DIOXIN EMISSIONS FROM SMALL-SCALE COMBUSTION OF BIO-FUEL AND

HOUSEHOLD WASTE

Björn Hedman

Akademisk avhandling

Som med tillstånd av rektorsämbetet vid Umeå universitet för erhållande av Filosofie Doktorsexamen vid Teknisk-naturvetenskapliga fakulteten i Umeå, framlägges till offentlig granskning vid Kemiska institutionen, hörsal KB3B1 i KBC, fredagen den 14 oktober 2005, kl. 10.00 Fakultetsopponent: Brian Gullett, Ph.D., U.S. Environmental Protection Agency, Office of Research and Development, National Risk Management Research Laboratory, Research Triangle Park, North Carolina, USA

ii

Copyright © 2005 Björn Hedman

Department of Chemistry Environmental Chemistry

Umeå University SE-901 87 Umeå

Sweden

ISBN 91-7305-956-0

Printed in Sweden by VMC, KBC, Umeå University, 2005.

iii

Title:

Dioxin emissions from small-scale combustion of bio-fuel and household waste

Author: Björn Hedman, Department of Chemistry, Environmental Chemistry, Umeå University, SE-901 87 Umeå, Sweden

ABSTRACT This thesis deals with emissions of persistent organic pollutants, primarily dioxins, from the combustion of solid biofuels and dry combustible household waste in relatively small facilities, 5-600 kW, without advanced air pollution controls. Co-combustion of waste and biofuel in effective small boilers was tested as an alternative to prevailing large-scale management and combustion strategies for handling municipal solid waste. This approach includes no advanced air pollution control systems, but offers two advantages: limiting transport and providing scope to use local biofuel resources. Source-sorted, dry, combustible household waste was collected from households in a sparsely populated area and co-combusted as briquettes together with reed canary-grass in 150-600 kW biofuel boilers. Most trials showed difficulties to meet regulative limits for the emissions of dioxins valid for incineration of MSW and the regulated limits for emissions of hydrochloric acid were exceeded manifold. It was concluded that additional flue-gas cleaning will be needed to ensure that emissions are sufficiently low. Dioxins were also found in the waste, especially in the textile fraction. The mass of dioxins in the flue-gas emissions was generally lower than the mass in the fuel input. Intermittent combustion of wood pellets in a residential boiler resulted in an unexpectedly high dioxin emissions factor of 28 ng (WHO-TEQ)/kg fuel. Combustion of wood in a modern environmentally certified boiler yielded considerably lower dioxin emissions than combustion in an old boiler, and combustion with a full air supply, i.e. with use of heat storage tank, resulted in up to 90% reductions in dioxin emission factors compared to combustion with reduced air supply. Combustion of plastic waste in a residential wood boiler resulted in high emissions of dioxins. Tests of uncontrolled combustion of garden and household waste in barrels or open fires, ‘backyard burnings’, resulted in emissions with large variations that could only be partly correlated to the waste constituents. The results imply that this may be an important source of dioxins in the environment and an emission factor range of 4-72 ng (WHO-TEQ)/kg is suggested for estimating emissions from backyard burnings of lightly and moderately chlorine-contaminated waste. A summarized conclusion from all of the experiments is that predicting emission levels from waste contents is not straightforward (except that fuels with very high chlorine levels will usually result in high levels of dioxins in flue-gas emissions). Moderate differences in chlorine levels will usually be masked by the effect of variations in combustion conditions. Keywords:

Source-sorted waste, Household waste, Textile, Biofuel, Firewood, Pellets, Combustion, Wood boiler, Pellet boiler, Backyard burning, Garden waste, PCDD/F, dioxin, PCB, Emission factor

ISBN 91-7305-956-0

iv

Svensk titel :

Dioxin emissioner från småskalig förbränning av biobränsle och hushållsavfall

Författare: Björn Hedman, Kemiska institutionen, Miljökemi, Umeå Universitet, 901 87 Umeå

SAMMANFATTNING Denna avhandling behandlar rökgasutsläpp av persistenta organiska föroreningar, framförallt dioxiner, vid förbränning av fasta biobränslen och torrt hushållsavfall i relativt små anläggningar (5-600 kW) utan avancerad rökgasreningsteknik. Samförbränning av avfall och biobränsle i effektiva mindre biobränslepannor testades som en alternativ strategi till den vanligen förekommande storskaliga hanteringen och förbränningen fast hushållsavfall. Medan storskalig förbränning av avfall ger investeringsmässiga fördelar med rökgastenik etc. kan små lokala anläggningar ha transportmässiga fördelar och möjligheter till utnyttjande av lokala biobränsletillgångar. Källsorterat, torrt, brännbart hushållsavfall insamlades från hushåll i glesbygd och samförbrändes i brikettform med energigräset rörflen i 150-600 kW biobränslepannor. Endast undantagsvis understeg dioxinemissionerna gällande gränsvärden för avfallsförbränning och nivåerna av väteklorid i rökgas översteg gränsvärdena flerfaldigt. Det bedömdes att någon form av extra rökgasrening är nödvändig för att säkerställa nivågränserna. Dioxiner hittades också i det eldade avfallet, framförallt i textilfraktionen. Dioxinmängderna i rökgaserna var oftast lägre än i det ingående bränslet. Intermittent pelletseldning gav oväntat höga utsläpp av dioxiner med en emissionsfaktor på 28 ng(WHO-TEQ)/kg. Vedeldning i en modern miljömärkt villapanna gav betydligt lägre utsläpp av dioxiner än eldning i en gammal kombipanna och eldning med full lufttillförsel, som kan jämföras med användning av ackumulatortank, resulterade i up till 90% minskning av utsläpp av dioxiner jämfört med eldning med reducerat lufttillskott (’pyreldning’). Eldning av plastavfall i en vedpanna gav höga utsläpp av dioxiner. Okontrollerad förbränning av trädgårdsavfall och hushålls avfall i tunna eller som öppen eld ’bakgårdsbränning’, gav stora variationer i utsläppsnivåer som bara delvis kunde kopplas till avfallsinnehåll. Resultaten visar att denna typ av förbränning kan vara en betydande källa till dioxiner i miljön, och en emissionsfaktorskala på 4-72 ng (WHO-TEQ)/kg föreslås för bedömningar av utsläpp från backgårdsbränning av avfall med låga eller måttliga klorhalter. En sammanfattande slutsats av alla försök är att dioxin utsläpp beror på komplicerade samband mellan bränsle innehåll och förbränningsbetingelser. Bränslen med mycket höga klorhalter av ger oftast högre utsläpp av dioxiner än bränslen med låga klorhalter meden små skillnader döljs av variationer i förbränningsbetingelser. Nyckelord:

Källsorterat avfall, Hushållsavfall, Textil, Biobränsle, Ved, Pellets, Förbränning, Vedpanna, Pelletpanna, Trädgårdsavfall, PCDD/F, dioxin, PCB, Emissionsfaktor

ISBN 91-7305-956-0

v

List of abbreviations

APC Air pollution control CE Combustion efficieny EF Emission factor FTIR Fourier transform infrared spectrometry GWP Global warming potential HCB Hexachlorobenzene HCl Hydrgen chloride IR Infrared LCA Life cycle assessment MSW Municipal solid waste MWI Muncipal waste incinerator PAH Polycyclic aromatic hydrocarbons PCA Principal component analysis PCB Polychlorinated biphenyl PCBz Polychlorinated benzenes PCDD Polychlorinated dibenzo-p-dioxins PCDD/F "Dioxins" (PCDD + PCDF) PCDF Polychlorinated dibenzofurans PCP Pentachlorophenol PCPh Polychlorinated phenols PLS Projection to latent structures POP Persistent organic pollutant PVC Polyvinyl chloride RCG Reed canary-grass RDF Refuse derived fuel TEF Toxic equivalency factor TE Toxic equivalents (general) WHO-TEQ Toxic equivalens according to the World Health Organization The dioxin homologues TCDD Tetrachlorinated dibenzo-p-dioxins PeCDD Pentachlorinated dibenzo-p-dioxins HxCDD Hexachlorinated dibenzo-p-dioxins HpCDD Heptachlorinated dibenzo-p-dioxins OCDD Octachlorinated dibenzo-p-dioxins TCDF Tetrachlorinated dibenzofurans PeCDF Pentachlorinated dibenzofurans HxCDF Hexachlorinated dibenzofurans HpCDF Heptahlorinated dibenzofurans OCDF Octachlorinated dibenzofurans

vi

List of papers This thesis is based on the following papers, which are referred to in the text by their respective Roman numerals.

I. Hedman B, Burvall J, Nilsson C and Marklund S. Emissions from small-scale energy production using co-combustion of biofuel and the dry fraction of household waste. Waste Management 2005 Vol. 5 (3), 311-321

II. Hedman B, Nilsson C and Marklund S. PCDD/F in source-sorted

waste fractions and emissions from their co-combustion with reed canary-grass. Submitted to Environmental Science & Technology

III. Hedman B, Näslund M and Marklund S. Emission of PCDD/F,

PCB and HCB from combustion of firewood and pellets in residential stoves and boilers. Submitted

IV. Hedman B, Näslund M, Nilsson C and Marklund S. Emissions of

PCDD/F and PCB from uncontrolled burning of garden- and domestic waste (back-yard burning). Accepted for publication in Environmental Science & Technology

vii

TABLE OF CONTENTS 1 INTRODUCTION .................................................................................................................1 2 BACKGROUND....................................................................................................................2

2.1 Dioxin nomenclature, structures and environmental relevance..................................2 2.2 Formation of dioxins...................................................................................................3 2.2.1 Temperature.............................................................................................................5 2.2.2 Oxygen. ....................................................................................................................5 2.2.3 Particles (soot and fly ash) ......................................................................................6 2.2.4 Catalysts and inhibitors ...........................................................................................6 2.2.5 How does the chlorine content affect dioxin formation? .........................................7 2.2.5 Profiles.....................................................................................................................8 2.3 Waste management .....................................................................................................8 2.3.1 Waste characterizations.........................................................................................10 2.4 Waste combustion .....................................................................................................12 2.5 Dioxins in waste........................................................................................................12 2.6 Biofuel.......................................................................................................................12 2.6.1 Reed canary-grass .................................................................................................13 2.6.2 Environmental problems associated with biofuels.................................................14 2.7 Combustion principles ..............................................................................................15 2.8 Backyard burning......................................................................................................15

3 EXPERIMENTAL ..............................................................................................................16 3.1 General outline of trials............................................................................................16 3.2 Materials...................................................................................................................17 3.2.1 Waste fuels .............................................................................................................17 3.2.2 Biofuels ..................................................................................................................17 3.2.3 Combustion equipments .........................................................................................18 3.3 Methods.....................................................................................................................19 3.3.1 Sorting analysis......................................................................................................19 3.3.2 Logging equipment.................................................................................................19 3.3.3 Sampling flue-gas...................................................................................................20 3.3.4 Analysis..................................................................................................................21

4 RESULTS AND DISCUSSION..........................................................................................23 4.1 Sorting analysis of source-sorted waste (Papers I and II) ........................................23 4.1.1 Chlorine and PCDD/F in waste fractions and fuel................................................23 4.1.2 Flow of dioxins, Recycle or incinerate – contaminated waste? .............................28 4.2 Co-combustion of biofuel and RDF (Papers I and II)...............................................28 4.3 Combustion of pellets and fire-wood in residential stoves and boilers (Paper III) ..29 4.4 Backyard burning (Paper IV) ...................................................................................30 4.5 Contents of fuels and their significance for combustion emissions (Papers I-IV) ....32 4.5.1 The significance of chlorine contents for dioxin emissions....................................32 4.5.2 Information provided from homologue profiles .....................................................37

5 CONCLUSIONS AND SUGGESTIONS FOR FUTURE WORK..................................40 6 ACKNOWLEDGEMENTS ................................................................................................42 7 REFERENCES ....................................................................................................................43

viii

1

1 Introduction Once upon a time, the waste from human societies amounted to no more than usually could be left on and around settlements, delighting future archaeologists and not usually doing any great harm to the environment. It is not so anymore. Industrialized society demands huge amounts of raw materials for the production of energy and consumer products, leading to wastes of similar scale. In recent decades there has been increasing concern about the consequences of irreversible linear movements of material and energy, and the concepts of sustainability and recycling have gained increasing acceptance. However, even if the problems are often obvious, the solutions aren’t always. The environmental impact of different waste management strategies has been debated, not least for the emissions of dioxins and other pollutants from waste incineration. The increasing demands for energy have to a large extent been solved by exploiting increasing amounts of fossil energy; the resulting emissions of greenhouse gases being one of the most serious consequences. Another tendency is for large-scale solutions to be preferred. Both large-scale and small-scale systems can have advantages as well as disadvantages. Combustion in large-scale facilities often permits the use of more technically advanced (and costly) systems for handling residues, such as air pollution controls (APC) in waste combustion, whereas small-scale, local energy and waste handling systems can offer advantages by minimizing the transport of material . This thesis deals primarily with some examples of small-scale combustion of waste and biofuel. Issues considered include the effects of the composition of the fuel and combustion parameters on the emission of dioxins and other chloro-organic pollutants. The thesis also contributes to the discussion on the relative merits of different waste handling strategies by comparing pollutants in combustion inputs and the outputs they generate.

2

2 Background

2.1 Dioxin nomenclature, structures and environmental relevance The term ‘dioxin’ is often used as a general term encompassing more than true dioxins. Instead, it usually refers to two groups of structurally related polychlorinated aromatic compounds; polychlorinated dibenzo-p-dioxins, PCDD (of which there are 75 species) and polychlorinated dibenzofurans, PCDF (135 species). The structures of the species within each group differ in terms of the number and spatial arrangements of chlorine atoms (Figure 1). The individual compounds within the groups are often referred to as congeners and congeners with the same number of chlorine substituents are called homologues. Abbreviations for the compounds sometimes differ between literature sources, but those used in this thesis and the papers it is based upon are listed on page v.

O

O

O

12

3

45

7 8

6

1

2

3

45

7

6

8

Figure 1. Structural formulas of PCDD (above) and PCDF (below). The numbers (1-8) indicate possible positions of chlorine substituents. Dioxins are chemically stable and, like many other chlorinated organic compounds, are included in a group given the generic term persistent organic pollutants (POPs). Because of their highly lipophilic character they tend to accumulate in fat and, thus, throughout the food chain. The highest levels are often found in fatty tissues of animals at the top of the chain. Organisms that obtain most of their food from the sea have been subjected to especially high levels of dioxins. These include fishes, and sea-living predators [1-3]. Humans that consume large amounts of fish have also shown elevated levels of dioxins [4]

3

Dioxins with chlorines in all the positions 2,3,7 and 8 (Figure 1) have been shown to bind strongly to the Ah-receptor, a protein located in the cellular plasma. This bonding results in a chain of reactions that affect DNA transcription and can eventually lead to complex pathological alterations, including mutagenic and carcinogenic phenomena [5]. 2,3,7,8-TCDD is considered one of the most toxic compounds known. However the toxic effects differ substantially between species and generally humans are considered to show comparably low sensitivity to dioxins [5]. When evaluating the occurrence of dioxins, information on their toxic effects is often required. The toxicity however differs greatly between congeners. Thus, dioxin levels are often presented as combined values, toxic equivalents (TEQs). In order to calculate these TEQs each of the species has been attributed a toxic equivalent factor (TEF) that corresponds to its toxicity relative to that of the most toxic dioxin congener (2,3,7,8-TCDD). A number of different TEF scales have been developed in recent decades, some of which are presented in Table 1, as more information on the toxicity of different species has become available, and varying measures of species-related toxic effects have been applied. Lately, the scale promoted by the World Health Organization (WHO) in 1997 [6], which also includes some polychlorinated biphenyls (PCB) with similar toxic actions to the PCDD/F, has been the most commonly used. This scale, which currently is under revision, has different versions for humans and mammals, fish and birds. The WHO-TEQ scale was also generally used in the work underlying this thesis, except in paper I, where the I-TEQ scale (Table 1) was used.

