# Biorefineries and platform chemicals: Conversion routes: how the plants work

Source page: https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/p3/

[Understanding Chemicals](https://www.materialinsights.org/learn/chemicals/) / [Module 13: Biorefineries and platform chemicals](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/) Part 3 of 5

## Conversion routes: how the plants work

Between the field and the product sit four families of process. Pretreatment and enzymes release sugars. Fermentation turns sugars into molecules. Chemical catalysis rearranges sugars directly. Heat breaks biomass into gas or oil. Each has its own conditions, yields and failure modes, and the economics of a biorefinery are set largely inside these vessels.

### 3.1 Pretreatment and enzymes: the expensive key to cellulose

Without pretreatment, Tolan writes, the enzyme demand for cellulose is too high to be practical. Iogen's answer was steam explosion with a little acid. Milled wheat straw is hit with high-pressure steam and 0.5 to 2% sulphuric acid at 180 to 260 C for 0.5 to 5 minutes, then the pressure is released suddenly. The hemicellulose hydrolyses to xylose, opening surface area; the straw comes out as a dark, muddy mass. Cooking too hard turns xylose into furfural and glucose into hydroxymethylfurfural, losing sugar and making fermentation inhibitors, so Iogen aims to hydrolyse less than 20% of the cellulose at this stage. Tolan compares it to roasting a turkey: the longer the time, the lower the temperature.

Older routes skipped enzymes. Concentrated acid (72% sulphuric, 40% hydrochloric) gives high yields at low temperature but needs about 99.9% acid recovery and exotic materials, which has blocked commercial use. Dilute acid at 180 to 200 C for 1 to 4 hours gives about 50% glucose yield, and Kent cites kinetic models capping it at 65 to 70%, because glucose degrades to HMF and levulinic acid. Plants built this way in wartime Germany and Russia suffered corrosion, low yields and poor returns.

Enzymatic hydrolysis is slow. Iogen runs a 5 to 15% solids slurry at pH 5 and 50 C in 200,000 gallon tanks for 5 to 7 days, reaching 90 to 98% cellulose conversion. The rate falls to under 2% of its initial value within 24 hours, and cellobiose, the glucose dimer, inhibits the enzymes at as little as 5 g/L. On-site enzyme production saves stabilisers and transport.

Process flowchart Iogen's wheat straw to ethanol demonstration process (Ottawa, 2004) Inputs **Wheat straw (45.5% cellulose, 19% arabino-xylan, 20% lignin)**enough for up to 2,000 gal/day of ethanol at demonstration scale **Sulphuric acid and steam**0.5-2% acid **Cellulase broth (Trichoderma, made on site)**100 L per t cellulose **Recombinant Saccharomyces yeast**ferments glucose and xylose Process steps 1 **Bale handling, chopping and milling**

Small particles for even acid distribution
 2 **Steam explosion pretreatment**

180-260 C, 0.5-5 min, rapid pressure release; hemicellulose → xylose; under 20% of cellulose hydrolysed
 3 **Enzyme production**

Trichoderma in 50,000 gal fermenters, 30 C, pH 4-5, aerobic, about one week
 4 **Enzymatic hydrolysis**

5-15% solids, pH 5, 50 C, 5-7 days; 90-98% of cellulose to glucose
 5 **Lignin separation**

Plate and frame filter; cake washed to recover sugar
 6 **Fermentation and distillation**

Glucose and xylose to ethanol; distil, dehydrate, denature with 1% gasoline
 Outputs **Fuel ethanol**about 75 US gal per t straw (about 284 L/t) **Lignin cake**burned for steam and power in a full plant

A full plant was planned at about 170,000 gal/day (60 million gal/yr). Dale and Kim expect about 400 L/t for mature technology.
 Source: Kamm (2006), Vol 1 Ch 9 Iogen's demonstration process (Tolan), pdf pp. 221-235  Common misreading

#### Is enzyme cost the barrier?

In 2004 Wittcoff wrote that cellulases cost about 150 dollars per tonne of ethanol, against 33 dollars needed for feasibility. Genencor and Novozymes, under DOE contracts, then cut cellulase cost per gallon by a factor of about ten, to an estimated 0.50 dollar per gallon, with a DOE target of 0.10. Dean and colleagues concluded enzyme cost could no longer be regarded as the largest barrier. The remaining problems are pretreatment capital, slow rates and mixed, inhibitor-laden sugar streams.

