6. Innovative and Sustainable Production of Biopolymers

Plastic materials from non-renewable feedstock (e.g. petroleum) have a crucial and undeniable role in our everyday life. However, the chemically engineered durability and the slow rate of biodegradation of most of the fully fossil-based plastics [e.g. polyethylene (PE), polyethylene terephthalate (PET), polypropylene (PP), polyvinyl chloride (PVC)], enable these synthetic polymers to withstand the ocean environment and terrestrial ecosystems for several years, thus affecting organisms at multiple trophic levels [1]. Polyhydroxyalkanoates (PHAs) are bio-based biodegradable polyesters produced by aerobic bacteria from various carbon sources. According to the carbon source, different monomers are produced, whose combination provides different chemo-physical properties. Thus, PHA can have properties ranging from polypropylene to polyethylene, being at the same time biodegradable and bio-based. This brings the interest towards the application of PHAs as packaging films and containers, biodegradable carriers for controlled chemical and/or drug release, disposable items, surgical pins and sutures, wound dressings, and bone replacements [2]. With respect to other bioplastics which are already commercially available and currently produced on industrial scale, e.g. polylactic acid (PLA) or starch-based polymers (Mater-Bi^®), PHAs are still not widely used because of the high production costs, more expensive than usual petrochemical plastics (~5–6 vs. 0.5–2 $/kg) [3]. The high costs’ production is mainly due to a combination of three factors: (i) absolutely sterile operation conditions are necessary for pure microbial fermentation; (ii) the use of expensive substrates as carbon source which represents approximately 40% of total production costs; (iii) the extraction and purification process for obtaining PHAs. Thus, the commercialization of PHAs is now limited to applications with high added value. Given the need to reduce costs, a multilevel approach to tackle all bottlenecks is mandatory to develop more cost-effective processes for PHAs production. In principle, the use of microbial mixed cultures (which do not require sterility) combined with the utilization of low value substrates, as agro-industrial waste and by-products, and with safe and cheap extraction and purification methodologies would allow lower investments and operating costs for the whole production process.

This work aimed to develop a new, cost-effective, and environmentally sustainable technology for isolating PHAs from bacteria mixed cultures, by combining:

The current technology limitations are tackled by optimizing the main process steps: microbial fermentation for PHAs purification, biomass pre-treatment, PHAs extraction and purification. An effort to implement the whole process was done by improving the monitoring and automation tools. Finally, on the basis of the results obtained by the process optimization, the study of a scaled-up prototype was performed leading to the realization of PHAs production semi-pilot plant as well as the construction of photocatalytic reactors for bioreactor wastewater purification with up to 100 L capacity.

Polyhydroxyalkanoates (PHAs) are bio-based and biodegradable polyesters produced by aerobic bacteria from various carbon sources. These polymers have tuneable elastomeric/thermoplastic properties according to the monomer composition. Despite the high potential and the efforts to put forward the development of cost effective fermentative systems, the PHA production costs remain considerably high (~5–6 $/kg), thus hampering the exploitation of these biopolymers as commodity materials [4]. The main limitations could be removed by:

