanez recalls that there was a belief that Brazil could use technologies developed abroad and only need to make “tropicalizing” adjustments. “We have always believed that adaptation to the characteristics of Brazilian biomass would be a challenge, and Raízen’s success shows that it was essential to create solutions locally,” he says.

DuPont DowDuPont cellulosic ethanol plant in the United States, which is for saleDuPont

GranBio has purchased licenses for multiple technologies to establish its Bioflex mill in Alagoas. The first yeast tested was provided by DSM, a Dutch company, while the Danish Novozymes furnished the enzymes for hydrolysis and the Italian Mossi Ghisolfi (MG) was responsible for the pretreatment and hydrolysis systems. The technology did not work as promised, particularly the package purchased from the Italian group. “What we saw initially was a true collapse of the system,” says Bernardo Gradin, president of GranBio. To make things more complicated, GranBio lost its main channel to MG with the tragic death of one of the group’s owners, Guido Ghisolfi; he apparently committed suicide in March 2015, just four months after Bioflex began testing. “For nearly a year we were unable to run the plant, which was involved in the investigation. Innovation required much more patience than we imagined.”

The dispute landed in a London arbitration court, and a negotiated solution is expected soon. At the end of 2017, the Italian company entered into judicial recovery and activities were interrupted at the mill in the city of Crescentino, which was built to produce ethanol from rice and wheat straw and giant reed (Arundo donax). In addition to the R$750 million spent to build the plant and the energy cogeneration system, GranBio invested R$150 million in R&D, which involved a team of 45 researchers and the acquisition of complementary patents, with financial support from FINEP. They also invested more than R$40 million in R&D to develop sugarcane varieties for energy, which can provide more than 2.5 times the biomass of traditional sugarcane.

Today, little of the Italian package remains at Bioflex. A new pretreatment process was developed, and enzymatic hydrolysis extended. MG promised to promote hydrolysis in just 19 hours, while the current technology requires between 48 and 90 hours. To support the new process, six fermentation tanks needed to be built in addition to the previously existing two, to hold the biomass for a longer period and tailor the wastewater system. GranBio invested in its own technology (for example, creating yeasts in partnership with UNICAMP) and went looking for technology abroad: it invested in a US company, American Process Inc. (API), which had developed platforms for biomass pretreatment to produce cellulosic sugar. Today it holds more than a hundred patents in the United States.

Iowa State University Bales of cornhusks collected in the state of Iowa for fuel productionIowa State University

Equipment corrosion was also a serious problem for the company. In a process known as steam explosion, the biomass is subjected to very high temperatures and pressures and then suddenly decompressed; initially, when this material slammed against the walls of the pre-treatment equipment it caused failures and production stoppages almost every day. The process had to be simplified, eliminating one of its steps. “We moved from a technology that proved unfeasible, with two pretreatment stages and one involving steam explosion, to a technology with a single, less intense cooking stage, known as liquid hot water (LHW), plus a mechanical treatment step for the fibers using equipment specifically made for this purpose,” says Gradin. After this, the plant began operations in 2017, reaching production capacity of 28 million liters of second-generation ethanol, 5 million liters of which were exported to the United States.

More resources are needed to ensure commercial viability—R$35 million this year and R$45 million in 2019—to reach production of 45 million liters and 60 million liters, respectively. In August, Gradin will present GranBio at an event in the Netherlands as a company which has changed since 2014, now with technologies developed in Brazil and the ability to license them worldwide.

Although it is slight, increased interest in second-generation ethanol is stimulated by a new escalation of oil prices, which were at approximately US$75 dollars per barrel in recent months. In an interview with the Financial Times, Feike Sijbesma, an executive with DSM, a Dutch company that produces cellulosic ethanol in Iowa in partnership with the American company Poet, said that the oil prices boost the company’s chances of offering a competitive product. “What level is comfortable for us? Around US$70 a barrel,” said Sijbesma.

The Poet-DSM plant, dubbed Project Liberty, was inaugurated in 2014. Its goal is to produce 72 gallons (272.5 liters) per ton of corn residue, and it is getting close to reaching this target: the mill has already attained the level of 70 gallons (265 liters) per ton. A year ago, the Poet-DSM consortium announced the construction of a unit to produce enzymes used to break down cellulose from corn waste. According to Sijbesma, the main challenge has been to organize collection of this raw material; as with sugarcane, there have been difficulties removing dirt and sand.

The robust nature of sugarcane leaves, which are rich in silica, eroded the pretreatment equipment

In the opinion of CTBE’s Antonio Bonomi, part of the fragility in the research efforts involving alcohol from cellulose comes from the fact that there is a global fuel market that can press for technological advances. “And even in Brazil, second-generation fuel faces competition from first-generation, a successful route that works very well.” A more favorable situation emerged two years ago, when the Paris Accord established commitments to limit global temperature increases and proposed low-carbon emissions scenarios in which biomass energy plays a central role. In Brazil, the launching of a new National Policy for Biofuels, which rewards sustainable ethanol production, may serve as an additional stimulus for second-generation fuel (see Pesquisa FAPESP issue no. 266).

A pioneer in the study of biomass use for energy production, American biologist Lee Lynd, is a professor at Dartmouth College’s Thayer School of Engineering. Lynd maintains that one mistake committed by governments and investors has been to bet heavily on the construction of large power plants and worry less about financing technological advances that can reduce production costs. In developing renewable solar and wind energy, he observes, investments were first made in niches and then more ambitious targets were sought. Initial small-scale applications permit faster learning, Lynd stated in an article published in October in the journal Nature Biotechnology. “Battery technologies were employed in electronic products before they were used in hybrid cars,” he offered as an example.

Lynd was one of the founders of Mascoma, a biofuel research company which in 2005 received contributions from investors such as Vinod Khosla, founder of Sun Microsystems, and the US Department of Energy. Like other startups, Mascoma was not able to meet the goal of converting inedible biomass into ethanol. In 2014, Mascoma was sold to a Canadian company, Lallemand, which was interested in yeasts developed using a technique Lyn created. However, the biologist did not abandon his original plans. He now works in another company, Enchi, which pursues goals similar to Mascoma.