Keywords

1 Introduction

The global production of PET in the year of 2018 was about 78 million tons. Around 98% of its production was obtained from non-renewable resources and just 2% from secondary raw materials (recyclates) [1]. PET is most used in the manufacture of high-quality plastic packaging and polyester textiles. Due to its mechanical properties, this polymer shows a good resistance against many chemicals and for this reason it meets the highest quality for food packaging. For some packaging applications, there is a need of protecting the inner content against light and air exposure. Therefore, the combination of different polymers (multilayer structure) is a viable solution for the packaging industry to fulfill specifications concerning low material consumption, high propensity against O2 permeation and mechanical strength [2]. Since raw materials are primarily derived from non-renewable sources, a sustainable approach for recycling PET-containing waste is essential. Furthermore, almost all the economic value of these materials are lost in their short-lived value chain. Recycling technologies present a solution for this challenge, as they provide a secondary raw material that can substitute and thus reduce the use of non-renewable resources to produce new PET. The depolymerization products can be recovered in high quality, so that PET can be produced in novel quality, which is suitable e.g. for food-contact applications.

The most established recycling technologies for PET are mainly thermomechanical processes, which are only applicable for high purity PET feed materials. Mechanical recycling of PET composite waste is technically and economically not viable due to the required polymer separation. As a result, these materials are mostly sent to energy recovery. However, the economic value as well as the resources of the waste stream cannot be recovered, and the product comes shortly to the end of its life cycle. Targeting these “hard-to-recycle” PET materials, the company RITTEC Umwelttechnik GmbH developed the innovative “back-to-monomer recycling” (BMR) technology revolPET.

1.1 State of the Art (Depolymerization of PET)

Chemical methods can supplement the mechanical recycling technologies to treat the “hard-to-recycle” waste fractions. PET feedstock recycled with pyrolysis, requires mostly high operation temperatures (450 °C…900 °C) and none of the developed technologies was able to prove itself economically viable. Furthermore, PET has not become the focus of pyrolysis research because of unfavorable composition of the decomposition products (bulk petrochemicals) [3, 4]. In addition to pyrolysis, the BMR methods, based on depolymerization, offers a possible solution for the aforementioned challenge. BMR processes are focused on recovering the monomers or other constitutional repetition units of the polymer for a new PET production [5]. These processes are, regarding their reaction pathways, divided into: ammonolysis, glycolysis, methanolysis and hydrolysis [6]. This contribution will focus on the hydrolysis processes.

The hydrolysis of PET is carried out in the presence of water either utilizing catalysts, high pressure, and high temperature or with strong bases and acids [7, 8]. In presence of strong bases, the ester linkages of PET are dissolved, resulting on EG and two salts (monosodium terephthalate and disodium terephthalate (DST)) [9]. After a further treatment of DST with a strong acid, both valuable monomers EG and TA can be recovered [9, 10]. Normally, alkaline hydrolysis processes are operated with high base concentrations (up to 20 wt.%), pressures around 1.4…2.0 MPa and high temperatures (up to 250 °C) [11, 12]. These last-mentioned points combined with a batch operation make these hydrolysis processes economically unviable for the PET depolymerization.

Given the growing volumes of PET being produced, there is a great economic and environmental need of developing a sustainable process solution for the depolymerization of PET. Compared to batch processes, continuous production offers many benefits regarding energy and resource efficiency, constant product quality and scalability. Moreover, the difficulty to increase the heat transfer area on scaling batch processes leads to longer residence time and lower reaction rates [13]. The most promising continuous alkaline depolymerization process was presented by Benzaria et al. [14]. The reaction took place in a kneading extruder at 120–160 °C with a residence time of approx. 6 min with a depolymerization degree above 60%. After the implementation of a further downstream process step operating at 80…130 °C for 30 min, a higher depolymerization degree (>95%) was achieved [14].

1.2 The Background of the revolPET Technology

In a joint R&D activity, RITTEC Umwelttechnik GmbH together with the Institute for Chemical and Thermal Process Engineering (TU Braunschweig) developed a continuous PET depolymerization technology. The newly developed revolPET process builds on the continuous operation of a twin-screw extruder leading to a residence time below 1 min and to high depolymerization degrees [15]. Furthermore, hydrolysis depolymerization has a great advantage in comparison to other processes because the produced TA can be directly used for a new PET production without further processing [16].

