Chemical Engineering Journal 472 (2023) 145007 Contents lists available at ScienceDirect Chemical Engineering Journal journal homepage: www.elsevier.com/locate/cej Polyhydroxyalkanoates production by an advanced food-on-demand strategy: The effect of operational conditions Antonio Mineo a, Laura Isern-Cazorla b, Carla Rizzo c, Antonio Palumbo Piccionello c, María Eugenia Suárez-Ojeda b, *, Giorgio Mannina a a Engineering Department, Palermo University, Viale delle Scienze ed. 8, 90128 Palermo, Italy b GENOCOV Research Group, Department of Chemical, Biological and Environmental Engineering, School of Engineering, Universitat Autònoma de Barcelona, Escola d’Enginyeria-Edifici Q, Campus UAB, 08193 Bellaterra, Barcelona, Spain c Department of Biological, Chemical and Pharmaceutical Sciences and Technologies, Palermo University, Viale delle Scienze Ed.17, 90128 Palermo, Italy A R T I C L E I N F O A B S T R A C T Keywords: Despite the increasing number of studies related to the polyhydroxyalkanoate (PHA) production from sewage Automated accumulation software sludge of wastewater treatment plants, there is still a gap in the correlation between the operational conditions, Bioplastic production such as the organic loading rate (OLR), and the intracellular polyhydroxyalkanoate (PHA) content, productivity Biopolymers extraction and final recovery of the polymer. Therefore, this work aims to provide experimental data on PHA productivity Water resource recovery and purity in view of scaling up the process to an industrial level taking into account process parameters. In view of that, three OLR were applied during the selection of PHA-accumulating bacteria in sewage sludge. Then, the biomass was harvested and subjected to batch accumulation experiments at two organic loads per dosage by employing a tailor-made software to adopt an automated feed-on-demand strategy, which allowed for 30–56 h of accumulation tests in stand-alone mode. Finally, an improved protocol for PHA extraction has been applied. Experimental results show that the maximum PHA content (60% w/w) was achieved using the highest organic load per dosage during the accumulation test with the biomass selected at the highest OLR (1.8 g COD L− 1 d− 1). Also, the extraction protocol efficiency was proven with four samples with different PHA content, achieving recovery yield as high as 78 ± 3 % with a purity of 89 ± 2 %, thus demonstrating that the adopted strategy might be beneficial for industrial use. 1. Introduction of the PHA production is generally based on a four-step process: i. the acidogenic fermentation of sewage sludge to produce volatile fatty acids Being nowadays globally dependent on fossil-fuel-based plastic ma­ that serve as a carbon source for MMC; ii. the selection of PHA producing terials, biodegradable polymers are assuming an increasingly significant microorganisms through the application of a feast-famine process; iii. role in the challenge to replace conventional plastics by decoupling their the enhancement of the amount of PHA produced, commonly referred to production from fossil fuel extraction [1]. Among the different types of as “accumulation” and iv. the PHA extraction by disrupting the cells polymers, polyhydroxyalkanoates (PHA) have garnered growing atten­ employing tailored protocols [3–8]. Despite the recent research on PHA tion as they possess similar properties to conventional fossil-fuel-based production using sewage sludge [4,5], few works still focus on plastics and they can be bioproduced by various microorganisms [2]. improving the PHA accumulation and downstream process. However, In addition, PHA production by mixed microbial cultures (MMC) can be biopolymer content in the biomass, productivity and purity are pivotal envisaged using the organic matter present in wastewater as a substrate. parameters for achieving sustainability in the overall process [5]. In this The utilisation of MMCs and wastewater are crucial factors that have context, Bengtsson et al. [8] stated that the minimum PHA content in the contributed to the rising popularity of PHA, particularly when consid­ biomass necessary for a commercially sustainable production process ering the circular economy approach, since sludge disposal is the main (referred to as the cut-off point) was equal to 0.4 g PHA g− 1 VSS (Volatile economic cost within the management of wastewater treatment plants, Suspended Solids). However, as far as the authors are aware, no existing which is still based on a linear economy model [3]. The implementation protocol has been found to consistently enable the production of such a * Corresponding author. E-mail address: mariaeugenia.suarez@uab.cat (M.E. Suárez-Ojeda). https://doi.org/10.1016/j.cej.2023.145007 Received 28 April 2023; Received in revised form 21 July 2023; Accepted 23 July 2023 Available online 24 July 2023 1385-8947/© 2023 The Author(s). Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by- nc-nd/4.0/). A. Mineo et al. Chemical Engineering Journal 472 (2023) 145007 significant content of PHA from sewage sludge. [3,7]. The lack of con­ Table 1 sistency in achieving a substantial quantity of PHA, coupled with the Details of selection step and accumulation tests. manual execution of the accumulation step, arguably the primary stage OLR applied during the Accumulation Organic load per dosage (g in PHA production, significantly hampers the potential scalability of the selection step (g COD L− 1 d− 1) tests COD L− 1 dosage− 1) process. [8]. Usually, when pilot plant experiments are performed 0.8 T1 7.5 [9–11], there is often a lack of emphasis on other parameters such as the T2 15.0 PHA productivity, that can be prominent for the industrial PHA pro­ 1.3 T3 7.5 duction process. T4 15.0 1.8 T5 7.5 Finally, the PHA extraction is another crucial step of the process T6 15.0 since it can contribute to more than 50% of the polymer production cost. The main current recovery processes are energy-intensive and rely on *COD: Chemical Oxygen Demand. the use of chemicals, such as chloroform, which are efficient but still potentially hazardous to the environment, particularly at the industrial (Corio CD-CB6) maintained at 20 ◦ C. The biomass (3 L) was harvested level [12–19]. Numerous studies have been published focusing on from the selection reactor, washed with tap water, and let to settle down research of less expensive and more environmentally friendly down­ to discard 1 L of supernatant. Then, it was mixed with 1 L mineral me­ stream processing methods for industrial implementation [20–24]. On dium without NH4Cl. Finally, the biomass was left aerated and stirred the one hand, Colombo et al. [19] successfully implemented a non-ionic overnight before starting the experiments to ensure endogenous condi­ surfactant for dissolving the non-polymeric cellular mass (NPCM), fol­ tions. The carbon source used was the same as the one used during the lowed by extraction using dimethyl carbonate. Results showed increased selection step and the feeding was automatically controlled by a tailor- recovery when the non-ionic surfactant pre-treatment was employed. made software (PHACC). For each OLR used in the selection step, two Nevertheless, despite the high polymer purity (>90%), the recovery different accumulation experiments (1.5 L each) were carried out to efficiency was less than 60%, which highlights how challenging it is to evaluate the influence of two different organic loads per dosage perform an environmentally friendly PHA extraction without compro­ (Table 1). Tests were stopped when no substrate consumption was mising the overall performance. On the other hand, as suggested by detected after about 2 h from the last injection. Depending on the effect Kurian et al. [16], the use of green solvents such as ethylene carbonate of the different organic loads per dosage evaluated in this work, the tests on large-scale production can be regarded as another alternative since lasted 24 to 56 h. During the experiments, samples (ca. 20 mL) were this environmentally friendly approach achieved excellent product re­ collected and immediately mixed with formaldehyde (36% w/v in H2O, covery and purity, reaching 98% and 99%, respectively. Sigma Aldrich) to stop the biological activity. All the samples were As far as authors are aware, studies have yet to be conducted on the centrifuged (40 min, 4046 RCF) and the obtained pellets were stored role played by the organic loading rate (OLR) in the biomass selection overnight at − 80 ◦ C and then lyophilised. step and by the organic load per dosage during the accumulation step. Building upon the identified research gap, the objective of this study 2.3. Automatically controlled pulsate-feeding was to evaluate the impact of biomass selection under three different OLR and the influence of using two different organic loads per dosage The accumulation tests were run with an automatic pulsate feeding employed during the accumulation step. To achieve this, a two-stage based on a feed-on-demand strategy [21] controlled by tailor-made process was implemented, consisting of PHA accumulation batch tests software implemented in a Raspberry Pi 4 model B. The automatic conducted automatically using a tailor-made software, to fully automate control was adapted from the literature [20] and implemented in the the process, followed by PHA extraction. An advanced extraction pro­ software. The feeding inputs were controlled by continuously moni­ tocol previously described by Mannina et al. [17] was modified to boost toring the DO. At the same time, two pumps (Watson-Marlow Qdos 30, the PHA obtention. One of the gaps in current PHA production at the Falmouth, United Kingdom) were used to feed the synthetic substrate in industrial scale is the automation of the accumulation process as few the two reactors simultaneously. The software functioning is described studies [20,21] have been published. by Eq. (1): DOt − DOm 2. Materials and methods δ= ⩾Tf (1) DOr − DOm 2.1. Selection step for enriching the sewage sludge in PHA-accumulating bacteria • DOr is the initial background respiration rate of the source biomass. • DOm is a weighted average of the minimum DO level monitored. The reactor used for the selection step was part of the new pilot plant • DOt is the current DO value read by the probe. established at the Water Resource Recovery Facility at Palermo Uni­ • δ is the trigger value to feed the substrate. versity [22,25]. The strategy adopted for biomass selection was aerobic dynamic feeding [13,25] and a synthetic substrate was used to After performing different tests and analysing data available from comprehensively compare three OLR and two organic loads per dosage. Mannina et al. [17] two new conditions were added: The carbon source was composed of acetic acid (70%) and propionic acid (30%) (Sigma Aldrich), as reported in the literature [10], while • DOm threshold which is the difference DOt - DOm ; to perform the KH2PO4 and NH4Cl (Sigma Aldrich) were added as a mineral medium. feeding, this value must be ≥ Tf . Table 1 reports the ORL applied during the selection step. • Tf is the threshold value for feeding or not, this value range from 0.01 to 0.3 and it is related to the type of microorganisms. 