Introduction

As growing environmental responsibility and sustainability issues have led to an increasing awareness in lightweight engineering [1], composites with natural fibres as reinforcement (NFRP) have recently received significant attention due to their potential to replace petroleum-based composites [2,3,4]. Compared to glass fibre reinforced polymers (GFRP), of which the glass fibres demand high energy consumption during the glass melting and formation process, NFRP are advantageous due to their lower energy consumption requirement [5] and less environmental impacts [6] during fibre production. Meanwhile, natural fibres exhibit attractive mechanical properties, such as specific tensile modulus and tensile strength [7]. As composites are usually lighter due to low density and high specific strength of natural fibres, NFRPs used in automotive applications enable lower fuel consumption and lower emissions during use stage, thereby providing a lower carbon footprint [8]. The higher energy recovery and lower carbon emissions during the end-of-life stage of NFRPs compared to glass fiber reinforced polypropylene (PP-GF) also contribute to better environmental performance [9]. Cellulosic fibres from biomass like flax, sisal and miscanthus show advantages in terms of carbon dioxide (CO2) emission or ozone depletion from cultivation to production of composites, when compared to glass fibre reinforced composites with the same mechanical properties [5, 6]. An exception is the raw material extraction stage due to a possibly higher eutrophication effect caused by the use of fertilizers as well as pesticides during cultivation [9].

Miscanthus is a biomass plant originating from Asia and grows even under difficult conditions. Its genotype Miscanthus × giganteus is predominately cultivated in Europe as a resource-efficient, low-input crop, which can achieve yields of about 20–30 t ha−1 a−1 (dry matter) under different conditions [10,11,12]. Due to its perennial nature (15–20 year cultivation period) and its high nitrogen- and water-use efficiency, Miscanthus has a comparatively low impact on the environment [11]. Miscanthus as a perennial grass was found to have a high average modulus and hardness values comparable to those of other bast fibres such as sisal [13]. Fibres from Miscanthus grass in different morphologies used as fibre reinforcement or fillers in composites are investigated in several studies [7, 13,14,15]. They also showed potentials as impact modifier in composites [14].

An efficient way to fibrillate the natural fibres can be achieved by a mechanical pulping extrusion, which is applied in form of a co-rotating twin-screw extruder [16]. During this process, individual cellulose fibres are fibrillated into different particle extents in micrometer scale from biomass, depending on the screw configuration and extrusion parameters. This enables better properties of composites with long fibres at high aspect ratio (the ratio of fibre length L to fibre diameter D), a large fibre surface and an advantageous dispersion or distribution in the polymer matrix [17]. With the aid of twin-screw extrusion, Miscanthus fibre bundles can be fibrillated or modified by adding agents, chemicals and catalysts directly during the extrusion process [18]. In twin-screw extrusion only small amounts of chemicals are needed and the output up to 10 kg/h with pilot-extrusion [19] is much larger compared to other methods like low consistency batch modification. Therefore it has the potential to be used for industrial application as an environmental friendly and flexible fibre treatment process [20].

In recent research, different swelling agents were used in a twin-screw extruder to fibrillate the fibres from the miscanthus shoot axis of sieve fraction [18]. The fibres prepared this way were then used as reinforcement for cellulose acetate (CA) to manufacture all cellulose-based composites in a compounding process. It was demonstrated that dimethylsulfoxide (DMSO) as swelling agent enables aspect ratios up to 29. Consequently, CA-composites with mechanical fibrillated miscanthus fibres (CA-meMis-DMSO) show higher tensile strength and modulus compared to composites using the basic miscanthus sieve fraction (CA-Mis). The properties of CA-meMis-DMSO such as tensile strength and tensile modulus are reaching the level of glass fibre reinforced polypropylene with comparable fibre content. The CA-meMis composites using sodium hydroxide as swelling agent during fibre fibrillation (CA-meMis-NaOH) also show good mechanical properties, making it interesting as alternative material for PP-GF (20 wt.%) in lightweight applications like door panels, dashboard components and seat cover in automotive industry. However, the environmental impacts of these composites are unknown. In this study, a life cycle assessment (LCA) of the new material and its conventional alternative PP-GF is conducted. It helps engineers and decision makers to consider this aspect during material selection.