2.2 Formation of dioxins Ideally, combustion of organic materials yields nothing but CO2 and H2O, provided that only carbon, hydrogen and oxygen are involved and the combustion is 100% effective. In reality, immense numbers of different reactions take place, and whenever chlorine is present, chloro-organic compounds are bound to form. Several factors influence the total amounts of dioxins, and the relative amounts of different homologues and congeners, that are formed. Obviously C, H, O and Cl have to be available, but variables like temperature, carbon source, oxygen level, the presence of particles and catalytic elements etc. also play important roles. Formation can take place either from pre-existing precursor molecules like polychlorinated benzenes (PCBz) and polychlorinated phenols (PCPh) or from newly formed precursors originating from aliphatic hydrocarbons [7,8]. These reactions can take place either in the gas-phase (homogeneous reactions) or on particle surfaces (heterogeneous reactions) [9]. Another possible route of formation is ‘de novo’ formation from elemental carbon available in soot or fly ash particles. The ‘de novo’ route has been defined as the breakdown reactions of a carbon matrix [10]. The formed carbon skeleton can be further chlorinated in following steps [11,12]. The importance of the ‘precursor’ and ‘de novo’ models has been disputed and the relevance of this division of formation routes has also been questioned [13,14].

4

Table 1. Toxic equivalency factors (TEFs) for PCDD/Fs and PCBs Congeners WHO

(Human and mammals)

International (I-TEQ)

Nordic Eadon

2,3,7,8 -TeCDD 1 1 1 1 1,2,3,7,8 -PeCDD 1 0.5 0.5 1 1,2,3,4,7,8 -HxCDD 0.1 0.1 0.1 0.033 1,2,3,6,7,8 -HxCDD 0.1 0.1 0.1 0.033 1,2,3,7,8,9 -HxCDD 0.1 0.1 0.1 0.033 1,2,3,4,6,7,8 -HpCDD 0.01 0.01 0.01 OCDD 0.0001 0.001 0.001 2,3,7,8 -TeCDF 0.1 0.1 0.1 0.33 1,2,3,7,8 -PeCDF 0.05 0.05 0.01 0.33 2,3,4,7,8 -PeCDF 0.5 0.5 0.5 0.33 1,2,3,4,7,8 -HxCDF 0.1 0.1 0.1 0.01 1,2,3,6,7,8 -HxCDF 0.1 0.1 0.1 0.01 1,2,3,7,8,9 -HxCDF 0.1 0.1 0.1 0.01 2,3,4,6,7,8 -HxCDF 0.1 0.1 0.1 0.01 1,2,3,4,6,7,8 -HpCDF 0.01 0.01 0.01 1,2,3,4,7,8,9 -HpCDF 0.01 0.01 0.01 OCDF 0.0001 0.001 0.001 3,4,4’,5 -TeCB (# 81)1) 0.0001 3,3’,4,4’ -TeCB (# 77) 0.0001 3,3’,4,4’,5 -PeCB (# 126) 0.1 3,3’,4,4’,5,5’ -HxCB (# 169) 0.01 2,3,3’,4,4’ -PeCB (# 105) 0.0001 2,3,4,4’,5 -PeCB (# 114) 0.0005 2,3’,4,4’,5 -PeCB (# 118) 0.0001 2’,3,4,4’,5 -PeCB (# 123) 0.0001 2,3,3’,4,4’,5 -HxCB (# 156) 0.0005 2,3,3’,4,4’,5’ -HxCB (# 157) 0.0005 2,3’,4,4’,5,5’ -HxCB (# 167) 0.00001 2,3,3’,4,4’,5,5’ -HpCB (# 189) 0.0001 1) Numbering according to International Union of Pure and Applied Chemistry (IUPAC) are shown in parantheses However, the precursor route is considered to mainly produce PCDD, and the PCDF/PCDD ratio [15] has been used as an indictor of the dominating reaction; a ratio of less than unity being regarded as a sign of precursor predominance. Another proposed measure is the dioxin fraction, DF=PCDD/ ∑PCDD/F, which gives better graphic representation, since the values range from zero to one instead of between zero and infinity [13].

5

This thesis is not intended to contribute to the debate regarding formation mechanisms, but an understanding of the factors underlying the mechanisms that influence the formation of PCDD/Fs and other POPs are important in the interpretation of the experimental results. A survey of such factors follows.

2.2.1 Temperature Both formation and destruction of dioxins will normally take place in combustion. Since the activation energies for the various destruction and formation reactions differ, the temperature at which maximal net production occurs also varies [16]. On fly ash particles the destruction has been shown to be virtually complete at temperatures as low as 500-600 °C [16]. However, other experiments have indicated that PCDD/F can occur at temperatures up to 900 °C [17]. Formation has been shown to take place mainly in two temperature regions; 500-800 °C for homogenous reactions and 200-400 °C for heterogeneous reactions [16]. Since temperatures in furnaces are usually higher than 800 °C, the production of dioxins mainly occurs during the cooling of the flue-gas. Consequently, in the development of combustion facilities, measures are often taken to make the cooling as rapid as possible to shorten the residence time of the flue gas in these temperature regions.

2.2.2 Oxygen. Apart from being one of the elements included in the dioxins, oxygen levels affect combustion efficiency and thus, indirectly, dioxin formation. However, oxygen is also involved in other processes. Gaseous oxygen is considered necessary for the de novo reactions in the initiation of carbon gasification and rearrangement [16,18]. It is also needed in the Deacon reaction, which is suggested to produce the active chlorinating agent Cl2 from the less active HCl [19]. 2HCl + ½ O2 → Cl2 + H2O (1) Oxygen also plays a role in oxidizing reduced copper, Cu(I), in a shuttle mechanism of chlorination [7]. Up to 10% [O2(g)] have been shown to increase PCDD (and to a lesser extent PCDF) formation [18,20,21]. Gaseous oxygen has also been shown to be incorporated to a greater degree in highly chlorinated PCDD/F homologues, whereas the formation of low chlorinated homologues appears to depend more on oxygen that is chemically bonded or adsorbed to the solid matrix [18]. Oxygen deficiency is one of the factors that can lower combustion efficiency (CE), defined as the ratio of organic material destroyed, divided by the total amount of organic material available [22,23]. It is also often accompanied by increasing levels of PAH and soot, which may increase dioxin formation by supplying carbon sources. These findings may seem to conflict with abovementioned requirement for oxygen for dioxin formation, but the relationships between oxygen level, combustion efficiency and dioxin formation are more complex than outlined here [23] and formation of carbon skeletons is only one step in dioxin formation. Wikström showed that low CE gives unchlorinated and low chlorinated PCDD/F, and that high O2 levels are needed for the formation of higher chlorinated dioxins [22].

6

2.2.3 Particles (soot and fly ash) Combustion of solid fuels releases various amounts of particles composed of fly ash, incombustible mineral residues, that are sufficiently low in density or size to co-migrate with the flue gas, and/or soot, which consists of carbonaceous particles formed in combustion. Fly ash particles from MSW incinerators usually contain 2-5 % carbon of poorly investigated nature [16]. Soot consists of roughly spherical particles with a graphitic structure and the approximate empirical formula (C8H)n [24]. Soot particles are formed around nuclei, and with increasing age the particles grow and aggregate in chainlike structures up to 200 nm long or more. Soot formation is accompanied by the formation of polycyclic aromatic hydrocarbons (PAH). PAH often condense on particle surfaces and they are also considered to be intermediates in the formation of soot [24,25]. Soot and fly ash can act as carbon sources in dioxin formation. They can also provide surface structures for heterogeneous reactions, but only fly ash provides the necessary catalytic elements for these reactions. The proximity of the carbon source and catalyst has been used to explain the observation that PCCD/F are usually more abundantly produced in combustion experiments in the presence of fly ash than with synthetic mixtures of carbon and catalysts [16,26]. However, in other studies [27,28]it has been found that the PCDF, and to some degree PCDD, can be formed from molecular backbone structures derived from the soot matrix, catalysed by present fly ash. Dioxins emitted in flue gas are mainly adsorbed to particles in amounts that have been shown to correlate with the external surface area of the particles [16,29]. The largest amounts of emitted dioxins are found on relatively small particles, which have a relatively large specific area [30,31]. To ensure effective flue gas cleaning, this must be taken into account.

2.2.4 Catalysts and inhibitors The heterogeneous dioxin formation mechanisms (precursor and de novo) are catalysed reactions, in which the catalyst is active both in the oxidation of carbon and the chlorination and dechlorination of organic compounds. The most active catalyst is copper, but other metals, such as iron and nickel [32,33], also have showed some activity. Copper, which is suggested to be equally active in metallic, chloride and oxide forms [14], catalyses the Deacon reaction (Equations 2 and 3, summarized in Equation 1) and by alternating between its two oxidation states it also acts as a shuttle for chlorine between the gas phase and the solid carbonaceous material [7]. This is energetically more favourable than the Deacon reaction and can work at lower temperatures. The catalytic activity can be inhibited by sulphur and nitrogen compounds, via the formation of sulphates (Equation 4) and nitrides. Sulphur can also inhibit it by reducing the chlorinating agent Cl2 to HCl (Equation 5). These inhibitors have been added in pilot scale combustions as sulphuric acid, ammonia, urea etc., and in full scale combustions of MSW in fuels (e.g. coal) with high sulphur contents [14,34].

7

2CuCl2 + ½O2 Cu2OCl2 + Cl2 (2) Cu2OCl2 + 2HCl 2CuCl2 +H2O (3) CuO + SO2 + ½O2 CuSO4 (4) Cl2 + SO2 + H2O SO3 + 2HCl (5)

2.2.5 How does the chlorine content affect dioxin formation? Of particular interest when discussing the combustion of waste are the effects of the level and forms of chlorine in the waste fuel on emissions of PCDD/Fs. However, these effects are not easy to elucidate. Chlorine can occur in the fuel in both inorganic and organic forms (e.g. chlorides and PVC, respectively). More than 90% of fuel chlorine is emitted as HCl in combustion [23], and of the remaining chlorine, some will be found as chlorides, in the bottom ash or fly ash, while a small amount will be found as chlorinated organic compounds in the ashes or gaseous emissions. Emitted amounts of HCl may be correlated to Cl levels in the fuel [35], but HCl does not function as a chlorinating agent. To react with stable organic compound in the gas phase the atomic Cl radical is needed. This is most likely supplied from the Cl2 molecule, although the reaction of HCl with the OH radical has also been suggested as a possible contributor [14]. The formation of atomic Cl is possible at high temperatures, and high concentrations of it can remain at lower temperatures after rapid cooling [14,16]. In solid phase reactions ash-bound chlorine (e.g. CuCl2) is also a possible source of chlorine for chlorination reactions [7]. It has been suggested that CuCl2 is the main chlorinating agent and ash-bound chlorine has been calculated to be a sufficient source for chlorination in de novo reactions [12,36]. If ash-bound chlorine is depleted, further chlorine can be supplied from the gas phase by the action of Cu. After CuCl2 has donated one of its chlorines to an aromatic compound, it may be re-chlorinated by gaseous chlorine [7,16,37]. Gullet et al. [14] suggests several other possible chlorination mechanisms derived from other areas of chemistry. Attempts to correlate the amount of PCDD/F formed with chlorine levels in fuels have been made with varying degrees of success. Several authors have stated that other variables are more important than the chlorine levels of fuels [14,23,38], some have found correlation [39-42] and other s have found correlation above certain level of chlorine [30,43,44]. Stanmore explained these variations by postulating that freely available chlorine chlorinates 3-ringed phenanthrene structures. In low Cl conditions, the rate of this reaction may be controlled by oxygen availability, but under excess Cl it may be proportional to the concentration of free chlorine [16]. A tendency for high levels of chlorine to inhibit de novo formation has also been shown [16,45,46]. Very little of the available chlorine is needed to form considerable amounts of dioxins. A rough calculation, based on an emission factor of 0.6 ng/kg for combustion of MSW at the limit value 0.1 ng/m3

(n) [47] suggests that less than 1/1010 of the chlorine in fuel is needed to form the dioxin molecules. Chemical reactions are to a great deal a matter of statistical probabilities. Variations in the amounts of reacting compounds influence the relative distribution of competing reactions, but when a very small fraction of an available compound is needed, like the amounts of chlorine

8

needed for the formation of dioxin, there is greater scope for other variables to influence the reactions. The probability that fuel chlorine will be incorporated in dioxin or furan molecules can be divided in two steps. Firstly, there is the probability of fuel chlorine becoming an active chlorinating agent, as species in either the ash (e.g. CuCl2) or the gas phase. If the ash-related chlorines are the most important chlorinating agents, the formation of HCl is part of a less important alternative route, and the correlation between HCl and PCDD/F is not a very relevant measure. Secondly, there are the opportunities for the chlorinating agent to chlorinate appropriate molecules. Both steps are regulated by external conditions that may act on the steps in the same or opposite directions. Consequently, even if it is possible to correlate chlorine contents in fuel to the amounts of dioxins produced in studies where these factors can be regulated, such correlations can rarely be used in a predictive manner in other situations where the combustion conditions are different or unfamiliar.

2.2.5 Profiles The variety of formation routes for PCDD/F, each favouring specific distributions of PCDDs and PCDFs with different levels of chlorination, has led to attempts to use the relative distributions of congeners to extract information about the conditions prevailing during their formation [3,48-50]. They have also been used to identify sources, since different formation sources can give different profiles [51-57]. Some researchers have considered the pattern of congeners within homologue groups found to be relatively consistent, and have found the relative amounts of homologue groups to be more informative [58-60] while others has gained more information from the individual congeners [50,53,55,61]. Profiles from combustion sources usually are dominated by dioxins with a low number of chlorines, or have a fairly even distribution [50,54,59,62] while highly chlorinated dioxins, especially OCDD, tend to persist longer, or even are formed [54]in the environment and thus, they often will dominate profiles of dioxins found in soils and sediments [54,63], human tissues [61,64-66] and ambient air [53,55,67]. Homologue and congener profiles from atmospheric depositions has presented characteristics between that from combustion sources and environmental sinks, however they showed more similarities with the former [62].

2.3 Waste management Waste is a term that can have somewhat different meanings, e.g. in technical and legislative contexts and everyday use. A definition used by the EU is that “…‘waste’ means any substance or object which the holder disposes of or is required to dispose …” [68]. A useful common definition is that waste consists of objects or materials that have lost their utility value and have to be disposed of, unlike reusable and recyclable products that still have utility value. The dividing lines between waste and recyclables are not always clear, and may shift depending on the circumstances. This thesis mainly deals with the solid household waste that is handled by municipal waste

9

management systems. This is a category that sometimes includes objects that could also be classified as recyclable or reusable products. For several decades, landfill has been a common means of waste disposal. However, problems associated with increasing flows of waste and awakening environmental awareness have led to the development of multiple waste treatment options, all with different environmental implications. To judge the relative merits of the possible alternatives various ‘rule of thumb’ waste treatment hierarchies have been proposed [68], which generally rank the possibilities as follows:

1) Reduce generation of waste 2) Reuse products 3) Recycle material 4) Recovery of energy by incineration. 5) Landfill.

The rankings of organic treatments, like anaerobic digestion and composting, in the hierarchy are more controversial. Anaerobic digestion, as a method of energy recovery, has usually been placed on a level with incineration, whereas composting has been more questioned due to potential problems associated with methane production [69]. The waste treatment hierarchy, as a general principle for all waste, has been questioned, with the argument that it cannot provide any guidance with respect to using combinations of treatment technologies [70]. Furthermore, the rankings of different options may depend on specific circumstances (e.g. transport considerations) even for single items of waste [69]. A more sophisticated concept is Integrated Solid Waste Management (ISWM). This can be defined in various ways, but generally it is an optimised waste management system in which the environmentally and economically best solution for each individual case is sought, and the hierarchy plays a less important role [70,71] The outcomes of different waste management strategies can be evaluated by system analytic tools, like Life Cycle Assessment (LCA) [69,71]. LCA studies on waste incineration have dealt with energy parameters and emissions of greenhouse gases and heavy metals and also PCDD/F emissions [71-74]. Comparing incineration to other waste management methods with respect to dioxin emissions is however considered problematic since these are generally regarded as being process-related rather than product-related (that is, not related to the amount or content of incinerated waste[71]. In a synthesis of three system analysis studies of different waste management strategies [75], it is concluded that:

Material recovery is generally environmentally preferable compared to combustion, but not necessarily for some material (e.g. some types of paper).

