Correspondence to: Aloijsius G. J. van der Ham, Sustainable Process Technology, Faculty of Science and Technology, University of Twente, P.O. Box 217, 7500 AE Enschede, the Netherlands. E-mail: a.g.j.vanderham@utwente.nl
Furfural to FDCA: systematic process
design and techno-economic evaluation
Guus H C Dubbink , Thomas R J Geverink, Bas Haar, Harald W Koets, Abhay Kumar, Henk van den Berg, Aloijsius G J van der Ham, Sustainable Process Technology, Faculty of Science andTechnology, University of Twente, Enschede, The Netherlands
Jean-Paul Lange , Sustainable Process Technology, Faculty of Science and Technology, University of
Twente, Enschede, The Netherlands; Shell Technology Center Amsterdam, Amsterdam, The Netherlands
Received November 07 2020; Revised January 15 2021; Accepted February 01 2021; View online at Wiley Online Library (wileyonlinelibrary.com);
DOI: 10.1002/bbb.2204; Biofuels. Bioprod. Bioref. (2021)
Abstract: 2,5-Furan dicarboxylic acid (FDCA) is a promising intermediate for producing polyethylene furan dicarboxylate, an alternative to polyethylene terephthalate that combines a significantly lower greenhouse gas footprint with better mechanical and gas barrier properties. This work presents a process design and techno-economic evaluation for producing FDCA from non-edible biomass via the oxidation of furfural to furoate salt, and subsequent carboxylation to furandicarboxylate salt. Major technical uncertainties are associated with the possible polymerization of furfural in the oxidation step and the state of salt phase in the carboxylation step. Based on the furfural market price of $1400/ton this process requires a minimum selling price of 2000 ± 500 $/ton FDCA. To compete with purified terephthalic acid (PTA), it requires a premium of 100% for better performance and sustainability, or a combination of much cheaper furfural and a much lower capital expenditures (CAPEX). © 2021 Society of Chemical Industry and John Wiley & Sons, Ltd
Supporting information may be found in the online version of this article.
Keywords: 2,5-furandicarboxylic acid (FDCA); furfural; techno-economic analysis; process design; electrodialysis
Introduction
P
olyethylene furan dicarboxylate (PEF) is a polymer alternative to polyethylene terephthalate (PET), which is produced using a different monomer, 2,5-furan dicarboxylic acid (FDCA). The chemical structures of both polymers are very comparable, although the linearity of the p-xylene segment in PET induces increased mobility compared to the furan ring in PEF, which lowers the mechanical and gas barrier properties from PET compared to PEF.1 Besides, PEF shows strong sustainability potentialby having a 45–55% lower greenhouse gas (GHG) footprint, when calculated from cradle to grave.2 These characteristics have energized the industry to develop and demonstrate the manufacture and use of PEF over the past decade.
The most advanced route that is being piloted to produce FDCA converts edible fructose to hydroxy-methyl furfural (HMF) or an ether of HMF and oxidizes it to FDCA.3 This HMF route is treated as a benchmark process, as there are plans for a 5 kt/a pilot plant operational in 2023.3 Recently Kanan et al.4 proposed an alternative route based on furfural (FF) with high conversion and selectivity at laboratory scale.
2
Table 1. Economic values used in this article.
Value Units Source
Furfural intake 50 Kt/a —
Furfural price* 1400 $/ton 14
Oxidation catalyst price** 43 000 $/ton 15–17
Cs2CO3 price 3000 $/ton 18
Membrane price 282 $/m2 19
Electricity price 0.04 $/kWh 8
Steam price 12.1 $/ton 8
Operator wage 62 000 $/operator/year 8
No. operators 5 Operators/shift 8
Number of shifts 5 8
Operation hours 8000 h/year
Location factor (EU) 1.19 8
Lang factor 4.6 8 Depreciation 10 Years 8 CEPCI value (2018) 610 20 Minimal return on investment (ROI) 10.0 % 8
Income tax rate 30.0 % 8
*Price rationale in SI.
