Twisting biomaterials around your little finger: environmental impacts of bio-based wrappings

The functional unit chosen for this study is 1 m² of packaging film. This is (mostly) laminated, printed film that is delivered on reels to the food factory, where the laminate is cut, sealed and filled. Cutting of the sheets and sealing and filling of the bags are excluded from the analysis because this step can be assumed to be identical across all packs of the same size and function, and to produce the same level of waste. The results of the environmental assessment are reported both for the system ‘cradle-to-factory gate’ (CF) and for the system ‘cradle-to-grave’ (CG): The system CF includes all activities in the process chain starting from the extraction & processing of non-renewable resources (e.g. oil and gas) or agricultural & silvicultural production (e.g. maize from seeds, including fertiliser and machinery use) up to and including film production, lamination and printing; it also covers all transportation activities and treatment of any process waste up and until the laminated film is delivered on reels to the food producer. The system CG includes the system CF plus waste management of the post-consumer packaging waste² where all key options are studied, i.e. incineration (with and without energy recovery), landfilling, composting, and digestion. We calculate GWP for the system cradle-to-factory gate for all bio-based products by adding all emissions of fossil greenhouse gas emissions and subtracting the biogenic carbon that is physically embedded in the product. As a consequence, both fossil and biogenic emissions of greenhouse gases from the waste treatment stages are considered.

In this study, we applied system expansion (also referred to as ‘avoided burdens’) to account for the co-generation of electricity and heat. To determine the credit, electricity that is co-produced, e.g. during the incineration of waste, is assumed to replace electricity produced according to the average power generation in Europe (see Section 3.3) and heat is assumed to replace average production in a gas-fired boiler.

In this study, we used the CML 2 baseline 2000 method (Guinée et al. 2001) for calculating the mid-point results, adding water use and land use as impact categories. For CF, the so-called LCA mid-point results are presented for the following impact categories: Non-renewable energy use, Total energy use (total of non-renewable and renewable energy use), Global warming potential, Depletion of abiotic resources, Photo-oxidant formation, Acidification, Eutrophication, Water use, and Land use;³ for CG, only results for the categories Non-renewable energy use and Global warming potential are shown because of the large uncertainties related to estimating individual process emissions other than CO₂, CH₄ and water during the waste treatment phase, especially of novel materials such as polylactic acid (PLA).

So far, water use as an impact category has not been very common in LCA studies but is receiving more and more attention (Mila i Canals et al. 2009; Pfister et al. 2009). Within the category of water use, we consider process water, cooling water and irrigation water. Process water includes all the water used during the production of raw materials, films and laminates but excludes the water used for electricity production from hydropower. The subcategory cooling water includes water used for cooling in any of the process steps. Irrigation water means water fed to the agricultural system during crop growth; it excludes rainfall because there is no generally accepted methodology that would allow a consistent comparison of rainfall quantities across agricultural as well as silvicultural crops and across geographical regions of production.⁴ Energy consumption for irrigation is excluded because it only contributes to a small extent to the total energy consumption of agricultural crops (Mila i Canals 2003).

Land for agriculture and forestry will be increasingly important in the future because of increasing land requirements not only for the production of food and feed but also of bio-energy, bio-fuels, and bio-materials. A growing number of environmental assessments of bio-materials include land use in their analyses (see Dornburg et al. 2003; Hermann et al. 2007; Kim and Dale 2005b) and methodological work is under way (see e.g. Jolliet et al. 2003; Kloverpris et al. 2008; Mila i Canals et al. 2007) but has not yet established one single accepted methodology in terms of how different types of land use should be compared. In this study, we focus on land use for agriculture and for (sustainable) forestry. The different types of land use (e.g. agriculture and forestry) are aggregated 1:1. The rotation period of forestry is taken into account. Land use for industrial plants, transportation infrastructure and waste management is comparable for different types of material and thus not taken into account.

As can be observed among private consumers, companies are also increasingly purchasing ‘green electricity’ in order to reduce their environmental footprint; one such company is NatureWorks, the most important producer of PLA. Purchasing green electricity is an option for any producer - and this raises the question if such environmental credits should be considered when conducting an LCA. In order to describe the different views, we distinguish between the company perspective and the technology perspective. An LCA carried out from a company perspective includes not only the processes operated by the company itself but also the purchasing decisions the company makes regarding materials and energy inputs. As a consequence, a study using the company perspective gives credits for the avoided environmental burden related to green electricity. We argue that the company perspective should be applied for decisions concerning business relations. A comparison from a technology perspective of materials production and processing strives to eliminate the effect of whether the company uses power generated by using wind energy, natural gas or coal, focussing instead on the core technology. (Here, core technology refers to the production of materials and their subsequent processing to produce packaging film, but excludes the generation of green power as an optimisation strategy.) The technology perspective is the adequate choice when making decisions concerning material choice or R&D strategies. In this paper, we focus on the technology perspective because we are mainly interested in material choice; but we include a case of company perspective to show the difference between the two.

In the following, we describe the effect of a company’s decision to purchase ‘green electricity’, using the example of wind energy. The accounting practice adopted by the International Energy Agency (IEA 2003) for their energy balance tables is that the primary energy form should be the first energy form downstream in the production process for which multiple energy uses are practical. The application of this principle leads to the following primary energy forms: 1) Heat for nuclear electricity, for geothermal heat or electricity and for solar heat production 2) Electricity for hydro, wind, wave/ocean and photovoltaic electricity production. This convention has been agreed upon internationally for energy balances. As there is no such agreement for LCAs in general, we follow the IEA convention. As a consequence, the primary energy consumption related to the production of 1 kWh of electricity is lower if it is generated from hydropower, wind, wave/ocean and/or photovoltaics compared to its generation from fossil resources (oil, gas, coal) or nuclear, geothermal or solar heat.

