3.2.1 Economic Considerations

Total cost can be divided into capital expenditures (CAPEX) and operating expenditure (OPEX). CAPEX can be subdivided into plant costs, off-site costs, and engineering costs. Plant costs represent the capital necessary for the installed process equipment with all auxiliaries that are needed for complete process operation (ie, piping, instrumentation, insulation, foundations, and site preparation). Off-site costs are not directly related to the process operation, they rather include costs of the addition to the site infrastructure, for example, power generation units, boilers, pipelines, offices, etc. According to the American Association of Cost Engineers, fixed capital investment can be divided into outside battery limits (OBL) or off-site and inside battery limits (IBL). IBL comprises one or more geographic boundaries, to specify the area where production takes place, enclosing all associated equipment and production facilities. OBL includes facilities such as storage, utilities, administration buildings, or auxiliary facilities. Engineering costs include costs for detailed design of equipment, construction supervision, project management, etc.

OPEX consists of fixed and variable costs. Variable costs comprise cost of feedstock and supplies, waste management, product packaging, finished and semifinished products in stock, etc. Fixed costs comprise salaries, taxes, license fees, interest payments, marketing costs, etc.

A generic approach on how to conduct the economic assessment, the most relevant cost parameters, and ballpark figures for first-generation (1G) feedstock is depicted in Fig. 3.1.

Investment costs are an important contributor to the economic feasibility of lignocellulosic biorefineries (Anex et al., 2010; Brown, 2015). Investment costs are more important for second-generation (2G) biorefineries than for 1G biorefineries because feedstock costs are lower and processing lignocellulosic biomass is more challenging (Hytönen, 2009; Zinoviev et al., 2010; Brown, 2015); hence capital investment costs are higher.

Innovative new conversion technologies usually follow a development pathway from the lab, to piloting, then demonstration, and finally the construction of a commercial plant. The number of years for a biobased product to reach commercialization depends heavily on economics and hence on drop-in versus nondrop-in (existing demand and infrastructure), type of conversion technology, and supply chain integration. Various methods are applied for estimating investment costs when plants or processes are scaled-up.

3.2.1.1 Factorial Method

Preliminary cost estimates can be based on the purchased equipment cost with all additional cost components based on it and estimated through different “factors,” that is, certain percentages of the purchased equipment cost. The factorial cost estimation method was developed by Lang (1947, 1948) and frequently used in chemical engineering in early design stages. Although biorefineries may differ from chemical plants, a similar approach is appropriate (Stanil, 2011). The accuracy of the result of the factorial method depends on what stage the design has reached at the time the estimate is made, and on the reliability of the data available on equipment costs. Quantified predictions can only be done for a limited amount of time ahead, since uncertain market conditions and rapid technology development result in sharp price inflation.

The equation for the cost estimation is:

$C_{\text{c}}=f_{\text{L}}*C_{\text{e}},$ (3.1)

(3.1)

where C_c = fixed capital costs, f_L = Lang factor, can be ob_tained from chemical engineers design books (eg, Peters et al., 2002; Towler, 2007), and C_e = total costs of major equipment.

The “Lang factor” depends on the type of process, that is, solids processing (f_L = 3.1), liquid processing (f_L = 4.2) and liquids–solids processing (f_L = 3.6).²

If more detailed information is available, compounds such as piping, electrical power, control instruments, etc., can be considered individually. The individual costs are summed to generate the total capital investment costs.

3.2.1.3 Exponential Method

The cost of a future plant can be estimated based on historical cost data of similar plants. The same applies for estimating the costs of specific process equipment, for example, heat exchanger, distillation units, etc.

$\frac{AC{C}_1}{AC{C}_2}={\left(\frac{Q_1}{Q_2}\right)}^m*cor{r}_F,$ (3.2)

(3.2)

where ACC₁ = annual capital cost of unit or scale (Q₁), for example, large scale; ACC₂ = annual capital cost of unit or scale (Q₂), for example, demo scale; m = scale exponent; and corr_F = correction factor, if the design is highly comparable corr_F = 1.