Methodology

Goal, Scope and Functional Unit

To assess the environmental performance of miscanthus-reinforced cellulose acetate (CA-Miscanthus) composites and compare it with their conventional reference product – PP-GF – an attributional LCA as defined in ISO 14040 and 14044 is conducted. For this, the production of the composites in Germany is considered. The CA-composites differ in terms of the treatment of the miscanthus fibres after the sieve fractionation. Three processing setups are considered resulting in three CA composite alternatives with a miscanthus fibre content of 25 wt.%. Their tensile modulus and tensile strength are comparable with PP-GF at 20 wt.% glass fibres. Table 1 presents the compositions, density, tensile modulus and tensile strength of each scenario.

Table 1 Characteristics and properties of the composites under study (properties for CA-composites are determined according to [18], for PP-GF20 are arithmetic mean of commercial PP-GF20 compounds from MOCOM Compounds GmbH & Co. KG and LyondellBasell, datasheets see supplementary material)
Full size table

The choice of the functional unit is based on the equivalent properties, which we assume that they could enable similar applications, e.g. in automotive industry. Exemplary we chose the equivalent flexural stiffness and tensile failure load scenarios using 1 kg PP-GF (20 wt.%) as reference material. To achieve equal flexural stiffness, the mass of the CA composites is calculated by Ashby’s material index [21, 22]:

$$m_{F} = \left( {\frac{{E_{r} }}{E}} \right)^{\frac{1}{3}} \frac{\rho }{{\rho_{r} }}m_{r}$$
(1)

where \({m}_{F}\) is the mass of the CA material with equivalent flexural stiffness, \({m}_{r}\) is the mass of the reference material, \(E\) is the Young’s Modulus and \(\rho\) is the material density with r indicating the reference material. For the equal tensile failure load scenario, the equivalent mass of the CA material is calculated assuming a tension rod loaded by a defined force \(F\):

$$F = \sigma A = \sigma_{r} A_{r}$$
(2)
$$m = \rho Al$$
(3)
$$m_{t} = \frac{{m_{r} \sigma_{r} \rho }}{{\sigma \rho_{r} }}$$
(4)

where \(F\) is the tensile load, \(A\) the cross section of the tension rod and \(\sigma\) the ultimate tensile strength of the materials (Eq. 2). The mass \({m}_{t}\) of the CA tensile bar can be calculated via its volume (length \(l\)) and density \(\rho\) ((3), so that the equivalent mass of CA composites can be referred from the reference material \({m}_{r}\) with (4).

According to this, two functional units are chosen as the production of the corresponding mass of the compounds, first for equivalent flexural stiffness and second for equivalent tensile failure load, each considering 1 kg of PP-GF (20 wt.%) as the reference. Table 2 summarizes the mass of the composites for both functional units.

Table 2 Masses calculated for equivalent flexural stiffness and equivalent tensile failure load serving as functional units
Full size table

System Boundaries and Description

The system boundaries are set to cradle-to-gate, meaning that the product is evaluated from the resource extraction (cradle) to the factory gate (i.e., before transport of the composites to further production facilities). The investigated product system is shown in Fig. 1, including the cultivation and the chipping of miscanthus, the production of matrices as well as all other required inputs, the fibre fractionation and fibrillation as well as the compounding. The use and end-of-life phase are excluded as it is expected that these are similar for all considered alternatives. For this reason, packaging materials are omitted. In our study, we assume a similar lifespan for both materials due to their comparable properties. The carbon absorption and storage of Miscanthus during the cultivation period are accounted for by subtracting the biogenic carbon in the results. This is important for the “beyond-gate” comparison with fossile-based GF-PP, as it involves extracting “fossil” carbon and brings additional environmental impact in the medium term. The transportation is not under consideration.

Fig. 1
figure 1

System boundary of the investigated miscanthus fibres reinforced biocomposites

Full size image

Per default, the cultivation and biomass fractionation were assumed to take place close to Goeda, Saxony in the east of Germany and includes soil preparation, crop establishment, fertilization, weed control and harvest of the biomass. For miscanthus, a 20-years cultivation period is taken. Average biomass yields were assumed in accordance with the geographical focus. Information of the fractionation steps were derived from measurements at a commercial pilot plant in Göda (personal communication with Uwe Kuehn, TSK GmbH). Production of the compounds was assumed to occur in Dresden, Germany and corresponding inventory data are based on lab experiments at TU Dresden and expert estimates. The shares of electricity in Germany are applied in this study.

Calculations were conducted using openLCA 1.11.0 (Green Delta GmbH) relying on background inventory datasets from ecoinvent 3.8 cut-off. For few inputs like acetyltributylcitrate (ATBC), celluloseacetate (CA), and sodium acetate (required for CA production), inventory information was not available in ecoinvent. For this reason, information from literature about production processes was used to complete the inventories.

