Bioplastics innovation: commercialization strategies for polyethylene furanoate (PEF) and polyhydroxy alkanoates (PHA)

Plastics have become ubiquitous materials essential for modern life and sustaining the global economy. On the other hand, their dependence on dwindling fossil fuel sources and their durability has raised concerns about their sustainability. Biobased plastics, which include biodegradable plastics, appear to be an appealing solution, yet their uptake has been slow. This paper specifically looks at the commercialization challenges for two bioplastics: polyethylene furanoate (PEF) and polyhydroxy alkanoate (PHA), that have been claimed by the industry as potentially disruptive. The superior barrier, mechanical and thermal properties of PEF present a suitable competition to polyethylene terephthalate (PET). The versatility and marine biodegradability of PHAs have attracted interest from several potential adopters. However, the high production costs of both pose a serious commercialization challenge when considering the established production, supply chain and recycling infrastructure for fossil‐based plastics. This study illustrates how firms are mitigating technical and market uncertainties by strategies such as adding value to other products, utilizing existing industrial assets, balancing exploration of more efficient and sustainable production pathways with commercial exploitation, lowering transaction costs by utilizing existing knowledge base, by developing formulation capabilities and forward integrating into end consumer products, or by entering collaborative alliances. These strategies could inevitably overcome inertia posed by the current market structure for fossil‐based plastics. © 2022 The Authors. Biofuels, Bioproducts and Biorefining published by Society of Industrial Chemistry and John Wiley & Sons Ltd.


Introduction
A lmost 368 million tonnes of plastics are consumed each year, 1 and they require only about 4% of the fossil resources extracted from Earth. 2 However, if the current growth in the use of plastics continues, the plastics sector will account for 20% of total oil consumption by 2050. 3,4 Growing scarcity and the rising cost of raw materials and serious concerns around climate change from fossil fuels has put the manufacture of plastics based on renewable raw materials 422 firmly back at the centre stage. Currently, about 2.1 million tonnes, or less than 1%, of plastics come from renewable bio sources. 5 Biobased plastics are not a universal solution to all of the world's problems -but if people must move someday (either by choice or by necessity) to a world without fossil resources, plastics will have to be made from some other sources, most probably agricultural, including waste biomass. It is one thing to make industrial use of materials already available in nature such as Jute fibers, but it is quite another thing to learn nature's principles of processing materials and apply them in novel ways untried in natural ecosystems to make bio-based materials and develop applications for them.
The success of bioplastics is contingent upon breakthrough radical materials and technologies, as well as the ability to compete with well-established incumbent materials and production methods with more than a century of technological development, sophisticated production and conversion, complex yet established supply chains and market channels, and growing infrastructure for recycling. Yet the long lag time for new materials and chemicals to realize returns is a major source of frustration among companies. In some cases, it can take 15-20 years for a new material or chemical to yield substantial revenue from the development of new markets and/or applications, 6,7 and bioplastics are no exception. 8 Therefore, firms producing or using incumbent materials could be more interested in finding bio-based routes to established plastics ('drop-in' plastics) that fit in with the current conversion and disposal infrastructure than developing novel bioplastics that carry high technical and market uncertainties. 9,10 Polyethylene terephthalate (PET) is the resin of choice for most beverage bottles. Bio-PET that is either partly or wholly derived from renewable sources is a 'drop-in' plastic because it is like for like with fossil derived PET. 11 Bio-PET still requires new routes and competencies including feedstock production, sourcing and production of the plastic's basic building blocks, but no change would be required in converting the building blocks to PET and the recycling infrastructure. Other than 'drop-in' plastics, there are new to the world materials such as polyethylene furanoate (PEF), 12 which are still in their pre-competitive stages. Compared with PET, PEF offers numerous benefits such as a superior oxygen and carbon dioxide (CO 2 ) barrier, hence it can preserve a soda's fizz longer by keeping oxygen out and CO 2 in; it is lightweight and it offers superior mechanical and thermal properties, and potentially long-term cost competitiveness at industrial scale. 13 However, there are no PEF bottles today as such and disrupting PET will not only require new routes and competencies in producing the material, but also new methods and competencies in conversion and recycling.
Polyhydroxy alkanoates (PHAs) that combine high functionality (tuneable mechanical and physical properties) with low environmental impact (biodegradability and nontoxicity) have the potential to disrupt across the industry. They can potentially substitute for polypropylene (PP), polyethylene (PE) and polystyrene (PS), which are the three main polymers of the global polymer market. 14 PHAs are the only known bioplastics that are entirely produced and degraded by living cells. Currently it is the only material certified for marine degradation. However, their cost, despite nearly four decades of commercial availability, is 2-4 times that of traditional plastics. 15,16 This limits their scale-up and they occupy only 2% of the 1% share of bioplastics. 17 In seeking to explain the commercialization of bioplastics innovations, competence destruction risk reflects the technical and market uncertainty in their scale-up. The technical uncertainties surrounding the development and production of new materials are compounded by market uncertainty because it is far from clear what benefits they provide. When technical and market uncertainties are high, it is difficult to predict which materials, routes to production and applications will be profitable. Therefore, the aim of this paper is to illustrate the innovation and commercialization strategies that firms adopt to mitigate technical and market uncertainties. It specifically focuses on PEF and PHA as they have been touted as potentially game changing.

