Biorefineries and platform chemicals: Biomass as a chemical feedstock

1.1 Plant matter is mostly sugar polymers, locked in a tough composite

Kamm and colleagues estimate that plants build about 170 billion tonnes of biomass a year. Roughly 75% is carbohydrate (mainly cellulose, starch and sucrose), 20% is lignin and only about 5% is everything else: oils, proteins, dyes and flavours. Of the total, people use about 6 billion tonnes, and only 3 to 3.5% of that goes to non-food uses such as chemicals.

The carbohydrates come in very different forms. Starch, the energy store in grain and tubers, is a chain of glucose joined by alpha links. Enzymes in our own gut can cut it, so industry can too, cheaply. Cellulose is also pure glucose, but joined by beta links that let neighbouring chains line up and hydrogen-bond into crystals, rather like strands of a rope pressed into a solid bar. Tolan of Iogen gives cellulose as 35 to 50% of most plant material, with chain lengths of 1,000 to 10,000 glucose units.

Hemicellulose (15 to 25%) is a branched polymer of mostly five-carbon sugars, chiefly xylose and arabinose. It is easy to break: dilute acid at about 120 C dissolves it. Lignin (15 to 30%) is not a sugar at all but a three-dimensional network of phenolic propane units. It glues the fibres together, keeps water out and resists microbes. No natural enzyme splits it cleanly into monomers.

The practical upshot is that grain and sugar cane hand over their sugar easily, while straw, stover and wood guard theirs. That single difference separates first-generation plants, which are commercial, from lignocellulosic ones, which mostly are not.

• Kamm (2006), Vol 1 Ch 1 Biorefinery systems - an overview · pdf pp. 44, 49, 58
• Kamm (2006), Vol 1 Ch 9 Iogen's demonstration process · pdf pp. 223, 233
• Kamm (2006), Vol 1 Ch 3 Development of biorefineries (Dean et al.) · pdf p. 104

1.2 Oil lacks oxygen; sugar carries too much, and removing it costs mass

Petroleum is mostly chains of CH₂ units with almost no oxygen. Petrochemistry therefore works by adding function: oxygen, nitrogen or chlorine is bolted onto ethylene, propylene or benzene to make useful intermediates. Lichtenthaler, writing in Kamm, puts the contrast neatly. Hydrocarbons are oxygen-free and devoid of functionality; carbohydrates are 'overfunctionalized' with hydroxyl groups. Making industrial chemicals from sugar is mostly the opposite job: taking oxygen out.

The DOE screening report (Werpy and Petersen 2004) found exactly this. Because biomass is already highly oxidised, few of its routes need oxidation, while many need hydrogenation or dehydration, that is, adding hydrogen or stripping water.

Removing oxygen costs mass. When yeast turns glucose into ethanol, Kent gives the maximum yield as 0.51 g of ethanol per g of sugar; the rest leaves as carbon dioxide. Dale and Kim note the upside: over 90% of the glucose's energy stays in the ethanol. So sugar-to-fuel conversion is efficient in energy terms but loses half the weight.

Triglycerides (fats and oils) are the exception. Their long fatty chains look like hydrocarbons, so Task 42 calls fatty acid derivatives the easiest biomass components to handle in today's chemical infrastructure. That is why oleochemicals are the oldest and largest bio-based chemical industry, and why Clements counts them as biomass's largest contribution to today's chemical industry.

The Lichtenthaler table below makes a surprising point: by the kilogram, the cheapest sugars cost about the same as basic petrochemicals. The catch is the mass lost on the way to a product.

• Kamm (2006), Vol 2 Ch 1 The key sugars of biomass · pdf pp. 467-470
• Werpy & Petersen (2004), Section 7 Pathways and challenges · pdf p. 25
• Kent (2012), Ch 33 Biomass conversion · pdf p. 1260
• Kamm (2006), Vol 1 Ch 2 Biomass refining global impact (Dale & Kim) · pdf p. 82
• Task 42 (2012), Sections 9.2 and 9.7 · pdf pp. 18, 28

1.3 Wet, bulky and seasonal: why all biomass is local

Crude oil is a dense liquid that can be piped across continents and stored in tanks. Biomass is a solid with a lot of water. Godshall notes in Kent that bagasse leaves the cane mill at about 50% water; Hettenhaus reports corn stover at harvest is typically 30 to 50% moisture and must dry in the field. Baled stover has a dry density of only 112 to 160 kg per cubic metre.

It also arrives all at once. Hettenhaus gives a harvest window of 20 to 40 days, while a lignocellulosic biorefinery needs 2,000 to 3,000 dry tonnes every day of the year. Delivering a million dry tonnes to one site during harvest would mean about 100 trucks an hour on a 20-hour day. Trucking is charged by distance, 1.00 to 1.50 dollars per km, and bulky loads hit the volume limit before the weight limit: round bales on load-and-go wagons carried 9 dry tonnes against a normal 18.2 t, doubling transport cost per tonne.

Storage is a trade-off too. Dry bales lose over 10% of their mass and burn easily; wet piles above 60% moisture lose under 5% and cannot catch fire, but water then rides along to the plant.

Dale and Kim draw the conclusion: because biomass is bulky, it pays to refine it near where it grows. Clements and Van Dyne quantify the limit. For a combined ethanol and furfural plant the best size was 4,360 tonnes of feedstock a day; below about 750 t/d capital and operating costs were too high, and above about 4,400 t/d gathering enough feedstock cost too much.

• Kamm (2006), Vol 1 Ch 14 Biomass commercialization and agriculture residue collection · pdf pp. 353-354, 361-366
• Kamm (2006), Vol 1 Ch 2 Biomass refining global impact · pdf pp. 88-89
• Kamm (2006), Vol 1 Ch 5 The lignocellulosic biorefinery (Clements & Van Dyne) · pdf pp. 152-154
• Kent (2012), Ch 35 Sugar and other sweeteners · pdf p. 1412

See it in the table

Check your understanding

Why does Werpy and Petersen's screening find that many sugar-to-chemical routes need hydrogenation or dehydration?

Glucose costs about the same per kilogram as ethylene, yet sugar-based ethylene is expensive. Why?

Clements and Van Dyne found an ethanol and furfural plant became less profitable above about 4,400 t/day. What drove this?