The impact assessment is conducted using the Product Environmental Footprint (PEF) methodology of the European Commission [23] and the EF 3.0 Method (adapted) was chosen for the modelling. Accordingly, results are presented as single scores. In addition, eight impact categories were considered in more detail providing contribution analyses. This includes Acidification (AC), Climate change (CC), Eutrophication, freshwater (EF), Particulate matter (PM), Water use (WU), Resource use, fossils (RF), and Resource use, minerals and metals (RM). These impact categories were selected, as they have been identified as the ones contributing at least 90% of the total impacts on single score level (excl. toxicity categories, due to insufficient robustness as suggested [24]).

Life-Cycle-Inventory

The life-cycle inventory (LCI) identifies and quantifies material and energy flow, emissions to air and water, as well as waste generation during the system boundaries.

Primary Production

Table 3 presents key inventory information for commercial cultivation of miscanthus. Inventory data for the cultivation procedures were taken from [25]. Direct (chipping) instead of indirect harvesting (cutting to swath and baling) were added for our case. Field emissions were modelled as suggested in Zampori & Pant [23], omitting potential soil carbon increases due to miscanthus cultivation. The full inventory is provided in Appendix in the Supplementary Material.

Table 3 Key inventory parameters of biomass cultivation (entire dataset in Supplementary Material)
Full size table

Fractionation & Refining of Chipped Miscanthus

After harvest, the chipped miscanthus is separated in different fractions. The particles are separated by size in a demonstration plant via two sieving steps (round holes of 4 mm), alternated by a refining step (disc distance of 0.1 mm). Data for the considered plant were provided by Technical Service Kuehn GmbH. It includes mass flows and the electricity consumption. The particles of interest for the CA-Miscanthus composites are characterized by a size between 40 and 100 µm and make up about 13.3 % of the harvested chopped biomass. The remaining fractions are useful for a wide range of applications. For this reason, allocation of the upstream impacts as well as the electricity consumption of the separation was required. Assuming a mass-based allocation approach, electricity consumption amounted to 0.067 kWh kg−1 of S40-100 (refined).

Fibrillation

The homogenous miscanthus fraction S40-100 (refined) are then pre-treated in a twin-screw extruder Process 11 (Thermo Fisher Scientific) plant in PTS, Heidenau (Fig. 2).

Fig. 2
figure 2

Two-step composites manufacturing process with fibillation (top) and compounding (bottom) with LH long helix screw, DS distributive mixing element, KS kneading screw and RFS reverse screw [18]

Full size image

Pure thermo-mechanical extrusion (Fig. 2, top) is used to fibrillate fibres in a small scale with high fibre length to width ratio. These fibrillated fibres (signed as “meMis”) were produced by adding solvent medium DMSO or NaOH in the process. The screw temperatures were 175 °C and 65 °C respectively. Biomass and electricity requirements were measured and provided by PTS. Table 4 details the inventory. As DMSO and NaOH are washed out after fibrillation and can be re-used (almost) entirely, their replacement was not taken into account in this study. This step does only apply to the composites CA-meMis_DMSO and CA-meMis_NaOH.

Table 4 Inventory of fibrillation (1 kg fibrillated fibres)
Full size table

Compounding

Finally, the composites are produced using a twin-screw extruder (Prism Eurolab, Thermo Fisher Scientific, Fig. 2, bottom). The biodegradable citric acid-based plasticizer acetyl tributyl citrate (ATBC) CITROFOL® BII is used as an aid for compounding of cellulose acetate in order to improve the processing and usage properties [26]. The liquid plasticizer is from Jungbunzlauer Ladenburg GmbH. The fibres and cellulose acetate are dried at 80 °C in an air-circulating oven for 8 h in order to ensure a humidity content below 1% whilst preventing further changes in the chemical composition of the plants before compounding. Inventory of this compounding step encompasses electricity and material requirements, including the matrices cellulose acetate, and is derived from experiments at ILK, TU Dresden. For comparison and energy measurement, processing studies on compounding PP-GF are performed using the same extruder. In addition, the amount of the biogenic carbon embodied in the composites is estimated from the carbon content of the fibres, the matrices and their composition in the final products. Table 4 gives an overview of the inputs required in the compounding step.

The inputs required for Polypropylene and the glass fibres are taken from the ecoinvent database. The ATBC inventory is based on the dataset “Acetyl tributyl citrate, Production mix, at plant”, which is available in the product environmental footprint (PEF) database. Data for CA are derived from literature [27]. The used datasets are given in the Supplementary Material (Table 5).

