In The Lab

The Need for Speed: Rapid Biomass Analysis Makes Better Breeding Possible
By Jerry W. Kram
"What's in it?" That's the question that analytical labs have had to answer for as long as they have existed. For biomass processors, it is a serious-even critical-question.

Biomass can be tricky stuff. It is a mixture of cellulose, hemicellulose, lignin, protein, sugars and other components. A process that works well with corn stover as a feedstock might not work as well with switchgrass and might not work at all with softwoods such as pine. Therefore, biomass producers are interested in learning what is in their feedstocks. "If you are going to understand conversion process yields, you need to know what is going into the front end of the process," says Bonnie Hames, senior chemistry manager for Ceres Inc. There are standardized methods for determining the composition and abundance of all the constituents of biomass. The problem is that the conventional wet chemistry methods for determining the composition of biomass are slow and expensive-between $1,000 and $3,000 a sample, she says.

Ceres is a plant-breeding company specializing in the development of cellulosic feedstock crops. Developing new plant varieties is a volume business, requiring the screening of tens of thousands of individual plants for valuable traits. Even when genetic engineering techniques such as those developed by Ceres are used, the resulting plant progenies still have to be screened for their value to the customer. "If you're trying to improve feedstock quality or assess the risk that you might encounter in your process, you really need to understand that variability," Hames says.

To get around the bottleneck, Hames adapted an off-the-shelf technology widely used in the food and feed industries, building on work she pioneered at the National Renewable Energy Laboratory. Near-infrared spectroscopy (NIRS) has been around since the 1970s, but her innovation was to apply that technology to cellulosic biomass. "It's really a high-impact, low-risk technology," Hames says. "It's taking something that has been demonstrated in many other industries and applying it to this new field." Hames and the other researchers at Ceres found they could analyze samples in minutes instead of days for about $20 instead of thousands of dollars with close to the same precision and accuracy as the standard wet chemistry methods.

Because it is a common analytical tool, there are a large number of vendors providing a wide range of NIRS systems. Some are optimized for use in the lab, while others have been miniaturized-some as small as a digital camera, powered by a handheld computer-for use in the field or the processing facility.

The industry isn't quite done with wet chemistry, however. NIRS works by comparing a spectrum of a sample to a spectrum calibrated to known samples. According to Hames, a good NIRS method needs at least 100 to 300 samples analyzed by standard wet chemistry methods. This has to be done for each major category of biomass, such as switchgrass, corn stover, bagasse or wood chips. "An uncalibrated NIRS instrument is like a car without gasoline," Hames says. "You need the ability to interpret the spectrum for the parameters you are interested in."

Having these tools in-house is a significant advantage for Ceres. By traditional methods, analyzing 500 samples with the company's current staff would take one year and $1 million. Using NIRS to analyze 500 samples takes about three days and $10,000. "We can keep an eye on yield per acre and the quality of that product for individual processes as we develop our energy crops," Hames says.

NIRS is an important component of Ceres' quest to develop energy crops tailored to the needs of biomass processors. "We are using this technology to assess what the natural variability would be in these plants," Hames says. "We can try to figure out what changes are genetic versus environmental. We are looking at how feedstocks can change with storage. It allows us to evaluate crop yields not just in tons per acre but in actual gallons of [final] product per acre."