*2.6. Experimental Setup*

The thermochemical experiments were conducted in a Linseis STA PT1600 thermobalance. The furnace chamber can be purged with different gases and thus different processing atmospheres can be set. The sample material was placed in a crucible, which rested on a thermocouple measuring the sample temperature through the ceramic bottom. All experiments were conducted with a heating rate of 30 ◦C/min up to 850 ◦C sample temperature. To establish the atmospheric conditions of pyrolysis, gasification and incineration a mass-flow-controller for nitrogen and synthetic air was used. During the experiments, the weighing system of the thermobalance was protected by nitrogen as purge gas flow at 75 mL/min, not interfering with the conversion process. The sample gas, which determines the thermochemical process, was injected separately and flushed the sample at 50 mL/min. Pyrolysis was performed in nitrogen atmosphere while synthetic air was used for incineration. For gasification, an air ratio of 0.3 was calculated based on data from fuel analysis and nitrogen and synthetic air flow set accordingly.

The sample materials created for laboratory analysis were also used as sample material in the TGA. The sample weight was 100 mg per trial and at least three repetitions (n) were carried out.

Pyrolysis trials were repeated for all three materials in a retort furnace with an increased sample size of 0.5 kg. As shown in Figure 5, the setup consisted of a modified Thermconcept KM 70/13 chamber furnace containing a stainless steel retort as fixed-bed reactor.

**Figure 5.** Retort furnace for pyrolysis (Figure: Kai Schlögel).

To ensure oxygen exclusion and thus pyrolytic conditions, the retort was continuously flushed with nitrogen. The furnace was heated up at a rate of ca. 8–10 K/min. To minimise the risk of sudden pressure rises the heating program involved a 30 min plateau at 250 ◦C before heating up to the target temperature of 700 ◦C and holding for another 30 min. The pyrolysis gas produced in the retort was led through steel containers tempered to −20 ◦C to collect all condensable components. In case of blockage this condenser can be bypassed through a bursting disc which was not necessary in the presented trials. The remaining permanent gas fraction was burned in a post-combustion chamber to minimise harmful gas emissions. After each trial, char and condensable product fractions were collected from the retort and the condenser respectively and analysed for proximate and elemental composition. As mentioned above, the condensate was additionally used for pre-trials in biotechnological processing. On this basis, an outlook is given for its potential to produce precursors for chemical industries.

### *2.7. Biotechnological Upcycling*

Microorganisms can colonise various habitats by metabolising a variety of molecules. In biotechnology, this potential is applied by converting inexpensive substrates into economically interesting molecules. For this purpose, microbes are isolated from diverse habitats. Hydrocarbon-degrading microbes can be found for example in soils and sediments and even enriched in the event of an oil-spill [16]. Lignolytic microbes often have the properties to degrade polycyclic aromatic hydrocarbons [17]. The idea of biotechnological upcycling is to use those abilities to metabolize pyrolysis condensate components and convert them into molecules that can be further used in the sense of a circular economy [18].

The approach to utilise pyrolysis condensate as feedstock has already been demonstrated for pure fractions. Ward et al. [19] were able to produce the biodegradable plastic polyhydroxyalkanoate (PHA) by feeding *Pseudomonas putida* CA-3 with pyrolysis oil from polystyrene (PS). For the pure fraction of polyethylene (PE), a pyrolysis condensate consisting of 99% alkanes and alkenes was also converted to PHA with *Pseudomonas aeruginosa* PAO-1 [20]. Pure polyethylene terephthalate (PET) pyrolysis condensate, which mainly consisted of oligomers of terephthalic acid, was added to a sodium hydroxide solution, resulting in 97% terephthalic acid of the solid fraction. Kenny et al. [21] were able to isolate several bacteria, also of the genus *Pseudomonas*, which metabolise terephthalic acid and thus also capable produce PHA from the PET pyrolysis condensate.

These studies each used a pure plastic fraction as starting material. In the case of mixed polymer fractions, the composition of the pyrolysis condensate can differ considerably from the pyrolysate of the pure fraction. It can consist of a mixture of all the substrates described above and additional molecules. The organism growing on this mixture must be able to ideally convert a wide range of substrates and tolerate the presence of potentially toxic substances. Furthermore, the following requirements apply when using mixed-fraction pyrolysis condensate as a substrate:


### **3. Results and Discussion**

*3.1. Thermogravimetric Analysis*

In total 39 marine litter trials were conducted with TGA to simulate pyrolysis (n = 14), gasification (n = 11) and incineration (n = 14).