- Kamm (2006), Vol 1 Ch 9 Iogen's demonstration process · pdf pp. 226-235
- Kent (2012), Ch 33 Biomass conversion · pdf p. 1289
- Kamm (2006), Vol 1 Ch 6 Lignocellulosic feedstock biorefinery: history (Katzen & Schell) · pdf p. 165
- Kamm (2006), Vol 1 Ch 3 Development of biorefineries · pdf pp. 104, 109
- Wittcoff (2004), Ch 14 Carbohydrates · pdf p. 483

### 3.2 Fermentation is judged on titre, rate and yield

Kent's chapter on industrial biotechnology names the three numbers that define a fermentation: rate (productivity, grams per litre per hour), yield (carbon conversion, grams of product per gram of sugar) and titre (final concentration, grams per litre). Each drives a different cost. Yield sets raw material cost. Rate sets how many tanks you must build. Titre sets how much water you must remove.

Dale and Kim give the arithmetic. With glucose at 0.10 dollar per pound and a 90% yield of lactic acid, sugar costs 0.11 dollar per pound of product; at 50% yield it costs 0.20 dollar. A lower yield also nearly doubles reactor volume for the same output, and separation costs rise even faster because they scale with the volume of dilute broth handled.

Werpy and Petersen set targets for commodity diacids: productivity of at least 2.5 g/L/h and a fermentation cost at or below 0.25 dollar per pound. Their Frye and Holladay chapter in Kamm gives 2 to 3 g/L/h for succinic acid.

Acidity is the hidden cost. Most organisms prefer pH 6 to 8, but organic acids have pK values of 3 to 5, so the broth must be neutralised with lime, ammonia or caustic. The product comes out as a salt that must be converted back to acid, and Werpy, Frye and Holladay note acidification can be up to 30% of total cost. In lactic acid plants, sulphuric acid turns calcium lactate into lactic acid plus gypsum, a low-value by-product to dispose of. Enzymes add another mismatch: cellulases work at 50 to 60 C and pH 3 to 5, most production strains at 30 to 40 C and pH 6 to 8.

Data Fermentation performance reported in the sources

| Product | Titre | Rate | Yield | Notes |
|---|---|---|---|---|
| Ethanol (yeast) | 8-10% (about 80-100 g/L) | 0.5-2 kg/m3/h free cells; 10-30 immobilised | max 0.51 g/g sugar | Zymomonas bacteria: 4-5 free, 50-80 immobilised |
| 1,3-Propanediol (engineered E. coli) | 135 g/L | 3.5 g/L/h | 51 wt% on glucose | 18 genes modified |
| 1,3-Propanediol from glycerol (natural strains) | 50-75 g/L (over 100 with new strains) | about 2 g/L/h; 30 immobilised | 0.44-0.69 mol/mol | glycerol feed |
| Lactic acid (engineered yeast, low pH) | over 100 g/L | 0.8 g/L/h | 1.19 mol/mol glucose | free acid at low pH |
| Succinic acid (MBI strains, vials) | 68-96 g/L | target 2-3 g/L/h | 85-94 wt% | acetate and pyruvate co-products |
| Itaconic acid (A. terreus) | 180-200 g/L | 3-6 day batch | 50-70% | 39-42 C, pH 2-4 |
| PHB (R. eutropha, sugar mill) | 150 kg/m3 cells, 75% polymer | 1.89 kg/m3/h | 98% fermentation yield | fed batch, 32-34 C |