Calculating the capital costs for a biorefinery or specific equipment is not straightforward because comparable facilities are not built yet and thus appropriate factors, for example, Lang factors, for future biorefineries are hardly available. If future biorefineries are comparable to established chemical plants, then the cost estimation methods described above can be also used to estimate the product price depending on the scale-up. They can also be used to define optimization targets for the process design and the scale-up process. An example of how the exponential method can be used is shown below.

Example:

For the scale-up the residence time and the conversion rates (cf_1–n) are the most relevant parameters. As first approximation product costs (PC) can be estimated based on throughput (TP), conversion factor (cf), annual capital costs (ACC), operation costs (OC), and revenue (R) according to the following equation:

$PC=\frac{1}{TP*cf}*(OC+ACC)-R$ (3.3)

(3.3)

The selling price (SP) can be expressed as function of the residence time if the throughput and the conversion rate are assumed to be constant.

$PC(rt)=\frac{1}{TP*cf}*(OC(rt)+ACC(rt))-R(rt)\le SP$ (3.4)

(3.4)

The acceptable residence time (rt) for a plant at technical scale can be calculated using the SP as a constraint. The revenue from byproducts depends linearly on the throughput. Hence the revenue can be expressed as function of the throughput when the SP for each byproduct (n) is used:

$R={\displaystyle \sum _1^{n}TP*c{f}_n*S{P}_n}$ (3.5)

(3.5)

ACC does not increase linearly with scale; therefore the acceptable residence is higher at large scale than at technical scale for a given minimum selling price (MSP). The exponential method can be used to estimate the ACC of large-scale plants (LS) based on ACC of technical scale plants (TS). The exponential method is frequently used in chemical engineering and is expressed as:

$AC{C}_{LS}=AC{C}_{TS}*{\left(\frac{LS}{TS}\right)}^m$ (3.6)

(3.6)

where the exponent (m) usually varies between 0.6 and 0.7.

Using Eqs. (3.5) and (3.6) in Eq. (3.4) allows calculating of the acceptable residence time for a given scale-up (TS→LS). The equations above also allow defining optimization targets, for example, conversion factors for the main and the byproduct(s) under consideration of the MSP of the product(s) for a given process design. However, those calculations are based on point estimates and incur a substantial uncertainty due to price fluctuations. One method to reduce the uncertainty is to model the statistical distribution of prices in combination with sensitivity analysis methods such as Monte Carlo simulation.

The methods described above are applicable if comparable plants exist. In the last decade many plants have been built for the production of biofuels and bioenergy, therefore reliable cost data exist for those technologies. Table 3.1 provides an overview of typical investment and operation costs of matured technologies at different scales.

Table 3.1

Overview of investment and operation costs of bioenergy-related technologies^a

Technology	Plant production capacityb			Investment costs (range)	Operation costs (range)	Preferred feedstock
	Minimum capacity	Typical capacity	Maximum Capacity
Pyrolysis (fast)	15,000 t/year	45,000 t/year	300,000 t/year	600–3000 € per kWinst	11% of investment	Wood and dry biomass
Anaerobic digestion	0.3 MWe	0.5 MWe	4 MWe	1800–4000  € per kWinst (0.5–2 MW)	11% of investment	Nearly all kinds of organic biomass
First-generation bioethanol	10,000 t/year	250,000 t/year	1,000,000 t/year	700–1000 € per t/year	11% of investment	Starch (wheat, potatoes, etc.)
Esterification (biodiesel)	10,000 t/year	100,000 t/year	500,000 t/year	400–600 € per t/year	11% of investment	Oils and fats
Direct liquefaction	5000 t/year	50,000 t/year	200,000 t/year	900 € pert/year	11% of investment	All kinds of organic waste streams

^aThis information is kindly provided by NOVIS GmbH.

^b1 t=1000 kg=1 Megagram (Mg).