The curves of the TGA (Figure 6A, Figure 7A, Figure 8A) and first derivative of TG (DTG) (Figure 6B, Figure 7B, Figure 8B) show similar trends for pyrolysis, gasification and incineration. As expected, the curves for the homogeneous reference material mulch foils are significantly smoother in all trials. However, the curves for the individual materials differ in relation to the respective thermochemical process. Pyrolysis mass losses were determined for ML Sylt at 79 wt%, Norderney at 88 wt% and mulch foils at 91 wt%. For all materials, the TGA and DTG signal show gradual progress of further mass loss above 500 ◦C until the end of the programme, which indicates a migration of fixed carbon to volatile phase.

For ML Sylt, mass loss for gasification (Figure 6) increases (compared to pyrolysis) insignificantly up to 80 wt%. The increase in mass loss of the other two materials is larger and an indication that the char remaining after pyrolysis is gasified. In the case of incineration, the analysis results of the ash content listed in Table 4 differ from those of the TGA for ML. As shown in Figure 7, the ash content can be determined for ML Sylt at 14 wt%, ML Norderney at 5 wt% and mulch foil at 0.1 wt%. Reason for this difference is the heterogeneity of the collected sample or the calcination of contained carbonates at temperatures above 550 ◦C in TGA.

TGA measurements can vary greatly depending on the sampling. Figure 9 shows a comparison of determined sample mass constancy of all conducted TGA trials. In order to evaluate the significance of the TGA trials independently of the heterogeneity of the sample, the deviation from the reference material mulch foil can be considered. The complexity of the investigated thermochemical processes pyrolysis (A), gasification (B) and incineration (C) are clearly visible. Further oxidation decreases deviation in total.

**Figure 6.** (**A**) shows the thermogravimetric analysis (TGA) mean value of pyrolysis. (**B**) shows the standardised derivative of TG (DTG) signal. (ML Sylt n = 5, ML Norderney n = 5, Mulch foil n = 4).

**Figure 7.** (**A**) shows the TGA mean value of gasification. (**B**) shows the standardised DTG signal. (ML Sylt n = 4, ML Norderney n = 4, Mulch foil n = 3).

**Figure 8.** (**A**) shows the TGA mean value of incineration. (**B**) shows the standardised DTG signal. (ML Sylt n = 5, ML Norderney n = 5, Mulch foil n = 4).

**Figure 9.** Comparison of determined sample mass constancy of all conducted TGA trials. (**A**) shows deviation for pyrolysis: ML Sylt n = 5, ML Norderney n = 5, Mulch foil n = 4. (**B**) shows deviation for gasification, ML Sylt n = 4, ML Norderney n = 4, Mulch foil n = 3. (**C**) shows deviation for incineration: ML Sylt n = 5, ML Norderney n = 5, Mulch foil n = 4.

Table 5 shows the measured reaction temperatures and calculated standard deviation, respectively. Ti indicates the initial temperature when 1 wt% mass loss is quantified while Tc shows the end of the reaction and a 99 wt% mass constancy. Tp marks the temperature of the highest relative mass loss as calculated by the derivation of the TGA signal. Although no moisture could be quantified in the sample of ML Sylt, initial mass loss can be observed at temperatures below 150 ◦C. The temperature of the reaction peak ranges from 438–467 ◦C. Tp differs slightly between the ML fractions, while Tp for mulch foil is very constant. Based on Tc, it can be observed that devolatilization of ML during pyrolysis is significantly slower and occurs up to high temperatures. The reaction with oxygen during incineration is completed earlier that is, at lower temperatures. In conclusion, it can be stated that the standard deviation for the results of the ML fractions is consistently higher than for the mulch foils.

**Table 5.** Overview of the specific reaction temperatures and calculated standard deviation. Ti indicates the initial temperature when 1 wt% mass loss is quantified while Tc shows the end of the reaction and a 99 wt% mass constancy. Tp marks the temperature of the highest relative mass loss as calculated by the derivation of the TGA signal.