Rates and yields come from different organisms, scales and years; compare with care.
 Source: Kent (2012), Ch 30 Industrial biotechnology; Kamm (2006), Vol 1 Ch 10, Ch 17 and Vol 2 Ch 13, Kent pdf pp. 1159-1166; Kamm pdf pp. 246, 415, 418-419, 824-825  Worked example

#### Reading a titre

Kent reports itaconic acid fermentations reaching 180 to 200 g/L from a 30% sugar medium. Thirty per cent sugar is 300 g/L, so the yield is 180/300 to 200/300, or 60 to 67%, consistent with his 50 to 70% range. Compare ethanol: Kent gives 8 to 10% ethanol after 28 to 72 hours, and Thoen and Busch 6.5 to 11%. Even the best broth is roughly nine parts water to one part product, and all that water must be boiled off or otherwise separated.

- Kent (2012), Ch 30 Industrial biotechnology: discovery to delivery · pdf pp. 1137, 1159-1166
- Kamm (2006), Vol 1 Ch 2 Biomass refining global impact · pdf p. 85
- Werpy & Petersen (2004), Section 9.1 · pdf pp. 31-32
- Kamm (2006), Vol 2 Ch 13 Succinic acid (Werpy, Frye & Holladay) · pdf pp. 823-824
- Kamm (2006), Vol 1 Ch 3 Development of biorefineries · pdf p. 111
- Kamm (2006), Vol 2 Ch 14 Polylactic acid (Gruber et al.) · pdf pp. 838-839

### 3.3 Chemical catalysis rearranges sugars directly

Not every route needs a microbe. Some of the largest bio-based chemicals are made with metal or acid catalysts acting on sugars, and Werpy and Petersen note that chemistry dominates the second step, turning building blocks into derivatives.

Hydrogenation adds hydrogen to the sugar's aldehyde group. Glucose becomes sorbitol over Raney nickel in batch reactors, batch because food uses demand almost no leftover sugar; Engelhard showed a continuous ruthenium-on-carbon process at near 99% yield. Xylose gives xylitol the same way. Further steps are harder: sorbitol to isosorbide reaches about 76% (90% desired), and hydrogenolysis of sorbitol to propylene glycol gives about 35 lb per 100 lb, against 60 needed.

Dehydration strips water. Pentoses give furfural with commercial yields around 50%, or 85% with improved hydrochloric acid processes. Hexoses give HMF, which in water rapidly goes on to levulinic acid and formic acid. The Biofine process exploits that sequence in two reactors, removing products quickly to limit tar. Its theoretical ceiling is 71.6% levulinic acid by weight of hexose; in practice it converts about 50% of hexose mass to levulinic acid, 20% to formic acid and 30% to tar, and 50% of pentose mass to furfural.

Selective hydrogenation upgrades fermentation acids. PNNL turned succinic acid to gamma-butyrolactone at over 95% yield on palladium-on-carbon at 150 to 175 C and 800 to 2,000 psig, leaving acetic acid untouched, then to 1,4-butanediol at over 95% selectivity. The recurring challenge is catalysts that tolerate the impurities in fermentation broths and prove their lifetime in long continuous runs.

Process flowchart The Biofine process: lignocellulose to levulinic acid, formic acid and furfural Inputs **Lignocellulose (paper sludge, residues, waste paper)**1,000 dry t/day design (50% cellulose, 20% hemicellulose, 20% lignin) **Dilute sulphuric acid**1.5-3%, recycled; make-up 3,500 t/yr **Ethanol (for ethyl levulinate)**35,000 t/yr Process steps 1 **Shredding and mixing**

Particles 0.5-1 cm, conveyed by air into mixing tank with recycled acid
 2 **Reactor 1 (plug flow)**

210-220 C, 25 bar, about 12 s; polysaccharides → soluble intermediates such as HMF
 3 **Reactor 2 (mixed)**

190-200 C, 14 bar, about 20 min; HMF → levulinic + formic acid; furfural and volatiles taken off
 4 **Separation and drying**

