Environmental assessment of bio-based polymers and natural fibres
Dr. Martin Patel*), Dr. Catia Bastioli**), Dr. Luigi Marini**), Dipl.-Geoökol. Eduard Würdinger***)

*)

Utrecht University, Department of Science, Technology and Society (STS), Copernicus Institute,
Padualaan 14, NL-3584 CH Utrecht, Netherlands; Phone: +31 30 253 7634; Fax: +31 30 253
7601; E-mail: m.patel@chem.uu.nl
**) Novamont, Via Fauser 8, I-28100 Novara, Italy; Phone: +39 0321 699 611; Fax: +39 0321 699
600; E-mail: bastioli@materbi.com, marini@materbi.com
***) BIFA (Bavarian Institute of Applied Environmental Research and Technology), Am Mittleren
Moos 46, D-86167 Augsburg, Germany; Phone: +49 821 7000-181; Fax: +49 821 7000-100; Email: ewuerdinger@bifa.de

Table of Contents
1.

INTRODUCTION

2.

HISTORICAL OUTLINE

3.

METHODOLOGY OF LCA

4.

PRESENTATION OF COMPARATIVE DATA
4.1
STARCH POLYMERS
4.1.1
Starch polymer pellets
4.1.2
Starch polymer loose fills
4.1.3
Starch polymer films and bags
4.1.4
Starch nanoparticles as fillers in tyres
4.2 POLYHYDROXYALKANOATES (PHA)
4.3 POLYLACTIDES (PLA)
4.4 OTHER POLYMERS BASED ON RENEWABLE RESOURCES
4.5 NATURAL FIBRES

5. SUMMARISING COMPARISON
6. DISCUSSION
7. CONCLUSIONS
7.1
7.2

SUMMARY AND FURTHER ELABORATION OF FINDINGS
OUTLOOK AND PERSPECTIVES

8. ACKNOWLEDGEMENTS

1

9. REFERENCES
ANNEX 1: OVERVIEW OF ENVIRONMENTAL LIFE CYCLE COMPARISONS FOR BIODEGRADABLE
POLYMERS INCLUDED IN THIS REVIEW
ANNEX 2: CHECKLIST FOR THE PREPARATION OF AN LCA FOR BIODEGRADABLE PLASTICS

a
ABS
CH4 cm CO2
CR pallet d ECCP
EPS
eq. g GF pallet
GHG
GJ ha HDPE kg l
LCA
LDPE
LLDPE
MJ m3 MSWI
N2O
PCL
PE
PET
PHA
PHB
PLA
PVOH
PWB
PO4
PP
PS
PVOH
PE
R&D
SO2
TPS

year acrylonitrile-butadiene-styrene methane centimetre carbon dioxide pallet made of chinareed as reinforcement day European Climate Change Programme expanded polystyrene equivalents grammes pallet made of fibreglass as reinforcement greenhouse gas emissions
Gigajoule (109 joule) hectare high density polyethylene kilogramme liter life cycle assessment low density polyethylene linear low density polyethylene
Megajoules (106 joule) cubic metre municipal solid waste incineration plant nitrous oxide polycaprolactone polyethylene polyethylene terephthalate polyhydroxyalkanoates polyhydroxybutyrates polylactides polyvinyl alcohol printed wiring board phosphate polypropylene polystyrene polyvinly alcohol polyethylene Research and Development sulphur dioxide thermoplastic starch
2

Abstract
Since the 1980ies and especially in the 1990ies, bio-based polymers and natural fibres have been playing a increasingly important role in various applications. Bio-based raw materials are being used to produce both biodegradable materials and materials that do not biodegrade. Superiority in environmental terms has been an important driver for the increased use of bio-based polymers and natural fibres and this can be expected to hold also for the future. As a consequence, thorough comparisons of bio-based products and their petrochemical counterparts are required. To this end, life cycle assessment
(LCA) can be applied, which is a standardised method to quantify environmental impacts.
In this paper twenty life cycle assessments are reviewed. Seven of the studies reviewed deal with starch polymers, five with polyhydroxyalkanoates (PHA), two with polylactides (PLA), three with other bio-based polymers (lignin-epoxy resins, epoxidised linseed oil) and another three with composites based on flax, hemp and china reed (miscanthus). All of the materials studied are manufactured exclusively or – in most cases - partially from renewable resources. The first three materials are biodegradable while this is not the case for the remaining materials studied. The types of end products covered are primary plastic materials (mainly pellets), loose-fill packaging material (packaging chips), films, bags, mulch films, printed wiring boards (for electronics), thickener for lacquer, two different panels for passenger cars and transport pallets. These products are compared with equivalent products made from petrochemical polymers – in many cases polyethylene, polypropylene or polystyrene. Some of the studies reviewed are rather limited in scope by assessing only energy use and CO2 emissions. They are nevertheless included because they contribute to a better understanding of the environmental aspects by addressing additional types of materials and by providing an indication of the uncertainty of the results.
The results show that bio-based polymers can contribute substantially to reduce the environmental impacts related to material use. In the case of starch polymer pellets energy requirements are mostly 25%-75% below those for polyethylene and greenhouse gas emissions are 20%-80% lower.
These ranges originate from the comparison of different starch/copolymer blends, different waste treatment and different polyolefin materials used as reference. The cradle-to-factory gate energy requirements for PLA are 20%-30% below those for polyethylene, while GHG emissions are about 15%-25% lower. The results for PHA vary greatly (only energy data are available): Cradle-to-factory gate energy requirements in the best case (66 GJ/t) are 10%-20% lower than those for polyethylene. For more energy intensive production processes PHA does not compare well with petrochemical polymers. Very attractive to attractive potentials were found for epoxidised linseed oil as thickener for lacquers.
(around 90% for energy and GHG emissions) and for the substitution of flax fibre mats for fibreglass mats (above 80% for energy). Considerable savings are also amenable by using other natural fibres
(between 14% for an under-floor panel for cars and 45%-50% energy for interior side panels and transport pallets).
The review also revealed a number of questionable assumptions and data uncertainties which need to be addressed by future research and LCA studies. One important finding is that various waste management treatment options should be included in LCAs for biopolymers and natural fibres due to their strong impact on the final results.

3

In spite of the uncertainties and diverse assumptions it is safe to conclude that biopolymers and natural fibres offer important environmental benefits today and for the future. This is particularly obvious for starch polymers. Compared to conventional plastics and fibres, the products studied generally contribute clearly to the goals of saving energy resources and mitigating GHG emissions. At the same time, none of the biopolymers studied performs better than its fossil fuel-based counterparts in all categories.
To summarize, the available LCA studies and environmental assessments support the further development of biopolymers and composites based on natural fibres. Regular monitoring of as many environmental impacts as possible forms a crucial basis for the continuous improvement of the environmental performance of biopolymers. For some materials the environmental benefits achieved are substantial already today. In many other cases the potentials are very promising and need to be exploited.

Keywords: Environmental life cycle assessment, LCA, energy, greenhouse gas emissions, GHG, uncertainties, renewable raw materials, bio-based polymers, natural fibre composites, starch, polyhydroxyalkanoates, polylactides, end products

4

1.

Introduction

Environmental considerations have been and will continue to be an important motivation to develop and introduce bio-based polymers and natural fibre composites. This calls for a comparison of their environmental performance with their petrochemical counterparts. To this end, life cycle assessment (LCA) can be applied, which is a standardised method to quantify environmental impacts (ISO, 1997-1999).
LCA studies, however, do not address environmental risks (e.g. related to outcrossing of genetically modified species) and they neither cover ethical, social and economic aspects.
This book chapter presents and discusses results from LCA studies for the commercially most important bio-based polymeric materials: Starch polymers, polyhydroxyalkanoates, polylactides, ligninepoxy resins, epoxidised linseed oil and composites reinforced with natural fibres sich as flax, hemp and china reed (miscanthus). The first three materials are biodegradable while this is not the case for the remaining materials studied.
The types of end products covered are primary plastic materials,1 loose-fill packaging material
(packaging chips), films, bags, mulch films, printed wiring boards (for electronics), thickener for lacquer, two different panels for passenger cars and transport pallets. These products are compared with equivalent products made from petrochemical polymers – in many cases polyethylene, polypropylene or polystyrene. 2.

Historical outline

While the first man-made polymers were derived from biomass resources, they were more and more displaced by petrochemical polymers parallel to the growth of the petrochemical industry since the
1930ies. Since the 1980ies and especially in the 1990ies, a comeback of bio-based polymers is observable in certain application areas. One of the main drivers for this development in the last two decades was the goal to provide the market with polymers that are biodegradable. In principle, biodegradable polymers can also be manufactured from petrochemical raw materials. But bio-based polymers, defined here as polymers that are fully or partially produced from renewable raw materials, have so far played the most important role in the domain of biodegradable polymers. These developments have also been a stimulus for R&D on bio-based polymers which are not biodegradable.
In Europe, biodegradable polymers were originally developed and introduced to the markets for two main reasons. Firstly, the limited volume of landfill capacity became more and more a threat and secondly, the bad general image of plastics in public called for more environmentally friendly products.
While the first issue has largely disappeared from the top of the agendas due to the introduction of plastics recycling schemes and due to newly built incineration plants, the environmental performance is currently the main argument for bio-based polymers, including their biodegradable representatives. Apart

1 Mainly pellets, i.e. granules; not to be confused with transportation pallets.

5

from consumer demand for environmentally friendly polymers ("market-pull"), technological progress is offering more and more possibilities ("technology push").
For many decades, cellulose polymers played a key role in a wide range of applications, for example apparel, food (e.g. for sausages) and non-plastics (e.g. varnishes). In the meantime, these biobased polymers have lost important markets mainly to polyolefins. On the other hand, attempts are made to develop new cellulose polymer markets in the area of films, fibres and non-plastics and for natural fibre composites (N.N., 2002).
Since the 1980s, more and more types of starch polymer have been introduced. To date, starch polymers represent by far the largest group of commercially available biodegradable materials. At the outset, simple products such as pure thermoplastic starch and starch/polyolefin blends were introduced.
Due to their incomplete degradability, the latter had a negative impact on the public attitude towards biodegradable polymers. It took many years to compensate this damage, which suceeded largely by introduction of more advanced copolymers consisting of thermoplastic starch and petrochemical copolymers. Research and development has continued in the meantime and embraces nowadays also nanotechnology (e.g. nanoparticle starch fillers as a substitute for carbon black in tyres). Given the historical development it is also interesting to note that polyethylene in non-stabilised and modified forms has returned to the market of biodegradable materials for certain applications (e.g. mulch films).
Widespread R&D activities were conducted to develop cheaper and simpler ways of producing polyhydroxyalkanoates (PHA), reaching from production by fermentation to direct synthesis in crops.
While considerable progress was undoubtedly made (Kopf, 1999), Monsanto terminated their activities in this area in 1999 since the envisioned PHA yields for the production in crops (e.g. maize) were not reached. Being one of the most important players in the field at that time, Monsanto's drawback revived principal doubts about the feasibility and the sensibleness of commercialising large-volume bio-based polymers (e.g. Gerngross and Slater, 2000). Nevertheless R&D has continued in multiple public and private organisations.
Major progress has been made concerning the production of other types of bio-based polymers at industrial scale. In 2001, Cargill Dow, a joint venture of Cargill and Dow, started up a plant in Nebraska for the manufacture of polylactic acid (PLA) with a total capacity of 140 kt per year; by the end of 2002 a lactic acid production facility with a capacity of 180 kt will start up. Apart from being the monomer for PLA, lactic acid has also the potential to become a new (bio-based) bulk chemical from which a variety of other chemicals and polymers can be produced (acrylic acid, propylene glycol, propylene oxide and others). In 2003, DuPont intends to commercialise its biotechnology process for the production of polytrimethylene terephthalate (PTT). While the mentioned processes for PLA and PTT run on starch or sugar plants, advance in R&D on vegetable oils (lower costs, taylored characteristics) offers further opportunities, e.g. for the production of nylon (nylon 9,9; Krämer, 1992). Moreover, R&D for the production of bio-based polyols for polyurethane synthesis is quite advanced (Schmidt and Langer, 2001). For polyethylene made of bioethanol and for several other products (e.g. succinic acid and levulinic acid), first assessments on the techno-economic feasibility indicate that these might be interesing candidates for the medium to longer term (Bozell and Landucci, 1993; Nossin et al., 2002).
Biotechnology, including genetic engineering, may play a key role in the further development.
Moreover, emerging nanotechnology entails interesting potentials for bio-based polymers. Biotechnology and nanotechnology offer new opportunities both for biodegradable materials and for non6

degradable polymers. It may well be that current developments will lead to an increasing importance of bio-based, non-degradable polymers while degradable polymers might be limited to relatively small markets. 3.

Methodology of LCA

A life cycle assessment (LCA) consists of four independent elements (ISO, 1997-1999; CML, 2001), i.e.
(i)
the definition of goal and scope,
(ii)
the life cycle inventory analysis,
(iii) the life cycle impact assessment and
(iv) the life cycle interpretation.
The definition of the goal and scope (i) includes a decision about the functional unit which forms the basis of comparison, the product system to be studied, system boundaries, allocation procedures, assumptions made and limitations. The functional unit can either be a certain service or a product, with the latter being the usual choice for the type of studies reviewed here (e.g., comparison of 1 m3 loose-fill packaging material made of starch polymer versus polystyrene). Critical LCA issues regarding bio-based polymers are, among others, the share of renewable versus fossil raw materials, the way of growing the agricultural raw materials (intensive versus extensive cultivation), the type of conventional polymer that is chosen as a reference and the mix of waste management processes assumed for both the bio-based and the petrochemical polymer (landfilling, incineration, recycling, composting and digestion). It is generally assumed that the carbon dioxide originating from biomass is equivalent to the amount which was previously withdrawn from the atmosphere during growth and that it therefore does not contribute to global warming (fossil fuels required for transport, processing the crops and producing auxiliaries, e.g. fertilizers, are accounted for separately).
The life cycle inventory analysis (ii) involves data collection and calculation procedures to quantify the total system's inputs and outputs that are relevant from an environmental point of view, i.e. mainly resource use, atmospheric emissions, aqueous emissions, solid waste and land use.
The life cycle impact assessment (iii) aims at evaluating the significance of potential environmental impacts using the results of the life cycle inventory analysis. One important goal of the life cycle impact assessment is to aggregate outputs with comparable effects (e.g. all greenhouse gases or all acidifying components) by use of so-called characterisation factors2. This leads to a limited number of parameters, called impact categories. As an optional step, the results by impact categories can be divided by a reference value (e.g., total greenhouse gas emissions of a country) in order to better understand the relative importance of the various impacts; this step is referred to as normalisation. Finally it is, in principle, possible to aggregate the results determined for the various impact categories. However, this valuation step is based not only on scientific facts but also on subjective choices and societal values. So far, there is no generally accepted methodology to translate life cycle inventory data to highly aggre2 Characterisation factors hence serve for aggregation within the impact categories - e.g., to determine the total greenhouse gas effect of a gas mixture containing CO2, CH4 and N2O. Characterisation factors are sometimes also referred to as equivalence factors.

