Bridging sustainability and intensified flow processing within process design for sustainable future factories

3.1 Case study A: analysis of catalyst plate reuse options

Due to the process design activities within CoPIRIDE focusing on flow conditions, reactor and especially catalyst development activities aiming for, e.g., long-life time and reusability, play an important role in the context of sustainability. In particular, within the project consortium, catalyst plates consisting of Ru/alumina catalytic coatings on stainless steel microstructured substrates were of interest. In order to identify the most eco-efficient option of catalyst plate treatment after use, a simplified cost analysis and LCA for the catalytic coating of microchannels, including the removal of the coating, and the recycling of the plates as well as of the catalytic compounds, was done (more detailed information can be found in Kressirer et al. [10]).

Here, the results of an eco-efficiency analysis based on experimental scale investigations are briefly summarized for catalytic plates with a size of 150×225×2 mm and with channel depth of 400 μm and diameter of 600 μm. After thermal pretreatment of fresh stainless steel plates, the catalyst was coated on the walls of the microchannels. Prior to use, the coated plates were dried and calcined. After use, there were several possibilities of treatment:

• Disposal of the plates and preparation of new ones.

• Removal of the coating and use of a new coating slurry on the recycled stainless steel plates.

• Removal of coating and recycling of stainless steel plates as well as of the catalytic components.

The analysis is based on the full capacity and under consideration of the life time of each device used within this process. The removal of the coating was done in four different ways, of which two were suited to recycle the catalytic components (Table 2). The term “base case” was used for the disposal of plates after use without removal of the coating or the recycling of the plates.

Figure 3 shows an outranking of the environmental efficiency of the base case scenario versus Methods 1–4, taking into account typical environmental impact categories evaluated within the LCIA step (see Table 1 for explanation of the abbreviations). As an example, the base case was found to be preferential in the case of GWP and abiotic resource depletion potential, whereas Method 2 should be preferred concerning eutrophication potential, acidification potential, photochemical ozone creation potential concerning low NO_x, ozone depletion potential and terrestrial eco toxicity potential. These outranking results of environmental efficiency provided the basis for the eco-efficiency analysis depicted in Figure 4.

Figure 3

Multi-criteria outranking of base case versus Methods 1–4 concerning environmental efficiency, taking into account several life cycle impact assessment (LCIA) categories.

Figure 4

Multi-criteria outranking of different methods for catalyst plate treatment after use according to their environmental and cost efficiency.

Considering both, environmental aspects and costs, Methods 2 and 3 for catalyst plate treatment after use were found to be the most sustainable options. This outcome influenced further catalyst and reactor design activities within the CoPIRIDE project in a substantial manner.

3.2 Case study B: reaction under harsh conditions from batch to flow – epoxidation of SBO

Case study B deals with process intensification of the epoxidation of SBO. Epoxidized SBO (ESBO) is bio-based, non-toxic, non-corrosive and biodegradable and can substitute phthalates, legally prohibited in some applications [34]. Thus, it is the most common oleochemical used for PVC compounding. Due to its excellent market and green product expectations, the improvement of its current production process has been chosen as one development case within the CoPIRIDE project.

In industry, the reaction is usually performed in pulse-fed-batch or fed-batch reactors at a reaction temperature between 60°C and 70°C, gradually adding the oxidant reagents to the oil. ESBO is produced by reacting hydrogen peroxide with the double bonds contained in the SBO oil molecules under the presence of formic acid (Scheme 1) and a mineral acid, such as sulfuric or phosphoric acid. Due to its highly exothermic nature, the reaction requires a sufficient control addressing corresponding safety issues. To overcome the limitations of the time and energy consuming batch process, the synthesis of ESBO was planned to be transferred to a flow process within the project [6, 35] in order to increase the productivity as well as its eco-efficiency [9].

Scheme 1

Soybean oil epoxidation: sequence of the occurring reactions placed at the double bounds of oil compounds.

Within the initial development stage, process simulation was found to be a powerful tool for screening of different alternatives [6]. It was apparent that due to the interfacial mass transfer and heat management, the reaction can be speeded up by utilizing microreaction technology and possibly also by Novel Process Windows (e.g., high temperatures).