Methodology
The analysis of PEF and PHA included semi-structured interviews with technical and managerial staff in bioplastics firms, patent analysis, and secondary data. Qualitative analyses including interviews and secondary data can effectively complement patent data, thereby providing in-depth case analysis through gathering data widely from multiple resources.
In total seven semi-structured interviews were conducted with various technical and managerial staff in bioplastics firms (see Table 1). Four interviewees asked that their name and the firm's identity are kept anonymous because of the business-sensitive nature of commercialization efforts. The interviews were approximately 30-45 min in length. The interview goal was to understand the technical and market challenges in commercializing bioplastics. The main topic of semi-structured interviews was how firms developed new capabilities and competencies through their innovation and production activities. Any quotes and the authors' own interpretations were verified for accuracy by the interviewees before disclosure.

423
Patent analysis was conducted to understand the technology trends of the specific bioplastics in this study. Patent data are suggested in the literature as one measure for innovation and diffusion and can thus provide a proxy for industry interest in particular technologies and their future trajectory. 18,19 Patent data were retrieved and analyzed using the LENS database (lens.org) using a combination of keyword-and classificationbased searches [both Cooperative Patent Classifications (CPC) and International Patent Classifications (IPC)]. The search was restricted to the triadic patents and grouped by simple families. The triadic patent family is a set of patents taken at the European Patent Office (EPO), the Japanese Patent Office (JPO) and the US Patent and Trademark Office (USPTO) that share one or more priorities. In comparison with patents filed in a single patent office that have a domestic bias, triadic patent families provide high-quality indicators as the most important inventions are filed at the EPO, JPO and USPTO. 20,21 For PHAs, we first conducted a search using keywords in the title and abstract and CPC and IPC classifications C12 (biochemistry beer spirits wine vinegar microbiology enzymology mutation or genetic engineering) and C08 (organic macromolecular compounds, their preparation or chemical working-up compositions based thereon). The C12 classification was used for PHAs because the polymers are directly produced by microbial fermentation. Searches were also made for bioplastic applications of PHAs in medical devices (CPC/IPC A61), the shaping of plastics (CPC/IPC B29) and multi-layered packaging (CPC/IPC B32). For furanic polymers PEF and polytrimethylene furandicarboxylate (PTF), searches also included their chemical building blocks and precursors, namely furandicarboxylic acid (FDCA), furan dicarboxylic methyl ester (FDME), chloromethyl furfural (CMF) and hydroxymethyl furfural (HMF). Because building blocks and precursors are included in our search, therefore CPC and IPC classification C07 (Organic Chemistry) was also included in addition to C08. Table 2 shows the search criteria for the two bioplastics.
For secondary data, scholarly and publicly available information such as peer-reviewed articles, industry news and magazines, government agency reports, newswire distribution, science and technology news and company websites were collected. According to Lee,22 the unobtrusive access that secondary data can present may help overcome sensitive matters that are problematic in the interview. Moreover, secondary data can also be used to provide 'triangulation' , increasing the credibility of research findings using primary data. 23

Polyethylene furanoate
Developing value-added applications The upstream position of bioplastics producers means that the creation of a prototype product will also depend  424 on downstream product design and complementary innovations. For example, a thin layer of Avantium's PEF will provide the Paboco paper bottle project with the high gas barrier properties needed for beverages such as beer and carbonated soft drinks. 24 Substituting complete glass or PET bottles with PEF has much greater market uncertainty than adding value to a paper beverage bottle; currently the cost of PEF is 8-10 times that of PET, 25 and even at economic scales it is predicted to be twice that of PET. 26 Presumably, the cost-performance of PEF for beverage bottles is not sufficient to overcome the cost-performance of PET even if all the PET were 100% bio-based. 27  Avantium has therefore perhaps decided not to compete with PET like for like. 28 Yet the gas barrier properties of PET bottles can be improved with a PEF coating 29,30 and according to Avantium's webpage, 31 this would also enable a PET bottle to be reused five times more than pure PET. Moreover, it can lead to reducing the overall weight per bottle and potentially lead to a reduction in costs. However, this would still interfere with a relatively good value chain for PET that is 100% recyclable. And though Avantium had received interim approval for including PEF in the European PET recycling stream, this is limited to a low 2% maximum and also subject to PEF producers and manufacturers developing a separate recycling stream for collected and sorted PEF-based bottles. 32 Further, Avantium is also targeting flexible packaging applications where multilayers can be replaced by one or two layers, with the PEF layer providing added functionalities from its improved barrier, physical and mechanical properties. 28,33,34 Existing multilayer packaging applications have three to 12 layers, 35 making them difficult to recycle. Proposed solutions such as delamination and selective dissolution of one polymer phase are still not technically or economically viable, and therefore multi-material flexible packaging films are currently landfilled or incinerated for energy recovery. 36 The current trends are towards mono-material-based packaging, defined as a packaging having 90-95% of one resin type like PE, PP or PET and ideally less than 5% of other materials or resins including barrier layers, adhesives and inks. [37][38][39] The main properties required for the successful adoption of mono-material films in the packaging sector are sufficient mechanical strength, appropriate gas/moisture barrier, low material and processing cost and recyclability using conventional methods. 36 PEF which has excellent barrier properties could support mono-material-based flexible packaging that is lighter and designed for recyclability.