7

gated - let alone, single-score - indicators3. In some of the LCAs reviewed in this chapter, two singlescore aggregation methods have been applied (Eco-indicator '95, Umweltbelastungspunkte; see e.g.
Dinkel et al., 1996). Given the missing general acceptance of these approaches, the results will, however, not be discussed.
The life cycle interpretation (iv) is the final step of the LCA where conclusions are drawn from both the life cycle inventory analysis and the life cycle impact assessment or, in the case of life cycle inventory studies, from the inventory analysis only. As an outcome of the interpretation stage, recommendations can be formulated which, for example, may be directed to producers or policy makers.
The main objective of this chapter is to review full-sized LCAs. However, a few more studies were taken into account that are much more limited in scope (e.g., by studying only non-renewable energy use and CO2 emissions). It was felt that these studies nevertheless contribute to a better understanding of the environmental aspects because they either address materials that have not been studied from this angle so far or because they provide an indication about how certain or uncertain the results are.

4.

Presentation of comparative data

In total, twenty publications were reviewed seven of which deal with starch polymers, five with polyhydroxyalkanoates (PHA), two with polylactides (PLA), three with other bio-based polymers and another three with composites based on natural fibres. The dominance and the size of studies analysing starch polymers reflect the current economic importance of this type of material among the bio-based polymers.4 Annex 1 provides an overview of the key features for most of the reviewed publications.
Regarding the choice of the functional unit some of the studies only analyse the production and waste management of materials in the form of pellets without referring to a specific application of use while other studies refer to a certain type of end use. Analyses for pellets have the advantage that they provide a first impression about the environmental advantages or disadvantages. For example, if the environmental performance is not attractive at the material level (pellets), there is a good chance that this will also be true at the product level. On the other hand, studies that exclusively analyse the materials level suffer from the drawback of not taking decisive parameters at the end use level into account, mainly concerning
- materials processing, where the amount of material required to manufacture a certain end product might be higher or lower than for petrochemical polymers
- transportation, which can be substantial for end products with a low density such as loose fill packaging material
- the use phase, where consumer behaviour can play a role (e.g., in the case of compost bins without an inliner where the way of cleaning the bin has a large influence on the overall environmental impact)

3 According to the draft ISO standard for the life cycle impact assessment "weighting, as described in [paragraph] 6.4, shall not be used for

comparative assertions, disclosed to public" (EN ISO 14 042, draft 1998, paragraph 9).
4 A tabular overview of the object, regional and temporal scope, system boundaries (production, use phase and waste management) and

parameters (e.g. the type of impact categories) of the most detailed studies covered in this chapter can be found in Patel (2003).
8

- the waste stage where logistics and recycling processes can be tailored to a specific product or product group.
These aspects will be explained in more detail in the following sections. The aspects listed show that both approaches – the analysis at the material level and the end product level – provide valuable insight.
They are therefore both included in this review.
The regional scope (compare Annex 1) is relevant due primarily to national differences in CO2 emissions from electricity generation (power is used for plastics production) and the type of waste management infrastructure in place (e.g., treatment of municipal solid waste by incineration, landfilling, recycling and/or composting).
Studies for biodegradable polymers usually include composting as a waste management option
(Annex 1). The output of the composting process, i.e. compost, can be used for soil amelioration. Pathogenic microorganisms are eliminated. Organic nitrogen and phosphorous compounds are converted into inorganic compounds which can be made use of by plants (Estermann, 1998). The option of waste management by digestion is taken into consideration only by one of the studies reviewed (Würdinger et al.
(2001).
The number of impact categories varies greatly (see Annex 1), giving an indication of the differences in depth of analysis among the various studies. The studies differ considerably also in the amount of published background data and the degree of detail regarding explanations about the methodology and results. In some cases the results are given in natural units while in others, they are expressed as an index relative to the reference case, which makes it more difficult for the user of this information to draw comparisons with other sources.

4.1

Starch polymers

All studies discussed in this section deal with thermoplastic starch (TPS) which is manufactured through destructurization of starch in presence of specific amounts of plasticizers and under certain extrusion conditions. Depending on the type of application either pure starch polymers or various types of blends with different shares of petrochemical copolymers are used.
4.1.1 Starch polymer pellets
Dinkel et al. (1996)
The Swiss Federal Agency for the Environment, Forests and Landscape (BUWAL) commissioned
Dinkel et al. (1996) to prepare the first detailed and publicly available LCA for bio-based polymers. This study analyses starch polymers and compares them to polyolefins. The LCA refers to Switzerland. The thermoplastic starch polymers are based on two starch sources, i.e. potatoes (85% of input) and maize
(15%). The system studied covers the entire production process (cradle-to factory gate) and the waste management stage (compare Annex 1). Two types of waste management have been distinguished, i.e. firstly a combination of combustion in municipal solid waste incineration (MSWI) plants and landfilling and secondly, composting. For most impact categories, this difference has only little impact on the data
(Table 1). The largest difference is found for the indicator "deposited waste" which is clearly lower in
9

the case of composting. The use of compost contributes to soil amelioration and it may replace synthetic fertilizers to some extent. If incinerated in waste-to-energy facilities, starch polymers yield certain amounts of electricity and/or steam. In contrast to other LCAs (see below) the study prepared by Dinkel et al. (1996) does not ascribe any credits to these benefits.
Dinkel et al. compare thermoplastic starch to virgin Low Density Polyethylene (LDPE; see Table 1). These data originate from an ealier study commissioned by BUWAL (Habersatter, 1991). The waste management assumed for LDPE consists of 80% incineration and 20% landfilling. Based on these assumptions the Dinkel et al. come to the conclusion that thermoplastic starch performs better than
LDPE in all impact categories except for eutrophication (see row 4). The use of TPS is particularly advantageous with regard to energy resources, greenhouse gas (GHG) emissions, human toxicity and salinization. GHG emissions are reported to be dominated by CO2 while N2O emissions (from agriculture) and methane emissions (from energy supply) are of minor importance (5% and 1-2% respectively of the total GHG emission potential; Dinkel et al., 1996, p.58). The impacts on biodiversity and soil quality were assessed in qualitative terms. Here, no additional negative impacts were determined if starch crops are grown on areas which are currently used for agricultural purposes. In contrast, the effects are clearly negative if natural areas are displaced (Dinkel et al., 1996, p.12).
Dinkel et al. (1996) conclude that the preferences among the environmental targets determine whether starch polymers are found to be environmentally attractive. If the reduction of eutrophication was the prime objective then starch polymers would not represent an attractive option. Regarding biodiversity, the type of land used plays an important role. For all the other parameters the results are in favour of starch polymers.
Table 1:

LCA results for thermoplastic starch (TPS) and Low Density Polyethylene (LDPE) (Functional unit = 100 kg of plastic material; Dinkel et al., 1996, p.13 and pp.51-53)
All data refer to
100 kg plastics

Energy resources [MJ]

Ozone
GHG
precursors emissions [kg
[kg ethylene
CO2 eq.] eq.] Human toxicity [a m3]

Acidification
[kg SO2 eq.]

Eutrophication
[kg PO4 eq.]

Ecotoxicity
[d l]

Salinization
+
[H /mol]

Deposited waste [10-3 EPS]*)
5.1 +/- 10%

(1)

TPS (80% MSWI,
20% landfilling)

2550 +/- 5% 120 +/- 15% 0.47 +/- 20%

20 +/- 40%

1.09 +/- 5% 0.47 +/- 40% 2.8 +/- 55%

180 +/- 15%

(2)

TPS
(100% composting)

2540 +/- 5% 114 +/- 15% 0.50 +/- 20%

20 +/- 35%

1.06 +/- 5% 0.47 +/- 40% 2.8 +/- 55%

180 +/- 15% 0.72 +/- 10%

(3)

LDPE (80% MSWI,
20% landfilling)

9170 +/- 5% 520 +/- 20%

1.3 +/- 15%

70 +/- 60%

1.74 +/- 5% 0.11 +/- 55% 4.6 +/- 25%

860 +/- 10%

5.5 +/- 5%

36%

29%

21%

93%

(4) Ratio (1)/(3)

28%

23%

63%

427%

61%

*) EPS: Environmental Point System

Patel et al. (1999)
As the LCA just discussed, the analysis by Patel et al. (1999) is restricted to starch polymer pellets and compares them with pellets made of polyethylene. The main difference to the other studies discussed in this chapter is that Patel et al. (1999) compare various blends with different shares of petrochemical copolymers. Information about the composition of the blends were provided by starch polymer manufacturers (Novamont, Biotec). It was assumed that both the starch polymers and polyethylene are incinerated in MSWI plants after their useful life. No credits have been assigned to steam and/or electricity generated in waste-to-energy facilities. The results of this analysis that is restricted to energy and CO2
10

are shown in Table 2. According to this comparison starch polymers offer saving potentials relative to polyethylene in the range of 28-55 GJ/t plastic and 1.4-3.9 t CO2/t plastic depending on the share of petrochemical co-polymers5.
Table 2:

Energy requirements and CO2 emissions for different types of starch polymers and for
LDPE (Patel et al., 1999)
Share of petrochemical compounds

Cradle-to-factory gate energy use 1)

Fossil CO2 emissions throughout life-cycle
(production and waste incineration) % (wt)

GJ/t product

kg CO2/t product

0%

25.4

1140

15%

24.9

1730

52.5%

48.3

3360

TPS/polycaprolacton 3)

60%

52.3

3600

LDPE 4)

100%

80.6

4840

Type of plastic

TPS 2)
TPS/polyvinyl alcohol
TPS/polycaprolacton

1)
2)

3)
4)

3)

3)

Non-renewable energy (fossil and nuclear)
Source of data in this row: Dinkel et al. (1996), p.51 (without waste management).
The CO2:energy ratio according to this dataset is very low (45 kg CO2/GJ). The reason might be co-firing of biomass waste.
Patel et al.(1999)
Boustead (1999). Embodied carbon: 3140 kg CO2/t PE.

4.1.2 Starch polymer loose fills
Estermann et al. (2000)
The Italian starch polymer manufacturer Novamont commissioned Estermann et al. (2000) to conduct an
LCA for loose fill packaging material made from Novamont's product Mater-Bi PE01U and from expanded polystyrene (EPS). Mater-Bi PE01U is a starch polymer containing about 15% of polyvinylalcohol. The study evaluates the use of Mater-Bi loose fills in Switzerland. The entire production chain and waste management is included in the system boundaries while the use phase is excluded (Annex 1). The transportation of loose fills is also included in the LCA. This is of general importance for products with low density due to the relatively high energy requirements for transportation. It is, moreover, of particular importance for the products studied here since the density of starch polymer loose fills is about twice as high than that of EPS loose fills. Regarding waste management, incineration has been assumed for
EPS loose fills while composting has been assumed for Mater-Bi loose fills (more than 90% of all organic waste delivered to commercial plants in Switzerland is processed by composting, the rest is fed to digestion plants). The calculations are based on composting in open stacks since more than 80% of the organic waste in Switzerland is composted in this way while the remainder is treated in boxes located in
5 It must be borne in mind that there are still considerable uncertainties even for conventional, fossil fuel-based polymers (see Chapter 6).

11

buildings (Estermann, 1998). An important assumption made here is that 60% of the carbon absorbed in the vegetable material is released to the atmosphere during composting (97% as CO2, 3% as CH4) and that the rest (40%) is sequestered in the compost.6 The authors of the LCA consider these data to be particularly uncertain (Estermann and Schwarzwälder, 1998, p.19; Estermann et al., 2000, p.16).
The polvinylalcohol (PVOH) required for the manufacture of Mater-Bi PE01U is of petrochemical origin. The production of PVOH is reported to be the largest consumer of energy resources throughout the life cycle of starch polymer loose fills (Estermann et al., 2000, p.19). The authors consider the data used for PVOH to be another major source of uncertainty (Estermann et al., 2000, p.28).
To account for these data uncertainties when comparing the results for starch and EPS loose fills, threshold values were determined. For example, the ecological damage is considered "significantly higher" if the impact is at least double as high in the case of energy and waste and if there is a difference of at least a factor of five for all the other impact categories. Similarly, threshold values are determined to quantify the relationships "higher", "comparable", "lower" and "significantly lower" (Estermann et al.,
2000, p.15).
Two different approaches were used to generate aggregated values for the various impact categories (Figure 1; compare also Annex 1). One of them is based on Dinkel et al. (1996) who use the characterisation factors generated by Heijungs (1992) and various other sources (9 impact categories); in the other approach characterisation factors according to Eco-indicator '95 (Goedkoop, 1995) are used (8 impact categories). Since there is some overlap between the two there are 13 impact categories in total
(Estermann et al., 2000, p.26). In eight of these 13 impact categories the production and disposal of Mater-Bi loose fill causes less environmental damage than EPS loose fill. The environmental impact of Mater-Bi loose fill is reported to be significantly lower for the categories winter smog, air toxicity and carcinogeneity. The impact of Mater-Bi loose fills is lower than EPS loose fills with regard to energy use, global warming, acidification, ozone creation/summer smog and heavy metals. In two categories, MaterBi loose fill has a larger environmental impact than EPS loose fill (salinization and deposited waste) while the effects are comparable for the three remaining categories (eutrophication, toxicity water and ozone layer depletion). The overall conclusion of this LCA is that Mater-Bi loose fills are ecologically less damaging than EPS loose fills (Estermann et al., 2000, p.27).

6 These assumptions have been described in the LCA study for bags prepared by Estermann and Schwarzwälder (1998), p.21.

12

Figure 1:

LCA for 1 m3 loose fills: Environmental performance using two different methodologies in the impact assessment stage (Estermann et al., 2000)
Impact categories based on: Heijungs et al. (1992); Dinkel et al. (1996)
500

Mater-Bi loose fill
400

EPS loose fill

%

300
200
100

Deposited waste Salinization

Toxicity water Toxicity air

Eutro phication Acidification

Ozone precursors GHG emissions Energy resources 0

Impact categories based on: Eco-indicator '95 (Goedkoop, 1995)
500
400

Mater-Bi loose fill
EPS loose fill

%

300
200
100

Summer smog

Winter smog

Carcinogeneity

Heavy metals

Eutrophication

Acidification

Ozone layer depletion GHG emissions 0

Würdinger et al. (2001)
Together with the LCA on starch polymers prepared by Dinkel et al. (1996) the LCA prepared by
Würdinger et al. (2001) is the most detailed study that is publicly available. This study was prepared by a project consortium comprising the organisations BIFA, IFEU and Flo-Pak. Two types of loose fills are compared: one based on starch polymers and the other on EPS. The study follows a 2-step-approach:
- In the first step, various options of production, use and waste management are evaluated sequentially: First, the effects of variations in production are studied while the assumptions made for the use phase and for waste management remain unchanged. This is followed by similar sensitivity analyses for the use phase and for waste management. All these analyses are prepared both for starch polymer (Table 3) and EPS loose fills (not reported here).
13

- In the second step, various scenarios are formed by combining selected options in the production, use and waste management stage (Table 4).
For starch polymer loose fills, the first column of Table 3 provides a ranking of the various options studied in production, in the use phase and in waste management. Twenty enviromental parameters were determined. The ranking shown in Table 3 has been determined by comparing for how many impact categories the environmental damage is lower in one case compared to the other without normalisation and valuation. Similar comparisons were also made for EPS loose fills (results not displayed).
The results of the scenarios studied in the second step (Table 4) are shown in Table 5. The main results can be summarized as follows:
- Among the scenarios for starch polymer loose fills, the environmental impacts are lowest for scenario Starch IV, followed by Starch III, Starch II and Starch I (see Table 4).
- Among the scenarios for EPS loose fills, the environmental effects are lowest for scenario EPS IV, followed EPS III, EPS II and EPS I.
- As shown in Table 5 loose-fill production from maize (Starch I) and from virgin EPS (EPS I) assuming no recycling of the products in neither of the two cases and waste management according to current practices are roughly comparable in environmental terms (9 impact categories in favour of starch, 11 impact categories in favour of EPS).
- EPS loose fills exclusively produced from post-consumer waste score better than starch polymer loose fills in most cases (Table 5). Exceptions are the comparisons Starch III – EPS II, Starch IV – EPS
II and Starch IV – EPS III.
The assumption that all EPS loose fills can be manufactured exclusively from post-consumer waste might not be realistic at the national scale and over longer periods of time due to logistics and cost restrictions. From this point of view the choice of the production process might be too optimistic in the scenarios EPS II, EPS III and EPS IV (see Table 4). However, the comparison of these results with those for the starch polymer scenarios show that important tradeoffs exist between recycling and bio-based polymers: According to Würdinger et al. (2001) starch polymers can hardly compete with petrochemical polymers on environmental grounds if the latter have been produced from recycled materials and if the loose-fills (from petrochemical polymers) are also recycled after use. It can be concluded that the use of renewable raw materials for the production of recyclable materials offers unexploited potentials which deserve further R&D.

