Multi-aspect evaluation of integrated forest-based biofuel
production pathways: Part 2. Economics, GHG Emissions,
Technology Maturity and Production Potentials
Yawer Jafria*, Elisabeth Wetterlunda, Marie Anhedenb1, Ida Kulanderb, Åsa Håkanssonc, Erik Furusjöa,d2
aLuleå University of Technology, Energy Engineering, Division of Energy Science 971 87 Luleå, Sweden. bRISE Research Institutes of Sweden, P. O Box 5604, 114 86 Stockholm, Sweden
cPreem AB 112 80 Stockholm, Sweden
dIVL Swedish Environmental Research Institute Ltd., P.O Box 210 60, 100 31 Stockholm, Sweden
*Corresponding Author (yawer.jafri@ltu.se)
ABSTRACT
Promoting the deployment of forest-based drop-in and high blend biofuels is considered strategically important in Sweden but many aspects of the overall performance of the foremost production technologies are as yet unexamined. This paper evaluates the technology maturity, profitability, investment requirements, GHG performance and Swedish biofuel production potential of six commercially interesting forest-based biofuel production pathways.
Significant heterogeneity in technology maturity was observed. Lack of technical demonstration in industrially representative scales renders the liquefaction-hydrotreatment route for drop-in biofuels less mature than its gasification-catalytic upgrading counterpart. It is a paradox that short-term priority being accorded to pathways with the lowest technology maturity. Nth-of-a-kind investments in (a) gasification-based methanol, (b) hydropyrolysis-based petrol/diesel, and (c) lignin depolymerization-based petrol/diesel were profitable for a range of plant sizes. The profitability of pulp mill-integrated small gasification units (< 100 MW) goes against the common perception of gasification being economically feasible only in large scales. New low-cost options for debottlenecking production at recovery boiler-limited kraft mills appear worth investigating. GHG emission reductions ranged from 66 to 95%; a penalty was incurred for high consumption of natural gas-based hydrogen. Swedish
1Current Address: Vattenfall AB, SE-169 92 Stockholm, Sweden
biofuel production potentials ranged from 4 to 27 TWh/y but a more feasible upper limit is 12-15 TWh/y.
KEYWORDS
Biofuels; gasification; lignin oil; pyrolysis oil; black liquor; forest residues.
1 Introduction
Commercial deployment of economically competitive biofuels that can meet stringent sustainability criteria is a central tenant in the most recent Swedish climate policy framework, which stipulates a reduction in domestic vehicle transport emissions of 70% by 2030 compared with 2010 levels [1–3]. This entails, among other measures, a switch from fossil fuels to biofuels produced from forest biomass feedstocks and industrial by-products like black liquor (BL) lignin, which have undergone extensive R&D in recent decades. Such fuels offer, in general, good performance both in terms of GHG footprint and from a resource utilization perspective [3–5]. Similarly, co-locating biofuel production with existing industries like pulp mills and oil refineries can produce synergies that improve energy and material utilization, thereby leading to better economic and resource efficiency performance, as has been studied in a number of previous studies, e.g. [6–8].
The development of drop-in biofuels such as renewable petrol and diesel, which can be distributed using existing fuel supply infrastructure, has been identified by the Swedish Energy Agency as a short-term priority. Substantial R&D support is currently being provided to this route [9] and drop-in fuels are also favored by new legislation featuring a GHG reduction mandate for both distributed diesel and petrol [10]. Reasons of energy and resource efficiency mean that the long-term focus is on high blend or pure biofuels such as methane, methanol, dimethyl ether (DME) and ligno-cellulosic ethanol [11].
Bearing in mind the above-mentioned strategic considerations and the relevance of forest-based feedstocks in a Swedish context, four technology tracks that offer a range of feasible production pathways are outlined in Figure 1.
Figure 1. Strategically relevant technology tracks for drop-in (petrol and diesel) and high blend (methanol) biofuel production from forest-based feedstocks in Sweden.
Despite the fact that industrially integrated forest-based biofuel production would appear to be attractive from a Swedish perspective, so far, these types of concepts have not been realized on any significant scale, but rather have been limited mainly to technology development activities. A number of issues, such as policy instability, “short-termism”, low predictability, lack of financing, and
knowledge deficiency among policy makers regarding e.g. the current biorefinery development status, have been identified as barriers to the advancement of biofuel initiatives [12–15]. In addition, market formation is also weak, both because of the need to compete with fossil fuels and because of internal competition from alternative use of raw materials [13] as well as from alternative investments in incumbent technologies[16]. Further uncertainties lie in investment estimates for emerging technologies [17,18]. Estimated optimal plant sizes for minimum production costs for mature technology (nth-of-a-kind, NOAK) are often several orders of magnitude larger than the currently largest demonstrated scales [19]. Ultimately, testing in commercially representative scales is required to mitigate the technical risk. The high cost of full-scale first-of-a-kind (FOAK) biorefinery
investments, however, poses a significant barrier to demonstration and commercialization [14,20]. Thus, large investments in e.g. pulp and paper-based biorefineries are blocked or postponed due to incompatibility with prevalent business models in the industry, resulting in what can be described as a systemic lock-in [16]. In order to mitigate financial risks related to the high initial capital costs of biorefinery investments, the trade-off between economies-of-scale and ability to mobilize sufficient capital can be explored to identify “breakeven points”, such as the minimum investment size for economic viability.
When assessing the performance of emerging biofuel production pathways, profitability and investment size are only two elements in the combination of aspects that govern overall viability. Some other relevant dimensions are technology readiness for commercial deployment, GHG
performance and biofuel production potential. The authors are unaware of publically available studies that systematically and consistently compare the cross-dimensional performance of commercially relevant forest-based drop-in and high-blend production pathways utilizing the same feedstocks. As of yet, the open literature appears to lack any evaluation of the relatively recent technology track based on the upgrading of kraft lignin to drop-in fuels.
This paper thus aims to fill the above-mentioned gaps by providing a comprehensive multi-aspect evaluation of six forest-based biofuel production pathways that are seen as commercially relevant in relatively near future. A particular focus is put on material and energy integration with existing kraft pulp mills and crude oil refineries. The following aspects are covered: (a) technology maturity, (b) economic viability, including profitability and investment requirement, (c) GHG performance and (d) Swedish biofuel production potential.
This paper is the second in a two-part series that comparatively evaluates the performance of forest-based drop-in and high-blend biofuel production routes, with the aim of facilitating consistent
cross-technology comparisons. The first part [21] provided an examination of product yields and energy efficiencies based on detailed material and energy balances, as well as more in-depth process descriptions. This part begins by briefly reiterating the studies process configurations. An explanation of the assessment methodologies follows, after which the results are presented and discussed. The paper ends with a series of conclusions centered on an overall performance evaluation.
2 Methods
The selection of biofuel production pathways outlined in Table 1 was based on a survey of relevant literature and commercial Research & Development & Demonstration initiatives in which alternatives were appraised on the basis of their technological maturity, commercial relevance and data availability. Energy balances and product yields were calculated based on detailed spreadsheet models of the integrated process configurations, based on best available data. Relevant process data were obtained primarily from the open literature and supplemented with information from technology developers and assumptions based on in-house knowledge. The procedure has been described in detail in Part 1 [21], where also the detailed product yields and energy balance calculations that form the basis for the evaluations detailed in this section can be found, as well as a thorough survey of pertinent literature regarding each evaluated pathway.
2.1 Process Configurations
Each pathway listed in Table 1 is identified by a number and an abbreviation, where the number indicates the technology track (1, 2, 3 and 4). In all but one (4-BMG) pathway, different parts of the biofuel production process are co-located and integrated with a pulp mill and/or a crude oil refinery. Process configurations and normalized energy balances are presented in Figures 2-5.
Table 1. Outline of evaluated biofuel production pathways
Pathway Description Feedstock Conversion route Integration
Pulp mill Refinery 1-MSL-HDO Hydrodeoxygenation of membrane-separated depolymerised lignin Kraft lignin Liquefaction-hydrotreatment
2-BLG Entrained-flow gasification of black liquor Kraft lignina Gasification-catalytic synthesis
3a-Pyr-HDO Fast pyrolysis and hydrodexoxygenation Forest residues Liquefaction-hydrotreatment
3b-Pyr-FCC Fast pyrolysis and fluidized catalytic cracking (FCC) Forest residues Liquefaction-hydrotreatment
3c-Hydropyr Catalytic hydropyrolysis and hydrotreatment Forest residues Liquefaction-hydrotreatment
4-BMG Fluidised-bed gasification Forest residues Gasification-catalytic synthesis
aKraft lignin is not separated from black liquor, all of which is processed in the gasifier, unlike in 1-MSL-HDO.