In general, the first-of-its-kind commercial plant shows disadvantages with regard to the total capital investment but also operational costs in comparison to succeeding installations. Several installations are frequently needed to enhance efficiencies due to technical learning. High-value material utilization of all product streams is essential to achieve profitability of complex biorefineries.

The production of bioethanol is the most matured biotechnological process. Historical data allow estimating the cost reduction potential due to technological progress, which has been made for ethanol production from woody material and agricultural residues. The techno-economic development of ethanol from cellulosic feedstock is depicted below.

In the early technology development stage ethanol production costs can be used as a proxy for other products derived from cellulosic sugars, for example, butanol, furan derivatives, etc. However, that is just a rough estimation because equipment costs and energy costs differ, although similar equipment is needed for purification and upgrading.

3.2.2 Economic Lessons Learned From Bioethanol and Bio-Oil Derived From Lignocellulosic Biomass

The National Renewable Energy Laboratory (NREL), Golden, CO, USA, has developed case models that document progress and cost targets for energy carrier production from cellulosic feedstock (Anex et al., 2010; Kazi et al., 2010; Swanson et al., 2010; Wright et al., 2010; Davis et al., 2013). The economics analysis includes at a conceptual level of process a design to develop flow diagrams using commercial simulation tools, for example, ASPEN PLUS. The capital investment costs are estimated based on the plant design using the factorial method; working costs are estimated based on similar chemical plants (Wright et al., 2010). The simulation software is also used to calculate the mass and energy flows in order to estimate the operation costs. Detailed information of the technology model is available (Aden, 2008, 2009).

The NREL cost model for ethanol and/or hydrocarbons from cellulosic feedstock is based on a plant capacity of 2000 t (biomass demand) per day (t/day). For bioethanol the model includes the following process steps:

Kumar and Murthy (2011) have modeled the conversion of grass straw to ethanol; grass straw is a byproduct of grass seed production and has a cellulose content of 31%. They used dilute acid, dilute alkali, hot water, and steam explosion as pretreatment in their model to estimate the production costs of all pretreatment methods followed by fermentation. Lignin residue was sufficient to cover the energy demand of pretreatment, conversion, and upgrading. The authors estimated ethanol production costs as US$ 1.06, 1.12, 1.02, and 1.09 per kg ethanol for dilute acid, dilute alkali, hot water, and steam explosion pretreatment, respectively, at a plant capacity of approximately 750 t/day (Kumar and Murthy, 2011). The main factors were feedstock costs and enzyme costs. Kazi et al. (2010) investigated the conversion of corn stover to ethanol and came to the same conclusion that enzyme costs and feedstock costs are the most important factors.

There are several optimization options that are suitable to reduce the production costs, for example, improved conversion efficiency of underutilized biomass fractions (mainly hemicellulose) with production of additional value-added products. Besides feedstock and enzyme costs, the pentose conversion rate is another important option for reducing ethanol production costs. Xylose is the most important pentose obtained via acid hydrolysis and can be thermochemically converted into furfural. Furfural, a common industrial chemical, is a potential platform chemical for biopolymers (Choudhary et al., 2012).

Interest in drop-in biofuels, especially for aviation, continues to grow. Tao has conducted a techno-economic assessment for the production of n-butanol, iso-butanol, and ethanol from corn stover. In contrast to cellulosic ethanol fermentation technology, n-butanol and especially iso-butanol fermentation technology, has not yet been fully demonstrated even in bench studies (Tao et al., 2014). However, the authors concluded that if fermentation of xylose and glucose to butanol would reach 85% yield, the minimum SP would then be comparable to that of ethanol.

Results from various NREL techno-economic studies about lignocellulosic bioethanol production in the period from 2001 to 2012, shown in Fig. 3.2, reveal that cost reduction up to a factor of four to five could be achieved due to technical learning over such a long time period. That may also be achievable for the production of higher value-added products in the next decade.

For the production of cellulosic ethanol, cost reduction was mainly achieved through technological improvements in the areas of pretreatment and conditioning, enzymatic hydrolysis (and associated improvements in enzyme performance reflecting commercial enzyme package improvements), and fermentation, while the feedstock cost contributions to ethanol SP remained relatively constant.