14

Table 3:

Overall environmental ranking for starch polymer loose fills according to step 1 of the
LCA study by Würdinger et al. (2001)
Phases throughout the life cycle

Environmental performance Best

Production

Use

Waste management

Wheat, extensive
(13/1/2/2) *)
Potato, with effluent use (6/7/4/1)

4 cycles (within company) (20/0/0/0)
4 cycles (recycling station) (9/10/0/0)

Optimized MSWI plant
(12/3/0/3)
Standard MSWI plant (2/4/0/3)

3rd

Corn (maize)
(4/3/6/3)

2 cycles (within company) (0/1/0/0)

4th

Potato
(0/3/0/8)
Wheat, intensive
(2/2/3/11)

- Current MSW m'mt practice
(3/2/3/7)
- Digestion (4/2/3/8)
- DSD (blast furnace) (2/1/6/5)

1 cycle (basis sc.)
(0/0/12/8)
1 cycle (w/o allocation) Composting (3/0/3/8)
(0/0/0/20)

2nd

Worst

In Step 1, the options of production, use and waste management listed in this table are studied separately and sequentially in a ceteris paribus approach; ranking of the environmental performance has been determined by comparing for how many of the 20 impact categories the environmental damage is lower in one case compared to the other (comparison without weighting). The following 20 impact assessment have been taken into account: GHG emissions, carcinogeneity, eutrophication, acidification, diesel particles, ozone precursors, ozone precursors
N-corrected, use of natural land, ozone depletion, eutrophication, cumulative fossil energy demand, cumulative nuclear energy demand, oil equivalents, lead, sulphur dioxide, fluorinated hydrocarbons, ammonia, nitrous oxide, adsorbable halogenated organic compounds, biocide use. Example: Among all the production scenarios studied, the option marked with *) scores best in
13 impact categories, it obtains the second-best score for one impact category, the second-toworst position in two cases and the worst score also for two impact categories. This result is abbreviated by (13/1/2/2). The number of scores does not add up to a total of 20 in many cases because partly 5 and partly 6 grades have been distinguished in the BIFA study while only 4 grades are reported here.

Table 4:

Definition of four starch polymer scenarios and four EPS scenarios in step 2 of the LCA study on loose fill packaging material by Würdinger et al. (2001)
Starch I

Starch II

Starch III

Starch IV

EPS II

EPS III

EPS IV

Virgin
Recycled
Polystyrene PS1)
(PS)

Recycled
PS1)

Recycled
PS2)

2 cycles

1 cycle

2 cycles

2 cycles

2 cycles

Optimized
MSWI plant

Current
MSW m'mt practice4) DSD (blast Open-loop- Optimized furnace) recycling
MSWI plant

Production

Maize

Wheat, intensive Potato, with Wheat, effluent use extensive

Use3)

1 cycle

2 cycles

2 cycles

Waste
Current
Compos- Digestion management MSW m'mt ting practice4) 1)

EPS I

1/3 EPS packaging, 1/3 MC/CD covers, 1/3 pre-consumer waste.
MC/CD covers
3)
1 cycle = single use of loose fill packaging material; 2 cycles = re-use of loose-fill packaging material.
4)
Refers to the average situation in Germany: 30% incineration (including waste-to-energy facilities) and 70% landfilling.
2)

15

Table 5:

Comparison of scores between the four starch polymer scenarios and the four EPS scenarios (based on Würdinger et al., 2001)
Number of impact categories with lower damage caused by ...
...starch polymers

...EPS

Overall judgement Starch I - EPS I
9
11 comparable Starch II - EPS II
5
15
EPS better
Starch III - EPS III
7
12
EPS better
Starch IV - EPS IV
7
12
EPS better
Starch II - EPS III
3
17
EPS better
Starch III - EPS II
11
9 comparable 3
17
EPS better
Starch II - EPS IV
Starch III - EPS IV
4
16
EPS better
Starch IV - EPS II
16
4
Starch better
9
10 comparable Starch IV - EPS III
The abbreviations of the scenarios used in the first columns are described in Table 5.
The number of scores do not add up to a total of 20 if there is at least one impact category for which the results are identical for starch polymers and EPS.

4.1.3 Starch polymer films and bags
Dinkel et al. (1996)
The study prepared by Dinkel et al. (1996), which was partly discussed in Section 4.1.1, contains also a comparative LCA for films made from thermoplastic starch (TPS) and polyethylene (LDPE). The study refers to Switzerland and assumes that 80% of the waste is incinerated and the remaining 20% is landfilled. The results are again lower for TPS compared to LDPE for most impact categories with eutrophication and deposited waste being the main exceptions; for acidification and exotoxicity the impacts are practically identical (Table 6). The results for human toxicity are reported to be subject to major uncertainties (Dinkel et al., 1996, p.60). Compared to the analysis prepared by Dinkel et al. for materials (Section 4.1.1) the starch polymer's advantage compared to LDPE is smaller. This is due to the assumption that the polymer input required to manufacture a given area of film is about 60% larger for starch polymers compared to LDPE (22.1 kg TPS compared to 13.8 kg LDPE for 100 m2 of film of 150µm). In the meantime, the raw material requirements for starch polymer films have decreased and now exceed that for LDPE by only 30% (personal communication Novamont, 2001).

16

Table 6:

LCA results for films (100 m2, 150µm) made from TPS compared to LDPE; assumed waste management: 80% incineration, 20% landfilling (Dinkel et al., 1996)
Energy
resources
[MJ]

Ozone
GHG
precursors emissions [kg ethylene
[kg CO2 eq.] eq.] (1) TPS Film

649 +/- 5%

25 +/- 15%

0.10 +/- 20% 4.3 +/- 40% 0.24 +/- 5% 0.13 +/- 40% 0.62 +/- 75% 40 +/- 15% 1.1 +/- 10%

(2) LDPE Film

1340 +/- 5%

67 +/- 20%

0.18 +/- 15% 9.7 +/- 60% 0.24 +/- 5% 0.02 +/- 50% 0.65 +/- 40% 120 +/- 8%

(3) Ratio (1)/(2)

48%

38%

56%

Human toxicity 3
[a m ]

44%

Acidification
[kg SO2 eq.]

100%

Eutrophication
[kg PO4 eq.]

687%

Ecotoxicity
[d l]

95%

Salinization
+
[H /mol]

33%

Deposited waste -3
[10 EPS]

0.8 +/- 5%
138%

Estermann and Schwarzwälder (1998)
In this study prepared by Estermann and Schwarzwälder (1998) for Novamont, biodegradable waste bags made from Mater-Bi ZF03U/A material are compared to High Density Polyethylene (HDPE) bags and Kraft paper bags. Mater-Bi ZF03U/A is a blend of TPS and polycaprolacton. It is assumed that these bags are used as liners for compost bins. The comparison is made for the smallest bags which were commercially available in Switzerland in 1998 and which could be used for a 10-litre bin. The considerable difference in the size of the bags assumed (Mater-Bi bag: 16.6 l; paper bag: 13.6 l; PE bag: 35.6 l) hence reflects the standard products available on the Swiss market. Most energy resources for the production of Mater-Bi bags are required to manufacture polycaprolactone (PCL). Mater-Bi bags and paper bags are assumed to be composted while PE bags are incinerated. Data for composting are considered to be particularly uncertain (see Section 4.1.2 and Estermann and Schwarzwälder, 1998, p.19).
To account for the data uncertainties when comparing the results for the three materials, threshold values were determined by analogy to the study for loose fills prepared by Estermann and colleagues
(2000; Section 4.1.2). The impact categories distinguished here are also identical with those for loose fills (Section 4.1.2). In eleven of these 13 impact categories Mater-Bi compost bags cause less environmental damage than paper compost bags (for energy resources, GHG emissions, ozone precursors/summer smog, acidification, eutrophication, toxicity air, toxicity water, deposited waste, heavy metals, carcinogeneity, winter smog; Figure 2). In the two remaining impact categories, Mater-Bi bags cause a comparable or a greater degree of environmental damage (salinization; ozone layer depletion).
The Mater-Bi compost bags and the PE multipurpose bags are equivalent in seven impact categories
(energy resources, acidification, eutrophication, toxicity water, ozone layer depletion, carcinogeneity, winter smog). The Mater-Bi bag achieves better scores in four categories (GHG emissions, toxicity air, ozone precursors, heavy metals) but worse results in the two remaining categories (salinization and deposited waste). However, Mater-Bi bags have a smaller environmental impact than PE multipurpose bags in ten categories if one considers that the waste adhering to the bags is incinerated together with the bags (ozone layer depletion and carcinogeneity). It is not specified in the LCA whether the organic waste is considered as neutral in CO2 terms (this would be expected due to its predominantly biogenic origin) and whether an energy yield according to its heating value has also been taken into account.

17

A sensitivity analysis was conducted to study whether the production of maize in France instead of Switzerland changes the final results. It is concluded that this is not the case since maize production has a relatively small influence on the total life cycle of Mater-Bi bags and since there are only slight differences between maize production in France and Switzerland. In another sensitivity analysis it was taken into account that the organic waste adhering to the PE bags is co-combusted in MSWI plants.
These calculations show a clearer environmental advantage for Mater-Bi bags ("significantly better").
However, as above, it is unclear whether the CO2 neutrality of natural organic waste and the additional energy yield have been taken into account.
No sensitivity analysis was conducted to determine how the availability of PE bags of the same size as Mater-Bi and paper bags would influence the results.
LCA results for bags made from thermoplastic starch (TPS) compared to polyethylene multipurpose bags and compost paper bags (Estermann and Schwarzwälder, 1998)
Impact categories based on: Heijungs et al. (1992); Dinkel et al. (1996)
500

Mater-Bi bag
400

Paper bag
PE multipurpose bag

%

300
200
100

Deposited waste Salinization

Toxicity water Toxicity air

Eutro phication Acidification

Ozone precursors GHG emissions Energy resources 0

Impact categories based on: Eco-indicator '95 (Goedkoop, 1995)
500

Mater-Bi bag

400

Paper bag
300

%

PE multipurpose bag

200
100

18

Summer smog

Winter smog

Carcinogeneity

Heavy metals

Eutrophication

Acidification

Ozone layer depletion 0

GHG emissions Figure 2:

Estermann (1998)
In this study prepared by Estermann (1998) for the Kompostforum Schweiz, various types of options for the collection of organic household waste were compared, among them three bags made from biodegradable polymers, one PE bag and finally cleaning of the compost bin instead of the use of an inlet7:
The collection of organic kitchen waste in a compost bin without inlet results in the lowest environmental impact of all options if the bin is cleaned after use with cold water or with washing-up water.
Since, however, 54% of the surveyed Swiss households use hot fresh water and 38% use, in addition, detergents for this purpose, the average environmental impact is highest for those households that do not use an inliner. Compared to these impacts in the use phase, the environmental damage originating from the production and waste management of the inlets is relatively small. The production of biodegradable inlets has a larger environmental impact than PE inlets. On the other hand the use of PE inliners results in environmental impacts especially due to incineration of the bag and adhering compostable waste (toxicity air and toxicity water). Interestingly, one of the compostable inlets (CompoBag) performs very well compared to the other biodegradable products in spite of being produced exclusively from petrochemical raw materials (the impact categories salinization and deposited waste are exceptions, see Figure 3). It is also interesting to note that the results for air and water toxicity differ considerably for the COMPOSAC and ecosac although both are produced from the same material (this difference is explained only partly by the difference in size). It is unclear whether any credits have been allocated to the co-production of electricity and/or steam when incinerating the PE compost bags. It is also unclear whether any credits have been ascribed to composting due to carbon sequestration in the compost. To summarize, the authors of the LCA recommend to clean the compost bin with cold water or with washing-up water. For consumers with higher standards of cleanliness they recommend biodegradable inliners.
Similar calculations were conducted for container inlets with a volume of 240 l (CompoBag
240 l, MaterBi bag 240 l, PE inlet 240 l). According to the LCA results the use of these inlets is comparable in environmental terms with the generally practised cleaning with cold water. In six out of nine impact categories, the biodegradable inliners score better than the PE inliner.

7 In more detail, the five options studied are:

1.) CompoBag 9 l, made from polycaprolacton (PCL) and polyester amide (these raw materials are both made from petrochemical resources). 2.) COMPOSAC 14 l, made from Mater-Bi Z, i.e. a blend of maize starch, PCL and additives.
3.) ecosac 6.5 l, also made from Mater-Bi Z.
4.) PE-Inlet 30x45cm
5.) No inlet (instead the compost bin is cleaned).
19

Figure 3:

LCA for household composting inlets (including average consumer behaviour in the use phase; Estermann, 1998)
100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Energy resources GHG emissions Ozone precursors Acidification

Eutro phication Toxicity air

Toxicity water Salinization

Deposited waste 4-litre container with CompoBag 9 l (PCL + Polyesteramid)

with COMPOSAC 14 l (MaterBi Z = PCL + corn starch)

with ecosac 6.5 l (MaterBi Z = PCL + corn starch)

with PE inlet 30x45cm

4-litre container without inlet (av. cleaning: hot water/detergents)
Note: The highest value has been set to 100%.

4.1.4 Starch nanoparticles as fillers in tyres
Corvasce (1999)
Since 2001, Goodyear has been producing tyres containing biopolymeric fillers derived from starch (a joint development by Goodyear and Novamont). These tyres are reported to have various functional advantages, the most important being controlled stiffness, improved wet skid performance, lower weight and reduced rolling resistance. As Table 7 shows, especially the latter feature leads to lower CO2 emissions: Savings due to lower rolling resistance, which result in fuel savings in the use phase, exceed cradle-to-factory gate emission reduction by factors of 23 to 26. The total savings according to Table 7 represent about 2% (for 3.53 g CO2/km) to 5% (for 9.52 g CO2/km) of the average CO2 emissions of a passenger car. It has to be pointed out that these results have so far not been published as a report or extensive article but only in the form of a presentation by Corvasce (1999). This case study is nevertheless discussed here it sheds light on a new, pr omising bulk application of biopolymers.

20

Table 7:

CO2 emission reduction potential of tyres with biopolymeric fillers (Corvasce, 1999)
CO2 reduction compared to conventional tyres1) g CO2/km
20% weight
50% weight replacement of replacement of carbon black carbon black
Use of starch-based raw materials2)

0.15

0.35

Tyre weight reduction3)

0.03

0.25

Tyre rolling resistance reduction3)

3.35

8.92

Total

3.53

9.52

1)

Averaged values over 30 000 km; tread weight 3.0 kg.