The liquefaction-hydrotreatment route for drop-in biofuel production is represented by hydrodeoxygenation of membrane-separated alkali-depolymerized lignin oil (LO) (1-MSL-HDO), and
by three pyrolysis-based alternatives that use forest residues as feedstock: fast pyrolysis and hydrodeoxygenation (3a-Pyr-HDO), fast pyrolysis and fluidized catalytic cracking (3b-Pyr-FCC), and hydropyrolysis (3c-Hydropyr). The gasification-catalytic upgrading route for the production of high blend biofuels like methanol is represented by two pathways: O2-blown pressurized entrained-flow
gasification of kraft black liquor (2-BLG) and steam/O2-blown pressurized fluidized-bed gasification of
forest residues (4-BMG), the latter being a standalone pathway. Further technical details can be found in Part 1 [21].
Figure 2. Energy balance for 1-MSL-HDO normalized to 1 MW HHV of biofuel product. Dotted and dashed boxes demarcate pulp mill and refinery boundaries, respectively. NCG stands for non-condensable gases, while HDO stands for hydrodeoxygenation.
Figure 3. Energy balance for 2-BLG normalized to 1 MW HHV of biofuel product. Dotted and dashed boxes
demarcate pulp mill and refinery boundaries, respectively. ASU stands for air separation unit, WGS stands for water gas shift, while AGR stands for acid gas removal.
Figure 4. Energy balance for (a) 3a-Pyr-HDO, (b) 3b-Pyr-FCC and (c) 3c-Hydropyr normalized to 1 MW HHV of biofuel product. Dotted and dashed boxes demarcate pulp mill and refinery boundaries, respectively. NCG stands for
non-condensable gases, HDO stands for hydrodeoxygenation, FCC stands for fluidized catalytic cracking, while SynSat (synergetic saturation) is a catalytic hydro-refining process.
Figure 5. Energy balance for 4-BMG normalized to 1 MW HHV of biofuel product. Dotted and dashed boxes demarcate pulp mill and refinery boundaries, respectively. ASU stands for air separation unit, WGS stands for water gas shift, while AGR stands for acid gas removal.
2.2 Technology Maturity
Process configurations in the studied pathways were broken down into 3-6 smaller steps. Integration into existing operations at the refinery and the pulp mill was treated as a step in its own right. The individual process steps were assigned a score based on the Technology Readiness Level (TRL) scale used by the European Commission [22]. Since the commission’s TRL scale is aimed at a general audience, the more topical TRL definitions from the US Department of Energy Clean Coal Program were used as an additional source of guidance to improve analysis granularity [23]. Technology maturity was assessed using two complementary approaches: (a) the weighted average approach, (b) the weakest link approach. In the weighted average approach, the assignment of weights was based on the importance and complexity of each step. The weakest link approach simply used the lowest TRL among the major steps (weight>0.2) in the pathway. See Appendix A in “Supplementary Material” for an overview of the relevant TRL definitions; TRL scores and assigned weights for individual process steps are provided in Appendix B.
2.3 Economic Evaluation
The economic evaluation was based around a four-step approach. In the first step, a specific investment margin (EUR/MWhbiofuel) was calculated under a baseline near-future (2018) and two
medium-term (2030) energy market scenarios for each of the pathways. This was done by dividing net annual operating income, defined as biofuel product revenues minus direct costs, by annual biofuel production volumes. Table 2 provides an overview of energy prices in the applied energy market scenarios. The near-future scenario was based on market data from 2016 and 2017. The two
medium-term scenarios, namely, 2030-450: a scenario for limiting atmospheric GHG concentrations to 450 ppm CO2 equivalents, and 2030-CP: a current policies scenario, were generated using the ENPAC tool, which
uses International Energy Agency forecasts for global fuel prices and projected policy instruments as inputs to calculate energy prices for large-volume customers. Details on the ENPAC model can be found elsewhere [24–26]. The prices of consumables such as catalysts and chemicals that did not vary between scenarios are listed under Appendix C in “Supplementary Material”. A sensitivity analysis was performed to assess the effects of selected cost and performance-related uncertainties on specific investment margin in the 2018 energy market scenario (see Table 3).
Table 2. Energy prices in the three studied energy market scenarios. General sales tax is not included. Near-future scenario Medium-term scenarios [ENPAC]
Scenario 2018 Scenario 2030-450i Scenario 2030-CPj
[EUR/MWh] Note [EUR/MWh] [EUR/MWh]
Forestry residues 20 a 41 41
Heating oil [E10/EO1] 39 b 56 83
Natural gas 35 c 58 49 Ethanol [T2] 85 d 115 122 Methanol 85 d 115 122 HVO diesel 75 d 115 123 Fossil diesel 37 e 85 95 Fossil petrol 49 e 79 85 Hydrogen 42 f 76 64 Electricity (purchased) 33 g 51 51 Electricity (sold) 30 h 51 51
a Average Swedish price in 2016 for wet woodchips from both coniferous and deciduous forest residues [27]. Includes transportation costs.
b Average Swedish price in week 37 of year 2016 for Preem AB. Excludes energy taxes and transportation costs.
c Average Swedish price in 2016 for industrial customers in the I4 category [3000 – 30000 MWh] [28]. Includes distribution costs and taxes.
d Average European price of ethanol [T2] and FAME, respectively, in 2016 [29]. FOB ARA: excluding insurance and transportation costs. For methanol, the price is assumed to be the same as for ethanol on an energy basis since a market for trading methanol as a transport fuel does not currently exist.
e Reference fossil fuel price in 2016 (Sweden) [29]. Includes production cost and gross margin but excludes all taxes. f Calculated as 3.564 times the price of natural gas produced using steam-methane reforming, on a mass basis. The multiple
represents the average of a low and a high estimate in the literature.
g Average Swedish price in the first half of 2017 for industry customers [70000 – 150000 MWh] [30]. Includes transmission costs and tax but excludes electricity certificates.
h Average hourly spot price in the first half of 2017 for South Central Sweden [31]. Excludes energy tax, electricity certificates mark up and transmission.
i Input data to ENPAC 2030-450: Crude oil (Brent): 85 USD/bbl.; Natural gas (EU import): 9 USD/Mbtu; Coal (OECD import): 57 USD/tonne; CO2 charge (EU ETS): 90 EUR/tonne CO2; support “diesel biofuels”: 28 EUR/MWh; support “petrol biofuels”: 37 EUR/MWh [32,33].
j Input data to ENPAC 2030-CP: Crude oil (Brent): 127 USD/bbl.; Natural gas (EU import): 11 USD/Mbtu; Coal (OECD import): 80 USD/tonne; CO2 charge (EU ETS): 27 EUR/tonne CO2; support “diesel biofuels”: 28 EUR/MWh; support “petrol biofuels”: 37 EUR/MWh [32,33].
Table 3. Factors studied in the sensitivity analysis of specific investment margin under the 2018 energy market scenario.
Parameter Range Pathway(s) Description
Hydrogen cost ±26%
1-MSL-HDO, 3a-Pyr-HDO, 3b-Pyr-FCC, 3c-Hydropyr
Range of autothermal reforming-based H2-to-natural gas cost ratios: 3.56±26% [34–36]; alternatively, the span can also be interpreted as the variation in natural gas price.
Catalyst cost -50%, +100% 1-MSL-HDO, 3a-Pyr-HDO, 3b-Pyr-FCC, 3c-Hydropyr
From Jones et al [37]. The upper limit represents the 2015 price and the lower limit the 2017 projection.
Chemicals cost +182
SEK/MWh 1-MSL-HDO
20% of the chemicals cost in the LBL-HTL case in Anheden et al. [38], see discussion under “Sensitivity Analysis”
Cold gas
efficiency -7.1 p.p. 2-BLG, 4-BMG
The same range is used for both 4-BMG and 2-BLG. Lower limit denotes hot gas filtration at 550 °C in 4-BMG, compared with the base case 850 °C demonstrated in pilot-scale[39]. HDO petrol
yield ±25% 1-MSL-HDO, 3a-Pyr-HDO
Evaluation of uncertainty in product distribution. Upper limit signifies the high petrol/low NCGs scenario and the lower limit the low petrol/high NCG case.