### *3.2. Retort Furnace*

The pyrolysis trials in the retort furnace allow separate capture of char and condensate products and subsequent analysis of samples. The results are presented in Table 6. The trials resulted in char yields ranging from 9.3–18.2 wt% and condensate yields of 21.5– 32.3 wt%. As suggested by raw material analysis the char fractions from marine litter educts contain higher amounts of ash and thus lower carbon. Elementary composition in general reflects the results of ultimate analyses as shown in Table 4 with heteroatoms of hydrocarbons concentrating in the char fractions up to 2.0 wt% of sulphur and 4.4 wt% of chlorine. Ultimate analysis shows high contents of carbon and hydrogen for marine litter condensates likely deriving form organic hydrocarbons while mulch foil condensate leaves a significant gap in the balance of detectable elements. This does not match its comparatively high calorific value suggesting an imprecision in elementary analysis. For all educts calorific values of condensate exceed their respective char fractions, demonstrating the release of energy-rich compounds in the pyrolysis process. These results appear plausible considering the raw materials consisting mainly of ash and volatile matter.


**Table 6.** Characteristics of pyrolysis products derived from different plastic wastes (as received) (\* n = 2; \*\* n = 3).

For further elemental analysis of the pyrolysis products an X-ray fluorescence (XRF) analysis was performed and the detected elements with the highest mass fractions were plotted as shown in Figures 10A, 11A and 12A. Furthermore, the content of heavy metals mentioned in the German ordinance for waste incineration (17. BImSchV) was determined in the samples and plotted on the right side (B). Due to the methodical detection limits of the XRF analyser, sodium (Na) and magnesium (Ma) were not included for condensate products. The XRF analysis of the ML raw materials shows significantly increased contents of sodium, silicon and calcium compared to the reference material (Figure 10A). These elements are used in plastics manufacturing as additives with the function of for example, filler and reinforcing material [25]. However, the cause of the increased concentration in ML could be explained by the origin from sand and soil before collection. As shown in Figure 12 almost all elements are concentrated in the pyrolysis char. Amounts of potential harmful substances were found in low quantities in all the samples, peaking at 0.015 wt% lead for ML Sylt raw material and 0.35 wt% in the char, which had been expected higher for material containing fishing gear. The lead is partly built into the ropes as a weighting for the nets to improve sinking in the water. Previous studies showed lead contents of up to 0.75 wt% in the raw material and 2.25 wt% in the char of ML [26]. Norderney condensate and mulch foil char evince unexpected measurements, as contents of iron, chromium and nickel show the same ratio, which indicates cross-contamination with stainless steel in the sample pre-treatment.

**Figure 10.** Mass fraction of selected elements in raw material determined by X-ray fluorescence (XRF) (n = 2). (**A**) shows detected elements with the highest mass fractions. (**B**) shows heavy metals mentioned in the German ordinance for waste incineration (17. BImSchV).

**Figure 11.** Mass fraction of selected elements in pyrolysis condensate determined by XRF (n = 2). (**A**) shows detected elements with the highest mass fractions. (**B**) shows heavy metals mentioned in the German ordinance for waste incineration (17. BImSchV).

Combining the results of pyrolysis product analysis with those of gas measurements allows examining the distribution of selected elements. Carbon and hydrogen were balanced according to their recovery rate in the product fractions, results are presented in Figure 13. Comparing elementary content in raw material with the accumulated content in pyrolysis products shows minor deviations in relation to the methodically inevitable imprecisions especially in pyrolytic processes. Carbon distributes across all three product fractions while hydrogen almost completely releases from the solid phase. In both elemental balances the gas fraction represents by far the highest fraction, due to slow cooling rates and thus high residence time in the gas phase allowing further cracking of condensable components into permanent gases. The distribution of elements across the product fractions is similar in all trials conducted.

**Figure 12.** Mass fraction of selected elements in pyrolysis char determined by XRF (n = 2). (**A**) shows detected elements with the highest mass fractions. (**B**) shows heavy metals mentioned in the German ordinance for waste incineration (17. BImSchV).

**Figure 13.** Elementary balances of pyrolysis products compared to analysed content in raw material. (**A**) carbon distribution across product fractions in wt%. (**B**) hydrogen distribution across product fractions in wt%.

To rank the transferability of results from the different experimental setups the distribution of pyrolysis products from the retort furnace are compared to the residual mass from the TGA setup. The latter one represents the char fraction while the mass loss is ascribed to the volatile compounds as a sum parameter consisting of condensate and gas as they correspond to the retort furnace. As presented in Figure 14 the char yields of the materials show very similar results in retort furnace and TGA setup. Both experiments used low heating rates of 8–10 ◦C/min and 30 ◦C/min thus returning virtually equal amounts of char. Target temperature (700 ◦C for retort furnace, 850 ◦C for TGA) appears to have an insignificant influence. Overall results validate both experimental setups used, suiting them for general material characterisation.