Mixed microbial population (MMCs) can be exploited for producing PHAs in a cheaper and more sustainable way [5]. In fact, compared to single strains (e.g. Cupriavidus necator or genetically modified Escherichia coli), MMCs do not require sterile conditions: this drastically reduces equipment costs and allows the exploitation of many cheaper substrates (e.g. wastes) thanks to the wider metabolic potential of MMCs in comparison to single strains such as waste biomass (e.g. fermented molasses), wastewater and pyrolysis products. Currently, in the literature there are few publications on PHAs recovery from MMCs, since the research effort is mainly focused on the set-up of efficient and sustainable cultivation systems. Little information is available on the characterization of PHAs extracted from MMCs and, above all, on the efficiency of the polymer recovery in the extraction step [6]. As it is well described in the literature for single strain cultures, obtaining the polymer from bacteria through a series of downstream steps (e.g. microbial biomass pre-treatment, polymer extraction, and post-treatment purification), could be challenging [7]. Indeed, it is reported that the downstream costs can represent up to 50% of the total production cost because non-recyclable chemicals/materials and high energy demands are required. However, since MMCs are more resistant than single strains, the extraction/purification of PHA can become challenging. The protocols reported in the literature [8, 9] for the extraction of PHA from MMCs include (i) an optional cell pre-treatment under acidic conditions followed by a solvent extraction (e.g. with acetone or chlorinated compounds) to dissolve the polymer or (ii) a treatment with additives that break microbial membranes enabling the release of the polymer inside the cells. The use of organic solvents helps in extracting high molecular weight PHAs but with a low yield (18–30%), whereas strong oxidants decrease the molecular weight but provide high extraction yields (close to 100%). Feed-back control strategies have already been studied for fed-batch systems, mainly to determine on-line when to feed the next pulse of substrate in the production reactor. Continuous feeding is also possible but it is used most successfully with single substrates and high cell densities [10]. Whiffin et al. [10] have proposed an algorithm that alternates continuous feeding and famine periods in order to prevent acetate accumulation and allow shorter feedback control intervals. The dissolved oxygen (DO) is closely related to the substrate concentration so that when the substrate is exhausted and oxygen consumption decreases, the DO level increases rapidly, suggesting to exploit the DO signal to control pulse feeding in the fed-batch [11, 12]. Westerberg [12] showed that inhibition of the biomass by the stored polymer is an important factor reducing the substrate and oxygen uptake rates and modelled the DO response to repeated substrate pulses. Another recent approach determines when to feed a pulse of substrate by using it as a control variable to regulate the pH [13]. Popova [14] designed a Kalman filter to estimate unmeasured variables in a mixed culture system, based on measurements of the substrate concentration. Vargas et al. [15] have explored the use of a model-based control scheme to enhance the productivity of polyhydroxyalkanoate (PHA) production in a mixed culture two-stage system. Oehmen et al. [16] operated a three-stage bioreactor with anaerobic fermentation, culture selection and PHA accumulation. Some adopted sensory systems, e.g. on-line gas chromatography in [17], strongly affect plant economic sustainability. In this work, a plant supervisor and controller has been developed exploiting the actual trend toward intelligent and interconnected production systems that integrate low cost distribute processing.

A systematic analysis and consequent improvement strategies of the PHAs production process is needed to reduce costs and limit the use of hazardous solvents for extraction and purification technologies. The main actions to enhance sustainability and cost-effectiveness are schematised in Fig. 6.1 and better described in the remaining part of this section.

Wastes such as starch, cellulose and hemicellulose deriving materials, plant oils, molasses, whey, and various industrial by-products have been proposed as potentially useful and cheap carbon sources. As a matter of fact, a wide range of agriculture and forestry wastes can be converted into carbon sources for microbial fermentation if coupled with suitable pre-treatments like thermochemical conversions (e.g. pyrolysis or hydrothermal liquefaction) or acidogenic fermentation steps. These specific pre-treatments allow to convert substrates which could not be directly utilized by microorganisms to produce PHAs, into PHAs chemical building blocks (volatile fatty acids, VFA), with a consequent abatement of the costs related to the carbon source and the accessibility to a widespread pool of substrates without any competition with food resources.

As regards microbial fermentation for PHA production, a feast (carbon excess) and famine (substrate exhaustion) approach was followed for growing MMCs. During the feast phase, bacteria able to accumulate PHA are selected since during the famine phase PHA can be exploited as energy and carbon source.

Highly skilled workforce for monitoring the system state and solving possible problems are required to ensure the achievement of satisfactory yield within an ecologically stable process. For this reason, a key aspect is to develop a complete monitoring of the process variables (physical and biological) with programmable logic controller that can readily react to the change observed (e.g. natural oscillation of the bacteria population, bulking problems, clogging) providing feedback in completely automated way and without the intervention of skilled labour. Therefore, the design and manufacturing of a compact all in one system able to convert a specific waste into ready-to-use bioplastic has been conceived. Another aspect investigated is the substitution of standard equipment series (usual approach of chemical plant) with one that can be tuned to provide multi-purpose function (e.g. bacteria growth, harvesting and extraction in the same apparatus), in order to reduce the cost (per unit of output).

As for the biomass pre-treatment, the challenge is the achievement of satisfactory yield in the PHAs extraction process despite the higher resistance of MMC than pure culture. For this reason, after the fermentation step, the bacterial broth is usually pre-treated by means of a freezing/heating approach or salt/alkaline treatment, aiming at denaturing genetic material and proteins, and thus enhancing the cell disruption and polymer recovery.

The novelty, at this point, is the application of a nanoparticles (NPs)-based pre-treatment step to weaken bacterial membranes without degrading the polymer stored inside the cells, exploiting photocatalytic TiO₂ NPs and antibacterial Ag⁰ NPs. This pre-treatment requires lower energy consumption than the traditional approaches (freezing and heating) and a lower consumption of chemicals by using supported recyclable nanomaterials.