Biofuels yield ±10% 3c-Hydropyr Limits reflect uncertainty in biofuel product yields owing to limited technology demonstration in representative scales. FCC bottoms
distribution
-50% renewable
“bottoms” 3b-Pyr-FCC
Evaluation of the uncertainty in FCC product distribution owing to the absence of conclusive experimental evidence. The alternative case represents an increase in fossil diesel yield at the expense of green diesel.
In the second step, annualized specific investment costs (also in units of EUR/MWhbiofuel) were
calculated. This was done as follows: (a) total capital investment (TCI) estimates from the open literature [39–48] and, for the liquefaction step in 1-MSL-HDO, a technology vendor [38], were divided by the product of the specified biofuel production capacity and an on-stream factor of 0.95 to determined specific investment costs; (b) The specific investment costs were multiplied with an annuity factor of 0.16, based on an internal rate of return (IRR) of 15% and an estimated plant life of 20 years. The resulting annualized specific investment costs include an operations & maintenance (O&M) supplement, set at 4% of the specific investment cost.
In the third step, the specific investment margins available under the chosen energy market scenarios were compared with the annualized specific investments costs for each of the pathways. The effect of economies-of-scale on profitability was evaluated by scaling the TCIs with a uniform scale exponent for the entire estimate. Given the uncertainties embedded in the source estimates, the use of a more granular approach was not considered meaningful. As several key technologies are still under development, the evaluated pathways are likely to benefit from technology learning. Hence, a distinction was drawn between first-of-a-kind (FOAK) and nth-of-a-kind (NOAK) investment estimates.
In the fourth and final step, an analysis of breakeven plant sizes and required minimum investments was performed, based on the calculated specific investment margins and the scaled annualized TCIs. The breakeven point can be viewed as the smallest scale at which biofuel production for a given pathway may be economically viable, and is thus not synonymous with an optimal production scale. The accompanying total investment costs provide an idea of the investment required in absolute terms.
As mentioned earlier, it was assumed that existing equipment could be used to some extent in both the pulp mill and the refinery, thereby lowering the required investment. This led to a question that is important for the economic evaluation: what would be the cost of using this equipment? For the refinery, it was assumed that all process units and tanks are already being operated at full capacity prior to the introduction of LO or fast pyrolysis oil (FPO). Although it may be possible to debottleneck existing equipment to handle the increased volumes of petrol and diesel from the biogenic feedstock, the technical viability and cost of doing so is very case specific and cannot be examined generally. The approach used in order to solve this issue assumed a reduction in fossil petrol and diesel production by an amount equal in volume to the additional renewable diesel and petrol yielded by the added bio-oil. This enabled the estimation of a “(reduced) fossil production” cost as a substitute for the cost of modification, calculated using the gross margins for Swedish petrol (0.20 EUR/kg) and diesel (0.13 EUR/kg). For the pulp mill, the steam turbine was assumed to have a free capacity of 10% that can be used for steam produced in the integrated biofuel process units. For 2-BLG, required pulp mill modifications are more extensive and the relevant costs are included in the source TCI estimate [47]. All prices and cost estimates used in this study were adjusted for inflation, where appropriate, before being converted to 2017 EUR using exchange rates of 9.5 SEK/EUR, 8.5 SEK/USD and 10.5 SEK/GBP.
2.4 GHG Performance
The GHG footprints of the studied production pathways were estimated using a simplified
approach based on the Renewable Energy Directive (RED) guidelines [49]. The emission factors used to estimate the footprints (Table 4) using the energy balances given in Figures 2-5 were selected in accordance with literature findings [50] [51–53], which have shown that emissions associated with forest residues feedstock supply, electricity and H2 constitute the vast bulk of the climatic impact of
the studied technologies. An exception to the above exists when biomass is studied from a marginal perspective with coal-fired production as alternative use [53], which is not the case in this study. Since both the feedstock studied are classified as residues, direct or indirect land use change effects were excluded from the analysis.
The issue of allocation in life cycle analyses and GHG evaluations of biofuel production and biorefineries has been the subject of constant debate for more than a decade (see e.g. [54–57]). RED stipulates energy allocation (based on LHV) to attribute the emissions between biofuel and co-products. This allocation method prohibits the allocation of GHG emissions to co-produced heat, and thus do not, for instance, encourage the implementation of energy integrated biofuel production, even though this could improve the overall efficiency. In addition to this, RED is associated with several other methodological issues and inconsistencies, as shown by e.g. [58,59]. In contrast, the ISO 14044 standard [60]advocates the avoidance of allocation, whenever possible, in favor of the system
processes can be avoided by taking into account removed burdens from replaced conventional products and services.
In this study, the heat released during the HDO step in 1-MSL-HDO and 3a-Pyr-HDO was assumed to substitute heating oil at the refinery with an efficiency of 80%, which cannot be accounted for using the RED method. Correspondingly, for 3b-Pyr-FCC, the influence of co-feeding renewable FPO and fossil VGO on product distribution represents an additional complicating factor that not covered by the RED method. For these cases, the GHG footprint was also calculated using the principle of system expansion.
Table 4. Emission factors used as inputs to GHG footprint evaluations.
Input GHG footprint [gCO2 eq/MJ
LHV] Comments [Source]
Forest
residues 2.2
Emissions associated with “typical” technology and transport distance. All pathways except 1-MSL-HDO. [61]
Hydrogen 91.4 From steam reforming of natural gas. Liquefaction-hydrotreatment (1, 3a-c) pathways only. [62]
Electricity 13.1 Swedish electricity mix in accordance with Swedish Energy Agency recommendation. All pathways. [63].
Heating oil 80.0 Based on the reference value for heat production in the proposed RED II directive. 1-MSL-HDO, 3a-Pyr-HDO, 3b-Pyr-FCC only. [64] Petrol 93.5 Used in the system expansion for 3b-Pyr-FCC only. [64] Diesel 95.5 Used in the system expansion for 3b-Pyr-FCC only. [64]
2.5 Biofuel Production Potentials
The feedstock potentials and processing constraints presented in Table 5 were used in conjunction with biofuel-to-biomass conversion efficiencies to determine Swedish biofuel production potentials of the evaluated pathways. For each feedstock, upper and lower potential estimates were made in order to obtain overall production potentials as a range. In constructing the upper and lower estimates some key technology constraints, such as maximum possible lignin extraction rates and maximum blend-in potential in the refinery fluidized catalytic cracker (FCC), were also considered.
Table 5. Summary of estimated ranges for feedstock potentials and technology constraints (FCC), with brief descriptions of key assumptions.
Source Range
(TWh/y)
Upper estimate Lower estimate Pathway(s) Sources
Black liquor (BL)
24-54 All BL in all kraft pulp mills, with annual prod. increase of 1.3% (to 2030)
All BL in kraft pulp mills (current prod.) with recovery boilers (RB) built before 1995
2-BLG [65–67]
Black liquor
retentate 6.7-16
a 30% of max BL potential assumed extractable (all mills, increased prod.)
Individual max extraction potential, to ensure no negative effects on RB operation 1-MSL-HDO [66] Forest residues (branches and tops) 13-37 “Techno-ecological“ harvesting potential from the Swedish Forest Agency’s forest impacts assessment (SKA 15), reduced by estimated use in other sectors (2030)
Same as for max, but reduced by further ecological restrictions to avoid conflicts with environmental quality objectives 3a-Pyr-HDO, 3b-Pyr-FCC, 3c-Hydropyr, 4-BMG [66,68,69] FCC co-processing capacity 0.30-0.50 b 5 wt.% co-feed in existing FCC 3 wt% co-feed in existing FCC 3b-Pyr-FCC c
a Both estimates would require the import of additional biomass fuel to the pulp mill to account for the energy carried away in the extracted lignin. This was not included in the potential estimations. Instead, a decrease in electricity production was assumed.
b PO feed to the FCC. c [70,71].