**Figure 14.** Comparison of product yields from pyrolysis trials in retort furnace and TGA in wt%.

The three pyrolysis condensates of this study are unexpectedly of wax-like consistency. Previous studies on ML pyrolysis in the same retort furnace produced liquid condensates of honey-like consistency for the temperature investigated [26]. The current focus of the upcycling-efforts therefore is on requirement 1 (Section 2.7) how to make the condensate bioavailable by introduction into the liquid culture. Solvents and biosurfactants are tested for the ability to dissolve the condensate and their biocompatibility. As shown in Figure 15C,D, simply supplementing the liquid medium with the condensate and rhamnolipids, as in the study of Guzik et al. [20], will not be sufficient with the condensates of this study. One promising solvent is ethyldecanoate (Figure 15A), which has recently been shown to not affect the growth of *Pseudomonas putida* KT2440 but to be suitable for biotechnological application [27].

**Figure 15.** Pyrolysis condensate in the solvents (**A**) ethyldecanoate; (**B**) hexadecane; and water supplemented with each 10% of the emulsifiers; (**C**) rhamnolipids and (**D**) Tween80 (Picture: Kristina Bitter).

Biocompatibility is tested with several strains of *Pseudomonas* but also other metabolic versatile organisms with affinity towards hydrocarbons as *Rhodococcus opacus* who harbours

genes for the *Alk*B monooxygenase [28] and known alkane-degrader *Alcanivorax borkumensis* [29] fulfilling requirements (2) and (3) (Section 2.7). In the following step, potential growth of microbes on pyrolysis condensate will be monitored via CO2-development. When bacterial growth on the mixed-plastic pyrolysis condensate can be confirmed, the production of valuable compounds will be investigated.

### **4. Conclusions**

The investigations performed and results presented refer to the collected material of the ML Norderney and ML Sylt. Since the collection was carried out randomly and without standardized conditions, it cannot be considered representative of all the waste washed up on the beach on Norderney and Sylt or other locations. The ML presented in this study rather shows the randomness of the composition of waste samples.

The results presented in this study demonstrate the principal possibility of thermochemical treatment of ML. The determined fluctuations in reaction temperatures, ash content or pollutants do not pose any problems for thermal treatment plants with state of the art process and emission control systems. Process selection depends on legislative incentives and possible cascaded recycling of intermediate and end products.

Conclusively, the necessity of tackling the pollution of the World Ocean with ML becomes obvious. Pyrolysis as part of a chemical recycling process needs extensive pretreatment to remove impurities, for example, by crushing and sorting. The obtained condensable fractions present a valuable product for material recovery. Combustion of pyrolysis gases supplies the necessary process heat. The pyrolysis char is contaminated with pollutants such as heavy metals and must be post-processed safely in waste incineration plants for energy recovery and discharge of pollutants.

The use of condensable fractions for biotechnological upcycling presents an innovative approach to access the material potential of plastics unsuitable for mechanical recycling sustainably. Biotechnological use of pyrolysis condensate poses many challenges but versatile microbes as *Pseudomonas* might be able to meet them. They offer opportunities to produce valuable molecules for further industrial applications, thus reintegrating plastic waste fractions into a circular economy.

**Author Contributions:** J.H.: Project Administration; Conceptualization, Methodology, Writing— Original Draft, Review and Editing; K.S.: Conceptualization, Methodology, Writing—Original Draft, Review and Editing; S.L.: Writing—Original Draft, Review and Editing; J.P.: Writing—Original Draft, Review and Editing; K.B.: Writing—Original Draft, Review and Editing; L.M.B.: Supervision; P.Q.: Supervision, Writing—Review and Editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research conducted by Institute of Applied Microbiology, Aachen Biology and Biotechnology was funded by iMULCH (FKZ: EFRE-0801169). Pacific Garbage Screening GmbH has received funding from the European Union's Horizon 2020 research and innovation programme under grant agreement No 870294.

**Acknowledgments:** The authors gratefully acknowledge the volunteers of the Wadden Sea World UNESCO World Natural Heritage Wadden Sea Visitor Centre Norderney and the Centre of Experience Natural Forces Sylt gGmbH who collected the material and provided the material to Pacific Garbage Screening GmbH (Future everwave) for research purposes.

**Conflicts of Interest:** The authors declare no conflict of interests.

### **References**