2)

Cradle-to-factory gate: Emission of fossil CO2 during processing minus carbon sequestration in starch during plant growth.
Use phase

3)

4.2 Polyhydroxyalkanoates (PHA)
Gerngross, Slater et al. (1999, 2000, 2001)
The main representatives of polyhydroxyalkanoates (PHA) are polyhydroxybutyrate (PHB) and polyhydroxyvalerate (PHV). For this family of polymers, no full LCAs are available but only studies and estimates comparing the energy requirements and CO2 or greenhouse gas emissions, among them the papers by Gerngross and Slater (2000), Gerngross (1999) and Kurdikar et al. (2001; with Gerngross and Slater as co-authors). In Table 8 their data are compared to LCA data for petrochemical polymers according to
Boustead (commissioned by APME).8 The table shows that the total cradle-to-factory gate fossil energy requirements of PHA can compete with polyethylene (HDPE) depending on the type of the PHA production process. Compared to polyethylene terephthalate (PET), the minimum total energy input for
PHA production (fermentation) is somewhat higher while it is lower compared to polystyrene (PS). In contrast, the process energy requirements of PHA are two to three times higher than for petrochemical polymers (Table 8). Limiting the discussion to these process energy data Gerngross and Slater drew the conclusion that polyhydroxyalkanoates do not offer any opportunities for emission reduction (Gerngross and Slater, 2000; Gerngross, 1999). This finding is valid for certain system boundaries, e.g. for the system 'cradle-to-factory gate' the output of which are plastics pellets. The conclusion is also correct if all plastic waste is deposited in landfills. In contrast, the finding is not correct if other types of waste management processes are assumed within the 'cradle-to-grave' concept. As the last column of Table 8 shows the total fossil energy requirements are practically identical for PE and PHA manufactured by bacterial fermentation. Hence, if combusted in a waste incinerator (without energy recovery), both plastics result in comparable CO2 emissions throughout the life cycle.

8 As mentioned earlier there are still considerable uncertainties even for conventional, fossil fuel-based polymers (see Chapter 5).

21

Table 8:

Energy requirements for plastics production (Gerngross and Slater, 2000; Boustead,
1999)
Cradle-to-factory gate fossil energy requirements, in GJ/tonne plastic
Process
energy

Feedstock energy Total

PHA grown in corn plants

90

0

90

PHA by bacterial fermentation

81

0

81

HDPE

31

49

80

PET (bottle grade)

38

39

77

PS (general purpose)

39

48

87

Data for PHA from Gerngross and Slater (2000).
Data for petrochemical polymers from Boustead (1999).

A more recent publication, co-authored by Gerngross and Slater, studies in more detail the greenhouse gas profile of PHA production in genetically modified corn (Kurdikar et al., 2001). Two alternative energy systems were studied. In one case process energy requirements are covered by natural gas and in the other, biomass energy from the corn stover is used as fuel. The publication focusses primarily on the system cradle-to-factory gate but some data on waste management is also provided. This information has been used in Table 9 to estimate also GHG emissions for two cradle-to-grave systems. It can be concluded that PHA production with integrated steam and electricity generation based on biomass scores better than conventional PE production in all cases, while the opposite is the case if natural gas is used to provide the PHA production process with steam and electricity9. The authors conclude that it is the biomass power and not the renewable feedstock that makes the product preferable to PE from a
GHG point of view.
When drawing conclusions from these results, it must also be taken into consideration that PHA production both by bacterial fermentation and by production in the plants, is in an early stage of development compared to the manufacture of polyolefins and that efficiency improvements are likely to accrue from upscaling of production and from process optimisation. In the medium term, this might result in a better environmental performance of PHA throughout its life cycle compared to PE and PET.

9 Note that the underlying process energy requirements for "PHA, natural gas" in Table 9 is around 100 GJ/t, while the the respective value

for "PHA grown in corn plants" in Table 8 is 90 GJ/t.
22

Table 9:

CO2 emissions from the life cycle of polyhydroxyalkanoates (PHA) and polyethylene
(PE) (Kurdikar et al., 2001; complemented with own assumptions)
Cradle-togate fossil
CO2 eq.

CO2 eq. uptake in biopolymers1) (A)

All values in kg
CO2 eq. / kg polymer (B)

(C)

(D)6)

(E)

(F)7)

(G)

2.0

-

ca. 3.8

2.0

ca. 5.8

ca. 4.8

2.0

1.5

-4.0

2.0

-2.0

-3.0

-

-

1.8

3.1

4.9

2.8

8)

PHA, natural gas

ca. 5.8

PHA, bioenergy

-0.5

PE

1.8

1)
2)
3)
4)
5)

6)
7)
8)
9)

9)

Cradle-toCradle-toCO2 eq. Cradle-toCO2 eq. grave CO2 eq. grave CO2 eq. uptake in gate net embodied in without energy with energy
CO2 eq. ash2) polymer3) recovery4) recovery4) 5)

Uptake of carbon from the atmosphere and fixation in biopolymer.
Carbon fixed in the ash from the boiler (due to incomplete combustion).
Both fossil and biogeneous CO2 is accounted for here. For PHA values in column B and E are identical.
Waste incineration in a plant without resp. with energy recovery
Estimated CO2 credits for 20% electricity yield from waste-to-energy recovery: 1 kg CO2/kg PHA, 2.1 kg
CO2/kg PE (underlying assumptions: Efficiency of electricity generation in average power station = 30%;
CO2 emission factor of fuel mix used = 74 kg CO2/GJ; Heating value, PHA = 18 MJ/kg; Heating value, PE =
42 MJ/kg).
(D) = (A) - (B) - (C)
(F) = (D) + (E)
Including energy use for smaller consumers, i.e. compounding, farming etc.
Small fossil energy input minus credit for surplus electricity produced from biomass

Heyde and Luck (1998, 1996)
Heyde (1998) compared the energy requirements of PHB production by bacterial fermentation using various feedstocks and processes to those of High Density Polyethylene (HDPE) and polystyrene (PS).
The PHB options studied include substrate supply from sugar beet, starch, fossil methane and fossilbased methanol and moreover, in the processing stage, the options of enzymatic treatment and solvent extraction. As Figure 4 shows the energy requirements for biotechnological PHB production can substantially exceed the requirements for conventional plastics, but on the other hand there is also scope to outpace fossil-based polymers in terms of energy requirements (PHB Best Case). An earlier publication by Luck (1996) shows that the choice in the waste management process also has a decisive influence on the results. For example, PHB manufactured in an efficient way and disposed of with MSW (German average) requires more energy resources and leads to higher GHG emissions than HDPE if the latter is recycled according to the German 1995 Packaging Ordinance (64% material recycling). If, on the other hand, the plastics waste is fed to average MSWI plants in both cases, then the results are comparable for energy and GHG emissions.
It can be concluded that energy use and CO2 emissions are nowadays often larger for PHB than for conventional polymers but that there is also scope to avoid this disadvantage if the entire system covering all stages of the life cycle is carefully optimised. In spite of these prospects, Monsanto, a frontrunner regarding PHA, decided in 1999 to postpone further research and to close down their production line based on fermentation. The original strategy had been to produce PHA by fermentation as an in23

terim step on the way towards PHA production in genetically modified plants. The goal of PHA production by fermentation was hence to gain experience with this product and to develop the market. The overall strategy was given up when it turned out that it would not be possible to reach in the short term the target of increasing PHA yields in genetically modified plants from around 3% of dry weight to at least 15% (Anonymous, 1999).
Cradle-to-factory gate requirements of non-renewable energy for the production of various polymers (Heyde, 1998)
Non-renewable energy resources, in MJ/kg

Figure 4:

700
573.2

600
500
400
300
200
100

66.1

73.8

91.7

PHB Best
Case

HDPE

PS

0
PHB Worst
Case

4.3 Polylactides (PLA)
Vink (2001, 2002)
LCA data for polylactides are very scarce. Cargill Dow Polymers, the major manufacturer of this type of bio-based polymer, has only published energy and CO2 data but no comprehensive dataset so far: As shown in Table 10 total fossil energy requirements of PLA are clearly below the respective values for the petrochemical polymers while the process energy requirements are higher for the first commercial
PLA plant (termed PLA-Year 1 in Table 10). Additional fossil energy savings are envisaged for future plants (PLA - Target, Year 5) by use of renewable energy as a fuel source (for power and heat), by energy integration of the PLA unit with the lactic acid facility and by optimised product separation. Further options are changes in feedstocks and in the production processes, such as the direct use of agricultural waste and biomass without the intermediate step of isolating dextrose and the use of improved biocatalysts (Vink, 2001, 2002).

24

Table 10:

Cradle-to-factory gate energy requirements and CO2 emissions for plastics production
(Vink, 2001 and 2002; Boustead, 1999/2000)
Process
Feedstock energy, fossil energy, fossil
[GJ/t plastic] [GJ/t plastic]

Net
Total fossil Fossil CO2 from CO2 absorption,
CO2
energy process energy plant growth
[GJ/t plastic]
[kg/t plastic]
[kg/t plastic]
[kg/t PLA] *)

PLA - Year 1

54

0

54

3450

-2190

1260

PLA - Target, Year 5

7

0

7

520

-2280

-1760

HDPE

31

49

80

1700

0

1700

PET (bottle grade)

38

39

77

4300

0

4300

Nylon 6

81

39

120

5500

0

5500

*) Equals the sum of "Fossil CO2 from process energy" (positive value) and "CO2 absorption, plant growth" (negative value). Data for PLA by personal communication with E. Vink, 2002.
Data for petrochemical polymers from Boustead (1999/2000).

Hakala et al. (1997)
In cooperation with the Neste company, the Technical Research Centre of Finland prepared a comparative LCA for two diaper systems one of which is based on polyolefins (PP, PE) while the other uses
PLA (made from maize, wheat and sugar beet). For most of the parameters studied, the polyolefin-based diaper shows better results than its bio-based counterpart. However, the differences in environmental impacts between the two systems are small. Moreover, the lactide and PLA production process were still under development when the LCA was prepared (Hakala et al., 1997, p.28) which gives the study a more preliminary, indicative quality.
Summarising the information that is currently available one can expect the environmental performance of PLA according to current production methods to be less advantageous than most starch polymers but clearly more beneficial than for PHA. Substantial improvements are expected for PLA for the future.

4.4 Other polymers based on renewable resources
Wolfensberger and Dinkel (1997)
This study was jointly prepared by the Swiss Research Institute for Agriculture (Eidgenössische Forschungsanstalt für Agrarwirtschaft und Landtechnik, FAT) and the environmental consultancy CARBOTECH for the Swiss Federal Office of Agriculture. It contains LCA results for three products:10
(i) Mulch films made of kenaf could be used as an alternative to PE mulch films.
(ii) Loose-fill chips made of china reed represent a potential substitute for EPS chips.

10 In the study co-ordinated by Wolfensberger and Dinkel (1997), these three products are referred to as No.3, No.10 and No.12.

25

(iii) Shreddered and ground china reed, combined with a matrix based on renewable resources, could be used as a substitute for PE in injection moulding.
For kenaf mulch films (i), the LCA shows clear advantages compared to the PE alternatives with regard to the use of energy resources, GHG emissions, ozone precursors, acidification and toxicity. Disadvantages have been identified for eutrophication, biodiversity and soil fertility (the last two indicators were not studied as part of the LCA). Moreover, the economics are less favourable than for the conventional product.
The two products made of china reed (ii, iii) score clearly better than or at least as good as their petrochemical counterparts for most of the indicators studied. For these two products, unresolved technical issues and the forthcoming commercialisation represent the main challenges for the future.
Kosbar et al. (2001)
This LCA compares a lignin-epoxy blend with a pure epoxy resin for the manufacture of printed wiring boards (PWB) as being used in the microelectronics industry. Nowadays most PWBs are disposed of by incineration to reclaim the precious metals content. The intractable nature of epoxy fibreglass laminates combined with the features of the embedded wiring make PWB unlikely candidates for either recycling or reuse. These were important reasons for IBM to analyse bio-based options. Lignin was chosen since its aromatic structure is similar to the phenolic resins in current PWBs. Lignin is available as a byproduct of paper making but it has only few industrial uses. Very little additional energy or processing is reported to be required for making lignin available for resin manufacture. The primary sources for lignin considered in this study were kraft pulping mills in the U.S. and Europe and an Organosolve pilot-scale pulping mill in Canada. As shown in Table 11 the organosolve lignin/epoxy resin is estimated to require
38% less total energy and the kraft lignin/epoxy resin 27% less total energy than currently used resins. If a combination of 50% incineration (for metal reclamation) and 50% disposal as municipal solid waste11 is assumed (not shown in Table 11) the energy requirements are about 2% smaller in both cases. With regard to atmospheric and waterborne emissions it was found that the release can be significantly reduced by lignin-based resins (Table 12). The findings of this project were not exploited by IBM for commercial purposes, mainly because of the varying purity of the available lignin. Organosolve lignin is an attractive raw material due to its low content of ionic contaminants. However, the organosolve pilot plant in Canada was recently shut down. Kraft lignin can also be used for PWB production but it requires substantial treatment to remove ionic contaminants. This imposes a financial disadvantage which was the main reason for IBM not to proceed with industrial application.

11 MSW management in the U.S. in 1996: 55.5% landfilling, 27.3% recycling/composting, 17.2% combustion (Subramanian, 2000)

26

Table 11: Energy requirements for wiring board systems assuming 100% metal reclamation (Kosbar et al., 2001)
GJ/100 kg resin solids

Process energy Feedstock Transporenergy tation Total energy Conventional epoxy*)

10.0

7.16

0.68

17.8

Kraft lignin/epoxy

9.54

2.91

0.44

12.9

Organosolve lignin/epoxy

7.64

2.91

0.45

11.0

Note: Apart from fossil and nuclear energy these data include smaller amounts of renewable energy including also wood energy which is used in the pulp and paper industry.
*) Epoxy resin (FR4) cured with dicyandiamide (DICY)

Table 12: Comparison of air and waterborne emissions from conventional epoxy and lignin/epoxy resins (Kosbar et al., 2001)
Conventional epoxy resin emissions at least 50% greater than lignin resin emissions

Lignin resin emissions at least 50% greater than conventional epoxy resin emissions

Nitrogen oxides
Hydrocarbons
Sulphur oxides
Aldehydes
Methane
Ammonia
Fossil CO2

Nonfossil CO2
Metals
Acid
Sulphuric acid
Suspended solids
Phosphates
Nitrogen

Metal ions
Dissolved solids
Phenols
Sulphides
Oil
Chlorides
Sulphates

Diehlmann and Kreisel (2000a, 2000b)
At the Institute of Technical Chemistry and Environmental Chemistry (ITUC) at Jena University (Germany) Diehlmann and Kreisel (2000a, 2000b) compared a lacquer based on epoxidised linseed oil as thickener with a conventional lacquer based on tripropylene glycol diacrylate / bisphenol-Adiglycidetheracrylate (50:50, on a weight basis). The linseed-oil based lacquer is being produced by two companies in Germany (Dreisol and Lott Lacke). The authors adopt the cradle-to-factory gate concept since the use and waste management are identical for the bio-based and the petrochemical products and therefore have no influence on the conclusions. As Figure 5 shows the linseed oil-based lacquer (average for centralised and decentralised production) scores clearly better than its purely petrochemical counterpart for the indicators presented. The authors conclude that, given the large advantage of epoxidised linseed thickeners, even substantial optimisation of the petrochemical route is very unlikely to change the overall conclusion about the environmental benefits of the bio-based product.