3 Results and Discussion
3.1 Technology Maturity
Figure 6 shows the TRL scores of all evaluated pathways based on weighted average and weakest link approaches. In general, the pathways in the refinery-integrated liquefaction-hydrotreatment route for drop-in biofuels (pathways 1, 3a-c) were at a lower level of technology maturity than those in the gasification-catalytic synthesis route for high blend biofuels (pathways 2, 4). This is most clearly seen in the contrasting performance of the two routes on the weakest link metric. While four pathways, 2-BLG, 4-BMG, 3b-Pyr-FCC and 3a-Pyr-HDO, scored ˃ 6 on the weighted average metric, thereby placing them at the “Engineering/pilot-scale prototypical system demonstrated in relevant
environment” stage as per US DOE definition, only 4-BMG reached the same TRL level following the weakest link approach, although 2-BLG came close.
1-MSL-HDO was the least technologically mature pathway with a score of ≤ 4 on both metrics. Several important process steps, such as lignin depolymerisation, lignin oil purification,
hydrodeoxygenation and renewable fraction cracking, had low individual scores. This clearly indicates that 1-MSL-HDO still needs substantial R&D work before commercial deployment. There is,
however, a distinct possibility that technology development is at a more advanced stage than can be ascertained from the open literature and from what the developers have chosen to put out in the public domain.
For 3a-Pyr-HDO and 3b-Pyr-FCC, the initial upgrading at the refinery is the least mature step in the process chain. In 3a-Pyr-HDO, this is the two-step catalytic HDO unit, which was awarded a TRL score of 3, and where small-scale catalytic experiments in 30-400 mL reactors represent the state-of-the-art [37]. Temperature control in the exothermal deoxygenation process, which is crucial for scale up, is typically not considered in the small, lab-scale studies. It can be noted that a commercial actor, UOP, owns several relevant patents that cover HDO product recycling and mild initial HDO stabilization [72]. In addition, much of the research in this area has been done on FPO from stem wood, which generally has lower alkali content than forest residue-based FPO. On the other hand, TRL scores of other process steps in 3a-Pyr-HDO are fairly high, thereby leading to a higher weighted average score. Concerning 3b-Pyr-FCC, while co-feeding of fossil VGO and small quantities of FPO has been experimentally demonstrated in moderately representative scales, (~ 200 kg/hr) the documented accumulated on-stream time is relatively low. Accordingly, the step was awarded a score of 4-5, which indicates that full demonstration in pilot-scale is yet to be achieved. Note that 3b-Pyr-FCC is also being actively developed by commercial actors and much like 1-MSL-HDO, technology development may well have a reached a more advanced stage than has been documented in the literature.
Figure 6. TRL scores based on weighted average (left) and weakest link (right). Red < 4 ≤ orange < 5 ≤ light green < 6 ≤ dark green
3c-Hydropyr differs from other pathways in that refinery integration is limited to final upgrading to finished biofuel since deoxygenation is built into the hydropyrolysis process itself. The hydropyrolysis step is therefore the single most important step in the pathway, which, to the best of authors’ knowledge, has so far only been demonstrated in a small pilot scale unit of 50 kg/d at the GTI facility in Chicago, US [73,74]. The continuous operation of a 5 t/d demo plant currently being commissioned in Bangalore, India could, if successful, lead to a significant climb up the TRL scale [75]. For 2-BLG and 4-BMG, which belong to the high blend gasification-catalytic-synthesis route, it is the gasification step itself, including primary gas cleaning, which had the lowest TRL scores. Even so, both pressurized fluidized-bed gasification (4-BMG) and pressurized entrained-flow gasification (2-BLG) have been demonstrated
in meaningfully large scales (20 t/d) for extended periods of time (>3000 h and >26000 h, respectively) that are substantially greater than those achieved for any of the refinery integrated pathways (1, 3a-3c).
3.2 Economic Evaluation
3.2.1 Specific Investment Margins
Near-future Energy Market Scenario
Figure 7 gives a breakdown of direct costs, avoided costs, by-product revenues and reduced revenues for all pathways in the 2018 energy market scenario. Biofuel prices are shown as dotted lines. As the results show, the income from biofuel sales is enough to cover total production costs in all the pathways, and which therefore all have a positive specific investment margin.
For 2-BLG, which has the largest specific investment margin, expenditure is divided roughly equally between the cost of biomass needed as BL replacement for covering the mill steam
requirement and the cost of electricity imported to cover the increase in own use at the pulp mill. 4-BMG also has a large specific investment margin. For this pathway the cost of biomass was by some way the dominant cost component. On the other hand, costs were spread more evenly for 1-MSL-HDO, which does not require the import of additional biomass. There is a large cost saving associated with the replacement of heating oil at the refinery with the NCGs and heat released during the HDO step, which improves the specific investment margin significantly. Two other significant cost components for 1-MSL-HDO are H2 costs and catalyst costs, both of which have relatively large
uncertainties that are enhanced by the lack of pertinent data on costs (catalysts) and consumption (catalysts and hydrogen) in the literature.
The distribution of costs in 3a-Pyr-HDO is broadly similar to that in 1 MSL-HDO, with the added penalty of relatively high biomass use that reduces the specific investment margin considerably. 3b-Pyr-FCC has the smallest available specific investment margin and the pathway benefits greatly from the extra revenue generated by the shift in the fossil yields towards petrol and diesel as a result of co-processing (labelled as “fossil production” in Figure 7). This shift occurs at the expense of char and NCGs, which would otherwise be combusted to produce process steam for the refinery. To
compensate, more heating oil is used, which can also be seen clearly in the cost column. As discussed in detail in Part 1 of this article series [21], the resulting product distribution from the FCC must, however, be seen as highly uncertain. In general, the cost of reduced fossil production represents a notable proportion of the total costs for all pathways belonging to the liquefaction-hydrotreatment route (1, 3a-c). 3c-Hydropyr, however, benefits from needing only small amounts of H2 for upgrading
Figure 7. Breakdown of direct costs, avoided costs, by-product revenues and reduced revenues for all pathways in the baseline 2018 energy market scenario. The difference between biofuel price (dotted red lines) and total production costs represents the specific investment margin for covering capital expenditure and O&M costs. HVO is used as a stand-in for the price of renewable diesel. The price of ethanol [T2] is used as an analogue for methanol
(produced in 2-BLG and 4-BMG) and renewable petrol (produced in all pathways except 2-BLG and 4-BMG) prices.
Medium-Term Energy Market Scenarios
Cost and revenue breakdowns in the 2030-CP and 2030-450 energy market scenarios are provided in Figure 8. Compared to the baseline 2018 scenario, all pathways show larger specific investment margins in the 2030-CP scenario. The increased revenue stemming from higher projected biofuel prices (122 EUR/MWh) counteracts the projected increase in energy commodity costs (see Table 2). Meanwhile, the picture for the 2030-450 scenario is more mixed, although the specific investment margin is still positive for all the pathways.
Comparing the contrasting performance of 3a-Pyr-HDO and 3b-Pyr-FCC in the two medium-term scenarios helps illustrate some of the trade-offs involved. As discussed earlier, co-locating the HDO step at the refinery causes a significant reduction in heating oil use. The resulting economic benefit increases with increasing heating oil price. The associated savings offset the projected rise in H2 price,
although biomass cost is also a significant parameter and it plays an important role in setting the final specific investment margin. Conversely, 3b-Pyr-FCC must bear the cost of an increase in heating oil usage at the refinery. The margin is largest when, such as in the 2030-450 scenario, an increase in revenue from both fossil and renewable sales outstrips the increase in energy commodity costs.
