Bulletin of Engineering Geology and the Environment

, Volume 63, Issue 3, pp 191–207

Limestone and dolomite powder as binders for wood ash agglomeration

Authors

    • Department of Biology and Environmental ScienceUniversity of Kalmar
  • T. Claesson
    • Department of Biology and Environmental ScienceUniversity of Kalmar
Article

DOI: 10.1007/s10064-003-0223-4

Cite this article as:
Sarenbo, S.L. & Claesson, T. Bull Eng Geol Environ (2004) 63: 191. doi:10.1007/s10064-003-0223-4

Abstract

Limestone and dolomite powder were tested as binders during wood ash agglomeration on an industrial and a laboratory scale. Two agglomeration methods are compared. Dolomite from Estonia is commonly used as a binder/additive during automatic production of agglomerated wood ash at the central heating plant of Kalmar, Sweden. Swedish limestones from Öland and Ignaberga as well as Swedish dolomite from Glanshammar were used as binders in the production of test agglomerates of wood ash. The chemical and mineralogical composition of the binders as well as of the resulting granules and pellets is presented. The structural, chemical and leaching properties of the hardened ash/binder agglomerates are discussed in relation to their possible environmental impact on forest soil. The environmental acceptance of recycling of agglomerated ashes to forest soils is also discussed in relation to the new recommendations.

Keywords

AgglomerationWood ashAsh recyclingAcidificationHardening

Résumé

De la poudre de calcaire et de dolomite a été testée comme liant pour agglomérer de la cendre de bois à l’échelle du laboratoire et à une échelle industrielle. Deux méthodes d’agglomération sont comparées. De la dolomite d’Estonie est communément utilisée comme additif et liant lors de la production industrielle de cendres de bois agglomérées à la centrale thermique de Kalmar en Suède. Par ailleurs, les calcaires suédois de Öland et Ignaberga ainsi que les dolomies suédoises de Glanshammar ont été utilisés comme liants tests dans la production de cendres de bois agglomérées. Les compositions chimique et minéralogique des liants ainsi que des granules et pellets produits sont présentées. Les propriétés structurales, chimiques et de drainage des cendres de bois agglomérées et affermies par le liant sont analysées, par rapport à un impact environnemental possible sur des sols forestiers. Cet impact est discuté par référence aux nouvelles recommandations environnementales.

Mots clés

AgglomérationCendres de boisRecyclage de cendresAcidificationAffermissement

Introduction

Sawdust is a common biofuel in European heating plants. In Sweden during 2002, sawdust was the main fuel in nine heating plants, and several plants are planned to be converted to burn sawdust in the near future (Daniel Jedfelt, personal communication). In Kalmar, Sweden, 50,000 tons of sawdust from a nearby flooring industry are used for heat production yearly. The burning of sawdust results in approximately 300 tons of ash every year. Nutrient losses caused by whole tree harvesting can be compensated for by recycling ash from burning of clean biofuel back to forest soils. Additionally, if wood ash is spread together with limestone/dolomite, the decrease of soil pH due to acid precipitation may be counteracted in the long term.

Combustion residues from wood burning (for convenience, called ash or wood ash in this article) have a highly variable composition, but are mainly composed of nutrient elements Ca, Mg, K, S and P. The main elements are present in the form of oxides (CaO, MgO, P2O3), sulphates (K2SO4) and chlorides (KCl, NaCl). Wood ash also includes variable amounts of micronutrients as well as heavy metals and organic compounds, such as polycyclic aromatic hydrocarbons (PAH).

Approximately 150,000 tons per year wood ash are deposited as waste in Sweden, although they could be recycled back to forest soils. Research is going on in order to find methods for suitable ash treatments (Bjurström and Sjöblom 1997; Windelhed 1998), since handling of untreated wood ash is problematic and stabilisation is necessary to make spreading feasible. Chemical stabilisation (hardening) includes hydration and carbonation of the ash. Hydration leads to the formation of portlandite, Ca(OH)2, which is transformed into calcite (CaCO3) in the subsequent process called carbonation. The carbonation process lowers the solubility of Ca significantly and reduces the alkalinity (Steenari et al. 1998). The mechanical stabilisation can be achieved by producing ash agglomerates by different techniques, such as granulation or roll pelleting. The agglomerates should ideally be stable over a period of several decades, especially if they are spread on newly harvested areas or in young forests.

It is difficult to produce agglomerates with extended dissolution times without additives. Cement and lime are used in several cases as bonding agents in coal or municipal solid waste ash to give satisfactory long-term compressive strength for the solidified products (Baykal and Döven 2000; Medici et al. 2000; Alba et al. 2001). Binders tested for wood ashes are cement, lignosulfonate and molasses (Börjesson 1992), but agglomeration without additives is also possible if the ash is of high quality. Agglomeration without additives is based on the self-hardening capacity of wood ash and requires an ash with a very low content of unburned residue (Steenari and Lindqvist 1997). Because wood ash should be spread together with limestone or dolomite on the most acidified forest soils (Ström 1994), it would probably be less time-consuming to mix ash and limestone/dolomite into agglomerates before spreading on forest soils instead of mixing them after hardening, crushing and sieving of the ash separately.

The ash from combustion of sawdust at the heating plant of Kalmar has been agglomerated by granulation during 1996–2000, and by pelletising since 2001. Crushed dolomite with a particle size <4 mm is used as binder during agglomeration. This dolomite is called Anelema dolomite, originating from a quarry in Anelema, Estonia. Importing a binder is expensive. Because local rock industry residues often are available, the environmental cost of transport could be reduced. In the search for alternative binders for ash agglomeration, some important aspects are emphasised: the production of granules or pellets must be efficient, the binder must not give any non-favourable properties to the agglomerates, and the final chemical composition and the stability of the agglomerates must comply with the recommendations set up by the Swedish forestry authorities. Some new compliance test methods designed for wood ash recycling are applied in this work and the results are discussed in relation to other laboratory methods.

In this work, two limestones and a dolomite are tested as binders and compared to the dolomite from Estonia, Anelema dolomite. The binders are limestones from Ignaberga and Öland in southern Sweden and dolomite from Glanshammar in central Sweden. The Anelema quarry was documented between 1969 and 1983 (Arvo Suppi, personal communication) and several hundred boreholes studied and described, although the chemistry was not described in detail. In order to shed light on the properties that might explain the binding ability of Anelema dolomite, a careful examination of its chemical and mineralogical composition was performed and the results compared to those from analysis of the other limestones and dolomites. The origin and geology of Anelema dolomite are described elsewhere (Raukas and Teedumäe 1997; Teedumäe et al. 1999) and are not further discussed in this paper.