27

Figure 5:

Comparison of environmental impacts of lacquers based on epoxidised linseed oil as thickener versus petrochemical thickener (tripropylene glycol diacrylate / bisphenol-Adiglycidetheracrylate) (Diehlmann and Kreisel, 2000a, 2000b)
100%

~19 MJ/kg

40%
30%
20%

~4 g NOx/kg

50%

36 g NOx/kg

60%

~1.2 kg CO2/kg

70%

217 MJ/kg

80%

9.8 kg CO2/kg

90%

10%
0%

Energy

CO2
Petrochemical

4.5

NOx
Linseed oil

Natural fibre composites

Diener and Siehler (1999)
This brief publication deals with an under-floor panel for the Mercedes A class. The main purposes of this panel are protection of the chassis and improved aerodynamics. In one case the panel is made from fibreglass reinforced polypropylene (PP) and in the other from PP reinforced with flax. The flaxreinforced panel has passed successfully all the technical tests which are very demanding for an underfloor panel due to its exposure to mechanical and thermal stress. The introduction of the flax-reinforced component instead of the fibreglass-based panel in series production had not been decided upon when this chapter was written. As shown in Figure 6 the flax reinforced panel scores better for all environmental impacts studied. For seven out of the ten impact categories, the environmental impact is reduced by close to 10% to 20%, in the remaining three cases the reduction of impacts is higher (30-80%). These reductions of environmental impacts reflect the fact that the manufacture of flax fibre mats requires 80% less energy than fibreglass mats (Table 13). The total energy savings for the entire component are smaller (14%) since the overall environmental impact is dominated by the PP input.

28

Figure 6:

Cradle-to-factory gate environmental impacts for the production of a fibre-reinforced automotive panel (Diener and Siehler, 1999)

Flax & PP

Nonrenewable energy Resources

Waste

Toxicity water

Toxicity air

Eutrophication

Glass fibre & PP

Acidification

Ozone precursors Ozone depletion GWP

100%
90%
80%
70%
60%
50%
40%
30%
20%
10%
0%

Table 13: Energy requirements for the manufacture of mats from fibreglass versus (Diener and Siehler,
1999)
Non-renewable energy requirements (MJ/kg)
Fibreglass mat

Flax fibre mat

Raw materials
Mixture
Transport
Melting
Spinning
Mat production

1.7
1.0
1.6
21.5
5.9
23.0

Seed production
Fertilizers
Transport
Cultivation
Fibre separation
Mat production

0.05
1.0
0.9
2.0
2.7
2.9

Total

54.7

Total

9.55

Wötzel et al. (1999)
In this publication prepared by Braunschweig University and Seeber, which is a manufacturer of automotive components, LCA results for the interior side panel of an Audi A3 are presented.12 13 The conventional side panel is manufactured by injection moulding of virgin acrylonitrile-butadiene-styrene (ABS) and weighs 1125 grammes. The bio-based component is made of a composite consisting of 66 vol-% hemp fibres and 34 vol-% epoxy resin plus hardener; this component weighs 820 grammes. This hempreinforced composite is nowadays manufactured in series production.
Apart from energy use the authors have established emission data for a whole range of solid, liquid and gaseous compounds; however, these have not been aggregated to impact categories14. For the purpose of this review, the GHG emissions have therefore been calculated on the basis of the available data. As Table 14 shows the requirements of non-renewable energy, are by 59 MJ or 45% lower for the hemp/epoxy composite and GHG emissions are 13% lower (table 14).
12 More recently, LCE Consulting and AUDI presented results on a conference for the same component (Flake et al., 2000a). While the system studied differs somewhat compared to the publication by Wötzel et al. (1999), the overall findings are identical. The paper by
Wötzel et al. (1999) was chosen for inclusion in this review since it contains more details of the LCA analysis.
13 One of the authors is co-author of a book on the use of renewable raw materials in passenger cars (Flake et al., 2000b).
14 The single-score indicator "Eco-indicator 95" has been determined in order to derive overall conclusions. As discussed in the

introduction of this chapter, these results will not be discussed here.
29

The results presented so far were based on the cradle-to-factory gate perspective. For this reason they did not address the possibilities of energy recovery and they neither accounted for the use phase of a car where weight reduction can reduce the environmental impacts substantially (modelled by so-called fuel reduction coefficients which depend on the weight of the car). As Figure 7 shows, energy recovery reduces the net energy requirements for the hemp component to 55 MJ (=73-18) and for the ABS component to 87 MJ (=132-45). Compared to the cradle-to-factory gate perspective, inclusion of energy recovery nearly halves the hemp composite's advantage in energy terms (from 59 MJ to {87-55} MJ = 32
MJ). However, the higher weight of the ABS component results in an additional burden during the use phase in the range of 71 MJ (light car) to 118 MJ (heavy car). The savings which are obtainable by natural fibres during the use phase (equivalent to the additional burden) are hence 2.2 to 3.7 times higher than the net benefit including energy recovery (32 MJ). If the use phase is included the overall advantage for the hemp composite amounts 103 to 150 MJ.15 This is equivalent to 1.4 to 2.1-fold of the cradleto-factory gate energy requirements of the hemp composite (73 MJ).

Table 14: Energy requirements and greenhouse gas emissions for the production of one interior side panel of an Audi A3 (Wötzel et al., 1999; own calculations)
All data for prod. of one interior side panel

Unit

ABS copolymer Hemp fibre/epoxy composite

Non-renewable energy MJ

132

73

CO2, fossil

kg CO2eq

5.0

4.2

CH4
N2O

kg CO2eq kg CO2eq

0.4

0.4

0.003

0.160

Total GHG emissions*) kg CO2eq

5.4

4.7

*) Only CO2, CH4 and N2O have been taken into account.
The GWP values (100 year period) assumed are 23 for
CH4 and 296 for N2O.

15 103 MJ = (132 + 71 – 45) MJ – (73 – 18) MJ

150 MJ = (132 + 118 – 45) MJ – (73 – 18) MJ
30

Non-renewable energy requirements related to the life cycle of an interior side panel of an
Audi A3 (Wötzel et al., 1999)
MJ energy / component

Figure 7:

250 max 200

118

Production

150

Energy recovery

100

min
71

132
73

50

18

45

Hemp part

ABS part

Burden use phase

132

132 45

45

0

ABS part with use phase light car ABS part with use phase heavy car Corbière-Nicollier et al. (2001)
This life cycle assessment prepared by Corbière-Nicollier et al. (2001) at the Swiss Federal Instute of
Technology in Lausanne compares a pallet made from fibreglass reinforced polypropylene (GF pallet) with a pallet manufactured from a polypropylene/china reed composite (CR pallet). The GF pallet weighs 15 kg (fibreglass content: 42 weight-%), while the CF pallet weighs only 11.8 kg (china reed fibre content: 53 weight-%). The lifetime of both products has been assumed to be 5 years. However, poor contact between matrix and biofibres currently leads to a strong decrease in material stiffness, calling for further R&D.
Table 15 shows the LCA results including all processes related to production and to incineration of the pallets in waste-to-energy facilities (the use phase is excluded and will be dealt with below). The
CR pallets score better or even much better than the GF pallets for all eight environmental impacts studied.
Table 15: Results by impact categories for fibreglass pallets (GF) and China reed (CR) pallets - System boundaries: production plus incineration with energy recovery (Corbière-Nicollier et al.,
2001)
GF pallet
Energy

CR pallet CR pallet (% GF pallet)

MJ/pal.

1400

717

51%

kg CO2eq./pal.

75.3

40.4

54%

kg ethyleneeq./pal.

0.208

0.133

64%

Acidification

kg SO2eq./pal.

0.653

0.432

66%

Eutrophication

kg PO4eq./pal.

0.0682

0.0628

92%

Human toxicity

kg 1,4 dichleq./pal.

21.2

9.04

43%

Terrestrial ecotoxicity

kg 1,4 dichleq./pal.

5250

4480

85%

Aquatic ecotoxicity

kg 1,4 dichleq./pal.

1.09

0.665

61%

GHG emissions
Ozone precursors

31

Apart from incineration also mechanical recycling was studied. The authors consider a rate of
20% to be realistic for pallets. Since the technical feasibility of recycling CR pallets is unknown, the comparison has been made both with and without recycling of CR pallets. It is concluded that for all impact categories except for human toxicity a recycling level of 20% is not sufficient for the GF pallets to match the low environmental impact of the CR pallets without recycling. The recycling rate for GF pallets would need to be clearly beyond 40% to reach the (environmental) break-even point (such a high recycling rate is not considered reachable by the authors under present conditions). If, on the other hand, both the GF and the CR pallet are recycled, then the difference between the two pallets decreases with higher recycling rates.
Since the real lifetime of the CR pallets is not known (base case assumption: 5 years) CorbièreNicollier et al. conducted a sensitivity analysis. The minimum lifetime necessary for the CR pallet to have an environmental impact lower than or equal to that of the GF pallet was found to be approximately 3 years.
By analogy to the preceding LCA the fuel savings in the use phase due to the lower weight of the
CR pallet were also assessed. Assuming a transportation distance of 5000 km the savings amount to
660 MJ, which is equivalent to more than 90% of the non-renewable energy required to produce a CR pallet (compare Table 15). For a distance of 200 000 km, 2300 MJ are saved, which is more than three times of the production energy required for manufacturing a new CR pallet.
As Table 16 shows the use of China reed as a replacement for fibreglass in pallets is four to ten times more efficient in terms of energy savings per land use than direct combustion of China reed for heat production (substitution of 1000-2300 GJ/ha). Fibreglass substitution in pallets is also a clearly more efficient type of use for China reed than the substitution of polypropylene chips. CR pallets also outscores the starch polymer product and the biofuel options studied. However, caveats that have to be kept in mind are the shortcomings of mechanical properties of CRF pallets according to the current state-of-the-art and their unknown lifetime.
Table 16: Comparison of potential energy savings by biomass cultivation for biomaterials versus bioenergy (Corbière-Nicollier et al., 2001)
Renewable raw material

Substituted product China reed transport pallet,
Glass fibre pallet
100 000 km
China reed transport pallet,
Glass fibre pallet
5000 km

Potential for energy savings
(GJ/ha)
2500
1200

Starch-based polymer films LDPE films

300 - 750

China reed for packaging chips Polystyrene chips

600 - 700

China reed for heat production Fuel oil

200 - 240

Rapeseed biofuel

Diesel

40 - 60

Rapeseed biofuel and heat
Diesel and fuel oil production from straw

32

110 - 120

Source
Corbière-Nicollier et al. (2001)
Corbière-Nicollier et al. (2001)
Dinkel et al. (1996)
Wolfensberger and
Dinkel (1997)
Wolfensberger and
Dinkel (1997),
Hartmann and
Strehler (1995)
Hartmann and
Strehler (1995)
Wolfensberger and
Dinkel (1997)

5. Summarising comparison
An overview of the life cycle inventory results for pellets and end products is given in Table 17a and
17b. Only those impact categories are listed for which data were available from several sources. All materials listed in Table 17a are commercially available and they are produced according to the current state-of-the-art. In contrast, future options, e.g. PHA production in plants and the medium and long-term
PLA production processes (Tables 8, 9 and 10), are not listed. In Table 17b, inventory data are presented for end products. All products listed in this table except for the printed wiring boards, the underfloor panel for the Mercedes A Class and the transport pallets are currently on the market. In the case of printed wiring boards, the lack of low-priced, pure lignin represents an obstacle to industrial implementation while, in the case of the underfloor panel the management decision about introduction in series production was pending when this chapter was written. For the transportation pallets reinforced with china reed, technical problems still need to be resolved (material stiffness).
Table 17a provides data for polycaprolactone (PCL) and polyvinyl alcohol (PVOH) which are both used as co-polymers for starch plastics. Life-cycle practitioners consider these data to be subject to major uncertainities. This is supported by the considerable range of values for energy use in the case of polyvinyl alcohol and for CO2 emissions for both PCL and PVOH.
In the case of starch polymer pellets energy requirements are mostly 25%-75% below those for polyethylene and greenhouse gas emissions are 20%-80% lower. These ranges originate from the comparison of different starch/copolymer blends, different waste treatment and different polyolefin materials used as reference. For the latter, Boustead's data for LLDPE (72.3 MJ/kg) and LDPE (80.6 MJ/kg) were assumed, which are lower than the value according to Dinkel et al. (91.7 MJ/kg, see Table 17a). The lower values according to Boustead serve also as reference for the comparison with the other bioplastics
(below).
Starch polymers score better than PE also for all the other indicators listed in Table 17a with eutrophication being the sole exception. The lower the share of petrochemical copolymers, the smaller the environmental impact of starch polymers generally is. However, the application areas for pure starch polymers and blends with small amounts of copolymers are limited due to inferior material properties.
Hence, blending can extend the applicability of starch polymers and thus lower the overall environmental impact at the macroeconomic level.
The cradle-to-factory gate energy requirements for PLA are 20%-30% below those for polyethylene, while GHG emissions are about 15%-25% lower. The results for PHA vary greatly (only energy data are available). Cradle-to-factory gate energy requirements in the best case (66.1 GJ/t) are 10%-20% lower than those for polyethylene. For more energy intensive production processes PHA does not compare well with petrochemical polymers. As pointed out earlier, especially the process energy requirements are higher for PHA than for conventional polymers (see Section 4.2).
Since all data in Table 17a refer to the current state-of-the-art, technological progress, improved process integration and various other possibilities for optimisation are likely to result in more favourable results for bio-based polymers in the future.
The results for starch polymer loose fills (Table 17b) differ decisively depending on the source.
Much of these differences can be explained by different assumptions regarding the bulk density of the
33

loose fills (see second column in Table 17b) and different approaches for the quantification of the ozone depletion potential (inclusion versus exclusion of NOx). It therefore seems more useful to compare the results of each study separately. One can conclude from both the study by Estermann (2000) and
Würdinger et al. (2001) that starch polymer loose fills generally score better than their equivalents made of virgin EPS. GHG emissions represent an exception where the release of CH4 emissions from biodegradable compounds in landfills results in a disadvantage for starch polymers (only according to
Würdinger et al., 2001). Loose fills produced from recycled polystyrene may represent an attractive option in environmental terms compared to starch polymers (Würdinger et al, 2001).
By analogy to loose fills, the range of results for starch polymer films and bags is to a large extent understandable from the differences in film thickness. Taking this factor into account, the environmental impacts of the starch films/bags are lower with regard to energy, GHG emissions and ozone precursors. The situation is less clear for acidification. For eutrophication, PE films tend to score better.
For Printed Wiring Boards (PWBs), about 30%-40% energy can be saved and the GHG mitigation potential is estimated to lie in a similar range. Exceptionally high savings (around 90% for energy and GHG emissions) have been established for epoxidised linseed oil as thickener for lacquers. Similar savings (above 80% for energy) are only reached by substituting flax fibre mats for fibreglass mats (Table 17b). For complete components (end products), the use of natural fibres is reported to save between
14% (under-floor panel) and 45%-50% energy (interior side panel and transport pallet).
Leaving PHA aside as the only exception, it can be summarised that bio-based polymers and natural fibres typically enable savings of around 20% (energy and CO2). Substantially higher savings up to 50% and beyond are considered feasible for certain starch polymers, printed wiring boards, certain lacquers and natural fibre composites.

34

Table 17a:

Summary of LCA key indicators for plastic pellets (only commercialised products manufactured by state-of-the-art technologies are listed)

35

Table 17b: see text)

Summary of LCA key indicators for end products (some of the products listed are commercialised, others not;

36

Apart from assessing bio-based materials in terms of the relative decrease of environmental impacts (in %, as just discussed), the comparison of savings per kg of bio-based polymer as shown in Table 18 can provide additional insight. These results show that printed wiring boards offer relatively low saving potentials and that GHG emissions for fibre composites can be disadvantageous.16 Otherwise, Table 17 confirms the finding that very attractive potentials for energy saving and GHG emission reduction exist for bio-based plastics (pellets), nonplastics (lacquer) and fibre composites.