2.2. Accumulation step for improving PHA contents in selected biomass • The number of samples allows deciding the number of DO values to build the weighted average DOm . The PHA-accumulation tests were run in duplicated in fed-batch mode with pulsating substrate feeding. Fig. 1 shows a schematic rep­ The DOm threshold allows identifying the slope in the DO shifting, resentation of the experimental set-up. Two 1.5 L glass reactors were thus letting the software acknowledge when an increase/decrease of the used, which were magnetically stirred and aerated. The reactors were biomass respiration rate is happening. Introducing this condition pre­ equipped with probe ports to monitor the temperature and the dissolved vented an error in feeding the substrate when the biomass has not yet oxygen (DO, WTW). The temperature was controlled using a thermostat consumed all the substrate. The number of samples used to build the 2 A. Mineo et al. Chemical Engineering Journal 472 (2023) 145007 Fig. 1. Schematic representation of the experimental set-up for the accumulation tests. weighted average allows a more precise calculation when the software 2.6. Calculations identifies the end of the feast. These conditions and values were found to fit more than ten different accumulation tests performed with a syn­ The calculation of PHA amount in microbial cells, yield, purity and thetic substrate under different conditions. recovery were calculated as reported [17]: MassPHA 2.4. PHA extraction protocol from biomass PHA amount in microbial cells = (3) Masslyophilized biomass The extraction protocol proposed by Mannina et al. [17] was adopted Extracted polymer weight using four samples collected during T6 at 4.6, 7.2, 25.5 and 56.2 h (T6- PHA yield = (4) Biomass weight S1, T6-S2, T6-S3 and T6-S4, respectively). The extraction was performed in triplicate. Briefly, 50 mg of lyophilised biomass was weighted in a MassPHA glass tube and mixed with 5 mL of NaOCl (15 % in H2O). The tube was PHA purity = (5) Massextracted polymer then mixed (11 RCF) and heated (75 ◦ C, 1 h) in an Eppendorf block heater (ThermoStat™ C). After centrifugation (1694 RCF, 10 min), the PHA yield × PHA purity PHA recovery = (6) pellet was washed twice with MilliQ Grade water (5 mL). Afterwards, PHA amount in microbial cells the pre-treated biomass was mixed (11 RCF, 75 ◦ C, 3 h) with an PHA concentration as g PHA g− 1 VSS and PHA storage yield were ammonium laurate solution prepared by adding 0.75 mL of ammonia (1 calculated as reported by Conca et al. [13]. M, Sigma Aldrich, Merck, Darmstadt, Germany) to a 100 mg lauric acid (Sigma Aldrich) suspension in 3 mL of water. The pellet obtained after the centrifugation was dried at 60 ◦ C overnight. 2.7. Biopolymer characterisation 2.5. Analytical methods The biopolymer structure and percentage of each monomer were confirmed by 1H NMR analysis performed on a Bruker magnetic reso­ The total PHA concentration was calculated as the sum of poly(3- nance spectrometer operating at 400 MHz. Extracted biopolymers (10 hydroxy)butyrate (PHB) and poly(3-hydroxy)valerate (PHV) mono­ mg, section 1.4) were dissolved in CDCl3 to carry out the above analysis. mers concentrations inside the lyophilised biomass, measured by NMR spectra were recorded at 25 ◦ C and solvent residual peak was used adopting the protocol described by Werker et al. [23]. Briefly, butanol as reference. Molecular mass (Mw), molecular mass distribution (Mn), (1.5 mL) and hydrochloric acid (0.5 mL) were added to the lyophilised and polydispersity index (PDI; Mw/Mn) were determined by means of samples, previously weighted, and the tubes were incubated at 100 ◦ C SEC-MALS performed on an HPLC-SEC Agilent 1200 series using a for 8–20 h. After, 2.5 mL of hexane and 4 mL of MilliQ Grade water were Phenogel organic SEC column (104 Å with cut-off 5 to 500 kDa), iso­ added to the test tubes and then vortex mixed. The organic phase was cratic runs were performed for 25 min with 100% THF as eluent and a transferred to the centrifuge tube and another aliquot of MilliQ Grade flow rate of 0.5 mL/min. Biopolymers (1 mg) were dissolved in 1 mL of water was added. After the centrifugation (350 RCF, 10 min), the THF and filtered before injection. organic phase was extracted, filtered through 0.22 µm filters and finally collected into gas chromatography (GC) vials. The analysis of the sam­ 3. Results and discussion ples was carried out as reported by Montiel-Jarillo et al. [24]. 