2 Materials and Methods

2.1 Process Concept and Experimental Setup

The process concept and the experimental setup were first presented in Biermann and Brepohl [15]. In a pilot plant, 6.96 kg/h PET waste material and 3.04 kg/h sodium hydroxide (NaOH) is conveyed into the twin screw extruder. The stoichiometric ratio of PET waste to NaOH is adjusted to 2.1 molNaOH/molPET calculated assuming 100% PET content on the PET waste stream.

The post-consumer multilayer PET trays were mixed with post-consumer monolayer trays and bottles. The trays were extracted from the waste stream and a second sorting was applied to extract the multilayer trays only. The stream contained a minimum 80% of multilayer trays and was ground to <3 mm in diameter. The material was processed in the extruder (ZSE 27 MAXX, Leistritz Extrusionstechnik GmbH) and NaOH (purity ≥ 98%, diameter < 3 mm) was added through a side feed in cylinder number 2. The temperature of the reaction zone was set to 130 ℃, see Fig. 1.

Following the depolymerization, the extruder output was conveyed in a dissolution tank where 66.9 l/h water is added. After the complete dissolution of all dissolvable materials, the suspension is sent to a filtration cascade where solid impurities are removed in a two-stage filtration. The particle free solution of DST and EG in water is purified from soluble impurities. The isolation of recycled TA (rTA) is conducted by the addition of a strong acid.

Fig. 1.
figure 1

Continuous revolPET depolymerization process (adapted from [15]).

Full size image

By solid/liquid separation the rTA is extracted from the product suspension and sent to the drying process. The filtrate is sent to the recycled EG (rEG) purification process where the water is removed and the rEG is distilled to obtain the required purity. Figure 2 presents a block diagram of the revolPET process. To date, the process has focused on recovering the monomers rTA and rEG, future technology developments will also address the sodium salt (Na-X).

Fig. 2.
figure 2

Block diagram of the revolPET process.

Full size image

After the isolation, rTA and rEG were repolymerized in a lab scale experiment to test the resulting polymer quality. Up front rTA (1.98 mol), virgin isophthalic acid (IA) (0.02 mol), diethylene glycol (DEG) (0.015 mol), a polymerization catalyst and the rEG (2.4 mol) were premixed and filled into the reactor. Subsequently, the reactor has been heated up to start the polymerization. After reaching a clear prepolymer melt, the polycondensation conditions were applied to remove water from the reaction equilibrium.

2.2 Environmental Assessment of the revolPET Process

For the revolPET process, a Life Cycle Assessment (LCA) was conducted following a cradle-to-gate approach based on the framework of ISO 14040/44 [17]. For the LCA considerations, holistic investigations were performed to identify parameters with a strong influence on the environmental impact. As the current technology readiness level (TRL) of the process holds a lot of potential for improvement (TRL 4), various studies were conducted to identify hotspots and opportunities that must be considered in the ongoing process development. It builds on analyses of the operation of a single process step, apparatus, or unit operation and the entire process as well as the site infrastructure.

Following the approach of Wesche [18], all inputs and outputs of the process are included in the investigation, i.e. the expenditures for providing the reactants, utilities, and auxiliaries, as well as the electricity, process emissions and the waste disposal.

The system boundaries include the expenditures of the recycling process, except for impacts of service and maintenance as well as the production plant and connected infrastructure itself, since the design is still not finalized. The production of 1 kg rTA is used as the functional unit. The inventory is mainly based on primary data collected from measurements and experiments at the pilot scale regarding the continuous processing of 14 kg/h of PET/PE waste fraction, complemented with calculations. The calculation of the elementary flows for the life cycle inventory is based on generic datasets from the ecoinvent database v3.7.1 [19]. For the life cycle impact assessment (LCIA), the characterization models described in the ReCiPe method [20] were used.

3 Results and Discussion

3.1 Process and Technology-Related Results

The PET waste material was depolymerized in the twin screw extruder as described in Sect. 2.1. Since the PET content of the waste material is unknown, no degree of depolymerization could be calculated [15]. The analytical results are displayed in Table 1 and show the obtained quality of the rTA in comparison with industrially produced purified TA (PTA) from crude oil.