Table 18:

Energy and GHG savings by bio-based polymers relative to their petrochemical counterparts GHG savings,
Energy savings,
MJ/kg bio-based kg CO2 eq./kg biopolymer*) based polymer*)
Bio-based plastics (pellets)
TPS
51
3.7
TPS + 15% PVOH
52
3.1
TPS + 52.5% PCL
28
1.4
TPS + 60% PCL
24
1.2
Mater-Bi foam grade
42
3.6
Mater-Bi film grade
23
3.6
PLA
19
1.0
Printed wiring boards
5
n/a
Lacquer
195
8.3
Flax fibre mat
45
n/a
Interior side panel for pass. car
28
-0.9
Transport pallet
33
1.6
*) Max. +/- 15% depending on whether LDPE or LLDPE according to
Boustead (1999/2000) is chosen as reference (see Table 17a)

6. Discussion
The comparison of the main assumptions made in the various studies and the comparison with the current state of the art reveals a number of uncertainties and caveats which are discussed in this chapter:
•
LCA data for polycaprolactone (PCL) and polyvinyl alcohol (PVOH) are generally considered to be subject to major uncertainties. In view of the widespread use of these compounds in biodegradable materials and given the strong impact on the final results especially for some starch polymers, reliable LCA data need to be generated.
•
A literature survey revealed that the process energy requirements for propylene oxide differ substantially depending on the source. The value used by Diehlmann and Kreisel
(2000b) is in the upper range. Since the production of propylene oxide contributes more than
55% to the total energy requirements for producing the petrochemical thickener (Diehlmann and Kreisel, 2000b) the energy input for the petrochemical thickener as a whole might be
16 Note that this statement - derived from the GHG data for the interior side panel of a passenger car in Table 17b - refers to

the use of this material for other purposes where substitution may occur on a 1:1 mass basis. As apparent from Table 14 the lower weight of the bio-based panel overcompensates this disadvantage.

- 37 -

overestimated. This might explain to a large extent why the saving potential related to the replacement of petrochemical by bio-based lacquer is exceptionally high compared to the other products (Table 18). In-depth analysis of the key assumptions would be required to understand the reasons and, if required, to reduce the concomitant uncertainties.
•
The data used for composting are subject to major uncertainties. This is partly explicitly stated by the authors (Estermann et al., 2000, p.16), partly it becomes obvious by comparing the assumptions made in the various studies (wherever these are described in detail). According to Esterman et al. (1998) 40-60% of the carbon absorbed in the vegetable material is released to the atmosphere during composting. To avoid the underestimation of GHG emissions, Esterman and colleagues (1998, 2000) assume in their studies that 60% of the absorbed carbon is released. The assumption can be considered as safe if compared to Schleiss and
Chardonnens (1994) according to whom the average carbon dissipation in the form of CO2 amounts to 40% (average of all composting plants in Switzerland). While these data refer to the average of all inputs and outputs of a composting plant, the question arises whether it also holds true for biodegradable starch polymers: Since the quality of biodegradable polymers is that they decompose to a large extent within a short period of time the question arises whether the approach chosen by Würdinger et al. (2001) might be more accurate where it was assumed that the buildup of organic matter and hence, the effect of carbon sequestration is negligible.
According to biodegradation tests conducted by several institutes, the degradation of starch polymers during composting (59ºC, 45 days) amounts to about 80% to 90% (test refers to mixture of 15% starch polymers and 85% pure cellulose). Since biodegradation in the subsequent maturation phase is negligible Novamont draws the conclusion that an average conversion rate of 80% is realistic (personal communication, L. Marini, 2001). The specific characteristics of the starch polymer considered and the type of composting technology applied may influence the biodegradation fraction.
•
The various studies differ in the accounting method for waste incineration of bio-based polymers. Even though the detailed assumptions are hardly ever spelled out it is quite obvious that the chosen approaches are not comparable. For example, Würdinger et al. (2001) assume that incineration takes place in waste-to-energy facilities, resulting in a net output of electricity and/or heat. Credits are assigned to these useful products. In contrast, Esterman and colleagues (1998, 2000) do not account for co-produced electricity/steam. It is unlikely that this reflects the differences in the share of energy recovery (waste-to-energy facilities versus simple incineration without energy recovery) among the countries studied; it rather represents different choices of system boundaries.
•
The environmental assessment of the incineration of mulch films with adhering organic waste (soil) raises particular questions. In one of the sensitivity analyses, Estermann and
Schwarzwälder (1998) introduce a CO2 penalty in order to account for the emissions resulting from the incineration of adhering organic waste. This may be justified if the moisture of the organic waste is so high that the vaporisation of the water contained requires more energy than the calorific value of the organic waste. In this case the incineration of the adhering waste represents a net energy sink. In practice, this is typically compensated by co-firing of fossil fuels or of other high-calorific combustible waste leading to CO2 and other environmental impacts. On the other hand it is also possible that the moisture content of the adhering

- 38 -

waste is low, resulting in a net energy yield in the incineration process. Moreover, if the organic waste is of biogenic origin, its incineration is neutral in CO2 terms (due to extraction of
CO2 from the atmosphere during plant growth). These considerations show that the specific circumstances determine whether the co-combustion of adhering organic waste – be it soil, organic kitchen waste or any other type of biogeneous waste – results in net environmental benefits or disadvantages.
•
Bio-based polymers generally have lower heating values than most petrochemical bulk polymers (Table 19). In some cases the difference is negligible (e.g., Polyhydroxybutyrate versus PET), while in other cases it is substantial (starch polymers versus PE). In practice, the difference in recoverable heat may be even larger than indicated by Table 19 due to the feature of most bio-based polymers to absorb water rather easily. The choice of the waste management system may therefore have a considerable impact on the overall conclusions. Regarding energy use, cradle-to-factory gate analyses, landfilling and waste incineration without energy recovery are in favour of bio-based polymers; on the other hand, incineration in wasteto-energy facilities, especially with high energy recovery yields, is in favour of petrochemical polymers (in energy terms). This calls for studying the material options by types of waste management technologies individually. Moreover, the actual situation in the country and region studied should be analysed. It should be taken into account here that energy recovery yields from waste-to-energy facilities are generally low at present: It is estimated that, one quarter of the heating value of the waste is converted to final energy in the form of power and useable heat.17 The generation of the same amount of final energy from regular fuels in power plants and district heating plants requires only half of the energy input. As a consequence, the credit for energy recovery is only half of the heating value. The advantage of petrochemical over bio-based polymers is therefore only half of the difference of their heating values. Depending on the petrochemical and the bio-based polymer studied this difference can still be substantial but it may also be negligible.
•
In the case of landfilling some studies account for methane (CH4) emissions due to anaerobic emissions while others do not take this into consideration. This can have a considerable impact on the results due to the relatively strong greenhouse gas effect of CH4. As a consequence the overall GHG emissions from biodegradable polymers manufactured from renewable raw materials may be higher than for petrochemical plastics depending on the waste management system chosen for the latter (Würdinger et al., 2001). Moreover, N2O emissions can have a sizable influence (compare e.g. Table 14).
•
Environmental comparisons including recycling as a waste management option are rarely made. Moreover, most of the bio-based polymers except for starch can be processed by mechanical or even feedstock recycling (back to monomer). Mechanical recycling, is in principle even possible for thermoplastic polymers reinforced with natural fibres (Hauspurg et al.,
2000). More attention must be paid to these options in future studies.

17 This estimate is based on an analysis for Germany (12% efficiency for both electricity and heat generation from

combustible waste; Patel, 1999) and for Western Europe (personal communication, Pezetta, 2001).

- 39 -

•
The characterisation factors for global warming used in most of the studies reviewed are outdated in the meantime (Global Warming Potentials for methane and nitrous oxide).18
Since the contribution of CO2 dominates the overall GHG effect, this uncertainty is considered to be less important.
•
When making comparisons with conventional fossil fuel-based polymers it must be borne in mind that LCA data for these products are also uncertain and continue to be corrected. This is in spite of the fact that petrochemical polymers are manufactured by use of mature technologies that are applied globally with only limited variations. For example, energy data for polyethylene (PE) production range between ca. 65 GJ/t and 85 GJ/t according to a comparison of various sources (Patel, 1999) while Dinkel et al. (1996) assume about
92 GJ/t. If, for example, compared to the TPS data determined by Dinkel et al. (1996), the wide range of the values for PE does not change anything about the conclusion that TPS is most beneficial in terms of energy use and CO2 emissions. However, it is unclear whether the final conclusions for the other environmental parameters covered by Dinkel et al. (e.g., emissions) are also insensitive to larger variations in the PE data.

Table 19: Heating value of bio-based and petrochemical polymers (heating values calculated according to Boie, compare Reimann and Hämmerli, 1995)
Polymer
Starch polymers

Lower heating value, GJ/tonne
13.6

Polyhydroxybutyrate (P3HB)

22.0

Polyhydroxyvalerate (P3HV)

25.0

Polylactic acid

17.9

Lignin (picea abies)

24.2

China reed

18.0

Flax

16.3

Hemp
Kenaf

17.4

PE

43.3

PS

39.4

PET

22.1

PVC

17.9

16.5

The problem related to these uncertainties can be resolved to some extent by taking into account the significance of the difference in values for the systems compared (thresholds for the categories "significantly higher", "higher", "comparable" etc., see Estermann et al., 2000, Section 4.1.2). In addition, it is an important goal of future research to reduce further the existing uncertainties. In all of the studies reviewed ecological ranking was determined by comparing for how many indicators the environmental impact is lower for bio-based polymers compared to

18 The Global Warming Potentials (GWP) used in the various studies are 11 or 21 for CH and 270 or 310 for N O, while 4
2

according to the current state of research - more accurate figures are 23 (CH4) and 296 (N2O) (Houghton et al., 2001). All
GWP values refer to a 100 year period.

- 40 -

the petrochemical polymers.19 The disadvantage of this approach is that the selection of the indicators compared can have an influence on the final conclusions. Together with the fact that the relative difference in the results for the various impact categories (a few per cent versus a a few hundred per cent) is hardly ever accounted for this shows the urgent need for the further development of the LCA methodology (e.g., by introduction of significance thresholds). When interpreting the results, it must finally be taken into account that the studies reviewed partly differ in regional scope. Since the results are to some extent subject to country specific circumstances (e.g., GHG emissions from national power production) care must be taken when drawing more general conclusions. On the other hand, the uncertainties related to conclusions can be reduced if several independent analyses for different countries arrive at similar conclusions. A summary of the aspects to be considered in future LCA studies for biobased polymers is given in the checklist in Annex 2 (with a special focus on biodegradable polymers). 7. Conclusions
7.1

Summary and further elaboration of findings

The number of published LCAs for bioplastics is very limited. This seems to be in contrast to the general public interest for this issue and the more recent interest by policy makers. For example, no comprehensive LCAs have been published so far for PLA (plant-based), cellulose polymers (plant-based) and petrochemical copolymers such as BASF's product Ecoflex.
The existing LCAs contain uncertainties which should be addressed by future research and analysis. A prominent example is the environmental assessment of the composting process for biodegradable polymers. In some studies further sensitivity analyses would be required to ensure that the final findings are well underpinned (e.g. for smaller PE bags in Estermann and Schwarzwälder, 1998). Moreover, many of the environmental analyses choose a cradleto-factory gate perspective (i.e., the analysis ends with the product under consideration).
While this approach provides valuable results, additional analyses taking a cradle-to-grave perspective by inclusion of the waste management stage should also be conducted. Due to their strong impact on the final results all major waste management options should be studied
(landfilling, composting, MSWI plants, waste-to-energy facilities, digestion and recycling).
To assist life cycle practitioners in making use of the lessons learnt from this review, a checklist has been prepared which will be published by RAPRA (Patel, 2003). In addition to this checklist, any LCA study must comply with the requirements specified in the ISO standards 14 040 to 14 043 (ISO, 1997-1999).
Apart from some methodological shortcomings three LCAs evaluate products for which considerable technical problems related to production and product properties still need
19 In addition, single-score parameters were used; as indicated earlier these results are not described here since the method

applied is not generally accepted.

- 41 -

to be overcome. This is the case for the two china reed-based products studied by Wolfensberger and Dinkel (1997) and for the transport pallet studied by Corbière-Nicollier et al.
(2001). This calls firstly, for further R&D and secondly, for caution when making use of the
LCA results for these specific applications.
In spite of these restrictions and the uncertainties and the information gaps mentioned above the body of work analysed overwhelmingly indicates that bio-based polymers offer important environmental benefits today and for the future. This applies to all three goups studied, i.e. to bio-based plastics, bio-based non-plastics (with one type of lacquers as the only representative) and composites based on natural fibres.
Of all bio-based plastics (pellets) studied, starch polymers are considered to perform best in environmental terms under the current state of the art – with some differences among the various types of starch polymers. Compared to starch polymers the environmental benefits seem to be smaller for PLA (LCA results only available for energy and CO2). For PHA, the environmental advantage currently seems to be very small compared to conventional polymers (LCA results only available for energy use). For both PLA and PHA, the production method, the scale of production and the type of waste management treatment can influence decisively the ultimate conclusion about the overall environmental balance.
The only available analysis for non-plastics (lacquer thickener based on linseed oil) revealed an exceptionally high saving potential, which calls for further analysis.
For natural fibres, the extent to which these can replace fibreglass (which is heavy and energy intensive to produce) determines mainly the net environmental benefits. The advantages according to cradle-to-factory gate analyses were rather limited in one case (-14% for under-floor panel) and very attractive in the two other cases (-45% to –50% for interior side panel and transport pallet).
The case studies for reinforced products demonstrate that savings in the use phase – sometimes also referred to as "secondary savings" – are as high or even clearly higher than the "primary savings" (typically related to cradle-to-factory gate systems) and that they may even be the main driver for substitution. In the case of the china reed pallet and the hemp fibre-based interior side panel secondary energy savings are up to three times as high. Exceptionally high secondary savings are reported to be obtainable with tyres containing starchbased fillers where secondary energy savings exceed primary savings by a factor of more than
20. The conclusion of the ECCP Working Group on Renewable Raw materials (European
Commission, 2001), according to which total secondary savings may exceed the primary savings by about one order of magnitude, is hence clearly confirmed by the case study for tyres while it seems somewhat too optimistic if compared to the results for natural fibres.
Starch polymers are currently the only type of bio-based polymer for which several comprehensive LCA studies are available. According to these assessments starch polymers do not perform better than their fossil fuel-based counterparts in all environmental categories, including biodiversity and soil quality, which are generally outside the scope of LCAs. However, most studies come to the conclusion that starch polymers (pellets and end products) are more beneficial in environmental terms than their petrochemical counterparts; this conclusion is drawn without weighting and in most cases without significance thresholds. The preferences among the environmental targets determine whether bio-based polymers are considered