It is unlikely that PEF will replace PET in the near future given all the commercialization challenges around competing with an established PET infrastructure. 40 Therefore, adding value to new downstream innovations such as the Paboco paper bottle project and existing applications such as barrier films that do not interfere with PET recycling is a suitable strategy. Yet firms engaged in green innovation may be willing to embrace market-based risks in the hope of creating disruptive innovation and gaining market advantage. 41 Avantium hopes in its investor report 42 that it will ultimately be able to license its technology to large industrial-scale polymer producers, enabling mass applications including bottles and textiles.
Utilizing existing PET assets for PEF production Dupont in collaboration with major agri-food player Archer Daniels Midland Co. (ADM) is focusing on FDME, a derivative of FDCA. FDME is a building block not only for PEF, but also for PTF. 43,44 Whereas both PTF and PEF offer better barrier properties compared with PET, PEF is a better choice for rigid packaging such as bottles and PTF is a more suitable choice for flexible packaging. Polymer production from FDCA requires high-purity FDCA and therefore Dupont is focusing on the cost-effective FDME route allowing for higher volumes. Another strategy for increasing production of PEF and creating demand is to use existing PET assets. As per respondent RA: It's easy to drop in PEF production into PET assets if you are going to solely convert a plant to make PEF, there are small adjustments to equipment that may be required, but yes by and large you can just drop in the technology into existing assets.
Currently Avantium has a pilot plant for FDCA at Geleen, Netherlands with a capacity of 10 t a −1 but is significantly ramping up capacity by building a 5 kt a −1 flagship FDCA demonstration plant at Delfzigl, Netherlands which is to be completed by 2023. 42 The final polymerization into PEF will happen at the assets of specialty polyester firm Selenis in Portugal. Smaller specialty polyester producers like Selenis have continuous lines and/or flexible batch processes that are inherently smaller and a significant portion of them can Perspective: Bioplastics commercialization S Kunamaneni 425 be dedicated to PEF production. Respondent RA suggested that many exclusive smaller PET lines in Europe and North America can also be perhaps dedicated for PEF production. These lines are unable to compete with larger modern PET plants in Asia that have economies of scale and are more efficient, but respondent RC believes differently: it's probably more attractive to build, it's expensive, but its more attractive to build dedicated volume, both the economy of scale perspective, but also the technical features, rather than to try and make it work in a capacity that's not purposely built for it.
Although technically it is feasible to produce both PET and PEF in the same continuous production plant, this would however lead to wastage of material because of off spec polymer and hence may not be economically viable. Also, synthesis of PEF using the conventional polycondensation process that is used to produce PET is challenging because the reaction time leads to undesired degradation and discoloration of the product. Rosenboom et al. 45 therefore propose a novel ring-opening polymerization process that is faster, more sustainable and provides a better material quality. Perhaps Avantium's strategy of using existing PET assets for PEF is due to its capital investments in core FDCA technology, thus locking them into developing distinctive competencies in FDCA production. Although Avantium has signed offtake agreements for its future PEF output, 46 it is challenging to forecast the future market potential for PEF or predict if and how industrial scale-up will develop through utilizing existing PET assets and building dedicated PEF production facilities.
Avantium has also recently licensed a portfolio of FDCA technology from Eastman to be combined with its own technology in the demonstration plant. 47 Eastman had also previously offered a non-exclusive license to Canadian firm Origin materials to efficiently convert HMF to FDCA. 48 Origin also purchased a pilot plant from Eastman as part of the deal, but Origin's primary focus is producing CMF from cellulosic feedstock. Given the technical and market uncertainties in FDCA and PEF commercialization, outlicensing on a non-exclusive basis is low risk and will have minimal impact on the core business of Eastman. The positive affect of licensing revenues based on manufactured volumes could dissipate any negative affect on their core business so long as the global share of bioplastics is low.