**1 wt% AuPd/mg(OH)2, assuming 90% AuPd recovery.
Furfural is produced by acid digestion of inedible lignocellulosic biomass. Its global market represented approximately 490 kt/a in 2016 and is still growing with an annual estimated growth rate of 4.3%.5 This market is relatively small compared to the enormous worldwide demand for PET (15 000 kt/a in 2009).2 Nevertheless, the furfural route promises the advantage of using a more sustainable feedstock, namely lignocellulose, and offers a better overall yield.4
Kanan’s route consists of (i) oxidizing furfural to furoate salt (FA) and (ii) reacting a furoate salt with CO2 to furandicarboxylate salt in the presence of Cs2CO3 at 200 °C and 8 bar pressure, as shown in Fig. 1. These two conversion steps are then complemented with a work-up section to recover and purify FDCA (after acidification) and to regenerate the base for recycling.
This paper presents a complete conceptual process design, including reactor design, heat integration and an environmental analysis. Major equipment has been sized, construction material has been selected, and this information has been used for a high-level economic estimate of capital expenditure (CAPEX), operational expenditure (OPEX), and minimum selling price for the product, including sensitivity analysis. Finally, key areas for technical improvements are identified and discussed. This research was performed by a team of five master’s students as part of a course on process plant design. The paper will focus on the final design and its techno-economic analysis. For further information on the process alternatives considered or on design details, the reader is invited to consider the supplementary information or the original design report of 190 pages, available upon request to the corresponding author (A.G.J. van der Ham).
Methodology
The process was designed following the systematic approach proposed by Douglas7 and further elaborated since.8 Various conceptual process designs were developed containing functional blocks. The least promising conceptual process designs were rejected, and the more promising designs were worked out in more detail. Finally, the most promising process
Figure 1. Overview of the reaction steps for the conversion of furfural to FDCA using the carboxylation route.6
design was designed in detail, comprising reactor engineering, separation technologies, and process utilities. This resulted in a process flow diagram, which was modeled with Aspen Plus V10 to gather data regarding mass and energy flows. This data were used as input to provide for the information required for heat integration, equipment design, the techno-economic viability, and the sustainability of the process.
In Aspen, the ELECNRTL (Electrolyte-NTRL) property method was used for predicting the properties of the electrolyte system including salt precipitation. ELECNRTL calculates molecular interactions identical to NRTL and can therefore also make use of that database with molecular interaction parameters. The solubility of the gasses H2, O2, and CO2 was calculated using Henry’s law. The Redlich–Kwong equation of state was applied to model the vapor-phase properties. Components not present in the
Aspen Plus database were added as pseudo-components and their properties were retrieved from the literature9,10 and complemented with estimation software, ICAS V19. Nevertheless, the behavior of certain electrolyte species could not be modeled with confidence (e.g. furoate salt, FDCA salt and cesium bicarbonate) and were therefore replaced by their respective protonated counterparts. As phase change behavior differs strongly for electrolytes in comparison with their respective protonated counterparts, forced separation blocks were applied. These separations were constructed with spreadsheet and MATLAB (R2019A) calculations based on data from Aspen and literature combined. The data retrieved from the Aspen Plus model were used as input to determine the energy flows for all streams that required a temperature change or a phase transition. The heat management tool for process integration FI2EPI11 was used to determine the composite curves and design the heat-exchanger network.
The process was designed for a relatively small furfural intake of 50 kt/a, which represents about 10% of the furfural market size in 2016.5 Main equipment was sized based on mass or energy flows. The purchased equipment costs were
estimated with cost curves.12,13 These costs estimates were converted into inside battery limit (ISBL) cost by applying the installation (Lang), location and inflation correction (CEPCI) factors presented in Table 1. The resulting ISBL cost was then converted to total capital investment by adding contributions for outside battery limits (OSBL), design and engineering, contingency, land, plant start up, and working capital, as presented in the work of Seider et al.8 and reported in the supplementary information. The annual production cost based on variable and fixed operational expenditure is supplemented with 10% of the total investment cost to account for depreciation over a 10-year period. The minimum selling price is set such that the profit before tax, the difference between annual production cost, and annual sales income, afford 30% tax and still deliver a profit after tax, which provides 10% return on the total investment cost (ROI) to pay the bank interest and still deliver some benefits. This calculation should correspond to a net present value around zero. Parameters used for raw materials and energy prices, and calculations details (operational costs, total investment cost and depreciation) are also summarized in Table 1.
Figure 2. Simplified process flow diagram for converting furfural to FDCA. Pumps, compressors, and heat exchangers are not shown. Table 2 provides a condensed stream table; a detailed one is available in the supplementary information.