For 1G biofuels, feedstock significantly influences production costs, whereas for 2G biofuels, the share decreases and becomes less than 40% (Hamelinck et al., 2005). The production cost of ethanol from cellulosic biomass is sensitive to economy of scale (Argo et al., 2013; Muth et al., 2014). However, the optimal size of biorefineries depends also upon the nature of the feedstock. Preprocessed feedstock in depot systems enables larger biorefinery sizes—at least conceptually but might compromise greenhouse gas emission (GHG)-reduction targets (Muth et al., 2014).

For processes dealing with high volumes of raw material and high capital costs, marginal changes in feedstock cost (including transport) can make the difference. Therefore, in assessing the economic viability of a lignocellulose biorefinery, trade-off between plant size and feedstock delivery costs must be taken into account. Ringer et al. (2006) provided scale-dependeny cost information on the production of bio-oil from wood chips.

Fig. 3.3 shows that the specific production costs decrease with scale, although the feedstock costs increases mainly due to transport costs. There is obviously a limit, where cost reduction due to technological progress is compensated by increasing cellulosic feedstock (and transport) cost (Argo et al., 2013). Transport costs can vary substantially. Estimation of feedstock cost is not straightforward due to the lack of formal markets for a large part of possible feedstock and due to site-specific availability and procurement constraints. This aspect is discussed in detail in chapters “Biomass Supply and Trade Opportunities of Preprocessed Biomass for Power Generation” and “Commodity Scale Biomass Trade and Integration With Other Supply Chains.”

Currently, the production of hydrocarbons from cellulosic feedstock is not cost-competitive to hydrocarbons from crude oil. Davis et al.(2013) summarize their findings as: “Tailoring the hydrolysate stream to the microorganism tolerance will be essential for improving overall yields and lowering production costs. Reduction of hydrolysate conditioning costs could also be realized through a better understanding of the tolerance of hydrocarbon-producing microbes to lignin and other cellulosic sugar substrate impurities and other potential inhibitors. There are several optimization possibilities that are suitable to reduce the production costs, for example, improving conversion efficiency of underutilised fractions of the biomass (mainly hemicellulose) or conversion of lignin to higher value-added products.” However, the latter would require the replacement of lignin by other energy carriers and that might lead to a trade-off between maximizing the GHG savings and reducing the production costs. Currently, lignin is mostly used as fuel on-site to cover the heat and/or power demand of the biorefinery.

For the techno-economic evaluation of innovative industrial processes, often limited data are available. The economic assessment of biorefineries depends on a number of factors such as feedstock costs (Argo et al., 2013; Muth et al., 2014), specific preprocessing and process equipment (Muth et al., 2014), plant capacity, etc. Moreover, some of those factors are site-specific, for example, feedstock logistics, process waste utilization, integration in existing installations, etc. Economic estimations have been conducted for green biorefineries in Ireland (O’Keeffe et al., 2011, 2012) and elsewhere (Höltinger et al., 2014), sugar cane biorefineries in Brazil (Cavalett et al., 2012), for thermochemical production of biofuels (Bridgwater, 2009; Reyes Valle et al., 2015), the production of advanced biofuels (Turley et al., 2013), lignocellulosic biorefineries (Kazi et al., 2010; Benali et al., 2014; Cheali et al., 2015), integrating bioethanol production into Kraft pulp and paper mills (Hytönen, 2009) but also for biorefinery-relevant separation techniques (Sievers et al., 2014) and pretreatment techniques (García et al., 2011). The results of the studies are hardly comparable because assumptions such as interest rate, equipment lifetime, and modeling approaches are different.

