As promised, here are my post-publication thoughts on Tanger et al. 2013, Biomass for Thermochemical Conversion:
This topic was somewhat of a departure for me, as my own research focuses on greenhouse gas (GHG) accounting for bioenergy production from a system optimization perspective. I find thermochemical conversion (TC) systems especially interesting because of how widely applicable they are, particularly in developing-country settings, and I had some basic familiarity with gasification and pyrolysis technologies in the context of distributed electricity generation, clean cooking technologies, and biochar production. However, all of my previous work has focused on optimization of feedstock production landscapes or system configurations considering embodied GHGs as the metric of ‘quality’, to be balanced against the economic cost of different system designs.
The actual physical qualities of biomass from a conversion perspective is a dimension that I had always overlooked. While one hears a lot about the importance of low-lignin feedstocks for biochemical conversion, I had always assumed that most TC processes were feedstock-flexible enough that there wasn’t much of an analogous issue there. However, a presentation at last year’s Agronomy Society of America meeting by these folks highlighted the importance of proper agronomic management on the physical characteristics of the resulting TC feedstock. When Paul started talking about a review article on breeding to achieve the same goals, my interest was piqued.
One of my big take-aways from this effort was an appreciation for how ‘messy’ thermochemical conversion is, in the sense of being complex and somewhat ill-defined. Slow pyrolysis, fast pyrolysis, and gasification all involve heating biomass to transform it into solid, liquid, and gaseous products, but they are optimized for different product fractions and differ radically in terms of process conditions (temperature, pressure, heating rate, etc.) and energy balance (endothermic to exothermic) involved. However, as far as I can tell, the associated terminology is actually quite imprecise! Different terms are used for the same types of products from different processes; one man’s ‘bio-oil’ is another man’s ‘tar’, likewise for soot/biochar. Even more fundamentally, how to distinguish between pyrolysis and gasification varies- some define the threshold in terms of temperature, others in terms of whether or not the process involves autothermal partial oxidation. I have always thought the latter was the clearest way to differentiate (i.e., pyrolysis implies an inert atmosphere and gasification/carbonization imply a partial oxidizing atmosphere), but there are examples of just the opposite terminology (‘pyrolysis’ in an oxidizing atmosphere and ‘gasification’ in an inert) in the literature.
The terminology associated with biomass characterization is equally disorienting. While biologists typically talk about cell wall composition in terms of the cellulose/hemicellulose/lignin and other bio-polymers present (‘biochemical analysis’), engineers typically focus on either the raw elemental composition (the fractions of C, H & O, termed ‘ultimate analysis’) of the biomass, or the fractions that volatilize or remain after controlled heating (‘proximate analysis’, see the figure from Paul’s post). All three paradigms are equally valid for describing the same pile of biomass, and breeding to change the cellulose:lignin content will also change the C:H ratio and volatility of the biomass as well. On of our primary tasks in producing this review was to try and bridge this difference in a clear but precise manner, explaining the science of the conversion in a way that someone with a biology background could easily grasp, without completely glossing over the points that a TC engineer would find important.
In the process, we came up with two insights I’m proud of, which were not obvious by inspection. First, considering that the different biomass characterization paradigms are alternate but complementary ways to describe the same thing, we came to the realization that proximate/ultimate analysis may be the more appropriate paradigm for rapid high-throughput biomass characterization for breeding. These measurements are typically conceptually simpler and more amenable to automation than their biochemical analysis counterparts. Additionally, since proximate analysis is a more direct evaluation of the performance of biomass during TC, measuring it directly ‘cuts out the middle man’ and alleviates the need to try and predict TC performance from biochemical properties. I doubt we’re the first to suggest this, but I think we did a nice job of covering it in a thorough and systematic way.
The second insight was to highlight the potential breeding strategy of targeting high-lignin, low-ash feedstocks. In the course of our readings we noticed that multiple studies* have shown an inverse relationship between lignin and mineral content across large biomass sample sets (perhaps not an unexpected result, as these are though to have similar functional roles in plant cell walls). However, to the extent that mineral (i.e., ash) content is universally bad for TC and lignin is typically beneficial (greater heating value than an equal mass of cellulose), trading minerals for lignin should improve TC conversion properties, perhaps with minimal effect on overall plant viability. This hypothesis is a complete departure from the biochemical conversion literature, where the exact opposite is true- lignin is something to be avoided and ash content is of little consequence.
All-in-all, participating in this paper was a very interesting and educational experience, and I’m very pleased with the response thus far.
* there are a few studies showing no relationship or even a weak positive correlation, but we found the other view to be more common and compelling