Table 1. Comparison of the rTA and rEG (revolPET) with virgin PTA and EG.
Full size table

It is shown that the values for the color and moisture of rTA match the industrial specifications and the typical values of PTA. Furthermore, the concentration of 4-Carboxybenzaldehyde (4-CBA) and 4-Methylbenzoic acid (4-MBS) deceeds the typical values of PTA. In contrast to the PTA, the rTA is a mixture of the isomers TA and IA. IA is commonly added during the production of PET and influences the crystallization behavior of the material. It is therefore not considered an impurity. Concerning the rEG, the industrial specifications are as well met, except for the moisture content of the material. This may be tolerable during the polymerization process because water is removed from the polycondensation reaction equilibrium of PET.

The recycled monomers were repolymerized in a lab scale experiment to assess the quality of the resulting PET. For comparison purposes, a polymerization experiment with the same reaction conditions was applied with virgin PTA and EG as reference. The results are displayed in Table 2 and show that the values deviate from the reference by 5.5 at L*, 0.8 at a* and by 3.1 at b*. The IV value of the recycled PET (rPET) is with 0.599 dl/g in a comparable regime.

Table 2. Comparison of the rPET from the chemically recycled monomers with virgin PET under identical lab conditions.
Full size table

The rPET was furthermore analyzed for Non-Intended Added Substances (NIAS) by the Fraunhofer Institute of Process Engineering and Packaging. As a result, the recycled material follows the safety requirements of Article 3 of Regulation (EC) Nº 1935/2004 for food-contact applications. In conclusion it was shown that mono- and multilayer food packaging can be recycled by the revolPET process to produce rTA and rEG in high quality, sufficient to produce rPET that meets the industrial requirements.

3.2 Environmental Assessment of the revolPET Process in the Current TRL

The results of the LCIA were subdivided in two parts: (1)-significantly influenced by the technology; (2)-mainly influenced by the type of energy supply employed, correlating with the available site infrastructure. This structure allows the presentation of the influence from TRL and the site infrastructure on the environmental assessment of the new developed recycling process. Figure 3 shows an excerpt of the LCA consideration, here the assignment of the environmental impacts with respect to the technology itself and the site infrastructure in the different impact categories.

Fig. 3.
figure 3

Excerpt of the LCA results using the characterization models of the ReCiPe method [20].

Full size image

The share of the technology part shows a variety of trends. For the impact category climate change (CC) less than 30% of the potential environmental impacts results from the process itself. The hotspot analyses also shows that the infrastructure share is largely influenced by the heat demand. The heat supply here in TRL 4 is ensured by electricity due to the existing infrastructure in the pilot plant (Braunschweig, Germany). The distribution in the category of water depletion (WD) is reversed, more than 70% of the total potential environmental impact is caused by the process, here mainly by the auxiliary materials. The results of the hotspot and the scenario analyses are used as a guidance for an environmentally beneficial process development and design. Furthermore, the influence of boundary conditions and parameters are highly dependent on the actual development stage.

Similarly, studies consider improving the TRL (TRL 4 à TRL 7) of the revolPET process as well as by changing the production site from the technical lab to an industrial park. Nevertheless, studies comparing the rTA production from the revolPET process with the fossil route are also conducted. Both, the increased TRL as well as the changed production site resulted in a significant reduction (14…34%) of the potential environmental impacts in all impact categories. The comparison of the revolPET process with the fossil production route of TA shows environmental benefits in the current development stage of the revolPET technology. For example, in a TRL 7 scenario, the potential environmental impacts in the category CC can be reduced by 33% by the revolPET process in an industrial park compared to the fossil production route.

4 Conclusions and Outlook

The innovative BMR revolPET technology was presented and discussed in this contribution. Regarding the process itself, revolPET presented a possible solution for PET composite materials, as the PET reacts selectively and continuously in a twin-screw extruder with depolymerization quotas up to 95% based on the PET fraction of the feed material under mild conditions (ambient pressure and 130 °C) and in less than 1 min. The produced monomers were analyzed, and it has been stablished that they have a similar quality as the virgin monomers. Looking at the environmental benefits of such innovative technology, the recycled monomers can reach a 33% reduction of CO2-emissions if their production in an industrial park is compared to the crude oil route. Furthermore, the caused environmental impact of the revolPET process can be reduced after scaling and optimizing the process.

For the further development of the revolPET process, a miniplant with a nominal throughput of 18 kg/h of PET composite waste is being built. It is planned to get a fully automated production, with heat and water recycling in the process, as well as the recovery of the valuable produced salt (Na-X) on his market specifications. In addition to these points, the application of the revolPET technology is being extended also for polyester composite textiles and should be addressed in following studies.