3.1. Accumulation experiments The accumulation experiments were conducted until the substrate 3 A. Mineo et al. Chemical Engineering Journal 472 (2023) 145007 fed was no longer consumed (generally after 30 h). To prevent PHA highlights the significant influence of the organic load per dosage on the accumulation inhibition by excessive COD injection [16], 5 and 10 mL of accumulation process. It indicates that a higher organic load per dosage substrate were fed in 3 min for T1, T3, T5 and T2, T4 and T6, resulted in a greater amount of PHA produced within the same time respectively. frame, even with fewer substrate injections being performed. The soft­ Fig. 2 shows the final PHA content achieved for each accumulation ware implementation allowed to continuously run and monitor the test and the PHB:PHV ratio calculated by analysing the lyophilised accumulation test up to 56 h during T6. The feed-on-demand strategy biomass, as reported in the analytical methods (section 2.5). Simple with an automatic control increased the final PHA content up to 20% w/ main effects analysis showed that both ORL and the organic load per w compared to the regular 6–10 h accumulation test performed without dosage shows a statistically significant influence on PHA content (p = the stand-alone injection [31]. These results prove the feasibility of 0.018, p = 0.036, respectively). T5 and T6 (ORL of 1.8 g COD L− 1 d− 1 implementing this software in a scaled-up process: the accumulation during the selection step) led to the highest amount of PHA concentra­ could be easily managed and monitored while maintaining its reliability tion (42 and 60 % w/w, 0.54 and 0.77 g PHA g.1 VSS respectively) due to the ability to acknowledge biomass’ respiration rate shift. Also, within the biomass cells compared to the other tests. The biomass har­ thanks to the number of parameters that can be set up and the flexibility vested from the selection reactor working at an ORL of 1.3 g COD L− 1 to make changes in them (trigger value, DOm threshold and several d− 1 reached 20 (T3) and 31 (T4) % w/w (0.24 and 0.36 g PHA g.1 VSS, samples for DOm), the software can be used regardless of the substrate for T3 and T4, respectively) while the accumulation performed with the and the selection strategy adopted, both at bench and pilot plant scale. biomass selected at an ORL of 0.8 g COD L− 1 d− 1 exhibited the lowest Fig. 4 shows the PHA productivity, expressed as g PHA L− 1h− 1, and PHA contents (17 and 27 % w/w, 0.19 and 0.29 g PHA g.1 VSS for T5 and the storage yield in COD equivalents of PHA per COD substrate equiv­ T6, respectively). When comparing accumulation tests with the same alents. Fig. 4A points out the key role played by the OLR applied in the organic load per dosage (T1, T3 and T5, T2, T4 and T6), the higher the selection step and the organic load per dosage used in the accumulation OLR applied during the selection step, the more effective is the step, which strongly influences the PHA productivity peak as well as the enhancement of the microorganisms’ PHA storage and production final PHA content in the biomass. Specifically, at the three OLRs capability [23]. employed (0.8, 1.3, and 1.8 g COD L− 1 d− 1) during the biomass selection Moreover, high OLR in the selection step enhances PHA production stage, the peak productivity values were of 0.40 (T1), 0.20 (T3), and mostly by favouring PHV synthesis, as can be seen by the PHB:PHV ratio 0.72 (T5) g PHA L− 1h− 1, while the corresponding PHA biomass contents [13]. The amount of PHV produced was affected by several factors, such were 18%, 20%, and 42% w/w as shown in Fig. 2 for T1, T3, and T5, as the feeding strategy or the organic load per dosage [25]. A continuous respectively. Regarding the organic load per dosage, results indicate regime of increase and decrease of feed results in higher HV content (8% productivity values of 0.4 and 0.6 g PHA L− 1h− 1, with PHA biomass of the increase) compared to an hourly pulse feeding regime when a contents of 18% and 28% w/w (Fig. 2) for T1 (7.5 g COD L− 1 dosage− 1) synthetic substrate is applied [26–28]. Also, higher organic load per and T2 (15.0 g COD L− 1 dosage− 1) at the first OLR used in the selection dosage, due to the use of propionic acid in the substrate, enhanced the (0.8 g COD L− 1 d− 1). Similar trends were observed for the other OLR: T4 PHV production [15]. Also, Vazquez et al. [29] obtained a similar HB: and T6 reached productivity peaks of 0.23 and 0.26 g PHA L− 1h− 1, HV ratio since the carbon source composition (acetate, butyrate:propi­ respectively, with final PHA contents of 31% and 60% w/w (Fig. 2). onate, valerate) was similar to the one used in this work. Instead, when a Also, increasing the OLR enhanced the microorganism storage more complex substrate is used [28], the synthesised PHV amount capability as reported in Fig. 4B. Moreover, simple main effects analysis drastically decreases. The PHV:PHB ratio in T6 could suggest that this showed that both, ORL and the organic load per dosage have a statis­ polymer may be a promising alternative to PHB homopolymers, espe­ tically significant effect on storage yield (p = 0.004 and p = 0.184, cially regarding its industrial applications [16,29,30]. respectively). This means that the biomass could increase its PHA con­ Fig. 3 shows the DO profile and PHA contents during T5 and T6. In tents from less than 1% (T5) up to 32% w/w (T6), but at the expense of more detail, over the course of the initial 30 h, T5 exhibited a total of 17 decreasing the productivity peak by 43% comparing T5 (0.7 g PHA DO peaks (Fig. 3A), while T6 had 11 DO