-100 -80 -60 -40 -20 0 20 40 60 80 100 120 140 160 180 200 220 -100 -80 -60 -40 -20 0 20 40 60 80 100 120 140 160 180 200 220
1-MSL-HDO 2-BLG 3a-Pyr-HDO 3b-Pyr-FCC 3c-Hydropyr 4-BMG
E U R /M W h b io fu el E U R /M W h b io fu el
Electricty Hydrogen Chemicals Fossil Production Biomass Fuel oil Catalysts Total Prod. Costs
Methanol/Ethanol [T2] ~ 85 EUR/MWh HVO ~ 75 EUR/MWh
Figure 8. Breakdown of direct costs, avoided costs, by-product revenues and reduced revenues for all pathways in (a) the 2030-CP and (b) the 2030-450 medium-term energy market scenario. The difference between biofuel price (dotted red lines) and total production costs represents the specific investment margin for capital expenditure and O&M costs. HVO is used as a stand-in for the price of renewable diesel. The price of ethanol [T2] is used as an analogue for
-100 -80 -60 -40 -20 0 20 40 60 80 100 120 140 160 180 200 220 -100 -80 -60 -40 -20 0 20 40 60 80 100 120 140 160 180 200 220
1-MSL-HDO 2-BLG 3a-Pyr-HDO 3b-Pyr-FCC 3c-Hydropyr 4-BMG
E U R /M W h b io fu el E U R /M W h b io fu el
Electricty Hydrogen Chemicals Fossil Production Biomass Fuel oil Catalysts Total Prod. Costs
HVO/Methanol/Ethanol [T2] ~ 120 EUR/MWh (a) -100 -80 -60 -40 -20 0 20 40 60 80 100 120 140 160 180 200 220 -100 -80 -60 -40 -20 0 20 40 60 80 100 120 140 160 180 200 220
1-MSL-HDO 2-BLG 3a-Pyr-HDO 3b-Pyr-FCC 3c-Hydropyr 4-BMG
E U R /M W h b io fu el E U R /M W h b io fu el
Electricty Hydrogen Chemicals Fossil Production Biomass Fuel oil Catalysts Fuel oil
methanol (produced in 2-BLG and BMG) and renewable petrol (produced in all pathways except 2-BLG and 4-BMG) prices.
Sensitivity Analysis
The results of the sensitivity analysis (Figure 9) suggest that specific investment margins are in general more sensitive to changes in the costs of energy commodities and catalysts than to degradation in product yields. The two HDO-based pathways 1-MSL-HDO and 3a-Pyr-HDO require significant quantities of H2 for oxygen removal and are thus more exposed to fluctuations in H2 price. Another
notable cost component that concerns the HDO-based pathways is HDO catalyst costs, which have been projected to decrease by a factor of 2 annually [37]. If realized, this would boost future economic viability. However, in the absence of recent price documentation and given that the HDO step has not yet been demonstrated in industrially-relevant scales and on-stream times, HDO catalyst cost
represents one of the more uncertain parameters in this study. Similar issues also hinder a more accurate assessment of the current cost of hydropyrolysis catalyst for 3c-Hydropyr, which, however, makes up a much smaller proportion of the total cost, based on data likely out-of-date today.
In the studied 1-MSL-HDO configuration, chemical addition to the lignin depolymerization step is not considered. However, other similar processes require the addition of a capping agent, [76] which can be costly, as shown recently by Anheden et al. [38]. The investigation of chemical cost
uncertainty, motivated by the low TRL score of 1-MSL-HDO, revealed that if a cost escalation were to occur, it could have a notable impact on the available specific investment margin.
As noted above, the product distribution from the FCC is subject to significant uncertainty, and as the sensitivity analysis shows, a different approach for estimating product yields in 3b-Pyr-FCC, i.e. greater fossil diesel yield at the expense of green diesel, can reduce the available specific investment margin by more than half. Thus, not only is the additional revenue from the increased diesel and petrol yields vital to economic viability of this pathway, but the distribution of these yields between the fossil and the renewable fractions is also of tangible importance. Dispelling the surrounding uncertainty requires experimental demonstration in a representative scale over a suitably long time period.
Figure 9. Sensitivity analysis of specific investment margin for selected factors in the 2018 energy market scenario. Black bars represent the reference specific investment margin.
3.2.2. Investment Cost Estimates
Overviews of TCI estimates that were used to calculate annualized specific investment costs are given in Table 6, Table 7 and Table 8. Since the lignin depolymerization step in 1-MSL-HDO represents a relatively novel technology with low technology maturity (see Section 3.1), a suitable NOAK TCI estimates could not be found. The only TCI estimate available for the lignin depolymerization step was an early stage FOAK estimate put together by the technology developer [38]. This was used in conjunction with a NOAK estimate for the HDO step from the literature [40] to obtain an overall TCI estimate for the 1-MSL-HDO pathway. Since the cost of the LO production step is significantly greater than the cost of the HDO step, the overall estimate was treated as a first-of-a-kind (FOAK) estimate in this study.
0 10 20 30 40 50 60 70
Catalyst Cost Cold Gas Efficiency Hydrogen/Natural Gas Catalyst Cost Biofuel Yield Hydrogen/Natural Gas Catalyst Cost FCC Bottoms Distribution Hydrogen/Natural Gas Catalyst Cost HDO Petrol Yield Catalyst Cost Cold Gas Efficiency Hydrogen/Natural Gas Catalyst Cost Chemicals Cost HDO Petrol Yield
4-B M G 3c -H yd ro py r 3b -P yr -F C C 3a -P yr -H D O 2-B L G 1-M S L -H D O
2-BLG NOAK estimates are based on TCIs given as a function of plant size [10] in the source study, with the smallest plant size as the scaling reference. For 3a-Pyr-HDO, TCIs by Jones et al. [77] and Dutta et al. [42] were modified to exclude the cost of the H2 production, as an existing natural gas (NG)
reformer was assumed to be available at the refinery. Cost estimates for modification to existing FCC infrastructure required for co-processing fossil VGO and renewable FPO at the refinery were not available in the open literature. The extent of the modifications required, such as changes to the in-take, the holding tanks and the feed system, was adjudged to be relatively limited and the cost was estimated to be ~ 7.5 MEUR2017 for a 60 MW FPO feed. This estimate is tentative, but as can be noted from Table
7, the investment cost share of FPO production is significantly larger and dominates the combined TCI for 3b-Pyr-FCC.
Table 6. Estimates of Total Capital Investment for 1-MSL-HDO and 3a-Pyr-HDO
Pathway 1-MSL-HDO 3a-Pyr-HDO
Source Study Anheden et al[38], Tews et al. [40] Jones et al. [77]
Dutta et al. [42] Total Capital Investment
(from source) 400 MSEKdepolymerisation) + 2017 (lignin 6.4 MUSD2014
(hydrodeoxygenation)a
578 MUSD2011 480 MUSD2014
Total Capital Investment
(composite)b 54 MEUR2017 563 MEUR2017 444 MEUR2017
Biofuel Production [MW HHV] 34 275 270 Specific Capital Investment
[MEUR2017/MW]
1.59 2.05 1.64
Technology Learning FOAK/NOAK NOAK NOAK
a Calculated by multiplying the uninstalled costs for the HDO-HTL upgrading unit with the installation factor of another hydrotreating unit from the same source; bCombined total cost of all the process steps. 2017 prices were estimated
using inflation rates of 3.4% for the hydrodeoxygenation part of 1-MSL-HDO (based on Tews et al.) and 8.8% and 3.4% for the 3a-Pyr-HDO Jones et al. and Dutta et al. estimates, respectively.
Table 7. Estimates of Total Capital Investment for 3b-Pyr-FCC and 3c-Hydropyr
Pathway 3b-Pyr-FCCa 3c-Hydropyr
Source Study Benjaminsson et al [43]-modified Jones et al [77]-modified Tan et al [44] Meerman and Larson [45] Meerman and Larson [45] Total Capital Investment (from source) 301 MSEK2013+ 67 MSEK2017a 301 MUSD2011+ 67 MSEK2017a 264 MUSD2007 306 MUSD2014 612 MUSD2014 Total Capital Investment (composite)b 72 MEUR2017 313 MEUR2017 311 MEUR2017 283 MEUR2017 566 MEUR2017 Biofuel Production Capacity [MW HHV] 56 275 260 493 493 Specific capital investment [MEUR2017/MW] 1.28 1.14 1.07 0.57 1.15
Technology Learning NOAK/NOAK NOAK/NOAK NOAK NOAK FOAK
aDivided between production of fast pyrolysis oil and refinery modification; bCombined total cost of all the process steps. 2017 prices were estimated using inflation rates of 8.8% for Jones et al.-modified, 1.8% for Benjaminsson et al.-modified, 3.4% for the two Meerman and Larson et al. estimates and 18.1% for Tan et al.
Table 8. Estimates of Total Capital Investment for 2-BLG and 4-BMG
Pathway 2-BLG 4-BMG
Source Study Andersson et al.