Table 5 Inventory of compounding (1 kg compounds through twin-screw extruder)
Full size table

Results

Comparative LCIA of the Composites

In the LCIA stage, the emissions and resource extraction are translated into the specified environmental impact scores by multiplying with the corresponding characterization factors. Table 6 summarizes the absolute results for the production of 1 kg CA-meMis compounds for the three different fibre treatment scenarios and their reference product PP-GF (Fig. 1). Figure 3 and Fig. 4 gives a graphical representation of this comparison for the assessed impact categories in the case of equivalent flexural stiffness and equivalent tensile failure load.

Table 6 Impact results for the production of 1 kg CA-miscanthus and 1 kg PP-GF
Full size table
Fig. 3
figure 3

Comparison of the environmental performance of CA-rMis, CA-meMis_DMSO, CA-meMis_NaOH and PP-GF for equivalent flexural stiffness. The value scoring highest in each category equals 100%. Abbreviations are: Acidification (AC), Climate change (CC), Eutrophication, freshwater (EF), Particulate matter (PM), Water use (WU), Resource use, fossils (RF), Resource use, minerals and metals (RM) and single score (SS)

Full size image
Fig. 4
figure 4

Comparison of the environmental performance of CA_rMis, CA-meMis_DMSO, CA-meMis_NaOH and PP-GF with the equivalent load. The value scoring highest in each category equals 100%

Full size image

In the case of equivalent flexural stiffness, the miscanthus-based composites result in higher environmental impacts than the conventional reference product PP-GF (Fig. 3). This applies to all impact categories assessed including also the single score indicator, combining the results of impact categories for the total environmental sustainability [28].

In the case of equivalent tensile failure load, a similar tendency is observed (Fig. 4). The miscanthus-based composites result in higher environmental impacts than the conventional reference product PP-GF and the differences are higher than for equivalent flexural stiffness.

Contribution Analysis

Figure 5 shows a contribution analysis for the CA composites.

Fig. 5
figure 5

Contribution analysis of the composites

Full size image

It highlights cellulose acetate as the dominant contributor, accounting for more than three quarter of the impacts in all categories besides EF. Electricity for compounding causes the most significant Freshwater Eutrophication and is the second hot spot in several other impact categories. Further significant contributors are ATBC and electricity for compounding.

Performance Against PP-GF and Other NFRP

The environmental performances of the three pathways to manufacture cellulose composites are almost equal, irrespective of the considered setups. Miscanthus composites perform worse than conventional reference based on PP and GF in all assessed impact categories.

Given these results, it can be concluded that cellulose-based composites using cellulose acetate as matrix and extrusion for further fibre fibrillation are less favorable than the conventional reference. This result bases on the fact, that CA-production has a much higher impact than PP, though the miscanthus fibres are more environmentally friendly than glass fibres. Compared to the analyses in other studies with kenaf or wood fibres and other biopolymers [29,30,31], higher impacts of our cellulose-based composites study are observed.

Discussion

Inherently, these findings are strongly dependent on the data used for the life cycle inventory. For this reason, possible limitations and uncertainties due to the LCI data quality are discussed in the following. The significant inputs and processes found in contribution analysis are further assessed to identify options to improve the environmental performance of the miscanthus-based biocomposites. This section follows the general structure of bio-based value chains, ranging from the biomass cultivation to the conversion. Finally, uncertainties surrounding the missing data and scale of manufacture are discussed to provide context to the conclusions and recommendations.

Influence/Variability of Biomass Cultivation

Agricultural processes strongly depend on the context, given bio-geophysical and climatic conditions, resulting for instance in variation in the yield levels. In addition, management parameters might be specific in each region. This was previously shown for GHG calculations for miscanthus cultivation, e.g., by Lask et al. [25]. However, as can be seen in Fig. 4, miscanthus cultivation accounts for a marginal contribution overall. Even substantial shifts in the cultivation intensity will not affect the outcome of the bio-based/fossil comparison.

Influence/Variability of Miscanthus Fractionation

Data for the fractionation and associated energy requirements were derived from a demonstration plant in Goeda, Germany. The plant has been set up recently, and does not necessarily represent commercial conditions. For this reason, further efficiency gains could be expected here. Mass-based allocation has been used to solve the multi-functionality problem in the fractionation process. This approach was taken as reliable economic values were missing. However, due to its low contribution to the overall impacts, this stage is also of low relevance.

Influence/Variability of Fibrillation and Compounding

Together these two stages account for the majority of miscanthus-related impacts.