As a first step, the influence of different parameter variations was screened by a set-up of 13 partly theoretical scenarios. They were based on information of the existing production process under batch conditions, results from process simulation, pilot plant engineering and expert knowledge of the partners involved in the process development (more detailed information can be found in Kralisch et al. [9]). Based on these early investigations, the following key criteria for the development of a sustainable epoxidation process were defined: i) maximizing the yield of the epoxidized product, ii) minimizing the amount of H₂O₂ and HCOOH, and iii) best suited reaction temperature range: 100°C–120°C (in cases where digestion times of 1 h and more are foreseen in the process) or synthesis at high temperatures (150°C–180°C) and residence times within seconds, if technically feasible. The conjoint implementation of all optimization ideas in the pilot process can lead, for example, to a reduction of environmental impacts up to 12% in terms of the LCIA category GWP, or 11% in terms of human toxicity potential. The results also showed that the environmental benefits of the synthesis under Novel Process Windows and short residence times will strongly depend on the resulting ESBO yield and on the hydrogen peroxide decomposition rate [6]. In a case where the amount of hydrogen peroxide would have to be increased due to a significant decomposition at high temperatures, the human toxicity potential would escalate, resulting in a worse environmental balance compared to process alternatives at lower reaction temperatures [9]. An optimum between temperature, reaction time, yield, decomposition rate and technical feasibility had to be found here.

In the next step, the existing industrial fed-batch process for SBO epoxidation at Mythen S.p.a., Italy, covering an annual production volume of 18,500 t, was taken as benchmark for a sustainability evaluation of the new process under development at pilot scale (planned annual production capacity: 2,500 t). For this, detailed information about the production and post-processing steps, including hydrogen peroxide decomposition, neutralization, and separation, were provided by the company. Two pilot scale process alternatives (pilot process, base case: continuous flow at elevated temperature and short residence time; pilot process, best case: pilot process, base case and reduced material and energy demand) were taken into account including again, the environmental burden related to all educts, solvents and auxiliaries, energy supply, synthesis, work-up and treatment of wastes. The process parameters of these SBO processing alternatives are summarized in Table 3. A more detailed insight into the optimization task and support for process design decisions could be gained from this. Figure 5 shows an outranking of the environmental efficiency of the process alternatives under consideration for pilot scale, compared to the industrial fed-batch process against different LCIA criteria. A clear preference for an intensified flow process operating at elevated temperatures with reduced H₂O₂ and HCOOH excess and heat recovery (pilot process, best case) was found among the three candidates considered.

Figure 5

Multi-criteria ranking of epoxidized soybean oil (ESBO) process alternatives (industrial fed-batch, 18,500 t/a; pilot process, base case, 2,500 t/a; pilot process, best case, 2,500 t/a) concerning environmental efficiency (saving potential) taking into account several LCIA categories.

In addition to the comparative environmental assessment, an evaluation of the future production costs of utilizing the new flow chemistry concept as alternative to the established fed-batch production plant was also performed (see Figures 6–9). The ESBO production costs are composed of variable and fixed cost components. Investment costs, calculatory depreciation and imputed interests considered as fixed costs were estimated from average values of typical equipment costs, in the case of the established production process. In Figure 6, the composition of typical investment costs taken into account for the hypothetical installation of a new fed-batch process is illustrated. They sum up to overall investment costs of €1.5 million.

Figure 6

Composition of theoretical investment costs for an epoxidized soybean oil (ESBO) fed-batch production plant.

Figure 7

Composition of production costs for the epoxidized soybean oil (ESBO) fed-batch production plant.

Figure 8

Composition of investment costs for the epoxidized soybean oil (ESBO) pilot plant, best case scenario.

Figure 9

Composition of future production costs for the epoxidized soybean oil (ESBO) pilot plant, best case scenario.

The composition of the overall production costs for the ESBO fed-batch production plant is shown in Figure 7. The considered cost categories for the variable costs are those for materials, energy, personnel and disposal. These cost figures are obtained, e.g., by assessing the quantitative consumption of raw materials, auxiliaries and supplies with a corresponding price. Energy costs (in the form of supply costs for electrical current and natural gas) are those costs that are incurred in the production of the product and are required for the operation of the plant. Personnel costs consist of salaries, wages and social costs. Also, waste arises during the production of the product. It was disposed of separately, for which additional costs incurred. The variable costs were dominated by the costs of the feedstock SBO (€0.45/l) followed by the costs for hydrogen peroxide and formic acid.