Neither combustion nor anaerobic digestion is unambiguously better than the other in cases where there is a choice between them, but both are generally better than composting and landfill.

10

Landfill of waste is a worse alternative than other possible treatments. Transport of waste, once it is collected, is of minor energetic, environmental

and economic significance. A study comparing waste management strategies in different municipalities [72,76] suggested that in sparsely populated areas transport of waste, e.g. to recycling facilities was relatively significant, but its impact was difficult to assess. Comparisons of different waste combustion strategies showed that if a clean, quality assured, source-sorted waste fuel could be produced, it could be combusted in relatively unsophisticated plants with approximately the same environmental impact as combustion of unsorted waste in plants with advanced cleaning technology [76]. Few LCA studies have included data on pollutants in the waste and no LCA that considered PCDD/F in waste fractions has been found in the scientific literature, which is a consequence of a general lack of data related to this subject. The household waste generated in Sweden in 2004 amounted to 4.2 million tons [77], corresponding to 463 kg per person. The proportions of waste treated by different methods are shown in Table 2. Table 2. Amounts of Swedish household waste treated by different methods in 2004

Treatment

Kg/capita %

Material recovery (recycling) 153.7 33.2 Biological treatment 48.1 10.4 Landfill 42.2 9.1 Incineration with energy recovery 215.8 46.7 Hazardous waste 2.9 0.6 Total 462.7

2.3.1 Waste characterizations Sometimes there may be a need to characterize the waste generated by households, for several possible reasons, e.g. to assess its control, to plan collection and transport schedules, to estimate the required dimensions of waste treatment plants and to control the purity of recycled material [78]. Waste is characterised by sorting analyses, which can be performed at several points, depending on the purpose; e.g. sometimes it is done at the source, by examining the material collected in specific collection rounds, and sometimes by examining the material arriving at a waste treatment plants. The number and classification of waste sorting categories also depends on the objectives of the analysis and the design of the investigation determines the size of the samples required [78,79]. Sampling at the source, to evaluate the generation and composition of waste from different types of residential areas, generally demands considerably larger samples than sampling performed to evaluate the composition of the waste arriving at a waste treatment plant [78].

11

Table 3 Results of a sorting analysis of Swedish waste, 2005

Categories

%

Papers 7,8 Soft plastics 7,2 Hard plastics 3,3 Paper packages 8,5 Remaining plastics 1,1 Textiles 2,3 Wood 0,5 Remaining combustibles 4,2 Glass, metal, electronics 5,6 Diapers 5,5 Food waste 42,8 Garden waste 6,7 Hazardous waste 0,3 Remainder 4,3

A series of studies has been undertaken in seven Swedish municipalities, and the most recently published found the composition shown in Table 3 [80]. The eight categories at the top of the table are those that could be conceivably used as fuel in co-combustion. This fraction accounts for about 35% of the total, a large proportion of which is recyclable material (27 % of the total). A contributory investigation performed in local waste collection rooms, with facilities for handling both recycling and ordinary waste, assessed the proportions of recyclables that were actually separated as recyclable material. The fraction of combustible recyclable sorted into the recyclables containers, was, on average, 78%. These percentages have been applied to the results from the whole collection to calculate the theoretical composition of the source-sorted combustible waste with contributions from both combustible recyclable and non recyclable material (Table 4). Table 4. Calculated composition of combustible waste

Categories

%

Papers 56,0 Soft plastics 7,7 Hard plastics 4,9 Paper packages 24,9 Remaining plastics 0,7 Textiles 1,9 Wood 0,4 Remaining combustibles 3,5

12

2.4 Waste combustion The high combustion temperatures (850-1100 °C) in large-scale municipal waste incinerators (MWI) will cause virtually all of the organic pollutants, including dioxins, introduced with the waste to be destroyed. However, new formation takes places when the flue gas is cooled. Studies of dioxin levels in individual units of an MWI have shown that levels, PCDF/PCDD-ratios, congener distributions and the distribution between gas and particle phases can vary considerably between them [48,81-83]. Hunsinger et al. [49] noticed that levels above the fuel bed (but not the levels found in the raw gas downstream of the boiler) were influenced by the composition of the fuel. The lowest levels of dioxin they found were above the fuel bed in an area depleted of oxygen and at the maximum temperature (900 °C). Most of the dioxins formed are eventually caught with the fly ash in the air pollution control (APC) systems; between two thirds and sometimes nearly all (99.9%) of the dioxins emitted from the furnace according to some mass balance studies [81,82,84-86]. Dioxins has also shown to form in the APC, e.g. the fabric filter, and be caught in the filter ash [82]. In modern MWI there is usually no problem meeting regulatory limits for flue-gas emissions, e.g. 0.1 ng(WHO-TEQ)/m3n [87] but the disposal of fly ash may cause problems. The slag, or furnace ash, is usually deposited in landfills or used in a stabilized form in different applications and remnants from flue-gas cleaning, fly-ash, are treated as hazardous waste and are deposited, usually after some kind of stabilization [77]. Dioxins are generally considered to be so tightly bound to the ash that leakage will not constitute a major problem [88-90].

2.5 Dioxins in waste A few studies have investigated dioxin contents in combusted waste, usually focusing on the mass balances of dioxins in MWIs [47,82,84,91-93]. Most of the reported levels are listed in Table 5. However, it should be noted that some reported levels seem to derive from the same source, a study done in the Federal Republic of Germany around 1980 [15,82,94] and others used in mass balance studies are second hand [82,95] information cited from other sources [82,95,96].

2.6 Biofuel Biomass-fuels were the dominant energy sources globally before industrialization and in many areas of the world, especially developing countries, they still are. They are also important in many industrialized countries that have appropriate natural resources. In these countries they were displaced to varying degrees by cheap and user-friendly fossil fuels, but the oil crisis of the 1970s led to a come-back, and their use is also being stimulated by the development of more convenient fuels and combustion techniques.

13

Table 5. Reported levels of dioxins in waste

Dioxins in waste Reference Indicated type of waste TE, ng/kg ∑PCDD/F ng/kg MSW 30-40 8790-15510 [92] Mixed waste 23-32 [92] Solid waste (collected -80 41-207 [94] Waste (collected -89-91) 2,1-50 [94] MSW 27 3485 [93] MSW/RDF 45 [49] MSW 11-255 [97,98] USW (Urban Solid Waste) 46-87 [47,84] MSW 2,2-7,1 [84] MSW 90 [15,82] MSW 50 [82,95] MSW 1,3-16 [82,95,96] Total waste 6,3 [99] MSW 1,4-2,6 390-530 [83] MSW 0,34-0,571) [83] 1) Toxiq equivalents (WHO) of PCB

They are also being strongly promoted in the ongoing, intense debate on emissions of greenhouse gases. Biofuels is a term that includes all kinds of fuels derived from biomass: solids, liquids (generally ethanol or methanol) and gaseous (biogas). These fuels are considered favourable alternatives, in the context of greenhouse effects, because when they are combusted the resulting emissions of the greenhouse gas carbon dioxide are derived not from fossil deposits, but from carbon that is recycled in plant material, thus they cause no net increase in greenhouse gases in the atmosphere. In modern practice, solid biofuels are often formed into briquettes, pellets or powder to facilitate their transportation and to give the fuel greater uniformity and energy density. This increases the efficiency and consistency of combustion [100]and allows mixtures of fuels with desired qualities to be composed.

2.6.1 Reed canary-grass Reed canary-grass (RCG) (Phalaris arundinacea L) is a perennial rhizomatous, native in Sweden and many other countries. It shows good fuel and fiber properties if harvested in early spring, just before the start of the next growing season (spring-harvest or delayed-harvest) [101-104]. RCG as a grass has high ash contents (3-12%). The ash contents and compositions vary considerably depending on soil type and fertilization in growing [105-107]. Spring-harvesting reduces contents of elements (e.g. K and Cl) that can cause problems with melting behaviour and agglomeration (i.e. slagging, fouling and sintering) and the often shown high melting temperatures of

14

ash from this fuel has made it a conceivable candidate in co-combustion with fuels uncertain qualities [108].

Figure 1. Emission factors from combustion of biofuel (wood and straw). Logarithmic scales. Several of the larger facilities were equipped with air pollution control (APC) systems, mostly cyclones.

2.6.2 Environmental problems associated with biofuels However, despite its good qualities, the use of biofuel is not without environmental loads. Apart from problems related to cultivation, transportation and processing of the refined fuel, which may be significant in a life-cycle analysis perspective, of particular concern here are factors that affect its use in combustion, which can be divided into two categories.

1) Combustion-related: Biofuel is often combusted in ineffective, small-scale units, which can lead to unnecessarily high levels of products of incomplete combustion (PICs). It has even been reported that combustion in old wood boilers can lead to higher global warming potentials (GWP) than use of oil boilers, due to production of the greenhouse gas methane, [109].

2) Fuel-related: The levels of ash, heavy metals and chlorine etc. vary in different fuels such as wood, waste wood and straw. Wood may also contain residues of biocides, e.g. pentachlorophenol (PCP), used as a fungicide, which sometimes contains traces of dioxins [110-112].

Emission factors from biofuel combustion

0.001

0.01

0.1

1

10

100

1 10 100 1000 10000 100000

Effect ( kW)

PCD

D/F

(ng

TEQ

/kg

fuel

)

no APC

APC

15

Combustion of uncontaminated biofuels mostly results in relatively low emissions of dioxins, although there is some variation in the levels [113-115], as is shown in Figure 1.

2.7 Combustion principles The design of the fire chamber and associated systems of residential wood boilers and stoves can vary considerably [116]. The most important differences lie in the way the primary air is supplied to the fire. Older boilers and some modern stoves are constructed according to the up-draught principle, in which the air is supplied from under the fuel bed, usually through the ash lid, and passes upwards through the fuel bed. This causes the whole stack of wood to burn simultaneously, and the coefficient of utilization rarely exceeds 50%. Modern boilers are often designed to pass the primary air through the wood from above, resulting in so-called down-draught combustion [109]. This results in only the lower part of the wood burning, while the stack gradually sinks. With a ceramic secondary combustion chamber, such designs can provide 70-80% utilization coefficients. This type of construction usually requires a fan to maintain sufficient draught. Pellets can be combusted in various types of systems, including large boilers designed for more than one house, pellet stoves and burners intended for integration in existing boilers. Pellet burners can regulate the power output by means of fuel feeding, usually with intermittent operation. They show a variety of construction solutions of fuel supply (e.g. under-fed, over-fed and horizontal-fed), ignition after combustion breaks, air supply and draught dynamics etc.

2.8 Backyard burning As a consequence of the improved effectiveness of much of the waste incineration and combustion for energy generation, and dramatically diminished emissions of pollutants in large-scale combustions, burnings under uncontrolled conditions may now be the major source of PCDD/F in developed countries. Combustion in these sources has been little studied. The reported experiments generally show the same tendencies, with broad ranges of emission factors from most wastes and very high levels from PVC-containing wastes [117-122].

16

3 Experimental The equipment and methods used in the experiments underlying this thesis have mainly been described in the respective papers, and repetitive information will be avoided here. However, some overviews, comparisons and complementary data are presented in the following chapter.

3.1 General outline of trials The trials were generally small- or medium- scale (approximately 5-500 kW) combustion experiments conducted to evaluate emissions from the combustion of waste and biofuel with small-scale equipment without advanced cleaning systems. The experiments, equipment and analyses are summarized in Table 6. Table 6. Outline of work underlying this thesis.

Paper

Equipment Fuels Analysis

I Laboratory reactor (5kW); Briquette-fed biofuel boiler (150 kW)

Reed canary-grass and mixtures with source-sorted waste (0 %, 21% and 36% RDF)

Waste sorting analysis (minor); Heat value, C, H, N, Cl, ash in waste and fuels; Inorganic gases, POPs, particles, elements in dust in emissions; Unburned remains, POPs, Cl, elements in ashes.

II Briquette-fed biofuel boiler (150 kW); Briquette-fed biofuel boiler (600 kW)

Mixes of reed canary-grass and source-sorted waste (43% and 30% RDF)

Waste sorting analysis; Cl and PCDD/F in waste fractions; Heat value, C, H, N, Cl, ash in waste and fuels; Inorganic gases, POPs, particles, elements in dust in emissions; Unburned remains, POPs, Cl, elements in ashes:

III Pellet boiler (15 kW); Old wood boiler (12 kW); Modern eco-certified wood boiler (30 kW); wood stove (9 kW)

Pellets of wood and straw; Wood from birch, pine and spruce; Paper and plastic waste.

Moisture content of wood; Inorganic gases and POPs in emissions; POPs in ash.

IV Combustion in steel barrel and on a steel plate.

Garden waste, paper, paper and plastic packaging, RDF, PVC and electronic scrap.

Moisture content of waste; Inorganic gases and POPs in emissions; POPs in ash.

17

3.2 Materials

3.2.1 Waste fuels In order to obtain suitable material for the co-combustions described in Papers I and II, two collections of source-sorted dry combustible waste were made. First, a small amount of material (approximately two tonnes) was collected by approximately 30 employees at the Unit for Biomass Technology and Chemistry (BTK) at the Swedish University of Agricultural Sciences (SLU) and the Environmental Chemistry section at Umeå University (Paper I). Secondly, waste was collected in a 19-month collection campaign by a refuse collection lorry from 200-290 households in villages and residential areas in the municipality of Bjurholm (Paper II), which resulted in more than 80 tonnes material. The contents of the wastes were characterized by sorting analysis. The waste was shredded and mixed with reed canary-grass in different proportions in order to evaluate the effect of such mixtures on emissions from waste.

3.2.2 Biofuels Most fuels used were various sorts of biofuels. The source-sorted waste consisted of approximately 85% biofuel, and the fuels used to supply the balance in co-combustion trials, or as single fuels were; reed canary grass (Papers I and II) pellets of wood and wheat-straw, and firewood from birch, pine and spruce (Paper III). In the experiments reported in Paper IV, various biomass materials were burned in uncontrolled outdoor experiments (backyard burnings), including garden waste (branches and leaves), paper and straw. Table 7 shows the contents of key chemical constituents and characteristics of the fuels used in the different experiments. Table 7. Characteristics of some of the fuels used. Values from Papers I-IV and the literature [123]

Fuel Cal.heat (MJ/kg)

Moisture (%)

Ash (%)

Cl (%)

C (%)

RCG 17.5 8.5 11.6 0.04 46 RDF 17.9 6-12 11-17 0.13-0.52 46-48 Wood pellets 6 0.3 <0.01 51.9 Straw pellets 8.1 5.9 0.19 46.2 Wood 16.2-16.9 0.16 0.15 51.9 Garden waste, leaves grass 32-58 5 0.4 46 Garden waste, branches 14 1 0.02 52

Reed canary grass was chosen as the counterpart in co-combustion trials with waste partly because it performed well in earlier (unpublished) experiments compared to other alternatives, with a low tendency to cause ash sintering, and partly to find out if its alkaline content would promote bonding of the chlorine in the ash.