Experimental procedures

Production of pellets with different binders at the heating plant

Pellets were produced with roll-pelleting equipment at the central heating plant of Kalmar on 12 occasions between 29 May and 21 June 2001. 100 kg of ash were homogenised in a 500 Fehmert ash mixer. Five 100 g sub-samples of the ash were taken before adding binder and water. 1 g of each ash sub-sample was dried at 105 °C for moisture content determination and burned at 550 °C for 2 h for determination of LOI (loss on ignition). Sub-samples were combined for chemical characterisation.

After collection of ash sub-samples, 100 kg of binder (Anelema dolomite, Glanshammar dolomite or Ignaberga limestone) were added to the ash and mixed. (The Glanshammar dolomite and the Ignaberga limestone were purchased in 25 kg sacks.) Finally, water was added in small amounts during continuous mixing until a suitable moisture content was achieved. One 100 kg ash batch was pelletised without binder and three batches with 100 kg of ash and 50 kg of Anelema dolomite. The resulting pellets were dried at 50 °C for about 2 days indoors before bringing them to the laboratory. Samples of pellets weighing about 2 kg were taken for determination of the particle size distribution and for chemical and mineralogical testing. The granule samples were produced with the former equipment, drum-granulator, at the heating plant 2 years earlier, and stored in double plastic bags in the laboratory until testing with the pelletised materials. The pellet/granule samples were designated as for (coarse) particle size fraction 0.5–1 mm F (fine) 0.25–0.5 mm, Gran=granules, Pel=pellets, 1/1=ash/binder weight ratio). Anel=dolomite from Anelema, Glans=dolomite from Glanshammar, Igna=limestone from Ignaberga, PelAsh=Pellets made of ash, without binder.

Dolomite and limestone sampling

Rock samples weighing about 25 kg were taken to the laboratory from five levels in the stratigraphy at the Anelema quarry. Four of the intervals are shown in Fig. 1. One sample consisted of clay cement (1–2-cm thick layer on the solid rock) occurring in Level 2. The Ignaberga limestone sample (~20 kg) was taken from piles at the quarry and from a commercial 25 kg bag. The Glanshammar dolomite was taken from a commercial 25 kg bag and the Öland limestone (a limestone powder, 25 kg) was taken from piles of rock sawing residue at Söderströms quarry, Öland. The samples were stored in double plastic bags until analysed.
Fig. 1

Sampling site, Anelema quarry

Production of mixtures with different binders

Because the same ash (the composition may vary over the course of days/weeks) could not be used in producing pellets with different binders during the study period of 3 weeks at the heating plant, a special test procedure was carried out in the laboratory. 5 g of ash (stored in double plastic bags in the laboratory) with known composition were mixed with 5 g (dry weight) of <2 mm sized binder (Anelema dolomite, Glanshammar dolomite or Ignaberga limestone) and 4 g of water. The pastes (a total of three samples) were spread on polystyrene weight boats and covered with watch glasses. Thus, the samples were stored in contact with air but with a cover in order to prevent contamination during hardening in the laboratory. After 4.5 months, the samples were powdered by milling for powder XRD analysis. 5 g of ash (the same ash as above) were mixed with 5 g (dry weight basis, <2 mm) of binder (Anelema, Glanshammar, Ignaberga and Öland) and 4 g of water. Duplicate samples were prepared by the same method and analysed by XRD after 2 weeks of hardening. In this sample series, the Öland limestone was included as well.

Analytical methods

The elemental compositions of ashes, binders and agglomerates were analysed by XRAL Laboratories in Ontario, Canada (ICP80, multi-acid digestion). Water content was determined according to CHM116 and LOI (loss on ignition) at 550 °C for 2 h. S was analysed according to CHM112 (by Leco), B according to ES-4 and Cl according to CHM113 (by ISE). The Ca concentration was analysed according to ICAY50. Anelema and Glanshammar dolomites and the Ignaberga limestone were analysed for tot-N and S by SGAB Analytica, Sweden. S-analysis was performed according to EPA method 200.7 (modified). Determination of N-tot was carried out by Analycen, Sweden (not accredited). The Swedish Geotechnical Institute, Linköping, Sweden determined the grain density and the grain size distribution of the Anelema dolomite by sieving and sedimentation. The pH and electrical conductivity (EC) were measured after agitating samples (dry weight basis) in deionised water and sedimentation or filtering (specified in the discussion) with a 0.45-µm membrane filter. The pH was measured according to Swedish standard SS 028122 and the EC in the supernatant/filtrate liquid according to Swedish standard SIS 028123 (since 1994 SS-EN 27 888).

The sample set, including dolomites, limestones, pellets and hardened laboratory mixtures was milled for 2 min with a Rocklabs Ring Mill. The powdered samples were analysed using a Siemens D5005 X-ray Diffractometer using Cu Kα radiation. Four pellet samples were also analysed by SGAB Analytica AB, Sweden, for phases exceeding 2% (Philips XRD, Cu Kα). Thin section slides of several binders and one granule were studied using a Leica DM RXP microscope with polarised light.

Serial batch leaching test

We used a serial batch leaching test as a tool for determination and for making comparisons of the properties of ash materials. Samples were fractionated into the size intervals 0.25–0.5 mm and 0.5–1 mm using nylon mesh, and are designated F (fine) and C (coarse) below. Each of the treatments was repeated twice. 1 g sub-samples of pellets and granules were leached by 200 ml of leachant, deionised water with H2SO4 (pH 4), in six consecutive steps in order to study leaching of K, Ca and Mg (cumulative L/S 200, 400, 600, 800, 1,000 and 1,200). Reaction of the ash with acidified (pH 4) water simulates the long term leaching under natural conditions and is based on the assumption that chemical equilibrium is reached during the test. The L/S ratio is defined as the accumulated amount of leachant in contact with the sample material related to the amount of dry matter (weight/weight). The results are presented as specific leached amounts as a function of L/S. The Ca and Mg concentrations in leachates were analysed by flame AAS (Perkin Elmer 4100), according to Swedish standard SS 028161. The K concentrations were determined by flame photometry. The pH and the EC were determined directly in the supernatant liquid after each leaching step with methods described previously.