4
Process design
The final process design (Fig. 2) consists of two reaction segments, namely furfural oxidation to furoate salt (R1) and its subsequent carboxylation to FDCA salt (R2). It also comprises a product work-up section to recover FDCA (after acidification and precipitation/filtration in R3), to remove by-products by L/L extraction, and to regenerate the Cs2CO3 and K2CO3 bases by means of bipolar membrane electrodialysis (BME) and CO2 addition (R4). The conceptual design is based on the carboxylation reaction proposed by Kanan et al. and the line-up roughly sketched in their article.4 The price of raw furfural requires high recoveries, high selectivities, or high value by-products for the product to be competitive. Process parameters for the reaction segments are summarized in Table 3.
Oxidation of furfural to FA
−In the oxidation reactor (R-1) the furfural is converted to furoate salt by reacting with oxygen in the presence of a
base in water. Minor side products are also formed, e.g. furfuryl alcohol made through a Cannizzaro reaction,21 and polymerized furfural. Furfural is diluted by a large amount of water due to its low solubility of 83 g L−1.22 Choices were made between various alternatives proposed in the literature; these are summarized below and discussed in more depth in the original design report. Pure oxygen was chosen over air to avoid contaminants, thereby accepting a modest cost penalty for purchasing the oxygen. An AuPd bimetallic catalyst (AuPd/Mg(OH)2)23 was selected rather than a homogeneous metal-free catalyst.24 The heterogeneous gold catalyst offers the advantage of a smaller separation sequence and operates without co-solvent.
A slurry tank reactor (R-1) was chosen over a packed bed reactor, due to the relatively short catalyst lifetime21 and lower risks of possible hot spots. The reactor was modeled using mass transfer correlations of Krishna25 and the reaction kinetics from Douthwaite.21 As the kinetics are determined in highly diluted batch experiments, the reactor productivity was extrapolated by assuming linear dependency on catalyst,
Table 2. Stream table of main streams, M are both the potassium and cesium ions combined.
Feed R-1 out Oxygen R-2 in Carbon R-3 in Product By-product Acid Metalsalt
Stream no. 1 2 5 9 13 15 16 19 26 36
Temp (°C) 30 30 30 260 260 25 25 80 25 30
Press. (bar) 3 3 3 8 8 1 1 1 1 3
Mass flow (ton/day)
Furfural 136 5.0 — — — — — — — — Residu 0.36 0.36 — 0.36 — 0.36 0.36 — — — Water 0.05 1634 — — — — — 21 630 1634 FOH — 2.9 — — — — — — — — O2 — 11 32 — — — — — — — Furoic acid — — — — — — — 6.0 — — M_FA — 325 — 325 — 13 — — — — M2CO3 — 22 — 478 — 254 — — — 456 CO2 — — — — 24 — — — — — M2_FDCA — — — — — 498 — — — — M_AA — — — — — 18 — — — — M2_MA — — — — — 34 — — — — FDCA — — — — — — 185 — — — Char — — — — — 0.56 0.56 — — — Ethyl acetate — — — — — — — 0.11 — — H2SO4 — — — — — — — — 628 — AA — — — — — — — 5.6 — — MA — — — — — — — 9.7 — — MHCO3 — 259 — — — — — — — —
Table 3. Design characteristics of the oxidation and carboxylation reactors.
Parameter Unit Oxidation
(R-1) Carboxylation (R-2) Temperature K 303 533 Pressure Bar 3 8 Reactor type — G/L/S Slurry reactor G/L Bubble reactor Conversion % 96.3 96.0 Selectivity % 97.8 92.7
Mass inflow Ton/day 2260 1330
Reactant inflow conc. Wt% 6.0 in water 24.4 in metal salt Catalyst — AuPd/mg(OH)2 — Catalyst loading m3/m3 0.01 — Reactor volume m3 98 19 Reactor productivity Kg product/m3 Reactor/h 138 409 Heat of reaction kJ/Mol −365 −144 Adiabatic temperature rise K 79 62 Cooling duty MW 4.8 0.5
furfural, and base concentration. The definitive design characteristics are available in Table 3.
Downstream of the oxidation reactor, a phase separation releases most of the oxygen (F-1). The effluent is fed to an evaporator (E-1) to reduce the water content and remove the residual furfuryl alcohol prior to feeding to the carboxylation reactor (R-2).