The companies involved were supplied with surveys on the inputs and outputs regarding that part of the production process which takes place at their respective production sites. The data they provided was on the level of process inputs (materials and energy) and outputs (materials, waste and emissions). We then went on to perform a plausibility check and to benchmark the received data against other company data as well as database data (where available). When large discrepancies were found, feedback was provided to the company. In some instances this led to a correction of the material inputs or outputs, but in other cases there was a technical explanation for certain differences. This resulted in a consistent data-set which was then incorporated in our calculations. At the end of the project and before submitting the final results, all contributors to the data-sets received a copy of the preliminary report and an opportunity to give feedback.

It is important to note that despite all these efforts, there can still be substantial differences between individual production sites, which are caused by differences in the qualtity of data or the level of technology. In the first case, data quality varied because only some of the suppliers had individual meters installed to measure energy consumption by individual process steps of the entire production process. Most suppliers broke down data from an entire production site or from the average yearly use of an installation to individual films or laminates. In the second case, the level of technology can lead to higher energy use for novel materials compared with ‘conventional’ films for the following reasons: 1) old production lines may be used for small scale production of these films and, in general, these production lines are less energy-efficient than state-of-the-art production lines as used for current conventional materials, and 2) production processes for conventional materials have been fine-tuned over decades, leading to higher throughput, yields and/or energy efficiency of production. Both cases apply especially to novel materials that have not yet been converted into films on a large scale and where, as a consequence, fine-tuning has not yet occurred. We took this inherent bias in the data into account in a sensitivity analysis (see Section 5.1).

We distinguished two levels of specificity, i.e. producer-specific⁷ data and regional (supranational) data. Producer-specific data was collected from production sites and can be grouped into information on materials production and processing. Producer-specific data was used for some polymers in primary form (e.g. PLA) and regional data for others (e.g. PP) and similarly for polymer processing such as the production of films from these materials. For example, regional data was used for OPP film but producer-specific data for PLA film because worldwide, there is only one large-scale plant to produce PLA. Similarly, also cellulose films for food packaging (using a novel technology) and a specific bio-based polyester (BBP) represent very specific products that are currently being produced at one or very few locations only. All these materials are novel and unique, the technology is not widely available and considerable progress can be made from one year to another. For paper we also used producer-specific data because of the multitude of different types of paper and their related LCA profiles, which make it difficult to decide which of the existing data sets are representative. On the other hand, bulk materials such as PP granulate are generally purchased from a wide range of sources, nationally and internationally, and therefore no close link exists to a specific producer or production site.

All regional data (such as electricity, heat, plastic granulate production) were assumed for a European setting, with the exception of PLA raw material produced in the USA and transported to Europe. Technical specifications such as tie layer thickness in extrusion lamination also vary among regions. Results were therefore only calculated for packs used and produced in Europe and may not be identical, nor lead to the same conclusions for the USA, Asia or other regions.

Transportation was considered during all stages of production, i.e. from the transportation of raw materials (such as wood for paper) to the delivery of laminates to the food producers. All manufacturing activities (production of materials, films and laminates) were assumed to occur within Europe, with the exception of PLA granulate production, for which the only large-scale industrial process is located in the U.S. As a consequence, the transportation distances and therefore also the environmental impacts from transportation are much larger for PLA than for the other materials.¹¹ Generic, regional data-sets were used for transport by lorry, rail and ship. Lorries were assumed to have a total maximum weight of 40 tonnes and a conservative load factor of approximately 50%. Ships were also considered to transport loads in large quantities and over long distances. Trains were assumed to run predominantly (70%) on electricity in Europe. A separate data-set was used for the transportation of PLA by train in the USA with higher primary energy use to account for lower energy efficiency.

Compared to the impacts of material production, transportation generally causes comparatively small environmental impacts. Due to the short collection distances this is particularly the case for collection & transportation of mixed household waste (including Inner and/or Outer Packs) in municipalities and their surroundings. We therefore excluded post-consumer waste collection from the LCA calculations. For all post-consumer waste treatment options, readily available data from databases and publications were compiled per type of material embodied in the waste. Based on this available information and additional input by experts from the field, we estimated the water and carbon released (as carbon dioxide or methane) from the material during the waste phase for all materials (Hermann and Patel 2007). For landfilling, composting and digestion, values in literature diverge significantly with respect to the carbon released over time (see e.g. 0–100% for PLA in landfill in Bohlmann 2004, 55% for composting of PLA in Iovino et al. 2008, 80% for composting of PLA in Kale et al. 2007). We take this uncertainty into account through uncertainty ranges for the degradation levels of these materials, i.e. high and low carbon storage (resp. low and high level of degradation, see Figs. 4 and 7).

The uncertainties related to the assessment of the waste management stage are very large, given the very wide range of the values reported in literature. These uncertainties concern both carbon storage in the solid phase (i.e. in compost, digestate or stored inside a landfill) and the composition of CO₂ and CH₄ in the gaseous phase (landfill gas, biogas from digestion and CH₄ emissions to the atmosphere). The overall difference between different types of waste-treatment with respect to GWP is rather small for biodegradable laminates and larger for non-degradable laminates where essentially no GHGs are released during landfilling. These differences are entirely due to the degradability (or lack thereof) of the material and as such cannot be reduced. Although the uncertainty regarding the degree of degradation of paper, cellulose, PLA and BBP is large, this uncertainty is not decisive for the conclusions: the difference between the materials in terms of environmental advantages is large enough for the conclusions to still hold (compare Figs. 4 and 7).