Results and discussion

Ash characterisation

Fly ashes extracted from electric filters are generally fine grained compared to bottom ashes. The grain size character of the fly ash from Kalmar does not vary very much and, in general, 95% of the grains are less than 0.5 mm and between 15 and 30% is less than 0.125 mm in size. The analytical data in Fig. 2 illustrates the wide variety in composition of the ash during a sampling period of 3 weeks. The K and the Mn contents vary substantially. Ca, K, Mg, S, Mn and P are the principal constituents, agreeing with earlier results (Holmberg 2000).
Fig. 2

Variations in the major element concentrations, the moisture content and the content of unburned matter in the ash, measured as LOI (Loss on Ignition), in fly ash from the central heating plant of Kalmar, Sweden. The sampling occasions (1–12) covered a period of 3 weeks. Element concentrations are in mg kg−1, H2O and LOI are given as %

There are negative correlations between the LOI and Mg, Mn, P, and Sr and positive correlations between the LOI and Zr, Si and Al (Fig. 3). When the proportion of unburned matter increases in the ash, the contents of Mg, Mn, P, Ni and Sr proportionally decrease and the contents of Si, Al and Zr proportionally increase. As this ash is intended for nutrient recycling, it can be stated that the quality, in terms of nutrient content, decreases with increasing content of unburned matter. In addition, there is a strong correlation between the LOI and the moisture content of the ash (Figs. 2 and 3), indicating that the wetter the ash the less complete the combustion is. The amounts of certain elements seem to be strongly related to others in the ash, for example, Na to Cl, Ni to Mn (Fig. 4) and Si to Al (not shown). These correlations indicate that sodium occurs bound to chloride in the ash, and also that aluminium silicates are probably present. Mn oxides are known to have a strong heavy metal absorption capacity (Kabata-Pendias and Pendias 1984), and this may explain the Ni–Mn association. If compared to the new recommended main nutrient levels (updated by the Board of Forestry in 2001, displayed in Table 1), the P concentrations are mostly too low, but in accordance with or slightly lower than common P levels of wood ashes reported in the literature (Lundborg 1998; Korpilahti et al. 1998).
Fig. 3

Correlations between the amounts of unburned matter (expressed as loss on ignition, LOI) and the element concentrations and the H2O content of the fly ash from combustion of sawdust at the heating plant of Kalmar, Sweden

Fig. 4

Relationship between certain elements in the fly ash

Table 1

The Swedish recommendations (SR) of the elemental composition of ashes for nutrient recycling in 1996 and 2001 and the composition of ash from the heating plant of Kalmar in June 2001 (conc. as mg kg−1 dry weight if not otherwise indicated)

Ca %

Mg %

P %

K %

Mn %

Zn

B

Cu

Co

Mo

SR 1996

≥10

≥1.5

≥0.5

≥2

0.5–3

500–6,000

100–500

50–500

2–100

0.3–10

SR 2001

≥12.5

≥2

≥1

≥3

1,000–7,000

≤500

≤400

Ash n=12

>15

3.7–5.35

0.69–1.07

3.19–9.25

1.01–1.78

835–1,400

410–626

189–249

44–58

1–3

PAH

Hg

As

Cd

V

Ni

Pb

Cr

LOI %

137Cs Bq/kg

1996

≤5

≤20

≤30

≤100

≤200

≤200

≤250

10

2001

2

≤3

≤30

≤30

≤70

≤70

≤300

≤100

≤5,000

Ash

n=12

7.2a

23–41b ppb

3–4

13–19

5–9

75–125

57–89

22–50

7.64–24.4

1,070–1,170b

aOne ash sample analysed September 2001

bThree samples analysed

Trace element concentrations also displayed variability (Fig. 5), but mostly within the range of values reported earlier (Holmberg 2000). However, higher contents of As (up to 29 mg kg−1) and Cd (up to 31 mg kg−1) have occasionally been observed previously. The amount of Zn is sometimes insufficient and the B levels are higher than or very close to the upper limit value, 500 mg kg−1. Ni, commonly redistributed from burning of fossil fuels, oversteps the Swedish limit value, 70 mg kg−1, continuously. Ni probably originates from the combustion of low-sulphur, heavy-duty firing oil at the heating plant, which is used to restart the boiler after production stops. The maximum content of micronutrients and heavy metals in the new, national recommendations is set so that the maximum heavy metal input for spruce forests in southern Sweden is not exceeded when the application of ashes is in the order of 3 t of dry matter per hectare. Common Ni concentrations in wood ashes are between 30 and 40 mg kg−1 (Lundborg 1998), <12 mg kg−1 in straw ash (Bertelsen 1998) and 71.9 mg kg−1 in sawdust ash (Holzner 1998).
Fig. 5

Variation in the trace element concentrations of the fly ash during a sampling period of 3 weeks, 12 occasions, concentration as mg kg−1

Results from various methods and L/S ratios used in the literature make a comparison of the results difficult. Suggested determinations of ash stability can be made from their electrical conductivity (National Board of Forestry 2002) or a leaching test (Larsson and Westling 1999). It is desirable that the ashes or products made from ashes dissolve slowly, over a period of 5–25 years in the field, and the preliminary target EC value is 10 mS cm−1 if 3 tons of ashes are going to be spread on 1 ha of forest soil. The EC is determined from a clear solution after shaking ash and water (L/S 4) for 1 h and sedimentation in a sealed vessel for 15–30 h. The electrical conductivity (EC) of the fly ash in our study is very high, 45–62 mS cm−1 (L/S 5, shaking 2 h in deionised water, measurement in the filtrate, n=5, duplicate samples). Untreated fly ash from Nybro (from combustion of the same fuel as at the heating plant of Kalmar) has a conductivity of 188 and 31.9 mS cm−1 at L/S ratios of 2 and 10, respectively (Bjurström 1999). An EC of 28.1–37.6 mS cm−1 was determined by Torbjörn Nilsson (personal comm.) for untreated ashes (L/S 4, shaking 1 h in deionised water, measurement in the filtrate). The very high EC values in our samples indicate the presence of easily soluble salts in the ash. The measurement is done near the liquid surface, in the clear solution. However, some ashes do not sediment properly because unburned particles float on the surface and contaminate the electrode. There is a concentration gradient in the liquid after sedimentation, and the conductivity increases downwards. We suggest EC should be directly measured in the filtrate, instead of the clear solution after up to 25 h sedimentation.

Properties of the binders

The composition and the solubility of a binder may be important for both hardening reactions and for the composition of the final agglomerates. The dissolution of limestone and dolomites depends mainly on their geological origin, the particle size distribution, density, grain shape and porosity. Also temperature may affect the dissolution of carbonates. Porosity affects the water absorption capacity and the moisture content of the material.

The grain density of the <4 mm Anelema dolomite (which is the size fraction used at the heating plant) is 2.83 t/m3. About 70% of the Anelema dolomite sample passes a 1 mm sieve and less than 8% of the particles are greater than 2 mm in size. The silt and clay fraction (<0.06 mm) of the crushed dolomite constitutes 17% of the total and almost 10% is less than 0.006 mm in size. The different contents of SiO2 and Mg within the rock profile at the Anelema quarry indicate changing sedimentation environments. The Mg content is highest in the two uppermost levels. Further, Level 1 contains 1.92% SiO2, the proportion of SiO2 increases downwards and is at a maximum of 7.58% in Level 3.