Carboxylation of FA
−to FDCA
2−The feed of the carboxylation reactor (R-2) consists of furoate salt and CO2 dissolved in a K2CO3/Cs2CO3 salt mixture with a molar ratio of Cs+/K+ 2:1 for optimal conversion and selectivity.6 The salt at a temperature above 200 °C enables4 the deprotonation of the 5th position of furan-2-carboxylate, which can react with CO2 to FDCA2−. Side-products include acetate, malonate, and char. We assume no effect of the presence of the heavy contaminants (polymerized furfural and residue) already in the system. Experimental work26 showed the salt mixture to be liquid at 260 °C. Based on this result and literature,6 the reactor was chosen to operate at 260 °C and 8 bar of pressure and designed as a bubble column, with high solubility of CO2 in the molten salt
predicted by Aspen simulations. Design and performance data on the carboxylation reactor are summarized in Table 3.
The unconverted CO2 is recycled back to the reactor after being dried by distillation under elevated pressure (D-1). Adsorption is a competitive alternative to distillation due to the expected low water concentration. Both water removal units (E-1 and D-1) are crucial for the process because water spoils the carboxylation reaction.4,6 The ratio of water content to yield reduction was not known and needs to be experimentally determined.
FDCA and by-product recovery (R-3 and
LL extraction)
The products leaving the carboxylation reactor (R-2) are the Cs+ and K+ salt of CH
3COO− (acetate), CH2(COO)2− (malonate), and FDCA2− as well as some heavy contaminants. As FDCA is insoluble in water in acidic form,27,28 we chose to recover it by acidification (R-3) followed by filtration. Among various acids, H2SO4 was selected by balancing the pKa, the bond dissociation enthalpy, price, corrosiveness, and safety of various candidates (see supplementary information). Sulfuric acid will also protonate acetate and malonate salts but the resulting acids are water soluble and will not precipitate with FDCA. By applying rotary drum filters (not shown in Fig. 2), which are well developed in industry, large throughputs can be achieved. In this way solid FDCA is obtained and separated from its by-products. The filtrate is fed to the liquid–liquid extraction column (LLE), in which the by-products, acetic acid and malonic acid, are recovered from water to prevent accumulation in the process. Ethyl acetate was selected as an extraction solvent based on selectivity, safety, price, and ease of recovery (see supplementary information). Ethyl acetate is then recovered by evaporation (E-2) and recycled back to the extraction column. The acetic and malonic acid by-products are assumed to be a waste stream.
Acid/base regeneration (bipolar
membrane electrodialysis)
The Cs/K base mixture and sulfuric acid are regenerated by bipolar membrane electrodialysis (BME) of the aqueous stream leaving the liquid–liquid extraction column. The BME converts Cs2SO4/K2SO4 in separated sulfuric acid and CsOH/KOH streams by first separating cations (Cs+, K+) and anions (SO42−) and combining them with OH− and H+ ions created by water dissociation.29 The regenerated sulfuric acid can then be recycled to the acidification reactor (R-3). The Cs/K hydroxides are converted to Cs/K carbonates upon reacting with CO2 gas (R-4) and are partly recycled back to the oxidation reactor (R-1) to ensure proper dissolution of
6
furfural in the feed (with 10% excess of solvent). The rest of the carbonate is dried in E-1 and then recycled to the carboxylation reactor (R-2).
Heat integration
In the overall process, 70.4 MW of cooling energy and 72.2 MW of heating energy is required. Heat integration saved 20.6 MW of both heating and cooling energy for a
∆Tmin of 5 °C. The largest reduction was achieved by using stream 8 to heat up stream 7 and unit E-2. The bottleneck for further integration presents itself in the 48.7 MW needed to evaporate water (E-1), which is located around the pinch temperature of 100 °C. A multi-stage evaporator could lead to a more efficient heat integration.
Major technical uncertainties
The present design has been built on a number of uncertainties that need to be recognized and evaluated experimentally at a later stage. The major ones are discussed here.Direct oxidation reaction
The polymerization of furfural in reactor R-1 largely depends on the nature of the base and reaction temperature (source: interview with M. Douthwaite).21,30 An increase in furfural polymer will lead to lower product purity and / or additional separation steps. The extent of polymerization should be validated by experiments and eventually reduced by lowering the temperature or reducing the concentration / strength of the base. Further research could also focus on identifying a solvent alternative that offers a higher solubility to furfural and a lower heat of vaporization.