Improving building block production through alliances
While alliances such as the EU funded PEFerence consortium 49 and the Paboco paper bottle project that 'exploit' commercial applications may mitigate some market uncertainties by establishing a value-chain for PEF, alliances that 'explore' new processes that can reduce both the environmental impact and the cost of producing the building blocks and the bioplastic from a variety of feedstock sources are equally, if not more, important. Higher operating costs from expensive sugar sources and catalysts, high energy demands and higher health and environmental impacts from the use of toxic solvents make the production of the building blocks HMF and FDCA less economically viable and sustainably feasible. 50,51 Firms generally enter exploitation alliances because they exhibit less uncertainty and thus require fewer resources. 52,53 However, for upstream chemicals and materials firms like Avantium, an exploitation strategy for developing commercial application for PEF needs to be balanced with technological exploration for more efficient building block production, otherwise it will be challenging to compete with fossil-based plastics.
After 2004 when USDOE named FDCA as one of the top 12 bio-based building blocks, there has been an exponential growth in patents surrounding furanic building blocks, i.e. FDCA, FDME, HMF and CMF (see Fig. 1). These serve as building blocks not only for polyesters PEF and PTF, but also for other key commercial polymers such as polyamides and polyurethanes. In total 434 patent records organized by simple families were found when all the keywords related to the building blocks and the specific furanic polyesters PEF and PTF as shown in Table 2 were used in the title and  Table 2. 'Only Building blocks' excludes PEF and PTF in the title and abstract and keywords polyesters, polyamides and polyurethanes in the title. CMF, chloromethyl furfural; FDCA, furandicarboxylic acid; FDME, furan dicarboxylic methyl ester; HMF, hydroxymethyl furfural; PEF, polyethylene furanoate; PTF, polytrimethylene furandicarboxylate.

S Kunamaneni
Perspective: Bioplastics commercialization 426 abstract. The number of patent records found for building blocks was 347. Specific searches for furanic polyesters PEF and PTF only yielded 29 records. Therefore, much of the focus has been on pathways for building blocks rather than 'polymerization' processes of building blocks into bioplastics or the conversion of bioplastics into end use applications. Avantium is involved in several EU-funded collaborative efforts with other firms and knowledge partners in developing novel routes for bio-based chemicals and building blocks. For example, Avantium is leading an EU-funded project IMPRESS (Integration of Efficient Downstream Processes for Sugars and Sugar Alcohols), a consortium of nine partners including three public science institutions. 54 The project is exploring technologies for refining renewable non-food resources such as forestry and agricultural residues into multiple sustainable chemicals and materials that can replace fossil-based products.
Another firm active in bio-based building blocks is Swiss-based AVA Biochem, which has been producing HMF for research purposes as well for sale at its 20 t a −1 facility in Muttenz. 55 In 2020, it entered into a joint development agreement with Michelin to build the world's first commercial-scale production facility for HMF. 56 Like Avantium, AVA Biochem is also involved in several EU-funded projects. For example, the Enzox2 (Enzymatic Oxidation/oxyfunctionalization Technologies) project that consisted of 12 partners including six public science institutions developed biochemical routes for the efficient conversion of HMF into FDCA. 57, 58 A joint EU project in which both Avantium and AVA Biochem are involved along with nine other partners including two public science institutions is PERFORM (Power Platform). 59,60 The main objective of the consortium is to show that bio-based building blocks can be obtained efficiently and sustainably in a single-step chemical process rather than multiple steps. Avantium's role is mainly to develop the catalysts for the single-step process whereas AVA Biochem's role is in conducting an economic feasibility analysis of the processes and identifying competitive routes to market. The consortium therefore entails both exploratory research and exploitative commercialization activities coordinated by different partners.
Although recognizably Avantium's proprietary YXY technology converts biomass feedstock directly into FDCA and Ava Biochem's novel 'hydrothermal processing technology' converts biomass to HMF, which is a precursor to FDCA, the PERFORM project illustrates that commercialization of bio-based chemicals and plastics is challenged by economic and sustainability concerns in biomass conversion, thus necessitating collaborative expertise in the value chain further upstream for 'exploring' new production pathways to bio-based building blocks and chemicals. Exploratory activities very importantly draw upon the knowledge base of public science institutions. 61 Whereas 'polymerization' of FDCA into PEF could utilize incremental innovations of existing PET assets, the production of the building blocks HMF and FDCA will require new and capital-intensive technologies that will have a long lead time, hence the need to explore pre-competitive technologies in collaboration with other early-stage competitors and public science institutions. Figure 2 summarizes how firms mitigate technical and market uncertainties for PEF and its building blocks. Firms face market uncertainties from having to compete with an established market and infrastructure for PET. Therefore, value-added applications that exploit the excellent barrier properties of PEF provide a niche market for PEF. Further PEF in tolerable quantities could meet the recyclable monomaterial based design for flexible packaging. Technical uncertainty for PEF stems from investment required to develop and build dedicated PEF production facilities. PEF can be synthesized using the conventional polycondensation process that is used to produce PET as well as a novel ringopening polymerization process that is more sustainable, faster and does not cause undesirable degradation and discoloration. Utilizing existing small-scale PET assets, however, mitigates any technical uncertainties in developing and building dedicated PEF production facilities and allows focus on building block production which will clearly require new capital investment.