peaks (Fig. 3B). This outcome L− 1h− 1) to T6 (0.3 g PHA L− 1h− 1) and of employing more time for Fig. 2. Final PHA content and HV:HB polymer ratio. Filled bars correspond to PHB. Striped bars correspond to PHV. 4 A. Mineo et al. Chemical Engineering Journal 472 (2023) 145007 Fig. 3. PHA content and DO profile during T5 (A) and T6 (B). accumulating PHA (Fig. 4A) [32]. Considering the same OLR used, when during T6 (Table 2). As described in the materials and methods section, a higher organic load per dosage is applied, the productivity peak is PHA content, purity and recovery were calculated. The protocol used always higher [33], the same as the amount of PHA produced. There­ herein comprises a two-step solvent extraction: NaClO pre-treatment fore, if the target is to achieve the highest productivity, it is advised to and NPCM destruction using an organic surfactant (ammonium laure­ use low OLR in the selection step, otherwise, high OLR should be used if ate). Table 2 shows an overall high purity (89 ± 2 %) and recovery (78 the target is to achieve the highest amount of PHA. Regarding the ± 3 %), demonstrating that the protocol’s efficiency is not influenced by organic load per dosage, results indicate that using high values is rec­ the PHA content in the lyophilised samples if higher than 32 % w/w ommended but, at the same time, substrate inhibition must be avoided (cut-off value). [34]. Table 3 shows the extraction efficiency obtained in this work These results highlight crucial parameters to consider when con­ compared to other literature studies. As it can be observed, high-purity ducting a pilot or full-scale accumulation test. Still, few works focus on results were obtained through the NPCM dissolution, as reported by optimising the organic load per dosage in the accumulation test while other studies [17]. Compared to the literature, Lorini and colleagues [5] considering the PHA productivity. These parameters could be used in an reported high PHA purity and recovery percentages (>80%) applying energy-mass balance to express, at an industrial level, where it is more alkaline treatment with NaOH, however, they obtained lower recovery convenient to stop the accumulation process and extract the stored PHA. yields using SDS surfactant (<50%) as a solvent for the extraction. Although the sample preparation and extraction process can seem highly energy-consuming compared to other works [4,5,9,25,35,36], the re­ 3.2. Extraction protocol and biopolymers characterisation sults consistently demonstrate the feasibility of applying this protocol for samples containing different amounts of PHA with different The extraction protocol [15] was applied to four samples collected 5 A. Mineo et al. Chemical Engineering Journal 472 (2023) 145007 standard. Signals corresponding to the protons of the PHB units are labelled as 1–3 while signals corresponding to PHV units are labelled as 4–7. Thus, deshielded peaks above 5 ppm (labelled as 2 and 5) are indicative of C-H protons directly linked to an electronegative nucleus such as O. The multiplicity of the signals indicates that the peak at 5.25 ppm is a sextuplet integrating for 1 proton, so this peak can be assigned to the C-H of the PHB (labelled as 2) as it is linked to a CH2 and CH3. On the other hand, the signal at 5.15 ppm is a quintuplet and it can be assigned to the C-H of the PHV (labelled as 5) as it is linked to two CH2. Signals around 2.5 ppm are assigned to CH2 at the alfa position with respect to a carbonyl group. These signals are overlapped in PHB and PHV. In the aliphatic region, a signal around 1.3 ppm (labelled as 3) is assigned to CH3 linked to C-H of the PHB unit and therefore is a doublet, while two signals at 1.6 ppm and 0.9 ppm (labelled as 6 and 7) were assigned to the ethyl group of the PHV units and their multiplicities, as expected, are (broad) quintuplet and triplet, respectively. Obtained PHA signals are in good agreement with the literature [37]. The HB-HV monomers percent present in the obtained PHA samples were calculated by comparison of the intensity, hence of integrals’ area, of diagnostic C-H peaks 2 and 5, respectively, as the integrals area of these two peaks is directly related to the percentage of PHB and PHV present in PHA sample. The NMR analysis is performed in the extracted polymer samples and not on the lyophilised biomass. In particular, the commercial PHA confirmed a monomer composition of 92% in PHB and 8% in PHV, while the composition of extracted polymers for T6-S4, T6-S3, T6-S2 and T6- S1 are stated in Table 4. In general, the PHV content of extracted PHA is larger than the one observed in the commercial sample as can be inferred from the higher intensity of PHV signals (4–7) compared to the commercial one (Fig. 5). From NMR analysis, sample purity around 90% was confirmed and agrees with previously reported data [38–40]. Moreover, from the observed peaks, could be assessed that impurities are ascribed to the presence of residual lauric acid from the extraction process. Molecular weight data obtained via SEC-MALS analysis revealed the Fig. 4. PHA productivity as a function of PHA content (A) and storage yield for presence of two peaks in each sample, the first one with larger molecular all the experiments (B). weight indicative of the polymer and the second one with smaller