The EC of the Anelema dolomite (L/S 5) Levels 1–5 varies between 0.23 and 0.39 mS cm−1 (Table 2). An EC of 2.6 mS cm−1 was measured for a mixed sample of Anelema dolomite (Table 3), ten times as high as that for Levels 1–5 separately or for the clay. The EC of the clay cement (from Level 2) is 0.21 mS cm−1, lower than the values for the solid rock. The clay cement is most probably a solid weathering product, reflected in part by Mg, Ca, Cl and S contents that are lower and the contents of relatively immobile elements Si, Al, P, K, Cr, V, Zn, Y, Zr, Ba and Li, which are higher than the concentrations of those in the solid rock. The dominating constituent is SiO2 (40.7%).
Table 2

Composition of dolomite from the Anelema quarry, Estonia. The measuring uncertainty is given for sample from Level 1. The clay cement sample is from Level 2. Concentrations are given as mg kg−1 dry weight if not otherwise indicated

Level 1

Level 2

Level 3

Level 4

Level 5

Clay cement

Be

<0.5

<0.5

<0.5

<0.5

<0.5

2.1

Na %

0.02±0.0

0.02

0.02

0.02

0.02

0.08

Mg %

12.2±0.6

11.2

10.8

10.8

10.9

4.43

Al %

0.21±0.0

0.59

0.61

0.52

0.42

5.96

P %

0.02±0.0

0.05

0.02

0.01

<0.01

0.32

K %

0.16±0.01

0.47

0.54

0.41

0.34

3.63

Ca %

>15

>15

>15

>15

>15

6.52

Sc

0.7±0.1

1.4

1.3

1.1

1

8.2

Ti %

0.01±0.0

0.03

0.03

0.03

0.02

0.34

V

8±1

17

6

6

4

119

Cr

6±1

7

27

9

10

103

Mn

291±14

305

261

305

381

142

Fe %

0.21±0.02

0.31

0.46

0.46

0.42

1.78

Co

3±1

3

3

3

1

5

Ni

2±1

4

6

3

4

18

Cu

0.7±0.1

1.7

4.1

2.3

2.1

4.7

Zn

3.7±0.6

10.4

4.2

4

2.8

66.8

Sr

69.6±4

62.1

56.9

49.2

44.3

74.8

Y

4.8±0.2

7.2

5.8

6.4

6.0

28.8

Zr

5.7±0.5

15.3

20

14.2

12.4

118

Ba

13±1

35

60

32

32

278

La

3.3±0.2

7.0

4.9

5.5

4.3

44

Pb

<2

<2

<2

4

3

<2

Li

3±1

4

4

3

2

37

S

<100

300

1300

1500

400

400

SiO2 %

1.92±0.15

5.62

7.58

5.09

4.72

40.7

LOI %

45.7±0.1

43.2

42.3

43.5

44.1

16.9

Cl

582±9

456

455

500

500

238

H2O %

0.1±0.1

0.2

0.3

0.3

0.2

0.3

pH

9.61

9.62

9.38

9.31

9.44

9.04

EC mS cm−1

0.291

0.240

0.307

0.394

0.234

0.213

The following metals were analysed but the concentrations in all samples were below detection limits (as mg kg−1): As (<3), Mo (<1), Ag (<0.2), Sn (<10), Cd (<1), Sb (<5), W (<10) and Bi (<5)

Table 

3 Compositions of dolostones from Glanshammar (Sweden) and Anelema (Estonia) and Swedish limestones from Ignaberga and Öland. Concentrations are given as mg kg−1 (dry weight basis) if not otherwise indicated. Calcium concentrations (not shown) are all >15%

Be

Na %

Mg %

Al %

P %

K %

Sc

Ti %

Anelema

<0.5

0.05

11

1.21

0.03

0.94

2.1

0.06

Glanshammar

<0.5

0.01

11.2

0.55

<0.01

0.21

0.6

<0.01

Ignaberga

<0.5

0.09

0.24

0.46

0.06

0.44

<0.5

0.07

Öland

0.6

0.09

0.60

1.34

0.04

0.68

3.0

0.09

V

Cr

Mn

Fe %

Co

Ni

Cu

Zn

Anelema

23

30

295

0.59

3

6

5.2

14.9

Glanshammar

<2

5

732

0.46

1

3

2.2

32.3

Ignaberga

6

26

253

0.27

2

6

2.2

8.9

Öland

15

17

1150

0.96

20

11

12.2

21.1

Sr

Y

Zr

Mo

Ag

Sn

Ba

La

Anelema

68.8

7.9

31.2

<1

<0.2

<10

86

8.8

Glanshammar

28.6

4.1

14.5

<1

0.3

<10

13

4.2

Ignaberga

185

4.5

19.6

1

<0.2

12

82

4.4

Öland

209

9.7

18.4

1

0.5

<10

102

8.4

Pb

Li

B

H2O %

LOI %

SiO2 %

S

Cl

Anelema

<2

7

40

0.4

7.92

10.0

1840

1420

Glanshammar

<2

7

Na

0.1

41.9

6.51

137

Na

Ignaberga

<2

<1

Na

0.2

34.1

20.2

288

Na

Öland

11

9

Na

4.8

40.7

8.25

Na

Na

pH

Conductivity mS cm−1

N-tot %

Anelema

8.9

2.64

0.01

Glanshammar

9.8

0.181

<0.01

Ignaberga

9.2

0.109

<0.01

Öland

8.9

1.58

Na

Na=not analysed. The following metals were analysed but the concentrations in all samples were below detection limits (as ppm): As (<3), Cd (<1), Sb (<5), W (<10) and Bi (<5)

Compositions of all the binders analysed within this work are displayed in Table 3. Note that the crushed Anelema dolomite (<4 mm) analysed contained material from all the Levels (1–5) and the clay.

Comparing compositions of different levels of the rock, the clay and “the mixed sample”, it is apparent that the clay significantly contributes to the final composition of the fraction that is used at the heating plant, mainly in terms of Si, Al and Fe (Tables 2 and 3). Further, the clay cement probably increases the content of fines in the fraction used in Kalmar. The “mixed sample dolomite” contains much more Cl and S (consistent with its high EC) than do Levels 1–5 above or the clay. The high S and Cl contents in the mixed sample indicate that these elements are somehow added during processing of the dolomite at the quarry. The very low LOI of the mixed sample is difficult to explain and may be due to analytical error. An unexpected finding is that traces of N are found in the Anelema dolomite (0.01%).