Based on the production costs of the existing production process, a prognosis for the impact on potential cost savings of a scale-up of the new processing concept to pilot scale was done. In Figures 8 and 9, the composition of the costs estimated for installing and operating is shown. Here, an annual outcome of 2,500 t per year is assumed. The calculation is based on the same market prices for materials, energy, natural gas and disposal as well as typical personnel costs for a production in Italy; in this case, it is also based on a detailed plan of investment costs.

As shown in Figure 9, the share of personnel costs on the overall production costs is higher for the pilot process, due to the significantly lower annual production capacity. However, in the case of increasing production capacities, the personnel costs can be expected to significantly fall below the personnel costs of the reference fed-batch process [9]. In combination with further cost savings gained by an increase of the annual production capacity of the microreaction plant concept, advantages of the overall production costs against the fed-batch process are likely. In summary, the results of this comparative evaluation pointed out that the overall ESBO production costs are dominated by variable costs independent from batch or continuous processing. In the case of the new pilot plant, the estimated share of variable costs is 97%.

Combining the results of the economic and the environmental assessments of the three case scenarios discussed, an eco-efficiency ranking of the alternative process options was created (see Figure 10). The overall cost efficiencies were interrelated to the environmental efficiencies (see Figure 5). In all cases, the same yield of 90% high-quality ESBO (oxirane number>6) was assumed. From the current state of knowledge, the most favorable scenario would be a continuously operating production process with a reduced amount of required substances and heat recovery (pilot process, best case scenario). Figure 10 reveals that this result was found due to improvements in environmental efficiency. Thus, further process intensification activities should mainly concentrate on a minimization of the production costs, in order to come out with an innovation (scenarios located in the up-right part of the eco-portfolio). As stated before, this can be supported by further capacity increase. But the main key criteria will be the most efficient use of the materials, demanding for: i) maximized yield and ii) minimized oxidative media excess. In order to meet these critical targets, recently promising progress has been made in the development of a highly active, selective heterogeneous catalyst, suited for continuous processing [36].

Figure 10

Eco-efficiency ranking of three discontinuously or continuously running epoxidized soybean oil (ESBO) process scenarios.

3.3 Case study C: reaction under harsh conditions from batch to flow – process design for intensified biodiesel production

The third case study deals with the development of a novel, intensified biodiesel production alternative [11, 37–40]. Biodiesel, a biodegradable, nontoxic diesel fuel substitute, has shown a significant added value compared to petrodiesel, because of its higher lubricity, lack of sulfur and aromatics and its carbon footprint benefits. However, in current comparative environmental studies between biodiesel and petroleum-based diesel, some significant drawbacks, such as direct and indirect land use changes, nitrous oxide emissions, the decrease of biodiversity and the competition to food, are also mentioned [41]. The CoPIRIDE project rose to the challenge of an intensified and more sustainable biodiesel production process (Scheme 2):

Scheme 2

Transesterification reaction of vegetable oil and methanol.

Economic and ecological screening analyses of 18 process parameter variations at an early stage of process design revealed:

• A clear advantage for waste vegetable oil as biodiesel feedstock.

• The advantages of methanol over ethanol or methyl acetate as a reactant as well as solvent [11].

In the next step, detailed LCA and cost analyses comparing five alternative process routes (pretreated alkali-catalyzed, acid-catalyzed, heterogeneous acid-catalyzed, and two supercritical alternatives) utilizing waste vegetable oil and methanol, were carried out for industrial scale processing (production capacity 8,000 t/a). Four of these routes were based on the work of West et al. published in 2008 [42]. The fifth was based on a simulation of an improved supercritical process alternative developed within the project, based on the results of the initial screening procedure applying ASPEN Tech 11.1 as software tool (see Table 4). The analysis included pretreatment, product separation and purification, and the recovery of methanol.