[46] Andersson et al. [47] Hannula et al. [39] Udengaard et al. [48] Total Capital Investment (from
source)
345 MEUR2014 141 MEUR2014 328 MEUR2010 410 MEUR2014
Total Capital Investment
(composite)a 353 MEUR2017 144 MEUR2017 359 MEUR2017 424 MEUR2017
Biofuel Production Capacity
[MW HHV] 110 82 229 447
Specific capital investment [MEUR2017/MW]
3.19 1.76 1.57 0.95
Technology Learning FOAK NOAK NOAK NOAK
aCombined total cost of all the process steps. 2017 prices were estimated using inflation rates of 2.1% for the two Andersson et al. estimates, 9.4% for Hannula et al., and 3.4% for Udengaard et al.
It should be noted that when comparing independently prepared TCI estimates, including for the same technology, contingencies and indirect cost allocations may differ markedly because of methodological differences. Such potential discrepancies thus represent an additional source of uncertainty, which should be kept in mind when interpreting the results of the economic evaluation.
3.2.3. Comparison of Specific Investment Margins and Annualized Specific
Investment Costs
An assessment of profitability is presented in Figure 10, which offers side-by-side comparison of specific investment margins (Section 3.2.1) and annualized specific investment costs (Section 3.2.2) based on both NOAK (Figure 10a) and FOAK estimates (Figure 10b). The error bars show the effect of scale on annualized specific investment costs of plants in the range 25-600 MW. NOAK
investments for 4-BMG, 2-BLG and 3c-Hydropyr (Figure 10a) all appear to demonstrate relatively robust profitability in all three energy market scenarios. The relatively good performance of these pathways is aided partly by the economies-of-scale associated with the assumption of large production units in the source studies, in particular for 4-BMG and 3c-Hydropyr. Interestingly, the profitability of FOAK 1-MSL-HDO (Figure 10b) plants also appears to be quite good in all three energy market scenarios. The only available TCI estimate for this pathway assumes FOAK costs for the lignin depolymerisation step, thereby indicating that technology learning can improve the case further. Equally, and by analogy with the similarly technologically immature 3a-Pyr-HDO, the risk for cost escalation is considered significant. Better and more granular estimates grounded in detailed design are therefore needed for a more conclusive assessment. FOAK 2-BLG (Figure 10b) plants only appear profitable in large scales. Note that the NOAK 2-BLG (Figure 10a) estimate includes a so-called recovery boiler (RB) replacement credit given for the replacement of a RB approaching the end of its service life, which improves the profitability of this pathway.
The refinery integrated pathway 3a-Pyr-HDO (Figure 10a) is not profitable in the near-future energy market scenario. However, since the specific investment margin for 3a-Pyr-HDO is much
greater in the 2030-CP energy market scenario as discussed earlier, the medium-term profitability appears to be better for plants that are suitably large. Furthermore, the difference between the two estimates for which annualized specific investment costs were calculated is greater than can be explained by scale alone. Consequently, the annualized specific costs for these pathways are, perhaps unsurprisingly given the low technology maturity, very uncertain. Similarly, 3b-Pyr-FCC (Figure 10a) also does not appear to be economically viable in the near-future, although a larger plant could be economically viable in the medium-term. However, since the specific investment margin for this pathway is quite sensitive to uncertainties in fuel yields, the same uncertainties are also carried through to the profitability estimation.
0 20 40 60 80 100 120 A n de rs so n e t al [ 4 3] Jo n e s e t a l [ 7 4 ] D u tt a e t a l [ 38 ] B e nj a m in ss on e t a l [ 3 9 ] Jo n e s e t a l [ 3 7 ] T a n e t a l [ 40 ] M e e rm a n a n d L a rs o n [4 1 ] H a n nu la a n d K u rk e la [ 4 4 ] U d e ng a ar d e t a l [ 4 5 ] 2-BLG
3a-Pyr-HDO 3a-Pyr-HDO 3b-Pyr-FCC 3b-Pyr-FCC HydroPyr3c- HydroPyr3c- 4-BMG 4-BMG
[E U R /M W h b io fu el ]
Specific Investment Margin 2018 Specific Investment Margin 2030-CP Specific Investment Margin 2030-450 Annual. Specific Investment Cost
82 MW 270 MW 275 MW 260 MW MW493 229 MW MW447 57 MW 275 MW (a)
Figure 10. Comparison of (a) nth-of-a-kind (NOAK) and (b) first-of-a-kind (FOAK) Annualized Specific Investment Costs with available Specific Investment Margins in the near-future (2018) and medium-term (2030) energy market scenarios. Boldface type denotes the biofuel production capacity specified in the source study and error bars show the effect of scale on annualized specific investment costs of plants in the range 25-600 MW.
3.2.4 Breakeven Plant Sizes and Required Minimum Investments
Figure 11 shows the estimated breakeven plant sizes and required investments for FOAK and NOAK plants. Given the uncertainties in TCI estimates discussed earlier, the absolute numbers should be treated with a high degree of caution. Nevertheless, it is clear from Figure 11 that the economic cases for NOAK 2-BLG, 4-BMG and 3c-Hydropyr plants are convincing over a wide range of plant and investment sizes. Interestingly, even if the costs were to scale more steeply, relatively small 2-BLG units would still appear to be economically viable under NOAK assumptions. This represents an opportunity for pulp mills to de-bottleneck their existing RBs by using the surplus BL to produce biofuels for the transport market. Looking at the less promising of the two NOAK 3c-Hydropyr estimates, the breakeven plant size still appears to be ~ 100 MW, which backs the case for medium-sized facilities in areas where biomass may not be easily or cheaply available in large quantities. On the other hand, the case for 3a-Pyr-HDO looks less encouraging as neither of the estimates is economically viable at realistically achievable plant sizes.
The most promising FOAK case is that for 1-MSL-HDO. Breakeven plant size is ~30 MW, which translates to a relatively modest required minimum investment of ~10 MEUR. Similar to a NOAK 2-BLG facility, the economic viability of 1-MSL-HDO in the medium-scale 50-100 MW range can be an opportunity for pulp mills wishing to expand their revenue base without necessarily exposing themselves to the risk of very large investments. Conversely, the breakeven plant sizes and required
0 20 40 60 80 100 120 Anheden et al [34]/Tews et al [36]
Andersson et al [42] Meerman and Larson [41]
1a-MSL-HDO 2-BLG 3c-HydroPyr [E U R /M W h b io fu el ]
Specific Investment Margin 2018 Specific Investment Margin 2030-CP Specific Investment Margin 2030-450 Annual. Specific Investment Cost
34 MW 110 MW 493 MW (b) 18 0 EU R/ M W h bi of ue l
minimum investments for FOAK 2-BLG and 3c-Hydropyr plants are larger than what would typically be considered viable for a pioneer plant. This indicates that any FOAK plants based on these pathways would need financial support.
0 200 400 600 800 1000 1200 1400 1600 0 200 400 600 800 1000 1200 1400 1600 A n de rs so n e t al [ 4 3] Jo n e s e t a l [ 7 4 ] D u tt a e t a l [ 38 ] B e nj a m in ss on e t a l [ 3 9 ] Jo n e s e t a l [ 3 7 ] T a n e t a l [ 40 ] M e e rm a n a n d L a rs o n [4 1 ] H a n nu la a n d K u rk e la [ 4 4 ] U d e ng a ar d e t a l [ 4 5 ] 2-BLG
3a-Pyr-HDO 3a-Pyr-HDO 3b-Pyr-FCC 3b-Pyr-FCC HydroPyr3c- HydroPyr3c- 4-BMG 4-BMG
R e q u ir ed M in im u m In v e st m en t [M E U R ] B re a ke v e n P la n t S iz e [ M W ]
Breakeven Plant Size [MW] Required Minimum Investment [MEUR] (a) 38 00 M W 58 50 M EU R 20 25 M W 31 18 M EU R
Figure 11. Breakeven plant sizes and required minimum investments for (a) nth-of-a-kind (NOAK), (b) first-of-a-kind (FOAK) investment estimates.