The composition of the Glanshammar dolomite differs from that of the Anelema dolomite mainly in the lower contents of Al, Ba and K and higher Mn content (Table 3). The EC of the Glanshammar dolomite is 0.18 mS cm−1, indicating that it is less soluble than the dolomite from Anelema. The mineral phases identified by XRD in both dolomite rocks are dolomite and quartz. Additionally, muscovite [KAl2(AlSi3O10)(OH)2] and a trace of vermiculite (Mg11Al5FeSi11O42·40H2O) are observed in the clay cement sample from Level 2 at Anelema.

Comparing the composition of the limestones from Öland and Ignaberga, it is apparent that the Mg, Al, Fe and Mn concentrations are higher and the SiO2 content lower in the Öland limestone. The high Fe content is indicated because of its brownish-red colour. The Öland limestone has an EC of 1.6 mS cm−1, almost as high as the mixed sample from Anelema. The Ignaberga limestone has an EC of 0.11 mS cm−1, indicating that it could be a much less soluble type of carbonate.

Examination of carbonates by microscopy

Dolomite from five levels in the Anelema quarry was examined by microscopy (Fig. 6). Level 1 is yellowish. The grains are quite spherical, 20–50 µm, and almost of the same size. A slight porosity is observed and the space between the grains is up to about 5 µm. A smaller amount of calcite, quartz and orthoclase are present. Iron precipitation is visible as a dark haze (or a broad stripe) on the left side of the image. This precipitation can be related to the weathering of ankerite, a Fe-rich variety of dolomite. Ankerite is physically and chemically very similar to dolomite. Its colour, however, ranges from yellow-brown to brown. During a metamorphic phase, the Fe-rich ankerite could transform into limonite or hematite.
Fig. 6

Anelema dolomite. Crossed nicols (cross-polarised light) in all magnification photos above. 20× magnification, scale bar 50 µm

The sample from Level 2 is light grey. The dark, brownish stripes in this level can also be associated with ankerite. The stripes, however, seem to be more concentrated and dense in Level 2. Occasional orthoclase and quartz grains are present. The structure of the dolomite seems to be of a slightly higher metamorphic grade in Level 2 compared to Level 1.

Samples collected from Levels 3 and 4 are dark grey. Despite this slight change of colour the dolomite from Level 3 seems to be rather similar to that in Level 2. Contrary to findings in the microscopy study, the SiO2 content, chemically analysed, is higher in Level 3 than in Level 2.

In Level 3, ankerite is present as orientated oblong structures but the porosity of the dolomite has decreased. In Level 3 there is more Fe-rich precipitation and also more ankerite crystals. Samples from Level 4 contain even more orthoclase and ankerite than Level 3 and it has a very dense texture. Chemical analysis supports the microscopy results in the sense that the highest Fe concentrations were found in Level 4. According to chemical analyses Level 4 also has the highest S concentration and EC value.

Common sulphur minerals in marine evaporates are anhydrite (CaSO4), gypsum (CaSO4·2H2O), langbeinite (K2Mg2(SO4)3), polyhalite (K2Ca2Mg(SO4)3·2H2O) and kieserite (MgSO4·H2O). No crystals were found in the microscopy study. Another S-containing mineral, pyrite (FeS2), has been observed in the Anelema dolomite elsewhere (Raukas and Teedumäe 1997; Teedumäe et al. 1997). The sample from the very bottom, Level 5, is slightly brownish. The Fe-rich precipitation is only rarely visible. The grain size is a bit smaller than in the above levels and the porosity is low.

In polarised light microscopy photos (2.5 times magnification) of the Cretaceous Ignaberga and Ordovician Öland limestones (Fig. 7), calcareous fossils such as shells, clams and different algae or parts of them are visible. The Ignaberga limestone is composed of grains up to 1,600 µm in size and the space between the grains are up to 400 µm. This cretaceous limestone has a high porosity. The Ordovician limestone from Öland is very well crystallised with a low porosity. The Glanshammar dolomite is a strongly metamorphosed dolomite, i.e. a marble. It is composed of crystals of similar size of about 200 µm. There are only few inter-grain spaces. The material is very compact and very few micro fissures can be seen in the microscope. The Anelema dolomite (Fig. 7) is extremely dense. It is fine grained, with a rather high inter-crystal porosity, and very homogeneous compared to the other materials studied.
Fig. 7

Thin sections showing different carbonate textures and compositions. Crossed nicols (cross-polarised light) in all magnification photos above, 2.5× magnification, scale bar 400 µm

Characteristics of pellets

Pellets with or without binders were experimentally produced by the roll-pelletiser at the heating plant of Kalmar. One pellet batch per day was produced. The chemical composition of the pellets varied due to the ash as well as the binder compositions (Table 4).
Table 4

Composition of pellets produced at the central heating plant of Kalmar in June 2001. Concentrations are given as mg kg−1 dry weight if not otherwise indicated. A50=pellets with 50% Anelema dolomite, G=Glanshammar dolomite, I=Ignaberga limestone