mo­ lecular weight probably indicative of the impurity of lauric acid quan­ Table 2 tified by NMR analysis (Table 5). Molecular weight data agree with PHA content, recovery and purity percentage for samples collected from T6. previously reported biopolymers even if the value is lower than recent studies with different extraction procedures [9]. PDI was close to 1, Sample Sampling time PHA content (% w/ Purity Recovery (h) w) (%) (%) indicating a homogeneous polymer, a pivotal feature for other polymer’s applications [15]. These results, despite the low molecular weights are a T6-S1 4.6 31.6 ± 0.3 90 ± 1 81 ± 2 further demonstration of the reliability of the extraction protocol which T6-S2 7.2 42.8 ± 0.6 90 ± 1 74 ± 1 T6-S3 25.5 55.7 ± 0.8 87 ± 1 78 ± 2 can be further optimised and implemented it for pilot-scale experiments T6-S4 56.2 60 ± 1 89 ± 1 78 ± 1 towards automation. 4. Conclusions Table 3 Extraction efficiency compared to other literature studies. High OLR in the selection step (1.8 g COD L− 1 d− 1) and high organic load per dosage (15.0 g COD L− 1 dosage− 1) in the accumulation test Sample PHA content range (%w/ Purity range Recovery range w) (%) (%) were the best conditions, resulting in the highest PHA content (60% w/ w) and the highest HV:HB ratio (0.66). These results were obtained by Present 32–60 87–90 74–81 study using a newly tailor-made software which applies a feed-on-demand [15] 35–54 42–100 57–82 strategy to perform the accumulation tests in stand-alone mode. This [17] 50 57–92 26–64 allowed to perform 56 h long accumulation tests, thus achieving high [35] 42–60 75–100 52–100 final PHA content (up to 60% w/w). Also, it was shown how high OLR [36] 55–60 56–82 48–74 and organic load per dosage decreased the PHA productivity during the test, despite increasing the biomass storage capability. Finally, an proportions between PHB and PHV monomers. extraction protocol was tested with different PHA content samples To confirm the polymer obtainment and composition, PHA recov­ demonstrating that the recovered polymer purity (89 ± 2 %) and re­ ered after extraction in T6 was dissolved in deuterated chloroform for covery yield (78 ± 3 %) are not affected by the PHA content in the characterisation through 1H NMR measurements. Fig. 5 shows the sample, thus proving the reliability and consistency of the overall representative spectrum of recovered PHA compared to a commercial strategy which can be beneficial for industrial use. 6 A. Mineo et al. Chemical Engineering Journal 472 (2023) 145007 Fig. 5. 1H NMR spectra in CDCl3 of commercial PHA (bottom in red) and extracted PHA (upper in green). Representative PHA structure and peak assignment were reported in the inset (sample T6-S4). Table 4 Table 5 PHA composition, as determined using 1H NMR. Retention time (RT), Mw, Mn and PDI. Sample %PHB %PHV Sample RT time (min) Mw (Da) Mn (Da) PDI T6-S4 39 61 T6-S4 17.13 18,998 14,871 1.278 T6-S3 38 62 21.43 765 763 1.002 T6-S2 50 50 T6-S1 18.95 53,166 32,494 1.636 T6-S1 31 69 21.48 1737 1733 1.002 7 A. Mineo et al. Chemical Engineering Journal 472 (2023) 145007 Declaration of Competing Interest [10] G.-Y.-A. Tan, C.-L. Chen, L. Li, L. Ge, L. Wang, I.M.N. Razaad, Y. Li, L. Zhao, Y. Mo, J.-Y. Wang, Start a research on biopolymer polyhydroxyalkanoate (PHA): a review, Polymers (Basel). 6 (2014) 706–754, https://doi.org/10.3390/polym6030706. The authors declare that they have no known competing financial [11] L. Lorini, G. Munarin, G. Salvatori, S. Alfano, P. Pavan, M. Majone, F. Valentino, interests or personal relationships that could have appeared to influence Sewage sludge as carbon source for polyhydroxyalkanoates: a holistic approach at the work reported in this paper. pilot scale level, J Clean Prod. 354 (2022), 131728, https://doi.org/10.1016/j. jclepro.2022.131728. [12] N. Frison, M. Andreolli, A. Botturi, S. Lampis, F. Fatone, Effects of the Sludge Data availability Retention Time and Carbon Source on Polyhydroxyalkanoate-Storing Biomass Selection under Aerobic-Feast and Anoxic-Famine Conditions, ACS Sustain Chem Eng. 9 (2021) 9455–9464, https://doi.org/10.1021/acssuschemeng.1c02973. Data will be made available on request. [13] V. Conca, C. da Ros, F. Valentino, A.L. Eusebi, N. Frison, F. Fatone, Long-term validation of polyhydroxyalkanoates production potential from the sidestream of Acknowledgements municipal wastewater treatment plant at pilot scale, Chemical Engineering Journal. 390 (2020), 124627, https://doi.org/10.1016/j.cej.2020.124627. [14] C. Collet, A.A. Vaidya, M. Gaugler, M. West, G. Lloyd-Jones, Extrusion of PHA- This work was funded by the project “Achieving wider uptake of containing bacterial biomass and the fate of endotoxins: A cost-reducing platform water-smart solutions—WIDER UPTAKE” (grant agreement number: for applications in molding, coating and 3D printing, Mater Today Commun. 33 (2022), 104162, https://doi.org/10.1016/j.mtcomm.2022.104162. 