18

3.2.3 Combustion equipments The common factor in the combustion experiments was that they were performed at a scale that reduced the scope to control emissions with cleaning systems at reasonable costs, compared to large-scale combustions of biofuel and MSW for instance. This shifted the focus to the contents of the waste and combustion performance. The combustion equipment used in these experiments also covered a wide range of several important parameters, including: the scale of fuel consumption, energy recovery and the scope to control emissions. The combustion systems examined encompassed back-yard burning, where few things can be controlled once the fire is started, stoves (pellet- and wood-fuelled) and boilers where the air supply (draft) can usually be regulated, and lab-scale reactor and biofuel boilers (Papers I and II) with various monitoring systems that can help to optimize the supply of fuel, primary air and secondary air. In the preliminary experiments reported in Paper I, the combustions were performed in a laboratory-scale fluidized bed reactor, offering substantial scope to control supplies of fuel and air and to control temperature with preheated air and heating elements [23,124]. The combustion was based on the bubbling fluidized bed principle, in which effective combustion is promoted by a good air supply and the contact of fuel with hot bubbling sand. The Ökotherm 150 kW and 600 kW biofuel boilers used for most of the combustion tests described in Papers I and II are different scale versions of the same construction and design (Figure I in Paper I). These boilers are constructed for the combustion of ash-rich fuels and are supplied with equipment for removing large quantities of ash; an ash stoker and a slag scraper remove ash from the primary combustion chamber and an ash screw feeds ash out from the ash chamber. In the wood boilers and wood stove used in the studies described in Paper III combustion also occurs on grates with underlying ash chambers. A major difference between them is in the route taken by the primary air through the fuel. Whereas the air in the old boiler and the stove is fed from under the fuel bed, leading to so-called up-draught combustion, the modern boiler exploits the down-draught combustion principle by feeding air from above. While most of the appliances have some kind of fan to maintain a controllable supply of air, the old wood boiler and wood stove are regulated through self draught. The down-draught combustion in the modern boiler generally requires a fan to ensure that enough air is forced downwards through the fuel bed [125]. The pellet burner used in the study described in Paper III is of a type where the pellets are fed from underneath into a burner cup. The ash formed is pushed over the edge of the cup down to the ash chamber, but some ash will remain in the cup and may affect the combustion. The rate at which generated heat is removed influences the rate at which fuel can be supplied and combusted. The laboratory reactor is constructed to allow advanced

19

control of heat generation by adjusting the supplies of fuel and air. The heat formed in batch-wise fuel feeding, as in the wood boiler experiments, was transported away in some of the experiments through heat exchangers, while in other experiments heat generation was regulated instead by adjusting the air supply. The biofuel boilers used in the studies reported in Papers I and II, and the pellet boiler in Paper III, allowed both the fuel supply and heat removal to be regulated. In the case of the pellet burner the fuel supply was intermittently regulated. The backyard burning experiments in barrels and as open fires are special cases, as the simplest possible types of combustion. The biggest difference between them is that the barrel retains more of the generated heat, thus promoting higher combustion temperatures. Both share the disadvantage that ash is not removed and will eventually inhibit combustion.

3.3 Methods

3.3.1 Sorting analysis Randomly selected waste sacks were sorted into different categories. In Paper I, where the sorting was done mainly according to material content, there were eight categories suitable for combustion and two that were unsuitable. In Paper II, because it also was considered of interest to receive information of the fraction size of different recyclable categories, a plastic packaging category was introduced, thus making it nine combustible categories. A greater diversity of material in the larger study and the introduction of a category for hazardous waste made the categories unsuitable for combustion to finally reach a number of five. The sampling size in the major sorting analysis in Paper II was initially set at 10%, or more, of the number of sacks received, which initially usually ranged between 80 and 120, and after an extension of the collection area averaged around 160. In the subsequent sorting, 20 sacks were characterized in order to check for when a stable distribution between the different sorting fractions were obtained. Fifteen were found to be a sufficient number, and this was the standard sampling size in the following waste sorting exercises. The selection of sacks was purely random on the first two occasions, when all sampling was done in the countryside, but stratified random sampling was applied to the following samplings since collection was extended to the centre of the municipality, where the waste generated could have been significantly different (due to differences in lifestyle). The two categories (countryside and municipal centre) was evaluated with a t-test, but as no difference was detected, they were thereafter treated as a single category. Analysis of variance (ANOVA) was used to evaluate differences between all sorting analyses, but here too no significant differences were detected.

3.3.2 Logging equipment Logging measured variables, such as temperatures and emissions, has several advantages in addition to facilitating a general survey of combustion performance,

20

allowing factors that may influence levels of dioxin emissions to be detected and evaluated, such as differences between combustion phases, disturbances and other events. The extent of temperature logging depended on the equipment used, and different positions were monitored in different experiments. In addition, emissions were monitored by various instruments exploiting diverse measuring principles, including:

Electrochemical detection: an ECOM-SG-Plus selective gas analyzer for analysis of O2, SO2, NO and CO levels > 1000 ppm in combustion tests in the laboratory reactor (Paper I).

IR detection: a Siemens Ultramat 22P for analysis of CO2 and CO at levels < 1000 ppm in combustions in the laboratory reactor (Paper I) and Maihak Finor and Maihak Unor 600 instruments for analysis of CO2, CO and SO2 in combustion tests in the 150 kW and 600 kW boilers (Papers 1 and II).

Chemoluminescence: an ECO Physics CL D 700 EL for analysis of NO, NO2 and NOx in combustions in the 150 kW and 600 kW boilers (Papers 1 and II)

Paramagnetic analysis: An M&C Products MA 25 for O2 analysis in combustion tests in the 150 kW and 600 kW boilers (Papers 1 and II)

FTIR: A Bomem 9100 FTIR – spectrometer for analysis of CH4, CO2, CO, NO2, NO, N2O, HCl, HF, HBr and SO2 in combustion tests in the 150 kW and 600 kW boilers (Paper II) and in pellet and wood combustions (Paper III).

In addition a Multi Gas Analyzer (Electra control) was used in combustions of wood (Paper III) and backyard burning (Paper IV) tests. This instrument incorporates both electrochemical cells (O2) and IR (CO2 and CO) systems.

All of these gas detectors except the FITR spectrometer generally have to be checked or calibrated using gas with known concentrations, a routine which also was performed.

3.3.3 Sampling flue-gas Organic pollutants were sampled with a sampling train, as illustrated in Figure 2. The sampled flue-gas is cooled quickly in the probe to avoid further reactions. It is then led to an ice-water cooled flask where the condensate and most particles are collected, then to an impinger flask with ethylene glycol where mainly non polar compounds are collected and to a polyurethane foam adsorbent and filter where aerosols and finer particles are collected. The sampling was basically performed according to standard method EN 1948:1 [126], but sometimes the sampling volumes were lower than those prescribed, due to limitations in the fuel supply or other constraints. The sampling was performed isokinetically, which means that the velocity of the flow in the probe was kept equal to the velocity in the flue gas duct, in order to sample a representative amount of particles. Thus, the sampling flow had to be adjusted to the measured flow in the flue gas duct. In most experiments fuel was fed steadily, there was a fairly constant combustion rate and fans helped keep the flow constant. In the wood combustions the application of a dilution channel [127] facilitated a constant flow.

21

Figure 2. Cooled probe sampling train.

3.3.4 Analysis PCDD/F, PCB, PCPh, PCBz and PAH in flue-gas and ash were identified and quantified by isotope dilution methods, in which known amounts of appropriate isotope-labelled surrogate compounds were added to the sample before extraction and analysis. Ideally, a surrogate compound should be chemically identical to the analyte, but contain an isotope that is not naturally abundant (e.g. C13, Cl37 or H2), resulting in a difference in mass between the analyte and the surrogate that allows them to be distinguished. Losses of the surrogate during extraction and clean-up can be accurately quantified. Thus, losses of the analyte can be estimated with a high degree of certainty since their proportional losses should be virtually identical due to their chemical similarity.’ The general procedure used, in which selective clean-up after extraction enables analysis of several classes of compounds from a single sample is described in Liljelind et. al [128] and is based on a European standard method [126]. A few of the clean-up steps were modified between the first and last of the studies. The emission samples from combustions of pellets and firewood, presented in Paper III, were analysed by the Institute of Public Health, Ostrava, in the Czech Republic according to European standard methods [126]. The extraction methods used for analysing combustion-related samples could easily be applied to the fuels, waste and most sorted waste fractions analysed in Paper II because most sorted materials consist of porous, fibrous or fine-grained materials

Cooling- water

Ice

Water Organic solvent

Air-pump

Air- volume

Filter

Absorbent

22

which allow the extraction solvent to come into contact with the analytes without dissolving the matrix. The plastic waste, however, included some unknown polymers with varying solubility properties; and extraction with organic solvents may have dissolved some polymers but left others unaffected. In addition, the dissolved polymers may have precipitated in subsequent clean-up steps and encapsulated previously dissolved analytes. Published methods for analysing additives in plastics often involve repeated cycles of dissolution and precipitation with solvents suitable for specific polymers [129]. This is an approach that also has been used for the analysis of dioxins in PVC [130]. However, since the sorted plastic waste fractions in our studies were mixed, and to a large extent consisted of unknown polymers, this method would not have been applicable. The approach adopted to overcome these problems was to include serial extractions of solvents with different solubility characteristics, followed by dialysis in polyethylene tubes (Paper II). To characterize solubility the “Hildebrand solubility parameter, δ” is often used, which indicates the relative solvency behaviour of specific solvents and polymers. This solubility parameter (δ, MPa1/2) is defined as the square root of the internal energy of vaporization divided by the molar volume of a substance [131]. As a general rule, the closer the values for a specific solvent and polymer are, the more of the polymer will dissolve in a given volume of the solvent. Three solvents were selected for the serial extractions (n-hexane, toluene and terahydrofuran, THF), following consideration of the solubility parameters of polymers that are likely to be common or environmentally significant (Paper II), together with reported experiences of using solvents in practice [129,130] and the results of solubility tests on the plastic waste fractions.’ 3.3.5 Multivariate data analysis Multivariate data analysis is a common term for a number of methods used to analyse patterns and to evaluate data matrixes with many variables statistically [132]. In Principal Component Analysis (PCA), every observation is represented by a point in a multidimensional space with as many co-ordinates as the number of variables. These points are then projected onto planes of lower dimensions, giving orthogonal principal components (PC´s) describing most of the variation in the dataset and are visualized in score-plots. The first PC, PC1, describes the largest contribution to the variation in the data and so on. The corresponding loading-plots show the variables that have the strongest influence on the distribution of the observations. Projection to latent structures (PLS) is a regression technique in which multidimensional spaces of x-variables and y-variables are fitted to each other in order to predict y-variables from the x-variables. In Papers III and IV, PCA was used to distinguish between profiles of dioxin homologues emitted in different tests, and in this thesis it is used to assess the differences in emission levels and homologues from all experiments performed and their relationships to fuel compositions and combustion conditions. PLS was used in Paper IV in an attempt to predict emission levels of dioxins from the combustion temperature and the composition of the combusted waste.

23

4 Results and discussion

4.1 Sorting analysis of source-sorted waste (Papers I and II) The waste sorting analyses were done to establish the composition of the fuel used, which is essential for identifying factors that influence emission levels. The results do not provide a sound basis for general conclusions about waste composition, due to the limited collection area involved. However, comparison of the results with data obtained from larger waste sorting analyses can give indications regarding the likely composition of combustible fractions sorted on a more general scale. In paper II, similarities and differences between these and previous analyses are discussed

4.1.1 Chlorine and PCDD/F in waste fractions and fuel Figure 3 and 5 show how chlorine and PCDD/F are distributed in the waste fractions. As can be seen in the figures, the majority of the chlorine is derived from the remaining plastic fraction, which constitutes a small proportion of the waste (Figure 4)Figure 3. The importance of this fraction can be explained by the fact that it may contain PVC, which usually has greater than 50% chlorine content. Most dioxin is found in the textile fraction (Figure 5) and to give a reasonable explanation for this, we have to take a closer look at differences in levels between dioxins with different degrees of chlorination, i.e. homologues.

Plastic

Textile Other Paper

Figure 3. Distribution of chlorine in waste fractions

24

Sorting analysis A, WHO-TEQ

Textile 74%

Paper20%

Other3% Plastic

3%

Sorting analysis E, WHO-TEQ

Plastic5%

Paper2%

Other0%

Textile 93%

Sorting analysis A, ∑PCDD/F

Textile 91%

Plastic1%

Paper8%

Other0%

Sorting analysis E, ∑PCDD/F

Textile 95%

Paper0% Plastic

4%Other1%

Figure 4. Distribution of dioxins in waste fractions from two sorting analyses. In Figure 6 the homologue profiles of all sorted waste fractions, fuel and fuel constituents are presented. They are also presented as PCA-plots in Figure 7, while Table 8 shows the labels used in the plots. Only the first component of the PCA is significant, the second component has been included in the figures mainly for the sake of clarity. The figures show profiles that can be mainly divided into two characteristic groups, one where the OCDD is the dominant homologue, and one with a more even distribution of homologues. In the score-plot of Figure 7, we can also see a couple of samples that seem to be more strongly influenced by PeCDF than the others; one sample of uncoated paper and one of remaining combustibles. They are however differentiated along the insignificant second component. The samples with an OCDD-dominated profile, also generally have higher levels of PCDD/F. The difference of profiles (and levels) is cut across categories, which shows that the profiles do not necessarily or generally show a main characteristic associated with the manufacture of the majority of products within a category. Instead the profiles may show 1) the characteristic profile of sub categories, products that is not represented in all samples, or 2) the effect of contamination affecting a minor group of the sample. The first suggestion may well be case in one ore two of the categories, but it seems to be more reasonable with common mechanism that can explain the same difference in several categories

25

Coated paper

0,00

0,10

0,20

0,30

0,40

0,50

0,60

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

s.o. As.o. E

Uncoated paper

0,000,010,020,030,040,050,060,070,08

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

s.o. As.o. E

Paper packages

0,00

0,05

0,10

0,15

0,20

0,25

0,30

0,35

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

s.o. As.o. E

Soft plastics

0,000,020,040,060,080,100,120,140,160,180,20

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

s.o. As.o. E

Hard plastics

0,00

0,02

0,04

0,06

0,08

0,10

0,12

0,14

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

s.o. As.o. E

Rem.plastics

0,00

5,00

10,00

15,00

20,00

25,00

30,00

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

s.o. As.o. E

Textile

0,00

20,00

40,00

60,00

80,00

100,00

120,00

140,00

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g s.o. A

s.o. Bs.o. E

Wood

0,00

0,50

1,00

1,50

2,00

2,50

3,00

3,50

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

s.o. As.o. E

Rem. Comb.

0,00

1,00

2,00

3,00

4,00

5,00

6,00

7,00

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

s.o. As.o. E

Reed canary-grass

0,0000,0010,0020,0030,0040,0050,0060,0070,0080,0090,010

4 5 6 7 8 4 5 6 7 8

PCDF PCDD

ng/g

Figure 5. Profiles of dioxin homologues in waste fractions and RCG.

26

-0,30

-0,20

-0,10

0,00

0,10

0,20

0,30

0,40

0,50

-0,90-0,80-0,70-0,60-0,50-0,40-0,30-0,20-0,100,00 0,10 0,20 0,30 0,40 0,50 0,60 0,70 0,80 0,90

t[2]

t[1]

UPaA

CPaAPPaAHPlAaRPlAa

WA

TA1

RCoA

RCG1UPaE

CPaEPPaE

HPlE

SPlERPlE

WETERCoEFuel1RDF1RDF2

HPlAb

SPlAb

RPlAbTAbTBFuel2

RCGb TEb

Figure 6. Score plot (top panel) and loading plot (bottom panel) of relative levels of dioxin homologues in waste fractions and fuels. The first component describes 79 % of the distribution and the second (non-significant) component describes another 10 %. Designations according to Table 8.