A50 29/5

A50 5/6

A50 6/6

A50 7/6

A50 8/6

a33 11/6

A33 12/6

ash100 14/6

a33 15/6

G50 18/6

I50 20/6

Be

<0.5

<0.5

<0.5

<0.5

<0.5

<0.5

<0.5

<0.5

<0.5

<0.5

<0.5

Na %

0.41

0.28

0.32

0.32

0.27

0.30

0.31

0.47

0.31

0.28

0.29

Mg %

8.70

7.55

7.31

7.33

7.59

6.42

7.04

5.28

6.79

8.19

2.89

Al %

0.76

0.68

0.81

0.90

0.85

0.68

0.64

0.44

0.64

0.55

0.45

P %

0.57

0.44

0.40

0.42

0.39

0.48

0.64

1.04

0.57

0.45

0.44

K %

3.17

7.16

6.15

5.34

4.73

6.41

9.51

12.9

8.29

6.26

5.49

Ca %

>15

>15

>15

>15

>15

>15

>15

>15

>15

>15

>15

Sc

1.2

1.1

1.2

1.2

1.3

1.0

0.9

<0.5

1.0

<0.5

<0.5

Ti %

0.06

0.04

0.05

0.05

0.05

0.05

0.04

0.03

0.04

0.02

0.04

V

11

10

13

15

15

11

11

6

12

3

6

Cr

27

41

52

43

52

34

29

33

40

33

64

Mn

7430

5720

5330

5340

5470

6500

10520

15730

8540

6350

6070

Fe %

0.51

0.44

0.45

0.46

0.47

0.37

0.36

0.21

0.39

0.28

0.24

Co

25

20

23

25

25

27

27

44

26

23

26

Ni

46

37

36

35

35

38

57

89

49

34

38

Cu

99.5

101

95.7

93.7

101

129

126

178

106

87.9

93.1

Zn

559

518

477

462

432

596

795

1130

690

453

420

As

<3

<3

<3

<3

<3

<3

<3

<3

<3

<3

<3

Sr

649

533

481

503

650

791

826

1220

746

629

673

Y

5.3

4.9

5.1

5.2

21.5

4.1

4.0

2.0

4.0

3.2

21.8

Zr

21.1

18.9

20.0

5.1

24.2

14.7

31.6

10.0

13.8

13.5

14.7

Mo

<1

1

1

2

1

2

1

2

1

<1

1

Ag

1.8

1.8

2.1

1.6

1.5

2.7

2.6

3.9

2.5

1.8

1.2

Cd

8

6

7

6

6

8

9

15

8

7

7

Sn

<10

<10

<10

<10

<10

<10

<10

<10

<10

<10

26

Sb

<5

<5

<5

<5

<5

<5

<5

<5

<5

<5

<5

Ba

1030

881

843

817

789

828

537

267

731

869

882

La

7.4

7.0

6.9

7.3

7.3

6.9

6.1

4.0

6.1

5.0

6.5

W

<10

<10

<10

<10

<10

<10

<10

<10

<10

<10

<10

Pb

41

33

32

30

30

42

38

58

39

28

26

Bi

<5

<5

<5

<5

<5

<5

<5

<5

<5

<5

<5

Li

10

8

9

10

9

9

10

12

9

9

7

SiO2 %

8.99

9.25

9.29

10.2

9.99

9.29

8.79

8.17

8.30

7.85

10.1

LOI %

21.5

19.2

21.3

18.9

18.0

17.1

12.1

9.66

17.3

16.5

9.42

H2O %

2.0

0.9

1.0

1.1

1.2

1.4

0.9

1.5

1.2

0.8

0.8

pH (L/S 200)

Na

Na

Na

11.5

Na

11.8

Na

11.7

Na

11.5

11.3

pH (L/S 4)

Na

Na

Na

12.8

Na

Na

Na

13.2C, 13.2F

Na

12.9C, 13.0F

12.8 C, F

EC mS cm−1

(L/S 200)

Na

Na

Na

1.58C, 1.78F

Na

2.83C, 3.02F

Na

2.99C, 3.01F

Na

1.66C, 1.61F

1.33C, 1.32F

EC mS cm−1

(L/S 4)

Na

Na

Na

29.5C, 29.1F

Na

Na

Na

66.5C

66.5F

Na

38.8C, 39.1F

32.4C, 33.3F

C=coarse particle size fraction. F=fine fraction. Na=not analysed

The particle size distributions of pellets varied with different binders (Fig. 8). Pellets produced without a binder contain much fewer fine particles than the other pellets and more than 70% of grains are coarser than 4 mm. Ideally, less than 5% of the material should be larger than 3 mm in order to not cause damage to the forest vegetation (e.g. bark of the trees) during spreading. At the same time, the content of particles smaller than 0.25 mm should be less than 30%. Small-sized particles may increase dust problems and dissolve rapidly causing negative effects on the soil ecosystem.
Fig. 8

Particle size distribution of pelletised ash

The amount of unburned matter (expressed as LOI) affected the proportion of elements in the ash, but the correlations are not as strong as for the ash alone. The proportion of Al in the pelletised material increases with increasing LOI (Fig. 9). Further, there are correlations between the contents of Al and Fe.
Fig. 9

Relationships between the proportion of unburned matter, Al and Si contents of pellets and granules

The ash mixtures containing dolomite as binder were pelletised without problems. The production of ~250 kg pellets took about 10 min to carry out (Svantesson and Olsson 2002). It is interesting to note that the chemical composition of the pellets varies with particle size in some cases. There is a significant difference in the composition of coarse (0.5–1 mm) and fine (0.25–0.5 mm) pellets with the Ignaberga limestone and in the pellets made without binder. In pelletised materials, the Ba concentrations are up to 800–900 mg kg−1, 2–7 times higher than the concentration in the ash. Ba is enriched during pellet production is probably a contaminant from the equipment. Bjurström (1999) reports the Ba content of untreated ash is 1,210 mg kg−1, 1,510 mg kg−1 for the same ash pelletised and 1,920 mg kg−1 for self-hardened ash. Ba is a trace element that belongs to the alkaline earths and behaves similarly to Ca and Mg. Common Ba concentrations in limestone and dolomites are 50–200 mg kg−1 (Kabata-Pendias and Pendias 1984). Ba is commonly present in plants, but it is not an essential component and is seldom toxic to plants.

Compared to the other kind of pellets studied in this work, the pellets without binder are the product that best fulfil the nutrient requirements. However, Ni and B concentrations are too high, making this kind of pellet hazardous to use as forest soil fertiliser. The main crystalline phases in this pellet type are arcanite, calcite, quartz and portlandite.

The pellets with 50% Anelema dolomite contain 9–26% grains that are smaller than 0.5 mm in size (Fig. 8) and it is anticipated that the content of fines will increase during packaging and transport of the pellets before spreading them on forest soils. Minerals present are dolomite, calcite, quartz, arcanite and portlandite. When pouring water onto the dry mixture of Anelema dolomite and ash, a smell of ammonia sometimes comes off (Thomas Svantesson, personal communication). This phenomenon does not occur with the other binders.

The overall LOI of pellets varied between 9.4 and 21.5% and the moisture content between 0.8 and 2.9% (Table 4). If wood ash is mixed with ashes from combustion of other kind of fuels, or if binder is added, the final composition must also be evaluated against environmental regulations. The P and Zn concentrations are about half of the required minimum levels in the pellets with 50% Anelema dolomite. Zn is weakly bound to the organic matter and is easily leached out by acidification. There is a risk of Zn deficiency in acidified forest soils and sufficient amounts of Zn should be returned with ashes. Therefore, it is suggested that Zn, and probably also P, should be added to the ash/binder mixture.

The Ignaberga limestone produced the following grain size distributions in the pellets: approximately 60% were larger than 4 mm and about 10% less than 0.5 mm. Production of the pellets was problematic because the wet ash/limestone mixture adhered to the surface of the equipment, the mixture consisted of small grains and the material did not hold together. The amount of water required to achieve a suitable consistence was more than for any of the other mixtures produced (Table 4). P and Zn contents in this pellet type are also inadequate compared to the recommendations. Minerals calcite, dolomite, quartz and arcanite are recognised with XRD in the pellets with the Ignaberga limestone.

Pellets with the Glanshammar dolomite were produced without problems. The grain size distributions were close to the ones with the Ignaberga limestone. Dolomite, calcite and arcanite are the mineral phases present.