Summary
Building block production also requires exploring more efficient and sustainable pathways. Collaborative alliances with public science institutions supported by government grants not only mitigates technical uncertainties, but also market uncertainties for PEF because more efficient and sustainable routes to building blocks can lead to opportunities in bio-based polyurethane and polyamide markets. In addition, more cost-effective FDME can also relieve technical and market uncertainties around the FDCA route to PEF that requires high-purity FDCA and therefore limits volumes.

Polyhydroxy alkanoate
Tipping point for market exploitation Polyhydroxy alkanoates are not new. ICI produced the first ever commercially available PHA 'Biopol' in 1990 after 15 years of development. 62,63 The first commercial application Perspective: Bioplastics commercialization S Kunamaneni 427 was for shampoo bottles. 64 ICI sold its rights to Zeneca in 1993 which then passed it onto Monsanto in 1996. 65 Metabolix which was founded in 1992 then acquired the rights from Monsanto in 2001, but it had to stop production of its 'Mirel' brand of PHAs in 2012 when Telles, its 6-yearold joint venture with ADM, ended. 66 According to ADM, uncertainty around projected capital and production costs, combined with the rate of market adoption, were the main hurdles in PHA adoption. All the intellectual property related to PHAs was retained by Metabolix whereas ADM retained the plant. In 2016, Metabolix which held the largest number of patents in PHAs, sold all its biopolymer intellectual property to South Korean firm CJ CheilJedang in 2016 for $10 million. 67 In 2017, Metabolix changed its name to 'Yield10 Bioscience' , focusing on the development of genetic technologies to improve yields for food and feed crops to enhance global food security. 68 Its platform project on Camelina crop aims to achieve increased biodiversity, high-protein meals, the production of edible oils, biodiesel and omega-3 oils, and also increased yield of PHA within Camelina seeds. 69 Science-based firms such as Metabolix attempting to produce highly radical innovations have to be flexible in their use of key organizational resources such as scientists, thus enabling them to change direction when faced with high technological volatility. 70 Arguably, the failure of Metabolix could perhaps be attributed to the lesser focus on application pathways despite its production ramp up. In the opinion of respondent RF: There are over 150 different PHA monomers that we know of, and so when you think about combinations of those

monomers, there is a myriad of different PHAs that exist and that we could utilise, but ICI and Monsanto and Metabolix were all focused in a very narrow range of actual PHA molecules and polymers that exist. Unfortunately for them, those ones they focused on don't have very broad utility or applicability.
However, given the long lead time and market uncertainty for PHA, it would have been challenging for Metabolix to predict which technologies, scientific expertise, investment and commercialization strategies will be effective. When faced with technological and market uncertainties, such firms have to change direction at relatively short notice. 70 Figure 3 shows the patent records for PHAs. In the late 1990s and 2000s, there was an increase in patenting activity mainly from Metabolix, Canon, P&G, Tepha and Kaneka. Despite the failure of Metabolix in 2012, which held one of the largest patent portfolios around PHAs, the growth rate of patenting activity between 2012 and 2021 is twice that in the 5 years preceding 2012 (see Table 3). Although primary enzymatic synthesis and related patents belonging to CPC and IPC classifications C12 are still growing, for example because of interest in utilizing waste sources such as wastewater, the growth in composition patents under CPC and IPC C08 is higher. Indeed, pure composition patents that do not include C12 classification overtook synthesis-related patents in 2017. Patents involving medical devices and other bioplastic applications are also gradually growing, but it is the higher growth in composition patents that indicates a tipping point. There is more focus on blends and formulations of PHAs and other functional entities to obtain unique properties. Also, according to respondent RF:

428
There is a very large body of intellectual property that was embodied in patents, starting in the mid-80s and pretty active all the way up until a decade ago. All of those patents that were in 80s and 90s and early 2000 are now expiring. So, anybody could utilize that information now that patents are expired.