- 0 . 20

0 . 00

0 . 20

0 . 40

0 . 60

0 . 80

- 0 . 30 - 0. 20 - 0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90

p[2]

p[1]

OCDD

OCDF

HpCDDHpCDF

HxCDD

HxCDF

PeCDD

PeCDF

TCDD TCDF

TEQ

27

Table 8. Designations used in the score plot, Figure 7. The beginning of the designation, upper- and lower-case letters, refers to the waste fraction or type of fuel as in the upper part of the table. The last upper-case letter, A and E refers, to respective sorting analysis (A or E) and the lower-case letters, a and b, refers to duplicate analyses of the same sample.

First part of designation

UP Uncoated paper CP Coated paper PP Paper packages SPl Soft plastics HPl Hard plastics RPl Remaining plastics W Wood T Textile RCo Remaining combustible RCG Reed canary-grass RDF Refuse-derived fuel (shredded waste) Fuel1 Mixture of 30% RDF and RCG, comb trial 2 Fuel2 Mixture of 30% RDF and RCG, comb trial 3 Last part of designation

A From sorting analysis A E From sorting analysis E a,b Duplicate analyses of the same sample

In Paper II possible sources of contamination are discussed. To summarize the discussion, the OCDD-dominated profiles may derive from deposition of dioxins from ambient surroundings [55,56]. The dioxins originate from different sources of which different combustion sources are considered to be the most important. When newly produced, the dioxins often show profiles with a fairly even distribution or domination by low chlorinated homologues [50,54,59] but OCDD is the most common dioxin in the environment and it usually characterizes environmental sinks. The dioxins are deposited on dust, and similar dioxin profiles have been found in air filters and vacuum cleaner bags [55]. The ‘remaining combustible’ fraction, which is likely to contain this sort of material, e.g. vacuum cleaner bags, filters, sweepings etc., had relatively high levels of dioxins with this characteristic profile. Although textiles are likely to have a high capacity to collect dust and adsorb pollutants, the dioxin levels in the textile samples were too high relative to the others (by 1 to 4 orders of magnitudes) to make this their probable main source. In Paper II it is argued that their high levels mainly originate from features of textile and leather production processes, and that the sources may be pentachlorophenol (PCP), a fungicide used in cotton and leather production [133-

28

135]or textile dying agents [136]. High levels have been found in samples of new clothing [133,137] effluents from the textile industry [138] and textile waste [91]. The ‘other plastics’ fraction also showed quite high levels of dioxins. PVC is a probable constituent in this fraction, but the reported levels of dioxin in PVC usually are low [130,139].

4.1.2 Flow of dioxins, Recycle or incinerate – contaminated waste? One aspect of recycling that has been little studied is that some products may collect pollutants from the environment during their lifetime. Specific issues that need to be addressed include identification of products that are especially exposed, acceptable levels of pollutants and the optimal number of recycling rounds. For some materials other sources of contamination may be important, as illustrated by the dioxin levels in textiles. The risk for human exposure in most cases is probably insignificant compared to other sources, e.g. food. But for some products, e.g. baby clothes and toys made of textile or plastics, there may be a case for identifying and maybe regulating the levels because of the habit of small children of putting things in their mouths.

4.2 Co-combustion of biofuel and RDF (Papers I and II) No major hygienic problems were experiences in the handling of the source sorted waste and no or major technical problems were encountered in the fuel preparation and co-combustion with biofuel. Combustions were both efficient and generated low emissions of carbon monoxide. The mass of dioxins (∑PCDD/F or WHO-TEQ) in the flue-gas emissions were generally lower than the mass in the fuel input. Comparison of the dioxin input and the total output levels; i.e. the summed dioxins contents in ashes and flue-gas, reveals that that the combustion resulted mostly in increases of both toxicity equivalents and total dioxin contents but also in a decrease in one experiment (30% RDF in the 150 kW-reactor). However levels of hydrogen chloride were an order of magnitude too high compared to regulated limits, although levels of equal magnitude were also generated from combustion of the pure biofuel. It was also difficult to meet the 0.1 ng (WHO-TEQ)/Nm3 limit [87] for emissions of dioxins in the co-combustions of waste and bio-fuel. Further developments of fuel-mixes are possible. Introducing more components in the fuel mixture would increase the possibilities for optimization of qualities by means of experimental design. A Cl2 suppressing component may contribute to a reduction in dioxin emissions. Co-combustion of RDF with sulphur-containing peat has shown such effects [41,140]. There are also possibilities to for a further development of boiler technology with respect to dioxin formation; e.g. with air-staging and reduction of residence time for flue gas in critical temperature zones. However, some additional cleaning probably has to be added to meet the regulated limit, and if the small-scale approach is to be maintained the additional processes cannot be too expensive. As emissions from MWI today are generally well below the limit value, an extensive changeover to small-scale co-combustion systems risks increasing total emissions. Thus, other environmental benefits must be gained to

29

justify such a change. In the regions where this has primarily been suggested (sparsely populated areas in northern Sweden; [108]), the replacement of old oil-boilers has provided one motive for changing the systems. Reductions in transportation, such as trips in private cars to recycling stations and perhaps collections rounds by a refuse lorry around rural areas with a scattered population, may also be beneficial. A locally based waste management and energy generation system may increase the motivation to avoid back-yard burning. The phasing out of polluted material may be a fourth factor, although this measure should be applied on a larger scale, if it is considered necessary. The choice of systems to use demands careful consideration and probably with the assistance of system analysis tools, like LCA. The small-scale co-combustion model of waste management and energy production may have potential in regions with two main characteristics. Firstly, it may be a viable alternative in sparsely populated areas where there are functioning alternative waste management systems, but long-distance transport may be a problem, e.g. parts of Scandinavia and North America. Secondly, it may be adjusted for use in areas in developing countries with poorly developed waste management and recycling infrastructure [141]. Such areas often have problems with increasing amounts of waste not that are not adequately treated, especially plastics, furthermore, in the same areas there can problems associated with a lack of energy supply (e.g. electricity). The approach of co-combusting waste with a suitable locally available biofuel may have an advantage in this context, if it can be done in an environmentally acceptable way. It is of course important not to introduce new environmental problems associated with emissions and ash disposal.

4.3 Combustion of pellets and fire-wood in residential stoves and boilers (Paper III) Intermittent combustion of wood pellets resulted in emissions of 28 ng (WHO-TEQ)/kg combusted fuel (dry weight). Emissions of the investigated chlorinated organic compounds in the wood boilers were considerably higher with reduced air supply than with full air supply. A modern environmentally certified boiler yielded considerably lower emissions than an old boiler. Chlorine-containing plastic waste gave rise to high emissions; 3 µg (WHO-TEQ)/kg for the whole combustion cycle. Besides the high emissions of dioxins obtained in combustion of plastic waste in a wood boiler, which can be related to fuel content, high emissions also resulted from smouldering combustion and intermittent combustion of fuels containing low levels of chlorine. This shows the importance of combustion efficiency, further emphasized by the low emissions obtained from the combustions with full air supply in the modern environmentally certified boiler. The differences in emissions of PCDD/F, PCB and HCB between boilers and operational routines are fully comparable with those obtained from emissions of CO, particles, PAH, CH4 and other volatile organic compounds in a similar experimental setup [109]. Thus, the positive effect on unwanted emissions generally attributed to combustion in boilers connected to heat

30

storage tanks also seems to be valid for chlorinated organic compounds. The high emissions from the intermittent pellet combustion are phenomena that should be further examined. As pointed out in Paper III, the rises and falls of temperature during intermittent operational cycles will provide temperature conditions that favour dioxin formation. However, other reported trials on pellet combustion [44,113,115,142] have not found such high levels of dioxin emission, and their possible dependence on the burners used and other factors should be examined. It is conceivable that differences in the designs, materials used and other variables such as the draught dynamics of pellet burners may lead to major differences in the variables affecting dioxin formation. Burners are often designed for use in specific boilers, and the use of burners in other boilers may cause especially unfavourable conditions.

4.4 Backyard burning (Paper IV) Emission factors of dioxins and PCB from the backyard burning ranged from 2.2 to 13000 ng(WHO-TEQ)/kg. The levels found in ash were usually less than 5% of the total. Apart from when burning waste with extreme contents, such as computer waste and PVC, the emissions of PCDD/F and PCB could not be related with any certainty to waste fuel content. The other possible waste emission correlations were masked by factors related to combustion performance. In Table 9, selected data from Paper IV and literature are compiled for comparison. The general tendencies from the reported experiments, with broad ranges of emission factors from most wastes, and exceptionally high levels from PVC-containing wastes, are confirmed in Paper IV. The uncertainty of the typicality of the experimental waste compositions and the large variety of reported levels makes it difficult to establish a single valid emission factor. In Paper IV, a rather broad range emission factor, 4-72 ng (WHO-TEQ)/kg, is proposed for future estimates of emissions from moderately contaminated waste. These figures imply that combusting waste in the backyard could contribute substantially to total emissions, even if the amounts of fuel involved are equivalent to just a few tenths of a percent of the amounts combusted in municipal waste incinerators. The emission factor could be estimated with greater precision by increasing the number of experiments. Furthermore, this approach of coupling emission factors to sub-groups of wastes could be extended to cover all cases of backyard burnings. Estimations of total emissions of chlorinated organic pollutants on a national basis based on one emission factor related to the mass of fuel, chlorine content or some other expectedly relevant basis may be one applicable method, but it may also be difficult because emissions levels are probably related to combinations of waste constituents as well as amounts, e.g., due to saturation of catalytic sites, one piece of PVC may cause more emissions of PCDD/F if divided between several fires than if it is burned in a single fire. A more convenient way to estimate emission levels may be to sum likely emissions, based on ‘model fire’ emission factors. This would require the following three steps.

31

1) Obtaining a better understanding of materials that are burnt in backyards through enquiries or observations. 2) Establishing a limited number of ‘model-fires’ based on this information. This could be done with a simple approach, e.g. with three or four levels with large degrees of randomness within them; e.g. lightly contaminated (just garden waste), moderately contaminated (garden waste, RDF, paper and packages) and highly contaminated (moderately contaminated + PVC and metals) waste, or a more sophisticated approach in which the model fires are based on an experimental design (high-low chlorine, high-low metal etc). In both cases, each model fire, with some degree of randomness, should be repeated to establish means with reasonably high significance representing the specific model. 3) Assessing the extent of fires corresponding to each model fire performed in reality at local, regional or national levels, again based on enquires and observations. This would still be the most problematic step, involving many sources of errors, but with the mean levels of model fires well established this approach would be more open to future adjustments. Table 9. Emission factors of dioxins (ng TEQ/kg fuel) in emissions from backyard burnings in Paper IV and reported in literature.

References Waste fuel

IV [118] [119] [122] [120] [117,121] [121]

Garden waste, tree, leaves

12-100 4.6 5-140 4-20

Paper 6-420 Paper+ polyethylene or polystyrene

7-13 3-40

PVC 130001) 6554 1670-280002) Scrap tire 203) 221 Electronic scrap (incl. PVC)

930 4) 10335)

Straw 4-10 20 Household waste 16-65 6) 35 14 2-28 Household waste + 0,2%PVC

759-903 79 9-139

Household waste + 1%PVC

201 179-242

Household waste + 4,5%PVC

1230-5400

Household waste + 7,5%PVC

4916 3543-6655

Houshold .waste + 7,5% inorg. Cl

734

1) Incl. garden waste, 2) Incl. paper, 3) Incl. RDF 4) Computer scrap + RDF+ paper pack. 5) Electric wire tube, 6)

RDF+ garden waste or RDF +paper

32

4.5 Contents of fuels and their significance for combustion emissions (Papers I-IV) One aim of the co-combustion project was to evaluate the scope to reduce emissions of POPs by improving waste-fuel quality, with respect to its contents of energy and environmentally undesirable compounds. In combustions of pellets and wood the quality of fuel also plays an essential role. Pellets are commercially available in various specified qualities. In residential wood combustion the qualities of different wood species varies, but in practice the moisture content of the fuel may be the variable that the common user has the most ability to influence. In backyard burning, the composition of the burned material is the only variable that can be influenced to a significant degree.

4.5.1 The significance of chlorine contents for dioxin emissions When all values of PCDD/F emissions from all experiments are plotted versus the levels of chlorine (Figure 8) there seems to be some correlation. The correlation factor of the logarithmic values is 0.7, but this factor drops dramatically if one or two of the sample fractions with the highest chlorine levels (those including PVC and computer scrap from backyard burning) are excluded. On the other hand, if the highest values in the low chlorine end of the point cluster (derived from intermittent pellet combustion) in Figure 8 are regarded as outliers and excluded, the correlation factors will be restored to similar levels. A tendency may be distinguished that there is an increase in PCDD/F emissions beginning from a chlorine level just below 0.1 %. This can be compared with previously reported ‘threshold values’ of approximately 0.1 % [43] and 1 % [44] for correlation between chlorine levels in fuel and dioxin emission. There is however a large spread of values in the y-direction that makes predicting dioxin emissions from chlorine contents uncertain. A conclusion that can be drawn with some degree of certainty is that combustion of fuels with very high levels of chlorine usually results in higher emissions of dioxins than combustion of fuels with very low levels of chlorine. The values from backyard barrel experiments alone also show a similar correlation, entirely due to the two most contaminated fuels.

33

Figure 7. Dioxins in emissions from experiments described in Papers I-IV. In the data from the co-combustion experiments (Papers I and II), a correlation between fuel chlorine and logarithmic PCDD/F emission values can be detected that is weak (correlation coefficients 0.5 and 0.7 for ∑PCDD/F and WHO-TEQ, respectively), but more robust in the sense that it does not change much if values are withdrawn. This may be because these experiments were designed to test their relationship. The combustion conditions were comparably controlled, in similar facilities (except for the laboratory reactor tests of Paper I) and similar fuels were used, adjusted to contain different levels of chlorine. In contrast, in the pellet and wood combustion experiments reported in Paper III there were small differences in chlorine levels (with some exceptions) but large differences in combustion conditions. In these experiments no correlation between fuel chlorine contents and PCDD/F emissions were found, but between combustion efficiency and logarithms of emission factors there was a correlation factor of -0.7 (Figure 9).

PCDD/F emissions vs chlorine in fuel in all experiments

0.01

0.1

1

10

100

1000

10000

100000

0.001 0.01 0.1 1 10 100 Cl, %

ng/N

m3

∑ PCDD/F

WHO-TEQ

34

Figure 8. Emission factors of dioxins in wood and pellet combustions (Paper III) in relation to combustion efficiency. Using PCA, the difference between the two experimental setups can be shown as plots with other variables included (Figure 10). In the score plot the observations from co-combustions (labelled L and B, according to Table 10) are shifted down and to the left, towards the fuel-related variables (in the loading plot) ash, Cl, and HCl (which is largely dependent on the chlorine level in the fuel). The observations from the pellet and wood combustions (labelled P and W, Table 10) are spread in a left-right direction and are mainly influenced by the relationships between the combustion-related variables CO2, O2 and CO. The observation from the smouldering combustions of plastic waste and wood (W8) are strongly influenced by both combustion-related and fuel-related variables, and since their dioxin emission levels were higher than those of any of the other combustions, this observation is located, in isolation, in the lower right corner of the score plot.

EF vs CE in wood and pellet boilers and stoves

1

10

100

1000

10000

100000

1000000

10000000

0.00 0.20 0.40 0.60 0,80 1.00 1.20

CE

EF,

ng/k

g fu

el

∑ PCDD/F

WHO-TEQ

35

-8

-6

-4

-2

0

2

4

-6 -5 -4 -3 -2 -1 0 1 2 3 4 5 6 7 8 9

t[2]

t[1]

L1L2L3

B1B2

B3B4B5B6B7B8

B9B10

B11B12B13B14B15aB15b

B16aB16b

B17aB17bB18aB18b

B19aB19b

B20aB20bB21aB21b

P1P2

P3

W1

W2

W3

W4

W5

W6

W7

W8

W9W10

W11W12W13

W14W15

Figure 10. Score and loading plots of fuel-related and emission-related variables for co-combustions of waste and biofuel (Papers I and II) and combustions of pellets and wood (Paper III). 48 % of the distribution is described in the first component and 25 % in the second. Both components are significant. Designation of scores according to Table 10.