Contaminants in product stream
Other solid particles and soluble impurities couldcontaminate the product stream of the rotary drum filters. We assumed that this level of contamination is low enough not to affect the polymerization of FDCA to PEF. All by-products should be tested to assess the level of formation of undesired solids. Again, additional separation steps will be detrimental to the economics. Experimental work26 has shown that there is already a high level of char deposition, which confirms the importance of this issue.
Molten phase
Measures need likely to be identified to ensure that the salt remains at molten state in all places and at all time
in the carboxylation reactor section. The occurrence of a different phase could affect the reactor design. Moreover, transporting an eventual salt after water evaporation could be a technical issue. Finally, the corrosiveness of the molten salt should be investigated to allow selection of the material for construction.
Bipolar membrane electrodialysis
The choice of BME membrane, electrode, and spacers and the resulting performance of the BME should be demonstrated. For instance, the recovery rate of 99.9% assumed here may not be technical and / or economically achievable.
Economic evaluation
Economic benchmark
The minimum FDCA selling price, which we will discuss in the next section, will need to be compared with representative benchmark prices. We tried to develop two independent ones, namely the market price of purified terephthalic acid (PTA), with which FDCA would be competing, and the manufacturing cost of the most advanced route to FDCA, which is based on fructose dehydration to HMF (or an HMF derivative) and its subsequent oxidation to FDCA.
As a first reference, we set a target selling price of FDCA at 1200 $/ton, based on a PTA market price of 960 $/ton (average price at the time crude oil was 100 $/bbl)31 and an arbitrary premium of 25% to account for FDCA’s better gas barrier properties. A larger premium can also be envisaged, e.g. to monetize the bio-based origin and low carbon footprint of FDCA over fossil-based PTA. In 2020, a survey by the Boston Consulting Group over 15 000 consumers stated that 74% would accept paying a premium for green packaging.32 The extent of such a ‘green’ premium will be investigated and covered in a sensitivity analysis (Fig. 4).
When looking for the second benchmark, we found no reliable economic analysis of the HMF route. Triebl et al.33 reported an FDCA minimum selling price of ∼$2500/t, assuming HMF priced at $1000–1400/t and a plant capacity of ∼3 kt/a, i.e. 1/10 of the capacity assumed here.33 A few objections can be formulated against using this analysis as benchmark here. Most importantly, HMF has no reliable market price because it is not a commercial product yet. Nevertheless, the HMF price used here accounts for 39% of the minimum sale price; a determining impact. To serve as a benchmark, such analysis should start from glucose or high-fructose syrup. Beyond that, the evaluation could be rescaled to the desired scale of 50 kt/a for proper comparison. But it may also need further corrections, e.g. including OSBL cost,
which seem to have been omitted, and revisiting the catalyst cost that has been considered as CAPEX and accounts for an unusual share (19%) of the sale price.
Another publication, by Dessbesell et al.,34 seemed to offer a more solid basis as it starts from glucose and high-fructose syrup as feedstock. In contrast with HMF, glucose is commercially available and therefore has published market prices. The minimum selling price is reported to amount to 1800 $/ton FDCA and is dominated by the raw material cost (77%). The CAPEX contribution is estimated at 1.6%, i.e. 13 M$ for a 50 kt/a FDCA plant, a very low figure for a three-step process for producing and isolating fructose, dehydrating it to HMF and then oxidizing it to FDCA. In fact, the total CAPEX is much smaller that the CAPEX estimated by Triebl
et al.33 for the last step of the process, the conversion of HMF to FDCA, after correction for scale and catalyst cost.
Figure 3. Breakdown of FDCA selling price.
-80% -70% -60% -50% -40% -30% -20% -10% +0% +10%
Change in Estimated Value
1000 1200 1400 1600 1800 2000 2200
Sales price [$/ ton FDCA]
Furfural price CAPEX Utilities
PTA +25% properties premium + 25% green premium
PTA +25% properties premium
Figure 4. Sensitivity analysis on main cost drivers. PTA price is added as indicator and is not affected by the x-axis.
In short, we could not develop a reliable benchmark for the HMF route and, therefore, invite the community to rerun the analysis with more details and in a way that is more traceable.