The PHAs extraction from cultures can be done by adopting a solvent extraction approach (mainly with chlorinated solvents, highly effective in solubilizing PHA granules) or a cellular-digestion approach thus breaking cellular constituents apart with additives. The final polymer purity is higher when solvents are used, however both approaches present some environmental/economic issues such as toxicity and non-recyclability. A new and green extraction process with the aim of increasing the sustainability of the entire process was designed through: (i) an abatement of the use of toxic and hazardous chemical compounds by using green solvents; (ii) an increase in the recyclability of the chemicals used in the process and a reduction in downstream flows which must be treated; (iii) an improvement in the polymer quality ensuring the retention of mechanical properties.

Moreover, PHAs purification was required since MMCs are composed of gram-negative bacteria that can release endotoxins in the form of lipopolysaccharides, from the outer cell wall. Endotoxin can cause several diseases if introduced into the bloodstream of humans or other animals, and for this reason, in view of biomedical applications, the extracted polymer must be purified. Common methods of purification involve sodium hypochlorite or hydrogen peroxide treatment combined with the action of enzymes or chelating agents. An alternative procedure was taken into consideration for the treatment of the biopolymer contaminated by dangerous cellular residues, based on heterogeneous photocatalysis, involving photoinduced redox reactions at the surface of semiconductor ceramics. Furthermore, this approach can reduce the use of compounds that could affect the quality of the resulting biopolymer and the costs of the purification process thanks the application of a mature technology in the field of organic compounds detoxification processes.

On the basis of the sustainability principles, we ad hoc designed two differently sized plants. As starting point, we developed a working lab-scale system (1 g PHA/day) to produce completely biodegradable and biocompatible PHAs biopolymers from volatile fatty acids (VFA) obtainable from different waste sources.

Subsequently, the lab scale prototype was scaled up to 100 g PHA/day demonstration micro-plant. The system was operated for 150 days showing reliable operations with minimal maintenance. The extraction system showed to be effective in obtaining high purity PHA from bacterial slurries with acceptable yield. PHA can be recovered from MMCs by using dimethyl carbonate (DMC) or DMC plus a pre-treatment with sodium hypochlorite (NaClO). In the first case an overall polymer recovery of 66% can be achieved, with a very high polymer purity and molecular weight (98% and 1.2 MDa); in the second case, NaClO helps in increasing the recovery (82%) but decreases the molecular weight to 0.6–0.2 MDa.

The scaled-up prototype consists of a module formed by two pieces of equipment that can be added downstream to a wastewater treatment plant or to a PHA producing bioreactor. The module is composed of (i) a biological reactor which allows the production of a PHA rich sludge (25–40 g_PHA/kg_slurry) and (ii) a complete dimethyl carbonate (DMC) extraction system able to provide high purity PHA (Fig. 6.2).

The micro-plant controller has been designed in order to assure autonomous operations, with local data logging, and to support remote supervisory interaction via a standard Internet browser. In order to deliver such functionalities with a limited plant cost, a solution based on open source micro-controllers has been developed. Given the required internet connectivity and control logic, the Arduino TIAN^TM micro-controller has been chosen. The control architecture is depicted in Fig. 6.3. Different types of sensor are foreseen for each plant tank: Dissolved Oxygen (DO) and pH sensors (connected via an I2C digital bus), temperature sensors (connected via One-wire digital bus). Digital buses allow the addition, for future developments, of additional sensors, identified by a unique address on the same bus line, without the need to upgrade the controller input/outputs. Plant valves are controlled via 18 digital outputs, whereas plant discrete states are read via 20 digital inputs. The Arduino TIAN has been selected because it joins a deterministic SAMD21 Atmel Cortex^® M0+ 32-bit microcontroller (MCU) with 32 KB of SRAM program memory and a Qualcomm Atheros AR9342 processor running a Linux operative system (OpenWrt) that supports networking capabilities. A user interface based on a web page, that shows the plant state and allows the user to enter all required reference values and commands, has been implemented by exploiting Arduino libraries for web server creation. Given that web server functionalities are still not available on the Arduino TIAN two alternative solutions have been developed for the user interface. The first one via a web browser, using an additional Arduino YUN, connected to the TIAN, acting as web server, and the second one where the user interacts with the TIAN via a serial connection to a local micro-PC, located in the plant electric cabinet.