The granules (made in a rotating drum) have less than 10% particles that are smaller than 0.5 mm or larger than 4 mm in size (Lindahl and Claesson 1996). The granules analysed in this work were produced during two different periods in 1998 and 1999. Granules No 1 contain twice as much unburned matter than Granules No 2. An earlier study showed that the granules are very durable in both a storing and a spreading test and changes of only a few percent were measured in the particle size classes (Lindqvist 1999). Compared to spherical granules, the pellets are non-regularly shaped and sharp-edged. It should be investigated if the shape of the material affects its resistance to breakdown.

Electrical conductivity (EC) of the agglomerates

After shaking the granules and pellets (L/S 4) for 2 h and after 24 h sedimentation, the EC is significantly lower compared to the untreated ash, except for the pellets without binder (Table 4), which have a higher EC than the original untreated ash. Pellets and granules with binder have a significantly lower EC compared to the original ash and to the pellets made without binder (because there is now half as much ash present), but the EC is 3–4 times as high as the preliminary target value. After shaking the granules and pellets (L/S 4) for 1 h and after 24 h sedimentation and filtering, similar results were obtained (Fig. 10). This time, additional particle size fractions were included. This suggests that dolomite lowers the EC values of the agglomerates, but does not particularly prevent the release of readily soluble species from the ash fraction. Hence, it is evident that high EC values are related to the high K content of the fly ash (Fig. 11).
Fig. 10

The EC of granulated and pelleted ash in different L/S ratios. Shaking 1 h, measurement in the filtrate

Fig. 11

Relationship between the K content of the ash and the EC

However, the pellets with 50% Anelema dolomite have the lowest EC values among the pelletised or granulated materials. The EC of the pelletised fly ash from Nybro (Bjurström 1999) was 147 and 26.2 mS cm−1 at L/S ratios of 2 and 10, respectively. The K content of those pellets is also high, almost 11%.

In addition to the sample:water weight ratio, the EC and the pH are dependent on the time of shaking; granulated ash (made without binder) shaken for 1 h and for 8 h had EC (L/S 4) values of 3.60 and 7.29 mS cm−1, respectively (Eriksson et al. 1998). The content of alkali metals was low in those granules. After shaking wood ash granules (L/S 9) for 1 h followed by 30 min sedimentation, the following result was obtained (Börjesson 1992): 4.8 mS cm−1 (granules with 4.5% cement as binder), 4.4 mS cm−1 (granules without binder), and 2.8 mS cm−1 (granules with 6% molass and 2% Na-lignosulphate as binder). These were compared to the EC of 26 untreated wood fly ash samples that had a mean EC of 18.6 mS cm−1. The pH, ANC and EC were also measured during the leaching test with L/S 200–1,200. Comparing pellets and granules with Anelema dolomite it is evident that the pellets containing only 33% dolomite result in more reactive material in terms of EC, pH and ANC. Granules with a greater amount of unburned matter (No 1) had lower pH than the ones with lower amounts of unburned matter (Fig. 12a). After the first leaching step, pellets with 33% (coarse and fine) and pellets with 50% Anelema dolomite had the highest pH, and granules No 1 (fine) the lowest. After the last step (L/S 1,200) granules No 2 (coarse) had the highest pH and pellets with 33% (coarse) the lowest. The acid neutralising capacity (ANC) is highest for the pellets with 33% dolomite and lowest for granules (No 1) after the first step (Fig. 12b). After the last step the ANC is still highest for the pellets with 33% dolomite and lowest for the granules No 2 (fine) and pellets with 50% (coarse). The EC values are highest for pellets with 33% (fine) and lowest for granules No 1 (coarse) after the first leaching step. After the last step, the EC values are between 0.06 and 0.07 mS cm−1 for all samples (Fig. 12c). The pH, ANC and EC of other pellets are displayed in Fig. 13. It is evident that the pellets without binder have the highest pH after the first leaching step, which remains relatively high until the sixth and last step. Pellets with Ignaberga have the lowest pH and ANC values. Pellets with Glanshammar seem to be least soluble in terms of EC in the first leaching step, and after the last step, all have approximately same EC, between 0.06 and 0.09 mS cm−1. However, the pellets without binder have the highest EC values through the whole leaching procedure.
Fig. 12

Granules and pellets with the Anelema dolomite as binder. a pH. b Acid neutralising capacity (ANC) mmol l−1. c Electrical conductivity (EC) mS cm−1

Fig. 13

ANC, pH and the conductivity during the leaching test a pH. b ANC. c EC

Leaching of nutrients

Most Ca is released during the first two leaching steps with an accumulated L/S ratio of 200 (Fig. 14). Thereafter, the leaching rate slows down. The solubility of CaCO3 and Ca(OH)2 is very low at pH levels higher than 12. Between 2 and 3% of the total Ca is released from the granules (No 2), and the pellets with the Ignaberga limestone and the pellets without binder in the first step. The pellets with the Glanshammar dolomite release 5–5.5% of their total Ca in the first leaching step. Between 8 and 10% of the total Ca is released from granules (No 1) and from the pellets with the Anelema dolomite.
Fig. 14

Accumulated percentage of Ca leached from pelleted and granulated ashes as a function of cumulative liquid to solid ratio

After the sixth leaching step, only 7–8% of the total Ca is released from the pellets with Ignaberga limestone, about 10% from pellets without binder and 11–13% from the rest of the materials, except the granules (No 1) which release the most Ca of all, 16–17%, and pellets with 33% Anelema dolomite which release 28–30%. Granules (No 1) pellets with Glanshammar dolomite and the pellets without binder released more Ca from large particles than from the smaller ones.
Fig. 15

Relationship between leached K and leached Ca for pellets and granules

Potassium is largely or completely released from the agglomerates upon first contact with water, during the first leaching step with L/S 200 (Fig. 17c). According to Baxter et al. (1998), between 80 and 90% of the total K occurs in the form of mobile (water-soluble or ion exchangeable) species in the biomass, contrasting with about 5% of the K in high rank coals and about 50% of the K in low rank coals. Potassium evidently occurs in the form of mobile species only in the sawdust ash from Kalmar. The granules and pellets with the Anelema dolomite released 90–100% of their K and the rest of the materials, about 110%. Abbas (2002) explained the differences in the total contents and leached amounts to be caused by the digestion methods of the materials. K is believed to be lost during the melting of filter ash (from the combustion of municipal solid waste) with borate at 1,000 °C. However, the samples in this work were prepared by multi-acid digestion, and it is more likely that the K in the ash was incompletely dissolved. Obernberger et al. (1996) indicate that the high K content might have been the reason for the incomplete dissolution of K during digestion of a straw fly ash due to insufficient amounts of digestion solvents added, yielding low K values (a Round Robin on biomass fuel and ash analysis).