869283) funded by the European Union’s Horizon 2020 Research and [15] K.L. Macagnan, M.I. Alves, A. da S. Moreira, Approaches for Enhancing Extraction Innovation Programme, in which the last author of this paper, prof. of Bacterial Polyhydroxyalkanoates for Industrial Applications, in: V.C. Kalia (Ed.), Giorgio Mannina, is the principal investigator for the University of Biotechnological Applications of Polyhydroxyalkanoates, Springer Singapore, Singapore, 2019, pp. 389–408, https://doi.org/10.1007/978-981-13-3759-8_15. Palermo. The Unipa project website can be found at: https://wider­ [16] N.S. Kurian, B. Das, Comparative analysis of various extraction processes based on uptake.unipa.it/. AtenCenter at the University of Palermo for SEC-MALS economy, eco-friendly, purity and recovery of polyhydroxyalkanoate: a review, Int. experiments. Carla Rizzo thanks PNR next generation EU-DM737/2021- J. Biol. Macromol. 183 (2021) 1881–1890, https://doi.org/10.1016/j. CUP B79J21038330001 for funding. Laura Isern-Cazorla thanks to the ijbiomac.2021.06.007. [17] G. Mannina, D. Presti, G. Montiel-Jarillo, M.E. Suárez-Ojeda, Bioplastic recovery Erasmus + Traineeship program of the European Union for the funds from wastewater: A new protocol for polyhydroxyalkanoates (PHA) extraction received to perform the stay at Palermo University in the framework of from mixed microbial cultures, Bioresour Technol. 282 (2019) 361–369, https:// her master studies in Biological and Environmental Engineering at doi.org/10.1016/j.biortech.2019.03.037. [18] C. Kourmentza, J. Plácido, N. Venetsaneas, A. Burniol-Figols, C. Varrone, H. Universitat Autònoma de Barcelona. Dr. María Eugenia Suárez-Ojeda N. Gavala, M.A.M. Reis, Recent Advances and Challenges towards sustainable thanks to the funds received under the grant ORIGEN PID2021- Polyhydroxyalkanoate (PHA) Production, Bioengineering 4 (4) (2017), https://doi. 126102OB-I00 funded by MCIN/AEI/10.13039/501100011033 and by org/10.3390/bioengineering4020055. [19] B. Colombo, J. Pereira, M. Martins, M.A. Torres-Acosta, A.C.R.V. Dias, P.C. Lemos, ERDF A way of making Europe from the European Union. S.P.M. Ventura, G. Eisele, A. Alekseeva, F. Adani, L.S. Serafim, Recovering PHA from mixed microbial biomass: using non-ionic surfactants as a pretreatment step, Appendix A. Supplementary data Sep Purif Technol. 253 (2020), 117521, https://doi.org/10.1016/j. seppur.2020.117521. [20] F. Valentino, L. Karabegovic, M. Majone, F. Morgan-Sagastume, A. Werker, Supplementary data to this article can be found online at https://doi. Polyhydroxyalkanoate (PHA) storage within a mixed-culture biomass with org/10.1016/j.cej.2023.145007. simultaneous growth as a function of accumulation substrate nitrogen and phosphorus levels, Water Res. 77 (2015) 49–63, https://doi.org/10.1016/j. watres.2015.03.016. References [21] A.G. Werker, S.O.H. Bengtsson, C.A.B. Karlsson, Method for accumulation of polyhydroxyalkanoates in biomass with on-line monitoring for feed rate control [1] A.S. Mathuriya, J.V. Yakhmi, Polyhydroxyalkanoates: Biodegradable Plastics and and process termination. United States Patent US 8,748,138 B2, Jun. 10 (2014). Their Applications, in: L.M.T. Martínez, O.V. Kharissova, B.I. Kharisov (Eds.), [22] G. Mannina, R. Alduina, L. Badalucco, L. Barbara, F.C. Capri, A. Cosenza, D. Di Handbook of Ecomaterials, Springer International Publishing, Cham, 2017, Trapani, G. Gallo, V.A. Laudicina, S.M. Muscarella, D. Presti, Water resource pp. 1–29, https://doi.org/10.1007/978-3-319-48281-1_84-1. recovery facilities (Wrrfs): The case study of palermo university (Italy), Water [2] R. Yukesh Kannah, M. Dinesh Kumar, S. Kavitha, J. Rajesh Banu, V. Kumar Tyagi, (Switzerland). 13 (23) (2021), https://doi.org/10.3390/w13233413. P. Rajaguru, G. Kumar, Production and recovery of polyhydroxyalkanoates (PHA) [23] A. Werker, P. Lind, S. Bengtsson, F. Nordström, Chlorinated-solvent-free gas from waste streams – A review, Bioresour Technol. 366 (2022), 128203, https:// chromatographic analysis of biomass containing polyhydroxyalkanoates, Water doi.org/10.1016/j.biortech.2022.128203. Res. 42 (2008) 2517–2526, https://doi.org/10.1016/j.watres.2008.02.011. [3] G. Mannina, D. Presti, G. Montiel-Jarillo, J. Carrera, M.E. Suárez-Ojeda, Recovery [24] G. Montiel-Jarillo, J. Carrera, M.E. Suárez-Ojeda, Enrichment of a mixed microbial of polyhydroxyalkanoates (PHAs) from wastewater: a review, Bioresour Technol. culture for polyhydroxyalkanoates production: Effect of pH and N and P 297 (2020), 122478, https://doi.org/10.1016/j.biortech.2019.122478. concentrations, Sci. Total Environ. 583 (2017) 300–307, https://doi.org/10.1016/ [4] G. Pagliano, P. Galletti, C. Samorì, A. Zaghini, C. Torri, Recovery of j.scitotenv.2017.01.069. polyhydroxyalkanoates from single and mixed microbial cultures: a review, Front [25] L. Isern-Cazorla, A. Mineo, M.E. Suárez-Ojeda, G. Mannina, Effect of organic Bioeng Biotechnol. 9 (2021), https://doi.org/10.3389/fbioe.2021.624021. loading rate on the production of Polyhydroxyalkanoates from sewage sludge, [5] L. Lorini, G. Salvatori, L.N. Tayou, F. Valentino, M. Villano, Innovative strategy for J. Environ. Manage. 343 (2023), 118272, https://doi.org/10.1016/j. polyhydroxyalkanoates recovery from mixed microbial cultures: effects of aqueous jenvman.2023.118272. phase and solvent extraction on polymer properties, Chem Eng Trans. 92 (2022) [26] F. Pardelha, M.G.E. Albuquerque, M.A.M. Reis, J.M.L. Dias, R. Oliveira, Flux 529–534, https://doi.org/10.3303/CET2292089. balance analysis of mixed microbial cultures: Application to the production of [6] V. Kumar, S. Srivastava, I.S. Thakur, Enhanced recovery of polyhydroxyalkanoates polyhydroxyalkanoates from complex mixtures of volatile fatty acids, from secondary wastewater sludge of sewage treatment plant: analysis and process J. Biotechnol. 