- 42 -

to be environmentally attractive. Full-sized LCA studies for further bio-based materials are indispensible to allow deriving conclusions and recommendations that are better underpinned and more focussed.
For the time being, it is not possible to make a concluding general judgement whether bio-based plastics should be preferred to petrochemical polymers from an environmental point of view. This has partly to do with the limited availability of comprehensive LCAs. But even if more LCA studies were available one would be left with considerable uncertainties, e.g. because it will never be feasible to cover all possible products and all possible impact categories (compare Finnveden, 2000 and Box 1 in Annex). In spite of these limitations one can conclude that the results for the use of fossil energy resources and GHG emissions are more favourable for most bio-based polymers already today. As an exception, landfilling of biodegradable polymers can result in methane emissions (unless landfill gas is captured) which may make the system unattractive in terms of reducing greenhouse gas emissions. As a potential source of N2O emissions, fertilizers also require special attention.
By comparing the use of biomass for the manufacture of materials (polymers and fibres) on the one hand and for energy purposes (bioenergy) on the other insight can be gained about the most effective options for land use and cultivation. Important findings of Dinkel et al. (1996) and the LCA prepared by Corbière-Nicollier et al. (2001) are hence that materials based on starch, kenaf and china reed offer larger opportunities for energy saving and GHG mitigation than bioenergy (Dinkel et al., 1996, p.12 and p.92; partly based on Wolfensberger and Dinkel, 1997). In contrast, Kurdikar et al. (2001) argue that bioenergy contributes more to
GHG emission reduction than biomass-derived feedstocks. The main reason for this contrasting finding seems to be that the product and process Kurdikar et al. (2001) studied – i.e., the production of polyhydroxyalkanoates in plants – currently cannot compete with conventional products in energy terms. Wherever the opposite applies – and this is the case for most of the other products analysed – the available results indicate that biomaterials offer higher environmental gains than bioenergy. This issue will systematically have to be studied for biobased polymers by comparing the benefits per km2 of cultivated land. In other words, comparative assessments will continue to be needed in order to keep track of the aspects of competition and complementarity between bioenergy and biomaterials. This is also necessary in order to account for innovations in both areas. It would ease such comparisons and the usefulness for decision-makers if future studies dealing with bioenergy and biomaterials always also studied the land use requirements of the various options.
7.2

Outlook and perspectives

To maximise the environmental benefits from bio-based polymers further R&D will be necessary in order to optimise the production by increasing the efficiencies of the various unit processes involved (e.g. separation processes) and by process integration. Substantial scope for improvement can be expected here considering economies of scale and given the fact that all bio-based polymers are still in their infancy while the manufacture of petrochemical polymers has been optimized for decades. Some of the LCA discussed above were already outdated when these conclusions were drawn since substantial progress had been made in

- 43 -

manufacturing and processing bio-based polymers (e.g., for films). This means that the real environmental impacts caused by bio-based polymers tend to be lower than established in the
LCA studies reviewed.
As a guide for future R&D, good practice targets for "environmentally correct" biobased products could be very useful. Based on the results presented in the preceding sections, a first attempt is made here to specify such targets: It is recommended that, relative to their petrochemical counterparts, bio-based polymers should
• save at least 20 MJ (non-renewable) energy per kg polymer,
• avoid at least 1 kg CO2 per kg polymer and
• reduce most other environmental impacts by at least 20%.
A good practice target will also have to be specified for land use, i.e. in terms of GJ energy saved per ha land cultivated. In parallel to these environmental targets, cost reduction must continue to be a priority.
A promising line for R&D in the longer term could be the development of biomassderived polymers that can be recycled mechanically and/or back to feedstocks/monomers.
Preferably this should be possible also in combination with petrochemical polymers. Such recyclable polymers made from renewable raw materials have good chances to be unrivalled in environmental terms provided that their manufacture is not too resource-intensive in the first place. This may offer longer term prospects to PHA, PLA and some other bio-based polymers. To summarize, the available LCA studies and environmental assessments strongly support the further development of bio-based polymers. Careful monitoring of the various environmental impacts continues to be necessary both for decision makers in companies and in policy. If combined with good-practice targets, this may accelerate and focus the ongoing product and process innovation. For some materials the environmental benefits achieved are substantial already today. In many other cases the potentials are very promising and need to be exploited.

- 44 -

8. Acknowledgements
The preparation of this chapter was only possible due to the large support from various experts in the field, among them Mrs. Tourane Corbière-Nicollier (Swiss Fed. Inst. of Technology), Mr. Joachim Diener (Daimler-Chrysler, Germany), Mr. Achim Diehlmann (ITUC, Jena
University), Mr. Fredy Dinkel (CARBOTECH, Switzerland), Mr. Thomas Fleißner (Technical University of Munich), Mr. Urs Hänggi (Biomer, Germany), Mr. Olivier Jolliet (Swiss
Fed. Inst. of Technology), Mr. Urs Lübker (Dreisol Coatings), Mr. Nitin Mutha (visiting fellow at Utrecht University, Netherlands), Mr. V. Premnath (National Chemical Laboratory,
India), Mrs. Bea Schwarzwälder (Composto, Switzerland) and Mr. Erwin Vink (Cargill Dow,
Netherlands).

9. References
Anonymous (1999) GM plants offer 'green' plastic production, Chemistry&Industry, p. 729.
Boustead, I. (1999, 2000) Eco-profiles of plastics and related intermediates (about 55 products). Prepared for the Association of Plastics Manufacturers in Europe (APME). Downloadable from the internet (http://www.apme.org), Brussels, Belgium.
Bozell, J.J.; Landucci, R. (1993) Alternative feedstocks program. Technical and economic assessment. Thermal/chemical and bioprocessing components. National Renewable Energy
Laboratory (NREL), Golden, Colorado, USA.
CML (Centre of Environmental Science Leiden, 2001) Life cycle assessment – An operational guide to the ISO standards. Report prepared by Guinée, J.B., Gorrée, M., Heijungs, R.,
Huppes, G., Kleijn, R., Koning, A. de, Oers, L. van, Wegener Sleeswijk, A., Suh, S., Udo de
Haes, A. (all CML), Bruin, H. de (TUD/TB), Duin, R. van (Bureau B&G), Huijbregts,
M.A.J. (UVA). Part 1-3, downloadable from http://www.leidenuniv.nl/cml/lca2/index.html.
Leiden, Netherlands.
Corbière-Nicollier, T., Gfeller Laban, B., Lundquist, L., Leterrier, Y. Manson, J.-A.E. and
Jolliet. O. (2001) Life cycle assessment of biofibres replacing glass fibres as reinforcement in plastics. Swiss Federal Institute of Technology, Lausanne, Switzerland. Resources, Conservation and Recycling, 33, 267-287.
Corvasce, F. (1999) Environment friendly tire concepts using a biopolymeric filler derived from starch. Goodyear Tires. Bioplastic Conference, 24.6.1999.Diehlmann, A., Kreisel, G.
(2000a) Ökologische Bilanz von UV-härtenden Lacken auf Leinölbasis. ITUC (Institute for
Technical Chemistry and Environmental Chemistry), Jena University, Germany.
Farbe&Lack 106, 12/2000, 101-105.
Diehlmann, A., Kreisel, G. (2000b) Ökologische Bilanzierung ausgewählter Lackrohstoffe:
Vergleich von Bindemitteln auf nativer und petrochemischer Basis. ITUC (Institute for
Technical Chemistry and Environmental Chemistry), Jena University, Germany, pp.95.
Diener, J., Siehler, U. (1999) Ökologischer Vergleich von NMT- und GMT-Bauteilen.
Angew. Makromol. Chem. 272 (Nr. 4744), 1-4.

- 45 -

Dinkel, F., Pohl, C., Ros, M., Waldeck, B. (1996) Ökobilanz stärkehaltiger Kunststoffe (Nr.
271), 2 volumes. Study prepared by CARBOTECH, Basel, for the Bundesamt für Umwelt und Landschaft (BUWAL), Bern, Switzerland.
Estermann, R. (1998) Test von Säcken aus biologisch abbaubaren Werkstoffen für die
Grünabfallsammlung. Study prepared by COMPOSTO for the Kompostforum Schweiz. Olten, Uerikon, Switzerland.
Estermann R., Schwarzwälder, B. (1998) Life cycle assessment of Mater-Bi bags for the collection of compostable waste. Study prepared by COMPOSTO , for Novamont, Novara, Italy. Olten, Uerikon, Switzerland.
Estermann, R., Schwarzwälder, B., Gysin, B. (2000) Life cycle assessment of Mater-Bi and
EPS loose fills. Study prepared by COMPOSTO for Novamont, Novara, Italy. Olten, Switzerland.
European Commission (2001) ECCP (European Climate Change Programme) –Long report. http://europa.eu.int/comm/environment/climat/eccp_longreport_0106.pdf. Brussels, Belgium.
Finnveden, G. (2000) On the limitations of life cycle assessment and environmental system analysis tools in general. International Journal of Life Cycle Assessment 5 (4), 229-238.
Flake, M., Hansen, A., Fischhaber, G., Schäper, S. (2000a) Assessment of natural fibre components for automotive parts - Ecological and economical decision making. Conference paper prepared by LCE Consulting GmbH, Braunschweig and Audi AG, Ingolstadt. In: ACTIN (Alternative Crops Technology Interaction Network): Proceedings of the conference
"Natural Fibres for the Automotive Industry". The Manchester Conference Centre, Weston
Building, University of Manchester Institute of Science and Technology (UMIST), United
Kingdom, 28 November 2000.
Flake, M., Fleissner, T., Hansen A. (2000b) Ökologische Bewertung des Einsatzes nachwachsender Rohstoffe für Verkleidungskomponenten im Automobilbau. ISSN 0170-7299.
Braunschweig, Germany.
Gerngross, T. U. (1999) Can biotechnology move us toward a sustainable society? Nature
Biotechnology 17 (June 1999), 541-544.
Gerngross, T. U., Slater, S. (2000) How Green are Green Plastics? Scientific American, August, 2000, 37-41.
Goedkoop, M. (1995) Eco-indicator 95 - Weighting method for environmental effects that damage ecosystems or human health on a European scale. Final report prepared by Pré
Consultants and DUIF Consultancy bv, commissioned by NOVEM and RIVM, Netherlands.
Habersatter, K. (1991) Ökobilanzen von Packstoffen. Bundesamt für Umwelt, Wald und
Landschaft (BUWAL), Schriftenreihe Umweltschutz, Nr.132. Bern, Switzerland.
Hakala, S., Virtanen, Y., Meinander, K., Tanner, T. (1997) Life-cycle assessment, comparison of biopolymer and traditional diaper systems. Technical Research Centre of Finland (VTT),
Finland, pp. 91
Hauspurg, Ch., Reußmann, T., Mieck, K.-P. (2000) Recycling von Naturfaserverbundwerkstoffen im Plastifizier-Pressverfahren. GAK 8/2000, Jahrgang 53, 523-527.
Heijungs, R., Guinee, J. B., Huppes, G., Lankreijer, R. M., Haes, U. de, Wegener Sleeswijk,
A. (1992) Environmental Life-Cycle Assessment of Products – Guide and Background.
CML (Centre for Environmental Science Leiden), Leiden, Netherlands; Netherlands Organisation for Applied Scientific Research (TNO); Fuels and Raw Materials Bureau (B&G).

- 46 -

Heyde, M. (1998) Ecological considerations on the use and production of biosynthetic and synthetic biodegradable polymers. Polym. Degrad. Stab. 59 (1-3), 3-6.
Houghton, J. T., Ding, Y., Griggs, D. J., Noguer, M., van der Linden, P. J., Da, X., Maskell,
K., Johnson, C. A. (editors, 2001) Climate Change 2001 – The scientific basis. Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change (IPCC). Full text downloadable from www.ipcc.ch (November 2001), 2001.
ISO (International Organisation for Standardisation, 1997-1999) 
- ISO 14 040 (DIN EN ISO 14040) Environmental Management – Life Cycle Assessment –
Principles and Framework. Berlin: DIN (Deutsches Institut für Normung), 1997.
- EN ISO 14 041 (1998) Environmental Management – Life Cycle Assessment – Goal and scope definition and inventory analysis. Berlin: DIN (Deutsches Institut für Normung).
- EN ISO 14 042 (1998) Environmental Management – Life Cycle Assessment – Life Cycle
Impact Assessment. Draft. Berlin: DIN (Deutsches Institut für Normung).
- EN ISO 14 043 (1999) Environmental Management – Life Cycle Assessment – Life Cycle
Interpretation. Draft. Berlin: DIN (Deutsches Institut für Normung).
Kopf, N. (1999) Kunststoffe aus nachwachsenden Rohstoffen – Polyhydroxybutyrat und
Stärkekunststoffe – Abschätzungen zum kumulierten Energieaufwand und zu CO2Emissionen. Student's report, prepared at the Fraunhofer Institute for Systems and Innovation Research (FhG-ISI), Karlsruhe for Engler-Bunte-Institut at Karlsruhe University.
Karlsruhe, Germany.
Kosbar, L. L., Gelorme, J. D., Japp, R. M., Fotorny, W. T. (2001) Introducing biobased materials into the electronics industry. Journal of Industrial Ecology, Volume 4, No. 3, 93-105.
Krämer, M.: Nylon-9,9 und Nylon-6,6 – Ein Vergleich der Synthese von Kunststoffen aus nachwachsenden Rohstoffen und Petrochemikalien, Kosten, Energie und CO2-Bilanzen.
Diplomarbeit (M.Sc. thesis), Chemiehochschule Fresenius, Wiesbaden, 1992
Kurdikar, D., Paster, M., Gruys, K. J., Fournet, L., Gerngross, T. U., Slater, S. C., Coulon, R.
(2001) Greengouse gas profile of a plastic derived from a genetically modified plant. Journal of Industrial Ecology. Volume 4, No. 3, 107-122.
Luck, T. (1996) Feasability-Studie Polyhydroxyfettsäuren – Kurzdarstellung der Ergebnisse der Studien 'Feasibility-Studie zur Abschätzung des Marktpotentials neuer Polyhydroxyfettsäuren (PHF)' und 'Abschätzende Ökobilanzen zu Polymerwerkstoffen auf der Basis biologisch erzeugter Polyhydroxyfettsäuren', Fraunhofer Institut für Lebensmitteltechnologie und
Verpackung (ILV), Freising, Germany.
Marini, L. (2001) Personal communication with Mr. L. Marini, Novamont, Novara, Italy,
2001.
N.N. (2001): Native Polymere. Chapter in Fraunhofer IAP Annual Report 2001 (Kapitel 3).
Fraunhofer Institute for Applied Polymer Research (IAP). Downloadable from http://www.iap.fhg.de/. Potsdam, Germany, 2001
Nossin, P., Joosten, J., Bruggink, A. (2002): Future feedstocks for commodity polymers:
EcotheneTM – Sustainability in the 21st century. Presentation at Utrecht University, Utrecht,
May 2002
Patel, M. (1999) Closing Carbon Cycles: Carbon use for materials in the context of resource efficiency and climate change.
PhD
thesis
(ISBN
90-73958-51-2, http://www.library.uu.nl/digiarchief/dip/diss/1894529/inhoud.htm), Utrecht University,
Utrecht, Netherlands.