Gravity separator, vacuum dehydration; tar cracked to dry char; crude 75% levulinic acid purified to 98%
 5 **Acid recovery and esterification**

Acid recycled; levulinic acid esterified with ethanol
 Outputs **Ethyl levulinate**133,000 t/yr **Formic acid**38,500 t/yr **Char (lignin and ash)**burned on site; 3.1 MW electricity exported **Ash**17,500 t/yr

Kent gives 210-230 C and 195-215 C for 15-30 min, and a levulinic yield of 75% or more of C6 content.
 Source: Kamm (2006), Vol 1 Ch 7 The Biofine process (Hayes et al.); Kent (2012), Ch 33, Kamm pdf pp. 173-175, 189; Kent pdf p. 1291  Case

#### Belle Glade furfural, 1966-1997

Quaker Oats and Katzen built the world's largest furfural plant at Belle Glade, Florida, on bagasse from neighbouring sugar mills. Expanded, it processed 1,800 dry tonnes of bagasse a day to make 137 t/day of furfural, a mass yield of about 7.6%. By-product methanol and HMF were burned with the residue for steam and power, and the sugar mills stored up to 227,000 dry tonnes of bagasse for year-round running. It closed in 1997 after 31 years, undercut by cheaper crude furfural from China.

- Werpy & Petersen (2004), Sections 7, 9.11 and 9.12 · pdf pp. 25, 65-68
- Kent (2012), Ch 33 Biomass conversion · pdf pp. 1290-1291
- Kamm (2006), Vol 1 Ch 7 The Biofine process · pdf pp. 173-175
- Kamm (2006), Vol 2 Ch 13 Succinic acid · pdf pp. 829-832
- Kamm (2006), Vol 1 Ch 6 Lignocellulosic feedstock biorefinery: history · pdf p. 162

### 3.4 Heat instead of microbes: gasification and pyrolysis

Thermochemical routes break biomass with heat. The product depends on temperature, time and oxygen. Kent's table for wood gives fast pyrolysis about 75% liquid, 12% char and 13% gas; slow carbonisation 30% liquid, 35% char and 35% gas; gasification 5% liquid, 10% char and 85% gas.

Gasification runs at about 750 to 850 C. Biomass dries, pyrolyses, then its char reacts with steam and carbon dioxide to give carbon monoxide and hydrogen, while the water-gas shift (CO + H2O → H2 + CO2) adjusts the ratio. Overall the process absorbs heat, so either some fuel is burned inside with air, diluting the gas with nitrogen, or pure oxygen is used, which Brown calls extremely expensive, or heat is supplied indirectly. Raw biomass syngas carries tar, up to 10% by weight in some gasifiers, plus ash and char, and Werpy and Petersen identified gas cleanliness as the key barrier.

Fast pyrolysis heats finely ground biomass (under 1 to 2 mm, dried below 10% water) at 100 to 1,000 C per second to about 500 C, with vapour residence under 2 seconds and rapid quenching. It yields up to 75% bio-oil, a dark liquid with 15 to 30% water, pH about 2.5, 37% oxygen and a heating value of 16 to 19 MJ/kg, under half that of fuel oil. It is corrosive, unstable and immiscible with hydrocarbons, so upgrading it to transport fuel was not economic. Small chemical outlets exist: liquid smoke flavouring, pyrolytic lignin as a phenol substitute in resins, and potentially levoglucosan.

So and Brown compared routes to ethanol (1997 dollars): pyrolysis plus fermentation 0.42 dollar per litre, enzymatic SSF 0.34, acid hydrolysis 0.35.