3.3 GHG Performance
Simplified GHG emissions per unit of biofuel that were calculated using an allocation method consistent with RED are shown in Figure 12. Compared to fossil-based petrol and diesel references of 93.5 and 95.5 gCO2eq/MJ, respectively, all the pathways have lower emissions in the range 5-30
CO2eq/MJ. The estimated footprint of 5 g CO2eq/MJ for 2-BLG and 4-BMG represents a saving of
almost 95%. Similar savings are seen for 3b-Pyr-FCC (88%) and for 3c-Hydropyr (93%). Based on these numbers, the aforementioned pathways appear able to meet the stipulation of 70% GHG savings for new-built plants in the RED II proposal [64,78]. At greater risk are 1-MSL-HDO and 3a-Pyr-HDO, with savings of 66% and 70%, respectively. In both cases, the primary reason for the comparatively larger footprint is the use of NG for H2 production, although the reduction in renewable electricity
generation at the pulp mill because of lignin extraction (1-MSL-HDO) and feedstock supply (3a-Pyr-HDO) also play a role. The use of water electrolysis or of carbon capture in conjunction with NG reforming can reduce the footprint, albeit at, likely, a higher cost. It is also technically possible to convert a portion of the NCGs released during the HDO process to H2. However, as these are already
credited for replacing heating oil, the GHG footprint would not be greatly affected.
Without the mitigating effect of NCGs replacing heating oil at the refinery (negative violet bars in Figure 12), the GHG savings from 1-MSL-HDO and 3a-Pyr-HDO would be reduced to 60% and 50%, respectively. There is no scope in the RED for crediting the CO2 savings associated with the
0 50 100 150 200 250 0 50 100 150 200 250 300 350 400 450 500 Anheden et al [34]/Tews et
al [36] Andersson et al [42] Meerman and Larson [41]
1a-MSL-HDO 2-BLG 3c-HydroPyr R e q u ir ed M in im u m In v e st m en t [M E U R ] B re a ke v e n P la n t S iz e [ M W ]
Breakeven Plant Size [MW] Required Minimum Investment [MEUR] (b)
replacement of heating oil with heat released during HDO. Similarly, the RED-compliant allocation method seems unable to cope with the uncertainties surrounding fossil yield shifts in 3b-Pyr-FCC.
Figure 12. GHG footprint of the studied production pathways estimated using RED methodology.
Figure 13 shows the RED-based GHG footprints of 1-MSL-HDO, 3a-Pyr-HDO and 3b-Pyr-FCC side-by-side with those calculated using the principle of system expansion. In stark contrast to the RED method, system expansion shows that the two HDO-based pathways have negative net
emissions, which illustrates the impact of allocation method choice on GHG performance, as has been discussed in numerous previous studies [79,80]. 3b-Pyr-FCC also gives a negative footprint if the system is expanded to include fossil yield shifts. This is primarily because emissions from the resulting increase in heating oil usage are lower than if the additional petrol and diesel obtained are produced using fossil crude. Hence, the emission credit associated with increased petrol and diesel production (cyan and orange in Figure 13) is larger than the burden from the increased heating oil usage (purple in Figure 13). As discussed earlier, a great deal of uncertainty still surrounds the phenomenon of fossil yield shifts. Accordingly, the negative footprint for 3a-Pyr-FCC given by system expansion should be considered highly uncertain.
-20 -10 0 10 20 30 40 50 -20 -10 0 10 20 30 40 50 g CO 2e q/ M J LH V Heating oil Forest residues Hydrogen Electricity Total
Figure 13. Alternative GHG accounting methodologies for 1-MSL-HDO, 3a-Pyr-HDO and 3b-Pyr-FCC
3.4 Biofuel Production Potentials
Figure 14 shows the upper and lower biofuel production potential estimates for the evaluated pathways, which, with one exception, range from 4.4 TWh/y to 27 TWh/y. As is evident, the ranges are very broad, thereby highlighting the large uncertainties inherent in estimating feedstock potentials. Likely, a more feasible upper limit over the timeframe studied would be around 12-15 TWh/y of biofuels. A pairwise comparison over the feedstocks shows that, with the exception of 3c-Pyr-FCC, pathways reliant on forest residues all exhibit similar potentials, subject to the same limitation: feedstock availability. Biofuel production potentials are as low as 6-9 TWh/y when assuming the presence of harvesting restrictions aimed at avoiding conflict with environment quality objectives [69]. Conversely, the upper limit of 18-25 TWh/y assumes that the full technical potential of
harvesting residues from final felling and commercial thinning can be realized in practice. The lower limit for 3a-Pyr-HDO is mainly due to lower overall product yields (as noted in [21]).
3b-Pyr-FCC has the lowest production potential by a large margin. The estimates for this pathway take into account existing FCC capacity in Sweden and limit the share of FPO in the co-feed to 3-5 wt.%, as noted earlier in section 2.5 and in Part 1. Without these constraints, the potential is similar to that of 3a-Pyr-HDO. An upper ceiling on lignin extraction rates means that the technical potential of 1-MSL-HDO is not as large as that of 2-BLG. Where a key assumption for 2-BLG is the transition from RB to gasification-based chemical recovery, which means that the number of mills that can be converted constitutes the principal restriction, the corresponding key assumption for 1-MSL-HDO is the preservation of core pulp mill functionality (RB), which makes the technical extraction potential
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 100 150
Syst exp RED Syst exp RED Syst exp RED 1-MSL-HDO 3a-Pyr-HDO 3b-Pyr-FCC
g CO 2e q/ M J LH V Fossil diesel Fossil petrol Heating oil Forest residues Hydrogen Electricity Total
the principal restriction. The upper estimate for 1-MSL-HDO assumes that 30% of all lignin in the entire Swedish kraft BL throughput can be extracted. This corresponds to a feedstock potential of 16 TWh/y BL retentate. In comparison, the lower estimate of 6.7 TWh/y is obtained by using a mill-by-mill approach, where the prevention of negative effects on chemical recovery operations takes
precedence over lignin extraction [66]. Consequently, a figure in the mid-to-lower end (4-8 TWh/y) of the technical biofuel potential range may represent a more realistic estimate. Theoretically, the
feedstock potential of 2-BLG encompasses the entire Swedish kraft BL output. Since it is extremely unlikely that every pulp mill would replace its RB with a gasifier, a more reasonable estimate is ~ 12 TWh/y, which assumes that only pulp mills with old boilers needing imminent replacement would make the switch.
Figure 14. Swedish biofuel production potentials of evaluated pathways. The hatched column for 3b-Pyr-FCC shows the total potential assuming no FCC capacity restriction
4 Conclusions
Six commercially-relevant forest-based drop-in (short-term priority) and high blend (long-term priority) biofuel production pathways were comparatively evaluated across a number of
complementary aspects: (a) technology maturity, (b) economic viability, including profitability and investment requirement, (c) GHG performance and (d) Swedish biofuel production potential. An overall assessment of performance is given in Figure 15. Each pathway is assigned a rating of -, 0 or + relative to others for each of evaluated aspects with the exception of GHG performance, which is assessed in absolute terms against RED sustainability criteria.
Nth-of-a-kind (NOAK) investments in 1-MSL-HDO, 2-BLG, 3c-Hydropyr and 4-BLG appear economically feasible in both the near-future (2018) and in the medium-term (2030) perspective. The viability of less mature pathways such as 1-MSL-HDO in relatively small scales can be seen as an
argument in favor of accelerated commercial demonstration aimed at improving technology maturity for a relatively modest investment outlay. It is worth noting that the required minimum investment for 3b-Pyr-FCC is not simply a matter of economies-of-scale but is dictated by available capacity in the FCC unit of a crude oil refinery.
2-BLG is awarded a dual rating on the minimum required investment metric shown in Figure 15, as the underlying cost estimates assume that the gasification unit would entirely replace the RB (recovery boiler) at the mill. Based on the results presented above, significantly smaller gasification units may also be profitable. The profitability of relatively small NOAK plants (below 100 MW) for 2-BLG and 4-BMG goes against the common perception that gasification is only economically feasible in large scales. For 2-BLG this motivates further research into the feasibility of so-called booster gasifiers, which could, at a relatively low cost and with less technical risk than wholesale RB replacement, allow pulp mills to expand their pulping capacity. However, it should be noted that a simplified scaling approach was used here and that the BLG NOAK estimate benefits from an investment credit for an end-of-life RB replacement. By analogy, booster units may also benefit from a so-called debottlenecking credit.
Figure 15. Overall evaluation of the selected pathways. All ratings are relative except for GHG performance, which is assessed again RED (Renewable Energy Directive) criteria.