LCI data on the electricity requirements of the extruders have been derived from lab experiments (ca. 4 kWh/kg for fibrillation and 3.6 kWh/kg for compounding). A comparable energy consumption is documented in [31]. A much lower electricity consumption 0.3 – 0.5 kWh is given in other publications [32, 33]. Such differences typically occur when using pilot plant or industrial data. With regard to an upscaling of the production processes [34], some efficiency gains can be expected. For comparison between PP-GF and the CA-compounds, energy consumption during extrusion is measured for all materials in compounding studies on the same extruder.

The LCI contains information on material inputs for the compounding, including CA and ATBC. The inventory of the latter is based on an existing PEF dataset. Despite some uncertainty related to this dataset, the impact on the overall conclusion of the study seems marginal in comparison with CA, as the ATBC-related impacts range between 4 and 11% across impact categories.

Comparison Miscanthus and Glass Fibres

The environmental impact of 1 kg miscanthus and glass fibres is compared in Fig. 6. Glass fibers show a higher impact than refined miscanthus fibers across all observed categories. When compared to fibrillated miscanthus fibers, glass fibers exhibit a higher impact in every category except for Climate Change, Freshwater Eutrophication and Fossil Resource. In these categories, the significant electricity usage during fiber fibrillation contributes to the increased impact. The results indicate that refined miscanthus fibers offer environmental benefits. However, the fibrillation process should be made more energy-efficient to ensure environmentally-friendly production.

Fig. 6
figure 6

Comparison of the environmental performance of 1 kg rMis, meMis_DMSO, meMis_NaOH and glass fibres. The value scoring highest in each category equals 100%

Full size image

Cellulose Acetate as Key Dataset

In contrast, CA is the major contributor for all but one impact categories, contributing consistently more than 75%. Only exception are eutrophication impacts where the electricity requirements for compounding are the major driver. Thus, the CA inventory needs to be critically discussed to estimate its potential as a matrix material in sustainable composites.

The CA inventory used in this study is derived from literature [27], which builds on insights from a collection of studies and patents for industrial CA production [35]. Major contribution to the overall impacts is related to the use of acetic anhydride, which is produced by ketene route or by acetaldehyde oxidation, accounting for instance for two-thirds of the climate change impact (Fig. 7). This traditional method was identified to be environmentally unfavorable with high energy consumption [36]. Much research and patenting focused on producing more eco-friendly CA, such as using ionic liquids or alternative cellulose origins instead of wood pulp and cotton linters [37,38,39,40,41]. However, they are still in laboratory scale and inventories are not published yet.

Fig. 7
figure 7

Impact contribution for climate change during CA-production

Full size image

Consequently, another cellulose derivate – cellulose aceto-butyrate (CAB) – is compared with CA to determine if it is a potential alternative matrix for further product development. The LCI is taken from PEF database. Although the study is only valid until 2020, it is selected because it is the only transparent, publicly available dataset. Additionally, petrol-based PP is taken into consideration as reference matrix. It can be seen, that CAB shows lower impact than CA in all categories. Since CAB-data from PEF are based on stoichiometric calculations, some deviations are possible. However, compared to polypropylene, CA and CAB have both much higher impacts (Fig. 8).

Fig. 8
figure 8

Comparison of the environmental performance of different matrices according to the PEF methodology. The option scoring highest in each category equals 100%

Full size image

Beyond the Gate

In our study, cradle to gate system is considered. Potential environmental impact differences through fuel reduction value [42, 43] caused by mass difference between the CA-composites and PP-GF in their use phase could not be identified by now, since no specific product was investigated. A future study could focus on a specific product and also consider an end-of-life system for better understanding. Since the mass of a CA-product is always slightly higher than its PP-GF reference, the relations of the results are expected to remain the same, at least in the use phase.

Conclusion

In this study, the environmental impact of a cellulose-based composite is investigated and compared to a conventional PP-GF. For the miscanthus fibre reinforcement, three scenarios with different fibre modification are considered. The results of this study indicate that CA-Mis composites have overall higher environmental impacts than the reference material PP-GF in all considered categories. The high impacts result from the production of CA, with acetic anhydride being the major contributor. It can be concluded, that the cellulose ester reinforced with cellulose fibres, despite its biomass origin, may not be as eco-friendly as eventually assumed. Based on that, using industrial grade cellulose ester have no environmental advantages to polypropylene. However, the bio-based miscanthus fibres are still interesting as fibre reinforcement. Hence, for further development of cellulose-based composites/products other matrix systems should be taken into account. Above that, research should concentrate on the scale-up of promising CA-routes with lower impact or other bio-based matrices for an environmental friendly industrial transfer.