- 47 -

Patel, M. (2003) Environmental life cycle comparisons of biodegradable plastics. In: C. Bastioli (ed.): Handbook on Biodegradable Materials. RAPRA Technology Ltd., forthcoming in 2003.
Patel, M., Jochem, E., Marscheider-Weidemann, F., Radgen, P., von Thienen, N. (1999) CSTREAMS - Estimation of material, energy and CO2 flows for model systems in the context of non-energy use, from a life cycle perspective (Volume I) (in German; English abstract).
Report by Fraunhofer ISI Karlsruhe, Germany.
Pezetta, O.: Personal communication with Mr. O. Pezetta, TN-Sofres, Paris, France, 2001.
Reimann, D. O., Hämmerli, H. (1995) Verbrennungstechnik für Abfälle in Theorie und
Praxis. Schriftenreihe Umweltschutz, Bamberg, pp.246
Schleiss, K., Chardonnens, M. (1994) Stand und Entwicklung der Kompostierung in der
Schweiz 1993. BUWAL Umwelt-Materialien Nr. 21. Bundesamt für Umwelt, Wald und
Landschaft (BUWAL), Bern, Switzerland.
Schmidt, B., Langer, E. (2001) Biomass for industry: German strategies for the 21st century.
Abstracts of the 5th International Biomass Conference of the Americas. Downloadable from http://bioproducts-bioenergy.gov/pdfs/bcota/default.html. Orlando, Florida, 2001
Subramanian, P. M. (2000) Plastics recycling and waste management in the US. Resources,
Conservation and Recycling 28, 253-263.
Vink, E. (2001) NatureWorks – A new generation of biopolymers. Presentation by E. Vink,
Cargill Dow on 29 March 2001, Birmingham, United Kingdom.
Vink, E. (2002) Personal communication with E. Vink, Cargill Dow, Netherlands, March
2002
Wötzel, K., Wirth, R., Flake, R. (1999) Life cycle studies on hemp fibre reinforced components and ABS for automotive parts. Braunschweig University and Seeber. Angew. Makromol. Chem. 272 (4763), 121-127.
Wolfensberger, U., Dinkel, F. (co-ordinators, 1997) Beurteilung nachwachsender Rohstoffe in der Schweiz in den Jahren 1993-1996 – Vergleichende Betrachtung von Produkten aus ausgewählten nachwachsenden Rohstoffen und entsprechenden konventionellen Produkten bezüglich Umweltwirkungen und Wirtschaftlichkeit. Prepared by FAT (Eidgenössische Forschungsanstalt für Agrarwirtschaft und Landtechnik) and CARBOTECH for the Federal Office of Agriculture, Bern, Switzerland.
Würdinger, E.; Wegener, A.; Roth, U.; Peche, R.; Reinhardt, G.A.; Detzel, A.; Giegrich, J.;
Möhler, S.; Borken, J.; Patyk, A.; Fehrenbach, H.; Vogt, R.; Mühlberger, D.; Wante, J.;
Klein, M.; Straver, F. C. M. (2001) Kunststoffe aus nachwachsenden Rohstoffen - Vergleichende Ökobilanz für Loose-fill-Packmittel aus Stärke bzw. aus Polystyrol (interim report).
Bayrisches Institut für angewandte Umweltforschung und –technik, Augsburg (BIFA; project leader), Institut für Energie- und Umweltforschung Heidelberg (IFEU), Flo-Pak GmbH,
Germany (part of the information presented in this chapter originates from the interim report published in April 2000; wherever required, more recent calculation results were made available by E. Würdinger, BIFA).

- 48 -

Annex 1:

Table A1-1:

Overview of environmental life cycle comparisons for biodegradable polymers included in this review

LCA studies for starch polymer pellets and films

- 49 -

Table A1-2:

LCA studies for starch polymer loose fill packaging material

- 50 -

Table A1-3:

LCA studies for compost bags

- 51 -

Table A1-4: LCA studies for polyhydroxyalkanoates (PHA), polylactides (PLA) and other biodegradable polymers

- 52 -

Annex 2:

Checklist for the preparation of an LCA for biodegradable plastics

The following aspects should be taken into account when preparing a life cycle assessment for biodegradable plastics. All the methodological decisions, assumptions and key data should be specified in the text. In addition to the aspects listed in this checklist the life cycle assessment must comply with the requirements specified in the ISO standards 14 040 to 14 043 (ISO,
1997-1999).

1.

Biomass production

1.1
1.2
1.3

Country of origin: Where is the biomass used grown?
Type of cultivation: Is the biomass grown by intensive or extensive cultivation?
Fertilizers: Have the effects related to the production of fertilizers been taken into account? Carbon balance plant growth: Is carbon uptake during plant growth
a)
considered as a separate process which is therefore reflected in the LCA calculations as negative CO2 emissions or is it
b)
combined with the process of decomposition (after the use of the product) resulting in overall net zero emissions?
Note: Both concepts are possible and the aggregated results throughout the life cycle are identical; however, differences in approaches result in difficulties when comparing disaggregated results of studies (results for subsystems). It is therefore recommended to apply approach a) since this is the more differentiated approach by breaking down the entire activity into more subprocesses.

1.4

2.

Plastics production and use

2.1

Country: In which country is the biodegradable plastic (or the end product) manufactured? Power generation: Are the assumptions regarding the average efficiency of power generation and the specific emissions (e.g. in kg per MWhel) stated?
By-products: If any by-products are produced (materials or energy), how are these taken into account (by means of credits or by extension of the system)?
System boundaries: Does the LCA refers to the system "cradle-to-factory gate" or to the system "cradle-to-grave"?
Note: "Cradle-to-factory gate" refers to the entire production system from the extraction of the required resources to the production of the product under consideration. The system "Cradle-to-grave" includes, moreover, the waste management after the useful life of the product.

2.2
2.3
2.4

- 53 -

2.5

2.6

3.

3.1

3.2

3.3
3.4

3.5

3.6

Functional unit: Is the functional unit a certain amount of polymer (in mass terms or in terms of volume?), a semi-finished product (e.g. 1m2 of film) or an end product
(e.g. 100 plastic bags)? A further option is to choose the product service as the functional unit, e.g. 100m3 of packed goods in the case of packaging materials. Has it been taken into account that the amount of material required (in kg) for a given functional unit might differ for biodegradable polymers and their potential alternatives
(e.g., non-degradable polymers, paper)?
Use phase (for end products only): Is it clearly specified whether the use phase is or is not included in the system boundaries? If so, have the assumptions been specified?
Note: For example, the inclusion of the use phase for compost bags means that the functional unit is the collection of biodegradable household waste; in this case, comparisons are generally made with the collection in a compost bin without a bag; the compost bin is therefore cleaned after use; this implies the use of water and detergents which are included in the system boundaries while this would, for example, not be the case for the system "cradle-to-factory gate".

Plastics waste management
The following aspects are only relevant if the system "cradle-to-grave" has been chosen, otherwise they are irrelevant.
Waste management system: Which waste treatment process/es has/have been assumed? I.e., what are the shares of waste landfilled, recycled, incinerated without energy recovery, fed to waste-to-energy facilities, composted, digested and/or disposed of via sewage treatment?
Note: Sewage treatment is a practical option e.g. for loose-fill polymers.
Landfill emissions: Have emissions from biodegradation in landfills been taken into account (especially: methane which orginates from inaerobic processes in landfills)?
How high are the emissions?
Composting: Has any sequestration of carbon in the compost been assumed and if so, to which extent?
Waste-to-energy: In the case of waste-to-energy facilities, what are the yields of power and/or heat? Have these useful outputs been accounted for by credits (for inputrelated impact categories like energy resources and output-related impact categories such as greenhouse gas emissions)? And if so, which assumptions have been made when establishing these credits?
Recycling processes: In the case of recycling, have the types of technologies been specified? - Mechanical or feedstock recycling?
- Which type(s) of feedstock recycling?
Mechanical recycling: Which substitution factor has been assumed in the case of mechanical recycling?
Note: It may be necessary to blend the recycled plastics with virgin material in order to obtain the desired material properties; it may also be necessary to use more

- 54 -

3.7

recyclates than virgin polymers for the same functional unit. In both cases the substitution factors is less than 100%, i.e. each kilogramme recycled plastics substitutes less than one kilogramme virgin material.
Waste scenarios: Have separate scenarios been developed for waste management, e.g. in order to account for different practices depending on the country or to account for
(future) changes in waste policy?

4.

Transportation

4.1

Particularly light products: Has transportation been taken into account for products with a particularly high volume/mass ratio?
Note: Transportation energy generally does not play any major role, neither for the production of plastics (including biodegradable ones) nor for final products made thereof. Particularly light products, e.g. loose fill packaging material, are exempted from this general rule.
Assumptions: Have all the assumptions been made clear in these cases (transportation mode, transportation distances, load factors, fuel efficiencies)?

4.2

5.

Overall assessment

5.1

Impact categories and impact subcategories: The choice of the impact subcategories specified in addition to the impact categories can have a major impact on the final conclusions. Has the selection of the impact subcategories been throughly reflected and is the choice justified in the text?
Note: Background information is given in Box 1. In Table A2-1, a set of impact categories is listed that are considered to be particularly relevant for biodegradable polymers. Significance thresholds: In comparative LCAs, the uncertainty of the results and the importance of the differences in the results can be taken into account by distinguishing between significance thresholds. Has any approach of this type been been applied?
Note: Table A2-2 gives an example of how this can be done.
Characterisation factors: Have updated characterisation factors been used for aggregation within the impact categories? (Characterisation factors are sometimes also referred to as equivalence factors.)
Note: The characterization factors for climate change are referred to as GWP values
(Global Warming Potential). The most recent figures can be found in Houghton et al.
(2001).
Weighting: Have the aggregated scores of the various impact categories been weighted? 5.2

5.3

5.4

- 55 -

5.5

Note: If not - i.e. if the conclusions are drawn without an weighting procedure - even more attention must be paid to the selection of the impact categories (see above, No
5.1).
Non-biodegradable polymer: Has the non-biodegradable (conventional) polymer been named (in the case of comparative LCAs)? Are the approach and the assumptions consistent with those used for biodegradable polymers and are all these assumptions clearly described?

6.

Further aspects (consider only if relevant)

6.1

Alternative use of biomass: Has the alternative use of renewable raw materials for other material purposes or as an energy source been studied in order to put the results for biodegradable polymers into perspective?
Note: The background of this question is that it is already known that biodegradable polymers based on renewable raw materials generally score better than petrochemical polymers with regard to fossil energy use and greenhouse gas emissions while they score worse with regard to land use, ecotoxicity and eutrophication. Given this knowledge and considering the limited availability of biomass-derived raw materials it might be of interest to study the environmental impacts of other options of using the same renewable raw materials as used for the manufacture of biodegradable polymers. National level: Have the results been translated to the national level?
Note: For materials that are or that could be used in bulk quantities this can be relevant for strategy development in companies and governmental policy.

6.2

- 56 -

Box 1:

LCA results for biodegradable polymers - Consequences of the choice of impact categories and impact subcategories and limitations of LCAs

The choice of the impact categories can have an important impact on the final conclusions drawn from an LCA study. So far, only suggestions for the choice of impact categories to be included have been made (Udo de Haes, 1996, Lindfors et al, 1995) but no obligatory set has been decided upon by SETAC (Society of Environmental Toxicology and Chemistry). Neither does a minimum or a maximum list exist. The following criteria have been put forward for the choice of the impact categories (Udo de Haes, 1996):
• Completeness: The list should include all relevant environmental problems.
• Independence: The categories should be as independent of each other as possible in order to avoid double counting of indirect effects.
• Practicality: For practical reasons the list should be as concise as possible.
A distinction can be made between input related and output related impact categories.
Input related impact categories refer to resource depletion or competition while output related categories are metrics for emission and pollution.
Apart from the main impact categories, subcategories can be defined. This is usually done in those cases where the overall impact within one main category is caused by two or more factors that differ decisively (e.g. input of materials and energy as two subcategories of the main category "resources"). The definition of subcategories raises the question how to aggregate this information in the further course of an LCA. There are two options (Udo de Haes, 1996)
• Impact subcategories may be aggregated in the "characterisation step" (e.g., aggregation of the subcategories "energy" and "materials" to the main category "resources") if the distinction of subcategories was mainly made due to lack of information for the impact category as a whole.
• The impact subcategories should be aggregated as a part of the valuation step if the effects and the underlying mechanisms are so diverse that the aggregation is primarily determined by the value system (this means that subcategories are upgraded to categories; example: ecotoxicity versus human toxicity).
The following table provides an overview of firstly, the main impact categories proposed by Udo de Haes (1996) and secondly, a selection of subcategories (and other impact indicators) encountered in LCAs for biodegradable polymers. The crosses in the columns give a rough indication for which of the categories either polymers based on renewable resources or on petrochemical raw materials score better and in which cases the difference tends to be insignificant (column "Neutral"). The table points out once more the potential importance of the
…
selection of impact categories for the findings of an LCA. If, for example, the selection is such that it includes many impact categories that are relatively insignificant (column
"Neutral"), comparative LCAs may come to the conclusion that the differences between the options are relatively small. This type of misinterpretation can be avoided to some
- 57 -

extent by introducing "Significance thresholds" (see No 5.2 in checklist) and by conducting a full-sized LCA including normalisation and valuation.
Table:

Main impact categories and subcategories in LCA studies for biodegradable polymers (developed on the basis of Udo de Haes, 1996)
Advantage
renewables
I.)

Advantage petrochemicals Neutral

Impact categories

I.A) Input related categories
- Abiotic resources
- Biotic resources
- Land use

X
X
X

I.B) Output related categories
- Global warming
- Depletion of stratospheric ozone
- Human toxicity
- Ecotoxicity
- Photo-oxidant formation
- Acidification
- Eutrophication
- Odour
- Noise
- Radiation
- Casualties
II.)

X
X
X
1)

X

X
X
X
X
X
X
X

Impact subcategories and
2)
other impact indicators
- Carbon resources (renewable & non-renew.)
- Non-renewable resources
- Non-renewable energy (fossil & nuclear)
- Nuclear energy
- AOX
- Lead
- Carcinogeneity
- Diesel particulates
- Total waste
- Hazardous waste

1)
2)

X
X
X
X
X
X
X
X
X
X

Mainly due to biocide and pesticide use
Selection of indicators used in LCA studies on biodegradable polymers: some of these indicators are categorised by Udo de
Haes (1996) as "Pro Memoria Categories". Udo de Haes defines these as "truncated flows" that cannot be allocated to the categories extraction or emissions. Examples named are energy and waste.

Finally, it is important to note the limitations of an LCA as a tool for decision support.
Finnveden (2000) points out that it is - strictly spoken – impossible to show by means of an LCA that one product is environmentally preferable to another. This has to do with the fact that universal statements are logically impossible to prove. Let us, for example, assume that a product A is (objectively) preferable to product B in environmental terms. Even if there is an LCA showing this, it is likely to contain some methodological and empirical choices that are uncertain to some extent. For example, it will probably be impossible to show that all relevant impact categories have been taken into account. It will therefore not be possible to prove the general environmental superiority of product A. If such a proof must be provided as a precondition for a decision at the company or governmental level, it is very likely that no action will ever take place. If, on the other hand, society wants to be able to act then it is inevitable to make decisions on a less rigid basis (Finnveden, 2000).

- 58 -

Table A2-1:

List of LCA impact categories of particular relevance for biodegradable polymers 1.
2.
3.

Non-renewable energy (fossil & nuclear)
Land use*)
Inorganic resources

4.
5.
6.
7.
8.
9.
10.
11.

Global warming
Depletion of stratospheric ozone
Human toxicity
Ecotoxicity
Photo-oxidant formation
Acidification
Eutrophication
Hazardous waste

*) The subcategories biodiversity and soil quality may be accounted for in qualitative terms.

Table A2-2:

Significance thresholds used in an LCA study on biodegradable polymers
(Estermann et al., 2000)
The environmental impact is…

Environmental categories

Symbol

Energy, waste

All others

> 200%

> 500%

++

…higher

125% - 200%

167% - 500%

+

…comparable

80% - 125%

60% - 167%

o

…lower

50% - 80%

20% - 60%

-

< 50%

< 20%

--

…much higher

…much lower

- 59 -