Composition of waste materials and recyclables

Abstract

As the valorization of waste has become a main objective of modern waste management a variety of waste technologies were developed and the complexity of management systems increased. Maximizing environmental benefits in one part of the system may lead to burdens in another. Environmentally sound decisions in waste planning thus require a holistic and systematic assessment of environmental impacts of different waste management options. Such assessment requires reliable information on the physical and chemical waste properties to model the flows of waste materials and substances throughout the entire waste management system. The aim of this PhD project was to improve the understanding of factors influencing the quality of waste composition data during waste characterization and application in the environmental assessment of waste management systems. Reviewing existing waste characterization studies revealed that a large variety of waste characterization approaches and analytical methods has been employed. The most frequently used approach was the chemical analysis of directly sampled waste materials which offers the highest flexibility for waste characterization studies. Direct waste analysis involves several steps to prepare the samples mechanically and/or chemically for final analysis. Not all sample preparation methods are equally well suited for specific waste characterization purposes. The correctness of results and practical feasibility of sample preparation was strongly affected by the material type of the sample and the physico-chemical parameter to be analyzed. For example, studies examining mechanical sample preparation methods suggest that plastic fractions are especially prone to de-mixing effects and that differing mechanical properties within a sample (e.g. plastic and metal) can lead to biased results. In the experimental part of this PhD project the milling of plastics and metals was especially challenging and alternative methods for preparation and analysis should be investigated. Furthermore, chemical sample preparation by means of acid digestion was found to severely affect the element content resulting from chemical analysis. Although the suitability of standardized HF-containing methods can be generally confirmed, these methods led to considerable underestimations of the element content for some combinations of element and waste material. Appropriate selection of acid digestion methods thus needs to take the waste material and the elements to be analyzed into account. The dataset obtained during this PhD project provides information on the performance for six relevant acid mixtures for nine different waste material fractions and 64 elements and can support the selection of appropriate acid digestion method for future waste characterization studies and the comparison of data across existing studies. A consistent dataset for 73 physico-chemical parameters in 49 residual and 24 source-segregated Danish household waste fractions was obtained and is now available for future modelling and assessment of waste management systems. The analyzed fractions were selected based on material properties with relevance for potential recycling processes. The physico-chemical analysis revealed chemical differences between residual and source-segregated samples for several fractions. The results for parameters associated with organic matter confirmed the idea of cross-contaminated recyclables in residual waste, whereas the results for heavy metals and trace elements were more complex. For many fractions rather high metal contents were found to be intrinsic properties of the recyclables. For example, the Sb content in PET packaging was 250-270 mg/kgTS. In some cases metal contents in source-segregated fractions were higher than in the respective fractions from residual waste. Rare earth elements (RRE) were quantified in all analyzed material fractions and considerably high concentrations (e.g. the Ce content in ceramics was 72 mg/kgTS) were associated with mineral and soil-like materials. This “natural background” concentrations need to be considered when concluding on the origin of REE which are typically associated with electronic/hazardous waste. In general the use of primary physico-chemical waste characterization data for the environmental assessment of waste management systems is always preferable because many factors such as fraction definition and sampling point can be controlled. Anyhow, value ranges (as opposed to single values) should be considered due to the possibility of systematic bias (e.g. resulting from specific characterization methods) and “natural” variation. Commonly reported measurement uncertainties are not suitable to capture such effects. Thus secondary data should be considered when deriving uncertainty ranges and more research is necessary to quantify systematic effects of different characterization methods. Considering the extensive time and costs related to physico-chemical waste characterization it is likely that environmental assessment of waste management systems will continue to draw on secondary data from literature. For some parameters the values reported in literature were found to differ significantly. The wide range of values in literature for physico-chemical properties of individual waste materials were shown to severely affect the results of the life cycle assessment (LCA) of waste management systems, although the parameters had relatively low sensitivity. This emphasizes that sensitivity alone is not an appropriate indicator to identify critical parameters for LCA modelling. Using the literature value ranges for complete uncertainty analysis physico-chemical parameters contributed substantially to the output uncertainty of the modelling results and were in many cases more important than technology and scenario specific parameters. By selecting well-fitting data from literature the input value ranges could be potentially narrowed. Suggestions for physico-chemical data selection for environmental assessment and related uncertainty analysis are presented. At this point, however, low data availability and the large variety of used waste characterization methods limit the quantification and ranking of influencing factors via statistical data analysis of literature data. Nevertheless, it was found that the regional context of the data origin appeared to be less or equally important than other potential influencing factors (such as e.g. analytical methods, waste management system, natural variation, etc.). To consider the influence of such factors on secondary data but also primary data, LCA of waste management systems should always include a systematic uncertainty analysis for physico-chemical waste properties which needs to be based on both sensitivity and realistic uncertainty ranges for the individual physico-chemical parameters.