Investment cost estimations are mostly point estimates and do not take the uncertainties inherent in the methodology into account. The accuracy of the result depends on the quality of the input data and the calculation approach. Brown (2015) has recently reviewed techno-economic studies of thermochemical cellulosic biorefinery and conducted a sensitivity analysis based on economic data available in the public domain. He concluded that the choice of estimation methodology differs across the techno-economic assessments and has a substantial impact on the final result. The difference of capital investment costs of a biorefinery with a capacity of 2000 t/day can be as high as US$ 300,000 million depending on the chosen estimation method and assumptions (Brown, 2015). Therefore, it is important to assess the uncertainty by sensitivity analysis if estimation methods are used for supporting decision making.

Daugaard et al. (2015) discuss how increasing biorefinery capital and feedstock learning rates could significantly reduce optimal size and production costs of biorefineries. The authors suggest that there is an economic incentive to invest in strategies that increase the learning rate for producing advanced biofuels.

3.3.1 Fast Pyrolysis: Current Status

Currently, the main objective of fast pyrolysis is to produce bio-oil that can be directly used in stationary boilers or in the near future after hydrogenation as transportation fuel. In the mid-term bio-oil could also be used as drop-in (after partial upgrading) in naphtha-crackers of conventional refineries or even as source for chemicals. The technological development in the area of fast pyrolysis of mainly woody biomass is driven by the demand for 2G biofuels.

3.3.1.2 KIT (Bioliq)³

The bioliq process, developed at the Karlsruhe Institute of Technology (KIT), aims to produce synthetic fuels and chemicals from biomass (Dahmen et al., 2012). This process requires a feed that can easily be fed to the gasifier at elevated pressures being atomized by oxygen as the gasification agent. Fast pyrolysis was chosen as the most promising process to produce this feed, by mixing pyrolysis condensates and char to a so-called bioliqSyncrude, exhibiting a sufficiently high heating value, and being suitable for transport, storage, and processing. This slurry-gasification concept has been extended to a complete process chain via an on-site pilot plant erected KIT.

The three subsequently constructed parts of the plant, funded by the German Ministry of Food, Agriculture, and Consumer Protection, consist of the fast pyrolysis and biosyncrude production, the 5 MW_th high-pressure entrained-flow gasifier operated at up to 8 MPa (both in cooperation with Lurgi GmbH, Frankfurt), as well as the hot gas cleaning (MUT Advanced Heating GmbH, Jena), dimethylether, and final gasoline synthesis (Chemieanlagenbau Chemnitz GmbH).

The technical concept of the pilot plant is based on the Lurgi-Ruhrgas mixer reactor, previously devoted to coal degassing or heavy crude coking. The actual flow scheme is depicted in Fig. 3.4. In November 2014, the whole process chain was demonstrated successfully at KIT and the first synthetic fuel was produced.

3.3.1.3 Fortum

The new technology has been developed into a commercial-scale concept in cooperation between Fortum, Metso, UPM, and VTT as part of TEKES Biorefine research program. Fortum commissioned a first-of-its-kind fast pyrolysis plant in 2013, which has been operating since 2014. It was the first-of-its-kind at commercial scale and is shown in Fig. 3.5. The bio-oil plant is integrated with existing combined heat and power production and an urban district heating network in Joensuu, Finland. Wood chips from forest residues are dried and then fine-milled before they are fed into the pyrolysis reactor together with hot sand. The bio-oil production capacity is 50,000 t/year. The bio-oil and the char are conveyed to the CHP unit, where the bio-oil replaces heavy fuel oil. The steam is used the district heating network and surplus electricity is fed into the national grid. This helps to reduce CO₂ emissions by 59,000 t/year and SO₂ emissions by 320 t/year.

3.3.2 Biochemical Conversion

Theoretically, the source of sugar is irrelevant for the fermentation process (whether derived from sugar cane, sugar beet, or lignocellulose), provided the sugar solution contains a suitable composition of sugars and does not contain inhibitors. However, specific microorganisms have specific requirements; they cannot use all sugars, are differently robust and produce distinguished products. Therefore, there are a number of possible combinations of feedstock, pretreatment options, and microorganisms, sometimes even to produce the same product.