-0 .40

-0 .30

-0 .20

-0 .10

0 .00

0 .10

-0 .40 -0..30 -0.20 -0.10 0.00 0.10 0.20 0.3 0

p[2]

p[1]

TEQ SumPCDDF

RatioFD

Ash

Cl

O2

CO2

CO

NO

NO2 SO2

HCl

Table 10. List of the combustion experiments performed and the designations used in the scores-plots. See respective papers for further information.

Paper I

Design. Paper II Design. Paper III Design. Paper IV Design.

Lab reactor 150 kW-boiler Pellets Backyard 0% RDF L1 1, 43% RDF B15 1.1 P1 Garden waste Y1 20% RDF L2 2a, 0%RDF B16 1.2 P2 Garden waste Y2 40% RDF L3 2b, 30% RDF B17 1.3 P3 Garden waste Y3 150kW Run1 2c, 30% RDF B18 Wood, old boiler Garden waste, Paper pack, Plastic pack. Y4 0% RDF B1 600 kW-boiler 2.1a W1 Garden waste, Plastic pack. Y5 0% RDF B2 3a, kW, 0% RDF B19 2.1b W2 Garden waste, RDF Y6 21% RDF B3 3b, 30% RDF B20 2.2a W3 Garden waste, RDF Y7 21% RDF B4 3c, 30% RDF B21 2.2b W4 Garden waste, Motor oil Y8 36% RDF B5 2.3a W5 Garden waste, Silage film Y9 36% RDF B6 2.3b W6 Garden waste, PVC Y10 150kW Run2 2.4a W7 Papers, Paper pack, Plastic pack. Y11 100%RCG B7 2.4.b W8 Papers, Paper pack., Plastic pack. Y12 100%RCG B8 Wood, modern boiler RDF, Paper pack. Y13 36%RDF-b.c B9 3.1 W9 RDF, Paper pack. Y14 36%RDF-a.c B10 3.2a W10 RDF, Paper pack. Car tyre , Motor oil Y15 36%RDF-b.c B11 3.2b W11 RDF, Paper pack. Computer scrap Y16 36%RDF-a.c B12 3.3a W12 Straw Y17 41% RDF B13 3.3b W13 Straw Y18 41% RDF B14 Wood, stove Straw, Silage film Y19 4.1 W14 Garden waste, open fire Y20 4.2 W15 Garden waste, RDF, open fire Y21

4.5.2 Information provided from homologue profiles Figure 9 shows score and loading plots of the homologue levels, related to ∑PCDD/F. Only the first component is significant but the second has component has been included for sake of clarity and because it still seems to reveal some information. As can be seen from the loading plot, the relative abundance of TCDF dominates the distribution, spreading out the observations from left to right along the first component. The observations to the left will have a more even homologue profile than the observations to the right. Along the second component there is a distribution in which the remaining low chlorinated PCDFs appear at the top and TCDD at the bottom. The observations are to a large extent grouped according to the experiments and the different combustion facilities used; with observations from the laboratory reactor (L1-L3), biofuel boilers (B1-B14), wood and pellet boilers (W1-W15, P1-P3) and backyard burnings (Y1-Y21) distributed in the cited order from the upper left to the lower right in the score plot, except for the backyard burning with PVC (Y10), which is located at the top. This distribution seems to be influenced by the relation between the PCDF levels; mainly between PeCDF and TCDF. The distribution in the score-plot seems to be related to the degree of control of the combustion, with the laboratory reactor and the biofuel boilers offering relatively stable conditions at one end (except for Y10), the uncontrolled backyard burnings at the other, and the two open burnings (without a barrel) at the very bottom. The temperature in combustion and post-combustion zones are parameters that might differ distinctly between the combustion facilities, for several of which (e.g. the wood boilers) temperatures in the combustion zones were not measured in a way that allows strict comparisons. Bearing this shortcoming in mind, the available information still suggest that the temperatures follow a similar pattern; with temperatures higher than 800 °C occurring in the post-combustion zone of the laboratory reactor, the same or somewhat lower temperatures in the biofuel boilers, and measured temperatures occasionally reaching 600-700° in the barrels, but varying substantially. The temperature of the fire in the open combustion was not measured, but the temperature of the flue-gas in the hood above it never reached 150 °C. The amount and quality of particles generated also presumably differed between the experiments. In the biofuel boilers, where the bulk of particles presumably are fly ash, levels were measured but not in the others. Compared to combustion of RDF and RCG, wood combustion does not produce much fly ash, but in the smouldering phase, and especially in the old wood boiler, considerable amounts of soot were produced. The backyard burnings produced a lot of unburned particles, both ash and soot. These parameters, which are associated with combustion performance and important in dioxin formation, may contribute to the differences amongst homologue profiles.

38

Figure 9. Score and loading plots of the relative abundance of different homologue groups in the dioxin profiles in emissions from combustions described in Papers I-IV. The first component describes 65 % of the distribution and the second (non-significant) component describes another 16 %. Designations of scores according to Table 10.

- 0 . 40

- 0 . 20

0 . 00

0 . 20

0 . 40

0 . 60

- 0 . 2 0 - 0 . 10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90

p[2]

p[1]

TCDF

PeCDF

HxCDF

HpCDF

OCDF

TCDD

PeCDD HxCDD HpCDDOCDD

- 0.20

- 0.10

0.00

0.10

0.20

- 0 . 4 0 - 0.30 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40

t[2]

t[1]

L1

L2L3

L4B1

B2

B3

B4 B5B6

B7B8B9B10B11

B12

B13

B14

B15aB15b

B16a

B16b B17aB17b

B18aB18b

B19a

B19b

B20aB20b

B21aB21b

P1P2

P3

W1W2 W3

W4

W5

W6

W7W8

W9 W10 W11W12

W13 W14W15

Y1Y6

Y4Y13Y3 Y12 Y14Y11

Y18Y17

Y19Y8Y15

Y16

Y20

Y21

Y2Y7

Y5Y9

Y10

39

Within each group observations are spread from lower left to upper right in the score plot, mainly reflecting relative contents of dioxins and TCDF, i.e. since TCDF is the dominating PCDF, the distribution also reflects the PCDF/PCDD-ratio. This distribution seems to be attributable to different factors for the different groups. The distribution of backyard burning samples is to some extent related to waste contents. Thus, several of the garden waste-dominated samples (Y1-Y9) are located centre top within the group, the straws (Y17, Y18) are located together in the middle and RDF-dominated samples (Y13-Y15) to the right. It may not be possible to identify the reasons for these homologue distributions and the anomalies (e.g. Y10, Y16 and Y14) from the available experimental information. In the wood combustion in boilers and stoves there is a tendency for the combustions with full air supply to be located to the right, especially those in the modern environmentally-certified boiler (W9, W10 and to some extent W12) and the stove (W14), whereas the samples to the far left (W6, W13 and W11) are examples of smouldering combustions. The distribution of profiles from the biofuel boilers (B1-B14) seems to change over time, shifting more to the right with each experimental round, and not to relate noticeably to the contents of RDF and RCG. A notable feature is that the combustions in the 600 kW boiler before (B20a and B20b) and after an instrumental malfunction (B21a and B21b), generated similar profiles (located to the right within the group), despite very different levels of dioxin emissions. Average profiles for each of the groups are shown in Figure 12, where the general dominance of TCDF can be seen. This figure also shows the dominance of PeCDF in the profiles from laboratory reactor and back-yard experiment Y10.

TCDF

PeCDF

HxCDF

HpCDF

OCDF

TCDD

PeCDD

HxCDD

HpCDD

OCDD

0

0,1

0,2

0,3

0,4

0,5

0,6

Average homologue profiles

Lab reactor, LBiofuel boiler, BWood boilers, WPellet boiler, PBackyard, Y (-Y10)Backyard PVC, Y10

Figure 10. Average homologue profiles in dioxin profiles in emissions from combustions described in Papers I-IV.

40

5 Conclusions and suggestions for future work Source-sorted waste Some source-sorted combustible waste fractions, and especially textile fractions, had considerable levels of PCDD/F. Future work: Source-sorted waste has to be further examined with respect not only to dioxins, but also other pollutants that may contaminate materials during their lifetime. There is also a need for follow-up studies and extension of the rather sparse investigations of dioxin contamination originating from the production of textiles and leather.

Co-combustion of source-sorted waste and biofuel Combustion of source-sorted waste with reed canary-grass in biofuel boilers was both efficient and generated low emissions of carbon monoxide, although it was difficult to keep emission levels of PCDD/F and hydrogen chloride within regulative limit valid for incineration of MSW. However, the mass of dioxins in the flue-gas emissions was generally lower than the mass in the fuel input. Future work: Development of fuel mixes and combustion technology may reduce dioxin formation, but there is also a need for development of inexpensive cleaning technology adjusted to small-scale combustion. Combustion of pellets and firewood in residential stoves and boilers Emissions of PCDD/F, PCB and HCB in small-scale combustions of biofuel depend largely on combustion efficiency. The modern, environmentally certified boiler generated considerably lower emissions than old boilers and the simulated use of heat storage tanks connected to both old and new boilers further reduced emissions. Combustion of plastic waste in small-scale wood boilers yielded very high emissions of PCDD/F, PCB and HCB. Intermittent combustion of wood pellets can cause high emissions of PCDD/F, PCB and HCB. Future work: Emissions from intermittent combustion should be further examined with respect to differences in the design, materials used, and variables such as draught dynamics in the pellet burners. Different combinations of burners and boilers should also be tested. Backyard burning In backyard combustions with low or moderately contaminated waste (dominated by garden waste) it is difficult to relate emissions of PCDD/F and PCB to waste contents. Other uncontrollable factors, i.e. variables related to combustion conditions, are more important, resulting in a large range of values. A suggested emission factor range for

41

low and moderately contaminated wastes implies that backyard burning may make significant contributions to total emissions. Future work: The emission factors need to be estimated with greater precision, but establishment of a single emission factor covering all possible compositions of wastes in backyard burnings is not practicable. One approach that could be more valid may be to establish emission factors for low contaminated waste, and to relate any further emissions to the estimated total amount of waste chlorine combusted. However, since other variables also strongly influence emissions a more convenient approach may to be establish emission factors with greater precision from a number of ‘model fires’ with differing degrees of contamination, as presented in Chapter 4.4. This may provide useful values for assessment e.g. on a national basis where the total emissions can be adjusted according to theoretical considerations of the relative abundance of different types of fires. The significance of fuel content for dioxin emission. Emissions of PCDD/F depend on complex interactions between fuel-related and combustion-related variables. Although experiments can be designed to evaluate the influence of individual variables, the influence on dioxin emissions solely based on knowledge of the fuel content are difficult to apply in a predictive manner. A conclusion that however can be drawn is that combustion of fuels with very high levels of chlorine usually results in higher emissions of dioxins than combustion of fuels with very low levels of chlorine.

42

6 Acknowledgements Det tar gärna tid innan jag kommer igång med saker. Då tänker jag både på den mogna ålder då jag äntligen kom på att jag kanske ändå skulle satsa på doktorandstudier och på den rätt sena tidpunkt då jag på allvar kom igång med avhandlingsskrivandet. Nu är den i alla fall klar, tror jag, och jag hoppas det bär ända fram. Det finns många att tacka för att det kom ända hit. Först Rolf och Janne och andra på Röbäcksdalen som trodde på mig och uppmuntrade mig att söka doktorandtjänsten. Stellan och Calle som, rätt utnyttjat, är det ideala handledarparet. Om man behöver någon som säger ”det där blir bra, kör på!”, eller någon som säger ”vänta lite, stämmer det här?” så är det bara att välja vem man ska gå till. Lite som gas och broms. De som känner dem vet nog vem som är vem. Sen alla som har varit delaktiga eller bistått vid alla försök och allt som oftast bidragit med kunnande som jag själv saknat. Gunnar och Håkan på BTC, Micke numera också på BTC, Morgan som väl egentligen har dragit det tyngsta lasset i ett par av försöken, Esbjörn och alla andra på ETC. Alla människor i Bjurhom, Bredträsk, Karlsbäck, Mjösjöby, Balsjö, Balfors med flera byar, mina barndomstrakter, som levererat utmärkt sorterade sopor till samförbränningförsöken. Gunilla L och Maria som visat stort tålamod med att lära mig rutinerna på labbet, och som också har gjort en del av analysarbetet, vilket även gäller Anna och många som passerat labbet i kortare anställningar. Gunilla S och Per har bidragit med sina erfarenheter och alla andra i förbränningsgruppen med stöd, hjälp och input av olika slag;, Lisa, Stina, Johanna och David m fl. Speciellt tack till dem som har kommit med synpunkter på artiklar och på avhandlingen. Alla andra på Miljökemi vill jag också tacka. Jag är väl inte överdrivet social av mig, men jag har verkligen uppskattat den vänliga och inkluderande atmosfär som råder här. Mina barn, Lisa, Anna och Nils som visat förståelse för att jag inte har hunnit umgås så mycket med dem i sommar. Tur också att ni klarar er så bra själva numera. Och slutligen Yvonne, tack för ditt stöd och din uppmuntran!

43

7 References

[1] Kumar KS,Kannan K,Corsolini S,Evans T,Giesy JP,Nakanishi J,Masunaga S.

Polychlorinated Dibenzo-p-Dioxins, Dibenzofurans and Polychlorinated Biphenyls in Polar Bear, Penguin and South Polar Skua, Environmental Pollution 2002; 119,(2) :151-161.

[2] Lamparski LL,Nestrick TJ,Stehl RH. Determination of Part-Per-Trillion

Concentrations of 2,3,7,8-Tetrachlorodibenzo-p-Dioxin in Fish, Anal. Chem. 1979; 51,(9) :1453-1458.

[3] Oehme M,Biseth A,Schlabach M,Wiig O. Concentrations of Polychlorinated

Dibenzo-p-Dioxins, Dibenzofurans and Non-Ortho Substituted Biphenyls in Polar Bear Milk From Svalbard (Norway), Environmental Pollution 1995; 90,(3) :401-407.

[4] Svensson BG,Nilsson A,Hansson M,Rappe C,Akesson B,Skerfving S. Exposure to

Dioxins and Dibenzofurans Through the Consumption of Fish, New England Journal of Medicine 1991; 324,(1) :8-12.

[5] Bernes, C. Persistent Organic Pollutants; Swedish Environmental Protection Agency:

1998. [6] Van den BM,Birnbaum L,Bosveld AT,Brunstrom B,Cook P,Feeley M,Giesy

JP,Hanberg A,Hasegawa R,Kennedy SW,Kubiak T,Larsen JC,van Leeuwen FX,Liem AK,Nolt C,Peterson RE,Poellinger L,Safe S,Schrenk D,Tillitt D,Tysklind M,Younes M,Waern F,Zacharewski T. Toxic Equivalency Factors (TEFs) for PCBs, PCDDs, PCDFs for Humans and Wildlife, Environ. Health Perspect. 1998; 106,(12) :775-792.

[7] Taylor PH,Sidhu SS,Rubey WA,Dellinger B,Wehrmeier A,Lenoir D,Schramm KW.

Evidence for a Unified Pathway of Dioxin Formation From Aliphatic Hydrocarbons, Symposium (International) on Combustion 1998; 2,1769-1775.

[8] Wehrmeier A,Lenoir D,Sidhu SS,Taylor PH,Rubey WA,Kettrup A,Dellinger B. Role

of Copper Species in Chlorination and Condensation Reactions of Acetylene, Environ. Sci. Technol. 1998; 32,(18) :2741-2748.

[9] Addink R,Olie K. Mechanisms of Formation and Destruction of Polychlorinated

Dibenzo-P-Dioxins and Dibenzofurans in Heterogeneous Systems, Environ. Sci. Technol. 1995; 29,(6) :1425-1435.