The chemical industry has a well-functioning market for intellectual property where process technology patents are sold through arm's length license contracts. 71 While patents are an important indicator of innovative activity, they are not the only indicator. There is also a growing base of academic literature on the genetics and biochemistry behind PHA synthesis. 72 If the biochemical synthesis pathways for PHAs are well laid out in the patents and academic literature, a skilfull combination of in-licensing, academic collaborations, new composition patents and secrecy around yield optimization could provide a competitive edge to existing players. In 2007, P&G licensed its intellectual property to Meredian (now Danimer Scientific), 73 implying that the earlier intellectual property that was created perhaps covered most of the basic principles and know-how of PHA synthesis. There are newer and more players today 74 despite the failure of Metabolix. An early player in PHA Innovation was Canon, which was mainly interested in their potential applications for printing toners and not bioplastics. Although Canon has a large body of patents in the synthesis of PHAs, its interest had subsided by 2013, and many patents have either expired or are inactive (see Fig. 4). Therefore, newer players could utilize this knowledge base. The propensity to explore new synthesis methods, exploit applications or both is hence dependent not just on organizational factors such as resources and governance structures, but also on environmental factors such as competitive intensity 75 and on transaction costs for utilizing the knowledge base.

Forward integration into end consumer products
Polyhydroxy alkanoates have inherent economic drawbacks compared with plastics from fossil resources. They typically have higher input costs, and the yields are inadequate owing to challenges in maintaining optimal bacterial growth conditions. 16,[76][77][78] Even after four decades of development, PHAs are still in their early stages of technical and market progress. They have a paltry 2% of the roughly 1% share of bioplastics. 17 It is no trivial task to displace the advantages of established fossil-based plastics, but the cost of utilizing a material does not consist in just simply the cost of the primary material, one also needs to consider the conversion and processing cost. As per respondent RF: I can enable a tonne of paper to be coated with PHA at the same or lower cost than polyethylene, not because the PHA is less expensive than the polyethylene. It is actually the opposite, but the cost to put the polyethylene onto paper is six or seven times greater than the cost to put PHA onto paper.
Although ADM had built a plant for producing 50 million kg at its Clinton facility in the US, Telles had only shipped less than 250 000 kg in 2011, hardly a fraction of the plant's capacity. 79 ADM spent $300 million constructing the plant over 4 years, which started production in 2010, 80 only to stop 2 years later. Plausibly, early capacity may have been built because of the often held notion that production drives demand for innovative products. 81 This is relevant for example in the case of consumer electronics such as new mobile phones incorporating the latest technology. However, for bioplastics innovation upstream of their value chain, the adoption challenges are in demonstrating a differential value over established plastics to justify switching costs. Although reportedly, the customer base for Mirel continued to grow, reaching 57 in 2011, of which 26 had placed repeat orders, 79 the volumes were still too low to meet the cost-performance  As discussed in the previous subsection, PHAs have recently received renewed interest, and the demand for PHAs has started to increase recently and it appears that the demand has outstripped the small supply base now dominated by start-ups. 82 According to respondent RE: When Metabolix was around, probably not that much from a regulation point of view or from a consumer pull, so there was no market pull for biodegradable packaging materials at all. But now those who advanced the PHA technology like Danimer and Kaneka, they are growing a lot, because this is definitely a solution for food packaging.
Films for food packaging are an attractive proposition for PHAs. Hoping to meet sustainability targets set by national plastic pacts, 83,84 many retailers are reducing packaging film weight or promoting in-store recycling. But in-store recycling has come under criticism because the films have ended up in incineration facilities abroad or re-exported again. 85 In the EU, only 17% of films are recycled into new material. 86 Even if infrastructure is in place by retailers for the short term before many local councils include films in kerbside recycling, it will not fix the expensive recycling process because of high-quality input required by end markets such as plastic lumber. 87 It is indeed cheaper to produce new film than to recycle film. Further, lidding films for ready-meal solutions cannot be recycled because of contamination. 88 Therefore, PHAs which are home compostable are a viable solution. Recently EarthFirst Films successfully completed the first commercial run of PHA films made from Danimer Scientific's Nodax PHA. 89 The film is designed for a wide range of applications across food, beverage, grocery retail, quick service restaurant and stadium foodservice. Today many start-ups and pilot technology programmes are either exploring or commercially producing PHAs in small quantities utilizing waste feedstocks that lowers feedstock costs and avoids competition with the food chain, for example using waste methane gas, 90 sequestered atmospheric CO 2 , 91 wastewater streams, 92 organic waste, 93,94 etc. Even though the feedstock costs can be significantly reduced using waste carbon sources, the costs of 'unit processes' such as separation and purification are higher compared with using pure sugars such as glucose for feedstock. 95,96 In 2015, Newlight signed a 20 year deal with Vinmar, a leading distributor of plastics and chemicals 97 and in 2016, it entered a supply, collaboration and technology license agreement with Ikea. 98 There has been no news of the supply deals since, but recently Newlight has commissioned a new commercial production facility. 