Minimum selling price
Based on the economic analysis that we will discuss below, the present process appears to require a minimum selling price of 2000 ± 500 $/ton (inherent 25% uncertainty, cost sheet reported in the supplementary information), 52% of which is accounted for by the raw material cost, mostly by the furfural cost, which was priced at $1400/t (Fig. 3). The remaining variable costs are utilities (steam, electricity) and they contribute 10% to the selling price. The various contributions that are estimated as factors of the CAPEX, i.e. the annual fixed cost, the return on investment and the depreciation and royalties, account together for 38% of the required minimum selling price. The CAPEX was estimated to be 120 ± 30 million $ or 2400 $ per annual ton of production. The ISBL and OSBL account each for 32% of the total CAPEX. The remaining 1/3 consists of design, contingency, land, and working capital. Obviously, the required minimum selling price of ∼$2000/t is well above the benchmark of $1200/t derived from PTA market price with premium. In fact, the cost of furfural alone has already reached the required minimum selling price (Fig. 3). Cost reduction is likely needed in all elements to make the furfural route economically competitive.
A first cost-reduction opportunity is in the cost of the raw material, furfural (Fig. 4). The furfural price has to drop from 1400 $/ton to below 300 $/ton to reach the target selling price of 1200 $/ton. Such a low cost is not very plausible but a severe reduction, e.g. to $700/t, is not impossible. Today’s furfural production is based on small-scale processes with
8
low yield of valuable product (∼10 wt% on biomass intake) and high energy consumption (30 t steam/t furfural).35 It therefore consumes most of the biomass to raise the steam needed for the process. More recently, however, numerous process concepts based on biomass fractionation are being developed to combine furfural production with the production of valuable glucose for further valorization.35 Beyond delivering more valuable products, these processes also promise to be more energy efficient and operate at larger scale. Hence, one can reasonably expect significant drop in furfural price in the coming decades.
Moreover, the furfural is bought at high purity (99% furfural) but is afterwards diluted with water. The use of cheaper wet furfural is expected to decrease the furfural cost further. Increasing the furfural efficiency is always beneficial but will not have a dramatic impact here. An opportunity to increase FDCA yield is to separate and recycle unconverted furfural after the oxidation reactor (R-1), e.g. using adsorption at high recovery rates up to 93%.36,37 This approach could increase the overall FDCA yield from 83.8% to 86.8% and, consequently, lower the selling price by approximately 2%.
A second area for cost reduction is the CAPEX, as it covers 35% of the FDCA price. The carboxylation and the base regeneration sections are the largest contributors to the ISBL (Fig. 5). The main cost driver in the base regeneration is the BME unit. Aiming for less sharp recovery (now assumed 99.9%) and compensating losses with acid makeup may bring significant savings. But an aggressive CAPEX reduction by 50% alone will still not allow the process to meet the target selling price. But it is getting there if it can be combined with
Figure 5. Breakdown of capital inside battery limits (ISBL) expenditures.
a furfural price of $700/t, a price that is challenging but not impossible, as discussed above.
Conclusion
A process design for converting furfural at 50 kt/a intake to FDCA has been developed and evaluated. The design consists of a slurry-type oxidation reactor followed by a carboxylation bubble column reactor and a work-up section to recover FDCA by precipitation and to regenerate its chemicals (acid and base) by bipolar membrane electrodialysis. This process delivers an FDCA yield of 83 mol%, based on furfural intake, and requires 52 MW of heating energy (24 MJ/kg FDCA).
Critical technical uncertainties have been identified and require experimental validation. These uncertainties include the polymerization of furfural in the oxidation reactor, the presence of solid impurities in the FDCA product stream, the behavior of the molten salt in the carboxylation reactor and the large-scale membrane electrodialysis.
Considering a furfural price of $1400/ton, such a process requires a minimum selling price for FDCA of 2000 ± 500 $/ ton, which is 67% over our target price of 1200 $/ton based on the PTA market price and 25% premium. The largest contribution towards the high selling price is the cost of the raw material, furfural. A cost cut of 75% is required to meet the target FDCA selling price for a profitable design.
References
1. Lacerda T, Gandini A. Polymers from renewable resources. Kirk-Othmer Encyclopedia of Chemical Technology 2016. 2. Eerhart AJ, Faaij AP and Patel MK, Replacing fossil based
PET with biobased PEF; process analysis, energy and GHG balance. Energy Environ Sci 5:6407–6422 (2012).