When applying the accounting method set out in the RED, all but two pathways (1-MSL-HDO and 3a-Pyr-HDO) are able to deliver GHG savings of 83-95% compared to a fossil reference. Both exceptions require substantial amounts of H2, which is assumed to be produced from NG by steam
reforming. Further research on renewable H2 alternatives appears to be well warranted. A more
holistic alternative to the RED accounting method may also be needed to assess the GHG performance of biofuel production that is partially co-located at a crude oil refinery. The differences between the two approaches can be dramatic, as evidenced by the negative footprint for 1-MSL-HDO, 3a-Pyr-HDO and 3b-Pyr-FCC under system expansion.
It is noteworthy that the pathways with the highest TRL score, with the exception of 3b-Pyr-FCC, also show the largest technical biofuel production potentials (10-25 TWh/y) in Sweden. Two pathways have lower potentials, which are attributed to restricted lignin supply (1-MSL-HDO) and a
combination of FCC capacity and co-processing limits (3a-Pyr-FCC). These potentials align well with expected demand and show that the pathways can make a very meaningful contribution to the
establishment of a fossil-independent transport fleet in Sweden. Note that future competition for forest residues from other CO2-intensive industrial sectors such as steel, cement and chemicals has not been
considered in the potential estimates. Quantifying the demand competition from these sectors is challenging as it is strongly influenced by future technological development and strategic policymaking choices.
In conclusion, it is a paradox that the pathways for drop-in biofuels being prioritized in the short-term are based on technologies with the lowest level of maturity. Several demonstration efforts, financially backed by both influential commercial actors and policymakers, are presently underway in Sweden. It remains to be seen whether sufficient progress can be made within the narrowing window of time consistent with a short-term perspective.
Acknowledgements
Funding Sources
This project was carried out within the collaborative research program Renewable Transportation Fuels and Systems [grant number 42406-1], financed by the Swedish Energy Agency and f3—the Swedish Knowledge Centre for Renewable Transportation Fuels. Economic support from Bio4Energy and from the Swedish Research Council FORMAS [dnr. 213-2014-184] is also gratefully
acknowledged. Notes
The authors declare no competing financial interests.
References
[1] SOU/Statens Offentliga Utredningar. SOU 2016:21 Ett klimatpolitiskt ramverk -
Delbetänkande av Miljömålsberedningen [A climate policy framework for Sweden]. Sweden: Swedish Government Official Reports; 2016.
[2] Government Offices of Sweden: Ministry of the Envoirnment & Energy. The Swedish climate policy framework 2018.
http://www.government.se/495f60/contentassets/883ae8e123bc4e42aa8d59296ebe0478/the-swedish-climate-policy-framework.pdf (accessed May 24, 2018).
[3] SOU/Statens Offentliga Utredningar. SOU 2016:47 En klimat- och luftvårdsstrategi för
Sverige, Del 1 (A climate and air pollution treatment strategy for Sweden – Part 1, in Swedish). Sweden: Swedish Government Official Reports; 2016.
[4] SOU/Statens Offentliga Utredningar. SOU 2013:84 Fossilfrihet på väg (A vehicle fleet
independent of fossil fuels, in Swedish). Sweden: Swedish Government Official Reports; 2013. [5] Börjesson P, Lundgren J, Ahlgren S, Nyström I. Dagens och framtidens hållbara biodrivmedel.
2013.
[6] Holmgren KM, Berntsson TS, Andersson E, Rydberg T. Comparison of integration options for gasification-based biofuel production systems – Economic and greenhouse gas emission implications. Energy 2016;111:272–94. doi:10.1016/j.energy.2016.05.059.
[7] Börjesson Hagberg M, Pettersson K, Ahlgren EO. Bioenergy futures in Sweden – Modeling integration scenarios for biofuel production. Energy 2016;109:1026–39.
doi:10.1016/J.ENERGY.2016.04.044.
[8] IEA Bioenergy. Biorefineries: adding value to the sustainable utilisation of biomass. IEA Bioenergy: task 42, 2009:01; 2009.
[9] NordicPaper. Production test factory for fossil free fuel is built in Bäckhammar 2016. https://www.nordic-paper.com/press/press-releases/production-test-factory-fossil-free-fuel-built-backhammar (accessed May 5, 2018).
[10] Svergies Riksdag. Lag (2017: 1201) om reduktion av växthusgasutsläpp genom inblandning av biodrivsmedel i bensin och dieselbränslen. Sweden: 2017.
[11] Energimyndigheten. ER2015:24 UP-Rapport Transportsystemet - Rådgivande underlag från utvecklingsplattform Transportsystemet till Energimyndighetens FOKUS-process. 2015. [12] Peck P, Grönkvist S, Hansson J, Lönnqvist T, Voytenko Y. Systemic constraints and drivers
for production of forest-derived transport biofuels in Sweden – Part A: Report. Gothenburg, Sweden: Report No 2016:09A, f3 The Swedish Knowledge Centre for Renewable
Transportation Fuels; 2016.
[13] Hellsmark H, Mossberg J, Söderholm P, Frishammar J. Innovation system strengths and weaknesses in progressing sustainable technology: the case of Swedish biorefinery development. J Clean Prod 2016;131:702–15. doi:10.1016/j.jclepro.2016.04.109. [14] Näyhä A, Pesonen H-L. Strategic change in the forest industry towards the biorefining
business. Technol Forecast Soc Change 2014;81:259–71. doi:http://dx.doi.org/10.1016/j.techfore.2013.04.014.
[15] de Assis CA, Gonzalez R, Kelley S, Jameel H, Bilek T, Daystar J, et al. Risk management consideration in the bioeconomy. Biofuels, Bioprod Biorefining 2017;11:549–66.
doi:10.1002/bbb.1765.
role of micro-level factors for technological innovation system weaknesses. Technol Anal Strateg Manag 2017;29:500–13. doi:10.1080/09537325.2016.1249838.
[17] Petter R, Tyner WE. Technoeconomic and Policy Analysis for Corn Stover Biofuels. ISRN Econ 2014;2014:1–13. doi:10.1155/2014/515898.
[18] Hannula I. Making Sense of Cost and Performance Estimates for Thermochemical Biofuel Plants. tcbiomass2017 Tech. Progr., Chicago: Gas Technology Institute; 2017.
[19] Wright M, Brown RC. Establishing the optimal sizes of different kinds of biorefineries. Biofuels, Bioprod Biorefining 2007;1:191–200. doi:10.1002/bbb.25.
[20] Bauer F, Coenen L, Hansen T, McCormick K, Palgan YV. Technological innovation systems for biorefineries: a review of the literature. Biofuels, Bioprod Biorefining 2017;11:534–48. doi:10.1002/bbb.1767.
[21] Jafri Y, Wetterlund E, Anheden M, Kulander I. Multi-aspect evaluation of integrated forest-based biofuel production pathways : Part 1 . Product yields & energetic performance 2019;166. doi:10.1016/j.energy.2018.10.008.
[22] European Commission. Technology readiness levels (TRL). Horiz 2020 – Work Program 2014-2015 2014.
https://ec.europa.eu/research/participants/data/ref/h2020/wp/2014_2015/annexes/h2020-wp1415-annex-g-trl_en.pdf (accessed May 24, 2018).
[23] US DOE. 2014 Technology readiness assessment. 2015. doi:10.1080/08956308.2010.11657640.
[24] Axelsson E, Harvey S, Berntsson T. A tool for creating energy market scenarios for evaluation of investments in energy intensive industry. Energy 2009;34:2069–74.
doi:10.1016/j.energy.2008.08.017.
[25] Axelsson E, Harvey S. Scenarios for assessing profitability and carbon balances of energy investments in industry. Göteborg: Pathways to sustainable european energy systems – AGS, the Alliance for Global Sustainability; 2010.
[26] Axelsson E, Pettersson K. Energy price and carbon balances scenarios tool (ENPAC) - a summary of recent updates. Göteborg: Chalmers; 2014.
[27] Energimyndigheten. Trädbränsle- och torvpriser, nr 3/2017 2017.
http://www.energimyndigheten.se/nyhetsarkiv/2017/tradbransle--och-torvpriser-for-tredje-kvartalet-2017/ (accessed May 24, 2018).
[28] SCB/Energimyndigheten. Priser på naturgas för industrikunder 2007– 2018.