99 The PHA from the facility is being used for its own new brands of end consumer products, 'Restore' and 'Covalent' . Restore offers reusable, compostable and marine degradable foodware such as straws and premium cutlery. Covalent offers luxury fashion accessories such as eyewear, wallets, cases and bags. Newlight sells both product ranges to end consumers through its own online portals and also has a direct distribution arrangement with major retailer Target in the US. 100 As Newlight had already spent 17 years on innovating the production of PHAs from waste greenhouse gases (GHG) s, shifting to exploitative markets would strengthen its capabilities in overcoming market uncertainties. Also, commercialization requires complementary assets like distribution channels that are dominated by established fossil-based plastics. Risks are higher for distributors when they must invest in new distribution capabilities that cannot be redeployed if the demand is not met. Early-stage ventures then may have to forward integrate into new markets, such as that adopted by Newlight. Therefore plausibly, Newlight has been unable to scale as per its agreement with distributor Vinmar. Although Newlight claims that its PHA is cost-competitive with fossil-based plastics, the scale-up would, however, be defined by supply deals for carbon from relatively concentrated sources and co-location of its production facilities. 101 Therefore, a major challenge for the large-scale production of PHA from GHGs would be co-locating near a reliable and large volume waste feedstock. Also the dependence on specific local conditions can make it difficult to ascertain the true environmental impact of PHA production. 102 Creating legitimacy and preparing the market through alliance formation In 2019, a global alliance GO!PHA was launched to provide a platform for sharing experiences, knowledge and developments in PHA, and to 'facilitate' joint development initiatives. 103 Its 38 members include biotech ventures Newlight, Danimer Scientific, RWDC Industries and Bioextrax, FMCG brands Unilever and P&G, petrochemical and energy major 'Total' , convertors and polymer processors Wittenburg and Maip, and five academic research institutions. Arguably, alliances promote the appeal and suitability of bioplastics technology by signaling intent and stimulating legitimacy. 104 Respondent RG comments: Something that we take very seriously within Go!PHA is that we are not allowed to discuss commercial opportunities because of anti-competition laws. But it is a very good place to meet other PHA producers, in particular potential buyers of PHA. There needs to be very close collaboration between the different steps in the value chain, where PHA producers speak to the formulators, compounders, brand owners and Go!PHA is a very good way of doing that. And of course, lobbying is a very important aspect as well, because legislation is key for PHA industry's future.
Alliances therefore vitally act as institutional entrepreneurs in the precompetitive stage of the value chain by promoting standards, certificates, labels, and legislation for the technology through collaboration. For example GO!PHA is lobbying with the European Commission to label PHA as a natural polymer that has not been chemically modified. 105 Not only producers, customers and users, but also distributors, complimentary technology developers, public science organizations, regulators and opinion leaders, can contribute to commercialization through an alliance by sharing knowledge, creating markets and facilitating innovation adoption/diffusion. Yet despite the strategic relevance of early-stage technology alliances, larger firms need to evolve and survive as they can provide the economies of scale in both production and innovation to compete with established fossil-based plastics. This is a challenge for PHAs, Perspective: Bioplastics commercialization S Kunamaneni 431 as has been shown by the failure of Metabolix, and seemingly appears to be the case for Newlight that would require reliable supply of GHGs from concentrated sources. Newlight is not the only firm exploring PHAs from waste GHGs. Another early stage California-based firm, Mango Materials, that started in 2012 as a spinoff from Stanford University, has also been exploring the production of PHAs from waste methane. 90 Perhaps the forward integration strategy of Newlight is to reduce the risks posed by opportunism in PHAs from waste GHGs, in addition to managing the uncertainties around distribution as discussed in the previous subsection. In the opinion of respondent RG: It makes sense for PHA producers to be more vertically integrated, to have a special expertise within polymer formulation and so on in house. Because what you see is that a lot of the large potential buyers of PHAs, they get the PHA, and they do not know what to do with it. So as a PHA producer, it is not enough to just produce PHA granules, you need to have the capacity to formulate it with other polymers or various additives into something that can be used directly by the customers.
The above statement correlates with the observation from patent trends on the growth of composition patents. While the various buyer actors in an alliance including convertors, distributors and brands can facilitate commercialization with knowledge, dynamism and relationships, early-stage bioplastics ventures cannot depend on alliance actors for their innovation, production and market moves. A cooperative knowledge alliance could intensify competition because of the availability of valuable information. And especially when alliances such as GO!PHA are restrictive for forming commercial opportunities, firms could engage in forward integration for accelerating consumer adoption and market position through new compositions and end-products, thus providing more valuable information and dynamism to the alliance.
Finally, as the history of PHA has shown, even for large incumbents in the petrochemical and biotechnology sectors with the resources for innovation, the incentives for change are little if the status quo is more profitable. Given the more than four decades of bumpy development for PHAs, the current producers of PHA are small and limited in number with a total capacity of ca. 66 kt a −1 . 106 Danimer Scientific is building a commercial plant that will have a capacity of 110 kt a −1 to be completed by 2024 in Bainbridge, Georgia. RWDC Industries has plans to build a 50 kt a −1 plant in Athens, Georgia and is building a 25 kt a −1 commercial plant to be completed by 2023 in Singapore. 107 And according to respondent RF, PHAs have significant potential: PHA is not right choice for every single application. But it is the right polymer for a great many applications. If you think about the total single use plastics market today, you know it depends on whose literature you read, you see ranges from 380 m tonnes to 600 m tonnes annually. Say its 380 m, 350 m, on the low side, little more than half of that is a good application space for PHAs. So, if you are being conservative and you said, PHAs have an applicability for 150 m tonnes of material on an annual basis, that's a lot.