44

[10] Iino F,Imagawa T,Takeuchi M,Sadakata M. De Novo Synthesis Mechanism of Polychlorinated Dibenzofurans From Polycyclic Aromatic Hydrocarbons and the Characteristic Isomers of Polychlorinated Naphthalenes, Environmental-Science-and-Technology. April 1999; 33,(7) :1038-1043.

[11] Ryu JY,Mulholland JA,Dunn JE,Iino F,Gullett BK. Potential Role of Chlorination on

Pathways in PCDD/F Formation in a Municipal Waste Incinerator, Environ. Sci. Technol. 2004; 38,(19) :5112-5119.

[12] Wikstrom E,Ryan S,Touati A,Telfer M,Tabor D,Gullett BK. Importance of Chlorine

Speciation on De Novo Formation of Polychlorinated Dihenzo-p-Dioxins and Polychlorinated Dibenzofurans, Environ. Sci. Technol. 2003; 37,(6) :1108-1113.

[13] Liljelind, P. Studies of formation and catalytic decomposition of chlorinated dioxins

and related compounds in lab-scale experiments. Thesis. Department of Chemistry, Environmental Chemistry. Umeå University. 2002.

[14] Gullett BK,Sarofim AF,Smith KA,Procaccini C. The Role of Chlorine in Dioxin Formation, Process Safety and Environmental Protection 2000; 78,(B1) :47-52

. [15] Huang H,Buekens A. On the Mechanisms of Dioxin Formation in Combustion

Processes, Chemosphere 1995; 31,(9) :4099-4117. [16] Stanmore BR. The Formation of Dioxins in Combustion Systems, Combustion and

Flame 2004; 136,(3) :398-427. [17] Hatanaka T,Imagawa T,Kitajima A,Takeuchi M. Effects of Combustion Temperature

on PCDD/Fs Formation in Laboratory- Scale Fluidized-Bed Incineration, Environ. Sci. Technol. 2001; 35,(24) :4936-4940.

[18] Addink R,Olie K. Role of Oxygen in Formation of Polychlorinated Dibenzo-P-

Dioxins/Dibenzofurans From Carbon on Fly-Ash, Environ. Sci. Technol. 1995; 29,(6) :1586-1590.

[19] Griffin RD. A New Theory of Dioxin Formation in Municipal Solid Waste

Combustion, Chemosphere 1986; 15,(9-12) :1987-1990. [20] Pekarek V,Grabic R,Marklund S,Puncochar M,Ullrich J. Effects of Oxygen on

Formation of PCB and PCDD/F on Extracted Fly Ash in the Presence of Carbon and Cupric Salt, Chemosphere 2001; 43,(4-7) :777-782.

[21] Vogg H,Metzger M,Stieglitz L. Recent Findings on the Formation and

Decomposition of Pcdd/Pcdf in Municipal Solid-Waste Incineration, Waste Manag. & Res. 1987; 5,(3) :285-294.

[22] Wikstrom E,Tysklind M,Marklund S. Influence of Variation in Combustion

Conditions on the Primary Formation of Chlorinated Organic Micropollutants During Municipal Solid Waste Combustion, Environ. Sci. Technol. 1999; 33,(23) :4263-4269.

45

[23] Wikstrom, E. The Role of Chlorine during Waste Combustion. Thesis. Department of Chemistry, Environmental Chemistry. Umeå University 1999

[24] Griffiths, J. F.; Barnard, J. A. Flame and Combustion; Blackie Academic &

Professional: 1995. [25] Huang H,Buekens A. Chlorinated Dioxins and Furans As Trace Products of

Combustion: Some Theoretical Aspects, Toxicol. Environ. Chem. 2000; 74,(3-4) :179-193.

[26] Hell K,Stieglitz L,Dinjus E. Mechanistic Aspects of the De-Novo Synthesis of

PCDD/PCDF on Model Mixtures and MSWI Fly Ashes Using Amorphous 12C- and 13C-Labeled Carbon, Environ. Sci. Technol. 2001; 35,(19) :3892-3898.

[27] Wikstrom E,Ryan S,Touati A,Tabor D,Gullett BK. Origin of Carbon in

Polychlorinated Dioxins and Furans Formed During Sooting Combustion, Environ. Sci. Technol. 2004; 38,(13) :3778-3784.

[28] Wikstrom E,Ryan S,Touati A,Gullett BK. In Situ Formed Soot Deposit As a Carbon

Source for Polychlorinated Dibenzo-p-Dioxins and Dibenzofurans, Environ. Sci. Technol. 2004; 38,(7) :2097-2101.

[29] Stanmore BR. Adsorption and Partitioning of PCDD/F on Flyash From Incinerators,

Environmental Engineering Science 2002; 19,(2) :69-78. [30] Fangmark I,Stromberg B,Berge N,Rappe C. The Influence of Fly Ash Load and

Particle Size on the Formation of PCDD, PCDF, PCBz and PCB in a Pilot Incinerator, Waste Manag. & Res. 1995; 13,(3) :259-272.

[31] Kurokawa Y,Takahiko M,Matayoshi N,Satoshi T,Kazumi F. Distribution of

Polychlorinated Dibenzo-p-Dioxins and Dibenzofurans in Various Sizes of Airborne Particles, Chemosphere 1998; 37,(9-12) :2161-2171.

[32] Gullett BK,Bruce KR,Beach LO,Drago AM. Mechanistic Steps in the Production of

PCDD and PCDF During Waste Combustion, Chemosphere 1992; 25,(7-10) :1387-1392.

[33] Halonen I,Tarhanen J,Ruokojarvi P,Tuppurainen K,Ruuskanen J. Effect of Catalysts

and Chlorine Source on the Formation of Organic Chlorinated Compounds, Chemosphere 1995; 30,(7) :1261-1273.

[34] Gullett BK,Dunn JE,Raghunathan K. Effect of Cofiring Coal on Formation of

Polychlorinated Dibenzo-p-Dioxins and Dibenzofurans During Waste Combustion, Environ. Sci. Technol. 2000; 34,(2) :282-290.

[35] Wang KS,Chiang KY,Lin SM,Tsai CC,Sun CJ. Effects of Chlorides on Emissions of

Hydrogen Chloride Formation in Waste Incineration, Chemosphere 1999; 38,(7) :1571-1582.

46

[36] Wikstrom E,Ryan S,Touati A,Gullett BK. Key Parameters for De Novo Formation of Polychlorinated Dibenzo-p-Dioxins and Dibenzofurans, Environ. Sci. Technol. 2003; 37,(9) :1962-1970.

[37] Stieglitz L,Vogg H,Zwick G,Beck J,Bautz H. On Formation Conditions of

Organohalogen Compounds From Particulate Carbon of Fly Ash, Chemosphere 1991; 23,(8-10) :1255-1264.

[38] Hatanaka T,Imagawa T,Takeuchi M. Formation of PCDD/Fs in Artificial Solid

Waste Incineration in a Laboratory-Scale Fluidized-Bed Reactor: Influence of Contents and Forms of Chlorine Sources in High-Temperature Combustion, Environ. Sci. Technol. 2000; 34,(18) :3920-3924.

[39] Halonen I,Tarhanen J,Kopsa T,Palonen J,Vilokki H,Ruuskanen J. Formation of

Polychlorinated Dioxins and Dibenzofurans in Incineration of Refuse Derived Fuel and Biosludge, Chemosphere 1993; 26,(10) :1869-1880.

[40] Huotari J,Vesterinen R. PCDD/F Emissions From Co-Combustion of RDF with Peat,

Wood Waste, and Coal in FBC Boilers, Hazard. Waste Hazard. Mater. 1996; 13,(1) :1-10.

[41] Vesterinen R,Flyktman M. Organic Emissions From Co-Combustion of RDF with

Wood Chips and Milled Peat in a Bubbling Fluidized Bed Boiler, Chemosphere 1996; 32,(4) :681-689.

[42] Wikstrom E,Marklund S. The Influence of Level and Chlorine Source on the

Formation of Mono- to Octa-Chlorinated Dibenzo-p-Dioxins, Dibenzofurans and Coplanar Polychlorinated Biphenyls During Combustion of an Artificial Municipal Waste, Chemosphere 2001; 43,(2) :227-234.

[43] Launhardt T,Thoma H. Investigation on Organic Pollutants From a Domestic Heating

System Using Various Solid Biofuels, Chemosphere 2000; 40,(9-11) :1149-1157 . [44] Wikstrom E,Lofvenius G,Rappe C,Marklund S. Influence of Level and Form of

Chlorine on the Formation of Chlorinated Dioxins, Dibenzofurans, and Benzenes During Combustion of an Artificial Fuel in a Laboratory Reactor, Environ. Sci. Technol. 1996; 30,1637-1644.

[45] Gullett BK,Lemieux PM. Role of Combustion and Sorbent Parameters in Prevention

of Polychlorinated Dibenzo-P-Dioxin and Polychlorinated Dibenzofuran Formation During Waste Combustion, Environ. Sci. Technol. 1994; 28,(1) :107-118.

[46] Luijk R,Akkerman DM,Slot P,Olie K,Kapteijn F. Mechanism of Formation of

Polychlorinated Dibenzo-P-Dioxins and Dibenzofurans in the Catalyzed Combustion of Carbon, Environ. Sci. Technol. 1994; 28,(2) :312-321.

[47] Fabrellas B,Sanz P,Abad E,Rivera J. The Spanish Dioxin Inventory Part I:

Incineration As Municipal Waste Management System, Chemosphere 2001; 43,(4-7) :683-688.

47

[48] Chi KH,Chang MB,Chang-Chien GP,Lin C. Characteristics of PCDD/F Congener Distributions in Gas/Particulate Phases and Emissions From Two Municipal Solid Waste Incinerators in Taiwan, Science of The Total Environment 2005; 347,(1-3) :148-162.

[49] Hunsinger H,Jay K,Vehlow J. Formation and Destruction of PCDD/F Inside a Grate

Furnace, Chemosphere 2002; 46,(9-10) :1263-1272. [50] Ryu JY,Mulholland JA,Kim DH,Takeuchi M. Homoloque and Isomer Patterns of

Polychlorinated Dibenzo-p-Dioxins and Dibenzofurans From Phenol Precursors: Comparison With Municipal Waste Incinerator Data, Environ. Sci. Technol. 2005; 39,(12) :4398-4406.

[51] Abad E,Caixach J,Rivera J. PCDD/PCDF From Emission Sources and Ambient Air

in Northeast Spain, Chemosphere 1997; 35,(3) :453-463. [52] Rappe C,Swanson SE,Glas B,Kringstad KP,Desousa P,Abe Z. Formation of PCDDs

and PCDFs by the Chlorination of Water, Chemosphere 1989; 19,(12) :1875-1880. [53] Tysklind M,Fangmark I,Marklund S,Lindskog A,Thaning L,Rappe C. Atmospheric

Transport and Transformation of Polychlorinated Dibenzo-P-Dioxins and Dibenzofurans, Environ. Sci. Technol. 1993; 27,(10) :2190-2197.

[54] Baker J, I,Hites RA. Is Combustion the Major Source of Polychlorinated Dibenzo-p-

Dioxins and Dibenzofurans to the Environment? A Mass Balance Investigation, Environ. Sci. Technol. 2000; 34,(14) :2879-2886.

[55] Berry RM,Luthe CE,Voss RH. Ubiquitous Nature of Dioxins - A Comparison of the

Dioxins Content of Common Everyday Materials With That of Pulps and Papers, Environ. Sci. Technol. 1993; 27,(6) :1164-1168.

[56] Shariff A,Nguyen XT. Characterization and Possible Sources of Chlorinated Dioxins

and Furans in Recycled Fiber, Pulp & Paper-Canada 1992; 93,(9) :29-37. [57] Swanson SE,Rappe C,Malmstrom J,Kringstad KP. Emissions of PCDDs and PCDFs

From the Pulp Industry, Chemosphere 1988; 17,(4) :681-691. [58] Iino F,Takasuga T,Touati A,Gullett BK. Correlations Between PCDD/F Homologue

Concentrations and TEQ Values in Laboratory-, Boiler-, and Practical-Scale Incinerators, Organohalogen Compd. 2001; 54,(Dioxin 2001) :351-354.

[59] Oehme M,Muller MD. Levels and Congener Patterns of Polychlorinated Dibenzo-p-

Dioxins and Dibenzofurans in Solid Residues From Wood-Fired Boilers. Influence of Combustion Conditions and Fuel Type, Chemosphere 1995; 30,(8) :1527-1539.

[60] Eitzer BD,Hites RA. Dioxins and Furans in the Ambient Atmosphere: A Baseline

Study, Chemosphere 1989; 18,(1-6) :593-598.

48

[61] Wingfors H,Lindstrom G,van Bavel B,Schuhmacher M,Hardell L. Multivariate Data Evaluation of PCB and Dioxin Profiles in the General Population in Sweden and Spain, Chemosphere 2000; 40,(9-11) :1083-1088.

[62] Ogura I,Masunaga S,Nakanishi J. Congener-Specific Characterization of

PCDDs/PCDFs in Atmospheric Deposition: Comparison of Profiles Among Deposition, Source, and Environmental Sink, Chemosphere 2001; 45,(2) :173-183.

[63] Brzuzy LP,Hites RA. Global Mass Balance for Polychlorinated Dibenzo-p-Dioxins

and Dibenzofurans, Environ. Sci. Technol. 1996; 30,(6) :1797-1804. [64] Chen HL,Su HJ,Liao PC,Chen CH,Lee CC. Serum PCDD/F Concentration

Distribution in Residents Living in the Vicinity of an Incinerator and Its Association With Predicted Ambient Dioxin Exposure, Chemosphere 2004; 54,(10) :1421-1429.

[65] Hansson M,Barregard L,Sallsten G,Svensson BG,Rappe C. Polychlorinated Dibenzo-

p-Dioxin and Dibenzofuran Levels and Patterns in Polyvinylchloride and Chloralkali Industry Workers, International Archives of Occupational and Environmental Health 1997; 70,(1) :51-56.

[66] Matsueda, T.; Hirakawa, H.; Iida, T.; Nakamura, M.; Nagayama, J. Concentration of

PCDDs/PCDFs and coplanar PCBs in human skin lipids. Organohalogen compounds. 1995; 26, 219-222.

[67] Alcock RE,Sweetman AJ,Jones KC. A Congener-Specific PCDD/F Emissions

Inventory for the UK: Do Current Estimates Account for the Measured Atmospheric Burden?, Chemosphere 2001; 43,(2) :183-194.

[68] The Council of the European communities. Council Directive 75/442/EEC of 15 July

1975 on waste. 1975.

[69] Finnveden G,Johansson J,Lind P,Moberg A. Life Cycle Assessment of Energy From Solid Waste - Part 1: General Methodology and Results, Journal of Cleaner Production 2005; 13,(3) :213-229.

[70] McDougall FR,Hruska JP. The Use of Life Cycle Inventory Tools to Support an

Integrated Approach to Solid Waste Management, Waste Manag. & Res. 2000; 18,(6) :590-594.

[71] Sundqvist, J.-O. Life cycles assessments and solid wastes - Guidelines for solid waste

treatment and disposal in LCA. Final report. 279, 1-133. IVL, Swedish Environmental Research Institute. AFR-Report. 1999.

[72] Eriksson O,Reich MC,Frostell B,Bjorklund A,Assefa G,Sundqvist JO,Granath J,Baky A,Thyselius L. Municipal Solid Waste Management From a Systems Perspective, Journal of Cleaner Production 2005; 13,(3) :241-252.

[73] Harrison KW,Dumas RD,Barlaz MA,Nishtala SR. A Life-Cycle Inventory Model of

Municipal Solid Waste Combustion, J. Air Waste Manag. Assoc. 2000; 50,(6) :993-1003.

49

[74] Otoma S,Mori Y,Terazono A,Aso T,Sameshima R. Estimation of Energy Recovery and Reduction of CO2 Emissions in Municipal Solid Waste Power Generation, Resources, Conservation and Recycling 1997; 20,(2) :95-117.