The outranking of the resulting environmental efficiency of these process alternatives is shown in Figure 11 as the basis for eco-efficiency determination. The supercritical processes D and E are characterized by more or less the same environmental impacts (overlaying yellow and pink points and areas). More importantly, conventional process alternatives have benefits in none of the LCIA criteria considered. All in all, a reduction of the GWP (as a representative and relevant environmental impact) by nearly 70% can be gained utilizing the new CoPIRIDE process concept, compared to traditional heterogeneously or homogeneously catalyzed transesterification processes at moderate process conditions. Emission savings of approximately 6,900 t/CO₂-equivalents per year can be expected from the technology transfer at a production capacity of 8,000 t biodiesel/a.

Figure 11

Multi-criteria outranking of different biodiesel process alternatives utilizing waste vegetable oil (Process A: pretreated alkali-catalyzed; Process B: acid-catalyzed; Process C: heterogeneous acid-catalyzed; Process D: supercritical process, mass and energy flows considered according to West et al. [42] and Process E: supercritical process developed within CoPIRIDE) concerning environmental efficiency, taking into account several life cycle impact assessment (LCIA) categories.

Further, a cost analysis of the new developed process E in comparison to the results published by West et al. [42] (A–D) was performed. This cost analysis included pretreatment, product separation and purification, and the recovery of methanol. The resulting general expenses, direct (required for the production) as well as indirect manufacturing costs (overhead costs) are shown in Figure 12. More detailed information cannot be published due to confidentiality reasons. The conventional, heterogeneously catalyzed process C was found to be the cheapest. However, the total manufacturing costs of process E were between 12% and 20% less than the total manufacturing costs of the processes A and B, mostly applied in industry (see Figure 12).

Figure 12

Comparison of production costs for different biodiesel production alternatives (Process A: pretreated alkali-catalyzed; Process B: acid-catalyzed; Process C: heterogeneous acid-catalyzed; Process D: supercritical process, all published by West et al. [42] and Process E: supercritical process developed within CoPIRIDE) feedstock in all cases: waste vegetable oil.

In Figure 13, the results of the LCA investigation and the cost analysis were used as basis for an eco-efficiency portfolio, pointing out the outranking results under consideration of both key criteria for sustainability. It can clearly be seen that, although process C was found to be more cost efficient than the supercritical processes D and E (CoPIRIDE process), the latter is more environmentally benign and more eco-efficient (Pareto-optimal).

Figure 13

Eco-efficiency ranking of different biodiesel process alternatives utilizing waste oil (Process A: pretreated alkali-catalyzed; Process B: acid-catalyzed; Process C: heterogeneous acid-catalyzed; Process D: supercritical process, all published by West et al. [42] and Process E: supercritical process developed within CoPIRIDE) ranked according to their environmental and cost efficiency.

The following general lessons were learned from case study C for biodiesel production: i) supercritical is beneficial against conventional processing in the case of waste vegetable oil utilization, but not preferred for fresh oil, ii) the ratio between oil and alcohol or acetate should be kept as low as possible, and iii) heterogeneous should be preferred against homogeneous catalysis, both due to decreased efforts for downstream processing.

An explanation of the background of the results shown in Figure 13 is given in Figure 14. Whereas the good eco-efficiency of process C is mainly reasoned by its cost efficiency, environmental issues were mainly addressed in the cases of process D and E.

Figure 14

Criteria contribution on overall eco-efficiency ranking of different biodiesel process alternatives utilizing waste oil (Process A: pretreated alkali-catalyzed; Process B: acid-catalyzed; Process C: heterogeneous acid-catalyzed; Process D: supercritical process, all published by West et al. [42] and Process E: supercritical process developed within CoPIRIDE).

Based on the outcome of the eco-efficiency analysis, a mini-plant for a supercritical production process of biodiesel, through a transesterification of waste vegetable oil using methanol in intensifying continuous flow reactors, was designed by Chemtex Italia, Politecnico di Torino and further partners. Here, a supercritical environment enhances the contact among the above-mentioned reagents and intensifies the process. Clear safety benefits can be gained, due to the processing in continuous micro-flow compared to a continuous stirred tank reactor (CSTR) [11]. The mini-plant will allow for a continuous production of about 6 l/h biodiesel.