The controller is implemented as a Discrete-State Controller. For each possible state of the plant, a control state is defined. When the state is activated, all requested control actions are performed (e.g. opening/closing valves, running a pump), then the associated activities (e.g. monitoring the solution DO or pH or counting a predefined time lap) are performed until the exiting conditions are satisfied and a new state become active. The defined states are: VFA dosing to R1 (SBR), Nutrients dosing to R2, Dilution with tap water, Mixing, VFA dosing to R2 (Accumulation Reactor), Mixing, Sedimentation, Collect the biomass, Remove excess water, Backflush for filter cleaning, and Standby. In a concurrent, parallel, control loop, the reactors are temperature controlled by reading their temperatures and switching on/off the corresponding heaters.

The main experimental activities were carried out to:

The main goal addressed in this work about the PHA production, including the scale-up study, was to develop an original, cost-effective, fully-automated and industrial competitive downstream process for obtaining PHA from microbial cultures. This is in fact still an open challenge in the lack of automated PHA extraction processes and integrated advanced oxidation photo-catalytic systems.

The already existing patents related to PHA (more than 150) mainly regard the production of the polymer, ranging from original approaches for microbial cultivation to novel extraction methodologies. Among the latter, various pre-treatments (e.g. physical treatment, alkali/acids, freezing or heating), polymer recovery (e.g. through cyclic carbonates, ionic liquids or surfactants) and polymer purification (e.g. through bleaching solution) have been investigated (the most relevant examples are reported in [27‐31]).

No patents or scientific publications are available for the specific applications related to PHA involving the proposed technologies: (i) the applicability of nanoparticles to destabilize bacterial membranes and purify the extracted PHA, and (ii) Pulse Electric Field (PEF)-assisted extraction processes. The available patents regard the use of TiO₂ or Ag⁰ nanoparticles for wastewater treatment, purification and decontamination, and the use of PEF for achieving the permeabilization of membranes to extract various valuable chemicals (lipids, pigments, juice, proteins) from plant or algal tissues.

The use of developed photocatalytic tools, in particular TiO₂ coated textiles, was tested at lab scale in the purification treatment of PHAWW. With the aim of improving environmental sustainability and saving water/energy consuming of the whole system, the integration of the waste-to-bioplastic pilot system is strongly encouraged. From a life cycle perspective, the abatement of pollutant load from the very large amount of water disposed during the concentration of biomass (around 98% by volume) represents a key step for the sustainability of the overall process. The produced know-how allowed the scale up of both bio and photocatalytic reactors at pilot scale level, ensuring the achievement of expected TRL 5/6 and providing a big chance for transferring plants and knowledge in a European research and industrial application environments.

The main outcomes achieved in this work beyond the state of the art contributed to make the overall process more attractive for the companies and even transferable on an industrial level. The process optimization provided safe and cheap PHAs extraction without toxic and harmful chemicals, moreover the improvement performed on the monitoring and automation tools represented a good starting point to lower investments and operating costs. Specifically, in the PHA production system, the biomass production showed to be effective in producing PHA enriched biomass with MMC. The biological reactor is easily scalable with acceptable cost.

Nonetheless, the extraction should be improved in order to obtain higher yield of PHA without significant degradation. This could be done by fine tuning the extraction or use of additives. Another option could be to thermally decompose the PHA inside the cell for the obtainment of monomeric chemicals (e.g. crotonic acid or propylene). Moreover, a complete system including the fermentation of raw wastewater to volatile fatty acids (VFA) should be demonstrated. Due to the difficulties found in the integration of TiO₂ and Ag based catalytic tools in PHAs production process, TiO₂ coated textiles were exploited as photocatalyst in purification process of wastewaters, originated from PHAs production process. Photocatalytic tests showed a great reduction of organic C, measured by TOC analysis, after 2 h of UV irradiation in the presence of TiO₂ coated textiles, as expected by a very efficient photocatalytic process. Thus, the photodegradation of organic pollutants present in wastewaters was demonstrated. The use of developed catalytic tools, in particular TiO₂ coated textiles, in purification treatment of wastewaters represent a promising alternative to their integration within PHAs production process. The very good results encouraged the introduction of such catalytic components in analogous processes, improving environmental sustainability and saving water/energy consuming.

More generally, the developed strategy contributes to allocate the PHAs production process as very promising green biotechnology. The fully automatized prototype realized for PHA cultivation is a further step towards the achievement of a compact all in one system able to convert a specific waste into ready-to-use bioplastic. Further studies will be focused on to the implementation, tuning and testing of the system to produce PHA with different carbon sources obtained from different wastes. A particular attention will be paid to the possibility to combine ceramic active phase such as supported nano-TiO₂ and Ag as green and sustainable remediation against any possible bacterial contamination that bio-reactor can generate.