3. Avantium N.V. Avantium to build FDCA flagship plant at Chemie Park Delfzijl, Netherlands. 2020; https://www. avantium.com/2020/avantium-to-build-fdca-flagship-plant-at-chemie-park-delfzijl-netherlands/.
4. Banerjee A, Dick G, Yoshino T and Kanan MW, Carbon dioxide utilization via carbonate-promoted C-H carboxylation. Nature 531:215–219 (2016).
5. Huang K, Brentzel ZJ, Barnett KJ, Dumesic JA, Huber GW and Maravelias CT, Conversion of furfural to 1,5-pentanediol: process synthesis and analysis. ACS Sustain Chem Eng 5:4699–4706 (2017).
6. Dick GR, Frankhouser AD, Banerjee A and Kanan MW, A scalable carboxylation route to furan-2,5-dicarboxylic acid.
Green Chem 19:2966–2972 (2017).
7. Douglas JM, Conceptual Design of Chemical Processes. McGraw-Hill, New York, p. 601 (1988).
8. Seider WD, Seader JD and Lewin DR, Product and Process
Design Principles, 4th edn. Wiley, Hoboken, p. 764 (2009).
9. Nist Webbook. Carbon dioxide, cesium carbonate, DMSO, furfural, furoin, furoic acid, hydrogenchloride, oxygen, potasium carbonate, water. (2019). https://webbook.nist.gov/chemistry/. 10. ChemSpider - 2,5-Furandicarboxylic acid,
11. Pereira PM, Fernandes MC, Matos HA and Nunes CP, FI2EPI: a heat management tool for process integration. Appl Therm
Eng 114:523–536 (2017).
12. Loh HP, Lyons J, Charles W, White, I II. Process Equipment Cost Estimation, Final Report; 2002.
13. Matches, Matches engineering to chemical energy manufacturing metallurgical industries. 2020; https://www. matche.com/.
14. DalinYebo, Furfural market prices (2014). https://dalinyebo. com/furfural-market-prices-2014/.
15. APMEX, Palladium Price (2019). https://www.apmex.com/ spotprices/palladium-price.
16. APMEX, Gold Prices (2019). https://www.apmex.com/ spotprices/gold-price.
17. Alibaba, Activated Carbon Price (2019). https://www. alibaba.com/product-detail/Granular-activated-carbon-price_62157540325.html.
18. Alibaba, Cesium Carbonate (2019). https://www.alibaba.com/ product-detail/Cesium-Carbonate_1255836447.html. 19. Palmen L. Protocols for electrodialysis membrane selection,
membrane fouling and design; 2017.
20. Chemical Engineering - Chemical Engineering essentials for the global chemical processing industries (CPI). https://www. chemengonline.com/.
21. Douthwaite M, Huang X, Iqbal S, Miedziak PJ., Brett GL, Kondrat S, Sankar M, Bethell D, Hutchings GJ. Supplementary Information: The Controlled Catalytic Oxidation of Furfural to Furoic acid using AuPd/Mg(OH)2; 2017.
22. GESTIS Substance database, 2-Furaldehyde. http://gestis-n. itrust.de/nxt/gateway.dll/gestis_en/025010.xml.
23. Paul S, Wojcieszak R, Dumeignil F, Santarelli F, Reghizzi B, Roselli A, Cavani F, Dick GR, Kanan MW. A New Two-Step Green Process for 2,5-Furandicarboxylic Acid Production from Furfural, Oxygen and Carbon Dioxide.
24. Gupta NK, Fukuoka A and Nakajima K, Metal-free and selective oxidation of furfural to Furoic acid with an N-heterocyclic carbene catalyst. ACS Sustain Chem Eng 6:3434–3442 (2018).
25. Maretto C and Krishna R, Modelling of a bubble column slurry reactor for Fischer–Tropsch synthesis. Catal Today 52:279–289 (2002).
26. Wiegerinck H. An Alternative Sustainable Route to
Polyethylene 2,5 Furandicarboxylate: The Molten Carbonate Promoted Carboxylation of Furoic Acid. Ph.D. thesis, University of Twente, 2020.