162 (2012) 336–345, https://doi.org/10.1016/j. parameters optimisation, Bioresour Technol Rep. 15 (2021), https://doi.org/ jbiotec.2012.08.017. 10.1016/j.biteb.2021.100783. [27] S. Wijeyekoon, C.R. Carere, M. West, S. Nath, D. Gapes, Mixed culture [7] G. Moretto, L. Lorini, P. Pavan, S. Crognale, B. Tonanzi, S. Rossetti, M. Majone, polyhydroxyalkanoate (PHA) synthesis from nutrient rich wet oxidation liquors, F. Valentino, Biopolymers from urban organic waste: Influence of the solid Water Res. 140 (2018) 1–11, https://doi.org/10.1016/j.watres.2018.04.017. retention time to cycle length ratio in the enrichment of a mixed microbial culture [28] M.G.E. Albuquerque, G. Carvalho, C. Kragelund, A.F. Silva, M.T. Barreto Crespo, (MMC), ACS Sustain Chem Eng 8 (38) (2020) 14531–14539, https://doi.org/ M.A.M. Reis, P.H. Nielsen, Link between microbial composition and carbon 10.1021/acssuschemeng.0c04980. substrate-uptake preferences in a PHA-storing community, ISME J. 7 (2013) 1–12, [8] S. Bengtsson, A. Karlsson, T. Alexandersson, L. Quadri, M. Hjort, P. Johansson, https://doi.org/10.1038/ismej.2012.74. F. Morgan-Sagastume, S. Anterrieu, M. Arcos-Hernandez, L. Karabegovic, [29] A. Fra-Vázquez, S. Santorio, T. Palmeiro-Sánchez, Á. Val del Río, A. Mosquera- P. Magnusson, A. Werker, A process for polyhydroxyalkanoate (PHA) production Corral, PHA accumulation of a mixed microbial culture co-exists with ammonia from municipal wastewater treatment with biological carbon and nitrogen removal partial nitritation, Chem. Eng. J. 360 (2019) 1255–1261, https://doi.org/10.1016/ demonstrated at pilot-scale, N Biotechnol. 35 (2017) 42–53, https://doi.org/ j.cej.2018.10.207. 10.1016/j.nbt.2016.11.005. [30] W. Tu, D. Zhang, H. Wang, Polyhydroxyalkanoates (PHA) production from [9] B. Yadav, S. Chavan, A. Atmakuri, R.D. Tyagi, P. Drogui, A review on recovery of fermented thermal-hydrolysed sludge by mixed microbial cultures: The link proteins from industrial wastewaters with special emphasis on PHA production between phosphorus and PHA yields, Waste Manage. 96 (2019) 149–157, https:// process: sustainable circular bioeconomy process development, Bioresour Technol. doi.org/10.1016/j.wasman.2019.07.021. 317 (2020), 124006, https://doi.org/10.1016/j.biortech.2020.124006. [31] L. de Donno Novelli, S. Moreno Sayavedra, E.R. Rene, Polyhydroxyalkanoate (PHA) production via resource recovery from industrial waste streams: A review of 8 A. Mineo et al. Chemical Engineering Journal 472 (2023) 145007 techniques and perspectives, Bioresour Technol. 331 (2021), 124985, https://doi. [37] Y. Zou, M. Yang, Q. Tao, K. Zhu, X. Liu, C. Wan, M.K. Harder, Q. Yan, B. Liang, org/10.1016/j.biortech.2021.124985. I. Ntaikou, G. Antonopoulou, G. Lyberatos, Y. Zhang, Recovery of [32] M. Singh, P. Kumar, S. Ray, V.C. Kalia, Challenges and opportunities for polyhydroxyalkanoates (PHAs) polymers from a mixed microbial culture through customizing polyhydroxyalkanoates, Indian J. Microbiol. 55 (2015) 235–249, combined ultrasonic disruption and alkaline digestion, J. Environ. Manage. 326 https://doi.org/10.1007/s12088-015-0528-6. (2023), 116786, https://doi.org/10.1016/j.jenvman.2022.116786. [33] R. Pei, G. Vicente-Venegas, M.C.M. van Loosdrecht, R. Kleerebezem, A. Werker, [38] M. Abbasi, E.R. Coats, A.G. McDonald, Green solvent extraction and properties Quantification of polyhydroxyalkanoate accumulated in waste activated sludge, characterisation of Poly(3-hydroxybutyrate-co-3-hydroxyvalerate) biosynthesised Water Res. 221 (2022), 118795, https://doi.org/10.1016/j.watres.2022.118795. by mixed microbial consortia fed fermented dairy manure, Bioresour. Technol. [34] F. Pinto-Ibieta, A. Serrano, M. Cea, G. Ciudad, F.G. Fermoso, Beyond PHA: Rep. 18 (2022), 101065, https://doi.org/10.1016/j.biteb.2022.101065. Stimulating intracellular accumulation of added-value compounds in mixed [39] B. Johnston, G. Jiang, D. Hill, G. Adamus, I. Kwiecień, M. Zięba, W. Sikorska, microbial cultures, Bioresour. Technol. 337 (2021), 125381, https://doi.org/ M. Green, M. Kowalczuk, I. Radecka, The molecular level characterization of 10.1016/j.biortech.2021.125381. biodegradable polymers originated from polyethylene using non-oxygenated [35] S. Heo, Y.-Q. Liu, Dependence of poly-β-hydroxybutyrate accumulation in sludge polyethylene wax as a carbon source for polyhydroxyalkanoate production, on biomass concentration in SBRs, Sci. Total Environ. 797 (2021), 149138, https:// Bioengineering 4 (4) (2017) 73, https://doi.org/10.3390/bioengineering4030073. doi.org/10.1016/j.scitotenv.2021.149138. [40] A. Akgül, T. Palmeiro-Sanchez, H. Lange, D. Magalhaes, S. Moore, A. Paiva, [36] T. Palmeiro-Sánchez, A. Val del Rio, A. Fra-Vázquez, J. Luis Campos, A. Mosquera- F. Kazanç, A. Trubetskaya, Characterisation of tars from recycling of PHA Corral, High-Yield Synthesis of Poly(3-hydroxybutyrate-co-3-hydroxyvalerate) bioplastic and synthetic plastics using fast pyrolysis, J. Hazard Mater. 439 (2022), Copolymers in a Mixed Microbial Culture: Effect of Substrate Switching and F/M 129696, https://doi.org/10.1016/j.jhazmat.2022.129696. Ratio, Ind. Eng. Chem. Res. 58 (2019) 21921–21926, https://doi.org/10.1021/acs. iecr.9b03514. 9