Data Biomass syngas composition by gasifier type

| Gasifier | H2% | CO% | CO2% | CH4% | N2% | Heating value (MJ/m3) | Tars |
|---|---|---|---|---|---|---|---|
| Air-blown updraft | 11 | 24 | 9 | 3 | 53 | 5.5 | High |
| Air-blown downdraft | 17 | 21 | 13 | 1 | 48 | 5.7 | Low |
| Air-blown fluidised bed | 9 | 14 | 20 | 7 | 50 | 5.4 | Medium |
| Oxygen-blown downdraft | 32 | 48 | 15 | 2 | 3 | 10.4 | Low |
| Indirectly heated fluid bed | 31 | 48 | 0 | 21 | 0 | 17.4 | Medium |

Dry volume per cent. Air dilutes syngas with about half nitrogen; downdraft units are limited to about 2 MW thermal.
 Source: Kamm (2006), Vol 1 Ch 11 Hybrid thermochemical-biological processing (Brown), Table 11.1, pdf p. 260  In the data

#### Three ethanol routes, one feedstock price

So and Brown sized each plant at 95 million litres a year using about 240,000 dry tonnes of biomass at 46 dollars per tonne. Capital was 64 to 69 million dollars and raw materials about 11 million dollars a year in every case; capital charges (12.8 to 13.8 million dollars) were the largest single item. Per litre, biomass cost about 11.1/95 = 0.12 dollar, roughly a third of the 0.34 to 0.42 dollar total.

- Kent (2012), Ch 33 Biomass conversion · pdf pp. 1294-1297
- Kamm (2006), Vol 1 Ch 11 Hybrid thermochemical-biological processing · pdf pp. 258-260, 269-275
- Werpy & Petersen (2004), Section 6 · pdf p. 24

### See it in the table
[H₂O Water](https://www.materialinsights.org/commodities/water/)[Ni Nickel](https://www.materialinsights.org/commodities/nickel/)[Pd Palladium](https://www.materialinsights.org/commodities/palladium/)[NaCl Salt](https://www.materialinsights.org/commodities/salt/)[Gyp Gypsum](https://www.materialinsights.org/commodities/gypsum/)[CaO Lime](https://www.materialinsights.org/commodities/lime/)[Wh Wheat](https://www.materialinsights.org/commodities/wheat/)[Oa Oats](https://www.materialinsights.org/commodities/oats/)[Su Sugar](https://www.materialinsights.org/commodities/sugar/)[H2SO4 Sulphuric acid](https://www.materialinsights.org/chemicals/sulphuric_acid/)[HCl Hydrochloric acid](https://www.materialinsights.org/chemicals/hydrochloric_acid/)[H2 Hydrogen](https://www.materialinsights.org/chemicals/hydrogen/)[O2 Oxygen](https://www.materialinsights.org/chemicals/oxygen/)[CO2 Carbon dioxide](https://www.materialinsights.org/chemicals/carbon_dioxide/)

### Check your understanding

Why does Iogen deliberately hydrolyse less than 20% of the cellulose during steam pretreatment?
 Pretreatment aims to open the structure. More severe conditions convert glucose to HMF and xylose to furfural, losing sugar and poisoning yeast.

A fermentation reaches the same titre and rate but its yield falls from 90% to 50%. What happens to raw material cost per kg of product?
 Raw material cost per unit product is sugar price divided by yield: 0.10/0.9 = 0.11, 0.10/0.5 = 0.20.

Which statement about fast pyrolysis bio-oil is supported by the sources?
 Bio-oil's elemental composition resembles biomass, not petroleum; its oxygen makes it acidic, unstable and low in energy.

[← Part 2: The biorefinery concept and its types](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/p2/)[Mark read · Part 4: Platform chemicals: the DOE top twelve and their families →](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/p4/)

## Other sections

- [Module overview](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/index.md)
- [Biomass as a chemical feedstock](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/p1/index.md)
- [The biorefinery concept and its types](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/p2/index.md)
- [Platform chemicals: the DOE top twelve and their families](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/p4/index.md)
- [Economics and real cases](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/p5/index.md)
- [Key terms, quiz and sources](https://www.materialinsights.org/learn/chemicals/biorefineries-platform-chemicals/review/index.md)