The long lead time for PHA and interest from large incumbent plastic producers and brand owners is impacted both by exogenous factors such as recognition of PHA as a natural polymer and endogenous factors such as complexities associated with feedstock supply and optimizing yields. The planned production capacity is still remote from the potential future scenario suggested by respondent RF above. It is still early to predict whether GO!PHA can create enough external influence leading to radical changes in firms' strategies, particularly more commitment from larger established firms to increase production capacity and diffusion of applications. But alliances can play a key role in breaking resistance to adoption by preparing markets for an innovation. They can create a supportive environment by guiding strategic direction of established firms, building credibility, ensuring the delivery and existence of supportive complementary offerings, educating customers and mobilizing them to ensure market pull. 108 Summary Figure 5 summarizes how firms mitigate technical and market uncertainties for PHA. Firms face market uncertainties because seemingly distributors do not want to invest in new capabilities for biodegradable plastics. Therefore, Newlight had to forward integrate into end consumer products to accelerate market adoption. Also, because the share of PHA is just 2% of the 1% share of bioplastics despite nearly four decades of development, the GO!PHA alliance can ease market diffusion by effecting legitimacy.
Many PHA patents have expired; this reduces transaction cost in acquiring basic knowledge of PHA synthesis, thus allaying some technical uncertainties around development and production. Yet yields are limited by challenges in maintaining optimal bacterial growth conditions. Therefore, firms still need to develop proprietary processes for yield optimization which can give them a competitive cost advantage. Further, input costs can be reduced by developing and exploiting routes to PHA using waste feedstocks, thus easing market uncertainties around utilizing food-based feedstock. Producers are also mitigating technical and

Conclusions
Despite many decades of development and significant yet fragmented effort at commercialization, mostly by earlystage ventures, the share of bioplastics is still insignificant and <1% of the global plastics market. This study clarifies the dynamics that underlie their commercialization challenges and shows how firms are mitigating technical and market uncertainties by employing strategies such as adding value to other products, utilizing existing industrial assets, balancing exploration of more efficient and sustainable production pathways with commercial exploitation, lowering transaction costs by utilizing existing knowledge base, by developing formulation capabilities and forward integrating into end consumer products, or by entering collaborative alliances. Although PEF can be synthesized in existing PET assets, the main challenge is in efficiently producing the building blocks FDCA, FDME, CMF and HMF from biomass. There are no bottles or added-value applications such as multilayer packaging products yet on the market today utilizing PEF. Patenting trends suggest that new pathways for building blocks are still the primary focus. Also, firms producing PEF and its building blocks can be vertically separated like the PET value chain, especially because PEF is not the only possible product for the building blocks. Avantium is perhaps therefore building a pre-competitive FDCA demonstrator plant while still exploring more efficient pathways through in-licensing and collaborative alliances involving publicscience institutions. Its goal is to out-license its FDCA technology for scale-up and eventually to be able to establish a value chain for PEF so that it can compete with PET. For PHAs, the main challenge is in maintaining optimal bacterial growth conditions and optimizing yields in bioreactors. Patenting trends and industry dynamism indicate that the existing knowledge base on PHA synthesis may be sufficient to develop compositions of PHAs with unique properties and create proprietary optimization processes for improving PHA yields. In the case of PHA, there are products available on the market especially owing to their biodegradability, e.g., Newlight's Restore foodware products. Newlight has forward integrated into end consumer products, seemingly because the risks are higher for existing distributors when they must invest in new capabilities that cannot be redeployed if the demand is not met. In addition, users downstream may lack the knowledge and expertise to produce suitable formulations of PHAs.
More exciting development is yet to come, for example reducing the cost of sequestering atmospheric GHGs could reduce the cost of PHAs and more durable bioplastics using PEF could promote reuse; these hold the potential to disrupt the feedstock supply infrastructure, distribution channels and our consumption patterns. A specific limitation of this study as regards scope is that it does not categorically analyze the policies and regulatory aspects surrounding bioplastics development; this is particularly important given the food security debate and controversial nature of synthetic biology. 109 With regards government support for innovation efforts, the article touches on government support for PEFand FDCA-related R&D projects such as PEFerence and

433
IMPRESS, but future research could explore in detail the impact of legislations, the research priorities of governments and funding in stimulating new bioplastics discovery and new synthesis methods, and the role of governments in creating an ecosystem for bioplastics commercialization and entrepreneurial ventures.