Earlier batch leaching tests (Holmberg 2000) with L/S ratios from 16 to 64 also showed rapid leaching of K from the granules. After the first leaching step with L/S ratio 16, about 10% of the K was removed, and after the fourth step, 15% of the K was lost. Compared to decomposition of the granules at a field study site, the loss of K was approximately 60% after seven months (Holmberg 2000). About 50–60% of the original amounts of K and Na had been lost from self-hardened CFB and GF boiler ashes after 1.5–2.5 years after spreading (Steenari et al. 1998). Of the remaining alkali metal species, only a few per cent (0–10%) were available for leaching, because they were present in phases with low solubility such as feldspars, KAlSi3O8. The sand used as bed material in the fluid bed may contain both quartz and feldspar and contributes to the total ash content of the combustion residue.

XRD analysis shows that in the pellets K occurs in the form of arcanite, K2SO4, also consistent with the positive correlation between the contents of K and S (r=0.893). K can also be bound in compounds such as syngenite, with dissolution rates much lower than, for example, sylvite (KCl). The granules and the pellets released more K from fine-sized particles, except the pellets without binder. Pellets with Ignaberga limestone released similar amounts of K from both sizes. There is a negative correlation between the release of Ca and K from the agglomerates (Fig. 15). However, the K release is very rapid from all types of agglomerated pellets. Leaching of K from an ash is comparable to decomposition of organic material (e.g. leaves) on the forest soil. Organic debris on forest soil releases its K very rapidly. About 80% of the K is leached out from Fraxinus excelsior (ash) leaves if they are soaked in water for 24 h (Troedsson and Nykvist 1973). In the literature it is stated that it would also be desirable for the K to be leached slowly from ashes in order to hinder rapid leaching of K from the soil profile. However, Levula et al. (2000) state that K was the only nutrient that increased significantly in the berries of lingonberry, Vaccinium vitis-ideae, after ash fertilisation with high doses (5 tons ha−1) of bark ash. It was stated that because K is a highly mobile nutrient, it moves rapidly from the soil into the plants.
Fig. 16

Mg (mg kg−1) released from the pellets and the granules as a function of L/S ratios

Mg is released very slowly from the agglomerates (Fig. 16, Fig. 17b). More Mg is released from the small particles, except the pellets with Anelema. More Mg is released from the pellets with 33% Anelema than from those with 50%. After six leaching steps the pellets without binder released approximately 14% of their Mg (Fig. 17b). The granules released least Mg; only 3–5% of the total is lost during six leaching steps. The pellets with the Glanshammar dolomite released 6–9% and the ones with the Ignaberga limestone, 9–12%. After the final step, the granules and the pellets with the Glanshammar dolomite had released more Mg from large particles than smaller ones.
Fig. 17

Total percentage of nutrients leached from granules and pellets (L/S 1,200). a Calcium. b Magnesium. c Potassium

Are there differences in the properties of agglomerates within the binders used? The differences within binders are very small compared to the variations of the ash itself. The drawback of using binder is that the P and Zn contents are diluted. Since dolomites and limestone contain low levels of these elements, they should be added to the mixture during agglomeration. However, the element release rates are slower from agglomerates with binder than from those without a binder. The pellets were produced during experimental conditions with the new equipment at the heating plant of Kalmar. The next step should be to investigate possibilities to refine the compaction grade of the pellets in order to decrease the element release rates.

Conclusions

Combustion of sawdust at the central heating plant of Kalmar results in a very reactive residue composed of <0.5 mm fly ash particles. The main constituents of the ash are Ca, K, Mg, S, Mn and P. The ash has a conductivity of 45–65 mS cm−1, and its composition varies within the amount of unburned matter, expressed as Loss on Ignition (LOI). The proportion of Mg, P, Ni, Zr and Sr decreases and the proportion of Si, Zr and Al increases when the proportion of LOI increases. The Ni content of the ash is strongly correlated to its Mn content, probably due to the heavy metal affinity with Mn oxide. Further, Na is correlated to Cl and Si to Al, indicating the presence of NaCl and aluminium silicates in the ash. Compared to the compositional requirements for ash recycling it is evident that the Ni and B contents exceed the maximum allowed concentrations. Therefore, it is important to use a binder that makes a positive contribution to the composition and properties of the final agglomerated material.

Dolomite from Anelema, Estonia, is a very fine-grained dolomite. Its composition varies at the point of extraction, the most significant changes are in SiO2 content. The presence of ankerite and a decrease of the inter-grain porosity down the rock profile are documented. The clay cement present in the rock contributes Si, Al, and trace metals to the composition of the final <4 mm fraction that is used as binder at the heating plant. Additionally, it is probable that the clay cement increases the amount of fine material. The Anelema dolomite is high in S, compared to the Swedish dolomite, and it also contains trace amounts of N.

The dolomite from Glanshammar, Sweden, is more crystalline and less soluble than the dolomite from Estonia. It contains much less Al, K and Ba and twice as much Mn as does the Estonian dolomite. The automated agglomeration process worked well with the dolomites, but the Ignaberga limestone required a lot of water and the material adhered to the surfaces of the equipment. The limestone from Ignaberga is low in Mg, Al, Fe and Mn contents, and high in SiO2 content. It is less soluble compared with the limestone from Öland, which is composed of smaller grains and has a low inter-grain porosity. The pellets with the Ignaberga agglomerate released most Mg and least K and Ca of the materials studied.

Both granules and pellets have insufficient P and Zn concentrations. The stability of all the agglomerates is high compared to the recommendations. The lowest EC, 30 mS cm−1, is measured in granules made with Anelema dolomite as binder. The mineralogical analysis of granules and pellets showed the presence of dolomite, calcite, arcanite and quartz. In the Ignaberga and Glanshammar agglomerates, the calcite forming dominates, but in the pellets with Anelema dolomite more arcanite than calcite is present. K is almost completely or completely leached out from the granules and the pellets in the leaching test. Pellets without binder released nutrients K, Ca and Mg most rapidly, and have the highest solubility in terms of conductivity. Pellets with 33% dolomite release more Ca, Mg and K than the ones with 50% dolomite. Pelletised materials are more reactive than granulated materials. With increasing amounts of binder and also unburned matter (measured as LOI) the reactivity of agglomerates decreases.

Acknowledgements

The financial support from the KK Foundation and Graninge Kalmar Energi AB is acknowledged. Thanks to Mr Per Brändström for help in the laboratory and to Mr Arvo Suppi for hospitality during our visit in Estonia.

Copyright information

© Springer-Verlag 2004