[75] Sundqvist, J.-O.; Finnveden, G.; Sundberg, J. Syntes av systemanalyser av

avfallshantering. B1491. 2002. IVL Rapport. (In Swedish) 2002.

[76] Sundqvist, J.-O.; Baky, A.; Björklund, A.; Carlsson, M.; Eriksson, O.; Frostell, B.; Granath, J.; Thyselius, L. Systemanalys av energiutnyttjande från avfall - utvärdering av energi, miljö och ekonomi. IVL.Rapport (In Swedish) 1999

[77] RVF. Svensk avfallshanterinhg 2005, Årsskrift. (In Swedish) 2005

[78] Ohlsson, T. Plockanalys av hushållsavfall. Metoder och trender. 1998:226 CIV. 1998. : Institutionen för Samhällsbyggnadsteknik, Avdelningen för Upplagsteknik (senare avfallsteknik). Luleå University of Technology (LTU) (In Swedish). 1998

[79] Leroy D,Giovanni JM,Maystre LY. Sampling Method to Determine a Household

Waste Composition Variance., Waste Manag. & Res. 1992; 10,(1) :3-12. [80] RVF Utveckling. Trender och variationer i hushållsavfallets sammansättning.

2005:05. Report (In Swedish) 2005.

[81] Giugliano M,Cernuschi S,Grosso M,Aloigi E,Miglio R. The Flux and Mass Balance of PCDD/F in a MSW Incineration Full Scale Plant, Chemosphere 2001; 43,(4-7) :743-750.

[82] Giugliano M,Cernuschi S,Grosso M,Miglio R,Aloigi E. PCDD/F Mass Balance in the

Flue Gas Cleaning Units of a MSW Incineration Plant, Chemosphere 2002; 46,(9-10) :1321-1328.

[83] Sakai SI,Hayakawa K,Takatsuki H,Kawakami I. Dioxin-Like PCBs Released From

Waste Incineration and Their Deposition Flux, Environ. Sci. Technol. 2001; 35,(18) :3601-3607.

[84] Abad E,Adrados MA,Caixach J,Rivera J. Dioxin Abatement Strategies and Mass

Balance at a Municipal Waste Management Plant, Environ. Sci. Technol. 2002; 36,(1) :92-99.

[85] Hamai, M.; Kondoh, M.; Yamaguchi, M.; Mori, S. Emission characteristics of

dioxins from RDF during incineration in internal-circulation-type fluidized bed boiler. 459-467. 2000. Irvine, Office of Environmental, Health & Safety, Irvine, Calif. Proceedings of the International Conference on Incineration and Thermal Treatment Technologies, Portland, OR, United States, May 8-12, 2000.

[86] Shin D,Choi S,Oh JE. Evaluation of Polychlorinated Dibenzo-p-Dioxin/Dibenzofuran (PCDD/F) Emission in Municipal Solid Waste Incinerators, Environ. Sci. Technol. 1999; 33,(15) :2657-2666.

50

[87] European Parliament; European Council. Directive 2000/76/EC of the European Parliament and of the Council of 4 December 2000 on the incineration of waste. 2002.

[88] Ahlgren, N.; Marklund, S. Förbränning av avfall. En kunskapssammanställning om dioxiner. 01:13. 2001. RVF Rapport.(In Swedish) 2001.

[89] Sakai S,Urano S,Takatsuki H. Leaching Behavior of PCBs and PCDDs/DFs From Some Waste Materials, Waste Management 2000; 20,(2-3) :241-247.

[90] Shin KJ,Chang YS. Characterization of Polychlorinated Dibenzo-p-Dioxins,

Dibenzofurans, Biphenyls, and Heavy Metals in Fly Ash Produced From Korean Municipal Solid Waste Incinerators, Chemosphere 1999; 38,(11) :2655-2666.

[91] Abad E,Adrados MA,Caixach J,Fabrellas B,Rivera J. Dioxin Mass Balance in a

Municipal Waste Incinerator, Chemosphere 2000; 40,(9-11) :1143-1147. [92] Lahl U,Wilken M,Zeschmarlahl B,Jager J. PCDD/PCDF Balance of Different

Municipal Waste Management Methods, Chemosphere 1991; 23,(8-10) :1481-1489. [93] Mohr K,Nonn C,JAGER J. Behaviour of PCDD/F Under Pyrolysis Conditions,

Chemosphere 1997; 34,(5-7) :1053-1064. [94] Wilken M,Cornelsen B,Zeschmarlahl B,Jager J. Distribution of PCDD/PCDFand

Other Organochlorine Compounds in Different Municipal Waste Fractions, Chemosphere 1992; 25,(7-10) :1517-1523.

[95] Vehlow, J. Simple, Reliable and Yet Efficient - Modern strategies in waste

incineration. UTA International 2, 144-166. 1996.

[96] Makoto, S.; Yoji, S.; Yasuhiro, I.; Toru, K.; Teruaki, T.; Osamu, F. Reduction of Total Dioxin Emission from MSW Incinerators. Organohalogen compounds.1998. 36, 325-328.

[97] Johnke B,Stelzner E. Results of the German Dioxin Measurement Programme at

MSW Incinerators, Waste Manag. & Res. 1992; 10,345-355. [98] Johnke B. Waste Incineration – an Important Element of the Integrated Waste

Management System in Germany, Waste Manag. & Res. 1992; 10,(4) :303-315. [99] Eduljee GH,Dyke P. An Updated Inventory of Potential PCDD and PCDF Emission

Sources in the UK, Science of The Total Environment 1996; 177,(1-3) :303-321. [100] Paulrud, S. Uppgraded biofuels - effects of quality on processing, handling

characteristics, combustion and ash melting. Thesis. Unit of Biomass Technology and Chemistry, Swedish University of Agricultural Sciences (SLU) 2004.

[101] Finell M,Nilsson C,Olsson R,Agnemo R,Svensson S. Briquetting of Fractionated Reed Canary-Grass for Pulp Production, Industrial Crops and Products 2002; 16,(3) :185-192.

51

[102] Paulrud S,Nilsson C. Briquetting and Combustion of Spring-Harvested Reed Canary-Grass: Effect of Fuel Composition, Biomass and Bioenergy 2001; 20,(1) :25-35.

[103] Hadders G,Olsson R. Harvest of Grass for Combustion in Late Summer and in

Spring, Biomass and Bioenergy 1997; 12,(3) :171-175. [104] Landstrom S,Lomakka L,Andersson S. Harvest in Spring Improves Yield and Quality

of Reed Canary Grass As a Bioenergy Crop, Biomass and Bioenergy 1996; 11,(4) :333-341.

[105] Burvall J. Influence of Harvest Time and Soil Type on Fuel Quality in Reed Canary

Grass (Phalaris Arundinacea L.), Biomass and Bioenergy 1997; 12,(3) :149-154. [106] Finell M,Nilsson C. Variations in Ash Content, Pulp Yield, and Fibre Properties of

Reed Canary-Grass, Industrial Crops and Products 2005; 22,(2) :157-167. [107] Paulrud S,Nilsson C,Ohman M. Reed Canary-Grass Ash Composition and Its

Melting Behaviour During Combustion, Fuel 2001; 80,(10) :1391-1398. [108] Burvall, J.; Örberg, H.; Kalén, G.; Marklund, S.; Hedman, B. Sameldning av

biobränslen och källsorterat hushållsavfall – Fullskaleförsök Bjurholm . FBT-04/10, 1-29. Energimyndigheten. (In Swedish) 2004.

[109] Johansson LS,Leckner B,Gustavsson L,Cooper D,Tullin C,Potter A. Emission Characteristics of Modern and Old-Type Residential Boilers Fired With Wood Logs and Wood Pellets, Atmospheric Environment 2004; 38,(25) :4183-4195.

[110] Asari M,Takatsuki H,Sakai S. Polycyclic Aromatic Hydrocarbons,

Pentachlorophenol, Dioxins and Furans in Waste Wooden Sleepers on the Recycling Market, Organohalogen Compd. 2002; 59,199-202.

[111] Bright DA,Cretney WJ,MacDonald RW,Ikonomou MG,Grundy SL. Differentiation

of Polychlorinated Dibenzo-p-Dioxin and Dibenzofuran Sources in Coastal British Columbia, Canada, Environmental Toxicology and Chemistry 1999; 18,(6) :1097-1108.

[112] Freeman MC,O'Dowd WJ,Brown TD,HargisJr. RA,James RA,Plasynski SI,Walbert GF,Lowe AF,BattistaJr. JJ. Pilot-Scale Air Toxics R&D Assessment of Creosote-Treated and PCP-Treated Wood Cofiring for Pulverized Coal Utility Boiler Applications, Biomass and Bioenergy 2000; 19,(6) :447-456.

[113] Hubner C,Boos R,Prey T. In-Field Measurements of PCDD/F Emissions From

Domestic Heating Appliances for Solid Fuels, Chemosphere 2005; 58,(3) :367-372. [114] Lavric ED,Konnov AA,De Ruyck J. Dioxin Levels in Wood Combustion - a Review,

Biomass & Bioenergy 2004; 26,(2) :115-145. [115] Schleicher, O.; Jensen, A. A.; Blinksbjerg, P.; Thomsen, E.; Schilling, B. Dioxin

emissions from biomass fired energy plants and other sources in Denmark. Organohalogen Compounds 2002, 56, 147-150.

52

[116] Baxter, L.; Coombes, P.; Gifford, J.; Guiu, G.; Houmann Jakobssen, H.; Koppejan, J.; Livingston, W.; Logie, R.; Loo van, S.; Nussbaumer, T.; Obenberger, I.; Oravainen, H.; Schenkel, Y.; Skreiberg, Ö.; Tullin, C. Handbook of Biomass Combustion and Co-Firing; Twente University Press: 2002

. [117] Gullett BK,Lemieux PM,Lutes CC,Winterrowd CK,Winters DL. Emissions of

PCDD/F From Uncontrolled, Domestic Waste Burning, Chemosphere 2001; 43,(4-7) :721-725.

[118] Ikeguchi T,Tanaka M. Experimental Study on Dioxins Emissions From Open

Burning of Selected Wastes, Organohalogen Compd. 1999; 41,507-510. [119] Ikeguchi T,Tanaka M. Dioxin Emisions From an Open-Burning-Like Waste

Incineration: Small Incinerators for Household Use, Organohalogen Compd. 2000; 46,298-301.

[120] Lemieux PM,Lutes CC,Abbot JA,Aldous KM. Emissions of Polychlorinated

Dibenzo-p-Dioxins and Polychlorinated Dibenzofurans From the Open Burning of Household Waste in Barrels., Environ. Sci. Technol. 2000; 34,377-384.

[121] Lemieux PM,Gullett BK,Lutes CC,Winterrowd CK,Winters DL. Variables Affecting

Emissions of PCDD/Fs From Uncontrolled Combustion of Household Waste in Barrels, Journal of the Air & Waste Management Association 2003; 53,(5) :523-531.

[122] Wevers M,De Fre R,Desmedt M. Effect of Backyard Burning on Dioxin Deposition

and Air Concentrations, Chemosphere 2004; 54,(9) :1351-1356. [123] Nordin A. Chemical Elemental Characteristics of Biomass Fuels, Biomass and

Bioenergy 1994; 6,(5) :339-347 . [124] Wikstrom E,Andersson P,Marklund S. Design of a Laboratory Scale Fluidized Bed

Reactor, Review of Scientific Instruments 1998; 69,(4) :1850-1859. [125] Novator föreningskonsult. Vedpärmen. Ljungblom, Lennart. Novator Media AB. (In

Swedish) 1996

[126] European Committee for Standardization. Stationary source emissions- Determination of the mass concentration of PCDDs/PCDFs, EN:1948:1-3. 1997.

[127] Boman C,Nordin A,Westerholm R,Pettersson E. Evaluation of a Constant Volume Sampling Setup for Residential Biomass Fired Appliances--Influence of Dilution Conditions on Particulate and PAH Emissions, Biomass and Bioenergy 2005; 29, (4): 258-268

[128] Liljelind P,Söderström G,Hedman B,Karlsson S,Lundin L,Marklund S. Method for

Multi-Residue Determination of Halogenated Aromatics and PAHs in Combustion Related Samples, Environ. Sci. Technol. 2003; 37,(16) :3680-3686.

53

[129] Vandenburg HJ,Clifford AA,Bartle KD,Carroll J,Newton I,Garden LM,Dean JR,Costley CT. Analytical Extraction of Additives From Polymers, Analyst 1997; 122,(9) :R101-R115.

[130] Wagenaar H,Langeland K,Hardman R,Sergeant Y,Brenner K,Sandra P,Rappe

C,Fernandes A,Tiernan T. Analysis of PCDDs and PCDFs in Virgin Suspension PVC Resin, Chemosphere 1998; 36,(1) :1-12.

[131] Vandenburg HJ,Clifford AA,Bartle KD,Zhu SA,Carroll J,Newton ID,Garden LM.

Factors Affecting High-Pressure Solvent Extraction (Accelerated Solvent Extraction) of Additives From Polymers, Anal. Chem. 1998; 70,(9) :1943-1948.

[132] Eriksson, L.; Johansson, E.; Kettaneh-Wold, N.; Wold, S. Multi- and Megavariate

Data Analysis; Umetrics AB: 2001. [133] Horstmann M,McLachlan MS. Results of an Initial Survey of Polychlorinated

Dibenzo-p-Dioxins (PCDD) and Dibenzofurans (PCDF) in Textiles, Chemosphere 1995; 31,(2) :2579-2589.

[134] Klasmeier, J.; McLachlan, M. S. Pentachlorophenol Mediated Contamination of

Leather and Leather Goods with Polychlorinated Dibenzo-p-doxins and Dibenzofurans. Organohalogen compounds 1999; 40, 169-172.

[135] Malisch, R. Determination of PCDD/PCDF in PCP-Contaminated Leather Samples. Organohalogen Compound 1994; 19, 73-76.

[136] Krizanec B,Le Marechal AM,Voncina E,Brodnjak-Voncina D. Presence of Dioxins

in Textile Dyes and Their Fate During the Dyeing Processes, Acta Chimica Slovenica 2005; 52,(2) :111-118.

[137] Klasmeier J,McLachlan MS. PCDD/Fs in Textiles - Part I: A Screening Method for

Detection of Octachlorodibenzo-p-Dioxin and Octachlorodibenzofuran, Chemosphere 1998; 36,(7) :1627-1635.

[138] Lexen K,de Wit C,Jansson B,Kjeller LO,Kulp SE,Ljung K,Soderstrom G,Rappe C.

Polychlorinated Dibenzo-p-Dioxin and Dibenzofuran Levels and Patterns in Samples From Different Swedish Industries Analyzed Within the Swedish Dioxin Survey, Chemosphere 1993; 27,(1-3) :163-170.

[139] Carroll WF,Berger TC,Borrelli FE,Garrity PJ,Jacobs RA,Lewis JW,McCreedy

RL,Tuhovak DR,Weston AF. Characterization of Emissions of Dioxins and Furans From Ethylene Dichloride (EDC), Vinyl Chloride (VCM) and Polyvinylchloride) (PVC) Manufacturing Facilities in the United States. I. Resin, Treated Wastewater, and Ethylene Dichloride, Chemosphere 1998; 37,(9-12) :1957-1972.

[140] Sinkkonen S,Makela R,Vesterinen R,Lahtipera M. Chlorinated Dioxins and

Dibenzothiophenes in Fly Ash Samples From Combustion of Peat, Wood Chips, Refuse Derived Fuel and Liquid Packaging Boards, Chemosphere 1995; 31,(2) :2629-2635.

54

[141] Olsson, R. Unit for Biofuel Technology and Chemistry (BTK), Swedish University of Agricultural Sciences. Personal Communication 2005.

[142] Johansson; Linda. Utsläpp av dioxiner, furaner, hexaklorbensen och PCB från småskalig pelletseldning, avtal nr 505 0405. 1-6. 2005. SP-rapport. Report (In Swedish, 2005