3.3.3 Biorefineries: Starch/Sugar-Based

Biobased chemicals and biofuels must compete on cost and performance with petrochemicals and petroleum fuels. At present, most chemical-producing biorefineries use starch or sugar as the major feedstock.

BASF, Novozymes, and Cargill jointly developed 3-hydroxypropionicacid a raw material for the production of biobased acrylic acid, in pilot scale.⁴ On the other hand, OPX Biotechnologies with its partner Dow Chemicals is aiming to commercialize bio acrylic acid by 2017. Their process is based on the direct conversion of dextrose (corn) or sucrose (cane) through genetically engineered E-Coli1/2 via OPXBIO EDGE Technology Platform. The global acrylic acid market is forecast to reach US$18.8 billion by 2020 from US$11 billion in 2013, corresponding to a global consumption of acrylic acid of 8.2 million t/year.⁵ OPXBio said it was already able to produce 38 cents/lb bioacrylic acid using sucrose feedstock at 8 cents/lb. The company plans to have a demonstration plant with a capacity of 600,000 lb/year in 2013.⁶ A commercial plant with a capacity of 100 million lb/year is expected by 2015.

Succinic acid has received increasing attention for the production of new polyesters with good mechanical properties combined with full biodegradability. The main actors are probably Succinity GmbH (a Corbion Purac and BASF joint venture) with a commercial production facility in Montmeló, Spain, and Reverdia (a joint venture between Roquette and DMS) with a commercial facility in CassanoSpinola, Italy. Both plants have a production capacity of about 10,000 t/year. In Canada, BioAmber, in collaboration with Mitsui & Co., is constructing a biobased succinic acid production plant at a bioindustrial park owned by Lanxess in Sarnia, Ontario. The plant will initially have a production capacity of 17,000 t/year. The capacity will be increased by a further 35,000 t/year and at the same site it is scheduled to produce 23,000 t/year of 1,4-butanediol.⁷ The estimated cost for the construction of the plant is US$110 million. For the production of succinic acid BioAmber wants to use a variety of feedstock, for example, corn, wheat, cassava, rice, sugarcane, sugar beet, and forest waste. The feedstock will initially be transported by trucks to a wet mill facility, where it will be processed into glucose (dextrose) sirup and other valuable coproducts. The glucose will then be transported to the Sarnia plant, where it will be fermented to raw biobased succinic acid and later purified to crystalline biobased succinic acid before being shipped. Costs of succinic acid production strongly depend on the plant size. Some investigations indicate a cost of US$ 2.20 (€ 1.96) per kg at the 5000 t/year. Liu (2000) estimated that the price can be lowered to US$ 0.55 (€ 0.50) per kg at a production capacity of 75,000 t/year. At present, succinic acid is mostly produced by the chemical process from n-butane through maleic anhydride. The incidence of the raw material cost can be estimated in US$ 1.027 per kg. On the other hand, glucose is sold at a price of about $0.39 per kg, and assuming the succinic acid yield of 91% (w/w) on glucose, the raw material cost in the bioprocess is then $0.428 per kg (Song and Lee, 2006). Thus fermentative production of succinic acid from renewable resources can compete with the chemical process.

NatureWorks produces polylactic acid (PLA) at a plant with a designed capacity of 140,000 t/year in 2003; its current capacity is 150,000 t/year. Total investment costs have been US$ 300 million in hardware and US$450 million in R&D, process development, and technical support.⁸ The current price of PLA is higher than the average cost of polyolefin (0.4–0.7 € per kg).⁹ On an industrial scale, producers are seeking a target manufacturing cost of lactic acid monomer to less than US$ 0.8 per kg because the SP of PLA should decrease roughly by half from its present price of US$ 2.2 per kg.¹⁰ This example shows that biobased materials are perceived as greener and therefore bought despite a higher price.

The European Commission Directorate-General Energy has released a report titled “From the Sugar Platform to Biofuels and Biochemical,” where product prices and market volumes are depicted; some of these are shown in Table 3.3.

Selected chemicals produced at a large scale and briefly described above are shown in Table 3.4.