27. Hussain Motagamwala A, Won W, Sener C, Alonso DM, Maravelias CT, Dumesic JA. Toward biomass-derived renewable plastics: Production of 2,5-furandicarboxylic acid from fructose. 2018.
28. Zhang Y, Guo X, Tang P, Xu J. Solubility of
2,5-furandicarboxylic acid in eight pure solvents and two binary solvent systems at 313.15–363.15 K; 2018.
29. Wilhelm FG. Bipolar Membrane Electrodialysis, Ph.D. thesis; 2001.
30. Douthwaite M, Huang X, Iqbal S, Miedziak PJ, Brett GL, Kondrat SA et al., The controlled catalytic oxidation of furfural to furoic acid using AuPd/mg(OH)2. Catal Sci Technol
7:5284–5293 (2017).
31. Straathof AJJ and Bampouli A, Potential of commodity chemicals to become bio based according to maximum yields and petrochemical prices. Biofuels Bioprod Biorefin 11:798–810 (2017).
32. Trivium Packaging, 2020 global buying green report. 2020; https://triviumpackaging.com/sustainability/2020BuyingGre enReport.pdf.
33. Triebl C, Nikolakis V and Ierapetritou M, Simulation and economic analysis of 5-hydroxymethylfurfural conversion to 2,5-furandicarboxylic acid. Comput Chem Eng 52:26–34 (2013). 34. Dessbesell L, Souzanchi S, Venkateswara Rao KT,
Carrillo AA, Bekker D, Hall KA et al., Production of 2,5 furandicarboxylic acid (FDCA) from starch, glucose, or high fructose corn syrup: techno economic analysis. Biofuels
Bioprod Biorefin 13:1234–1245 (2019).
35. Lange JP, van der Heide E, van Buijtenen J and Price R, Furfural - a promising platform for lignocellulosic biofuels.
ChemSusChem 5:150–166 (2012).
36. Cuevas M, Quero SM, Hodaifa G, L’opez AJ and S’anchez S, Furfural removal from liquid effluents by adsorption onto commercial activated carbon in a batch heterogeneous reactor. Ecol Eng 68:241–250 (2014).
37. Sulaymon AH and Ahmed KW, Competitive adsorption of furfural and phenolic compounds onto activated carbon in fixed bed column. Environ Sci Technol 42:392–397 (2008).
Guus H.C. Dubbink
Guus H.C. Dubbink has a bachelor’s degree in industrial engineering and a
cum laude master’s degree in chemical
engineering from the University of Twente, the Netherlands. He is currently working as chemical industry business analyst for an investment firm.
Bas Haar
Bas Haar is a student at the University of Twente, Enschede. He has a bachelor’s degree in advanced technology and is pursuing a master’s degree in chemical process engineering. He is currently working on his thesis, optimizing the electrocatalytic synthesis of urea from carbon dioxide and nitrate.
Harald W. Koets
Harald W. Koets has a bachelor’s degree in chemical engineering from Hanzehogeschool University of Applied Sciences, and is currently a master graduating student in chemical engineering at the University of Twente, the Netherlands.
Thomas R.J. Geverink
Thomas R.J. Geverink is finalizing his two master’s degrees at the University of Twente. He is combining a master’s degree in chemical engineering with a master’s in chemistry education. Thomas is passionate about
sustainable chemistry, which he wishes to pursue during the remainder of his career.
10
Abhay Kumar
Abhay Kumar has a bachelor’s and master’s degree in chemical engineering from the Birla Institute of Technology and the Indian Institute of Technology, both in India. He is currently working as manager under the future leadership programme of a chemical manufacturing company in India.
Henk van den Berg
Henk van den Berg is professor emeritus at the University of Twente. He has an industrial background and is still involved in teaching and research. His special interests are the innovation needed for chemical processes given the challenges related to sustainability, raw materials, and pollution.
Jean-Paul Lange
Jean-Paul Lange is principal scientist at Shell and professor at the University of Twente, the Netherlands, where he explores novel catalytic processes for producing fuels and chemicals from natural gas, oil, biomass, and waste plastic. He is co-author of 100 patents, 60 papers, and seven book chapters.
Aloijsius G.J. van der Ham
Aloijsius G.J. van der Ham is assistant professor in the research group Sustainable Process Technology at the University of Twente, where he obtained his PhD in chemical engineering. His main expertise is research and education in the field of process design, process simulation, and evaluation.
